Adding upstream version 6.6.15.upstream/6.6.15

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

17 files changed, 6087 insertions, 0 deletions

diff --git a/drivers/hwtracing/intel_th/Kconfig b/drivers/hwtracing/intel_th/Kconfig
new file mode 100644
index 0000000000..4b6359326e
--- /dev/null
+++ b/drivers/hwtracing/intel_th/Kconfig
@@ -0,0 +1,86 @@
+# SPDX-License-Identifier: GPL-2.0-only
+config INTEL_TH
+	tristate "Intel(R) Trace Hub controller"
+	depends on HAS_DMA && HAS_IOMEM
+	help
+	  Intel(R) Trace Hub (TH) is a set of hardware blocks (subdevices) that
+	  produce, switch and output trace data from multiple hardware and
+	  software sources over several types of trace output ports encoded
+	  in System Trace Protocol (MIPI STPv2) and is intended to perform
+	  full system debugging.
+
+	  This option enables intel_th bus and common code used by TH
+	  subdevices to interact with each other and hardware and for
+	  platform glue layers to drive Intel TH devices.
+
+	  Say Y here to enable Intel(R) Trace Hub controller support.
+
+if INTEL_TH
+
+config INTEL_TH_PCI
+	tristate "Intel(R) Trace Hub PCI controller"
+	depends on PCI
+	help
+	  Intel(R) Trace Hub may exist as a PCI device. This option enables
+	  support glue layer for PCI-based Intel TH.
+
+	  Say Y here to enable PCI Intel TH support.
+
+config INTEL_TH_ACPI
+	tristate "Intel(R) Trace Hub ACPI controller"
+	depends on ACPI
+	help
+	  Intel(R) Trace Hub may exist as an ACPI device. This option enables
+	  support glue layer for ACPI-based Intel TH. This typically implies
+	  'host debugger' mode, that is, the trace configuration and capture
+	  is handled by an external debug host and corresponding controls will
+	  not be available on the target.
+
+	  Say Y here to enable ACPI Intel TH support.
+
+config INTEL_TH_GTH
+	tristate "Intel(R) Trace Hub Global Trace Hub"
+	help
+	  Global Trace Hub (GTH) is the central component of the
+	  Intel TH infrastructure and acts as a switch for source
+	  and output devices. This driver is required for other
+	  Intel TH subdevices to initialize.
+
+	  Say Y here to enable GTH subdevice of Intel(R) Trace Hub.
+
+config INTEL_TH_STH
+	tristate "Intel(R) Trace Hub Software Trace Hub support"
+	depends on STM
+	help
+	  Software Trace Hub (STH) enables trace data from software
+	  trace sources to be sent out via Intel(R) Trace Hub. It
+	  uses stm class device to interface with its sources.
+
+	  Say Y here to enable STH subdevice of Intel(R) Trace Hub.
+
+config INTEL_TH_MSU
+	tristate "Intel(R) Trace Hub Memory Storage Unit"
+	help
+	  Memory Storage Unit (MSU) trace output device enables
+	  storing STP traces to system memory. It supports single
+	  and multiblock modes of operation and provides read()
+	  and mmap() access to the collected data.
+
+	  Say Y here to enable MSU output device for Intel TH.
+
+config INTEL_TH_PTI
+	tristate "Intel(R) Trace Hub PTI output"
+	help
+	  Parallel Trace Interface unit (PTI) is a trace output device
+	  of Intel TH architecture that facilitates STP trace output via
+	  a PTI port.
+
+	  Say Y to enable PTI output of Intel TH data.
+
+config INTEL_TH_DEBUG
+	bool "Intel(R) Trace Hub debugging"
+	depends on DEBUG_FS
+	help
+	  Say Y here to enable debugging.
+
+endif
diff --git a/drivers/hwtracing/intel_th/Makefile b/drivers/hwtracing/intel_th/Makefile
new file mode 100644
index 0000000000..b63eb8f309
--- /dev/null
+++ b/drivers/hwtracing/intel_th/Makefile
@@ -0,0 +1,25 @@
+# SPDX-License-Identifier: GPL-2.0
+obj-$(CONFIG_INTEL_TH)		+= intel_th.o
+intel_th-y			:= core.o
+intel_th-$(CONFIG_INTEL_TH_DEBUG) += debug.o
+
+obj-$(CONFIG_INTEL_TH_PCI)	+= intel_th_pci.o
+intel_th_pci-y			:= pci.o
+
+obj-$(CONFIG_INTEL_TH_ACPI)	+= intel_th_acpi.o
+intel_th_acpi-y			:= acpi.o
+
+obj-$(CONFIG_INTEL_TH_GTH)	+= intel_th_gth.o
+intel_th_gth-y			:= gth.o
+
+obj-$(CONFIG_INTEL_TH_STH)	+= intel_th_sth.o
+intel_th_sth-y			:= sth.o
+
+obj-$(CONFIG_INTEL_TH_MSU)	+= intel_th_msu.o
+intel_th_msu-y			:= msu.o
+
+obj-$(CONFIG_INTEL_TH_PTI)	+= intel_th_pti.o
+intel_th_pti-y			:= pti.o
+
+obj-$(CONFIG_INTEL_TH_MSU)	+= intel_th_msu_sink.o
+intel_th_msu_sink-y		:= msu-sink.o
diff --git a/drivers/hwtracing/intel_th/acpi.c b/drivers/hwtracing/intel_th/acpi.c
new file mode 100644
index 0000000000..87f9024e4b
--- /dev/null
+++ b/drivers/hwtracing/intel_th/acpi.c
@@ -0,0 +1,85 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Intel(R) Trace Hub ACPI driver
+ *
+ * Copyright (C) 2017 Intel Corporation.
+ */
+
+#define pr_fmt(fmt)	KBUILD_MODNAME ": " fmt
+
+#include <linux/types.h>
+#include <linux/module.h>
+#include <linux/device.h>
+#include <linux/sysfs.h>
+#include <linux/platform_device.h>
+#include <linux/acpi.h>
+
+#include "intel_th.h"
+
+#define DRIVER_NAME "intel_th_acpi"
+
+static const struct intel_th_drvdata intel_th_acpi_pch = {
+	.host_mode_only	= 1,
+};
+
+static const struct intel_th_drvdata intel_th_acpi_uncore = {
+	.host_mode_only	= 1,
+};
+
+static const struct acpi_device_id intel_th_acpi_ids[] = {
+	{ "INTC1000",	(kernel_ulong_t)&intel_th_acpi_uncore },
+	{ "INTC1001",	(kernel_ulong_t)&intel_th_acpi_pch },
+	{ "",		0 },
+};
+
+MODULE_DEVICE_TABLE(acpi, intel_th_acpi_ids);
+
+static int intel_th_acpi_probe(struct platform_device *pdev)
+{
+	struct acpi_device *adev = ACPI_COMPANION(&pdev->dev);
+	struct resource resource[TH_MMIO_END];
+	const struct acpi_device_id *id;
+	struct intel_th *th;
+	int i, r;
+
+	id = acpi_match_device(intel_th_acpi_ids, &pdev->dev);
+	if (!id)
+		return -ENODEV;
+
+	for (i = 0, r = 0; i < pdev->num_resources && r < TH_MMIO_END; i++)
+		if (pdev->resource[i].flags &
+		    (IORESOURCE_IRQ | IORESOURCE_MEM))
+			resource[r++] = pdev->resource[i];
+
+	th = intel_th_alloc(&pdev->dev, (void *)id->driver_data, resource, r);
+	if (IS_ERR(th))
+		return PTR_ERR(th);
+
+	adev->driver_data = th;
+
+	return 0;
+}
+
+static int intel_th_acpi_remove(struct platform_device *pdev)
+{
+	struct intel_th *th = platform_get_drvdata(pdev);
+
+	intel_th_free(th);
+
+	return 0;
+}
+
+static struct platform_driver intel_th_acpi_driver = {
+	.probe		= intel_th_acpi_probe,
+	.remove		= intel_th_acpi_remove,
+	.driver		= {
+		.name			= DRIVER_NAME,
+		.acpi_match_table	= intel_th_acpi_ids,
+	},
+};
+
+module_platform_driver(intel_th_acpi_driver);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("Intel(R) Trace Hub ACPI controller driver");
+MODULE_AUTHOR("Alexander Shishkin <alexander.shishkin@intel.com>");
diff --git a/drivers/hwtracing/intel_th/core.c b/drivers/hwtracing/intel_th/core.c
new file mode 100644
index 0000000000..cc7f879bb1
--- /dev/null
+++ b/drivers/hwtracing/intel_th/core.c
@@ -0,0 +1,1086 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Intel(R) Trace Hub driver core
+ *
+ * Copyright (C) 2014-2015 Intel Corporation.
+ */
+
+#define pr_fmt(fmt)	KBUILD_MODNAME ": " fmt
+
+#include <linux/types.h>
+#include <linux/module.h>
+#include <linux/device.h>
+#include <linux/sysfs.h>
+#include <linux/kdev_t.h>
+#include <linux/debugfs.h>
+#include <linux/idr.h>
+#include <linux/pci.h>
+#include <linux/pm_runtime.h>
+#include <linux/dma-mapping.h>
+
+#include "intel_th.h"
+#include "debug.h"
+
+static bool host_mode __read_mostly;
+module_param(host_mode, bool, 0444);
+
+static DEFINE_IDA(intel_th_ida);
+
+static int intel_th_match(struct device *dev, struct device_driver *driver)
+{
+	struct intel_th_driver *thdrv = to_intel_th_driver(driver);
+	struct intel_th_device *thdev = to_intel_th_device(dev);
+
+	if (thdev->type == INTEL_TH_SWITCH &&
+	    (!thdrv->enable || !thdrv->disable))
+		return 0;
+
+	return !strcmp(thdev->name, driver->name);
+}
+
+static int intel_th_child_remove(struct device *dev, void *data)
+{
+	device_release_driver(dev);
+
+	return 0;
+}
+
+static int intel_th_probe(struct device *dev)
+{
+	struct intel_th_driver *thdrv = to_intel_th_driver(dev->driver);
+	struct intel_th_device *thdev = to_intel_th_device(dev);
+	struct intel_th_driver *hubdrv;
+	struct intel_th_device *hub = NULL;
+	int ret;
+
+	if (thdev->type == INTEL_TH_SWITCH)
+		hub = thdev;
+	else if (dev->parent)
+		hub = to_intel_th_device(dev->parent);
+
+	if (!hub || !hub->dev.driver)
+		return -EPROBE_DEFER;
+
+	hubdrv = to_intel_th_driver(hub->dev.driver);
+
+	pm_runtime_set_active(dev);
+	pm_runtime_no_callbacks(dev);
+	pm_runtime_enable(dev);
+
+	ret = thdrv->probe(to_intel_th_device(dev));
+	if (ret)
+		goto out_pm;
+
+	if (thdrv->attr_group) {
+		ret = sysfs_create_group(&thdev->dev.kobj, thdrv->attr_group);
+		if (ret)
+			goto out;
+	}
+
+	if (thdev->type == INTEL_TH_OUTPUT &&
+	    !intel_th_output_assigned(thdev))
+		/* does not talk to hardware */
+		ret = hubdrv->assign(hub, thdev);
+
+out:
+	if (ret)
+		thdrv->remove(thdev);
+
+out_pm:
+	if (ret)
+		pm_runtime_disable(dev);
+
+	return ret;
+}
+
+static void intel_th_device_remove(struct intel_th_device *thdev);
+
+static void intel_th_remove(struct device *dev)
+{
+	struct intel_th_driver *thdrv = to_intel_th_driver(dev->driver);
+	struct intel_th_device *thdev = to_intel_th_device(dev);
+	struct intel_th_device *hub = to_intel_th_hub(thdev);
+
+	if (thdev->type == INTEL_TH_SWITCH) {
+		struct intel_th *th = to_intel_th(hub);
+		int i, lowest;
+
+		/*
+		 * disconnect outputs
+		 *
+		 * intel_th_child_remove returns 0 unconditionally, so there is
+		 * no need to check the return value of device_for_each_child.
+		 */
+		device_for_each_child(dev, thdev, intel_th_child_remove);
+
+		/*
+		 * Remove outputs, that is, hub's children: they are created
+		 * at hub's probe time by having the hub call
+		 * intel_th_output_enable() for each of them.
+		 */
+		for (i = 0, lowest = -1; i < th->num_thdevs; i++) {
+			/*
+			 * Move the non-output devices from higher up the
+			 * th->thdev[] array to lower positions to maintain
+			 * a contiguous array.
+			 */
+			if (th->thdev[i]->type != INTEL_TH_OUTPUT) {
+				if (lowest >= 0) {
+					th->thdev[lowest] = th->thdev[i];
+					th->thdev[i] = NULL;
+					++lowest;
+				}
+
+				continue;
+			}
+
+			if (lowest == -1)
+				lowest = i;
+
+			intel_th_device_remove(th->thdev[i]);
+			th->thdev[i] = NULL;
+		}
+
+		if (lowest >= 0)
+			th->num_thdevs = lowest;
+	}
+
+	if (thdrv->attr_group)
+		sysfs_remove_group(&thdev->dev.kobj, thdrv->attr_group);
+
+	pm_runtime_get_sync(dev);
+
+	thdrv->remove(thdev);
+
+	if (intel_th_output_assigned(thdev)) {
+		struct intel_th_driver *hubdrv =
+			to_intel_th_driver(dev->parent->driver);
+
+		if (hub->dev.driver)
+			/* does not talk to hardware */
+			hubdrv->unassign(hub, thdev);
+	}
+
+	pm_runtime_disable(dev);
+	pm_runtime_set_active(dev);
+	pm_runtime_enable(dev);
+}
+
+static struct bus_type intel_th_bus = {
+	.name		= "intel_th",
+	.match		= intel_th_match,
+	.probe		= intel_th_probe,
+	.remove		= intel_th_remove,
+};
+
+static void intel_th_device_free(struct intel_th_device *thdev);
+
+static void intel_th_device_release(struct device *dev)
+{
+	intel_th_device_free(to_intel_th_device(dev));
+}
+
+static struct device_type intel_th_source_device_type = {
+	.name		= "intel_th_source_device",
+	.release	= intel_th_device_release,
+};
+
+static char *intel_th_output_devnode(const struct device *dev, umode_t *mode,
+				     kuid_t *uid, kgid_t *gid)
+{
+	const struct intel_th_device *thdev = to_intel_th_device(dev);
+	const struct intel_th *th = to_intel_th(thdev);
+	char *node;
+
+	if (thdev->id >= 0)
+		node = kasprintf(GFP_KERNEL, "intel_th%d/%s%d", th->id,
+				 thdev->name, thdev->id);
+	else
+		node = kasprintf(GFP_KERNEL, "intel_th%d/%s", th->id,
+				 thdev->name);
+
+	return node;
+}
+
+static ssize_t port_show(struct device *dev, struct device_attribute *attr,
+			 char *buf)
+{
+	struct intel_th_device *thdev = to_intel_th_device(dev);
+
+	if (thdev->output.port >= 0)
+		return scnprintf(buf, PAGE_SIZE, "%u\n", thdev->output.port);
+
+	return scnprintf(buf, PAGE_SIZE, "unassigned\n");
+}
+
+static DEVICE_ATTR_RO(port);
+
+static void intel_th_trace_prepare(struct intel_th_device *thdev)
+{
+	struct intel_th_device *hub = to_intel_th_hub(thdev);
+	struct intel_th_driver *hubdrv = to_intel_th_driver(hub->dev.driver);
+
+	if (hub->type != INTEL_TH_SWITCH)
+		return;
+
+	if (thdev->type != INTEL_TH_OUTPUT)
+		return;
+
+	pm_runtime_get_sync(&thdev->dev);
+	hubdrv->prepare(hub, &thdev->output);
+	pm_runtime_put(&thdev->dev);
+}
+
+static int intel_th_output_activate(struct intel_th_device *thdev)
+{
+	struct intel_th_driver *thdrv =
+		to_intel_th_driver_or_null(thdev->dev.driver);
+	struct intel_th *th = to_intel_th(thdev);
+	int ret = 0;
+
+	if (!thdrv)
+		return -ENODEV;
+
+	if (!try_module_get(thdrv->driver.owner))
+		return -ENODEV;
+
+	pm_runtime_get_sync(&thdev->dev);
+
+	if (th->activate)
+		ret = th->activate(th);
+	if (ret)
+		goto fail_put;
+
+	intel_th_trace_prepare(thdev);
+	if (thdrv->activate)
+		ret = thdrv->activate(thdev);
+	else
+		intel_th_trace_enable(thdev);
+
+	if (ret)
+		goto fail_deactivate;
+
+	return 0;
+
+fail_deactivate:
+	if (th->deactivate)
+		th->deactivate(th);
+
+fail_put:
+	pm_runtime_put(&thdev->dev);
+	module_put(thdrv->driver.owner);
+
+	return ret;
+}
+
+static void intel_th_output_deactivate(struct intel_th_device *thdev)
+{
+	struct intel_th_driver *thdrv =
+		to_intel_th_driver_or_null(thdev->dev.driver);
+	struct intel_th *th = to_intel_th(thdev);
+
+	if (!thdrv)
+		return;
+
+	if (thdrv->deactivate)
+		thdrv->deactivate(thdev);
+	else
+		intel_th_trace_disable(thdev);
+
+	if (th->deactivate)
+		th->deactivate(th);
+
+	pm_runtime_put(&thdev->dev);
+	module_put(thdrv->driver.owner);
+}
+
+static ssize_t active_show(struct device *dev, struct device_attribute *attr,
+			   char *buf)
+{
+	struct intel_th_device *thdev = to_intel_th_device(dev);
+
+	return scnprintf(buf, PAGE_SIZE, "%d\n", thdev->output.active);
+}
+
+static ssize_t active_store(struct device *dev, struct device_attribute *attr,
+			    const char *buf, size_t size)
+{
+	struct intel_th_device *thdev = to_intel_th_device(dev);
+	unsigned long val;
+	int ret;
+
+	ret = kstrtoul(buf, 10, &val);
+	if (ret)
+		return ret;
+
+	if (!!val != thdev->output.active) {
+		if (val)
+			ret = intel_th_output_activate(thdev);
+		else
+			intel_th_output_deactivate(thdev);
+	}
+
+	return ret ? ret : size;
+}
+
+static DEVICE_ATTR_RW(active);
+
+static struct attribute *intel_th_output_attrs[] = {
+	&dev_attr_port.attr,
+	&dev_attr_active.attr,
+	NULL,
+};
+
+ATTRIBUTE_GROUPS(intel_th_output);
+
+static struct device_type intel_th_output_device_type = {
+	.name		= "intel_th_output_device",
+	.groups		= intel_th_output_groups,
+	.release	= intel_th_device_release,
+	.devnode	= intel_th_output_devnode,
+};
+
+static struct device_type intel_th_switch_device_type = {
+	.name		= "intel_th_switch_device",
+	.release	= intel_th_device_release,
+};
+
+static struct device_type *intel_th_device_type[] = {
+	[INTEL_TH_SOURCE]	= &intel_th_source_device_type,
+	[INTEL_TH_OUTPUT]	= &intel_th_output_device_type,
+	[INTEL_TH_SWITCH]	= &intel_th_switch_device_type,
+};
+
+int intel_th_driver_register(struct intel_th_driver *thdrv)
+{
+	if (!thdrv->probe || !thdrv->remove)
+		return -EINVAL;
+
+	thdrv->driver.bus = &intel_th_bus;
+
+	return driver_register(&thdrv->driver);
+}
+EXPORT_SYMBOL_GPL(intel_th_driver_register);
+
+void intel_th_driver_unregister(struct intel_th_driver *thdrv)
+{
+	driver_unregister(&thdrv->driver);
+}
+EXPORT_SYMBOL_GPL(intel_th_driver_unregister);
+
+static struct intel_th_device *
+intel_th_device_alloc(struct intel_th *th, unsigned int type, const char *name,
+		      int id)
+{
+	struct device *parent;
+	struct intel_th_device *thdev;
+
+	if (type == INTEL_TH_OUTPUT)
+		parent = &th->hub->dev;
+	else
+		parent = th->dev;
+
+	thdev = kzalloc(sizeof(*thdev) + strlen(name) + 1, GFP_KERNEL);
+	if (!thdev)
+		return NULL;
+
+	thdev->id = id;
+	thdev->type = type;
+
+	strcpy(thdev->name, name);
+	device_initialize(&thdev->dev);
+	thdev->dev.bus = &intel_th_bus;
+	thdev->dev.type = intel_th_device_type[type];
+	thdev->dev.parent = parent;
+	thdev->dev.dma_mask = parent->dma_mask;
+	thdev->dev.dma_parms = parent->dma_parms;
+	dma_set_coherent_mask(&thdev->dev, parent->coherent_dma_mask);
+	if (id >= 0)
+		dev_set_name(&thdev->dev, "%d-%s%d", th->id, name, id);
+	else
+		dev_set_name(&thdev->dev, "%d-%s", th->id, name);
+
+	return thdev;
+}
+
+static int intel_th_device_add_resources(struct intel_th_device *thdev,
+					 struct resource *res, int nres)
+{
+	struct resource *r;
+
+	r = kmemdup(res, sizeof(*res) * nres, GFP_KERNEL);
+	if (!r)
+		return -ENOMEM;
+
+	thdev->resource = r;
+	thdev->num_resources = nres;
+
+	return 0;
+}
+
+static void intel_th_device_remove(struct intel_th_device *thdev)
+{
+	device_del(&thdev->dev);
+	put_device(&thdev->dev);
+}
+
+static void intel_th_device_free(struct intel_th_device *thdev)
+{
+	kfree(thdev->resource);
+	kfree(thdev);
+}
+
+/*
+ * Intel(R) Trace Hub subdevices
+ */
+static const struct intel_th_subdevice {
+	const char		*name;
+	struct resource		res[3];
+	unsigned		nres;
+	unsigned		type;
+	unsigned		otype;
+	bool			mknode;
+	unsigned		scrpd;
+	int			id;
+} intel_th_subdevices[] = {
+	{
+		.nres	= 1,
+		.res	= {
+			{
+				/* Handle TSCU and CTS from GTH driver */
+				.start	= REG_GTH_OFFSET,
+				.end	= REG_CTS_OFFSET + REG_CTS_LENGTH - 1,
+				.flags	= IORESOURCE_MEM,
+			},
+		},
+		.name	= "gth",
+		.type	= INTEL_TH_SWITCH,
+		.id	= -1,
+	},
+	{
+		.nres	= 2,
+		.res	= {
+			{
+				.start	= REG_MSU_OFFSET,
+				.end	= REG_MSU_OFFSET + REG_MSU_LENGTH - 1,
+				.flags	= IORESOURCE_MEM,
+			},
+			{
+				.start	= BUF_MSU_OFFSET,
+				.end	= BUF_MSU_OFFSET + BUF_MSU_LENGTH - 1,
+				.flags	= IORESOURCE_MEM,
+			},
+		},
+		.name	= "msc",
+		.id	= 0,
+		.type	= INTEL_TH_OUTPUT,
+		.mknode	= true,
+		.otype	= GTH_MSU,
+		.scrpd	= SCRPD_MEM_IS_PRIM_DEST | SCRPD_MSC0_IS_ENABLED,
+	},
+	{
+		.nres	= 2,
+		.res	= {
+			{
+				.start	= REG_MSU_OFFSET,
+				.end	= REG_MSU_OFFSET + REG_MSU_LENGTH - 1,
+				.flags	= IORESOURCE_MEM,
+			},
+			{
+				.start	= BUF_MSU_OFFSET,
+				.end	= BUF_MSU_OFFSET + BUF_MSU_LENGTH - 1,
+				.flags	= IORESOURCE_MEM,
+			},
+		},
+		.name	= "msc",
+		.id	= 1,
+		.type	= INTEL_TH_OUTPUT,
+		.mknode	= true,
+		.otype	= GTH_MSU,
+		.scrpd	= SCRPD_MEM_IS_PRIM_DEST | SCRPD_MSC1_IS_ENABLED,
+	},
+	{
+		.nres	= 2,
+		.res	= {
+			{
+				.start	= REG_STH_OFFSET,
+				.end	= REG_STH_OFFSET + REG_STH_LENGTH - 1,
+				.flags	= IORESOURCE_MEM,
+			},
+			{
+				.start	= TH_MMIO_SW,
+				.end	= 0,
+				.flags	= IORESOURCE_MEM,
+			},
+		},
+		.id	= -1,
+		.name	= "sth",
+		.type	= INTEL_TH_SOURCE,
+	},
+	{
+		.nres	= 2,
+		.res	= {
+			{
+				.start	= REG_STH_OFFSET,
+				.end	= REG_STH_OFFSET + REG_STH_LENGTH - 1,
+				.flags	= IORESOURCE_MEM,
+			},
+			{
+				.start	= TH_MMIO_RTIT,
+				.end	= 0,
+				.flags	= IORESOURCE_MEM,
+			},
+		},
+		.id	= -1,
+		.name	= "rtit",
+		.type	= INTEL_TH_SOURCE,
+	},
+	{
+		.nres	= 1,
+		.res	= {
+			{
+				.start	= REG_PTI_OFFSET,
+				.end	= REG_PTI_OFFSET + REG_PTI_LENGTH - 1,
+				.flags	= IORESOURCE_MEM,
+			},
+		},
+		.id	= -1,
+		.name	= "pti",
+		.type	= INTEL_TH_OUTPUT,
+		.otype	= GTH_PTI,
+		.scrpd	= SCRPD_PTI_IS_PRIM_DEST,
+	},
+	{
+		.nres	= 1,
+		.res	= {
+			{
+				.start	= REG_PTI_OFFSET,
+				.end	= REG_PTI_OFFSET + REG_PTI_LENGTH - 1,
+				.flags	= IORESOURCE_MEM,
+			},
+		},
+		.id	= -1,
+		.name	= "lpp",
+		.type	= INTEL_TH_OUTPUT,
+		.otype	= GTH_LPP,
+		.scrpd	= SCRPD_PTI_IS_PRIM_DEST,
+	},
+	{
+		.nres	= 1,
+		.res	= {
+			{
+				.start	= REG_DCIH_OFFSET,
+				.end	= REG_DCIH_OFFSET + REG_DCIH_LENGTH - 1,
+				.flags	= IORESOURCE_MEM,
+			},
+		},
+		.id	= -1,
+		.name	= "dcih",
+		.type	= INTEL_TH_OUTPUT,
+	},
+};
+
+#ifdef CONFIG_MODULES
+static void __intel_th_request_hub_module(struct work_struct *work)
+{
+	struct intel_th *th = container_of(work, struct intel_th,
+					   request_module_work);
+
+	request_module("intel_th_%s", th->hub->name);
+}
+
+static int intel_th_request_hub_module(struct intel_th *th)
+{
+	INIT_WORK(&th->request_module_work, __intel_th_request_hub_module);
+	schedule_work(&th->request_module_work);
+
+	return 0;
+}
+
+static void intel_th_request_hub_module_flush(struct intel_th *th)
+{
+	flush_work(&th->request_module_work);
+}
+#else
+static inline int intel_th_request_hub_module(struct intel_th *th)
+{
+	return -EINVAL;
+}
+
+static inline void intel_th_request_hub_module_flush(struct intel_th *th)
+{
+}
+#endif /* CONFIG_MODULES */
+
+static struct intel_th_device *
+intel_th_subdevice_alloc(struct intel_th *th,
+			 const struct intel_th_subdevice *subdev)
+{
+	struct intel_th_device *thdev;
+	struct resource res[3];
+	unsigned int req = 0;
+	int r, err;
+
+	thdev = intel_th_device_alloc(th, subdev->type, subdev->name,
+				      subdev->id);
+	if (!thdev)
+		return ERR_PTR(-ENOMEM);
+
+	thdev->drvdata = th->drvdata;
+
+	memcpy(res, subdev->res,
+	       sizeof(struct resource) * subdev->nres);
+
+	for (r = 0; r < subdev->nres; r++) {
+		struct resource *devres = th->resource;
+		int bar = TH_MMIO_CONFIG;
+
+		/*
+		 * Take .end == 0 to mean 'take the whole bar',
+		 * .start then tells us which bar it is. Default to
+		 * TH_MMIO_CONFIG.
+		 */
+		if (!res[r].end && res[r].flags == IORESOURCE_MEM) {
+			bar = res[r].start;
+			err = -ENODEV;
+			if (bar >= th->num_resources)
+				goto fail_put_device;
+			res[r].start = 0;
+			res[r].end = resource_size(&devres[bar]) - 1;
+		}
+
+		if (res[r].flags & IORESOURCE_MEM) {
+			res[r].start	+= devres[bar].start;
+			res[r].end	+= devres[bar].start;
+
+			dev_dbg(th->dev, "%s:%d @ %pR\n",
+				subdev->name, r, &res[r]);
+		} else if (res[r].flags & IORESOURCE_IRQ) {
+			/*
+			 * Only pass on the IRQ if we have useful interrupts:
+			 * the ones that can be configured via MINTCTL.
+			 */
+			if (INTEL_TH_CAP(th, has_mintctl) && th->irq != -1)
+				res[r].start = th->irq;
+		}
+	}
+
+	err = intel_th_device_add_resources(thdev, res, subdev->nres);
+	if (err)
+		goto fail_put_device;
+
+	if (subdev->type == INTEL_TH_OUTPUT) {
+		if (subdev->mknode)
+			thdev->dev.devt = MKDEV(th->major, th->num_thdevs);
+		thdev->output.type = subdev->otype;
+		thdev->output.port = -1;
+		thdev->output.scratchpad = subdev->scrpd;
+	} else if (subdev->type == INTEL_TH_SWITCH) {
+		thdev->host_mode =
+			INTEL_TH_CAP(th, host_mode_only) ? true : host_mode;
+		th->hub = thdev;
+	}
+
+	err = device_add(&thdev->dev);
+	if (err)
+		goto fail_free_res;
+
+	/* need switch driver to be loaded to enumerate the rest */
+	if (subdev->type == INTEL_TH_SWITCH && !req) {
+		err = intel_th_request_hub_module(th);
+		if (!err)
+			req++;
+	}
+
+	return thdev;
+
+fail_free_res:
+	kfree(thdev->resource);
+
+fail_put_device:
+	put_device(&thdev->dev);
+
+	return ERR_PTR(err);
+}
+
+/**
+ * intel_th_output_enable() - find and enable a device for a given output type
+ * @th:		Intel TH instance
+ * @otype:	output type
+ *
+ * Go through the unallocated output devices, find the first one whos type
+ * matches @otype and instantiate it. These devices are removed when the hub
+ * device is removed, see intel_th_remove().
+ */
+int intel_th_output_enable(struct intel_th *th, unsigned int otype)
+{
+	struct intel_th_device *thdev;
+	int src = 0, dst = 0;
+
+	for (src = 0, dst = 0; dst <= th->num_thdevs; src++, dst++) {
+		for (; src < ARRAY_SIZE(intel_th_subdevices); src++) {
+			if (intel_th_subdevices[src].type != INTEL_TH_OUTPUT)
+				continue;
+
+			if (intel_th_subdevices[src].otype != otype)
+				continue;
+
+			break;
+		}
+
+		/* no unallocated matching subdevices */
+		if (src == ARRAY_SIZE(intel_th_subdevices))
+			return -ENODEV;
+
+		for (; dst < th->num_thdevs; dst++) {
+			if (th->thdev[dst]->type != INTEL_TH_OUTPUT)
+				continue;
+
+			if (th->thdev[dst]->output.type != otype)
+				continue;
+
+			break;
+		}
+
+		/*
+		 * intel_th_subdevices[src] matches our requirements and is
+		 * not matched in th::thdev[]
+		 */
+		if (dst == th->num_thdevs)
+			goto found;
+	}
+
+	return -ENODEV;
+
+found:
+	thdev = intel_th_subdevice_alloc(th, &intel_th_subdevices[src]);
+	if (IS_ERR(thdev))
+		return PTR_ERR(thdev);
+
+	th->thdev[th->num_thdevs++] = thdev;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(intel_th_output_enable);
+
+static int intel_th_populate(struct intel_th *th)
+{
+	int src;
+
+	/* create devices for each intel_th_subdevice */
+	for (src = 0; src < ARRAY_SIZE(intel_th_subdevices); src++) {
+		const struct intel_th_subdevice *subdev =
+			&intel_th_subdevices[src];
+		struct intel_th_device *thdev;
+
+		/* only allow SOURCE and SWITCH devices in host mode */
+		if ((INTEL_TH_CAP(th, host_mode_only) || host_mode) &&
+		    subdev->type == INTEL_TH_OUTPUT)
+			continue;
+
+		/*
+		 * don't enable port OUTPUTs in this path; SWITCH enables them
+		 * via intel_th_output_enable()
+		 */
+		if (subdev->type == INTEL_TH_OUTPUT &&
+		    subdev->otype != GTH_NONE)
+			continue;
+
+		thdev = intel_th_subdevice_alloc(th, subdev);
+		/* note: caller should free subdevices from th::thdev[] */
+		if (IS_ERR(thdev)) {
+			/* ENODEV for individual subdevices is allowed */
+			if (PTR_ERR(thdev) == -ENODEV)
+				continue;
+
+			return PTR_ERR(thdev);
+		}
+
+		th->thdev[th->num_thdevs++] = thdev;
+	}
+
+	return 0;
+}
+
+static int intel_th_output_open(struct inode *inode, struct file *file)
+{
+	const struct file_operations *fops;
+	struct intel_th_driver *thdrv;
+	struct device *dev;
+	int err;
+
+	dev = bus_find_device_by_devt(&intel_th_bus, inode->i_rdev);
+	if (!dev || !dev->driver)
+		return -ENODEV;
+
+	thdrv = to_intel_th_driver(dev->driver);
+	fops = fops_get(thdrv->fops);
+	if (!fops)
+		return -ENODEV;
+
+	replace_fops(file, fops);
+
+	file->private_data = to_intel_th_device(dev);
+
+	if (file->f_op->open) {
+		err = file->f_op->open(inode, file);
+		return err;
+	}
+
+	return 0;
+}
+
+static const struct file_operations intel_th_output_fops = {
+	.open	= intel_th_output_open,
+	.llseek	= noop_llseek,
+};
+
+static irqreturn_t intel_th_irq(int irq, void *data)
+{
+	struct intel_th *th = data;
+	irqreturn_t ret = IRQ_NONE;
+	struct intel_th_driver *d;
+	int i;
+
+	for (i = 0; i < th->num_thdevs; i++) {
+		if (th->thdev[i]->type != INTEL_TH_OUTPUT)
+			continue;
+
+		d = to_intel_th_driver(th->thdev[i]->dev.driver);
+		if (d && d->irq)
+			ret |= d->irq(th->thdev[i]);
+	}
+
+	return ret;
+}
+
+/**
+ * intel_th_alloc() - allocate a new Intel TH device and its subdevices
+ * @dev:	parent device
+ * @devres:	resources indexed by th_mmio_idx
+ * @irq:	irq number
+ */
+struct intel_th *
+intel_th_alloc(struct device *dev, const struct intel_th_drvdata *drvdata,
+	       struct resource *devres, unsigned int ndevres)
+{
+	int err, r, nr_mmios = 0;
+	struct intel_th *th;
+
+	th = kzalloc(sizeof(*th), GFP_KERNEL);
+	if (!th)
+		return ERR_PTR(-ENOMEM);
+
+	th->id = ida_simple_get(&intel_th_ida, 0, 0, GFP_KERNEL);
+	if (th->id < 0) {
+		err = th->id;
+		goto err_alloc;
+	}
+
+	th->major = __register_chrdev(0, 0, TH_POSSIBLE_OUTPUTS,
+				      "intel_th/output", &intel_th_output_fops);
+	if (th->major < 0) {
+		err = th->major;
+		goto err_ida;
+	}
+	th->irq = -1;
+	th->dev = dev;
+	th->drvdata = drvdata;
+
+	for (r = 0; r < ndevres; r++)
+		switch (devres[r].flags & IORESOURCE_TYPE_BITS) {
+		case IORESOURCE_MEM:
+			th->resource[nr_mmios++] = devres[r];
+			break;
+		case IORESOURCE_IRQ:
+			err = devm_request_irq(dev, devres[r].start,
+					       intel_th_irq, IRQF_SHARED,
+					       dev_name(dev), th);
+			if (err)
+				goto err_chrdev;
+
+			if (th->irq == -1)
+				th->irq = devres[r].start;
+			th->num_irqs++;
+			break;
+		default:
+			dev_warn(dev, "Unknown resource type %lx\n",
+				 devres[r].flags);
+			break;
+		}
+
+	th->num_resources = nr_mmios;
+
+	dev_set_drvdata(dev, th);
+
+	pm_runtime_no_callbacks(dev);
+	pm_runtime_put(dev);
+	pm_runtime_allow(dev);
+
+	err = intel_th_populate(th);
+	if (err) {
+		/* free the subdevices and undo everything */
+		intel_th_free(th);
+		return ERR_PTR(err);
+	}
+
+	return th;
+
+err_chrdev:
+	__unregister_chrdev(th->major, 0, TH_POSSIBLE_OUTPUTS,
+			    "intel_th/output");
+
+err_ida:
+	ida_simple_remove(&intel_th_ida, th->id);
+
+err_alloc:
+	kfree(th);
+
+	return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(intel_th_alloc);
+
+void intel_th_free(struct intel_th *th)
+{
+	int i;
+
+	intel_th_request_hub_module_flush(th);
+
+	intel_th_device_remove(th->hub);
+	for (i = 0; i < th->num_thdevs; i++) {
+		if (th->thdev[i] != th->hub)
+			intel_th_device_remove(th->thdev[i]);
+		th->thdev[i] = NULL;
+	}
+
+	th->num_thdevs = 0;
+
+	for (i = 0; i < th->num_irqs; i++)
+		devm_free_irq(th->dev, th->irq + i, th);
+
+	pm_runtime_get_sync(th->dev);
+	pm_runtime_forbid(th->dev);
+
+	__unregister_chrdev(th->major, 0, TH_POSSIBLE_OUTPUTS,
+			    "intel_th/output");
+
+	ida_simple_remove(&intel_th_ida, th->id);
+
+	kfree(th);
+}
+EXPORT_SYMBOL_GPL(intel_th_free);
+
+/**
+ * intel_th_trace_enable() - enable tracing for an output device
+ * @thdev:	output device that requests tracing be enabled
+ */
+int intel_th_trace_enable(struct intel_th_device *thdev)
+{
+	struct intel_th_device *hub = to_intel_th_device(thdev->dev.parent);
+	struct intel_th_driver *hubdrv = to_intel_th_driver(hub->dev.driver);
+
+	if (WARN_ON_ONCE(hub->type != INTEL_TH_SWITCH))
+		return -EINVAL;
+
+	if (WARN_ON_ONCE(thdev->type != INTEL_TH_OUTPUT))
+		return -EINVAL;
+
+	pm_runtime_get_sync(&thdev->dev);
+	hubdrv->enable(hub, &thdev->output);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(intel_th_trace_enable);
+
+/**
+ * intel_th_trace_switch() - execute a switch sequence
+ * @thdev:	output device that requests tracing switch
+ */
+int intel_th_trace_switch(struct intel_th_device *thdev)
+{
+	struct intel_th_device *hub = to_intel_th_device(thdev->dev.parent);
+	struct intel_th_driver *hubdrv = to_intel_th_driver(hub->dev.driver);
+
+	if (WARN_ON_ONCE(hub->type != INTEL_TH_SWITCH))
+		return -EINVAL;
+
+	if (WARN_ON_ONCE(thdev->type != INTEL_TH_OUTPUT))
+		return -EINVAL;
+
+	hubdrv->trig_switch(hub, &thdev->output);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(intel_th_trace_switch);
+
+/**
+ * intel_th_trace_disable() - disable tracing for an output device
+ * @thdev:	output device that requests tracing be disabled
+ */
+int intel_th_trace_disable(struct intel_th_device *thdev)
+{
+	struct intel_th_device *hub = to_intel_th_device(thdev->dev.parent);
+	struct intel_th_driver *hubdrv = to_intel_th_driver(hub->dev.driver);
+
+	WARN_ON_ONCE(hub->type != INTEL_TH_SWITCH);
+	if (WARN_ON_ONCE(thdev->type != INTEL_TH_OUTPUT))
+		return -EINVAL;
+
+	hubdrv->disable(hub, &thdev->output);
+	pm_runtime_put(&thdev->dev);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(intel_th_trace_disable);
+
+int intel_th_set_output(struct intel_th_device *thdev,
+			unsigned int master)
+{
+	struct intel_th_device *hub = to_intel_th_hub(thdev);
+	struct intel_th_driver *hubdrv = to_intel_th_driver(hub->dev.driver);
+	int ret;
+
+	/* In host mode, this is up to the external debugger, do nothing. */
+	if (hub->host_mode)
+		return 0;
+
+	/*
+	 * hub is instantiated together with the source device that
+	 * calls here, so guaranteed to be present.
+	 */
+	hubdrv = to_intel_th_driver(hub->dev.driver);
+	if (!hubdrv || !try_module_get(hubdrv->driver.owner))
+		return -EINVAL;
+
+	if (!hubdrv->set_output) {
+		ret = -ENOTSUPP;
+		goto out;
+	}
+
+	ret = hubdrv->set_output(hub, master);
+
+out:
+	module_put(hubdrv->driver.owner);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(intel_th_set_output);
+
+static int __init intel_th_init(void)
+{
+	intel_th_debug_init();
+
+	return bus_register(&intel_th_bus);
+}
+subsys_initcall(intel_th_init);
+
+static void __exit intel_th_exit(void)
+{
+	intel_th_debug_done();
+
+	bus_unregister(&intel_th_bus);
+}
+module_exit(intel_th_exit);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("Intel(R) Trace Hub controller driver");
+MODULE_AUTHOR("Alexander Shishkin <alexander.shishkin@linux.intel.com>");
diff --git a/drivers/hwtracing/intel_th/debug.c b/drivers/hwtracing/intel_th/debug.c
new file mode 100644
index 0000000000..ff79063118
--- /dev/null
+++ b/drivers/hwtracing/intel_th/debug.c
@@ -0,0 +1,28 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Intel(R) Trace Hub driver debugging
+ *
+ * Copyright (C) 2014-2015 Intel Corporation.
+ */
+
+#include <linux/types.h>
+#include <linux/device.h>
+#include <linux/debugfs.h>
+
+#include "intel_th.h"
+#include "debug.h"
+
+struct dentry *intel_th_dbg;
+
+void intel_th_debug_init(void)
+{
+	intel_th_dbg = debugfs_create_dir("intel_th", NULL);
+	if (IS_ERR(intel_th_dbg))
+		intel_th_dbg = NULL;
+}
+
+void intel_th_debug_done(void)
+{
+	debugfs_remove(intel_th_dbg);
+	intel_th_dbg = NULL;
+}
diff --git a/drivers/hwtracing/intel_th/debug.h b/drivers/hwtracing/intel_th/debug.h
new file mode 100644
index 0000000000..78bd7e4bf9
--- /dev/null
+++ b/drivers/hwtracing/intel_th/debug.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Intel(R) Trace Hub driver debugging
+ *
+ * Copyright (C) 2014-2015 Intel Corporation.
+ */
+
+#ifndef __INTEL_TH_DEBUG_H__
+#define __INTEL_TH_DEBUG_H__
+
+#ifdef CONFIG_INTEL_TH_DEBUG
+extern struct dentry *intel_th_dbg;
+
+void intel_th_debug_init(void);
+void intel_th_debug_done(void);
+#else
+static inline void intel_th_debug_init(void)
+{
+}
+
+static inline void intel_th_debug_done(void)
+{
+}
+#endif
+
+#endif /* __INTEL_TH_DEBUG_H__ */
diff --git a/drivers/hwtracing/intel_th/gth.c b/drivers/hwtracing/intel_th/gth.c
new file mode 100644
index 0000000000..b3308934a6
--- /dev/null
+++ b/drivers/hwtracing/intel_th/gth.c
@@ -0,0 +1,850 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Intel(R) Trace Hub Global Trace Hub
+ *
+ * Copyright (C) 2014-2015 Intel Corporation.
+ */
+
+#define pr_fmt(fmt)	KBUILD_MODNAME ": " fmt
+
+#include <linux/types.h>
+#include <linux/module.h>
+#include <linux/device.h>
+#include <linux/io.h>
+#include <linux/mm.h>
+#include <linux/slab.h>
+#include <linux/bitmap.h>
+#include <linux/pm_runtime.h>
+
+#include "intel_th.h"
+#include "gth.h"
+
+struct gth_device;
+
+/**
+ * struct gth_output - GTH view on an output port
+ * @gth:	backlink to the GTH device
+ * @output:	link to output device's output descriptor
+ * @index:	output port number
+ * @port_type:	one of GTH_* port type values
+ * @master:	bitmap of masters configured for this output
+ */
+struct gth_output {
+	struct gth_device	*gth;
+	struct intel_th_output	*output;
+	unsigned int		index;
+	unsigned int		port_type;
+	DECLARE_BITMAP(master, TH_CONFIGURABLE_MASTERS + 1);
+};
+
+/**
+ * struct gth_device - GTH device
+ * @dev:	driver core's device
+ * @base:	register window base address
+ * @output_group:	attributes describing output ports
+ * @master_group:	attributes describing master assignments
+ * @output:		output ports
+ * @master:		master/output port assignments
+ * @gth_lock:		serializes accesses to GTH bits
+ */
+struct gth_device {
+	struct device		*dev;
+	void __iomem		*base;
+
+	struct attribute_group	output_group;
+	struct attribute_group	master_group;
+	struct gth_output	output[TH_POSSIBLE_OUTPUTS];
+	signed char		master[TH_CONFIGURABLE_MASTERS + 1];
+	spinlock_t		gth_lock;
+};
+
+static void gth_output_set(struct gth_device *gth, int port,
+			   unsigned int config)
+{
+	unsigned long reg = port & 4 ? REG_GTH_GTHOPT1 : REG_GTH_GTHOPT0;
+	u32 val;
+	int shift = (port & 3) * 8;
+
+	val = ioread32(gth->base + reg);
+	val &= ~(0xff << shift);
+	val |= config << shift;
+	iowrite32(val, gth->base + reg);
+}
+
+static unsigned int gth_output_get(struct gth_device *gth, int port)
+{
+	unsigned long reg = port & 4 ? REG_GTH_GTHOPT1 : REG_GTH_GTHOPT0;
+	u32 val;
+	int shift = (port & 3) * 8;
+
+	val = ioread32(gth->base + reg);
+	val &= 0xff << shift;
+	val >>= shift;
+
+	return val;
+}
+
+static void gth_smcfreq_set(struct gth_device *gth, int port,
+			    unsigned int freq)
+{
+	unsigned long reg = REG_GTH_SMCR0 + ((port / 2) * 4);
+	int shift = (port & 1) * 16;
+	u32 val;
+
+	val = ioread32(gth->base + reg);
+	val &= ~(0xffff << shift);
+	val |= freq << shift;
+	iowrite32(val, gth->base + reg);
+}
+
+static unsigned int gth_smcfreq_get(struct gth_device *gth, int port)
+{
+	unsigned long reg = REG_GTH_SMCR0 + ((port / 2) * 4);
+	int shift = (port & 1) * 16;
+	u32 val;
+
+	val = ioread32(gth->base + reg);
+	val &= 0xffff << shift;
+	val >>= shift;
+
+	return val;
+}
+
+/*
+ * "masters" attribute group
+ */
+
+struct master_attribute {
+	struct device_attribute	attr;
+	struct gth_device	*gth;
+	unsigned int		master;
+};
+
+static void
+gth_master_set(struct gth_device *gth, unsigned int master, int port)
+{
+	unsigned int reg = REG_GTH_SWDEST0 + ((master >> 1) & ~3u);
+	unsigned int shift = (master & 0x7) * 4;
+	u32 val;
+
+	if (master >= 256) {
+		reg = REG_GTH_GSWTDEST;
+		shift = 0;
+	}
+
+	val = ioread32(gth->base + reg);
+	val &= ~(0xf << shift);
+	if (port >= 0)
+		val |= (0x8 | port) << shift;
+	iowrite32(val, gth->base + reg);
+}
+
+static ssize_t master_attr_show(struct device *dev,
+				struct device_attribute *attr,
+				char *buf)
+{
+	struct master_attribute *ma =
+		container_of(attr, struct master_attribute, attr);
+	struct gth_device *gth = ma->gth;
+	size_t count;
+	int port;
+
+	spin_lock(&gth->gth_lock);
+	port = gth->master[ma->master];
+	spin_unlock(&gth->gth_lock);
+
+	if (port >= 0)
+		count = snprintf(buf, PAGE_SIZE, "%x\n", port);
+	else
+		count = snprintf(buf, PAGE_SIZE, "disabled\n");
+
+	return count;
+}
+
+static ssize_t master_attr_store(struct device *dev,
+				 struct device_attribute *attr,
+				 const char *buf, size_t count)
+{
+	struct master_attribute *ma =
+		container_of(attr, struct master_attribute, attr);
+	struct gth_device *gth = ma->gth;
+	int old_port, port;
+
+	if (kstrtoint(buf, 10, &port) < 0)
+		return -EINVAL;
+
+	if (port >= TH_POSSIBLE_OUTPUTS || port < -1)
+		return -EINVAL;
+
+	spin_lock(&gth->gth_lock);
+
+	/* disconnect from the previous output port, if any */
+	old_port = gth->master[ma->master];
+	if (old_port >= 0) {
+		gth->master[ma->master] = -1;
+		clear_bit(ma->master, gth->output[old_port].master);
+
+		/*
+		 * if the port is active, program this setting,
+		 * implies that runtime PM is on
+		 */
+		if (gth->output[old_port].output->active)
+			gth_master_set(gth, ma->master, -1);
+	}
+
+	/* connect to the new output port, if any */
+	if (port >= 0) {
+		/* check if there's a driver for this port */
+		if (!gth->output[port].output) {
+			count = -ENODEV;
+			goto unlock;
+		}
+
+		set_bit(ma->master, gth->output[port].master);
+
+		/* if the port is active, program this setting, see above */
+		if (gth->output[port].output->active)
+			gth_master_set(gth, ma->master, port);
+	}
+
+	gth->master[ma->master] = port;
+
+unlock:
+	spin_unlock(&gth->gth_lock);
+
+	return count;
+}
+
+struct output_attribute {
+	struct device_attribute attr;
+	struct gth_device	*gth;
+	unsigned int		port;
+	unsigned int		parm;
+};
+
+#define OUTPUT_PARM(_name, _mask, _r, _w, _what)			\
+	[TH_OUTPUT_PARM(_name)] = { .name = __stringify(_name),		\
+				    .get = gth_ ## _what ## _get,	\
+				    .set = gth_ ## _what ## _set,	\
+				    .mask = (_mask),			\
+				    .readable = (_r),			\
+				    .writable = (_w) }
+
+static const struct output_parm {
+	const char	*name;
+	unsigned int	(*get)(struct gth_device *gth, int port);
+	void		(*set)(struct gth_device *gth, int port,
+			       unsigned int val);
+	unsigned int	mask;
+	unsigned int	readable : 1,
+			writable : 1;
+} output_parms[] = {
+	OUTPUT_PARM(port,	0x7,	1, 0, output),
+	OUTPUT_PARM(null,	BIT(3),	1, 1, output),
+	OUTPUT_PARM(drop,	BIT(4),	1, 1, output),
+	OUTPUT_PARM(reset,	BIT(5),	1, 0, output),
+	OUTPUT_PARM(flush,	BIT(7),	0, 1, output),
+	OUTPUT_PARM(smcfreq,	0xffff,	1, 1, smcfreq),
+};
+
+static void
+gth_output_parm_set(struct gth_device *gth, int port, unsigned int parm,
+		    unsigned int val)
+{
+	unsigned int config = output_parms[parm].get(gth, port);
+	unsigned int mask = output_parms[parm].mask;
+	unsigned int shift = __ffs(mask);
+
+	config &= ~mask;
+	config |= (val << shift) & mask;
+	output_parms[parm].set(gth, port, config);
+}
+
+static unsigned int
+gth_output_parm_get(struct gth_device *gth, int port, unsigned int parm)
+{
+	unsigned int config = output_parms[parm].get(gth, port);
+	unsigned int mask = output_parms[parm].mask;
+	unsigned int shift = __ffs(mask);
+
+	config &= mask;
+	config >>= shift;
+	return config;
+}
+
+/*
+ * Reset outputs and sources
+ */
+static int intel_th_gth_reset(struct gth_device *gth)
+{
+	u32 reg;
+	int port, i;
+
+	reg = ioread32(gth->base + REG_GTH_SCRPD0);
+	if (reg & SCRPD_DEBUGGER_IN_USE)
+		return -EBUSY;
+
+	/* Always save/restore STH and TU registers in S0ix entry/exit */
+	reg |= SCRPD_STH_IS_ENABLED | SCRPD_TRIGGER_IS_ENABLED;
+	iowrite32(reg, gth->base + REG_GTH_SCRPD0);
+
+	/* output ports */
+	for (port = 0; port < 8; port++) {
+		if (gth_output_parm_get(gth, port, TH_OUTPUT_PARM(port)) ==
+		    GTH_NONE)
+			continue;
+
+		gth_output_set(gth, port, 0);
+		gth_smcfreq_set(gth, port, 16);
+	}
+	/* disable overrides */
+	iowrite32(0, gth->base + REG_GTH_DESTOVR);
+
+	/* masters swdest_0~31 and gswdest */
+	for (i = 0; i < 33; i++)
+		iowrite32(0, gth->base + REG_GTH_SWDEST0 + i * 4);
+
+	/* sources */
+	iowrite32(0, gth->base + REG_GTH_SCR);
+	iowrite32(0xfc, gth->base + REG_GTH_SCR2);
+
+	/* setup CTS for single trigger */
+	iowrite32(CTS_EVENT_ENABLE_IF_ANYTHING, gth->base + REG_CTS_C0S0_EN);
+	iowrite32(CTS_ACTION_CONTROL_SET_STATE(CTS_STATE_IDLE) |
+		  CTS_ACTION_CONTROL_TRIGGER, gth->base + REG_CTS_C0S0_ACT);
+
+	return 0;
+}
+
+/*
+ * "outputs" attribute group
+ */
+
+static ssize_t output_attr_show(struct device *dev,
+				struct device_attribute *attr,
+				char *buf)
+{
+	struct output_attribute *oa =
+		container_of(attr, struct output_attribute, attr);
+	struct gth_device *gth = oa->gth;
+	size_t count;
+
+	pm_runtime_get_sync(dev);
+
+	spin_lock(&gth->gth_lock);
+	count = snprintf(buf, PAGE_SIZE, "%x\n",
+			 gth_output_parm_get(gth, oa->port, oa->parm));
+	spin_unlock(&gth->gth_lock);
+
+	pm_runtime_put(dev);
+
+	return count;
+}
+
+static ssize_t output_attr_store(struct device *dev,
+				 struct device_attribute *attr,
+				 const char *buf, size_t count)
+{
+	struct output_attribute *oa =
+		container_of(attr, struct output_attribute, attr);
+	struct gth_device *gth = oa->gth;
+	unsigned int config;
+
+	if (kstrtouint(buf, 16, &config) < 0)
+		return -EINVAL;
+
+	pm_runtime_get_sync(dev);
+
+	spin_lock(&gth->gth_lock);
+	gth_output_parm_set(gth, oa->port, oa->parm, config);
+	spin_unlock(&gth->gth_lock);
+
+	pm_runtime_put(dev);
+
+	return count;
+}
+
+static int intel_th_master_attributes(struct gth_device *gth)
+{
+	struct master_attribute *master_attrs;
+	struct attribute **attrs;
+	int i, nattrs = TH_CONFIGURABLE_MASTERS + 2;
+
+	attrs = devm_kcalloc(gth->dev, nattrs, sizeof(void *), GFP_KERNEL);
+	if (!attrs)
+		return -ENOMEM;
+
+	master_attrs = devm_kcalloc(gth->dev, nattrs,
+				    sizeof(struct master_attribute),
+				    GFP_KERNEL);
+	if (!master_attrs)
+		return -ENOMEM;
+
+	for (i = 0; i < TH_CONFIGURABLE_MASTERS + 1; i++) {
+		char *name;
+
+		name = devm_kasprintf(gth->dev, GFP_KERNEL, "%d%s", i,
+				      i == TH_CONFIGURABLE_MASTERS ? "+" : "");
+		if (!name)
+			return -ENOMEM;
+
+		master_attrs[i].attr.attr.name = name;
+		master_attrs[i].attr.attr.mode = S_IRUGO | S_IWUSR;
+		master_attrs[i].attr.show = master_attr_show;
+		master_attrs[i].attr.store = master_attr_store;
+
+		sysfs_attr_init(&master_attrs[i].attr.attr);
+		attrs[i] = &master_attrs[i].attr.attr;
+
+		master_attrs[i].gth = gth;
+		master_attrs[i].master = i;
+	}
+
+	gth->master_group.name	= "masters";
+	gth->master_group.attrs = attrs;
+
+	return sysfs_create_group(&gth->dev->kobj, &gth->master_group);
+}
+
+static int intel_th_output_attributes(struct gth_device *gth)
+{
+	struct output_attribute *out_attrs;
+	struct attribute **attrs;
+	int i, j, nouts = TH_POSSIBLE_OUTPUTS;
+	int nparms = ARRAY_SIZE(output_parms);
+	int nattrs = nouts * nparms + 1;
+
+	attrs = devm_kcalloc(gth->dev, nattrs, sizeof(void *), GFP_KERNEL);
+	if (!attrs)
+		return -ENOMEM;
+
+	out_attrs = devm_kcalloc(gth->dev, nattrs,
+				 sizeof(struct output_attribute),
+				 GFP_KERNEL);
+	if (!out_attrs)
+		return -ENOMEM;
+
+	for (i = 0; i < nouts; i++) {
+		for (j = 0; j < nparms; j++) {
+			unsigned int idx = i * nparms + j;
+			char *name;
+
+			name = devm_kasprintf(gth->dev, GFP_KERNEL, "%d_%s", i,
+					      output_parms[j].name);
+			if (!name)
+				return -ENOMEM;
+
+			out_attrs[idx].attr.attr.name = name;
+
+			if (output_parms[j].readable) {
+				out_attrs[idx].attr.attr.mode |= S_IRUGO;
+				out_attrs[idx].attr.show = output_attr_show;
+			}
+
+			if (output_parms[j].writable) {
+				out_attrs[idx].attr.attr.mode |= S_IWUSR;
+				out_attrs[idx].attr.store = output_attr_store;
+			}
+
+			sysfs_attr_init(&out_attrs[idx].attr.attr);
+			attrs[idx] = &out_attrs[idx].attr.attr;
+
+			out_attrs[idx].gth = gth;
+			out_attrs[idx].port = i;
+			out_attrs[idx].parm = j;
+		}
+	}
+
+	gth->output_group.name	= "outputs";
+	gth->output_group.attrs = attrs;
+
+	return sysfs_create_group(&gth->dev->kobj, &gth->output_group);
+}
+
+/**
+ * intel_th_gth_stop() - stop tracing to an output device
+ * @gth:		GTH device
+ * @output:		output device's descriptor
+ * @capture_done:	set when no more traces will be captured
+ *
+ * This will stop tracing using force storeEn off signal and wait for the
+ * pipelines to be empty for the corresponding output port.
+ */
+static void intel_th_gth_stop(struct gth_device *gth,
+			      struct intel_th_output *output,
+			      bool capture_done)
+{
+	struct intel_th_device *outdev =
+		container_of(output, struct intel_th_device, output);
+	struct intel_th_driver *outdrv =
+		to_intel_th_driver(outdev->dev.driver);
+	unsigned long count;
+	u32 reg;
+	u32 scr2 = 0xfc | (capture_done ? 1 : 0);
+
+	iowrite32(0, gth->base + REG_GTH_SCR);
+	iowrite32(scr2, gth->base + REG_GTH_SCR2);
+
+	/* wait on pipeline empty for the given port */
+	for (reg = 0, count = GTH_PLE_WAITLOOP_DEPTH;
+	     count && !(reg & BIT(output->port)); count--) {
+		reg = ioread32(gth->base + REG_GTH_STAT);
+		cpu_relax();
+	}
+
+	if (!count)
+		dev_dbg(gth->dev, "timeout waiting for GTH[%d] PLE\n",
+			output->port);
+
+	/* wait on output piepline empty */
+	if (outdrv->wait_empty)
+		outdrv->wait_empty(outdev);
+
+	/* clear force capture done for next captures */
+	iowrite32(0xfc, gth->base + REG_GTH_SCR2);
+}
+
+/**
+ * intel_th_gth_start() - start tracing to an output device
+ * @gth:	GTH device
+ * @output:	output device's descriptor
+ *
+ * This will start tracing using force storeEn signal.
+ */
+static void intel_th_gth_start(struct gth_device *gth,
+			       struct intel_th_output *output)
+{
+	u32 scr = 0xfc0000;
+
+	if (output->multiblock)
+		scr |= 0xff;
+
+	iowrite32(scr, gth->base + REG_GTH_SCR);
+	iowrite32(0, gth->base + REG_GTH_SCR2);
+}
+
+/**
+ * intel_th_gth_disable() - disable tracing to an output device
+ * @thdev:	GTH device
+ * @output:	output device's descriptor
+ *
+ * This will deconfigure all masters set to output to this device,
+ * disable tracing using force storeEn off signal and wait for the
+ * "pipeline empty" bit for corresponding output port.
+ */
+static void intel_th_gth_disable(struct intel_th_device *thdev,
+				 struct intel_th_output *output)
+{
+	struct gth_device *gth = dev_get_drvdata(&thdev->dev);
+	int master;
+	u32 reg;
+
+	spin_lock(&gth->gth_lock);
+	output->active = false;
+
+	for_each_set_bit(master, gth->output[output->port].master,
+			 TH_CONFIGURABLE_MASTERS + 1) {
+		gth_master_set(gth, master, -1);
+	}
+	spin_unlock(&gth->gth_lock);
+
+	intel_th_gth_stop(gth, output, true);
+
+	reg = ioread32(gth->base + REG_GTH_SCRPD0);
+	reg &= ~output->scratchpad;
+	iowrite32(reg, gth->base + REG_GTH_SCRPD0);
+}
+
+static void gth_tscu_resync(struct gth_device *gth)
+{
+	u32 reg;
+
+	reg = ioread32(gth->base + REG_TSCU_TSUCTRL);
+	reg &= ~TSUCTRL_CTCRESYNC;
+	iowrite32(reg, gth->base + REG_TSCU_TSUCTRL);
+}
+
+static void intel_th_gth_prepare(struct intel_th_device *thdev,
+				 struct intel_th_output *output)
+{
+	struct gth_device *gth = dev_get_drvdata(&thdev->dev);
+	int count;
+
+	/*
+	 * Wait until the output port is in reset before we start
+	 * programming it.
+	 */
+	for (count = GTH_PLE_WAITLOOP_DEPTH;
+	     count && !(gth_output_get(gth, output->port) & BIT(5)); count--)
+		cpu_relax();
+}
+
+/**
+ * intel_th_gth_enable() - enable tracing to an output device
+ * @thdev:	GTH device
+ * @output:	output device's descriptor
+ *
+ * This will configure all masters set to output to this device and
+ * enable tracing using force storeEn signal.
+ */
+static void intel_th_gth_enable(struct intel_th_device *thdev,
+				struct intel_th_output *output)
+{
+	struct gth_device *gth = dev_get_drvdata(&thdev->dev);
+	struct intel_th *th = to_intel_th(thdev);
+	int master;
+	u32 scrpd;
+
+	spin_lock(&gth->gth_lock);
+	for_each_set_bit(master, gth->output[output->port].master,
+			 TH_CONFIGURABLE_MASTERS + 1) {
+		gth_master_set(gth, master, output->port);
+	}
+
+	output->active = true;
+	spin_unlock(&gth->gth_lock);
+
+	if (INTEL_TH_CAP(th, tscu_enable))
+		gth_tscu_resync(gth);
+
+	scrpd = ioread32(gth->base + REG_GTH_SCRPD0);
+	scrpd |= output->scratchpad;
+	iowrite32(scrpd, gth->base + REG_GTH_SCRPD0);
+
+	intel_th_gth_start(gth, output);
+}
+
+/**
+ * intel_th_gth_switch() - execute a switch sequence
+ * @thdev:	GTH device
+ * @output:	output device's descriptor
+ *
+ * This will execute a switch sequence that will trigger a switch window
+ * when tracing to MSC in multi-block mode.
+ */
+static void intel_th_gth_switch(struct intel_th_device *thdev,
+				struct intel_th_output *output)
+{
+	struct gth_device *gth = dev_get_drvdata(&thdev->dev);
+	unsigned long count;
+	u32 reg;
+
+	/* trigger */
+	iowrite32(0, gth->base + REG_CTS_CTL);
+	iowrite32(CTS_CTL_SEQUENCER_ENABLE, gth->base + REG_CTS_CTL);
+	/* wait on trigger status */
+	for (reg = 0, count = CTS_TRIG_WAITLOOP_DEPTH;
+	     count && !(reg & BIT(4)); count--) {
+		reg = ioread32(gth->base + REG_CTS_STAT);
+		cpu_relax();
+	}
+	if (!count)
+		dev_dbg(&thdev->dev, "timeout waiting for CTS Trigger\n");
+
+	/* De-assert the trigger */
+	iowrite32(0, gth->base + REG_CTS_CTL);
+
+	intel_th_gth_stop(gth, output, false);
+	intel_th_gth_start(gth, output);
+}
+
+/**
+ * intel_th_gth_assign() - assign output device to a GTH output port
+ * @thdev:	GTH device
+ * @othdev:	output device
+ *
+ * This will match a given output device parameters against present
+ * output ports on the GTH and fill out relevant bits in output device's
+ * descriptor.
+ *
+ * Return:	0 on success, -errno on error.
+ */
+static int intel_th_gth_assign(struct intel_th_device *thdev,
+			       struct intel_th_device *othdev)
+{
+	struct gth_device *gth = dev_get_drvdata(&thdev->dev);
+	int i, id;
+
+	if (thdev->host_mode)
+		return -EBUSY;
+
+	if (othdev->type != INTEL_TH_OUTPUT)
+		return -EINVAL;
+
+	for (i = 0, id = 0; i < TH_POSSIBLE_OUTPUTS; i++) {
+		if (gth->output[i].port_type != othdev->output.type)
+			continue;
+
+		if (othdev->id == -1 || othdev->id == id)
+			goto found;
+
+		id++;
+	}
+
+	return -ENOENT;
+
+found:
+	spin_lock(&gth->gth_lock);
+	othdev->output.port = i;
+	othdev->output.active = false;
+	gth->output[i].output = &othdev->output;
+	spin_unlock(&gth->gth_lock);
+
+	return 0;
+}
+
+/**
+ * intel_th_gth_unassign() - deassociate an output device from its output port
+ * @thdev:	GTH device
+ * @othdev:	output device
+ */
+static void intel_th_gth_unassign(struct intel_th_device *thdev,
+				  struct intel_th_device *othdev)
+{
+	struct gth_device *gth = dev_get_drvdata(&thdev->dev);
+	int port = othdev->output.port;
+	int master;
+
+	if (thdev->host_mode)
+		return;
+
+	spin_lock(&gth->gth_lock);
+	othdev->output.port = -1;
+	othdev->output.active = false;
+	gth->output[port].output = NULL;
+	for (master = 0; master < TH_CONFIGURABLE_MASTERS + 1; master++)
+		if (gth->master[master] == port)
+			gth->master[master] = -1;
+	spin_unlock(&gth->gth_lock);
+}
+
+static int
+intel_th_gth_set_output(struct intel_th_device *thdev, unsigned int master)
+{
+	struct gth_device *gth = dev_get_drvdata(&thdev->dev);
+	int port = 0; /* FIXME: make default output configurable */
+
+	/*
+	 * everything above TH_CONFIGURABLE_MASTERS is controlled by the
+	 * same register
+	 */
+	if (master > TH_CONFIGURABLE_MASTERS)
+		master = TH_CONFIGURABLE_MASTERS;
+
+	spin_lock(&gth->gth_lock);
+	if (gth->master[master] == -1) {
+		set_bit(master, gth->output[port].master);
+		gth->master[master] = port;
+	}
+	spin_unlock(&gth->gth_lock);
+
+	return 0;
+}
+
+static int intel_th_gth_probe(struct intel_th_device *thdev)
+{
+	struct device *dev = &thdev->dev;
+	struct intel_th *th = dev_get_drvdata(dev->parent);
+	struct gth_device *gth;
+	struct resource *res;
+	void __iomem *base;
+	int i, ret;
+
+	res = intel_th_device_get_resource(thdev, IORESOURCE_MEM, 0);
+	if (!res)
+		return -ENODEV;
+
+	base = devm_ioremap(dev, res->start, resource_size(res));
+	if (!base)
+		return -ENOMEM;
+
+	gth = devm_kzalloc(dev, sizeof(*gth), GFP_KERNEL);
+	if (!gth)
+		return -ENOMEM;
+
+	gth->dev = dev;
+	gth->base = base;
+	spin_lock_init(&gth->gth_lock);
+
+	dev_set_drvdata(dev, gth);
+
+	/*
+	 * Host mode can be signalled via SW means or via SCRPD_DEBUGGER_IN_USE
+	 * bit. Either way, don't reset HW in this case, and don't export any
+	 * capture configuration attributes. Also, refuse to assign output
+	 * drivers to ports, see intel_th_gth_assign().
+	 */
+	if (thdev->host_mode)
+		return 0;
+
+	ret = intel_th_gth_reset(gth);
+	if (ret) {
+		if (ret != -EBUSY)
+			return ret;
+
+		thdev->host_mode = true;
+
+		return 0;
+	}
+
+	for (i = 0; i < TH_CONFIGURABLE_MASTERS + 1; i++)
+		gth->master[i] = -1;
+
+	for (i = 0; i < TH_POSSIBLE_OUTPUTS; i++) {
+		gth->output[i].gth = gth;
+		gth->output[i].index = i;
+		gth->output[i].port_type =
+			gth_output_parm_get(gth, i, TH_OUTPUT_PARM(port));
+		if (gth->output[i].port_type == GTH_NONE)
+			continue;
+
+		ret = intel_th_output_enable(th, gth->output[i].port_type);
+		/* -ENODEV is ok, we just won't have that device enumerated */
+		if (ret && ret != -ENODEV)
+			return ret;
+	}
+
+	if (intel_th_output_attributes(gth) ||
+	    intel_th_master_attributes(gth)) {
+		pr_warn("Can't initialize sysfs attributes\n");
+
+		if (gth->output_group.attrs)
+			sysfs_remove_group(&gth->dev->kobj, &gth->output_group);
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static void intel_th_gth_remove(struct intel_th_device *thdev)
+{
+	struct gth_device *gth = dev_get_drvdata(&thdev->dev);
+
+	sysfs_remove_group(&gth->dev->kobj, &gth->output_group);
+	sysfs_remove_group(&gth->dev->kobj, &gth->master_group);
+}
+
+static struct intel_th_driver intel_th_gth_driver = {
+	.probe		= intel_th_gth_probe,
+	.remove		= intel_th_gth_remove,
+	.assign		= intel_th_gth_assign,
+	.unassign	= intel_th_gth_unassign,
+	.set_output	= intel_th_gth_set_output,
+	.prepare	= intel_th_gth_prepare,
+	.enable		= intel_th_gth_enable,
+	.trig_switch	= intel_th_gth_switch,
+	.disable	= intel_th_gth_disable,
+	.driver	= {
+		.name	= "gth",
+		.owner	= THIS_MODULE,
+	},
+};
+
+module_driver(intel_th_gth_driver,
+	      intel_th_driver_register,
+	      intel_th_driver_unregister);
+
+MODULE_ALIAS("intel_th_switch");
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("Intel(R) Trace Hub Global Trace Hub driver");
+MODULE_AUTHOR("Alexander Shishkin <alexander.shishkin@linux.intel.com>");
diff --git a/drivers/hwtracing/intel_th/gth.h b/drivers/hwtracing/intel_th/gth.h
new file mode 100644
index 0000000000..bfcc0fd011
--- /dev/null
+++ b/drivers/hwtracing/intel_th/gth.h
@@ -0,0 +1,79 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Intel(R) Trace Hub Global Trace Hub (GTH) data structures
+ *
+ * Copyright (C) 2014-2015 Intel Corporation.
+ */
+
+#ifndef __INTEL_TH_GTH_H__
+#define __INTEL_TH_GTH_H__
+
+/* Map output port parameter bits to symbolic names */
+#define TH_OUTPUT_PARM(name)			\
+	TH_OUTPUT_ ## name
+
+enum intel_th_output_parm {
+	/* output port type */
+	TH_OUTPUT_PARM(port),
+	/* generate NULL packet */
+	TH_OUTPUT_PARM(null),
+	/* packet drop */
+	TH_OUTPUT_PARM(drop),
+	/* port in reset state */
+	TH_OUTPUT_PARM(reset),
+	/* flush out data */
+	TH_OUTPUT_PARM(flush),
+	/* mainenance packet frequency */
+	TH_OUTPUT_PARM(smcfreq),
+};
+
+/*
+ * Register offsets
+ */
+enum {
+	REG_GTH_GTHOPT0		= 0x00, /* Output ports 0..3 config */
+	REG_GTH_GTHOPT1		= 0x04, /* Output ports 4..7 config */
+	REG_GTH_SWDEST0		= 0x08, /* Switching destination masters 0..7 */
+	REG_GTH_GSWTDEST	= 0x88, /* Global sw trace destination */
+	REG_GTH_SMCR0		= 0x9c, /* STP mainenance for ports 0/1 */
+	REG_GTH_SMCR1		= 0xa0, /* STP mainenance for ports 2/3 */
+	REG_GTH_SMCR2		= 0xa4, /* STP mainenance for ports 4/5 */
+	REG_GTH_SMCR3		= 0xa8, /* STP mainenance for ports 6/7 */
+	REG_GTH_SCR		= 0xc8, /* Source control (storeEn override) */
+	REG_GTH_STAT		= 0xd4, /* GTH status */
+	REG_GTH_SCR2		= 0xd8, /* Source control (force storeEn off) */
+	REG_GTH_DESTOVR		= 0xdc, /* Destination override */
+	REG_GTH_SCRPD0		= 0xe0, /* ScratchPad[0] */
+	REG_GTH_SCRPD1		= 0xe4, /* ScratchPad[1] */
+	REG_GTH_SCRPD2		= 0xe8, /* ScratchPad[2] */
+	REG_GTH_SCRPD3		= 0xec, /* ScratchPad[3] */
+	REG_TSCU_TSUCTRL	= 0x2000, /* TSCU control register */
+	REG_TSCU_TSCUSTAT	= 0x2004, /* TSCU status register */
+
+	/* Common Capture Sequencer (CTS) registers */
+	REG_CTS_C0S0_EN		= 0x30c0, /* clause_event_enable_c0s0 */
+	REG_CTS_C0S0_ACT	= 0x3180, /* clause_action_control_c0s0 */
+	REG_CTS_STAT		= 0x32a0, /* cts_status */
+	REG_CTS_CTL		= 0x32a4, /* cts_control */
+};
+
+/* waiting for Pipeline Empty bit(s) to assert for GTH */
+#define GTH_PLE_WAITLOOP_DEPTH	10000
+
+#define TSUCTRL_CTCRESYNC	BIT(0)
+#define TSCUSTAT_CTCSYNCING	BIT(1)
+
+/* waiting for Trigger status to assert for CTS */
+#define CTS_TRIG_WAITLOOP_DEPTH	10000
+
+#define CTS_EVENT_ENABLE_IF_ANYTHING	BIT(31)
+#define CTS_ACTION_CONTROL_STATE_OFF	27
+#define CTS_ACTION_CONTROL_SET_STATE(x)	\
+	(((x) & 0x1f) << CTS_ACTION_CONTROL_STATE_OFF)
+#define CTS_ACTION_CONTROL_TRIGGER	BIT(4)
+
+#define CTS_STATE_IDLE			0x10u
+
+#define CTS_CTL_SEQUENCER_ENABLE	BIT(0)
+
+#endif /* __INTEL_TH_GTH_H__ */
diff --git a/drivers/hwtracing/intel_th/intel_th.h b/drivers/hwtracing/intel_th/intel_th.h
new file mode 100644
index 0000000000..6cbba733f2
--- /dev/null
+++ b/drivers/hwtracing/intel_th/intel_th.h
@@ -0,0 +1,384 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Intel(R) Trace Hub data structures
+ *
+ * Copyright (C) 2014-2015 Intel Corporation.
+ */
+
+#ifndef __INTEL_TH_H__
+#define __INTEL_TH_H__
+
+#include <linux/irqreturn.h>
+
+/* intel_th_device device types */
+enum {
+	/* Devices that generate trace data */
+	INTEL_TH_SOURCE = 0,
+	/* Output ports (MSC, PTI) */
+	INTEL_TH_OUTPUT,
+	/* Switch, the Global Trace Hub (GTH) */
+	INTEL_TH_SWITCH,
+};
+
+struct intel_th_device;
+
+/**
+ * struct intel_th_output - descriptor INTEL_TH_OUTPUT type devices
+ * @port:	output port number, assigned by the switch
+ * @type:	GTH_{MSU,CTP,PTI}
+ * @scratchpad:	scratchpad bits to flag when this output is enabled
+ * @multiblock:	true for multiblock output configuration
+ * @active:	true when this output is enabled
+ * @wait_empty:	wait for device pipeline to be empty
+ *
+ * Output port descriptor, used by switch driver to tell which output
+ * port this output device corresponds to. Filled in at output device's
+ * probe time by switch::assign(). Passed from output device driver to
+ * switch related code to enable/disable its port.
+ */
+struct intel_th_output {
+	int		port;
+	unsigned int	type;
+	unsigned int	scratchpad;
+	bool		multiblock;
+	bool		active;
+};
+
+/**
+ * struct intel_th_drvdata - describes hardware capabilities and quirks
+ * @tscu_enable:	device needs SW to enable time stamping unit
+ * @multi_is_broken:	device has multiblock mode is broken
+ * @has_mintctl:	device has interrupt control (MINTCTL) register
+ * @host_mode_only:	device can only operate in 'host debugger' mode
+ */
+struct intel_th_drvdata {
+	unsigned int	tscu_enable        : 1,
+			multi_is_broken    : 1,
+			has_mintctl        : 1,
+			host_mode_only     : 1;
+};
+
+#define INTEL_TH_CAP(_th, _cap) ((_th)->drvdata ? (_th)->drvdata->_cap : 0)
+
+/**
+ * struct intel_th_device - device on the intel_th bus
+ * @dev:		device
+ * @drvdata:		hardware capabilities/quirks
+ * @resource:		array of resources available to this device
+ * @num_resources:	number of resources in @resource array
+ * @type:		INTEL_TH_{SOURCE,OUTPUT,SWITCH}
+ * @id:			device instance or -1
+ * @host_mode:		Intel TH is controlled by an external debug host
+ * @output:		output descriptor for INTEL_TH_OUTPUT devices
+ * @name:		device name to match the driver
+ */
+struct intel_th_device {
+	struct device		dev;
+	const struct intel_th_drvdata *drvdata;
+	struct resource		*resource;
+	unsigned int		num_resources;
+	unsigned int		type;
+	int			id;
+
+	/* INTEL_TH_SWITCH specific */
+	bool			host_mode;
+
+	/* INTEL_TH_OUTPUT specific */
+	struct intel_th_output	output;
+
+	char		name[];
+};
+
+#define to_intel_th_device(_d)				\
+	container_of((_d), struct intel_th_device, dev)
+
+/**
+ * intel_th_device_get_resource() - obtain @num'th resource of type @type
+ * @thdev:	the device to search the resource for
+ * @type:	resource type
+ * @num:	number of the resource
+ */
+static inline struct resource *
+intel_th_device_get_resource(struct intel_th_device *thdev, unsigned int type,
+			     unsigned int num)
+{
+	int i;
+
+	for (i = 0; i < thdev->num_resources; i++)
+		if (resource_type(&thdev->resource[i]) == type && !num--)
+			return &thdev->resource[i];
+
+	return NULL;
+}
+
+/*
+ * GTH, output ports configuration
+ */
+enum {
+	GTH_NONE = 0,
+	GTH_MSU,	/* memory/usb */
+	GTH_CTP,	/* Common Trace Port */
+	GTH_LPP,	/* Low Power Path */
+	GTH_PTI,	/* MIPI-PTI */
+};
+
+/**
+ * intel_th_output_assigned() - if an output device is assigned to a switch port
+ * @thdev:	the output device
+ *
+ * Return:	true if the device is INTEL_TH_OUTPUT *and* is assigned a port
+ */
+static inline bool
+intel_th_output_assigned(struct intel_th_device *thdev)
+{
+	return thdev->type == INTEL_TH_OUTPUT &&
+		(thdev->output.port >= 0 ||
+		 thdev->output.type == GTH_NONE);
+}
+
+/**
+ * struct intel_th_driver - driver for an intel_th_device device
+ * @driver:	generic driver
+ * @probe:	probe method
+ * @remove:	remove method
+ * @assign:	match a given output type device against available outputs
+ * @unassign:	deassociate an output type device from an output port
+ * @prepare:	prepare output port for tracing
+ * @enable:	enable tracing for a given output device
+ * @disable:	disable tracing for a given output device
+ * @irq:	interrupt callback
+ * @activate:	enable tracing on the output's side
+ * @deactivate:	disable tracing on the output's side
+ * @fops:	file operations for device nodes
+ * @attr_group:	attributes provided by the driver
+ *
+ * Callbacks @probe and @remove are required for all device types.
+ * Switch device driver needs to fill in @assign, @enable and @disable
+ * callbacks.
+ */
+struct intel_th_driver {
+	struct device_driver	driver;
+	int			(*probe)(struct intel_th_device *thdev);
+	void			(*remove)(struct intel_th_device *thdev);
+	/* switch (GTH) ops */
+	int			(*assign)(struct intel_th_device *thdev,
+					  struct intel_th_device *othdev);
+	void			(*unassign)(struct intel_th_device *thdev,
+					    struct intel_th_device *othdev);
+	void			(*prepare)(struct intel_th_device *thdev,
+					   struct intel_th_output *output);
+	void			(*enable)(struct intel_th_device *thdev,
+					  struct intel_th_output *output);
+	void			(*trig_switch)(struct intel_th_device *thdev,
+					       struct intel_th_output *output);
+	void			(*disable)(struct intel_th_device *thdev,
+					   struct intel_th_output *output);
+	/* output ops */
+	irqreturn_t		(*irq)(struct intel_th_device *thdev);
+	void			(*wait_empty)(struct intel_th_device *thdev);
+	int			(*activate)(struct intel_th_device *thdev);
+	void			(*deactivate)(struct intel_th_device *thdev);
+	/* file_operations for those who want a device node */
+	const struct file_operations *fops;
+	/* optional attributes */
+	const struct attribute_group *attr_group;
+
+	/* source ops */
+	int			(*set_output)(struct intel_th_device *thdev,
+					      unsigned int master);
+};
+
+#define to_intel_th_driver(_d)					\
+	container_of((_d), struct intel_th_driver, driver)
+
+#define to_intel_th_driver_or_null(_d)		\
+	((_d) ? to_intel_th_driver(_d) : NULL)
+
+/*
+ * Subdevice tree structure is as follows:
+ * + struct intel_th device (pci; dev_{get,set}_drvdata()
+ *   + struct intel_th_device INTEL_TH_SWITCH (GTH)
+ *     + struct intel_th_device INTEL_TH_OUTPUT (MSU, PTI)
+ *   + struct intel_th_device INTEL_TH_SOURCE (STH)
+ *
+ * In other words, INTEL_TH_OUTPUT devices are children of INTEL_TH_SWITCH;
+ * INTEL_TH_SWITCH and INTEL_TH_SOURCE are children of the intel_th device.
+ */
+static inline struct intel_th_device *
+to_intel_th_parent(const struct intel_th_device *thdev)
+{
+	struct device *parent = thdev->dev.parent;
+
+	if (!parent)
+		return NULL;
+
+	return to_intel_th_device(parent);
+}
+
+static inline struct intel_th *to_intel_th(const struct intel_th_device *thdev)
+{
+	if (thdev->type == INTEL_TH_OUTPUT)
+		thdev = to_intel_th_parent(thdev);
+
+	if (WARN_ON_ONCE(!thdev || thdev->type == INTEL_TH_OUTPUT))
+		return NULL;
+
+	return dev_get_drvdata(thdev->dev.parent);
+}
+
+struct intel_th *
+intel_th_alloc(struct device *dev, const struct intel_th_drvdata *drvdata,
+	       struct resource *devres, unsigned int ndevres);
+void intel_th_free(struct intel_th *th);
+
+int intel_th_driver_register(struct intel_th_driver *thdrv);
+void intel_th_driver_unregister(struct intel_th_driver *thdrv);
+
+int intel_th_trace_enable(struct intel_th_device *thdev);
+int intel_th_trace_switch(struct intel_th_device *thdev);
+int intel_th_trace_disable(struct intel_th_device *thdev);
+int intel_th_set_output(struct intel_th_device *thdev,
+			unsigned int master);
+int intel_th_output_enable(struct intel_th *th, unsigned int otype);
+
+enum th_mmio_idx {
+	TH_MMIO_CONFIG = 0,
+	TH_MMIO_SW = 1,
+	TH_MMIO_RTIT = 2,
+	TH_MMIO_END,
+};
+
+#define TH_POSSIBLE_OUTPUTS	8
+/* Total number of possible subdevices: outputs + GTH + STH */
+#define TH_SUBDEVICE_MAX	(TH_POSSIBLE_OUTPUTS + 2)
+#define TH_CONFIGURABLE_MASTERS 256
+#define TH_MSC_MAX		2
+
+/* Maximum IRQ vectors */
+#define TH_NVEC_MAX		8
+
+/**
+ * struct intel_th - Intel TH controller
+ * @dev:	driver core's device
+ * @thdev:	subdevices
+ * @hub:	"switch" subdevice (GTH)
+ * @resource:	resources of the entire controller
+ * @num_thdevs:	number of devices in the @thdev array
+ * @num_resources:	number of resources in the @resource array
+ * @irq:	irq number
+ * @num_irqs:	number of IRQs is use
+ * @id:		this Intel TH controller's device ID in the system
+ * @major:	device node major for output devices
+ */
+struct intel_th {
+	struct device		*dev;
+
+	struct intel_th_device	*thdev[TH_SUBDEVICE_MAX];
+	struct intel_th_device	*hub;
+	const struct intel_th_drvdata	*drvdata;
+
+	struct resource		resource[TH_MMIO_END];
+	int			(*activate)(struct intel_th *);
+	void			(*deactivate)(struct intel_th *);
+	unsigned int		num_thdevs;
+	unsigned int		num_resources;
+	int			irq;
+	int			num_irqs;
+
+	int			id;
+	int			major;
+#ifdef CONFIG_MODULES
+	struct work_struct	request_module_work;
+#endif /* CONFIG_MODULES */
+#ifdef CONFIG_INTEL_TH_DEBUG
+	struct dentry		*dbg;
+#endif
+};
+
+static inline struct intel_th_device *
+to_intel_th_hub(struct intel_th_device *thdev)
+{
+	if (thdev->type == INTEL_TH_SWITCH)
+		return thdev;
+	else if (thdev->type == INTEL_TH_OUTPUT)
+		return to_intel_th_parent(thdev);
+
+	return to_intel_th(thdev)->hub;
+}
+
+/*
+ * Register windows
+ */
+enum {
+	/* Global Trace Hub (GTH) */
+	REG_GTH_OFFSET		= 0x0000,
+	REG_GTH_LENGTH		= 0x2000,
+
+	/* Timestamp counter unit (TSCU) */
+	REG_TSCU_OFFSET		= 0x2000,
+	REG_TSCU_LENGTH		= 0x1000,
+
+	REG_CTS_OFFSET		= 0x3000,
+	REG_CTS_LENGTH		= 0x1000,
+
+	/* Software Trace Hub (STH) [0x4000..0x4fff] */
+	REG_STH_OFFSET		= 0x4000,
+	REG_STH_LENGTH		= 0x2000,
+
+	/* Memory Storage Unit (MSU) [0xa0000..0xa1fff] */
+	REG_MSU_OFFSET		= 0xa0000,
+	REG_MSU_LENGTH		= 0x02000,
+
+	/* Internal MSU trace buffer [0x80000..0x9ffff] */
+	BUF_MSU_OFFSET		= 0x80000,
+	BUF_MSU_LENGTH		= 0x20000,
+
+	/* PTI output == same window as GTH */
+	REG_PTI_OFFSET		= REG_GTH_OFFSET,
+	REG_PTI_LENGTH		= REG_GTH_LENGTH,
+
+	/* DCI Handler (DCIH) == some window as MSU */
+	REG_DCIH_OFFSET		= REG_MSU_OFFSET,
+	REG_DCIH_LENGTH		= REG_MSU_LENGTH,
+};
+
+/*
+ * Scratchpad bits: tell firmware and external debuggers
+ * what we are up to.
+ */
+enum {
+	/* Memory is the primary destination */
+	SCRPD_MEM_IS_PRIM_DEST		= BIT(0),
+	/* XHCI DbC is the primary destination */
+	SCRPD_DBC_IS_PRIM_DEST		= BIT(1),
+	/* PTI is the primary destination */
+	SCRPD_PTI_IS_PRIM_DEST		= BIT(2),
+	/* BSSB is the primary destination */
+	SCRPD_BSSB_IS_PRIM_DEST		= BIT(3),
+	/* PTI is the alternate destination */
+	SCRPD_PTI_IS_ALT_DEST		= BIT(4),
+	/* BSSB is the alternate destination */
+	SCRPD_BSSB_IS_ALT_DEST		= BIT(5),
+	/* DeepSx exit occurred */
+	SCRPD_DEEPSX_EXIT		= BIT(6),
+	/* S4 exit occurred */
+	SCRPD_S4_EXIT			= BIT(7),
+	/* S5 exit occurred */
+	SCRPD_S5_EXIT			= BIT(8),
+	/* MSU controller 0/1 is enabled */
+	SCRPD_MSC0_IS_ENABLED		= BIT(9),
+	SCRPD_MSC1_IS_ENABLED		= BIT(10),
+	/* Sx exit occurred */
+	SCRPD_SX_EXIT			= BIT(11),
+	/* Trigger Unit is enabled */
+	SCRPD_TRIGGER_IS_ENABLED	= BIT(12),
+	SCRPD_ODLA_IS_ENABLED		= BIT(13),
+	SCRPD_SOCHAP_IS_ENABLED		= BIT(14),
+	SCRPD_STH_IS_ENABLED		= BIT(15),
+	SCRPD_DCIH_IS_ENABLED		= BIT(16),
+	SCRPD_VER_IS_ENABLED		= BIT(17),
+	/* External debugger is using Intel TH */
+	SCRPD_DEBUGGER_IN_USE		= BIT(24),
+};
+
+#endif
diff --git a/drivers/hwtracing/intel_th/msu-sink.c b/drivers/hwtracing/intel_th/msu-sink.c
new file mode 100644
index 0000000000..891b28ea25
--- /dev/null
+++ b/drivers/hwtracing/intel_th/msu-sink.c
@@ -0,0 +1,119 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * An example software sink buffer for Intel TH MSU.
+ *
+ * Copyright (C) 2019 Intel Corporation.
+ */
+
+#include <linux/intel_th.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+
+#define MAX_SGTS 16
+
+struct msu_sink_private {
+	struct device	*dev;
+	struct sg_table **sgts;
+	unsigned int	nr_sgts;
+};
+
+static void *msu_sink_assign(struct device *dev, int *mode)
+{
+	struct msu_sink_private *priv;
+
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return NULL;
+
+	priv->sgts = kcalloc(MAX_SGTS, sizeof(void *), GFP_KERNEL);
+	if (!priv->sgts) {
+		kfree(priv);
+		return NULL;
+	}
+
+	priv->dev = dev;
+	*mode = MSC_MODE_MULTI;
+
+	return priv;
+}
+
+static void msu_sink_unassign(void *data)
+{
+	struct msu_sink_private *priv = data;
+
+	kfree(priv->sgts);
+	kfree(priv);
+}
+
+/* See also: msc.c: __msc_buffer_win_alloc() */
+static int msu_sink_alloc_window(void *data, struct sg_table **sgt, size_t size)
+{
+	struct msu_sink_private *priv = data;
+	unsigned int nents;
+	struct scatterlist *sg_ptr;
+	void *block;
+	int ret, i;
+
+	if (priv->nr_sgts == MAX_SGTS)
+		return -ENOMEM;
+
+	nents = DIV_ROUND_UP(size, PAGE_SIZE);
+
+	ret = sg_alloc_table(*sgt, nents, GFP_KERNEL);
+	if (ret)
+		return -ENOMEM;
+
+	priv->sgts[priv->nr_sgts++] = *sgt;
+
+	for_each_sg((*sgt)->sgl, sg_ptr, nents, i) {
+		block = dma_alloc_coherent(priv->dev->parent->parent,
+					   PAGE_SIZE, &sg_dma_address(sg_ptr),
+					   GFP_KERNEL);
+		if (!block)
+			return -ENOMEM;
+
+		sg_set_buf(sg_ptr, block, PAGE_SIZE);
+	}
+
+	return nents;
+}
+
+/* See also: msc.c: __msc_buffer_win_free() */
+static void msu_sink_free_window(void *data, struct sg_table *sgt)
+{
+	struct msu_sink_private *priv = data;
+	struct scatterlist *sg_ptr;
+	int i;
+
+	for_each_sg(sgt->sgl, sg_ptr, sgt->nents, i) {
+		dma_free_coherent(priv->dev->parent->parent, PAGE_SIZE,
+				  sg_virt(sg_ptr), sg_dma_address(sg_ptr));
+	}
+
+	sg_free_table(sgt);
+	priv->nr_sgts--;
+}
+
+static int msu_sink_ready(void *data, struct sg_table *sgt, size_t bytes)
+{
+	struct msu_sink_private *priv = data;
+
+	intel_th_msc_window_unlock(priv->dev, sgt);
+
+	return 0;
+}
+
+static const struct msu_buffer sink_mbuf = {
+	.name		= "sink",
+	.assign		= msu_sink_assign,
+	.unassign	= msu_sink_unassign,
+	.alloc_window	= msu_sink_alloc_window,
+	.free_window	= msu_sink_free_window,
+	.ready		= msu_sink_ready,
+};
+
+module_intel_th_msu_buffer(sink_mbuf);
+
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/hwtracing/intel_th/msu.c b/drivers/hwtracing/intel_th/msu.c
new file mode 100644
index 0000000000..9621efe0e9
--- /dev/null
+++ b/drivers/hwtracing/intel_th/msu.c
@@ -0,0 +1,2206 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Intel(R) Trace Hub Memory Storage Unit
+ *
+ * Copyright (C) 2014-2015 Intel Corporation.
+ */
+
+#define pr_fmt(fmt)	KBUILD_MODNAME ": " fmt
+
+#include <linux/types.h>
+#include <linux/module.h>
+#include <linux/device.h>
+#include <linux/uaccess.h>
+#include <linux/sizes.h>
+#include <linux/printk.h>
+#include <linux/slab.h>
+#include <linux/mm.h>
+#include <linux/fs.h>
+#include <linux/io.h>
+#include <linux/workqueue.h>
+#include <linux/dma-mapping.h>
+
+#ifdef CONFIG_X86
+#include <asm/set_memory.h>
+#endif
+
+#include <linux/intel_th.h>
+#include "intel_th.h"
+#include "msu.h"
+
+#define msc_dev(x) (&(x)->thdev->dev)
+
+/*
+ * Lockout state transitions:
+ *   READY -> INUSE -+-> LOCKED -+-> READY -> etc.
+ *                   \-----------/
+ * WIN_READY:	window can be used by HW
+ * WIN_INUSE:	window is in use
+ * WIN_LOCKED:	window is filled up and is being processed by the buffer
+ * handling code
+ *
+ * All state transitions happen automatically, except for the LOCKED->READY,
+ * which needs to be signalled by the buffer code by calling
+ * intel_th_msc_window_unlock().
+ *
+ * When the interrupt handler has to switch to the next window, it checks
+ * whether it's READY, and if it is, it performs the switch and tracing
+ * continues. If it's LOCKED, it stops the trace.
+ */
+enum lockout_state {
+	WIN_READY = 0,
+	WIN_INUSE,
+	WIN_LOCKED
+};
+
+/**
+ * struct msc_window - multiblock mode window descriptor
+ * @entry:	window list linkage (msc::win_list)
+ * @pgoff:	page offset into the buffer that this window starts at
+ * @lockout:	lockout state, see comment below
+ * @lo_lock:	lockout state serialization
+ * @nr_blocks:	number of blocks (pages) in this window
+ * @nr_segs:	number of segments in this window (<= @nr_blocks)
+ * @_sgt:	array of block descriptors
+ * @sgt:	array of block descriptors
+ */
+struct msc_window {
+	struct list_head	entry;
+	unsigned long		pgoff;
+	enum lockout_state	lockout;
+	spinlock_t		lo_lock;
+	unsigned int		nr_blocks;
+	unsigned int		nr_segs;
+	struct msc		*msc;
+	struct sg_table		_sgt;
+	struct sg_table		*sgt;
+};
+
+/**
+ * struct msc_iter - iterator for msc buffer
+ * @entry:		msc::iter_list linkage
+ * @msc:		pointer to the MSC device
+ * @start_win:		oldest window
+ * @win:		current window
+ * @offset:		current logical offset into the buffer
+ * @start_block:	oldest block in the window
+ * @block:		block number in the window
+ * @block_off:		offset into current block
+ * @wrap_count:		block wrapping handling
+ * @eof:		end of buffer reached
+ */
+struct msc_iter {
+	struct list_head	entry;
+	struct msc		*msc;
+	struct msc_window	*start_win;
+	struct msc_window	*win;
+	unsigned long		offset;
+	struct scatterlist	*start_block;
+	struct scatterlist	*block;
+	unsigned int		block_off;
+	unsigned int		wrap_count;
+	unsigned int		eof;
+};
+
+/**
+ * struct msc - MSC device representation
+ * @reg_base:		register window base address
+ * @thdev:		intel_th_device pointer
+ * @mbuf:		MSU buffer, if assigned
+ * @mbuf_priv		MSU buffer's private data, if @mbuf
+ * @win_list:		list of windows in multiblock mode
+ * @single_sgt:		single mode buffer
+ * @cur_win:		current window
+ * @nr_pages:		total number of pages allocated for this buffer
+ * @single_sz:		amount of data in single mode
+ * @single_wrap:	single mode wrap occurred
+ * @base:		buffer's base pointer
+ * @base_addr:		buffer's base address
+ * @user_count:		number of users of the buffer
+ * @mmap_count:		number of mappings
+ * @buf_mutex:		mutex to serialize access to buffer-related bits
+
+ * @enabled:		MSC is enabled
+ * @wrap:		wrapping is enabled
+ * @mode:		MSC operating mode
+ * @burst_len:		write burst length
+ * @index:		number of this MSC in the MSU
+ */
+struct msc {
+	void __iomem		*reg_base;
+	void __iomem		*msu_base;
+	struct intel_th_device	*thdev;
+
+	const struct msu_buffer	*mbuf;
+	void			*mbuf_priv;
+
+	struct work_struct	work;
+	struct list_head	win_list;
+	struct sg_table		single_sgt;
+	struct msc_window	*cur_win;
+	struct msc_window	*switch_on_unlock;
+	unsigned long		nr_pages;
+	unsigned long		single_sz;
+	unsigned int		single_wrap : 1;
+	void			*base;
+	dma_addr_t		base_addr;
+	u32			orig_addr;
+	u32			orig_sz;
+
+	/* <0: no buffer, 0: no users, >0: active users */
+	atomic_t		user_count;
+
+	atomic_t		mmap_count;
+	struct mutex		buf_mutex;
+
+	struct list_head	iter_list;
+
+	bool			stop_on_full;
+
+	/* config */
+	unsigned int		enabled : 1,
+				wrap	: 1,
+				do_irq	: 1,
+				multi_is_broken : 1;
+	unsigned int		mode;
+	unsigned int		burst_len;
+	unsigned int		index;
+};
+
+static LIST_HEAD(msu_buffer_list);
+static DEFINE_MUTEX(msu_buffer_mutex);
+
+/**
+ * struct msu_buffer_entry - internal MSU buffer bookkeeping
+ * @entry:	link to msu_buffer_list
+ * @mbuf:	MSU buffer object
+ * @owner:	module that provides this MSU buffer
+ */
+struct msu_buffer_entry {
+	struct list_head	entry;
+	const struct msu_buffer	*mbuf;
+	struct module		*owner;
+};
+
+static struct msu_buffer_entry *__msu_buffer_entry_find(const char *name)
+{
+	struct msu_buffer_entry *mbe;
+
+	lockdep_assert_held(&msu_buffer_mutex);
+
+	list_for_each_entry(mbe, &msu_buffer_list, entry) {
+		if (!strcmp(mbe->mbuf->name, name))
+			return mbe;
+	}
+
+	return NULL;
+}
+
+static const struct msu_buffer *
+msu_buffer_get(const char *name)
+{
+	struct msu_buffer_entry *mbe;
+
+	mutex_lock(&msu_buffer_mutex);
+	mbe = __msu_buffer_entry_find(name);
+	if (mbe && !try_module_get(mbe->owner))
+		mbe = NULL;
+	mutex_unlock(&msu_buffer_mutex);
+
+	return mbe ? mbe->mbuf : NULL;
+}
+
+static void msu_buffer_put(const struct msu_buffer *mbuf)
+{
+	struct msu_buffer_entry *mbe;
+
+	mutex_lock(&msu_buffer_mutex);
+	mbe = __msu_buffer_entry_find(mbuf->name);
+	if (mbe)
+		module_put(mbe->owner);
+	mutex_unlock(&msu_buffer_mutex);
+}
+
+int intel_th_msu_buffer_register(const struct msu_buffer *mbuf,
+				 struct module *owner)
+{
+	struct msu_buffer_entry *mbe;
+	int ret = 0;
+
+	mbe = kzalloc(sizeof(*mbe), GFP_KERNEL);
+	if (!mbe)
+		return -ENOMEM;
+
+	mutex_lock(&msu_buffer_mutex);
+	if (__msu_buffer_entry_find(mbuf->name)) {
+		ret = -EEXIST;
+		kfree(mbe);
+		goto unlock;
+	}
+
+	mbe->mbuf = mbuf;
+	mbe->owner = owner;
+	list_add_tail(&mbe->entry, &msu_buffer_list);
+unlock:
+	mutex_unlock(&msu_buffer_mutex);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(intel_th_msu_buffer_register);
+
+void intel_th_msu_buffer_unregister(const struct msu_buffer *mbuf)
+{
+	struct msu_buffer_entry *mbe;
+
+	mutex_lock(&msu_buffer_mutex);
+	mbe = __msu_buffer_entry_find(mbuf->name);
+	if (mbe) {
+		list_del(&mbe->entry);
+		kfree(mbe);
+	}
+	mutex_unlock(&msu_buffer_mutex);
+}
+EXPORT_SYMBOL_GPL(intel_th_msu_buffer_unregister);
+
+static inline bool msc_block_is_empty(struct msc_block_desc *bdesc)
+{
+	/* header hasn't been written */
+	if (!bdesc->valid_dw)
+		return true;
+
+	/* valid_dw includes the header */
+	if (!msc_data_sz(bdesc))
+		return true;
+
+	return false;
+}
+
+static inline struct scatterlist *msc_win_base_sg(struct msc_window *win)
+{
+	return win->sgt->sgl;
+}
+
+static inline struct msc_block_desc *msc_win_base(struct msc_window *win)
+{
+	return sg_virt(msc_win_base_sg(win));
+}
+
+static inline dma_addr_t msc_win_base_dma(struct msc_window *win)
+{
+	return sg_dma_address(msc_win_base_sg(win));
+}
+
+static inline unsigned long
+msc_win_base_pfn(struct msc_window *win)
+{
+	return PFN_DOWN(msc_win_base_dma(win));
+}
+
+/**
+ * msc_is_last_win() - check if a window is the last one for a given MSC
+ * @win:	window
+ * Return:	true if @win is the last window in MSC's multiblock buffer
+ */
+static inline bool msc_is_last_win(struct msc_window *win)
+{
+	return win->entry.next == &win->msc->win_list;
+}
+
+/**
+ * msc_next_window() - return next window in the multiblock buffer
+ * @win:	current window
+ *
+ * Return:	window following the current one
+ */
+static struct msc_window *msc_next_window(struct msc_window *win)
+{
+	if (msc_is_last_win(win))
+		return list_first_entry(&win->msc->win_list, struct msc_window,
+					entry);
+
+	return list_next_entry(win, entry);
+}
+
+static size_t msc_win_total_sz(struct msc_window *win)
+{
+	struct scatterlist *sg;
+	unsigned int blk;
+	size_t size = 0;
+
+	for_each_sg(win->sgt->sgl, sg, win->nr_segs, blk) {
+		struct msc_block_desc *bdesc = sg_virt(sg);
+
+		if (msc_block_wrapped(bdesc))
+			return (size_t)win->nr_blocks << PAGE_SHIFT;
+
+		size += msc_total_sz(bdesc);
+		if (msc_block_last_written(bdesc))
+			break;
+	}
+
+	return size;
+}
+
+/**
+ * msc_find_window() - find a window matching a given sg_table
+ * @msc:	MSC device
+ * @sgt:	SG table of the window
+ * @nonempty:	skip over empty windows
+ *
+ * Return:	MSC window structure pointer or NULL if the window
+ *		could not be found.
+ */
+static struct msc_window *
+msc_find_window(struct msc *msc, struct sg_table *sgt, bool nonempty)
+{
+	struct msc_window *win;
+	unsigned int found = 0;
+
+	if (list_empty(&msc->win_list))
+		return NULL;
+
+	/*
+	 * we might need a radix tree for this, depending on how
+	 * many windows a typical user would allocate; ideally it's
+	 * something like 2, in which case we're good
+	 */
+	list_for_each_entry(win, &msc->win_list, entry) {
+		if (win->sgt == sgt)
+			found++;
+
+		/* skip the empty ones */
+		if (nonempty && msc_block_is_empty(msc_win_base(win)))
+			continue;
+
+		if (found)
+			return win;
+	}
+
+	return NULL;
+}
+
+/**
+ * msc_oldest_window() - locate the window with oldest data
+ * @msc:	MSC device
+ *
+ * This should only be used in multiblock mode. Caller should hold the
+ * msc::user_count reference.
+ *
+ * Return:	the oldest window with valid data
+ */
+static struct msc_window *msc_oldest_window(struct msc *msc)
+{
+	struct msc_window *win;
+
+	if (list_empty(&msc->win_list))
+		return NULL;
+
+	win = msc_find_window(msc, msc_next_window(msc->cur_win)->sgt, true);
+	if (win)
+		return win;
+
+	return list_first_entry(&msc->win_list, struct msc_window, entry);
+}
+
+/**
+ * msc_win_oldest_sg() - locate the oldest block in a given window
+ * @win:	window to look at
+ *
+ * Return:	index of the block with the oldest data
+ */
+static struct scatterlist *msc_win_oldest_sg(struct msc_window *win)
+{
+	unsigned int blk;
+	struct scatterlist *sg;
+	struct msc_block_desc *bdesc = msc_win_base(win);
+
+	/* without wrapping, first block is the oldest */
+	if (!msc_block_wrapped(bdesc))
+		return msc_win_base_sg(win);
+
+	/*
+	 * with wrapping, last written block contains both the newest and the
+	 * oldest data for this window.
+	 */
+	for_each_sg(win->sgt->sgl, sg, win->nr_segs, blk) {
+		struct msc_block_desc *bdesc = sg_virt(sg);
+
+		if (msc_block_last_written(bdesc))
+			return sg;
+	}
+
+	return msc_win_base_sg(win);
+}
+
+static struct msc_block_desc *msc_iter_bdesc(struct msc_iter *iter)
+{
+	return sg_virt(iter->block);
+}
+
+static struct msc_iter *msc_iter_install(struct msc *msc)
+{
+	struct msc_iter *iter;
+
+	iter = kzalloc(sizeof(*iter), GFP_KERNEL);
+	if (!iter)
+		return ERR_PTR(-ENOMEM);
+
+	mutex_lock(&msc->buf_mutex);
+
+	/*
+	 * Reading and tracing are mutually exclusive; if msc is
+	 * enabled, open() will fail; otherwise existing readers
+	 * will prevent enabling the msc and the rest of fops don't
+	 * need to worry about it.
+	 */
+	if (msc->enabled) {
+		kfree(iter);
+		iter = ERR_PTR(-EBUSY);
+		goto unlock;
+	}
+
+	iter->msc = msc;
+
+	list_add_tail(&iter->entry, &msc->iter_list);
+unlock:
+	mutex_unlock(&msc->buf_mutex);
+
+	return iter;
+}
+
+static void msc_iter_remove(struct msc_iter *iter, struct msc *msc)
+{
+	mutex_lock(&msc->buf_mutex);
+	list_del(&iter->entry);
+	mutex_unlock(&msc->buf_mutex);
+
+	kfree(iter);
+}
+
+static void msc_iter_block_start(struct msc_iter *iter)
+{
+	if (iter->start_block)
+		return;
+
+	iter->start_block = msc_win_oldest_sg(iter->win);
+	iter->block = iter->start_block;
+	iter->wrap_count = 0;
+
+	/*
+	 * start with the block with oldest data; if data has wrapped
+	 * in this window, it should be in this block
+	 */
+	if (msc_block_wrapped(msc_iter_bdesc(iter)))
+		iter->wrap_count = 2;
+
+}
+
+static int msc_iter_win_start(struct msc_iter *iter, struct msc *msc)
+{
+	/* already started, nothing to do */
+	if (iter->start_win)
+		return 0;
+
+	iter->start_win = msc_oldest_window(msc);
+	if (!iter->start_win)
+		return -EINVAL;
+
+	iter->win = iter->start_win;
+	iter->start_block = NULL;
+
+	msc_iter_block_start(iter);
+
+	return 0;
+}
+
+static int msc_iter_win_advance(struct msc_iter *iter)
+{
+	iter->win = msc_next_window(iter->win);
+	iter->start_block = NULL;
+
+	if (iter->win == iter->start_win) {
+		iter->eof++;
+		return 1;
+	}
+
+	msc_iter_block_start(iter);
+
+	return 0;
+}
+
+static int msc_iter_block_advance(struct msc_iter *iter)
+{
+	iter->block_off = 0;
+
+	/* wrapping */
+	if (iter->wrap_count && iter->block == iter->start_block) {
+		iter->wrap_count--;
+		if (!iter->wrap_count)
+			/* copied newest data from the wrapped block */
+			return msc_iter_win_advance(iter);
+	}
+
+	/* no wrapping, check for last written block */
+	if (!iter->wrap_count && msc_block_last_written(msc_iter_bdesc(iter)))
+		/* copied newest data for the window */
+		return msc_iter_win_advance(iter);
+
+	/* block advance */
+	if (sg_is_last(iter->block))
+		iter->block = msc_win_base_sg(iter->win);
+	else
+		iter->block = sg_next(iter->block);
+
+	/* no wrapping, sanity check in case there is no last written block */
+	if (!iter->wrap_count && iter->block == iter->start_block)
+		return msc_iter_win_advance(iter);
+
+	return 0;
+}
+
+/**
+ * msc_buffer_iterate() - go through multiblock buffer's data
+ * @iter:	iterator structure
+ * @size:	amount of data to scan
+ * @data:	callback's private data
+ * @fn:		iterator callback
+ *
+ * This will start at the window which will be written to next (containing
+ * the oldest data) and work its way to the current window, calling @fn
+ * for each chunk of data as it goes.
+ *
+ * Caller should have msc::user_count reference to make sure the buffer
+ * doesn't disappear from under us.
+ *
+ * Return:	amount of data actually scanned.
+ */
+static ssize_t
+msc_buffer_iterate(struct msc_iter *iter, size_t size, void *data,
+		   unsigned long (*fn)(void *, void *, size_t))
+{
+	struct msc *msc = iter->msc;
+	size_t len = size;
+	unsigned int advance;
+
+	if (iter->eof)
+		return 0;
+
+	/* start with the oldest window */
+	if (msc_iter_win_start(iter, msc))
+		return 0;
+
+	do {
+		unsigned long data_bytes = msc_data_sz(msc_iter_bdesc(iter));
+		void *src = (void *)msc_iter_bdesc(iter) + MSC_BDESC;
+		size_t tocopy = data_bytes, copied = 0;
+		size_t remaining = 0;
+
+		advance = 1;
+
+		/*
+		 * If block wrapping happened, we need to visit the last block
+		 * twice, because it contains both the oldest and the newest
+		 * data in this window.
+		 *
+		 * First time (wrap_count==2), in the very beginning, to collect
+		 * the oldest data, which is in the range
+		 * (data_bytes..DATA_IN_PAGE).
+		 *
+		 * Second time (wrap_count==1), it's just like any other block,
+		 * containing data in the range of [MSC_BDESC..data_bytes].
+		 */
+		if (iter->block == iter->start_block && iter->wrap_count == 2) {
+			tocopy = DATA_IN_PAGE - data_bytes;
+			src += data_bytes;
+		}
+
+		if (!tocopy)
+			goto next_block;
+
+		tocopy -= iter->block_off;
+		src += iter->block_off;
+
+		if (len < tocopy) {
+			tocopy = len;
+			advance = 0;
+		}
+
+		remaining = fn(data, src, tocopy);
+
+		if (remaining)
+			advance = 0;
+
+		copied = tocopy - remaining;
+		len -= copied;
+		iter->block_off += copied;
+		iter->offset += copied;
+
+		if (!advance)
+			break;
+
+next_block:
+		if (msc_iter_block_advance(iter))
+			break;
+
+	} while (len);
+
+	return size - len;
+}
+
+/**
+ * msc_buffer_clear_hw_header() - clear hw header for multiblock
+ * @msc:	MSC device
+ */
+static void msc_buffer_clear_hw_header(struct msc *msc)
+{
+	struct msc_window *win;
+	struct scatterlist *sg;
+
+	list_for_each_entry(win, &msc->win_list, entry) {
+		unsigned int blk;
+
+		for_each_sg(win->sgt->sgl, sg, win->nr_segs, blk) {
+			struct msc_block_desc *bdesc = sg_virt(sg);
+
+			memset_startat(bdesc, 0, hw_tag);
+		}
+	}
+}
+
+static int intel_th_msu_init(struct msc *msc)
+{
+	u32 mintctl, msusts;
+
+	if (!msc->do_irq)
+		return 0;
+
+	if (!msc->mbuf)
+		return 0;
+
+	mintctl = ioread32(msc->msu_base + REG_MSU_MINTCTL);
+	mintctl |= msc->index ? M1BLIE : M0BLIE;
+	iowrite32(mintctl, msc->msu_base + REG_MSU_MINTCTL);
+	if (mintctl != ioread32(msc->msu_base + REG_MSU_MINTCTL)) {
+		dev_info(msc_dev(msc), "MINTCTL ignores writes: no usable interrupts\n");
+		msc->do_irq = 0;
+		return 0;
+	}
+
+	msusts = ioread32(msc->msu_base + REG_MSU_MSUSTS);
+	iowrite32(msusts, msc->msu_base + REG_MSU_MSUSTS);
+
+	return 0;
+}
+
+static void intel_th_msu_deinit(struct msc *msc)
+{
+	u32 mintctl;
+
+	if (!msc->do_irq)
+		return;
+
+	mintctl = ioread32(msc->msu_base + REG_MSU_MINTCTL);
+	mintctl &= msc->index ? ~M1BLIE : ~M0BLIE;
+	iowrite32(mintctl, msc->msu_base + REG_MSU_MINTCTL);
+}
+
+static int msc_win_set_lockout(struct msc_window *win,
+			       enum lockout_state expect,
+			       enum lockout_state new)
+{
+	enum lockout_state old;
+	unsigned long flags;
+	int ret = 0;
+
+	if (!win->msc->mbuf)
+		return 0;
+
+	spin_lock_irqsave(&win->lo_lock, flags);
+	old = win->lockout;
+
+	if (old != expect) {
+		ret = -EINVAL;
+		goto unlock;
+	}
+
+	win->lockout = new;
+
+	if (old == expect && new == WIN_LOCKED)
+		atomic_inc(&win->msc->user_count);
+	else if (old == expect && old == WIN_LOCKED)
+		atomic_dec(&win->msc->user_count);
+
+unlock:
+	spin_unlock_irqrestore(&win->lo_lock, flags);
+
+	if (ret) {
+		if (expect == WIN_READY && old == WIN_LOCKED)
+			return -EBUSY;
+
+		/* from intel_th_msc_window_unlock(), don't warn if not locked */
+		if (expect == WIN_LOCKED && old == new)
+			return 0;
+
+		dev_warn_ratelimited(msc_dev(win->msc),
+				     "expected lockout state %d, got %d\n",
+				     expect, old);
+	}
+
+	return ret;
+}
+/**
+ * msc_configure() - set up MSC hardware
+ * @msc:	the MSC device to configure
+ *
+ * Program storage mode, wrapping, burst length and trace buffer address
+ * into a given MSC. Then, enable tracing and set msc::enabled.
+ * The latter is serialized on msc::buf_mutex, so make sure to hold it.
+ */
+static int msc_configure(struct msc *msc)
+{
+	u32 reg;
+
+	lockdep_assert_held(&msc->buf_mutex);
+
+	if (msc->mode > MSC_MODE_MULTI)
+		return -EINVAL;
+
+	if (msc->mode == MSC_MODE_MULTI) {
+		if (msc_win_set_lockout(msc->cur_win, WIN_READY, WIN_INUSE))
+			return -EBUSY;
+
+		msc_buffer_clear_hw_header(msc);
+	}
+
+	msc->orig_addr = ioread32(msc->reg_base + REG_MSU_MSC0BAR);
+	msc->orig_sz   = ioread32(msc->reg_base + REG_MSU_MSC0SIZE);
+
+	reg = msc->base_addr >> PAGE_SHIFT;
+	iowrite32(reg, msc->reg_base + REG_MSU_MSC0BAR);
+
+	if (msc->mode == MSC_MODE_SINGLE) {
+		reg = msc->nr_pages;
+		iowrite32(reg, msc->reg_base + REG_MSU_MSC0SIZE);
+	}
+
+	reg = ioread32(msc->reg_base + REG_MSU_MSC0CTL);
+	reg &= ~(MSC_MODE | MSC_WRAPEN | MSC_EN | MSC_RD_HDR_OVRD);
+
+	reg |= MSC_EN;
+	reg |= msc->mode << __ffs(MSC_MODE);
+	reg |= msc->burst_len << __ffs(MSC_LEN);
+
+	if (msc->wrap)
+		reg |= MSC_WRAPEN;
+
+	iowrite32(reg, msc->reg_base + REG_MSU_MSC0CTL);
+
+	intel_th_msu_init(msc);
+
+	msc->thdev->output.multiblock = msc->mode == MSC_MODE_MULTI;
+	intel_th_trace_enable(msc->thdev);
+	msc->enabled = 1;
+
+	if (msc->mbuf && msc->mbuf->activate)
+		msc->mbuf->activate(msc->mbuf_priv);
+
+	return 0;
+}
+
+/**
+ * msc_disable() - disable MSC hardware
+ * @msc:	MSC device to disable
+ *
+ * If @msc is enabled, disable tracing on the switch and then disable MSC
+ * storage. Caller must hold msc::buf_mutex.
+ */
+static void msc_disable(struct msc *msc)
+{
+	struct msc_window *win = msc->cur_win;
+	u32 reg;
+
+	lockdep_assert_held(&msc->buf_mutex);
+
+	if (msc->mode == MSC_MODE_MULTI)
+		msc_win_set_lockout(win, WIN_INUSE, WIN_LOCKED);
+
+	if (msc->mbuf && msc->mbuf->deactivate)
+		msc->mbuf->deactivate(msc->mbuf_priv);
+	intel_th_msu_deinit(msc);
+	intel_th_trace_disable(msc->thdev);
+
+	if (msc->mode == MSC_MODE_SINGLE) {
+		reg = ioread32(msc->reg_base + REG_MSU_MSC0STS);
+		msc->single_wrap = !!(reg & MSCSTS_WRAPSTAT);
+
+		reg = ioread32(msc->reg_base + REG_MSU_MSC0MWP);
+		msc->single_sz = reg & ((msc->nr_pages << PAGE_SHIFT) - 1);
+		dev_dbg(msc_dev(msc), "MSCnMWP: %08x/%08lx, wrap: %d\n",
+			reg, msc->single_sz, msc->single_wrap);
+	}
+
+	reg = ioread32(msc->reg_base + REG_MSU_MSC0CTL);
+	reg &= ~MSC_EN;
+	iowrite32(reg, msc->reg_base + REG_MSU_MSC0CTL);
+
+	if (msc->mbuf && msc->mbuf->ready)
+		msc->mbuf->ready(msc->mbuf_priv, win->sgt,
+				 msc_win_total_sz(win));
+
+	msc->enabled = 0;
+
+	iowrite32(msc->orig_addr, msc->reg_base + REG_MSU_MSC0BAR);
+	iowrite32(msc->orig_sz, msc->reg_base + REG_MSU_MSC0SIZE);
+
+	dev_dbg(msc_dev(msc), "MSCnNWSA: %08x\n",
+		ioread32(msc->reg_base + REG_MSU_MSC0NWSA));
+
+	reg = ioread32(msc->reg_base + REG_MSU_MSC0STS);
+	dev_dbg(msc_dev(msc), "MSCnSTS: %08x\n", reg);
+
+	reg = ioread32(msc->reg_base + REG_MSU_MSUSTS);
+	reg &= msc->index ? MSUSTS_MSC1BLAST : MSUSTS_MSC0BLAST;
+	iowrite32(reg, msc->reg_base + REG_MSU_MSUSTS);
+}
+
+static int intel_th_msc_activate(struct intel_th_device *thdev)
+{
+	struct msc *msc = dev_get_drvdata(&thdev->dev);
+	int ret = -EBUSY;
+
+	if (!atomic_inc_unless_negative(&msc->user_count))
+		return -ENODEV;
+
+	mutex_lock(&msc->buf_mutex);
+
+	/* if there are readers, refuse */
+	if (list_empty(&msc->iter_list))
+		ret = msc_configure(msc);
+
+	mutex_unlock(&msc->buf_mutex);
+
+	if (ret)
+		atomic_dec(&msc->user_count);
+
+	return ret;
+}
+
+static void intel_th_msc_deactivate(struct intel_th_device *thdev)
+{
+	struct msc *msc = dev_get_drvdata(&thdev->dev);
+
+	mutex_lock(&msc->buf_mutex);
+	if (msc->enabled) {
+		msc_disable(msc);
+		atomic_dec(&msc->user_count);
+	}
+	mutex_unlock(&msc->buf_mutex);
+}
+
+/**
+ * msc_buffer_contig_alloc() - allocate a contiguous buffer for SINGLE mode
+ * @msc:	MSC device
+ * @size:	allocation size in bytes
+ *
+ * This modifies msc::base, which requires msc::buf_mutex to serialize, so the
+ * caller is expected to hold it.
+ *
+ * Return:	0 on success, -errno otherwise.
+ */
+static int msc_buffer_contig_alloc(struct msc *msc, unsigned long size)
+{
+	unsigned long nr_pages = size >> PAGE_SHIFT;
+	unsigned int order = get_order(size);
+	struct page *page;
+	int ret;
+
+	if (!size)
+		return 0;
+
+	ret = sg_alloc_table(&msc->single_sgt, 1, GFP_KERNEL);
+	if (ret)
+		goto err_out;
+
+	ret = -ENOMEM;
+	page = alloc_pages(GFP_KERNEL | __GFP_ZERO | GFP_DMA32, order);
+	if (!page)
+		goto err_free_sgt;
+
+	split_page(page, order);
+	sg_set_buf(msc->single_sgt.sgl, page_address(page), size);
+
+	ret = dma_map_sg(msc_dev(msc)->parent->parent, msc->single_sgt.sgl, 1,
+			 DMA_FROM_DEVICE);
+	if (ret < 0)
+		goto err_free_pages;
+
+	msc->nr_pages = nr_pages;
+	msc->base = page_address(page);
+	msc->base_addr = sg_dma_address(msc->single_sgt.sgl);
+
+	return 0;
+
+err_free_pages:
+	__free_pages(page, order);
+
+err_free_sgt:
+	sg_free_table(&msc->single_sgt);
+
+err_out:
+	return ret;
+}
+
+/**
+ * msc_buffer_contig_free() - free a contiguous buffer
+ * @msc:	MSC configured in SINGLE mode
+ */
+static void msc_buffer_contig_free(struct msc *msc)
+{
+	unsigned long off;
+
+	dma_unmap_sg(msc_dev(msc)->parent->parent, msc->single_sgt.sgl,
+		     1, DMA_FROM_DEVICE);
+	sg_free_table(&msc->single_sgt);
+
+	for (off = 0; off < msc->nr_pages << PAGE_SHIFT; off += PAGE_SIZE) {
+		struct page *page = virt_to_page(msc->base + off);
+
+		page->mapping = NULL;
+		__free_page(page);
+	}
+
+	msc->nr_pages = 0;
+}
+
+/**
+ * msc_buffer_contig_get_page() - find a page at a given offset
+ * @msc:	MSC configured in SINGLE mode
+ * @pgoff:	page offset
+ *
+ * Return:	page, if @pgoff is within the range, NULL otherwise.
+ */
+static struct page *msc_buffer_contig_get_page(struct msc *msc,
+					       unsigned long pgoff)
+{
+	if (pgoff >= msc->nr_pages)
+		return NULL;
+
+	return virt_to_page(msc->base + (pgoff << PAGE_SHIFT));
+}
+
+static int __msc_buffer_win_alloc(struct msc_window *win,
+				  unsigned int nr_segs)
+{
+	struct scatterlist *sg_ptr;
+	void *block;
+	int i, ret;
+
+	ret = sg_alloc_table(win->sgt, nr_segs, GFP_KERNEL);
+	if (ret)
+		return -ENOMEM;
+
+	for_each_sg(win->sgt->sgl, sg_ptr, nr_segs, i) {
+		block = dma_alloc_coherent(msc_dev(win->msc)->parent->parent,
+					  PAGE_SIZE, &sg_dma_address(sg_ptr),
+					  GFP_KERNEL);
+		if (!block)
+			goto err_nomem;
+
+		sg_set_buf(sg_ptr, block, PAGE_SIZE);
+	}
+
+	return nr_segs;
+
+err_nomem:
+	for_each_sg(win->sgt->sgl, sg_ptr, i, ret)
+		dma_free_coherent(msc_dev(win->msc)->parent->parent, PAGE_SIZE,
+				  sg_virt(sg_ptr), sg_dma_address(sg_ptr));
+
+	sg_free_table(win->sgt);
+
+	return -ENOMEM;
+}
+
+#ifdef CONFIG_X86
+static void msc_buffer_set_uc(struct msc *msc)
+{
+	struct scatterlist *sg_ptr;
+	struct msc_window *win;
+	int i;
+
+	if (msc->mode == MSC_MODE_SINGLE) {
+		set_memory_uc((unsigned long)msc->base, msc->nr_pages);
+		return;
+	}
+
+	list_for_each_entry(win, &msc->win_list, entry) {
+		for_each_sg(win->sgt->sgl, sg_ptr, win->nr_segs, i) {
+			/* Set the page as uncached */
+			set_memory_uc((unsigned long)sg_virt(sg_ptr),
+					PFN_DOWN(sg_ptr->length));
+		}
+	}
+}
+
+static void msc_buffer_set_wb(struct msc *msc)
+{
+	struct scatterlist *sg_ptr;
+	struct msc_window *win;
+	int i;
+
+	if (msc->mode == MSC_MODE_SINGLE) {
+		set_memory_wb((unsigned long)msc->base, msc->nr_pages);
+		return;
+	}
+
+	list_for_each_entry(win, &msc->win_list, entry) {
+		for_each_sg(win->sgt->sgl, sg_ptr, win->nr_segs, i) {
+			/* Reset the page to write-back */
+			set_memory_wb((unsigned long)sg_virt(sg_ptr),
+					PFN_DOWN(sg_ptr->length));
+		}
+	}
+}
+#else /* !X86 */
+static inline void
+msc_buffer_set_uc(struct msc *msc) {}
+static inline void msc_buffer_set_wb(struct msc *msc) {}
+#endif /* CONFIG_X86 */
+
+static struct page *msc_sg_page(struct scatterlist *sg)
+{
+	void *addr = sg_virt(sg);
+
+	if (is_vmalloc_addr(addr))
+		return vmalloc_to_page(addr);
+
+	return sg_page(sg);
+}
+
+/**
+ * msc_buffer_win_alloc() - alloc a window for a multiblock mode
+ * @msc:	MSC device
+ * @nr_blocks:	number of pages in this window
+ *
+ * This modifies msc::win_list and msc::base, which requires msc::buf_mutex
+ * to serialize, so the caller is expected to hold it.
+ *
+ * Return:	0 on success, -errno otherwise.
+ */
+static int msc_buffer_win_alloc(struct msc *msc, unsigned int nr_blocks)
+{
+	struct msc_window *win;
+	int ret = -ENOMEM;
+
+	if (!nr_blocks)
+		return 0;
+
+	win = kzalloc(sizeof(*win), GFP_KERNEL);
+	if (!win)
+		return -ENOMEM;
+
+	win->msc = msc;
+	win->sgt = &win->_sgt;
+	win->lockout = WIN_READY;
+	spin_lock_init(&win->lo_lock);
+
+	if (!list_empty(&msc->win_list)) {
+		struct msc_window *prev = list_last_entry(&msc->win_list,
+							  struct msc_window,
+							  entry);
+
+		win->pgoff = prev->pgoff + prev->nr_blocks;
+	}
+
+	if (msc->mbuf && msc->mbuf->alloc_window)
+		ret = msc->mbuf->alloc_window(msc->mbuf_priv, &win->sgt,
+					      nr_blocks << PAGE_SHIFT);
+	else
+		ret = __msc_buffer_win_alloc(win, nr_blocks);
+
+	if (ret <= 0)
+		goto err_nomem;
+
+	win->nr_segs = ret;
+	win->nr_blocks = nr_blocks;
+
+	if (list_empty(&msc->win_list)) {
+		msc->base = msc_win_base(win);
+		msc->base_addr = msc_win_base_dma(win);
+		msc->cur_win = win;
+	}
+
+	list_add_tail(&win->entry, &msc->win_list);
+	msc->nr_pages += nr_blocks;
+
+	return 0;
+
+err_nomem:
+	kfree(win);
+
+	return ret;
+}
+
+static void __msc_buffer_win_free(struct msc *msc, struct msc_window *win)
+{
+	struct scatterlist *sg;
+	int i;
+
+	for_each_sg(win->sgt->sgl, sg, win->nr_segs, i) {
+		struct page *page = msc_sg_page(sg);
+
+		page->mapping = NULL;
+		dma_free_coherent(msc_dev(win->msc)->parent->parent, PAGE_SIZE,
+				  sg_virt(sg), sg_dma_address(sg));
+	}
+	sg_free_table(win->sgt);
+}
+
+/**
+ * msc_buffer_win_free() - free a window from MSC's window list
+ * @msc:	MSC device
+ * @win:	window to free
+ *
+ * This modifies msc::win_list and msc::base, which requires msc::buf_mutex
+ * to serialize, so the caller is expected to hold it.
+ */
+static void msc_buffer_win_free(struct msc *msc, struct msc_window *win)
+{
+	msc->nr_pages -= win->nr_blocks;
+
+	list_del(&win->entry);
+	if (list_empty(&msc->win_list)) {
+		msc->base = NULL;
+		msc->base_addr = 0;
+	}
+
+	if (msc->mbuf && msc->mbuf->free_window)
+		msc->mbuf->free_window(msc->mbuf_priv, win->sgt);
+	else
+		__msc_buffer_win_free(msc, win);
+
+	kfree(win);
+}
+
+/**
+ * msc_buffer_relink() - set up block descriptors for multiblock mode
+ * @msc:	MSC device
+ *
+ * This traverses msc::win_list, which requires msc::buf_mutex to serialize,
+ * so the caller is expected to hold it.
+ */
+static void msc_buffer_relink(struct msc *msc)
+{
+	struct msc_window *win, *next_win;
+
+	/* call with msc::mutex locked */
+	list_for_each_entry(win, &msc->win_list, entry) {
+		struct scatterlist *sg;
+		unsigned int blk;
+		u32 sw_tag = 0;
+
+		/*
+		 * Last window's next_win should point to the first window
+		 * and MSC_SW_TAG_LASTWIN should be set.
+		 */
+		if (msc_is_last_win(win)) {
+			sw_tag |= MSC_SW_TAG_LASTWIN;
+			next_win = list_first_entry(&msc->win_list,
+						    struct msc_window, entry);
+		} else {
+			next_win = list_next_entry(win, entry);
+		}
+
+		for_each_sg(win->sgt->sgl, sg, win->nr_segs, blk) {
+			struct msc_block_desc *bdesc = sg_virt(sg);
+
+			memset(bdesc, 0, sizeof(*bdesc));
+
+			bdesc->next_win = msc_win_base_pfn(next_win);
+
+			/*
+			 * Similarly to last window, last block should point
+			 * to the first one.
+			 */
+			if (blk == win->nr_segs - 1) {
+				sw_tag |= MSC_SW_TAG_LASTBLK;
+				bdesc->next_blk = msc_win_base_pfn(win);
+			} else {
+				dma_addr_t addr = sg_dma_address(sg_next(sg));
+
+				bdesc->next_blk = PFN_DOWN(addr);
+			}
+
+			bdesc->sw_tag = sw_tag;
+			bdesc->block_sz = sg->length / 64;
+		}
+	}
+
+	/*
+	 * Make the above writes globally visible before tracing is
+	 * enabled to make sure hardware sees them coherently.
+	 */
+	wmb();
+}
+
+static void msc_buffer_multi_free(struct msc *msc)
+{
+	struct msc_window *win, *iter;
+
+	list_for_each_entry_safe(win, iter, &msc->win_list, entry)
+		msc_buffer_win_free(msc, win);
+}
+
+static int msc_buffer_multi_alloc(struct msc *msc, unsigned long *nr_pages,
+				  unsigned int nr_wins)
+{
+	int ret, i;
+
+	for (i = 0; i < nr_wins; i++) {
+		ret = msc_buffer_win_alloc(msc, nr_pages[i]);
+		if (ret) {
+			msc_buffer_multi_free(msc);
+			return ret;
+		}
+	}
+
+	msc_buffer_relink(msc);
+
+	return 0;
+}
+
+/**
+ * msc_buffer_free() - free buffers for MSC
+ * @msc:	MSC device
+ *
+ * Free MSC's storage buffers.
+ *
+ * This modifies msc::win_list and msc::base, which requires msc::buf_mutex to
+ * serialize, so the caller is expected to hold it.
+ */
+static void msc_buffer_free(struct msc *msc)
+{
+	msc_buffer_set_wb(msc);
+
+	if (msc->mode == MSC_MODE_SINGLE)
+		msc_buffer_contig_free(msc);
+	else if (msc->mode == MSC_MODE_MULTI)
+		msc_buffer_multi_free(msc);
+}
+
+/**
+ * msc_buffer_alloc() - allocate a buffer for MSC
+ * @msc:	MSC device
+ * @size:	allocation size in bytes
+ *
+ * Allocate a storage buffer for MSC, depending on the msc::mode, it will be
+ * either done via msc_buffer_contig_alloc() for SINGLE operation mode or
+ * msc_buffer_win_alloc() for multiblock operation. The latter allocates one
+ * window per invocation, so in multiblock mode this can be called multiple
+ * times for the same MSC to allocate multiple windows.
+ *
+ * This modifies msc::win_list and msc::base, which requires msc::buf_mutex
+ * to serialize, so the caller is expected to hold it.
+ *
+ * Return:	0 on success, -errno otherwise.
+ */
+static int msc_buffer_alloc(struct msc *msc, unsigned long *nr_pages,
+			    unsigned int nr_wins)
+{
+	int ret;
+
+	/* -1: buffer not allocated */
+	if (atomic_read(&msc->user_count) != -1)
+		return -EBUSY;
+
+	if (msc->mode == MSC_MODE_SINGLE) {
+		if (nr_wins != 1)
+			return -EINVAL;
+
+		ret = msc_buffer_contig_alloc(msc, nr_pages[0] << PAGE_SHIFT);
+	} else if (msc->mode == MSC_MODE_MULTI) {
+		ret = msc_buffer_multi_alloc(msc, nr_pages, nr_wins);
+	} else {
+		ret = -EINVAL;
+	}
+
+	if (!ret) {
+		msc_buffer_set_uc(msc);
+
+		/* allocation should be visible before the counter goes to 0 */
+		smp_mb__before_atomic();
+
+		if (WARN_ON_ONCE(atomic_cmpxchg(&msc->user_count, -1, 0) != -1))
+			return -EINVAL;
+	}
+
+	return ret;
+}
+
+/**
+ * msc_buffer_unlocked_free_unless_used() - free a buffer unless it's in use
+ * @msc:	MSC device
+ *
+ * This will free MSC buffer unless it is in use or there is no allocated
+ * buffer.
+ * Caller needs to hold msc::buf_mutex.
+ *
+ * Return:	0 on successful deallocation or if there was no buffer to
+ *		deallocate, -EBUSY if there are active users.
+ */
+static int msc_buffer_unlocked_free_unless_used(struct msc *msc)
+{
+	int count, ret = 0;
+
+	count = atomic_cmpxchg(&msc->user_count, 0, -1);
+
+	/* > 0: buffer is allocated and has users */
+	if (count > 0)
+		ret = -EBUSY;
+	/* 0: buffer is allocated, no users */
+	else if (!count)
+		msc_buffer_free(msc);
+	/* < 0: no buffer, nothing to do */
+
+	return ret;
+}
+
+/**
+ * msc_buffer_free_unless_used() - free a buffer unless it's in use
+ * @msc:	MSC device
+ *
+ * This is a locked version of msc_buffer_unlocked_free_unless_used().
+ */
+static int msc_buffer_free_unless_used(struct msc *msc)
+{
+	int ret;
+
+	mutex_lock(&msc->buf_mutex);
+	ret = msc_buffer_unlocked_free_unless_used(msc);
+	mutex_unlock(&msc->buf_mutex);
+
+	return ret;
+}
+
+/**
+ * msc_buffer_get_page() - get MSC buffer page at a given offset
+ * @msc:	MSC device
+ * @pgoff:	page offset into the storage buffer
+ *
+ * This traverses msc::win_list, so holding msc::buf_mutex is expected from
+ * the caller.
+ *
+ * Return:	page if @pgoff corresponds to a valid buffer page or NULL.
+ */
+static struct page *msc_buffer_get_page(struct msc *msc, unsigned long pgoff)
+{
+	struct msc_window *win;
+	struct scatterlist *sg;
+	unsigned int blk;
+
+	if (msc->mode == MSC_MODE_SINGLE)
+		return msc_buffer_contig_get_page(msc, pgoff);
+
+	list_for_each_entry(win, &msc->win_list, entry)
+		if (pgoff >= win->pgoff && pgoff < win->pgoff + win->nr_blocks)
+			goto found;
+
+	return NULL;
+
+found:
+	pgoff -= win->pgoff;
+
+	for_each_sg(win->sgt->sgl, sg, win->nr_segs, blk) {
+		struct page *page = msc_sg_page(sg);
+		size_t pgsz = PFN_DOWN(sg->length);
+
+		if (pgoff < pgsz)
+			return page + pgoff;
+
+		pgoff -= pgsz;
+	}
+
+	return NULL;
+}
+
+/**
+ * struct msc_win_to_user_struct - data for copy_to_user() callback
+ * @buf:	userspace buffer to copy data to
+ * @offset:	running offset
+ */
+struct msc_win_to_user_struct {
+	char __user	*buf;
+	unsigned long	offset;
+};
+
+/**
+ * msc_win_to_user() - iterator for msc_buffer_iterate() to copy data to user
+ * @data:	callback's private data
+ * @src:	source buffer
+ * @len:	amount of data to copy from the source buffer
+ */
+static unsigned long msc_win_to_user(void *data, void *src, size_t len)
+{
+	struct msc_win_to_user_struct *u = data;
+	unsigned long ret;
+
+	ret = copy_to_user(u->buf + u->offset, src, len);
+	u->offset += len - ret;
+
+	return ret;
+}
+
+
+/*
+ * file operations' callbacks
+ */
+
+static int intel_th_msc_open(struct inode *inode, struct file *file)
+{
+	struct intel_th_device *thdev = file->private_data;
+	struct msc *msc = dev_get_drvdata(&thdev->dev);
+	struct msc_iter *iter;
+
+	if (!capable(CAP_SYS_RAWIO))
+		return -EPERM;
+
+	iter = msc_iter_install(msc);
+	if (IS_ERR(iter))
+		return PTR_ERR(iter);
+
+	file->private_data = iter;
+
+	return nonseekable_open(inode, file);
+}
+
+static int intel_th_msc_release(struct inode *inode, struct file *file)
+{
+	struct msc_iter *iter = file->private_data;
+	struct msc *msc = iter->msc;
+
+	msc_iter_remove(iter, msc);
+
+	return 0;
+}
+
+static ssize_t
+msc_single_to_user(struct msc *msc, char __user *buf, loff_t off, size_t len)
+{
+	unsigned long size = msc->nr_pages << PAGE_SHIFT, rem = len;
+	unsigned long start = off, tocopy = 0;
+
+	if (msc->single_wrap) {
+		start += msc->single_sz;
+		if (start < size) {
+			tocopy = min(rem, size - start);
+			if (copy_to_user(buf, msc->base + start, tocopy))
+				return -EFAULT;
+
+			buf += tocopy;
+			rem -= tocopy;
+			start += tocopy;
+		}
+
+		start &= size - 1;
+		if (rem) {
+			tocopy = min(rem, msc->single_sz - start);
+			if (copy_to_user(buf, msc->base + start, tocopy))
+				return -EFAULT;
+
+			rem -= tocopy;
+		}
+
+		return len - rem;
+	}
+
+	if (copy_to_user(buf, msc->base + start, rem))
+		return -EFAULT;
+
+	return len;
+}
+
+static ssize_t intel_th_msc_read(struct file *file, char __user *buf,
+				 size_t len, loff_t *ppos)
+{
+	struct msc_iter *iter = file->private_data;
+	struct msc *msc = iter->msc;
+	size_t size;
+	loff_t off = *ppos;
+	ssize_t ret = 0;
+
+	if (!atomic_inc_unless_negative(&msc->user_count))
+		return 0;
+
+	if (msc->mode == MSC_MODE_SINGLE && !msc->single_wrap)
+		size = msc->single_sz;
+	else
+		size = msc->nr_pages << PAGE_SHIFT;
+
+	if (!size)
+		goto put_count;
+
+	if (off >= size)
+		goto put_count;
+
+	if (off + len >= size)
+		len = size - off;
+
+	if (msc->mode == MSC_MODE_SINGLE) {
+		ret = msc_single_to_user(msc, buf, off, len);
+		if (ret >= 0)
+			*ppos += ret;
+	} else if (msc->mode == MSC_MODE_MULTI) {
+		struct msc_win_to_user_struct u = {
+			.buf	= buf,
+			.offset	= 0,
+		};
+
+		ret = msc_buffer_iterate(iter, len, &u, msc_win_to_user);
+		if (ret >= 0)
+			*ppos = iter->offset;
+	} else {
+		ret = -EINVAL;
+	}
+
+put_count:
+	atomic_dec(&msc->user_count);
+
+	return ret;
+}
+
+/*
+ * vm operations callbacks (vm_ops)
+ */
+
+static void msc_mmap_open(struct vm_area_struct *vma)
+{
+	struct msc_iter *iter = vma->vm_file->private_data;
+	struct msc *msc = iter->msc;
+
+	atomic_inc(&msc->mmap_count);
+}
+
+static void msc_mmap_close(struct vm_area_struct *vma)
+{
+	struct msc_iter *iter = vma->vm_file->private_data;
+	struct msc *msc = iter->msc;
+	unsigned long pg;
+
+	if (!atomic_dec_and_mutex_lock(&msc->mmap_count, &msc->buf_mutex))
+		return;
+
+	/* drop page _refcounts */
+	for (pg = 0; pg < msc->nr_pages; pg++) {
+		struct page *page = msc_buffer_get_page(msc, pg);
+
+		if (WARN_ON_ONCE(!page))
+			continue;
+
+		if (page->mapping)
+			page->mapping = NULL;
+	}
+
+	/* last mapping -- drop user_count */
+	atomic_dec(&msc->user_count);
+	mutex_unlock(&msc->buf_mutex);
+}
+
+static vm_fault_t msc_mmap_fault(struct vm_fault *vmf)
+{
+	struct msc_iter *iter = vmf->vma->vm_file->private_data;
+	struct msc *msc = iter->msc;
+
+	vmf->page = msc_buffer_get_page(msc, vmf->pgoff);
+	if (!vmf->page)
+		return VM_FAULT_SIGBUS;
+
+	get_page(vmf->page);
+	vmf->page->mapping = vmf->vma->vm_file->f_mapping;
+	vmf->page->index = vmf->pgoff;
+
+	return 0;
+}
+
+static const struct vm_operations_struct msc_mmap_ops = {
+	.open	= msc_mmap_open,
+	.close	= msc_mmap_close,
+	.fault	= msc_mmap_fault,
+};
+
+static int intel_th_msc_mmap(struct file *file, struct vm_area_struct *vma)
+{
+	unsigned long size = vma->vm_end - vma->vm_start;
+	struct msc_iter *iter = vma->vm_file->private_data;
+	struct msc *msc = iter->msc;
+	int ret = -EINVAL;
+
+	if (!size || offset_in_page(size))
+		return -EINVAL;
+
+	if (vma->vm_pgoff)
+		return -EINVAL;
+
+	/* grab user_count once per mmap; drop in msc_mmap_close() */
+	if (!atomic_inc_unless_negative(&msc->user_count))
+		return -EINVAL;
+
+	if (msc->mode != MSC_MODE_SINGLE &&
+	    msc->mode != MSC_MODE_MULTI)
+		goto out;
+
+	if (size >> PAGE_SHIFT != msc->nr_pages)
+		goto out;
+
+	atomic_set(&msc->mmap_count, 1);
+	ret = 0;
+
+out:
+	if (ret)
+		atomic_dec(&msc->user_count);
+
+	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
+	vm_flags_set(vma, VM_DONTEXPAND | VM_DONTCOPY);
+	vma->vm_ops = &msc_mmap_ops;
+	return ret;
+}
+
+static const struct file_operations intel_th_msc_fops = {
+	.open		= intel_th_msc_open,
+	.release	= intel_th_msc_release,
+	.read		= intel_th_msc_read,
+	.mmap		= intel_th_msc_mmap,
+	.llseek		= no_llseek,
+	.owner		= THIS_MODULE,
+};
+
+static void intel_th_msc_wait_empty(struct intel_th_device *thdev)
+{
+	struct msc *msc = dev_get_drvdata(&thdev->dev);
+	unsigned long count;
+	u32 reg;
+
+	for (reg = 0, count = MSC_PLE_WAITLOOP_DEPTH;
+	     count && !(reg & MSCSTS_PLE); count--) {
+		reg = __raw_readl(msc->reg_base + REG_MSU_MSC0STS);
+		cpu_relax();
+	}
+
+	if (!count)
+		dev_dbg(msc_dev(msc), "timeout waiting for MSC0 PLE\n");
+}
+
+static int intel_th_msc_init(struct msc *msc)
+{
+	atomic_set(&msc->user_count, -1);
+
+	msc->mode = msc->multi_is_broken ? MSC_MODE_SINGLE : MSC_MODE_MULTI;
+	mutex_init(&msc->buf_mutex);
+	INIT_LIST_HEAD(&msc->win_list);
+	INIT_LIST_HEAD(&msc->iter_list);
+
+	msc->burst_len =
+		(ioread32(msc->reg_base + REG_MSU_MSC0CTL) & MSC_LEN) >>
+		__ffs(MSC_LEN);
+
+	return 0;
+}
+
+static int msc_win_switch(struct msc *msc)
+{
+	struct msc_window *first;
+
+	if (list_empty(&msc->win_list))
+		return -EINVAL;
+
+	first = list_first_entry(&msc->win_list, struct msc_window, entry);
+
+	if (msc_is_last_win(msc->cur_win))
+		msc->cur_win = first;
+	else
+		msc->cur_win = list_next_entry(msc->cur_win, entry);
+
+	msc->base = msc_win_base(msc->cur_win);
+	msc->base_addr = msc_win_base_dma(msc->cur_win);
+
+	intel_th_trace_switch(msc->thdev);
+
+	return 0;
+}
+
+/**
+ * intel_th_msc_window_unlock - put the window back in rotation
+ * @dev:	MSC device to which this relates
+ * @sgt:	buffer's sg_table for the window, does nothing if NULL
+ */
+void intel_th_msc_window_unlock(struct device *dev, struct sg_table *sgt)
+{
+	struct msc *msc = dev_get_drvdata(dev);
+	struct msc_window *win;
+
+	if (!sgt)
+		return;
+
+	win = msc_find_window(msc, sgt, false);
+	if (!win)
+		return;
+
+	msc_win_set_lockout(win, WIN_LOCKED, WIN_READY);
+	if (msc->switch_on_unlock == win) {
+		msc->switch_on_unlock = NULL;
+		msc_win_switch(msc);
+	}
+}
+EXPORT_SYMBOL_GPL(intel_th_msc_window_unlock);
+
+static void msc_work(struct work_struct *work)
+{
+	struct msc *msc = container_of(work, struct msc, work);
+
+	intel_th_msc_deactivate(msc->thdev);
+}
+
+static irqreturn_t intel_th_msc_interrupt(struct intel_th_device *thdev)
+{
+	struct msc *msc = dev_get_drvdata(&thdev->dev);
+	u32 msusts = ioread32(msc->msu_base + REG_MSU_MSUSTS);
+	u32 mask = msc->index ? MSUSTS_MSC1BLAST : MSUSTS_MSC0BLAST;
+	struct msc_window *win, *next_win;
+
+	if (!msc->do_irq || !msc->mbuf)
+		return IRQ_NONE;
+
+	msusts &= mask;
+
+	if (!msusts)
+		return msc->enabled ? IRQ_HANDLED : IRQ_NONE;
+
+	iowrite32(msusts, msc->msu_base + REG_MSU_MSUSTS);
+
+	if (!msc->enabled)
+		return IRQ_NONE;
+
+	/* grab the window before we do the switch */
+	win = msc->cur_win;
+	if (!win)
+		return IRQ_HANDLED;
+	next_win = msc_next_window(win);
+	if (!next_win)
+		return IRQ_HANDLED;
+
+	/* next window: if READY, proceed, if LOCKED, stop the trace */
+	if (msc_win_set_lockout(next_win, WIN_READY, WIN_INUSE)) {
+		if (msc->stop_on_full)
+			schedule_work(&msc->work);
+		else
+			msc->switch_on_unlock = next_win;
+
+		return IRQ_HANDLED;
+	}
+
+	/* current window: INUSE -> LOCKED */
+	msc_win_set_lockout(win, WIN_INUSE, WIN_LOCKED);
+
+	msc_win_switch(msc);
+
+	if (msc->mbuf && msc->mbuf->ready)
+		msc->mbuf->ready(msc->mbuf_priv, win->sgt,
+				 msc_win_total_sz(win));
+
+	return IRQ_HANDLED;
+}
+
+static const char * const msc_mode[] = {
+	[MSC_MODE_SINGLE]	= "single",
+	[MSC_MODE_MULTI]	= "multi",
+	[MSC_MODE_EXI]		= "ExI",
+	[MSC_MODE_DEBUG]	= "debug",
+};
+
+static ssize_t
+wrap_show(struct device *dev, struct device_attribute *attr, char *buf)
+{
+	struct msc *msc = dev_get_drvdata(dev);
+
+	return scnprintf(buf, PAGE_SIZE, "%d\n", msc->wrap);
+}
+
+static ssize_t
+wrap_store(struct device *dev, struct device_attribute *attr, const char *buf,
+	   size_t size)
+{
+	struct msc *msc = dev_get_drvdata(dev);
+	unsigned long val;
+	int ret;
+
+	ret = kstrtoul(buf, 10, &val);
+	if (ret)
+		return ret;
+
+	msc->wrap = !!val;
+
+	return size;
+}
+
+static DEVICE_ATTR_RW(wrap);
+
+static void msc_buffer_unassign(struct msc *msc)
+{
+	lockdep_assert_held(&msc->buf_mutex);
+
+	if (!msc->mbuf)
+		return;
+
+	msc->mbuf->unassign(msc->mbuf_priv);
+	msu_buffer_put(msc->mbuf);
+	msc->mbuf_priv = NULL;
+	msc->mbuf = NULL;
+}
+
+static ssize_t
+mode_show(struct device *dev, struct device_attribute *attr, char *buf)
+{
+	struct msc *msc = dev_get_drvdata(dev);
+	const char *mode = msc_mode[msc->mode];
+	ssize_t ret;
+
+	mutex_lock(&msc->buf_mutex);
+	if (msc->mbuf)
+		mode = msc->mbuf->name;
+	ret = scnprintf(buf, PAGE_SIZE, "%s\n", mode);
+	mutex_unlock(&msc->buf_mutex);
+
+	return ret;
+}
+
+static ssize_t
+mode_store(struct device *dev, struct device_attribute *attr, const char *buf,
+	   size_t size)
+{
+	const struct msu_buffer *mbuf = NULL;
+	struct msc *msc = dev_get_drvdata(dev);
+	size_t len = size;
+	char *cp, *mode;
+	int i, ret;
+
+	if (!capable(CAP_SYS_RAWIO))
+		return -EPERM;
+
+	cp = memchr(buf, '\n', len);
+	if (cp)
+		len = cp - buf;
+
+	mode = kstrndup(buf, len, GFP_KERNEL);
+	if (!mode)
+		return -ENOMEM;
+
+	i = match_string(msc_mode, ARRAY_SIZE(msc_mode), mode);
+	if (i >= 0) {
+		kfree(mode);
+		goto found;
+	}
+
+	/* Buffer sinks only work with a usable IRQ */
+	if (!msc->do_irq) {
+		kfree(mode);
+		return -EINVAL;
+	}
+
+	mbuf = msu_buffer_get(mode);
+	kfree(mode);
+	if (mbuf)
+		goto found;
+
+	return -EINVAL;
+
+found:
+	if (i == MSC_MODE_MULTI && msc->multi_is_broken)
+		return -EOPNOTSUPP;
+
+	mutex_lock(&msc->buf_mutex);
+	ret = 0;
+
+	/* Same buffer: do nothing */
+	if (mbuf && mbuf == msc->mbuf) {
+		/* put the extra reference we just got */
+		msu_buffer_put(mbuf);
+		goto unlock;
+	}
+
+	ret = msc_buffer_unlocked_free_unless_used(msc);
+	if (ret)
+		goto unlock;
+
+	if (mbuf) {
+		void *mbuf_priv = mbuf->assign(dev, &i);
+
+		if (!mbuf_priv) {
+			ret = -ENOMEM;
+			goto unlock;
+		}
+
+		msc_buffer_unassign(msc);
+		msc->mbuf_priv = mbuf_priv;
+		msc->mbuf = mbuf;
+	} else {
+		msc_buffer_unassign(msc);
+	}
+
+	msc->mode = i;
+
+unlock:
+	if (ret && mbuf)
+		msu_buffer_put(mbuf);
+	mutex_unlock(&msc->buf_mutex);
+
+	return ret ? ret : size;
+}
+
+static DEVICE_ATTR_RW(mode);
+
+static ssize_t
+nr_pages_show(struct device *dev, struct device_attribute *attr, char *buf)
+{
+	struct msc *msc = dev_get_drvdata(dev);
+	struct msc_window *win;
+	size_t count = 0;
+
+	mutex_lock(&msc->buf_mutex);
+
+	if (msc->mode == MSC_MODE_SINGLE)
+		count = scnprintf(buf, PAGE_SIZE, "%ld\n", msc->nr_pages);
+	else if (msc->mode == MSC_MODE_MULTI) {
+		list_for_each_entry(win, &msc->win_list, entry) {
+			count += scnprintf(buf + count, PAGE_SIZE - count,
+					   "%d%c", win->nr_blocks,
+					   msc_is_last_win(win) ? '\n' : ',');
+		}
+	} else {
+		count = scnprintf(buf, PAGE_SIZE, "unsupported\n");
+	}
+
+	mutex_unlock(&msc->buf_mutex);
+
+	return count;
+}
+
+static ssize_t
+nr_pages_store(struct device *dev, struct device_attribute *attr,
+	       const char *buf, size_t size)
+{
+	struct msc *msc = dev_get_drvdata(dev);
+	unsigned long val, *win = NULL, *rewin;
+	size_t len = size;
+	const char *p = buf;
+	char *end, *s;
+	int ret, nr_wins = 0;
+
+	if (!capable(CAP_SYS_RAWIO))
+		return -EPERM;
+
+	ret = msc_buffer_free_unless_used(msc);
+	if (ret)
+		return ret;
+
+	/* scan the comma-separated list of allocation sizes */
+	end = memchr(buf, '\n', len);
+	if (end)
+		len = end - buf;
+
+	do {
+		end = memchr(p, ',', len);
+		s = kstrndup(p, end ? end - p : len, GFP_KERNEL);
+		if (!s) {
+			ret = -ENOMEM;
+			goto free_win;
+		}
+
+		ret = kstrtoul(s, 10, &val);
+		kfree(s);
+
+		if (ret || !val)
+			goto free_win;
+
+		if (nr_wins && msc->mode == MSC_MODE_SINGLE) {
+			ret = -EINVAL;
+			goto free_win;
+		}
+
+		nr_wins++;
+		rewin = krealloc_array(win, nr_wins, sizeof(*win), GFP_KERNEL);
+		if (!rewin) {
+			kfree(win);
+			return -ENOMEM;
+		}
+
+		win = rewin;
+		win[nr_wins - 1] = val;
+
+		if (!end)
+			break;
+
+		/* consume the number and the following comma, hence +1 */
+		len -= end - p + 1;
+		p = end + 1;
+	} while (len);
+
+	mutex_lock(&msc->buf_mutex);
+	ret = msc_buffer_alloc(msc, win, nr_wins);
+	mutex_unlock(&msc->buf_mutex);
+
+free_win:
+	kfree(win);
+
+	return ret ? ret : size;
+}
+
+static DEVICE_ATTR_RW(nr_pages);
+
+static ssize_t
+win_switch_store(struct device *dev, struct device_attribute *attr,
+		 const char *buf, size_t size)
+{
+	struct msc *msc = dev_get_drvdata(dev);
+	unsigned long val;
+	int ret;
+
+	ret = kstrtoul(buf, 10, &val);
+	if (ret)
+		return ret;
+
+	if (val != 1)
+		return -EINVAL;
+
+	ret = -EINVAL;
+	mutex_lock(&msc->buf_mutex);
+	/*
+	 * Window switch can only happen in the "multi" mode.
+	 * If a external buffer is engaged, they have the full
+	 * control over window switching.
+	 */
+	if (msc->mode == MSC_MODE_MULTI && !msc->mbuf)
+		ret = msc_win_switch(msc);
+	mutex_unlock(&msc->buf_mutex);
+
+	return ret ? ret : size;
+}
+
+static DEVICE_ATTR_WO(win_switch);
+
+static ssize_t stop_on_full_show(struct device *dev,
+				 struct device_attribute *attr, char *buf)
+{
+	struct msc *msc = dev_get_drvdata(dev);
+
+	return sprintf(buf, "%d\n", msc->stop_on_full);
+}
+
+static ssize_t stop_on_full_store(struct device *dev,
+				  struct device_attribute *attr,
+				  const char *buf, size_t size)
+{
+	struct msc *msc = dev_get_drvdata(dev);
+	int ret;
+
+	ret = kstrtobool(buf, &msc->stop_on_full);
+	if (ret)
+		return ret;
+
+	return size;
+}
+
+static DEVICE_ATTR_RW(stop_on_full);
+
+static struct attribute *msc_output_attrs[] = {
+	&dev_attr_wrap.attr,
+	&dev_attr_mode.attr,
+	&dev_attr_nr_pages.attr,
+	&dev_attr_win_switch.attr,
+	&dev_attr_stop_on_full.attr,
+	NULL,
+};
+
+static const struct attribute_group msc_output_group = {
+	.attrs	= msc_output_attrs,
+};
+
+static int intel_th_msc_probe(struct intel_th_device *thdev)
+{
+	struct device *dev = &thdev->dev;
+	struct resource *res;
+	struct msc *msc;
+	void __iomem *base;
+	int err;
+
+	res = intel_th_device_get_resource(thdev, IORESOURCE_MEM, 0);
+	if (!res)
+		return -ENODEV;
+
+	base = devm_ioremap(dev, res->start, resource_size(res));
+	if (!base)
+		return -ENOMEM;
+
+	msc = devm_kzalloc(dev, sizeof(*msc), GFP_KERNEL);
+	if (!msc)
+		return -ENOMEM;
+
+	res = intel_th_device_get_resource(thdev, IORESOURCE_IRQ, 1);
+	if (!res)
+		msc->do_irq = 1;
+
+	if (INTEL_TH_CAP(to_intel_th(thdev), multi_is_broken))
+		msc->multi_is_broken = 1;
+
+	msc->index = thdev->id;
+
+	msc->thdev = thdev;
+	msc->reg_base = base + msc->index * 0x100;
+	msc->msu_base = base;
+
+	INIT_WORK(&msc->work, msc_work);
+	err = intel_th_msc_init(msc);
+	if (err)
+		return err;
+
+	dev_set_drvdata(dev, msc);
+
+	return 0;
+}
+
+static void intel_th_msc_remove(struct intel_th_device *thdev)
+{
+	struct msc *msc = dev_get_drvdata(&thdev->dev);
+	int ret;
+
+	intel_th_msc_deactivate(thdev);
+
+	/*
+	 * Buffers should not be used at this point except if the
+	 * output character device is still open and the parent
+	 * device gets detached from its bus, which is a FIXME.
+	 */
+	ret = msc_buffer_free_unless_used(msc);
+	WARN_ON_ONCE(ret);
+}
+
+static struct intel_th_driver intel_th_msc_driver = {
+	.probe	= intel_th_msc_probe,
+	.remove	= intel_th_msc_remove,
+	.irq		= intel_th_msc_interrupt,
+	.wait_empty	= intel_th_msc_wait_empty,
+	.activate	= intel_th_msc_activate,
+	.deactivate	= intel_th_msc_deactivate,
+	.fops	= &intel_th_msc_fops,
+	.attr_group	= &msc_output_group,
+	.driver	= {
+		.name	= "msc",
+		.owner	= THIS_MODULE,
+	},
+};
+
+module_driver(intel_th_msc_driver,
+	      intel_th_driver_register,
+	      intel_th_driver_unregister);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("Intel(R) Trace Hub Memory Storage Unit driver");
+MODULE_AUTHOR("Alexander Shishkin <alexander.shishkin@linux.intel.com>");
diff --git a/drivers/hwtracing/intel_th/msu.h b/drivers/hwtracing/intel_th/msu.h
new file mode 100644
index 0000000000..e771f509bd
--- /dev/null
+++ b/drivers/hwtracing/intel_th/msu.h
@@ -0,0 +1,118 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Intel(R) Trace Hub Memory Storage Unit (MSU) data structures
+ *
+ * Copyright (C) 2014-2015 Intel Corporation.
+ */
+
+#ifndef __INTEL_TH_MSU_H__
+#define __INTEL_TH_MSU_H__
+
+enum {
+	REG_MSU_MSUPARAMS	= 0x0000,
+	REG_MSU_MSUSTS		= 0x0008,
+	REG_MSU_MINTCTL		= 0x0004, /* MSU-global interrupt control */
+	REG_MSU_MSC0CTL		= 0x0100, /* MSC0 control */
+	REG_MSU_MSC0STS		= 0x0104, /* MSC0 status */
+	REG_MSU_MSC0BAR		= 0x0108, /* MSC0 output base address */
+	REG_MSU_MSC0SIZE	= 0x010c, /* MSC0 output size */
+	REG_MSU_MSC0MWP		= 0x0110, /* MSC0 write pointer */
+	REG_MSU_MSC0NWSA	= 0x011c, /* MSC0 next window start address */
+
+	REG_MSU_MSC1CTL		= 0x0200, /* MSC1 control */
+	REG_MSU_MSC1STS		= 0x0204, /* MSC1 status */
+	REG_MSU_MSC1BAR		= 0x0208, /* MSC1 output base address */
+	REG_MSU_MSC1SIZE	= 0x020c, /* MSC1 output size */
+	REG_MSU_MSC1MWP		= 0x0210, /* MSC1 write pointer */
+	REG_MSU_MSC1NWSA	= 0x021c, /* MSC1 next window start address */
+};
+
+/* MSUSTS bits */
+#define MSUSTS_MSU_INT	BIT(0)
+#define MSUSTS_MSC0BLAST	BIT(16)
+#define MSUSTS_MSC1BLAST	BIT(24)
+
+/* MSCnCTL bits */
+#define MSC_EN		BIT(0)
+#define MSC_WRAPEN	BIT(1)
+#define MSC_RD_HDR_OVRD	BIT(2)
+#define MSC_MODE	(BIT(4) | BIT(5))
+#define MSC_LEN		(BIT(8) | BIT(9) | BIT(10))
+
+/* MINTCTL bits */
+#define MICDE		BIT(0)
+#define M0BLIE		BIT(16)
+#define M1BLIE		BIT(24)
+
+/* MSCnSTS bits */
+#define MSCSTS_WRAPSTAT	BIT(1)	/* Wrap occurred */
+#define MSCSTS_PLE	BIT(2)	/* Pipeline Empty */
+
+/*
+ * Multiblock/multiwindow block descriptor
+ */
+struct msc_block_desc {
+	u32	sw_tag;
+	u32	block_sz;
+	u32	next_blk;
+	u32	next_win;
+	u32	res0[4];
+	u32	hw_tag;
+	u32	valid_dw;
+	u32	ts_low;
+	u32	ts_high;
+	u32	res1[4];
+} __packed;
+
+#define MSC_BDESC	sizeof(struct msc_block_desc)
+#define DATA_IN_PAGE	(PAGE_SIZE - MSC_BDESC)
+
+/* MSC multiblock sw tag bits */
+#define MSC_SW_TAG_LASTBLK	BIT(0)
+#define MSC_SW_TAG_LASTWIN	BIT(1)
+
+/* MSC multiblock hw tag bits */
+#define MSC_HW_TAG_TRIGGER	BIT(0)
+#define MSC_HW_TAG_BLOCKWRAP	BIT(1)
+#define MSC_HW_TAG_WINWRAP	BIT(2)
+#define MSC_HW_TAG_ENDBIT	BIT(3)
+
+static inline unsigned long msc_data_sz(struct msc_block_desc *bdesc)
+{
+	if (!bdesc->valid_dw)
+		return 0;
+
+	return bdesc->valid_dw * 4 - MSC_BDESC;
+}
+
+static inline unsigned long msc_total_sz(struct msc_block_desc *bdesc)
+{
+	return bdesc->valid_dw * 4;
+}
+
+static inline unsigned long msc_block_sz(struct msc_block_desc *bdesc)
+{
+	return bdesc->block_sz * 64 - MSC_BDESC;
+}
+
+static inline bool msc_block_wrapped(struct msc_block_desc *bdesc)
+{
+	if (bdesc->hw_tag & (MSC_HW_TAG_BLOCKWRAP | MSC_HW_TAG_WINWRAP))
+		return true;
+
+	return false;
+}
+
+static inline bool msc_block_last_written(struct msc_block_desc *bdesc)
+{
+	if ((bdesc->hw_tag & MSC_HW_TAG_ENDBIT) ||
+	    (msc_data_sz(bdesc) != msc_block_sz(bdesc)))
+		return true;
+
+	return false;
+}
+
+/* waiting for Pipeline Empty bit(s) to assert for MSC */
+#define MSC_PLE_WAITLOOP_DEPTH	10000
+
+#endif /* __INTEL_TH_MSU_H__ */
diff --git a/drivers/hwtracing/intel_th/pci.c b/drivers/hwtracing/intel_th/pci.c
new file mode 100644
index 0000000000..147d338c19
--- /dev/null
+++ b/drivers/hwtracing/intel_th/pci.c
@@ -0,0 +1,328 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Intel(R) Trace Hub pci driver
+ *
+ * Copyright (C) 2014-2015 Intel Corporation.
+ */
+
+#define pr_fmt(fmt)	KBUILD_MODNAME ": " fmt
+
+#include <linux/types.h>
+#include <linux/module.h>
+#include <linux/device.h>
+#include <linux/sysfs.h>
+#include <linux/pci.h>
+
+#include "intel_th.h"
+
+#define DRIVER_NAME "intel_th_pci"
+
+enum {
+	TH_PCI_CONFIG_BAR	= 0,
+	TH_PCI_STH_SW_BAR	= 2,
+	TH_PCI_RTIT_BAR		= 4,
+};
+
+#define BAR_MASK (BIT(TH_PCI_CONFIG_BAR) | BIT(TH_PCI_STH_SW_BAR))
+
+#define PCI_REG_NPKDSC	0x80
+#define NPKDSC_TSACT	BIT(5)
+
+static int intel_th_pci_activate(struct intel_th *th)
+{
+	struct pci_dev *pdev = to_pci_dev(th->dev);
+	u32 npkdsc;
+	int err;
+
+	if (!INTEL_TH_CAP(th, tscu_enable))
+		return 0;
+
+	err = pci_read_config_dword(pdev, PCI_REG_NPKDSC, &npkdsc);
+	if (!err) {
+		npkdsc |= NPKDSC_TSACT;
+		err = pci_write_config_dword(pdev, PCI_REG_NPKDSC, npkdsc);
+	}
+
+	if (err)
+		dev_err(&pdev->dev, "failed to read NPKDSC register\n");
+
+	return err;
+}
+
+static void intel_th_pci_deactivate(struct intel_th *th)
+{
+	struct pci_dev *pdev = to_pci_dev(th->dev);
+	u32 npkdsc;
+	int err;
+
+	if (!INTEL_TH_CAP(th, tscu_enable))
+		return;
+
+	err = pci_read_config_dword(pdev, PCI_REG_NPKDSC, &npkdsc);
+	if (!err) {
+		npkdsc |= NPKDSC_TSACT;
+		err = pci_write_config_dword(pdev, PCI_REG_NPKDSC, npkdsc);
+	}
+
+	if (err)
+		dev_err(&pdev->dev, "failed to read NPKDSC register\n");
+}
+
+static int intel_th_pci_probe(struct pci_dev *pdev,
+			      const struct pci_device_id *id)
+{
+	const struct intel_th_drvdata *drvdata = (void *)id->driver_data;
+	struct resource resource[TH_MMIO_END + TH_NVEC_MAX] = {
+		[TH_MMIO_CONFIG]	= pdev->resource[TH_PCI_CONFIG_BAR],
+		[TH_MMIO_SW]		= pdev->resource[TH_PCI_STH_SW_BAR],
+	};
+	int err, r = TH_MMIO_SW + 1, i;
+	struct intel_th *th;
+
+	err = pcim_enable_device(pdev);
+	if (err)
+		return err;
+
+	err = pcim_iomap_regions_request_all(pdev, BAR_MASK, DRIVER_NAME);
+	if (err)
+		return err;
+
+	if (pdev->resource[TH_PCI_RTIT_BAR].start) {
+		resource[TH_MMIO_RTIT] = pdev->resource[TH_PCI_RTIT_BAR];
+		r++;
+	}
+
+	err = pci_alloc_irq_vectors(pdev, 1, 8, PCI_IRQ_ALL_TYPES);
+	if (err > 0)
+		for (i = 0; i < err; i++, r++) {
+			resource[r].flags = IORESOURCE_IRQ;
+			resource[r].start = pci_irq_vector(pdev, i);
+		}
+
+	th = intel_th_alloc(&pdev->dev, drvdata, resource, r);
+	if (IS_ERR(th)) {
+		err = PTR_ERR(th);
+		goto err_free_irq;
+	}
+
+	th->activate   = intel_th_pci_activate;
+	th->deactivate = intel_th_pci_deactivate;
+
+	pci_set_master(pdev);
+
+	return 0;
+
+err_free_irq:
+	pci_free_irq_vectors(pdev);
+	return err;
+}
+
+static void intel_th_pci_remove(struct pci_dev *pdev)
+{
+	struct intel_th *th = pci_get_drvdata(pdev);
+
+	intel_th_free(th);
+
+	pci_free_irq_vectors(pdev);
+}
+
+static const struct intel_th_drvdata intel_th_1x_multi_is_broken = {
+	.multi_is_broken	= 1,
+};
+
+static const struct intel_th_drvdata intel_th_2x = {
+	.tscu_enable	= 1,
+	.has_mintctl	= 1,
+};
+
+static const struct pci_device_id intel_th_pci_id_table[] = {
+	{
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x9d26),
+		.driver_data = (kernel_ulong_t)0,
+	},
+	{
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0xa126),
+		.driver_data = (kernel_ulong_t)0,
+	},
+	{
+		/* Apollo Lake */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x5a8e),
+		.driver_data = (kernel_ulong_t)0,
+	},
+	{
+		/* Broxton */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x0a80),
+		.driver_data = (kernel_ulong_t)0,
+	},
+	{
+		/* Broxton B-step */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x1a8e),
+		.driver_data = (kernel_ulong_t)0,
+	},
+	{
+		/* Kaby Lake PCH-H */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0xa2a6),
+		.driver_data = (kernel_ulong_t)&intel_th_1x_multi_is_broken,
+	},
+	{
+		/* Denverton */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x19e1),
+		.driver_data = (kernel_ulong_t)0,
+	},
+	{
+		/* Lewisburg PCH */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0xa1a6),
+		.driver_data = (kernel_ulong_t)0,
+	},
+	{
+		/* Lewisburg PCH */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0xa226),
+		.driver_data = (kernel_ulong_t)0,
+	},
+	{
+		/* Gemini Lake */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x318e),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Cannon Lake H */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0xa326),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Cannon Lake LP */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x9da6),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Cedar Fork PCH */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x18e1),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Ice Lake PCH */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x34a6),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Comet Lake */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x02a6),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Comet Lake PCH */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x06a6),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Comet Lake PCH-V */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0xa3a6),
+		.driver_data = (kernel_ulong_t)&intel_th_1x_multi_is_broken,
+	},
+	{
+		/* Ice Lake NNPI */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x45c5),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Ice Lake CPU */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x8a29),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Tiger Lake CPU */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x9a33),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Tiger Lake PCH */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0xa0a6),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Tiger Lake PCH-H */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x43a6),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Jasper Lake PCH */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x4da6),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Jasper Lake CPU */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x4e29),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Elkhart Lake CPU */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x4529),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Elkhart Lake */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x4b26),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Emmitsburg PCH */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x1bcc),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Alder Lake */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7aa6),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Alder Lake-P */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x51a6),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Alder Lake-M */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x54a6),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Meteor Lake-P */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7e24),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Raptor Lake-S */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x7a26),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Raptor Lake-S CPU */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0xa76f),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Alder Lake CPU */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x466f),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{
+		/* Rocket Lake CPU */
+		PCI_DEVICE(PCI_VENDOR_ID_INTEL, 0x4c19),
+		.driver_data = (kernel_ulong_t)&intel_th_2x,
+	},
+	{ 0 },
+};
+
+MODULE_DEVICE_TABLE(pci, intel_th_pci_id_table);
+
+static struct pci_driver intel_th_pci_driver = {
+	.name		= DRIVER_NAME,
+	.id_table	= intel_th_pci_id_table,
+	.probe		= intel_th_pci_probe,
+	.remove		= intel_th_pci_remove,
+};
+
+module_pci_driver(intel_th_pci_driver);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("Intel(R) Trace Hub PCI controller driver");
+MODULE_AUTHOR("Alexander Shishkin <alexander.shishkin@intel.com>");
diff --git a/drivers/hwtracing/intel_th/pti.c b/drivers/hwtracing/intel_th/pti.c
new file mode 100644
index 0000000000..09132ab8bc
--- /dev/null
+++ b/drivers/hwtracing/intel_th/pti.c
@@ -0,0 +1,343 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Intel(R) Trace Hub PTI output driver
+ *
+ * Copyright (C) 2014-2016 Intel Corporation.
+ */
+
+#define pr_fmt(fmt)	KBUILD_MODNAME ": " fmt
+
+#include <linux/types.h>
+#include <linux/module.h>
+#include <linux/device.h>
+#include <linux/sizes.h>
+#include <linux/printk.h>
+#include <linux/slab.h>
+#include <linux/mm.h>
+#include <linux/io.h>
+
+#include "intel_th.h"
+#include "pti.h"
+
+struct pti_device {
+	void __iomem		*base;
+	struct intel_th_device	*thdev;
+	unsigned int		mode;
+	unsigned int		freeclk;
+	unsigned int		clkdiv;
+	unsigned int		patgen;
+	unsigned int		lpp_dest_mask;
+	unsigned int		lpp_dest;
+};
+
+/* map PTI widths to MODE settings of PTI_CTL register */
+static const unsigned int pti_mode[] = {
+	0, 4, 8, 0, 12, 0, 0, 0, 16, 0, 0, 0, 0, 0, 0, 0,
+};
+
+static int pti_width_mode(unsigned int width)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(pti_mode); i++)
+		if (pti_mode[i] == width)
+			return i;
+
+	return -EINVAL;
+}
+
+static ssize_t mode_show(struct device *dev, struct device_attribute *attr,
+			 char *buf)
+{
+	struct pti_device *pti = dev_get_drvdata(dev);
+
+	return scnprintf(buf, PAGE_SIZE, "%d\n", pti_mode[pti->mode]);
+}
+
+static ssize_t mode_store(struct device *dev, struct device_attribute *attr,
+			  const char *buf, size_t size)
+{
+	struct pti_device *pti = dev_get_drvdata(dev);
+	unsigned long val;
+	int ret;
+
+	ret = kstrtoul(buf, 10, &val);
+	if (ret)
+		return ret;
+
+	ret = pti_width_mode(val);
+	if (ret < 0)
+		return ret;
+
+	pti->mode = ret;
+
+	return size;
+}
+
+static DEVICE_ATTR_RW(mode);
+
+static ssize_t
+freerunning_clock_show(struct device *dev, struct device_attribute *attr,
+		       char *buf)
+{
+	struct pti_device *pti = dev_get_drvdata(dev);
+
+	return scnprintf(buf, PAGE_SIZE, "%d\n", pti->freeclk);
+}
+
+static ssize_t
+freerunning_clock_store(struct device *dev, struct device_attribute *attr,
+			const char *buf, size_t size)
+{
+	struct pti_device *pti = dev_get_drvdata(dev);
+	unsigned long val;
+	int ret;
+
+	ret = kstrtoul(buf, 10, &val);
+	if (ret)
+		return ret;
+
+	pti->freeclk = !!val;
+
+	return size;
+}
+
+static DEVICE_ATTR_RW(freerunning_clock);
+
+static ssize_t
+clock_divider_show(struct device *dev, struct device_attribute *attr,
+		   char *buf)
+{
+	struct pti_device *pti = dev_get_drvdata(dev);
+
+	return scnprintf(buf, PAGE_SIZE, "%d\n", 1u << pti->clkdiv);
+}
+
+static ssize_t
+clock_divider_store(struct device *dev, struct device_attribute *attr,
+		    const char *buf, size_t size)
+{
+	struct pti_device *pti = dev_get_drvdata(dev);
+	unsigned long val;
+	int ret;
+
+	ret = kstrtoul(buf, 10, &val);
+	if (ret)
+		return ret;
+
+	if (!is_power_of_2(val) || val > 8 || !val)
+		return -EINVAL;
+
+	pti->clkdiv = val;
+
+	return size;
+}
+
+static DEVICE_ATTR_RW(clock_divider);
+
+static struct attribute *pti_output_attrs[] = {
+	&dev_attr_mode.attr,
+	&dev_attr_freerunning_clock.attr,
+	&dev_attr_clock_divider.attr,
+	NULL,
+};
+
+static const struct attribute_group pti_output_group = {
+	.attrs	= pti_output_attrs,
+};
+
+static int intel_th_pti_activate(struct intel_th_device *thdev)
+{
+	struct pti_device *pti = dev_get_drvdata(&thdev->dev);
+	u32 ctl = PTI_EN;
+
+	if (pti->patgen)
+		ctl |= pti->patgen << __ffs(PTI_PATGENMODE);
+	if (pti->freeclk)
+		ctl |= PTI_FCEN;
+	ctl |= pti->mode << __ffs(PTI_MODE);
+	ctl |= pti->clkdiv << __ffs(PTI_CLKDIV);
+	ctl |= pti->lpp_dest << __ffs(LPP_DEST);
+
+	iowrite32(ctl, pti->base + REG_PTI_CTL);
+
+	intel_th_trace_enable(thdev);
+
+	return 0;
+}
+
+static void intel_th_pti_deactivate(struct intel_th_device *thdev)
+{
+	struct pti_device *pti = dev_get_drvdata(&thdev->dev);
+
+	intel_th_trace_disable(thdev);
+
+	iowrite32(0, pti->base + REG_PTI_CTL);
+}
+
+static void read_hw_config(struct pti_device *pti)
+{
+	u32 ctl = ioread32(pti->base + REG_PTI_CTL);
+
+	pti->mode	= (ctl & PTI_MODE) >> __ffs(PTI_MODE);
+	pti->clkdiv	= (ctl & PTI_CLKDIV) >> __ffs(PTI_CLKDIV);
+	pti->freeclk	= !!(ctl & PTI_FCEN);
+
+	if (!pti_mode[pti->mode])
+		pti->mode = pti_width_mode(4);
+	if (!pti->clkdiv)
+		pti->clkdiv = 1;
+
+	if (pti->thdev->output.type == GTH_LPP) {
+		if (ctl & LPP_PTIPRESENT)
+			pti->lpp_dest_mask |= LPP_DEST_PTI;
+		if (ctl & LPP_BSSBPRESENT)
+			pti->lpp_dest_mask |= LPP_DEST_EXI;
+		if (ctl & LPP_DEST)
+			pti->lpp_dest = 1;
+	}
+}
+
+static int intel_th_pti_probe(struct intel_th_device *thdev)
+{
+	struct device *dev = &thdev->dev;
+	struct resource *res;
+	struct pti_device *pti;
+	void __iomem *base;
+
+	res = intel_th_device_get_resource(thdev, IORESOURCE_MEM, 0);
+	if (!res)
+		return -ENODEV;
+
+	base = devm_ioremap(dev, res->start, resource_size(res));
+	if (!base)
+		return -ENOMEM;
+
+	pti = devm_kzalloc(dev, sizeof(*pti), GFP_KERNEL);
+	if (!pti)
+		return -ENOMEM;
+
+	pti->thdev = thdev;
+	pti->base = base;
+
+	read_hw_config(pti);
+
+	dev_set_drvdata(dev, pti);
+
+	return 0;
+}
+
+static void intel_th_pti_remove(struct intel_th_device *thdev)
+{
+}
+
+static struct intel_th_driver intel_th_pti_driver = {
+	.probe	= intel_th_pti_probe,
+	.remove	= intel_th_pti_remove,
+	.activate	= intel_th_pti_activate,
+	.deactivate	= intel_th_pti_deactivate,
+	.attr_group	= &pti_output_group,
+	.driver	= {
+		.name	= "pti",
+		.owner	= THIS_MODULE,
+	},
+};
+
+static const char * const lpp_dest_str[] = { "pti", "exi" };
+
+static ssize_t lpp_dest_show(struct device *dev, struct device_attribute *attr,
+			     char *buf)
+{
+	struct pti_device *pti = dev_get_drvdata(dev);
+	ssize_t ret = 0;
+	int i;
+
+	for (i = ARRAY_SIZE(lpp_dest_str) - 1; i >= 0; i--) {
+		const char *fmt = pti->lpp_dest == i ? "[%s] " : "%s ";
+
+		if (!(pti->lpp_dest_mask & BIT(i)))
+			continue;
+
+		ret += scnprintf(buf + ret, PAGE_SIZE - ret,
+				 fmt, lpp_dest_str[i]);
+	}
+
+	if (ret)
+		buf[ret - 1] = '\n';
+
+	return ret;
+}
+
+static ssize_t lpp_dest_store(struct device *dev, struct device_attribute *attr,
+			      const char *buf, size_t size)
+{
+	struct pti_device *pti = dev_get_drvdata(dev);
+	int i;
+
+	i = sysfs_match_string(lpp_dest_str, buf);
+	if (i < 0)
+		return i;
+
+	if (!(pti->lpp_dest_mask & BIT(i)))
+		return -EINVAL;
+
+	pti->lpp_dest = i;
+	return size;
+}
+
+static DEVICE_ATTR_RW(lpp_dest);
+
+static struct attribute *lpp_output_attrs[] = {
+	&dev_attr_mode.attr,
+	&dev_attr_freerunning_clock.attr,
+	&dev_attr_clock_divider.attr,
+	&dev_attr_lpp_dest.attr,
+	NULL,
+};
+
+static const struct attribute_group lpp_output_group = {
+	.attrs	= lpp_output_attrs,
+};
+
+static struct intel_th_driver intel_th_lpp_driver = {
+	.probe		= intel_th_pti_probe,
+	.remove		= intel_th_pti_remove,
+	.activate	= intel_th_pti_activate,
+	.deactivate	= intel_th_pti_deactivate,
+	.attr_group	= &lpp_output_group,
+	.driver	= {
+		.name	= "lpp",
+		.owner	= THIS_MODULE,
+	},
+};
+
+static int __init intel_th_pti_lpp_init(void)
+{
+	int err;
+
+	err = intel_th_driver_register(&intel_th_pti_driver);
+	if (err)
+		return err;
+
+	err = intel_th_driver_register(&intel_th_lpp_driver);
+	if (err) {
+		intel_th_driver_unregister(&intel_th_pti_driver);
+		return err;
+	}
+
+	return 0;
+}
+
+module_init(intel_th_pti_lpp_init);
+
+static void __exit intel_th_pti_lpp_exit(void)
+{
+	intel_th_driver_unregister(&intel_th_pti_driver);
+	intel_th_driver_unregister(&intel_th_lpp_driver);
+}
+
+module_exit(intel_th_pti_lpp_exit);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("Intel(R) Trace Hub PTI/LPP output driver");
+MODULE_AUTHOR("Alexander Shishkin <alexander.shishkin@linux.intel.com>");
diff --git a/drivers/hwtracing/intel_th/pti.h b/drivers/hwtracing/intel_th/pti.h
new file mode 100644
index 0000000000..7dfc043133
--- /dev/null
+++ b/drivers/hwtracing/intel_th/pti.h
@@ -0,0 +1,29 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Intel(R) Trace Hub PTI output data structures
+ *
+ * Copyright (C) 2014-2015 Intel Corporation.
+ */
+
+#ifndef __INTEL_TH_STH_H__
+#define __INTEL_TH_STH_H__
+
+enum {
+	REG_PTI_CTL	= 0x1c00,
+};
+
+#define PTI_EN		BIT(0)
+#define PTI_FCEN	BIT(1)
+#define PTI_MODE	0xf0
+#define LPP_PTIPRESENT	BIT(8)
+#define LPP_BSSBPRESENT	BIT(9)
+#define PTI_CLKDIV	0x000f0000
+#define PTI_PATGENMODE	0x00f00000
+#define LPP_DEST	BIT(25)
+#define LPP_BSSBACT	BIT(30)
+#define LPP_LPPBUSY	BIT(31)
+
+#define LPP_DEST_PTI	BIT(0)
+#define LPP_DEST_EXI	BIT(1)
+
+#endif /* __INTEL_TH_STH_H__ */
diff --git a/drivers/hwtracing/intel_th/sth.c b/drivers/hwtracing/intel_th/sth.c
new file mode 100644
index 0000000000..9ca8c4e045
--- /dev/null
+++ b/drivers/hwtracing/intel_th/sth.c
@@ -0,0 +1,261 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Intel(R) Trace Hub Software Trace Hub support
+ *
+ * Copyright (C) 2014-2015 Intel Corporation.
+ */
+
+#define pr_fmt(fmt)	KBUILD_MODNAME ": " fmt
+
+#include <linux/types.h>
+#include <linux/module.h>
+#include <linux/device.h>
+#include <linux/io.h>
+#include <linux/mm.h>
+#include <linux/slab.h>
+#include <linux/stm.h>
+
+#include "intel_th.h"
+#include "sth.h"
+
+struct sth_device {
+	void __iomem	*base;
+	void __iomem	*channels;
+	phys_addr_t	channels_phys;
+	struct device	*dev;
+	struct stm_data	stm;
+	unsigned int	sw_nmasters;
+};
+
+static struct intel_th_channel __iomem *
+sth_channel(struct sth_device *sth, unsigned int master, unsigned int channel)
+{
+	struct intel_th_channel __iomem *sw_map = sth->channels;
+
+	return &sw_map[(master - sth->stm.sw_start) * sth->stm.sw_nchannels +
+		       channel];
+}
+
+static void sth_iowrite(void __iomem *dest, const unsigned char *payload,
+			unsigned int size)
+{
+	switch (size) {
+#ifdef CONFIG_64BIT
+	case 8:
+		writeq_relaxed(*(u64 *)payload, dest);
+		break;
+#endif
+	case 4:
+		writel_relaxed(*(u32 *)payload, dest);
+		break;
+	case 2:
+		writew_relaxed(*(u16 *)payload, dest);
+		break;
+	case 1:
+		writeb_relaxed(*(u8 *)payload, dest);
+		break;
+	default:
+		break;
+	}
+}
+
+static ssize_t notrace sth_stm_packet(struct stm_data *stm_data,
+				      unsigned int master,
+				      unsigned int channel,
+				      unsigned int packet,
+				      unsigned int flags,
+				      unsigned int size,
+				      const unsigned char *payload)
+{
+	struct sth_device *sth = container_of(stm_data, struct sth_device, stm);
+	struct intel_th_channel __iomem *out =
+		sth_channel(sth, master, channel);
+	u64 __iomem *outp = &out->Dn;
+	unsigned long reg = REG_STH_TRIG;
+
+#ifndef CONFIG_64BIT
+	if (size > 4)
+		size = 4;
+#endif
+
+	size = rounddown_pow_of_two(size);
+
+	switch (packet) {
+	/* Global packets (GERR, XSYNC, TRIG) are sent with register writes */
+	case STP_PACKET_GERR:
+		reg += 4;
+		fallthrough;
+
+	case STP_PACKET_XSYNC:
+		reg += 8;
+		fallthrough;
+
+	case STP_PACKET_TRIG:
+		if (flags & STP_PACKET_TIMESTAMPED)
+			reg += 4;
+		writeb_relaxed(*payload, sth->base + reg);
+		break;
+
+	case STP_PACKET_MERR:
+		if (size > 4)
+			size = 4;
+
+		sth_iowrite(&out->MERR, payload, size);
+		break;
+
+	case STP_PACKET_FLAG:
+		if (flags & STP_PACKET_TIMESTAMPED)
+			outp = (u64 __iomem *)&out->FLAG_TS;
+		else
+			outp = (u64 __iomem *)&out->FLAG;
+
+		size = 0;
+		writeb_relaxed(0, outp);
+		break;
+
+	case STP_PACKET_USER:
+		if (flags & STP_PACKET_TIMESTAMPED)
+			outp = &out->USER_TS;
+		else
+			outp = &out->USER;
+		sth_iowrite(outp, payload, size);
+		break;
+
+	case STP_PACKET_DATA:
+		outp = &out->Dn;
+
+		if (flags & STP_PACKET_TIMESTAMPED)
+			outp += 2;
+		if (flags & STP_PACKET_MARKED)
+			outp++;
+
+		sth_iowrite(outp, payload, size);
+		break;
+	default:
+		return -ENOTSUPP;
+	}
+
+	return size;
+}
+
+static phys_addr_t
+sth_stm_mmio_addr(struct stm_data *stm_data, unsigned int master,
+		  unsigned int channel, unsigned int nr_chans)
+{
+	struct sth_device *sth = container_of(stm_data, struct sth_device, stm);
+	phys_addr_t addr;
+
+	master -= sth->stm.sw_start;
+	addr = sth->channels_phys + (master * sth->stm.sw_nchannels + channel) *
+		sizeof(struct intel_th_channel);
+
+	if (offset_in_page(addr) ||
+	    offset_in_page(nr_chans * sizeof(struct intel_th_channel)))
+		return 0;
+
+	return addr;
+}
+
+static int sth_stm_link(struct stm_data *stm_data, unsigned int master,
+			 unsigned int channel)
+{
+	struct sth_device *sth = container_of(stm_data, struct sth_device, stm);
+
+	return intel_th_set_output(to_intel_th_device(sth->dev), master);
+}
+
+static int intel_th_sw_init(struct sth_device *sth)
+{
+	u32 reg;
+
+	reg = ioread32(sth->base + REG_STH_STHCAP1);
+	sth->stm.sw_nchannels = reg & 0xff;
+
+	reg = ioread32(sth->base + REG_STH_STHCAP0);
+	sth->stm.sw_start = reg & 0xffff;
+	sth->stm.sw_end = reg >> 16;
+
+	sth->sw_nmasters = sth->stm.sw_end - sth->stm.sw_start;
+	dev_dbg(sth->dev, "sw_start: %x sw_end: %x masters: %x nchannels: %x\n",
+		sth->stm.sw_start, sth->stm.sw_end, sth->sw_nmasters,
+		sth->stm.sw_nchannels);
+
+	return 0;
+}
+
+static int intel_th_sth_probe(struct intel_th_device *thdev)
+{
+	struct device *dev = &thdev->dev;
+	struct sth_device *sth;
+	struct resource *res;
+	void __iomem *base, *channels;
+	int err;
+
+	res = intel_th_device_get_resource(thdev, IORESOURCE_MEM, 0);
+	if (!res)
+		return -ENODEV;
+
+	base = devm_ioremap(dev, res->start, resource_size(res));
+	if (!base)
+		return -ENOMEM;
+
+	res = intel_th_device_get_resource(thdev, IORESOURCE_MEM, 1);
+	if (!res)
+		return -ENODEV;
+
+	channels = devm_ioremap(dev, res->start, resource_size(res));
+	if (!channels)
+		return -ENOMEM;
+
+	sth = devm_kzalloc(dev, sizeof(*sth), GFP_KERNEL);
+	if (!sth)
+		return -ENOMEM;
+
+	sth->dev = dev;
+	sth->base = base;
+	sth->channels = channels;
+	sth->channels_phys = res->start;
+	sth->stm.name = dev_name(dev);
+	sth->stm.packet = sth_stm_packet;
+	sth->stm.mmio_addr = sth_stm_mmio_addr;
+	sth->stm.sw_mmiosz = sizeof(struct intel_th_channel);
+	sth->stm.link = sth_stm_link;
+
+	err = intel_th_sw_init(sth);
+	if (err)
+		return err;
+
+	err = stm_register_device(dev, &sth->stm, THIS_MODULE);
+	if (err) {
+		dev_err(dev, "stm_register_device failed\n");
+		return err;
+	}
+
+	dev_set_drvdata(dev, sth);
+
+	return 0;
+}
+
+static void intel_th_sth_remove(struct intel_th_device *thdev)
+{
+	struct sth_device *sth = dev_get_drvdata(&thdev->dev);
+
+	stm_unregister_device(&sth->stm);
+}
+
+static struct intel_th_driver intel_th_sth_driver = {
+	.probe	= intel_th_sth_probe,
+	.remove	= intel_th_sth_remove,
+	.driver	= {
+		.name	= "sth",
+		.owner	= THIS_MODULE,
+	},
+};
+
+module_driver(intel_th_sth_driver,
+	      intel_th_driver_register,
+	      intel_th_driver_unregister);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("Intel(R) Trace Hub Software Trace Hub driver");
+MODULE_AUTHOR("Alexander Shishkin <alexander.shishkin@intel.com>");
diff --git a/drivers/hwtracing/intel_th/sth.h b/drivers/hwtracing/intel_th/sth.h
new file mode 100644
index 0000000000..f97fc0c517
--- /dev/null
+++ b/drivers/hwtracing/intel_th/sth.h
@@ -0,0 +1,34 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Intel(R) Trace Hub Software Trace Hub (STH) data structures
+ *
+ * Copyright (C) 2014-2015 Intel Corporation.
+ */
+
+#ifndef __INTEL_TH_STH_H__
+#define __INTEL_TH_STH_H__
+
+enum {
+	REG_STH_STHCAP0		= 0x0000, /* capabilities pt1 */
+	REG_STH_STHCAP1		= 0x0004, /* capabilities pt2 */
+	REG_STH_TRIG		= 0x0008, /* TRIG packet payload */
+	REG_STH_TRIG_TS		= 0x000c, /* TRIG_TS packet payload */
+	REG_STH_XSYNC		= 0x0010, /* XSYNC packet payload */
+	REG_STH_XSYNC_TS	= 0x0014, /* XSYNC_TS packet payload */
+	REG_STH_GERR		= 0x0018, /* GERR packet payload */
+};
+
+struct intel_th_channel {
+	u64	Dn;
+	u64	DnM;
+	u64	DnTS;
+	u64	DnMTS;
+	u64	USER;
+	u64	USER_TS;
+	u32	FLAG;
+	u32	FLAG_TS;
+	u32	MERR;
+	u32	__unused;
+} __packed;
+
+#endif /* __INTEL_TH_STH_H__ */