Adding upstream version 5.10.209.upstream/5.10.209upstream

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

126 files changed, 100506 insertions, 0 deletions

diff --git a/drivers/mmc/host/Kconfig b/drivers/mmc/host/Kconfig
new file mode 100644
index 000000000..8fe4a0fd6
--- /dev/null
+++ b/drivers/mmc/host/Kconfig
@@ -0,0 +1,1107 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# MMC/SD host controller drivers
+#
+
+comment "MMC/SD/SDIO Host Controller Drivers"
+
+config MMC_DEBUG
+	bool "MMC host drivers debugging"
+	depends on MMC != n
+	help
+	  This is an option for use by developers; most people should
+	  say N here. This enables MMC host driver debugging. And further
+	  added host drivers please don't invent their private macro for
+	  debugging.
+
+config MMC_ARMMMCI
+	tristate "ARM AMBA Multimedia Card Interface support"
+	depends on ARM_AMBA
+	help
+	  This selects the ARM(R) AMBA(R) PrimeCell Multimedia Card
+	  Interface (PL180 and PL181) support.  If you have an ARM(R)
+	  platform with a Multimedia Card slot, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_QCOM_DML
+	bool "Qualcomm Data Mover for SD Card Controller"
+	depends on MMC_ARMMMCI && QCOM_BAM_DMA
+	default y
+	help
+	  This selects the Qualcomm Data Mover lite/local on SD Card controller.
+	  This option will enable the dma to work correctly, if you are using
+	  Qcom SOCs and MMC, you would probably need this option to get DMA working.
+
+	  if unsure, say N.
+
+config MMC_STM32_SDMMC
+	bool "STMicroelectronics STM32 SDMMC Controller"
+	depends on MMC_ARMMMCI
+	default y
+	help
+	  This selects the STMicroelectronics STM32 SDMMC host controller.
+	  If you have a STM32 sdmmc host with internal DMA say Y here.
+
+	  If unsure, say N.
+
+config MMC_PXA
+	tristate "Intel PXA25x/26x/27x Multimedia Card Interface support"
+	depends on ARCH_PXA
+	help
+	  This selects the Intel(R) PXA(R) Multimedia card Interface.
+	  If you have a PXA(R) platform with a Multimedia Card slot,
+	  say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI
+	tristate "Secure Digital Host Controller Interface support"
+	depends on HAS_DMA
+	help
+	  This selects the generic Secure Digital Host Controller Interface.
+	  It is used by manufacturers such as Texas Instruments(R), Ricoh(R)
+	  and Toshiba(R). Most controllers found in laptops are of this type.
+
+	  If you have a controller with this interface, say Y or M here. You
+	  also need to enable an appropriate bus interface.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_IO_ACCESSORS
+	bool
+	depends on MMC_SDHCI
+	help
+	  This is silent Kconfig symbol that is selected by the drivers that
+	  need to overwrite SDHCI IO memory accessors.
+
+config MMC_SDHCI_BIG_ENDIAN_32BIT_BYTE_SWAPPER
+	bool
+	depends on MMC_SDHCI
+	select MMC_SDHCI_IO_ACCESSORS
+	help
+	  This option is selected by drivers running on big endian hosts
+	  and performing I/O to a SDHCI controller through a bus that
+	  implements a hardware byte swapper using a 32-bit datum.
+	  This endian mapping mode is called "data invariance" and
+	  has the effect of scrambling the addresses and formats of data
+	  accessed in sizes other than the datum size.
+
+	  This is the case for the Nintendo Wii SDHCI.
+
+config MMC_SDHCI_PCI
+	tristate "SDHCI support on PCI bus"
+	depends on MMC_SDHCI && PCI
+	select MMC_CQHCI
+	select IOSF_MBI if X86
+	select MMC_SDHCI_IO_ACCESSORS
+	help
+	  This selects the PCI Secure Digital Host Controller Interface.
+	  Most controllers found today are PCI devices.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_RICOH_MMC
+	bool "Ricoh MMC Controller Disabler"
+	depends on MMC_SDHCI_PCI
+	default y
+	help
+	  This adds a pci quirk to disable Ricoh MMC Controller. This
+	  proprietary controller is unnecessary because the SDHCI driver
+	  supports MMC cards on the SD controller, but if it is not
+	  disabled, it will steal the MMC cards away - rendering them
+	  useless. It is safe to select this even if you don't
+	  have a Ricoh based card reader.
+
+	  If unsure, say Y.
+
+config MMC_SDHCI_ACPI
+	tristate "SDHCI support for ACPI enumerated SDHCI controllers"
+	depends on MMC_SDHCI && ACPI && PCI
+	select IOSF_MBI if X86
+	help
+	  This selects support for ACPI enumerated SDHCI controllers,
+	  identified by ACPI Compatibility ID PNP0D40 or specific
+	  ACPI Hardware IDs.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_PLTFM
+	tristate "SDHCI platform and OF driver helper"
+	depends on MMC_SDHCI
+	help
+	  This selects the common helper functions support for Secure Digital
+	  Host Controller Interface based platform and OF drivers.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_OF_ARASAN
+	tristate "SDHCI OF support for the Arasan SDHCI controllers"
+	depends on MMC_SDHCI_PLTFM
+	depends on OF
+	depends on COMMON_CLK
+	select MMC_CQHCI
+	help
+	  This selects the Arasan Secure Digital Host Controller Interface
+	  (SDHCI). This hardware is found e.g. in Xilinx' Zynq SoC.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_OF_ASPEED
+	tristate "SDHCI OF support for the ASPEED SDHCI controller"
+	depends on MMC_SDHCI_PLTFM
+	depends on OF && OF_ADDRESS
+	select MMC_SDHCI_IO_ACCESSORS
+	help
+	  This selects the ASPEED Secure Digital Host Controller Interface.
+
+	  If you have a controller with this interface, say Y or M here. You
+	  also need to enable an appropriate bus interface.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_OF_AT91
+	tristate "SDHCI OF support for the Atmel SDMMC controller"
+	depends on MMC_SDHCI_PLTFM
+	depends on OF && HAVE_CLK
+	help
+	  This selects the Atmel SDMMC driver
+
+config MMC_SDHCI_OF_ESDHC
+	tristate "SDHCI OF support for the Freescale eSDHC controller"
+	depends on MMC_SDHCI_PLTFM
+	depends on PPC || ARCH_MXC || ARCH_LAYERSCAPE || COMPILE_TEST
+	select MMC_SDHCI_IO_ACCESSORS
+	select FSL_GUTS
+	help
+	  This selects the Freescale eSDHC controller support.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_OF_HLWD
+	tristate "SDHCI OF support for the Nintendo Wii SDHCI controllers"
+	depends on MMC_SDHCI_PLTFM
+	depends on PPC
+	select MMC_SDHCI_BIG_ENDIAN_32BIT_BYTE_SWAPPER
+	help
+	  This selects the Secure Digital Host Controller Interface (SDHCI)
+	  found in the "Hollywood" chipset of the Nintendo Wii video game
+	  console.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_OF_DWCMSHC
+	tristate "SDHCI OF support for the Synopsys DWC MSHC"
+	depends on MMC_SDHCI_PLTFM
+	depends on OF
+	depends on COMMON_CLK
+	help
+	  This selects Synopsys DesignWare Cores Mobile Storage Controller
+	  support.
+	  If you have a controller with this interface, say Y or M here.
+	  If unsure, say N.
+
+config MMC_SDHCI_OF_SPARX5
+	tristate "SDHCI OF support for the MCHP Sparx5 SoC"
+	depends on MMC_SDHCI_PLTFM
+	depends on ARCH_SPARX5 || COMPILE_TEST
+	help
+	  This selects the Secure Digital Host Controller Interface (SDHCI)
+	  found in the MCHP Sparx5 SoC.
+
+	  If you have a Sparx5 SoC with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_CADENCE
+	tristate "SDHCI support for the Cadence SD/SDIO/eMMC controller"
+	depends on MMC_SDHCI_PLTFM
+	depends on OF
+	help
+	  This selects the Cadence SD/SDIO/eMMC driver.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_CNS3XXX
+	tristate "SDHCI support on the Cavium Networks CNS3xxx SoC"
+	depends on ARCH_CNS3XXX || COMPILE_TEST
+	depends on MMC_SDHCI_PLTFM
+	help
+	  This selects the SDHCI support for CNS3xxx System-on-Chip devices.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_ESDHC_MCF
+	tristate "SDHCI support for the Freescale eSDHC ColdFire controller"
+	depends on M5441x
+	depends on MMC_SDHCI_PLTFM
+	select MMC_SDHCI_IO_ACCESSORS
+	help
+	  This selects the Freescale eSDHC controller support for
+	  ColdFire mcf5441x devices.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_ESDHC_IMX
+	tristate "SDHCI support for the Freescale eSDHC/uSDHC i.MX controller"
+	depends on ARCH_MXC || COMPILE_TEST
+	depends on MMC_SDHCI_PLTFM
+	select MMC_SDHCI_IO_ACCESSORS
+	select MMC_CQHCI
+	help
+	  This selects the Freescale eSDHC/uSDHC controller support
+	  found on i.MX25, i.MX35 i.MX5x and i.MX6x.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_DOVE
+	tristate "SDHCI support on Marvell's Dove SoC"
+	depends on ARCH_DOVE || MACH_DOVE || COMPILE_TEST
+	depends on MMC_SDHCI_PLTFM
+	select MMC_SDHCI_IO_ACCESSORS
+	help
+	  This selects the Secure Digital Host Controller Interface in
+	  Marvell's Dove SoC.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_TEGRA
+	tristate "SDHCI platform support for the Tegra SD/MMC Controller"
+	depends on ARCH_TEGRA || COMPILE_TEST
+	depends on MMC_SDHCI_PLTFM
+	select MMC_SDHCI_IO_ACCESSORS
+	select MMC_CQHCI
+	help
+	  This selects the Tegra SD/MMC controller. If you have a Tegra
+	  platform with SD or MMC devices, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_S3C
+	tristate "SDHCI support on Samsung S3C SoC"
+	depends on MMC_SDHCI
+	depends on PLAT_SAMSUNG || ARCH_S5PV210 || ARCH_EXYNOS || COMPILE_TEST
+	help
+	  This selects the Secure Digital Host Controller Interface (SDHCI)
+	  often referrered to as the HSMMC block in some of the Samsung S3C
+	  range of SoC.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_SIRF
+	tristate "SDHCI support on CSR SiRFprimaII and SiRFmarco SoCs"
+	depends on ARCH_SIRF || COMPILE_TEST
+	depends on MMC_SDHCI_PLTFM
+	select MMC_SDHCI_IO_ACCESSORS
+	help
+	  This selects the SDHCI support for SiRF System-on-Chip devices.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_PXAV3
+	tristate "Marvell MMP2 SD Host Controller support (PXAV3)"
+	depends on CLKDEV_LOOKUP
+	depends on MMC_SDHCI_PLTFM
+	depends on ARCH_BERLIN || ARCH_MMP || ARCH_MVEBU || COMPILE_TEST
+	default CPU_MMP2
+	help
+	  This selects the Marvell(R) PXAV3 SD Host Controller.
+	  If you have a MMP2 platform with SD Host Controller
+	  and a card slot, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_PXAV2
+	tristate "Marvell PXA9XX SD Host Controller support (PXAV2)"
+	depends on CLKDEV_LOOKUP
+	depends on MMC_SDHCI_PLTFM
+	depends on ARCH_MMP || COMPILE_TEST
+	default CPU_PXA910
+	help
+	  This selects the Marvell(R) PXAV2 SD Host Controller.
+	  If you have a PXA9XX platform with SD Host Controller
+	  and a card slot, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_SPEAR
+	tristate "SDHCI support on ST SPEAr platform"
+	depends on MMC_SDHCI
+	depends on PLAT_SPEAR || COMPILE_TEST
+	depends on OF
+	help
+	  This selects the Secure Digital Host Controller Interface (SDHCI)
+	  often referrered to as the HSMMC block in some of the ST SPEAR range
+	  of SoC
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_S3C_DMA
+	bool "DMA support on S3C SDHCI"
+	depends on MMC_SDHCI_S3C
+	help
+	  Enable DMA support on the Samsung S3C SDHCI glue. The DMA
+	  has proved to be problematic if the controller encounters
+	  certain errors, and thus should be treated with care.
+
+	  YMMV.
+
+config MMC_SDHCI_BCM_KONA
+	tristate "SDHCI support on Broadcom KONA platform"
+	depends on ARCH_BCM_MOBILE || COMPILE_TEST
+	depends on MMC_SDHCI_PLTFM
+	help
+	  This selects the Broadcom Kona Secure Digital Host Controller
+	  Interface(SDHCI) support.
+	  This is used in Broadcom mobile SoCs.
+
+	  If you have a controller with this interface, say Y or M here.
+
+config MMC_SDHCI_F_SDH30
+	tristate "SDHCI support for Fujitsu Semiconductor F_SDH30"
+	depends on MMC_SDHCI_PLTFM
+	depends on OF || ACPI
+	help
+	  This selects the Secure Digital Host Controller Interface (SDHCI)
+	  Needed by some Fujitsu SoC for MMC / SD / SDIO support.
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_MILBEAUT
+	tristate "SDHCI support for Socionext Milbeaut Serieas using F_SDH30"
+	depends on MMC_SDHCI_PLTFM
+	depends on OF
+	help
+	  This selects the Secure Digital Host Controller Interface (SDHCI)
+	  Needed by Milbeaut SoC for MMC / SD / SDIO support.
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_IPROC
+	tristate "SDHCI support for the BCM2835 & iProc SD/MMC Controller"
+	depends on ARCH_BCM2835 || ARCH_BCM_IPROC || COMPILE_TEST
+	depends on MMC_SDHCI_PLTFM
+	depends on OF || ACPI
+	default ARCH_BCM_IPROC
+	select MMC_SDHCI_IO_ACCESSORS
+	help
+	  This selects the iProc SD/MMC controller.
+
+	  If you have a BCM2835 or IPROC platform with SD or MMC devices,
+	  say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_MESON_GX
+	tristate "Amlogic S905/GX*/AXG SD/MMC Host Controller support"
+	depends on ARCH_MESON|| COMPILE_TEST
+	depends on COMMON_CLK
+	help
+	  This selects support for the Amlogic SD/MMC Host Controller
+	  found on the S905/GX*/AXG family of SoCs.  This controller is
+	  MMC 5.1 compliant and supports SD, eMMC and SDIO interfaces.
+
+	  If you have a controller with this interface, say Y here.
+
+config MMC_MESON_MX_SDHC
+	tristate "Amlogic Meson SDHC Host Controller support"
+	depends on (ARM && ARCH_MESON) || COMPILE_TEST
+	depends on COMMON_CLK
+	depends on OF
+	help
+	  This selects support for the SDHC Host Controller on
+	  Amlogic Meson6, Meson8, Meson8b and Meson8m2 SoCs.
+	  The controller supports the SD/SDIO Spec 3.x and eMMC Spec 4.5x
+	  with 1, 4, and 8 bit bus widths.
+
+	  If you have a controller with this interface, say Y or M here.
+	  If unsure, say N.
+
+config MMC_MESON_MX_SDIO
+	tristate "Amlogic Meson6/Meson8/Meson8b SD/MMC Host Controller support"
+	depends on ARCH_MESON || COMPILE_TEST
+	depends on COMMON_CLK
+	depends on OF_ADDRESS
+	help
+	  This selects support for the SD/MMC Host Controller on
+	  Amlogic Meson6, Meson8 and Meson8b SoCs.
+
+	  If you have a controller with this interface, say Y or M here.
+	  If unsure, say N.
+
+config MMC_MOXART
+	tristate "MOXART SD/MMC Host Controller support"
+	depends on ARCH_MOXART || COMPILE_TEST
+	help
+	  This selects support for the MOXART SD/MMC Host Controller.
+	  MOXA provides one multi-functional card reader which can
+	  be found on some embedded hardware such as UC-7112-LX.
+	  If you have a controller with this interface, say Y here.
+
+config MMC_SDHCI_ST
+	tristate "SDHCI support on STMicroelectronics SoC"
+	depends on ARCH_STI || FSP2 || COMPILE_TEST
+	depends on MMC_SDHCI_PLTFM
+	select MMC_SDHCI_IO_ACCESSORS
+	help
+	  This selects the Secure Digital Host Controller Interface in
+	  STMicroelectronics SoCs.
+
+	  If you have a controller with this interface, say Y or M here.
+	  If unsure, say N.
+
+config MMC_OMAP
+	tristate "TI OMAP Multimedia Card Interface support"
+	depends on ARCH_OMAP
+	depends on TPS65010 || !MACH_OMAP_H2
+	help
+	  This selects the TI OMAP Multimedia card Interface.
+	  If you have an OMAP board with a Multimedia Card slot,
+	  say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_OMAP_HS
+	tristate "TI OMAP High Speed Multimedia Card Interface support"
+	depends on ARCH_OMAP2PLUS || ARCH_KEYSTONE || COMPILE_TEST
+	help
+	  This selects the TI OMAP High Speed Multimedia card Interface.
+	  If you have an omap2plus board with a Multimedia Card slot,
+	  say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_WBSD
+	tristate "Winbond W83L51xD SD/MMC Card Interface support"
+	depends on ISA_DMA_API && !M68K
+	help
+	  This selects the Winbond(R) W83L51xD Secure digital and
+	  Multimedia card Interface.
+	  If you have a machine with a integrated W83L518D or W83L519D
+	  SD/MMC card reader, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_ALCOR
+	tristate "Alcor Micro/Alcor Link SD/MMC controller"
+	depends on MISC_ALCOR_PCI
+	help
+	  Say Y here to include driver code to support SD/MMC card interface
+	  of Alcor Micro PCI-E card reader
+
+config MMC_AU1X
+	bool "Alchemy AU1XX0 MMC Card Interface support"
+	depends on MIPS_ALCHEMY
+	depends on MMC=y
+	help
+	  This selects the AMD Alchemy(R) Multimedia card interface.
+	  If you have a Alchemy platform with a MMC slot, say Y here.
+
+	  If unsure, say N.
+
+config MMC_ATMELMCI
+	tristate "Atmel SD/MMC Driver (Multimedia Card Interface)"
+	depends on ARCH_AT91
+	help
+	  This selects the Atmel Multimedia Card Interface driver.
+	  If you have an AT91 platform with a Multimedia Card slot,
+	  say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_MSM
+	tristate "Qualcomm SDHCI Controller Support"
+	depends on ARCH_QCOM || COMPILE_TEST
+	depends on MMC_SDHCI_PLTFM
+	select MMC_SDHCI_IO_ACCESSORS
+	select MMC_CQHCI
+	help
+	  This selects the Secure Digital Host Controller Interface (SDHCI)
+	  support present in Qualcomm SOCs. The controller supports
+	  SD/MMC/SDIO devices.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_MXC
+	tristate "Freescale i.MX21/27/31 or MPC512x Multimedia Card support"
+	depends on ARCH_MXC || PPC_MPC512x
+	help
+	  This selects the Freescale i.MX21, i.MX27, i.MX31 or MPC512x
+	  Multimedia Card Interface. If you have an i.MX or MPC512x platform
+	  with a Multimedia Card slot, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_MXS
+	tristate "Freescale MXS Multimedia Card Interface support"
+	depends on ARCH_MXS && MXS_DMA
+	help
+	  This selects the Freescale SSP MMC controller found on MXS based
+	  platforms like mx23/28.
+
+	  If unsure, say N.
+
+config MMC_TIFM_SD
+	tristate "TI Flash Media MMC/SD Interface support"
+	depends on PCI
+	select TIFM_CORE
+	help
+	  Say Y here if you want to be able to access MMC/SD cards with
+	  the Texas Instruments(R) Flash Media card reader, found in many
+	  laptops.
+	  This option 'selects' (turns on, enables) 'TIFM_CORE', but you
+	  probably also need appropriate card reader host adapter, such as
+	  'Misc devices: TI Flash Media PCI74xx/PCI76xx host adapter support
+	  (TIFM_7XX1)'.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called tifm_sd.
+
+config MMC_MVSDIO
+	tristate "Marvell MMC/SD/SDIO host driver"
+	depends on PLAT_ORION || (COMPILE_TEST && ARM)
+	depends on OF
+	help
+	  This selects the Marvell SDIO host driver.
+	  SDIO may currently be found on the Kirkwood 88F6281 and 88F6192
+	  SoC controllers.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called mvsdio.
+
+config MMC_DAVINCI
+	tristate "TI DAVINCI Multimedia Card Interface support"
+	depends on ARCH_DAVINCI || COMPILE_TEST
+	help
+	  This selects the TI DAVINCI Multimedia card Interface.
+	  If you have an DAVINCI board with a Multimedia Card slot,
+	  say Y or M here.  If unsure, say N.
+
+config MMC_GOLDFISH
+	tristate "goldfish qemu Multimedia Card Interface support"
+	depends on GOLDFISH || COMPILE_TEST
+	help
+	  This selects the Goldfish Multimedia card Interface emulation
+	  found on the Goldfish Android virtual device emulation.
+
+config MMC_SPI
+	tristate "MMC/SD/SDIO over SPI"
+	depends on SPI_MASTER
+	select CRC7
+	select CRC_ITU_T
+	help
+	  Some systems access MMC/SD/SDIO cards using a SPI controller
+	  instead of using a "native" MMC/SD/SDIO controller.  This has a
+	  disadvantage of being relatively high overhead, but a compensating
+	  advantage of working on many systems without dedicated MMC/SD/SDIO
+	  controllers.
+
+	  If unsure, or if your system has no SPI master driver, say N.
+
+config MMC_S3C
+	tristate "Samsung S3C SD/MMC Card Interface support"
+	depends on ARCH_S3C24XX
+	depends on S3C24XX_DMAC
+	help
+	  This selects a driver for the MCI interface found in
+	  Samsung's S3C2410, S3C2412, S3C2440, S3C2442 CPUs.
+	  If you have a board based on one of those and a MMC/SD
+	  slot, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_S3C_HW_SDIO_IRQ
+	bool "Hardware support for SDIO IRQ"
+	depends on MMC_S3C
+	help
+	  Enable the hardware support for SDIO interrupts instead of using
+	  the generic polling code.
+
+choice
+	prompt "Samsung S3C SD/MMC transfer code"
+	depends on MMC_S3C
+
+config MMC_S3C_PIO
+	bool "Use PIO transfers only"
+	help
+	  Use PIO to transfer data between memory and the hardware.
+
+	  PIO is slower than DMA as it requires CPU instructions to
+	  move the data. This has been the traditional default for
+	  the S3C MCI driver.
+
+config MMC_S3C_DMA
+	bool "Use DMA transfers only"
+	help
+	  Use DMA to transfer data between memory and the hardare.
+
+	  Currently, the DMA support in this driver seems to not be
+	  working properly and needs to be debugged before this
+	  option is useful.
+
+endchoice
+
+config MMC_SDRICOH_CS
+	tristate "MMC/SD driver for Ricoh Bay1Controllers"
+	depends on PCI && PCMCIA
+	help
+	  Say Y here if your Notebook reports a Ricoh Bay1Controller PCMCIA
+	  card whenever you insert a MMC or SD card into the card slot.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called sdricoh_cs.
+
+config MMC_SDHCI_SPRD
+	tristate "Spreadtrum SDIO host Controller"
+	depends on ARCH_SPRD || COMPILE_TEST
+	depends on MMC_SDHCI_PLTFM
+	select MMC_SDHCI_IO_ACCESSORS
+	select MMC_HSQ
+	help
+	  This selects the SDIO Host Controller in Spreadtrum
+	  SoCs, this driver supports R11(IP version: R11P0).
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_TMIO_CORE
+	tristate
+
+config MMC_TMIO
+	tristate "Toshiba Mobile IO Controller (TMIO) MMC/SD function support"
+	depends on MFD_TMIO || MFD_ASIC3 || COMPILE_TEST
+	select MMC_TMIO_CORE
+	help
+	  This provides support for the SD/MMC cell found in TC6393XB,
+	  T7L66XB and also HTC ASIC3
+
+config MMC_SDHI
+	tristate "Renesas SDHI SD/SDIO controller support"
+	depends on SUPERH || ARCH_RENESAS || COMPILE_TEST
+	select MMC_TMIO_CORE
+	help
+	  This provides support for the SDHI SD/SDIO controller found in
+	  Renesas SuperH, ARM and ARM64 based SoCs
+
+config MMC_SDHI_SYS_DMAC
+	tristate "DMA for SDHI SD/SDIO controllers using SYS-DMAC"
+	depends on MMC_SDHI
+	default MMC_SDHI if (SUPERH || ARM)
+	help
+	  This provides DMA support for SDHI SD/SDIO controllers
+	  using SYS-DMAC via DMA Engine. This supports the controllers
+	  found in SuperH and Renesas ARM based SoCs.
+
+config MMC_SDHI_INTERNAL_DMAC
+	tristate "DMA for SDHI SD/SDIO controllers using on-chip bus mastering"
+	depends on ARM64 || ARCH_R7S9210 || ARCH_R8A77470 || COMPILE_TEST
+	depends on MMC_SDHI
+	default MMC_SDHI if (ARM64 || ARCH_R7S9210 || ARCH_R8A77470)
+	help
+	  This provides DMA support for SDHI SD/SDIO controllers
+	  using on-chip bus mastering. This supports the controllers
+	  found in arm64 based SoCs. This controller is also found in
+	  some RZ family SoCs.
+
+config MMC_UNIPHIER
+	tristate "UniPhier SD/eMMC Host Controller support"
+	depends on ARCH_UNIPHIER || COMPILE_TEST
+	depends on OF
+	select MMC_TMIO_CORE
+	help
+	  This provides support for the SD/eMMC controller found in
+	  UniPhier SoCs. The eMMC variant of this controller is used
+	  only for 32-bit SoCs.
+
+config MMC_CB710
+	tristate "ENE CB710 MMC/SD Interface support"
+	depends on PCI
+	select CB710_CORE
+	help
+	  This option enables support for MMC/SD part of ENE CB710/720 Flash
+	  memory card reader found in some laptops (ie. some versions of
+	  HP Compaq nx9500).
+
+	  This driver can also be built as a module. If so, the module
+	  will be called cb710-mmc.
+
+config MMC_VIA_SDMMC
+	tristate "VIA SD/MMC Card Reader Driver"
+	depends on PCI
+	help
+	  This selects the VIA SD/MMC Card Reader driver, say Y or M here.
+	  VIA provides one multi-functional card reader which integrated into
+	  some motherboards manufactured by VIA. This card reader supports
+	  SD/MMC/SDHC.
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_CAVIUM_OCTEON
+	tristate "Cavium OCTEON SD/MMC Card Interface support"
+	depends on CAVIUM_OCTEON_SOC
+	help
+	  This selects Cavium OCTEON SD/MMC card Interface.
+	  If you have an OCTEON board with a Multimedia Card slot,
+	  say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_CAVIUM_THUNDERX
+	tristate "Cavium ThunderX SD/MMC Card Interface support"
+	depends on PCI && 64BIT && (ARM64 || COMPILE_TEST)
+	depends on GPIO_THUNDERX
+	depends on OF_ADDRESS
+	help
+	  This selects Cavium ThunderX SD/MMC Card Interface.
+	  If you have an Cavium ARM64 board with a Multimedia Card slot
+	  or builtin eMMC chip say Y or M here. If built as a module
+	  the module will be called thunderx_mmc.ko.
+
+config MMC_DW
+	tristate "Synopsys DesignWare Memory Card Interface"
+	depends on ARC || ARM || ARM64 || MIPS || RISCV || CSKY || COMPILE_TEST
+	help
+	  This selects support for the Synopsys DesignWare Mobile Storage IP
+	  block, this provides host support for SD and MMC interfaces, in both
+	  PIO, internal DMA mode and external DMA mode.
+
+config MMC_DW_PLTFM
+	tristate "Synopsys Designware MCI Support as platform device"
+	depends on MMC_DW
+	default y
+	help
+	  This selects the common helper functions support for Host Controller
+	  Interface based platform driver. Please select this option if the IP
+	  is present as a platform device. This is the common interface for the
+	  Synopsys Designware IP.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say Y.
+
+config MMC_DW_BLUEFIELD
+	tristate "BlueField specific extensions for Synopsys DW Memory Card Interface"
+	depends on MMC_DW
+	select MMC_DW_PLTFM
+	help
+	  This selects support for Mellanox BlueField SoC specific extensions to
+	  the Synopsys DesignWare Memory Card Interface driver. Select this
+	  option for platforms based on Mellanox BlueField SoC's.
+
+config MMC_DW_EXYNOS
+	tristate "Exynos specific extensions for Synopsys DW Memory Card Interface"
+	depends on MMC_DW
+	select MMC_DW_PLTFM
+	help
+	  This selects support for Samsung Exynos SoC specific extensions to the
+	  Synopsys DesignWare Memory Card Interface driver. Select this option
+	  for platforms based on Exynos4 and Exynos5 SoC's.
+
+config MMC_DW_HI3798CV200
+	tristate "Hi3798CV200 specific extensions for Synopsys DW Memory Card Interface"
+	depends on MMC_DW
+	select MMC_DW_PLTFM
+	help
+	  This selects support for HiSilicon Hi3798CV200 SoC specific extensions to the
+	  Synopsys DesignWare Memory Card Interface driver. Select this option
+	  for platforms based on HiSilicon Hi3798CV200 SoC.
+
+config MMC_DW_K3
+	tristate "K3 specific extensions for Synopsys DW Memory Card Interface"
+	depends on MMC_DW
+	select MMC_DW_PLTFM
+	help
+	  This selects support for Hisilicon K3 SoC specific extensions to the
+	  Synopsys DesignWare Memory Card Interface driver. Select this option
+	  for platforms based on Hisilicon K3 SoC's.
+
+config MMC_DW_PCI
+	tristate "Synopsys Designware MCI support on PCI bus"
+	depends on MMC_DW && PCI
+	help
+	  This selects the PCI bus for the Synopsys Designware Mobile Storage IP.
+	  Select this option if the IP is present on PCI platform.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_DW_ROCKCHIP
+	tristate "Rockchip specific extensions for Synopsys DW Memory Card Interface"
+	depends on MMC_DW && ARCH_ROCKCHIP
+	select MMC_DW_PLTFM
+	help
+	  This selects support for Rockchip SoC specific extensions to the
+	  Synopsys DesignWare Memory Card Interface driver. Select this option
+	  for platforms based on RK3066, RK3188 and RK3288 SoC's.
+
+config MMC_DW_ZX
+	tristate "ZTE specific extensions for Synopsys DW Memory Card Interface"
+	depends on MMC_DW && ARCH_ZX
+	select MMC_DW_PLTFM
+	help
+	  This selects support for ZTE SoC specific extensions to the
+	  Synopsys DesignWare Memory Card Interface driver. Select this option
+	  for platforms based on ZX296718 SoC's.
+
+config MMC_SH_MMCIF
+	tristate "SuperH Internal MMCIF support"
+	depends on SUPERH || ARCH_RENESAS || COMPILE_TEST
+	help
+	  This selects the MMC Host Interface controller (MMCIF) found in various
+	  Renesas SoCs for SH and ARM architectures.
+
+
+config MMC_JZ4740
+	tristate "Ingenic JZ47xx SD/Multimedia Card Interface support"
+	depends on MIPS
+	help
+	  This selects support for the SD/MMC controller on Ingenic
+	  JZ4740, JZ4750, JZ4770 and JZ4780 SoCs.
+
+	  If you have a board based on such a SoC and with a SD/MMC slot,
+	  say Y or M here.
+
+config MMC_VUB300
+	tristate "VUB300 USB to SDIO/SD/MMC Host Controller support"
+	depends on USB
+	help
+	  This selects support for Elan Digital Systems' VUB300 chip.
+
+	  The VUB300 is a USB-SDIO Host Controller Interface chip
+	  that enables the host computer to use SDIO/SD/MMC cards
+	  via a USB 2.0 or USB 1.1 host.
+
+	  The VUB300 chip will be found in both physically separate
+	  USB to SDIO/SD/MMC adapters and embedded on some motherboards.
+
+	  The VUB300 chip supports SD and MMC memory cards in addition
+	  to single and multifunction SDIO cards.
+
+	  Some SDIO cards will need a firmware file to be loaded and
+	  sent to VUB300 chip in order to achieve better data throughput.
+	  Download these "Offload Pseudocode" from Elan Digital Systems'
+	  web-site http://www.elandigitalsystems.com/support/downloads.php
+	  and put them in /lib/firmware. Note that without these additional
+	  firmware files the VUB300 chip will still function, but not at
+	  the best obtainable data rate.
+
+	  To compile this mmc host controller driver as a module,
+	  choose M here: the module will be called vub300.
+
+	  If you have a computer with an embedded VUB300 chip
+	  or if you intend connecting a USB adapter based on a
+	  VUB300 chip say Y or M here.
+
+config MMC_USHC
+	tristate "USB SD Host Controller (USHC) support"
+	depends on USB
+	help
+	  This selects support for USB SD Host Controllers based on
+	  the Cypress Astoria chip with firmware compliant with CSR's
+	  USB SD Host Controller specification (CS-118793-SP).
+
+	  CSR boards with this device include: USB<>SDIO (M1985v2),
+	  and Ultrasira.
+
+	  Note: These controllers only support SDIO cards and do not
+	  support MMC or SD memory cards.
+
+config MMC_WMT
+	tristate "Wondermedia SD/MMC Host Controller support"
+	depends on ARCH_VT8500
+	default y
+	help
+	  This selects support for the SD/MMC Host Controller on
+	  Wondermedia WM8505/WM8650 based SoCs.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called wmt-sdmmc.
+
+config MMC_USDHI6ROL0
+	tristate "Renesas USDHI6ROL0 SD/SDIO Host Controller support"
+	depends on HAS_DMA
+	help
+	  This selects support for the Renesas USDHI6ROL0 SD/SDIO
+	  Host Controller
+
+config MMC_REALTEK_PCI
+	tristate "Realtek PCI-E SD/MMC Card Interface Driver"
+	depends on MISC_RTSX_PCI
+	help
+	  Say Y here to include driver code to support SD/MMC card interface
+	  of Realtek PCI-E card reader
+
+config MMC_REALTEK_USB
+	tristate "Realtek USB SD/MMC Card Interface Driver"
+	depends on MISC_RTSX_USB
+	help
+	  Say Y here to include driver code to support SD/MMC card interface
+	  of Realtek RTS5129/39 series card reader
+
+config MMC_SUNXI
+	tristate "Allwinner sunxi SD/MMC Host Controller support"
+	depends on ARCH_SUNXI || COMPILE_TEST
+	help
+	  This selects support for the SD/MMC Host Controller on
+	  Allwinner sunxi SoCs.
+
+config MMC_CQHCI
+	tristate "Command Queue Host Controller Interface support"
+	depends on HAS_DMA
+	help
+	  This selects the Command Queue Host Controller Interface (CQHCI)
+	  support present in host controllers of Qualcomm Technologies, Inc
+	  amongst others.
+	  This controller supports eMMC devices with command queue support.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_HSQ
+	tristate "MMC Host Software Queue support"
+	help
+	  This selects the MMC Host Software Queue support. This may increase
+	  performance, if the host controller and its driver supports it.
+
+	  If you have a controller/driver supporting this interface, say Y or M
+	  here.
+
+	  If unsure, say N.
+
+config MMC_TOSHIBA_PCI
+	tristate "Toshiba Type A SD/MMC Card Interface Driver"
+	depends on PCI
+
+config MMC_BCM2835
+	tristate "Broadcom BCM2835 SDHOST MMC Controller support"
+	depends on ARCH_BCM2835 || COMPILE_TEST
+	help
+	  This selects the BCM2835 SDHOST MMC controller. If you have
+	  a BCM2835 platform with SD or MMC devices, say Y or M here.
+
+	  Note that the BCM2835 has two SD controllers: The Arasan
+	  sdhci controller (supported by MMC_SDHCI_IPROC) and a custom
+	  sdhost controller (supported by this driver).
+
+	  If unsure, say N.
+
+config MMC_MTK
+	tristate "MediaTek SD/MMC Card Interface support"
+	depends on HAS_DMA
+	select REGULATOR
+	select MMC_CQHCI
+	help
+	  This selects the MediaTek(R) Secure digital and Multimedia card Interface.
+	  If you have a machine with a integrated SD/MMC card reader, say Y or M here.
+	  This is needed if support for any SD/SDIO/MMC devices is required.
+	  If unsure, say N.
+
+config MMC_SDHCI_MICROCHIP_PIC32
+	tristate "Microchip PIC32MZDA SDHCI support"
+	depends on MMC_SDHCI && PIC32MZDA && MMC_SDHCI_PLTFM
+	help
+	  This selects the Secure Digital Host Controller Interface (SDHCI)
+	  for PIC32MZDA platform.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_BRCMSTB
+	tristate "Broadcom SDIO/SD/MMC support"
+	depends on ARCH_BRCMSTB || BMIPS_GENERIC
+	depends on MMC_SDHCI_PLTFM
+	select MMC_CQHCI
+	default y
+	help
+	  This selects support for the SDIO/SD/MMC Host Controller on
+	  Broadcom STB SoCs.
+
+	  If unsure, say Y.
+
+config MMC_SDHCI_XENON
+	tristate "Marvell Xenon eMMC/SD/SDIO SDHCI driver"
+	depends on MMC_SDHCI_PLTFM
+	help
+	  This selects Marvell Xenon eMMC/SD/SDIO SDHCI.
+	  If you have a controller with this interface, say Y or M here.
+	  If unsure, say N.
+
+config MMC_SDHCI_OMAP
+	tristate "TI SDHCI Controller Support"
+	depends on ARCH_OMAP2PLUS || ARCH_KEYSTONE || COMPILE_TEST
+	depends on MMC_SDHCI_PLTFM && OF
+	select THERMAL
+	imply TI_SOC_THERMAL
+	select MMC_SDHCI_EXTERNAL_DMA if DMA_ENGINE
+	help
+	  This selects the Secure Digital Host Controller Interface (SDHCI)
+	  support present in TI's Keystone/OMAP2+/DRA7 SOCs. The controller
+	  supports SD/MMC/SDIO devices.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_AM654
+	tristate "Support for the SDHCI Controller in TI's AM654 SOCs"
+	depends on ARCH_K3 || COMPILE_TEST
+	depends on MMC_SDHCI_PLTFM && OF
+	select MMC_SDHCI_IO_ACCESSORS
+	select MMC_CQHCI
+	select REGMAP_MMIO
+	help
+	  This selects the Secure Digital Host Controller Interface (SDHCI)
+	  support present in TI's AM65x/AM64x/AM62x/J721E SOCs. The controller
+	  supports SD/MMC/SDIO devices.
+
+	  If you have a controller with this interface, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_OWL
+	tristate "Actions Semi Owl SD/MMC Host Controller support"
+	depends on HAS_DMA
+	depends on ARCH_ACTIONS || COMPILE_TEST
+	help
+	  This selects support for the SD/MMC Host Controller on
+	  Actions Semi Owl SoCs.
+
+config MMC_SDHCI_EXTERNAL_DMA
+	bool
diff --git a/drivers/mmc/host/Makefile b/drivers/mmc/host/Makefile
new file mode 100644
index 000000000..451c25fc2
--- /dev/null
+++ b/drivers/mmc/host/Makefile
@@ -0,0 +1,114 @@
+# SPDX-License-Identifier: GPL-2.0
+#
+# Makefile for MMC/SD host controller drivers
+#
+
+obj-$(CONFIG_MMC_ARMMMCI) += armmmci.o
+armmmci-y := mmci.o
+armmmci-$(CONFIG_MMC_QCOM_DML) += mmci_qcom_dml.o
+armmmci-$(CONFIG_MMC_STM32_SDMMC) += mmci_stm32_sdmmc.o
+obj-$(CONFIG_MMC_PXA)		+= pxamci.o
+obj-$(CONFIG_MMC_MXC)		+= mxcmmc.o
+obj-$(CONFIG_MMC_MXS)		+= mxs-mmc.o
+obj-$(CONFIG_MMC_SDHCI)		+= sdhci.o
+obj-$(CONFIG_MMC_SDHCI_PCI)	+= sdhci-pci.o
+sdhci-pci-y			+= sdhci-pci-core.o sdhci-pci-o2micro.o sdhci-pci-arasan.o \
+				   sdhci-pci-dwc-mshc.o sdhci-pci-gli.o
+obj-$(subst m,y,$(CONFIG_MMC_SDHCI_PCI))	+= sdhci-pci-data.o
+obj-$(CONFIG_MMC_SDHCI_ACPI)	+= sdhci-acpi.o
+obj-$(CONFIG_MMC_SDHCI_PXAV3)	+= sdhci-pxav3.o
+obj-$(CONFIG_MMC_SDHCI_PXAV2)	+= sdhci-pxav2.o
+obj-$(CONFIG_MMC_SDHCI_S3C)	+= sdhci-s3c.o
+obj-$(CONFIG_MMC_SDHCI_SIRF)   	+= sdhci-sirf.o
+obj-$(CONFIG_MMC_SDHCI_F_SDH30)	+= sdhci_f_sdh30.o
+obj-$(CONFIG_MMC_SDHCI_MILBEAUT)	+= sdhci-milbeaut.o
+obj-$(CONFIG_MMC_SDHCI_SPEAR)	+= sdhci-spear.o
+obj-$(CONFIG_MMC_SDHCI_AM654)	+= sdhci_am654.o
+obj-$(CONFIG_MMC_WBSD)		+= wbsd.o
+obj-$(CONFIG_MMC_AU1X)		+= au1xmmc.o
+obj-$(CONFIG_MMC_ALCOR)	+= alcor.o
+obj-$(CONFIG_MMC_MTK)		+= mtk-sd.o
+obj-$(CONFIG_MMC_OMAP)		+= omap.o
+obj-$(CONFIG_MMC_OMAP_HS)	+= omap_hsmmc.o
+obj-$(CONFIG_MMC_ATMELMCI)	+= atmel-mci.o
+obj-$(CONFIG_MMC_TIFM_SD)	+= tifm_sd.o
+obj-$(CONFIG_MMC_MVSDIO)	+= mvsdio.o
+obj-$(CONFIG_MMC_DAVINCI)       += davinci_mmc.o
+obj-$(CONFIG_MMC_GOLDFISH)	+= android-goldfish.o
+obj-$(CONFIG_MMC_SPI)		+= mmc_spi.o
+ifeq ($(CONFIG_OF),y)
+obj-$(CONFIG_MMC_SPI)		+= of_mmc_spi.o
+endif
+obj-$(CONFIG_MMC_S3C)   	+= s3cmci.o
+obj-$(CONFIG_MMC_SDRICOH_CS)	+= sdricoh_cs.o
+obj-$(CONFIG_MMC_TMIO)		+= tmio_mmc.o
+obj-$(CONFIG_MMC_TMIO_CORE)	+= tmio_mmc_core.o
+obj-$(CONFIG_MMC_SDHI)		+= renesas_sdhi_core.o
+obj-$(CONFIG_MMC_SDHI_SYS_DMAC)		+= renesas_sdhi_sys_dmac.o
+obj-$(CONFIG_MMC_SDHI_INTERNAL_DMAC)	+= renesas_sdhi_internal_dmac.o
+obj-$(CONFIG_MMC_UNIPHIER)	+= uniphier-sd.o
+obj-$(CONFIG_MMC_CB710)		+= cb710-mmc.o
+obj-$(CONFIG_MMC_VIA_SDMMC)	+= via-sdmmc.o
+octeon-mmc-objs := cavium.o cavium-octeon.o
+obj-$(CONFIG_MMC_CAVIUM_OCTEON) += octeon-mmc.o
+thunderx-mmc-objs := cavium.o cavium-thunderx.o
+obj-$(CONFIG_MMC_CAVIUM_THUNDERX) += thunderx-mmc.o
+obj-$(CONFIG_MMC_DW)		+= dw_mmc.o
+obj-$(CONFIG_MMC_DW_PLTFM)	+= dw_mmc-pltfm.o
+obj-$(CONFIG_MMC_DW_BLUEFIELD)	+= dw_mmc-bluefield.o
+obj-$(CONFIG_MMC_DW_EXYNOS)	+= dw_mmc-exynos.o
+obj-$(CONFIG_MMC_DW_HI3798CV200) += dw_mmc-hi3798cv200.o
+obj-$(CONFIG_MMC_DW_K3)		+= dw_mmc-k3.o
+obj-$(CONFIG_MMC_DW_PCI)	+= dw_mmc-pci.o
+obj-$(CONFIG_MMC_DW_ROCKCHIP)	+= dw_mmc-rockchip.o
+obj-$(CONFIG_MMC_DW_ZX)		+= dw_mmc-zx.o
+obj-$(CONFIG_MMC_SH_MMCIF)	+= sh_mmcif.o
+obj-$(CONFIG_MMC_JZ4740)	+= jz4740_mmc.o
+obj-$(CONFIG_MMC_VUB300)	+= vub300.o
+obj-$(CONFIG_MMC_USHC)		+= ushc.o
+obj-$(CONFIG_MMC_WMT)		+= wmt-sdmmc.o
+obj-$(CONFIG_MMC_MESON_GX)	+= meson-gx-mmc.o
+meson-mx-sdhc-objs 		:= meson-mx-sdhc-clkc.o meson-mx-sdhc-mmc.o
+obj-$(CONFIG_MMC_MESON_MX_SDHC)	+= meson-mx-sdhc.o
+obj-$(CONFIG_MMC_MESON_MX_SDIO)	+= meson-mx-sdio.o
+obj-$(CONFIG_MMC_MOXART)	+= moxart-mmc.o
+obj-$(CONFIG_MMC_SUNXI)		+= sunxi-mmc.o
+obj-$(CONFIG_MMC_USDHI6ROL0)	+= usdhi6rol0.o
+obj-$(CONFIG_MMC_TOSHIBA_PCI)	+= toshsd.o
+obj-$(CONFIG_MMC_BCM2835)	+= bcm2835.o
+obj-$(CONFIG_MMC_OWL)		+= owl-mmc.o
+
+obj-$(CONFIG_MMC_REALTEK_PCI)	+= rtsx_pci_sdmmc.o
+obj-$(CONFIG_MMC_REALTEK_USB)	+= rtsx_usb_sdmmc.o
+
+obj-$(CONFIG_MMC_SDHCI_PLTFM)		+= sdhci-pltfm.o
+obj-$(CONFIG_MMC_SDHCI_CADENCE)		+= sdhci-cadence.o
+obj-$(CONFIG_MMC_SDHCI_CNS3XXX)		+= sdhci-cns3xxx.o
+obj-$(CONFIG_MMC_SDHCI_ESDHC_MCF)       += sdhci-esdhc-mcf.o
+obj-$(CONFIG_MMC_SDHCI_ESDHC_IMX)	+= sdhci-esdhc-imx.o
+obj-$(CONFIG_MMC_SDHCI_DOVE)		+= sdhci-dove.o
+obj-$(CONFIG_MMC_SDHCI_TEGRA)		+= sdhci-tegra.o
+obj-$(CONFIG_MMC_SDHCI_OF_ARASAN)	+= sdhci-of-arasan.o
+obj-$(CONFIG_MMC_SDHCI_OF_ASPEED)	+= sdhci-of-aspeed.o
+obj-$(CONFIG_MMC_SDHCI_OF_AT91)		+= sdhci-of-at91.o
+obj-$(CONFIG_MMC_SDHCI_OF_ESDHC)	+= sdhci-of-esdhc.o
+obj-$(CONFIG_MMC_SDHCI_OF_HLWD)		+= sdhci-of-hlwd.o
+obj-$(CONFIG_MMC_SDHCI_OF_DWCMSHC)	+= sdhci-of-dwcmshc.o
+obj-$(CONFIG_MMC_SDHCI_OF_SPARX5)	+= sdhci-of-sparx5.o
+obj-$(CONFIG_MMC_SDHCI_BCM_KONA)	+= sdhci-bcm-kona.o
+obj-$(CONFIG_MMC_SDHCI_IPROC)		+= sdhci-iproc.o
+obj-$(CONFIG_MMC_SDHCI_MSM)		+= sdhci-msm.o
+obj-$(CONFIG_MMC_SDHCI_ST)		+= sdhci-st.o
+obj-$(CONFIG_MMC_SDHCI_MICROCHIP_PIC32)	+= sdhci-pic32.o
+obj-$(CONFIG_MMC_SDHCI_BRCMSTB)		+= sdhci-brcmstb.o
+obj-$(CONFIG_MMC_SDHCI_OMAP)		+= sdhci-omap.o
+obj-$(CONFIG_MMC_SDHCI_SPRD)		+= sdhci-sprd.o
+obj-$(CONFIG_MMC_CQHCI)			+= cqhci.o
+obj-$(CONFIG_MMC_HSQ)			+= mmc_hsq.o
+
+ifeq ($(CONFIG_CB710_DEBUG),y)
+	CFLAGS-cb710-mmc	+= -DDEBUG
+endif
+
+obj-$(CONFIG_MMC_SDHCI_XENON)	+= sdhci-xenon-driver.o
+sdhci-xenon-driver-y		+= sdhci-xenon.o sdhci-xenon-phy.o
diff --git a/drivers/mmc/host/alcor.c b/drivers/mmc/host/alcor.c
new file mode 100644
index 000000000..d01df01d4
--- /dev/null
+++ b/drivers/mmc/host/alcor.c
@@ -0,0 +1,1192 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright (C) 2018 Oleksij Rempel <linux@rempel-privat.de>
+ *
+ * Driver for Alcor Micro AU6601 and AU6621 controllers
+ */
+
+/* Note: this driver was created without any documentation. Based
+ * on sniffing, testing and in some cases mimic of original driver.
+ * As soon as some one with documentation or more experience in SD/MMC, or
+ * reverse engineering then me, please review this driver and question every
+ * thing what I did. 2018 Oleksij Rempel <linux@rempel-privat.de>
+ */
+
+#include <linux/delay.h>
+#include <linux/pci.h>
+#include <linux/module.h>
+#include <linux/io.h>
+#include <linux/pm.h>
+#include <linux/irq.h>
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+
+#include <linux/alcor_pci.h>
+
+enum alcor_cookie {
+	COOKIE_UNMAPPED,
+	COOKIE_PRE_MAPPED,
+	COOKIE_MAPPED,
+};
+
+struct alcor_pll_conf {
+	unsigned int clk_src_freq;
+	unsigned int clk_src_reg;
+	unsigned int min_div;
+	unsigned int max_div;
+};
+
+struct alcor_sdmmc_host {
+	struct  device *dev;
+	struct alcor_pci_priv *alcor_pci;
+
+	struct mmc_request *mrq;
+	struct mmc_command *cmd;
+	struct mmc_data *data;
+	unsigned int dma_on:1;
+
+	struct mutex cmd_mutex;
+
+	struct delayed_work timeout_work;
+
+	struct sg_mapping_iter sg_miter;	/* SG state for PIO */
+	struct scatterlist *sg;
+	unsigned int blocks;		/* remaining PIO blocks */
+	int sg_count;
+
+	u32			irq_status_sd;
+	unsigned char		cur_power_mode;
+};
+
+static const struct alcor_pll_conf alcor_pll_cfg[] = {
+	/* MHZ,		CLK src,		max div, min div */
+	{ 31250000,	AU6601_CLK_31_25_MHZ,	1,	511},
+	{ 48000000,	AU6601_CLK_48_MHZ,	1,	511},
+	{125000000,	AU6601_CLK_125_MHZ,	1,	511},
+	{384000000,	AU6601_CLK_384_MHZ,	1,	511},
+};
+
+static inline void alcor_rmw8(struct alcor_sdmmc_host *host, unsigned int addr,
+			       u8 clear, u8 set)
+{
+	struct alcor_pci_priv *priv = host->alcor_pci;
+	u32 var;
+
+	var = alcor_read8(priv, addr);
+	var &= ~clear;
+	var |= set;
+	alcor_write8(priv, var, addr);
+}
+
+/* As soon as irqs are masked, some status updates may be missed.
+ * Use this with care.
+ */
+static inline void alcor_mask_sd_irqs(struct alcor_sdmmc_host *host)
+{
+	struct alcor_pci_priv *priv = host->alcor_pci;
+
+	alcor_write32(priv, 0, AU6601_REG_INT_ENABLE);
+}
+
+static inline void alcor_unmask_sd_irqs(struct alcor_sdmmc_host *host)
+{
+	struct alcor_pci_priv *priv = host->alcor_pci;
+
+	alcor_write32(priv, AU6601_INT_CMD_MASK | AU6601_INT_DATA_MASK |
+		  AU6601_INT_CARD_INSERT | AU6601_INT_CARD_REMOVE |
+		  AU6601_INT_OVER_CURRENT_ERR,
+		  AU6601_REG_INT_ENABLE);
+}
+
+static void alcor_reset(struct alcor_sdmmc_host *host, u8 val)
+{
+	struct alcor_pci_priv *priv = host->alcor_pci;
+	int i;
+
+	alcor_write8(priv, val | AU6601_BUF_CTRL_RESET,
+		      AU6601_REG_SW_RESET);
+	for (i = 0; i < 100; i++) {
+		if (!(alcor_read8(priv, AU6601_REG_SW_RESET) & val))
+			return;
+		udelay(50);
+	}
+	dev_err(host->dev, "%s: timeout\n", __func__);
+}
+
+/*
+ * Perform DMA I/O of a single page.
+ */
+static void alcor_data_set_dma(struct alcor_sdmmc_host *host)
+{
+	struct alcor_pci_priv *priv = host->alcor_pci;
+	u32 addr;
+
+	if (!host->sg_count)
+		return;
+
+	if (!host->sg) {
+		dev_err(host->dev, "have blocks, but no SG\n");
+		return;
+	}
+
+	if (!sg_dma_len(host->sg)) {
+		dev_err(host->dev, "DMA SG len == 0\n");
+		return;
+	}
+
+
+	addr = (u32)sg_dma_address(host->sg);
+
+	alcor_write32(priv, addr, AU6601_REG_SDMA_ADDR);
+	host->sg = sg_next(host->sg);
+	host->sg_count--;
+}
+
+static void alcor_trigger_data_transfer(struct alcor_sdmmc_host *host)
+{
+	struct alcor_pci_priv *priv = host->alcor_pci;
+	struct mmc_data *data = host->data;
+	u8 ctrl = 0;
+
+	if (data->flags & MMC_DATA_WRITE)
+		ctrl |= AU6601_DATA_WRITE;
+
+	if (data->host_cookie == COOKIE_MAPPED) {
+		/*
+		 * For DMA transfers, this function is called just once,
+		 * at the start of the operation. The hardware can only
+		 * perform DMA I/O on a single page at a time, so here
+		 * we kick off the transfer with the first page, and expect
+		 * subsequent pages to be transferred upon IRQ events
+		 * indicating that the single-page DMA was completed.
+		 */
+		alcor_data_set_dma(host);
+		ctrl |= AU6601_DATA_DMA_MODE;
+		host->dma_on = 1;
+		alcor_write32(priv, data->sg_count * 0x1000,
+			       AU6601_REG_BLOCK_SIZE);
+	} else {
+		/*
+		 * For PIO transfers, we break down each operation
+		 * into several sector-sized transfers. When one sector has
+		 * complete, the IRQ handler will call this function again
+		 * to kick off the transfer of the next sector.
+		 */
+		alcor_write32(priv, data->blksz, AU6601_REG_BLOCK_SIZE);
+	}
+
+	alcor_write8(priv, ctrl | AU6601_DATA_START_XFER,
+		      AU6601_DATA_XFER_CTRL);
+}
+
+static void alcor_trf_block_pio(struct alcor_sdmmc_host *host, bool read)
+{
+	struct alcor_pci_priv *priv = host->alcor_pci;
+	size_t blksize, len;
+	u8 *buf;
+
+	if (!host->blocks)
+		return;
+
+	if (host->dma_on) {
+		dev_err(host->dev, "configured DMA but got PIO request.\n");
+		return;
+	}
+
+	if (!!(host->data->flags & MMC_DATA_READ) != read) {
+		dev_err(host->dev, "got unexpected direction %i != %i\n",
+			!!(host->data->flags & MMC_DATA_READ), read);
+	}
+
+	if (!sg_miter_next(&host->sg_miter))
+		return;
+
+	blksize = host->data->blksz;
+	len = min(host->sg_miter.length, blksize);
+
+	dev_dbg(host->dev, "PIO, %s block size: 0x%zx\n",
+		read ? "read" : "write", blksize);
+
+	host->sg_miter.consumed = len;
+	host->blocks--;
+
+	buf = host->sg_miter.addr;
+
+	if (read)
+		ioread32_rep(priv->iobase + AU6601_REG_BUFFER, buf, len >> 2);
+	else
+		iowrite32_rep(priv->iobase + AU6601_REG_BUFFER, buf, len >> 2);
+
+	sg_miter_stop(&host->sg_miter);
+}
+
+static void alcor_prepare_sg_miter(struct alcor_sdmmc_host *host)
+{
+	unsigned int flags = SG_MITER_ATOMIC;
+	struct mmc_data *data = host->data;
+
+	if (data->flags & MMC_DATA_READ)
+		flags |= SG_MITER_TO_SG;
+	else
+		flags |= SG_MITER_FROM_SG;
+	sg_miter_start(&host->sg_miter, data->sg, data->sg_len, flags);
+}
+
+static void alcor_prepare_data(struct alcor_sdmmc_host *host,
+			       struct mmc_command *cmd)
+{
+	struct alcor_pci_priv *priv = host->alcor_pci;
+	struct mmc_data *data = cmd->data;
+
+	if (!data)
+		return;
+
+
+	host->data = data;
+	host->data->bytes_xfered = 0;
+	host->blocks = data->blocks;
+	host->sg = data->sg;
+	host->sg_count = data->sg_count;
+	dev_dbg(host->dev, "prepare DATA: sg %i, blocks: %i\n",
+			host->sg_count, host->blocks);
+
+	if (data->host_cookie != COOKIE_MAPPED)
+		alcor_prepare_sg_miter(host);
+
+	alcor_write8(priv, 0, AU6601_DATA_XFER_CTRL);
+}
+
+static void alcor_send_cmd(struct alcor_sdmmc_host *host,
+			   struct mmc_command *cmd, bool set_timeout)
+{
+	struct alcor_pci_priv *priv = host->alcor_pci;
+	unsigned long timeout = 0;
+	u8 ctrl = 0;
+
+	host->cmd = cmd;
+	alcor_prepare_data(host, cmd);
+
+	dev_dbg(host->dev, "send CMD. opcode: 0x%02x, arg; 0x%08x\n",
+		cmd->opcode, cmd->arg);
+	alcor_write8(priv, cmd->opcode | 0x40, AU6601_REG_CMD_OPCODE);
+	alcor_write32be(priv, cmd->arg, AU6601_REG_CMD_ARG);
+
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE:
+		ctrl = AU6601_CMD_NO_RESP;
+		break;
+	case MMC_RSP_R1:
+		ctrl = AU6601_CMD_6_BYTE_CRC;
+		break;
+	case MMC_RSP_R1B:
+		ctrl = AU6601_CMD_6_BYTE_CRC | AU6601_CMD_STOP_WAIT_RDY;
+		break;
+	case MMC_RSP_R2:
+		ctrl = AU6601_CMD_17_BYTE_CRC;
+		break;
+	case MMC_RSP_R3:
+		ctrl = AU6601_CMD_6_BYTE_WO_CRC;
+		break;
+	default:
+		dev_err(host->dev, "%s: cmd->flag (0x%02x) is not valid\n",
+			mmc_hostname(mmc_from_priv(host)), mmc_resp_type(cmd));
+		break;
+	}
+
+	if (set_timeout) {
+		if (!cmd->data && cmd->busy_timeout)
+			timeout = cmd->busy_timeout;
+		else
+			timeout = 10000;
+
+		schedule_delayed_work(&host->timeout_work,
+				      msecs_to_jiffies(timeout));
+	}
+
+	dev_dbg(host->dev, "xfer ctrl: 0x%02x; timeout: %lu\n", ctrl, timeout);
+	alcor_write8(priv, ctrl | AU6601_CMD_START_XFER,
+				 AU6601_CMD_XFER_CTRL);
+}
+
+static void alcor_request_complete(struct alcor_sdmmc_host *host,
+				   bool cancel_timeout)
+{
+	struct mmc_request *mrq;
+
+	/*
+	 * If this work gets rescheduled while running, it will
+	 * be run again afterwards but without any active request.
+	 */
+	if (!host->mrq)
+		return;
+
+	if (cancel_timeout)
+		cancel_delayed_work(&host->timeout_work);
+
+	mrq = host->mrq;
+
+	host->mrq = NULL;
+	host->cmd = NULL;
+	host->data = NULL;
+	host->dma_on = 0;
+
+	mmc_request_done(mmc_from_priv(host), mrq);
+}
+
+static void alcor_finish_data(struct alcor_sdmmc_host *host)
+{
+	struct mmc_data *data;
+
+	data = host->data;
+	host->data = NULL;
+	host->dma_on = 0;
+
+	/*
+	 * The specification states that the block count register must
+	 * be updated, but it does not specify at what point in the
+	 * data flow. That makes the register entirely useless to read
+	 * back so we have to assume that nothing made it to the card
+	 * in the event of an error.
+	 */
+	if (data->error)
+		data->bytes_xfered = 0;
+	else
+		data->bytes_xfered = data->blksz * data->blocks;
+
+	/*
+	 * Need to send CMD12 if -
+	 * a) open-ended multiblock transfer (no CMD23)
+	 * b) error in multiblock transfer
+	 */
+	if (data->stop &&
+	    (data->error ||
+	     !host->mrq->sbc)) {
+
+		/*
+		 * The controller needs a reset of internal state machines
+		 * upon error conditions.
+		 */
+		if (data->error)
+			alcor_reset(host, AU6601_RESET_CMD | AU6601_RESET_DATA);
+
+		alcor_unmask_sd_irqs(host);
+		alcor_send_cmd(host, data->stop, false);
+		return;
+	}
+
+	alcor_request_complete(host, 1);
+}
+
+static void alcor_err_irq(struct alcor_sdmmc_host *host, u32 intmask)
+{
+	dev_dbg(host->dev, "ERR IRQ %x\n", intmask);
+
+	if (host->cmd) {
+		if (intmask & AU6601_INT_CMD_TIMEOUT_ERR)
+			host->cmd->error = -ETIMEDOUT;
+		else
+			host->cmd->error = -EILSEQ;
+	}
+
+	if (host->data) {
+		if (intmask & AU6601_INT_DATA_TIMEOUT_ERR)
+			host->data->error = -ETIMEDOUT;
+		else
+			host->data->error = -EILSEQ;
+
+		host->data->bytes_xfered = 0;
+	}
+
+	alcor_reset(host, AU6601_RESET_CMD | AU6601_RESET_DATA);
+	alcor_request_complete(host, 1);
+}
+
+static int alcor_cmd_irq_done(struct alcor_sdmmc_host *host, u32 intmask)
+{
+	struct alcor_pci_priv *priv = host->alcor_pci;
+
+	intmask &= AU6601_INT_CMD_END;
+
+	if (!intmask)
+		return true;
+
+	/* got CMD_END but no CMD is in progress, wake thread an process the
+	 * error
+	 */
+	if (!host->cmd)
+		return false;
+
+	if (host->cmd->flags & MMC_RSP_PRESENT) {
+		struct mmc_command *cmd = host->cmd;
+
+		cmd->resp[0] = alcor_read32be(priv, AU6601_REG_CMD_RSP0);
+		dev_dbg(host->dev, "RSP0: 0x%04x\n", cmd->resp[0]);
+		if (host->cmd->flags & MMC_RSP_136) {
+			cmd->resp[1] =
+				alcor_read32be(priv, AU6601_REG_CMD_RSP1);
+			cmd->resp[2] =
+				alcor_read32be(priv, AU6601_REG_CMD_RSP2);
+			cmd->resp[3] =
+				alcor_read32be(priv, AU6601_REG_CMD_RSP3);
+			dev_dbg(host->dev, "RSP1,2,3: 0x%04x 0x%04x 0x%04x\n",
+				cmd->resp[1], cmd->resp[2], cmd->resp[3]);
+		}
+
+	}
+
+	host->cmd->error = 0;
+
+	/* Processed actual command. */
+	if (!host->data)
+		return false;
+
+	alcor_trigger_data_transfer(host);
+	host->cmd = NULL;
+	return true;
+}
+
+static void alcor_cmd_irq_thread(struct alcor_sdmmc_host *host, u32 intmask)
+{
+	intmask &= AU6601_INT_CMD_END;
+
+	if (!intmask)
+		return;
+
+	if (!host->cmd && intmask & AU6601_INT_CMD_END) {
+		dev_dbg(host->dev, "Got command interrupt 0x%08x even though no command operation was in progress.\n",
+			intmask);
+	}
+
+	/* Processed actual command. */
+	if (!host->data)
+		alcor_request_complete(host, 1);
+	else
+		alcor_trigger_data_transfer(host);
+	host->cmd = NULL;
+}
+
+static int alcor_data_irq_done(struct alcor_sdmmc_host *host, u32 intmask)
+{
+	u32 tmp;
+
+	intmask &= AU6601_INT_DATA_MASK;
+
+	/* nothing here to do */
+	if (!intmask)
+		return 1;
+
+	/* we was too fast and got DATA_END after it was processed?
+	 * lets ignore it for now.
+	 */
+	if (!host->data && intmask == AU6601_INT_DATA_END)
+		return 1;
+
+	/* looks like an error, so lets handle it. */
+	if (!host->data)
+		return 0;
+
+	tmp = intmask & (AU6601_INT_READ_BUF_RDY | AU6601_INT_WRITE_BUF_RDY
+			 | AU6601_INT_DMA_END);
+	switch (tmp) {
+	case 0:
+		break;
+	case AU6601_INT_READ_BUF_RDY:
+		alcor_trf_block_pio(host, true);
+		return 1;
+	case AU6601_INT_WRITE_BUF_RDY:
+		alcor_trf_block_pio(host, false);
+		return 1;
+	case AU6601_INT_DMA_END:
+		if (!host->sg_count)
+			break;
+
+		alcor_data_set_dma(host);
+		break;
+	default:
+		dev_err(host->dev, "Got READ_BUF_RDY and WRITE_BUF_RDY at same time\n");
+		break;
+	}
+
+	if (intmask & AU6601_INT_DATA_END) {
+		if (!host->dma_on && host->blocks) {
+			alcor_trigger_data_transfer(host);
+			return 1;
+		} else {
+			return 0;
+		}
+	}
+
+	return 1;
+}
+
+static void alcor_data_irq_thread(struct alcor_sdmmc_host *host, u32 intmask)
+{
+	intmask &= AU6601_INT_DATA_MASK;
+
+	if (!intmask)
+		return;
+
+	if (!host->data) {
+		dev_dbg(host->dev, "Got data interrupt 0x%08x even though no data operation was in progress.\n",
+			intmask);
+		alcor_reset(host, AU6601_RESET_DATA);
+		return;
+	}
+
+	if (alcor_data_irq_done(host, intmask))
+		return;
+
+	if ((intmask & AU6601_INT_DATA_END) || !host->blocks ||
+	    (host->dma_on && !host->sg_count))
+		alcor_finish_data(host);
+}
+
+static void alcor_cd_irq(struct alcor_sdmmc_host *host, u32 intmask)
+{
+	dev_dbg(host->dev, "card %s\n",
+		intmask & AU6601_INT_CARD_REMOVE ? "removed" : "inserted");
+
+	if (host->mrq) {
+		dev_dbg(host->dev, "cancel all pending tasks.\n");
+
+		if (host->data)
+			host->data->error = -ENOMEDIUM;
+
+		if (host->cmd)
+			host->cmd->error = -ENOMEDIUM;
+		else
+			host->mrq->cmd->error = -ENOMEDIUM;
+
+		alcor_request_complete(host, 1);
+	}
+
+	mmc_detect_change(mmc_from_priv(host), msecs_to_jiffies(1));
+}
+
+static irqreturn_t alcor_irq_thread(int irq, void *d)
+{
+	struct alcor_sdmmc_host *host = d;
+	irqreturn_t ret = IRQ_HANDLED;
+	u32 intmask, tmp;
+
+	mutex_lock(&host->cmd_mutex);
+
+	intmask = host->irq_status_sd;
+
+	/* some thing bad */
+	if (unlikely(!intmask || AU6601_INT_ALL_MASK == intmask)) {
+		dev_dbg(host->dev, "unexpected IRQ: 0x%04x\n", intmask);
+		ret = IRQ_NONE;
+		goto exit;
+	}
+
+	tmp = intmask & (AU6601_INT_CMD_MASK | AU6601_INT_DATA_MASK);
+	if (tmp) {
+		if (tmp & AU6601_INT_ERROR_MASK)
+			alcor_err_irq(host, tmp);
+		else {
+			alcor_cmd_irq_thread(host, tmp);
+			alcor_data_irq_thread(host, tmp);
+		}
+		intmask &= ~(AU6601_INT_CMD_MASK | AU6601_INT_DATA_MASK);
+	}
+
+	if (intmask & (AU6601_INT_CARD_INSERT | AU6601_INT_CARD_REMOVE)) {
+		alcor_cd_irq(host, intmask);
+		intmask &= ~(AU6601_INT_CARD_INSERT | AU6601_INT_CARD_REMOVE);
+	}
+
+	if (intmask & AU6601_INT_OVER_CURRENT_ERR) {
+		dev_warn(host->dev,
+			 "warning: over current detected!\n");
+		intmask &= ~AU6601_INT_OVER_CURRENT_ERR;
+	}
+
+	if (intmask)
+		dev_dbg(host->dev, "got not handled IRQ: 0x%04x\n", intmask);
+
+exit:
+	mutex_unlock(&host->cmd_mutex);
+	alcor_unmask_sd_irqs(host);
+	return ret;
+}
+
+
+static irqreturn_t alcor_irq(int irq, void *d)
+{
+	struct alcor_sdmmc_host *host = d;
+	struct alcor_pci_priv *priv = host->alcor_pci;
+	u32 status, tmp;
+	irqreturn_t ret;
+	int cmd_done, data_done;
+
+	status = alcor_read32(priv, AU6601_REG_INT_STATUS);
+	if (!status)
+		return IRQ_NONE;
+
+	alcor_write32(priv, status, AU6601_REG_INT_STATUS);
+
+	tmp = status & (AU6601_INT_READ_BUF_RDY | AU6601_INT_WRITE_BUF_RDY
+			| AU6601_INT_DATA_END | AU6601_INT_DMA_END
+			| AU6601_INT_CMD_END);
+	if (tmp == status) {
+		cmd_done = alcor_cmd_irq_done(host, tmp);
+		data_done = alcor_data_irq_done(host, tmp);
+		/* use fast path for simple tasks */
+		if (cmd_done && data_done) {
+			ret = IRQ_HANDLED;
+			goto alcor_irq_done;
+		}
+	}
+
+	host->irq_status_sd = status;
+	ret = IRQ_WAKE_THREAD;
+	alcor_mask_sd_irqs(host);
+alcor_irq_done:
+	return ret;
+}
+
+static void alcor_set_clock(struct alcor_sdmmc_host *host, unsigned int clock)
+{
+	struct alcor_pci_priv *priv = host->alcor_pci;
+	int i, diff = 0x7fffffff, tmp_clock = 0;
+	u16 clk_src = 0;
+	u8 clk_div = 0;
+
+	if (clock == 0) {
+		alcor_write16(priv, 0, AU6601_CLK_SELECT);
+		return;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(alcor_pll_cfg); i++) {
+		unsigned int tmp_div, tmp_diff;
+		const struct alcor_pll_conf *cfg = &alcor_pll_cfg[i];
+
+		tmp_div = DIV_ROUND_UP(cfg->clk_src_freq, clock);
+		if (cfg->min_div > tmp_div || tmp_div > cfg->max_div)
+			continue;
+
+		tmp_clock = DIV_ROUND_UP(cfg->clk_src_freq, tmp_div);
+		tmp_diff = abs(clock - tmp_clock);
+
+		if (tmp_diff < diff) {
+			diff = tmp_diff;
+			clk_src = cfg->clk_src_reg;
+			clk_div = tmp_div;
+		}
+	}
+
+	clk_src |= ((clk_div - 1) << 8);
+	clk_src |= AU6601_CLK_ENABLE;
+
+	dev_dbg(host->dev, "set freq %d cal freq %d, use div %d, mod %x\n",
+			clock, tmp_clock, clk_div, clk_src);
+
+	alcor_write16(priv, clk_src, AU6601_CLK_SELECT);
+
+}
+
+static void alcor_set_timing(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct alcor_sdmmc_host *host = mmc_priv(mmc);
+
+	if (ios->timing == MMC_TIMING_LEGACY) {
+		alcor_rmw8(host, AU6601_CLK_DELAY,
+			    AU6601_CLK_POSITIVE_EDGE_ALL, 0);
+	} else {
+		alcor_rmw8(host, AU6601_CLK_DELAY,
+			    0, AU6601_CLK_POSITIVE_EDGE_ALL);
+	}
+}
+
+static void alcor_set_bus_width(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct alcor_sdmmc_host *host = mmc_priv(mmc);
+	struct alcor_pci_priv *priv = host->alcor_pci;
+
+	if (ios->bus_width == MMC_BUS_WIDTH_1) {
+		alcor_write8(priv, 0, AU6601_REG_BUS_CTRL);
+	} else if (ios->bus_width == MMC_BUS_WIDTH_4) {
+		alcor_write8(priv, AU6601_BUS_WIDTH_4BIT,
+			      AU6601_REG_BUS_CTRL);
+	} else
+		dev_err(host->dev, "Unknown BUS mode\n");
+
+}
+
+static int alcor_card_busy(struct mmc_host *mmc)
+{
+	struct alcor_sdmmc_host *host = mmc_priv(mmc);
+	struct alcor_pci_priv *priv = host->alcor_pci;
+	u8 status;
+
+	/* Check whether dat[0:3] low */
+	status = alcor_read8(priv, AU6601_DATA_PIN_STATE);
+
+	return !(status & AU6601_BUS_STAT_DAT_MASK);
+}
+
+static int alcor_get_cd(struct mmc_host *mmc)
+{
+	struct alcor_sdmmc_host *host = mmc_priv(mmc);
+	struct alcor_pci_priv *priv = host->alcor_pci;
+	u8 detect;
+
+	detect = alcor_read8(priv, AU6601_DETECT_STATUS)
+		& AU6601_DETECT_STATUS_M;
+	/* check if card is present then send command and data */
+	return (detect == AU6601_SD_DETECTED);
+}
+
+static int alcor_get_ro(struct mmc_host *mmc)
+{
+	struct alcor_sdmmc_host *host = mmc_priv(mmc);
+	struct alcor_pci_priv *priv = host->alcor_pci;
+	u8 status;
+
+	/* get write protect pin status */
+	status = alcor_read8(priv, AU6601_INTERFACE_MODE_CTRL);
+
+	return !!(status & AU6601_SD_CARD_WP);
+}
+
+static void alcor_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct alcor_sdmmc_host *host = mmc_priv(mmc);
+
+	mutex_lock(&host->cmd_mutex);
+
+	host->mrq = mrq;
+
+	/* check if card is present then send command and data */
+	if (alcor_get_cd(mmc))
+		alcor_send_cmd(host, mrq->cmd, true);
+	else {
+		mrq->cmd->error = -ENOMEDIUM;
+		alcor_request_complete(host, 1);
+	}
+
+	mutex_unlock(&host->cmd_mutex);
+}
+
+static void alcor_pre_req(struct mmc_host *mmc,
+			   struct mmc_request *mrq)
+{
+	struct alcor_sdmmc_host *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+	struct mmc_command *cmd = mrq->cmd;
+	struct scatterlist *sg;
+	unsigned int i, sg_len;
+
+	if (!data || !cmd)
+		return;
+
+	data->host_cookie = COOKIE_UNMAPPED;
+
+	/* FIXME: looks like the DMA engine works only with CMD18 */
+	if (cmd->opcode != MMC_READ_MULTIPLE_BLOCK
+			&& cmd->opcode != MMC_WRITE_MULTIPLE_BLOCK)
+		return;
+	/*
+	 * We don't do DMA on "complex" transfers, i.e. with
+	 * non-word-aligned buffers or lengths. A future improvement
+	 * could be made to use temporary DMA bounce-buffers when these
+	 * requirements are not met.
+	 *
+	 * Also, we don't bother with all the DMA setup overhead for
+	 * short transfers.
+	 */
+	if (data->blocks * data->blksz < AU6601_MAX_DMA_BLOCK_SIZE)
+		return;
+
+	if (data->blksz & 3)
+		return;
+
+	for_each_sg(data->sg, sg, data->sg_len, i) {
+		if (sg->length != AU6601_MAX_DMA_BLOCK_SIZE)
+			return;
+		if (sg->offset != 0)
+			return;
+	}
+
+	/* This data might be unmapped at this time */
+
+	sg_len = dma_map_sg(host->dev, data->sg, data->sg_len,
+			    mmc_get_dma_dir(data));
+	if (sg_len)
+		data->host_cookie = COOKIE_MAPPED;
+
+	data->sg_count = sg_len;
+}
+
+static void alcor_post_req(struct mmc_host *mmc,
+			    struct mmc_request *mrq,
+			    int err)
+{
+	struct alcor_sdmmc_host *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+
+	if (!data)
+		return;
+
+	if (data->host_cookie == COOKIE_MAPPED) {
+		dma_unmap_sg(host->dev,
+			     data->sg,
+			     data->sg_len,
+			     mmc_get_dma_dir(data));
+	}
+
+	data->host_cookie = COOKIE_UNMAPPED;
+}
+
+static void alcor_set_power_mode(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct alcor_sdmmc_host *host = mmc_priv(mmc);
+	struct alcor_pci_priv *priv = host->alcor_pci;
+
+	switch (ios->power_mode) {
+	case MMC_POWER_OFF:
+		alcor_set_clock(host, ios->clock);
+		/* set all pins to input */
+		alcor_write8(priv, 0, AU6601_OUTPUT_ENABLE);
+		/* turn of VDD */
+		alcor_write8(priv, 0, AU6601_POWER_CONTROL);
+		break;
+	case MMC_POWER_UP:
+		break;
+	case MMC_POWER_ON:
+		/* This is most trickiest part. The order and timings of
+		 * instructions seems to play important role. Any changes may
+		 * confuse internal state engine if this HW.
+		 * FIXME: If we will ever get access to documentation, then this
+		 * part should be reviewed again.
+		 */
+
+		/* enable SD card mode */
+		alcor_write8(priv, AU6601_SD_CARD,
+			      AU6601_ACTIVE_CTRL);
+		/* set signal voltage to 3.3V */
+		alcor_write8(priv, 0, AU6601_OPT);
+		/* no documentation about clk delay, for now just try to mimic
+		 * original driver.
+		 */
+		alcor_write8(priv, 0x20, AU6601_CLK_DELAY);
+		/* set BUS width to 1 bit */
+		alcor_write8(priv, 0, AU6601_REG_BUS_CTRL);
+		/* set CLK first time */
+		alcor_set_clock(host, ios->clock);
+		/* power on VDD */
+		alcor_write8(priv, AU6601_SD_CARD,
+			      AU6601_POWER_CONTROL);
+		/* wait until the CLK will get stable */
+		mdelay(20);
+		/* set CLK again, mimic original driver. */
+		alcor_set_clock(host, ios->clock);
+
+		/* enable output */
+		alcor_write8(priv, AU6601_SD_CARD,
+			      AU6601_OUTPUT_ENABLE);
+		/* The clk will not work on au6621. We need to trigger data
+		 * transfer.
+		 */
+		alcor_write8(priv, AU6601_DATA_WRITE,
+			      AU6601_DATA_XFER_CTRL);
+		/* configure timeout. Not clear what exactly it means. */
+		alcor_write8(priv, 0x7d, AU6601_TIME_OUT_CTRL);
+		mdelay(100);
+		break;
+	default:
+		dev_err(host->dev, "Unknown power parameter\n");
+	}
+}
+
+static void alcor_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct alcor_sdmmc_host *host = mmc_priv(mmc);
+
+	mutex_lock(&host->cmd_mutex);
+
+	dev_dbg(host->dev, "set ios. bus width: %x, power mode: %x\n",
+		ios->bus_width, ios->power_mode);
+
+	if (ios->power_mode != host->cur_power_mode) {
+		alcor_set_power_mode(mmc, ios);
+		host->cur_power_mode = ios->power_mode;
+	} else {
+		alcor_set_timing(mmc, ios);
+		alcor_set_bus_width(mmc, ios);
+		alcor_set_clock(host, ios->clock);
+	}
+
+	mutex_unlock(&host->cmd_mutex);
+}
+
+static int alcor_signal_voltage_switch(struct mmc_host *mmc,
+				       struct mmc_ios *ios)
+{
+	struct alcor_sdmmc_host *host = mmc_priv(mmc);
+
+	mutex_lock(&host->cmd_mutex);
+
+	switch (ios->signal_voltage) {
+	case MMC_SIGNAL_VOLTAGE_330:
+		alcor_rmw8(host, AU6601_OPT, AU6601_OPT_SD_18V, 0);
+		break;
+	case MMC_SIGNAL_VOLTAGE_180:
+		alcor_rmw8(host, AU6601_OPT, 0, AU6601_OPT_SD_18V);
+		break;
+	default:
+		/* No signal voltage switch required */
+		break;
+	}
+
+	mutex_unlock(&host->cmd_mutex);
+	return 0;
+}
+
+static const struct mmc_host_ops alcor_sdc_ops = {
+	.card_busy	= alcor_card_busy,
+	.get_cd		= alcor_get_cd,
+	.get_ro		= alcor_get_ro,
+	.post_req	= alcor_post_req,
+	.pre_req	= alcor_pre_req,
+	.request	= alcor_request,
+	.set_ios	= alcor_set_ios,
+	.start_signal_voltage_switch = alcor_signal_voltage_switch,
+};
+
+static void alcor_timeout_timer(struct work_struct *work)
+{
+	struct delayed_work *d = to_delayed_work(work);
+	struct alcor_sdmmc_host *host = container_of(d, struct alcor_sdmmc_host,
+						timeout_work);
+	mutex_lock(&host->cmd_mutex);
+
+	dev_dbg(host->dev, "triggered timeout\n");
+	if (host->mrq) {
+		dev_err(host->dev, "Timeout waiting for hardware interrupt.\n");
+
+		if (host->data) {
+			host->data->error = -ETIMEDOUT;
+		} else {
+			if (host->cmd)
+				host->cmd->error = -ETIMEDOUT;
+			else
+				host->mrq->cmd->error = -ETIMEDOUT;
+		}
+
+		alcor_reset(host, AU6601_RESET_CMD | AU6601_RESET_DATA);
+		alcor_request_complete(host, 0);
+	}
+
+	mutex_unlock(&host->cmd_mutex);
+}
+
+static void alcor_hw_init(struct alcor_sdmmc_host *host)
+{
+	struct alcor_pci_priv *priv = host->alcor_pci;
+	struct alcor_dev_cfg *cfg = priv->cfg;
+
+	/* FIXME: This part is a mimics HW init of original driver.
+	 * If we will ever get access to documentation, then this part
+	 * should be reviewed again.
+	 */
+
+	/* reset command state engine */
+	alcor_reset(host, AU6601_RESET_CMD);
+
+	alcor_write8(priv, 0, AU6601_DMA_BOUNDARY);
+	/* enable sd card mode */
+	alcor_write8(priv, AU6601_SD_CARD, AU6601_ACTIVE_CTRL);
+
+	/* set BUS width to 1 bit */
+	alcor_write8(priv, 0, AU6601_REG_BUS_CTRL);
+
+	/* reset data state engine */
+	alcor_reset(host, AU6601_RESET_DATA);
+	/* Not sure if a voodoo with AU6601_DMA_BOUNDARY is really needed */
+	alcor_write8(priv, 0, AU6601_DMA_BOUNDARY);
+
+	alcor_write8(priv, 0, AU6601_INTERFACE_MODE_CTRL);
+	/* not clear what we are doing here. */
+	alcor_write8(priv, 0x44, AU6601_PAD_DRIVE0);
+	alcor_write8(priv, 0x44, AU6601_PAD_DRIVE1);
+	alcor_write8(priv, 0x00, AU6601_PAD_DRIVE2);
+
+	/* for 6601 - dma_boundary; for 6621 - dma_page_cnt
+	 * exact meaning of this register is not clear.
+	 */
+	alcor_write8(priv, cfg->dma, AU6601_DMA_BOUNDARY);
+
+	/* make sure all pins are set to input and VDD is off */
+	alcor_write8(priv, 0, AU6601_OUTPUT_ENABLE);
+	alcor_write8(priv, 0, AU6601_POWER_CONTROL);
+
+	alcor_write8(priv, AU6601_DETECT_EN, AU6601_DETECT_STATUS);
+	/* now we should be safe to enable IRQs */
+	alcor_unmask_sd_irqs(host);
+}
+
+static void alcor_hw_uninit(struct alcor_sdmmc_host *host)
+{
+	struct alcor_pci_priv *priv = host->alcor_pci;
+
+	alcor_mask_sd_irqs(host);
+	alcor_reset(host, AU6601_RESET_CMD | AU6601_RESET_DATA);
+
+	alcor_write8(priv, 0, AU6601_DETECT_STATUS);
+
+	alcor_write8(priv, 0, AU6601_OUTPUT_ENABLE);
+	alcor_write8(priv, 0, AU6601_POWER_CONTROL);
+
+	alcor_write8(priv, 0, AU6601_OPT);
+}
+
+static void alcor_init_mmc(struct alcor_sdmmc_host *host)
+{
+	struct mmc_host *mmc = mmc_from_priv(host);
+
+	mmc->f_min = AU6601_MIN_CLOCK;
+	mmc->f_max = AU6601_MAX_CLOCK;
+	mmc->ocr_avail = MMC_VDD_33_34;
+	mmc->caps = MMC_CAP_4_BIT_DATA | MMC_CAP_SD_HIGHSPEED
+		| MMC_CAP_UHS_SDR12 | MMC_CAP_UHS_SDR25 | MMC_CAP_UHS_SDR50
+		| MMC_CAP_UHS_SDR104 | MMC_CAP_UHS_DDR50;
+	mmc->caps2 = MMC_CAP2_NO_SDIO;
+	mmc->ops = &alcor_sdc_ops;
+
+	/* The hardware does DMA data transfer of 4096 bytes to/from a single
+	 * buffer address. Scatterlists are not supported at the hardware
+	 * level, however we can work with them at the driver level,
+	 * provided that each segment is exactly 4096 bytes in size.
+	 * Upon DMA completion of a single segment (signalled via IRQ), we
+	 * immediately proceed to transfer the next segment from the
+	 * scatterlist.
+	 *
+	 * The overall request is limited to 240 sectors, matching the
+	 * original vendor driver.
+	 */
+	mmc->max_segs = AU6601_MAX_DMA_SEGMENTS;
+	mmc->max_seg_size = AU6601_MAX_DMA_BLOCK_SIZE;
+	mmc->max_blk_count = 240;
+	mmc->max_req_size = mmc->max_blk_count * mmc->max_blk_size;
+	dma_set_max_seg_size(host->dev, mmc->max_seg_size);
+}
+
+static int alcor_pci_sdmmc_drv_probe(struct platform_device *pdev)
+{
+	struct alcor_pci_priv *priv = pdev->dev.platform_data;
+	struct mmc_host *mmc;
+	struct alcor_sdmmc_host *host;
+	int ret;
+
+	mmc = mmc_alloc_host(sizeof(*host), &pdev->dev);
+	if (!mmc) {
+		dev_err(&pdev->dev, "Can't allocate MMC\n");
+		return -ENOMEM;
+	}
+
+	host = mmc_priv(mmc);
+	host->dev = &pdev->dev;
+	host->cur_power_mode = MMC_POWER_UNDEFINED;
+	host->alcor_pci = priv;
+
+	/* make sure irqs are disabled */
+	alcor_write32(priv, 0, AU6601_REG_INT_ENABLE);
+	alcor_write32(priv, 0, AU6601_MS_INT_ENABLE);
+
+	ret = devm_request_threaded_irq(&pdev->dev, priv->irq,
+			alcor_irq, alcor_irq_thread, IRQF_SHARED,
+			DRV_NAME_ALCOR_PCI_SDMMC, host);
+
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to get irq for data line\n");
+		goto free_host;
+	}
+
+	mutex_init(&host->cmd_mutex);
+	INIT_DELAYED_WORK(&host->timeout_work, alcor_timeout_timer);
+
+	alcor_init_mmc(host);
+	alcor_hw_init(host);
+
+	dev_set_drvdata(&pdev->dev, host);
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto free_host;
+
+	return 0;
+
+free_host:
+	mmc_free_host(mmc);
+	return ret;
+}
+
+static int alcor_pci_sdmmc_drv_remove(struct platform_device *pdev)
+{
+	struct alcor_sdmmc_host *host = dev_get_drvdata(&pdev->dev);
+	struct mmc_host *mmc = mmc_from_priv(host);
+
+	if (cancel_delayed_work_sync(&host->timeout_work))
+		alcor_request_complete(host, 0);
+
+	alcor_hw_uninit(host);
+	mmc_remove_host(mmc);
+	mmc_free_host(mmc);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int alcor_pci_sdmmc_suspend(struct device *dev)
+{
+	struct alcor_sdmmc_host *host = dev_get_drvdata(dev);
+
+	if (cancel_delayed_work_sync(&host->timeout_work))
+		alcor_request_complete(host, 0);
+
+	alcor_hw_uninit(host);
+
+	return 0;
+}
+
+static int alcor_pci_sdmmc_resume(struct device *dev)
+{
+	struct alcor_sdmmc_host *host = dev_get_drvdata(dev);
+
+	alcor_hw_init(host);
+
+	return 0;
+}
+#endif /* CONFIG_PM_SLEEP */
+
+static SIMPLE_DEV_PM_OPS(alcor_mmc_pm_ops, alcor_pci_sdmmc_suspend,
+			 alcor_pci_sdmmc_resume);
+
+static const struct platform_device_id alcor_pci_sdmmc_ids[] = {
+	{
+		.name = DRV_NAME_ALCOR_PCI_SDMMC,
+	}, {
+		/* sentinel */
+	}
+};
+MODULE_DEVICE_TABLE(platform, alcor_pci_sdmmc_ids);
+
+static struct platform_driver alcor_pci_sdmmc_driver = {
+	.probe		= alcor_pci_sdmmc_drv_probe,
+	.remove		= alcor_pci_sdmmc_drv_remove,
+	.id_table	= alcor_pci_sdmmc_ids,
+	.driver		= {
+		.name	= DRV_NAME_ALCOR_PCI_SDMMC,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= &alcor_mmc_pm_ops
+	},
+};
+module_platform_driver(alcor_pci_sdmmc_driver);
+
+MODULE_AUTHOR("Oleksij Rempel <linux@rempel-privat.de>");
+MODULE_DESCRIPTION("PCI driver for Alcor Micro AU6601 Secure Digital Host Controller Interface");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mmc/host/android-goldfish.c b/drivers/mmc/host/android-goldfish.c
new file mode 100644
index 000000000..e878fdf8f
--- /dev/null
+++ b/drivers/mmc/host/android-goldfish.c
@@ -0,0 +1,545 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ *  Copyright 2007, Google Inc.
+ *  Copyright 2012, Intel Inc.
+ *
+ *  based on omap.c driver, which was
+ *  Copyright (C) 2004 Nokia Corporation
+ *  Written by Tuukka Tikkanen and Juha Yrjölä <juha.yrjola@nokia.com>
+ *  Misc hacks here and there by Tony Lindgren <tony@atomide.com>
+ *  Other hacks (DMA, SD, etc) by David Brownell
+ */
+
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/major.h>
+
+#include <linux/types.h>
+#include <linux/pci.h>
+#include <linux/interrupt.h>
+
+#include <linux/kernel.h>
+#include <linux/fs.h>
+#include <linux/errno.h>
+#include <linux/hdreg.h>
+#include <linux/kdev_t.h>
+#include <linux/blkdev.h>
+#include <linux/mutex.h>
+#include <linux/scatterlist.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/card.h>
+
+#include <linux/moduleparam.h>
+#include <linux/init.h>
+#include <linux/ioport.h>
+#include <linux/dma-mapping.h>
+#include <linux/delay.h>
+#include <linux/spinlock.h>
+#include <linux/timer.h>
+#include <linux/clk.h>
+
+#include <asm/io.h>
+#include <asm/irq.h>
+
+#include <asm/types.h>
+#include <linux/uaccess.h>
+
+#define DRIVER_NAME "goldfish_mmc"
+
+#define BUFFER_SIZE   16384
+
+#define GOLDFISH_MMC_READ(host, addr)   (readl(host->reg_base + addr))
+#define GOLDFISH_MMC_WRITE(host, addr, x)   (writel(x, host->reg_base + addr))
+
+enum {
+	/* status register */
+	MMC_INT_STATUS	        = 0x00,
+	/* set this to enable IRQ */
+	MMC_INT_ENABLE	        = 0x04,
+	/* set this to specify buffer address */
+	MMC_SET_BUFFER          = 0x08,
+
+	/* MMC command number */
+	MMC_CMD	                = 0x0C,
+
+	/* MMC argument */
+	MMC_ARG	                = 0x10,
+
+	/* MMC response (or R2 bits 0 - 31) */
+	MMC_RESP_0		        = 0x14,
+
+	/* MMC R2 response bits 32 - 63 */
+	MMC_RESP_1		        = 0x18,
+
+	/* MMC R2 response bits 64 - 95 */
+	MMC_RESP_2		        = 0x1C,
+
+	/* MMC R2 response bits 96 - 127 */
+	MMC_RESP_3		        = 0x20,
+
+	MMC_BLOCK_LENGTH        = 0x24,
+	MMC_BLOCK_COUNT         = 0x28,
+
+	/* MMC state flags */
+	MMC_STATE               = 0x2C,
+
+	/* MMC_INT_STATUS bits */
+
+	MMC_STAT_END_OF_CMD     = 1U << 0,
+	MMC_STAT_END_OF_DATA    = 1U << 1,
+	MMC_STAT_STATE_CHANGE   = 1U << 2,
+	MMC_STAT_CMD_TIMEOUT    = 1U << 3,
+
+	/* MMC_STATE bits */
+	MMC_STATE_INSERTED     = 1U << 0,
+	MMC_STATE_READ_ONLY    = 1U << 1,
+};
+
+/*
+ * Command types
+ */
+#define OMAP_MMC_CMDTYPE_BC	0
+#define OMAP_MMC_CMDTYPE_BCR	1
+#define OMAP_MMC_CMDTYPE_AC	2
+#define OMAP_MMC_CMDTYPE_ADTC	3
+
+
+struct goldfish_mmc_host {
+	struct mmc_request	*mrq;
+	struct mmc_command	*cmd;
+	struct mmc_data		*data;
+	struct device		*dev;
+	unsigned char		id; /* 16xx chips have 2 MMC blocks */
+	void			*virt_base;
+	unsigned int		phys_base;
+	int			irq;
+	unsigned char		bus_mode;
+	unsigned char		hw_bus_mode;
+
+	unsigned int		sg_len;
+	unsigned		dma_done:1;
+	unsigned		dma_in_use:1;
+
+	void __iomem		*reg_base;
+};
+
+static inline int
+goldfish_mmc_cover_is_open(struct goldfish_mmc_host *host)
+{
+	return 0;
+}
+
+static ssize_t
+goldfish_mmc_show_cover_switch(struct device *dev,
+			       struct device_attribute *attr, char *buf)
+{
+	struct goldfish_mmc_host *host = dev_get_drvdata(dev);
+
+	return sprintf(buf, "%s\n", goldfish_mmc_cover_is_open(host) ? "open" :
+		       "closed");
+}
+
+static DEVICE_ATTR(cover_switch, S_IRUGO, goldfish_mmc_show_cover_switch, NULL);
+
+static void
+goldfish_mmc_start_command(struct goldfish_mmc_host *host, struct mmc_command *cmd)
+{
+	u32 cmdreg;
+	u32 resptype;
+	u32 cmdtype;
+
+	host->cmd = cmd;
+
+	resptype = 0;
+	cmdtype = 0;
+
+	/* Our hardware needs to know exact type */
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE:
+		break;
+	case MMC_RSP_R1:
+	case MMC_RSP_R1B:
+		/* resp 1, 1b, 6, 7 */
+		resptype = 1;
+		break;
+	case MMC_RSP_R2:
+		resptype = 2;
+		break;
+	case MMC_RSP_R3:
+		resptype = 3;
+		break;
+	default:
+		dev_err(mmc_dev(mmc_from_priv(host)),
+			"Invalid response type: %04x\n", mmc_resp_type(cmd));
+		break;
+	}
+
+	if (mmc_cmd_type(cmd) == MMC_CMD_ADTC)
+		cmdtype = OMAP_MMC_CMDTYPE_ADTC;
+	else if (mmc_cmd_type(cmd) == MMC_CMD_BC)
+		cmdtype = OMAP_MMC_CMDTYPE_BC;
+	else if (mmc_cmd_type(cmd) == MMC_CMD_BCR)
+		cmdtype = OMAP_MMC_CMDTYPE_BCR;
+	else
+		cmdtype = OMAP_MMC_CMDTYPE_AC;
+
+	cmdreg = cmd->opcode | (resptype << 8) | (cmdtype << 12);
+
+	if (host->bus_mode == MMC_BUSMODE_OPENDRAIN)
+		cmdreg |= 1 << 6;
+
+	if (cmd->flags & MMC_RSP_BUSY)
+		cmdreg |= 1 << 11;
+
+	if (host->data && !(host->data->flags & MMC_DATA_WRITE))
+		cmdreg |= 1 << 15;
+
+	GOLDFISH_MMC_WRITE(host, MMC_ARG, cmd->arg);
+	GOLDFISH_MMC_WRITE(host, MMC_CMD, cmdreg);
+}
+
+static void goldfish_mmc_xfer_done(struct goldfish_mmc_host *host,
+				   struct mmc_data *data)
+{
+	if (host->dma_in_use) {
+		enum dma_data_direction dma_data_dir;
+
+		dma_data_dir = mmc_get_dma_dir(data);
+
+		if (dma_data_dir == DMA_FROM_DEVICE) {
+			/*
+			 * We don't really have DMA, so we need
+			 * to copy from our platform driver buffer
+			 */
+			sg_copy_from_buffer(data->sg, 1, host->virt_base,
+					data->sg->length);
+		}
+		host->data->bytes_xfered += data->sg->length;
+		dma_unmap_sg(mmc_dev(mmc_from_priv(host)), data->sg,
+			     host->sg_len, dma_data_dir);
+	}
+
+	host->data = NULL;
+	host->sg_len = 0;
+
+	/*
+	 * NOTE:  MMC layer will sometimes poll-wait CMD13 next, issuing
+	 * dozens of requests until the card finishes writing data.
+	 * It'd be cheaper to just wait till an EOFB interrupt arrives...
+	 */
+
+	if (!data->stop) {
+		host->mrq = NULL;
+		mmc_request_done(mmc_from_priv(host), data->mrq);
+		return;
+	}
+
+	goldfish_mmc_start_command(host, data->stop);
+}
+
+static void goldfish_mmc_end_of_data(struct goldfish_mmc_host *host,
+				     struct mmc_data *data)
+{
+	if (!host->dma_in_use) {
+		goldfish_mmc_xfer_done(host, data);
+		return;
+	}
+	if (host->dma_done)
+		goldfish_mmc_xfer_done(host, data);
+}
+
+static void goldfish_mmc_cmd_done(struct goldfish_mmc_host *host,
+				  struct mmc_command *cmd)
+{
+	host->cmd = NULL;
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136) {
+			/* response type 2 */
+			cmd->resp[3] =
+				GOLDFISH_MMC_READ(host, MMC_RESP_0);
+			cmd->resp[2] =
+				GOLDFISH_MMC_READ(host, MMC_RESP_1);
+			cmd->resp[1] =
+				GOLDFISH_MMC_READ(host, MMC_RESP_2);
+			cmd->resp[0] =
+				GOLDFISH_MMC_READ(host, MMC_RESP_3);
+		} else {
+			/* response types 1, 1b, 3, 4, 5, 6 */
+			cmd->resp[0] =
+				GOLDFISH_MMC_READ(host, MMC_RESP_0);
+		}
+	}
+
+	if (host->data == NULL || cmd->error) {
+		host->mrq = NULL;
+		mmc_request_done(mmc_from_priv(host), cmd->mrq);
+	}
+}
+
+static irqreturn_t goldfish_mmc_irq(int irq, void *dev_id)
+{
+	struct goldfish_mmc_host *host = (struct goldfish_mmc_host *)dev_id;
+	u16 status;
+	int end_command = 0;
+	int end_transfer = 0;
+	int state_changed = 0;
+	int cmd_timeout = 0;
+
+	while ((status = GOLDFISH_MMC_READ(host, MMC_INT_STATUS)) != 0) {
+		GOLDFISH_MMC_WRITE(host, MMC_INT_STATUS, status);
+
+		if (status & MMC_STAT_END_OF_CMD)
+			end_command = 1;
+
+		if (status & MMC_STAT_END_OF_DATA)
+			end_transfer = 1;
+
+		if (status & MMC_STAT_STATE_CHANGE)
+			state_changed = 1;
+
+                if (status & MMC_STAT_CMD_TIMEOUT) {
+			end_command = 0;
+			cmd_timeout = 1;
+                }
+	}
+
+	if (cmd_timeout) {
+		struct mmc_request *mrq = host->mrq;
+		mrq->cmd->error = -ETIMEDOUT;
+		host->mrq = NULL;
+		mmc_request_done(mmc_from_priv(host), mrq);
+	}
+
+	if (end_command)
+		goldfish_mmc_cmd_done(host, host->cmd);
+
+	if (end_transfer) {
+		host->dma_done = 1;
+		goldfish_mmc_end_of_data(host, host->data);
+	} else if (host->data != NULL) {
+		/*
+		 * WORKAROUND -- after porting this driver from 2.6 to 3.4,
+		 * during device initialization, cases where host->data is
+		 * non-null but end_transfer is false would occur. Doing
+		 * nothing in such cases results in no further interrupts,
+		 * and initialization failure.
+		 * TODO -- find the real cause.
+		 */
+		host->dma_done = 1;
+		goldfish_mmc_end_of_data(host, host->data);
+	}
+
+	if (state_changed) {
+		u32 state = GOLDFISH_MMC_READ(host, MMC_STATE);
+		pr_info("%s: Card detect now %d\n", __func__,
+			(state & MMC_STATE_INSERTED));
+		mmc_detect_change(mmc_from_priv(host), 0);
+	}
+
+	if (!end_command && !end_transfer && !state_changed && !cmd_timeout) {
+		status = GOLDFISH_MMC_READ(host, MMC_INT_STATUS);
+		dev_info(mmc_dev(mmc_from_priv(host)), "spurious irq 0x%04x\n",
+			 status);
+		if (status != 0) {
+			GOLDFISH_MMC_WRITE(host, MMC_INT_STATUS, status);
+			GOLDFISH_MMC_WRITE(host, MMC_INT_ENABLE, 0);
+		}
+	}
+
+	return IRQ_HANDLED;
+}
+
+static void goldfish_mmc_prepare_data(struct goldfish_mmc_host *host,
+				      struct mmc_request *req)
+{
+	struct mmc_data *data = req->data;
+	int block_size;
+	unsigned sg_len;
+	enum dma_data_direction dma_data_dir;
+
+	host->data = data;
+	if (data == NULL) {
+		GOLDFISH_MMC_WRITE(host, MMC_BLOCK_LENGTH, 0);
+		GOLDFISH_MMC_WRITE(host, MMC_BLOCK_COUNT, 0);
+		host->dma_in_use = 0;
+		return;
+	}
+
+	block_size = data->blksz;
+
+	GOLDFISH_MMC_WRITE(host, MMC_BLOCK_COUNT, data->blocks - 1);
+	GOLDFISH_MMC_WRITE(host, MMC_BLOCK_LENGTH, block_size - 1);
+
+	/*
+	 * Cope with calling layer confusion; it issues "single
+	 * block" writes using multi-block scatterlists.
+	 */
+	sg_len = (data->blocks == 1) ? 1 : data->sg_len;
+
+	dma_data_dir = mmc_get_dma_dir(data);
+
+	host->sg_len = dma_map_sg(mmc_dev(mmc_from_priv(host)), data->sg,
+				  sg_len, dma_data_dir);
+	host->dma_done = 0;
+	host->dma_in_use = 1;
+
+	if (dma_data_dir == DMA_TO_DEVICE) {
+		/*
+		 * We don't really have DMA, so we need to copy to our
+		 * platform driver buffer
+		 */
+		sg_copy_to_buffer(data->sg, 1, host->virt_base,
+				data->sg->length);
+	}
+}
+
+static void goldfish_mmc_request(struct mmc_host *mmc, struct mmc_request *req)
+{
+	struct goldfish_mmc_host *host = mmc_priv(mmc);
+
+	WARN_ON(host->mrq != NULL);
+
+	host->mrq = req;
+	goldfish_mmc_prepare_data(host, req);
+	goldfish_mmc_start_command(host, req->cmd);
+}
+
+static void goldfish_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct goldfish_mmc_host *host = mmc_priv(mmc);
+
+	host->bus_mode = ios->bus_mode;
+	host->hw_bus_mode = host->bus_mode;
+}
+
+static int goldfish_mmc_get_ro(struct mmc_host *mmc)
+{
+	uint32_t state;
+	struct goldfish_mmc_host *host = mmc_priv(mmc);
+
+	state = GOLDFISH_MMC_READ(host, MMC_STATE);
+	return ((state & MMC_STATE_READ_ONLY) != 0);
+}
+
+static const struct mmc_host_ops goldfish_mmc_ops = {
+	.request	= goldfish_mmc_request,
+	.set_ios	= goldfish_mmc_set_ios,
+	.get_ro		= goldfish_mmc_get_ro,
+};
+
+static int goldfish_mmc_probe(struct platform_device *pdev)
+{
+	struct mmc_host *mmc;
+	struct goldfish_mmc_host *host = NULL;
+	struct resource *res;
+	int ret = 0;
+	int irq;
+	dma_addr_t buf_addr;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	irq = platform_get_irq(pdev, 0);
+	if (res == NULL || irq < 0)
+		return -ENXIO;
+
+	mmc = mmc_alloc_host(sizeof(struct goldfish_mmc_host), &pdev->dev);
+	if (mmc == NULL) {
+		ret = -ENOMEM;
+		goto err_alloc_host_failed;
+	}
+
+	host = mmc_priv(mmc);
+
+	pr_err("mmc: Mapping %lX to %lX\n", (long)res->start, (long)res->end);
+	host->reg_base = ioremap(res->start, resource_size(res));
+	if (host->reg_base == NULL) {
+		ret = -ENOMEM;
+		goto ioremap_failed;
+	}
+	host->virt_base = dma_alloc_coherent(&pdev->dev, BUFFER_SIZE,
+					     &buf_addr, GFP_KERNEL);
+
+	if (host->virt_base == 0) {
+		ret = -ENOMEM;
+		goto dma_alloc_failed;
+	}
+	host->phys_base = buf_addr;
+
+	host->id = pdev->id;
+	host->irq = irq;
+
+	mmc->ops = &goldfish_mmc_ops;
+	mmc->f_min = 400000;
+	mmc->f_max = 24000000;
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+	mmc->caps = MMC_CAP_4_BIT_DATA;
+	mmc->caps2 = MMC_CAP2_NO_SDIO;
+
+	/* Use scatterlist DMA to reduce per-transfer costs.
+	 * NOTE max_seg_size assumption that small blocks aren't
+	 * normally used (except e.g. for reading SD registers).
+	 */
+	mmc->max_segs = 32;
+	mmc->max_blk_size = 2048;	/* MMC_BLOCK_LENGTH is 11 bits (+1) */
+	mmc->max_blk_count = 2048;	/* MMC_BLOCK_COUNT is 11 bits (+1) */
+	mmc->max_req_size = BUFFER_SIZE;
+	mmc->max_seg_size = mmc->max_req_size;
+
+	ret = request_irq(host->irq, goldfish_mmc_irq, 0, DRIVER_NAME, host);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed IRQ Adding goldfish MMC\n");
+		goto err_request_irq_failed;
+	}
+
+	host->dev = &pdev->dev;
+	platform_set_drvdata(pdev, host);
+
+	ret = device_create_file(&pdev->dev, &dev_attr_cover_switch);
+	if (ret)
+		dev_warn(mmc_dev(mmc), "Unable to create sysfs attributes\n");
+
+	GOLDFISH_MMC_WRITE(host, MMC_SET_BUFFER, host->phys_base);
+	GOLDFISH_MMC_WRITE(host, MMC_INT_ENABLE,
+			   MMC_STAT_END_OF_CMD | MMC_STAT_END_OF_DATA |
+			   MMC_STAT_STATE_CHANGE | MMC_STAT_CMD_TIMEOUT);
+
+	mmc_add_host(mmc);
+	return 0;
+
+err_request_irq_failed:
+	dma_free_coherent(&pdev->dev, BUFFER_SIZE, host->virt_base,
+			  host->phys_base);
+dma_alloc_failed:
+	iounmap(host->reg_base);
+ioremap_failed:
+	mmc_free_host(mmc);
+err_alloc_host_failed:
+	return ret;
+}
+
+static int goldfish_mmc_remove(struct platform_device *pdev)
+{
+	struct goldfish_mmc_host *host = platform_get_drvdata(pdev);
+	struct mmc_host *mmc = mmc_from_priv(host);
+
+	BUG_ON(host == NULL);
+
+	mmc_remove_host(mmc);
+	free_irq(host->irq, host);
+	dma_free_coherent(&pdev->dev, BUFFER_SIZE, host->virt_base, host->phys_base);
+	iounmap(host->reg_base);
+	mmc_free_host(mmc);
+	return 0;
+}
+
+static struct platform_driver goldfish_mmc_driver = {
+	.probe		= goldfish_mmc_probe,
+	.remove		= goldfish_mmc_remove,
+	.driver		= {
+		.name	= DRIVER_NAME,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+	},
+};
+
+module_platform_driver(goldfish_mmc_driver);
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/atmel-mci.c b/drivers/mmc/host/atmel-mci.c
new file mode 100644
index 000000000..c468f9a02
--- /dev/null
+++ b/drivers/mmc/host/atmel-mci.c
@@ -0,0 +1,2680 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Atmel MultiMedia Card Interface driver
+ *
+ * Copyright (C) 2004-2008 Atmel Corporation
+ */
+#include <linux/blkdev.h>
+#include <linux/clk.h>
+#include <linux/debugfs.h>
+#include <linux/device.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/err.h>
+#include <linux/gpio.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/ioport.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/of_gpio.h>
+#include <linux/platform_device.h>
+#include <linux/scatterlist.h>
+#include <linux/seq_file.h>
+#include <linux/slab.h>
+#include <linux/stat.h>
+#include <linux/types.h>
+
+#include <linux/mmc/host.h>
+#include <linux/mmc/sdio.h>
+
+#include <linux/atmel-mci.h>
+#include <linux/atmel_pdc.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+#include <linux/pinctrl/consumer.h>
+
+#include <asm/cacheflush.h>
+#include <asm/io.h>
+#include <asm/unaligned.h>
+
+/*
+ * Superset of MCI IP registers integrated in Atmel AT91 Processor
+ * Registers and bitfields marked with [2] are only available in MCI2
+ */
+
+/* MCI Register Definitions */
+#define	ATMCI_CR			0x0000	/* Control */
+#define		ATMCI_CR_MCIEN			BIT(0)		/* MCI Enable */
+#define		ATMCI_CR_MCIDIS			BIT(1)		/* MCI Disable */
+#define		ATMCI_CR_PWSEN			BIT(2)		/* Power Save Enable */
+#define		ATMCI_CR_PWSDIS			BIT(3)		/* Power Save Disable */
+#define		ATMCI_CR_SWRST			BIT(7)		/* Software Reset */
+#define	ATMCI_MR			0x0004	/* Mode */
+#define		ATMCI_MR_CLKDIV(x)		((x) <<  0)	/* Clock Divider */
+#define		ATMCI_MR_PWSDIV(x)		((x) <<  8)	/* Power Saving Divider */
+#define		ATMCI_MR_RDPROOF		BIT(11)		/* Read Proof */
+#define		ATMCI_MR_WRPROOF		BIT(12)		/* Write Proof */
+#define		ATMCI_MR_PDCFBYTE		BIT(13)		/* Force Byte Transfer */
+#define		ATMCI_MR_PDCPADV		BIT(14)		/* Padding Value */
+#define		ATMCI_MR_PDCMODE		BIT(15)		/* PDC-oriented Mode */
+#define		ATMCI_MR_CLKODD(x)		((x) << 16)	/* LSB of Clock Divider */
+#define	ATMCI_DTOR			0x0008	/* Data Timeout */
+#define		ATMCI_DTOCYC(x)			((x) <<  0)	/* Data Timeout Cycles */
+#define		ATMCI_DTOMUL(x)			((x) <<  4)	/* Data Timeout Multiplier */
+#define	ATMCI_SDCR			0x000c	/* SD Card / SDIO */
+#define		ATMCI_SDCSEL_SLOT_A		(0 <<  0)	/* Select SD slot A */
+#define		ATMCI_SDCSEL_SLOT_B		(1 <<  0)	/* Select SD slot A */
+#define		ATMCI_SDCSEL_MASK		(3 <<  0)
+#define		ATMCI_SDCBUS_1BIT		(0 <<  6)	/* 1-bit data bus */
+#define		ATMCI_SDCBUS_4BIT		(2 <<  6)	/* 4-bit data bus */
+#define		ATMCI_SDCBUS_8BIT		(3 <<  6)	/* 8-bit data bus[2] */
+#define		ATMCI_SDCBUS_MASK		(3 <<  6)
+#define	ATMCI_ARGR			0x0010	/* Command Argument */
+#define	ATMCI_CMDR			0x0014	/* Command */
+#define		ATMCI_CMDR_CMDNB(x)		((x) <<  0)	/* Command Opcode */
+#define		ATMCI_CMDR_RSPTYP_NONE		(0 <<  6)	/* No response */
+#define		ATMCI_CMDR_RSPTYP_48BIT		(1 <<  6)	/* 48-bit response */
+#define		ATMCI_CMDR_RSPTYP_136BIT	(2 <<  6)	/* 136-bit response */
+#define		ATMCI_CMDR_SPCMD_INIT		(1 <<  8)	/* Initialization command */
+#define		ATMCI_CMDR_SPCMD_SYNC		(2 <<  8)	/* Synchronized command */
+#define		ATMCI_CMDR_SPCMD_INT		(4 <<  8)	/* Interrupt command */
+#define		ATMCI_CMDR_SPCMD_INTRESP	(5 <<  8)	/* Interrupt response */
+#define		ATMCI_CMDR_OPDCMD		(1 << 11)	/* Open Drain */
+#define		ATMCI_CMDR_MAXLAT_5CYC		(0 << 12)	/* Max latency 5 cycles */
+#define		ATMCI_CMDR_MAXLAT_64CYC		(1 << 12)	/* Max latency 64 cycles */
+#define		ATMCI_CMDR_START_XFER		(1 << 16)	/* Start data transfer */
+#define		ATMCI_CMDR_STOP_XFER		(2 << 16)	/* Stop data transfer */
+#define		ATMCI_CMDR_TRDIR_WRITE		(0 << 18)	/* Write data */
+#define		ATMCI_CMDR_TRDIR_READ		(1 << 18)	/* Read data */
+#define		ATMCI_CMDR_BLOCK		(0 << 19)	/* Single-block transfer */
+#define		ATMCI_CMDR_MULTI_BLOCK		(1 << 19)	/* Multi-block transfer */
+#define		ATMCI_CMDR_STREAM		(2 << 19)	/* MMC Stream transfer */
+#define		ATMCI_CMDR_SDIO_BYTE		(4 << 19)	/* SDIO Byte transfer */
+#define		ATMCI_CMDR_SDIO_BLOCK		(5 << 19)	/* SDIO Block transfer */
+#define		ATMCI_CMDR_SDIO_SUSPEND		(1 << 24)	/* SDIO Suspend Command */
+#define		ATMCI_CMDR_SDIO_RESUME		(2 << 24)	/* SDIO Resume Command */
+#define	ATMCI_BLKR			0x0018	/* Block */
+#define		ATMCI_BCNT(x)			((x) <<  0)	/* Data Block Count */
+#define		ATMCI_BLKLEN(x)			((x) << 16)	/* Data Block Length */
+#define	ATMCI_CSTOR			0x001c	/* Completion Signal Timeout[2] */
+#define		ATMCI_CSTOCYC(x)		((x) <<  0)	/* CST cycles */
+#define		ATMCI_CSTOMUL(x)		((x) <<  4)	/* CST multiplier */
+#define	ATMCI_RSPR			0x0020	/* Response 0 */
+#define	ATMCI_RSPR1			0x0024	/* Response 1 */
+#define	ATMCI_RSPR2			0x0028	/* Response 2 */
+#define	ATMCI_RSPR3			0x002c	/* Response 3 */
+#define	ATMCI_RDR			0x0030	/* Receive Data */
+#define	ATMCI_TDR			0x0034	/* Transmit Data */
+#define	ATMCI_SR			0x0040	/* Status */
+#define	ATMCI_IER			0x0044	/* Interrupt Enable */
+#define	ATMCI_IDR			0x0048	/* Interrupt Disable */
+#define	ATMCI_IMR			0x004c	/* Interrupt Mask */
+#define		ATMCI_CMDRDY			BIT(0)		/* Command Ready */
+#define		ATMCI_RXRDY			BIT(1)		/* Receiver Ready */
+#define		ATMCI_TXRDY			BIT(2)		/* Transmitter Ready */
+#define		ATMCI_BLKE			BIT(3)		/* Data Block Ended */
+#define		ATMCI_DTIP			BIT(4)		/* Data Transfer In Progress */
+#define		ATMCI_NOTBUSY			BIT(5)		/* Data Not Busy */
+#define		ATMCI_ENDRX			BIT(6)		/* End of RX Buffer */
+#define		ATMCI_ENDTX			BIT(7)		/* End of TX Buffer */
+#define		ATMCI_SDIOIRQA			BIT(8)		/* SDIO IRQ in slot A */
+#define		ATMCI_SDIOIRQB			BIT(9)		/* SDIO IRQ in slot B */
+#define		ATMCI_SDIOWAIT			BIT(12)		/* SDIO Read Wait Operation Status */
+#define		ATMCI_CSRCV			BIT(13)		/* CE-ATA Completion Signal Received */
+#define		ATMCI_RXBUFF			BIT(14)		/* RX Buffer Full */
+#define		ATMCI_TXBUFE			BIT(15)		/* TX Buffer Empty */
+#define		ATMCI_RINDE			BIT(16)		/* Response Index Error */
+#define		ATMCI_RDIRE			BIT(17)		/* Response Direction Error */
+#define		ATMCI_RCRCE			BIT(18)		/* Response CRC Error */
+#define		ATMCI_RENDE			BIT(19)		/* Response End Bit Error */
+#define		ATMCI_RTOE			BIT(20)		/* Response Time-Out Error */
+#define		ATMCI_DCRCE			BIT(21)		/* Data CRC Error */
+#define		ATMCI_DTOE			BIT(22)		/* Data Time-Out Error */
+#define		ATMCI_CSTOE			BIT(23)		/* Completion Signal Time-out Error */
+#define		ATMCI_BLKOVRE			BIT(24)		/* DMA Block Overrun Error */
+#define		ATMCI_DMADONE			BIT(25)		/* DMA Transfer Done */
+#define		ATMCI_FIFOEMPTY			BIT(26)		/* FIFO Empty Flag */
+#define		ATMCI_XFRDONE			BIT(27)		/* Transfer Done Flag */
+#define		ATMCI_ACKRCV			BIT(28)		/* Boot Operation Acknowledge Received */
+#define		ATMCI_ACKRCVE			BIT(29)		/* Boot Operation Acknowledge Error */
+#define		ATMCI_OVRE			BIT(30)		/* RX Overrun Error */
+#define		ATMCI_UNRE			BIT(31)		/* TX Underrun Error */
+#define	ATMCI_DMA			0x0050	/* DMA Configuration[2] */
+#define		ATMCI_DMA_OFFSET(x)		((x) <<  0)	/* DMA Write Buffer Offset */
+#define		ATMCI_DMA_CHKSIZE(x)		((x) <<  4)	/* DMA Channel Read and Write Chunk Size */
+#define		ATMCI_DMAEN			BIT(8)	/* DMA Hardware Handshaking Enable */
+#define	ATMCI_CFG			0x0054	/* Configuration[2] */
+#define		ATMCI_CFG_FIFOMODE_1DATA	BIT(0)		/* MCI Internal FIFO control mode */
+#define		ATMCI_CFG_FERRCTRL_COR		BIT(4)		/* Flow Error flag reset control mode */
+#define		ATMCI_CFG_HSMODE		BIT(8)		/* High Speed Mode */
+#define		ATMCI_CFG_LSYNC			BIT(12)		/* Synchronize on the last block */
+#define	ATMCI_WPMR			0x00e4	/* Write Protection Mode[2] */
+#define		ATMCI_WP_EN			BIT(0)		/* WP Enable */
+#define		ATMCI_WP_KEY			(0x4d4349 << 8)	/* WP Key */
+#define	ATMCI_WPSR			0x00e8	/* Write Protection Status[2] */
+#define		ATMCI_GET_WP_VS(x)		((x) & 0x0f)
+#define		ATMCI_GET_WP_VSRC(x)		(((x) >> 8) & 0xffff)
+#define	ATMCI_VERSION			0x00FC  /* Version */
+#define	ATMCI_FIFO_APERTURE		0x0200	/* FIFO Aperture[2] */
+
+/* This is not including the FIFO Aperture on MCI2 */
+#define	ATMCI_REGS_SIZE		0x100
+
+/* Register access macros */
+#define	atmci_readl(port, reg)				\
+	__raw_readl((port)->regs + reg)
+#define	atmci_writel(port, reg, value)			\
+	__raw_writel((value), (port)->regs + reg)
+
+#define ATMCI_CMD_TIMEOUT_MS	2000
+#define AUTOSUSPEND_DELAY	50
+
+#define ATMCI_DATA_ERROR_FLAGS	(ATMCI_DCRCE | ATMCI_DTOE | ATMCI_OVRE | ATMCI_UNRE)
+#define ATMCI_DMA_THRESHOLD	16
+
+enum {
+	EVENT_CMD_RDY = 0,
+	EVENT_XFER_COMPLETE,
+	EVENT_NOTBUSY,
+	EVENT_DATA_ERROR,
+};
+
+enum atmel_mci_state {
+	STATE_IDLE = 0,
+	STATE_SENDING_CMD,
+	STATE_DATA_XFER,
+	STATE_WAITING_NOTBUSY,
+	STATE_SENDING_STOP,
+	STATE_END_REQUEST,
+};
+
+enum atmci_xfer_dir {
+	XFER_RECEIVE = 0,
+	XFER_TRANSMIT,
+};
+
+enum atmci_pdc_buf {
+	PDC_FIRST_BUF = 0,
+	PDC_SECOND_BUF,
+};
+
+struct atmel_mci_caps {
+	bool    has_dma_conf_reg;
+	bool    has_pdc;
+	bool    has_cfg_reg;
+	bool    has_cstor_reg;
+	bool    has_highspeed;
+	bool    has_rwproof;
+	bool	has_odd_clk_div;
+	bool	has_bad_data_ordering;
+	bool	need_reset_after_xfer;
+	bool	need_blksz_mul_4;
+	bool	need_notbusy_for_read_ops;
+};
+
+struct atmel_mci_dma {
+	struct dma_chan			*chan;
+	struct dma_async_tx_descriptor	*data_desc;
+};
+
+/**
+ * struct atmel_mci - MMC controller state shared between all slots
+ * @lock: Spinlock protecting the queue and associated data.
+ * @regs: Pointer to MMIO registers.
+ * @sg: Scatterlist entry currently being processed by PIO or PDC code.
+ * @sg_len: Size of the scatterlist
+ * @pio_offset: Offset into the current scatterlist entry.
+ * @buffer: Buffer used if we don't have the r/w proof capability. We
+ *      don't have the time to switch pdc buffers so we have to use only
+ *      one buffer for the full transaction.
+ * @buf_size: size of the buffer.
+ * @buf_phys_addr: buffer address needed for pdc.
+ * @cur_slot: The slot which is currently using the controller.
+ * @mrq: The request currently being processed on @cur_slot,
+ *	or NULL if the controller is idle.
+ * @cmd: The command currently being sent to the card, or NULL.
+ * @data: The data currently being transferred, or NULL if no data
+ *	transfer is in progress.
+ * @data_size: just data->blocks * data->blksz.
+ * @dma: DMA client state.
+ * @data_chan: DMA channel being used for the current data transfer.
+ * @dma_conf: Configuration for the DMA slave
+ * @cmd_status: Snapshot of SR taken upon completion of the current
+ *	command. Only valid when EVENT_CMD_COMPLETE is pending.
+ * @data_status: Snapshot of SR taken upon completion of the current
+ *	data transfer. Only valid when EVENT_DATA_COMPLETE or
+ *	EVENT_DATA_ERROR is pending.
+ * @stop_cmdr: Value to be loaded into CMDR when the stop command is
+ *	to be sent.
+ * @tasklet: Tasklet running the request state machine.
+ * @pending_events: Bitmask of events flagged by the interrupt handler
+ *	to be processed by the tasklet.
+ * @completed_events: Bitmask of events which the state machine has
+ *	processed.
+ * @state: Tasklet state.
+ * @queue: List of slots waiting for access to the controller.
+ * @need_clock_update: Update the clock rate before the next request.
+ * @need_reset: Reset controller before next request.
+ * @timer: Timer to balance the data timeout error flag which cannot rise.
+ * @mode_reg: Value of the MR register.
+ * @cfg_reg: Value of the CFG register.
+ * @bus_hz: The rate of @mck in Hz. This forms the basis for MMC bus
+ *	rate and timeout calculations.
+ * @mapbase: Physical address of the MMIO registers.
+ * @mck: The peripheral bus clock hooked up to the MMC controller.
+ * @pdev: Platform device associated with the MMC controller.
+ * @slot: Slots sharing this MMC controller.
+ * @caps: MCI capabilities depending on MCI version.
+ * @prepare_data: function to setup MCI before data transfer which
+ * depends on MCI capabilities.
+ * @submit_data: function to start data transfer which depends on MCI
+ * capabilities.
+ * @stop_transfer: function to stop data transfer which depends on MCI
+ * capabilities.
+ *
+ * Locking
+ * =======
+ *
+ * @lock is a softirq-safe spinlock protecting @queue as well as
+ * @cur_slot, @mrq and @state. These must always be updated
+ * at the same time while holding @lock.
+ *
+ * @lock also protects mode_reg and need_clock_update since these are
+ * used to synchronize mode register updates with the queue
+ * processing.
+ *
+ * The @mrq field of struct atmel_mci_slot is also protected by @lock,
+ * and must always be written at the same time as the slot is added to
+ * @queue.
+ *
+ * @pending_events and @completed_events are accessed using atomic bit
+ * operations, so they don't need any locking.
+ *
+ * None of the fields touched by the interrupt handler need any
+ * locking. However, ordering is important: Before EVENT_DATA_ERROR or
+ * EVENT_DATA_COMPLETE is set in @pending_events, all data-related
+ * interrupts must be disabled and @data_status updated with a
+ * snapshot of SR. Similarly, before EVENT_CMD_COMPLETE is set, the
+ * CMDRDY interrupt must be disabled and @cmd_status updated with a
+ * snapshot of SR, and before EVENT_XFER_COMPLETE can be set, the
+ * bytes_xfered field of @data must be written. This is ensured by
+ * using barriers.
+ */
+struct atmel_mci {
+	spinlock_t		lock;
+	void __iomem		*regs;
+
+	struct scatterlist	*sg;
+	unsigned int		sg_len;
+	unsigned int		pio_offset;
+	unsigned int		*buffer;
+	unsigned int		buf_size;
+	dma_addr_t		buf_phys_addr;
+
+	struct atmel_mci_slot	*cur_slot;
+	struct mmc_request	*mrq;
+	struct mmc_command	*cmd;
+	struct mmc_data		*data;
+	unsigned int		data_size;
+
+	struct atmel_mci_dma	dma;
+	struct dma_chan		*data_chan;
+	struct dma_slave_config	dma_conf;
+
+	u32			cmd_status;
+	u32			data_status;
+	u32			stop_cmdr;
+
+	struct tasklet_struct	tasklet;
+	unsigned long		pending_events;
+	unsigned long		completed_events;
+	enum atmel_mci_state	state;
+	struct list_head	queue;
+
+	bool			need_clock_update;
+	bool			need_reset;
+	struct timer_list	timer;
+	u32			mode_reg;
+	u32			cfg_reg;
+	unsigned long		bus_hz;
+	unsigned long		mapbase;
+	struct clk		*mck;
+	struct platform_device	*pdev;
+
+	struct atmel_mci_slot	*slot[ATMCI_MAX_NR_SLOTS];
+
+	struct atmel_mci_caps   caps;
+
+	u32 (*prepare_data)(struct atmel_mci *host, struct mmc_data *data);
+	void (*submit_data)(struct atmel_mci *host, struct mmc_data *data);
+	void (*stop_transfer)(struct atmel_mci *host);
+};
+
+/**
+ * struct atmel_mci_slot - MMC slot state
+ * @mmc: The mmc_host representing this slot.
+ * @host: The MMC controller this slot is using.
+ * @sdc_reg: Value of SDCR to be written before using this slot.
+ * @sdio_irq: SDIO irq mask for this slot.
+ * @mrq: mmc_request currently being processed or waiting to be
+ *	processed, or NULL when the slot is idle.
+ * @queue_node: List node for placing this node in the @queue list of
+ *	&struct atmel_mci.
+ * @clock: Clock rate configured by set_ios(). Protected by host->lock.
+ * @flags: Random state bits associated with the slot.
+ * @detect_pin: GPIO pin used for card detection, or negative if not
+ *	available.
+ * @wp_pin: GPIO pin used for card write protect sending, or negative
+ *	if not available.
+ * @detect_is_active_high: The state of the detect pin when it is active.
+ * @detect_timer: Timer used for debouncing @detect_pin interrupts.
+ */
+struct atmel_mci_slot {
+	struct mmc_host		*mmc;
+	struct atmel_mci	*host;
+
+	u32			sdc_reg;
+	u32			sdio_irq;
+
+	struct mmc_request	*mrq;
+	struct list_head	queue_node;
+
+	unsigned int		clock;
+	unsigned long		flags;
+#define ATMCI_CARD_PRESENT	0
+#define ATMCI_CARD_NEED_INIT	1
+#define ATMCI_SHUTDOWN		2
+
+	int			detect_pin;
+	int			wp_pin;
+	bool			detect_is_active_high;
+
+	struct timer_list	detect_timer;
+};
+
+#define atmci_test_and_clear_pending(host, event)		\
+	test_and_clear_bit(event, &host->pending_events)
+#define atmci_set_completed(host, event)			\
+	set_bit(event, &host->completed_events)
+#define atmci_set_pending(host, event)				\
+	set_bit(event, &host->pending_events)
+
+/*
+ * The debugfs stuff below is mostly optimized away when
+ * CONFIG_DEBUG_FS is not set.
+ */
+static int atmci_req_show(struct seq_file *s, void *v)
+{
+	struct atmel_mci_slot	*slot = s->private;
+	struct mmc_request	*mrq;
+	struct mmc_command	*cmd;
+	struct mmc_command	*stop;
+	struct mmc_data		*data;
+
+	/* Make sure we get a consistent snapshot */
+	spin_lock_bh(&slot->host->lock);
+	mrq = slot->mrq;
+
+	if (mrq) {
+		cmd = mrq->cmd;
+		data = mrq->data;
+		stop = mrq->stop;
+
+		if (cmd)
+			seq_printf(s,
+				"CMD%u(0x%x) flg %x rsp %x %x %x %x err %d\n",
+				cmd->opcode, cmd->arg, cmd->flags,
+				cmd->resp[0], cmd->resp[1], cmd->resp[2],
+				cmd->resp[3], cmd->error);
+		if (data)
+			seq_printf(s, "DATA %u / %u * %u flg %x err %d\n",
+				data->bytes_xfered, data->blocks,
+				data->blksz, data->flags, data->error);
+		if (stop)
+			seq_printf(s,
+				"CMD%u(0x%x) flg %x rsp %x %x %x %x err %d\n",
+				stop->opcode, stop->arg, stop->flags,
+				stop->resp[0], stop->resp[1], stop->resp[2],
+				stop->resp[3], stop->error);
+	}
+
+	spin_unlock_bh(&slot->host->lock);
+
+	return 0;
+}
+
+DEFINE_SHOW_ATTRIBUTE(atmci_req);
+
+static void atmci_show_status_reg(struct seq_file *s,
+		const char *regname, u32 value)
+{
+	static const char	*sr_bit[] = {
+		[0]	= "CMDRDY",
+		[1]	= "RXRDY",
+		[2]	= "TXRDY",
+		[3]	= "BLKE",
+		[4]	= "DTIP",
+		[5]	= "NOTBUSY",
+		[6]	= "ENDRX",
+		[7]	= "ENDTX",
+		[8]	= "SDIOIRQA",
+		[9]	= "SDIOIRQB",
+		[12]	= "SDIOWAIT",
+		[14]	= "RXBUFF",
+		[15]	= "TXBUFE",
+		[16]	= "RINDE",
+		[17]	= "RDIRE",
+		[18]	= "RCRCE",
+		[19]	= "RENDE",
+		[20]	= "RTOE",
+		[21]	= "DCRCE",
+		[22]	= "DTOE",
+		[23]	= "CSTOE",
+		[24]	= "BLKOVRE",
+		[25]	= "DMADONE",
+		[26]	= "FIFOEMPTY",
+		[27]	= "XFRDONE",
+		[30]	= "OVRE",
+		[31]	= "UNRE",
+	};
+	unsigned int		i;
+
+	seq_printf(s, "%s:\t0x%08x", regname, value);
+	for (i = 0; i < ARRAY_SIZE(sr_bit); i++) {
+		if (value & (1 << i)) {
+			if (sr_bit[i])
+				seq_printf(s, " %s", sr_bit[i]);
+			else
+				seq_puts(s, " UNKNOWN");
+		}
+	}
+	seq_putc(s, '\n');
+}
+
+static int atmci_regs_show(struct seq_file *s, void *v)
+{
+	struct atmel_mci	*host = s->private;
+	u32			*buf;
+	int			ret = 0;
+
+
+	buf = kmalloc(ATMCI_REGS_SIZE, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	pm_runtime_get_sync(&host->pdev->dev);
+
+	/*
+	 * Grab a more or less consistent snapshot. Note that we're
+	 * not disabling interrupts, so IMR and SR may not be
+	 * consistent.
+	 */
+	spin_lock_bh(&host->lock);
+	memcpy_fromio(buf, host->regs, ATMCI_REGS_SIZE);
+	spin_unlock_bh(&host->lock);
+
+	pm_runtime_mark_last_busy(&host->pdev->dev);
+	pm_runtime_put_autosuspend(&host->pdev->dev);
+
+	seq_printf(s, "MR:\t0x%08x%s%s ",
+			buf[ATMCI_MR / 4],
+			buf[ATMCI_MR / 4] & ATMCI_MR_RDPROOF ? " RDPROOF" : "",
+			buf[ATMCI_MR / 4] & ATMCI_MR_WRPROOF ? " WRPROOF" : "");
+	if (host->caps.has_odd_clk_div)
+		seq_printf(s, "{CLKDIV,CLKODD}=%u\n",
+				((buf[ATMCI_MR / 4] & 0xff) << 1)
+				| ((buf[ATMCI_MR / 4] >> 16) & 1));
+	else
+		seq_printf(s, "CLKDIV=%u\n",
+				(buf[ATMCI_MR / 4] & 0xff));
+	seq_printf(s, "DTOR:\t0x%08x\n", buf[ATMCI_DTOR / 4]);
+	seq_printf(s, "SDCR:\t0x%08x\n", buf[ATMCI_SDCR / 4]);
+	seq_printf(s, "ARGR:\t0x%08x\n", buf[ATMCI_ARGR / 4]);
+	seq_printf(s, "BLKR:\t0x%08x BCNT=%u BLKLEN=%u\n",
+			buf[ATMCI_BLKR / 4],
+			buf[ATMCI_BLKR / 4] & 0xffff,
+			(buf[ATMCI_BLKR / 4] >> 16) & 0xffff);
+	if (host->caps.has_cstor_reg)
+		seq_printf(s, "CSTOR:\t0x%08x\n", buf[ATMCI_CSTOR / 4]);
+
+	/* Don't read RSPR and RDR; it will consume the data there */
+
+	atmci_show_status_reg(s, "SR", buf[ATMCI_SR / 4]);
+	atmci_show_status_reg(s, "IMR", buf[ATMCI_IMR / 4]);
+
+	if (host->caps.has_dma_conf_reg) {
+		u32 val;
+
+		val = buf[ATMCI_DMA / 4];
+		seq_printf(s, "DMA:\t0x%08x OFFSET=%u CHKSIZE=%u%s\n",
+				val, val & 3,
+				((val >> 4) & 3) ?
+					1 << (((val >> 4) & 3) + 1) : 1,
+				val & ATMCI_DMAEN ? " DMAEN" : "");
+	}
+	if (host->caps.has_cfg_reg) {
+		u32 val;
+
+		val = buf[ATMCI_CFG / 4];
+		seq_printf(s, "CFG:\t0x%08x%s%s%s%s\n",
+				val,
+				val & ATMCI_CFG_FIFOMODE_1DATA ? " FIFOMODE_ONE_DATA" : "",
+				val & ATMCI_CFG_FERRCTRL_COR ? " FERRCTRL_CLEAR_ON_READ" : "",
+				val & ATMCI_CFG_HSMODE ? " HSMODE" : "",
+				val & ATMCI_CFG_LSYNC ? " LSYNC" : "");
+	}
+
+	kfree(buf);
+
+	return ret;
+}
+
+DEFINE_SHOW_ATTRIBUTE(atmci_regs);
+
+static void atmci_init_debugfs(struct atmel_mci_slot *slot)
+{
+	struct mmc_host		*mmc = slot->mmc;
+	struct atmel_mci	*host = slot->host;
+	struct dentry		*root;
+
+	root = mmc->debugfs_root;
+	if (!root)
+		return;
+
+	debugfs_create_file("regs", S_IRUSR, root, host, &atmci_regs_fops);
+	debugfs_create_file("req", S_IRUSR, root, slot, &atmci_req_fops);
+	debugfs_create_u32("state", S_IRUSR, root, &host->state);
+	debugfs_create_xul("pending_events", S_IRUSR, root,
+			   &host->pending_events);
+	debugfs_create_xul("completed_events", S_IRUSR, root,
+			   &host->completed_events);
+}
+
+#if defined(CONFIG_OF)
+static const struct of_device_id atmci_dt_ids[] = {
+	{ .compatible = "atmel,hsmci" },
+	{ /* sentinel */ }
+};
+
+MODULE_DEVICE_TABLE(of, atmci_dt_ids);
+
+static struct mci_platform_data*
+atmci_of_init(struct platform_device *pdev)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct device_node *cnp;
+	struct mci_platform_data *pdata;
+	u32 slot_id;
+
+	if (!np) {
+		dev_err(&pdev->dev, "device node not found\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata)
+		return ERR_PTR(-ENOMEM);
+
+	for_each_child_of_node(np, cnp) {
+		if (of_property_read_u32(cnp, "reg", &slot_id)) {
+			dev_warn(&pdev->dev, "reg property is missing for %pOF\n",
+				 cnp);
+			continue;
+		}
+
+		if (slot_id >= ATMCI_MAX_NR_SLOTS) {
+			dev_warn(&pdev->dev, "can't have more than %d slots\n",
+			         ATMCI_MAX_NR_SLOTS);
+			of_node_put(cnp);
+			break;
+		}
+
+		if (of_property_read_u32(cnp, "bus-width",
+		                         &pdata->slot[slot_id].bus_width))
+			pdata->slot[slot_id].bus_width = 1;
+
+		pdata->slot[slot_id].detect_pin =
+			of_get_named_gpio(cnp, "cd-gpios", 0);
+
+		pdata->slot[slot_id].detect_is_active_high =
+			of_property_read_bool(cnp, "cd-inverted");
+
+		pdata->slot[slot_id].non_removable =
+			of_property_read_bool(cnp, "non-removable");
+
+		pdata->slot[slot_id].wp_pin =
+			of_get_named_gpio(cnp, "wp-gpios", 0);
+	}
+
+	return pdata;
+}
+#else /* CONFIG_OF */
+static inline struct mci_platform_data*
+atmci_of_init(struct platform_device *dev)
+{
+	return ERR_PTR(-EINVAL);
+}
+#endif
+
+static inline unsigned int atmci_get_version(struct atmel_mci *host)
+{
+	return atmci_readl(host, ATMCI_VERSION) & 0x00000fff;
+}
+
+/*
+ * Fix sconfig's burst size according to atmel MCI. We need to convert them as:
+ * 1 -> 0, 4 -> 1, 8 -> 2, 16 -> 3.
+ * With version 0x600, we need to convert them as: 1 -> 0, 2 -> 1, 4 -> 2,
+ * 8 -> 3, 16 -> 4.
+ *
+ * This can be done by finding most significant bit set.
+ */
+static inline unsigned int atmci_convert_chksize(struct atmel_mci *host,
+						 unsigned int maxburst)
+{
+	unsigned int version = atmci_get_version(host);
+	unsigned int offset = 2;
+
+	if (version >= 0x600)
+		offset = 1;
+
+	if (maxburst > 1)
+		return fls(maxburst) - offset;
+	else
+		return 0;
+}
+
+static void atmci_timeout_timer(struct timer_list *t)
+{
+	struct atmel_mci *host;
+
+	host = from_timer(host, t, timer);
+
+	dev_dbg(&host->pdev->dev, "software timeout\n");
+
+	if (host->mrq->cmd->data) {
+		host->mrq->cmd->data->error = -ETIMEDOUT;
+		host->data = NULL;
+		/*
+		 * With some SDIO modules, sometimes DMA transfer hangs. If
+		 * stop_transfer() is not called then the DMA request is not
+		 * removed, following ones are queued and never computed.
+		 */
+		if (host->state == STATE_DATA_XFER)
+			host->stop_transfer(host);
+	} else {
+		host->mrq->cmd->error = -ETIMEDOUT;
+		host->cmd = NULL;
+	}
+	host->need_reset = 1;
+	host->state = STATE_END_REQUEST;
+	smp_wmb();
+	tasklet_schedule(&host->tasklet);
+}
+
+static inline unsigned int atmci_ns_to_clocks(struct atmel_mci *host,
+					unsigned int ns)
+{
+	/*
+	 * It is easier here to use us instead of ns for the timeout,
+	 * it prevents from overflows during calculation.
+	 */
+	unsigned int us = DIV_ROUND_UP(ns, 1000);
+
+	/* Maximum clock frequency is host->bus_hz/2 */
+	return us * (DIV_ROUND_UP(host->bus_hz, 2000000));
+}
+
+static void atmci_set_timeout(struct atmel_mci *host,
+		struct atmel_mci_slot *slot, struct mmc_data *data)
+{
+	static unsigned	dtomul_to_shift[] = {
+		0, 4, 7, 8, 10, 12, 16, 20
+	};
+	unsigned	timeout;
+	unsigned	dtocyc;
+	unsigned	dtomul;
+
+	timeout = atmci_ns_to_clocks(host, data->timeout_ns)
+		+ data->timeout_clks;
+
+	for (dtomul = 0; dtomul < 8; dtomul++) {
+		unsigned shift = dtomul_to_shift[dtomul];
+		dtocyc = (timeout + (1 << shift) - 1) >> shift;
+		if (dtocyc < 15)
+			break;
+	}
+
+	if (dtomul >= 8) {
+		dtomul = 7;
+		dtocyc = 15;
+	}
+
+	dev_vdbg(&slot->mmc->class_dev, "setting timeout to %u cycles\n",
+			dtocyc << dtomul_to_shift[dtomul]);
+	atmci_writel(host, ATMCI_DTOR, (ATMCI_DTOMUL(dtomul) | ATMCI_DTOCYC(dtocyc)));
+}
+
+/*
+ * Return mask with command flags to be enabled for this command.
+ */
+static u32 atmci_prepare_command(struct mmc_host *mmc,
+				 struct mmc_command *cmd)
+{
+	struct mmc_data	*data;
+	u32		cmdr;
+
+	cmd->error = -EINPROGRESS;
+
+	cmdr = ATMCI_CMDR_CMDNB(cmd->opcode);
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136)
+			cmdr |= ATMCI_CMDR_RSPTYP_136BIT;
+		else
+			cmdr |= ATMCI_CMDR_RSPTYP_48BIT;
+	}
+
+	/*
+	 * This should really be MAXLAT_5 for CMD2 and ACMD41, but
+	 * it's too difficult to determine whether this is an ACMD or
+	 * not. Better make it 64.
+	 */
+	cmdr |= ATMCI_CMDR_MAXLAT_64CYC;
+
+	if (mmc->ios.bus_mode == MMC_BUSMODE_OPENDRAIN)
+		cmdr |= ATMCI_CMDR_OPDCMD;
+
+	data = cmd->data;
+	if (data) {
+		cmdr |= ATMCI_CMDR_START_XFER;
+
+		if (cmd->opcode == SD_IO_RW_EXTENDED) {
+			cmdr |= ATMCI_CMDR_SDIO_BLOCK;
+		} else {
+			if (data->blocks > 1)
+				cmdr |= ATMCI_CMDR_MULTI_BLOCK;
+			else
+				cmdr |= ATMCI_CMDR_BLOCK;
+		}
+
+		if (data->flags & MMC_DATA_READ)
+			cmdr |= ATMCI_CMDR_TRDIR_READ;
+	}
+
+	return cmdr;
+}
+
+static void atmci_send_command(struct atmel_mci *host,
+		struct mmc_command *cmd, u32 cmd_flags)
+{
+	unsigned int timeout_ms = cmd->busy_timeout ? cmd->busy_timeout :
+		ATMCI_CMD_TIMEOUT_MS;
+
+	WARN_ON(host->cmd);
+	host->cmd = cmd;
+
+	dev_vdbg(&host->pdev->dev,
+			"start command: ARGR=0x%08x CMDR=0x%08x\n",
+			cmd->arg, cmd_flags);
+
+	atmci_writel(host, ATMCI_ARGR, cmd->arg);
+	atmci_writel(host, ATMCI_CMDR, cmd_flags);
+
+	mod_timer(&host->timer, jiffies + msecs_to_jiffies(timeout_ms));
+}
+
+static void atmci_send_stop_cmd(struct atmel_mci *host, struct mmc_data *data)
+{
+	dev_dbg(&host->pdev->dev, "send stop command\n");
+	atmci_send_command(host, data->stop, host->stop_cmdr);
+	atmci_writel(host, ATMCI_IER, ATMCI_CMDRDY);
+}
+
+/*
+ * Configure given PDC buffer taking care of alignement issues.
+ * Update host->data_size and host->sg.
+ */
+static void atmci_pdc_set_single_buf(struct atmel_mci *host,
+	enum atmci_xfer_dir dir, enum atmci_pdc_buf buf_nb)
+{
+	u32 pointer_reg, counter_reg;
+	unsigned int buf_size;
+
+	if (dir == XFER_RECEIVE) {
+		pointer_reg = ATMEL_PDC_RPR;
+		counter_reg = ATMEL_PDC_RCR;
+	} else {
+		pointer_reg = ATMEL_PDC_TPR;
+		counter_reg = ATMEL_PDC_TCR;
+	}
+
+	if (buf_nb == PDC_SECOND_BUF) {
+		pointer_reg += ATMEL_PDC_SCND_BUF_OFF;
+		counter_reg += ATMEL_PDC_SCND_BUF_OFF;
+	}
+
+	if (!host->caps.has_rwproof) {
+		buf_size = host->buf_size;
+		atmci_writel(host, pointer_reg, host->buf_phys_addr);
+	} else {
+		buf_size = sg_dma_len(host->sg);
+		atmci_writel(host, pointer_reg, sg_dma_address(host->sg));
+	}
+
+	if (host->data_size <= buf_size) {
+		if (host->data_size & 0x3) {
+			/* If size is different from modulo 4, transfer bytes */
+			atmci_writel(host, counter_reg, host->data_size);
+			atmci_writel(host, ATMCI_MR, host->mode_reg | ATMCI_MR_PDCFBYTE);
+		} else {
+			/* Else transfer 32-bits words */
+			atmci_writel(host, counter_reg, host->data_size / 4);
+		}
+		host->data_size = 0;
+	} else {
+		/* We assume the size of a page is 32-bits aligned */
+		atmci_writel(host, counter_reg, sg_dma_len(host->sg) / 4);
+		host->data_size -= sg_dma_len(host->sg);
+		if (host->data_size)
+			host->sg = sg_next(host->sg);
+	}
+}
+
+/*
+ * Configure PDC buffer according to the data size ie configuring one or two
+ * buffers. Don't use this function if you want to configure only the second
+ * buffer. In this case, use atmci_pdc_set_single_buf.
+ */
+static void atmci_pdc_set_both_buf(struct atmel_mci *host, int dir)
+{
+	atmci_pdc_set_single_buf(host, dir, PDC_FIRST_BUF);
+	if (host->data_size)
+		atmci_pdc_set_single_buf(host, dir, PDC_SECOND_BUF);
+}
+
+/*
+ * Unmap sg lists, called when transfer is finished.
+ */
+static void atmci_pdc_cleanup(struct atmel_mci *host)
+{
+	struct mmc_data         *data = host->data;
+
+	if (data)
+		dma_unmap_sg(&host->pdev->dev,
+				data->sg, data->sg_len,
+				mmc_get_dma_dir(data));
+}
+
+/*
+ * Disable PDC transfers. Update pending flags to EVENT_XFER_COMPLETE after
+ * having received ATMCI_TXBUFE or ATMCI_RXBUFF interrupt. Enable ATMCI_NOTBUSY
+ * interrupt needed for both transfer directions.
+ */
+static void atmci_pdc_complete(struct atmel_mci *host)
+{
+	int transfer_size = host->data->blocks * host->data->blksz;
+	int i;
+
+	atmci_writel(host, ATMEL_PDC_PTCR, ATMEL_PDC_RXTDIS | ATMEL_PDC_TXTDIS);
+
+	if ((!host->caps.has_rwproof)
+	    && (host->data->flags & MMC_DATA_READ)) {
+		if (host->caps.has_bad_data_ordering)
+			for (i = 0; i < transfer_size; i++)
+				host->buffer[i] = swab32(host->buffer[i]);
+		sg_copy_from_buffer(host->data->sg, host->data->sg_len,
+		                    host->buffer, transfer_size);
+	}
+
+	atmci_pdc_cleanup(host);
+
+	dev_dbg(&host->pdev->dev, "(%s) set pending xfer complete\n", __func__);
+	atmci_set_pending(host, EVENT_XFER_COMPLETE);
+	tasklet_schedule(&host->tasklet);
+}
+
+static void atmci_dma_cleanup(struct atmel_mci *host)
+{
+	struct mmc_data                 *data = host->data;
+
+	if (data)
+		dma_unmap_sg(host->dma.chan->device->dev,
+				data->sg, data->sg_len,
+				mmc_get_dma_dir(data));
+}
+
+/*
+ * This function is called by the DMA driver from tasklet context.
+ */
+static void atmci_dma_complete(void *arg)
+{
+	struct atmel_mci	*host = arg;
+	struct mmc_data		*data = host->data;
+
+	dev_vdbg(&host->pdev->dev, "DMA complete\n");
+
+	if (host->caps.has_dma_conf_reg)
+		/* Disable DMA hardware handshaking on MCI */
+		atmci_writel(host, ATMCI_DMA, atmci_readl(host, ATMCI_DMA) & ~ATMCI_DMAEN);
+
+	atmci_dma_cleanup(host);
+
+	/*
+	 * If the card was removed, data will be NULL. No point trying
+	 * to send the stop command or waiting for NBUSY in this case.
+	 */
+	if (data) {
+		dev_dbg(&host->pdev->dev,
+		        "(%s) set pending xfer complete\n", __func__);
+		atmci_set_pending(host, EVENT_XFER_COMPLETE);
+		tasklet_schedule(&host->tasklet);
+
+		/*
+		 * Regardless of what the documentation says, we have
+		 * to wait for NOTBUSY even after block read
+		 * operations.
+		 *
+		 * When the DMA transfer is complete, the controller
+		 * may still be reading the CRC from the card, i.e.
+		 * the data transfer is still in progress and we
+		 * haven't seen all the potential error bits yet.
+		 *
+		 * The interrupt handler will schedule a different
+		 * tasklet to finish things up when the data transfer
+		 * is completely done.
+		 *
+		 * We may not complete the mmc request here anyway
+		 * because the mmc layer may call back and cause us to
+		 * violate the "don't submit new operations from the
+		 * completion callback" rule of the dma engine
+		 * framework.
+		 */
+		atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
+	}
+}
+
+/*
+ * Returns a mask of interrupt flags to be enabled after the whole
+ * request has been prepared.
+ */
+static u32 atmci_prepare_data(struct atmel_mci *host, struct mmc_data *data)
+{
+	u32 iflags;
+
+	data->error = -EINPROGRESS;
+
+	host->sg = data->sg;
+	host->sg_len = data->sg_len;
+	host->data = data;
+	host->data_chan = NULL;
+
+	iflags = ATMCI_DATA_ERROR_FLAGS;
+
+	/*
+	 * Errata: MMC data write operation with less than 12
+	 * bytes is impossible.
+	 *
+	 * Errata: MCI Transmit Data Register (TDR) FIFO
+	 * corruption when length is not multiple of 4.
+	 */
+	if (data->blocks * data->blksz < 12
+			|| (data->blocks * data->blksz) & 3)
+		host->need_reset = true;
+
+	host->pio_offset = 0;
+	if (data->flags & MMC_DATA_READ)
+		iflags |= ATMCI_RXRDY;
+	else
+		iflags |= ATMCI_TXRDY;
+
+	return iflags;
+}
+
+/*
+ * Set interrupt flags and set block length into the MCI mode register even
+ * if this value is also accessible in the MCI block register. It seems to be
+ * necessary before the High Speed MCI version. It also map sg and configure
+ * PDC registers.
+ */
+static u32
+atmci_prepare_data_pdc(struct atmel_mci *host, struct mmc_data *data)
+{
+	u32 iflags, tmp;
+	int i;
+
+	data->error = -EINPROGRESS;
+
+	host->data = data;
+	host->sg = data->sg;
+	iflags = ATMCI_DATA_ERROR_FLAGS;
+
+	/* Enable pdc mode */
+	atmci_writel(host, ATMCI_MR, host->mode_reg | ATMCI_MR_PDCMODE);
+
+	if (data->flags & MMC_DATA_READ)
+		iflags |= ATMCI_ENDRX | ATMCI_RXBUFF;
+	else
+		iflags |= ATMCI_ENDTX | ATMCI_TXBUFE | ATMCI_BLKE;
+
+	/* Set BLKLEN */
+	tmp = atmci_readl(host, ATMCI_MR);
+	tmp &= 0x0000ffff;
+	tmp |= ATMCI_BLKLEN(data->blksz);
+	atmci_writel(host, ATMCI_MR, tmp);
+
+	/* Configure PDC */
+	host->data_size = data->blocks * data->blksz;
+	dma_map_sg(&host->pdev->dev, data->sg, data->sg_len,
+		   mmc_get_dma_dir(data));
+
+	if ((!host->caps.has_rwproof)
+	    && (host->data->flags & MMC_DATA_WRITE)) {
+		sg_copy_to_buffer(host->data->sg, host->data->sg_len,
+		                  host->buffer, host->data_size);
+		if (host->caps.has_bad_data_ordering)
+			for (i = 0; i < host->data_size; i++)
+				host->buffer[i] = swab32(host->buffer[i]);
+	}
+
+	if (host->data_size)
+		atmci_pdc_set_both_buf(host, data->flags & MMC_DATA_READ ?
+				       XFER_RECEIVE : XFER_TRANSMIT);
+	return iflags;
+}
+
+static u32
+atmci_prepare_data_dma(struct atmel_mci *host, struct mmc_data *data)
+{
+	struct dma_chan			*chan;
+	struct dma_async_tx_descriptor	*desc;
+	struct scatterlist		*sg;
+	unsigned int			i;
+	enum dma_transfer_direction	slave_dirn;
+	unsigned int			sglen;
+	u32				maxburst;
+	u32 iflags;
+
+	data->error = -EINPROGRESS;
+
+	WARN_ON(host->data);
+	host->sg = NULL;
+	host->data = data;
+
+	iflags = ATMCI_DATA_ERROR_FLAGS;
+
+	/*
+	 * We don't do DMA on "complex" transfers, i.e. with
+	 * non-word-aligned buffers or lengths. Also, we don't bother
+	 * with all the DMA setup overhead for short transfers.
+	 */
+	if (data->blocks * data->blksz < ATMCI_DMA_THRESHOLD)
+		return atmci_prepare_data(host, data);
+	if (data->blksz & 3)
+		return atmci_prepare_data(host, data);
+
+	for_each_sg(data->sg, sg, data->sg_len, i) {
+		if (sg->offset & 3 || sg->length & 3)
+			return atmci_prepare_data(host, data);
+	}
+
+	/* If we don't have a channel, we can't do DMA */
+	chan = host->dma.chan;
+	if (chan)
+		host->data_chan = chan;
+
+	if (!chan)
+		return -ENODEV;
+
+	if (data->flags & MMC_DATA_READ) {
+		host->dma_conf.direction = slave_dirn = DMA_DEV_TO_MEM;
+		maxburst = atmci_convert_chksize(host,
+						 host->dma_conf.src_maxburst);
+	} else {
+		host->dma_conf.direction = slave_dirn = DMA_MEM_TO_DEV;
+		maxburst = atmci_convert_chksize(host,
+						 host->dma_conf.dst_maxburst);
+	}
+
+	if (host->caps.has_dma_conf_reg)
+		atmci_writel(host, ATMCI_DMA, ATMCI_DMA_CHKSIZE(maxburst) |
+			ATMCI_DMAEN);
+
+	sglen = dma_map_sg(chan->device->dev, data->sg,
+			data->sg_len, mmc_get_dma_dir(data));
+
+	dmaengine_slave_config(chan, &host->dma_conf);
+	desc = dmaengine_prep_slave_sg(chan,
+			data->sg, sglen, slave_dirn,
+			DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc)
+		goto unmap_exit;
+
+	host->dma.data_desc = desc;
+	desc->callback = atmci_dma_complete;
+	desc->callback_param = host;
+
+	return iflags;
+unmap_exit:
+	dma_unmap_sg(chan->device->dev, data->sg, data->sg_len,
+		     mmc_get_dma_dir(data));
+	return -ENOMEM;
+}
+
+static void
+atmci_submit_data(struct atmel_mci *host, struct mmc_data *data)
+{
+	return;
+}
+
+/*
+ * Start PDC according to transfer direction.
+ */
+static void
+atmci_submit_data_pdc(struct atmel_mci *host, struct mmc_data *data)
+{
+	if (data->flags & MMC_DATA_READ)
+		atmci_writel(host, ATMEL_PDC_PTCR, ATMEL_PDC_RXTEN);
+	else
+		atmci_writel(host, ATMEL_PDC_PTCR, ATMEL_PDC_TXTEN);
+}
+
+static void
+atmci_submit_data_dma(struct atmel_mci *host, struct mmc_data *data)
+{
+	struct dma_chan			*chan = host->data_chan;
+	struct dma_async_tx_descriptor	*desc = host->dma.data_desc;
+
+	if (chan) {
+		dmaengine_submit(desc);
+		dma_async_issue_pending(chan);
+	}
+}
+
+static void atmci_stop_transfer(struct atmel_mci *host)
+{
+	dev_dbg(&host->pdev->dev,
+	        "(%s) set pending xfer complete\n", __func__);
+	atmci_set_pending(host, EVENT_XFER_COMPLETE);
+	atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
+}
+
+/*
+ * Stop data transfer because error(s) occurred.
+ */
+static void atmci_stop_transfer_pdc(struct atmel_mci *host)
+{
+	atmci_writel(host, ATMEL_PDC_PTCR, ATMEL_PDC_RXTDIS | ATMEL_PDC_TXTDIS);
+}
+
+static void atmci_stop_transfer_dma(struct atmel_mci *host)
+{
+	struct dma_chan *chan = host->data_chan;
+
+	if (chan) {
+		dmaengine_terminate_all(chan);
+		atmci_dma_cleanup(host);
+	} else {
+		/* Data transfer was stopped by the interrupt handler */
+		dev_dbg(&host->pdev->dev,
+		        "(%s) set pending xfer complete\n", __func__);
+		atmci_set_pending(host, EVENT_XFER_COMPLETE);
+		atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
+	}
+}
+
+/*
+ * Start a request: prepare data if needed, prepare the command and activate
+ * interrupts.
+ */
+static void atmci_start_request(struct atmel_mci *host,
+		struct atmel_mci_slot *slot)
+{
+	struct mmc_request	*mrq;
+	struct mmc_command	*cmd;
+	struct mmc_data		*data;
+	u32			iflags;
+	u32			cmdflags;
+
+	mrq = slot->mrq;
+	host->cur_slot = slot;
+	host->mrq = mrq;
+
+	host->pending_events = 0;
+	host->completed_events = 0;
+	host->cmd_status = 0;
+	host->data_status = 0;
+
+	dev_dbg(&host->pdev->dev, "start request: cmd %u\n", mrq->cmd->opcode);
+
+	if (host->need_reset || host->caps.need_reset_after_xfer) {
+		iflags = atmci_readl(host, ATMCI_IMR);
+		iflags &= (ATMCI_SDIOIRQA | ATMCI_SDIOIRQB);
+		atmci_writel(host, ATMCI_CR, ATMCI_CR_SWRST);
+		atmci_writel(host, ATMCI_CR, ATMCI_CR_MCIEN);
+		atmci_writel(host, ATMCI_MR, host->mode_reg);
+		if (host->caps.has_cfg_reg)
+			atmci_writel(host, ATMCI_CFG, host->cfg_reg);
+		atmci_writel(host, ATMCI_IER, iflags);
+		host->need_reset = false;
+	}
+	atmci_writel(host, ATMCI_SDCR, slot->sdc_reg);
+
+	iflags = atmci_readl(host, ATMCI_IMR);
+	if (iflags & ~(ATMCI_SDIOIRQA | ATMCI_SDIOIRQB))
+		dev_dbg(&slot->mmc->class_dev, "WARNING: IMR=0x%08x\n",
+				iflags);
+
+	if (unlikely(test_and_clear_bit(ATMCI_CARD_NEED_INIT, &slot->flags))) {
+		/* Send init sequence (74 clock cycles) */
+		atmci_writel(host, ATMCI_CMDR, ATMCI_CMDR_SPCMD_INIT);
+		while (!(atmci_readl(host, ATMCI_SR) & ATMCI_CMDRDY))
+			cpu_relax();
+	}
+	iflags = 0;
+	data = mrq->data;
+	if (data) {
+		atmci_set_timeout(host, slot, data);
+
+		/* Must set block count/size before sending command */
+		atmci_writel(host, ATMCI_BLKR, ATMCI_BCNT(data->blocks)
+				| ATMCI_BLKLEN(data->blksz));
+		dev_vdbg(&slot->mmc->class_dev, "BLKR=0x%08x\n",
+			ATMCI_BCNT(data->blocks) | ATMCI_BLKLEN(data->blksz));
+
+		iflags |= host->prepare_data(host, data);
+	}
+
+	iflags |= ATMCI_CMDRDY;
+	cmd = mrq->cmd;
+	cmdflags = atmci_prepare_command(slot->mmc, cmd);
+
+	/*
+	 * DMA transfer should be started before sending the command to avoid
+	 * unexpected errors especially for read operations in SDIO mode.
+	 * Unfortunately, in PDC mode, command has to be sent before starting
+	 * the transfer.
+	 */
+	if (host->submit_data != &atmci_submit_data_dma)
+		atmci_send_command(host, cmd, cmdflags);
+
+	if (data)
+		host->submit_data(host, data);
+
+	if (host->submit_data == &atmci_submit_data_dma)
+		atmci_send_command(host, cmd, cmdflags);
+
+	if (mrq->stop) {
+		host->stop_cmdr = atmci_prepare_command(slot->mmc, mrq->stop);
+		host->stop_cmdr |= ATMCI_CMDR_STOP_XFER;
+		if (!(data->flags & MMC_DATA_WRITE))
+			host->stop_cmdr |= ATMCI_CMDR_TRDIR_READ;
+		host->stop_cmdr |= ATMCI_CMDR_MULTI_BLOCK;
+	}
+
+	/*
+	 * We could have enabled interrupts earlier, but I suspect
+	 * that would open up a nice can of interesting race
+	 * conditions (e.g. command and data complete, but stop not
+	 * prepared yet.)
+	 */
+	atmci_writel(host, ATMCI_IER, iflags);
+}
+
+static void atmci_queue_request(struct atmel_mci *host,
+		struct atmel_mci_slot *slot, struct mmc_request *mrq)
+{
+	dev_vdbg(&slot->mmc->class_dev, "queue request: state=%d\n",
+			host->state);
+
+	spin_lock_bh(&host->lock);
+	slot->mrq = mrq;
+	if (host->state == STATE_IDLE) {
+		host->state = STATE_SENDING_CMD;
+		atmci_start_request(host, slot);
+	} else {
+		dev_dbg(&host->pdev->dev, "queue request\n");
+		list_add_tail(&slot->queue_node, &host->queue);
+	}
+	spin_unlock_bh(&host->lock);
+}
+
+static void atmci_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct atmel_mci_slot	*slot = mmc_priv(mmc);
+	struct atmel_mci	*host = slot->host;
+	struct mmc_data		*data;
+
+	WARN_ON(slot->mrq);
+	dev_dbg(&host->pdev->dev, "MRQ: cmd %u\n", mrq->cmd->opcode);
+
+	/*
+	 * We may "know" the card is gone even though there's still an
+	 * electrical connection. If so, we really need to communicate
+	 * this to the MMC core since there won't be any more
+	 * interrupts as the card is completely removed. Otherwise,
+	 * the MMC core might believe the card is still there even
+	 * though the card was just removed very slowly.
+	 */
+	if (!test_bit(ATMCI_CARD_PRESENT, &slot->flags)) {
+		mrq->cmd->error = -ENOMEDIUM;
+		mmc_request_done(mmc, mrq);
+		return;
+	}
+
+	/* We don't support multiple blocks of weird lengths. */
+	data = mrq->data;
+	if (data && data->blocks > 1 && data->blksz & 3) {
+		mrq->cmd->error = -EINVAL;
+		mmc_request_done(mmc, mrq);
+	}
+
+	atmci_queue_request(host, slot, mrq);
+}
+
+static void atmci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct atmel_mci_slot	*slot = mmc_priv(mmc);
+	struct atmel_mci	*host = slot->host;
+	unsigned int		i;
+
+	slot->sdc_reg &= ~ATMCI_SDCBUS_MASK;
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_1:
+		slot->sdc_reg |= ATMCI_SDCBUS_1BIT;
+		break;
+	case MMC_BUS_WIDTH_4:
+		slot->sdc_reg |= ATMCI_SDCBUS_4BIT;
+		break;
+	case MMC_BUS_WIDTH_8:
+		slot->sdc_reg |= ATMCI_SDCBUS_8BIT;
+		break;
+	}
+
+	if (ios->clock) {
+		unsigned int clock_min = ~0U;
+		int clkdiv;
+
+		spin_lock_bh(&host->lock);
+		if (!host->mode_reg) {
+			atmci_writel(host, ATMCI_CR, ATMCI_CR_SWRST);
+			atmci_writel(host, ATMCI_CR, ATMCI_CR_MCIEN);
+			if (host->caps.has_cfg_reg)
+				atmci_writel(host, ATMCI_CFG, host->cfg_reg);
+		}
+
+		/*
+		 * Use mirror of ios->clock to prevent race with mmc
+		 * core ios update when finding the minimum.
+		 */
+		slot->clock = ios->clock;
+		for (i = 0; i < ATMCI_MAX_NR_SLOTS; i++) {
+			if (host->slot[i] && host->slot[i]->clock
+					&& host->slot[i]->clock < clock_min)
+				clock_min = host->slot[i]->clock;
+		}
+
+		/* Calculate clock divider */
+		if (host->caps.has_odd_clk_div) {
+			clkdiv = DIV_ROUND_UP(host->bus_hz, clock_min) - 2;
+			if (clkdiv < 0) {
+				dev_warn(&mmc->class_dev,
+					 "clock %u too fast; using %lu\n",
+					 clock_min, host->bus_hz / 2);
+				clkdiv = 0;
+			} else if (clkdiv > 511) {
+				dev_warn(&mmc->class_dev,
+				         "clock %u too slow; using %lu\n",
+				         clock_min, host->bus_hz / (511 + 2));
+				clkdiv = 511;
+			}
+			host->mode_reg = ATMCI_MR_CLKDIV(clkdiv >> 1)
+			                 | ATMCI_MR_CLKODD(clkdiv & 1);
+		} else {
+			clkdiv = DIV_ROUND_UP(host->bus_hz, 2 * clock_min) - 1;
+			if (clkdiv > 255) {
+				dev_warn(&mmc->class_dev,
+				         "clock %u too slow; using %lu\n",
+				         clock_min, host->bus_hz / (2 * 256));
+				clkdiv = 255;
+			}
+			host->mode_reg = ATMCI_MR_CLKDIV(clkdiv);
+		}
+
+		/*
+		 * WRPROOF and RDPROOF prevent overruns/underruns by
+		 * stopping the clock when the FIFO is full/empty.
+		 * This state is not expected to last for long.
+		 */
+		if (host->caps.has_rwproof)
+			host->mode_reg |= (ATMCI_MR_WRPROOF | ATMCI_MR_RDPROOF);
+
+		if (host->caps.has_cfg_reg) {
+			/* setup High Speed mode in relation with card capacity */
+			if (ios->timing == MMC_TIMING_SD_HS)
+				host->cfg_reg |= ATMCI_CFG_HSMODE;
+			else
+				host->cfg_reg &= ~ATMCI_CFG_HSMODE;
+		}
+
+		if (list_empty(&host->queue)) {
+			atmci_writel(host, ATMCI_MR, host->mode_reg);
+			if (host->caps.has_cfg_reg)
+				atmci_writel(host, ATMCI_CFG, host->cfg_reg);
+		} else {
+			host->need_clock_update = true;
+		}
+
+		spin_unlock_bh(&host->lock);
+	} else {
+		bool any_slot_active = false;
+
+		spin_lock_bh(&host->lock);
+		slot->clock = 0;
+		for (i = 0; i < ATMCI_MAX_NR_SLOTS; i++) {
+			if (host->slot[i] && host->slot[i]->clock) {
+				any_slot_active = true;
+				break;
+			}
+		}
+		if (!any_slot_active) {
+			atmci_writel(host, ATMCI_CR, ATMCI_CR_MCIDIS);
+			if (host->mode_reg) {
+				atmci_readl(host, ATMCI_MR);
+			}
+			host->mode_reg = 0;
+		}
+		spin_unlock_bh(&host->lock);
+	}
+
+	switch (ios->power_mode) {
+	case MMC_POWER_OFF:
+		if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+		break;
+	case MMC_POWER_UP:
+		set_bit(ATMCI_CARD_NEED_INIT, &slot->flags);
+		if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, ios->vdd);
+		break;
+	default:
+		break;
+	}
+}
+
+static int atmci_get_ro(struct mmc_host *mmc)
+{
+	int			read_only = -ENOSYS;
+	struct atmel_mci_slot	*slot = mmc_priv(mmc);
+
+	if (gpio_is_valid(slot->wp_pin)) {
+		read_only = gpio_get_value(slot->wp_pin);
+		dev_dbg(&mmc->class_dev, "card is %s\n",
+				read_only ? "read-only" : "read-write");
+	}
+
+	return read_only;
+}
+
+static int atmci_get_cd(struct mmc_host *mmc)
+{
+	int			present = -ENOSYS;
+	struct atmel_mci_slot	*slot = mmc_priv(mmc);
+
+	if (gpio_is_valid(slot->detect_pin)) {
+		present = !(gpio_get_value(slot->detect_pin) ^
+			    slot->detect_is_active_high);
+		dev_dbg(&mmc->class_dev, "card is %spresent\n",
+				present ? "" : "not ");
+	}
+
+	return present;
+}
+
+static void atmci_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct atmel_mci_slot	*slot = mmc_priv(mmc);
+	struct atmel_mci	*host = slot->host;
+
+	if (enable)
+		atmci_writel(host, ATMCI_IER, slot->sdio_irq);
+	else
+		atmci_writel(host, ATMCI_IDR, slot->sdio_irq);
+}
+
+static const struct mmc_host_ops atmci_ops = {
+	.request	= atmci_request,
+	.set_ios	= atmci_set_ios,
+	.get_ro		= atmci_get_ro,
+	.get_cd		= atmci_get_cd,
+	.enable_sdio_irq = atmci_enable_sdio_irq,
+};
+
+/* Called with host->lock held */
+static void atmci_request_end(struct atmel_mci *host, struct mmc_request *mrq)
+	__releases(&host->lock)
+	__acquires(&host->lock)
+{
+	struct atmel_mci_slot	*slot = NULL;
+	struct mmc_host		*prev_mmc = host->cur_slot->mmc;
+
+	WARN_ON(host->cmd || host->data);
+
+	del_timer(&host->timer);
+
+	/*
+	 * Update the MMC clock rate if necessary. This may be
+	 * necessary if set_ios() is called when a different slot is
+	 * busy transferring data.
+	 */
+	if (host->need_clock_update) {
+		atmci_writel(host, ATMCI_MR, host->mode_reg);
+		if (host->caps.has_cfg_reg)
+			atmci_writel(host, ATMCI_CFG, host->cfg_reg);
+	}
+
+	host->cur_slot->mrq = NULL;
+	host->mrq = NULL;
+	if (!list_empty(&host->queue)) {
+		slot = list_entry(host->queue.next,
+				struct atmel_mci_slot, queue_node);
+		list_del(&slot->queue_node);
+		dev_vdbg(&host->pdev->dev, "list not empty: %s is next\n",
+				mmc_hostname(slot->mmc));
+		host->state = STATE_SENDING_CMD;
+		atmci_start_request(host, slot);
+	} else {
+		dev_vdbg(&host->pdev->dev, "list empty\n");
+		host->state = STATE_IDLE;
+	}
+
+	spin_unlock(&host->lock);
+	mmc_request_done(prev_mmc, mrq);
+	spin_lock(&host->lock);
+}
+
+static void atmci_command_complete(struct atmel_mci *host,
+			struct mmc_command *cmd)
+{
+	u32		status = host->cmd_status;
+
+	/* Read the response from the card (up to 16 bytes) */
+	cmd->resp[0] = atmci_readl(host, ATMCI_RSPR);
+	cmd->resp[1] = atmci_readl(host, ATMCI_RSPR);
+	cmd->resp[2] = atmci_readl(host, ATMCI_RSPR);
+	cmd->resp[3] = atmci_readl(host, ATMCI_RSPR);
+
+	if (status & ATMCI_RTOE)
+		cmd->error = -ETIMEDOUT;
+	else if ((cmd->flags & MMC_RSP_CRC) && (status & ATMCI_RCRCE))
+		cmd->error = -EILSEQ;
+	else if (status & (ATMCI_RINDE | ATMCI_RDIRE | ATMCI_RENDE))
+		cmd->error = -EIO;
+	else if (host->mrq->data && (host->mrq->data->blksz & 3)) {
+		if (host->caps.need_blksz_mul_4) {
+			cmd->error = -EINVAL;
+			host->need_reset = 1;
+		}
+	} else
+		cmd->error = 0;
+}
+
+static void atmci_detect_change(struct timer_list *t)
+{
+	struct atmel_mci_slot	*slot = from_timer(slot, t, detect_timer);
+	bool			present;
+	bool			present_old;
+
+	/*
+	 * atmci_cleanup_slot() sets the ATMCI_SHUTDOWN flag before
+	 * freeing the interrupt. We must not re-enable the interrupt
+	 * if it has been freed, and if we're shutting down, it
+	 * doesn't really matter whether the card is present or not.
+	 */
+	smp_rmb();
+	if (test_bit(ATMCI_SHUTDOWN, &slot->flags))
+		return;
+
+	enable_irq(gpio_to_irq(slot->detect_pin));
+	present = !(gpio_get_value(slot->detect_pin) ^
+		    slot->detect_is_active_high);
+	present_old = test_bit(ATMCI_CARD_PRESENT, &slot->flags);
+
+	dev_vdbg(&slot->mmc->class_dev, "detect change: %d (was %d)\n",
+			present, present_old);
+
+	if (present != present_old) {
+		struct atmel_mci	*host = slot->host;
+		struct mmc_request	*mrq;
+
+		dev_dbg(&slot->mmc->class_dev, "card %s\n",
+			present ? "inserted" : "removed");
+
+		spin_lock(&host->lock);
+
+		if (!present)
+			clear_bit(ATMCI_CARD_PRESENT, &slot->flags);
+		else
+			set_bit(ATMCI_CARD_PRESENT, &slot->flags);
+
+		/* Clean up queue if present */
+		mrq = slot->mrq;
+		if (mrq) {
+			if (mrq == host->mrq) {
+				/*
+				 * Reset controller to terminate any ongoing
+				 * commands or data transfers.
+				 */
+				atmci_writel(host, ATMCI_CR, ATMCI_CR_SWRST);
+				atmci_writel(host, ATMCI_CR, ATMCI_CR_MCIEN);
+				atmci_writel(host, ATMCI_MR, host->mode_reg);
+				if (host->caps.has_cfg_reg)
+					atmci_writel(host, ATMCI_CFG, host->cfg_reg);
+
+				host->data = NULL;
+				host->cmd = NULL;
+
+				switch (host->state) {
+				case STATE_IDLE:
+					break;
+				case STATE_SENDING_CMD:
+					mrq->cmd->error = -ENOMEDIUM;
+					if (mrq->data)
+						host->stop_transfer(host);
+					break;
+				case STATE_DATA_XFER:
+					mrq->data->error = -ENOMEDIUM;
+					host->stop_transfer(host);
+					break;
+				case STATE_WAITING_NOTBUSY:
+					mrq->data->error = -ENOMEDIUM;
+					break;
+				case STATE_SENDING_STOP:
+					mrq->stop->error = -ENOMEDIUM;
+					break;
+				case STATE_END_REQUEST:
+					break;
+				}
+
+				atmci_request_end(host, mrq);
+			} else {
+				list_del(&slot->queue_node);
+				mrq->cmd->error = -ENOMEDIUM;
+				if (mrq->data)
+					mrq->data->error = -ENOMEDIUM;
+				if (mrq->stop)
+					mrq->stop->error = -ENOMEDIUM;
+
+				spin_unlock(&host->lock);
+				mmc_request_done(slot->mmc, mrq);
+				spin_lock(&host->lock);
+			}
+		}
+		spin_unlock(&host->lock);
+
+		mmc_detect_change(slot->mmc, 0);
+	}
+}
+
+static void atmci_tasklet_func(unsigned long priv)
+{
+	struct atmel_mci	*host = (struct atmel_mci *)priv;
+	struct mmc_request	*mrq = host->mrq;
+	struct mmc_data		*data = host->data;
+	enum atmel_mci_state	state = host->state;
+	enum atmel_mci_state	prev_state;
+	u32			status;
+
+	spin_lock(&host->lock);
+
+	state = host->state;
+
+	dev_vdbg(&host->pdev->dev,
+		"tasklet: state %u pending/completed/mask %lx/%lx/%x\n",
+		state, host->pending_events, host->completed_events,
+		atmci_readl(host, ATMCI_IMR));
+
+	do {
+		prev_state = state;
+		dev_dbg(&host->pdev->dev, "FSM: state=%d\n", state);
+
+		switch (state) {
+		case STATE_IDLE:
+			break;
+
+		case STATE_SENDING_CMD:
+			/*
+			 * Command has been sent, we are waiting for command
+			 * ready. Then we have three next states possible:
+			 * END_REQUEST by default, WAITING_NOTBUSY if it's a
+			 * command needing it or DATA_XFER if there is data.
+			 */
+			dev_dbg(&host->pdev->dev, "FSM: cmd ready?\n");
+			if (!atmci_test_and_clear_pending(host,
+						EVENT_CMD_RDY))
+				break;
+
+			dev_dbg(&host->pdev->dev, "set completed cmd ready\n");
+			host->cmd = NULL;
+			atmci_set_completed(host, EVENT_CMD_RDY);
+			atmci_command_complete(host, mrq->cmd);
+			if (mrq->data) {
+				dev_dbg(&host->pdev->dev,
+				        "command with data transfer");
+				/*
+				 * If there is a command error don't start
+				 * data transfer.
+				 */
+				if (mrq->cmd->error) {
+					host->stop_transfer(host);
+					host->data = NULL;
+					atmci_writel(host, ATMCI_IDR,
+					             ATMCI_TXRDY | ATMCI_RXRDY
+					             | ATMCI_DATA_ERROR_FLAGS);
+					state = STATE_END_REQUEST;
+				} else
+					state = STATE_DATA_XFER;
+			} else if ((!mrq->data) && (mrq->cmd->flags & MMC_RSP_BUSY)) {
+				dev_dbg(&host->pdev->dev,
+				        "command response need waiting notbusy");
+				atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
+				state = STATE_WAITING_NOTBUSY;
+			} else
+				state = STATE_END_REQUEST;
+
+			break;
+
+		case STATE_DATA_XFER:
+			if (atmci_test_and_clear_pending(host,
+						EVENT_DATA_ERROR)) {
+				dev_dbg(&host->pdev->dev, "set completed data error\n");
+				atmci_set_completed(host, EVENT_DATA_ERROR);
+				state = STATE_END_REQUEST;
+				break;
+			}
+
+			/*
+			 * A data transfer is in progress. The event expected
+			 * to move to the next state depends of data transfer
+			 * type (PDC or DMA). Once transfer done we can move
+			 * to the next step which is WAITING_NOTBUSY in write
+			 * case and directly SENDING_STOP in read case.
+			 */
+			dev_dbg(&host->pdev->dev, "FSM: xfer complete?\n");
+			if (!atmci_test_and_clear_pending(host,
+						EVENT_XFER_COMPLETE))
+				break;
+
+			dev_dbg(&host->pdev->dev,
+			        "(%s) set completed xfer complete\n",
+				__func__);
+			atmci_set_completed(host, EVENT_XFER_COMPLETE);
+
+			if (host->caps.need_notbusy_for_read_ops ||
+			   (host->data->flags & MMC_DATA_WRITE)) {
+				atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
+				state = STATE_WAITING_NOTBUSY;
+			} else if (host->mrq->stop) {
+				atmci_send_stop_cmd(host, data);
+				state = STATE_SENDING_STOP;
+			} else {
+				host->data = NULL;
+				data->bytes_xfered = data->blocks * data->blksz;
+				data->error = 0;
+				state = STATE_END_REQUEST;
+			}
+			break;
+
+		case STATE_WAITING_NOTBUSY:
+			/*
+			 * We can be in the state for two reasons: a command
+			 * requiring waiting not busy signal (stop command
+			 * included) or a write operation. In the latest case,
+			 * we need to send a stop command.
+			 */
+			dev_dbg(&host->pdev->dev, "FSM: not busy?\n");
+			if (!atmci_test_and_clear_pending(host,
+						EVENT_NOTBUSY))
+				break;
+
+			dev_dbg(&host->pdev->dev, "set completed not busy\n");
+			atmci_set_completed(host, EVENT_NOTBUSY);
+
+			if (host->data) {
+				/*
+				 * For some commands such as CMD53, even if
+				 * there is data transfer, there is no stop
+				 * command to send.
+				 */
+				if (host->mrq->stop) {
+					atmci_send_stop_cmd(host, data);
+					state = STATE_SENDING_STOP;
+				} else {
+					host->data = NULL;
+					data->bytes_xfered = data->blocks
+					                     * data->blksz;
+					data->error = 0;
+					state = STATE_END_REQUEST;
+				}
+			} else
+				state = STATE_END_REQUEST;
+			break;
+
+		case STATE_SENDING_STOP:
+			/*
+			 * In this state, it is important to set host->data to
+			 * NULL (which is tested in the waiting notbusy state)
+			 * in order to go to the end request state instead of
+			 * sending stop again.
+			 */
+			dev_dbg(&host->pdev->dev, "FSM: cmd ready?\n");
+			if (!atmci_test_and_clear_pending(host,
+						EVENT_CMD_RDY))
+				break;
+
+			dev_dbg(&host->pdev->dev, "FSM: cmd ready\n");
+			host->cmd = NULL;
+			data->bytes_xfered = data->blocks * data->blksz;
+			data->error = 0;
+			atmci_command_complete(host, mrq->stop);
+			if (mrq->stop->error) {
+				host->stop_transfer(host);
+				atmci_writel(host, ATMCI_IDR,
+				             ATMCI_TXRDY | ATMCI_RXRDY
+				             | ATMCI_DATA_ERROR_FLAGS);
+				state = STATE_END_REQUEST;
+			} else {
+				atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
+				state = STATE_WAITING_NOTBUSY;
+			}
+			host->data = NULL;
+			break;
+
+		case STATE_END_REQUEST:
+			atmci_writel(host, ATMCI_IDR, ATMCI_TXRDY | ATMCI_RXRDY
+			                   | ATMCI_DATA_ERROR_FLAGS);
+			status = host->data_status;
+			if (unlikely(status)) {
+				host->stop_transfer(host);
+				host->data = NULL;
+				if (data) {
+					if (status & ATMCI_DTOE) {
+						data->error = -ETIMEDOUT;
+					} else if (status & ATMCI_DCRCE) {
+						data->error = -EILSEQ;
+					} else {
+						data->error = -EIO;
+					}
+				}
+			}
+
+			atmci_request_end(host, host->mrq);
+			goto unlock; /* atmci_request_end() sets host->state */
+			break;
+		}
+	} while (state != prev_state);
+
+	host->state = state;
+
+unlock:
+	spin_unlock(&host->lock);
+}
+
+static void atmci_read_data_pio(struct atmel_mci *host)
+{
+	struct scatterlist	*sg = host->sg;
+	unsigned int		offset = host->pio_offset;
+	struct mmc_data		*data = host->data;
+	u32			value;
+	u32			status;
+	unsigned int		nbytes = 0;
+
+	do {
+		value = atmci_readl(host, ATMCI_RDR);
+		if (likely(offset + 4 <= sg->length)) {
+			sg_pcopy_from_buffer(sg, 1, &value, sizeof(u32), offset);
+
+			offset += 4;
+			nbytes += 4;
+
+			if (offset == sg->length) {
+				flush_dcache_page(sg_page(sg));
+				host->sg = sg = sg_next(sg);
+				host->sg_len--;
+				if (!sg || !host->sg_len)
+					goto done;
+
+				offset = 0;
+			}
+		} else {
+			unsigned int remaining = sg->length - offset;
+
+			sg_pcopy_from_buffer(sg, 1, &value, remaining, offset);
+			nbytes += remaining;
+
+			flush_dcache_page(sg_page(sg));
+			host->sg = sg = sg_next(sg);
+			host->sg_len--;
+			if (!sg || !host->sg_len)
+				goto done;
+
+			offset = 4 - remaining;
+			sg_pcopy_from_buffer(sg, 1, (u8 *)&value + remaining,
+					offset, 0);
+			nbytes += offset;
+		}
+
+		status = atmci_readl(host, ATMCI_SR);
+		if (status & ATMCI_DATA_ERROR_FLAGS) {
+			atmci_writel(host, ATMCI_IDR, (ATMCI_NOTBUSY | ATMCI_RXRDY
+						| ATMCI_DATA_ERROR_FLAGS));
+			host->data_status = status;
+			data->bytes_xfered += nbytes;
+			return;
+		}
+	} while (status & ATMCI_RXRDY);
+
+	host->pio_offset = offset;
+	data->bytes_xfered += nbytes;
+
+	return;
+
+done:
+	atmci_writel(host, ATMCI_IDR, ATMCI_RXRDY);
+	atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
+	data->bytes_xfered += nbytes;
+	smp_wmb();
+	atmci_set_pending(host, EVENT_XFER_COMPLETE);
+}
+
+static void atmci_write_data_pio(struct atmel_mci *host)
+{
+	struct scatterlist	*sg = host->sg;
+	unsigned int		offset = host->pio_offset;
+	struct mmc_data		*data = host->data;
+	u32			value;
+	u32			status;
+	unsigned int		nbytes = 0;
+
+	do {
+		if (likely(offset + 4 <= sg->length)) {
+			sg_pcopy_to_buffer(sg, 1, &value, sizeof(u32), offset);
+			atmci_writel(host, ATMCI_TDR, value);
+
+			offset += 4;
+			nbytes += 4;
+			if (offset == sg->length) {
+				host->sg = sg = sg_next(sg);
+				host->sg_len--;
+				if (!sg || !host->sg_len)
+					goto done;
+
+				offset = 0;
+			}
+		} else {
+			unsigned int remaining = sg->length - offset;
+
+			value = 0;
+			sg_pcopy_to_buffer(sg, 1, &value, remaining, offset);
+			nbytes += remaining;
+
+			host->sg = sg = sg_next(sg);
+			host->sg_len--;
+			if (!sg || !host->sg_len) {
+				atmci_writel(host, ATMCI_TDR, value);
+				goto done;
+			}
+
+			offset = 4 - remaining;
+			sg_pcopy_to_buffer(sg, 1, (u8 *)&value + remaining,
+					offset, 0);
+			atmci_writel(host, ATMCI_TDR, value);
+			nbytes += offset;
+		}
+
+		status = atmci_readl(host, ATMCI_SR);
+		if (status & ATMCI_DATA_ERROR_FLAGS) {
+			atmci_writel(host, ATMCI_IDR, (ATMCI_NOTBUSY | ATMCI_TXRDY
+						| ATMCI_DATA_ERROR_FLAGS));
+			host->data_status = status;
+			data->bytes_xfered += nbytes;
+			return;
+		}
+	} while (status & ATMCI_TXRDY);
+
+	host->pio_offset = offset;
+	data->bytes_xfered += nbytes;
+
+	return;
+
+done:
+	atmci_writel(host, ATMCI_IDR, ATMCI_TXRDY);
+	atmci_writel(host, ATMCI_IER, ATMCI_NOTBUSY);
+	data->bytes_xfered += nbytes;
+	smp_wmb();
+	atmci_set_pending(host, EVENT_XFER_COMPLETE);
+}
+
+static void atmci_sdio_interrupt(struct atmel_mci *host, u32 status)
+{
+	int	i;
+
+	for (i = 0; i < ATMCI_MAX_NR_SLOTS; i++) {
+		struct atmel_mci_slot *slot = host->slot[i];
+		if (slot && (status & slot->sdio_irq)) {
+			mmc_signal_sdio_irq(slot->mmc);
+		}
+	}
+}
+
+
+static irqreturn_t atmci_interrupt(int irq, void *dev_id)
+{
+	struct atmel_mci	*host = dev_id;
+	u32			status, mask, pending;
+	unsigned int		pass_count = 0;
+
+	do {
+		status = atmci_readl(host, ATMCI_SR);
+		mask = atmci_readl(host, ATMCI_IMR);
+		pending = status & mask;
+		if (!pending)
+			break;
+
+		if (pending & ATMCI_DATA_ERROR_FLAGS) {
+			dev_dbg(&host->pdev->dev, "IRQ: data error\n");
+			atmci_writel(host, ATMCI_IDR, ATMCI_DATA_ERROR_FLAGS
+					| ATMCI_RXRDY | ATMCI_TXRDY
+					| ATMCI_ENDRX | ATMCI_ENDTX
+					| ATMCI_RXBUFF | ATMCI_TXBUFE);
+
+			host->data_status = status;
+			dev_dbg(&host->pdev->dev, "set pending data error\n");
+			smp_wmb();
+			atmci_set_pending(host, EVENT_DATA_ERROR);
+			tasklet_schedule(&host->tasklet);
+		}
+
+		if (pending & ATMCI_TXBUFE) {
+			dev_dbg(&host->pdev->dev, "IRQ: tx buffer empty\n");
+			atmci_writel(host, ATMCI_IDR, ATMCI_TXBUFE);
+			atmci_writel(host, ATMCI_IDR, ATMCI_ENDTX);
+			/*
+			 * We can receive this interruption before having configured
+			 * the second pdc buffer, so we need to reconfigure first and
+			 * second buffers again
+			 */
+			if (host->data_size) {
+				atmci_pdc_set_both_buf(host, XFER_TRANSMIT);
+				atmci_writel(host, ATMCI_IER, ATMCI_ENDTX);
+				atmci_writel(host, ATMCI_IER, ATMCI_TXBUFE);
+			} else {
+				atmci_pdc_complete(host);
+			}
+		} else if (pending & ATMCI_ENDTX) {
+			dev_dbg(&host->pdev->dev, "IRQ: end of tx buffer\n");
+			atmci_writel(host, ATMCI_IDR, ATMCI_ENDTX);
+
+			if (host->data_size) {
+				atmci_pdc_set_single_buf(host,
+						XFER_TRANSMIT, PDC_SECOND_BUF);
+				atmci_writel(host, ATMCI_IER, ATMCI_ENDTX);
+			}
+		}
+
+		if (pending & ATMCI_RXBUFF) {
+			dev_dbg(&host->pdev->dev, "IRQ: rx buffer full\n");
+			atmci_writel(host, ATMCI_IDR, ATMCI_RXBUFF);
+			atmci_writel(host, ATMCI_IDR, ATMCI_ENDRX);
+			/*
+			 * We can receive this interruption before having configured
+			 * the second pdc buffer, so we need to reconfigure first and
+			 * second buffers again
+			 */
+			if (host->data_size) {
+				atmci_pdc_set_both_buf(host, XFER_RECEIVE);
+				atmci_writel(host, ATMCI_IER, ATMCI_ENDRX);
+				atmci_writel(host, ATMCI_IER, ATMCI_RXBUFF);
+			} else {
+				atmci_pdc_complete(host);
+			}
+		} else if (pending & ATMCI_ENDRX) {
+			dev_dbg(&host->pdev->dev, "IRQ: end of rx buffer\n");
+			atmci_writel(host, ATMCI_IDR, ATMCI_ENDRX);
+
+			if (host->data_size) {
+				atmci_pdc_set_single_buf(host,
+						XFER_RECEIVE, PDC_SECOND_BUF);
+				atmci_writel(host, ATMCI_IER, ATMCI_ENDRX);
+			}
+		}
+
+		/*
+		 * First mci IPs, so mainly the ones having pdc, have some
+		 * issues with the notbusy signal. You can't get it after
+		 * data transmission if you have not sent a stop command.
+		 * The appropriate workaround is to use the BLKE signal.
+		 */
+		if (pending & ATMCI_BLKE) {
+			dev_dbg(&host->pdev->dev, "IRQ: blke\n");
+			atmci_writel(host, ATMCI_IDR, ATMCI_BLKE);
+			smp_wmb();
+			dev_dbg(&host->pdev->dev, "set pending notbusy\n");
+			atmci_set_pending(host, EVENT_NOTBUSY);
+			tasklet_schedule(&host->tasklet);
+		}
+
+		if (pending & ATMCI_NOTBUSY) {
+			dev_dbg(&host->pdev->dev, "IRQ: not_busy\n");
+			atmci_writel(host, ATMCI_IDR, ATMCI_NOTBUSY);
+			smp_wmb();
+			dev_dbg(&host->pdev->dev, "set pending notbusy\n");
+			atmci_set_pending(host, EVENT_NOTBUSY);
+			tasklet_schedule(&host->tasklet);
+		}
+
+		if (pending & ATMCI_RXRDY)
+			atmci_read_data_pio(host);
+		if (pending & ATMCI_TXRDY)
+			atmci_write_data_pio(host);
+
+		if (pending & ATMCI_CMDRDY) {
+			dev_dbg(&host->pdev->dev, "IRQ: cmd ready\n");
+			atmci_writel(host, ATMCI_IDR, ATMCI_CMDRDY);
+			host->cmd_status = status;
+			smp_wmb();
+			dev_dbg(&host->pdev->dev, "set pending cmd rdy\n");
+			atmci_set_pending(host, EVENT_CMD_RDY);
+			tasklet_schedule(&host->tasklet);
+		}
+
+		if (pending & (ATMCI_SDIOIRQA | ATMCI_SDIOIRQB))
+			atmci_sdio_interrupt(host, status);
+
+	} while (pass_count++ < 5);
+
+	return pass_count ? IRQ_HANDLED : IRQ_NONE;
+}
+
+static irqreturn_t atmci_detect_interrupt(int irq, void *dev_id)
+{
+	struct atmel_mci_slot	*slot = dev_id;
+
+	/*
+	 * Disable interrupts until the pin has stabilized and check
+	 * the state then. Use mod_timer() since we may be in the
+	 * middle of the timer routine when this interrupt triggers.
+	 */
+	disable_irq_nosync(irq);
+	mod_timer(&slot->detect_timer, jiffies + msecs_to_jiffies(20));
+
+	return IRQ_HANDLED;
+}
+
+static int atmci_init_slot(struct atmel_mci *host,
+		struct mci_slot_pdata *slot_data, unsigned int id,
+		u32 sdc_reg, u32 sdio_irq)
+{
+	struct mmc_host			*mmc;
+	struct atmel_mci_slot		*slot;
+	int ret;
+
+	mmc = mmc_alloc_host(sizeof(struct atmel_mci_slot), &host->pdev->dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	slot = mmc_priv(mmc);
+	slot->mmc = mmc;
+	slot->host = host;
+	slot->detect_pin = slot_data->detect_pin;
+	slot->wp_pin = slot_data->wp_pin;
+	slot->detect_is_active_high = slot_data->detect_is_active_high;
+	slot->sdc_reg = sdc_reg;
+	slot->sdio_irq = sdio_irq;
+
+	dev_dbg(&mmc->class_dev,
+	        "slot[%u]: bus_width=%u, detect_pin=%d, "
+		"detect_is_active_high=%s, wp_pin=%d\n",
+		id, slot_data->bus_width, slot_data->detect_pin,
+		slot_data->detect_is_active_high ? "true" : "false",
+		slot_data->wp_pin);
+
+	mmc->ops = &atmci_ops;
+	mmc->f_min = DIV_ROUND_UP(host->bus_hz, 512);
+	mmc->f_max = host->bus_hz / 2;
+	mmc->ocr_avail	= MMC_VDD_32_33 | MMC_VDD_33_34;
+	if (sdio_irq)
+		mmc->caps |= MMC_CAP_SDIO_IRQ;
+	if (host->caps.has_highspeed)
+		mmc->caps |= MMC_CAP_SD_HIGHSPEED;
+	/*
+	 * Without the read/write proof capability, it is strongly suggested to
+	 * use only one bit for data to prevent fifo underruns and overruns
+	 * which will corrupt data.
+	 */
+	if ((slot_data->bus_width >= 4) && host->caps.has_rwproof) {
+		mmc->caps |= MMC_CAP_4_BIT_DATA;
+		if (slot_data->bus_width >= 8)
+			mmc->caps |= MMC_CAP_8_BIT_DATA;
+	}
+
+	if (atmci_get_version(host) < 0x200) {
+		mmc->max_segs = 256;
+		mmc->max_blk_size = 4095;
+		mmc->max_blk_count = 256;
+		mmc->max_req_size = mmc->max_blk_size * mmc->max_blk_count;
+		mmc->max_seg_size = mmc->max_blk_size * mmc->max_segs;
+	} else {
+		mmc->max_segs = 64;
+		mmc->max_req_size = 32768 * 512;
+		mmc->max_blk_size = 32768;
+		mmc->max_blk_count = 512;
+	}
+
+	/* Assume card is present initially */
+	set_bit(ATMCI_CARD_PRESENT, &slot->flags);
+	if (gpio_is_valid(slot->detect_pin)) {
+		if (devm_gpio_request(&host->pdev->dev, slot->detect_pin,
+				      "mmc_detect")) {
+			dev_dbg(&mmc->class_dev, "no detect pin available\n");
+			slot->detect_pin = -EBUSY;
+		} else if (gpio_get_value(slot->detect_pin) ^
+				slot->detect_is_active_high) {
+			clear_bit(ATMCI_CARD_PRESENT, &slot->flags);
+		}
+	}
+
+	if (!gpio_is_valid(slot->detect_pin)) {
+		if (slot_data->non_removable)
+			mmc->caps |= MMC_CAP_NONREMOVABLE;
+		else
+			mmc->caps |= MMC_CAP_NEEDS_POLL;
+	}
+
+	if (gpio_is_valid(slot->wp_pin)) {
+		if (devm_gpio_request(&host->pdev->dev, slot->wp_pin,
+				      "mmc_wp")) {
+			dev_dbg(&mmc->class_dev, "no WP pin available\n");
+			slot->wp_pin = -EBUSY;
+		}
+	}
+
+	host->slot[id] = slot;
+	mmc_regulator_get_supply(mmc);
+	ret = mmc_add_host(mmc);
+	if (ret) {
+		mmc_free_host(mmc);
+		return ret;
+	}
+
+	if (gpio_is_valid(slot->detect_pin)) {
+		timer_setup(&slot->detect_timer, atmci_detect_change, 0);
+
+		ret = request_irq(gpio_to_irq(slot->detect_pin),
+				atmci_detect_interrupt,
+				IRQF_TRIGGER_FALLING | IRQF_TRIGGER_RISING,
+				"mmc-detect", slot);
+		if (ret) {
+			dev_dbg(&mmc->class_dev,
+				"could not request IRQ %d for detect pin\n",
+				gpio_to_irq(slot->detect_pin));
+			slot->detect_pin = -EBUSY;
+		}
+	}
+
+	atmci_init_debugfs(slot);
+
+	return 0;
+}
+
+static void atmci_cleanup_slot(struct atmel_mci_slot *slot,
+		unsigned int id)
+{
+	/* Debugfs stuff is cleaned up by mmc core */
+
+	set_bit(ATMCI_SHUTDOWN, &slot->flags);
+	smp_wmb();
+
+	mmc_remove_host(slot->mmc);
+
+	if (gpio_is_valid(slot->detect_pin)) {
+		int pin = slot->detect_pin;
+
+		free_irq(gpio_to_irq(pin), slot);
+		del_timer_sync(&slot->detect_timer);
+	}
+
+	slot->host->slot[id] = NULL;
+	mmc_free_host(slot->mmc);
+}
+
+static int atmci_configure_dma(struct atmel_mci *host)
+{
+	host->dma.chan = dma_request_chan(&host->pdev->dev, "rxtx");
+
+	if (PTR_ERR(host->dma.chan) == -ENODEV) {
+		struct mci_platform_data *pdata = host->pdev->dev.platform_data;
+		dma_cap_mask_t mask;
+
+		if (!pdata || !pdata->dma_filter)
+			return -ENODEV;
+
+		dma_cap_zero(mask);
+		dma_cap_set(DMA_SLAVE, mask);
+
+		host->dma.chan = dma_request_channel(mask, pdata->dma_filter,
+						     pdata->dma_slave);
+		if (!host->dma.chan)
+			host->dma.chan = ERR_PTR(-ENODEV);
+	}
+
+	if (IS_ERR(host->dma.chan))
+		return PTR_ERR(host->dma.chan);
+
+	dev_info(&host->pdev->dev, "using %s for DMA transfers\n",
+		 dma_chan_name(host->dma.chan));
+
+	host->dma_conf.src_addr = host->mapbase + ATMCI_RDR;
+	host->dma_conf.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	host->dma_conf.src_maxburst = 1;
+	host->dma_conf.dst_addr = host->mapbase + ATMCI_TDR;
+	host->dma_conf.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	host->dma_conf.dst_maxburst = 1;
+	host->dma_conf.device_fc = false;
+
+	return 0;
+}
+
+/*
+ * HSMCI (High Speed MCI) module is not fully compatible with MCI module.
+ * HSMCI provides DMA support and a new config register but no more supports
+ * PDC.
+ */
+static void atmci_get_cap(struct atmel_mci *host)
+{
+	unsigned int version;
+
+	version = atmci_get_version(host);
+	dev_info(&host->pdev->dev,
+			"version: 0x%x\n", version);
+
+	host->caps.has_dma_conf_reg = 0;
+	host->caps.has_pdc = 1;
+	host->caps.has_cfg_reg = 0;
+	host->caps.has_cstor_reg = 0;
+	host->caps.has_highspeed = 0;
+	host->caps.has_rwproof = 0;
+	host->caps.has_odd_clk_div = 0;
+	host->caps.has_bad_data_ordering = 1;
+	host->caps.need_reset_after_xfer = 1;
+	host->caps.need_blksz_mul_4 = 1;
+	host->caps.need_notbusy_for_read_ops = 0;
+
+	/* keep only major version number */
+	switch (version & 0xf00) {
+	case 0x600:
+	case 0x500:
+		host->caps.has_odd_clk_div = 1;
+		fallthrough;
+	case 0x400:
+	case 0x300:
+		host->caps.has_dma_conf_reg = 1;
+		host->caps.has_pdc = 0;
+		host->caps.has_cfg_reg = 1;
+		host->caps.has_cstor_reg = 1;
+		host->caps.has_highspeed = 1;
+		fallthrough;
+	case 0x200:
+		host->caps.has_rwproof = 1;
+		host->caps.need_blksz_mul_4 = 0;
+		host->caps.need_notbusy_for_read_ops = 1;
+		fallthrough;
+	case 0x100:
+		host->caps.has_bad_data_ordering = 0;
+		host->caps.need_reset_after_xfer = 0;
+		fallthrough;
+	case 0x0:
+		break;
+	default:
+		host->caps.has_pdc = 0;
+		dev_warn(&host->pdev->dev,
+				"Unmanaged mci version, set minimum capabilities\n");
+		break;
+	}
+}
+
+static int atmci_probe(struct platform_device *pdev)
+{
+	struct mci_platform_data	*pdata;
+	struct atmel_mci		*host;
+	struct resource			*regs;
+	unsigned int			nr_slots;
+	int				irq;
+	int				ret, i;
+
+	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!regs)
+		return -ENXIO;
+	pdata = pdev->dev.platform_data;
+	if (!pdata) {
+		pdata = atmci_of_init(pdev);
+		if (IS_ERR(pdata)) {
+			dev_err(&pdev->dev, "platform data not available\n");
+			return PTR_ERR(pdata);
+		}
+	}
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return irq;
+
+	host = devm_kzalloc(&pdev->dev, sizeof(*host), GFP_KERNEL);
+	if (!host)
+		return -ENOMEM;
+
+	host->pdev = pdev;
+	spin_lock_init(&host->lock);
+	INIT_LIST_HEAD(&host->queue);
+
+	host->mck = devm_clk_get(&pdev->dev, "mci_clk");
+	if (IS_ERR(host->mck))
+		return PTR_ERR(host->mck);
+
+	host->regs = devm_ioremap(&pdev->dev, regs->start, resource_size(regs));
+	if (!host->regs)
+		return -ENOMEM;
+
+	ret = clk_prepare_enable(host->mck);
+	if (ret)
+		return ret;
+
+	atmci_writel(host, ATMCI_CR, ATMCI_CR_SWRST);
+	host->bus_hz = clk_get_rate(host->mck);
+
+	host->mapbase = regs->start;
+
+	tasklet_init(&host->tasklet, atmci_tasklet_func, (unsigned long)host);
+
+	ret = request_irq(irq, atmci_interrupt, 0, dev_name(&pdev->dev), host);
+	if (ret) {
+		clk_disable_unprepare(host->mck);
+		return ret;
+	}
+
+	/* Get MCI capabilities and set operations according to it */
+	atmci_get_cap(host);
+	ret = atmci_configure_dma(host);
+	if (ret == -EPROBE_DEFER)
+		goto err_dma_probe_defer;
+	if (ret == 0) {
+		host->prepare_data = &atmci_prepare_data_dma;
+		host->submit_data = &atmci_submit_data_dma;
+		host->stop_transfer = &atmci_stop_transfer_dma;
+	} else if (host->caps.has_pdc) {
+		dev_info(&pdev->dev, "using PDC\n");
+		host->prepare_data = &atmci_prepare_data_pdc;
+		host->submit_data = &atmci_submit_data_pdc;
+		host->stop_transfer = &atmci_stop_transfer_pdc;
+	} else {
+		dev_info(&pdev->dev, "using PIO\n");
+		host->prepare_data = &atmci_prepare_data;
+		host->submit_data = &atmci_submit_data;
+		host->stop_transfer = &atmci_stop_transfer;
+	}
+
+	platform_set_drvdata(pdev, host);
+
+	timer_setup(&host->timer, atmci_timeout_timer, 0);
+
+	pm_runtime_get_noresume(&pdev->dev);
+	pm_runtime_set_active(&pdev->dev);
+	pm_runtime_set_autosuspend_delay(&pdev->dev, AUTOSUSPEND_DELAY);
+	pm_runtime_use_autosuspend(&pdev->dev);
+	pm_runtime_enable(&pdev->dev);
+
+	/* We need at least one slot to succeed */
+	nr_slots = 0;
+	ret = -ENODEV;
+	if (pdata->slot[0].bus_width) {
+		ret = atmci_init_slot(host, &pdata->slot[0],
+				0, ATMCI_SDCSEL_SLOT_A, ATMCI_SDIOIRQA);
+		if (!ret) {
+			nr_slots++;
+			host->buf_size = host->slot[0]->mmc->max_req_size;
+		}
+	}
+	if (pdata->slot[1].bus_width) {
+		ret = atmci_init_slot(host, &pdata->slot[1],
+				1, ATMCI_SDCSEL_SLOT_B, ATMCI_SDIOIRQB);
+		if (!ret) {
+			nr_slots++;
+			if (host->slot[1]->mmc->max_req_size > host->buf_size)
+				host->buf_size =
+					host->slot[1]->mmc->max_req_size;
+		}
+	}
+
+	if (!nr_slots) {
+		dev_err(&pdev->dev, "init failed: no slot defined\n");
+		goto err_init_slot;
+	}
+
+	if (!host->caps.has_rwproof) {
+		host->buffer = dma_alloc_coherent(&pdev->dev, host->buf_size,
+		                                  &host->buf_phys_addr,
+						  GFP_KERNEL);
+		if (!host->buffer) {
+			ret = -ENOMEM;
+			dev_err(&pdev->dev, "buffer allocation failed\n");
+			goto err_dma_alloc;
+		}
+	}
+
+	dev_info(&pdev->dev,
+			"Atmel MCI controller at 0x%08lx irq %d, %u slots\n",
+			host->mapbase, irq, nr_slots);
+
+	pm_runtime_mark_last_busy(&host->pdev->dev);
+	pm_runtime_put_autosuspend(&pdev->dev);
+
+	return 0;
+
+err_dma_alloc:
+	for (i = 0; i < ATMCI_MAX_NR_SLOTS; i++) {
+		if (host->slot[i])
+			atmci_cleanup_slot(host->slot[i], i);
+	}
+err_init_slot:
+	clk_disable_unprepare(host->mck);
+
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+
+	del_timer_sync(&host->timer);
+	if (!IS_ERR(host->dma.chan))
+		dma_release_channel(host->dma.chan);
+err_dma_probe_defer:
+	free_irq(irq, host);
+	return ret;
+}
+
+static int atmci_remove(struct platform_device *pdev)
+{
+	struct atmel_mci	*host = platform_get_drvdata(pdev);
+	unsigned int		i;
+
+	pm_runtime_get_sync(&pdev->dev);
+
+	if (host->buffer)
+		dma_free_coherent(&pdev->dev, host->buf_size,
+		                  host->buffer, host->buf_phys_addr);
+
+	for (i = 0; i < ATMCI_MAX_NR_SLOTS; i++) {
+		if (host->slot[i])
+			atmci_cleanup_slot(host->slot[i], i);
+	}
+
+	atmci_writel(host, ATMCI_IDR, ~0UL);
+	atmci_writel(host, ATMCI_CR, ATMCI_CR_MCIDIS);
+	atmci_readl(host, ATMCI_SR);
+
+	del_timer_sync(&host->timer);
+	if (!IS_ERR(host->dma.chan))
+		dma_release_channel(host->dma.chan);
+
+	free_irq(platform_get_irq(pdev, 0), host);
+
+	clk_disable_unprepare(host->mck);
+
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static int atmci_runtime_suspend(struct device *dev)
+{
+	struct atmel_mci *host = dev_get_drvdata(dev);
+
+	clk_disable_unprepare(host->mck);
+
+	pinctrl_pm_select_sleep_state(dev);
+
+	return 0;
+}
+
+static int atmci_runtime_resume(struct device *dev)
+{
+	struct atmel_mci *host = dev_get_drvdata(dev);
+
+	pinctrl_select_default_state(dev);
+
+	return clk_prepare_enable(host->mck);
+}
+#endif
+
+static const struct dev_pm_ops atmci_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(atmci_runtime_suspend, atmci_runtime_resume, NULL)
+};
+
+static struct platform_driver atmci_driver = {
+	.probe		= atmci_probe,
+	.remove		= atmci_remove,
+	.driver		= {
+		.name		= "atmel_mci",
+		.probe_type	= PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table	= of_match_ptr(atmci_dt_ids),
+		.pm		= &atmci_dev_pm_ops,
+	},
+};
+module_platform_driver(atmci_driver);
+
+MODULE_DESCRIPTION("Atmel Multimedia Card Interface driver");
+MODULE_AUTHOR("Haavard Skinnemoen (Atmel)");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/au1xmmc.c b/drivers/mmc/host/au1xmmc.c
new file mode 100644
index 000000000..56a3bf51d
--- /dev/null
+++ b/drivers/mmc/host/au1xmmc.c
@@ -0,0 +1,1225 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * linux/drivers/mmc/host/au1xmmc.c - AU1XX0 MMC driver
+ *
+ *  Copyright (c) 2005, Advanced Micro Devices, Inc.
+ *
+ *  Developed with help from the 2.4.30 MMC AU1XXX controller including
+ *  the following copyright notices:
+ *     Copyright (c) 2003-2004 Embedded Edge, LLC.
+ *     Portions Copyright (C) 2002 Embedix, Inc
+ *     Copyright 2002 Hewlett-Packard Company
+
+ *  2.6 version of this driver inspired by:
+ *     (drivers/mmc/wbsd.c) Copyright (C) 2004-2005 Pierre Ossman,
+ *     All Rights Reserved.
+ *     (drivers/mmc/pxa.c) Copyright (C) 2003 Russell King,
+ *     All Rights Reserved.
+ *
+
+ */
+
+/* Why don't we use the SD controllers' carddetect feature?
+ *
+ * From the AU1100 MMC application guide:
+ * If the Au1100-based design is intended to support both MultiMediaCards
+ * and 1- or 4-data bit SecureDigital cards, then the solution is to
+ * connect a weak (560KOhm) pull-up resistor to connector pin 1.
+ * In doing so, a MMC card never enters SPI-mode communications,
+ * but now the SecureDigital card-detect feature of CD/DAT3 is ineffective
+ * (the low to high transition will not occur).
+ */
+
+#include <linux/clk.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/platform_device.h>
+#include <linux/mm.h>
+#include <linux/interrupt.h>
+#include <linux/dma-mapping.h>
+#include <linux/scatterlist.h>
+#include <linux/highmem.h>
+#include <linux/leds.h>
+#include <linux/mmc/host.h>
+#include <linux/slab.h>
+
+#include <asm/io.h>
+#include <asm/mach-au1x00/au1000.h>
+#include <asm/mach-au1x00/au1xxx_dbdma.h>
+#include <asm/mach-au1x00/au1100_mmc.h>
+
+#define DRIVER_NAME "au1xxx-mmc"
+
+/* Set this to enable special debugging macros */
+/* #define DEBUG */
+
+#ifdef DEBUG
+#define DBG(fmt, idx, args...)	\
+	pr_debug("au1xmmc(%d): DEBUG: " fmt, idx, ##args)
+#else
+#define DBG(fmt, idx, args...) do {} while (0)
+#endif
+
+/* Hardware definitions */
+#define AU1XMMC_DESCRIPTOR_COUNT 1
+
+/* max DMA seg size: 64KB on Au1100, 4MB on Au1200 */
+#define AU1100_MMC_DESCRIPTOR_SIZE 0x0000ffff
+#define AU1200_MMC_DESCRIPTOR_SIZE 0x003fffff
+
+#define AU1XMMC_OCR (MMC_VDD_27_28 | MMC_VDD_28_29 | MMC_VDD_29_30 | \
+		     MMC_VDD_30_31 | MMC_VDD_31_32 | MMC_VDD_32_33 | \
+		     MMC_VDD_33_34 | MMC_VDD_34_35 | MMC_VDD_35_36)
+
+/* This gives us a hard value for the stop command that we can write directly
+ * to the command register.
+ */
+#define STOP_CMD	\
+	(SD_CMD_RT_1B | SD_CMD_CT_7 | (0xC << SD_CMD_CI_SHIFT) | SD_CMD_GO)
+
+/* This is the set of interrupts that we configure by default. */
+#define AU1XMMC_INTERRUPTS 				\
+	(SD_CONFIG_SC | SD_CONFIG_DT | SD_CONFIG_RAT |	\
+	 SD_CONFIG_CR | SD_CONFIG_I)
+
+/* The poll event (looking for insert/remove events runs twice a second. */
+#define AU1XMMC_DETECT_TIMEOUT (HZ/2)
+
+struct au1xmmc_host {
+	struct mmc_host *mmc;
+	struct mmc_request *mrq;
+
+	u32 flags;
+	void __iomem *iobase;
+	u32 clock;
+	u32 bus_width;
+	u32 power_mode;
+
+	int status;
+
+	struct {
+		int len;
+		int dir;
+	} dma;
+
+	struct {
+		int index;
+		int offset;
+		int len;
+	} pio;
+
+	u32 tx_chan;
+	u32 rx_chan;
+
+	int irq;
+
+	struct tasklet_struct finish_task;
+	struct tasklet_struct data_task;
+	struct au1xmmc_platform_data *platdata;
+	struct platform_device *pdev;
+	struct resource *ioarea;
+	struct clk *clk;
+};
+
+/* Status flags used by the host structure */
+#define HOST_F_XMIT	0x0001
+#define HOST_F_RECV	0x0002
+#define HOST_F_DMA	0x0010
+#define HOST_F_DBDMA	0x0020
+#define HOST_F_ACTIVE	0x0100
+#define HOST_F_STOP	0x1000
+
+#define HOST_S_IDLE	0x0001
+#define HOST_S_CMD	0x0002
+#define HOST_S_DATA	0x0003
+#define HOST_S_STOP	0x0004
+
+/* Easy access macros */
+#define HOST_STATUS(h)	((h)->iobase + SD_STATUS)
+#define HOST_CONFIG(h)	((h)->iobase + SD_CONFIG)
+#define HOST_ENABLE(h)	((h)->iobase + SD_ENABLE)
+#define HOST_TXPORT(h)	((h)->iobase + SD_TXPORT)
+#define HOST_RXPORT(h)	((h)->iobase + SD_RXPORT)
+#define HOST_CMDARG(h)	((h)->iobase + SD_CMDARG)
+#define HOST_BLKSIZE(h)	((h)->iobase + SD_BLKSIZE)
+#define HOST_CMD(h)	((h)->iobase + SD_CMD)
+#define HOST_CONFIG2(h)	((h)->iobase + SD_CONFIG2)
+#define HOST_TIMEOUT(h)	((h)->iobase + SD_TIMEOUT)
+#define HOST_DEBUG(h)	((h)->iobase + SD_DEBUG)
+
+#define DMA_CHANNEL(h)	\
+	(((h)->flags & HOST_F_XMIT) ? (h)->tx_chan : (h)->rx_chan)
+
+static inline int has_dbdma(void)
+{
+	switch (alchemy_get_cputype()) {
+	case ALCHEMY_CPU_AU1200:
+	case ALCHEMY_CPU_AU1300:
+		return 1;
+	default:
+		return 0;
+	}
+}
+
+static inline void IRQ_ON(struct au1xmmc_host *host, u32 mask)
+{
+	u32 val = __raw_readl(HOST_CONFIG(host));
+	val |= mask;
+	__raw_writel(val, HOST_CONFIG(host));
+	wmb(); /* drain writebuffer */
+}
+
+static inline void FLUSH_FIFO(struct au1xmmc_host *host)
+{
+	u32 val = __raw_readl(HOST_CONFIG2(host));
+
+	__raw_writel(val | SD_CONFIG2_FF, HOST_CONFIG2(host));
+	wmb(); /* drain writebuffer */
+	mdelay(1);
+
+	/* SEND_STOP will turn off clock control - this re-enables it */
+	val &= ~SD_CONFIG2_DF;
+
+	__raw_writel(val, HOST_CONFIG2(host));
+	wmb(); /* drain writebuffer */
+}
+
+static inline void IRQ_OFF(struct au1xmmc_host *host, u32 mask)
+{
+	u32 val = __raw_readl(HOST_CONFIG(host));
+	val &= ~mask;
+	__raw_writel(val, HOST_CONFIG(host));
+	wmb(); /* drain writebuffer */
+}
+
+static inline void SEND_STOP(struct au1xmmc_host *host)
+{
+	u32 config2;
+
+	WARN_ON(host->status != HOST_S_DATA);
+	host->status = HOST_S_STOP;
+
+	config2 = __raw_readl(HOST_CONFIG2(host));
+	__raw_writel(config2 | SD_CONFIG2_DF, HOST_CONFIG2(host));
+	wmb(); /* drain writebuffer */
+
+	/* Send the stop command */
+	__raw_writel(STOP_CMD, HOST_CMD(host));
+	wmb(); /* drain writebuffer */
+}
+
+static void au1xmmc_set_power(struct au1xmmc_host *host, int state)
+{
+	if (host->platdata && host->platdata->set_power)
+		host->platdata->set_power(host->mmc, state);
+}
+
+static int au1xmmc_card_inserted(struct mmc_host *mmc)
+{
+	struct au1xmmc_host *host = mmc_priv(mmc);
+
+	if (host->platdata && host->platdata->card_inserted)
+		return !!host->platdata->card_inserted(host->mmc);
+
+	return -ENOSYS;
+}
+
+static int au1xmmc_card_readonly(struct mmc_host *mmc)
+{
+	struct au1xmmc_host *host = mmc_priv(mmc);
+
+	if (host->platdata && host->platdata->card_readonly)
+		return !!host->platdata->card_readonly(mmc);
+
+	return -ENOSYS;
+}
+
+static void au1xmmc_finish_request(struct au1xmmc_host *host)
+{
+	struct mmc_request *mrq = host->mrq;
+
+	host->mrq = NULL;
+	host->flags &= HOST_F_ACTIVE | HOST_F_DMA;
+
+	host->dma.len = 0;
+	host->dma.dir = 0;
+
+	host->pio.index  = 0;
+	host->pio.offset = 0;
+	host->pio.len = 0;
+
+	host->status = HOST_S_IDLE;
+
+	mmc_request_done(host->mmc, mrq);
+}
+
+static void au1xmmc_tasklet_finish(unsigned long param)
+{
+	struct au1xmmc_host *host = (struct au1xmmc_host *) param;
+	au1xmmc_finish_request(host);
+}
+
+static int au1xmmc_send_command(struct au1xmmc_host *host,
+				struct mmc_command *cmd, struct mmc_data *data)
+{
+	u32 mmccmd = (cmd->opcode << SD_CMD_CI_SHIFT);
+
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE:
+		break;
+	case MMC_RSP_R1:
+		mmccmd |= SD_CMD_RT_1;
+		break;
+	case MMC_RSP_R1B:
+		mmccmd |= SD_CMD_RT_1B;
+		break;
+	case MMC_RSP_R2:
+		mmccmd |= SD_CMD_RT_2;
+		break;
+	case MMC_RSP_R3:
+		mmccmd |= SD_CMD_RT_3;
+		break;
+	default:
+		pr_info("au1xmmc: unhandled response type %02x\n",
+			mmc_resp_type(cmd));
+		return -EINVAL;
+	}
+
+	if (data) {
+		if (data->flags & MMC_DATA_READ) {
+			if (data->blocks > 1)
+				mmccmd |= SD_CMD_CT_4;
+			else
+				mmccmd |= SD_CMD_CT_2;
+		} else if (data->flags & MMC_DATA_WRITE) {
+			if (data->blocks > 1)
+				mmccmd |= SD_CMD_CT_3;
+			else
+				mmccmd |= SD_CMD_CT_1;
+		}
+	}
+
+	__raw_writel(cmd->arg, HOST_CMDARG(host));
+	wmb(); /* drain writebuffer */
+
+	__raw_writel((mmccmd | SD_CMD_GO), HOST_CMD(host));
+	wmb(); /* drain writebuffer */
+
+	/* Wait for the command to go on the line */
+	while (__raw_readl(HOST_CMD(host)) & SD_CMD_GO)
+		/* nop */;
+
+	return 0;
+}
+
+static void au1xmmc_data_complete(struct au1xmmc_host *host, u32 status)
+{
+	struct mmc_request *mrq = host->mrq;
+	struct mmc_data *data;
+	u32 crc;
+
+	WARN_ON((host->status != HOST_S_DATA) && (host->status != HOST_S_STOP));
+
+	if (host->mrq == NULL)
+		return;
+
+	data = mrq->cmd->data;
+
+	if (status == 0)
+		status = __raw_readl(HOST_STATUS(host));
+
+	/* The transaction is really over when the SD_STATUS_DB bit is clear */
+	while ((host->flags & HOST_F_XMIT) && (status & SD_STATUS_DB))
+		status = __raw_readl(HOST_STATUS(host));
+
+	data->error = 0;
+	dma_unmap_sg(mmc_dev(host->mmc), data->sg, data->sg_len, host->dma.dir);
+
+        /* Process any errors */
+	crc = (status & (SD_STATUS_WC | SD_STATUS_RC));
+	if (host->flags & HOST_F_XMIT)
+		crc |= ((status & 0x07) == 0x02) ? 0 : 1;
+
+	if (crc)
+		data->error = -EILSEQ;
+
+	/* Clear the CRC bits */
+	__raw_writel(SD_STATUS_WC | SD_STATUS_RC, HOST_STATUS(host));
+
+	data->bytes_xfered = 0;
+
+	if (!data->error) {
+		if (host->flags & (HOST_F_DMA | HOST_F_DBDMA)) {
+			u32 chan = DMA_CHANNEL(host);
+
+			chan_tab_t *c = *((chan_tab_t **)chan);
+			au1x_dma_chan_t *cp = c->chan_ptr;
+			data->bytes_xfered = cp->ddma_bytecnt;
+		} else
+			data->bytes_xfered =
+				(data->blocks * data->blksz) - host->pio.len;
+	}
+
+	au1xmmc_finish_request(host);
+}
+
+static void au1xmmc_tasklet_data(unsigned long param)
+{
+	struct au1xmmc_host *host = (struct au1xmmc_host *)param;
+
+	u32 status = __raw_readl(HOST_STATUS(host));
+	au1xmmc_data_complete(host, status);
+}
+
+#define AU1XMMC_MAX_TRANSFER 8
+
+static void au1xmmc_send_pio(struct au1xmmc_host *host)
+{
+	struct mmc_data *data;
+	int sg_len, max, count;
+	unsigned char *sg_ptr, val;
+	u32 status;
+	struct scatterlist *sg;
+
+	data = host->mrq->data;
+
+	if (!(host->flags & HOST_F_XMIT))
+		return;
+
+	/* This is the pointer to the data buffer */
+	sg = &data->sg[host->pio.index];
+	sg_ptr = kmap_atomic(sg_page(sg)) + sg->offset + host->pio.offset;
+
+	/* This is the space left inside the buffer */
+	sg_len = data->sg[host->pio.index].length - host->pio.offset;
+
+	/* Check if we need less than the size of the sg_buffer */
+	max = (sg_len > host->pio.len) ? host->pio.len : sg_len;
+	if (max > AU1XMMC_MAX_TRANSFER)
+		max = AU1XMMC_MAX_TRANSFER;
+
+	for (count = 0; count < max; count++) {
+		status = __raw_readl(HOST_STATUS(host));
+
+		if (!(status & SD_STATUS_TH))
+			break;
+
+		val = sg_ptr[count];
+
+		__raw_writel((unsigned long)val, HOST_TXPORT(host));
+		wmb(); /* drain writebuffer */
+	}
+	kunmap_atomic(sg_ptr);
+
+	host->pio.len -= count;
+	host->pio.offset += count;
+
+	if (count == sg_len) {
+		host->pio.index++;
+		host->pio.offset = 0;
+	}
+
+	if (host->pio.len == 0) {
+		IRQ_OFF(host, SD_CONFIG_TH);
+
+		if (host->flags & HOST_F_STOP)
+			SEND_STOP(host);
+
+		tasklet_schedule(&host->data_task);
+	}
+}
+
+static void au1xmmc_receive_pio(struct au1xmmc_host *host)
+{
+	struct mmc_data *data;
+	int max, count, sg_len = 0;
+	unsigned char *sg_ptr = NULL;
+	u32 status, val;
+	struct scatterlist *sg;
+
+	data = host->mrq->data;
+
+	if (!(host->flags & HOST_F_RECV))
+		return;
+
+	max = host->pio.len;
+
+	if (host->pio.index < host->dma.len) {
+		sg = &data->sg[host->pio.index];
+		sg_ptr = kmap_atomic(sg_page(sg)) + sg->offset + host->pio.offset;
+
+		/* This is the space left inside the buffer */
+		sg_len = sg_dma_len(&data->sg[host->pio.index]) - host->pio.offset;
+
+		/* Check if we need less than the size of the sg_buffer */
+		if (sg_len < max)
+			max = sg_len;
+	}
+
+	if (max > AU1XMMC_MAX_TRANSFER)
+		max = AU1XMMC_MAX_TRANSFER;
+
+	for (count = 0; count < max; count++) {
+		status = __raw_readl(HOST_STATUS(host));
+
+		if (!(status & SD_STATUS_NE))
+			break;
+
+		if (status & SD_STATUS_RC) {
+			DBG("RX CRC Error [%d + %d].\n", host->pdev->id,
+					host->pio.len, count);
+			break;
+		}
+
+		if (status & SD_STATUS_RO) {
+			DBG("RX Overrun [%d + %d]\n", host->pdev->id,
+					host->pio.len, count);
+			break;
+		}
+		else if (status & SD_STATUS_RU) {
+			DBG("RX Underrun [%d + %d]\n", host->pdev->id,
+					host->pio.len,	count);
+			break;
+		}
+
+		val = __raw_readl(HOST_RXPORT(host));
+
+		if (sg_ptr)
+			sg_ptr[count] = (unsigned char)(val & 0xFF);
+	}
+	if (sg_ptr)
+		kunmap_atomic(sg_ptr);
+
+	host->pio.len -= count;
+	host->pio.offset += count;
+
+	if (sg_len && count == sg_len) {
+		host->pio.index++;
+		host->pio.offset = 0;
+	}
+
+	if (host->pio.len == 0) {
+		/* IRQ_OFF(host, SD_CONFIG_RA | SD_CONFIG_RF); */
+		IRQ_OFF(host, SD_CONFIG_NE);
+
+		if (host->flags & HOST_F_STOP)
+			SEND_STOP(host);
+
+		tasklet_schedule(&host->data_task);
+	}
+}
+
+/* This is called when a command has been completed - grab the response
+ * and check for errors.  Then start the data transfer if it is indicated.
+ */
+static void au1xmmc_cmd_complete(struct au1xmmc_host *host, u32 status)
+{
+	struct mmc_request *mrq = host->mrq;
+	struct mmc_command *cmd;
+	u32 r[4];
+	int i, trans;
+
+	if (!host->mrq)
+		return;
+
+	cmd = mrq->cmd;
+	cmd->error = 0;
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136) {
+			r[0] = __raw_readl(host->iobase + SD_RESP3);
+			r[1] = __raw_readl(host->iobase + SD_RESP2);
+			r[2] = __raw_readl(host->iobase + SD_RESP1);
+			r[3] = __raw_readl(host->iobase + SD_RESP0);
+
+			/* The CRC is omitted from the response, so really
+			 * we only got 120 bytes, but the engine expects
+			 * 128 bits, so we have to shift things up.
+			 */
+			for (i = 0; i < 4; i++) {
+				cmd->resp[i] = (r[i] & 0x00FFFFFF) << 8;
+				if (i != 3)
+					cmd->resp[i] |= (r[i + 1] & 0xFF000000) >> 24;
+			}
+		} else {
+			/* Techincally, we should be getting all 48 bits of
+			 * the response (SD_RESP1 + SD_RESP2), but because
+			 * our response omits the CRC, our data ends up
+			 * being shifted 8 bits to the right.  In this case,
+			 * that means that the OSR data starts at bit 31,
+			 * so we can just read RESP0 and return that.
+			 */
+			cmd->resp[0] = __raw_readl(host->iobase + SD_RESP0);
+		}
+	}
+
+        /* Figure out errors */
+	if (status & (SD_STATUS_SC | SD_STATUS_WC | SD_STATUS_RC))
+		cmd->error = -EILSEQ;
+
+	trans = host->flags & (HOST_F_XMIT | HOST_F_RECV);
+
+	if (!trans || cmd->error) {
+		IRQ_OFF(host, SD_CONFIG_TH | SD_CONFIG_RA | SD_CONFIG_RF);
+		tasklet_schedule(&host->finish_task);
+		return;
+	}
+
+	host->status = HOST_S_DATA;
+
+	if ((host->flags & (HOST_F_DMA | HOST_F_DBDMA))) {
+		u32 channel = DMA_CHANNEL(host);
+
+		/* Start the DBDMA as soon as the buffer gets something in it */
+
+		if (host->flags & HOST_F_RECV) {
+			u32 mask = SD_STATUS_DB | SD_STATUS_NE;
+
+			while((status & mask) != mask)
+				status = __raw_readl(HOST_STATUS(host));
+		}
+
+		au1xxx_dbdma_start(channel);
+	}
+}
+
+static void au1xmmc_set_clock(struct au1xmmc_host *host, int rate)
+{
+	unsigned int pbus = clk_get_rate(host->clk);
+	unsigned int divisor = ((pbus / rate) / 2) - 1;
+	u32 config;
+
+	config = __raw_readl(HOST_CONFIG(host));
+
+	config &= ~(SD_CONFIG_DIV);
+	config |= (divisor & SD_CONFIG_DIV) | SD_CONFIG_DE;
+
+	__raw_writel(config, HOST_CONFIG(host));
+	wmb(); /* drain writebuffer */
+}
+
+static int au1xmmc_prepare_data(struct au1xmmc_host *host,
+				struct mmc_data *data)
+{
+	int datalen = data->blocks * data->blksz;
+
+	if (data->flags & MMC_DATA_READ)
+		host->flags |= HOST_F_RECV;
+	else
+		host->flags |= HOST_F_XMIT;
+
+	if (host->mrq->stop)
+		host->flags |= HOST_F_STOP;
+
+	host->dma.dir = DMA_BIDIRECTIONAL;
+
+	host->dma.len = dma_map_sg(mmc_dev(host->mmc), data->sg,
+				   data->sg_len, host->dma.dir);
+
+	if (host->dma.len == 0)
+		return -ETIMEDOUT;
+
+	__raw_writel(data->blksz - 1, HOST_BLKSIZE(host));
+
+	if (host->flags & (HOST_F_DMA | HOST_F_DBDMA)) {
+		int i;
+		u32 channel = DMA_CHANNEL(host);
+
+		au1xxx_dbdma_stop(channel);
+
+		for (i = 0; i < host->dma.len; i++) {
+			u32 ret = 0, flags = DDMA_FLAGS_NOIE;
+			struct scatterlist *sg = &data->sg[i];
+			int sg_len = sg->length;
+
+			int len = (datalen > sg_len) ? sg_len : datalen;
+
+			if (i == host->dma.len - 1)
+				flags = DDMA_FLAGS_IE;
+
+			if (host->flags & HOST_F_XMIT) {
+				ret = au1xxx_dbdma_put_source(channel,
+					sg_phys(sg), len, flags);
+			} else {
+				ret = au1xxx_dbdma_put_dest(channel,
+					sg_phys(sg), len, flags);
+			}
+
+			if (!ret)
+				goto dataerr;
+
+			datalen -= len;
+		}
+	} else {
+		host->pio.index = 0;
+		host->pio.offset = 0;
+		host->pio.len = datalen;
+
+		if (host->flags & HOST_F_XMIT)
+			IRQ_ON(host, SD_CONFIG_TH);
+		else
+			IRQ_ON(host, SD_CONFIG_NE);
+			/* IRQ_ON(host, SD_CONFIG_RA | SD_CONFIG_RF); */
+	}
+
+	return 0;
+
+dataerr:
+	dma_unmap_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+			host->dma.dir);
+	return -ETIMEDOUT;
+}
+
+/* This actually starts a command or data transaction */
+static void au1xmmc_request(struct mmc_host* mmc, struct mmc_request* mrq)
+{
+	struct au1xmmc_host *host = mmc_priv(mmc);
+	int ret = 0;
+
+	WARN_ON(irqs_disabled());
+	WARN_ON(host->status != HOST_S_IDLE);
+
+	host->mrq = mrq;
+	host->status = HOST_S_CMD;
+
+	/* fail request immediately if no card is present */
+	if (0 == au1xmmc_card_inserted(mmc)) {
+		mrq->cmd->error = -ENOMEDIUM;
+		au1xmmc_finish_request(host);
+		return;
+	}
+
+	if (mrq->data) {
+		FLUSH_FIFO(host);
+		ret = au1xmmc_prepare_data(host, mrq->data);
+	}
+
+	if (!ret)
+		ret = au1xmmc_send_command(host, mrq->cmd, mrq->data);
+
+	if (ret) {
+		mrq->cmd->error = ret;
+		au1xmmc_finish_request(host);
+	}
+}
+
+static void au1xmmc_reset_controller(struct au1xmmc_host *host)
+{
+	/* Apply the clock */
+	__raw_writel(SD_ENABLE_CE, HOST_ENABLE(host));
+	wmb(); /* drain writebuffer */
+	mdelay(1);
+
+	__raw_writel(SD_ENABLE_R | SD_ENABLE_CE, HOST_ENABLE(host));
+	wmb(); /* drain writebuffer */
+	mdelay(5);
+
+	__raw_writel(~0, HOST_STATUS(host));
+	wmb(); /* drain writebuffer */
+
+	__raw_writel(0, HOST_BLKSIZE(host));
+	__raw_writel(0x001fffff, HOST_TIMEOUT(host));
+	wmb(); /* drain writebuffer */
+
+	__raw_writel(SD_CONFIG2_EN, HOST_CONFIG2(host));
+	wmb(); /* drain writebuffer */
+
+	__raw_writel(SD_CONFIG2_EN | SD_CONFIG2_FF, HOST_CONFIG2(host));
+	wmb(); /* drain writebuffer */
+	mdelay(1);
+
+	__raw_writel(SD_CONFIG2_EN, HOST_CONFIG2(host));
+	wmb(); /* drain writebuffer */
+
+	/* Configure interrupts */
+	__raw_writel(AU1XMMC_INTERRUPTS, HOST_CONFIG(host));
+	wmb(); /* drain writebuffer */
+}
+
+
+static void au1xmmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct au1xmmc_host *host = mmc_priv(mmc);
+	u32 config2;
+
+	if (ios->power_mode == MMC_POWER_OFF)
+		au1xmmc_set_power(host, 0);
+	else if (ios->power_mode == MMC_POWER_ON) {
+		au1xmmc_set_power(host, 1);
+	}
+
+	if (ios->clock && ios->clock != host->clock) {
+		au1xmmc_set_clock(host, ios->clock);
+		host->clock = ios->clock;
+	}
+
+	config2 = __raw_readl(HOST_CONFIG2(host));
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_8:
+		config2 |= SD_CONFIG2_BB;
+		break;
+	case MMC_BUS_WIDTH_4:
+		config2 &= ~SD_CONFIG2_BB;
+		config2 |= SD_CONFIG2_WB;
+		break;
+	case MMC_BUS_WIDTH_1:
+		config2 &= ~(SD_CONFIG2_WB | SD_CONFIG2_BB);
+		break;
+	}
+	__raw_writel(config2, HOST_CONFIG2(host));
+	wmb(); /* drain writebuffer */
+}
+
+#define STATUS_TIMEOUT (SD_STATUS_RAT | SD_STATUS_DT)
+#define STATUS_DATA_IN  (SD_STATUS_NE)
+#define STATUS_DATA_OUT (SD_STATUS_TH)
+
+static irqreturn_t au1xmmc_irq(int irq, void *dev_id)
+{
+	struct au1xmmc_host *host = dev_id;
+	u32 status;
+
+	status = __raw_readl(HOST_STATUS(host));
+
+	if (!(status & SD_STATUS_I))
+		return IRQ_NONE;	/* not ours */
+
+	if (status & SD_STATUS_SI)	/* SDIO */
+		mmc_signal_sdio_irq(host->mmc);
+
+	if (host->mrq && (status & STATUS_TIMEOUT)) {
+		if (status & SD_STATUS_RAT)
+			host->mrq->cmd->error = -ETIMEDOUT;
+		else if (status & SD_STATUS_DT)
+			host->mrq->data->error = -ETIMEDOUT;
+
+		/* In PIO mode, interrupts might still be enabled */
+		IRQ_OFF(host, SD_CONFIG_NE | SD_CONFIG_TH);
+
+		/* IRQ_OFF(host, SD_CONFIG_TH | SD_CONFIG_RA | SD_CONFIG_RF); */
+		tasklet_schedule(&host->finish_task);
+	}
+#if 0
+	else if (status & SD_STATUS_DD) {
+		/* Sometimes we get a DD before a NE in PIO mode */
+		if (!(host->flags & HOST_F_DMA) && (status & SD_STATUS_NE))
+			au1xmmc_receive_pio(host);
+		else {
+			au1xmmc_data_complete(host, status);
+			/* tasklet_schedule(&host->data_task); */
+		}
+	}
+#endif
+	else if (status & SD_STATUS_CR) {
+		if (host->status == HOST_S_CMD)
+			au1xmmc_cmd_complete(host, status);
+
+	} else if (!(host->flags & HOST_F_DMA)) {
+		if ((host->flags & HOST_F_XMIT) && (status & STATUS_DATA_OUT))
+			au1xmmc_send_pio(host);
+		else if ((host->flags & HOST_F_RECV) && (status & STATUS_DATA_IN))
+			au1xmmc_receive_pio(host);
+
+	} else if (status & 0x203F3C70) {
+			DBG("Unhandled status %8.8x\n", host->pdev->id,
+				status);
+	}
+
+	__raw_writel(status, HOST_STATUS(host));
+	wmb(); /* drain writebuffer */
+
+	return IRQ_HANDLED;
+}
+
+/* 8bit memory DMA device */
+static dbdev_tab_t au1xmmc_mem_dbdev = {
+	.dev_id		= DSCR_CMD0_ALWAYS,
+	.dev_flags	= DEV_FLAGS_ANYUSE,
+	.dev_tsize	= 0,
+	.dev_devwidth	= 8,
+	.dev_physaddr	= 0x00000000,
+	.dev_intlevel	= 0,
+	.dev_intpolarity = 0,
+};
+static int memid;
+
+static void au1xmmc_dbdma_callback(int irq, void *dev_id)
+{
+	struct au1xmmc_host *host = (struct au1xmmc_host *)dev_id;
+
+	/* Avoid spurious interrupts */
+	if (!host->mrq)
+		return;
+
+	if (host->flags & HOST_F_STOP)
+		SEND_STOP(host);
+
+	tasklet_schedule(&host->data_task);
+}
+
+static int au1xmmc_dbdma_init(struct au1xmmc_host *host)
+{
+	struct resource *res;
+	int txid, rxid;
+
+	res = platform_get_resource(host->pdev, IORESOURCE_DMA, 0);
+	if (!res)
+		return -ENODEV;
+	txid = res->start;
+
+	res = platform_get_resource(host->pdev, IORESOURCE_DMA, 1);
+	if (!res)
+		return -ENODEV;
+	rxid = res->start;
+
+	if (!memid)
+		return -ENODEV;
+
+	host->tx_chan = au1xxx_dbdma_chan_alloc(memid, txid,
+				au1xmmc_dbdma_callback, (void *)host);
+	if (!host->tx_chan) {
+		dev_err(&host->pdev->dev, "cannot allocate TX DMA\n");
+		return -ENODEV;
+	}
+
+	host->rx_chan = au1xxx_dbdma_chan_alloc(rxid, memid,
+				au1xmmc_dbdma_callback, (void *)host);
+	if (!host->rx_chan) {
+		dev_err(&host->pdev->dev, "cannot allocate RX DMA\n");
+		au1xxx_dbdma_chan_free(host->tx_chan);
+		return -ENODEV;
+	}
+
+	au1xxx_dbdma_set_devwidth(host->tx_chan, 8);
+	au1xxx_dbdma_set_devwidth(host->rx_chan, 8);
+
+	au1xxx_dbdma_ring_alloc(host->tx_chan, AU1XMMC_DESCRIPTOR_COUNT);
+	au1xxx_dbdma_ring_alloc(host->rx_chan, AU1XMMC_DESCRIPTOR_COUNT);
+
+	/* DBDMA is good to go */
+	host->flags |= HOST_F_DMA | HOST_F_DBDMA;
+
+	return 0;
+}
+
+static void au1xmmc_dbdma_shutdown(struct au1xmmc_host *host)
+{
+	if (host->flags & HOST_F_DMA) {
+		host->flags &= ~HOST_F_DMA;
+		au1xxx_dbdma_chan_free(host->tx_chan);
+		au1xxx_dbdma_chan_free(host->rx_chan);
+	}
+}
+
+static void au1xmmc_enable_sdio_irq(struct mmc_host *mmc, int en)
+{
+	struct au1xmmc_host *host = mmc_priv(mmc);
+
+	if (en)
+		IRQ_ON(host, SD_CONFIG_SI);
+	else
+		IRQ_OFF(host, SD_CONFIG_SI);
+}
+
+static const struct mmc_host_ops au1xmmc_ops = {
+	.request	= au1xmmc_request,
+	.set_ios	= au1xmmc_set_ios,
+	.get_ro		= au1xmmc_card_readonly,
+	.get_cd		= au1xmmc_card_inserted,
+	.enable_sdio_irq = au1xmmc_enable_sdio_irq,
+};
+
+static int au1xmmc_probe(struct platform_device *pdev)
+{
+	struct mmc_host *mmc;
+	struct au1xmmc_host *host;
+	struct resource *r;
+	int ret, iflag;
+
+	mmc = mmc_alloc_host(sizeof(struct au1xmmc_host), &pdev->dev);
+	if (!mmc) {
+		dev_err(&pdev->dev, "no memory for mmc_host\n");
+		ret = -ENOMEM;
+		goto out0;
+	}
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+	host->platdata = pdev->dev.platform_data;
+	host->pdev = pdev;
+
+	ret = -ENODEV;
+	r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!r) {
+		dev_err(&pdev->dev, "no mmio defined\n");
+		goto out1;
+	}
+
+	host->ioarea = request_mem_region(r->start, resource_size(r),
+					   pdev->name);
+	if (!host->ioarea) {
+		dev_err(&pdev->dev, "mmio already in use\n");
+		goto out1;
+	}
+
+	host->iobase = ioremap(r->start, 0x3c);
+	if (!host->iobase) {
+		dev_err(&pdev->dev, "cannot remap mmio\n");
+		goto out2;
+	}
+
+	host->irq = platform_get_irq(pdev, 0);
+	if (host->irq < 0)
+		goto out3;
+
+	mmc->ops = &au1xmmc_ops;
+
+	mmc->f_min =   450000;
+	mmc->f_max = 24000000;
+
+	mmc->max_blk_size = 2048;
+	mmc->max_blk_count = 512;
+
+	mmc->ocr_avail = AU1XMMC_OCR;
+	mmc->caps = MMC_CAP_4_BIT_DATA | MMC_CAP_SDIO_IRQ;
+	mmc->max_segs = AU1XMMC_DESCRIPTOR_COUNT;
+
+	iflag = IRQF_SHARED;	/* Au1100/Au1200: one int for both ctrls */
+
+	switch (alchemy_get_cputype()) {
+	case ALCHEMY_CPU_AU1100:
+		mmc->max_seg_size = AU1100_MMC_DESCRIPTOR_SIZE;
+		break;
+	case ALCHEMY_CPU_AU1200:
+		mmc->max_seg_size = AU1200_MMC_DESCRIPTOR_SIZE;
+		break;
+	case ALCHEMY_CPU_AU1300:
+		iflag = 0;	/* nothing is shared */
+		mmc->max_seg_size = AU1200_MMC_DESCRIPTOR_SIZE;
+		mmc->f_max = 52000000;
+		if (host->ioarea->start == AU1100_SD0_PHYS_ADDR)
+			mmc->caps |= MMC_CAP_8_BIT_DATA;
+		break;
+	}
+
+	ret = request_irq(host->irq, au1xmmc_irq, iflag, DRIVER_NAME, host);
+	if (ret) {
+		dev_err(&pdev->dev, "cannot grab IRQ\n");
+		goto out3;
+	}
+
+	host->clk = clk_get(&pdev->dev, ALCHEMY_PERIPH_CLK);
+	if (IS_ERR(host->clk)) {
+		dev_err(&pdev->dev, "cannot find clock\n");
+		ret = PTR_ERR(host->clk);
+		goto out_irq;
+	}
+
+	ret = clk_prepare_enable(host->clk);
+	if (ret) {
+		dev_err(&pdev->dev, "cannot enable clock\n");
+		goto out_clk;
+	}
+
+	host->status = HOST_S_IDLE;
+
+	/* board-specific carddetect setup, if any */
+	if (host->platdata && host->platdata->cd_setup) {
+		ret = host->platdata->cd_setup(mmc, 1);
+		if (ret) {
+			dev_warn(&pdev->dev, "board CD setup failed\n");
+			mmc->caps |= MMC_CAP_NEEDS_POLL;
+		}
+	} else
+		mmc->caps |= MMC_CAP_NEEDS_POLL;
+
+	/* platform may not be able to use all advertised caps */
+	if (host->platdata)
+		mmc->caps &= ~(host->platdata->mask_host_caps);
+
+	tasklet_init(&host->data_task, au1xmmc_tasklet_data,
+			(unsigned long)host);
+
+	tasklet_init(&host->finish_task, au1xmmc_tasklet_finish,
+			(unsigned long)host);
+
+	if (has_dbdma()) {
+		ret = au1xmmc_dbdma_init(host);
+		if (ret)
+			pr_info(DRIVER_NAME ": DBDMA init failed; using PIO\n");
+	}
+
+#ifdef CONFIG_LEDS_CLASS
+	if (host->platdata && host->platdata->led) {
+		struct led_classdev *led = host->platdata->led;
+		led->name = mmc_hostname(mmc);
+		led->brightness = LED_OFF;
+		led->default_trigger = mmc_hostname(mmc);
+		ret = led_classdev_register(mmc_dev(mmc), led);
+		if (ret)
+			goto out5;
+	}
+#endif
+
+	au1xmmc_reset_controller(host);
+
+	ret = mmc_add_host(mmc);
+	if (ret) {
+		dev_err(&pdev->dev, "cannot add mmc host\n");
+		goto out6;
+	}
+
+	platform_set_drvdata(pdev, host);
+
+	pr_info(DRIVER_NAME ": MMC Controller %d set up at %p"
+		" (mode=%s)\n", pdev->id, host->iobase,
+		host->flags & HOST_F_DMA ? "dma" : "pio");
+
+	return 0;	/* all ok */
+
+out6:
+#ifdef CONFIG_LEDS_CLASS
+	if (host->platdata && host->platdata->led)
+		led_classdev_unregister(host->platdata->led);
+out5:
+#endif
+	__raw_writel(0, HOST_ENABLE(host));
+	__raw_writel(0, HOST_CONFIG(host));
+	__raw_writel(0, HOST_CONFIG2(host));
+	wmb(); /* drain writebuffer */
+
+	if (host->flags & HOST_F_DBDMA)
+		au1xmmc_dbdma_shutdown(host);
+
+	tasklet_kill(&host->data_task);
+	tasklet_kill(&host->finish_task);
+
+	if (host->platdata && host->platdata->cd_setup &&
+	    !(mmc->caps & MMC_CAP_NEEDS_POLL))
+		host->platdata->cd_setup(mmc, 0);
+
+	clk_disable_unprepare(host->clk);
+out_clk:
+	clk_put(host->clk);
+out_irq:
+	free_irq(host->irq, host);
+out3:
+	iounmap((void *)host->iobase);
+out2:
+	release_resource(host->ioarea);
+	kfree(host->ioarea);
+out1:
+	mmc_free_host(mmc);
+out0:
+	return ret;
+}
+
+static int au1xmmc_remove(struct platform_device *pdev)
+{
+	struct au1xmmc_host *host = platform_get_drvdata(pdev);
+
+	if (host) {
+		mmc_remove_host(host->mmc);
+
+#ifdef CONFIG_LEDS_CLASS
+		if (host->platdata && host->platdata->led)
+			led_classdev_unregister(host->platdata->led);
+#endif
+
+		if (host->platdata && host->platdata->cd_setup &&
+		    !(host->mmc->caps & MMC_CAP_NEEDS_POLL))
+			host->platdata->cd_setup(host->mmc, 0);
+
+		__raw_writel(0, HOST_ENABLE(host));
+		__raw_writel(0, HOST_CONFIG(host));
+		__raw_writel(0, HOST_CONFIG2(host));
+		wmb(); /* drain writebuffer */
+
+		tasklet_kill(&host->data_task);
+		tasklet_kill(&host->finish_task);
+
+		if (host->flags & HOST_F_DBDMA)
+			au1xmmc_dbdma_shutdown(host);
+
+		au1xmmc_set_power(host, 0);
+
+		clk_disable_unprepare(host->clk);
+		clk_put(host->clk);
+
+		free_irq(host->irq, host);
+		iounmap((void *)host->iobase);
+		release_resource(host->ioarea);
+		kfree(host->ioarea);
+
+		mmc_free_host(host->mmc);
+	}
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static int au1xmmc_suspend(struct platform_device *pdev, pm_message_t state)
+{
+	struct au1xmmc_host *host = platform_get_drvdata(pdev);
+
+	__raw_writel(0, HOST_CONFIG2(host));
+	__raw_writel(0, HOST_CONFIG(host));
+	__raw_writel(0xffffffff, HOST_STATUS(host));
+	__raw_writel(0, HOST_ENABLE(host));
+	wmb(); /* drain writebuffer */
+
+	return 0;
+}
+
+static int au1xmmc_resume(struct platform_device *pdev)
+{
+	struct au1xmmc_host *host = platform_get_drvdata(pdev);
+
+	au1xmmc_reset_controller(host);
+
+	return 0;
+}
+#else
+#define au1xmmc_suspend NULL
+#define au1xmmc_resume NULL
+#endif
+
+static struct platform_driver au1xmmc_driver = {
+	.probe         = au1xmmc_probe,
+	.remove        = au1xmmc_remove,
+	.suspend       = au1xmmc_suspend,
+	.resume        = au1xmmc_resume,
+	.driver        = {
+		.name  = DRIVER_NAME,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+	},
+};
+
+static int __init au1xmmc_init(void)
+{
+	if (has_dbdma()) {
+		/* DSCR_CMD0_ALWAYS has a stride of 32 bits, we need a stride
+		* of 8 bits.  And since devices are shared, we need to create
+		* our own to avoid freaking out other devices.
+		*/
+		memid = au1xxx_ddma_add_device(&au1xmmc_mem_dbdev);
+		if (!memid)
+			pr_err("au1xmmc: cannot add memory dbdma\n");
+	}
+	return platform_driver_register(&au1xmmc_driver);
+}
+
+static void __exit au1xmmc_exit(void)
+{
+	if (has_dbdma() && memid)
+		au1xxx_ddma_del_device(memid);
+
+	platform_driver_unregister(&au1xmmc_driver);
+}
+
+module_init(au1xmmc_init);
+module_exit(au1xmmc_exit);
+
+MODULE_AUTHOR("Advanced Micro Devices, Inc");
+MODULE_DESCRIPTION("MMC/SD driver for the Alchemy Au1XXX");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:au1xxx-mmc");
diff --git a/drivers/mmc/host/bcm2835.c b/drivers/mmc/host/bcm2835.c
new file mode 100644
index 000000000..985079943
--- /dev/null
+++ b/drivers/mmc/host/bcm2835.c
@@ -0,0 +1,1486 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * bcm2835 sdhost driver.
+ *
+ * The 2835 has two SD controllers: The Arasan sdhci controller
+ * (supported by the iproc driver) and a custom sdhost controller
+ * (supported by this driver).
+ *
+ * The sdhci controller supports both sdcard and sdio.  The sdhost
+ * controller supports the sdcard only, but has better performance.
+ * Also note that the rpi3 has sdio wifi, so driving the sdcard with
+ * the sdhost controller allows to use the sdhci controller for wifi
+ * support.
+ *
+ * The configuration is done by devicetree via pin muxing.  Both
+ * SD controller are available on the same pins (2 pin groups = pin 22
+ * to 27 + pin 48 to 53).  So it's possible to use both SD controllers
+ * at the same time with different pin groups.
+ *
+ * Author:      Phil Elwell <phil@raspberrypi.org>
+ *              Copyright (C) 2015-2016 Raspberry Pi (Trading) Ltd.
+ *
+ * Based on
+ *  mmc-bcm2835.c by Gellert Weisz
+ * which is, in turn, based on
+ *  sdhci-bcm2708.c by Broadcom
+ *  sdhci-bcm2835.c by Stephen Warren and Oleksandr Tymoshenko
+ *  sdhci.c and sdhci-pci.c by Pierre Ossman
+ */
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/err.h>
+#include <linux/highmem.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/iopoll.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/platform_device.h>
+#include <linux/scatterlist.h>
+#include <linux/time.h>
+#include <linux/workqueue.h>
+
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sd.h>
+
+#define SDCMD  0x00 /* Command to SD card              - 16 R/W */
+#define SDARG  0x04 /* Argument to SD card             - 32 R/W */
+#define SDTOUT 0x08 /* Start value for timeout counter - 32 R/W */
+#define SDCDIV 0x0c /* Start value for clock divider   - 11 R/W */
+#define SDRSP0 0x10 /* SD card response (31:0)         - 32 R   */
+#define SDRSP1 0x14 /* SD card response (63:32)        - 32 R   */
+#define SDRSP2 0x18 /* SD card response (95:64)        - 32 R   */
+#define SDRSP3 0x1c /* SD card response (127:96)       - 32 R   */
+#define SDHSTS 0x20 /* SD host status                  - 11 R/W */
+#define SDVDD  0x30 /* SD card power control           -  1 R/W */
+#define SDEDM  0x34 /* Emergency Debug Mode            - 13 R/W */
+#define SDHCFG 0x38 /* Host configuration              -  2 R/W */
+#define SDHBCT 0x3c /* Host byte count (debug)         - 32 R/W */
+#define SDDATA 0x40 /* Data to/from SD card            - 32 R/W */
+#define SDHBLC 0x50 /* Host block count (SDIO/SDHC)    -  9 R/W */
+
+#define SDCMD_NEW_FLAG			0x8000
+#define SDCMD_FAIL_FLAG			0x4000
+#define SDCMD_BUSYWAIT			0x800
+#define SDCMD_NO_RESPONSE		0x400
+#define SDCMD_LONG_RESPONSE		0x200
+#define SDCMD_WRITE_CMD			0x80
+#define SDCMD_READ_CMD			0x40
+#define SDCMD_CMD_MASK			0x3f
+
+#define SDCDIV_MAX_CDIV			0x7ff
+
+#define SDHSTS_BUSY_IRPT		0x400
+#define SDHSTS_BLOCK_IRPT		0x200
+#define SDHSTS_SDIO_IRPT		0x100
+#define SDHSTS_REW_TIME_OUT		0x80
+#define SDHSTS_CMD_TIME_OUT		0x40
+#define SDHSTS_CRC16_ERROR		0x20
+#define SDHSTS_CRC7_ERROR		0x10
+#define SDHSTS_FIFO_ERROR		0x08
+/* Reserved */
+/* Reserved */
+#define SDHSTS_DATA_FLAG		0x01
+
+#define SDHSTS_TRANSFER_ERROR_MASK	(SDHSTS_CRC7_ERROR | \
+					 SDHSTS_CRC16_ERROR | \
+					 SDHSTS_REW_TIME_OUT | \
+					 SDHSTS_FIFO_ERROR)
+
+#define SDHSTS_ERROR_MASK		(SDHSTS_CMD_TIME_OUT | \
+					 SDHSTS_TRANSFER_ERROR_MASK)
+
+#define SDHCFG_BUSY_IRPT_EN	BIT(10)
+#define SDHCFG_BLOCK_IRPT_EN	BIT(8)
+#define SDHCFG_SDIO_IRPT_EN	BIT(5)
+#define SDHCFG_DATA_IRPT_EN	BIT(4)
+#define SDHCFG_SLOW_CARD	BIT(3)
+#define SDHCFG_WIDE_EXT_BUS	BIT(2)
+#define SDHCFG_WIDE_INT_BUS	BIT(1)
+#define SDHCFG_REL_CMD_LINE	BIT(0)
+
+#define SDVDD_POWER_OFF		0
+#define SDVDD_POWER_ON		1
+
+#define SDEDM_FORCE_DATA_MODE	BIT(19)
+#define SDEDM_CLOCK_PULSE	BIT(20)
+#define SDEDM_BYPASS		BIT(21)
+
+#define SDEDM_WRITE_THRESHOLD_SHIFT	9
+#define SDEDM_READ_THRESHOLD_SHIFT	14
+#define SDEDM_THRESHOLD_MASK		0x1f
+
+#define SDEDM_FSM_MASK		0xf
+#define SDEDM_FSM_IDENTMODE	0x0
+#define SDEDM_FSM_DATAMODE	0x1
+#define SDEDM_FSM_READDATA	0x2
+#define SDEDM_FSM_WRITEDATA	0x3
+#define SDEDM_FSM_READWAIT	0x4
+#define SDEDM_FSM_READCRC	0x5
+#define SDEDM_FSM_WRITECRC	0x6
+#define SDEDM_FSM_WRITEWAIT1	0x7
+#define SDEDM_FSM_POWERDOWN	0x8
+#define SDEDM_FSM_POWERUP	0x9
+#define SDEDM_FSM_WRITESTART1	0xa
+#define SDEDM_FSM_WRITESTART2	0xb
+#define SDEDM_FSM_GENPULSES	0xc
+#define SDEDM_FSM_WRITEWAIT2	0xd
+#define SDEDM_FSM_STARTPOWDOWN	0xf
+
+#define SDDATA_FIFO_WORDS	16
+
+#define FIFO_READ_THRESHOLD	4
+#define FIFO_WRITE_THRESHOLD	4
+#define SDDATA_FIFO_PIO_BURST	8
+
+#define PIO_THRESHOLD	1  /* Maximum block count for PIO (0 = always DMA) */
+
+struct bcm2835_host {
+	spinlock_t		lock;
+	struct mutex		mutex;
+
+	void __iomem		*ioaddr;
+	u32			phys_addr;
+
+	struct platform_device	*pdev;
+
+	int			clock;		/* Current clock speed */
+	unsigned int		max_clk;	/* Max possible freq */
+	struct work_struct	dma_work;
+	struct delayed_work	timeout_work;	/* Timer for timeouts */
+	struct sg_mapping_iter	sg_miter;	/* SG state for PIO */
+	unsigned int		blocks;		/* remaining PIO blocks */
+	int			irq;		/* Device IRQ */
+
+	u32			ns_per_fifo_word;
+
+	/* cached registers */
+	u32			hcfg;
+	u32			cdiv;
+
+	struct mmc_request	*mrq;		/* Current request */
+	struct mmc_command	*cmd;		/* Current command */
+	struct mmc_data		*data;		/* Current data request */
+	bool			data_complete:1;/* Data finished before cmd */
+	bool			use_busy:1;	/* Wait for busy interrupt */
+	bool			use_sbc:1;	/* Send CMD23 */
+
+	/* for threaded irq handler */
+	bool			irq_block;
+	bool			irq_busy;
+	bool			irq_data;
+
+	/* DMA part */
+	struct dma_chan		*dma_chan_rxtx;
+	struct dma_chan		*dma_chan;
+	struct dma_slave_config dma_cfg_rx;
+	struct dma_slave_config dma_cfg_tx;
+	struct dma_async_tx_descriptor	*dma_desc;
+	u32			dma_dir;
+	u32			drain_words;
+	struct page		*drain_page;
+	u32			drain_offset;
+	bool			use_dma;
+};
+
+static void bcm2835_dumpcmd(struct bcm2835_host *host, struct mmc_command *cmd,
+			    const char *label)
+{
+	struct device *dev = &host->pdev->dev;
+
+	if (!cmd)
+		return;
+
+	dev_dbg(dev, "%c%s op %d arg 0x%x flags 0x%x - resp %08x %08x %08x %08x, err %d\n",
+		(cmd == host->cmd) ? '>' : ' ',
+		label, cmd->opcode, cmd->arg, cmd->flags,
+		cmd->resp[0], cmd->resp[1], cmd->resp[2], cmd->resp[3],
+		cmd->error);
+}
+
+static void bcm2835_dumpregs(struct bcm2835_host *host)
+{
+	struct mmc_request *mrq = host->mrq;
+	struct device *dev = &host->pdev->dev;
+
+	if (mrq) {
+		bcm2835_dumpcmd(host, mrq->sbc, "sbc");
+		bcm2835_dumpcmd(host, mrq->cmd, "cmd");
+		if (mrq->data) {
+			dev_dbg(dev, "data blocks %x blksz %x - err %d\n",
+				mrq->data->blocks,
+				mrq->data->blksz,
+				mrq->data->error);
+		}
+		bcm2835_dumpcmd(host, mrq->stop, "stop");
+	}
+
+	dev_dbg(dev, "=========== REGISTER DUMP ===========\n");
+	dev_dbg(dev, "SDCMD  0x%08x\n", readl(host->ioaddr + SDCMD));
+	dev_dbg(dev, "SDARG  0x%08x\n", readl(host->ioaddr + SDARG));
+	dev_dbg(dev, "SDTOUT 0x%08x\n", readl(host->ioaddr + SDTOUT));
+	dev_dbg(dev, "SDCDIV 0x%08x\n", readl(host->ioaddr + SDCDIV));
+	dev_dbg(dev, "SDRSP0 0x%08x\n", readl(host->ioaddr + SDRSP0));
+	dev_dbg(dev, "SDRSP1 0x%08x\n", readl(host->ioaddr + SDRSP1));
+	dev_dbg(dev, "SDRSP2 0x%08x\n", readl(host->ioaddr + SDRSP2));
+	dev_dbg(dev, "SDRSP3 0x%08x\n", readl(host->ioaddr + SDRSP3));
+	dev_dbg(dev, "SDHSTS 0x%08x\n", readl(host->ioaddr + SDHSTS));
+	dev_dbg(dev, "SDVDD  0x%08x\n", readl(host->ioaddr + SDVDD));
+	dev_dbg(dev, "SDEDM  0x%08x\n", readl(host->ioaddr + SDEDM));
+	dev_dbg(dev, "SDHCFG 0x%08x\n", readl(host->ioaddr + SDHCFG));
+	dev_dbg(dev, "SDHBCT 0x%08x\n", readl(host->ioaddr + SDHBCT));
+	dev_dbg(dev, "SDHBLC 0x%08x\n", readl(host->ioaddr + SDHBLC));
+	dev_dbg(dev, "===========================================\n");
+}
+
+static void bcm2835_reset_internal(struct bcm2835_host *host)
+{
+	u32 temp;
+
+	writel(SDVDD_POWER_OFF, host->ioaddr + SDVDD);
+	writel(0, host->ioaddr + SDCMD);
+	writel(0, host->ioaddr + SDARG);
+	writel(0xf00000, host->ioaddr + SDTOUT);
+	writel(0, host->ioaddr + SDCDIV);
+	writel(0x7f8, host->ioaddr + SDHSTS); /* Write 1s to clear */
+	writel(0, host->ioaddr + SDHCFG);
+	writel(0, host->ioaddr + SDHBCT);
+	writel(0, host->ioaddr + SDHBLC);
+
+	/* Limit fifo usage due to silicon bug */
+	temp = readl(host->ioaddr + SDEDM);
+	temp &= ~((SDEDM_THRESHOLD_MASK << SDEDM_READ_THRESHOLD_SHIFT) |
+		  (SDEDM_THRESHOLD_MASK << SDEDM_WRITE_THRESHOLD_SHIFT));
+	temp |= (FIFO_READ_THRESHOLD << SDEDM_READ_THRESHOLD_SHIFT) |
+		(FIFO_WRITE_THRESHOLD << SDEDM_WRITE_THRESHOLD_SHIFT);
+	writel(temp, host->ioaddr + SDEDM);
+	msleep(20);
+	writel(SDVDD_POWER_ON, host->ioaddr + SDVDD);
+	msleep(20);
+	host->clock = 0;
+	writel(host->hcfg, host->ioaddr + SDHCFG);
+	writel(host->cdiv, host->ioaddr + SDCDIV);
+}
+
+static void bcm2835_reset(struct mmc_host *mmc)
+{
+	struct bcm2835_host *host = mmc_priv(mmc);
+
+	if (host->dma_chan)
+		dmaengine_terminate_sync(host->dma_chan);
+	host->dma_chan = NULL;
+	bcm2835_reset_internal(host);
+}
+
+static void bcm2835_finish_command(struct bcm2835_host *host);
+
+static void bcm2835_wait_transfer_complete(struct bcm2835_host *host)
+{
+	int timediff;
+	u32 alternate_idle;
+
+	alternate_idle = (host->mrq->data->flags & MMC_DATA_READ) ?
+		SDEDM_FSM_READWAIT : SDEDM_FSM_WRITESTART1;
+
+	timediff = 0;
+
+	while (1) {
+		u32 edm, fsm;
+
+		edm = readl(host->ioaddr + SDEDM);
+		fsm = edm & SDEDM_FSM_MASK;
+
+		if ((fsm == SDEDM_FSM_IDENTMODE) ||
+		    (fsm == SDEDM_FSM_DATAMODE))
+			break;
+		if (fsm == alternate_idle) {
+			writel(edm | SDEDM_FORCE_DATA_MODE,
+			       host->ioaddr + SDEDM);
+			break;
+		}
+
+		timediff++;
+		if (timediff == 100000) {
+			dev_err(&host->pdev->dev,
+				"wait_transfer_complete - still waiting after %d retries\n",
+				timediff);
+			bcm2835_dumpregs(host);
+			host->mrq->data->error = -ETIMEDOUT;
+			return;
+		}
+		cpu_relax();
+	}
+}
+
+static void bcm2835_dma_complete(void *param)
+{
+	struct bcm2835_host *host = param;
+
+	schedule_work(&host->dma_work);
+}
+
+static void bcm2835_transfer_block_pio(struct bcm2835_host *host, bool is_read)
+{
+	unsigned long flags;
+	size_t blksize;
+	unsigned long wait_max;
+
+	blksize = host->data->blksz;
+
+	wait_max = jiffies + msecs_to_jiffies(500);
+
+	local_irq_save(flags);
+
+	while (blksize) {
+		int copy_words;
+		u32 hsts = 0;
+		size_t len;
+		u32 *buf;
+
+		if (!sg_miter_next(&host->sg_miter)) {
+			host->data->error = -EINVAL;
+			break;
+		}
+
+		len = min(host->sg_miter.length, blksize);
+		if (len % 4) {
+			host->data->error = -EINVAL;
+			break;
+		}
+
+		blksize -= len;
+		host->sg_miter.consumed = len;
+
+		buf = (u32 *)host->sg_miter.addr;
+
+		copy_words = len / 4;
+
+		while (copy_words) {
+			int burst_words, words;
+			u32 edm;
+
+			burst_words = min(SDDATA_FIFO_PIO_BURST, copy_words);
+			edm = readl(host->ioaddr + SDEDM);
+			if (is_read)
+				words = ((edm >> 4) & 0x1f);
+			else
+				words = SDDATA_FIFO_WORDS - ((edm >> 4) & 0x1f);
+
+			if (words < burst_words) {
+				int fsm_state = (edm & SDEDM_FSM_MASK);
+				struct device *dev = &host->pdev->dev;
+
+				if ((is_read &&
+				     (fsm_state != SDEDM_FSM_READDATA &&
+				      fsm_state != SDEDM_FSM_READWAIT &&
+				      fsm_state != SDEDM_FSM_READCRC)) ||
+				    (!is_read &&
+				     (fsm_state != SDEDM_FSM_WRITEDATA &&
+				      fsm_state != SDEDM_FSM_WRITESTART1 &&
+				      fsm_state != SDEDM_FSM_WRITESTART2))) {
+					hsts = readl(host->ioaddr + SDHSTS);
+					dev_err(dev, "fsm %x, hsts %08x\n",
+						fsm_state, hsts);
+					if (hsts & SDHSTS_ERROR_MASK)
+						break;
+				}
+
+				if (time_after(jiffies, wait_max)) {
+					dev_err(dev, "PIO %s timeout - EDM %08x\n",
+						is_read ? "read" : "write",
+						edm);
+					hsts = SDHSTS_REW_TIME_OUT;
+					break;
+				}
+				ndelay((burst_words - words) *
+				       host->ns_per_fifo_word);
+				continue;
+			} else if (words > copy_words) {
+				words = copy_words;
+			}
+
+			copy_words -= words;
+
+			while (words) {
+				if (is_read)
+					*(buf++) = readl(host->ioaddr + SDDATA);
+				else
+					writel(*(buf++), host->ioaddr + SDDATA);
+				words--;
+			}
+		}
+
+		if (hsts & SDHSTS_ERROR_MASK)
+			break;
+	}
+
+	sg_miter_stop(&host->sg_miter);
+
+	local_irq_restore(flags);
+}
+
+static void bcm2835_transfer_pio(struct bcm2835_host *host)
+{
+	struct device *dev = &host->pdev->dev;
+	u32 sdhsts;
+	bool is_read;
+
+	is_read = (host->data->flags & MMC_DATA_READ) != 0;
+	bcm2835_transfer_block_pio(host, is_read);
+
+	sdhsts = readl(host->ioaddr + SDHSTS);
+	if (sdhsts & (SDHSTS_CRC16_ERROR |
+		      SDHSTS_CRC7_ERROR |
+		      SDHSTS_FIFO_ERROR)) {
+		dev_err(dev, "%s transfer error - HSTS %08x\n",
+			is_read ? "read" : "write", sdhsts);
+		host->data->error = -EILSEQ;
+	} else if ((sdhsts & (SDHSTS_CMD_TIME_OUT |
+			      SDHSTS_REW_TIME_OUT))) {
+		dev_err(dev, "%s timeout error - HSTS %08x\n",
+			is_read ? "read" : "write", sdhsts);
+		host->data->error = -ETIMEDOUT;
+	}
+}
+
+static
+void bcm2835_prepare_dma(struct bcm2835_host *host, struct mmc_data *data)
+{
+	int sg_len, dir_data, dir_slave;
+	struct dma_async_tx_descriptor *desc = NULL;
+	struct dma_chan *dma_chan;
+
+	dma_chan = host->dma_chan_rxtx;
+	if (data->flags & MMC_DATA_READ) {
+		dir_data = DMA_FROM_DEVICE;
+		dir_slave = DMA_DEV_TO_MEM;
+	} else {
+		dir_data = DMA_TO_DEVICE;
+		dir_slave = DMA_MEM_TO_DEV;
+	}
+
+	/* The block doesn't manage the FIFO DREQs properly for
+	 * multi-block transfers, so don't attempt to DMA the final
+	 * few words.  Unfortunately this requires the final sg entry
+	 * to be trimmed.  N.B. This code demands that the overspill
+	 * is contained in a single sg entry.
+	 */
+
+	host->drain_words = 0;
+	if ((data->blocks > 1) && (dir_data == DMA_FROM_DEVICE)) {
+		struct scatterlist *sg;
+		u32 len;
+		int i;
+
+		len = min((u32)(FIFO_READ_THRESHOLD - 1) * 4,
+			  (u32)data->blocks * data->blksz);
+
+		for_each_sg(data->sg, sg, data->sg_len, i) {
+			if (sg_is_last(sg)) {
+				WARN_ON(sg->length < len);
+				sg->length -= len;
+				host->drain_page = sg_page(sg);
+				host->drain_offset = sg->offset + sg->length;
+			}
+		}
+		host->drain_words = len / 4;
+	}
+
+	/* The parameters have already been validated, so this will not fail */
+	(void)dmaengine_slave_config(dma_chan,
+				     (dir_data == DMA_FROM_DEVICE) ?
+				     &host->dma_cfg_rx :
+				     &host->dma_cfg_tx);
+
+	sg_len = dma_map_sg(dma_chan->device->dev, data->sg, data->sg_len,
+			    dir_data);
+	if (!sg_len)
+		return;
+
+	desc = dmaengine_prep_slave_sg(dma_chan, data->sg, sg_len, dir_slave,
+				       DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+
+	if (!desc) {
+		dma_unmap_sg(dma_chan->device->dev, data->sg, sg_len, dir_data);
+		return;
+	}
+
+	desc->callback = bcm2835_dma_complete;
+	desc->callback_param = host;
+	host->dma_desc = desc;
+	host->dma_chan = dma_chan;
+	host->dma_dir = dir_data;
+}
+
+static void bcm2835_start_dma(struct bcm2835_host *host)
+{
+	dmaengine_submit(host->dma_desc);
+	dma_async_issue_pending(host->dma_chan);
+}
+
+static void bcm2835_set_transfer_irqs(struct bcm2835_host *host)
+{
+	u32 all_irqs = SDHCFG_DATA_IRPT_EN | SDHCFG_BLOCK_IRPT_EN |
+		SDHCFG_BUSY_IRPT_EN;
+
+	if (host->dma_desc) {
+		host->hcfg = (host->hcfg & ~all_irqs) |
+			SDHCFG_BUSY_IRPT_EN;
+	} else {
+		host->hcfg = (host->hcfg & ~all_irqs) |
+			SDHCFG_DATA_IRPT_EN |
+			SDHCFG_BUSY_IRPT_EN;
+	}
+
+	writel(host->hcfg, host->ioaddr + SDHCFG);
+}
+
+static
+void bcm2835_prepare_data(struct bcm2835_host *host, struct mmc_command *cmd)
+{
+	struct mmc_data *data = cmd->data;
+
+	WARN_ON(host->data);
+
+	host->data = data;
+	if (!data)
+		return;
+
+	host->data_complete = false;
+	host->data->bytes_xfered = 0;
+
+	if (!host->dma_desc) {
+		/* Use PIO */
+		int flags = SG_MITER_ATOMIC;
+
+		if (data->flags & MMC_DATA_READ)
+			flags |= SG_MITER_TO_SG;
+		else
+			flags |= SG_MITER_FROM_SG;
+		sg_miter_start(&host->sg_miter, data->sg, data->sg_len, flags);
+		host->blocks = data->blocks;
+	}
+
+	bcm2835_set_transfer_irqs(host);
+
+	writel(data->blksz, host->ioaddr + SDHBCT);
+	writel(data->blocks, host->ioaddr + SDHBLC);
+}
+
+static u32 bcm2835_read_wait_sdcmd(struct bcm2835_host *host, u32 max_ms)
+{
+	struct device *dev = &host->pdev->dev;
+	u32 value;
+	int ret;
+
+	ret = readl_poll_timeout(host->ioaddr + SDCMD, value,
+				 !(value & SDCMD_NEW_FLAG), 1, 10);
+	if (ret == -ETIMEDOUT)
+		/* if it takes a while make poll interval bigger */
+		ret = readl_poll_timeout(host->ioaddr + SDCMD, value,
+					 !(value & SDCMD_NEW_FLAG),
+					 10, max_ms * 1000);
+	if (ret == -ETIMEDOUT)
+		dev_err(dev, "%s: timeout (%d ms)\n", __func__, max_ms);
+
+	return value;
+}
+
+static void bcm2835_finish_request(struct bcm2835_host *host)
+{
+	struct dma_chan *terminate_chan = NULL;
+	struct mmc_request *mrq;
+
+	cancel_delayed_work(&host->timeout_work);
+
+	mrq = host->mrq;
+
+	host->mrq = NULL;
+	host->cmd = NULL;
+	host->data = NULL;
+
+	host->dma_desc = NULL;
+	terminate_chan = host->dma_chan;
+	host->dma_chan = NULL;
+
+	if (terminate_chan) {
+		int err = dmaengine_terminate_all(terminate_chan);
+
+		if (err)
+			dev_err(&host->pdev->dev,
+				"failed to terminate DMA (%d)\n", err);
+	}
+
+	mmc_request_done(mmc_from_priv(host), mrq);
+}
+
+static
+bool bcm2835_send_command(struct bcm2835_host *host, struct mmc_command *cmd)
+{
+	struct device *dev = &host->pdev->dev;
+	u32 sdcmd, sdhsts;
+	unsigned long timeout;
+
+	WARN_ON(host->cmd);
+
+	sdcmd = bcm2835_read_wait_sdcmd(host, 100);
+	if (sdcmd & SDCMD_NEW_FLAG) {
+		dev_err(dev, "previous command never completed.\n");
+		bcm2835_dumpregs(host);
+		cmd->error = -EILSEQ;
+		bcm2835_finish_request(host);
+		return false;
+	}
+
+	if (!cmd->data && cmd->busy_timeout > 9000)
+		timeout = DIV_ROUND_UP(cmd->busy_timeout, 1000) * HZ + HZ;
+	else
+		timeout = 10 * HZ;
+	schedule_delayed_work(&host->timeout_work, timeout);
+
+	host->cmd = cmd;
+
+	/* Clear any error flags */
+	sdhsts = readl(host->ioaddr + SDHSTS);
+	if (sdhsts & SDHSTS_ERROR_MASK)
+		writel(sdhsts, host->ioaddr + SDHSTS);
+
+	if ((cmd->flags & MMC_RSP_136) && (cmd->flags & MMC_RSP_BUSY)) {
+		dev_err(dev, "unsupported response type!\n");
+		cmd->error = -EINVAL;
+		bcm2835_finish_request(host);
+		return false;
+	}
+
+	bcm2835_prepare_data(host, cmd);
+
+	writel(cmd->arg, host->ioaddr + SDARG);
+
+	sdcmd = cmd->opcode & SDCMD_CMD_MASK;
+
+	host->use_busy = false;
+	if (!(cmd->flags & MMC_RSP_PRESENT)) {
+		sdcmd |= SDCMD_NO_RESPONSE;
+	} else {
+		if (cmd->flags & MMC_RSP_136)
+			sdcmd |= SDCMD_LONG_RESPONSE;
+		if (cmd->flags & MMC_RSP_BUSY) {
+			sdcmd |= SDCMD_BUSYWAIT;
+			host->use_busy = true;
+		}
+	}
+
+	if (cmd->data) {
+		if (cmd->data->flags & MMC_DATA_WRITE)
+			sdcmd |= SDCMD_WRITE_CMD;
+		if (cmd->data->flags & MMC_DATA_READ)
+			sdcmd |= SDCMD_READ_CMD;
+	}
+
+	writel(sdcmd | SDCMD_NEW_FLAG, host->ioaddr + SDCMD);
+
+	return true;
+}
+
+static void bcm2835_transfer_complete(struct bcm2835_host *host)
+{
+	struct mmc_data *data;
+
+	WARN_ON(!host->data_complete);
+
+	data = host->data;
+	host->data = NULL;
+
+	/* Need to send CMD12 if -
+	 * a) open-ended multiblock transfer (no CMD23)
+	 * b) error in multiblock transfer
+	 */
+	if (host->mrq->stop && (data->error || !host->use_sbc)) {
+		if (bcm2835_send_command(host, host->mrq->stop)) {
+			/* No busy, so poll for completion */
+			if (!host->use_busy)
+				bcm2835_finish_command(host);
+		}
+	} else {
+		bcm2835_wait_transfer_complete(host);
+		bcm2835_finish_request(host);
+	}
+}
+
+static void bcm2835_finish_data(struct bcm2835_host *host)
+{
+	struct device *dev = &host->pdev->dev;
+	struct mmc_data *data;
+
+	data = host->data;
+
+	host->hcfg &= ~(SDHCFG_DATA_IRPT_EN | SDHCFG_BLOCK_IRPT_EN);
+	writel(host->hcfg, host->ioaddr + SDHCFG);
+
+	data->bytes_xfered = data->error ? 0 : (data->blksz * data->blocks);
+
+	host->data_complete = true;
+
+	if (host->cmd) {
+		/* Data managed to finish before the
+		 * command completed. Make sure we do
+		 * things in the proper order.
+		 */
+		dev_dbg(dev, "Finished early - HSTS %08x\n",
+			readl(host->ioaddr + SDHSTS));
+	} else {
+		bcm2835_transfer_complete(host);
+	}
+}
+
+static void bcm2835_finish_command(struct bcm2835_host *host)
+{
+	struct device *dev = &host->pdev->dev;
+	struct mmc_command *cmd = host->cmd;
+	u32 sdcmd;
+
+	sdcmd = bcm2835_read_wait_sdcmd(host, 100);
+
+	/* Check for errors */
+	if (sdcmd & SDCMD_NEW_FLAG) {
+		dev_err(dev, "command never completed.\n");
+		bcm2835_dumpregs(host);
+		host->cmd->error = -EIO;
+		bcm2835_finish_request(host);
+		return;
+	} else if (sdcmd & SDCMD_FAIL_FLAG) {
+		u32 sdhsts = readl(host->ioaddr + SDHSTS);
+
+		/* Clear the errors */
+		writel(SDHSTS_ERROR_MASK, host->ioaddr + SDHSTS);
+
+		if (!(sdhsts & SDHSTS_CRC7_ERROR) ||
+		    (host->cmd->opcode != MMC_SEND_OP_COND)) {
+			u32 edm, fsm;
+
+			if (sdhsts & SDHSTS_CMD_TIME_OUT) {
+				host->cmd->error = -ETIMEDOUT;
+			} else {
+				dev_err(dev, "unexpected command %d error\n",
+					host->cmd->opcode);
+				bcm2835_dumpregs(host);
+				host->cmd->error = -EILSEQ;
+			}
+			edm = readl(host->ioaddr + SDEDM);
+			fsm = edm & SDEDM_FSM_MASK;
+			if (fsm == SDEDM_FSM_READWAIT ||
+			    fsm == SDEDM_FSM_WRITESTART1)
+				/* Kick the FSM out of its wait */
+				writel(edm | SDEDM_FORCE_DATA_MODE,
+				       host->ioaddr + SDEDM);
+			bcm2835_finish_request(host);
+			return;
+		}
+	}
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136) {
+			int i;
+
+			for (i = 0; i < 4; i++) {
+				cmd->resp[3 - i] =
+					readl(host->ioaddr + SDRSP0 + i * 4);
+			}
+		} else {
+			cmd->resp[0] = readl(host->ioaddr + SDRSP0);
+		}
+	}
+
+	if (cmd == host->mrq->sbc) {
+		/* Finished CMD23, now send actual command. */
+		host->cmd = NULL;
+		if (bcm2835_send_command(host, host->mrq->cmd)) {
+			if (host->data && host->dma_desc)
+				/* DMA transfer starts now, PIO starts
+				 * after irq
+				 */
+				bcm2835_start_dma(host);
+
+			if (!host->use_busy)
+				bcm2835_finish_command(host);
+		}
+	} else if (cmd == host->mrq->stop) {
+		/* Finished CMD12 */
+		bcm2835_finish_request(host);
+	} else {
+		/* Processed actual command. */
+		host->cmd = NULL;
+		if (!host->data)
+			bcm2835_finish_request(host);
+		else if (host->data_complete)
+			bcm2835_transfer_complete(host);
+	}
+}
+
+static void bcm2835_timeout(struct work_struct *work)
+{
+	struct delayed_work *d = to_delayed_work(work);
+	struct bcm2835_host *host =
+		container_of(d, struct bcm2835_host, timeout_work);
+	struct device *dev = &host->pdev->dev;
+
+	mutex_lock(&host->mutex);
+
+	if (host->mrq) {
+		dev_err(dev, "timeout waiting for hardware interrupt.\n");
+		bcm2835_dumpregs(host);
+
+		bcm2835_reset(mmc_from_priv(host));
+
+		if (host->data) {
+			host->data->error = -ETIMEDOUT;
+			bcm2835_finish_data(host);
+		} else {
+			if (host->cmd)
+				host->cmd->error = -ETIMEDOUT;
+			else
+				host->mrq->cmd->error = -ETIMEDOUT;
+
+			bcm2835_finish_request(host);
+		}
+	}
+
+	mutex_unlock(&host->mutex);
+}
+
+static bool bcm2835_check_cmd_error(struct bcm2835_host *host, u32 intmask)
+{
+	struct device *dev = &host->pdev->dev;
+
+	if (!(intmask & SDHSTS_ERROR_MASK))
+		return false;
+
+	if (!host->cmd)
+		return true;
+
+	dev_err(dev, "sdhost_busy_irq: intmask %08x\n", intmask);
+	if (intmask & SDHSTS_CRC7_ERROR) {
+		host->cmd->error = -EILSEQ;
+	} else if (intmask & (SDHSTS_CRC16_ERROR |
+			      SDHSTS_FIFO_ERROR)) {
+		if (host->mrq->data)
+			host->mrq->data->error = -EILSEQ;
+		else
+			host->cmd->error = -EILSEQ;
+	} else if (intmask & SDHSTS_REW_TIME_OUT) {
+		if (host->mrq->data)
+			host->mrq->data->error = -ETIMEDOUT;
+		else
+			host->cmd->error = -ETIMEDOUT;
+	} else if (intmask & SDHSTS_CMD_TIME_OUT) {
+		host->cmd->error = -ETIMEDOUT;
+	}
+	bcm2835_dumpregs(host);
+	return true;
+}
+
+static void bcm2835_check_data_error(struct bcm2835_host *host, u32 intmask)
+{
+	if (!host->data)
+		return;
+	if (intmask & (SDHSTS_CRC16_ERROR | SDHSTS_FIFO_ERROR))
+		host->data->error = -EILSEQ;
+	if (intmask & SDHSTS_REW_TIME_OUT)
+		host->data->error = -ETIMEDOUT;
+}
+
+static void bcm2835_busy_irq(struct bcm2835_host *host)
+{
+	if (WARN_ON(!host->cmd)) {
+		bcm2835_dumpregs(host);
+		return;
+	}
+
+	if (WARN_ON(!host->use_busy)) {
+		bcm2835_dumpregs(host);
+		return;
+	}
+	host->use_busy = false;
+
+	bcm2835_finish_command(host);
+}
+
+static void bcm2835_data_irq(struct bcm2835_host *host, u32 intmask)
+{
+	/* There are no dedicated data/space available interrupt
+	 * status bits, so it is necessary to use the single shared
+	 * data/space available FIFO status bits. It is therefore not
+	 * an error to get here when there is no data transfer in
+	 * progress.
+	 */
+	if (!host->data)
+		return;
+
+	bcm2835_check_data_error(host, intmask);
+	if (host->data->error)
+		goto finished;
+
+	if (host->data->flags & MMC_DATA_WRITE) {
+		/* Use the block interrupt for writes after the first block */
+		host->hcfg &= ~(SDHCFG_DATA_IRPT_EN);
+		host->hcfg |= SDHCFG_BLOCK_IRPT_EN;
+		writel(host->hcfg, host->ioaddr + SDHCFG);
+		bcm2835_transfer_pio(host);
+	} else {
+		bcm2835_transfer_pio(host);
+		host->blocks--;
+		if ((host->blocks == 0) || host->data->error)
+			goto finished;
+	}
+	return;
+
+finished:
+	host->hcfg &= ~(SDHCFG_DATA_IRPT_EN | SDHCFG_BLOCK_IRPT_EN);
+	writel(host->hcfg, host->ioaddr + SDHCFG);
+}
+
+static void bcm2835_data_threaded_irq(struct bcm2835_host *host)
+{
+	if (!host->data)
+		return;
+	if ((host->blocks == 0) || host->data->error)
+		bcm2835_finish_data(host);
+}
+
+static void bcm2835_block_irq(struct bcm2835_host *host)
+{
+	if (WARN_ON(!host->data)) {
+		bcm2835_dumpregs(host);
+		return;
+	}
+
+	if (!host->dma_desc) {
+		WARN_ON(!host->blocks);
+		if (host->data->error || (--host->blocks == 0))
+			bcm2835_finish_data(host);
+		else
+			bcm2835_transfer_pio(host);
+	} else if (host->data->flags & MMC_DATA_WRITE) {
+		bcm2835_finish_data(host);
+	}
+}
+
+static irqreturn_t bcm2835_irq(int irq, void *dev_id)
+{
+	irqreturn_t result = IRQ_NONE;
+	struct bcm2835_host *host = dev_id;
+	u32 intmask;
+
+	spin_lock(&host->lock);
+
+	intmask = readl(host->ioaddr + SDHSTS);
+
+	writel(SDHSTS_BUSY_IRPT |
+	       SDHSTS_BLOCK_IRPT |
+	       SDHSTS_SDIO_IRPT |
+	       SDHSTS_DATA_FLAG,
+	       host->ioaddr + SDHSTS);
+
+	if (intmask & SDHSTS_BLOCK_IRPT) {
+		bcm2835_check_data_error(host, intmask);
+		host->irq_block = true;
+		result = IRQ_WAKE_THREAD;
+	}
+
+	if (intmask & SDHSTS_BUSY_IRPT) {
+		if (!bcm2835_check_cmd_error(host, intmask)) {
+			host->irq_busy = true;
+			result = IRQ_WAKE_THREAD;
+		} else {
+			result = IRQ_HANDLED;
+		}
+	}
+
+	/* There is no true data interrupt status bit, so it is
+	 * necessary to qualify the data flag with the interrupt
+	 * enable bit.
+	 */
+	if ((intmask & SDHSTS_DATA_FLAG) &&
+	    (host->hcfg & SDHCFG_DATA_IRPT_EN)) {
+		bcm2835_data_irq(host, intmask);
+		host->irq_data = true;
+		result = IRQ_WAKE_THREAD;
+	}
+
+	spin_unlock(&host->lock);
+
+	return result;
+}
+
+static irqreturn_t bcm2835_threaded_irq(int irq, void *dev_id)
+{
+	struct bcm2835_host *host = dev_id;
+	unsigned long flags;
+	bool block, busy, data;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	block = host->irq_block;
+	busy  = host->irq_busy;
+	data  = host->irq_data;
+	host->irq_block = false;
+	host->irq_busy  = false;
+	host->irq_data  = false;
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	mutex_lock(&host->mutex);
+
+	if (block)
+		bcm2835_block_irq(host);
+	if (busy)
+		bcm2835_busy_irq(host);
+	if (data)
+		bcm2835_data_threaded_irq(host);
+
+	mutex_unlock(&host->mutex);
+
+	return IRQ_HANDLED;
+}
+
+static void bcm2835_dma_complete_work(struct work_struct *work)
+{
+	struct bcm2835_host *host =
+		container_of(work, struct bcm2835_host, dma_work);
+	struct mmc_data *data;
+
+	mutex_lock(&host->mutex);
+
+	data = host->data;
+
+	if (host->dma_chan) {
+		dma_unmap_sg(host->dma_chan->device->dev,
+			     data->sg, data->sg_len,
+			     host->dma_dir);
+
+		host->dma_chan = NULL;
+	}
+
+	if (host->drain_words) {
+		unsigned long flags;
+		void *page;
+		u32 *buf;
+
+		if (host->drain_offset & PAGE_MASK) {
+			host->drain_page += host->drain_offset >> PAGE_SHIFT;
+			host->drain_offset &= ~PAGE_MASK;
+		}
+		local_irq_save(flags);
+		page = kmap_atomic(host->drain_page);
+		buf = page + host->drain_offset;
+
+		while (host->drain_words) {
+			u32 edm = readl(host->ioaddr + SDEDM);
+
+			if ((edm >> 4) & 0x1f)
+				*(buf++) = readl(host->ioaddr + SDDATA);
+			host->drain_words--;
+		}
+
+		kunmap_atomic(page);
+		local_irq_restore(flags);
+	}
+
+	bcm2835_finish_data(host);
+
+	mutex_unlock(&host->mutex);
+}
+
+static void bcm2835_set_clock(struct bcm2835_host *host, unsigned int clock)
+{
+	struct mmc_host *mmc = mmc_from_priv(host);
+	int div;
+
+	/* The SDCDIV register has 11 bits, and holds (div - 2).  But
+	 * in data mode the max is 50MHz wihout a minimum, and only
+	 * the bottom 3 bits are used. Since the switch over is
+	 * automatic (unless we have marked the card as slow...),
+	 * chosen values have to make sense in both modes.  Ident mode
+	 * must be 100-400KHz, so can range check the requested
+	 * clock. CMD15 must be used to return to data mode, so this
+	 * can be monitored.
+	 *
+	 * clock 250MHz -> 0->125MHz, 1->83.3MHz, 2->62.5MHz, 3->50.0MHz
+	 *                 4->41.7MHz, 5->35.7MHz, 6->31.3MHz, 7->27.8MHz
+	 *
+	 *		 623->400KHz/27.8MHz
+	 *		 reset value (507)->491159/50MHz
+	 *
+	 * BUT, the 3-bit clock divisor in data mode is too small if
+	 * the core clock is higher than 250MHz, so instead use the
+	 * SLOW_CARD configuration bit to force the use of the ident
+	 * clock divisor at all times.
+	 */
+
+	if (clock < 100000) {
+		/* Can't stop the clock, but make it as slow as possible
+		 * to show willing
+		 */
+		host->cdiv = SDCDIV_MAX_CDIV;
+		writel(host->cdiv, host->ioaddr + SDCDIV);
+		return;
+	}
+
+	div = host->max_clk / clock;
+	if (div < 2)
+		div = 2;
+	if ((host->max_clk / div) > clock)
+		div++;
+	div -= 2;
+
+	if (div > SDCDIV_MAX_CDIV)
+		div = SDCDIV_MAX_CDIV;
+
+	clock = host->max_clk / (div + 2);
+	mmc->actual_clock = clock;
+
+	/* Calibrate some delays */
+
+	host->ns_per_fifo_word = (1000000000 / clock) *
+		((mmc->caps & MMC_CAP_4_BIT_DATA) ? 8 : 32);
+
+	host->cdiv = div;
+	writel(host->cdiv, host->ioaddr + SDCDIV);
+
+	/* Set the timeout to 500ms */
+	writel(mmc->actual_clock / 2, host->ioaddr + SDTOUT);
+}
+
+static void bcm2835_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct bcm2835_host *host = mmc_priv(mmc);
+	struct device *dev = &host->pdev->dev;
+	u32 edm, fsm;
+
+	/* Reset the error statuses in case this is a retry */
+	if (mrq->sbc)
+		mrq->sbc->error = 0;
+	if (mrq->cmd)
+		mrq->cmd->error = 0;
+	if (mrq->data)
+		mrq->data->error = 0;
+	if (mrq->stop)
+		mrq->stop->error = 0;
+
+	if (mrq->data && !is_power_of_2(mrq->data->blksz)) {
+		dev_err(dev, "unsupported block size (%d bytes)\n",
+			mrq->data->blksz);
+
+		if (mrq->cmd)
+			mrq->cmd->error = -EINVAL;
+
+		mmc_request_done(mmc, mrq);
+		return;
+	}
+
+	mutex_lock(&host->mutex);
+
+	WARN_ON(host->mrq);
+	host->mrq = mrq;
+
+	edm = readl(host->ioaddr + SDEDM);
+	fsm = edm & SDEDM_FSM_MASK;
+
+	if ((fsm != SDEDM_FSM_IDENTMODE) &&
+	    (fsm != SDEDM_FSM_DATAMODE)) {
+		dev_err(dev, "previous command (%d) not complete (EDM %08x)\n",
+			readl(host->ioaddr + SDCMD) & SDCMD_CMD_MASK,
+			edm);
+		bcm2835_dumpregs(host);
+
+		if (mrq->cmd)
+			mrq->cmd->error = -EILSEQ;
+
+		bcm2835_finish_request(host);
+		mutex_unlock(&host->mutex);
+		return;
+	}
+
+	if (host->use_dma && mrq->data && (mrq->data->blocks > PIO_THRESHOLD))
+		bcm2835_prepare_dma(host, mrq->data);
+
+	host->use_sbc = !!mrq->sbc && host->mrq->data &&
+			(host->mrq->data->flags & MMC_DATA_READ);
+	if (host->use_sbc) {
+		if (bcm2835_send_command(host, mrq->sbc)) {
+			if (!host->use_busy)
+				bcm2835_finish_command(host);
+		}
+	} else if (mrq->cmd && bcm2835_send_command(host, mrq->cmd)) {
+		if (host->data && host->dma_desc) {
+			/* DMA transfer starts now, PIO starts after irq */
+			bcm2835_start_dma(host);
+		}
+
+		if (!host->use_busy)
+			bcm2835_finish_command(host);
+	}
+
+	mutex_unlock(&host->mutex);
+}
+
+static void bcm2835_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct bcm2835_host *host = mmc_priv(mmc);
+
+	mutex_lock(&host->mutex);
+
+	if (!ios->clock || ios->clock != host->clock) {
+		bcm2835_set_clock(host, ios->clock);
+		host->clock = ios->clock;
+	}
+
+	/* set bus width */
+	host->hcfg &= ~SDHCFG_WIDE_EXT_BUS;
+	if (ios->bus_width == MMC_BUS_WIDTH_4)
+		host->hcfg |= SDHCFG_WIDE_EXT_BUS;
+
+	host->hcfg |= SDHCFG_WIDE_INT_BUS;
+
+	/* Disable clever clock switching, to cope with fast core clocks */
+	host->hcfg |= SDHCFG_SLOW_CARD;
+
+	writel(host->hcfg, host->ioaddr + SDHCFG);
+
+	mutex_unlock(&host->mutex);
+}
+
+static const struct mmc_host_ops bcm2835_ops = {
+	.request = bcm2835_request,
+	.set_ios = bcm2835_set_ios,
+	.hw_reset = bcm2835_reset,
+};
+
+static int bcm2835_add_host(struct bcm2835_host *host)
+{
+	struct mmc_host *mmc = mmc_from_priv(host);
+	struct device *dev = &host->pdev->dev;
+	char pio_limit_string[20];
+	int ret;
+
+	if (!mmc->f_max || mmc->f_max > host->max_clk)
+		mmc->f_max = host->max_clk;
+	mmc->f_min = host->max_clk / SDCDIV_MAX_CDIV;
+
+	mmc->max_busy_timeout = ~0 / (mmc->f_max / 1000);
+
+	dev_dbg(dev, "f_max %d, f_min %d, max_busy_timeout %d\n",
+		mmc->f_max, mmc->f_min, mmc->max_busy_timeout);
+
+	/* host controller capabilities */
+	mmc->caps |= MMC_CAP_SD_HIGHSPEED | MMC_CAP_MMC_HIGHSPEED |
+		     MMC_CAP_NEEDS_POLL | MMC_CAP_HW_RESET | MMC_CAP_CMD23;
+
+	spin_lock_init(&host->lock);
+	mutex_init(&host->mutex);
+
+	if (!host->dma_chan_rxtx) {
+		dev_warn(dev, "unable to initialise DMA channel. Falling back to PIO\n");
+		host->use_dma = false;
+	} else {
+		host->use_dma = true;
+
+		host->dma_cfg_tx.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+		host->dma_cfg_tx.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+		host->dma_cfg_tx.slave_id = 13;		/* DREQ channel */
+		host->dma_cfg_tx.direction = DMA_MEM_TO_DEV;
+		host->dma_cfg_tx.src_addr = 0;
+		host->dma_cfg_tx.dst_addr = host->phys_addr + SDDATA;
+
+		host->dma_cfg_rx.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+		host->dma_cfg_rx.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+		host->dma_cfg_rx.slave_id = 13;		/* DREQ channel */
+		host->dma_cfg_rx.direction = DMA_DEV_TO_MEM;
+		host->dma_cfg_rx.src_addr = host->phys_addr + SDDATA;
+		host->dma_cfg_rx.dst_addr = 0;
+
+		if (dmaengine_slave_config(host->dma_chan_rxtx,
+					   &host->dma_cfg_tx) != 0 ||
+		    dmaengine_slave_config(host->dma_chan_rxtx,
+					   &host->dma_cfg_rx) != 0)
+			host->use_dma = false;
+	}
+
+	mmc->max_segs = 128;
+	mmc->max_req_size = min_t(size_t, 524288, dma_max_mapping_size(dev));
+	mmc->max_seg_size = mmc->max_req_size;
+	mmc->max_blk_size = 1024;
+	mmc->max_blk_count =  65535;
+
+	/* report supported voltage ranges */
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+
+	INIT_WORK(&host->dma_work, bcm2835_dma_complete_work);
+	INIT_DELAYED_WORK(&host->timeout_work, bcm2835_timeout);
+
+	/* Set interrupt enables */
+	host->hcfg = SDHCFG_BUSY_IRPT_EN;
+
+	bcm2835_reset_internal(host);
+
+	ret = request_threaded_irq(host->irq, bcm2835_irq,
+				   bcm2835_threaded_irq,
+				   0, mmc_hostname(mmc), host);
+	if (ret) {
+		dev_err(dev, "failed to request IRQ %d: %d\n", host->irq, ret);
+		return ret;
+	}
+
+	ret = mmc_add_host(mmc);
+	if (ret) {
+		free_irq(host->irq, host);
+		return ret;
+	}
+
+	pio_limit_string[0] = '\0';
+	if (host->use_dma && (PIO_THRESHOLD > 0))
+		sprintf(pio_limit_string, " (>%d)", PIO_THRESHOLD);
+	dev_info(dev, "loaded - DMA %s%s\n",
+		 host->use_dma ? "enabled" : "disabled", pio_limit_string);
+
+	return 0;
+}
+
+static int bcm2835_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct clk *clk;
+	struct bcm2835_host *host;
+	struct mmc_host *mmc;
+	const __be32 *regaddr_p;
+	int ret;
+
+	dev_dbg(dev, "%s\n", __func__);
+	mmc = mmc_alloc_host(sizeof(*host), dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	mmc->ops = &bcm2835_ops;
+	host = mmc_priv(mmc);
+	host->pdev = pdev;
+	spin_lock_init(&host->lock);
+
+	host->ioaddr = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(host->ioaddr)) {
+		ret = PTR_ERR(host->ioaddr);
+		goto err;
+	}
+
+	/* Parse OF address directly to get the physical address for
+	 * DMA to our registers.
+	 */
+	regaddr_p = of_get_address(pdev->dev.of_node, 0, NULL, NULL);
+	if (!regaddr_p) {
+		dev_err(dev, "Can't get phys address\n");
+		ret = -EINVAL;
+		goto err;
+	}
+
+	host->phys_addr = be32_to_cpup(regaddr_p);
+
+	host->dma_chan = NULL;
+	host->dma_desc = NULL;
+
+	host->dma_chan_rxtx = dma_request_chan(dev, "rx-tx");
+	if (IS_ERR(host->dma_chan_rxtx)) {
+		ret = PTR_ERR(host->dma_chan_rxtx);
+		host->dma_chan_rxtx = NULL;
+
+		if (ret == -EPROBE_DEFER)
+			goto err;
+
+		/* Ignore errors to fall back to PIO mode */
+	}
+
+
+	clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(clk)) {
+		ret = dev_err_probe(dev, PTR_ERR(clk), "could not get clk\n");
+		goto err;
+	}
+
+	host->max_clk = clk_get_rate(clk);
+
+	host->irq = platform_get_irq(pdev, 0);
+	if (host->irq < 0) {
+		ret = host->irq;
+		goto err;
+	}
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto err;
+
+	ret = bcm2835_add_host(host);
+	if (ret)
+		goto err;
+
+	platform_set_drvdata(pdev, host);
+
+	dev_dbg(dev, "%s -> OK\n", __func__);
+
+	return 0;
+
+err:
+	dev_dbg(dev, "%s -> err %d\n", __func__, ret);
+	if (host->dma_chan_rxtx)
+		dma_release_channel(host->dma_chan_rxtx);
+	mmc_free_host(mmc);
+
+	return ret;
+}
+
+static int bcm2835_remove(struct platform_device *pdev)
+{
+	struct bcm2835_host *host = platform_get_drvdata(pdev);
+	struct mmc_host *mmc = mmc_from_priv(host);
+
+	mmc_remove_host(mmc);
+
+	writel(SDVDD_POWER_OFF, host->ioaddr + SDVDD);
+
+	free_irq(host->irq, host);
+
+	cancel_work_sync(&host->dma_work);
+	cancel_delayed_work_sync(&host->timeout_work);
+
+	if (host->dma_chan_rxtx)
+		dma_release_channel(host->dma_chan_rxtx);
+
+	mmc_free_host(mmc);
+
+	return 0;
+}
+
+static const struct of_device_id bcm2835_match[] = {
+	{ .compatible = "brcm,bcm2835-sdhost" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, bcm2835_match);
+
+static struct platform_driver bcm2835_driver = {
+	.probe      = bcm2835_probe,
+	.remove     = bcm2835_remove,
+	.driver     = {
+		.name		= "sdhost-bcm2835",
+		.probe_type	= PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table	= bcm2835_match,
+	},
+};
+module_platform_driver(bcm2835_driver);
+
+MODULE_ALIAS("platform:sdhost-bcm2835");
+MODULE_DESCRIPTION("BCM2835 SDHost driver");
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Phil Elwell");
diff --git a/drivers/mmc/host/cavium-octeon.c b/drivers/mmc/host/cavium-octeon.c
new file mode 100644
index 000000000..12dca91a8
--- /dev/null
+++ b/drivers/mmc/host/cavium-octeon.c
@@ -0,0 +1,340 @@
+/*
+ * Driver for MMC and SSD cards for Cavium OCTEON SOCs.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 2012-2017 Cavium Inc.
+ */
+#include <linux/dma-mapping.h>
+#include <linux/gpio/consumer.h>
+#include <linux/interrupt.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/module.h>
+#include <linux/of_platform.h>
+#include <asm/octeon/octeon.h>
+#include "cavium.h"
+
+#define CVMX_MIO_BOOT_CTL CVMX_ADD_IO_SEG(0x00011800000000D0ull)
+
+/*
+ * The l2c* functions below are used for the EMMC-17978 workaround.
+ *
+ * Due to a bug in the design of the MMC bus hardware, the 2nd to last
+ * cache block of a DMA read must be locked into the L2 Cache.
+ * Otherwise, data corruption may occur.
+ */
+static inline void *phys_to_ptr(u64 address)
+{
+	return (void *)(address | (1ull << 63)); /* XKPHYS */
+}
+
+/*
+ * Lock a single line into L2. The line is zeroed before locking
+ * to make sure no dram accesses are made.
+ */
+static void l2c_lock_line(u64 addr)
+{
+	char *addr_ptr = phys_to_ptr(addr);
+
+	asm volatile (
+		"cache 31, %[line]"	/* Unlock the line */
+		::[line] "m" (*addr_ptr));
+}
+
+/* Unlock a single line in the L2 cache. */
+static void l2c_unlock_line(u64 addr)
+{
+	char *addr_ptr = phys_to_ptr(addr);
+
+	asm volatile (
+		"cache 23, %[line]"	/* Unlock the line */
+		::[line] "m" (*addr_ptr));
+}
+
+/* Locks a memory region in the L2 cache. */
+static void l2c_lock_mem_region(u64 start, u64 len)
+{
+	u64 end;
+
+	/* Round start/end to cache line boundaries */
+	end = ALIGN(start + len - 1, CVMX_CACHE_LINE_SIZE);
+	start = ALIGN(start, CVMX_CACHE_LINE_SIZE);
+
+	while (start <= end) {
+		l2c_lock_line(start);
+		start += CVMX_CACHE_LINE_SIZE;
+	}
+	asm volatile("sync");
+}
+
+/* Unlock a memory region in the L2 cache. */
+static void l2c_unlock_mem_region(u64 start, u64 len)
+{
+	u64 end;
+
+	/* Round start/end to cache line boundaries */
+	end = ALIGN(start + len - 1, CVMX_CACHE_LINE_SIZE);
+	start = ALIGN(start, CVMX_CACHE_LINE_SIZE);
+
+	while (start <= end) {
+		l2c_unlock_line(start);
+		start += CVMX_CACHE_LINE_SIZE;
+	}
+}
+
+static void octeon_mmc_acquire_bus(struct cvm_mmc_host *host)
+{
+	if (!host->has_ciu3) {
+		down(&octeon_bootbus_sem);
+		/* For CN70XX, switch the MMC controller onto the bus. */
+		if (OCTEON_IS_MODEL(OCTEON_CN70XX))
+			writeq(0, (void __iomem *)CVMX_MIO_BOOT_CTL);
+	} else {
+		down(&host->mmc_serializer);
+	}
+}
+
+static void octeon_mmc_release_bus(struct cvm_mmc_host *host)
+{
+	if (!host->has_ciu3)
+		up(&octeon_bootbus_sem);
+	else
+		up(&host->mmc_serializer);
+}
+
+static void octeon_mmc_int_enable(struct cvm_mmc_host *host, u64 val)
+{
+	writeq(val, host->base + MIO_EMM_INT(host));
+	if (!host->has_ciu3)
+		writeq(val, host->base + MIO_EMM_INT_EN(host));
+}
+
+static void octeon_mmc_set_shared_power(struct cvm_mmc_host *host, int dir)
+{
+	if (dir == 0)
+		if (!atomic_dec_return(&host->shared_power_users))
+			gpiod_set_value_cansleep(host->global_pwr_gpiod, 0);
+	if (dir == 1)
+		if (atomic_inc_return(&host->shared_power_users) == 1)
+			gpiod_set_value_cansleep(host->global_pwr_gpiod, 1);
+}
+
+static void octeon_mmc_dmar_fixup(struct cvm_mmc_host *host,
+				  struct mmc_command *cmd,
+				  struct mmc_data *data,
+				  u64 addr)
+{
+	if (cmd->opcode != MMC_WRITE_MULTIPLE_BLOCK)
+		return;
+	if (data->blksz * data->blocks <= 1024)
+		return;
+
+	host->n_minus_one = addr + (data->blksz * data->blocks) - 1024;
+	l2c_lock_mem_region(host->n_minus_one, 512);
+}
+
+static void octeon_mmc_dmar_fixup_done(struct cvm_mmc_host *host)
+{
+	if (!host->n_minus_one)
+		return;
+	l2c_unlock_mem_region(host->n_minus_one, 512);
+	host->n_minus_one = 0;
+}
+
+static int octeon_mmc_probe(struct platform_device *pdev)
+{
+	struct device_node *cn, *node = pdev->dev.of_node;
+	struct cvm_mmc_host *host;
+	void __iomem *base;
+	int mmc_irq[9];
+	int i, ret = 0;
+	u64 val;
+
+	host = devm_kzalloc(&pdev->dev, sizeof(*host), GFP_KERNEL);
+	if (!host)
+		return -ENOMEM;
+
+	spin_lock_init(&host->irq_handler_lock);
+	sema_init(&host->mmc_serializer, 1);
+
+	host->dev = &pdev->dev;
+	host->acquire_bus = octeon_mmc_acquire_bus;
+	host->release_bus = octeon_mmc_release_bus;
+	host->int_enable = octeon_mmc_int_enable;
+	host->set_shared_power = octeon_mmc_set_shared_power;
+	if (OCTEON_IS_MODEL(OCTEON_CN6XXX) ||
+	    OCTEON_IS_MODEL(OCTEON_CNF7XXX)) {
+		host->dmar_fixup = octeon_mmc_dmar_fixup;
+		host->dmar_fixup_done = octeon_mmc_dmar_fixup_done;
+	}
+
+	host->sys_freq = octeon_get_io_clock_rate();
+
+	if (of_device_is_compatible(node, "cavium,octeon-7890-mmc")) {
+		host->big_dma_addr = true;
+		host->need_irq_handler_lock = true;
+		host->has_ciu3 = true;
+		host->use_sg = true;
+		/*
+		 * First seven are the EMM_INT bits 0..6, then two for
+		 * the EMM_DMA_INT bits
+		 */
+		for (i = 0; i < 9; i++) {
+			mmc_irq[i] = platform_get_irq(pdev, i);
+			if (mmc_irq[i] < 0)
+				return mmc_irq[i];
+
+			/* work around legacy u-boot device trees */
+			irq_set_irq_type(mmc_irq[i], IRQ_TYPE_EDGE_RISING);
+		}
+	} else {
+		host->big_dma_addr = false;
+		host->need_irq_handler_lock = false;
+		host->has_ciu3 = false;
+		/* First one is EMM second DMA */
+		for (i = 0; i < 2; i++) {
+			mmc_irq[i] = platform_get_irq(pdev, i);
+			if (mmc_irq[i] < 0)
+				return mmc_irq[i];
+		}
+	}
+
+	host->last_slot = -1;
+
+	base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(base))
+		return PTR_ERR(base);
+	host->base = base;
+	host->reg_off = 0;
+
+	base = devm_platform_ioremap_resource(pdev, 1);
+	if (IS_ERR(base))
+		return PTR_ERR(base);
+	host->dma_base = base;
+	/*
+	 * To keep the register addresses shared we intentionaly use
+	 * a negative offset here, first register used on Octeon therefore
+	 * starts at 0x20 (MIO_EMM_DMA_CFG).
+	 */
+	host->reg_off_dma = -0x20;
+
+	ret = dma_set_mask(&pdev->dev, DMA_BIT_MASK(64));
+	if (ret)
+		return ret;
+
+	/*
+	 * Clear out any pending interrupts that may be left over from
+	 * bootloader.
+	 */
+	val = readq(host->base + MIO_EMM_INT(host));
+	writeq(val, host->base + MIO_EMM_INT(host));
+
+	if (host->has_ciu3) {
+		/* Only CMD_DONE, DMA_DONE, CMD_ERR, DMA_ERR */
+		for (i = 1; i <= 4; i++) {
+			ret = devm_request_irq(&pdev->dev, mmc_irq[i],
+					       cvm_mmc_interrupt,
+					       0, cvm_mmc_irq_names[i], host);
+			if (ret < 0) {
+				dev_err(&pdev->dev, "Error: devm_request_irq %d\n",
+					mmc_irq[i]);
+				return ret;
+			}
+		}
+	} else {
+		ret = devm_request_irq(&pdev->dev, mmc_irq[0],
+				       cvm_mmc_interrupt, 0, KBUILD_MODNAME,
+				       host);
+		if (ret < 0) {
+			dev_err(&pdev->dev, "Error: devm_request_irq %d\n",
+				mmc_irq[0]);
+			return ret;
+		}
+	}
+
+	host->global_pwr_gpiod = devm_gpiod_get_optional(&pdev->dev,
+							 "power",
+							 GPIOD_OUT_HIGH);
+	if (IS_ERR(host->global_pwr_gpiod)) {
+		dev_err(&pdev->dev, "Invalid power GPIO\n");
+		return PTR_ERR(host->global_pwr_gpiod);
+	}
+
+	platform_set_drvdata(pdev, host);
+
+	i = 0;
+	for_each_child_of_node(node, cn) {
+		host->slot_pdev[i] =
+			of_platform_device_create(cn, NULL, &pdev->dev);
+		if (!host->slot_pdev[i]) {
+			i++;
+			continue;
+		}
+		ret = cvm_mmc_of_slot_probe(&host->slot_pdev[i]->dev, host);
+		if (ret) {
+			dev_err(&pdev->dev, "Error populating slots\n");
+			octeon_mmc_set_shared_power(host, 0);
+			of_node_put(cn);
+			goto error;
+		}
+		i++;
+	}
+	return 0;
+
+error:
+	for (i = 0; i < CAVIUM_MAX_MMC; i++) {
+		if (host->slot[i])
+			cvm_mmc_of_slot_remove(host->slot[i]);
+		if (host->slot_pdev[i])
+			of_platform_device_destroy(&host->slot_pdev[i]->dev, NULL);
+	}
+	return ret;
+}
+
+static int octeon_mmc_remove(struct platform_device *pdev)
+{
+	struct cvm_mmc_host *host = platform_get_drvdata(pdev);
+	u64 dma_cfg;
+	int i;
+
+	for (i = 0; i < CAVIUM_MAX_MMC; i++)
+		if (host->slot[i])
+			cvm_mmc_of_slot_remove(host->slot[i]);
+
+	dma_cfg = readq(host->dma_base + MIO_EMM_DMA_CFG(host));
+	dma_cfg &= ~MIO_EMM_DMA_CFG_EN;
+	writeq(dma_cfg, host->dma_base + MIO_EMM_DMA_CFG(host));
+
+	octeon_mmc_set_shared_power(host, 0);
+	return 0;
+}
+
+static const struct of_device_id octeon_mmc_match[] = {
+	{
+		.compatible = "cavium,octeon-6130-mmc",
+	},
+	{
+		.compatible = "cavium,octeon-7890-mmc",
+	},
+	{},
+};
+MODULE_DEVICE_TABLE(of, octeon_mmc_match);
+
+static struct platform_driver octeon_mmc_driver = {
+	.probe		= octeon_mmc_probe,
+	.remove		= octeon_mmc_remove,
+	.driver		= {
+		.name	= KBUILD_MODNAME,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = octeon_mmc_match,
+	},
+};
+
+module_platform_driver(octeon_mmc_driver);
+
+MODULE_AUTHOR("Cavium Inc. <support@cavium.com>");
+MODULE_DESCRIPTION("Low-level driver for Cavium OCTEON MMC/SSD card");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mmc/host/cavium-thunderx.c b/drivers/mmc/host/cavium-thunderx.c
new file mode 100644
index 000000000..202b1d6da
--- /dev/null
+++ b/drivers/mmc/host/cavium-thunderx.c
@@ -0,0 +1,205 @@
+/*
+ * Driver for MMC and SSD cards for Cavium ThunderX SOCs.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 2016 Cavium Inc.
+ */
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/interrupt.h>
+#include <linux/mmc/mmc.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_platform.h>
+#include <linux/pci.h>
+#include "cavium.h"
+
+static void thunder_mmc_acquire_bus(struct cvm_mmc_host *host)
+{
+	down(&host->mmc_serializer);
+}
+
+static void thunder_mmc_release_bus(struct cvm_mmc_host *host)
+{
+	up(&host->mmc_serializer);
+}
+
+static void thunder_mmc_int_enable(struct cvm_mmc_host *host, u64 val)
+{
+	writeq(val, host->base + MIO_EMM_INT(host));
+	writeq(val, host->base + MIO_EMM_INT_EN_SET(host));
+}
+
+static int thunder_mmc_register_interrupts(struct cvm_mmc_host *host,
+					   struct pci_dev *pdev)
+{
+	int nvec, ret, i;
+
+	nvec = pci_alloc_irq_vectors(pdev, 1, 9, PCI_IRQ_MSIX);
+	if (nvec < 0)
+		return nvec;
+
+	/* register interrupts */
+	for (i = 0; i < nvec; i++) {
+		ret = devm_request_irq(&pdev->dev, pci_irq_vector(pdev, i),
+				       cvm_mmc_interrupt,
+				       0, cvm_mmc_irq_names[i], host);
+		if (ret)
+			return ret;
+	}
+	return 0;
+}
+
+static int thunder_mmc_probe(struct pci_dev *pdev,
+			     const struct pci_device_id *id)
+{
+	struct device_node *node = pdev->dev.of_node;
+	struct device *dev = &pdev->dev;
+	struct device_node *child_node;
+	struct cvm_mmc_host *host;
+	int ret, i = 0;
+
+	host = devm_kzalloc(dev, sizeof(*host), GFP_KERNEL);
+	if (!host)
+		return -ENOMEM;
+
+	pci_set_drvdata(pdev, host);
+	ret = pcim_enable_device(pdev);
+	if (ret)
+		return ret;
+
+	ret = pci_request_regions(pdev, KBUILD_MODNAME);
+	if (ret)
+		return ret;
+
+	host->base = pcim_iomap(pdev, 0, pci_resource_len(pdev, 0));
+	if (!host->base) {
+		ret = -EINVAL;
+		goto error;
+	}
+
+	/* On ThunderX these are identical */
+	host->dma_base = host->base;
+
+	host->reg_off = 0x2000;
+	host->reg_off_dma = 0x160;
+
+	host->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(host->clk)) {
+		ret = PTR_ERR(host->clk);
+		goto error;
+	}
+
+	ret = clk_prepare_enable(host->clk);
+	if (ret)
+		goto error;
+	host->sys_freq = clk_get_rate(host->clk);
+
+	spin_lock_init(&host->irq_handler_lock);
+	sema_init(&host->mmc_serializer, 1);
+
+	host->dev = dev;
+	host->acquire_bus = thunder_mmc_acquire_bus;
+	host->release_bus = thunder_mmc_release_bus;
+	host->int_enable = thunder_mmc_int_enable;
+
+	host->use_sg = true;
+	host->big_dma_addr = true;
+	host->need_irq_handler_lock = true;
+	host->last_slot = -1;
+
+	ret = dma_set_mask(dev, DMA_BIT_MASK(48));
+	if (ret)
+		goto error;
+
+	/*
+	 * Clear out any pending interrupts that may be left over from
+	 * bootloader. Writing 1 to the bits clears them.
+	 */
+	writeq(127, host->base + MIO_EMM_INT_EN(host));
+	writeq(3, host->base + MIO_EMM_DMA_INT_ENA_W1C(host));
+	/* Clear DMA FIFO */
+	writeq(BIT_ULL(16), host->base + MIO_EMM_DMA_FIFO_CFG(host));
+
+	ret = thunder_mmc_register_interrupts(host, pdev);
+	if (ret)
+		goto error;
+
+	for_each_child_of_node(node, child_node) {
+		/*
+		 * mmc_of_parse and devm* require one device per slot.
+		 * Create a dummy device per slot and set the node pointer to
+		 * the slot. The easiest way to get this is using
+		 * of_platform_device_create.
+		 */
+		if (of_device_is_compatible(child_node, "mmc-slot")) {
+			host->slot_pdev[i] = of_platform_device_create(child_node, NULL,
+								       &pdev->dev);
+			if (!host->slot_pdev[i])
+				continue;
+
+			ret = cvm_mmc_of_slot_probe(&host->slot_pdev[i]->dev, host);
+			if (ret) {
+				of_node_put(child_node);
+				goto error;
+			}
+		}
+		i++;
+	}
+	dev_info(dev, "probed\n");
+	return 0;
+
+error:
+	for (i = 0; i < CAVIUM_MAX_MMC; i++) {
+		if (host->slot[i])
+			cvm_mmc_of_slot_remove(host->slot[i]);
+		if (host->slot_pdev[i]) {
+			get_device(&host->slot_pdev[i]->dev);
+			of_platform_device_destroy(&host->slot_pdev[i]->dev, NULL);
+			put_device(&host->slot_pdev[i]->dev);
+		}
+	}
+	clk_disable_unprepare(host->clk);
+	pci_release_regions(pdev);
+	return ret;
+}
+
+static void thunder_mmc_remove(struct pci_dev *pdev)
+{
+	struct cvm_mmc_host *host = pci_get_drvdata(pdev);
+	u64 dma_cfg;
+	int i;
+
+	for (i = 0; i < CAVIUM_MAX_MMC; i++)
+		if (host->slot[i])
+			cvm_mmc_of_slot_remove(host->slot[i]);
+
+	dma_cfg = readq(host->dma_base + MIO_EMM_DMA_CFG(host));
+	dma_cfg &= ~MIO_EMM_DMA_CFG_EN;
+	writeq(dma_cfg, host->dma_base + MIO_EMM_DMA_CFG(host));
+
+	clk_disable_unprepare(host->clk);
+	pci_release_regions(pdev);
+}
+
+static const struct pci_device_id thunder_mmc_id_table[] = {
+	{ PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, 0xa010) },
+	{ 0, }  /* end of table */
+};
+
+static struct pci_driver thunder_mmc_driver = {
+	.name = KBUILD_MODNAME,
+	.id_table = thunder_mmc_id_table,
+	.probe = thunder_mmc_probe,
+	.remove = thunder_mmc_remove,
+};
+
+module_pci_driver(thunder_mmc_driver);
+
+MODULE_AUTHOR("Cavium Inc.");
+MODULE_DESCRIPTION("Cavium ThunderX eMMC Driver");
+MODULE_LICENSE("GPL");
+MODULE_DEVICE_TABLE(pci, thunder_mmc_id_table);
diff --git a/drivers/mmc/host/cavium.c b/drivers/mmc/host/cavium.c
new file mode 100644
index 000000000..c5da3aaee
--- /dev/null
+++ b/drivers/mmc/host/cavium.c
@@ -0,0 +1,1086 @@
+/*
+ * Shared part of driver for MMC/SDHC controller on Cavium OCTEON and
+ * ThunderX SOCs.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 2012-2017 Cavium Inc.
+ * Authors:
+ *   David Daney <david.daney@cavium.com>
+ *   Peter Swain <pswain@cavium.com>
+ *   Steven J. Hill <steven.hill@cavium.com>
+ *   Jan Glauber <jglauber@cavium.com>
+ */
+#include <linux/bitfield.h>
+#include <linux/delay.h>
+#include <linux/dma-direction.h>
+#include <linux/dma-mapping.h>
+#include <linux/gpio/consumer.h>
+#include <linux/interrupt.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/module.h>
+#include <linux/regulator/consumer.h>
+#include <linux/scatterlist.h>
+#include <linux/time.h>
+
+#include "cavium.h"
+
+const char *cvm_mmc_irq_names[] = {
+	"MMC Buffer",
+	"MMC Command",
+	"MMC DMA",
+	"MMC Command Error",
+	"MMC DMA Error",
+	"MMC Switch",
+	"MMC Switch Error",
+	"MMC DMA int Fifo",
+	"MMC DMA int",
+};
+
+/*
+ * The Cavium MMC host hardware assumes that all commands have fixed
+ * command and response types.  These are correct if MMC devices are
+ * being used.  However, non-MMC devices like SD use command and
+ * response types that are unexpected by the host hardware.
+ *
+ * The command and response types can be overridden by supplying an
+ * XOR value that is applied to the type.  We calculate the XOR value
+ * from the values in this table and the flags passed from the MMC
+ * core.
+ */
+static struct cvm_mmc_cr_type cvm_mmc_cr_types[] = {
+	{0, 0},		/* CMD0 */
+	{0, 3},		/* CMD1 */
+	{0, 2},		/* CMD2 */
+	{0, 1},		/* CMD3 */
+	{0, 0},		/* CMD4 */
+	{0, 1},		/* CMD5 */
+	{0, 1},		/* CMD6 */
+	{0, 1},		/* CMD7 */
+	{1, 1},		/* CMD8 */
+	{0, 2},		/* CMD9 */
+	{0, 2},		/* CMD10 */
+	{1, 1},		/* CMD11 */
+	{0, 1},		/* CMD12 */
+	{0, 1},		/* CMD13 */
+	{1, 1},		/* CMD14 */
+	{0, 0},		/* CMD15 */
+	{0, 1},		/* CMD16 */
+	{1, 1},		/* CMD17 */
+	{1, 1},		/* CMD18 */
+	{3, 1},		/* CMD19 */
+	{2, 1},		/* CMD20 */
+	{0, 0},		/* CMD21 */
+	{0, 0},		/* CMD22 */
+	{0, 1},		/* CMD23 */
+	{2, 1},		/* CMD24 */
+	{2, 1},		/* CMD25 */
+	{2, 1},		/* CMD26 */
+	{2, 1},		/* CMD27 */
+	{0, 1},		/* CMD28 */
+	{0, 1},		/* CMD29 */
+	{1, 1},		/* CMD30 */
+	{1, 1},		/* CMD31 */
+	{0, 0},		/* CMD32 */
+	{0, 0},		/* CMD33 */
+	{0, 0},		/* CMD34 */
+	{0, 1},		/* CMD35 */
+	{0, 1},		/* CMD36 */
+	{0, 0},		/* CMD37 */
+	{0, 1},		/* CMD38 */
+	{0, 4},		/* CMD39 */
+	{0, 5},		/* CMD40 */
+	{0, 0},		/* CMD41 */
+	{2, 1},		/* CMD42 */
+	{0, 0},		/* CMD43 */
+	{0, 0},		/* CMD44 */
+	{0, 0},		/* CMD45 */
+	{0, 0},		/* CMD46 */
+	{0, 0},		/* CMD47 */
+	{0, 0},		/* CMD48 */
+	{0, 0},		/* CMD49 */
+	{0, 0},		/* CMD50 */
+	{0, 0},		/* CMD51 */
+	{0, 0},		/* CMD52 */
+	{0, 0},		/* CMD53 */
+	{0, 0},		/* CMD54 */
+	{0, 1},		/* CMD55 */
+	{0xff, 0xff},	/* CMD56 */
+	{0, 0},		/* CMD57 */
+	{0, 0},		/* CMD58 */
+	{0, 0},		/* CMD59 */
+	{0, 0},		/* CMD60 */
+	{0, 0},		/* CMD61 */
+	{0, 0},		/* CMD62 */
+	{0, 0}		/* CMD63 */
+};
+
+static struct cvm_mmc_cr_mods cvm_mmc_get_cr_mods(struct mmc_command *cmd)
+{
+	struct cvm_mmc_cr_type *cr;
+	u8 hardware_ctype, hardware_rtype;
+	u8 desired_ctype = 0, desired_rtype = 0;
+	struct cvm_mmc_cr_mods r;
+
+	cr = cvm_mmc_cr_types + (cmd->opcode & 0x3f);
+	hardware_ctype = cr->ctype;
+	hardware_rtype = cr->rtype;
+	if (cmd->opcode == MMC_GEN_CMD)
+		hardware_ctype = (cmd->arg & 1) ? 1 : 2;
+
+	switch (mmc_cmd_type(cmd)) {
+	case MMC_CMD_ADTC:
+		desired_ctype = (cmd->data->flags & MMC_DATA_WRITE) ? 2 : 1;
+		break;
+	case MMC_CMD_AC:
+	case MMC_CMD_BC:
+	case MMC_CMD_BCR:
+		desired_ctype = 0;
+		break;
+	}
+
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE:
+		desired_rtype = 0;
+		break;
+	case MMC_RSP_R1:/* MMC_RSP_R5, MMC_RSP_R6, MMC_RSP_R7 */
+	case MMC_RSP_R1B:
+		desired_rtype = 1;
+		break;
+	case MMC_RSP_R2:
+		desired_rtype = 2;
+		break;
+	case MMC_RSP_R3: /* MMC_RSP_R4 */
+		desired_rtype = 3;
+		break;
+	}
+	r.ctype_xor = desired_ctype ^ hardware_ctype;
+	r.rtype_xor = desired_rtype ^ hardware_rtype;
+	return r;
+}
+
+static void check_switch_errors(struct cvm_mmc_host *host)
+{
+	u64 emm_switch;
+
+	emm_switch = readq(host->base + MIO_EMM_SWITCH(host));
+	if (emm_switch & MIO_EMM_SWITCH_ERR0)
+		dev_err(host->dev, "Switch power class error\n");
+	if (emm_switch & MIO_EMM_SWITCH_ERR1)
+		dev_err(host->dev, "Switch hs timing error\n");
+	if (emm_switch & MIO_EMM_SWITCH_ERR2)
+		dev_err(host->dev, "Switch bus width error\n");
+}
+
+static void clear_bus_id(u64 *reg)
+{
+	u64 bus_id_mask = GENMASK_ULL(61, 60);
+
+	*reg &= ~bus_id_mask;
+}
+
+static void set_bus_id(u64 *reg, int bus_id)
+{
+	clear_bus_id(reg);
+	*reg |= FIELD_PREP(GENMASK(61, 60), bus_id);
+}
+
+static int get_bus_id(u64 reg)
+{
+	return FIELD_GET(GENMASK_ULL(61, 60), reg);
+}
+
+/*
+ * We never set the switch_exe bit since that would interfere
+ * with the commands send by the MMC core.
+ */
+static void do_switch(struct cvm_mmc_host *host, u64 emm_switch)
+{
+	int retries = 100;
+	u64 rsp_sts;
+	int bus_id;
+
+	/*
+	 * Modes setting only taken from slot 0. Work around that hardware
+	 * issue by first switching to slot 0.
+	 */
+	bus_id = get_bus_id(emm_switch);
+	clear_bus_id(&emm_switch);
+	writeq(emm_switch, host->base + MIO_EMM_SWITCH(host));
+
+	set_bus_id(&emm_switch, bus_id);
+	writeq(emm_switch, host->base + MIO_EMM_SWITCH(host));
+
+	/* wait for the switch to finish */
+	do {
+		rsp_sts = readq(host->base + MIO_EMM_RSP_STS(host));
+		if (!(rsp_sts & MIO_EMM_RSP_STS_SWITCH_VAL))
+			break;
+		udelay(10);
+	} while (--retries);
+
+	check_switch_errors(host);
+}
+
+static bool switch_val_changed(struct cvm_mmc_slot *slot, u64 new_val)
+{
+	/* Match BUS_ID, HS_TIMING, BUS_WIDTH, POWER_CLASS, CLK_HI, CLK_LO */
+	u64 match = 0x3001070fffffffffull;
+
+	return (slot->cached_switch & match) != (new_val & match);
+}
+
+static void set_wdog(struct cvm_mmc_slot *slot, unsigned int ns)
+{
+	u64 timeout;
+
+	if (!slot->clock)
+		return;
+
+	if (ns)
+		timeout = (slot->clock * ns) / NSEC_PER_SEC;
+	else
+		timeout = (slot->clock * 850ull) / 1000ull;
+	writeq(timeout, slot->host->base + MIO_EMM_WDOG(slot->host));
+}
+
+static void cvm_mmc_reset_bus(struct cvm_mmc_slot *slot)
+{
+	struct cvm_mmc_host *host = slot->host;
+	u64 emm_switch, wdog;
+
+	emm_switch = readq(slot->host->base + MIO_EMM_SWITCH(host));
+	emm_switch &= ~(MIO_EMM_SWITCH_EXE | MIO_EMM_SWITCH_ERR0 |
+			MIO_EMM_SWITCH_ERR1 | MIO_EMM_SWITCH_ERR2);
+	set_bus_id(&emm_switch, slot->bus_id);
+
+	wdog = readq(slot->host->base + MIO_EMM_WDOG(host));
+	do_switch(slot->host, emm_switch);
+
+	slot->cached_switch = emm_switch;
+
+	msleep(20);
+
+	writeq(wdog, slot->host->base + MIO_EMM_WDOG(host));
+}
+
+/* Switch to another slot if needed */
+static void cvm_mmc_switch_to(struct cvm_mmc_slot *slot)
+{
+	struct cvm_mmc_host *host = slot->host;
+	struct cvm_mmc_slot *old_slot;
+	u64 emm_sample, emm_switch;
+
+	if (slot->bus_id == host->last_slot)
+		return;
+
+	if (host->last_slot >= 0 && host->slot[host->last_slot]) {
+		old_slot = host->slot[host->last_slot];
+		old_slot->cached_switch = readq(host->base + MIO_EMM_SWITCH(host));
+		old_slot->cached_rca = readq(host->base + MIO_EMM_RCA(host));
+	}
+
+	writeq(slot->cached_rca, host->base + MIO_EMM_RCA(host));
+	emm_switch = slot->cached_switch;
+	set_bus_id(&emm_switch, slot->bus_id);
+	do_switch(host, emm_switch);
+
+	emm_sample = FIELD_PREP(MIO_EMM_SAMPLE_CMD_CNT, slot->cmd_cnt) |
+		     FIELD_PREP(MIO_EMM_SAMPLE_DAT_CNT, slot->dat_cnt);
+	writeq(emm_sample, host->base + MIO_EMM_SAMPLE(host));
+
+	host->last_slot = slot->bus_id;
+}
+
+static void do_read(struct cvm_mmc_host *host, struct mmc_request *req,
+		    u64 dbuf)
+{
+	struct sg_mapping_iter *smi = &host->smi;
+	int data_len = req->data->blocks * req->data->blksz;
+	int bytes_xfered, shift = -1;
+	u64 dat = 0;
+
+	/* Auto inc from offset zero */
+	writeq((0x10000 | (dbuf << 6)), host->base + MIO_EMM_BUF_IDX(host));
+
+	for (bytes_xfered = 0; bytes_xfered < data_len;) {
+		if (smi->consumed >= smi->length) {
+			if (!sg_miter_next(smi))
+				break;
+			smi->consumed = 0;
+		}
+
+		if (shift < 0) {
+			dat = readq(host->base + MIO_EMM_BUF_DAT(host));
+			shift = 56;
+		}
+
+		while (smi->consumed < smi->length && shift >= 0) {
+			((u8 *)smi->addr)[smi->consumed] = (dat >> shift) & 0xff;
+			bytes_xfered++;
+			smi->consumed++;
+			shift -= 8;
+		}
+	}
+
+	sg_miter_stop(smi);
+	req->data->bytes_xfered = bytes_xfered;
+	req->data->error = 0;
+}
+
+static void do_write(struct mmc_request *req)
+{
+	req->data->bytes_xfered = req->data->blocks * req->data->blksz;
+	req->data->error = 0;
+}
+
+static void set_cmd_response(struct cvm_mmc_host *host, struct mmc_request *req,
+			     u64 rsp_sts)
+{
+	u64 rsp_hi, rsp_lo;
+
+	if (!(rsp_sts & MIO_EMM_RSP_STS_RSP_VAL))
+		return;
+
+	rsp_lo = readq(host->base + MIO_EMM_RSP_LO(host));
+
+	switch (FIELD_GET(MIO_EMM_RSP_STS_RSP_TYPE, rsp_sts)) {
+	case 1:
+	case 3:
+		req->cmd->resp[0] = (rsp_lo >> 8) & 0xffffffff;
+		req->cmd->resp[1] = 0;
+		req->cmd->resp[2] = 0;
+		req->cmd->resp[3] = 0;
+		break;
+	case 2:
+		req->cmd->resp[3] = rsp_lo & 0xffffffff;
+		req->cmd->resp[2] = (rsp_lo >> 32) & 0xffffffff;
+		rsp_hi = readq(host->base + MIO_EMM_RSP_HI(host));
+		req->cmd->resp[1] = rsp_hi & 0xffffffff;
+		req->cmd->resp[0] = (rsp_hi >> 32) & 0xffffffff;
+		break;
+	}
+}
+
+static int get_dma_dir(struct mmc_data *data)
+{
+	return (data->flags & MMC_DATA_WRITE) ? DMA_TO_DEVICE : DMA_FROM_DEVICE;
+}
+
+static int finish_dma_single(struct cvm_mmc_host *host, struct mmc_data *data)
+{
+	data->bytes_xfered = data->blocks * data->blksz;
+	data->error = 0;
+	dma_unmap_sg(host->dev, data->sg, data->sg_len, get_dma_dir(data));
+	return 1;
+}
+
+static int finish_dma_sg(struct cvm_mmc_host *host, struct mmc_data *data)
+{
+	u64 fifo_cfg;
+	int count;
+
+	/* Check if there are any pending requests left */
+	fifo_cfg = readq(host->dma_base + MIO_EMM_DMA_FIFO_CFG(host));
+	count = FIELD_GET(MIO_EMM_DMA_FIFO_CFG_COUNT, fifo_cfg);
+	if (count)
+		dev_err(host->dev, "%u requests still pending\n", count);
+
+	data->bytes_xfered = data->blocks * data->blksz;
+	data->error = 0;
+
+	/* Clear and disable FIFO */
+	writeq(BIT_ULL(16), host->dma_base + MIO_EMM_DMA_FIFO_CFG(host));
+	dma_unmap_sg(host->dev, data->sg, data->sg_len, get_dma_dir(data));
+	return 1;
+}
+
+static int finish_dma(struct cvm_mmc_host *host, struct mmc_data *data)
+{
+	if (host->use_sg && data->sg_len > 1)
+		return finish_dma_sg(host, data);
+	else
+		return finish_dma_single(host, data);
+}
+
+static int check_status(u64 rsp_sts)
+{
+	if (rsp_sts & MIO_EMM_RSP_STS_RSP_BAD_STS ||
+	    rsp_sts & MIO_EMM_RSP_STS_RSP_CRC_ERR ||
+	    rsp_sts & MIO_EMM_RSP_STS_BLK_CRC_ERR)
+		return -EILSEQ;
+	if (rsp_sts & MIO_EMM_RSP_STS_RSP_TIMEOUT ||
+	    rsp_sts & MIO_EMM_RSP_STS_BLK_TIMEOUT)
+		return -ETIMEDOUT;
+	if (rsp_sts & MIO_EMM_RSP_STS_DBUF_ERR)
+		return -EIO;
+	return 0;
+}
+
+/* Try to clean up failed DMA. */
+static void cleanup_dma(struct cvm_mmc_host *host, u64 rsp_sts)
+{
+	u64 emm_dma;
+
+	emm_dma = readq(host->base + MIO_EMM_DMA(host));
+	emm_dma |= FIELD_PREP(MIO_EMM_DMA_VAL, 1) |
+		   FIELD_PREP(MIO_EMM_DMA_DAT_NULL, 1);
+	set_bus_id(&emm_dma, get_bus_id(rsp_sts));
+	writeq(emm_dma, host->base + MIO_EMM_DMA(host));
+}
+
+irqreturn_t cvm_mmc_interrupt(int irq, void *dev_id)
+{
+	struct cvm_mmc_host *host = dev_id;
+	struct mmc_request *req;
+	unsigned long flags = 0;
+	u64 emm_int, rsp_sts;
+	bool host_done;
+
+	if (host->need_irq_handler_lock)
+		spin_lock_irqsave(&host->irq_handler_lock, flags);
+	else
+		__acquire(&host->irq_handler_lock);
+
+	/* Clear interrupt bits (write 1 clears ). */
+	emm_int = readq(host->base + MIO_EMM_INT(host));
+	writeq(emm_int, host->base + MIO_EMM_INT(host));
+
+	if (emm_int & MIO_EMM_INT_SWITCH_ERR)
+		check_switch_errors(host);
+
+	req = host->current_req;
+	if (!req)
+		goto out;
+
+	rsp_sts = readq(host->base + MIO_EMM_RSP_STS(host));
+	/*
+	 * dma_val set means DMA is still in progress. Don't touch
+	 * the request and wait for the interrupt indicating that
+	 * the DMA is finished.
+	 */
+	if ((rsp_sts & MIO_EMM_RSP_STS_DMA_VAL) && host->dma_active)
+		goto out;
+
+	if (!host->dma_active && req->data &&
+	    (emm_int & MIO_EMM_INT_BUF_DONE)) {
+		unsigned int type = (rsp_sts >> 7) & 3;
+
+		if (type == 1)
+			do_read(host, req, rsp_sts & MIO_EMM_RSP_STS_DBUF);
+		else if (type == 2)
+			do_write(req);
+	}
+
+	host_done = emm_int & MIO_EMM_INT_CMD_DONE ||
+		    emm_int & MIO_EMM_INT_DMA_DONE ||
+		    emm_int & MIO_EMM_INT_CMD_ERR  ||
+		    emm_int & MIO_EMM_INT_DMA_ERR;
+
+	if (!(host_done && req->done))
+		goto no_req_done;
+
+	req->cmd->error = check_status(rsp_sts);
+
+	if (host->dma_active && req->data)
+		if (!finish_dma(host, req->data))
+			goto no_req_done;
+
+	set_cmd_response(host, req, rsp_sts);
+	if ((emm_int & MIO_EMM_INT_DMA_ERR) &&
+	    (rsp_sts & MIO_EMM_RSP_STS_DMA_PEND))
+		cleanup_dma(host, rsp_sts);
+
+	host->current_req = NULL;
+	req->done(req);
+
+no_req_done:
+	if (host->dmar_fixup_done)
+		host->dmar_fixup_done(host);
+	if (host_done)
+		host->release_bus(host);
+out:
+	if (host->need_irq_handler_lock)
+		spin_unlock_irqrestore(&host->irq_handler_lock, flags);
+	else
+		__release(&host->irq_handler_lock);
+	return IRQ_RETVAL(emm_int != 0);
+}
+
+/*
+ * Program DMA_CFG and if needed DMA_ADR.
+ * Returns 0 on error, DMA address otherwise.
+ */
+static u64 prepare_dma_single(struct cvm_mmc_host *host, struct mmc_data *data)
+{
+	u64 dma_cfg, addr;
+	int count, rw;
+
+	count = dma_map_sg(host->dev, data->sg, data->sg_len,
+			   get_dma_dir(data));
+	if (!count)
+		return 0;
+
+	rw = (data->flags & MMC_DATA_WRITE) ? 1 : 0;
+	dma_cfg = FIELD_PREP(MIO_EMM_DMA_CFG_EN, 1) |
+		  FIELD_PREP(MIO_EMM_DMA_CFG_RW, rw);
+#ifdef __LITTLE_ENDIAN
+	dma_cfg |= FIELD_PREP(MIO_EMM_DMA_CFG_ENDIAN, 1);
+#endif
+	dma_cfg |= FIELD_PREP(MIO_EMM_DMA_CFG_SIZE,
+			      (sg_dma_len(&data->sg[0]) / 8) - 1);
+
+	addr = sg_dma_address(&data->sg[0]);
+	if (!host->big_dma_addr)
+		dma_cfg |= FIELD_PREP(MIO_EMM_DMA_CFG_ADR, addr);
+	writeq(dma_cfg, host->dma_base + MIO_EMM_DMA_CFG(host));
+
+	pr_debug("[%s] sg_dma_len: %u  total sg_elem: %d\n",
+		 (rw) ? "W" : "R", sg_dma_len(&data->sg[0]), count);
+
+	if (host->big_dma_addr)
+		writeq(addr, host->dma_base + MIO_EMM_DMA_ADR(host));
+	return addr;
+}
+
+/*
+ * Queue complete sg list into the FIFO.
+ * Returns 0 on error, 1 otherwise.
+ */
+static u64 prepare_dma_sg(struct cvm_mmc_host *host, struct mmc_data *data)
+{
+	struct scatterlist *sg;
+	u64 fifo_cmd, addr;
+	int count, i, rw;
+
+	count = dma_map_sg(host->dev, data->sg, data->sg_len,
+			   get_dma_dir(data));
+	if (!count)
+		return 0;
+	if (count > 16)
+		goto error;
+
+	/* Enable FIFO by removing CLR bit */
+	writeq(0, host->dma_base + MIO_EMM_DMA_FIFO_CFG(host));
+
+	for_each_sg(data->sg, sg, count, i) {
+		/* Program DMA address */
+		addr = sg_dma_address(sg);
+		if (addr & 7)
+			goto error;
+		writeq(addr, host->dma_base + MIO_EMM_DMA_FIFO_ADR(host));
+
+		/*
+		 * If we have scatter-gather support we also have an extra
+		 * register for the DMA addr, so no need to check
+		 * host->big_dma_addr here.
+		 */
+		rw = (data->flags & MMC_DATA_WRITE) ? 1 : 0;
+		fifo_cmd = FIELD_PREP(MIO_EMM_DMA_FIFO_CMD_RW, rw);
+
+		/* enable interrupts on the last element */
+		fifo_cmd |= FIELD_PREP(MIO_EMM_DMA_FIFO_CMD_INTDIS,
+				       (i + 1 == count) ? 0 : 1);
+
+#ifdef __LITTLE_ENDIAN
+		fifo_cmd |= FIELD_PREP(MIO_EMM_DMA_FIFO_CMD_ENDIAN, 1);
+#endif
+		fifo_cmd |= FIELD_PREP(MIO_EMM_DMA_FIFO_CMD_SIZE,
+				       sg_dma_len(sg) / 8 - 1);
+		/*
+		 * The write copies the address and the command to the FIFO
+		 * and increments the FIFO's COUNT field.
+		 */
+		writeq(fifo_cmd, host->dma_base + MIO_EMM_DMA_FIFO_CMD(host));
+		pr_debug("[%s] sg_dma_len: %u  sg_elem: %d/%d\n",
+			 (rw) ? "W" : "R", sg_dma_len(sg), i, count);
+	}
+
+	/*
+	 * In difference to prepare_dma_single we don't return the
+	 * address here, as it would not make sense for scatter-gather.
+	 * The dma fixup is only required on models that don't support
+	 * scatter-gather, so that is not a problem.
+	 */
+	return 1;
+
+error:
+	WARN_ON_ONCE(1);
+	dma_unmap_sg(host->dev, data->sg, data->sg_len, get_dma_dir(data));
+	/* Disable FIFO */
+	writeq(BIT_ULL(16), host->dma_base + MIO_EMM_DMA_FIFO_CFG(host));
+	return 0;
+}
+
+static u64 prepare_dma(struct cvm_mmc_host *host, struct mmc_data *data)
+{
+	if (host->use_sg && data->sg_len > 1)
+		return prepare_dma_sg(host, data);
+	else
+		return prepare_dma_single(host, data);
+}
+
+static u64 prepare_ext_dma(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct cvm_mmc_slot *slot = mmc_priv(mmc);
+	u64 emm_dma;
+
+	emm_dma = FIELD_PREP(MIO_EMM_DMA_VAL, 1) |
+		  FIELD_PREP(MIO_EMM_DMA_SECTOR,
+			     mmc_card_is_blockaddr(mmc->card) ? 1 : 0) |
+		  FIELD_PREP(MIO_EMM_DMA_RW,
+			     (mrq->data->flags & MMC_DATA_WRITE) ? 1 : 0) |
+		  FIELD_PREP(MIO_EMM_DMA_BLOCK_CNT, mrq->data->blocks) |
+		  FIELD_PREP(MIO_EMM_DMA_CARD_ADDR, mrq->cmd->arg);
+	set_bus_id(&emm_dma, slot->bus_id);
+
+	if (mmc_card_mmc(mmc->card) || (mmc_card_sd(mmc->card) &&
+	    (mmc->card->scr.cmds & SD_SCR_CMD23_SUPPORT)))
+		emm_dma |= FIELD_PREP(MIO_EMM_DMA_MULTI, 1);
+
+	pr_debug("[%s] blocks: %u  multi: %d\n",
+		(emm_dma & MIO_EMM_DMA_RW) ? "W" : "R",
+		 mrq->data->blocks, (emm_dma & MIO_EMM_DMA_MULTI) ? 1 : 0);
+	return emm_dma;
+}
+
+static void cvm_mmc_dma_request(struct mmc_host *mmc,
+				struct mmc_request *mrq)
+{
+	struct cvm_mmc_slot *slot = mmc_priv(mmc);
+	struct cvm_mmc_host *host = slot->host;
+	struct mmc_data *data;
+	u64 emm_dma, addr;
+
+	if (!mrq->data || !mrq->data->sg || !mrq->data->sg_len ||
+	    !mrq->stop || mrq->stop->opcode != MMC_STOP_TRANSMISSION) {
+		dev_err(&mmc->card->dev,
+			"Error: cmv_mmc_dma_request no data\n");
+		goto error;
+	}
+
+	cvm_mmc_switch_to(slot);
+
+	data = mrq->data;
+	pr_debug("DMA request  blocks: %d  block_size: %d  total_size: %d\n",
+		 data->blocks, data->blksz, data->blocks * data->blksz);
+	if (data->timeout_ns)
+		set_wdog(slot, data->timeout_ns);
+
+	WARN_ON(host->current_req);
+	host->current_req = mrq;
+
+	emm_dma = prepare_ext_dma(mmc, mrq);
+	addr = prepare_dma(host, data);
+	if (!addr) {
+		dev_err(host->dev, "prepare_dma failed\n");
+		goto error;
+	}
+
+	host->dma_active = true;
+	host->int_enable(host, MIO_EMM_INT_CMD_ERR | MIO_EMM_INT_DMA_DONE |
+			 MIO_EMM_INT_DMA_ERR);
+
+	if (host->dmar_fixup)
+		host->dmar_fixup(host, mrq->cmd, data, addr);
+
+	/*
+	 * If we have a valid SD card in the slot, we set the response
+	 * bit mask to check for CRC errors and timeouts only.
+	 * Otherwise, use the default power reset value.
+	 */
+	if (mmc_card_sd(mmc->card))
+		writeq(0x00b00000ull, host->base + MIO_EMM_STS_MASK(host));
+	else
+		writeq(0xe4390080ull, host->base + MIO_EMM_STS_MASK(host));
+	writeq(emm_dma, host->base + MIO_EMM_DMA(host));
+	return;
+
+error:
+	mrq->cmd->error = -EINVAL;
+	if (mrq->done)
+		mrq->done(mrq);
+	host->release_bus(host);
+}
+
+static void do_read_request(struct cvm_mmc_host *host, struct mmc_request *mrq)
+{
+	sg_miter_start(&host->smi, mrq->data->sg, mrq->data->sg_len,
+		       SG_MITER_ATOMIC | SG_MITER_TO_SG);
+}
+
+static void do_write_request(struct cvm_mmc_host *host, struct mmc_request *mrq)
+{
+	unsigned int data_len = mrq->data->blocks * mrq->data->blksz;
+	struct sg_mapping_iter *smi = &host->smi;
+	unsigned int bytes_xfered;
+	int shift = 56;
+	u64 dat = 0;
+
+	/* Copy data to the xmit buffer before issuing the command. */
+	sg_miter_start(smi, mrq->data->sg, mrq->data->sg_len, SG_MITER_FROM_SG);
+
+	/* Auto inc from offset zero, dbuf zero */
+	writeq(0x10000ull, host->base + MIO_EMM_BUF_IDX(host));
+
+	for (bytes_xfered = 0; bytes_xfered < data_len;) {
+		if (smi->consumed >= smi->length) {
+			if (!sg_miter_next(smi))
+				break;
+			smi->consumed = 0;
+		}
+
+		while (smi->consumed < smi->length && shift >= 0) {
+			dat |= (u64)((u8 *)smi->addr)[smi->consumed] << shift;
+			bytes_xfered++;
+			smi->consumed++;
+			shift -= 8;
+		}
+
+		if (shift < 0) {
+			writeq(dat, host->base + MIO_EMM_BUF_DAT(host));
+			shift = 56;
+			dat = 0;
+		}
+	}
+	sg_miter_stop(smi);
+}
+
+static void cvm_mmc_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct cvm_mmc_slot *slot = mmc_priv(mmc);
+	struct cvm_mmc_host *host = slot->host;
+	struct mmc_command *cmd = mrq->cmd;
+	struct cvm_mmc_cr_mods mods;
+	u64 emm_cmd, rsp_sts;
+	int retries = 100;
+
+	/*
+	 * Note about locking:
+	 * All MMC devices share the same bus and controller. Allow only a
+	 * single user of the bootbus/MMC bus at a time. The lock is acquired
+	 * on all entry points from the MMC layer.
+	 *
+	 * For requests the lock is only released after the completion
+	 * interrupt!
+	 */
+	host->acquire_bus(host);
+
+	if (cmd->opcode == MMC_READ_MULTIPLE_BLOCK ||
+	    cmd->opcode == MMC_WRITE_MULTIPLE_BLOCK)
+		return cvm_mmc_dma_request(mmc, mrq);
+
+	cvm_mmc_switch_to(slot);
+
+	mods = cvm_mmc_get_cr_mods(cmd);
+
+	WARN_ON(host->current_req);
+	host->current_req = mrq;
+
+	if (cmd->data) {
+		if (cmd->data->flags & MMC_DATA_READ)
+			do_read_request(host, mrq);
+		else
+			do_write_request(host, mrq);
+
+		if (cmd->data->timeout_ns)
+			set_wdog(slot, cmd->data->timeout_ns);
+	} else
+		set_wdog(slot, 0);
+
+	host->dma_active = false;
+	host->int_enable(host, MIO_EMM_INT_CMD_DONE | MIO_EMM_INT_CMD_ERR);
+
+	emm_cmd = FIELD_PREP(MIO_EMM_CMD_VAL, 1) |
+		  FIELD_PREP(MIO_EMM_CMD_CTYPE_XOR, mods.ctype_xor) |
+		  FIELD_PREP(MIO_EMM_CMD_RTYPE_XOR, mods.rtype_xor) |
+		  FIELD_PREP(MIO_EMM_CMD_IDX, cmd->opcode) |
+		  FIELD_PREP(MIO_EMM_CMD_ARG, cmd->arg);
+	set_bus_id(&emm_cmd, slot->bus_id);
+	if (cmd->data && mmc_cmd_type(cmd) == MMC_CMD_ADTC)
+		emm_cmd |= FIELD_PREP(MIO_EMM_CMD_OFFSET,
+				64 - ((cmd->data->blocks * cmd->data->blksz) / 8));
+
+	writeq(0, host->base + MIO_EMM_STS_MASK(host));
+
+retry:
+	rsp_sts = readq(host->base + MIO_EMM_RSP_STS(host));
+	if (rsp_sts & MIO_EMM_RSP_STS_DMA_VAL ||
+	    rsp_sts & MIO_EMM_RSP_STS_CMD_VAL ||
+	    rsp_sts & MIO_EMM_RSP_STS_SWITCH_VAL ||
+	    rsp_sts & MIO_EMM_RSP_STS_DMA_PEND) {
+		udelay(10);
+		if (--retries)
+			goto retry;
+	}
+	if (!retries)
+		dev_err(host->dev, "Bad status: %llx before command write\n", rsp_sts);
+	writeq(emm_cmd, host->base + MIO_EMM_CMD(host));
+}
+
+static void cvm_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct cvm_mmc_slot *slot = mmc_priv(mmc);
+	struct cvm_mmc_host *host = slot->host;
+	int clk_period = 0, power_class = 10, bus_width = 0;
+	u64 clock, emm_switch;
+
+	host->acquire_bus(host);
+	cvm_mmc_switch_to(slot);
+
+	/* Set the power state */
+	switch (ios->power_mode) {
+	case MMC_POWER_ON:
+		break;
+
+	case MMC_POWER_OFF:
+		cvm_mmc_reset_bus(slot);
+		if (host->global_pwr_gpiod)
+			host->set_shared_power(host, 0);
+		else if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+		break;
+
+	case MMC_POWER_UP:
+		if (host->global_pwr_gpiod)
+			host->set_shared_power(host, 1);
+		else if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, ios->vdd);
+		break;
+	}
+
+	/* Convert bus width to HW definition */
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_8:
+		bus_width = 2;
+		break;
+	case MMC_BUS_WIDTH_4:
+		bus_width = 1;
+		break;
+	case MMC_BUS_WIDTH_1:
+		bus_width = 0;
+		break;
+	}
+
+	/* DDR is available for 4/8 bit bus width */
+	if (ios->bus_width && ios->timing == MMC_TIMING_MMC_DDR52)
+		bus_width |= 4;
+
+	/* Change the clock frequency. */
+	clock = ios->clock;
+	if (clock > 52000000)
+		clock = 52000000;
+	slot->clock = clock;
+
+	if (clock)
+		clk_period = (host->sys_freq + clock - 1) / (2 * clock);
+
+	emm_switch = FIELD_PREP(MIO_EMM_SWITCH_HS_TIMING,
+				(ios->timing == MMC_TIMING_MMC_HS)) |
+		     FIELD_PREP(MIO_EMM_SWITCH_BUS_WIDTH, bus_width) |
+		     FIELD_PREP(MIO_EMM_SWITCH_POWER_CLASS, power_class) |
+		     FIELD_PREP(MIO_EMM_SWITCH_CLK_HI, clk_period) |
+		     FIELD_PREP(MIO_EMM_SWITCH_CLK_LO, clk_period);
+	set_bus_id(&emm_switch, slot->bus_id);
+
+	if (!switch_val_changed(slot, emm_switch))
+		goto out;
+
+	set_wdog(slot, 0);
+	do_switch(host, emm_switch);
+	slot->cached_switch = emm_switch;
+out:
+	host->release_bus(host);
+}
+
+static const struct mmc_host_ops cvm_mmc_ops = {
+	.request        = cvm_mmc_request,
+	.set_ios        = cvm_mmc_set_ios,
+	.get_ro		= mmc_gpio_get_ro,
+	.get_cd		= mmc_gpio_get_cd,
+};
+
+static void cvm_mmc_set_clock(struct cvm_mmc_slot *slot, unsigned int clock)
+{
+	struct mmc_host *mmc = slot->mmc;
+
+	clock = min(clock, mmc->f_max);
+	clock = max(clock, mmc->f_min);
+	slot->clock = clock;
+}
+
+static int cvm_mmc_init_lowlevel(struct cvm_mmc_slot *slot)
+{
+	struct cvm_mmc_host *host = slot->host;
+	u64 emm_switch;
+
+	/* Enable this bus slot. */
+	host->emm_cfg |= (1ull << slot->bus_id);
+	writeq(host->emm_cfg, slot->host->base + MIO_EMM_CFG(host));
+	udelay(10);
+
+	/* Program initial clock speed and power. */
+	cvm_mmc_set_clock(slot, slot->mmc->f_min);
+	emm_switch = FIELD_PREP(MIO_EMM_SWITCH_POWER_CLASS, 10);
+	emm_switch |= FIELD_PREP(MIO_EMM_SWITCH_CLK_HI,
+				 (host->sys_freq / slot->clock) / 2);
+	emm_switch |= FIELD_PREP(MIO_EMM_SWITCH_CLK_LO,
+				 (host->sys_freq / slot->clock) / 2);
+
+	/* Make the changes take effect on this bus slot. */
+	set_bus_id(&emm_switch, slot->bus_id);
+	do_switch(host, emm_switch);
+
+	slot->cached_switch = emm_switch;
+
+	/*
+	 * Set watchdog timeout value and default reset value
+	 * for the mask register. Finally, set the CARD_RCA
+	 * bit so that we can get the card address relative
+	 * to the CMD register for CMD7 transactions.
+	 */
+	set_wdog(slot, 0);
+	writeq(0xe4390080ull, host->base + MIO_EMM_STS_MASK(host));
+	writeq(1, host->base + MIO_EMM_RCA(host));
+	return 0;
+}
+
+static int cvm_mmc_of_parse(struct device *dev, struct cvm_mmc_slot *slot)
+{
+	u32 id, cmd_skew = 0, dat_skew = 0, bus_width = 0;
+	struct device_node *node = dev->of_node;
+	struct mmc_host *mmc = slot->mmc;
+	u64 clock_period;
+	int ret;
+
+	ret = of_property_read_u32(node, "reg", &id);
+	if (ret) {
+		dev_err(dev, "Missing or invalid reg property on %pOF\n", node);
+		return ret;
+	}
+
+	if (id >= CAVIUM_MAX_MMC || slot->host->slot[id]) {
+		dev_err(dev, "Invalid reg property on %pOF\n", node);
+		return -EINVAL;
+	}
+
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret)
+		return ret;
+	/*
+	 * Legacy Octeon firmware has no regulator entry, fall-back to
+	 * a hard-coded voltage to get a sane OCR.
+	 */
+	if (IS_ERR(mmc->supply.vmmc))
+		mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+
+	/* Common MMC bindings */
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		return ret;
+
+	/* Set bus width */
+	if (!(mmc->caps & (MMC_CAP_8_BIT_DATA | MMC_CAP_4_BIT_DATA))) {
+		of_property_read_u32(node, "cavium,bus-max-width", &bus_width);
+		if (bus_width == 8)
+			mmc->caps |= MMC_CAP_8_BIT_DATA | MMC_CAP_4_BIT_DATA;
+		else if (bus_width == 4)
+			mmc->caps |= MMC_CAP_4_BIT_DATA;
+	}
+
+	/* Set maximum and minimum frequency */
+	if (!mmc->f_max)
+		of_property_read_u32(node, "spi-max-frequency", &mmc->f_max);
+	if (!mmc->f_max || mmc->f_max > 52000000)
+		mmc->f_max = 52000000;
+	mmc->f_min = 400000;
+
+	/* Sampling register settings, period in picoseconds */
+	clock_period = 1000000000000ull / slot->host->sys_freq;
+	of_property_read_u32(node, "cavium,cmd-clk-skew", &cmd_skew);
+	of_property_read_u32(node, "cavium,dat-clk-skew", &dat_skew);
+	slot->cmd_cnt = (cmd_skew + clock_period / 2) / clock_period;
+	slot->dat_cnt = (dat_skew + clock_period / 2) / clock_period;
+
+	return id;
+}
+
+int cvm_mmc_of_slot_probe(struct device *dev, struct cvm_mmc_host *host)
+{
+	struct cvm_mmc_slot *slot;
+	struct mmc_host *mmc;
+	int ret, id;
+
+	mmc = mmc_alloc_host(sizeof(struct cvm_mmc_slot), dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	slot = mmc_priv(mmc);
+	slot->mmc = mmc;
+	slot->host = host;
+
+	ret = cvm_mmc_of_parse(dev, slot);
+	if (ret < 0)
+		goto error;
+	id = ret;
+
+	/* Set up host parameters */
+	mmc->ops = &cvm_mmc_ops;
+
+	/*
+	 * We only have a 3.3v supply, we cannot support any
+	 * of the UHS modes. We do support the high speed DDR
+	 * modes up to 52MHz.
+	 *
+	 * Disable bounce buffers for max_segs = 1
+	 */
+	mmc->caps |= MMC_CAP_MMC_HIGHSPEED | MMC_CAP_SD_HIGHSPEED |
+		     MMC_CAP_CMD23 | MMC_CAP_POWER_OFF_CARD | MMC_CAP_3_3V_DDR;
+
+	if (host->use_sg)
+		mmc->max_segs = 16;
+	else
+		mmc->max_segs = 1;
+
+	/* DMA size field can address up to 8 MB */
+	mmc->max_seg_size = min_t(unsigned int, 8 * 1024 * 1024,
+				  dma_get_max_seg_size(host->dev));
+	mmc->max_req_size = mmc->max_seg_size;
+	/* External DMA is in 512 byte blocks */
+	mmc->max_blk_size = 512;
+	/* DMA block count field is 15 bits */
+	mmc->max_blk_count = 32767;
+
+	slot->clock = mmc->f_min;
+	slot->bus_id = id;
+	slot->cached_rca = 1;
+
+	host->acquire_bus(host);
+	host->slot[id] = slot;
+	cvm_mmc_switch_to(slot);
+	cvm_mmc_init_lowlevel(slot);
+	host->release_bus(host);
+
+	ret = mmc_add_host(mmc);
+	if (ret) {
+		dev_err(dev, "mmc_add_host() returned %d\n", ret);
+		slot->host->slot[id] = NULL;
+		goto error;
+	}
+	return 0;
+
+error:
+	mmc_free_host(slot->mmc);
+	return ret;
+}
+
+int cvm_mmc_of_slot_remove(struct cvm_mmc_slot *slot)
+{
+	mmc_remove_host(slot->mmc);
+	slot->host->slot[slot->bus_id] = NULL;
+	mmc_free_host(slot->mmc);
+	return 0;
+}
diff --git a/drivers/mmc/host/cavium.h b/drivers/mmc/host/cavium.h
new file mode 100644
index 000000000..f3eea5eaa
--- /dev/null
+++ b/drivers/mmc/host/cavium.h
@@ -0,0 +1,215 @@
+/*
+ * Driver for MMC and SSD cards for Cavium OCTEON and ThunderX SOCs.
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 2012-2017 Cavium Inc.
+ */
+
+#ifndef _CAVIUM_MMC_H_
+#define _CAVIUM_MMC_H_
+
+#include <linux/bitops.h>
+#include <linux/clk.h>
+#include <linux/gpio/consumer.h>
+#include <linux/io.h>
+#include <linux/mmc/host.h>
+#include <linux/of.h>
+#include <linux/scatterlist.h>
+#include <linux/semaphore.h>
+
+#define CAVIUM_MAX_MMC		4
+
+/* DMA register addresses */
+#define MIO_EMM_DMA_FIFO_CFG(x)	(0x00 + x->reg_off_dma)
+#define MIO_EMM_DMA_FIFO_ADR(x)	(0x10 + x->reg_off_dma)
+#define MIO_EMM_DMA_FIFO_CMD(x)	(0x18 + x->reg_off_dma)
+#define MIO_EMM_DMA_CFG(x)	(0x20 + x->reg_off_dma)
+#define MIO_EMM_DMA_ADR(x)	(0x28 + x->reg_off_dma)
+#define MIO_EMM_DMA_INT(x)	(0x30 + x->reg_off_dma)
+#define MIO_EMM_DMA_INT_W1S(x)	(0x38 + x->reg_off_dma)
+#define MIO_EMM_DMA_INT_ENA_W1S(x) (0x40 + x->reg_off_dma)
+#define MIO_EMM_DMA_INT_ENA_W1C(x) (0x48 + x->reg_off_dma)
+
+/* register addresses */
+#define MIO_EMM_CFG(x)		(0x00 + x->reg_off)
+#define MIO_EMM_SWITCH(x)	(0x48 + x->reg_off)
+#define MIO_EMM_DMA(x)		(0x50 + x->reg_off)
+#define MIO_EMM_CMD(x)		(0x58 + x->reg_off)
+#define MIO_EMM_RSP_STS(x)	(0x60 + x->reg_off)
+#define MIO_EMM_RSP_LO(x)	(0x68 + x->reg_off)
+#define MIO_EMM_RSP_HI(x)	(0x70 + x->reg_off)
+#define MIO_EMM_INT(x)		(0x78 + x->reg_off)
+#define MIO_EMM_INT_EN(x)	(0x80 + x->reg_off)
+#define MIO_EMM_WDOG(x)		(0x88 + x->reg_off)
+#define MIO_EMM_SAMPLE(x)	(0x90 + x->reg_off)
+#define MIO_EMM_STS_MASK(x)	(0x98 + x->reg_off)
+#define MIO_EMM_RCA(x)		(0xa0 + x->reg_off)
+#define MIO_EMM_INT_EN_SET(x)	(0xb0 + x->reg_off)
+#define MIO_EMM_INT_EN_CLR(x)	(0xb8 + x->reg_off)
+#define MIO_EMM_BUF_IDX(x)	(0xe0 + x->reg_off)
+#define MIO_EMM_BUF_DAT(x)	(0xe8 + x->reg_off)
+
+struct cvm_mmc_host {
+	struct device *dev;
+	void __iomem *base;
+	void __iomem *dma_base;
+	int reg_off;
+	int reg_off_dma;
+	u64 emm_cfg;
+	u64 n_minus_one;	/* OCTEON II workaround location */
+	int last_slot;
+	struct clk *clk;
+	int sys_freq;
+
+	struct mmc_request *current_req;
+	struct sg_mapping_iter smi;
+	bool dma_active;
+	bool use_sg;
+
+	bool has_ciu3;
+	bool big_dma_addr;
+	bool need_irq_handler_lock;
+	spinlock_t irq_handler_lock;
+	struct semaphore mmc_serializer;
+
+	struct gpio_desc *global_pwr_gpiod;
+	atomic_t shared_power_users;
+
+	struct cvm_mmc_slot *slot[CAVIUM_MAX_MMC];
+	struct platform_device *slot_pdev[CAVIUM_MAX_MMC];
+
+	void (*set_shared_power)(struct cvm_mmc_host *, int);
+	void (*acquire_bus)(struct cvm_mmc_host *);
+	void (*release_bus)(struct cvm_mmc_host *);
+	void (*int_enable)(struct cvm_mmc_host *, u64);
+	/* required on some MIPS models */
+	void (*dmar_fixup)(struct cvm_mmc_host *, struct mmc_command *,
+			   struct mmc_data *, u64);
+	void (*dmar_fixup_done)(struct cvm_mmc_host *);
+};
+
+struct cvm_mmc_slot {
+	struct mmc_host *mmc;		/* slot-level mmc_core object */
+	struct cvm_mmc_host *host;	/* common hw for all slots */
+
+	u64 clock;
+
+	u64 cached_switch;
+	u64 cached_rca;
+
+	unsigned int cmd_cnt;		/* sample delay */
+	unsigned int dat_cnt;		/* sample delay */
+
+	int bus_id;
+};
+
+struct cvm_mmc_cr_type {
+	u8 ctype;
+	u8 rtype;
+};
+
+struct cvm_mmc_cr_mods {
+	u8 ctype_xor;
+	u8 rtype_xor;
+};
+
+/* Bitfield definitions */
+#define MIO_EMM_DMA_FIFO_CFG_CLR	BIT_ULL(16)
+#define MIO_EMM_DMA_FIFO_CFG_INT_LVL	GENMASK_ULL(12, 8)
+#define MIO_EMM_DMA_FIFO_CFG_COUNT	GENMASK_ULL(4, 0)
+
+#define MIO_EMM_DMA_FIFO_CMD_RW		BIT_ULL(62)
+#define MIO_EMM_DMA_FIFO_CMD_INTDIS	BIT_ULL(60)
+#define MIO_EMM_DMA_FIFO_CMD_SWAP32	BIT_ULL(59)
+#define MIO_EMM_DMA_FIFO_CMD_SWAP16	BIT_ULL(58)
+#define MIO_EMM_DMA_FIFO_CMD_SWAP8	BIT_ULL(57)
+#define MIO_EMM_DMA_FIFO_CMD_ENDIAN	BIT_ULL(56)
+#define MIO_EMM_DMA_FIFO_CMD_SIZE	GENMASK_ULL(55, 36)
+
+#define MIO_EMM_CMD_SKIP_BUSY		BIT_ULL(62)
+#define MIO_EMM_CMD_BUS_ID		GENMASK_ULL(61, 60)
+#define MIO_EMM_CMD_VAL			BIT_ULL(59)
+#define MIO_EMM_CMD_DBUF		BIT_ULL(55)
+#define MIO_EMM_CMD_OFFSET		GENMASK_ULL(54, 49)
+#define MIO_EMM_CMD_CTYPE_XOR		GENMASK_ULL(42, 41)
+#define MIO_EMM_CMD_RTYPE_XOR		GENMASK_ULL(40, 38)
+#define MIO_EMM_CMD_IDX			GENMASK_ULL(37, 32)
+#define MIO_EMM_CMD_ARG			GENMASK_ULL(31, 0)
+
+#define MIO_EMM_DMA_SKIP_BUSY		BIT_ULL(62)
+#define MIO_EMM_DMA_BUS_ID		GENMASK_ULL(61, 60)
+#define MIO_EMM_DMA_VAL			BIT_ULL(59)
+#define MIO_EMM_DMA_SECTOR		BIT_ULL(58)
+#define MIO_EMM_DMA_DAT_NULL		BIT_ULL(57)
+#define MIO_EMM_DMA_THRES		GENMASK_ULL(56, 51)
+#define MIO_EMM_DMA_REL_WR		BIT_ULL(50)
+#define MIO_EMM_DMA_RW			BIT_ULL(49)
+#define MIO_EMM_DMA_MULTI		BIT_ULL(48)
+#define MIO_EMM_DMA_BLOCK_CNT		GENMASK_ULL(47, 32)
+#define MIO_EMM_DMA_CARD_ADDR		GENMASK_ULL(31, 0)
+
+#define MIO_EMM_DMA_CFG_EN		BIT_ULL(63)
+#define MIO_EMM_DMA_CFG_RW		BIT_ULL(62)
+#define MIO_EMM_DMA_CFG_CLR		BIT_ULL(61)
+#define MIO_EMM_DMA_CFG_SWAP32		BIT_ULL(59)
+#define MIO_EMM_DMA_CFG_SWAP16		BIT_ULL(58)
+#define MIO_EMM_DMA_CFG_SWAP8		BIT_ULL(57)
+#define MIO_EMM_DMA_CFG_ENDIAN		BIT_ULL(56)
+#define MIO_EMM_DMA_CFG_SIZE		GENMASK_ULL(55, 36)
+#define MIO_EMM_DMA_CFG_ADR		GENMASK_ULL(35, 0)
+
+#define MIO_EMM_INT_SWITCH_ERR		BIT_ULL(6)
+#define MIO_EMM_INT_SWITCH_DONE		BIT_ULL(5)
+#define MIO_EMM_INT_DMA_ERR		BIT_ULL(4)
+#define MIO_EMM_INT_CMD_ERR		BIT_ULL(3)
+#define MIO_EMM_INT_DMA_DONE		BIT_ULL(2)
+#define MIO_EMM_INT_CMD_DONE		BIT_ULL(1)
+#define MIO_EMM_INT_BUF_DONE		BIT_ULL(0)
+
+#define MIO_EMM_RSP_STS_BUS_ID		GENMASK_ULL(61, 60)
+#define MIO_EMM_RSP_STS_CMD_VAL		BIT_ULL(59)
+#define MIO_EMM_RSP_STS_SWITCH_VAL	BIT_ULL(58)
+#define MIO_EMM_RSP_STS_DMA_VAL		BIT_ULL(57)
+#define MIO_EMM_RSP_STS_DMA_PEND	BIT_ULL(56)
+#define MIO_EMM_RSP_STS_DBUF_ERR	BIT_ULL(28)
+#define MIO_EMM_RSP_STS_DBUF		BIT_ULL(23)
+#define MIO_EMM_RSP_STS_BLK_TIMEOUT	BIT_ULL(22)
+#define MIO_EMM_RSP_STS_BLK_CRC_ERR	BIT_ULL(21)
+#define MIO_EMM_RSP_STS_RSP_BUSYBIT	BIT_ULL(20)
+#define MIO_EMM_RSP_STS_STP_TIMEOUT	BIT_ULL(19)
+#define MIO_EMM_RSP_STS_STP_CRC_ERR	BIT_ULL(18)
+#define MIO_EMM_RSP_STS_STP_BAD_STS	BIT_ULL(17)
+#define MIO_EMM_RSP_STS_STP_VAL		BIT_ULL(16)
+#define MIO_EMM_RSP_STS_RSP_TIMEOUT	BIT_ULL(15)
+#define MIO_EMM_RSP_STS_RSP_CRC_ERR	BIT_ULL(14)
+#define MIO_EMM_RSP_STS_RSP_BAD_STS	BIT_ULL(13)
+#define MIO_EMM_RSP_STS_RSP_VAL		BIT_ULL(12)
+#define MIO_EMM_RSP_STS_RSP_TYPE	GENMASK_ULL(11, 9)
+#define MIO_EMM_RSP_STS_CMD_TYPE	GENMASK_ULL(8, 7)
+#define MIO_EMM_RSP_STS_CMD_IDX		GENMASK_ULL(6, 1)
+#define MIO_EMM_RSP_STS_CMD_DONE	BIT_ULL(0)
+
+#define MIO_EMM_SAMPLE_CMD_CNT		GENMASK_ULL(25, 16)
+#define MIO_EMM_SAMPLE_DAT_CNT		GENMASK_ULL(9, 0)
+
+#define MIO_EMM_SWITCH_BUS_ID		GENMASK_ULL(61, 60)
+#define MIO_EMM_SWITCH_EXE		BIT_ULL(59)
+#define MIO_EMM_SWITCH_ERR0		BIT_ULL(58)
+#define MIO_EMM_SWITCH_ERR1		BIT_ULL(57)
+#define MIO_EMM_SWITCH_ERR2		BIT_ULL(56)
+#define MIO_EMM_SWITCH_HS_TIMING	BIT_ULL(48)
+#define MIO_EMM_SWITCH_BUS_WIDTH	GENMASK_ULL(42, 40)
+#define MIO_EMM_SWITCH_POWER_CLASS	GENMASK_ULL(35, 32)
+#define MIO_EMM_SWITCH_CLK_HI		GENMASK_ULL(31, 16)
+#define MIO_EMM_SWITCH_CLK_LO		GENMASK_ULL(15, 0)
+
+/* Protoypes */
+irqreturn_t cvm_mmc_interrupt(int irq, void *dev_id);
+int cvm_mmc_of_slot_probe(struct device *dev, struct cvm_mmc_host *host);
+int cvm_mmc_of_slot_remove(struct cvm_mmc_slot *slot);
+extern const char *cvm_mmc_irq_names[];
+
+#endif
diff --git a/drivers/mmc/host/cb710-mmc.c b/drivers/mmc/host/cb710-mmc.c
new file mode 100644
index 000000000..e84ed84ea
--- /dev/null
+++ b/drivers/mmc/host/cb710-mmc.c
@@ -0,0 +1,787 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ *  cb710/mmc.c
+ *
+ *  Copyright by Michał Mirosław, 2008-2009
+ */
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/delay.h>
+#include "cb710-mmc.h"
+
+#define CB710_MMC_REQ_TIMEOUT_MS	2000
+
+static const u8 cb710_clock_divider_log2[8] = {
+/*	1, 2, 4, 8, 16, 32, 128, 512 */
+	0, 1, 2, 3,  4,  5,   7,   9
+};
+#define CB710_MAX_DIVIDER_IDX	\
+	(ARRAY_SIZE(cb710_clock_divider_log2) - 1)
+
+static const u8 cb710_src_freq_mhz[16] = {
+	33, 10, 20, 25, 30, 35, 40, 45,
+	50, 55, 60, 65, 70, 75, 80, 85
+};
+
+static void cb710_mmc_select_clock_divider(struct mmc_host *mmc, int hz)
+{
+	struct cb710_slot *slot = cb710_mmc_to_slot(mmc);
+	struct pci_dev *pdev = cb710_slot_to_chip(slot)->pdev;
+	u32 src_freq_idx;
+	u32 divider_idx;
+	int src_hz;
+
+	/* on CB710 in HP nx9500:
+	 *   src_freq_idx == 0
+	 *   indexes 1-7 work as written in the table
+	 *   indexes 0,8-15 give no clock output
+	 */
+	pci_read_config_dword(pdev, 0x48, &src_freq_idx);
+	src_freq_idx = (src_freq_idx >> 16) & 0xF;
+	src_hz = cb710_src_freq_mhz[src_freq_idx] * 1000000;
+
+	for (divider_idx = 0; divider_idx < CB710_MAX_DIVIDER_IDX; ++divider_idx) {
+		if (hz >= src_hz >> cb710_clock_divider_log2[divider_idx])
+			break;
+	}
+
+	if (src_freq_idx)
+		divider_idx |= 0x8;
+	else if (divider_idx == 0)
+		divider_idx = 1;
+
+	cb710_pci_update_config_reg(pdev, 0x40, ~0xF0000000, divider_idx << 28);
+
+	dev_dbg(cb710_slot_dev(slot),
+		"clock set to %d Hz, wanted %d Hz; src_freq_idx = %d, divider_idx = %d|%d\n",
+		src_hz >> cb710_clock_divider_log2[divider_idx & 7],
+		hz, src_freq_idx, divider_idx & 7, divider_idx & 8);
+}
+
+static void __cb710_mmc_enable_irq(struct cb710_slot *slot,
+	unsigned short enable, unsigned short mask)
+{
+	/* clear global IE
+	 * - it gets set later if any interrupt sources are enabled */
+	mask |= CB710_MMC_IE_IRQ_ENABLE;
+
+	/* look like interrupt is fired whenever
+	 * WORD[0x0C] & WORD[0x10] != 0;
+	 * -> bit 15 port 0x0C seems to be global interrupt enable
+	 */
+
+	enable = (cb710_read_port_16(slot, CB710_MMC_IRQ_ENABLE_PORT)
+		& ~mask) | enable;
+
+	if (enable)
+		enable |= CB710_MMC_IE_IRQ_ENABLE;
+
+	cb710_write_port_16(slot, CB710_MMC_IRQ_ENABLE_PORT, enable);
+}
+
+static void cb710_mmc_enable_irq(struct cb710_slot *slot,
+	unsigned short enable, unsigned short mask)
+{
+	struct cb710_mmc_reader *reader = mmc_priv(cb710_slot_to_mmc(slot));
+	unsigned long flags;
+
+	spin_lock_irqsave(&reader->irq_lock, flags);
+	/* this is the only thing irq_lock protects */
+	__cb710_mmc_enable_irq(slot, enable, mask);
+	spin_unlock_irqrestore(&reader->irq_lock, flags);
+}
+
+static void cb710_mmc_reset_events(struct cb710_slot *slot)
+{
+	cb710_write_port_8(slot, CB710_MMC_STATUS0_PORT, 0xFF);
+	cb710_write_port_8(slot, CB710_MMC_STATUS1_PORT, 0xFF);
+	cb710_write_port_8(slot, CB710_MMC_STATUS2_PORT, 0xFF);
+}
+
+static void cb710_mmc_enable_4bit_data(struct cb710_slot *slot, int enable)
+{
+	if (enable)
+		cb710_modify_port_8(slot, CB710_MMC_CONFIG1_PORT,
+			CB710_MMC_C1_4BIT_DATA_BUS, 0);
+	else
+		cb710_modify_port_8(slot, CB710_MMC_CONFIG1_PORT,
+			0, CB710_MMC_C1_4BIT_DATA_BUS);
+}
+
+static int cb710_check_event(struct cb710_slot *slot, u8 what)
+{
+	u16 status;
+
+	status = cb710_read_port_16(slot, CB710_MMC_STATUS_PORT);
+
+	if (status & CB710_MMC_S0_FIFO_UNDERFLOW) {
+		/* it is just a guess, so log it */
+		dev_dbg(cb710_slot_dev(slot),
+			"CHECK : ignoring bit 6 in status %04X\n", status);
+		cb710_write_port_8(slot, CB710_MMC_STATUS0_PORT,
+			CB710_MMC_S0_FIFO_UNDERFLOW);
+		status &= ~CB710_MMC_S0_FIFO_UNDERFLOW;
+	}
+
+	if (status & CB710_MMC_STATUS_ERROR_EVENTS) {
+		dev_dbg(cb710_slot_dev(slot),
+			"CHECK : returning EIO on status %04X\n", status);
+		cb710_write_port_8(slot, CB710_MMC_STATUS0_PORT, status & 0xFF);
+		cb710_write_port_8(slot, CB710_MMC_STATUS1_PORT,
+			CB710_MMC_S1_RESET);
+		return -EIO;
+	}
+
+	/* 'what' is a bit in MMC_STATUS1 */
+	if ((status >> 8) & what) {
+		cb710_write_port_8(slot, CB710_MMC_STATUS1_PORT, what);
+		return 1;
+	}
+
+	return 0;
+}
+
+static int cb710_wait_for_event(struct cb710_slot *slot, u8 what)
+{
+	int err = 0;
+	unsigned limit = 2000000;	/* FIXME: real timeout */
+
+#ifdef CONFIG_CB710_DEBUG
+	u32 e, x;
+	e = cb710_read_port_32(slot, CB710_MMC_STATUS_PORT);
+#endif
+
+	while (!(err = cb710_check_event(slot, what))) {
+		if (!--limit) {
+			cb710_dump_regs(cb710_slot_to_chip(slot),
+				CB710_DUMP_REGS_MMC);
+			err = -ETIMEDOUT;
+			break;
+		}
+		udelay(1);
+	}
+
+#ifdef CONFIG_CB710_DEBUG
+	x = cb710_read_port_32(slot, CB710_MMC_STATUS_PORT);
+
+	limit = 2000000 - limit;
+	if (limit > 100)
+		dev_dbg(cb710_slot_dev(slot),
+			"WAIT10: waited %d loops, what %d, entry val %08X, exit val %08X\n",
+			limit, what, e, x);
+#endif
+	return err < 0 ? err : 0;
+}
+
+
+static int cb710_wait_while_busy(struct cb710_slot *slot, uint8_t mask)
+{
+	unsigned limit = 500000;	/* FIXME: real timeout */
+	int err = 0;
+
+#ifdef CONFIG_CB710_DEBUG
+	u32 e, x;
+	e = cb710_read_port_32(slot, CB710_MMC_STATUS_PORT);
+#endif
+
+	while (cb710_read_port_8(slot, CB710_MMC_STATUS2_PORT) & mask) {
+		if (!--limit) {
+			cb710_dump_regs(cb710_slot_to_chip(slot),
+				CB710_DUMP_REGS_MMC);
+			err = -ETIMEDOUT;
+			break;
+		}
+		udelay(1);
+	}
+
+#ifdef CONFIG_CB710_DEBUG
+	x = cb710_read_port_32(slot, CB710_MMC_STATUS_PORT);
+
+	limit = 500000 - limit;
+	if (limit > 100)
+		dev_dbg(cb710_slot_dev(slot),
+			"WAIT12: waited %d loops, mask %02X, entry val %08X, exit val %08X\n",
+			limit, mask, e, x);
+#endif
+	return err;
+}
+
+static void cb710_mmc_set_transfer_size(struct cb710_slot *slot,
+	size_t count, size_t blocksize)
+{
+	cb710_wait_while_busy(slot, CB710_MMC_S2_BUSY_20);
+	cb710_write_port_32(slot, CB710_MMC_TRANSFER_SIZE_PORT,
+		((count - 1) << 16)|(blocksize - 1));
+
+	dev_vdbg(cb710_slot_dev(slot), "set up for %zu block%s of %zu bytes\n",
+		count, count == 1 ? "" : "s", blocksize);
+}
+
+static void cb710_mmc_fifo_hack(struct cb710_slot *slot)
+{
+	/* without this, received data is prepended with 8-bytes of zeroes */
+	u32 r1, r2;
+	int ok = 0;
+
+	r1 = cb710_read_port_32(slot, CB710_MMC_DATA_PORT);
+	r2 = cb710_read_port_32(slot, CB710_MMC_DATA_PORT);
+	if (cb710_read_port_8(slot, CB710_MMC_STATUS0_PORT)
+	    & CB710_MMC_S0_FIFO_UNDERFLOW) {
+		cb710_write_port_8(slot, CB710_MMC_STATUS0_PORT,
+			CB710_MMC_S0_FIFO_UNDERFLOW);
+		ok = 1;
+	}
+
+	dev_dbg(cb710_slot_dev(slot),
+		"FIFO-read-hack: expected STATUS0 bit was %s\n",
+		ok ? "set." : "NOT SET!");
+	dev_dbg(cb710_slot_dev(slot),
+		"FIFO-read-hack: dwords ignored: %08X %08X - %s\n",
+		r1, r2, (r1|r2) ? "BAD (NOT ZERO)!" : "ok");
+}
+
+static int cb710_mmc_receive_pio(struct cb710_slot *slot,
+	struct sg_mapping_iter *miter, size_t dw_count)
+{
+	if (!(cb710_read_port_8(slot, CB710_MMC_STATUS2_PORT) & CB710_MMC_S2_FIFO_READY)) {
+		int err = cb710_wait_for_event(slot,
+			CB710_MMC_S1_PIO_TRANSFER_DONE);
+		if (err)
+			return err;
+	}
+
+	cb710_sg_dwiter_write_from_io(miter,
+		slot->iobase + CB710_MMC_DATA_PORT, dw_count);
+
+	return 0;
+}
+
+static bool cb710_is_transfer_size_supported(struct mmc_data *data)
+{
+	return !(data->blksz & 15 && (data->blocks != 1 || data->blksz != 8));
+}
+
+static int cb710_mmc_receive(struct cb710_slot *slot, struct mmc_data *data)
+{
+	struct sg_mapping_iter miter;
+	size_t len, blocks = data->blocks;
+	int err = 0;
+
+	/* TODO: I don't know how/if the hardware handles non-16B-boundary blocks
+	 * except single 8B block */
+	if (unlikely(data->blksz & 15 && (data->blocks != 1 || data->blksz != 8)))
+		return -EINVAL;
+
+	sg_miter_start(&miter, data->sg, data->sg_len, SG_MITER_TO_SG);
+
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG2_PORT,
+		15, CB710_MMC_C2_READ_PIO_SIZE_MASK);
+
+	cb710_mmc_fifo_hack(slot);
+
+	while (blocks-- > 0) {
+		len = data->blksz;
+
+		while (len >= 16) {
+			err = cb710_mmc_receive_pio(slot, &miter, 4);
+			if (err)
+				goto out;
+			len -= 16;
+		}
+
+		if (!len)
+			continue;
+
+		cb710_modify_port_8(slot, CB710_MMC_CONFIG2_PORT,
+			len - 1, CB710_MMC_C2_READ_PIO_SIZE_MASK);
+
+		len = (len >= 8) ? 4 : 2;
+		err = cb710_mmc_receive_pio(slot, &miter, len);
+		if (err)
+			goto out;
+	}
+out:
+	sg_miter_stop(&miter);
+	return err;
+}
+
+static int cb710_mmc_send(struct cb710_slot *slot, struct mmc_data *data)
+{
+	struct sg_mapping_iter miter;
+	size_t len, blocks = data->blocks;
+	int err = 0;
+
+	/* TODO: I don't know how/if the hardware handles multiple
+	 * non-16B-boundary blocks */
+	if (unlikely(data->blocks > 1 && data->blksz & 15))
+		return -EINVAL;
+
+	sg_miter_start(&miter, data->sg, data->sg_len, SG_MITER_FROM_SG);
+
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG2_PORT,
+		0, CB710_MMC_C2_READ_PIO_SIZE_MASK);
+
+	while (blocks-- > 0) {
+		len = (data->blksz + 15) >> 4;
+		do {
+			if (!(cb710_read_port_8(slot, CB710_MMC_STATUS2_PORT)
+			    & CB710_MMC_S2_FIFO_EMPTY)) {
+				err = cb710_wait_for_event(slot,
+					CB710_MMC_S1_PIO_TRANSFER_DONE);
+				if (err)
+					goto out;
+			}
+			cb710_sg_dwiter_read_to_io(&miter,
+				slot->iobase + CB710_MMC_DATA_PORT, 4);
+		} while (--len);
+	}
+out:
+	sg_miter_stop(&miter);
+	return err;
+}
+
+static u16 cb710_encode_cmd_flags(struct cb710_mmc_reader *reader,
+	struct mmc_command *cmd)
+{
+	unsigned int flags = cmd->flags;
+	u16 cb_flags = 0;
+
+	/* Windows driver returned 0 for commands for which no response
+	 * is expected. It happened that there were only two such commands
+	 * used: MMC_GO_IDLE_STATE and MMC_GO_INACTIVE_STATE so it might
+	 * as well be a bug in that driver.
+	 *
+	 * Original driver set bit 14 for MMC/SD application
+	 * commands. There's no difference 'on the wire' and
+	 * it apparently works without it anyway.
+	 */
+
+	switch (flags & MMC_CMD_MASK) {
+	case MMC_CMD_AC:	cb_flags = CB710_MMC_CMD_AC;	break;
+	case MMC_CMD_ADTC:	cb_flags = CB710_MMC_CMD_ADTC;	break;
+	case MMC_CMD_BC:	cb_flags = CB710_MMC_CMD_BC;	break;
+	case MMC_CMD_BCR:	cb_flags = CB710_MMC_CMD_BCR;	break;
+	}
+
+	if (flags & MMC_RSP_BUSY)
+		cb_flags |= CB710_MMC_RSP_BUSY;
+
+	cb_flags |= cmd->opcode << CB710_MMC_CMD_CODE_SHIFT;
+
+	if (cmd->data && (cmd->data->flags & MMC_DATA_READ))
+		cb_flags |= CB710_MMC_DATA_READ;
+
+	if (flags & MMC_RSP_PRESENT) {
+		/* Windows driver set 01 at bits 4,3 except for
+		 * MMC_SET_BLOCKLEN where it set 10. Maybe the
+		 * hardware can do something special about this
+		 * command? The original driver looks buggy/incomplete
+		 * anyway so we ignore this for now.
+		 *
+		 * I assume that 00 here means no response is expected.
+		 */
+		cb_flags |= CB710_MMC_RSP_PRESENT;
+
+		if (flags & MMC_RSP_136)
+			cb_flags |= CB710_MMC_RSP_136;
+		if (!(flags & MMC_RSP_CRC))
+			cb_flags |= CB710_MMC_RSP_NO_CRC;
+	}
+
+	return cb_flags;
+}
+
+static void cb710_receive_response(struct cb710_slot *slot,
+	struct mmc_command *cmd)
+{
+	unsigned rsp_opcode, wanted_opcode;
+
+	/* Looks like final byte with CRC is always stripped (same as SDHCI) */
+	if (cmd->flags & MMC_RSP_136) {
+		u32 resp[4];
+
+		resp[0] = cb710_read_port_32(slot, CB710_MMC_RESPONSE3_PORT);
+		resp[1] = cb710_read_port_32(slot, CB710_MMC_RESPONSE2_PORT);
+		resp[2] = cb710_read_port_32(slot, CB710_MMC_RESPONSE1_PORT);
+		resp[3] = cb710_read_port_32(slot, CB710_MMC_RESPONSE0_PORT);
+		rsp_opcode = resp[0] >> 24;
+
+		cmd->resp[0] = (resp[0] << 8)|(resp[1] >> 24);
+		cmd->resp[1] = (resp[1] << 8)|(resp[2] >> 24);
+		cmd->resp[2] = (resp[2] << 8)|(resp[3] >> 24);
+		cmd->resp[3] = (resp[3] << 8);
+	} else {
+		rsp_opcode = cb710_read_port_32(slot, CB710_MMC_RESPONSE1_PORT) & 0x3F;
+		cmd->resp[0] = cb710_read_port_32(slot, CB710_MMC_RESPONSE0_PORT);
+	}
+
+	wanted_opcode = (cmd->flags & MMC_RSP_OPCODE) ? cmd->opcode : 0x3F;
+	if (rsp_opcode != wanted_opcode)
+		cmd->error = -EILSEQ;
+}
+
+static int cb710_mmc_transfer_data(struct cb710_slot *slot,
+	struct mmc_data *data)
+{
+	int error, to;
+
+	if (data->flags & MMC_DATA_READ)
+		error = cb710_mmc_receive(slot, data);
+	else
+		error = cb710_mmc_send(slot, data);
+
+	to = cb710_wait_for_event(slot, CB710_MMC_S1_DATA_TRANSFER_DONE);
+	if (!error)
+		error = to;
+
+	if (!error)
+		data->bytes_xfered = data->blksz * data->blocks;
+	return error;
+}
+
+static int cb710_mmc_command(struct mmc_host *mmc, struct mmc_command *cmd)
+{
+	struct cb710_slot *slot = cb710_mmc_to_slot(mmc);
+	struct cb710_mmc_reader *reader = mmc_priv(mmc);
+	struct mmc_data *data = cmd->data;
+
+	u16 cb_cmd = cb710_encode_cmd_flags(reader, cmd);
+	dev_dbg(cb710_slot_dev(slot), "cmd request: 0x%04X\n", cb_cmd);
+
+	if (data) {
+		if (!cb710_is_transfer_size_supported(data)) {
+			data->error = -EINVAL;
+			return -1;
+		}
+		cb710_mmc_set_transfer_size(slot, data->blocks, data->blksz);
+	}
+
+	cb710_wait_while_busy(slot, CB710_MMC_S2_BUSY_20|CB710_MMC_S2_BUSY_10);
+	cb710_write_port_16(slot, CB710_MMC_CMD_TYPE_PORT, cb_cmd);
+	cb710_wait_while_busy(slot, CB710_MMC_S2_BUSY_20);
+	cb710_write_port_32(slot, CB710_MMC_CMD_PARAM_PORT, cmd->arg);
+	cb710_mmc_reset_events(slot);
+	cb710_wait_while_busy(slot, CB710_MMC_S2_BUSY_20);
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG0_PORT, 0x01, 0);
+
+	cmd->error = cb710_wait_for_event(slot, CB710_MMC_S1_COMMAND_SENT);
+	if (cmd->error)
+		return -1;
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		cb710_receive_response(slot, cmd);
+		if (cmd->error)
+			return -1;
+	}
+
+	if (data)
+		data->error = cb710_mmc_transfer_data(slot, data);
+	return 0;
+}
+
+static void cb710_mmc_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct cb710_slot *slot = cb710_mmc_to_slot(mmc);
+	struct cb710_mmc_reader *reader = mmc_priv(mmc);
+
+	WARN_ON(reader->mrq != NULL);
+
+	reader->mrq = mrq;
+	cb710_mmc_enable_irq(slot, CB710_MMC_IE_TEST_MASK, 0);
+
+	if (!cb710_mmc_command(mmc, mrq->cmd) && mrq->stop)
+		cb710_mmc_command(mmc, mrq->stop);
+
+	tasklet_schedule(&reader->finish_req_tasklet);
+}
+
+static int cb710_mmc_powerup(struct cb710_slot *slot)
+{
+#ifdef CONFIG_CB710_DEBUG
+	struct cb710_chip *chip = cb710_slot_to_chip(slot);
+#endif
+	int err;
+
+	/* a lot of magic for now */
+	dev_dbg(cb710_slot_dev(slot), "bus powerup\n");
+	cb710_dump_regs(chip, CB710_DUMP_REGS_MMC);
+	err = cb710_wait_while_busy(slot, CB710_MMC_S2_BUSY_20);
+	if (unlikely(err))
+		return err;
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG1_PORT, 0x80, 0);
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG3_PORT, 0x80, 0);
+	cb710_dump_regs(chip, CB710_DUMP_REGS_MMC);
+	mdelay(1);
+	dev_dbg(cb710_slot_dev(slot), "after delay 1\n");
+	cb710_dump_regs(chip, CB710_DUMP_REGS_MMC);
+	err = cb710_wait_while_busy(slot, CB710_MMC_S2_BUSY_20);
+	if (unlikely(err))
+		return err;
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG1_PORT, 0x09, 0);
+	cb710_dump_regs(chip, CB710_DUMP_REGS_MMC);
+	mdelay(1);
+	dev_dbg(cb710_slot_dev(slot), "after delay 2\n");
+	cb710_dump_regs(chip, CB710_DUMP_REGS_MMC);
+	err = cb710_wait_while_busy(slot, CB710_MMC_S2_BUSY_20);
+	if (unlikely(err))
+		return err;
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG1_PORT, 0, 0x08);
+	cb710_dump_regs(chip, CB710_DUMP_REGS_MMC);
+	mdelay(2);
+	dev_dbg(cb710_slot_dev(slot), "after delay 3\n");
+	cb710_dump_regs(chip, CB710_DUMP_REGS_MMC);
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG0_PORT, 0x06, 0);
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG1_PORT, 0x70, 0);
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG2_PORT, 0x80, 0);
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG3_PORT, 0x03, 0);
+	cb710_dump_regs(chip, CB710_DUMP_REGS_MMC);
+	err = cb710_wait_while_busy(slot, CB710_MMC_S2_BUSY_20);
+	if (unlikely(err))
+		return err;
+	/* This port behaves weird: quick byte reads of 0x08,0x09 return
+	 * 0xFF,0x00 after writing 0xFFFF to 0x08; it works correctly when
+	 * read/written from userspace...  What am I missing here?
+	 * (it doesn't depend on write-to-read delay) */
+	cb710_write_port_16(slot, CB710_MMC_CONFIGB_PORT, 0xFFFF);
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG0_PORT, 0x06, 0);
+	cb710_dump_regs(chip, CB710_DUMP_REGS_MMC);
+	dev_dbg(cb710_slot_dev(slot), "bus powerup finished\n");
+
+	return cb710_check_event(slot, 0);
+}
+
+static void cb710_mmc_powerdown(struct cb710_slot *slot)
+{
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG1_PORT, 0, 0x81);
+	cb710_modify_port_8(slot, CB710_MMC_CONFIG3_PORT, 0, 0x80);
+}
+
+static void cb710_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct cb710_slot *slot = cb710_mmc_to_slot(mmc);
+	struct cb710_mmc_reader *reader = mmc_priv(mmc);
+	int err;
+
+	cb710_mmc_select_clock_divider(mmc, ios->clock);
+
+	if (ios->power_mode != reader->last_power_mode) {
+		switch (ios->power_mode) {
+		case MMC_POWER_ON:
+			err = cb710_mmc_powerup(slot);
+			if (err) {
+				dev_warn(cb710_slot_dev(slot),
+					"powerup failed (%d)- retrying\n", err);
+				cb710_mmc_powerdown(slot);
+				udelay(1);
+				err = cb710_mmc_powerup(slot);
+				if (err)
+					dev_warn(cb710_slot_dev(slot),
+						"powerup retry failed (%d) - expect errors\n",
+					err);
+			}
+			reader->last_power_mode = MMC_POWER_ON;
+			break;
+		case MMC_POWER_OFF:
+			cb710_mmc_powerdown(slot);
+			reader->last_power_mode = MMC_POWER_OFF;
+			break;
+		case MMC_POWER_UP:
+		default:
+			/* ignore */
+			break;
+		}
+	}
+
+	cb710_mmc_enable_4bit_data(slot, ios->bus_width != MMC_BUS_WIDTH_1);
+
+	cb710_mmc_enable_irq(slot, CB710_MMC_IE_TEST_MASK, 0);
+}
+
+static int cb710_mmc_get_ro(struct mmc_host *mmc)
+{
+	struct cb710_slot *slot = cb710_mmc_to_slot(mmc);
+
+	return cb710_read_port_8(slot, CB710_MMC_STATUS3_PORT)
+		& CB710_MMC_S3_WRITE_PROTECTED;
+}
+
+static int cb710_mmc_get_cd(struct mmc_host *mmc)
+{
+	struct cb710_slot *slot = cb710_mmc_to_slot(mmc);
+
+	return cb710_read_port_8(slot, CB710_MMC_STATUS3_PORT)
+		& CB710_MMC_S3_CARD_DETECTED;
+}
+
+static int cb710_mmc_irq_handler(struct cb710_slot *slot)
+{
+	struct mmc_host *mmc = cb710_slot_to_mmc(slot);
+	struct cb710_mmc_reader *reader = mmc_priv(mmc);
+	u32 status, config1, config2, irqen;
+
+	status = cb710_read_port_32(slot, CB710_MMC_STATUS_PORT);
+	irqen = cb710_read_port_32(slot, CB710_MMC_IRQ_ENABLE_PORT);
+	config2 = cb710_read_port_32(slot, CB710_MMC_CONFIGB_PORT);
+	config1 = cb710_read_port_32(slot, CB710_MMC_CONFIG_PORT);
+
+	dev_dbg(cb710_slot_dev(slot), "interrupt; status: %08X, "
+		"ie: %08X, c2: %08X, c1: %08X\n",
+		status, irqen, config2, config1);
+
+	if (status & (CB710_MMC_S1_CARD_CHANGED << 8)) {
+		/* ack the event */
+		cb710_write_port_8(slot, CB710_MMC_STATUS1_PORT,
+			CB710_MMC_S1_CARD_CHANGED);
+		if ((irqen & CB710_MMC_IE_CISTATUS_MASK)
+		    == CB710_MMC_IE_CISTATUS_MASK)
+			mmc_detect_change(mmc, HZ/5);
+	} else {
+		dev_dbg(cb710_slot_dev(slot), "unknown interrupt (test)\n");
+		spin_lock(&reader->irq_lock);
+		__cb710_mmc_enable_irq(slot, 0, CB710_MMC_IE_TEST_MASK);
+		spin_unlock(&reader->irq_lock);
+	}
+
+	return 1;
+}
+
+static void cb710_mmc_finish_request_tasklet(unsigned long data)
+{
+	struct mmc_host *mmc = (void *)data;
+	struct cb710_mmc_reader *reader = mmc_priv(mmc);
+	struct mmc_request *mrq = reader->mrq;
+
+	reader->mrq = NULL;
+	mmc_request_done(mmc, mrq);
+}
+
+static const struct mmc_host_ops cb710_mmc_host = {
+	.request = cb710_mmc_request,
+	.set_ios = cb710_mmc_set_ios,
+	.get_ro = cb710_mmc_get_ro,
+	.get_cd = cb710_mmc_get_cd,
+};
+
+#ifdef CONFIG_PM
+
+static int cb710_mmc_suspend(struct platform_device *pdev, pm_message_t state)
+{
+	struct cb710_slot *slot = cb710_pdev_to_slot(pdev);
+
+	cb710_mmc_enable_irq(slot, 0, ~0);
+	return 0;
+}
+
+static int cb710_mmc_resume(struct platform_device *pdev)
+{
+	struct cb710_slot *slot = cb710_pdev_to_slot(pdev);
+
+	cb710_mmc_enable_irq(slot, 0, ~0);
+	return 0;
+}
+
+#endif /* CONFIG_PM */
+
+static int cb710_mmc_init(struct platform_device *pdev)
+{
+	struct cb710_slot *slot = cb710_pdev_to_slot(pdev);
+	struct cb710_chip *chip = cb710_slot_to_chip(slot);
+	struct mmc_host *mmc;
+	struct cb710_mmc_reader *reader;
+	int err;
+	u32 val;
+
+	mmc = mmc_alloc_host(sizeof(*reader), cb710_slot_dev(slot));
+	if (!mmc)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, mmc);
+
+	/* harmless (maybe) magic */
+	pci_read_config_dword(chip->pdev, 0x48, &val);
+	val = cb710_src_freq_mhz[(val >> 16) & 0xF];
+	dev_dbg(cb710_slot_dev(slot), "source frequency: %dMHz\n", val);
+	val *= 1000000;
+
+	mmc->ops = &cb710_mmc_host;
+	mmc->f_max = val;
+	mmc->f_min = val >> cb710_clock_divider_log2[CB710_MAX_DIVIDER_IDX];
+	mmc->ocr_avail = MMC_VDD_32_33|MMC_VDD_33_34;
+	mmc->caps = MMC_CAP_4_BIT_DATA;
+	/*
+	 * In cb710_wait_for_event() we use a fixed timeout of ~2s, hence let's
+	 * inform the core about it. A future improvement should instead make
+	 * use of the cmd->busy_timeout.
+	 */
+	mmc->max_busy_timeout = CB710_MMC_REQ_TIMEOUT_MS;
+
+	reader = mmc_priv(mmc);
+
+	tasklet_init(&reader->finish_req_tasklet,
+		cb710_mmc_finish_request_tasklet, (unsigned long)mmc);
+	spin_lock_init(&reader->irq_lock);
+	cb710_dump_regs(chip, CB710_DUMP_REGS_MMC);
+
+	cb710_mmc_enable_irq(slot, 0, ~0);
+	cb710_set_irq_handler(slot, cb710_mmc_irq_handler);
+
+	err = mmc_add_host(mmc);
+	if (unlikely(err))
+		goto err_free_mmc;
+
+	dev_dbg(cb710_slot_dev(slot), "mmc_hostname is %s\n",
+		mmc_hostname(mmc));
+
+	cb710_mmc_enable_irq(slot, CB710_MMC_IE_CARD_INSERTION_STATUS, 0);
+
+	return 0;
+
+err_free_mmc:
+	dev_dbg(cb710_slot_dev(slot), "mmc_add_host() failed: %d\n", err);
+
+	cb710_set_irq_handler(slot, NULL);
+	mmc_free_host(mmc);
+	return err;
+}
+
+static int cb710_mmc_exit(struct platform_device *pdev)
+{
+	struct cb710_slot *slot = cb710_pdev_to_slot(pdev);
+	struct mmc_host *mmc = cb710_slot_to_mmc(slot);
+	struct cb710_mmc_reader *reader = mmc_priv(mmc);
+
+	cb710_mmc_enable_irq(slot, 0, CB710_MMC_IE_CARD_INSERTION_STATUS);
+
+	mmc_remove_host(mmc);
+
+	/* IRQs should be disabled now, but let's stay on the safe side */
+	cb710_mmc_enable_irq(slot, 0, ~0);
+	cb710_set_irq_handler(slot, NULL);
+
+	/* clear config ports - just in case */
+	cb710_write_port_32(slot, CB710_MMC_CONFIG_PORT, 0);
+	cb710_write_port_16(slot, CB710_MMC_CONFIGB_PORT, 0);
+
+	tasklet_kill(&reader->finish_req_tasklet);
+
+	mmc_free_host(mmc);
+	return 0;
+}
+
+static struct platform_driver cb710_mmc_driver = {
+	.driver.name = "cb710-mmc",
+	.probe = cb710_mmc_init,
+	.remove = cb710_mmc_exit,
+#ifdef CONFIG_PM
+	.suspend = cb710_mmc_suspend,
+	.resume = cb710_mmc_resume,
+#endif
+};
+
+module_platform_driver(cb710_mmc_driver);
+
+MODULE_AUTHOR("Michał Mirosław <mirq-linux@rere.qmqm.pl>");
+MODULE_DESCRIPTION("ENE CB710 memory card reader driver - MMC/SD part");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:cb710-mmc");
diff --git a/drivers/mmc/host/cb710-mmc.h b/drivers/mmc/host/cb710-mmc.h
new file mode 100644
index 000000000..5e053077d
--- /dev/null
+++ b/drivers/mmc/host/cb710-mmc.h
@@ -0,0 +1,100 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ *  cb710/cb710-mmc.h
+ *
+ *  Copyright by Michał Mirosław, 2008-2009
+ */
+#ifndef LINUX_CB710_MMC_H
+#define LINUX_CB710_MMC_H
+
+#include <linux/cb710.h>
+
+/* per-MMC-reader structure */
+struct cb710_mmc_reader {
+	struct tasklet_struct finish_req_tasklet;
+	struct mmc_request *mrq;
+	spinlock_t irq_lock;
+	unsigned char last_power_mode;
+};
+
+/* some device struct walking */
+
+static inline struct mmc_host *cb710_slot_to_mmc(struct cb710_slot *slot)
+{
+	return platform_get_drvdata(&slot->pdev);
+}
+
+static inline struct cb710_slot *cb710_mmc_to_slot(struct mmc_host *mmc)
+{
+	struct platform_device *pdev = to_platform_device(mmc_dev(mmc));
+	return cb710_pdev_to_slot(pdev);
+}
+
+/* registers (this might be all wrong ;) */
+
+#define CB710_MMC_DATA_PORT		0x00
+
+#define CB710_MMC_CONFIG_PORT		0x04
+#define CB710_MMC_CONFIG0_PORT		0x04
+#define CB710_MMC_CONFIG1_PORT		0x05
+#define   CB710_MMC_C1_4BIT_DATA_BUS		0x40
+#define CB710_MMC_CONFIG2_PORT		0x06
+#define   CB710_MMC_C2_READ_PIO_SIZE_MASK	0x0F	/* N-1 */
+#define CB710_MMC_CONFIG3_PORT		0x07
+
+#define CB710_MMC_CONFIGB_PORT		0x08
+
+#define CB710_MMC_IRQ_ENABLE_PORT	0x0C
+#define   CB710_MMC_IE_TEST_MASK		0x00BF
+#define   CB710_MMC_IE_CARD_INSERTION_STATUS	0x1000
+#define   CB710_MMC_IE_IRQ_ENABLE		0x8000
+#define   CB710_MMC_IE_CISTATUS_MASK		\
+		(CB710_MMC_IE_CARD_INSERTION_STATUS|CB710_MMC_IE_IRQ_ENABLE)
+
+#define CB710_MMC_STATUS_PORT		0x10
+#define   CB710_MMC_STATUS_ERROR_EVENTS		0x60FF
+#define CB710_MMC_STATUS0_PORT		0x10
+#define   CB710_MMC_S0_FIFO_UNDERFLOW		0x40
+#define CB710_MMC_STATUS1_PORT		0x11
+#define   CB710_MMC_S1_COMMAND_SENT		0x01
+#define   CB710_MMC_S1_DATA_TRANSFER_DONE	0x02
+#define   CB710_MMC_S1_PIO_TRANSFER_DONE	0x04
+#define   CB710_MMC_S1_CARD_CHANGED		0x10
+#define   CB710_MMC_S1_RESET			0x20
+#define CB710_MMC_STATUS2_PORT		0x12
+#define   CB710_MMC_S2_FIFO_READY		0x01
+#define   CB710_MMC_S2_FIFO_EMPTY		0x02
+#define   CB710_MMC_S2_BUSY_10			0x10
+#define   CB710_MMC_S2_BUSY_20			0x20
+#define CB710_MMC_STATUS3_PORT		0x13
+#define   CB710_MMC_S3_CARD_DETECTED		0x02
+#define   CB710_MMC_S3_WRITE_PROTECTED		0x04
+
+#define CB710_MMC_CMD_TYPE_PORT		0x14
+#define   CB710_MMC_RSP_TYPE_MASK		0x0007
+#define     CB710_MMC_RSP_R1			(0)
+#define     CB710_MMC_RSP_136			(5)
+#define     CB710_MMC_RSP_NO_CRC		(2)
+#define   CB710_MMC_RSP_PRESENT_MASK		0x0018
+#define     CB710_MMC_RSP_NONE			(0 << 3)
+#define     CB710_MMC_RSP_PRESENT		(1 << 3)
+#define     CB710_MMC_RSP_PRESENT_X		(2 << 3)
+#define   CB710_MMC_CMD_TYPE_MASK		0x0060
+#define     CB710_MMC_CMD_BC			(0 << 5)
+#define     CB710_MMC_CMD_BCR			(1 << 5)
+#define     CB710_MMC_CMD_AC			(2 << 5)
+#define     CB710_MMC_CMD_ADTC			(3 << 5)
+#define   CB710_MMC_DATA_READ			0x0080
+#define   CB710_MMC_CMD_CODE_MASK		0x3F00
+#define   CB710_MMC_CMD_CODE_SHIFT		8
+#define   CB710_MMC_IS_APP_CMD			0x4000
+#define   CB710_MMC_RSP_BUSY			0x8000
+
+#define CB710_MMC_CMD_PARAM_PORT	0x18
+#define CB710_MMC_TRANSFER_SIZE_PORT	0x1C
+#define CB710_MMC_RESPONSE0_PORT	0x20
+#define CB710_MMC_RESPONSE1_PORT	0x24
+#define CB710_MMC_RESPONSE2_PORT	0x28
+#define CB710_MMC_RESPONSE3_PORT	0x2C
+
+#endif /* LINUX_CB710_MMC_H */
diff --git a/drivers/mmc/host/cqhci.c b/drivers/mmc/host/cqhci.c
new file mode 100644
index 000000000..23cf7912c
--- /dev/null
+++ b/drivers/mmc/host/cqhci.c
@@ -0,0 +1,1168 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (c) 2015, The Linux Foundation. All rights reserved.
+ */
+
+#include <linux/delay.h>
+#include <linux/highmem.h>
+#include <linux/io.h>
+#include <linux/iopoll.h>
+#include <linux/module.h>
+#include <linux/dma-mapping.h>
+#include <linux/slab.h>
+#include <linux/scatterlist.h>
+#include <linux/platform_device.h>
+#include <linux/ktime.h>
+
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/card.h>
+
+#include "cqhci.h"
+
+#define DCMD_SLOT 31
+#define NUM_SLOTS 32
+
+struct cqhci_slot {
+	struct mmc_request *mrq;
+	unsigned int flags;
+#define CQHCI_EXTERNAL_TIMEOUT	BIT(0)
+#define CQHCI_COMPLETED		BIT(1)
+#define CQHCI_HOST_CRC		BIT(2)
+#define CQHCI_HOST_TIMEOUT	BIT(3)
+#define CQHCI_HOST_OTHER	BIT(4)
+};
+
+static inline u8 *get_desc(struct cqhci_host *cq_host, u8 tag)
+{
+	return cq_host->desc_base + (tag * cq_host->slot_sz);
+}
+
+static inline u8 *get_link_desc(struct cqhci_host *cq_host, u8 tag)
+{
+	u8 *desc = get_desc(cq_host, tag);
+
+	return desc + cq_host->task_desc_len;
+}
+
+static inline dma_addr_t get_trans_desc_dma(struct cqhci_host *cq_host, u8 tag)
+{
+	return cq_host->trans_desc_dma_base +
+		(cq_host->mmc->max_segs * tag *
+		 cq_host->trans_desc_len);
+}
+
+static inline u8 *get_trans_desc(struct cqhci_host *cq_host, u8 tag)
+{
+	return cq_host->trans_desc_base +
+		(cq_host->trans_desc_len * cq_host->mmc->max_segs * tag);
+}
+
+static void setup_trans_desc(struct cqhci_host *cq_host, u8 tag)
+{
+	u8 *link_temp;
+	dma_addr_t trans_temp;
+
+	link_temp = get_link_desc(cq_host, tag);
+	trans_temp = get_trans_desc_dma(cq_host, tag);
+
+	memset(link_temp, 0, cq_host->link_desc_len);
+	if (cq_host->link_desc_len > 8)
+		*(link_temp + 8) = 0;
+
+	if (tag == DCMD_SLOT && (cq_host->mmc->caps2 & MMC_CAP2_CQE_DCMD)) {
+		*link_temp = CQHCI_VALID(0) | CQHCI_ACT(0) | CQHCI_END(1);
+		return;
+	}
+
+	*link_temp = CQHCI_VALID(1) | CQHCI_ACT(0x6) | CQHCI_END(0);
+
+	if (cq_host->dma64) {
+		__le64 *data_addr = (__le64 __force *)(link_temp + 4);
+
+		data_addr[0] = cpu_to_le64(trans_temp);
+	} else {
+		__le32 *data_addr = (__le32 __force *)(link_temp + 4);
+
+		data_addr[0] = cpu_to_le32(trans_temp);
+	}
+}
+
+static void cqhci_set_irqs(struct cqhci_host *cq_host, u32 set)
+{
+	cqhci_writel(cq_host, set, CQHCI_ISTE);
+	cqhci_writel(cq_host, set, CQHCI_ISGE);
+}
+
+#define DRV_NAME "cqhci"
+
+#define CQHCI_DUMP(f, x...) \
+	pr_err("%s: " DRV_NAME ": " f, mmc_hostname(mmc), ## x)
+
+static void cqhci_dumpregs(struct cqhci_host *cq_host)
+{
+	struct mmc_host *mmc = cq_host->mmc;
+
+	CQHCI_DUMP("============ CQHCI REGISTER DUMP ===========\n");
+
+	CQHCI_DUMP("Caps:      0x%08x | Version:  0x%08x\n",
+		   cqhci_readl(cq_host, CQHCI_CAP),
+		   cqhci_readl(cq_host, CQHCI_VER));
+	CQHCI_DUMP("Config:    0x%08x | Control:  0x%08x\n",
+		   cqhci_readl(cq_host, CQHCI_CFG),
+		   cqhci_readl(cq_host, CQHCI_CTL));
+	CQHCI_DUMP("Int stat:  0x%08x | Int enab: 0x%08x\n",
+		   cqhci_readl(cq_host, CQHCI_IS),
+		   cqhci_readl(cq_host, CQHCI_ISTE));
+	CQHCI_DUMP("Int sig:   0x%08x | Int Coal: 0x%08x\n",
+		   cqhci_readl(cq_host, CQHCI_ISGE),
+		   cqhci_readl(cq_host, CQHCI_IC));
+	CQHCI_DUMP("TDL base:  0x%08x | TDL up32: 0x%08x\n",
+		   cqhci_readl(cq_host, CQHCI_TDLBA),
+		   cqhci_readl(cq_host, CQHCI_TDLBAU));
+	CQHCI_DUMP("Doorbell:  0x%08x | TCN:      0x%08x\n",
+		   cqhci_readl(cq_host, CQHCI_TDBR),
+		   cqhci_readl(cq_host, CQHCI_TCN));
+	CQHCI_DUMP("Dev queue: 0x%08x | Dev Pend: 0x%08x\n",
+		   cqhci_readl(cq_host, CQHCI_DQS),
+		   cqhci_readl(cq_host, CQHCI_DPT));
+	CQHCI_DUMP("Task clr:  0x%08x | SSC1:     0x%08x\n",
+		   cqhci_readl(cq_host, CQHCI_TCLR),
+		   cqhci_readl(cq_host, CQHCI_SSC1));
+	CQHCI_DUMP("SSC2:      0x%08x | DCMD rsp: 0x%08x\n",
+		   cqhci_readl(cq_host, CQHCI_SSC2),
+		   cqhci_readl(cq_host, CQHCI_CRDCT));
+	CQHCI_DUMP("RED mask:  0x%08x | TERRI:    0x%08x\n",
+		   cqhci_readl(cq_host, CQHCI_RMEM),
+		   cqhci_readl(cq_host, CQHCI_TERRI));
+	CQHCI_DUMP("Resp idx:  0x%08x | Resp arg: 0x%08x\n",
+		   cqhci_readl(cq_host, CQHCI_CRI),
+		   cqhci_readl(cq_host, CQHCI_CRA));
+
+	if (cq_host->ops->dumpregs)
+		cq_host->ops->dumpregs(mmc);
+	else
+		CQHCI_DUMP(": ===========================================\n");
+}
+
+/*
+ * The allocated descriptor table for task, link & transfer descritors
+ * looks like:
+ * |----------|
+ * |task desc |  |->|----------|
+ * |----------|  |  |trans desc|
+ * |link desc-|->|  |----------|
+ * |----------|          .
+ *      .                .
+ *  no. of slots      max-segs
+ *      .           |----------|
+ * |----------|
+ * The idea here is to create the [task+trans] table and mark & point the
+ * link desc to the transfer desc table on a per slot basis.
+ */
+static int cqhci_host_alloc_tdl(struct cqhci_host *cq_host)
+{
+	int i = 0;
+
+	/* task descriptor can be 64/128 bit irrespective of arch */
+	if (cq_host->caps & CQHCI_TASK_DESC_SZ_128) {
+		cqhci_writel(cq_host, cqhci_readl(cq_host, CQHCI_CFG) |
+			       CQHCI_TASK_DESC_SZ, CQHCI_CFG);
+		cq_host->task_desc_len = 16;
+	} else {
+		cq_host->task_desc_len = 8;
+	}
+
+	/*
+	 * 96 bits length of transfer desc instead of 128 bits which means
+	 * ADMA would expect next valid descriptor at the 96th bit
+	 * or 128th bit
+	 */
+	if (cq_host->dma64) {
+		if (cq_host->quirks & CQHCI_QUIRK_SHORT_TXFR_DESC_SZ)
+			cq_host->trans_desc_len = 12;
+		else
+			cq_host->trans_desc_len = 16;
+		cq_host->link_desc_len = 16;
+	} else {
+		cq_host->trans_desc_len = 8;
+		cq_host->link_desc_len = 8;
+	}
+
+	/* total size of a slot: 1 task & 1 transfer (link) */
+	cq_host->slot_sz = cq_host->task_desc_len + cq_host->link_desc_len;
+
+	cq_host->desc_size = cq_host->slot_sz * cq_host->num_slots;
+
+	cq_host->data_size = cq_host->trans_desc_len * cq_host->mmc->max_segs *
+		cq_host->mmc->cqe_qdepth;
+
+	pr_debug("%s: cqhci: desc_size: %zu data_sz: %zu slot-sz: %d\n",
+		 mmc_hostname(cq_host->mmc), cq_host->desc_size, cq_host->data_size,
+		 cq_host->slot_sz);
+
+	/*
+	 * allocate a dma-mapped chunk of memory for the descriptors
+	 * allocate a dma-mapped chunk of memory for link descriptors
+	 * setup each link-desc memory offset per slot-number to
+	 * the descriptor table.
+	 */
+	cq_host->desc_base = dmam_alloc_coherent(mmc_dev(cq_host->mmc),
+						 cq_host->desc_size,
+						 &cq_host->desc_dma_base,
+						 GFP_KERNEL);
+	if (!cq_host->desc_base)
+		return -ENOMEM;
+
+	cq_host->trans_desc_base = dmam_alloc_coherent(mmc_dev(cq_host->mmc),
+					      cq_host->data_size,
+					      &cq_host->trans_desc_dma_base,
+					      GFP_KERNEL);
+	if (!cq_host->trans_desc_base) {
+		dmam_free_coherent(mmc_dev(cq_host->mmc), cq_host->desc_size,
+				   cq_host->desc_base,
+				   cq_host->desc_dma_base);
+		cq_host->desc_base = NULL;
+		cq_host->desc_dma_base = 0;
+		return -ENOMEM;
+	}
+
+	pr_debug("%s: cqhci: desc-base: 0x%p trans-base: 0x%p\n desc_dma 0x%llx trans_dma: 0x%llx\n",
+		 mmc_hostname(cq_host->mmc), cq_host->desc_base, cq_host->trans_desc_base,
+		(unsigned long long)cq_host->desc_dma_base,
+		(unsigned long long)cq_host->trans_desc_dma_base);
+
+	for (; i < (cq_host->num_slots); i++)
+		setup_trans_desc(cq_host, i);
+
+	return 0;
+}
+
+static void __cqhci_enable(struct cqhci_host *cq_host)
+{
+	struct mmc_host *mmc = cq_host->mmc;
+	u32 cqcfg;
+
+	cqcfg = cqhci_readl(cq_host, CQHCI_CFG);
+
+	/* Configuration must not be changed while enabled */
+	if (cqcfg & CQHCI_ENABLE) {
+		cqcfg &= ~CQHCI_ENABLE;
+		cqhci_writel(cq_host, cqcfg, CQHCI_CFG);
+	}
+
+	cqcfg &= ~(CQHCI_DCMD | CQHCI_TASK_DESC_SZ);
+
+	if (mmc->caps2 & MMC_CAP2_CQE_DCMD)
+		cqcfg |= CQHCI_DCMD;
+
+	if (cq_host->caps & CQHCI_TASK_DESC_SZ_128)
+		cqcfg |= CQHCI_TASK_DESC_SZ;
+
+	cqhci_writel(cq_host, cqcfg, CQHCI_CFG);
+
+	cqhci_writel(cq_host, lower_32_bits(cq_host->desc_dma_base),
+		     CQHCI_TDLBA);
+	cqhci_writel(cq_host, upper_32_bits(cq_host->desc_dma_base),
+		     CQHCI_TDLBAU);
+
+	cqhci_writel(cq_host, cq_host->rca, CQHCI_SSC2);
+
+	cqhci_set_irqs(cq_host, 0);
+
+	cqcfg |= CQHCI_ENABLE;
+
+	cqhci_writel(cq_host, cqcfg, CQHCI_CFG);
+
+	if (cqhci_readl(cq_host, CQHCI_CTL) & CQHCI_HALT)
+		cqhci_writel(cq_host, 0, CQHCI_CTL);
+
+	mmc->cqe_on = true;
+
+	if (cq_host->ops->enable)
+		cq_host->ops->enable(mmc);
+
+	/* Ensure all writes are done before interrupts are enabled */
+	wmb();
+
+	cqhci_set_irqs(cq_host, CQHCI_IS_MASK);
+
+	cq_host->activated = true;
+}
+
+static void __cqhci_disable(struct cqhci_host *cq_host)
+{
+	u32 cqcfg;
+
+	cqcfg = cqhci_readl(cq_host, CQHCI_CFG);
+	cqcfg &= ~CQHCI_ENABLE;
+	cqhci_writel(cq_host, cqcfg, CQHCI_CFG);
+
+	cq_host->mmc->cqe_on = false;
+
+	cq_host->activated = false;
+}
+
+int cqhci_deactivate(struct mmc_host *mmc)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+
+	if (cq_host->enabled && cq_host->activated)
+		__cqhci_disable(cq_host);
+
+	return 0;
+}
+EXPORT_SYMBOL(cqhci_deactivate);
+
+int cqhci_resume(struct mmc_host *mmc)
+{
+	/* Re-enable is done upon first request */
+	return 0;
+}
+EXPORT_SYMBOL(cqhci_resume);
+
+static int cqhci_enable(struct mmc_host *mmc, struct mmc_card *card)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	int err;
+
+	if (!card->ext_csd.cmdq_en)
+		return -EINVAL;
+
+	if (cq_host->enabled)
+		return 0;
+
+	cq_host->rca = card->rca;
+
+	err = cqhci_host_alloc_tdl(cq_host);
+	if (err) {
+		pr_err("%s: Failed to enable CQE, error %d\n",
+		       mmc_hostname(mmc), err);
+		return err;
+	}
+
+	__cqhci_enable(cq_host);
+
+	cq_host->enabled = true;
+
+#ifdef DEBUG
+	cqhci_dumpregs(cq_host);
+#endif
+	return 0;
+}
+
+/* CQHCI is idle and should halt immediately, so set a small timeout */
+#define CQHCI_OFF_TIMEOUT 100
+
+static u32 cqhci_read_ctl(struct cqhci_host *cq_host)
+{
+	return cqhci_readl(cq_host, CQHCI_CTL);
+}
+
+static void cqhci_off(struct mmc_host *mmc)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	u32 reg;
+	int err;
+
+	if (!cq_host->enabled || !mmc->cqe_on || cq_host->recovery_halt)
+		return;
+
+	if (cq_host->ops->disable)
+		cq_host->ops->disable(mmc, false);
+
+	cqhci_writel(cq_host, CQHCI_HALT, CQHCI_CTL);
+
+	err = readx_poll_timeout(cqhci_read_ctl, cq_host, reg,
+				 reg & CQHCI_HALT, 0, CQHCI_OFF_TIMEOUT);
+	if (err < 0)
+		pr_err("%s: cqhci: CQE stuck on\n", mmc_hostname(mmc));
+	else
+		pr_debug("%s: cqhci: CQE off\n", mmc_hostname(mmc));
+
+	if (cq_host->ops->post_disable)
+		cq_host->ops->post_disable(mmc);
+
+	mmc->cqe_on = false;
+}
+
+static void cqhci_disable(struct mmc_host *mmc)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+
+	if (!cq_host->enabled)
+		return;
+
+	cqhci_off(mmc);
+
+	__cqhci_disable(cq_host);
+
+	dmam_free_coherent(mmc_dev(mmc), cq_host->data_size,
+			   cq_host->trans_desc_base,
+			   cq_host->trans_desc_dma_base);
+
+	dmam_free_coherent(mmc_dev(mmc), cq_host->desc_size,
+			   cq_host->desc_base,
+			   cq_host->desc_dma_base);
+
+	cq_host->trans_desc_base = NULL;
+	cq_host->desc_base = NULL;
+
+	cq_host->enabled = false;
+}
+
+static void cqhci_prep_task_desc(struct mmc_request *mrq,
+					u64 *data, bool intr)
+{
+	u32 req_flags = mrq->data->flags;
+
+	*data = CQHCI_VALID(1) |
+		CQHCI_END(1) |
+		CQHCI_INT(intr) |
+		CQHCI_ACT(0x5) |
+		CQHCI_FORCED_PROG(!!(req_flags & MMC_DATA_FORCED_PRG)) |
+		CQHCI_DATA_TAG(!!(req_flags & MMC_DATA_DAT_TAG)) |
+		CQHCI_DATA_DIR(!!(req_flags & MMC_DATA_READ)) |
+		CQHCI_PRIORITY(!!(req_flags & MMC_DATA_PRIO)) |
+		CQHCI_QBAR(!!(req_flags & MMC_DATA_QBR)) |
+		CQHCI_REL_WRITE(!!(req_flags & MMC_DATA_REL_WR)) |
+		CQHCI_BLK_COUNT(mrq->data->blocks) |
+		CQHCI_BLK_ADDR((u64)mrq->data->blk_addr);
+
+	pr_debug("%s: cqhci: tag %d task descriptor 0x%016llx\n",
+		 mmc_hostname(mrq->host), mrq->tag, (unsigned long long)*data);
+}
+
+static int cqhci_dma_map(struct mmc_host *host, struct mmc_request *mrq)
+{
+	int sg_count;
+	struct mmc_data *data = mrq->data;
+
+	if (!data)
+		return -EINVAL;
+
+	sg_count = dma_map_sg(mmc_dev(host), data->sg,
+			      data->sg_len,
+			      (data->flags & MMC_DATA_WRITE) ?
+			      DMA_TO_DEVICE : DMA_FROM_DEVICE);
+	if (!sg_count) {
+		pr_err("%s: sg-len: %d\n", __func__, data->sg_len);
+		return -ENOMEM;
+	}
+
+	return sg_count;
+}
+
+static void cqhci_set_tran_desc(u8 *desc, dma_addr_t addr, int len, bool end,
+				bool dma64)
+{
+	__le32 *attr = (__le32 __force *)desc;
+
+	*attr = (CQHCI_VALID(1) |
+		 CQHCI_END(end ? 1 : 0) |
+		 CQHCI_INT(0) |
+		 CQHCI_ACT(0x4) |
+		 CQHCI_DAT_LENGTH(len));
+
+	if (dma64) {
+		__le64 *dataddr = (__le64 __force *)(desc + 4);
+
+		dataddr[0] = cpu_to_le64(addr);
+	} else {
+		__le32 *dataddr = (__le32 __force *)(desc + 4);
+
+		dataddr[0] = cpu_to_le32(addr);
+	}
+}
+
+static int cqhci_prep_tran_desc(struct mmc_request *mrq,
+			       struct cqhci_host *cq_host, int tag)
+{
+	struct mmc_data *data = mrq->data;
+	int i, sg_count, len;
+	bool end = false;
+	bool dma64 = cq_host->dma64;
+	dma_addr_t addr;
+	u8 *desc;
+	struct scatterlist *sg;
+
+	sg_count = cqhci_dma_map(mrq->host, mrq);
+	if (sg_count < 0) {
+		pr_err("%s: %s: unable to map sg lists, %d\n",
+				mmc_hostname(mrq->host), __func__, sg_count);
+		return sg_count;
+	}
+
+	desc = get_trans_desc(cq_host, tag);
+
+	for_each_sg(data->sg, sg, sg_count, i) {
+		addr = sg_dma_address(sg);
+		len = sg_dma_len(sg);
+
+		if ((i+1) == sg_count)
+			end = true;
+		cqhci_set_tran_desc(desc, addr, len, end, dma64);
+		desc += cq_host->trans_desc_len;
+	}
+
+	return 0;
+}
+
+static void cqhci_prep_dcmd_desc(struct mmc_host *mmc,
+				   struct mmc_request *mrq)
+{
+	u64 *task_desc = NULL;
+	u64 data = 0;
+	u8 resp_type;
+	u8 *desc;
+	__le64 *dataddr;
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	u8 timing;
+
+	if (!(mrq->cmd->flags & MMC_RSP_PRESENT)) {
+		resp_type = 0x0;
+		timing = 0x1;
+	} else {
+		if (mrq->cmd->flags & MMC_RSP_R1B) {
+			resp_type = 0x3;
+			timing = 0x0;
+		} else {
+			resp_type = 0x2;
+			timing = 0x1;
+		}
+	}
+
+	task_desc = (__le64 __force *)get_desc(cq_host, cq_host->dcmd_slot);
+	memset(task_desc, 0, cq_host->task_desc_len);
+	data |= (CQHCI_VALID(1) |
+		 CQHCI_END(1) |
+		 CQHCI_INT(1) |
+		 CQHCI_QBAR(1) |
+		 CQHCI_ACT(0x5) |
+		 CQHCI_CMD_INDEX(mrq->cmd->opcode) |
+		 CQHCI_CMD_TIMING(timing) | CQHCI_RESP_TYPE(resp_type));
+	if (cq_host->ops->update_dcmd_desc)
+		cq_host->ops->update_dcmd_desc(mmc, mrq, &data);
+	*task_desc |= data;
+	desc = (u8 *)task_desc;
+	pr_debug("%s: cqhci: dcmd: cmd: %d timing: %d resp: %d\n",
+		 mmc_hostname(mmc), mrq->cmd->opcode, timing, resp_type);
+	dataddr = (__le64 __force *)(desc + 4);
+	dataddr[0] = cpu_to_le64((u64)mrq->cmd->arg);
+
+}
+
+static void cqhci_post_req(struct mmc_host *host, struct mmc_request *mrq)
+{
+	struct mmc_data *data = mrq->data;
+
+	if (data) {
+		dma_unmap_sg(mmc_dev(host), data->sg, data->sg_len,
+			     (data->flags & MMC_DATA_READ) ?
+			     DMA_FROM_DEVICE : DMA_TO_DEVICE);
+	}
+}
+
+static inline int cqhci_tag(struct mmc_request *mrq)
+{
+	return mrq->cmd ? DCMD_SLOT : mrq->tag;
+}
+
+static int cqhci_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	int err = 0;
+	u64 data = 0;
+	u64 *task_desc = NULL;
+	int tag = cqhci_tag(mrq);
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	unsigned long flags;
+
+	if (!cq_host->enabled) {
+		pr_err("%s: cqhci: not enabled\n", mmc_hostname(mmc));
+		return -EINVAL;
+	}
+
+	/* First request after resume has to re-enable */
+	if (!cq_host->activated)
+		__cqhci_enable(cq_host);
+
+	if (!mmc->cqe_on) {
+		if (cq_host->ops->pre_enable)
+			cq_host->ops->pre_enable(mmc);
+
+		cqhci_writel(cq_host, 0, CQHCI_CTL);
+		mmc->cqe_on = true;
+		pr_debug("%s: cqhci: CQE on\n", mmc_hostname(mmc));
+		if (cqhci_readl(cq_host, CQHCI_CTL) && CQHCI_HALT) {
+			pr_err("%s: cqhci: CQE failed to exit halt state\n",
+			       mmc_hostname(mmc));
+		}
+		if (cq_host->ops->enable)
+			cq_host->ops->enable(mmc);
+	}
+
+	if (mrq->data) {
+		task_desc = (__le64 __force *)get_desc(cq_host, tag);
+		cqhci_prep_task_desc(mrq, &data, 1);
+		*task_desc = cpu_to_le64(data);
+		err = cqhci_prep_tran_desc(mrq, cq_host, tag);
+		if (err) {
+			pr_err("%s: cqhci: failed to setup tx desc: %d\n",
+			       mmc_hostname(mmc), err);
+			return err;
+		}
+	} else {
+		cqhci_prep_dcmd_desc(mmc, mrq);
+	}
+
+	spin_lock_irqsave(&cq_host->lock, flags);
+
+	if (cq_host->recovery_halt) {
+		err = -EBUSY;
+		goto out_unlock;
+	}
+
+	cq_host->slot[tag].mrq = mrq;
+	cq_host->slot[tag].flags = 0;
+
+	cq_host->qcnt += 1;
+	/* Make sure descriptors are ready before ringing the doorbell */
+	wmb();
+	cqhci_writel(cq_host, 1 << tag, CQHCI_TDBR);
+	if (!(cqhci_readl(cq_host, CQHCI_TDBR) & (1 << tag)))
+		pr_debug("%s: cqhci: doorbell not set for tag %d\n",
+			 mmc_hostname(mmc), tag);
+out_unlock:
+	spin_unlock_irqrestore(&cq_host->lock, flags);
+
+	if (err)
+		cqhci_post_req(mmc, mrq);
+
+	return err;
+}
+
+static void cqhci_recovery_needed(struct mmc_host *mmc, struct mmc_request *mrq,
+				  bool notify)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+
+	if (!cq_host->recovery_halt) {
+		cq_host->recovery_halt = true;
+		pr_debug("%s: cqhci: recovery needed\n", mmc_hostname(mmc));
+		wake_up(&cq_host->wait_queue);
+		if (notify && mrq->recovery_notifier)
+			mrq->recovery_notifier(mrq);
+	}
+}
+
+static unsigned int cqhci_error_flags(int error1, int error2)
+{
+	int error = error1 ? error1 : error2;
+
+	switch (error) {
+	case -EILSEQ:
+		return CQHCI_HOST_CRC;
+	case -ETIMEDOUT:
+		return CQHCI_HOST_TIMEOUT;
+	default:
+		return CQHCI_HOST_OTHER;
+	}
+}
+
+static void cqhci_error_irq(struct mmc_host *mmc, u32 status, int cmd_error,
+			    int data_error)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	struct cqhci_slot *slot;
+	u32 terri;
+	int tag;
+
+	spin_lock(&cq_host->lock);
+
+	terri = cqhci_readl(cq_host, CQHCI_TERRI);
+
+	pr_debug("%s: cqhci: error IRQ status: 0x%08x cmd error %d data error %d TERRI: 0x%08x\n",
+		 mmc_hostname(mmc), status, cmd_error, data_error, terri);
+
+	/* Forget about errors when recovery has already been triggered */
+	if (cq_host->recovery_halt)
+		goto out_unlock;
+
+	if (!cq_host->qcnt) {
+		WARN_ONCE(1, "%s: cqhci: error when idle. IRQ status: 0x%08x cmd error %d data error %d TERRI: 0x%08x\n",
+			  mmc_hostname(mmc), status, cmd_error, data_error,
+			  terri);
+		goto out_unlock;
+	}
+
+	if (CQHCI_TERRI_C_VALID(terri)) {
+		tag = CQHCI_TERRI_C_TASK(terri);
+		slot = &cq_host->slot[tag];
+		if (slot->mrq) {
+			slot->flags = cqhci_error_flags(cmd_error, data_error);
+			cqhci_recovery_needed(mmc, slot->mrq, true);
+		}
+	}
+
+	if (CQHCI_TERRI_D_VALID(terri)) {
+		tag = CQHCI_TERRI_D_TASK(terri);
+		slot = &cq_host->slot[tag];
+		if (slot->mrq) {
+			slot->flags = cqhci_error_flags(data_error, cmd_error);
+			cqhci_recovery_needed(mmc, slot->mrq, true);
+		}
+	}
+
+	if (!cq_host->recovery_halt) {
+		/*
+		 * The only way to guarantee forward progress is to mark at
+		 * least one task in error, so if none is indicated, pick one.
+		 */
+		for (tag = 0; tag < NUM_SLOTS; tag++) {
+			slot = &cq_host->slot[tag];
+			if (!slot->mrq)
+				continue;
+			slot->flags = cqhci_error_flags(data_error, cmd_error);
+			cqhci_recovery_needed(mmc, slot->mrq, true);
+			break;
+		}
+	}
+
+out_unlock:
+	spin_unlock(&cq_host->lock);
+}
+
+static void cqhci_finish_mrq(struct mmc_host *mmc, unsigned int tag)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	struct cqhci_slot *slot = &cq_host->slot[tag];
+	struct mmc_request *mrq = slot->mrq;
+	struct mmc_data *data;
+
+	if (!mrq) {
+		WARN_ONCE(1, "%s: cqhci: spurious TCN for tag %d\n",
+			  mmc_hostname(mmc), tag);
+		return;
+	}
+
+	/* No completions allowed during recovery */
+	if (cq_host->recovery_halt) {
+		slot->flags |= CQHCI_COMPLETED;
+		return;
+	}
+
+	slot->mrq = NULL;
+
+	cq_host->qcnt -= 1;
+
+	data = mrq->data;
+	if (data) {
+		if (data->error)
+			data->bytes_xfered = 0;
+		else
+			data->bytes_xfered = data->blksz * data->blocks;
+	}
+
+	mmc_cqe_request_done(mmc, mrq);
+}
+
+irqreturn_t cqhci_irq(struct mmc_host *mmc, u32 intmask, int cmd_error,
+		      int data_error)
+{
+	u32 status;
+	unsigned long tag = 0, comp_status;
+	struct cqhci_host *cq_host = mmc->cqe_private;
+
+	status = cqhci_readl(cq_host, CQHCI_IS);
+	cqhci_writel(cq_host, status, CQHCI_IS);
+
+	pr_debug("%s: cqhci: IRQ status: 0x%08x\n", mmc_hostname(mmc), status);
+
+	if ((status & CQHCI_IS_RED) || cmd_error || data_error)
+		cqhci_error_irq(mmc, status, cmd_error, data_error);
+
+	if (status & CQHCI_IS_TCC) {
+		/* read TCN and complete the request */
+		comp_status = cqhci_readl(cq_host, CQHCI_TCN);
+		cqhci_writel(cq_host, comp_status, CQHCI_TCN);
+		pr_debug("%s: cqhci: TCN: 0x%08lx\n",
+			 mmc_hostname(mmc), comp_status);
+
+		spin_lock(&cq_host->lock);
+
+		for_each_set_bit(tag, &comp_status, cq_host->num_slots) {
+			/* complete the corresponding mrq */
+			pr_debug("%s: cqhci: completing tag %lu\n",
+				 mmc_hostname(mmc), tag);
+			cqhci_finish_mrq(mmc, tag);
+		}
+
+		if (cq_host->waiting_for_idle && !cq_host->qcnt) {
+			cq_host->waiting_for_idle = false;
+			wake_up(&cq_host->wait_queue);
+		}
+
+		spin_unlock(&cq_host->lock);
+	}
+
+	if (status & CQHCI_IS_TCL)
+		wake_up(&cq_host->wait_queue);
+
+	if (status & CQHCI_IS_HAC)
+		wake_up(&cq_host->wait_queue);
+
+	return IRQ_HANDLED;
+}
+EXPORT_SYMBOL(cqhci_irq);
+
+static bool cqhci_is_idle(struct cqhci_host *cq_host, int *ret)
+{
+	unsigned long flags;
+	bool is_idle;
+
+	spin_lock_irqsave(&cq_host->lock, flags);
+	is_idle = !cq_host->qcnt || cq_host->recovery_halt;
+	*ret = cq_host->recovery_halt ? -EBUSY : 0;
+	cq_host->waiting_for_idle = !is_idle;
+	spin_unlock_irqrestore(&cq_host->lock, flags);
+
+	return is_idle;
+}
+
+static int cqhci_wait_for_idle(struct mmc_host *mmc)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	int ret;
+
+	wait_event(cq_host->wait_queue, cqhci_is_idle(cq_host, &ret));
+
+	return ret;
+}
+
+static bool cqhci_timeout(struct mmc_host *mmc, struct mmc_request *mrq,
+			  bool *recovery_needed)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	int tag = cqhci_tag(mrq);
+	struct cqhci_slot *slot = &cq_host->slot[tag];
+	unsigned long flags;
+	bool timed_out;
+
+	spin_lock_irqsave(&cq_host->lock, flags);
+	timed_out = slot->mrq == mrq;
+	if (timed_out) {
+		slot->flags |= CQHCI_EXTERNAL_TIMEOUT;
+		cqhci_recovery_needed(mmc, mrq, false);
+		*recovery_needed = cq_host->recovery_halt;
+	}
+	spin_unlock_irqrestore(&cq_host->lock, flags);
+
+	if (timed_out) {
+		pr_err("%s: cqhci: timeout for tag %d\n",
+		       mmc_hostname(mmc), tag);
+		cqhci_dumpregs(cq_host);
+	}
+
+	return timed_out;
+}
+
+static bool cqhci_tasks_cleared(struct cqhci_host *cq_host)
+{
+	return !(cqhci_readl(cq_host, CQHCI_CTL) & CQHCI_CLEAR_ALL_TASKS);
+}
+
+static bool cqhci_clear_all_tasks(struct mmc_host *mmc, unsigned int timeout)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	bool ret;
+	u32 ctl;
+
+	cqhci_set_irqs(cq_host, CQHCI_IS_TCL);
+
+	ctl = cqhci_readl(cq_host, CQHCI_CTL);
+	ctl |= CQHCI_CLEAR_ALL_TASKS;
+	cqhci_writel(cq_host, ctl, CQHCI_CTL);
+
+	wait_event_timeout(cq_host->wait_queue, cqhci_tasks_cleared(cq_host),
+			   msecs_to_jiffies(timeout) + 1);
+
+	cqhci_set_irqs(cq_host, 0);
+
+	ret = cqhci_tasks_cleared(cq_host);
+
+	if (!ret)
+		pr_warn("%s: cqhci: Failed to clear tasks\n",
+			mmc_hostname(mmc));
+
+	return ret;
+}
+
+static bool cqhci_halted(struct cqhci_host *cq_host)
+{
+	return cqhci_readl(cq_host, CQHCI_CTL) & CQHCI_HALT;
+}
+
+static bool cqhci_halt(struct mmc_host *mmc, unsigned int timeout)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	bool ret;
+	u32 ctl;
+
+	if (cqhci_halted(cq_host))
+		return true;
+
+	cqhci_set_irqs(cq_host, CQHCI_IS_HAC);
+
+	ctl = cqhci_readl(cq_host, CQHCI_CTL);
+	ctl |= CQHCI_HALT;
+	cqhci_writel(cq_host, ctl, CQHCI_CTL);
+
+	wait_event_timeout(cq_host->wait_queue, cqhci_halted(cq_host),
+			   msecs_to_jiffies(timeout) + 1);
+
+	cqhci_set_irqs(cq_host, 0);
+
+	ret = cqhci_halted(cq_host);
+
+	if (!ret)
+		pr_warn("%s: cqhci: Failed to halt\n", mmc_hostname(mmc));
+
+	return ret;
+}
+
+/*
+ * After halting we expect to be able to use the command line. We interpret the
+ * failure to halt to mean the data lines might still be in use (and the upper
+ * layers will need to send a STOP command), however failing to halt complicates
+ * the recovery, so set a timeout that would reasonably allow I/O to complete.
+ */
+#define CQHCI_START_HALT_TIMEOUT	500
+
+static void cqhci_recovery_start(struct mmc_host *mmc)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+
+	pr_debug("%s: cqhci: %s\n", mmc_hostname(mmc), __func__);
+
+	WARN_ON(!cq_host->recovery_halt);
+
+	cqhci_halt(mmc, CQHCI_START_HALT_TIMEOUT);
+
+	if (cq_host->ops->disable)
+		cq_host->ops->disable(mmc, true);
+
+	mmc->cqe_on = false;
+}
+
+static int cqhci_error_from_flags(unsigned int flags)
+{
+	if (!flags)
+		return 0;
+
+	/* CRC errors might indicate re-tuning so prefer to report that */
+	if (flags & CQHCI_HOST_CRC)
+		return -EILSEQ;
+
+	if (flags & (CQHCI_EXTERNAL_TIMEOUT | CQHCI_HOST_TIMEOUT))
+		return -ETIMEDOUT;
+
+	return -EIO;
+}
+
+static void cqhci_recover_mrq(struct cqhci_host *cq_host, unsigned int tag)
+{
+	struct cqhci_slot *slot = &cq_host->slot[tag];
+	struct mmc_request *mrq = slot->mrq;
+	struct mmc_data *data;
+
+	if (!mrq)
+		return;
+
+	slot->mrq = NULL;
+
+	cq_host->qcnt -= 1;
+
+	data = mrq->data;
+	if (data) {
+		data->bytes_xfered = 0;
+		data->error = cqhci_error_from_flags(slot->flags);
+	} else {
+		mrq->cmd->error = cqhci_error_from_flags(slot->flags);
+	}
+
+	mmc_cqe_request_done(cq_host->mmc, mrq);
+}
+
+static void cqhci_recover_mrqs(struct cqhci_host *cq_host)
+{
+	int i;
+
+	for (i = 0; i < cq_host->num_slots; i++)
+		cqhci_recover_mrq(cq_host, i);
+}
+
+/*
+ * By now the command and data lines should be unused so there is no reason for
+ * CQHCI to take a long time to halt, but if it doesn't halt there could be
+ * problems clearing tasks, so be generous.
+ */
+#define CQHCI_FINISH_HALT_TIMEOUT	20
+
+/* CQHCI could be expected to clear it's internal state pretty quickly */
+#define CQHCI_CLEAR_TIMEOUT		20
+
+static void cqhci_recovery_finish(struct mmc_host *mmc)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	unsigned long flags;
+	u32 cqcfg;
+	bool ok;
+
+	pr_debug("%s: cqhci: %s\n", mmc_hostname(mmc), __func__);
+
+	WARN_ON(!cq_host->recovery_halt);
+
+	ok = cqhci_halt(mmc, CQHCI_FINISH_HALT_TIMEOUT);
+
+	/*
+	 * The specification contradicts itself, by saying that tasks cannot be
+	 * cleared if CQHCI does not halt, but if CQHCI does not halt, it should
+	 * be disabled/re-enabled, but not to disable before clearing tasks.
+	 * Have a go anyway.
+	 */
+	if (!cqhci_clear_all_tasks(mmc, CQHCI_CLEAR_TIMEOUT))
+		ok = false;
+
+	/* Disable to make sure tasks really are cleared */
+	cqcfg = cqhci_readl(cq_host, CQHCI_CFG);
+	cqcfg &= ~CQHCI_ENABLE;
+	cqhci_writel(cq_host, cqcfg, CQHCI_CFG);
+
+	cqcfg = cqhci_readl(cq_host, CQHCI_CFG);
+	cqcfg |= CQHCI_ENABLE;
+	cqhci_writel(cq_host, cqcfg, CQHCI_CFG);
+
+	cqhci_halt(mmc, CQHCI_FINISH_HALT_TIMEOUT);
+
+	if (!ok)
+		cqhci_clear_all_tasks(mmc, CQHCI_CLEAR_TIMEOUT);
+
+	cqhci_recover_mrqs(cq_host);
+
+	WARN_ON(cq_host->qcnt);
+
+	spin_lock_irqsave(&cq_host->lock, flags);
+	cq_host->qcnt = 0;
+	cq_host->recovery_halt = false;
+	mmc->cqe_on = false;
+	spin_unlock_irqrestore(&cq_host->lock, flags);
+
+	/* Ensure all writes are done before interrupts are re-enabled */
+	wmb();
+
+	cqhci_writel(cq_host, CQHCI_IS_HAC | CQHCI_IS_TCL, CQHCI_IS);
+
+	cqhci_set_irqs(cq_host, CQHCI_IS_MASK);
+
+	pr_debug("%s: cqhci: recovery done\n", mmc_hostname(mmc));
+}
+
+static const struct mmc_cqe_ops cqhci_cqe_ops = {
+	.cqe_enable = cqhci_enable,
+	.cqe_disable = cqhci_disable,
+	.cqe_request = cqhci_request,
+	.cqe_post_req = cqhci_post_req,
+	.cqe_off = cqhci_off,
+	.cqe_wait_for_idle = cqhci_wait_for_idle,
+	.cqe_timeout = cqhci_timeout,
+	.cqe_recovery_start = cqhci_recovery_start,
+	.cqe_recovery_finish = cqhci_recovery_finish,
+};
+
+struct cqhci_host *cqhci_pltfm_init(struct platform_device *pdev)
+{
+	struct cqhci_host *cq_host;
+	struct resource *cqhci_memres = NULL;
+
+	/* check and setup CMDQ interface */
+	cqhci_memres = platform_get_resource_byname(pdev, IORESOURCE_MEM,
+						   "cqhci");
+	if (!cqhci_memres) {
+		dev_dbg(&pdev->dev, "CMDQ not supported\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	cq_host = devm_kzalloc(&pdev->dev, sizeof(*cq_host), GFP_KERNEL);
+	if (!cq_host)
+		return ERR_PTR(-ENOMEM);
+	cq_host->mmio = devm_ioremap(&pdev->dev,
+				     cqhci_memres->start,
+				     resource_size(cqhci_memres));
+	if (!cq_host->mmio) {
+		dev_err(&pdev->dev, "failed to remap cqhci regs\n");
+		return ERR_PTR(-EBUSY);
+	}
+	dev_dbg(&pdev->dev, "CMDQ ioremap: done\n");
+
+	return cq_host;
+}
+EXPORT_SYMBOL(cqhci_pltfm_init);
+
+static unsigned int cqhci_ver_major(struct cqhci_host *cq_host)
+{
+	return CQHCI_VER_MAJOR(cqhci_readl(cq_host, CQHCI_VER));
+}
+
+static unsigned int cqhci_ver_minor(struct cqhci_host *cq_host)
+{
+	u32 ver = cqhci_readl(cq_host, CQHCI_VER);
+
+	return CQHCI_VER_MINOR1(ver) * 10 + CQHCI_VER_MINOR2(ver);
+}
+
+int cqhci_init(struct cqhci_host *cq_host, struct mmc_host *mmc,
+	      bool dma64)
+{
+	int err;
+
+	cq_host->dma64 = dma64;
+	cq_host->mmc = mmc;
+	cq_host->mmc->cqe_private = cq_host;
+
+	cq_host->num_slots = NUM_SLOTS;
+	cq_host->dcmd_slot = DCMD_SLOT;
+
+	mmc->cqe_ops = &cqhci_cqe_ops;
+
+	mmc->cqe_qdepth = NUM_SLOTS;
+	if (mmc->caps2 & MMC_CAP2_CQE_DCMD)
+		mmc->cqe_qdepth -= 1;
+
+	cq_host->slot = devm_kcalloc(mmc_dev(mmc), cq_host->num_slots,
+				     sizeof(*cq_host->slot), GFP_KERNEL);
+	if (!cq_host->slot) {
+		err = -ENOMEM;
+		goto out_err;
+	}
+
+	spin_lock_init(&cq_host->lock);
+
+	init_completion(&cq_host->halt_comp);
+	init_waitqueue_head(&cq_host->wait_queue);
+
+	pr_info("%s: CQHCI version %u.%02u\n",
+		mmc_hostname(mmc), cqhci_ver_major(cq_host),
+		cqhci_ver_minor(cq_host));
+
+	return 0;
+
+out_err:
+	pr_err("%s: CQHCI version %u.%02u failed to initialize, error %d\n",
+	       mmc_hostname(mmc), cqhci_ver_major(cq_host),
+	       cqhci_ver_minor(cq_host), err);
+	return err;
+}
+EXPORT_SYMBOL(cqhci_init);
+
+MODULE_AUTHOR("Venkat Gopalakrishnan <venkatg@codeaurora.org>");
+MODULE_DESCRIPTION("Command Queue Host Controller Interface driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/cqhci.h b/drivers/mmc/host/cqhci.h
new file mode 100644
index 000000000..89bf6adbc
--- /dev/null
+++ b/drivers/mmc/host/cqhci.h
@@ -0,0 +1,242 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/* Copyright (c) 2015, The Linux Foundation. All rights reserved.
+ */
+#ifndef LINUX_MMC_CQHCI_H
+#define LINUX_MMC_CQHCI_H
+
+#include <linux/compiler.h>
+#include <linux/bitops.h>
+#include <linux/spinlock_types.h>
+#include <linux/types.h>
+#include <linux/completion.h>
+#include <linux/wait.h>
+#include <linux/irqreturn.h>
+#include <asm/io.h>
+
+/* registers */
+/* version */
+#define CQHCI_VER			0x00
+#define CQHCI_VER_MAJOR(x)		(((x) & GENMASK(11, 8)) >> 8)
+#define CQHCI_VER_MINOR1(x)		(((x) & GENMASK(7, 4)) >> 4)
+#define CQHCI_VER_MINOR2(x)		((x) & GENMASK(3, 0))
+
+/* capabilities */
+#define CQHCI_CAP			0x04
+/* configuration */
+#define CQHCI_CFG			0x08
+#define CQHCI_DCMD			0x00001000
+#define CQHCI_TASK_DESC_SZ		0x00000100
+#define CQHCI_ENABLE			0x00000001
+
+/* control */
+#define CQHCI_CTL			0x0C
+#define CQHCI_CLEAR_ALL_TASKS		0x00000100
+#define CQHCI_HALT			0x00000001
+
+/* interrupt status */
+#define CQHCI_IS			0x10
+#define CQHCI_IS_HAC			BIT(0)
+#define CQHCI_IS_TCC			BIT(1)
+#define CQHCI_IS_RED			BIT(2)
+#define CQHCI_IS_TCL			BIT(3)
+
+#define CQHCI_IS_MASK (CQHCI_IS_TCC | CQHCI_IS_RED)
+
+/* interrupt status enable */
+#define CQHCI_ISTE			0x14
+
+/* interrupt signal enable */
+#define CQHCI_ISGE			0x18
+
+/* interrupt coalescing */
+#define CQHCI_IC			0x1C
+#define CQHCI_IC_ENABLE			BIT(31)
+#define CQHCI_IC_RESET			BIT(16)
+#define CQHCI_IC_ICCTHWEN		BIT(15)
+#define CQHCI_IC_ICCTH(x)		(((x) & 0x1F) << 8)
+#define CQHCI_IC_ICTOVALWEN		BIT(7)
+#define CQHCI_IC_ICTOVAL(x)		((x) & 0x7F)
+
+/* task list base address */
+#define CQHCI_TDLBA			0x20
+
+/* task list base address upper */
+#define CQHCI_TDLBAU			0x24
+
+/* door-bell */
+#define CQHCI_TDBR			0x28
+
+/* task completion notification */
+#define CQHCI_TCN			0x2C
+
+/* device queue status */
+#define CQHCI_DQS			0x30
+
+/* device pending tasks */
+#define CQHCI_DPT			0x34
+
+/* task clear */
+#define CQHCI_TCLR			0x38
+
+/* send status config 1 */
+#define CQHCI_SSC1			0x40
+#define CQHCI_SSC1_CBC_MASK		GENMASK(19, 16)
+
+/* send status config 2 */
+#define CQHCI_SSC2			0x44
+
+/* response for dcmd */
+#define CQHCI_CRDCT			0x48
+
+/* response mode error mask */
+#define CQHCI_RMEM			0x50
+
+/* task error info */
+#define CQHCI_TERRI			0x54
+
+#define CQHCI_TERRI_C_INDEX(x)		((x) & GENMASK(5, 0))
+#define CQHCI_TERRI_C_TASK(x)		(((x) & GENMASK(12, 8)) >> 8)
+#define CQHCI_TERRI_C_VALID(x)		((x) & BIT(15))
+#define CQHCI_TERRI_D_INDEX(x)		(((x) & GENMASK(21, 16)) >> 16)
+#define CQHCI_TERRI_D_TASK(x)		(((x) & GENMASK(28, 24)) >> 24)
+#define CQHCI_TERRI_D_VALID(x)		((x) & BIT(31))
+
+/* command response index */
+#define CQHCI_CRI			0x58
+
+/* command response argument */
+#define CQHCI_CRA			0x5C
+
+#define CQHCI_INT_ALL			0xF
+#define CQHCI_IC_DEFAULT_ICCTH		31
+#define CQHCI_IC_DEFAULT_ICTOVAL	1
+
+/* attribute fields */
+#define CQHCI_VALID(x)			(((x) & 1) << 0)
+#define CQHCI_END(x)			(((x) & 1) << 1)
+#define CQHCI_INT(x)			(((x) & 1) << 2)
+#define CQHCI_ACT(x)			(((x) & 0x7) << 3)
+
+/* data command task descriptor fields */
+#define CQHCI_FORCED_PROG(x)		(((x) & 1) << 6)
+#define CQHCI_CONTEXT(x)		(((x) & 0xF) << 7)
+#define CQHCI_DATA_TAG(x)		(((x) & 1) << 11)
+#define CQHCI_DATA_DIR(x)		(((x) & 1) << 12)
+#define CQHCI_PRIORITY(x)		(((x) & 1) << 13)
+#define CQHCI_QBAR(x)			(((x) & 1) << 14)
+#define CQHCI_REL_WRITE(x)		(((x) & 1) << 15)
+#define CQHCI_BLK_COUNT(x)		(((x) & 0xFFFF) << 16)
+#define CQHCI_BLK_ADDR(x)		(((x) & 0xFFFFFFFF) << 32)
+
+/* direct command task descriptor fields */
+#define CQHCI_CMD_INDEX(x)		(((x) & 0x3F) << 16)
+#define CQHCI_CMD_TIMING(x)		(((x) & 1) << 22)
+#define CQHCI_RESP_TYPE(x)		(((x) & 0x3) << 23)
+
+/* transfer descriptor fields */
+#define CQHCI_DAT_LENGTH(x)		(((x) & 0xFFFF) << 16)
+#define CQHCI_DAT_ADDR_LO(x)		(((x) & 0xFFFFFFFF) << 32)
+#define CQHCI_DAT_ADDR_HI(x)		(((x) & 0xFFFFFFFF) << 0)
+
+struct cqhci_host_ops;
+struct mmc_host;
+struct mmc_request;
+struct cqhci_slot;
+
+struct cqhci_host {
+	const struct cqhci_host_ops *ops;
+	void __iomem *mmio;
+	struct mmc_host *mmc;
+
+	spinlock_t lock;
+
+	/* relative card address of device */
+	unsigned int rca;
+
+	/* 64 bit DMA */
+	bool dma64;
+	int num_slots;
+	int qcnt;
+
+	u32 dcmd_slot;
+	u32 caps;
+#define CQHCI_TASK_DESC_SZ_128		0x1
+
+	u32 quirks;
+#define CQHCI_QUIRK_SHORT_TXFR_DESC_SZ	0x1
+
+	bool enabled;
+	bool halted;
+	bool init_done;
+	bool activated;
+	bool waiting_for_idle;
+	bool recovery_halt;
+
+	size_t desc_size;
+	size_t data_size;
+
+	u8 *desc_base;
+
+	/* total descriptor size */
+	u8 slot_sz;
+
+	/* 64/128 bit depends on CQHCI_CFG */
+	u8 task_desc_len;
+
+	/* 64 bit on 32-bit arch, 128 bit on 64-bit */
+	u8 link_desc_len;
+
+	u8 *trans_desc_base;
+	/* same length as transfer descriptor */
+	u8 trans_desc_len;
+
+	dma_addr_t desc_dma_base;
+	dma_addr_t trans_desc_dma_base;
+
+	struct completion halt_comp;
+	wait_queue_head_t wait_queue;
+	struct cqhci_slot *slot;
+};
+
+struct cqhci_host_ops {
+	void (*dumpregs)(struct mmc_host *mmc);
+	void (*write_l)(struct cqhci_host *host, u32 val, int reg);
+	u32 (*read_l)(struct cqhci_host *host, int reg);
+	void (*enable)(struct mmc_host *mmc);
+	void (*disable)(struct mmc_host *mmc, bool recovery);
+	void (*update_dcmd_desc)(struct mmc_host *mmc, struct mmc_request *mrq,
+				 u64 *data);
+	void (*pre_enable)(struct mmc_host *mmc);
+	void (*post_disable)(struct mmc_host *mmc);
+};
+
+static inline void cqhci_writel(struct cqhci_host *host, u32 val, int reg)
+{
+	if (unlikely(host->ops->write_l))
+		host->ops->write_l(host, val, reg);
+	else
+		writel_relaxed(val, host->mmio + reg);
+}
+
+static inline u32 cqhci_readl(struct cqhci_host *host, int reg)
+{
+	if (unlikely(host->ops->read_l))
+		return host->ops->read_l(host, reg);
+	else
+		return readl_relaxed(host->mmio + reg);
+}
+
+struct platform_device;
+
+irqreturn_t cqhci_irq(struct mmc_host *mmc, u32 intmask, int cmd_error,
+		      int data_error);
+int cqhci_init(struct cqhci_host *cq_host, struct mmc_host *mmc, bool dma64);
+struct cqhci_host *cqhci_pltfm_init(struct platform_device *pdev);
+int cqhci_deactivate(struct mmc_host *mmc);
+static inline int cqhci_suspend(struct mmc_host *mmc)
+{
+	return cqhci_deactivate(mmc);
+}
+int cqhci_resume(struct mmc_host *mmc);
+
+#endif
diff --git a/drivers/mmc/host/davinci_mmc.c b/drivers/mmc/host/davinci_mmc.c
new file mode 100644
index 000000000..647928ab0
--- /dev/null
+++ b/drivers/mmc/host/davinci_mmc.c
@@ -0,0 +1,1417 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * davinci_mmc.c - TI DaVinci MMC/SD/SDIO driver
+ *
+ * Copyright (C) 2006 Texas Instruments.
+ *       Original author: Purushotam Kumar
+ * Copyright (C) 2009 David Brownell
+ */
+
+#include <linux/module.h>
+#include <linux/ioport.h>
+#include <linux/platform_device.h>
+#include <linux/clk.h>
+#include <linux/err.h>
+#include <linux/cpufreq.h>
+#include <linux/mmc/host.h>
+#include <linux/io.h>
+#include <linux/irq.h>
+#include <linux/delay.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/mmc/mmc.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/interrupt.h>
+
+#include <linux/platform_data/mmc-davinci.h>
+
+/*
+ * Register Definitions
+ */
+#define DAVINCI_MMCCTL       0x00 /* Control Register                  */
+#define DAVINCI_MMCCLK       0x04 /* Memory Clock Control Register     */
+#define DAVINCI_MMCST0       0x08 /* Status Register 0                 */
+#define DAVINCI_MMCST1       0x0C /* Status Register 1                 */
+#define DAVINCI_MMCIM        0x10 /* Interrupt Mask Register           */
+#define DAVINCI_MMCTOR       0x14 /* Response Time-Out Register        */
+#define DAVINCI_MMCTOD       0x18 /* Data Read Time-Out Register       */
+#define DAVINCI_MMCBLEN      0x1C /* Block Length Register             */
+#define DAVINCI_MMCNBLK      0x20 /* Number of Blocks Register         */
+#define DAVINCI_MMCNBLC      0x24 /* Number of Blocks Counter Register */
+#define DAVINCI_MMCDRR       0x28 /* Data Receive Register             */
+#define DAVINCI_MMCDXR       0x2C /* Data Transmit Register            */
+#define DAVINCI_MMCCMD       0x30 /* Command Register                  */
+#define DAVINCI_MMCARGHL     0x34 /* Argument Register                 */
+#define DAVINCI_MMCRSP01     0x38 /* Response Register 0 and 1         */
+#define DAVINCI_MMCRSP23     0x3C /* Response Register 0 and 1         */
+#define DAVINCI_MMCRSP45     0x40 /* Response Register 0 and 1         */
+#define DAVINCI_MMCRSP67     0x44 /* Response Register 0 and 1         */
+#define DAVINCI_MMCDRSP      0x48 /* Data Response Register            */
+#define DAVINCI_MMCETOK      0x4C
+#define DAVINCI_MMCCIDX      0x50 /* Command Index Register            */
+#define DAVINCI_MMCCKC       0x54
+#define DAVINCI_MMCTORC      0x58
+#define DAVINCI_MMCTODC      0x5C
+#define DAVINCI_MMCBLNC      0x60
+#define DAVINCI_SDIOCTL      0x64
+#define DAVINCI_SDIOST0      0x68
+#define DAVINCI_SDIOIEN      0x6C
+#define DAVINCI_SDIOIST      0x70
+#define DAVINCI_MMCFIFOCTL   0x74 /* FIFO Control Register             */
+
+/* DAVINCI_MMCCTL definitions */
+#define MMCCTL_DATRST         (1 << 0)
+#define MMCCTL_CMDRST         (1 << 1)
+#define MMCCTL_WIDTH_8_BIT    (1 << 8)
+#define MMCCTL_WIDTH_4_BIT    (1 << 2)
+#define MMCCTL_DATEG_DISABLED (0 << 6)
+#define MMCCTL_DATEG_RISING   (1 << 6)
+#define MMCCTL_DATEG_FALLING  (2 << 6)
+#define MMCCTL_DATEG_BOTH     (3 << 6)
+#define MMCCTL_PERMDR_LE      (0 << 9)
+#define MMCCTL_PERMDR_BE      (1 << 9)
+#define MMCCTL_PERMDX_LE      (0 << 10)
+#define MMCCTL_PERMDX_BE      (1 << 10)
+
+/* DAVINCI_MMCCLK definitions */
+#define MMCCLK_CLKEN          (1 << 8)
+#define MMCCLK_CLKRT_MASK     (0xFF << 0)
+
+/* IRQ bit definitions, for DAVINCI_MMCST0 and DAVINCI_MMCIM */
+#define MMCST0_DATDNE         BIT(0)	/* data done */
+#define MMCST0_BSYDNE         BIT(1)	/* busy done */
+#define MMCST0_RSPDNE         BIT(2)	/* command done */
+#define MMCST0_TOUTRD         BIT(3)	/* data read timeout */
+#define MMCST0_TOUTRS         BIT(4)	/* command response timeout */
+#define MMCST0_CRCWR          BIT(5)	/* data write CRC error */
+#define MMCST0_CRCRD          BIT(6)	/* data read CRC error */
+#define MMCST0_CRCRS          BIT(7)	/* command response CRC error */
+#define MMCST0_DXRDY          BIT(9)	/* data transmit ready (fifo empty) */
+#define MMCST0_DRRDY          BIT(10)	/* data receive ready (data in fifo)*/
+#define MMCST0_DATED          BIT(11)	/* DAT3 edge detect */
+#define MMCST0_TRNDNE         BIT(12)	/* transfer done */
+
+/* DAVINCI_MMCST1 definitions */
+#define MMCST1_BUSY           (1 << 0)
+
+/* DAVINCI_MMCCMD definitions */
+#define MMCCMD_CMD_MASK       (0x3F << 0)
+#define MMCCMD_PPLEN          (1 << 7)
+#define MMCCMD_BSYEXP         (1 << 8)
+#define MMCCMD_RSPFMT_MASK    (3 << 9)
+#define MMCCMD_RSPFMT_NONE    (0 << 9)
+#define MMCCMD_RSPFMT_R1456   (1 << 9)
+#define MMCCMD_RSPFMT_R2      (2 << 9)
+#define MMCCMD_RSPFMT_R3      (3 << 9)
+#define MMCCMD_DTRW           (1 << 11)
+#define MMCCMD_STRMTP         (1 << 12)
+#define MMCCMD_WDATX          (1 << 13)
+#define MMCCMD_INITCK         (1 << 14)
+#define MMCCMD_DCLR           (1 << 15)
+#define MMCCMD_DMATRIG        (1 << 16)
+
+/* DAVINCI_MMCFIFOCTL definitions */
+#define MMCFIFOCTL_FIFORST    (1 << 0)
+#define MMCFIFOCTL_FIFODIR_WR (1 << 1)
+#define MMCFIFOCTL_FIFODIR_RD (0 << 1)
+#define MMCFIFOCTL_FIFOLEV    (1 << 2) /* 0 = 128 bits, 1 = 256 bits */
+#define MMCFIFOCTL_ACCWD_4    (0 << 3) /* access width of 4 bytes    */
+#define MMCFIFOCTL_ACCWD_3    (1 << 3) /* access width of 3 bytes    */
+#define MMCFIFOCTL_ACCWD_2    (2 << 3) /* access width of 2 bytes    */
+#define MMCFIFOCTL_ACCWD_1    (3 << 3) /* access width of 1 byte     */
+
+/* DAVINCI_SDIOST0 definitions */
+#define SDIOST0_DAT1_HI       BIT(0)
+
+/* DAVINCI_SDIOIEN definitions */
+#define SDIOIEN_IOINTEN       BIT(0)
+
+/* DAVINCI_SDIOIST definitions */
+#define SDIOIST_IOINT         BIT(0)
+
+/* MMCSD Init clock in Hz in opendrain mode */
+#define MMCSD_INIT_CLOCK		200000
+
+/*
+ * One scatterlist dma "segment" is at most MAX_CCNT rw_threshold units,
+ * and we handle up to MAX_NR_SG segments.  MMC_BLOCK_BOUNCE kicks in only
+ * for drivers with max_segs == 1, making the segments bigger (64KB)
+ * than the page or two that's otherwise typical. nr_sg (passed from
+ * platform data) == 16 gives at least the same throughput boost, using
+ * EDMA transfer linkage instead of spending CPU time copying pages.
+ */
+#define MAX_CCNT	((1 << 16) - 1)
+
+#define MAX_NR_SG	16
+
+static unsigned rw_threshold = 32;
+module_param(rw_threshold, uint, S_IRUGO);
+MODULE_PARM_DESC(rw_threshold,
+		"Read/Write threshold. Default = 32");
+
+static unsigned poll_threshold = 128;
+module_param(poll_threshold, uint, S_IRUGO);
+MODULE_PARM_DESC(poll_threshold,
+		 "Polling transaction size threshold. Default = 128");
+
+static unsigned poll_loopcount = 32;
+module_param(poll_loopcount, uint, S_IRUGO);
+MODULE_PARM_DESC(poll_loopcount,
+		 "Maximum polling loop count. Default = 32");
+
+static unsigned use_dma = 1;
+module_param(use_dma, uint, 0);
+MODULE_PARM_DESC(use_dma, "Whether to use DMA or not. Default = 1");
+
+struct mmc_davinci_host {
+	struct mmc_command *cmd;
+	struct mmc_data *data;
+	struct mmc_host *mmc;
+	struct clk *clk;
+	unsigned int mmc_input_clk;
+	void __iomem *base;
+	struct resource *mem_res;
+	int mmc_irq, sdio_irq;
+	unsigned char bus_mode;
+
+#define DAVINCI_MMC_DATADIR_NONE	0
+#define DAVINCI_MMC_DATADIR_READ	1
+#define DAVINCI_MMC_DATADIR_WRITE	2
+	unsigned char data_dir;
+
+	/* buffer is used during PIO of one scatterlist segment, and
+	 * is updated along with buffer_bytes_left.  bytes_left applies
+	 * to all N blocks of the PIO transfer.
+	 */
+	u8 *buffer;
+	u32 buffer_bytes_left;
+	u32 bytes_left;
+
+	struct dma_chan *dma_tx;
+	struct dma_chan *dma_rx;
+	bool use_dma;
+	bool do_dma;
+	bool sdio_int;
+	bool active_request;
+
+	/* For PIO we walk scatterlists one segment at a time. */
+	unsigned int		sg_len;
+	struct scatterlist *sg;
+
+	/* Version of the MMC/SD controller */
+	u8 version;
+	/* for ns in one cycle calculation */
+	unsigned ns_in_one_cycle;
+	/* Number of sg segments */
+	u8 nr_sg;
+#ifdef CONFIG_CPU_FREQ
+	struct notifier_block	freq_transition;
+#endif
+};
+
+static irqreturn_t mmc_davinci_irq(int irq, void *dev_id);
+
+/* PIO only */
+static void mmc_davinci_sg_to_buf(struct mmc_davinci_host *host)
+{
+	host->buffer_bytes_left = sg_dma_len(host->sg);
+	host->buffer = sg_virt(host->sg);
+	if (host->buffer_bytes_left > host->bytes_left)
+		host->buffer_bytes_left = host->bytes_left;
+}
+
+static void davinci_fifo_data_trans(struct mmc_davinci_host *host,
+					unsigned int n)
+{
+	u8 *p;
+	unsigned int i;
+
+	if (host->buffer_bytes_left == 0) {
+		host->sg = sg_next(host->data->sg);
+		mmc_davinci_sg_to_buf(host);
+	}
+
+	p = host->buffer;
+	if (n > host->buffer_bytes_left)
+		n = host->buffer_bytes_left;
+	host->buffer_bytes_left -= n;
+	host->bytes_left -= n;
+
+	/* NOTE:  we never transfer more than rw_threshold bytes
+	 * to/from the fifo here; there's no I/O overlap.
+	 * This also assumes that access width( i.e. ACCWD) is 4 bytes
+	 */
+	if (host->data_dir == DAVINCI_MMC_DATADIR_WRITE) {
+		for (i = 0; i < (n >> 2); i++) {
+			writel(*((u32 *)p), host->base + DAVINCI_MMCDXR);
+			p = p + 4;
+		}
+		if (n & 3) {
+			iowrite8_rep(host->base + DAVINCI_MMCDXR, p, (n & 3));
+			p = p + (n & 3);
+		}
+	} else {
+		for (i = 0; i < (n >> 2); i++) {
+			*((u32 *)p) = readl(host->base + DAVINCI_MMCDRR);
+			p  = p + 4;
+		}
+		if (n & 3) {
+			ioread8_rep(host->base + DAVINCI_MMCDRR, p, (n & 3));
+			p = p + (n & 3);
+		}
+	}
+	host->buffer = p;
+}
+
+static void mmc_davinci_start_command(struct mmc_davinci_host *host,
+		struct mmc_command *cmd)
+{
+	u32 cmd_reg = 0;
+	u32 im_val;
+
+	dev_dbg(mmc_dev(host->mmc), "CMD%d, arg 0x%08x%s\n",
+		cmd->opcode, cmd->arg,
+		({ char *s;
+		switch (mmc_resp_type(cmd)) {
+		case MMC_RSP_R1:
+			s = ", R1/R5/R6/R7 response";
+			break;
+		case MMC_RSP_R1B:
+			s = ", R1b response";
+			break;
+		case MMC_RSP_R2:
+			s = ", R2 response";
+			break;
+		case MMC_RSP_R3:
+			s = ", R3/R4 response";
+			break;
+		default:
+			s = ", (R? response)";
+			break;
+		}; s; }));
+	host->cmd = cmd;
+
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_R1B:
+		/* There's some spec confusion about when R1B is
+		 * allowed, but if the card doesn't issue a BUSY
+		 * then it's harmless for us to allow it.
+		 */
+		cmd_reg |= MMCCMD_BSYEXP;
+		fallthrough;
+	case MMC_RSP_R1:		/* 48 bits, CRC */
+		cmd_reg |= MMCCMD_RSPFMT_R1456;
+		break;
+	case MMC_RSP_R2:		/* 136 bits, CRC */
+		cmd_reg |= MMCCMD_RSPFMT_R2;
+		break;
+	case MMC_RSP_R3:		/* 48 bits, no CRC */
+		cmd_reg |= MMCCMD_RSPFMT_R3;
+		break;
+	default:
+		cmd_reg |= MMCCMD_RSPFMT_NONE;
+		dev_dbg(mmc_dev(host->mmc), "unknown resp_type %04x\n",
+			mmc_resp_type(cmd));
+		break;
+	}
+
+	/* Set command index */
+	cmd_reg |= cmd->opcode;
+
+	/* Enable EDMA transfer triggers */
+	if (host->do_dma)
+		cmd_reg |= MMCCMD_DMATRIG;
+
+	if (host->version == MMC_CTLR_VERSION_2 && host->data != NULL &&
+			host->data_dir == DAVINCI_MMC_DATADIR_READ)
+		cmd_reg |= MMCCMD_DMATRIG;
+
+	/* Setting whether command involves data transfer or not */
+	if (cmd->data)
+		cmd_reg |= MMCCMD_WDATX;
+
+	/* Setting whether data read or write */
+	if (host->data_dir == DAVINCI_MMC_DATADIR_WRITE)
+		cmd_reg |= MMCCMD_DTRW;
+
+	if (host->bus_mode == MMC_BUSMODE_PUSHPULL)
+		cmd_reg |= MMCCMD_PPLEN;
+
+	/* set Command timeout */
+	writel(0x1FFF, host->base + DAVINCI_MMCTOR);
+
+	/* Enable interrupt (calculate here, defer until FIFO is stuffed). */
+	im_val =  MMCST0_RSPDNE | MMCST0_CRCRS | MMCST0_TOUTRS;
+	if (host->data_dir == DAVINCI_MMC_DATADIR_WRITE) {
+		im_val |= MMCST0_DATDNE | MMCST0_CRCWR;
+
+		if (!host->do_dma)
+			im_val |= MMCST0_DXRDY;
+	} else if (host->data_dir == DAVINCI_MMC_DATADIR_READ) {
+		im_val |= MMCST0_DATDNE | MMCST0_CRCRD | MMCST0_TOUTRD;
+
+		if (!host->do_dma)
+			im_val |= MMCST0_DRRDY;
+	}
+
+	/*
+	 * Before non-DMA WRITE commands the controller needs priming:
+	 * FIFO should be populated with 32 bytes i.e. whatever is the FIFO size
+	 */
+	if (!host->do_dma && (host->data_dir == DAVINCI_MMC_DATADIR_WRITE))
+		davinci_fifo_data_trans(host, rw_threshold);
+
+	writel(cmd->arg, host->base + DAVINCI_MMCARGHL);
+	writel(cmd_reg,  host->base + DAVINCI_MMCCMD);
+
+	host->active_request = true;
+
+	if (!host->do_dma && host->bytes_left <= poll_threshold) {
+		u32 count = poll_loopcount;
+
+		while (host->active_request && count--) {
+			mmc_davinci_irq(0, host);
+			cpu_relax();
+		}
+	}
+
+	if (host->active_request)
+		writel(im_val, host->base + DAVINCI_MMCIM);
+}
+
+/*----------------------------------------------------------------------*/
+
+/* DMA infrastructure */
+
+static void davinci_abort_dma(struct mmc_davinci_host *host)
+{
+	struct dma_chan *sync_dev;
+
+	if (host->data_dir == DAVINCI_MMC_DATADIR_READ)
+		sync_dev = host->dma_rx;
+	else
+		sync_dev = host->dma_tx;
+
+	dmaengine_terminate_all(sync_dev);
+}
+
+static int mmc_davinci_send_dma_request(struct mmc_davinci_host *host,
+		struct mmc_data *data)
+{
+	struct dma_chan *chan;
+	struct dma_async_tx_descriptor *desc;
+	int ret = 0;
+
+	if (host->data_dir == DAVINCI_MMC_DATADIR_WRITE) {
+		struct dma_slave_config dma_tx_conf = {
+			.direction = DMA_MEM_TO_DEV,
+			.dst_addr = host->mem_res->start + DAVINCI_MMCDXR,
+			.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+			.dst_maxburst =
+				rw_threshold / DMA_SLAVE_BUSWIDTH_4_BYTES,
+		};
+		chan = host->dma_tx;
+		dmaengine_slave_config(host->dma_tx, &dma_tx_conf);
+
+		desc = dmaengine_prep_slave_sg(host->dma_tx,
+				data->sg,
+				host->sg_len,
+				DMA_MEM_TO_DEV,
+				DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+		if (!desc) {
+			dev_dbg(mmc_dev(host->mmc),
+				"failed to allocate DMA TX descriptor");
+			ret = -1;
+			goto out;
+		}
+	} else {
+		struct dma_slave_config dma_rx_conf = {
+			.direction = DMA_DEV_TO_MEM,
+			.src_addr = host->mem_res->start + DAVINCI_MMCDRR,
+			.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+			.src_maxburst =
+				rw_threshold / DMA_SLAVE_BUSWIDTH_4_BYTES,
+		};
+		chan = host->dma_rx;
+		dmaengine_slave_config(host->dma_rx, &dma_rx_conf);
+
+		desc = dmaengine_prep_slave_sg(host->dma_rx,
+				data->sg,
+				host->sg_len,
+				DMA_DEV_TO_MEM,
+				DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+		if (!desc) {
+			dev_dbg(mmc_dev(host->mmc),
+				"failed to allocate DMA RX descriptor");
+			ret = -1;
+			goto out;
+		}
+	}
+
+	dmaengine_submit(desc);
+	dma_async_issue_pending(chan);
+
+out:
+	return ret;
+}
+
+static int mmc_davinci_start_dma_transfer(struct mmc_davinci_host *host,
+		struct mmc_data *data)
+{
+	int i;
+	int mask = rw_threshold - 1;
+	int ret = 0;
+
+	host->sg_len = dma_map_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+				  mmc_get_dma_dir(data));
+
+	/* no individual DMA segment should need a partial FIFO */
+	for (i = 0; i < host->sg_len; i++) {
+		if (sg_dma_len(data->sg + i) & mask) {
+			dma_unmap_sg(mmc_dev(host->mmc),
+				     data->sg, data->sg_len,
+				     mmc_get_dma_dir(data));
+			return -1;
+		}
+	}
+
+	host->do_dma = 1;
+	ret = mmc_davinci_send_dma_request(host, data);
+
+	return ret;
+}
+
+static void davinci_release_dma_channels(struct mmc_davinci_host *host)
+{
+	if (!host->use_dma)
+		return;
+
+	dma_release_channel(host->dma_tx);
+	dma_release_channel(host->dma_rx);
+}
+
+static int davinci_acquire_dma_channels(struct mmc_davinci_host *host)
+{
+	host->dma_tx = dma_request_chan(mmc_dev(host->mmc), "tx");
+	if (IS_ERR(host->dma_tx)) {
+		dev_err(mmc_dev(host->mmc), "Can't get dma_tx channel\n");
+		return PTR_ERR(host->dma_tx);
+	}
+
+	host->dma_rx = dma_request_chan(mmc_dev(host->mmc), "rx");
+	if (IS_ERR(host->dma_rx)) {
+		dev_err(mmc_dev(host->mmc), "Can't get dma_rx channel\n");
+		dma_release_channel(host->dma_tx);
+		return PTR_ERR(host->dma_rx);
+	}
+
+	return 0;
+}
+
+/*----------------------------------------------------------------------*/
+
+static void
+mmc_davinci_prepare_data(struct mmc_davinci_host *host, struct mmc_request *req)
+{
+	int fifo_lev = (rw_threshold == 32) ? MMCFIFOCTL_FIFOLEV : 0;
+	int timeout;
+	struct mmc_data *data = req->data;
+
+	if (host->version == MMC_CTLR_VERSION_2)
+		fifo_lev = (rw_threshold == 64) ? MMCFIFOCTL_FIFOLEV : 0;
+
+	host->data = data;
+	if (data == NULL) {
+		host->data_dir = DAVINCI_MMC_DATADIR_NONE;
+		writel(0, host->base + DAVINCI_MMCBLEN);
+		writel(0, host->base + DAVINCI_MMCNBLK);
+		return;
+	}
+
+	dev_dbg(mmc_dev(host->mmc), "%s, %d blocks of %d bytes\n",
+		(data->flags & MMC_DATA_WRITE) ? "write" : "read",
+		data->blocks, data->blksz);
+	dev_dbg(mmc_dev(host->mmc), "  DTO %d cycles + %d ns\n",
+		data->timeout_clks, data->timeout_ns);
+	timeout = data->timeout_clks +
+		(data->timeout_ns / host->ns_in_one_cycle);
+	if (timeout > 0xffff)
+		timeout = 0xffff;
+
+	writel(timeout, host->base + DAVINCI_MMCTOD);
+	writel(data->blocks, host->base + DAVINCI_MMCNBLK);
+	writel(data->blksz, host->base + DAVINCI_MMCBLEN);
+
+	/* Configure the FIFO */
+	if (data->flags & MMC_DATA_WRITE) {
+		host->data_dir = DAVINCI_MMC_DATADIR_WRITE;
+		writel(fifo_lev | MMCFIFOCTL_FIFODIR_WR | MMCFIFOCTL_FIFORST,
+			host->base + DAVINCI_MMCFIFOCTL);
+		writel(fifo_lev | MMCFIFOCTL_FIFODIR_WR,
+			host->base + DAVINCI_MMCFIFOCTL);
+	} else {
+		host->data_dir = DAVINCI_MMC_DATADIR_READ;
+		writel(fifo_lev | MMCFIFOCTL_FIFODIR_RD | MMCFIFOCTL_FIFORST,
+			host->base + DAVINCI_MMCFIFOCTL);
+		writel(fifo_lev | MMCFIFOCTL_FIFODIR_RD,
+			host->base + DAVINCI_MMCFIFOCTL);
+	}
+
+	host->buffer = NULL;
+	host->bytes_left = data->blocks * data->blksz;
+
+	/* For now we try to use DMA whenever we won't need partial FIFO
+	 * reads or writes, either for the whole transfer (as tested here)
+	 * or for any individual scatterlist segment (tested when we call
+	 * start_dma_transfer).
+	 *
+	 * While we *could* change that, unusual block sizes are rarely
+	 * used.  The occasional fallback to PIO should't hurt.
+	 */
+	if (host->use_dma && (host->bytes_left & (rw_threshold - 1)) == 0
+			&& mmc_davinci_start_dma_transfer(host, data) == 0) {
+		/* zero this to ensure we take no PIO paths */
+		host->bytes_left = 0;
+	} else {
+		/* Revert to CPU Copy */
+		host->sg_len = data->sg_len;
+		host->sg = host->data->sg;
+		mmc_davinci_sg_to_buf(host);
+	}
+}
+
+static void mmc_davinci_request(struct mmc_host *mmc, struct mmc_request *req)
+{
+	struct mmc_davinci_host *host = mmc_priv(mmc);
+	unsigned long timeout = jiffies + msecs_to_jiffies(900);
+	u32 mmcst1 = 0;
+
+	/* Card may still be sending BUSY after a previous operation,
+	 * typically some kind of write.  If so, we can't proceed yet.
+	 */
+	while (time_before(jiffies, timeout)) {
+		mmcst1  = readl(host->base + DAVINCI_MMCST1);
+		if (!(mmcst1 & MMCST1_BUSY))
+			break;
+		cpu_relax();
+	}
+	if (mmcst1 & MMCST1_BUSY) {
+		dev_err(mmc_dev(host->mmc), "still BUSY? bad ... \n");
+		req->cmd->error = -ETIMEDOUT;
+		mmc_request_done(mmc, req);
+		return;
+	}
+
+	host->do_dma = 0;
+	mmc_davinci_prepare_data(host, req);
+	mmc_davinci_start_command(host, req->cmd);
+}
+
+static unsigned int calculate_freq_for_card(struct mmc_davinci_host *host,
+	unsigned int mmc_req_freq)
+{
+	unsigned int mmc_freq = 0, mmc_pclk = 0, mmc_push_pull_divisor = 0;
+
+	mmc_pclk = host->mmc_input_clk;
+	if (mmc_req_freq && mmc_pclk > (2 * mmc_req_freq))
+		mmc_push_pull_divisor = ((unsigned int)mmc_pclk
+				/ (2 * mmc_req_freq)) - 1;
+	else
+		mmc_push_pull_divisor = 0;
+
+	mmc_freq = (unsigned int)mmc_pclk
+		/ (2 * (mmc_push_pull_divisor + 1));
+
+	if (mmc_freq > mmc_req_freq)
+		mmc_push_pull_divisor = mmc_push_pull_divisor + 1;
+	/* Convert ns to clock cycles */
+	if (mmc_req_freq <= 400000)
+		host->ns_in_one_cycle = (1000000) / (((mmc_pclk
+				/ (2 * (mmc_push_pull_divisor + 1)))/1000));
+	else
+		host->ns_in_one_cycle = (1000000) / (((mmc_pclk
+				/ (2 * (mmc_push_pull_divisor + 1)))/1000000));
+
+	return mmc_push_pull_divisor;
+}
+
+static void calculate_clk_divider(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	unsigned int open_drain_freq = 0, mmc_pclk = 0;
+	unsigned int mmc_push_pull_freq = 0;
+	struct mmc_davinci_host *host = mmc_priv(mmc);
+
+	if (ios->bus_mode == MMC_BUSMODE_OPENDRAIN) {
+		u32 temp;
+
+		/* Ignoring the init clock value passed for fixing the inter
+		 * operability with different cards.
+		 */
+		open_drain_freq = ((unsigned int)mmc_pclk
+				/ (2 * MMCSD_INIT_CLOCK)) - 1;
+
+		if (open_drain_freq > 0xFF)
+			open_drain_freq = 0xFF;
+
+		temp = readl(host->base + DAVINCI_MMCCLK) & ~MMCCLK_CLKRT_MASK;
+		temp |= open_drain_freq;
+		writel(temp, host->base + DAVINCI_MMCCLK);
+
+		/* Convert ns to clock cycles */
+		host->ns_in_one_cycle = (1000000) / (MMCSD_INIT_CLOCK/1000);
+	} else {
+		u32 temp;
+		mmc_push_pull_freq = calculate_freq_for_card(host, ios->clock);
+
+		if (mmc_push_pull_freq > 0xFF)
+			mmc_push_pull_freq = 0xFF;
+
+		temp = readl(host->base + DAVINCI_MMCCLK) & ~MMCCLK_CLKEN;
+		writel(temp, host->base + DAVINCI_MMCCLK);
+
+		udelay(10);
+
+		temp = readl(host->base + DAVINCI_MMCCLK) & ~MMCCLK_CLKRT_MASK;
+		temp |= mmc_push_pull_freq;
+		writel(temp, host->base + DAVINCI_MMCCLK);
+
+		writel(temp | MMCCLK_CLKEN, host->base + DAVINCI_MMCCLK);
+
+		udelay(10);
+	}
+}
+
+static void mmc_davinci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct mmc_davinci_host *host = mmc_priv(mmc);
+	struct platform_device *pdev = to_platform_device(mmc->parent);
+	struct davinci_mmc_config *config = pdev->dev.platform_data;
+
+	dev_dbg(mmc_dev(host->mmc),
+		"clock %dHz busmode %d powermode %d Vdd %04x\n",
+		ios->clock, ios->bus_mode, ios->power_mode,
+		ios->vdd);
+
+	switch (ios->power_mode) {
+	case MMC_POWER_OFF:
+		if (config && config->set_power)
+			config->set_power(pdev->id, false);
+		break;
+	case MMC_POWER_UP:
+		if (config && config->set_power)
+			config->set_power(pdev->id, true);
+		break;
+	}
+
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_8:
+		dev_dbg(mmc_dev(host->mmc), "Enabling 8 bit mode\n");
+		writel((readl(host->base + DAVINCI_MMCCTL) &
+			~MMCCTL_WIDTH_4_BIT) | MMCCTL_WIDTH_8_BIT,
+			host->base + DAVINCI_MMCCTL);
+		break;
+	case MMC_BUS_WIDTH_4:
+		dev_dbg(mmc_dev(host->mmc), "Enabling 4 bit mode\n");
+		if (host->version == MMC_CTLR_VERSION_2)
+			writel((readl(host->base + DAVINCI_MMCCTL) &
+				~MMCCTL_WIDTH_8_BIT) | MMCCTL_WIDTH_4_BIT,
+				host->base + DAVINCI_MMCCTL);
+		else
+			writel(readl(host->base + DAVINCI_MMCCTL) |
+				MMCCTL_WIDTH_4_BIT,
+				host->base + DAVINCI_MMCCTL);
+		break;
+	case MMC_BUS_WIDTH_1:
+		dev_dbg(mmc_dev(host->mmc), "Enabling 1 bit mode\n");
+		if (host->version == MMC_CTLR_VERSION_2)
+			writel(readl(host->base + DAVINCI_MMCCTL) &
+				~(MMCCTL_WIDTH_8_BIT | MMCCTL_WIDTH_4_BIT),
+				host->base + DAVINCI_MMCCTL);
+		else
+			writel(readl(host->base + DAVINCI_MMCCTL) &
+				~MMCCTL_WIDTH_4_BIT,
+				host->base + DAVINCI_MMCCTL);
+		break;
+	}
+
+	calculate_clk_divider(mmc, ios);
+
+	host->bus_mode = ios->bus_mode;
+	if (ios->power_mode == MMC_POWER_UP) {
+		unsigned long timeout = jiffies + msecs_to_jiffies(50);
+		bool lose = true;
+
+		/* Send clock cycles, poll completion */
+		writel(0, host->base + DAVINCI_MMCARGHL);
+		writel(MMCCMD_INITCK, host->base + DAVINCI_MMCCMD);
+		while (time_before(jiffies, timeout)) {
+			u32 tmp = readl(host->base + DAVINCI_MMCST0);
+
+			if (tmp & MMCST0_RSPDNE) {
+				lose = false;
+				break;
+			}
+			cpu_relax();
+		}
+		if (lose)
+			dev_warn(mmc_dev(host->mmc), "powerup timeout\n");
+	}
+
+	/* FIXME on power OFF, reset things ... */
+}
+
+static void
+mmc_davinci_xfer_done(struct mmc_davinci_host *host, struct mmc_data *data)
+{
+	host->data = NULL;
+
+	if (host->mmc->caps & MMC_CAP_SDIO_IRQ) {
+		/*
+		 * SDIO Interrupt Detection work-around as suggested by
+		 * Davinci Errata (TMS320DM355 Silicon Revision 1.1 Errata
+		 * 2.1.6): Signal SDIO interrupt only if it is enabled by core
+		 */
+		if (host->sdio_int && !(readl(host->base + DAVINCI_SDIOST0) &
+					SDIOST0_DAT1_HI)) {
+			writel(SDIOIST_IOINT, host->base + DAVINCI_SDIOIST);
+			mmc_signal_sdio_irq(host->mmc);
+		}
+	}
+
+	if (host->do_dma) {
+		davinci_abort_dma(host);
+
+		dma_unmap_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+			     mmc_get_dma_dir(data));
+		host->do_dma = false;
+	}
+	host->data_dir = DAVINCI_MMC_DATADIR_NONE;
+
+	if (!data->stop || (host->cmd && host->cmd->error)) {
+		mmc_request_done(host->mmc, data->mrq);
+		writel(0, host->base + DAVINCI_MMCIM);
+		host->active_request = false;
+	} else
+		mmc_davinci_start_command(host, data->stop);
+}
+
+static void mmc_davinci_cmd_done(struct mmc_davinci_host *host,
+				 struct mmc_command *cmd)
+{
+	host->cmd = NULL;
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136) {
+			/* response type 2 */
+			cmd->resp[3] = readl(host->base + DAVINCI_MMCRSP01);
+			cmd->resp[2] = readl(host->base + DAVINCI_MMCRSP23);
+			cmd->resp[1] = readl(host->base + DAVINCI_MMCRSP45);
+			cmd->resp[0] = readl(host->base + DAVINCI_MMCRSP67);
+		} else {
+			/* response types 1, 1b, 3, 4, 5, 6 */
+			cmd->resp[0] = readl(host->base + DAVINCI_MMCRSP67);
+		}
+	}
+
+	if (host->data == NULL || cmd->error) {
+		if (cmd->error == -ETIMEDOUT)
+			cmd->mrq->cmd->retries = 0;
+		mmc_request_done(host->mmc, cmd->mrq);
+		writel(0, host->base + DAVINCI_MMCIM);
+		host->active_request = false;
+	}
+}
+
+static inline void mmc_davinci_reset_ctrl(struct mmc_davinci_host *host,
+								int val)
+{
+	u32 temp;
+
+	temp = readl(host->base + DAVINCI_MMCCTL);
+	if (val)	/* reset */
+		temp |= MMCCTL_CMDRST | MMCCTL_DATRST;
+	else		/* enable */
+		temp &= ~(MMCCTL_CMDRST | MMCCTL_DATRST);
+
+	writel(temp, host->base + DAVINCI_MMCCTL);
+	udelay(10);
+}
+
+static void
+davinci_abort_data(struct mmc_davinci_host *host, struct mmc_data *data)
+{
+	mmc_davinci_reset_ctrl(host, 1);
+	mmc_davinci_reset_ctrl(host, 0);
+}
+
+static irqreturn_t mmc_davinci_sdio_irq(int irq, void *dev_id)
+{
+	struct mmc_davinci_host *host = dev_id;
+	unsigned int status;
+
+	status = readl(host->base + DAVINCI_SDIOIST);
+	if (status & SDIOIST_IOINT) {
+		dev_dbg(mmc_dev(host->mmc),
+			"SDIO interrupt status %x\n", status);
+		writel(status | SDIOIST_IOINT, host->base + DAVINCI_SDIOIST);
+		mmc_signal_sdio_irq(host->mmc);
+	}
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t mmc_davinci_irq(int irq, void *dev_id)
+{
+	struct mmc_davinci_host *host = (struct mmc_davinci_host *)dev_id;
+	unsigned int status, qstatus;
+	int end_command = 0;
+	int end_transfer = 0;
+	struct mmc_data *data = host->data;
+
+	if (host->cmd == NULL && host->data == NULL) {
+		status = readl(host->base + DAVINCI_MMCST0);
+		dev_dbg(mmc_dev(host->mmc),
+			"Spurious interrupt 0x%04x\n", status);
+		/* Disable the interrupt from mmcsd */
+		writel(0, host->base + DAVINCI_MMCIM);
+		return IRQ_NONE;
+	}
+
+	status = readl(host->base + DAVINCI_MMCST0);
+	qstatus = status;
+
+	/* handle FIFO first when using PIO for data.
+	 * bytes_left will decrease to zero as I/O progress and status will
+	 * read zero over iteration because this controller status
+	 * register(MMCST0) reports any status only once and it is cleared
+	 * by read. So, it is not unbouned loop even in the case of
+	 * non-dma.
+	 */
+	if (host->bytes_left && (status & (MMCST0_DXRDY | MMCST0_DRRDY))) {
+		unsigned long im_val;
+
+		/*
+		 * If interrupts fire during the following loop, they will be
+		 * handled by the handler, but the PIC will still buffer these.
+		 * As a result, the handler will be called again to serve these
+		 * needlessly. In order to avoid these spurious interrupts,
+		 * keep interrupts masked during the loop.
+		 */
+		im_val = readl(host->base + DAVINCI_MMCIM);
+		writel(0, host->base + DAVINCI_MMCIM);
+
+		do {
+			davinci_fifo_data_trans(host, rw_threshold);
+			status = readl(host->base + DAVINCI_MMCST0);
+			qstatus |= status;
+		} while (host->bytes_left &&
+			 (status & (MMCST0_DXRDY | MMCST0_DRRDY)));
+
+		/*
+		 * If an interrupt is pending, it is assumed it will fire when
+		 * it is unmasked. This assumption is also taken when the MMCIM
+		 * is first set. Otherwise, writing to MMCIM after reading the
+		 * status is race-prone.
+		 */
+		writel(im_val, host->base + DAVINCI_MMCIM);
+	}
+
+	if (qstatus & MMCST0_DATDNE) {
+		/* All blocks sent/received, and CRC checks passed */
+		if (data != NULL) {
+			if ((host->do_dma == 0) && (host->bytes_left > 0)) {
+				/* if datasize < rw_threshold
+				 * no RX ints are generated
+				 */
+				davinci_fifo_data_trans(host, host->bytes_left);
+			}
+			end_transfer = 1;
+			data->bytes_xfered = data->blocks * data->blksz;
+		} else {
+			dev_err(mmc_dev(host->mmc),
+					"DATDNE with no host->data\n");
+		}
+	}
+
+	if (qstatus & MMCST0_TOUTRD) {
+		/* Read data timeout */
+		data->error = -ETIMEDOUT;
+		end_transfer = 1;
+
+		dev_dbg(mmc_dev(host->mmc),
+			"read data timeout, status %x\n",
+			qstatus);
+
+		davinci_abort_data(host, data);
+	}
+
+	if (qstatus & (MMCST0_CRCWR | MMCST0_CRCRD)) {
+		/* Data CRC error */
+		data->error = -EILSEQ;
+		end_transfer = 1;
+
+		/* NOTE:  this controller uses CRCWR to report both CRC
+		 * errors and timeouts (on writes).  MMCDRSP values are
+		 * only weakly documented, but 0x9f was clearly a timeout
+		 * case and the two three-bit patterns in various SD specs
+		 * (101, 010) aren't part of it ...
+		 */
+		if (qstatus & MMCST0_CRCWR) {
+			u32 temp = readb(host->base + DAVINCI_MMCDRSP);
+
+			if (temp == 0x9f)
+				data->error = -ETIMEDOUT;
+		}
+		dev_dbg(mmc_dev(host->mmc), "data %s %s error\n",
+			(qstatus & MMCST0_CRCWR) ? "write" : "read",
+			(data->error == -ETIMEDOUT) ? "timeout" : "CRC");
+
+		davinci_abort_data(host, data);
+	}
+
+	if (qstatus & MMCST0_TOUTRS) {
+		/* Command timeout */
+		if (host->cmd) {
+			dev_dbg(mmc_dev(host->mmc),
+				"CMD%d timeout, status %x\n",
+				host->cmd->opcode, qstatus);
+			host->cmd->error = -ETIMEDOUT;
+			if (data) {
+				end_transfer = 1;
+				davinci_abort_data(host, data);
+			} else
+				end_command = 1;
+		}
+	}
+
+	if (qstatus & MMCST0_CRCRS) {
+		/* Command CRC error */
+		dev_dbg(mmc_dev(host->mmc), "Command CRC error\n");
+		if (host->cmd) {
+			host->cmd->error = -EILSEQ;
+			end_command = 1;
+		}
+	}
+
+	if (qstatus & MMCST0_RSPDNE) {
+		/* End of command phase */
+		end_command = host->cmd ? 1 : 0;
+	}
+
+	if (end_command)
+		mmc_davinci_cmd_done(host, host->cmd);
+	if (end_transfer)
+		mmc_davinci_xfer_done(host, data);
+	return IRQ_HANDLED;
+}
+
+static int mmc_davinci_get_cd(struct mmc_host *mmc)
+{
+	struct platform_device *pdev = to_platform_device(mmc->parent);
+	struct davinci_mmc_config *config = pdev->dev.platform_data;
+
+	if (config && config->get_cd)
+		return config->get_cd(pdev->id);
+
+	return mmc_gpio_get_cd(mmc);
+}
+
+static int mmc_davinci_get_ro(struct mmc_host *mmc)
+{
+	struct platform_device *pdev = to_platform_device(mmc->parent);
+	struct davinci_mmc_config *config = pdev->dev.platform_data;
+
+	if (config && config->get_ro)
+		return config->get_ro(pdev->id);
+
+	return mmc_gpio_get_ro(mmc);
+}
+
+static void mmc_davinci_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct mmc_davinci_host *host = mmc_priv(mmc);
+
+	if (enable) {
+		if (!(readl(host->base + DAVINCI_SDIOST0) & SDIOST0_DAT1_HI)) {
+			writel(SDIOIST_IOINT, host->base + DAVINCI_SDIOIST);
+			mmc_signal_sdio_irq(host->mmc);
+		} else {
+			host->sdio_int = true;
+			writel(readl(host->base + DAVINCI_SDIOIEN) |
+			       SDIOIEN_IOINTEN, host->base + DAVINCI_SDIOIEN);
+		}
+	} else {
+		host->sdio_int = false;
+		writel(readl(host->base + DAVINCI_SDIOIEN) & ~SDIOIEN_IOINTEN,
+		       host->base + DAVINCI_SDIOIEN);
+	}
+}
+
+static const struct mmc_host_ops mmc_davinci_ops = {
+	.request	= mmc_davinci_request,
+	.set_ios	= mmc_davinci_set_ios,
+	.get_cd		= mmc_davinci_get_cd,
+	.get_ro		= mmc_davinci_get_ro,
+	.enable_sdio_irq = mmc_davinci_enable_sdio_irq,
+};
+
+/*----------------------------------------------------------------------*/
+
+#ifdef CONFIG_CPU_FREQ
+static int mmc_davinci_cpufreq_transition(struct notifier_block *nb,
+				     unsigned long val, void *data)
+{
+	struct mmc_davinci_host *host;
+	unsigned int mmc_pclk;
+	struct mmc_host *mmc;
+	unsigned long flags;
+
+	host = container_of(nb, struct mmc_davinci_host, freq_transition);
+	mmc = host->mmc;
+	mmc_pclk = clk_get_rate(host->clk);
+
+	if (val == CPUFREQ_POSTCHANGE) {
+		spin_lock_irqsave(&mmc->lock, flags);
+		host->mmc_input_clk = mmc_pclk;
+		calculate_clk_divider(mmc, &mmc->ios);
+		spin_unlock_irqrestore(&mmc->lock, flags);
+	}
+
+	return 0;
+}
+
+static inline int mmc_davinci_cpufreq_register(struct mmc_davinci_host *host)
+{
+	host->freq_transition.notifier_call = mmc_davinci_cpufreq_transition;
+
+	return cpufreq_register_notifier(&host->freq_transition,
+					 CPUFREQ_TRANSITION_NOTIFIER);
+}
+
+static inline void mmc_davinci_cpufreq_deregister(struct mmc_davinci_host *host)
+{
+	cpufreq_unregister_notifier(&host->freq_transition,
+				    CPUFREQ_TRANSITION_NOTIFIER);
+}
+#else
+static inline int mmc_davinci_cpufreq_register(struct mmc_davinci_host *host)
+{
+	return 0;
+}
+
+static inline void mmc_davinci_cpufreq_deregister(struct mmc_davinci_host *host)
+{
+}
+#endif
+static void init_mmcsd_host(struct mmc_davinci_host *host)
+{
+
+	mmc_davinci_reset_ctrl(host, 1);
+
+	writel(0, host->base + DAVINCI_MMCCLK);
+	writel(MMCCLK_CLKEN, host->base + DAVINCI_MMCCLK);
+
+	writel(0x1FFF, host->base + DAVINCI_MMCTOR);
+	writel(0xFFFF, host->base + DAVINCI_MMCTOD);
+
+	mmc_davinci_reset_ctrl(host, 0);
+}
+
+static const struct platform_device_id davinci_mmc_devtype[] = {
+	{
+		.name	= "dm6441-mmc",
+		.driver_data = MMC_CTLR_VERSION_1,
+	}, {
+		.name	= "da830-mmc",
+		.driver_data = MMC_CTLR_VERSION_2,
+	},
+	{},
+};
+MODULE_DEVICE_TABLE(platform, davinci_mmc_devtype);
+
+static const struct of_device_id davinci_mmc_dt_ids[] = {
+	{
+		.compatible = "ti,dm6441-mmc",
+		.data = &davinci_mmc_devtype[MMC_CTLR_VERSION_1],
+	},
+	{
+		.compatible = "ti,da830-mmc",
+		.data = &davinci_mmc_devtype[MMC_CTLR_VERSION_2],
+	},
+	{},
+};
+MODULE_DEVICE_TABLE(of, davinci_mmc_dt_ids);
+
+static int mmc_davinci_parse_pdata(struct mmc_host *mmc)
+{
+	struct platform_device *pdev = to_platform_device(mmc->parent);
+	struct davinci_mmc_config *pdata = pdev->dev.platform_data;
+	struct mmc_davinci_host *host;
+	int ret;
+
+	if (!pdata)
+		return -EINVAL;
+
+	host = mmc_priv(mmc);
+	if (!host)
+		return -EINVAL;
+
+	if (pdata && pdata->nr_sg)
+		host->nr_sg = pdata->nr_sg - 1;
+
+	if (pdata && (pdata->wires == 4 || pdata->wires == 0))
+		mmc->caps |= MMC_CAP_4_BIT_DATA;
+
+	if (pdata && (pdata->wires == 8))
+		mmc->caps |= (MMC_CAP_4_BIT_DATA | MMC_CAP_8_BIT_DATA);
+
+	mmc->f_min = 312500;
+	mmc->f_max = 25000000;
+	if (pdata && pdata->max_freq)
+		mmc->f_max = pdata->max_freq;
+	if (pdata && pdata->caps)
+		mmc->caps |= pdata->caps;
+
+	/* Register a cd gpio, if there is not one, enable polling */
+	ret = mmc_gpiod_request_cd(mmc, "cd", 0, false, 0);
+	if (ret == -EPROBE_DEFER)
+		return ret;
+	else if (ret)
+		mmc->caps |= MMC_CAP_NEEDS_POLL;
+
+	ret = mmc_gpiod_request_ro(mmc, "wp", 0, 0);
+	if (ret == -EPROBE_DEFER)
+		return ret;
+
+	return 0;
+}
+
+static int davinci_mmcsd_probe(struct platform_device *pdev)
+{
+	const struct of_device_id *match;
+	struct mmc_davinci_host *host = NULL;
+	struct mmc_host *mmc = NULL;
+	struct resource *r, *mem = NULL;
+	int ret, irq;
+	size_t mem_size;
+	const struct platform_device_id *id_entry;
+
+	r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!r)
+		return -ENODEV;
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return irq;
+
+	mem_size = resource_size(r);
+	mem = devm_request_mem_region(&pdev->dev, r->start, mem_size,
+				      pdev->name);
+	if (!mem)
+		return -EBUSY;
+
+	mmc = mmc_alloc_host(sizeof(struct mmc_davinci_host), &pdev->dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;	/* Important */
+
+	host->mem_res = mem;
+	host->base = devm_ioremap(&pdev->dev, mem->start, mem_size);
+	if (!host->base) {
+		ret = -ENOMEM;
+		goto ioremap_fail;
+	}
+
+	host->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(host->clk)) {
+		ret = PTR_ERR(host->clk);
+		goto clk_get_fail;
+	}
+	ret = clk_prepare_enable(host->clk);
+	if (ret)
+		goto clk_prepare_enable_fail;
+
+	host->mmc_input_clk = clk_get_rate(host->clk);
+
+	match = of_match_device(davinci_mmc_dt_ids, &pdev->dev);
+	if (match) {
+		pdev->id_entry = match->data;
+		ret = mmc_of_parse(mmc);
+		if (ret) {
+			dev_err_probe(&pdev->dev, ret,
+				      "could not parse of data\n");
+			goto parse_fail;
+		}
+	} else {
+		ret = mmc_davinci_parse_pdata(mmc);
+		if (ret) {
+			dev_err(&pdev->dev,
+				"could not parse platform data: %d\n", ret);
+			goto parse_fail;
+	}	}
+
+	if (host->nr_sg > MAX_NR_SG || !host->nr_sg)
+		host->nr_sg = MAX_NR_SG;
+
+	init_mmcsd_host(host);
+
+	host->use_dma = use_dma;
+	host->mmc_irq = irq;
+	host->sdio_irq = platform_get_irq(pdev, 1);
+
+	if (host->use_dma) {
+		ret = davinci_acquire_dma_channels(host);
+		if (ret == -EPROBE_DEFER)
+			goto dma_probe_defer;
+		else if (ret)
+			host->use_dma = 0;
+	}
+
+	mmc->caps |= MMC_CAP_WAIT_WHILE_BUSY;
+
+	id_entry = platform_get_device_id(pdev);
+	if (id_entry)
+		host->version = id_entry->driver_data;
+
+	mmc->ops = &mmc_davinci_ops;
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+
+	/* With no iommu coalescing pages, each phys_seg is a hw_seg.
+	 * Each hw_seg uses one EDMA parameter RAM slot, always one
+	 * channel and then usually some linked slots.
+	 */
+	mmc->max_segs		= MAX_NR_SG;
+
+	/* EDMA limit per hw segment (one or two MBytes) */
+	mmc->max_seg_size	= MAX_CCNT * rw_threshold;
+
+	/* MMC/SD controller limits for multiblock requests */
+	mmc->max_blk_size	= 4095;  /* BLEN is 12 bits */
+	mmc->max_blk_count	= 65535; /* NBLK is 16 bits */
+	mmc->max_req_size	= mmc->max_blk_size * mmc->max_blk_count;
+
+	dev_dbg(mmc_dev(host->mmc), "max_segs=%d\n", mmc->max_segs);
+	dev_dbg(mmc_dev(host->mmc), "max_blk_size=%d\n", mmc->max_blk_size);
+	dev_dbg(mmc_dev(host->mmc), "max_req_size=%d\n", mmc->max_req_size);
+	dev_dbg(mmc_dev(host->mmc), "max_seg_size=%d\n", mmc->max_seg_size);
+
+	platform_set_drvdata(pdev, host);
+
+	ret = mmc_davinci_cpufreq_register(host);
+	if (ret) {
+		dev_err(&pdev->dev, "failed to register cpufreq\n");
+		goto cpu_freq_fail;
+	}
+
+	ret = mmc_add_host(mmc);
+	if (ret < 0)
+		goto mmc_add_host_fail;
+
+	ret = devm_request_irq(&pdev->dev, irq, mmc_davinci_irq, 0,
+			       mmc_hostname(mmc), host);
+	if (ret)
+		goto request_irq_fail;
+
+	if (host->sdio_irq >= 0) {
+		ret = devm_request_irq(&pdev->dev, host->sdio_irq,
+				       mmc_davinci_sdio_irq, 0,
+				       mmc_hostname(mmc), host);
+		if (!ret)
+			mmc->caps |= MMC_CAP_SDIO_IRQ;
+	}
+
+	rename_region(mem, mmc_hostname(mmc));
+
+	dev_info(mmc_dev(host->mmc), "Using %s, %d-bit mode\n",
+		host->use_dma ? "DMA" : "PIO",
+		(mmc->caps & MMC_CAP_4_BIT_DATA) ? 4 : 1);
+
+	return 0;
+
+request_irq_fail:
+	mmc_remove_host(mmc);
+mmc_add_host_fail:
+	mmc_davinci_cpufreq_deregister(host);
+cpu_freq_fail:
+	davinci_release_dma_channels(host);
+parse_fail:
+dma_probe_defer:
+	clk_disable_unprepare(host->clk);
+clk_prepare_enable_fail:
+clk_get_fail:
+ioremap_fail:
+	mmc_free_host(mmc);
+
+	return ret;
+}
+
+static int __exit davinci_mmcsd_remove(struct platform_device *pdev)
+{
+	struct mmc_davinci_host *host = platform_get_drvdata(pdev);
+
+	mmc_remove_host(host->mmc);
+	mmc_davinci_cpufreq_deregister(host);
+	davinci_release_dma_channels(host);
+	clk_disable_unprepare(host->clk);
+	mmc_free_host(host->mmc);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static int davinci_mmcsd_suspend(struct device *dev)
+{
+	struct mmc_davinci_host *host = dev_get_drvdata(dev);
+
+	writel(0, host->base + DAVINCI_MMCIM);
+	mmc_davinci_reset_ctrl(host, 1);
+	clk_disable(host->clk);
+
+	return 0;
+}
+
+static int davinci_mmcsd_resume(struct device *dev)
+{
+	struct mmc_davinci_host *host = dev_get_drvdata(dev);
+	int ret;
+
+	ret = clk_enable(host->clk);
+	if (ret)
+		return ret;
+
+	mmc_davinci_reset_ctrl(host, 0);
+
+	return 0;
+}
+
+static const struct dev_pm_ops davinci_mmcsd_pm = {
+	.suspend        = davinci_mmcsd_suspend,
+	.resume         = davinci_mmcsd_resume,
+};
+
+#define davinci_mmcsd_pm_ops (&davinci_mmcsd_pm)
+#else
+#define davinci_mmcsd_pm_ops NULL
+#endif
+
+static struct platform_driver davinci_mmcsd_driver = {
+	.driver		= {
+		.name	= "davinci_mmc",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= davinci_mmcsd_pm_ops,
+		.of_match_table = davinci_mmc_dt_ids,
+	},
+	.probe		= davinci_mmcsd_probe,
+	.remove		= __exit_p(davinci_mmcsd_remove),
+	.id_table	= davinci_mmc_devtype,
+};
+
+module_platform_driver(davinci_mmcsd_driver);
+
+MODULE_AUTHOR("Texas Instruments India");
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("MMC/SD driver for Davinci MMC controller");
+MODULE_ALIAS("platform:davinci_mmc");
+
diff --git a/drivers/mmc/host/dw_mmc-bluefield.c b/drivers/mmc/host/dw_mmc-bluefield.c
new file mode 100644
index 000000000..10baf122b
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-bluefield.c
@@ -0,0 +1,68 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2018 Mellanox Technologies.
+ */
+
+#include <linux/bitfield.h>
+#include <linux/bitops.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+
+#include "dw_mmc.h"
+#include "dw_mmc-pltfm.h"
+
+#define UHS_REG_EXT_SAMPLE_MASK		GENMASK(22, 16)
+#define UHS_REG_EXT_DRIVE_MASK		GENMASK(29, 23)
+#define BLUEFIELD_UHS_REG_EXT_SAMPLE	2
+#define BLUEFIELD_UHS_REG_EXT_DRIVE	4
+
+static void dw_mci_bluefield_set_ios(struct dw_mci *host, struct mmc_ios *ios)
+{
+	u32 reg;
+
+	/* Update the Drive and Sample fields in register UHS_REG_EXT. */
+	reg = mci_readl(host, UHS_REG_EXT);
+	reg &= ~UHS_REG_EXT_SAMPLE_MASK;
+	reg |= FIELD_PREP(UHS_REG_EXT_SAMPLE_MASK,
+			  BLUEFIELD_UHS_REG_EXT_SAMPLE);
+	reg &= ~UHS_REG_EXT_DRIVE_MASK;
+	reg |= FIELD_PREP(UHS_REG_EXT_DRIVE_MASK, BLUEFIELD_UHS_REG_EXT_DRIVE);
+	mci_writel(host, UHS_REG_EXT, reg);
+}
+
+static const struct dw_mci_drv_data bluefield_drv_data = {
+	.set_ios		= dw_mci_bluefield_set_ios
+};
+
+static const struct of_device_id dw_mci_bluefield_match[] = {
+	{ .compatible = "mellanox,bluefield-dw-mshc",
+	  .data = &bluefield_drv_data },
+	{},
+};
+MODULE_DEVICE_TABLE(of, dw_mci_bluefield_match);
+
+static int dw_mci_bluefield_probe(struct platform_device *pdev)
+{
+	return dw_mci_pltfm_register(pdev, &bluefield_drv_data);
+}
+
+static struct platform_driver dw_mci_bluefield_pltfm_driver = {
+	.probe		= dw_mci_bluefield_probe,
+	.remove		= dw_mci_pltfm_remove,
+	.driver		= {
+		.name		= "dwmmc_bluefield",
+		.probe_type	= PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table	= dw_mci_bluefield_match,
+		.pm		= &dw_mci_pltfm_pmops,
+	},
+};
+
+module_platform_driver(dw_mci_bluefield_pltfm_driver);
+
+MODULE_DESCRIPTION("BlueField DW Multimedia Card driver");
+MODULE_AUTHOR("Mellanox Technologies");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/dw_mmc-exynos.c b/drivers/mmc/host/dw_mmc-exynos.c
new file mode 100644
index 000000000..1f8a3c0dd
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-exynos.c
@@ -0,0 +1,620 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Exynos Specific Extensions for Synopsys DW Multimedia Card Interface driver
+ *
+ * Copyright (C) 2012, Samsung Electronics Co., Ltd.
+ */
+
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/clk.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/of.h>
+#include <linux/of_gpio.h>
+#include <linux/pm_runtime.h>
+#include <linux/slab.h>
+
+#include "dw_mmc.h"
+#include "dw_mmc-pltfm.h"
+#include "dw_mmc-exynos.h"
+
+/* Variations in Exynos specific dw-mshc controller */
+enum dw_mci_exynos_type {
+	DW_MCI_TYPE_EXYNOS4210,
+	DW_MCI_TYPE_EXYNOS4412,
+	DW_MCI_TYPE_EXYNOS5250,
+	DW_MCI_TYPE_EXYNOS5420,
+	DW_MCI_TYPE_EXYNOS5420_SMU,
+	DW_MCI_TYPE_EXYNOS7,
+	DW_MCI_TYPE_EXYNOS7_SMU,
+};
+
+/* Exynos implementation specific driver private data */
+struct dw_mci_exynos_priv_data {
+	enum dw_mci_exynos_type		ctrl_type;
+	u8				ciu_div;
+	u32				sdr_timing;
+	u32				ddr_timing;
+	u32				hs400_timing;
+	u32				tuned_sample;
+	u32				cur_speed;
+	u32				dqs_delay;
+	u32				saved_dqs_en;
+	u32				saved_strobe_ctrl;
+};
+
+static struct dw_mci_exynos_compatible {
+	char				*compatible;
+	enum dw_mci_exynos_type		ctrl_type;
+} exynos_compat[] = {
+	{
+		.compatible	= "samsung,exynos4210-dw-mshc",
+		.ctrl_type	= DW_MCI_TYPE_EXYNOS4210,
+	}, {
+		.compatible	= "samsung,exynos4412-dw-mshc",
+		.ctrl_type	= DW_MCI_TYPE_EXYNOS4412,
+	}, {
+		.compatible	= "samsung,exynos5250-dw-mshc",
+		.ctrl_type	= DW_MCI_TYPE_EXYNOS5250,
+	}, {
+		.compatible	= "samsung,exynos5420-dw-mshc",
+		.ctrl_type	= DW_MCI_TYPE_EXYNOS5420,
+	}, {
+		.compatible	= "samsung,exynos5420-dw-mshc-smu",
+		.ctrl_type	= DW_MCI_TYPE_EXYNOS5420_SMU,
+	}, {
+		.compatible	= "samsung,exynos7-dw-mshc",
+		.ctrl_type	= DW_MCI_TYPE_EXYNOS7,
+	}, {
+		.compatible	= "samsung,exynos7-dw-mshc-smu",
+		.ctrl_type	= DW_MCI_TYPE_EXYNOS7_SMU,
+	},
+};
+
+static inline u8 dw_mci_exynos_get_ciu_div(struct dw_mci *host)
+{
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+
+	if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS4412)
+		return EXYNOS4412_FIXED_CIU_CLK_DIV;
+	else if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS4210)
+		return EXYNOS4210_FIXED_CIU_CLK_DIV;
+	else if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS7 ||
+			priv->ctrl_type == DW_MCI_TYPE_EXYNOS7_SMU)
+		return SDMMC_CLKSEL_GET_DIV(mci_readl(host, CLKSEL64)) + 1;
+	else
+		return SDMMC_CLKSEL_GET_DIV(mci_readl(host, CLKSEL)) + 1;
+}
+
+static void dw_mci_exynos_config_smu(struct dw_mci *host)
+{
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+
+	/*
+	 * If Exynos is provided the Security management,
+	 * set for non-ecryption mode at this time.
+	 */
+	if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS5420_SMU ||
+		priv->ctrl_type == DW_MCI_TYPE_EXYNOS7_SMU) {
+		mci_writel(host, MPSBEGIN0, 0);
+		mci_writel(host, MPSEND0, SDMMC_ENDING_SEC_NR_MAX);
+		mci_writel(host, MPSCTRL0, SDMMC_MPSCTRL_SECURE_WRITE_BIT |
+			   SDMMC_MPSCTRL_NON_SECURE_READ_BIT |
+			   SDMMC_MPSCTRL_VALID |
+			   SDMMC_MPSCTRL_NON_SECURE_WRITE_BIT);
+	}
+}
+
+static int dw_mci_exynos_priv_init(struct dw_mci *host)
+{
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+
+	dw_mci_exynos_config_smu(host);
+
+	if (priv->ctrl_type >= DW_MCI_TYPE_EXYNOS5420) {
+		priv->saved_strobe_ctrl = mci_readl(host, HS400_DLINE_CTRL);
+		priv->saved_dqs_en = mci_readl(host, HS400_DQS_EN);
+		priv->saved_dqs_en |= AXI_NON_BLOCKING_WR;
+		mci_writel(host, HS400_DQS_EN, priv->saved_dqs_en);
+		if (!priv->dqs_delay)
+			priv->dqs_delay =
+				DQS_CTRL_GET_RD_DELAY(priv->saved_strobe_ctrl);
+	}
+
+	host->bus_hz /= (priv->ciu_div + 1);
+
+	return 0;
+}
+
+static void dw_mci_exynos_set_clksel_timing(struct dw_mci *host, u32 timing)
+{
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+	u32 clksel;
+
+	if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS7 ||
+		priv->ctrl_type == DW_MCI_TYPE_EXYNOS7_SMU)
+		clksel = mci_readl(host, CLKSEL64);
+	else
+		clksel = mci_readl(host, CLKSEL);
+
+	clksel = (clksel & ~SDMMC_CLKSEL_TIMING_MASK) | timing;
+
+	if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS7 ||
+		priv->ctrl_type == DW_MCI_TYPE_EXYNOS7_SMU)
+		mci_writel(host, CLKSEL64, clksel);
+	else
+		mci_writel(host, CLKSEL, clksel);
+
+	/*
+	 * Exynos4412 and Exynos5250 extends the use of CMD register with the
+	 * use of bit 29 (which is reserved on standard MSHC controllers) for
+	 * optionally bypassing the HOLD register for command and data. The
+	 * HOLD register should be bypassed in case there is no phase shift
+	 * applied on CMD/DATA that is sent to the card.
+	 */
+	if (!SDMMC_CLKSEL_GET_DRV_WD3(clksel) && host->slot)
+		set_bit(DW_MMC_CARD_NO_USE_HOLD, &host->slot->flags);
+}
+
+#ifdef CONFIG_PM
+static int dw_mci_exynos_runtime_resume(struct device *dev)
+{
+	struct dw_mci *host = dev_get_drvdata(dev);
+	int ret;
+
+	ret = dw_mci_runtime_resume(dev);
+	if (ret)
+		return ret;
+
+	dw_mci_exynos_config_smu(host);
+
+	return ret;
+}
+#endif /* CONFIG_PM */
+
+#ifdef CONFIG_PM_SLEEP
+/**
+ * dw_mci_exynos_suspend_noirq - Exynos-specific suspend code
+ * @dev: Device to suspend (this device)
+ *
+ * This ensures that device will be in runtime active state in
+ * dw_mci_exynos_resume_noirq after calling pm_runtime_force_resume()
+ */
+static int dw_mci_exynos_suspend_noirq(struct device *dev)
+{
+	pm_runtime_get_noresume(dev);
+	return pm_runtime_force_suspend(dev);
+}
+
+/**
+ * dw_mci_exynos_resume_noirq - Exynos-specific resume code
+ * @dev: Device to resume (this device)
+ *
+ * On exynos5420 there is a silicon errata that will sometimes leave the
+ * WAKEUP_INT bit in the CLKSEL register asserted.  This bit is 1 to indicate
+ * that it fired and we can clear it by writing a 1 back.  Clear it to prevent
+ * interrupts from going off constantly.
+ *
+ * We run this code on all exynos variants because it doesn't hurt.
+ */
+static int dw_mci_exynos_resume_noirq(struct device *dev)
+{
+	struct dw_mci *host = dev_get_drvdata(dev);
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+	u32 clksel;
+	int ret;
+
+	ret = pm_runtime_force_resume(dev);
+	if (ret)
+		return ret;
+
+	if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS7 ||
+		priv->ctrl_type == DW_MCI_TYPE_EXYNOS7_SMU)
+		clksel = mci_readl(host, CLKSEL64);
+	else
+		clksel = mci_readl(host, CLKSEL);
+
+	if (clksel & SDMMC_CLKSEL_WAKEUP_INT) {
+		if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS7 ||
+			priv->ctrl_type == DW_MCI_TYPE_EXYNOS7_SMU)
+			mci_writel(host, CLKSEL64, clksel);
+		else
+			mci_writel(host, CLKSEL, clksel);
+	}
+
+	pm_runtime_put(dev);
+
+	return 0;
+}
+#endif /* CONFIG_PM_SLEEP */
+
+static void dw_mci_exynos_config_hs400(struct dw_mci *host, u32 timing)
+{
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+	u32 dqs, strobe;
+
+	/*
+	 * Not supported to configure register
+	 * related to HS400
+	 */
+	if (priv->ctrl_type < DW_MCI_TYPE_EXYNOS5420) {
+		if (timing == MMC_TIMING_MMC_HS400)
+			dev_warn(host->dev,
+				 "cannot configure HS400, unsupported chipset\n");
+		return;
+	}
+
+	dqs = priv->saved_dqs_en;
+	strobe = priv->saved_strobe_ctrl;
+
+	if (timing == MMC_TIMING_MMC_HS400) {
+		dqs |= DATA_STROBE_EN;
+		strobe = DQS_CTRL_RD_DELAY(strobe, priv->dqs_delay);
+	} else if (timing == MMC_TIMING_UHS_SDR104) {
+		dqs &= 0xffffff00;
+	} else {
+		dqs &= ~DATA_STROBE_EN;
+	}
+
+	mci_writel(host, HS400_DQS_EN, dqs);
+	mci_writel(host, HS400_DLINE_CTRL, strobe);
+}
+
+static void dw_mci_exynos_adjust_clock(struct dw_mci *host, unsigned int wanted)
+{
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+	unsigned long actual;
+	u8 div;
+	int ret;
+	/*
+	 * Don't care if wanted clock is zero or
+	 * ciu clock is unavailable
+	 */
+	if (!wanted || IS_ERR(host->ciu_clk))
+		return;
+
+	/* Guaranteed minimum frequency for cclkin */
+	if (wanted < EXYNOS_CCLKIN_MIN)
+		wanted = EXYNOS_CCLKIN_MIN;
+
+	if (wanted == priv->cur_speed)
+		return;
+
+	div = dw_mci_exynos_get_ciu_div(host);
+	ret = clk_set_rate(host->ciu_clk, wanted * div);
+	if (ret)
+		dev_warn(host->dev,
+			"failed to set clk-rate %u error: %d\n",
+			wanted * div, ret);
+	actual = clk_get_rate(host->ciu_clk);
+	host->bus_hz = actual / div;
+	priv->cur_speed = wanted;
+	host->current_speed = 0;
+}
+
+static void dw_mci_exynos_set_ios(struct dw_mci *host, struct mmc_ios *ios)
+{
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+	unsigned int wanted = ios->clock;
+	u32 timing = ios->timing, clksel;
+
+	switch (timing) {
+	case MMC_TIMING_MMC_HS400:
+		/* Update tuned sample timing */
+		clksel = SDMMC_CLKSEL_UP_SAMPLE(
+				priv->hs400_timing, priv->tuned_sample);
+		wanted <<= 1;
+		break;
+	case MMC_TIMING_MMC_DDR52:
+		clksel = priv->ddr_timing;
+		/* Should be double rate for DDR mode */
+		if (ios->bus_width == MMC_BUS_WIDTH_8)
+			wanted <<= 1;
+		break;
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_UHS_SDR50:
+		clksel = (priv->sdr_timing & 0xfff8ffff) |
+			(priv->ciu_div << 16);
+		break;
+	case MMC_TIMING_UHS_DDR50:
+		clksel = (priv->ddr_timing & 0xfff8ffff) |
+			(priv->ciu_div << 16);
+		break;
+	default:
+		clksel = priv->sdr_timing;
+	}
+
+	/* Set clock timing for the requested speed mode*/
+	dw_mci_exynos_set_clksel_timing(host, clksel);
+
+	/* Configure setting for HS400 */
+	dw_mci_exynos_config_hs400(host, timing);
+
+	/* Configure clock rate */
+	dw_mci_exynos_adjust_clock(host, wanted);
+}
+
+static int dw_mci_exynos_parse_dt(struct dw_mci *host)
+{
+	struct dw_mci_exynos_priv_data *priv;
+	struct device_node *np = host->dev->of_node;
+	u32 timing[2];
+	u32 div = 0;
+	int idx;
+	int ret;
+
+	priv = devm_kzalloc(host->dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	for (idx = 0; idx < ARRAY_SIZE(exynos_compat); idx++) {
+		if (of_device_is_compatible(np, exynos_compat[idx].compatible))
+			priv->ctrl_type = exynos_compat[idx].ctrl_type;
+	}
+
+	if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS4412)
+		priv->ciu_div = EXYNOS4412_FIXED_CIU_CLK_DIV - 1;
+	else if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS4210)
+		priv->ciu_div = EXYNOS4210_FIXED_CIU_CLK_DIV - 1;
+	else {
+		of_property_read_u32(np, "samsung,dw-mshc-ciu-div", &div);
+		priv->ciu_div = div;
+	}
+
+	ret = of_property_read_u32_array(np,
+			"samsung,dw-mshc-sdr-timing", timing, 2);
+	if (ret)
+		return ret;
+
+	priv->sdr_timing = SDMMC_CLKSEL_TIMING(timing[0], timing[1], div);
+
+	ret = of_property_read_u32_array(np,
+			"samsung,dw-mshc-ddr-timing", timing, 2);
+	if (ret)
+		return ret;
+
+	priv->ddr_timing = SDMMC_CLKSEL_TIMING(timing[0], timing[1], div);
+
+	ret = of_property_read_u32_array(np,
+			"samsung,dw-mshc-hs400-timing", timing, 2);
+	if (!ret && of_property_read_u32(np,
+				"samsung,read-strobe-delay", &priv->dqs_delay))
+		dev_dbg(host->dev,
+			"read-strobe-delay is not found, assuming usage of default value\n");
+
+	priv->hs400_timing = SDMMC_CLKSEL_TIMING(timing[0], timing[1],
+						HS400_FIXED_CIU_CLK_DIV);
+	host->priv = priv;
+	return 0;
+}
+
+static inline u8 dw_mci_exynos_get_clksmpl(struct dw_mci *host)
+{
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+
+	if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS7 ||
+		priv->ctrl_type == DW_MCI_TYPE_EXYNOS7_SMU)
+		return SDMMC_CLKSEL_CCLK_SAMPLE(mci_readl(host, CLKSEL64));
+	else
+		return SDMMC_CLKSEL_CCLK_SAMPLE(mci_readl(host, CLKSEL));
+}
+
+static inline void dw_mci_exynos_set_clksmpl(struct dw_mci *host, u8 sample)
+{
+	u32 clksel;
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+
+	if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS7 ||
+		priv->ctrl_type == DW_MCI_TYPE_EXYNOS7_SMU)
+		clksel = mci_readl(host, CLKSEL64);
+	else
+		clksel = mci_readl(host, CLKSEL);
+	clksel = SDMMC_CLKSEL_UP_SAMPLE(clksel, sample);
+	if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS7 ||
+		priv->ctrl_type == DW_MCI_TYPE_EXYNOS7_SMU)
+		mci_writel(host, CLKSEL64, clksel);
+	else
+		mci_writel(host, CLKSEL, clksel);
+}
+
+static inline u8 dw_mci_exynos_move_next_clksmpl(struct dw_mci *host)
+{
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+	u32 clksel;
+	u8 sample;
+
+	if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS7 ||
+		priv->ctrl_type == DW_MCI_TYPE_EXYNOS7_SMU)
+		clksel = mci_readl(host, CLKSEL64);
+	else
+		clksel = mci_readl(host, CLKSEL);
+
+	sample = (clksel + 1) & 0x7;
+	clksel = SDMMC_CLKSEL_UP_SAMPLE(clksel, sample);
+
+	if (priv->ctrl_type == DW_MCI_TYPE_EXYNOS7 ||
+		priv->ctrl_type == DW_MCI_TYPE_EXYNOS7_SMU)
+		mci_writel(host, CLKSEL64, clksel);
+	else
+		mci_writel(host, CLKSEL, clksel);
+
+	return sample;
+}
+
+static s8 dw_mci_exynos_get_best_clksmpl(u8 candiates)
+{
+	const u8 iter = 8;
+	u8 __c;
+	s8 i, loc = -1;
+
+	for (i = 0; i < iter; i++) {
+		__c = ror8(candiates, i);
+		if ((__c & 0xc7) == 0xc7) {
+			loc = i;
+			goto out;
+		}
+	}
+
+	for (i = 0; i < iter; i++) {
+		__c = ror8(candiates, i);
+		if ((__c & 0x83) == 0x83) {
+			loc = i;
+			goto out;
+		}
+	}
+
+	/*
+	 * If there is no cadiates value, then it needs to return -EIO.
+	 * If there are candiates values and don't find bset clk sample value,
+	 * then use a first candiates clock sample value.
+	 */
+	for (i = 0; i < iter; i++) {
+		__c = ror8(candiates, i);
+		if ((__c & 0x1) == 0x1) {
+			loc = i;
+			goto out;
+		}
+	}
+out:
+	return loc;
+}
+
+static int dw_mci_exynos_execute_tuning(struct dw_mci_slot *slot, u32 opcode)
+{
+	struct dw_mci *host = slot->host;
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+	struct mmc_host *mmc = slot->mmc;
+	u8 start_smpl, smpl, candiates = 0;
+	s8 found;
+	int ret = 0;
+
+	start_smpl = dw_mci_exynos_get_clksmpl(host);
+
+	do {
+		mci_writel(host, TMOUT, ~0);
+		smpl = dw_mci_exynos_move_next_clksmpl(host);
+
+		if (!mmc_send_tuning(mmc, opcode, NULL))
+			candiates |= (1 << smpl);
+
+	} while (start_smpl != smpl);
+
+	found = dw_mci_exynos_get_best_clksmpl(candiates);
+	if (found >= 0) {
+		dw_mci_exynos_set_clksmpl(host, found);
+		priv->tuned_sample = found;
+	} else {
+		ret = -EIO;
+		dev_warn(&mmc->class_dev,
+			"There is no candiates value about clksmpl!\n");
+	}
+
+	return ret;
+}
+
+static int dw_mci_exynos_prepare_hs400_tuning(struct dw_mci *host,
+					struct mmc_ios *ios)
+{
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+
+	dw_mci_exynos_set_clksel_timing(host, priv->hs400_timing);
+	dw_mci_exynos_adjust_clock(host, (ios->clock) << 1);
+
+	return 0;
+}
+
+/* Common capabilities of Exynos4/Exynos5 SoC */
+static unsigned long exynos_dwmmc_caps[4] = {
+	MMC_CAP_1_8V_DDR | MMC_CAP_8_BIT_DATA | MMC_CAP_CMD23,
+	MMC_CAP_CMD23,
+	MMC_CAP_CMD23,
+	MMC_CAP_CMD23,
+};
+
+static const struct dw_mci_drv_data exynos_drv_data = {
+	.caps			= exynos_dwmmc_caps,
+	.num_caps		= ARRAY_SIZE(exynos_dwmmc_caps),
+	.init			= dw_mci_exynos_priv_init,
+	.set_ios		= dw_mci_exynos_set_ios,
+	.parse_dt		= dw_mci_exynos_parse_dt,
+	.execute_tuning		= dw_mci_exynos_execute_tuning,
+	.prepare_hs400_tuning	= dw_mci_exynos_prepare_hs400_tuning,
+};
+
+static const struct of_device_id dw_mci_exynos_match[] = {
+	{ .compatible = "samsung,exynos4412-dw-mshc",
+			.data = &exynos_drv_data, },
+	{ .compatible = "samsung,exynos5250-dw-mshc",
+			.data = &exynos_drv_data, },
+	{ .compatible = "samsung,exynos5420-dw-mshc",
+			.data = &exynos_drv_data, },
+	{ .compatible = "samsung,exynos5420-dw-mshc-smu",
+			.data = &exynos_drv_data, },
+	{ .compatible = "samsung,exynos7-dw-mshc",
+			.data = &exynos_drv_data, },
+	{ .compatible = "samsung,exynos7-dw-mshc-smu",
+			.data = &exynos_drv_data, },
+	{},
+};
+MODULE_DEVICE_TABLE(of, dw_mci_exynos_match);
+
+static int dw_mci_exynos_probe(struct platform_device *pdev)
+{
+	const struct dw_mci_drv_data *drv_data;
+	const struct of_device_id *match;
+	int ret;
+
+	match = of_match_node(dw_mci_exynos_match, pdev->dev.of_node);
+	drv_data = match->data;
+
+	pm_runtime_get_noresume(&pdev->dev);
+	pm_runtime_set_active(&pdev->dev);
+	pm_runtime_enable(&pdev->dev);
+
+	ret = dw_mci_pltfm_register(pdev, drv_data);
+	if (ret) {
+		pm_runtime_disable(&pdev->dev);
+		pm_runtime_set_suspended(&pdev->dev);
+		pm_runtime_put_noidle(&pdev->dev);
+
+		return ret;
+	}
+
+	return 0;
+}
+
+static int dw_mci_exynos_remove(struct platform_device *pdev)
+{
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_set_suspended(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+
+	return dw_mci_pltfm_remove(pdev);
+}
+
+static const struct dev_pm_ops dw_mci_exynos_pmops = {
+	SET_NOIRQ_SYSTEM_SLEEP_PM_OPS(dw_mci_exynos_suspend_noirq,
+				      dw_mci_exynos_resume_noirq)
+	SET_RUNTIME_PM_OPS(dw_mci_runtime_suspend,
+			   dw_mci_exynos_runtime_resume,
+			   NULL)
+};
+
+static struct platform_driver dw_mci_exynos_pltfm_driver = {
+	.probe		= dw_mci_exynos_probe,
+	.remove		= dw_mci_exynos_remove,
+	.driver		= {
+		.name		= "dwmmc_exynos",
+		.probe_type	= PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table	= dw_mci_exynos_match,
+		.pm		= &dw_mci_exynos_pmops,
+	},
+};
+
+module_platform_driver(dw_mci_exynos_pltfm_driver);
+
+MODULE_DESCRIPTION("Samsung Specific DW-MSHC Driver Extension");
+MODULE_AUTHOR("Thomas Abraham <thomas.ab@samsung.com");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:dwmmc_exynos");
diff --git a/drivers/mmc/host/dw_mmc-exynos.h b/drivers/mmc/host/dw_mmc-exynos.h
new file mode 100644
index 000000000..0280d394a
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-exynos.h
@@ -0,0 +1,69 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Exynos Specific Extensions for Synopsys DW Multimedia Card Interface driver
+ *
+ * Copyright (C) 2012-2014 Samsung Electronics Co., Ltd.
+ */
+
+#ifndef _DW_MMC_EXYNOS_H_
+#define _DW_MMC_EXYNOS_H_
+
+#define SDMMC_CLKSEL			0x09C
+#define SDMMC_CLKSEL64			0x0A8
+
+/* Extended Register's Offset */
+#define SDMMC_HS400_DQS_EN		0x180
+#define SDMMC_HS400_ASYNC_FIFO_CTRL	0x184
+#define SDMMC_HS400_DLINE_CTRL		0x188
+
+/* CLKSEL register defines */
+#define SDMMC_CLKSEL_CCLK_SAMPLE(x)	(((x) & 7) << 0)
+#define SDMMC_CLKSEL_CCLK_DRIVE(x)	(((x) & 7) << 16)
+#define SDMMC_CLKSEL_CCLK_DIVIDER(x)	(((x) & 7) << 24)
+#define SDMMC_CLKSEL_GET_DRV_WD3(x)	(((x) >> 16) & 0x7)
+#define SDMMC_CLKSEL_GET_DIV(x)		(((x) >> 24) & 0x7)
+#define SDMMC_CLKSEL_UP_SAMPLE(x, y)	(((x) & ~SDMMC_CLKSEL_CCLK_SAMPLE(7)) |\
+					 SDMMC_CLKSEL_CCLK_SAMPLE(y))
+#define SDMMC_CLKSEL_TIMING(x, y, z)	(SDMMC_CLKSEL_CCLK_SAMPLE(x) |	\
+					 SDMMC_CLKSEL_CCLK_DRIVE(y) |	\
+					 SDMMC_CLKSEL_CCLK_DIVIDER(z))
+#define SDMMC_CLKSEL_TIMING_MASK	SDMMC_CLKSEL_TIMING(0x7, 0x7, 0x7)
+#define SDMMC_CLKSEL_WAKEUP_INT		BIT(11)
+
+/* RCLK_EN register defines */
+#define DATA_STROBE_EN			BIT(0)
+#define AXI_NON_BLOCKING_WR	BIT(7)
+
+/* DLINE_CTRL register defines */
+#define DQS_CTRL_RD_DELAY(x, y)		(((x) & ~0x3FF) | ((y) & 0x3FF))
+#define DQS_CTRL_GET_RD_DELAY(x)	((x) & 0x3FF)
+
+/* Protector Register */
+#define SDMMC_EMMCP_BASE	0x1000
+#define SDMMC_MPSECURITY	(SDMMC_EMMCP_BASE + 0x0010)
+#define SDMMC_MPSBEGIN0		(SDMMC_EMMCP_BASE + 0x0200)
+#define SDMMC_MPSEND0		(SDMMC_EMMCP_BASE + 0x0204)
+#define SDMMC_MPSCTRL0		(SDMMC_EMMCP_BASE + 0x020C)
+
+/* SMU control defines */
+#define SDMMC_MPSCTRL_SECURE_READ_BIT		BIT(7)
+#define SDMMC_MPSCTRL_SECURE_WRITE_BIT		BIT(6)
+#define SDMMC_MPSCTRL_NON_SECURE_READ_BIT	BIT(5)
+#define SDMMC_MPSCTRL_NON_SECURE_WRITE_BIT	BIT(4)
+#define SDMMC_MPSCTRL_USE_FUSE_KEY		BIT(3)
+#define SDMMC_MPSCTRL_ECB_MODE			BIT(2)
+#define SDMMC_MPSCTRL_ENCRYPTION		BIT(1)
+#define SDMMC_MPSCTRL_VALID			BIT(0)
+
+/* Maximum number of Ending sector */
+#define SDMMC_ENDING_SEC_NR_MAX	0xFFFFFFFF
+
+/* Fixed clock divider */
+#define EXYNOS4210_FIXED_CIU_CLK_DIV	2
+#define EXYNOS4412_FIXED_CIU_CLK_DIV	4
+#define HS400_FIXED_CIU_CLK_DIV		1
+
+/* Minimal required clock frequency for cclkin, unit: HZ */
+#define EXYNOS_CCLKIN_MIN	50000000
+
+#endif /* _DW_MMC_EXYNOS_H_ */
diff --git a/drivers/mmc/host/dw_mmc-hi3798cv200.c b/drivers/mmc/host/dw_mmc-hi3798cv200.c
new file mode 100644
index 000000000..39794f938
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-hi3798cv200.c
@@ -0,0 +1,211 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2018 HiSilicon Technologies Co., Ltd.
+ */
+
+#include <linux/clk.h>
+#include <linux/mfd/syscon.h>
+#include <linux/mmc/host.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/regmap.h>
+#include <linux/regulator/consumer.h>
+
+#include "dw_mmc.h"
+#include "dw_mmc-pltfm.h"
+
+#define ALL_INT_CLR		0x1ffff
+
+struct hi3798cv200_priv {
+	struct clk *sample_clk;
+	struct clk *drive_clk;
+};
+
+static unsigned long dw_mci_hi3798cv200_caps[] = {
+	MMC_CAP_CMD23,
+	MMC_CAP_CMD23,
+	MMC_CAP_CMD23
+};
+
+static void dw_mci_hi3798cv200_set_ios(struct dw_mci *host, struct mmc_ios *ios)
+{
+	struct hi3798cv200_priv *priv = host->priv;
+	u32 val;
+
+	val = mci_readl(host, UHS_REG);
+	if (ios->timing == MMC_TIMING_MMC_DDR52 ||
+	    ios->timing == MMC_TIMING_UHS_DDR50)
+		val |= SDMMC_UHS_DDR;
+	else
+		val &= ~SDMMC_UHS_DDR;
+	mci_writel(host, UHS_REG, val);
+
+	val = mci_readl(host, ENABLE_SHIFT);
+	if (ios->timing == MMC_TIMING_MMC_DDR52)
+		val |= SDMMC_ENABLE_PHASE;
+	else
+		val &= ~SDMMC_ENABLE_PHASE;
+	mci_writel(host, ENABLE_SHIFT, val);
+
+	val = mci_readl(host, DDR_REG);
+	if (ios->timing == MMC_TIMING_MMC_HS400)
+		val |= SDMMC_DDR_HS400;
+	else
+		val &= ~SDMMC_DDR_HS400;
+	mci_writel(host, DDR_REG, val);
+
+	if (ios->timing == MMC_TIMING_MMC_HS ||
+	    ios->timing == MMC_TIMING_LEGACY)
+		clk_set_phase(priv->drive_clk, 180);
+	else if (ios->timing == MMC_TIMING_MMC_HS200)
+		clk_set_phase(priv->drive_clk, 135);
+}
+
+static int dw_mci_hi3798cv200_execute_tuning(struct dw_mci_slot *slot,
+					     u32 opcode)
+{
+	static const int degrees[] = { 0, 45, 90, 135, 180, 225, 270, 315 };
+	struct dw_mci *host = slot->host;
+	struct hi3798cv200_priv *priv = host->priv;
+	int raise_point = -1, fall_point = -1;
+	int err, prev_err = -1;
+	int found = 0;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(degrees); i++) {
+		clk_set_phase(priv->sample_clk, degrees[i]);
+		mci_writel(host, RINTSTS, ALL_INT_CLR);
+
+		err = mmc_send_tuning(slot->mmc, opcode, NULL);
+		if (!err)
+			found = 1;
+
+		if (i > 0) {
+			if (err && !prev_err)
+				fall_point = i - 1;
+			if (!err && prev_err)
+				raise_point = i;
+		}
+
+		if (raise_point != -1 && fall_point != -1)
+			goto tuning_out;
+
+		prev_err = err;
+		err = 0;
+	}
+
+tuning_out:
+	if (found) {
+		if (raise_point == -1)
+			raise_point = 0;
+		if (fall_point == -1)
+			fall_point = ARRAY_SIZE(degrees) - 1;
+		if (fall_point < raise_point) {
+			if ((raise_point + fall_point) >
+			    (ARRAY_SIZE(degrees) - 1))
+				i = fall_point / 2;
+			else
+				i = (raise_point + ARRAY_SIZE(degrees) - 1) / 2;
+		} else {
+			i = (raise_point + fall_point) / 2;
+		}
+
+		clk_set_phase(priv->sample_clk, degrees[i]);
+		dev_dbg(host->dev, "Tuning clk_sample[%d, %d], set[%d]\n",
+			raise_point, fall_point, degrees[i]);
+	} else {
+		dev_err(host->dev, "No valid clk_sample shift! use default\n");
+		err = -EINVAL;
+	}
+
+	mci_writel(host, RINTSTS, ALL_INT_CLR);
+	return err;
+}
+
+static int dw_mci_hi3798cv200_init(struct dw_mci *host)
+{
+	struct hi3798cv200_priv *priv;
+	int ret;
+
+	priv = devm_kzalloc(host->dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->sample_clk = devm_clk_get(host->dev, "ciu-sample");
+	if (IS_ERR(priv->sample_clk)) {
+		dev_err(host->dev, "failed to get ciu-sample clock\n");
+		return PTR_ERR(priv->sample_clk);
+	}
+
+	priv->drive_clk = devm_clk_get(host->dev, "ciu-drive");
+	if (IS_ERR(priv->drive_clk)) {
+		dev_err(host->dev, "failed to get ciu-drive clock\n");
+		return PTR_ERR(priv->drive_clk);
+	}
+
+	ret = clk_prepare_enable(priv->sample_clk);
+	if (ret) {
+		dev_err(host->dev, "failed to enable ciu-sample clock\n");
+		return ret;
+	}
+
+	ret = clk_prepare_enable(priv->drive_clk);
+	if (ret) {
+		dev_err(host->dev, "failed to enable ciu-drive clock\n");
+		goto disable_sample_clk;
+	}
+
+	host->priv = priv;
+	return 0;
+
+disable_sample_clk:
+	clk_disable_unprepare(priv->sample_clk);
+	return ret;
+}
+
+static const struct dw_mci_drv_data hi3798cv200_data = {
+	.caps = dw_mci_hi3798cv200_caps,
+	.num_caps = ARRAY_SIZE(dw_mci_hi3798cv200_caps),
+	.init = dw_mci_hi3798cv200_init,
+	.set_ios = dw_mci_hi3798cv200_set_ios,
+	.execute_tuning = dw_mci_hi3798cv200_execute_tuning,
+};
+
+static int dw_mci_hi3798cv200_probe(struct platform_device *pdev)
+{
+	return dw_mci_pltfm_register(pdev, &hi3798cv200_data);
+}
+
+static int dw_mci_hi3798cv200_remove(struct platform_device *pdev)
+{
+	struct dw_mci *host = platform_get_drvdata(pdev);
+	struct hi3798cv200_priv *priv = host->priv;
+
+	clk_disable_unprepare(priv->drive_clk);
+	clk_disable_unprepare(priv->sample_clk);
+
+	return dw_mci_pltfm_remove(pdev);
+}
+
+static const struct of_device_id dw_mci_hi3798cv200_match[] = {
+	{ .compatible = "hisilicon,hi3798cv200-dw-mshc", },
+	{},
+};
+
+MODULE_DEVICE_TABLE(of, dw_mci_hi3798cv200_match);
+static struct platform_driver dw_mci_hi3798cv200_driver = {
+	.probe = dw_mci_hi3798cv200_probe,
+	.remove = dw_mci_hi3798cv200_remove,
+	.driver = {
+		.name = "dwmmc_hi3798cv200",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = dw_mci_hi3798cv200_match,
+	},
+};
+module_platform_driver(dw_mci_hi3798cv200_driver);
+
+MODULE_DESCRIPTION("HiSilicon Hi3798CV200 Specific DW-MSHC Driver Extension");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:dwmmc_hi3798cv200");
diff --git a/drivers/mmc/host/dw_mmc-k3.c b/drivers/mmc/host/dw_mmc-k3.c
new file mode 100644
index 000000000..29d2494eb
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-k3.c
@@ -0,0 +1,486 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2013 Linaro Ltd.
+ * Copyright (c) 2013 Hisilicon Limited.
+ */
+
+#include <linux/bitops.h>
+#include <linux/bitfield.h>
+#include <linux/clk.h>
+#include <linux/mfd/syscon.h>
+#include <linux/mmc/host.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/regmap.h>
+#include <linux/regulator/consumer.h>
+
+#include "dw_mmc.h"
+#include "dw_mmc-pltfm.h"
+
+/*
+ * hi6220 sd only support io voltage 1.8v and 3v
+ * Also need config AO_SCTRL_SEL18 accordingly
+ */
+#define AO_SCTRL_SEL18		BIT(10)
+#define AO_SCTRL_CTRL3		0x40C
+
+#define DWMMC_SDIO_ID 2
+
+#define SOC_SCTRL_SCPERCTRL5    (0x314)
+#define SDCARD_IO_SEL18         BIT(2)
+
+#define SDCARD_RD_THRESHOLD  (512)
+
+#define GENCLK_DIV (7)
+
+#define GPIO_CLK_ENABLE                   BIT(16)
+#define GPIO_CLK_DIV_MASK                 GENMASK(11, 8)
+#define GPIO_USE_SAMPLE_DLY_MASK          GENMASK(13, 13)
+#define UHS_REG_EXT_SAMPLE_PHASE_MASK     GENMASK(20, 16)
+#define UHS_REG_EXT_SAMPLE_DRVPHASE_MASK  GENMASK(25, 21)
+#define UHS_REG_EXT_SAMPLE_DLY_MASK       GENMASK(30, 26)
+
+#define TIMING_MODE     3
+#define TIMING_CFG_NUM 10
+
+#define NUM_PHASES (40)
+
+#define ENABLE_SHIFT_MIN_SMPL (4)
+#define ENABLE_SHIFT_MAX_SMPL (12)
+#define USE_DLY_MIN_SMPL (11)
+#define USE_DLY_MAX_SMPL (14)
+
+struct k3_priv {
+	int ctrl_id;
+	u32 cur_speed;
+	struct regmap	*reg;
+};
+
+static unsigned long dw_mci_hi6220_caps[] = {
+	MMC_CAP_CMD23,
+	MMC_CAP_CMD23,
+	0
+};
+
+struct hs_timing {
+	u32 drv_phase;
+	u32 smpl_dly;
+	u32 smpl_phase_max;
+	u32 smpl_phase_min;
+};
+
+static struct hs_timing hs_timing_cfg[TIMING_MODE][TIMING_CFG_NUM] = {
+	{ /* reserved */ },
+	{ /* SD */
+		{7, 0, 15, 15,},  /* 0: LEGACY 400k */
+		{6, 0,  4,  4,},  /* 1: MMC_HS */
+		{6, 0,  3,  3,},  /* 2: SD_HS */
+		{6, 0, 15, 15,},  /* 3: SDR12 */
+		{6, 0,  2,  2,},  /* 4: SDR25 */
+		{4, 0, 11,  0,},  /* 5: SDR50 */
+		{6, 4, 15,  0,},  /* 6: SDR104 */
+		{0},              /* 7: DDR50 */
+		{0},              /* 8: DDR52 */
+		{0},              /* 9: HS200 */
+	},
+	{ /* SDIO */
+		{7, 0, 15, 15,},  /* 0: LEGACY 400k */
+		{0},              /* 1: MMC_HS */
+		{6, 0, 15, 15,},  /* 2: SD_HS */
+		{6, 0, 15, 15,},  /* 3: SDR12 */
+		{6, 0,  0,  0,},  /* 4: SDR25 */
+		{4, 0, 12,  0,},  /* 5: SDR50 */
+		{5, 4, 15,  0,},  /* 6: SDR104 */
+		{0},              /* 7: DDR50 */
+		{0},              /* 8: DDR52 */
+		{0},              /* 9: HS200 */
+	}
+};
+
+static void dw_mci_k3_set_ios(struct dw_mci *host, struct mmc_ios *ios)
+{
+	int ret;
+
+	ret = clk_set_rate(host->ciu_clk, ios->clock);
+	if (ret)
+		dev_warn(host->dev, "failed to set rate %uHz\n", ios->clock);
+
+	host->bus_hz = clk_get_rate(host->ciu_clk);
+}
+
+static const struct dw_mci_drv_data k3_drv_data = {
+	.set_ios		= dw_mci_k3_set_ios,
+};
+
+static int dw_mci_hi6220_parse_dt(struct dw_mci *host)
+{
+	struct k3_priv *priv;
+
+	priv = devm_kzalloc(host->dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->reg = syscon_regmap_lookup_by_phandle(host->dev->of_node,
+					 "hisilicon,peripheral-syscon");
+	if (IS_ERR(priv->reg))
+		priv->reg = NULL;
+
+	priv->ctrl_id = of_alias_get_id(host->dev->of_node, "mshc");
+	if (priv->ctrl_id < 0)
+		priv->ctrl_id = 0;
+
+	if (priv->ctrl_id >= TIMING_MODE)
+		return -EINVAL;
+
+	host->priv = priv;
+	return 0;
+}
+
+static int dw_mci_hi6220_switch_voltage(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct k3_priv *priv;
+	struct dw_mci *host;
+	int min_uv, max_uv;
+	int ret;
+
+	host = slot->host;
+	priv = host->priv;
+
+	if (!priv || !priv->reg)
+		return 0;
+
+	if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_330) {
+		ret = regmap_update_bits(priv->reg, AO_SCTRL_CTRL3,
+					 AO_SCTRL_SEL18, 0);
+		min_uv = 3000000;
+		max_uv = 3000000;
+	} else if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_180) {
+		ret = regmap_update_bits(priv->reg, AO_SCTRL_CTRL3,
+					 AO_SCTRL_SEL18, AO_SCTRL_SEL18);
+		min_uv = 1800000;
+		max_uv = 1800000;
+	} else {
+		dev_dbg(host->dev, "voltage not supported\n");
+		return -EINVAL;
+	}
+
+	if (ret) {
+		dev_dbg(host->dev, "switch voltage failed\n");
+		return ret;
+	}
+
+	if (IS_ERR_OR_NULL(mmc->supply.vqmmc))
+		return 0;
+
+	ret = regulator_set_voltage(mmc->supply.vqmmc, min_uv, max_uv);
+	if (ret) {
+		dev_dbg(host->dev, "Regulator set error %d: %d - %d\n",
+				 ret, min_uv, max_uv);
+		return ret;
+	}
+
+	return 0;
+}
+
+static void dw_mci_hi6220_set_ios(struct dw_mci *host, struct mmc_ios *ios)
+{
+	int ret;
+	unsigned int clock;
+
+	clock = (ios->clock <= 25000000) ? 25000000 : ios->clock;
+
+	ret = clk_set_rate(host->biu_clk, clock);
+	if (ret)
+		dev_warn(host->dev, "failed to set rate %uHz\n", clock);
+
+	host->bus_hz = clk_get_rate(host->biu_clk);
+}
+
+static int dw_mci_hi6220_execute_tuning(struct dw_mci_slot *slot, u32 opcode)
+{
+	return 0;
+}
+
+static const struct dw_mci_drv_data hi6220_data = {
+	.caps			= dw_mci_hi6220_caps,
+	.num_caps		= ARRAY_SIZE(dw_mci_hi6220_caps),
+	.switch_voltage		= dw_mci_hi6220_switch_voltage,
+	.set_ios		= dw_mci_hi6220_set_ios,
+	.parse_dt		= dw_mci_hi6220_parse_dt,
+	.execute_tuning		= dw_mci_hi6220_execute_tuning,
+};
+
+static void dw_mci_hs_set_timing(struct dw_mci *host, int timing,
+				     int smpl_phase)
+{
+	u32 drv_phase;
+	u32 smpl_dly;
+	u32 use_smpl_dly = 0;
+	u32 enable_shift = 0;
+	u32 reg_value;
+	int ctrl_id;
+	struct k3_priv *priv;
+
+	priv = host->priv;
+	ctrl_id = priv->ctrl_id;
+
+	drv_phase = hs_timing_cfg[ctrl_id][timing].drv_phase;
+	smpl_dly   = hs_timing_cfg[ctrl_id][timing].smpl_dly;
+	if (smpl_phase == -1)
+		smpl_phase = (hs_timing_cfg[ctrl_id][timing].smpl_phase_max +
+			     hs_timing_cfg[ctrl_id][timing].smpl_phase_min) / 2;
+
+	switch (timing) {
+	case MMC_TIMING_UHS_SDR104:
+		if (smpl_phase >= USE_DLY_MIN_SMPL &&
+				smpl_phase <= USE_DLY_MAX_SMPL)
+			use_smpl_dly = 1;
+		fallthrough;
+	case MMC_TIMING_UHS_SDR50:
+		if (smpl_phase >= ENABLE_SHIFT_MIN_SMPL &&
+				smpl_phase <= ENABLE_SHIFT_MAX_SMPL)
+			enable_shift = 1;
+		break;
+	}
+
+	mci_writel(host, GPIO, 0x0);
+	usleep_range(5, 10);
+
+	reg_value = FIELD_PREP(UHS_REG_EXT_SAMPLE_PHASE_MASK, smpl_phase) |
+		    FIELD_PREP(UHS_REG_EXT_SAMPLE_DLY_MASK, smpl_dly) |
+		    FIELD_PREP(UHS_REG_EXT_SAMPLE_DRVPHASE_MASK, drv_phase);
+	mci_writel(host, UHS_REG_EXT, reg_value);
+
+	mci_writel(host, ENABLE_SHIFT, enable_shift);
+
+	reg_value = FIELD_PREP(GPIO_CLK_DIV_MASK, GENCLK_DIV) |
+			     FIELD_PREP(GPIO_USE_SAMPLE_DLY_MASK, use_smpl_dly);
+	mci_writel(host, GPIO, (unsigned int)reg_value | GPIO_CLK_ENABLE);
+
+	/* We should delay 1ms wait for timing setting finished. */
+	usleep_range(1000, 2000);
+}
+
+static int dw_mci_hi3660_init(struct dw_mci *host)
+{
+	mci_writel(host, CDTHRCTL, SDMMC_SET_THLD(SDCARD_RD_THRESHOLD,
+		    SDMMC_CARD_RD_THR_EN));
+
+	dw_mci_hs_set_timing(host, MMC_TIMING_LEGACY, -1);
+	host->bus_hz /= (GENCLK_DIV + 1);
+
+	return 0;
+}
+
+static int dw_mci_set_sel18(struct dw_mci *host, bool set)
+{
+	int ret;
+	unsigned int val;
+	struct k3_priv *priv;
+
+	priv = host->priv;
+
+	val = set ? SDCARD_IO_SEL18 : 0;
+	ret = regmap_update_bits(priv->reg, SOC_SCTRL_SCPERCTRL5,
+				 SDCARD_IO_SEL18, val);
+	if (ret) {
+		dev_err(host->dev, "sel18 %u error\n", val);
+		return ret;
+	}
+
+	return 0;
+}
+
+static void dw_mci_hi3660_set_ios(struct dw_mci *host, struct mmc_ios *ios)
+{
+	int ret;
+	unsigned long wanted;
+	unsigned long actual;
+	struct k3_priv *priv = host->priv;
+
+	if (!ios->clock || ios->clock == priv->cur_speed)
+		return;
+
+	wanted = ios->clock * (GENCLK_DIV + 1);
+	ret = clk_set_rate(host->ciu_clk, wanted);
+	if (ret) {
+		dev_err(host->dev, "failed to set rate %luHz\n", wanted);
+		return;
+	}
+	actual = clk_get_rate(host->ciu_clk);
+
+	dw_mci_hs_set_timing(host, ios->timing, -1);
+	host->bus_hz = actual / (GENCLK_DIV + 1);
+	host->current_speed = 0;
+	priv->cur_speed = host->bus_hz;
+}
+
+static int dw_mci_get_best_clksmpl(unsigned int sample_flag)
+{
+	int i;
+	int interval;
+	unsigned int v;
+	unsigned int len;
+	unsigned int range_start = 0;
+	unsigned int range_length = 0;
+	unsigned int middle_range = 0;
+
+	if (!sample_flag)
+		return -EIO;
+
+	if (~sample_flag == 0)
+		return 0;
+
+	i = ffs(sample_flag) - 1;
+
+	/*
+	* A clock cycle is divided into 32 phases,
+	* each of which is represented by a bit,
+	* finding the optimal phase.
+	*/
+	while (i < 32) {
+		v = ror32(sample_flag, i);
+		len = ffs(~v) - 1;
+
+		if (len > range_length) {
+			range_length = len;
+			range_start = i;
+		}
+
+		interval = ffs(v >> len) - 1;
+		if (interval < 0)
+			break;
+
+		i += len + interval;
+	}
+
+	middle_range = range_start + range_length / 2;
+	if (middle_range >= 32)
+		middle_range %= 32;
+
+	return middle_range;
+}
+
+static int dw_mci_hi3660_execute_tuning(struct dw_mci_slot *slot, u32 opcode)
+{
+	int i = 0;
+	struct dw_mci *host = slot->host;
+	struct mmc_host *mmc = slot->mmc;
+	int smpl_phase = 0;
+	u32 tuning_sample_flag = 0;
+	int best_clksmpl = 0;
+
+	for (i = 0; i < NUM_PHASES; ++i, ++smpl_phase) {
+		smpl_phase %= 32;
+
+		mci_writel(host, TMOUT, ~0);
+		dw_mci_hs_set_timing(host, mmc->ios.timing, smpl_phase);
+
+		if (!mmc_send_tuning(mmc, opcode, NULL))
+			tuning_sample_flag |= (1 << smpl_phase);
+		else
+			tuning_sample_flag &= ~(1 << smpl_phase);
+	}
+
+	best_clksmpl = dw_mci_get_best_clksmpl(tuning_sample_flag);
+	if (best_clksmpl < 0) {
+		dev_err(host->dev, "All phases bad!\n");
+		return -EIO;
+	}
+
+	dw_mci_hs_set_timing(host, mmc->ios.timing, best_clksmpl);
+
+	dev_info(host->dev, "tuning ok best_clksmpl %u tuning_sample_flag %x\n",
+		 best_clksmpl, tuning_sample_flag);
+	return 0;
+}
+
+static int dw_mci_hi3660_switch_voltage(struct mmc_host *mmc,
+					struct mmc_ios *ios)
+{
+	int ret = 0;
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct k3_priv *priv;
+	struct dw_mci *host;
+
+	host = slot->host;
+	priv = host->priv;
+
+	if (!priv || !priv->reg)
+		return 0;
+
+	if (priv->ctrl_id == DWMMC_SDIO_ID)
+		return 0;
+
+	if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_330)
+		ret = dw_mci_set_sel18(host, 0);
+	else if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_180)
+		ret = dw_mci_set_sel18(host, 1);
+	if (ret)
+		return ret;
+
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+		ret = mmc_regulator_set_vqmmc(mmc, ios);
+		if (ret < 0) {
+			dev_err(host->dev, "Regulator set error %d\n", ret);
+			return ret;
+		}
+	}
+
+	return 0;
+}
+
+static const struct dw_mci_drv_data hi3660_data = {
+	.init = dw_mci_hi3660_init,
+	.set_ios = dw_mci_hi3660_set_ios,
+	.parse_dt = dw_mci_hi6220_parse_dt,
+	.execute_tuning = dw_mci_hi3660_execute_tuning,
+	.switch_voltage  = dw_mci_hi3660_switch_voltage,
+};
+
+static const struct of_device_id dw_mci_k3_match[] = {
+	{ .compatible = "hisilicon,hi3660-dw-mshc", .data = &hi3660_data, },
+	{ .compatible = "hisilicon,hi4511-dw-mshc", .data = &k3_drv_data, },
+	{ .compatible = "hisilicon,hi6220-dw-mshc", .data = &hi6220_data, },
+	{},
+};
+MODULE_DEVICE_TABLE(of, dw_mci_k3_match);
+
+static int dw_mci_k3_probe(struct platform_device *pdev)
+{
+	const struct dw_mci_drv_data *drv_data;
+	const struct of_device_id *match;
+
+	match = of_match_node(dw_mci_k3_match, pdev->dev.of_node);
+	drv_data = match->data;
+
+	return dw_mci_pltfm_register(pdev, drv_data);
+}
+
+static const struct dev_pm_ops dw_mci_k3_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(dw_mci_runtime_suspend,
+			   dw_mci_runtime_resume,
+			   NULL)
+};
+
+static struct platform_driver dw_mci_k3_pltfm_driver = {
+	.probe		= dw_mci_k3_probe,
+	.remove		= dw_mci_pltfm_remove,
+	.driver		= {
+		.name		= "dwmmc_k3",
+		.probe_type	= PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table	= dw_mci_k3_match,
+		.pm		= &dw_mci_k3_dev_pm_ops,
+	},
+};
+
+module_platform_driver(dw_mci_k3_pltfm_driver);
+
+MODULE_DESCRIPTION("K3 Specific DW-MSHC Driver Extension");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:dwmmc_k3");
diff --git a/drivers/mmc/host/dw_mmc-pci.c b/drivers/mmc/host/dw_mmc-pci.c
new file mode 100644
index 000000000..e7ab699f4
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-pci.c
@@ -0,0 +1,105 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Synopsys DesignWare Multimedia Card PCI Interface driver
+ *
+ * Copyright (C) 2012 Vayavya Labs Pvt. Ltd.
+ */
+
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/io.h>
+#include <linux/irq.h>
+#include <linux/pci.h>
+#include <linux/pm_runtime.h>
+#include <linux/slab.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include "dw_mmc.h"
+
+#define PCI_BAR_NO 2
+#define SYNOPSYS_DW_MCI_VENDOR_ID 0x700
+#define SYNOPSYS_DW_MCI_DEVICE_ID 0x1107
+/* Defining the Capabilities */
+#define DW_MCI_CAPABILITIES (MMC_CAP_4_BIT_DATA | MMC_CAP_MMC_HIGHSPEED |\
+				MMC_CAP_SD_HIGHSPEED | MMC_CAP_8_BIT_DATA |\
+				MMC_CAP_SDIO_IRQ)
+
+static struct dw_mci_board pci_board_data = {
+	.caps				= DW_MCI_CAPABILITIES,
+	.bus_hz				= 33 * 1000 * 1000,
+	.detect_delay_ms		= 200,
+	.fifo_depth			= 32,
+};
+
+static int dw_mci_pci_probe(struct pci_dev *pdev,
+			    const struct pci_device_id *entries)
+{
+	struct dw_mci *host;
+	int ret;
+
+	ret = pcim_enable_device(pdev);
+	if (ret)
+		return ret;
+
+	host = devm_kzalloc(&pdev->dev, sizeof(struct dw_mci), GFP_KERNEL);
+	if (!host)
+		return -ENOMEM;
+
+	host->irq = pdev->irq;
+	host->irq_flags = IRQF_SHARED;
+	host->dev = &pdev->dev;
+	host->pdata = &pci_board_data;
+
+	ret = pcim_iomap_regions(pdev, 1 << PCI_BAR_NO, pci_name(pdev));
+	if (ret)
+		return ret;
+
+	host->regs = pcim_iomap_table(pdev)[PCI_BAR_NO];
+
+	pci_set_master(pdev);
+
+	ret = dw_mci_probe(host);
+	if (ret)
+		return ret;
+
+	pci_set_drvdata(pdev, host);
+
+	return 0;
+}
+
+static void dw_mci_pci_remove(struct pci_dev *pdev)
+{
+	struct dw_mci *host = pci_get_drvdata(pdev);
+
+	dw_mci_remove(host);
+}
+
+static const struct dev_pm_ops dw_mci_pci_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(dw_mci_runtime_suspend,
+			   dw_mci_runtime_resume,
+			   NULL)
+};
+
+static const struct pci_device_id dw_mci_pci_id[] = {
+	{ PCI_DEVICE(SYNOPSYS_DW_MCI_VENDOR_ID, SYNOPSYS_DW_MCI_DEVICE_ID) },
+	{}
+};
+MODULE_DEVICE_TABLE(pci, dw_mci_pci_id);
+
+static struct pci_driver dw_mci_pci_driver = {
+	.name		= "dw_mmc_pci",
+	.id_table	= dw_mci_pci_id,
+	.probe		= dw_mci_pci_probe,
+	.remove		= dw_mci_pci_remove,
+	.driver		=	{
+		.pm =   &dw_mci_pci_dev_pm_ops,
+	},
+};
+
+module_pci_driver(dw_mci_pci_driver);
+
+MODULE_DESCRIPTION("DW Multimedia Card PCI Interface driver");
+MODULE_AUTHOR("Shashidhar Hiremath <shashidharh@vayavyalabs.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/dw_mmc-pltfm.c b/drivers/mmc/host/dw_mmc-pltfm.c
new file mode 100644
index 000000000..73731cd3b
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-pltfm.c
@@ -0,0 +1,112 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Synopsys DesignWare Multimedia Card Interface driver
+ *
+ * Copyright (C) 2009 NXP Semiconductors
+ * Copyright (C) 2009, 2010 Imagination Technologies Ltd.
+ */
+
+#include <linux/err.h>
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/io.h>
+#include <linux/irq.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/slab.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/of.h>
+#include <linux/clk.h>
+
+#include "dw_mmc.h"
+#include "dw_mmc-pltfm.h"
+
+int dw_mci_pltfm_register(struct platform_device *pdev,
+			  const struct dw_mci_drv_data *drv_data)
+{
+	struct dw_mci *host;
+	struct resource	*regs;
+
+	host = devm_kzalloc(&pdev->dev, sizeof(struct dw_mci), GFP_KERNEL);
+	if (!host)
+		return -ENOMEM;
+
+	host->irq = platform_get_irq(pdev, 0);
+	if (host->irq < 0)
+		return host->irq;
+
+	host->drv_data = drv_data;
+	host->dev = &pdev->dev;
+	host->irq_flags = 0;
+	host->pdata = pdev->dev.platform_data;
+
+	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	host->regs = devm_ioremap_resource(&pdev->dev, regs);
+	if (IS_ERR(host->regs))
+		return PTR_ERR(host->regs);
+
+	/* Get registers' physical base address */
+	host->phy_regs = regs->start;
+
+	platform_set_drvdata(pdev, host);
+	return dw_mci_probe(host);
+}
+EXPORT_SYMBOL_GPL(dw_mci_pltfm_register);
+
+const struct dev_pm_ops dw_mci_pltfm_pmops = {
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(dw_mci_runtime_suspend,
+			   dw_mci_runtime_resume,
+			   NULL)
+};
+EXPORT_SYMBOL_GPL(dw_mci_pltfm_pmops);
+
+static const struct of_device_id dw_mci_pltfm_match[] = {
+	{ .compatible = "snps,dw-mshc", },
+	{ .compatible = "altr,socfpga-dw-mshc", },
+	{ .compatible = "img,pistachio-dw-mshc", },
+	{},
+};
+MODULE_DEVICE_TABLE(of, dw_mci_pltfm_match);
+
+static int dw_mci_pltfm_probe(struct platform_device *pdev)
+{
+	const struct dw_mci_drv_data *drv_data = NULL;
+	const struct of_device_id *match;
+
+	if (pdev->dev.of_node) {
+		match = of_match_node(dw_mci_pltfm_match, pdev->dev.of_node);
+		drv_data = match->data;
+	}
+
+	return dw_mci_pltfm_register(pdev, drv_data);
+}
+
+int dw_mci_pltfm_remove(struct platform_device *pdev)
+{
+	struct dw_mci *host = platform_get_drvdata(pdev);
+
+	dw_mci_remove(host);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(dw_mci_pltfm_remove);
+
+static struct platform_driver dw_mci_pltfm_driver = {
+	.probe		= dw_mci_pltfm_probe,
+	.remove		= dw_mci_pltfm_remove,
+	.driver		= {
+		.name		= "dw_mmc",
+		.probe_type	= PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table	= dw_mci_pltfm_match,
+		.pm		= &dw_mci_pltfm_pmops,
+	},
+};
+
+module_platform_driver(dw_mci_pltfm_driver);
+
+MODULE_DESCRIPTION("DW Multimedia Card Interface driver");
+MODULE_AUTHOR("NXP Semiconductor VietNam");
+MODULE_AUTHOR("Imagination Technologies Ltd");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/dw_mmc-pltfm.h b/drivers/mmc/host/dw_mmc-pltfm.h
new file mode 100644
index 000000000..2d50d7da2
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-pltfm.h
@@ -0,0 +1,16 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Synopsys DesignWare Multimedia Card Interface Platform driver
+ *
+ * Copyright (C) 2012, Samsung Electronics Co., Ltd.
+ */
+
+#ifndef _DW_MMC_PLTFM_H_
+#define _DW_MMC_PLTFM_H_
+
+extern int dw_mci_pltfm_register(struct platform_device *pdev,
+				const struct dw_mci_drv_data *drv_data);
+extern int dw_mci_pltfm_remove(struct platform_device *pdev);
+extern const struct dev_pm_ops dw_mci_pltfm_pmops;
+
+#endif /* _DW_MMC_PLTFM_H_ */
diff --git a/drivers/mmc/host/dw_mmc-rockchip.c b/drivers/mmc/host/dw_mmc-rockchip.c
new file mode 100644
index 000000000..753502ce3
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-rockchip.c
@@ -0,0 +1,397 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (c) 2014, Fuzhou Rockchip Electronics Co., Ltd
+ */
+
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/clk.h>
+#include <linux/mmc/host.h>
+#include <linux/of_address.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/pm_runtime.h>
+#include <linux/slab.h>
+
+#include "dw_mmc.h"
+#include "dw_mmc-pltfm.h"
+
+#define RK3288_CLKGEN_DIV       2
+
+struct dw_mci_rockchip_priv_data {
+	struct clk		*drv_clk;
+	struct clk		*sample_clk;
+	int			default_sample_phase;
+	int			num_phases;
+};
+
+static void dw_mci_rk3288_set_ios(struct dw_mci *host, struct mmc_ios *ios)
+{
+	struct dw_mci_rockchip_priv_data *priv = host->priv;
+	int ret;
+	unsigned int cclkin;
+	u32 bus_hz;
+
+	if (ios->clock == 0)
+		return;
+
+	/*
+	 * cclkin: source clock of mmc controller
+	 * bus_hz: card interface clock generated by CLKGEN
+	 * bus_hz = cclkin / RK3288_CLKGEN_DIV
+	 * ios->clock = (div == 0) ? bus_hz : (bus_hz / (2 * div))
+	 *
+	 * Note: div can only be 0 or 1, but div must be set to 1 for eMMC
+	 * DDR52 8-bit mode.
+	 */
+	if (ios->bus_width == MMC_BUS_WIDTH_8 &&
+	    ios->timing == MMC_TIMING_MMC_DDR52)
+		cclkin = 2 * ios->clock * RK3288_CLKGEN_DIV;
+	else
+		cclkin = ios->clock * RK3288_CLKGEN_DIV;
+
+	ret = clk_set_rate(host->ciu_clk, cclkin);
+	if (ret)
+		dev_warn(host->dev, "failed to set rate %uHz\n", ios->clock);
+
+	bus_hz = clk_get_rate(host->ciu_clk) / RK3288_CLKGEN_DIV;
+	if (bus_hz != host->bus_hz) {
+		host->bus_hz = bus_hz;
+		/* force dw_mci_setup_bus() */
+		host->current_speed = 0;
+	}
+
+	/* Make sure we use phases which we can enumerate with */
+	if (!IS_ERR(priv->sample_clk))
+		clk_set_phase(priv->sample_clk, priv->default_sample_phase);
+
+	/*
+	 * Set the drive phase offset based on speed mode to achieve hold times.
+	 *
+	 * NOTE: this is _not_ a value that is dynamically tuned and is also
+	 * _not_ a value that will vary from board to board.  It is a value
+	 * that could vary between different SoC models if they had massively
+	 * different output clock delays inside their dw_mmc IP block (delay_o),
+	 * but since it's OK to overshoot a little we don't need to do complex
+	 * calculations and can pick values that will just work for everyone.
+	 *
+	 * When picking values we'll stick with picking 0/90/180/270 since
+	 * those can be made very accurately on all known Rockchip SoCs.
+	 *
+	 * Note that these values match values from the DesignWare Databook
+	 * tables for the most part except for SDR12 and "ID mode".  For those
+	 * two modes the databook calculations assume a clock in of 50MHz.  As
+	 * seen above, we always use a clock in rate that is exactly the
+	 * card's input clock (times RK3288_CLKGEN_DIV, but that gets divided
+	 * back out before the controller sees it).
+	 *
+	 * From measurement of a single device, it appears that delay_o is
+	 * about .5 ns.  Since we try to leave a bit of margin, it's expected
+	 * that numbers here will be fine even with much larger delay_o
+	 * (the 1.4 ns assumed by the DesignWare Databook would result in the
+	 * same results, for instance).
+	 */
+	if (!IS_ERR(priv->drv_clk)) {
+		int phase;
+
+		/*
+		 * In almost all cases a 90 degree phase offset will provide
+		 * sufficient hold times across all valid input clock rates
+		 * assuming delay_o is not absurd for a given SoC.  We'll use
+		 * that as a default.
+		 */
+		phase = 90;
+
+		switch (ios->timing) {
+		case MMC_TIMING_MMC_DDR52:
+			/*
+			 * Since clock in rate with MMC_DDR52 is doubled when
+			 * bus width is 8 we need to double the phase offset
+			 * to get the same timings.
+			 */
+			if (ios->bus_width == MMC_BUS_WIDTH_8)
+				phase = 180;
+			break;
+		case MMC_TIMING_UHS_SDR104:
+		case MMC_TIMING_MMC_HS200:
+			/*
+			 * In the case of 150 MHz clock (typical max for
+			 * Rockchip SoCs), 90 degree offset will add a delay
+			 * of 1.67 ns.  That will meet min hold time of .8 ns
+			 * as long as clock output delay is < .87 ns.  On
+			 * SoCs measured this seems to be OK, but it doesn't
+			 * hurt to give margin here, so we use 180.
+			 */
+			phase = 180;
+			break;
+		}
+
+		clk_set_phase(priv->drv_clk, phase);
+	}
+}
+
+#define TUNING_ITERATION_TO_PHASE(i, num_phases) \
+		(DIV_ROUND_UP((i) * 360, num_phases))
+
+static int dw_mci_rk3288_execute_tuning(struct dw_mci_slot *slot, u32 opcode)
+{
+	struct dw_mci *host = slot->host;
+	struct dw_mci_rockchip_priv_data *priv = host->priv;
+	struct mmc_host *mmc = slot->mmc;
+	int ret = 0;
+	int i;
+	bool v, prev_v = 0, first_v;
+	struct range_t {
+		int start;
+		int end; /* inclusive */
+	};
+	struct range_t *ranges;
+	unsigned int range_count = 0;
+	int longest_range_len = -1;
+	int longest_range = -1;
+	int middle_phase;
+
+	if (IS_ERR(priv->sample_clk)) {
+		dev_err(host->dev, "Tuning clock (sample_clk) not defined.\n");
+		return -EIO;
+	}
+
+	ranges = kmalloc_array(priv->num_phases / 2 + 1,
+			       sizeof(*ranges), GFP_KERNEL);
+	if (!ranges)
+		return -ENOMEM;
+
+	/* Try each phase and extract good ranges */
+	for (i = 0; i < priv->num_phases; ) {
+		clk_set_phase(priv->sample_clk,
+			      TUNING_ITERATION_TO_PHASE(i, priv->num_phases));
+
+		v = !mmc_send_tuning(mmc, opcode, NULL);
+
+		if (i == 0)
+			first_v = v;
+
+		if ((!prev_v) && v) {
+			range_count++;
+			ranges[range_count-1].start = i;
+		}
+		if (v) {
+			ranges[range_count-1].end = i;
+			i++;
+		} else if (i == priv->num_phases - 1) {
+			/* No extra skipping rules if we're at the end */
+			i++;
+		} else {
+			/*
+			 * No need to check too close to an invalid
+			 * one since testing bad phases is slow.  Skip
+			 * 20 degrees.
+			 */
+			i += DIV_ROUND_UP(20 * priv->num_phases, 360);
+
+			/* Always test the last one */
+			if (i >= priv->num_phases)
+				i = priv->num_phases - 1;
+		}
+
+		prev_v = v;
+	}
+
+	if (range_count == 0) {
+		dev_warn(host->dev, "All phases bad!");
+		ret = -EIO;
+		goto free;
+	}
+
+	/* wrap around case, merge the end points */
+	if ((range_count > 1) && first_v && v) {
+		ranges[0].start = ranges[range_count-1].start;
+		range_count--;
+	}
+
+	if (ranges[0].start == 0 && ranges[0].end == priv->num_phases - 1) {
+		clk_set_phase(priv->sample_clk, priv->default_sample_phase);
+		dev_info(host->dev, "All phases work, using default phase %d.",
+			 priv->default_sample_phase);
+		goto free;
+	}
+
+	/* Find the longest range */
+	for (i = 0; i < range_count; i++) {
+		int len = (ranges[i].end - ranges[i].start + 1);
+
+		if (len < 0)
+			len += priv->num_phases;
+
+		if (longest_range_len < len) {
+			longest_range_len = len;
+			longest_range = i;
+		}
+
+		dev_dbg(host->dev, "Good phase range %d-%d (%d len)\n",
+			TUNING_ITERATION_TO_PHASE(ranges[i].start,
+						  priv->num_phases),
+			TUNING_ITERATION_TO_PHASE(ranges[i].end,
+						  priv->num_phases),
+			len
+		);
+	}
+
+	dev_dbg(host->dev, "Best phase range %d-%d (%d len)\n",
+		TUNING_ITERATION_TO_PHASE(ranges[longest_range].start,
+					  priv->num_phases),
+		TUNING_ITERATION_TO_PHASE(ranges[longest_range].end,
+					  priv->num_phases),
+		longest_range_len
+	);
+
+	middle_phase = ranges[longest_range].start + longest_range_len / 2;
+	middle_phase %= priv->num_phases;
+	dev_info(host->dev, "Successfully tuned phase to %d\n",
+		 TUNING_ITERATION_TO_PHASE(middle_phase, priv->num_phases));
+
+	clk_set_phase(priv->sample_clk,
+		      TUNING_ITERATION_TO_PHASE(middle_phase,
+						priv->num_phases));
+
+free:
+	kfree(ranges);
+	return ret;
+}
+
+static int dw_mci_rk3288_parse_dt(struct dw_mci *host)
+{
+	struct device_node *np = host->dev->of_node;
+	struct dw_mci_rockchip_priv_data *priv;
+
+	priv = devm_kzalloc(host->dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	if (of_property_read_u32(np, "rockchip,desired-num-phases",
+					&priv->num_phases))
+		priv->num_phases = 360;
+
+	if (of_property_read_u32(np, "rockchip,default-sample-phase",
+					&priv->default_sample_phase))
+		priv->default_sample_phase = 0;
+
+	priv->drv_clk = devm_clk_get(host->dev, "ciu-drive");
+	if (IS_ERR(priv->drv_clk))
+		dev_dbg(host->dev, "ciu-drive not available\n");
+
+	priv->sample_clk = devm_clk_get(host->dev, "ciu-sample");
+	if (IS_ERR(priv->sample_clk))
+		dev_dbg(host->dev, "ciu-sample not available\n");
+
+	host->priv = priv;
+
+	return 0;
+}
+
+static int dw_mci_rockchip_init(struct dw_mci *host)
+{
+	/* It is slot 8 on Rockchip SoCs */
+	host->sdio_id0 = 8;
+
+	if (of_device_is_compatible(host->dev->of_node,
+				    "rockchip,rk3288-dw-mshc"))
+		host->bus_hz /= RK3288_CLKGEN_DIV;
+
+	return 0;
+}
+
+/* Common capabilities of RK3288 SoC */
+static unsigned long dw_mci_rk3288_dwmmc_caps[4] = {
+	MMC_CAP_CMD23,
+	MMC_CAP_CMD23,
+	MMC_CAP_CMD23,
+	MMC_CAP_CMD23,
+};
+
+static const struct dw_mci_drv_data rk2928_drv_data = {
+	.init			= dw_mci_rockchip_init,
+};
+
+static const struct dw_mci_drv_data rk3288_drv_data = {
+	.caps			= dw_mci_rk3288_dwmmc_caps,
+	.num_caps		= ARRAY_SIZE(dw_mci_rk3288_dwmmc_caps),
+	.set_ios		= dw_mci_rk3288_set_ios,
+	.execute_tuning		= dw_mci_rk3288_execute_tuning,
+	.parse_dt		= dw_mci_rk3288_parse_dt,
+	.init			= dw_mci_rockchip_init,
+};
+
+static const struct of_device_id dw_mci_rockchip_match[] = {
+	{ .compatible = "rockchip,rk2928-dw-mshc",
+		.data = &rk2928_drv_data },
+	{ .compatible = "rockchip,rk3288-dw-mshc",
+		.data = &rk3288_drv_data },
+	{},
+};
+MODULE_DEVICE_TABLE(of, dw_mci_rockchip_match);
+
+static int dw_mci_rockchip_probe(struct platform_device *pdev)
+{
+	const struct dw_mci_drv_data *drv_data;
+	const struct of_device_id *match;
+	int ret;
+
+	if (!pdev->dev.of_node)
+		return -ENODEV;
+
+	match = of_match_node(dw_mci_rockchip_match, pdev->dev.of_node);
+	drv_data = match->data;
+
+	pm_runtime_get_noresume(&pdev->dev);
+	pm_runtime_set_active(&pdev->dev);
+	pm_runtime_enable(&pdev->dev);
+	pm_runtime_set_autosuspend_delay(&pdev->dev, 50);
+	pm_runtime_use_autosuspend(&pdev->dev);
+
+	ret = dw_mci_pltfm_register(pdev, drv_data);
+	if (ret) {
+		pm_runtime_disable(&pdev->dev);
+		pm_runtime_set_suspended(&pdev->dev);
+		pm_runtime_put_noidle(&pdev->dev);
+		return ret;
+	}
+
+	pm_runtime_put_autosuspend(&pdev->dev);
+
+	return 0;
+}
+
+static int dw_mci_rockchip_remove(struct platform_device *pdev)
+{
+	pm_runtime_get_sync(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+
+	return dw_mci_pltfm_remove(pdev);
+}
+
+static const struct dev_pm_ops dw_mci_rockchip_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(dw_mci_runtime_suspend,
+			   dw_mci_runtime_resume,
+			   NULL)
+};
+
+static struct platform_driver dw_mci_rockchip_pltfm_driver = {
+	.probe		= dw_mci_rockchip_probe,
+	.remove		= dw_mci_rockchip_remove,
+	.driver		= {
+		.name		= "dwmmc_rockchip",
+		.probe_type	= PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table	= dw_mci_rockchip_match,
+		.pm		= &dw_mci_rockchip_dev_pm_ops,
+	},
+};
+
+module_platform_driver(dw_mci_rockchip_pltfm_driver);
+
+MODULE_AUTHOR("Addy Ke <addy.ke@rock-chips.com>");
+MODULE_DESCRIPTION("Rockchip Specific DW-MSHC Driver Extension");
+MODULE_ALIAS("platform:dwmmc_rockchip");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/dw_mmc-zx.c b/drivers/mmc/host/dw_mmc-zx.c
new file mode 100644
index 000000000..51bcc6332
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-zx.c
@@ -0,0 +1,234 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * ZX Specific Extensions for Synopsys DW Multimedia Card Interface driver
+ *
+ * Copyright (C) 2016, Linaro Ltd.
+ * Copyright (C) 2016, ZTE Corp.
+ */
+
+#include <linux/clk.h>
+#include <linux/mfd/syscon.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/regmap.h>
+#include <linux/slab.h>
+
+#include "dw_mmc.h"
+#include "dw_mmc-pltfm.h"
+#include "dw_mmc-zx.h"
+
+struct dw_mci_zx_priv_data {
+	struct regmap	*sysc_base;
+};
+
+enum delay_type {
+	DELAY_TYPE_READ,	/* read dqs delay */
+	DELAY_TYPE_CLK,		/* clk sample delay */
+};
+
+static int dw_mci_zx_emmc_set_delay(struct dw_mci *host, unsigned int delay,
+				    enum delay_type dflag)
+{
+	struct dw_mci_zx_priv_data *priv = host->priv;
+	struct regmap *sysc_base = priv->sysc_base;
+	unsigned int clksel;
+	unsigned int loop = 1000;
+	int ret;
+
+	if (!sysc_base)
+		return -EINVAL;
+
+	ret = regmap_update_bits(sysc_base, LB_AON_EMMC_CFG_REG0,
+				 PARA_HALF_CLK_MODE | PARA_DLL_BYPASS_MODE |
+				 PARA_PHASE_DET_SEL_MASK |
+				 PARA_DLL_LOCK_NUM_MASK |
+				 DLL_REG_SET | PARA_DLL_START_MASK,
+				 PARA_DLL_START(4) | PARA_DLL_LOCK_NUM(4));
+	if (ret)
+		return ret;
+
+	ret = regmap_read(sysc_base, LB_AON_EMMC_CFG_REG1, &clksel);
+	if (ret)
+		return ret;
+
+	if (dflag == DELAY_TYPE_CLK) {
+		clksel &= ~CLK_SAMP_DELAY_MASK;
+		clksel |= CLK_SAMP_DELAY(delay);
+	} else {
+		clksel &= ~READ_DQS_DELAY_MASK;
+		clksel |= READ_DQS_DELAY(delay);
+	}
+
+	regmap_write(sysc_base, LB_AON_EMMC_CFG_REG1, clksel);
+	regmap_update_bits(sysc_base, LB_AON_EMMC_CFG_REG0,
+			   PARA_DLL_START_MASK | PARA_DLL_LOCK_NUM_MASK |
+			   DLL_REG_SET,
+			   PARA_DLL_START(4) | PARA_DLL_LOCK_NUM(4) |
+			   DLL_REG_SET);
+
+	do {
+		ret = regmap_read(sysc_base, LB_AON_EMMC_CFG_REG2, &clksel);
+		if (ret)
+			return ret;
+
+	} while (--loop && !(clksel & ZX_DLL_LOCKED));
+
+	if (!loop) {
+		dev_err(host->dev, "Error: %s dll lock fail\n", __func__);
+		return -EIO;
+	}
+
+	return 0;
+}
+
+static int dw_mci_zx_emmc_execute_tuning(struct dw_mci_slot *slot, u32 opcode)
+{
+	struct dw_mci *host = slot->host;
+	struct mmc_host *mmc = slot->mmc;
+	int ret, len = 0, start = 0, end = 0, delay, best = 0;
+
+	for (delay = 1; delay < 128; delay++) {
+		ret = dw_mci_zx_emmc_set_delay(host, delay, DELAY_TYPE_CLK);
+		if (!ret && mmc_send_tuning(mmc, opcode, NULL)) {
+			if (start >= 0) {
+				end = delay - 1;
+				/* check and update longest good range */
+				if ((end - start) > len) {
+					best = (start + end) >> 1;
+					len = end - start;
+				}
+			}
+			start = -1;
+			end = 0;
+			continue;
+		}
+		if (start < 0)
+			start = delay;
+	}
+
+	if (start >= 0) {
+		end = delay - 1;
+		if ((end - start) > len) {
+			best = (start + end) >> 1;
+			len = end - start;
+		}
+	}
+	if (best < 0)
+		return -EIO;
+
+	dev_info(host->dev, "%s best range: start %d end %d\n", __func__,
+		 start, end);
+	return dw_mci_zx_emmc_set_delay(host, best, DELAY_TYPE_CLK);
+}
+
+static int dw_mci_zx_prepare_hs400_tuning(struct dw_mci *host,
+					  struct mmc_ios *ios)
+{
+	int ret;
+
+	/* config phase shift as 90 degree */
+	ret = dw_mci_zx_emmc_set_delay(host, 32, DELAY_TYPE_READ);
+	if (ret < 0)
+		return -EIO;
+
+	return 0;
+}
+
+static int dw_mci_zx_execute_tuning(struct dw_mci_slot *slot, u32 opcode)
+{
+	struct dw_mci *host = slot->host;
+
+	if (host->verid == 0x290a) /* only for emmc */
+		return dw_mci_zx_emmc_execute_tuning(slot, opcode);
+	/* TODO: Add 0x210a dedicated tuning for sd/sdio */
+
+	return 0;
+}
+
+static int dw_mci_zx_parse_dt(struct dw_mci *host)
+{
+	struct device_node *np = host->dev->of_node;
+	struct device_node *node;
+	struct dw_mci_zx_priv_data *priv;
+	struct regmap *sysc_base;
+
+	/* syscon is needed only by emmc */
+	node = of_parse_phandle(np, "zte,aon-syscon", 0);
+	if (node) {
+		sysc_base = syscon_node_to_regmap(node);
+		of_node_put(node);
+
+		if (IS_ERR(sysc_base))
+			return dev_err_probe(host->dev, PTR_ERR(sysc_base),
+					     "Can't get syscon\n");
+	} else {
+		return 0;
+	}
+
+	priv = devm_kzalloc(host->dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+	priv->sysc_base = sysc_base;
+	host->priv = priv;
+
+	return 0;
+}
+
+static unsigned long zx_dwmmc_caps[3] = {
+	MMC_CAP_CMD23,
+	MMC_CAP_CMD23,
+	MMC_CAP_CMD23,
+};
+
+static const struct dw_mci_drv_data zx_drv_data = {
+	.caps			= zx_dwmmc_caps,
+	.num_caps		= ARRAY_SIZE(zx_dwmmc_caps),
+	.execute_tuning		= dw_mci_zx_execute_tuning,
+	.prepare_hs400_tuning	= dw_mci_zx_prepare_hs400_tuning,
+	.parse_dt               = dw_mci_zx_parse_dt,
+};
+
+static const struct of_device_id dw_mci_zx_match[] = {
+	{ .compatible = "zte,zx296718-dw-mshc", .data = &zx_drv_data},
+	{},
+};
+MODULE_DEVICE_TABLE(of, dw_mci_zx_match);
+
+static int dw_mci_zx_probe(struct platform_device *pdev)
+{
+	const struct dw_mci_drv_data *drv_data;
+	const struct of_device_id *match;
+
+	match = of_match_node(dw_mci_zx_match, pdev->dev.of_node);
+	drv_data = match->data;
+
+	return dw_mci_pltfm_register(pdev, drv_data);
+}
+
+static const struct dev_pm_ops dw_mci_zx_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(dw_mci_runtime_suspend,
+			   dw_mci_runtime_resume,
+			   NULL)
+};
+
+static struct platform_driver dw_mci_zx_pltfm_driver = {
+	.probe		= dw_mci_zx_probe,
+	.remove		= dw_mci_pltfm_remove,
+	.driver		= {
+		.name		= "dwmmc_zx",
+		.probe_type	= PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table	= dw_mci_zx_match,
+		.pm		= &dw_mci_zx_dev_pm_ops,
+	},
+};
+
+module_platform_driver(dw_mci_zx_pltfm_driver);
+
+MODULE_DESCRIPTION("ZTE emmc/sd driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/dw_mmc-zx.h b/drivers/mmc/host/dw_mmc-zx.h
new file mode 100644
index 000000000..09ac52766
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-zx.h
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _DW_MMC_ZX_H_
+#define _DW_MMC_ZX_H_
+
+/* ZX296718 SoC specific DLL register offset. */
+#define LB_AON_EMMC_CFG_REG0  0x1B0
+#define LB_AON_EMMC_CFG_REG1  0x1B4
+#define LB_AON_EMMC_CFG_REG2  0x1B8
+
+/* LB_AON_EMMC_CFG_REG0 register defines */
+#define PARA_DLL_START(x)	((x) & 0xFF)
+#define PARA_DLL_START_MASK	0xFF
+#define DLL_REG_SET		BIT(8)
+#define PARA_DLL_LOCK_NUM(x)	(((x) & 7) << 16)
+#define PARA_DLL_LOCK_NUM_MASK  (7 << 16)
+#define PARA_PHASE_DET_SEL(x)	(((x) & 7) << 20)
+#define PARA_PHASE_DET_SEL_MASK	(7 << 20)
+#define PARA_DLL_BYPASS_MODE	BIT(23)
+#define PARA_HALF_CLK_MODE	BIT(24)
+
+/* LB_AON_EMMC_CFG_REG1 register defines */
+#define READ_DQS_DELAY(x)	((x) & 0x7F)
+#define READ_DQS_DELAY_MASK	(0x7F)
+#define READ_DQS_BYPASS_MODE	BIT(7)
+#define CLK_SAMP_DELAY(x)	(((x) & 0x7F) << 8)
+#define CLK_SAMP_DELAY_MASK	(0x7F << 8)
+#define CLK_SAMP_BYPASS_MODE	BIT(15)
+
+/* LB_AON_EMMC_CFG_REG2 register defines */
+#define ZX_DLL_LOCKED		BIT(2)
+
+#endif /* _DW_MMC_ZX_H_ */
diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c
new file mode 100644
index 000000000..a6170f80b
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc.c
@@ -0,0 +1,3493 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Synopsys DesignWare Multimedia Card Interface driver
+ *  (Based on NXP driver for lpc 31xx)
+ *
+ * Copyright (C) 2009 NXP Semiconductors
+ * Copyright (C) 2009, 2010 Imagination Technologies Ltd.
+ */
+
+#include <linux/blkdev.h>
+#include <linux/clk.h>
+#include <linux/debugfs.h>
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/iopoll.h>
+#include <linux/ioport.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/seq_file.h>
+#include <linux/slab.h>
+#include <linux/stat.h>
+#include <linux/delay.h>
+#include <linux/irq.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sd.h>
+#include <linux/mmc/sdio.h>
+#include <linux/bitops.h>
+#include <linux/regulator/consumer.h>
+#include <linux/of.h>
+#include <linux/of_gpio.h>
+#include <linux/mmc/slot-gpio.h>
+
+#include "dw_mmc.h"
+
+/* Common flag combinations */
+#define DW_MCI_DATA_ERROR_FLAGS	(SDMMC_INT_DRTO | SDMMC_INT_DCRC | \
+				 SDMMC_INT_HTO | SDMMC_INT_SBE  | \
+				 SDMMC_INT_EBE | SDMMC_INT_HLE)
+#define DW_MCI_CMD_ERROR_FLAGS	(SDMMC_INT_RTO | SDMMC_INT_RCRC | \
+				 SDMMC_INT_RESP_ERR | SDMMC_INT_HLE)
+#define DW_MCI_ERROR_FLAGS	(DW_MCI_DATA_ERROR_FLAGS | \
+				 DW_MCI_CMD_ERROR_FLAGS)
+#define DW_MCI_SEND_STATUS	1
+#define DW_MCI_RECV_STATUS	2
+#define DW_MCI_DMA_THRESHOLD	16
+
+#define DW_MCI_FREQ_MAX	200000000	/* unit: HZ */
+#define DW_MCI_FREQ_MIN	100000		/* unit: HZ */
+
+#define IDMAC_INT_CLR		(SDMMC_IDMAC_INT_AI | SDMMC_IDMAC_INT_NI | \
+				 SDMMC_IDMAC_INT_CES | SDMMC_IDMAC_INT_DU | \
+				 SDMMC_IDMAC_INT_FBE | SDMMC_IDMAC_INT_RI | \
+				 SDMMC_IDMAC_INT_TI)
+
+#define DESC_RING_BUF_SZ	PAGE_SIZE
+
+struct idmac_desc_64addr {
+	u32		des0;	/* Control Descriptor */
+#define IDMAC_OWN_CLR64(x) \
+	!((x) & cpu_to_le32(IDMAC_DES0_OWN))
+
+	u32		des1;	/* Reserved */
+
+	u32		des2;	/*Buffer sizes */
+#define IDMAC_64ADDR_SET_BUFFER1_SIZE(d, s) \
+	((d)->des2 = ((d)->des2 & cpu_to_le32(0x03ffe000)) | \
+	 ((cpu_to_le32(s)) & cpu_to_le32(0x1fff)))
+
+	u32		des3;	/* Reserved */
+
+	u32		des4;	/* Lower 32-bits of Buffer Address Pointer 1*/
+	u32		des5;	/* Upper 32-bits of Buffer Address Pointer 1*/
+
+	u32		des6;	/* Lower 32-bits of Next Descriptor Address */
+	u32		des7;	/* Upper 32-bits of Next Descriptor Address */
+};
+
+struct idmac_desc {
+	__le32		des0;	/* Control Descriptor */
+#define IDMAC_DES0_DIC	BIT(1)
+#define IDMAC_DES0_LD	BIT(2)
+#define IDMAC_DES0_FD	BIT(3)
+#define IDMAC_DES0_CH	BIT(4)
+#define IDMAC_DES0_ER	BIT(5)
+#define IDMAC_DES0_CES	BIT(30)
+#define IDMAC_DES0_OWN	BIT(31)
+
+	__le32		des1;	/* Buffer sizes */
+#define IDMAC_SET_BUFFER1_SIZE(d, s) \
+	((d)->des1 = ((d)->des1 & cpu_to_le32(0x03ffe000)) | (cpu_to_le32((s) & 0x1fff)))
+
+	__le32		des2;	/* buffer 1 physical address */
+
+	__le32		des3;	/* buffer 2 physical address */
+};
+
+/* Each descriptor can transfer up to 4KB of data in chained mode */
+#define DW_MCI_DESC_DATA_LENGTH	0x1000
+
+#if defined(CONFIG_DEBUG_FS)
+static int dw_mci_req_show(struct seq_file *s, void *v)
+{
+	struct dw_mci_slot *slot = s->private;
+	struct mmc_request *mrq;
+	struct mmc_command *cmd;
+	struct mmc_command *stop;
+	struct mmc_data	*data;
+
+	/* Make sure we get a consistent snapshot */
+	spin_lock_bh(&slot->host->lock);
+	mrq = slot->mrq;
+
+	if (mrq) {
+		cmd = mrq->cmd;
+		data = mrq->data;
+		stop = mrq->stop;
+
+		if (cmd)
+			seq_printf(s,
+				   "CMD%u(0x%x) flg %x rsp %x %x %x %x err %d\n",
+				   cmd->opcode, cmd->arg, cmd->flags,
+				   cmd->resp[0], cmd->resp[1], cmd->resp[2],
+				   cmd->resp[2], cmd->error);
+		if (data)
+			seq_printf(s, "DATA %u / %u * %u flg %x err %d\n",
+				   data->bytes_xfered, data->blocks,
+				   data->blksz, data->flags, data->error);
+		if (stop)
+			seq_printf(s,
+				   "CMD%u(0x%x) flg %x rsp %x %x %x %x err %d\n",
+				   stop->opcode, stop->arg, stop->flags,
+				   stop->resp[0], stop->resp[1], stop->resp[2],
+				   stop->resp[2], stop->error);
+	}
+
+	spin_unlock_bh(&slot->host->lock);
+
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(dw_mci_req);
+
+static int dw_mci_regs_show(struct seq_file *s, void *v)
+{
+	struct dw_mci *host = s->private;
+
+	pm_runtime_get_sync(host->dev);
+
+	seq_printf(s, "STATUS:\t0x%08x\n", mci_readl(host, STATUS));
+	seq_printf(s, "RINTSTS:\t0x%08x\n", mci_readl(host, RINTSTS));
+	seq_printf(s, "CMD:\t0x%08x\n", mci_readl(host, CMD));
+	seq_printf(s, "CTRL:\t0x%08x\n", mci_readl(host, CTRL));
+	seq_printf(s, "INTMASK:\t0x%08x\n", mci_readl(host, INTMASK));
+	seq_printf(s, "CLKENA:\t0x%08x\n", mci_readl(host, CLKENA));
+
+	pm_runtime_put_autosuspend(host->dev);
+
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(dw_mci_regs);
+
+static void dw_mci_init_debugfs(struct dw_mci_slot *slot)
+{
+	struct mmc_host	*mmc = slot->mmc;
+	struct dw_mci *host = slot->host;
+	struct dentry *root;
+
+	root = mmc->debugfs_root;
+	if (!root)
+		return;
+
+	debugfs_create_file("regs", S_IRUSR, root, host, &dw_mci_regs_fops);
+	debugfs_create_file("req", S_IRUSR, root, slot, &dw_mci_req_fops);
+	debugfs_create_u32("state", S_IRUSR, root, &host->state);
+	debugfs_create_xul("pending_events", S_IRUSR, root,
+			   &host->pending_events);
+	debugfs_create_xul("completed_events", S_IRUSR, root,
+			   &host->completed_events);
+}
+#endif /* defined(CONFIG_DEBUG_FS) */
+
+static bool dw_mci_ctrl_reset(struct dw_mci *host, u32 reset)
+{
+	u32 ctrl;
+
+	ctrl = mci_readl(host, CTRL);
+	ctrl |= reset;
+	mci_writel(host, CTRL, ctrl);
+
+	/* wait till resets clear */
+	if (readl_poll_timeout_atomic(host->regs + SDMMC_CTRL, ctrl,
+				      !(ctrl & reset),
+				      1, 500 * USEC_PER_MSEC)) {
+		dev_err(host->dev,
+			"Timeout resetting block (ctrl reset %#x)\n",
+			ctrl & reset);
+		return false;
+	}
+
+	return true;
+}
+
+static void dw_mci_wait_while_busy(struct dw_mci *host, u32 cmd_flags)
+{
+	u32 status;
+
+	/*
+	 * Databook says that before issuing a new data transfer command
+	 * we need to check to see if the card is busy.  Data transfer commands
+	 * all have SDMMC_CMD_PRV_DAT_WAIT set, so we'll key off that.
+	 *
+	 * ...also allow sending for SDMMC_CMD_VOLT_SWITCH where busy is
+	 * expected.
+	 */
+	if ((cmd_flags & SDMMC_CMD_PRV_DAT_WAIT) &&
+	    !(cmd_flags & SDMMC_CMD_VOLT_SWITCH)) {
+		if (readl_poll_timeout_atomic(host->regs + SDMMC_STATUS,
+					      status,
+					      !(status & SDMMC_STATUS_BUSY),
+					      10, 500 * USEC_PER_MSEC))
+			dev_err(host->dev, "Busy; trying anyway\n");
+	}
+}
+
+static void mci_send_cmd(struct dw_mci_slot *slot, u32 cmd, u32 arg)
+{
+	struct dw_mci *host = slot->host;
+	unsigned int cmd_status = 0;
+
+	mci_writel(host, CMDARG, arg);
+	wmb(); /* drain writebuffer */
+	dw_mci_wait_while_busy(host, cmd);
+	mci_writel(host, CMD, SDMMC_CMD_START | cmd);
+
+	if (readl_poll_timeout_atomic(host->regs + SDMMC_CMD, cmd_status,
+				      !(cmd_status & SDMMC_CMD_START),
+				      1, 500 * USEC_PER_MSEC))
+		dev_err(&slot->mmc->class_dev,
+			"Timeout sending command (cmd %#x arg %#x status %#x)\n",
+			cmd, arg, cmd_status);
+}
+
+static u32 dw_mci_prepare_command(struct mmc_host *mmc, struct mmc_command *cmd)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct dw_mci *host = slot->host;
+	u32 cmdr;
+
+	cmd->error = -EINPROGRESS;
+	cmdr = cmd->opcode;
+
+	if (cmd->opcode == MMC_STOP_TRANSMISSION ||
+	    cmd->opcode == MMC_GO_IDLE_STATE ||
+	    cmd->opcode == MMC_GO_INACTIVE_STATE ||
+	    (cmd->opcode == SD_IO_RW_DIRECT &&
+	     ((cmd->arg >> 9) & 0x1FFFF) == SDIO_CCCR_ABORT))
+		cmdr |= SDMMC_CMD_STOP;
+	else if (cmd->opcode != MMC_SEND_STATUS && cmd->data)
+		cmdr |= SDMMC_CMD_PRV_DAT_WAIT;
+
+	if (cmd->opcode == SD_SWITCH_VOLTAGE) {
+		u32 clk_en_a;
+
+		/* Special bit makes CMD11 not die */
+		cmdr |= SDMMC_CMD_VOLT_SWITCH;
+
+		/* Change state to continue to handle CMD11 weirdness */
+		WARN_ON(slot->host->state != STATE_SENDING_CMD);
+		slot->host->state = STATE_SENDING_CMD11;
+
+		/*
+		 * We need to disable low power mode (automatic clock stop)
+		 * while doing voltage switch so we don't confuse the card,
+		 * since stopping the clock is a specific part of the UHS
+		 * voltage change dance.
+		 *
+		 * Note that low power mode (SDMMC_CLKEN_LOW_PWR) will be
+		 * unconditionally turned back on in dw_mci_setup_bus() if it's
+		 * ever called with a non-zero clock.  That shouldn't happen
+		 * until the voltage change is all done.
+		 */
+		clk_en_a = mci_readl(host, CLKENA);
+		clk_en_a &= ~(SDMMC_CLKEN_LOW_PWR << slot->id);
+		mci_writel(host, CLKENA, clk_en_a);
+		mci_send_cmd(slot, SDMMC_CMD_UPD_CLK |
+			     SDMMC_CMD_PRV_DAT_WAIT, 0);
+	}
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		/* We expect a response, so set this bit */
+		cmdr |= SDMMC_CMD_RESP_EXP;
+		if (cmd->flags & MMC_RSP_136)
+			cmdr |= SDMMC_CMD_RESP_LONG;
+	}
+
+	if (cmd->flags & MMC_RSP_CRC)
+		cmdr |= SDMMC_CMD_RESP_CRC;
+
+	if (cmd->data) {
+		cmdr |= SDMMC_CMD_DAT_EXP;
+		if (cmd->data->flags & MMC_DATA_WRITE)
+			cmdr |= SDMMC_CMD_DAT_WR;
+	}
+
+	if (!test_bit(DW_MMC_CARD_NO_USE_HOLD, &slot->flags))
+		cmdr |= SDMMC_CMD_USE_HOLD_REG;
+
+	return cmdr;
+}
+
+static u32 dw_mci_prep_stop_abort(struct dw_mci *host, struct mmc_command *cmd)
+{
+	struct mmc_command *stop;
+	u32 cmdr;
+
+	if (!cmd->data)
+		return 0;
+
+	stop = &host->stop_abort;
+	cmdr = cmd->opcode;
+	memset(stop, 0, sizeof(struct mmc_command));
+
+	if (cmdr == MMC_READ_SINGLE_BLOCK ||
+	    cmdr == MMC_READ_MULTIPLE_BLOCK ||
+	    cmdr == MMC_WRITE_BLOCK ||
+	    cmdr == MMC_WRITE_MULTIPLE_BLOCK ||
+	    cmdr == MMC_SEND_TUNING_BLOCK ||
+	    cmdr == MMC_SEND_TUNING_BLOCK_HS200) {
+		stop->opcode = MMC_STOP_TRANSMISSION;
+		stop->arg = 0;
+		stop->flags = MMC_RSP_R1B | MMC_CMD_AC;
+	} else if (cmdr == SD_IO_RW_EXTENDED) {
+		stop->opcode = SD_IO_RW_DIRECT;
+		stop->arg |= (1 << 31) | (0 << 28) | (SDIO_CCCR_ABORT << 9) |
+			     ((cmd->arg >> 28) & 0x7);
+		stop->flags = MMC_RSP_SPI_R5 | MMC_RSP_R5 | MMC_CMD_AC;
+	} else {
+		return 0;
+	}
+
+	cmdr = stop->opcode | SDMMC_CMD_STOP |
+		SDMMC_CMD_RESP_CRC | SDMMC_CMD_RESP_EXP;
+
+	if (!test_bit(DW_MMC_CARD_NO_USE_HOLD, &host->slot->flags))
+		cmdr |= SDMMC_CMD_USE_HOLD_REG;
+
+	return cmdr;
+}
+
+static inline void dw_mci_set_cto(struct dw_mci *host)
+{
+	unsigned int cto_clks;
+	unsigned int cto_div;
+	unsigned int cto_ms;
+	unsigned long irqflags;
+
+	cto_clks = mci_readl(host, TMOUT) & 0xff;
+	cto_div = (mci_readl(host, CLKDIV) & 0xff) * 2;
+	if (cto_div == 0)
+		cto_div = 1;
+
+	cto_ms = DIV_ROUND_UP_ULL((u64)MSEC_PER_SEC * cto_clks * cto_div,
+				  host->bus_hz);
+
+	/* add a bit spare time */
+	cto_ms += 10;
+
+	/*
+	 * The durations we're working with are fairly short so we have to be
+	 * extra careful about synchronization here.  Specifically in hardware a
+	 * command timeout is _at most_ 5.1 ms, so that means we expect an
+	 * interrupt (either command done or timeout) to come rather quickly
+	 * after the mci_writel.  ...but just in case we have a long interrupt
+	 * latency let's add a bit of paranoia.
+	 *
+	 * In general we'll assume that at least an interrupt will be asserted
+	 * in hardware by the time the cto_timer runs.  ...and if it hasn't
+	 * been asserted in hardware by that time then we'll assume it'll never
+	 * come.
+	 */
+	spin_lock_irqsave(&host->irq_lock, irqflags);
+	if (!test_bit(EVENT_CMD_COMPLETE, &host->pending_events))
+		mod_timer(&host->cto_timer,
+			jiffies + msecs_to_jiffies(cto_ms) + 1);
+	spin_unlock_irqrestore(&host->irq_lock, irqflags);
+}
+
+static void dw_mci_start_command(struct dw_mci *host,
+				 struct mmc_command *cmd, u32 cmd_flags)
+{
+	host->cmd = cmd;
+	dev_vdbg(host->dev,
+		 "start command: ARGR=0x%08x CMDR=0x%08x\n",
+		 cmd->arg, cmd_flags);
+
+	mci_writel(host, CMDARG, cmd->arg);
+	wmb(); /* drain writebuffer */
+	dw_mci_wait_while_busy(host, cmd_flags);
+
+	mci_writel(host, CMD, cmd_flags | SDMMC_CMD_START);
+
+	/* response expected command only */
+	if (cmd_flags & SDMMC_CMD_RESP_EXP)
+		dw_mci_set_cto(host);
+}
+
+static inline void send_stop_abort(struct dw_mci *host, struct mmc_data *data)
+{
+	struct mmc_command *stop = &host->stop_abort;
+
+	dw_mci_start_command(host, stop, host->stop_cmdr);
+}
+
+/* DMA interface functions */
+static void dw_mci_stop_dma(struct dw_mci *host)
+{
+	if (host->using_dma) {
+		host->dma_ops->stop(host);
+		host->dma_ops->cleanup(host);
+	}
+
+	/* Data transfer was stopped by the interrupt handler */
+	set_bit(EVENT_XFER_COMPLETE, &host->pending_events);
+}
+
+static void dw_mci_dma_cleanup(struct dw_mci *host)
+{
+	struct mmc_data *data = host->data;
+
+	if (data && data->host_cookie == COOKIE_MAPPED) {
+		dma_unmap_sg(host->dev,
+			     data->sg,
+			     data->sg_len,
+			     mmc_get_dma_dir(data));
+		data->host_cookie = COOKIE_UNMAPPED;
+	}
+}
+
+static void dw_mci_idmac_reset(struct dw_mci *host)
+{
+	u32 bmod = mci_readl(host, BMOD);
+	/* Software reset of DMA */
+	bmod |= SDMMC_IDMAC_SWRESET;
+	mci_writel(host, BMOD, bmod);
+}
+
+static void dw_mci_idmac_stop_dma(struct dw_mci *host)
+{
+	u32 temp;
+
+	/* Disable and reset the IDMAC interface */
+	temp = mci_readl(host, CTRL);
+	temp &= ~SDMMC_CTRL_USE_IDMAC;
+	temp |= SDMMC_CTRL_DMA_RESET;
+	mci_writel(host, CTRL, temp);
+
+	/* Stop the IDMAC running */
+	temp = mci_readl(host, BMOD);
+	temp &= ~(SDMMC_IDMAC_ENABLE | SDMMC_IDMAC_FB);
+	temp |= SDMMC_IDMAC_SWRESET;
+	mci_writel(host, BMOD, temp);
+}
+
+static void dw_mci_dmac_complete_dma(void *arg)
+{
+	struct dw_mci *host = arg;
+	struct mmc_data *data = host->data;
+
+	dev_vdbg(host->dev, "DMA complete\n");
+
+	if ((host->use_dma == TRANS_MODE_EDMAC) &&
+	    data && (data->flags & MMC_DATA_READ))
+		/* Invalidate cache after read */
+		dma_sync_sg_for_cpu(mmc_dev(host->slot->mmc),
+				    data->sg,
+				    data->sg_len,
+				    DMA_FROM_DEVICE);
+
+	host->dma_ops->cleanup(host);
+
+	/*
+	 * If the card was removed, data will be NULL. No point in trying to
+	 * send the stop command or waiting for NBUSY in this case.
+	 */
+	if (data) {
+		set_bit(EVENT_XFER_COMPLETE, &host->pending_events);
+		tasklet_schedule(&host->tasklet);
+	}
+}
+
+static int dw_mci_idmac_init(struct dw_mci *host)
+{
+	int i;
+
+	if (host->dma_64bit_address == 1) {
+		struct idmac_desc_64addr *p;
+		/* Number of descriptors in the ring buffer */
+		host->ring_size =
+			DESC_RING_BUF_SZ / sizeof(struct idmac_desc_64addr);
+
+		/* Forward link the descriptor list */
+		for (i = 0, p = host->sg_cpu; i < host->ring_size - 1;
+								i++, p++) {
+			p->des6 = (host->sg_dma +
+					(sizeof(struct idmac_desc_64addr) *
+							(i + 1))) & 0xffffffff;
+
+			p->des7 = (u64)(host->sg_dma +
+					(sizeof(struct idmac_desc_64addr) *
+							(i + 1))) >> 32;
+			/* Initialize reserved and buffer size fields to "0" */
+			p->des0 = 0;
+			p->des1 = 0;
+			p->des2 = 0;
+			p->des3 = 0;
+		}
+
+		/* Set the last descriptor as the end-of-ring descriptor */
+		p->des6 = host->sg_dma & 0xffffffff;
+		p->des7 = (u64)host->sg_dma >> 32;
+		p->des0 = IDMAC_DES0_ER;
+
+	} else {
+		struct idmac_desc *p;
+		/* Number of descriptors in the ring buffer */
+		host->ring_size =
+			DESC_RING_BUF_SZ / sizeof(struct idmac_desc);
+
+		/* Forward link the descriptor list */
+		for (i = 0, p = host->sg_cpu;
+		     i < host->ring_size - 1;
+		     i++, p++) {
+			p->des3 = cpu_to_le32(host->sg_dma +
+					(sizeof(struct idmac_desc) * (i + 1)));
+			p->des0 = 0;
+			p->des1 = 0;
+		}
+
+		/* Set the last descriptor as the end-of-ring descriptor */
+		p->des3 = cpu_to_le32(host->sg_dma);
+		p->des0 = cpu_to_le32(IDMAC_DES0_ER);
+	}
+
+	dw_mci_idmac_reset(host);
+
+	if (host->dma_64bit_address == 1) {
+		/* Mask out interrupts - get Tx & Rx complete only */
+		mci_writel(host, IDSTS64, IDMAC_INT_CLR);
+		mci_writel(host, IDINTEN64, SDMMC_IDMAC_INT_NI |
+				SDMMC_IDMAC_INT_RI | SDMMC_IDMAC_INT_TI);
+
+		/* Set the descriptor base address */
+		mci_writel(host, DBADDRL, host->sg_dma & 0xffffffff);
+		mci_writel(host, DBADDRU, (u64)host->sg_dma >> 32);
+
+	} else {
+		/* Mask out interrupts - get Tx & Rx complete only */
+		mci_writel(host, IDSTS, IDMAC_INT_CLR);
+		mci_writel(host, IDINTEN, SDMMC_IDMAC_INT_NI |
+				SDMMC_IDMAC_INT_RI | SDMMC_IDMAC_INT_TI);
+
+		/* Set the descriptor base address */
+		mci_writel(host, DBADDR, host->sg_dma);
+	}
+
+	return 0;
+}
+
+static inline int dw_mci_prepare_desc64(struct dw_mci *host,
+					 struct mmc_data *data,
+					 unsigned int sg_len)
+{
+	unsigned int desc_len;
+	struct idmac_desc_64addr *desc_first, *desc_last, *desc;
+	u32 val;
+	int i;
+
+	desc_first = desc_last = desc = host->sg_cpu;
+
+	for (i = 0; i < sg_len; i++) {
+		unsigned int length = sg_dma_len(&data->sg[i]);
+
+		u64 mem_addr = sg_dma_address(&data->sg[i]);
+
+		for ( ; length ; desc++) {
+			desc_len = (length <= DW_MCI_DESC_DATA_LENGTH) ?
+				   length : DW_MCI_DESC_DATA_LENGTH;
+
+			length -= desc_len;
+
+			/*
+			 * Wait for the former clear OWN bit operation
+			 * of IDMAC to make sure that this descriptor
+			 * isn't still owned by IDMAC as IDMAC's write
+			 * ops and CPU's read ops are asynchronous.
+			 */
+			if (readl_poll_timeout_atomic(&desc->des0, val,
+						!(val & IDMAC_DES0_OWN),
+						10, 100 * USEC_PER_MSEC))
+				goto err_own_bit;
+
+			/*
+			 * Set the OWN bit and disable interrupts
+			 * for this descriptor
+			 */
+			desc->des0 = IDMAC_DES0_OWN | IDMAC_DES0_DIC |
+						IDMAC_DES0_CH;
+
+			/* Buffer length */
+			IDMAC_64ADDR_SET_BUFFER1_SIZE(desc, desc_len);
+
+			/* Physical address to DMA to/from */
+			desc->des4 = mem_addr & 0xffffffff;
+			desc->des5 = mem_addr >> 32;
+
+			/* Update physical address for the next desc */
+			mem_addr += desc_len;
+
+			/* Save pointer to the last descriptor */
+			desc_last = desc;
+		}
+	}
+
+	/* Set first descriptor */
+	desc_first->des0 |= IDMAC_DES0_FD;
+
+	/* Set last descriptor */
+	desc_last->des0 &= ~(IDMAC_DES0_CH | IDMAC_DES0_DIC);
+	desc_last->des0 |= IDMAC_DES0_LD;
+
+	return 0;
+err_own_bit:
+	/* restore the descriptor chain as it's polluted */
+	dev_dbg(host->dev, "descriptor is still owned by IDMAC.\n");
+	memset(host->sg_cpu, 0, DESC_RING_BUF_SZ);
+	dw_mci_idmac_init(host);
+	return -EINVAL;
+}
+
+
+static inline int dw_mci_prepare_desc32(struct dw_mci *host,
+					 struct mmc_data *data,
+					 unsigned int sg_len)
+{
+	unsigned int desc_len;
+	struct idmac_desc *desc_first, *desc_last, *desc;
+	u32 val;
+	int i;
+
+	desc_first = desc_last = desc = host->sg_cpu;
+
+	for (i = 0; i < sg_len; i++) {
+		unsigned int length = sg_dma_len(&data->sg[i]);
+
+		u32 mem_addr = sg_dma_address(&data->sg[i]);
+
+		for ( ; length ; desc++) {
+			desc_len = (length <= DW_MCI_DESC_DATA_LENGTH) ?
+				   length : DW_MCI_DESC_DATA_LENGTH;
+
+			length -= desc_len;
+
+			/*
+			 * Wait for the former clear OWN bit operation
+			 * of IDMAC to make sure that this descriptor
+			 * isn't still owned by IDMAC as IDMAC's write
+			 * ops and CPU's read ops are asynchronous.
+			 */
+			if (readl_poll_timeout_atomic(&desc->des0, val,
+						      IDMAC_OWN_CLR64(val),
+						      10,
+						      100 * USEC_PER_MSEC))
+				goto err_own_bit;
+
+			/*
+			 * Set the OWN bit and disable interrupts
+			 * for this descriptor
+			 */
+			desc->des0 = cpu_to_le32(IDMAC_DES0_OWN |
+						 IDMAC_DES0_DIC |
+						 IDMAC_DES0_CH);
+
+			/* Buffer length */
+			IDMAC_SET_BUFFER1_SIZE(desc, desc_len);
+
+			/* Physical address to DMA to/from */
+			desc->des2 = cpu_to_le32(mem_addr);
+
+			/* Update physical address for the next desc */
+			mem_addr += desc_len;
+
+			/* Save pointer to the last descriptor */
+			desc_last = desc;
+		}
+	}
+
+	/* Set first descriptor */
+	desc_first->des0 |= cpu_to_le32(IDMAC_DES0_FD);
+
+	/* Set last descriptor */
+	desc_last->des0 &= cpu_to_le32(~(IDMAC_DES0_CH |
+				       IDMAC_DES0_DIC));
+	desc_last->des0 |= cpu_to_le32(IDMAC_DES0_LD);
+
+	return 0;
+err_own_bit:
+	/* restore the descriptor chain as it's polluted */
+	dev_dbg(host->dev, "descriptor is still owned by IDMAC.\n");
+	memset(host->sg_cpu, 0, DESC_RING_BUF_SZ);
+	dw_mci_idmac_init(host);
+	return -EINVAL;
+}
+
+static int dw_mci_idmac_start_dma(struct dw_mci *host, unsigned int sg_len)
+{
+	u32 temp;
+	int ret;
+
+	if (host->dma_64bit_address == 1)
+		ret = dw_mci_prepare_desc64(host, host->data, sg_len);
+	else
+		ret = dw_mci_prepare_desc32(host, host->data, sg_len);
+
+	if (ret)
+		goto out;
+
+	/* drain writebuffer */
+	wmb();
+
+	/* Make sure to reset DMA in case we did PIO before this */
+	dw_mci_ctrl_reset(host, SDMMC_CTRL_DMA_RESET);
+	dw_mci_idmac_reset(host);
+
+	/* Select IDMAC interface */
+	temp = mci_readl(host, CTRL);
+	temp |= SDMMC_CTRL_USE_IDMAC;
+	mci_writel(host, CTRL, temp);
+
+	/* drain writebuffer */
+	wmb();
+
+	/* Enable the IDMAC */
+	temp = mci_readl(host, BMOD);
+	temp |= SDMMC_IDMAC_ENABLE | SDMMC_IDMAC_FB;
+	mci_writel(host, BMOD, temp);
+
+	/* Start it running */
+	mci_writel(host, PLDMND, 1);
+
+out:
+	return ret;
+}
+
+static const struct dw_mci_dma_ops dw_mci_idmac_ops = {
+	.init = dw_mci_idmac_init,
+	.start = dw_mci_idmac_start_dma,
+	.stop = dw_mci_idmac_stop_dma,
+	.complete = dw_mci_dmac_complete_dma,
+	.cleanup = dw_mci_dma_cleanup,
+};
+
+static void dw_mci_edmac_stop_dma(struct dw_mci *host)
+{
+	dmaengine_terminate_async(host->dms->ch);
+}
+
+static int dw_mci_edmac_start_dma(struct dw_mci *host,
+					    unsigned int sg_len)
+{
+	struct dma_slave_config cfg;
+	struct dma_async_tx_descriptor *desc = NULL;
+	struct scatterlist *sgl = host->data->sg;
+	static const u32 mszs[] = {1, 4, 8, 16, 32, 64, 128, 256};
+	u32 sg_elems = host->data->sg_len;
+	u32 fifoth_val;
+	u32 fifo_offset = host->fifo_reg - host->regs;
+	int ret = 0;
+
+	/* Set external dma config: burst size, burst width */
+	memset(&cfg, 0, sizeof(cfg));
+	cfg.dst_addr = host->phy_regs + fifo_offset;
+	cfg.src_addr = cfg.dst_addr;
+	cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+
+	/* Match burst msize with external dma config */
+	fifoth_val = mci_readl(host, FIFOTH);
+	cfg.dst_maxburst = mszs[(fifoth_val >> 28) & 0x7];
+	cfg.src_maxburst = cfg.dst_maxburst;
+
+	if (host->data->flags & MMC_DATA_WRITE)
+		cfg.direction = DMA_MEM_TO_DEV;
+	else
+		cfg.direction = DMA_DEV_TO_MEM;
+
+	ret = dmaengine_slave_config(host->dms->ch, &cfg);
+	if (ret) {
+		dev_err(host->dev, "Failed to config edmac.\n");
+		return -EBUSY;
+	}
+
+	desc = dmaengine_prep_slave_sg(host->dms->ch, sgl,
+				       sg_len, cfg.direction,
+				       DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc) {
+		dev_err(host->dev, "Can't prepare slave sg.\n");
+		return -EBUSY;
+	}
+
+	/* Set dw_mci_dmac_complete_dma as callback */
+	desc->callback = dw_mci_dmac_complete_dma;
+	desc->callback_param = (void *)host;
+	dmaengine_submit(desc);
+
+	/* Flush cache before write */
+	if (host->data->flags & MMC_DATA_WRITE)
+		dma_sync_sg_for_device(mmc_dev(host->slot->mmc), sgl,
+				       sg_elems, DMA_TO_DEVICE);
+
+	dma_async_issue_pending(host->dms->ch);
+
+	return 0;
+}
+
+static int dw_mci_edmac_init(struct dw_mci *host)
+{
+	/* Request external dma channel */
+	host->dms = kzalloc(sizeof(struct dw_mci_dma_slave), GFP_KERNEL);
+	if (!host->dms)
+		return -ENOMEM;
+
+	host->dms->ch = dma_request_chan(host->dev, "rx-tx");
+	if (IS_ERR(host->dms->ch)) {
+		int ret = PTR_ERR(host->dms->ch);
+
+		dev_err(host->dev, "Failed to get external DMA channel.\n");
+		kfree(host->dms);
+		host->dms = NULL;
+		return ret;
+	}
+
+	return 0;
+}
+
+static void dw_mci_edmac_exit(struct dw_mci *host)
+{
+	if (host->dms) {
+		if (host->dms->ch) {
+			dma_release_channel(host->dms->ch);
+			host->dms->ch = NULL;
+		}
+		kfree(host->dms);
+		host->dms = NULL;
+	}
+}
+
+static const struct dw_mci_dma_ops dw_mci_edmac_ops = {
+	.init = dw_mci_edmac_init,
+	.exit = dw_mci_edmac_exit,
+	.start = dw_mci_edmac_start_dma,
+	.stop = dw_mci_edmac_stop_dma,
+	.complete = dw_mci_dmac_complete_dma,
+	.cleanup = dw_mci_dma_cleanup,
+};
+
+static int dw_mci_pre_dma_transfer(struct dw_mci *host,
+				   struct mmc_data *data,
+				   int cookie)
+{
+	struct scatterlist *sg;
+	unsigned int i, sg_len;
+
+	if (data->host_cookie == COOKIE_PRE_MAPPED)
+		return data->sg_len;
+
+	/*
+	 * We don't do DMA on "complex" transfers, i.e. with
+	 * non-word-aligned buffers or lengths. Also, we don't bother
+	 * with all the DMA setup overhead for short transfers.
+	 */
+	if (data->blocks * data->blksz < DW_MCI_DMA_THRESHOLD)
+		return -EINVAL;
+
+	if (data->blksz & 3)
+		return -EINVAL;
+
+	for_each_sg(data->sg, sg, data->sg_len, i) {
+		if (sg->offset & 3 || sg->length & 3)
+			return -EINVAL;
+	}
+
+	sg_len = dma_map_sg(host->dev,
+			    data->sg,
+			    data->sg_len,
+			    mmc_get_dma_dir(data));
+	if (sg_len == 0)
+		return -EINVAL;
+
+	data->host_cookie = cookie;
+
+	return sg_len;
+}
+
+static void dw_mci_pre_req(struct mmc_host *mmc,
+			   struct mmc_request *mrq)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+
+	if (!slot->host->use_dma || !data)
+		return;
+
+	/* This data might be unmapped at this time */
+	data->host_cookie = COOKIE_UNMAPPED;
+
+	if (dw_mci_pre_dma_transfer(slot->host, mrq->data,
+				COOKIE_PRE_MAPPED) < 0)
+		data->host_cookie = COOKIE_UNMAPPED;
+}
+
+static void dw_mci_post_req(struct mmc_host *mmc,
+			    struct mmc_request *mrq,
+			    int err)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+
+	if (!slot->host->use_dma || !data)
+		return;
+
+	if (data->host_cookie != COOKIE_UNMAPPED)
+		dma_unmap_sg(slot->host->dev,
+			     data->sg,
+			     data->sg_len,
+			     mmc_get_dma_dir(data));
+	data->host_cookie = COOKIE_UNMAPPED;
+}
+
+static int dw_mci_get_cd(struct mmc_host *mmc)
+{
+	int present;
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct dw_mci *host = slot->host;
+	int gpio_cd = mmc_gpio_get_cd(mmc);
+
+	/* Use platform get_cd function, else try onboard card detect */
+	if (((mmc->caps & MMC_CAP_NEEDS_POLL)
+				|| !mmc_card_is_removable(mmc))) {
+		present = 1;
+
+		if (!test_bit(DW_MMC_CARD_PRESENT, &slot->flags)) {
+			if (mmc->caps & MMC_CAP_NEEDS_POLL) {
+				dev_info(&mmc->class_dev,
+					"card is polling.\n");
+			} else {
+				dev_info(&mmc->class_dev,
+					"card is non-removable.\n");
+			}
+			set_bit(DW_MMC_CARD_PRESENT, &slot->flags);
+		}
+
+		return present;
+	} else if (gpio_cd >= 0)
+		present = gpio_cd;
+	else
+		present = (mci_readl(slot->host, CDETECT) & (1 << slot->id))
+			== 0 ? 1 : 0;
+
+	spin_lock_bh(&host->lock);
+	if (present && !test_and_set_bit(DW_MMC_CARD_PRESENT, &slot->flags))
+		dev_dbg(&mmc->class_dev, "card is present\n");
+	else if (!present &&
+			!test_and_clear_bit(DW_MMC_CARD_PRESENT, &slot->flags))
+		dev_dbg(&mmc->class_dev, "card is not present\n");
+	spin_unlock_bh(&host->lock);
+
+	return present;
+}
+
+static void dw_mci_adjust_fifoth(struct dw_mci *host, struct mmc_data *data)
+{
+	unsigned int blksz = data->blksz;
+	static const u32 mszs[] = {1, 4, 8, 16, 32, 64, 128, 256};
+	u32 fifo_width = 1 << host->data_shift;
+	u32 blksz_depth = blksz / fifo_width, fifoth_val;
+	u32 msize = 0, rx_wmark = 1, tx_wmark, tx_wmark_invers;
+	int idx = ARRAY_SIZE(mszs) - 1;
+
+	/* pio should ship this scenario */
+	if (!host->use_dma)
+		return;
+
+	tx_wmark = (host->fifo_depth) / 2;
+	tx_wmark_invers = host->fifo_depth - tx_wmark;
+
+	/*
+	 * MSIZE is '1',
+	 * if blksz is not a multiple of the FIFO width
+	 */
+	if (blksz % fifo_width)
+		goto done;
+
+	do {
+		if (!((blksz_depth % mszs[idx]) ||
+		     (tx_wmark_invers % mszs[idx]))) {
+			msize = idx;
+			rx_wmark = mszs[idx] - 1;
+			break;
+		}
+	} while (--idx > 0);
+	/*
+	 * If idx is '0', it won't be tried
+	 * Thus, initial values are uesed
+	 */
+done:
+	fifoth_val = SDMMC_SET_FIFOTH(msize, rx_wmark, tx_wmark);
+	mci_writel(host, FIFOTH, fifoth_val);
+}
+
+static void dw_mci_ctrl_thld(struct dw_mci *host, struct mmc_data *data)
+{
+	unsigned int blksz = data->blksz;
+	u32 blksz_depth, fifo_depth;
+	u16 thld_size;
+	u8 enable;
+
+	/*
+	 * CDTHRCTL doesn't exist prior to 240A (in fact that register offset is
+	 * in the FIFO region, so we really shouldn't access it).
+	 */
+	if (host->verid < DW_MMC_240A ||
+		(host->verid < DW_MMC_280A && data->flags & MMC_DATA_WRITE))
+		return;
+
+	/*
+	 * Card write Threshold is introduced since 2.80a
+	 * It's used when HS400 mode is enabled.
+	 */
+	if (data->flags & MMC_DATA_WRITE &&
+		host->timing != MMC_TIMING_MMC_HS400)
+		goto disable;
+
+	if (data->flags & MMC_DATA_WRITE)
+		enable = SDMMC_CARD_WR_THR_EN;
+	else
+		enable = SDMMC_CARD_RD_THR_EN;
+
+	if (host->timing != MMC_TIMING_MMC_HS200 &&
+	    host->timing != MMC_TIMING_UHS_SDR104 &&
+	    host->timing != MMC_TIMING_MMC_HS400)
+		goto disable;
+
+	blksz_depth = blksz / (1 << host->data_shift);
+	fifo_depth = host->fifo_depth;
+
+	if (blksz_depth > fifo_depth)
+		goto disable;
+
+	/*
+	 * If (blksz_depth) >= (fifo_depth >> 1), should be 'thld_size <= blksz'
+	 * If (blksz_depth) <  (fifo_depth >> 1), should be thld_size = blksz
+	 * Currently just choose blksz.
+	 */
+	thld_size = blksz;
+	mci_writel(host, CDTHRCTL, SDMMC_SET_THLD(thld_size, enable));
+	return;
+
+disable:
+	mci_writel(host, CDTHRCTL, 0);
+}
+
+static int dw_mci_submit_data_dma(struct dw_mci *host, struct mmc_data *data)
+{
+	unsigned long irqflags;
+	int sg_len;
+	u32 temp;
+
+	host->using_dma = 0;
+
+	/* If we don't have a channel, we can't do DMA */
+	if (!host->use_dma)
+		return -ENODEV;
+
+	sg_len = dw_mci_pre_dma_transfer(host, data, COOKIE_MAPPED);
+	if (sg_len < 0) {
+		host->dma_ops->stop(host);
+		return sg_len;
+	}
+
+	host->using_dma = 1;
+
+	if (host->use_dma == TRANS_MODE_IDMAC)
+		dev_vdbg(host->dev,
+			 "sd sg_cpu: %#lx sg_dma: %#lx sg_len: %d\n",
+			 (unsigned long)host->sg_cpu,
+			 (unsigned long)host->sg_dma,
+			 sg_len);
+
+	/*
+	 * Decide the MSIZE and RX/TX Watermark.
+	 * If current block size is same with previous size,
+	 * no need to update fifoth.
+	 */
+	if (host->prev_blksz != data->blksz)
+		dw_mci_adjust_fifoth(host, data);
+
+	/* Enable the DMA interface */
+	temp = mci_readl(host, CTRL);
+	temp |= SDMMC_CTRL_DMA_ENABLE;
+	mci_writel(host, CTRL, temp);
+
+	/* Disable RX/TX IRQs, let DMA handle it */
+	spin_lock_irqsave(&host->irq_lock, irqflags);
+	temp = mci_readl(host, INTMASK);
+	temp  &= ~(SDMMC_INT_RXDR | SDMMC_INT_TXDR);
+	mci_writel(host, INTMASK, temp);
+	spin_unlock_irqrestore(&host->irq_lock, irqflags);
+
+	if (host->dma_ops->start(host, sg_len)) {
+		host->dma_ops->stop(host);
+		/* We can't do DMA, try PIO for this one */
+		dev_dbg(host->dev,
+			"%s: fall back to PIO mode for current transfer\n",
+			__func__);
+		return -ENODEV;
+	}
+
+	return 0;
+}
+
+static void dw_mci_submit_data(struct dw_mci *host, struct mmc_data *data)
+{
+	unsigned long irqflags;
+	int flags = SG_MITER_ATOMIC;
+	u32 temp;
+
+	data->error = -EINPROGRESS;
+
+	WARN_ON(host->data);
+	host->sg = NULL;
+	host->data = data;
+
+	if (data->flags & MMC_DATA_READ)
+		host->dir_status = DW_MCI_RECV_STATUS;
+	else
+		host->dir_status = DW_MCI_SEND_STATUS;
+
+	dw_mci_ctrl_thld(host, data);
+
+	if (dw_mci_submit_data_dma(host, data)) {
+		if (host->data->flags & MMC_DATA_READ)
+			flags |= SG_MITER_TO_SG;
+		else
+			flags |= SG_MITER_FROM_SG;
+
+		sg_miter_start(&host->sg_miter, data->sg, data->sg_len, flags);
+		host->sg = data->sg;
+		host->part_buf_start = 0;
+		host->part_buf_count = 0;
+
+		mci_writel(host, RINTSTS, SDMMC_INT_TXDR | SDMMC_INT_RXDR);
+
+		spin_lock_irqsave(&host->irq_lock, irqflags);
+		temp = mci_readl(host, INTMASK);
+		temp |= SDMMC_INT_TXDR | SDMMC_INT_RXDR;
+		mci_writel(host, INTMASK, temp);
+		spin_unlock_irqrestore(&host->irq_lock, irqflags);
+
+		temp = mci_readl(host, CTRL);
+		temp &= ~SDMMC_CTRL_DMA_ENABLE;
+		mci_writel(host, CTRL, temp);
+
+		/*
+		 * Use the initial fifoth_val for PIO mode. If wm_algined
+		 * is set, we set watermark same as data size.
+		 * If next issued data may be transfered by DMA mode,
+		 * prev_blksz should be invalidated.
+		 */
+		if (host->wm_aligned)
+			dw_mci_adjust_fifoth(host, data);
+		else
+			mci_writel(host, FIFOTH, host->fifoth_val);
+		host->prev_blksz = 0;
+	} else {
+		/*
+		 * Keep the current block size.
+		 * It will be used to decide whether to update
+		 * fifoth register next time.
+		 */
+		host->prev_blksz = data->blksz;
+	}
+}
+
+static void dw_mci_setup_bus(struct dw_mci_slot *slot, bool force_clkinit)
+{
+	struct dw_mci *host = slot->host;
+	unsigned int clock = slot->clock;
+	u32 div;
+	u32 clk_en_a;
+	u32 sdmmc_cmd_bits = SDMMC_CMD_UPD_CLK | SDMMC_CMD_PRV_DAT_WAIT;
+
+	/* We must continue to set bit 28 in CMD until the change is complete */
+	if (host->state == STATE_WAITING_CMD11_DONE)
+		sdmmc_cmd_bits |= SDMMC_CMD_VOLT_SWITCH;
+
+	slot->mmc->actual_clock = 0;
+
+	if (!clock) {
+		mci_writel(host, CLKENA, 0);
+		mci_send_cmd(slot, sdmmc_cmd_bits, 0);
+	} else if (clock != host->current_speed || force_clkinit) {
+		div = host->bus_hz / clock;
+		if (host->bus_hz % clock && host->bus_hz > clock)
+			/*
+			 * move the + 1 after the divide to prevent
+			 * over-clocking the card.
+			 */
+			div += 1;
+
+		div = (host->bus_hz != clock) ? DIV_ROUND_UP(div, 2) : 0;
+
+		if ((clock != slot->__clk_old &&
+			!test_bit(DW_MMC_CARD_NEEDS_POLL, &slot->flags)) ||
+			force_clkinit) {
+			/* Silent the verbose log if calling from PM context */
+			if (!force_clkinit)
+				dev_info(&slot->mmc->class_dev,
+					 "Bus speed (slot %d) = %dHz (slot req %dHz, actual %dHZ div = %d)\n",
+					 slot->id, host->bus_hz, clock,
+					 div ? ((host->bus_hz / div) >> 1) :
+					 host->bus_hz, div);
+
+			/*
+			 * If card is polling, display the message only
+			 * one time at boot time.
+			 */
+			if (slot->mmc->caps & MMC_CAP_NEEDS_POLL &&
+					slot->mmc->f_min == clock)
+				set_bit(DW_MMC_CARD_NEEDS_POLL, &slot->flags);
+		}
+
+		/* disable clock */
+		mci_writel(host, CLKENA, 0);
+		mci_writel(host, CLKSRC, 0);
+
+		/* inform CIU */
+		mci_send_cmd(slot, sdmmc_cmd_bits, 0);
+
+		/* set clock to desired speed */
+		mci_writel(host, CLKDIV, div);
+
+		/* inform CIU */
+		mci_send_cmd(slot, sdmmc_cmd_bits, 0);
+
+		/* enable clock; only low power if no SDIO */
+		clk_en_a = SDMMC_CLKEN_ENABLE << slot->id;
+		if (!test_bit(DW_MMC_CARD_NO_LOW_PWR, &slot->flags))
+			clk_en_a |= SDMMC_CLKEN_LOW_PWR << slot->id;
+		mci_writel(host, CLKENA, clk_en_a);
+
+		/* inform CIU */
+		mci_send_cmd(slot, sdmmc_cmd_bits, 0);
+
+		/* keep the last clock value that was requested from core */
+		slot->__clk_old = clock;
+		slot->mmc->actual_clock = div ? ((host->bus_hz / div) >> 1) :
+					  host->bus_hz;
+	}
+
+	host->current_speed = clock;
+
+	/* Set the current slot bus width */
+	mci_writel(host, CTYPE, (slot->ctype << slot->id));
+}
+
+static void __dw_mci_start_request(struct dw_mci *host,
+				   struct dw_mci_slot *slot,
+				   struct mmc_command *cmd)
+{
+	struct mmc_request *mrq;
+	struct mmc_data	*data;
+	u32 cmdflags;
+
+	mrq = slot->mrq;
+
+	host->mrq = mrq;
+
+	host->pending_events = 0;
+	host->completed_events = 0;
+	host->cmd_status = 0;
+	host->data_status = 0;
+	host->dir_status = 0;
+
+	data = cmd->data;
+	if (data) {
+		mci_writel(host, TMOUT, 0xFFFFFFFF);
+		mci_writel(host, BYTCNT, data->blksz*data->blocks);
+		mci_writel(host, BLKSIZ, data->blksz);
+	}
+
+	cmdflags = dw_mci_prepare_command(slot->mmc, cmd);
+
+	/* this is the first command, send the initialization clock */
+	if (test_and_clear_bit(DW_MMC_CARD_NEED_INIT, &slot->flags))
+		cmdflags |= SDMMC_CMD_INIT;
+
+	if (data) {
+		dw_mci_submit_data(host, data);
+		wmb(); /* drain writebuffer */
+	}
+
+	dw_mci_start_command(host, cmd, cmdflags);
+
+	if (cmd->opcode == SD_SWITCH_VOLTAGE) {
+		unsigned long irqflags;
+
+		/*
+		 * Databook says to fail after 2ms w/ no response, but evidence
+		 * shows that sometimes the cmd11 interrupt takes over 130ms.
+		 * We'll set to 500ms, plus an extra jiffy just in case jiffies
+		 * is just about to roll over.
+		 *
+		 * We do this whole thing under spinlock and only if the
+		 * command hasn't already completed (indicating the the irq
+		 * already ran so we don't want the timeout).
+		 */
+		spin_lock_irqsave(&host->irq_lock, irqflags);
+		if (!test_bit(EVENT_CMD_COMPLETE, &host->pending_events))
+			mod_timer(&host->cmd11_timer,
+				jiffies + msecs_to_jiffies(500) + 1);
+		spin_unlock_irqrestore(&host->irq_lock, irqflags);
+	}
+
+	host->stop_cmdr = dw_mci_prep_stop_abort(host, cmd);
+}
+
+static void dw_mci_start_request(struct dw_mci *host,
+				 struct dw_mci_slot *slot)
+{
+	struct mmc_request *mrq = slot->mrq;
+	struct mmc_command *cmd;
+
+	cmd = mrq->sbc ? mrq->sbc : mrq->cmd;
+	__dw_mci_start_request(host, slot, cmd);
+}
+
+/* must be called with host->lock held */
+static void dw_mci_queue_request(struct dw_mci *host, struct dw_mci_slot *slot,
+				 struct mmc_request *mrq)
+{
+	dev_vdbg(&slot->mmc->class_dev, "queue request: state=%d\n",
+		 host->state);
+
+	slot->mrq = mrq;
+
+	if (host->state == STATE_WAITING_CMD11_DONE) {
+		dev_warn(&slot->mmc->class_dev,
+			 "Voltage change didn't complete\n");
+		/*
+		 * this case isn't expected to happen, so we can
+		 * either crash here or just try to continue on
+		 * in the closest possible state
+		 */
+		host->state = STATE_IDLE;
+	}
+
+	if (host->state == STATE_IDLE) {
+		host->state = STATE_SENDING_CMD;
+		dw_mci_start_request(host, slot);
+	} else {
+		list_add_tail(&slot->queue_node, &host->queue);
+	}
+}
+
+static void dw_mci_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct dw_mci *host = slot->host;
+
+	WARN_ON(slot->mrq);
+
+	/*
+	 * The check for card presence and queueing of the request must be
+	 * atomic, otherwise the card could be removed in between and the
+	 * request wouldn't fail until another card was inserted.
+	 */
+
+	if (!dw_mci_get_cd(mmc)) {
+		mrq->cmd->error = -ENOMEDIUM;
+		mmc_request_done(mmc, mrq);
+		return;
+	}
+
+	spin_lock_bh(&host->lock);
+
+	dw_mci_queue_request(host, slot, mrq);
+
+	spin_unlock_bh(&host->lock);
+}
+
+static void dw_mci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	const struct dw_mci_drv_data *drv_data = slot->host->drv_data;
+	u32 regs;
+	int ret;
+
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_4:
+		slot->ctype = SDMMC_CTYPE_4BIT;
+		break;
+	case MMC_BUS_WIDTH_8:
+		slot->ctype = SDMMC_CTYPE_8BIT;
+		break;
+	default:
+		/* set default 1 bit mode */
+		slot->ctype = SDMMC_CTYPE_1BIT;
+	}
+
+	regs = mci_readl(slot->host, UHS_REG);
+
+	/* DDR mode set */
+	if (ios->timing == MMC_TIMING_MMC_DDR52 ||
+	    ios->timing == MMC_TIMING_UHS_DDR50 ||
+	    ios->timing == MMC_TIMING_MMC_HS400)
+		regs |= ((0x1 << slot->id) << 16);
+	else
+		regs &= ~((0x1 << slot->id) << 16);
+
+	mci_writel(slot->host, UHS_REG, regs);
+	slot->host->timing = ios->timing;
+
+	/*
+	 * Use mirror of ios->clock to prevent race with mmc
+	 * core ios update when finding the minimum.
+	 */
+	slot->clock = ios->clock;
+
+	if (drv_data && drv_data->set_ios)
+		drv_data->set_ios(slot->host, ios);
+
+	switch (ios->power_mode) {
+	case MMC_POWER_UP:
+		if (!IS_ERR(mmc->supply.vmmc)) {
+			ret = mmc_regulator_set_ocr(mmc, mmc->supply.vmmc,
+					ios->vdd);
+			if (ret) {
+				dev_err(slot->host->dev,
+					"failed to enable vmmc regulator\n");
+				/*return, if failed turn on vmmc*/
+				return;
+			}
+		}
+		set_bit(DW_MMC_CARD_NEED_INIT, &slot->flags);
+		regs = mci_readl(slot->host, PWREN);
+		regs |= (1 << slot->id);
+		mci_writel(slot->host, PWREN, regs);
+		break;
+	case MMC_POWER_ON:
+		if (!slot->host->vqmmc_enabled) {
+			if (!IS_ERR(mmc->supply.vqmmc)) {
+				ret = regulator_enable(mmc->supply.vqmmc);
+				if (ret < 0)
+					dev_err(slot->host->dev,
+						"failed to enable vqmmc\n");
+				else
+					slot->host->vqmmc_enabled = true;
+
+			} else {
+				/* Keep track so we don't reset again */
+				slot->host->vqmmc_enabled = true;
+			}
+
+			/* Reset our state machine after powering on */
+			dw_mci_ctrl_reset(slot->host,
+					  SDMMC_CTRL_ALL_RESET_FLAGS);
+		}
+
+		/* Adjust clock / bus width after power is up */
+		dw_mci_setup_bus(slot, false);
+
+		break;
+	case MMC_POWER_OFF:
+		/* Turn clock off before power goes down */
+		dw_mci_setup_bus(slot, false);
+
+		if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+
+		if (!IS_ERR(mmc->supply.vqmmc) && slot->host->vqmmc_enabled)
+			regulator_disable(mmc->supply.vqmmc);
+		slot->host->vqmmc_enabled = false;
+
+		regs = mci_readl(slot->host, PWREN);
+		regs &= ~(1 << slot->id);
+		mci_writel(slot->host, PWREN, regs);
+		break;
+	default:
+		break;
+	}
+
+	if (slot->host->state == STATE_WAITING_CMD11_DONE && ios->clock != 0)
+		slot->host->state = STATE_IDLE;
+}
+
+static int dw_mci_card_busy(struct mmc_host *mmc)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	u32 status;
+
+	/*
+	 * Check the busy bit which is low when DAT[3:0]
+	 * (the data lines) are 0000
+	 */
+	status = mci_readl(slot->host, STATUS);
+
+	return !!(status & SDMMC_STATUS_BUSY);
+}
+
+static int dw_mci_switch_voltage(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct dw_mci *host = slot->host;
+	const struct dw_mci_drv_data *drv_data = host->drv_data;
+	u32 uhs;
+	u32 v18 = SDMMC_UHS_18V << slot->id;
+	int ret;
+
+	if (drv_data && drv_data->switch_voltage)
+		return drv_data->switch_voltage(mmc, ios);
+
+	/*
+	 * Program the voltage.  Note that some instances of dw_mmc may use
+	 * the UHS_REG for this.  For other instances (like exynos) the UHS_REG
+	 * does no harm but you need to set the regulator directly.  Try both.
+	 */
+	uhs = mci_readl(host, UHS_REG);
+	if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_330)
+		uhs &= ~v18;
+	else
+		uhs |= v18;
+
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+		ret = mmc_regulator_set_vqmmc(mmc, ios);
+		if (ret < 0) {
+			dev_dbg(&mmc->class_dev,
+					 "Regulator set error %d - %s V\n",
+					 ret, uhs & v18 ? "1.8" : "3.3");
+			return ret;
+		}
+	}
+	mci_writel(host, UHS_REG, uhs);
+
+	return 0;
+}
+
+static int dw_mci_get_ro(struct mmc_host *mmc)
+{
+	int read_only;
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	int gpio_ro = mmc_gpio_get_ro(mmc);
+
+	/* Use platform get_ro function, else try on board write protect */
+	if (gpio_ro >= 0)
+		read_only = gpio_ro;
+	else
+		read_only =
+			mci_readl(slot->host, WRTPRT) & (1 << slot->id) ? 1 : 0;
+
+	dev_dbg(&mmc->class_dev, "card is %s\n",
+		read_only ? "read-only" : "read-write");
+
+	return read_only;
+}
+
+static void dw_mci_hw_reset(struct mmc_host *mmc)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct dw_mci *host = slot->host;
+	int reset;
+
+	if (host->use_dma == TRANS_MODE_IDMAC)
+		dw_mci_idmac_reset(host);
+
+	if (!dw_mci_ctrl_reset(host, SDMMC_CTRL_DMA_RESET |
+				     SDMMC_CTRL_FIFO_RESET))
+		return;
+
+	/*
+	 * According to eMMC spec, card reset procedure:
+	 * tRstW >= 1us:   RST_n pulse width
+	 * tRSCA >= 200us: RST_n to Command time
+	 * tRSTH >= 1us:   RST_n high period
+	 */
+	reset = mci_readl(host, RST_N);
+	reset &= ~(SDMMC_RST_HWACTIVE << slot->id);
+	mci_writel(host, RST_N, reset);
+	usleep_range(1, 2);
+	reset |= SDMMC_RST_HWACTIVE << slot->id;
+	mci_writel(host, RST_N, reset);
+	usleep_range(200, 300);
+}
+
+static void dw_mci_init_card(struct mmc_host *mmc, struct mmc_card *card)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct dw_mci *host = slot->host;
+
+	/*
+	 * Low power mode will stop the card clock when idle.  According to the
+	 * description of the CLKENA register we should disable low power mode
+	 * for SDIO cards if we need SDIO interrupts to work.
+	 */
+	if (mmc->caps & MMC_CAP_SDIO_IRQ) {
+		const u32 clken_low_pwr = SDMMC_CLKEN_LOW_PWR << slot->id;
+		u32 clk_en_a_old;
+		u32 clk_en_a;
+
+		clk_en_a_old = mci_readl(host, CLKENA);
+
+		if (card->type == MMC_TYPE_SDIO ||
+		    card->type == MMC_TYPE_SD_COMBO) {
+			set_bit(DW_MMC_CARD_NO_LOW_PWR, &slot->flags);
+			clk_en_a = clk_en_a_old & ~clken_low_pwr;
+		} else {
+			clear_bit(DW_MMC_CARD_NO_LOW_PWR, &slot->flags);
+			clk_en_a = clk_en_a_old | clken_low_pwr;
+		}
+
+		if (clk_en_a != clk_en_a_old) {
+			mci_writel(host, CLKENA, clk_en_a);
+			mci_send_cmd(slot, SDMMC_CMD_UPD_CLK |
+				     SDMMC_CMD_PRV_DAT_WAIT, 0);
+		}
+	}
+}
+
+static void __dw_mci_enable_sdio_irq(struct dw_mci_slot *slot, int enb)
+{
+	struct dw_mci *host = slot->host;
+	unsigned long irqflags;
+	u32 int_mask;
+
+	spin_lock_irqsave(&host->irq_lock, irqflags);
+
+	/* Enable/disable Slot Specific SDIO interrupt */
+	int_mask = mci_readl(host, INTMASK);
+	if (enb)
+		int_mask |= SDMMC_INT_SDIO(slot->sdio_id);
+	else
+		int_mask &= ~SDMMC_INT_SDIO(slot->sdio_id);
+	mci_writel(host, INTMASK, int_mask);
+
+	spin_unlock_irqrestore(&host->irq_lock, irqflags);
+}
+
+static void dw_mci_enable_sdio_irq(struct mmc_host *mmc, int enb)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct dw_mci *host = slot->host;
+
+	__dw_mci_enable_sdio_irq(slot, enb);
+
+	/* Avoid runtime suspending the device when SDIO IRQ is enabled */
+	if (enb)
+		pm_runtime_get_noresume(host->dev);
+	else
+		pm_runtime_put_noidle(host->dev);
+}
+
+static void dw_mci_ack_sdio_irq(struct mmc_host *mmc)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+
+	__dw_mci_enable_sdio_irq(slot, 1);
+}
+
+static int dw_mci_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct dw_mci *host = slot->host;
+	const struct dw_mci_drv_data *drv_data = host->drv_data;
+	int err = -EINVAL;
+
+	if (drv_data && drv_data->execute_tuning)
+		err = drv_data->execute_tuning(slot, opcode);
+	return err;
+}
+
+static int dw_mci_prepare_hs400_tuning(struct mmc_host *mmc,
+				       struct mmc_ios *ios)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct dw_mci *host = slot->host;
+	const struct dw_mci_drv_data *drv_data = host->drv_data;
+
+	if (drv_data && drv_data->prepare_hs400_tuning)
+		return drv_data->prepare_hs400_tuning(host, ios);
+
+	return 0;
+}
+
+static bool dw_mci_reset(struct dw_mci *host)
+{
+	u32 flags = SDMMC_CTRL_RESET | SDMMC_CTRL_FIFO_RESET;
+	bool ret = false;
+	u32 status = 0;
+
+	/*
+	 * Resetting generates a block interrupt, hence setting
+	 * the scatter-gather pointer to NULL.
+	 */
+	if (host->sg) {
+		sg_miter_stop(&host->sg_miter);
+		host->sg = NULL;
+	}
+
+	if (host->use_dma)
+		flags |= SDMMC_CTRL_DMA_RESET;
+
+	if (dw_mci_ctrl_reset(host, flags)) {
+		/*
+		 * In all cases we clear the RAWINTS
+		 * register to clear any interrupts.
+		 */
+		mci_writel(host, RINTSTS, 0xFFFFFFFF);
+
+		if (!host->use_dma) {
+			ret = true;
+			goto ciu_out;
+		}
+
+		/* Wait for dma_req to be cleared */
+		if (readl_poll_timeout_atomic(host->regs + SDMMC_STATUS,
+					      status,
+					      !(status & SDMMC_STATUS_DMA_REQ),
+					      1, 500 * USEC_PER_MSEC)) {
+			dev_err(host->dev,
+				"%s: Timeout waiting for dma_req to be cleared\n",
+				__func__);
+			goto ciu_out;
+		}
+
+		/* when using DMA next we reset the fifo again */
+		if (!dw_mci_ctrl_reset(host, SDMMC_CTRL_FIFO_RESET))
+			goto ciu_out;
+	} else {
+		/* if the controller reset bit did clear, then set clock regs */
+		if (!(mci_readl(host, CTRL) & SDMMC_CTRL_RESET)) {
+			dev_err(host->dev,
+				"%s: fifo/dma reset bits didn't clear but ciu was reset, doing clock update\n",
+				__func__);
+			goto ciu_out;
+		}
+	}
+
+	if (host->use_dma == TRANS_MODE_IDMAC)
+		/* It is also required that we reinit idmac */
+		dw_mci_idmac_init(host);
+
+	ret = true;
+
+ciu_out:
+	/* After a CTRL reset we need to have CIU set clock registers  */
+	mci_send_cmd(host->slot, SDMMC_CMD_UPD_CLK, 0);
+
+	return ret;
+}
+
+static const struct mmc_host_ops dw_mci_ops = {
+	.request		= dw_mci_request,
+	.pre_req		= dw_mci_pre_req,
+	.post_req		= dw_mci_post_req,
+	.set_ios		= dw_mci_set_ios,
+	.get_ro			= dw_mci_get_ro,
+	.get_cd			= dw_mci_get_cd,
+	.hw_reset               = dw_mci_hw_reset,
+	.enable_sdio_irq	= dw_mci_enable_sdio_irq,
+	.ack_sdio_irq		= dw_mci_ack_sdio_irq,
+	.execute_tuning		= dw_mci_execute_tuning,
+	.card_busy		= dw_mci_card_busy,
+	.start_signal_voltage_switch = dw_mci_switch_voltage,
+	.init_card		= dw_mci_init_card,
+	.prepare_hs400_tuning	= dw_mci_prepare_hs400_tuning,
+};
+
+static void dw_mci_request_end(struct dw_mci *host, struct mmc_request *mrq)
+	__releases(&host->lock)
+	__acquires(&host->lock)
+{
+	struct dw_mci_slot *slot;
+	struct mmc_host	*prev_mmc = host->slot->mmc;
+
+	WARN_ON(host->cmd || host->data);
+
+	host->slot->mrq = NULL;
+	host->mrq = NULL;
+	if (!list_empty(&host->queue)) {
+		slot = list_entry(host->queue.next,
+				  struct dw_mci_slot, queue_node);
+		list_del(&slot->queue_node);
+		dev_vdbg(host->dev, "list not empty: %s is next\n",
+			 mmc_hostname(slot->mmc));
+		host->state = STATE_SENDING_CMD;
+		dw_mci_start_request(host, slot);
+	} else {
+		dev_vdbg(host->dev, "list empty\n");
+
+		if (host->state == STATE_SENDING_CMD11)
+			host->state = STATE_WAITING_CMD11_DONE;
+		else
+			host->state = STATE_IDLE;
+	}
+
+	spin_unlock(&host->lock);
+	mmc_request_done(prev_mmc, mrq);
+	spin_lock(&host->lock);
+}
+
+static int dw_mci_command_complete(struct dw_mci *host, struct mmc_command *cmd)
+{
+	u32 status = host->cmd_status;
+
+	host->cmd_status = 0;
+
+	/* Read the response from the card (up to 16 bytes) */
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136) {
+			cmd->resp[3] = mci_readl(host, RESP0);
+			cmd->resp[2] = mci_readl(host, RESP1);
+			cmd->resp[1] = mci_readl(host, RESP2);
+			cmd->resp[0] = mci_readl(host, RESP3);
+		} else {
+			cmd->resp[0] = mci_readl(host, RESP0);
+			cmd->resp[1] = 0;
+			cmd->resp[2] = 0;
+			cmd->resp[3] = 0;
+		}
+	}
+
+	if (status & SDMMC_INT_RTO)
+		cmd->error = -ETIMEDOUT;
+	else if ((cmd->flags & MMC_RSP_CRC) && (status & SDMMC_INT_RCRC))
+		cmd->error = -EILSEQ;
+	else if (status & SDMMC_INT_RESP_ERR)
+		cmd->error = -EIO;
+	else
+		cmd->error = 0;
+
+	return cmd->error;
+}
+
+static int dw_mci_data_complete(struct dw_mci *host, struct mmc_data *data)
+{
+	u32 status = host->data_status;
+
+	if (status & DW_MCI_DATA_ERROR_FLAGS) {
+		if (status & SDMMC_INT_DRTO) {
+			data->error = -ETIMEDOUT;
+		} else if (status & SDMMC_INT_DCRC) {
+			data->error = -EILSEQ;
+		} else if (status & SDMMC_INT_EBE) {
+			if (host->dir_status ==
+				DW_MCI_SEND_STATUS) {
+				/*
+				 * No data CRC status was returned.
+				 * The number of bytes transferred
+				 * will be exaggerated in PIO mode.
+				 */
+				data->bytes_xfered = 0;
+				data->error = -ETIMEDOUT;
+			} else if (host->dir_status ==
+					DW_MCI_RECV_STATUS) {
+				data->error = -EILSEQ;
+			}
+		} else {
+			/* SDMMC_INT_SBE is included */
+			data->error = -EILSEQ;
+		}
+
+		dev_dbg(host->dev, "data error, status 0x%08x\n", status);
+
+		/*
+		 * After an error, there may be data lingering
+		 * in the FIFO
+		 */
+		dw_mci_reset(host);
+	} else {
+		data->bytes_xfered = data->blocks * data->blksz;
+		data->error = 0;
+	}
+
+	return data->error;
+}
+
+static void dw_mci_set_drto(struct dw_mci *host)
+{
+	unsigned int drto_clks;
+	unsigned int drto_div;
+	unsigned int drto_ms;
+	unsigned long irqflags;
+
+	drto_clks = mci_readl(host, TMOUT) >> 8;
+	drto_div = (mci_readl(host, CLKDIV) & 0xff) * 2;
+	if (drto_div == 0)
+		drto_div = 1;
+
+	drto_ms = DIV_ROUND_UP_ULL((u64)MSEC_PER_SEC * drto_clks * drto_div,
+				   host->bus_hz);
+
+	/* add a bit spare time */
+	drto_ms += 10;
+
+	spin_lock_irqsave(&host->irq_lock, irqflags);
+	if (!test_bit(EVENT_DATA_COMPLETE, &host->pending_events))
+		mod_timer(&host->dto_timer,
+			  jiffies + msecs_to_jiffies(drto_ms));
+	spin_unlock_irqrestore(&host->irq_lock, irqflags);
+}
+
+static bool dw_mci_clear_pending_cmd_complete(struct dw_mci *host)
+{
+	if (!test_bit(EVENT_CMD_COMPLETE, &host->pending_events))
+		return false;
+
+	/*
+	 * Really be certain that the timer has stopped.  This is a bit of
+	 * paranoia and could only really happen if we had really bad
+	 * interrupt latency and the interrupt routine and timeout were
+	 * running concurrently so that the del_timer() in the interrupt
+	 * handler couldn't run.
+	 */
+	WARN_ON(del_timer_sync(&host->cto_timer));
+	clear_bit(EVENT_CMD_COMPLETE, &host->pending_events);
+
+	return true;
+}
+
+static bool dw_mci_clear_pending_data_complete(struct dw_mci *host)
+{
+	if (!test_bit(EVENT_DATA_COMPLETE, &host->pending_events))
+		return false;
+
+	/* Extra paranoia just like dw_mci_clear_pending_cmd_complete() */
+	WARN_ON(del_timer_sync(&host->dto_timer));
+	clear_bit(EVENT_DATA_COMPLETE, &host->pending_events);
+
+	return true;
+}
+
+static void dw_mci_tasklet_func(unsigned long priv)
+{
+	struct dw_mci *host = (struct dw_mci *)priv;
+	struct mmc_data	*data;
+	struct mmc_command *cmd;
+	struct mmc_request *mrq;
+	enum dw_mci_state state;
+	enum dw_mci_state prev_state;
+	unsigned int err;
+
+	spin_lock(&host->lock);
+
+	state = host->state;
+	data = host->data;
+	mrq = host->mrq;
+
+	do {
+		prev_state = state;
+
+		switch (state) {
+		case STATE_IDLE:
+		case STATE_WAITING_CMD11_DONE:
+			break;
+
+		case STATE_SENDING_CMD11:
+		case STATE_SENDING_CMD:
+			if (!dw_mci_clear_pending_cmd_complete(host))
+				break;
+
+			cmd = host->cmd;
+			host->cmd = NULL;
+			set_bit(EVENT_CMD_COMPLETE, &host->completed_events);
+			err = dw_mci_command_complete(host, cmd);
+			if (cmd == mrq->sbc && !err) {
+				__dw_mci_start_request(host, host->slot,
+						       mrq->cmd);
+				goto unlock;
+			}
+
+			if (cmd->data && err) {
+				/*
+				 * During UHS tuning sequence, sending the stop
+				 * command after the response CRC error would
+				 * throw the system into a confused state
+				 * causing all future tuning phases to report
+				 * failure.
+				 *
+				 * In such case controller will move into a data
+				 * transfer state after a response error or
+				 * response CRC error. Let's let that finish
+				 * before trying to send a stop, so we'll go to
+				 * STATE_SENDING_DATA.
+				 *
+				 * Although letting the data transfer take place
+				 * will waste a bit of time (we already know
+				 * the command was bad), it can't cause any
+				 * errors since it's possible it would have
+				 * taken place anyway if this tasklet got
+				 * delayed. Allowing the transfer to take place
+				 * avoids races and keeps things simple.
+				 */
+				if (err != -ETIMEDOUT &&
+				    host->dir_status == DW_MCI_RECV_STATUS) {
+					state = STATE_SENDING_DATA;
+					continue;
+				}
+
+				send_stop_abort(host, data);
+				dw_mci_stop_dma(host);
+				state = STATE_SENDING_STOP;
+				break;
+			}
+
+			if (!cmd->data || err) {
+				dw_mci_request_end(host, mrq);
+				goto unlock;
+			}
+
+			prev_state = state = STATE_SENDING_DATA;
+			fallthrough;
+
+		case STATE_SENDING_DATA:
+			/*
+			 * We could get a data error and never a transfer
+			 * complete so we'd better check for it here.
+			 *
+			 * Note that we don't really care if we also got a
+			 * transfer complete; stopping the DMA and sending an
+			 * abort won't hurt.
+			 */
+			if (test_and_clear_bit(EVENT_DATA_ERROR,
+					       &host->pending_events)) {
+				if (!(host->data_status & (SDMMC_INT_DRTO |
+							   SDMMC_INT_EBE)))
+					send_stop_abort(host, data);
+				dw_mci_stop_dma(host);
+				state = STATE_DATA_ERROR;
+				break;
+			}
+
+			if (!test_and_clear_bit(EVENT_XFER_COMPLETE,
+						&host->pending_events)) {
+				/*
+				 * If all data-related interrupts don't come
+				 * within the given time in reading data state.
+				 */
+				if (host->dir_status == DW_MCI_RECV_STATUS)
+					dw_mci_set_drto(host);
+				break;
+			}
+
+			set_bit(EVENT_XFER_COMPLETE, &host->completed_events);
+
+			/*
+			 * Handle an EVENT_DATA_ERROR that might have shown up
+			 * before the transfer completed.  This might not have
+			 * been caught by the check above because the interrupt
+			 * could have gone off between the previous check and
+			 * the check for transfer complete.
+			 *
+			 * Technically this ought not be needed assuming we
+			 * get a DATA_COMPLETE eventually (we'll notice the
+			 * error and end the request), but it shouldn't hurt.
+			 *
+			 * This has the advantage of sending the stop command.
+			 */
+			if (test_and_clear_bit(EVENT_DATA_ERROR,
+					       &host->pending_events)) {
+				if (!(host->data_status & (SDMMC_INT_DRTO |
+							   SDMMC_INT_EBE)))
+					send_stop_abort(host, data);
+				dw_mci_stop_dma(host);
+				state = STATE_DATA_ERROR;
+				break;
+			}
+			prev_state = state = STATE_DATA_BUSY;
+
+			fallthrough;
+
+		case STATE_DATA_BUSY:
+			if (!dw_mci_clear_pending_data_complete(host)) {
+				/*
+				 * If data error interrupt comes but data over
+				 * interrupt doesn't come within the given time.
+				 * in reading data state.
+				 */
+				if (host->dir_status == DW_MCI_RECV_STATUS)
+					dw_mci_set_drto(host);
+				break;
+			}
+
+			host->data = NULL;
+			set_bit(EVENT_DATA_COMPLETE, &host->completed_events);
+			err = dw_mci_data_complete(host, data);
+
+			if (!err) {
+				if (!data->stop || mrq->sbc) {
+					if (mrq->sbc && data->stop)
+						data->stop->error = 0;
+					dw_mci_request_end(host, mrq);
+					goto unlock;
+				}
+
+				/* stop command for open-ended transfer*/
+				if (data->stop)
+					send_stop_abort(host, data);
+			} else {
+				/*
+				 * If we don't have a command complete now we'll
+				 * never get one since we just reset everything;
+				 * better end the request.
+				 *
+				 * If we do have a command complete we'll fall
+				 * through to the SENDING_STOP command and
+				 * everything will be peachy keen.
+				 */
+				if (!test_bit(EVENT_CMD_COMPLETE,
+					      &host->pending_events)) {
+					host->cmd = NULL;
+					dw_mci_request_end(host, mrq);
+					goto unlock;
+				}
+			}
+
+			/*
+			 * If err has non-zero,
+			 * stop-abort command has been already issued.
+			 */
+			prev_state = state = STATE_SENDING_STOP;
+
+			fallthrough;
+
+		case STATE_SENDING_STOP:
+			if (!dw_mci_clear_pending_cmd_complete(host))
+				break;
+
+			/* CMD error in data command */
+			if (mrq->cmd->error && mrq->data)
+				dw_mci_reset(host);
+
+			host->cmd = NULL;
+			host->data = NULL;
+
+			if (!mrq->sbc && mrq->stop)
+				dw_mci_command_complete(host, mrq->stop);
+			else
+				host->cmd_status = 0;
+
+			dw_mci_request_end(host, mrq);
+			goto unlock;
+
+		case STATE_DATA_ERROR:
+			if (!test_and_clear_bit(EVENT_XFER_COMPLETE,
+						&host->pending_events))
+				break;
+
+			state = STATE_DATA_BUSY;
+			break;
+		}
+	} while (state != prev_state);
+
+	host->state = state;
+unlock:
+	spin_unlock(&host->lock);
+
+}
+
+/* push final bytes to part_buf, only use during push */
+static void dw_mci_set_part_bytes(struct dw_mci *host, void *buf, int cnt)
+{
+	memcpy((void *)&host->part_buf, buf, cnt);
+	host->part_buf_count = cnt;
+}
+
+/* append bytes to part_buf, only use during push */
+static int dw_mci_push_part_bytes(struct dw_mci *host, void *buf, int cnt)
+{
+	cnt = min(cnt, (1 << host->data_shift) - host->part_buf_count);
+	memcpy((void *)&host->part_buf + host->part_buf_count, buf, cnt);
+	host->part_buf_count += cnt;
+	return cnt;
+}
+
+/* pull first bytes from part_buf, only use during pull */
+static int dw_mci_pull_part_bytes(struct dw_mci *host, void *buf, int cnt)
+{
+	cnt = min_t(int, cnt, host->part_buf_count);
+	if (cnt) {
+		memcpy(buf, (void *)&host->part_buf + host->part_buf_start,
+		       cnt);
+		host->part_buf_count -= cnt;
+		host->part_buf_start += cnt;
+	}
+	return cnt;
+}
+
+/* pull final bytes from the part_buf, assuming it's just been filled */
+static void dw_mci_pull_final_bytes(struct dw_mci *host, void *buf, int cnt)
+{
+	memcpy(buf, &host->part_buf, cnt);
+	host->part_buf_start = cnt;
+	host->part_buf_count = (1 << host->data_shift) - cnt;
+}
+
+static void dw_mci_push_data16(struct dw_mci *host, void *buf, int cnt)
+{
+	struct mmc_data *data = host->data;
+	int init_cnt = cnt;
+
+	/* try and push anything in the part_buf */
+	if (unlikely(host->part_buf_count)) {
+		int len = dw_mci_push_part_bytes(host, buf, cnt);
+
+		buf += len;
+		cnt -= len;
+		if (host->part_buf_count == 2) {
+			mci_fifo_writew(host->fifo_reg, host->part_buf16);
+			host->part_buf_count = 0;
+		}
+	}
+#ifndef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
+	if (unlikely((unsigned long)buf & 0x1)) {
+		while (cnt >= 2) {
+			u16 aligned_buf[64];
+			int len = min(cnt & -2, (int)sizeof(aligned_buf));
+			int items = len >> 1;
+			int i;
+			/* memcpy from input buffer into aligned buffer */
+			memcpy(aligned_buf, buf, len);
+			buf += len;
+			cnt -= len;
+			/* push data from aligned buffer into fifo */
+			for (i = 0; i < items; ++i)
+				mci_fifo_writew(host->fifo_reg, aligned_buf[i]);
+		}
+	} else
+#endif
+	{
+		u16 *pdata = buf;
+
+		for (; cnt >= 2; cnt -= 2)
+			mci_fifo_writew(host->fifo_reg, *pdata++);
+		buf = pdata;
+	}
+	/* put anything remaining in the part_buf */
+	if (cnt) {
+		dw_mci_set_part_bytes(host, buf, cnt);
+		 /* Push data if we have reached the expected data length */
+		if ((data->bytes_xfered + init_cnt) ==
+		    (data->blksz * data->blocks))
+			mci_fifo_writew(host->fifo_reg, host->part_buf16);
+	}
+}
+
+static void dw_mci_pull_data16(struct dw_mci *host, void *buf, int cnt)
+{
+#ifndef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
+	if (unlikely((unsigned long)buf & 0x1)) {
+		while (cnt >= 2) {
+			/* pull data from fifo into aligned buffer */
+			u16 aligned_buf[64];
+			int len = min(cnt & -2, (int)sizeof(aligned_buf));
+			int items = len >> 1;
+			int i;
+
+			for (i = 0; i < items; ++i)
+				aligned_buf[i] = mci_fifo_readw(host->fifo_reg);
+			/* memcpy from aligned buffer into output buffer */
+			memcpy(buf, aligned_buf, len);
+			buf += len;
+			cnt -= len;
+		}
+	} else
+#endif
+	{
+		u16 *pdata = buf;
+
+		for (; cnt >= 2; cnt -= 2)
+			*pdata++ = mci_fifo_readw(host->fifo_reg);
+		buf = pdata;
+	}
+	if (cnt) {
+		host->part_buf16 = mci_fifo_readw(host->fifo_reg);
+		dw_mci_pull_final_bytes(host, buf, cnt);
+	}
+}
+
+static void dw_mci_push_data32(struct dw_mci *host, void *buf, int cnt)
+{
+	struct mmc_data *data = host->data;
+	int init_cnt = cnt;
+
+	/* try and push anything in the part_buf */
+	if (unlikely(host->part_buf_count)) {
+		int len = dw_mci_push_part_bytes(host, buf, cnt);
+
+		buf += len;
+		cnt -= len;
+		if (host->part_buf_count == 4) {
+			mci_fifo_writel(host->fifo_reg,	host->part_buf32);
+			host->part_buf_count = 0;
+		}
+	}
+#ifndef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
+	if (unlikely((unsigned long)buf & 0x3)) {
+		while (cnt >= 4) {
+			u32 aligned_buf[32];
+			int len = min(cnt & -4, (int)sizeof(aligned_buf));
+			int items = len >> 2;
+			int i;
+			/* memcpy from input buffer into aligned buffer */
+			memcpy(aligned_buf, buf, len);
+			buf += len;
+			cnt -= len;
+			/* push data from aligned buffer into fifo */
+			for (i = 0; i < items; ++i)
+				mci_fifo_writel(host->fifo_reg,	aligned_buf[i]);
+		}
+	} else
+#endif
+	{
+		u32 *pdata = buf;
+
+		for (; cnt >= 4; cnt -= 4)
+			mci_fifo_writel(host->fifo_reg, *pdata++);
+		buf = pdata;
+	}
+	/* put anything remaining in the part_buf */
+	if (cnt) {
+		dw_mci_set_part_bytes(host, buf, cnt);
+		 /* Push data if we have reached the expected data length */
+		if ((data->bytes_xfered + init_cnt) ==
+		    (data->blksz * data->blocks))
+			mci_fifo_writel(host->fifo_reg, host->part_buf32);
+	}
+}
+
+static void dw_mci_pull_data32(struct dw_mci *host, void *buf, int cnt)
+{
+#ifndef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
+	if (unlikely((unsigned long)buf & 0x3)) {
+		while (cnt >= 4) {
+			/* pull data from fifo into aligned buffer */
+			u32 aligned_buf[32];
+			int len = min(cnt & -4, (int)sizeof(aligned_buf));
+			int items = len >> 2;
+			int i;
+
+			for (i = 0; i < items; ++i)
+				aligned_buf[i] = mci_fifo_readl(host->fifo_reg);
+			/* memcpy from aligned buffer into output buffer */
+			memcpy(buf, aligned_buf, len);
+			buf += len;
+			cnt -= len;
+		}
+	} else
+#endif
+	{
+		u32 *pdata = buf;
+
+		for (; cnt >= 4; cnt -= 4)
+			*pdata++ = mci_fifo_readl(host->fifo_reg);
+		buf = pdata;
+	}
+	if (cnt) {
+		host->part_buf32 = mci_fifo_readl(host->fifo_reg);
+		dw_mci_pull_final_bytes(host, buf, cnt);
+	}
+}
+
+static void dw_mci_push_data64(struct dw_mci *host, void *buf, int cnt)
+{
+	struct mmc_data *data = host->data;
+	int init_cnt = cnt;
+
+	/* try and push anything in the part_buf */
+	if (unlikely(host->part_buf_count)) {
+		int len = dw_mci_push_part_bytes(host, buf, cnt);
+
+		buf += len;
+		cnt -= len;
+
+		if (host->part_buf_count == 8) {
+			mci_fifo_writeq(host->fifo_reg,	host->part_buf);
+			host->part_buf_count = 0;
+		}
+	}
+#ifndef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
+	if (unlikely((unsigned long)buf & 0x7)) {
+		while (cnt >= 8) {
+			u64 aligned_buf[16];
+			int len = min(cnt & -8, (int)sizeof(aligned_buf));
+			int items = len >> 3;
+			int i;
+			/* memcpy from input buffer into aligned buffer */
+			memcpy(aligned_buf, buf, len);
+			buf += len;
+			cnt -= len;
+			/* push data from aligned buffer into fifo */
+			for (i = 0; i < items; ++i)
+				mci_fifo_writeq(host->fifo_reg,	aligned_buf[i]);
+		}
+	} else
+#endif
+	{
+		u64 *pdata = buf;
+
+		for (; cnt >= 8; cnt -= 8)
+			mci_fifo_writeq(host->fifo_reg, *pdata++);
+		buf = pdata;
+	}
+	/* put anything remaining in the part_buf */
+	if (cnt) {
+		dw_mci_set_part_bytes(host, buf, cnt);
+		/* Push data if we have reached the expected data length */
+		if ((data->bytes_xfered + init_cnt) ==
+		    (data->blksz * data->blocks))
+			mci_fifo_writeq(host->fifo_reg, host->part_buf);
+	}
+}
+
+static void dw_mci_pull_data64(struct dw_mci *host, void *buf, int cnt)
+{
+#ifndef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
+	if (unlikely((unsigned long)buf & 0x7)) {
+		while (cnt >= 8) {
+			/* pull data from fifo into aligned buffer */
+			u64 aligned_buf[16];
+			int len = min(cnt & -8, (int)sizeof(aligned_buf));
+			int items = len >> 3;
+			int i;
+
+			for (i = 0; i < items; ++i)
+				aligned_buf[i] = mci_fifo_readq(host->fifo_reg);
+
+			/* memcpy from aligned buffer into output buffer */
+			memcpy(buf, aligned_buf, len);
+			buf += len;
+			cnt -= len;
+		}
+	} else
+#endif
+	{
+		u64 *pdata = buf;
+
+		for (; cnt >= 8; cnt -= 8)
+			*pdata++ = mci_fifo_readq(host->fifo_reg);
+		buf = pdata;
+	}
+	if (cnt) {
+		host->part_buf = mci_fifo_readq(host->fifo_reg);
+		dw_mci_pull_final_bytes(host, buf, cnt);
+	}
+}
+
+static void dw_mci_pull_data(struct dw_mci *host, void *buf, int cnt)
+{
+	int len;
+
+	/* get remaining partial bytes */
+	len = dw_mci_pull_part_bytes(host, buf, cnt);
+	if (unlikely(len == cnt))
+		return;
+	buf += len;
+	cnt -= len;
+
+	/* get the rest of the data */
+	host->pull_data(host, buf, cnt);
+}
+
+static void dw_mci_read_data_pio(struct dw_mci *host, bool dto)
+{
+	struct sg_mapping_iter *sg_miter = &host->sg_miter;
+	void *buf;
+	unsigned int offset;
+	struct mmc_data	*data = host->data;
+	int shift = host->data_shift;
+	u32 status;
+	unsigned int len;
+	unsigned int remain, fcnt;
+
+	do {
+		if (!sg_miter_next(sg_miter))
+			goto done;
+
+		host->sg = sg_miter->piter.sg;
+		buf = sg_miter->addr;
+		remain = sg_miter->length;
+		offset = 0;
+
+		do {
+			fcnt = (SDMMC_GET_FCNT(mci_readl(host, STATUS))
+					<< shift) + host->part_buf_count;
+			len = min(remain, fcnt);
+			if (!len)
+				break;
+			dw_mci_pull_data(host, (void *)(buf + offset), len);
+			data->bytes_xfered += len;
+			offset += len;
+			remain -= len;
+		} while (remain);
+
+		sg_miter->consumed = offset;
+		status = mci_readl(host, MINTSTS);
+		mci_writel(host, RINTSTS, SDMMC_INT_RXDR);
+	/* if the RXDR is ready read again */
+	} while ((status & SDMMC_INT_RXDR) ||
+		 (dto && SDMMC_GET_FCNT(mci_readl(host, STATUS))));
+
+	if (!remain) {
+		if (!sg_miter_next(sg_miter))
+			goto done;
+		sg_miter->consumed = 0;
+	}
+	sg_miter_stop(sg_miter);
+	return;
+
+done:
+	sg_miter_stop(sg_miter);
+	host->sg = NULL;
+	smp_wmb(); /* drain writebuffer */
+	set_bit(EVENT_XFER_COMPLETE, &host->pending_events);
+}
+
+static void dw_mci_write_data_pio(struct dw_mci *host)
+{
+	struct sg_mapping_iter *sg_miter = &host->sg_miter;
+	void *buf;
+	unsigned int offset;
+	struct mmc_data	*data = host->data;
+	int shift = host->data_shift;
+	u32 status;
+	unsigned int len;
+	unsigned int fifo_depth = host->fifo_depth;
+	unsigned int remain, fcnt;
+
+	do {
+		if (!sg_miter_next(sg_miter))
+			goto done;
+
+		host->sg = sg_miter->piter.sg;
+		buf = sg_miter->addr;
+		remain = sg_miter->length;
+		offset = 0;
+
+		do {
+			fcnt = ((fifo_depth -
+				 SDMMC_GET_FCNT(mci_readl(host, STATUS)))
+					<< shift) - host->part_buf_count;
+			len = min(remain, fcnt);
+			if (!len)
+				break;
+			host->push_data(host, (void *)(buf + offset), len);
+			data->bytes_xfered += len;
+			offset += len;
+			remain -= len;
+		} while (remain);
+
+		sg_miter->consumed = offset;
+		status = mci_readl(host, MINTSTS);
+		mci_writel(host, RINTSTS, SDMMC_INT_TXDR);
+	} while (status & SDMMC_INT_TXDR); /* if TXDR write again */
+
+	if (!remain) {
+		if (!sg_miter_next(sg_miter))
+			goto done;
+		sg_miter->consumed = 0;
+	}
+	sg_miter_stop(sg_miter);
+	return;
+
+done:
+	sg_miter_stop(sg_miter);
+	host->sg = NULL;
+	smp_wmb(); /* drain writebuffer */
+	set_bit(EVENT_XFER_COMPLETE, &host->pending_events);
+}
+
+static void dw_mci_cmd_interrupt(struct dw_mci *host, u32 status)
+{
+	del_timer(&host->cto_timer);
+
+	if (!host->cmd_status)
+		host->cmd_status = status;
+
+	smp_wmb(); /* drain writebuffer */
+
+	set_bit(EVENT_CMD_COMPLETE, &host->pending_events);
+	tasklet_schedule(&host->tasklet);
+}
+
+static void dw_mci_handle_cd(struct dw_mci *host)
+{
+	struct dw_mci_slot *slot = host->slot;
+
+	if (slot->mmc->ops->card_event)
+		slot->mmc->ops->card_event(slot->mmc);
+	mmc_detect_change(slot->mmc,
+		msecs_to_jiffies(host->pdata->detect_delay_ms));
+}
+
+static irqreturn_t dw_mci_interrupt(int irq, void *dev_id)
+{
+	struct dw_mci *host = dev_id;
+	u32 pending;
+	struct dw_mci_slot *slot = host->slot;
+	unsigned long irqflags;
+
+	pending = mci_readl(host, MINTSTS); /* read-only mask reg */
+
+	if (pending) {
+		/* Check volt switch first, since it can look like an error */
+		if ((host->state == STATE_SENDING_CMD11) &&
+		    (pending & SDMMC_INT_VOLT_SWITCH)) {
+			mci_writel(host, RINTSTS, SDMMC_INT_VOLT_SWITCH);
+			pending &= ~SDMMC_INT_VOLT_SWITCH;
+
+			/*
+			 * Hold the lock; we know cmd11_timer can't be kicked
+			 * off after the lock is released, so safe to delete.
+			 */
+			spin_lock_irqsave(&host->irq_lock, irqflags);
+			dw_mci_cmd_interrupt(host, pending);
+			spin_unlock_irqrestore(&host->irq_lock, irqflags);
+
+			del_timer(&host->cmd11_timer);
+		}
+
+		if (pending & DW_MCI_CMD_ERROR_FLAGS) {
+			spin_lock_irqsave(&host->irq_lock, irqflags);
+
+			del_timer(&host->cto_timer);
+			mci_writel(host, RINTSTS, DW_MCI_CMD_ERROR_FLAGS);
+			host->cmd_status = pending;
+			smp_wmb(); /* drain writebuffer */
+			set_bit(EVENT_CMD_COMPLETE, &host->pending_events);
+
+			spin_unlock_irqrestore(&host->irq_lock, irqflags);
+		}
+
+		if (pending & DW_MCI_DATA_ERROR_FLAGS) {
+			/* if there is an error report DATA_ERROR */
+			mci_writel(host, RINTSTS, DW_MCI_DATA_ERROR_FLAGS);
+			host->data_status = pending;
+			smp_wmb(); /* drain writebuffer */
+			set_bit(EVENT_DATA_ERROR, &host->pending_events);
+			tasklet_schedule(&host->tasklet);
+		}
+
+		if (pending & SDMMC_INT_DATA_OVER) {
+			spin_lock_irqsave(&host->irq_lock, irqflags);
+
+			del_timer(&host->dto_timer);
+
+			mci_writel(host, RINTSTS, SDMMC_INT_DATA_OVER);
+			if (!host->data_status)
+				host->data_status = pending;
+			smp_wmb(); /* drain writebuffer */
+			if (host->dir_status == DW_MCI_RECV_STATUS) {
+				if (host->sg != NULL)
+					dw_mci_read_data_pio(host, true);
+			}
+			set_bit(EVENT_DATA_COMPLETE, &host->pending_events);
+			tasklet_schedule(&host->tasklet);
+
+			spin_unlock_irqrestore(&host->irq_lock, irqflags);
+		}
+
+		if (pending & SDMMC_INT_RXDR) {
+			mci_writel(host, RINTSTS, SDMMC_INT_RXDR);
+			if (host->dir_status == DW_MCI_RECV_STATUS && host->sg)
+				dw_mci_read_data_pio(host, false);
+		}
+
+		if (pending & SDMMC_INT_TXDR) {
+			mci_writel(host, RINTSTS, SDMMC_INT_TXDR);
+			if (host->dir_status == DW_MCI_SEND_STATUS && host->sg)
+				dw_mci_write_data_pio(host);
+		}
+
+		if (pending & SDMMC_INT_CMD_DONE) {
+			spin_lock_irqsave(&host->irq_lock, irqflags);
+
+			mci_writel(host, RINTSTS, SDMMC_INT_CMD_DONE);
+			dw_mci_cmd_interrupt(host, pending);
+
+			spin_unlock_irqrestore(&host->irq_lock, irqflags);
+		}
+
+		if (pending & SDMMC_INT_CD) {
+			mci_writel(host, RINTSTS, SDMMC_INT_CD);
+			dw_mci_handle_cd(host);
+		}
+
+		if (pending & SDMMC_INT_SDIO(slot->sdio_id)) {
+			mci_writel(host, RINTSTS,
+				   SDMMC_INT_SDIO(slot->sdio_id));
+			__dw_mci_enable_sdio_irq(slot, 0);
+			sdio_signal_irq(slot->mmc);
+		}
+
+	}
+
+	if (host->use_dma != TRANS_MODE_IDMAC)
+		return IRQ_HANDLED;
+
+	/* Handle IDMA interrupts */
+	if (host->dma_64bit_address == 1) {
+		pending = mci_readl(host, IDSTS64);
+		if (pending & (SDMMC_IDMAC_INT_TI | SDMMC_IDMAC_INT_RI)) {
+			mci_writel(host, IDSTS64, SDMMC_IDMAC_INT_TI |
+							SDMMC_IDMAC_INT_RI);
+			mci_writel(host, IDSTS64, SDMMC_IDMAC_INT_NI);
+			if (!test_bit(EVENT_DATA_ERROR, &host->pending_events))
+				host->dma_ops->complete((void *)host);
+		}
+	} else {
+		pending = mci_readl(host, IDSTS);
+		if (pending & (SDMMC_IDMAC_INT_TI | SDMMC_IDMAC_INT_RI)) {
+			mci_writel(host, IDSTS, SDMMC_IDMAC_INT_TI |
+							SDMMC_IDMAC_INT_RI);
+			mci_writel(host, IDSTS, SDMMC_IDMAC_INT_NI);
+			if (!test_bit(EVENT_DATA_ERROR, &host->pending_events))
+				host->dma_ops->complete((void *)host);
+		}
+	}
+
+	return IRQ_HANDLED;
+}
+
+static int dw_mci_init_slot_caps(struct dw_mci_slot *slot)
+{
+	struct dw_mci *host = slot->host;
+	const struct dw_mci_drv_data *drv_data = host->drv_data;
+	struct mmc_host *mmc = slot->mmc;
+	int ctrl_id;
+
+	if (host->pdata->caps)
+		mmc->caps = host->pdata->caps;
+
+	if (host->pdata->pm_caps)
+		mmc->pm_caps = host->pdata->pm_caps;
+
+	if (host->dev->of_node) {
+		ctrl_id = of_alias_get_id(host->dev->of_node, "mshc");
+		if (ctrl_id < 0)
+			ctrl_id = 0;
+	} else {
+		ctrl_id = to_platform_device(host->dev)->id;
+	}
+
+	if (drv_data && drv_data->caps) {
+		if (ctrl_id >= drv_data->num_caps) {
+			dev_err(host->dev, "invalid controller id %d\n",
+				ctrl_id);
+			return -EINVAL;
+		}
+		mmc->caps |= drv_data->caps[ctrl_id];
+	}
+
+	if (host->pdata->caps2)
+		mmc->caps2 = host->pdata->caps2;
+
+	mmc->f_min = DW_MCI_FREQ_MIN;
+	if (!mmc->f_max)
+		mmc->f_max = DW_MCI_FREQ_MAX;
+
+	/* Process SDIO IRQs through the sdio_irq_work. */
+	if (mmc->caps & MMC_CAP_SDIO_IRQ)
+		mmc->caps2 |= MMC_CAP2_SDIO_IRQ_NOTHREAD;
+
+	return 0;
+}
+
+static int dw_mci_init_slot(struct dw_mci *host)
+{
+	struct mmc_host *mmc;
+	struct dw_mci_slot *slot;
+	int ret;
+
+	mmc = mmc_alloc_host(sizeof(struct dw_mci_slot), host->dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	slot = mmc_priv(mmc);
+	slot->id = 0;
+	slot->sdio_id = host->sdio_id0 + slot->id;
+	slot->mmc = mmc;
+	slot->host = host;
+	host->slot = slot;
+
+	mmc->ops = &dw_mci_ops;
+
+	/*if there are external regulators, get them*/
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret)
+		goto err_host_allocated;
+
+	if (!mmc->ocr_avail)
+		mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto err_host_allocated;
+
+	ret = dw_mci_init_slot_caps(slot);
+	if (ret)
+		goto err_host_allocated;
+
+	/* Useful defaults if platform data is unset. */
+	if (host->use_dma == TRANS_MODE_IDMAC) {
+		mmc->max_segs = host->ring_size;
+		mmc->max_blk_size = 65535;
+		mmc->max_seg_size = 0x1000;
+		mmc->max_req_size = mmc->max_seg_size * host->ring_size;
+		mmc->max_blk_count = mmc->max_req_size / 512;
+	} else if (host->use_dma == TRANS_MODE_EDMAC) {
+		mmc->max_segs = 64;
+		mmc->max_blk_size = 65535;
+		mmc->max_blk_count = 65535;
+		mmc->max_req_size =
+				mmc->max_blk_size * mmc->max_blk_count;
+		mmc->max_seg_size = mmc->max_req_size;
+	} else {
+		/* TRANS_MODE_PIO */
+		mmc->max_segs = 64;
+		mmc->max_blk_size = 65535; /* BLKSIZ is 16 bits */
+		mmc->max_blk_count = 512;
+		mmc->max_req_size = mmc->max_blk_size *
+				    mmc->max_blk_count;
+		mmc->max_seg_size = mmc->max_req_size;
+	}
+
+	dw_mci_get_cd(mmc);
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto err_host_allocated;
+
+#if defined(CONFIG_DEBUG_FS)
+	dw_mci_init_debugfs(slot);
+#endif
+
+	return 0;
+
+err_host_allocated:
+	mmc_free_host(mmc);
+	return ret;
+}
+
+static void dw_mci_cleanup_slot(struct dw_mci_slot *slot)
+{
+	/* Debugfs stuff is cleaned up by mmc core */
+	mmc_remove_host(slot->mmc);
+	slot->host->slot = NULL;
+	mmc_free_host(slot->mmc);
+}
+
+static void dw_mci_init_dma(struct dw_mci *host)
+{
+	int addr_config;
+	struct device *dev = host->dev;
+
+	/*
+	* Check tansfer mode from HCON[17:16]
+	* Clear the ambiguous description of dw_mmc databook:
+	* 2b'00: No DMA Interface -> Actually means using Internal DMA block
+	* 2b'01: DesignWare DMA Interface -> Synopsys DW-DMA block
+	* 2b'10: Generic DMA Interface -> non-Synopsys generic DMA block
+	* 2b'11: Non DW DMA Interface -> pio only
+	* Compared to DesignWare DMA Interface, Generic DMA Interface has a
+	* simpler request/acknowledge handshake mechanism and both of them
+	* are regarded as external dma master for dw_mmc.
+	*/
+	host->use_dma = SDMMC_GET_TRANS_MODE(mci_readl(host, HCON));
+	if (host->use_dma == DMA_INTERFACE_IDMA) {
+		host->use_dma = TRANS_MODE_IDMAC;
+	} else if (host->use_dma == DMA_INTERFACE_DWDMA ||
+		   host->use_dma == DMA_INTERFACE_GDMA) {
+		host->use_dma = TRANS_MODE_EDMAC;
+	} else {
+		goto no_dma;
+	}
+
+	/* Determine which DMA interface to use */
+	if (host->use_dma == TRANS_MODE_IDMAC) {
+		/*
+		* Check ADDR_CONFIG bit in HCON to find
+		* IDMAC address bus width
+		*/
+		addr_config = SDMMC_GET_ADDR_CONFIG(mci_readl(host, HCON));
+
+		if (addr_config == 1) {
+			/* host supports IDMAC in 64-bit address mode */
+			host->dma_64bit_address = 1;
+			dev_info(host->dev,
+				 "IDMAC supports 64-bit address mode.\n");
+			if (!dma_set_mask(host->dev, DMA_BIT_MASK(64)))
+				dma_set_coherent_mask(host->dev,
+						      DMA_BIT_MASK(64));
+		} else {
+			/* host supports IDMAC in 32-bit address mode */
+			host->dma_64bit_address = 0;
+			dev_info(host->dev,
+				 "IDMAC supports 32-bit address mode.\n");
+		}
+
+		/* Alloc memory for sg translation */
+		host->sg_cpu = dmam_alloc_coherent(host->dev,
+						   DESC_RING_BUF_SZ,
+						   &host->sg_dma, GFP_KERNEL);
+		if (!host->sg_cpu) {
+			dev_err(host->dev,
+				"%s: could not alloc DMA memory\n",
+				__func__);
+			goto no_dma;
+		}
+
+		host->dma_ops = &dw_mci_idmac_ops;
+		dev_info(host->dev, "Using internal DMA controller.\n");
+	} else {
+		/* TRANS_MODE_EDMAC: check dma bindings again */
+		if ((device_property_read_string_array(dev, "dma-names",
+						       NULL, 0) < 0) ||
+		    !device_property_present(dev, "dmas")) {
+			goto no_dma;
+		}
+		host->dma_ops = &dw_mci_edmac_ops;
+		dev_info(host->dev, "Using external DMA controller.\n");
+	}
+
+	if (host->dma_ops->init && host->dma_ops->start &&
+	    host->dma_ops->stop && host->dma_ops->cleanup) {
+		if (host->dma_ops->init(host)) {
+			dev_err(host->dev, "%s: Unable to initialize DMA Controller.\n",
+				__func__);
+			goto no_dma;
+		}
+	} else {
+		dev_err(host->dev, "DMA initialization not found.\n");
+		goto no_dma;
+	}
+
+	return;
+
+no_dma:
+	dev_info(host->dev, "Using PIO mode.\n");
+	host->use_dma = TRANS_MODE_PIO;
+}
+
+static void dw_mci_cmd11_timer(struct timer_list *t)
+{
+	struct dw_mci *host = from_timer(host, t, cmd11_timer);
+
+	if (host->state != STATE_SENDING_CMD11) {
+		dev_warn(host->dev, "Unexpected CMD11 timeout\n");
+		return;
+	}
+
+	host->cmd_status = SDMMC_INT_RTO;
+	set_bit(EVENT_CMD_COMPLETE, &host->pending_events);
+	tasklet_schedule(&host->tasklet);
+}
+
+static void dw_mci_cto_timer(struct timer_list *t)
+{
+	struct dw_mci *host = from_timer(host, t, cto_timer);
+	unsigned long irqflags;
+	u32 pending;
+
+	spin_lock_irqsave(&host->irq_lock, irqflags);
+
+	/*
+	 * If somehow we have very bad interrupt latency it's remotely possible
+	 * that the timer could fire while the interrupt is still pending or
+	 * while the interrupt is midway through running.  Let's be paranoid
+	 * and detect those two cases.  Note that this is paranoia is somewhat
+	 * justified because in this function we don't actually cancel the
+	 * pending command in the controller--we just assume it will never come.
+	 */
+	pending = mci_readl(host, MINTSTS); /* read-only mask reg */
+	if (pending & (DW_MCI_CMD_ERROR_FLAGS | SDMMC_INT_CMD_DONE)) {
+		/* The interrupt should fire; no need to act but we can warn */
+		dev_warn(host->dev, "Unexpected interrupt latency\n");
+		goto exit;
+	}
+	if (test_bit(EVENT_CMD_COMPLETE, &host->pending_events)) {
+		/* Presumably interrupt handler couldn't delete the timer */
+		dev_warn(host->dev, "CTO timeout when already completed\n");
+		goto exit;
+	}
+
+	/*
+	 * Continued paranoia to make sure we're in the state we expect.
+	 * This paranoia isn't really justified but it seems good to be safe.
+	 */
+	switch (host->state) {
+	case STATE_SENDING_CMD11:
+	case STATE_SENDING_CMD:
+	case STATE_SENDING_STOP:
+		/*
+		 * If CMD_DONE interrupt does NOT come in sending command
+		 * state, we should notify the driver to terminate current
+		 * transfer and report a command timeout to the core.
+		 */
+		host->cmd_status = SDMMC_INT_RTO;
+		set_bit(EVENT_CMD_COMPLETE, &host->pending_events);
+		tasklet_schedule(&host->tasklet);
+		break;
+	default:
+		dev_warn(host->dev, "Unexpected command timeout, state %d\n",
+			 host->state);
+		break;
+	}
+
+exit:
+	spin_unlock_irqrestore(&host->irq_lock, irqflags);
+}
+
+static void dw_mci_dto_timer(struct timer_list *t)
+{
+	struct dw_mci *host = from_timer(host, t, dto_timer);
+	unsigned long irqflags;
+	u32 pending;
+
+	spin_lock_irqsave(&host->irq_lock, irqflags);
+
+	/*
+	 * The DTO timer is much longer than the CTO timer, so it's even less
+	 * likely that we'll these cases, but it pays to be paranoid.
+	 */
+	pending = mci_readl(host, MINTSTS); /* read-only mask reg */
+	if (pending & SDMMC_INT_DATA_OVER) {
+		/* The interrupt should fire; no need to act but we can warn */
+		dev_warn(host->dev, "Unexpected data interrupt latency\n");
+		goto exit;
+	}
+	if (test_bit(EVENT_DATA_COMPLETE, &host->pending_events)) {
+		/* Presumably interrupt handler couldn't delete the timer */
+		dev_warn(host->dev, "DTO timeout when already completed\n");
+		goto exit;
+	}
+
+	/*
+	 * Continued paranoia to make sure we're in the state we expect.
+	 * This paranoia isn't really justified but it seems good to be safe.
+	 */
+	switch (host->state) {
+	case STATE_SENDING_DATA:
+	case STATE_DATA_BUSY:
+		/*
+		 * If DTO interrupt does NOT come in sending data state,
+		 * we should notify the driver to terminate current transfer
+		 * and report a data timeout to the core.
+		 */
+		host->data_status = SDMMC_INT_DRTO;
+		set_bit(EVENT_DATA_ERROR, &host->pending_events);
+		set_bit(EVENT_DATA_COMPLETE, &host->pending_events);
+		tasklet_schedule(&host->tasklet);
+		break;
+	default:
+		dev_warn(host->dev, "Unexpected data timeout, state %d\n",
+			 host->state);
+		break;
+	}
+
+exit:
+	spin_unlock_irqrestore(&host->irq_lock, irqflags);
+}
+
+#ifdef CONFIG_OF
+static struct dw_mci_board *dw_mci_parse_dt(struct dw_mci *host)
+{
+	struct dw_mci_board *pdata;
+	struct device *dev = host->dev;
+	const struct dw_mci_drv_data *drv_data = host->drv_data;
+	int ret;
+	u32 clock_frequency;
+
+	pdata = devm_kzalloc(dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata)
+		return ERR_PTR(-ENOMEM);
+
+	/* find reset controller when exist */
+	pdata->rstc = devm_reset_control_get_optional_exclusive(dev, "reset");
+	if (IS_ERR(pdata->rstc)) {
+		if (PTR_ERR(pdata->rstc) == -EPROBE_DEFER)
+			return ERR_PTR(-EPROBE_DEFER);
+	}
+
+	if (device_property_read_u32(dev, "fifo-depth", &pdata->fifo_depth))
+		dev_info(dev,
+			 "fifo-depth property not found, using value of FIFOTH register as default\n");
+
+	device_property_read_u32(dev, "card-detect-delay",
+				 &pdata->detect_delay_ms);
+
+	device_property_read_u32(dev, "data-addr", &host->data_addr_override);
+
+	if (device_property_present(dev, "fifo-watermark-aligned"))
+		host->wm_aligned = true;
+
+	if (!device_property_read_u32(dev, "clock-frequency", &clock_frequency))
+		pdata->bus_hz = clock_frequency;
+
+	if (drv_data && drv_data->parse_dt) {
+		ret = drv_data->parse_dt(host);
+		if (ret)
+			return ERR_PTR(ret);
+	}
+
+	return pdata;
+}
+
+#else /* CONFIG_OF */
+static struct dw_mci_board *dw_mci_parse_dt(struct dw_mci *host)
+{
+	return ERR_PTR(-EINVAL);
+}
+#endif /* CONFIG_OF */
+
+static void dw_mci_enable_cd(struct dw_mci *host)
+{
+	unsigned long irqflags;
+	u32 temp;
+
+	/*
+	 * No need for CD if all slots have a non-error GPIO
+	 * as well as broken card detection is found.
+	 */
+	if (host->slot->mmc->caps & MMC_CAP_NEEDS_POLL)
+		return;
+
+	if (mmc_gpio_get_cd(host->slot->mmc) < 0) {
+		spin_lock_irqsave(&host->irq_lock, irqflags);
+		temp = mci_readl(host, INTMASK);
+		temp  |= SDMMC_INT_CD;
+		mci_writel(host, INTMASK, temp);
+		spin_unlock_irqrestore(&host->irq_lock, irqflags);
+	}
+}
+
+int dw_mci_probe(struct dw_mci *host)
+{
+	const struct dw_mci_drv_data *drv_data = host->drv_data;
+	int width, i, ret = 0;
+	u32 fifo_size;
+
+	if (!host->pdata) {
+		host->pdata = dw_mci_parse_dt(host);
+		if (IS_ERR(host->pdata))
+			return dev_err_probe(host->dev, PTR_ERR(host->pdata),
+					     "platform data not available\n");
+	}
+
+	host->biu_clk = devm_clk_get(host->dev, "biu");
+	if (IS_ERR(host->biu_clk)) {
+		dev_dbg(host->dev, "biu clock not available\n");
+	} else {
+		ret = clk_prepare_enable(host->biu_clk);
+		if (ret) {
+			dev_err(host->dev, "failed to enable biu clock\n");
+			return ret;
+		}
+	}
+
+	host->ciu_clk = devm_clk_get(host->dev, "ciu");
+	if (IS_ERR(host->ciu_clk)) {
+		dev_dbg(host->dev, "ciu clock not available\n");
+		host->bus_hz = host->pdata->bus_hz;
+	} else {
+		ret = clk_prepare_enable(host->ciu_clk);
+		if (ret) {
+			dev_err(host->dev, "failed to enable ciu clock\n");
+			goto err_clk_biu;
+		}
+
+		if (host->pdata->bus_hz) {
+			ret = clk_set_rate(host->ciu_clk, host->pdata->bus_hz);
+			if (ret)
+				dev_warn(host->dev,
+					 "Unable to set bus rate to %uHz\n",
+					 host->pdata->bus_hz);
+		}
+		host->bus_hz = clk_get_rate(host->ciu_clk);
+	}
+
+	if (!host->bus_hz) {
+		dev_err(host->dev,
+			"Platform data must supply bus speed\n");
+		ret = -ENODEV;
+		goto err_clk_ciu;
+	}
+
+	if (!IS_ERR(host->pdata->rstc)) {
+		reset_control_assert(host->pdata->rstc);
+		usleep_range(10, 50);
+		reset_control_deassert(host->pdata->rstc);
+	}
+
+	if (drv_data && drv_data->init) {
+		ret = drv_data->init(host);
+		if (ret) {
+			dev_err(host->dev,
+				"implementation specific init failed\n");
+			goto err_clk_ciu;
+		}
+	}
+
+	timer_setup(&host->cmd11_timer, dw_mci_cmd11_timer, 0);
+	timer_setup(&host->cto_timer, dw_mci_cto_timer, 0);
+	timer_setup(&host->dto_timer, dw_mci_dto_timer, 0);
+
+	spin_lock_init(&host->lock);
+	spin_lock_init(&host->irq_lock);
+	INIT_LIST_HEAD(&host->queue);
+
+	/*
+	 * Get the host data width - this assumes that HCON has been set with
+	 * the correct values.
+	 */
+	i = SDMMC_GET_HDATA_WIDTH(mci_readl(host, HCON));
+	if (!i) {
+		host->push_data = dw_mci_push_data16;
+		host->pull_data = dw_mci_pull_data16;
+		width = 16;
+		host->data_shift = 1;
+	} else if (i == 2) {
+		host->push_data = dw_mci_push_data64;
+		host->pull_data = dw_mci_pull_data64;
+		width = 64;
+		host->data_shift = 3;
+	} else {
+		/* Check for a reserved value, and warn if it is */
+		WARN((i != 1),
+		     "HCON reports a reserved host data width!\n"
+		     "Defaulting to 32-bit access.\n");
+		host->push_data = dw_mci_push_data32;
+		host->pull_data = dw_mci_pull_data32;
+		width = 32;
+		host->data_shift = 2;
+	}
+
+	/* Reset all blocks */
+	if (!dw_mci_ctrl_reset(host, SDMMC_CTRL_ALL_RESET_FLAGS)) {
+		ret = -ENODEV;
+		goto err_clk_ciu;
+	}
+
+	host->dma_ops = host->pdata->dma_ops;
+	dw_mci_init_dma(host);
+
+	/* Clear the interrupts for the host controller */
+	mci_writel(host, RINTSTS, 0xFFFFFFFF);
+	mci_writel(host, INTMASK, 0); /* disable all mmc interrupt first */
+
+	/* Put in max timeout */
+	mci_writel(host, TMOUT, 0xFFFFFFFF);
+
+	/*
+	 * FIFO threshold settings  RxMark  = fifo_size / 2 - 1,
+	 *                          Tx Mark = fifo_size / 2 DMA Size = 8
+	 */
+	if (!host->pdata->fifo_depth) {
+		/*
+		 * Power-on value of RX_WMark is FIFO_DEPTH-1, but this may
+		 * have been overwritten by the bootloader, just like we're
+		 * about to do, so if you know the value for your hardware, you
+		 * should put it in the platform data.
+		 */
+		fifo_size = mci_readl(host, FIFOTH);
+		fifo_size = 1 + ((fifo_size >> 16) & 0xfff);
+	} else {
+		fifo_size = host->pdata->fifo_depth;
+	}
+	host->fifo_depth = fifo_size;
+	host->fifoth_val =
+		SDMMC_SET_FIFOTH(0x2, fifo_size / 2 - 1, fifo_size / 2);
+	mci_writel(host, FIFOTH, host->fifoth_val);
+
+	/* disable clock to CIU */
+	mci_writel(host, CLKENA, 0);
+	mci_writel(host, CLKSRC, 0);
+
+	/*
+	 * In 2.40a spec, Data offset is changed.
+	 * Need to check the version-id and set data-offset for DATA register.
+	 */
+	host->verid = SDMMC_GET_VERID(mci_readl(host, VERID));
+	dev_info(host->dev, "Version ID is %04x\n", host->verid);
+
+	if (host->data_addr_override)
+		host->fifo_reg = host->regs + host->data_addr_override;
+	else if (host->verid < DW_MMC_240A)
+		host->fifo_reg = host->regs + DATA_OFFSET;
+	else
+		host->fifo_reg = host->regs + DATA_240A_OFFSET;
+
+	tasklet_init(&host->tasklet, dw_mci_tasklet_func, (unsigned long)host);
+	ret = devm_request_irq(host->dev, host->irq, dw_mci_interrupt,
+			       host->irq_flags, "dw-mci", host);
+	if (ret)
+		goto err_dmaunmap;
+
+	/*
+	 * Enable interrupts for command done, data over, data empty,
+	 * receive ready and error such as transmit, receive timeout, crc error
+	 */
+	mci_writel(host, INTMASK, SDMMC_INT_CMD_DONE | SDMMC_INT_DATA_OVER |
+		   SDMMC_INT_TXDR | SDMMC_INT_RXDR |
+		   DW_MCI_ERROR_FLAGS);
+	/* Enable mci interrupt */
+	mci_writel(host, CTRL, SDMMC_CTRL_INT_ENABLE);
+
+	dev_info(host->dev,
+		 "DW MMC controller at irq %d,%d bit host data width,%u deep fifo\n",
+		 host->irq, width, fifo_size);
+
+	/* We need at least one slot to succeed */
+	ret = dw_mci_init_slot(host);
+	if (ret) {
+		dev_dbg(host->dev, "slot %d init failed\n", i);
+		goto err_dmaunmap;
+	}
+
+	/* Now that slots are all setup, we can enable card detect */
+	dw_mci_enable_cd(host);
+
+	return 0;
+
+err_dmaunmap:
+	if (host->use_dma && host->dma_ops->exit)
+		host->dma_ops->exit(host);
+
+	if (!IS_ERR(host->pdata->rstc))
+		reset_control_assert(host->pdata->rstc);
+
+err_clk_ciu:
+	clk_disable_unprepare(host->ciu_clk);
+
+err_clk_biu:
+	clk_disable_unprepare(host->biu_clk);
+
+	return ret;
+}
+EXPORT_SYMBOL(dw_mci_probe);
+
+void dw_mci_remove(struct dw_mci *host)
+{
+	dev_dbg(host->dev, "remove slot\n");
+	if (host->slot)
+		dw_mci_cleanup_slot(host->slot);
+
+	mci_writel(host, RINTSTS, 0xFFFFFFFF);
+	mci_writel(host, INTMASK, 0); /* disable all mmc interrupt first */
+
+	/* disable clock to CIU */
+	mci_writel(host, CLKENA, 0);
+	mci_writel(host, CLKSRC, 0);
+
+	if (host->use_dma && host->dma_ops->exit)
+		host->dma_ops->exit(host);
+
+	if (!IS_ERR(host->pdata->rstc))
+		reset_control_assert(host->pdata->rstc);
+
+	clk_disable_unprepare(host->ciu_clk);
+	clk_disable_unprepare(host->biu_clk);
+}
+EXPORT_SYMBOL(dw_mci_remove);
+
+
+
+#ifdef CONFIG_PM
+int dw_mci_runtime_suspend(struct device *dev)
+{
+	struct dw_mci *host = dev_get_drvdata(dev);
+
+	if (host->use_dma && host->dma_ops->exit)
+		host->dma_ops->exit(host);
+
+	clk_disable_unprepare(host->ciu_clk);
+
+	if (host->slot &&
+	    (mmc_can_gpio_cd(host->slot->mmc) ||
+	     !mmc_card_is_removable(host->slot->mmc)))
+		clk_disable_unprepare(host->biu_clk);
+
+	return 0;
+}
+EXPORT_SYMBOL(dw_mci_runtime_suspend);
+
+int dw_mci_runtime_resume(struct device *dev)
+{
+	int ret = 0;
+	struct dw_mci *host = dev_get_drvdata(dev);
+
+	if (host->slot &&
+	    (mmc_can_gpio_cd(host->slot->mmc) ||
+	     !mmc_card_is_removable(host->slot->mmc))) {
+		ret = clk_prepare_enable(host->biu_clk);
+		if (ret)
+			return ret;
+	}
+
+	ret = clk_prepare_enable(host->ciu_clk);
+	if (ret)
+		goto err;
+
+	if (!dw_mci_ctrl_reset(host, SDMMC_CTRL_ALL_RESET_FLAGS)) {
+		clk_disable_unprepare(host->ciu_clk);
+		ret = -ENODEV;
+		goto err;
+	}
+
+	if (host->use_dma && host->dma_ops->init)
+		host->dma_ops->init(host);
+
+	/*
+	 * Restore the initial value at FIFOTH register
+	 * And Invalidate the prev_blksz with zero
+	 */
+	mci_writel(host, FIFOTH, host->fifoth_val);
+	host->prev_blksz = 0;
+
+	/* Put in max timeout */
+	mci_writel(host, TMOUT, 0xFFFFFFFF);
+
+	mci_writel(host, RINTSTS, 0xFFFFFFFF);
+	mci_writel(host, INTMASK, SDMMC_INT_CMD_DONE | SDMMC_INT_DATA_OVER |
+		   SDMMC_INT_TXDR | SDMMC_INT_RXDR |
+		   DW_MCI_ERROR_FLAGS);
+	mci_writel(host, CTRL, SDMMC_CTRL_INT_ENABLE);
+
+
+	if (host->slot->mmc->pm_flags & MMC_PM_KEEP_POWER)
+		dw_mci_set_ios(host->slot->mmc, &host->slot->mmc->ios);
+
+	/* Force setup bus to guarantee available clock output */
+	dw_mci_setup_bus(host->slot, true);
+
+	/* Re-enable SDIO interrupts. */
+	if (sdio_irq_claimed(host->slot->mmc))
+		__dw_mci_enable_sdio_irq(host->slot, 1);
+
+	/* Now that slots are all setup, we can enable card detect */
+	dw_mci_enable_cd(host);
+
+	return 0;
+
+err:
+	if (host->slot &&
+	    (mmc_can_gpio_cd(host->slot->mmc) ||
+	     !mmc_card_is_removable(host->slot->mmc)))
+		clk_disable_unprepare(host->biu_clk);
+
+	return ret;
+}
+EXPORT_SYMBOL(dw_mci_runtime_resume);
+#endif /* CONFIG_PM */
+
+static int __init dw_mci_init(void)
+{
+	pr_info("Synopsys Designware Multimedia Card Interface Driver\n");
+	return 0;
+}
+
+static void __exit dw_mci_exit(void)
+{
+}
+
+module_init(dw_mci_init);
+module_exit(dw_mci_exit);
+
+MODULE_DESCRIPTION("DW Multimedia Card Interface driver");
+MODULE_AUTHOR("NXP Semiconductor VietNam");
+MODULE_AUTHOR("Imagination Technologies Ltd");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/dw_mmc.h b/drivers/mmc/host/dw_mmc.h
new file mode 100644
index 000000000..da5923a92
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc.h
@@ -0,0 +1,567 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Synopsys DesignWare Multimedia Card Interface driver
+ *  (Based on NXP driver for lpc 31xx)
+ *
+ * Copyright (C) 2009 NXP Semiconductors
+ * Copyright (C) 2009, 2010 Imagination Technologies Ltd.
+ */
+
+#ifndef _DW_MMC_H_
+#define _DW_MMC_H_
+
+#include <linux/scatterlist.h>
+#include <linux/mmc/core.h>
+#include <linux/dmaengine.h>
+#include <linux/reset.h>
+#include <linux/interrupt.h>
+
+enum dw_mci_state {
+	STATE_IDLE = 0,
+	STATE_SENDING_CMD,
+	STATE_SENDING_DATA,
+	STATE_DATA_BUSY,
+	STATE_SENDING_STOP,
+	STATE_DATA_ERROR,
+	STATE_SENDING_CMD11,
+	STATE_WAITING_CMD11_DONE,
+};
+
+enum {
+	EVENT_CMD_COMPLETE = 0,
+	EVENT_XFER_COMPLETE,
+	EVENT_DATA_COMPLETE,
+	EVENT_DATA_ERROR,
+};
+
+enum dw_mci_cookie {
+	COOKIE_UNMAPPED,
+	COOKIE_PRE_MAPPED,	/* mapped by pre_req() of dwmmc */
+	COOKIE_MAPPED,		/* mapped by prepare_data() of dwmmc */
+};
+
+struct mmc_data;
+
+enum {
+	TRANS_MODE_PIO = 0,
+	TRANS_MODE_IDMAC,
+	TRANS_MODE_EDMAC
+};
+
+struct dw_mci_dma_slave {
+	struct dma_chan *ch;
+	enum dma_transfer_direction direction;
+};
+
+/**
+ * struct dw_mci - MMC controller state shared between all slots
+ * @lock: Spinlock protecting the queue and associated data.
+ * @irq_lock: Spinlock protecting the INTMASK setting.
+ * @regs: Pointer to MMIO registers.
+ * @fifo_reg: Pointer to MMIO registers for data FIFO
+ * @sg: Scatterlist entry currently being processed by PIO code, if any.
+ * @sg_miter: PIO mapping scatterlist iterator.
+ * @mrq: The request currently being processed on @slot,
+ *	or NULL if the controller is idle.
+ * @cmd: The command currently being sent to the card, or NULL.
+ * @data: The data currently being transferred, or NULL if no data
+ *	transfer is in progress.
+ * @stop_abort: The command currently prepared for stoping transfer.
+ * @prev_blksz: The former transfer blksz record.
+ * @timing: Record of current ios timing.
+ * @use_dma: Which DMA channel is in use for the current transfer, zero
+ *	denotes PIO mode.
+ * @using_dma: Whether DMA is in use for the current transfer.
+ * @dma_64bit_address: Whether DMA supports 64-bit address mode or not.
+ * @sg_dma: Bus address of DMA buffer.
+ * @sg_cpu: Virtual address of DMA buffer.
+ * @dma_ops: Pointer to platform-specific DMA callbacks.
+ * @cmd_status: Snapshot of SR taken upon completion of the current
+ * @ring_size: Buffer size for idma descriptors.
+ *	command. Only valid when EVENT_CMD_COMPLETE is pending.
+ * @dms: structure of slave-dma private data.
+ * @phy_regs: physical address of controller's register map
+ * @data_status: Snapshot of SR taken upon completion of the current
+ *	data transfer. Only valid when EVENT_DATA_COMPLETE or
+ *	EVENT_DATA_ERROR is pending.
+ * @stop_cmdr: Value to be loaded into CMDR when the stop command is
+ *	to be sent.
+ * @dir_status: Direction of current transfer.
+ * @tasklet: Tasklet running the request state machine.
+ * @pending_events: Bitmask of events flagged by the interrupt handler
+ *	to be processed by the tasklet.
+ * @completed_events: Bitmask of events which the state machine has
+ *	processed.
+ * @state: Tasklet state.
+ * @queue: List of slots waiting for access to the controller.
+ * @bus_hz: The rate of @mck in Hz. This forms the basis for MMC bus
+ *	rate and timeout calculations.
+ * @current_speed: Configured rate of the controller.
+ * @fifoth_val: The value of FIFOTH register.
+ * @verid: Denote Version ID.
+ * @dev: Device associated with the MMC controller.
+ * @pdata: Platform data associated with the MMC controller.
+ * @drv_data: Driver specific data for identified variant of the controller
+ * @priv: Implementation defined private data.
+ * @biu_clk: Pointer to bus interface unit clock instance.
+ * @ciu_clk: Pointer to card interface unit clock instance.
+ * @slot: Slots sharing this MMC controller.
+ * @fifo_depth: depth of FIFO.
+ * @data_addr_override: override fifo reg offset with this value.
+ * @wm_aligned: force fifo watermark equal with data length in PIO mode.
+ *	Set as true if alignment is needed.
+ * @data_shift: log2 of FIFO item size.
+ * @part_buf_start: Start index in part_buf.
+ * @part_buf_count: Bytes of partial data in part_buf.
+ * @part_buf: Simple buffer for partial fifo reads/writes.
+ * @push_data: Pointer to FIFO push function.
+ * @pull_data: Pointer to FIFO pull function.
+ * @vqmmc_enabled: Status of vqmmc, should be true or false.
+ * @irq_flags: The flags to be passed to request_irq.
+ * @irq: The irq value to be passed to request_irq.
+ * @sdio_id0: Number of slot0 in the SDIO interrupt registers.
+ * @cmd11_timer: Timer for SD3.0 voltage switch over scheme.
+ * @cto_timer: Timer for broken command transfer over scheme.
+ * @dto_timer: Timer for broken data transfer over scheme.
+ *
+ * Locking
+ * =======
+ *
+ * @lock is a softirq-safe spinlock protecting @queue as well as
+ * @slot, @mrq and @state. These must always be updated
+ * at the same time while holding @lock.
+ * The @mrq field of struct dw_mci_slot is also protected by @lock,
+ * and must always be written at the same time as the slot is added to
+ * @queue.
+ *
+ * @irq_lock is an irq-safe spinlock protecting the INTMASK register
+ * to allow the interrupt handler to modify it directly.  Held for only long
+ * enough to read-modify-write INTMASK and no other locks are grabbed when
+ * holding this one.
+ *
+ * @pending_events and @completed_events are accessed using atomic bit
+ * operations, so they don't need any locking.
+ *
+ * None of the fields touched by the interrupt handler need any
+ * locking. However, ordering is important: Before EVENT_DATA_ERROR or
+ * EVENT_DATA_COMPLETE is set in @pending_events, all data-related
+ * interrupts must be disabled and @data_status updated with a
+ * snapshot of SR. Similarly, before EVENT_CMD_COMPLETE is set, the
+ * CMDRDY interrupt must be disabled and @cmd_status updated with a
+ * snapshot of SR, and before EVENT_XFER_COMPLETE can be set, the
+ * bytes_xfered field of @data must be written. This is ensured by
+ * using barriers.
+ */
+struct dw_mci {
+	spinlock_t		lock;
+	spinlock_t		irq_lock;
+	void __iomem		*regs;
+	void __iomem		*fifo_reg;
+	u32			data_addr_override;
+	bool			wm_aligned;
+
+	struct scatterlist	*sg;
+	struct sg_mapping_iter	sg_miter;
+
+	struct mmc_request	*mrq;
+	struct mmc_command	*cmd;
+	struct mmc_data		*data;
+	struct mmc_command	stop_abort;
+	unsigned int		prev_blksz;
+	unsigned char		timing;
+
+	/* DMA interface members*/
+	int			use_dma;
+	int			using_dma;
+	int			dma_64bit_address;
+
+	dma_addr_t		sg_dma;
+	void			*sg_cpu;
+	const struct dw_mci_dma_ops	*dma_ops;
+	/* For idmac */
+	unsigned int		ring_size;
+
+	/* For edmac */
+	struct dw_mci_dma_slave *dms;
+	/* Registers's physical base address */
+	resource_size_t		phy_regs;
+
+	u32			cmd_status;
+	u32			data_status;
+	u32			stop_cmdr;
+	u32			dir_status;
+	struct tasklet_struct	tasklet;
+	unsigned long		pending_events;
+	unsigned long		completed_events;
+	enum dw_mci_state	state;
+	struct list_head	queue;
+
+	u32			bus_hz;
+	u32			current_speed;
+	u32			fifoth_val;
+	u16			verid;
+	struct device		*dev;
+	struct dw_mci_board	*pdata;
+	const struct dw_mci_drv_data	*drv_data;
+	void			*priv;
+	struct clk		*biu_clk;
+	struct clk		*ciu_clk;
+	struct dw_mci_slot	*slot;
+
+	/* FIFO push and pull */
+	int			fifo_depth;
+	int			data_shift;
+	u8			part_buf_start;
+	u8			part_buf_count;
+	union {
+		u16		part_buf16;
+		u32		part_buf32;
+		u64		part_buf;
+	};
+	void (*push_data)(struct dw_mci *host, void *buf, int cnt);
+	void (*pull_data)(struct dw_mci *host, void *buf, int cnt);
+
+	bool			vqmmc_enabled;
+	unsigned long		irq_flags; /* IRQ flags */
+	int			irq;
+
+	int			sdio_id0;
+
+	struct timer_list       cmd11_timer;
+	struct timer_list       cto_timer;
+	struct timer_list       dto_timer;
+};
+
+/* DMA ops for Internal/External DMAC interface */
+struct dw_mci_dma_ops {
+	/* DMA Ops */
+	int (*init)(struct dw_mci *host);
+	int (*start)(struct dw_mci *host, unsigned int sg_len);
+	void (*complete)(void *host);
+	void (*stop)(struct dw_mci *host);
+	void (*cleanup)(struct dw_mci *host);
+	void (*exit)(struct dw_mci *host);
+};
+
+struct dma_pdata;
+
+/* Board platform data */
+struct dw_mci_board {
+	unsigned int bus_hz; /* Clock speed at the cclk_in pad */
+
+	u32 caps;	/* Capabilities */
+	u32 caps2;	/* More capabilities */
+	u32 pm_caps;	/* PM capabilities */
+	/*
+	 * Override fifo depth. If 0, autodetect it from the FIFOTH register,
+	 * but note that this may not be reliable after a bootloader has used
+	 * it.
+	 */
+	unsigned int fifo_depth;
+
+	/* delay in mS before detecting cards after interrupt */
+	u32 detect_delay_ms;
+
+	struct reset_control *rstc;
+	struct dw_mci_dma_ops *dma_ops;
+	struct dma_pdata *data;
+};
+
+#define DW_MMC_240A		0x240a
+#define DW_MMC_280A		0x280a
+
+#define SDMMC_CTRL		0x000
+#define SDMMC_PWREN		0x004
+#define SDMMC_CLKDIV		0x008
+#define SDMMC_CLKSRC		0x00c
+#define SDMMC_CLKENA		0x010
+#define SDMMC_TMOUT		0x014
+#define SDMMC_CTYPE		0x018
+#define SDMMC_BLKSIZ		0x01c
+#define SDMMC_BYTCNT		0x020
+#define SDMMC_INTMASK		0x024
+#define SDMMC_CMDARG		0x028
+#define SDMMC_CMD		0x02c
+#define SDMMC_RESP0		0x030
+#define SDMMC_RESP1		0x034
+#define SDMMC_RESP2		0x038
+#define SDMMC_RESP3		0x03c
+#define SDMMC_MINTSTS		0x040
+#define SDMMC_RINTSTS		0x044
+#define SDMMC_STATUS		0x048
+#define SDMMC_FIFOTH		0x04c
+#define SDMMC_CDETECT		0x050
+#define SDMMC_WRTPRT		0x054
+#define SDMMC_GPIO		0x058
+#define SDMMC_TCBCNT		0x05c
+#define SDMMC_TBBCNT		0x060
+#define SDMMC_DEBNCE		0x064
+#define SDMMC_USRID		0x068
+#define SDMMC_VERID		0x06c
+#define SDMMC_HCON		0x070
+#define SDMMC_UHS_REG		0x074
+#define SDMMC_RST_N		0x078
+#define SDMMC_BMOD		0x080
+#define SDMMC_PLDMND		0x084
+#define SDMMC_DBADDR		0x088
+#define SDMMC_IDSTS		0x08c
+#define SDMMC_IDINTEN		0x090
+#define SDMMC_DSCADDR		0x094
+#define SDMMC_BUFADDR		0x098
+#define SDMMC_CDTHRCTL		0x100
+#define SDMMC_UHS_REG_EXT	0x108
+#define SDMMC_DDR_REG		0x10c
+#define SDMMC_ENABLE_SHIFT	0x110
+#define SDMMC_DATA(x)		(x)
+/*
+ * Registers to support idmac 64-bit address mode
+ */
+#define SDMMC_DBADDRL		0x088
+#define SDMMC_DBADDRU		0x08c
+#define SDMMC_IDSTS64		0x090
+#define SDMMC_IDINTEN64		0x094
+#define SDMMC_DSCADDRL		0x098
+#define SDMMC_DSCADDRU		0x09c
+#define SDMMC_BUFADDRL		0x0A0
+#define SDMMC_BUFADDRU		0x0A4
+
+/*
+ * Data offset is difference according to Version
+ * Lower than 2.40a : data register offest is 0x100
+ */
+#define DATA_OFFSET		0x100
+#define DATA_240A_OFFSET	0x200
+
+/* shift bit field */
+#define _SBF(f, v)		((v) << (f))
+
+/* Control register defines */
+#define SDMMC_CTRL_USE_IDMAC		BIT(25)
+#define SDMMC_CTRL_CEATA_INT_EN		BIT(11)
+#define SDMMC_CTRL_SEND_AS_CCSD		BIT(10)
+#define SDMMC_CTRL_SEND_CCSD		BIT(9)
+#define SDMMC_CTRL_ABRT_READ_DATA	BIT(8)
+#define SDMMC_CTRL_SEND_IRQ_RESP	BIT(7)
+#define SDMMC_CTRL_READ_WAIT		BIT(6)
+#define SDMMC_CTRL_DMA_ENABLE		BIT(5)
+#define SDMMC_CTRL_INT_ENABLE		BIT(4)
+#define SDMMC_CTRL_DMA_RESET		BIT(2)
+#define SDMMC_CTRL_FIFO_RESET		BIT(1)
+#define SDMMC_CTRL_RESET		BIT(0)
+/* Clock Enable register defines */
+#define SDMMC_CLKEN_LOW_PWR		BIT(16)
+#define SDMMC_CLKEN_ENABLE		BIT(0)
+/* time-out register defines */
+#define SDMMC_TMOUT_DATA(n)		_SBF(8, (n))
+#define SDMMC_TMOUT_DATA_MSK		0xFFFFFF00
+#define SDMMC_TMOUT_RESP(n)		((n) & 0xFF)
+#define SDMMC_TMOUT_RESP_MSK		0xFF
+/* card-type register defines */
+#define SDMMC_CTYPE_8BIT		BIT(16)
+#define SDMMC_CTYPE_4BIT		BIT(0)
+#define SDMMC_CTYPE_1BIT		0
+/* Interrupt status & mask register defines */
+#define SDMMC_INT_SDIO(n)		BIT(16 + (n))
+#define SDMMC_INT_EBE			BIT(15)
+#define SDMMC_INT_ACD			BIT(14)
+#define SDMMC_INT_SBE			BIT(13)
+#define SDMMC_INT_HLE			BIT(12)
+#define SDMMC_INT_FRUN			BIT(11)
+#define SDMMC_INT_HTO			BIT(10)
+#define SDMMC_INT_VOLT_SWITCH		BIT(10) /* overloads bit 10! */
+#define SDMMC_INT_DRTO			BIT(9)
+#define SDMMC_INT_RTO			BIT(8)
+#define SDMMC_INT_DCRC			BIT(7)
+#define SDMMC_INT_RCRC			BIT(6)
+#define SDMMC_INT_RXDR			BIT(5)
+#define SDMMC_INT_TXDR			BIT(4)
+#define SDMMC_INT_DATA_OVER		BIT(3)
+#define SDMMC_INT_CMD_DONE		BIT(2)
+#define SDMMC_INT_RESP_ERR		BIT(1)
+#define SDMMC_INT_CD			BIT(0)
+#define SDMMC_INT_ERROR			0xbfc2
+/* Command register defines */
+#define SDMMC_CMD_START			BIT(31)
+#define SDMMC_CMD_USE_HOLD_REG	BIT(29)
+#define SDMMC_CMD_VOLT_SWITCH		BIT(28)
+#define SDMMC_CMD_CCS_EXP		BIT(23)
+#define SDMMC_CMD_CEATA_RD		BIT(22)
+#define SDMMC_CMD_UPD_CLK		BIT(21)
+#define SDMMC_CMD_INIT			BIT(15)
+#define SDMMC_CMD_STOP			BIT(14)
+#define SDMMC_CMD_PRV_DAT_WAIT		BIT(13)
+#define SDMMC_CMD_SEND_STOP		BIT(12)
+#define SDMMC_CMD_STRM_MODE		BIT(11)
+#define SDMMC_CMD_DAT_WR		BIT(10)
+#define SDMMC_CMD_DAT_EXP		BIT(9)
+#define SDMMC_CMD_RESP_CRC		BIT(8)
+#define SDMMC_CMD_RESP_LONG		BIT(7)
+#define SDMMC_CMD_RESP_EXP		BIT(6)
+#define SDMMC_CMD_INDX(n)		((n) & 0x1F)
+/* Status register defines */
+#define SDMMC_GET_FCNT(x)		(((x)>>17) & 0x1FFF)
+#define SDMMC_STATUS_DMA_REQ		BIT(31)
+#define SDMMC_STATUS_BUSY		BIT(9)
+/* FIFOTH register defines */
+#define SDMMC_SET_FIFOTH(m, r, t)	(((m) & 0x7) << 28 | \
+					 ((r) & 0xFFF) << 16 | \
+					 ((t) & 0xFFF))
+/* HCON register defines */
+#define DMA_INTERFACE_IDMA		(0x0)
+#define DMA_INTERFACE_DWDMA		(0x1)
+#define DMA_INTERFACE_GDMA		(0x2)
+#define DMA_INTERFACE_NODMA		(0x3)
+#define SDMMC_GET_TRANS_MODE(x)		(((x)>>16) & 0x3)
+#define SDMMC_GET_SLOT_NUM(x)		((((x)>>1) & 0x1F) + 1)
+#define SDMMC_GET_HDATA_WIDTH(x)	(((x)>>7) & 0x7)
+#define SDMMC_GET_ADDR_CONFIG(x)	(((x)>>27) & 0x1)
+/* Internal DMAC interrupt defines */
+#define SDMMC_IDMAC_INT_AI		BIT(9)
+#define SDMMC_IDMAC_INT_NI		BIT(8)
+#define SDMMC_IDMAC_INT_CES		BIT(5)
+#define SDMMC_IDMAC_INT_DU		BIT(4)
+#define SDMMC_IDMAC_INT_FBE		BIT(2)
+#define SDMMC_IDMAC_INT_RI		BIT(1)
+#define SDMMC_IDMAC_INT_TI		BIT(0)
+/* Internal DMAC bus mode bits */
+#define SDMMC_IDMAC_ENABLE		BIT(7)
+#define SDMMC_IDMAC_FB			BIT(1)
+#define SDMMC_IDMAC_SWRESET		BIT(0)
+/* H/W reset */
+#define SDMMC_RST_HWACTIVE		0x1
+/* Version ID register define */
+#define SDMMC_GET_VERID(x)		((x) & 0xFFFF)
+/* Card read threshold */
+#define SDMMC_SET_THLD(v, x)		(((v) & 0xFFF) << 16 | (x))
+#define SDMMC_CARD_WR_THR_EN		BIT(2)
+#define SDMMC_CARD_RD_THR_EN		BIT(0)
+/* UHS-1 register defines */
+#define SDMMC_UHS_DDR			BIT(16)
+#define SDMMC_UHS_18V			BIT(0)
+/* DDR register defines */
+#define SDMMC_DDR_HS400			BIT(31)
+/* Enable shift register defines */
+#define SDMMC_ENABLE_PHASE		BIT(0)
+/* All ctrl reset bits */
+#define SDMMC_CTRL_ALL_RESET_FLAGS \
+	(SDMMC_CTRL_RESET | SDMMC_CTRL_FIFO_RESET | SDMMC_CTRL_DMA_RESET)
+
+/* FIFO register access macros. These should not change the data endian-ness
+ * as they are written to memory to be dealt with by the upper layers
+ */
+#define mci_fifo_readw(__reg)	__raw_readw(__reg)
+#define mci_fifo_readl(__reg)	__raw_readl(__reg)
+#define mci_fifo_readq(__reg)	__raw_readq(__reg)
+
+#define mci_fifo_writew(__value, __reg)	__raw_writew(__reg, __value)
+#define mci_fifo_writel(__value, __reg)	__raw_writel(__reg, __value)
+#define mci_fifo_writeq(__value, __reg)	__raw_writeq(__reg, __value)
+
+/* Register access macros */
+#define mci_readl(dev, reg)			\
+	readl_relaxed((dev)->regs + SDMMC_##reg)
+#define mci_writel(dev, reg, value)			\
+	writel_relaxed((value), (dev)->regs + SDMMC_##reg)
+
+/* 16-bit FIFO access macros */
+#define mci_readw(dev, reg)			\
+	readw_relaxed((dev)->regs + SDMMC_##reg)
+#define mci_writew(dev, reg, value)			\
+	writew_relaxed((value), (dev)->regs + SDMMC_##reg)
+
+/* 64-bit FIFO access macros */
+#ifdef readq
+#define mci_readq(dev, reg)			\
+	readq_relaxed((dev)->regs + SDMMC_##reg)
+#define mci_writeq(dev, reg, value)			\
+	writeq_relaxed((value), (dev)->regs + SDMMC_##reg)
+#else
+/*
+ * Dummy readq implementation for architectures that don't define it.
+ *
+ * We would assume that none of these architectures would configure
+ * the IP block with a 64bit FIFO width, so this code will never be
+ * executed on those machines. Defining these macros here keeps the
+ * rest of the code free from ifdefs.
+ */
+#define mci_readq(dev, reg)			\
+	(*(volatile u64 __force *)((dev)->regs + SDMMC_##reg))
+#define mci_writeq(dev, reg, value)			\
+	(*(volatile u64 __force *)((dev)->regs + SDMMC_##reg) = (value))
+
+#define __raw_writeq(__value, __reg) \
+	(*(volatile u64 __force *)(__reg) = (__value))
+#define __raw_readq(__reg) (*(volatile u64 __force *)(__reg))
+#endif
+
+extern int dw_mci_probe(struct dw_mci *host);
+extern void dw_mci_remove(struct dw_mci *host);
+#ifdef CONFIG_PM
+extern int dw_mci_runtime_suspend(struct device *device);
+extern int dw_mci_runtime_resume(struct device *device);
+#endif
+
+/**
+ * struct dw_mci_slot - MMC slot state
+ * @mmc: The mmc_host representing this slot.
+ * @host: The MMC controller this slot is using.
+ * @ctype: Card type for this slot.
+ * @mrq: mmc_request currently being processed or waiting to be
+ *	processed, or NULL when the slot is idle.
+ * @queue_node: List node for placing this node in the @queue list of
+ *	&struct dw_mci.
+ * @clock: Clock rate configured by set_ios(). Protected by host->lock.
+ * @__clk_old: The last clock value that was requested from core.
+ *	Keeping track of this helps us to avoid spamming the console.
+ * @flags: Random state bits associated with the slot.
+ * @id: Number of this slot.
+ * @sdio_id: Number of this slot in the SDIO interrupt registers.
+ */
+struct dw_mci_slot {
+	struct mmc_host		*mmc;
+	struct dw_mci		*host;
+
+	u32			ctype;
+
+	struct mmc_request	*mrq;
+	struct list_head	queue_node;
+
+	unsigned int		clock;
+	unsigned int		__clk_old;
+
+	unsigned long		flags;
+#define DW_MMC_CARD_PRESENT	0
+#define DW_MMC_CARD_NEED_INIT	1
+#define DW_MMC_CARD_NO_LOW_PWR	2
+#define DW_MMC_CARD_NO_USE_HOLD 3
+#define DW_MMC_CARD_NEEDS_POLL	4
+	int			id;
+	int			sdio_id;
+};
+
+/**
+ * dw_mci driver data - dw-mshc implementation specific driver data.
+ * @caps: mmc subsystem specified capabilities of the controller(s).
+ * @num_caps: number of capabilities specified by @caps.
+ * @init: early implementation specific initialization.
+ * @set_ios: handle bus specific extensions.
+ * @parse_dt: parse implementation specific device tree properties.
+ * @execute_tuning: implementation specific tuning procedure.
+ *
+ * Provide controller implementation specific extensions. The usage of this
+ * data structure is fully optional and usage of each member in this structure
+ * is optional as well.
+ */
+struct dw_mci_drv_data {
+	unsigned long	*caps;
+	u32		num_caps;
+	int		(*init)(struct dw_mci *host);
+	void		(*set_ios)(struct dw_mci *host, struct mmc_ios *ios);
+	int		(*parse_dt)(struct dw_mci *host);
+	int		(*execute_tuning)(struct dw_mci_slot *slot, u32 opcode);
+	int		(*prepare_hs400_tuning)(struct dw_mci *host,
+						struct mmc_ios *ios);
+	int		(*switch_voltage)(struct mmc_host *mmc,
+					  struct mmc_ios *ios);
+};
+#endif /* _DW_MMC_H_ */
diff --git a/drivers/mmc/host/jz4740_mmc.c b/drivers/mmc/host/jz4740_mmc.c
new file mode 100644
index 000000000..62d00232f
--- /dev/null
+++ b/drivers/mmc/host/jz4740_mmc.c
@@ -0,0 +1,1167 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ *  Copyright (C) 2009-2010, Lars-Peter Clausen <lars@metafoo.de>
+ *  Copyright (C) 2013, Imagination Technologies
+ *
+ *  JZ4740 SD/MMC controller driver
+ */
+
+#include <linux/bitops.h>
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/err.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/irq.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/platform_device.h>
+#include <linux/scatterlist.h>
+
+#include <asm/cacheflush.h>
+
+#define JZ_REG_MMC_STRPCL	0x00
+#define JZ_REG_MMC_STATUS	0x04
+#define JZ_REG_MMC_CLKRT	0x08
+#define JZ_REG_MMC_CMDAT	0x0C
+#define JZ_REG_MMC_RESTO	0x10
+#define JZ_REG_MMC_RDTO		0x14
+#define JZ_REG_MMC_BLKLEN	0x18
+#define JZ_REG_MMC_NOB		0x1C
+#define JZ_REG_MMC_SNOB		0x20
+#define JZ_REG_MMC_IMASK	0x24
+#define JZ_REG_MMC_IREG		0x28
+#define JZ_REG_MMC_CMD		0x2C
+#define JZ_REG_MMC_ARG		0x30
+#define JZ_REG_MMC_RESP_FIFO	0x34
+#define JZ_REG_MMC_RXFIFO	0x38
+#define JZ_REG_MMC_TXFIFO	0x3C
+#define JZ_REG_MMC_LPM		0x40
+#define JZ_REG_MMC_DMAC		0x44
+
+#define JZ_MMC_STRPCL_EXIT_MULTIPLE BIT(7)
+#define JZ_MMC_STRPCL_EXIT_TRANSFER BIT(6)
+#define JZ_MMC_STRPCL_START_READWAIT BIT(5)
+#define JZ_MMC_STRPCL_STOP_READWAIT BIT(4)
+#define JZ_MMC_STRPCL_RESET BIT(3)
+#define JZ_MMC_STRPCL_START_OP BIT(2)
+#define JZ_MMC_STRPCL_CLOCK_CONTROL (BIT(1) | BIT(0))
+#define JZ_MMC_STRPCL_CLOCK_STOP BIT(0)
+#define JZ_MMC_STRPCL_CLOCK_START BIT(1)
+
+
+#define JZ_MMC_STATUS_IS_RESETTING BIT(15)
+#define JZ_MMC_STATUS_SDIO_INT_ACTIVE BIT(14)
+#define JZ_MMC_STATUS_PRG_DONE BIT(13)
+#define JZ_MMC_STATUS_DATA_TRAN_DONE BIT(12)
+#define JZ_MMC_STATUS_END_CMD_RES BIT(11)
+#define JZ_MMC_STATUS_DATA_FIFO_AFULL BIT(10)
+#define JZ_MMC_STATUS_IS_READWAIT BIT(9)
+#define JZ_MMC_STATUS_CLK_EN BIT(8)
+#define JZ_MMC_STATUS_DATA_FIFO_FULL BIT(7)
+#define JZ_MMC_STATUS_DATA_FIFO_EMPTY BIT(6)
+#define JZ_MMC_STATUS_CRC_RES_ERR BIT(5)
+#define JZ_MMC_STATUS_CRC_READ_ERROR BIT(4)
+#define JZ_MMC_STATUS_TIMEOUT_WRITE BIT(3)
+#define JZ_MMC_STATUS_CRC_WRITE_ERROR BIT(2)
+#define JZ_MMC_STATUS_TIMEOUT_RES BIT(1)
+#define JZ_MMC_STATUS_TIMEOUT_READ BIT(0)
+
+#define JZ_MMC_STATUS_READ_ERROR_MASK (BIT(4) | BIT(0))
+#define JZ_MMC_STATUS_WRITE_ERROR_MASK (BIT(3) | BIT(2))
+
+
+#define JZ_MMC_CMDAT_IO_ABORT BIT(11)
+#define JZ_MMC_CMDAT_BUS_WIDTH_4BIT BIT(10)
+#define JZ_MMC_CMDAT_BUS_WIDTH_8BIT (BIT(10) | BIT(9))
+#define	JZ_MMC_CMDAT_BUS_WIDTH_MASK (BIT(10) | BIT(9))
+#define JZ_MMC_CMDAT_DMA_EN BIT(8)
+#define JZ_MMC_CMDAT_INIT BIT(7)
+#define JZ_MMC_CMDAT_BUSY BIT(6)
+#define JZ_MMC_CMDAT_STREAM BIT(5)
+#define JZ_MMC_CMDAT_WRITE BIT(4)
+#define JZ_MMC_CMDAT_DATA_EN BIT(3)
+#define JZ_MMC_CMDAT_RESPONSE_FORMAT (BIT(2) | BIT(1) | BIT(0))
+#define JZ_MMC_CMDAT_RSP_R1 1
+#define JZ_MMC_CMDAT_RSP_R2 2
+#define JZ_MMC_CMDAT_RSP_R3 3
+
+#define JZ_MMC_IRQ_SDIO BIT(7)
+#define JZ_MMC_IRQ_TXFIFO_WR_REQ BIT(6)
+#define JZ_MMC_IRQ_RXFIFO_RD_REQ BIT(5)
+#define JZ_MMC_IRQ_END_CMD_RES BIT(2)
+#define JZ_MMC_IRQ_PRG_DONE BIT(1)
+#define JZ_MMC_IRQ_DATA_TRAN_DONE BIT(0)
+
+#define JZ_MMC_DMAC_DMA_SEL BIT(1)
+#define JZ_MMC_DMAC_DMA_EN BIT(0)
+
+#define	JZ_MMC_LPM_DRV_RISING BIT(31)
+#define	JZ_MMC_LPM_DRV_RISING_QTR_PHASE_DLY BIT(31)
+#define	JZ_MMC_LPM_DRV_RISING_1NS_DLY BIT(30)
+#define	JZ_MMC_LPM_SMP_RISING_QTR_OR_HALF_PHASE_DLY BIT(29)
+#define	JZ_MMC_LPM_LOW_POWER_MODE_EN BIT(0)
+
+#define JZ_MMC_CLK_RATE 24000000
+#define JZ_MMC_REQ_TIMEOUT_MS 5000
+
+enum jz4740_mmc_version {
+	JZ_MMC_JZ4740,
+	JZ_MMC_JZ4725B,
+	JZ_MMC_JZ4760,
+	JZ_MMC_JZ4780,
+	JZ_MMC_X1000,
+};
+
+enum jz4740_mmc_state {
+	JZ4740_MMC_STATE_READ_RESPONSE,
+	JZ4740_MMC_STATE_TRANSFER_DATA,
+	JZ4740_MMC_STATE_SEND_STOP,
+	JZ4740_MMC_STATE_DONE,
+};
+
+/*
+ * The MMC core allows to prepare a mmc_request while another mmc_request
+ * is in-flight. This is used via the pre_req/post_req hooks.
+ * This driver uses the pre_req/post_req hooks to map/unmap the mmc_request.
+ * Following what other drivers do (sdhci, dw_mmc) we use the following cookie
+ * flags to keep track of the mmc_request mapping state.
+ *
+ * COOKIE_UNMAPPED: the request is not mapped.
+ * COOKIE_PREMAPPED: the request was mapped in pre_req,
+ * and should be unmapped in post_req.
+ * COOKIE_MAPPED: the request was mapped in the irq handler,
+ * and should be unmapped before mmc_request_done is called..
+ */
+enum jz4780_cookie {
+	COOKIE_UNMAPPED = 0,
+	COOKIE_PREMAPPED,
+	COOKIE_MAPPED,
+};
+
+struct jz4740_mmc_host {
+	struct mmc_host *mmc;
+	struct platform_device *pdev;
+	struct clk *clk;
+
+	enum jz4740_mmc_version version;
+
+	int irq;
+	int card_detect_irq;
+
+	void __iomem *base;
+	struct resource *mem_res;
+	struct mmc_request *req;
+	struct mmc_command *cmd;
+
+	unsigned long waiting;
+
+	uint32_t cmdat;
+
+	uint32_t irq_mask;
+
+	spinlock_t lock;
+
+	struct timer_list timeout_timer;
+	struct sg_mapping_iter miter;
+	enum jz4740_mmc_state state;
+
+	/* DMA support */
+	struct dma_chan *dma_rx;
+	struct dma_chan *dma_tx;
+	bool use_dma;
+
+/* The DMA trigger level is 8 words, that is to say, the DMA read
+ * trigger is when data words in MSC_RXFIFO is >= 8 and the DMA write
+ * trigger is when data words in MSC_TXFIFO is < 8.
+ */
+#define JZ4740_MMC_FIFO_HALF_SIZE 8
+};
+
+static void jz4740_mmc_write_irq_mask(struct jz4740_mmc_host *host,
+				      uint32_t val)
+{
+	if (host->version >= JZ_MMC_JZ4725B)
+		return writel(val, host->base + JZ_REG_MMC_IMASK);
+	else
+		return writew(val, host->base + JZ_REG_MMC_IMASK);
+}
+
+static void jz4740_mmc_write_irq_reg(struct jz4740_mmc_host *host,
+				     uint32_t val)
+{
+	if (host->version >= JZ_MMC_JZ4780)
+		writel(val, host->base + JZ_REG_MMC_IREG);
+	else
+		writew(val, host->base + JZ_REG_MMC_IREG);
+}
+
+static uint32_t jz4740_mmc_read_irq_reg(struct jz4740_mmc_host *host)
+{
+	if (host->version >= JZ_MMC_JZ4780)
+		return readl(host->base + JZ_REG_MMC_IREG);
+	else
+		return readw(host->base + JZ_REG_MMC_IREG);
+}
+
+/*----------------------------------------------------------------------------*/
+/* DMA infrastructure */
+
+static void jz4740_mmc_release_dma_channels(struct jz4740_mmc_host *host)
+{
+	if (!host->use_dma)
+		return;
+
+	dma_release_channel(host->dma_tx);
+	dma_release_channel(host->dma_rx);
+}
+
+static int jz4740_mmc_acquire_dma_channels(struct jz4740_mmc_host *host)
+{
+	host->dma_tx = dma_request_chan(mmc_dev(host->mmc), "tx");
+	if (IS_ERR(host->dma_tx)) {
+		dev_err(mmc_dev(host->mmc), "Failed to get dma_tx channel\n");
+		return PTR_ERR(host->dma_tx);
+	}
+
+	host->dma_rx = dma_request_chan(mmc_dev(host->mmc), "rx");
+	if (IS_ERR(host->dma_rx)) {
+		dev_err(mmc_dev(host->mmc), "Failed to get dma_rx channel\n");
+		dma_release_channel(host->dma_tx);
+		return PTR_ERR(host->dma_rx);
+	}
+
+	/*
+	 * Limit the maximum segment size in any SG entry according to
+	 * the parameters of the DMA engine device.
+	 */
+	if (host->dma_tx) {
+		struct device *dev = host->dma_tx->device->dev;
+		unsigned int max_seg_size = dma_get_max_seg_size(dev);
+
+		if (max_seg_size < host->mmc->max_seg_size)
+			host->mmc->max_seg_size = max_seg_size;
+	}
+
+	if (host->dma_rx) {
+		struct device *dev = host->dma_rx->device->dev;
+		unsigned int max_seg_size = dma_get_max_seg_size(dev);
+
+		if (max_seg_size < host->mmc->max_seg_size)
+			host->mmc->max_seg_size = max_seg_size;
+	}
+
+	return 0;
+}
+
+static inline struct dma_chan *jz4740_mmc_get_dma_chan(struct jz4740_mmc_host *host,
+						       struct mmc_data *data)
+{
+	return (data->flags & MMC_DATA_READ) ? host->dma_rx : host->dma_tx;
+}
+
+static void jz4740_mmc_dma_unmap(struct jz4740_mmc_host *host,
+				 struct mmc_data *data)
+{
+	struct dma_chan *chan = jz4740_mmc_get_dma_chan(host, data);
+	enum dma_data_direction dir = mmc_get_dma_dir(data);
+
+	dma_unmap_sg(chan->device->dev, data->sg, data->sg_len, dir);
+	data->host_cookie = COOKIE_UNMAPPED;
+}
+
+/* Prepares DMA data for current or next transfer.
+ * A request can be in-flight when this is called.
+ */
+static int jz4740_mmc_prepare_dma_data(struct jz4740_mmc_host *host,
+				       struct mmc_data *data,
+				       int cookie)
+{
+	struct dma_chan *chan = jz4740_mmc_get_dma_chan(host, data);
+	enum dma_data_direction dir = mmc_get_dma_dir(data);
+	int sg_count;
+
+	if (data->host_cookie == COOKIE_PREMAPPED)
+		return data->sg_count;
+
+	sg_count = dma_map_sg(chan->device->dev,
+			data->sg,
+			data->sg_len,
+			dir);
+
+	if (sg_count <= 0) {
+		dev_err(mmc_dev(host->mmc),
+			"Failed to map scatterlist for DMA operation\n");
+		return -EINVAL;
+	}
+
+	data->sg_count = sg_count;
+	data->host_cookie = cookie;
+
+	return data->sg_count;
+}
+
+static int jz4740_mmc_start_dma_transfer(struct jz4740_mmc_host *host,
+					 struct mmc_data *data)
+{
+	struct dma_chan *chan = jz4740_mmc_get_dma_chan(host, data);
+	struct dma_async_tx_descriptor *desc;
+	struct dma_slave_config conf = {
+		.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+		.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+		.src_maxburst = JZ4740_MMC_FIFO_HALF_SIZE,
+		.dst_maxburst = JZ4740_MMC_FIFO_HALF_SIZE,
+	};
+	int sg_count;
+
+	if (data->flags & MMC_DATA_WRITE) {
+		conf.direction = DMA_MEM_TO_DEV;
+		conf.dst_addr = host->mem_res->start + JZ_REG_MMC_TXFIFO;
+	} else {
+		conf.direction = DMA_DEV_TO_MEM;
+		conf.src_addr = host->mem_res->start + JZ_REG_MMC_RXFIFO;
+	}
+
+	sg_count = jz4740_mmc_prepare_dma_data(host, data, COOKIE_MAPPED);
+	if (sg_count < 0)
+		return sg_count;
+
+	dmaengine_slave_config(chan, &conf);
+	desc = dmaengine_prep_slave_sg(chan, data->sg, sg_count,
+			conf.direction,
+			DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc) {
+		dev_err(mmc_dev(host->mmc),
+			"Failed to allocate DMA %s descriptor",
+			 conf.direction == DMA_MEM_TO_DEV ? "TX" : "RX");
+		goto dma_unmap;
+	}
+
+	dmaengine_submit(desc);
+	dma_async_issue_pending(chan);
+
+	return 0;
+
+dma_unmap:
+	if (data->host_cookie == COOKIE_MAPPED)
+		jz4740_mmc_dma_unmap(host, data);
+	return -ENOMEM;
+}
+
+static void jz4740_mmc_pre_request(struct mmc_host *mmc,
+				   struct mmc_request *mrq)
+{
+	struct jz4740_mmc_host *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+
+	if (!host->use_dma)
+		return;
+
+	data->host_cookie = COOKIE_UNMAPPED;
+	if (jz4740_mmc_prepare_dma_data(host, data, COOKIE_PREMAPPED) < 0)
+		data->host_cookie = COOKIE_UNMAPPED;
+}
+
+static void jz4740_mmc_post_request(struct mmc_host *mmc,
+				    struct mmc_request *mrq,
+				    int err)
+{
+	struct jz4740_mmc_host *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+
+	if (data && data->host_cookie != COOKIE_UNMAPPED)
+		jz4740_mmc_dma_unmap(host, data);
+
+	if (err) {
+		struct dma_chan *chan = jz4740_mmc_get_dma_chan(host, data);
+
+		dmaengine_terminate_all(chan);
+	}
+}
+
+/*----------------------------------------------------------------------------*/
+
+static void jz4740_mmc_set_irq_enabled(struct jz4740_mmc_host *host,
+	unsigned int irq, bool enabled)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (enabled)
+		host->irq_mask &= ~irq;
+	else
+		host->irq_mask |= irq;
+
+	jz4740_mmc_write_irq_mask(host, host->irq_mask);
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void jz4740_mmc_clock_enable(struct jz4740_mmc_host *host,
+	bool start_transfer)
+{
+	uint16_t val = JZ_MMC_STRPCL_CLOCK_START;
+
+	if (start_transfer)
+		val |= JZ_MMC_STRPCL_START_OP;
+
+	writew(val, host->base + JZ_REG_MMC_STRPCL);
+}
+
+static void jz4740_mmc_clock_disable(struct jz4740_mmc_host *host)
+{
+	uint32_t status;
+	unsigned int timeout = 1000;
+
+	writew(JZ_MMC_STRPCL_CLOCK_STOP, host->base + JZ_REG_MMC_STRPCL);
+	do {
+		status = readl(host->base + JZ_REG_MMC_STATUS);
+	} while (status & JZ_MMC_STATUS_CLK_EN && --timeout);
+}
+
+static void jz4740_mmc_reset(struct jz4740_mmc_host *host)
+{
+	uint32_t status;
+	unsigned int timeout = 1000;
+
+	writew(JZ_MMC_STRPCL_RESET, host->base + JZ_REG_MMC_STRPCL);
+	udelay(10);
+	do {
+		status = readl(host->base + JZ_REG_MMC_STATUS);
+	} while (status & JZ_MMC_STATUS_IS_RESETTING && --timeout);
+}
+
+static void jz4740_mmc_request_done(struct jz4740_mmc_host *host)
+{
+	struct mmc_request *req;
+	struct mmc_data *data;
+
+	req = host->req;
+	data = req->data;
+	host->req = NULL;
+
+	if (data && data->host_cookie == COOKIE_MAPPED)
+		jz4740_mmc_dma_unmap(host, data);
+	mmc_request_done(host->mmc, req);
+}
+
+static unsigned int jz4740_mmc_poll_irq(struct jz4740_mmc_host *host,
+	unsigned int irq)
+{
+	unsigned int timeout = 0x800;
+	uint32_t status;
+
+	do {
+		status = jz4740_mmc_read_irq_reg(host);
+	} while (!(status & irq) && --timeout);
+
+	if (timeout == 0) {
+		set_bit(0, &host->waiting);
+		mod_timer(&host->timeout_timer,
+			  jiffies + msecs_to_jiffies(JZ_MMC_REQ_TIMEOUT_MS));
+		jz4740_mmc_set_irq_enabled(host, irq, true);
+		return true;
+	}
+
+	return false;
+}
+
+static void jz4740_mmc_transfer_check_state(struct jz4740_mmc_host *host,
+	struct mmc_data *data)
+{
+	int status;
+
+	status = readl(host->base + JZ_REG_MMC_STATUS);
+	if (status & JZ_MMC_STATUS_WRITE_ERROR_MASK) {
+		if (status & (JZ_MMC_STATUS_TIMEOUT_WRITE)) {
+			host->req->cmd->error = -ETIMEDOUT;
+			data->error = -ETIMEDOUT;
+		} else {
+			host->req->cmd->error = -EIO;
+			data->error = -EIO;
+		}
+	} else if (status & JZ_MMC_STATUS_READ_ERROR_MASK) {
+		if (status & (JZ_MMC_STATUS_TIMEOUT_READ)) {
+			host->req->cmd->error = -ETIMEDOUT;
+			data->error = -ETIMEDOUT;
+		} else {
+			host->req->cmd->error = -EIO;
+			data->error = -EIO;
+		}
+	}
+}
+
+static bool jz4740_mmc_write_data(struct jz4740_mmc_host *host,
+	struct mmc_data *data)
+{
+	struct sg_mapping_iter *miter = &host->miter;
+	void __iomem *fifo_addr = host->base + JZ_REG_MMC_TXFIFO;
+	uint32_t *buf;
+	bool timeout;
+	size_t i, j;
+
+	while (sg_miter_next(miter)) {
+		buf = miter->addr;
+		i = miter->length / 4;
+		j = i / 8;
+		i = i & 0x7;
+		while (j) {
+			timeout = jz4740_mmc_poll_irq(host, JZ_MMC_IRQ_TXFIFO_WR_REQ);
+			if (unlikely(timeout))
+				goto poll_timeout;
+
+			writel(buf[0], fifo_addr);
+			writel(buf[1], fifo_addr);
+			writel(buf[2], fifo_addr);
+			writel(buf[3], fifo_addr);
+			writel(buf[4], fifo_addr);
+			writel(buf[5], fifo_addr);
+			writel(buf[6], fifo_addr);
+			writel(buf[7], fifo_addr);
+			buf += 8;
+			--j;
+		}
+		if (unlikely(i)) {
+			timeout = jz4740_mmc_poll_irq(host, JZ_MMC_IRQ_TXFIFO_WR_REQ);
+			if (unlikely(timeout))
+				goto poll_timeout;
+
+			while (i) {
+				writel(*buf, fifo_addr);
+				++buf;
+				--i;
+			}
+		}
+		data->bytes_xfered += miter->length;
+	}
+	sg_miter_stop(miter);
+
+	return false;
+
+poll_timeout:
+	miter->consumed = (void *)buf - miter->addr;
+	data->bytes_xfered += miter->consumed;
+	sg_miter_stop(miter);
+
+	return true;
+}
+
+static bool jz4740_mmc_read_data(struct jz4740_mmc_host *host,
+				struct mmc_data *data)
+{
+	struct sg_mapping_iter *miter = &host->miter;
+	void __iomem *fifo_addr = host->base + JZ_REG_MMC_RXFIFO;
+	uint32_t *buf;
+	uint32_t d;
+	uint32_t status;
+	size_t i, j;
+	unsigned int timeout;
+
+	while (sg_miter_next(miter)) {
+		buf = miter->addr;
+		i = miter->length;
+		j = i / 32;
+		i = i & 0x1f;
+		while (j) {
+			timeout = jz4740_mmc_poll_irq(host, JZ_MMC_IRQ_RXFIFO_RD_REQ);
+			if (unlikely(timeout))
+				goto poll_timeout;
+
+			buf[0] = readl(fifo_addr);
+			buf[1] = readl(fifo_addr);
+			buf[2] = readl(fifo_addr);
+			buf[3] = readl(fifo_addr);
+			buf[4] = readl(fifo_addr);
+			buf[5] = readl(fifo_addr);
+			buf[6] = readl(fifo_addr);
+			buf[7] = readl(fifo_addr);
+
+			buf += 8;
+			--j;
+		}
+
+		if (unlikely(i)) {
+			timeout = jz4740_mmc_poll_irq(host, JZ_MMC_IRQ_RXFIFO_RD_REQ);
+			if (unlikely(timeout))
+				goto poll_timeout;
+
+			while (i >= 4) {
+				*buf++ = readl(fifo_addr);
+				i -= 4;
+			}
+			if (unlikely(i > 0)) {
+				d = readl(fifo_addr);
+				memcpy(buf, &d, i);
+			}
+		}
+		data->bytes_xfered += miter->length;
+
+		/* This can go away once MIPS implements
+		 * flush_kernel_dcache_page */
+		flush_dcache_page(miter->page);
+	}
+	sg_miter_stop(miter);
+
+	/* For whatever reason there is sometime one word more in the fifo then
+	 * requested */
+	timeout = 1000;
+	status = readl(host->base + JZ_REG_MMC_STATUS);
+	while (!(status & JZ_MMC_STATUS_DATA_FIFO_EMPTY) && --timeout) {
+		d = readl(fifo_addr);
+		status = readl(host->base + JZ_REG_MMC_STATUS);
+	}
+
+	return false;
+
+poll_timeout:
+	miter->consumed = (void *)buf - miter->addr;
+	data->bytes_xfered += miter->consumed;
+	sg_miter_stop(miter);
+
+	return true;
+}
+
+static void jz4740_mmc_timeout(struct timer_list *t)
+{
+	struct jz4740_mmc_host *host = from_timer(host, t, timeout_timer);
+
+	if (!test_and_clear_bit(0, &host->waiting))
+		return;
+
+	jz4740_mmc_set_irq_enabled(host, JZ_MMC_IRQ_END_CMD_RES, false);
+
+	host->req->cmd->error = -ETIMEDOUT;
+	jz4740_mmc_request_done(host);
+}
+
+static void jz4740_mmc_read_response(struct jz4740_mmc_host *host,
+	struct mmc_command *cmd)
+{
+	int i;
+	uint16_t tmp;
+	void __iomem *fifo_addr = host->base + JZ_REG_MMC_RESP_FIFO;
+
+	if (cmd->flags & MMC_RSP_136) {
+		tmp = readw(fifo_addr);
+		for (i = 0; i < 4; ++i) {
+			cmd->resp[i] = tmp << 24;
+			tmp = readw(fifo_addr);
+			cmd->resp[i] |= tmp << 8;
+			tmp = readw(fifo_addr);
+			cmd->resp[i] |= tmp >> 8;
+		}
+	} else {
+		cmd->resp[0] = readw(fifo_addr) << 24;
+		cmd->resp[0] |= readw(fifo_addr) << 8;
+		cmd->resp[0] |= readw(fifo_addr) & 0xff;
+	}
+}
+
+static void jz4740_mmc_send_command(struct jz4740_mmc_host *host,
+	struct mmc_command *cmd)
+{
+	uint32_t cmdat = host->cmdat;
+
+	host->cmdat &= ~JZ_MMC_CMDAT_INIT;
+	jz4740_mmc_clock_disable(host);
+
+	host->cmd = cmd;
+
+	if (cmd->flags & MMC_RSP_BUSY)
+		cmdat |= JZ_MMC_CMDAT_BUSY;
+
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_R1B:
+	case MMC_RSP_R1:
+		cmdat |= JZ_MMC_CMDAT_RSP_R1;
+		break;
+	case MMC_RSP_R2:
+		cmdat |= JZ_MMC_CMDAT_RSP_R2;
+		break;
+	case MMC_RSP_R3:
+		cmdat |= JZ_MMC_CMDAT_RSP_R3;
+		break;
+	default:
+		break;
+	}
+
+	if (cmd->data) {
+		cmdat |= JZ_MMC_CMDAT_DATA_EN;
+		if (cmd->data->flags & MMC_DATA_WRITE)
+			cmdat |= JZ_MMC_CMDAT_WRITE;
+		if (host->use_dma) {
+			/*
+			 * The 4780's MMC controller has integrated DMA ability
+			 * in addition to being able to use the external DMA
+			 * controller. It moves DMA control bits to a separate
+			 * register. The DMA_SEL bit chooses the external
+			 * controller over the integrated one. Earlier SoCs
+			 * can only use the external controller, and have a
+			 * single DMA enable bit in CMDAT.
+			 */
+			if (host->version >= JZ_MMC_JZ4780) {
+				writel(JZ_MMC_DMAC_DMA_EN | JZ_MMC_DMAC_DMA_SEL,
+				       host->base + JZ_REG_MMC_DMAC);
+			} else {
+				cmdat |= JZ_MMC_CMDAT_DMA_EN;
+			}
+		} else if (host->version >= JZ_MMC_JZ4780) {
+			writel(0, host->base + JZ_REG_MMC_DMAC);
+		}
+
+		writew(cmd->data->blksz, host->base + JZ_REG_MMC_BLKLEN);
+		writew(cmd->data->blocks, host->base + JZ_REG_MMC_NOB);
+	}
+
+	writeb(cmd->opcode, host->base + JZ_REG_MMC_CMD);
+	writel(cmd->arg, host->base + JZ_REG_MMC_ARG);
+	writel(cmdat, host->base + JZ_REG_MMC_CMDAT);
+
+	jz4740_mmc_clock_enable(host, 1);
+}
+
+static void jz_mmc_prepare_data_transfer(struct jz4740_mmc_host *host)
+{
+	struct mmc_command *cmd = host->req->cmd;
+	struct mmc_data *data = cmd->data;
+	int direction;
+
+	if (data->flags & MMC_DATA_READ)
+		direction = SG_MITER_TO_SG;
+	else
+		direction = SG_MITER_FROM_SG;
+
+	sg_miter_start(&host->miter, data->sg, data->sg_len, direction);
+}
+
+
+static irqreturn_t jz_mmc_irq_worker(int irq, void *devid)
+{
+	struct jz4740_mmc_host *host = (struct jz4740_mmc_host *)devid;
+	struct mmc_command *cmd = host->req->cmd;
+	struct mmc_request *req = host->req;
+	struct mmc_data *data = cmd->data;
+	bool timeout = false;
+
+	if (cmd->error)
+		host->state = JZ4740_MMC_STATE_DONE;
+
+	switch (host->state) {
+	case JZ4740_MMC_STATE_READ_RESPONSE:
+		if (cmd->flags & MMC_RSP_PRESENT)
+			jz4740_mmc_read_response(host, cmd);
+
+		if (!data)
+			break;
+
+		jz_mmc_prepare_data_transfer(host);
+		fallthrough;
+
+	case JZ4740_MMC_STATE_TRANSFER_DATA:
+		if (host->use_dma) {
+			/* Use DMA if enabled.
+			 * Data transfer direction is defined later by
+			 * relying on data flags in
+			 * jz4740_mmc_prepare_dma_data() and
+			 * jz4740_mmc_start_dma_transfer().
+			 */
+			timeout = jz4740_mmc_start_dma_transfer(host, data);
+			data->bytes_xfered = data->blocks * data->blksz;
+		} else if (data->flags & MMC_DATA_READ)
+			/* Use PIO if DMA is not enabled.
+			 * Data transfer direction was defined before
+			 * by relying on data flags in
+			 * jz_mmc_prepare_data_transfer().
+			 */
+			timeout = jz4740_mmc_read_data(host, data);
+		else
+			timeout = jz4740_mmc_write_data(host, data);
+
+		if (unlikely(timeout)) {
+			host->state = JZ4740_MMC_STATE_TRANSFER_DATA;
+			break;
+		}
+
+		jz4740_mmc_transfer_check_state(host, data);
+
+		timeout = jz4740_mmc_poll_irq(host, JZ_MMC_IRQ_DATA_TRAN_DONE);
+		if (unlikely(timeout)) {
+			host->state = JZ4740_MMC_STATE_SEND_STOP;
+			break;
+		}
+		jz4740_mmc_write_irq_reg(host, JZ_MMC_IRQ_DATA_TRAN_DONE);
+		fallthrough;
+
+	case JZ4740_MMC_STATE_SEND_STOP:
+		if (!req->stop)
+			break;
+
+		jz4740_mmc_send_command(host, req->stop);
+
+		if (mmc_resp_type(req->stop) & MMC_RSP_BUSY) {
+			timeout = jz4740_mmc_poll_irq(host,
+						      JZ_MMC_IRQ_PRG_DONE);
+			if (timeout) {
+				host->state = JZ4740_MMC_STATE_DONE;
+				break;
+			}
+		}
+	case JZ4740_MMC_STATE_DONE:
+		break;
+	}
+
+	if (!timeout)
+		jz4740_mmc_request_done(host);
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t jz_mmc_irq(int irq, void *devid)
+{
+	struct jz4740_mmc_host *host = devid;
+	struct mmc_command *cmd = host->cmd;
+	uint32_t irq_reg, status, tmp;
+
+	status = readl(host->base + JZ_REG_MMC_STATUS);
+	irq_reg = jz4740_mmc_read_irq_reg(host);
+
+	tmp = irq_reg;
+	irq_reg &= ~host->irq_mask;
+
+	tmp &= ~(JZ_MMC_IRQ_TXFIFO_WR_REQ | JZ_MMC_IRQ_RXFIFO_RD_REQ |
+		JZ_MMC_IRQ_PRG_DONE | JZ_MMC_IRQ_DATA_TRAN_DONE);
+
+	if (tmp != irq_reg)
+		jz4740_mmc_write_irq_reg(host, tmp & ~irq_reg);
+
+	if (irq_reg & JZ_MMC_IRQ_SDIO) {
+		jz4740_mmc_write_irq_reg(host, JZ_MMC_IRQ_SDIO);
+		mmc_signal_sdio_irq(host->mmc);
+		irq_reg &= ~JZ_MMC_IRQ_SDIO;
+	}
+
+	if (host->req && cmd && irq_reg) {
+		if (test_and_clear_bit(0, &host->waiting)) {
+			del_timer(&host->timeout_timer);
+
+			if (status & JZ_MMC_STATUS_TIMEOUT_RES) {
+				cmd->error = -ETIMEDOUT;
+			} else if (status & JZ_MMC_STATUS_CRC_RES_ERR) {
+				cmd->error = -EIO;
+			} else if (status & (JZ_MMC_STATUS_CRC_READ_ERROR |
+				    JZ_MMC_STATUS_CRC_WRITE_ERROR)) {
+				if (cmd->data)
+					cmd->data->error = -EIO;
+				cmd->error = -EIO;
+			}
+
+			jz4740_mmc_set_irq_enabled(host, irq_reg, false);
+			jz4740_mmc_write_irq_reg(host, irq_reg);
+
+			return IRQ_WAKE_THREAD;
+		}
+	}
+
+	return IRQ_HANDLED;
+}
+
+static int jz4740_mmc_set_clock_rate(struct jz4740_mmc_host *host, int rate)
+{
+	int div = 0;
+	int real_rate;
+
+	jz4740_mmc_clock_disable(host);
+	clk_set_rate(host->clk, host->mmc->f_max);
+
+	real_rate = clk_get_rate(host->clk);
+
+	while (real_rate > rate && div < 7) {
+		++div;
+		real_rate >>= 1;
+	}
+
+	writew(div, host->base + JZ_REG_MMC_CLKRT);
+
+	if (real_rate > 25000000) {
+		if (host->version >= JZ_MMC_X1000) {
+			writel(JZ_MMC_LPM_DRV_RISING_QTR_PHASE_DLY |
+				   JZ_MMC_LPM_SMP_RISING_QTR_OR_HALF_PHASE_DLY |
+				   JZ_MMC_LPM_LOW_POWER_MODE_EN,
+				   host->base + JZ_REG_MMC_LPM);
+		} else if (host->version >= JZ_MMC_JZ4760) {
+			writel(JZ_MMC_LPM_DRV_RISING |
+				   JZ_MMC_LPM_LOW_POWER_MODE_EN,
+				   host->base + JZ_REG_MMC_LPM);
+		} else if (host->version >= JZ_MMC_JZ4725B)
+			writel(JZ_MMC_LPM_LOW_POWER_MODE_EN,
+				   host->base + JZ_REG_MMC_LPM);
+	}
+
+	return real_rate;
+}
+
+static void jz4740_mmc_request(struct mmc_host *mmc, struct mmc_request *req)
+{
+	struct jz4740_mmc_host *host = mmc_priv(mmc);
+
+	host->req = req;
+
+	jz4740_mmc_write_irq_reg(host, ~0);
+	jz4740_mmc_set_irq_enabled(host, JZ_MMC_IRQ_END_CMD_RES, true);
+
+	host->state = JZ4740_MMC_STATE_READ_RESPONSE;
+	set_bit(0, &host->waiting);
+	mod_timer(&host->timeout_timer,
+		  jiffies + msecs_to_jiffies(JZ_MMC_REQ_TIMEOUT_MS));
+	jz4740_mmc_send_command(host, req->cmd);
+}
+
+static void jz4740_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct jz4740_mmc_host *host = mmc_priv(mmc);
+	if (ios->clock)
+		jz4740_mmc_set_clock_rate(host, ios->clock);
+
+	switch (ios->power_mode) {
+	case MMC_POWER_UP:
+		jz4740_mmc_reset(host);
+		if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, ios->vdd);
+		host->cmdat |= JZ_MMC_CMDAT_INIT;
+		clk_prepare_enable(host->clk);
+		break;
+	case MMC_POWER_ON:
+		break;
+	default:
+		if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+		clk_disable_unprepare(host->clk);
+		break;
+	}
+
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_1:
+		host->cmdat &= ~JZ_MMC_CMDAT_BUS_WIDTH_MASK;
+		break;
+	case MMC_BUS_WIDTH_4:
+		host->cmdat &= ~JZ_MMC_CMDAT_BUS_WIDTH_MASK;
+		host->cmdat |= JZ_MMC_CMDAT_BUS_WIDTH_4BIT;
+		break;
+	case MMC_BUS_WIDTH_8:
+		host->cmdat &= ~JZ_MMC_CMDAT_BUS_WIDTH_MASK;
+		host->cmdat |= JZ_MMC_CMDAT_BUS_WIDTH_8BIT;
+		break;
+	default:
+		break;
+	}
+}
+
+static void jz4740_mmc_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct jz4740_mmc_host *host = mmc_priv(mmc);
+	jz4740_mmc_set_irq_enabled(host, JZ_MMC_IRQ_SDIO, enable);
+}
+
+static const struct mmc_host_ops jz4740_mmc_ops = {
+	.request	= jz4740_mmc_request,
+	.pre_req	= jz4740_mmc_pre_request,
+	.post_req	= jz4740_mmc_post_request,
+	.set_ios	= jz4740_mmc_set_ios,
+	.get_ro		= mmc_gpio_get_ro,
+	.get_cd		= mmc_gpio_get_cd,
+	.enable_sdio_irq = jz4740_mmc_enable_sdio_irq,
+};
+
+static const struct of_device_id jz4740_mmc_of_match[] = {
+	{ .compatible = "ingenic,jz4740-mmc", .data = (void *) JZ_MMC_JZ4740 },
+	{ .compatible = "ingenic,jz4725b-mmc", .data = (void *)JZ_MMC_JZ4725B },
+	{ .compatible = "ingenic,jz4760-mmc", .data = (void *) JZ_MMC_JZ4760 },
+	{ .compatible = "ingenic,jz4780-mmc", .data = (void *) JZ_MMC_JZ4780 },
+	{ .compatible = "ingenic,x1000-mmc", .data = (void *) JZ_MMC_X1000 },
+	{},
+};
+MODULE_DEVICE_TABLE(of, jz4740_mmc_of_match);
+
+static int jz4740_mmc_probe(struct platform_device* pdev)
+{
+	int ret;
+	struct mmc_host *mmc;
+	struct jz4740_mmc_host *host;
+	const struct of_device_id *match;
+
+	mmc = mmc_alloc_host(sizeof(struct jz4740_mmc_host), &pdev->dev);
+	if (!mmc) {
+		dev_err(&pdev->dev, "Failed to alloc mmc host structure\n");
+		return -ENOMEM;
+	}
+
+	host = mmc_priv(mmc);
+
+	match = of_match_device(jz4740_mmc_of_match, &pdev->dev);
+	if (match) {
+		host->version = (enum jz4740_mmc_version)match->data;
+	} else {
+		/* JZ4740 should be the only one using legacy probe */
+		host->version = JZ_MMC_JZ4740;
+	}
+
+	ret = mmc_of_parse(mmc);
+	if (ret) {
+		dev_err_probe(&pdev->dev, ret, "could not parse device properties\n");
+		goto err_free_host;
+	}
+
+	mmc_regulator_get_supply(mmc);
+
+	host->irq = platform_get_irq(pdev, 0);
+	if (host->irq < 0) {
+		ret = host->irq;
+		goto err_free_host;
+	}
+
+	host->clk = devm_clk_get(&pdev->dev, "mmc");
+	if (IS_ERR(host->clk)) {
+		ret = PTR_ERR(host->clk);
+		dev_err(&pdev->dev, "Failed to get mmc clock\n");
+		goto err_free_host;
+	}
+
+	host->mem_res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	host->base = devm_ioremap_resource(&pdev->dev, host->mem_res);
+	if (IS_ERR(host->base)) {
+		ret = PTR_ERR(host->base);
+		dev_err(&pdev->dev, "Failed to ioremap base memory\n");
+		goto err_free_host;
+	}
+
+	mmc->ops = &jz4740_mmc_ops;
+	if (!mmc->f_max)
+		mmc->f_max = JZ_MMC_CLK_RATE;
+
+	/*
+	 * There seems to be a problem with this driver on the JZ4760 and
+	 * JZ4760B SoCs. There, when using the maximum rate supported (50 MHz),
+	 * the communication fails with many SD cards.
+	 * Until this bug is sorted out, limit the maximum rate to 24 MHz.
+	 */
+	if (host->version == JZ_MMC_JZ4760 && mmc->f_max > JZ_MMC_CLK_RATE)
+		mmc->f_max = JZ_MMC_CLK_RATE;
+
+	mmc->f_min = mmc->f_max / 128;
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+
+	/*
+	 * We use a fixed timeout of 5s, hence inform the core about it. A
+	 * future improvement should instead respect the cmd->busy_timeout.
+	 */
+	mmc->max_busy_timeout = JZ_MMC_REQ_TIMEOUT_MS;
+
+	mmc->max_blk_size = (1 << 10) - 1;
+	mmc->max_blk_count = (1 << 15) - 1;
+	mmc->max_req_size = mmc->max_blk_size * mmc->max_blk_count;
+
+	mmc->max_segs = 128;
+	mmc->max_seg_size = mmc->max_req_size;
+
+	host->mmc = mmc;
+	host->pdev = pdev;
+	spin_lock_init(&host->lock);
+	host->irq_mask = ~0;
+
+	jz4740_mmc_reset(host);
+
+	ret = request_threaded_irq(host->irq, jz_mmc_irq, jz_mmc_irq_worker, 0,
+			dev_name(&pdev->dev), host);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to request irq: %d\n", ret);
+		goto err_free_host;
+	}
+
+	jz4740_mmc_clock_disable(host);
+	timer_setup(&host->timeout_timer, jz4740_mmc_timeout, 0);
+
+	ret = jz4740_mmc_acquire_dma_channels(host);
+	if (ret == -EPROBE_DEFER)
+		goto err_free_irq;
+	host->use_dma = !ret;
+
+	platform_set_drvdata(pdev, host);
+	ret = mmc_add_host(mmc);
+
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to add mmc host: %d\n", ret);
+		goto err_release_dma;
+	}
+	dev_info(&pdev->dev, "Ingenic SD/MMC card driver registered\n");
+
+	dev_info(&pdev->dev, "Using %s, %d-bit mode\n",
+		 host->use_dma ? "DMA" : "PIO",
+		 (mmc->caps & MMC_CAP_8_BIT_DATA) ? 8 :
+		 ((mmc->caps & MMC_CAP_4_BIT_DATA) ? 4 : 1));
+
+	return 0;
+
+err_release_dma:
+	if (host->use_dma)
+		jz4740_mmc_release_dma_channels(host);
+err_free_irq:
+	free_irq(host->irq, host);
+err_free_host:
+	mmc_free_host(mmc);
+
+	return ret;
+}
+
+static int jz4740_mmc_remove(struct platform_device *pdev)
+{
+	struct jz4740_mmc_host *host = platform_get_drvdata(pdev);
+
+	del_timer_sync(&host->timeout_timer);
+	jz4740_mmc_set_irq_enabled(host, 0xff, false);
+	jz4740_mmc_reset(host);
+
+	mmc_remove_host(host->mmc);
+
+	free_irq(host->irq, host);
+
+	if (host->use_dma)
+		jz4740_mmc_release_dma_channels(host);
+
+	mmc_free_host(host->mmc);
+
+	return 0;
+}
+
+static int __maybe_unused jz4740_mmc_suspend(struct device *dev)
+{
+	return pinctrl_pm_select_sleep_state(dev);
+}
+
+static int __maybe_unused jz4740_mmc_resume(struct device *dev)
+{
+	return pinctrl_select_default_state(dev);
+}
+
+static SIMPLE_DEV_PM_OPS(jz4740_mmc_pm_ops, jz4740_mmc_suspend,
+	jz4740_mmc_resume);
+
+static struct platform_driver jz4740_mmc_driver = {
+	.probe = jz4740_mmc_probe,
+	.remove = jz4740_mmc_remove,
+	.driver = {
+		.name = "jz4740-mmc",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(jz4740_mmc_of_match),
+		.pm = pm_ptr(&jz4740_mmc_pm_ops),
+	},
+};
+
+module_platform_driver(jz4740_mmc_driver);
+
+MODULE_DESCRIPTION("JZ4740 SD/MMC controller driver");
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Lars-Peter Clausen <lars@metafoo.de>");
diff --git a/drivers/mmc/host/meson-gx-mmc.c b/drivers/mmc/host/meson-gx-mmc.c
new file mode 100644
index 000000000..78c9a56b0
--- /dev/null
+++ b/drivers/mmc/host/meson-gx-mmc.c
@@ -0,0 +1,1341 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Amlogic SD/eMMC driver for the GX/S905 family SoCs
+ *
+ * Copyright (c) 2016 BayLibre, SAS.
+ * Author: Kevin Hilman <khilman@baylibre.com>
+ */
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/iopoll.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <linux/ioport.h>
+#include <linux/dma-mapping.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sdio.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/io.h>
+#include <linux/clk.h>
+#include <linux/clk-provider.h>
+#include <linux/regulator/consumer.h>
+#include <linux/reset.h>
+#include <linux/interrupt.h>
+#include <linux/bitfield.h>
+#include <linux/pinctrl/consumer.h>
+
+#define DRIVER_NAME "meson-gx-mmc"
+
+#define SD_EMMC_CLOCK 0x0
+#define   CLK_DIV_MASK GENMASK(5, 0)
+#define   CLK_SRC_MASK GENMASK(7, 6)
+#define   CLK_CORE_PHASE_MASK GENMASK(9, 8)
+#define   CLK_TX_PHASE_MASK GENMASK(11, 10)
+#define   CLK_RX_PHASE_MASK GENMASK(13, 12)
+#define   CLK_PHASE_0 0
+#define   CLK_PHASE_180 2
+#define   CLK_V2_TX_DELAY_MASK GENMASK(19, 16)
+#define   CLK_V2_RX_DELAY_MASK GENMASK(23, 20)
+#define   CLK_V2_ALWAYS_ON BIT(24)
+
+#define   CLK_V3_TX_DELAY_MASK GENMASK(21, 16)
+#define   CLK_V3_RX_DELAY_MASK GENMASK(27, 22)
+#define   CLK_V3_ALWAYS_ON BIT(28)
+
+#define   CLK_TX_DELAY_MASK(h)		(h->data->tx_delay_mask)
+#define   CLK_RX_DELAY_MASK(h)		(h->data->rx_delay_mask)
+#define   CLK_ALWAYS_ON(h)		(h->data->always_on)
+
+#define SD_EMMC_DELAY 0x4
+#define SD_EMMC_ADJUST 0x8
+#define   ADJUST_ADJ_DELAY_MASK GENMASK(21, 16)
+#define   ADJUST_DS_EN BIT(15)
+#define   ADJUST_ADJ_EN BIT(13)
+
+#define SD_EMMC_DELAY1 0x4
+#define SD_EMMC_DELAY2 0x8
+#define SD_EMMC_V3_ADJUST 0xc
+
+#define SD_EMMC_CALOUT 0x10
+#define SD_EMMC_START 0x40
+#define   START_DESC_INIT BIT(0)
+#define   START_DESC_BUSY BIT(1)
+#define   START_DESC_ADDR_MASK GENMASK(31, 2)
+
+#define SD_EMMC_CFG 0x44
+#define   CFG_BUS_WIDTH_MASK GENMASK(1, 0)
+#define   CFG_BUS_WIDTH_1 0x0
+#define   CFG_BUS_WIDTH_4 0x1
+#define   CFG_BUS_WIDTH_8 0x2
+#define   CFG_DDR BIT(2)
+#define   CFG_BLK_LEN_MASK GENMASK(7, 4)
+#define   CFG_RESP_TIMEOUT_MASK GENMASK(11, 8)
+#define   CFG_RC_CC_MASK GENMASK(15, 12)
+#define   CFG_STOP_CLOCK BIT(22)
+#define   CFG_CLK_ALWAYS_ON BIT(18)
+#define   CFG_CHK_DS BIT(20)
+#define   CFG_AUTO_CLK BIT(23)
+#define   CFG_ERR_ABORT BIT(27)
+
+#define SD_EMMC_STATUS 0x48
+#define   STATUS_BUSY BIT(31)
+#define   STATUS_DESC_BUSY BIT(30)
+#define   STATUS_DATI GENMASK(23, 16)
+
+#define SD_EMMC_IRQ_EN 0x4c
+#define   IRQ_RXD_ERR_MASK GENMASK(7, 0)
+#define   IRQ_TXD_ERR BIT(8)
+#define   IRQ_DESC_ERR BIT(9)
+#define   IRQ_RESP_ERR BIT(10)
+#define   IRQ_CRC_ERR \
+	(IRQ_RXD_ERR_MASK | IRQ_TXD_ERR | IRQ_DESC_ERR | IRQ_RESP_ERR)
+#define   IRQ_RESP_TIMEOUT BIT(11)
+#define   IRQ_DESC_TIMEOUT BIT(12)
+#define   IRQ_TIMEOUTS \
+	(IRQ_RESP_TIMEOUT | IRQ_DESC_TIMEOUT)
+#define   IRQ_END_OF_CHAIN BIT(13)
+#define   IRQ_RESP_STATUS BIT(14)
+#define   IRQ_SDIO BIT(15)
+#define   IRQ_EN_MASK \
+	(IRQ_CRC_ERR | IRQ_TIMEOUTS | IRQ_END_OF_CHAIN | IRQ_RESP_STATUS |\
+	 IRQ_SDIO)
+
+#define SD_EMMC_CMD_CFG 0x50
+#define SD_EMMC_CMD_ARG 0x54
+#define SD_EMMC_CMD_DAT 0x58
+#define SD_EMMC_CMD_RSP 0x5c
+#define SD_EMMC_CMD_RSP1 0x60
+#define SD_EMMC_CMD_RSP2 0x64
+#define SD_EMMC_CMD_RSP3 0x68
+
+#define SD_EMMC_RXD 0x94
+#define SD_EMMC_TXD 0x94
+#define SD_EMMC_LAST_REG SD_EMMC_TXD
+
+#define SD_EMMC_SRAM_DATA_BUF_LEN 1536
+#define SD_EMMC_SRAM_DATA_BUF_OFF 0x200
+
+#define SD_EMMC_CFG_BLK_SIZE 512 /* internal buffer max: 512 bytes */
+#define SD_EMMC_CFG_RESP_TIMEOUT 256 /* in clock cycles */
+#define SD_EMMC_CMD_TIMEOUT 1024 /* in ms */
+#define SD_EMMC_CMD_TIMEOUT_DATA 4096 /* in ms */
+#define SD_EMMC_CFG_CMD_GAP 16 /* in clock cycles */
+#define SD_EMMC_DESC_BUF_LEN PAGE_SIZE
+
+#define SD_EMMC_PRE_REQ_DONE BIT(0)
+#define SD_EMMC_DESC_CHAIN_MODE BIT(1)
+
+#define MUX_CLK_NUM_PARENTS 2
+
+struct meson_mmc_data {
+	unsigned int tx_delay_mask;
+	unsigned int rx_delay_mask;
+	unsigned int always_on;
+	unsigned int adjust;
+};
+
+struct sd_emmc_desc {
+	u32 cmd_cfg;
+	u32 cmd_arg;
+	u32 cmd_data;
+	u32 cmd_resp;
+};
+
+struct meson_host {
+	struct	device		*dev;
+	struct	meson_mmc_data *data;
+	struct	mmc_host	*mmc;
+	struct	mmc_command	*cmd;
+
+	void __iomem *regs;
+	struct clk *core_clk;
+	struct clk *mux_clk;
+	struct clk *mmc_clk;
+	unsigned long req_rate;
+	bool ddr;
+
+	bool dram_access_quirk;
+
+	struct pinctrl *pinctrl;
+	struct pinctrl_state *pins_clk_gate;
+
+	unsigned int bounce_buf_size;
+	void *bounce_buf;
+	void __iomem *bounce_iomem_buf;
+	dma_addr_t bounce_dma_addr;
+	struct sd_emmc_desc *descs;
+	dma_addr_t descs_dma_addr;
+
+	int irq;
+
+	bool vqmmc_enabled;
+	bool needs_pre_post_req;
+
+};
+
+#define CMD_CFG_LENGTH_MASK GENMASK(8, 0)
+#define CMD_CFG_BLOCK_MODE BIT(9)
+#define CMD_CFG_R1B BIT(10)
+#define CMD_CFG_END_OF_CHAIN BIT(11)
+#define CMD_CFG_TIMEOUT_MASK GENMASK(15, 12)
+#define CMD_CFG_NO_RESP BIT(16)
+#define CMD_CFG_NO_CMD BIT(17)
+#define CMD_CFG_DATA_IO BIT(18)
+#define CMD_CFG_DATA_WR BIT(19)
+#define CMD_CFG_RESP_NOCRC BIT(20)
+#define CMD_CFG_RESP_128 BIT(21)
+#define CMD_CFG_RESP_NUM BIT(22)
+#define CMD_CFG_DATA_NUM BIT(23)
+#define CMD_CFG_CMD_INDEX_MASK GENMASK(29, 24)
+#define CMD_CFG_ERROR BIT(30)
+#define CMD_CFG_OWNER BIT(31)
+
+#define CMD_DATA_MASK GENMASK(31, 2)
+#define CMD_DATA_BIG_ENDIAN BIT(1)
+#define CMD_DATA_SRAM BIT(0)
+#define CMD_RESP_MASK GENMASK(31, 1)
+#define CMD_RESP_SRAM BIT(0)
+
+static unsigned int meson_mmc_get_timeout_msecs(struct mmc_data *data)
+{
+	unsigned int timeout = data->timeout_ns / NSEC_PER_MSEC;
+
+	if (!timeout)
+		return SD_EMMC_CMD_TIMEOUT_DATA;
+
+	timeout = roundup_pow_of_two(timeout);
+
+	return min(timeout, 32768U); /* max. 2^15 ms */
+}
+
+static struct mmc_command *meson_mmc_get_next_command(struct mmc_command *cmd)
+{
+	if (cmd->opcode == MMC_SET_BLOCK_COUNT && !cmd->error)
+		return cmd->mrq->cmd;
+	else if (mmc_op_multi(cmd->opcode) &&
+		 (!cmd->mrq->sbc || cmd->error || cmd->data->error))
+		return cmd->mrq->stop;
+	else
+		return NULL;
+}
+
+static void meson_mmc_get_transfer_mode(struct mmc_host *mmc,
+					struct mmc_request *mrq)
+{
+	struct meson_host *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+	struct scatterlist *sg;
+	int i;
+	bool use_desc_chain_mode = true;
+
+	/*
+	 * When Controller DMA cannot directly access DDR memory, disable
+	 * support for Chain Mode to directly use the internal SRAM using
+	 * the bounce buffer mode.
+	 */
+	if (host->dram_access_quirk)
+		return;
+
+	/*
+	 * Broken SDIO with AP6255-based WiFi on Khadas VIM Pro has been
+	 * reported. For some strange reason this occurs in descriptor
+	 * chain mode only. So let's fall back to bounce buffer mode
+	 * for command SD_IO_RW_EXTENDED.
+	 */
+	if (mrq->cmd->opcode == SD_IO_RW_EXTENDED)
+		return;
+
+	for_each_sg(data->sg, sg, data->sg_len, i)
+		/* check for 8 byte alignment */
+		if (sg->offset & 7) {
+			WARN_ONCE(1, "unaligned scatterlist buffer\n");
+			use_desc_chain_mode = false;
+			break;
+		}
+
+	if (use_desc_chain_mode)
+		data->host_cookie |= SD_EMMC_DESC_CHAIN_MODE;
+}
+
+static inline bool meson_mmc_desc_chain_mode(const struct mmc_data *data)
+{
+	return data->host_cookie & SD_EMMC_DESC_CHAIN_MODE;
+}
+
+static inline bool meson_mmc_bounce_buf_read(const struct mmc_data *data)
+{
+	return data && data->flags & MMC_DATA_READ &&
+	       !meson_mmc_desc_chain_mode(data);
+}
+
+static void meson_mmc_pre_req(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct mmc_data *data = mrq->data;
+
+	if (!data)
+		return;
+
+	meson_mmc_get_transfer_mode(mmc, mrq);
+	data->host_cookie |= SD_EMMC_PRE_REQ_DONE;
+
+	if (!meson_mmc_desc_chain_mode(data))
+		return;
+
+	data->sg_count = dma_map_sg(mmc_dev(mmc), data->sg, data->sg_len,
+                                   mmc_get_dma_dir(data));
+	if (!data->sg_count)
+		dev_err(mmc_dev(mmc), "dma_map_sg failed");
+}
+
+static void meson_mmc_post_req(struct mmc_host *mmc, struct mmc_request *mrq,
+			       int err)
+{
+	struct mmc_data *data = mrq->data;
+
+	if (data && meson_mmc_desc_chain_mode(data) && data->sg_count)
+		dma_unmap_sg(mmc_dev(mmc), data->sg, data->sg_len,
+			     mmc_get_dma_dir(data));
+}
+
+/*
+ * Gating the clock on this controller is tricky.  It seems the mmc clock
+ * is also used by the controller.  It may crash during some operation if the
+ * clock is stopped.  The safest thing to do, whenever possible, is to keep
+ * clock running at stop it at the pad using the pinmux.
+ */
+static void meson_mmc_clk_gate(struct meson_host *host)
+{
+	u32 cfg;
+
+	if (host->pins_clk_gate) {
+		pinctrl_select_state(host->pinctrl, host->pins_clk_gate);
+	} else {
+		/*
+		 * If the pinmux is not provided - default to the classic and
+		 * unsafe method
+		 */
+		cfg = readl(host->regs + SD_EMMC_CFG);
+		cfg |= CFG_STOP_CLOCK;
+		writel(cfg, host->regs + SD_EMMC_CFG);
+	}
+}
+
+static void meson_mmc_clk_ungate(struct meson_host *host)
+{
+	u32 cfg;
+
+	if (host->pins_clk_gate)
+		pinctrl_select_default_state(host->dev);
+
+	/* Make sure the clock is not stopped in the controller */
+	cfg = readl(host->regs + SD_EMMC_CFG);
+	cfg &= ~CFG_STOP_CLOCK;
+	writel(cfg, host->regs + SD_EMMC_CFG);
+}
+
+static int meson_mmc_clk_set(struct meson_host *host, unsigned long rate,
+			     bool ddr)
+{
+	struct mmc_host *mmc = host->mmc;
+	int ret;
+	u32 cfg;
+
+	/* Same request - bail-out */
+	if (host->ddr == ddr && host->req_rate == rate)
+		return 0;
+
+	/* stop clock */
+	meson_mmc_clk_gate(host);
+	host->req_rate = 0;
+	mmc->actual_clock = 0;
+
+	/* return with clock being stopped */
+	if (!rate)
+		return 0;
+
+	/* Stop the clock during rate change to avoid glitches */
+	cfg = readl(host->regs + SD_EMMC_CFG);
+	cfg |= CFG_STOP_CLOCK;
+	writel(cfg, host->regs + SD_EMMC_CFG);
+
+	if (ddr) {
+		/* DDR modes require higher module clock */
+		rate <<= 1;
+		cfg |= CFG_DDR;
+	} else {
+		cfg &= ~CFG_DDR;
+	}
+	writel(cfg, host->regs + SD_EMMC_CFG);
+	host->ddr = ddr;
+
+	ret = clk_set_rate(host->mmc_clk, rate);
+	if (ret) {
+		dev_err(host->dev, "Unable to set cfg_div_clk to %lu. ret=%d\n",
+			rate, ret);
+		return ret;
+	}
+
+	host->req_rate = rate;
+	mmc->actual_clock = clk_get_rate(host->mmc_clk);
+
+	/* We should report the real output frequency of the controller */
+	if (ddr) {
+		host->req_rate >>= 1;
+		mmc->actual_clock >>= 1;
+	}
+
+	dev_dbg(host->dev, "clk rate: %u Hz\n", mmc->actual_clock);
+	if (rate != mmc->actual_clock)
+		dev_dbg(host->dev, "requested rate was %lu\n", rate);
+
+	/* (re)start clock */
+	meson_mmc_clk_ungate(host);
+
+	return 0;
+}
+
+/*
+ * The SD/eMMC IP block has an internal mux and divider used for
+ * generating the MMC clock.  Use the clock framework to create and
+ * manage these clocks.
+ */
+static int meson_mmc_clk_init(struct meson_host *host)
+{
+	struct clk_init_data init;
+	struct clk_mux *mux;
+	struct clk_divider *div;
+	char clk_name[32];
+	int i, ret = 0;
+	const char *mux_parent_names[MUX_CLK_NUM_PARENTS];
+	const char *clk_parent[1];
+	u32 clk_reg;
+
+	/* init SD_EMMC_CLOCK to sane defaults w/min clock rate */
+	clk_reg = CLK_ALWAYS_ON(host);
+	clk_reg |= CLK_DIV_MASK;
+	clk_reg |= FIELD_PREP(CLK_CORE_PHASE_MASK, CLK_PHASE_180);
+	clk_reg |= FIELD_PREP(CLK_TX_PHASE_MASK, CLK_PHASE_0);
+	clk_reg |= FIELD_PREP(CLK_RX_PHASE_MASK, CLK_PHASE_0);
+	writel(clk_reg, host->regs + SD_EMMC_CLOCK);
+
+	/* get the mux parents */
+	for (i = 0; i < MUX_CLK_NUM_PARENTS; i++) {
+		struct clk *clk;
+		char name[16];
+
+		snprintf(name, sizeof(name), "clkin%d", i);
+		clk = devm_clk_get(host->dev, name);
+		if (IS_ERR(clk))
+			return dev_err_probe(host->dev, PTR_ERR(clk),
+					     "Missing clock %s\n", name);
+
+		mux_parent_names[i] = __clk_get_name(clk);
+	}
+
+	/* create the mux */
+	mux = devm_kzalloc(host->dev, sizeof(*mux), GFP_KERNEL);
+	if (!mux)
+		return -ENOMEM;
+
+	snprintf(clk_name, sizeof(clk_name), "%s#mux", dev_name(host->dev));
+	init.name = clk_name;
+	init.ops = &clk_mux_ops;
+	init.flags = 0;
+	init.parent_names = mux_parent_names;
+	init.num_parents = MUX_CLK_NUM_PARENTS;
+
+	mux->reg = host->regs + SD_EMMC_CLOCK;
+	mux->shift = __ffs(CLK_SRC_MASK);
+	mux->mask = CLK_SRC_MASK >> mux->shift;
+	mux->hw.init = &init;
+
+	host->mux_clk = devm_clk_register(host->dev, &mux->hw);
+	if (WARN_ON(IS_ERR(host->mux_clk)))
+		return PTR_ERR(host->mux_clk);
+
+	/* create the divider */
+	div = devm_kzalloc(host->dev, sizeof(*div), GFP_KERNEL);
+	if (!div)
+		return -ENOMEM;
+
+	snprintf(clk_name, sizeof(clk_name), "%s#div", dev_name(host->dev));
+	init.name = clk_name;
+	init.ops = &clk_divider_ops;
+	init.flags = CLK_SET_RATE_PARENT;
+	clk_parent[0] = __clk_get_name(host->mux_clk);
+	init.parent_names = clk_parent;
+	init.num_parents = 1;
+
+	div->reg = host->regs + SD_EMMC_CLOCK;
+	div->shift = __ffs(CLK_DIV_MASK);
+	div->width = __builtin_popcountl(CLK_DIV_MASK);
+	div->hw.init = &init;
+	div->flags = CLK_DIVIDER_ONE_BASED;
+
+	host->mmc_clk = devm_clk_register(host->dev, &div->hw);
+	if (WARN_ON(IS_ERR(host->mmc_clk)))
+		return PTR_ERR(host->mmc_clk);
+
+	/* init SD_EMMC_CLOCK to sane defaults w/min clock rate */
+	host->mmc->f_min = clk_round_rate(host->mmc_clk, 400000);
+	ret = clk_set_rate(host->mmc_clk, host->mmc->f_min);
+	if (ret)
+		return ret;
+
+	return clk_prepare_enable(host->mmc_clk);
+}
+
+static void meson_mmc_disable_resampling(struct meson_host *host)
+{
+	unsigned int val = readl(host->regs + host->data->adjust);
+
+	val &= ~ADJUST_ADJ_EN;
+	writel(val, host->regs + host->data->adjust);
+}
+
+static void meson_mmc_reset_resampling(struct meson_host *host)
+{
+	unsigned int val;
+
+	meson_mmc_disable_resampling(host);
+
+	val = readl(host->regs + host->data->adjust);
+	val &= ~ADJUST_ADJ_DELAY_MASK;
+	writel(val, host->regs + host->data->adjust);
+}
+
+static int meson_mmc_resampling_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct meson_host *host = mmc_priv(mmc);
+	unsigned int val, dly, max_dly, i;
+	int ret;
+
+	/* Resampling is done using the source clock */
+	max_dly = DIV_ROUND_UP(clk_get_rate(host->mux_clk),
+			       clk_get_rate(host->mmc_clk));
+
+	val = readl(host->regs + host->data->adjust);
+	val |= ADJUST_ADJ_EN;
+	writel(val, host->regs + host->data->adjust);
+
+	if (mmc_doing_retune(mmc))
+		dly = FIELD_GET(ADJUST_ADJ_DELAY_MASK, val) + 1;
+	else
+		dly = 0;
+
+	for (i = 0; i < max_dly; i++) {
+		val &= ~ADJUST_ADJ_DELAY_MASK;
+		val |= FIELD_PREP(ADJUST_ADJ_DELAY_MASK, (dly + i) % max_dly);
+		writel(val, host->regs + host->data->adjust);
+
+		ret = mmc_send_tuning(mmc, opcode, NULL);
+		if (!ret) {
+			dev_dbg(mmc_dev(mmc), "resampling delay: %u\n",
+				(dly + i) % max_dly);
+			return 0;
+		}
+	}
+
+	meson_mmc_reset_resampling(host);
+	return -EIO;
+}
+
+static int meson_mmc_prepare_ios_clock(struct meson_host *host,
+				       struct mmc_ios *ios)
+{
+	bool ddr;
+
+	switch (ios->timing) {
+	case MMC_TIMING_MMC_DDR52:
+	case MMC_TIMING_UHS_DDR50:
+		ddr = true;
+		break;
+
+	default:
+		ddr = false;
+		break;
+	}
+
+	return meson_mmc_clk_set(host, ios->clock, ddr);
+}
+
+static void meson_mmc_check_resampling(struct meson_host *host,
+				       struct mmc_ios *ios)
+{
+	switch (ios->timing) {
+	case MMC_TIMING_LEGACY:
+	case MMC_TIMING_MMC_HS:
+	case MMC_TIMING_SD_HS:
+	case MMC_TIMING_MMC_DDR52:
+		meson_mmc_disable_resampling(host);
+		break;
+	}
+}
+
+static void meson_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct meson_host *host = mmc_priv(mmc);
+	u32 bus_width, val;
+	int err;
+
+	/*
+	 * GPIO regulator, only controls switching between 1v8 and
+	 * 3v3, doesn't support MMC_POWER_OFF, MMC_POWER_ON.
+	 */
+	switch (ios->power_mode) {
+	case MMC_POWER_OFF:
+		if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+
+		if (!IS_ERR(mmc->supply.vqmmc) && host->vqmmc_enabled) {
+			regulator_disable(mmc->supply.vqmmc);
+			host->vqmmc_enabled = false;
+		}
+
+		break;
+
+	case MMC_POWER_UP:
+		if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, ios->vdd);
+
+		break;
+
+	case MMC_POWER_ON:
+		if (!IS_ERR(mmc->supply.vqmmc) && !host->vqmmc_enabled) {
+			int ret = regulator_enable(mmc->supply.vqmmc);
+
+			if (ret < 0)
+				dev_err(host->dev,
+					"failed to enable vqmmc regulator\n");
+			else
+				host->vqmmc_enabled = true;
+		}
+
+		break;
+	}
+
+	/* Bus width */
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_1:
+		bus_width = CFG_BUS_WIDTH_1;
+		break;
+	case MMC_BUS_WIDTH_4:
+		bus_width = CFG_BUS_WIDTH_4;
+		break;
+	case MMC_BUS_WIDTH_8:
+		bus_width = CFG_BUS_WIDTH_8;
+		break;
+	default:
+		dev_err(host->dev, "Invalid ios->bus_width: %u.  Setting to 4.\n",
+			ios->bus_width);
+		bus_width = CFG_BUS_WIDTH_4;
+	}
+
+	val = readl(host->regs + SD_EMMC_CFG);
+	val &= ~CFG_BUS_WIDTH_MASK;
+	val |= FIELD_PREP(CFG_BUS_WIDTH_MASK, bus_width);
+	writel(val, host->regs + SD_EMMC_CFG);
+
+	meson_mmc_check_resampling(host, ios);
+	err = meson_mmc_prepare_ios_clock(host, ios);
+	if (err)
+		dev_err(host->dev, "Failed to set clock: %d\n,", err);
+
+	dev_dbg(host->dev, "SD_EMMC_CFG:  0x%08x\n", val);
+}
+
+static void meson_mmc_request_done(struct mmc_host *mmc,
+				   struct mmc_request *mrq)
+{
+	struct meson_host *host = mmc_priv(mmc);
+
+	host->cmd = NULL;
+	if (host->needs_pre_post_req)
+		meson_mmc_post_req(mmc, mrq, 0);
+	mmc_request_done(host->mmc, mrq);
+}
+
+static void meson_mmc_set_blksz(struct mmc_host *mmc, unsigned int blksz)
+{
+	struct meson_host *host = mmc_priv(mmc);
+	u32 cfg, blksz_old;
+
+	cfg = readl(host->regs + SD_EMMC_CFG);
+	blksz_old = FIELD_GET(CFG_BLK_LEN_MASK, cfg);
+
+	if (!is_power_of_2(blksz))
+		dev_err(host->dev, "blksz %u is not a power of 2\n", blksz);
+
+	blksz = ilog2(blksz);
+
+	/* check if block-size matches, if not update */
+	if (blksz == blksz_old)
+		return;
+
+	dev_dbg(host->dev, "%s: update blk_len %d -> %d\n", __func__,
+		blksz_old, blksz);
+
+	cfg &= ~CFG_BLK_LEN_MASK;
+	cfg |= FIELD_PREP(CFG_BLK_LEN_MASK, blksz);
+	writel(cfg, host->regs + SD_EMMC_CFG);
+}
+
+static void meson_mmc_set_response_bits(struct mmc_command *cmd, u32 *cmd_cfg)
+{
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136)
+			*cmd_cfg |= CMD_CFG_RESP_128;
+		*cmd_cfg |= CMD_CFG_RESP_NUM;
+
+		if (!(cmd->flags & MMC_RSP_CRC))
+			*cmd_cfg |= CMD_CFG_RESP_NOCRC;
+
+		if (cmd->flags & MMC_RSP_BUSY)
+			*cmd_cfg |= CMD_CFG_R1B;
+	} else {
+		*cmd_cfg |= CMD_CFG_NO_RESP;
+	}
+}
+
+static void meson_mmc_desc_chain_transfer(struct mmc_host *mmc, u32 cmd_cfg)
+{
+	struct meson_host *host = mmc_priv(mmc);
+	struct sd_emmc_desc *desc = host->descs;
+	struct mmc_data *data = host->cmd->data;
+	struct scatterlist *sg;
+	u32 start;
+	int i;
+
+	if (data->flags & MMC_DATA_WRITE)
+		cmd_cfg |= CMD_CFG_DATA_WR;
+
+	if (data->blocks > 1) {
+		cmd_cfg |= CMD_CFG_BLOCK_MODE;
+		meson_mmc_set_blksz(mmc, data->blksz);
+	}
+
+	for_each_sg(data->sg, sg, data->sg_count, i) {
+		unsigned int len = sg_dma_len(sg);
+
+		if (data->blocks > 1)
+			len /= data->blksz;
+
+		desc[i].cmd_cfg = cmd_cfg;
+		desc[i].cmd_cfg |= FIELD_PREP(CMD_CFG_LENGTH_MASK, len);
+		if (i > 0)
+			desc[i].cmd_cfg |= CMD_CFG_NO_CMD;
+		desc[i].cmd_arg = host->cmd->arg;
+		desc[i].cmd_resp = 0;
+		desc[i].cmd_data = sg_dma_address(sg);
+	}
+	desc[data->sg_count - 1].cmd_cfg |= CMD_CFG_END_OF_CHAIN;
+
+	dma_wmb(); /* ensure descriptor is written before kicked */
+	start = host->descs_dma_addr | START_DESC_BUSY;
+	writel(start, host->regs + SD_EMMC_START);
+}
+
+/* local sg copy for dram_access_quirk */
+static void meson_mmc_copy_buffer(struct meson_host *host, struct mmc_data *data,
+				  size_t buflen, bool to_buffer)
+{
+	unsigned int sg_flags = SG_MITER_ATOMIC;
+	struct scatterlist *sgl = data->sg;
+	unsigned int nents = data->sg_len;
+	struct sg_mapping_iter miter;
+	unsigned int offset = 0;
+
+	if (to_buffer)
+		sg_flags |= SG_MITER_FROM_SG;
+	else
+		sg_flags |= SG_MITER_TO_SG;
+
+	sg_miter_start(&miter, sgl, nents, sg_flags);
+
+	while ((offset < buflen) && sg_miter_next(&miter)) {
+		unsigned int buf_offset = 0;
+		unsigned int len, left;
+		u32 *buf = miter.addr;
+
+		len = min(miter.length, buflen - offset);
+		left = len;
+
+		if (to_buffer) {
+			do {
+				writel(*buf++, host->bounce_iomem_buf + offset + buf_offset);
+
+				buf_offset += 4;
+				left -= 4;
+			} while (left);
+		} else {
+			do {
+				*buf++ = readl(host->bounce_iomem_buf + offset + buf_offset);
+
+				buf_offset += 4;
+				left -= 4;
+			} while (left);
+		}
+
+		offset += len;
+	}
+
+	sg_miter_stop(&miter);
+}
+
+static void meson_mmc_start_cmd(struct mmc_host *mmc, struct mmc_command *cmd)
+{
+	struct meson_host *host = mmc_priv(mmc);
+	struct mmc_data *data = cmd->data;
+	u32 cmd_cfg = 0, cmd_data = 0;
+	unsigned int xfer_bytes = 0;
+
+	/* Setup descriptors */
+	dma_rmb();
+
+	host->cmd = cmd;
+
+	cmd_cfg |= FIELD_PREP(CMD_CFG_CMD_INDEX_MASK, cmd->opcode);
+	cmd_cfg |= CMD_CFG_OWNER;  /* owned by CPU */
+
+	meson_mmc_set_response_bits(cmd, &cmd_cfg);
+
+	/* data? */
+	if (data) {
+		data->bytes_xfered = 0;
+		cmd_cfg |= CMD_CFG_DATA_IO;
+		cmd_cfg |= FIELD_PREP(CMD_CFG_TIMEOUT_MASK,
+				      ilog2(meson_mmc_get_timeout_msecs(data)));
+
+		if (meson_mmc_desc_chain_mode(data)) {
+			meson_mmc_desc_chain_transfer(mmc, cmd_cfg);
+			return;
+		}
+
+		if (data->blocks > 1) {
+			cmd_cfg |= CMD_CFG_BLOCK_MODE;
+			cmd_cfg |= FIELD_PREP(CMD_CFG_LENGTH_MASK,
+					      data->blocks);
+			meson_mmc_set_blksz(mmc, data->blksz);
+		} else {
+			cmd_cfg |= FIELD_PREP(CMD_CFG_LENGTH_MASK, data->blksz);
+		}
+
+		xfer_bytes = data->blksz * data->blocks;
+		if (data->flags & MMC_DATA_WRITE) {
+			cmd_cfg |= CMD_CFG_DATA_WR;
+			WARN_ON(xfer_bytes > host->bounce_buf_size);
+			if (host->dram_access_quirk)
+				meson_mmc_copy_buffer(host, data, xfer_bytes, true);
+			else
+				sg_copy_to_buffer(data->sg, data->sg_len,
+						  host->bounce_buf, xfer_bytes);
+			dma_wmb();
+		}
+
+		cmd_data = host->bounce_dma_addr & CMD_DATA_MASK;
+	} else {
+		cmd_cfg |= FIELD_PREP(CMD_CFG_TIMEOUT_MASK,
+				      ilog2(SD_EMMC_CMD_TIMEOUT));
+	}
+
+	/* Last descriptor */
+	cmd_cfg |= CMD_CFG_END_OF_CHAIN;
+	writel(cmd_cfg, host->regs + SD_EMMC_CMD_CFG);
+	writel(cmd_data, host->regs + SD_EMMC_CMD_DAT);
+	writel(0, host->regs + SD_EMMC_CMD_RSP);
+	wmb(); /* ensure descriptor is written before kicked */
+	writel(cmd->arg, host->regs + SD_EMMC_CMD_ARG);
+}
+
+static int meson_mmc_validate_dram_access(struct mmc_host *mmc, struct mmc_data *data)
+{
+	struct scatterlist *sg;
+	int i;
+
+	/* Reject request if any element offset or size is not 32bit aligned */
+	for_each_sg(data->sg, sg, data->sg_len, i) {
+		if (!IS_ALIGNED(sg->offset, sizeof(u32)) ||
+		    !IS_ALIGNED(sg->length, sizeof(u32))) {
+			dev_err(mmc_dev(mmc), "unaligned sg offset %u len %u\n",
+				data->sg->offset, data->sg->length);
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static void meson_mmc_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct meson_host *host = mmc_priv(mmc);
+	host->needs_pre_post_req = mrq->data &&
+			!(mrq->data->host_cookie & SD_EMMC_PRE_REQ_DONE);
+
+	/*
+	 * The memory at the end of the controller used as bounce buffer for
+	 * the dram_access_quirk only accepts 32bit read/write access,
+	 * check the aligment and length of the data before starting the request.
+	 */
+	if (host->dram_access_quirk && mrq->data) {
+		mrq->cmd->error = meson_mmc_validate_dram_access(mmc, mrq->data);
+		if (mrq->cmd->error) {
+			mmc_request_done(mmc, mrq);
+			return;
+		}
+	}
+
+	if (host->needs_pre_post_req) {
+		meson_mmc_get_transfer_mode(mmc, mrq);
+		if (!meson_mmc_desc_chain_mode(mrq->data))
+			host->needs_pre_post_req = false;
+	}
+
+	if (host->needs_pre_post_req)
+		meson_mmc_pre_req(mmc, mrq);
+
+	/* Stop execution */
+	writel(0, host->regs + SD_EMMC_START);
+
+	meson_mmc_start_cmd(mmc, mrq->sbc ?: mrq->cmd);
+}
+
+static void meson_mmc_read_resp(struct mmc_host *mmc, struct mmc_command *cmd)
+{
+	struct meson_host *host = mmc_priv(mmc);
+
+	if (cmd->flags & MMC_RSP_136) {
+		cmd->resp[0] = readl(host->regs + SD_EMMC_CMD_RSP3);
+		cmd->resp[1] = readl(host->regs + SD_EMMC_CMD_RSP2);
+		cmd->resp[2] = readl(host->regs + SD_EMMC_CMD_RSP1);
+		cmd->resp[3] = readl(host->regs + SD_EMMC_CMD_RSP);
+	} else if (cmd->flags & MMC_RSP_PRESENT) {
+		cmd->resp[0] = readl(host->regs + SD_EMMC_CMD_RSP);
+	}
+}
+
+static irqreturn_t meson_mmc_irq(int irq, void *dev_id)
+{
+	struct meson_host *host = dev_id;
+	struct mmc_command *cmd;
+	struct mmc_data *data;
+	u32 irq_en, status, raw_status;
+	irqreturn_t ret = IRQ_NONE;
+
+	irq_en = readl(host->regs + SD_EMMC_IRQ_EN);
+	raw_status = readl(host->regs + SD_EMMC_STATUS);
+	status = raw_status & irq_en;
+
+	if (!status) {
+		dev_dbg(host->dev,
+			"Unexpected IRQ! irq_en 0x%08x - status 0x%08x\n",
+			 irq_en, raw_status);
+		return IRQ_NONE;
+	}
+
+	if (WARN_ON(!host) || WARN_ON(!host->cmd))
+		return IRQ_NONE;
+
+	/* ack all raised interrupts */
+	writel(status, host->regs + SD_EMMC_STATUS);
+
+	cmd = host->cmd;
+	data = cmd->data;
+	cmd->error = 0;
+	if (status & IRQ_CRC_ERR) {
+		dev_dbg(host->dev, "CRC Error - status 0x%08x\n", status);
+		cmd->error = -EILSEQ;
+		ret = IRQ_WAKE_THREAD;
+		goto out;
+	}
+
+	if (status & IRQ_TIMEOUTS) {
+		dev_dbg(host->dev, "Timeout - status 0x%08x\n", status);
+		cmd->error = -ETIMEDOUT;
+		ret = IRQ_WAKE_THREAD;
+		goto out;
+	}
+
+	meson_mmc_read_resp(host->mmc, cmd);
+
+	if (status & IRQ_SDIO) {
+		dev_dbg(host->dev, "IRQ: SDIO TODO.\n");
+		ret = IRQ_HANDLED;
+	}
+
+	if (status & (IRQ_END_OF_CHAIN | IRQ_RESP_STATUS)) {
+		if (data && !cmd->error)
+			data->bytes_xfered = data->blksz * data->blocks;
+
+		return IRQ_WAKE_THREAD;
+	}
+
+out:
+	if (cmd->error) {
+		/* Stop desc in case of errors */
+		u32 start = readl(host->regs + SD_EMMC_START);
+
+		start &= ~START_DESC_BUSY;
+		writel(start, host->regs + SD_EMMC_START);
+	}
+
+	return ret;
+}
+
+static int meson_mmc_wait_desc_stop(struct meson_host *host)
+{
+	u32 status;
+
+	/*
+	 * It may sometimes take a while for it to actually halt. Here, we
+	 * are giving it 5ms to comply
+	 *
+	 * If we don't confirm the descriptor is stopped, it might raise new
+	 * IRQs after we have called mmc_request_done() which is bad.
+	 */
+
+	return readl_poll_timeout(host->regs + SD_EMMC_STATUS, status,
+				  !(status & (STATUS_BUSY | STATUS_DESC_BUSY)),
+				  100, 5000);
+}
+
+static irqreturn_t meson_mmc_irq_thread(int irq, void *dev_id)
+{
+	struct meson_host *host = dev_id;
+	struct mmc_command *next_cmd, *cmd = host->cmd;
+	struct mmc_data *data;
+	unsigned int xfer_bytes;
+
+	if (WARN_ON(!cmd))
+		return IRQ_NONE;
+
+	if (cmd->error) {
+		meson_mmc_wait_desc_stop(host);
+		meson_mmc_request_done(host->mmc, cmd->mrq);
+
+		return IRQ_HANDLED;
+	}
+
+	data = cmd->data;
+	if (meson_mmc_bounce_buf_read(data)) {
+		xfer_bytes = data->blksz * data->blocks;
+		WARN_ON(xfer_bytes > host->bounce_buf_size);
+		if (host->dram_access_quirk)
+			meson_mmc_copy_buffer(host, data, xfer_bytes, false);
+		else
+			sg_copy_from_buffer(data->sg, data->sg_len,
+					    host->bounce_buf, xfer_bytes);
+	}
+
+	next_cmd = meson_mmc_get_next_command(cmd);
+	if (next_cmd)
+		meson_mmc_start_cmd(host->mmc, next_cmd);
+	else
+		meson_mmc_request_done(host->mmc, cmd->mrq);
+
+	return IRQ_HANDLED;
+}
+
+/*
+ * NOTE: we only need this until the GPIO/pinctrl driver can handle
+ * interrupts.  For now, the MMC core will use this for polling.
+ */
+static int meson_mmc_get_cd(struct mmc_host *mmc)
+{
+	int status = mmc_gpio_get_cd(mmc);
+
+	if (status == -ENOSYS)
+		return 1; /* assume present */
+
+	return status;
+}
+
+static void meson_mmc_cfg_init(struct meson_host *host)
+{
+	u32 cfg = 0;
+
+	cfg |= FIELD_PREP(CFG_RESP_TIMEOUT_MASK,
+			  ilog2(SD_EMMC_CFG_RESP_TIMEOUT));
+	cfg |= FIELD_PREP(CFG_RC_CC_MASK, ilog2(SD_EMMC_CFG_CMD_GAP));
+	cfg |= FIELD_PREP(CFG_BLK_LEN_MASK, ilog2(SD_EMMC_CFG_BLK_SIZE));
+
+	/* abort chain on R/W errors */
+	cfg |= CFG_ERR_ABORT;
+
+	writel(cfg, host->regs + SD_EMMC_CFG);
+}
+
+static int meson_mmc_card_busy(struct mmc_host *mmc)
+{
+	struct meson_host *host = mmc_priv(mmc);
+	u32 regval;
+
+	regval = readl(host->regs + SD_EMMC_STATUS);
+
+	/* We are only interrested in lines 0 to 3, so mask the other ones */
+	return !(FIELD_GET(STATUS_DATI, regval) & 0xf);
+}
+
+static int meson_mmc_voltage_switch(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	int ret;
+
+	/* vqmmc regulator is available */
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+		/*
+		 * The usual amlogic setup uses a GPIO to switch from one
+		 * regulator to the other. While the voltage ramp up is
+		 * pretty fast, care must be taken when switching from 3.3v
+		 * to 1.8v. Please make sure the regulator framework is aware
+		 * of your own regulator constraints
+		 */
+		ret = mmc_regulator_set_vqmmc(mmc, ios);
+		return ret < 0 ? ret : 0;
+	}
+
+	/* no vqmmc regulator, assume fixed regulator at 3/3.3V */
+	if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_330)
+		return 0;
+
+	return -EINVAL;
+}
+
+static const struct mmc_host_ops meson_mmc_ops = {
+	.request	= meson_mmc_request,
+	.set_ios	= meson_mmc_set_ios,
+	.get_cd         = meson_mmc_get_cd,
+	.pre_req	= meson_mmc_pre_req,
+	.post_req	= meson_mmc_post_req,
+	.execute_tuning = meson_mmc_resampling_tuning,
+	.card_busy	= meson_mmc_card_busy,
+	.start_signal_voltage_switch = meson_mmc_voltage_switch,
+};
+
+static int meson_mmc_probe(struct platform_device *pdev)
+{
+	struct resource *res;
+	struct meson_host *host;
+	struct mmc_host *mmc;
+	int ret;
+
+	mmc = devm_mmc_alloc_host(&pdev->dev, sizeof(struct meson_host));
+	if (!mmc)
+		return -ENOMEM;
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+	host->dev = &pdev->dev;
+	dev_set_drvdata(&pdev->dev, host);
+
+	/* The G12A SDIO Controller needs an SRAM bounce buffer */
+	host->dram_access_quirk = device_property_read_bool(&pdev->dev,
+					"amlogic,dram-access-quirk");
+
+	/* Get regulators and the supported OCR mask */
+	host->vqmmc_enabled = false;
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret)
+		return ret;
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		return dev_err_probe(&pdev->dev, ret, "error parsing DT\n");
+
+	host->data = (struct meson_mmc_data *)
+		of_device_get_match_data(&pdev->dev);
+	if (!host->data)
+		return -EINVAL;
+
+	ret = device_reset_optional(&pdev->dev);
+	if (ret)
+		return dev_err_probe(&pdev->dev, ret, "device reset failed\n");
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	host->regs = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(host->regs))
+		return PTR_ERR(host->regs);
+
+	host->irq = platform_get_irq(pdev, 0);
+	if (host->irq < 0)
+		return host->irq;
+
+	host->pinctrl = devm_pinctrl_get(&pdev->dev);
+	if (IS_ERR(host->pinctrl))
+		return PTR_ERR(host->pinctrl);
+
+	host->pins_clk_gate = pinctrl_lookup_state(host->pinctrl,
+						   "clk-gate");
+	if (IS_ERR(host->pins_clk_gate)) {
+		dev_warn(&pdev->dev,
+			 "can't get clk-gate pinctrl, using clk_stop bit\n");
+		host->pins_clk_gate = NULL;
+	}
+
+	host->core_clk = devm_clk_get(&pdev->dev, "core");
+	if (IS_ERR(host->core_clk))
+		return PTR_ERR(host->core_clk);
+
+	ret = clk_prepare_enable(host->core_clk);
+	if (ret)
+		return ret;
+
+	ret = meson_mmc_clk_init(host);
+	if (ret)
+		goto err_core_clk;
+
+	/* set config to sane default */
+	meson_mmc_cfg_init(host);
+
+	/* Stop execution */
+	writel(0, host->regs + SD_EMMC_START);
+
+	/* clear, ack and enable interrupts */
+	writel(0, host->regs + SD_EMMC_IRQ_EN);
+	writel(IRQ_CRC_ERR | IRQ_TIMEOUTS | IRQ_END_OF_CHAIN,
+	       host->regs + SD_EMMC_STATUS);
+	writel(IRQ_CRC_ERR | IRQ_TIMEOUTS | IRQ_END_OF_CHAIN,
+	       host->regs + SD_EMMC_IRQ_EN);
+
+	ret = request_threaded_irq(host->irq, meson_mmc_irq,
+				   meson_mmc_irq_thread, IRQF_ONESHOT,
+				   dev_name(&pdev->dev), host);
+	if (ret)
+		goto err_init_clk;
+
+	mmc->caps |= MMC_CAP_CMD23;
+	if (host->dram_access_quirk) {
+		/* Limit segments to 1 due to low available sram memory */
+		mmc->max_segs = 1;
+		/* Limit to the available sram memory */
+		mmc->max_blk_count = SD_EMMC_SRAM_DATA_BUF_LEN /
+				     mmc->max_blk_size;
+	} else {
+		mmc->max_blk_count = CMD_CFG_LENGTH_MASK;
+		mmc->max_segs = SD_EMMC_DESC_BUF_LEN /
+				sizeof(struct sd_emmc_desc);
+	}
+	mmc->max_req_size = mmc->max_blk_count * mmc->max_blk_size;
+	mmc->max_seg_size = mmc->max_req_size;
+
+	/*
+	 * At the moment, we don't know how to reliably enable HS400.
+	 * From the different datasheets, it is not even clear if this mode
+	 * is officially supported by any of the SoCs
+	 */
+	mmc->caps2 &= ~MMC_CAP2_HS400;
+
+	if (host->dram_access_quirk) {
+		/*
+		 * The MMC Controller embeds 1,5KiB of internal SRAM
+		 * that can be used to be used as bounce buffer.
+		 * In the case of the G12A SDIO controller, use these
+		 * instead of the DDR memory
+		 */
+		host->bounce_buf_size = SD_EMMC_SRAM_DATA_BUF_LEN;
+		host->bounce_iomem_buf = host->regs + SD_EMMC_SRAM_DATA_BUF_OFF;
+		host->bounce_dma_addr = res->start + SD_EMMC_SRAM_DATA_BUF_OFF;
+	} else {
+		/* data bounce buffer */
+		host->bounce_buf_size = mmc->max_req_size;
+		host->bounce_buf =
+			dma_alloc_coherent(host->dev, host->bounce_buf_size,
+					   &host->bounce_dma_addr, GFP_KERNEL);
+		if (host->bounce_buf == NULL) {
+			dev_err(host->dev, "Unable to map allocate DMA bounce buffer.\n");
+			ret = -ENOMEM;
+			goto err_free_irq;
+		}
+	}
+
+	host->descs = dma_alloc_coherent(host->dev, SD_EMMC_DESC_BUF_LEN,
+		      &host->descs_dma_addr, GFP_KERNEL);
+	if (!host->descs) {
+		dev_err(host->dev, "Allocating descriptor DMA buffer failed\n");
+		ret = -ENOMEM;
+		goto err_bounce_buf;
+	}
+
+	mmc->ops = &meson_mmc_ops;
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto err_free_irq;
+
+	return 0;
+
+err_bounce_buf:
+	if (!host->dram_access_quirk)
+		dma_free_coherent(host->dev, host->bounce_buf_size,
+				  host->bounce_buf, host->bounce_dma_addr);
+err_free_irq:
+	free_irq(host->irq, host);
+err_init_clk:
+	clk_disable_unprepare(host->mmc_clk);
+err_core_clk:
+	clk_disable_unprepare(host->core_clk);
+	return ret;
+}
+
+static int meson_mmc_remove(struct platform_device *pdev)
+{
+	struct meson_host *host = dev_get_drvdata(&pdev->dev);
+
+	mmc_remove_host(host->mmc);
+
+	/* disable interrupts */
+	writel(0, host->regs + SD_EMMC_IRQ_EN);
+	free_irq(host->irq, host);
+
+	dma_free_coherent(host->dev, SD_EMMC_DESC_BUF_LEN,
+			  host->descs, host->descs_dma_addr);
+
+	if (!host->dram_access_quirk)
+		dma_free_coherent(host->dev, host->bounce_buf_size,
+				  host->bounce_buf, host->bounce_dma_addr);
+
+	clk_disable_unprepare(host->mmc_clk);
+	clk_disable_unprepare(host->core_clk);
+
+	return 0;
+}
+
+static const struct meson_mmc_data meson_gx_data = {
+	.tx_delay_mask	= CLK_V2_TX_DELAY_MASK,
+	.rx_delay_mask	= CLK_V2_RX_DELAY_MASK,
+	.always_on	= CLK_V2_ALWAYS_ON,
+	.adjust		= SD_EMMC_ADJUST,
+};
+
+static const struct meson_mmc_data meson_axg_data = {
+	.tx_delay_mask	= CLK_V3_TX_DELAY_MASK,
+	.rx_delay_mask	= CLK_V3_RX_DELAY_MASK,
+	.always_on	= CLK_V3_ALWAYS_ON,
+	.adjust		= SD_EMMC_V3_ADJUST,
+};
+
+static const struct of_device_id meson_mmc_of_match[] = {
+	{ .compatible = "amlogic,meson-gx-mmc",		.data = &meson_gx_data },
+	{ .compatible = "amlogic,meson-gxbb-mmc", 	.data = &meson_gx_data },
+	{ .compatible = "amlogic,meson-gxl-mmc",	.data = &meson_gx_data },
+	{ .compatible = "amlogic,meson-gxm-mmc",	.data = &meson_gx_data },
+	{ .compatible = "amlogic,meson-axg-mmc",	.data = &meson_axg_data },
+	{}
+};
+MODULE_DEVICE_TABLE(of, meson_mmc_of_match);
+
+static struct platform_driver meson_mmc_driver = {
+	.probe		= meson_mmc_probe,
+	.remove		= meson_mmc_remove,
+	.driver		= {
+		.name = DRIVER_NAME,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(meson_mmc_of_match),
+	},
+};
+
+module_platform_driver(meson_mmc_driver);
+
+MODULE_DESCRIPTION("Amlogic S905*/GX*/AXG SD/eMMC driver");
+MODULE_AUTHOR("Kevin Hilman <khilman@baylibre.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/meson-mx-sdhc-clkc.c b/drivers/mmc/host/meson-mx-sdhc-clkc.c
new file mode 100644
index 000000000..e1f29b279
--- /dev/null
+++ b/drivers/mmc/host/meson-mx-sdhc-clkc.c
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Amlogic Meson SDHC clock controller
+ *
+ * Copyright (C) 2020 Martin Blumenstingl <martin.blumenstingl@googlemail.com>
+ */
+
+#include <linux/clk.h>
+#include <linux/clk-provider.h>
+#include <linux/device.h>
+#include <linux/platform_device.h>
+
+#include "meson-mx-sdhc.h"
+
+#define MESON_SDHC_NUM_BUILTIN_CLKS	6
+
+struct meson_mx_sdhc_clkc {
+	struct clk_mux			src_sel;
+	struct clk_divider		div;
+	struct clk_gate			mod_clk_en;
+	struct clk_gate			tx_clk_en;
+	struct clk_gate			rx_clk_en;
+	struct clk_gate			sd_clk_en;
+};
+
+static const struct clk_parent_data meson_mx_sdhc_src_sel_parents[4] = {
+	{ .fw_name = "clkin0" },
+	{ .fw_name = "clkin1" },
+	{ .fw_name = "clkin2" },
+	{ .fw_name = "clkin3" },
+};
+
+static const struct clk_div_table meson_mx_sdhc_div_table[] = {
+	{ .div = 6, .val = 5, },
+	{ .div = 8, .val = 7, },
+	{ .div = 9, .val = 8, },
+	{ .div = 10, .val = 9, },
+	{ .div = 12, .val = 11, },
+	{ .div = 16, .val = 15, },
+	{ .div = 18, .val = 17, },
+	{ .div = 34, .val = 33, },
+	{ .div = 142, .val = 141, },
+	{ .div = 850, .val = 849, },
+	{ .div = 2126, .val = 2125, },
+	{ .div = 4096, .val = 4095, },
+	{ /* sentinel */ }
+};
+
+static int meson_mx_sdhc_clk_hw_register(struct device *dev,
+					 const char *name_suffix,
+					 const struct clk_parent_data *parents,
+					 unsigned int num_parents,
+					 const struct clk_ops *ops,
+					 struct clk_hw *hw)
+{
+	struct clk_init_data init = { };
+	char clk_name[32];
+
+	snprintf(clk_name, sizeof(clk_name), "%s#%s", dev_name(dev),
+		 name_suffix);
+
+	init.name = clk_name;
+	init.ops = ops;
+	init.flags = CLK_SET_RATE_PARENT;
+	init.parent_data = parents;
+	init.num_parents = num_parents;
+
+	hw->init = &init;
+
+	return devm_clk_hw_register(dev, hw);
+}
+
+static int meson_mx_sdhc_gate_clk_hw_register(struct device *dev,
+					      const char *name_suffix,
+					      struct clk_hw *parent,
+					      struct clk_hw *hw)
+{
+	struct clk_parent_data parent_data = { .hw = parent };
+
+	return meson_mx_sdhc_clk_hw_register(dev, name_suffix, &parent_data, 1,
+					     &clk_gate_ops, hw);
+}
+
+int meson_mx_sdhc_register_clkc(struct device *dev, void __iomem *base,
+				struct clk_bulk_data *clk_bulk_data)
+{
+	struct clk_parent_data div_parent = { };
+	struct meson_mx_sdhc_clkc *clkc_data;
+	int ret;
+
+	clkc_data = devm_kzalloc(dev, sizeof(*clkc_data), GFP_KERNEL);
+	if (!clkc_data)
+		return -ENOMEM;
+
+	clkc_data->src_sel.reg = base + MESON_SDHC_CLKC;
+	clkc_data->src_sel.mask = 0x3;
+	clkc_data->src_sel.shift = 16;
+	ret = meson_mx_sdhc_clk_hw_register(dev, "src_sel",
+					    meson_mx_sdhc_src_sel_parents, 4,
+					    &clk_mux_ops,
+					    &clkc_data->src_sel.hw);
+	if (ret)
+		return ret;
+
+	clkc_data->div.reg = base + MESON_SDHC_CLKC;
+	clkc_data->div.shift = 0;
+	clkc_data->div.width = 12;
+	clkc_data->div.table = meson_mx_sdhc_div_table;
+	div_parent.hw = &clkc_data->src_sel.hw;
+	ret = meson_mx_sdhc_clk_hw_register(dev, "div", &div_parent, 1,
+					    &clk_divider_ops,
+					    &clkc_data->div.hw);
+	if (ret)
+		return ret;
+
+	clkc_data->mod_clk_en.reg = base + MESON_SDHC_CLKC;
+	clkc_data->mod_clk_en.bit_idx = 15;
+	ret = meson_mx_sdhc_gate_clk_hw_register(dev, "mod_clk_on",
+						 &clkc_data->div.hw,
+						 &clkc_data->mod_clk_en.hw);
+	if (ret)
+		return ret;
+
+	clkc_data->tx_clk_en.reg = base + MESON_SDHC_CLKC;
+	clkc_data->tx_clk_en.bit_idx = 14;
+	ret = meson_mx_sdhc_gate_clk_hw_register(dev, "tx_clk_on",
+						 &clkc_data->div.hw,
+						 &clkc_data->tx_clk_en.hw);
+	if (ret)
+		return ret;
+
+	clkc_data->rx_clk_en.reg = base + MESON_SDHC_CLKC;
+	clkc_data->rx_clk_en.bit_idx = 13;
+	ret = meson_mx_sdhc_gate_clk_hw_register(dev, "rx_clk_on",
+						 &clkc_data->div.hw,
+						 &clkc_data->rx_clk_en.hw);
+	if (ret)
+		return ret;
+
+	clkc_data->sd_clk_en.reg = base + MESON_SDHC_CLKC;
+	clkc_data->sd_clk_en.bit_idx = 12;
+	ret = meson_mx_sdhc_gate_clk_hw_register(dev, "sd_clk_on",
+						 &clkc_data->div.hw,
+						 &clkc_data->sd_clk_en.hw);
+	if (ret)
+		return ret;
+
+	/*
+	 * TODO: Replace clk_hw.clk with devm_clk_hw_get_clk() once that is
+	 * available.
+	 */
+	clk_bulk_data[0].clk = clkc_data->mod_clk_en.hw.clk;
+	clk_bulk_data[1].clk = clkc_data->sd_clk_en.hw.clk;
+	clk_bulk_data[2].clk = clkc_data->tx_clk_en.hw.clk;
+	clk_bulk_data[3].clk = clkc_data->rx_clk_en.hw.clk;
+
+	return 0;
+}
diff --git a/drivers/mmc/host/meson-mx-sdhc-mmc.c b/drivers/mmc/host/meson-mx-sdhc-mmc.c
new file mode 100644
index 000000000..ba59061fe
--- /dev/null
+++ b/drivers/mmc/host/meson-mx-sdhc-mmc.c
@@ -0,0 +1,920 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Amlogic Meson6/Meson8/Meson8b/Meson8m2 SDHC MMC host controller driver.
+ *
+ * Copyright (C) 2020 Martin Blumenstingl <martin.blumenstingl@googlemail.com>
+ */
+
+#include <linux/clk.h>
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/interrupt.h>
+#include <linux/iopoll.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/property.h>
+#include <linux/regmap.h>
+#include <linux/regulator/consumer.h>
+#include <linux/types.h>
+
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sdio.h>
+#include <linux/mmc/slot-gpio.h>
+
+#include "meson-mx-sdhc.h"
+
+#define MESON_SDHC_NUM_BULK_CLKS				4
+#define MESON_SDHC_MAX_BLK_SIZE					512
+#define MESON_SDHC_NUM_TUNING_TRIES				10
+
+#define MESON_SDHC_WAIT_CMD_READY_SLEEP_US			1
+#define MESON_SDHC_WAIT_CMD_READY_TIMEOUT_US			100000
+#define MESON_SDHC_WAIT_BEFORE_SEND_SLEEP_US			1
+#define MESON_SDHC_WAIT_BEFORE_SEND_TIMEOUT_US			200
+
+struct meson_mx_sdhc_data {
+	void		(*init_hw)(struct mmc_host *mmc);
+	void		(*set_pdma)(struct mmc_host *mmc);
+	void		(*wait_before_send)(struct mmc_host *mmc);
+	bool		hardware_flush_all_cmds;
+};
+
+struct meson_mx_sdhc_host {
+	struct mmc_host			*mmc;
+
+	struct mmc_request		*mrq;
+	struct mmc_command		*cmd;
+	int				error;
+
+	struct regmap			*regmap;
+
+	struct clk			*pclk;
+	struct clk			*sd_clk;
+	struct clk_bulk_data		bulk_clks[MESON_SDHC_NUM_BULK_CLKS];
+	bool				bulk_clks_enabled;
+
+	const struct meson_mx_sdhc_data	*platform;
+};
+
+static const struct regmap_config meson_mx_sdhc_regmap_config = {
+	.reg_bits = 8,
+	.val_bits = 32,
+	.reg_stride = 4,
+	.max_register = MESON_SDHC_CLK2,
+};
+
+static void meson_mx_sdhc_hw_reset(struct mmc_host *mmc)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+
+	regmap_write(host->regmap, MESON_SDHC_SRST, MESON_SDHC_SRST_MAIN_CTRL |
+		     MESON_SDHC_SRST_RXFIFO | MESON_SDHC_SRST_TXFIFO |
+		     MESON_SDHC_SRST_DPHY_RX | MESON_SDHC_SRST_DPHY_TX |
+		     MESON_SDHC_SRST_DMA_IF);
+	usleep_range(10, 100);
+
+	regmap_write(host->regmap, MESON_SDHC_SRST, 0);
+	usleep_range(10, 100);
+}
+
+static void meson_mx_sdhc_clear_fifo(struct mmc_host *mmc)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+	u32 stat;
+
+	regmap_read(host->regmap, MESON_SDHC_STAT, &stat);
+	if (!FIELD_GET(MESON_SDHC_STAT_RXFIFO_CNT, stat) &&
+	    !FIELD_GET(MESON_SDHC_STAT_TXFIFO_CNT, stat))
+		return;
+
+	regmap_write(host->regmap, MESON_SDHC_SRST, MESON_SDHC_SRST_RXFIFO |
+		     MESON_SDHC_SRST_TXFIFO | MESON_SDHC_SRST_MAIN_CTRL);
+	udelay(5);
+
+	regmap_read(host->regmap, MESON_SDHC_STAT, &stat);
+	if (FIELD_GET(MESON_SDHC_STAT_RXFIFO_CNT, stat) ||
+	    FIELD_GET(MESON_SDHC_STAT_TXFIFO_CNT, stat))
+		dev_warn(mmc_dev(host->mmc),
+			 "Failed to clear FIFOs, RX: %lu, TX: %lu\n",
+			 FIELD_GET(MESON_SDHC_STAT_RXFIFO_CNT, stat),
+			 FIELD_GET(MESON_SDHC_STAT_TXFIFO_CNT, stat));
+}
+
+static void meson_mx_sdhc_wait_cmd_ready(struct mmc_host *mmc)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+	u32 stat, esta;
+	int ret;
+
+	ret = regmap_read_poll_timeout(host->regmap, MESON_SDHC_STAT, stat,
+				       !(stat & MESON_SDHC_STAT_CMD_BUSY),
+				       MESON_SDHC_WAIT_CMD_READY_SLEEP_US,
+				       MESON_SDHC_WAIT_CMD_READY_TIMEOUT_US);
+	if (ret) {
+		dev_warn(mmc_dev(mmc),
+			 "Failed to poll for CMD_BUSY while processing CMD%d\n",
+			 host->cmd->opcode);
+		meson_mx_sdhc_hw_reset(mmc);
+	}
+
+	ret = regmap_read_poll_timeout(host->regmap, MESON_SDHC_ESTA, esta,
+				       !(esta & MESON_SDHC_ESTA_11_13),
+				       MESON_SDHC_WAIT_CMD_READY_SLEEP_US,
+				       MESON_SDHC_WAIT_CMD_READY_TIMEOUT_US);
+	if (ret) {
+		dev_warn(mmc_dev(mmc),
+			 "Failed to poll for ESTA[13:11] while processing CMD%d\n",
+			 host->cmd->opcode);
+		meson_mx_sdhc_hw_reset(mmc);
+	}
+}
+
+static void meson_mx_sdhc_start_cmd(struct mmc_host *mmc,
+				    struct mmc_command *cmd)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+	bool manual_stop = false;
+	u32 ictl, send;
+	int pack_len;
+
+	host->cmd = cmd;
+
+	ictl = MESON_SDHC_ICTL_DATA_TIMEOUT | MESON_SDHC_ICTL_DATA_ERR_CRC |
+	       MESON_SDHC_ICTL_RXFIFO_FULL | MESON_SDHC_ICTL_TXFIFO_EMPTY |
+	       MESON_SDHC_ICTL_RESP_TIMEOUT | MESON_SDHC_ICTL_RESP_ERR_CRC;
+
+	send = FIELD_PREP(MESON_SDHC_SEND_CMD_INDEX, cmd->opcode);
+
+	if (cmd->data) {
+		send |= MESON_SDHC_SEND_CMD_HAS_DATA;
+		send |= FIELD_PREP(MESON_SDHC_SEND_TOTAL_PACK,
+				   cmd->data->blocks - 1);
+
+		if (cmd->data->blksz < MESON_SDHC_MAX_BLK_SIZE)
+			pack_len = cmd->data->blksz;
+		else
+			pack_len = 0;
+
+		if (cmd->data->flags & MMC_DATA_WRITE)
+			send |= MESON_SDHC_SEND_DATA_DIR;
+
+		/*
+		 * If command with no data, just wait response done
+		 * interrupt(int[0]), and if command with data transfer, just
+		 * wait dma done interrupt(int[11]), don't need care about
+		 * dat0 busy or not.
+		 */
+		if (host->platform->hardware_flush_all_cmds ||
+		    cmd->data->flags & MMC_DATA_WRITE)
+			/* hardware flush: */
+			ictl |= MESON_SDHC_ICTL_DMA_DONE;
+		else
+			/* software flush: */
+			ictl |= MESON_SDHC_ICTL_DATA_XFER_OK;
+
+		/*
+		 * Mimic the logic from the vendor driver where (only)
+		 * SD_IO_RW_EXTENDED commands with more than one block set the
+		 * MESON_SDHC_MISC_MANUAL_STOP bit. This fixes the firmware
+		 * download in the brcmfmac driver for a BCM43362/1 card.
+		 * Without this sdio_memcpy_toio() (with a size of 219557
+		 * bytes) times out if MESON_SDHC_MISC_MANUAL_STOP is not set.
+		 */
+		manual_stop = cmd->data->blocks > 1 &&
+			      cmd->opcode == SD_IO_RW_EXTENDED;
+	} else {
+		pack_len = 0;
+
+		ictl |= MESON_SDHC_ICTL_RESP_OK;
+	}
+
+	regmap_update_bits(host->regmap, MESON_SDHC_MISC,
+			   MESON_SDHC_MISC_MANUAL_STOP,
+			   manual_stop ? MESON_SDHC_MISC_MANUAL_STOP : 0);
+
+	if (cmd->opcode == MMC_STOP_TRANSMISSION)
+		send |= MESON_SDHC_SEND_DATA_STOP;
+
+	if (cmd->flags & MMC_RSP_PRESENT)
+		send |= MESON_SDHC_SEND_CMD_HAS_RESP;
+
+	if (cmd->flags & MMC_RSP_136) {
+		send |= MESON_SDHC_SEND_RESP_LEN;
+		send |= MESON_SDHC_SEND_RESP_NO_CRC;
+	}
+
+	if (!(cmd->flags & MMC_RSP_CRC))
+		send |= MESON_SDHC_SEND_RESP_NO_CRC;
+
+	if (cmd->flags & MMC_RSP_BUSY)
+		send |= MESON_SDHC_SEND_R1B;
+
+	/* enable the new IRQs and mask all pending ones */
+	regmap_write(host->regmap, MESON_SDHC_ICTL, ictl);
+	regmap_write(host->regmap, MESON_SDHC_ISTA, MESON_SDHC_ISTA_ALL_IRQS);
+
+	regmap_write(host->regmap, MESON_SDHC_ARGU, cmd->arg);
+
+	regmap_update_bits(host->regmap, MESON_SDHC_CTRL,
+			   MESON_SDHC_CTRL_PACK_LEN,
+			   FIELD_PREP(MESON_SDHC_CTRL_PACK_LEN, pack_len));
+
+	if (cmd->data)
+		regmap_write(host->regmap, MESON_SDHC_ADDR,
+			     sg_dma_address(cmd->data->sg));
+
+	meson_mx_sdhc_wait_cmd_ready(mmc);
+
+	if (cmd->data)
+		host->platform->set_pdma(mmc);
+
+	if (host->platform->wait_before_send)
+		host->platform->wait_before_send(mmc);
+
+	regmap_write(host->regmap, MESON_SDHC_SEND, send);
+}
+
+static void meson_mx_sdhc_disable_clks(struct mmc_host *mmc)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+
+	if (!host->bulk_clks_enabled)
+		return;
+
+	clk_bulk_disable_unprepare(MESON_SDHC_NUM_BULK_CLKS, host->bulk_clks);
+
+	host->bulk_clks_enabled = false;
+}
+
+static int meson_mx_sdhc_enable_clks(struct mmc_host *mmc)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+	int ret;
+
+	if (host->bulk_clks_enabled)
+		return 0;
+
+	ret = clk_bulk_prepare_enable(MESON_SDHC_NUM_BULK_CLKS,
+				      host->bulk_clks);
+	if (ret)
+		return ret;
+
+	host->bulk_clks_enabled = true;
+
+	return 0;
+}
+
+static int meson_mx_sdhc_set_clk(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+	u32 val, rx_clk_phase;
+	int ret;
+
+	meson_mx_sdhc_disable_clks(mmc);
+
+	if (ios->clock) {
+		ret = clk_set_rate(host->sd_clk, ios->clock);
+		if (ret) {
+			dev_warn(mmc_dev(mmc),
+				 "Failed to set MMC clock to %uHz: %d\n",
+				 ios->clock, host->error);
+			return ret;
+		}
+
+		ret = meson_mx_sdhc_enable_clks(mmc);
+		if (ret)
+			return ret;
+
+		mmc->actual_clock = clk_get_rate(host->sd_clk);
+
+		/*
+		 * Phase 90 should work in most cases. For data transmission,
+		 * meson_mx_sdhc_execute_tuning() will find a accurate value
+		 */
+		regmap_read(host->regmap, MESON_SDHC_CLKC, &val);
+		rx_clk_phase = FIELD_GET(MESON_SDHC_CLKC_CLK_DIV, val) / 4;
+		regmap_update_bits(host->regmap, MESON_SDHC_CLK2,
+				   MESON_SDHC_CLK2_RX_CLK_PHASE,
+				   FIELD_PREP(MESON_SDHC_CLK2_RX_CLK_PHASE,
+					      rx_clk_phase));
+	} else {
+		mmc->actual_clock = 0;
+	}
+
+	return 0;
+}
+
+static void meson_mx_sdhc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+	unsigned short vdd = ios->vdd;
+
+	switch (ios->power_mode) {
+	case MMC_POWER_OFF:
+		vdd = 0;
+		fallthrough;
+
+	case MMC_POWER_UP:
+		if (!IS_ERR(mmc->supply.vmmc)) {
+			host->error = mmc_regulator_set_ocr(mmc,
+							    mmc->supply.vmmc,
+							    vdd);
+			if (host->error)
+				return;
+		}
+
+		break;
+
+	case MMC_POWER_ON:
+		break;
+	}
+
+	host->error = meson_mx_sdhc_set_clk(mmc, ios);
+	if (host->error)
+		return;
+
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_1:
+		regmap_update_bits(host->regmap, MESON_SDHC_CTRL,
+				   MESON_SDHC_CTRL_DAT_TYPE,
+				   FIELD_PREP(MESON_SDHC_CTRL_DAT_TYPE, 0));
+		break;
+
+	case MMC_BUS_WIDTH_4:
+		regmap_update_bits(host->regmap, MESON_SDHC_CTRL,
+				   MESON_SDHC_CTRL_DAT_TYPE,
+				   FIELD_PREP(MESON_SDHC_CTRL_DAT_TYPE, 1));
+		break;
+
+	case MMC_BUS_WIDTH_8:
+		regmap_update_bits(host->regmap, MESON_SDHC_CTRL,
+				   MESON_SDHC_CTRL_DAT_TYPE,
+				   FIELD_PREP(MESON_SDHC_CTRL_DAT_TYPE, 2));
+		break;
+
+	default:
+		dev_err(mmc_dev(mmc), "unsupported bus width: %d\n",
+			ios->bus_width);
+		host->error = -EINVAL;
+		return;
+	}
+}
+
+static int meson_mx_sdhc_map_dma(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct mmc_data *data = mrq->data;
+	int dma_len;
+
+	if (!data)
+		return 0;
+
+	dma_len = dma_map_sg(mmc_dev(mmc), data->sg, data->sg_len,
+			     mmc_get_dma_dir(data));
+	if (dma_len <= 0) {
+		dev_err(mmc_dev(mmc), "dma_map_sg failed\n");
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static void meson_mx_sdhc_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+	struct mmc_command *cmd = mrq->cmd;
+
+	if (!host->error)
+		host->error = meson_mx_sdhc_map_dma(mmc, mrq);
+
+	if (host->error) {
+		cmd->error = host->error;
+		mmc_request_done(mmc, mrq);
+		return;
+	}
+
+	host->mrq = mrq;
+
+	meson_mx_sdhc_start_cmd(mmc, mrq->cmd);
+}
+
+static int meson_mx_sdhc_card_busy(struct mmc_host *mmc)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+	u32 stat;
+
+	regmap_read(host->regmap, MESON_SDHC_STAT, &stat);
+	return FIELD_GET(MESON_SDHC_STAT_DAT3_0, stat) == 0;
+}
+
+static bool meson_mx_sdhc_tuning_point_matches(struct mmc_host *mmc,
+					       u32 opcode)
+{
+	unsigned int i, num_matches = 0;
+	int ret;
+
+	for (i = 0; i < MESON_SDHC_NUM_TUNING_TRIES; i++) {
+		ret = mmc_send_tuning(mmc, opcode, NULL);
+		if (!ret)
+			num_matches++;
+	}
+
+	return num_matches == MESON_SDHC_NUM_TUNING_TRIES;
+}
+
+static int meson_mx_sdhc_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+	int div, start, len, best_start, best_len;
+	int curr_phase, old_phase, new_phase;
+	u32 val;
+
+	len = 0;
+	start = 0;
+	best_len = 0;
+
+	regmap_read(host->regmap, MESON_SDHC_CLK2, &val);
+	old_phase = FIELD_GET(MESON_SDHC_CLK2_RX_CLK_PHASE, val);
+
+	regmap_read(host->regmap, MESON_SDHC_CLKC, &val);
+	div = FIELD_GET(MESON_SDHC_CLKC_CLK_DIV, val);
+
+	for (curr_phase = 0; curr_phase <= div; curr_phase++) {
+		regmap_update_bits(host->regmap, MESON_SDHC_CLK2,
+				   MESON_SDHC_CLK2_RX_CLK_PHASE,
+				   FIELD_PREP(MESON_SDHC_CLK2_RX_CLK_PHASE,
+					      curr_phase));
+
+		if (meson_mx_sdhc_tuning_point_matches(mmc, opcode)) {
+			if (!len) {
+				start = curr_phase;
+
+				dev_dbg(mmc_dev(mmc),
+					"New RX phase window starts at %u\n",
+					start);
+			}
+
+			len++;
+		} else {
+			if (len > best_len) {
+				best_start = start;
+				best_len = len;
+
+				dev_dbg(mmc_dev(mmc),
+					"New best RX phase window: %u - %u\n",
+					best_start, best_start + best_len);
+			}
+
+			/* reset the current window */
+			len = 0;
+		}
+	}
+
+	if (len > best_len)
+		/* the last window is the best (or possibly only) window */
+		new_phase = start + (len / 2);
+	else if (best_len)
+		/* there was a better window than the last */
+		new_phase = best_start + (best_len / 2);
+	else
+		/* no window was found at all, reset to the original phase */
+		new_phase = old_phase;
+
+	regmap_update_bits(host->regmap, MESON_SDHC_CLK2,
+			   MESON_SDHC_CLK2_RX_CLK_PHASE,
+			   FIELD_PREP(MESON_SDHC_CLK2_RX_CLK_PHASE,
+				      new_phase));
+
+	if (!len && !best_len)
+		return -EIO;
+
+	dev_dbg(mmc_dev(mmc), "Tuned RX clock phase to %u\n", new_phase);
+
+	return 0;
+}
+
+static const struct mmc_host_ops meson_mx_sdhc_ops = {
+	.hw_reset			= meson_mx_sdhc_hw_reset,
+	.request			= meson_mx_sdhc_request,
+	.set_ios			= meson_mx_sdhc_set_ios,
+	.card_busy			= meson_mx_sdhc_card_busy,
+	.execute_tuning			= meson_mx_sdhc_execute_tuning,
+	.get_cd				= mmc_gpio_get_cd,
+	.get_ro				= mmc_gpio_get_ro,
+};
+
+static void meson_mx_sdhc_request_done(struct meson_mx_sdhc_host *host)
+{
+	struct mmc_request *mrq = host->mrq;
+	struct mmc_host *mmc = host->mmc;
+
+	/* disable interrupts and mask all pending ones */
+	regmap_update_bits(host->regmap, MESON_SDHC_ICTL,
+			   MESON_SDHC_ICTL_ALL_IRQS, 0);
+	regmap_update_bits(host->regmap, MESON_SDHC_ISTA,
+			   MESON_SDHC_ISTA_ALL_IRQS, MESON_SDHC_ISTA_ALL_IRQS);
+
+	host->mrq = NULL;
+	host->cmd = NULL;
+
+	mmc_request_done(mmc, mrq);
+}
+
+static u32 meson_mx_sdhc_read_response(struct meson_mx_sdhc_host *host, u8 idx)
+{
+	u32 val;
+
+	regmap_update_bits(host->regmap, MESON_SDHC_PDMA,
+			   MESON_SDHC_PDMA_DMA_MODE, 0);
+
+	regmap_update_bits(host->regmap, MESON_SDHC_PDMA,
+			   MESON_SDHC_PDMA_PIO_RDRESP,
+			   FIELD_PREP(MESON_SDHC_PDMA_PIO_RDRESP, idx));
+
+	regmap_read(host->regmap, MESON_SDHC_ARGU, &val);
+
+	return val;
+}
+
+static irqreturn_t meson_mx_sdhc_irq(int irq, void *data)
+{
+	struct meson_mx_sdhc_host *host = data;
+	struct mmc_command *cmd = host->cmd;
+	u32 ictl, ista;
+
+	regmap_read(host->regmap, MESON_SDHC_ICTL, &ictl);
+	regmap_read(host->regmap, MESON_SDHC_ISTA, &ista);
+
+	if (!(ictl & ista))
+		return IRQ_NONE;
+
+	if (ista & MESON_SDHC_ISTA_RXFIFO_FULL ||
+	    ista & MESON_SDHC_ISTA_TXFIFO_EMPTY)
+		cmd->error = -EIO;
+	else if (ista & MESON_SDHC_ISTA_RESP_ERR_CRC)
+		cmd->error = -EILSEQ;
+	else if (ista & MESON_SDHC_ISTA_RESP_TIMEOUT)
+		cmd->error = -ETIMEDOUT;
+
+	if (cmd->data) {
+		if (ista & MESON_SDHC_ISTA_DATA_ERR_CRC)
+			cmd->data->error = -EILSEQ;
+		else if (ista & MESON_SDHC_ISTA_DATA_TIMEOUT)
+			cmd->data->error = -ETIMEDOUT;
+	}
+
+	if (cmd->error || (cmd->data && cmd->data->error))
+		dev_dbg(mmc_dev(host->mmc), "CMD%d error, ISTA: 0x%08x\n",
+			cmd->opcode, ista);
+
+	return IRQ_WAKE_THREAD;
+}
+
+static irqreturn_t meson_mx_sdhc_irq_thread(int irq, void *irq_data)
+{
+	struct meson_mx_sdhc_host *host = irq_data;
+	struct mmc_command *cmd;
+	u32 val;
+
+	cmd = host->cmd;
+	if (WARN_ON(!cmd))
+		return IRQ_HANDLED;
+
+	if (cmd->data && !cmd->data->error) {
+		if (!host->platform->hardware_flush_all_cmds &&
+		    cmd->data->flags & MMC_DATA_READ) {
+			meson_mx_sdhc_wait_cmd_ready(host->mmc);
+
+			/*
+			 * If MESON_SDHC_PDMA_RXFIFO_MANUAL_FLUSH was
+			 * previously 0x1 then it has to be set to 0x3. If it
+			 * was 0x0 before then it has to be set to 0x2. Without
+			 * this reading SD cards sometimes transfers garbage,
+			 * which results in cards not being detected due to:
+			 *   unrecognised SCR structure version <random number>
+			 */
+			val = FIELD_PREP(MESON_SDHC_PDMA_RXFIFO_MANUAL_FLUSH,
+					 2);
+			regmap_update_bits(host->regmap, MESON_SDHC_PDMA, val,
+					   val);
+		}
+
+		dma_unmap_sg(mmc_dev(host->mmc), cmd->data->sg,
+			     cmd->data->sg_len, mmc_get_dma_dir(cmd->data));
+
+		cmd->data->bytes_xfered = cmd->data->blksz * cmd->data->blocks;
+	}
+
+	meson_mx_sdhc_wait_cmd_ready(host->mmc);
+
+	if (cmd->flags & MMC_RSP_136) {
+		cmd->resp[0] = meson_mx_sdhc_read_response(host, 4);
+		cmd->resp[1] = meson_mx_sdhc_read_response(host, 3);
+		cmd->resp[2] = meson_mx_sdhc_read_response(host, 2);
+		cmd->resp[3] = meson_mx_sdhc_read_response(host, 1);
+	} else {
+		cmd->resp[0] = meson_mx_sdhc_read_response(host, 0);
+	}
+
+	if (cmd->error == -EIO || cmd->error == -ETIMEDOUT)
+		meson_mx_sdhc_hw_reset(host->mmc);
+	else if (cmd->data)
+		/*
+		 * Clear the FIFOs after completing data transfers to prevent
+		 * corrupting data on write access. It's not clear why this is
+		 * needed (for reads and writes), but it mimics what the BSP
+		 * kernel did.
+		 */
+		meson_mx_sdhc_clear_fifo(host->mmc);
+
+	meson_mx_sdhc_request_done(host);
+
+	return IRQ_HANDLED;
+}
+
+static void meson_mx_sdhc_init_hw_meson8(struct mmc_host *mmc)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+
+	regmap_write(host->regmap, MESON_SDHC_MISC,
+		     FIELD_PREP(MESON_SDHC_MISC_TXSTART_THRES, 7) |
+		     FIELD_PREP(MESON_SDHC_MISC_WCRC_ERR_PATT, 5) |
+		     FIELD_PREP(MESON_SDHC_MISC_WCRC_OK_PATT, 2));
+
+	regmap_write(host->regmap, MESON_SDHC_ENHC,
+		     FIELD_PREP(MESON_SDHC_ENHC_RXFIFO_TH, 63) |
+		     MESON_SDHC_ENHC_MESON6_DMA_WR_RESP |
+		     FIELD_PREP(MESON_SDHC_ENHC_MESON6_RX_TIMEOUT, 255) |
+		     FIELD_PREP(MESON_SDHC_ENHC_SDIO_IRQ_PERIOD, 12));
+};
+
+static void meson_mx_sdhc_set_pdma_meson8(struct mmc_host *mmc)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+
+	if (host->cmd->data->flags & MMC_DATA_WRITE)
+		regmap_update_bits(host->regmap, MESON_SDHC_PDMA,
+				   MESON_SDHC_PDMA_DMA_MODE |
+				   MESON_SDHC_PDMA_RD_BURST |
+				   MESON_SDHC_PDMA_TXFIFO_FILL,
+				   MESON_SDHC_PDMA_DMA_MODE |
+				   FIELD_PREP(MESON_SDHC_PDMA_RD_BURST, 31) |
+				   MESON_SDHC_PDMA_TXFIFO_FILL);
+	else
+		regmap_update_bits(host->regmap, MESON_SDHC_PDMA,
+				   MESON_SDHC_PDMA_DMA_MODE |
+				   MESON_SDHC_PDMA_RXFIFO_MANUAL_FLUSH,
+				   MESON_SDHC_PDMA_DMA_MODE |
+				   FIELD_PREP(MESON_SDHC_PDMA_RXFIFO_MANUAL_FLUSH,
+					      1));
+
+	if (host->cmd->data->flags & MMC_DATA_WRITE)
+		regmap_update_bits(host->regmap, MESON_SDHC_PDMA,
+				   MESON_SDHC_PDMA_RD_BURST,
+				   FIELD_PREP(MESON_SDHC_PDMA_RD_BURST, 15));
+}
+
+static void meson_mx_sdhc_wait_before_send_meson8(struct mmc_host *mmc)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+	u32 val;
+	int ret;
+
+	ret = regmap_read_poll_timeout(host->regmap, MESON_SDHC_ESTA, val,
+				       val == 0,
+				       MESON_SDHC_WAIT_BEFORE_SEND_SLEEP_US,
+				       MESON_SDHC_WAIT_BEFORE_SEND_TIMEOUT_US);
+	if (ret)
+		dev_warn(mmc_dev(mmc),
+			 "Failed to wait for ESTA to clear: 0x%08x\n", val);
+
+	if (host->cmd->data && host->cmd->data->flags & MMC_DATA_WRITE) {
+		ret = regmap_read_poll_timeout(host->regmap, MESON_SDHC_STAT,
+					val, val & MESON_SDHC_STAT_TXFIFO_CNT,
+					MESON_SDHC_WAIT_BEFORE_SEND_SLEEP_US,
+					MESON_SDHC_WAIT_BEFORE_SEND_TIMEOUT_US);
+		if (ret)
+			dev_warn(mmc_dev(mmc),
+				 "Failed to wait for TX FIFO to fill\n");
+	}
+}
+
+static void meson_mx_sdhc_init_hw_meson8m2(struct mmc_host *mmc)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+
+	regmap_write(host->regmap, MESON_SDHC_MISC,
+		     FIELD_PREP(MESON_SDHC_MISC_TXSTART_THRES, 6) |
+		     FIELD_PREP(MESON_SDHC_MISC_WCRC_ERR_PATT, 5) |
+		     FIELD_PREP(MESON_SDHC_MISC_WCRC_OK_PATT, 2));
+
+	regmap_write(host->regmap, MESON_SDHC_ENHC,
+		     FIELD_PREP(MESON_SDHC_ENHC_RXFIFO_TH, 64) |
+		     FIELD_PREP(MESON_SDHC_ENHC_MESON8M2_DEBUG, 1) |
+		     MESON_SDHC_ENHC_MESON8M2_WRRSP_MODE |
+		     FIELD_PREP(MESON_SDHC_ENHC_SDIO_IRQ_PERIOD, 12));
+}
+
+static void meson_mx_sdhc_set_pdma_meson8m2(struct mmc_host *mmc)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+
+	regmap_update_bits(host->regmap, MESON_SDHC_PDMA,
+			   MESON_SDHC_PDMA_DMA_MODE, MESON_SDHC_PDMA_DMA_MODE);
+}
+
+static void meson_mx_sdhc_init_hw(struct mmc_host *mmc)
+{
+	struct meson_mx_sdhc_host *host = mmc_priv(mmc);
+
+	meson_mx_sdhc_hw_reset(mmc);
+
+	regmap_write(host->regmap, MESON_SDHC_CTRL,
+		     FIELD_PREP(MESON_SDHC_CTRL_RX_PERIOD, 0xf) |
+		     FIELD_PREP(MESON_SDHC_CTRL_RX_TIMEOUT, 0x7f) |
+		     FIELD_PREP(MESON_SDHC_CTRL_RX_ENDIAN, 0x7) |
+		     FIELD_PREP(MESON_SDHC_CTRL_TX_ENDIAN, 0x7));
+
+	/*
+	 * start with a valid divider and enable the memory (un-setting
+	 * MESON_SDHC_CLKC_MEM_PWR_OFF).
+	 */
+	regmap_write(host->regmap, MESON_SDHC_CLKC, MESON_SDHC_CLKC_CLK_DIV);
+
+	regmap_write(host->regmap, MESON_SDHC_CLK2,
+		     FIELD_PREP(MESON_SDHC_CLK2_SD_CLK_PHASE, 1));
+
+	regmap_write(host->regmap, MESON_SDHC_PDMA,
+		     MESON_SDHC_PDMA_DMA_URGENT |
+		     FIELD_PREP(MESON_SDHC_PDMA_WR_BURST, 7) |
+		     FIELD_PREP(MESON_SDHC_PDMA_TXFIFO_TH, 49) |
+		     FIELD_PREP(MESON_SDHC_PDMA_RD_BURST, 15) |
+		     FIELD_PREP(MESON_SDHC_PDMA_RXFIFO_TH, 7));
+
+	/* some initialization bits depend on the SoC: */
+	host->platform->init_hw(mmc);
+
+	/* disable and mask all interrupts: */
+	regmap_write(host->regmap, MESON_SDHC_ICTL, 0);
+	regmap_write(host->regmap, MESON_SDHC_ISTA, MESON_SDHC_ISTA_ALL_IRQS);
+}
+
+static int meson_mx_sdhc_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct meson_mx_sdhc_host *host;
+	struct mmc_host *mmc;
+	void __iomem *base;
+	int ret, irq;
+
+	mmc = mmc_alloc_host(sizeof(*host), dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	ret = devm_add_action_or_reset(dev, (void(*)(void *))mmc_free_host,
+				       mmc);
+	if (ret) {
+		dev_err(dev, "Failed to register mmc_free_host action\n");
+		return ret;
+	}
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+
+	platform_set_drvdata(pdev, host);
+
+	host->platform = device_get_match_data(dev);
+	if (!host->platform)
+		return -EINVAL;
+
+	base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(base))
+		return PTR_ERR(base);
+
+	host->regmap = devm_regmap_init_mmio(dev, base,
+					     &meson_mx_sdhc_regmap_config);
+	if (IS_ERR(host->regmap))
+		return PTR_ERR(host->regmap);
+
+	host->pclk = devm_clk_get(dev, "pclk");
+	if (IS_ERR(host->pclk))
+		return PTR_ERR(host->pclk);
+
+	/* accessing any register requires the module clock to be enabled: */
+	ret = clk_prepare_enable(host->pclk);
+	if (ret) {
+		dev_err(dev, "Failed to enable 'pclk' clock\n");
+		return ret;
+	}
+
+	meson_mx_sdhc_init_hw(mmc);
+
+	ret = meson_mx_sdhc_register_clkc(dev, base, host->bulk_clks);
+	if (ret)
+		goto err_disable_pclk;
+
+	host->sd_clk = host->bulk_clks[1].clk;
+
+	/* Get regulators and the supported OCR mask */
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret)
+		goto err_disable_pclk;
+
+	mmc->max_req_size = SZ_128K;
+	mmc->max_seg_size = mmc->max_req_size;
+	mmc->max_blk_count = FIELD_GET(MESON_SDHC_SEND_TOTAL_PACK, ~0);
+	mmc->max_blk_size = MESON_SDHC_MAX_BLK_SIZE;
+	mmc->max_busy_timeout = 30 * MSEC_PER_SEC;
+	mmc->f_min = clk_round_rate(host->sd_clk, 1);
+	mmc->f_max = clk_round_rate(host->sd_clk, ULONG_MAX);
+	mmc->max_current_180 = 300;
+	mmc->max_current_330 = 300;
+	mmc->caps |= MMC_CAP_WAIT_WHILE_BUSY | MMC_CAP_HW_RESET;
+	mmc->ops = &meson_mx_sdhc_ops;
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto err_disable_pclk;
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0) {
+		ret = irq;
+		goto err_disable_pclk;
+	}
+
+	ret = devm_request_threaded_irq(dev, irq, meson_mx_sdhc_irq,
+					meson_mx_sdhc_irq_thread, IRQF_ONESHOT,
+					NULL, host);
+	if (ret)
+		goto err_disable_pclk;
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto err_disable_pclk;
+
+	return 0;
+
+err_disable_pclk:
+	clk_disable_unprepare(host->pclk);
+	return ret;
+}
+
+static int meson_mx_sdhc_remove(struct platform_device *pdev)
+{
+	struct meson_mx_sdhc_host *host = platform_get_drvdata(pdev);
+
+	mmc_remove_host(host->mmc);
+
+	meson_mx_sdhc_disable_clks(host->mmc);
+
+	clk_disable_unprepare(host->pclk);
+
+	return 0;
+}
+
+static const struct meson_mx_sdhc_data meson_mx_sdhc_data_meson8 = {
+	.init_hw			= meson_mx_sdhc_init_hw_meson8,
+	.set_pdma			= meson_mx_sdhc_set_pdma_meson8,
+	.wait_before_send		= meson_mx_sdhc_wait_before_send_meson8,
+	.hardware_flush_all_cmds	= false,
+};
+
+static const struct meson_mx_sdhc_data meson_mx_sdhc_data_meson8m2 = {
+	.init_hw			= meson_mx_sdhc_init_hw_meson8m2,
+	.set_pdma			= meson_mx_sdhc_set_pdma_meson8m2,
+	.hardware_flush_all_cmds	= true,
+};
+
+static const struct of_device_id meson_mx_sdhc_of_match[] = {
+	{
+		.compatible = "amlogic,meson8-sdhc",
+		.data = &meson_mx_sdhc_data_meson8
+	},
+	{
+		.compatible = "amlogic,meson8b-sdhc",
+		.data = &meson_mx_sdhc_data_meson8
+	},
+	{
+		.compatible = "amlogic,meson8m2-sdhc",
+		.data = &meson_mx_sdhc_data_meson8m2
+	},
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, meson_mx_sdhc_of_match);
+
+static struct platform_driver meson_mx_sdhc_driver = {
+	.probe   = meson_mx_sdhc_probe,
+	.remove  = meson_mx_sdhc_remove,
+	.driver  = {
+		.name = "meson-mx-sdhc",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(meson_mx_sdhc_of_match),
+	},
+};
+
+module_platform_driver(meson_mx_sdhc_driver);
+
+MODULE_DESCRIPTION("Meson6, Meson8, Meson8b and Meson8m2 SDHC Host Driver");
+MODULE_AUTHOR("Martin Blumenstingl <martin.blumenstingl@googlemail.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/meson-mx-sdhc.h b/drivers/mmc/host/meson-mx-sdhc.h
new file mode 100644
index 000000000..230e8fbe6
--- /dev/null
+++ b/drivers/mmc/host/meson-mx-sdhc.h
@@ -0,0 +1,141 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * Copyright (C) 2020 Martin Blumenstingl <martin.blumenstingl@googlemail.com>
+ */
+
+#ifndef _MESON_MX_SDHC_H_
+#define _MESON_MX_SDHC_H_
+
+#include <linux/bitfield.h>
+
+#define MESON_SDHC_ARGU						0x00
+
+#define MESON_SDHC_SEND						0x04
+	#define MESON_SDHC_SEND_CMD_INDEX			GENMASK(5, 0)
+	#define MESON_SDHC_SEND_CMD_HAS_RESP			BIT(6)
+	#define MESON_SDHC_SEND_CMD_HAS_DATA			BIT(7)
+	#define MESON_SDHC_SEND_RESP_LEN			BIT(8)
+	#define MESON_SDHC_SEND_RESP_NO_CRC			BIT(9)
+	#define MESON_SDHC_SEND_DATA_DIR			BIT(10)
+	#define MESON_SDHC_SEND_DATA_STOP			BIT(11)
+	#define MESON_SDHC_SEND_R1B				BIT(12)
+	#define MESON_SDHC_SEND_TOTAL_PACK			GENMASK(31, 16)
+
+#define MESON_SDHC_CTRL						0x08
+	#define MESON_SDHC_CTRL_DAT_TYPE			GENMASK(1, 0)
+	#define MESON_SDHC_CTRL_DDR_MODE			BIT(2)
+	#define MESON_SDHC_CTRL_TX_CRC_NOCHECK			BIT(3)
+	#define MESON_SDHC_CTRL_PACK_LEN			GENMASK(12, 4)
+	#define MESON_SDHC_CTRL_RX_TIMEOUT			GENMASK(19, 13)
+	#define MESON_SDHC_CTRL_RX_PERIOD			GENMASK(23, 20)
+	#define MESON_SDHC_CTRL_RX_ENDIAN			GENMASK(26, 24)
+	#define MESON_SDHC_CTRL_SDIO_IRQ_MODE			BIT(27)
+	#define MESON_SDHC_CTRL_DAT0_IRQ_SEL			BIT(28)
+	#define MESON_SDHC_CTRL_TX_ENDIAN			GENMASK(31, 29)
+
+#define MESON_SDHC_STAT						0x0c
+	#define MESON_SDHC_STAT_CMD_BUSY			BIT(0)
+	#define MESON_SDHC_STAT_DAT3_0				GENMASK(4, 1)
+	#define MESON_SDHC_STAT_CMD				BIT(5)
+	#define MESON_SDHC_STAT_RXFIFO_CNT			GENMASK(12, 6)
+	#define MESON_SDHC_STAT_TXFIFO_CNT			GENMASK(19, 13)
+	#define MESON_SDHC_STAT_DAT7_4				GENMASK(23, 20)
+
+#define MESON_SDHC_CLKC						0x10
+	#define MESON_SDHC_CLKC_CLK_DIV				GENMASK(11, 0)
+	#define MESON_SDHC_CLKC_CLK_JIC				BIT(24)
+	#define MESON_SDHC_CLKC_MEM_PWR_OFF			GENMASK(26, 25)
+
+#define MESON_SDHC_ADDR						0x14
+
+#define MESON_SDHC_PDMA						0x18
+	#define MESON_SDHC_PDMA_DMA_MODE			BIT(0)
+	#define MESON_SDHC_PDMA_PIO_RDRESP			GENMASK(3, 1)
+	#define MESON_SDHC_PDMA_DMA_URGENT			BIT(4)
+	#define MESON_SDHC_PDMA_WR_BURST			GENMASK(9, 5)
+	#define MESON_SDHC_PDMA_RD_BURST			GENMASK(14, 10)
+	#define MESON_SDHC_PDMA_RXFIFO_TH			GENMASK(21, 15)
+	#define MESON_SDHC_PDMA_TXFIFO_TH			GENMASK(28, 22)
+	#define MESON_SDHC_PDMA_RXFIFO_MANUAL_FLUSH		GENMASK(30, 29)
+	#define MESON_SDHC_PDMA_TXFIFO_FILL			BIT(31)
+
+#define MESON_SDHC_MISC						0x1c
+	#define MESON_SDHC_MISC_WCRC_ERR_PATT			GENMASK(6, 4)
+	#define MESON_SDHC_MISC_WCRC_OK_PATT			GENMASK(9, 7)
+	#define MESON_SDHC_MISC_BURST_NUM			GENMASK(21, 16)
+	#define MESON_SDHC_MISC_THREAD_ID			GENMASK(27, 22)
+	#define MESON_SDHC_MISC_MANUAL_STOP			BIT(28)
+	#define MESON_SDHC_MISC_TXSTART_THRES			GENMASK(31, 29)
+
+#define MESON_SDHC_DATA						0x20
+
+#define MESON_SDHC_ICTL						0x24
+	#define MESON_SDHC_ICTL_RESP_OK				BIT(0)
+	#define MESON_SDHC_ICTL_RESP_TIMEOUT			BIT(1)
+	#define MESON_SDHC_ICTL_RESP_ERR_CRC			BIT(2)
+	#define MESON_SDHC_ICTL_RESP_OK_NOCLEAR			BIT(3)
+	#define MESON_SDHC_ICTL_DATA_1PACK_OK			BIT(4)
+	#define MESON_SDHC_ICTL_DATA_TIMEOUT			BIT(5)
+	#define MESON_SDHC_ICTL_DATA_ERR_CRC			BIT(6)
+	#define MESON_SDHC_ICTL_DATA_XFER_OK			BIT(7)
+	#define MESON_SDHC_ICTL_RX_HIGHER			BIT(8)
+	#define MESON_SDHC_ICTL_RX_LOWER			BIT(9)
+	#define MESON_SDHC_ICTL_DAT1_IRQ			BIT(10)
+	#define MESON_SDHC_ICTL_DMA_DONE			BIT(11)
+	#define MESON_SDHC_ICTL_RXFIFO_FULL			BIT(12)
+	#define MESON_SDHC_ICTL_TXFIFO_EMPTY			BIT(13)
+	#define MESON_SDHC_ICTL_ADDI_DAT1_IRQ			BIT(14)
+	#define MESON_SDHC_ICTL_ALL_IRQS			GENMASK(14, 0)
+	#define MESON_SDHC_ICTL_DAT1_IRQ_DELAY			GENMASK(17, 16)
+
+#define MESON_SDHC_ISTA						0x28
+	#define MESON_SDHC_ISTA_RESP_OK				BIT(0)
+	#define MESON_SDHC_ISTA_RESP_TIMEOUT			BIT(1)
+	#define MESON_SDHC_ISTA_RESP_ERR_CRC			BIT(2)
+	#define MESON_SDHC_ISTA_RESP_OK_NOCLEAR			BIT(3)
+	#define MESON_SDHC_ISTA_DATA_1PACK_OK			BIT(4)
+	#define MESON_SDHC_ISTA_DATA_TIMEOUT			BIT(5)
+	#define MESON_SDHC_ISTA_DATA_ERR_CRC			BIT(6)
+	#define MESON_SDHC_ISTA_DATA_XFER_OK			BIT(7)
+	#define MESON_SDHC_ISTA_RX_HIGHER			BIT(8)
+	#define MESON_SDHC_ISTA_RX_LOWER			BIT(9)
+	#define MESON_SDHC_ISTA_DAT1_IRQ			BIT(10)
+	#define MESON_SDHC_ISTA_DMA_DONE			BIT(11)
+	#define MESON_SDHC_ISTA_RXFIFO_FULL			BIT(12)
+	#define MESON_SDHC_ISTA_TXFIFO_EMPTY			BIT(13)
+	#define MESON_SDHC_ISTA_ADDI_DAT1_IRQ			BIT(14)
+	#define MESON_SDHC_ISTA_ALL_IRQS			GENMASK(14, 0)
+
+#define MESON_SDHC_SRST						0x2c
+	#define MESON_SDHC_SRST_MAIN_CTRL			BIT(0)
+	#define MESON_SDHC_SRST_RXFIFO				BIT(1)
+	#define MESON_SDHC_SRST_TXFIFO				BIT(2)
+	#define MESON_SDHC_SRST_DPHY_RX				BIT(3)
+	#define MESON_SDHC_SRST_DPHY_TX				BIT(4)
+	#define MESON_SDHC_SRST_DMA_IF				BIT(5)
+
+#define MESON_SDHC_ESTA						0x30
+	#define MESON_SDHC_ESTA_11_13				GENMASK(13, 11)
+
+#define MESON_SDHC_ENHC						0x34
+	#define MESON_SDHC_ENHC_MESON8M2_WRRSP_MODE		BIT(0)
+	#define MESON_SDHC_ENHC_MESON8M2_CHK_WRRSP		BIT(1)
+	#define MESON_SDHC_ENHC_MESON8M2_CHK_DMA		BIT(2)
+	#define MESON_SDHC_ENHC_MESON8M2_DEBUG			GENMASK(5, 3)
+	#define MESON_SDHC_ENHC_MESON6_RX_TIMEOUT		GENMASK(7, 0)
+	#define MESON_SDHC_ENHC_MESON6_DMA_RD_RESP		BIT(16)
+	#define MESON_SDHC_ENHC_MESON6_DMA_WR_RESP		BIT(17)
+	#define MESON_SDHC_ENHC_SDIO_IRQ_PERIOD			GENMASK(15, 8)
+	#define MESON_SDHC_ENHC_RXFIFO_TH			GENMASK(24, 18)
+	#define MESON_SDHC_ENHC_TXFIFO_TH			GENMASK(31, 25)
+
+#define MESON_SDHC_CLK2						0x38
+	#define MESON_SDHC_CLK2_RX_CLK_PHASE			GENMASK(11, 0)
+	#define MESON_SDHC_CLK2_SD_CLK_PHASE			GENMASK(23, 12)
+
+struct clk_bulk_data;
+
+int meson_mx_sdhc_register_clkc(struct device *dev, void __iomem *base,
+				struct clk_bulk_data *clk_bulk_data);
+
+#endif /* _MESON_MX_SDHC_H_ */
diff --git a/drivers/mmc/host/meson-mx-sdio.c b/drivers/mmc/host/meson-mx-sdio.c
new file mode 100644
index 000000000..264aae2a2
--- /dev/null
+++ b/drivers/mmc/host/meson-mx-sdio.c
@@ -0,0 +1,773 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * meson-mx-sdio.c - Meson6, Meson8 and Meson8b SDIO/MMC Host Controller
+ *
+ * Copyright (C) 2015 Endless Mobile, Inc.
+ * Author: Carlo Caione <carlo@endlessm.com>
+ * Copyright (C) 2017 Martin Blumenstingl <martin.blumenstingl@googlemail.com>
+ */
+
+#include <linux/bitfield.h>
+#include <linux/clk.h>
+#include <linux/clk-provider.h>
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/module.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/ioport.h>
+#include <linux/platform_device.h>
+#include <linux/of_platform.h>
+#include <linux/timer.h>
+#include <linux/types.h>
+
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sdio.h>
+#include <linux/mmc/slot-gpio.h>
+
+#define MESON_MX_SDIO_ARGU					0x00
+
+#define MESON_MX_SDIO_SEND					0x04
+	#define MESON_MX_SDIO_SEND_COMMAND_INDEX_MASK		GENMASK(7, 0)
+	#define MESON_MX_SDIO_SEND_CMD_RESP_BITS_MASK		GENMASK(15, 8)
+	#define MESON_MX_SDIO_SEND_RESP_WITHOUT_CRC7		BIT(16)
+	#define MESON_MX_SDIO_SEND_RESP_HAS_DATA		BIT(17)
+	#define MESON_MX_SDIO_SEND_RESP_CRC7_FROM_8		BIT(18)
+	#define MESON_MX_SDIO_SEND_CHECK_DAT0_BUSY		BIT(19)
+	#define MESON_MX_SDIO_SEND_DATA				BIT(20)
+	#define MESON_MX_SDIO_SEND_USE_INT_WINDOW		BIT(21)
+	#define MESON_MX_SDIO_SEND_REPEAT_PACKAGE_TIMES_MASK	GENMASK(31, 24)
+
+#define MESON_MX_SDIO_CONF					0x08
+	#define MESON_MX_SDIO_CONF_CMD_CLK_DIV_SHIFT		0
+	#define MESON_MX_SDIO_CONF_CMD_CLK_DIV_WIDTH		10
+	#define MESON_MX_SDIO_CONF_CMD_DISABLE_CRC		BIT(10)
+	#define MESON_MX_SDIO_CONF_CMD_OUT_AT_POSITIVE_EDGE	BIT(11)
+	#define MESON_MX_SDIO_CONF_CMD_ARGUMENT_BITS_MASK	GENMASK(17, 12)
+	#define MESON_MX_SDIO_CONF_RESP_LATCH_AT_NEGATIVE_EDGE	BIT(18)
+	#define MESON_MX_SDIO_CONF_DATA_LATCH_AT_NEGATIVE_EDGE	BIT(19)
+	#define MESON_MX_SDIO_CONF_BUS_WIDTH			BIT(20)
+	#define MESON_MX_SDIO_CONF_M_ENDIAN_MASK		GENMASK(22, 21)
+	#define MESON_MX_SDIO_CONF_WRITE_NWR_MASK		GENMASK(28, 23)
+	#define MESON_MX_SDIO_CONF_WRITE_CRC_OK_STATUS_MASK	GENMASK(31, 29)
+
+#define MESON_MX_SDIO_IRQS					0x0c
+	#define MESON_MX_SDIO_IRQS_STATUS_STATE_MACHINE_MASK	GENMASK(3, 0)
+	#define MESON_MX_SDIO_IRQS_CMD_BUSY			BIT(4)
+	#define MESON_MX_SDIO_IRQS_RESP_CRC7_OK			BIT(5)
+	#define MESON_MX_SDIO_IRQS_DATA_READ_CRC16_OK		BIT(6)
+	#define MESON_MX_SDIO_IRQS_DATA_WRITE_CRC16_OK		BIT(7)
+	#define MESON_MX_SDIO_IRQS_IF_INT			BIT(8)
+	#define MESON_MX_SDIO_IRQS_CMD_INT			BIT(9)
+	#define MESON_MX_SDIO_IRQS_STATUS_INFO_MASK		GENMASK(15, 12)
+	#define MESON_MX_SDIO_IRQS_TIMING_OUT_INT		BIT(16)
+	#define MESON_MX_SDIO_IRQS_AMRISC_TIMING_OUT_INT_EN	BIT(17)
+	#define MESON_MX_SDIO_IRQS_ARC_TIMING_OUT_INT_EN	BIT(18)
+	#define MESON_MX_SDIO_IRQS_TIMING_OUT_COUNT_MASK	GENMASK(31, 19)
+
+#define MESON_MX_SDIO_IRQC					0x10
+	#define MESON_MX_SDIO_IRQC_ARC_IF_INT_EN		BIT(3)
+	#define MESON_MX_SDIO_IRQC_ARC_CMD_INT_EN		BIT(4)
+	#define MESON_MX_SDIO_IRQC_IF_CONFIG_MASK		GENMASK(7, 6)
+	#define MESON_MX_SDIO_IRQC_FORCE_DATA_CLK		BIT(8)
+	#define MESON_MX_SDIO_IRQC_FORCE_DATA_CMD		BIT(9)
+	#define MESON_MX_SDIO_IRQC_FORCE_DATA_DAT_MASK		GENMASK(13, 10)
+	#define MESON_MX_SDIO_IRQC_SOFT_RESET			BIT(15)
+	#define MESON_MX_SDIO_IRQC_FORCE_HALT			BIT(30)
+	#define MESON_MX_SDIO_IRQC_HALT_HOLE			BIT(31)
+
+#define MESON_MX_SDIO_MULT					0x14
+	#define MESON_MX_SDIO_MULT_PORT_SEL_MASK		GENMASK(1, 0)
+	#define MESON_MX_SDIO_MULT_MEMORY_STICK_ENABLE		BIT(2)
+	#define MESON_MX_SDIO_MULT_MEMORY_STICK_SCLK_ALWAYS	BIT(3)
+	#define MESON_MX_SDIO_MULT_STREAM_ENABLE		BIT(4)
+	#define MESON_MX_SDIO_MULT_STREAM_8BITS_MODE		BIT(5)
+	#define MESON_MX_SDIO_MULT_WR_RD_OUT_INDEX		BIT(8)
+	#define MESON_MX_SDIO_MULT_DAT0_DAT1_SWAPPED		BIT(10)
+	#define MESON_MX_SDIO_MULT_DAT1_DAT0_SWAPPED		BIT(11)
+	#define MESON_MX_SDIO_MULT_RESP_READ_INDEX_MASK		GENMASK(15, 12)
+
+#define MESON_MX_SDIO_ADDR					0x18
+
+#define MESON_MX_SDIO_EXT					0x1c
+	#define MESON_MX_SDIO_EXT_DATA_RW_NUMBER_MASK		GENMASK(29, 16)
+
+#define MESON_MX_SDIO_BOUNCE_REQ_SIZE				(128 * 1024)
+#define MESON_MX_SDIO_RESPONSE_CRC16_BITS			(16 - 1)
+#define MESON_MX_SDIO_MAX_SLOTS					3
+
+struct meson_mx_mmc_host {
+	struct device			*controller_dev;
+
+	struct clk			*parent_clk;
+	struct clk			*core_clk;
+	struct clk_divider		cfg_div;
+	struct clk			*cfg_div_clk;
+	struct clk_fixed_factor		fixed_factor;
+	struct clk			*fixed_factor_clk;
+
+	void __iomem			*base;
+	int				irq;
+	spinlock_t			irq_lock;
+
+	struct timer_list		cmd_timeout;
+
+	unsigned int			slot_id;
+	struct mmc_host			*mmc;
+
+	struct mmc_request		*mrq;
+	struct mmc_command		*cmd;
+	int				error;
+};
+
+static void meson_mx_mmc_mask_bits(struct mmc_host *mmc, char reg, u32 mask,
+				   u32 val)
+{
+	struct meson_mx_mmc_host *host = mmc_priv(mmc);
+	u32 regval;
+
+	regval = readl(host->base + reg);
+	regval &= ~mask;
+	regval |= (val & mask);
+
+	writel(regval, host->base + reg);
+}
+
+static void meson_mx_mmc_soft_reset(struct meson_mx_mmc_host *host)
+{
+	writel(MESON_MX_SDIO_IRQC_SOFT_RESET, host->base + MESON_MX_SDIO_IRQC);
+	udelay(2);
+}
+
+static struct mmc_command *meson_mx_mmc_get_next_cmd(struct mmc_command *cmd)
+{
+	if (cmd->opcode == MMC_SET_BLOCK_COUNT && !cmd->error)
+		return cmd->mrq->cmd;
+	else if (mmc_op_multi(cmd->opcode) &&
+		 (!cmd->mrq->sbc || cmd->error || cmd->data->error))
+		return cmd->mrq->stop;
+	else
+		return NULL;
+}
+
+static void meson_mx_mmc_start_cmd(struct mmc_host *mmc,
+				   struct mmc_command *cmd)
+{
+	struct meson_mx_mmc_host *host = mmc_priv(mmc);
+	unsigned int pack_size;
+	unsigned long irqflags, timeout;
+	u32 mult, send = 0, ext = 0;
+
+	host->cmd = cmd;
+
+	if (cmd->busy_timeout)
+		timeout = msecs_to_jiffies(cmd->busy_timeout);
+	else
+		timeout = msecs_to_jiffies(1000);
+
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_R1:
+	case MMC_RSP_R1B:
+	case MMC_RSP_R3:
+		/* 7 (CMD) + 32 (response) + 7 (CRC) -1 */
+		send |= FIELD_PREP(MESON_MX_SDIO_SEND_CMD_RESP_BITS_MASK, 45);
+		break;
+	case MMC_RSP_R2:
+		/* 7 (CMD) + 120 (response) + 7 (CRC) -1 */
+		send |= FIELD_PREP(MESON_MX_SDIO_SEND_CMD_RESP_BITS_MASK, 133);
+		send |= MESON_MX_SDIO_SEND_RESP_CRC7_FROM_8;
+		break;
+	default:
+		break;
+	}
+
+	if (!(cmd->flags & MMC_RSP_CRC))
+		send |= MESON_MX_SDIO_SEND_RESP_WITHOUT_CRC7;
+
+	if (cmd->flags & MMC_RSP_BUSY)
+		send |= MESON_MX_SDIO_SEND_CHECK_DAT0_BUSY;
+
+	if (cmd->data) {
+		send |= FIELD_PREP(MESON_MX_SDIO_SEND_REPEAT_PACKAGE_TIMES_MASK,
+				   (cmd->data->blocks - 1));
+
+		pack_size = cmd->data->blksz * BITS_PER_BYTE;
+		if (mmc->ios.bus_width == MMC_BUS_WIDTH_4)
+			pack_size += MESON_MX_SDIO_RESPONSE_CRC16_BITS * 4;
+		else
+			pack_size += MESON_MX_SDIO_RESPONSE_CRC16_BITS * 1;
+
+		ext |= FIELD_PREP(MESON_MX_SDIO_EXT_DATA_RW_NUMBER_MASK,
+				  pack_size);
+
+		if (cmd->data->flags & MMC_DATA_WRITE)
+			send |= MESON_MX_SDIO_SEND_DATA;
+		else
+			send |= MESON_MX_SDIO_SEND_RESP_HAS_DATA;
+
+		cmd->data->bytes_xfered = 0;
+	}
+
+	send |= FIELD_PREP(MESON_MX_SDIO_SEND_COMMAND_INDEX_MASK,
+			   (0x40 | cmd->opcode));
+
+	spin_lock_irqsave(&host->irq_lock, irqflags);
+
+	mult = readl(host->base + MESON_MX_SDIO_MULT);
+	mult &= ~MESON_MX_SDIO_MULT_PORT_SEL_MASK;
+	mult |= FIELD_PREP(MESON_MX_SDIO_MULT_PORT_SEL_MASK, host->slot_id);
+	mult |= BIT(31);
+	writel(mult, host->base + MESON_MX_SDIO_MULT);
+
+	/* enable the CMD done interrupt */
+	meson_mx_mmc_mask_bits(mmc, MESON_MX_SDIO_IRQC,
+			       MESON_MX_SDIO_IRQC_ARC_CMD_INT_EN,
+			       MESON_MX_SDIO_IRQC_ARC_CMD_INT_EN);
+
+	/* clear pending interrupts */
+	meson_mx_mmc_mask_bits(mmc, MESON_MX_SDIO_IRQS,
+			       MESON_MX_SDIO_IRQS_CMD_INT,
+			       MESON_MX_SDIO_IRQS_CMD_INT);
+
+	writel(cmd->arg, host->base + MESON_MX_SDIO_ARGU);
+	writel(ext, host->base + MESON_MX_SDIO_EXT);
+	writel(send, host->base + MESON_MX_SDIO_SEND);
+
+	spin_unlock_irqrestore(&host->irq_lock, irqflags);
+
+	mod_timer(&host->cmd_timeout, jiffies + timeout);
+}
+
+static void meson_mx_mmc_request_done(struct meson_mx_mmc_host *host)
+{
+	struct mmc_request *mrq;
+
+	mrq = host->mrq;
+
+	if (host->cmd->error)
+		meson_mx_mmc_soft_reset(host);
+
+	host->mrq = NULL;
+	host->cmd = NULL;
+
+	mmc_request_done(host->mmc, mrq);
+}
+
+static void meson_mx_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct meson_mx_mmc_host *host = mmc_priv(mmc);
+	unsigned short vdd = ios->vdd;
+	unsigned long clk_rate = ios->clock;
+
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_1:
+		meson_mx_mmc_mask_bits(mmc, MESON_MX_SDIO_CONF,
+				       MESON_MX_SDIO_CONF_BUS_WIDTH, 0);
+		break;
+
+	case MMC_BUS_WIDTH_4:
+		meson_mx_mmc_mask_bits(mmc, MESON_MX_SDIO_CONF,
+				       MESON_MX_SDIO_CONF_BUS_WIDTH,
+				       MESON_MX_SDIO_CONF_BUS_WIDTH);
+		break;
+
+	case MMC_BUS_WIDTH_8:
+	default:
+		dev_err(mmc_dev(mmc), "unsupported bus width: %d\n",
+			ios->bus_width);
+		host->error = -EINVAL;
+		return;
+	}
+
+	host->error = clk_set_rate(host->cfg_div_clk, ios->clock);
+	if (host->error) {
+		dev_warn(mmc_dev(mmc),
+				"failed to set MMC clock to %lu: %d\n",
+				clk_rate, host->error);
+		return;
+	}
+
+	mmc->actual_clock = clk_get_rate(host->cfg_div_clk);
+
+	switch (ios->power_mode) {
+	case MMC_POWER_OFF:
+		vdd = 0;
+		fallthrough;
+	case MMC_POWER_UP:
+		if (!IS_ERR(mmc->supply.vmmc)) {
+			host->error = mmc_regulator_set_ocr(mmc,
+							    mmc->supply.vmmc,
+							    vdd);
+			if (host->error)
+				return;
+		}
+		break;
+	}
+}
+
+static int meson_mx_mmc_map_dma(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct mmc_data *data = mrq->data;
+	int dma_len;
+	struct scatterlist *sg;
+
+	if (!data)
+		return 0;
+
+	sg = data->sg;
+	if (sg->offset & 3 || sg->length & 3) {
+		dev_err(mmc_dev(mmc),
+			"unaligned scatterlist: offset %x length %d\n",
+			sg->offset, sg->length);
+		return -EINVAL;
+	}
+
+	dma_len = dma_map_sg(mmc_dev(mmc), data->sg, data->sg_len,
+			     mmc_get_dma_dir(data));
+	if (dma_len <= 0) {
+		dev_err(mmc_dev(mmc), "dma_map_sg failed\n");
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static void meson_mx_mmc_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct meson_mx_mmc_host *host = mmc_priv(mmc);
+	struct mmc_command *cmd = mrq->cmd;
+
+	if (!host->error)
+		host->error = meson_mx_mmc_map_dma(mmc, mrq);
+
+	if (host->error) {
+		cmd->error = host->error;
+		mmc_request_done(mmc, mrq);
+		return;
+	}
+
+	host->mrq = mrq;
+
+	if (mrq->data)
+		writel(sg_dma_address(mrq->data->sg),
+		       host->base + MESON_MX_SDIO_ADDR);
+
+	if (mrq->sbc)
+		meson_mx_mmc_start_cmd(mmc, mrq->sbc);
+	else
+		meson_mx_mmc_start_cmd(mmc, mrq->cmd);
+}
+
+static void meson_mx_mmc_read_response(struct mmc_host *mmc,
+				       struct mmc_command *cmd)
+{
+	struct meson_mx_mmc_host *host = mmc_priv(mmc);
+	u32 mult;
+	int i, resp[4];
+
+	mult = readl(host->base + MESON_MX_SDIO_MULT);
+	mult |= MESON_MX_SDIO_MULT_WR_RD_OUT_INDEX;
+	mult &= ~MESON_MX_SDIO_MULT_RESP_READ_INDEX_MASK;
+	mult |= FIELD_PREP(MESON_MX_SDIO_MULT_RESP_READ_INDEX_MASK, 0);
+	writel(mult, host->base + MESON_MX_SDIO_MULT);
+
+	if (cmd->flags & MMC_RSP_136) {
+		for (i = 0; i <= 3; i++)
+			resp[3 - i] = readl(host->base + MESON_MX_SDIO_ARGU);
+		cmd->resp[0] = (resp[0] << 8) | ((resp[1] >> 24) & 0xff);
+		cmd->resp[1] = (resp[1] << 8) | ((resp[2] >> 24) & 0xff);
+		cmd->resp[2] = (resp[2] << 8) | ((resp[3] >> 24) & 0xff);
+		cmd->resp[3] = (resp[3] << 8);
+	} else if (cmd->flags & MMC_RSP_PRESENT) {
+		cmd->resp[0] = readl(host->base + MESON_MX_SDIO_ARGU);
+	}
+}
+
+static irqreturn_t meson_mx_mmc_process_cmd_irq(struct meson_mx_mmc_host *host,
+						u32 irqs, u32 send)
+{
+	struct mmc_command *cmd = host->cmd;
+
+	/*
+	 * NOTE: even though it shouldn't happen we sometimes get command
+	 * interrupts twice (at least this is what it looks like). Ideally
+	 * we find out why this happens and warn here as soon as it occurs.
+	 */
+	if (!cmd)
+		return IRQ_HANDLED;
+
+	cmd->error = 0;
+	meson_mx_mmc_read_response(host->mmc, cmd);
+
+	if (cmd->data) {
+		if (!((irqs & MESON_MX_SDIO_IRQS_DATA_READ_CRC16_OK) ||
+		      (irqs & MESON_MX_SDIO_IRQS_DATA_WRITE_CRC16_OK)))
+			cmd->error = -EILSEQ;
+	} else {
+		if (!((irqs & MESON_MX_SDIO_IRQS_RESP_CRC7_OK) ||
+		      (send & MESON_MX_SDIO_SEND_RESP_WITHOUT_CRC7)))
+			cmd->error = -EILSEQ;
+	}
+
+	return IRQ_WAKE_THREAD;
+}
+
+static irqreturn_t meson_mx_mmc_irq(int irq, void *data)
+{
+	struct meson_mx_mmc_host *host = (void *) data;
+	u32 irqs, send;
+	unsigned long irqflags;
+	irqreturn_t ret;
+
+	spin_lock_irqsave(&host->irq_lock, irqflags);
+
+	irqs = readl(host->base + MESON_MX_SDIO_IRQS);
+	send = readl(host->base + MESON_MX_SDIO_SEND);
+
+	if (irqs & MESON_MX_SDIO_IRQS_CMD_INT)
+		ret = meson_mx_mmc_process_cmd_irq(host, irqs, send);
+	else
+		ret = IRQ_HANDLED;
+
+	/* finally ACK all pending interrupts */
+	writel(irqs, host->base + MESON_MX_SDIO_IRQS);
+
+	spin_unlock_irqrestore(&host->irq_lock, irqflags);
+
+	return ret;
+}
+
+static irqreturn_t meson_mx_mmc_irq_thread(int irq, void *irq_data)
+{
+	struct meson_mx_mmc_host *host = (void *) irq_data;
+	struct mmc_command *cmd = host->cmd, *next_cmd;
+
+	if (WARN_ON(!cmd))
+		return IRQ_HANDLED;
+
+	del_timer_sync(&host->cmd_timeout);
+
+	if (cmd->data) {
+		dma_unmap_sg(mmc_dev(host->mmc), cmd->data->sg,
+				cmd->data->sg_len,
+				mmc_get_dma_dir(cmd->data));
+
+		cmd->data->bytes_xfered = cmd->data->blksz * cmd->data->blocks;
+	}
+
+	next_cmd = meson_mx_mmc_get_next_cmd(cmd);
+	if (next_cmd)
+		meson_mx_mmc_start_cmd(host->mmc, next_cmd);
+	else
+		meson_mx_mmc_request_done(host);
+
+	return IRQ_HANDLED;
+}
+
+static void meson_mx_mmc_timeout(struct timer_list *t)
+{
+	struct meson_mx_mmc_host *host = from_timer(host, t, cmd_timeout);
+	unsigned long irqflags;
+	u32 irqc;
+
+	spin_lock_irqsave(&host->irq_lock, irqflags);
+
+	/* disable the CMD interrupt */
+	irqc = readl(host->base + MESON_MX_SDIO_IRQC);
+	irqc &= ~MESON_MX_SDIO_IRQC_ARC_CMD_INT_EN;
+	writel(irqc, host->base + MESON_MX_SDIO_IRQC);
+
+	spin_unlock_irqrestore(&host->irq_lock, irqflags);
+
+	/*
+	 * skip the timeout handling if the interrupt handler already processed
+	 * the command.
+	 */
+	if (!host->cmd)
+		return;
+
+	dev_dbg(mmc_dev(host->mmc),
+		"Timeout on CMD%u (IRQS = 0x%08x, ARGU = 0x%08x)\n",
+		host->cmd->opcode, readl(host->base + MESON_MX_SDIO_IRQS),
+		readl(host->base + MESON_MX_SDIO_ARGU));
+
+	host->cmd->error = -ETIMEDOUT;
+
+	meson_mx_mmc_request_done(host);
+}
+
+static struct mmc_host_ops meson_mx_mmc_ops = {
+	.request		= meson_mx_mmc_request,
+	.set_ios		= meson_mx_mmc_set_ios,
+	.get_cd			= mmc_gpio_get_cd,
+	.get_ro			= mmc_gpio_get_ro,
+};
+
+static struct platform_device *meson_mx_mmc_slot_pdev(struct device *parent)
+{
+	struct device_node *slot_node;
+	struct platform_device *pdev;
+
+	/*
+	 * TODO: the MMC core framework currently does not support
+	 * controllers with multiple slots properly. So we only register
+	 * the first slot for now
+	 */
+	slot_node = of_get_compatible_child(parent->of_node, "mmc-slot");
+	if (!slot_node) {
+		dev_warn(parent, "no 'mmc-slot' sub-node found\n");
+		return ERR_PTR(-ENOENT);
+	}
+
+	pdev = of_platform_device_create(slot_node, NULL, parent);
+	of_node_put(slot_node);
+
+	return pdev;
+}
+
+static int meson_mx_mmc_add_host(struct meson_mx_mmc_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	struct device *slot_dev = mmc_dev(mmc);
+	int ret;
+
+	if (of_property_read_u32(slot_dev->of_node, "reg", &host->slot_id)) {
+		dev_err(slot_dev, "missing 'reg' property\n");
+		return -EINVAL;
+	}
+
+	if (host->slot_id >= MESON_MX_SDIO_MAX_SLOTS) {
+		dev_err(slot_dev, "invalid 'reg' property value %d\n",
+			host->slot_id);
+		return -EINVAL;
+	}
+
+	/* Get regulators and the supported OCR mask */
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret)
+		return ret;
+
+	mmc->max_req_size = MESON_MX_SDIO_BOUNCE_REQ_SIZE;
+	mmc->max_seg_size = mmc->max_req_size;
+	mmc->max_blk_count =
+		FIELD_GET(MESON_MX_SDIO_SEND_REPEAT_PACKAGE_TIMES_MASK,
+			  0xffffffff);
+	mmc->max_blk_size = FIELD_GET(MESON_MX_SDIO_EXT_DATA_RW_NUMBER_MASK,
+				      0xffffffff);
+	mmc->max_blk_size -= (4 * MESON_MX_SDIO_RESPONSE_CRC16_BITS);
+	mmc->max_blk_size /= BITS_PER_BYTE;
+
+	/* Get the min and max supported clock rates */
+	mmc->f_min = clk_round_rate(host->cfg_div_clk, 1);
+	mmc->f_max = clk_round_rate(host->cfg_div_clk,
+				    clk_get_rate(host->parent_clk));
+
+	mmc->caps |= MMC_CAP_CMD23 | MMC_CAP_WAIT_WHILE_BUSY;
+	mmc->ops = &meson_mx_mmc_ops;
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		return ret;
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int meson_mx_mmc_register_clks(struct meson_mx_mmc_host *host)
+{
+	struct clk_init_data init;
+	const char *clk_div_parent, *clk_fixed_factor_parent;
+
+	clk_fixed_factor_parent = __clk_get_name(host->parent_clk);
+	init.name = devm_kasprintf(host->controller_dev, GFP_KERNEL,
+				   "%s#fixed_factor",
+				   dev_name(host->controller_dev));
+	if (!init.name)
+		return -ENOMEM;
+
+	init.ops = &clk_fixed_factor_ops;
+	init.flags = 0;
+	init.parent_names = &clk_fixed_factor_parent;
+	init.num_parents = 1;
+	host->fixed_factor.div = 2;
+	host->fixed_factor.mult = 1;
+	host->fixed_factor.hw.init = &init;
+
+	host->fixed_factor_clk = devm_clk_register(host->controller_dev,
+						 &host->fixed_factor.hw);
+	if (WARN_ON(IS_ERR(host->fixed_factor_clk)))
+		return PTR_ERR(host->fixed_factor_clk);
+
+	clk_div_parent = __clk_get_name(host->fixed_factor_clk);
+	init.name = devm_kasprintf(host->controller_dev, GFP_KERNEL,
+				   "%s#div", dev_name(host->controller_dev));
+	if (!init.name)
+		return -ENOMEM;
+
+	init.ops = &clk_divider_ops;
+	init.flags = CLK_SET_RATE_PARENT;
+	init.parent_names = &clk_div_parent;
+	init.num_parents = 1;
+	host->cfg_div.reg = host->base + MESON_MX_SDIO_CONF;
+	host->cfg_div.shift = MESON_MX_SDIO_CONF_CMD_CLK_DIV_SHIFT;
+	host->cfg_div.width = MESON_MX_SDIO_CONF_CMD_CLK_DIV_WIDTH;
+	host->cfg_div.hw.init = &init;
+	host->cfg_div.flags = CLK_DIVIDER_ALLOW_ZERO;
+
+	host->cfg_div_clk = devm_clk_register(host->controller_dev,
+					      &host->cfg_div.hw);
+	if (WARN_ON(IS_ERR(host->cfg_div_clk)))
+		return PTR_ERR(host->cfg_div_clk);
+
+	return 0;
+}
+
+static int meson_mx_mmc_probe(struct platform_device *pdev)
+{
+	struct platform_device *slot_pdev;
+	struct mmc_host *mmc;
+	struct meson_mx_mmc_host *host;
+	int ret, irq;
+	u32 conf;
+
+	slot_pdev = meson_mx_mmc_slot_pdev(&pdev->dev);
+	if (!slot_pdev)
+		return -ENODEV;
+	else if (IS_ERR(slot_pdev))
+		return PTR_ERR(slot_pdev);
+
+	mmc = mmc_alloc_host(sizeof(*host), &slot_pdev->dev);
+	if (!mmc) {
+		ret = -ENOMEM;
+		goto error_unregister_slot_pdev;
+	}
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+	host->controller_dev = &pdev->dev;
+
+	spin_lock_init(&host->irq_lock);
+	timer_setup(&host->cmd_timeout, meson_mx_mmc_timeout, 0);
+
+	platform_set_drvdata(pdev, host);
+
+	host->base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(host->base)) {
+		ret = PTR_ERR(host->base);
+		goto error_free_mmc;
+	}
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0) {
+		ret = irq;
+		goto error_free_mmc;
+	}
+
+	ret = devm_request_threaded_irq(host->controller_dev, irq,
+					meson_mx_mmc_irq,
+					meson_mx_mmc_irq_thread, IRQF_ONESHOT,
+					NULL, host);
+	if (ret)
+		goto error_free_mmc;
+
+	host->core_clk = devm_clk_get(host->controller_dev, "core");
+	if (IS_ERR(host->core_clk)) {
+		ret = PTR_ERR(host->core_clk);
+		goto error_free_mmc;
+	}
+
+	host->parent_clk = devm_clk_get(host->controller_dev, "clkin");
+	if (IS_ERR(host->parent_clk)) {
+		ret = PTR_ERR(host->parent_clk);
+		goto error_free_mmc;
+	}
+
+	ret = meson_mx_mmc_register_clks(host);
+	if (ret)
+		goto error_free_mmc;
+
+	ret = clk_prepare_enable(host->core_clk);
+	if (ret) {
+		dev_err(host->controller_dev, "Failed to enable core clock\n");
+		goto error_free_mmc;
+	}
+
+	ret = clk_prepare_enable(host->cfg_div_clk);
+	if (ret) {
+		dev_err(host->controller_dev, "Failed to enable MMC clock\n");
+		goto error_disable_core_clk;
+	}
+
+	conf = 0;
+	conf |= FIELD_PREP(MESON_MX_SDIO_CONF_CMD_ARGUMENT_BITS_MASK, 39);
+	conf |= FIELD_PREP(MESON_MX_SDIO_CONF_M_ENDIAN_MASK, 0x3);
+	conf |= FIELD_PREP(MESON_MX_SDIO_CONF_WRITE_NWR_MASK, 0x2);
+	conf |= FIELD_PREP(MESON_MX_SDIO_CONF_WRITE_CRC_OK_STATUS_MASK, 0x2);
+	writel(conf, host->base + MESON_MX_SDIO_CONF);
+
+	meson_mx_mmc_soft_reset(host);
+
+	ret = meson_mx_mmc_add_host(host);
+	if (ret)
+		goto error_disable_clks;
+
+	return 0;
+
+error_disable_clks:
+	clk_disable_unprepare(host->cfg_div_clk);
+error_disable_core_clk:
+	clk_disable_unprepare(host->core_clk);
+error_free_mmc:
+	mmc_free_host(mmc);
+error_unregister_slot_pdev:
+	of_platform_device_destroy(&slot_pdev->dev, NULL);
+	return ret;
+}
+
+static int meson_mx_mmc_remove(struct platform_device *pdev)
+{
+	struct meson_mx_mmc_host *host = platform_get_drvdata(pdev);
+	struct device *slot_dev = mmc_dev(host->mmc);
+
+	del_timer_sync(&host->cmd_timeout);
+
+	mmc_remove_host(host->mmc);
+
+	of_platform_device_destroy(slot_dev, NULL);
+
+	clk_disable_unprepare(host->cfg_div_clk);
+	clk_disable_unprepare(host->core_clk);
+
+	mmc_free_host(host->mmc);
+
+	return 0;
+}
+
+static const struct of_device_id meson_mx_mmc_of_match[] = {
+	{ .compatible = "amlogic,meson8-sdio", },
+	{ .compatible = "amlogic,meson8b-sdio", },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, meson_mx_mmc_of_match);
+
+static struct platform_driver meson_mx_mmc_driver = {
+	.probe   = meson_mx_mmc_probe,
+	.remove  = meson_mx_mmc_remove,
+	.driver  = {
+		.name = "meson-mx-sdio",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(meson_mx_mmc_of_match),
+	},
+};
+
+module_platform_driver(meson_mx_mmc_driver);
+
+MODULE_DESCRIPTION("Meson6, Meson8 and Meson8b SDIO/MMC Host Driver");
+MODULE_AUTHOR("Carlo Caione <carlo@endlessm.com>");
+MODULE_AUTHOR("Martin Blumenstingl <martin.blumenstingl@googlemail.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/mmc_hsq.c b/drivers/mmc/host/mmc_hsq.c
new file mode 100644
index 000000000..9d35453e7
--- /dev/null
+++ b/drivers/mmc/host/mmc_hsq.c
@@ -0,0 +1,375 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ *
+ * MMC software queue support based on command queue interfaces
+ *
+ * Copyright (C) 2019 Linaro, Inc.
+ * Author: Baolin Wang <baolin.wang@linaro.org>
+ */
+
+#include <linux/mmc/card.h>
+#include <linux/mmc/host.h>
+#include <linux/module.h>
+
+#include "mmc_hsq.h"
+
+#define HSQ_NUM_SLOTS	64
+#define HSQ_INVALID_TAG	HSQ_NUM_SLOTS
+
+static void mmc_hsq_retry_handler(struct work_struct *work)
+{
+	struct mmc_hsq *hsq = container_of(work, struct mmc_hsq, retry_work);
+	struct mmc_host *mmc = hsq->mmc;
+
+	mmc->ops->request(mmc, hsq->mrq);
+}
+
+static void mmc_hsq_pump_requests(struct mmc_hsq *hsq)
+{
+	struct mmc_host *mmc = hsq->mmc;
+	struct hsq_slot *slot;
+	unsigned long flags;
+	int ret = 0;
+
+	spin_lock_irqsave(&hsq->lock, flags);
+
+	/* Make sure we are not already running a request now */
+	if (hsq->mrq || hsq->recovery_halt) {
+		spin_unlock_irqrestore(&hsq->lock, flags);
+		return;
+	}
+
+	/* Make sure there are remain requests need to pump */
+	if (!hsq->qcnt || !hsq->enabled) {
+		spin_unlock_irqrestore(&hsq->lock, flags);
+		return;
+	}
+
+	slot = &hsq->slot[hsq->next_tag];
+	hsq->mrq = slot->mrq;
+	hsq->qcnt--;
+
+	spin_unlock_irqrestore(&hsq->lock, flags);
+
+	if (mmc->ops->request_atomic)
+		ret = mmc->ops->request_atomic(mmc, hsq->mrq);
+	else
+		mmc->ops->request(mmc, hsq->mrq);
+
+	/*
+	 * If returning BUSY from request_atomic(), which means the card
+	 * may be busy now, and we should change to non-atomic context to
+	 * try again for this unusual case, to avoid time-consuming operations
+	 * in the atomic context.
+	 *
+	 * Note: we just give a warning for other error cases, since the host
+	 * driver will handle them.
+	 */
+	if (ret == -EBUSY)
+		schedule_work(&hsq->retry_work);
+	else
+		WARN_ON_ONCE(ret);
+}
+
+static void mmc_hsq_update_next_tag(struct mmc_hsq *hsq, int remains)
+{
+	struct hsq_slot *slot;
+	int tag;
+
+	/*
+	 * If there are no remain requests in software queue, then set a invalid
+	 * tag.
+	 */
+	if (!remains) {
+		hsq->next_tag = HSQ_INVALID_TAG;
+		return;
+	}
+
+	/*
+	 * Increasing the next tag and check if the corresponding request is
+	 * available, if yes, then we found a candidate request.
+	 */
+	if (++hsq->next_tag != HSQ_INVALID_TAG) {
+		slot = &hsq->slot[hsq->next_tag];
+		if (slot->mrq)
+			return;
+	}
+
+	/* Othersie we should iterate all slots to find a available tag. */
+	for (tag = 0; tag < HSQ_NUM_SLOTS; tag++) {
+		slot = &hsq->slot[tag];
+		if (slot->mrq)
+			break;
+	}
+
+	if (tag == HSQ_NUM_SLOTS)
+		tag = HSQ_INVALID_TAG;
+
+	hsq->next_tag = tag;
+}
+
+static void mmc_hsq_post_request(struct mmc_hsq *hsq)
+{
+	unsigned long flags;
+	int remains;
+
+	spin_lock_irqsave(&hsq->lock, flags);
+
+	remains = hsq->qcnt;
+	hsq->mrq = NULL;
+
+	/* Update the next available tag to be queued. */
+	mmc_hsq_update_next_tag(hsq, remains);
+
+	if (hsq->waiting_for_idle && !remains) {
+		hsq->waiting_for_idle = false;
+		wake_up(&hsq->wait_queue);
+	}
+
+	/* Do not pump new request in recovery mode. */
+	if (hsq->recovery_halt) {
+		spin_unlock_irqrestore(&hsq->lock, flags);
+		return;
+	}
+
+	spin_unlock_irqrestore(&hsq->lock, flags);
+
+	 /*
+	  * Try to pump new request to host controller as fast as possible,
+	  * after completing previous request.
+	  */
+	if (remains > 0)
+		mmc_hsq_pump_requests(hsq);
+}
+
+/**
+ * mmc_hsq_finalize_request - finalize one request if the request is done
+ * @mmc: the host controller
+ * @mrq: the request need to be finalized
+ *
+ * Return true if we finalized the corresponding request in software queue,
+ * otherwise return false.
+ */
+bool mmc_hsq_finalize_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct mmc_hsq *hsq = mmc->cqe_private;
+	unsigned long flags;
+
+	spin_lock_irqsave(&hsq->lock, flags);
+
+	if (!hsq->enabled || !hsq->mrq || hsq->mrq != mrq) {
+		spin_unlock_irqrestore(&hsq->lock, flags);
+		return false;
+	}
+
+	/*
+	 * Clear current completed slot request to make a room for new request.
+	 */
+	hsq->slot[hsq->next_tag].mrq = NULL;
+
+	spin_unlock_irqrestore(&hsq->lock, flags);
+
+	mmc_cqe_request_done(mmc, hsq->mrq);
+
+	mmc_hsq_post_request(hsq);
+
+	return true;
+}
+EXPORT_SYMBOL_GPL(mmc_hsq_finalize_request);
+
+static void mmc_hsq_recovery_start(struct mmc_host *mmc)
+{
+	struct mmc_hsq *hsq = mmc->cqe_private;
+	unsigned long flags;
+
+	spin_lock_irqsave(&hsq->lock, flags);
+
+	hsq->recovery_halt = true;
+
+	spin_unlock_irqrestore(&hsq->lock, flags);
+}
+
+static void mmc_hsq_recovery_finish(struct mmc_host *mmc)
+{
+	struct mmc_hsq *hsq = mmc->cqe_private;
+	int remains;
+
+	spin_lock_irq(&hsq->lock);
+
+	hsq->recovery_halt = false;
+	remains = hsq->qcnt;
+
+	spin_unlock_irq(&hsq->lock);
+
+	/*
+	 * Try to pump new request if there are request pending in software
+	 * queue after finishing recovery.
+	 */
+	if (remains > 0)
+		mmc_hsq_pump_requests(hsq);
+}
+
+static int mmc_hsq_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct mmc_hsq *hsq = mmc->cqe_private;
+	int tag = mrq->tag;
+
+	spin_lock_irq(&hsq->lock);
+
+	if (!hsq->enabled) {
+		spin_unlock_irq(&hsq->lock);
+		return -ESHUTDOWN;
+	}
+
+	/* Do not queue any new requests in recovery mode. */
+	if (hsq->recovery_halt) {
+		spin_unlock_irq(&hsq->lock);
+		return -EBUSY;
+	}
+
+	hsq->slot[tag].mrq = mrq;
+
+	/*
+	 * Set the next tag as current request tag if no available
+	 * next tag.
+	 */
+	if (hsq->next_tag == HSQ_INVALID_TAG)
+		hsq->next_tag = tag;
+
+	hsq->qcnt++;
+
+	spin_unlock_irq(&hsq->lock);
+
+	mmc_hsq_pump_requests(hsq);
+
+	return 0;
+}
+
+static void mmc_hsq_post_req(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	if (mmc->ops->post_req)
+		mmc->ops->post_req(mmc, mrq, 0);
+}
+
+static bool mmc_hsq_queue_is_idle(struct mmc_hsq *hsq, int *ret)
+{
+	bool is_idle;
+
+	spin_lock_irq(&hsq->lock);
+
+	is_idle = (!hsq->mrq && !hsq->qcnt) ||
+		hsq->recovery_halt;
+
+	*ret = hsq->recovery_halt ? -EBUSY : 0;
+	hsq->waiting_for_idle = !is_idle;
+
+	spin_unlock_irq(&hsq->lock);
+
+	return is_idle;
+}
+
+static int mmc_hsq_wait_for_idle(struct mmc_host *mmc)
+{
+	struct mmc_hsq *hsq = mmc->cqe_private;
+	int ret;
+
+	wait_event(hsq->wait_queue,
+		   mmc_hsq_queue_is_idle(hsq, &ret));
+
+	return ret;
+}
+
+static void mmc_hsq_disable(struct mmc_host *mmc)
+{
+	struct mmc_hsq *hsq = mmc->cqe_private;
+	u32 timeout = 500;
+	int ret;
+
+	spin_lock_irq(&hsq->lock);
+
+	if (!hsq->enabled) {
+		spin_unlock_irq(&hsq->lock);
+		return;
+	}
+
+	spin_unlock_irq(&hsq->lock);
+
+	ret = wait_event_timeout(hsq->wait_queue,
+				 mmc_hsq_queue_is_idle(hsq, &ret),
+				 msecs_to_jiffies(timeout));
+	if (ret == 0) {
+		pr_warn("could not stop mmc software queue\n");
+		return;
+	}
+
+	spin_lock_irq(&hsq->lock);
+
+	hsq->enabled = false;
+
+	spin_unlock_irq(&hsq->lock);
+}
+
+static int mmc_hsq_enable(struct mmc_host *mmc, struct mmc_card *card)
+{
+	struct mmc_hsq *hsq = mmc->cqe_private;
+
+	spin_lock_irq(&hsq->lock);
+
+	if (hsq->enabled) {
+		spin_unlock_irq(&hsq->lock);
+		return -EBUSY;
+	}
+
+	hsq->enabled = true;
+
+	spin_unlock_irq(&hsq->lock);
+
+	return 0;
+}
+
+static const struct mmc_cqe_ops mmc_hsq_ops = {
+	.cqe_enable = mmc_hsq_enable,
+	.cqe_disable = mmc_hsq_disable,
+	.cqe_request = mmc_hsq_request,
+	.cqe_post_req = mmc_hsq_post_req,
+	.cqe_wait_for_idle = mmc_hsq_wait_for_idle,
+	.cqe_recovery_start = mmc_hsq_recovery_start,
+	.cqe_recovery_finish = mmc_hsq_recovery_finish,
+};
+
+int mmc_hsq_init(struct mmc_hsq *hsq, struct mmc_host *mmc)
+{
+	hsq->num_slots = HSQ_NUM_SLOTS;
+	hsq->next_tag = HSQ_INVALID_TAG;
+
+	hsq->slot = devm_kcalloc(mmc_dev(mmc), hsq->num_slots,
+				 sizeof(struct hsq_slot), GFP_KERNEL);
+	if (!hsq->slot)
+		return -ENOMEM;
+
+	hsq->mmc = mmc;
+	hsq->mmc->cqe_private = hsq;
+	mmc->cqe_ops = &mmc_hsq_ops;
+
+	INIT_WORK(&hsq->retry_work, mmc_hsq_retry_handler);
+	spin_lock_init(&hsq->lock);
+	init_waitqueue_head(&hsq->wait_queue);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mmc_hsq_init);
+
+void mmc_hsq_suspend(struct mmc_host *mmc)
+{
+	mmc_hsq_disable(mmc);
+}
+EXPORT_SYMBOL_GPL(mmc_hsq_suspend);
+
+int mmc_hsq_resume(struct mmc_host *mmc)
+{
+	return mmc_hsq_enable(mmc, NULL);
+}
+EXPORT_SYMBOL_GPL(mmc_hsq_resume);
+
+MODULE_DESCRIPTION("MMC Host Software Queue support");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/mmc_hsq.h b/drivers/mmc/host/mmc_hsq.h
new file mode 100644
index 000000000..ffdd9cd17
--- /dev/null
+++ b/drivers/mmc/host/mmc_hsq.h
@@ -0,0 +1,31 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef LINUX_MMC_HSQ_H
+#define LINUX_MMC_HSQ_H
+
+struct hsq_slot {
+	struct mmc_request *mrq;
+};
+
+struct mmc_hsq {
+	struct mmc_host *mmc;
+	struct mmc_request *mrq;
+	wait_queue_head_t wait_queue;
+	struct hsq_slot *slot;
+	spinlock_t lock;
+	struct work_struct retry_work;
+
+	int next_tag;
+	int num_slots;
+	int qcnt;
+
+	bool enabled;
+	bool waiting_for_idle;
+	bool recovery_halt;
+};
+
+int mmc_hsq_init(struct mmc_hsq *hsq, struct mmc_host *mmc);
+void mmc_hsq_suspend(struct mmc_host *mmc);
+int mmc_hsq_resume(struct mmc_host *mmc);
+bool mmc_hsq_finalize_request(struct mmc_host *mmc, struct mmc_request *mrq);
+
+#endif
diff --git a/drivers/mmc/host/mmc_spi.c b/drivers/mmc/host/mmc_spi.c
new file mode 100644
index 000000000..1d814919e
--- /dev/null
+++ b/drivers/mmc/host/mmc_spi.c
@@ -0,0 +1,1546 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Access SD/MMC cards through SPI master controllers
+ *
+ * (C) Copyright 2005, Intec Automation,
+ *		Mike Lavender (mike@steroidmicros)
+ * (C) Copyright 2006-2007, David Brownell
+ * (C) Copyright 2007, Axis Communications,
+ *		Hans-Peter Nilsson (hp@axis.com)
+ * (C) Copyright 2007, ATRON electronic GmbH,
+ *		Jan Nikitenko <jan.nikitenko@gmail.com>
+ */
+#include <linux/sched.h>
+#include <linux/delay.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <linux/bio.h>
+#include <linux/dma-mapping.h>
+#include <linux/crc7.h>
+#include <linux/crc-itu-t.h>
+#include <linux/scatterlist.h>
+
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>		/* for R1_SPI_* bit values */
+#include <linux/mmc/slot-gpio.h>
+
+#include <linux/spi/spi.h>
+#include <linux/spi/mmc_spi.h>
+
+#include <asm/unaligned.h>
+
+
+/* NOTES:
+ *
+ * - For now, we won't try to interoperate with a real mmc/sd/sdio
+ *   controller, although some of them do have hardware support for
+ *   SPI protocol.  The main reason for such configs would be mmc-ish
+ *   cards like DataFlash, which don't support that "native" protocol.
+ *
+ *   We don't have a "DataFlash/MMC/SD/SDIO card slot" abstraction to
+ *   switch between driver stacks, and in any case if "native" mode
+ *   is available, it will be faster and hence preferable.
+ *
+ * - MMC depends on a different chipselect management policy than the
+ *   SPI interface currently supports for shared bus segments:  it needs
+ *   to issue multiple spi_message requests with the chipselect active,
+ *   using the results of one message to decide the next one to issue.
+ *
+ *   Pending updates to the programming interface, this driver expects
+ *   that it not share the bus with other drivers (precluding conflicts).
+ *
+ * - We tell the controller to keep the chipselect active from the
+ *   beginning of an mmc_host_ops.request until the end.  So beware
+ *   of SPI controller drivers that mis-handle the cs_change flag!
+ *
+ *   However, many cards seem OK with chipselect flapping up/down
+ *   during that time ... at least on unshared bus segments.
+ */
+
+
+/*
+ * Local protocol constants, internal to data block protocols.
+ */
+
+/* Response tokens used to ack each block written: */
+#define SPI_MMC_RESPONSE_CODE(x)	((x) & 0x1f)
+#define SPI_RESPONSE_ACCEPTED		((2 << 1)|1)
+#define SPI_RESPONSE_CRC_ERR		((5 << 1)|1)
+#define SPI_RESPONSE_WRITE_ERR		((6 << 1)|1)
+
+/* Read and write blocks start with these tokens and end with crc;
+ * on error, read tokens act like a subset of R2_SPI_* values.
+ */
+#define SPI_TOKEN_SINGLE	0xfe	/* single block r/w, multiblock read */
+#define SPI_TOKEN_MULTI_WRITE	0xfc	/* multiblock write */
+#define SPI_TOKEN_STOP_TRAN	0xfd	/* terminate multiblock write */
+
+#define MMC_SPI_BLOCKSIZE	512
+
+#define MMC_SPI_R1B_TIMEOUT_MS	3000
+#define MMC_SPI_INIT_TIMEOUT_MS	3000
+
+/* One of the critical speed parameters is the amount of data which may
+ * be transferred in one command. If this value is too low, the SD card
+ * controller has to do multiple partial block writes (argggh!). With
+ * today (2008) SD cards there is little speed gain if we transfer more
+ * than 64 KBytes at a time. So use this value until there is any indication
+ * that we should do more here.
+ */
+#define MMC_SPI_BLOCKSATONCE	128
+
+/****************************************************************************/
+
+/*
+ * Local Data Structures
+ */
+
+/* "scratch" is per-{command,block} data exchanged with the card */
+struct scratch {
+	u8			status[29];
+	u8			data_token;
+	__be16			crc_val;
+};
+
+struct mmc_spi_host {
+	struct mmc_host		*mmc;
+	struct spi_device	*spi;
+
+	unsigned char		power_mode;
+	u16			powerup_msecs;
+
+	struct mmc_spi_platform_data	*pdata;
+
+	/* for bulk data transfers */
+	struct spi_transfer	token, t, crc, early_status;
+	struct spi_message	m;
+
+	/* for status readback */
+	struct spi_transfer	status;
+	struct spi_message	readback;
+
+	/* underlying DMA-aware controller, or null */
+	struct device		*dma_dev;
+
+	/* buffer used for commands and for message "overhead" */
+	struct scratch		*data;
+	dma_addr_t		data_dma;
+
+	/* Specs say to write ones most of the time, even when the card
+	 * has no need to read its input data; and many cards won't care.
+	 * This is our source of those ones.
+	 */
+	void			*ones;
+	dma_addr_t		ones_dma;
+};
+
+
+/****************************************************************************/
+
+/*
+ * MMC-over-SPI protocol glue, used by the MMC stack interface
+ */
+
+static inline int mmc_cs_off(struct mmc_spi_host *host)
+{
+	/* chipselect will always be inactive after setup() */
+	return spi_setup(host->spi);
+}
+
+static int
+mmc_spi_readbytes(struct mmc_spi_host *host, unsigned len)
+{
+	int status;
+
+	if (len > sizeof(*host->data)) {
+		WARN_ON(1);
+		return -EIO;
+	}
+
+	host->status.len = len;
+
+	if (host->dma_dev)
+		dma_sync_single_for_device(host->dma_dev,
+				host->data_dma, sizeof(*host->data),
+				DMA_FROM_DEVICE);
+
+	status = spi_sync_locked(host->spi, &host->readback);
+
+	if (host->dma_dev)
+		dma_sync_single_for_cpu(host->dma_dev,
+				host->data_dma, sizeof(*host->data),
+				DMA_FROM_DEVICE);
+
+	return status;
+}
+
+static int mmc_spi_skip(struct mmc_spi_host *host, unsigned long timeout,
+			unsigned n, u8 byte)
+{
+	u8 *cp = host->data->status;
+	unsigned long start = jiffies;
+
+	while (1) {
+		int		status;
+		unsigned	i;
+
+		status = mmc_spi_readbytes(host, n);
+		if (status < 0)
+			return status;
+
+		for (i = 0; i < n; i++) {
+			if (cp[i] != byte)
+				return cp[i];
+		}
+
+		if (time_is_before_jiffies(start + timeout))
+			break;
+
+		/* If we need long timeouts, we may release the CPU.
+		 * We use jiffies here because we want to have a relation
+		 * between elapsed time and the blocking of the scheduler.
+		 */
+		if (time_is_before_jiffies(start + 1))
+			schedule();
+	}
+	return -ETIMEDOUT;
+}
+
+static inline int
+mmc_spi_wait_unbusy(struct mmc_spi_host *host, unsigned long timeout)
+{
+	return mmc_spi_skip(host, timeout, sizeof(host->data->status), 0);
+}
+
+static int mmc_spi_readtoken(struct mmc_spi_host *host, unsigned long timeout)
+{
+	return mmc_spi_skip(host, timeout, 1, 0xff);
+}
+
+
+/*
+ * Note that for SPI, cmd->resp[0] is not the same data as "native" protocol
+ * hosts return!  The low byte holds R1_SPI bits.  The next byte may hold
+ * R2_SPI bits ... for SEND_STATUS, or after data read errors.
+ *
+ * cmd->resp[1] holds any four-byte response, for R3 (READ_OCR) and on
+ * newer cards R7 (IF_COND).
+ */
+
+static char *maptype(struct mmc_command *cmd)
+{
+	switch (mmc_spi_resp_type(cmd)) {
+	case MMC_RSP_SPI_R1:	return "R1";
+	case MMC_RSP_SPI_R1B:	return "R1B";
+	case MMC_RSP_SPI_R2:	return "R2/R5";
+	case MMC_RSP_SPI_R3:	return "R3/R4/R7";
+	default:		return "?";
+	}
+}
+
+/* return zero, else negative errno after setting cmd->error */
+static int mmc_spi_response_get(struct mmc_spi_host *host,
+		struct mmc_command *cmd, int cs_on)
+{
+	unsigned long timeout_ms;
+	u8	*cp = host->data->status;
+	u8	*end = cp + host->t.len;
+	int	value = 0;
+	int	bitshift;
+	u8 	leftover = 0;
+	unsigned short rotator;
+	int 	i;
+	char	tag[32];
+
+	snprintf(tag, sizeof(tag), "  ... CMD%d response SPI_%s",
+		cmd->opcode, maptype(cmd));
+
+	/* Except for data block reads, the whole response will already
+	 * be stored in the scratch buffer.  It's somewhere after the
+	 * command and the first byte we read after it.  We ignore that
+	 * first byte.  After STOP_TRANSMISSION command it may include
+	 * two data bits, but otherwise it's all ones.
+	 */
+	cp += 8;
+	while (cp < end && *cp == 0xff)
+		cp++;
+
+	/* Data block reads (R1 response types) may need more data... */
+	if (cp == end) {
+		cp = host->data->status;
+		end = cp+1;
+
+		/* Card sends N(CR) (== 1..8) bytes of all-ones then one
+		 * status byte ... and we already scanned 2 bytes.
+		 *
+		 * REVISIT block read paths use nasty byte-at-a-time I/O
+		 * so it can always DMA directly into the target buffer.
+		 * It'd probably be better to memcpy() the first chunk and
+		 * avoid extra i/o calls...
+		 *
+		 * Note we check for more than 8 bytes, because in practice,
+		 * some SD cards are slow...
+		 */
+		for (i = 2; i < 16; i++) {
+			value = mmc_spi_readbytes(host, 1);
+			if (value < 0)
+				goto done;
+			if (*cp != 0xff)
+				goto checkstatus;
+		}
+		value = -ETIMEDOUT;
+		goto done;
+	}
+
+checkstatus:
+	bitshift = 0;
+	if (*cp & 0x80)	{
+		/* Houston, we have an ugly card with a bit-shifted response */
+		rotator = *cp++ << 8;
+		/* read the next byte */
+		if (cp == end) {
+			value = mmc_spi_readbytes(host, 1);
+			if (value < 0)
+				goto done;
+			cp = host->data->status;
+			end = cp+1;
+		}
+		rotator |= *cp++;
+		while (rotator & 0x8000) {
+			bitshift++;
+			rotator <<= 1;
+		}
+		cmd->resp[0] = rotator >> 8;
+		leftover = rotator;
+	} else {
+		cmd->resp[0] = *cp++;
+	}
+	cmd->error = 0;
+
+	/* Status byte: the entire seven-bit R1 response.  */
+	if (cmd->resp[0] != 0) {
+		if ((R1_SPI_PARAMETER | R1_SPI_ADDRESS)
+				& cmd->resp[0])
+			value = -EFAULT; /* Bad address */
+		else if (R1_SPI_ILLEGAL_COMMAND & cmd->resp[0])
+			value = -ENOSYS; /* Function not implemented */
+		else if (R1_SPI_COM_CRC & cmd->resp[0])
+			value = -EILSEQ; /* Illegal byte sequence */
+		else if ((R1_SPI_ERASE_SEQ | R1_SPI_ERASE_RESET)
+				& cmd->resp[0])
+			value = -EIO;    /* I/O error */
+		/* else R1_SPI_IDLE, "it's resetting" */
+	}
+
+	switch (mmc_spi_resp_type(cmd)) {
+
+	/* SPI R1B == R1 + busy; STOP_TRANSMISSION (for multiblock reads)
+	 * and less-common stuff like various erase operations.
+	 */
+	case MMC_RSP_SPI_R1B:
+		/* maybe we read all the busy tokens already */
+		while (cp < end && *cp == 0)
+			cp++;
+		if (cp == end) {
+			timeout_ms = cmd->busy_timeout ? cmd->busy_timeout :
+				MMC_SPI_R1B_TIMEOUT_MS;
+			mmc_spi_wait_unbusy(host, msecs_to_jiffies(timeout_ms));
+		}
+		break;
+
+	/* SPI R2 == R1 + second status byte; SEND_STATUS
+	 * SPI R5 == R1 + data byte; IO_RW_DIRECT
+	 */
+	case MMC_RSP_SPI_R2:
+		/* read the next byte */
+		if (cp == end) {
+			value = mmc_spi_readbytes(host, 1);
+			if (value < 0)
+				goto done;
+			cp = host->data->status;
+			end = cp+1;
+		}
+		if (bitshift) {
+			rotator = leftover << 8;
+			rotator |= *cp << bitshift;
+			cmd->resp[0] |= (rotator & 0xFF00);
+		} else {
+			cmd->resp[0] |= *cp << 8;
+		}
+		break;
+
+	/* SPI R3, R4, or R7 == R1 + 4 bytes */
+	case MMC_RSP_SPI_R3:
+		rotator = leftover << 8;
+		cmd->resp[1] = 0;
+		for (i = 0; i < 4; i++) {
+			cmd->resp[1] <<= 8;
+			/* read the next byte */
+			if (cp == end) {
+				value = mmc_spi_readbytes(host, 1);
+				if (value < 0)
+					goto done;
+				cp = host->data->status;
+				end = cp+1;
+			}
+			if (bitshift) {
+				rotator |= *cp++ << bitshift;
+				cmd->resp[1] |= (rotator >> 8);
+				rotator <<= 8;
+			} else {
+				cmd->resp[1] |= *cp++;
+			}
+		}
+		break;
+
+	/* SPI R1 == just one status byte */
+	case MMC_RSP_SPI_R1:
+		break;
+
+	default:
+		dev_dbg(&host->spi->dev, "bad response type %04x\n",
+			mmc_spi_resp_type(cmd));
+		if (value >= 0)
+			value = -EINVAL;
+		goto done;
+	}
+
+	if (value < 0)
+		dev_dbg(&host->spi->dev, "%s: resp %04x %08x\n",
+			tag, cmd->resp[0], cmd->resp[1]);
+
+	/* disable chipselect on errors and some success cases */
+	if (value >= 0 && cs_on)
+		return value;
+done:
+	if (value < 0)
+		cmd->error = value;
+	mmc_cs_off(host);
+	return value;
+}
+
+/* Issue command and read its response.
+ * Returns zero on success, negative for error.
+ *
+ * On error, caller must cope with mmc core retry mechanism.  That
+ * means immediate low-level resubmit, which affects the bus lock...
+ */
+static int
+mmc_spi_command_send(struct mmc_spi_host *host,
+		struct mmc_request *mrq,
+		struct mmc_command *cmd, int cs_on)
+{
+	struct scratch		*data = host->data;
+	u8			*cp = data->status;
+	int			status;
+	struct spi_transfer	*t;
+
+	/* We can handle most commands (except block reads) in one full
+	 * duplex I/O operation before either starting the next transfer
+	 * (data block or command) or else deselecting the card.
+	 *
+	 * First, write 7 bytes:
+	 *  - an all-ones byte to ensure the card is ready
+	 *  - opcode byte (plus start and transmission bits)
+	 *  - four bytes of big-endian argument
+	 *  - crc7 (plus end bit) ... always computed, it's cheap
+	 *
+	 * We init the whole buffer to all-ones, which is what we need
+	 * to write while we're reading (later) response data.
+	 */
+	memset(cp, 0xff, sizeof(data->status));
+
+	cp[1] = 0x40 | cmd->opcode;
+	put_unaligned_be32(cmd->arg, cp + 2);
+	cp[6] = crc7_be(0, cp + 1, 5) | 0x01;
+	cp += 7;
+
+	/* Then, read up to 13 bytes (while writing all-ones):
+	 *  - N(CR) (== 1..8) bytes of all-ones
+	 *  - status byte (for all response types)
+	 *  - the rest of the response, either:
+	 *      + nothing, for R1 or R1B responses
+	 *	+ second status byte, for R2 responses
+	 *	+ four data bytes, for R3 and R7 responses
+	 *
+	 * Finally, read some more bytes ... in the nice cases we know in
+	 * advance how many, and reading 1 more is always OK:
+	 *  - N(EC) (== 0..N) bytes of all-ones, before deselect/finish
+	 *  - N(RC) (== 1..N) bytes of all-ones, before next command
+	 *  - N(WR) (== 1..N) bytes of all-ones, before data write
+	 *
+	 * So in those cases one full duplex I/O of at most 21 bytes will
+	 * handle the whole command, leaving the card ready to receive a
+	 * data block or new command.  We do that whenever we can, shaving
+	 * CPU and IRQ costs (especially when using DMA or FIFOs).
+	 *
+	 * There are two other cases, where it's not generally practical
+	 * to rely on a single I/O:
+	 *
+	 *  - R1B responses need at least N(EC) bytes of all-zeroes.
+	 *
+	 *    In this case we can *try* to fit it into one I/O, then
+	 *    maybe read more data later.
+	 *
+	 *  - Data block reads are more troublesome, since a variable
+	 *    number of padding bytes precede the token and data.
+	 *      + N(CX) (== 0..8) bytes of all-ones, before CSD or CID
+	 *      + N(AC) (== 1..many) bytes of all-ones
+	 *
+	 *    In this case we currently only have minimal speedups here:
+	 *    when N(CR) == 1 we can avoid I/O in response_get().
+	 */
+	if (cs_on && (mrq->data->flags & MMC_DATA_READ)) {
+		cp += 2;	/* min(N(CR)) + status */
+		/* R1 */
+	} else {
+		cp += 10;	/* max(N(CR)) + status + min(N(RC),N(WR)) */
+		if (cmd->flags & MMC_RSP_SPI_S2)	/* R2/R5 */
+			cp++;
+		else if (cmd->flags & MMC_RSP_SPI_B4)	/* R3/R4/R7 */
+			cp += 4;
+		else if (cmd->flags & MMC_RSP_BUSY)	/* R1B */
+			cp = data->status + sizeof(data->status);
+		/* else:  R1 (most commands) */
+	}
+
+	dev_dbg(&host->spi->dev, "  mmc_spi: CMD%d, resp %s\n",
+		cmd->opcode, maptype(cmd));
+
+	/* send command, leaving chipselect active */
+	spi_message_init(&host->m);
+
+	t = &host->t;
+	memset(t, 0, sizeof(*t));
+	t->tx_buf = t->rx_buf = data->status;
+	t->tx_dma = t->rx_dma = host->data_dma;
+	t->len = cp - data->status;
+	t->cs_change = 1;
+	spi_message_add_tail(t, &host->m);
+
+	if (host->dma_dev) {
+		host->m.is_dma_mapped = 1;
+		dma_sync_single_for_device(host->dma_dev,
+				host->data_dma, sizeof(*host->data),
+				DMA_BIDIRECTIONAL);
+	}
+	status = spi_sync_locked(host->spi, &host->m);
+
+	if (host->dma_dev)
+		dma_sync_single_for_cpu(host->dma_dev,
+				host->data_dma, sizeof(*host->data),
+				DMA_BIDIRECTIONAL);
+	if (status < 0) {
+		dev_dbg(&host->spi->dev, "  ... write returned %d\n", status);
+		cmd->error = status;
+		return status;
+	}
+
+	/* after no-data commands and STOP_TRANSMISSION, chipselect off */
+	return mmc_spi_response_get(host, cmd, cs_on);
+}
+
+/* Build data message with up to four separate transfers.  For TX, we
+ * start by writing the data token.  And in most cases, we finish with
+ * a status transfer.
+ *
+ * We always provide TX data for data and CRC.  The MMC/SD protocol
+ * requires us to write ones; but Linux defaults to writing zeroes;
+ * so we explicitly initialize it to all ones on RX paths.
+ *
+ * We also handle DMA mapping, so the underlying SPI controller does
+ * not need to (re)do it for each message.
+ */
+static void
+mmc_spi_setup_data_message(
+	struct mmc_spi_host	*host,
+	int			multiple,
+	enum dma_data_direction	direction)
+{
+	struct spi_transfer	*t;
+	struct scratch		*scratch = host->data;
+	dma_addr_t		dma = host->data_dma;
+
+	spi_message_init(&host->m);
+	if (dma)
+		host->m.is_dma_mapped = 1;
+
+	/* for reads, readblock() skips 0xff bytes before finding
+	 * the token; for writes, this transfer issues that token.
+	 */
+	if (direction == DMA_TO_DEVICE) {
+		t = &host->token;
+		memset(t, 0, sizeof(*t));
+		t->len = 1;
+		if (multiple)
+			scratch->data_token = SPI_TOKEN_MULTI_WRITE;
+		else
+			scratch->data_token = SPI_TOKEN_SINGLE;
+		t->tx_buf = &scratch->data_token;
+		if (dma)
+			t->tx_dma = dma + offsetof(struct scratch, data_token);
+		spi_message_add_tail(t, &host->m);
+	}
+
+	/* Body of transfer is buffer, then CRC ...
+	 * either TX-only, or RX with TX-ones.
+	 */
+	t = &host->t;
+	memset(t, 0, sizeof(*t));
+	t->tx_buf = host->ones;
+	t->tx_dma = host->ones_dma;
+	/* length and actual buffer info are written later */
+	spi_message_add_tail(t, &host->m);
+
+	t = &host->crc;
+	memset(t, 0, sizeof(*t));
+	t->len = 2;
+	if (direction == DMA_TO_DEVICE) {
+		/* the actual CRC may get written later */
+		t->tx_buf = &scratch->crc_val;
+		if (dma)
+			t->tx_dma = dma + offsetof(struct scratch, crc_val);
+	} else {
+		t->tx_buf = host->ones;
+		t->tx_dma = host->ones_dma;
+		t->rx_buf = &scratch->crc_val;
+		if (dma)
+			t->rx_dma = dma + offsetof(struct scratch, crc_val);
+	}
+	spi_message_add_tail(t, &host->m);
+
+	/*
+	 * A single block read is followed by N(EC) [0+] all-ones bytes
+	 * before deselect ... don't bother.
+	 *
+	 * Multiblock reads are followed by N(AC) [1+] all-ones bytes before
+	 * the next block is read, or a STOP_TRANSMISSION is issued.  We'll
+	 * collect that single byte, so readblock() doesn't need to.
+	 *
+	 * For a write, the one-byte data response follows immediately, then
+	 * come zero or more busy bytes, then N(WR) [1+] all-ones bytes.
+	 * Then single block reads may deselect, and multiblock ones issue
+	 * the next token (next data block, or STOP_TRAN).  We can try to
+	 * minimize I/O ops by using a single read to collect end-of-busy.
+	 */
+	if (multiple || direction == DMA_TO_DEVICE) {
+		t = &host->early_status;
+		memset(t, 0, sizeof(*t));
+		t->len = (direction == DMA_TO_DEVICE) ? sizeof(scratch->status) : 1;
+		t->tx_buf = host->ones;
+		t->tx_dma = host->ones_dma;
+		t->rx_buf = scratch->status;
+		if (dma)
+			t->rx_dma = dma + offsetof(struct scratch, status);
+		t->cs_change = 1;
+		spi_message_add_tail(t, &host->m);
+	}
+}
+
+/*
+ * Write one block:
+ *  - caller handled preceding N(WR) [1+] all-ones bytes
+ *  - data block
+ *	+ token
+ *	+ data bytes
+ *	+ crc16
+ *  - an all-ones byte ... card writes a data-response byte
+ *  - followed by N(EC) [0+] all-ones bytes, card writes zero/'busy'
+ *
+ * Return negative errno, else success.
+ */
+static int
+mmc_spi_writeblock(struct mmc_spi_host *host, struct spi_transfer *t,
+	unsigned long timeout)
+{
+	struct spi_device	*spi = host->spi;
+	int			status, i;
+	struct scratch		*scratch = host->data;
+	u32			pattern;
+
+	if (host->mmc->use_spi_crc)
+		scratch->crc_val = cpu_to_be16(crc_itu_t(0, t->tx_buf, t->len));
+	if (host->dma_dev)
+		dma_sync_single_for_device(host->dma_dev,
+				host->data_dma, sizeof(*scratch),
+				DMA_BIDIRECTIONAL);
+
+	status = spi_sync_locked(spi, &host->m);
+
+	if (status != 0) {
+		dev_dbg(&spi->dev, "write error (%d)\n", status);
+		return status;
+	}
+
+	if (host->dma_dev)
+		dma_sync_single_for_cpu(host->dma_dev,
+				host->data_dma, sizeof(*scratch),
+				DMA_BIDIRECTIONAL);
+
+	/*
+	 * Get the transmission data-response reply.  It must follow
+	 * immediately after the data block we transferred.  This reply
+	 * doesn't necessarily tell whether the write operation succeeded;
+	 * it just says if the transmission was ok and whether *earlier*
+	 * writes succeeded; see the standard.
+	 *
+	 * In practice, there are (even modern SDHC-)cards which are late
+	 * in sending the response, and miss the time frame by a few bits,
+	 * so we have to cope with this situation and check the response
+	 * bit-by-bit. Arggh!!!
+	 */
+	pattern = get_unaligned_be32(scratch->status);
+
+	/* First 3 bit of pattern are undefined */
+	pattern |= 0xE0000000;
+
+	/* left-adjust to leading 0 bit */
+	while (pattern & 0x80000000)
+		pattern <<= 1;
+	/* right-adjust for pattern matching. Code is in bit 4..0 now. */
+	pattern >>= 27;
+
+	switch (pattern) {
+	case SPI_RESPONSE_ACCEPTED:
+		status = 0;
+		break;
+	case SPI_RESPONSE_CRC_ERR:
+		/* host shall then issue MMC_STOP_TRANSMISSION */
+		status = -EILSEQ;
+		break;
+	case SPI_RESPONSE_WRITE_ERR:
+		/* host shall then issue MMC_STOP_TRANSMISSION,
+		 * and should MMC_SEND_STATUS to sort it out
+		 */
+		status = -EIO;
+		break;
+	default:
+		status = -EPROTO;
+		break;
+	}
+	if (status != 0) {
+		dev_dbg(&spi->dev, "write error %02x (%d)\n",
+			scratch->status[0], status);
+		return status;
+	}
+
+	t->tx_buf += t->len;
+	if (host->dma_dev)
+		t->tx_dma += t->len;
+
+	/* Return when not busy.  If we didn't collect that status yet,
+	 * we'll need some more I/O.
+	 */
+	for (i = 4; i < sizeof(scratch->status); i++) {
+		/* card is non-busy if the most recent bit is 1 */
+		if (scratch->status[i] & 0x01)
+			return 0;
+	}
+	return mmc_spi_wait_unbusy(host, timeout);
+}
+
+/*
+ * Read one block:
+ *  - skip leading all-ones bytes ... either
+ *      + N(AC) [1..f(clock,CSD)] usually, else
+ *      + N(CX) [0..8] when reading CSD or CID
+ *  - data block
+ *	+ token ... if error token, no data or crc
+ *	+ data bytes
+ *	+ crc16
+ *
+ * After single block reads, we're done; N(EC) [0+] all-ones bytes follow
+ * before dropping chipselect.
+ *
+ * For multiblock reads, caller either reads the next block or issues a
+ * STOP_TRANSMISSION command.
+ */
+static int
+mmc_spi_readblock(struct mmc_spi_host *host, struct spi_transfer *t,
+	unsigned long timeout)
+{
+	struct spi_device	*spi = host->spi;
+	int			status;
+	struct scratch		*scratch = host->data;
+	unsigned int 		bitshift;
+	u8			leftover;
+
+	/* At least one SD card sends an all-zeroes byte when N(CX)
+	 * applies, before the all-ones bytes ... just cope with that.
+	 */
+	status = mmc_spi_readbytes(host, 1);
+	if (status < 0)
+		return status;
+	status = scratch->status[0];
+	if (status == 0xff || status == 0)
+		status = mmc_spi_readtoken(host, timeout);
+
+	if (status < 0) {
+		dev_dbg(&spi->dev, "read error %02x (%d)\n", status, status);
+		return status;
+	}
+
+	/* The token may be bit-shifted...
+	 * the first 0-bit precedes the data stream.
+	 */
+	bitshift = 7;
+	while (status & 0x80) {
+		status <<= 1;
+		bitshift--;
+	}
+	leftover = status << 1;
+
+	if (host->dma_dev) {
+		dma_sync_single_for_device(host->dma_dev,
+				host->data_dma, sizeof(*scratch),
+				DMA_BIDIRECTIONAL);
+		dma_sync_single_for_device(host->dma_dev,
+				t->rx_dma, t->len,
+				DMA_FROM_DEVICE);
+	}
+
+	status = spi_sync_locked(spi, &host->m);
+	if (status < 0) {
+		dev_dbg(&spi->dev, "read error %d\n", status);
+		return status;
+	}
+
+	if (host->dma_dev) {
+		dma_sync_single_for_cpu(host->dma_dev,
+				host->data_dma, sizeof(*scratch),
+				DMA_BIDIRECTIONAL);
+		dma_sync_single_for_cpu(host->dma_dev,
+				t->rx_dma, t->len,
+				DMA_FROM_DEVICE);
+	}
+
+	if (bitshift) {
+		/* Walk through the data and the crc and do
+		 * all the magic to get byte-aligned data.
+		 */
+		u8 *cp = t->rx_buf;
+		unsigned int len;
+		unsigned int bitright = 8 - bitshift;
+		u8 temp;
+		for (len = t->len; len; len--) {
+			temp = *cp;
+			*cp++ = leftover | (temp >> bitshift);
+			leftover = temp << bitright;
+		}
+		cp = (u8 *) &scratch->crc_val;
+		temp = *cp;
+		*cp++ = leftover | (temp >> bitshift);
+		leftover = temp << bitright;
+		temp = *cp;
+		*cp = leftover | (temp >> bitshift);
+	}
+
+	if (host->mmc->use_spi_crc) {
+		u16 crc = crc_itu_t(0, t->rx_buf, t->len);
+
+		be16_to_cpus(&scratch->crc_val);
+		if (scratch->crc_val != crc) {
+			dev_dbg(&spi->dev,
+				"read - crc error: crc_val=0x%04x, computed=0x%04x len=%d\n",
+				scratch->crc_val, crc, t->len);
+			return -EILSEQ;
+		}
+	}
+
+	t->rx_buf += t->len;
+	if (host->dma_dev)
+		t->rx_dma += t->len;
+
+	return 0;
+}
+
+/*
+ * An MMC/SD data stage includes one or more blocks, optional CRCs,
+ * and inline handshaking.  That handhaking makes it unlike most
+ * other SPI protocol stacks.
+ */
+static void
+mmc_spi_data_do(struct mmc_spi_host *host, struct mmc_command *cmd,
+		struct mmc_data *data, u32 blk_size)
+{
+	struct spi_device	*spi = host->spi;
+	struct device		*dma_dev = host->dma_dev;
+	struct spi_transfer	*t;
+	enum dma_data_direction	direction;
+	struct scatterlist	*sg;
+	unsigned		n_sg;
+	int			multiple = (data->blocks > 1);
+	u32			clock_rate;
+	unsigned long		timeout;
+
+	direction = mmc_get_dma_dir(data);
+	mmc_spi_setup_data_message(host, multiple, direction);
+	t = &host->t;
+
+	if (t->speed_hz)
+		clock_rate = t->speed_hz;
+	else
+		clock_rate = spi->max_speed_hz;
+
+	timeout = data->timeout_ns / 1000 +
+		  data->timeout_clks * 1000000 / clock_rate;
+	timeout = usecs_to_jiffies((unsigned int)timeout) + 1;
+
+	/* Handle scatterlist segments one at a time, with synch for
+	 * each 512-byte block
+	 */
+	for_each_sg(data->sg, sg, data->sg_len, n_sg) {
+		int			status = 0;
+		dma_addr_t		dma_addr = 0;
+		void			*kmap_addr;
+		unsigned		length = sg->length;
+		enum dma_data_direction	dir = direction;
+
+		/* set up dma mapping for controller drivers that might
+		 * use DMA ... though they may fall back to PIO
+		 */
+		if (dma_dev) {
+			/* never invalidate whole *shared* pages ... */
+			if ((sg->offset != 0 || length != PAGE_SIZE)
+					&& dir == DMA_FROM_DEVICE)
+				dir = DMA_BIDIRECTIONAL;
+
+			dma_addr = dma_map_page(dma_dev, sg_page(sg), 0,
+						PAGE_SIZE, dir);
+			if (dma_mapping_error(dma_dev, dma_addr)) {
+				data->error = -EFAULT;
+				break;
+			}
+			if (direction == DMA_TO_DEVICE)
+				t->tx_dma = dma_addr + sg->offset;
+			else
+				t->rx_dma = dma_addr + sg->offset;
+		}
+
+		/* allow pio too; we don't allow highmem */
+		kmap_addr = kmap(sg_page(sg));
+		if (direction == DMA_TO_DEVICE)
+			t->tx_buf = kmap_addr + sg->offset;
+		else
+			t->rx_buf = kmap_addr + sg->offset;
+
+		/* transfer each block, and update request status */
+		while (length) {
+			t->len = min(length, blk_size);
+
+			dev_dbg(&host->spi->dev,
+				"    mmc_spi: %s block, %d bytes\n",
+				(direction == DMA_TO_DEVICE) ? "write" : "read",
+				t->len);
+
+			if (direction == DMA_TO_DEVICE)
+				status = mmc_spi_writeblock(host, t, timeout);
+			else
+				status = mmc_spi_readblock(host, t, timeout);
+			if (status < 0)
+				break;
+
+			data->bytes_xfered += t->len;
+			length -= t->len;
+
+			if (!multiple)
+				break;
+		}
+
+		/* discard mappings */
+		if (direction == DMA_FROM_DEVICE)
+			flush_kernel_dcache_page(sg_page(sg));
+		kunmap(sg_page(sg));
+		if (dma_dev)
+			dma_unmap_page(dma_dev, dma_addr, PAGE_SIZE, dir);
+
+		if (status < 0) {
+			data->error = status;
+			dev_dbg(&spi->dev, "%s status %d\n",
+				(direction == DMA_TO_DEVICE) ? "write" : "read",
+				status);
+			break;
+		}
+	}
+
+	/* NOTE some docs describe an MMC-only SET_BLOCK_COUNT (CMD23) that
+	 * can be issued before multiblock writes.  Unlike its more widely
+	 * documented analogue for SD cards (SET_WR_BLK_ERASE_COUNT, ACMD23),
+	 * that can affect the STOP_TRAN logic.   Complete (and current)
+	 * MMC specs should sort that out before Linux starts using CMD23.
+	 */
+	if (direction == DMA_TO_DEVICE && multiple) {
+		struct scratch	*scratch = host->data;
+		int		tmp;
+		const unsigned	statlen = sizeof(scratch->status);
+
+		dev_dbg(&spi->dev, "    mmc_spi: STOP_TRAN\n");
+
+		/* Tweak the per-block message we set up earlier by morphing
+		 * it to hold single buffer with the token followed by some
+		 * all-ones bytes ... skip N(BR) (0..1), scan the rest for
+		 * "not busy any longer" status, and leave chip selected.
+		 */
+		INIT_LIST_HEAD(&host->m.transfers);
+		list_add(&host->early_status.transfer_list,
+				&host->m.transfers);
+
+		memset(scratch->status, 0xff, statlen);
+		scratch->status[0] = SPI_TOKEN_STOP_TRAN;
+
+		host->early_status.tx_buf = host->early_status.rx_buf;
+		host->early_status.tx_dma = host->early_status.rx_dma;
+		host->early_status.len = statlen;
+
+		if (host->dma_dev)
+			dma_sync_single_for_device(host->dma_dev,
+					host->data_dma, sizeof(*scratch),
+					DMA_BIDIRECTIONAL);
+
+		tmp = spi_sync_locked(spi, &host->m);
+
+		if (host->dma_dev)
+			dma_sync_single_for_cpu(host->dma_dev,
+					host->data_dma, sizeof(*scratch),
+					DMA_BIDIRECTIONAL);
+
+		if (tmp < 0) {
+			if (!data->error)
+				data->error = tmp;
+			return;
+		}
+
+		/* Ideally we collected "not busy" status with one I/O,
+		 * avoiding wasteful byte-at-a-time scanning... but more
+		 * I/O is often needed.
+		 */
+		for (tmp = 2; tmp < statlen; tmp++) {
+			if (scratch->status[tmp] != 0)
+				return;
+		}
+		tmp = mmc_spi_wait_unbusy(host, timeout);
+		if (tmp < 0 && !data->error)
+			data->error = tmp;
+	}
+}
+
+/****************************************************************************/
+
+/*
+ * MMC driver implementation -- the interface to the MMC stack
+ */
+
+static void mmc_spi_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct mmc_spi_host	*host = mmc_priv(mmc);
+	int			status = -EINVAL;
+	int			crc_retry = 5;
+	struct mmc_command	stop;
+
+#ifdef DEBUG
+	/* MMC core and layered drivers *MUST* issue SPI-aware commands */
+	{
+		struct mmc_command	*cmd;
+		int			invalid = 0;
+
+		cmd = mrq->cmd;
+		if (!mmc_spi_resp_type(cmd)) {
+			dev_dbg(&host->spi->dev, "bogus command\n");
+			cmd->error = -EINVAL;
+			invalid = 1;
+		}
+
+		cmd = mrq->stop;
+		if (cmd && !mmc_spi_resp_type(cmd)) {
+			dev_dbg(&host->spi->dev, "bogus STOP command\n");
+			cmd->error = -EINVAL;
+			invalid = 1;
+		}
+
+		if (invalid) {
+			dump_stack();
+			mmc_request_done(host->mmc, mrq);
+			return;
+		}
+	}
+#endif
+
+	/* request exclusive bus access */
+	spi_bus_lock(host->spi->master);
+
+crc_recover:
+	/* issue command; then optionally data and stop */
+	status = mmc_spi_command_send(host, mrq, mrq->cmd, mrq->data != NULL);
+	if (status == 0 && mrq->data) {
+		mmc_spi_data_do(host, mrq->cmd, mrq->data, mrq->data->blksz);
+
+		/*
+		 * The SPI bus is not always reliable for large data transfers.
+		 * If an occasional crc error is reported by the SD device with
+		 * data read/write over SPI, it may be recovered by repeating
+		 * the last SD command again. The retry count is set to 5 to
+		 * ensure the driver passes stress tests.
+		 */
+		if (mrq->data->error == -EILSEQ && crc_retry) {
+			stop.opcode = MMC_STOP_TRANSMISSION;
+			stop.arg = 0;
+			stop.flags = MMC_RSP_SPI_R1B | MMC_RSP_R1B | MMC_CMD_AC;
+			status = mmc_spi_command_send(host, mrq, &stop, 0);
+			crc_retry--;
+			mrq->data->error = 0;
+			goto crc_recover;
+		}
+
+		if (mrq->stop)
+			status = mmc_spi_command_send(host, mrq, mrq->stop, 0);
+		else
+			mmc_cs_off(host);
+	}
+
+	/* release the bus */
+	spi_bus_unlock(host->spi->master);
+
+	mmc_request_done(host->mmc, mrq);
+}
+
+/* See Section 6.4.1, in SD "Simplified Physical Layer Specification 2.0"
+ *
+ * NOTE that here we can't know that the card has just been powered up;
+ * not all MMC/SD sockets support power switching.
+ *
+ * FIXME when the card is still in SPI mode, e.g. from a previous kernel,
+ * this doesn't seem to do the right thing at all...
+ */
+static void mmc_spi_initsequence(struct mmc_spi_host *host)
+{
+	/* Try to be very sure any previous command has completed;
+	 * wait till not-busy, skip debris from any old commands.
+	 */
+	mmc_spi_wait_unbusy(host, msecs_to_jiffies(MMC_SPI_INIT_TIMEOUT_MS));
+	mmc_spi_readbytes(host, 10);
+
+	/*
+	 * Do a burst with chipselect active-high.  We need to do this to
+	 * meet the requirement of 74 clock cycles with both chipselect
+	 * and CMD (MOSI) high before CMD0 ... after the card has been
+	 * powered up to Vdd(min), and so is ready to take commands.
+	 *
+	 * Some cards are particularly needy of this (e.g. Viking "SD256")
+	 * while most others don't seem to care.
+	 *
+	 * Note that this is one of the places MMC/SD plays games with the
+	 * SPI protocol.  Another is that when chipselect is released while
+	 * the card returns BUSY status, the clock must issue several cycles
+	 * with chipselect high before the card will stop driving its output.
+	 *
+	 * SPI_CS_HIGH means "asserted" here. In some cases like when using
+	 * GPIOs for chip select, SPI_CS_HIGH is set but this will be logically
+	 * inverted by gpiolib, so if we want to ascertain to drive it high
+	 * we should toggle the default with an XOR as we do here.
+	 */
+	host->spi->mode ^= SPI_CS_HIGH;
+	if (spi_setup(host->spi) != 0) {
+		/* Just warn; most cards work without it. */
+		dev_warn(&host->spi->dev,
+				"can't change chip-select polarity\n");
+		host->spi->mode ^= SPI_CS_HIGH;
+	} else {
+		mmc_spi_readbytes(host, 18);
+
+		host->spi->mode ^= SPI_CS_HIGH;
+		if (spi_setup(host->spi) != 0) {
+			/* Wot, we can't get the same setup we had before? */
+			dev_err(&host->spi->dev,
+					"can't restore chip-select polarity\n");
+		}
+	}
+}
+
+static char *mmc_powerstring(u8 power_mode)
+{
+	switch (power_mode) {
+	case MMC_POWER_OFF: return "off";
+	case MMC_POWER_UP:  return "up";
+	case MMC_POWER_ON:  return "on";
+	}
+	return "?";
+}
+
+static void mmc_spi_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct mmc_spi_host *host = mmc_priv(mmc);
+
+	if (host->power_mode != ios->power_mode) {
+		int		canpower;
+
+		canpower = host->pdata && host->pdata->setpower;
+
+		dev_dbg(&host->spi->dev, "mmc_spi: power %s (%d)%s\n",
+				mmc_powerstring(ios->power_mode),
+				ios->vdd,
+				canpower ? ", can switch" : "");
+
+		/* switch power on/off if possible, accounting for
+		 * max 250msec powerup time if needed.
+		 */
+		if (canpower) {
+			switch (ios->power_mode) {
+			case MMC_POWER_OFF:
+			case MMC_POWER_UP:
+				host->pdata->setpower(&host->spi->dev,
+						ios->vdd);
+				if (ios->power_mode == MMC_POWER_UP)
+					msleep(host->powerup_msecs);
+			}
+		}
+
+		/* See 6.4.1 in the simplified SD card physical spec 2.0 */
+		if (ios->power_mode == MMC_POWER_ON)
+			mmc_spi_initsequence(host);
+
+		/* If powering down, ground all card inputs to avoid power
+		 * delivery from data lines!  On a shared SPI bus, this
+		 * will probably be temporary; 6.4.2 of the simplified SD
+		 * spec says this must last at least 1msec.
+		 *
+		 *   - Clock low means CPOL 0, e.g. mode 0
+		 *   - MOSI low comes from writing zero
+		 *   - Chipselect is usually active low...
+		 */
+		if (canpower && ios->power_mode == MMC_POWER_OFF) {
+			int mres;
+			u8 nullbyte = 0;
+
+			host->spi->mode &= ~(SPI_CPOL|SPI_CPHA);
+			mres = spi_setup(host->spi);
+			if (mres < 0)
+				dev_dbg(&host->spi->dev,
+					"switch to SPI mode 0 failed\n");
+
+			if (spi_write(host->spi, &nullbyte, 1) < 0)
+				dev_dbg(&host->spi->dev,
+					"put spi signals to low failed\n");
+
+			/*
+			 * Now clock should be low due to spi mode 0;
+			 * MOSI should be low because of written 0x00;
+			 * chipselect should be low (it is active low)
+			 * power supply is off, so now MMC is off too!
+			 *
+			 * FIXME no, chipselect can be high since the
+			 * device is inactive and SPI_CS_HIGH is clear...
+			 */
+			msleep(10);
+			if (mres == 0) {
+				host->spi->mode |= (SPI_CPOL|SPI_CPHA);
+				mres = spi_setup(host->spi);
+				if (mres < 0)
+					dev_dbg(&host->spi->dev,
+						"switch back to SPI mode 3 failed\n");
+			}
+		}
+
+		host->power_mode = ios->power_mode;
+	}
+
+	if (host->spi->max_speed_hz != ios->clock && ios->clock != 0) {
+		int		status;
+
+		host->spi->max_speed_hz = ios->clock;
+		status = spi_setup(host->spi);
+		dev_dbg(&host->spi->dev,
+			"mmc_spi:  clock to %d Hz, %d\n",
+			host->spi->max_speed_hz, status);
+	}
+}
+
+static const struct mmc_host_ops mmc_spi_ops = {
+	.request	= mmc_spi_request,
+	.set_ios	= mmc_spi_set_ios,
+	.get_ro		= mmc_gpio_get_ro,
+	.get_cd		= mmc_gpio_get_cd,
+};
+
+
+/****************************************************************************/
+
+/*
+ * SPI driver implementation
+ */
+
+static irqreturn_t
+mmc_spi_detect_irq(int irq, void *mmc)
+{
+	struct mmc_spi_host *host = mmc_priv(mmc);
+	u16 delay_msec = max(host->pdata->detect_delay, (u16)100);
+
+	mmc_detect_change(mmc, msecs_to_jiffies(delay_msec));
+	return IRQ_HANDLED;
+}
+
+#ifdef CONFIG_HAS_DMA
+static int mmc_spi_dma_alloc(struct mmc_spi_host *host)
+{
+	struct spi_device *spi = host->spi;
+	struct device *dev;
+
+	if (!spi->master->dev.parent->dma_mask)
+		return 0;
+
+	dev = spi->master->dev.parent;
+
+	host->ones_dma = dma_map_single(dev, host->ones, MMC_SPI_BLOCKSIZE,
+					DMA_TO_DEVICE);
+	if (dma_mapping_error(dev, host->ones_dma))
+		return -ENOMEM;
+
+	host->data_dma = dma_map_single(dev, host->data, sizeof(*host->data),
+					DMA_BIDIRECTIONAL);
+	if (dma_mapping_error(dev, host->data_dma)) {
+		dma_unmap_single(dev, host->ones_dma, MMC_SPI_BLOCKSIZE,
+				 DMA_TO_DEVICE);
+		return -ENOMEM;
+	}
+
+	dma_sync_single_for_cpu(dev, host->data_dma, sizeof(*host->data),
+				DMA_BIDIRECTIONAL);
+
+	host->dma_dev = dev;
+	return 0;
+}
+
+static void mmc_spi_dma_free(struct mmc_spi_host *host)
+{
+	if (!host->dma_dev)
+		return;
+
+	dma_unmap_single(host->dma_dev, host->ones_dma, MMC_SPI_BLOCKSIZE,
+			 DMA_TO_DEVICE);
+	dma_unmap_single(host->dma_dev, host->data_dma,	sizeof(*host->data),
+			 DMA_BIDIRECTIONAL);
+}
+#else
+static inline int mmc_spi_dma_alloc(struct mmc_spi_host *host) { return 0; }
+static inline void mmc_spi_dma_free(struct mmc_spi_host *host) {}
+#endif
+
+static int mmc_spi_probe(struct spi_device *spi)
+{
+	void			*ones;
+	struct mmc_host		*mmc;
+	struct mmc_spi_host	*host;
+	int			status;
+	bool			has_ro = false;
+
+	/* We rely on full duplex transfers, mostly to reduce
+	 * per-transfer overheads (by making fewer transfers).
+	 */
+	if (spi->master->flags & SPI_MASTER_HALF_DUPLEX)
+		return -EINVAL;
+
+	/* MMC and SD specs only seem to care that sampling is on the
+	 * rising edge ... meaning SPI modes 0 or 3.  So either SPI mode
+	 * should be legit.  We'll use mode 0 since the steady state is 0,
+	 * which is appropriate for hotplugging, unless the platform data
+	 * specify mode 3 (if hardware is not compatible to mode 0).
+	 */
+	if (spi->mode != SPI_MODE_3)
+		spi->mode = SPI_MODE_0;
+	spi->bits_per_word = 8;
+
+	status = spi_setup(spi);
+	if (status < 0) {
+		dev_dbg(&spi->dev, "needs SPI mode %02x, %d KHz; %d\n",
+				spi->mode, spi->max_speed_hz / 1000,
+				status);
+		return status;
+	}
+
+	/* We need a supply of ones to transmit.  This is the only time
+	 * the CPU touches these, so cache coherency isn't a concern.
+	 *
+	 * NOTE if many systems use more than one MMC-over-SPI connector
+	 * it'd save some memory to share this.  That's evidently rare.
+	 */
+	status = -ENOMEM;
+	ones = kmalloc(MMC_SPI_BLOCKSIZE, GFP_KERNEL);
+	if (!ones)
+		goto nomem;
+	memset(ones, 0xff, MMC_SPI_BLOCKSIZE);
+
+	mmc = mmc_alloc_host(sizeof(*host), &spi->dev);
+	if (!mmc)
+		goto nomem;
+
+	mmc->ops = &mmc_spi_ops;
+	mmc->max_blk_size = MMC_SPI_BLOCKSIZE;
+	mmc->max_segs = MMC_SPI_BLOCKSATONCE;
+	mmc->max_req_size = MMC_SPI_BLOCKSATONCE * MMC_SPI_BLOCKSIZE;
+	mmc->max_blk_count = MMC_SPI_BLOCKSATONCE;
+
+	mmc->caps = MMC_CAP_SPI;
+
+	/* SPI doesn't need the lowspeed device identification thing for
+	 * MMC or SD cards, since it never comes up in open drain mode.
+	 * That's good; some SPI masters can't handle very low speeds!
+	 *
+	 * However, low speed SDIO cards need not handle over 400 KHz;
+	 * that's the only reason not to use a few MHz for f_min (until
+	 * the upper layer reads the target frequency from the CSD).
+	 */
+	mmc->f_min = 400000;
+	mmc->f_max = spi->max_speed_hz;
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+	host->spi = spi;
+
+	host->ones = ones;
+
+	/* Platform data is used to hook up things like card sensing
+	 * and power switching gpios.
+	 */
+	host->pdata = mmc_spi_get_pdata(spi);
+	if (host->pdata)
+		mmc->ocr_avail = host->pdata->ocr_mask;
+	if (!mmc->ocr_avail) {
+		dev_warn(&spi->dev, "ASSUMING 3.2-3.4 V slot power\n");
+		mmc->ocr_avail = MMC_VDD_32_33|MMC_VDD_33_34;
+	}
+	if (host->pdata && host->pdata->setpower) {
+		host->powerup_msecs = host->pdata->powerup_msecs;
+		if (!host->powerup_msecs || host->powerup_msecs > 250)
+			host->powerup_msecs = 250;
+	}
+
+	dev_set_drvdata(&spi->dev, mmc);
+
+	/* preallocate dma buffers */
+	host->data = kmalloc(sizeof(*host->data), GFP_KERNEL);
+	if (!host->data)
+		goto fail_nobuf1;
+
+	status = mmc_spi_dma_alloc(host);
+	if (status)
+		goto fail_dma;
+
+	/* setup message for status/busy readback */
+	spi_message_init(&host->readback);
+	host->readback.is_dma_mapped = (host->dma_dev != NULL);
+
+	spi_message_add_tail(&host->status, &host->readback);
+	host->status.tx_buf = host->ones;
+	host->status.tx_dma = host->ones_dma;
+	host->status.rx_buf = &host->data->status;
+	host->status.rx_dma = host->data_dma + offsetof(struct scratch, status);
+	host->status.cs_change = 1;
+
+	/* register card detect irq */
+	if (host->pdata && host->pdata->init) {
+		status = host->pdata->init(&spi->dev, mmc_spi_detect_irq, mmc);
+		if (status != 0)
+			goto fail_glue_init;
+	}
+
+	/* pass platform capabilities, if any */
+	if (host->pdata) {
+		mmc->caps |= host->pdata->caps;
+		mmc->caps2 |= host->pdata->caps2;
+	}
+
+	status = mmc_add_host(mmc);
+	if (status != 0)
+		goto fail_glue_init;
+
+	/*
+	 * Index 0 is card detect
+	 * Old boardfiles were specifying 1 ms as debounce
+	 */
+	status = mmc_gpiod_request_cd(mmc, NULL, 0, false, 1000);
+	if (status == -EPROBE_DEFER)
+		goto fail_gpiod_request;
+	if (!status) {
+		/*
+		 * The platform has a CD GPIO signal that may support
+		 * interrupts, so let mmc_gpiod_request_cd_irq() decide
+		 * if polling is needed or not.
+		 */
+		mmc->caps &= ~MMC_CAP_NEEDS_POLL;
+		mmc_gpiod_request_cd_irq(mmc);
+	}
+	mmc_detect_change(mmc, 0);
+
+	/* Index 1 is write protect/read only */
+	status = mmc_gpiod_request_ro(mmc, NULL, 1, 0);
+	if (status == -EPROBE_DEFER)
+		goto fail_gpiod_request;
+	if (!status)
+		has_ro = true;
+
+	dev_info(&spi->dev, "SD/MMC host %s%s%s%s%s\n",
+			dev_name(&mmc->class_dev),
+			host->dma_dev ? "" : ", no DMA",
+			has_ro ? "" : ", no WP",
+			(host->pdata && host->pdata->setpower)
+				? "" : ", no poweroff",
+			(mmc->caps & MMC_CAP_NEEDS_POLL)
+				? ", cd polling" : "");
+	return 0;
+
+fail_gpiod_request:
+	mmc_remove_host(mmc);
+fail_glue_init:
+	mmc_spi_dma_free(host);
+fail_dma:
+	kfree(host->data);
+fail_nobuf1:
+	mmc_free_host(mmc);
+	mmc_spi_put_pdata(spi);
+nomem:
+	kfree(ones);
+	return status;
+}
+
+
+static int mmc_spi_remove(struct spi_device *spi)
+{
+	struct mmc_host		*mmc = dev_get_drvdata(&spi->dev);
+	struct mmc_spi_host	*host = mmc_priv(mmc);
+
+	/* prevent new mmc_detect_change() calls */
+	if (host->pdata && host->pdata->exit)
+		host->pdata->exit(&spi->dev, mmc);
+
+	mmc_remove_host(mmc);
+
+	mmc_spi_dma_free(host);
+	kfree(host->data);
+	kfree(host->ones);
+
+	spi->max_speed_hz = mmc->f_max;
+	mmc_free_host(mmc);
+	mmc_spi_put_pdata(spi);
+	return 0;
+}
+
+static const struct of_device_id mmc_spi_of_match_table[] = {
+	{ .compatible = "mmc-spi-slot", },
+	{},
+};
+MODULE_DEVICE_TABLE(of, mmc_spi_of_match_table);
+
+static struct spi_driver mmc_spi_driver = {
+	.driver = {
+		.name =		"mmc_spi",
+		.of_match_table = mmc_spi_of_match_table,
+	},
+	.probe =	mmc_spi_probe,
+	.remove =	mmc_spi_remove,
+};
+
+module_spi_driver(mmc_spi_driver);
+
+MODULE_AUTHOR("Mike Lavender, David Brownell, Hans-Peter Nilsson, Jan Nikitenko");
+MODULE_DESCRIPTION("SPI SD/MMC host driver");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("spi:mmc_spi");
diff --git a/drivers/mmc/host/mmci.c b/drivers/mmc/host/mmci.c
new file mode 100644
index 000000000..a72ca7416
--- /dev/null
+++ b/drivers/mmc/host/mmci.c
@@ -0,0 +1,2401 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ *  linux/drivers/mmc/host/mmci.c - ARM PrimeCell MMCI PL180/1 driver
+ *
+ *  Copyright (C) 2003 Deep Blue Solutions, Ltd, All Rights Reserved.
+ *  Copyright (C) 2010 ST-Ericsson SA
+ */
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/init.h>
+#include <linux/ioport.h>
+#include <linux/device.h>
+#include <linux/io.h>
+#include <linux/interrupt.h>
+#include <linux/kernel.h>
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/highmem.h>
+#include <linux/log2.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/pm.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/sd.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/amba/bus.h>
+#include <linux/clk.h>
+#include <linux/scatterlist.h>
+#include <linux/of.h>
+#include <linux/regulator/consumer.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/amba/mmci.h>
+#include <linux/pm_runtime.h>
+#include <linux/types.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/reset.h>
+
+#include <asm/div64.h>
+#include <asm/io.h>
+
+#include "mmci.h"
+
+#define DRIVER_NAME "mmci-pl18x"
+
+static void mmci_variant_init(struct mmci_host *host);
+static void ux500_variant_init(struct mmci_host *host);
+static void ux500v2_variant_init(struct mmci_host *host);
+
+static unsigned int fmax = 515633;
+
+static struct variant_data variant_arm = {
+	.fifosize		= 16 * 4,
+	.fifohalfsize		= 8 * 4,
+	.cmdreg_cpsm_enable	= MCI_CPSM_ENABLE,
+	.cmdreg_lrsp_crc	= MCI_CPSM_RESPONSE | MCI_CPSM_LONGRSP,
+	.cmdreg_srsp_crc	= MCI_CPSM_RESPONSE,
+	.cmdreg_srsp		= MCI_CPSM_RESPONSE,
+	.datalength_bits	= 16,
+	.datactrl_blocksz	= 11,
+	.pwrreg_powerup		= MCI_PWR_UP,
+	.f_max			= 100000000,
+	.reversed_irq_handling	= true,
+	.mmcimask1		= true,
+	.irq_pio_mask		= MCI_IRQ_PIO_MASK,
+	.start_err		= MCI_STARTBITERR,
+	.opendrain		= MCI_ROD,
+	.init			= mmci_variant_init,
+};
+
+static struct variant_data variant_arm_extended_fifo = {
+	.fifosize		= 128 * 4,
+	.fifohalfsize		= 64 * 4,
+	.cmdreg_cpsm_enable	= MCI_CPSM_ENABLE,
+	.cmdreg_lrsp_crc	= MCI_CPSM_RESPONSE | MCI_CPSM_LONGRSP,
+	.cmdreg_srsp_crc	= MCI_CPSM_RESPONSE,
+	.cmdreg_srsp		= MCI_CPSM_RESPONSE,
+	.datalength_bits	= 16,
+	.datactrl_blocksz	= 11,
+	.pwrreg_powerup		= MCI_PWR_UP,
+	.f_max			= 100000000,
+	.mmcimask1		= true,
+	.irq_pio_mask		= MCI_IRQ_PIO_MASK,
+	.start_err		= MCI_STARTBITERR,
+	.opendrain		= MCI_ROD,
+	.init			= mmci_variant_init,
+};
+
+static struct variant_data variant_arm_extended_fifo_hwfc = {
+	.fifosize		= 128 * 4,
+	.fifohalfsize		= 64 * 4,
+	.clkreg_enable		= MCI_ARM_HWFCEN,
+	.cmdreg_cpsm_enable	= MCI_CPSM_ENABLE,
+	.cmdreg_lrsp_crc	= MCI_CPSM_RESPONSE | MCI_CPSM_LONGRSP,
+	.cmdreg_srsp_crc	= MCI_CPSM_RESPONSE,
+	.cmdreg_srsp		= MCI_CPSM_RESPONSE,
+	.datalength_bits	= 16,
+	.datactrl_blocksz	= 11,
+	.pwrreg_powerup		= MCI_PWR_UP,
+	.f_max			= 100000000,
+	.mmcimask1		= true,
+	.irq_pio_mask		= MCI_IRQ_PIO_MASK,
+	.start_err		= MCI_STARTBITERR,
+	.opendrain		= MCI_ROD,
+	.init			= mmci_variant_init,
+};
+
+static struct variant_data variant_u300 = {
+	.fifosize		= 16 * 4,
+	.fifohalfsize		= 8 * 4,
+	.clkreg_enable		= MCI_ST_U300_HWFCEN,
+	.clkreg_8bit_bus_enable = MCI_ST_8BIT_BUS,
+	.cmdreg_cpsm_enable	= MCI_CPSM_ENABLE,
+	.cmdreg_lrsp_crc	= MCI_CPSM_RESPONSE | MCI_CPSM_LONGRSP,
+	.cmdreg_srsp_crc	= MCI_CPSM_RESPONSE,
+	.cmdreg_srsp		= MCI_CPSM_RESPONSE,
+	.datalength_bits	= 16,
+	.datactrl_blocksz	= 11,
+	.datactrl_mask_sdio	= MCI_DPSM_ST_SDIOEN,
+	.st_sdio			= true,
+	.pwrreg_powerup		= MCI_PWR_ON,
+	.f_max			= 100000000,
+	.signal_direction	= true,
+	.pwrreg_clkgate		= true,
+	.pwrreg_nopower		= true,
+	.mmcimask1		= true,
+	.irq_pio_mask		= MCI_IRQ_PIO_MASK,
+	.start_err		= MCI_STARTBITERR,
+	.opendrain		= MCI_OD,
+	.init			= mmci_variant_init,
+};
+
+static struct variant_data variant_nomadik = {
+	.fifosize		= 16 * 4,
+	.fifohalfsize		= 8 * 4,
+	.clkreg			= MCI_CLK_ENABLE,
+	.clkreg_8bit_bus_enable = MCI_ST_8BIT_BUS,
+	.cmdreg_cpsm_enable	= MCI_CPSM_ENABLE,
+	.cmdreg_lrsp_crc	= MCI_CPSM_RESPONSE | MCI_CPSM_LONGRSP,
+	.cmdreg_srsp_crc	= MCI_CPSM_RESPONSE,
+	.cmdreg_srsp		= MCI_CPSM_RESPONSE,
+	.datalength_bits	= 24,
+	.datactrl_blocksz	= 11,
+	.datactrl_mask_sdio	= MCI_DPSM_ST_SDIOEN,
+	.st_sdio		= true,
+	.st_clkdiv		= true,
+	.pwrreg_powerup		= MCI_PWR_ON,
+	.f_max			= 100000000,
+	.signal_direction	= true,
+	.pwrreg_clkgate		= true,
+	.pwrreg_nopower		= true,
+	.mmcimask1		= true,
+	.irq_pio_mask		= MCI_IRQ_PIO_MASK,
+	.start_err		= MCI_STARTBITERR,
+	.opendrain		= MCI_OD,
+	.init			= mmci_variant_init,
+};
+
+static struct variant_data variant_ux500 = {
+	.fifosize		= 30 * 4,
+	.fifohalfsize		= 8 * 4,
+	.clkreg			= MCI_CLK_ENABLE,
+	.clkreg_enable		= MCI_ST_UX500_HWFCEN,
+	.clkreg_8bit_bus_enable = MCI_ST_8BIT_BUS,
+	.clkreg_neg_edge_enable	= MCI_ST_UX500_NEG_EDGE,
+	.cmdreg_cpsm_enable	= MCI_CPSM_ENABLE,
+	.cmdreg_lrsp_crc	= MCI_CPSM_RESPONSE | MCI_CPSM_LONGRSP,
+	.cmdreg_srsp_crc	= MCI_CPSM_RESPONSE,
+	.cmdreg_srsp		= MCI_CPSM_RESPONSE,
+	.datalength_bits	= 24,
+	.datactrl_blocksz	= 11,
+	.datactrl_any_blocksz	= true,
+	.dma_power_of_2		= true,
+	.datactrl_mask_sdio	= MCI_DPSM_ST_SDIOEN,
+	.st_sdio		= true,
+	.st_clkdiv		= true,
+	.pwrreg_powerup		= MCI_PWR_ON,
+	.f_max			= 100000000,
+	.signal_direction	= true,
+	.pwrreg_clkgate		= true,
+	.busy_detect		= true,
+	.busy_dpsm_flag		= MCI_DPSM_ST_BUSYMODE,
+	.busy_detect_flag	= MCI_ST_CARDBUSY,
+	.busy_detect_mask	= MCI_ST_BUSYENDMASK,
+	.pwrreg_nopower		= true,
+	.mmcimask1		= true,
+	.irq_pio_mask		= MCI_IRQ_PIO_MASK,
+	.start_err		= MCI_STARTBITERR,
+	.opendrain		= MCI_OD,
+	.init			= ux500_variant_init,
+};
+
+static struct variant_data variant_ux500v2 = {
+	.fifosize		= 30 * 4,
+	.fifohalfsize		= 8 * 4,
+	.clkreg			= MCI_CLK_ENABLE,
+	.clkreg_enable		= MCI_ST_UX500_HWFCEN,
+	.clkreg_8bit_bus_enable = MCI_ST_8BIT_BUS,
+	.clkreg_neg_edge_enable	= MCI_ST_UX500_NEG_EDGE,
+	.cmdreg_cpsm_enable	= MCI_CPSM_ENABLE,
+	.cmdreg_lrsp_crc	= MCI_CPSM_RESPONSE | MCI_CPSM_LONGRSP,
+	.cmdreg_srsp_crc	= MCI_CPSM_RESPONSE,
+	.cmdreg_srsp		= MCI_CPSM_RESPONSE,
+	.datactrl_mask_ddrmode	= MCI_DPSM_ST_DDRMODE,
+	.datalength_bits	= 24,
+	.datactrl_blocksz	= 11,
+	.datactrl_any_blocksz	= true,
+	.dma_power_of_2		= true,
+	.datactrl_mask_sdio	= MCI_DPSM_ST_SDIOEN,
+	.st_sdio		= true,
+	.st_clkdiv		= true,
+	.pwrreg_powerup		= MCI_PWR_ON,
+	.f_max			= 100000000,
+	.signal_direction	= true,
+	.pwrreg_clkgate		= true,
+	.busy_detect		= true,
+	.busy_dpsm_flag		= MCI_DPSM_ST_BUSYMODE,
+	.busy_detect_flag	= MCI_ST_CARDBUSY,
+	.busy_detect_mask	= MCI_ST_BUSYENDMASK,
+	.pwrreg_nopower		= true,
+	.mmcimask1		= true,
+	.irq_pio_mask		= MCI_IRQ_PIO_MASK,
+	.start_err		= MCI_STARTBITERR,
+	.opendrain		= MCI_OD,
+	.init			= ux500v2_variant_init,
+};
+
+static struct variant_data variant_stm32 = {
+	.fifosize		= 32 * 4,
+	.fifohalfsize		= 8 * 4,
+	.clkreg			= MCI_CLK_ENABLE,
+	.clkreg_enable		= MCI_ST_UX500_HWFCEN,
+	.clkreg_8bit_bus_enable = MCI_ST_8BIT_BUS,
+	.clkreg_neg_edge_enable	= MCI_ST_UX500_NEG_EDGE,
+	.cmdreg_cpsm_enable	= MCI_CPSM_ENABLE,
+	.cmdreg_lrsp_crc	= MCI_CPSM_RESPONSE | MCI_CPSM_LONGRSP,
+	.cmdreg_srsp_crc	= MCI_CPSM_RESPONSE,
+	.cmdreg_srsp		= MCI_CPSM_RESPONSE,
+	.irq_pio_mask		= MCI_IRQ_PIO_MASK,
+	.datalength_bits	= 24,
+	.datactrl_blocksz	= 11,
+	.datactrl_mask_sdio	= MCI_DPSM_ST_SDIOEN,
+	.st_sdio		= true,
+	.st_clkdiv		= true,
+	.pwrreg_powerup		= MCI_PWR_ON,
+	.f_max			= 48000000,
+	.pwrreg_clkgate		= true,
+	.pwrreg_nopower		= true,
+	.init			= mmci_variant_init,
+};
+
+static struct variant_data variant_stm32_sdmmc = {
+	.fifosize		= 16 * 4,
+	.fifohalfsize		= 8 * 4,
+	.f_max			= 208000000,
+	.stm32_clkdiv		= true,
+	.cmdreg_cpsm_enable	= MCI_CPSM_STM32_ENABLE,
+	.cmdreg_lrsp_crc	= MCI_CPSM_STM32_LRSP_CRC,
+	.cmdreg_srsp_crc	= MCI_CPSM_STM32_SRSP_CRC,
+	.cmdreg_srsp		= MCI_CPSM_STM32_SRSP,
+	.cmdreg_stop		= MCI_CPSM_STM32_CMDSTOP,
+	.data_cmd_enable	= MCI_CPSM_STM32_CMDTRANS,
+	.irq_pio_mask		= MCI_IRQ_PIO_STM32_MASK,
+	.datactrl_first		= true,
+	.datacnt_useless	= true,
+	.datalength_bits	= 25,
+	.datactrl_blocksz	= 14,
+	.datactrl_any_blocksz	= true,
+	.datactrl_mask_sdio	= MCI_DPSM_ST_SDIOEN,
+	.stm32_idmabsize_mask	= GENMASK(12, 5),
+	.busy_timeout		= true,
+	.busy_detect		= true,
+	.busy_detect_flag	= MCI_STM32_BUSYD0,
+	.busy_detect_mask	= MCI_STM32_BUSYD0ENDMASK,
+	.init			= sdmmc_variant_init,
+};
+
+static struct variant_data variant_stm32_sdmmcv2 = {
+	.fifosize		= 16 * 4,
+	.fifohalfsize		= 8 * 4,
+	.f_max			= 208000000,
+	.stm32_clkdiv		= true,
+	.cmdreg_cpsm_enable	= MCI_CPSM_STM32_ENABLE,
+	.cmdreg_lrsp_crc	= MCI_CPSM_STM32_LRSP_CRC,
+	.cmdreg_srsp_crc	= MCI_CPSM_STM32_SRSP_CRC,
+	.cmdreg_srsp		= MCI_CPSM_STM32_SRSP,
+	.cmdreg_stop		= MCI_CPSM_STM32_CMDSTOP,
+	.data_cmd_enable	= MCI_CPSM_STM32_CMDTRANS,
+	.irq_pio_mask		= MCI_IRQ_PIO_STM32_MASK,
+	.datactrl_first		= true,
+	.datacnt_useless	= true,
+	.datalength_bits	= 25,
+	.datactrl_blocksz	= 14,
+	.datactrl_any_blocksz	= true,
+	.datactrl_mask_sdio	= MCI_DPSM_ST_SDIOEN,
+	.stm32_idmabsize_mask	= GENMASK(16, 5),
+	.dma_lli		= true,
+	.busy_timeout		= true,
+	.busy_detect		= true,
+	.busy_detect_flag	= MCI_STM32_BUSYD0,
+	.busy_detect_mask	= MCI_STM32_BUSYD0ENDMASK,
+	.init			= sdmmc_variant_init,
+};
+
+static struct variant_data variant_qcom = {
+	.fifosize		= 16 * 4,
+	.fifohalfsize		= 8 * 4,
+	.clkreg			= MCI_CLK_ENABLE,
+	.clkreg_enable		= MCI_QCOM_CLK_FLOWENA |
+				  MCI_QCOM_CLK_SELECT_IN_FBCLK,
+	.clkreg_8bit_bus_enable = MCI_QCOM_CLK_WIDEBUS_8,
+	.datactrl_mask_ddrmode	= MCI_QCOM_CLK_SELECT_IN_DDR_MODE,
+	.cmdreg_cpsm_enable	= MCI_CPSM_ENABLE,
+	.cmdreg_lrsp_crc	= MCI_CPSM_RESPONSE | MCI_CPSM_LONGRSP,
+	.cmdreg_srsp_crc	= MCI_CPSM_RESPONSE,
+	.cmdreg_srsp		= MCI_CPSM_RESPONSE,
+	.data_cmd_enable	= MCI_CPSM_QCOM_DATCMD,
+	.datalength_bits	= 24,
+	.datactrl_blocksz	= 11,
+	.datactrl_any_blocksz	= true,
+	.pwrreg_powerup		= MCI_PWR_UP,
+	.f_max			= 208000000,
+	.explicit_mclk_control	= true,
+	.qcom_fifo		= true,
+	.qcom_dml		= true,
+	.mmcimask1		= true,
+	.irq_pio_mask		= MCI_IRQ_PIO_MASK,
+	.start_err		= MCI_STARTBITERR,
+	.opendrain		= MCI_ROD,
+	.init			= qcom_variant_init,
+};
+
+/* Busy detection for the ST Micro variant */
+static int mmci_card_busy(struct mmc_host *mmc)
+{
+	struct mmci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+	int busy = 0;
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (readl(host->base + MMCISTATUS) & host->variant->busy_detect_flag)
+		busy = 1;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	return busy;
+}
+
+static void mmci_reg_delay(struct mmci_host *host)
+{
+	/*
+	 * According to the spec, at least three feedback clock cycles
+	 * of max 52 MHz must pass between two writes to the MMCICLOCK reg.
+	 * Three MCLK clock cycles must pass between two MMCIPOWER reg writes.
+	 * Worst delay time during card init is at 100 kHz => 30 us.
+	 * Worst delay time when up and running is at 25 MHz => 120 ns.
+	 */
+	if (host->cclk < 25000000)
+		udelay(30);
+	else
+		ndelay(120);
+}
+
+/*
+ * This must be called with host->lock held
+ */
+void mmci_write_clkreg(struct mmci_host *host, u32 clk)
+{
+	if (host->clk_reg != clk) {
+		host->clk_reg = clk;
+		writel(clk, host->base + MMCICLOCK);
+	}
+}
+
+/*
+ * This must be called with host->lock held
+ */
+void mmci_write_pwrreg(struct mmci_host *host, u32 pwr)
+{
+	if (host->pwr_reg != pwr) {
+		host->pwr_reg = pwr;
+		writel(pwr, host->base + MMCIPOWER);
+	}
+}
+
+/*
+ * This must be called with host->lock held
+ */
+static void mmci_write_datactrlreg(struct mmci_host *host, u32 datactrl)
+{
+	/* Keep busy mode in DPSM if enabled */
+	datactrl |= host->datactrl_reg & host->variant->busy_dpsm_flag;
+
+	if (host->datactrl_reg != datactrl) {
+		host->datactrl_reg = datactrl;
+		writel(datactrl, host->base + MMCIDATACTRL);
+	}
+}
+
+/*
+ * This must be called with host->lock held
+ */
+static void mmci_set_clkreg(struct mmci_host *host, unsigned int desired)
+{
+	struct variant_data *variant = host->variant;
+	u32 clk = variant->clkreg;
+
+	/* Make sure cclk reflects the current calculated clock */
+	host->cclk = 0;
+
+	if (desired) {
+		if (variant->explicit_mclk_control) {
+			host->cclk = host->mclk;
+		} else if (desired >= host->mclk) {
+			clk = MCI_CLK_BYPASS;
+			if (variant->st_clkdiv)
+				clk |= MCI_ST_UX500_NEG_EDGE;
+			host->cclk = host->mclk;
+		} else if (variant->st_clkdiv) {
+			/*
+			 * DB8500 TRM says f = mclk / (clkdiv + 2)
+			 * => clkdiv = (mclk / f) - 2
+			 * Round the divider up so we don't exceed the max
+			 * frequency
+			 */
+			clk = DIV_ROUND_UP(host->mclk, desired) - 2;
+			if (clk >= 256)
+				clk = 255;
+			host->cclk = host->mclk / (clk + 2);
+		} else {
+			/*
+			 * PL180 TRM says f = mclk / (2 * (clkdiv + 1))
+			 * => clkdiv = mclk / (2 * f) - 1
+			 */
+			clk = host->mclk / (2 * desired) - 1;
+			if (clk >= 256)
+				clk = 255;
+			host->cclk = host->mclk / (2 * (clk + 1));
+		}
+
+		clk |= variant->clkreg_enable;
+		clk |= MCI_CLK_ENABLE;
+		/* This hasn't proven to be worthwhile */
+		/* clk |= MCI_CLK_PWRSAVE; */
+	}
+
+	/* Set actual clock for debug */
+	host->mmc->actual_clock = host->cclk;
+
+	if (host->mmc->ios.bus_width == MMC_BUS_WIDTH_4)
+		clk |= MCI_4BIT_BUS;
+	if (host->mmc->ios.bus_width == MMC_BUS_WIDTH_8)
+		clk |= variant->clkreg_8bit_bus_enable;
+
+	if (host->mmc->ios.timing == MMC_TIMING_UHS_DDR50 ||
+	    host->mmc->ios.timing == MMC_TIMING_MMC_DDR52)
+		clk |= variant->clkreg_neg_edge_enable;
+
+	mmci_write_clkreg(host, clk);
+}
+
+static void mmci_dma_release(struct mmci_host *host)
+{
+	if (host->ops && host->ops->dma_release)
+		host->ops->dma_release(host);
+
+	host->use_dma = false;
+}
+
+static void mmci_dma_setup(struct mmci_host *host)
+{
+	if (!host->ops || !host->ops->dma_setup)
+		return;
+
+	if (host->ops->dma_setup(host))
+		return;
+
+	/* initialize pre request cookie */
+	host->next_cookie = 1;
+
+	host->use_dma = true;
+}
+
+/*
+ * Validate mmc prerequisites
+ */
+static int mmci_validate_data(struct mmci_host *host,
+			      struct mmc_data *data)
+{
+	struct variant_data *variant = host->variant;
+
+	if (!data)
+		return 0;
+	if (!is_power_of_2(data->blksz) && !variant->datactrl_any_blocksz) {
+		dev_err(mmc_dev(host->mmc),
+			"unsupported block size (%d bytes)\n", data->blksz);
+		return -EINVAL;
+	}
+
+	if (host->ops && host->ops->validate_data)
+		return host->ops->validate_data(host, data);
+
+	return 0;
+}
+
+static int mmci_prep_data(struct mmci_host *host, struct mmc_data *data, bool next)
+{
+	int err;
+
+	if (!host->ops || !host->ops->prep_data)
+		return 0;
+
+	err = host->ops->prep_data(host, data, next);
+
+	if (next && !err)
+		data->host_cookie = ++host->next_cookie < 0 ?
+			1 : host->next_cookie;
+
+	return err;
+}
+
+static void mmci_unprep_data(struct mmci_host *host, struct mmc_data *data,
+		      int err)
+{
+	if (host->ops && host->ops->unprep_data)
+		host->ops->unprep_data(host, data, err);
+
+	data->host_cookie = 0;
+}
+
+static void mmci_get_next_data(struct mmci_host *host, struct mmc_data *data)
+{
+	WARN_ON(data->host_cookie && data->host_cookie != host->next_cookie);
+
+	if (host->ops && host->ops->get_next_data)
+		host->ops->get_next_data(host, data);
+}
+
+static int mmci_dma_start(struct mmci_host *host, unsigned int datactrl)
+{
+	struct mmc_data *data = host->data;
+	int ret;
+
+	if (!host->use_dma)
+		return -EINVAL;
+
+	ret = mmci_prep_data(host, data, false);
+	if (ret)
+		return ret;
+
+	if (!host->ops || !host->ops->dma_start)
+		return -EINVAL;
+
+	/* Okay, go for it. */
+	dev_vdbg(mmc_dev(host->mmc),
+		 "Submit MMCI DMA job, sglen %d blksz %04x blks %04x flags %08x\n",
+		 data->sg_len, data->blksz, data->blocks, data->flags);
+
+	ret = host->ops->dma_start(host, &datactrl);
+	if (ret)
+		return ret;
+
+	/* Trigger the DMA transfer */
+	mmci_write_datactrlreg(host, datactrl);
+
+	/*
+	 * Let the MMCI say when the data is ended and it's time
+	 * to fire next DMA request. When that happens, MMCI will
+	 * call mmci_data_end()
+	 */
+	writel(readl(host->base + MMCIMASK0) | MCI_DATAENDMASK,
+	       host->base + MMCIMASK0);
+	return 0;
+}
+
+static void mmci_dma_finalize(struct mmci_host *host, struct mmc_data *data)
+{
+	if (!host->use_dma)
+		return;
+
+	if (host->ops && host->ops->dma_finalize)
+		host->ops->dma_finalize(host, data);
+}
+
+static void mmci_dma_error(struct mmci_host *host)
+{
+	if (!host->use_dma)
+		return;
+
+	if (host->ops && host->ops->dma_error)
+		host->ops->dma_error(host);
+}
+
+static void
+mmci_request_end(struct mmci_host *host, struct mmc_request *mrq)
+{
+	writel(0, host->base + MMCICOMMAND);
+
+	BUG_ON(host->data);
+
+	host->mrq = NULL;
+	host->cmd = NULL;
+
+	mmc_request_done(host->mmc, mrq);
+}
+
+static void mmci_set_mask1(struct mmci_host *host, unsigned int mask)
+{
+	void __iomem *base = host->base;
+	struct variant_data *variant = host->variant;
+
+	if (host->singleirq) {
+		unsigned int mask0 = readl(base + MMCIMASK0);
+
+		mask0 &= ~variant->irq_pio_mask;
+		mask0 |= mask;
+
+		writel(mask0, base + MMCIMASK0);
+	}
+
+	if (variant->mmcimask1)
+		writel(mask, base + MMCIMASK1);
+
+	host->mask1_reg = mask;
+}
+
+static void mmci_stop_data(struct mmci_host *host)
+{
+	mmci_write_datactrlreg(host, 0);
+	mmci_set_mask1(host, 0);
+	host->data = NULL;
+}
+
+static void mmci_init_sg(struct mmci_host *host, struct mmc_data *data)
+{
+	unsigned int flags = SG_MITER_ATOMIC;
+
+	if (data->flags & MMC_DATA_READ)
+		flags |= SG_MITER_TO_SG;
+	else
+		flags |= SG_MITER_FROM_SG;
+
+	sg_miter_start(&host->sg_miter, data->sg, data->sg_len, flags);
+}
+
+static u32 mmci_get_dctrl_cfg(struct mmci_host *host)
+{
+	return MCI_DPSM_ENABLE | mmci_dctrl_blksz(host);
+}
+
+static u32 ux500v2_get_dctrl_cfg(struct mmci_host *host)
+{
+	return MCI_DPSM_ENABLE | (host->data->blksz << 16);
+}
+
+static bool ux500_busy_complete(struct mmci_host *host, u32 status, u32 err_msk)
+{
+	void __iomem *base = host->base;
+
+	/*
+	 * Before unmasking for the busy end IRQ, confirm that the
+	 * command was sent successfully. To keep track of having a
+	 * command in-progress, waiting for busy signaling to end,
+	 * store the status in host->busy_status.
+	 *
+	 * Note that, the card may need a couple of clock cycles before
+	 * it starts signaling busy on DAT0, hence re-read the
+	 * MMCISTATUS register here, to allow the busy bit to be set.
+	 * Potentially we may even need to poll the register for a
+	 * while, to allow it to be set, but tests indicates that it
+	 * isn't needed.
+	 */
+	if (!host->busy_status && !(status & err_msk) &&
+	    (readl(base + MMCISTATUS) & host->variant->busy_detect_flag)) {
+		writel(readl(base + MMCIMASK0) |
+		       host->variant->busy_detect_mask,
+		       base + MMCIMASK0);
+
+		host->busy_status = status & (MCI_CMDSENT | MCI_CMDRESPEND);
+		return false;
+	}
+
+	/*
+	 * If there is a command in-progress that has been successfully
+	 * sent, then bail out if busy status is set and wait for the
+	 * busy end IRQ.
+	 *
+	 * Note that, the HW triggers an IRQ on both edges while
+	 * monitoring DAT0 for busy completion, but there is only one
+	 * status bit in MMCISTATUS for the busy state. Therefore
+	 * both the start and the end interrupts needs to be cleared,
+	 * one after the other. So, clear the busy start IRQ here.
+	 */
+	if (host->busy_status &&
+	    (status & host->variant->busy_detect_flag)) {
+		writel(host->variant->busy_detect_mask, base + MMCICLEAR);
+		return false;
+	}
+
+	/*
+	 * If there is a command in-progress that has been successfully
+	 * sent and the busy bit isn't set, it means we have received
+	 * the busy end IRQ. Clear and mask the IRQ, then continue to
+	 * process the command.
+	 */
+	if (host->busy_status) {
+		writel(host->variant->busy_detect_mask, base + MMCICLEAR);
+
+		writel(readl(base + MMCIMASK0) &
+		       ~host->variant->busy_detect_mask, base + MMCIMASK0);
+		host->busy_status = 0;
+	}
+
+	return true;
+}
+
+/*
+ * All the DMA operation mode stuff goes inside this ifdef.
+ * This assumes that you have a generic DMA device interface,
+ * no custom DMA interfaces are supported.
+ */
+#ifdef CONFIG_DMA_ENGINE
+struct mmci_dmae_next {
+	struct dma_async_tx_descriptor *desc;
+	struct dma_chan	*chan;
+};
+
+struct mmci_dmae_priv {
+	struct dma_chan	*cur;
+	struct dma_chan	*rx_channel;
+	struct dma_chan	*tx_channel;
+	struct dma_async_tx_descriptor	*desc_current;
+	struct mmci_dmae_next next_data;
+};
+
+int mmci_dmae_setup(struct mmci_host *host)
+{
+	const char *rxname, *txname;
+	struct mmci_dmae_priv *dmae;
+
+	dmae = devm_kzalloc(mmc_dev(host->mmc), sizeof(*dmae), GFP_KERNEL);
+	if (!dmae)
+		return -ENOMEM;
+
+	host->dma_priv = dmae;
+
+	dmae->rx_channel = dma_request_chan(mmc_dev(host->mmc), "rx");
+	if (IS_ERR(dmae->rx_channel)) {
+		int ret = PTR_ERR(dmae->rx_channel);
+		dmae->rx_channel = NULL;
+		return ret;
+	}
+
+	dmae->tx_channel = dma_request_chan(mmc_dev(host->mmc), "tx");
+	if (IS_ERR(dmae->tx_channel)) {
+		if (PTR_ERR(dmae->tx_channel) == -EPROBE_DEFER)
+			dev_warn(mmc_dev(host->mmc),
+				 "Deferred probe for TX channel ignored\n");
+		dmae->tx_channel = NULL;
+	}
+
+	/*
+	 * If only an RX channel is specified, the driver will
+	 * attempt to use it bidirectionally, however if it is
+	 * is specified but cannot be located, DMA will be disabled.
+	 */
+	if (dmae->rx_channel && !dmae->tx_channel)
+		dmae->tx_channel = dmae->rx_channel;
+
+	if (dmae->rx_channel)
+		rxname = dma_chan_name(dmae->rx_channel);
+	else
+		rxname = "none";
+
+	if (dmae->tx_channel)
+		txname = dma_chan_name(dmae->tx_channel);
+	else
+		txname = "none";
+
+	dev_info(mmc_dev(host->mmc), "DMA channels RX %s, TX %s\n",
+		 rxname, txname);
+
+	/*
+	 * Limit the maximum segment size in any SG entry according to
+	 * the parameters of the DMA engine device.
+	 */
+	if (dmae->tx_channel) {
+		struct device *dev = dmae->tx_channel->device->dev;
+		unsigned int max_seg_size = dma_get_max_seg_size(dev);
+
+		if (max_seg_size < host->mmc->max_seg_size)
+			host->mmc->max_seg_size = max_seg_size;
+	}
+	if (dmae->rx_channel) {
+		struct device *dev = dmae->rx_channel->device->dev;
+		unsigned int max_seg_size = dma_get_max_seg_size(dev);
+
+		if (max_seg_size < host->mmc->max_seg_size)
+			host->mmc->max_seg_size = max_seg_size;
+	}
+
+	if (!dmae->tx_channel || !dmae->rx_channel) {
+		mmci_dmae_release(host);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+/*
+ * This is used in or so inline it
+ * so it can be discarded.
+ */
+void mmci_dmae_release(struct mmci_host *host)
+{
+	struct mmci_dmae_priv *dmae = host->dma_priv;
+
+	if (dmae->rx_channel)
+		dma_release_channel(dmae->rx_channel);
+	if (dmae->tx_channel)
+		dma_release_channel(dmae->tx_channel);
+	dmae->rx_channel = dmae->tx_channel = NULL;
+}
+
+static void mmci_dma_unmap(struct mmci_host *host, struct mmc_data *data)
+{
+	struct mmci_dmae_priv *dmae = host->dma_priv;
+	struct dma_chan *chan;
+
+	if (data->flags & MMC_DATA_READ)
+		chan = dmae->rx_channel;
+	else
+		chan = dmae->tx_channel;
+
+	dma_unmap_sg(chan->device->dev, data->sg, data->sg_len,
+		     mmc_get_dma_dir(data));
+}
+
+void mmci_dmae_error(struct mmci_host *host)
+{
+	struct mmci_dmae_priv *dmae = host->dma_priv;
+
+	if (!dma_inprogress(host))
+		return;
+
+	dev_err(mmc_dev(host->mmc), "error during DMA transfer!\n");
+	dmaengine_terminate_all(dmae->cur);
+	host->dma_in_progress = false;
+	dmae->cur = NULL;
+	dmae->desc_current = NULL;
+	host->data->host_cookie = 0;
+
+	mmci_dma_unmap(host, host->data);
+}
+
+void mmci_dmae_finalize(struct mmci_host *host, struct mmc_data *data)
+{
+	struct mmci_dmae_priv *dmae = host->dma_priv;
+	u32 status;
+	int i;
+
+	if (!dma_inprogress(host))
+		return;
+
+	/* Wait up to 1ms for the DMA to complete */
+	for (i = 0; ; i++) {
+		status = readl(host->base + MMCISTATUS);
+		if (!(status & MCI_RXDATAAVLBLMASK) || i >= 100)
+			break;
+		udelay(10);
+	}
+
+	/*
+	 * Check to see whether we still have some data left in the FIFO -
+	 * this catches DMA controllers which are unable to monitor the
+	 * DMALBREQ and DMALSREQ signals while allowing us to DMA to non-
+	 * contiguous buffers.  On TX, we'll get a FIFO underrun error.
+	 */
+	if (status & MCI_RXDATAAVLBLMASK) {
+		mmci_dma_error(host);
+		if (!data->error)
+			data->error = -EIO;
+	} else if (!data->host_cookie) {
+		mmci_dma_unmap(host, data);
+	}
+
+	/*
+	 * Use of DMA with scatter-gather is impossible.
+	 * Give up with DMA and switch back to PIO mode.
+	 */
+	if (status & MCI_RXDATAAVLBLMASK) {
+		dev_err(mmc_dev(host->mmc), "buggy DMA detected. Taking evasive action.\n");
+		mmci_dma_release(host);
+	}
+
+	host->dma_in_progress = false;
+	dmae->cur = NULL;
+	dmae->desc_current = NULL;
+}
+
+/* prepares DMA channel and DMA descriptor, returns non-zero on failure */
+static int _mmci_dmae_prep_data(struct mmci_host *host, struct mmc_data *data,
+				struct dma_chan **dma_chan,
+				struct dma_async_tx_descriptor **dma_desc)
+{
+	struct mmci_dmae_priv *dmae = host->dma_priv;
+	struct variant_data *variant = host->variant;
+	struct dma_slave_config conf = {
+		.src_addr = host->phybase + MMCIFIFO,
+		.dst_addr = host->phybase + MMCIFIFO,
+		.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+		.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+		.src_maxburst = variant->fifohalfsize >> 2, /* # of words */
+		.dst_maxburst = variant->fifohalfsize >> 2, /* # of words */
+		.device_fc = false,
+	};
+	struct dma_chan *chan;
+	struct dma_device *device;
+	struct dma_async_tx_descriptor *desc;
+	int nr_sg;
+	unsigned long flags = DMA_CTRL_ACK;
+
+	if (data->flags & MMC_DATA_READ) {
+		conf.direction = DMA_DEV_TO_MEM;
+		chan = dmae->rx_channel;
+	} else {
+		conf.direction = DMA_MEM_TO_DEV;
+		chan = dmae->tx_channel;
+	}
+
+	/* If there's no DMA channel, fall back to PIO */
+	if (!chan)
+		return -EINVAL;
+
+	/* If less than or equal to the fifo size, don't bother with DMA */
+	if (data->blksz * data->blocks <= variant->fifosize)
+		return -EINVAL;
+
+	/*
+	 * This is necessary to get SDIO working on the Ux500. We do not yet
+	 * know if this is a bug in:
+	 * - The Ux500 DMA controller (DMA40)
+	 * - The MMCI DMA interface on the Ux500
+	 * some power of two blocks (such as 64 bytes) are sent regularly
+	 * during SDIO traffic and those work fine so for these we enable DMA
+	 * transfers.
+	 */
+	if (host->variant->dma_power_of_2 && !is_power_of_2(data->blksz))
+		return -EINVAL;
+
+	device = chan->device;
+	nr_sg = dma_map_sg(device->dev, data->sg, data->sg_len,
+			   mmc_get_dma_dir(data));
+	if (nr_sg == 0)
+		return -EINVAL;
+
+	if (host->variant->qcom_dml)
+		flags |= DMA_PREP_INTERRUPT;
+
+	dmaengine_slave_config(chan, &conf);
+	desc = dmaengine_prep_slave_sg(chan, data->sg, nr_sg,
+					    conf.direction, flags);
+	if (!desc)
+		goto unmap_exit;
+
+	*dma_chan = chan;
+	*dma_desc = desc;
+
+	return 0;
+
+ unmap_exit:
+	dma_unmap_sg(device->dev, data->sg, data->sg_len,
+		     mmc_get_dma_dir(data));
+	return -ENOMEM;
+}
+
+int mmci_dmae_prep_data(struct mmci_host *host,
+			struct mmc_data *data,
+			bool next)
+{
+	struct mmci_dmae_priv *dmae = host->dma_priv;
+	struct mmci_dmae_next *nd = &dmae->next_data;
+
+	if (!host->use_dma)
+		return -EINVAL;
+
+	if (next)
+		return _mmci_dmae_prep_data(host, data, &nd->chan, &nd->desc);
+	/* Check if next job is already prepared. */
+	if (dmae->cur && dmae->desc_current)
+		return 0;
+
+	/* No job were prepared thus do it now. */
+	return _mmci_dmae_prep_data(host, data, &dmae->cur,
+				    &dmae->desc_current);
+}
+
+int mmci_dmae_start(struct mmci_host *host, unsigned int *datactrl)
+{
+	struct mmci_dmae_priv *dmae = host->dma_priv;
+	int ret;
+
+	host->dma_in_progress = true;
+	ret = dma_submit_error(dmaengine_submit(dmae->desc_current));
+	if (ret < 0) {
+		host->dma_in_progress = false;
+		return ret;
+	}
+	dma_async_issue_pending(dmae->cur);
+
+	*datactrl |= MCI_DPSM_DMAENABLE;
+
+	return 0;
+}
+
+void mmci_dmae_get_next_data(struct mmci_host *host, struct mmc_data *data)
+{
+	struct mmci_dmae_priv *dmae = host->dma_priv;
+	struct mmci_dmae_next *next = &dmae->next_data;
+
+	if (!host->use_dma)
+		return;
+
+	WARN_ON(!data->host_cookie && (next->desc || next->chan));
+
+	dmae->desc_current = next->desc;
+	dmae->cur = next->chan;
+	next->desc = NULL;
+	next->chan = NULL;
+}
+
+void mmci_dmae_unprep_data(struct mmci_host *host,
+			   struct mmc_data *data, int err)
+
+{
+	struct mmci_dmae_priv *dmae = host->dma_priv;
+
+	if (!host->use_dma)
+		return;
+
+	mmci_dma_unmap(host, data);
+
+	if (err) {
+		struct mmci_dmae_next *next = &dmae->next_data;
+		struct dma_chan *chan;
+		if (data->flags & MMC_DATA_READ)
+			chan = dmae->rx_channel;
+		else
+			chan = dmae->tx_channel;
+		dmaengine_terminate_all(chan);
+
+		if (dmae->desc_current == next->desc)
+			dmae->desc_current = NULL;
+
+		if (dmae->cur == next->chan) {
+			host->dma_in_progress = false;
+			dmae->cur = NULL;
+		}
+
+		next->desc = NULL;
+		next->chan = NULL;
+	}
+}
+
+static struct mmci_host_ops mmci_variant_ops = {
+	.prep_data = mmci_dmae_prep_data,
+	.unprep_data = mmci_dmae_unprep_data,
+	.get_datactrl_cfg = mmci_get_dctrl_cfg,
+	.get_next_data = mmci_dmae_get_next_data,
+	.dma_setup = mmci_dmae_setup,
+	.dma_release = mmci_dmae_release,
+	.dma_start = mmci_dmae_start,
+	.dma_finalize = mmci_dmae_finalize,
+	.dma_error = mmci_dmae_error,
+};
+#else
+static struct mmci_host_ops mmci_variant_ops = {
+	.get_datactrl_cfg = mmci_get_dctrl_cfg,
+};
+#endif
+
+static void mmci_variant_init(struct mmci_host *host)
+{
+	host->ops = &mmci_variant_ops;
+}
+
+static void ux500_variant_init(struct mmci_host *host)
+{
+	host->ops = &mmci_variant_ops;
+	host->ops->busy_complete = ux500_busy_complete;
+}
+
+static void ux500v2_variant_init(struct mmci_host *host)
+{
+	host->ops = &mmci_variant_ops;
+	host->ops->busy_complete = ux500_busy_complete;
+	host->ops->get_datactrl_cfg = ux500v2_get_dctrl_cfg;
+}
+
+static void mmci_pre_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct mmci_host *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+
+	if (!data)
+		return;
+
+	WARN_ON(data->host_cookie);
+
+	if (mmci_validate_data(host, data))
+		return;
+
+	mmci_prep_data(host, data, true);
+}
+
+static void mmci_post_request(struct mmc_host *mmc, struct mmc_request *mrq,
+			      int err)
+{
+	struct mmci_host *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+
+	if (!data || !data->host_cookie)
+		return;
+
+	mmci_unprep_data(host, data, err);
+}
+
+static void mmci_start_data(struct mmci_host *host, struct mmc_data *data)
+{
+	struct variant_data *variant = host->variant;
+	unsigned int datactrl, timeout, irqmask;
+	unsigned long long clks;
+	void __iomem *base;
+
+	dev_dbg(mmc_dev(host->mmc), "blksz %04x blks %04x flags %08x\n",
+		data->blksz, data->blocks, data->flags);
+
+	host->data = data;
+	host->size = data->blksz * data->blocks;
+	data->bytes_xfered = 0;
+
+	clks = (unsigned long long)data->timeout_ns * host->cclk;
+	do_div(clks, NSEC_PER_SEC);
+
+	timeout = data->timeout_clks + (unsigned int)clks;
+
+	base = host->base;
+	writel(timeout, base + MMCIDATATIMER);
+	writel(host->size, base + MMCIDATALENGTH);
+
+	datactrl = host->ops->get_datactrl_cfg(host);
+	datactrl |= host->data->flags & MMC_DATA_READ ? MCI_DPSM_DIRECTION : 0;
+
+	if (host->mmc->card && mmc_card_sdio(host->mmc->card)) {
+		u32 clk;
+
+		datactrl |= variant->datactrl_mask_sdio;
+
+		/*
+		 * The ST Micro variant for SDIO small write transfers
+		 * needs to have clock H/W flow control disabled,
+		 * otherwise the transfer will not start. The threshold
+		 * depends on the rate of MCLK.
+		 */
+		if (variant->st_sdio && data->flags & MMC_DATA_WRITE &&
+		    (host->size < 8 ||
+		     (host->size <= 8 && host->mclk > 50000000)))
+			clk = host->clk_reg & ~variant->clkreg_enable;
+		else
+			clk = host->clk_reg | variant->clkreg_enable;
+
+		mmci_write_clkreg(host, clk);
+	}
+
+	if (host->mmc->ios.timing == MMC_TIMING_UHS_DDR50 ||
+	    host->mmc->ios.timing == MMC_TIMING_MMC_DDR52)
+		datactrl |= variant->datactrl_mask_ddrmode;
+
+	/*
+	 * Attempt to use DMA operation mode, if this
+	 * should fail, fall back to PIO mode
+	 */
+	if (!mmci_dma_start(host, datactrl))
+		return;
+
+	/* IRQ mode, map the SG list for CPU reading/writing */
+	mmci_init_sg(host, data);
+
+	if (data->flags & MMC_DATA_READ) {
+		irqmask = MCI_RXFIFOHALFFULLMASK;
+
+		/*
+		 * If we have less than the fifo 'half-full' threshold to
+		 * transfer, trigger a PIO interrupt as soon as any data
+		 * is available.
+		 */
+		if (host->size < variant->fifohalfsize)
+			irqmask |= MCI_RXDATAAVLBLMASK;
+	} else {
+		/*
+		 * We don't actually need to include "FIFO empty" here
+		 * since its implicit in "FIFO half empty".
+		 */
+		irqmask = MCI_TXFIFOHALFEMPTYMASK;
+	}
+
+	mmci_write_datactrlreg(host, datactrl);
+	writel(readl(base + MMCIMASK0) & ~MCI_DATAENDMASK, base + MMCIMASK0);
+	mmci_set_mask1(host, irqmask);
+}
+
+static void
+mmci_start_command(struct mmci_host *host, struct mmc_command *cmd, u32 c)
+{
+	void __iomem *base = host->base;
+	unsigned long long clks;
+
+	dev_dbg(mmc_dev(host->mmc), "op %02x arg %08x flags %08x\n",
+	    cmd->opcode, cmd->arg, cmd->flags);
+
+	if (readl(base + MMCICOMMAND) & host->variant->cmdreg_cpsm_enable) {
+		writel(0, base + MMCICOMMAND);
+		mmci_reg_delay(host);
+	}
+
+	if (host->variant->cmdreg_stop &&
+	    cmd->opcode == MMC_STOP_TRANSMISSION)
+		c |= host->variant->cmdreg_stop;
+
+	c |= cmd->opcode | host->variant->cmdreg_cpsm_enable;
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136)
+			c |= host->variant->cmdreg_lrsp_crc;
+		else if (cmd->flags & MMC_RSP_CRC)
+			c |= host->variant->cmdreg_srsp_crc;
+		else
+			c |= host->variant->cmdreg_srsp;
+	}
+
+	if (host->variant->busy_timeout && cmd->flags & MMC_RSP_BUSY) {
+		if (!cmd->busy_timeout)
+			cmd->busy_timeout = 10 * MSEC_PER_SEC;
+
+		if (cmd->busy_timeout > host->mmc->max_busy_timeout)
+			clks = (unsigned long long)host->mmc->max_busy_timeout * host->cclk;
+		else
+			clks = (unsigned long long)cmd->busy_timeout * host->cclk;
+
+		do_div(clks, MSEC_PER_SEC);
+		writel_relaxed(clks, host->base + MMCIDATATIMER);
+	}
+
+	if (host->ops->pre_sig_volt_switch && cmd->opcode == SD_SWITCH_VOLTAGE)
+		host->ops->pre_sig_volt_switch(host);
+
+	if (/*interrupt*/0)
+		c |= MCI_CPSM_INTERRUPT;
+
+	if (mmc_cmd_type(cmd) == MMC_CMD_ADTC)
+		c |= host->variant->data_cmd_enable;
+
+	host->cmd = cmd;
+
+	writel(cmd->arg, base + MMCIARGUMENT);
+	writel(c, base + MMCICOMMAND);
+}
+
+static void mmci_stop_command(struct mmci_host *host)
+{
+	host->stop_abort.error = 0;
+	mmci_start_command(host, &host->stop_abort, 0);
+}
+
+static void
+mmci_data_irq(struct mmci_host *host, struct mmc_data *data,
+	      unsigned int status)
+{
+	unsigned int status_err;
+
+	/* Make sure we have data to handle */
+	if (!data)
+		return;
+
+	/* First check for errors */
+	status_err = status & (host->variant->start_err |
+			       MCI_DATACRCFAIL | MCI_DATATIMEOUT |
+			       MCI_TXUNDERRUN | MCI_RXOVERRUN);
+
+	if (status_err) {
+		u32 remain, success;
+
+		/* Terminate the DMA transfer */
+		mmci_dma_error(host);
+
+		/*
+		 * Calculate how far we are into the transfer.  Note that
+		 * the data counter gives the number of bytes transferred
+		 * on the MMC bus, not on the host side.  On reads, this
+		 * can be as much as a FIFO-worth of data ahead.  This
+		 * matters for FIFO overruns only.
+		 */
+		if (!host->variant->datacnt_useless) {
+			remain = readl(host->base + MMCIDATACNT);
+			success = data->blksz * data->blocks - remain;
+		} else {
+			success = 0;
+		}
+
+		dev_dbg(mmc_dev(host->mmc), "MCI ERROR IRQ, status 0x%08x at 0x%08x\n",
+			status_err, success);
+		if (status_err & MCI_DATACRCFAIL) {
+			/* Last block was not successful */
+			success -= 1;
+			data->error = -EILSEQ;
+		} else if (status_err & MCI_DATATIMEOUT) {
+			data->error = -ETIMEDOUT;
+		} else if (status_err & MCI_STARTBITERR) {
+			data->error = -ECOMM;
+		} else if (status_err & MCI_TXUNDERRUN) {
+			data->error = -EIO;
+		} else if (status_err & MCI_RXOVERRUN) {
+			if (success > host->variant->fifosize)
+				success -= host->variant->fifosize;
+			else
+				success = 0;
+			data->error = -EIO;
+		}
+		data->bytes_xfered = round_down(success, data->blksz);
+	}
+
+	if (status & MCI_DATABLOCKEND)
+		dev_err(mmc_dev(host->mmc), "stray MCI_DATABLOCKEND interrupt\n");
+
+	if (status & MCI_DATAEND || data->error) {
+		mmci_dma_finalize(host, data);
+
+		mmci_stop_data(host);
+
+		if (!data->error)
+			/* The error clause is handled above, success! */
+			data->bytes_xfered = data->blksz * data->blocks;
+
+		if (!data->stop) {
+			if (host->variant->cmdreg_stop && data->error)
+				mmci_stop_command(host);
+			else
+				mmci_request_end(host, data->mrq);
+		} else if (host->mrq->sbc && !data->error) {
+			mmci_request_end(host, data->mrq);
+		} else {
+			mmci_start_command(host, data->stop, 0);
+		}
+	}
+}
+
+static void
+mmci_cmd_irq(struct mmci_host *host, struct mmc_command *cmd,
+	     unsigned int status)
+{
+	u32 err_msk = MCI_CMDCRCFAIL | MCI_CMDTIMEOUT;
+	void __iomem *base = host->base;
+	bool sbc, busy_resp;
+
+	if (!cmd)
+		return;
+
+	sbc = (cmd == host->mrq->sbc);
+	busy_resp = !!(cmd->flags & MMC_RSP_BUSY);
+
+	/*
+	 * We need to be one of these interrupts to be considered worth
+	 * handling. Note that we tag on any latent IRQs postponed
+	 * due to waiting for busy status.
+	 */
+	if (host->variant->busy_timeout && busy_resp)
+		err_msk |= MCI_DATATIMEOUT;
+
+	if (!((status | host->busy_status) &
+	      (err_msk | MCI_CMDSENT | MCI_CMDRESPEND)))
+		return;
+
+	/* Handle busy detection on DAT0 if the variant supports it. */
+	if (busy_resp && host->variant->busy_detect)
+		if (!host->ops->busy_complete(host, status, err_msk))
+			return;
+
+	host->cmd = NULL;
+
+	if (status & MCI_CMDTIMEOUT) {
+		cmd->error = -ETIMEDOUT;
+	} else if (status & MCI_CMDCRCFAIL && cmd->flags & MMC_RSP_CRC) {
+		cmd->error = -EILSEQ;
+	} else if (host->variant->busy_timeout && busy_resp &&
+		   status & MCI_DATATIMEOUT) {
+		cmd->error = -ETIMEDOUT;
+		host->irq_action = IRQ_WAKE_THREAD;
+	} else {
+		cmd->resp[0] = readl(base + MMCIRESPONSE0);
+		cmd->resp[1] = readl(base + MMCIRESPONSE1);
+		cmd->resp[2] = readl(base + MMCIRESPONSE2);
+		cmd->resp[3] = readl(base + MMCIRESPONSE3);
+	}
+
+	if ((!sbc && !cmd->data) || cmd->error) {
+		if (host->data) {
+			/* Terminate the DMA transfer */
+			mmci_dma_error(host);
+
+			mmci_stop_data(host);
+			if (host->variant->cmdreg_stop && cmd->error) {
+				mmci_stop_command(host);
+				return;
+			}
+		}
+
+		if (host->irq_action != IRQ_WAKE_THREAD)
+			mmci_request_end(host, host->mrq);
+
+	} else if (sbc) {
+		mmci_start_command(host, host->mrq->cmd, 0);
+	} else if (!host->variant->datactrl_first &&
+		   !(cmd->data->flags & MMC_DATA_READ)) {
+		mmci_start_data(host, cmd->data);
+	}
+}
+
+static int mmci_get_rx_fifocnt(struct mmci_host *host, u32 status, int remain)
+{
+	return remain - (readl(host->base + MMCIFIFOCNT) << 2);
+}
+
+static int mmci_qcom_get_rx_fifocnt(struct mmci_host *host, u32 status, int r)
+{
+	/*
+	 * on qcom SDCC4 only 8 words are used in each burst so only 8 addresses
+	 * from the fifo range should be used
+	 */
+	if (status & MCI_RXFIFOHALFFULL)
+		return host->variant->fifohalfsize;
+	else if (status & MCI_RXDATAAVLBL)
+		return 4;
+
+	return 0;
+}
+
+static int mmci_pio_read(struct mmci_host *host, char *buffer, unsigned int remain)
+{
+	void __iomem *base = host->base;
+	char *ptr = buffer;
+	u32 status = readl(host->base + MMCISTATUS);
+	int host_remain = host->size;
+
+	do {
+		int count = host->get_rx_fifocnt(host, status, host_remain);
+
+		if (count > remain)
+			count = remain;
+
+		if (count <= 0)
+			break;
+
+		/*
+		 * SDIO especially may want to send something that is
+		 * not divisible by 4 (as opposed to card sectors
+		 * etc). Therefore make sure to always read the last bytes
+		 * while only doing full 32-bit reads towards the FIFO.
+		 */
+		if (unlikely(count & 0x3)) {
+			if (count < 4) {
+				unsigned char buf[4];
+				ioread32_rep(base + MMCIFIFO, buf, 1);
+				memcpy(ptr, buf, count);
+			} else {
+				ioread32_rep(base + MMCIFIFO, ptr, count >> 2);
+				count &= ~0x3;
+			}
+		} else {
+			ioread32_rep(base + MMCIFIFO, ptr, count >> 2);
+		}
+
+		ptr += count;
+		remain -= count;
+		host_remain -= count;
+
+		if (remain == 0)
+			break;
+
+		status = readl(base + MMCISTATUS);
+	} while (status & MCI_RXDATAAVLBL);
+
+	return ptr - buffer;
+}
+
+static int mmci_pio_write(struct mmci_host *host, char *buffer, unsigned int remain, u32 status)
+{
+	struct variant_data *variant = host->variant;
+	void __iomem *base = host->base;
+	char *ptr = buffer;
+
+	do {
+		unsigned int count, maxcnt;
+
+		maxcnt = status & MCI_TXFIFOEMPTY ?
+			 variant->fifosize : variant->fifohalfsize;
+		count = min(remain, maxcnt);
+
+		/*
+		 * SDIO especially may want to send something that is
+		 * not divisible by 4 (as opposed to card sectors
+		 * etc), and the FIFO only accept full 32-bit writes.
+		 * So compensate by adding +3 on the count, a single
+		 * byte become a 32bit write, 7 bytes will be two
+		 * 32bit writes etc.
+		 */
+		iowrite32_rep(base + MMCIFIFO, ptr, (count + 3) >> 2);
+
+		ptr += count;
+		remain -= count;
+
+		if (remain == 0)
+			break;
+
+		status = readl(base + MMCISTATUS);
+	} while (status & MCI_TXFIFOHALFEMPTY);
+
+	return ptr - buffer;
+}
+
+/*
+ * PIO data transfer IRQ handler.
+ */
+static irqreturn_t mmci_pio_irq(int irq, void *dev_id)
+{
+	struct mmci_host *host = dev_id;
+	struct sg_mapping_iter *sg_miter = &host->sg_miter;
+	struct variant_data *variant = host->variant;
+	void __iomem *base = host->base;
+	u32 status;
+
+	status = readl(base + MMCISTATUS);
+
+	dev_dbg(mmc_dev(host->mmc), "irq1 (pio) %08x\n", status);
+
+	do {
+		unsigned int remain, len;
+		char *buffer;
+
+		/*
+		 * For write, we only need to test the half-empty flag
+		 * here - if the FIFO is completely empty, then by
+		 * definition it is more than half empty.
+		 *
+		 * For read, check for data available.
+		 */
+		if (!(status & (MCI_TXFIFOHALFEMPTY|MCI_RXDATAAVLBL)))
+			break;
+
+		if (!sg_miter_next(sg_miter))
+			break;
+
+		buffer = sg_miter->addr;
+		remain = sg_miter->length;
+
+		len = 0;
+		if (status & MCI_RXACTIVE)
+			len = mmci_pio_read(host, buffer, remain);
+		if (status & MCI_TXACTIVE)
+			len = mmci_pio_write(host, buffer, remain, status);
+
+		sg_miter->consumed = len;
+
+		host->size -= len;
+		remain -= len;
+
+		if (remain)
+			break;
+
+		status = readl(base + MMCISTATUS);
+	} while (1);
+
+	sg_miter_stop(sg_miter);
+
+	/*
+	 * If we have less than the fifo 'half-full' threshold to transfer,
+	 * trigger a PIO interrupt as soon as any data is available.
+	 */
+	if (status & MCI_RXACTIVE && host->size < variant->fifohalfsize)
+		mmci_set_mask1(host, MCI_RXDATAAVLBLMASK);
+
+	/*
+	 * If we run out of data, disable the data IRQs; this
+	 * prevents a race where the FIFO becomes empty before
+	 * the chip itself has disabled the data path, and
+	 * stops us racing with our data end IRQ.
+	 */
+	if (host->size == 0) {
+		mmci_set_mask1(host, 0);
+		writel(readl(base + MMCIMASK0) | MCI_DATAENDMASK, base + MMCIMASK0);
+	}
+
+	return IRQ_HANDLED;
+}
+
+/*
+ * Handle completion of command and data transfers.
+ */
+static irqreturn_t mmci_irq(int irq, void *dev_id)
+{
+	struct mmci_host *host = dev_id;
+	u32 status;
+
+	spin_lock(&host->lock);
+	host->irq_action = IRQ_HANDLED;
+
+	do {
+		status = readl(host->base + MMCISTATUS);
+
+		if (host->singleirq) {
+			if (status & host->mask1_reg)
+				mmci_pio_irq(irq, dev_id);
+
+			status &= ~host->variant->irq_pio_mask;
+		}
+
+		/*
+		 * Busy detection is managed by mmci_cmd_irq(), including to
+		 * clear the corresponding IRQ.
+		 */
+		status &= readl(host->base + MMCIMASK0);
+		if (host->variant->busy_detect)
+			writel(status & ~host->variant->busy_detect_mask,
+			       host->base + MMCICLEAR);
+		else
+			writel(status, host->base + MMCICLEAR);
+
+		dev_dbg(mmc_dev(host->mmc), "irq0 (data+cmd) %08x\n", status);
+
+		if (host->variant->reversed_irq_handling) {
+			mmci_data_irq(host, host->data, status);
+			mmci_cmd_irq(host, host->cmd, status);
+		} else {
+			mmci_cmd_irq(host, host->cmd, status);
+			mmci_data_irq(host, host->data, status);
+		}
+
+		/*
+		 * Busy detection has been handled by mmci_cmd_irq() above.
+		 * Clear the status bit to prevent polling in IRQ context.
+		 */
+		if (host->variant->busy_detect_flag)
+			status &= ~host->variant->busy_detect_flag;
+
+	} while (status);
+
+	spin_unlock(&host->lock);
+
+	return host->irq_action;
+}
+
+/*
+ * mmci_irq_thread() - A threaded IRQ handler that manages a reset of the HW.
+ *
+ * A reset is needed for some variants, where a datatimeout for a R1B request
+ * causes the DPSM to stay busy (non-functional).
+ */
+static irqreturn_t mmci_irq_thread(int irq, void *dev_id)
+{
+	struct mmci_host *host = dev_id;
+	unsigned long flags;
+
+	if (host->rst) {
+		reset_control_assert(host->rst);
+		udelay(2);
+		reset_control_deassert(host->rst);
+	}
+
+	spin_lock_irqsave(&host->lock, flags);
+	writel(host->clk_reg, host->base + MMCICLOCK);
+	writel(host->pwr_reg, host->base + MMCIPOWER);
+	writel(MCI_IRQENABLE | host->variant->start_err,
+	       host->base + MMCIMASK0);
+
+	host->irq_action = IRQ_HANDLED;
+	mmci_request_end(host, host->mrq);
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	return host->irq_action;
+}
+
+static void mmci_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct mmci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+
+	WARN_ON(host->mrq != NULL);
+
+	mrq->cmd->error = mmci_validate_data(host, mrq->data);
+	if (mrq->cmd->error) {
+		mmc_request_done(mmc, mrq);
+		return;
+	}
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	host->mrq = mrq;
+
+	if (mrq->data)
+		mmci_get_next_data(host, mrq->data);
+
+	if (mrq->data &&
+	    (host->variant->datactrl_first || mrq->data->flags & MMC_DATA_READ))
+		mmci_start_data(host, mrq->data);
+
+	if (mrq->sbc)
+		mmci_start_command(host, mrq->sbc, 0);
+	else
+		mmci_start_command(host, mrq->cmd, 0);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void mmci_set_max_busy_timeout(struct mmc_host *mmc)
+{
+	struct mmci_host *host = mmc_priv(mmc);
+	u32 max_busy_timeout = 0;
+
+	if (!host->variant->busy_detect)
+		return;
+
+	if (host->variant->busy_timeout && mmc->actual_clock)
+		max_busy_timeout = U32_MAX / DIV_ROUND_UP(mmc->actual_clock,
+							  MSEC_PER_SEC);
+
+	mmc->max_busy_timeout = max_busy_timeout;
+}
+
+static void mmci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct mmci_host *host = mmc_priv(mmc);
+	struct variant_data *variant = host->variant;
+	u32 pwr = 0;
+	unsigned long flags;
+	int ret;
+
+	if (host->plat->ios_handler &&
+		host->plat->ios_handler(mmc_dev(mmc), ios))
+			dev_err(mmc_dev(mmc), "platform ios_handler failed\n");
+
+	switch (ios->power_mode) {
+	case MMC_POWER_OFF:
+		if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+
+		if (!IS_ERR(mmc->supply.vqmmc) && host->vqmmc_enabled) {
+			regulator_disable(mmc->supply.vqmmc);
+			host->vqmmc_enabled = false;
+		}
+
+		break;
+	case MMC_POWER_UP:
+		if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, ios->vdd);
+
+		/*
+		 * The ST Micro variant doesn't have the PL180s MCI_PWR_UP
+		 * and instead uses MCI_PWR_ON so apply whatever value is
+		 * configured in the variant data.
+		 */
+		pwr |= variant->pwrreg_powerup;
+
+		break;
+	case MMC_POWER_ON:
+		if (!IS_ERR(mmc->supply.vqmmc) && !host->vqmmc_enabled) {
+			ret = regulator_enable(mmc->supply.vqmmc);
+			if (ret < 0)
+				dev_err(mmc_dev(mmc),
+					"failed to enable vqmmc regulator\n");
+			else
+				host->vqmmc_enabled = true;
+		}
+
+		pwr |= MCI_PWR_ON;
+		break;
+	}
+
+	if (variant->signal_direction && ios->power_mode != MMC_POWER_OFF) {
+		/*
+		 * The ST Micro variant has some additional bits
+		 * indicating signal direction for the signals in
+		 * the SD/MMC bus and feedback-clock usage.
+		 */
+		pwr |= host->pwr_reg_add;
+
+		if (ios->bus_width == MMC_BUS_WIDTH_4)
+			pwr &= ~MCI_ST_DATA74DIREN;
+		else if (ios->bus_width == MMC_BUS_WIDTH_1)
+			pwr &= (~MCI_ST_DATA74DIREN &
+				~MCI_ST_DATA31DIREN &
+				~MCI_ST_DATA2DIREN);
+	}
+
+	if (variant->opendrain) {
+		if (ios->bus_mode == MMC_BUSMODE_OPENDRAIN)
+			pwr |= variant->opendrain;
+	} else {
+		/*
+		 * If the variant cannot configure the pads by its own, then we
+		 * expect the pinctrl to be able to do that for us
+		 */
+		if (ios->bus_mode == MMC_BUSMODE_OPENDRAIN)
+			pinctrl_select_state(host->pinctrl, host->pins_opendrain);
+		else
+			pinctrl_select_default_state(mmc_dev(mmc));
+	}
+
+	/*
+	 * If clock = 0 and the variant requires the MMCIPOWER to be used for
+	 * gating the clock, the MCI_PWR_ON bit is cleared.
+	 */
+	if (!ios->clock && variant->pwrreg_clkgate)
+		pwr &= ~MCI_PWR_ON;
+
+	if (host->variant->explicit_mclk_control &&
+	    ios->clock != host->clock_cache) {
+		ret = clk_set_rate(host->clk, ios->clock);
+		if (ret < 0)
+			dev_err(mmc_dev(host->mmc),
+				"Error setting clock rate (%d)\n", ret);
+		else
+			host->mclk = clk_get_rate(host->clk);
+	}
+	host->clock_cache = ios->clock;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (host->ops && host->ops->set_clkreg)
+		host->ops->set_clkreg(host, ios->clock);
+	else
+		mmci_set_clkreg(host, ios->clock);
+
+	mmci_set_max_busy_timeout(mmc);
+
+	if (host->ops && host->ops->set_pwrreg)
+		host->ops->set_pwrreg(host, pwr);
+	else
+		mmci_write_pwrreg(host, pwr);
+
+	mmci_reg_delay(host);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static int mmci_get_cd(struct mmc_host *mmc)
+{
+	struct mmci_host *host = mmc_priv(mmc);
+	struct mmci_platform_data *plat = host->plat;
+	unsigned int status = mmc_gpio_get_cd(mmc);
+
+	if (status == -ENOSYS) {
+		if (!plat->status)
+			return 1; /* Assume always present */
+
+		status = plat->status(mmc_dev(host->mmc));
+	}
+	return status;
+}
+
+static int mmci_sig_volt_switch(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct mmci_host *host = mmc_priv(mmc);
+	int ret;
+
+	ret = mmc_regulator_set_vqmmc(mmc, ios);
+
+	if (!ret && host->ops && host->ops->post_sig_volt_switch)
+		ret = host->ops->post_sig_volt_switch(host, ios);
+	else if (ret)
+		ret = 0;
+
+	if (ret < 0)
+		dev_warn(mmc_dev(mmc), "Voltage switch failed\n");
+
+	return ret;
+}
+
+static struct mmc_host_ops mmci_ops = {
+	.request	= mmci_request,
+	.pre_req	= mmci_pre_request,
+	.post_req	= mmci_post_request,
+	.set_ios	= mmci_set_ios,
+	.get_ro		= mmc_gpio_get_ro,
+	.get_cd		= mmci_get_cd,
+	.start_signal_voltage_switch = mmci_sig_volt_switch,
+};
+
+static int mmci_of_parse(struct device_node *np, struct mmc_host *mmc)
+{
+	struct mmci_host *host = mmc_priv(mmc);
+	int ret = mmc_of_parse(mmc);
+
+	if (ret)
+		return ret;
+
+	if (of_get_property(np, "st,sig-dir-dat0", NULL))
+		host->pwr_reg_add |= MCI_ST_DATA0DIREN;
+	if (of_get_property(np, "st,sig-dir-dat2", NULL))
+		host->pwr_reg_add |= MCI_ST_DATA2DIREN;
+	if (of_get_property(np, "st,sig-dir-dat31", NULL))
+		host->pwr_reg_add |= MCI_ST_DATA31DIREN;
+	if (of_get_property(np, "st,sig-dir-dat74", NULL))
+		host->pwr_reg_add |= MCI_ST_DATA74DIREN;
+	if (of_get_property(np, "st,sig-dir-cmd", NULL))
+		host->pwr_reg_add |= MCI_ST_CMDDIREN;
+	if (of_get_property(np, "st,sig-pin-fbclk", NULL))
+		host->pwr_reg_add |= MCI_ST_FBCLKEN;
+	if (of_get_property(np, "st,sig-dir", NULL))
+		host->pwr_reg_add |= MCI_STM32_DIRPOL;
+	if (of_get_property(np, "st,neg-edge", NULL))
+		host->clk_reg_add |= MCI_STM32_CLK_NEGEDGE;
+	if (of_get_property(np, "st,use-ckin", NULL))
+		host->clk_reg_add |= MCI_STM32_CLK_SELCKIN;
+
+	if (of_get_property(np, "mmc-cap-mmc-highspeed", NULL))
+		mmc->caps |= MMC_CAP_MMC_HIGHSPEED;
+	if (of_get_property(np, "mmc-cap-sd-highspeed", NULL))
+		mmc->caps |= MMC_CAP_SD_HIGHSPEED;
+
+	return 0;
+}
+
+static int mmci_probe(struct amba_device *dev,
+	const struct amba_id *id)
+{
+	struct mmci_platform_data *plat = dev->dev.platform_data;
+	struct device_node *np = dev->dev.of_node;
+	struct variant_data *variant = id->data;
+	struct mmci_host *host;
+	struct mmc_host *mmc;
+	int ret;
+
+	/* Must have platform data or Device Tree. */
+	if (!plat && !np) {
+		dev_err(&dev->dev, "No plat data or DT found\n");
+		return -EINVAL;
+	}
+
+	if (!plat) {
+		plat = devm_kzalloc(&dev->dev, sizeof(*plat), GFP_KERNEL);
+		if (!plat)
+			return -ENOMEM;
+	}
+
+	mmc = mmc_alloc_host(sizeof(struct mmci_host), &dev->dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	ret = mmci_of_parse(np, mmc);
+	if (ret)
+		goto host_free;
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+	host->mmc_ops = &mmci_ops;
+	mmc->ops = &mmci_ops;
+
+	/*
+	 * Some variant (STM32) doesn't have opendrain bit, nevertheless
+	 * pins can be set accordingly using pinctrl
+	 */
+	if (!variant->opendrain) {
+		host->pinctrl = devm_pinctrl_get(&dev->dev);
+		if (IS_ERR(host->pinctrl)) {
+			dev_err(&dev->dev, "failed to get pinctrl");
+			ret = PTR_ERR(host->pinctrl);
+			goto host_free;
+		}
+
+		host->pins_opendrain = pinctrl_lookup_state(host->pinctrl,
+							    MMCI_PINCTRL_STATE_OPENDRAIN);
+		if (IS_ERR(host->pins_opendrain)) {
+			dev_err(mmc_dev(mmc), "Can't select opendrain pins\n");
+			ret = PTR_ERR(host->pins_opendrain);
+			goto host_free;
+		}
+	}
+
+	host->hw_designer = amba_manf(dev);
+	host->hw_revision = amba_rev(dev);
+	dev_dbg(mmc_dev(mmc), "designer ID = 0x%02x\n", host->hw_designer);
+	dev_dbg(mmc_dev(mmc), "revision = 0x%01x\n", host->hw_revision);
+
+	host->clk = devm_clk_get(&dev->dev, NULL);
+	if (IS_ERR(host->clk)) {
+		ret = PTR_ERR(host->clk);
+		goto host_free;
+	}
+
+	ret = clk_prepare_enable(host->clk);
+	if (ret)
+		goto host_free;
+
+	if (variant->qcom_fifo)
+		host->get_rx_fifocnt = mmci_qcom_get_rx_fifocnt;
+	else
+		host->get_rx_fifocnt = mmci_get_rx_fifocnt;
+
+	host->plat = plat;
+	host->variant = variant;
+	host->mclk = clk_get_rate(host->clk);
+	/*
+	 * According to the spec, mclk is max 100 MHz,
+	 * so we try to adjust the clock down to this,
+	 * (if possible).
+	 */
+	if (host->mclk > variant->f_max) {
+		ret = clk_set_rate(host->clk, variant->f_max);
+		if (ret < 0)
+			goto clk_disable;
+		host->mclk = clk_get_rate(host->clk);
+		dev_dbg(mmc_dev(mmc), "eventual mclk rate: %u Hz\n",
+			host->mclk);
+	}
+
+	host->phybase = dev->res.start;
+	host->base = devm_ioremap_resource(&dev->dev, &dev->res);
+	if (IS_ERR(host->base)) {
+		ret = PTR_ERR(host->base);
+		goto clk_disable;
+	}
+
+	if (variant->init)
+		variant->init(host);
+
+	/*
+	 * The ARM and ST versions of the block have slightly different
+	 * clock divider equations which means that the minimum divider
+	 * differs too.
+	 * on Qualcomm like controllers get the nearest minimum clock to 100Khz
+	 */
+	if (variant->st_clkdiv)
+		mmc->f_min = DIV_ROUND_UP(host->mclk, 257);
+	else if (variant->stm32_clkdiv)
+		mmc->f_min = DIV_ROUND_UP(host->mclk, 2046);
+	else if (variant->explicit_mclk_control)
+		mmc->f_min = clk_round_rate(host->clk, 100000);
+	else
+		mmc->f_min = DIV_ROUND_UP(host->mclk, 512);
+	/*
+	 * If no maximum operating frequency is supplied, fall back to use
+	 * the module parameter, which has a (low) default value in case it
+	 * is not specified. Either value must not exceed the clock rate into
+	 * the block, of course.
+	 */
+	if (mmc->f_max)
+		mmc->f_max = variant->explicit_mclk_control ?
+				min(variant->f_max, mmc->f_max) :
+				min(host->mclk, mmc->f_max);
+	else
+		mmc->f_max = variant->explicit_mclk_control ?
+				fmax : min(host->mclk, fmax);
+
+
+	dev_dbg(mmc_dev(mmc), "clocking block at %u Hz\n", mmc->f_max);
+
+	host->rst = devm_reset_control_get_optional_exclusive(&dev->dev, NULL);
+	if (IS_ERR(host->rst)) {
+		ret = PTR_ERR(host->rst);
+		goto clk_disable;
+	}
+
+	/* Get regulators and the supported OCR mask */
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret)
+		goto clk_disable;
+
+	if (!mmc->ocr_avail)
+		mmc->ocr_avail = plat->ocr_mask;
+	else if (plat->ocr_mask)
+		dev_warn(mmc_dev(mmc), "Platform OCR mask is ignored\n");
+
+	/* We support these capabilities. */
+	mmc->caps |= MMC_CAP_CMD23;
+
+	/*
+	 * Enable busy detection.
+	 */
+	if (variant->busy_detect) {
+		mmci_ops.card_busy = mmci_card_busy;
+		/*
+		 * Not all variants have a flag to enable busy detection
+		 * in the DPSM, but if they do, set it here.
+		 */
+		if (variant->busy_dpsm_flag)
+			mmci_write_datactrlreg(host,
+					       host->variant->busy_dpsm_flag);
+		mmc->caps |= MMC_CAP_WAIT_WHILE_BUSY;
+	}
+
+	/* Variants with mandatory busy timeout in HW needs R1B responses. */
+	if (variant->busy_timeout)
+		mmc->caps |= MMC_CAP_NEED_RSP_BUSY;
+
+	/* Prepare a CMD12 - needed to clear the DPSM on some variants. */
+	host->stop_abort.opcode = MMC_STOP_TRANSMISSION;
+	host->stop_abort.arg = 0;
+	host->stop_abort.flags = MMC_RSP_R1B | MMC_CMD_AC;
+
+	/* We support these PM capabilities. */
+	mmc->pm_caps |= MMC_PM_KEEP_POWER;
+
+	/*
+	 * We can do SGIO
+	 */
+	mmc->max_segs = NR_SG;
+
+	/*
+	 * Since only a certain number of bits are valid in the data length
+	 * register, we must ensure that we don't exceed 2^num-1 bytes in a
+	 * single request.
+	 */
+	mmc->max_req_size = (1 << variant->datalength_bits) - 1;
+
+	/*
+	 * Set the maximum segment size.  Since we aren't doing DMA
+	 * (yet) we are only limited by the data length register.
+	 */
+	mmc->max_seg_size = mmc->max_req_size;
+
+	/*
+	 * Block size can be up to 2048 bytes, but must be a power of two.
+	 */
+	mmc->max_blk_size = 1 << variant->datactrl_blocksz;
+
+	/*
+	 * Limit the number of blocks transferred so that we don't overflow
+	 * the maximum request size.
+	 */
+	mmc->max_blk_count = mmc->max_req_size >> variant->datactrl_blocksz;
+
+	spin_lock_init(&host->lock);
+
+	writel(0, host->base + MMCIMASK0);
+
+	if (variant->mmcimask1)
+		writel(0, host->base + MMCIMASK1);
+
+	writel(0xfff, host->base + MMCICLEAR);
+
+	/*
+	 * If:
+	 * - not using DT but using a descriptor table, or
+	 * - using a table of descriptors ALONGSIDE DT, or
+	 * look up these descriptors named "cd" and "wp" right here, fail
+	 * silently of these do not exist
+	 */
+	if (!np) {
+		ret = mmc_gpiod_request_cd(mmc, "cd", 0, false, 0);
+		if (ret == -EPROBE_DEFER)
+			goto clk_disable;
+
+		ret = mmc_gpiod_request_ro(mmc, "wp", 0, 0);
+		if (ret == -EPROBE_DEFER)
+			goto clk_disable;
+	}
+
+	ret = devm_request_threaded_irq(&dev->dev, dev->irq[0], mmci_irq,
+					mmci_irq_thread, IRQF_SHARED,
+					DRIVER_NAME " (cmd)", host);
+	if (ret)
+		goto clk_disable;
+
+	if (!dev->irq[1])
+		host->singleirq = true;
+	else {
+		ret = devm_request_irq(&dev->dev, dev->irq[1], mmci_pio_irq,
+				IRQF_SHARED, DRIVER_NAME " (pio)", host);
+		if (ret)
+			goto clk_disable;
+	}
+
+	writel(MCI_IRQENABLE | variant->start_err, host->base + MMCIMASK0);
+
+	amba_set_drvdata(dev, mmc);
+
+	dev_info(&dev->dev, "%s: PL%03x manf %x rev%u at 0x%08llx irq %d,%d (pio)\n",
+		 mmc_hostname(mmc), amba_part(dev), amba_manf(dev),
+		 amba_rev(dev), (unsigned long long)dev->res.start,
+		 dev->irq[0], dev->irq[1]);
+
+	mmci_dma_setup(host);
+
+	pm_runtime_set_autosuspend_delay(&dev->dev, 50);
+	pm_runtime_use_autosuspend(&dev->dev);
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto clk_disable;
+
+	pm_runtime_put(&dev->dev);
+	return 0;
+
+ clk_disable:
+	clk_disable_unprepare(host->clk);
+ host_free:
+	mmc_free_host(mmc);
+	return ret;
+}
+
+static void mmci_remove(struct amba_device *dev)
+{
+	struct mmc_host *mmc = amba_get_drvdata(dev);
+
+	if (mmc) {
+		struct mmci_host *host = mmc_priv(mmc);
+		struct variant_data *variant = host->variant;
+
+		/*
+		 * Undo pm_runtime_put() in probe.  We use the _sync
+		 * version here so that we can access the primecell.
+		 */
+		pm_runtime_get_sync(&dev->dev);
+
+		mmc_remove_host(mmc);
+
+		writel(0, host->base + MMCIMASK0);
+
+		if (variant->mmcimask1)
+			writel(0, host->base + MMCIMASK1);
+
+		writel(0, host->base + MMCICOMMAND);
+		writel(0, host->base + MMCIDATACTRL);
+
+		mmci_dma_release(host);
+		clk_disable_unprepare(host->clk);
+		mmc_free_host(mmc);
+	}
+}
+
+#ifdef CONFIG_PM
+static void mmci_save(struct mmci_host *host)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	writel(0, host->base + MMCIMASK0);
+	if (host->variant->pwrreg_nopower) {
+		writel(0, host->base + MMCIDATACTRL);
+		writel(0, host->base + MMCIPOWER);
+		writel(0, host->base + MMCICLOCK);
+	}
+	mmci_reg_delay(host);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void mmci_restore(struct mmci_host *host)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (host->variant->pwrreg_nopower) {
+		writel(host->clk_reg, host->base + MMCICLOCK);
+		writel(host->datactrl_reg, host->base + MMCIDATACTRL);
+		writel(host->pwr_reg, host->base + MMCIPOWER);
+	}
+	writel(MCI_IRQENABLE | host->variant->start_err,
+	       host->base + MMCIMASK0);
+	mmci_reg_delay(host);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static int mmci_runtime_suspend(struct device *dev)
+{
+	struct amba_device *adev = to_amba_device(dev);
+	struct mmc_host *mmc = amba_get_drvdata(adev);
+
+	if (mmc) {
+		struct mmci_host *host = mmc_priv(mmc);
+		pinctrl_pm_select_sleep_state(dev);
+		mmci_save(host);
+		clk_disable_unprepare(host->clk);
+	}
+
+	return 0;
+}
+
+static int mmci_runtime_resume(struct device *dev)
+{
+	struct amba_device *adev = to_amba_device(dev);
+	struct mmc_host *mmc = amba_get_drvdata(adev);
+
+	if (mmc) {
+		struct mmci_host *host = mmc_priv(mmc);
+		clk_prepare_enable(host->clk);
+		mmci_restore(host);
+		pinctrl_select_default_state(dev);
+	}
+
+	return 0;
+}
+#endif
+
+static const struct dev_pm_ops mmci_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(mmci_runtime_suspend, mmci_runtime_resume, NULL)
+};
+
+static const struct amba_id mmci_ids[] = {
+	{
+		.id	= 0x00041180,
+		.mask	= 0xff0fffff,
+		.data	= &variant_arm,
+	},
+	{
+		.id	= 0x01041180,
+		.mask	= 0xff0fffff,
+		.data	= &variant_arm_extended_fifo,
+	},
+	{
+		.id	= 0x02041180,
+		.mask	= 0xff0fffff,
+		.data	= &variant_arm_extended_fifo_hwfc,
+	},
+	{
+		.id	= 0x00041181,
+		.mask	= 0x000fffff,
+		.data	= &variant_arm,
+	},
+	/* ST Micro variants */
+	{
+		.id     = 0x00180180,
+		.mask   = 0x00ffffff,
+		.data	= &variant_u300,
+	},
+	{
+		.id     = 0x10180180,
+		.mask   = 0xf0ffffff,
+		.data	= &variant_nomadik,
+	},
+	{
+		.id     = 0x00280180,
+		.mask   = 0x00ffffff,
+		.data	= &variant_nomadik,
+	},
+	{
+		.id     = 0x00480180,
+		.mask   = 0xf0ffffff,
+		.data	= &variant_ux500,
+	},
+	{
+		.id     = 0x10480180,
+		.mask   = 0xf0ffffff,
+		.data	= &variant_ux500v2,
+	},
+	{
+		.id     = 0x00880180,
+		.mask   = 0x00ffffff,
+		.data	= &variant_stm32,
+	},
+	{
+		.id     = 0x10153180,
+		.mask	= 0xf0ffffff,
+		.data	= &variant_stm32_sdmmc,
+	},
+	{
+		.id     = 0x00253180,
+		.mask	= 0xf0ffffff,
+		.data	= &variant_stm32_sdmmcv2,
+	},
+	/* Qualcomm variants */
+	{
+		.id     = 0x00051180,
+		.mask	= 0x000fffff,
+		.data	= &variant_qcom,
+	},
+	{ 0, 0 },
+};
+
+MODULE_DEVICE_TABLE(amba, mmci_ids);
+
+static struct amba_driver mmci_driver = {
+	.drv		= {
+		.name	= DRIVER_NAME,
+		.pm	= &mmci_dev_pm_ops,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+	},
+	.probe		= mmci_probe,
+	.remove		= mmci_remove,
+	.id_table	= mmci_ids,
+};
+
+module_amba_driver(mmci_driver);
+
+module_param(fmax, uint, 0444);
+
+MODULE_DESCRIPTION("ARM PrimeCell PL180/181 Multimedia Card Interface driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mmc/host/mmci.h b/drivers/mmc/host/mmci.h
new file mode 100644
index 000000000..e1a9b96a3
--- /dev/null
+++ b/drivers/mmc/host/mmci.h
@@ -0,0 +1,475 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ *  linux/drivers/mmc/host/mmci.h - ARM PrimeCell MMCI PL180/1 driver
+ *
+ *  Copyright (C) 2003 Deep Blue Solutions, Ltd, All Rights Reserved.
+ */
+#define MMCIPOWER		0x000
+#define MCI_PWR_OFF		0x00
+#define MCI_PWR_UP		0x02
+#define MCI_PWR_ON		0x03
+#define MCI_OD			(1 << 6)
+#define MCI_ROD			(1 << 7)
+/*
+ * The ST Micro version does not have ROD and reuse the voltage registers for
+ * direction settings.
+ */
+#define MCI_ST_DATA2DIREN	(1 << 2)
+#define MCI_ST_CMDDIREN		(1 << 3)
+#define MCI_ST_DATA0DIREN	(1 << 4)
+#define MCI_ST_DATA31DIREN	(1 << 5)
+#define MCI_ST_FBCLKEN		(1 << 7)
+#define MCI_ST_DATA74DIREN	(1 << 8)
+/*
+ * The STM32 sdmmc does not have PWR_UP/OD/ROD
+ * and uses the power register for
+ */
+#define MCI_STM32_PWR_CYC	0x02
+#define MCI_STM32_VSWITCH	BIT(2)
+#define MCI_STM32_VSWITCHEN	BIT(3)
+#define MCI_STM32_DIRPOL	BIT(4)
+
+#define MMCICLOCK		0x004
+#define MCI_CLK_ENABLE		(1 << 8)
+#define MCI_CLK_PWRSAVE		(1 << 9)
+#define MCI_CLK_BYPASS		(1 << 10)
+#define MCI_4BIT_BUS		(1 << 11)
+/*
+ * 8bit wide buses, hardware flow contronl, negative edges and clock inversion
+ * supported in ST Micro U300 and Ux500 versions
+ */
+#define MCI_ST_8BIT_BUS		(1 << 12)
+#define MCI_ST_U300_HWFCEN	(1 << 13)
+#define MCI_ST_UX500_NEG_EDGE	(1 << 13)
+#define MCI_ST_UX500_HWFCEN	(1 << 14)
+#define MCI_ST_UX500_CLK_INV	(1 << 15)
+/* Modified PL180 on Versatile Express platform */
+#define MCI_ARM_HWFCEN		(1 << 12)
+
+/* Modified on Qualcomm Integrations */
+#define MCI_QCOM_CLK_WIDEBUS_8	(BIT(10) | BIT(11))
+#define MCI_QCOM_CLK_FLOWENA	BIT(12)
+#define MCI_QCOM_CLK_INVERTOUT	BIT(13)
+
+/* select in latch data and command in */
+#define MCI_QCOM_CLK_SELECT_IN_FBCLK	BIT(15)
+#define MCI_QCOM_CLK_SELECT_IN_DDR_MODE	(BIT(14) | BIT(15))
+
+/* Modified on STM32 sdmmc */
+#define MCI_STM32_CLK_CLKDIV_MSK	GENMASK(9, 0)
+#define MCI_STM32_CLK_WIDEBUS_4		BIT(14)
+#define MCI_STM32_CLK_WIDEBUS_8		BIT(15)
+#define MCI_STM32_CLK_NEGEDGE		BIT(16)
+#define MCI_STM32_CLK_HWFCEN		BIT(17)
+#define MCI_STM32_CLK_DDR		BIT(18)
+#define MCI_STM32_CLK_BUSSPEED		BIT(19)
+#define MCI_STM32_CLK_SEL_MSK		GENMASK(21, 20)
+#define MCI_STM32_CLK_SELCK		(0 << 20)
+#define MCI_STM32_CLK_SELCKIN		(1 << 20)
+#define MCI_STM32_CLK_SELFBCK		(2 << 20)
+
+#define MMCIARGUMENT		0x008
+
+/* The command register controls the Command Path State Machine (CPSM) */
+#define MMCICOMMAND		0x00c
+#define MCI_CPSM_RESPONSE	BIT(6)
+#define MCI_CPSM_LONGRSP	BIT(7)
+#define MCI_CPSM_INTERRUPT	BIT(8)
+#define MCI_CPSM_PENDING	BIT(9)
+#define MCI_CPSM_ENABLE		BIT(10)
+/* Command register flag extenstions in the ST Micro versions */
+#define MCI_CPSM_ST_SDIO_SUSP		BIT(11)
+#define MCI_CPSM_ST_ENCMD_COMPL		BIT(12)
+#define MCI_CPSM_ST_NIEN		BIT(13)
+#define MCI_CPSM_ST_CE_ATACMD		BIT(14)
+/* Command register flag extensions in the Qualcomm versions */
+#define MCI_CPSM_QCOM_PROGENA		BIT(11)
+#define MCI_CPSM_QCOM_DATCMD		BIT(12)
+#define MCI_CPSM_QCOM_MCIABORT		BIT(13)
+#define MCI_CPSM_QCOM_CCSENABLE		BIT(14)
+#define MCI_CPSM_QCOM_CCSDISABLE	BIT(15)
+#define MCI_CPSM_QCOM_AUTO_CMD19	BIT(16)
+#define MCI_CPSM_QCOM_AUTO_CMD21	BIT(21)
+/* Command register in STM32 sdmmc versions */
+#define MCI_CPSM_STM32_CMDTRANS		BIT(6)
+#define MCI_CPSM_STM32_CMDSTOP		BIT(7)
+#define MCI_CPSM_STM32_WAITRESP_MASK	GENMASK(9, 8)
+#define MCI_CPSM_STM32_NORSP		(0 << 8)
+#define MCI_CPSM_STM32_SRSP_CRC		(1 << 8)
+#define MCI_CPSM_STM32_SRSP		(2 << 8)
+#define MCI_CPSM_STM32_LRSP_CRC		(3 << 8)
+#define MCI_CPSM_STM32_ENABLE		BIT(12)
+
+#define MMCIRESPCMD		0x010
+#define MMCIRESPONSE0		0x014
+#define MMCIRESPONSE1		0x018
+#define MMCIRESPONSE2		0x01c
+#define MMCIRESPONSE3		0x020
+#define MMCIDATATIMER		0x024
+#define MMCIDATALENGTH		0x028
+
+/* The data control register controls the Data Path State Machine (DPSM) */
+#define MMCIDATACTRL		0x02c
+#define MCI_DPSM_ENABLE		BIT(0)
+#define MCI_DPSM_DIRECTION	BIT(1)
+#define MCI_DPSM_MODE		BIT(2)
+#define MCI_DPSM_DMAENABLE	BIT(3)
+#define MCI_DPSM_BLOCKSIZE	BIT(4)
+/* Control register extensions in the ST Micro U300 and Ux500 versions */
+#define MCI_DPSM_ST_RWSTART	BIT(8)
+#define MCI_DPSM_ST_RWSTOP	BIT(9)
+#define MCI_DPSM_ST_RWMOD	BIT(10)
+#define MCI_DPSM_ST_SDIOEN	BIT(11)
+/* Control register extensions in the ST Micro Ux500 versions */
+#define MCI_DPSM_ST_DMAREQCTL	BIT(12)
+#define MCI_DPSM_ST_DBOOTMODEEN	BIT(13)
+#define MCI_DPSM_ST_BUSYMODE	BIT(14)
+#define MCI_DPSM_ST_DDRMODE	BIT(15)
+/* Control register extensions in the Qualcomm versions */
+#define MCI_DPSM_QCOM_DATA_PEND	BIT(17)
+#define MCI_DPSM_QCOM_RX_DATA_PEND BIT(20)
+/* Control register extensions in STM32 versions */
+#define MCI_DPSM_STM32_MODE_BLOCK	(0 << 2)
+#define MCI_DPSM_STM32_MODE_SDIO	(1 << 2)
+#define MCI_DPSM_STM32_MODE_STREAM	(2 << 2)
+#define MCI_DPSM_STM32_MODE_BLOCK_STOP	(3 << 2)
+
+#define MMCIDATACNT		0x030
+#define MMCISTATUS		0x034
+#define MCI_CMDCRCFAIL		(1 << 0)
+#define MCI_DATACRCFAIL		(1 << 1)
+#define MCI_CMDTIMEOUT		(1 << 2)
+#define MCI_DATATIMEOUT		(1 << 3)
+#define MCI_TXUNDERRUN		(1 << 4)
+#define MCI_RXOVERRUN		(1 << 5)
+#define MCI_CMDRESPEND		(1 << 6)
+#define MCI_CMDSENT		(1 << 7)
+#define MCI_DATAEND		(1 << 8)
+#define MCI_STARTBITERR		(1 << 9)
+#define MCI_DATABLOCKEND	(1 << 10)
+#define MCI_CMDACTIVE		(1 << 11)
+#define MCI_TXACTIVE		(1 << 12)
+#define MCI_RXACTIVE		(1 << 13)
+#define MCI_TXFIFOHALFEMPTY	(1 << 14)
+#define MCI_RXFIFOHALFFULL	(1 << 15)
+#define MCI_TXFIFOFULL		(1 << 16)
+#define MCI_RXFIFOFULL		(1 << 17)
+#define MCI_TXFIFOEMPTY		(1 << 18)
+#define MCI_RXFIFOEMPTY		(1 << 19)
+#define MCI_TXDATAAVLBL		(1 << 20)
+#define MCI_RXDATAAVLBL		(1 << 21)
+/* Extended status bits for the ST Micro variants */
+#define MCI_ST_SDIOIT		(1 << 22)
+#define MCI_ST_CEATAEND		(1 << 23)
+#define MCI_ST_CARDBUSY		(1 << 24)
+/* Extended status bits for the STM32 variants */
+#define MCI_STM32_BUSYD0	BIT(20)
+#define MCI_STM32_BUSYD0END	BIT(21)
+#define MCI_STM32_VSWEND	BIT(25)
+
+#define MMCICLEAR		0x038
+#define MCI_CMDCRCFAILCLR	(1 << 0)
+#define MCI_DATACRCFAILCLR	(1 << 1)
+#define MCI_CMDTIMEOUTCLR	(1 << 2)
+#define MCI_DATATIMEOUTCLR	(1 << 3)
+#define MCI_TXUNDERRUNCLR	(1 << 4)
+#define MCI_RXOVERRUNCLR	(1 << 5)
+#define MCI_CMDRESPENDCLR	(1 << 6)
+#define MCI_CMDSENTCLR		(1 << 7)
+#define MCI_DATAENDCLR		(1 << 8)
+#define MCI_STARTBITERRCLR	(1 << 9)
+#define MCI_DATABLOCKENDCLR	(1 << 10)
+/* Extended status bits for the ST Micro variants */
+#define MCI_ST_SDIOITC		(1 << 22)
+#define MCI_ST_CEATAENDC	(1 << 23)
+#define MCI_ST_BUSYENDC		(1 << 24)
+/* Extended clear bits for the STM32 variants */
+#define MCI_STM32_VSWENDC	BIT(25)
+#define MCI_STM32_CKSTOPC	BIT(26)
+
+#define MMCIMASK0		0x03c
+#define MCI_CMDCRCFAILMASK	(1 << 0)
+#define MCI_DATACRCFAILMASK	(1 << 1)
+#define MCI_CMDTIMEOUTMASK	(1 << 2)
+#define MCI_DATATIMEOUTMASK	(1 << 3)
+#define MCI_TXUNDERRUNMASK	(1 << 4)
+#define MCI_RXOVERRUNMASK	(1 << 5)
+#define MCI_CMDRESPENDMASK	(1 << 6)
+#define MCI_CMDSENTMASK		(1 << 7)
+#define MCI_DATAENDMASK		(1 << 8)
+#define MCI_STARTBITERRMASK	(1 << 9)
+#define MCI_DATABLOCKENDMASK	(1 << 10)
+#define MCI_CMDACTIVEMASK	(1 << 11)
+#define MCI_TXACTIVEMASK	(1 << 12)
+#define MCI_RXACTIVEMASK	(1 << 13)
+#define MCI_TXFIFOHALFEMPTYMASK	(1 << 14)
+#define MCI_RXFIFOHALFFULLMASK	(1 << 15)
+#define MCI_TXFIFOFULLMASK	(1 << 16)
+#define MCI_RXFIFOFULLMASK	(1 << 17)
+#define MCI_TXFIFOEMPTYMASK	(1 << 18)
+#define MCI_RXFIFOEMPTYMASK	(1 << 19)
+#define MCI_TXDATAAVLBLMASK	(1 << 20)
+#define MCI_RXDATAAVLBLMASK	(1 << 21)
+/* Extended status bits for the ST Micro variants */
+#define MCI_ST_SDIOITMASK	(1 << 22)
+#define MCI_ST_CEATAENDMASK	(1 << 23)
+#define MCI_ST_BUSYENDMASK	(1 << 24)
+/* Extended status bits for the STM32 variants */
+#define MCI_STM32_BUSYD0ENDMASK	BIT(21)
+
+#define MMCIMASK1		0x040
+#define MMCIFIFOCNT		0x048
+#define MMCIFIFO		0x080 /* to 0x0bc */
+
+/* STM32 sdmmc registers for IDMA (Internal DMA) */
+#define MMCI_STM32_IDMACTRLR	0x050
+#define MMCI_STM32_IDMAEN	BIT(0)
+#define MMCI_STM32_IDMALLIEN	BIT(1)
+
+#define MMCI_STM32_IDMABSIZER		0x054
+#define MMCI_STM32_IDMABNDT_SHIFT	5
+#define MMCI_STM32_IDMABNDT_MASK	GENMASK(12, 5)
+
+#define MMCI_STM32_IDMABASE0R	0x058
+
+#define MMCI_STM32_IDMALAR	0x64
+#define MMCI_STM32_IDMALA_MASK	GENMASK(13, 0)
+#define MMCI_STM32_ABR		BIT(29)
+#define MMCI_STM32_ULS		BIT(30)
+#define MMCI_STM32_ULA		BIT(31)
+
+#define MMCI_STM32_IDMABAR	0x68
+
+#define MCI_IRQENABLE	\
+	(MCI_CMDCRCFAILMASK | MCI_DATACRCFAILMASK | MCI_CMDTIMEOUTMASK | \
+	MCI_DATATIMEOUTMASK | MCI_TXUNDERRUNMASK | MCI_RXOVERRUNMASK |	\
+	MCI_CMDRESPENDMASK | MCI_CMDSENTMASK)
+
+/* These interrupts are directed to IRQ1 when two IRQ lines are available */
+#define MCI_IRQ_PIO_MASK \
+	(MCI_RXFIFOHALFFULLMASK | MCI_RXDATAAVLBLMASK | \
+	 MCI_TXFIFOHALFEMPTYMASK)
+
+#define MCI_IRQ_PIO_STM32_MASK \
+	(MCI_RXFIFOHALFFULLMASK | MCI_TXFIFOHALFEMPTYMASK)
+
+#define NR_SG		128
+
+#define MMCI_PINCTRL_STATE_OPENDRAIN "opendrain"
+
+struct clk;
+struct dma_chan;
+struct mmci_host;
+
+/**
+ * struct variant_data - MMCI variant-specific quirks
+ * @clkreg: default value for MCICLOCK register
+ * @clkreg_enable: enable value for MMCICLOCK register
+ * @clkreg_8bit_bus_enable: enable value for 8 bit bus
+ * @clkreg_neg_edge_enable: enable value for inverted data/cmd output
+ * @cmdreg_cpsm_enable: enable value for CPSM
+ * @cmdreg_lrsp_crc: enable value for long response with crc
+ * @cmdreg_srsp_crc: enable value for short response with crc
+ * @cmdreg_srsp: enable value for short response without crc
+ * @cmdreg_stop: enable value for stop and abort transmission
+ * @datalength_bits: number of bits in the MMCIDATALENGTH register
+ * @fifosize: number of bytes that can be written when MMCI_TXFIFOEMPTY
+ *	      is asserted (likewise for RX)
+ * @fifohalfsize: number of bytes that can be written when MCI_TXFIFOHALFEMPTY
+ *		  is asserted (likewise for RX)
+ * @data_cmd_enable: enable value for data commands.
+ * @st_sdio: enable ST specific SDIO logic
+ * @st_clkdiv: true if using a ST-specific clock divider algorithm
+ * @stm32_clkdiv: true if using a STM32-specific clock divider algorithm
+ * @datactrl_mask_ddrmode: ddr mode mask in datactrl register.
+ * @datactrl_mask_sdio: SDIO enable mask in datactrl register
+ * @datactrl_blocksz: block size in power of two
+ * @datactrl_any_blocksz: true if block any block sizes are accepted by
+ *		  hardware, such as with some SDIO traffic that send
+ *		  odd packets.
+ * @dma_power_of_2: DMA only works with blocks that are a power of 2.
+ * @datactrl_first: true if data must be setup before send command
+ * @datacnt_useless: true if you could not use datacnt register to read
+ *		     remaining data
+ * @pwrreg_powerup: power up value for MMCIPOWER register
+ * @f_max: maximum clk frequency supported by the controller.
+ * @signal_direction: input/out direction of bus signals can be indicated
+ * @pwrreg_clkgate: MMCIPOWER register must be used to gate the clock
+ * @busy_detect: true if the variant supports busy detection on DAT0.
+ * @busy_timeout: true if the variant starts data timer when the DPSM
+ *		  enter in Wait_R or Busy state.
+ * @busy_dpsm_flag: bitmask enabling busy detection in the DPSM
+ * @busy_detect_flag: bitmask identifying the bit in the MMCISTATUS register
+ *		      indicating that the card is busy
+ * @busy_detect_mask: bitmask identifying the bit in the MMCIMASK0 to mask for
+ *		      getting busy end detection interrupts
+ * @pwrreg_nopower: bits in MMCIPOWER don't controls ext. power supply
+ * @explicit_mclk_control: enable explicit mclk control in driver.
+ * @qcom_fifo: enables qcom specific fifo pio read logic.
+ * @qcom_dml: enables qcom specific dma glue for dma transfers.
+ * @reversed_irq_handling: handle data irq before cmd irq.
+ * @mmcimask1: true if variant have a MMCIMASK1 register.
+ * @irq_pio_mask: bitmask used to manage interrupt pio transfert in mmcimask
+ *		  register
+ * @start_err: bitmask identifying the STARTBITERR bit inside MMCISTATUS
+ *	       register.
+ * @opendrain: bitmask identifying the OPENDRAIN bit inside MMCIPOWER register
+ * @dma_lli: true if variant has dma link list feature.
+ * @stm32_idmabsize_mask: stm32 sdmmc idma buffer size.
+ */
+struct variant_data {
+	unsigned int		clkreg;
+	unsigned int		clkreg_enable;
+	unsigned int		clkreg_8bit_bus_enable;
+	unsigned int		clkreg_neg_edge_enable;
+	unsigned int		cmdreg_cpsm_enable;
+	unsigned int		cmdreg_lrsp_crc;
+	unsigned int		cmdreg_srsp_crc;
+	unsigned int		cmdreg_srsp;
+	unsigned int		cmdreg_stop;
+	unsigned int		datalength_bits;
+	unsigned int		fifosize;
+	unsigned int		fifohalfsize;
+	unsigned int		data_cmd_enable;
+	unsigned int		datactrl_mask_ddrmode;
+	unsigned int		datactrl_mask_sdio;
+	unsigned int		datactrl_blocksz;
+	u8			datactrl_any_blocksz:1;
+	u8			dma_power_of_2:1;
+	u8			datactrl_first:1;
+	u8			datacnt_useless:1;
+	u8			st_sdio:1;
+	u8			st_clkdiv:1;
+	u8			stm32_clkdiv:1;
+	u32			pwrreg_powerup;
+	u32			f_max;
+	u8			signal_direction:1;
+	u8			pwrreg_clkgate:1;
+	u8			busy_detect:1;
+	u8			busy_timeout:1;
+	u32			busy_dpsm_flag;
+	u32			busy_detect_flag;
+	u32			busy_detect_mask;
+	u8			pwrreg_nopower:1;
+	u8			explicit_mclk_control:1;
+	u8			qcom_fifo:1;
+	u8			qcom_dml:1;
+	u8			reversed_irq_handling:1;
+	u8			mmcimask1:1;
+	unsigned int		irq_pio_mask;
+	u32			start_err;
+	u32			opendrain;
+	u8			dma_lli:1;
+	u32			stm32_idmabsize_mask;
+	void (*init)(struct mmci_host *host);
+};
+
+/* mmci variant callbacks */
+struct mmci_host_ops {
+	int (*validate_data)(struct mmci_host *host, struct mmc_data *data);
+	int (*prep_data)(struct mmci_host *host, struct mmc_data *data,
+			 bool next);
+	void (*unprep_data)(struct mmci_host *host, struct mmc_data *data,
+			    int err);
+	u32 (*get_datactrl_cfg)(struct mmci_host *host);
+	void (*get_next_data)(struct mmci_host *host, struct mmc_data *data);
+	int (*dma_setup)(struct mmci_host *host);
+	void (*dma_release)(struct mmci_host *host);
+	int (*dma_start)(struct mmci_host *host, unsigned int *datactrl);
+	void (*dma_finalize)(struct mmci_host *host, struct mmc_data *data);
+	void (*dma_error)(struct mmci_host *host);
+	void (*set_clkreg)(struct mmci_host *host, unsigned int desired);
+	void (*set_pwrreg)(struct mmci_host *host, unsigned int pwr);
+	bool (*busy_complete)(struct mmci_host *host, u32 status, u32 err_msk);
+	void (*pre_sig_volt_switch)(struct mmci_host *host);
+	int (*post_sig_volt_switch)(struct mmci_host *host, struct mmc_ios *ios);
+};
+
+struct mmci_host {
+	phys_addr_t		phybase;
+	void __iomem		*base;
+	struct mmc_request	*mrq;
+	struct mmc_command	*cmd;
+	struct mmc_command	stop_abort;
+	struct mmc_data		*data;
+	struct mmc_host		*mmc;
+	struct clk		*clk;
+	u8			singleirq:1;
+
+	struct reset_control	*rst;
+
+	spinlock_t		lock;
+
+	unsigned int		mclk;
+	/* cached value of requested clk in set_ios */
+	unsigned int		clock_cache;
+	unsigned int		cclk;
+	u32			pwr_reg;
+	u32			pwr_reg_add;
+	u32			clk_reg;
+	u32			clk_reg_add;
+	u32			datactrl_reg;
+	u32			busy_status;
+	u32			mask1_reg;
+	u8			vqmmc_enabled:1;
+	struct mmci_platform_data *plat;
+	struct mmc_host_ops	*mmc_ops;
+	struct mmci_host_ops	*ops;
+	struct variant_data	*variant;
+	void			*variant_priv;
+	struct pinctrl		*pinctrl;
+	struct pinctrl_state	*pins_opendrain;
+
+	u8			hw_designer;
+	u8			hw_revision:4;
+
+	struct timer_list	timer;
+	unsigned int		oldstat;
+	u32			irq_action;
+
+	/* pio stuff */
+	struct sg_mapping_iter	sg_miter;
+	unsigned int		size;
+	int (*get_rx_fifocnt)(struct mmci_host *h, u32 status, int remain);
+
+	u8			use_dma:1;
+	u8			dma_in_progress:1;
+	void			*dma_priv;
+
+	s32			next_cookie;
+};
+
+#define dma_inprogress(host)	((host)->dma_in_progress)
+
+void mmci_write_clkreg(struct mmci_host *host, u32 clk);
+void mmci_write_pwrreg(struct mmci_host *host, u32 pwr);
+
+static inline u32 mmci_dctrl_blksz(struct mmci_host *host)
+{
+	return (ffs(host->data->blksz) - 1) << 4;
+}
+
+#ifdef CONFIG_DMA_ENGINE
+int mmci_dmae_prep_data(struct mmci_host *host, struct mmc_data *data,
+			bool next);
+void mmci_dmae_unprep_data(struct mmci_host *host, struct mmc_data *data,
+			   int err);
+void mmci_dmae_get_next_data(struct mmci_host *host, struct mmc_data *data);
+int mmci_dmae_setup(struct mmci_host *host);
+void mmci_dmae_release(struct mmci_host *host);
+int mmci_dmae_start(struct mmci_host *host, unsigned int *datactrl);
+void mmci_dmae_finalize(struct mmci_host *host, struct mmc_data *data);
+void mmci_dmae_error(struct mmci_host *host);
+#endif
+
+#ifdef CONFIG_MMC_QCOM_DML
+void qcom_variant_init(struct mmci_host *host);
+#else
+static inline void qcom_variant_init(struct mmci_host *host) {}
+#endif
+
+#ifdef CONFIG_MMC_STM32_SDMMC
+void sdmmc_variant_init(struct mmci_host *host);
+#else
+static inline void sdmmc_variant_init(struct mmci_host *host) {}
+#endif
diff --git a/drivers/mmc/host/mmci_qcom_dml.c b/drivers/mmc/host/mmci_qcom_dml.c
new file mode 100644
index 000000000..3da6112fb
--- /dev/null
+++ b/drivers/mmc/host/mmci_qcom_dml.c
@@ -0,0 +1,202 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ *
+ * Copyright (c) 2011, The Linux Foundation. All rights reserved.
+ */
+#include <linux/of.h>
+#include <linux/of_dma.h>
+#include <linux/bitops.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/card.h>
+#include "mmci.h"
+
+/* Registers */
+#define DML_CONFIG			0x00
+#define PRODUCER_CRCI_MSK		GENMASK(1, 0)
+#define PRODUCER_CRCI_DISABLE		0
+#define PRODUCER_CRCI_X_SEL		BIT(0)
+#define PRODUCER_CRCI_Y_SEL		BIT(1)
+#define CONSUMER_CRCI_MSK		GENMASK(3, 2)
+#define CONSUMER_CRCI_DISABLE		0
+#define CONSUMER_CRCI_X_SEL		BIT(2)
+#define CONSUMER_CRCI_Y_SEL		BIT(3)
+#define PRODUCER_TRANS_END_EN		BIT(4)
+#define BYPASS				BIT(16)
+#define DIRECT_MODE			BIT(17)
+#define INFINITE_CONS_TRANS		BIT(18)
+
+#define DML_SW_RESET			0x08
+#define DML_PRODUCER_START		0x0c
+#define DML_CONSUMER_START		0x10
+#define DML_PRODUCER_PIPE_LOGICAL_SIZE	0x14
+#define DML_CONSUMER_PIPE_LOGICAL_SIZE	0x18
+#define DML_PIPE_ID			0x1c
+#define PRODUCER_PIPE_ID_SHFT		0
+#define PRODUCER_PIPE_ID_MSK		GENMASK(4, 0)
+#define CONSUMER_PIPE_ID_SHFT		16
+#define CONSUMER_PIPE_ID_MSK		GENMASK(20, 16)
+
+#define DML_PRODUCER_BAM_BLOCK_SIZE	0x24
+#define DML_PRODUCER_BAM_TRANS_SIZE	0x28
+
+/* other definitions */
+#define PRODUCER_PIPE_LOGICAL_SIZE	4096
+#define CONSUMER_PIPE_LOGICAL_SIZE	4096
+
+#define DML_OFFSET			0x800
+
+static int qcom_dma_start(struct mmci_host *host, unsigned int *datactrl)
+{
+	u32 config;
+	void __iomem *base = host->base + DML_OFFSET;
+	struct mmc_data *data = host->data;
+	int ret = mmci_dmae_start(host, datactrl);
+
+	if (ret)
+		return ret;
+
+	if (data->flags & MMC_DATA_READ) {
+		/* Read operation: configure DML for producer operation */
+		/* Set producer CRCI-x and disable consumer CRCI */
+		config = readl_relaxed(base + DML_CONFIG);
+		config = (config & ~PRODUCER_CRCI_MSK) | PRODUCER_CRCI_X_SEL;
+		config = (config & ~CONSUMER_CRCI_MSK) | CONSUMER_CRCI_DISABLE;
+		writel_relaxed(config, base + DML_CONFIG);
+
+		/* Set the Producer BAM block size */
+		writel_relaxed(data->blksz, base + DML_PRODUCER_BAM_BLOCK_SIZE);
+
+		/* Set Producer BAM Transaction size */
+		writel_relaxed(data->blocks * data->blksz,
+			       base + DML_PRODUCER_BAM_TRANS_SIZE);
+		/* Set Producer Transaction End bit */
+		config = readl_relaxed(base + DML_CONFIG);
+		config |= PRODUCER_TRANS_END_EN;
+		writel_relaxed(config, base + DML_CONFIG);
+		/* Trigger producer */
+		writel_relaxed(1, base + DML_PRODUCER_START);
+	} else {
+		/* Write operation: configure DML for consumer operation */
+		/* Set consumer CRCI-x and disable producer CRCI*/
+		config = readl_relaxed(base + DML_CONFIG);
+		config = (config & ~CONSUMER_CRCI_MSK) | CONSUMER_CRCI_X_SEL;
+		config = (config & ~PRODUCER_CRCI_MSK) | PRODUCER_CRCI_DISABLE;
+		writel_relaxed(config, base + DML_CONFIG);
+		/* Clear Producer Transaction End bit */
+		config = readl_relaxed(base + DML_CONFIG);
+		config &= ~PRODUCER_TRANS_END_EN;
+		writel_relaxed(config, base + DML_CONFIG);
+		/* Trigger consumer */
+		writel_relaxed(1, base + DML_CONSUMER_START);
+	}
+
+	/* make sure the dml is configured before dma is triggered */
+	wmb();
+	return 0;
+}
+
+static int of_get_dml_pipe_index(struct device_node *np, const char *name)
+{
+	int index;
+	struct of_phandle_args	dma_spec;
+
+	index = of_property_match_string(np, "dma-names", name);
+
+	if (index < 0)
+		return -ENODEV;
+
+	if (of_parse_phandle_with_args(np, "dmas", "#dma-cells", index,
+				       &dma_spec))
+		return -ENODEV;
+
+	if (dma_spec.args_count)
+		return dma_spec.args[0];
+
+	return -ENODEV;
+}
+
+/* Initialize the dml hardware connected to SD Card controller */
+static int qcom_dma_setup(struct mmci_host *host)
+{
+	u32 config;
+	void __iomem *base;
+	int consumer_id, producer_id;
+	struct device_node *np = host->mmc->parent->of_node;
+
+	if (mmci_dmae_setup(host))
+		return -EINVAL;
+
+	consumer_id = of_get_dml_pipe_index(np, "tx");
+	producer_id = of_get_dml_pipe_index(np, "rx");
+
+	if (producer_id < 0 || consumer_id < 0) {
+		mmci_dmae_release(host);
+		return -EINVAL;
+	}
+
+	base = host->base + DML_OFFSET;
+
+	/* Reset the DML block */
+	writel_relaxed(1, base + DML_SW_RESET);
+
+	/* Disable the producer and consumer CRCI */
+	config = (PRODUCER_CRCI_DISABLE | CONSUMER_CRCI_DISABLE);
+	/*
+	 * Disable the bypass mode. Bypass mode will only be used
+	 * if data transfer is to happen in PIO mode and don't
+	 * want the BAM interface to connect with SDCC-DML.
+	 */
+	config &= ~BYPASS;
+	/*
+	 * Disable direct mode as we don't DML to MASTER the AHB bus.
+	 * BAM connected with DML should MASTER the AHB bus.
+	 */
+	config &= ~DIRECT_MODE;
+	/*
+	 * Disable infinite mode transfer as we won't be doing any
+	 * infinite size data transfers. All data transfer will be
+	 * of finite data size.
+	 */
+	config &= ~INFINITE_CONS_TRANS;
+	writel_relaxed(config, base + DML_CONFIG);
+
+	/*
+	 * Initialize the logical BAM pipe size for producer
+	 * and consumer.
+	 */
+	writel_relaxed(PRODUCER_PIPE_LOGICAL_SIZE,
+		       base + DML_PRODUCER_PIPE_LOGICAL_SIZE);
+	writel_relaxed(CONSUMER_PIPE_LOGICAL_SIZE,
+		       base + DML_CONSUMER_PIPE_LOGICAL_SIZE);
+
+	/* Initialize Producer/consumer pipe id */
+	writel_relaxed(producer_id | (consumer_id << CONSUMER_PIPE_ID_SHFT),
+		       base + DML_PIPE_ID);
+
+	/* Make sure dml initialization is finished */
+	mb();
+
+	return 0;
+}
+
+static u32 qcom_get_dctrl_cfg(struct mmci_host *host)
+{
+	return MCI_DPSM_ENABLE | (host->data->blksz << 4);
+}
+
+static struct mmci_host_ops qcom_variant_ops = {
+	.prep_data = mmci_dmae_prep_data,
+	.unprep_data = mmci_dmae_unprep_data,
+	.get_datactrl_cfg = qcom_get_dctrl_cfg,
+	.get_next_data = mmci_dmae_get_next_data,
+	.dma_setup = qcom_dma_setup,
+	.dma_release = mmci_dmae_release,
+	.dma_start = qcom_dma_start,
+	.dma_finalize = mmci_dmae_finalize,
+	.dma_error = mmci_dmae_error,
+};
+
+void qcom_variant_init(struct mmci_host *host)
+{
+	host->ops = &qcom_variant_ops;
+}
diff --git a/drivers/mmc/host/mmci_stm32_sdmmc.c b/drivers/mmc/host/mmci_stm32_sdmmc.c
new file mode 100644
index 000000000..4cceb9bab
--- /dev/null
+++ b/drivers/mmc/host/mmci_stm32_sdmmc.c
@@ -0,0 +1,542 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) STMicroelectronics 2018 - All Rights Reserved
+ * Author: Ludovic.barre@st.com for STMicroelectronics.
+ */
+#include <linux/bitfield.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/iopoll.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/card.h>
+#include <linux/of_address.h>
+#include <linux/reset.h>
+#include <linux/scatterlist.h>
+#include "mmci.h"
+
+#define SDMMC_LLI_BUF_LEN	PAGE_SIZE
+#define SDMMC_IDMA_BURST	BIT(MMCI_STM32_IDMABNDT_SHIFT)
+
+#define DLYB_CR			0x0
+#define DLYB_CR_DEN		BIT(0)
+#define DLYB_CR_SEN		BIT(1)
+
+#define DLYB_CFGR		0x4
+#define DLYB_CFGR_SEL_MASK	GENMASK(3, 0)
+#define DLYB_CFGR_UNIT_MASK	GENMASK(14, 8)
+#define DLYB_CFGR_LNG_MASK	GENMASK(27, 16)
+#define DLYB_CFGR_LNGF		BIT(31)
+
+#define DLYB_NB_DELAY		11
+#define DLYB_CFGR_SEL_MAX	(DLYB_NB_DELAY + 1)
+#define DLYB_CFGR_UNIT_MAX	127
+
+#define DLYB_LNG_TIMEOUT_US	1000
+#define SDMMC_VSWEND_TIMEOUT_US 10000
+
+struct sdmmc_lli_desc {
+	u32 idmalar;
+	u32 idmabase;
+	u32 idmasize;
+};
+
+struct sdmmc_idma {
+	dma_addr_t sg_dma;
+	void *sg_cpu;
+};
+
+struct sdmmc_dlyb {
+	void __iomem *base;
+	u32 unit;
+	u32 max;
+};
+
+static int sdmmc_idma_validate_data(struct mmci_host *host,
+				    struct mmc_data *data)
+{
+	struct scatterlist *sg;
+	int i;
+
+	/*
+	 * idma has constraints on idmabase & idmasize for each element
+	 * excepted the last element which has no constraint on idmasize
+	 */
+	for_each_sg(data->sg, sg, data->sg_len - 1, i) {
+		if (!IS_ALIGNED(sg->offset, sizeof(u32)) ||
+		    !IS_ALIGNED(sg->length, SDMMC_IDMA_BURST)) {
+			dev_err(mmc_dev(host->mmc),
+				"unaligned scatterlist: ofst:%x length:%d\n",
+				data->sg->offset, data->sg->length);
+			return -EINVAL;
+		}
+	}
+
+	if (!IS_ALIGNED(sg->offset, sizeof(u32))) {
+		dev_err(mmc_dev(host->mmc),
+			"unaligned last scatterlist: ofst:%x length:%d\n",
+			data->sg->offset, data->sg->length);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int _sdmmc_idma_prep_data(struct mmci_host *host,
+				 struct mmc_data *data)
+{
+	int n_elem;
+
+	n_elem = dma_map_sg(mmc_dev(host->mmc),
+			    data->sg,
+			    data->sg_len,
+			    mmc_get_dma_dir(data));
+
+	if (!n_elem) {
+		dev_err(mmc_dev(host->mmc), "dma_map_sg failed\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int sdmmc_idma_prep_data(struct mmci_host *host,
+				struct mmc_data *data, bool next)
+{
+	/* Check if job is already prepared. */
+	if (!next && data->host_cookie == host->next_cookie)
+		return 0;
+
+	return _sdmmc_idma_prep_data(host, data);
+}
+
+static void sdmmc_idma_unprep_data(struct mmci_host *host,
+				   struct mmc_data *data, int err)
+{
+	dma_unmap_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+		     mmc_get_dma_dir(data));
+}
+
+static int sdmmc_idma_setup(struct mmci_host *host)
+{
+	struct sdmmc_idma *idma;
+	struct device *dev = mmc_dev(host->mmc);
+
+	idma = devm_kzalloc(dev, sizeof(*idma), GFP_KERNEL);
+	if (!idma)
+		return -ENOMEM;
+
+	host->dma_priv = idma;
+
+	if (host->variant->dma_lli) {
+		idma->sg_cpu = dmam_alloc_coherent(dev, SDMMC_LLI_BUF_LEN,
+						   &idma->sg_dma, GFP_KERNEL);
+		if (!idma->sg_cpu) {
+			dev_err(dev, "Failed to alloc IDMA descriptor\n");
+			return -ENOMEM;
+		}
+		host->mmc->max_segs = SDMMC_LLI_BUF_LEN /
+			sizeof(struct sdmmc_lli_desc);
+		host->mmc->max_seg_size = host->variant->stm32_idmabsize_mask;
+	} else {
+		host->mmc->max_segs = 1;
+		host->mmc->max_seg_size = host->mmc->max_req_size;
+	}
+
+	return dma_set_max_seg_size(dev, host->mmc->max_seg_size);
+}
+
+static int sdmmc_idma_start(struct mmci_host *host, unsigned int *datactrl)
+
+{
+	struct sdmmc_idma *idma = host->dma_priv;
+	struct sdmmc_lli_desc *desc = (struct sdmmc_lli_desc *)idma->sg_cpu;
+	struct mmc_data *data = host->data;
+	struct scatterlist *sg;
+	int i;
+
+	if (!host->variant->dma_lli || data->sg_len == 1) {
+		writel_relaxed(sg_dma_address(data->sg),
+			       host->base + MMCI_STM32_IDMABASE0R);
+		writel_relaxed(MMCI_STM32_IDMAEN,
+			       host->base + MMCI_STM32_IDMACTRLR);
+		return 0;
+	}
+
+	for_each_sg(data->sg, sg, data->sg_len, i) {
+		desc[i].idmalar = (i + 1) * sizeof(struct sdmmc_lli_desc);
+		desc[i].idmalar |= MMCI_STM32_ULA | MMCI_STM32_ULS
+			| MMCI_STM32_ABR;
+		desc[i].idmabase = sg_dma_address(sg);
+		desc[i].idmasize = sg_dma_len(sg);
+	}
+
+	/* notice the end of link list */
+	desc[data->sg_len - 1].idmalar &= ~MMCI_STM32_ULA;
+
+	dma_wmb();
+	writel_relaxed(idma->sg_dma, host->base + MMCI_STM32_IDMABAR);
+	writel_relaxed(desc[0].idmalar, host->base + MMCI_STM32_IDMALAR);
+	writel_relaxed(desc[0].idmabase, host->base + MMCI_STM32_IDMABASE0R);
+	writel_relaxed(desc[0].idmasize, host->base + MMCI_STM32_IDMABSIZER);
+	writel_relaxed(MMCI_STM32_IDMAEN | MMCI_STM32_IDMALLIEN,
+		       host->base + MMCI_STM32_IDMACTRLR);
+
+	return 0;
+}
+
+static void sdmmc_idma_finalize(struct mmci_host *host, struct mmc_data *data)
+{
+	writel_relaxed(0, host->base + MMCI_STM32_IDMACTRLR);
+
+	if (!data->host_cookie)
+		sdmmc_idma_unprep_data(host, data, 0);
+}
+
+static void mmci_sdmmc_set_clkreg(struct mmci_host *host, unsigned int desired)
+{
+	unsigned int clk = 0, ddr = 0;
+
+	if (host->mmc->ios.timing == MMC_TIMING_MMC_DDR52 ||
+	    host->mmc->ios.timing == MMC_TIMING_UHS_DDR50)
+		ddr = MCI_STM32_CLK_DDR;
+
+	/*
+	 * cclk = mclk / (2 * clkdiv)
+	 * clkdiv 0 => bypass
+	 * in ddr mode bypass is not possible
+	 */
+	if (desired) {
+		if (desired >= host->mclk && !ddr) {
+			host->cclk = host->mclk;
+		} else {
+			clk = DIV_ROUND_UP(host->mclk, 2 * desired);
+			if (clk > MCI_STM32_CLK_CLKDIV_MSK)
+				clk = MCI_STM32_CLK_CLKDIV_MSK;
+			host->cclk = host->mclk / (2 * clk);
+		}
+	} else {
+		/*
+		 * while power-on phase the clock can't be define to 0,
+		 * Only power-off and power-cyc deactivate the clock.
+		 * if desired clock is 0, set max divider
+		 */
+		clk = MCI_STM32_CLK_CLKDIV_MSK;
+		host->cclk = host->mclk / (2 * clk);
+	}
+
+	/* Set actual clock for debug */
+	if (host->mmc->ios.power_mode == MMC_POWER_ON)
+		host->mmc->actual_clock = host->cclk;
+	else
+		host->mmc->actual_clock = 0;
+
+	if (host->mmc->ios.bus_width == MMC_BUS_WIDTH_4)
+		clk |= MCI_STM32_CLK_WIDEBUS_4;
+	if (host->mmc->ios.bus_width == MMC_BUS_WIDTH_8)
+		clk |= MCI_STM32_CLK_WIDEBUS_8;
+
+	clk |= MCI_STM32_CLK_HWFCEN;
+	clk |= host->clk_reg_add;
+	clk |= ddr;
+
+	/*
+	 * SDMMC_FBCK is selected when an external Delay Block is needed
+	 * with SDR104.
+	 */
+	if (host->mmc->ios.timing >= MMC_TIMING_UHS_SDR50) {
+		clk |= MCI_STM32_CLK_BUSSPEED;
+		if (host->mmc->ios.timing == MMC_TIMING_UHS_SDR104) {
+			clk &= ~MCI_STM32_CLK_SEL_MSK;
+			clk |= MCI_STM32_CLK_SELFBCK;
+		}
+	}
+
+	mmci_write_clkreg(host, clk);
+}
+
+static void sdmmc_dlyb_input_ck(struct sdmmc_dlyb *dlyb)
+{
+	if (!dlyb || !dlyb->base)
+		return;
+
+	/* Output clock = Input clock */
+	writel_relaxed(0, dlyb->base + DLYB_CR);
+}
+
+static void mmci_sdmmc_set_pwrreg(struct mmci_host *host, unsigned int pwr)
+{
+	struct mmc_ios ios = host->mmc->ios;
+	struct sdmmc_dlyb *dlyb = host->variant_priv;
+
+	/* adds OF options */
+	pwr = host->pwr_reg_add;
+
+	sdmmc_dlyb_input_ck(dlyb);
+
+	if (ios.power_mode == MMC_POWER_OFF) {
+		/* Only a reset could power-off sdmmc */
+		reset_control_assert(host->rst);
+		udelay(2);
+		reset_control_deassert(host->rst);
+
+		/*
+		 * Set the SDMMC in Power-cycle state.
+		 * This will make that the SDMMC_D[7:0], SDMMC_CMD and SDMMC_CK
+		 * are driven low, to prevent the Card from being supplied
+		 * through the signal lines.
+		 */
+		mmci_write_pwrreg(host, MCI_STM32_PWR_CYC | pwr);
+	} else if (ios.power_mode == MMC_POWER_ON) {
+		/*
+		 * After power-off (reset): the irq mask defined in probe
+		 * functionis lost
+		 * ault irq mask (probe) must be activated
+		 */
+		writel(MCI_IRQENABLE | host->variant->start_err,
+		       host->base + MMCIMASK0);
+
+		/* preserves voltage switch bits */
+		pwr |= host->pwr_reg & (MCI_STM32_VSWITCHEN |
+					MCI_STM32_VSWITCH);
+
+		/*
+		 * After a power-cycle state, we must set the SDMMC in
+		 * Power-off. The SDMMC_D[7:0], SDMMC_CMD and SDMMC_CK are
+		 * driven high. Then we can set the SDMMC to Power-on state
+		 */
+		mmci_write_pwrreg(host, MCI_PWR_OFF | pwr);
+		mdelay(1);
+		mmci_write_pwrreg(host, MCI_PWR_ON | pwr);
+	}
+}
+
+static u32 sdmmc_get_dctrl_cfg(struct mmci_host *host)
+{
+	u32 datactrl;
+
+	datactrl = mmci_dctrl_blksz(host);
+
+	if (host->mmc->card && mmc_card_sdio(host->mmc->card) &&
+	    host->data->blocks == 1)
+		datactrl |= MCI_DPSM_STM32_MODE_SDIO;
+	else if (host->data->stop && !host->mrq->sbc)
+		datactrl |= MCI_DPSM_STM32_MODE_BLOCK_STOP;
+	else
+		datactrl |= MCI_DPSM_STM32_MODE_BLOCK;
+
+	return datactrl;
+}
+
+static bool sdmmc_busy_complete(struct mmci_host *host, u32 status, u32 err_msk)
+{
+	void __iomem *base = host->base;
+	u32 busy_d0, busy_d0end, mask, sdmmc_status;
+
+	mask = readl_relaxed(base + MMCIMASK0);
+	sdmmc_status = readl_relaxed(base + MMCISTATUS);
+	busy_d0end = sdmmc_status & MCI_STM32_BUSYD0END;
+	busy_d0 = sdmmc_status & MCI_STM32_BUSYD0;
+
+	/* complete if there is an error or busy_d0end */
+	if ((status & err_msk) || busy_d0end)
+		goto complete;
+
+	/*
+	 * On response the busy signaling is reflected in the BUSYD0 flag.
+	 * if busy_d0 is in-progress we must activate busyd0end interrupt
+	 * to wait this completion. Else this request has no busy step.
+	 */
+	if (busy_d0) {
+		if (!host->busy_status) {
+			writel_relaxed(mask | host->variant->busy_detect_mask,
+				       base + MMCIMASK0);
+			host->busy_status = status &
+				(MCI_CMDSENT | MCI_CMDRESPEND);
+		}
+		return false;
+	}
+
+complete:
+	if (host->busy_status) {
+		writel_relaxed(mask & ~host->variant->busy_detect_mask,
+			       base + MMCIMASK0);
+		host->busy_status = 0;
+	}
+
+	writel_relaxed(host->variant->busy_detect_mask, base + MMCICLEAR);
+
+	return true;
+}
+
+static void sdmmc_dlyb_set_cfgr(struct sdmmc_dlyb *dlyb,
+				int unit, int phase, bool sampler)
+{
+	u32 cfgr;
+
+	writel_relaxed(DLYB_CR_SEN | DLYB_CR_DEN, dlyb->base + DLYB_CR);
+
+	cfgr = FIELD_PREP(DLYB_CFGR_UNIT_MASK, unit) |
+	       FIELD_PREP(DLYB_CFGR_SEL_MASK, phase);
+	writel_relaxed(cfgr, dlyb->base + DLYB_CFGR);
+
+	if (!sampler)
+		writel_relaxed(DLYB_CR_DEN, dlyb->base + DLYB_CR);
+}
+
+static int sdmmc_dlyb_lng_tuning(struct mmci_host *host)
+{
+	struct sdmmc_dlyb *dlyb = host->variant_priv;
+	u32 cfgr;
+	int i, lng, ret;
+
+	for (i = 0; i <= DLYB_CFGR_UNIT_MAX; i++) {
+		sdmmc_dlyb_set_cfgr(dlyb, i, DLYB_CFGR_SEL_MAX, true);
+
+		ret = readl_relaxed_poll_timeout(dlyb->base + DLYB_CFGR, cfgr,
+						 (cfgr & DLYB_CFGR_LNGF),
+						 1, DLYB_LNG_TIMEOUT_US);
+		if (ret) {
+			dev_warn(mmc_dev(host->mmc),
+				 "delay line cfg timeout unit:%d cfgr:%d\n",
+				 i, cfgr);
+			continue;
+		}
+
+		lng = FIELD_GET(DLYB_CFGR_LNG_MASK, cfgr);
+		if (lng < BIT(DLYB_NB_DELAY) && lng > 0)
+			break;
+	}
+
+	if (i > DLYB_CFGR_UNIT_MAX)
+		return -EINVAL;
+
+	dlyb->unit = i;
+	dlyb->max = __fls(lng);
+
+	return 0;
+}
+
+static int sdmmc_dlyb_phase_tuning(struct mmci_host *host, u32 opcode)
+{
+	struct sdmmc_dlyb *dlyb = host->variant_priv;
+	int cur_len = 0, max_len = 0, end_of_len = 0;
+	int phase;
+
+	for (phase = 0; phase <= dlyb->max; phase++) {
+		sdmmc_dlyb_set_cfgr(dlyb, dlyb->unit, phase, false);
+
+		if (mmc_send_tuning(host->mmc, opcode, NULL)) {
+			cur_len = 0;
+		} else {
+			cur_len++;
+			if (cur_len > max_len) {
+				max_len = cur_len;
+				end_of_len = phase;
+			}
+		}
+	}
+
+	if (!max_len) {
+		dev_err(mmc_dev(host->mmc), "no tuning point found\n");
+		return -EINVAL;
+	}
+
+	writel_relaxed(0, dlyb->base + DLYB_CR);
+
+	phase = end_of_len - max_len / 2;
+	sdmmc_dlyb_set_cfgr(dlyb, dlyb->unit, phase, false);
+
+	dev_dbg(mmc_dev(host->mmc), "unit:%d max_dly:%d phase:%d\n",
+		dlyb->unit, dlyb->max, phase);
+
+	return 0;
+}
+
+static int sdmmc_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct mmci_host *host = mmc_priv(mmc);
+	struct sdmmc_dlyb *dlyb = host->variant_priv;
+
+	if (!dlyb || !dlyb->base)
+		return -EINVAL;
+
+	if (sdmmc_dlyb_lng_tuning(host))
+		return -EINVAL;
+
+	return sdmmc_dlyb_phase_tuning(host, opcode);
+}
+
+static void sdmmc_pre_sig_volt_vswitch(struct mmci_host *host)
+{
+	/* clear the voltage switch completion flag */
+	writel_relaxed(MCI_STM32_VSWENDC, host->base + MMCICLEAR);
+	/* enable Voltage switch procedure */
+	mmci_write_pwrreg(host, host->pwr_reg | MCI_STM32_VSWITCHEN);
+}
+
+static int sdmmc_post_sig_volt_switch(struct mmci_host *host,
+				      struct mmc_ios *ios)
+{
+	unsigned long flags;
+	u32 status;
+	int ret = 0;
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_180 &&
+	    host->pwr_reg & MCI_STM32_VSWITCHEN) {
+		mmci_write_pwrreg(host, host->pwr_reg | MCI_STM32_VSWITCH);
+		spin_unlock_irqrestore(&host->lock, flags);
+
+		/* wait voltage switch completion while 10ms */
+		ret = readl_relaxed_poll_timeout(host->base + MMCISTATUS,
+						 status,
+						 (status & MCI_STM32_VSWEND),
+						 10, SDMMC_VSWEND_TIMEOUT_US);
+
+		writel_relaxed(MCI_STM32_VSWENDC | MCI_STM32_CKSTOPC,
+			       host->base + MMCICLEAR);
+		spin_lock_irqsave(&host->lock, flags);
+		mmci_write_pwrreg(host, host->pwr_reg &
+				  ~(MCI_STM32_VSWITCHEN | MCI_STM32_VSWITCH));
+	}
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	return ret;
+}
+
+static struct mmci_host_ops sdmmc_variant_ops = {
+	.validate_data = sdmmc_idma_validate_data,
+	.prep_data = sdmmc_idma_prep_data,
+	.unprep_data = sdmmc_idma_unprep_data,
+	.get_datactrl_cfg = sdmmc_get_dctrl_cfg,
+	.dma_setup = sdmmc_idma_setup,
+	.dma_start = sdmmc_idma_start,
+	.dma_finalize = sdmmc_idma_finalize,
+	.set_clkreg = mmci_sdmmc_set_clkreg,
+	.set_pwrreg = mmci_sdmmc_set_pwrreg,
+	.busy_complete = sdmmc_busy_complete,
+	.pre_sig_volt_switch = sdmmc_pre_sig_volt_vswitch,
+	.post_sig_volt_switch = sdmmc_post_sig_volt_switch,
+};
+
+void sdmmc_variant_init(struct mmci_host *host)
+{
+	struct device_node *np = host->mmc->parent->of_node;
+	void __iomem *base_dlyb;
+	struct sdmmc_dlyb *dlyb;
+
+	host->ops = &sdmmc_variant_ops;
+	host->pwr_reg = readl_relaxed(host->base + MMCIPOWER);
+
+	base_dlyb = devm_of_iomap(mmc_dev(host->mmc), np, 1, NULL);
+	if (IS_ERR(base_dlyb))
+		return;
+
+	dlyb = devm_kzalloc(mmc_dev(host->mmc), sizeof(*dlyb), GFP_KERNEL);
+	if (!dlyb)
+		return;
+
+	dlyb->base = base_dlyb;
+	host->variant_priv = dlyb;
+	host->mmc_ops->execute_tuning = sdmmc_execute_tuning;
+}
diff --git a/drivers/mmc/host/moxart-mmc.c b/drivers/mmc/host/moxart-mmc.c
new file mode 100644
index 000000000..c4baf8afb
--- /dev/null
+++ b/drivers/mmc/host/moxart-mmc.c
@@ -0,0 +1,727 @@
+/*
+ * MOXA ART MMC host driver.
+ *
+ * Copyright (C) 2014 Jonas Jensen
+ *
+ * Jonas Jensen <jonas.jensen@gmail.com>
+ *
+ * Based on code from
+ * Moxa Technologies Co., Ltd. <www.moxa.com>
+ *
+ * This file is licensed under the terms of the GNU General Public
+ * License version 2.  This program is licensed "as is" without any
+ * warranty of any kind, whether express or implied.
+ */
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/platform_device.h>
+#include <linux/delay.h>
+#include <linux/errno.h>
+#include <linux/interrupt.h>
+#include <linux/blkdev.h>
+#include <linux/dma-mapping.h>
+#include <linux/dmaengine.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/sd.h>
+#include <linux/sched.h>
+#include <linux/io.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/clk.h>
+#include <linux/bitops.h>
+#include <linux/of_dma.h>
+#include <linux/spinlock.h>
+
+#define REG_COMMAND		0
+#define REG_ARGUMENT		4
+#define REG_RESPONSE0		8
+#define REG_RESPONSE1		12
+#define REG_RESPONSE2		16
+#define REG_RESPONSE3		20
+#define REG_RESPONSE_COMMAND	24
+#define REG_DATA_CONTROL	28
+#define REG_DATA_TIMER		32
+#define REG_DATA_LENGTH		36
+#define REG_STATUS		40
+#define REG_CLEAR		44
+#define REG_INTERRUPT_MASK	48
+#define REG_POWER_CONTROL	52
+#define REG_CLOCK_CONTROL	56
+#define REG_BUS_WIDTH		60
+#define REG_DATA_WINDOW		64
+#define REG_FEATURE		68
+#define REG_REVISION		72
+
+/* REG_COMMAND */
+#define CMD_SDC_RESET		BIT(10)
+#define CMD_EN			BIT(9)
+#define CMD_APP_CMD		BIT(8)
+#define CMD_LONG_RSP		BIT(7)
+#define CMD_NEED_RSP		BIT(6)
+#define CMD_IDX_MASK		0x3f
+
+/* REG_RESPONSE_COMMAND */
+#define RSP_CMD_APP		BIT(6)
+#define RSP_CMD_IDX_MASK	0x3f
+
+/* REG_DATA_CONTROL */
+#define DCR_DATA_FIFO_RESET     BIT(8)
+#define DCR_DATA_THRES          BIT(7)
+#define DCR_DATA_EN		BIT(6)
+#define DCR_DMA_EN		BIT(5)
+#define DCR_DATA_WRITE		BIT(4)
+#define DCR_BLK_SIZE		0x0f
+
+/* REG_DATA_LENGTH */
+#define DATA_LEN_MASK		0xffffff
+
+/* REG_STATUS */
+#define WRITE_PROT		BIT(12)
+#define CARD_DETECT		BIT(11)
+/* 1-10 below can be sent to either registers, interrupt or clear. */
+#define CARD_CHANGE		BIT(10)
+#define FIFO_ORUN		BIT(9)
+#define FIFO_URUN		BIT(8)
+#define DATA_END		BIT(7)
+#define CMD_SENT		BIT(6)
+#define DATA_CRC_OK		BIT(5)
+#define RSP_CRC_OK		BIT(4)
+#define DATA_TIMEOUT		BIT(3)
+#define RSP_TIMEOUT		BIT(2)
+#define DATA_CRC_FAIL		BIT(1)
+#define RSP_CRC_FAIL		BIT(0)
+
+#define MASK_RSP		(RSP_TIMEOUT | RSP_CRC_FAIL | \
+				 RSP_CRC_OK  | CARD_DETECT  | CMD_SENT)
+
+#define MASK_DATA		(DATA_CRC_OK   | DATA_END | \
+				 DATA_CRC_FAIL | DATA_TIMEOUT)
+
+#define MASK_INTR_PIO		(FIFO_URUN | FIFO_ORUN | CARD_CHANGE)
+
+/* REG_POWER_CONTROL */
+#define SD_POWER_ON		BIT(4)
+#define SD_POWER_MASK		0x0f
+
+/* REG_CLOCK_CONTROL */
+#define CLK_HISPD		BIT(9)
+#define CLK_OFF			BIT(8)
+#define CLK_SD			BIT(7)
+#define CLK_DIV_MASK		0x7f
+
+/* REG_BUS_WIDTH */
+#define BUS_WIDTH_4_SUPPORT	BIT(3)
+#define BUS_WIDTH_4		BIT(2)
+#define BUS_WIDTH_1		BIT(0)
+
+#define MMC_VDD_360		23
+#define MIN_POWER		(MMC_VDD_360 - SD_POWER_MASK)
+#define MAX_RETRIES		500000
+
+struct moxart_host {
+	spinlock_t			lock;
+
+	void __iomem			*base;
+
+	phys_addr_t			reg_phys;
+
+	struct dma_chan			*dma_chan_tx;
+	struct dma_chan                 *dma_chan_rx;
+	struct dma_async_tx_descriptor	*tx_desc;
+	struct mmc_host			*mmc;
+	struct mmc_request		*mrq;
+	struct scatterlist		*cur_sg;
+	struct completion		dma_complete;
+	struct completion		pio_complete;
+
+	u32				num_sg;
+	u32				data_remain;
+	u32				data_len;
+	u32				fifo_width;
+	u32				timeout;
+	u32				rate;
+
+	long				sysclk;
+
+	bool				have_dma;
+	bool				is_removed;
+};
+
+static inline void moxart_init_sg(struct moxart_host *host,
+				  struct mmc_data *data)
+{
+	host->cur_sg = data->sg;
+	host->num_sg = data->sg_len;
+	host->data_remain = host->cur_sg->length;
+
+	if (host->data_remain > host->data_len)
+		host->data_remain = host->data_len;
+}
+
+static inline int moxart_next_sg(struct moxart_host *host)
+{
+	int remain;
+	struct mmc_data *data = host->mrq->cmd->data;
+
+	host->cur_sg++;
+	host->num_sg--;
+
+	if (host->num_sg > 0) {
+		host->data_remain = host->cur_sg->length;
+		remain = host->data_len - data->bytes_xfered;
+		if (remain > 0 && remain < host->data_remain)
+			host->data_remain = remain;
+	}
+
+	return host->num_sg;
+}
+
+static int moxart_wait_for_status(struct moxart_host *host,
+				  u32 mask, u32 *status)
+{
+	int ret = -ETIMEDOUT;
+	u32 i;
+
+	for (i = 0; i < MAX_RETRIES; i++) {
+		*status = readl(host->base + REG_STATUS);
+		if (!(*status & mask)) {
+			udelay(5);
+			continue;
+		}
+		writel(*status & mask, host->base + REG_CLEAR);
+		ret = 0;
+		break;
+	}
+
+	if (ret)
+		dev_err(mmc_dev(host->mmc), "timed out waiting for status\n");
+
+	return ret;
+}
+
+
+static void moxart_send_command(struct moxart_host *host,
+	struct mmc_command *cmd)
+{
+	u32 status, cmdctrl;
+
+	writel(RSP_TIMEOUT  | RSP_CRC_OK |
+	       RSP_CRC_FAIL | CMD_SENT, host->base + REG_CLEAR);
+	writel(cmd->arg, host->base + REG_ARGUMENT);
+
+	cmdctrl = cmd->opcode & CMD_IDX_MASK;
+	if (cmdctrl == SD_APP_SET_BUS_WIDTH    || cmdctrl == SD_APP_OP_COND   ||
+	    cmdctrl == SD_APP_SEND_SCR         || cmdctrl == SD_APP_SD_STATUS ||
+	    cmdctrl == SD_APP_SEND_NUM_WR_BLKS)
+		cmdctrl |= CMD_APP_CMD;
+
+	if (cmd->flags & MMC_RSP_PRESENT)
+		cmdctrl |= CMD_NEED_RSP;
+
+	if (cmd->flags & MMC_RSP_136)
+		cmdctrl |= CMD_LONG_RSP;
+
+	writel(cmdctrl | CMD_EN, host->base + REG_COMMAND);
+
+	if (moxart_wait_for_status(host, MASK_RSP, &status) == -ETIMEDOUT)
+		cmd->error = -ETIMEDOUT;
+
+	if (status & RSP_TIMEOUT) {
+		cmd->error = -ETIMEDOUT;
+		return;
+	}
+	if (status & RSP_CRC_FAIL) {
+		cmd->error = -EIO;
+		return;
+	}
+	if (status & RSP_CRC_OK) {
+		if (cmd->flags & MMC_RSP_136) {
+			cmd->resp[3] = readl(host->base + REG_RESPONSE0);
+			cmd->resp[2] = readl(host->base + REG_RESPONSE1);
+			cmd->resp[1] = readl(host->base + REG_RESPONSE2);
+			cmd->resp[0] = readl(host->base + REG_RESPONSE3);
+		} else {
+			cmd->resp[0] = readl(host->base + REG_RESPONSE0);
+		}
+	}
+}
+
+static void moxart_dma_complete(void *param)
+{
+	struct moxart_host *host = param;
+
+	complete(&host->dma_complete);
+}
+
+static void moxart_transfer_dma(struct mmc_data *data, struct moxart_host *host)
+{
+	u32 len, dir_slave;
+	long dma_time;
+	struct dma_async_tx_descriptor *desc = NULL;
+	struct dma_chan *dma_chan;
+
+	if (host->data_len == data->bytes_xfered)
+		return;
+
+	if (data->flags & MMC_DATA_WRITE) {
+		dma_chan = host->dma_chan_tx;
+		dir_slave = DMA_MEM_TO_DEV;
+	} else {
+		dma_chan = host->dma_chan_rx;
+		dir_slave = DMA_DEV_TO_MEM;
+	}
+
+	len = dma_map_sg(dma_chan->device->dev, data->sg,
+			 data->sg_len, mmc_get_dma_dir(data));
+
+	if (len > 0) {
+		desc = dmaengine_prep_slave_sg(dma_chan, data->sg,
+					       len, dir_slave,
+					       DMA_PREP_INTERRUPT |
+					       DMA_CTRL_ACK);
+	} else {
+		dev_err(mmc_dev(host->mmc), "dma_map_sg returned zero length\n");
+	}
+
+	if (desc) {
+		host->tx_desc = desc;
+		desc->callback = moxart_dma_complete;
+		desc->callback_param = host;
+		dmaengine_submit(desc);
+		dma_async_issue_pending(dma_chan);
+	}
+
+	data->bytes_xfered += host->data_remain;
+
+	dma_time = wait_for_completion_interruptible_timeout(
+		   &host->dma_complete, host->timeout);
+
+	dma_unmap_sg(dma_chan->device->dev,
+		     data->sg, data->sg_len,
+		     mmc_get_dma_dir(data));
+}
+
+
+static void moxart_transfer_pio(struct moxart_host *host)
+{
+	struct mmc_data *data = host->mrq->cmd->data;
+	u32 *sgp, len = 0, remain, status;
+
+	if (host->data_len == data->bytes_xfered)
+		return;
+
+	sgp = sg_virt(host->cur_sg);
+	remain = host->data_remain;
+
+	if (data->flags & MMC_DATA_WRITE) {
+		while (remain > 0) {
+			if (moxart_wait_for_status(host, FIFO_URUN, &status)
+			     == -ETIMEDOUT) {
+				data->error = -ETIMEDOUT;
+				complete(&host->pio_complete);
+				return;
+			}
+			for (len = 0; len < remain && len < host->fifo_width;) {
+				iowrite32(*sgp, host->base + REG_DATA_WINDOW);
+				sgp++;
+				len += 4;
+			}
+			remain -= len;
+		}
+
+	} else {
+		while (remain > 0) {
+			if (moxart_wait_for_status(host, FIFO_ORUN, &status)
+			    == -ETIMEDOUT) {
+				data->error = -ETIMEDOUT;
+				complete(&host->pio_complete);
+				return;
+			}
+			for (len = 0; len < remain && len < host->fifo_width;) {
+				*sgp = ioread32(host->base + REG_DATA_WINDOW);
+				sgp++;
+				len += 4;
+			}
+			remain -= len;
+		}
+	}
+
+	data->bytes_xfered += host->data_remain - remain;
+	host->data_remain = remain;
+
+	if (host->data_len != data->bytes_xfered)
+		moxart_next_sg(host);
+	else
+		complete(&host->pio_complete);
+}
+
+static void moxart_prepare_data(struct moxart_host *host)
+{
+	struct mmc_data *data = host->mrq->cmd->data;
+	u32 datactrl;
+	int blksz_bits;
+
+	if (!data)
+		return;
+
+	host->data_len = data->blocks * data->blksz;
+	blksz_bits = ffs(data->blksz) - 1;
+	BUG_ON(1 << blksz_bits != data->blksz);
+
+	moxart_init_sg(host, data);
+
+	datactrl = DCR_DATA_EN | (blksz_bits & DCR_BLK_SIZE);
+
+	if (data->flags & MMC_DATA_WRITE)
+		datactrl |= DCR_DATA_WRITE;
+
+	if ((host->data_len > host->fifo_width) && host->have_dma)
+		datactrl |= DCR_DMA_EN;
+
+	writel(DCR_DATA_FIFO_RESET, host->base + REG_DATA_CONTROL);
+	writel(MASK_DATA | FIFO_URUN | FIFO_ORUN, host->base + REG_CLEAR);
+	writel(host->rate, host->base + REG_DATA_TIMER);
+	writel(host->data_len, host->base + REG_DATA_LENGTH);
+	writel(datactrl, host->base + REG_DATA_CONTROL);
+}
+
+static void moxart_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct moxart_host *host = mmc_priv(mmc);
+	long pio_time;
+	unsigned long flags;
+	u32 status;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	init_completion(&host->dma_complete);
+	init_completion(&host->pio_complete);
+
+	host->mrq = mrq;
+
+	if (readl(host->base + REG_STATUS) & CARD_DETECT) {
+		mrq->cmd->error = -ETIMEDOUT;
+		goto request_done;
+	}
+
+	moxart_prepare_data(host);
+	moxart_send_command(host, host->mrq->cmd);
+
+	if (mrq->cmd->data) {
+		if ((host->data_len > host->fifo_width) && host->have_dma) {
+
+			writel(CARD_CHANGE, host->base + REG_INTERRUPT_MASK);
+
+			spin_unlock_irqrestore(&host->lock, flags);
+
+			moxart_transfer_dma(mrq->cmd->data, host);
+
+			spin_lock_irqsave(&host->lock, flags);
+		} else {
+
+			writel(MASK_INTR_PIO, host->base + REG_INTERRUPT_MASK);
+
+			spin_unlock_irqrestore(&host->lock, flags);
+
+			/* PIO transfers start from interrupt. */
+			pio_time = wait_for_completion_interruptible_timeout(
+				   &host->pio_complete, host->timeout);
+
+			spin_lock_irqsave(&host->lock, flags);
+		}
+
+		if (host->is_removed) {
+			dev_err(mmc_dev(host->mmc), "card removed\n");
+			mrq->cmd->error = -ETIMEDOUT;
+			goto request_done;
+		}
+
+		if (moxart_wait_for_status(host, MASK_DATA, &status)
+		    == -ETIMEDOUT) {
+			mrq->cmd->data->error = -ETIMEDOUT;
+			goto request_done;
+		}
+
+		if (status & DATA_CRC_FAIL)
+			mrq->cmd->data->error = -ETIMEDOUT;
+
+		if (mrq->cmd->data->stop)
+			moxart_send_command(host, mrq->cmd->data->stop);
+	}
+
+request_done:
+	spin_unlock_irqrestore(&host->lock, flags);
+	mmc_request_done(host->mmc, mrq);
+}
+
+static irqreturn_t moxart_irq(int irq, void *devid)
+{
+	struct moxart_host *host = (struct moxart_host *)devid;
+	u32 status;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	status = readl(host->base + REG_STATUS);
+	if (status & CARD_CHANGE) {
+		host->is_removed = status & CARD_DETECT;
+		if (host->is_removed && host->have_dma) {
+			dmaengine_terminate_all(host->dma_chan_tx);
+			dmaengine_terminate_all(host->dma_chan_rx);
+		}
+		host->mrq = NULL;
+		writel(MASK_INTR_PIO, host->base + REG_CLEAR);
+		writel(CARD_CHANGE, host->base + REG_INTERRUPT_MASK);
+		mmc_detect_change(host->mmc, 0);
+	}
+	if (status & (FIFO_ORUN | FIFO_URUN) && host->mrq)
+		moxart_transfer_pio(host);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	return IRQ_HANDLED;
+}
+
+static void moxart_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct moxart_host *host = mmc_priv(mmc);
+	unsigned long flags;
+	u8 power, div;
+	u32 ctrl;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (ios->clock) {
+		for (div = 0; div < CLK_DIV_MASK; ++div) {
+			if (ios->clock >= host->sysclk / (2 * (div + 1)))
+				break;
+		}
+		ctrl = CLK_SD | div;
+		host->rate = host->sysclk / (2 * (div + 1));
+		if (host->rate > host->sysclk)
+			ctrl |= CLK_HISPD;
+		writel(ctrl, host->base + REG_CLOCK_CONTROL);
+	}
+
+	if (ios->power_mode == MMC_POWER_OFF) {
+		writel(readl(host->base + REG_POWER_CONTROL) & ~SD_POWER_ON,
+		       host->base + REG_POWER_CONTROL);
+	} else {
+		if (ios->vdd < MIN_POWER)
+			power = 0;
+		else
+			power = ios->vdd - MIN_POWER;
+
+		writel(SD_POWER_ON | (u32) power,
+		       host->base + REG_POWER_CONTROL);
+	}
+
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_4:
+		writel(BUS_WIDTH_4, host->base + REG_BUS_WIDTH);
+		break;
+	default:
+		writel(BUS_WIDTH_1, host->base + REG_BUS_WIDTH);
+		break;
+	}
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+
+static int moxart_get_ro(struct mmc_host *mmc)
+{
+	struct moxart_host *host = mmc_priv(mmc);
+
+	return !!(readl(host->base + REG_STATUS) & WRITE_PROT);
+}
+
+static const struct mmc_host_ops moxart_ops = {
+	.request = moxart_request,
+	.set_ios = moxart_set_ios,
+	.get_ro = moxart_get_ro,
+};
+
+static int moxart_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct device_node *node = dev->of_node;
+	struct resource res_mmc;
+	struct mmc_host *mmc;
+	struct moxart_host *host = NULL;
+	struct dma_slave_config cfg;
+	struct clk *clk;
+	void __iomem *reg_mmc;
+	int irq, ret;
+	u32 i;
+
+	mmc = mmc_alloc_host(sizeof(struct moxart_host), dev);
+	if (!mmc) {
+		dev_err(dev, "mmc_alloc_host failed\n");
+		ret = -ENOMEM;
+		goto out_mmc;
+	}
+
+	ret = of_address_to_resource(node, 0, &res_mmc);
+	if (ret) {
+		dev_err(dev, "of_address_to_resource failed\n");
+		goto out_mmc;
+	}
+
+	irq = irq_of_parse_and_map(node, 0);
+	if (irq <= 0) {
+		dev_err(dev, "irq_of_parse_and_map failed\n");
+		ret = -EINVAL;
+		goto out_mmc;
+	}
+
+	clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(clk)) {
+		ret = PTR_ERR(clk);
+		goto out_mmc;
+	}
+
+	reg_mmc = devm_ioremap_resource(dev, &res_mmc);
+	if (IS_ERR(reg_mmc)) {
+		ret = PTR_ERR(reg_mmc);
+		goto out_mmc;
+	}
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto out_mmc;
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+	host->base = reg_mmc;
+	host->reg_phys = res_mmc.start;
+	host->timeout = msecs_to_jiffies(1000);
+	host->sysclk = clk_get_rate(clk);
+	host->fifo_width = readl(host->base + REG_FEATURE) << 2;
+	host->dma_chan_tx = dma_request_chan(dev, "tx");
+	host->dma_chan_rx = dma_request_chan(dev, "rx");
+
+	spin_lock_init(&host->lock);
+
+	mmc->ops = &moxart_ops;
+	mmc->f_max = DIV_ROUND_CLOSEST(host->sysclk, 2);
+	mmc->f_min = DIV_ROUND_CLOSEST(host->sysclk, CLK_DIV_MASK * 2);
+	mmc->ocr_avail = 0xffff00;	/* Support 2.0v - 3.6v power. */
+
+	if (IS_ERR(host->dma_chan_tx) || IS_ERR(host->dma_chan_rx)) {
+		if (PTR_ERR(host->dma_chan_tx) == -EPROBE_DEFER ||
+		    PTR_ERR(host->dma_chan_rx) == -EPROBE_DEFER) {
+			ret = -EPROBE_DEFER;
+			goto out;
+		}
+		if (!IS_ERR(host->dma_chan_tx)) {
+			dma_release_channel(host->dma_chan_tx);
+			host->dma_chan_tx = NULL;
+		}
+		if (!IS_ERR(host->dma_chan_rx)) {
+			dma_release_channel(host->dma_chan_rx);
+			host->dma_chan_rx = NULL;
+		}
+		dev_dbg(dev, "PIO mode transfer enabled\n");
+		host->have_dma = false;
+	} else {
+		dev_dbg(dev, "DMA channels found (%p,%p)\n",
+			 host->dma_chan_tx, host->dma_chan_rx);
+		host->have_dma = true;
+
+		memset(&cfg, 0, sizeof(cfg));
+		cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+		cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+
+		cfg.direction = DMA_MEM_TO_DEV;
+		cfg.src_addr = 0;
+		cfg.dst_addr = host->reg_phys + REG_DATA_WINDOW;
+		dmaengine_slave_config(host->dma_chan_tx, &cfg);
+
+		cfg.direction = DMA_DEV_TO_MEM;
+		cfg.src_addr = host->reg_phys + REG_DATA_WINDOW;
+		cfg.dst_addr = 0;
+		dmaengine_slave_config(host->dma_chan_rx, &cfg);
+	}
+
+	if (readl(host->base + REG_BUS_WIDTH) & BUS_WIDTH_4_SUPPORT)
+		mmc->caps |= MMC_CAP_4_BIT_DATA;
+
+	writel(0, host->base + REG_INTERRUPT_MASK);
+
+	writel(CMD_SDC_RESET, host->base + REG_COMMAND);
+	for (i = 0; i < MAX_RETRIES; i++) {
+		if (!(readl(host->base + REG_COMMAND) & CMD_SDC_RESET))
+			break;
+		udelay(5);
+	}
+
+	ret = devm_request_irq(dev, irq, moxart_irq, 0, "moxart-mmc", host);
+	if (ret)
+		goto out;
+
+	dev_set_drvdata(dev, mmc);
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto out;
+
+	dev_dbg(dev, "IRQ=%d, FIFO is %d bytes\n", irq, host->fifo_width);
+
+	return 0;
+
+out:
+	if (!IS_ERR_OR_NULL(host->dma_chan_tx))
+		dma_release_channel(host->dma_chan_tx);
+	if (!IS_ERR_OR_NULL(host->dma_chan_rx))
+		dma_release_channel(host->dma_chan_rx);
+out_mmc:
+	if (mmc)
+		mmc_free_host(mmc);
+	return ret;
+}
+
+static int moxart_remove(struct platform_device *pdev)
+{
+	struct mmc_host *mmc = dev_get_drvdata(&pdev->dev);
+	struct moxart_host *host = mmc_priv(mmc);
+
+	dev_set_drvdata(&pdev->dev, NULL);
+
+	if (!IS_ERR_OR_NULL(host->dma_chan_tx))
+		dma_release_channel(host->dma_chan_tx);
+	if (!IS_ERR_OR_NULL(host->dma_chan_rx))
+		dma_release_channel(host->dma_chan_rx);
+	mmc_remove_host(mmc);
+
+	writel(0, host->base + REG_INTERRUPT_MASK);
+	writel(0, host->base + REG_POWER_CONTROL);
+	writel(readl(host->base + REG_CLOCK_CONTROL) | CLK_OFF,
+	       host->base + REG_CLOCK_CONTROL);
+	mmc_free_host(mmc);
+
+	return 0;
+}
+
+static const struct of_device_id moxart_mmc_match[] = {
+	{ .compatible = "moxa,moxart-mmc" },
+	{ .compatible = "faraday,ftsdc010" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, moxart_mmc_match);
+
+static struct platform_driver moxart_mmc_driver = {
+	.probe      = moxart_probe,
+	.remove     = moxart_remove,
+	.driver     = {
+		.name		= "mmc-moxart",
+		.probe_type	= PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table	= moxart_mmc_match,
+	},
+};
+module_platform_driver(moxart_mmc_driver);
+
+MODULE_ALIAS("platform:mmc-moxart");
+MODULE_DESCRIPTION("MOXA ART MMC driver");
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Jonas Jensen <jonas.jensen@gmail.com>");
diff --git a/drivers/mmc/host/mtk-sd.c b/drivers/mmc/host/mtk-sd.c
new file mode 100644
index 000000000..2c9ea5ed0
--- /dev/null
+++ b/drivers/mmc/host/mtk-sd.c
@@ -0,0 +1,2735 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2014-2015 MediaTek Inc.
+ * Author: Chaotian.Jing <chaotian.jing@mediatek.com>
+ */
+
+#include <linux/module.h>
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/iopoll.h>
+#include <linux/ioport.h>
+#include <linux/irq.h>
+#include <linux/of_address.h>
+#include <linux/of_device.h>
+#include <linux/of_irq.h>
+#include <linux/of_gpio.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/platform_device.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+#include <linux/regulator/consumer.h>
+#include <linux/slab.h>
+#include <linux/spinlock.h>
+#include <linux/interrupt.h>
+#include <linux/reset.h>
+
+#include <linux/mmc/card.h>
+#include <linux/mmc/core.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sd.h>
+#include <linux/mmc/sdio.h>
+#include <linux/mmc/slot-gpio.h>
+
+#include "cqhci.h"
+
+#define MAX_BD_NUM          1024
+
+/*--------------------------------------------------------------------------*/
+/* Common Definition                                                        */
+/*--------------------------------------------------------------------------*/
+#define MSDC_BUS_1BITS          0x0
+#define MSDC_BUS_4BITS          0x1
+#define MSDC_BUS_8BITS          0x2
+
+#define MSDC_BURST_64B          0x6
+
+/*--------------------------------------------------------------------------*/
+/* Register Offset                                                          */
+/*--------------------------------------------------------------------------*/
+#define MSDC_CFG         0x0
+#define MSDC_IOCON       0x04
+#define MSDC_PS          0x08
+#define MSDC_INT         0x0c
+#define MSDC_INTEN       0x10
+#define MSDC_FIFOCS      0x14
+#define SDC_CFG          0x30
+#define SDC_CMD          0x34
+#define SDC_ARG          0x38
+#define SDC_STS          0x3c
+#define SDC_RESP0        0x40
+#define SDC_RESP1        0x44
+#define SDC_RESP2        0x48
+#define SDC_RESP3        0x4c
+#define SDC_BLK_NUM      0x50
+#define SDC_ADV_CFG0     0x64
+#define EMMC_IOCON       0x7c
+#define SDC_ACMD_RESP    0x80
+#define DMA_SA_H4BIT     0x8c
+#define MSDC_DMA_SA      0x90
+#define MSDC_DMA_CTRL    0x98
+#define MSDC_DMA_CFG     0x9c
+#define MSDC_PATCH_BIT   0xb0
+#define MSDC_PATCH_BIT1  0xb4
+#define MSDC_PATCH_BIT2  0xb8
+#define MSDC_PAD_TUNE    0xec
+#define MSDC_PAD_TUNE0   0xf0
+#define PAD_DS_TUNE      0x188
+#define PAD_CMD_TUNE     0x18c
+#define EMMC50_CFG0      0x208
+#define EMMC50_CFG3      0x220
+#define SDC_FIFO_CFG     0x228
+
+/*--------------------------------------------------------------------------*/
+/* Top Pad Register Offset                                                  */
+/*--------------------------------------------------------------------------*/
+#define EMMC_TOP_CONTROL	0x00
+#define EMMC_TOP_CMD		0x04
+#define EMMC50_PAD_DS_TUNE	0x0c
+
+/*--------------------------------------------------------------------------*/
+/* Register Mask                                                            */
+/*--------------------------------------------------------------------------*/
+
+/* MSDC_CFG mask */
+#define MSDC_CFG_MODE           (0x1 << 0)	/* RW */
+#define MSDC_CFG_CKPDN          (0x1 << 1)	/* RW */
+#define MSDC_CFG_RST            (0x1 << 2)	/* RW */
+#define MSDC_CFG_PIO            (0x1 << 3)	/* RW */
+#define MSDC_CFG_CKDRVEN        (0x1 << 4)	/* RW */
+#define MSDC_CFG_BV18SDT        (0x1 << 5)	/* RW */
+#define MSDC_CFG_BV18PSS        (0x1 << 6)	/* R  */
+#define MSDC_CFG_CKSTB          (0x1 << 7)	/* R  */
+#define MSDC_CFG_CKDIV          (0xff << 8)	/* RW */
+#define MSDC_CFG_CKMOD          (0x3 << 16)	/* RW */
+#define MSDC_CFG_HS400_CK_MODE  (0x1 << 18)	/* RW */
+#define MSDC_CFG_HS400_CK_MODE_EXTRA  (0x1 << 22)	/* RW */
+#define MSDC_CFG_CKDIV_EXTRA    (0xfff << 8)	/* RW */
+#define MSDC_CFG_CKMOD_EXTRA    (0x3 << 20)	/* RW */
+
+/* MSDC_IOCON mask */
+#define MSDC_IOCON_SDR104CKS    (0x1 << 0)	/* RW */
+#define MSDC_IOCON_RSPL         (0x1 << 1)	/* RW */
+#define MSDC_IOCON_DSPL         (0x1 << 2)	/* RW */
+#define MSDC_IOCON_DDLSEL       (0x1 << 3)	/* RW */
+#define MSDC_IOCON_DDR50CKD     (0x1 << 4)	/* RW */
+#define MSDC_IOCON_DSPLSEL      (0x1 << 5)	/* RW */
+#define MSDC_IOCON_W_DSPL       (0x1 << 8)	/* RW */
+#define MSDC_IOCON_D0SPL        (0x1 << 16)	/* RW */
+#define MSDC_IOCON_D1SPL        (0x1 << 17)	/* RW */
+#define MSDC_IOCON_D2SPL        (0x1 << 18)	/* RW */
+#define MSDC_IOCON_D3SPL        (0x1 << 19)	/* RW */
+#define MSDC_IOCON_D4SPL        (0x1 << 20)	/* RW */
+#define MSDC_IOCON_D5SPL        (0x1 << 21)	/* RW */
+#define MSDC_IOCON_D6SPL        (0x1 << 22)	/* RW */
+#define MSDC_IOCON_D7SPL        (0x1 << 23)	/* RW */
+#define MSDC_IOCON_RISCSZ       (0x3 << 24)	/* RW */
+
+/* MSDC_PS mask */
+#define MSDC_PS_CDEN            (0x1 << 0)	/* RW */
+#define MSDC_PS_CDSTS           (0x1 << 1)	/* R  */
+#define MSDC_PS_CDDEBOUNCE      (0xf << 12)	/* RW */
+#define MSDC_PS_DAT             (0xff << 16)	/* R  */
+#define MSDC_PS_DATA1           (0x1 << 17)	/* R  */
+#define MSDC_PS_CMD             (0x1 << 24)	/* R  */
+#define MSDC_PS_WP              (0x1 << 31)	/* R  */
+
+/* MSDC_INT mask */
+#define MSDC_INT_MMCIRQ         (0x1 << 0)	/* W1C */
+#define MSDC_INT_CDSC           (0x1 << 1)	/* W1C */
+#define MSDC_INT_ACMDRDY        (0x1 << 3)	/* W1C */
+#define MSDC_INT_ACMDTMO        (0x1 << 4)	/* W1C */
+#define MSDC_INT_ACMDCRCERR     (0x1 << 5)	/* W1C */
+#define MSDC_INT_DMAQ_EMPTY     (0x1 << 6)	/* W1C */
+#define MSDC_INT_SDIOIRQ        (0x1 << 7)	/* W1C */
+#define MSDC_INT_CMDRDY         (0x1 << 8)	/* W1C */
+#define MSDC_INT_CMDTMO         (0x1 << 9)	/* W1C */
+#define MSDC_INT_RSPCRCERR      (0x1 << 10)	/* W1C */
+#define MSDC_INT_CSTA           (0x1 << 11)	/* R */
+#define MSDC_INT_XFER_COMPL     (0x1 << 12)	/* W1C */
+#define MSDC_INT_DXFER_DONE     (0x1 << 13)	/* W1C */
+#define MSDC_INT_DATTMO         (0x1 << 14)	/* W1C */
+#define MSDC_INT_DATCRCERR      (0x1 << 15)	/* W1C */
+#define MSDC_INT_ACMD19_DONE    (0x1 << 16)	/* W1C */
+#define MSDC_INT_DMA_BDCSERR    (0x1 << 17)	/* W1C */
+#define MSDC_INT_DMA_GPDCSERR   (0x1 << 18)	/* W1C */
+#define MSDC_INT_DMA_PROTECT    (0x1 << 19)	/* W1C */
+#define MSDC_INT_CMDQ           (0x1 << 28)	/* W1C */
+
+/* MSDC_INTEN mask */
+#define MSDC_INTEN_MMCIRQ       (0x1 << 0)	/* RW */
+#define MSDC_INTEN_CDSC         (0x1 << 1)	/* RW */
+#define MSDC_INTEN_ACMDRDY      (0x1 << 3)	/* RW */
+#define MSDC_INTEN_ACMDTMO      (0x1 << 4)	/* RW */
+#define MSDC_INTEN_ACMDCRCERR   (0x1 << 5)	/* RW */
+#define MSDC_INTEN_DMAQ_EMPTY   (0x1 << 6)	/* RW */
+#define MSDC_INTEN_SDIOIRQ      (0x1 << 7)	/* RW */
+#define MSDC_INTEN_CMDRDY       (0x1 << 8)	/* RW */
+#define MSDC_INTEN_CMDTMO       (0x1 << 9)	/* RW */
+#define MSDC_INTEN_RSPCRCERR    (0x1 << 10)	/* RW */
+#define MSDC_INTEN_CSTA         (0x1 << 11)	/* RW */
+#define MSDC_INTEN_XFER_COMPL   (0x1 << 12)	/* RW */
+#define MSDC_INTEN_DXFER_DONE   (0x1 << 13)	/* RW */
+#define MSDC_INTEN_DATTMO       (0x1 << 14)	/* RW */
+#define MSDC_INTEN_DATCRCERR    (0x1 << 15)	/* RW */
+#define MSDC_INTEN_ACMD19_DONE  (0x1 << 16)	/* RW */
+#define MSDC_INTEN_DMA_BDCSERR  (0x1 << 17)	/* RW */
+#define MSDC_INTEN_DMA_GPDCSERR (0x1 << 18)	/* RW */
+#define MSDC_INTEN_DMA_PROTECT  (0x1 << 19)	/* RW */
+
+/* MSDC_FIFOCS mask */
+#define MSDC_FIFOCS_RXCNT       (0xff << 0)	/* R */
+#define MSDC_FIFOCS_TXCNT       (0xff << 16)	/* R */
+#define MSDC_FIFOCS_CLR         (0x1 << 31)	/* RW */
+
+/* SDC_CFG mask */
+#define SDC_CFG_SDIOINTWKUP     (0x1 << 0)	/* RW */
+#define SDC_CFG_INSWKUP         (0x1 << 1)	/* RW */
+#define SDC_CFG_WRDTOC          (0x1fff  << 2)  /* RW */
+#define SDC_CFG_BUSWIDTH        (0x3 << 16)	/* RW */
+#define SDC_CFG_SDIO            (0x1 << 19)	/* RW */
+#define SDC_CFG_SDIOIDE         (0x1 << 20)	/* RW */
+#define SDC_CFG_INTATGAP        (0x1 << 21)	/* RW */
+#define SDC_CFG_DTOC            (0xff << 24)	/* RW */
+
+/* SDC_STS mask */
+#define SDC_STS_SDCBUSY         (0x1 << 0)	/* RW */
+#define SDC_STS_CMDBUSY         (0x1 << 1)	/* RW */
+#define SDC_STS_SWR_COMPL       (0x1 << 31)	/* RW */
+
+#define SDC_DAT1_IRQ_TRIGGER	(0x1 << 19)	/* RW */
+/* SDC_ADV_CFG0 mask */
+#define SDC_RX_ENHANCE_EN	(0x1 << 20)	/* RW */
+
+/* DMA_SA_H4BIT mask */
+#define DMA_ADDR_HIGH_4BIT      (0xf << 0)      /* RW */
+
+/* MSDC_DMA_CTRL mask */
+#define MSDC_DMA_CTRL_START     (0x1 << 0)	/* W */
+#define MSDC_DMA_CTRL_STOP      (0x1 << 1)	/* W */
+#define MSDC_DMA_CTRL_RESUME    (0x1 << 2)	/* W */
+#define MSDC_DMA_CTRL_MODE      (0x1 << 8)	/* RW */
+#define MSDC_DMA_CTRL_LASTBUF   (0x1 << 10)	/* RW */
+#define MSDC_DMA_CTRL_BRUSTSZ   (0x7 << 12)	/* RW */
+
+/* MSDC_DMA_CFG mask */
+#define MSDC_DMA_CFG_STS        (0x1 << 0)	/* R */
+#define MSDC_DMA_CFG_DECSEN     (0x1 << 1)	/* RW */
+#define MSDC_DMA_CFG_AHBHPROT2  (0x2 << 8)	/* RW */
+#define MSDC_DMA_CFG_ACTIVEEN   (0x2 << 12)	/* RW */
+#define MSDC_DMA_CFG_CS12B16B   (0x1 << 16)	/* RW */
+
+/* MSDC_PATCH_BIT mask */
+#define MSDC_PATCH_BIT_ODDSUPP    (0x1 <<  1)	/* RW */
+#define MSDC_INT_DAT_LATCH_CK_SEL (0x7 <<  7)
+#define MSDC_CKGEN_MSDC_DLY_SEL   (0x1f << 10)
+#define MSDC_PATCH_BIT_IODSSEL    (0x1 << 16)	/* RW */
+#define MSDC_PATCH_BIT_IOINTSEL   (0x1 << 17)	/* RW */
+#define MSDC_PATCH_BIT_BUSYDLY    (0xf << 18)	/* RW */
+#define MSDC_PATCH_BIT_WDOD       (0xf << 22)	/* RW */
+#define MSDC_PATCH_BIT_IDRTSEL    (0x1 << 26)	/* RW */
+#define MSDC_PATCH_BIT_CMDFSEL    (0x1 << 27)	/* RW */
+#define MSDC_PATCH_BIT_INTDLSEL   (0x1 << 28)	/* RW */
+#define MSDC_PATCH_BIT_SPCPUSH    (0x1 << 29)	/* RW */
+#define MSDC_PATCH_BIT_DECRCTMO   (0x1 << 30)	/* RW */
+
+#define MSDC_PATCH_BIT1_CMDTA     (0x7 << 3)    /* RW */
+#define MSDC_PB1_BUSY_CHECK_SEL   (0x1 << 7)    /* RW */
+#define MSDC_PATCH_BIT1_STOP_DLY  (0xf << 8)    /* RW */
+
+#define MSDC_PATCH_BIT2_CFGRESP   (0x1 << 15)   /* RW */
+#define MSDC_PATCH_BIT2_CFGCRCSTS (0x1 << 28)   /* RW */
+#define MSDC_PB2_SUPPORT_64G      (0x1 << 1)    /* RW */
+#define MSDC_PB2_RESPWAIT         (0x3 << 2)    /* RW */
+#define MSDC_PB2_RESPSTSENSEL     (0x7 << 16)   /* RW */
+#define MSDC_PB2_CRCSTSENSEL      (0x7 << 29)   /* RW */
+
+#define MSDC_PAD_TUNE_DATWRDLY	  (0x1f <<  0)	/* RW */
+#define MSDC_PAD_TUNE_DATRRDLY	  (0x1f <<  8)	/* RW */
+#define MSDC_PAD_TUNE_CMDRDLY	  (0x1f << 16)  /* RW */
+#define MSDC_PAD_TUNE_CMDRRDLY	  (0x1f << 22)	/* RW */
+#define MSDC_PAD_TUNE_CLKTDLY	  (0x1f << 27)  /* RW */
+#define MSDC_PAD_TUNE_RXDLYSEL	  (0x1 << 15)   /* RW */
+#define MSDC_PAD_TUNE_RD_SEL	  (0x1 << 13)   /* RW */
+#define MSDC_PAD_TUNE_CMD_SEL	  (0x1 << 21)   /* RW */
+
+#define PAD_DS_TUNE_DLY1	  (0x1f << 2)   /* RW */
+#define PAD_DS_TUNE_DLY2	  (0x1f << 7)   /* RW */
+#define PAD_DS_TUNE_DLY3	  (0x1f << 12)  /* RW */
+
+#define PAD_CMD_TUNE_RX_DLY3	  (0x1f << 1)  /* RW */
+
+#define EMMC50_CFG_PADCMD_LATCHCK (0x1 << 0)   /* RW */
+#define EMMC50_CFG_CRCSTS_EDGE    (0x1 << 3)   /* RW */
+#define EMMC50_CFG_CFCSTS_SEL     (0x1 << 4)   /* RW */
+
+#define EMMC50_CFG3_OUTS_WR       (0x1f << 0)  /* RW */
+
+#define SDC_FIFO_CFG_WRVALIDSEL   (0x1 << 24)  /* RW */
+#define SDC_FIFO_CFG_RDVALIDSEL   (0x1 << 25)  /* RW */
+
+/* EMMC_TOP_CONTROL mask */
+#define PAD_RXDLY_SEL           (0x1 << 0)      /* RW */
+#define DELAY_EN                (0x1 << 1)      /* RW */
+#define PAD_DAT_RD_RXDLY2       (0x1f << 2)     /* RW */
+#define PAD_DAT_RD_RXDLY        (0x1f << 7)     /* RW */
+#define PAD_DAT_RD_RXDLY2_SEL   (0x1 << 12)     /* RW */
+#define PAD_DAT_RD_RXDLY_SEL    (0x1 << 13)     /* RW */
+#define DATA_K_VALUE_SEL        (0x1 << 14)     /* RW */
+#define SDC_RX_ENH_EN           (0x1 << 15)     /* TW */
+
+/* EMMC_TOP_CMD mask */
+#define PAD_CMD_RXDLY2          (0x1f << 0)     /* RW */
+#define PAD_CMD_RXDLY           (0x1f << 5)     /* RW */
+#define PAD_CMD_RD_RXDLY2_SEL   (0x1 << 10)     /* RW */
+#define PAD_CMD_RD_RXDLY_SEL    (0x1 << 11)     /* RW */
+#define PAD_CMD_TX_DLY          (0x1f << 12)    /* RW */
+
+#define REQ_CMD_EIO  (0x1 << 0)
+#define REQ_CMD_TMO  (0x1 << 1)
+#define REQ_DAT_ERR  (0x1 << 2)
+#define REQ_STOP_EIO (0x1 << 3)
+#define REQ_STOP_TMO (0x1 << 4)
+#define REQ_CMD_BUSY (0x1 << 5)
+
+#define MSDC_PREPARE_FLAG (0x1 << 0)
+#define MSDC_ASYNC_FLAG (0x1 << 1)
+#define MSDC_MMAP_FLAG (0x1 << 2)
+
+#define MTK_MMC_AUTOSUSPEND_DELAY	50
+#define CMD_TIMEOUT         (HZ/10 * 5)	/* 100ms x5 */
+#define DAT_TIMEOUT         (HZ    * 5)	/* 1000ms x5 */
+
+#define DEFAULT_DEBOUNCE	(8)	/* 8 cycles CD debounce */
+
+#define PAD_DELAY_MAX	32 /* PAD delay cells */
+/*--------------------------------------------------------------------------*/
+/* Descriptor Structure                                                     */
+/*--------------------------------------------------------------------------*/
+struct mt_gpdma_desc {
+	u32 gpd_info;
+#define GPDMA_DESC_HWO		(0x1 << 0)
+#define GPDMA_DESC_BDP		(0x1 << 1)
+#define GPDMA_DESC_CHECKSUM	(0xff << 8) /* bit8 ~ bit15 */
+#define GPDMA_DESC_INT		(0x1 << 16)
+#define GPDMA_DESC_NEXT_H4	(0xf << 24)
+#define GPDMA_DESC_PTR_H4	(0xf << 28)
+	u32 next;
+	u32 ptr;
+	u32 gpd_data_len;
+#define GPDMA_DESC_BUFLEN	(0xffff) /* bit0 ~ bit15 */
+#define GPDMA_DESC_EXTLEN	(0xff << 16) /* bit16 ~ bit23 */
+	u32 arg;
+	u32 blknum;
+	u32 cmd;
+};
+
+struct mt_bdma_desc {
+	u32 bd_info;
+#define BDMA_DESC_EOL		(0x1 << 0)
+#define BDMA_DESC_CHECKSUM	(0xff << 8) /* bit8 ~ bit15 */
+#define BDMA_DESC_BLKPAD	(0x1 << 17)
+#define BDMA_DESC_DWPAD		(0x1 << 18)
+#define BDMA_DESC_NEXT_H4	(0xf << 24)
+#define BDMA_DESC_PTR_H4	(0xf << 28)
+	u32 next;
+	u32 ptr;
+	u32 bd_data_len;
+#define BDMA_DESC_BUFLEN	(0xffff) /* bit0 ~ bit15 */
+#define BDMA_DESC_BUFLEN_EXT	(0xffffff) /* bit0 ~ bit23 */
+};
+
+struct msdc_dma {
+	struct scatterlist *sg;	/* I/O scatter list */
+	struct mt_gpdma_desc *gpd;		/* pointer to gpd array */
+	struct mt_bdma_desc *bd;		/* pointer to bd array */
+	dma_addr_t gpd_addr;	/* the physical address of gpd array */
+	dma_addr_t bd_addr;	/* the physical address of bd array */
+};
+
+struct msdc_save_para {
+	u32 msdc_cfg;
+	u32 iocon;
+	u32 sdc_cfg;
+	u32 pad_tune;
+	u32 patch_bit0;
+	u32 patch_bit1;
+	u32 patch_bit2;
+	u32 pad_ds_tune;
+	u32 pad_cmd_tune;
+	u32 emmc50_cfg0;
+	u32 emmc50_cfg3;
+	u32 sdc_fifo_cfg;
+	u32 emmc_top_control;
+	u32 emmc_top_cmd;
+	u32 emmc50_pad_ds_tune;
+};
+
+struct mtk_mmc_compatible {
+	u8 clk_div_bits;
+	bool recheck_sdio_irq;
+	bool hs400_tune; /* only used for MT8173 */
+	u32 pad_tune_reg;
+	bool async_fifo;
+	bool data_tune;
+	bool busy_check;
+	bool stop_clk_fix;
+	bool enhance_rx;
+	bool support_64g;
+	bool use_internal_cd;
+};
+
+struct msdc_tune_para {
+	u32 iocon;
+	u32 pad_tune;
+	u32 pad_cmd_tune;
+	u32 emmc_top_control;
+	u32 emmc_top_cmd;
+};
+
+struct msdc_delay_phase {
+	u8 maxlen;
+	u8 start;
+	u8 final_phase;
+};
+
+struct msdc_host {
+	struct device *dev;
+	const struct mtk_mmc_compatible *dev_comp;
+	int cmd_rsp;
+
+	spinlock_t lock;
+	struct mmc_request *mrq;
+	struct mmc_command *cmd;
+	struct mmc_data *data;
+	int error;
+
+	void __iomem *base;		/* host base address */
+	void __iomem *top_base;		/* host top register base address */
+
+	struct msdc_dma dma;	/* dma channel */
+	u64 dma_mask;
+
+	u32 timeout_ns;		/* data timeout ns */
+	u32 timeout_clks;	/* data timeout clks */
+
+	struct pinctrl *pinctrl;
+	struct pinctrl_state *pins_default;
+	struct pinctrl_state *pins_uhs;
+	struct delayed_work req_timeout;
+	int irq;		/* host interrupt */
+	struct reset_control *reset;
+
+	struct clk *src_clk;	/* msdc source clock */
+	struct clk *h_clk;      /* msdc h_clk */
+	struct clk *bus_clk;	/* bus clock which used to access register */
+	struct clk *src_clk_cg; /* msdc source clock control gate */
+	u32 mclk;		/* mmc subsystem clock frequency */
+	u32 src_clk_freq;	/* source clock frequency */
+	unsigned char timing;
+	bool vqmmc_enabled;
+	u32 latch_ck;
+	u32 hs400_ds_delay;
+	u32 hs200_cmd_int_delay; /* cmd internal delay for HS200/SDR104 */
+	u32 hs400_cmd_int_delay; /* cmd internal delay for HS400 */
+	bool hs400_cmd_resp_sel_rising;
+				 /* cmd response sample selection for HS400 */
+	bool hs400_mode;	/* current eMMC will run at hs400 mode */
+	bool internal_cd;	/* Use internal card-detect logic */
+	bool cqhci;		/* support eMMC hw cmdq */
+	struct msdc_save_para save_para; /* used when gate HCLK */
+	struct msdc_tune_para def_tune_para; /* default tune setting */
+	struct msdc_tune_para saved_tune_para; /* tune result of CMD21/CMD19 */
+	struct cqhci_host *cq_host;
+};
+
+static const struct mtk_mmc_compatible mt8135_compat = {
+	.clk_div_bits = 8,
+	.recheck_sdio_irq = true,
+	.hs400_tune = false,
+	.pad_tune_reg = MSDC_PAD_TUNE,
+	.async_fifo = false,
+	.data_tune = false,
+	.busy_check = false,
+	.stop_clk_fix = false,
+	.enhance_rx = false,
+	.support_64g = false,
+};
+
+static const struct mtk_mmc_compatible mt8173_compat = {
+	.clk_div_bits = 8,
+	.recheck_sdio_irq = true,
+	.hs400_tune = true,
+	.pad_tune_reg = MSDC_PAD_TUNE,
+	.async_fifo = false,
+	.data_tune = false,
+	.busy_check = false,
+	.stop_clk_fix = false,
+	.enhance_rx = false,
+	.support_64g = false,
+};
+
+static const struct mtk_mmc_compatible mt8183_compat = {
+	.clk_div_bits = 12,
+	.recheck_sdio_irq = false,
+	.hs400_tune = false,
+	.pad_tune_reg = MSDC_PAD_TUNE0,
+	.async_fifo = true,
+	.data_tune = true,
+	.busy_check = true,
+	.stop_clk_fix = true,
+	.enhance_rx = true,
+	.support_64g = true,
+};
+
+static const struct mtk_mmc_compatible mt2701_compat = {
+	.clk_div_bits = 12,
+	.recheck_sdio_irq = true,
+	.hs400_tune = false,
+	.pad_tune_reg = MSDC_PAD_TUNE0,
+	.async_fifo = true,
+	.data_tune = true,
+	.busy_check = false,
+	.stop_clk_fix = false,
+	.enhance_rx = false,
+	.support_64g = false,
+};
+
+static const struct mtk_mmc_compatible mt2712_compat = {
+	.clk_div_bits = 12,
+	.recheck_sdio_irq = false,
+	.hs400_tune = false,
+	.pad_tune_reg = MSDC_PAD_TUNE0,
+	.async_fifo = true,
+	.data_tune = true,
+	.busy_check = true,
+	.stop_clk_fix = true,
+	.enhance_rx = true,
+	.support_64g = true,
+};
+
+static const struct mtk_mmc_compatible mt7622_compat = {
+	.clk_div_bits = 12,
+	.recheck_sdio_irq = true,
+	.hs400_tune = false,
+	.pad_tune_reg = MSDC_PAD_TUNE0,
+	.async_fifo = true,
+	.data_tune = true,
+	.busy_check = true,
+	.stop_clk_fix = true,
+	.enhance_rx = true,
+	.support_64g = false,
+};
+
+static const struct mtk_mmc_compatible mt8516_compat = {
+	.clk_div_bits = 12,
+	.recheck_sdio_irq = true,
+	.hs400_tune = false,
+	.pad_tune_reg = MSDC_PAD_TUNE0,
+	.async_fifo = true,
+	.data_tune = true,
+	.busy_check = true,
+	.stop_clk_fix = true,
+};
+
+static const struct mtk_mmc_compatible mt7620_compat = {
+	.clk_div_bits = 8,
+	.recheck_sdio_irq = true,
+	.hs400_tune = false,
+	.pad_tune_reg = MSDC_PAD_TUNE,
+	.async_fifo = false,
+	.data_tune = false,
+	.busy_check = false,
+	.stop_clk_fix = false,
+	.enhance_rx = false,
+	.use_internal_cd = true,
+};
+
+static const struct mtk_mmc_compatible mt6779_compat = {
+	.clk_div_bits = 12,
+	.recheck_sdio_irq = false,
+	.hs400_tune = false,
+	.pad_tune_reg = MSDC_PAD_TUNE0,
+	.async_fifo = true,
+	.data_tune = true,
+	.busy_check = true,
+	.stop_clk_fix = true,
+	.enhance_rx = true,
+	.support_64g = true,
+};
+
+static const struct of_device_id msdc_of_ids[] = {
+	{ .compatible = "mediatek,mt8135-mmc", .data = &mt8135_compat},
+	{ .compatible = "mediatek,mt8173-mmc", .data = &mt8173_compat},
+	{ .compatible = "mediatek,mt8183-mmc", .data = &mt8183_compat},
+	{ .compatible = "mediatek,mt2701-mmc", .data = &mt2701_compat},
+	{ .compatible = "mediatek,mt2712-mmc", .data = &mt2712_compat},
+	{ .compatible = "mediatek,mt7622-mmc", .data = &mt7622_compat},
+	{ .compatible = "mediatek,mt8516-mmc", .data = &mt8516_compat},
+	{ .compatible = "mediatek,mt7620-mmc", .data = &mt7620_compat},
+	{ .compatible = "mediatek,mt6779-mmc", .data = &mt6779_compat},
+	{}
+};
+MODULE_DEVICE_TABLE(of, msdc_of_ids);
+
+static void sdr_set_bits(void __iomem *reg, u32 bs)
+{
+	u32 val = readl(reg);
+
+	val |= bs;
+	writel(val, reg);
+}
+
+static void sdr_clr_bits(void __iomem *reg, u32 bs)
+{
+	u32 val = readl(reg);
+
+	val &= ~bs;
+	writel(val, reg);
+}
+
+static void sdr_set_field(void __iomem *reg, u32 field, u32 val)
+{
+	unsigned int tv = readl(reg);
+
+	tv &= ~field;
+	tv |= ((val) << (ffs((unsigned int)field) - 1));
+	writel(tv, reg);
+}
+
+static void sdr_get_field(void __iomem *reg, u32 field, u32 *val)
+{
+	unsigned int tv = readl(reg);
+
+	*val = ((tv & field) >> (ffs((unsigned int)field) - 1));
+}
+
+static void msdc_reset_hw(struct msdc_host *host)
+{
+	u32 val;
+
+	sdr_set_bits(host->base + MSDC_CFG, MSDC_CFG_RST);
+	while (readl(host->base + MSDC_CFG) & MSDC_CFG_RST)
+		cpu_relax();
+
+	sdr_set_bits(host->base + MSDC_FIFOCS, MSDC_FIFOCS_CLR);
+	while (readl(host->base + MSDC_FIFOCS) & MSDC_FIFOCS_CLR)
+		cpu_relax();
+
+	val = readl(host->base + MSDC_INT);
+	writel(val, host->base + MSDC_INT);
+}
+
+static void msdc_cmd_next(struct msdc_host *host,
+		struct mmc_request *mrq, struct mmc_command *cmd);
+static void __msdc_enable_sdio_irq(struct msdc_host *host, int enb);
+
+static const u32 cmd_ints_mask = MSDC_INTEN_CMDRDY | MSDC_INTEN_RSPCRCERR |
+			MSDC_INTEN_CMDTMO | MSDC_INTEN_ACMDRDY |
+			MSDC_INTEN_ACMDCRCERR | MSDC_INTEN_ACMDTMO;
+static const u32 data_ints_mask = MSDC_INTEN_XFER_COMPL | MSDC_INTEN_DATTMO |
+			MSDC_INTEN_DATCRCERR | MSDC_INTEN_DMA_BDCSERR |
+			MSDC_INTEN_DMA_GPDCSERR | MSDC_INTEN_DMA_PROTECT;
+
+static u8 msdc_dma_calcs(u8 *buf, u32 len)
+{
+	u32 i, sum = 0;
+
+	for (i = 0; i < len; i++)
+		sum += buf[i];
+	return 0xff - (u8) sum;
+}
+
+static inline void msdc_dma_setup(struct msdc_host *host, struct msdc_dma *dma,
+		struct mmc_data *data)
+{
+	unsigned int j, dma_len;
+	dma_addr_t dma_address;
+	u32 dma_ctrl;
+	struct scatterlist *sg;
+	struct mt_gpdma_desc *gpd;
+	struct mt_bdma_desc *bd;
+
+	sg = data->sg;
+
+	gpd = dma->gpd;
+	bd = dma->bd;
+
+	/* modify gpd */
+	gpd->gpd_info |= GPDMA_DESC_HWO;
+	gpd->gpd_info |= GPDMA_DESC_BDP;
+	/* need to clear first. use these bits to calc checksum */
+	gpd->gpd_info &= ~GPDMA_DESC_CHECKSUM;
+	gpd->gpd_info |= msdc_dma_calcs((u8 *) gpd, 16) << 8;
+
+	/* modify bd */
+	for_each_sg(data->sg, sg, data->sg_count, j) {
+		dma_address = sg_dma_address(sg);
+		dma_len = sg_dma_len(sg);
+
+		/* init bd */
+		bd[j].bd_info &= ~BDMA_DESC_BLKPAD;
+		bd[j].bd_info &= ~BDMA_DESC_DWPAD;
+		bd[j].ptr = lower_32_bits(dma_address);
+		if (host->dev_comp->support_64g) {
+			bd[j].bd_info &= ~BDMA_DESC_PTR_H4;
+			bd[j].bd_info |= (upper_32_bits(dma_address) & 0xf)
+					 << 28;
+		}
+
+		if (host->dev_comp->support_64g) {
+			bd[j].bd_data_len &= ~BDMA_DESC_BUFLEN_EXT;
+			bd[j].bd_data_len |= (dma_len & BDMA_DESC_BUFLEN_EXT);
+		} else {
+			bd[j].bd_data_len &= ~BDMA_DESC_BUFLEN;
+			bd[j].bd_data_len |= (dma_len & BDMA_DESC_BUFLEN);
+		}
+
+		if (j == data->sg_count - 1) /* the last bd */
+			bd[j].bd_info |= BDMA_DESC_EOL;
+		else
+			bd[j].bd_info &= ~BDMA_DESC_EOL;
+
+		/* checksume need to clear first */
+		bd[j].bd_info &= ~BDMA_DESC_CHECKSUM;
+		bd[j].bd_info |= msdc_dma_calcs((u8 *)(&bd[j]), 16) << 8;
+	}
+
+	sdr_set_field(host->base + MSDC_DMA_CFG, MSDC_DMA_CFG_DECSEN, 1);
+	dma_ctrl = readl_relaxed(host->base + MSDC_DMA_CTRL);
+	dma_ctrl &= ~(MSDC_DMA_CTRL_BRUSTSZ | MSDC_DMA_CTRL_MODE);
+	dma_ctrl |= (MSDC_BURST_64B << 12 | 1 << 8);
+	writel_relaxed(dma_ctrl, host->base + MSDC_DMA_CTRL);
+	if (host->dev_comp->support_64g)
+		sdr_set_field(host->base + DMA_SA_H4BIT, DMA_ADDR_HIGH_4BIT,
+			      upper_32_bits(dma->gpd_addr) & 0xf);
+	writel(lower_32_bits(dma->gpd_addr), host->base + MSDC_DMA_SA);
+}
+
+static void msdc_prepare_data(struct msdc_host *host, struct mmc_request *mrq)
+{
+	struct mmc_data *data = mrq->data;
+
+	if (!(data->host_cookie & MSDC_PREPARE_FLAG)) {
+		data->host_cookie |= MSDC_PREPARE_FLAG;
+		data->sg_count = dma_map_sg(host->dev, data->sg, data->sg_len,
+					    mmc_get_dma_dir(data));
+	}
+}
+
+static void msdc_unprepare_data(struct msdc_host *host, struct mmc_request *mrq)
+{
+	struct mmc_data *data = mrq->data;
+
+	if (data->host_cookie & MSDC_ASYNC_FLAG)
+		return;
+
+	if (data->host_cookie & MSDC_PREPARE_FLAG) {
+		dma_unmap_sg(host->dev, data->sg, data->sg_len,
+			     mmc_get_dma_dir(data));
+		data->host_cookie &= ~MSDC_PREPARE_FLAG;
+	}
+}
+
+static u64 msdc_timeout_cal(struct msdc_host *host, u64 ns, u64 clks)
+{
+	struct mmc_host *mmc = mmc_from_priv(host);
+	u64 timeout, clk_ns;
+	u32 mode = 0;
+
+	if (mmc->actual_clock == 0) {
+		timeout = 0;
+	} else {
+		clk_ns  = 1000000000ULL;
+		do_div(clk_ns, mmc->actual_clock);
+		timeout = ns + clk_ns - 1;
+		do_div(timeout, clk_ns);
+		timeout += clks;
+		/* in 1048576 sclk cycle unit */
+		timeout = DIV_ROUND_UP(timeout, (0x1 << 20));
+		if (host->dev_comp->clk_div_bits == 8)
+			sdr_get_field(host->base + MSDC_CFG,
+				      MSDC_CFG_CKMOD, &mode);
+		else
+			sdr_get_field(host->base + MSDC_CFG,
+				      MSDC_CFG_CKMOD_EXTRA, &mode);
+		/*DDR mode will double the clk cycles for data timeout */
+		timeout = mode >= 2 ? timeout * 2 : timeout;
+		timeout = timeout > 1 ? timeout - 1 : 0;
+	}
+	return timeout;
+}
+
+/* clock control primitives */
+static void msdc_set_timeout(struct msdc_host *host, u64 ns, u64 clks)
+{
+	u64 timeout;
+
+	host->timeout_ns = ns;
+	host->timeout_clks = clks;
+
+	timeout = msdc_timeout_cal(host, ns, clks);
+	sdr_set_field(host->base + SDC_CFG, SDC_CFG_DTOC,
+		      (u32)(timeout > 255 ? 255 : timeout));
+}
+
+static void msdc_set_busy_timeout(struct msdc_host *host, u64 ns, u64 clks)
+{
+	u64 timeout;
+
+	timeout = msdc_timeout_cal(host, ns, clks);
+	sdr_set_field(host->base + SDC_CFG, SDC_CFG_WRDTOC,
+		      (u32)(timeout > 8191 ? 8191 : timeout));
+}
+
+static void msdc_gate_clock(struct msdc_host *host)
+{
+	clk_disable_unprepare(host->src_clk_cg);
+	clk_disable_unprepare(host->src_clk);
+	clk_disable_unprepare(host->bus_clk);
+	clk_disable_unprepare(host->h_clk);
+}
+
+static void msdc_ungate_clock(struct msdc_host *host)
+{
+	clk_prepare_enable(host->h_clk);
+	clk_prepare_enable(host->bus_clk);
+	clk_prepare_enable(host->src_clk);
+	clk_prepare_enable(host->src_clk_cg);
+	while (!(readl(host->base + MSDC_CFG) & MSDC_CFG_CKSTB))
+		cpu_relax();
+}
+
+static void msdc_set_mclk(struct msdc_host *host, unsigned char timing, u32 hz)
+{
+	struct mmc_host *mmc = mmc_from_priv(host);
+	u32 mode;
+	u32 flags;
+	u32 div;
+	u32 sclk;
+	u32 tune_reg = host->dev_comp->pad_tune_reg;
+
+	if (!hz) {
+		dev_dbg(host->dev, "set mclk to 0\n");
+		host->mclk = 0;
+		mmc->actual_clock = 0;
+		sdr_clr_bits(host->base + MSDC_CFG, MSDC_CFG_CKPDN);
+		return;
+	}
+
+	flags = readl(host->base + MSDC_INTEN);
+	sdr_clr_bits(host->base + MSDC_INTEN, flags);
+	if (host->dev_comp->clk_div_bits == 8)
+		sdr_clr_bits(host->base + MSDC_CFG, MSDC_CFG_HS400_CK_MODE);
+	else
+		sdr_clr_bits(host->base + MSDC_CFG,
+			     MSDC_CFG_HS400_CK_MODE_EXTRA);
+	if (timing == MMC_TIMING_UHS_DDR50 ||
+	    timing == MMC_TIMING_MMC_DDR52 ||
+	    timing == MMC_TIMING_MMC_HS400) {
+		if (timing == MMC_TIMING_MMC_HS400)
+			mode = 0x3;
+		else
+			mode = 0x2; /* ddr mode and use divisor */
+
+		if (hz >= (host->src_clk_freq >> 2)) {
+			div = 0; /* mean div = 1/4 */
+			sclk = host->src_clk_freq >> 2; /* sclk = clk / 4 */
+		} else {
+			div = (host->src_clk_freq + ((hz << 2) - 1)) / (hz << 2);
+			sclk = (host->src_clk_freq >> 2) / div;
+			div = (div >> 1);
+		}
+
+		if (timing == MMC_TIMING_MMC_HS400 &&
+		    hz >= (host->src_clk_freq >> 1)) {
+			if (host->dev_comp->clk_div_bits == 8)
+				sdr_set_bits(host->base + MSDC_CFG,
+					     MSDC_CFG_HS400_CK_MODE);
+			else
+				sdr_set_bits(host->base + MSDC_CFG,
+					     MSDC_CFG_HS400_CK_MODE_EXTRA);
+			sclk = host->src_clk_freq >> 1;
+			div = 0; /* div is ignore when bit18 is set */
+		}
+	} else if (hz >= host->src_clk_freq) {
+		mode = 0x1; /* no divisor */
+		div = 0;
+		sclk = host->src_clk_freq;
+	} else {
+		mode = 0x0; /* use divisor */
+		if (hz >= (host->src_clk_freq >> 1)) {
+			div = 0; /* mean div = 1/2 */
+			sclk = host->src_clk_freq >> 1; /* sclk = clk / 2 */
+		} else {
+			div = (host->src_clk_freq + ((hz << 2) - 1)) / (hz << 2);
+			sclk = (host->src_clk_freq >> 2) / div;
+		}
+	}
+	sdr_clr_bits(host->base + MSDC_CFG, MSDC_CFG_CKPDN);
+	/*
+	 * As src_clk/HCLK use the same bit to gate/ungate,
+	 * So if want to only gate src_clk, need gate its parent(mux).
+	 */
+	if (host->src_clk_cg)
+		clk_disable_unprepare(host->src_clk_cg);
+	else
+		clk_disable_unprepare(clk_get_parent(host->src_clk));
+	if (host->dev_comp->clk_div_bits == 8)
+		sdr_set_field(host->base + MSDC_CFG,
+			      MSDC_CFG_CKMOD | MSDC_CFG_CKDIV,
+			      (mode << 8) | div);
+	else
+		sdr_set_field(host->base + MSDC_CFG,
+			      MSDC_CFG_CKMOD_EXTRA | MSDC_CFG_CKDIV_EXTRA,
+			      (mode << 12) | div);
+	if (host->src_clk_cg)
+		clk_prepare_enable(host->src_clk_cg);
+	else
+		clk_prepare_enable(clk_get_parent(host->src_clk));
+
+	while (!(readl(host->base + MSDC_CFG) & MSDC_CFG_CKSTB))
+		cpu_relax();
+	sdr_set_bits(host->base + MSDC_CFG, MSDC_CFG_CKPDN);
+	mmc->actual_clock = sclk;
+	host->mclk = hz;
+	host->timing = timing;
+	/* need because clk changed. */
+	msdc_set_timeout(host, host->timeout_ns, host->timeout_clks);
+	sdr_set_bits(host->base + MSDC_INTEN, flags);
+
+	/*
+	 * mmc_select_hs400() will drop to 50Mhz and High speed mode,
+	 * tune result of hs200/200Mhz is not suitable for 50Mhz
+	 */
+	if (mmc->actual_clock <= 52000000) {
+		writel(host->def_tune_para.iocon, host->base + MSDC_IOCON);
+		if (host->top_base) {
+			writel(host->def_tune_para.emmc_top_control,
+			       host->top_base + EMMC_TOP_CONTROL);
+			writel(host->def_tune_para.emmc_top_cmd,
+			       host->top_base + EMMC_TOP_CMD);
+		} else {
+			writel(host->def_tune_para.pad_tune,
+			       host->base + tune_reg);
+		}
+	} else {
+		writel(host->saved_tune_para.iocon, host->base + MSDC_IOCON);
+		writel(host->saved_tune_para.pad_cmd_tune,
+		       host->base + PAD_CMD_TUNE);
+		if (host->top_base) {
+			writel(host->saved_tune_para.emmc_top_control,
+			       host->top_base + EMMC_TOP_CONTROL);
+			writel(host->saved_tune_para.emmc_top_cmd,
+			       host->top_base + EMMC_TOP_CMD);
+		} else {
+			writel(host->saved_tune_para.pad_tune,
+			       host->base + tune_reg);
+		}
+	}
+
+	if (timing == MMC_TIMING_MMC_HS400 &&
+	    host->dev_comp->hs400_tune)
+		sdr_set_field(host->base + tune_reg,
+			      MSDC_PAD_TUNE_CMDRRDLY,
+			      host->hs400_cmd_int_delay);
+	dev_dbg(host->dev, "sclk: %d, timing: %d\n", mmc->actual_clock,
+		timing);
+}
+
+static inline u32 msdc_cmd_find_resp(struct msdc_host *host,
+		struct mmc_request *mrq, struct mmc_command *cmd)
+{
+	u32 resp;
+
+	switch (mmc_resp_type(cmd)) {
+		/* Actually, R1, R5, R6, R7 are the same */
+	case MMC_RSP_R1:
+		resp = 0x1;
+		break;
+	case MMC_RSP_R1B:
+		resp = 0x7;
+		break;
+	case MMC_RSP_R2:
+		resp = 0x2;
+		break;
+	case MMC_RSP_R3:
+		resp = 0x3;
+		break;
+	case MMC_RSP_NONE:
+	default:
+		resp = 0x0;
+		break;
+	}
+
+	return resp;
+}
+
+static inline u32 msdc_cmd_prepare_raw_cmd(struct msdc_host *host,
+		struct mmc_request *mrq, struct mmc_command *cmd)
+{
+	struct mmc_host *mmc = mmc_from_priv(host);
+	/* rawcmd :
+	 * vol_swt << 30 | auto_cmd << 28 | blklen << 16 | go_irq << 15 |
+	 * stop << 14 | rw << 13 | dtype << 11 | rsptyp << 7 | brk << 6 | opcode
+	 */
+	u32 opcode = cmd->opcode;
+	u32 resp = msdc_cmd_find_resp(host, mrq, cmd);
+	u32 rawcmd = (opcode & 0x3f) | ((resp & 0x7) << 7);
+
+	host->cmd_rsp = resp;
+
+	if ((opcode == SD_IO_RW_DIRECT && cmd->flags == (unsigned int) -1) ||
+	    opcode == MMC_STOP_TRANSMISSION)
+		rawcmd |= (0x1 << 14);
+	else if (opcode == SD_SWITCH_VOLTAGE)
+		rawcmd |= (0x1 << 30);
+	else if (opcode == SD_APP_SEND_SCR ||
+		 opcode == SD_APP_SEND_NUM_WR_BLKS ||
+		 (opcode == SD_SWITCH && mmc_cmd_type(cmd) == MMC_CMD_ADTC) ||
+		 (opcode == SD_APP_SD_STATUS && mmc_cmd_type(cmd) == MMC_CMD_ADTC) ||
+		 (opcode == MMC_SEND_EXT_CSD && mmc_cmd_type(cmd) == MMC_CMD_ADTC))
+		rawcmd |= (0x1 << 11);
+
+	if (cmd->data) {
+		struct mmc_data *data = cmd->data;
+
+		if (mmc_op_multi(opcode)) {
+			if (mmc_card_mmc(mmc->card) && mrq->sbc &&
+			    !(mrq->sbc->arg & 0xFFFF0000))
+				rawcmd |= 0x2 << 28; /* AutoCMD23 */
+		}
+
+		rawcmd |= ((data->blksz & 0xFFF) << 16);
+		if (data->flags & MMC_DATA_WRITE)
+			rawcmd |= (0x1 << 13);
+		if (data->blocks > 1)
+			rawcmd |= (0x2 << 11);
+		else
+			rawcmd |= (0x1 << 11);
+		/* Always use dma mode */
+		sdr_clr_bits(host->base + MSDC_CFG, MSDC_CFG_PIO);
+
+		if (host->timeout_ns != data->timeout_ns ||
+		    host->timeout_clks != data->timeout_clks)
+			msdc_set_timeout(host, data->timeout_ns,
+					data->timeout_clks);
+
+		writel(data->blocks, host->base + SDC_BLK_NUM);
+	}
+	return rawcmd;
+}
+
+static void msdc_start_data(struct msdc_host *host, struct mmc_request *mrq,
+			    struct mmc_command *cmd, struct mmc_data *data)
+{
+	bool read;
+
+	WARN_ON(host->data);
+	host->data = data;
+	read = data->flags & MMC_DATA_READ;
+
+	mod_delayed_work(system_wq, &host->req_timeout, DAT_TIMEOUT);
+	msdc_dma_setup(host, &host->dma, data);
+	sdr_set_bits(host->base + MSDC_INTEN, data_ints_mask);
+	sdr_set_field(host->base + MSDC_DMA_CTRL, MSDC_DMA_CTRL_START, 1);
+	dev_dbg(host->dev, "DMA start\n");
+	dev_dbg(host->dev, "%s: cmd=%d DMA data: %d blocks; read=%d\n",
+			__func__, cmd->opcode, data->blocks, read);
+}
+
+static int msdc_auto_cmd_done(struct msdc_host *host, int events,
+		struct mmc_command *cmd)
+{
+	u32 *rsp = cmd->resp;
+
+	rsp[0] = readl(host->base + SDC_ACMD_RESP);
+
+	if (events & MSDC_INT_ACMDRDY) {
+		cmd->error = 0;
+	} else {
+		msdc_reset_hw(host);
+		if (events & MSDC_INT_ACMDCRCERR) {
+			cmd->error = -EILSEQ;
+			host->error |= REQ_STOP_EIO;
+		} else if (events & MSDC_INT_ACMDTMO) {
+			cmd->error = -ETIMEDOUT;
+			host->error |= REQ_STOP_TMO;
+		}
+		dev_err(host->dev,
+			"%s: AUTO_CMD%d arg=%08X; rsp %08X; cmd_error=%d\n",
+			__func__, cmd->opcode, cmd->arg, rsp[0], cmd->error);
+	}
+	return cmd->error;
+}
+
+/*
+ * msdc_recheck_sdio_irq - recheck whether the SDIO irq is lost
+ *
+ * Host controller may lost interrupt in some special case.
+ * Add SDIO irq recheck mechanism to make sure all interrupts
+ * can be processed immediately
+ */
+static void msdc_recheck_sdio_irq(struct msdc_host *host)
+{
+	struct mmc_host *mmc = mmc_from_priv(host);
+	u32 reg_int, reg_inten, reg_ps;
+
+	if (mmc->caps & MMC_CAP_SDIO_IRQ) {
+		reg_inten = readl(host->base + MSDC_INTEN);
+		if (reg_inten & MSDC_INTEN_SDIOIRQ) {
+			reg_int = readl(host->base + MSDC_INT);
+			reg_ps = readl(host->base + MSDC_PS);
+			if (!(reg_int & MSDC_INT_SDIOIRQ ||
+			      reg_ps & MSDC_PS_DATA1)) {
+				__msdc_enable_sdio_irq(host, 0);
+				sdio_signal_irq(mmc);
+			}
+		}
+	}
+}
+
+static void msdc_track_cmd_data(struct msdc_host *host,
+				struct mmc_command *cmd, struct mmc_data *data)
+{
+	if (host->error)
+		dev_dbg(host->dev, "%s: cmd=%d arg=%08X; host->error=0x%08X\n",
+			__func__, cmd->opcode, cmd->arg, host->error);
+}
+
+static void msdc_request_done(struct msdc_host *host, struct mmc_request *mrq)
+{
+	unsigned long flags;
+
+	/*
+	 * No need check the return value of cancel_delayed_work, as only ONE
+	 * path will go here!
+	 */
+	cancel_delayed_work(&host->req_timeout);
+
+	spin_lock_irqsave(&host->lock, flags);
+	host->mrq = NULL;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	msdc_track_cmd_data(host, mrq->cmd, mrq->data);
+	if (mrq->data)
+		msdc_unprepare_data(host, mrq);
+	if (host->error)
+		msdc_reset_hw(host);
+	mmc_request_done(mmc_from_priv(host), mrq);
+	if (host->dev_comp->recheck_sdio_irq)
+		msdc_recheck_sdio_irq(host);
+}
+
+/* returns true if command is fully handled; returns false otherwise */
+static bool msdc_cmd_done(struct msdc_host *host, int events,
+			  struct mmc_request *mrq, struct mmc_command *cmd)
+{
+	bool done = false;
+	bool sbc_error;
+	unsigned long flags;
+	u32 *rsp;
+
+	if (mrq->sbc && cmd == mrq->cmd &&
+	    (events & (MSDC_INT_ACMDRDY | MSDC_INT_ACMDCRCERR
+				   | MSDC_INT_ACMDTMO)))
+		msdc_auto_cmd_done(host, events, mrq->sbc);
+
+	sbc_error = mrq->sbc && mrq->sbc->error;
+
+	if (!sbc_error && !(events & (MSDC_INT_CMDRDY
+					| MSDC_INT_RSPCRCERR
+					| MSDC_INT_CMDTMO)))
+		return done;
+
+	spin_lock_irqsave(&host->lock, flags);
+	done = !host->cmd;
+	host->cmd = NULL;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (done)
+		return true;
+	rsp = cmd->resp;
+
+	sdr_clr_bits(host->base + MSDC_INTEN, cmd_ints_mask);
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136) {
+			rsp[0] = readl(host->base + SDC_RESP3);
+			rsp[1] = readl(host->base + SDC_RESP2);
+			rsp[2] = readl(host->base + SDC_RESP1);
+			rsp[3] = readl(host->base + SDC_RESP0);
+		} else {
+			rsp[0] = readl(host->base + SDC_RESP0);
+		}
+	}
+
+	if (!sbc_error && !(events & MSDC_INT_CMDRDY)) {
+		if (events & MSDC_INT_CMDTMO ||
+		    (cmd->opcode != MMC_SEND_TUNING_BLOCK &&
+		     cmd->opcode != MMC_SEND_TUNING_BLOCK_HS200))
+			/*
+			 * should not clear fifo/interrupt as the tune data
+			 * may have alreay come when cmd19/cmd21 gets response
+			 * CRC error.
+			 */
+			msdc_reset_hw(host);
+		if (events & MSDC_INT_RSPCRCERR) {
+			cmd->error = -EILSEQ;
+			host->error |= REQ_CMD_EIO;
+		} else if (events & MSDC_INT_CMDTMO) {
+			cmd->error = -ETIMEDOUT;
+			host->error |= REQ_CMD_TMO;
+		}
+	}
+	if (cmd->error)
+		dev_dbg(host->dev,
+				"%s: cmd=%d arg=%08X; rsp %08X; cmd_error=%d\n",
+				__func__, cmd->opcode, cmd->arg, rsp[0],
+				cmd->error);
+
+	msdc_cmd_next(host, mrq, cmd);
+	return true;
+}
+
+/* It is the core layer's responsibility to ensure card status
+ * is correct before issue a request. but host design do below
+ * checks recommended.
+ */
+static inline bool msdc_cmd_is_ready(struct msdc_host *host,
+		struct mmc_request *mrq, struct mmc_command *cmd)
+{
+	/* The max busy time we can endure is 20ms */
+	unsigned long tmo = jiffies + msecs_to_jiffies(20);
+
+	while ((readl(host->base + SDC_STS) & SDC_STS_CMDBUSY) &&
+			time_before(jiffies, tmo))
+		cpu_relax();
+	if (readl(host->base + SDC_STS) & SDC_STS_CMDBUSY) {
+		dev_err(host->dev, "CMD bus busy detected\n");
+		host->error |= REQ_CMD_BUSY;
+		msdc_cmd_done(host, MSDC_INT_CMDTMO, mrq, cmd);
+		return false;
+	}
+
+	if (mmc_resp_type(cmd) == MMC_RSP_R1B || cmd->data) {
+		tmo = jiffies + msecs_to_jiffies(20);
+		/* R1B or with data, should check SDCBUSY */
+		while ((readl(host->base + SDC_STS) & SDC_STS_SDCBUSY) &&
+				time_before(jiffies, tmo))
+			cpu_relax();
+		if (readl(host->base + SDC_STS) & SDC_STS_SDCBUSY) {
+			dev_err(host->dev, "Controller busy detected\n");
+			host->error |= REQ_CMD_BUSY;
+			msdc_cmd_done(host, MSDC_INT_CMDTMO, mrq, cmd);
+			return false;
+		}
+	}
+	return true;
+}
+
+static void msdc_start_command(struct msdc_host *host,
+		struct mmc_request *mrq, struct mmc_command *cmd)
+{
+	u32 rawcmd;
+	unsigned long flags;
+
+	WARN_ON(host->cmd);
+	host->cmd = cmd;
+
+	mod_delayed_work(system_wq, &host->req_timeout, DAT_TIMEOUT);
+	if (!msdc_cmd_is_ready(host, mrq, cmd))
+		return;
+
+	if ((readl(host->base + MSDC_FIFOCS) & MSDC_FIFOCS_TXCNT) >> 16 ||
+	    readl(host->base + MSDC_FIFOCS) & MSDC_FIFOCS_RXCNT) {
+		dev_err(host->dev, "TX/RX FIFO non-empty before start of IO. Reset\n");
+		msdc_reset_hw(host);
+	}
+
+	cmd->error = 0;
+	rawcmd = msdc_cmd_prepare_raw_cmd(host, mrq, cmd);
+
+	spin_lock_irqsave(&host->lock, flags);
+	sdr_set_bits(host->base + MSDC_INTEN, cmd_ints_mask);
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	writel(cmd->arg, host->base + SDC_ARG);
+	writel(rawcmd, host->base + SDC_CMD);
+}
+
+static void msdc_cmd_next(struct msdc_host *host,
+		struct mmc_request *mrq, struct mmc_command *cmd)
+{
+	if ((cmd->error &&
+	    !(cmd->error == -EILSEQ &&
+	      (cmd->opcode == MMC_SEND_TUNING_BLOCK ||
+	       cmd->opcode == MMC_SEND_TUNING_BLOCK_HS200))) ||
+	    (mrq->sbc && mrq->sbc->error))
+		msdc_request_done(host, mrq);
+	else if (cmd == mrq->sbc)
+		msdc_start_command(host, mrq, mrq->cmd);
+	else if (!cmd->data)
+		msdc_request_done(host, mrq);
+	else
+		msdc_start_data(host, mrq, cmd, cmd->data);
+}
+
+static void msdc_ops_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+
+	host->error = 0;
+	WARN_ON(host->mrq);
+	host->mrq = mrq;
+
+	if (mrq->data)
+		msdc_prepare_data(host, mrq);
+
+	/* if SBC is required, we have HW option and SW option.
+	 * if HW option is enabled, and SBC does not have "special" flags,
+	 * use HW option,  otherwise use SW option
+	 */
+	if (mrq->sbc && (!mmc_card_mmc(mmc->card) ||
+	    (mrq->sbc->arg & 0xFFFF0000)))
+		msdc_start_command(host, mrq, mrq->sbc);
+	else
+		msdc_start_command(host, mrq, mrq->cmd);
+}
+
+static void msdc_pre_req(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+
+	if (!data)
+		return;
+
+	msdc_prepare_data(host, mrq);
+	data->host_cookie |= MSDC_ASYNC_FLAG;
+}
+
+static void msdc_post_req(struct mmc_host *mmc, struct mmc_request *mrq,
+		int err)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+	struct mmc_data *data;
+
+	data = mrq->data;
+	if (!data)
+		return;
+	if (data->host_cookie) {
+		data->host_cookie &= ~MSDC_ASYNC_FLAG;
+		msdc_unprepare_data(host, mrq);
+	}
+}
+
+static void msdc_data_xfer_next(struct msdc_host *host,
+				struct mmc_request *mrq, struct mmc_data *data)
+{
+	if (mmc_op_multi(mrq->cmd->opcode) && mrq->stop && !mrq->stop->error &&
+	    !mrq->sbc)
+		msdc_start_command(host, mrq, mrq->stop);
+	else
+		msdc_request_done(host, mrq);
+}
+
+static bool msdc_data_xfer_done(struct msdc_host *host, u32 events,
+				struct mmc_request *mrq, struct mmc_data *data)
+{
+	struct mmc_command *stop;
+	unsigned long flags;
+	bool done;
+	unsigned int check_data = events &
+	    (MSDC_INT_XFER_COMPL | MSDC_INT_DATCRCERR | MSDC_INT_DATTMO
+	     | MSDC_INT_DMA_BDCSERR | MSDC_INT_DMA_GPDCSERR
+	     | MSDC_INT_DMA_PROTECT);
+
+	spin_lock_irqsave(&host->lock, flags);
+	done = !host->data;
+	if (check_data)
+		host->data = NULL;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (done)
+		return true;
+	stop = data->stop;
+
+	if (check_data || (stop && stop->error)) {
+		dev_dbg(host->dev, "DMA status: 0x%8X\n",
+				readl(host->base + MSDC_DMA_CFG));
+		sdr_set_field(host->base + MSDC_DMA_CTRL, MSDC_DMA_CTRL_STOP,
+				1);
+		while (readl(host->base + MSDC_DMA_CFG) & MSDC_DMA_CFG_STS)
+			cpu_relax();
+		sdr_clr_bits(host->base + MSDC_INTEN, data_ints_mask);
+		dev_dbg(host->dev, "DMA stop\n");
+
+		if ((events & MSDC_INT_XFER_COMPL) && (!stop || !stop->error)) {
+			data->bytes_xfered = data->blocks * data->blksz;
+		} else {
+			dev_dbg(host->dev, "interrupt events: %x\n", events);
+			msdc_reset_hw(host);
+			host->error |= REQ_DAT_ERR;
+			data->bytes_xfered = 0;
+
+			if (events & MSDC_INT_DATTMO)
+				data->error = -ETIMEDOUT;
+			else if (events & MSDC_INT_DATCRCERR)
+				data->error = -EILSEQ;
+
+			dev_dbg(host->dev, "%s: cmd=%d; blocks=%d",
+				__func__, mrq->cmd->opcode, data->blocks);
+			dev_dbg(host->dev, "data_error=%d xfer_size=%d\n",
+				(int)data->error, data->bytes_xfered);
+		}
+
+		msdc_data_xfer_next(host, mrq, data);
+		done = true;
+	}
+	return done;
+}
+
+static void msdc_set_buswidth(struct msdc_host *host, u32 width)
+{
+	u32 val = readl(host->base + SDC_CFG);
+
+	val &= ~SDC_CFG_BUSWIDTH;
+
+	switch (width) {
+	default:
+	case MMC_BUS_WIDTH_1:
+		val |= (MSDC_BUS_1BITS << 16);
+		break;
+	case MMC_BUS_WIDTH_4:
+		val |= (MSDC_BUS_4BITS << 16);
+		break;
+	case MMC_BUS_WIDTH_8:
+		val |= (MSDC_BUS_8BITS << 16);
+		break;
+	}
+
+	writel(val, host->base + SDC_CFG);
+	dev_dbg(host->dev, "Bus Width = %d", width);
+}
+
+static int msdc_ops_switch_volt(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+	int ret;
+
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+		if (ios->signal_voltage != MMC_SIGNAL_VOLTAGE_330 &&
+		    ios->signal_voltage != MMC_SIGNAL_VOLTAGE_180) {
+			dev_err(host->dev, "Unsupported signal voltage!\n");
+			return -EINVAL;
+		}
+
+		ret = mmc_regulator_set_vqmmc(mmc, ios);
+		if (ret < 0) {
+			dev_dbg(host->dev, "Regulator set error %d (%d)\n",
+				ret, ios->signal_voltage);
+			return ret;
+		}
+
+		/* Apply different pinctrl settings for different signal voltage */
+		if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_180)
+			pinctrl_select_state(host->pinctrl, host->pins_uhs);
+		else
+			pinctrl_select_state(host->pinctrl, host->pins_default);
+	}
+	return 0;
+}
+
+static int msdc_card_busy(struct mmc_host *mmc)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+	u32 status = readl(host->base + MSDC_PS);
+
+	/* only check if data0 is low */
+	return !(status & BIT(16));
+}
+
+static void msdc_request_timeout(struct work_struct *work)
+{
+	struct msdc_host *host = container_of(work, struct msdc_host,
+			req_timeout.work);
+
+	/* simulate HW timeout status */
+	dev_err(host->dev, "%s: aborting cmd/data/mrq\n", __func__);
+	if (host->mrq) {
+		dev_err(host->dev, "%s: aborting mrq=%p cmd=%d\n", __func__,
+				host->mrq, host->mrq->cmd->opcode);
+		if (host->cmd) {
+			dev_err(host->dev, "%s: aborting cmd=%d\n",
+					__func__, host->cmd->opcode);
+			msdc_cmd_done(host, MSDC_INT_CMDTMO, host->mrq,
+					host->cmd);
+		} else if (host->data) {
+			dev_err(host->dev, "%s: abort data: cmd%d; %d blocks\n",
+					__func__, host->mrq->cmd->opcode,
+					host->data->blocks);
+			msdc_data_xfer_done(host, MSDC_INT_DATTMO, host->mrq,
+					host->data);
+		}
+	}
+}
+
+static void __msdc_enable_sdio_irq(struct msdc_host *host, int enb)
+{
+	if (enb) {
+		sdr_set_bits(host->base + MSDC_INTEN, MSDC_INTEN_SDIOIRQ);
+		sdr_set_bits(host->base + SDC_CFG, SDC_CFG_SDIOIDE);
+		if (host->dev_comp->recheck_sdio_irq)
+			msdc_recheck_sdio_irq(host);
+	} else {
+		sdr_clr_bits(host->base + MSDC_INTEN, MSDC_INTEN_SDIOIRQ);
+		sdr_clr_bits(host->base + SDC_CFG, SDC_CFG_SDIOIDE);
+	}
+}
+
+static void msdc_enable_sdio_irq(struct mmc_host *mmc, int enb)
+{
+	unsigned long flags;
+	struct msdc_host *host = mmc_priv(mmc);
+
+	spin_lock_irqsave(&host->lock, flags);
+	__msdc_enable_sdio_irq(host, enb);
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (enb)
+		pm_runtime_get_noresume(host->dev);
+	else
+		pm_runtime_put_noidle(host->dev);
+}
+
+static irqreturn_t msdc_cmdq_irq(struct msdc_host *host, u32 intsts)
+{
+	struct mmc_host *mmc = mmc_from_priv(host);
+	int cmd_err = 0, dat_err = 0;
+
+	if (intsts & MSDC_INT_RSPCRCERR) {
+		cmd_err = -EILSEQ;
+		dev_err(host->dev, "%s: CMD CRC ERR", __func__);
+	} else if (intsts & MSDC_INT_CMDTMO) {
+		cmd_err = -ETIMEDOUT;
+		dev_err(host->dev, "%s: CMD TIMEOUT ERR", __func__);
+	}
+
+	if (intsts & MSDC_INT_DATCRCERR) {
+		dat_err = -EILSEQ;
+		dev_err(host->dev, "%s: DATA CRC ERR", __func__);
+	} else if (intsts & MSDC_INT_DATTMO) {
+		dat_err = -ETIMEDOUT;
+		dev_err(host->dev, "%s: DATA TIMEOUT ERR", __func__);
+	}
+
+	if (cmd_err || dat_err) {
+		dev_err(host->dev, "cmd_err = %d, dat_err =%d, intsts = 0x%x",
+			cmd_err, dat_err, intsts);
+	}
+
+	return cqhci_irq(mmc, 0, cmd_err, dat_err);
+}
+
+static irqreturn_t msdc_irq(int irq, void *dev_id)
+{
+	struct msdc_host *host = (struct msdc_host *) dev_id;
+	struct mmc_host *mmc = mmc_from_priv(host);
+
+	while (true) {
+		unsigned long flags;
+		struct mmc_request *mrq;
+		struct mmc_command *cmd;
+		struct mmc_data *data;
+		u32 events, event_mask;
+
+		spin_lock_irqsave(&host->lock, flags);
+		events = readl(host->base + MSDC_INT);
+		event_mask = readl(host->base + MSDC_INTEN);
+		if ((events & event_mask) & MSDC_INT_SDIOIRQ)
+			__msdc_enable_sdio_irq(host, 0);
+		/* clear interrupts */
+		writel(events & event_mask, host->base + MSDC_INT);
+
+		mrq = host->mrq;
+		cmd = host->cmd;
+		data = host->data;
+		spin_unlock_irqrestore(&host->lock, flags);
+
+		if ((events & event_mask) & MSDC_INT_SDIOIRQ)
+			sdio_signal_irq(mmc);
+
+		if ((events & event_mask) & MSDC_INT_CDSC) {
+			if (host->internal_cd)
+				mmc_detect_change(mmc, msecs_to_jiffies(20));
+			events &= ~MSDC_INT_CDSC;
+		}
+
+		if (!(events & (event_mask & ~MSDC_INT_SDIOIRQ)))
+			break;
+
+		if ((mmc->caps2 & MMC_CAP2_CQE) &&
+		    (events & MSDC_INT_CMDQ)) {
+			msdc_cmdq_irq(host, events);
+			/* clear interrupts */
+			writel(events, host->base + MSDC_INT);
+			return IRQ_HANDLED;
+		}
+
+		if (!mrq) {
+			dev_err(host->dev,
+				"%s: MRQ=NULL; events=%08X; event_mask=%08X\n",
+				__func__, events, event_mask);
+			WARN_ON(1);
+			break;
+		}
+
+		dev_dbg(host->dev, "%s: events=%08X\n", __func__, events);
+
+		if (cmd)
+			msdc_cmd_done(host, events, mrq, cmd);
+		else if (data)
+			msdc_data_xfer_done(host, events, mrq, data);
+	}
+
+	return IRQ_HANDLED;
+}
+
+static void msdc_init_hw(struct msdc_host *host)
+{
+	u32 val;
+	u32 tune_reg = host->dev_comp->pad_tune_reg;
+
+	if (host->reset) {
+		reset_control_assert(host->reset);
+		usleep_range(10, 50);
+		reset_control_deassert(host->reset);
+	}
+
+	/* Configure to MMC/SD mode, clock free running */
+	sdr_set_bits(host->base + MSDC_CFG, MSDC_CFG_MODE | MSDC_CFG_CKPDN);
+
+	/* Reset */
+	msdc_reset_hw(host);
+
+	/* Disable and clear all interrupts */
+	writel(0, host->base + MSDC_INTEN);
+	val = readl(host->base + MSDC_INT);
+	writel(val, host->base + MSDC_INT);
+
+	/* Configure card detection */
+	if (host->internal_cd) {
+		sdr_set_field(host->base + MSDC_PS, MSDC_PS_CDDEBOUNCE,
+			      DEFAULT_DEBOUNCE);
+		sdr_set_bits(host->base + MSDC_PS, MSDC_PS_CDEN);
+		sdr_set_bits(host->base + MSDC_INTEN, MSDC_INTEN_CDSC);
+		sdr_set_bits(host->base + SDC_CFG, SDC_CFG_INSWKUP);
+	} else {
+		sdr_clr_bits(host->base + SDC_CFG, SDC_CFG_INSWKUP);
+		sdr_clr_bits(host->base + MSDC_PS, MSDC_PS_CDEN);
+		sdr_clr_bits(host->base + MSDC_INTEN, MSDC_INTEN_CDSC);
+	}
+
+	if (host->top_base) {
+		writel(0, host->top_base + EMMC_TOP_CONTROL);
+		writel(0, host->top_base + EMMC_TOP_CMD);
+	} else {
+		writel(0, host->base + tune_reg);
+	}
+	writel(0, host->base + MSDC_IOCON);
+	sdr_set_field(host->base + MSDC_IOCON, MSDC_IOCON_DDLSEL, 0);
+	writel(0x403c0046, host->base + MSDC_PATCH_BIT);
+	sdr_set_field(host->base + MSDC_PATCH_BIT, MSDC_CKGEN_MSDC_DLY_SEL, 1);
+	writel(0xffff4089, host->base + MSDC_PATCH_BIT1);
+	sdr_set_bits(host->base + EMMC50_CFG0, EMMC50_CFG_CFCSTS_SEL);
+
+	if (host->dev_comp->stop_clk_fix) {
+		sdr_set_field(host->base + MSDC_PATCH_BIT1,
+			      MSDC_PATCH_BIT1_STOP_DLY, 3);
+		sdr_clr_bits(host->base + SDC_FIFO_CFG,
+			     SDC_FIFO_CFG_WRVALIDSEL);
+		sdr_clr_bits(host->base + SDC_FIFO_CFG,
+			     SDC_FIFO_CFG_RDVALIDSEL);
+	}
+
+	if (host->dev_comp->busy_check)
+		sdr_clr_bits(host->base + MSDC_PATCH_BIT1, (1 << 7));
+
+	if (host->dev_comp->async_fifo) {
+		sdr_set_field(host->base + MSDC_PATCH_BIT2,
+			      MSDC_PB2_RESPWAIT, 3);
+		if (host->dev_comp->enhance_rx) {
+			if (host->top_base)
+				sdr_set_bits(host->top_base + EMMC_TOP_CONTROL,
+					     SDC_RX_ENH_EN);
+			else
+				sdr_set_bits(host->base + SDC_ADV_CFG0,
+					     SDC_RX_ENHANCE_EN);
+		} else {
+			sdr_set_field(host->base + MSDC_PATCH_BIT2,
+				      MSDC_PB2_RESPSTSENSEL, 2);
+			sdr_set_field(host->base + MSDC_PATCH_BIT2,
+				      MSDC_PB2_CRCSTSENSEL, 2);
+		}
+		/* use async fifo, then no need tune internal delay */
+		sdr_clr_bits(host->base + MSDC_PATCH_BIT2,
+			     MSDC_PATCH_BIT2_CFGRESP);
+		sdr_set_bits(host->base + MSDC_PATCH_BIT2,
+			     MSDC_PATCH_BIT2_CFGCRCSTS);
+	}
+
+	if (host->dev_comp->support_64g)
+		sdr_set_bits(host->base + MSDC_PATCH_BIT2,
+			     MSDC_PB2_SUPPORT_64G);
+	if (host->dev_comp->data_tune) {
+		if (host->top_base) {
+			sdr_set_bits(host->top_base + EMMC_TOP_CONTROL,
+				     PAD_DAT_RD_RXDLY_SEL);
+			sdr_clr_bits(host->top_base + EMMC_TOP_CONTROL,
+				     DATA_K_VALUE_SEL);
+			sdr_set_bits(host->top_base + EMMC_TOP_CMD,
+				     PAD_CMD_RD_RXDLY_SEL);
+		} else {
+			sdr_set_bits(host->base + tune_reg,
+				     MSDC_PAD_TUNE_RD_SEL |
+				     MSDC_PAD_TUNE_CMD_SEL);
+		}
+	} else {
+		/* choose clock tune */
+		if (host->top_base)
+			sdr_set_bits(host->top_base + EMMC_TOP_CONTROL,
+				     PAD_RXDLY_SEL);
+		else
+			sdr_set_bits(host->base + tune_reg,
+				     MSDC_PAD_TUNE_RXDLYSEL);
+	}
+
+	/* Configure to enable SDIO mode.
+	 * it's must otherwise sdio cmd5 failed
+	 */
+	sdr_set_bits(host->base + SDC_CFG, SDC_CFG_SDIO);
+
+	/* Config SDIO device detect interrupt function */
+	sdr_clr_bits(host->base + SDC_CFG, SDC_CFG_SDIOIDE);
+	sdr_set_bits(host->base + SDC_ADV_CFG0, SDC_DAT1_IRQ_TRIGGER);
+
+	/* Configure to default data timeout */
+	sdr_set_field(host->base + SDC_CFG, SDC_CFG_DTOC, 3);
+
+	host->def_tune_para.iocon = readl(host->base + MSDC_IOCON);
+	host->saved_tune_para.iocon = readl(host->base + MSDC_IOCON);
+	if (host->top_base) {
+		host->def_tune_para.emmc_top_control =
+			readl(host->top_base + EMMC_TOP_CONTROL);
+		host->def_tune_para.emmc_top_cmd =
+			readl(host->top_base + EMMC_TOP_CMD);
+		host->saved_tune_para.emmc_top_control =
+			readl(host->top_base + EMMC_TOP_CONTROL);
+		host->saved_tune_para.emmc_top_cmd =
+			readl(host->top_base + EMMC_TOP_CMD);
+	} else {
+		host->def_tune_para.pad_tune = readl(host->base + tune_reg);
+		host->saved_tune_para.pad_tune = readl(host->base + tune_reg);
+	}
+	dev_dbg(host->dev, "init hardware done!");
+}
+
+static void msdc_deinit_hw(struct msdc_host *host)
+{
+	u32 val;
+
+	if (host->internal_cd) {
+		/* Disabled card-detect */
+		sdr_clr_bits(host->base + MSDC_PS, MSDC_PS_CDEN);
+		sdr_clr_bits(host->base + SDC_CFG, SDC_CFG_INSWKUP);
+	}
+
+	/* Disable and clear all interrupts */
+	writel(0, host->base + MSDC_INTEN);
+
+	val = readl(host->base + MSDC_INT);
+	writel(val, host->base + MSDC_INT);
+}
+
+/* init gpd and bd list in msdc_drv_probe */
+static void msdc_init_gpd_bd(struct msdc_host *host, struct msdc_dma *dma)
+{
+	struct mt_gpdma_desc *gpd = dma->gpd;
+	struct mt_bdma_desc *bd = dma->bd;
+	dma_addr_t dma_addr;
+	int i;
+
+	memset(gpd, 0, sizeof(struct mt_gpdma_desc) * 2);
+
+	dma_addr = dma->gpd_addr + sizeof(struct mt_gpdma_desc);
+	gpd->gpd_info = GPDMA_DESC_BDP; /* hwo, cs, bd pointer */
+	/* gpd->next is must set for desc DMA
+	 * That's why must alloc 2 gpd structure.
+	 */
+	gpd->next = lower_32_bits(dma_addr);
+	if (host->dev_comp->support_64g)
+		gpd->gpd_info |= (upper_32_bits(dma_addr) & 0xf) << 24;
+
+	dma_addr = dma->bd_addr;
+	gpd->ptr = lower_32_bits(dma->bd_addr); /* physical address */
+	if (host->dev_comp->support_64g)
+		gpd->gpd_info |= (upper_32_bits(dma_addr) & 0xf) << 28;
+
+	memset(bd, 0, sizeof(struct mt_bdma_desc) * MAX_BD_NUM);
+	for (i = 0; i < (MAX_BD_NUM - 1); i++) {
+		dma_addr = dma->bd_addr + sizeof(*bd) * (i + 1);
+		bd[i].next = lower_32_bits(dma_addr);
+		if (host->dev_comp->support_64g)
+			bd[i].bd_info |= (upper_32_bits(dma_addr) & 0xf) << 24;
+	}
+}
+
+static void msdc_ops_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+	int ret;
+
+	msdc_set_buswidth(host, ios->bus_width);
+
+	/* Suspend/Resume will do power off/on */
+	switch (ios->power_mode) {
+	case MMC_POWER_UP:
+		if (!IS_ERR(mmc->supply.vmmc)) {
+			msdc_init_hw(host);
+			ret = mmc_regulator_set_ocr(mmc, mmc->supply.vmmc,
+					ios->vdd);
+			if (ret) {
+				dev_err(host->dev, "Failed to set vmmc power!\n");
+				return;
+			}
+		}
+		break;
+	case MMC_POWER_ON:
+		if (!IS_ERR(mmc->supply.vqmmc) && !host->vqmmc_enabled) {
+			ret = regulator_enable(mmc->supply.vqmmc);
+			if (ret)
+				dev_err(host->dev, "Failed to set vqmmc power!\n");
+			else
+				host->vqmmc_enabled = true;
+		}
+		break;
+	case MMC_POWER_OFF:
+		if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+
+		if (!IS_ERR(mmc->supply.vqmmc) && host->vqmmc_enabled) {
+			regulator_disable(mmc->supply.vqmmc);
+			host->vqmmc_enabled = false;
+		}
+		break;
+	default:
+		break;
+	}
+
+	if (host->mclk != ios->clock || host->timing != ios->timing)
+		msdc_set_mclk(host, ios->timing, ios->clock);
+}
+
+static u32 test_delay_bit(u32 delay, u32 bit)
+{
+	bit %= PAD_DELAY_MAX;
+	return delay & (1 << bit);
+}
+
+static int get_delay_len(u32 delay, u32 start_bit)
+{
+	int i;
+
+	for (i = 0; i < (PAD_DELAY_MAX - start_bit); i++) {
+		if (test_delay_bit(delay, start_bit + i) == 0)
+			return i;
+	}
+	return PAD_DELAY_MAX - start_bit;
+}
+
+static struct msdc_delay_phase get_best_delay(struct msdc_host *host, u32 delay)
+{
+	int start = 0, len = 0;
+	int start_final = 0, len_final = 0;
+	u8 final_phase = 0xff;
+	struct msdc_delay_phase delay_phase = { 0, };
+
+	if (delay == 0) {
+		dev_err(host->dev, "phase error: [map:%x]\n", delay);
+		delay_phase.final_phase = final_phase;
+		return delay_phase;
+	}
+
+	while (start < PAD_DELAY_MAX) {
+		len = get_delay_len(delay, start);
+		if (len_final < len) {
+			start_final = start;
+			len_final = len;
+		}
+		start += len ? len : 1;
+		if (len >= 12 && start_final < 4)
+			break;
+	}
+
+	/* The rule is that to find the smallest delay cell */
+	if (start_final == 0)
+		final_phase = (start_final + len_final / 3) % PAD_DELAY_MAX;
+	else
+		final_phase = (start_final + len_final / 2) % PAD_DELAY_MAX;
+	dev_info(host->dev, "phase: [map:%x] [maxlen:%d] [final:%d]\n",
+		 delay, len_final, final_phase);
+
+	delay_phase.maxlen = len_final;
+	delay_phase.start = start_final;
+	delay_phase.final_phase = final_phase;
+	return delay_phase;
+}
+
+static inline void msdc_set_cmd_delay(struct msdc_host *host, u32 value)
+{
+	u32 tune_reg = host->dev_comp->pad_tune_reg;
+
+	if (host->top_base)
+		sdr_set_field(host->top_base + EMMC_TOP_CMD, PAD_CMD_RXDLY,
+			      value);
+	else
+		sdr_set_field(host->base + tune_reg, MSDC_PAD_TUNE_CMDRDLY,
+			      value);
+}
+
+static inline void msdc_set_data_delay(struct msdc_host *host, u32 value)
+{
+	u32 tune_reg = host->dev_comp->pad_tune_reg;
+
+	if (host->top_base)
+		sdr_set_field(host->top_base + EMMC_TOP_CONTROL,
+			      PAD_DAT_RD_RXDLY, value);
+	else
+		sdr_set_field(host->base + tune_reg, MSDC_PAD_TUNE_DATRRDLY,
+			      value);
+}
+
+static int msdc_tune_response(struct mmc_host *mmc, u32 opcode)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+	u32 rise_delay = 0, fall_delay = 0;
+	struct msdc_delay_phase final_rise_delay, final_fall_delay = { 0,};
+	struct msdc_delay_phase internal_delay_phase;
+	u8 final_delay, final_maxlen;
+	u32 internal_delay = 0;
+	u32 tune_reg = host->dev_comp->pad_tune_reg;
+	int cmd_err;
+	int i, j;
+
+	if (mmc->ios.timing == MMC_TIMING_MMC_HS200 ||
+	    mmc->ios.timing == MMC_TIMING_UHS_SDR104)
+		sdr_set_field(host->base + tune_reg,
+			      MSDC_PAD_TUNE_CMDRRDLY,
+			      host->hs200_cmd_int_delay);
+
+	sdr_clr_bits(host->base + MSDC_IOCON, MSDC_IOCON_RSPL);
+	for (i = 0 ; i < PAD_DELAY_MAX; i++) {
+		msdc_set_cmd_delay(host, i);
+		/*
+		 * Using the same parameters, it may sometimes pass the test,
+		 * but sometimes it may fail. To make sure the parameters are
+		 * more stable, we test each set of parameters 3 times.
+		 */
+		for (j = 0; j < 3; j++) {
+			mmc_send_tuning(mmc, opcode, &cmd_err);
+			if (!cmd_err) {
+				rise_delay |= (1 << i);
+			} else {
+				rise_delay &= ~(1 << i);
+				break;
+			}
+		}
+	}
+	final_rise_delay = get_best_delay(host, rise_delay);
+	/* if rising edge has enough margin, then do not scan falling edge */
+	if (final_rise_delay.maxlen >= 12 ||
+	    (final_rise_delay.start == 0 && final_rise_delay.maxlen >= 4))
+		goto skip_fall;
+
+	sdr_set_bits(host->base + MSDC_IOCON, MSDC_IOCON_RSPL);
+	for (i = 0; i < PAD_DELAY_MAX; i++) {
+		msdc_set_cmd_delay(host, i);
+		/*
+		 * Using the same parameters, it may sometimes pass the test,
+		 * but sometimes it may fail. To make sure the parameters are
+		 * more stable, we test each set of parameters 3 times.
+		 */
+		for (j = 0; j < 3; j++) {
+			mmc_send_tuning(mmc, opcode, &cmd_err);
+			if (!cmd_err) {
+				fall_delay |= (1 << i);
+			} else {
+				fall_delay &= ~(1 << i);
+				break;
+			}
+		}
+	}
+	final_fall_delay = get_best_delay(host, fall_delay);
+
+skip_fall:
+	final_maxlen = max(final_rise_delay.maxlen, final_fall_delay.maxlen);
+	if (final_fall_delay.maxlen >= 12 && final_fall_delay.start < 4)
+		final_maxlen = final_fall_delay.maxlen;
+	if (final_maxlen == final_rise_delay.maxlen) {
+		sdr_clr_bits(host->base + MSDC_IOCON, MSDC_IOCON_RSPL);
+		final_delay = final_rise_delay.final_phase;
+	} else {
+		sdr_set_bits(host->base + MSDC_IOCON, MSDC_IOCON_RSPL);
+		final_delay = final_fall_delay.final_phase;
+	}
+	msdc_set_cmd_delay(host, final_delay);
+
+	if (host->dev_comp->async_fifo || host->hs200_cmd_int_delay)
+		goto skip_internal;
+
+	for (i = 0; i < PAD_DELAY_MAX; i++) {
+		sdr_set_field(host->base + tune_reg,
+			      MSDC_PAD_TUNE_CMDRRDLY, i);
+		mmc_send_tuning(mmc, opcode, &cmd_err);
+		if (!cmd_err)
+			internal_delay |= (1 << i);
+	}
+	dev_dbg(host->dev, "Final internal delay: 0x%x\n", internal_delay);
+	internal_delay_phase = get_best_delay(host, internal_delay);
+	sdr_set_field(host->base + tune_reg, MSDC_PAD_TUNE_CMDRRDLY,
+		      internal_delay_phase.final_phase);
+skip_internal:
+	dev_dbg(host->dev, "Final cmd pad delay: %x\n", final_delay);
+	return final_delay == 0xff ? -EIO : 0;
+}
+
+static int hs400_tune_response(struct mmc_host *mmc, u32 opcode)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+	u32 cmd_delay = 0;
+	struct msdc_delay_phase final_cmd_delay = { 0,};
+	u8 final_delay;
+	int cmd_err;
+	int i, j;
+
+	/* select EMMC50 PAD CMD tune */
+	sdr_set_bits(host->base + PAD_CMD_TUNE, BIT(0));
+	sdr_set_field(host->base + MSDC_PATCH_BIT1, MSDC_PATCH_BIT1_CMDTA, 2);
+
+	if (mmc->ios.timing == MMC_TIMING_MMC_HS200 ||
+	    mmc->ios.timing == MMC_TIMING_UHS_SDR104)
+		sdr_set_field(host->base + MSDC_PAD_TUNE,
+			      MSDC_PAD_TUNE_CMDRRDLY,
+			      host->hs200_cmd_int_delay);
+
+	if (host->hs400_cmd_resp_sel_rising)
+		sdr_clr_bits(host->base + MSDC_IOCON, MSDC_IOCON_RSPL);
+	else
+		sdr_set_bits(host->base + MSDC_IOCON, MSDC_IOCON_RSPL);
+	for (i = 0 ; i < PAD_DELAY_MAX; i++) {
+		sdr_set_field(host->base + PAD_CMD_TUNE,
+			      PAD_CMD_TUNE_RX_DLY3, i);
+		/*
+		 * Using the same parameters, it may sometimes pass the test,
+		 * but sometimes it may fail. To make sure the parameters are
+		 * more stable, we test each set of parameters 3 times.
+		 */
+		for (j = 0; j < 3; j++) {
+			mmc_send_tuning(mmc, opcode, &cmd_err);
+			if (!cmd_err) {
+				cmd_delay |= (1 << i);
+			} else {
+				cmd_delay &= ~(1 << i);
+				break;
+			}
+		}
+	}
+	final_cmd_delay = get_best_delay(host, cmd_delay);
+	sdr_set_field(host->base + PAD_CMD_TUNE, PAD_CMD_TUNE_RX_DLY3,
+		      final_cmd_delay.final_phase);
+	final_delay = final_cmd_delay.final_phase;
+
+	dev_dbg(host->dev, "Final cmd pad delay: %x\n", final_delay);
+	return final_delay == 0xff ? -EIO : 0;
+}
+
+static int msdc_tune_data(struct mmc_host *mmc, u32 opcode)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+	u32 rise_delay = 0, fall_delay = 0;
+	struct msdc_delay_phase final_rise_delay, final_fall_delay = { 0,};
+	u8 final_delay, final_maxlen;
+	int i, ret;
+
+	sdr_set_field(host->base + MSDC_PATCH_BIT, MSDC_INT_DAT_LATCH_CK_SEL,
+		      host->latch_ck);
+	sdr_clr_bits(host->base + MSDC_IOCON, MSDC_IOCON_DSPL);
+	sdr_clr_bits(host->base + MSDC_IOCON, MSDC_IOCON_W_DSPL);
+	for (i = 0 ; i < PAD_DELAY_MAX; i++) {
+		msdc_set_data_delay(host, i);
+		ret = mmc_send_tuning(mmc, opcode, NULL);
+		if (!ret)
+			rise_delay |= (1 << i);
+	}
+	final_rise_delay = get_best_delay(host, rise_delay);
+	/* if rising edge has enough margin, then do not scan falling edge */
+	if (final_rise_delay.maxlen >= 12 ||
+	    (final_rise_delay.start == 0 && final_rise_delay.maxlen >= 4))
+		goto skip_fall;
+
+	sdr_set_bits(host->base + MSDC_IOCON, MSDC_IOCON_DSPL);
+	sdr_set_bits(host->base + MSDC_IOCON, MSDC_IOCON_W_DSPL);
+	for (i = 0; i < PAD_DELAY_MAX; i++) {
+		msdc_set_data_delay(host, i);
+		ret = mmc_send_tuning(mmc, opcode, NULL);
+		if (!ret)
+			fall_delay |= (1 << i);
+	}
+	final_fall_delay = get_best_delay(host, fall_delay);
+
+skip_fall:
+	final_maxlen = max(final_rise_delay.maxlen, final_fall_delay.maxlen);
+	if (final_maxlen == final_rise_delay.maxlen) {
+		sdr_clr_bits(host->base + MSDC_IOCON, MSDC_IOCON_DSPL);
+		sdr_clr_bits(host->base + MSDC_IOCON, MSDC_IOCON_W_DSPL);
+		final_delay = final_rise_delay.final_phase;
+	} else {
+		sdr_set_bits(host->base + MSDC_IOCON, MSDC_IOCON_DSPL);
+		sdr_set_bits(host->base + MSDC_IOCON, MSDC_IOCON_W_DSPL);
+		final_delay = final_fall_delay.final_phase;
+	}
+	msdc_set_data_delay(host, final_delay);
+
+	dev_dbg(host->dev, "Final data pad delay: %x\n", final_delay);
+	return final_delay == 0xff ? -EIO : 0;
+}
+
+/*
+ * MSDC IP which supports data tune + async fifo can do CMD/DAT tune
+ * together, which can save the tuning time.
+ */
+static int msdc_tune_together(struct mmc_host *mmc, u32 opcode)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+	u32 rise_delay = 0, fall_delay = 0;
+	struct msdc_delay_phase final_rise_delay, final_fall_delay = { 0,};
+	u8 final_delay, final_maxlen;
+	int i, ret;
+
+	sdr_set_field(host->base + MSDC_PATCH_BIT, MSDC_INT_DAT_LATCH_CK_SEL,
+		      host->latch_ck);
+
+	sdr_clr_bits(host->base + MSDC_IOCON, MSDC_IOCON_RSPL);
+	sdr_clr_bits(host->base + MSDC_IOCON,
+		     MSDC_IOCON_DSPL | MSDC_IOCON_W_DSPL);
+	for (i = 0 ; i < PAD_DELAY_MAX; i++) {
+		msdc_set_cmd_delay(host, i);
+		msdc_set_data_delay(host, i);
+		ret = mmc_send_tuning(mmc, opcode, NULL);
+		if (!ret)
+			rise_delay |= (1 << i);
+	}
+	final_rise_delay = get_best_delay(host, rise_delay);
+	/* if rising edge has enough margin, then do not scan falling edge */
+	if (final_rise_delay.maxlen >= 12 ||
+	    (final_rise_delay.start == 0 && final_rise_delay.maxlen >= 4))
+		goto skip_fall;
+
+	sdr_set_bits(host->base + MSDC_IOCON, MSDC_IOCON_RSPL);
+	sdr_set_bits(host->base + MSDC_IOCON,
+		     MSDC_IOCON_DSPL | MSDC_IOCON_W_DSPL);
+	for (i = 0; i < PAD_DELAY_MAX; i++) {
+		msdc_set_cmd_delay(host, i);
+		msdc_set_data_delay(host, i);
+		ret = mmc_send_tuning(mmc, opcode, NULL);
+		if (!ret)
+			fall_delay |= (1 << i);
+	}
+	final_fall_delay = get_best_delay(host, fall_delay);
+
+skip_fall:
+	final_maxlen = max(final_rise_delay.maxlen, final_fall_delay.maxlen);
+	if (final_maxlen == final_rise_delay.maxlen) {
+		sdr_clr_bits(host->base + MSDC_IOCON, MSDC_IOCON_RSPL);
+		sdr_clr_bits(host->base + MSDC_IOCON,
+			     MSDC_IOCON_DSPL | MSDC_IOCON_W_DSPL);
+		final_delay = final_rise_delay.final_phase;
+	} else {
+		sdr_set_bits(host->base + MSDC_IOCON, MSDC_IOCON_RSPL);
+		sdr_set_bits(host->base + MSDC_IOCON,
+			     MSDC_IOCON_DSPL | MSDC_IOCON_W_DSPL);
+		final_delay = final_fall_delay.final_phase;
+	}
+
+	msdc_set_cmd_delay(host, final_delay);
+	msdc_set_data_delay(host, final_delay);
+
+	dev_dbg(host->dev, "Final pad delay: %x\n", final_delay);
+	return final_delay == 0xff ? -EIO : 0;
+}
+
+static int msdc_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+	int ret;
+	u32 tune_reg = host->dev_comp->pad_tune_reg;
+
+	if (host->dev_comp->data_tune && host->dev_comp->async_fifo) {
+		ret = msdc_tune_together(mmc, opcode);
+		if (host->hs400_mode) {
+			sdr_clr_bits(host->base + MSDC_IOCON,
+				     MSDC_IOCON_DSPL | MSDC_IOCON_W_DSPL);
+			msdc_set_data_delay(host, 0);
+		}
+		goto tune_done;
+	}
+	if (host->hs400_mode &&
+	    host->dev_comp->hs400_tune)
+		ret = hs400_tune_response(mmc, opcode);
+	else
+		ret = msdc_tune_response(mmc, opcode);
+	if (ret == -EIO) {
+		dev_err(host->dev, "Tune response fail!\n");
+		return ret;
+	}
+	if (host->hs400_mode == false) {
+		ret = msdc_tune_data(mmc, opcode);
+		if (ret == -EIO)
+			dev_err(host->dev, "Tune data fail!\n");
+	}
+
+tune_done:
+	host->saved_tune_para.iocon = readl(host->base + MSDC_IOCON);
+	host->saved_tune_para.pad_tune = readl(host->base + tune_reg);
+	host->saved_tune_para.pad_cmd_tune = readl(host->base + PAD_CMD_TUNE);
+	if (host->top_base) {
+		host->saved_tune_para.emmc_top_control = readl(host->top_base +
+				EMMC_TOP_CONTROL);
+		host->saved_tune_para.emmc_top_cmd = readl(host->top_base +
+				EMMC_TOP_CMD);
+	}
+	return ret;
+}
+
+static int msdc_prepare_hs400_tuning(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+	host->hs400_mode = true;
+
+	if (host->top_base)
+		writel(host->hs400_ds_delay,
+		       host->top_base + EMMC50_PAD_DS_TUNE);
+	else
+		writel(host->hs400_ds_delay, host->base + PAD_DS_TUNE);
+	/* hs400 mode must set it to 0 */
+	sdr_clr_bits(host->base + MSDC_PATCH_BIT2, MSDC_PATCH_BIT2_CFGCRCSTS);
+	/* to improve read performance, set outstanding to 2 */
+	sdr_set_field(host->base + EMMC50_CFG3, EMMC50_CFG3_OUTS_WR, 2);
+
+	return 0;
+}
+
+static void msdc_hw_reset(struct mmc_host *mmc)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+
+	sdr_set_bits(host->base + EMMC_IOCON, 1);
+	udelay(10); /* 10us is enough */
+	sdr_clr_bits(host->base + EMMC_IOCON, 1);
+}
+
+static void msdc_ack_sdio_irq(struct mmc_host *mmc)
+{
+	unsigned long flags;
+	struct msdc_host *host = mmc_priv(mmc);
+
+	spin_lock_irqsave(&host->lock, flags);
+	__msdc_enable_sdio_irq(host, 1);
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static int msdc_get_cd(struct mmc_host *mmc)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+	int val;
+
+	if (mmc->caps & MMC_CAP_NONREMOVABLE)
+		return 1;
+
+	if (!host->internal_cd)
+		return mmc_gpio_get_cd(mmc);
+
+	val = readl(host->base + MSDC_PS) & MSDC_PS_CDSTS;
+	if (mmc->caps2 & MMC_CAP2_CD_ACTIVE_HIGH)
+		return !!val;
+	else
+		return !val;
+}
+
+static void msdc_cqe_enable(struct mmc_host *mmc)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+
+	/* enable cmdq irq */
+	writel(MSDC_INT_CMDQ, host->base + MSDC_INTEN);
+	/* enable busy check */
+	sdr_set_bits(host->base + MSDC_PATCH_BIT1, MSDC_PB1_BUSY_CHECK_SEL);
+	/* default write data / busy timeout 20s */
+	msdc_set_busy_timeout(host, 20 * 1000000000ULL, 0);
+	/* default read data timeout 1s */
+	msdc_set_timeout(host, 1000000000ULL, 0);
+}
+
+static void msdc_cqe_disable(struct mmc_host *mmc, bool recovery)
+{
+	struct msdc_host *host = mmc_priv(mmc);
+	unsigned int val = 0;
+
+	/* disable cmdq irq */
+	sdr_clr_bits(host->base + MSDC_INTEN, MSDC_INT_CMDQ);
+	/* disable busy check */
+	sdr_clr_bits(host->base + MSDC_PATCH_BIT1, MSDC_PB1_BUSY_CHECK_SEL);
+
+	val = readl(host->base + MSDC_INT);
+	writel(val, host->base + MSDC_INT);
+
+	if (recovery) {
+		sdr_set_field(host->base + MSDC_DMA_CTRL,
+			      MSDC_DMA_CTRL_STOP, 1);
+		if (WARN_ON(readl_poll_timeout(host->base + MSDC_DMA_CFG, val,
+			!(val & MSDC_DMA_CFG_STS), 1, 3000)))
+			return;
+		msdc_reset_hw(host);
+	}
+}
+
+static void msdc_cqe_pre_enable(struct mmc_host *mmc)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	u32 reg;
+
+	reg = cqhci_readl(cq_host, CQHCI_CFG);
+	reg |= CQHCI_ENABLE;
+	cqhci_writel(cq_host, reg, CQHCI_CFG);
+}
+
+static void msdc_cqe_post_disable(struct mmc_host *mmc)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	u32 reg;
+
+	reg = cqhci_readl(cq_host, CQHCI_CFG);
+	reg &= ~CQHCI_ENABLE;
+	cqhci_writel(cq_host, reg, CQHCI_CFG);
+}
+
+static const struct mmc_host_ops mt_msdc_ops = {
+	.post_req = msdc_post_req,
+	.pre_req = msdc_pre_req,
+	.request = msdc_ops_request,
+	.set_ios = msdc_ops_set_ios,
+	.get_ro = mmc_gpio_get_ro,
+	.get_cd = msdc_get_cd,
+	.enable_sdio_irq = msdc_enable_sdio_irq,
+	.ack_sdio_irq = msdc_ack_sdio_irq,
+	.start_signal_voltage_switch = msdc_ops_switch_volt,
+	.card_busy = msdc_card_busy,
+	.execute_tuning = msdc_execute_tuning,
+	.prepare_hs400_tuning = msdc_prepare_hs400_tuning,
+	.hw_reset = msdc_hw_reset,
+};
+
+static const struct cqhci_host_ops msdc_cmdq_ops = {
+	.enable         = msdc_cqe_enable,
+	.disable        = msdc_cqe_disable,
+	.pre_enable = msdc_cqe_pre_enable,
+	.post_disable = msdc_cqe_post_disable,
+};
+
+static void msdc_of_property_parse(struct platform_device *pdev,
+				   struct msdc_host *host)
+{
+	of_property_read_u32(pdev->dev.of_node, "mediatek,latch-ck",
+			     &host->latch_ck);
+
+	of_property_read_u32(pdev->dev.of_node, "hs400-ds-delay",
+			     &host->hs400_ds_delay);
+
+	of_property_read_u32(pdev->dev.of_node, "mediatek,hs200-cmd-int-delay",
+			     &host->hs200_cmd_int_delay);
+
+	of_property_read_u32(pdev->dev.of_node, "mediatek,hs400-cmd-int-delay",
+			     &host->hs400_cmd_int_delay);
+
+	if (of_property_read_bool(pdev->dev.of_node,
+				  "mediatek,hs400-cmd-resp-sel-rising"))
+		host->hs400_cmd_resp_sel_rising = true;
+	else
+		host->hs400_cmd_resp_sel_rising = false;
+
+	if (of_property_read_bool(pdev->dev.of_node,
+				  "supports-cqe"))
+		host->cqhci = true;
+	else
+		host->cqhci = false;
+}
+
+static int msdc_drv_probe(struct platform_device *pdev)
+{
+	struct mmc_host *mmc;
+	struct msdc_host *host;
+	struct resource *res;
+	int ret;
+
+	if (!pdev->dev.of_node) {
+		dev_err(&pdev->dev, "No DT found\n");
+		return -EINVAL;
+	}
+
+	/* Allocate MMC host for this device */
+	mmc = mmc_alloc_host(sizeof(struct msdc_host), &pdev->dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	host = mmc_priv(mmc);
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto host_free;
+
+	host->base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(host->base)) {
+		ret = PTR_ERR(host->base);
+		goto host_free;
+	}
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+	if (res) {
+		host->top_base = devm_ioremap_resource(&pdev->dev, res);
+		if (IS_ERR(host->top_base))
+			host->top_base = NULL;
+	}
+
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret)
+		goto host_free;
+
+	host->src_clk = devm_clk_get(&pdev->dev, "source");
+	if (IS_ERR(host->src_clk)) {
+		ret = PTR_ERR(host->src_clk);
+		goto host_free;
+	}
+
+	host->h_clk = devm_clk_get(&pdev->dev, "hclk");
+	if (IS_ERR(host->h_clk)) {
+		ret = PTR_ERR(host->h_clk);
+		goto host_free;
+	}
+
+	host->bus_clk = devm_clk_get(&pdev->dev, "bus_clk");
+	if (IS_ERR(host->bus_clk))
+		host->bus_clk = NULL;
+	/*source clock control gate is optional clock*/
+	host->src_clk_cg = devm_clk_get(&pdev->dev, "source_cg");
+	if (IS_ERR(host->src_clk_cg))
+		host->src_clk_cg = NULL;
+
+	host->reset = devm_reset_control_get_optional_exclusive(&pdev->dev,
+								"hrst");
+	if (IS_ERR(host->reset))
+		return PTR_ERR(host->reset);
+
+	host->irq = platform_get_irq(pdev, 0);
+	if (host->irq < 0) {
+		ret = host->irq;
+		goto host_free;
+	}
+
+	host->pinctrl = devm_pinctrl_get(&pdev->dev);
+	if (IS_ERR(host->pinctrl)) {
+		ret = PTR_ERR(host->pinctrl);
+		dev_err(&pdev->dev, "Cannot find pinctrl!\n");
+		goto host_free;
+	}
+
+	host->pins_default = pinctrl_lookup_state(host->pinctrl, "default");
+	if (IS_ERR(host->pins_default)) {
+		ret = PTR_ERR(host->pins_default);
+		dev_err(&pdev->dev, "Cannot find pinctrl default!\n");
+		goto host_free;
+	}
+
+	host->pins_uhs = pinctrl_lookup_state(host->pinctrl, "state_uhs");
+	if (IS_ERR(host->pins_uhs)) {
+		ret = PTR_ERR(host->pins_uhs);
+		dev_err(&pdev->dev, "Cannot find pinctrl uhs!\n");
+		goto host_free;
+	}
+
+	msdc_of_property_parse(pdev, host);
+
+	host->dev = &pdev->dev;
+	host->dev_comp = of_device_get_match_data(&pdev->dev);
+	host->src_clk_freq = clk_get_rate(host->src_clk);
+	/* Set host parameters to mmc */
+	mmc->ops = &mt_msdc_ops;
+	if (host->dev_comp->clk_div_bits == 8)
+		mmc->f_min = DIV_ROUND_UP(host->src_clk_freq, 4 * 255);
+	else
+		mmc->f_min = DIV_ROUND_UP(host->src_clk_freq, 4 * 4095);
+
+	if (!(mmc->caps & MMC_CAP_NONREMOVABLE) &&
+	    !mmc_can_gpio_cd(mmc) &&
+	    host->dev_comp->use_internal_cd) {
+		/*
+		 * Is removable but no GPIO declared, so
+		 * use internal functionality.
+		 */
+		host->internal_cd = true;
+	}
+
+	if (mmc->caps & MMC_CAP_SDIO_IRQ)
+		mmc->caps2 |= MMC_CAP2_SDIO_IRQ_NOTHREAD;
+
+	mmc->caps |= MMC_CAP_CMD23;
+	if (host->cqhci)
+		mmc->caps2 |= MMC_CAP2_CQE | MMC_CAP2_CQE_DCMD;
+	/* MMC core transfer sizes tunable parameters */
+	mmc->max_segs = MAX_BD_NUM;
+	if (host->dev_comp->support_64g)
+		mmc->max_seg_size = BDMA_DESC_BUFLEN_EXT;
+	else
+		mmc->max_seg_size = BDMA_DESC_BUFLEN;
+	mmc->max_blk_size = 2048;
+	mmc->max_req_size = 512 * 1024;
+	mmc->max_blk_count = mmc->max_req_size / 512;
+	if (host->dev_comp->support_64g)
+		host->dma_mask = DMA_BIT_MASK(36);
+	else
+		host->dma_mask = DMA_BIT_MASK(32);
+	mmc_dev(mmc)->dma_mask = &host->dma_mask;
+
+	host->timeout_clks = 3 * 1048576;
+	host->dma.gpd = dma_alloc_coherent(&pdev->dev,
+				2 * sizeof(struct mt_gpdma_desc),
+				&host->dma.gpd_addr, GFP_KERNEL);
+	host->dma.bd = dma_alloc_coherent(&pdev->dev,
+				MAX_BD_NUM * sizeof(struct mt_bdma_desc),
+				&host->dma.bd_addr, GFP_KERNEL);
+	if (!host->dma.gpd || !host->dma.bd) {
+		ret = -ENOMEM;
+		goto release_mem;
+	}
+	msdc_init_gpd_bd(host, &host->dma);
+	INIT_DELAYED_WORK(&host->req_timeout, msdc_request_timeout);
+	spin_lock_init(&host->lock);
+
+	platform_set_drvdata(pdev, mmc);
+	msdc_ungate_clock(host);
+	msdc_init_hw(host);
+
+	if (mmc->caps2 & MMC_CAP2_CQE) {
+		host->cq_host = devm_kzalloc(mmc->parent,
+					     sizeof(*host->cq_host),
+					     GFP_KERNEL);
+		if (!host->cq_host) {
+			ret = -ENOMEM;
+			goto host_free;
+		}
+		host->cq_host->caps |= CQHCI_TASK_DESC_SZ_128;
+		host->cq_host->mmio = host->base + 0x800;
+		host->cq_host->ops = &msdc_cmdq_ops;
+		ret = cqhci_init(host->cq_host, mmc, true);
+		if (ret)
+			goto host_free;
+		mmc->max_segs = 128;
+		/* cqhci 16bit length */
+		/* 0 size, means 65536 so we don't have to -1 here */
+		mmc->max_seg_size = 64 * 1024;
+	}
+
+	ret = devm_request_irq(&pdev->dev, host->irq, msdc_irq,
+			       IRQF_TRIGGER_NONE, pdev->name, host);
+	if (ret)
+		goto release;
+
+	pm_runtime_set_active(host->dev);
+	pm_runtime_set_autosuspend_delay(host->dev, MTK_MMC_AUTOSUSPEND_DELAY);
+	pm_runtime_use_autosuspend(host->dev);
+	pm_runtime_enable(host->dev);
+	ret = mmc_add_host(mmc);
+
+	if (ret)
+		goto end;
+
+	return 0;
+end:
+	pm_runtime_disable(host->dev);
+release:
+	platform_set_drvdata(pdev, NULL);
+	msdc_deinit_hw(host);
+	msdc_gate_clock(host);
+release_mem:
+	if (host->dma.gpd)
+		dma_free_coherent(&pdev->dev,
+			2 * sizeof(struct mt_gpdma_desc),
+			host->dma.gpd, host->dma.gpd_addr);
+	if (host->dma.bd)
+		dma_free_coherent(&pdev->dev,
+			MAX_BD_NUM * sizeof(struct mt_bdma_desc),
+			host->dma.bd, host->dma.bd_addr);
+host_free:
+	mmc_free_host(mmc);
+
+	return ret;
+}
+
+static int msdc_drv_remove(struct platform_device *pdev)
+{
+	struct mmc_host *mmc;
+	struct msdc_host *host;
+
+	mmc = platform_get_drvdata(pdev);
+	host = mmc_priv(mmc);
+
+	pm_runtime_get_sync(host->dev);
+
+	platform_set_drvdata(pdev, NULL);
+	mmc_remove_host(mmc);
+	msdc_deinit_hw(host);
+	msdc_gate_clock(host);
+
+	pm_runtime_disable(host->dev);
+	pm_runtime_put_noidle(host->dev);
+	dma_free_coherent(&pdev->dev,
+			2 * sizeof(struct mt_gpdma_desc),
+			host->dma.gpd, host->dma.gpd_addr);
+	dma_free_coherent(&pdev->dev, MAX_BD_NUM * sizeof(struct mt_bdma_desc),
+			host->dma.bd, host->dma.bd_addr);
+
+	mmc_free_host(mmc);
+
+	return 0;
+}
+
+static void msdc_save_reg(struct msdc_host *host)
+{
+	u32 tune_reg = host->dev_comp->pad_tune_reg;
+
+	host->save_para.msdc_cfg = readl(host->base + MSDC_CFG);
+	host->save_para.iocon = readl(host->base + MSDC_IOCON);
+	host->save_para.sdc_cfg = readl(host->base + SDC_CFG);
+	host->save_para.patch_bit0 = readl(host->base + MSDC_PATCH_BIT);
+	host->save_para.patch_bit1 = readl(host->base + MSDC_PATCH_BIT1);
+	host->save_para.patch_bit2 = readl(host->base + MSDC_PATCH_BIT2);
+	host->save_para.pad_ds_tune = readl(host->base + PAD_DS_TUNE);
+	host->save_para.pad_cmd_tune = readl(host->base + PAD_CMD_TUNE);
+	host->save_para.emmc50_cfg0 = readl(host->base + EMMC50_CFG0);
+	host->save_para.emmc50_cfg3 = readl(host->base + EMMC50_CFG3);
+	host->save_para.sdc_fifo_cfg = readl(host->base + SDC_FIFO_CFG);
+	if (host->top_base) {
+		host->save_para.emmc_top_control =
+			readl(host->top_base + EMMC_TOP_CONTROL);
+		host->save_para.emmc_top_cmd =
+			readl(host->top_base + EMMC_TOP_CMD);
+		host->save_para.emmc50_pad_ds_tune =
+			readl(host->top_base + EMMC50_PAD_DS_TUNE);
+	} else {
+		host->save_para.pad_tune = readl(host->base + tune_reg);
+	}
+}
+
+static void msdc_restore_reg(struct msdc_host *host)
+{
+	struct mmc_host *mmc = mmc_from_priv(host);
+	u32 tune_reg = host->dev_comp->pad_tune_reg;
+
+	writel(host->save_para.msdc_cfg, host->base + MSDC_CFG);
+	writel(host->save_para.iocon, host->base + MSDC_IOCON);
+	writel(host->save_para.sdc_cfg, host->base + SDC_CFG);
+	writel(host->save_para.patch_bit0, host->base + MSDC_PATCH_BIT);
+	writel(host->save_para.patch_bit1, host->base + MSDC_PATCH_BIT1);
+	writel(host->save_para.patch_bit2, host->base + MSDC_PATCH_BIT2);
+	writel(host->save_para.pad_ds_tune, host->base + PAD_DS_TUNE);
+	writel(host->save_para.pad_cmd_tune, host->base + PAD_CMD_TUNE);
+	writel(host->save_para.emmc50_cfg0, host->base + EMMC50_CFG0);
+	writel(host->save_para.emmc50_cfg3, host->base + EMMC50_CFG3);
+	writel(host->save_para.sdc_fifo_cfg, host->base + SDC_FIFO_CFG);
+	if (host->top_base) {
+		writel(host->save_para.emmc_top_control,
+		       host->top_base + EMMC_TOP_CONTROL);
+		writel(host->save_para.emmc_top_cmd,
+		       host->top_base + EMMC_TOP_CMD);
+		writel(host->save_para.emmc50_pad_ds_tune,
+		       host->top_base + EMMC50_PAD_DS_TUNE);
+	} else {
+		writel(host->save_para.pad_tune, host->base + tune_reg);
+	}
+
+	if (sdio_irq_claimed(mmc))
+		__msdc_enable_sdio_irq(host, 1);
+}
+
+static int __maybe_unused msdc_runtime_suspend(struct device *dev)
+{
+	struct mmc_host *mmc = dev_get_drvdata(dev);
+	struct msdc_host *host = mmc_priv(mmc);
+
+	msdc_save_reg(host);
+	msdc_gate_clock(host);
+	return 0;
+}
+
+static int __maybe_unused msdc_runtime_resume(struct device *dev)
+{
+	struct mmc_host *mmc = dev_get_drvdata(dev);
+	struct msdc_host *host = mmc_priv(mmc);
+
+	msdc_ungate_clock(host);
+	msdc_restore_reg(host);
+	return 0;
+}
+
+static int __maybe_unused msdc_suspend(struct device *dev)
+{
+	struct mmc_host *mmc = dev_get_drvdata(dev);
+	int ret;
+	u32 val;
+
+	if (mmc->caps2 & MMC_CAP2_CQE) {
+		ret = cqhci_suspend(mmc);
+		if (ret)
+			return ret;
+		val = readl(((struct msdc_host *)mmc_priv(mmc))->base + MSDC_INT);
+		writel(val, ((struct msdc_host *)mmc_priv(mmc))->base + MSDC_INT);
+	}
+
+	return pm_runtime_force_suspend(dev);
+}
+
+static int __maybe_unused msdc_resume(struct device *dev)
+{
+	return pm_runtime_force_resume(dev);
+}
+
+static const struct dev_pm_ops msdc_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(msdc_suspend, msdc_resume)
+	SET_RUNTIME_PM_OPS(msdc_runtime_suspend, msdc_runtime_resume, NULL)
+};
+
+static struct platform_driver mt_msdc_driver = {
+	.probe = msdc_drv_probe,
+	.remove = msdc_drv_remove,
+	.driver = {
+		.name = "mtk-msdc",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = msdc_of_ids,
+		.pm = &msdc_dev_pm_ops,
+	},
+};
+
+module_platform_driver(mt_msdc_driver);
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("MediaTek SD/MMC Card Driver");
diff --git a/drivers/mmc/host/mvsdio.c b/drivers/mmc/host/mvsdio.c
new file mode 100644
index 000000000..b4f6a0a2f
--- /dev/null
+++ b/drivers/mmc/host/mvsdio.c
@@ -0,0 +1,843 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Marvell MMC/SD/SDIO driver
+ *
+ * Authors: Maen Suleiman, Nicolas Pitre
+ * Copyright (C) 2008-2009 Marvell Ltd.
+ */
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/io.h>
+#include <linux/platform_device.h>
+#include <linux/mbus.h>
+#include <linux/delay.h>
+#include <linux/interrupt.h>
+#include <linux/dma-mapping.h>
+#include <linux/scatterlist.h>
+#include <linux/irq.h>
+#include <linux/clk.h>
+#include <linux/of_irq.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/slot-gpio.h>
+
+#include <linux/sizes.h>
+#include <asm/unaligned.h>
+
+#include "mvsdio.h"
+
+#define DRIVER_NAME	"mvsdio"
+
+static int maxfreq;
+static int nodma;
+
+struct mvsd_host {
+	void __iomem *base;
+	struct mmc_request *mrq;
+	spinlock_t lock;
+	unsigned int xfer_mode;
+	unsigned int intr_en;
+	unsigned int ctrl;
+	unsigned int pio_size;
+	void *pio_ptr;
+	unsigned int sg_frags;
+	unsigned int ns_per_clk;
+	unsigned int clock;
+	unsigned int base_clock;
+	struct timer_list timer;
+	struct mmc_host *mmc;
+	struct device *dev;
+	struct clk *clk;
+};
+
+#define mvsd_write(offs, val)	writel(val, iobase + (offs))
+#define mvsd_read(offs)		readl(iobase + (offs))
+
+static int mvsd_setup_data(struct mvsd_host *host, struct mmc_data *data)
+{
+	void __iomem *iobase = host->base;
+	unsigned int tmout;
+	int tmout_index;
+
+	/*
+	 * Hardware weirdness.  The FIFO_EMPTY bit of the HW_STATE
+	 * register is sometimes not set before a while when some
+	 * "unusual" data block sizes are used (such as with the SWITCH
+	 * command), even despite the fact that the XFER_DONE interrupt
+	 * was raised.  And if another data transfer starts before
+	 * this bit comes to good sense (which eventually happens by
+	 * itself) then the new transfer simply fails with a timeout.
+	 */
+	if (!(mvsd_read(MVSD_HW_STATE) & (1 << 13))) {
+		unsigned long t = jiffies + HZ;
+		unsigned int hw_state,  count = 0;
+		do {
+			hw_state = mvsd_read(MVSD_HW_STATE);
+			if (time_after(jiffies, t)) {
+				dev_warn(host->dev, "FIFO_EMPTY bit missing\n");
+				break;
+			}
+			count++;
+		} while (!(hw_state & (1 << 13)));
+		dev_dbg(host->dev, "*** wait for FIFO_EMPTY bit "
+				   "(hw=0x%04x, count=%d, jiffies=%ld)\n",
+				   hw_state, count, jiffies - (t - HZ));
+	}
+
+	/* If timeout=0 then maximum timeout index is used. */
+	tmout = DIV_ROUND_UP(data->timeout_ns, host->ns_per_clk);
+	tmout += data->timeout_clks;
+	tmout_index = fls(tmout - 1) - 12;
+	if (tmout_index < 0)
+		tmout_index = 0;
+	if (tmout_index > MVSD_HOST_CTRL_TMOUT_MAX)
+		tmout_index = MVSD_HOST_CTRL_TMOUT_MAX;
+
+	dev_dbg(host->dev, "data %s at 0x%08x: blocks=%d blksz=%d tmout=%u (%d)\n",
+		(data->flags & MMC_DATA_READ) ? "read" : "write",
+		(u32)sg_virt(data->sg), data->blocks, data->blksz,
+		tmout, tmout_index);
+
+	host->ctrl &= ~MVSD_HOST_CTRL_TMOUT_MASK;
+	host->ctrl |= MVSD_HOST_CTRL_TMOUT(tmout_index);
+	mvsd_write(MVSD_HOST_CTRL, host->ctrl);
+	mvsd_write(MVSD_BLK_COUNT, data->blocks);
+	mvsd_write(MVSD_BLK_SIZE, data->blksz);
+
+	if (nodma || (data->blksz | data->sg->offset) & 3 ||
+	    ((!(data->flags & MMC_DATA_READ) && data->sg->offset & 0x3f))) {
+		/*
+		 * We cannot do DMA on a buffer which offset or size
+		 * is not aligned on a 4-byte boundary.
+		 *
+		 * It also appears the host to card DMA can corrupt
+		 * data when the buffer is not aligned on a 64 byte
+		 * boundary.
+		 */
+		host->pio_size = data->blocks * data->blksz;
+		host->pio_ptr = sg_virt(data->sg);
+		if (!nodma)
+			dev_dbg(host->dev, "fallback to PIO for data at 0x%p size %d\n",
+				host->pio_ptr, host->pio_size);
+		return 1;
+	} else {
+		dma_addr_t phys_addr;
+
+		host->sg_frags = dma_map_sg(mmc_dev(host->mmc),
+					    data->sg, data->sg_len,
+					    mmc_get_dma_dir(data));
+		phys_addr = sg_dma_address(data->sg);
+		mvsd_write(MVSD_SYS_ADDR_LOW, (u32)phys_addr & 0xffff);
+		mvsd_write(MVSD_SYS_ADDR_HI,  (u32)phys_addr >> 16);
+		return 0;
+	}
+}
+
+static void mvsd_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct mvsd_host *host = mmc_priv(mmc);
+	void __iomem *iobase = host->base;
+	struct mmc_command *cmd = mrq->cmd;
+	u32 cmdreg = 0, xfer = 0, intr = 0;
+	unsigned long flags;
+	unsigned int timeout;
+
+	BUG_ON(host->mrq != NULL);
+	host->mrq = mrq;
+
+	dev_dbg(host->dev, "cmd %d (hw state 0x%04x)\n",
+		cmd->opcode, mvsd_read(MVSD_HW_STATE));
+
+	cmdreg = MVSD_CMD_INDEX(cmd->opcode);
+
+	if (cmd->flags & MMC_RSP_BUSY)
+		cmdreg |= MVSD_CMD_RSP_48BUSY;
+	else if (cmd->flags & MMC_RSP_136)
+		cmdreg |= MVSD_CMD_RSP_136;
+	else if (cmd->flags & MMC_RSP_PRESENT)
+		cmdreg |= MVSD_CMD_RSP_48;
+	else
+		cmdreg |= MVSD_CMD_RSP_NONE;
+
+	if (cmd->flags & MMC_RSP_CRC)
+		cmdreg |= MVSD_CMD_CHECK_CMDCRC;
+
+	if (cmd->flags & MMC_RSP_OPCODE)
+		cmdreg |= MVSD_CMD_INDX_CHECK;
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		cmdreg |= MVSD_UNEXPECTED_RESP;
+		intr |= MVSD_NOR_UNEXP_RSP;
+	}
+
+	if (mrq->data) {
+		struct mmc_data *data = mrq->data;
+		int pio;
+
+		cmdreg |= MVSD_CMD_DATA_PRESENT | MVSD_CMD_CHECK_DATACRC16;
+		xfer |= MVSD_XFER_MODE_HW_WR_DATA_EN;
+		if (data->flags & MMC_DATA_READ)
+			xfer |= MVSD_XFER_MODE_TO_HOST;
+
+		pio = mvsd_setup_data(host, data);
+		if (pio) {
+			xfer |= MVSD_XFER_MODE_PIO;
+			/* PIO section of mvsd_irq has comments on those bits */
+			if (data->flags & MMC_DATA_WRITE)
+				intr |= MVSD_NOR_TX_AVAIL;
+			else if (host->pio_size > 32)
+				intr |= MVSD_NOR_RX_FIFO_8W;
+			else
+				intr |= MVSD_NOR_RX_READY;
+		}
+
+		if (data->stop) {
+			struct mmc_command *stop = data->stop;
+			u32 cmd12reg = 0;
+
+			mvsd_write(MVSD_AUTOCMD12_ARG_LOW, stop->arg & 0xffff);
+			mvsd_write(MVSD_AUTOCMD12_ARG_HI,  stop->arg >> 16);
+
+			if (stop->flags & MMC_RSP_BUSY)
+				cmd12reg |= MVSD_AUTOCMD12_BUSY;
+			if (stop->flags & MMC_RSP_OPCODE)
+				cmd12reg |= MVSD_AUTOCMD12_INDX_CHECK;
+			cmd12reg |= MVSD_AUTOCMD12_INDEX(stop->opcode);
+			mvsd_write(MVSD_AUTOCMD12_CMD, cmd12reg);
+
+			xfer |= MVSD_XFER_MODE_AUTO_CMD12;
+			intr |= MVSD_NOR_AUTOCMD12_DONE;
+		} else {
+			intr |= MVSD_NOR_XFER_DONE;
+		}
+	} else {
+		intr |= MVSD_NOR_CMD_DONE;
+	}
+
+	mvsd_write(MVSD_ARG_LOW, cmd->arg & 0xffff);
+	mvsd_write(MVSD_ARG_HI,  cmd->arg >> 16);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	host->xfer_mode &= MVSD_XFER_MODE_INT_CHK_EN;
+	host->xfer_mode |= xfer;
+	mvsd_write(MVSD_XFER_MODE, host->xfer_mode);
+
+	mvsd_write(MVSD_NOR_INTR_STATUS, ~MVSD_NOR_CARD_INT);
+	mvsd_write(MVSD_ERR_INTR_STATUS, 0xffff);
+	mvsd_write(MVSD_CMD, cmdreg);
+
+	host->intr_en &= MVSD_NOR_CARD_INT;
+	host->intr_en |= intr | MVSD_NOR_ERROR;
+	mvsd_write(MVSD_NOR_INTR_EN, host->intr_en);
+	mvsd_write(MVSD_ERR_INTR_EN, 0xffff);
+
+	timeout = cmd->busy_timeout ? cmd->busy_timeout : 5000;
+	mod_timer(&host->timer, jiffies + msecs_to_jiffies(timeout));
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static u32 mvsd_finish_cmd(struct mvsd_host *host, struct mmc_command *cmd,
+			   u32 err_status)
+{
+	void __iomem *iobase = host->base;
+
+	if (cmd->flags & MMC_RSP_136) {
+		unsigned int response[8], i;
+		for (i = 0; i < 8; i++)
+			response[i] = mvsd_read(MVSD_RSP(i));
+		cmd->resp[0] =		((response[0] & 0x03ff) << 22) |
+					((response[1] & 0xffff) << 6) |
+					((response[2] & 0xfc00) >> 10);
+		cmd->resp[1] =		((response[2] & 0x03ff) << 22) |
+					((response[3] & 0xffff) << 6) |
+					((response[4] & 0xfc00) >> 10);
+		cmd->resp[2] =		((response[4] & 0x03ff) << 22) |
+					((response[5] & 0xffff) << 6) |
+					((response[6] & 0xfc00) >> 10);
+		cmd->resp[3] =		((response[6] & 0x03ff) << 22) |
+					((response[7] & 0x3fff) << 8);
+	} else if (cmd->flags & MMC_RSP_PRESENT) {
+		unsigned int response[3], i;
+		for (i = 0; i < 3; i++)
+			response[i] = mvsd_read(MVSD_RSP(i));
+		cmd->resp[0] =		((response[2] & 0x003f) << (8 - 8)) |
+					((response[1] & 0xffff) << (14 - 8)) |
+					((response[0] & 0x03ff) << (30 - 8));
+		cmd->resp[1] =		((response[0] & 0xfc00) >> 10);
+		cmd->resp[2] = 0;
+		cmd->resp[3] = 0;
+	}
+
+	if (err_status & MVSD_ERR_CMD_TIMEOUT) {
+		cmd->error = -ETIMEDOUT;
+	} else if (err_status & (MVSD_ERR_CMD_CRC | MVSD_ERR_CMD_ENDBIT |
+				 MVSD_ERR_CMD_INDEX | MVSD_ERR_CMD_STARTBIT)) {
+		cmd->error = -EILSEQ;
+	}
+	err_status &= ~(MVSD_ERR_CMD_TIMEOUT | MVSD_ERR_CMD_CRC |
+			MVSD_ERR_CMD_ENDBIT | MVSD_ERR_CMD_INDEX |
+			MVSD_ERR_CMD_STARTBIT);
+
+	return err_status;
+}
+
+static u32 mvsd_finish_data(struct mvsd_host *host, struct mmc_data *data,
+			    u32 err_status)
+{
+	void __iomem *iobase = host->base;
+
+	if (host->pio_ptr) {
+		host->pio_ptr = NULL;
+		host->pio_size = 0;
+	} else {
+		dma_unmap_sg(mmc_dev(host->mmc), data->sg, host->sg_frags,
+			     mmc_get_dma_dir(data));
+	}
+
+	if (err_status & MVSD_ERR_DATA_TIMEOUT)
+		data->error = -ETIMEDOUT;
+	else if (err_status & (MVSD_ERR_DATA_CRC | MVSD_ERR_DATA_ENDBIT))
+		data->error = -EILSEQ;
+	else if (err_status & MVSD_ERR_XFER_SIZE)
+		data->error = -EBADE;
+	err_status &= ~(MVSD_ERR_DATA_TIMEOUT | MVSD_ERR_DATA_CRC |
+			MVSD_ERR_DATA_ENDBIT | MVSD_ERR_XFER_SIZE);
+
+	dev_dbg(host->dev, "data done: blocks_left=%d, bytes_left=%d\n",
+		mvsd_read(MVSD_CURR_BLK_LEFT), mvsd_read(MVSD_CURR_BYTE_LEFT));
+	data->bytes_xfered =
+		(data->blocks - mvsd_read(MVSD_CURR_BLK_LEFT)) * data->blksz;
+	/* We can't be sure about the last block when errors are detected */
+	if (data->bytes_xfered && data->error)
+		data->bytes_xfered -= data->blksz;
+
+	/* Handle Auto cmd 12 response */
+	if (data->stop) {
+		unsigned int response[3], i;
+		for (i = 0; i < 3; i++)
+			response[i] = mvsd_read(MVSD_AUTO_RSP(i));
+		data->stop->resp[0] =	((response[2] & 0x003f) << (8 - 8)) |
+					((response[1] & 0xffff) << (14 - 8)) |
+					((response[0] & 0x03ff) << (30 - 8));
+		data->stop->resp[1] =	((response[0] & 0xfc00) >> 10);
+		data->stop->resp[2] = 0;
+		data->stop->resp[3] = 0;
+
+		if (err_status & MVSD_ERR_AUTOCMD12) {
+			u32 err_cmd12 = mvsd_read(MVSD_AUTOCMD12_ERR_STATUS);
+			dev_dbg(host->dev, "c12err 0x%04x\n", err_cmd12);
+			if (err_cmd12 & MVSD_AUTOCMD12_ERR_NOTEXE)
+				data->stop->error = -ENOEXEC;
+			else if (err_cmd12 & MVSD_AUTOCMD12_ERR_TIMEOUT)
+				data->stop->error = -ETIMEDOUT;
+			else if (err_cmd12)
+				data->stop->error = -EILSEQ;
+			err_status &= ~MVSD_ERR_AUTOCMD12;
+		}
+	}
+
+	return err_status;
+}
+
+static irqreturn_t mvsd_irq(int irq, void *dev)
+{
+	struct mvsd_host *host = dev;
+	void __iomem *iobase = host->base;
+	u32 intr_status, intr_done_mask;
+	int irq_handled = 0;
+
+	intr_status = mvsd_read(MVSD_NOR_INTR_STATUS);
+	dev_dbg(host->dev, "intr 0x%04x intr_en 0x%04x hw_state 0x%04x\n",
+		intr_status, mvsd_read(MVSD_NOR_INTR_EN),
+		mvsd_read(MVSD_HW_STATE));
+
+	/*
+	 * It looks like, SDIO IP can issue one late, spurious irq
+	 * although all irqs should be disabled. To work around this,
+	 * bail out early, if we didn't expect any irqs to occur.
+	 */
+	if (!mvsd_read(MVSD_NOR_INTR_EN) && !mvsd_read(MVSD_ERR_INTR_EN)) {
+		dev_dbg(host->dev, "spurious irq detected intr 0x%04x intr_en 0x%04x erri 0x%04x erri_en 0x%04x\n",
+			mvsd_read(MVSD_NOR_INTR_STATUS),
+			mvsd_read(MVSD_NOR_INTR_EN),
+			mvsd_read(MVSD_ERR_INTR_STATUS),
+			mvsd_read(MVSD_ERR_INTR_EN));
+		return IRQ_HANDLED;
+	}
+
+	spin_lock(&host->lock);
+
+	/* PIO handling, if needed. Messy business... */
+	if (host->pio_size &&
+	    (intr_status & host->intr_en &
+	     (MVSD_NOR_RX_READY | MVSD_NOR_RX_FIFO_8W))) {
+		u16 *p = host->pio_ptr;
+		int s = host->pio_size;
+		while (s >= 32 && (intr_status & MVSD_NOR_RX_FIFO_8W)) {
+			readsw(iobase + MVSD_FIFO, p, 16);
+			p += 16;
+			s -= 32;
+			intr_status = mvsd_read(MVSD_NOR_INTR_STATUS);
+		}
+		/*
+		 * Normally we'd use < 32 here, but the RX_FIFO_8W bit
+		 * doesn't appear to assert when there is exactly 32 bytes
+		 * (8 words) left to fetch in a transfer.
+		 */
+		if (s <= 32) {
+			while (s >= 4 && (intr_status & MVSD_NOR_RX_READY)) {
+				put_unaligned(mvsd_read(MVSD_FIFO), p++);
+				put_unaligned(mvsd_read(MVSD_FIFO), p++);
+				s -= 4;
+				intr_status = mvsd_read(MVSD_NOR_INTR_STATUS);
+			}
+			if (s && s < 4 && (intr_status & MVSD_NOR_RX_READY)) {
+				u16 val[2] = {0, 0};
+				val[0] = mvsd_read(MVSD_FIFO);
+				val[1] = mvsd_read(MVSD_FIFO);
+				memcpy(p, ((void *)&val) + 4 - s, s);
+				s = 0;
+				intr_status = mvsd_read(MVSD_NOR_INTR_STATUS);
+			}
+			if (s == 0) {
+				host->intr_en &=
+				     ~(MVSD_NOR_RX_READY | MVSD_NOR_RX_FIFO_8W);
+				mvsd_write(MVSD_NOR_INTR_EN, host->intr_en);
+			} else if (host->intr_en & MVSD_NOR_RX_FIFO_8W) {
+				host->intr_en &= ~MVSD_NOR_RX_FIFO_8W;
+				host->intr_en |= MVSD_NOR_RX_READY;
+				mvsd_write(MVSD_NOR_INTR_EN, host->intr_en);
+			}
+		}
+		dev_dbg(host->dev, "pio %d intr 0x%04x hw_state 0x%04x\n",
+			s, intr_status, mvsd_read(MVSD_HW_STATE));
+		host->pio_ptr = p;
+		host->pio_size = s;
+		irq_handled = 1;
+	} else if (host->pio_size &&
+		   (intr_status & host->intr_en &
+		    (MVSD_NOR_TX_AVAIL | MVSD_NOR_TX_FIFO_8W))) {
+		u16 *p = host->pio_ptr;
+		int s = host->pio_size;
+		/*
+		 * The TX_FIFO_8W bit is unreliable. When set, bursting
+		 * 16 halfwords all at once in the FIFO drops data. Actually
+		 * TX_AVAIL does go off after only one word is pushed even if
+		 * TX_FIFO_8W remains set.
+		 */
+		while (s >= 4 && (intr_status & MVSD_NOR_TX_AVAIL)) {
+			mvsd_write(MVSD_FIFO, get_unaligned(p++));
+			mvsd_write(MVSD_FIFO, get_unaligned(p++));
+			s -= 4;
+			intr_status = mvsd_read(MVSD_NOR_INTR_STATUS);
+		}
+		if (s < 4) {
+			if (s && (intr_status & MVSD_NOR_TX_AVAIL)) {
+				u16 val[2] = {0, 0};
+				memcpy(((void *)&val) + 4 - s, p, s);
+				mvsd_write(MVSD_FIFO, val[0]);
+				mvsd_write(MVSD_FIFO, val[1]);
+				s = 0;
+				intr_status = mvsd_read(MVSD_NOR_INTR_STATUS);
+			}
+			if (s == 0) {
+				host->intr_en &=
+				     ~(MVSD_NOR_TX_AVAIL | MVSD_NOR_TX_FIFO_8W);
+				mvsd_write(MVSD_NOR_INTR_EN, host->intr_en);
+			}
+		}
+		dev_dbg(host->dev, "pio %d intr 0x%04x hw_state 0x%04x\n",
+			s, intr_status, mvsd_read(MVSD_HW_STATE));
+		host->pio_ptr = p;
+		host->pio_size = s;
+		irq_handled = 1;
+	}
+
+	mvsd_write(MVSD_NOR_INTR_STATUS, intr_status);
+
+	intr_done_mask = MVSD_NOR_CARD_INT | MVSD_NOR_RX_READY |
+			 MVSD_NOR_RX_FIFO_8W | MVSD_NOR_TX_FIFO_8W;
+	if (intr_status & host->intr_en & ~intr_done_mask) {
+		struct mmc_request *mrq = host->mrq;
+		struct mmc_command *cmd = mrq->cmd;
+		u32 err_status = 0;
+
+		del_timer(&host->timer);
+		host->mrq = NULL;
+
+		host->intr_en &= MVSD_NOR_CARD_INT;
+		mvsd_write(MVSD_NOR_INTR_EN, host->intr_en);
+		mvsd_write(MVSD_ERR_INTR_EN, 0);
+
+		spin_unlock(&host->lock);
+
+		if (intr_status & MVSD_NOR_UNEXP_RSP) {
+			cmd->error = -EPROTO;
+		} else if (intr_status & MVSD_NOR_ERROR) {
+			err_status = mvsd_read(MVSD_ERR_INTR_STATUS);
+			dev_dbg(host->dev, "err 0x%04x\n", err_status);
+		}
+
+		err_status = mvsd_finish_cmd(host, cmd, err_status);
+		if (mrq->data)
+			err_status = mvsd_finish_data(host, mrq->data, err_status);
+		if (err_status) {
+			dev_err(host->dev, "unhandled error status %#04x\n",
+				err_status);
+			cmd->error = -ENOMSG;
+		}
+
+		mmc_request_done(host->mmc, mrq);
+		irq_handled = 1;
+	} else
+		spin_unlock(&host->lock);
+
+	if (intr_status & MVSD_NOR_CARD_INT) {
+		mmc_signal_sdio_irq(host->mmc);
+		irq_handled = 1;
+	}
+
+	if (irq_handled)
+		return IRQ_HANDLED;
+
+	dev_err(host->dev, "unhandled interrupt status=0x%04x en=0x%04x pio=%d\n",
+		intr_status, host->intr_en, host->pio_size);
+	return IRQ_NONE;
+}
+
+static void mvsd_timeout_timer(struct timer_list *t)
+{
+	struct mvsd_host *host = from_timer(host, t, timer);
+	void __iomem *iobase = host->base;
+	struct mmc_request *mrq;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	mrq = host->mrq;
+	if (mrq) {
+		dev_err(host->dev, "Timeout waiting for hardware interrupt.\n");
+		dev_err(host->dev, "hw_state=0x%04x, intr_status=0x%04x intr_en=0x%04x\n",
+			mvsd_read(MVSD_HW_STATE),
+			mvsd_read(MVSD_NOR_INTR_STATUS),
+			mvsd_read(MVSD_NOR_INTR_EN));
+
+		host->mrq = NULL;
+
+		mvsd_write(MVSD_SW_RESET, MVSD_SW_RESET_NOW);
+
+		host->xfer_mode &= MVSD_XFER_MODE_INT_CHK_EN;
+		mvsd_write(MVSD_XFER_MODE, host->xfer_mode);
+
+		host->intr_en &= MVSD_NOR_CARD_INT;
+		mvsd_write(MVSD_NOR_INTR_EN, host->intr_en);
+		mvsd_write(MVSD_ERR_INTR_EN, 0);
+		mvsd_write(MVSD_ERR_INTR_STATUS, 0xffff);
+
+		mrq->cmd->error = -ETIMEDOUT;
+		mvsd_finish_cmd(host, mrq->cmd, 0);
+		if (mrq->data) {
+			mrq->data->error = -ETIMEDOUT;
+			mvsd_finish_data(host, mrq->data, 0);
+		}
+	}
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (mrq)
+		mmc_request_done(host->mmc, mrq);
+}
+
+static void mvsd_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct mvsd_host *host = mmc_priv(mmc);
+	void __iomem *iobase = host->base;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (enable) {
+		host->xfer_mode |= MVSD_XFER_MODE_INT_CHK_EN;
+		host->intr_en |= MVSD_NOR_CARD_INT;
+	} else {
+		host->xfer_mode &= ~MVSD_XFER_MODE_INT_CHK_EN;
+		host->intr_en &= ~MVSD_NOR_CARD_INT;
+	}
+	mvsd_write(MVSD_XFER_MODE, host->xfer_mode);
+	mvsd_write(MVSD_NOR_INTR_EN, host->intr_en);
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void mvsd_power_up(struct mvsd_host *host)
+{
+	void __iomem *iobase = host->base;
+	dev_dbg(host->dev, "power up\n");
+	mvsd_write(MVSD_NOR_INTR_EN, 0);
+	mvsd_write(MVSD_ERR_INTR_EN, 0);
+	mvsd_write(MVSD_SW_RESET, MVSD_SW_RESET_NOW);
+	mvsd_write(MVSD_XFER_MODE, 0);
+	mvsd_write(MVSD_NOR_STATUS_EN, 0xffff);
+	mvsd_write(MVSD_ERR_STATUS_EN, 0xffff);
+	mvsd_write(MVSD_NOR_INTR_STATUS, 0xffff);
+	mvsd_write(MVSD_ERR_INTR_STATUS, 0xffff);
+}
+
+static void mvsd_power_down(struct mvsd_host *host)
+{
+	void __iomem *iobase = host->base;
+	dev_dbg(host->dev, "power down\n");
+	mvsd_write(MVSD_NOR_INTR_EN, 0);
+	mvsd_write(MVSD_ERR_INTR_EN, 0);
+	mvsd_write(MVSD_SW_RESET, MVSD_SW_RESET_NOW);
+	mvsd_write(MVSD_XFER_MODE, MVSD_XFER_MODE_STOP_CLK);
+	mvsd_write(MVSD_NOR_STATUS_EN, 0);
+	mvsd_write(MVSD_ERR_STATUS_EN, 0);
+	mvsd_write(MVSD_NOR_INTR_STATUS, 0xffff);
+	mvsd_write(MVSD_ERR_INTR_STATUS, 0xffff);
+}
+
+static void mvsd_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct mvsd_host *host = mmc_priv(mmc);
+	void __iomem *iobase = host->base;
+	u32 ctrl_reg = 0;
+
+	if (ios->power_mode == MMC_POWER_UP)
+		mvsd_power_up(host);
+
+	if (ios->clock == 0) {
+		mvsd_write(MVSD_XFER_MODE, MVSD_XFER_MODE_STOP_CLK);
+		mvsd_write(MVSD_CLK_DIV, MVSD_BASE_DIV_MAX);
+		host->clock = 0;
+		dev_dbg(host->dev, "clock off\n");
+	} else if (ios->clock != host->clock) {
+		u32 m = DIV_ROUND_UP(host->base_clock, ios->clock) - 1;
+		if (m > MVSD_BASE_DIV_MAX)
+			m = MVSD_BASE_DIV_MAX;
+		mvsd_write(MVSD_CLK_DIV, m);
+		host->clock = ios->clock;
+		host->ns_per_clk = 1000000000 / (host->base_clock / (m+1));
+		dev_dbg(host->dev, "clock=%d (%d), div=0x%04x\n",
+			ios->clock, host->base_clock / (m+1), m);
+	}
+
+	/* default transfer mode */
+	ctrl_reg |= MVSD_HOST_CTRL_BIG_ENDIAN;
+	ctrl_reg &= ~MVSD_HOST_CTRL_LSB_FIRST;
+
+	/* default to maximum timeout */
+	ctrl_reg |= MVSD_HOST_CTRL_TMOUT_MASK;
+	ctrl_reg |= MVSD_HOST_CTRL_TMOUT_EN;
+
+	if (ios->bus_mode == MMC_BUSMODE_PUSHPULL)
+		ctrl_reg |= MVSD_HOST_CTRL_PUSH_PULL_EN;
+
+	if (ios->bus_width == MMC_BUS_WIDTH_4)
+		ctrl_reg |= MVSD_HOST_CTRL_DATA_WIDTH_4_BITS;
+
+	/*
+	 * The HI_SPEED_EN bit is causing trouble with many (but not all)
+	 * high speed SD, SDHC and SDIO cards.  Not enabling that bit
+	 * makes all cards work.  So let's just ignore that bit for now
+	 * and revisit this issue if problems for not enabling this bit
+	 * are ever reported.
+	 */
+#if 0
+	if (ios->timing == MMC_TIMING_MMC_HS ||
+	    ios->timing == MMC_TIMING_SD_HS)
+		ctrl_reg |= MVSD_HOST_CTRL_HI_SPEED_EN;
+#endif
+
+	host->ctrl = ctrl_reg;
+	mvsd_write(MVSD_HOST_CTRL, ctrl_reg);
+	dev_dbg(host->dev, "ctrl 0x%04x: %s %s %s\n", ctrl_reg,
+		(ctrl_reg & MVSD_HOST_CTRL_PUSH_PULL_EN) ?
+			"push-pull" : "open-drain",
+		(ctrl_reg & MVSD_HOST_CTRL_DATA_WIDTH_4_BITS) ?
+			"4bit-width" : "1bit-width",
+		(ctrl_reg & MVSD_HOST_CTRL_HI_SPEED_EN) ?
+			"high-speed" : "");
+
+	if (ios->power_mode == MMC_POWER_OFF)
+		mvsd_power_down(host);
+}
+
+static const struct mmc_host_ops mvsd_ops = {
+	.request		= mvsd_request,
+	.get_ro			= mmc_gpio_get_ro,
+	.set_ios		= mvsd_set_ios,
+	.enable_sdio_irq	= mvsd_enable_sdio_irq,
+};
+
+static void
+mv_conf_mbus_windows(struct mvsd_host *host,
+		     const struct mbus_dram_target_info *dram)
+{
+	void __iomem *iobase = host->base;
+	int i;
+
+	for (i = 0; i < 4; i++) {
+		writel(0, iobase + MVSD_WINDOW_CTRL(i));
+		writel(0, iobase + MVSD_WINDOW_BASE(i));
+	}
+
+	for (i = 0; i < dram->num_cs; i++) {
+		const struct mbus_dram_window *cs = dram->cs + i;
+		writel(((cs->size - 1) & 0xffff0000) |
+		       (cs->mbus_attr << 8) |
+		       (dram->mbus_dram_target_id << 4) | 1,
+		       iobase + MVSD_WINDOW_CTRL(i));
+		writel(cs->base, iobase + MVSD_WINDOW_BASE(i));
+	}
+}
+
+static int mvsd_probe(struct platform_device *pdev)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct mmc_host *mmc = NULL;
+	struct mvsd_host *host = NULL;
+	const struct mbus_dram_target_info *dram;
+	int ret, irq;
+
+	if (!np) {
+		dev_err(&pdev->dev, "no DT node\n");
+		return -ENODEV;
+	}
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return irq;
+
+	mmc = mmc_alloc_host(sizeof(struct mvsd_host), &pdev->dev);
+	if (!mmc) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+	host->dev = &pdev->dev;
+
+	/*
+	 * Some non-DT platforms do not pass a clock, and the clock
+	 * frequency is passed through platform_data. On DT platforms,
+	 * a clock must always be passed, even if there is no gatable
+	 * clock associated to the SDIO interface (it can simply be a
+	 * fixed rate clock).
+	 */
+	host->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(host->clk)) {
+		dev_err(&pdev->dev, "no clock associated\n");
+		ret = -EINVAL;
+		goto out;
+	}
+	clk_prepare_enable(host->clk);
+
+	mmc->ops = &mvsd_ops;
+
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+
+	mmc->f_min = DIV_ROUND_UP(host->base_clock, MVSD_BASE_DIV_MAX);
+	mmc->f_max = MVSD_CLOCKRATE_MAX;
+
+	mmc->max_blk_size = 2048;
+	mmc->max_blk_count = 65535;
+
+	mmc->max_segs = 1;
+	mmc->max_seg_size = mmc->max_blk_size * mmc->max_blk_count;
+	mmc->max_req_size = mmc->max_blk_size * mmc->max_blk_count;
+
+	host->base_clock = clk_get_rate(host->clk) / 2;
+	ret = mmc_of_parse(mmc);
+	if (ret < 0)
+		goto out;
+	if (maxfreq)
+		mmc->f_max = maxfreq;
+
+	spin_lock_init(&host->lock);
+
+	host->base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(host->base)) {
+		ret = PTR_ERR(host->base);
+		goto out;
+	}
+
+	/* (Re-)program MBUS remapping windows if we are asked to. */
+	dram = mv_mbus_dram_info();
+	if (dram)
+		mv_conf_mbus_windows(host, dram);
+
+	mvsd_power_down(host);
+
+	ret = devm_request_irq(&pdev->dev, irq, mvsd_irq, 0, DRIVER_NAME, host);
+	if (ret) {
+		dev_err(&pdev->dev, "cannot assign irq %d\n", irq);
+		goto out;
+	}
+
+	timer_setup(&host->timer, mvsd_timeout_timer, 0);
+	platform_set_drvdata(pdev, mmc);
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto out;
+
+	if (!(mmc->caps & MMC_CAP_NEEDS_POLL))
+		dev_dbg(&pdev->dev, "using GPIO for card detection\n");
+	else
+		dev_dbg(&pdev->dev, "lacking card detect (fall back to polling)\n");
+
+	return 0;
+
+out:
+	if (mmc) {
+		if (!IS_ERR(host->clk))
+			clk_disable_unprepare(host->clk);
+		mmc_free_host(mmc);
+	}
+
+	return ret;
+}
+
+static int mvsd_remove(struct platform_device *pdev)
+{
+	struct mmc_host *mmc = platform_get_drvdata(pdev);
+
+	struct mvsd_host *host = mmc_priv(mmc);
+
+	mmc_remove_host(mmc);
+	del_timer_sync(&host->timer);
+	mvsd_power_down(host);
+
+	if (!IS_ERR(host->clk))
+		clk_disable_unprepare(host->clk);
+	mmc_free_host(mmc);
+
+	return 0;
+}
+
+static const struct of_device_id mvsdio_dt_ids[] = {
+	{ .compatible = "marvell,orion-sdio" },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, mvsdio_dt_ids);
+
+static struct platform_driver mvsd_driver = {
+	.probe		= mvsd_probe,
+	.remove		= mvsd_remove,
+	.driver		= {
+		.name	= DRIVER_NAME,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = mvsdio_dt_ids,
+	},
+};
+
+module_platform_driver(mvsd_driver);
+
+/* maximum card clock frequency (default 50MHz) */
+module_param(maxfreq, int, 0);
+
+/* force PIO transfers all the time */
+module_param(nodma, int, 0);
+
+MODULE_AUTHOR("Maen Suleiman, Nicolas Pitre");
+MODULE_DESCRIPTION("Marvell MMC,SD,SDIO Host Controller driver");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:mvsdio");
diff --git a/drivers/mmc/host/mvsdio.h b/drivers/mmc/host/mvsdio.h
new file mode 100644
index 000000000..2f1458ac6
--- /dev/null
+++ b/drivers/mmc/host/mvsdio.h
@@ -0,0 +1,187 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ *  Copyright (C) 2008 Marvell Semiconductors, All Rights Reserved.
+ */
+
+#ifndef __MVSDIO_H
+#define __MVSDIO_H
+
+/*
+ * Clock rates
+ */
+
+#define MVSD_CLOCKRATE_MAX			50000000
+#define MVSD_BASE_DIV_MAX			0x7ff
+
+
+/*
+ * Register offsets
+ */
+
+#define MVSD_SYS_ADDR_LOW			0x000
+#define MVSD_SYS_ADDR_HI			0x004
+#define MVSD_BLK_SIZE				0x008
+#define MVSD_BLK_COUNT				0x00c
+#define MVSD_ARG_LOW				0x010
+#define MVSD_ARG_HI				0x014
+#define MVSD_XFER_MODE				0x018
+#define MVSD_CMD				0x01c
+#define MVSD_RSP(i)				(0x020 + ((i)<<2))
+#define MVSD_RSP0				0x020
+#define MVSD_RSP1				0x024
+#define MVSD_RSP2				0x028
+#define MVSD_RSP3				0x02c
+#define MVSD_RSP4				0x030
+#define MVSD_RSP5				0x034
+#define MVSD_RSP6				0x038
+#define MVSD_RSP7				0x03c
+#define MVSD_FIFO				0x040
+#define MVSD_RSP_CRC7				0x044
+#define MVSD_HW_STATE				0x048
+#define MVSD_HOST_CTRL				0x050
+#define MVSD_BLK_GAP_CTRL			0x054
+#define MVSD_CLK_CTRL				0x058
+#define MVSD_SW_RESET				0x05c
+#define MVSD_NOR_INTR_STATUS			0x060
+#define MVSD_ERR_INTR_STATUS			0x064
+#define MVSD_NOR_STATUS_EN			0x068
+#define MVSD_ERR_STATUS_EN			0x06c
+#define MVSD_NOR_INTR_EN			0x070
+#define MVSD_ERR_INTR_EN			0x074
+#define MVSD_AUTOCMD12_ERR_STATUS		0x078
+#define MVSD_CURR_BYTE_LEFT			0x07c
+#define MVSD_CURR_BLK_LEFT			0x080
+#define MVSD_AUTOCMD12_ARG_LOW			0x084
+#define MVSD_AUTOCMD12_ARG_HI			0x088
+#define MVSD_AUTOCMD12_CMD			0x08c
+#define MVSD_AUTO_RSP(i)			(0x090 + ((i)<<2))
+#define MVSD_AUTO_RSP0				0x090
+#define MVSD_AUTO_RSP1				0x094
+#define MVSD_AUTO_RSP2				0x098
+#define MVSD_CLK_DIV				0x128
+
+#define MVSD_WINDOW_CTRL(i)			(0x108 + ((i) << 3))
+#define MVSD_WINDOW_BASE(i)			(0x10c + ((i) << 3))
+
+
+/*
+ * MVSD_CMD
+ */
+
+#define MVSD_CMD_RSP_NONE			(0 << 0)
+#define MVSD_CMD_RSP_136			(1 << 0)
+#define MVSD_CMD_RSP_48				(2 << 0)
+#define MVSD_CMD_RSP_48BUSY			(3 << 0)
+
+#define MVSD_CMD_CHECK_DATACRC16		(1 << 2)
+#define MVSD_CMD_CHECK_CMDCRC			(1 << 3)
+#define MVSD_CMD_INDX_CHECK			(1 << 4)
+#define MVSD_CMD_DATA_PRESENT			(1 << 5)
+#define MVSD_UNEXPECTED_RESP			(1 << 7)
+#define MVSD_CMD_INDEX(x)			((x) << 8)
+
+
+/*
+ * MVSD_AUTOCMD12_CMD
+ */
+
+#define MVSD_AUTOCMD12_BUSY			(1 << 0)
+#define MVSD_AUTOCMD12_INDX_CHECK		(1 << 1)
+#define MVSD_AUTOCMD12_INDEX(x)			((x) << 8)
+
+/*
+ * MVSD_XFER_MODE
+ */
+
+#define MVSD_XFER_MODE_WR_DATA_START		(1 << 0)
+#define MVSD_XFER_MODE_HW_WR_DATA_EN		(1 << 1)
+#define MVSD_XFER_MODE_AUTO_CMD12		(1 << 2)
+#define MVSD_XFER_MODE_INT_CHK_EN		(1 << 3)
+#define MVSD_XFER_MODE_TO_HOST			(1 << 4)
+#define MVSD_XFER_MODE_STOP_CLK			(1 << 5)
+#define MVSD_XFER_MODE_PIO			(1 << 6)
+
+
+/*
+ * MVSD_HOST_CTRL
+ */
+
+#define MVSD_HOST_CTRL_PUSH_PULL_EN 		(1 << 0)
+
+#define MVSD_HOST_CTRL_CARD_TYPE_MEM_ONLY 	(0 << 1)
+#define MVSD_HOST_CTRL_CARD_TYPE_IO_ONLY 	(1 << 1)
+#define MVSD_HOST_CTRL_CARD_TYPE_IO_MEM_COMBO 	(2 << 1)
+#define MVSD_HOST_CTRL_CARD_TYPE_IO_MMC 	(3 << 1)
+#define MVSD_HOST_CTRL_CARD_TYPE_MASK	 	(3 << 1)
+
+#define MVSD_HOST_CTRL_BIG_ENDIAN 		(1 << 3)
+#define MVSD_HOST_CTRL_LSB_FIRST 		(1 << 4)
+#define MVSD_HOST_CTRL_DATA_WIDTH_4_BITS 	(1 << 9)
+#define MVSD_HOST_CTRL_HI_SPEED_EN 		(1 << 10)
+
+#define MVSD_HOST_CTRL_TMOUT_MAX 		0xf
+#define MVSD_HOST_CTRL_TMOUT_MASK 		(0xf << 11)
+#define MVSD_HOST_CTRL_TMOUT(x) 		((x) << 11)
+#define MVSD_HOST_CTRL_TMOUT_EN 		(1 << 15)
+
+
+/*
+ * MVSD_SW_RESET
+ */
+
+#define MVSD_SW_RESET_NOW			(1 << 8)
+
+
+/*
+ * Normal interrupt status bits
+ */
+
+#define MVSD_NOR_CMD_DONE			(1 << 0)
+#define MVSD_NOR_XFER_DONE			(1 << 1)
+#define MVSD_NOR_BLK_GAP_EVT			(1 << 2)
+#define MVSD_NOR_DMA_DONE			(1 << 3)
+#define MVSD_NOR_TX_AVAIL			(1 << 4)
+#define MVSD_NOR_RX_READY			(1 << 5)
+#define MVSD_NOR_CARD_INT			(1 << 8)
+#define MVSD_NOR_READ_WAIT_ON			(1 << 9)
+#define MVSD_NOR_RX_FIFO_8W			(1 << 10)
+#define MVSD_NOR_TX_FIFO_8W			(1 << 11)
+#define MVSD_NOR_SUSPEND_ON			(1 << 12)
+#define MVSD_NOR_AUTOCMD12_DONE			(1 << 13)
+#define MVSD_NOR_UNEXP_RSP			(1 << 14)
+#define MVSD_NOR_ERROR				(1 << 15)
+
+
+/*
+ * Error status bits
+ */
+
+#define MVSD_ERR_CMD_TIMEOUT			(1 << 0)
+#define MVSD_ERR_CMD_CRC			(1 << 1)
+#define MVSD_ERR_CMD_ENDBIT			(1 << 2)
+#define MVSD_ERR_CMD_INDEX			(1 << 3)
+#define MVSD_ERR_DATA_TIMEOUT			(1 << 4)
+#define MVSD_ERR_DATA_CRC			(1 << 5)
+#define MVSD_ERR_DATA_ENDBIT			(1 << 6)
+#define MVSD_ERR_AUTOCMD12			(1 << 8)
+#define MVSD_ERR_CMD_STARTBIT			(1 << 9)
+#define MVSD_ERR_XFER_SIZE			(1 << 10)
+#define MVSD_ERR_RESP_T_BIT			(1 << 11)
+#define MVSD_ERR_CRC_ENDBIT			(1 << 12)
+#define MVSD_ERR_CRC_STARTBIT			(1 << 13)
+#define MVSD_ERR_CRC_STATUS			(1 << 14)
+
+
+/*
+ * CMD12 error status bits
+ */
+
+#define MVSD_AUTOCMD12_ERR_NOTEXE		(1 << 0)
+#define MVSD_AUTOCMD12_ERR_TIMEOUT		(1 << 1)
+#define MVSD_AUTOCMD12_ERR_CRC			(1 << 2)
+#define MVSD_AUTOCMD12_ERR_ENDBIT		(1 << 3)
+#define MVSD_AUTOCMD12_ERR_INDEX		(1 << 4)
+#define MVSD_AUTOCMD12_ERR_RESP_T_BIT		(1 << 5)
+#define MVSD_AUTOCMD12_ERR_RESP_STARTBIT	(1 << 6)
+
+#endif
diff --git a/drivers/mmc/host/mxcmmc.c b/drivers/mmc/host/mxcmmc.c
new file mode 100644
index 000000000..93a105b07
--- /dev/null
+++ b/drivers/mmc/host/mxcmmc.c
@@ -0,0 +1,1260 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ *  linux/drivers/mmc/host/mxcmmc.c - Freescale i.MX MMCI driver
+ *
+ *  This is a driver for the SDHC controller found in Freescale MX2/MX3
+ *  SoCs. It is basically the same hardware as found on MX1 (imxmmc.c).
+ *  Unlike the hardware found on MX1, this hardware just works and does
+ *  not need all the quirks found in imxmmc.c, hence the separate driver.
+ *
+ *  Copyright (C) 2008 Sascha Hauer, Pengutronix <s.hauer@pengutronix.de>
+ *  Copyright (C) 2006 Pavel Pisa, PiKRON <ppisa@pikron.com>
+ *
+ *  derived from pxamci.c by Russell King
+ */
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/ioport.h>
+#include <linux/platform_device.h>
+#include <linux/highmem.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/blkdev.h>
+#include <linux/dma-mapping.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/card.h>
+#include <linux/delay.h>
+#include <linux/clk.h>
+#include <linux/io.h>
+#include <linux/regulator/consumer.h>
+#include <linux/dmaengine.h>
+#include <linux/types.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/of_dma.h>
+#include <linux/mmc/slot-gpio.h>
+
+#include <asm/dma.h>
+#include <asm/irq.h>
+#include <linux/platform_data/mmc-mxcmmc.h>
+
+#include <linux/platform_data/dma-imx.h>
+
+#define DRIVER_NAME "mxc-mmc"
+#define MXCMCI_TIMEOUT_MS 10000
+
+#define MMC_REG_STR_STP_CLK		0x00
+#define MMC_REG_STATUS			0x04
+#define MMC_REG_CLK_RATE		0x08
+#define MMC_REG_CMD_DAT_CONT		0x0C
+#define MMC_REG_RES_TO			0x10
+#define MMC_REG_READ_TO			0x14
+#define MMC_REG_BLK_LEN			0x18
+#define MMC_REG_NOB			0x1C
+#define MMC_REG_REV_NO			0x20
+#define MMC_REG_INT_CNTR		0x24
+#define MMC_REG_CMD			0x28
+#define MMC_REG_ARG			0x2C
+#define MMC_REG_RES_FIFO		0x34
+#define MMC_REG_BUFFER_ACCESS		0x38
+
+#define STR_STP_CLK_RESET               (1 << 3)
+#define STR_STP_CLK_START_CLK           (1 << 1)
+#define STR_STP_CLK_STOP_CLK            (1 << 0)
+
+#define STATUS_CARD_INSERTION		(1 << 31)
+#define STATUS_CARD_REMOVAL		(1 << 30)
+#define STATUS_YBUF_EMPTY		(1 << 29)
+#define STATUS_XBUF_EMPTY		(1 << 28)
+#define STATUS_YBUF_FULL		(1 << 27)
+#define STATUS_XBUF_FULL		(1 << 26)
+#define STATUS_BUF_UND_RUN		(1 << 25)
+#define STATUS_BUF_OVFL			(1 << 24)
+#define STATUS_SDIO_INT_ACTIVE		(1 << 14)
+#define STATUS_END_CMD_RESP		(1 << 13)
+#define STATUS_WRITE_OP_DONE		(1 << 12)
+#define STATUS_DATA_TRANS_DONE		(1 << 11)
+#define STATUS_READ_OP_DONE		(1 << 11)
+#define STATUS_WR_CRC_ERROR_CODE_MASK	(3 << 10)
+#define STATUS_CARD_BUS_CLK_RUN		(1 << 8)
+#define STATUS_BUF_READ_RDY		(1 << 7)
+#define STATUS_BUF_WRITE_RDY		(1 << 6)
+#define STATUS_RESP_CRC_ERR		(1 << 5)
+#define STATUS_CRC_READ_ERR		(1 << 3)
+#define STATUS_CRC_WRITE_ERR		(1 << 2)
+#define STATUS_TIME_OUT_RESP		(1 << 1)
+#define STATUS_TIME_OUT_READ		(1 << 0)
+#define STATUS_ERR_MASK			0x2f
+
+#define CMD_DAT_CONT_CMD_RESP_LONG_OFF	(1 << 12)
+#define CMD_DAT_CONT_STOP_READWAIT	(1 << 11)
+#define CMD_DAT_CONT_START_READWAIT	(1 << 10)
+#define CMD_DAT_CONT_BUS_WIDTH_4	(2 << 8)
+#define CMD_DAT_CONT_INIT		(1 << 7)
+#define CMD_DAT_CONT_WRITE		(1 << 4)
+#define CMD_DAT_CONT_DATA_ENABLE	(1 << 3)
+#define CMD_DAT_CONT_RESPONSE_48BIT_CRC	(1 << 0)
+#define CMD_DAT_CONT_RESPONSE_136BIT	(2 << 0)
+#define CMD_DAT_CONT_RESPONSE_48BIT	(3 << 0)
+
+#define INT_SDIO_INT_WKP_EN		(1 << 18)
+#define INT_CARD_INSERTION_WKP_EN	(1 << 17)
+#define INT_CARD_REMOVAL_WKP_EN		(1 << 16)
+#define INT_CARD_INSERTION_EN		(1 << 15)
+#define INT_CARD_REMOVAL_EN		(1 << 14)
+#define INT_SDIO_IRQ_EN			(1 << 13)
+#define INT_DAT0_EN			(1 << 12)
+#define INT_BUF_READ_EN			(1 << 4)
+#define INT_BUF_WRITE_EN		(1 << 3)
+#define INT_END_CMD_RES_EN		(1 << 2)
+#define INT_WRITE_OP_DONE_EN		(1 << 1)
+#define INT_READ_OP_EN			(1 << 0)
+
+enum mxcmci_type {
+	IMX21_MMC,
+	IMX31_MMC,
+	MPC512X_MMC,
+};
+
+struct mxcmci_host {
+	struct mmc_host		*mmc;
+	void __iomem		*base;
+	dma_addr_t		phys_base;
+	int			detect_irq;
+	struct dma_chan		*dma;
+	struct dma_async_tx_descriptor *desc;
+	int			do_dma;
+	int			default_irq_mask;
+	int			use_sdio;
+	unsigned int		power_mode;
+	struct imxmmc_platform_data *pdata;
+
+	struct mmc_request	*req;
+	struct mmc_command	*cmd;
+	struct mmc_data		*data;
+
+	unsigned int		datasize;
+	unsigned int		dma_dir;
+
+	u16			rev_no;
+	unsigned int		cmdat;
+
+	struct clk		*clk_ipg;
+	struct clk		*clk_per;
+
+	int			clock;
+
+	struct work_struct	datawork;
+	spinlock_t		lock;
+
+	int			burstlen;
+	int			dmareq;
+	struct dma_slave_config dma_slave_config;
+	struct imx_dma_data	dma_data;
+
+	struct timer_list	watchdog;
+	enum mxcmci_type	devtype;
+};
+
+static const struct platform_device_id mxcmci_devtype[] = {
+	{
+		.name = "imx21-mmc",
+		.driver_data = IMX21_MMC,
+	}, {
+		.name = "imx31-mmc",
+		.driver_data = IMX31_MMC,
+	}, {
+		.name = "mpc512x-sdhc",
+		.driver_data = MPC512X_MMC,
+	}, {
+		/* sentinel */
+	}
+};
+MODULE_DEVICE_TABLE(platform, mxcmci_devtype);
+
+static const struct of_device_id mxcmci_of_match[] = {
+	{
+		.compatible = "fsl,imx21-mmc",
+		.data = &mxcmci_devtype[IMX21_MMC],
+	}, {
+		.compatible = "fsl,imx31-mmc",
+		.data = &mxcmci_devtype[IMX31_MMC],
+	}, {
+		.compatible = "fsl,mpc5121-sdhc",
+		.data = &mxcmci_devtype[MPC512X_MMC],
+	}, {
+		/* sentinel */
+	}
+};
+MODULE_DEVICE_TABLE(of, mxcmci_of_match);
+
+static inline int is_imx31_mmc(struct mxcmci_host *host)
+{
+	return host->devtype == IMX31_MMC;
+}
+
+static inline int is_mpc512x_mmc(struct mxcmci_host *host)
+{
+	return host->devtype == MPC512X_MMC;
+}
+
+static inline u32 mxcmci_readl(struct mxcmci_host *host, int reg)
+{
+	if (IS_ENABLED(CONFIG_PPC_MPC512x))
+		return ioread32be(host->base + reg);
+	else
+		return readl(host->base + reg);
+}
+
+static inline void mxcmci_writel(struct mxcmci_host *host, u32 val, int reg)
+{
+	if (IS_ENABLED(CONFIG_PPC_MPC512x))
+		iowrite32be(val, host->base + reg);
+	else
+		writel(val, host->base + reg);
+}
+
+static inline u16 mxcmci_readw(struct mxcmci_host *host, int reg)
+{
+	if (IS_ENABLED(CONFIG_PPC_MPC512x))
+		return ioread32be(host->base + reg);
+	else
+		return readw(host->base + reg);
+}
+
+static inline void mxcmci_writew(struct mxcmci_host *host, u16 val, int reg)
+{
+	if (IS_ENABLED(CONFIG_PPC_MPC512x))
+		iowrite32be(val, host->base + reg);
+	else
+		writew(val, host->base + reg);
+}
+
+static void mxcmci_set_clk_rate(struct mxcmci_host *host, unsigned int clk_ios);
+
+static void mxcmci_set_power(struct mxcmci_host *host, unsigned int vdd)
+{
+	if (!IS_ERR(host->mmc->supply.vmmc)) {
+		if (host->power_mode == MMC_POWER_UP)
+			mmc_regulator_set_ocr(host->mmc,
+					      host->mmc->supply.vmmc, vdd);
+		else if (host->power_mode == MMC_POWER_OFF)
+			mmc_regulator_set_ocr(host->mmc,
+					      host->mmc->supply.vmmc, 0);
+	}
+
+	if (host->pdata && host->pdata->setpower)
+		host->pdata->setpower(mmc_dev(host->mmc), vdd);
+}
+
+static inline int mxcmci_use_dma(struct mxcmci_host *host)
+{
+	return host->do_dma;
+}
+
+static void mxcmci_softreset(struct mxcmci_host *host)
+{
+	int i;
+
+	dev_dbg(mmc_dev(host->mmc), "mxcmci_softreset\n");
+
+	/* reset sequence */
+	mxcmci_writew(host, STR_STP_CLK_RESET, MMC_REG_STR_STP_CLK);
+	mxcmci_writew(host, STR_STP_CLK_RESET | STR_STP_CLK_START_CLK,
+			MMC_REG_STR_STP_CLK);
+
+	for (i = 0; i < 8; i++)
+		mxcmci_writew(host, STR_STP_CLK_START_CLK, MMC_REG_STR_STP_CLK);
+
+	mxcmci_writew(host, 0xff, MMC_REG_RES_TO);
+}
+
+#if IS_ENABLED(CONFIG_PPC_MPC512x)
+static inline void buffer_swap32(u32 *buf, int len)
+{
+	int i;
+
+	for (i = 0; i < ((len + 3) / 4); i++) {
+		*buf = swab32(*buf);
+		buf++;
+	}
+}
+
+static void mxcmci_swap_buffers(struct mmc_data *data)
+{
+	struct scatterlist *sg;
+	int i;
+
+	for_each_sg(data->sg, sg, data->sg_len, i)
+		buffer_swap32(sg_virt(sg), sg->length);
+}
+#else
+static inline void mxcmci_swap_buffers(struct mmc_data *data) {}
+#endif
+
+static int mxcmci_setup_data(struct mxcmci_host *host, struct mmc_data *data)
+{
+	unsigned int nob = data->blocks;
+	unsigned int blksz = data->blksz;
+	unsigned int datasize = nob * blksz;
+	struct scatterlist *sg;
+	enum dma_transfer_direction slave_dirn;
+	int i, nents;
+
+	host->data = data;
+	data->bytes_xfered = 0;
+
+	mxcmci_writew(host, nob, MMC_REG_NOB);
+	mxcmci_writew(host, blksz, MMC_REG_BLK_LEN);
+	host->datasize = datasize;
+
+	if (!mxcmci_use_dma(host))
+		return 0;
+
+	for_each_sg(data->sg, sg, data->sg_len, i) {
+		if (sg->offset & 3 || sg->length & 3 || sg->length < 512) {
+			host->do_dma = 0;
+			return 0;
+		}
+	}
+
+	if (data->flags & MMC_DATA_READ) {
+		host->dma_dir = DMA_FROM_DEVICE;
+		slave_dirn = DMA_DEV_TO_MEM;
+	} else {
+		host->dma_dir = DMA_TO_DEVICE;
+		slave_dirn = DMA_MEM_TO_DEV;
+
+		mxcmci_swap_buffers(data);
+	}
+
+	nents = dma_map_sg(host->dma->device->dev, data->sg,
+				     data->sg_len,  host->dma_dir);
+	if (nents != data->sg_len)
+		return -EINVAL;
+
+	host->desc = dmaengine_prep_slave_sg(host->dma,
+		data->sg, data->sg_len, slave_dirn,
+		DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+
+	if (!host->desc) {
+		dma_unmap_sg(host->dma->device->dev, data->sg, data->sg_len,
+				host->dma_dir);
+		host->do_dma = 0;
+		return 0; /* Fall back to PIO */
+	}
+	wmb();
+
+	dmaengine_submit(host->desc);
+	dma_async_issue_pending(host->dma);
+
+	mod_timer(&host->watchdog, jiffies + msecs_to_jiffies(MXCMCI_TIMEOUT_MS));
+
+	return 0;
+}
+
+static void mxcmci_cmd_done(struct mxcmci_host *host, unsigned int stat);
+static void mxcmci_data_done(struct mxcmci_host *host, unsigned int stat);
+
+static void mxcmci_dma_callback(void *data)
+{
+	struct mxcmci_host *host = data;
+	u32 stat;
+
+	del_timer(&host->watchdog);
+
+	stat = mxcmci_readl(host, MMC_REG_STATUS);
+
+	dev_dbg(mmc_dev(host->mmc), "%s: 0x%08x\n", __func__, stat);
+
+	mxcmci_data_done(host, stat);
+}
+
+static int mxcmci_start_cmd(struct mxcmci_host *host, struct mmc_command *cmd,
+		unsigned int cmdat)
+{
+	u32 int_cntr = host->default_irq_mask;
+	unsigned long flags;
+
+	WARN_ON(host->cmd != NULL);
+	host->cmd = cmd;
+
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_R1: /* short CRC, OPCODE */
+	case MMC_RSP_R1B:/* short CRC, OPCODE, BUSY */
+		cmdat |= CMD_DAT_CONT_RESPONSE_48BIT_CRC;
+		break;
+	case MMC_RSP_R2: /* long 136 bit + CRC */
+		cmdat |= CMD_DAT_CONT_RESPONSE_136BIT;
+		break;
+	case MMC_RSP_R3: /* short */
+		cmdat |= CMD_DAT_CONT_RESPONSE_48BIT;
+		break;
+	case MMC_RSP_NONE:
+		break;
+	default:
+		dev_err(mmc_dev(host->mmc), "unhandled response type 0x%x\n",
+				mmc_resp_type(cmd));
+		cmd->error = -EINVAL;
+		return -EINVAL;
+	}
+
+	int_cntr = INT_END_CMD_RES_EN;
+
+	if (mxcmci_use_dma(host)) {
+		if (host->dma_dir == DMA_FROM_DEVICE) {
+			host->desc->callback = mxcmci_dma_callback;
+			host->desc->callback_param = host;
+		} else {
+			int_cntr |= INT_WRITE_OP_DONE_EN;
+		}
+	}
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (host->use_sdio)
+		int_cntr |= INT_SDIO_IRQ_EN;
+	mxcmci_writel(host, int_cntr, MMC_REG_INT_CNTR);
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	mxcmci_writew(host, cmd->opcode, MMC_REG_CMD);
+	mxcmci_writel(host, cmd->arg, MMC_REG_ARG);
+	mxcmci_writew(host, cmdat, MMC_REG_CMD_DAT_CONT);
+
+	return 0;
+}
+
+static void mxcmci_finish_request(struct mxcmci_host *host,
+		struct mmc_request *req)
+{
+	u32 int_cntr = host->default_irq_mask;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (host->use_sdio)
+		int_cntr |= INT_SDIO_IRQ_EN;
+	mxcmci_writel(host, int_cntr, MMC_REG_INT_CNTR);
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	host->req = NULL;
+	host->cmd = NULL;
+	host->data = NULL;
+
+	mmc_request_done(host->mmc, req);
+}
+
+static int mxcmci_finish_data(struct mxcmci_host *host, unsigned int stat)
+{
+	struct mmc_data *data = host->data;
+	int data_error;
+
+	if (mxcmci_use_dma(host)) {
+		dma_unmap_sg(host->dma->device->dev, data->sg, data->sg_len,
+				host->dma_dir);
+		mxcmci_swap_buffers(data);
+	}
+
+	if (stat & STATUS_ERR_MASK) {
+		dev_dbg(mmc_dev(host->mmc), "request failed. status: 0x%08x\n",
+				stat);
+		if (stat & STATUS_CRC_READ_ERR) {
+			dev_err(mmc_dev(host->mmc), "%s: -EILSEQ\n", __func__);
+			data->error = -EILSEQ;
+		} else if (stat & STATUS_CRC_WRITE_ERR) {
+			u32 err_code = (stat >> 9) & 0x3;
+			if (err_code == 2) { /* No CRC response */
+				dev_err(mmc_dev(host->mmc),
+					"%s: No CRC -ETIMEDOUT\n", __func__);
+				data->error = -ETIMEDOUT;
+			} else {
+				dev_err(mmc_dev(host->mmc),
+					"%s: -EILSEQ\n", __func__);
+				data->error = -EILSEQ;
+			}
+		} else if (stat & STATUS_TIME_OUT_READ) {
+			dev_err(mmc_dev(host->mmc),
+				"%s: read -ETIMEDOUT\n", __func__);
+			data->error = -ETIMEDOUT;
+		} else {
+			dev_err(mmc_dev(host->mmc), "%s: -EIO\n", __func__);
+			data->error = -EIO;
+		}
+	} else {
+		data->bytes_xfered = host->datasize;
+	}
+
+	data_error = data->error;
+
+	host->data = NULL;
+
+	return data_error;
+}
+
+static void mxcmci_read_response(struct mxcmci_host *host, unsigned int stat)
+{
+	struct mmc_command *cmd = host->cmd;
+	int i;
+	u32 a, b, c;
+
+	if (!cmd)
+		return;
+
+	if (stat & STATUS_TIME_OUT_RESP) {
+		dev_dbg(mmc_dev(host->mmc), "CMD TIMEOUT\n");
+		cmd->error = -ETIMEDOUT;
+	} else if (stat & STATUS_RESP_CRC_ERR && cmd->flags & MMC_RSP_CRC) {
+		dev_dbg(mmc_dev(host->mmc), "cmd crc error\n");
+		cmd->error = -EILSEQ;
+	}
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136) {
+			for (i = 0; i < 4; i++) {
+				a = mxcmci_readw(host, MMC_REG_RES_FIFO);
+				b = mxcmci_readw(host, MMC_REG_RES_FIFO);
+				cmd->resp[i] = a << 16 | b;
+			}
+		} else {
+			a = mxcmci_readw(host, MMC_REG_RES_FIFO);
+			b = mxcmci_readw(host, MMC_REG_RES_FIFO);
+			c = mxcmci_readw(host, MMC_REG_RES_FIFO);
+			cmd->resp[0] = a << 24 | b << 8 | c >> 8;
+		}
+	}
+}
+
+static int mxcmci_poll_status(struct mxcmci_host *host, u32 mask)
+{
+	u32 stat;
+	unsigned long timeout = jiffies + HZ;
+
+	do {
+		stat = mxcmci_readl(host, MMC_REG_STATUS);
+		if (stat & STATUS_ERR_MASK)
+			return stat;
+		if (time_after(jiffies, timeout)) {
+			mxcmci_softreset(host);
+			mxcmci_set_clk_rate(host, host->clock);
+			return STATUS_TIME_OUT_READ;
+		}
+		if (stat & mask)
+			return 0;
+		cpu_relax();
+	} while (1);
+}
+
+static int mxcmci_pull(struct mxcmci_host *host, void *_buf, int bytes)
+{
+	unsigned int stat;
+	u32 *buf = _buf;
+
+	while (bytes > 3) {
+		stat = mxcmci_poll_status(host,
+				STATUS_BUF_READ_RDY | STATUS_READ_OP_DONE);
+		if (stat)
+			return stat;
+		*buf++ = cpu_to_le32(mxcmci_readl(host, MMC_REG_BUFFER_ACCESS));
+		bytes -= 4;
+	}
+
+	if (bytes) {
+		u8 *b = (u8 *)buf;
+		u32 tmp;
+
+		stat = mxcmci_poll_status(host,
+				STATUS_BUF_READ_RDY | STATUS_READ_OP_DONE);
+		if (stat)
+			return stat;
+		tmp = cpu_to_le32(mxcmci_readl(host, MMC_REG_BUFFER_ACCESS));
+		memcpy(b, &tmp, bytes);
+	}
+
+	return 0;
+}
+
+static int mxcmci_push(struct mxcmci_host *host, void *_buf, int bytes)
+{
+	unsigned int stat;
+	u32 *buf = _buf;
+
+	while (bytes > 3) {
+		stat = mxcmci_poll_status(host, STATUS_BUF_WRITE_RDY);
+		if (stat)
+			return stat;
+		mxcmci_writel(host, cpu_to_le32(*buf++), MMC_REG_BUFFER_ACCESS);
+		bytes -= 4;
+	}
+
+	if (bytes) {
+		u8 *b = (u8 *)buf;
+		u32 tmp;
+
+		stat = mxcmci_poll_status(host, STATUS_BUF_WRITE_RDY);
+		if (stat)
+			return stat;
+
+		memcpy(&tmp, b, bytes);
+		mxcmci_writel(host, cpu_to_le32(tmp), MMC_REG_BUFFER_ACCESS);
+	}
+
+	return mxcmci_poll_status(host, STATUS_BUF_WRITE_RDY);
+}
+
+static int mxcmci_transfer_data(struct mxcmci_host *host)
+{
+	struct mmc_data *data = host->req->data;
+	struct scatterlist *sg;
+	int stat, i;
+
+	host->data = data;
+	host->datasize = 0;
+
+	if (data->flags & MMC_DATA_READ) {
+		for_each_sg(data->sg, sg, data->sg_len, i) {
+			stat = mxcmci_pull(host, sg_virt(sg), sg->length);
+			if (stat)
+				return stat;
+			host->datasize += sg->length;
+		}
+	} else {
+		for_each_sg(data->sg, sg, data->sg_len, i) {
+			stat = mxcmci_push(host, sg_virt(sg), sg->length);
+			if (stat)
+				return stat;
+			host->datasize += sg->length;
+		}
+		stat = mxcmci_poll_status(host, STATUS_WRITE_OP_DONE);
+		if (stat)
+			return stat;
+	}
+	return 0;
+}
+
+static void mxcmci_datawork(struct work_struct *work)
+{
+	struct mxcmci_host *host = container_of(work, struct mxcmci_host,
+						  datawork);
+	int datastat = mxcmci_transfer_data(host);
+
+	mxcmci_writel(host, STATUS_READ_OP_DONE | STATUS_WRITE_OP_DONE,
+		MMC_REG_STATUS);
+	mxcmci_finish_data(host, datastat);
+
+	if (host->req->stop) {
+		if (mxcmci_start_cmd(host, host->req->stop, 0)) {
+			mxcmci_finish_request(host, host->req);
+			return;
+		}
+	} else {
+		mxcmci_finish_request(host, host->req);
+	}
+}
+
+static void mxcmci_data_done(struct mxcmci_host *host, unsigned int stat)
+{
+	struct mmc_request *req;
+	int data_error;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (!host->data) {
+		spin_unlock_irqrestore(&host->lock, flags);
+		return;
+	}
+
+	if (!host->req) {
+		spin_unlock_irqrestore(&host->lock, flags);
+		return;
+	}
+
+	req = host->req;
+	if (!req->stop)
+		host->req = NULL; /* we will handle finish req below */
+
+	data_error = mxcmci_finish_data(host, stat);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (data_error)
+		return;
+
+	mxcmci_read_response(host, stat);
+	host->cmd = NULL;
+
+	if (req->stop) {
+		if (mxcmci_start_cmd(host, req->stop, 0)) {
+			mxcmci_finish_request(host, req);
+			return;
+		}
+	} else {
+		mxcmci_finish_request(host, req);
+	}
+}
+
+static void mxcmci_cmd_done(struct mxcmci_host *host, unsigned int stat)
+{
+	mxcmci_read_response(host, stat);
+	host->cmd = NULL;
+
+	if (!host->data && host->req) {
+		mxcmci_finish_request(host, host->req);
+		return;
+	}
+
+	/* For the DMA case the DMA engine handles the data transfer
+	 * automatically. For non DMA we have to do it ourselves.
+	 * Don't do it in interrupt context though.
+	 */
+	if (!mxcmci_use_dma(host) && host->data)
+		schedule_work(&host->datawork);
+
+}
+
+static irqreturn_t mxcmci_irq(int irq, void *devid)
+{
+	struct mxcmci_host *host = devid;
+	bool sdio_irq;
+	u32 stat;
+
+	stat = mxcmci_readl(host, MMC_REG_STATUS);
+	mxcmci_writel(host,
+		stat & ~(STATUS_SDIO_INT_ACTIVE | STATUS_DATA_TRANS_DONE |
+			 STATUS_WRITE_OP_DONE),
+		MMC_REG_STATUS);
+
+	dev_dbg(mmc_dev(host->mmc), "%s: 0x%08x\n", __func__, stat);
+
+	spin_lock(&host->lock);
+	sdio_irq = (stat & STATUS_SDIO_INT_ACTIVE) && host->use_sdio;
+	spin_unlock(&host->lock);
+
+	if (mxcmci_use_dma(host) && (stat & (STATUS_WRITE_OP_DONE)))
+		mxcmci_writel(host, STATUS_WRITE_OP_DONE, MMC_REG_STATUS);
+
+	if (sdio_irq) {
+		mxcmci_writel(host, STATUS_SDIO_INT_ACTIVE, MMC_REG_STATUS);
+		mmc_signal_sdio_irq(host->mmc);
+	}
+
+	if (stat & STATUS_END_CMD_RESP)
+		mxcmci_cmd_done(host, stat);
+
+	if (mxcmci_use_dma(host) && (stat & STATUS_WRITE_OP_DONE)) {
+		del_timer(&host->watchdog);
+		mxcmci_data_done(host, stat);
+	}
+
+	if (host->default_irq_mask &&
+		  (stat & (STATUS_CARD_INSERTION | STATUS_CARD_REMOVAL)))
+		mmc_detect_change(host->mmc, msecs_to_jiffies(200));
+
+	return IRQ_HANDLED;
+}
+
+static void mxcmci_request(struct mmc_host *mmc, struct mmc_request *req)
+{
+	struct mxcmci_host *host = mmc_priv(mmc);
+	unsigned int cmdat = host->cmdat;
+	int error;
+
+	WARN_ON(host->req != NULL);
+
+	host->req = req;
+	host->cmdat &= ~CMD_DAT_CONT_INIT;
+
+	if (host->dma)
+		host->do_dma = 1;
+
+	if (req->data) {
+		error = mxcmci_setup_data(host, req->data);
+		if (error) {
+			req->cmd->error = error;
+			goto out;
+		}
+
+
+		cmdat |= CMD_DAT_CONT_DATA_ENABLE;
+
+		if (req->data->flags & MMC_DATA_WRITE)
+			cmdat |= CMD_DAT_CONT_WRITE;
+	}
+
+	error = mxcmci_start_cmd(host, req->cmd, cmdat);
+
+out:
+	if (error)
+		mxcmci_finish_request(host, req);
+}
+
+static void mxcmci_set_clk_rate(struct mxcmci_host *host, unsigned int clk_ios)
+{
+	unsigned int divider;
+	int prescaler = 0;
+	unsigned int clk_in = clk_get_rate(host->clk_per);
+
+	while (prescaler <= 0x800) {
+		for (divider = 1; divider <= 0xF; divider++) {
+			int x;
+
+			x = (clk_in / (divider + 1));
+
+			if (prescaler)
+				x /= (prescaler * 2);
+
+			if (x <= clk_ios)
+				break;
+		}
+		if (divider < 0x10)
+			break;
+
+		if (prescaler == 0)
+			prescaler = 1;
+		else
+			prescaler <<= 1;
+	}
+
+	mxcmci_writew(host, (prescaler << 4) | divider, MMC_REG_CLK_RATE);
+
+	dev_dbg(mmc_dev(host->mmc), "scaler: %d divider: %d in: %d out: %d\n",
+			prescaler, divider, clk_in, clk_ios);
+}
+
+static int mxcmci_setup_dma(struct mmc_host *mmc)
+{
+	struct mxcmci_host *host = mmc_priv(mmc);
+	struct dma_slave_config *config = &host->dma_slave_config;
+
+	config->dst_addr = host->phys_base + MMC_REG_BUFFER_ACCESS;
+	config->src_addr = host->phys_base + MMC_REG_BUFFER_ACCESS;
+	config->dst_addr_width = 4;
+	config->src_addr_width = 4;
+	config->dst_maxburst = host->burstlen;
+	config->src_maxburst = host->burstlen;
+	config->device_fc = false;
+
+	return dmaengine_slave_config(host->dma, config);
+}
+
+static void mxcmci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct mxcmci_host *host = mmc_priv(mmc);
+	int burstlen, ret;
+
+	/*
+	 * use burstlen of 64 (16 words) in 4 bit mode (--> reg value  0)
+	 * use burstlen of 16 (4 words) in 1 bit mode (--> reg value 16)
+	 */
+	if (ios->bus_width == MMC_BUS_WIDTH_4)
+		burstlen = 16;
+	else
+		burstlen = 4;
+
+	if (mxcmci_use_dma(host) && burstlen != host->burstlen) {
+		host->burstlen = burstlen;
+		ret = mxcmci_setup_dma(mmc);
+		if (ret) {
+			dev_err(mmc_dev(host->mmc),
+				"failed to config DMA channel. Falling back to PIO\n");
+			dma_release_channel(host->dma);
+			host->do_dma = 0;
+			host->dma = NULL;
+		}
+	}
+
+	if (ios->bus_width == MMC_BUS_WIDTH_4)
+		host->cmdat |= CMD_DAT_CONT_BUS_WIDTH_4;
+	else
+		host->cmdat &= ~CMD_DAT_CONT_BUS_WIDTH_4;
+
+	if (host->power_mode != ios->power_mode) {
+		host->power_mode = ios->power_mode;
+		mxcmci_set_power(host, ios->vdd);
+
+		if (ios->power_mode == MMC_POWER_ON)
+			host->cmdat |= CMD_DAT_CONT_INIT;
+	}
+
+	if (ios->clock) {
+		mxcmci_set_clk_rate(host, ios->clock);
+		mxcmci_writew(host, STR_STP_CLK_START_CLK, MMC_REG_STR_STP_CLK);
+	} else {
+		mxcmci_writew(host, STR_STP_CLK_STOP_CLK, MMC_REG_STR_STP_CLK);
+	}
+
+	host->clock = ios->clock;
+}
+
+static irqreturn_t mxcmci_detect_irq(int irq, void *data)
+{
+	struct mmc_host *mmc = data;
+
+	dev_dbg(mmc_dev(mmc), "%s\n", __func__);
+
+	mmc_detect_change(mmc, msecs_to_jiffies(250));
+	return IRQ_HANDLED;
+}
+
+static int mxcmci_get_ro(struct mmc_host *mmc)
+{
+	struct mxcmci_host *host = mmc_priv(mmc);
+
+	if (host->pdata && host->pdata->get_ro)
+		return !!host->pdata->get_ro(mmc_dev(mmc));
+	/*
+	 * If board doesn't support read only detection (no mmc_gpio
+	 * context or gpio is invalid), then let the mmc core decide
+	 * what to do.
+	 */
+	return mmc_gpio_get_ro(mmc);
+}
+
+static void mxcmci_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct mxcmci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+	u32 int_cntr;
+
+	spin_lock_irqsave(&host->lock, flags);
+	host->use_sdio = enable;
+	int_cntr = mxcmci_readl(host, MMC_REG_INT_CNTR);
+
+	if (enable)
+		int_cntr |= INT_SDIO_IRQ_EN;
+	else
+		int_cntr &= ~INT_SDIO_IRQ_EN;
+
+	mxcmci_writel(host, int_cntr, MMC_REG_INT_CNTR);
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void mxcmci_init_card(struct mmc_host *host, struct mmc_card *card)
+{
+	struct mxcmci_host *mxcmci = mmc_priv(host);
+
+	/*
+	 * MX3 SoCs have a silicon bug which corrupts CRC calculation of
+	 * multi-block transfers when connected SDIO peripheral doesn't
+	 * drive the BUSY line as required by the specs.
+	 * One way to prevent this is to only allow 1-bit transfers.
+	 */
+
+	if (is_imx31_mmc(mxcmci) && card->type == MMC_TYPE_SDIO)
+		host->caps &= ~MMC_CAP_4_BIT_DATA;
+	else
+		host->caps |= MMC_CAP_4_BIT_DATA;
+}
+
+static bool filter(struct dma_chan *chan, void *param)
+{
+	struct mxcmci_host *host = param;
+
+	if (!imx_dma_is_general_purpose(chan))
+		return false;
+
+	chan->private = &host->dma_data;
+
+	return true;
+}
+
+static void mxcmci_watchdog(struct timer_list *t)
+{
+	struct mxcmci_host *host = from_timer(host, t, watchdog);
+	struct mmc_request *req = host->req;
+	unsigned int stat = mxcmci_readl(host, MMC_REG_STATUS);
+
+	if (host->dma_dir == DMA_FROM_DEVICE) {
+		dmaengine_terminate_all(host->dma);
+		dev_err(mmc_dev(host->mmc),
+			"%s: read time out (status = 0x%08x)\n",
+			__func__, stat);
+	} else {
+		dev_err(mmc_dev(host->mmc),
+			"%s: write time out (status = 0x%08x)\n",
+			__func__, stat);
+		mxcmci_softreset(host);
+	}
+
+	/* Mark transfer as erroneus and inform the upper layers */
+
+	if (host->data)
+		host->data->error = -ETIMEDOUT;
+	host->req = NULL;
+	host->cmd = NULL;
+	host->data = NULL;
+	mmc_request_done(host->mmc, req);
+}
+
+static const struct mmc_host_ops mxcmci_ops = {
+	.request		= mxcmci_request,
+	.set_ios		= mxcmci_set_ios,
+	.get_ro			= mxcmci_get_ro,
+	.enable_sdio_irq	= mxcmci_enable_sdio_irq,
+	.init_card		= mxcmci_init_card,
+};
+
+static int mxcmci_probe(struct platform_device *pdev)
+{
+	struct mmc_host *mmc;
+	struct mxcmci_host *host;
+	struct resource *res;
+	int ret = 0, irq;
+	bool dat3_card_detect = false;
+	dma_cap_mask_t mask;
+	const struct of_device_id *of_id;
+	struct imxmmc_platform_data *pdata = pdev->dev.platform_data;
+
+	pr_info("i.MX/MPC512x SDHC driver\n");
+
+	of_id = of_match_device(mxcmci_of_match, &pdev->dev);
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return irq;
+
+	mmc = mmc_alloc_host(sizeof(*host), &pdev->dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	host = mmc_priv(mmc);
+
+	host->base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(host->base)) {
+		ret = PTR_ERR(host->base);
+		goto out_free;
+	}
+
+	host->phys_base = res->start;
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto out_free;
+	mmc->ops = &mxcmci_ops;
+
+	/* For devicetree parsing, the bus width is read from devicetree */
+	if (pdata)
+		mmc->caps = MMC_CAP_4_BIT_DATA | MMC_CAP_SDIO_IRQ;
+	else
+		mmc->caps |= MMC_CAP_SDIO_IRQ;
+
+	/* MMC core transfer sizes tunable parameters */
+	mmc->max_blk_size = 2048;
+	mmc->max_blk_count = 65535;
+	mmc->max_req_size = mmc->max_blk_size * mmc->max_blk_count;
+	mmc->max_seg_size = mmc->max_req_size;
+
+	if (of_id) {
+		const struct platform_device_id *id_entry = of_id->data;
+		host->devtype = id_entry->driver_data;
+	} else {
+		host->devtype = pdev->id_entry->driver_data;
+	}
+
+	/* adjust max_segs after devtype detection */
+	if (!is_mpc512x_mmc(host))
+		mmc->max_segs = 64;
+
+	host->mmc = mmc;
+	host->pdata = pdata;
+	spin_lock_init(&host->lock);
+
+	if (pdata)
+		dat3_card_detect = pdata->dat3_card_detect;
+	else if (mmc_card_is_removable(mmc)
+			&& !of_property_read_bool(pdev->dev.of_node, "cd-gpios"))
+		dat3_card_detect = true;
+
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret)
+		goto out_free;
+
+	if (!mmc->ocr_avail) {
+		if (pdata && pdata->ocr_avail)
+			mmc->ocr_avail = pdata->ocr_avail;
+		else
+			mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+	}
+
+	if (dat3_card_detect)
+		host->default_irq_mask =
+			INT_CARD_INSERTION_EN | INT_CARD_REMOVAL_EN;
+	else
+		host->default_irq_mask = 0;
+
+	host->clk_ipg = devm_clk_get(&pdev->dev, "ipg");
+	if (IS_ERR(host->clk_ipg)) {
+		ret = PTR_ERR(host->clk_ipg);
+		goto out_free;
+	}
+
+	host->clk_per = devm_clk_get(&pdev->dev, "per");
+	if (IS_ERR(host->clk_per)) {
+		ret = PTR_ERR(host->clk_per);
+		goto out_free;
+	}
+
+	ret = clk_prepare_enable(host->clk_per);
+	if (ret)
+		goto out_free;
+
+	ret = clk_prepare_enable(host->clk_ipg);
+	if (ret)
+		goto out_clk_per_put;
+
+	mxcmci_softreset(host);
+
+	host->rev_no = mxcmci_readw(host, MMC_REG_REV_NO);
+	if (host->rev_no != 0x400) {
+		ret = -ENODEV;
+		dev_err(mmc_dev(host->mmc), "wrong rev.no. 0x%08x. aborting.\n",
+			host->rev_no);
+		goto out_clk_put;
+	}
+
+	mmc->f_min = clk_get_rate(host->clk_per) >> 16;
+	mmc->f_max = clk_get_rate(host->clk_per) >> 1;
+
+	/* recommended in data sheet */
+	mxcmci_writew(host, 0x2db4, MMC_REG_READ_TO);
+
+	mxcmci_writel(host, host->default_irq_mask, MMC_REG_INT_CNTR);
+
+	if (!host->pdata) {
+		host->dma = dma_request_chan(&pdev->dev, "rx-tx");
+		if (IS_ERR(host->dma)) {
+			if (PTR_ERR(host->dma) == -EPROBE_DEFER) {
+				ret = -EPROBE_DEFER;
+				goto out_clk_put;
+			}
+
+			/* Ignore errors to fall back to PIO mode */
+			host->dma = NULL;
+		}
+	} else {
+		res = platform_get_resource(pdev, IORESOURCE_DMA, 0);
+		if (res) {
+			host->dmareq = res->start;
+			host->dma_data.peripheral_type = IMX_DMATYPE_SDHC;
+			host->dma_data.priority = DMA_PRIO_LOW;
+			host->dma_data.dma_request = host->dmareq;
+			dma_cap_zero(mask);
+			dma_cap_set(DMA_SLAVE, mask);
+			host->dma = dma_request_channel(mask, filter, host);
+		}
+	}
+	if (host->dma)
+		mmc->max_seg_size = dma_get_max_seg_size(
+				host->dma->device->dev);
+	else
+		dev_info(mmc_dev(host->mmc), "dma not available. Using PIO\n");
+
+	INIT_WORK(&host->datawork, mxcmci_datawork);
+
+	ret = devm_request_irq(&pdev->dev, irq, mxcmci_irq, 0,
+			       dev_name(&pdev->dev), host);
+	if (ret)
+		goto out_free_dma;
+
+	platform_set_drvdata(pdev, mmc);
+
+	if (host->pdata && host->pdata->init) {
+		ret = host->pdata->init(&pdev->dev, mxcmci_detect_irq,
+				host->mmc);
+		if (ret)
+			goto out_free_dma;
+	}
+
+	timer_setup(&host->watchdog, mxcmci_watchdog, 0);
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto out_free_dma;
+
+	return 0;
+
+out_free_dma:
+	if (host->dma)
+		dma_release_channel(host->dma);
+
+out_clk_put:
+	clk_disable_unprepare(host->clk_ipg);
+out_clk_per_put:
+	clk_disable_unprepare(host->clk_per);
+
+out_free:
+	mmc_free_host(mmc);
+
+	return ret;
+}
+
+static int mxcmci_remove(struct platform_device *pdev)
+{
+	struct mmc_host *mmc = platform_get_drvdata(pdev);
+	struct mxcmci_host *host = mmc_priv(mmc);
+
+	mmc_remove_host(mmc);
+
+	if (host->pdata && host->pdata->exit)
+		host->pdata->exit(&pdev->dev, mmc);
+
+	if (host->dma)
+		dma_release_channel(host->dma);
+
+	clk_disable_unprepare(host->clk_per);
+	clk_disable_unprepare(host->clk_ipg);
+
+	mmc_free_host(mmc);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int mxcmci_suspend(struct device *dev)
+{
+	struct mmc_host *mmc = dev_get_drvdata(dev);
+	struct mxcmci_host *host = mmc_priv(mmc);
+
+	clk_disable_unprepare(host->clk_per);
+	clk_disable_unprepare(host->clk_ipg);
+	return 0;
+}
+
+static int mxcmci_resume(struct device *dev)
+{
+	struct mmc_host *mmc = dev_get_drvdata(dev);
+	struct mxcmci_host *host = mmc_priv(mmc);
+	int ret;
+
+	ret = clk_prepare_enable(host->clk_per);
+	if (ret)
+		return ret;
+
+	ret = clk_prepare_enable(host->clk_ipg);
+	if (ret)
+		clk_disable_unprepare(host->clk_per);
+
+	return ret;
+}
+#endif
+
+static SIMPLE_DEV_PM_OPS(mxcmci_pm_ops, mxcmci_suspend, mxcmci_resume);
+
+static struct platform_driver mxcmci_driver = {
+	.probe		= mxcmci_probe,
+	.remove		= mxcmci_remove,
+	.id_table	= mxcmci_devtype,
+	.driver		= {
+		.name		= DRIVER_NAME,
+		.probe_type	= PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= &mxcmci_pm_ops,
+		.of_match_table	= mxcmci_of_match,
+	}
+};
+
+module_platform_driver(mxcmci_driver);
+
+MODULE_DESCRIPTION("i.MX Multimedia Card Interface Driver");
+MODULE_AUTHOR("Sascha Hauer, Pengutronix");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:mxc-mmc");
diff --git a/drivers/mmc/host/mxs-mmc.c b/drivers/mmc/host/mxs-mmc.c
new file mode 100644
index 000000000..2ec3eb651
--- /dev/null
+++ b/drivers/mmc/host/mxs-mmc.c
@@ -0,0 +1,750 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Portions copyright (C) 2003 Russell King, PXA MMCI Driver
+ * Portions copyright (C) 2004-2005 Pierre Ossman, W83L51xD SD/MMC driver
+ *
+ * Copyright 2008 Embedded Alley Solutions, Inc.
+ * Copyright 2009-2011 Freescale Semiconductor, Inc.
+ */
+
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/ioport.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <linux/delay.h>
+#include <linux/interrupt.h>
+#include <linux/dma-mapping.h>
+#include <linux/dmaengine.h>
+#include <linux/dma/mxs-dma.h>
+#include <linux/highmem.h>
+#include <linux/clk.h>
+#include <linux/err.h>
+#include <linux/completion.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sdio.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/regulator/consumer.h>
+#include <linux/module.h>
+#include <linux/stmp_device.h>
+#include <linux/spi/mxs-spi.h>
+
+#define DRIVER_NAME	"mxs-mmc"
+
+#define MXS_MMC_IRQ_BITS	(BM_SSP_CTRL1_SDIO_IRQ		| \
+				 BM_SSP_CTRL1_RESP_ERR_IRQ	| \
+				 BM_SSP_CTRL1_RESP_TIMEOUT_IRQ	| \
+				 BM_SSP_CTRL1_DATA_TIMEOUT_IRQ	| \
+				 BM_SSP_CTRL1_DATA_CRC_IRQ	| \
+				 BM_SSP_CTRL1_FIFO_UNDERRUN_IRQ	| \
+				 BM_SSP_CTRL1_RECV_TIMEOUT_IRQ  | \
+				 BM_SSP_CTRL1_FIFO_OVERRUN_IRQ)
+
+/* card detect polling timeout */
+#define MXS_MMC_DETECT_TIMEOUT			(HZ/2)
+
+struct mxs_mmc_host {
+	struct mxs_ssp			ssp;
+
+	struct mmc_host			*mmc;
+	struct mmc_request		*mrq;
+	struct mmc_command		*cmd;
+	struct mmc_data			*data;
+
+	unsigned char			bus_width;
+	spinlock_t			lock;
+	int				sdio_irq_en;
+	bool				broken_cd;
+};
+
+static int mxs_mmc_get_cd(struct mmc_host *mmc)
+{
+	struct mxs_mmc_host *host = mmc_priv(mmc);
+	struct mxs_ssp *ssp = &host->ssp;
+	int present, ret;
+
+	if (host->broken_cd)
+		return -ENOSYS;
+
+	ret = mmc_gpio_get_cd(mmc);
+	if (ret >= 0)
+		return ret;
+
+	present = mmc->caps & MMC_CAP_NEEDS_POLL ||
+		!(readl(ssp->base + HW_SSP_STATUS(ssp)) &
+			BM_SSP_STATUS_CARD_DETECT);
+
+	if (mmc->caps2 & MMC_CAP2_CD_ACTIVE_HIGH)
+		present = !present;
+
+	return present;
+}
+
+static int mxs_mmc_reset(struct mxs_mmc_host *host)
+{
+	struct mxs_ssp *ssp = &host->ssp;
+	u32 ctrl0, ctrl1;
+	int ret;
+
+	ret = stmp_reset_block(ssp->base);
+	if (ret)
+		return ret;
+
+	ctrl0 = BM_SSP_CTRL0_IGNORE_CRC;
+	ctrl1 = BF_SSP(0x3, CTRL1_SSP_MODE) |
+		BF_SSP(0x7, CTRL1_WORD_LENGTH) |
+		BM_SSP_CTRL1_DMA_ENABLE |
+		BM_SSP_CTRL1_POLARITY |
+		BM_SSP_CTRL1_RECV_TIMEOUT_IRQ_EN |
+		BM_SSP_CTRL1_DATA_CRC_IRQ_EN |
+		BM_SSP_CTRL1_DATA_TIMEOUT_IRQ_EN |
+		BM_SSP_CTRL1_RESP_TIMEOUT_IRQ_EN |
+		BM_SSP_CTRL1_RESP_ERR_IRQ_EN;
+
+	writel(BF_SSP(0xffff, TIMING_TIMEOUT) |
+	       BF_SSP(2, TIMING_CLOCK_DIVIDE) |
+	       BF_SSP(0, TIMING_CLOCK_RATE),
+	       ssp->base + HW_SSP_TIMING(ssp));
+
+	if (host->sdio_irq_en) {
+		ctrl0 |= BM_SSP_CTRL0_SDIO_IRQ_CHECK;
+		ctrl1 |= BM_SSP_CTRL1_SDIO_IRQ_EN;
+	}
+
+	writel(ctrl0, ssp->base + HW_SSP_CTRL0);
+	writel(ctrl1, ssp->base + HW_SSP_CTRL1(ssp));
+	return 0;
+}
+
+static void mxs_mmc_start_cmd(struct mxs_mmc_host *host,
+			      struct mmc_command *cmd);
+
+static void mxs_mmc_request_done(struct mxs_mmc_host *host)
+{
+	struct mmc_command *cmd = host->cmd;
+	struct mmc_data *data = host->data;
+	struct mmc_request *mrq = host->mrq;
+	struct mxs_ssp *ssp = &host->ssp;
+
+	if (mmc_resp_type(cmd) & MMC_RSP_PRESENT) {
+		if (mmc_resp_type(cmd) & MMC_RSP_136) {
+			cmd->resp[3] = readl(ssp->base + HW_SSP_SDRESP0(ssp));
+			cmd->resp[2] = readl(ssp->base + HW_SSP_SDRESP1(ssp));
+			cmd->resp[1] = readl(ssp->base + HW_SSP_SDRESP2(ssp));
+			cmd->resp[0] = readl(ssp->base + HW_SSP_SDRESP3(ssp));
+		} else {
+			cmd->resp[0] = readl(ssp->base + HW_SSP_SDRESP0(ssp));
+		}
+	}
+
+	if (cmd == mrq->sbc) {
+		/* Finished CMD23, now send actual command. */
+		mxs_mmc_start_cmd(host, mrq->cmd);
+		return;
+	} else if (data) {
+		dma_unmap_sg(mmc_dev(host->mmc), data->sg,
+			     data->sg_len, ssp->dma_dir);
+		/*
+		 * If there was an error on any block, we mark all
+		 * data blocks as being in error.
+		 */
+		if (!data->error)
+			data->bytes_xfered = data->blocks * data->blksz;
+		else
+			data->bytes_xfered = 0;
+
+		host->data = NULL;
+		if (data->stop && (data->error || !mrq->sbc)) {
+			mxs_mmc_start_cmd(host, mrq->stop);
+			return;
+		}
+	}
+
+	host->mrq = NULL;
+	mmc_request_done(host->mmc, mrq);
+}
+
+static void mxs_mmc_dma_irq_callback(void *param)
+{
+	struct mxs_mmc_host *host = param;
+
+	mxs_mmc_request_done(host);
+}
+
+static irqreturn_t mxs_mmc_irq_handler(int irq, void *dev_id)
+{
+	struct mxs_mmc_host *host = dev_id;
+	struct mmc_command *cmd = host->cmd;
+	struct mmc_data *data = host->data;
+	struct mxs_ssp *ssp = &host->ssp;
+	u32 stat;
+
+	spin_lock(&host->lock);
+
+	stat = readl(ssp->base + HW_SSP_CTRL1(ssp));
+	writel(stat & MXS_MMC_IRQ_BITS,
+	       ssp->base + HW_SSP_CTRL1(ssp) + STMP_OFFSET_REG_CLR);
+
+	spin_unlock(&host->lock);
+
+	if ((stat & BM_SSP_CTRL1_SDIO_IRQ) && (stat & BM_SSP_CTRL1_SDIO_IRQ_EN))
+		mmc_signal_sdio_irq(host->mmc);
+
+	if (stat & BM_SSP_CTRL1_RESP_TIMEOUT_IRQ)
+		cmd->error = -ETIMEDOUT;
+	else if (stat & BM_SSP_CTRL1_RESP_ERR_IRQ)
+		cmd->error = -EIO;
+
+	if (data) {
+		if (stat & (BM_SSP_CTRL1_DATA_TIMEOUT_IRQ |
+			    BM_SSP_CTRL1_RECV_TIMEOUT_IRQ))
+			data->error = -ETIMEDOUT;
+		else if (stat & BM_SSP_CTRL1_DATA_CRC_IRQ)
+			data->error = -EILSEQ;
+		else if (stat & (BM_SSP_CTRL1_FIFO_UNDERRUN_IRQ |
+				 BM_SSP_CTRL1_FIFO_OVERRUN_IRQ))
+			data->error = -EIO;
+	}
+
+	return IRQ_HANDLED;
+}
+
+static struct dma_async_tx_descriptor *mxs_mmc_prep_dma(
+	struct mxs_mmc_host *host, unsigned long flags)
+{
+	struct mxs_ssp *ssp = &host->ssp;
+	struct dma_async_tx_descriptor *desc;
+	struct mmc_data *data = host->data;
+	struct scatterlist * sgl;
+	unsigned int sg_len;
+
+	if (data) {
+		/* data */
+		dma_map_sg(mmc_dev(host->mmc), data->sg,
+			   data->sg_len, ssp->dma_dir);
+		sgl = data->sg;
+		sg_len = data->sg_len;
+	} else {
+		/* pio */
+		sgl = (struct scatterlist *) ssp->ssp_pio_words;
+		sg_len = SSP_PIO_NUM;
+	}
+
+	desc = dmaengine_prep_slave_sg(ssp->dmach,
+				sgl, sg_len, ssp->slave_dirn, flags);
+	if (desc) {
+		desc->callback = mxs_mmc_dma_irq_callback;
+		desc->callback_param = host;
+	} else {
+		if (data)
+			dma_unmap_sg(mmc_dev(host->mmc), data->sg,
+				     data->sg_len, ssp->dma_dir);
+	}
+
+	return desc;
+}
+
+static void mxs_mmc_bc(struct mxs_mmc_host *host)
+{
+	struct mxs_ssp *ssp = &host->ssp;
+	struct mmc_command *cmd = host->cmd;
+	struct dma_async_tx_descriptor *desc;
+	u32 ctrl0, cmd0, cmd1;
+
+	ctrl0 = BM_SSP_CTRL0_ENABLE | BM_SSP_CTRL0_IGNORE_CRC;
+	cmd0 = BF_SSP(cmd->opcode, CMD0_CMD) | BM_SSP_CMD0_APPEND_8CYC;
+	cmd1 = cmd->arg;
+
+	if (host->sdio_irq_en) {
+		ctrl0 |= BM_SSP_CTRL0_SDIO_IRQ_CHECK;
+		cmd0 |= BM_SSP_CMD0_CONT_CLKING_EN | BM_SSP_CMD0_SLOW_CLKING_EN;
+	}
+
+	ssp->ssp_pio_words[0] = ctrl0;
+	ssp->ssp_pio_words[1] = cmd0;
+	ssp->ssp_pio_words[2] = cmd1;
+	ssp->dma_dir = DMA_NONE;
+	ssp->slave_dirn = DMA_TRANS_NONE;
+	desc = mxs_mmc_prep_dma(host, MXS_DMA_CTRL_WAIT4END);
+	if (!desc)
+		goto out;
+
+	dmaengine_submit(desc);
+	dma_async_issue_pending(ssp->dmach);
+	return;
+
+out:
+	dev_warn(mmc_dev(host->mmc),
+		 "%s: failed to prep dma\n", __func__);
+}
+
+static void mxs_mmc_ac(struct mxs_mmc_host *host)
+{
+	struct mxs_ssp *ssp = &host->ssp;
+	struct mmc_command *cmd = host->cmd;
+	struct dma_async_tx_descriptor *desc;
+	u32 ignore_crc, get_resp, long_resp;
+	u32 ctrl0, cmd0, cmd1;
+
+	ignore_crc = (mmc_resp_type(cmd) & MMC_RSP_CRC) ?
+			0 : BM_SSP_CTRL0_IGNORE_CRC;
+	get_resp = (mmc_resp_type(cmd) & MMC_RSP_PRESENT) ?
+			BM_SSP_CTRL0_GET_RESP : 0;
+	long_resp = (mmc_resp_type(cmd) & MMC_RSP_136) ?
+			BM_SSP_CTRL0_LONG_RESP : 0;
+
+	ctrl0 = BM_SSP_CTRL0_ENABLE | ignore_crc | get_resp | long_resp;
+	cmd0 = BF_SSP(cmd->opcode, CMD0_CMD);
+	cmd1 = cmd->arg;
+
+	if (cmd->opcode == MMC_STOP_TRANSMISSION)
+		cmd0 |= BM_SSP_CMD0_APPEND_8CYC;
+
+	if (host->sdio_irq_en) {
+		ctrl0 |= BM_SSP_CTRL0_SDIO_IRQ_CHECK;
+		cmd0 |= BM_SSP_CMD0_CONT_CLKING_EN | BM_SSP_CMD0_SLOW_CLKING_EN;
+	}
+
+	ssp->ssp_pio_words[0] = ctrl0;
+	ssp->ssp_pio_words[1] = cmd0;
+	ssp->ssp_pio_words[2] = cmd1;
+	ssp->dma_dir = DMA_NONE;
+	ssp->slave_dirn = DMA_TRANS_NONE;
+	desc = mxs_mmc_prep_dma(host, MXS_DMA_CTRL_WAIT4END);
+	if (!desc)
+		goto out;
+
+	dmaengine_submit(desc);
+	dma_async_issue_pending(ssp->dmach);
+	return;
+
+out:
+	dev_warn(mmc_dev(host->mmc),
+		 "%s: failed to prep dma\n", __func__);
+}
+
+static unsigned short mxs_ns_to_ssp_ticks(unsigned clock_rate, unsigned ns)
+{
+	const unsigned int ssp_timeout_mul = 4096;
+	/*
+	 * Calculate ticks in ms since ns are large numbers
+	 * and might overflow
+	 */
+	const unsigned int clock_per_ms = clock_rate / 1000;
+	const unsigned int ms = ns / 1000;
+	const unsigned int ticks = ms * clock_per_ms;
+	const unsigned int ssp_ticks = ticks / ssp_timeout_mul;
+
+	WARN_ON(ssp_ticks == 0);
+	return ssp_ticks;
+}
+
+static void mxs_mmc_adtc(struct mxs_mmc_host *host)
+{
+	struct mmc_command *cmd = host->cmd;
+	struct mmc_data *data = cmd->data;
+	struct dma_async_tx_descriptor *desc;
+	struct scatterlist *sgl = data->sg, *sg;
+	unsigned int sg_len = data->sg_len;
+	unsigned int i;
+
+	unsigned short dma_data_dir, timeout;
+	enum dma_transfer_direction slave_dirn;
+	unsigned int data_size = 0, log2_blksz;
+	unsigned int blocks = data->blocks;
+
+	struct mxs_ssp *ssp = &host->ssp;
+
+	u32 ignore_crc, get_resp, long_resp, read;
+	u32 ctrl0, cmd0, cmd1, val;
+
+	ignore_crc = (mmc_resp_type(cmd) & MMC_RSP_CRC) ?
+			0 : BM_SSP_CTRL0_IGNORE_CRC;
+	get_resp = (mmc_resp_type(cmd) & MMC_RSP_PRESENT) ?
+			BM_SSP_CTRL0_GET_RESP : 0;
+	long_resp = (mmc_resp_type(cmd) & MMC_RSP_136) ?
+			BM_SSP_CTRL0_LONG_RESP : 0;
+
+	if (data->flags & MMC_DATA_WRITE) {
+		dma_data_dir = DMA_TO_DEVICE;
+		slave_dirn = DMA_MEM_TO_DEV;
+		read = 0;
+	} else {
+		dma_data_dir = DMA_FROM_DEVICE;
+		slave_dirn = DMA_DEV_TO_MEM;
+		read = BM_SSP_CTRL0_READ;
+	}
+
+	ctrl0 = BF_SSP(host->bus_width, CTRL0_BUS_WIDTH) |
+		ignore_crc | get_resp | long_resp |
+		BM_SSP_CTRL0_DATA_XFER | read |
+		BM_SSP_CTRL0_WAIT_FOR_IRQ |
+		BM_SSP_CTRL0_ENABLE;
+
+	cmd0 = BF_SSP(cmd->opcode, CMD0_CMD);
+
+	/* get logarithm to base 2 of block size for setting register */
+	log2_blksz = ilog2(data->blksz);
+
+	/*
+	 * take special care of the case that data size from data->sg
+	 * is not equal to blocks x blksz
+	 */
+	for_each_sg(sgl, sg, sg_len, i)
+		data_size += sg->length;
+
+	if (data_size != data->blocks * data->blksz)
+		blocks = 1;
+
+	/* xfer count, block size and count need to be set differently */
+	if (ssp_is_old(ssp)) {
+		ctrl0 |= BF_SSP(data_size, CTRL0_XFER_COUNT);
+		cmd0 |= BF_SSP(log2_blksz, CMD0_BLOCK_SIZE) |
+			BF_SSP(blocks - 1, CMD0_BLOCK_COUNT);
+	} else {
+		writel(data_size, ssp->base + HW_SSP_XFER_SIZE);
+		writel(BF_SSP(log2_blksz, BLOCK_SIZE_BLOCK_SIZE) |
+		       BF_SSP(blocks - 1, BLOCK_SIZE_BLOCK_COUNT),
+		       ssp->base + HW_SSP_BLOCK_SIZE);
+	}
+
+	if (cmd->opcode == SD_IO_RW_EXTENDED)
+		cmd0 |= BM_SSP_CMD0_APPEND_8CYC;
+
+	cmd1 = cmd->arg;
+
+	if (host->sdio_irq_en) {
+		ctrl0 |= BM_SSP_CTRL0_SDIO_IRQ_CHECK;
+		cmd0 |= BM_SSP_CMD0_CONT_CLKING_EN | BM_SSP_CMD0_SLOW_CLKING_EN;
+	}
+
+	/* set the timeout count */
+	timeout = mxs_ns_to_ssp_ticks(ssp->clk_rate, data->timeout_ns);
+	val = readl(ssp->base + HW_SSP_TIMING(ssp));
+	val &= ~(BM_SSP_TIMING_TIMEOUT);
+	val |= BF_SSP(timeout, TIMING_TIMEOUT);
+	writel(val, ssp->base + HW_SSP_TIMING(ssp));
+
+	/* pio */
+	ssp->ssp_pio_words[0] = ctrl0;
+	ssp->ssp_pio_words[1] = cmd0;
+	ssp->ssp_pio_words[2] = cmd1;
+	ssp->dma_dir = DMA_NONE;
+	ssp->slave_dirn = DMA_TRANS_NONE;
+	desc = mxs_mmc_prep_dma(host, 0);
+	if (!desc)
+		goto out;
+
+	/* append data sg */
+	WARN_ON(host->data != NULL);
+	host->data = data;
+	ssp->dma_dir = dma_data_dir;
+	ssp->slave_dirn = slave_dirn;
+	desc = mxs_mmc_prep_dma(host, DMA_PREP_INTERRUPT | MXS_DMA_CTRL_WAIT4END);
+	if (!desc)
+		goto out;
+
+	dmaengine_submit(desc);
+	dma_async_issue_pending(ssp->dmach);
+	return;
+out:
+	dev_warn(mmc_dev(host->mmc),
+		 "%s: failed to prep dma\n", __func__);
+}
+
+static void mxs_mmc_start_cmd(struct mxs_mmc_host *host,
+			      struct mmc_command *cmd)
+{
+	host->cmd = cmd;
+
+	switch (mmc_cmd_type(cmd)) {
+	case MMC_CMD_BC:
+		mxs_mmc_bc(host);
+		break;
+	case MMC_CMD_BCR:
+		mxs_mmc_ac(host);
+		break;
+	case MMC_CMD_AC:
+		mxs_mmc_ac(host);
+		break;
+	case MMC_CMD_ADTC:
+		mxs_mmc_adtc(host);
+		break;
+	default:
+		dev_warn(mmc_dev(host->mmc),
+			 "%s: unknown MMC command\n", __func__);
+		break;
+	}
+}
+
+static void mxs_mmc_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct mxs_mmc_host *host = mmc_priv(mmc);
+
+	WARN_ON(host->mrq != NULL);
+	host->mrq = mrq;
+
+	if (mrq->sbc)
+		mxs_mmc_start_cmd(host, mrq->sbc);
+	else
+		mxs_mmc_start_cmd(host, mrq->cmd);
+}
+
+static void mxs_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct mxs_mmc_host *host = mmc_priv(mmc);
+
+	if (ios->bus_width == MMC_BUS_WIDTH_8)
+		host->bus_width = 2;
+	else if (ios->bus_width == MMC_BUS_WIDTH_4)
+		host->bus_width = 1;
+	else
+		host->bus_width = 0;
+
+	if (ios->clock)
+		mxs_ssp_set_clk_rate(&host->ssp, ios->clock);
+}
+
+static void mxs_mmc_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct mxs_mmc_host *host = mmc_priv(mmc);
+	struct mxs_ssp *ssp = &host->ssp;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	host->sdio_irq_en = enable;
+
+	if (enable) {
+		writel(BM_SSP_CTRL0_SDIO_IRQ_CHECK,
+		       ssp->base + HW_SSP_CTRL0 + STMP_OFFSET_REG_SET);
+		writel(BM_SSP_CTRL1_SDIO_IRQ_EN,
+		       ssp->base + HW_SSP_CTRL1(ssp) + STMP_OFFSET_REG_SET);
+	} else {
+		writel(BM_SSP_CTRL0_SDIO_IRQ_CHECK,
+		       ssp->base + HW_SSP_CTRL0 + STMP_OFFSET_REG_CLR);
+		writel(BM_SSP_CTRL1_SDIO_IRQ_EN,
+		       ssp->base + HW_SSP_CTRL1(ssp) + STMP_OFFSET_REG_CLR);
+	}
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (enable && readl(ssp->base + HW_SSP_STATUS(ssp)) &
+			BM_SSP_STATUS_SDIO_IRQ)
+		mmc_signal_sdio_irq(host->mmc);
+
+}
+
+static const struct mmc_host_ops mxs_mmc_ops = {
+	.request = mxs_mmc_request,
+	.get_ro = mmc_gpio_get_ro,
+	.get_cd = mxs_mmc_get_cd,
+	.set_ios = mxs_mmc_set_ios,
+	.enable_sdio_irq = mxs_mmc_enable_sdio_irq,
+};
+
+static const struct platform_device_id mxs_ssp_ids[] = {
+	{
+		.name = "imx23-mmc",
+		.driver_data = IMX23_SSP,
+	}, {
+		.name = "imx28-mmc",
+		.driver_data = IMX28_SSP,
+	}, {
+		/* sentinel */
+	}
+};
+MODULE_DEVICE_TABLE(platform, mxs_ssp_ids);
+
+static const struct of_device_id mxs_mmc_dt_ids[] = {
+	{ .compatible = "fsl,imx23-mmc", .data = (void *) IMX23_SSP, },
+	{ .compatible = "fsl,imx28-mmc", .data = (void *) IMX28_SSP, },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, mxs_mmc_dt_ids);
+
+static void mxs_mmc_regulator_disable(void *regulator)
+{
+	regulator_disable(regulator);
+}
+
+static int mxs_mmc_probe(struct platform_device *pdev)
+{
+	const struct of_device_id *of_id =
+			of_match_device(mxs_mmc_dt_ids, &pdev->dev);
+	struct device_node *np = pdev->dev.of_node;
+	struct mxs_mmc_host *host;
+	struct mmc_host *mmc;
+	int ret = 0, irq_err;
+	struct regulator *reg_vmmc;
+	struct mxs_ssp *ssp;
+
+	irq_err = platform_get_irq(pdev, 0);
+	if (irq_err < 0)
+		return irq_err;
+
+	mmc = mmc_alloc_host(sizeof(struct mxs_mmc_host), &pdev->dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	host = mmc_priv(mmc);
+	ssp = &host->ssp;
+	ssp->dev = &pdev->dev;
+	ssp->base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(ssp->base)) {
+		ret = PTR_ERR(ssp->base);
+		goto out_mmc_free;
+	}
+
+	ssp->devid = (enum mxs_ssp_id) of_id->data;
+
+	host->mmc = mmc;
+	host->sdio_irq_en = 0;
+
+	reg_vmmc = devm_regulator_get(&pdev->dev, "vmmc");
+	if (!IS_ERR(reg_vmmc)) {
+		ret = regulator_enable(reg_vmmc);
+		if (ret) {
+			dev_err(&pdev->dev,
+				"Failed to enable vmmc regulator: %d\n", ret);
+			goto out_mmc_free;
+		}
+
+		ret = devm_add_action_or_reset(&pdev->dev, mxs_mmc_regulator_disable,
+					       reg_vmmc);
+		if (ret)
+			goto out_mmc_free;
+	}
+
+	ssp->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(ssp->clk)) {
+		ret = PTR_ERR(ssp->clk);
+		goto out_mmc_free;
+	}
+	ret = clk_prepare_enable(ssp->clk);
+	if (ret)
+		goto out_mmc_free;
+
+	ret = mxs_mmc_reset(host);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to reset mmc: %d\n", ret);
+		goto out_clk_disable;
+	}
+
+	ssp->dmach = dma_request_chan(&pdev->dev, "rx-tx");
+	if (IS_ERR(ssp->dmach)) {
+		dev_err(mmc_dev(host->mmc),
+			"%s: failed to request dma\n", __func__);
+		ret = PTR_ERR(ssp->dmach);
+		goto out_clk_disable;
+	}
+
+	/* set mmc core parameters */
+	mmc->ops = &mxs_mmc_ops;
+	mmc->caps = MMC_CAP_SD_HIGHSPEED | MMC_CAP_MMC_HIGHSPEED |
+		    MMC_CAP_SDIO_IRQ | MMC_CAP_NEEDS_POLL | MMC_CAP_CMD23;
+
+	host->broken_cd = of_property_read_bool(np, "broken-cd");
+
+	mmc->f_min = 400000;
+	mmc->f_max = 288000000;
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto out_free_dma;
+
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+
+	mmc->max_segs = 52;
+	mmc->max_blk_size = 1 << 0xf;
+	mmc->max_blk_count = (ssp_is_old(ssp)) ? 0xff : 0xffffff;
+	mmc->max_req_size = (ssp_is_old(ssp)) ? 0xffff : 0xffffffff;
+	mmc->max_seg_size = dma_get_max_seg_size(ssp->dmach->device->dev);
+
+	platform_set_drvdata(pdev, mmc);
+
+	spin_lock_init(&host->lock);
+
+	ret = devm_request_irq(&pdev->dev, irq_err, mxs_mmc_irq_handler, 0,
+			       dev_name(&pdev->dev), host);
+	if (ret)
+		goto out_free_dma;
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto out_free_dma;
+
+	dev_info(mmc_dev(host->mmc), "initialized\n");
+
+	return 0;
+
+out_free_dma:
+	dma_release_channel(ssp->dmach);
+out_clk_disable:
+	clk_disable_unprepare(ssp->clk);
+out_mmc_free:
+	mmc_free_host(mmc);
+	return ret;
+}
+
+static int mxs_mmc_remove(struct platform_device *pdev)
+{
+	struct mmc_host *mmc = platform_get_drvdata(pdev);
+	struct mxs_mmc_host *host = mmc_priv(mmc);
+	struct mxs_ssp *ssp = &host->ssp;
+
+	mmc_remove_host(mmc);
+
+	if (ssp->dmach)
+		dma_release_channel(ssp->dmach);
+
+	clk_disable_unprepare(ssp->clk);
+
+	mmc_free_host(mmc);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int mxs_mmc_suspend(struct device *dev)
+{
+	struct mmc_host *mmc = dev_get_drvdata(dev);
+	struct mxs_mmc_host *host = mmc_priv(mmc);
+	struct mxs_ssp *ssp = &host->ssp;
+
+	clk_disable_unprepare(ssp->clk);
+	return 0;
+}
+
+static int mxs_mmc_resume(struct device *dev)
+{
+	struct mmc_host *mmc = dev_get_drvdata(dev);
+	struct mxs_mmc_host *host = mmc_priv(mmc);
+	struct mxs_ssp *ssp = &host->ssp;
+
+	return clk_prepare_enable(ssp->clk);
+}
+#endif
+
+static SIMPLE_DEV_PM_OPS(mxs_mmc_pm_ops, mxs_mmc_suspend, mxs_mmc_resume);
+
+static struct platform_driver mxs_mmc_driver = {
+	.probe		= mxs_mmc_probe,
+	.remove		= mxs_mmc_remove,
+	.id_table	= mxs_ssp_ids,
+	.driver		= {
+		.name	= DRIVER_NAME,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= &mxs_mmc_pm_ops,
+		.of_match_table = mxs_mmc_dt_ids,
+	},
+};
+
+module_platform_driver(mxs_mmc_driver);
+
+MODULE_DESCRIPTION("FREESCALE MXS MMC peripheral");
+MODULE_AUTHOR("Freescale Semiconductor");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:" DRIVER_NAME);
diff --git a/drivers/mmc/host/of_mmc_spi.c b/drivers/mmc/host/of_mmc_spi.c
new file mode 100644
index 000000000..3c4d950a4
--- /dev/null
+++ b/drivers/mmc/host/of_mmc_spi.c
@@ -0,0 +1,99 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * OpenFirmware bindings for the MMC-over-SPI driver
+ *
+ * Copyright (c) MontaVista Software, Inc. 2008.
+ *
+ * Author: Anton Vorontsov <avorontsov@ru.mvista.com>
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/device.h>
+#include <linux/slab.h>
+#include <linux/irq.h>
+#include <linux/of.h>
+#include <linux/of_irq.h>
+#include <linux/spi/spi.h>
+#include <linux/spi/mmc_spi.h>
+#include <linux/mmc/core.h>
+#include <linux/mmc/host.h>
+
+/* For archs that don't support NO_IRQ (such as mips), provide a dummy value */
+#ifndef NO_IRQ
+#define NO_IRQ 0
+#endif
+
+MODULE_LICENSE("GPL");
+
+struct of_mmc_spi {
+	int detect_irq;
+	struct mmc_spi_platform_data pdata;
+};
+
+static struct of_mmc_spi *to_of_mmc_spi(struct device *dev)
+{
+	return container_of(dev->platform_data, struct of_mmc_spi, pdata);
+}
+
+static int of_mmc_spi_init(struct device *dev,
+			   irqreturn_t (*irqhandler)(int, void *), void *mmc)
+{
+	struct of_mmc_spi *oms = to_of_mmc_spi(dev);
+
+	return request_threaded_irq(oms->detect_irq, NULL, irqhandler,
+					IRQF_ONESHOT, dev_name(dev), mmc);
+}
+
+static void of_mmc_spi_exit(struct device *dev, void *mmc)
+{
+	struct of_mmc_spi *oms = to_of_mmc_spi(dev);
+
+	free_irq(oms->detect_irq, mmc);
+}
+
+struct mmc_spi_platform_data *mmc_spi_get_pdata(struct spi_device *spi)
+{
+	struct device *dev = &spi->dev;
+	struct device_node *np = dev->of_node;
+	struct of_mmc_spi *oms;
+
+	if (dev->platform_data || !np)
+		return dev->platform_data;
+
+	oms = kzalloc(sizeof(*oms), GFP_KERNEL);
+	if (!oms)
+		return NULL;
+
+	if (mmc_of_parse_voltage(np, &oms->pdata.ocr_mask) <= 0)
+		goto err_ocr;
+
+	oms->detect_irq = irq_of_parse_and_map(np, 0);
+	if (oms->detect_irq != 0) {
+		oms->pdata.init = of_mmc_spi_init;
+		oms->pdata.exit = of_mmc_spi_exit;
+	} else {
+		oms->pdata.caps |= MMC_CAP_NEEDS_POLL;
+	}
+
+	dev->platform_data = &oms->pdata;
+	return dev->platform_data;
+err_ocr:
+	kfree(oms);
+	return NULL;
+}
+EXPORT_SYMBOL(mmc_spi_get_pdata);
+
+void mmc_spi_put_pdata(struct spi_device *spi)
+{
+	struct device *dev = &spi->dev;
+	struct device_node *np = dev->of_node;
+	struct of_mmc_spi *oms = to_of_mmc_spi(dev);
+
+	if (!dev->platform_data || !np)
+		return;
+
+	kfree(oms);
+	dev->platform_data = NULL;
+}
+EXPORT_SYMBOL(mmc_spi_put_pdata);
diff --git a/drivers/mmc/host/omap.c b/drivers/mmc/host/omap.c
new file mode 100644
index 000000000..eb978b75d
--- /dev/null
+++ b/drivers/mmc/host/omap.c
@@ -0,0 +1,1516 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ *  linux/drivers/mmc/host/omap.c
+ *
+ *  Copyright (C) 2004 Nokia Corporation
+ *  Written by Tuukka Tikkanen and Juha Yrjölä<juha.yrjola@nokia.com>
+ *  Misc hacks here and there by Tony Lindgren <tony@atomide.com>
+ *  Other hacks (DMA, SD, etc) by David Brownell
+ */
+
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/init.h>
+#include <linux/ioport.h>
+#include <linux/platform_device.h>
+#include <linux/interrupt.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/delay.h>
+#include <linux/spinlock.h>
+#include <linux/timer.h>
+#include <linux/of.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/mmc.h>
+#include <linux/clk.h>
+#include <linux/scatterlist.h>
+#include <linux/slab.h>
+#include <linux/platform_data/mmc-omap.h>
+
+
+#define	OMAP_MMC_REG_CMD	0x00
+#define	OMAP_MMC_REG_ARGL	0x01
+#define	OMAP_MMC_REG_ARGH	0x02
+#define	OMAP_MMC_REG_CON	0x03
+#define	OMAP_MMC_REG_STAT	0x04
+#define	OMAP_MMC_REG_IE		0x05
+#define	OMAP_MMC_REG_CTO	0x06
+#define	OMAP_MMC_REG_DTO	0x07
+#define	OMAP_MMC_REG_DATA	0x08
+#define	OMAP_MMC_REG_BLEN	0x09
+#define	OMAP_MMC_REG_NBLK	0x0a
+#define	OMAP_MMC_REG_BUF	0x0b
+#define	OMAP_MMC_REG_SDIO	0x0d
+#define	OMAP_MMC_REG_REV	0x0f
+#define	OMAP_MMC_REG_RSP0	0x10
+#define	OMAP_MMC_REG_RSP1	0x11
+#define	OMAP_MMC_REG_RSP2	0x12
+#define	OMAP_MMC_REG_RSP3	0x13
+#define	OMAP_MMC_REG_RSP4	0x14
+#define	OMAP_MMC_REG_RSP5	0x15
+#define	OMAP_MMC_REG_RSP6	0x16
+#define	OMAP_MMC_REG_RSP7	0x17
+#define	OMAP_MMC_REG_IOSR	0x18
+#define	OMAP_MMC_REG_SYSC	0x19
+#define	OMAP_MMC_REG_SYSS	0x1a
+
+#define	OMAP_MMC_STAT_CARD_ERR		(1 << 14)
+#define	OMAP_MMC_STAT_CARD_IRQ		(1 << 13)
+#define	OMAP_MMC_STAT_OCR_BUSY		(1 << 12)
+#define	OMAP_MMC_STAT_A_EMPTY		(1 << 11)
+#define	OMAP_MMC_STAT_A_FULL		(1 << 10)
+#define	OMAP_MMC_STAT_CMD_CRC		(1 <<  8)
+#define	OMAP_MMC_STAT_CMD_TOUT		(1 <<  7)
+#define	OMAP_MMC_STAT_DATA_CRC		(1 <<  6)
+#define	OMAP_MMC_STAT_DATA_TOUT		(1 <<  5)
+#define	OMAP_MMC_STAT_END_BUSY		(1 <<  4)
+#define	OMAP_MMC_STAT_END_OF_DATA	(1 <<  3)
+#define	OMAP_MMC_STAT_CARD_BUSY		(1 <<  2)
+#define	OMAP_MMC_STAT_END_OF_CMD	(1 <<  0)
+
+#define mmc_omap7xx()	(host->features & MMC_OMAP7XX)
+#define mmc_omap15xx()	(host->features & MMC_OMAP15XX)
+#define mmc_omap16xx()	(host->features & MMC_OMAP16XX)
+#define MMC_OMAP1_MASK	(MMC_OMAP7XX | MMC_OMAP15XX | MMC_OMAP16XX)
+#define mmc_omap1()	(host->features & MMC_OMAP1_MASK)
+#define mmc_omap2()	(!mmc_omap1())
+
+#define OMAP_MMC_REG(host, reg)		(OMAP_MMC_REG_##reg << (host)->reg_shift)
+#define OMAP_MMC_READ(host, reg)	__raw_readw((host)->virt_base + OMAP_MMC_REG(host, reg))
+#define OMAP_MMC_WRITE(host, reg, val)	__raw_writew((val), (host)->virt_base + OMAP_MMC_REG(host, reg))
+
+/*
+ * Command types
+ */
+#define OMAP_MMC_CMDTYPE_BC	0
+#define OMAP_MMC_CMDTYPE_BCR	1
+#define OMAP_MMC_CMDTYPE_AC	2
+#define OMAP_MMC_CMDTYPE_ADTC	3
+
+#define DRIVER_NAME "mmci-omap"
+
+/* Specifies how often in millisecs to poll for card status changes
+ * when the cover switch is open */
+#define OMAP_MMC_COVER_POLL_DELAY	500
+
+struct mmc_omap_host;
+
+struct mmc_omap_slot {
+	int			id;
+	unsigned int		vdd;
+	u16			saved_con;
+	u16			bus_mode;
+	u16			power_mode;
+	unsigned int		fclk_freq;
+
+	struct tasklet_struct	cover_tasklet;
+	struct timer_list       cover_timer;
+	unsigned		cover_open;
+
+	struct mmc_request      *mrq;
+	struct mmc_omap_host    *host;
+	struct mmc_host		*mmc;
+	struct omap_mmc_slot_data *pdata;
+};
+
+struct mmc_omap_host {
+	int			initialized;
+	struct mmc_request *	mrq;
+	struct mmc_command *	cmd;
+	struct mmc_data *	data;
+	struct mmc_host *	mmc;
+	struct device *		dev;
+	unsigned char		id; /* 16xx chips have 2 MMC blocks */
+	struct clk *		iclk;
+	struct clk *		fclk;
+	struct dma_chan		*dma_rx;
+	u32			dma_rx_burst;
+	struct dma_chan		*dma_tx;
+	u32			dma_tx_burst;
+	void __iomem		*virt_base;
+	unsigned int		phys_base;
+	int			irq;
+	unsigned char		bus_mode;
+	unsigned int		reg_shift;
+
+	struct work_struct	cmd_abort_work;
+	unsigned		abort:1;
+	struct timer_list	cmd_abort_timer;
+
+	struct work_struct      slot_release_work;
+	struct mmc_omap_slot    *next_slot;
+	struct work_struct      send_stop_work;
+	struct mmc_data		*stop_data;
+
+	unsigned int		sg_len;
+	int			sg_idx;
+	u16 *			buffer;
+	u32			buffer_bytes_left;
+	u32			total_bytes_left;
+
+	unsigned		features;
+	unsigned		brs_received:1, dma_done:1;
+	unsigned		dma_in_use:1;
+	spinlock_t		dma_lock;
+
+	struct mmc_omap_slot    *slots[OMAP_MMC_MAX_SLOTS];
+	struct mmc_omap_slot    *current_slot;
+	spinlock_t              slot_lock;
+	wait_queue_head_t       slot_wq;
+	int                     nr_slots;
+
+	struct timer_list       clk_timer;
+	spinlock_t		clk_lock;     /* for changing enabled state */
+	unsigned int            fclk_enabled:1;
+	struct workqueue_struct *mmc_omap_wq;
+
+	struct omap_mmc_platform_data *pdata;
+};
+
+
+static void mmc_omap_fclk_offdelay(struct mmc_omap_slot *slot)
+{
+	unsigned long tick_ns;
+
+	if (slot != NULL && slot->host->fclk_enabled && slot->fclk_freq > 0) {
+		tick_ns = DIV_ROUND_UP(NSEC_PER_SEC, slot->fclk_freq);
+		ndelay(8 * tick_ns);
+	}
+}
+
+static void mmc_omap_fclk_enable(struct mmc_omap_host *host, unsigned int enable)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->clk_lock, flags);
+	if (host->fclk_enabled != enable) {
+		host->fclk_enabled = enable;
+		if (enable)
+			clk_enable(host->fclk);
+		else
+			clk_disable(host->fclk);
+	}
+	spin_unlock_irqrestore(&host->clk_lock, flags);
+}
+
+static void mmc_omap_select_slot(struct mmc_omap_slot *slot, int claimed)
+{
+	struct mmc_omap_host *host = slot->host;
+	unsigned long flags;
+
+	if (claimed)
+		goto no_claim;
+	spin_lock_irqsave(&host->slot_lock, flags);
+	while (host->mmc != NULL) {
+		spin_unlock_irqrestore(&host->slot_lock, flags);
+		wait_event(host->slot_wq, host->mmc == NULL);
+		spin_lock_irqsave(&host->slot_lock, flags);
+	}
+	host->mmc = slot->mmc;
+	spin_unlock_irqrestore(&host->slot_lock, flags);
+no_claim:
+	del_timer(&host->clk_timer);
+	if (host->current_slot != slot || !claimed)
+		mmc_omap_fclk_offdelay(host->current_slot);
+
+	if (host->current_slot != slot) {
+		OMAP_MMC_WRITE(host, CON, slot->saved_con & 0xFC00);
+		if (host->pdata->switch_slot != NULL)
+			host->pdata->switch_slot(mmc_dev(slot->mmc), slot->id);
+		host->current_slot = slot;
+	}
+
+	if (claimed) {
+		mmc_omap_fclk_enable(host, 1);
+
+		/* Doing the dummy read here seems to work around some bug
+		 * at least in OMAP24xx silicon where the command would not
+		 * start after writing the CMD register. Sigh. */
+		OMAP_MMC_READ(host, CON);
+
+		OMAP_MMC_WRITE(host, CON, slot->saved_con);
+	} else
+		mmc_omap_fclk_enable(host, 0);
+}
+
+static void mmc_omap_start_request(struct mmc_omap_host *host,
+				   struct mmc_request *req);
+
+static void mmc_omap_slot_release_work(struct work_struct *work)
+{
+	struct mmc_omap_host *host = container_of(work, struct mmc_omap_host,
+						  slot_release_work);
+	struct mmc_omap_slot *next_slot = host->next_slot;
+	struct mmc_request *rq;
+
+	host->next_slot = NULL;
+	mmc_omap_select_slot(next_slot, 1);
+
+	rq = next_slot->mrq;
+	next_slot->mrq = NULL;
+	mmc_omap_start_request(host, rq);
+}
+
+static void mmc_omap_release_slot(struct mmc_omap_slot *slot, int clk_enabled)
+{
+	struct mmc_omap_host *host = slot->host;
+	unsigned long flags;
+	int i;
+
+	BUG_ON(slot == NULL || host->mmc == NULL);
+
+	if (clk_enabled)
+		/* Keeps clock running for at least 8 cycles on valid freq */
+		mod_timer(&host->clk_timer, jiffies  + HZ/10);
+	else {
+		del_timer(&host->clk_timer);
+		mmc_omap_fclk_offdelay(slot);
+		mmc_omap_fclk_enable(host, 0);
+	}
+
+	spin_lock_irqsave(&host->slot_lock, flags);
+	/* Check for any pending requests */
+	for (i = 0; i < host->nr_slots; i++) {
+		struct mmc_omap_slot *new_slot;
+
+		if (host->slots[i] == NULL || host->slots[i]->mrq == NULL)
+			continue;
+
+		BUG_ON(host->next_slot != NULL);
+		new_slot = host->slots[i];
+		/* The current slot should not have a request in queue */
+		BUG_ON(new_slot == host->current_slot);
+
+		host->next_slot = new_slot;
+		host->mmc = new_slot->mmc;
+		spin_unlock_irqrestore(&host->slot_lock, flags);
+		queue_work(host->mmc_omap_wq, &host->slot_release_work);
+		return;
+	}
+
+	host->mmc = NULL;
+	wake_up(&host->slot_wq);
+	spin_unlock_irqrestore(&host->slot_lock, flags);
+}
+
+static inline
+int mmc_omap_cover_is_open(struct mmc_omap_slot *slot)
+{
+	if (slot->pdata->get_cover_state)
+		return slot->pdata->get_cover_state(mmc_dev(slot->mmc),
+						    slot->id);
+	return 0;
+}
+
+static ssize_t
+mmc_omap_show_cover_switch(struct device *dev, struct device_attribute *attr,
+			   char *buf)
+{
+	struct mmc_host *mmc = container_of(dev, struct mmc_host, class_dev);
+	struct mmc_omap_slot *slot = mmc_priv(mmc);
+
+	return sprintf(buf, "%s\n", mmc_omap_cover_is_open(slot) ? "open" :
+		       "closed");
+}
+
+static DEVICE_ATTR(cover_switch, S_IRUGO, mmc_omap_show_cover_switch, NULL);
+
+static ssize_t
+mmc_omap_show_slot_name(struct device *dev, struct device_attribute *attr,
+			char *buf)
+{
+	struct mmc_host *mmc = container_of(dev, struct mmc_host, class_dev);
+	struct mmc_omap_slot *slot = mmc_priv(mmc);
+
+	return sprintf(buf, "%s\n", slot->pdata->name);
+}
+
+static DEVICE_ATTR(slot_name, S_IRUGO, mmc_omap_show_slot_name, NULL);
+
+static void
+mmc_omap_start_command(struct mmc_omap_host *host, struct mmc_command *cmd)
+{
+	u32 cmdreg;
+	u32 resptype;
+	u32 cmdtype;
+	u16 irq_mask;
+
+	host->cmd = cmd;
+
+	resptype = 0;
+	cmdtype = 0;
+
+	/* Our hardware needs to know exact type */
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE:
+		break;
+	case MMC_RSP_R1:
+	case MMC_RSP_R1B:
+		/* resp 1, 1b, 6, 7 */
+		resptype = 1;
+		break;
+	case MMC_RSP_R2:
+		resptype = 2;
+		break;
+	case MMC_RSP_R3:
+		resptype = 3;
+		break;
+	default:
+		dev_err(mmc_dev(host->mmc), "Invalid response type: %04x\n", mmc_resp_type(cmd));
+		break;
+	}
+
+	if (mmc_cmd_type(cmd) == MMC_CMD_ADTC) {
+		cmdtype = OMAP_MMC_CMDTYPE_ADTC;
+	} else if (mmc_cmd_type(cmd) == MMC_CMD_BC) {
+		cmdtype = OMAP_MMC_CMDTYPE_BC;
+	} else if (mmc_cmd_type(cmd) == MMC_CMD_BCR) {
+		cmdtype = OMAP_MMC_CMDTYPE_BCR;
+	} else {
+		cmdtype = OMAP_MMC_CMDTYPE_AC;
+	}
+
+	cmdreg = cmd->opcode | (resptype << 8) | (cmdtype << 12);
+
+	if (host->current_slot->bus_mode == MMC_BUSMODE_OPENDRAIN)
+		cmdreg |= 1 << 6;
+
+	if (cmd->flags & MMC_RSP_BUSY)
+		cmdreg |= 1 << 11;
+
+	if (host->data && !(host->data->flags & MMC_DATA_WRITE))
+		cmdreg |= 1 << 15;
+
+	mod_timer(&host->cmd_abort_timer, jiffies + HZ/2);
+
+	OMAP_MMC_WRITE(host, CTO, 200);
+	OMAP_MMC_WRITE(host, ARGL, cmd->arg & 0xffff);
+	OMAP_MMC_WRITE(host, ARGH, cmd->arg >> 16);
+	irq_mask = OMAP_MMC_STAT_A_EMPTY    | OMAP_MMC_STAT_A_FULL    |
+		   OMAP_MMC_STAT_CMD_CRC    | OMAP_MMC_STAT_CMD_TOUT  |
+		   OMAP_MMC_STAT_DATA_CRC   | OMAP_MMC_STAT_DATA_TOUT |
+		   OMAP_MMC_STAT_END_OF_CMD | OMAP_MMC_STAT_CARD_ERR  |
+		   OMAP_MMC_STAT_END_OF_DATA;
+	if (cmd->opcode == MMC_ERASE)
+		irq_mask &= ~OMAP_MMC_STAT_DATA_TOUT;
+	OMAP_MMC_WRITE(host, IE, irq_mask);
+	OMAP_MMC_WRITE(host, CMD, cmdreg);
+}
+
+static void
+mmc_omap_release_dma(struct mmc_omap_host *host, struct mmc_data *data,
+		     int abort)
+{
+	enum dma_data_direction dma_data_dir;
+	struct device *dev = mmc_dev(host->mmc);
+	struct dma_chan *c;
+
+	if (data->flags & MMC_DATA_WRITE) {
+		dma_data_dir = DMA_TO_DEVICE;
+		c = host->dma_tx;
+	} else {
+		dma_data_dir = DMA_FROM_DEVICE;
+		c = host->dma_rx;
+	}
+	if (c) {
+		if (data->error) {
+			dmaengine_terminate_all(c);
+			/* Claim nothing transferred on error... */
+			data->bytes_xfered = 0;
+		}
+		dev = c->device->dev;
+	}
+	dma_unmap_sg(dev, data->sg, host->sg_len, dma_data_dir);
+}
+
+static void mmc_omap_send_stop_work(struct work_struct *work)
+{
+	struct mmc_omap_host *host = container_of(work, struct mmc_omap_host,
+						  send_stop_work);
+	struct mmc_omap_slot *slot = host->current_slot;
+	struct mmc_data *data = host->stop_data;
+	unsigned long tick_ns;
+
+	tick_ns = DIV_ROUND_UP(NSEC_PER_SEC, slot->fclk_freq);
+	ndelay(8*tick_ns);
+
+	mmc_omap_start_command(host, data->stop);
+}
+
+static void
+mmc_omap_xfer_done(struct mmc_omap_host *host, struct mmc_data *data)
+{
+	if (host->dma_in_use)
+		mmc_omap_release_dma(host, data, data->error);
+
+	host->data = NULL;
+	host->sg_len = 0;
+
+	/* NOTE:  MMC layer will sometimes poll-wait CMD13 next, issuing
+	 * dozens of requests until the card finishes writing data.
+	 * It'd be cheaper to just wait till an EOFB interrupt arrives...
+	 */
+
+	if (!data->stop) {
+		struct mmc_host *mmc;
+
+		host->mrq = NULL;
+		mmc = host->mmc;
+		mmc_omap_release_slot(host->current_slot, 1);
+		mmc_request_done(mmc, data->mrq);
+		return;
+	}
+
+	host->stop_data = data;
+	queue_work(host->mmc_omap_wq, &host->send_stop_work);
+}
+
+static void
+mmc_omap_send_abort(struct mmc_omap_host *host, int maxloops)
+{
+	struct mmc_omap_slot *slot = host->current_slot;
+	unsigned int restarts, passes, timeout;
+	u16 stat = 0;
+
+	/* Sending abort takes 80 clocks. Have some extra and round up */
+	timeout = DIV_ROUND_UP(120 * USEC_PER_SEC, slot->fclk_freq);
+	restarts = 0;
+	while (restarts < maxloops) {
+		OMAP_MMC_WRITE(host, STAT, 0xFFFF);
+		OMAP_MMC_WRITE(host, CMD, (3 << 12) | (1 << 7));
+
+		passes = 0;
+		while (passes < timeout) {
+			stat = OMAP_MMC_READ(host, STAT);
+			if (stat & OMAP_MMC_STAT_END_OF_CMD)
+				goto out;
+			udelay(1);
+			passes++;
+		}
+
+		restarts++;
+	}
+out:
+	OMAP_MMC_WRITE(host, STAT, stat);
+}
+
+static void
+mmc_omap_abort_xfer(struct mmc_omap_host *host, struct mmc_data *data)
+{
+	if (host->dma_in_use)
+		mmc_omap_release_dma(host, data, 1);
+
+	host->data = NULL;
+	host->sg_len = 0;
+
+	mmc_omap_send_abort(host, 10000);
+}
+
+static void
+mmc_omap_end_of_data(struct mmc_omap_host *host, struct mmc_data *data)
+{
+	unsigned long flags;
+	int done;
+
+	if (!host->dma_in_use) {
+		mmc_omap_xfer_done(host, data);
+		return;
+	}
+	done = 0;
+	spin_lock_irqsave(&host->dma_lock, flags);
+	if (host->dma_done)
+		done = 1;
+	else
+		host->brs_received = 1;
+	spin_unlock_irqrestore(&host->dma_lock, flags);
+	if (done)
+		mmc_omap_xfer_done(host, data);
+}
+
+static void
+mmc_omap_dma_done(struct mmc_omap_host *host, struct mmc_data *data)
+{
+	unsigned long flags;
+	int done;
+
+	done = 0;
+	spin_lock_irqsave(&host->dma_lock, flags);
+	if (host->brs_received)
+		done = 1;
+	else
+		host->dma_done = 1;
+	spin_unlock_irqrestore(&host->dma_lock, flags);
+	if (done)
+		mmc_omap_xfer_done(host, data);
+}
+
+static void
+mmc_omap_cmd_done(struct mmc_omap_host *host, struct mmc_command *cmd)
+{
+	host->cmd = NULL;
+
+	del_timer(&host->cmd_abort_timer);
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136) {
+			/* response type 2 */
+			cmd->resp[3] =
+				OMAP_MMC_READ(host, RSP0) |
+				(OMAP_MMC_READ(host, RSP1) << 16);
+			cmd->resp[2] =
+				OMAP_MMC_READ(host, RSP2) |
+				(OMAP_MMC_READ(host, RSP3) << 16);
+			cmd->resp[1] =
+				OMAP_MMC_READ(host, RSP4) |
+				(OMAP_MMC_READ(host, RSP5) << 16);
+			cmd->resp[0] =
+				OMAP_MMC_READ(host, RSP6) |
+				(OMAP_MMC_READ(host, RSP7) << 16);
+		} else {
+			/* response types 1, 1b, 3, 4, 5, 6 */
+			cmd->resp[0] =
+				OMAP_MMC_READ(host, RSP6) |
+				(OMAP_MMC_READ(host, RSP7) << 16);
+		}
+	}
+
+	if (host->data == NULL || cmd->error) {
+		struct mmc_host *mmc;
+
+		if (host->data != NULL)
+			mmc_omap_abort_xfer(host, host->data);
+		host->mrq = NULL;
+		mmc = host->mmc;
+		mmc_omap_release_slot(host->current_slot, 1);
+		mmc_request_done(mmc, cmd->mrq);
+	}
+}
+
+/*
+ * Abort stuck command. Can occur when card is removed while it is being
+ * read.
+ */
+static void mmc_omap_abort_command(struct work_struct *work)
+{
+	struct mmc_omap_host *host = container_of(work, struct mmc_omap_host,
+						  cmd_abort_work);
+	BUG_ON(!host->cmd);
+
+	dev_dbg(mmc_dev(host->mmc), "Aborting stuck command CMD%d\n",
+		host->cmd->opcode);
+
+	if (host->cmd->error == 0)
+		host->cmd->error = -ETIMEDOUT;
+
+	if (host->data == NULL) {
+		struct mmc_command *cmd;
+		struct mmc_host    *mmc;
+
+		cmd = host->cmd;
+		host->cmd = NULL;
+		mmc_omap_send_abort(host, 10000);
+
+		host->mrq = NULL;
+		mmc = host->mmc;
+		mmc_omap_release_slot(host->current_slot, 1);
+		mmc_request_done(mmc, cmd->mrq);
+	} else
+		mmc_omap_cmd_done(host, host->cmd);
+
+	host->abort = 0;
+	enable_irq(host->irq);
+}
+
+static void
+mmc_omap_cmd_timer(struct timer_list *t)
+{
+	struct mmc_omap_host *host = from_timer(host, t, cmd_abort_timer);
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->slot_lock, flags);
+	if (host->cmd != NULL && !host->abort) {
+		OMAP_MMC_WRITE(host, IE, 0);
+		disable_irq(host->irq);
+		host->abort = 1;
+		queue_work(host->mmc_omap_wq, &host->cmd_abort_work);
+	}
+	spin_unlock_irqrestore(&host->slot_lock, flags);
+}
+
+/* PIO only */
+static void
+mmc_omap_sg_to_buf(struct mmc_omap_host *host)
+{
+	struct scatterlist *sg;
+
+	sg = host->data->sg + host->sg_idx;
+	host->buffer_bytes_left = sg->length;
+	host->buffer = sg_virt(sg);
+	if (host->buffer_bytes_left > host->total_bytes_left)
+		host->buffer_bytes_left = host->total_bytes_left;
+}
+
+static void
+mmc_omap_clk_timer(struct timer_list *t)
+{
+	struct mmc_omap_host *host = from_timer(host, t, clk_timer);
+
+	mmc_omap_fclk_enable(host, 0);
+}
+
+/* PIO only */
+static void
+mmc_omap_xfer_data(struct mmc_omap_host *host, int write)
+{
+	int n, nwords;
+
+	if (host->buffer_bytes_left == 0) {
+		host->sg_idx++;
+		BUG_ON(host->sg_idx == host->sg_len);
+		mmc_omap_sg_to_buf(host);
+	}
+	n = 64;
+	if (n > host->buffer_bytes_left)
+		n = host->buffer_bytes_left;
+
+	/* Round up to handle odd number of bytes to transfer */
+	nwords = DIV_ROUND_UP(n, 2);
+
+	host->buffer_bytes_left -= n;
+	host->total_bytes_left -= n;
+	host->data->bytes_xfered += n;
+
+	if (write) {
+		__raw_writesw(host->virt_base + OMAP_MMC_REG(host, DATA),
+			      host->buffer, nwords);
+	} else {
+		__raw_readsw(host->virt_base + OMAP_MMC_REG(host, DATA),
+			     host->buffer, nwords);
+	}
+
+	host->buffer += nwords;
+}
+
+#ifdef CONFIG_MMC_DEBUG
+static void mmc_omap_report_irq(struct mmc_omap_host *host, u16 status)
+{
+	static const char *mmc_omap_status_bits[] = {
+		"EOC", "CD", "CB", "BRS", "EOFB", "DTO", "DCRC", "CTO",
+		"CCRC", "CRW", "AF", "AE", "OCRB", "CIRQ", "CERR"
+	};
+	int i;
+	char res[64], *buf = res;
+
+	buf += sprintf(buf, "MMC IRQ 0x%x:", status);
+
+	for (i = 0; i < ARRAY_SIZE(mmc_omap_status_bits); i++)
+		if (status & (1 << i))
+			buf += sprintf(buf, " %s", mmc_omap_status_bits[i]);
+	dev_vdbg(mmc_dev(host->mmc), "%s\n", res);
+}
+#else
+static void mmc_omap_report_irq(struct mmc_omap_host *host, u16 status)
+{
+}
+#endif
+
+
+static irqreturn_t mmc_omap_irq(int irq, void *dev_id)
+{
+	struct mmc_omap_host * host = (struct mmc_omap_host *)dev_id;
+	u16 status;
+	int end_command;
+	int end_transfer;
+	int transfer_error, cmd_error;
+
+	if (host->cmd == NULL && host->data == NULL) {
+		status = OMAP_MMC_READ(host, STAT);
+		dev_info(mmc_dev(host->slots[0]->mmc),
+			 "Spurious IRQ 0x%04x\n", status);
+		if (status != 0) {
+			OMAP_MMC_WRITE(host, STAT, status);
+			OMAP_MMC_WRITE(host, IE, 0);
+		}
+		return IRQ_HANDLED;
+	}
+
+	end_command = 0;
+	end_transfer = 0;
+	transfer_error = 0;
+	cmd_error = 0;
+
+	while ((status = OMAP_MMC_READ(host, STAT)) != 0) {
+		int cmd;
+
+		OMAP_MMC_WRITE(host, STAT, status);
+		if (host->cmd != NULL)
+			cmd = host->cmd->opcode;
+		else
+			cmd = -1;
+		dev_dbg(mmc_dev(host->mmc), "MMC IRQ %04x (CMD %d): ",
+			status, cmd);
+		mmc_omap_report_irq(host, status);
+
+		if (host->total_bytes_left) {
+			if ((status & OMAP_MMC_STAT_A_FULL) ||
+			    (status & OMAP_MMC_STAT_END_OF_DATA))
+				mmc_omap_xfer_data(host, 0);
+			if (status & OMAP_MMC_STAT_A_EMPTY)
+				mmc_omap_xfer_data(host, 1);
+		}
+
+		if (status & OMAP_MMC_STAT_END_OF_DATA)
+			end_transfer = 1;
+
+		if (status & OMAP_MMC_STAT_DATA_TOUT) {
+			dev_dbg(mmc_dev(host->mmc), "data timeout (CMD%d)\n",
+				cmd);
+			if (host->data) {
+				host->data->error = -ETIMEDOUT;
+				transfer_error = 1;
+			}
+		}
+
+		if (status & OMAP_MMC_STAT_DATA_CRC) {
+			if (host->data) {
+				host->data->error = -EILSEQ;
+				dev_dbg(mmc_dev(host->mmc),
+					 "data CRC error, bytes left %d\n",
+					host->total_bytes_left);
+				transfer_error = 1;
+			} else {
+				dev_dbg(mmc_dev(host->mmc), "data CRC error\n");
+			}
+		}
+
+		if (status & OMAP_MMC_STAT_CMD_TOUT) {
+			/* Timeouts are routine with some commands */
+			if (host->cmd) {
+				struct mmc_omap_slot *slot =
+					host->current_slot;
+				if (slot == NULL ||
+				    !mmc_omap_cover_is_open(slot))
+					dev_err(mmc_dev(host->mmc),
+						"command timeout (CMD%d)\n",
+						cmd);
+				host->cmd->error = -ETIMEDOUT;
+				end_command = 1;
+				cmd_error = 1;
+			}
+		}
+
+		if (status & OMAP_MMC_STAT_CMD_CRC) {
+			if (host->cmd) {
+				dev_err(mmc_dev(host->mmc),
+					"command CRC error (CMD%d, arg 0x%08x)\n",
+					cmd, host->cmd->arg);
+				host->cmd->error = -EILSEQ;
+				end_command = 1;
+				cmd_error = 1;
+			} else
+				dev_err(mmc_dev(host->mmc),
+					"command CRC error without cmd?\n");
+		}
+
+		if (status & OMAP_MMC_STAT_CARD_ERR) {
+			dev_dbg(mmc_dev(host->mmc),
+				"ignoring card status error (CMD%d)\n",
+				cmd);
+			end_command = 1;
+		}
+
+		/*
+		 * NOTE: On 1610 the END_OF_CMD may come too early when
+		 * starting a write
+		 */
+		if ((status & OMAP_MMC_STAT_END_OF_CMD) &&
+		    (!(status & OMAP_MMC_STAT_A_EMPTY))) {
+			end_command = 1;
+		}
+	}
+
+	if (cmd_error && host->data) {
+		del_timer(&host->cmd_abort_timer);
+		host->abort = 1;
+		OMAP_MMC_WRITE(host, IE, 0);
+		disable_irq_nosync(host->irq);
+		queue_work(host->mmc_omap_wq, &host->cmd_abort_work);
+		return IRQ_HANDLED;
+	}
+
+	if (end_command && host->cmd)
+		mmc_omap_cmd_done(host, host->cmd);
+	if (host->data != NULL) {
+		if (transfer_error)
+			mmc_omap_xfer_done(host, host->data);
+		else if (end_transfer)
+			mmc_omap_end_of_data(host, host->data);
+	}
+
+	return IRQ_HANDLED;
+}
+
+void omap_mmc_notify_cover_event(struct device *dev, int num, int is_closed)
+{
+	int cover_open;
+	struct mmc_omap_host *host = dev_get_drvdata(dev);
+	struct mmc_omap_slot *slot = host->slots[num];
+
+	BUG_ON(num >= host->nr_slots);
+
+	/* Other subsystems can call in here before we're initialised. */
+	if (host->nr_slots == 0 || !host->slots[num])
+		return;
+
+	cover_open = mmc_omap_cover_is_open(slot);
+	if (cover_open != slot->cover_open) {
+		slot->cover_open = cover_open;
+		sysfs_notify(&slot->mmc->class_dev.kobj, NULL, "cover_switch");
+	}
+
+	tasklet_hi_schedule(&slot->cover_tasklet);
+}
+
+static void mmc_omap_cover_timer(struct timer_list *t)
+{
+	struct mmc_omap_slot *slot = from_timer(slot, t, cover_timer);
+	tasklet_schedule(&slot->cover_tasklet);
+}
+
+static void mmc_omap_cover_handler(unsigned long param)
+{
+	struct mmc_omap_slot *slot = (struct mmc_omap_slot *)param;
+	int cover_open = mmc_omap_cover_is_open(slot);
+
+	mmc_detect_change(slot->mmc, 0);
+	if (!cover_open)
+		return;
+
+	/*
+	 * If no card is inserted, we postpone polling until
+	 * the cover has been closed.
+	 */
+	if (slot->mmc->card == NULL)
+		return;
+
+	mod_timer(&slot->cover_timer,
+		  jiffies + msecs_to_jiffies(OMAP_MMC_COVER_POLL_DELAY));
+}
+
+static void mmc_omap_dma_callback(void *priv)
+{
+	struct mmc_omap_host *host = priv;
+	struct mmc_data *data = host->data;
+
+	/* If we got to the end of DMA, assume everything went well */
+	data->bytes_xfered += data->blocks * data->blksz;
+
+	mmc_omap_dma_done(host, data);
+}
+
+static inline void set_cmd_timeout(struct mmc_omap_host *host, struct mmc_request *req)
+{
+	u16 reg;
+
+	reg = OMAP_MMC_READ(host, SDIO);
+	reg &= ~(1 << 5);
+	OMAP_MMC_WRITE(host, SDIO, reg);
+	/* Set maximum timeout */
+	OMAP_MMC_WRITE(host, CTO, 0xfd);
+}
+
+static inline void set_data_timeout(struct mmc_omap_host *host, struct mmc_request *req)
+{
+	unsigned int timeout, cycle_ns;
+	u16 reg;
+
+	cycle_ns = 1000000000 / host->current_slot->fclk_freq;
+	timeout = req->data->timeout_ns / cycle_ns;
+	timeout += req->data->timeout_clks;
+
+	/* Check if we need to use timeout multiplier register */
+	reg = OMAP_MMC_READ(host, SDIO);
+	if (timeout > 0xffff) {
+		reg |= (1 << 5);
+		timeout /= 1024;
+	} else
+		reg &= ~(1 << 5);
+	OMAP_MMC_WRITE(host, SDIO, reg);
+	OMAP_MMC_WRITE(host, DTO, timeout);
+}
+
+static void
+mmc_omap_prepare_data(struct mmc_omap_host *host, struct mmc_request *req)
+{
+	struct mmc_data *data = req->data;
+	int i, use_dma = 1, block_size;
+	struct scatterlist *sg;
+	unsigned sg_len;
+
+	host->data = data;
+	if (data == NULL) {
+		OMAP_MMC_WRITE(host, BLEN, 0);
+		OMAP_MMC_WRITE(host, NBLK, 0);
+		OMAP_MMC_WRITE(host, BUF, 0);
+		host->dma_in_use = 0;
+		set_cmd_timeout(host, req);
+		return;
+	}
+
+	block_size = data->blksz;
+
+	OMAP_MMC_WRITE(host, NBLK, data->blocks - 1);
+	OMAP_MMC_WRITE(host, BLEN, block_size - 1);
+	set_data_timeout(host, req);
+
+	/* cope with calling layer confusion; it issues "single
+	 * block" writes using multi-block scatterlists.
+	 */
+	sg_len = (data->blocks == 1) ? 1 : data->sg_len;
+
+	/* Only do DMA for entire blocks */
+	for_each_sg(data->sg, sg, sg_len, i) {
+		if ((sg->length % block_size) != 0) {
+			use_dma = 0;
+			break;
+		}
+	}
+
+	host->sg_idx = 0;
+	if (use_dma) {
+		enum dma_data_direction dma_data_dir;
+		struct dma_async_tx_descriptor *tx;
+		struct dma_chan *c;
+		u32 burst, *bp;
+		u16 buf;
+
+		/*
+		 * FIFO is 16x2 bytes on 15xx, and 32x2 bytes on 16xx
+		 * and 24xx. Use 16 or 32 word frames when the
+		 * blocksize is at least that large. Blocksize is
+		 * usually 512 bytes; but not for some SD reads.
+		 */
+		burst = mmc_omap15xx() ? 32 : 64;
+		if (burst > data->blksz)
+			burst = data->blksz;
+
+		burst >>= 1;
+
+		if (data->flags & MMC_DATA_WRITE) {
+			c = host->dma_tx;
+			bp = &host->dma_tx_burst;
+			buf = 0x0f80 | (burst - 1) << 0;
+			dma_data_dir = DMA_TO_DEVICE;
+		} else {
+			c = host->dma_rx;
+			bp = &host->dma_rx_burst;
+			buf = 0x800f | (burst - 1) << 8;
+			dma_data_dir = DMA_FROM_DEVICE;
+		}
+
+		if (!c)
+			goto use_pio;
+
+		/* Only reconfigure if we have a different burst size */
+		if (*bp != burst) {
+			struct dma_slave_config cfg = {
+				.src_addr = host->phys_base +
+					    OMAP_MMC_REG(host, DATA),
+				.dst_addr = host->phys_base +
+					    OMAP_MMC_REG(host, DATA),
+				.src_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES,
+				.dst_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES,
+				.src_maxburst = burst,
+				.dst_maxburst = burst,
+			};
+
+			if (dmaengine_slave_config(c, &cfg))
+				goto use_pio;
+
+			*bp = burst;
+		}
+
+		host->sg_len = dma_map_sg(c->device->dev, data->sg, sg_len,
+					  dma_data_dir);
+		if (host->sg_len == 0)
+			goto use_pio;
+
+		tx = dmaengine_prep_slave_sg(c, data->sg, host->sg_len,
+			data->flags & MMC_DATA_WRITE ? DMA_MEM_TO_DEV : DMA_DEV_TO_MEM,
+			DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+		if (!tx)
+			goto use_pio;
+
+		OMAP_MMC_WRITE(host, BUF, buf);
+
+		tx->callback = mmc_omap_dma_callback;
+		tx->callback_param = host;
+		dmaengine_submit(tx);
+		host->brs_received = 0;
+		host->dma_done = 0;
+		host->dma_in_use = 1;
+		return;
+	}
+ use_pio:
+
+	/* Revert to PIO? */
+	OMAP_MMC_WRITE(host, BUF, 0x1f1f);
+	host->total_bytes_left = data->blocks * block_size;
+	host->sg_len = sg_len;
+	mmc_omap_sg_to_buf(host);
+	host->dma_in_use = 0;
+}
+
+static void mmc_omap_start_request(struct mmc_omap_host *host,
+				   struct mmc_request *req)
+{
+	BUG_ON(host->mrq != NULL);
+
+	host->mrq = req;
+
+	/* only touch fifo AFTER the controller readies it */
+	mmc_omap_prepare_data(host, req);
+	mmc_omap_start_command(host, req->cmd);
+	if (host->dma_in_use) {
+		struct dma_chan *c = host->data->flags & MMC_DATA_WRITE ?
+				host->dma_tx : host->dma_rx;
+
+		dma_async_issue_pending(c);
+	}
+}
+
+static void mmc_omap_request(struct mmc_host *mmc, struct mmc_request *req)
+{
+	struct mmc_omap_slot *slot = mmc_priv(mmc);
+	struct mmc_omap_host *host = slot->host;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->slot_lock, flags);
+	if (host->mmc != NULL) {
+		BUG_ON(slot->mrq != NULL);
+		slot->mrq = req;
+		spin_unlock_irqrestore(&host->slot_lock, flags);
+		return;
+	} else
+		host->mmc = mmc;
+	spin_unlock_irqrestore(&host->slot_lock, flags);
+	mmc_omap_select_slot(slot, 1);
+	mmc_omap_start_request(host, req);
+}
+
+static void mmc_omap_set_power(struct mmc_omap_slot *slot, int power_on,
+				int vdd)
+{
+	struct mmc_omap_host *host;
+
+	host = slot->host;
+
+	if (slot->pdata->set_power != NULL)
+		slot->pdata->set_power(mmc_dev(slot->mmc), slot->id, power_on,
+					vdd);
+	if (mmc_omap2()) {
+		u16 w;
+
+		if (power_on) {
+			w = OMAP_MMC_READ(host, CON);
+			OMAP_MMC_WRITE(host, CON, w | (1 << 11));
+		} else {
+			w = OMAP_MMC_READ(host, CON);
+			OMAP_MMC_WRITE(host, CON, w & ~(1 << 11));
+		}
+	}
+}
+
+static int mmc_omap_calc_divisor(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct mmc_omap_slot *slot = mmc_priv(mmc);
+	struct mmc_omap_host *host = slot->host;
+	int func_clk_rate = clk_get_rate(host->fclk);
+	int dsor;
+
+	if (ios->clock == 0)
+		return 0;
+
+	dsor = func_clk_rate / ios->clock;
+	if (dsor < 1)
+		dsor = 1;
+
+	if (func_clk_rate / dsor > ios->clock)
+		dsor++;
+
+	if (dsor > 250)
+		dsor = 250;
+
+	slot->fclk_freq = func_clk_rate / dsor;
+
+	if (ios->bus_width == MMC_BUS_WIDTH_4)
+		dsor |= 1 << 15;
+
+	return dsor;
+}
+
+static void mmc_omap_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct mmc_omap_slot *slot = mmc_priv(mmc);
+	struct mmc_omap_host *host = slot->host;
+	int i, dsor;
+	int clk_enabled, init_stream;
+
+	mmc_omap_select_slot(slot, 0);
+
+	dsor = mmc_omap_calc_divisor(mmc, ios);
+
+	if (ios->vdd != slot->vdd)
+		slot->vdd = ios->vdd;
+
+	clk_enabled = 0;
+	init_stream = 0;
+	switch (ios->power_mode) {
+	case MMC_POWER_OFF:
+		mmc_omap_set_power(slot, 0, ios->vdd);
+		break;
+	case MMC_POWER_UP:
+		/* Cannot touch dsor yet, just power up MMC */
+		mmc_omap_set_power(slot, 1, ios->vdd);
+		slot->power_mode = ios->power_mode;
+		goto exit;
+	case MMC_POWER_ON:
+		mmc_omap_fclk_enable(host, 1);
+		clk_enabled = 1;
+		dsor |= 1 << 11;
+		if (slot->power_mode != MMC_POWER_ON)
+			init_stream = 1;
+		break;
+	}
+	slot->power_mode = ios->power_mode;
+
+	if (slot->bus_mode != ios->bus_mode) {
+		if (slot->pdata->set_bus_mode != NULL)
+			slot->pdata->set_bus_mode(mmc_dev(mmc), slot->id,
+						  ios->bus_mode);
+		slot->bus_mode = ios->bus_mode;
+	}
+
+	/* On insanely high arm_per frequencies something sometimes
+	 * goes somehow out of sync, and the POW bit is not being set,
+	 * which results in the while loop below getting stuck.
+	 * Writing to the CON register twice seems to do the trick. */
+	for (i = 0; i < 2; i++)
+		OMAP_MMC_WRITE(host, CON, dsor);
+	slot->saved_con = dsor;
+	if (init_stream) {
+		/* worst case at 400kHz, 80 cycles makes 200 microsecs */
+		int usecs = 250;
+
+		/* Send clock cycles, poll completion */
+		OMAP_MMC_WRITE(host, IE, 0);
+		OMAP_MMC_WRITE(host, STAT, 0xffff);
+		OMAP_MMC_WRITE(host, CMD, 1 << 7);
+		while (usecs > 0 && (OMAP_MMC_READ(host, STAT) & 1) == 0) {
+			udelay(1);
+			usecs--;
+		}
+		OMAP_MMC_WRITE(host, STAT, 1);
+	}
+
+exit:
+	mmc_omap_release_slot(slot, clk_enabled);
+}
+
+static const struct mmc_host_ops mmc_omap_ops = {
+	.request	= mmc_omap_request,
+	.set_ios	= mmc_omap_set_ios,
+};
+
+static int mmc_omap_new_slot(struct mmc_omap_host *host, int id)
+{
+	struct mmc_omap_slot *slot = NULL;
+	struct mmc_host *mmc;
+	int r;
+
+	mmc = mmc_alloc_host(sizeof(struct mmc_omap_slot), host->dev);
+	if (mmc == NULL)
+		return -ENOMEM;
+
+	slot = mmc_priv(mmc);
+	slot->host = host;
+	slot->mmc = mmc;
+	slot->id = id;
+	slot->power_mode = MMC_POWER_UNDEFINED;
+	slot->pdata = &host->pdata->slots[id];
+
+	host->slots[id] = slot;
+
+	mmc->caps = 0;
+	if (host->pdata->slots[id].wires >= 4)
+		mmc->caps |= MMC_CAP_4_BIT_DATA;
+
+	mmc->ops = &mmc_omap_ops;
+	mmc->f_min = 400000;
+
+	if (mmc_omap2())
+		mmc->f_max = 48000000;
+	else
+		mmc->f_max = 24000000;
+	if (host->pdata->max_freq)
+		mmc->f_max = min(host->pdata->max_freq, mmc->f_max);
+	mmc->ocr_avail = slot->pdata->ocr_mask;
+
+	/* Use scatterlist DMA to reduce per-transfer costs.
+	 * NOTE max_seg_size assumption that small blocks aren't
+	 * normally used (except e.g. for reading SD registers).
+	 */
+	mmc->max_segs = 32;
+	mmc->max_blk_size = 2048;	/* BLEN is 11 bits (+1) */
+	mmc->max_blk_count = 2048;	/* NBLK is 11 bits (+1) */
+	mmc->max_req_size = mmc->max_blk_size * mmc->max_blk_count;
+	mmc->max_seg_size = mmc->max_req_size;
+
+	if (slot->pdata->get_cover_state != NULL) {
+		timer_setup(&slot->cover_timer, mmc_omap_cover_timer, 0);
+		tasklet_init(&slot->cover_tasklet, mmc_omap_cover_handler,
+			     (unsigned long)slot);
+	}
+
+	r = mmc_add_host(mmc);
+	if (r < 0)
+		goto err_remove_host;
+
+	if (slot->pdata->name != NULL) {
+		r = device_create_file(&mmc->class_dev,
+					&dev_attr_slot_name);
+		if (r < 0)
+			goto err_remove_host;
+	}
+
+	if (slot->pdata->get_cover_state != NULL) {
+		r = device_create_file(&mmc->class_dev,
+					&dev_attr_cover_switch);
+		if (r < 0)
+			goto err_remove_slot_name;
+		tasklet_schedule(&slot->cover_tasklet);
+	}
+
+	return 0;
+
+err_remove_slot_name:
+	if (slot->pdata->name != NULL)
+		device_remove_file(&mmc->class_dev, &dev_attr_slot_name);
+err_remove_host:
+	mmc_remove_host(mmc);
+	mmc_free_host(mmc);
+	return r;
+}
+
+static void mmc_omap_remove_slot(struct mmc_omap_slot *slot)
+{
+	struct mmc_host *mmc = slot->mmc;
+
+	if (slot->pdata->name != NULL)
+		device_remove_file(&mmc->class_dev, &dev_attr_slot_name);
+	if (slot->pdata->get_cover_state != NULL)
+		device_remove_file(&mmc->class_dev, &dev_attr_cover_switch);
+
+	tasklet_kill(&slot->cover_tasklet);
+	del_timer_sync(&slot->cover_timer);
+	flush_workqueue(slot->host->mmc_omap_wq);
+
+	mmc_remove_host(mmc);
+	mmc_free_host(mmc);
+}
+
+static int mmc_omap_probe(struct platform_device *pdev)
+{
+	struct omap_mmc_platform_data *pdata = pdev->dev.platform_data;
+	struct mmc_omap_host *host = NULL;
+	struct resource *res;
+	int i, ret = 0;
+	int irq;
+
+	if (pdata == NULL) {
+		dev_err(&pdev->dev, "platform data missing\n");
+		return -ENXIO;
+	}
+	if (pdata->nr_slots == 0) {
+		dev_err(&pdev->dev, "no slots\n");
+		return -EPROBE_DEFER;
+	}
+
+	host = devm_kzalloc(&pdev->dev, sizeof(struct mmc_omap_host),
+			    GFP_KERNEL);
+	if (host == NULL)
+		return -ENOMEM;
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return irq;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	host->virt_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(host->virt_base))
+		return PTR_ERR(host->virt_base);
+
+	INIT_WORK(&host->slot_release_work, mmc_omap_slot_release_work);
+	INIT_WORK(&host->send_stop_work, mmc_omap_send_stop_work);
+
+	INIT_WORK(&host->cmd_abort_work, mmc_omap_abort_command);
+	timer_setup(&host->cmd_abort_timer, mmc_omap_cmd_timer, 0);
+
+	spin_lock_init(&host->clk_lock);
+	timer_setup(&host->clk_timer, mmc_omap_clk_timer, 0);
+
+	spin_lock_init(&host->dma_lock);
+	spin_lock_init(&host->slot_lock);
+	init_waitqueue_head(&host->slot_wq);
+
+	host->pdata = pdata;
+	host->features = host->pdata->slots[0].features;
+	host->dev = &pdev->dev;
+	platform_set_drvdata(pdev, host);
+
+	host->id = pdev->id;
+	host->irq = irq;
+	host->phys_base = res->start;
+	host->iclk = clk_get(&pdev->dev, "ick");
+	if (IS_ERR(host->iclk))
+		return PTR_ERR(host->iclk);
+	clk_enable(host->iclk);
+
+	host->fclk = clk_get(&pdev->dev, "fck");
+	if (IS_ERR(host->fclk)) {
+		ret = PTR_ERR(host->fclk);
+		goto err_free_iclk;
+	}
+
+	host->dma_tx_burst = -1;
+	host->dma_rx_burst = -1;
+
+	host->dma_tx = dma_request_chan(&pdev->dev, "tx");
+	if (IS_ERR(host->dma_tx)) {
+		ret = PTR_ERR(host->dma_tx);
+		if (ret == -EPROBE_DEFER) {
+			clk_put(host->fclk);
+			goto err_free_iclk;
+		}
+
+		host->dma_tx = NULL;
+		dev_warn(host->dev, "TX DMA channel request failed\n");
+	}
+
+	host->dma_rx = dma_request_chan(&pdev->dev, "rx");
+	if (IS_ERR(host->dma_rx)) {
+		ret = PTR_ERR(host->dma_rx);
+		if (ret == -EPROBE_DEFER) {
+			if (host->dma_tx)
+				dma_release_channel(host->dma_tx);
+			clk_put(host->fclk);
+			goto err_free_iclk;
+		}
+
+		host->dma_rx = NULL;
+		dev_warn(host->dev, "RX DMA channel request failed\n");
+	}
+
+	ret = request_irq(host->irq, mmc_omap_irq, 0, DRIVER_NAME, host);
+	if (ret)
+		goto err_free_dma;
+
+	if (pdata->init != NULL) {
+		ret = pdata->init(&pdev->dev);
+		if (ret < 0)
+			goto err_free_irq;
+	}
+
+	host->nr_slots = pdata->nr_slots;
+	host->reg_shift = (mmc_omap7xx() ? 1 : 2);
+
+	host->mmc_omap_wq = alloc_workqueue("mmc_omap", 0, 0);
+	if (!host->mmc_omap_wq) {
+		ret = -ENOMEM;
+		goto err_plat_cleanup;
+	}
+
+	for (i = 0; i < pdata->nr_slots; i++) {
+		ret = mmc_omap_new_slot(host, i);
+		if (ret < 0) {
+			while (--i >= 0)
+				mmc_omap_remove_slot(host->slots[i]);
+
+			goto err_destroy_wq;
+		}
+	}
+
+	return 0;
+
+err_destroy_wq:
+	destroy_workqueue(host->mmc_omap_wq);
+err_plat_cleanup:
+	if (pdata->cleanup)
+		pdata->cleanup(&pdev->dev);
+err_free_irq:
+	free_irq(host->irq, host);
+err_free_dma:
+	if (host->dma_tx)
+		dma_release_channel(host->dma_tx);
+	if (host->dma_rx)
+		dma_release_channel(host->dma_rx);
+	clk_put(host->fclk);
+err_free_iclk:
+	clk_disable(host->iclk);
+	clk_put(host->iclk);
+	return ret;
+}
+
+static int mmc_omap_remove(struct platform_device *pdev)
+{
+	struct mmc_omap_host *host = platform_get_drvdata(pdev);
+	int i;
+
+	BUG_ON(host == NULL);
+
+	for (i = 0; i < host->nr_slots; i++)
+		mmc_omap_remove_slot(host->slots[i]);
+
+	if (host->pdata->cleanup)
+		host->pdata->cleanup(&pdev->dev);
+
+	mmc_omap_fclk_enable(host, 0);
+	free_irq(host->irq, host);
+	clk_put(host->fclk);
+	clk_disable(host->iclk);
+	clk_put(host->iclk);
+
+	if (host->dma_tx)
+		dma_release_channel(host->dma_tx);
+	if (host->dma_rx)
+		dma_release_channel(host->dma_rx);
+
+	destroy_workqueue(host->mmc_omap_wq);
+
+	return 0;
+}
+
+#if IS_BUILTIN(CONFIG_OF)
+static const struct of_device_id mmc_omap_match[] = {
+	{ .compatible = "ti,omap2420-mmc", },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, mmc_omap_match);
+#endif
+
+static struct platform_driver mmc_omap_driver = {
+	.probe		= mmc_omap_probe,
+	.remove		= mmc_omap_remove,
+	.driver		= {
+		.name	= DRIVER_NAME,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(mmc_omap_match),
+	},
+};
+
+module_platform_driver(mmc_omap_driver);
+MODULE_DESCRIPTION("OMAP Multimedia Card driver");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:" DRIVER_NAME);
+MODULE_AUTHOR("Juha Yrjölä");
diff --git a/drivers/mmc/host/omap_hsmmc.c b/drivers/mmc/host/omap_hsmmc.c
new file mode 100644
index 000000000..098075449
--- /dev/null
+++ b/drivers/mmc/host/omap_hsmmc.c
@@ -0,0 +1,2182 @@
+/*
+ * drivers/mmc/host/omap_hsmmc.c
+ *
+ * Driver for OMAP2430/3430 MMC controller.
+ *
+ * Copyright (C) 2007 Texas Instruments.
+ *
+ * Authors:
+ *	Syed Mohammed Khasim	<x0khasim@ti.com>
+ *	Madhusudhan		<madhu.cr@ti.com>
+ *	Mohit Jalori		<mjalori@ti.com>
+ *
+ * This file is licensed under the terms of the GNU General Public License
+ * version 2. This program is licensed "as is" without any warranty of any
+ * kind, whether express or implied.
+ */
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/debugfs.h>
+#include <linux/dmaengine.h>
+#include <linux/seq_file.h>
+#include <linux/sizes.h>
+#include <linux/interrupt.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/platform_device.h>
+#include <linux/timer.h>
+#include <linux/clk.h>
+#include <linux/of.h>
+#include <linux/of_irq.h>
+#include <linux/of_device.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/core.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/io.h>
+#include <linux/irq.h>
+#include <linux/regulator/consumer.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/pm_runtime.h>
+#include <linux/pm_wakeirq.h>
+#include <linux/platform_data/hsmmc-omap.h>
+
+/* OMAP HSMMC Host Controller Registers */
+#define OMAP_HSMMC_SYSSTATUS	0x0014
+#define OMAP_HSMMC_CON		0x002C
+#define OMAP_HSMMC_SDMASA	0x0100
+#define OMAP_HSMMC_BLK		0x0104
+#define OMAP_HSMMC_ARG		0x0108
+#define OMAP_HSMMC_CMD		0x010C
+#define OMAP_HSMMC_RSP10	0x0110
+#define OMAP_HSMMC_RSP32	0x0114
+#define OMAP_HSMMC_RSP54	0x0118
+#define OMAP_HSMMC_RSP76	0x011C
+#define OMAP_HSMMC_DATA		0x0120
+#define OMAP_HSMMC_PSTATE	0x0124
+#define OMAP_HSMMC_HCTL		0x0128
+#define OMAP_HSMMC_SYSCTL	0x012C
+#define OMAP_HSMMC_STAT		0x0130
+#define OMAP_HSMMC_IE		0x0134
+#define OMAP_HSMMC_ISE		0x0138
+#define OMAP_HSMMC_AC12		0x013C
+#define OMAP_HSMMC_CAPA		0x0140
+
+#define VS18			(1 << 26)
+#define VS30			(1 << 25)
+#define HSS			(1 << 21)
+#define SDVS18			(0x5 << 9)
+#define SDVS30			(0x6 << 9)
+#define SDVS33			(0x7 << 9)
+#define SDVS_MASK		0x00000E00
+#define SDVSCLR			0xFFFFF1FF
+#define SDVSDET			0x00000400
+#define AUTOIDLE		0x1
+#define SDBP			(1 << 8)
+#define DTO			0xe
+#define ICE			0x1
+#define ICS			0x2
+#define CEN			(1 << 2)
+#define CLKD_MAX		0x3FF		/* max clock divisor: 1023 */
+#define CLKD_MASK		0x0000FFC0
+#define CLKD_SHIFT		6
+#define DTO_MASK		0x000F0000
+#define DTO_SHIFT		16
+#define INIT_STREAM		(1 << 1)
+#define ACEN_ACMD23		(2 << 2)
+#define DP_SELECT		(1 << 21)
+#define DDIR			(1 << 4)
+#define DMAE			0x1
+#define MSBS			(1 << 5)
+#define BCE			(1 << 1)
+#define FOUR_BIT		(1 << 1)
+#define HSPE			(1 << 2)
+#define IWE			(1 << 24)
+#define DDR			(1 << 19)
+#define CLKEXTFREE		(1 << 16)
+#define CTPL			(1 << 11)
+#define DW8			(1 << 5)
+#define OD			0x1
+#define STAT_CLEAR		0xFFFFFFFF
+#define INIT_STREAM_CMD		0x00000000
+#define DUAL_VOLT_OCR_BIT	7
+#define SRC			(1 << 25)
+#define SRD			(1 << 26)
+#define SOFTRESET		(1 << 1)
+
+/* PSTATE */
+#define DLEV_DAT(x)		(1 << (20 + (x)))
+
+/* Interrupt masks for IE and ISE register */
+#define CC_EN			(1 << 0)
+#define TC_EN			(1 << 1)
+#define BWR_EN			(1 << 4)
+#define BRR_EN			(1 << 5)
+#define CIRQ_EN			(1 << 8)
+#define ERR_EN			(1 << 15)
+#define CTO_EN			(1 << 16)
+#define CCRC_EN			(1 << 17)
+#define CEB_EN			(1 << 18)
+#define CIE_EN			(1 << 19)
+#define DTO_EN			(1 << 20)
+#define DCRC_EN			(1 << 21)
+#define DEB_EN			(1 << 22)
+#define ACE_EN			(1 << 24)
+#define CERR_EN			(1 << 28)
+#define BADA_EN			(1 << 29)
+
+#define INT_EN_MASK (BADA_EN | CERR_EN | ACE_EN | DEB_EN | DCRC_EN |\
+		DTO_EN | CIE_EN | CEB_EN | CCRC_EN | CTO_EN | \
+		BRR_EN | BWR_EN | TC_EN | CC_EN)
+
+#define CNI	(1 << 7)
+#define ACIE	(1 << 4)
+#define ACEB	(1 << 3)
+#define ACCE	(1 << 2)
+#define ACTO	(1 << 1)
+#define ACNE	(1 << 0)
+
+#define MMC_AUTOSUSPEND_DELAY	100
+#define MMC_TIMEOUT_MS		20		/* 20 mSec */
+#define MMC_TIMEOUT_US		20000		/* 20000 micro Sec */
+#define OMAP_MMC_MIN_CLOCK	400000
+#define OMAP_MMC_MAX_CLOCK	52000000
+#define DRIVER_NAME		"omap_hsmmc"
+
+/*
+ * One controller can have multiple slots, like on some omap boards using
+ * omap.c controller driver. Luckily this is not currently done on any known
+ * omap_hsmmc.c device.
+ */
+#define mmc_pdata(host)		host->pdata
+
+/*
+ * MMC Host controller read/write API's
+ */
+#define OMAP_HSMMC_READ(base, reg)	\
+	__raw_readl((base) + OMAP_HSMMC_##reg)
+
+#define OMAP_HSMMC_WRITE(base, reg, val) \
+	__raw_writel((val), (base) + OMAP_HSMMC_##reg)
+
+struct omap_hsmmc_next {
+	unsigned int	dma_len;
+	s32		cookie;
+};
+
+struct omap_hsmmc_host {
+	struct	device		*dev;
+	struct	mmc_host	*mmc;
+	struct	mmc_request	*mrq;
+	struct	mmc_command	*cmd;
+	struct	mmc_data	*data;
+	struct	clk		*fclk;
+	struct	clk		*dbclk;
+	struct	regulator	*pbias;
+	bool			pbias_enabled;
+	void	__iomem		*base;
+	int			vqmmc_enabled;
+	resource_size_t		mapbase;
+	spinlock_t		irq_lock; /* Prevent races with irq handler */
+	unsigned int		dma_len;
+	unsigned int		dma_sg_idx;
+	unsigned char		bus_mode;
+	unsigned char		power_mode;
+	int			suspended;
+	u32			con;
+	u32			hctl;
+	u32			sysctl;
+	u32			capa;
+	int			irq;
+	int			wake_irq;
+	int			use_dma, dma_ch;
+	struct dma_chan		*tx_chan;
+	struct dma_chan		*rx_chan;
+	int			response_busy;
+	int			context_loss;
+	int			reqs_blocked;
+	int			req_in_progress;
+	unsigned long		clk_rate;
+	unsigned int		flags;
+#define AUTO_CMD23		(1 << 0)        /* Auto CMD23 support */
+#define HSMMC_SDIO_IRQ_ENABLED	(1 << 1)        /* SDIO irq enabled */
+	struct omap_hsmmc_next	next_data;
+	struct	omap_hsmmc_platform_data	*pdata;
+};
+
+struct omap_mmc_of_data {
+	u32 reg_offset;
+	u8 controller_flags;
+};
+
+static void omap_hsmmc_start_dma_transfer(struct omap_hsmmc_host *host);
+
+static int omap_hsmmc_enable_supply(struct mmc_host *mmc)
+{
+	int ret;
+	struct omap_hsmmc_host *host = mmc_priv(mmc);
+	struct mmc_ios *ios = &mmc->ios;
+
+	if (!IS_ERR(mmc->supply.vmmc)) {
+		ret = mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, ios->vdd);
+		if (ret)
+			return ret;
+	}
+
+	/* Enable interface voltage rail, if needed */
+	if (!IS_ERR(mmc->supply.vqmmc) && !host->vqmmc_enabled) {
+		ret = regulator_enable(mmc->supply.vqmmc);
+		if (ret) {
+			dev_err(mmc_dev(mmc), "vmmc_aux reg enable failed\n");
+			goto err_vqmmc;
+		}
+		host->vqmmc_enabled = 1;
+	}
+
+	return 0;
+
+err_vqmmc:
+	if (!IS_ERR(mmc->supply.vmmc))
+		mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+
+	return ret;
+}
+
+static int omap_hsmmc_disable_supply(struct mmc_host *mmc)
+{
+	int ret;
+	int status;
+	struct omap_hsmmc_host *host = mmc_priv(mmc);
+
+	if (!IS_ERR(mmc->supply.vqmmc) && host->vqmmc_enabled) {
+		ret = regulator_disable(mmc->supply.vqmmc);
+		if (ret) {
+			dev_err(mmc_dev(mmc), "vmmc_aux reg disable failed\n");
+			return ret;
+		}
+		host->vqmmc_enabled = 0;
+	}
+
+	if (!IS_ERR(mmc->supply.vmmc)) {
+		ret = mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+		if (ret)
+			goto err_set_ocr;
+	}
+
+	return 0;
+
+err_set_ocr:
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+		status = regulator_enable(mmc->supply.vqmmc);
+		if (status)
+			dev_err(mmc_dev(mmc), "vmmc_aux re-enable failed\n");
+	}
+
+	return ret;
+}
+
+static int omap_hsmmc_set_pbias(struct omap_hsmmc_host *host, bool power_on)
+{
+	int ret;
+
+	if (IS_ERR(host->pbias))
+		return 0;
+
+	if (power_on) {
+		if (host->pbias_enabled == 0) {
+			ret = regulator_enable(host->pbias);
+			if (ret) {
+				dev_err(host->dev, "pbias reg enable fail\n");
+				return ret;
+			}
+			host->pbias_enabled = 1;
+		}
+	} else {
+		if (host->pbias_enabled == 1) {
+			ret = regulator_disable(host->pbias);
+			if (ret) {
+				dev_err(host->dev, "pbias reg disable fail\n");
+				return ret;
+			}
+			host->pbias_enabled = 0;
+		}
+	}
+
+	return 0;
+}
+
+static int omap_hsmmc_set_power(struct omap_hsmmc_host *host, int power_on)
+{
+	struct mmc_host *mmc = host->mmc;
+	int ret = 0;
+
+	/*
+	 * If we don't see a Vcc regulator, assume it's a fixed
+	 * voltage always-on regulator.
+	 */
+	if (IS_ERR(mmc->supply.vmmc))
+		return 0;
+
+	ret = omap_hsmmc_set_pbias(host, false);
+	if (ret)
+		return ret;
+
+	/*
+	 * Assume Vcc regulator is used only to power the card ... OMAP
+	 * VDDS is used to power the pins, optionally with a transceiver to
+	 * support cards using voltages other than VDDS (1.8V nominal).  When a
+	 * transceiver is used, DAT3..7 are muxed as transceiver control pins.
+	 *
+	 * In some cases this regulator won't support enable/disable;
+	 * e.g. it's a fixed rail for a WLAN chip.
+	 *
+	 * In other cases vcc_aux switches interface power.  Example, for
+	 * eMMC cards it represents VccQ.  Sometimes transceivers or SDIO
+	 * chips/cards need an interface voltage rail too.
+	 */
+	if (power_on) {
+		ret = omap_hsmmc_enable_supply(mmc);
+		if (ret)
+			return ret;
+
+		ret = omap_hsmmc_set_pbias(host, true);
+		if (ret)
+			goto err_set_voltage;
+	} else {
+		ret = omap_hsmmc_disable_supply(mmc);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+
+err_set_voltage:
+	omap_hsmmc_disable_supply(mmc);
+
+	return ret;
+}
+
+static int omap_hsmmc_disable_boot_regulator(struct regulator *reg)
+{
+	int ret;
+
+	if (IS_ERR(reg))
+		return 0;
+
+	if (regulator_is_enabled(reg)) {
+		ret = regulator_enable(reg);
+		if (ret)
+			return ret;
+
+		ret = regulator_disable(reg);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int omap_hsmmc_disable_boot_regulators(struct omap_hsmmc_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	int ret;
+
+	/*
+	 * disable regulators enabled during boot and get the usecount
+	 * right so that regulators can be enabled/disabled by checking
+	 * the return value of regulator_is_enabled
+	 */
+	ret = omap_hsmmc_disable_boot_regulator(mmc->supply.vmmc);
+	if (ret) {
+		dev_err(host->dev, "fail to disable boot enabled vmmc reg\n");
+		return ret;
+	}
+
+	ret = omap_hsmmc_disable_boot_regulator(mmc->supply.vqmmc);
+	if (ret) {
+		dev_err(host->dev,
+			"fail to disable boot enabled vmmc_aux reg\n");
+		return ret;
+	}
+
+	ret = omap_hsmmc_disable_boot_regulator(host->pbias);
+	if (ret) {
+		dev_err(host->dev,
+			"failed to disable boot enabled pbias reg\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+static int omap_hsmmc_reg_get(struct omap_hsmmc_host *host)
+{
+	int ret;
+	struct mmc_host *mmc = host->mmc;
+
+
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret)
+		return ret;
+
+	/* Allow an aux regulator */
+	if (IS_ERR(mmc->supply.vqmmc)) {
+		mmc->supply.vqmmc = devm_regulator_get_optional(host->dev,
+								"vmmc_aux");
+		if (IS_ERR(mmc->supply.vqmmc)) {
+			ret = PTR_ERR(mmc->supply.vqmmc);
+			if ((ret != -ENODEV) && host->dev->of_node)
+				return ret;
+			dev_dbg(host->dev, "unable to get vmmc_aux regulator %ld\n",
+				PTR_ERR(mmc->supply.vqmmc));
+		}
+	}
+
+	host->pbias = devm_regulator_get_optional(host->dev, "pbias");
+	if (IS_ERR(host->pbias)) {
+		ret = PTR_ERR(host->pbias);
+		if ((ret != -ENODEV) && host->dev->of_node) {
+			dev_err(host->dev,
+			"SD card detect fail? enable CONFIG_REGULATOR_PBIAS\n");
+			return ret;
+		}
+		dev_dbg(host->dev, "unable to get pbias regulator %ld\n",
+			PTR_ERR(host->pbias));
+	}
+
+	/* For eMMC do not power off when not in sleep state */
+	if (mmc_pdata(host)->no_regulator_off_init)
+		return 0;
+
+	ret = omap_hsmmc_disable_boot_regulators(host);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+/*
+ * Start clock to the card
+ */
+static void omap_hsmmc_start_clock(struct omap_hsmmc_host *host)
+{
+	OMAP_HSMMC_WRITE(host->base, SYSCTL,
+		OMAP_HSMMC_READ(host->base, SYSCTL) | CEN);
+}
+
+/*
+ * Stop clock to the card
+ */
+static void omap_hsmmc_stop_clock(struct omap_hsmmc_host *host)
+{
+	OMAP_HSMMC_WRITE(host->base, SYSCTL,
+		OMAP_HSMMC_READ(host->base, SYSCTL) & ~CEN);
+	if ((OMAP_HSMMC_READ(host->base, SYSCTL) & CEN) != 0x0)
+		dev_dbg(mmc_dev(host->mmc), "MMC Clock is not stopped\n");
+}
+
+static void omap_hsmmc_enable_irq(struct omap_hsmmc_host *host,
+				  struct mmc_command *cmd)
+{
+	u32 irq_mask = INT_EN_MASK;
+	unsigned long flags;
+
+	if (host->use_dma)
+		irq_mask &= ~(BRR_EN | BWR_EN);
+
+	/* Disable timeout for erases */
+	if (cmd->opcode == MMC_ERASE)
+		irq_mask &= ~DTO_EN;
+
+	spin_lock_irqsave(&host->irq_lock, flags);
+	OMAP_HSMMC_WRITE(host->base, STAT, STAT_CLEAR);
+	OMAP_HSMMC_WRITE(host->base, ISE, irq_mask);
+
+	/* latch pending CIRQ, but don't signal MMC core */
+	if (host->flags & HSMMC_SDIO_IRQ_ENABLED)
+		irq_mask |= CIRQ_EN;
+	OMAP_HSMMC_WRITE(host->base, IE, irq_mask);
+	spin_unlock_irqrestore(&host->irq_lock, flags);
+}
+
+static void omap_hsmmc_disable_irq(struct omap_hsmmc_host *host)
+{
+	u32 irq_mask = 0;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->irq_lock, flags);
+	/* no transfer running but need to keep cirq if enabled */
+	if (host->flags & HSMMC_SDIO_IRQ_ENABLED)
+		irq_mask |= CIRQ_EN;
+	OMAP_HSMMC_WRITE(host->base, ISE, irq_mask);
+	OMAP_HSMMC_WRITE(host->base, IE, irq_mask);
+	OMAP_HSMMC_WRITE(host->base, STAT, STAT_CLEAR);
+	spin_unlock_irqrestore(&host->irq_lock, flags);
+}
+
+/* Calculate divisor for the given clock frequency */
+static u16 calc_divisor(struct omap_hsmmc_host *host, struct mmc_ios *ios)
+{
+	u16 dsor = 0;
+
+	if (ios->clock) {
+		dsor = DIV_ROUND_UP(clk_get_rate(host->fclk), ios->clock);
+		if (dsor > CLKD_MAX)
+			dsor = CLKD_MAX;
+	}
+
+	return dsor;
+}
+
+static void omap_hsmmc_set_clock(struct omap_hsmmc_host *host)
+{
+	struct mmc_ios *ios = &host->mmc->ios;
+	unsigned long regval;
+	unsigned long timeout;
+	unsigned long clkdiv;
+
+	dev_vdbg(mmc_dev(host->mmc), "Set clock to %uHz\n", ios->clock);
+
+	omap_hsmmc_stop_clock(host);
+
+	regval = OMAP_HSMMC_READ(host->base, SYSCTL);
+	regval = regval & ~(CLKD_MASK | DTO_MASK);
+	clkdiv = calc_divisor(host, ios);
+	regval = regval | (clkdiv << 6) | (DTO << 16);
+	OMAP_HSMMC_WRITE(host->base, SYSCTL, regval);
+	OMAP_HSMMC_WRITE(host->base, SYSCTL,
+		OMAP_HSMMC_READ(host->base, SYSCTL) | ICE);
+
+	/* Wait till the ICS bit is set */
+	timeout = jiffies + msecs_to_jiffies(MMC_TIMEOUT_MS);
+	while ((OMAP_HSMMC_READ(host->base, SYSCTL) & ICS) != ICS
+		&& time_before(jiffies, timeout))
+		cpu_relax();
+
+	/*
+	 * Enable High-Speed Support
+	 * Pre-Requisites
+	 *	- Controller should support High-Speed-Enable Bit
+	 *	- Controller should not be using DDR Mode
+	 *	- Controller should advertise that it supports High Speed
+	 *	  in capabilities register
+	 *	- MMC/SD clock coming out of controller > 25MHz
+	 */
+	if ((mmc_pdata(host)->features & HSMMC_HAS_HSPE_SUPPORT) &&
+	    (ios->timing != MMC_TIMING_MMC_DDR52) &&
+	    (ios->timing != MMC_TIMING_UHS_DDR50) &&
+	    ((OMAP_HSMMC_READ(host->base, CAPA) & HSS) == HSS)) {
+		regval = OMAP_HSMMC_READ(host->base, HCTL);
+		if (clkdiv && (clk_get_rate(host->fclk)/clkdiv) > 25000000)
+			regval |= HSPE;
+		else
+			regval &= ~HSPE;
+
+		OMAP_HSMMC_WRITE(host->base, HCTL, regval);
+	}
+
+	omap_hsmmc_start_clock(host);
+}
+
+static void omap_hsmmc_set_bus_width(struct omap_hsmmc_host *host)
+{
+	struct mmc_ios *ios = &host->mmc->ios;
+	u32 con;
+
+	con = OMAP_HSMMC_READ(host->base, CON);
+	if (ios->timing == MMC_TIMING_MMC_DDR52 ||
+	    ios->timing == MMC_TIMING_UHS_DDR50)
+		con |= DDR;	/* configure in DDR mode */
+	else
+		con &= ~DDR;
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_8:
+		OMAP_HSMMC_WRITE(host->base, CON, con | DW8);
+		break;
+	case MMC_BUS_WIDTH_4:
+		OMAP_HSMMC_WRITE(host->base, CON, con & ~DW8);
+		OMAP_HSMMC_WRITE(host->base, HCTL,
+			OMAP_HSMMC_READ(host->base, HCTL) | FOUR_BIT);
+		break;
+	case MMC_BUS_WIDTH_1:
+		OMAP_HSMMC_WRITE(host->base, CON, con & ~DW8);
+		OMAP_HSMMC_WRITE(host->base, HCTL,
+			OMAP_HSMMC_READ(host->base, HCTL) & ~FOUR_BIT);
+		break;
+	}
+}
+
+static void omap_hsmmc_set_bus_mode(struct omap_hsmmc_host *host)
+{
+	struct mmc_ios *ios = &host->mmc->ios;
+	u32 con;
+
+	con = OMAP_HSMMC_READ(host->base, CON);
+	if (ios->bus_mode == MMC_BUSMODE_OPENDRAIN)
+		OMAP_HSMMC_WRITE(host->base, CON, con | OD);
+	else
+		OMAP_HSMMC_WRITE(host->base, CON, con & ~OD);
+}
+
+#ifdef CONFIG_PM
+
+/*
+ * Restore the MMC host context, if it was lost as result of a
+ * power state change.
+ */
+static int omap_hsmmc_context_restore(struct omap_hsmmc_host *host)
+{
+	struct mmc_ios *ios = &host->mmc->ios;
+	u32 hctl, capa;
+	unsigned long timeout;
+
+	if (host->con == OMAP_HSMMC_READ(host->base, CON) &&
+	    host->hctl == OMAP_HSMMC_READ(host->base, HCTL) &&
+	    host->sysctl == OMAP_HSMMC_READ(host->base, SYSCTL) &&
+	    host->capa == OMAP_HSMMC_READ(host->base, CAPA))
+		return 0;
+
+	host->context_loss++;
+
+	if (host->pdata->controller_flags & OMAP_HSMMC_SUPPORTS_DUAL_VOLT) {
+		if (host->power_mode != MMC_POWER_OFF &&
+		    (1 << ios->vdd) <= MMC_VDD_23_24)
+			hctl = SDVS18;
+		else
+			hctl = SDVS30;
+		capa = VS30 | VS18;
+	} else {
+		hctl = SDVS18;
+		capa = VS18;
+	}
+
+	if (host->mmc->caps & MMC_CAP_SDIO_IRQ)
+		hctl |= IWE;
+
+	OMAP_HSMMC_WRITE(host->base, HCTL,
+			OMAP_HSMMC_READ(host->base, HCTL) | hctl);
+
+	OMAP_HSMMC_WRITE(host->base, CAPA,
+			OMAP_HSMMC_READ(host->base, CAPA) | capa);
+
+	OMAP_HSMMC_WRITE(host->base, HCTL,
+			OMAP_HSMMC_READ(host->base, HCTL) | SDBP);
+
+	timeout = jiffies + msecs_to_jiffies(MMC_TIMEOUT_MS);
+	while ((OMAP_HSMMC_READ(host->base, HCTL) & SDBP) != SDBP
+		&& time_before(jiffies, timeout))
+		;
+
+	OMAP_HSMMC_WRITE(host->base, ISE, 0);
+	OMAP_HSMMC_WRITE(host->base, IE, 0);
+	OMAP_HSMMC_WRITE(host->base, STAT, STAT_CLEAR);
+
+	/* Do not initialize card-specific things if the power is off */
+	if (host->power_mode == MMC_POWER_OFF)
+		goto out;
+
+	omap_hsmmc_set_bus_width(host);
+
+	omap_hsmmc_set_clock(host);
+
+	omap_hsmmc_set_bus_mode(host);
+
+out:
+	dev_dbg(mmc_dev(host->mmc), "context is restored: restore count %d\n",
+		host->context_loss);
+	return 0;
+}
+
+/*
+ * Save the MMC host context (store the number of power state changes so far).
+ */
+static void omap_hsmmc_context_save(struct omap_hsmmc_host *host)
+{
+	host->con =  OMAP_HSMMC_READ(host->base, CON);
+	host->hctl = OMAP_HSMMC_READ(host->base, HCTL);
+	host->sysctl =  OMAP_HSMMC_READ(host->base, SYSCTL);
+	host->capa = OMAP_HSMMC_READ(host->base, CAPA);
+}
+
+#else
+
+static int omap_hsmmc_context_restore(struct omap_hsmmc_host *host)
+{
+	return 0;
+}
+
+static void omap_hsmmc_context_save(struct omap_hsmmc_host *host)
+{
+}
+
+#endif
+
+/*
+ * Send init stream sequence to card
+ * before sending IDLE command
+ */
+static void send_init_stream(struct omap_hsmmc_host *host)
+{
+	int reg = 0;
+	unsigned long timeout;
+
+	disable_irq(host->irq);
+
+	OMAP_HSMMC_WRITE(host->base, IE, INT_EN_MASK);
+	OMAP_HSMMC_WRITE(host->base, CON,
+		OMAP_HSMMC_READ(host->base, CON) | INIT_STREAM);
+	OMAP_HSMMC_WRITE(host->base, CMD, INIT_STREAM_CMD);
+
+	timeout = jiffies + msecs_to_jiffies(MMC_TIMEOUT_MS);
+	while ((reg != CC_EN) && time_before(jiffies, timeout))
+		reg = OMAP_HSMMC_READ(host->base, STAT) & CC_EN;
+
+	OMAP_HSMMC_WRITE(host->base, CON,
+		OMAP_HSMMC_READ(host->base, CON) & ~INIT_STREAM);
+
+	OMAP_HSMMC_WRITE(host->base, STAT, STAT_CLEAR);
+	OMAP_HSMMC_READ(host->base, STAT);
+
+	enable_irq(host->irq);
+}
+
+static ssize_t
+omap_hsmmc_show_slot_name(struct device *dev, struct device_attribute *attr,
+			char *buf)
+{
+	struct mmc_host *mmc = container_of(dev, struct mmc_host, class_dev);
+	struct omap_hsmmc_host *host = mmc_priv(mmc);
+
+	return sprintf(buf, "%s\n", mmc_pdata(host)->name);
+}
+
+static DEVICE_ATTR(slot_name, S_IRUGO, omap_hsmmc_show_slot_name, NULL);
+
+/*
+ * Configure the response type and send the cmd.
+ */
+static void
+omap_hsmmc_start_command(struct omap_hsmmc_host *host, struct mmc_command *cmd,
+	struct mmc_data *data)
+{
+	int cmdreg = 0, resptype = 0, cmdtype = 0;
+
+	dev_vdbg(mmc_dev(host->mmc), "%s: CMD%d, argument 0x%08x\n",
+		mmc_hostname(host->mmc), cmd->opcode, cmd->arg);
+	host->cmd = cmd;
+
+	omap_hsmmc_enable_irq(host, cmd);
+
+	host->response_busy = 0;
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136)
+			resptype = 1;
+		else if (cmd->flags & MMC_RSP_BUSY) {
+			resptype = 3;
+			host->response_busy = 1;
+		} else
+			resptype = 2;
+	}
+
+	/*
+	 * Unlike OMAP1 controller, the cmdtype does not seem to be based on
+	 * ac, bc, adtc, bcr. Only commands ending an open ended transfer need
+	 * a val of 0x3, rest 0x0.
+	 */
+	if (cmd == host->mrq->stop)
+		cmdtype = 0x3;
+
+	cmdreg = (cmd->opcode << 24) | (resptype << 16) | (cmdtype << 22);
+
+	if ((host->flags & AUTO_CMD23) && mmc_op_multi(cmd->opcode) &&
+	    host->mrq->sbc) {
+		cmdreg |= ACEN_ACMD23;
+		OMAP_HSMMC_WRITE(host->base, SDMASA, host->mrq->sbc->arg);
+	}
+	if (data) {
+		cmdreg |= DP_SELECT | MSBS | BCE;
+		if (data->flags & MMC_DATA_READ)
+			cmdreg |= DDIR;
+		else
+			cmdreg &= ~(DDIR);
+	}
+
+	if (host->use_dma)
+		cmdreg |= DMAE;
+
+	host->req_in_progress = 1;
+
+	OMAP_HSMMC_WRITE(host->base, ARG, cmd->arg);
+	OMAP_HSMMC_WRITE(host->base, CMD, cmdreg);
+}
+
+static struct dma_chan *omap_hsmmc_get_dma_chan(struct omap_hsmmc_host *host,
+	struct mmc_data *data)
+{
+	return data->flags & MMC_DATA_WRITE ? host->tx_chan : host->rx_chan;
+}
+
+static void omap_hsmmc_request_done(struct omap_hsmmc_host *host, struct mmc_request *mrq)
+{
+	int dma_ch;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->irq_lock, flags);
+	host->req_in_progress = 0;
+	dma_ch = host->dma_ch;
+	spin_unlock_irqrestore(&host->irq_lock, flags);
+
+	omap_hsmmc_disable_irq(host);
+	/* Do not complete the request if DMA is still in progress */
+	if (mrq->data && host->use_dma && dma_ch != -1)
+		return;
+	host->mrq = NULL;
+	mmc_request_done(host->mmc, mrq);
+}
+
+/*
+ * Notify the transfer complete to MMC core
+ */
+static void
+omap_hsmmc_xfer_done(struct omap_hsmmc_host *host, struct mmc_data *data)
+{
+	if (!data) {
+		struct mmc_request *mrq = host->mrq;
+
+		/* TC before CC from CMD6 - don't know why, but it happens */
+		if (host->cmd && host->cmd->opcode == 6 &&
+		    host->response_busy) {
+			host->response_busy = 0;
+			return;
+		}
+
+		omap_hsmmc_request_done(host, mrq);
+		return;
+	}
+
+	host->data = NULL;
+
+	if (!data->error)
+		data->bytes_xfered += data->blocks * (data->blksz);
+	else
+		data->bytes_xfered = 0;
+
+	if (data->stop && (data->error || !host->mrq->sbc))
+		omap_hsmmc_start_command(host, data->stop, NULL);
+	else
+		omap_hsmmc_request_done(host, data->mrq);
+}
+
+/*
+ * Notify the core about command completion
+ */
+static void
+omap_hsmmc_cmd_done(struct omap_hsmmc_host *host, struct mmc_command *cmd)
+{
+	if (host->mrq->sbc && (host->cmd == host->mrq->sbc) &&
+	    !host->mrq->sbc->error && !(host->flags & AUTO_CMD23)) {
+		host->cmd = NULL;
+		omap_hsmmc_start_dma_transfer(host);
+		omap_hsmmc_start_command(host, host->mrq->cmd,
+						host->mrq->data);
+		return;
+	}
+
+	host->cmd = NULL;
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136) {
+			/* response type 2 */
+			cmd->resp[3] = OMAP_HSMMC_READ(host->base, RSP10);
+			cmd->resp[2] = OMAP_HSMMC_READ(host->base, RSP32);
+			cmd->resp[1] = OMAP_HSMMC_READ(host->base, RSP54);
+			cmd->resp[0] = OMAP_HSMMC_READ(host->base, RSP76);
+		} else {
+			/* response types 1, 1b, 3, 4, 5, 6 */
+			cmd->resp[0] = OMAP_HSMMC_READ(host->base, RSP10);
+		}
+	}
+	if ((host->data == NULL && !host->response_busy) || cmd->error)
+		omap_hsmmc_request_done(host, host->mrq);
+}
+
+/*
+ * DMA clean up for command errors
+ */
+static void omap_hsmmc_dma_cleanup(struct omap_hsmmc_host *host, int errno)
+{
+	int dma_ch;
+	unsigned long flags;
+
+	host->data->error = errno;
+
+	spin_lock_irqsave(&host->irq_lock, flags);
+	dma_ch = host->dma_ch;
+	host->dma_ch = -1;
+	spin_unlock_irqrestore(&host->irq_lock, flags);
+
+	if (host->use_dma && dma_ch != -1) {
+		struct dma_chan *chan = omap_hsmmc_get_dma_chan(host, host->data);
+
+		dmaengine_terminate_all(chan);
+		dma_unmap_sg(chan->device->dev,
+			host->data->sg, host->data->sg_len,
+			mmc_get_dma_dir(host->data));
+
+		host->data->host_cookie = 0;
+	}
+	host->data = NULL;
+}
+
+/*
+ * Readable error output
+ */
+#ifdef CONFIG_MMC_DEBUG
+static void omap_hsmmc_dbg_report_irq(struct omap_hsmmc_host *host, u32 status)
+{
+	/* --- means reserved bit without definition at documentation */
+	static const char *omap_hsmmc_status_bits[] = {
+		"CC"  , "TC"  , "BGE", "---", "BWR" , "BRR" , "---" , "---" ,
+		"CIRQ",	"OBI" , "---", "---", "---" , "---" , "---" , "ERRI",
+		"CTO" , "CCRC", "CEB", "CIE", "DTO" , "DCRC", "DEB" , "---" ,
+		"ACE" , "---" , "---", "---", "CERR", "BADA", "---" , "---"
+	};
+	char res[256];
+	char *buf = res;
+	int len, i;
+
+	len = sprintf(buf, "MMC IRQ 0x%x :", status);
+	buf += len;
+
+	for (i = 0; i < ARRAY_SIZE(omap_hsmmc_status_bits); i++)
+		if (status & (1 << i)) {
+			len = sprintf(buf, " %s", omap_hsmmc_status_bits[i]);
+			buf += len;
+		}
+
+	dev_vdbg(mmc_dev(host->mmc), "%s\n", res);
+}
+#else
+static inline void omap_hsmmc_dbg_report_irq(struct omap_hsmmc_host *host,
+					     u32 status)
+{
+}
+#endif  /* CONFIG_MMC_DEBUG */
+
+/*
+ * MMC controller internal state machines reset
+ *
+ * Used to reset command or data internal state machines, using respectively
+ *  SRC or SRD bit of SYSCTL register
+ * Can be called from interrupt context
+ */
+static inline void omap_hsmmc_reset_controller_fsm(struct omap_hsmmc_host *host,
+						   unsigned long bit)
+{
+	unsigned long i = 0;
+	unsigned long limit = MMC_TIMEOUT_US;
+
+	OMAP_HSMMC_WRITE(host->base, SYSCTL,
+			 OMAP_HSMMC_READ(host->base, SYSCTL) | bit);
+
+	/*
+	 * OMAP4 ES2 and greater has an updated reset logic.
+	 * Monitor a 0->1 transition first
+	 */
+	if (mmc_pdata(host)->features & HSMMC_HAS_UPDATED_RESET) {
+		while ((!(OMAP_HSMMC_READ(host->base, SYSCTL) & bit))
+					&& (i++ < limit))
+			udelay(1);
+	}
+	i = 0;
+
+	while ((OMAP_HSMMC_READ(host->base, SYSCTL) & bit) &&
+		(i++ < limit))
+		udelay(1);
+
+	if (OMAP_HSMMC_READ(host->base, SYSCTL) & bit)
+		dev_err(mmc_dev(host->mmc),
+			"Timeout waiting on controller reset in %s\n",
+			__func__);
+}
+
+static void hsmmc_command_incomplete(struct omap_hsmmc_host *host,
+					int err, int end_cmd)
+{
+	if (end_cmd) {
+		omap_hsmmc_reset_controller_fsm(host, SRC);
+		if (host->cmd)
+			host->cmd->error = err;
+	}
+
+	if (host->data) {
+		omap_hsmmc_reset_controller_fsm(host, SRD);
+		omap_hsmmc_dma_cleanup(host, err);
+	} else if (host->mrq && host->mrq->cmd)
+		host->mrq->cmd->error = err;
+}
+
+static void omap_hsmmc_do_irq(struct omap_hsmmc_host *host, int status)
+{
+	struct mmc_data *data;
+	int end_cmd = 0, end_trans = 0;
+	int error = 0;
+
+	data = host->data;
+	dev_vdbg(mmc_dev(host->mmc), "IRQ Status is %x\n", status);
+
+	if (status & ERR_EN) {
+		omap_hsmmc_dbg_report_irq(host, status);
+
+		if (status & (CTO_EN | CCRC_EN | CEB_EN))
+			end_cmd = 1;
+		if (host->data || host->response_busy) {
+			end_trans = !end_cmd;
+			host->response_busy = 0;
+		}
+		if (status & (CTO_EN | DTO_EN))
+			hsmmc_command_incomplete(host, -ETIMEDOUT, end_cmd);
+		else if (status & (CCRC_EN | DCRC_EN | DEB_EN | CEB_EN |
+				   BADA_EN))
+			hsmmc_command_incomplete(host, -EILSEQ, end_cmd);
+
+		if (status & ACE_EN) {
+			u32 ac12;
+			ac12 = OMAP_HSMMC_READ(host->base, AC12);
+			if (!(ac12 & ACNE) && host->mrq->sbc) {
+				end_cmd = 1;
+				if (ac12 & ACTO)
+					error =  -ETIMEDOUT;
+				else if (ac12 & (ACCE | ACEB | ACIE))
+					error = -EILSEQ;
+				host->mrq->sbc->error = error;
+				hsmmc_command_incomplete(host, error, end_cmd);
+			}
+			dev_dbg(mmc_dev(host->mmc), "AC12 err: 0x%x\n", ac12);
+		}
+	}
+
+	OMAP_HSMMC_WRITE(host->base, STAT, status);
+	if (end_cmd || ((status & CC_EN) && host->cmd))
+		omap_hsmmc_cmd_done(host, host->cmd);
+	if ((end_trans || (status & TC_EN)) && host->mrq)
+		omap_hsmmc_xfer_done(host, data);
+}
+
+/*
+ * MMC controller IRQ handler
+ */
+static irqreturn_t omap_hsmmc_irq(int irq, void *dev_id)
+{
+	struct omap_hsmmc_host *host = dev_id;
+	int status;
+
+	status = OMAP_HSMMC_READ(host->base, STAT);
+	while (status & (INT_EN_MASK | CIRQ_EN)) {
+		if (host->req_in_progress)
+			omap_hsmmc_do_irq(host, status);
+
+		if (status & CIRQ_EN)
+			mmc_signal_sdio_irq(host->mmc);
+
+		/* Flush posted write */
+		status = OMAP_HSMMC_READ(host->base, STAT);
+	}
+
+	return IRQ_HANDLED;
+}
+
+static void set_sd_bus_power(struct omap_hsmmc_host *host)
+{
+	unsigned long i;
+
+	OMAP_HSMMC_WRITE(host->base, HCTL,
+			 OMAP_HSMMC_READ(host->base, HCTL) | SDBP);
+	for (i = 0; i < loops_per_jiffy; i++) {
+		if (OMAP_HSMMC_READ(host->base, HCTL) & SDBP)
+			break;
+		cpu_relax();
+	}
+}
+
+/*
+ * Switch MMC interface voltage ... only relevant for MMC1.
+ *
+ * MMC2 and MMC3 use fixed 1.8V levels, and maybe a transceiver.
+ * The MMC2 transceiver controls are used instead of DAT4..DAT7.
+ * Some chips, like eMMC ones, use internal transceivers.
+ */
+static int omap_hsmmc_switch_opcond(struct omap_hsmmc_host *host, int vdd)
+{
+	u32 reg_val = 0;
+	int ret;
+
+	/* Disable the clocks */
+	clk_disable_unprepare(host->dbclk);
+
+	/* Turn the power off */
+	ret = omap_hsmmc_set_power(host, 0);
+
+	/* Turn the power ON with given VDD 1.8 or 3.0v */
+	if (!ret)
+		ret = omap_hsmmc_set_power(host, 1);
+	clk_prepare_enable(host->dbclk);
+
+	if (ret != 0)
+		goto err;
+
+	OMAP_HSMMC_WRITE(host->base, HCTL,
+		OMAP_HSMMC_READ(host->base, HCTL) & SDVSCLR);
+	reg_val = OMAP_HSMMC_READ(host->base, HCTL);
+
+	/*
+	 * If a MMC dual voltage card is detected, the set_ios fn calls
+	 * this fn with VDD bit set for 1.8V. Upon card removal from the
+	 * slot, omap_hsmmc_set_ios sets the VDD back to 3V on MMC_POWER_OFF.
+	 *
+	 * Cope with a bit of slop in the range ... per data sheets:
+	 *  - "1.8V" for vdds_mmc1/vdds_mmc1a can be up to 2.45V max,
+	 *    but recommended values are 1.71V to 1.89V
+	 *  - "3.0V" for vdds_mmc1/vdds_mmc1a can be up to 3.5V max,
+	 *    but recommended values are 2.7V to 3.3V
+	 *
+	 * Board setup code shouldn't permit anything very out-of-range.
+	 * TWL4030-family VMMC1 and VSIM regulators are fine (avoiding the
+	 * middle range) but VSIM can't power DAT4..DAT7 at more than 3V.
+	 */
+	if ((1 << vdd) <= MMC_VDD_23_24)
+		reg_val |= SDVS18;
+	else
+		reg_val |= SDVS30;
+
+	OMAP_HSMMC_WRITE(host->base, HCTL, reg_val);
+	set_sd_bus_power(host);
+
+	return 0;
+err:
+	dev_err(mmc_dev(host->mmc), "Unable to switch operating voltage\n");
+	return ret;
+}
+
+static void omap_hsmmc_dma_callback(void *param)
+{
+	struct omap_hsmmc_host *host = param;
+	struct dma_chan *chan;
+	struct mmc_data *data;
+	int req_in_progress;
+
+	spin_lock_irq(&host->irq_lock);
+	if (host->dma_ch < 0) {
+		spin_unlock_irq(&host->irq_lock);
+		return;
+	}
+
+	data = host->mrq->data;
+	chan = omap_hsmmc_get_dma_chan(host, data);
+	if (!data->host_cookie)
+		dma_unmap_sg(chan->device->dev,
+			     data->sg, data->sg_len,
+			     mmc_get_dma_dir(data));
+
+	req_in_progress = host->req_in_progress;
+	host->dma_ch = -1;
+	spin_unlock_irq(&host->irq_lock);
+
+	/* If DMA has finished after TC, complete the request */
+	if (!req_in_progress) {
+		struct mmc_request *mrq = host->mrq;
+
+		host->mrq = NULL;
+		mmc_request_done(host->mmc, mrq);
+	}
+}
+
+static int omap_hsmmc_pre_dma_transfer(struct omap_hsmmc_host *host,
+				       struct mmc_data *data,
+				       struct omap_hsmmc_next *next,
+				       struct dma_chan *chan)
+{
+	int dma_len;
+
+	if (!next && data->host_cookie &&
+	    data->host_cookie != host->next_data.cookie) {
+		dev_warn(host->dev, "[%s] invalid cookie: data->host_cookie %d"
+		       " host->next_data.cookie %d\n",
+		       __func__, data->host_cookie, host->next_data.cookie);
+		data->host_cookie = 0;
+	}
+
+	/* Check if next job is already prepared */
+	if (next || data->host_cookie != host->next_data.cookie) {
+		dma_len = dma_map_sg(chan->device->dev, data->sg, data->sg_len,
+				     mmc_get_dma_dir(data));
+
+	} else {
+		dma_len = host->next_data.dma_len;
+		host->next_data.dma_len = 0;
+	}
+
+
+	if (dma_len == 0)
+		return -EINVAL;
+
+	if (next) {
+		next->dma_len = dma_len;
+		data->host_cookie = ++next->cookie < 0 ? 1 : next->cookie;
+	} else
+		host->dma_len = dma_len;
+
+	return 0;
+}
+
+/*
+ * Routine to configure and start DMA for the MMC card
+ */
+static int omap_hsmmc_setup_dma_transfer(struct omap_hsmmc_host *host,
+					struct mmc_request *req)
+{
+	struct dma_async_tx_descriptor *tx;
+	int ret = 0, i;
+	struct mmc_data *data = req->data;
+	struct dma_chan *chan;
+	struct dma_slave_config cfg = {
+		.src_addr = host->mapbase + OMAP_HSMMC_DATA,
+		.dst_addr = host->mapbase + OMAP_HSMMC_DATA,
+		.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+		.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+		.src_maxburst = data->blksz / 4,
+		.dst_maxburst = data->blksz / 4,
+	};
+
+	/* Sanity check: all the SG entries must be aligned by block size. */
+	for (i = 0; i < data->sg_len; i++) {
+		struct scatterlist *sgl;
+
+		sgl = data->sg + i;
+		if (sgl->length % data->blksz)
+			return -EINVAL;
+	}
+	if ((data->blksz % 4) != 0)
+		/* REVISIT: The MMC buffer increments only when MSB is written.
+		 * Return error for blksz which is non multiple of four.
+		 */
+		return -EINVAL;
+
+	BUG_ON(host->dma_ch != -1);
+
+	chan = omap_hsmmc_get_dma_chan(host, data);
+
+	ret = dmaengine_slave_config(chan, &cfg);
+	if (ret)
+		return ret;
+
+	ret = omap_hsmmc_pre_dma_transfer(host, data, NULL, chan);
+	if (ret)
+		return ret;
+
+	tx = dmaengine_prep_slave_sg(chan, data->sg, data->sg_len,
+		data->flags & MMC_DATA_WRITE ? DMA_MEM_TO_DEV : DMA_DEV_TO_MEM,
+		DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!tx) {
+		dev_err(mmc_dev(host->mmc), "prep_slave_sg() failed\n");
+		/* FIXME: cleanup */
+		return -1;
+	}
+
+	tx->callback = omap_hsmmc_dma_callback;
+	tx->callback_param = host;
+
+	/* Does not fail */
+	dmaengine_submit(tx);
+
+	host->dma_ch = 1;
+
+	return 0;
+}
+
+static void set_data_timeout(struct omap_hsmmc_host *host,
+			     unsigned long long timeout_ns,
+			     unsigned int timeout_clks)
+{
+	unsigned long long timeout = timeout_ns;
+	unsigned int cycle_ns;
+	uint32_t reg, clkd, dto = 0;
+
+	reg = OMAP_HSMMC_READ(host->base, SYSCTL);
+	clkd = (reg & CLKD_MASK) >> CLKD_SHIFT;
+	if (clkd == 0)
+		clkd = 1;
+
+	cycle_ns = 1000000000 / (host->clk_rate / clkd);
+	do_div(timeout, cycle_ns);
+	timeout += timeout_clks;
+	if (timeout) {
+		while ((timeout & 0x80000000) == 0) {
+			dto += 1;
+			timeout <<= 1;
+		}
+		dto = 31 - dto;
+		timeout <<= 1;
+		if (timeout && dto)
+			dto += 1;
+		if (dto >= 13)
+			dto -= 13;
+		else
+			dto = 0;
+		if (dto > 14)
+			dto = 14;
+	}
+
+	reg &= ~DTO_MASK;
+	reg |= dto << DTO_SHIFT;
+	OMAP_HSMMC_WRITE(host->base, SYSCTL, reg);
+}
+
+static void omap_hsmmc_start_dma_transfer(struct omap_hsmmc_host *host)
+{
+	struct mmc_request *req = host->mrq;
+	struct dma_chan *chan;
+
+	if (!req->data)
+		return;
+	OMAP_HSMMC_WRITE(host->base, BLK, (req->data->blksz)
+				| (req->data->blocks << 16));
+	set_data_timeout(host, req->data->timeout_ns,
+				req->data->timeout_clks);
+	chan = omap_hsmmc_get_dma_chan(host, req->data);
+	dma_async_issue_pending(chan);
+}
+
+/*
+ * Configure block length for MMC/SD cards and initiate the transfer.
+ */
+static int
+omap_hsmmc_prepare_data(struct omap_hsmmc_host *host, struct mmc_request *req)
+{
+	int ret;
+	unsigned long long timeout;
+
+	host->data = req->data;
+
+	if (req->data == NULL) {
+		OMAP_HSMMC_WRITE(host->base, BLK, 0);
+		if (req->cmd->flags & MMC_RSP_BUSY) {
+			timeout = req->cmd->busy_timeout * NSEC_PER_MSEC;
+
+			/*
+			 * Set an arbitrary 100ms data timeout for commands with
+			 * busy signal and no indication of busy_timeout.
+			 */
+			if (!timeout)
+				timeout = 100000000U;
+
+			set_data_timeout(host, timeout, 0);
+		}
+		return 0;
+	}
+
+	if (host->use_dma) {
+		ret = omap_hsmmc_setup_dma_transfer(host, req);
+		if (ret != 0) {
+			dev_err(mmc_dev(host->mmc), "MMC start dma failure\n");
+			return ret;
+		}
+	}
+	return 0;
+}
+
+static void omap_hsmmc_post_req(struct mmc_host *mmc, struct mmc_request *mrq,
+				int err)
+{
+	struct omap_hsmmc_host *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+
+	if (host->use_dma && data->host_cookie) {
+		struct dma_chan *c = omap_hsmmc_get_dma_chan(host, data);
+
+		dma_unmap_sg(c->device->dev, data->sg, data->sg_len,
+			     mmc_get_dma_dir(data));
+		data->host_cookie = 0;
+	}
+}
+
+static void omap_hsmmc_pre_req(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct omap_hsmmc_host *host = mmc_priv(mmc);
+
+	if (mrq->data->host_cookie) {
+		mrq->data->host_cookie = 0;
+		return ;
+	}
+
+	if (host->use_dma) {
+		struct dma_chan *c = omap_hsmmc_get_dma_chan(host, mrq->data);
+
+		if (omap_hsmmc_pre_dma_transfer(host, mrq->data,
+						&host->next_data, c))
+			mrq->data->host_cookie = 0;
+	}
+}
+
+/*
+ * Request function. for read/write operation
+ */
+static void omap_hsmmc_request(struct mmc_host *mmc, struct mmc_request *req)
+{
+	struct omap_hsmmc_host *host = mmc_priv(mmc);
+	int err;
+
+	BUG_ON(host->req_in_progress);
+	BUG_ON(host->dma_ch != -1);
+	if (host->reqs_blocked)
+		host->reqs_blocked = 0;
+	WARN_ON(host->mrq != NULL);
+	host->mrq = req;
+	host->clk_rate = clk_get_rate(host->fclk);
+	err = omap_hsmmc_prepare_data(host, req);
+	if (err) {
+		req->cmd->error = err;
+		if (req->data)
+			req->data->error = err;
+		host->mrq = NULL;
+		mmc_request_done(mmc, req);
+		return;
+	}
+	if (req->sbc && !(host->flags & AUTO_CMD23)) {
+		omap_hsmmc_start_command(host, req->sbc, NULL);
+		return;
+	}
+
+	omap_hsmmc_start_dma_transfer(host);
+	omap_hsmmc_start_command(host, req->cmd, req->data);
+}
+
+/* Routine to configure clock values. Exposed API to core */
+static void omap_hsmmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct omap_hsmmc_host *host = mmc_priv(mmc);
+	int do_send_init_stream = 0;
+
+	if (ios->power_mode != host->power_mode) {
+		switch (ios->power_mode) {
+		case MMC_POWER_OFF:
+			omap_hsmmc_set_power(host, 0);
+			break;
+		case MMC_POWER_UP:
+			omap_hsmmc_set_power(host, 1);
+			break;
+		case MMC_POWER_ON:
+			do_send_init_stream = 1;
+			break;
+		}
+		host->power_mode = ios->power_mode;
+	}
+
+	/* FIXME: set registers based only on changes to ios */
+
+	omap_hsmmc_set_bus_width(host);
+
+	if (host->pdata->controller_flags & OMAP_HSMMC_SUPPORTS_DUAL_VOLT) {
+		/* Only MMC1 can interface at 3V without some flavor
+		 * of external transceiver; but they all handle 1.8V.
+		 */
+		if ((OMAP_HSMMC_READ(host->base, HCTL) & SDVSDET) &&
+			(ios->vdd == DUAL_VOLT_OCR_BIT)) {
+				/*
+				 * The mmc_select_voltage fn of the core does
+				 * not seem to set the power_mode to
+				 * MMC_POWER_UP upon recalculating the voltage.
+				 * vdd 1.8v.
+				 */
+			if (omap_hsmmc_switch_opcond(host, ios->vdd) != 0)
+				dev_dbg(mmc_dev(host->mmc),
+						"Switch operation failed\n");
+		}
+	}
+
+	omap_hsmmc_set_clock(host);
+
+	if (do_send_init_stream)
+		send_init_stream(host);
+
+	omap_hsmmc_set_bus_mode(host);
+}
+
+static void omap_hsmmc_init_card(struct mmc_host *mmc, struct mmc_card *card)
+{
+	struct omap_hsmmc_host *host = mmc_priv(mmc);
+
+	if (card->type == MMC_TYPE_SDIO || card->type == MMC_TYPE_SD_COMBO) {
+		struct device_node *np = mmc_dev(mmc)->of_node;
+
+		/*
+		 * REVISIT: should be moved to sdio core and made more
+		 * general e.g. by expanding the DT bindings of child nodes
+		 * to provide a mechanism to provide this information:
+		 * Documentation/devicetree/bindings/mmc/mmc-card.txt
+		 */
+
+		np = of_get_compatible_child(np, "ti,wl1251");
+		if (np) {
+			/*
+			 * We have TI wl1251 attached to MMC3. Pass this
+			 * information to the SDIO core because it can't be
+			 * probed by normal methods.
+			 */
+
+			dev_info(host->dev, "found wl1251\n");
+			card->quirks |= MMC_QUIRK_NONSTD_SDIO;
+			card->cccr.wide_bus = 1;
+			card->cis.vendor = 0x104c;
+			card->cis.device = 0x9066;
+			card->cis.blksize = 512;
+			card->cis.max_dtr = 24000000;
+			card->ocr = 0x80;
+			of_node_put(np);
+		}
+	}
+}
+
+static void omap_hsmmc_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct omap_hsmmc_host *host = mmc_priv(mmc);
+	u32 irq_mask, con;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->irq_lock, flags);
+
+	con = OMAP_HSMMC_READ(host->base, CON);
+	irq_mask = OMAP_HSMMC_READ(host->base, ISE);
+	if (enable) {
+		host->flags |= HSMMC_SDIO_IRQ_ENABLED;
+		irq_mask |= CIRQ_EN;
+		con |= CTPL | CLKEXTFREE;
+	} else {
+		host->flags &= ~HSMMC_SDIO_IRQ_ENABLED;
+		irq_mask &= ~CIRQ_EN;
+		con &= ~(CTPL | CLKEXTFREE);
+	}
+	OMAP_HSMMC_WRITE(host->base, CON, con);
+	OMAP_HSMMC_WRITE(host->base, IE, irq_mask);
+
+	/*
+	 * if enable, piggy back detection on current request
+	 * but always disable immediately
+	 */
+	if (!host->req_in_progress || !enable)
+		OMAP_HSMMC_WRITE(host->base, ISE, irq_mask);
+
+	/* flush posted write */
+	OMAP_HSMMC_READ(host->base, IE);
+
+	spin_unlock_irqrestore(&host->irq_lock, flags);
+}
+
+static int omap_hsmmc_configure_wake_irq(struct omap_hsmmc_host *host)
+{
+	int ret;
+
+	/*
+	 * For omaps with wake-up path, wakeirq will be irq from pinctrl and
+	 * for other omaps, wakeirq will be from GPIO (dat line remuxed to
+	 * gpio). wakeirq is needed to detect sdio irq in runtime suspend state
+	 * with functional clock disabled.
+	 */
+	if (!host->dev->of_node || !host->wake_irq)
+		return -ENODEV;
+
+	ret = dev_pm_set_dedicated_wake_irq(host->dev, host->wake_irq);
+	if (ret) {
+		dev_err(mmc_dev(host->mmc), "Unable to request wake IRQ\n");
+		goto err;
+	}
+
+	/*
+	 * Some omaps don't have wake-up path from deeper idle states
+	 * and need to remux SDIO DAT1 to GPIO for wake-up from idle.
+	 */
+	if (host->pdata->controller_flags & OMAP_HSMMC_SWAKEUP_MISSING) {
+		struct pinctrl *p = devm_pinctrl_get(host->dev);
+		if (IS_ERR(p)) {
+			ret = PTR_ERR(p);
+			goto err_free_irq;
+		}
+
+		if (IS_ERR(pinctrl_lookup_state(p, PINCTRL_STATE_IDLE))) {
+			dev_info(host->dev, "missing idle pinctrl state\n");
+			devm_pinctrl_put(p);
+			ret = -EINVAL;
+			goto err_free_irq;
+		}
+		devm_pinctrl_put(p);
+	}
+
+	OMAP_HSMMC_WRITE(host->base, HCTL,
+			 OMAP_HSMMC_READ(host->base, HCTL) | IWE);
+	return 0;
+
+err_free_irq:
+	dev_pm_clear_wake_irq(host->dev);
+err:
+	dev_warn(host->dev, "no SDIO IRQ support, falling back to polling\n");
+	host->wake_irq = 0;
+	return ret;
+}
+
+static void omap_hsmmc_conf_bus_power(struct omap_hsmmc_host *host)
+{
+	u32 hctl, capa, value;
+
+	/* Only MMC1 supports 3.0V */
+	if (host->pdata->controller_flags & OMAP_HSMMC_SUPPORTS_DUAL_VOLT) {
+		hctl = SDVS30;
+		capa = VS30 | VS18;
+	} else {
+		hctl = SDVS18;
+		capa = VS18;
+	}
+
+	value = OMAP_HSMMC_READ(host->base, HCTL) & ~SDVS_MASK;
+	OMAP_HSMMC_WRITE(host->base, HCTL, value | hctl);
+
+	value = OMAP_HSMMC_READ(host->base, CAPA);
+	OMAP_HSMMC_WRITE(host->base, CAPA, value | capa);
+
+	/* Set SD bus power bit */
+	set_sd_bus_power(host);
+}
+
+static int omap_hsmmc_multi_io_quirk(struct mmc_card *card,
+				     unsigned int direction, int blk_size)
+{
+	/* This controller can't do multiblock reads due to hw bugs */
+	if (direction == MMC_DATA_READ)
+		return 1;
+
+	return blk_size;
+}
+
+static struct mmc_host_ops omap_hsmmc_ops = {
+	.post_req = omap_hsmmc_post_req,
+	.pre_req = omap_hsmmc_pre_req,
+	.request = omap_hsmmc_request,
+	.set_ios = omap_hsmmc_set_ios,
+	.get_cd = mmc_gpio_get_cd,
+	.get_ro = mmc_gpio_get_ro,
+	.init_card = omap_hsmmc_init_card,
+	.enable_sdio_irq = omap_hsmmc_enable_sdio_irq,
+};
+
+#ifdef CONFIG_DEBUG_FS
+
+static int mmc_regs_show(struct seq_file *s, void *data)
+{
+	struct mmc_host *mmc = s->private;
+	struct omap_hsmmc_host *host = mmc_priv(mmc);
+
+	seq_printf(s, "mmc%d:\n", mmc->index);
+	seq_printf(s, "sdio irq mode\t%s\n",
+		   (mmc->caps & MMC_CAP_SDIO_IRQ) ? "interrupt" : "polling");
+
+	if (mmc->caps & MMC_CAP_SDIO_IRQ) {
+		seq_printf(s, "sdio irq \t%s\n",
+			   (host->flags & HSMMC_SDIO_IRQ_ENABLED) ?  "enabled"
+			   : "disabled");
+	}
+	seq_printf(s, "ctx_loss:\t%d\n", host->context_loss);
+
+	pm_runtime_get_sync(host->dev);
+	seq_puts(s, "\nregs:\n");
+	seq_printf(s, "CON:\t\t0x%08x\n",
+			OMAP_HSMMC_READ(host->base, CON));
+	seq_printf(s, "PSTATE:\t\t0x%08x\n",
+		   OMAP_HSMMC_READ(host->base, PSTATE));
+	seq_printf(s, "HCTL:\t\t0x%08x\n",
+			OMAP_HSMMC_READ(host->base, HCTL));
+	seq_printf(s, "SYSCTL:\t\t0x%08x\n",
+			OMAP_HSMMC_READ(host->base, SYSCTL));
+	seq_printf(s, "IE:\t\t0x%08x\n",
+			OMAP_HSMMC_READ(host->base, IE));
+	seq_printf(s, "ISE:\t\t0x%08x\n",
+			OMAP_HSMMC_READ(host->base, ISE));
+	seq_printf(s, "CAPA:\t\t0x%08x\n",
+			OMAP_HSMMC_READ(host->base, CAPA));
+
+	pm_runtime_mark_last_busy(host->dev);
+	pm_runtime_put_autosuspend(host->dev);
+
+	return 0;
+}
+
+DEFINE_SHOW_ATTRIBUTE(mmc_regs);
+
+static void omap_hsmmc_debugfs(struct mmc_host *mmc)
+{
+	if (mmc->debugfs_root)
+		debugfs_create_file("regs", S_IRUSR, mmc->debugfs_root,
+			mmc, &mmc_regs_fops);
+}
+
+#else
+
+static void omap_hsmmc_debugfs(struct mmc_host *mmc)
+{
+}
+
+#endif
+
+#ifdef CONFIG_OF
+static const struct omap_mmc_of_data omap3_pre_es3_mmc_of_data = {
+	/* See 35xx errata 2.1.1.128 in SPRZ278F */
+	.controller_flags = OMAP_HSMMC_BROKEN_MULTIBLOCK_READ,
+};
+
+static const struct omap_mmc_of_data omap4_mmc_of_data = {
+	.reg_offset = 0x100,
+};
+static const struct omap_mmc_of_data am33xx_mmc_of_data = {
+	.reg_offset = 0x100,
+	.controller_flags = OMAP_HSMMC_SWAKEUP_MISSING,
+};
+
+static const struct of_device_id omap_mmc_of_match[] = {
+	{
+		.compatible = "ti,omap2-hsmmc",
+	},
+	{
+		.compatible = "ti,omap3-pre-es3-hsmmc",
+		.data = &omap3_pre_es3_mmc_of_data,
+	},
+	{
+		.compatible = "ti,omap3-hsmmc",
+	},
+	{
+		.compatible = "ti,omap4-hsmmc",
+		.data = &omap4_mmc_of_data,
+	},
+	{
+		.compatible = "ti,am33xx-hsmmc",
+		.data = &am33xx_mmc_of_data,
+	},
+	{},
+};
+MODULE_DEVICE_TABLE(of, omap_mmc_of_match);
+
+static struct omap_hsmmc_platform_data *of_get_hsmmc_pdata(struct device *dev)
+{
+	struct omap_hsmmc_platform_data *pdata, *legacy;
+	struct device_node *np = dev->of_node;
+
+	pdata = devm_kzalloc(dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata)
+		return ERR_PTR(-ENOMEM); /* out of memory */
+
+	legacy = dev_get_platdata(dev);
+	if (legacy && legacy->name)
+		pdata->name = legacy->name;
+
+	if (of_find_property(np, "ti,dual-volt", NULL))
+		pdata->controller_flags |= OMAP_HSMMC_SUPPORTS_DUAL_VOLT;
+
+	if (of_find_property(np, "ti,non-removable", NULL)) {
+		pdata->nonremovable = true;
+		pdata->no_regulator_off_init = true;
+	}
+
+	if (of_find_property(np, "ti,needs-special-reset", NULL))
+		pdata->features |= HSMMC_HAS_UPDATED_RESET;
+
+	if (of_find_property(np, "ti,needs-special-hs-handling", NULL))
+		pdata->features |= HSMMC_HAS_HSPE_SUPPORT;
+
+	return pdata;
+}
+#else
+static inline struct omap_hsmmc_platform_data
+			*of_get_hsmmc_pdata(struct device *dev)
+{
+	return ERR_PTR(-EINVAL);
+}
+#endif
+
+static int omap_hsmmc_probe(struct platform_device *pdev)
+{
+	struct omap_hsmmc_platform_data *pdata = pdev->dev.platform_data;
+	struct mmc_host *mmc;
+	struct omap_hsmmc_host *host = NULL;
+	struct resource *res;
+	int ret, irq;
+	const struct of_device_id *match;
+	const struct omap_mmc_of_data *data;
+	void __iomem *base;
+
+	match = of_match_device(of_match_ptr(omap_mmc_of_match), &pdev->dev);
+	if (match) {
+		pdata = of_get_hsmmc_pdata(&pdev->dev);
+
+		if (IS_ERR(pdata))
+			return PTR_ERR(pdata);
+
+		if (match->data) {
+			data = match->data;
+			pdata->reg_offset = data->reg_offset;
+			pdata->controller_flags |= data->controller_flags;
+		}
+	}
+
+	if (pdata == NULL) {
+		dev_err(&pdev->dev, "Platform Data is missing\n");
+		return -ENXIO;
+	}
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res)
+		return -ENXIO;
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return irq;
+
+	base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(base))
+		return PTR_ERR(base);
+
+	mmc = mmc_alloc_host(sizeof(struct omap_hsmmc_host), &pdev->dev);
+	if (!mmc) {
+		ret = -ENOMEM;
+		goto err;
+	}
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto err1;
+
+	host		= mmc_priv(mmc);
+	host->mmc	= mmc;
+	host->pdata	= pdata;
+	host->dev	= &pdev->dev;
+	host->use_dma	= 1;
+	host->dma_ch	= -1;
+	host->irq	= irq;
+	host->mapbase	= res->start + pdata->reg_offset;
+	host->base	= base + pdata->reg_offset;
+	host->power_mode = MMC_POWER_OFF;
+	host->next_data.cookie = 1;
+	host->pbias_enabled = 0;
+	host->vqmmc_enabled = 0;
+
+	platform_set_drvdata(pdev, host);
+
+	if (pdev->dev.of_node)
+		host->wake_irq = irq_of_parse_and_map(pdev->dev.of_node, 1);
+
+	mmc->ops	= &omap_hsmmc_ops;
+
+	mmc->f_min = OMAP_MMC_MIN_CLOCK;
+
+	if (pdata->max_freq > 0)
+		mmc->f_max = pdata->max_freq;
+	else if (mmc->f_max == 0)
+		mmc->f_max = OMAP_MMC_MAX_CLOCK;
+
+	spin_lock_init(&host->irq_lock);
+
+	host->fclk = devm_clk_get(&pdev->dev, "fck");
+	if (IS_ERR(host->fclk)) {
+		ret = PTR_ERR(host->fclk);
+		host->fclk = NULL;
+		goto err1;
+	}
+
+	if (host->pdata->controller_flags & OMAP_HSMMC_BROKEN_MULTIBLOCK_READ) {
+		dev_info(&pdev->dev, "multiblock reads disabled due to 35xx erratum 2.1.1.128; MMC read performance may suffer\n");
+		omap_hsmmc_ops.multi_io_quirk = omap_hsmmc_multi_io_quirk;
+	}
+
+	device_init_wakeup(&pdev->dev, true);
+	pm_runtime_enable(host->dev);
+	pm_runtime_get_sync(host->dev);
+	pm_runtime_set_autosuspend_delay(host->dev, MMC_AUTOSUSPEND_DELAY);
+	pm_runtime_use_autosuspend(host->dev);
+
+	omap_hsmmc_context_save(host);
+
+	host->dbclk = devm_clk_get(&pdev->dev, "mmchsdb_fck");
+	/*
+	 * MMC can still work without debounce clock.
+	 */
+	if (IS_ERR(host->dbclk)) {
+		host->dbclk = NULL;
+	} else if (clk_prepare_enable(host->dbclk) != 0) {
+		dev_warn(mmc_dev(host->mmc), "Failed to enable debounce clk\n");
+		host->dbclk = NULL;
+	}
+
+	/* Set this to a value that allows allocating an entire descriptor
+	 * list within a page (zero order allocation). */
+	mmc->max_segs = 64;
+
+	mmc->max_blk_size = 512;       /* Block Length at max can be 1024 */
+	mmc->max_blk_count = 0xFFFF;    /* No. of Blocks is 16 bits */
+	mmc->max_req_size = mmc->max_blk_size * mmc->max_blk_count;
+
+	mmc->caps |= MMC_CAP_MMC_HIGHSPEED | MMC_CAP_SD_HIGHSPEED |
+		     MMC_CAP_WAIT_WHILE_BUSY | MMC_CAP_CMD23;
+
+	mmc->caps |= mmc_pdata(host)->caps;
+	if (mmc->caps & MMC_CAP_8_BIT_DATA)
+		mmc->caps |= MMC_CAP_4_BIT_DATA;
+
+	if (mmc_pdata(host)->nonremovable)
+		mmc->caps |= MMC_CAP_NONREMOVABLE;
+
+	mmc->pm_caps |= mmc_pdata(host)->pm_caps;
+
+	omap_hsmmc_conf_bus_power(host);
+
+	host->rx_chan = dma_request_chan(&pdev->dev, "rx");
+	if (IS_ERR(host->rx_chan)) {
+		dev_err(mmc_dev(host->mmc), "RX DMA channel request failed\n");
+		ret = PTR_ERR(host->rx_chan);
+		goto err_irq;
+	}
+
+	host->tx_chan = dma_request_chan(&pdev->dev, "tx");
+	if (IS_ERR(host->tx_chan)) {
+		dev_err(mmc_dev(host->mmc), "TX DMA channel request failed\n");
+		ret = PTR_ERR(host->tx_chan);
+		goto err_irq;
+	}
+
+	/*
+	 * Limit the maximum segment size to the lower of the request size
+	 * and the DMA engine device segment size limits.  In reality, with
+	 * 32-bit transfers, the DMA engine can do longer segments than this
+	 * but there is no way to represent that in the DMA model - if we
+	 * increase this figure here, we get warnings from the DMA API debug.
+	 */
+	mmc->max_seg_size = min3(mmc->max_req_size,
+			dma_get_max_seg_size(host->rx_chan->device->dev),
+			dma_get_max_seg_size(host->tx_chan->device->dev));
+
+	/* Request IRQ for MMC operations */
+	ret = devm_request_irq(&pdev->dev, host->irq, omap_hsmmc_irq, 0,
+			mmc_hostname(mmc), host);
+	if (ret) {
+		dev_err(mmc_dev(host->mmc), "Unable to grab HSMMC IRQ\n");
+		goto err_irq;
+	}
+
+	ret = omap_hsmmc_reg_get(host);
+	if (ret)
+		goto err_irq;
+
+	if (!mmc->ocr_avail)
+		mmc->ocr_avail = mmc_pdata(host)->ocr_mask;
+
+	omap_hsmmc_disable_irq(host);
+
+	/*
+	 * For now, only support SDIO interrupt if we have a separate
+	 * wake-up interrupt configured from device tree. This is because
+	 * the wake-up interrupt is needed for idle state and some
+	 * platforms need special quirks. And we don't want to add new
+	 * legacy mux platform init code callbacks any longer as we
+	 * are moving to DT based booting anyways.
+	 */
+	ret = omap_hsmmc_configure_wake_irq(host);
+	if (!ret)
+		mmc->caps |= MMC_CAP_SDIO_IRQ;
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto err_irq;
+
+	if (mmc_pdata(host)->name != NULL) {
+		ret = device_create_file(&mmc->class_dev, &dev_attr_slot_name);
+		if (ret < 0)
+			goto err_slot_name;
+	}
+
+	omap_hsmmc_debugfs(mmc);
+	pm_runtime_mark_last_busy(host->dev);
+	pm_runtime_put_autosuspend(host->dev);
+
+	return 0;
+
+err_slot_name:
+	mmc_remove_host(mmc);
+err_irq:
+	device_init_wakeup(&pdev->dev, false);
+	if (!IS_ERR_OR_NULL(host->tx_chan))
+		dma_release_channel(host->tx_chan);
+	if (!IS_ERR_OR_NULL(host->rx_chan))
+		dma_release_channel(host->rx_chan);
+	pm_runtime_dont_use_autosuspend(host->dev);
+	pm_runtime_put_sync(host->dev);
+	pm_runtime_disable(host->dev);
+	clk_disable_unprepare(host->dbclk);
+err1:
+	mmc_free_host(mmc);
+err:
+	return ret;
+}
+
+static int omap_hsmmc_remove(struct platform_device *pdev)
+{
+	struct omap_hsmmc_host *host = platform_get_drvdata(pdev);
+
+	pm_runtime_get_sync(host->dev);
+	mmc_remove_host(host->mmc);
+
+	dma_release_channel(host->tx_chan);
+	dma_release_channel(host->rx_chan);
+
+	dev_pm_clear_wake_irq(host->dev);
+	pm_runtime_dont_use_autosuspend(host->dev);
+	pm_runtime_put_sync(host->dev);
+	pm_runtime_disable(host->dev);
+	device_init_wakeup(&pdev->dev, false);
+	clk_disable_unprepare(host->dbclk);
+
+	mmc_free_host(host->mmc);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int omap_hsmmc_suspend(struct device *dev)
+{
+	struct omap_hsmmc_host *host = dev_get_drvdata(dev);
+
+	if (!host)
+		return 0;
+
+	pm_runtime_get_sync(host->dev);
+
+	if (!(host->mmc->pm_flags & MMC_PM_KEEP_POWER)) {
+		OMAP_HSMMC_WRITE(host->base, ISE, 0);
+		OMAP_HSMMC_WRITE(host->base, IE, 0);
+		OMAP_HSMMC_WRITE(host->base, STAT, STAT_CLEAR);
+		OMAP_HSMMC_WRITE(host->base, HCTL,
+				OMAP_HSMMC_READ(host->base, HCTL) & ~SDBP);
+	}
+
+	clk_disable_unprepare(host->dbclk);
+
+	pm_runtime_put_sync(host->dev);
+	return 0;
+}
+
+/* Routine to resume the MMC device */
+static int omap_hsmmc_resume(struct device *dev)
+{
+	struct omap_hsmmc_host *host = dev_get_drvdata(dev);
+
+	if (!host)
+		return 0;
+
+	pm_runtime_get_sync(host->dev);
+
+	clk_prepare_enable(host->dbclk);
+
+	if (!(host->mmc->pm_flags & MMC_PM_KEEP_POWER))
+		omap_hsmmc_conf_bus_power(host);
+
+	pm_runtime_mark_last_busy(host->dev);
+	pm_runtime_put_autosuspend(host->dev);
+	return 0;
+}
+#endif
+
+static int omap_hsmmc_runtime_suspend(struct device *dev)
+{
+	struct omap_hsmmc_host *host;
+	unsigned long flags;
+	int ret = 0;
+
+	host = dev_get_drvdata(dev);
+	omap_hsmmc_context_save(host);
+	dev_dbg(dev, "disabled\n");
+
+	spin_lock_irqsave(&host->irq_lock, flags);
+	if ((host->mmc->caps & MMC_CAP_SDIO_IRQ) &&
+	    (host->flags & HSMMC_SDIO_IRQ_ENABLED)) {
+		/* disable sdio irq handling to prevent race */
+		OMAP_HSMMC_WRITE(host->base, ISE, 0);
+		OMAP_HSMMC_WRITE(host->base, IE, 0);
+
+		if (!(OMAP_HSMMC_READ(host->base, PSTATE) & DLEV_DAT(1))) {
+			/*
+			 * dat1 line low, pending sdio irq
+			 * race condition: possible irq handler running on
+			 * multi-core, abort
+			 */
+			dev_dbg(dev, "pending sdio irq, abort suspend\n");
+			OMAP_HSMMC_WRITE(host->base, STAT, STAT_CLEAR);
+			OMAP_HSMMC_WRITE(host->base, ISE, CIRQ_EN);
+			OMAP_HSMMC_WRITE(host->base, IE, CIRQ_EN);
+			pm_runtime_mark_last_busy(dev);
+			ret = -EBUSY;
+			goto abort;
+		}
+
+		pinctrl_pm_select_idle_state(dev);
+	} else {
+		pinctrl_pm_select_idle_state(dev);
+	}
+
+abort:
+	spin_unlock_irqrestore(&host->irq_lock, flags);
+	return ret;
+}
+
+static int omap_hsmmc_runtime_resume(struct device *dev)
+{
+	struct omap_hsmmc_host *host;
+	unsigned long flags;
+
+	host = dev_get_drvdata(dev);
+	omap_hsmmc_context_restore(host);
+	dev_dbg(dev, "enabled\n");
+
+	spin_lock_irqsave(&host->irq_lock, flags);
+	if ((host->mmc->caps & MMC_CAP_SDIO_IRQ) &&
+	    (host->flags & HSMMC_SDIO_IRQ_ENABLED)) {
+
+		pinctrl_select_default_state(host->dev);
+
+		/* irq lost, if pinmux incorrect */
+		OMAP_HSMMC_WRITE(host->base, STAT, STAT_CLEAR);
+		OMAP_HSMMC_WRITE(host->base, ISE, CIRQ_EN);
+		OMAP_HSMMC_WRITE(host->base, IE, CIRQ_EN);
+	} else {
+		pinctrl_select_default_state(host->dev);
+	}
+	spin_unlock_irqrestore(&host->irq_lock, flags);
+	return 0;
+}
+
+static const struct dev_pm_ops omap_hsmmc_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(omap_hsmmc_suspend, omap_hsmmc_resume)
+	.runtime_suspend = omap_hsmmc_runtime_suspend,
+	.runtime_resume = omap_hsmmc_runtime_resume,
+};
+
+static struct platform_driver omap_hsmmc_driver = {
+	.probe		= omap_hsmmc_probe,
+	.remove		= omap_hsmmc_remove,
+	.driver		= {
+		.name = DRIVER_NAME,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm = &omap_hsmmc_dev_pm_ops,
+		.of_match_table = of_match_ptr(omap_mmc_of_match),
+	},
+};
+
+module_platform_driver(omap_hsmmc_driver);
+MODULE_DESCRIPTION("OMAP High Speed Multimedia Card driver");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:" DRIVER_NAME);
+MODULE_AUTHOR("Texas Instruments Inc");
diff --git a/drivers/mmc/host/owl-mmc.c b/drivers/mmc/host/owl-mmc.c
new file mode 100644
index 000000000..8a40cf8a9
--- /dev/null
+++ b/drivers/mmc/host/owl-mmc.c
@@ -0,0 +1,706 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Actions Semi Owl SoCs SD/MMC driver
+ *
+ * Copyright (c) 2014 Actions Semi Inc.
+ * Copyright (c) 2019 Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
+ *
+ * TODO: SDIO support
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-direction.h>
+#include <linux/dma-mapping.h>
+#include <linux/interrupt.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/module.h>
+#include <linux/of_platform.h>
+#include <linux/reset.h>
+#include <linux/spinlock.h>
+
+/*
+ * SDC registers
+ */
+#define OWL_REG_SD_EN			0x0000
+#define OWL_REG_SD_CTL			0x0004
+#define OWL_REG_SD_STATE		0x0008
+#define OWL_REG_SD_CMD			0x000c
+#define OWL_REG_SD_ARG			0x0010
+#define OWL_REG_SD_RSPBUF0		0x0014
+#define OWL_REG_SD_RSPBUF1		0x0018
+#define OWL_REG_SD_RSPBUF2		0x001c
+#define OWL_REG_SD_RSPBUF3		0x0020
+#define OWL_REG_SD_RSPBUF4		0x0024
+#define OWL_REG_SD_DAT			0x0028
+#define OWL_REG_SD_BLK_SIZE		0x002c
+#define OWL_REG_SD_BLK_NUM		0x0030
+#define OWL_REG_SD_BUF_SIZE		0x0034
+
+/* SD_EN Bits */
+#define OWL_SD_EN_RANE			BIT(31)
+#define OWL_SD_EN_RAN_SEED(x)		(((x) & 0x3f) << 24)
+#define OWL_SD_EN_S18EN			BIT(12)
+#define OWL_SD_EN_RESE			BIT(10)
+#define OWL_SD_EN_DAT1_S		BIT(9)
+#define OWL_SD_EN_CLK_S			BIT(8)
+#define OWL_SD_ENABLE			BIT(7)
+#define OWL_SD_EN_BSEL			BIT(6)
+#define OWL_SD_EN_SDIOEN		BIT(3)
+#define OWL_SD_EN_DDREN			BIT(2)
+#define OWL_SD_EN_DATAWID(x)		(((x) & 0x3) << 0)
+
+/* SD_CTL Bits */
+#define OWL_SD_CTL_TOUTEN		BIT(31)
+#define OWL_SD_CTL_TOUTCNT(x)		(((x) & 0x7f) << 24)
+#define OWL_SD_CTL_DELAY_MSK		GENMASK(23, 16)
+#define OWL_SD_CTL_RDELAY(x)		(((x) & 0xf) << 20)
+#define OWL_SD_CTL_WDELAY(x)		(((x) & 0xf) << 16)
+#define OWL_SD_CTL_CMDLEN		BIT(13)
+#define OWL_SD_CTL_SCC			BIT(12)
+#define OWL_SD_CTL_TCN(x)		(((x) & 0xf) << 8)
+#define OWL_SD_CTL_TS			BIT(7)
+#define OWL_SD_CTL_LBE			BIT(6)
+#define OWL_SD_CTL_C7EN			BIT(5)
+#define OWL_SD_CTL_TM(x)		(((x) & 0xf) << 0)
+
+#define OWL_SD_DELAY_LOW_CLK		0x0f
+#define OWL_SD_DELAY_MID_CLK		0x0a
+#define OWL_SD_DELAY_HIGH_CLK		0x09
+#define OWL_SD_RDELAY_DDR50		0x0a
+#define OWL_SD_WDELAY_DDR50		0x08
+
+/* SD_STATE Bits */
+#define OWL_SD_STATE_DAT1BS		BIT(18)
+#define OWL_SD_STATE_SDIOB_P		BIT(17)
+#define OWL_SD_STATE_SDIOB_EN		BIT(16)
+#define OWL_SD_STATE_TOUTE		BIT(15)
+#define OWL_SD_STATE_BAEP		BIT(14)
+#define OWL_SD_STATE_MEMRDY		BIT(12)
+#define OWL_SD_STATE_CMDS		BIT(11)
+#define OWL_SD_STATE_DAT1AS		BIT(10)
+#define OWL_SD_STATE_SDIOA_P		BIT(9)
+#define OWL_SD_STATE_SDIOA_EN		BIT(8)
+#define OWL_SD_STATE_DAT0S		BIT(7)
+#define OWL_SD_STATE_TEIE		BIT(6)
+#define OWL_SD_STATE_TEI		BIT(5)
+#define OWL_SD_STATE_CLNR		BIT(4)
+#define OWL_SD_STATE_CLC		BIT(3)
+#define OWL_SD_STATE_WC16ER		BIT(2)
+#define OWL_SD_STATE_RC16ER		BIT(1)
+#define OWL_SD_STATE_CRC7ER		BIT(0)
+
+#define OWL_CMD_TIMEOUT_MS		30000
+
+struct owl_mmc_host {
+	struct device *dev;
+	struct reset_control *reset;
+	void __iomem *base;
+	struct clk *clk;
+	struct completion sdc_complete;
+	spinlock_t lock;
+	int irq;
+	u32 clock;
+	bool ddr_50;
+
+	enum dma_data_direction dma_dir;
+	struct dma_chan *dma;
+	struct dma_async_tx_descriptor *desc;
+	struct dma_slave_config dma_cfg;
+	struct completion dma_complete;
+
+	struct mmc_host	*mmc;
+	struct mmc_request *mrq;
+	struct mmc_command *cmd;
+	struct mmc_data	*data;
+};
+
+static void owl_mmc_update_reg(void __iomem *reg, unsigned int val, bool state)
+{
+	unsigned int regval;
+
+	regval = readl(reg);
+
+	if (state)
+		regval |= val;
+	else
+		regval &= ~val;
+
+	writel(regval, reg);
+}
+
+static irqreturn_t owl_irq_handler(int irq, void *devid)
+{
+	struct owl_mmc_host *owl_host = devid;
+	unsigned long flags;
+	u32 state;
+
+	spin_lock_irqsave(&owl_host->lock, flags);
+
+	state = readl(owl_host->base + OWL_REG_SD_STATE);
+	if (state & OWL_SD_STATE_TEI) {
+		state = readl(owl_host->base + OWL_REG_SD_STATE);
+		state |= OWL_SD_STATE_TEI;
+		writel(state, owl_host->base + OWL_REG_SD_STATE);
+		complete(&owl_host->sdc_complete);
+	}
+
+	spin_unlock_irqrestore(&owl_host->lock, flags);
+
+	return IRQ_HANDLED;
+}
+
+static void owl_mmc_finish_request(struct owl_mmc_host *owl_host)
+{
+	struct mmc_request *mrq = owl_host->mrq;
+	struct mmc_data *data = mrq->data;
+
+	/* Should never be NULL */
+	WARN_ON(!mrq);
+
+	owl_host->mrq = NULL;
+
+	if (data)
+		dma_unmap_sg(owl_host->dma->device->dev, data->sg, data->sg_len,
+			     owl_host->dma_dir);
+
+	/* Finally finish request */
+	mmc_request_done(owl_host->mmc, mrq);
+}
+
+static void owl_mmc_send_cmd(struct owl_mmc_host *owl_host,
+			     struct mmc_command *cmd,
+			     struct mmc_data *data)
+{
+	unsigned long timeout;
+	u32 mode, state, resp[2];
+	u32 cmd_rsp_mask = 0;
+
+	init_completion(&owl_host->sdc_complete);
+
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE:
+		mode = OWL_SD_CTL_TM(0);
+		break;
+
+	case MMC_RSP_R1:
+		if (data) {
+			if (data->flags & MMC_DATA_READ)
+				mode = OWL_SD_CTL_TM(4);
+			else
+				mode = OWL_SD_CTL_TM(5);
+		} else {
+			mode = OWL_SD_CTL_TM(1);
+		}
+		cmd_rsp_mask = OWL_SD_STATE_CLNR | OWL_SD_STATE_CRC7ER;
+
+		break;
+
+	case MMC_RSP_R1B:
+		mode = OWL_SD_CTL_TM(3);
+		cmd_rsp_mask = OWL_SD_STATE_CLNR | OWL_SD_STATE_CRC7ER;
+		break;
+
+	case MMC_RSP_R2:
+		mode = OWL_SD_CTL_TM(2);
+		cmd_rsp_mask = OWL_SD_STATE_CLNR | OWL_SD_STATE_CRC7ER;
+		break;
+
+	case MMC_RSP_R3:
+		mode = OWL_SD_CTL_TM(1);
+		cmd_rsp_mask = OWL_SD_STATE_CLNR;
+		break;
+
+	default:
+		dev_warn(owl_host->dev, "Unknown MMC command\n");
+		cmd->error = -EINVAL;
+		return;
+	}
+
+	/* Keep current WDELAY and RDELAY */
+	mode |= (readl(owl_host->base + OWL_REG_SD_CTL) & (0xff << 16));
+
+	/* Start to send corresponding command type */
+	writel(cmd->arg, owl_host->base + OWL_REG_SD_ARG);
+	writel(cmd->opcode, owl_host->base + OWL_REG_SD_CMD);
+
+	/* Set LBE to send clk at the end of last read block */
+	if (data) {
+		mode |= (OWL_SD_CTL_TS | OWL_SD_CTL_LBE | 0x64000000);
+	} else {
+		mode &= ~(OWL_SD_CTL_TOUTEN | OWL_SD_CTL_LBE);
+		mode |= OWL_SD_CTL_TS;
+	}
+
+	owl_host->cmd = cmd;
+
+	/* Start transfer */
+	writel(mode, owl_host->base + OWL_REG_SD_CTL);
+
+	if (data)
+		return;
+
+	timeout = msecs_to_jiffies(cmd->busy_timeout ? cmd->busy_timeout :
+		OWL_CMD_TIMEOUT_MS);
+
+	if (!wait_for_completion_timeout(&owl_host->sdc_complete, timeout)) {
+		dev_err(owl_host->dev, "CMD interrupt timeout\n");
+		cmd->error = -ETIMEDOUT;
+		return;
+	}
+
+	state = readl(owl_host->base + OWL_REG_SD_STATE);
+	if (mmc_resp_type(cmd) & MMC_RSP_PRESENT) {
+		if (cmd_rsp_mask & state) {
+			if (state & OWL_SD_STATE_CLNR) {
+				dev_err(owl_host->dev, "Error CMD_NO_RSP\n");
+				cmd->error = -EILSEQ;
+				return;
+			}
+
+			if (state & OWL_SD_STATE_CRC7ER) {
+				dev_err(owl_host->dev, "Error CMD_RSP_CRC\n");
+				cmd->error = -EILSEQ;
+				return;
+			}
+		}
+
+		if (mmc_resp_type(cmd) & MMC_RSP_136) {
+			cmd->resp[3] = readl(owl_host->base + OWL_REG_SD_RSPBUF0);
+			cmd->resp[2] = readl(owl_host->base + OWL_REG_SD_RSPBUF1);
+			cmd->resp[1] = readl(owl_host->base + OWL_REG_SD_RSPBUF2);
+			cmd->resp[0] = readl(owl_host->base + OWL_REG_SD_RSPBUF3);
+		} else {
+			resp[0] = readl(owl_host->base + OWL_REG_SD_RSPBUF0);
+			resp[1] = readl(owl_host->base + OWL_REG_SD_RSPBUF1);
+			cmd->resp[0] = resp[1] << 24 | resp[0] >> 8;
+			cmd->resp[1] = resp[1] >> 8;
+		}
+	}
+}
+
+static void owl_mmc_dma_complete(void *param)
+{
+	struct owl_mmc_host *owl_host = param;
+	struct mmc_data *data = owl_host->data;
+
+	if (data)
+		complete(&owl_host->dma_complete);
+}
+
+static int owl_mmc_prepare_data(struct owl_mmc_host *owl_host,
+				struct mmc_data *data)
+{
+	u32 total;
+
+	owl_mmc_update_reg(owl_host->base + OWL_REG_SD_EN, OWL_SD_EN_BSEL,
+			   true);
+	writel(data->blocks, owl_host->base + OWL_REG_SD_BLK_NUM);
+	writel(data->blksz, owl_host->base + OWL_REG_SD_BLK_SIZE);
+	total = data->blksz * data->blocks;
+
+	if (total < 512)
+		writel(total, owl_host->base + OWL_REG_SD_BUF_SIZE);
+	else
+		writel(512, owl_host->base + OWL_REG_SD_BUF_SIZE);
+
+	if (data->flags & MMC_DATA_WRITE) {
+		owl_host->dma_dir = DMA_TO_DEVICE;
+		owl_host->dma_cfg.direction = DMA_MEM_TO_DEV;
+	} else {
+		owl_host->dma_dir = DMA_FROM_DEVICE;
+		owl_host->dma_cfg.direction = DMA_DEV_TO_MEM;
+	}
+
+	dma_map_sg(owl_host->dma->device->dev, data->sg,
+		   data->sg_len, owl_host->dma_dir);
+
+	dmaengine_slave_config(owl_host->dma, &owl_host->dma_cfg);
+	owl_host->desc = dmaengine_prep_slave_sg(owl_host->dma, data->sg,
+						 data->sg_len,
+						 owl_host->dma_cfg.direction,
+						 DMA_PREP_INTERRUPT |
+						 DMA_CTRL_ACK);
+	if (!owl_host->desc) {
+		dev_err(owl_host->dev, "Can't prepare slave sg\n");
+		return -EBUSY;
+	}
+
+	owl_host->data = data;
+
+	owl_host->desc->callback = owl_mmc_dma_complete;
+	owl_host->desc->callback_param = (void *)owl_host;
+	data->error = 0;
+
+	return 0;
+}
+
+static void owl_mmc_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct owl_mmc_host *owl_host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+	int ret;
+
+	owl_host->mrq = mrq;
+	if (mrq->data) {
+		ret = owl_mmc_prepare_data(owl_host, data);
+		if (ret < 0) {
+			data->error = ret;
+			goto err_out;
+		}
+
+		init_completion(&owl_host->dma_complete);
+		dmaengine_submit(owl_host->desc);
+		dma_async_issue_pending(owl_host->dma);
+	}
+
+	owl_mmc_send_cmd(owl_host, mrq->cmd, data);
+
+	if (data) {
+		if (!wait_for_completion_timeout(&owl_host->sdc_complete,
+						 10 * HZ)) {
+			dev_err(owl_host->dev, "CMD interrupt timeout\n");
+			mrq->cmd->error = -ETIMEDOUT;
+			dmaengine_terminate_all(owl_host->dma);
+			goto err_out;
+		}
+
+		if (!wait_for_completion_timeout(&owl_host->dma_complete,
+						 5 * HZ)) {
+			dev_err(owl_host->dev, "DMA interrupt timeout\n");
+			mrq->cmd->error = -ETIMEDOUT;
+			dmaengine_terminate_all(owl_host->dma);
+			goto err_out;
+		}
+
+		if (data->stop)
+			owl_mmc_send_cmd(owl_host, data->stop, NULL);
+
+		data->bytes_xfered = data->blocks * data->blksz;
+	}
+
+err_out:
+	owl_mmc_finish_request(owl_host);
+}
+
+static int owl_mmc_set_clk_rate(struct owl_mmc_host *owl_host,
+				unsigned int rate)
+{
+	unsigned long clk_rate;
+	int ret;
+	u32 reg;
+
+	reg = readl(owl_host->base + OWL_REG_SD_CTL);
+	reg &= ~OWL_SD_CTL_DELAY_MSK;
+
+	/* Set RDELAY and WDELAY based on the clock */
+	if (rate <= 1000000) {
+		writel(reg | OWL_SD_CTL_RDELAY(OWL_SD_DELAY_LOW_CLK) |
+		       OWL_SD_CTL_WDELAY(OWL_SD_DELAY_LOW_CLK),
+		       owl_host->base + OWL_REG_SD_CTL);
+	} else if ((rate > 1000000) && (rate <= 26000000)) {
+		writel(reg | OWL_SD_CTL_RDELAY(OWL_SD_DELAY_MID_CLK) |
+		       OWL_SD_CTL_WDELAY(OWL_SD_DELAY_MID_CLK),
+		       owl_host->base + OWL_REG_SD_CTL);
+	} else if ((rate > 26000000) && (rate <= 52000000) && !owl_host->ddr_50) {
+		writel(reg | OWL_SD_CTL_RDELAY(OWL_SD_DELAY_HIGH_CLK) |
+		       OWL_SD_CTL_WDELAY(OWL_SD_DELAY_HIGH_CLK),
+		       owl_host->base + OWL_REG_SD_CTL);
+	/* DDR50 mode has special delay chain */
+	} else if ((rate > 26000000) && (rate <= 52000000) && owl_host->ddr_50) {
+		writel(reg | OWL_SD_CTL_RDELAY(OWL_SD_RDELAY_DDR50) |
+		       OWL_SD_CTL_WDELAY(OWL_SD_WDELAY_DDR50),
+		       owl_host->base + OWL_REG_SD_CTL);
+	} else {
+		dev_err(owl_host->dev, "SD clock rate not supported\n");
+		return -EINVAL;
+	}
+
+	clk_rate = clk_round_rate(owl_host->clk, rate << 1);
+	ret = clk_set_rate(owl_host->clk, clk_rate);
+
+	return ret;
+}
+
+static void owl_mmc_set_clk(struct owl_mmc_host *owl_host, struct mmc_ios *ios)
+{
+	if (!ios->clock)
+		return;
+
+	owl_host->clock = ios->clock;
+	owl_mmc_set_clk_rate(owl_host, ios->clock);
+}
+
+static void owl_mmc_set_bus_width(struct owl_mmc_host *owl_host,
+				  struct mmc_ios *ios)
+{
+	u32 reg;
+
+	reg = readl(owl_host->base + OWL_REG_SD_EN);
+	reg &= ~0x03;
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_1:
+		break;
+	case MMC_BUS_WIDTH_4:
+		reg |= OWL_SD_EN_DATAWID(1);
+		break;
+	case MMC_BUS_WIDTH_8:
+		reg |= OWL_SD_EN_DATAWID(2);
+		break;
+	}
+
+	writel(reg, owl_host->base + OWL_REG_SD_EN);
+}
+
+static void owl_mmc_ctr_reset(struct owl_mmc_host *owl_host)
+{
+	reset_control_assert(owl_host->reset);
+	udelay(20);
+	reset_control_deassert(owl_host->reset);
+}
+
+static void owl_mmc_power_on(struct owl_mmc_host *owl_host)
+{
+	u32 mode;
+
+	init_completion(&owl_host->sdc_complete);
+
+	/* Enable transfer end IRQ */
+	owl_mmc_update_reg(owl_host->base + OWL_REG_SD_STATE,
+		       OWL_SD_STATE_TEIE, true);
+
+	/* Send init clk */
+	mode = (readl(owl_host->base + OWL_REG_SD_CTL) & (0xff << 16));
+	mode |= OWL_SD_CTL_TS | OWL_SD_CTL_TCN(5) | OWL_SD_CTL_TM(8);
+	writel(mode, owl_host->base + OWL_REG_SD_CTL);
+
+	if (!wait_for_completion_timeout(&owl_host->sdc_complete, HZ)) {
+		dev_err(owl_host->dev, "CMD interrupt timeout\n");
+		return;
+	}
+}
+
+static void owl_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct owl_mmc_host *owl_host = mmc_priv(mmc);
+
+	switch (ios->power_mode) {
+	case MMC_POWER_UP:
+		dev_dbg(owl_host->dev, "Powering card up\n");
+
+		/* Reset the SDC controller to clear all previous states */
+		owl_mmc_ctr_reset(owl_host);
+		clk_prepare_enable(owl_host->clk);
+		writel(OWL_SD_ENABLE | OWL_SD_EN_RESE,
+		       owl_host->base + OWL_REG_SD_EN);
+
+		break;
+
+	case MMC_POWER_ON:
+		dev_dbg(owl_host->dev, "Powering card on\n");
+		owl_mmc_power_on(owl_host);
+
+		break;
+
+	case MMC_POWER_OFF:
+		dev_dbg(owl_host->dev, "Powering card off\n");
+		clk_disable_unprepare(owl_host->clk);
+
+		return;
+
+	default:
+		dev_dbg(owl_host->dev, "Ignoring unknown card power state\n");
+		break;
+	}
+
+	if (ios->clock != owl_host->clock)
+		owl_mmc_set_clk(owl_host, ios);
+
+	owl_mmc_set_bus_width(owl_host, ios);
+
+	/* Enable DDR mode if requested */
+	if (ios->timing == MMC_TIMING_UHS_DDR50) {
+		owl_host->ddr_50 = 1;
+		owl_mmc_update_reg(owl_host->base + OWL_REG_SD_EN,
+			       OWL_SD_EN_DDREN, true);
+	} else {
+		owl_host->ddr_50 = 0;
+	}
+}
+
+static int owl_mmc_start_signal_voltage_switch(struct mmc_host *mmc,
+					       struct mmc_ios *ios)
+{
+	struct owl_mmc_host *owl_host = mmc_priv(mmc);
+
+	/* It is enough to change the pad ctrl bit for voltage switch */
+	switch (ios->signal_voltage) {
+	case MMC_SIGNAL_VOLTAGE_330:
+		owl_mmc_update_reg(owl_host->base + OWL_REG_SD_EN,
+			       OWL_SD_EN_S18EN, false);
+		break;
+	case MMC_SIGNAL_VOLTAGE_180:
+		owl_mmc_update_reg(owl_host->base + OWL_REG_SD_EN,
+			       OWL_SD_EN_S18EN, true);
+		break;
+	default:
+		return -ENOTSUPP;
+	}
+
+	return 0;
+}
+
+static const struct mmc_host_ops owl_mmc_ops = {
+	.request	= owl_mmc_request,
+	.set_ios	= owl_mmc_set_ios,
+	.get_ro		= mmc_gpio_get_ro,
+	.get_cd		= mmc_gpio_get_cd,
+	.start_signal_voltage_switch = owl_mmc_start_signal_voltage_switch,
+};
+
+static int owl_mmc_probe(struct platform_device *pdev)
+{
+	struct owl_mmc_host *owl_host;
+	struct mmc_host *mmc;
+	struct resource *res;
+	int ret;
+
+	mmc = mmc_alloc_host(sizeof(struct owl_mmc_host), &pdev->dev);
+	if (!mmc) {
+		dev_err(&pdev->dev, "mmc alloc host failed\n");
+		return -ENOMEM;
+	}
+	platform_set_drvdata(pdev, mmc);
+
+	owl_host = mmc_priv(mmc);
+	owl_host->dev = &pdev->dev;
+	owl_host->mmc = mmc;
+	spin_lock_init(&owl_host->lock);
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	owl_host->base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(owl_host->base)) {
+		dev_err(&pdev->dev, "Failed to remap registers\n");
+		ret = PTR_ERR(owl_host->base);
+		goto err_free_host;
+	}
+
+	owl_host->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(owl_host->clk)) {
+		dev_err(&pdev->dev, "No clock defined\n");
+		ret = PTR_ERR(owl_host->clk);
+		goto err_free_host;
+	}
+
+	owl_host->reset = devm_reset_control_get_exclusive(&pdev->dev, NULL);
+	if (IS_ERR(owl_host->reset)) {
+		dev_err(&pdev->dev, "Could not get reset control\n");
+		ret = PTR_ERR(owl_host->reset);
+		goto err_free_host;
+	}
+
+	mmc->ops		= &owl_mmc_ops;
+	mmc->max_blk_count	= 512;
+	mmc->max_blk_size	= 512;
+	mmc->max_segs		= 256;
+	mmc->max_seg_size	= 262144;
+	mmc->max_req_size	= 262144;
+	/* 100kHz ~ 52MHz */
+	mmc->f_min		= 100000;
+	mmc->f_max		= 52000000;
+	mmc->caps	       |= MMC_CAP_MMC_HIGHSPEED | MMC_CAP_SD_HIGHSPEED |
+				  MMC_CAP_4_BIT_DATA;
+	mmc->caps2		= (MMC_CAP2_BOOTPART_NOACC | MMC_CAP2_NO_SDIO);
+	mmc->ocr_avail		= MMC_VDD_32_33 | MMC_VDD_33_34 |
+				  MMC_VDD_165_195;
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto err_free_host;
+
+	pdev->dev.coherent_dma_mask = DMA_BIT_MASK(32);
+	pdev->dev.dma_mask = &pdev->dev.coherent_dma_mask;
+	owl_host->dma = dma_request_chan(&pdev->dev, "mmc");
+	if (IS_ERR(owl_host->dma)) {
+		dev_err(owl_host->dev, "Failed to get external DMA channel.\n");
+		ret = PTR_ERR(owl_host->dma);
+		goto err_free_host;
+	}
+
+	dev_info(&pdev->dev, "Using %s for DMA transfers\n",
+		 dma_chan_name(owl_host->dma));
+
+	owl_host->dma_cfg.src_addr = res->start + OWL_REG_SD_DAT;
+	owl_host->dma_cfg.dst_addr = res->start + OWL_REG_SD_DAT;
+	owl_host->dma_cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	owl_host->dma_cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	owl_host->dma_cfg.device_fc = false;
+
+	owl_host->irq = platform_get_irq(pdev, 0);
+	if (owl_host->irq < 0) {
+		ret = owl_host->irq;
+		goto err_release_channel;
+	}
+
+	ret = devm_request_irq(&pdev->dev, owl_host->irq, owl_irq_handler,
+			       0, dev_name(&pdev->dev), owl_host);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to request irq %d\n",
+			owl_host->irq);
+		goto err_release_channel;
+	}
+
+	ret = mmc_add_host(mmc);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to add host\n");
+		goto err_release_channel;
+	}
+
+	dev_dbg(&pdev->dev, "Owl MMC Controller Initialized\n");
+
+	return 0;
+
+err_release_channel:
+	dma_release_channel(owl_host->dma);
+err_free_host:
+	mmc_free_host(mmc);
+
+	return ret;
+}
+
+static int owl_mmc_remove(struct platform_device *pdev)
+{
+	struct mmc_host	*mmc = platform_get_drvdata(pdev);
+	struct owl_mmc_host *owl_host = mmc_priv(mmc);
+
+	mmc_remove_host(mmc);
+	disable_irq(owl_host->irq);
+	dma_release_channel(owl_host->dma);
+	mmc_free_host(mmc);
+
+	return 0;
+}
+
+static const struct of_device_id owl_mmc_of_match[] = {
+	{.compatible = "actions,owl-mmc",},
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, owl_mmc_of_match);
+
+static struct platform_driver owl_mmc_driver = {
+	.driver = {
+		.name	= "owl_mmc",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = owl_mmc_of_match,
+	},
+	.probe		= owl_mmc_probe,
+	.remove		= owl_mmc_remove,
+};
+module_platform_driver(owl_mmc_driver);
+
+MODULE_DESCRIPTION("Actions Semi Owl SoCs SD/MMC Driver");
+MODULE_AUTHOR("Actions Semi");
+MODULE_AUTHOR("Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mmc/host/pxamci.c b/drivers/mmc/host/pxamci.c
new file mode 100644
index 000000000..e25e9bb34
--- /dev/null
+++ b/drivers/mmc/host/pxamci.c
@@ -0,0 +1,829 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ *  linux/drivers/mmc/host/pxa.c - PXA MMCI driver
+ *
+ *  Copyright (C) 2003 Russell King, All Rights Reserved.
+ *
+ *  This hardware is really sick:
+ *   - No way to clear interrupts.
+ *   - Have to turn off the clock whenever we touch the device.
+ *   - Doesn't tell you how many data blocks were transferred.
+ *  Yuck!
+ *
+ *	1 and 3 byte data transfers not supported
+ *	max block length up to 1023
+ */
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/ioport.h>
+#include <linux/platform_device.h>
+#include <linux/delay.h>
+#include <linux/interrupt.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/clk.h>
+#include <linux/err.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/io.h>
+#include <linux/regulator/consumer.h>
+#include <linux/gpio/consumer.h>
+#include <linux/gfp.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+
+#include <linux/sizes.h>
+
+#include <mach/hardware.h>
+#include <linux/platform_data/mmc-pxamci.h>
+
+#include "pxamci.h"
+
+#define DRIVER_NAME	"pxa2xx-mci"
+
+#define NR_SG	1
+#define CLKRT_OFF	(~0)
+
+#define mmc_has_26MHz()		(cpu_is_pxa300() || cpu_is_pxa310() \
+				|| cpu_is_pxa935())
+
+struct pxamci_host {
+	struct mmc_host		*mmc;
+	spinlock_t		lock;
+	struct resource		*res;
+	void __iomem		*base;
+	struct clk		*clk;
+	unsigned long		clkrate;
+	unsigned int		clkrt;
+	unsigned int		cmdat;
+	unsigned int		imask;
+	unsigned int		power_mode;
+	unsigned long		detect_delay_ms;
+	bool			use_ro_gpio;
+	struct gpio_desc	*power;
+	struct pxamci_platform_data *pdata;
+
+	struct mmc_request	*mrq;
+	struct mmc_command	*cmd;
+	struct mmc_data		*data;
+
+	struct dma_chan		*dma_chan_rx;
+	struct dma_chan		*dma_chan_tx;
+	dma_cookie_t		dma_cookie;
+	unsigned int		dma_len;
+	unsigned int		dma_dir;
+};
+
+static int pxamci_init_ocr(struct pxamci_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	int ret;
+
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret < 0)
+		return ret;
+
+	if (IS_ERR(mmc->supply.vmmc)) {
+		/* fall-back to platform data */
+		mmc->ocr_avail = host->pdata ?
+			host->pdata->ocr_mask :
+			MMC_VDD_32_33 | MMC_VDD_33_34;
+	}
+
+	return 0;
+}
+
+static inline int pxamci_set_power(struct pxamci_host *host,
+				    unsigned char power_mode,
+				    unsigned int vdd)
+{
+	struct mmc_host *mmc = host->mmc;
+	struct regulator *supply = mmc->supply.vmmc;
+
+	if (!IS_ERR(supply))
+		return mmc_regulator_set_ocr(mmc, supply, vdd);
+
+	if (host->power) {
+		bool on = !!((1 << vdd) & host->pdata->ocr_mask);
+		gpiod_set_value(host->power, on);
+	}
+
+	if (host->pdata && host->pdata->setpower)
+		return host->pdata->setpower(mmc_dev(host->mmc), vdd);
+
+	return 0;
+}
+
+static void pxamci_stop_clock(struct pxamci_host *host)
+{
+	if (readl(host->base + MMC_STAT) & STAT_CLK_EN) {
+		unsigned long timeout = 10000;
+		unsigned int v;
+
+		writel(STOP_CLOCK, host->base + MMC_STRPCL);
+
+		do {
+			v = readl(host->base + MMC_STAT);
+			if (!(v & STAT_CLK_EN))
+				break;
+			udelay(1);
+		} while (timeout--);
+
+		if (v & STAT_CLK_EN)
+			dev_err(mmc_dev(host->mmc), "unable to stop clock\n");
+	}
+}
+
+static void pxamci_enable_irq(struct pxamci_host *host, unsigned int mask)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	host->imask &= ~mask;
+	writel(host->imask, host->base + MMC_I_MASK);
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void pxamci_disable_irq(struct pxamci_host *host, unsigned int mask)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	host->imask |= mask;
+	writel(host->imask, host->base + MMC_I_MASK);
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void pxamci_dma_irq(void *param);
+
+static void pxamci_setup_data(struct pxamci_host *host, struct mmc_data *data)
+{
+	struct dma_async_tx_descriptor *tx;
+	enum dma_transfer_direction direction;
+	struct dma_slave_config	config;
+	struct dma_chan *chan;
+	unsigned int nob = data->blocks;
+	unsigned long long clks;
+	unsigned int timeout;
+	int ret;
+
+	host->data = data;
+
+	writel(nob, host->base + MMC_NOB);
+	writel(data->blksz, host->base + MMC_BLKLEN);
+
+	clks = (unsigned long long)data->timeout_ns * host->clkrate;
+	do_div(clks, 1000000000UL);
+	timeout = (unsigned int)clks + (data->timeout_clks << host->clkrt);
+	writel((timeout + 255) / 256, host->base + MMC_RDTO);
+
+	memset(&config, 0, sizeof(config));
+	config.src_addr_width = DMA_SLAVE_BUSWIDTH_1_BYTE;
+	config.dst_addr_width = DMA_SLAVE_BUSWIDTH_1_BYTE;
+	config.src_addr = host->res->start + MMC_RXFIFO;
+	config.dst_addr = host->res->start + MMC_TXFIFO;
+	config.src_maxburst = 32;
+	config.dst_maxburst = 32;
+
+	if (data->flags & MMC_DATA_READ) {
+		host->dma_dir = DMA_FROM_DEVICE;
+		direction = DMA_DEV_TO_MEM;
+		chan = host->dma_chan_rx;
+	} else {
+		host->dma_dir = DMA_TO_DEVICE;
+		direction = DMA_MEM_TO_DEV;
+		chan = host->dma_chan_tx;
+	}
+
+	config.direction = direction;
+
+	ret = dmaengine_slave_config(chan, &config);
+	if (ret < 0) {
+		dev_err(mmc_dev(host->mmc), "dma slave config failed\n");
+		return;
+	}
+
+	host->dma_len = dma_map_sg(chan->device->dev, data->sg, data->sg_len,
+				   host->dma_dir);
+
+	tx = dmaengine_prep_slave_sg(chan, data->sg, host->dma_len, direction,
+				     DMA_PREP_INTERRUPT);
+	if (!tx) {
+		dev_err(mmc_dev(host->mmc), "prep_slave_sg() failed\n");
+		return;
+	}
+
+	if (!(data->flags & MMC_DATA_READ)) {
+		tx->callback = pxamci_dma_irq;
+		tx->callback_param = host;
+	}
+
+	host->dma_cookie = dmaengine_submit(tx);
+
+	/*
+	 * workaround for erratum #91:
+	 * only start DMA now if we are doing a read,
+	 * otherwise we wait until CMD/RESP has finished
+	 * before starting DMA.
+	 */
+	if (!cpu_is_pxa27x() || data->flags & MMC_DATA_READ)
+		dma_async_issue_pending(chan);
+}
+
+static void pxamci_start_cmd(struct pxamci_host *host, struct mmc_command *cmd, unsigned int cmdat)
+{
+	WARN_ON(host->cmd != NULL);
+	host->cmd = cmd;
+
+	if (cmd->flags & MMC_RSP_BUSY)
+		cmdat |= CMDAT_BUSY;
+
+#define RSP_TYPE(x)	((x) & ~(MMC_RSP_BUSY|MMC_RSP_OPCODE))
+	switch (RSP_TYPE(mmc_resp_type(cmd))) {
+	case RSP_TYPE(MMC_RSP_R1): /* r1, r1b, r6, r7 */
+		cmdat |= CMDAT_RESP_SHORT;
+		break;
+	case RSP_TYPE(MMC_RSP_R3):
+		cmdat |= CMDAT_RESP_R3;
+		break;
+	case RSP_TYPE(MMC_RSP_R2):
+		cmdat |= CMDAT_RESP_R2;
+		break;
+	default:
+		break;
+	}
+
+	writel(cmd->opcode, host->base + MMC_CMD);
+	writel(cmd->arg >> 16, host->base + MMC_ARGH);
+	writel(cmd->arg & 0xffff, host->base + MMC_ARGL);
+	writel(cmdat, host->base + MMC_CMDAT);
+	writel(host->clkrt, host->base + MMC_CLKRT);
+
+	writel(START_CLOCK, host->base + MMC_STRPCL);
+
+	pxamci_enable_irq(host, END_CMD_RES);
+}
+
+static void pxamci_finish_request(struct pxamci_host *host, struct mmc_request *mrq)
+{
+	host->mrq = NULL;
+	host->cmd = NULL;
+	host->data = NULL;
+	mmc_request_done(host->mmc, mrq);
+}
+
+static int pxamci_cmd_done(struct pxamci_host *host, unsigned int stat)
+{
+	struct mmc_command *cmd = host->cmd;
+	int i;
+	u32 v;
+
+	if (!cmd)
+		return 0;
+
+	host->cmd = NULL;
+
+	/*
+	 * Did I mention this is Sick.  We always need to
+	 * discard the upper 8 bits of the first 16-bit word.
+	 */
+	v = readl(host->base + MMC_RES) & 0xffff;
+	for (i = 0; i < 4; i++) {
+		u32 w1 = readl(host->base + MMC_RES) & 0xffff;
+		u32 w2 = readl(host->base + MMC_RES) & 0xffff;
+		cmd->resp[i] = v << 24 | w1 << 8 | w2 >> 8;
+		v = w2;
+	}
+
+	if (stat & STAT_TIME_OUT_RESPONSE) {
+		cmd->error = -ETIMEDOUT;
+	} else if (stat & STAT_RES_CRC_ERR && cmd->flags & MMC_RSP_CRC) {
+		/*
+		 * workaround for erratum #42:
+		 * Intel PXA27x Family Processor Specification Update Rev 001
+		 * A bogus CRC error can appear if the msb of a 136 bit
+		 * response is a one.
+		 */
+		if (cpu_is_pxa27x() &&
+		    (cmd->flags & MMC_RSP_136 && cmd->resp[0] & 0x80000000))
+			pr_debug("ignoring CRC from command %d - *risky*\n", cmd->opcode);
+		else
+			cmd->error = -EILSEQ;
+	}
+
+	pxamci_disable_irq(host, END_CMD_RES);
+	if (host->data && !cmd->error) {
+		pxamci_enable_irq(host, DATA_TRAN_DONE);
+		/*
+		 * workaround for erratum #91, if doing write
+		 * enable DMA late
+		 */
+		if (cpu_is_pxa27x() && host->data->flags & MMC_DATA_WRITE)
+			dma_async_issue_pending(host->dma_chan_tx);
+	} else {
+		pxamci_finish_request(host, host->mrq);
+	}
+
+	return 1;
+}
+
+static int pxamci_data_done(struct pxamci_host *host, unsigned int stat)
+{
+	struct mmc_data *data = host->data;
+	struct dma_chan *chan;
+
+	if (!data)
+		return 0;
+
+	if (data->flags & MMC_DATA_READ)
+		chan = host->dma_chan_rx;
+	else
+		chan = host->dma_chan_tx;
+	dma_unmap_sg(chan->device->dev,
+		     data->sg, data->sg_len, host->dma_dir);
+
+	if (stat & STAT_READ_TIME_OUT)
+		data->error = -ETIMEDOUT;
+	else if (stat & (STAT_CRC_READ_ERROR|STAT_CRC_WRITE_ERROR))
+		data->error = -EILSEQ;
+
+	/*
+	 * There appears to be a hardware design bug here.  There seems to
+	 * be no way to find out how much data was transferred to the card.
+	 * This means that if there was an error on any block, we mark all
+	 * data blocks as being in error.
+	 */
+	if (!data->error)
+		data->bytes_xfered = data->blocks * data->blksz;
+	else
+		data->bytes_xfered = 0;
+
+	pxamci_disable_irq(host, DATA_TRAN_DONE);
+
+	host->data = NULL;
+	if (host->mrq->stop) {
+		pxamci_stop_clock(host);
+		pxamci_start_cmd(host, host->mrq->stop, host->cmdat);
+	} else {
+		pxamci_finish_request(host, host->mrq);
+	}
+
+	return 1;
+}
+
+static irqreturn_t pxamci_irq(int irq, void *devid)
+{
+	struct pxamci_host *host = devid;
+	unsigned int ireg;
+	int handled = 0;
+
+	ireg = readl(host->base + MMC_I_REG) & ~readl(host->base + MMC_I_MASK);
+
+	if (ireg) {
+		unsigned stat = readl(host->base + MMC_STAT);
+
+		pr_debug("PXAMCI: irq %08x stat %08x\n", ireg, stat);
+
+		if (ireg & END_CMD_RES)
+			handled |= pxamci_cmd_done(host, stat);
+		if (ireg & DATA_TRAN_DONE)
+			handled |= pxamci_data_done(host, stat);
+		if (ireg & SDIO_INT) {
+			mmc_signal_sdio_irq(host->mmc);
+			handled = 1;
+		}
+	}
+
+	return IRQ_RETVAL(handled);
+}
+
+static void pxamci_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct pxamci_host *host = mmc_priv(mmc);
+	unsigned int cmdat;
+
+	WARN_ON(host->mrq != NULL);
+
+	host->mrq = mrq;
+
+	pxamci_stop_clock(host);
+
+	cmdat = host->cmdat;
+	host->cmdat &= ~CMDAT_INIT;
+
+	if (mrq->data) {
+		pxamci_setup_data(host, mrq->data);
+
+		cmdat &= ~CMDAT_BUSY;
+		cmdat |= CMDAT_DATAEN | CMDAT_DMAEN;
+		if (mrq->data->flags & MMC_DATA_WRITE)
+			cmdat |= CMDAT_WRITE;
+	}
+
+	pxamci_start_cmd(host, mrq->cmd, cmdat);
+}
+
+static int pxamci_get_ro(struct mmc_host *mmc)
+{
+	struct pxamci_host *host = mmc_priv(mmc);
+
+	if (host->use_ro_gpio)
+		return mmc_gpio_get_ro(mmc);
+	if (host->pdata && host->pdata->get_ro)
+		return !!host->pdata->get_ro(mmc_dev(mmc));
+	/*
+	 * Board doesn't support read only detection; let the mmc core
+	 * decide what to do.
+	 */
+	return -ENOSYS;
+}
+
+static void pxamci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct pxamci_host *host = mmc_priv(mmc);
+
+	if (ios->clock) {
+		unsigned long rate = host->clkrate;
+		unsigned int clk = rate / ios->clock;
+
+		if (host->clkrt == CLKRT_OFF)
+			clk_prepare_enable(host->clk);
+
+		if (ios->clock == 26000000) {
+			/* to support 26MHz */
+			host->clkrt = 7;
+		} else {
+			/* to handle (19.5MHz, 26MHz) */
+			if (!clk)
+				clk = 1;
+
+			/*
+			 * clk might result in a lower divisor than we
+			 * desire.  check for that condition and adjust
+			 * as appropriate.
+			 */
+			if (rate / clk > ios->clock)
+				clk <<= 1;
+			host->clkrt = fls(clk) - 1;
+		}
+
+		/*
+		 * we write clkrt on the next command
+		 */
+	} else {
+		pxamci_stop_clock(host);
+		if (host->clkrt != CLKRT_OFF) {
+			host->clkrt = CLKRT_OFF;
+			clk_disable_unprepare(host->clk);
+		}
+	}
+
+	if (host->power_mode != ios->power_mode) {
+		int ret;
+
+		host->power_mode = ios->power_mode;
+
+		ret = pxamci_set_power(host, ios->power_mode, ios->vdd);
+		if (ret) {
+			dev_err(mmc_dev(mmc), "unable to set power\n");
+			/*
+			 * The .set_ios() function in the mmc_host_ops
+			 * struct return void, and failing to set the
+			 * power should be rare so we print an error and
+			 * return here.
+			 */
+			return;
+		}
+
+		if (ios->power_mode == MMC_POWER_ON)
+			host->cmdat |= CMDAT_INIT;
+	}
+
+	if (ios->bus_width == MMC_BUS_WIDTH_4)
+		host->cmdat |= CMDAT_SD_4DAT;
+	else
+		host->cmdat &= ~CMDAT_SD_4DAT;
+
+	dev_dbg(mmc_dev(mmc), "PXAMCI: clkrt = %x cmdat = %x\n",
+		host->clkrt, host->cmdat);
+}
+
+static void pxamci_enable_sdio_irq(struct mmc_host *host, int enable)
+{
+	struct pxamci_host *pxa_host = mmc_priv(host);
+
+	if (enable)
+		pxamci_enable_irq(pxa_host, SDIO_INT);
+	else
+		pxamci_disable_irq(pxa_host, SDIO_INT);
+}
+
+static const struct mmc_host_ops pxamci_ops = {
+	.request		= pxamci_request,
+	.get_cd			= mmc_gpio_get_cd,
+	.get_ro			= pxamci_get_ro,
+	.set_ios		= pxamci_set_ios,
+	.enable_sdio_irq	= pxamci_enable_sdio_irq,
+};
+
+static void pxamci_dma_irq(void *param)
+{
+	struct pxamci_host *host = param;
+	struct dma_tx_state state;
+	enum dma_status status;
+	struct dma_chan *chan;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (!host->data)
+		goto out_unlock;
+
+	if (host->data->flags & MMC_DATA_READ)
+		chan = host->dma_chan_rx;
+	else
+		chan = host->dma_chan_tx;
+
+	status = dmaengine_tx_status(chan, host->dma_cookie, &state);
+
+	if (likely(status == DMA_COMPLETE)) {
+		writel(BUF_PART_FULL, host->base + MMC_PRTBUF);
+	} else {
+		pr_err("%s: DMA error on %s channel\n", mmc_hostname(host->mmc),
+			host->data->flags & MMC_DATA_READ ? "rx" : "tx");
+		host->data->error = -EIO;
+		pxamci_data_done(host, 0);
+	}
+
+out_unlock:
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static irqreturn_t pxamci_detect_irq(int irq, void *devid)
+{
+	struct pxamci_host *host = mmc_priv(devid);
+
+	mmc_detect_change(devid, msecs_to_jiffies(host->detect_delay_ms));
+	return IRQ_HANDLED;
+}
+
+#ifdef CONFIG_OF
+static const struct of_device_id pxa_mmc_dt_ids[] = {
+        { .compatible = "marvell,pxa-mmc" },
+        { }
+};
+
+MODULE_DEVICE_TABLE(of, pxa_mmc_dt_ids);
+
+static int pxamci_of_init(struct platform_device *pdev,
+			  struct mmc_host *mmc)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct pxamci_host *host = mmc_priv(mmc);
+	u32 tmp;
+	int ret;
+
+	if (!np)
+		return 0;
+
+	/* pxa-mmc specific */
+	if (of_property_read_u32(np, "pxa-mmc,detect-delay-ms", &tmp) == 0)
+		host->detect_delay_ms = tmp;
+
+	ret = mmc_of_parse(mmc);
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
+#else
+static int pxamci_of_init(struct platform_device *pdev,
+			  struct mmc_host *mmc)
+{
+        return 0;
+}
+#endif
+
+static int pxamci_probe(struct platform_device *pdev)
+{
+	struct mmc_host *mmc;
+	struct pxamci_host *host = NULL;
+	struct device *dev = &pdev->dev;
+	struct resource *r;
+	int ret, irq;
+
+	r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return irq;
+
+	mmc = mmc_alloc_host(sizeof(struct pxamci_host), dev);
+	if (!mmc) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	mmc->ops = &pxamci_ops;
+
+	/*
+	 * We can do SG-DMA, but we don't because we never know how much
+	 * data we successfully wrote to the card.
+	 */
+	mmc->max_segs = NR_SG;
+
+	/*
+	 * Our hardware DMA can handle a maximum of one page per SG entry.
+	 */
+	mmc->max_seg_size = PAGE_SIZE;
+
+	/*
+	 * Block length register is only 10 bits before PXA27x.
+	 */
+	mmc->max_blk_size = cpu_is_pxa25x() ? 1023 : 2048;
+
+	/*
+	 * Block count register is 16 bits.
+	 */
+	mmc->max_blk_count = 65535;
+
+	ret = pxamci_of_init(pdev, mmc);
+	if (ret)
+		goto out;
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+	host->pdata = pdev->dev.platform_data;
+	host->clkrt = CLKRT_OFF;
+
+	host->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(host->clk)) {
+		ret = PTR_ERR(host->clk);
+		host->clk = NULL;
+		goto out;
+	}
+
+	host->clkrate = clk_get_rate(host->clk);
+
+	/*
+	 * Calculate minimum clock rate, rounding up.
+	 */
+	mmc->f_min = (host->clkrate + 63) / 64;
+	mmc->f_max = (mmc_has_26MHz()) ? 26000000 : host->clkrate;
+
+	ret = pxamci_init_ocr(host);
+	if (ret < 0)
+		goto out;
+
+	mmc->caps = 0;
+	host->cmdat = 0;
+	if (!cpu_is_pxa25x()) {
+		mmc->caps |= MMC_CAP_4_BIT_DATA | MMC_CAP_SDIO_IRQ;
+		host->cmdat |= CMDAT_SDIO_INT_EN;
+		if (mmc_has_26MHz())
+			mmc->caps |= MMC_CAP_MMC_HIGHSPEED |
+				     MMC_CAP_SD_HIGHSPEED;
+	}
+
+	spin_lock_init(&host->lock);
+	host->res = r;
+	host->imask = MMC_I_MASK_ALL;
+
+	host->base = devm_ioremap_resource(dev, r);
+	if (IS_ERR(host->base)) {
+		ret = PTR_ERR(host->base);
+		goto out;
+	}
+
+	/*
+	 * Ensure that the host controller is shut down, and setup
+	 * with our defaults.
+	 */
+	pxamci_stop_clock(host);
+	writel(0, host->base + MMC_SPI);
+	writel(64, host->base + MMC_RESTO);
+	writel(host->imask, host->base + MMC_I_MASK);
+
+	ret = devm_request_irq(dev, irq, pxamci_irq, 0,
+			       DRIVER_NAME, host);
+	if (ret)
+		goto out;
+
+	platform_set_drvdata(pdev, mmc);
+
+	host->dma_chan_rx = dma_request_chan(dev, "rx");
+	if (IS_ERR(host->dma_chan_rx)) {
+		dev_err(dev, "unable to request rx dma channel\n");
+		ret = PTR_ERR(host->dma_chan_rx);
+		host->dma_chan_rx = NULL;
+		goto out;
+	}
+
+	host->dma_chan_tx = dma_request_chan(dev, "tx");
+	if (IS_ERR(host->dma_chan_tx)) {
+		dev_err(dev, "unable to request tx dma channel\n");
+		ret = PTR_ERR(host->dma_chan_tx);
+		host->dma_chan_tx = NULL;
+		goto out;
+	}
+
+	if (host->pdata) {
+		host->detect_delay_ms = host->pdata->detect_delay_ms;
+
+		host->power = devm_gpiod_get_optional(dev, "power", GPIOD_OUT_LOW);
+		if (IS_ERR(host->power)) {
+			ret = PTR_ERR(host->power);
+			dev_err(dev, "Failed requesting gpio_power\n");
+			goto out;
+		}
+
+		/* FIXME: should we pass detection delay to debounce? */
+		ret = mmc_gpiod_request_cd(mmc, "cd", 0, false, 0);
+		if (ret && ret != -ENOENT) {
+			dev_err(dev, "Failed requesting gpio_cd\n");
+			goto out;
+		}
+
+		if (!host->pdata->gpio_card_ro_invert)
+			mmc->caps2 |= MMC_CAP2_RO_ACTIVE_HIGH;
+
+		ret = mmc_gpiod_request_ro(mmc, "wp", 0, 0);
+		if (ret && ret != -ENOENT) {
+			dev_err(dev, "Failed requesting gpio_ro\n");
+			goto out;
+		}
+		if (!ret)
+			host->use_ro_gpio = true;
+
+		if (host->pdata->init)
+			host->pdata->init(dev, pxamci_detect_irq, mmc);
+
+		if (host->power && host->pdata->setpower)
+			dev_warn(dev, "gpio_power and setpower() both defined\n");
+		if (host->use_ro_gpio && host->pdata->get_ro)
+			dev_warn(dev, "gpio_ro and get_ro() both defined\n");
+	}
+
+	ret = mmc_add_host(mmc);
+	if (ret) {
+		if (host->pdata && host->pdata->exit)
+			host->pdata->exit(dev, mmc);
+		goto out;
+	}
+
+	return 0;
+
+out:
+	if (host) {
+		if (host->dma_chan_rx)
+			dma_release_channel(host->dma_chan_rx);
+		if (host->dma_chan_tx)
+			dma_release_channel(host->dma_chan_tx);
+	}
+	if (mmc)
+		mmc_free_host(mmc);
+	return ret;
+}
+
+static int pxamci_remove(struct platform_device *pdev)
+{
+	struct mmc_host *mmc = platform_get_drvdata(pdev);
+
+	if (mmc) {
+		struct pxamci_host *host = mmc_priv(mmc);
+
+		mmc_remove_host(mmc);
+
+		if (host->pdata && host->pdata->exit)
+			host->pdata->exit(&pdev->dev, mmc);
+
+		pxamci_stop_clock(host);
+		writel(TXFIFO_WR_REQ|RXFIFO_RD_REQ|CLK_IS_OFF|STOP_CMD|
+		       END_CMD_RES|PRG_DONE|DATA_TRAN_DONE,
+		       host->base + MMC_I_MASK);
+
+		dmaengine_terminate_all(host->dma_chan_rx);
+		dmaengine_terminate_all(host->dma_chan_tx);
+		dma_release_channel(host->dma_chan_rx);
+		dma_release_channel(host->dma_chan_tx);
+
+		mmc_free_host(mmc);
+	}
+
+	return 0;
+}
+
+static struct platform_driver pxamci_driver = {
+	.probe		= pxamci_probe,
+	.remove		= pxamci_remove,
+	.driver		= {
+		.name	= DRIVER_NAME,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(pxa_mmc_dt_ids),
+	},
+};
+
+module_platform_driver(pxamci_driver);
+
+MODULE_DESCRIPTION("PXA Multimedia Card Interface Driver");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:pxa2xx-mci");
diff --git a/drivers/mmc/host/pxamci.h b/drivers/mmc/host/pxamci.h
new file mode 100644
index 000000000..d301ca18c
--- /dev/null
+++ b/drivers/mmc/host/pxamci.h
@@ -0,0 +1,91 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#define MMC_STRPCL	0x0000
+#define STOP_CLOCK		(1 << 0)
+#define START_CLOCK		(2 << 0)
+
+#define MMC_STAT	0x0004
+#define STAT_END_CMD_RES		(1 << 13)
+#define STAT_PRG_DONE			(1 << 12)
+#define STAT_DATA_TRAN_DONE		(1 << 11)
+#define STAT_CLK_EN			(1 << 8)
+#define STAT_RECV_FIFO_FULL		(1 << 7)
+#define STAT_XMIT_FIFO_EMPTY		(1 << 6)
+#define STAT_RES_CRC_ERR		(1 << 5)
+#define STAT_SPI_READ_ERROR_TOKEN	(1 << 4)
+#define STAT_CRC_READ_ERROR		(1 << 3)
+#define STAT_CRC_WRITE_ERROR		(1 << 2)
+#define STAT_TIME_OUT_RESPONSE		(1 << 1)
+#define STAT_READ_TIME_OUT		(1 << 0)
+
+#define MMC_CLKRT	0x0008		/* 3 bit */
+
+#define MMC_SPI		0x000c
+#define SPI_CS_ADDRESS		(1 << 3)
+#define SPI_CS_EN		(1 << 2)
+#define CRC_ON			(1 << 1)
+#define SPI_EN			(1 << 0)
+
+#define MMC_CMDAT	0x0010
+#define CMDAT_SDIO_INT_EN	(1 << 11)
+#define CMDAT_SD_4DAT		(1 << 8)
+#define CMDAT_DMAEN		(1 << 7)
+#define CMDAT_INIT		(1 << 6)
+#define CMDAT_BUSY		(1 << 5)
+#define CMDAT_STREAM		(1 << 4)	/* 1 = stream */
+#define CMDAT_WRITE		(1 << 3)	/* 1 = write */
+#define CMDAT_DATAEN		(1 << 2)
+#define CMDAT_RESP_NONE		(0 << 0)
+#define CMDAT_RESP_SHORT	(1 << 0)
+#define CMDAT_RESP_R2		(2 << 0)
+#define CMDAT_RESP_R3		(3 << 0)
+
+#define MMC_RESTO	0x0014	/* 7 bit */
+
+#define MMC_RDTO	0x0018	/* 16 bit */
+
+#define MMC_BLKLEN	0x001c	/* 10 bit */
+
+#define MMC_NOB		0x0020	/* 16 bit */
+
+#define MMC_PRTBUF	0x0024
+#define BUF_PART_FULL		(1 << 0)
+
+#define MMC_I_MASK	0x0028
+
+/*PXA27x MMC interrupts*/
+#define SDIO_SUSPEND_ACK  	(1 << 12)
+#define SDIO_INT          	(1 << 11)
+#define RD_STALLED        	(1 << 10)
+#define RES_ERR           	(1 << 9)
+#define DAT_ERR           	(1 << 8)
+#define TINT              	(1 << 7)
+
+/*PXA2xx MMC interrupts*/
+#define TXFIFO_WR_REQ		(1 << 6)
+#define RXFIFO_RD_REQ		(1 << 5)
+#define CLK_IS_OFF		(1 << 4)
+#define STOP_CMD		(1 << 3)
+#define END_CMD_RES		(1 << 2)
+#define PRG_DONE		(1 << 1)
+#define DATA_TRAN_DONE		(1 << 0)
+
+#if defined(CONFIG_PXA27x) || defined(CONFIG_PXA3xx)
+#define MMC_I_MASK_ALL          0x00001fff
+#else
+#define MMC_I_MASK_ALL          0x0000007f
+#endif
+
+#define MMC_I_REG	0x002c
+/* same as MMC_I_MASK */
+
+#define MMC_CMD		0x0030
+
+#define MMC_ARGH	0x0034	/* 16 bit */
+
+#define MMC_ARGL	0x0038	/* 16 bit */
+
+#define MMC_RES		0x003c	/* 16 bit */
+
+#define MMC_RXFIFO	0x0040	/* 8 bit */
+
+#define MMC_TXFIFO	0x0044	/* 8 bit */
diff --git a/drivers/mmc/host/renesas_sdhi.h b/drivers/mmc/host/renesas_sdhi.h
new file mode 100644
index 000000000..cb962c788
--- /dev/null
+++ b/drivers/mmc/host/renesas_sdhi.h
@@ -0,0 +1,81 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Renesas Mobile SDHI
+ *
+ * Copyright (C) 2017 Horms Solutions Ltd., Simon Horman
+ * Copyright (C) 2017-19 Renesas Electronics Corporation
+ */
+
+#ifndef RENESAS_SDHI_H
+#define RENESAS_SDHI_H
+
+#include <linux/platform_device.h>
+#include "tmio_mmc.h"
+
+struct renesas_sdhi_scc {
+	unsigned long clk_rate;	/* clock rate for SDR104 */
+	u32 tap;		/* sampling clock position for SDR104/HS400 (8 TAP) */
+	u32 tap_hs400_4tap;	/* sampling clock position for HS400 (4 TAP) */
+};
+
+struct renesas_sdhi_of_data {
+	unsigned long tmio_flags;
+	u32	      tmio_ocr_mask;
+	unsigned long capabilities;
+	unsigned long capabilities2;
+	enum dma_slave_buswidth dma_buswidth;
+	dma_addr_t dma_rx_offset;
+	unsigned int bus_shift;
+	int scc_offset;
+	struct renesas_sdhi_scc *taps;
+	int taps_num;
+	unsigned int max_blk_count;
+	unsigned short max_segs;
+};
+
+#define SDHI_CALIB_TABLE_MAX 32
+
+struct renesas_sdhi_quirks {
+	bool hs400_disabled;
+	bool hs400_4taps;
+	u32 hs400_bad_taps;
+	const u8 (*hs400_calib_table)[SDHI_CALIB_TABLE_MAX];
+};
+
+struct tmio_mmc_dma {
+	enum dma_slave_buswidth dma_buswidth;
+	bool (*filter)(struct dma_chan *chan, void *arg);
+	void (*enable)(struct tmio_mmc_host *host, bool enable);
+	struct completion	dma_dataend;
+	struct tasklet_struct	dma_complete;
+};
+
+struct renesas_sdhi {
+	struct clk *clk;
+	struct clk *clk_cd;
+	struct tmio_mmc_data mmc_data;
+	struct tmio_mmc_dma dma_priv;
+	const struct renesas_sdhi_quirks *quirks;
+	struct pinctrl *pinctrl;
+	struct pinctrl_state *pins_default, *pins_uhs;
+	void __iomem *scc_ctl;
+	u32 scc_tappos;
+	u32 scc_tappos_hs400;
+	const u8 *adjust_hs400_calib_table;
+	bool needs_adjust_hs400;
+
+	/* Tuning values: 1 for success, 0 for failure */
+	DECLARE_BITMAP(taps, BITS_PER_LONG);
+	/* Sampling data comparison: 1 for match, 0 for mismatch */
+	DECLARE_BITMAP(smpcmp, BITS_PER_LONG);
+	unsigned int tap_num;
+	unsigned int tap_set;
+};
+
+#define host_to_priv(host) \
+	container_of((host)->pdata, struct renesas_sdhi, mmc_data)
+
+int renesas_sdhi_probe(struct platform_device *pdev,
+		       const struct tmio_mmc_dma_ops *dma_ops);
+int renesas_sdhi_remove(struct platform_device *pdev);
+#endif
diff --git a/drivers/mmc/host/renesas_sdhi_core.c b/drivers/mmc/host/renesas_sdhi_core.c
new file mode 100644
index 000000000..24e524a1b
--- /dev/null
+++ b/drivers/mmc/host/renesas_sdhi_core.c
@@ -0,0 +1,1165 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Renesas SDHI
+ *
+ * Copyright (C) 2015-19 Renesas Electronics Corporation
+ * Copyright (C) 2016-19 Sang Engineering, Wolfram Sang
+ * Copyright (C) 2016-17 Horms Solutions, Simon Horman
+ * Copyright (C) 2009 Magnus Damm
+ *
+ * Based on "Compaq ASIC3 support":
+ *
+ * Copyright 2001 Compaq Computer Corporation.
+ * Copyright 2004-2005 Phil Blundell
+ * Copyright 2007-2008 OpenedHand Ltd.
+ *
+ * Authors: Phil Blundell <pb@handhelds.org>,
+ *	    Samuel Ortiz <sameo@openedhand.com>
+ *
+ */
+
+#include <linux/kernel.h>
+#include <linux/clk.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <linux/pm_domain.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/mfd/tmio.h>
+#include <linux/sh_dma.h>
+#include <linux/delay.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/pinctrl/pinctrl-state.h>
+#include <linux/regulator/consumer.h>
+#include <linux/sys_soc.h>
+
+#include "renesas_sdhi.h"
+#include "tmio_mmc.h"
+
+#define HOST_MODE		0xe4
+
+#define SDHI_VER_GEN2_SDR50	0x490c
+#define SDHI_VER_RZ_A1		0x820b
+/* very old datasheets said 0x490c for SDR104, too. They are wrong! */
+#define SDHI_VER_GEN2_SDR104	0xcb0d
+#define SDHI_VER_GEN3_SD	0xcc10
+#define SDHI_VER_GEN3_SDMMC	0xcd10
+
+#define SDHI_GEN3_MMC0_ADDR	0xee140000
+
+static void renesas_sdhi_sdbuf_width(struct tmio_mmc_host *host, int width)
+{
+	u32 val;
+
+	/*
+	 * see also
+	 *	renesas_sdhi_of_data :: dma_buswidth
+	 */
+	switch (sd_ctrl_read16(host, CTL_VERSION)) {
+	case SDHI_VER_GEN2_SDR50:
+		val = (width == 32) ? 0x0001 : 0x0000;
+		break;
+	case SDHI_VER_GEN2_SDR104:
+		val = (width == 32) ? 0x0000 : 0x0001;
+		break;
+	case SDHI_VER_GEN3_SD:
+	case SDHI_VER_GEN3_SDMMC:
+		if (width == 64)
+			val = 0x0000;
+		else if (width == 32)
+			val = 0x0101;
+		else
+			val = 0x0001;
+		break;
+	default:
+		/* nothing to do */
+		return;
+	}
+
+	sd_ctrl_write16(host, HOST_MODE, val);
+}
+
+static int renesas_sdhi_clk_enable(struct tmio_mmc_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	struct renesas_sdhi *priv = host_to_priv(host);
+	int ret;
+
+	ret = clk_prepare_enable(priv->clk_cd);
+	if (ret < 0)
+		return ret;
+
+	/*
+	 * The clock driver may not know what maximum frequency
+	 * actually works, so it should be set with the max-frequency
+	 * property which will already have been read to f_max.  If it
+	 * was missing, assume the current frequency is the maximum.
+	 */
+	if (!mmc->f_max)
+		mmc->f_max = clk_get_rate(priv->clk);
+
+	/*
+	 * Minimum frequency is the minimum input clock frequency
+	 * divided by our maximum divider.
+	 */
+	mmc->f_min = max(clk_round_rate(priv->clk, 1) / 512, 1L);
+
+	/* enable 16bit data access on SDBUF as default */
+	renesas_sdhi_sdbuf_width(host, 16);
+
+	return 0;
+}
+
+static unsigned int renesas_sdhi_clk_update(struct tmio_mmc_host *host,
+					    unsigned int new_clock)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+	unsigned int freq, diff, best_freq = 0, diff_min = ~0;
+	int i;
+
+	/*
+	 * We simply return the current rate if a) we are not on a R-Car Gen2+
+	 * SoC (may work for others, but untested) or b) if the SCC needs its
+	 * clock during tuning, so we don't change the external clock setup.
+	 */
+	if (!(host->pdata->flags & TMIO_MMC_MIN_RCAR2) || mmc_doing_tune(host->mmc))
+		return clk_get_rate(priv->clk);
+
+	/*
+	 * We want the bus clock to be as close as possible to, but no
+	 * greater than, new_clock.  As we can divide by 1 << i for
+	 * any i in [0, 9] we want the input clock to be as close as
+	 * possible, but no greater than, new_clock << i.
+	 */
+	for (i = min(9, ilog2(UINT_MAX / new_clock)); i >= 0; i--) {
+		freq = clk_round_rate(priv->clk, new_clock << i);
+		if (freq > (new_clock << i)) {
+			/* Too fast; look for a slightly slower option */
+			freq = clk_round_rate(priv->clk,
+					      (new_clock << i) / 4 * 3);
+			if (freq > (new_clock << i))
+				continue;
+		}
+
+		diff = new_clock - (freq >> i);
+		if (diff <= diff_min) {
+			best_freq = freq;
+			diff_min = diff;
+		}
+	}
+
+	clk_set_rate(priv->clk, best_freq);
+
+	return clk_get_rate(priv->clk);
+}
+
+static void renesas_sdhi_set_clock(struct tmio_mmc_host *host,
+				   unsigned int new_clock)
+{
+	u32 clk = 0, clock;
+
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, ~CLK_CTL_SCLKEN &
+		sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+
+	if (new_clock == 0) {
+		host->mmc->actual_clock = 0;
+		goto out;
+	}
+
+	host->mmc->actual_clock = renesas_sdhi_clk_update(host, new_clock);
+	clock = host->mmc->actual_clock / 512;
+
+	for (clk = 0x80000080; new_clock >= (clock << 1); clk >>= 1)
+		clock <<= 1;
+
+	/* 1/1 clock is option */
+	if ((host->pdata->flags & TMIO_MMC_CLK_ACTUAL) && ((clk >> 22) & 0x1)) {
+		if (!(host->mmc->ios.timing == MMC_TIMING_MMC_HS400))
+			clk |= 0xff;
+		else
+			clk &= ~0xff;
+	}
+
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, clk & CLK_CTL_DIV_MASK);
+	if (!(host->pdata->flags & TMIO_MMC_MIN_RCAR2))
+		usleep_range(10000, 11000);
+
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, CLK_CTL_SCLKEN |
+		sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+
+out:
+	/* HW engineers overrode docs: no sleep needed on R-Car2+ */
+	if (!(host->pdata->flags & TMIO_MMC_MIN_RCAR2))
+		usleep_range(10000, 11000);
+}
+
+static void renesas_sdhi_clk_disable(struct tmio_mmc_host *host)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+
+	clk_disable_unprepare(priv->clk_cd);
+}
+
+static int renesas_sdhi_card_busy(struct mmc_host *mmc)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+
+	return !(sd_ctrl_read16_and_16_as_32(host, CTL_STATUS) &
+		 TMIO_STAT_DAT0);
+}
+
+static int renesas_sdhi_start_signal_voltage_switch(struct mmc_host *mmc,
+						    struct mmc_ios *ios)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+	struct renesas_sdhi *priv = host_to_priv(host);
+	struct pinctrl_state *pin_state;
+	int ret;
+
+	switch (ios->signal_voltage) {
+	case MMC_SIGNAL_VOLTAGE_330:
+		pin_state = priv->pins_default;
+		break;
+	case MMC_SIGNAL_VOLTAGE_180:
+		pin_state = priv->pins_uhs;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	/*
+	 * If anything is missing, assume signal voltage is fixed at
+	 * 3.3V and succeed/fail accordingly.
+	 */
+	if (IS_ERR(priv->pinctrl) || IS_ERR(pin_state))
+		return ios->signal_voltage ==
+			MMC_SIGNAL_VOLTAGE_330 ? 0 : -EINVAL;
+
+	ret = mmc_regulator_set_vqmmc(host->mmc, ios);
+	if (ret < 0)
+		return ret;
+
+	return pinctrl_select_state(priv->pinctrl, pin_state);
+}
+
+/* SCC registers */
+#define SH_MOBILE_SDHI_SCC_DTCNTL	0x000
+#define SH_MOBILE_SDHI_SCC_TAPSET	0x002
+#define SH_MOBILE_SDHI_SCC_DT2FF	0x004
+#define SH_MOBILE_SDHI_SCC_CKSEL	0x006
+#define SH_MOBILE_SDHI_SCC_RVSCNTL	0x008
+#define SH_MOBILE_SDHI_SCC_RVSREQ	0x00A
+#define SH_MOBILE_SDHI_SCC_SMPCMP       0x00C
+#define SH_MOBILE_SDHI_SCC_TMPPORT2	0x00E
+#define SH_MOBILE_SDHI_SCC_TMPPORT3	0x014
+#define SH_MOBILE_SDHI_SCC_TMPPORT4	0x016
+#define SH_MOBILE_SDHI_SCC_TMPPORT5	0x018
+#define SH_MOBILE_SDHI_SCC_TMPPORT6	0x01A
+#define SH_MOBILE_SDHI_SCC_TMPPORT7	0x01C
+
+#define SH_MOBILE_SDHI_SCC_DTCNTL_TAPEN		BIT(0)
+#define SH_MOBILE_SDHI_SCC_DTCNTL_TAPNUM_SHIFT	16
+#define SH_MOBILE_SDHI_SCC_DTCNTL_TAPNUM_MASK	0xff
+
+#define SH_MOBILE_SDHI_SCC_CKSEL_DTSEL		BIT(0)
+
+#define SH_MOBILE_SDHI_SCC_RVSCNTL_RVSEN	BIT(0)
+
+#define SH_MOBILE_SDHI_SCC_RVSREQ_REQTAPDOWN	BIT(0)
+#define SH_MOBILE_SDHI_SCC_RVSREQ_REQTAPUP	BIT(1)
+#define SH_MOBILE_SDHI_SCC_RVSREQ_RVSERR	BIT(2)
+
+#define SH_MOBILE_SDHI_SCC_SMPCMP_CMD_REQDOWN	BIT(8)
+#define SH_MOBILE_SDHI_SCC_SMPCMP_CMD_REQUP	BIT(24)
+#define SH_MOBILE_SDHI_SCC_SMPCMP_CMD_ERR	(BIT(8) | BIT(24))
+
+#define SH_MOBILE_SDHI_SCC_TMPPORT2_HS400OSEL	BIT(4)
+#define SH_MOBILE_SDHI_SCC_TMPPORT2_HS400EN	BIT(31)
+
+/* Definitions for values the SH_MOBILE_SDHI_SCC_TMPPORT4 register */
+#define SH_MOBILE_SDHI_SCC_TMPPORT4_DLL_ACC_START	BIT(0)
+
+/* Definitions for values the SH_MOBILE_SDHI_SCC_TMPPORT5 register */
+#define SH_MOBILE_SDHI_SCC_TMPPORT5_DLL_RW_SEL_R	BIT(8)
+#define SH_MOBILE_SDHI_SCC_TMPPORT5_DLL_RW_SEL_W	(0 << 8)
+#define SH_MOBILE_SDHI_SCC_TMPPORT5_DLL_ADR_MASK	0x3F
+
+/* Definitions for values the SH_MOBILE_SDHI_SCC register */
+#define SH_MOBILE_SDHI_SCC_TMPPORT_DISABLE_WP_CODE	0xa5000000
+#define SH_MOBILE_SDHI_SCC_TMPPORT_CALIB_CODE_MASK	0x1f
+#define SH_MOBILE_SDHI_SCC_TMPPORT_MANUAL_MODE		BIT(7)
+
+static const u8 r8a7796_es13_calib_table[2][SDHI_CALIB_TABLE_MAX] = {
+	{ 3,  3,  3,  3,  3,  3,  3,  4,  4,  5,  6,  7,  8,  9, 10, 15,
+	 16, 16, 16, 16, 16, 16, 17, 18, 18, 19, 20, 21, 22, 23, 24, 25 },
+	{ 5,  5,  5,  5,  5,  5,  5,  5,  5,  5,  5,  5,  6,  7,  8, 11,
+	 12, 17, 18, 18, 18, 18, 18, 18, 18, 19, 20, 21, 22, 23, 25, 25 }
+};
+
+static const u8 r8a77965_calib_table[2][SDHI_CALIB_TABLE_MAX] = {
+	{ 1,  2,  6,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 15, 15, 16,
+	 17, 18, 19, 20, 21, 22, 23, 24, 25, 25, 26, 27, 28, 29, 30, 31 },
+	{ 2,  3,  4,  4,  5,  6,  7,  9, 10, 11, 12, 13, 14, 15, 16, 17,
+	 17, 17, 20, 21, 22, 23, 24, 25, 27, 28, 29, 30, 31, 31, 31, 31 }
+};
+
+static const u8 r8a77990_calib_table[2][SDHI_CALIB_TABLE_MAX] = {
+	{ 0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
+	  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0 },
+	{ 0,  0,  0,  1,  2,  3,  3,  4,  4,  4,  5,  5,  6,  8,  9, 10,
+	 11, 12, 13, 15, 16, 17, 17, 18, 18, 19, 20, 22, 24, 25, 26, 26 }
+};
+
+static inline u32 sd_scc_read32(struct tmio_mmc_host *host,
+				struct renesas_sdhi *priv, int addr)
+{
+	return readl(priv->scc_ctl + (addr << host->bus_shift));
+}
+
+static inline void sd_scc_write32(struct tmio_mmc_host *host,
+				  struct renesas_sdhi *priv,
+				  int addr, u32 val)
+{
+	writel(val, priv->scc_ctl + (addr << host->bus_shift));
+}
+
+static unsigned int renesas_sdhi_init_tuning(struct tmio_mmc_host *host)
+{
+	struct renesas_sdhi *priv;
+
+	priv = host_to_priv(host);
+
+	/* Initialize SCC */
+	sd_ctrl_write32_as_16_and_16(host, CTL_STATUS, 0x0);
+
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, ~CLK_CTL_SCLKEN &
+			sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+
+	/* set sampling clock selection range */
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_DTCNTL,
+		       SH_MOBILE_SDHI_SCC_DTCNTL_TAPEN |
+		       0x8 << SH_MOBILE_SDHI_SCC_DTCNTL_TAPNUM_SHIFT);
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_CKSEL,
+		       SH_MOBILE_SDHI_SCC_CKSEL_DTSEL |
+		       sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_CKSEL));
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_RVSCNTL,
+		       ~SH_MOBILE_SDHI_SCC_RVSCNTL_RVSEN &
+		       sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_RVSCNTL));
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_DT2FF, priv->scc_tappos);
+
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, CLK_CTL_SCLKEN |
+			sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+
+	/* Read TAPNUM */
+	return (sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_DTCNTL) >>
+		SH_MOBILE_SDHI_SCC_DTCNTL_TAPNUM_SHIFT) &
+		SH_MOBILE_SDHI_SCC_DTCNTL_TAPNUM_MASK;
+}
+
+static void renesas_sdhi_hs400_complete(struct mmc_host *mmc)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+	struct renesas_sdhi *priv = host_to_priv(host);
+	u32 bad_taps = priv->quirks ? priv->quirks->hs400_bad_taps : 0;
+	bool use_4tap = priv->quirks && priv->quirks->hs400_4taps;
+
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, ~CLK_CTL_SCLKEN &
+		sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+
+	/* Set HS400 mode */
+	sd_ctrl_write16(host, CTL_SDIF_MODE, 0x0001 |
+			sd_ctrl_read16(host, CTL_SDIF_MODE));
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_DT2FF,
+		       priv->scc_tappos_hs400);
+
+	/* Gen3 can't do automatic tap correction with HS400, so disable it */
+	if (sd_ctrl_read16(host, CTL_VERSION) == SDHI_VER_GEN3_SDMMC)
+		sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_RVSCNTL,
+			       ~SH_MOBILE_SDHI_SCC_RVSCNTL_RVSEN &
+			       sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_RVSCNTL));
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT2,
+		       (SH_MOBILE_SDHI_SCC_TMPPORT2_HS400EN |
+			SH_MOBILE_SDHI_SCC_TMPPORT2_HS400OSEL) |
+			sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT2));
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_DTCNTL,
+		       SH_MOBILE_SDHI_SCC_DTCNTL_TAPEN |
+		       sd_scc_read32(host, priv,
+				     SH_MOBILE_SDHI_SCC_DTCNTL));
+
+	/* Avoid bad TAP */
+	if (bad_taps & BIT(priv->tap_set)) {
+		u32 new_tap = (priv->tap_set + 1) % priv->tap_num;
+
+		if (bad_taps & BIT(new_tap))
+			new_tap = (priv->tap_set - 1) % priv->tap_num;
+
+		if (bad_taps & BIT(new_tap)) {
+			new_tap = priv->tap_set;
+			dev_dbg(&host->pdev->dev, "Can't handle three bad tap in a row\n");
+		}
+
+		priv->tap_set = new_tap;
+	}
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TAPSET,
+		       priv->tap_set / (use_4tap ? 2 : 1));
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_CKSEL,
+		       SH_MOBILE_SDHI_SCC_CKSEL_DTSEL |
+		       sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_CKSEL));
+
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, CLK_CTL_SCLKEN |
+			sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+
+	if (priv->adjust_hs400_calib_table)
+		priv->needs_adjust_hs400 = true;
+}
+
+static void renesas_sdhi_reset_scc(struct tmio_mmc_host *host,
+				   struct renesas_sdhi *priv)
+{
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, ~CLK_CTL_SCLKEN &
+			sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_CKSEL,
+		       ~SH_MOBILE_SDHI_SCC_CKSEL_DTSEL &
+		       sd_scc_read32(host, priv,
+				     SH_MOBILE_SDHI_SCC_CKSEL));
+}
+
+static void renesas_sdhi_disable_scc(struct mmc_host *mmc)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+	struct renesas_sdhi *priv = host_to_priv(host);
+
+	renesas_sdhi_reset_scc(host, priv);
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_DTCNTL,
+		       ~SH_MOBILE_SDHI_SCC_DTCNTL_TAPEN &
+		       sd_scc_read32(host, priv,
+				     SH_MOBILE_SDHI_SCC_DTCNTL));
+
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, CLK_CTL_SCLKEN |
+			sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+}
+
+static u32 sd_scc_tmpport_read32(struct tmio_mmc_host *host,
+				 struct renesas_sdhi *priv, u32 addr)
+{
+	/* read mode */
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT5,
+		       SH_MOBILE_SDHI_SCC_TMPPORT5_DLL_RW_SEL_R |
+		       (SH_MOBILE_SDHI_SCC_TMPPORT5_DLL_ADR_MASK & addr));
+
+	/* access start and stop */
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT4,
+		       SH_MOBILE_SDHI_SCC_TMPPORT4_DLL_ACC_START);
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT4, 0);
+
+	return sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT7);
+}
+
+static void sd_scc_tmpport_write32(struct tmio_mmc_host *host,
+				   struct renesas_sdhi *priv, u32 addr, u32 val)
+{
+	/* write mode */
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT5,
+		       SH_MOBILE_SDHI_SCC_TMPPORT5_DLL_RW_SEL_W |
+		       (SH_MOBILE_SDHI_SCC_TMPPORT5_DLL_ADR_MASK & addr));
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT6, val);
+
+	/* access start and stop */
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT4,
+		       SH_MOBILE_SDHI_SCC_TMPPORT4_DLL_ACC_START);
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT4, 0);
+}
+
+static void renesas_sdhi_adjust_hs400_mode_enable(struct tmio_mmc_host *host)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+	u32 calib_code;
+
+	/* disable write protect */
+	sd_scc_tmpport_write32(host, priv, 0x00,
+			       SH_MOBILE_SDHI_SCC_TMPPORT_DISABLE_WP_CODE);
+	/* read calibration code and adjust */
+	calib_code = sd_scc_tmpport_read32(host, priv, 0x26);
+	calib_code &= SH_MOBILE_SDHI_SCC_TMPPORT_CALIB_CODE_MASK;
+
+	sd_scc_tmpport_write32(host, priv, 0x22,
+			       SH_MOBILE_SDHI_SCC_TMPPORT_MANUAL_MODE |
+			       priv->adjust_hs400_calib_table[calib_code]);
+
+	/* set offset value to TMPPORT3, hardcoded to OFFSET0 (= 0x3) for now */
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT3, 0x3);
+
+	/* adjustment done, clear flag */
+	priv->needs_adjust_hs400 = false;
+}
+
+static void renesas_sdhi_adjust_hs400_mode_disable(struct tmio_mmc_host *host)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+
+	/* disable write protect */
+	sd_scc_tmpport_write32(host, priv, 0x00,
+			       SH_MOBILE_SDHI_SCC_TMPPORT_DISABLE_WP_CODE);
+	/* disable manual calibration */
+	sd_scc_tmpport_write32(host, priv, 0x22, 0);
+	/* clear offset value of TMPPORT3 */
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT3, 0);
+}
+
+static void renesas_sdhi_reset_hs400_mode(struct tmio_mmc_host *host,
+					  struct renesas_sdhi *priv)
+{
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, ~CLK_CTL_SCLKEN &
+			sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+
+	/* Reset HS400 mode */
+	sd_ctrl_write16(host, CTL_SDIF_MODE, ~0x0001 &
+			sd_ctrl_read16(host, CTL_SDIF_MODE));
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_DT2FF, priv->scc_tappos);
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT2,
+		       ~(SH_MOBILE_SDHI_SCC_TMPPORT2_HS400EN |
+			 SH_MOBILE_SDHI_SCC_TMPPORT2_HS400OSEL) &
+			sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_TMPPORT2));
+
+	if (priv->quirks && (priv->quirks->hs400_calib_table || priv->quirks->hs400_bad_taps))
+		renesas_sdhi_adjust_hs400_mode_disable(host);
+
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, CLK_CTL_SCLKEN |
+			sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+}
+
+static int renesas_sdhi_prepare_hs400_tuning(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+
+	renesas_sdhi_reset_hs400_mode(host, host_to_priv(host));
+	return 0;
+}
+
+static void renesas_sdhi_reset(struct tmio_mmc_host *host)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+
+	if (priv->scc_ctl) {
+		renesas_sdhi_reset_scc(host, priv);
+		renesas_sdhi_reset_hs400_mode(host, priv);
+		priv->needs_adjust_hs400 = false;
+
+		sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, CLK_CTL_SCLKEN |
+				sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+
+		sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_RVSCNTL,
+			       ~SH_MOBILE_SDHI_SCC_RVSCNTL_RVSEN &
+			       sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_RVSCNTL));
+	}
+
+	if (host->pdata->flags & TMIO_MMC_MIN_RCAR2)
+		sd_ctrl_write32_as_16_and_16(host, CTL_IRQ_MASK,
+					     TMIO_MASK_ALL_RCAR2);
+}
+
+#define SH_MOBILE_SDHI_MIN_TAP_ROW 3
+
+static int renesas_sdhi_select_tuning(struct tmio_mmc_host *host)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+	unsigned int tap_start = 0, tap_end = 0, tap_cnt = 0, rs, re, i;
+	unsigned int taps_size = priv->tap_num * 2, min_tap_row;
+	unsigned long *bitmap;
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_RVSREQ, 0);
+
+	/*
+	 * When tuning CMD19 is issued twice for each tap, merge the
+	 * result requiring the tap to be good in both runs before
+	 * considering it for tuning selection.
+	 */
+	for (i = 0; i < taps_size; i++) {
+		int offset = priv->tap_num * (i < priv->tap_num ? 1 : -1);
+
+		if (!test_bit(i, priv->taps))
+			clear_bit(i + offset, priv->taps);
+
+		if (!test_bit(i, priv->smpcmp))
+			clear_bit(i + offset, priv->smpcmp);
+	}
+
+	/*
+	 * If all TAP are OK, the sampling clock position is selected by
+	 * identifying the change point of data.
+	 */
+	if (bitmap_full(priv->taps, taps_size)) {
+		bitmap = priv->smpcmp;
+		min_tap_row = 1;
+	} else {
+		bitmap = priv->taps;
+		min_tap_row = SH_MOBILE_SDHI_MIN_TAP_ROW;
+	}
+
+	/*
+	 * Find the longest consecutive run of successful probes. If that
+	 * is at least SH_MOBILE_SDHI_MIN_TAP_ROW probes long then use the
+	 * center index as the tap, otherwise bail out.
+	 */
+	bitmap_for_each_set_region(bitmap, rs, re, 0, taps_size) {
+		if (re - rs > tap_cnt) {
+			tap_end = re;
+			tap_start = rs;
+			tap_cnt = tap_end - tap_start;
+		}
+	}
+
+	if (tap_cnt >= min_tap_row)
+		priv->tap_set = (tap_start + tap_end) / 2 % priv->tap_num;
+	else
+		return -EIO;
+
+	/* Set SCC */
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TAPSET, priv->tap_set);
+
+	/* Enable auto re-tuning */
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_RVSCNTL,
+		       SH_MOBILE_SDHI_SCC_RVSCNTL_RVSEN |
+		       sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_RVSCNTL));
+
+	return 0;
+}
+
+static int renesas_sdhi_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+	struct renesas_sdhi *priv = host_to_priv(host);
+	int i, ret;
+
+	priv->tap_num = renesas_sdhi_init_tuning(host);
+	if (!priv->tap_num)
+		return 0; /* Tuning is not supported */
+
+	if (priv->tap_num * 2 >= sizeof(priv->taps) * BITS_PER_BYTE) {
+		dev_err(&host->pdev->dev,
+			"Too many taps, please update 'taps' in tmio_mmc_host!\n");
+		return -EINVAL;
+	}
+
+	bitmap_zero(priv->taps, priv->tap_num * 2);
+	bitmap_zero(priv->smpcmp, priv->tap_num * 2);
+
+	/* Issue CMD19 twice for each tap */
+	for (i = 0; i < 2 * priv->tap_num; i++) {
+		int cmd_error = 0;
+
+		/* Set sampling clock position */
+		sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TAPSET, i % priv->tap_num);
+
+		if (mmc_send_tuning(mmc, opcode, &cmd_error) == 0)
+			set_bit(i, priv->taps);
+
+		if (sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_SMPCMP) == 0)
+			set_bit(i, priv->smpcmp);
+
+		if (cmd_error)
+			mmc_abort_tuning(mmc, opcode);
+	}
+
+	ret = renesas_sdhi_select_tuning(host);
+	if (ret < 0)
+		renesas_sdhi_reset(host);
+	return ret;
+}
+
+static bool renesas_sdhi_manual_correction(struct tmio_mmc_host *host, bool use_4tap)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+	unsigned int new_tap = priv->tap_set, error_tap = priv->tap_set;
+	u32 val;
+
+	val = sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_RVSREQ);
+	if (!val)
+		return false;
+
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_RVSREQ, 0);
+
+	/* Change TAP position according to correction status */
+	if (sd_ctrl_read16(host, CTL_VERSION) == SDHI_VER_GEN3_SDMMC &&
+	    host->mmc->ios.timing == MMC_TIMING_MMC_HS400) {
+		u32 bad_taps = priv->quirks ? priv->quirks->hs400_bad_taps : 0;
+		/*
+		 * With HS400, the DAT signal is based on DS, not CLK.
+		 * Therefore, use only CMD status.
+		 */
+		u32 smpcmp = sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_SMPCMP) &
+					   SH_MOBILE_SDHI_SCC_SMPCMP_CMD_ERR;
+		if (!smpcmp) {
+			return false;	/* no error in CMD signal */
+		} else if (smpcmp == SH_MOBILE_SDHI_SCC_SMPCMP_CMD_REQUP) {
+			new_tap++;
+			error_tap--;
+		} else if (smpcmp == SH_MOBILE_SDHI_SCC_SMPCMP_CMD_REQDOWN) {
+			new_tap--;
+			error_tap++;
+		} else {
+			return true;	/* need retune */
+		}
+
+		/*
+		 * When new_tap is a bad tap, we cannot change. Then, we compare
+		 * with the HS200 tuning result. When smpcmp[error_tap] is OK,
+		 * we can at least retune.
+		 */
+		if (bad_taps & BIT(new_tap % priv->tap_num))
+			return test_bit(error_tap % priv->tap_num, priv->smpcmp);
+	} else {
+		if (val & SH_MOBILE_SDHI_SCC_RVSREQ_RVSERR)
+			return true;    /* need retune */
+		else if (val & SH_MOBILE_SDHI_SCC_RVSREQ_REQTAPUP)
+			new_tap++;
+		else if (val & SH_MOBILE_SDHI_SCC_RVSREQ_REQTAPDOWN)
+			new_tap--;
+		else
+			return false;
+	}
+
+	priv->tap_set = (new_tap % priv->tap_num);
+	sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_TAPSET,
+		       priv->tap_set / (use_4tap ? 2 : 1));
+
+	return false;
+}
+
+static bool renesas_sdhi_auto_correction(struct tmio_mmc_host *host)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+
+	/* Check SCC error */
+	if (sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_RVSREQ) &
+	    SH_MOBILE_SDHI_SCC_RVSREQ_RVSERR) {
+		sd_scc_write32(host, priv, SH_MOBILE_SDHI_SCC_RVSREQ, 0);
+		return true;
+	}
+
+	return false;
+}
+
+static bool renesas_sdhi_check_scc_error(struct tmio_mmc_host *host)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+	bool use_4tap = priv->quirks && priv->quirks->hs400_4taps;
+
+	/*
+	 * Skip checking SCC errors when running on 4 taps in HS400 mode as
+	 * any retuning would still result in the same 4 taps being used.
+	 */
+	if (!(host->mmc->ios.timing == MMC_TIMING_UHS_SDR104) &&
+	    !(host->mmc->ios.timing == MMC_TIMING_MMC_HS200) &&
+	    !(host->mmc->ios.timing == MMC_TIMING_MMC_HS400 && !use_4tap))
+		return false;
+
+	if (mmc_doing_tune(host->mmc))
+		return false;
+
+	if (sd_scc_read32(host, priv, SH_MOBILE_SDHI_SCC_RVSCNTL) &
+	    SH_MOBILE_SDHI_SCC_RVSCNTL_RVSEN)
+		return renesas_sdhi_auto_correction(host);
+
+	return renesas_sdhi_manual_correction(host, use_4tap);
+}
+
+static int renesas_sdhi_wait_idle(struct tmio_mmc_host *host, u32 bit)
+{
+	int timeout = 1000;
+	/* CBSY is set when busy, SCLKDIVEN is cleared when busy */
+	u32 wait_state = (bit == TMIO_STAT_CMD_BUSY ? TMIO_STAT_CMD_BUSY : 0);
+
+	while (--timeout && (sd_ctrl_read16_and_16_as_32(host, CTL_STATUS)
+			      & bit) == wait_state)
+		udelay(1);
+
+	if (!timeout) {
+		dev_warn(&host->pdev->dev, "timeout waiting for SD bus idle\n");
+		return -EBUSY;
+	}
+
+	return 0;
+}
+
+static int renesas_sdhi_write16_hook(struct tmio_mmc_host *host, int addr)
+{
+	u32 bit = TMIO_STAT_SCLKDIVEN;
+
+	switch (addr) {
+	case CTL_SD_CMD:
+	case CTL_STOP_INTERNAL_ACTION:
+	case CTL_XFER_BLK_COUNT:
+	case CTL_SD_XFER_LEN:
+	case CTL_SD_MEM_CARD_OPT:
+	case CTL_TRANSACTION_CTL:
+	case CTL_DMA_ENABLE:
+	case HOST_MODE:
+		if (host->pdata->flags & TMIO_MMC_HAVE_CBSY)
+			bit = TMIO_STAT_CMD_BUSY;
+		fallthrough;
+	case CTL_SD_CARD_CLK_CTL:
+		return renesas_sdhi_wait_idle(host, bit);
+	}
+
+	return 0;
+}
+
+static int renesas_sdhi_multi_io_quirk(struct mmc_card *card,
+				       unsigned int direction, int blk_size)
+{
+	/*
+	 * In Renesas controllers, when performing a
+	 * multiple block read of one or two blocks,
+	 * depending on the timing with which the
+	 * response register is read, the response
+	 * value may not be read properly.
+	 * Use single block read for this HW bug
+	 */
+	if ((direction == MMC_DATA_READ) &&
+	    blk_size == 2)
+		return 1;
+
+	return blk_size;
+}
+
+static void renesas_sdhi_fixup_request(struct tmio_mmc_host *host, struct mmc_request *mrq)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+
+	if (priv->needs_adjust_hs400 && mrq->cmd->opcode == MMC_SEND_STATUS)
+		renesas_sdhi_adjust_hs400_mode_enable(host);
+}
+static void renesas_sdhi_enable_dma(struct tmio_mmc_host *host, bool enable)
+{
+	/* Iff regs are 8 byte apart, sdbuf is 64 bit. Otherwise always 32. */
+	int width = (host->bus_shift == 2) ? 64 : 32;
+
+	sd_ctrl_write16(host, CTL_DMA_ENABLE, enable ? DMA_ENABLE_DMASDRW : 0);
+	renesas_sdhi_sdbuf_width(host, enable ? width : 16);
+}
+
+static const struct renesas_sdhi_quirks sdhi_quirks_4tap_nohs400 = {
+	.hs400_disabled = true,
+	.hs400_4taps = true,
+};
+
+static const struct renesas_sdhi_quirks sdhi_quirks_4tap = {
+	.hs400_4taps = true,
+	.hs400_bad_taps = BIT(2) | BIT(3) | BIT(6) | BIT(7),
+};
+
+static const struct renesas_sdhi_quirks sdhi_quirks_nohs400 = {
+	.hs400_disabled = true,
+};
+
+static const struct renesas_sdhi_quirks sdhi_quirks_bad_taps1357 = {
+	.hs400_bad_taps = BIT(1) | BIT(3) | BIT(5) | BIT(7),
+};
+
+static const struct renesas_sdhi_quirks sdhi_quirks_bad_taps2367 = {
+	.hs400_bad_taps = BIT(2) | BIT(3) | BIT(6) | BIT(7),
+};
+
+static const struct renesas_sdhi_quirks sdhi_quirks_r8a7796_es13 = {
+	.hs400_4taps = true,
+	.hs400_bad_taps = BIT(2) | BIT(3) | BIT(6) | BIT(7),
+	.hs400_calib_table = r8a7796_es13_calib_table,
+};
+
+static const struct renesas_sdhi_quirks sdhi_quirks_r8a77965 = {
+	.hs400_bad_taps = BIT(2) | BIT(3) | BIT(6) | BIT(7),
+	.hs400_calib_table = r8a77965_calib_table,
+};
+
+static const struct renesas_sdhi_quirks sdhi_quirks_r8a77990 = {
+	.hs400_calib_table = r8a77990_calib_table,
+};
+
+/*
+ * Note for r8a7796 / r8a774a1: we can't distinguish ES1.1 and 1.2 as of now.
+ * So, we want to treat them equally and only have a match for ES1.2 to enforce
+ * this if there ever will be a way to distinguish ES1.2.
+ */
+static const struct soc_device_attribute sdhi_quirks_match[]  = {
+	{ .soc_id = "r8a774a1", .revision = "ES1.[012]", .data = &sdhi_quirks_4tap_nohs400 },
+	{ .soc_id = "r8a7795", .revision = "ES1.*", .data = &sdhi_quirks_4tap_nohs400 },
+	{ .soc_id = "r8a7795", .revision = "ES2.0", .data = &sdhi_quirks_4tap },
+	{ .soc_id = "r8a7795", .revision = "ES3.*", .data = &sdhi_quirks_bad_taps2367 },
+	{ .soc_id = "r8a7796", .revision = "ES1.[012]", .data = &sdhi_quirks_4tap_nohs400 },
+	{ .soc_id = "r8a7796", .revision = "ES1.*", .data = &sdhi_quirks_r8a7796_es13 },
+	{ .soc_id = "r8a77961", .data = &sdhi_quirks_bad_taps1357 },
+	{ .soc_id = "r8a77965", .data = &sdhi_quirks_r8a77965 },
+	{ .soc_id = "r8a77980", .data = &sdhi_quirks_nohs400 },
+	{ .soc_id = "r8a77990", .data = &sdhi_quirks_r8a77990 },
+	{ /* Sentinel. */ },
+};
+
+int renesas_sdhi_probe(struct platform_device *pdev,
+		       const struct tmio_mmc_dma_ops *dma_ops)
+{
+	struct tmio_mmc_data *mmd = pdev->dev.platform_data;
+	const struct renesas_sdhi_quirks *quirks = NULL;
+	const struct renesas_sdhi_of_data *of_data;
+	const struct soc_device_attribute *attr;
+	struct tmio_mmc_data *mmc_data;
+	struct tmio_mmc_dma *dma_priv;
+	struct tmio_mmc_host *host;
+	struct renesas_sdhi *priv;
+	int num_irqs, irq, ret, i;
+	struct resource *res;
+	u16 ver;
+
+	of_data = of_device_get_match_data(&pdev->dev);
+
+	attr = soc_device_match(sdhi_quirks_match);
+	if (attr)
+		quirks = attr->data;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res)
+		return -EINVAL;
+
+	priv = devm_kzalloc(&pdev->dev, sizeof(struct renesas_sdhi),
+			    GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->quirks = quirks;
+	mmc_data = &priv->mmc_data;
+	dma_priv = &priv->dma_priv;
+
+	priv->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(priv->clk)) {
+		ret = PTR_ERR(priv->clk);
+		dev_err(&pdev->dev, "cannot get clock: %d\n", ret);
+		return ret;
+	}
+
+	/*
+	 * Some controllers provide a 2nd clock just to run the internal card
+	 * detection logic. Unfortunately, the existing driver architecture does
+	 * not support a separation of clocks for runtime PM usage. When
+	 * native hotplug is used, the tmio driver assumes that the core
+	 * must continue to run for card detect to stay active, so we cannot
+	 * disable it.
+	 * Additionally, it is prohibited to supply a clock to the core but not
+	 * to the card detect circuit. That leaves us with if separate clocks
+	 * are presented, we must treat them both as virtually 1 clock.
+	 */
+	priv->clk_cd = devm_clk_get(&pdev->dev, "cd");
+	if (IS_ERR(priv->clk_cd))
+		priv->clk_cd = NULL;
+
+	priv->pinctrl = devm_pinctrl_get(&pdev->dev);
+	if (!IS_ERR(priv->pinctrl)) {
+		priv->pins_default = pinctrl_lookup_state(priv->pinctrl,
+						PINCTRL_STATE_DEFAULT);
+		priv->pins_uhs = pinctrl_lookup_state(priv->pinctrl,
+						"state_uhs");
+	}
+
+	host = tmio_mmc_host_alloc(pdev, mmc_data);
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	if (of_data) {
+		mmc_data->flags |= of_data->tmio_flags;
+		mmc_data->ocr_mask = of_data->tmio_ocr_mask;
+		mmc_data->capabilities |= of_data->capabilities;
+		mmc_data->capabilities2 |= of_data->capabilities2;
+		mmc_data->dma_rx_offset = of_data->dma_rx_offset;
+		mmc_data->max_blk_count = of_data->max_blk_count;
+		mmc_data->max_segs = of_data->max_segs;
+		dma_priv->dma_buswidth = of_data->dma_buswidth;
+		host->bus_shift = of_data->bus_shift;
+	}
+
+	host->write16_hook	= renesas_sdhi_write16_hook;
+	host->clk_enable	= renesas_sdhi_clk_enable;
+	host->clk_disable	= renesas_sdhi_clk_disable;
+	host->set_clock		= renesas_sdhi_set_clock;
+	host->multi_io_quirk	= renesas_sdhi_multi_io_quirk;
+	host->dma_ops		= dma_ops;
+
+	if (quirks && quirks->hs400_disabled)
+		host->mmc->caps2 &= ~(MMC_CAP2_HS400 | MMC_CAP2_HS400_ES);
+
+	/* For some SoC, we disable internal WP. GPIO may override this */
+	if (mmc_can_gpio_ro(host->mmc))
+		mmc_data->capabilities2 &= ~MMC_CAP2_NO_WRITE_PROTECT;
+
+	/* SDR speeds are only available on Gen2+ */
+	if (mmc_data->flags & TMIO_MMC_MIN_RCAR2) {
+		/* card_busy caused issues on r8a73a4 (pre-Gen2) CD-less SDHI */
+		host->ops.card_busy = renesas_sdhi_card_busy;
+		host->ops.start_signal_voltage_switch =
+			renesas_sdhi_start_signal_voltage_switch;
+		host->sdcard_irq_setbit_mask = TMIO_STAT_ALWAYS_SET_27;
+		host->sdcard_irq_mask_all = TMIO_MASK_ALL_RCAR2;
+		host->reset = renesas_sdhi_reset;
+	} else {
+		host->sdcard_irq_mask_all = TMIO_MASK_ALL;
+	}
+
+	/* Orginally registers were 16 bit apart, could be 32 or 64 nowadays */
+	if (!host->bus_shift && resource_size(res) > 0x100) /* old way to determine the shift */
+		host->bus_shift = 1;
+
+	if (mmd)
+		*mmc_data = *mmd;
+
+	dma_priv->filter = shdma_chan_filter;
+	dma_priv->enable = renesas_sdhi_enable_dma;
+
+	mmc_data->alignment_shift = 1; /* 2-byte alignment */
+	mmc_data->capabilities |= MMC_CAP_MMC_HIGHSPEED;
+
+	/*
+	 * All SDHI blocks support 2-byte and larger block sizes in 4-bit
+	 * bus width mode.
+	 */
+	mmc_data->flags |= TMIO_MMC_BLKSZ_2BYTES;
+
+	/*
+	 * All SDHI blocks support SDIO IRQ signalling.
+	 */
+	mmc_data->flags |= TMIO_MMC_SDIO_IRQ;
+
+	/* All SDHI have CMD12 control bit */
+	mmc_data->flags |= TMIO_MMC_HAVE_CMD12_CTRL;
+
+	/* All SDHI have SDIO status bits which must be 1 */
+	mmc_data->flags |= TMIO_MMC_SDIO_STATUS_SETBITS;
+
+	dev_pm_domain_start(&pdev->dev);
+
+	ret = renesas_sdhi_clk_enable(host);
+	if (ret)
+		goto efree;
+
+	ver = sd_ctrl_read16(host, CTL_VERSION);
+	/* GEN2_SDR104 is first known SDHI to use 32bit block count */
+	if (ver < SDHI_VER_GEN2_SDR104 && mmc_data->max_blk_count > U16_MAX)
+		mmc_data->max_blk_count = U16_MAX;
+
+	/* One Gen2 SDHI incarnation does NOT have a CBSY bit */
+	if (ver == SDHI_VER_GEN2_SDR50)
+		mmc_data->flags &= ~TMIO_MMC_HAVE_CBSY;
+
+	if (ver == SDHI_VER_GEN3_SDMMC && quirks && quirks->hs400_calib_table) {
+		host->fixup_request = renesas_sdhi_fixup_request;
+		priv->adjust_hs400_calib_table = *(
+			res->start == SDHI_GEN3_MMC0_ADDR ?
+			quirks->hs400_calib_table :
+			quirks->hs400_calib_table + 1);
+	}
+
+	/* Enable tuning iff we have an SCC and a supported mode */
+	if (of_data && of_data->scc_offset &&
+	    (host->mmc->caps & MMC_CAP_UHS_SDR104 ||
+	     host->mmc->caps2 & (MMC_CAP2_HS200_1_8V_SDR |
+				 MMC_CAP2_HS400_1_8V))) {
+		const struct renesas_sdhi_scc *taps = of_data->taps;
+		bool use_4tap = priv->quirks && priv->quirks->hs400_4taps;
+		bool hit = false;
+
+		for (i = 0; i < of_data->taps_num; i++) {
+			if (taps[i].clk_rate == 0 ||
+			    taps[i].clk_rate == host->mmc->f_max) {
+				priv->scc_tappos = taps->tap;
+				priv->scc_tappos_hs400 = use_4tap ?
+							 taps->tap_hs400_4tap :
+							 taps->tap;
+				hit = true;
+				break;
+			}
+		}
+
+		if (!hit)
+			dev_warn(&host->pdev->dev, "Unknown clock rate for tuning\n");
+
+		priv->scc_ctl = host->ctl + of_data->scc_offset;
+		host->check_retune = renesas_sdhi_check_scc_error;
+		host->ops.execute_tuning = renesas_sdhi_execute_tuning;
+		host->ops.prepare_hs400_tuning = renesas_sdhi_prepare_hs400_tuning;
+		host->ops.hs400_downgrade = renesas_sdhi_disable_scc;
+		host->ops.hs400_complete = renesas_sdhi_hs400_complete;
+	}
+
+	sd_ctrl_write32_as_16_and_16(host, CTL_IRQ_MASK, host->sdcard_irq_mask_all);
+
+	num_irqs = platform_irq_count(pdev);
+	if (num_irqs < 0) {
+		ret = num_irqs;
+		goto eirq;
+	}
+
+	/* There must be at least one IRQ source */
+	if (!num_irqs) {
+		ret = -ENXIO;
+		goto eirq;
+	}
+
+	for (i = 0; i < num_irqs; i++) {
+		irq = platform_get_irq(pdev, i);
+		if (irq < 0) {
+			ret = irq;
+			goto eirq;
+		}
+
+		ret = devm_request_irq(&pdev->dev, irq, tmio_mmc_irq, 0,
+				       dev_name(&pdev->dev), host);
+		if (ret)
+			goto eirq;
+	}
+
+	ret = tmio_mmc_host_probe(host);
+	if (ret < 0)
+		goto edisclk;
+
+	dev_info(&pdev->dev, "%s base at %pa, max clock rate %u MHz\n",
+		 mmc_hostname(host->mmc), &res->start, host->mmc->f_max / 1000000);
+
+	return ret;
+
+eirq:
+	tmio_mmc_host_remove(host);
+edisclk:
+	renesas_sdhi_clk_disable(host);
+efree:
+	tmio_mmc_host_free(host);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(renesas_sdhi_probe);
+
+int renesas_sdhi_remove(struct platform_device *pdev)
+{
+	struct tmio_mmc_host *host = platform_get_drvdata(pdev);
+
+	tmio_mmc_host_remove(host);
+	renesas_sdhi_clk_disable(host);
+	tmio_mmc_host_free(host);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(renesas_sdhi_remove);
+
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/renesas_sdhi_internal_dmac.c b/drivers/mmc/host/renesas_sdhi_internal_dmac.c
new file mode 100644
index 000000000..f3e76d6b3
--- /dev/null
+++ b/drivers/mmc/host/renesas_sdhi_internal_dmac.c
@@ -0,0 +1,368 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * DMA support for Internal DMAC with SDHI SD/SDIO controller
+ *
+ * Copyright (C) 2016-19 Renesas Electronics Corporation
+ * Copyright (C) 2016-17 Horms Solutions, Simon Horman
+ * Copyright (C) 2018-19 Sang Engineering, Wolfram Sang
+ */
+
+#include <linux/bitops.h>
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/io-64-nonatomic-hi-lo.h>
+#include <linux/mfd/tmio.h>
+#include <linux/mmc/host.h>
+#include <linux/mod_devicetable.h>
+#include <linux/module.h>
+#include <linux/pagemap.h>
+#include <linux/scatterlist.h>
+#include <linux/sys_soc.h>
+
+#include "renesas_sdhi.h"
+#include "tmio_mmc.h"
+
+#define DM_CM_DTRAN_MODE	0x820
+#define DM_CM_DTRAN_CTRL	0x828
+#define DM_CM_RST		0x830
+#define DM_CM_INFO1		0x840
+#define DM_CM_INFO1_MASK	0x848
+#define DM_CM_INFO2		0x850
+#define DM_CM_INFO2_MASK	0x858
+#define DM_DTRAN_ADDR		0x880
+
+/* DM_CM_DTRAN_MODE */
+#define DTRAN_MODE_CH_NUM_CH0	0	/* "downstream" = for write commands */
+#define DTRAN_MODE_CH_NUM_CH1	BIT(16)	/* "upstream" = for read commands */
+#define DTRAN_MODE_BUS_WIDTH	(BIT(5) | BIT(4))
+#define DTRAN_MODE_ADDR_MODE	BIT(0)	/* 1 = Increment address, 0 = Fixed */
+
+/* DM_CM_DTRAN_CTRL */
+#define DTRAN_CTRL_DM_START	BIT(0)
+
+/* DM_CM_RST */
+#define RST_DTRANRST1		BIT(9)
+#define RST_DTRANRST0		BIT(8)
+#define RST_RESERVED_BITS	GENMASK_ULL(31, 0)
+
+/* DM_CM_INFO1 and DM_CM_INFO1_MASK */
+#define INFO1_CLEAR		0
+#define INFO1_MASK_CLEAR	GENMASK_ULL(31, 0)
+#define INFO1_DTRANEND1		BIT(17)
+#define INFO1_DTRANEND0		BIT(16)
+
+/* DM_CM_INFO2 and DM_CM_INFO2_MASK */
+#define INFO2_MASK_CLEAR	GENMASK_ULL(31, 0)
+#define INFO2_DTRANERR1		BIT(17)
+#define INFO2_DTRANERR0		BIT(16)
+
+/*
+ * Specification of this driver:
+ * - host->chan_{rx,tx} will be used as a flag of enabling/disabling the dma
+ * - Since this SDHI DMAC register set has 16 but 32-bit width, we
+ *   need a custom accessor.
+ */
+
+static unsigned long global_flags;
+/*
+ * Workaround for avoiding to use RX DMAC by multiple channels.
+ * On R-Car H3 ES1.* and M3-W ES1.0, when multiple SDHI channels use
+ * RX DMAC simultaneously, sometimes hundreds of bytes data are not
+ * stored into the system memory even if the DMAC interrupt happened.
+ * So, this driver then uses one RX DMAC channel only.
+ */
+#define SDHI_INTERNAL_DMAC_ONE_RX_ONLY	0
+#define SDHI_INTERNAL_DMAC_RX_IN_USE	1
+
+/* RZ/A2 does not have the ADRR_MODE bit */
+#define SDHI_INTERNAL_DMAC_ADDR_MODE_FIXED_ONLY 2
+
+/* Definitions for sampling clocks */
+static struct renesas_sdhi_scc rcar_gen3_scc_taps[] = {
+	{
+		.clk_rate = 0,
+		.tap = 0x00000300,
+		.tap_hs400_4tap = 0x00000100,
+	},
+};
+
+static const struct renesas_sdhi_of_data of_rza2_compatible = {
+	.tmio_flags	= TMIO_MMC_HAS_IDLE_WAIT | TMIO_MMC_CLK_ACTUAL |
+			  TMIO_MMC_HAVE_CBSY,
+	.tmio_ocr_mask	= MMC_VDD_32_33,
+	.capabilities	= MMC_CAP_SD_HIGHSPEED | MMC_CAP_SDIO_IRQ |
+			  MMC_CAP_CMD23,
+	.bus_shift	= 2,
+	.scc_offset	= 0 - 0x1000,
+	.taps		= rcar_gen3_scc_taps,
+	.taps_num	= ARRAY_SIZE(rcar_gen3_scc_taps),
+	/* DMAC can handle 32bit blk count but only 1 segment */
+	.max_blk_count	= UINT_MAX / TMIO_MAX_BLK_SIZE,
+	.max_segs	= 1,
+};
+
+static const struct renesas_sdhi_of_data of_rcar_gen3_compatible = {
+	.tmio_flags	= TMIO_MMC_HAS_IDLE_WAIT | TMIO_MMC_CLK_ACTUAL |
+			  TMIO_MMC_HAVE_CBSY | TMIO_MMC_MIN_RCAR2,
+	.capabilities	= MMC_CAP_SD_HIGHSPEED | MMC_CAP_SDIO_IRQ |
+			  MMC_CAP_CMD23,
+	.capabilities2	= MMC_CAP2_NO_WRITE_PROTECT | MMC_CAP2_MERGE_CAPABLE,
+	.bus_shift	= 2,
+	.scc_offset	= 0x1000,
+	.taps		= rcar_gen3_scc_taps,
+	.taps_num	= ARRAY_SIZE(rcar_gen3_scc_taps),
+	/* DMAC can handle 32bit blk count but only 1 segment */
+	.max_blk_count	= UINT_MAX / TMIO_MAX_BLK_SIZE,
+	.max_segs	= 1,
+};
+
+static const struct of_device_id renesas_sdhi_internal_dmac_of_match[] = {
+	{ .compatible = "renesas,sdhi-r7s9210", .data = &of_rza2_compatible, },
+	{ .compatible = "renesas,sdhi-mmc-r8a77470", .data = &of_rcar_gen3_compatible, },
+	{ .compatible = "renesas,sdhi-r8a7795", .data = &of_rcar_gen3_compatible, },
+	{ .compatible = "renesas,sdhi-r8a7796", .data = &of_rcar_gen3_compatible, },
+	{ .compatible = "renesas,rcar-gen3-sdhi", .data = &of_rcar_gen3_compatible, },
+	{},
+};
+MODULE_DEVICE_TABLE(of, renesas_sdhi_internal_dmac_of_match);
+
+static void
+renesas_sdhi_internal_dmac_dm_write(struct tmio_mmc_host *host,
+				    int addr, u64 val)
+{
+	writeq(val, host->ctl + addr);
+}
+
+static void
+renesas_sdhi_internal_dmac_enable_dma(struct tmio_mmc_host *host, bool enable)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+
+	if (!host->chan_tx || !host->chan_rx)
+		return;
+
+	if (!enable)
+		renesas_sdhi_internal_dmac_dm_write(host, DM_CM_INFO1,
+						    INFO1_CLEAR);
+
+	if (priv->dma_priv.enable)
+		priv->dma_priv.enable(host, enable);
+}
+
+static void
+renesas_sdhi_internal_dmac_abort_dma(struct tmio_mmc_host *host) {
+	u64 val = RST_DTRANRST1 | RST_DTRANRST0;
+
+	renesas_sdhi_internal_dmac_enable_dma(host, false);
+
+	renesas_sdhi_internal_dmac_dm_write(host, DM_CM_RST,
+					    RST_RESERVED_BITS & ~val);
+	renesas_sdhi_internal_dmac_dm_write(host, DM_CM_RST,
+					    RST_RESERVED_BITS | val);
+
+	clear_bit(SDHI_INTERNAL_DMAC_RX_IN_USE, &global_flags);
+
+	renesas_sdhi_internal_dmac_enable_dma(host, true);
+}
+
+static void
+renesas_sdhi_internal_dmac_dataend_dma(struct tmio_mmc_host *host) {
+	struct renesas_sdhi *priv = host_to_priv(host);
+
+	tasklet_schedule(&priv->dma_priv.dma_complete);
+}
+
+static void
+renesas_sdhi_internal_dmac_start_dma(struct tmio_mmc_host *host,
+				     struct mmc_data *data)
+{
+	struct scatterlist *sg = host->sg_ptr;
+	u32 dtran_mode = DTRAN_MODE_BUS_WIDTH;
+
+	if (!test_bit(SDHI_INTERNAL_DMAC_ADDR_MODE_FIXED_ONLY, &global_flags))
+		dtran_mode |= DTRAN_MODE_ADDR_MODE;
+
+	if (!dma_map_sg(&host->pdev->dev, sg, host->sg_len,
+			mmc_get_dma_dir(data)))
+		goto force_pio;
+
+	/* This DMAC cannot handle if buffer is not 128-bytes alignment */
+	if (!IS_ALIGNED(sg_dma_address(sg), 128))
+		goto force_pio_with_unmap;
+
+	if (data->flags & MMC_DATA_READ) {
+		dtran_mode |= DTRAN_MODE_CH_NUM_CH1;
+		if (test_bit(SDHI_INTERNAL_DMAC_ONE_RX_ONLY, &global_flags) &&
+		    test_and_set_bit(SDHI_INTERNAL_DMAC_RX_IN_USE, &global_flags))
+			goto force_pio_with_unmap;
+	} else {
+		dtran_mode |= DTRAN_MODE_CH_NUM_CH0;
+	}
+
+	renesas_sdhi_internal_dmac_enable_dma(host, true);
+
+	/* set dma parameters */
+	renesas_sdhi_internal_dmac_dm_write(host, DM_CM_DTRAN_MODE,
+					    dtran_mode);
+	renesas_sdhi_internal_dmac_dm_write(host, DM_DTRAN_ADDR,
+					    sg_dma_address(sg));
+
+	host->dma_on = true;
+
+	return;
+
+force_pio_with_unmap:
+	dma_unmap_sg(&host->pdev->dev, sg, host->sg_len, mmc_get_dma_dir(data));
+
+force_pio:
+	renesas_sdhi_internal_dmac_enable_dma(host, false);
+}
+
+static void renesas_sdhi_internal_dmac_issue_tasklet_fn(unsigned long arg)
+{
+	struct tmio_mmc_host *host = (struct tmio_mmc_host *)arg;
+
+	tmio_mmc_enable_mmc_irqs(host, TMIO_STAT_DATAEND);
+
+	/* start the DMAC */
+	renesas_sdhi_internal_dmac_dm_write(host, DM_CM_DTRAN_CTRL,
+					    DTRAN_CTRL_DM_START);
+}
+
+static bool renesas_sdhi_internal_dmac_complete(struct tmio_mmc_host *host)
+{
+	enum dma_data_direction dir;
+
+	if (!host->dma_on)
+		return false;
+
+	if (!host->data)
+		return false;
+
+	if (host->data->flags & MMC_DATA_READ)
+		dir = DMA_FROM_DEVICE;
+	else
+		dir = DMA_TO_DEVICE;
+
+	renesas_sdhi_internal_dmac_enable_dma(host, false);
+	dma_unmap_sg(&host->pdev->dev, host->sg_ptr, host->sg_len, dir);
+
+	if (dir == DMA_FROM_DEVICE)
+		clear_bit(SDHI_INTERNAL_DMAC_RX_IN_USE, &global_flags);
+
+	host->dma_on = false;
+
+	return true;
+}
+
+static void renesas_sdhi_internal_dmac_complete_tasklet_fn(unsigned long arg)
+{
+	struct tmio_mmc_host *host = (struct tmio_mmc_host *)arg;
+
+	spin_lock_irq(&host->lock);
+	if (!renesas_sdhi_internal_dmac_complete(host))
+		goto out;
+
+	tmio_mmc_do_data_irq(host);
+out:
+	spin_unlock_irq(&host->lock);
+}
+
+static void renesas_sdhi_internal_dmac_end_dma(struct tmio_mmc_host *host)
+{
+	if (host->data)
+		renesas_sdhi_internal_dmac_complete(host);
+}
+
+static void
+renesas_sdhi_internal_dmac_request_dma(struct tmio_mmc_host *host,
+				       struct tmio_mmc_data *pdata)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+
+	/* Disable DMAC interrupts, we don't use them */
+	renesas_sdhi_internal_dmac_dm_write(host, DM_CM_INFO1_MASK,
+					    INFO1_MASK_CLEAR);
+	renesas_sdhi_internal_dmac_dm_write(host, DM_CM_INFO2_MASK,
+					    INFO2_MASK_CLEAR);
+
+	/* Each value is set to non-zero to assume "enabling" each DMA */
+	host->chan_rx = host->chan_tx = (void *)0xdeadbeaf;
+
+	tasklet_init(&priv->dma_priv.dma_complete,
+		     renesas_sdhi_internal_dmac_complete_tasklet_fn,
+		     (unsigned long)host);
+	tasklet_init(&host->dma_issue,
+		     renesas_sdhi_internal_dmac_issue_tasklet_fn,
+		     (unsigned long)host);
+}
+
+static void
+renesas_sdhi_internal_dmac_release_dma(struct tmio_mmc_host *host)
+{
+	/* Each value is set to zero to assume "disabling" each DMA */
+	host->chan_rx = host->chan_tx = NULL;
+}
+
+static const struct tmio_mmc_dma_ops renesas_sdhi_internal_dmac_dma_ops = {
+	.start = renesas_sdhi_internal_dmac_start_dma,
+	.enable = renesas_sdhi_internal_dmac_enable_dma,
+	.request = renesas_sdhi_internal_dmac_request_dma,
+	.release = renesas_sdhi_internal_dmac_release_dma,
+	.abort = renesas_sdhi_internal_dmac_abort_dma,
+	.dataend = renesas_sdhi_internal_dmac_dataend_dma,
+	.end = renesas_sdhi_internal_dmac_end_dma,
+};
+
+/*
+ * Whitelist of specific R-Car Gen3 SoC ES versions to use this DMAC
+ * implementation as others may use a different implementation.
+ */
+static const struct soc_device_attribute soc_dma_quirks[] = {
+	{ .soc_id = "r7s9210",
+	  .data = (void *)BIT(SDHI_INTERNAL_DMAC_ADDR_MODE_FIXED_ONLY) },
+	{ .soc_id = "r8a7795", .revision = "ES1.*",
+	  .data = (void *)BIT(SDHI_INTERNAL_DMAC_ONE_RX_ONLY) },
+	{ .soc_id = "r8a7796", .revision = "ES1.0",
+	  .data = (void *)BIT(SDHI_INTERNAL_DMAC_ONE_RX_ONLY) },
+	{ /* sentinel */ }
+};
+
+static int renesas_sdhi_internal_dmac_probe(struct platform_device *pdev)
+{
+	const struct soc_device_attribute *soc = soc_device_match(soc_dma_quirks);
+	struct device *dev = &pdev->dev;
+
+	if (soc)
+		global_flags |= (unsigned long)soc->data;
+
+	/* value is max of SD_SECCNT. Confirmed by HW engineers */
+	dma_set_max_seg_size(dev, 0xffffffff);
+
+	return renesas_sdhi_probe(pdev, &renesas_sdhi_internal_dmac_dma_ops);
+}
+
+static const struct dev_pm_ops renesas_sdhi_internal_dmac_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(tmio_mmc_host_runtime_suspend,
+			   tmio_mmc_host_runtime_resume,
+			   NULL)
+};
+
+static struct platform_driver renesas_internal_dmac_sdhi_driver = {
+	.driver		= {
+		.name	= "renesas_sdhi_internal_dmac",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= &renesas_sdhi_internal_dmac_dev_pm_ops,
+		.of_match_table = renesas_sdhi_internal_dmac_of_match,
+	},
+	.probe		= renesas_sdhi_internal_dmac_probe,
+	.remove		= renesas_sdhi_remove,
+};
+
+module_platform_driver(renesas_internal_dmac_sdhi_driver);
+
+MODULE_DESCRIPTION("Renesas SDHI driver for internal DMAC");
+MODULE_AUTHOR("Yoshihiro Shimoda");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/renesas_sdhi_sys_dmac.c b/drivers/mmc/host/renesas_sdhi_sys_dmac.c
new file mode 100644
index 000000000..c5f789675
--- /dev/null
+++ b/drivers/mmc/host/renesas_sdhi_sys_dmac.c
@@ -0,0 +1,479 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * DMA support use of SYS DMAC with SDHI SD/SDIO controller
+ *
+ * Copyright (C) 2016-19 Renesas Electronics Corporation
+ * Copyright (C) 2016-19 Sang Engineering, Wolfram Sang
+ * Copyright (C) 2017 Horms Solutions, Simon Horman
+ * Copyright (C) 2010-2011 Guennadi Liakhovetski
+ */
+
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/dmaengine.h>
+#include <linux/mfd/tmio.h>
+#include <linux/mmc/host.h>
+#include <linux/mod_devicetable.h>
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/pagemap.h>
+#include <linux/scatterlist.h>
+#include <linux/sys_soc.h>
+
+#include "renesas_sdhi.h"
+#include "tmio_mmc.h"
+
+#define TMIO_MMC_MIN_DMA_LEN 8
+
+static const struct renesas_sdhi_of_data of_default_cfg = {
+	.tmio_flags = TMIO_MMC_HAS_IDLE_WAIT,
+};
+
+static const struct renesas_sdhi_of_data of_rz_compatible = {
+	.tmio_flags	= TMIO_MMC_HAS_IDLE_WAIT | TMIO_MMC_32BIT_DATA_PORT |
+			  TMIO_MMC_HAVE_CBSY,
+	.tmio_ocr_mask	= MMC_VDD_32_33,
+	.capabilities	= MMC_CAP_SD_HIGHSPEED | MMC_CAP_SDIO_IRQ,
+};
+
+static const struct renesas_sdhi_of_data of_rcar_gen1_compatible = {
+	.tmio_flags	= TMIO_MMC_HAS_IDLE_WAIT | TMIO_MMC_CLK_ACTUAL,
+	.capabilities	= MMC_CAP_SD_HIGHSPEED | MMC_CAP_SDIO_IRQ,
+	.capabilities2	= MMC_CAP2_NO_WRITE_PROTECT,
+};
+
+/* Definitions for sampling clocks */
+static struct renesas_sdhi_scc rcar_gen2_scc_taps[] = {
+	{
+		.clk_rate = 156000000,
+		.tap = 0x00000703,
+	},
+	{
+		.clk_rate = 0,
+		.tap = 0x00000300,
+	},
+};
+
+static const struct renesas_sdhi_of_data of_rcar_gen2_compatible = {
+	.tmio_flags	= TMIO_MMC_HAS_IDLE_WAIT | TMIO_MMC_CLK_ACTUAL |
+			  TMIO_MMC_HAVE_CBSY | TMIO_MMC_MIN_RCAR2,
+	.capabilities	= MMC_CAP_SD_HIGHSPEED | MMC_CAP_SDIO_IRQ |
+			  MMC_CAP_CMD23,
+	.capabilities2	= MMC_CAP2_NO_WRITE_PROTECT,
+	.dma_buswidth	= DMA_SLAVE_BUSWIDTH_4_BYTES,
+	.dma_rx_offset	= 0x2000,
+	.scc_offset	= 0x0300,
+	.taps		= rcar_gen2_scc_taps,
+	.taps_num	= ARRAY_SIZE(rcar_gen2_scc_taps),
+	.max_blk_count	= UINT_MAX / TMIO_MAX_BLK_SIZE,
+};
+
+static const struct of_device_id renesas_sdhi_sys_dmac_of_match[] = {
+	{ .compatible = "renesas,sdhi-sh73a0", .data = &of_default_cfg, },
+	{ .compatible = "renesas,sdhi-r8a73a4", .data = &of_default_cfg, },
+	{ .compatible = "renesas,sdhi-r8a7740", .data = &of_default_cfg, },
+	{ .compatible = "renesas,sdhi-r7s72100", .data = &of_rz_compatible, },
+	{ .compatible = "renesas,sdhi-r8a7778", .data = &of_rcar_gen1_compatible, },
+	{ .compatible = "renesas,sdhi-r8a7779", .data = &of_rcar_gen1_compatible, },
+	{ .compatible = "renesas,sdhi-r8a7743", .data = &of_rcar_gen2_compatible, },
+	{ .compatible = "renesas,sdhi-r8a7745", .data = &of_rcar_gen2_compatible, },
+	{ .compatible = "renesas,sdhi-r8a7790", .data = &of_rcar_gen2_compatible, },
+	{ .compatible = "renesas,sdhi-r8a7791", .data = &of_rcar_gen2_compatible, },
+	{ .compatible = "renesas,sdhi-r8a7792", .data = &of_rcar_gen2_compatible, },
+	{ .compatible = "renesas,sdhi-r8a7793", .data = &of_rcar_gen2_compatible, },
+	{ .compatible = "renesas,sdhi-r8a7794", .data = &of_rcar_gen2_compatible, },
+	{ .compatible = "renesas,rcar-gen1-sdhi", .data = &of_rcar_gen1_compatible, },
+	{ .compatible = "renesas,rcar-gen2-sdhi", .data = &of_rcar_gen2_compatible, },
+	{ .compatible = "renesas,sdhi-shmobile" },
+	{},
+};
+MODULE_DEVICE_TABLE(of, renesas_sdhi_sys_dmac_of_match);
+
+static void renesas_sdhi_sys_dmac_enable_dma(struct tmio_mmc_host *host,
+					     bool enable)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+
+	if (!host->chan_tx || !host->chan_rx)
+		return;
+
+	if (priv->dma_priv.enable)
+		priv->dma_priv.enable(host, enable);
+}
+
+static void renesas_sdhi_sys_dmac_abort_dma(struct tmio_mmc_host *host)
+{
+	renesas_sdhi_sys_dmac_enable_dma(host, false);
+
+	if (host->chan_rx)
+		dmaengine_terminate_all(host->chan_rx);
+	if (host->chan_tx)
+		dmaengine_terminate_all(host->chan_tx);
+
+	renesas_sdhi_sys_dmac_enable_dma(host, true);
+}
+
+static void renesas_sdhi_sys_dmac_dataend_dma(struct tmio_mmc_host *host)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+
+	complete(&priv->dma_priv.dma_dataend);
+}
+
+static void renesas_sdhi_sys_dmac_dma_callback(void *arg)
+{
+	struct tmio_mmc_host *host = arg;
+	struct renesas_sdhi *priv = host_to_priv(host);
+
+	spin_lock_irq(&host->lock);
+
+	if (!host->data)
+		goto out;
+
+	if (host->data->flags & MMC_DATA_READ)
+		dma_unmap_sg(host->chan_rx->device->dev,
+			     host->sg_ptr, host->sg_len,
+			     DMA_FROM_DEVICE);
+	else
+		dma_unmap_sg(host->chan_tx->device->dev,
+			     host->sg_ptr, host->sg_len,
+			     DMA_TO_DEVICE);
+
+	spin_unlock_irq(&host->lock);
+
+	wait_for_completion(&priv->dma_priv.dma_dataend);
+
+	spin_lock_irq(&host->lock);
+	tmio_mmc_do_data_irq(host);
+out:
+	spin_unlock_irq(&host->lock);
+}
+
+static void renesas_sdhi_sys_dmac_start_dma_rx(struct tmio_mmc_host *host)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+	struct scatterlist *sg = host->sg_ptr, *sg_tmp;
+	struct dma_async_tx_descriptor *desc = NULL;
+	struct dma_chan *chan = host->chan_rx;
+	dma_cookie_t cookie;
+	int ret, i;
+	bool aligned = true, multiple = true;
+	unsigned int align = (1 << host->pdata->alignment_shift) - 1;
+
+	for_each_sg(sg, sg_tmp, host->sg_len, i) {
+		if (sg_tmp->offset & align)
+			aligned = false;
+		if (sg_tmp->length & align) {
+			multiple = false;
+			break;
+		}
+	}
+
+	if ((!aligned && (host->sg_len > 1 || sg->length > PAGE_SIZE ||
+			  (align & PAGE_MASK))) || !multiple) {
+		ret = -EINVAL;
+		goto pio;
+	}
+
+	if (sg->length < TMIO_MMC_MIN_DMA_LEN)
+		return;
+
+	/* The only sg element can be unaligned, use our bounce buffer then */
+	if (!aligned) {
+		sg_init_one(&host->bounce_sg, host->bounce_buf, sg->length);
+		host->sg_ptr = &host->bounce_sg;
+		sg = host->sg_ptr;
+	}
+
+	ret = dma_map_sg(chan->device->dev, sg, host->sg_len, DMA_FROM_DEVICE);
+	if (ret > 0)
+		desc = dmaengine_prep_slave_sg(chan, sg, ret, DMA_DEV_TO_MEM,
+					       DMA_CTRL_ACK);
+
+	if (desc) {
+		reinit_completion(&priv->dma_priv.dma_dataend);
+		desc->callback = renesas_sdhi_sys_dmac_dma_callback;
+		desc->callback_param = host;
+
+		cookie = dmaengine_submit(desc);
+		if (cookie < 0) {
+			desc = NULL;
+			ret = cookie;
+		}
+		host->dma_on = true;
+	}
+pio:
+	if (!desc) {
+		/* DMA failed, fall back to PIO */
+		renesas_sdhi_sys_dmac_enable_dma(host, false);
+		if (ret >= 0)
+			ret = -EIO;
+		host->chan_rx = NULL;
+		dma_release_channel(chan);
+		/* Free the Tx channel too */
+		chan = host->chan_tx;
+		if (chan) {
+			host->chan_tx = NULL;
+			dma_release_channel(chan);
+		}
+		dev_warn(&host->pdev->dev,
+			 "DMA failed: %d, falling back to PIO\n", ret);
+	}
+}
+
+static void renesas_sdhi_sys_dmac_start_dma_tx(struct tmio_mmc_host *host)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+	struct scatterlist *sg = host->sg_ptr, *sg_tmp;
+	struct dma_async_tx_descriptor *desc = NULL;
+	struct dma_chan *chan = host->chan_tx;
+	dma_cookie_t cookie;
+	int ret, i;
+	bool aligned = true, multiple = true;
+	unsigned int align = (1 << host->pdata->alignment_shift) - 1;
+
+	for_each_sg(sg, sg_tmp, host->sg_len, i) {
+		if (sg_tmp->offset & align)
+			aligned = false;
+		if (sg_tmp->length & align) {
+			multiple = false;
+			break;
+		}
+	}
+
+	if ((!aligned && (host->sg_len > 1 || sg->length > PAGE_SIZE ||
+			  (align & PAGE_MASK))) || !multiple) {
+		ret = -EINVAL;
+		goto pio;
+	}
+
+	if (sg->length < TMIO_MMC_MIN_DMA_LEN)
+		return;
+
+	/* The only sg element can be unaligned, use our bounce buffer then */
+	if (!aligned) {
+		unsigned long flags;
+		void *sg_vaddr = tmio_mmc_kmap_atomic(sg, &flags);
+
+		sg_init_one(&host->bounce_sg, host->bounce_buf, sg->length);
+		memcpy(host->bounce_buf, sg_vaddr, host->bounce_sg.length);
+		tmio_mmc_kunmap_atomic(sg, &flags, sg_vaddr);
+		host->sg_ptr = &host->bounce_sg;
+		sg = host->sg_ptr;
+	}
+
+	ret = dma_map_sg(chan->device->dev, sg, host->sg_len, DMA_TO_DEVICE);
+	if (ret > 0)
+		desc = dmaengine_prep_slave_sg(chan, sg, ret, DMA_MEM_TO_DEV,
+					       DMA_CTRL_ACK);
+
+	if (desc) {
+		reinit_completion(&priv->dma_priv.dma_dataend);
+		desc->callback = renesas_sdhi_sys_dmac_dma_callback;
+		desc->callback_param = host;
+
+		cookie = dmaengine_submit(desc);
+		if (cookie < 0) {
+			desc = NULL;
+			ret = cookie;
+		}
+		host->dma_on = true;
+	}
+pio:
+	if (!desc) {
+		/* DMA failed, fall back to PIO */
+		renesas_sdhi_sys_dmac_enable_dma(host, false);
+		if (ret >= 0)
+			ret = -EIO;
+		host->chan_tx = NULL;
+		dma_release_channel(chan);
+		/* Free the Rx channel too */
+		chan = host->chan_rx;
+		if (chan) {
+			host->chan_rx = NULL;
+			dma_release_channel(chan);
+		}
+		dev_warn(&host->pdev->dev,
+			 "DMA failed: %d, falling back to PIO\n", ret);
+	}
+}
+
+static void renesas_sdhi_sys_dmac_start_dma(struct tmio_mmc_host *host,
+					    struct mmc_data *data)
+{
+	if (data->flags & MMC_DATA_READ) {
+		if (host->chan_rx)
+			renesas_sdhi_sys_dmac_start_dma_rx(host);
+	} else {
+		if (host->chan_tx)
+			renesas_sdhi_sys_dmac_start_dma_tx(host);
+	}
+}
+
+static void renesas_sdhi_sys_dmac_issue_tasklet_fn(unsigned long priv)
+{
+	struct tmio_mmc_host *host = (struct tmio_mmc_host *)priv;
+	struct dma_chan *chan = NULL;
+
+	spin_lock_irq(&host->lock);
+
+	if (host->data) {
+		if (host->data->flags & MMC_DATA_READ)
+			chan = host->chan_rx;
+		else
+			chan = host->chan_tx;
+	}
+
+	spin_unlock_irq(&host->lock);
+
+	tmio_mmc_enable_mmc_irqs(host, TMIO_STAT_DATAEND);
+
+	if (chan)
+		dma_async_issue_pending(chan);
+}
+
+static void renesas_sdhi_sys_dmac_request_dma(struct tmio_mmc_host *host,
+					      struct tmio_mmc_data *pdata)
+{
+	struct renesas_sdhi *priv = host_to_priv(host);
+
+	/* We can only either use DMA for both Tx and Rx or not use it at all */
+	if (!host->pdev->dev.of_node &&
+	    (!pdata->chan_priv_tx || !pdata->chan_priv_rx))
+		return;
+
+	if (!host->chan_tx && !host->chan_rx) {
+		struct resource *res = platform_get_resource(host->pdev,
+							     IORESOURCE_MEM, 0);
+		struct dma_slave_config cfg = {};
+		dma_cap_mask_t mask;
+		int ret;
+
+		if (!res)
+			return;
+
+		dma_cap_zero(mask);
+		dma_cap_set(DMA_SLAVE, mask);
+
+		host->chan_tx = dma_request_slave_channel_compat(mask,
+					priv->dma_priv.filter, pdata->chan_priv_tx,
+					&host->pdev->dev, "tx");
+		dev_dbg(&host->pdev->dev, "%s: TX: got channel %p\n", __func__,
+			host->chan_tx);
+
+		if (!host->chan_tx)
+			return;
+
+		cfg.direction = DMA_MEM_TO_DEV;
+		cfg.dst_addr = res->start +
+			(CTL_SD_DATA_PORT << host->bus_shift);
+		cfg.dst_addr_width = priv->dma_priv.dma_buswidth;
+		if (!cfg.dst_addr_width)
+			cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES;
+		cfg.src_addr = 0;
+		ret = dmaengine_slave_config(host->chan_tx, &cfg);
+		if (ret < 0)
+			goto ecfgtx;
+
+		host->chan_rx = dma_request_slave_channel_compat(mask,
+					priv->dma_priv.filter, pdata->chan_priv_rx,
+					&host->pdev->dev, "rx");
+		dev_dbg(&host->pdev->dev, "%s: RX: got channel %p\n", __func__,
+			host->chan_rx);
+
+		if (!host->chan_rx)
+			goto ereqrx;
+
+		cfg.direction = DMA_DEV_TO_MEM;
+		cfg.src_addr = cfg.dst_addr + host->pdata->dma_rx_offset;
+		cfg.src_addr_width = priv->dma_priv.dma_buswidth;
+		if (!cfg.src_addr_width)
+			cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES;
+		cfg.dst_addr = 0;
+		ret = dmaengine_slave_config(host->chan_rx, &cfg);
+		if (ret < 0)
+			goto ecfgrx;
+
+		host->bounce_buf = (u8 *)__get_free_page(GFP_KERNEL | GFP_DMA);
+		if (!host->bounce_buf)
+			goto ebouncebuf;
+
+		init_completion(&priv->dma_priv.dma_dataend);
+		tasklet_init(&host->dma_issue,
+			     renesas_sdhi_sys_dmac_issue_tasklet_fn,
+			     (unsigned long)host);
+	}
+
+	renesas_sdhi_sys_dmac_enable_dma(host, true);
+
+	return;
+
+ebouncebuf:
+ecfgrx:
+	dma_release_channel(host->chan_rx);
+	host->chan_rx = NULL;
+ereqrx:
+ecfgtx:
+	dma_release_channel(host->chan_tx);
+	host->chan_tx = NULL;
+}
+
+static void renesas_sdhi_sys_dmac_release_dma(struct tmio_mmc_host *host)
+{
+	if (host->chan_tx) {
+		struct dma_chan *chan = host->chan_tx;
+
+		host->chan_tx = NULL;
+		dma_release_channel(chan);
+	}
+	if (host->chan_rx) {
+		struct dma_chan *chan = host->chan_rx;
+
+		host->chan_rx = NULL;
+		dma_release_channel(chan);
+	}
+	if (host->bounce_buf) {
+		free_pages((unsigned long)host->bounce_buf, 0);
+		host->bounce_buf = NULL;
+	}
+}
+
+static const struct tmio_mmc_dma_ops renesas_sdhi_sys_dmac_dma_ops = {
+	.start = renesas_sdhi_sys_dmac_start_dma,
+	.enable = renesas_sdhi_sys_dmac_enable_dma,
+	.request = renesas_sdhi_sys_dmac_request_dma,
+	.release = renesas_sdhi_sys_dmac_release_dma,
+	.abort = renesas_sdhi_sys_dmac_abort_dma,
+	.dataend = renesas_sdhi_sys_dmac_dataend_dma,
+};
+
+static int renesas_sdhi_sys_dmac_probe(struct platform_device *pdev)
+{
+	return renesas_sdhi_probe(pdev, &renesas_sdhi_sys_dmac_dma_ops);
+}
+
+static const struct dev_pm_ops renesas_sdhi_sys_dmac_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(tmio_mmc_host_runtime_suspend,
+			   tmio_mmc_host_runtime_resume,
+			   NULL)
+};
+
+static struct platform_driver renesas_sys_dmac_sdhi_driver = {
+	.driver		= {
+		.name	= "sh_mobile_sdhi",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= &renesas_sdhi_sys_dmac_dev_pm_ops,
+		.of_match_table = renesas_sdhi_sys_dmac_of_match,
+	},
+	.probe		= renesas_sdhi_sys_dmac_probe,
+	.remove		= renesas_sdhi_remove,
+};
+
+module_platform_driver(renesas_sys_dmac_sdhi_driver);
+
+MODULE_DESCRIPTION("Renesas SDHI driver");
+MODULE_AUTHOR("Magnus Damm");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:sh_mobile_sdhi");
diff --git a/drivers/mmc/host/rtsx_pci_sdmmc.c b/drivers/mmc/host/rtsx_pci_sdmmc.c
new file mode 100644
index 000000000..b5cb83bb9
--- /dev/null
+++ b/drivers/mmc/host/rtsx_pci_sdmmc.c
@@ -0,0 +1,1502 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Realtek PCI-Express SD/MMC Card Interface driver
+ *
+ * Copyright(c) 2009-2013 Realtek Semiconductor Corp. All rights reserved.
+ *
+ * Author:
+ *   Wei WANG <wei_wang@realsil.com.cn>
+ */
+
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/highmem.h>
+#include <linux/delay.h>
+#include <linux/platform_device.h>
+#include <linux/workqueue.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sd.h>
+#include <linux/mmc/sdio.h>
+#include <linux/mmc/card.h>
+#include <linux/rtsx_pci.h>
+#include <asm/unaligned.h>
+
+struct realtek_pci_sdmmc {
+	struct platform_device	*pdev;
+	struct rtsx_pcr		*pcr;
+	struct mmc_host		*mmc;
+	struct mmc_request	*mrq;
+#define SDMMC_WORKQ_NAME	"rtsx_pci_sdmmc_workq"
+
+	struct work_struct	work;
+	struct mutex		host_mutex;
+
+	u8			ssc_depth;
+	unsigned int		clock;
+	bool			vpclk;
+	bool			double_clk;
+	bool			eject;
+	bool			initial_mode;
+	int			prev_power_state;
+	int			sg_count;
+	s32			cookie;
+	int			cookie_sg_count;
+	bool			using_cookie;
+};
+
+static inline struct device *sdmmc_dev(struct realtek_pci_sdmmc *host)
+{
+	return &(host->pdev->dev);
+}
+
+static inline void sd_clear_error(struct realtek_pci_sdmmc *host)
+{
+	rtsx_pci_write_register(host->pcr, CARD_STOP,
+			SD_STOP | SD_CLR_ERR, SD_STOP | SD_CLR_ERR);
+}
+
+#ifdef DEBUG
+static void dump_reg_range(struct realtek_pci_sdmmc *host, u16 start, u16 end)
+{
+	u16 len = end - start + 1;
+	int i;
+	u8 data[8];
+
+	for (i = 0; i < len; i += 8) {
+		int j;
+		int n = min(8, len - i);
+
+		memset(&data, 0, sizeof(data));
+		for (j = 0; j < n; j++)
+			rtsx_pci_read_register(host->pcr, start + i + j,
+				data + j);
+		dev_dbg(sdmmc_dev(host), "0x%04X(%d): %8ph\n",
+			start + i, n, data);
+	}
+}
+
+static void sd_print_debug_regs(struct realtek_pci_sdmmc *host)
+{
+	dump_reg_range(host, 0xFDA0, 0xFDB3);
+	dump_reg_range(host, 0xFD52, 0xFD69);
+}
+#else
+#define sd_print_debug_regs(host)
+#endif /* DEBUG */
+
+static inline int sd_get_cd_int(struct realtek_pci_sdmmc *host)
+{
+	return rtsx_pci_readl(host->pcr, RTSX_BIPR) & SD_EXIST;
+}
+
+static void sd_cmd_set_sd_cmd(struct rtsx_pcr *pcr, struct mmc_command *cmd)
+{
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_CMD0, 0xFF,
+		SD_CMD_START | cmd->opcode);
+	rtsx_pci_write_be32(pcr, SD_CMD1, cmd->arg);
+}
+
+static void sd_cmd_set_data_len(struct rtsx_pcr *pcr, u16 blocks, u16 blksz)
+{
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_BLOCK_CNT_L, 0xFF, blocks);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_BLOCK_CNT_H, 0xFF, blocks >> 8);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_BYTE_CNT_L, 0xFF, blksz);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_BYTE_CNT_H, 0xFF, blksz >> 8);
+}
+
+static int sd_response_type(struct mmc_command *cmd)
+{
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE:
+		return SD_RSP_TYPE_R0;
+	case MMC_RSP_R1:
+		return SD_RSP_TYPE_R1;
+	case MMC_RSP_R1_NO_CRC:
+		return SD_RSP_TYPE_R1 | SD_NO_CHECK_CRC7;
+	case MMC_RSP_R1B:
+		return SD_RSP_TYPE_R1b;
+	case MMC_RSP_R2:
+		return SD_RSP_TYPE_R2;
+	case MMC_RSP_R3:
+		return SD_RSP_TYPE_R3;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int sd_status_index(int resp_type)
+{
+	if (resp_type == SD_RSP_TYPE_R0)
+		return 0;
+	else if (resp_type == SD_RSP_TYPE_R2)
+		return 16;
+
+	return 5;
+}
+/*
+ * sd_pre_dma_transfer - do dma_map_sg() or using cookie
+ *
+ * @pre: if called in pre_req()
+ * return:
+ *	0 - do dma_map_sg()
+ *	1 - using cookie
+ */
+static int sd_pre_dma_transfer(struct realtek_pci_sdmmc *host,
+		struct mmc_data *data, bool pre)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	int read = data->flags & MMC_DATA_READ;
+	int count = 0;
+	int using_cookie = 0;
+
+	if (!pre && data->host_cookie && data->host_cookie != host->cookie) {
+		dev_err(sdmmc_dev(host),
+			"error: data->host_cookie = %d, host->cookie = %d\n",
+			data->host_cookie, host->cookie);
+		data->host_cookie = 0;
+	}
+
+	if (pre || data->host_cookie != host->cookie) {
+		count = rtsx_pci_dma_map_sg(pcr, data->sg, data->sg_len, read);
+	} else {
+		count = host->cookie_sg_count;
+		using_cookie = 1;
+	}
+
+	if (pre) {
+		host->cookie_sg_count = count;
+		if (++host->cookie < 0)
+			host->cookie = 1;
+		data->host_cookie = host->cookie;
+	} else {
+		host->sg_count = count;
+	}
+
+	return using_cookie;
+}
+
+static void sdmmc_pre_req(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct realtek_pci_sdmmc *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+
+	if (data->host_cookie) {
+		dev_err(sdmmc_dev(host),
+			"error: reset data->host_cookie = %d\n",
+			data->host_cookie);
+		data->host_cookie = 0;
+	}
+
+	sd_pre_dma_transfer(host, data, true);
+	dev_dbg(sdmmc_dev(host), "pre dma sg: %d\n", host->cookie_sg_count);
+}
+
+static void sdmmc_post_req(struct mmc_host *mmc, struct mmc_request *mrq,
+		int err)
+{
+	struct realtek_pci_sdmmc *host = mmc_priv(mmc);
+	struct rtsx_pcr *pcr = host->pcr;
+	struct mmc_data *data = mrq->data;
+	int read = data->flags & MMC_DATA_READ;
+
+	rtsx_pci_dma_unmap_sg(pcr, data->sg, data->sg_len, read);
+	data->host_cookie = 0;
+}
+
+static void sd_send_cmd_get_rsp(struct realtek_pci_sdmmc *host,
+		struct mmc_command *cmd)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	u8 cmd_idx = (u8)cmd->opcode;
+	u32 arg = cmd->arg;
+	int err = 0;
+	int timeout = 100;
+	int i;
+	u8 *ptr;
+	int rsp_type;
+	int stat_idx;
+	bool clock_toggled = false;
+
+	dev_dbg(sdmmc_dev(host), "%s: SD/MMC CMD %d, arg = 0x%08x\n",
+			__func__, cmd_idx, arg);
+
+	rsp_type = sd_response_type(cmd);
+	if (rsp_type < 0)
+		goto out;
+
+	stat_idx = sd_status_index(rsp_type);
+
+	if (rsp_type == SD_RSP_TYPE_R1b)
+		timeout = cmd->busy_timeout ? cmd->busy_timeout : 3000;
+
+	if (cmd->opcode == SD_SWITCH_VOLTAGE) {
+		err = rtsx_pci_write_register(pcr, SD_BUS_STAT,
+				0xFF, SD_CLK_TOGGLE_EN);
+		if (err < 0)
+			goto out;
+
+		clock_toggled = true;
+	}
+
+	rtsx_pci_init_cmd(pcr);
+	sd_cmd_set_sd_cmd(pcr, cmd);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_CFG2, 0xFF, rsp_type);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CARD_DATA_SOURCE,
+			0x01, PINGPONG_BUFFER);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_TRANSFER,
+			0xFF, SD_TM_CMD_RSP | SD_TRANSFER_START);
+	rtsx_pci_add_cmd(pcr, CHECK_REG_CMD, SD_TRANSFER,
+		     SD_TRANSFER_END | SD_STAT_IDLE,
+		     SD_TRANSFER_END | SD_STAT_IDLE);
+
+	if (rsp_type == SD_RSP_TYPE_R2) {
+		/* Read data from ping-pong buffer */
+		for (i = PPBUF_BASE2; i < PPBUF_BASE2 + 16; i++)
+			rtsx_pci_add_cmd(pcr, READ_REG_CMD, (u16)i, 0, 0);
+	} else if (rsp_type != SD_RSP_TYPE_R0) {
+		/* Read data from SD_CMDx registers */
+		for (i = SD_CMD0; i <= SD_CMD4; i++)
+			rtsx_pci_add_cmd(pcr, READ_REG_CMD, (u16)i, 0, 0);
+	}
+
+	rtsx_pci_add_cmd(pcr, READ_REG_CMD, SD_STAT1, 0, 0);
+
+	err = rtsx_pci_send_cmd(pcr, timeout);
+	if (err < 0) {
+		sd_print_debug_regs(host);
+		sd_clear_error(host);
+		dev_dbg(sdmmc_dev(host),
+			"rtsx_pci_send_cmd error (err = %d)\n", err);
+		goto out;
+	}
+
+	if (rsp_type == SD_RSP_TYPE_R0) {
+		err = 0;
+		goto out;
+	}
+
+	/* Eliminate returned value of CHECK_REG_CMD */
+	ptr = rtsx_pci_get_cmd_data(pcr) + 1;
+
+	/* Check (Start,Transmission) bit of Response */
+	if ((ptr[0] & 0xC0) != 0) {
+		err = -EILSEQ;
+		dev_dbg(sdmmc_dev(host), "Invalid response bit\n");
+		goto out;
+	}
+
+	/* Check CRC7 */
+	if (!(rsp_type & SD_NO_CHECK_CRC7)) {
+		if (ptr[stat_idx] & SD_CRC7_ERR) {
+			err = -EILSEQ;
+			dev_dbg(sdmmc_dev(host), "CRC7 error\n");
+			goto out;
+		}
+	}
+
+	if (rsp_type == SD_RSP_TYPE_R2) {
+		/*
+		 * The controller offloads the last byte {CRC-7, end bit 1'b1}
+		 * of response type R2. Assign dummy CRC, 0, and end bit to the
+		 * byte(ptr[16], goes into the LSB of resp[3] later).
+		 */
+		ptr[16] = 1;
+
+		for (i = 0; i < 4; i++) {
+			cmd->resp[i] = get_unaligned_be32(ptr + 1 + i * 4);
+			dev_dbg(sdmmc_dev(host), "cmd->resp[%d] = 0x%08x\n",
+					i, cmd->resp[i]);
+		}
+	} else {
+		cmd->resp[0] = get_unaligned_be32(ptr + 1);
+		dev_dbg(sdmmc_dev(host), "cmd->resp[0] = 0x%08x\n",
+				cmd->resp[0]);
+	}
+
+out:
+	cmd->error = err;
+
+	if (err && clock_toggled)
+		rtsx_pci_write_register(pcr, SD_BUS_STAT,
+				SD_CLK_TOGGLE_EN | SD_CLK_FORCE_STOP, 0);
+}
+
+static int sd_read_data(struct realtek_pci_sdmmc *host, struct mmc_command *cmd,
+	u16 byte_cnt, u8 *buf, int buf_len, int timeout)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	int err;
+	u8 trans_mode;
+
+	dev_dbg(sdmmc_dev(host), "%s: SD/MMC CMD %d, arg = 0x%08x\n",
+		__func__, cmd->opcode, cmd->arg);
+
+	if (!buf)
+		buf_len = 0;
+
+	if (cmd->opcode == MMC_SEND_TUNING_BLOCK)
+		trans_mode = SD_TM_AUTO_TUNING;
+	else
+		trans_mode = SD_TM_NORMAL_READ;
+
+	rtsx_pci_init_cmd(pcr);
+	sd_cmd_set_sd_cmd(pcr, cmd);
+	sd_cmd_set_data_len(pcr, 1, byte_cnt);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_CFG2, 0xFF,
+			SD_CALCULATE_CRC7 | SD_CHECK_CRC16 |
+			SD_NO_WAIT_BUSY_END | SD_CHECK_CRC7 | SD_RSP_LEN_6);
+	if (trans_mode != SD_TM_AUTO_TUNING)
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD,
+				CARD_DATA_SOURCE, 0x01, PINGPONG_BUFFER);
+
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_TRANSFER,
+			0xFF, trans_mode | SD_TRANSFER_START);
+	rtsx_pci_add_cmd(pcr, CHECK_REG_CMD, SD_TRANSFER,
+			SD_TRANSFER_END, SD_TRANSFER_END);
+
+	err = rtsx_pci_send_cmd(pcr, timeout);
+	if (err < 0) {
+		sd_print_debug_regs(host);
+		dev_dbg(sdmmc_dev(host),
+			"rtsx_pci_send_cmd fail (err = %d)\n", err);
+		return err;
+	}
+
+	if (buf && buf_len) {
+		err = rtsx_pci_read_ppbuf(pcr, buf, buf_len);
+		if (err < 0) {
+			dev_dbg(sdmmc_dev(host),
+				"rtsx_pci_read_ppbuf fail (err = %d)\n", err);
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+static int sd_write_data(struct realtek_pci_sdmmc *host,
+	struct mmc_command *cmd, u16 byte_cnt, u8 *buf, int buf_len,
+	int timeout)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	int err;
+
+	dev_dbg(sdmmc_dev(host), "%s: SD/MMC CMD %d, arg = 0x%08x\n",
+		__func__, cmd->opcode, cmd->arg);
+
+	if (!buf)
+		buf_len = 0;
+
+	sd_send_cmd_get_rsp(host, cmd);
+	if (cmd->error)
+		return cmd->error;
+
+	if (buf && buf_len) {
+		err = rtsx_pci_write_ppbuf(pcr, buf, buf_len);
+		if (err < 0) {
+			dev_dbg(sdmmc_dev(host),
+				"rtsx_pci_write_ppbuf fail (err = %d)\n", err);
+			return err;
+		}
+	}
+
+	rtsx_pci_init_cmd(pcr);
+	sd_cmd_set_data_len(pcr, 1, byte_cnt);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_CFG2, 0xFF,
+		SD_CALCULATE_CRC7 | SD_CHECK_CRC16 |
+		SD_NO_WAIT_BUSY_END | SD_CHECK_CRC7 | SD_RSP_LEN_0);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_TRANSFER, 0xFF,
+			SD_TRANSFER_START | SD_TM_AUTO_WRITE_3);
+	rtsx_pci_add_cmd(pcr, CHECK_REG_CMD, SD_TRANSFER,
+			SD_TRANSFER_END, SD_TRANSFER_END);
+
+	err = rtsx_pci_send_cmd(pcr, timeout);
+	if (err < 0) {
+		sd_print_debug_regs(host);
+		dev_dbg(sdmmc_dev(host),
+			"rtsx_pci_send_cmd fail (err = %d)\n", err);
+		return err;
+	}
+
+	return 0;
+}
+
+static int sd_read_long_data(struct realtek_pci_sdmmc *host,
+	struct mmc_request *mrq)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	struct mmc_host *mmc = host->mmc;
+	struct mmc_card *card = mmc->card;
+	struct mmc_command *cmd = mrq->cmd;
+	struct mmc_data *data = mrq->data;
+	int uhs = mmc_card_uhs(card);
+	u8 cfg2 = 0;
+	int err;
+	int resp_type;
+	size_t data_len = data->blksz * data->blocks;
+
+	dev_dbg(sdmmc_dev(host), "%s: SD/MMC CMD %d, arg = 0x%08x\n",
+		__func__, cmd->opcode, cmd->arg);
+
+	resp_type = sd_response_type(cmd);
+	if (resp_type < 0)
+		return resp_type;
+
+	if (!uhs)
+		cfg2 |= SD_NO_CHECK_WAIT_CRC_TO;
+
+	rtsx_pci_init_cmd(pcr);
+	sd_cmd_set_sd_cmd(pcr, cmd);
+	sd_cmd_set_data_len(pcr, data->blocks, data->blksz);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, IRQSTAT0,
+			DMA_DONE_INT, DMA_DONE_INT);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, DMATC3,
+		0xFF, (u8)(data_len >> 24));
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, DMATC2,
+		0xFF, (u8)(data_len >> 16));
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, DMATC1,
+		0xFF, (u8)(data_len >> 8));
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, DMATC0, 0xFF, (u8)data_len);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, DMACTL,
+		0x03 | DMA_PACK_SIZE_MASK,
+		DMA_DIR_FROM_CARD | DMA_EN | DMA_512);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CARD_DATA_SOURCE,
+			0x01, RING_BUFFER);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_CFG2, 0xFF, cfg2 | resp_type);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_TRANSFER, 0xFF,
+			SD_TRANSFER_START | SD_TM_AUTO_READ_2);
+	rtsx_pci_add_cmd(pcr, CHECK_REG_CMD, SD_TRANSFER,
+			SD_TRANSFER_END, SD_TRANSFER_END);
+	rtsx_pci_send_cmd_no_wait(pcr);
+
+	err = rtsx_pci_dma_transfer(pcr, data->sg, host->sg_count, 1, 10000);
+	if (err < 0) {
+		sd_print_debug_regs(host);
+		sd_clear_error(host);
+		return err;
+	}
+
+	return 0;
+}
+
+static int sd_write_long_data(struct realtek_pci_sdmmc *host,
+	struct mmc_request *mrq)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	struct mmc_host *mmc = host->mmc;
+	struct mmc_card *card = mmc->card;
+	struct mmc_command *cmd = mrq->cmd;
+	struct mmc_data *data = mrq->data;
+	int uhs = mmc_card_uhs(card);
+	u8 cfg2;
+	int err;
+	size_t data_len = data->blksz * data->blocks;
+
+	sd_send_cmd_get_rsp(host, cmd);
+	if (cmd->error)
+		return cmd->error;
+
+	dev_dbg(sdmmc_dev(host), "%s: SD/MMC CMD %d, arg = 0x%08x\n",
+		__func__, cmd->opcode, cmd->arg);
+
+	cfg2 = SD_NO_CALCULATE_CRC7 | SD_CHECK_CRC16 |
+		SD_NO_WAIT_BUSY_END | SD_NO_CHECK_CRC7 | SD_RSP_LEN_0;
+
+	if (!uhs)
+		cfg2 |= SD_NO_CHECK_WAIT_CRC_TO;
+
+	rtsx_pci_init_cmd(pcr);
+	sd_cmd_set_data_len(pcr, data->blocks, data->blksz);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, IRQSTAT0,
+			DMA_DONE_INT, DMA_DONE_INT);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, DMATC3,
+		0xFF, (u8)(data_len >> 24));
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, DMATC2,
+		0xFF, (u8)(data_len >> 16));
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, DMATC1,
+		0xFF, (u8)(data_len >> 8));
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, DMATC0, 0xFF, (u8)data_len);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, DMACTL,
+		0x03 | DMA_PACK_SIZE_MASK,
+		DMA_DIR_TO_CARD | DMA_EN | DMA_512);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CARD_DATA_SOURCE,
+			0x01, RING_BUFFER);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_CFG2, 0xFF, cfg2);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_TRANSFER, 0xFF,
+			SD_TRANSFER_START | SD_TM_AUTO_WRITE_3);
+	rtsx_pci_add_cmd(pcr, CHECK_REG_CMD, SD_TRANSFER,
+			SD_TRANSFER_END, SD_TRANSFER_END);
+	rtsx_pci_send_cmd_no_wait(pcr);
+	err = rtsx_pci_dma_transfer(pcr, data->sg, host->sg_count, 0, 10000);
+	if (err < 0) {
+		sd_clear_error(host);
+		return err;
+	}
+
+	return 0;
+}
+
+static inline void sd_enable_initial_mode(struct realtek_pci_sdmmc *host)
+{
+	rtsx_pci_write_register(host->pcr, SD_CFG1,
+			SD_CLK_DIVIDE_MASK, SD_CLK_DIVIDE_128);
+}
+
+static inline void sd_disable_initial_mode(struct realtek_pci_sdmmc *host)
+{
+	rtsx_pci_write_register(host->pcr, SD_CFG1,
+			SD_CLK_DIVIDE_MASK, SD_CLK_DIVIDE_0);
+}
+
+static int sd_rw_multi(struct realtek_pci_sdmmc *host, struct mmc_request *mrq)
+{
+	struct mmc_data *data = mrq->data;
+	int err;
+
+	if (host->sg_count < 0) {
+		data->error = host->sg_count;
+		dev_dbg(sdmmc_dev(host), "%s: sg_count = %d is invalid\n",
+			__func__, host->sg_count);
+		return data->error;
+	}
+
+	if (data->flags & MMC_DATA_READ) {
+		if (host->initial_mode)
+			sd_disable_initial_mode(host);
+
+		err = sd_read_long_data(host, mrq);
+
+		if (host->initial_mode)
+			sd_enable_initial_mode(host);
+
+		return err;
+	}
+
+	return sd_write_long_data(host, mrq);
+}
+
+static void sd_normal_rw(struct realtek_pci_sdmmc *host,
+		struct mmc_request *mrq)
+{
+	struct mmc_command *cmd = mrq->cmd;
+	struct mmc_data *data = mrq->data;
+	u8 *buf;
+
+	buf = kzalloc(data->blksz, GFP_NOIO);
+	if (!buf) {
+		cmd->error = -ENOMEM;
+		return;
+	}
+
+	if (data->flags & MMC_DATA_READ) {
+		if (host->initial_mode)
+			sd_disable_initial_mode(host);
+
+		cmd->error = sd_read_data(host, cmd, (u16)data->blksz, buf,
+				data->blksz, 200);
+
+		if (host->initial_mode)
+			sd_enable_initial_mode(host);
+
+		sg_copy_from_buffer(data->sg, data->sg_len, buf, data->blksz);
+	} else {
+		sg_copy_to_buffer(data->sg, data->sg_len, buf, data->blksz);
+
+		cmd->error = sd_write_data(host, cmd, (u16)data->blksz, buf,
+				data->blksz, 200);
+	}
+
+	kfree(buf);
+}
+
+static int sd_change_phase(struct realtek_pci_sdmmc *host,
+		u8 sample_point, bool rx)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	u16 SD_VP_CTL = 0;
+	dev_dbg(sdmmc_dev(host), "%s(%s): sample_point = %d\n",
+			__func__, rx ? "RX" : "TX", sample_point);
+
+	rtsx_pci_write_register(pcr, CLK_CTL, CHANGE_CLK, CHANGE_CLK);
+	if (rx) {
+		SD_VP_CTL = SD_VPRX_CTL;
+		rtsx_pci_write_register(pcr, SD_VPRX_CTL,
+			PHASE_SELECT_MASK, sample_point);
+	} else {
+		SD_VP_CTL = SD_VPTX_CTL;
+		rtsx_pci_write_register(pcr, SD_VPTX_CTL,
+			PHASE_SELECT_MASK, sample_point);
+	}
+	rtsx_pci_write_register(pcr, SD_VP_CTL, PHASE_NOT_RESET, 0);
+	rtsx_pci_write_register(pcr, SD_VP_CTL, PHASE_NOT_RESET,
+				PHASE_NOT_RESET);
+	rtsx_pci_write_register(pcr, CLK_CTL, CHANGE_CLK, 0);
+	rtsx_pci_write_register(pcr, SD_CFG1, SD_ASYNC_FIFO_NOT_RST, 0);
+
+	return 0;
+}
+
+static inline u32 test_phase_bit(u32 phase_map, unsigned int bit)
+{
+	bit %= RTSX_PHASE_MAX;
+	return phase_map & (1 << bit);
+}
+
+static int sd_get_phase_len(u32 phase_map, unsigned int start_bit)
+{
+	int i;
+
+	for (i = 0; i < RTSX_PHASE_MAX; i++) {
+		if (test_phase_bit(phase_map, start_bit + i) == 0)
+			return i;
+	}
+	return RTSX_PHASE_MAX;
+}
+
+static u8 sd_search_final_phase(struct realtek_pci_sdmmc *host, u32 phase_map)
+{
+	int start = 0, len = 0;
+	int start_final = 0, len_final = 0;
+	u8 final_phase = 0xFF;
+
+	if (phase_map == 0) {
+		dev_err(sdmmc_dev(host), "phase error: [map:%x]\n", phase_map);
+		return final_phase;
+	}
+
+	while (start < RTSX_PHASE_MAX) {
+		len = sd_get_phase_len(phase_map, start);
+		if (len_final < len) {
+			start_final = start;
+			len_final = len;
+		}
+		start += len ? len : 1;
+	}
+
+	final_phase = (start_final + len_final / 2) % RTSX_PHASE_MAX;
+	dev_dbg(sdmmc_dev(host), "phase: [map:%x] [maxlen:%d] [final:%d]\n",
+		phase_map, len_final, final_phase);
+
+	return final_phase;
+}
+
+static void sd_wait_data_idle(struct realtek_pci_sdmmc *host)
+{
+	int i;
+	u8 val = 0;
+
+	for (i = 0; i < 100; i++) {
+		rtsx_pci_read_register(host->pcr, SD_DATA_STATE, &val);
+		if (val & SD_DATA_IDLE)
+			return;
+
+		udelay(100);
+	}
+}
+
+static int sd_tuning_rx_cmd(struct realtek_pci_sdmmc *host,
+		u8 opcode, u8 sample_point)
+{
+	int err;
+	struct mmc_command cmd = {};
+	struct rtsx_pcr *pcr = host->pcr;
+
+	sd_change_phase(host, sample_point, true);
+
+	rtsx_pci_write_register(pcr, SD_CFG3, SD_RSP_80CLK_TIMEOUT_EN,
+		SD_RSP_80CLK_TIMEOUT_EN);
+
+	cmd.opcode = opcode;
+	err = sd_read_data(host, &cmd, 0x40, NULL, 0, 100);
+	if (err < 0) {
+		/* Wait till SD DATA IDLE */
+		sd_wait_data_idle(host);
+		sd_clear_error(host);
+		rtsx_pci_write_register(pcr, SD_CFG3,
+			SD_RSP_80CLK_TIMEOUT_EN, 0);
+		return err;
+	}
+
+	rtsx_pci_write_register(pcr, SD_CFG3, SD_RSP_80CLK_TIMEOUT_EN, 0);
+	return 0;
+}
+
+static int sd_tuning_phase(struct realtek_pci_sdmmc *host,
+		u8 opcode, u32 *phase_map)
+{
+	int err, i;
+	u32 raw_phase_map = 0;
+
+	for (i = 0; i < RTSX_PHASE_MAX; i++) {
+		err = sd_tuning_rx_cmd(host, opcode, (u8)i);
+		if (err == 0)
+			raw_phase_map |= 1 << i;
+	}
+
+	if (phase_map)
+		*phase_map = raw_phase_map;
+
+	return 0;
+}
+
+static int sd_tuning_rx(struct realtek_pci_sdmmc *host, u8 opcode)
+{
+	int err, i;
+	u32 raw_phase_map[RX_TUNING_CNT] = {0}, phase_map;
+	u8 final_phase;
+
+	for (i = 0; i < RX_TUNING_CNT; i++) {
+		err = sd_tuning_phase(host, opcode, &(raw_phase_map[i]));
+		if (err < 0)
+			return err;
+
+		if (raw_phase_map[i] == 0)
+			break;
+	}
+
+	phase_map = 0xFFFFFFFF;
+	for (i = 0; i < RX_TUNING_CNT; i++) {
+		dev_dbg(sdmmc_dev(host), "RX raw_phase_map[%d] = 0x%08x\n",
+				i, raw_phase_map[i]);
+		phase_map &= raw_phase_map[i];
+	}
+	dev_dbg(sdmmc_dev(host), "RX phase_map = 0x%08x\n", phase_map);
+
+	if (phase_map) {
+		final_phase = sd_search_final_phase(host, phase_map);
+		if (final_phase == 0xFF)
+			return -EINVAL;
+
+		err = sd_change_phase(host, final_phase, true);
+		if (err < 0)
+			return err;
+	} else {
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static inline int sdio_extblock_cmd(struct mmc_command *cmd,
+	struct mmc_data *data)
+{
+	return (cmd->opcode == SD_IO_RW_EXTENDED) && (data->blksz == 512);
+}
+
+static inline int sd_rw_cmd(struct mmc_command *cmd)
+{
+	return mmc_op_multi(cmd->opcode) ||
+		(cmd->opcode == MMC_READ_SINGLE_BLOCK) ||
+		(cmd->opcode == MMC_WRITE_BLOCK);
+}
+
+static void sd_request(struct work_struct *work)
+{
+	struct realtek_pci_sdmmc *host = container_of(work,
+			struct realtek_pci_sdmmc, work);
+	struct rtsx_pcr *pcr = host->pcr;
+
+	struct mmc_host *mmc = host->mmc;
+	struct mmc_request *mrq = host->mrq;
+	struct mmc_command *cmd = mrq->cmd;
+	struct mmc_data *data = mrq->data;
+
+	unsigned int data_size = 0;
+	int err;
+
+	if (host->eject || !sd_get_cd_int(host)) {
+		cmd->error = -ENOMEDIUM;
+		goto finish;
+	}
+
+	err = rtsx_pci_card_exclusive_check(host->pcr, RTSX_SD_CARD);
+	if (err) {
+		cmd->error = err;
+		goto finish;
+	}
+
+	mutex_lock(&pcr->pcr_mutex);
+
+	rtsx_pci_start_run(pcr);
+
+	rtsx_pci_switch_clock(pcr, host->clock, host->ssc_depth,
+			host->initial_mode, host->double_clk, host->vpclk);
+	rtsx_pci_write_register(pcr, CARD_SELECT, 0x07, SD_MOD_SEL);
+	rtsx_pci_write_register(pcr, CARD_SHARE_MODE,
+			CARD_SHARE_MASK, CARD_SHARE_48_SD);
+
+	mutex_lock(&host->host_mutex);
+	host->mrq = mrq;
+	mutex_unlock(&host->host_mutex);
+
+	if (mrq->data)
+		data_size = data->blocks * data->blksz;
+
+	if (!data_size) {
+		sd_send_cmd_get_rsp(host, cmd);
+	} else if (sd_rw_cmd(cmd) || sdio_extblock_cmd(cmd, data)) {
+		cmd->error = sd_rw_multi(host, mrq);
+		if (!host->using_cookie)
+			sdmmc_post_req(host->mmc, host->mrq, 0);
+
+		if (mmc_op_multi(cmd->opcode) && mrq->stop)
+			sd_send_cmd_get_rsp(host, mrq->stop);
+	} else {
+		sd_normal_rw(host, mrq);
+	}
+
+	if (mrq->data) {
+		if (cmd->error || data->error)
+			data->bytes_xfered = 0;
+		else
+			data->bytes_xfered = data->blocks * data->blksz;
+	}
+
+	mutex_unlock(&pcr->pcr_mutex);
+
+finish:
+	if (cmd->error) {
+		dev_dbg(sdmmc_dev(host), "CMD %d 0x%08x error(%d)\n",
+			cmd->opcode, cmd->arg, cmd->error);
+	}
+
+	mutex_lock(&host->host_mutex);
+	host->mrq = NULL;
+	mutex_unlock(&host->host_mutex);
+
+	mmc_request_done(mmc, mrq);
+}
+
+static void sdmmc_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct realtek_pci_sdmmc *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+
+	mutex_lock(&host->host_mutex);
+	host->mrq = mrq;
+	mutex_unlock(&host->host_mutex);
+
+	if (sd_rw_cmd(mrq->cmd) || sdio_extblock_cmd(mrq->cmd, data))
+		host->using_cookie = sd_pre_dma_transfer(host, data, false);
+
+	schedule_work(&host->work);
+}
+
+static int sd_set_bus_width(struct realtek_pci_sdmmc *host,
+		unsigned char bus_width)
+{
+	int err = 0;
+	u8 width[] = {
+		[MMC_BUS_WIDTH_1] = SD_BUS_WIDTH_1BIT,
+		[MMC_BUS_WIDTH_4] = SD_BUS_WIDTH_4BIT,
+		[MMC_BUS_WIDTH_8] = SD_BUS_WIDTH_8BIT,
+	};
+
+	if (bus_width <= MMC_BUS_WIDTH_8)
+		err = rtsx_pci_write_register(host->pcr, SD_CFG1,
+				0x03, width[bus_width]);
+
+	return err;
+}
+
+static int sd_power_on(struct realtek_pci_sdmmc *host, unsigned char power_mode)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	int err;
+
+	if (host->prev_power_state == MMC_POWER_ON)
+		return 0;
+
+	if (host->prev_power_state == MMC_POWER_UP) {
+		rtsx_pci_write_register(pcr, SD_BUS_STAT, SD_CLK_TOGGLE_EN, 0);
+		goto finish;
+	}
+
+	msleep(100);
+
+	rtsx_pci_init_cmd(pcr);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CARD_SELECT, 0x07, SD_MOD_SEL);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CARD_SHARE_MODE,
+			CARD_SHARE_MASK, CARD_SHARE_48_SD);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CARD_CLK_EN,
+			SD_CLK_EN, SD_CLK_EN);
+	err = rtsx_pci_send_cmd(pcr, 100);
+	if (err < 0)
+		return err;
+
+	err = rtsx_pci_card_pull_ctl_enable(pcr, RTSX_SD_CARD);
+	if (err < 0)
+		return err;
+
+	err = rtsx_pci_card_power_on(pcr, RTSX_SD_CARD);
+	if (err < 0)
+		return err;
+
+	mdelay(1);
+
+	err = rtsx_pci_write_register(pcr, CARD_OE, SD_OUTPUT_EN, SD_OUTPUT_EN);
+	if (err < 0)
+		return err;
+
+	/* send at least 74 clocks */
+	rtsx_pci_write_register(pcr, SD_BUS_STAT, SD_CLK_TOGGLE_EN, SD_CLK_TOGGLE_EN);
+
+finish:
+	host->prev_power_state = power_mode;
+	return 0;
+}
+
+static int sd_power_off(struct realtek_pci_sdmmc *host)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	int err;
+
+	host->prev_power_state = MMC_POWER_OFF;
+
+	rtsx_pci_init_cmd(pcr);
+
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CARD_CLK_EN, SD_CLK_EN, 0);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CARD_OE, SD_OUTPUT_EN, 0);
+
+	err = rtsx_pci_send_cmd(pcr, 100);
+	if (err < 0)
+		return err;
+
+	err = rtsx_pci_card_power_off(pcr, RTSX_SD_CARD);
+	if (err < 0)
+		return err;
+
+	return rtsx_pci_card_pull_ctl_disable(pcr, RTSX_SD_CARD);
+}
+
+static int sd_set_power_mode(struct realtek_pci_sdmmc *host,
+		unsigned char power_mode)
+{
+	int err;
+
+	if (power_mode == MMC_POWER_OFF)
+		err = sd_power_off(host);
+	else
+		err = sd_power_on(host, power_mode);
+
+	return err;
+}
+
+static int sd_set_timing(struct realtek_pci_sdmmc *host, unsigned char timing)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	int err = 0;
+
+	rtsx_pci_init_cmd(pcr);
+
+	switch (timing) {
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_UHS_SDR50:
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_CFG1,
+				0x0C | SD_ASYNC_FIFO_NOT_RST,
+				SD_30_MODE | SD_ASYNC_FIFO_NOT_RST);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CLK_CTL,
+				CLK_LOW_FREQ, CLK_LOW_FREQ);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CARD_CLK_SOURCE, 0xFF,
+				CRC_VAR_CLK0 | SD30_FIX_CLK | SAMPLE_VAR_CLK1);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CLK_CTL, CLK_LOW_FREQ, 0);
+		break;
+
+	case MMC_TIMING_MMC_DDR52:
+	case MMC_TIMING_UHS_DDR50:
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_CFG1,
+				0x0C | SD_ASYNC_FIFO_NOT_RST,
+				SD_DDR_MODE | SD_ASYNC_FIFO_NOT_RST);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CLK_CTL,
+				CLK_LOW_FREQ, CLK_LOW_FREQ);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CARD_CLK_SOURCE, 0xFF,
+				CRC_VAR_CLK0 | SD30_FIX_CLK | SAMPLE_VAR_CLK1);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CLK_CTL, CLK_LOW_FREQ, 0);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_PUSH_POINT_CTL,
+				DDR_VAR_TX_CMD_DAT, DDR_VAR_TX_CMD_DAT);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_SAMPLE_POINT_CTL,
+				DDR_VAR_RX_DAT | DDR_VAR_RX_CMD,
+				DDR_VAR_RX_DAT | DDR_VAR_RX_CMD);
+		break;
+
+	case MMC_TIMING_MMC_HS:
+	case MMC_TIMING_SD_HS:
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_CFG1,
+				0x0C, SD_20_MODE);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CLK_CTL,
+				CLK_LOW_FREQ, CLK_LOW_FREQ);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CARD_CLK_SOURCE, 0xFF,
+				CRC_FIX_CLK | SD30_VAR_CLK0 | SAMPLE_VAR_CLK1);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CLK_CTL, CLK_LOW_FREQ, 0);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_PUSH_POINT_CTL,
+				SD20_TX_SEL_MASK, SD20_TX_14_AHEAD);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_SAMPLE_POINT_CTL,
+				SD20_RX_SEL_MASK, SD20_RX_14_DELAY);
+		break;
+
+	default:
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD,
+				SD_CFG1, 0x0C, SD_20_MODE);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CLK_CTL,
+				CLK_LOW_FREQ, CLK_LOW_FREQ);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CARD_CLK_SOURCE, 0xFF,
+				CRC_FIX_CLK | SD30_VAR_CLK0 | SAMPLE_VAR_CLK1);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CLK_CTL, CLK_LOW_FREQ, 0);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD,
+				SD_PUSH_POINT_CTL, 0xFF, 0);
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_SAMPLE_POINT_CTL,
+				SD20_RX_SEL_MASK, SD20_RX_POS_EDGE);
+		break;
+	}
+
+	err = rtsx_pci_send_cmd(pcr, 100);
+
+	return err;
+}
+
+static void sdmmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct realtek_pci_sdmmc *host = mmc_priv(mmc);
+	struct rtsx_pcr *pcr = host->pcr;
+
+	if (host->eject)
+		return;
+
+	if (rtsx_pci_card_exclusive_check(host->pcr, RTSX_SD_CARD))
+		return;
+
+	mutex_lock(&pcr->pcr_mutex);
+
+	rtsx_pci_start_run(pcr);
+
+	sd_set_bus_width(host, ios->bus_width);
+	sd_set_power_mode(host, ios->power_mode);
+	sd_set_timing(host, ios->timing);
+
+	host->vpclk = false;
+	host->double_clk = true;
+
+	switch (ios->timing) {
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_UHS_SDR50:
+		host->ssc_depth = RTSX_SSC_DEPTH_2M;
+		host->vpclk = true;
+		host->double_clk = false;
+		break;
+	case MMC_TIMING_MMC_DDR52:
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_UHS_SDR25:
+		host->ssc_depth = RTSX_SSC_DEPTH_1M;
+		break;
+	default:
+		host->ssc_depth = RTSX_SSC_DEPTH_500K;
+		break;
+	}
+
+	host->initial_mode = (ios->clock <= 1000000) ? true : false;
+
+	host->clock = ios->clock;
+	rtsx_pci_switch_clock(pcr, ios->clock, host->ssc_depth,
+			host->initial_mode, host->double_clk, host->vpclk);
+
+	mutex_unlock(&pcr->pcr_mutex);
+}
+
+static int sdmmc_get_ro(struct mmc_host *mmc)
+{
+	struct realtek_pci_sdmmc *host = mmc_priv(mmc);
+	struct rtsx_pcr *pcr = host->pcr;
+	int ro = 0;
+	u32 val;
+
+	if (host->eject)
+		return -ENOMEDIUM;
+
+	mutex_lock(&pcr->pcr_mutex);
+
+	rtsx_pci_start_run(pcr);
+
+	/* Check SD mechanical write-protect switch */
+	val = rtsx_pci_readl(pcr, RTSX_BIPR);
+	dev_dbg(sdmmc_dev(host), "%s: RTSX_BIPR = 0x%08x\n", __func__, val);
+	if (val & SD_WRITE_PROTECT)
+		ro = 1;
+
+	mutex_unlock(&pcr->pcr_mutex);
+
+	return ro;
+}
+
+static int sdmmc_get_cd(struct mmc_host *mmc)
+{
+	struct realtek_pci_sdmmc *host = mmc_priv(mmc);
+	struct rtsx_pcr *pcr = host->pcr;
+	int cd = 0;
+	u32 val;
+
+	if (host->eject)
+		return cd;
+
+	mutex_lock(&pcr->pcr_mutex);
+
+	rtsx_pci_start_run(pcr);
+
+	/* Check SD card detect */
+	val = rtsx_pci_card_exist(pcr);
+	dev_dbg(sdmmc_dev(host), "%s: RTSX_BIPR = 0x%08x\n", __func__, val);
+	if (val & SD_EXIST)
+		cd = 1;
+
+	mutex_unlock(&pcr->pcr_mutex);
+
+	return cd;
+}
+
+static int sd_wait_voltage_stable_1(struct realtek_pci_sdmmc *host)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	int err;
+	u8 stat;
+
+	/* Reference to Signal Voltage Switch Sequence in SD spec.
+	 * Wait for a period of time so that the card can drive SD_CMD and
+	 * SD_DAT[3:0] to low after sending back CMD11 response.
+	 */
+	mdelay(1);
+
+	/* SD_CMD, SD_DAT[3:0] should be driven to low by card;
+	 * If either one of SD_CMD,SD_DAT[3:0] is not low,
+	 * abort the voltage switch sequence;
+	 */
+	err = rtsx_pci_read_register(pcr, SD_BUS_STAT, &stat);
+	if (err < 0)
+		return err;
+
+	if (stat & (SD_CMD_STATUS | SD_DAT3_STATUS | SD_DAT2_STATUS |
+				SD_DAT1_STATUS | SD_DAT0_STATUS))
+		return -EINVAL;
+
+	/* Stop toggle SD clock */
+	err = rtsx_pci_write_register(pcr, SD_BUS_STAT,
+			0xFF, SD_CLK_FORCE_STOP);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
+static int sd_wait_voltage_stable_2(struct realtek_pci_sdmmc *host)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	int err;
+	u8 stat, mask, val;
+
+	/* Wait 1.8V output of voltage regulator in card stable */
+	msleep(50);
+
+	/* Toggle SD clock again */
+	err = rtsx_pci_write_register(pcr, SD_BUS_STAT, 0xFF, SD_CLK_TOGGLE_EN);
+	if (err < 0)
+		return err;
+
+	/* Wait for a period of time so that the card can drive
+	 * SD_DAT[3:0] to high at 1.8V
+	 */
+	msleep(20);
+
+	/* SD_CMD, SD_DAT[3:0] should be pulled high by host */
+	err = rtsx_pci_read_register(pcr, SD_BUS_STAT, &stat);
+	if (err < 0)
+		return err;
+
+	mask = SD_CMD_STATUS | SD_DAT3_STATUS | SD_DAT2_STATUS |
+		SD_DAT1_STATUS | SD_DAT0_STATUS;
+	val = SD_CMD_STATUS | SD_DAT3_STATUS | SD_DAT2_STATUS |
+		SD_DAT1_STATUS | SD_DAT0_STATUS;
+	if ((stat & mask) != val) {
+		dev_dbg(sdmmc_dev(host),
+			"%s: SD_BUS_STAT = 0x%x\n", __func__, stat);
+		rtsx_pci_write_register(pcr, SD_BUS_STAT,
+				SD_CLK_TOGGLE_EN | SD_CLK_FORCE_STOP, 0);
+		rtsx_pci_write_register(pcr, CARD_CLK_EN, 0xFF, 0);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int sdmmc_switch_voltage(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct realtek_pci_sdmmc *host = mmc_priv(mmc);
+	struct rtsx_pcr *pcr = host->pcr;
+	int err = 0;
+	u8 voltage;
+
+	dev_dbg(sdmmc_dev(host), "%s: signal_voltage = %d\n",
+			__func__, ios->signal_voltage);
+
+	if (host->eject)
+		return -ENOMEDIUM;
+
+	err = rtsx_pci_card_exclusive_check(host->pcr, RTSX_SD_CARD);
+	if (err)
+		return err;
+
+	mutex_lock(&pcr->pcr_mutex);
+
+	rtsx_pci_start_run(pcr);
+
+	if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_330)
+		voltage = OUTPUT_3V3;
+	else
+		voltage = OUTPUT_1V8;
+
+	if (voltage == OUTPUT_1V8) {
+		err = sd_wait_voltage_stable_1(host);
+		if (err < 0)
+			goto out;
+	}
+
+	err = rtsx_pci_switch_output_voltage(pcr, voltage);
+	if (err < 0)
+		goto out;
+
+	if (voltage == OUTPUT_1V8) {
+		err = sd_wait_voltage_stable_2(host);
+		if (err < 0)
+			goto out;
+	}
+
+out:
+	/* Stop toggle SD clock in idle */
+	err = rtsx_pci_write_register(pcr, SD_BUS_STAT,
+			SD_CLK_TOGGLE_EN | SD_CLK_FORCE_STOP, 0);
+
+	mutex_unlock(&pcr->pcr_mutex);
+
+	return err;
+}
+
+static int sdmmc_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct realtek_pci_sdmmc *host = mmc_priv(mmc);
+	struct rtsx_pcr *pcr = host->pcr;
+	int err = 0;
+
+	if (host->eject)
+		return -ENOMEDIUM;
+
+	err = rtsx_pci_card_exclusive_check(host->pcr, RTSX_SD_CARD);
+	if (err)
+		return err;
+
+	mutex_lock(&pcr->pcr_mutex);
+
+	rtsx_pci_start_run(pcr);
+
+	/* Set initial TX phase */
+	switch (mmc->ios.timing) {
+	case MMC_TIMING_UHS_SDR104:
+		err = sd_change_phase(host, SDR104_TX_PHASE(pcr), false);
+		break;
+
+	case MMC_TIMING_UHS_SDR50:
+		err = sd_change_phase(host, SDR50_TX_PHASE(pcr), false);
+		break;
+
+	case MMC_TIMING_UHS_DDR50:
+		err = sd_change_phase(host, DDR50_TX_PHASE(pcr), false);
+		break;
+
+	default:
+		err = 0;
+	}
+
+	if (err)
+		goto out;
+
+	/* Tuning RX phase */
+	if ((mmc->ios.timing == MMC_TIMING_UHS_SDR104) ||
+			(mmc->ios.timing == MMC_TIMING_UHS_SDR50))
+		err = sd_tuning_rx(host, opcode);
+	else if (mmc->ios.timing == MMC_TIMING_UHS_DDR50)
+		err = sd_change_phase(host, DDR50_RX_PHASE(pcr), true);
+
+out:
+	mutex_unlock(&pcr->pcr_mutex);
+
+	return err;
+}
+
+static const struct mmc_host_ops realtek_pci_sdmmc_ops = {
+	.pre_req = sdmmc_pre_req,
+	.post_req = sdmmc_post_req,
+	.request = sdmmc_request,
+	.set_ios = sdmmc_set_ios,
+	.get_ro = sdmmc_get_ro,
+	.get_cd = sdmmc_get_cd,
+	.start_signal_voltage_switch = sdmmc_switch_voltage,
+	.execute_tuning = sdmmc_execute_tuning,
+};
+
+static void init_extra_caps(struct realtek_pci_sdmmc *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	struct rtsx_pcr *pcr = host->pcr;
+
+	dev_dbg(sdmmc_dev(host), "pcr->extra_caps = 0x%x\n", pcr->extra_caps);
+
+	if (pcr->extra_caps & EXTRA_CAPS_SD_SDR50)
+		mmc->caps |= MMC_CAP_UHS_SDR50;
+	if (pcr->extra_caps & EXTRA_CAPS_SD_SDR104)
+		mmc->caps |= MMC_CAP_UHS_SDR104;
+	if (pcr->extra_caps & EXTRA_CAPS_SD_DDR50)
+		mmc->caps |= MMC_CAP_UHS_DDR50;
+	if (pcr->extra_caps & EXTRA_CAPS_MMC_HSDDR)
+		mmc->caps |= MMC_CAP_1_8V_DDR;
+	if (pcr->extra_caps & EXTRA_CAPS_MMC_8BIT)
+		mmc->caps |= MMC_CAP_8_BIT_DATA;
+	if (pcr->extra_caps & EXTRA_CAPS_NO_MMC)
+		mmc->caps2 |= MMC_CAP2_NO_MMC;
+}
+
+static void realtek_init_host(struct realtek_pci_sdmmc *host)
+{
+	struct mmc_host *mmc = host->mmc;
+
+	mmc->f_min = 250000;
+	mmc->f_max = 208000000;
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34 | MMC_VDD_165_195;
+	mmc->caps = MMC_CAP_4_BIT_DATA | MMC_CAP_SD_HIGHSPEED |
+		MMC_CAP_MMC_HIGHSPEED | MMC_CAP_BUS_WIDTH_TEST |
+		MMC_CAP_UHS_SDR12 | MMC_CAP_UHS_SDR25;
+	mmc->caps2 = MMC_CAP2_NO_PRESCAN_POWERUP | MMC_CAP2_FULL_PWR_CYCLE;
+	mmc->max_current_330 = 400;
+	mmc->max_current_180 = 800;
+	mmc->ops = &realtek_pci_sdmmc_ops;
+
+	init_extra_caps(host);
+
+	mmc->max_segs = 256;
+	mmc->max_seg_size = 65536;
+	mmc->max_blk_size = 512;
+	mmc->max_blk_count = 65535;
+	mmc->max_req_size = 524288;
+}
+
+static void rtsx_pci_sdmmc_card_event(struct platform_device *pdev)
+{
+	struct realtek_pci_sdmmc *host = platform_get_drvdata(pdev);
+
+	host->cookie = -1;
+	mmc_detect_change(host->mmc, 0);
+}
+
+static int rtsx_pci_sdmmc_drv_probe(struct platform_device *pdev)
+{
+	struct mmc_host *mmc;
+	struct realtek_pci_sdmmc *host;
+	struct rtsx_pcr *pcr;
+	struct pcr_handle *handle = pdev->dev.platform_data;
+
+	if (!handle)
+		return -ENXIO;
+
+	pcr = handle->pcr;
+	if (!pcr)
+		return -ENXIO;
+
+	dev_dbg(&(pdev->dev), ": Realtek PCI-E SDMMC controller found\n");
+
+	mmc = mmc_alloc_host(sizeof(*host), &pdev->dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	host = mmc_priv(mmc);
+	host->pcr = pcr;
+	mmc->ios.power_delay_ms = 5;
+	host->mmc = mmc;
+	host->pdev = pdev;
+	host->cookie = -1;
+	host->prev_power_state = MMC_POWER_OFF;
+	INIT_WORK(&host->work, sd_request);
+	platform_set_drvdata(pdev, host);
+	pcr->slots[RTSX_SD_CARD].p_dev = pdev;
+	pcr->slots[RTSX_SD_CARD].card_event = rtsx_pci_sdmmc_card_event;
+
+	mutex_init(&host->host_mutex);
+
+	realtek_init_host(host);
+
+	mmc_add_host(mmc);
+
+	return 0;
+}
+
+static int rtsx_pci_sdmmc_drv_remove(struct platform_device *pdev)
+{
+	struct realtek_pci_sdmmc *host = platform_get_drvdata(pdev);
+	struct rtsx_pcr *pcr;
+	struct mmc_host *mmc;
+
+	if (!host)
+		return 0;
+
+	pcr = host->pcr;
+	pcr->slots[RTSX_SD_CARD].p_dev = NULL;
+	pcr->slots[RTSX_SD_CARD].card_event = NULL;
+	mmc = host->mmc;
+
+	cancel_work_sync(&host->work);
+
+	mutex_lock(&host->host_mutex);
+	if (host->mrq) {
+		dev_dbg(&(pdev->dev),
+			"%s: Controller removed during transfer\n",
+			mmc_hostname(mmc));
+
+		rtsx_pci_complete_unfinished_transfer(pcr);
+
+		host->mrq->cmd->error = -ENOMEDIUM;
+		if (host->mrq->stop)
+			host->mrq->stop->error = -ENOMEDIUM;
+		mmc_request_done(mmc, host->mrq);
+	}
+	mutex_unlock(&host->host_mutex);
+
+	mmc_remove_host(mmc);
+	host->eject = true;
+
+	flush_work(&host->work);
+
+	mmc_free_host(mmc);
+
+	dev_dbg(&(pdev->dev),
+		": Realtek PCI-E SDMMC controller has been removed\n");
+
+	return 0;
+}
+
+static const struct platform_device_id rtsx_pci_sdmmc_ids[] = {
+	{
+		.name = DRV_NAME_RTSX_PCI_SDMMC,
+	}, {
+		/* sentinel */
+	}
+};
+MODULE_DEVICE_TABLE(platform, rtsx_pci_sdmmc_ids);
+
+static struct platform_driver rtsx_pci_sdmmc_driver = {
+	.probe		= rtsx_pci_sdmmc_drv_probe,
+	.remove		= rtsx_pci_sdmmc_drv_remove,
+	.id_table       = rtsx_pci_sdmmc_ids,
+	.driver		= {
+		.name	= DRV_NAME_RTSX_PCI_SDMMC,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+	},
+};
+module_platform_driver(rtsx_pci_sdmmc_driver);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Wei WANG <wei_wang@realsil.com.cn>");
+MODULE_DESCRIPTION("Realtek PCI-E SD/MMC Card Host Driver");
diff --git a/drivers/mmc/host/rtsx_usb_sdmmc.c b/drivers/mmc/host/rtsx_usb_sdmmc.c
new file mode 100644
index 000000000..1be3a355f
--- /dev/null
+++ b/drivers/mmc/host/rtsx_usb_sdmmc.c
@@ -0,0 +1,1473 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Realtek USB SD/MMC Card Interface driver
+ *
+ * Copyright(c) 2009-2013 Realtek Semiconductor Corp. All rights reserved.
+ *
+ * Author:
+ *   Roger Tseng <rogerable@realtek.com>
+ */
+
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/platform_device.h>
+#include <linux/usb.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sd.h>
+#include <linux/mmc/card.h>
+#include <linux/scatterlist.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+
+#include <linux/rtsx_usb.h>
+#include <asm/unaligned.h>
+
+#if defined(CONFIG_LEDS_CLASS) || (defined(CONFIG_LEDS_CLASS_MODULE) && \
+		defined(CONFIG_MMC_REALTEK_USB_MODULE))
+#include <linux/leds.h>
+#include <linux/workqueue.h>
+#define RTSX_USB_USE_LEDS_CLASS
+#endif
+
+struct rtsx_usb_sdmmc {
+	struct platform_device	*pdev;
+	struct rtsx_ucr	*ucr;
+	struct mmc_host		*mmc;
+	struct mmc_request	*mrq;
+
+	struct mutex		host_mutex;
+
+	u8			ssc_depth;
+	unsigned int		clock;
+	bool			vpclk;
+	bool			double_clk;
+	bool			host_removal;
+	bool			card_exist;
+	bool			initial_mode;
+	bool			ddr_mode;
+
+	unsigned char		power_mode;
+
+#ifdef RTSX_USB_USE_LEDS_CLASS
+	struct led_classdev	led;
+	char			led_name[32];
+	struct work_struct	led_work;
+#endif
+};
+
+static inline struct device *sdmmc_dev(struct rtsx_usb_sdmmc *host)
+{
+	return &(host->pdev->dev);
+}
+
+static inline void sd_clear_error(struct rtsx_usb_sdmmc *host)
+{
+	struct rtsx_ucr *ucr = host->ucr;
+	rtsx_usb_ep0_write_register(ucr, CARD_STOP,
+				  SD_STOP | SD_CLR_ERR,
+				  SD_STOP | SD_CLR_ERR);
+
+	rtsx_usb_clear_dma_err(ucr);
+	rtsx_usb_clear_fsm_err(ucr);
+}
+
+#ifdef DEBUG
+static void sd_print_debug_regs(struct rtsx_usb_sdmmc *host)
+{
+	struct rtsx_ucr *ucr = host->ucr;
+	u8 val = 0;
+
+	rtsx_usb_ep0_read_register(ucr, SD_STAT1, &val);
+	dev_dbg(sdmmc_dev(host), "SD_STAT1: 0x%x\n", val);
+	rtsx_usb_ep0_read_register(ucr, SD_STAT2, &val);
+	dev_dbg(sdmmc_dev(host), "SD_STAT2: 0x%x\n", val);
+	rtsx_usb_ep0_read_register(ucr, SD_BUS_STAT, &val);
+	dev_dbg(sdmmc_dev(host), "SD_BUS_STAT: 0x%x\n", val);
+}
+#else
+#define sd_print_debug_regs(host)
+#endif /* DEBUG */
+
+static int sd_read_data(struct rtsx_usb_sdmmc *host, struct mmc_command *cmd,
+	       u16 byte_cnt, u8 *buf, int buf_len, int timeout)
+{
+	struct rtsx_ucr *ucr = host->ucr;
+	int err;
+	u8 trans_mode;
+
+	if (!buf)
+		buf_len = 0;
+
+	rtsx_usb_init_cmd(ucr);
+	if (cmd != NULL) {
+		dev_dbg(sdmmc_dev(host), "%s: SD/MMC CMD%d\n", __func__
+				, cmd->opcode);
+		if (cmd->opcode == MMC_SEND_TUNING_BLOCK)
+			trans_mode = SD_TM_AUTO_TUNING;
+		else
+			trans_mode = SD_TM_NORMAL_READ;
+
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+				SD_CMD0, 0xFF, (u8)(cmd->opcode) | 0x40);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+				SD_CMD1, 0xFF, (u8)(cmd->arg >> 24));
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+				SD_CMD2, 0xFF, (u8)(cmd->arg >> 16));
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+				SD_CMD3, 0xFF, (u8)(cmd->arg >> 8));
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+				SD_CMD4, 0xFF, (u8)cmd->arg);
+	} else {
+		trans_mode = SD_TM_AUTO_READ_3;
+	}
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_BYTE_CNT_L, 0xFF, (u8)byte_cnt);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_BYTE_CNT_H,
+			0xFF, (u8)(byte_cnt >> 8));
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_BLOCK_CNT_L, 0xFF, 1);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_BLOCK_CNT_H, 0xFF, 0);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_CFG2, 0xFF,
+			SD_CALCULATE_CRC7 | SD_CHECK_CRC16 |
+			SD_NO_WAIT_BUSY_END | SD_CHECK_CRC7 | SD_RSP_LEN_6);
+	if (trans_mode != SD_TM_AUTO_TUNING)
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+				CARD_DATA_SOURCE, 0x01, PINGPONG_BUFFER);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_TRANSFER,
+			0xFF, trans_mode | SD_TRANSFER_START);
+	rtsx_usb_add_cmd(ucr, CHECK_REG_CMD, SD_TRANSFER,
+			SD_TRANSFER_END, SD_TRANSFER_END);
+
+	if (cmd != NULL) {
+		rtsx_usb_add_cmd(ucr, READ_REG_CMD, SD_CMD1, 0, 0);
+		rtsx_usb_add_cmd(ucr, READ_REG_CMD, SD_CMD2, 0, 0);
+		rtsx_usb_add_cmd(ucr, READ_REG_CMD, SD_CMD3, 0, 0);
+		rtsx_usb_add_cmd(ucr, READ_REG_CMD, SD_CMD4, 0, 0);
+	}
+
+	err = rtsx_usb_send_cmd(ucr, MODE_CR, timeout);
+	if (err) {
+		dev_dbg(sdmmc_dev(host),
+			"rtsx_usb_send_cmd failed (err = %d)\n", err);
+		return err;
+	}
+
+	err = rtsx_usb_get_rsp(ucr, !cmd ? 1 : 5, timeout);
+	if (err || (ucr->rsp_buf[0] & SD_TRANSFER_ERR)) {
+		sd_print_debug_regs(host);
+
+		if (!err) {
+			dev_dbg(sdmmc_dev(host),
+				"Transfer failed (SD_TRANSFER = %02x)\n",
+				ucr->rsp_buf[0]);
+			err = -EIO;
+		} else {
+			dev_dbg(sdmmc_dev(host),
+				"rtsx_usb_get_rsp failed (err = %d)\n", err);
+		}
+
+		return err;
+	}
+
+	if (cmd != NULL) {
+		cmd->resp[0] = get_unaligned_be32(ucr->rsp_buf + 1);
+		dev_dbg(sdmmc_dev(host), "cmd->resp[0] = 0x%08x\n",
+				cmd->resp[0]);
+	}
+
+	if (buf && buf_len) {
+		/* 2-byte aligned part */
+		err = rtsx_usb_read_ppbuf(ucr, buf, byte_cnt - (byte_cnt % 2));
+		if (err) {
+			dev_dbg(sdmmc_dev(host),
+				"rtsx_usb_read_ppbuf failed (err = %d)\n", err);
+			return err;
+		}
+
+		/* unaligned byte */
+		if (byte_cnt % 2)
+			return rtsx_usb_read_register(ucr,
+					PPBUF_BASE2 + byte_cnt,
+					buf + byte_cnt - 1);
+	}
+
+	return 0;
+}
+
+static int sd_write_data(struct rtsx_usb_sdmmc *host, struct mmc_command *cmd,
+		u16 byte_cnt, u8 *buf, int buf_len, int timeout)
+{
+	struct rtsx_ucr *ucr = host->ucr;
+	int err;
+	u8 trans_mode;
+
+	if (!buf)
+		buf_len = 0;
+
+	if (buf && buf_len) {
+		err = rtsx_usb_write_ppbuf(ucr, buf, buf_len);
+		if (err) {
+			dev_dbg(sdmmc_dev(host),
+				"rtsx_usb_write_ppbuf failed (err = %d)\n",
+				err);
+			return err;
+		}
+	}
+
+	trans_mode = (cmd != NULL) ? SD_TM_AUTO_WRITE_2 : SD_TM_AUTO_WRITE_3;
+	rtsx_usb_init_cmd(ucr);
+
+	if (cmd != NULL) {
+		dev_dbg(sdmmc_dev(host), "%s: SD/MMC CMD%d\n", __func__,
+				cmd->opcode);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+				SD_CMD0, 0xFF, (u8)(cmd->opcode) | 0x40);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+				SD_CMD1, 0xFF, (u8)(cmd->arg >> 24));
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+				SD_CMD2, 0xFF, (u8)(cmd->arg >> 16));
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+				SD_CMD3, 0xFF, (u8)(cmd->arg >> 8));
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+				SD_CMD4, 0xFF, (u8)cmd->arg);
+	}
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_BYTE_CNT_L, 0xFF, (u8)byte_cnt);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_BYTE_CNT_H,
+			0xFF, (u8)(byte_cnt >> 8));
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_BLOCK_CNT_L, 0xFF, 1);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_BLOCK_CNT_H, 0xFF, 0);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_CFG2, 0xFF,
+		SD_CALCULATE_CRC7 | SD_CHECK_CRC16 |
+		SD_NO_WAIT_BUSY_END | SD_CHECK_CRC7 | SD_RSP_LEN_6);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+			CARD_DATA_SOURCE, 0x01, PINGPONG_BUFFER);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_TRANSFER, 0xFF,
+			trans_mode | SD_TRANSFER_START);
+	rtsx_usb_add_cmd(ucr, CHECK_REG_CMD, SD_TRANSFER,
+			SD_TRANSFER_END, SD_TRANSFER_END);
+
+	if (cmd != NULL) {
+		rtsx_usb_add_cmd(ucr, READ_REG_CMD, SD_CMD1, 0, 0);
+		rtsx_usb_add_cmd(ucr, READ_REG_CMD, SD_CMD2, 0, 0);
+		rtsx_usb_add_cmd(ucr, READ_REG_CMD, SD_CMD3, 0, 0);
+		rtsx_usb_add_cmd(ucr, READ_REG_CMD, SD_CMD4, 0, 0);
+	}
+
+	err = rtsx_usb_send_cmd(ucr, MODE_CR, timeout);
+	if (err) {
+		dev_dbg(sdmmc_dev(host),
+			"rtsx_usb_send_cmd failed (err = %d)\n", err);
+		return err;
+	}
+
+	err = rtsx_usb_get_rsp(ucr, !cmd ? 1 : 5, timeout);
+	if (err) {
+		sd_print_debug_regs(host);
+		dev_dbg(sdmmc_dev(host),
+			"rtsx_usb_get_rsp failed (err = %d)\n", err);
+		return err;
+	}
+
+	if (cmd != NULL) {
+		cmd->resp[0] = get_unaligned_be32(ucr->rsp_buf + 1);
+		dev_dbg(sdmmc_dev(host), "cmd->resp[0] = 0x%08x\n",
+				cmd->resp[0]);
+	}
+
+	return 0;
+}
+
+static void sd_send_cmd_get_rsp(struct rtsx_usb_sdmmc *host,
+		struct mmc_command *cmd)
+{
+	struct rtsx_ucr *ucr = host->ucr;
+	u8 cmd_idx = (u8)cmd->opcode;
+	u32 arg = cmd->arg;
+	int err = 0;
+	int timeout = 100;
+	int i;
+	u8 *ptr;
+	int stat_idx = 0;
+	int len = 2;
+	u8 rsp_type;
+
+	dev_dbg(sdmmc_dev(host), "%s: SD/MMC CMD %d, arg = 0x%08x\n",
+			__func__, cmd_idx, arg);
+
+	/* Response type:
+	 * R0
+	 * R1, R5, R6, R7
+	 * R1b
+	 * R2
+	 * R3, R4
+	 */
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE:
+		rsp_type = SD_RSP_TYPE_R0;
+		break;
+	case MMC_RSP_R1:
+		rsp_type = SD_RSP_TYPE_R1;
+		break;
+	case MMC_RSP_R1_NO_CRC:
+		rsp_type = SD_RSP_TYPE_R1 | SD_NO_CHECK_CRC7;
+		break;
+	case MMC_RSP_R1B:
+		rsp_type = SD_RSP_TYPE_R1b;
+		break;
+	case MMC_RSP_R2:
+		rsp_type = SD_RSP_TYPE_R2;
+		break;
+	case MMC_RSP_R3:
+		rsp_type = SD_RSP_TYPE_R3;
+		break;
+	default:
+		dev_dbg(sdmmc_dev(host), "cmd->flag is not valid\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	if (rsp_type == SD_RSP_TYPE_R1b)
+		timeout = cmd->busy_timeout ? cmd->busy_timeout : 3000;
+
+	if (cmd->opcode == SD_SWITCH_VOLTAGE) {
+		err = rtsx_usb_write_register(ucr, SD_BUS_STAT,
+				SD_CLK_TOGGLE_EN | SD_CLK_FORCE_STOP,
+				SD_CLK_TOGGLE_EN);
+		if (err)
+			goto out;
+	}
+
+	rtsx_usb_init_cmd(ucr);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_CMD0, 0xFF, 0x40 | cmd_idx);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_CMD1, 0xFF, (u8)(arg >> 24));
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_CMD2, 0xFF, (u8)(arg >> 16));
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_CMD3, 0xFF, (u8)(arg >> 8));
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_CMD4, 0xFF, (u8)arg);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_CFG2, 0xFF, rsp_type);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_DATA_SOURCE,
+			0x01, PINGPONG_BUFFER);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_TRANSFER,
+			0xFF, SD_TM_CMD_RSP | SD_TRANSFER_START);
+	rtsx_usb_add_cmd(ucr, CHECK_REG_CMD, SD_TRANSFER,
+		     SD_TRANSFER_END | SD_STAT_IDLE,
+		     SD_TRANSFER_END | SD_STAT_IDLE);
+
+	if (rsp_type == SD_RSP_TYPE_R2) {
+		/* Read data from ping-pong buffer */
+		for (i = PPBUF_BASE2; i < PPBUF_BASE2 + 16; i++)
+			rtsx_usb_add_cmd(ucr, READ_REG_CMD, (u16)i, 0, 0);
+		stat_idx = 16;
+	} else if (rsp_type != SD_RSP_TYPE_R0) {
+		/* Read data from SD_CMDx registers */
+		for (i = SD_CMD0; i <= SD_CMD4; i++)
+			rtsx_usb_add_cmd(ucr, READ_REG_CMD, (u16)i, 0, 0);
+		stat_idx = 5;
+	}
+	len += stat_idx;
+
+	rtsx_usb_add_cmd(ucr, READ_REG_CMD, SD_STAT1, 0, 0);
+
+	err = rtsx_usb_send_cmd(ucr, MODE_CR, 100);
+	if (err) {
+		dev_dbg(sdmmc_dev(host),
+			"rtsx_usb_send_cmd error (err = %d)\n", err);
+		goto out;
+	}
+
+	err = rtsx_usb_get_rsp(ucr, len, timeout);
+	if (err || (ucr->rsp_buf[0] & SD_TRANSFER_ERR)) {
+		sd_print_debug_regs(host);
+		sd_clear_error(host);
+
+		if (!err) {
+			dev_dbg(sdmmc_dev(host),
+				"Transfer failed (SD_TRANSFER = %02x)\n",
+					ucr->rsp_buf[0]);
+			err = -EIO;
+		} else {
+			dev_dbg(sdmmc_dev(host),
+				"rtsx_usb_get_rsp failed (err = %d)\n", err);
+		}
+
+		goto out;
+	}
+
+	if (rsp_type == SD_RSP_TYPE_R0) {
+		err = 0;
+		goto out;
+	}
+
+	/* Skip result of CHECK_REG_CMD */
+	ptr = ucr->rsp_buf + 1;
+
+	/* Check (Start,Transmission) bit of Response */
+	if ((ptr[0] & 0xC0) != 0) {
+		err = -EILSEQ;
+		dev_dbg(sdmmc_dev(host), "Invalid response bit\n");
+		goto out;
+	}
+
+	/* Check CRC7 */
+	if (!(rsp_type & SD_NO_CHECK_CRC7)) {
+		if (ptr[stat_idx] & SD_CRC7_ERR) {
+			err = -EILSEQ;
+			dev_dbg(sdmmc_dev(host), "CRC7 error\n");
+			goto out;
+		}
+	}
+
+	if (rsp_type == SD_RSP_TYPE_R2) {
+		/*
+		 * The controller offloads the last byte {CRC-7, end bit 1'b1}
+		 * of response type R2. Assign dummy CRC, 0, and end bit to the
+		 * byte(ptr[16], goes into the LSB of resp[3] later).
+		 */
+		ptr[16] = 1;
+
+		for (i = 0; i < 4; i++) {
+			cmd->resp[i] = get_unaligned_be32(ptr + 1 + i * 4);
+			dev_dbg(sdmmc_dev(host), "cmd->resp[%d] = 0x%08x\n",
+					i, cmd->resp[i]);
+		}
+	} else {
+		cmd->resp[0] = get_unaligned_be32(ptr + 1);
+		dev_dbg(sdmmc_dev(host), "cmd->resp[0] = 0x%08x\n",
+				cmd->resp[0]);
+	}
+
+out:
+	cmd->error = err;
+}
+
+static int sd_rw_multi(struct rtsx_usb_sdmmc *host, struct mmc_request *mrq)
+{
+	struct rtsx_ucr *ucr = host->ucr;
+	struct mmc_data *data = mrq->data;
+	int read = (data->flags & MMC_DATA_READ) ? 1 : 0;
+	u8 cfg2, trans_mode;
+	int err;
+	u8 flag;
+	size_t data_len = data->blksz * data->blocks;
+	unsigned int pipe;
+
+	if (read) {
+		dev_dbg(sdmmc_dev(host), "%s: read %zu bytes\n",
+				__func__, data_len);
+		cfg2 = SD_CALCULATE_CRC7 | SD_CHECK_CRC16 |
+			SD_NO_WAIT_BUSY_END | SD_CHECK_CRC7 | SD_RSP_LEN_0;
+		trans_mode = SD_TM_AUTO_READ_3;
+	} else {
+		dev_dbg(sdmmc_dev(host), "%s: write %zu bytes\n",
+				__func__, data_len);
+		cfg2 = SD_NO_CALCULATE_CRC7 | SD_CHECK_CRC16 |
+			SD_NO_WAIT_BUSY_END | SD_NO_CHECK_CRC7 | SD_RSP_LEN_0;
+		trans_mode = SD_TM_AUTO_WRITE_3;
+	}
+
+	rtsx_usb_init_cmd(ucr);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_BYTE_CNT_L, 0xFF, 0x00);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_BYTE_CNT_H, 0xFF, 0x02);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_BLOCK_CNT_L,
+			0xFF, (u8)data->blocks);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_BLOCK_CNT_H,
+			0xFF, (u8)(data->blocks >> 8));
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_DATA_SOURCE,
+			0x01, RING_BUFFER);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, MC_DMA_TC3,
+			0xFF, (u8)(data_len >> 24));
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, MC_DMA_TC2,
+			0xFF, (u8)(data_len >> 16));
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, MC_DMA_TC1,
+			0xFF, (u8)(data_len >> 8));
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, MC_DMA_TC0,
+			0xFF, (u8)data_len);
+	if (read) {
+		flag = MODE_CDIR;
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, MC_DMA_CTL,
+				0x03 | DMA_PACK_SIZE_MASK,
+				DMA_DIR_FROM_CARD | DMA_EN | DMA_512);
+	} else {
+		flag = MODE_CDOR;
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, MC_DMA_CTL,
+				0x03 | DMA_PACK_SIZE_MASK,
+				DMA_DIR_TO_CARD | DMA_EN | DMA_512);
+	}
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_CFG2, 0xFF, cfg2);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_TRANSFER, 0xFF,
+			trans_mode | SD_TRANSFER_START);
+	rtsx_usb_add_cmd(ucr, CHECK_REG_CMD, SD_TRANSFER,
+			SD_TRANSFER_END, SD_TRANSFER_END);
+
+	err = rtsx_usb_send_cmd(ucr, flag, 100);
+	if (err)
+		return err;
+
+	if (read)
+		pipe = usb_rcvbulkpipe(ucr->pusb_dev, EP_BULK_IN);
+	else
+		pipe = usb_sndbulkpipe(ucr->pusb_dev, EP_BULK_OUT);
+
+	err = rtsx_usb_transfer_data(ucr, pipe, data->sg, data_len,
+			data->sg_len,  NULL, 10000);
+	if (err) {
+		dev_dbg(sdmmc_dev(host), "rtsx_usb_transfer_data error %d\n"
+				, err);
+		sd_clear_error(host);
+		return err;
+	}
+
+	return rtsx_usb_get_rsp(ucr, 1, 2000);
+}
+
+static inline void sd_enable_initial_mode(struct rtsx_usb_sdmmc *host)
+{
+	rtsx_usb_write_register(host->ucr, SD_CFG1,
+			SD_CLK_DIVIDE_MASK, SD_CLK_DIVIDE_128);
+}
+
+static inline void sd_disable_initial_mode(struct rtsx_usb_sdmmc *host)
+{
+	rtsx_usb_write_register(host->ucr, SD_CFG1,
+			SD_CLK_DIVIDE_MASK, SD_CLK_DIVIDE_0);
+}
+
+static void sd_normal_rw(struct rtsx_usb_sdmmc *host,
+		struct mmc_request *mrq)
+{
+	struct mmc_command *cmd = mrq->cmd;
+	struct mmc_data *data = mrq->data;
+	u8 *buf;
+
+	buf = kzalloc(data->blksz, GFP_NOIO);
+	if (!buf) {
+		cmd->error = -ENOMEM;
+		return;
+	}
+
+	if (data->flags & MMC_DATA_READ) {
+		if (host->initial_mode)
+			sd_disable_initial_mode(host);
+
+		cmd->error = sd_read_data(host, cmd, (u16)data->blksz, buf,
+				data->blksz, 200);
+
+		if (host->initial_mode)
+			sd_enable_initial_mode(host);
+
+		sg_copy_from_buffer(data->sg, data->sg_len, buf, data->blksz);
+	} else {
+		sg_copy_to_buffer(data->sg, data->sg_len, buf, data->blksz);
+
+		cmd->error = sd_write_data(host, cmd, (u16)data->blksz, buf,
+				data->blksz, 200);
+	}
+
+	kfree(buf);
+}
+
+static int sd_change_phase(struct rtsx_usb_sdmmc *host, u8 sample_point, int tx)
+{
+	struct rtsx_ucr *ucr = host->ucr;
+
+	dev_dbg(sdmmc_dev(host), "%s: %s sample_point = %d\n",
+			__func__, tx ? "TX" : "RX", sample_point);
+
+	rtsx_usb_init_cmd(ucr);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CLK_DIV, CLK_CHANGE, CLK_CHANGE);
+
+	if (tx)
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_VPCLK0_CTL,
+				0x0F, sample_point);
+	else
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_VPCLK1_CTL,
+				0x0F, sample_point);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_VPCLK0_CTL, PHASE_NOT_RESET, 0);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_VPCLK0_CTL,
+			PHASE_NOT_RESET, PHASE_NOT_RESET);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CLK_DIV, CLK_CHANGE, 0);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_CFG1, SD_ASYNC_FIFO_RST, 0);
+
+	return rtsx_usb_send_cmd(ucr, MODE_C, 100);
+}
+
+static inline u32 get_phase_point(u32 phase_map, unsigned int idx)
+{
+	idx &= MAX_PHASE;
+	return phase_map & (1 << idx);
+}
+
+static int get_phase_len(u32 phase_map, unsigned int idx)
+{
+	int i;
+
+	for (i = 0; i < MAX_PHASE + 1; i++) {
+		if (get_phase_point(phase_map, idx + i) == 0)
+			return i;
+	}
+	return MAX_PHASE + 1;
+}
+
+static u8 sd_search_final_phase(struct rtsx_usb_sdmmc *host, u32 phase_map)
+{
+	int start = 0, len = 0;
+	int start_final = 0, len_final = 0;
+	u8 final_phase = 0xFF;
+
+	if (phase_map == 0) {
+		dev_dbg(sdmmc_dev(host), "Phase: [map:%x]\n", phase_map);
+		return final_phase;
+	}
+
+	while (start < MAX_PHASE + 1) {
+		len = get_phase_len(phase_map, start);
+		if (len_final < len) {
+			start_final = start;
+			len_final = len;
+		}
+		start += len ? len : 1;
+	}
+
+	final_phase = (start_final + len_final / 2) & MAX_PHASE;
+	dev_dbg(sdmmc_dev(host), "Phase: [map:%x] [maxlen:%d] [final:%d]\n",
+		phase_map, len_final, final_phase);
+
+	return final_phase;
+}
+
+static void sd_wait_data_idle(struct rtsx_usb_sdmmc *host)
+{
+	int i;
+	u8 val = 0;
+
+	for (i = 0; i < 100; i++) {
+		rtsx_usb_ep0_read_register(host->ucr, SD_DATA_STATE, &val);
+		if (val & SD_DATA_IDLE)
+			return;
+
+		usleep_range(100, 1000);
+	}
+}
+
+static int sd_tuning_rx_cmd(struct rtsx_usb_sdmmc *host,
+		u8 opcode, u8 sample_point)
+{
+	int err;
+	struct mmc_command cmd = {};
+
+	err = sd_change_phase(host, sample_point, 0);
+	if (err)
+		return err;
+
+	cmd.opcode = MMC_SEND_TUNING_BLOCK;
+	err = sd_read_data(host, &cmd, 0x40, NULL, 0, 100);
+	if (err) {
+		/* Wait till SD DATA IDLE */
+		sd_wait_data_idle(host);
+		sd_clear_error(host);
+		return err;
+	}
+
+	return 0;
+}
+
+static void sd_tuning_phase(struct rtsx_usb_sdmmc *host,
+		u8 opcode, u16 *phase_map)
+{
+	int err, i;
+	u16 raw_phase_map = 0;
+
+	for (i = MAX_PHASE; i >= 0; i--) {
+		err = sd_tuning_rx_cmd(host, opcode, (u8)i);
+		if (!err)
+			raw_phase_map |= 1 << i;
+	}
+
+	if (phase_map)
+		*phase_map = raw_phase_map;
+}
+
+static int sd_tuning_rx(struct rtsx_usb_sdmmc *host, u8 opcode)
+{
+	int err, i;
+	u16 raw_phase_map[RX_TUNING_CNT] = {0}, phase_map;
+	u8 final_phase;
+
+	/* setting fixed default TX phase */
+	err = sd_change_phase(host, 0x01, 1);
+	if (err) {
+		dev_dbg(sdmmc_dev(host), "TX phase setting failed\n");
+		return err;
+	}
+
+	/* tuning RX phase */
+	for (i = 0; i < RX_TUNING_CNT; i++) {
+		sd_tuning_phase(host, opcode, &(raw_phase_map[i]));
+
+		if (raw_phase_map[i] == 0)
+			break;
+	}
+
+	phase_map = 0xFFFF;
+	for (i = 0; i < RX_TUNING_CNT; i++) {
+		dev_dbg(sdmmc_dev(host), "RX raw_phase_map[%d] = 0x%04x\n",
+				i, raw_phase_map[i]);
+		phase_map &= raw_phase_map[i];
+	}
+	dev_dbg(sdmmc_dev(host), "RX phase_map = 0x%04x\n", phase_map);
+
+	if (phase_map) {
+		final_phase = sd_search_final_phase(host, phase_map);
+		if (final_phase == 0xFF)
+			return -EINVAL;
+
+		err = sd_change_phase(host, final_phase, 0);
+		if (err)
+			return err;
+	} else {
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int sdmmc_get_ro(struct mmc_host *mmc)
+{
+	struct rtsx_usb_sdmmc *host = mmc_priv(mmc);
+	struct rtsx_ucr *ucr = host->ucr;
+	int err;
+	u16 val;
+
+	if (host->host_removal)
+		return -ENOMEDIUM;
+
+	mutex_lock(&ucr->dev_mutex);
+
+	/* Check SD card detect */
+	err = rtsx_usb_get_card_status(ucr, &val);
+
+	mutex_unlock(&ucr->dev_mutex);
+
+
+	/* Treat failed detection as non-ro */
+	if (err)
+		return 0;
+
+	if (val & SD_WP)
+		return 1;
+
+	return 0;
+}
+
+static int sdmmc_get_cd(struct mmc_host *mmc)
+{
+	struct rtsx_usb_sdmmc *host = mmc_priv(mmc);
+	struct rtsx_ucr *ucr = host->ucr;
+	int err;
+	u16 val;
+
+	if (host->host_removal)
+		return -ENOMEDIUM;
+
+	mutex_lock(&ucr->dev_mutex);
+
+	/* Check SD card detect */
+	err = rtsx_usb_get_card_status(ucr, &val);
+
+	mutex_unlock(&ucr->dev_mutex);
+
+	/* Treat failed detection as non-exist */
+	if (err)
+		goto no_card;
+
+	if (val & SD_CD) {
+		host->card_exist = true;
+		return 1;
+	}
+
+no_card:
+	host->card_exist = false;
+	return 0;
+}
+
+static void sdmmc_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct rtsx_usb_sdmmc *host = mmc_priv(mmc);
+	struct rtsx_ucr *ucr = host->ucr;
+	struct mmc_command *cmd = mrq->cmd;
+	struct mmc_data *data = mrq->data;
+	unsigned int data_size = 0;
+
+	dev_dbg(sdmmc_dev(host), "%s\n", __func__);
+
+	if (host->host_removal) {
+		cmd->error = -ENOMEDIUM;
+		goto finish;
+	}
+
+	if ((!host->card_exist)) {
+		cmd->error = -ENOMEDIUM;
+		goto finish_detect_card;
+	}
+
+	mutex_lock(&ucr->dev_mutex);
+
+	mutex_lock(&host->host_mutex);
+	host->mrq = mrq;
+	mutex_unlock(&host->host_mutex);
+
+	if (mrq->data)
+		data_size = data->blocks * data->blksz;
+
+	if (!data_size) {
+		sd_send_cmd_get_rsp(host, cmd);
+	} else if ((!(data_size % 512) && cmd->opcode != MMC_SEND_EXT_CSD) ||
+		   mmc_op_multi(cmd->opcode)) {
+		sd_send_cmd_get_rsp(host, cmd);
+
+		if (!cmd->error) {
+			sd_rw_multi(host, mrq);
+
+			if (mmc_op_multi(cmd->opcode) && mrq->stop) {
+				sd_send_cmd_get_rsp(host, mrq->stop);
+				rtsx_usb_write_register(ucr, MC_FIFO_CTL,
+						FIFO_FLUSH, FIFO_FLUSH);
+			}
+		}
+	} else {
+		sd_normal_rw(host, mrq);
+	}
+
+	if (mrq->data) {
+		if (cmd->error || data->error)
+			data->bytes_xfered = 0;
+		else
+			data->bytes_xfered = data->blocks * data->blksz;
+	}
+
+	mutex_unlock(&ucr->dev_mutex);
+
+finish_detect_card:
+	if (cmd->error) {
+		/*
+		 * detect card when fail to update card existence state and
+		 * speed up card removal when retry
+		 */
+		sdmmc_get_cd(mmc);
+		dev_dbg(sdmmc_dev(host), "cmd->error = %d\n", cmd->error);
+	}
+
+finish:
+	mutex_lock(&host->host_mutex);
+	host->mrq = NULL;
+	mutex_unlock(&host->host_mutex);
+
+	mmc_request_done(mmc, mrq);
+}
+
+static int sd_set_bus_width(struct rtsx_usb_sdmmc *host,
+		unsigned char bus_width)
+{
+	int err = 0;
+	static const u8 width[] = {
+		[MMC_BUS_WIDTH_1] = SD_BUS_WIDTH_1BIT,
+		[MMC_BUS_WIDTH_4] = SD_BUS_WIDTH_4BIT,
+		[MMC_BUS_WIDTH_8] = SD_BUS_WIDTH_8BIT,
+	};
+
+	if (bus_width <= MMC_BUS_WIDTH_8)
+		err = rtsx_usb_write_register(host->ucr, SD_CFG1,
+				0x03, width[bus_width]);
+
+	return err;
+}
+
+static int sd_pull_ctl_disable_lqfp48(struct rtsx_ucr *ucr)
+{
+	rtsx_usb_init_cmd(ucr);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL1, 0xFF, 0x55);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL2, 0xFF, 0x55);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL3, 0xFF, 0x95);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL4, 0xFF, 0x55);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL5, 0xFF, 0x55);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL6, 0xFF, 0xA5);
+
+	return rtsx_usb_send_cmd(ucr, MODE_C, 100);
+}
+
+static int sd_pull_ctl_disable_qfn24(struct rtsx_ucr *ucr)
+{
+	rtsx_usb_init_cmd(ucr);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL1, 0xFF, 0x65);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL2, 0xFF, 0x55);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL3, 0xFF, 0x95);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL4, 0xFF, 0x55);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL5, 0xFF, 0x56);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL6, 0xFF, 0x59);
+
+	return rtsx_usb_send_cmd(ucr, MODE_C, 100);
+}
+
+static int sd_pull_ctl_enable_lqfp48(struct rtsx_ucr *ucr)
+{
+	rtsx_usb_init_cmd(ucr);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL1, 0xFF, 0xAA);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL2, 0xFF, 0xAA);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL3, 0xFF, 0xA9);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL4, 0xFF, 0x55);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL5, 0xFF, 0x55);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL6, 0xFF, 0xA5);
+
+	return rtsx_usb_send_cmd(ucr, MODE_C, 100);
+}
+
+static int sd_pull_ctl_enable_qfn24(struct rtsx_ucr *ucr)
+{
+	rtsx_usb_init_cmd(ucr);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL1, 0xFF, 0xA5);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL2, 0xFF, 0x9A);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL3, 0xFF, 0xA5);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL4, 0xFF, 0x9A);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL5, 0xFF, 0x65);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PULL_CTL6, 0xFF, 0x5A);
+
+	return rtsx_usb_send_cmd(ucr, MODE_C, 100);
+}
+
+static int sd_power_on(struct rtsx_usb_sdmmc *host)
+{
+	struct rtsx_ucr *ucr = host->ucr;
+	int err;
+
+	dev_dbg(sdmmc_dev(host), "%s\n", __func__);
+	rtsx_usb_init_cmd(ucr);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_SELECT, 0x07, SD_MOD_SEL);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_SHARE_MODE,
+			CARD_SHARE_MASK, CARD_SHARE_SD);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_CLK_EN,
+			SD_CLK_EN, SD_CLK_EN);
+	err = rtsx_usb_send_cmd(ucr, MODE_C, 100);
+	if (err)
+		return err;
+
+	if (CHECK_PKG(ucr, LQFP48))
+		err = sd_pull_ctl_enable_lqfp48(ucr);
+	else
+		err = sd_pull_ctl_enable_qfn24(ucr);
+	if (err)
+		return err;
+
+	err = rtsx_usb_write_register(ucr, CARD_PWR_CTL,
+			POWER_MASK, PARTIAL_POWER_ON);
+	if (err)
+		return err;
+
+	usleep_range(800, 1000);
+
+	rtsx_usb_init_cmd(ucr);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PWR_CTL,
+			POWER_MASK|LDO3318_PWR_MASK, POWER_ON|LDO_ON);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_OE,
+			SD_OUTPUT_EN, SD_OUTPUT_EN);
+
+	return rtsx_usb_send_cmd(ucr, MODE_C, 100);
+}
+
+static int sd_power_off(struct rtsx_usb_sdmmc *host)
+{
+	struct rtsx_ucr *ucr = host->ucr;
+	int err;
+
+	dev_dbg(sdmmc_dev(host), "%s\n", __func__);
+	rtsx_usb_init_cmd(ucr);
+
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_CLK_EN, SD_CLK_EN, 0);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_OE, SD_OUTPUT_EN, 0);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PWR_CTL,
+			POWER_MASK, POWER_OFF);
+	rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_PWR_CTL,
+			POWER_MASK|LDO3318_PWR_MASK, POWER_OFF|LDO_SUSPEND);
+
+	err = rtsx_usb_send_cmd(ucr, MODE_C, 100);
+	if (err)
+		return err;
+
+	if (CHECK_PKG(ucr, LQFP48))
+			return sd_pull_ctl_disable_lqfp48(ucr);
+	return sd_pull_ctl_disable_qfn24(ucr);
+}
+
+static int sd_set_power_mode(struct rtsx_usb_sdmmc *host,
+		unsigned char power_mode)
+{
+	int err;
+
+	if (power_mode != MMC_POWER_OFF)
+		power_mode = MMC_POWER_ON;
+
+	if (power_mode == host->power_mode)
+		return 0;
+
+	if (power_mode == MMC_POWER_OFF) {
+		err = sd_power_off(host);
+		pm_runtime_put_noidle(sdmmc_dev(host));
+	} else {
+		pm_runtime_get_noresume(sdmmc_dev(host));
+		err = sd_power_on(host);
+	}
+
+	if (!err)
+		host->power_mode = power_mode;
+
+	return err;
+}
+
+static int sd_set_timing(struct rtsx_usb_sdmmc *host,
+		unsigned char timing, bool *ddr_mode)
+{
+	struct rtsx_ucr *ucr = host->ucr;
+	int err;
+
+	*ddr_mode = false;
+
+	rtsx_usb_init_cmd(ucr);
+
+	switch (timing) {
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_UHS_SDR50:
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_CFG1,
+				0x0C | SD_ASYNC_FIFO_RST,
+				SD_30_MODE | SD_ASYNC_FIFO_RST);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_CLK_SOURCE, 0xFF,
+				CRC_VAR_CLK0 | SD30_FIX_CLK | SAMPLE_VAR_CLK1);
+		break;
+
+	case MMC_TIMING_UHS_DDR50:
+		*ddr_mode = true;
+
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_CFG1,
+				0x0C | SD_ASYNC_FIFO_RST,
+				SD_DDR_MODE | SD_ASYNC_FIFO_RST);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_CLK_SOURCE, 0xFF,
+				CRC_VAR_CLK0 | SD30_FIX_CLK | SAMPLE_VAR_CLK1);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_PUSH_POINT_CTL,
+				DDR_VAR_TX_CMD_DAT, DDR_VAR_TX_CMD_DAT);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_SAMPLE_POINT_CTL,
+				DDR_VAR_RX_DAT | DDR_VAR_RX_CMD,
+				DDR_VAR_RX_DAT | DDR_VAR_RX_CMD);
+		break;
+
+	case MMC_TIMING_MMC_HS:
+	case MMC_TIMING_SD_HS:
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_CFG1,
+				0x0C, SD_20_MODE);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_CLK_SOURCE, 0xFF,
+				CRC_FIX_CLK | SD30_VAR_CLK0 | SAMPLE_VAR_CLK1);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_PUSH_POINT_CTL,
+				SD20_TX_SEL_MASK, SD20_TX_14_AHEAD);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_SAMPLE_POINT_CTL,
+				SD20_RX_SEL_MASK, SD20_RX_14_DELAY);
+		break;
+
+	default:
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+				SD_CFG1, 0x0C, SD_20_MODE);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, CARD_CLK_SOURCE, 0xFF,
+				CRC_FIX_CLK | SD30_VAR_CLK0 | SAMPLE_VAR_CLK1);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD,
+				SD_PUSH_POINT_CTL, 0xFF, 0);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_SAMPLE_POINT_CTL,
+				SD20_RX_SEL_MASK, SD20_RX_POS_EDGE);
+		break;
+	}
+
+	err = rtsx_usb_send_cmd(ucr, MODE_C, 100);
+
+	return err;
+}
+
+static void sdmmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct rtsx_usb_sdmmc *host = mmc_priv(mmc);
+	struct rtsx_ucr *ucr = host->ucr;
+
+	dev_dbg(sdmmc_dev(host), "%s\n", __func__);
+	mutex_lock(&ucr->dev_mutex);
+
+	sd_set_power_mode(host, ios->power_mode);
+	sd_set_bus_width(host, ios->bus_width);
+	sd_set_timing(host, ios->timing, &host->ddr_mode);
+
+	host->vpclk = false;
+	host->double_clk = true;
+
+	switch (ios->timing) {
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_UHS_SDR50:
+		host->ssc_depth = SSC_DEPTH_2M;
+		host->vpclk = true;
+		host->double_clk = false;
+		break;
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_UHS_SDR25:
+		host->ssc_depth = SSC_DEPTH_1M;
+		break;
+	default:
+		host->ssc_depth = SSC_DEPTH_512K;
+		break;
+	}
+
+	host->initial_mode = (ios->clock <= 1000000) ? true : false;
+	host->clock = ios->clock;
+
+	rtsx_usb_switch_clock(host->ucr, host->clock, host->ssc_depth,
+			host->initial_mode, host->double_clk, host->vpclk);
+
+	mutex_unlock(&ucr->dev_mutex);
+	dev_dbg(sdmmc_dev(host), "%s end\n", __func__);
+}
+
+static int sdmmc_switch_voltage(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct rtsx_usb_sdmmc *host = mmc_priv(mmc);
+	struct rtsx_ucr *ucr = host->ucr;
+	int err = 0;
+
+	dev_dbg(sdmmc_dev(host), "%s: signal_voltage = %d\n",
+			__func__, ios->signal_voltage);
+
+	if (host->host_removal)
+		return -ENOMEDIUM;
+
+	if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_120)
+		return -EPERM;
+
+	mutex_lock(&ucr->dev_mutex);
+
+	err = rtsx_usb_card_exclusive_check(ucr, RTSX_USB_SD_CARD);
+	if (err) {
+		mutex_unlock(&ucr->dev_mutex);
+		return err;
+	}
+
+	/* Let mmc core do the busy checking, simply stop the forced-toggle
+	 * clock(while issuing CMD11) and switch voltage.
+	 */
+	rtsx_usb_init_cmd(ucr);
+
+	if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_330) {
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_PAD_CTL,
+				SD_IO_USING_1V8, SD_IO_USING_3V3);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, LDO_POWER_CFG,
+				TUNE_SD18_MASK, TUNE_SD18_3V3);
+	} else {
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_BUS_STAT,
+				SD_CLK_TOGGLE_EN | SD_CLK_FORCE_STOP,
+				SD_CLK_FORCE_STOP);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, SD_PAD_CTL,
+				SD_IO_USING_1V8, SD_IO_USING_1V8);
+		rtsx_usb_add_cmd(ucr, WRITE_REG_CMD, LDO_POWER_CFG,
+				TUNE_SD18_MASK, TUNE_SD18_1V8);
+	}
+
+	err = rtsx_usb_send_cmd(ucr, MODE_C, 100);
+	mutex_unlock(&ucr->dev_mutex);
+
+	return err;
+}
+
+static int sdmmc_card_busy(struct mmc_host *mmc)
+{
+	struct rtsx_usb_sdmmc *host = mmc_priv(mmc);
+	struct rtsx_ucr *ucr = host->ucr;
+	int err;
+	u8 stat;
+	u8 mask = SD_DAT3_STATUS | SD_DAT2_STATUS | SD_DAT1_STATUS
+		| SD_DAT0_STATUS;
+
+	dev_dbg(sdmmc_dev(host), "%s\n", __func__);
+
+	mutex_lock(&ucr->dev_mutex);
+
+	err = rtsx_usb_write_register(ucr, SD_BUS_STAT,
+			SD_CLK_TOGGLE_EN | SD_CLK_FORCE_STOP,
+			SD_CLK_TOGGLE_EN);
+	if (err)
+		goto out;
+
+	mdelay(1);
+
+	err = rtsx_usb_read_register(ucr, SD_BUS_STAT, &stat);
+	if (err)
+		goto out;
+
+	err = rtsx_usb_write_register(ucr, SD_BUS_STAT,
+			SD_CLK_TOGGLE_EN | SD_CLK_FORCE_STOP, 0);
+out:
+	mutex_unlock(&ucr->dev_mutex);
+
+	if (err)
+		return err;
+
+	/* check if any pin between dat[0:3] is low */
+	if ((stat & mask) != mask)
+		return 1;
+	else
+		return 0;
+}
+
+static int sdmmc_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct rtsx_usb_sdmmc *host = mmc_priv(mmc);
+	struct rtsx_ucr *ucr = host->ucr;
+	int err = 0;
+
+	if (host->host_removal)
+		return -ENOMEDIUM;
+
+	mutex_lock(&ucr->dev_mutex);
+
+	if (!host->ddr_mode)
+		err = sd_tuning_rx(host, MMC_SEND_TUNING_BLOCK);
+
+	mutex_unlock(&ucr->dev_mutex);
+
+	return err;
+}
+
+static const struct mmc_host_ops rtsx_usb_sdmmc_ops = {
+	.request = sdmmc_request,
+	.set_ios = sdmmc_set_ios,
+	.get_ro = sdmmc_get_ro,
+	.get_cd = sdmmc_get_cd,
+	.start_signal_voltage_switch = sdmmc_switch_voltage,
+	.card_busy = sdmmc_card_busy,
+	.execute_tuning = sdmmc_execute_tuning,
+};
+
+#ifdef RTSX_USB_USE_LEDS_CLASS
+static void rtsx_usb_led_control(struct led_classdev *led,
+	enum led_brightness brightness)
+{
+	struct rtsx_usb_sdmmc *host = container_of(led,
+			struct rtsx_usb_sdmmc, led);
+
+	if (host->host_removal)
+		return;
+
+	host->led.brightness = brightness;
+	schedule_work(&host->led_work);
+}
+
+static void rtsx_usb_update_led(struct work_struct *work)
+{
+	struct rtsx_usb_sdmmc *host =
+		container_of(work, struct rtsx_usb_sdmmc, led_work);
+	struct rtsx_ucr *ucr = host->ucr;
+
+	pm_runtime_get_noresume(sdmmc_dev(host));
+	mutex_lock(&ucr->dev_mutex);
+
+	if (host->power_mode == MMC_POWER_OFF)
+		goto out;
+
+	if (host->led.brightness == LED_OFF)
+		rtsx_usb_turn_off_led(ucr);
+	else
+		rtsx_usb_turn_on_led(ucr);
+
+out:
+	mutex_unlock(&ucr->dev_mutex);
+	pm_runtime_put_sync_suspend(sdmmc_dev(host));
+}
+#endif
+
+static void rtsx_usb_init_host(struct rtsx_usb_sdmmc *host)
+{
+	struct mmc_host *mmc = host->mmc;
+
+	mmc->f_min = 250000;
+	mmc->f_max = 208000000;
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34 | MMC_VDD_165_195;
+	mmc->caps = MMC_CAP_4_BIT_DATA | MMC_CAP_SD_HIGHSPEED |
+		MMC_CAP_MMC_HIGHSPEED | MMC_CAP_BUS_WIDTH_TEST |
+		MMC_CAP_UHS_SDR12 | MMC_CAP_UHS_SDR25 | MMC_CAP_UHS_SDR50 |
+		MMC_CAP_SYNC_RUNTIME_PM;
+	mmc->caps2 = MMC_CAP2_NO_PRESCAN_POWERUP | MMC_CAP2_FULL_PWR_CYCLE |
+		MMC_CAP2_NO_SDIO;
+
+	mmc->max_current_330 = 400;
+	mmc->max_current_180 = 800;
+	mmc->ops = &rtsx_usb_sdmmc_ops;
+	mmc->max_segs = 256;
+	mmc->max_seg_size = 65536;
+	mmc->max_blk_size = 512;
+	mmc->max_blk_count = 65535;
+	mmc->max_req_size = 524288;
+
+	host->power_mode = MMC_POWER_OFF;
+}
+
+static int rtsx_usb_sdmmc_drv_probe(struct platform_device *pdev)
+{
+	struct mmc_host *mmc;
+	struct rtsx_usb_sdmmc *host;
+	struct rtsx_ucr *ucr;
+#ifdef RTSX_USB_USE_LEDS_CLASS
+	int err;
+#endif
+	int ret;
+
+	ucr = usb_get_intfdata(to_usb_interface(pdev->dev.parent));
+	if (!ucr)
+		return -ENXIO;
+
+	dev_dbg(&(pdev->dev), ": Realtek USB SD/MMC controller found\n");
+
+	mmc = mmc_alloc_host(sizeof(*host), &pdev->dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	host = mmc_priv(mmc);
+	host->ucr = ucr;
+	host->mmc = mmc;
+	host->pdev = pdev;
+	platform_set_drvdata(pdev, host);
+
+	mutex_init(&host->host_mutex);
+	rtsx_usb_init_host(host);
+	pm_runtime_enable(&pdev->dev);
+
+#ifdef RTSX_USB_USE_LEDS_CLASS
+	snprintf(host->led_name, sizeof(host->led_name),
+		"%s::", mmc_hostname(mmc));
+	host->led.name = host->led_name;
+	host->led.brightness = LED_OFF;
+	host->led.default_trigger = mmc_hostname(mmc);
+	host->led.brightness_set = rtsx_usb_led_control;
+
+	err = led_classdev_register(mmc_dev(mmc), &host->led);
+	if (err)
+		dev_err(&(pdev->dev),
+				"Failed to register LED device: %d\n", err);
+	INIT_WORK(&host->led_work, rtsx_usb_update_led);
+
+#endif
+	ret = mmc_add_host(mmc);
+	if (ret) {
+#ifdef RTSX_USB_USE_LEDS_CLASS
+		led_classdev_unregister(&host->led);
+#endif
+		mmc_free_host(mmc);
+		pm_runtime_disable(&pdev->dev);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int rtsx_usb_sdmmc_drv_remove(struct platform_device *pdev)
+{
+	struct rtsx_usb_sdmmc *host = platform_get_drvdata(pdev);
+	struct mmc_host *mmc;
+
+	if (!host)
+		return 0;
+
+	mmc = host->mmc;
+	host->host_removal = true;
+
+	mutex_lock(&host->host_mutex);
+	if (host->mrq) {
+		dev_dbg(&(pdev->dev),
+			"%s: Controller removed during transfer\n",
+			mmc_hostname(mmc));
+		host->mrq->cmd->error = -ENOMEDIUM;
+		if (host->mrq->stop)
+			host->mrq->stop->error = -ENOMEDIUM;
+		mmc_request_done(mmc, host->mrq);
+	}
+	mutex_unlock(&host->host_mutex);
+
+	mmc_remove_host(mmc);
+
+#ifdef RTSX_USB_USE_LEDS_CLASS
+	cancel_work_sync(&host->led_work);
+	led_classdev_unregister(&host->led);
+#endif
+
+	mmc_free_host(mmc);
+	pm_runtime_disable(&pdev->dev);
+	platform_set_drvdata(pdev, NULL);
+
+	dev_dbg(&(pdev->dev),
+		": Realtek USB SD/MMC module has been removed\n");
+
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static int rtsx_usb_sdmmc_runtime_suspend(struct device *dev)
+{
+	struct rtsx_usb_sdmmc *host = dev_get_drvdata(dev);
+
+	host->mmc->caps &= ~MMC_CAP_NEEDS_POLL;
+	return 0;
+}
+
+static int rtsx_usb_sdmmc_runtime_resume(struct device *dev)
+{
+	struct rtsx_usb_sdmmc *host = dev_get_drvdata(dev);
+
+	host->mmc->caps |= MMC_CAP_NEEDS_POLL;
+	if (sdmmc_get_cd(host->mmc) == 1)
+		mmc_detect_change(host->mmc, 0);
+	return 0;
+}
+#endif
+
+static const struct dev_pm_ops rtsx_usb_sdmmc_dev_pm_ops = {
+	SET_RUNTIME_PM_OPS(rtsx_usb_sdmmc_runtime_suspend,
+			   rtsx_usb_sdmmc_runtime_resume, NULL)
+};
+
+static const struct platform_device_id rtsx_usb_sdmmc_ids[] = {
+	{
+		.name = "rtsx_usb_sdmmc",
+	}, {
+		/* sentinel */
+	}
+};
+MODULE_DEVICE_TABLE(platform, rtsx_usb_sdmmc_ids);
+
+static struct platform_driver rtsx_usb_sdmmc_driver = {
+	.probe		= rtsx_usb_sdmmc_drv_probe,
+	.remove		= rtsx_usb_sdmmc_drv_remove,
+	.id_table       = rtsx_usb_sdmmc_ids,
+	.driver		= {
+		.name	= "rtsx_usb_sdmmc",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= &rtsx_usb_sdmmc_dev_pm_ops,
+	},
+};
+module_platform_driver(rtsx_usb_sdmmc_driver);
+
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Roger Tseng <rogerable@realtek.com>");
+MODULE_DESCRIPTION("Realtek USB SD/MMC Card Host Driver");
diff --git a/drivers/mmc/host/s3cmci.c b/drivers/mmc/host/s3cmci.c
new file mode 100644
index 000000000..643d54ece
--- /dev/null
+++ b/drivers/mmc/host/s3cmci.c
@@ -0,0 +1,1782 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ *  linux/drivers/mmc/s3cmci.h - Samsung S3C MCI driver
+ *
+ *  Copyright (C) 2004-2006 maintech GmbH, Thomas Kleffel <tk@maintech.de>
+ *
+ * Current driver maintained by Ben Dooks and Simtec Electronics
+ *  Copyright (C) 2008 Simtec Electronics <ben-linux@fluff.org>
+ */
+
+#include <linux/module.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/clk.h>
+#include <linux/mmc/host.h>
+#include <linux/platform_device.h>
+#include <linux/cpufreq.h>
+#include <linux/debugfs.h>
+#include <linux/seq_file.h>
+#include <linux/gpio.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/io.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/platform_data/mmc-s3cmci.h>
+
+#include "s3cmci.h"
+
+#define DRIVER_NAME "s3c-mci"
+
+#define S3C2410_SDICON			(0x00)
+#define S3C2410_SDIPRE			(0x04)
+#define S3C2410_SDICMDARG		(0x08)
+#define S3C2410_SDICMDCON		(0x0C)
+#define S3C2410_SDICMDSTAT		(0x10)
+#define S3C2410_SDIRSP0			(0x14)
+#define S3C2410_SDIRSP1			(0x18)
+#define S3C2410_SDIRSP2			(0x1C)
+#define S3C2410_SDIRSP3			(0x20)
+#define S3C2410_SDITIMER		(0x24)
+#define S3C2410_SDIBSIZE		(0x28)
+#define S3C2410_SDIDCON			(0x2C)
+#define S3C2410_SDIDCNT			(0x30)
+#define S3C2410_SDIDSTA			(0x34)
+#define S3C2410_SDIFSTA			(0x38)
+
+#define S3C2410_SDIDATA			(0x3C)
+#define S3C2410_SDIIMSK			(0x40)
+
+#define S3C2440_SDIDATA			(0x40)
+#define S3C2440_SDIIMSK			(0x3C)
+
+#define S3C2440_SDICON_SDRESET		(1 << 8)
+#define S3C2410_SDICON_SDIOIRQ		(1 << 3)
+#define S3C2410_SDICON_FIFORESET	(1 << 1)
+#define S3C2410_SDICON_CLOCKTYPE	(1 << 0)
+
+#define S3C2410_SDICMDCON_LONGRSP	(1 << 10)
+#define S3C2410_SDICMDCON_WAITRSP	(1 << 9)
+#define S3C2410_SDICMDCON_CMDSTART	(1 << 8)
+#define S3C2410_SDICMDCON_SENDERHOST	(1 << 6)
+#define S3C2410_SDICMDCON_INDEX		(0x3f)
+
+#define S3C2410_SDICMDSTAT_CRCFAIL	(1 << 12)
+#define S3C2410_SDICMDSTAT_CMDSENT	(1 << 11)
+#define S3C2410_SDICMDSTAT_CMDTIMEOUT	(1 << 10)
+#define S3C2410_SDICMDSTAT_RSPFIN	(1 << 9)
+
+#define S3C2440_SDIDCON_DS_WORD		(2 << 22)
+#define S3C2410_SDIDCON_TXAFTERRESP	(1 << 20)
+#define S3C2410_SDIDCON_RXAFTERCMD	(1 << 19)
+#define S3C2410_SDIDCON_BLOCKMODE	(1 << 17)
+#define S3C2410_SDIDCON_WIDEBUS		(1 << 16)
+#define S3C2410_SDIDCON_DMAEN		(1 << 15)
+#define S3C2410_SDIDCON_STOP		(1 << 14)
+#define S3C2440_SDIDCON_DATSTART	(1 << 14)
+
+#define S3C2410_SDIDCON_XFER_RXSTART	(2 << 12)
+#define S3C2410_SDIDCON_XFER_TXSTART	(3 << 12)
+
+#define S3C2410_SDIDCON_BLKNUM_MASK	(0xFFF)
+
+#define S3C2410_SDIDSTA_SDIOIRQDETECT	(1 << 9)
+#define S3C2410_SDIDSTA_FIFOFAIL	(1 << 8)
+#define S3C2410_SDIDSTA_CRCFAIL		(1 << 7)
+#define S3C2410_SDIDSTA_RXCRCFAIL	(1 << 6)
+#define S3C2410_SDIDSTA_DATATIMEOUT	(1 << 5)
+#define S3C2410_SDIDSTA_XFERFINISH	(1 << 4)
+#define S3C2410_SDIDSTA_TXDATAON	(1 << 1)
+#define S3C2410_SDIDSTA_RXDATAON	(1 << 0)
+
+#define S3C2440_SDIFSTA_FIFORESET	(1 << 16)
+#define S3C2440_SDIFSTA_FIFOFAIL	(3 << 14)
+#define S3C2410_SDIFSTA_TFDET		(1 << 13)
+#define S3C2410_SDIFSTA_RFDET		(1 << 12)
+#define S3C2410_SDIFSTA_COUNTMASK	(0x7f)
+
+#define S3C2410_SDIIMSK_RESPONSECRC	(1 << 17)
+#define S3C2410_SDIIMSK_CMDSENT		(1 << 16)
+#define S3C2410_SDIIMSK_CMDTIMEOUT	(1 << 15)
+#define S3C2410_SDIIMSK_RESPONSEND	(1 << 14)
+#define S3C2410_SDIIMSK_SDIOIRQ		(1 << 12)
+#define S3C2410_SDIIMSK_FIFOFAIL	(1 << 11)
+#define S3C2410_SDIIMSK_CRCSTATUS	(1 << 10)
+#define S3C2410_SDIIMSK_DATACRC		(1 << 9)
+#define S3C2410_SDIIMSK_DATATIMEOUT	(1 << 8)
+#define S3C2410_SDIIMSK_DATAFINISH	(1 << 7)
+#define S3C2410_SDIIMSK_TXFIFOHALF	(1 << 4)
+#define S3C2410_SDIIMSK_RXFIFOLAST	(1 << 2)
+#define S3C2410_SDIIMSK_RXFIFOHALF	(1 << 0)
+
+enum dbg_channels {
+	dbg_err   = (1 << 0),
+	dbg_debug = (1 << 1),
+	dbg_info  = (1 << 2),
+	dbg_irq   = (1 << 3),
+	dbg_sg    = (1 << 4),
+	dbg_dma   = (1 << 5),
+	dbg_pio   = (1 << 6),
+	dbg_fail  = (1 << 7),
+	dbg_conf  = (1 << 8),
+};
+
+static const int dbgmap_err   = dbg_fail;
+static const int dbgmap_info  = dbg_info | dbg_conf;
+static const int dbgmap_debug = dbg_err | dbg_debug;
+
+#define dbg(host, channels, args...)		  \
+	do {					  \
+	if (dbgmap_err & channels) 		  \
+		dev_err(&host->pdev->dev, args);  \
+	else if (dbgmap_info & channels)	  \
+		dev_info(&host->pdev->dev, args); \
+	else if (dbgmap_debug & channels)	  \
+		dev_dbg(&host->pdev->dev, args);  \
+	} while (0)
+
+static void finalize_request(struct s3cmci_host *host);
+static void s3cmci_send_request(struct mmc_host *mmc);
+static void s3cmci_reset(struct s3cmci_host *host);
+
+#ifdef CONFIG_MMC_DEBUG
+
+static void dbg_dumpregs(struct s3cmci_host *host, char *prefix)
+{
+	u32 con, pre, cmdarg, cmdcon, cmdsta, r0, r1, r2, r3, timer;
+	u32 datcon, datcnt, datsta, fsta;
+
+	con 	= readl(host->base + S3C2410_SDICON);
+	pre 	= readl(host->base + S3C2410_SDIPRE);
+	cmdarg 	= readl(host->base + S3C2410_SDICMDARG);
+	cmdcon 	= readl(host->base + S3C2410_SDICMDCON);
+	cmdsta 	= readl(host->base + S3C2410_SDICMDSTAT);
+	r0 	= readl(host->base + S3C2410_SDIRSP0);
+	r1 	= readl(host->base + S3C2410_SDIRSP1);
+	r2 	= readl(host->base + S3C2410_SDIRSP2);
+	r3 	= readl(host->base + S3C2410_SDIRSP3);
+	timer 	= readl(host->base + S3C2410_SDITIMER);
+	datcon 	= readl(host->base + S3C2410_SDIDCON);
+	datcnt 	= readl(host->base + S3C2410_SDIDCNT);
+	datsta 	= readl(host->base + S3C2410_SDIDSTA);
+	fsta 	= readl(host->base + S3C2410_SDIFSTA);
+
+	dbg(host, dbg_debug, "%s  CON:[%08x]  PRE:[%08x]  TMR:[%08x]\n",
+				prefix, con, pre, timer);
+
+	dbg(host, dbg_debug, "%s CCON:[%08x] CARG:[%08x] CSTA:[%08x]\n",
+				prefix, cmdcon, cmdarg, cmdsta);
+
+	dbg(host, dbg_debug, "%s DCON:[%08x] FSTA:[%08x]"
+			       " DSTA:[%08x] DCNT:[%08x]\n",
+				prefix, datcon, fsta, datsta, datcnt);
+
+	dbg(host, dbg_debug, "%s   R0:[%08x]   R1:[%08x]"
+			       "   R2:[%08x]   R3:[%08x]\n",
+				prefix, r0, r1, r2, r3);
+}
+
+static void prepare_dbgmsg(struct s3cmci_host *host, struct mmc_command *cmd,
+			   int stop)
+{
+	snprintf(host->dbgmsg_cmd, 300,
+		 "#%u%s op:%i arg:0x%08x flags:0x08%x retries:%u",
+		 host->ccnt, (stop ? " (STOP)" : ""),
+		 cmd->opcode, cmd->arg, cmd->flags, cmd->retries);
+
+	if (cmd->data) {
+		snprintf(host->dbgmsg_dat, 300,
+			 "#%u bsize:%u blocks:%u bytes:%u",
+			 host->dcnt, cmd->data->blksz,
+			 cmd->data->blocks,
+			 cmd->data->blocks * cmd->data->blksz);
+	} else {
+		host->dbgmsg_dat[0] = '\0';
+	}
+}
+
+static void dbg_dumpcmd(struct s3cmci_host *host, struct mmc_command *cmd,
+			int fail)
+{
+	unsigned int dbglvl = fail ? dbg_fail : dbg_debug;
+
+	if (!cmd)
+		return;
+
+	if (cmd->error == 0) {
+		dbg(host, dbglvl, "CMD[OK] %s R0:0x%08x\n",
+			host->dbgmsg_cmd, cmd->resp[0]);
+	} else {
+		dbg(host, dbglvl, "CMD[ERR %i] %s Status:%s\n",
+			cmd->error, host->dbgmsg_cmd, host->status);
+	}
+
+	if (!cmd->data)
+		return;
+
+	if (cmd->data->error == 0) {
+		dbg(host, dbglvl, "DAT[OK] %s\n", host->dbgmsg_dat);
+	} else {
+		dbg(host, dbglvl, "DAT[ERR %i] %s DCNT:0x%08x\n",
+			cmd->data->error, host->dbgmsg_dat,
+			readl(host->base + S3C2410_SDIDCNT));
+	}
+}
+#else
+static void dbg_dumpcmd(struct s3cmci_host *host,
+			struct mmc_command *cmd, int fail) { }
+
+static void prepare_dbgmsg(struct s3cmci_host *host, struct mmc_command *cmd,
+			   int stop) { }
+
+static void dbg_dumpregs(struct s3cmci_host *host, char *prefix) { }
+
+#endif /* CONFIG_MMC_DEBUG */
+
+/**
+ * s3cmci_host_usedma - return whether the host is using dma or pio
+ * @host: The host state
+ *
+ * Return true if the host is using DMA to transfer data, else false
+ * to use PIO mode. Will return static data depending on the driver
+ * configuration.
+ */
+static inline bool s3cmci_host_usedma(struct s3cmci_host *host)
+{
+#ifdef CONFIG_MMC_S3C_PIO
+	return false;
+#else /* CONFIG_MMC_S3C_DMA */
+	return true;
+#endif
+}
+
+static inline u32 enable_imask(struct s3cmci_host *host, u32 imask)
+{
+	u32 newmask;
+
+	newmask = readl(host->base + host->sdiimsk);
+	newmask |= imask;
+
+	writel(newmask, host->base + host->sdiimsk);
+
+	return newmask;
+}
+
+static inline u32 disable_imask(struct s3cmci_host *host, u32 imask)
+{
+	u32 newmask;
+
+	newmask = readl(host->base + host->sdiimsk);
+	newmask &= ~imask;
+
+	writel(newmask, host->base + host->sdiimsk);
+
+	return newmask;
+}
+
+static inline void clear_imask(struct s3cmci_host *host)
+{
+	u32 mask = readl(host->base + host->sdiimsk);
+
+	/* preserve the SDIO IRQ mask state */
+	mask &= S3C2410_SDIIMSK_SDIOIRQ;
+	writel(mask, host->base + host->sdiimsk);
+}
+
+/**
+ * s3cmci_check_sdio_irq - test whether the SDIO IRQ is being signalled
+ * @host: The host to check.
+ *
+ * Test to see if the SDIO interrupt is being signalled in case the
+ * controller has failed to re-detect a card interrupt. Read GPE8 and
+ * see if it is low and if so, signal a SDIO interrupt.
+ *
+ * This is currently called if a request is finished (we assume that the
+ * bus is now idle) and when the SDIO IRQ is enabled in case the IRQ is
+ * already being indicated.
+*/
+static void s3cmci_check_sdio_irq(struct s3cmci_host *host)
+{
+	if (host->sdio_irqen) {
+		if (host->pdata->bus[3] &&
+		    gpiod_get_value(host->pdata->bus[3]) == 0) {
+			pr_debug("%s: signalling irq\n", __func__);
+			mmc_signal_sdio_irq(host->mmc);
+		}
+	}
+}
+
+static inline int get_data_buffer(struct s3cmci_host *host,
+				  u32 *bytes, u32 **pointer)
+{
+	struct scatterlist *sg;
+
+	if (host->pio_active == XFER_NONE)
+		return -EINVAL;
+
+	if ((!host->mrq) || (!host->mrq->data))
+		return -EINVAL;
+
+	if (host->pio_sgptr >= host->mrq->data->sg_len) {
+		dbg(host, dbg_debug, "no more buffers (%i/%i)\n",
+		      host->pio_sgptr, host->mrq->data->sg_len);
+		return -EBUSY;
+	}
+	sg = &host->mrq->data->sg[host->pio_sgptr];
+
+	*bytes = sg->length;
+	*pointer = sg_virt(sg);
+
+	host->pio_sgptr++;
+
+	dbg(host, dbg_sg, "new buffer (%i/%i)\n",
+	    host->pio_sgptr, host->mrq->data->sg_len);
+
+	return 0;
+}
+
+static inline u32 fifo_count(struct s3cmci_host *host)
+{
+	u32 fifostat = readl(host->base + S3C2410_SDIFSTA);
+
+	fifostat &= S3C2410_SDIFSTA_COUNTMASK;
+	return fifostat;
+}
+
+static inline u32 fifo_free(struct s3cmci_host *host)
+{
+	u32 fifostat = readl(host->base + S3C2410_SDIFSTA);
+
+	fifostat &= S3C2410_SDIFSTA_COUNTMASK;
+	return 63 - fifostat;
+}
+
+/**
+ * s3cmci_enable_irq - enable IRQ, after having disabled it.
+ * @host: The device state.
+ * @more: True if more IRQs are expected from transfer.
+ *
+ * Enable the main IRQ if needed after it has been disabled.
+ *
+ * The IRQ can be one of the following states:
+ *	- disabled during IDLE
+ *	- disabled whilst processing data
+ *	- enabled during transfer
+ *	- enabled whilst awaiting SDIO interrupt detection
+ */
+static void s3cmci_enable_irq(struct s3cmci_host *host, bool more)
+{
+	unsigned long flags;
+	bool enable = false;
+
+	local_irq_save(flags);
+
+	host->irq_enabled = more;
+	host->irq_disabled = false;
+
+	enable = more | host->sdio_irqen;
+
+	if (host->irq_state != enable) {
+		host->irq_state = enable;
+
+		if (enable)
+			enable_irq(host->irq);
+		else
+			disable_irq(host->irq);
+	}
+
+	local_irq_restore(flags);
+}
+
+static void s3cmci_disable_irq(struct s3cmci_host *host, bool transfer)
+{
+	unsigned long flags;
+
+	local_irq_save(flags);
+
+	/* pr_debug("%s: transfer %d\n", __func__, transfer); */
+
+	host->irq_disabled = transfer;
+
+	if (transfer && host->irq_state) {
+		host->irq_state = false;
+		disable_irq(host->irq);
+	}
+
+	local_irq_restore(flags);
+}
+
+static void do_pio_read(struct s3cmci_host *host)
+{
+	int res;
+	u32 fifo;
+	u32 *ptr;
+	u32 fifo_words;
+	void __iomem *from_ptr;
+
+	/* write real prescaler to host, it might be set slow to fix */
+	writel(host->prescaler, host->base + S3C2410_SDIPRE);
+
+	from_ptr = host->base + host->sdidata;
+
+	while ((fifo = fifo_count(host))) {
+		if (!host->pio_bytes) {
+			res = get_data_buffer(host, &host->pio_bytes,
+					      &host->pio_ptr);
+			if (res) {
+				host->pio_active = XFER_NONE;
+				host->complete_what = COMPLETION_FINALIZE;
+
+				dbg(host, dbg_pio, "pio_read(): "
+				    "complete (no more data).\n");
+				return;
+			}
+
+			dbg(host, dbg_pio,
+			    "pio_read(): new target: [%i]@[%p]\n",
+			    host->pio_bytes, host->pio_ptr);
+		}
+
+		dbg(host, dbg_pio,
+		    "pio_read(): fifo:[%02i] buffer:[%03i] dcnt:[%08X]\n",
+		    fifo, host->pio_bytes,
+		    readl(host->base + S3C2410_SDIDCNT));
+
+		/* If we have reached the end of the block, we can
+		 * read a word and get 1 to 3 bytes.  If we in the
+		 * middle of the block, we have to read full words,
+		 * otherwise we will write garbage, so round down to
+		 * an even multiple of 4. */
+		if (fifo >= host->pio_bytes)
+			fifo = host->pio_bytes;
+		else
+			fifo -= fifo & 3;
+
+		host->pio_bytes -= fifo;
+		host->pio_count += fifo;
+
+		fifo_words = fifo >> 2;
+		ptr = host->pio_ptr;
+		while (fifo_words--)
+			*ptr++ = readl(from_ptr);
+		host->pio_ptr = ptr;
+
+		if (fifo & 3) {
+			u32 n = fifo & 3;
+			u32 data = readl(from_ptr);
+			u8 *p = (u8 *)host->pio_ptr;
+
+			while (n--) {
+				*p++ = data;
+				data >>= 8;
+			}
+		}
+	}
+
+	if (!host->pio_bytes) {
+		res = get_data_buffer(host, &host->pio_bytes, &host->pio_ptr);
+		if (res) {
+			dbg(host, dbg_pio,
+			    "pio_read(): complete (no more buffers).\n");
+			host->pio_active = XFER_NONE;
+			host->complete_what = COMPLETION_FINALIZE;
+
+			return;
+		}
+	}
+
+	enable_imask(host,
+		     S3C2410_SDIIMSK_RXFIFOHALF | S3C2410_SDIIMSK_RXFIFOLAST);
+}
+
+static void do_pio_write(struct s3cmci_host *host)
+{
+	void __iomem *to_ptr;
+	int res;
+	u32 fifo;
+	u32 *ptr;
+
+	to_ptr = host->base + host->sdidata;
+
+	while ((fifo = fifo_free(host)) > 3) {
+		if (!host->pio_bytes) {
+			res = get_data_buffer(host, &host->pio_bytes,
+							&host->pio_ptr);
+			if (res) {
+				dbg(host, dbg_pio,
+				    "pio_write(): complete (no more data).\n");
+				host->pio_active = XFER_NONE;
+
+				return;
+			}
+
+			dbg(host, dbg_pio,
+			    "pio_write(): new source: [%i]@[%p]\n",
+			    host->pio_bytes, host->pio_ptr);
+
+		}
+
+		/* If we have reached the end of the block, we have to
+		 * write exactly the remaining number of bytes.  If we
+		 * in the middle of the block, we have to write full
+		 * words, so round down to an even multiple of 4. */
+		if (fifo >= host->pio_bytes)
+			fifo = host->pio_bytes;
+		else
+			fifo -= fifo & 3;
+
+		host->pio_bytes -= fifo;
+		host->pio_count += fifo;
+
+		fifo = (fifo + 3) >> 2;
+		ptr = host->pio_ptr;
+		while (fifo--)
+			writel(*ptr++, to_ptr);
+		host->pio_ptr = ptr;
+	}
+
+	enable_imask(host, S3C2410_SDIIMSK_TXFIFOHALF);
+}
+
+static void pio_tasklet(unsigned long data)
+{
+	struct s3cmci_host *host = (struct s3cmci_host *) data;
+
+	s3cmci_disable_irq(host, true);
+
+	if (host->pio_active == XFER_WRITE)
+		do_pio_write(host);
+
+	if (host->pio_active == XFER_READ)
+		do_pio_read(host);
+
+	if (host->complete_what == COMPLETION_FINALIZE) {
+		clear_imask(host);
+		if (host->pio_active != XFER_NONE) {
+			dbg(host, dbg_err, "unfinished %s "
+			    "- pio_count:[%u] pio_bytes:[%u]\n",
+			    (host->pio_active == XFER_READ) ? "read" : "write",
+			    host->pio_count, host->pio_bytes);
+
+			if (host->mrq->data)
+				host->mrq->data->error = -EINVAL;
+		}
+
+		s3cmci_enable_irq(host, false);
+		finalize_request(host);
+	} else
+		s3cmci_enable_irq(host, true);
+}
+
+/*
+ * ISR for SDI Interface IRQ
+ * Communication between driver and ISR works as follows:
+ *   host->mrq 			points to current request
+ *   host->complete_what	Indicates when the request is considered done
+ *     COMPLETION_CMDSENT	  when the command was sent
+ *     COMPLETION_RSPFIN          when a response was received
+ *     COMPLETION_XFERFINISH	  when the data transfer is finished
+ *     COMPLETION_XFERFINISH_RSPFIN both of the above.
+ *   host->complete_request	is the completion-object the driver waits for
+ *
+ * 1) Driver sets up host->mrq and host->complete_what
+ * 2) Driver prepares the transfer
+ * 3) Driver enables interrupts
+ * 4) Driver starts transfer
+ * 5) Driver waits for host->complete_rquest
+ * 6) ISR checks for request status (errors and success)
+ * 6) ISR sets host->mrq->cmd->error and host->mrq->data->error
+ * 7) ISR completes host->complete_request
+ * 8) ISR disables interrupts
+ * 9) Driver wakes up and takes care of the request
+ *
+ * Note: "->error"-fields are expected to be set to 0 before the request
+ *       was issued by mmc.c - therefore they are only set, when an error
+ *       contition comes up
+ */
+
+static irqreturn_t s3cmci_irq(int irq, void *dev_id)
+{
+	struct s3cmci_host *host = dev_id;
+	struct mmc_command *cmd;
+	u32 mci_csta, mci_dsta, mci_fsta, mci_dcnt, mci_imsk;
+	u32 mci_cclear = 0, mci_dclear;
+	unsigned long iflags;
+
+	mci_dsta = readl(host->base + S3C2410_SDIDSTA);
+	mci_imsk = readl(host->base + host->sdiimsk);
+
+	if (mci_dsta & S3C2410_SDIDSTA_SDIOIRQDETECT) {
+		if (mci_imsk & S3C2410_SDIIMSK_SDIOIRQ) {
+			mci_dclear = S3C2410_SDIDSTA_SDIOIRQDETECT;
+			writel(mci_dclear, host->base + S3C2410_SDIDSTA);
+
+			mmc_signal_sdio_irq(host->mmc);
+			return IRQ_HANDLED;
+		}
+	}
+
+	spin_lock_irqsave(&host->complete_lock, iflags);
+
+	mci_csta = readl(host->base + S3C2410_SDICMDSTAT);
+	mci_dcnt = readl(host->base + S3C2410_SDIDCNT);
+	mci_fsta = readl(host->base + S3C2410_SDIFSTA);
+	mci_dclear = 0;
+
+	if ((host->complete_what == COMPLETION_NONE) ||
+	    (host->complete_what == COMPLETION_FINALIZE)) {
+		host->status = "nothing to complete";
+		clear_imask(host);
+		goto irq_out;
+	}
+
+	if (!host->mrq) {
+		host->status = "no active mrq";
+		clear_imask(host);
+		goto irq_out;
+	}
+
+	cmd = host->cmd_is_stop ? host->mrq->stop : host->mrq->cmd;
+
+	if (!cmd) {
+		host->status = "no active cmd";
+		clear_imask(host);
+		goto irq_out;
+	}
+
+	if (!s3cmci_host_usedma(host)) {
+		if ((host->pio_active == XFER_WRITE) &&
+		    (mci_fsta & S3C2410_SDIFSTA_TFDET)) {
+
+			disable_imask(host, S3C2410_SDIIMSK_TXFIFOHALF);
+			tasklet_schedule(&host->pio_tasklet);
+			host->status = "pio tx";
+		}
+
+		if ((host->pio_active == XFER_READ) &&
+		    (mci_fsta & S3C2410_SDIFSTA_RFDET)) {
+
+			disable_imask(host,
+				      S3C2410_SDIIMSK_RXFIFOHALF |
+				      S3C2410_SDIIMSK_RXFIFOLAST);
+
+			tasklet_schedule(&host->pio_tasklet);
+			host->status = "pio rx";
+		}
+	}
+
+	if (mci_csta & S3C2410_SDICMDSTAT_CMDTIMEOUT) {
+		dbg(host, dbg_err, "CMDSTAT: error CMDTIMEOUT\n");
+		cmd->error = -ETIMEDOUT;
+		host->status = "error: command timeout";
+		goto fail_transfer;
+	}
+
+	if (mci_csta & S3C2410_SDICMDSTAT_CMDSENT) {
+		if (host->complete_what == COMPLETION_CMDSENT) {
+			host->status = "ok: command sent";
+			goto close_transfer;
+		}
+
+		mci_cclear |= S3C2410_SDICMDSTAT_CMDSENT;
+	}
+
+	if (mci_csta & S3C2410_SDICMDSTAT_CRCFAIL) {
+		if (cmd->flags & MMC_RSP_CRC) {
+			if (host->mrq->cmd->flags & MMC_RSP_136) {
+				dbg(host, dbg_irq,
+				    "fixup: ignore CRC fail with long rsp\n");
+			} else {
+				/* note, we used to fail the transfer
+				 * here, but it seems that this is just
+				 * the hardware getting it wrong.
+				 *
+				 * cmd->error = -EILSEQ;
+				 * host->status = "error: bad command crc";
+				 * goto fail_transfer;
+				*/
+			}
+		}
+
+		mci_cclear |= S3C2410_SDICMDSTAT_CRCFAIL;
+	}
+
+	if (mci_csta & S3C2410_SDICMDSTAT_RSPFIN) {
+		if (host->complete_what == COMPLETION_RSPFIN) {
+			host->status = "ok: command response received";
+			goto close_transfer;
+		}
+
+		if (host->complete_what == COMPLETION_XFERFINISH_RSPFIN)
+			host->complete_what = COMPLETION_XFERFINISH;
+
+		mci_cclear |= S3C2410_SDICMDSTAT_RSPFIN;
+	}
+
+	/* errors handled after this point are only relevant
+	   when a data transfer is in progress */
+
+	if (!cmd->data)
+		goto clear_status_bits;
+
+	/* Check for FIFO failure */
+	if (host->is2440) {
+		if (mci_fsta & S3C2440_SDIFSTA_FIFOFAIL) {
+			dbg(host, dbg_err, "FIFO failure\n");
+			host->mrq->data->error = -EILSEQ;
+			host->status = "error: 2440 fifo failure";
+			goto fail_transfer;
+		}
+	} else {
+		if (mci_dsta & S3C2410_SDIDSTA_FIFOFAIL) {
+			dbg(host, dbg_err, "FIFO failure\n");
+			cmd->data->error = -EILSEQ;
+			host->status = "error:  fifo failure";
+			goto fail_transfer;
+		}
+	}
+
+	if (mci_dsta & S3C2410_SDIDSTA_RXCRCFAIL) {
+		dbg(host, dbg_err, "bad data crc (outgoing)\n");
+		cmd->data->error = -EILSEQ;
+		host->status = "error: bad data crc (outgoing)";
+		goto fail_transfer;
+	}
+
+	if (mci_dsta & S3C2410_SDIDSTA_CRCFAIL) {
+		dbg(host, dbg_err, "bad data crc (incoming)\n");
+		cmd->data->error = -EILSEQ;
+		host->status = "error: bad data crc (incoming)";
+		goto fail_transfer;
+	}
+
+	if (mci_dsta & S3C2410_SDIDSTA_DATATIMEOUT) {
+		dbg(host, dbg_err, "data timeout\n");
+		cmd->data->error = -ETIMEDOUT;
+		host->status = "error: data timeout";
+		goto fail_transfer;
+	}
+
+	if (mci_dsta & S3C2410_SDIDSTA_XFERFINISH) {
+		if (host->complete_what == COMPLETION_XFERFINISH) {
+			host->status = "ok: data transfer completed";
+			goto close_transfer;
+		}
+
+		if (host->complete_what == COMPLETION_XFERFINISH_RSPFIN)
+			host->complete_what = COMPLETION_RSPFIN;
+
+		mci_dclear |= S3C2410_SDIDSTA_XFERFINISH;
+	}
+
+clear_status_bits:
+	writel(mci_cclear, host->base + S3C2410_SDICMDSTAT);
+	writel(mci_dclear, host->base + S3C2410_SDIDSTA);
+
+	goto irq_out;
+
+fail_transfer:
+	host->pio_active = XFER_NONE;
+
+close_transfer:
+	host->complete_what = COMPLETION_FINALIZE;
+
+	clear_imask(host);
+	tasklet_schedule(&host->pio_tasklet);
+
+	goto irq_out;
+
+irq_out:
+	dbg(host, dbg_irq,
+	    "csta:0x%08x dsta:0x%08x fsta:0x%08x dcnt:0x%08x status:%s.\n",
+	    mci_csta, mci_dsta, mci_fsta, mci_dcnt, host->status);
+
+	spin_unlock_irqrestore(&host->complete_lock, iflags);
+	return IRQ_HANDLED;
+
+}
+
+static void s3cmci_dma_done_callback(void *arg)
+{
+	struct s3cmci_host *host = arg;
+	unsigned long iflags;
+
+	BUG_ON(!host->mrq);
+	BUG_ON(!host->mrq->data);
+
+	spin_lock_irqsave(&host->complete_lock, iflags);
+
+	dbg(host, dbg_dma, "DMA FINISHED\n");
+
+	host->dma_complete = 1;
+	host->complete_what = COMPLETION_FINALIZE;
+
+	tasklet_schedule(&host->pio_tasklet);
+	spin_unlock_irqrestore(&host->complete_lock, iflags);
+
+}
+
+static void finalize_request(struct s3cmci_host *host)
+{
+	struct mmc_request *mrq = host->mrq;
+	struct mmc_command *cmd;
+	int debug_as_failure = 0;
+
+	if (host->complete_what != COMPLETION_FINALIZE)
+		return;
+
+	if (!mrq)
+		return;
+	cmd = host->cmd_is_stop ? mrq->stop : mrq->cmd;
+
+	if (cmd->data && (cmd->error == 0) &&
+	    (cmd->data->error == 0)) {
+		if (s3cmci_host_usedma(host) && (!host->dma_complete)) {
+			dbg(host, dbg_dma, "DMA Missing (%d)!\n",
+			    host->dma_complete);
+			return;
+		}
+	}
+
+	/* Read response from controller. */
+	cmd->resp[0] = readl(host->base + S3C2410_SDIRSP0);
+	cmd->resp[1] = readl(host->base + S3C2410_SDIRSP1);
+	cmd->resp[2] = readl(host->base + S3C2410_SDIRSP2);
+	cmd->resp[3] = readl(host->base + S3C2410_SDIRSP3);
+
+	writel(host->prescaler, host->base + S3C2410_SDIPRE);
+
+	if (cmd->error)
+		debug_as_failure = 1;
+
+	if (cmd->data && cmd->data->error)
+		debug_as_failure = 1;
+
+	dbg_dumpcmd(host, cmd, debug_as_failure);
+
+	/* Cleanup controller */
+	writel(0, host->base + S3C2410_SDICMDARG);
+	writel(S3C2410_SDIDCON_STOP, host->base + S3C2410_SDIDCON);
+	writel(0, host->base + S3C2410_SDICMDCON);
+	clear_imask(host);
+
+	if (cmd->data && cmd->error)
+		cmd->data->error = cmd->error;
+
+	if (cmd->data && cmd->data->stop && (!host->cmd_is_stop)) {
+		host->cmd_is_stop = 1;
+		s3cmci_send_request(host->mmc);
+		return;
+	}
+
+	/* If we have no data transfer we are finished here */
+	if (!mrq->data)
+		goto request_done;
+
+	/* Calculate the amout of bytes transfer if there was no error */
+	if (mrq->data->error == 0) {
+		mrq->data->bytes_xfered =
+			(mrq->data->blocks * mrq->data->blksz);
+	} else {
+		mrq->data->bytes_xfered = 0;
+	}
+
+	/* If we had an error while transferring data we flush the
+	 * DMA channel and the fifo to clear out any garbage. */
+	if (mrq->data->error != 0) {
+		if (s3cmci_host_usedma(host))
+			dmaengine_terminate_all(host->dma);
+
+		if (host->is2440) {
+			/* Clear failure register and reset fifo. */
+			writel(S3C2440_SDIFSTA_FIFORESET |
+			       S3C2440_SDIFSTA_FIFOFAIL,
+			       host->base + S3C2410_SDIFSTA);
+		} else {
+			u32 mci_con;
+
+			/* reset fifo */
+			mci_con = readl(host->base + S3C2410_SDICON);
+			mci_con |= S3C2410_SDICON_FIFORESET;
+
+			writel(mci_con, host->base + S3C2410_SDICON);
+		}
+	}
+
+request_done:
+	host->complete_what = COMPLETION_NONE;
+	host->mrq = NULL;
+
+	s3cmci_check_sdio_irq(host);
+	mmc_request_done(host->mmc, mrq);
+}
+
+static void s3cmci_send_command(struct s3cmci_host *host,
+					struct mmc_command *cmd)
+{
+	u32 ccon, imsk;
+
+	imsk  = S3C2410_SDIIMSK_CRCSTATUS | S3C2410_SDIIMSK_CMDTIMEOUT |
+		S3C2410_SDIIMSK_RESPONSEND | S3C2410_SDIIMSK_CMDSENT |
+		S3C2410_SDIIMSK_RESPONSECRC;
+
+	enable_imask(host, imsk);
+
+	if (cmd->data)
+		host->complete_what = COMPLETION_XFERFINISH_RSPFIN;
+	else if (cmd->flags & MMC_RSP_PRESENT)
+		host->complete_what = COMPLETION_RSPFIN;
+	else
+		host->complete_what = COMPLETION_CMDSENT;
+
+	writel(cmd->arg, host->base + S3C2410_SDICMDARG);
+
+	ccon  = cmd->opcode & S3C2410_SDICMDCON_INDEX;
+	ccon |= S3C2410_SDICMDCON_SENDERHOST | S3C2410_SDICMDCON_CMDSTART;
+
+	if (cmd->flags & MMC_RSP_PRESENT)
+		ccon |= S3C2410_SDICMDCON_WAITRSP;
+
+	if (cmd->flags & MMC_RSP_136)
+		ccon |= S3C2410_SDICMDCON_LONGRSP;
+
+	writel(ccon, host->base + S3C2410_SDICMDCON);
+}
+
+static int s3cmci_setup_data(struct s3cmci_host *host, struct mmc_data *data)
+{
+	u32 dcon, imsk, stoptries = 3;
+
+	if ((data->blksz & 3) != 0) {
+		/* We cannot deal with unaligned blocks with more than
+		 * one block being transferred. */
+
+		if (data->blocks > 1) {
+			pr_warn("%s: can't do non-word sized block transfers (blksz %d)\n",
+				__func__, data->blksz);
+			return -EINVAL;
+		}
+	}
+
+	while (readl(host->base + S3C2410_SDIDSTA) &
+	       (S3C2410_SDIDSTA_TXDATAON | S3C2410_SDIDSTA_RXDATAON)) {
+
+		dbg(host, dbg_err,
+		    "mci_setup_data() transfer stillin progress.\n");
+
+		writel(S3C2410_SDIDCON_STOP, host->base + S3C2410_SDIDCON);
+		s3cmci_reset(host);
+
+		if ((stoptries--) == 0) {
+			dbg_dumpregs(host, "DRF");
+			return -EINVAL;
+		}
+	}
+
+	dcon  = data->blocks & S3C2410_SDIDCON_BLKNUM_MASK;
+
+	if (s3cmci_host_usedma(host))
+		dcon |= S3C2410_SDIDCON_DMAEN;
+
+	if (host->bus_width == MMC_BUS_WIDTH_4)
+		dcon |= S3C2410_SDIDCON_WIDEBUS;
+
+	dcon |= S3C2410_SDIDCON_BLOCKMODE;
+
+	if (data->flags & MMC_DATA_WRITE) {
+		dcon |= S3C2410_SDIDCON_TXAFTERRESP;
+		dcon |= S3C2410_SDIDCON_XFER_TXSTART;
+	}
+
+	if (data->flags & MMC_DATA_READ) {
+		dcon |= S3C2410_SDIDCON_RXAFTERCMD;
+		dcon |= S3C2410_SDIDCON_XFER_RXSTART;
+	}
+
+	if (host->is2440) {
+		dcon |= S3C2440_SDIDCON_DS_WORD;
+		dcon |= S3C2440_SDIDCON_DATSTART;
+	}
+
+	writel(dcon, host->base + S3C2410_SDIDCON);
+
+	/* write BSIZE register */
+
+	writel(data->blksz, host->base + S3C2410_SDIBSIZE);
+
+	/* add to IMASK register */
+	imsk = S3C2410_SDIIMSK_FIFOFAIL | S3C2410_SDIIMSK_DATACRC |
+	       S3C2410_SDIIMSK_DATATIMEOUT | S3C2410_SDIIMSK_DATAFINISH;
+
+	enable_imask(host, imsk);
+
+	/* write TIMER register */
+
+	if (host->is2440) {
+		writel(0x007FFFFF, host->base + S3C2410_SDITIMER);
+	} else {
+		writel(0x0000FFFF, host->base + S3C2410_SDITIMER);
+
+		/* FIX: set slow clock to prevent timeouts on read */
+		if (data->flags & MMC_DATA_READ)
+			writel(0xFF, host->base + S3C2410_SDIPRE);
+	}
+
+	return 0;
+}
+
+#define BOTH_DIR (MMC_DATA_WRITE | MMC_DATA_READ)
+
+static int s3cmci_prepare_pio(struct s3cmci_host *host, struct mmc_data *data)
+{
+	int rw = (data->flags & MMC_DATA_WRITE) ? 1 : 0;
+
+	BUG_ON((data->flags & BOTH_DIR) == BOTH_DIR);
+
+	host->pio_sgptr = 0;
+	host->pio_bytes = 0;
+	host->pio_count = 0;
+	host->pio_active = rw ? XFER_WRITE : XFER_READ;
+
+	if (rw) {
+		do_pio_write(host);
+		enable_imask(host, S3C2410_SDIIMSK_TXFIFOHALF);
+	} else {
+		enable_imask(host, S3C2410_SDIIMSK_RXFIFOHALF
+			     | S3C2410_SDIIMSK_RXFIFOLAST);
+	}
+
+	return 0;
+}
+
+static int s3cmci_prepare_dma(struct s3cmci_host *host, struct mmc_data *data)
+{
+	int rw = data->flags & MMC_DATA_WRITE;
+	struct dma_async_tx_descriptor *desc;
+	struct dma_slave_config conf = {
+		.src_addr = host->mem->start + host->sdidata,
+		.dst_addr = host->mem->start + host->sdidata,
+		.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+		.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+	};
+
+	BUG_ON((data->flags & BOTH_DIR) == BOTH_DIR);
+
+	/* Restore prescaler value */
+	writel(host->prescaler, host->base + S3C2410_SDIPRE);
+
+	if (!rw)
+		conf.direction = DMA_DEV_TO_MEM;
+	else
+		conf.direction = DMA_MEM_TO_DEV;
+
+	dma_map_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+		   mmc_get_dma_dir(data));
+
+	dmaengine_slave_config(host->dma, &conf);
+	desc = dmaengine_prep_slave_sg(host->dma, data->sg, data->sg_len,
+		conf.direction,
+		DMA_CTRL_ACK | DMA_PREP_INTERRUPT);
+	if (!desc)
+		goto unmap_exit;
+	desc->callback = s3cmci_dma_done_callback;
+	desc->callback_param = host;
+	dmaengine_submit(desc);
+	dma_async_issue_pending(host->dma);
+
+	return 0;
+
+unmap_exit:
+	dma_unmap_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+		     mmc_get_dma_dir(data));
+	return -ENOMEM;
+}
+
+static void s3cmci_send_request(struct mmc_host *mmc)
+{
+	struct s3cmci_host *host = mmc_priv(mmc);
+	struct mmc_request *mrq = host->mrq;
+	struct mmc_command *cmd = host->cmd_is_stop ? mrq->stop : mrq->cmd;
+
+	host->ccnt++;
+	prepare_dbgmsg(host, cmd, host->cmd_is_stop);
+
+	/* Clear command, data and fifo status registers
+	   Fifo clear only necessary on 2440, but doesn't hurt on 2410
+	*/
+	writel(0xFFFFFFFF, host->base + S3C2410_SDICMDSTAT);
+	writel(0xFFFFFFFF, host->base + S3C2410_SDIDSTA);
+	writel(0xFFFFFFFF, host->base + S3C2410_SDIFSTA);
+
+	if (cmd->data) {
+		int res = s3cmci_setup_data(host, cmd->data);
+
+		host->dcnt++;
+
+		if (res) {
+			dbg(host, dbg_err, "setup data error %d\n", res);
+			cmd->error = res;
+			cmd->data->error = res;
+
+			mmc_request_done(mmc, mrq);
+			return;
+		}
+
+		if (s3cmci_host_usedma(host))
+			res = s3cmci_prepare_dma(host, cmd->data);
+		else
+			res = s3cmci_prepare_pio(host, cmd->data);
+
+		if (res) {
+			dbg(host, dbg_err, "data prepare error %d\n", res);
+			cmd->error = res;
+			cmd->data->error = res;
+
+			mmc_request_done(mmc, mrq);
+			return;
+		}
+	}
+
+	/* Send command */
+	s3cmci_send_command(host, cmd);
+
+	/* Enable Interrupt */
+	s3cmci_enable_irq(host, true);
+}
+
+static void s3cmci_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct s3cmci_host *host = mmc_priv(mmc);
+
+	host->status = "mmc request";
+	host->cmd_is_stop = 0;
+	host->mrq = mrq;
+
+	if (mmc_gpio_get_cd(mmc) == 0) {
+		dbg(host, dbg_err, "%s: no medium present\n", __func__);
+		host->mrq->cmd->error = -ENOMEDIUM;
+		mmc_request_done(mmc, mrq);
+	} else
+		s3cmci_send_request(mmc);
+}
+
+static void s3cmci_set_clk(struct s3cmci_host *host, struct mmc_ios *ios)
+{
+	u32 mci_psc;
+
+	/* Set clock */
+	for (mci_psc = 0; mci_psc < 255; mci_psc++) {
+		host->real_rate = host->clk_rate / (host->clk_div*(mci_psc+1));
+
+		if (host->real_rate <= ios->clock)
+			break;
+	}
+
+	if (mci_psc > 255)
+		mci_psc = 255;
+
+	host->prescaler = mci_psc;
+	writel(host->prescaler, host->base + S3C2410_SDIPRE);
+
+	/* If requested clock is 0, real_rate will be 0, too */
+	if (ios->clock == 0)
+		host->real_rate = 0;
+}
+
+static void s3cmci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct s3cmci_host *host = mmc_priv(mmc);
+	u32 mci_con;
+
+	/* Set the power state */
+
+	mci_con = readl(host->base + S3C2410_SDICON);
+
+	switch (ios->power_mode) {
+	case MMC_POWER_ON:
+	case MMC_POWER_UP:
+		if (!host->is2440)
+			mci_con |= S3C2410_SDICON_FIFORESET;
+		break;
+
+	case MMC_POWER_OFF:
+	default:
+		if (host->is2440)
+			mci_con |= S3C2440_SDICON_SDRESET;
+		break;
+	}
+
+	if (host->pdata->set_power)
+		host->pdata->set_power(ios->power_mode, ios->vdd);
+
+	s3cmci_set_clk(host, ios);
+
+	/* Set CLOCK_ENABLE */
+	if (ios->clock)
+		mci_con |= S3C2410_SDICON_CLOCKTYPE;
+	else
+		mci_con &= ~S3C2410_SDICON_CLOCKTYPE;
+
+	writel(mci_con, host->base + S3C2410_SDICON);
+
+	if ((ios->power_mode == MMC_POWER_ON) ||
+	    (ios->power_mode == MMC_POWER_UP)) {
+		dbg(host, dbg_conf, "running at %lukHz (requested: %ukHz).\n",
+			host->real_rate/1000, ios->clock/1000);
+	} else {
+		dbg(host, dbg_conf, "powered down.\n");
+	}
+
+	host->bus_width = ios->bus_width;
+}
+
+static void s3cmci_reset(struct s3cmci_host *host)
+{
+	u32 con = readl(host->base + S3C2410_SDICON);
+
+	con |= S3C2440_SDICON_SDRESET;
+	writel(con, host->base + S3C2410_SDICON);
+}
+
+static void s3cmci_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct s3cmci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+	u32 con;
+
+	local_irq_save(flags);
+
+	con = readl(host->base + S3C2410_SDICON);
+	host->sdio_irqen = enable;
+
+	if (enable == host->sdio_irqen)
+		goto same_state;
+
+	if (enable) {
+		con |= S3C2410_SDICON_SDIOIRQ;
+		enable_imask(host, S3C2410_SDIIMSK_SDIOIRQ);
+
+		if (!host->irq_state && !host->irq_disabled) {
+			host->irq_state = true;
+			enable_irq(host->irq);
+		}
+	} else {
+		disable_imask(host, S3C2410_SDIIMSK_SDIOIRQ);
+		con &= ~S3C2410_SDICON_SDIOIRQ;
+
+		if (!host->irq_enabled && host->irq_state) {
+			disable_irq_nosync(host->irq);
+			host->irq_state = false;
+		}
+	}
+
+	writel(con, host->base + S3C2410_SDICON);
+
+ same_state:
+	local_irq_restore(flags);
+
+	s3cmci_check_sdio_irq(host);
+}
+
+static const struct mmc_host_ops s3cmci_ops = {
+	.request	= s3cmci_request,
+	.set_ios	= s3cmci_set_ios,
+	.get_ro		= mmc_gpio_get_ro,
+	.get_cd		= mmc_gpio_get_cd,
+	.enable_sdio_irq = s3cmci_enable_sdio_irq,
+};
+
+#ifdef CONFIG_ARM_S3C24XX_CPUFREQ
+
+static int s3cmci_cpufreq_transition(struct notifier_block *nb,
+				     unsigned long val, void *data)
+{
+	struct s3cmci_host *host;
+	struct mmc_host *mmc;
+	unsigned long newclk;
+	unsigned long flags;
+
+	host = container_of(nb, struct s3cmci_host, freq_transition);
+	newclk = clk_get_rate(host->clk);
+	mmc = host->mmc;
+
+	if ((val == CPUFREQ_PRECHANGE && newclk > host->clk_rate) ||
+	    (val == CPUFREQ_POSTCHANGE && newclk < host->clk_rate)) {
+		spin_lock_irqsave(&mmc->lock, flags);
+
+		host->clk_rate = newclk;
+
+		if (mmc->ios.power_mode != MMC_POWER_OFF &&
+		    mmc->ios.clock != 0)
+			s3cmci_set_clk(host, &mmc->ios);
+
+		spin_unlock_irqrestore(&mmc->lock, flags);
+	}
+
+	return 0;
+}
+
+static inline int s3cmci_cpufreq_register(struct s3cmci_host *host)
+{
+	host->freq_transition.notifier_call = s3cmci_cpufreq_transition;
+
+	return cpufreq_register_notifier(&host->freq_transition,
+					 CPUFREQ_TRANSITION_NOTIFIER);
+}
+
+static inline void s3cmci_cpufreq_deregister(struct s3cmci_host *host)
+{
+	cpufreq_unregister_notifier(&host->freq_transition,
+				    CPUFREQ_TRANSITION_NOTIFIER);
+}
+
+#else
+static inline int s3cmci_cpufreq_register(struct s3cmci_host *host)
+{
+	return 0;
+}
+
+static inline void s3cmci_cpufreq_deregister(struct s3cmci_host *host)
+{
+}
+#endif
+
+
+#ifdef CONFIG_DEBUG_FS
+
+static int s3cmci_state_show(struct seq_file *seq, void *v)
+{
+	struct s3cmci_host *host = seq->private;
+
+	seq_printf(seq, "Register base = 0x%p\n", host->base);
+	seq_printf(seq, "Clock rate = %ld\n", host->clk_rate);
+	seq_printf(seq, "Prescale = %d\n", host->prescaler);
+	seq_printf(seq, "is2440 = %d\n", host->is2440);
+	seq_printf(seq, "IRQ = %d\n", host->irq);
+	seq_printf(seq, "IRQ enabled = %d\n", host->irq_enabled);
+	seq_printf(seq, "IRQ disabled = %d\n", host->irq_disabled);
+	seq_printf(seq, "IRQ state = %d\n", host->irq_state);
+	seq_printf(seq, "CD IRQ = %d\n", host->irq_cd);
+	seq_printf(seq, "Do DMA = %d\n", s3cmci_host_usedma(host));
+	seq_printf(seq, "SDIIMSK at %d\n", host->sdiimsk);
+	seq_printf(seq, "SDIDATA at %d\n", host->sdidata);
+
+	return 0;
+}
+
+DEFINE_SHOW_ATTRIBUTE(s3cmci_state);
+
+#define DBG_REG(_r) { .addr = S3C2410_SDI##_r, .name = #_r }
+
+struct s3cmci_reg {
+	unsigned short	addr;
+	unsigned char	*name;
+};
+
+static const struct s3cmci_reg debug_regs[] = {
+	DBG_REG(CON),
+	DBG_REG(PRE),
+	DBG_REG(CMDARG),
+	DBG_REG(CMDCON),
+	DBG_REG(CMDSTAT),
+	DBG_REG(RSP0),
+	DBG_REG(RSP1),
+	DBG_REG(RSP2),
+	DBG_REG(RSP3),
+	DBG_REG(TIMER),
+	DBG_REG(BSIZE),
+	DBG_REG(DCON),
+	DBG_REG(DCNT),
+	DBG_REG(DSTA),
+	DBG_REG(FSTA),
+	{}
+};
+
+static int s3cmci_regs_show(struct seq_file *seq, void *v)
+{
+	struct s3cmci_host *host = seq->private;
+	const struct s3cmci_reg *rptr = debug_regs;
+
+	for (; rptr->name; rptr++)
+		seq_printf(seq, "SDI%s\t=0x%08x\n", rptr->name,
+			   readl(host->base + rptr->addr));
+
+	seq_printf(seq, "SDIIMSK\t=0x%08x\n", readl(host->base + host->sdiimsk));
+
+	return 0;
+}
+
+DEFINE_SHOW_ATTRIBUTE(s3cmci_regs);
+
+static void s3cmci_debugfs_attach(struct s3cmci_host *host)
+{
+	struct device *dev = &host->pdev->dev;
+	struct dentry *root;
+
+	root = debugfs_create_dir(dev_name(dev), NULL);
+	host->debug_root = root;
+
+	debugfs_create_file("state", 0444, root, host, &s3cmci_state_fops);
+	debugfs_create_file("regs", 0444, root, host, &s3cmci_regs_fops);
+}
+
+static void s3cmci_debugfs_remove(struct s3cmci_host *host)
+{
+	debugfs_remove_recursive(host->debug_root);
+}
+
+#else
+static inline void s3cmci_debugfs_attach(struct s3cmci_host *host) { }
+static inline void s3cmci_debugfs_remove(struct s3cmci_host *host) { }
+
+#endif /* CONFIG_DEBUG_FS */
+
+static int s3cmci_probe_pdata(struct s3cmci_host *host)
+{
+	struct platform_device *pdev = host->pdev;
+	struct mmc_host *mmc = host->mmc;
+	struct s3c24xx_mci_pdata *pdata;
+	int i, ret;
+
+	host->is2440 = platform_get_device_id(pdev)->driver_data;
+	pdata = pdev->dev.platform_data;
+	if (!pdata) {
+		dev_err(&pdev->dev, "need platform data");
+		return -ENXIO;
+	}
+
+	for (i = 0; i < 6; i++) {
+		pdata->bus[i] = devm_gpiod_get_index(&pdev->dev, "bus", i,
+						     GPIOD_OUT_LOW);
+		if (IS_ERR(pdata->bus[i])) {
+			dev_err(&pdev->dev, "failed to get gpio %d\n", i);
+			return PTR_ERR(pdata->bus[i]);
+		}
+	}
+
+	if (pdata->no_wprotect)
+		mmc->caps2 |= MMC_CAP2_NO_WRITE_PROTECT;
+
+	if (pdata->no_detect)
+		mmc->caps |= MMC_CAP_NEEDS_POLL;
+
+	if (pdata->wprotect_invert)
+		mmc->caps2 |= MMC_CAP2_RO_ACTIVE_HIGH;
+
+	/* If we get -ENOENT we have no card detect GPIO line */
+	ret = mmc_gpiod_request_cd(mmc, "cd", 0, false, 0);
+	if (ret != -ENOENT) {
+		dev_err(&pdev->dev, "error requesting GPIO for CD %d\n",
+			ret);
+		return ret;
+	}
+
+	ret = mmc_gpiod_request_ro(host->mmc, "wp", 0, 0);
+	if (ret != -ENOENT) {
+		dev_err(&pdev->dev, "error requesting GPIO for WP %d\n",
+			ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int s3cmci_probe_dt(struct s3cmci_host *host)
+{
+	struct platform_device *pdev = host->pdev;
+	struct s3c24xx_mci_pdata *pdata;
+	struct mmc_host *mmc = host->mmc;
+	int ret;
+
+	host->is2440 = (long) of_device_get_match_data(&pdev->dev);
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		return ret;
+
+	pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata)
+		return -ENOMEM;
+
+	pdev->dev.platform_data = pdata;
+
+	return 0;
+}
+
+static int s3cmci_probe(struct platform_device *pdev)
+{
+	struct s3cmci_host *host;
+	struct mmc_host	*mmc;
+	int ret;
+
+	mmc = mmc_alloc_host(sizeof(struct s3cmci_host), &pdev->dev);
+	if (!mmc) {
+		ret = -ENOMEM;
+		goto probe_out;
+	}
+
+	host = mmc_priv(mmc);
+	host->mmc 	= mmc;
+	host->pdev	= pdev;
+
+	if (pdev->dev.of_node)
+		ret = s3cmci_probe_dt(host);
+	else
+		ret = s3cmci_probe_pdata(host);
+
+	if (ret)
+		goto probe_free_host;
+
+	host->pdata = pdev->dev.platform_data;
+
+	spin_lock_init(&host->complete_lock);
+	tasklet_init(&host->pio_tasklet, pio_tasklet, (unsigned long) host);
+
+	if (host->is2440) {
+		host->sdiimsk	= S3C2440_SDIIMSK;
+		host->sdidata	= S3C2440_SDIDATA;
+		host->clk_div	= 1;
+	} else {
+		host->sdiimsk	= S3C2410_SDIIMSK;
+		host->sdidata	= S3C2410_SDIDATA;
+		host->clk_div	= 2;
+	}
+
+	host->complete_what 	= COMPLETION_NONE;
+	host->pio_active 	= XFER_NONE;
+
+	host->mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!host->mem) {
+		dev_err(&pdev->dev,
+			"failed to get io memory region resource.\n");
+
+		ret = -ENOENT;
+		goto probe_free_host;
+	}
+
+	host->mem = request_mem_region(host->mem->start,
+				       resource_size(host->mem), pdev->name);
+
+	if (!host->mem) {
+		dev_err(&pdev->dev, "failed to request io memory region.\n");
+		ret = -ENOENT;
+		goto probe_free_host;
+	}
+
+	host->base = ioremap(host->mem->start, resource_size(host->mem));
+	if (!host->base) {
+		dev_err(&pdev->dev, "failed to ioremap() io memory region.\n");
+		ret = -EINVAL;
+		goto probe_free_mem_region;
+	}
+
+	host->irq = platform_get_irq(pdev, 0);
+	if (host->irq <= 0) {
+		ret = -EINVAL;
+		goto probe_iounmap;
+	}
+
+	if (request_irq(host->irq, s3cmci_irq, 0, DRIVER_NAME, host)) {
+		dev_err(&pdev->dev, "failed to request mci interrupt.\n");
+		ret = -ENOENT;
+		goto probe_iounmap;
+	}
+
+	/* We get spurious interrupts even when we have set the IMSK
+	 * register to ignore everything, so use disable_irq() to make
+	 * ensure we don't lock the system with un-serviceable requests. */
+
+	disable_irq(host->irq);
+	host->irq_state = false;
+
+	/* Depending on the dma state, get a DMA channel to use. */
+
+	if (s3cmci_host_usedma(host)) {
+		host->dma = dma_request_chan(&pdev->dev, "rx-tx");
+		ret = PTR_ERR_OR_ZERO(host->dma);
+		if (ret) {
+			dev_err(&pdev->dev, "cannot get DMA channel.\n");
+			goto probe_free_irq;
+		}
+	}
+
+	host->clk = clk_get(&pdev->dev, "sdi");
+	if (IS_ERR(host->clk)) {
+		dev_err(&pdev->dev, "failed to find clock source.\n");
+		ret = PTR_ERR(host->clk);
+		host->clk = NULL;
+		goto probe_free_dma;
+	}
+
+	ret = clk_prepare_enable(host->clk);
+	if (ret) {
+		dev_err(&pdev->dev, "failed to enable clock source.\n");
+		goto clk_free;
+	}
+
+	host->clk_rate = clk_get_rate(host->clk);
+
+	mmc->ops 	= &s3cmci_ops;
+	mmc->ocr_avail	= MMC_VDD_32_33 | MMC_VDD_33_34;
+#ifdef CONFIG_MMC_S3C_HW_SDIO_IRQ
+	mmc->caps	= MMC_CAP_4_BIT_DATA | MMC_CAP_SDIO_IRQ;
+#else
+	mmc->caps	= MMC_CAP_4_BIT_DATA;
+#endif
+	mmc->f_min 	= host->clk_rate / (host->clk_div * 256);
+	mmc->f_max 	= host->clk_rate / host->clk_div;
+
+	if (host->pdata->ocr_avail)
+		mmc->ocr_avail = host->pdata->ocr_avail;
+
+	mmc->max_blk_count	= 4095;
+	mmc->max_blk_size	= 4095;
+	mmc->max_req_size	= 4095 * 512;
+	mmc->max_seg_size	= mmc->max_req_size;
+
+	mmc->max_segs		= 128;
+
+	dbg(host, dbg_debug,
+	    "probe: mode:%s mapped mci_base:%p irq:%u irq_cd:%u dma:%p.\n",
+	    (host->is2440?"2440":""),
+	    host->base, host->irq, host->irq_cd, host->dma);
+
+	ret = s3cmci_cpufreq_register(host);
+	if (ret) {
+		dev_err(&pdev->dev, "failed to register cpufreq\n");
+		goto free_dmabuf;
+	}
+
+	ret = mmc_add_host(mmc);
+	if (ret) {
+		dev_err(&pdev->dev, "failed to add mmc host.\n");
+		goto free_cpufreq;
+	}
+
+	s3cmci_debugfs_attach(host);
+
+	platform_set_drvdata(pdev, mmc);
+	dev_info(&pdev->dev, "%s - using %s, %s SDIO IRQ\n", mmc_hostname(mmc),
+		 s3cmci_host_usedma(host) ? "dma" : "pio",
+		 mmc->caps & MMC_CAP_SDIO_IRQ ? "hw" : "sw");
+
+	return 0;
+
+ free_cpufreq:
+	s3cmci_cpufreq_deregister(host);
+
+ free_dmabuf:
+	clk_disable_unprepare(host->clk);
+
+ clk_free:
+	clk_put(host->clk);
+
+ probe_free_dma:
+	if (s3cmci_host_usedma(host))
+		dma_release_channel(host->dma);
+
+ probe_free_irq:
+	free_irq(host->irq, host);
+
+ probe_iounmap:
+	iounmap(host->base);
+
+ probe_free_mem_region:
+	release_mem_region(host->mem->start, resource_size(host->mem));
+
+ probe_free_host:
+	mmc_free_host(mmc);
+
+ probe_out:
+	return ret;
+}
+
+static void s3cmci_shutdown(struct platform_device *pdev)
+{
+	struct mmc_host	*mmc = platform_get_drvdata(pdev);
+	struct s3cmci_host *host = mmc_priv(mmc);
+
+	if (host->irq_cd >= 0)
+		free_irq(host->irq_cd, host);
+
+	s3cmci_debugfs_remove(host);
+	s3cmci_cpufreq_deregister(host);
+	mmc_remove_host(mmc);
+	clk_disable_unprepare(host->clk);
+}
+
+static int s3cmci_remove(struct platform_device *pdev)
+{
+	struct mmc_host		*mmc  = platform_get_drvdata(pdev);
+	struct s3cmci_host	*host = mmc_priv(mmc);
+
+	s3cmci_shutdown(pdev);
+
+	clk_put(host->clk);
+
+	tasklet_disable(&host->pio_tasklet);
+
+	if (s3cmci_host_usedma(host))
+		dma_release_channel(host->dma);
+
+	free_irq(host->irq, host);
+
+	iounmap(host->base);
+	release_mem_region(host->mem->start, resource_size(host->mem));
+
+	mmc_free_host(mmc);
+	return 0;
+}
+
+static const struct of_device_id s3cmci_dt_match[] = {
+	{
+		.compatible = "samsung,s3c2410-sdi",
+		.data = (void *)0,
+	},
+	{
+		.compatible = "samsung,s3c2412-sdi",
+		.data = (void *)1,
+	},
+	{
+		.compatible = "samsung,s3c2440-sdi",
+		.data = (void *)1,
+	},
+	{ /* sentinel */ },
+};
+MODULE_DEVICE_TABLE(of, s3cmci_dt_match);
+
+static const struct platform_device_id s3cmci_driver_ids[] = {
+	{
+		.name	= "s3c2410-sdi",
+		.driver_data	= 0,
+	}, {
+		.name	= "s3c2412-sdi",
+		.driver_data	= 1,
+	}, {
+		.name	= "s3c2440-sdi",
+		.driver_data	= 1,
+	},
+	{ }
+};
+
+MODULE_DEVICE_TABLE(platform, s3cmci_driver_ids);
+
+static struct platform_driver s3cmci_driver = {
+	.driver	= {
+		.name	= "s3c-sdi",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = s3cmci_dt_match,
+	},
+	.id_table	= s3cmci_driver_ids,
+	.probe		= s3cmci_probe,
+	.remove		= s3cmci_remove,
+	.shutdown	= s3cmci_shutdown,
+};
+
+module_platform_driver(s3cmci_driver);
+
+MODULE_DESCRIPTION("Samsung S3C MMC/SD Card Interface driver");
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Thomas Kleffel <tk@maintech.de>, Ben Dooks <ben-linux@fluff.org>");
diff --git a/drivers/mmc/host/s3cmci.h b/drivers/mmc/host/s3cmci.h
new file mode 100644
index 000000000..8b65d7ad9
--- /dev/null
+++ b/drivers/mmc/host/s3cmci.h
@@ -0,0 +1,75 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ *  linux/drivers/mmc/s3cmci.h - Samsung S3C MCI driver
+ *
+ *  Copyright (C) 2004-2006 Thomas Kleffel, All Rights Reserved.
+ */
+
+enum s3cmci_waitfor {
+	COMPLETION_NONE,
+	COMPLETION_FINALIZE,
+	COMPLETION_CMDSENT,
+	COMPLETION_RSPFIN,
+	COMPLETION_XFERFINISH,
+	COMPLETION_XFERFINISH_RSPFIN,
+};
+
+struct s3cmci_host {
+	struct platform_device	*pdev;
+	struct s3c24xx_mci_pdata *pdata;
+	struct mmc_host		*mmc;
+	struct resource		*mem;
+	struct clk		*clk;
+	void __iomem		*base;
+	int			irq;
+	int			irq_cd;
+	struct dma_chan		*dma;
+
+	unsigned long		clk_rate;
+	unsigned long		clk_div;
+	unsigned long		real_rate;
+	u8			prescaler;
+
+	int			is2440;
+	unsigned		sdiimsk;
+	unsigned		sdidata;
+
+	bool			irq_disabled;
+	bool			irq_enabled;
+	bool			irq_state;
+	int			sdio_irqen;
+
+	struct mmc_request	*mrq;
+	int			cmd_is_stop;
+
+	spinlock_t		complete_lock;
+	enum s3cmci_waitfor	complete_what;
+
+	int			dma_complete;
+
+	u32			pio_sgptr;
+	u32			pio_bytes;
+	u32			pio_count;
+	u32			*pio_ptr;
+#define XFER_NONE 0
+#define XFER_READ 1
+#define XFER_WRITE 2
+	u32			pio_active;
+
+	int			bus_width;
+
+	char 			dbgmsg_cmd[301];
+	char 			dbgmsg_dat[301];
+	char			*status;
+
+	unsigned int		ccnt, dcnt;
+	struct tasklet_struct	pio_tasklet;
+
+#ifdef CONFIG_DEBUG_FS
+	struct dentry		*debug_root;
+#endif
+
+#ifdef CONFIG_ARM_S3C24XX_CPUFREQ
+	struct notifier_block	freq_transition;
+#endif
+};
diff --git a/drivers/mmc/host/sdhci-acpi.c b/drivers/mmc/host/sdhci-acpi.c
new file mode 100644
index 000000000..2a2810177
--- /dev/null
+++ b/drivers/mmc/host/sdhci-acpi.c
@@ -0,0 +1,1090 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Secure Digital Host Controller Interface ACPI driver.
+ *
+ * Copyright (c) 2012, Intel Corporation.
+ */
+
+#include <linux/init.h>
+#include <linux/export.h>
+#include <linux/module.h>
+#include <linux/device.h>
+#include <linux/platform_device.h>
+#include <linux/ioport.h>
+#include <linux/io.h>
+#include <linux/dma-mapping.h>
+#include <linux/compiler.h>
+#include <linux/stddef.h>
+#include <linux/bitops.h>
+#include <linux/types.h>
+#include <linux/err.h>
+#include <linux/interrupt.h>
+#include <linux/acpi.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+#include <linux/delay.h>
+#include <linux/dmi.h>
+
+#include <linux/mmc/host.h>
+#include <linux/mmc/pm.h>
+#include <linux/mmc/slot-gpio.h>
+
+#ifdef CONFIG_X86
+#include <asm/cpu_device_id.h>
+#include <asm/intel-family.h>
+#include <asm/iosf_mbi.h>
+#include <linux/pci.h>
+#endif
+
+#include "sdhci.h"
+
+enum {
+	SDHCI_ACPI_SD_CD		= BIT(0),
+	SDHCI_ACPI_RUNTIME_PM		= BIT(1),
+	SDHCI_ACPI_SD_CD_OVERRIDE_LEVEL	= BIT(2),
+};
+
+struct sdhci_acpi_chip {
+	const struct	sdhci_ops *ops;
+	unsigned int	quirks;
+	unsigned int	quirks2;
+	unsigned long	caps;
+	unsigned int	caps2;
+	mmc_pm_flag_t	pm_caps;
+};
+
+struct sdhci_acpi_slot {
+	const struct	sdhci_acpi_chip *chip;
+	unsigned int	quirks;
+	unsigned int	quirks2;
+	unsigned long	caps;
+	unsigned int	caps2;
+	mmc_pm_flag_t	pm_caps;
+	unsigned int	flags;
+	size_t		priv_size;
+	int (*probe_slot)(struct platform_device *, struct acpi_device *);
+	int (*remove_slot)(struct platform_device *);
+	int (*free_slot)(struct platform_device *pdev);
+	int (*setup_host)(struct platform_device *pdev);
+};
+
+struct sdhci_acpi_host {
+	struct sdhci_host		*host;
+	const struct sdhci_acpi_slot	*slot;
+	struct platform_device		*pdev;
+	bool				use_runtime_pm;
+	bool				is_intel;
+	bool				reset_signal_volt_on_suspend;
+	unsigned long			private[] ____cacheline_aligned;
+};
+
+enum {
+	DMI_QUIRK_RESET_SD_SIGNAL_VOLT_ON_SUSP			= BIT(0),
+	DMI_QUIRK_SD_NO_WRITE_PROTECT				= BIT(1),
+};
+
+static inline void *sdhci_acpi_priv(struct sdhci_acpi_host *c)
+{
+	return (void *)c->private;
+}
+
+static inline bool sdhci_acpi_flag(struct sdhci_acpi_host *c, unsigned int flag)
+{
+	return c->slot && (c->slot->flags & flag);
+}
+
+#define INTEL_DSM_HS_CAPS_SDR25		BIT(0)
+#define INTEL_DSM_HS_CAPS_DDR50		BIT(1)
+#define INTEL_DSM_HS_CAPS_SDR50		BIT(2)
+#define INTEL_DSM_HS_CAPS_SDR104	BIT(3)
+
+enum {
+	INTEL_DSM_FNS		=  0,
+	INTEL_DSM_V18_SWITCH	=  3,
+	INTEL_DSM_V33_SWITCH	=  4,
+	INTEL_DSM_HS_CAPS	=  8,
+};
+
+struct intel_host {
+	u32	dsm_fns;
+	u32	hs_caps;
+};
+
+static const guid_t intel_dsm_guid =
+	GUID_INIT(0xF6C13EA5, 0x65CD, 0x461F,
+		  0xAB, 0x7A, 0x29, 0xF7, 0xE8, 0xD5, 0xBD, 0x61);
+
+static int __intel_dsm(struct intel_host *intel_host, struct device *dev,
+		       unsigned int fn, u32 *result)
+{
+	union acpi_object *obj;
+	int err = 0;
+
+	obj = acpi_evaluate_dsm(ACPI_HANDLE(dev), &intel_dsm_guid, 0, fn, NULL);
+	if (!obj)
+		return -EOPNOTSUPP;
+
+	if (obj->type == ACPI_TYPE_INTEGER) {
+		*result = obj->integer.value;
+	} else if (obj->type == ACPI_TYPE_BUFFER && obj->buffer.length > 0) {
+		size_t len = min_t(size_t, obj->buffer.length, 4);
+
+		*result = 0;
+		memcpy(result, obj->buffer.pointer, len);
+	} else {
+		dev_err(dev, "%s DSM fn %u obj->type %d obj->buffer.length %d\n",
+			__func__, fn, obj->type, obj->buffer.length);
+		err = -EINVAL;
+	}
+
+	ACPI_FREE(obj);
+
+	return err;
+}
+
+static int intel_dsm(struct intel_host *intel_host, struct device *dev,
+		     unsigned int fn, u32 *result)
+{
+	if (fn > 31 || !(intel_host->dsm_fns & (1 << fn)))
+		return -EOPNOTSUPP;
+
+	return __intel_dsm(intel_host, dev, fn, result);
+}
+
+static void intel_dsm_init(struct intel_host *intel_host, struct device *dev,
+			   struct mmc_host *mmc)
+{
+	int err;
+
+	intel_host->hs_caps = ~0;
+
+	err = __intel_dsm(intel_host, dev, INTEL_DSM_FNS, &intel_host->dsm_fns);
+	if (err) {
+		pr_debug("%s: DSM not supported, error %d\n",
+			 mmc_hostname(mmc), err);
+		return;
+	}
+
+	pr_debug("%s: DSM function mask %#x\n",
+		 mmc_hostname(mmc), intel_host->dsm_fns);
+
+	intel_dsm(intel_host, dev, INTEL_DSM_HS_CAPS, &intel_host->hs_caps);
+}
+
+static int intel_start_signal_voltage_switch(struct mmc_host *mmc,
+					     struct mmc_ios *ios)
+{
+	struct device *dev = mmc_dev(mmc);
+	struct sdhci_acpi_host *c = dev_get_drvdata(dev);
+	struct intel_host *intel_host = sdhci_acpi_priv(c);
+	unsigned int fn;
+	u32 result = 0;
+	int err;
+
+	err = sdhci_start_signal_voltage_switch(mmc, ios);
+	if (err)
+		return err;
+
+	switch (ios->signal_voltage) {
+	case MMC_SIGNAL_VOLTAGE_330:
+		fn = INTEL_DSM_V33_SWITCH;
+		break;
+	case MMC_SIGNAL_VOLTAGE_180:
+		fn = INTEL_DSM_V18_SWITCH;
+		break;
+	default:
+		return 0;
+	}
+
+	err = intel_dsm(intel_host, dev, fn, &result);
+	pr_debug("%s: %s DSM fn %u error %d result %u\n",
+		 mmc_hostname(mmc), __func__, fn, err, result);
+
+	return 0;
+}
+
+static void sdhci_acpi_int_hw_reset(struct sdhci_host *host)
+{
+	u8 reg;
+
+	reg = sdhci_readb(host, SDHCI_POWER_CONTROL);
+	reg |= 0x10;
+	sdhci_writeb(host, reg, SDHCI_POWER_CONTROL);
+	/* For eMMC, minimum is 1us but give it 9us for good measure */
+	udelay(9);
+	reg &= ~0x10;
+	sdhci_writeb(host, reg, SDHCI_POWER_CONTROL);
+	/* For eMMC, minimum is 200us but give it 300us for good measure */
+	usleep_range(300, 1000);
+}
+
+static const struct sdhci_ops sdhci_acpi_ops_dflt = {
+	.set_clock = sdhci_set_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static const struct sdhci_ops sdhci_acpi_ops_int = {
+	.set_clock = sdhci_set_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+	.hw_reset   = sdhci_acpi_int_hw_reset,
+};
+
+static const struct sdhci_acpi_chip sdhci_acpi_chip_int = {
+	.ops = &sdhci_acpi_ops_int,
+};
+
+#ifdef CONFIG_X86
+
+static bool sdhci_acpi_byt(void)
+{
+	static const struct x86_cpu_id byt[] = {
+		X86_MATCH_INTEL_FAM6_MODEL(ATOM_SILVERMONT, NULL),
+		{}
+	};
+
+	return x86_match_cpu(byt);
+}
+
+static bool sdhci_acpi_cht(void)
+{
+	static const struct x86_cpu_id cht[] = {
+		X86_MATCH_INTEL_FAM6_MODEL(ATOM_AIRMONT, NULL),
+		{}
+	};
+
+	return x86_match_cpu(cht);
+}
+
+#define BYT_IOSF_SCCEP			0x63
+#define BYT_IOSF_OCP_NETCTRL0		0x1078
+#define BYT_IOSF_OCP_TIMEOUT_BASE	GENMASK(10, 8)
+
+static void sdhci_acpi_byt_setting(struct device *dev)
+{
+	u32 val = 0;
+
+	if (!sdhci_acpi_byt())
+		return;
+
+	if (iosf_mbi_read(BYT_IOSF_SCCEP, MBI_CR_READ, BYT_IOSF_OCP_NETCTRL0,
+			  &val)) {
+		dev_err(dev, "%s read error\n", __func__);
+		return;
+	}
+
+	if (!(val & BYT_IOSF_OCP_TIMEOUT_BASE))
+		return;
+
+	val &= ~BYT_IOSF_OCP_TIMEOUT_BASE;
+
+	if (iosf_mbi_write(BYT_IOSF_SCCEP, MBI_CR_WRITE, BYT_IOSF_OCP_NETCTRL0,
+			   val)) {
+		dev_err(dev, "%s write error\n", __func__);
+		return;
+	}
+
+	dev_dbg(dev, "%s completed\n", __func__);
+}
+
+static bool sdhci_acpi_byt_defer(struct device *dev)
+{
+	if (!sdhci_acpi_byt())
+		return false;
+
+	if (!iosf_mbi_available())
+		return true;
+
+	sdhci_acpi_byt_setting(dev);
+
+	return false;
+}
+
+static bool sdhci_acpi_cht_pci_wifi(unsigned int vendor, unsigned int device,
+				    unsigned int slot, unsigned int parent_slot)
+{
+	struct pci_dev *dev, *parent, *from = NULL;
+
+	while (1) {
+		dev = pci_get_device(vendor, device, from);
+		pci_dev_put(from);
+		if (!dev)
+			break;
+		parent = pci_upstream_bridge(dev);
+		if (ACPI_COMPANION(&dev->dev) && PCI_SLOT(dev->devfn) == slot &&
+		    parent && PCI_SLOT(parent->devfn) == parent_slot &&
+		    !pci_upstream_bridge(parent)) {
+			pci_dev_put(dev);
+			return true;
+		}
+		from = dev;
+	}
+
+	return false;
+}
+
+/*
+ * GPDwin uses PCI wifi which conflicts with SDIO's use of
+ * acpi_device_fix_up_power() on child device nodes. Identifying GPDwin is
+ * problematic, but since SDIO is only used for wifi, the presence of the PCI
+ * wifi card in the expected slot with an ACPI companion node, is used to
+ * indicate that acpi_device_fix_up_power() should be avoided.
+ */
+static inline bool sdhci_acpi_no_fixup_child_power(struct acpi_device *adev)
+{
+	return sdhci_acpi_cht() &&
+	       acpi_dev_hid_uid_match(adev, "80860F14", "2") &&
+	       sdhci_acpi_cht_pci_wifi(0x14e4, 0x43ec, 0, 28);
+}
+
+#else
+
+static inline void sdhci_acpi_byt_setting(struct device *dev)
+{
+}
+
+static inline bool sdhci_acpi_byt_defer(struct device *dev)
+{
+	return false;
+}
+
+static inline bool sdhci_acpi_no_fixup_child_power(struct acpi_device *adev)
+{
+	return false;
+}
+
+#endif
+
+static int bxt_get_cd(struct mmc_host *mmc)
+{
+	int gpio_cd = mmc_gpio_get_cd(mmc);
+	struct sdhci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+	int ret = 0;
+
+	if (!gpio_cd)
+		return 0;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (host->flags & SDHCI_DEVICE_DEAD)
+		goto out;
+
+	ret = !!(sdhci_readl(host, SDHCI_PRESENT_STATE) & SDHCI_CARD_PRESENT);
+out:
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	return ret;
+}
+
+static int intel_probe_slot(struct platform_device *pdev, struct acpi_device *adev)
+{
+	struct sdhci_acpi_host *c = platform_get_drvdata(pdev);
+	struct intel_host *intel_host = sdhci_acpi_priv(c);
+	struct sdhci_host *host = c->host;
+
+	if (acpi_dev_hid_uid_match(adev, "80860F14", "1") &&
+	    sdhci_readl(host, SDHCI_CAPABILITIES) == 0x446cc8b2 &&
+	    sdhci_readl(host, SDHCI_CAPABILITIES_1) == 0x00000807)
+		host->timeout_clk = 1000; /* 1000 kHz i.e. 1 MHz */
+
+	if (acpi_dev_hid_uid_match(adev, "80865ACA", NULL))
+		host->mmc_host_ops.get_cd = bxt_get_cd;
+
+	intel_dsm_init(intel_host, &pdev->dev, host->mmc);
+
+	host->mmc_host_ops.start_signal_voltage_switch =
+					intel_start_signal_voltage_switch;
+
+	c->is_intel = true;
+
+	return 0;
+}
+
+static int intel_setup_host(struct platform_device *pdev)
+{
+	struct sdhci_acpi_host *c = platform_get_drvdata(pdev);
+	struct intel_host *intel_host = sdhci_acpi_priv(c);
+
+	if (!(intel_host->hs_caps & INTEL_DSM_HS_CAPS_SDR25))
+		c->host->mmc->caps &= ~MMC_CAP_UHS_SDR25;
+
+	if (!(intel_host->hs_caps & INTEL_DSM_HS_CAPS_SDR50))
+		c->host->mmc->caps &= ~MMC_CAP_UHS_SDR50;
+
+	if (!(intel_host->hs_caps & INTEL_DSM_HS_CAPS_DDR50))
+		c->host->mmc->caps &= ~MMC_CAP_UHS_DDR50;
+
+	if (!(intel_host->hs_caps & INTEL_DSM_HS_CAPS_SDR104))
+		c->host->mmc->caps &= ~MMC_CAP_UHS_SDR104;
+
+	return 0;
+}
+
+static const struct sdhci_acpi_slot sdhci_acpi_slot_int_emmc = {
+	.chip    = &sdhci_acpi_chip_int,
+	.caps    = MMC_CAP_8_BIT_DATA | MMC_CAP_NONREMOVABLE |
+		   MMC_CAP_HW_RESET | MMC_CAP_1_8V_DDR |
+		   MMC_CAP_CMD_DURING_TFR | MMC_CAP_WAIT_WHILE_BUSY,
+	.flags   = SDHCI_ACPI_RUNTIME_PM,
+	.quirks  = SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+		   SDHCI_QUIRK_NO_LED,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+		   SDHCI_QUIRK2_STOP_WITH_TC |
+		   SDHCI_QUIRK2_CAPS_BIT63_FOR_HS400,
+	.probe_slot	= intel_probe_slot,
+	.setup_host	= intel_setup_host,
+	.priv_size	= sizeof(struct intel_host),
+};
+
+static const struct sdhci_acpi_slot sdhci_acpi_slot_int_sdio = {
+	.quirks  = SDHCI_QUIRK_BROKEN_CARD_DETECTION |
+		   SDHCI_QUIRK_NO_LED |
+		   SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.quirks2 = SDHCI_QUIRK2_HOST_OFF_CARD_ON,
+	.caps    = MMC_CAP_NONREMOVABLE | MMC_CAP_POWER_OFF_CARD |
+		   MMC_CAP_WAIT_WHILE_BUSY,
+	.flags   = SDHCI_ACPI_RUNTIME_PM,
+	.pm_caps = MMC_PM_KEEP_POWER,
+	.probe_slot	= intel_probe_slot,
+	.setup_host	= intel_setup_host,
+	.priv_size	= sizeof(struct intel_host),
+};
+
+static const struct sdhci_acpi_slot sdhci_acpi_slot_int_sd = {
+	.flags   = SDHCI_ACPI_SD_CD | SDHCI_ACPI_SD_CD_OVERRIDE_LEVEL |
+		   SDHCI_ACPI_RUNTIME_PM,
+	.quirks  = SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+		   SDHCI_QUIRK_NO_LED,
+	.quirks2 = SDHCI_QUIRK2_CARD_ON_NEEDS_BUS_ON |
+		   SDHCI_QUIRK2_STOP_WITH_TC,
+	.caps    = MMC_CAP_WAIT_WHILE_BUSY | MMC_CAP_AGGRESSIVE_PM,
+	.probe_slot	= intel_probe_slot,
+	.setup_host	= intel_setup_host,
+	.priv_size	= sizeof(struct intel_host),
+};
+
+#define VENDOR_SPECIFIC_PWRCTL_CLEAR_REG	0x1a8
+#define VENDOR_SPECIFIC_PWRCTL_CTL_REG		0x1ac
+static irqreturn_t sdhci_acpi_qcom_handler(int irq, void *ptr)
+{
+	struct sdhci_host *host = ptr;
+
+	sdhci_writel(host, 0x3, VENDOR_SPECIFIC_PWRCTL_CLEAR_REG);
+	sdhci_writel(host, 0x1, VENDOR_SPECIFIC_PWRCTL_CTL_REG);
+
+	return IRQ_HANDLED;
+}
+
+static int qcom_probe_slot(struct platform_device *pdev, struct acpi_device *adev)
+{
+	struct sdhci_acpi_host *c = platform_get_drvdata(pdev);
+	struct sdhci_host *host = c->host;
+	int *irq = sdhci_acpi_priv(c);
+
+	*irq = -EINVAL;
+
+	if (!acpi_dev_hid_uid_match(adev, "QCOM8051", NULL))
+		return 0;
+
+	*irq = platform_get_irq(pdev, 1);
+	if (*irq < 0)
+		return 0;
+
+	return request_threaded_irq(*irq, NULL, sdhci_acpi_qcom_handler,
+				    IRQF_ONESHOT | IRQF_TRIGGER_HIGH,
+				    "sdhci_qcom", host);
+}
+
+static int qcom_free_slot(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct sdhci_acpi_host *c = platform_get_drvdata(pdev);
+	struct sdhci_host *host = c->host;
+	struct acpi_device *adev;
+	int *irq = sdhci_acpi_priv(c);
+
+	adev = ACPI_COMPANION(dev);
+	if (!adev)
+		return -ENODEV;
+
+	if (!acpi_dev_hid_uid_match(adev, "QCOM8051", NULL))
+		return 0;
+
+	if (*irq < 0)
+		return 0;
+
+	free_irq(*irq, host);
+	return 0;
+}
+
+static const struct sdhci_acpi_slot sdhci_acpi_slot_qcom_sd_3v = {
+	.quirks  = SDHCI_QUIRK_BROKEN_CARD_DETECTION,
+	.quirks2 = SDHCI_QUIRK2_NO_1_8_V,
+	.caps    = MMC_CAP_NONREMOVABLE,
+	.priv_size	= sizeof(int),
+	.probe_slot	= qcom_probe_slot,
+	.free_slot	= qcom_free_slot,
+};
+
+static const struct sdhci_acpi_slot sdhci_acpi_slot_qcom_sd = {
+	.quirks  = SDHCI_QUIRK_BROKEN_CARD_DETECTION,
+	.caps    = MMC_CAP_NONREMOVABLE,
+};
+
+struct amd_sdhci_host {
+	bool	tuned_clock;
+	bool	dll_enabled;
+};
+
+/* AMD sdhci reset dll register. */
+#define SDHCI_AMD_RESET_DLL_REGISTER    0x908
+
+static int amd_select_drive_strength(struct mmc_card *card,
+				     unsigned int max_dtr, int host_drv,
+				     int card_drv, int *drv_type)
+{
+	*drv_type = MMC_SET_DRIVER_TYPE_A;
+	return MMC_SET_DRIVER_TYPE_A;
+}
+
+static void sdhci_acpi_amd_hs400_dll(struct sdhci_host *host, bool enable)
+{
+	struct sdhci_acpi_host *acpi_host = sdhci_priv(host);
+	struct amd_sdhci_host *amd_host = sdhci_acpi_priv(acpi_host);
+
+	/* AMD Platform requires dll setting */
+	sdhci_writel(host, 0x40003210, SDHCI_AMD_RESET_DLL_REGISTER);
+	usleep_range(10, 20);
+	if (enable)
+		sdhci_writel(host, 0x40033210, SDHCI_AMD_RESET_DLL_REGISTER);
+
+	amd_host->dll_enabled = enable;
+}
+
+/*
+ * The initialization sequence for HS400 is:
+ *     HS->HS200->Perform Tuning->HS->HS400
+ *
+ * The re-tuning sequence is:
+ *     HS400->DDR52->HS->HS200->Perform Tuning->HS->HS400
+ *
+ * The AMD eMMC Controller can only use the tuned clock while in HS200 and HS400
+ * mode. If we switch to a different mode, we need to disable the tuned clock.
+ * If we have previously performed tuning and switch back to HS200 or
+ * HS400, we can re-enable the tuned clock.
+ *
+ */
+static void amd_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_acpi_host *acpi_host = sdhci_priv(host);
+	struct amd_sdhci_host *amd_host = sdhci_acpi_priv(acpi_host);
+	unsigned int old_timing = host->timing;
+	u16 val;
+
+	sdhci_set_ios(mmc, ios);
+
+	if (old_timing != host->timing && amd_host->tuned_clock) {
+		if (host->timing == MMC_TIMING_MMC_HS400 ||
+		    host->timing == MMC_TIMING_MMC_HS200) {
+			val = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+			val |= SDHCI_CTRL_TUNED_CLK;
+			sdhci_writew(host, val, SDHCI_HOST_CONTROL2);
+		} else {
+			val = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+			val &= ~SDHCI_CTRL_TUNED_CLK;
+			sdhci_writew(host, val, SDHCI_HOST_CONTROL2);
+		}
+
+		/* DLL is only required for HS400 */
+		if (host->timing == MMC_TIMING_MMC_HS400 &&
+		    !amd_host->dll_enabled)
+			sdhci_acpi_amd_hs400_dll(host, true);
+	}
+}
+
+static int amd_sdhci_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	int err;
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_acpi_host *acpi_host = sdhci_priv(host);
+	struct amd_sdhci_host *amd_host = sdhci_acpi_priv(acpi_host);
+
+	amd_host->tuned_clock = false;
+
+	err = sdhci_execute_tuning(mmc, opcode);
+
+	if (!err && !host->tuning_err)
+		amd_host->tuned_clock = true;
+
+	return err;
+}
+
+static void amd_sdhci_reset(struct sdhci_host *host, u8 mask)
+{
+	struct sdhci_acpi_host *acpi_host = sdhci_priv(host);
+	struct amd_sdhci_host *amd_host = sdhci_acpi_priv(acpi_host);
+
+	if (mask & SDHCI_RESET_ALL) {
+		amd_host->tuned_clock = false;
+		sdhci_acpi_amd_hs400_dll(host, false);
+	}
+
+	sdhci_reset(host, mask);
+}
+
+static const struct sdhci_ops sdhci_acpi_ops_amd = {
+	.set_clock	= sdhci_set_clock,
+	.set_bus_width	= sdhci_set_bus_width,
+	.reset		= amd_sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static const struct sdhci_acpi_chip sdhci_acpi_chip_amd = {
+	.ops = &sdhci_acpi_ops_amd,
+};
+
+static int sdhci_acpi_emmc_amd_probe_slot(struct platform_device *pdev,
+					  struct acpi_device *adev)
+{
+	struct sdhci_acpi_host *c = platform_get_drvdata(pdev);
+	struct sdhci_host *host   = c->host;
+
+	sdhci_read_caps(host);
+	if (host->caps1 & SDHCI_SUPPORT_DDR50)
+		host->mmc->caps = MMC_CAP_1_8V_DDR;
+
+	if ((host->caps1 & SDHCI_SUPPORT_SDR104) &&
+	    (host->mmc->caps & MMC_CAP_1_8V_DDR))
+		host->mmc->caps2 = MMC_CAP2_HS400_1_8V;
+
+	/*
+	 * There are two types of presets out in the wild:
+	 * 1) Default/broken presets.
+	 *    These presets have two sets of problems:
+	 *    a) The clock divisor for SDR12, SDR25, and SDR50 is too small.
+	 *       This results in clock frequencies that are 2x higher than
+	 *       acceptable. i.e., SDR12 = 25 MHz, SDR25 = 50 MHz, SDR50 =
+	 *       100 MHz.x
+	 *    b) The HS200 and HS400 driver strengths don't match.
+	 *       By default, the SDR104 preset register has a driver strength of
+	 *       A, but the (internal) HS400 preset register has a driver
+	 *       strength of B. As part of initializing HS400, HS200 tuning
+	 *       needs to be performed. Having different driver strengths
+	 *       between tuning and operation is wrong. It results in different
+	 *       rise/fall times that lead to incorrect sampling.
+	 * 2) Firmware with properly initialized presets.
+	 *    These presets have proper clock divisors. i.e., SDR12 => 12MHz,
+	 *    SDR25 => 25 MHz, SDR50 => 50 MHz. Additionally the HS200 and
+	 *    HS400 preset driver strengths match.
+	 *
+	 *    Enabling presets for HS400 doesn't work for the following reasons:
+	 *    1) sdhci_set_ios has a hard coded list of timings that are used
+	 *       to determine if presets should be enabled.
+	 *    2) sdhci_get_preset_value is using a non-standard register to
+	 *       read out HS400 presets. The AMD controller doesn't support this
+	 *       non-standard register. In fact, it doesn't expose the HS400
+	 *       preset register anywhere in the SDHCI memory map. This results
+	 *       in reading a garbage value and using the wrong presets.
+	 *
+	 *       Since HS400 and HS200 presets must be identical, we could
+	 *       instead use the the SDR104 preset register.
+	 *
+	 *    If the above issues are resolved we could remove this quirk for
+	 *    firmware that that has valid presets (i.e., SDR12 <= 12 MHz).
+	 */
+	host->quirks2 |= SDHCI_QUIRK2_PRESET_VALUE_BROKEN;
+
+	host->mmc_host_ops.select_drive_strength = amd_select_drive_strength;
+	host->mmc_host_ops.set_ios = amd_set_ios;
+	host->mmc_host_ops.execute_tuning = amd_sdhci_execute_tuning;
+	return 0;
+}
+
+static const struct sdhci_acpi_slot sdhci_acpi_slot_amd_emmc = {
+	.chip		= &sdhci_acpi_chip_amd,
+	.caps		= MMC_CAP_8_BIT_DATA | MMC_CAP_NONREMOVABLE,
+	.quirks		= SDHCI_QUIRK_32BIT_DMA_ADDR |
+			  SDHCI_QUIRK_32BIT_DMA_SIZE |
+			  SDHCI_QUIRK_32BIT_ADMA_SIZE,
+	.quirks2	= SDHCI_QUIRK2_BROKEN_64_BIT_DMA,
+	.probe_slot     = sdhci_acpi_emmc_amd_probe_slot,
+	.priv_size	= sizeof(struct amd_sdhci_host),
+};
+
+struct sdhci_acpi_uid_slot {
+	const char *hid;
+	const char *uid;
+	const struct sdhci_acpi_slot *slot;
+};
+
+static const struct sdhci_acpi_uid_slot sdhci_acpi_uids[] = {
+	{ "80865ACA", NULL, &sdhci_acpi_slot_int_sd },
+	{ "80865ACC", NULL, &sdhci_acpi_slot_int_emmc },
+	{ "80865AD0", NULL, &sdhci_acpi_slot_int_sdio },
+	{ "80860F14" , "1" , &sdhci_acpi_slot_int_emmc },
+	{ "80860F14" , "2" , &sdhci_acpi_slot_int_sdio },
+	{ "80860F14" , "3" , &sdhci_acpi_slot_int_sd   },
+	{ "80860F16" , NULL, &sdhci_acpi_slot_int_sd   },
+	{ "INT33BB"  , "2" , &sdhci_acpi_slot_int_sdio },
+	{ "INT33BB"  , "3" , &sdhci_acpi_slot_int_sd },
+	{ "INT33C6"  , NULL, &sdhci_acpi_slot_int_sdio },
+	{ "INT3436"  , NULL, &sdhci_acpi_slot_int_sdio },
+	{ "INT344D"  , NULL, &sdhci_acpi_slot_int_sdio },
+	{ "PNP0FFF"  , "3" , &sdhci_acpi_slot_int_sd   },
+	{ "PNP0D40"  },
+	{ "QCOM8051", NULL, &sdhci_acpi_slot_qcom_sd_3v },
+	{ "QCOM8052", NULL, &sdhci_acpi_slot_qcom_sd },
+	{ "AMDI0040", NULL, &sdhci_acpi_slot_amd_emmc },
+	{ },
+};
+
+static const struct acpi_device_id sdhci_acpi_ids[] = {
+	{ "80865ACA" },
+	{ "80865ACC" },
+	{ "80865AD0" },
+	{ "80860F14" },
+	{ "80860F16" },
+	{ "INT33BB"  },
+	{ "INT33C6"  },
+	{ "INT3436"  },
+	{ "INT344D"  },
+	{ "PNP0D40"  },
+	{ "QCOM8051" },
+	{ "QCOM8052" },
+	{ "AMDI0040" },
+	{ },
+};
+MODULE_DEVICE_TABLE(acpi, sdhci_acpi_ids);
+
+static const struct dmi_system_id sdhci_acpi_quirks[] = {
+	{
+		/*
+		 * The Lenovo Miix 320-10ICR has a bug in the _PS0 method of
+		 * the SHC1 ACPI device, this bug causes it to reprogram the
+		 * wrong LDO (DLDO3) to 1.8V if 1.8V modes are used and the
+		 * card is (runtime) suspended + resumed. DLDO3 is used for
+		 * the LCD and setting it to 1.8V causes the LCD to go black.
+		 */
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "LENOVO"),
+			DMI_MATCH(DMI_PRODUCT_VERSION, "Lenovo MIIX 320-10ICR"),
+		},
+		.driver_data = (void *)DMI_QUIRK_RESET_SD_SIGNAL_VOLT_ON_SUSP,
+	},
+	{
+		/*
+		 * The Acer Aspire Switch 10 (SW5-012) microSD slot always
+		 * reports the card being write-protected even though microSD
+		 * cards do not have a write-protect switch at all.
+		 */
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Acer"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "Aspire SW5-012"),
+		},
+		.driver_data = (void *)DMI_QUIRK_SD_NO_WRITE_PROTECT,
+	},
+	{
+		/*
+		 * The Toshiba WT8-B's microSD slot always reports the card being
+		 * write-protected.
+		 */
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "TOSHIBA"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "TOSHIBA ENCORE 2 WT8-B"),
+		},
+		.driver_data = (void *)DMI_QUIRK_SD_NO_WRITE_PROTECT,
+	},
+	{} /* Terminating entry */
+};
+
+static const struct sdhci_acpi_slot *sdhci_acpi_get_slot(struct acpi_device *adev)
+{
+	const struct sdhci_acpi_uid_slot *u;
+
+	for (u = sdhci_acpi_uids; u->hid; u++) {
+		if (acpi_dev_hid_uid_match(adev, u->hid, u->uid))
+			return u->slot;
+	}
+	return NULL;
+}
+
+static int sdhci_acpi_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	const struct sdhci_acpi_slot *slot;
+	struct acpi_device *device, *child;
+	const struct dmi_system_id *id;
+	struct sdhci_acpi_host *c;
+	struct sdhci_host *host;
+	struct resource *iomem;
+	resource_size_t len;
+	size_t priv_size;
+	int quirks = 0;
+	int err;
+
+	device = ACPI_COMPANION(dev);
+	if (!device)
+		return -ENODEV;
+
+	id = dmi_first_match(sdhci_acpi_quirks);
+	if (id)
+		quirks = (long)id->driver_data;
+
+	slot = sdhci_acpi_get_slot(device);
+
+	/* Power on the SDHCI controller and its children */
+	acpi_device_fix_up_power(device);
+	if (!sdhci_acpi_no_fixup_child_power(device)) {
+		list_for_each_entry(child, &device->children, node)
+			if (child->status.present && child->status.enabled)
+				acpi_device_fix_up_power(child);
+	}
+
+	if (sdhci_acpi_byt_defer(dev))
+		return -EPROBE_DEFER;
+
+	iomem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!iomem)
+		return -ENOMEM;
+
+	len = resource_size(iomem);
+	if (len < 0x100)
+		dev_err(dev, "Invalid iomem size!\n");
+
+	if (!devm_request_mem_region(dev, iomem->start, len, dev_name(dev)))
+		return -ENOMEM;
+
+	priv_size = slot ? slot->priv_size : 0;
+	host = sdhci_alloc_host(dev, sizeof(struct sdhci_acpi_host) + priv_size);
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	c = sdhci_priv(host);
+	c->host = host;
+	c->slot = slot;
+	c->pdev = pdev;
+	c->use_runtime_pm = sdhci_acpi_flag(c, SDHCI_ACPI_RUNTIME_PM);
+
+	platform_set_drvdata(pdev, c);
+
+	host->hw_name	= "ACPI";
+	host->ops	= &sdhci_acpi_ops_dflt;
+	host->irq	= platform_get_irq(pdev, 0);
+	if (host->irq < 0) {
+		err = host->irq;
+		goto err_free;
+	}
+
+	host->ioaddr = devm_ioremap(dev, iomem->start,
+					    resource_size(iomem));
+	if (host->ioaddr == NULL) {
+		err = -ENOMEM;
+		goto err_free;
+	}
+
+	if (c->slot) {
+		if (c->slot->probe_slot) {
+			err = c->slot->probe_slot(pdev, device);
+			if (err)
+				goto err_free;
+		}
+		if (c->slot->chip) {
+			host->ops            = c->slot->chip->ops;
+			host->quirks        |= c->slot->chip->quirks;
+			host->quirks2       |= c->slot->chip->quirks2;
+			host->mmc->caps     |= c->slot->chip->caps;
+			host->mmc->caps2    |= c->slot->chip->caps2;
+			host->mmc->pm_caps  |= c->slot->chip->pm_caps;
+		}
+		host->quirks        |= c->slot->quirks;
+		host->quirks2       |= c->slot->quirks2;
+		host->mmc->caps     |= c->slot->caps;
+		host->mmc->caps2    |= c->slot->caps2;
+		host->mmc->pm_caps  |= c->slot->pm_caps;
+	}
+
+	host->mmc->caps2 |= MMC_CAP2_NO_PRESCAN_POWERUP;
+
+	if (sdhci_acpi_flag(c, SDHCI_ACPI_SD_CD)) {
+		bool v = sdhci_acpi_flag(c, SDHCI_ACPI_SD_CD_OVERRIDE_LEVEL);
+
+		err = mmc_gpiod_request_cd(host->mmc, NULL, 0, v, 0);
+		if (err) {
+			if (err == -EPROBE_DEFER)
+				goto err_free;
+			dev_warn(dev, "failed to setup card detect gpio\n");
+			c->use_runtime_pm = false;
+		}
+
+		if (quirks & DMI_QUIRK_RESET_SD_SIGNAL_VOLT_ON_SUSP)
+			c->reset_signal_volt_on_suspend = true;
+
+		if (quirks & DMI_QUIRK_SD_NO_WRITE_PROTECT)
+			host->mmc->caps2 |= MMC_CAP2_NO_WRITE_PROTECT;
+	}
+
+	err = sdhci_setup_host(host);
+	if (err)
+		goto err_free;
+
+	if (c->slot && c->slot->setup_host) {
+		err = c->slot->setup_host(pdev);
+		if (err)
+			goto err_cleanup;
+	}
+
+	err = __sdhci_add_host(host);
+	if (err)
+		goto err_cleanup;
+
+	if (c->use_runtime_pm) {
+		pm_runtime_set_active(dev);
+		pm_suspend_ignore_children(dev, 1);
+		pm_runtime_set_autosuspend_delay(dev, 50);
+		pm_runtime_use_autosuspend(dev);
+		pm_runtime_enable(dev);
+	}
+
+	device_enable_async_suspend(dev);
+
+	return 0;
+
+err_cleanup:
+	sdhci_cleanup_host(c->host);
+err_free:
+	if (c->slot && c->slot->free_slot)
+		c->slot->free_slot(pdev);
+
+	sdhci_free_host(c->host);
+	return err;
+}
+
+static int sdhci_acpi_remove(struct platform_device *pdev)
+{
+	struct sdhci_acpi_host *c = platform_get_drvdata(pdev);
+	struct device *dev = &pdev->dev;
+	int dead;
+
+	if (c->use_runtime_pm) {
+		pm_runtime_get_sync(dev);
+		pm_runtime_disable(dev);
+		pm_runtime_put_noidle(dev);
+	}
+
+	if (c->slot && c->slot->remove_slot)
+		c->slot->remove_slot(pdev);
+
+	dead = (sdhci_readl(c->host, SDHCI_INT_STATUS) == ~0);
+	sdhci_remove_host(c->host, dead);
+
+	if (c->slot && c->slot->free_slot)
+		c->slot->free_slot(pdev);
+
+	sdhci_free_host(c->host);
+
+	return 0;
+}
+
+static void __maybe_unused sdhci_acpi_reset_signal_voltage_if_needed(
+	struct device *dev)
+{
+	struct sdhci_acpi_host *c = dev_get_drvdata(dev);
+	struct sdhci_host *host = c->host;
+
+	if (c->is_intel && c->reset_signal_volt_on_suspend &&
+	    host->mmc->ios.signal_voltage != MMC_SIGNAL_VOLTAGE_330) {
+		struct intel_host *intel_host = sdhci_acpi_priv(c);
+		unsigned int fn = INTEL_DSM_V33_SWITCH;
+		u32 result = 0;
+
+		intel_dsm(intel_host, dev, fn, &result);
+	}
+}
+
+#ifdef CONFIG_PM_SLEEP
+
+static int sdhci_acpi_suspend(struct device *dev)
+{
+	struct sdhci_acpi_host *c = dev_get_drvdata(dev);
+	struct sdhci_host *host = c->host;
+	int ret;
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	ret = sdhci_suspend_host(host);
+	if (ret)
+		return ret;
+
+	sdhci_acpi_reset_signal_voltage_if_needed(dev);
+	return 0;
+}
+
+static int sdhci_acpi_resume(struct device *dev)
+{
+	struct sdhci_acpi_host *c = dev_get_drvdata(dev);
+
+	sdhci_acpi_byt_setting(&c->pdev->dev);
+
+	return sdhci_resume_host(c->host);
+}
+
+#endif
+
+#ifdef CONFIG_PM
+
+static int sdhci_acpi_runtime_suspend(struct device *dev)
+{
+	struct sdhci_acpi_host *c = dev_get_drvdata(dev);
+	struct sdhci_host *host = c->host;
+	int ret;
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	ret = sdhci_runtime_suspend_host(host);
+	if (ret)
+		return ret;
+
+	sdhci_acpi_reset_signal_voltage_if_needed(dev);
+	return 0;
+}
+
+static int sdhci_acpi_runtime_resume(struct device *dev)
+{
+	struct sdhci_acpi_host *c = dev_get_drvdata(dev);
+
+	sdhci_acpi_byt_setting(&c->pdev->dev);
+
+	return sdhci_runtime_resume_host(c->host, 0);
+}
+
+#endif
+
+static const struct dev_pm_ops sdhci_acpi_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(sdhci_acpi_suspend, sdhci_acpi_resume)
+	SET_RUNTIME_PM_OPS(sdhci_acpi_runtime_suspend,
+			sdhci_acpi_runtime_resume, NULL)
+};
+
+static struct platform_driver sdhci_acpi_driver = {
+	.driver = {
+		.name			= "sdhci-acpi",
+		.probe_type		= PROBE_PREFER_ASYNCHRONOUS,
+		.acpi_match_table	= sdhci_acpi_ids,
+		.pm			= &sdhci_acpi_pm_ops,
+	},
+	.probe	= sdhci_acpi_probe,
+	.remove	= sdhci_acpi_remove,
+};
+
+module_platform_driver(sdhci_acpi_driver);
+
+MODULE_DESCRIPTION("Secure Digital Host Controller Interface ACPI driver");
+MODULE_AUTHOR("Adrian Hunter");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-bcm-kona.c b/drivers/mmc/host/sdhci-bcm-kona.c
new file mode 100644
index 000000000..4d4aac85c
--- /dev/null
+++ b/drivers/mmc/host/sdhci-bcm-kona.c
@@ -0,0 +1,338 @@
+/*
+ * Copyright (C) 2013 Broadcom Corporation
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/delay.h>
+#include <linux/highmem.h>
+#include <linux/platform_device.h>
+#include <linux/mmc/host.h>
+#include <linux/io.h>
+#include <linux/clk.h>
+#include <linux/regulator/consumer.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/mmc/slot-gpio.h>
+
+#include "sdhci-pltfm.h"
+#include "sdhci.h"
+
+#define SDHCI_SOFT_RESET			0x01000000
+#define KONA_SDHOST_CORECTRL			0x8000
+#define KONA_SDHOST_CD_PINCTRL			0x00000008
+#define KONA_SDHOST_STOP_HCLK			0x00000004
+#define KONA_SDHOST_RESET			0x00000002
+#define KONA_SDHOST_EN				0x00000001
+
+#define KONA_SDHOST_CORESTAT			0x8004
+#define KONA_SDHOST_WP				0x00000002
+#define KONA_SDHOST_CD_SW			0x00000001
+
+#define KONA_SDHOST_COREIMR			0x8008
+#define KONA_SDHOST_IP				0x00000001
+
+#define KONA_SDHOST_COREISR			0x800C
+#define KONA_SDHOST_COREIMSR			0x8010
+#define KONA_SDHOST_COREDBG1			0x8014
+#define KONA_SDHOST_COREGPO_MASK		0x8018
+
+#define SD_DETECT_GPIO_DEBOUNCE_128MS		128
+
+#define KONA_MMC_AUTOSUSPEND_DELAY		(50)
+
+struct sdhci_bcm_kona_dev {
+	struct mutex	write_lock; /* protect back to back writes */
+};
+
+
+static int sdhci_bcm_kona_sd_reset(struct sdhci_host *host)
+{
+	unsigned int val;
+	unsigned long timeout;
+
+	/* This timeout should be sufficent for core to reset */
+	timeout = jiffies + msecs_to_jiffies(100);
+
+	/* reset the host using the top level reset */
+	val = sdhci_readl(host, KONA_SDHOST_CORECTRL);
+	val |= KONA_SDHOST_RESET;
+	sdhci_writel(host, val, KONA_SDHOST_CORECTRL);
+
+	while (!(sdhci_readl(host, KONA_SDHOST_CORECTRL) & KONA_SDHOST_RESET)) {
+		if (time_is_before_jiffies(timeout)) {
+			pr_err("Error: sd host is stuck in reset!!!\n");
+			return -EFAULT;
+		}
+	}
+
+	/* bring the host out of reset */
+	val = sdhci_readl(host, KONA_SDHOST_CORECTRL);
+	val &= ~KONA_SDHOST_RESET;
+
+	/*
+	 * Back-to-Back register write needs a delay of 1ms at bootup (min 10uS)
+	 * Back-to-Back writes to same register needs delay when SD bus clock
+	 * is very low w.r.t AHB clock, mainly during boot-time and during card
+	 * insert-removal.
+	 */
+	usleep_range(1000, 5000);
+	sdhci_writel(host, val, KONA_SDHOST_CORECTRL);
+
+	return 0;
+}
+
+static void sdhci_bcm_kona_sd_init(struct sdhci_host *host)
+{
+	unsigned int val;
+
+	/* enable the interrupt from the IP core */
+	val = sdhci_readl(host, KONA_SDHOST_COREIMR);
+	val |= KONA_SDHOST_IP;
+	sdhci_writel(host, val, KONA_SDHOST_COREIMR);
+
+	/* Enable the AHB clock gating module to the host */
+	val = sdhci_readl(host, KONA_SDHOST_CORECTRL);
+	val |= KONA_SDHOST_EN;
+
+	/*
+	 * Back-to-Back register write needs a delay of 1ms at bootup (min 10uS)
+	 * Back-to-Back writes to same register needs delay when SD bus clock
+	 * is very low w.r.t AHB clock, mainly during boot-time and during card
+	 * insert-removal.
+	 */
+	usleep_range(1000, 5000);
+	sdhci_writel(host, val, KONA_SDHOST_CORECTRL);
+}
+
+/*
+ * Software emulation of the SD card insertion/removal. Set insert=1 for insert
+ * and insert=0 for removal. The card detection is done by GPIO. For Broadcom
+ * IP to function properly the bit 0 of CORESTAT register needs to be set/reset
+ * to generate the CD IRQ handled in sdhci.c which schedules card_tasklet.
+ */
+static int sdhci_bcm_kona_sd_card_emulate(struct sdhci_host *host, int insert)
+{
+	struct sdhci_pltfm_host *pltfm_priv = sdhci_priv(host);
+	struct sdhci_bcm_kona_dev *kona_dev = sdhci_pltfm_priv(pltfm_priv);
+	u32 val;
+
+	/*
+	 * Back-to-Back register write needs a delay of min 10uS.
+	 * Back-to-Back writes to same register needs delay when SD bus clock
+	 * is very low w.r.t AHB clock, mainly during boot-time and during card
+	 * insert-removal.
+	 * We keep 20uS
+	 */
+	mutex_lock(&kona_dev->write_lock);
+	udelay(20);
+	val = sdhci_readl(host, KONA_SDHOST_CORESTAT);
+
+	if (insert) {
+		int ret;
+
+		ret = mmc_gpio_get_ro(host->mmc);
+		if (ret >= 0)
+			val = (val & ~KONA_SDHOST_WP) |
+				((ret) ? KONA_SDHOST_WP : 0);
+
+		val |= KONA_SDHOST_CD_SW;
+		sdhci_writel(host, val, KONA_SDHOST_CORESTAT);
+	} else {
+		val &= ~KONA_SDHOST_CD_SW;
+		sdhci_writel(host, val, KONA_SDHOST_CORESTAT);
+	}
+	mutex_unlock(&kona_dev->write_lock);
+
+	return 0;
+}
+
+/*
+ * SD card interrupt event callback
+ */
+static void sdhci_bcm_kona_card_event(struct sdhci_host *host)
+{
+	if (mmc_gpio_get_cd(host->mmc) > 0) {
+		dev_dbg(mmc_dev(host->mmc),
+			"card inserted\n");
+		sdhci_bcm_kona_sd_card_emulate(host, 1);
+	} else {
+		dev_dbg(mmc_dev(host->mmc),
+			"card removed\n");
+		sdhci_bcm_kona_sd_card_emulate(host, 0);
+	}
+}
+
+static void sdhci_bcm_kona_init_74_clocks(struct sdhci_host *host,
+				u8 power_mode)
+{
+	/*
+	 *  JEDEC and SD spec specify supplying 74 continuous clocks to
+	 * device after power up. With minimum bus (100KHz) that
+	 * that translates to 740us
+	 */
+	if (power_mode != MMC_POWER_OFF)
+		udelay(740);
+}
+
+static const struct sdhci_ops sdhci_bcm_kona_ops = {
+	.set_clock = sdhci_set_clock,
+	.get_max_clock = sdhci_pltfm_clk_get_max_clock,
+	.get_timeout_clock = sdhci_pltfm_clk_get_max_clock,
+	.platform_send_init_74_clocks = sdhci_bcm_kona_init_74_clocks,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+	.card_event = sdhci_bcm_kona_card_event,
+};
+
+static const struct sdhci_pltfm_data sdhci_pltfm_data_kona = {
+	.ops    = &sdhci_bcm_kona_ops,
+	.quirks = SDHCI_QUIRK_NO_CARD_NO_RESET |
+		SDHCI_QUIRK_BROKEN_TIMEOUT_VAL | SDHCI_QUIRK_32BIT_DMA_ADDR |
+		SDHCI_QUIRK_32BIT_DMA_SIZE | SDHCI_QUIRK_32BIT_ADMA_SIZE |
+		SDHCI_QUIRK_FORCE_BLK_SZ_2048 |
+		SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+};
+
+static const struct of_device_id sdhci_bcm_kona_of_match[] = {
+	{ .compatible = "brcm,kona-sdhci"},
+	{ .compatible = "bcm,kona-sdhci"}, /* deprecated name */
+	{}
+};
+MODULE_DEVICE_TABLE(of, sdhci_bcm_kona_of_match);
+
+static int sdhci_bcm_kona_probe(struct platform_device *pdev)
+{
+	struct sdhci_bcm_kona_dev *kona_dev = NULL;
+	struct sdhci_pltfm_host *pltfm_priv;
+	struct device *dev = &pdev->dev;
+	struct sdhci_host *host;
+	int ret;
+
+	ret = 0;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_pltfm_data_kona,
+			sizeof(*kona_dev));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	dev_dbg(dev, "%s: inited. IOADDR=%p\n", __func__, host->ioaddr);
+
+	pltfm_priv = sdhci_priv(host);
+
+	kona_dev = sdhci_pltfm_priv(pltfm_priv);
+	mutex_init(&kona_dev->write_lock);
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto err_pltfm_free;
+
+	if (!host->mmc->f_max) {
+		dev_err(&pdev->dev, "Missing max-freq for SDHCI cfg\n");
+		ret = -ENXIO;
+		goto err_pltfm_free;
+	}
+
+	/* Get and enable the core clock */
+	pltfm_priv->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(pltfm_priv->clk)) {
+		dev_err(dev, "Failed to get core clock\n");
+		ret = PTR_ERR(pltfm_priv->clk);
+		goto err_pltfm_free;
+	}
+
+	ret = clk_set_rate(pltfm_priv->clk, host->mmc->f_max);
+	if (ret) {
+		dev_err(dev, "Failed to set rate core clock\n");
+		goto err_pltfm_free;
+	}
+
+	ret = clk_prepare_enable(pltfm_priv->clk);
+	if (ret) {
+		dev_err(dev, "Failed to enable core clock\n");
+		goto err_pltfm_free;
+	}
+
+	dev_dbg(dev, "non-removable=%c\n",
+		mmc_card_is_removable(host->mmc) ? 'N' : 'Y');
+	dev_dbg(dev, "cd_gpio %c, wp_gpio %c\n",
+		(mmc_gpio_get_cd(host->mmc) != -ENOSYS) ? 'Y' : 'N',
+		(mmc_gpio_get_ro(host->mmc) != -ENOSYS) ? 'Y' : 'N');
+
+	if (!mmc_card_is_removable(host->mmc))
+		host->quirks |= SDHCI_QUIRK_BROKEN_CARD_DETECTION;
+
+	dev_dbg(dev, "is_8bit=%c\n",
+		(host->mmc->caps & MMC_CAP_8_BIT_DATA) ? 'Y' : 'N');
+
+	ret = sdhci_bcm_kona_sd_reset(host);
+	if (ret)
+		goto err_clk_disable;
+
+	sdhci_bcm_kona_sd_init(host);
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err_reset;
+
+	/* if device is eMMC, emulate card insert right here */
+	if (!mmc_card_is_removable(host->mmc)) {
+		ret = sdhci_bcm_kona_sd_card_emulate(host, 1);
+		if (ret) {
+			dev_err(dev,
+				"unable to emulate card insertion\n");
+			goto err_remove_host;
+		}
+	}
+	/*
+	 * Since the card detection GPIO interrupt is configured to be
+	 * edge sensitive, check the initial GPIO value here, emulate
+	 * only if the card is present
+	 */
+	if (mmc_gpio_get_cd(host->mmc) > 0)
+		sdhci_bcm_kona_sd_card_emulate(host, 1);
+
+	dev_dbg(dev, "initialized properly\n");
+	return 0;
+
+err_remove_host:
+	sdhci_remove_host(host, 0);
+
+err_reset:
+	sdhci_bcm_kona_sd_reset(host);
+
+err_clk_disable:
+	clk_disable_unprepare(pltfm_priv->clk);
+
+err_pltfm_free:
+	sdhci_pltfm_free(pdev);
+
+	dev_err(dev, "Probing of sdhci-pltfm failed: %d\n", ret);
+	return ret;
+}
+
+static struct platform_driver sdhci_bcm_kona_driver = {
+	.driver		= {
+		.name	= "sdhci-kona",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= &sdhci_pltfm_pmops,
+		.of_match_table = sdhci_bcm_kona_of_match,
+	},
+	.probe		= sdhci_bcm_kona_probe,
+	.remove		= sdhci_pltfm_unregister,
+};
+module_platform_driver(sdhci_bcm_kona_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for Broadcom Kona platform");
+MODULE_AUTHOR("Broadcom");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-brcmstb.c b/drivers/mmc/host/sdhci-brcmstb.c
new file mode 100644
index 000000000..4d42b1810
--- /dev/null
+++ b/drivers/mmc/host/sdhci-brcmstb.c
@@ -0,0 +1,371 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * sdhci-brcmstb.c Support for SDHCI on Broadcom BRCMSTB SoC's
+ *
+ * Copyright (C) 2015 Broadcom Corporation
+ */
+
+#include <linux/io.h>
+#include <linux/mmc/host.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/bitops.h>
+#include <linux/delay.h>
+
+#include "sdhci-cqhci.h"
+#include "sdhci-pltfm.h"
+#include "cqhci.h"
+
+#define SDHCI_VENDOR 0x78
+#define  SDHCI_VENDOR_ENHANCED_STRB 0x1
+#define  SDHCI_VENDOR_GATE_SDCLK_EN 0x2
+
+#define BRCMSTB_MATCH_FLAGS_NO_64BIT		BIT(0)
+#define BRCMSTB_MATCH_FLAGS_BROKEN_TIMEOUT	BIT(1)
+#define BRCMSTB_MATCH_FLAGS_HAS_CLOCK_GATE	BIT(2)
+
+#define BRCMSTB_PRIV_FLAGS_HAS_CQE		BIT(0)
+#define BRCMSTB_PRIV_FLAGS_GATE_CLOCK		BIT(1)
+
+#define SDHCI_ARASAN_CQE_BASE_ADDR		0x200
+
+struct sdhci_brcmstb_priv {
+	void __iomem *cfg_regs;
+	unsigned int flags;
+};
+
+struct brcmstb_match_priv {
+	void (*hs400es)(struct mmc_host *mmc, struct mmc_ios *ios);
+	struct sdhci_ops *ops;
+	const unsigned int flags;
+};
+
+static inline void enable_clock_gating(struct sdhci_host *host)
+{
+	u32 reg;
+
+	reg = sdhci_readl(host, SDHCI_VENDOR);
+	reg |= SDHCI_VENDOR_GATE_SDCLK_EN;
+	sdhci_writel(host, reg, SDHCI_VENDOR);
+}
+
+void brcmstb_reset(struct sdhci_host *host, u8 mask)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_brcmstb_priv *priv = sdhci_pltfm_priv(pltfm_host);
+
+	sdhci_and_cqhci_reset(host, mask);
+
+	/* Reset will clear this, so re-enable it */
+	if (priv->flags & BRCMSTB_PRIV_FLAGS_GATE_CLOCK)
+		enable_clock_gating(host);
+}
+
+static void sdhci_brcmstb_hs400es(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	u32 reg;
+
+	dev_dbg(mmc_dev(mmc), "%s(): Setting HS400-Enhanced-Strobe mode\n",
+		__func__);
+	reg = readl(host->ioaddr + SDHCI_VENDOR);
+	if (ios->enhanced_strobe)
+		reg |= SDHCI_VENDOR_ENHANCED_STRB;
+	else
+		reg &= ~SDHCI_VENDOR_ENHANCED_STRB;
+	writel(reg, host->ioaddr + SDHCI_VENDOR);
+}
+
+static void sdhci_brcmstb_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	u16 clk;
+
+	host->mmc->actual_clock = 0;
+
+	clk = sdhci_calc_clk(host, clock, &host->mmc->actual_clock);
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+	if (clock == 0)
+		return;
+
+	sdhci_enable_clk(host, clk);
+}
+
+static void sdhci_brcmstb_set_uhs_signaling(struct sdhci_host *host,
+					    unsigned int timing)
+{
+	u16 ctrl_2;
+
+	dev_dbg(mmc_dev(host->mmc), "%s: Setting UHS signaling for %d timing\n",
+		__func__, timing);
+	ctrl_2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	/* Select Bus Speed Mode for host */
+	ctrl_2 &= ~SDHCI_CTRL_UHS_MASK;
+	if ((timing == MMC_TIMING_MMC_HS200) ||
+	    (timing == MMC_TIMING_UHS_SDR104))
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR104;
+	else if (timing == MMC_TIMING_UHS_SDR12)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR12;
+	else if (timing == MMC_TIMING_SD_HS ||
+		 timing == MMC_TIMING_MMC_HS ||
+		 timing == MMC_TIMING_UHS_SDR25)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR25;
+	else if (timing == MMC_TIMING_UHS_SDR50)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR50;
+	else if ((timing == MMC_TIMING_UHS_DDR50) ||
+		 (timing == MMC_TIMING_MMC_DDR52))
+		ctrl_2 |= SDHCI_CTRL_UHS_DDR50;
+	else if (timing == MMC_TIMING_MMC_HS400)
+		ctrl_2 |= SDHCI_CTRL_HS400; /* Non-standard */
+	sdhci_writew(host, ctrl_2, SDHCI_HOST_CONTROL2);
+}
+
+static void sdhci_brcmstb_dumpregs(struct mmc_host *mmc)
+{
+	sdhci_dumpregs(mmc_priv(mmc));
+}
+
+static void sdhci_brcmstb_cqe_enable(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	u32 reg;
+
+	reg = sdhci_readl(host, SDHCI_PRESENT_STATE);
+	while (reg & SDHCI_DATA_AVAILABLE) {
+		sdhci_readl(host, SDHCI_BUFFER);
+		reg = sdhci_readl(host, SDHCI_PRESENT_STATE);
+	}
+
+	sdhci_cqe_enable(mmc);
+}
+
+static const struct cqhci_host_ops sdhci_brcmstb_cqhci_ops = {
+	.enable         = sdhci_brcmstb_cqe_enable,
+	.disable        = sdhci_cqe_disable,
+	.dumpregs       = sdhci_brcmstb_dumpregs,
+};
+
+static struct sdhci_ops sdhci_brcmstb_ops = {
+	.set_clock = sdhci_set_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static struct sdhci_ops sdhci_brcmstb_ops_7216 = {
+	.set_clock = sdhci_brcmstb_set_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = brcmstb_reset,
+	.set_uhs_signaling = sdhci_brcmstb_set_uhs_signaling,
+};
+
+static struct brcmstb_match_priv match_priv_7425 = {
+	.flags = BRCMSTB_MATCH_FLAGS_NO_64BIT |
+	BRCMSTB_MATCH_FLAGS_BROKEN_TIMEOUT,
+	.ops = &sdhci_brcmstb_ops,
+};
+
+static struct brcmstb_match_priv match_priv_7445 = {
+	.flags = BRCMSTB_MATCH_FLAGS_BROKEN_TIMEOUT,
+	.ops = &sdhci_brcmstb_ops,
+};
+
+static const struct brcmstb_match_priv match_priv_7216 = {
+	.flags = BRCMSTB_MATCH_FLAGS_HAS_CLOCK_GATE,
+	.hs400es = sdhci_brcmstb_hs400es,
+	.ops = &sdhci_brcmstb_ops_7216,
+};
+
+static const struct of_device_id sdhci_brcm_of_match[] = {
+	{ .compatible = "brcm,bcm7425-sdhci", .data = &match_priv_7425 },
+	{ .compatible = "brcm,bcm7445-sdhci", .data = &match_priv_7445 },
+	{ .compatible = "brcm,bcm7216-sdhci", .data = &match_priv_7216 },
+	{},
+};
+
+static u32 sdhci_brcmstb_cqhci_irq(struct sdhci_host *host, u32 intmask)
+{
+	int cmd_error = 0;
+	int data_error = 0;
+
+	if (!sdhci_cqe_irq(host, intmask, &cmd_error, &data_error))
+		return intmask;
+
+	cqhci_irq(host->mmc, intmask, cmd_error, data_error);
+
+	return 0;
+}
+
+static int sdhci_brcmstb_add_host(struct sdhci_host *host,
+				  struct sdhci_brcmstb_priv *priv)
+{
+	struct cqhci_host *cq_host;
+	bool dma64;
+	int ret;
+
+	if ((priv->flags & BRCMSTB_PRIV_FLAGS_HAS_CQE) == 0)
+		return sdhci_add_host(host);
+
+	dev_dbg(mmc_dev(host->mmc), "CQE is enabled\n");
+	host->mmc->caps2 |= MMC_CAP2_CQE | MMC_CAP2_CQE_DCMD;
+	ret = sdhci_setup_host(host);
+	if (ret)
+		return ret;
+
+	cq_host = devm_kzalloc(mmc_dev(host->mmc),
+			       sizeof(*cq_host), GFP_KERNEL);
+	if (!cq_host) {
+		ret = -ENOMEM;
+		goto cleanup;
+	}
+
+	cq_host->mmio = host->ioaddr + SDHCI_ARASAN_CQE_BASE_ADDR;
+	cq_host->ops = &sdhci_brcmstb_cqhci_ops;
+
+	dma64 = host->flags & SDHCI_USE_64_BIT_DMA;
+	if (dma64) {
+		dev_dbg(mmc_dev(host->mmc), "Using 64 bit DMA\n");
+		cq_host->caps |= CQHCI_TASK_DESC_SZ_128;
+	}
+
+	ret = cqhci_init(cq_host, host->mmc, dma64);
+	if (ret)
+		goto cleanup;
+
+	ret = __sdhci_add_host(host);
+	if (ret)
+		goto cleanup;
+
+	return 0;
+
+cleanup:
+	sdhci_cleanup_host(host);
+	return ret;
+}
+
+static int sdhci_brcmstb_probe(struct platform_device *pdev)
+{
+	const struct brcmstb_match_priv *match_priv;
+	struct sdhci_pltfm_data brcmstb_pdata;
+	struct sdhci_pltfm_host *pltfm_host;
+	const struct of_device_id *match;
+	struct sdhci_brcmstb_priv *priv;
+	struct sdhci_host *host;
+	struct resource *iomem;
+	struct clk *clk;
+	int res;
+
+	match = of_match_node(sdhci_brcm_of_match, pdev->dev.of_node);
+	match_priv = match->data;
+
+	dev_dbg(&pdev->dev, "Probe found match for %s\n",  match->compatible);
+
+	clk = devm_clk_get_optional(&pdev->dev, NULL);
+	if (IS_ERR(clk))
+		return dev_err_probe(&pdev->dev, PTR_ERR(clk),
+				     "Failed to get clock from Device Tree\n");
+
+	res = clk_prepare_enable(clk);
+	if (res)
+		return res;
+
+	memset(&brcmstb_pdata, 0, sizeof(brcmstb_pdata));
+	brcmstb_pdata.ops = match_priv->ops;
+	host = sdhci_pltfm_init(pdev, &brcmstb_pdata,
+				sizeof(struct sdhci_brcmstb_priv));
+	if (IS_ERR(host)) {
+		res = PTR_ERR(host);
+		goto err_clk;
+	}
+
+	pltfm_host = sdhci_priv(host);
+	priv = sdhci_pltfm_priv(pltfm_host);
+	if (device_property_read_bool(&pdev->dev, "supports-cqe")) {
+		priv->flags |= BRCMSTB_PRIV_FLAGS_HAS_CQE;
+		match_priv->ops->irq = sdhci_brcmstb_cqhci_irq;
+	}
+
+	/* Map in the non-standard CFG registers */
+	iomem = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+	priv->cfg_regs = devm_ioremap_resource(&pdev->dev, iomem);
+	if (IS_ERR(priv->cfg_regs)) {
+		res = PTR_ERR(priv->cfg_regs);
+		goto err;
+	}
+
+	sdhci_get_of_property(pdev);
+	res = mmc_of_parse(host->mmc);
+	if (res)
+		goto err;
+
+	/*
+	 * Automatic clock gating does not work for SD cards that may
+	 * voltage switch so only enable it for non-removable devices.
+	 */
+	if ((match_priv->flags & BRCMSTB_MATCH_FLAGS_HAS_CLOCK_GATE) &&
+	    (host->mmc->caps & MMC_CAP_NONREMOVABLE))
+		priv->flags |= BRCMSTB_PRIV_FLAGS_GATE_CLOCK;
+
+	/*
+	 * If the chip has enhanced strobe and it's enabled, add
+	 * callback
+	 */
+	if (match_priv->hs400es &&
+	    (host->mmc->caps2 & MMC_CAP2_HS400_ES))
+		host->mmc_host_ops.hs400_enhanced_strobe = match_priv->hs400es;
+
+	/*
+	 * Supply the existing CAPS, but clear the UHS modes. This
+	 * will allow these modes to be specified by device tree
+	 * properties through mmc_of_parse().
+	 */
+	host->caps = sdhci_readl(host, SDHCI_CAPABILITIES);
+	if (match_priv->flags & BRCMSTB_MATCH_FLAGS_NO_64BIT)
+		host->caps &= ~SDHCI_CAN_64BIT;
+	host->caps1 = sdhci_readl(host, SDHCI_CAPABILITIES_1);
+	host->caps1 &= ~(SDHCI_SUPPORT_SDR50 | SDHCI_SUPPORT_SDR104 |
+			 SDHCI_SUPPORT_DDR50);
+	host->quirks |= SDHCI_QUIRK_MISSING_CAPS;
+
+	if (match_priv->flags & BRCMSTB_MATCH_FLAGS_BROKEN_TIMEOUT)
+		host->quirks |= SDHCI_QUIRK_BROKEN_TIMEOUT_VAL;
+
+	res = sdhci_brcmstb_add_host(host, priv);
+	if (res)
+		goto err;
+
+	pltfm_host->clk = clk;
+	return res;
+
+err:
+	sdhci_pltfm_free(pdev);
+err_clk:
+	clk_disable_unprepare(clk);
+	return res;
+}
+
+static void sdhci_brcmstb_shutdown(struct platform_device *pdev)
+{
+	sdhci_pltfm_suspend(&pdev->dev);
+}
+
+MODULE_DEVICE_TABLE(of, sdhci_brcm_of_match);
+
+static struct platform_driver sdhci_brcmstb_driver = {
+	.driver		= {
+		.name	= "sdhci-brcmstb",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= &sdhci_pltfm_pmops,
+		.of_match_table = of_match_ptr(sdhci_brcm_of_match),
+	},
+	.probe		= sdhci_brcmstb_probe,
+	.remove		= sdhci_pltfm_unregister,
+	.shutdown	= sdhci_brcmstb_shutdown,
+};
+
+module_platform_driver(sdhci_brcmstb_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for Broadcom BRCMSTB SoCs");
+MODULE_AUTHOR("Broadcom");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-cadence.c b/drivers/mmc/host/sdhci-cadence.c
new file mode 100644
index 000000000..6f2de54a5
--- /dev/null
+++ b/drivers/mmc/host/sdhci-cadence.c
@@ -0,0 +1,477 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (C) 2016 Socionext Inc.
+ *   Author: Masahiro Yamada <yamada.masahiro@socionext.com>
+ */
+
+#include <linux/bitfield.h>
+#include <linux/bits.h>
+#include <linux/iopoll.h>
+#include <linux/module.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+
+#include "sdhci-pltfm.h"
+
+/* HRS - Host Register Set (specific to Cadence) */
+#define SDHCI_CDNS_HRS04		0x10		/* PHY access port */
+#define   SDHCI_CDNS_HRS04_ACK			BIT(26)
+#define   SDHCI_CDNS_HRS04_RD			BIT(25)
+#define   SDHCI_CDNS_HRS04_WR			BIT(24)
+#define   SDHCI_CDNS_HRS04_RDATA		GENMASK(23, 16)
+#define   SDHCI_CDNS_HRS04_WDATA		GENMASK(15, 8)
+#define   SDHCI_CDNS_HRS04_ADDR			GENMASK(5, 0)
+
+#define SDHCI_CDNS_HRS06		0x18		/* eMMC control */
+#define   SDHCI_CDNS_HRS06_TUNE_UP		BIT(15)
+#define   SDHCI_CDNS_HRS06_TUNE			GENMASK(13, 8)
+#define   SDHCI_CDNS_HRS06_MODE			GENMASK(2, 0)
+#define   SDHCI_CDNS_HRS06_MODE_SD		0x0
+#define   SDHCI_CDNS_HRS06_MODE_MMC_SDR		0x2
+#define   SDHCI_CDNS_HRS06_MODE_MMC_DDR		0x3
+#define   SDHCI_CDNS_HRS06_MODE_MMC_HS200	0x4
+#define   SDHCI_CDNS_HRS06_MODE_MMC_HS400	0x5
+#define   SDHCI_CDNS_HRS06_MODE_MMC_HS400ES	0x6
+
+/* SRS - Slot Register Set (SDHCI-compatible) */
+#define SDHCI_CDNS_SRS_BASE		0x200
+
+/* PHY */
+#define SDHCI_CDNS_PHY_DLY_SD_HS	0x00
+#define SDHCI_CDNS_PHY_DLY_SD_DEFAULT	0x01
+#define SDHCI_CDNS_PHY_DLY_UHS_SDR12	0x02
+#define SDHCI_CDNS_PHY_DLY_UHS_SDR25	0x03
+#define SDHCI_CDNS_PHY_DLY_UHS_SDR50	0x04
+#define SDHCI_CDNS_PHY_DLY_UHS_DDR50	0x05
+#define SDHCI_CDNS_PHY_DLY_EMMC_LEGACY	0x06
+#define SDHCI_CDNS_PHY_DLY_EMMC_SDR	0x07
+#define SDHCI_CDNS_PHY_DLY_EMMC_DDR	0x08
+#define SDHCI_CDNS_PHY_DLY_SDCLK	0x0b
+#define SDHCI_CDNS_PHY_DLY_HSMMC	0x0c
+#define SDHCI_CDNS_PHY_DLY_STROBE	0x0d
+
+/*
+ * The tuned val register is 6 bit-wide, but not the whole of the range is
+ * available.  The range 0-42 seems to be available (then 43 wraps around to 0)
+ * but I am not quite sure if it is official.  Use only 0 to 39 for safety.
+ */
+#define SDHCI_CDNS_MAX_TUNING_LOOP	40
+
+struct sdhci_cdns_phy_param {
+	u8 addr;
+	u8 data;
+};
+
+struct sdhci_cdns_priv {
+	void __iomem *hrs_addr;
+	bool enhanced_strobe;
+	unsigned int nr_phy_params;
+	struct sdhci_cdns_phy_param phy_params[];
+};
+
+struct sdhci_cdns_phy_cfg {
+	const char *property;
+	u8 addr;
+};
+
+static const struct sdhci_cdns_phy_cfg sdhci_cdns_phy_cfgs[] = {
+	{ "cdns,phy-input-delay-sd-highspeed", SDHCI_CDNS_PHY_DLY_SD_HS, },
+	{ "cdns,phy-input-delay-legacy", SDHCI_CDNS_PHY_DLY_SD_DEFAULT, },
+	{ "cdns,phy-input-delay-sd-uhs-sdr12", SDHCI_CDNS_PHY_DLY_UHS_SDR12, },
+	{ "cdns,phy-input-delay-sd-uhs-sdr25", SDHCI_CDNS_PHY_DLY_UHS_SDR25, },
+	{ "cdns,phy-input-delay-sd-uhs-sdr50", SDHCI_CDNS_PHY_DLY_UHS_SDR50, },
+	{ "cdns,phy-input-delay-sd-uhs-ddr50", SDHCI_CDNS_PHY_DLY_UHS_DDR50, },
+	{ "cdns,phy-input-delay-mmc-highspeed", SDHCI_CDNS_PHY_DLY_EMMC_SDR, },
+	{ "cdns,phy-input-delay-mmc-ddr", SDHCI_CDNS_PHY_DLY_EMMC_DDR, },
+	{ "cdns,phy-dll-delay-sdclk", SDHCI_CDNS_PHY_DLY_SDCLK, },
+	{ "cdns,phy-dll-delay-sdclk-hsmmc", SDHCI_CDNS_PHY_DLY_HSMMC, },
+	{ "cdns,phy-dll-delay-strobe", SDHCI_CDNS_PHY_DLY_STROBE, },
+};
+
+static int sdhci_cdns_write_phy_reg(struct sdhci_cdns_priv *priv,
+				    u8 addr, u8 data)
+{
+	void __iomem *reg = priv->hrs_addr + SDHCI_CDNS_HRS04;
+	u32 tmp;
+	int ret;
+
+	ret = readl_poll_timeout(reg, tmp, !(tmp & SDHCI_CDNS_HRS04_ACK),
+				 0, 10);
+	if (ret)
+		return ret;
+
+	tmp = FIELD_PREP(SDHCI_CDNS_HRS04_WDATA, data) |
+	      FIELD_PREP(SDHCI_CDNS_HRS04_ADDR, addr);
+	writel(tmp, reg);
+
+	tmp |= SDHCI_CDNS_HRS04_WR;
+	writel(tmp, reg);
+
+	ret = readl_poll_timeout(reg, tmp, tmp & SDHCI_CDNS_HRS04_ACK, 0, 10);
+	if (ret)
+		return ret;
+
+	tmp &= ~SDHCI_CDNS_HRS04_WR;
+	writel(tmp, reg);
+
+	ret = readl_poll_timeout(reg, tmp, !(tmp & SDHCI_CDNS_HRS04_ACK),
+				 0, 10);
+
+	return ret;
+}
+
+static unsigned int sdhci_cdns_phy_param_count(struct device_node *np)
+{
+	unsigned int count = 0;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(sdhci_cdns_phy_cfgs); i++)
+		if (of_property_read_bool(np, sdhci_cdns_phy_cfgs[i].property))
+			count++;
+
+	return count;
+}
+
+static void sdhci_cdns_phy_param_parse(struct device_node *np,
+				       struct sdhci_cdns_priv *priv)
+{
+	struct sdhci_cdns_phy_param *p = priv->phy_params;
+	u32 val;
+	int ret, i;
+
+	for (i = 0; i < ARRAY_SIZE(sdhci_cdns_phy_cfgs); i++) {
+		ret = of_property_read_u32(np, sdhci_cdns_phy_cfgs[i].property,
+					   &val);
+		if (ret)
+			continue;
+
+		p->addr = sdhci_cdns_phy_cfgs[i].addr;
+		p->data = val;
+		p++;
+	}
+}
+
+static int sdhci_cdns_phy_init(struct sdhci_cdns_priv *priv)
+{
+	int ret, i;
+
+	for (i = 0; i < priv->nr_phy_params; i++) {
+		ret = sdhci_cdns_write_phy_reg(priv, priv->phy_params[i].addr,
+					       priv->phy_params[i].data);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static void *sdhci_cdns_priv(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	return sdhci_pltfm_priv(pltfm_host);
+}
+
+static unsigned int sdhci_cdns_get_timeout_clock(struct sdhci_host *host)
+{
+	/*
+	 * Cadence's spec says the Timeout Clock Frequency is the same as the
+	 * Base Clock Frequency.
+	 */
+	return host->max_clk;
+}
+
+static void sdhci_cdns_set_emmc_mode(struct sdhci_cdns_priv *priv, u32 mode)
+{
+	u32 tmp;
+
+	/* The speed mode for eMMC is selected by HRS06 register */
+	tmp = readl(priv->hrs_addr + SDHCI_CDNS_HRS06);
+	tmp &= ~SDHCI_CDNS_HRS06_MODE;
+	tmp |= FIELD_PREP(SDHCI_CDNS_HRS06_MODE, mode);
+	writel(tmp, priv->hrs_addr + SDHCI_CDNS_HRS06);
+}
+
+static u32 sdhci_cdns_get_emmc_mode(struct sdhci_cdns_priv *priv)
+{
+	u32 tmp;
+
+	tmp = readl(priv->hrs_addr + SDHCI_CDNS_HRS06);
+	return FIELD_GET(SDHCI_CDNS_HRS06_MODE, tmp);
+}
+
+static int sdhci_cdns_set_tune_val(struct sdhci_host *host, unsigned int val)
+{
+	struct sdhci_cdns_priv *priv = sdhci_cdns_priv(host);
+	void __iomem *reg = priv->hrs_addr + SDHCI_CDNS_HRS06;
+	u32 tmp;
+	int i, ret;
+
+	if (WARN_ON(!FIELD_FIT(SDHCI_CDNS_HRS06_TUNE, val)))
+		return -EINVAL;
+
+	tmp = readl(reg);
+	tmp &= ~SDHCI_CDNS_HRS06_TUNE;
+	tmp |= FIELD_PREP(SDHCI_CDNS_HRS06_TUNE, val);
+
+	/*
+	 * Workaround for IP errata:
+	 * The IP6116 SD/eMMC PHY design has a timing issue on receive data
+	 * path. Send tune request twice.
+	 */
+	for (i = 0; i < 2; i++) {
+		tmp |= SDHCI_CDNS_HRS06_TUNE_UP;
+		writel(tmp, reg);
+
+		ret = readl_poll_timeout(reg, tmp,
+					 !(tmp & SDHCI_CDNS_HRS06_TUNE_UP),
+					 0, 1);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+/*
+ * In SD mode, software must not use the hardware tuning and instead perform
+ * an almost identical procedure to eMMC.
+ */
+static int sdhci_cdns_execute_tuning(struct sdhci_host *host, u32 opcode)
+{
+	int cur_streak = 0;
+	int max_streak = 0;
+	int end_of_streak = 0;
+	int i;
+
+	/*
+	 * Do not execute tuning for UHS_SDR50 or UHS_DDR50.
+	 * The delay is set by probe, based on the DT properties.
+	 */
+	if (host->timing != MMC_TIMING_MMC_HS200 &&
+	    host->timing != MMC_TIMING_UHS_SDR104)
+		return 0;
+
+	for (i = 0; i < SDHCI_CDNS_MAX_TUNING_LOOP; i++) {
+		if (sdhci_cdns_set_tune_val(host, i) ||
+		    mmc_send_tuning(host->mmc, opcode, NULL)) { /* bad */
+			cur_streak = 0;
+		} else { /* good */
+			cur_streak++;
+			if (cur_streak > max_streak) {
+				max_streak = cur_streak;
+				end_of_streak = i;
+			}
+		}
+	}
+
+	if (!max_streak) {
+		dev_err(mmc_dev(host->mmc), "no tuning point found\n");
+		return -EIO;
+	}
+
+	return sdhci_cdns_set_tune_val(host, end_of_streak - max_streak / 2);
+}
+
+static void sdhci_cdns_set_uhs_signaling(struct sdhci_host *host,
+					 unsigned int timing)
+{
+	struct sdhci_cdns_priv *priv = sdhci_cdns_priv(host);
+	u32 mode;
+
+	switch (timing) {
+	case MMC_TIMING_MMC_HS:
+		mode = SDHCI_CDNS_HRS06_MODE_MMC_SDR;
+		break;
+	case MMC_TIMING_MMC_DDR52:
+		mode = SDHCI_CDNS_HRS06_MODE_MMC_DDR;
+		break;
+	case MMC_TIMING_MMC_HS200:
+		mode = SDHCI_CDNS_HRS06_MODE_MMC_HS200;
+		break;
+	case MMC_TIMING_MMC_HS400:
+		if (priv->enhanced_strobe)
+			mode = SDHCI_CDNS_HRS06_MODE_MMC_HS400ES;
+		else
+			mode = SDHCI_CDNS_HRS06_MODE_MMC_HS400;
+		break;
+	default:
+		mode = SDHCI_CDNS_HRS06_MODE_SD;
+		break;
+	}
+
+	sdhci_cdns_set_emmc_mode(priv, mode);
+
+	/* For SD, fall back to the default handler */
+	if (mode == SDHCI_CDNS_HRS06_MODE_SD)
+		sdhci_set_uhs_signaling(host, timing);
+}
+
+static const struct sdhci_ops sdhci_cdns_ops = {
+	.set_clock = sdhci_set_clock,
+	.get_timeout_clock = sdhci_cdns_get_timeout_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.platform_execute_tuning = sdhci_cdns_execute_tuning,
+	.set_uhs_signaling = sdhci_cdns_set_uhs_signaling,
+};
+
+static const struct sdhci_pltfm_data sdhci_cdns_uniphier_pltfm_data = {
+	.ops = &sdhci_cdns_ops,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+};
+
+static const struct sdhci_pltfm_data sdhci_cdns_pltfm_data = {
+	.ops = &sdhci_cdns_ops,
+};
+
+static void sdhci_cdns_hs400_enhanced_strobe(struct mmc_host *mmc,
+					     struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_cdns_priv *priv = sdhci_cdns_priv(host);
+	u32 mode;
+
+	priv->enhanced_strobe = ios->enhanced_strobe;
+
+	mode = sdhci_cdns_get_emmc_mode(priv);
+
+	if (mode == SDHCI_CDNS_HRS06_MODE_MMC_HS400 && ios->enhanced_strobe)
+		sdhci_cdns_set_emmc_mode(priv,
+					 SDHCI_CDNS_HRS06_MODE_MMC_HS400ES);
+
+	if (mode == SDHCI_CDNS_HRS06_MODE_MMC_HS400ES && !ios->enhanced_strobe)
+		sdhci_cdns_set_emmc_mode(priv,
+					 SDHCI_CDNS_HRS06_MODE_MMC_HS400);
+}
+
+static int sdhci_cdns_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	const struct sdhci_pltfm_data *data;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_cdns_priv *priv;
+	struct clk *clk;
+	unsigned int nr_phy_params;
+	int ret;
+	struct device *dev = &pdev->dev;
+	static const u16 version = SDHCI_SPEC_400 << SDHCI_SPEC_VER_SHIFT;
+
+	clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(clk))
+		return PTR_ERR(clk);
+
+	ret = clk_prepare_enable(clk);
+	if (ret)
+		return ret;
+
+	data = of_device_get_match_data(dev);
+	if (!data)
+		data = &sdhci_cdns_pltfm_data;
+
+	nr_phy_params = sdhci_cdns_phy_param_count(dev->of_node);
+	host = sdhci_pltfm_init(pdev, data,
+				struct_size(priv, phy_params, nr_phy_params));
+	if (IS_ERR(host)) {
+		ret = PTR_ERR(host);
+		goto disable_clk;
+	}
+
+	pltfm_host = sdhci_priv(host);
+	pltfm_host->clk = clk;
+
+	priv = sdhci_pltfm_priv(pltfm_host);
+	priv->nr_phy_params = nr_phy_params;
+	priv->hrs_addr = host->ioaddr;
+	priv->enhanced_strobe = false;
+	host->ioaddr += SDHCI_CDNS_SRS_BASE;
+	host->mmc_host_ops.hs400_enhanced_strobe =
+				sdhci_cdns_hs400_enhanced_strobe;
+	sdhci_enable_v4_mode(host);
+	__sdhci_read_caps(host, &version, NULL, NULL);
+
+	sdhci_get_of_property(pdev);
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto free;
+
+	sdhci_cdns_phy_param_parse(dev->of_node, priv);
+
+	ret = sdhci_cdns_phy_init(priv);
+	if (ret)
+		goto free;
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto free;
+
+	return 0;
+free:
+	sdhci_pltfm_free(pdev);
+disable_clk:
+	clk_disable_unprepare(clk);
+
+	return ret;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int sdhci_cdns_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_cdns_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	ret = clk_prepare_enable(pltfm_host->clk);
+	if (ret)
+		return ret;
+
+	ret = sdhci_cdns_phy_init(priv);
+	if (ret)
+		goto disable_clk;
+
+	ret = sdhci_resume_host(host);
+	if (ret)
+		goto disable_clk;
+
+	return 0;
+
+disable_clk:
+	clk_disable_unprepare(pltfm_host->clk);
+
+	return ret;
+}
+#endif
+
+static const struct dev_pm_ops sdhci_cdns_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(sdhci_pltfm_suspend, sdhci_cdns_resume)
+};
+
+static const struct of_device_id sdhci_cdns_match[] = {
+	{
+		.compatible = "socionext,uniphier-sd4hc",
+		.data = &sdhci_cdns_uniphier_pltfm_data,
+	},
+	{ .compatible = "cdns,sd4hc" },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, sdhci_cdns_match);
+
+static struct platform_driver sdhci_cdns_driver = {
+	.driver = {
+		.name = "sdhci-cdns",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm = &sdhci_cdns_pm_ops,
+		.of_match_table = sdhci_cdns_match,
+	},
+	.probe = sdhci_cdns_probe,
+	.remove = sdhci_pltfm_unregister,
+};
+module_platform_driver(sdhci_cdns_driver);
+
+MODULE_AUTHOR("Masahiro Yamada <yamada.masahiro@socionext.com>");
+MODULE_DESCRIPTION("Cadence SD/SDIO/eMMC Host Controller Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mmc/host/sdhci-cns3xxx.c b/drivers/mmc/host/sdhci-cns3xxx.c
new file mode 100644
index 000000000..2a29c7a4f
--- /dev/null
+++ b/drivers/mmc/host/sdhci-cns3xxx.c
@@ -0,0 +1,113 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * SDHCI support for CNS3xxx SoC
+ *
+ * Copyright 2008 Cavium Networks
+ * Copyright 2010 MontaVista Software, LLC.
+ *
+ * Authors: Scott Shu
+ *	    Anton Vorontsov <avorontsov@mvista.com>
+ */
+
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/mmc/host.h>
+#include <linux/module.h>
+#include "sdhci-pltfm.h"
+
+static unsigned int sdhci_cns3xxx_get_max_clk(struct sdhci_host *host)
+{
+	return 150000000;
+}
+
+static void sdhci_cns3xxx_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	struct device *dev = mmc_dev(host->mmc);
+	int div = 1;
+	u16 clk;
+	unsigned long timeout;
+
+	host->mmc->actual_clock = 0;
+
+	sdhci_writew(host, 0, SDHCI_CLOCK_CONTROL);
+
+	if (clock == 0)
+		return;
+
+	while (host->max_clk / div > clock) {
+		/*
+		 * On CNS3xxx divider grows linearly up to 4, and then
+		 * exponentially up to 256.
+		 */
+		if (div < 4)
+			div += 1;
+		else if (div < 256)
+			div *= 2;
+		else
+			break;
+	}
+
+	dev_dbg(dev, "desired SD clock: %d, actual: %d\n",
+		clock, host->max_clk / div);
+
+	/* Divide by 3 is special. */
+	if (div != 3)
+		div >>= 1;
+
+	clk = div << SDHCI_DIVIDER_SHIFT;
+	clk |= SDHCI_CLOCK_INT_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+	timeout = 20;
+	while (!((clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL))
+			& SDHCI_CLOCK_INT_STABLE)) {
+		if (timeout == 0) {
+			dev_warn(dev, "clock is unstable");
+			break;
+		}
+		timeout--;
+		mdelay(1);
+	}
+
+	clk |= SDHCI_CLOCK_CARD_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+}
+
+static const struct sdhci_ops sdhci_cns3xxx_ops = {
+	.get_max_clock	= sdhci_cns3xxx_get_max_clk,
+	.set_clock	= sdhci_cns3xxx_set_clock,
+	.set_bus_width	= sdhci_set_bus_width,
+	.reset          = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static const struct sdhci_pltfm_data sdhci_cns3xxx_pdata = {
+	.ops = &sdhci_cns3xxx_ops,
+	.quirks = SDHCI_QUIRK_BROKEN_DMA |
+		  SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK |
+		  SDHCI_QUIRK_INVERTED_WRITE_PROTECT |
+		  SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN |
+		  SDHCI_QUIRK_BROKEN_TIMEOUT_VAL,
+};
+
+static int sdhci_cns3xxx_probe(struct platform_device *pdev)
+{
+	return sdhci_pltfm_register(pdev, &sdhci_cns3xxx_pdata, 0);
+}
+
+static struct platform_driver sdhci_cns3xxx_driver = {
+	.driver		= {
+		.name	= "sdhci-cns3xxx",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= &sdhci_pltfm_pmops,
+	},
+	.probe		= sdhci_cns3xxx_probe,
+	.remove		= sdhci_pltfm_unregister,
+};
+
+module_platform_driver(sdhci_cns3xxx_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for CNS3xxx");
+MODULE_AUTHOR("Scott Shu, "
+	      "Anton Vorontsov <avorontsov@mvista.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-cqhci.h b/drivers/mmc/host/sdhci-cqhci.h
new file mode 100644
index 000000000..cf8e7ba71
--- /dev/null
+++ b/drivers/mmc/host/sdhci-cqhci.h
@@ -0,0 +1,24 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright 2022 The Chromium OS Authors
+ *
+ * Support that applies to the combination of SDHCI and CQHCI, while not
+ * expressing a dependency between the two modules.
+ */
+
+#ifndef __MMC_HOST_SDHCI_CQHCI_H__
+#define __MMC_HOST_SDHCI_CQHCI_H__
+
+#include "cqhci.h"
+#include "sdhci.h"
+
+static inline void sdhci_and_cqhci_reset(struct sdhci_host *host, u8 mask)
+{
+	if ((host->mmc->caps2 & MMC_CAP2_CQE) && (mask & SDHCI_RESET_ALL) &&
+	    host->mmc->cqe_private)
+		cqhci_deactivate(host->mmc);
+
+	sdhci_reset(host, mask);
+}
+
+#endif /* __MMC_HOST_SDHCI_CQHCI_H__ */
diff --git a/drivers/mmc/host/sdhci-dove.c b/drivers/mmc/host/sdhci-dove.c
new file mode 100644
index 000000000..5e5bf82e5
--- /dev/null
+++ b/drivers/mmc/host/sdhci-dove.c
@@ -0,0 +1,121 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * sdhci-dove.c Support for SDHCI on Marvell's Dove SoC
+ *
+ * Author: Saeed Bishara <saeed@marvell.com>
+ *	   Mike Rapoport <mike@compulab.co.il>
+ * Based on sdhci-cns3xxx.c
+ */
+
+#include <linux/clk.h>
+#include <linux/err.h>
+#include <linux/io.h>
+#include <linux/mmc/host.h>
+#include <linux/module.h>
+#include <linux/of.h>
+
+#include "sdhci-pltfm.h"
+
+static u16 sdhci_dove_readw(struct sdhci_host *host, int reg)
+{
+	u16 ret;
+
+	switch (reg) {
+	case SDHCI_HOST_VERSION:
+	case SDHCI_SLOT_INT_STATUS:
+		/* those registers don't exist */
+		return 0;
+	default:
+		ret = readw(host->ioaddr + reg);
+	}
+	return ret;
+}
+
+static u32 sdhci_dove_readl(struct sdhci_host *host, int reg)
+{
+	u32 ret;
+
+	ret = readl(host->ioaddr + reg);
+
+	switch (reg) {
+	case SDHCI_CAPABILITIES:
+		/* Mask the support for 3.0V */
+		ret &= ~SDHCI_CAN_VDD_300;
+		break;
+	}
+	return ret;
+}
+
+static const struct sdhci_ops sdhci_dove_ops = {
+	.read_w	= sdhci_dove_readw,
+	.read_l	= sdhci_dove_readl,
+	.set_clock = sdhci_set_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static const struct sdhci_pltfm_data sdhci_dove_pdata = {
+	.ops	= &sdhci_dove_ops,
+	.quirks	= SDHCI_QUIRK_NO_SIMULT_VDD_AND_POWER |
+		  SDHCI_QUIRK_NO_BUSY_IRQ |
+		  SDHCI_QUIRK_BROKEN_TIMEOUT_VAL |
+		  SDHCI_QUIRK_FORCE_DMA |
+		  SDHCI_QUIRK_NO_HISPD_BIT,
+};
+
+static int sdhci_dove_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct sdhci_pltfm_host *pltfm_host;
+	int ret;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_dove_pdata, 0);
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	pltfm_host->clk = devm_clk_get(&pdev->dev, NULL);
+
+	if (!IS_ERR(pltfm_host->clk))
+		clk_prepare_enable(pltfm_host->clk);
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto err_sdhci_add;
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err_sdhci_add;
+
+	return 0;
+
+err_sdhci_add:
+	clk_disable_unprepare(pltfm_host->clk);
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static const struct of_device_id sdhci_dove_of_match_table[] = {
+	{ .compatible = "marvell,dove-sdhci", },
+	{}
+};
+MODULE_DEVICE_TABLE(of, sdhci_dove_of_match_table);
+
+static struct platform_driver sdhci_dove_driver = {
+	.driver		= {
+		.name	= "sdhci-dove",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= &sdhci_pltfm_pmops,
+		.of_match_table = sdhci_dove_of_match_table,
+	},
+	.probe		= sdhci_dove_probe,
+	.remove		= sdhci_pltfm_unregister,
+};
+
+module_platform_driver(sdhci_dove_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for Dove");
+MODULE_AUTHOR("Saeed Bishara <saeed@marvell.com>, "
+	      "Mike Rapoport <mike@compulab.co.il>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-esdhc-imx.c b/drivers/mmc/host/sdhci-esdhc-imx.c
new file mode 100644
index 000000000..b030f657e
--- /dev/null
+++ b/drivers/mmc/host/sdhci-esdhc-imx.c
@@ -0,0 +1,1881 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Freescale eSDHC i.MX controller driver for the platform bus.
+ *
+ * derived from the OF-version.
+ *
+ * Copyright (c) 2010 Pengutronix e.K.
+ *   Author: Wolfram Sang <kernel@pengutronix.de>
+ */
+
+#include <linux/bitfield.h>
+#include <linux/io.h>
+#include <linux/iopoll.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/clk.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/pm_qos.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sdio.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/platform_data/mmc-esdhc-imx.h>
+#include <linux/pm_runtime.h>
+#include "sdhci-cqhci.h"
+#include "sdhci-pltfm.h"
+#include "sdhci-esdhc.h"
+#include "cqhci.h"
+
+#define ESDHC_SYS_CTRL_DTOCV_MASK	0x0f
+#define	ESDHC_CTRL_D3CD			0x08
+#define ESDHC_BURST_LEN_EN_INCR		(1 << 27)
+/* VENDOR SPEC register */
+#define ESDHC_VENDOR_SPEC		0xc0
+#define  ESDHC_VENDOR_SPEC_SDIO_QUIRK	(1 << 1)
+#define  ESDHC_VENDOR_SPEC_VSELECT	(1 << 1)
+#define  ESDHC_VENDOR_SPEC_FRC_SDCLK_ON	(1 << 8)
+#define ESDHC_DEBUG_SEL_AND_STATUS_REG		0xc2
+#define ESDHC_DEBUG_SEL_REG			0xc3
+#define ESDHC_DEBUG_SEL_MASK			0xf
+#define ESDHC_DEBUG_SEL_CMD_STATE		1
+#define ESDHC_DEBUG_SEL_DATA_STATE		2
+#define ESDHC_DEBUG_SEL_TRANS_STATE		3
+#define ESDHC_DEBUG_SEL_DMA_STATE		4
+#define ESDHC_DEBUG_SEL_ADMA_STATE		5
+#define ESDHC_DEBUG_SEL_FIFO_STATE		6
+#define ESDHC_DEBUG_SEL_ASYNC_FIFO_STATE	7
+#define ESDHC_WTMK_LVL			0x44
+#define  ESDHC_WTMK_DEFAULT_VAL		0x10401040
+#define  ESDHC_WTMK_LVL_RD_WML_MASK	0x000000FF
+#define  ESDHC_WTMK_LVL_RD_WML_SHIFT	0
+#define  ESDHC_WTMK_LVL_WR_WML_MASK	0x00FF0000
+#define  ESDHC_WTMK_LVL_WR_WML_SHIFT	16
+#define  ESDHC_WTMK_LVL_WML_VAL_DEF	64
+#define  ESDHC_WTMK_LVL_WML_VAL_MAX	128
+#define ESDHC_MIX_CTRL			0x48
+#define  ESDHC_MIX_CTRL_DDREN		(1 << 3)
+#define  ESDHC_MIX_CTRL_AC23EN		(1 << 7)
+#define  ESDHC_MIX_CTRL_EXE_TUNE	(1 << 22)
+#define  ESDHC_MIX_CTRL_SMPCLK_SEL	(1 << 23)
+#define  ESDHC_MIX_CTRL_AUTO_TUNE_EN	(1 << 24)
+#define  ESDHC_MIX_CTRL_FBCLK_SEL	(1 << 25)
+#define  ESDHC_MIX_CTRL_HS400_EN	(1 << 26)
+#define  ESDHC_MIX_CTRL_HS400_ES_EN	(1 << 27)
+/* Bits 3 and 6 are not SDHCI standard definitions */
+#define  ESDHC_MIX_CTRL_SDHCI_MASK	0xb7
+/* Tuning bits */
+#define  ESDHC_MIX_CTRL_TUNING_MASK	0x03c00000
+
+/* dll control register */
+#define ESDHC_DLL_CTRL			0x60
+#define ESDHC_DLL_OVERRIDE_VAL_SHIFT	9
+#define ESDHC_DLL_OVERRIDE_EN_SHIFT	8
+
+/* tune control register */
+#define ESDHC_TUNE_CTRL_STATUS		0x68
+#define  ESDHC_TUNE_CTRL_STEP		1
+#define  ESDHC_TUNE_CTRL_MIN		0
+#define  ESDHC_TUNE_CTRL_MAX		((1 << 7) - 1)
+
+/* strobe dll register */
+#define ESDHC_STROBE_DLL_CTRL		0x70
+#define ESDHC_STROBE_DLL_CTRL_ENABLE	(1 << 0)
+#define ESDHC_STROBE_DLL_CTRL_RESET	(1 << 1)
+#define ESDHC_STROBE_DLL_CTRL_SLV_DLY_TARGET_DEFAULT	0x7
+#define ESDHC_STROBE_DLL_CTRL_SLV_DLY_TARGET_SHIFT	3
+#define ESDHC_STROBE_DLL_CTRL_SLV_UPDATE_INT_DEFAULT	(4 << 20)
+
+#define ESDHC_STROBE_DLL_STATUS		0x74
+#define ESDHC_STROBE_DLL_STS_REF_LOCK	(1 << 1)
+#define ESDHC_STROBE_DLL_STS_SLV_LOCK	0x1
+
+#define ESDHC_VEND_SPEC2		0xc8
+#define ESDHC_VEND_SPEC2_EN_BUSY_IRQ	(1 << 8)
+
+#define ESDHC_TUNING_CTRL		0xcc
+#define ESDHC_STD_TUNING_EN		(1 << 24)
+/* NOTE: the minimum valid tuning start tap for mx6sl is 1 */
+#define ESDHC_TUNING_START_TAP_DEFAULT	0x1
+#define ESDHC_TUNING_START_TAP_MASK	0x7f
+#define ESDHC_TUNING_CMD_CRC_CHECK_DISABLE	(1 << 7)
+#define ESDHC_TUNING_STEP_DEFAULT	0x1
+#define ESDHC_TUNING_STEP_MASK		0x00070000
+#define ESDHC_TUNING_STEP_SHIFT		16
+
+/* pinctrl state */
+#define ESDHC_PINCTRL_STATE_100MHZ	"state_100mhz"
+#define ESDHC_PINCTRL_STATE_200MHZ	"state_200mhz"
+
+/*
+ * Our interpretation of the SDHCI_HOST_CONTROL register
+ */
+#define ESDHC_CTRL_4BITBUS		(0x1 << 1)
+#define ESDHC_CTRL_8BITBUS		(0x2 << 1)
+#define ESDHC_CTRL_BUSWIDTH_MASK	(0x3 << 1)
+
+/*
+ * There is an INT DMA ERR mismatch between eSDHC and STD SDHC SPEC:
+ * Bit25 is used in STD SPEC, and is reserved in fsl eSDHC design,
+ * but bit28 is used as the INT DMA ERR in fsl eSDHC design.
+ * Define this macro DMA error INT for fsl eSDHC
+ */
+#define ESDHC_INT_VENDOR_SPEC_DMA_ERR	(1 << 28)
+
+/* the address offset of CQHCI */
+#define ESDHC_CQHCI_ADDR_OFFSET		0x100
+
+/*
+ * The CMDTYPE of the CMD register (offset 0xE) should be set to
+ * "11" when the STOP CMD12 is issued on imx53 to abort one
+ * open ended multi-blk IO. Otherwise the TC INT wouldn't
+ * be generated.
+ * In exact block transfer, the controller doesn't complete the
+ * operations automatically as required at the end of the
+ * transfer and remains on hold if the abort command is not sent.
+ * As a result, the TC flag is not asserted and SW received timeout
+ * exception. Bit1 of Vendor Spec register is used to fix it.
+ */
+#define ESDHC_FLAG_MULTIBLK_NO_INT	BIT(1)
+/*
+ * The flag tells that the ESDHC controller is an USDHC block that is
+ * integrated on the i.MX6 series.
+ */
+#define ESDHC_FLAG_USDHC		BIT(3)
+/* The IP supports manual tuning process */
+#define ESDHC_FLAG_MAN_TUNING		BIT(4)
+/* The IP supports standard tuning process */
+#define ESDHC_FLAG_STD_TUNING		BIT(5)
+/* The IP has SDHCI_CAPABILITIES_1 register */
+#define ESDHC_FLAG_HAVE_CAP1		BIT(6)
+/*
+ * The IP has erratum ERR004536
+ * uSDHC: ADMA Length Mismatch Error occurs if the AHB read access is slow,
+ * when reading data from the card
+ * This flag is also set for i.MX25 and i.MX35 in order to get
+ * SDHCI_QUIRK_BROKEN_ADMA, but for different reasons (ADMA capability bits).
+ */
+#define ESDHC_FLAG_ERR004536		BIT(7)
+/* The IP supports HS200 mode */
+#define ESDHC_FLAG_HS200		BIT(8)
+/* The IP supports HS400 mode */
+#define ESDHC_FLAG_HS400		BIT(9)
+/*
+ * The IP has errata ERR010450
+ * uSDHC: At 1.8V due to the I/O timing limit, for SDR mode, SD card
+ * clock can't exceed 150MHz, for DDR mode, SD card clock can't exceed 45MHz.
+ */
+#define ESDHC_FLAG_ERR010450		BIT(10)
+/* The IP supports HS400ES mode */
+#define ESDHC_FLAG_HS400_ES		BIT(11)
+/* The IP has Host Controller Interface for Command Queuing */
+#define ESDHC_FLAG_CQHCI		BIT(12)
+/* need request pmqos during low power */
+#define ESDHC_FLAG_PMQOS		BIT(13)
+/* The IP state got lost in low power mode */
+#define ESDHC_FLAG_STATE_LOST_IN_LPMODE		BIT(14)
+/* The IP lost clock rate in PM_RUNTIME */
+#define ESDHC_FLAG_CLK_RATE_LOST_IN_PM_RUNTIME	BIT(15)
+/*
+ * The IP do not support the ACMD23 feature completely when use ADMA mode.
+ * In ADMA mode, it only use the 16 bit block count of the register 0x4
+ * (BLOCK_ATT) as the CMD23's argument for ACMD23 mode, which means it will
+ * ignore the upper 16 bit of the CMD23's argument. This will block the reliable
+ * write operation in RPMB, because RPMB reliable write need to set the bit31
+ * of the CMD23's argument.
+ * imx6qpdl/imx6sx/imx6sl/imx7d has this limitation only for ADMA mode, SDMA
+ * do not has this limitation. so when these SoC use ADMA mode, it need to
+ * disable the ACMD23 feature.
+ */
+#define ESDHC_FLAG_BROKEN_AUTO_CMD23	BIT(16)
+
+struct esdhc_soc_data {
+	u32 flags;
+};
+
+static const struct esdhc_soc_data esdhc_imx25_data = {
+	.flags = ESDHC_FLAG_ERR004536,
+};
+
+static const struct esdhc_soc_data esdhc_imx35_data = {
+	.flags = ESDHC_FLAG_ERR004536,
+};
+
+static const struct esdhc_soc_data esdhc_imx51_data = {
+	.flags = 0,
+};
+
+static const struct esdhc_soc_data esdhc_imx53_data = {
+	.flags = ESDHC_FLAG_MULTIBLK_NO_INT,
+};
+
+static const struct esdhc_soc_data usdhc_imx6q_data = {
+	.flags = ESDHC_FLAG_USDHC | ESDHC_FLAG_MAN_TUNING
+			| ESDHC_FLAG_BROKEN_AUTO_CMD23,
+};
+
+static const struct esdhc_soc_data usdhc_imx6sl_data = {
+	.flags = ESDHC_FLAG_USDHC | ESDHC_FLAG_STD_TUNING
+			| ESDHC_FLAG_HAVE_CAP1 | ESDHC_FLAG_ERR004536
+			| ESDHC_FLAG_HS200
+			| ESDHC_FLAG_BROKEN_AUTO_CMD23,
+};
+
+static const struct esdhc_soc_data usdhc_imx6sll_data = {
+	.flags = ESDHC_FLAG_USDHC | ESDHC_FLAG_STD_TUNING
+			| ESDHC_FLAG_HAVE_CAP1 | ESDHC_FLAG_HS200
+			| ESDHC_FLAG_HS400
+			| ESDHC_FLAG_STATE_LOST_IN_LPMODE,
+};
+
+static const struct esdhc_soc_data usdhc_imx6sx_data = {
+	.flags = ESDHC_FLAG_USDHC | ESDHC_FLAG_STD_TUNING
+			| ESDHC_FLAG_HAVE_CAP1 | ESDHC_FLAG_HS200
+			| ESDHC_FLAG_STATE_LOST_IN_LPMODE
+			| ESDHC_FLAG_BROKEN_AUTO_CMD23,
+};
+
+static const struct esdhc_soc_data usdhc_imx6ull_data = {
+	.flags = ESDHC_FLAG_USDHC | ESDHC_FLAG_STD_TUNING
+			| ESDHC_FLAG_HAVE_CAP1 | ESDHC_FLAG_HS200
+			| ESDHC_FLAG_ERR010450
+			| ESDHC_FLAG_STATE_LOST_IN_LPMODE,
+};
+
+static const struct esdhc_soc_data usdhc_imx7d_data = {
+	.flags = ESDHC_FLAG_USDHC | ESDHC_FLAG_STD_TUNING
+			| ESDHC_FLAG_HAVE_CAP1 | ESDHC_FLAG_HS200
+			| ESDHC_FLAG_HS400
+			| ESDHC_FLAG_STATE_LOST_IN_LPMODE
+			| ESDHC_FLAG_BROKEN_AUTO_CMD23,
+};
+
+static struct esdhc_soc_data usdhc_imx7ulp_data = {
+	.flags = ESDHC_FLAG_USDHC | ESDHC_FLAG_STD_TUNING
+			| ESDHC_FLAG_HAVE_CAP1 | ESDHC_FLAG_HS200
+			| ESDHC_FLAG_PMQOS | ESDHC_FLAG_HS400
+			| ESDHC_FLAG_STATE_LOST_IN_LPMODE,
+};
+
+static struct esdhc_soc_data usdhc_imx8qxp_data = {
+	.flags = ESDHC_FLAG_USDHC | ESDHC_FLAG_STD_TUNING
+			| ESDHC_FLAG_HAVE_CAP1 | ESDHC_FLAG_HS200
+			| ESDHC_FLAG_HS400 | ESDHC_FLAG_HS400_ES
+			| ESDHC_FLAG_STATE_LOST_IN_LPMODE
+			| ESDHC_FLAG_CLK_RATE_LOST_IN_PM_RUNTIME,
+};
+
+static struct esdhc_soc_data usdhc_imx8mm_data = {
+	.flags = ESDHC_FLAG_USDHC | ESDHC_FLAG_STD_TUNING
+			| ESDHC_FLAG_HAVE_CAP1 | ESDHC_FLAG_HS200
+			| ESDHC_FLAG_HS400 | ESDHC_FLAG_HS400_ES
+			| ESDHC_FLAG_STATE_LOST_IN_LPMODE,
+};
+
+struct pltfm_imx_data {
+	u32 scratchpad;
+	struct pinctrl *pinctrl;
+	struct pinctrl_state *pins_100mhz;
+	struct pinctrl_state *pins_200mhz;
+	const struct esdhc_soc_data *socdata;
+	struct esdhc_platform_data boarddata;
+	struct clk *clk_ipg;
+	struct clk *clk_ahb;
+	struct clk *clk_per;
+	unsigned int actual_clock;
+	enum {
+		NO_CMD_PENDING,      /* no multiblock command pending */
+		MULTIBLK_IN_PROCESS, /* exact multiblock cmd in process */
+		WAIT_FOR_INT,        /* sent CMD12, waiting for response INT */
+	} multiblock_status;
+	u32 is_ddr;
+	struct pm_qos_request pm_qos_req;
+};
+
+static const struct of_device_id imx_esdhc_dt_ids[] = {
+	{ .compatible = "fsl,imx25-esdhc", .data = &esdhc_imx25_data, },
+	{ .compatible = "fsl,imx35-esdhc", .data = &esdhc_imx35_data, },
+	{ .compatible = "fsl,imx51-esdhc", .data = &esdhc_imx51_data, },
+	{ .compatible = "fsl,imx53-esdhc", .data = &esdhc_imx53_data, },
+	{ .compatible = "fsl,imx6sx-usdhc", .data = &usdhc_imx6sx_data, },
+	{ .compatible = "fsl,imx6sl-usdhc", .data = &usdhc_imx6sl_data, },
+	{ .compatible = "fsl,imx6sll-usdhc", .data = &usdhc_imx6sll_data, },
+	{ .compatible = "fsl,imx6q-usdhc", .data = &usdhc_imx6q_data, },
+	{ .compatible = "fsl,imx6ull-usdhc", .data = &usdhc_imx6ull_data, },
+	{ .compatible = "fsl,imx7d-usdhc", .data = &usdhc_imx7d_data, },
+	{ .compatible = "fsl,imx7ulp-usdhc", .data = &usdhc_imx7ulp_data, },
+	{ .compatible = "fsl,imx8qxp-usdhc", .data = &usdhc_imx8qxp_data, },
+	{ .compatible = "fsl,imx8mm-usdhc", .data = &usdhc_imx8mm_data, },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, imx_esdhc_dt_ids);
+
+static inline int is_imx25_esdhc(struct pltfm_imx_data *data)
+{
+	return data->socdata == &esdhc_imx25_data;
+}
+
+static inline int is_imx53_esdhc(struct pltfm_imx_data *data)
+{
+	return data->socdata == &esdhc_imx53_data;
+}
+
+static inline int is_imx6q_usdhc(struct pltfm_imx_data *data)
+{
+	return data->socdata == &usdhc_imx6q_data;
+}
+
+static inline int esdhc_is_usdhc(struct pltfm_imx_data *data)
+{
+	return !!(data->socdata->flags & ESDHC_FLAG_USDHC);
+}
+
+static inline void esdhc_clrset_le(struct sdhci_host *host, u32 mask, u32 val, int reg)
+{
+	void __iomem *base = host->ioaddr + (reg & ~0x3);
+	u32 shift = (reg & 0x3) * 8;
+
+	writel(((readl(base) & ~(mask << shift)) | (val << shift)), base);
+}
+
+#define DRIVER_NAME "sdhci-esdhc-imx"
+#define ESDHC_IMX_DUMP(f, x...) \
+	pr_err("%s: " DRIVER_NAME ": " f, mmc_hostname(host->mmc), ## x)
+static void esdhc_dump_debug_regs(struct sdhci_host *host)
+{
+	int i;
+	char *debug_status[7] = {
+				 "cmd debug status",
+				 "data debug status",
+				 "trans debug status",
+				 "dma debug status",
+				 "adma debug status",
+				 "fifo debug status",
+				 "async fifo debug status"
+	};
+
+	ESDHC_IMX_DUMP("========= ESDHC IMX DEBUG STATUS DUMP =========\n");
+	for (i = 0; i < 7; i++) {
+		esdhc_clrset_le(host, ESDHC_DEBUG_SEL_MASK,
+			ESDHC_DEBUG_SEL_CMD_STATE + i, ESDHC_DEBUG_SEL_REG);
+		ESDHC_IMX_DUMP("%s:  0x%04x\n", debug_status[i],
+			readw(host->ioaddr + ESDHC_DEBUG_SEL_AND_STATUS_REG));
+	}
+
+	esdhc_clrset_le(host, ESDHC_DEBUG_SEL_MASK, 0, ESDHC_DEBUG_SEL_REG);
+
+}
+
+static inline void esdhc_wait_for_card_clock_gate_off(struct sdhci_host *host)
+{
+	u32 present_state;
+	int ret;
+
+	ret = readl_poll_timeout(host->ioaddr + ESDHC_PRSSTAT, present_state,
+				(present_state & ESDHC_CLOCK_GATE_OFF), 2, 100);
+	if (ret == -ETIMEDOUT)
+		dev_warn(mmc_dev(host->mmc), "%s: card clock still not gate off in 100us!.\n", __func__);
+}
+
+static u32 esdhc_readl_le(struct sdhci_host *host, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	u32 val = readl(host->ioaddr + reg);
+
+	if (unlikely(reg == SDHCI_PRESENT_STATE)) {
+		u32 fsl_prss = val;
+		/* save the least 20 bits */
+		val = fsl_prss & 0x000FFFFF;
+		/* move dat[0-3] bits */
+		val |= (fsl_prss & 0x0F000000) >> 4;
+		/* move cmd line bit */
+		val |= (fsl_prss & 0x00800000) << 1;
+	}
+
+	if (unlikely(reg == SDHCI_CAPABILITIES)) {
+		/* ignore bit[0-15] as it stores cap_1 register val for mx6sl */
+		if (imx_data->socdata->flags & ESDHC_FLAG_HAVE_CAP1)
+			val &= 0xffff0000;
+
+		/* In FSL esdhc IC module, only bit20 is used to indicate the
+		 * ADMA2 capability of esdhc, but this bit is messed up on
+		 * some SOCs (e.g. on MX25, MX35 this bit is set, but they
+		 * don't actually support ADMA2). So set the BROKEN_ADMA
+		 * quirk on MX25/35 platforms.
+		 */
+
+		if (val & SDHCI_CAN_DO_ADMA1) {
+			val &= ~SDHCI_CAN_DO_ADMA1;
+			val |= SDHCI_CAN_DO_ADMA2;
+		}
+	}
+
+	if (unlikely(reg == SDHCI_CAPABILITIES_1)) {
+		if (esdhc_is_usdhc(imx_data)) {
+			if (imx_data->socdata->flags & ESDHC_FLAG_HAVE_CAP1)
+				val = readl(host->ioaddr + SDHCI_CAPABILITIES) & 0xFFFF;
+			else
+				/* imx6q/dl does not have cap_1 register, fake one */
+				val = SDHCI_SUPPORT_DDR50 | SDHCI_SUPPORT_SDR104
+					| SDHCI_SUPPORT_SDR50
+					| SDHCI_USE_SDR50_TUNING
+					| FIELD_PREP(SDHCI_RETUNING_MODE_MASK,
+						     SDHCI_TUNING_MODE_3);
+
+			if (imx_data->socdata->flags & ESDHC_FLAG_HS400)
+				val |= SDHCI_SUPPORT_HS400;
+
+			/*
+			 * Do not advertise faster UHS modes if there are no
+			 * pinctrl states for 100MHz/200MHz.
+			 */
+			if (IS_ERR_OR_NULL(imx_data->pins_100mhz) ||
+			    IS_ERR_OR_NULL(imx_data->pins_200mhz))
+				val &= ~(SDHCI_SUPPORT_SDR50 | SDHCI_SUPPORT_DDR50
+					 | SDHCI_SUPPORT_SDR104 | SDHCI_SUPPORT_HS400);
+		}
+	}
+
+	if (unlikely(reg == SDHCI_MAX_CURRENT) && esdhc_is_usdhc(imx_data)) {
+		val = 0;
+		val |= FIELD_PREP(SDHCI_MAX_CURRENT_330_MASK, 0xFF);
+		val |= FIELD_PREP(SDHCI_MAX_CURRENT_300_MASK, 0xFF);
+		val |= FIELD_PREP(SDHCI_MAX_CURRENT_180_MASK, 0xFF);
+	}
+
+	if (unlikely(reg == SDHCI_INT_STATUS)) {
+		if (val & ESDHC_INT_VENDOR_SPEC_DMA_ERR) {
+			val &= ~ESDHC_INT_VENDOR_SPEC_DMA_ERR;
+			val |= SDHCI_INT_ADMA_ERROR;
+		}
+
+		/*
+		 * mask off the interrupt we get in response to the manually
+		 * sent CMD12
+		 */
+		if ((imx_data->multiblock_status == WAIT_FOR_INT) &&
+		    ((val & SDHCI_INT_RESPONSE) == SDHCI_INT_RESPONSE)) {
+			val &= ~SDHCI_INT_RESPONSE;
+			writel(SDHCI_INT_RESPONSE, host->ioaddr +
+						   SDHCI_INT_STATUS);
+			imx_data->multiblock_status = NO_CMD_PENDING;
+		}
+	}
+
+	return val;
+}
+
+static void esdhc_writel_le(struct sdhci_host *host, u32 val, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	u32 data;
+
+	if (unlikely(reg == SDHCI_INT_ENABLE || reg == SDHCI_SIGNAL_ENABLE ||
+			reg == SDHCI_INT_STATUS)) {
+		if ((val & SDHCI_INT_CARD_INT) && !esdhc_is_usdhc(imx_data)) {
+			/*
+			 * Clear and then set D3CD bit to avoid missing the
+			 * card interrupt. This is an eSDHC controller problem
+			 * so we need to apply the following workaround: clear
+			 * and set D3CD bit will make eSDHC re-sample the card
+			 * interrupt. In case a card interrupt was lost,
+			 * re-sample it by the following steps.
+			 */
+			data = readl(host->ioaddr + SDHCI_HOST_CONTROL);
+			data &= ~ESDHC_CTRL_D3CD;
+			writel(data, host->ioaddr + SDHCI_HOST_CONTROL);
+			data |= ESDHC_CTRL_D3CD;
+			writel(data, host->ioaddr + SDHCI_HOST_CONTROL);
+		}
+
+		if (val & SDHCI_INT_ADMA_ERROR) {
+			val &= ~SDHCI_INT_ADMA_ERROR;
+			val |= ESDHC_INT_VENDOR_SPEC_DMA_ERR;
+		}
+	}
+
+	if (unlikely((imx_data->socdata->flags & ESDHC_FLAG_MULTIBLK_NO_INT)
+				&& (reg == SDHCI_INT_STATUS)
+				&& (val & SDHCI_INT_DATA_END))) {
+			u32 v;
+			v = readl(host->ioaddr + ESDHC_VENDOR_SPEC);
+			v &= ~ESDHC_VENDOR_SPEC_SDIO_QUIRK;
+			writel(v, host->ioaddr + ESDHC_VENDOR_SPEC);
+
+			if (imx_data->multiblock_status == MULTIBLK_IN_PROCESS)
+			{
+				/* send a manual CMD12 with RESPTYP=none */
+				data = MMC_STOP_TRANSMISSION << 24 |
+				       SDHCI_CMD_ABORTCMD << 16;
+				writel(data, host->ioaddr + SDHCI_TRANSFER_MODE);
+				imx_data->multiblock_status = WAIT_FOR_INT;
+			}
+	}
+
+	writel(val, host->ioaddr + reg);
+}
+
+static u16 esdhc_readw_le(struct sdhci_host *host, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	u16 ret = 0;
+	u32 val;
+
+	if (unlikely(reg == SDHCI_HOST_VERSION)) {
+		reg ^= 2;
+		if (esdhc_is_usdhc(imx_data)) {
+			/*
+			 * The usdhc register returns a wrong host version.
+			 * Correct it here.
+			 */
+			return SDHCI_SPEC_300;
+		}
+	}
+
+	if (unlikely(reg == SDHCI_HOST_CONTROL2)) {
+		val = readl(host->ioaddr + ESDHC_VENDOR_SPEC);
+		if (val & ESDHC_VENDOR_SPEC_VSELECT)
+			ret |= SDHCI_CTRL_VDD_180;
+
+		if (esdhc_is_usdhc(imx_data)) {
+			if (imx_data->socdata->flags & ESDHC_FLAG_MAN_TUNING)
+				val = readl(host->ioaddr + ESDHC_MIX_CTRL);
+			else if (imx_data->socdata->flags & ESDHC_FLAG_STD_TUNING)
+				/* the std tuning bits is in ACMD12_ERR for imx6sl */
+				val = readl(host->ioaddr + SDHCI_AUTO_CMD_STATUS);
+		}
+
+		if (val & ESDHC_MIX_CTRL_EXE_TUNE)
+			ret |= SDHCI_CTRL_EXEC_TUNING;
+		if (val & ESDHC_MIX_CTRL_SMPCLK_SEL)
+			ret |= SDHCI_CTRL_TUNED_CLK;
+
+		ret &= ~SDHCI_CTRL_PRESET_VAL_ENABLE;
+
+		return ret;
+	}
+
+	if (unlikely(reg == SDHCI_TRANSFER_MODE)) {
+		if (esdhc_is_usdhc(imx_data)) {
+			u32 m = readl(host->ioaddr + ESDHC_MIX_CTRL);
+			ret = m & ESDHC_MIX_CTRL_SDHCI_MASK;
+			/* Swap AC23 bit */
+			if (m & ESDHC_MIX_CTRL_AC23EN) {
+				ret &= ~ESDHC_MIX_CTRL_AC23EN;
+				ret |= SDHCI_TRNS_AUTO_CMD23;
+			}
+		} else {
+			ret = readw(host->ioaddr + SDHCI_TRANSFER_MODE);
+		}
+
+		return ret;
+	}
+
+	return readw(host->ioaddr + reg);
+}
+
+static void esdhc_writew_le(struct sdhci_host *host, u16 val, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	u32 new_val = 0;
+
+	switch (reg) {
+	case SDHCI_CLOCK_CONTROL:
+		new_val = readl(host->ioaddr + ESDHC_VENDOR_SPEC);
+		if (val & SDHCI_CLOCK_CARD_EN)
+			new_val |= ESDHC_VENDOR_SPEC_FRC_SDCLK_ON;
+		else
+			new_val &= ~ESDHC_VENDOR_SPEC_FRC_SDCLK_ON;
+		writel(new_val, host->ioaddr + ESDHC_VENDOR_SPEC);
+		if (!(new_val & ESDHC_VENDOR_SPEC_FRC_SDCLK_ON))
+			esdhc_wait_for_card_clock_gate_off(host);
+		return;
+	case SDHCI_HOST_CONTROL2:
+		new_val = readl(host->ioaddr + ESDHC_VENDOR_SPEC);
+		if (val & SDHCI_CTRL_VDD_180)
+			new_val |= ESDHC_VENDOR_SPEC_VSELECT;
+		else
+			new_val &= ~ESDHC_VENDOR_SPEC_VSELECT;
+		writel(new_val, host->ioaddr + ESDHC_VENDOR_SPEC);
+		if (imx_data->socdata->flags & ESDHC_FLAG_MAN_TUNING) {
+			new_val = readl(host->ioaddr + ESDHC_MIX_CTRL);
+			if (val & SDHCI_CTRL_TUNED_CLK) {
+				new_val |= ESDHC_MIX_CTRL_SMPCLK_SEL;
+				new_val |= ESDHC_MIX_CTRL_AUTO_TUNE_EN;
+			} else {
+				new_val &= ~ESDHC_MIX_CTRL_SMPCLK_SEL;
+				new_val &= ~ESDHC_MIX_CTRL_AUTO_TUNE_EN;
+			}
+			writel(new_val , host->ioaddr + ESDHC_MIX_CTRL);
+		} else if (imx_data->socdata->flags & ESDHC_FLAG_STD_TUNING) {
+			u32 v = readl(host->ioaddr + SDHCI_AUTO_CMD_STATUS);
+			u32 m = readl(host->ioaddr + ESDHC_MIX_CTRL);
+			if (val & SDHCI_CTRL_TUNED_CLK) {
+				v |= ESDHC_MIX_CTRL_SMPCLK_SEL;
+			} else {
+				v &= ~ESDHC_MIX_CTRL_SMPCLK_SEL;
+				m &= ~ESDHC_MIX_CTRL_FBCLK_SEL;
+				m &= ~ESDHC_MIX_CTRL_AUTO_TUNE_EN;
+			}
+
+			if (val & SDHCI_CTRL_EXEC_TUNING) {
+				v |= ESDHC_MIX_CTRL_EXE_TUNE;
+				m |= ESDHC_MIX_CTRL_FBCLK_SEL;
+				m |= ESDHC_MIX_CTRL_AUTO_TUNE_EN;
+			} else {
+				v &= ~ESDHC_MIX_CTRL_EXE_TUNE;
+			}
+
+			writel(v, host->ioaddr + SDHCI_AUTO_CMD_STATUS);
+			writel(m, host->ioaddr + ESDHC_MIX_CTRL);
+		}
+		return;
+	case SDHCI_TRANSFER_MODE:
+		if ((imx_data->socdata->flags & ESDHC_FLAG_MULTIBLK_NO_INT)
+				&& (host->cmd->opcode == SD_IO_RW_EXTENDED)
+				&& (host->cmd->data->blocks > 1)
+				&& (host->cmd->data->flags & MMC_DATA_READ)) {
+			u32 v;
+			v = readl(host->ioaddr + ESDHC_VENDOR_SPEC);
+			v |= ESDHC_VENDOR_SPEC_SDIO_QUIRK;
+			writel(v, host->ioaddr + ESDHC_VENDOR_SPEC);
+		}
+
+		if (esdhc_is_usdhc(imx_data)) {
+			u32 wml;
+			u32 m = readl(host->ioaddr + ESDHC_MIX_CTRL);
+			/* Swap AC23 bit */
+			if (val & SDHCI_TRNS_AUTO_CMD23) {
+				val &= ~SDHCI_TRNS_AUTO_CMD23;
+				val |= ESDHC_MIX_CTRL_AC23EN;
+			}
+			m = val | (m & ~ESDHC_MIX_CTRL_SDHCI_MASK);
+			writel(m, host->ioaddr + ESDHC_MIX_CTRL);
+
+			/* Set watermark levels for PIO access to maximum value
+			 * (128 words) to accommodate full 512 bytes buffer.
+			 * For DMA access restore the levels to default value.
+			 */
+			m = readl(host->ioaddr + ESDHC_WTMK_LVL);
+			if (val & SDHCI_TRNS_DMA) {
+				wml = ESDHC_WTMK_LVL_WML_VAL_DEF;
+			} else {
+				u8 ctrl;
+				wml = ESDHC_WTMK_LVL_WML_VAL_MAX;
+
+				/*
+				 * Since already disable DMA mode, so also need
+				 * to clear the DMASEL. Otherwise, for standard
+				 * tuning, when send tuning command, usdhc will
+				 * still prefetch the ADMA script from wrong
+				 * DMA address, then we will see IOMMU report
+				 * some error which show lack of TLB mapping.
+				 */
+				ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+				ctrl &= ~SDHCI_CTRL_DMA_MASK;
+				sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+			}
+			m &= ~(ESDHC_WTMK_LVL_RD_WML_MASK |
+			       ESDHC_WTMK_LVL_WR_WML_MASK);
+			m |= (wml << ESDHC_WTMK_LVL_RD_WML_SHIFT) |
+			     (wml << ESDHC_WTMK_LVL_WR_WML_SHIFT);
+			writel(m, host->ioaddr + ESDHC_WTMK_LVL);
+		} else {
+			/*
+			 * Postpone this write, we must do it together with a
+			 * command write that is down below.
+			 */
+			imx_data->scratchpad = val;
+		}
+		return;
+	case SDHCI_COMMAND:
+		if (host->cmd->opcode == MMC_STOP_TRANSMISSION)
+			val |= SDHCI_CMD_ABORTCMD;
+
+		if ((host->cmd->opcode == MMC_SET_BLOCK_COUNT) &&
+		    (imx_data->socdata->flags & ESDHC_FLAG_MULTIBLK_NO_INT))
+			imx_data->multiblock_status = MULTIBLK_IN_PROCESS;
+
+		if (esdhc_is_usdhc(imx_data))
+			writel(val << 16,
+			       host->ioaddr + SDHCI_TRANSFER_MODE);
+		else
+			writel(val << 16 | imx_data->scratchpad,
+			       host->ioaddr + SDHCI_TRANSFER_MODE);
+		return;
+	case SDHCI_BLOCK_SIZE:
+		val &= ~SDHCI_MAKE_BLKSZ(0x7, 0);
+		break;
+	}
+	esdhc_clrset_le(host, 0xffff, val, reg);
+}
+
+static u8 esdhc_readb_le(struct sdhci_host *host, int reg)
+{
+	u8 ret;
+	u32 val;
+
+	switch (reg) {
+	case SDHCI_HOST_CONTROL:
+		val = readl(host->ioaddr + reg);
+
+		ret = val & SDHCI_CTRL_LED;
+		ret |= (val >> 5) & SDHCI_CTRL_DMA_MASK;
+		ret |= (val & ESDHC_CTRL_4BITBUS);
+		ret |= (val & ESDHC_CTRL_8BITBUS) << 3;
+		return ret;
+	}
+
+	return readb(host->ioaddr + reg);
+}
+
+static void esdhc_writeb_le(struct sdhci_host *host, u8 val, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	u32 new_val = 0;
+	u32 mask;
+
+	switch (reg) {
+	case SDHCI_POWER_CONTROL:
+		/*
+		 * FSL put some DMA bits here
+		 * If your board has a regulator, code should be here
+		 */
+		return;
+	case SDHCI_HOST_CONTROL:
+		/* FSL messed up here, so we need to manually compose it. */
+		new_val = val & SDHCI_CTRL_LED;
+		/* ensure the endianness */
+		new_val |= ESDHC_HOST_CONTROL_LE;
+		/* bits 8&9 are reserved on mx25 */
+		if (!is_imx25_esdhc(imx_data)) {
+			/* DMA mode bits are shifted */
+			new_val |= (val & SDHCI_CTRL_DMA_MASK) << 5;
+		}
+
+		/*
+		 * Do not touch buswidth bits here. This is done in
+		 * esdhc_pltfm_bus_width.
+		 * Do not touch the D3CD bit either which is used for the
+		 * SDIO interrupt erratum workaround.
+		 */
+		mask = 0xffff & ~(ESDHC_CTRL_BUSWIDTH_MASK | ESDHC_CTRL_D3CD);
+
+		esdhc_clrset_le(host, mask, new_val, reg);
+		return;
+	case SDHCI_SOFTWARE_RESET:
+		if (val & SDHCI_RESET_DATA)
+			new_val = readl(host->ioaddr + SDHCI_HOST_CONTROL);
+		break;
+	}
+	esdhc_clrset_le(host, 0xff, val, reg);
+
+	if (reg == SDHCI_SOFTWARE_RESET) {
+		if (val & SDHCI_RESET_ALL) {
+			/*
+			 * The esdhc has a design violation to SDHC spec which
+			 * tells that software reset should not affect card
+			 * detection circuit. But esdhc clears its SYSCTL
+			 * register bits [0..2] during the software reset. This
+			 * will stop those clocks that card detection circuit
+			 * relies on. To work around it, we turn the clocks on
+			 * back to keep card detection circuit functional.
+			 */
+			esdhc_clrset_le(host, 0x7, 0x7, ESDHC_SYSTEM_CONTROL);
+			/*
+			 * The reset on usdhc fails to clear MIX_CTRL register.
+			 * Do it manually here.
+			 */
+			if (esdhc_is_usdhc(imx_data)) {
+				/*
+				 * the tuning bits should be kept during reset
+				 */
+				new_val = readl(host->ioaddr + ESDHC_MIX_CTRL);
+				writel(new_val & ESDHC_MIX_CTRL_TUNING_MASK,
+						host->ioaddr + ESDHC_MIX_CTRL);
+				imx_data->is_ddr = 0;
+			}
+		} else if (val & SDHCI_RESET_DATA) {
+			/*
+			 * The eSDHC DAT line software reset clears at least the
+			 * data transfer width on i.MX25, so make sure that the
+			 * Host Control register is unaffected.
+			 */
+			esdhc_clrset_le(host, 0xff, new_val,
+					SDHCI_HOST_CONTROL);
+		}
+	}
+}
+
+static unsigned int esdhc_pltfm_get_max_clock(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	return pltfm_host->clock;
+}
+
+static unsigned int esdhc_pltfm_get_min_clock(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	return pltfm_host->clock / 256 / 16;
+}
+
+static inline void esdhc_pltfm_set_clock(struct sdhci_host *host,
+					 unsigned int clock)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	unsigned int host_clock = pltfm_host->clock;
+	int ddr_pre_div = imx_data->is_ddr ? 2 : 1;
+	int pre_div = 1;
+	int div = 1;
+	int ret;
+	u32 temp, val;
+
+	if (esdhc_is_usdhc(imx_data)) {
+		val = readl(host->ioaddr + ESDHC_VENDOR_SPEC);
+		writel(val & ~ESDHC_VENDOR_SPEC_FRC_SDCLK_ON,
+			host->ioaddr + ESDHC_VENDOR_SPEC);
+		esdhc_wait_for_card_clock_gate_off(host);
+	}
+
+	if (clock == 0) {
+		host->mmc->actual_clock = 0;
+		return;
+	}
+
+	/* For i.MX53 eSDHCv3, SYSCTL.SDCLKFS may not be set to 0. */
+	if (is_imx53_esdhc(imx_data)) {
+		/*
+		 * According to the i.MX53 reference manual, if DLLCTRL[10] can
+		 * be set, then the controller is eSDHCv3, else it is eSDHCv2.
+		 */
+		val = readl(host->ioaddr + ESDHC_DLL_CTRL);
+		writel(val | BIT(10), host->ioaddr + ESDHC_DLL_CTRL);
+		temp = readl(host->ioaddr + ESDHC_DLL_CTRL);
+		writel(val, host->ioaddr + ESDHC_DLL_CTRL);
+		if (temp & BIT(10))
+			pre_div = 2;
+	}
+
+	temp = sdhci_readl(host, ESDHC_SYSTEM_CONTROL);
+	temp &= ~(ESDHC_CLOCK_IPGEN | ESDHC_CLOCK_HCKEN | ESDHC_CLOCK_PEREN
+		| ESDHC_CLOCK_MASK);
+	sdhci_writel(host, temp, ESDHC_SYSTEM_CONTROL);
+
+	if ((imx_data->socdata->flags & ESDHC_FLAG_ERR010450) &&
+	    (!(host->quirks2 & SDHCI_QUIRK2_NO_1_8_V))) {
+		unsigned int max_clock;
+
+		max_clock = imx_data->is_ddr ? 45000000 : 150000000;
+
+		clock = min(clock, max_clock);
+	}
+
+	while (host_clock / (16 * pre_div * ddr_pre_div) > clock &&
+			pre_div < 256)
+		pre_div *= 2;
+
+	while (host_clock / (div * pre_div * ddr_pre_div) > clock && div < 16)
+		div++;
+
+	host->mmc->actual_clock = host_clock / (div * pre_div * ddr_pre_div);
+	dev_dbg(mmc_dev(host->mmc), "desired SD clock: %d, actual: %d\n",
+		clock, host->mmc->actual_clock);
+
+	pre_div >>= 1;
+	div--;
+
+	temp = sdhci_readl(host, ESDHC_SYSTEM_CONTROL);
+	temp |= (ESDHC_CLOCK_IPGEN | ESDHC_CLOCK_HCKEN | ESDHC_CLOCK_PEREN
+		| (div << ESDHC_DIVIDER_SHIFT)
+		| (pre_div << ESDHC_PREDIV_SHIFT));
+	sdhci_writel(host, temp, ESDHC_SYSTEM_CONTROL);
+
+	/* need to wait the bit 3 of the PRSSTAT to be set, make sure card clock is stable */
+	ret = readl_poll_timeout(host->ioaddr + ESDHC_PRSSTAT, temp,
+				(temp & ESDHC_CLOCK_STABLE), 2, 100);
+	if (ret == -ETIMEDOUT)
+		dev_warn(mmc_dev(host->mmc), "card clock still not stable in 100us!.\n");
+
+	if (esdhc_is_usdhc(imx_data)) {
+		val = readl(host->ioaddr + ESDHC_VENDOR_SPEC);
+		writel(val | ESDHC_VENDOR_SPEC_FRC_SDCLK_ON,
+			host->ioaddr + ESDHC_VENDOR_SPEC);
+	}
+
+}
+
+static unsigned int esdhc_pltfm_get_ro(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	struct esdhc_platform_data *boarddata = &imx_data->boarddata;
+
+	switch (boarddata->wp_type) {
+	case ESDHC_WP_GPIO:
+		return mmc_gpio_get_ro(host->mmc);
+	case ESDHC_WP_CONTROLLER:
+		return !(readl(host->ioaddr + SDHCI_PRESENT_STATE) &
+			       SDHCI_WRITE_PROTECT);
+	case ESDHC_WP_NONE:
+		break;
+	}
+
+	return -ENOSYS;
+}
+
+static void esdhc_pltfm_set_bus_width(struct sdhci_host *host, int width)
+{
+	u32 ctrl;
+
+	switch (width) {
+	case MMC_BUS_WIDTH_8:
+		ctrl = ESDHC_CTRL_8BITBUS;
+		break;
+	case MMC_BUS_WIDTH_4:
+		ctrl = ESDHC_CTRL_4BITBUS;
+		break;
+	default:
+		ctrl = 0;
+		break;
+	}
+
+	esdhc_clrset_le(host, ESDHC_CTRL_BUSWIDTH_MASK, ctrl,
+			SDHCI_HOST_CONTROL);
+}
+
+static int usdhc_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	/*
+	 * i.MX uSDHC internally already uses a fixed optimized timing for
+	 * DDR50, normally does not require tuning for DDR50 mode.
+	 */
+	if (host->timing == MMC_TIMING_UHS_DDR50)
+		return 0;
+
+	return sdhci_execute_tuning(mmc, opcode);
+}
+
+static void esdhc_prepare_tuning(struct sdhci_host *host, u32 val)
+{
+	u32 reg;
+	u8 sw_rst;
+	int ret;
+
+	/* FIXME: delay a bit for card to be ready for next tuning due to errors */
+	mdelay(1);
+
+	/* IC suggest to reset USDHC before every tuning command */
+	esdhc_clrset_le(host, 0xff, SDHCI_RESET_ALL, SDHCI_SOFTWARE_RESET);
+	ret = readb_poll_timeout(host->ioaddr + SDHCI_SOFTWARE_RESET, sw_rst,
+				!(sw_rst & SDHCI_RESET_ALL), 10, 100);
+	if (ret == -ETIMEDOUT)
+		dev_warn(mmc_dev(host->mmc),
+		"warning! RESET_ALL never complete before sending tuning command\n");
+
+	reg = readl(host->ioaddr + ESDHC_MIX_CTRL);
+	reg |= ESDHC_MIX_CTRL_EXE_TUNE | ESDHC_MIX_CTRL_SMPCLK_SEL |
+			ESDHC_MIX_CTRL_FBCLK_SEL;
+	writel(reg, host->ioaddr + ESDHC_MIX_CTRL);
+	writel(val << 8, host->ioaddr + ESDHC_TUNE_CTRL_STATUS);
+	dev_dbg(mmc_dev(host->mmc),
+		"tuning with delay 0x%x ESDHC_TUNE_CTRL_STATUS 0x%x\n",
+			val, readl(host->ioaddr + ESDHC_TUNE_CTRL_STATUS));
+}
+
+static void esdhc_post_tuning(struct sdhci_host *host)
+{
+	u32 reg;
+
+	reg = readl(host->ioaddr + ESDHC_MIX_CTRL);
+	reg &= ~ESDHC_MIX_CTRL_EXE_TUNE;
+	reg |= ESDHC_MIX_CTRL_AUTO_TUNE_EN;
+	writel(reg, host->ioaddr + ESDHC_MIX_CTRL);
+}
+
+static int esdhc_executing_tuning(struct sdhci_host *host, u32 opcode)
+{
+	int min, max, avg, ret;
+
+	/* find the mininum delay first which can pass tuning */
+	min = ESDHC_TUNE_CTRL_MIN;
+	while (min < ESDHC_TUNE_CTRL_MAX) {
+		esdhc_prepare_tuning(host, min);
+		if (!mmc_send_tuning(host->mmc, opcode, NULL))
+			break;
+		min += ESDHC_TUNE_CTRL_STEP;
+	}
+
+	/* find the maxinum delay which can not pass tuning */
+	max = min + ESDHC_TUNE_CTRL_STEP;
+	while (max < ESDHC_TUNE_CTRL_MAX) {
+		esdhc_prepare_tuning(host, max);
+		if (mmc_send_tuning(host->mmc, opcode, NULL)) {
+			max -= ESDHC_TUNE_CTRL_STEP;
+			break;
+		}
+		max += ESDHC_TUNE_CTRL_STEP;
+	}
+
+	/* use average delay to get the best timing */
+	avg = (min + max) / 2;
+	esdhc_prepare_tuning(host, avg);
+	ret = mmc_send_tuning(host->mmc, opcode, NULL);
+	esdhc_post_tuning(host);
+
+	dev_dbg(mmc_dev(host->mmc), "tuning %s at 0x%x ret %d\n",
+		ret ? "failed" : "passed", avg, ret);
+
+	return ret;
+}
+
+static void esdhc_hs400_enhanced_strobe(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	u32 m;
+
+	m = readl(host->ioaddr + ESDHC_MIX_CTRL);
+	if (ios->enhanced_strobe)
+		m |= ESDHC_MIX_CTRL_HS400_ES_EN;
+	else
+		m &= ~ESDHC_MIX_CTRL_HS400_ES_EN;
+	writel(m, host->ioaddr + ESDHC_MIX_CTRL);
+}
+
+static int esdhc_change_pinstate(struct sdhci_host *host,
+						unsigned int uhs)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	struct pinctrl_state *pinctrl;
+
+	dev_dbg(mmc_dev(host->mmc), "change pinctrl state for uhs %d\n", uhs);
+
+	if (IS_ERR(imx_data->pinctrl) ||
+		IS_ERR(imx_data->pins_100mhz) ||
+		IS_ERR(imx_data->pins_200mhz))
+		return -EINVAL;
+
+	switch (uhs) {
+	case MMC_TIMING_UHS_SDR50:
+	case MMC_TIMING_UHS_DDR50:
+		pinctrl = imx_data->pins_100mhz;
+		break;
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_MMC_HS200:
+	case MMC_TIMING_MMC_HS400:
+		pinctrl = imx_data->pins_200mhz;
+		break;
+	default:
+		/* back to default state for other legacy timing */
+		return pinctrl_select_default_state(mmc_dev(host->mmc));
+	}
+
+	return pinctrl_select_state(imx_data->pinctrl, pinctrl);
+}
+
+/*
+ * For HS400 eMMC, there is a data_strobe line. This signal is generated
+ * by the device and used for data output and CRC status response output
+ * in HS400 mode. The frequency of this signal follows the frequency of
+ * CLK generated by host. The host receives the data which is aligned to the
+ * edge of data_strobe line. Due to the time delay between CLK line and
+ * data_strobe line, if the delay time is larger than one clock cycle,
+ * then CLK and data_strobe line will be misaligned, read error shows up.
+ */
+static void esdhc_set_strobe_dll(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	u32 strobe_delay;
+	u32 v;
+	int ret;
+
+	/* disable clock before enabling strobe dll */
+	writel(readl(host->ioaddr + ESDHC_VENDOR_SPEC) &
+		~ESDHC_VENDOR_SPEC_FRC_SDCLK_ON,
+		host->ioaddr + ESDHC_VENDOR_SPEC);
+	esdhc_wait_for_card_clock_gate_off(host);
+
+	/* force a reset on strobe dll */
+	writel(ESDHC_STROBE_DLL_CTRL_RESET,
+		host->ioaddr + ESDHC_STROBE_DLL_CTRL);
+	/* clear the reset bit on strobe dll before any setting */
+	writel(0, host->ioaddr + ESDHC_STROBE_DLL_CTRL);
+
+	/*
+	 * enable strobe dll ctrl and adjust the delay target
+	 * for the uSDHC loopback read clock
+	 */
+	if (imx_data->boarddata.strobe_dll_delay_target)
+		strobe_delay = imx_data->boarddata.strobe_dll_delay_target;
+	else
+		strobe_delay = ESDHC_STROBE_DLL_CTRL_SLV_DLY_TARGET_DEFAULT;
+	v = ESDHC_STROBE_DLL_CTRL_ENABLE |
+		ESDHC_STROBE_DLL_CTRL_SLV_UPDATE_INT_DEFAULT |
+		(strobe_delay << ESDHC_STROBE_DLL_CTRL_SLV_DLY_TARGET_SHIFT);
+	writel(v, host->ioaddr + ESDHC_STROBE_DLL_CTRL);
+
+	/* wait max 50us to get the REF/SLV lock */
+	ret = readl_poll_timeout(host->ioaddr + ESDHC_STROBE_DLL_STATUS, v,
+		((v & ESDHC_STROBE_DLL_STS_REF_LOCK) && (v & ESDHC_STROBE_DLL_STS_SLV_LOCK)), 1, 50);
+	if (ret == -ETIMEDOUT)
+		dev_warn(mmc_dev(host->mmc),
+		"warning! HS400 strobe DLL status REF/SLV not lock in 50us, STROBE DLL status is %x!\n", v);
+}
+
+static void esdhc_reset_tuning(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	u32 ctrl;
+	int ret;
+
+	/* Reset the tuning circuit */
+	if (esdhc_is_usdhc(imx_data)) {
+		if (imx_data->socdata->flags & ESDHC_FLAG_MAN_TUNING) {
+			ctrl = readl(host->ioaddr + ESDHC_MIX_CTRL);
+			ctrl &= ~ESDHC_MIX_CTRL_SMPCLK_SEL;
+			ctrl &= ~ESDHC_MIX_CTRL_FBCLK_SEL;
+			writel(ctrl, host->ioaddr + ESDHC_MIX_CTRL);
+			writel(0, host->ioaddr + ESDHC_TUNE_CTRL_STATUS);
+		} else if (imx_data->socdata->flags & ESDHC_FLAG_STD_TUNING) {
+			ctrl = readl(host->ioaddr + SDHCI_AUTO_CMD_STATUS);
+			ctrl &= ~ESDHC_MIX_CTRL_SMPCLK_SEL;
+			ctrl &= ~ESDHC_MIX_CTRL_EXE_TUNE;
+			writel(ctrl, host->ioaddr + SDHCI_AUTO_CMD_STATUS);
+			/* Make sure ESDHC_MIX_CTRL_EXE_TUNE cleared */
+			ret = readl_poll_timeout(host->ioaddr + SDHCI_AUTO_CMD_STATUS,
+				ctrl, !(ctrl & ESDHC_MIX_CTRL_EXE_TUNE), 1, 50);
+			if (ret == -ETIMEDOUT)
+				dev_warn(mmc_dev(host->mmc),
+				 "Warning! clear execute tuning bit failed\n");
+			/*
+			 * SDHCI_INT_DATA_AVAIL is W1C bit, set this bit will clear the
+			 * usdhc IP internal logic flag execute_tuning_with_clr_buf, which
+			 * will finally make sure the normal data transfer logic correct.
+			 */
+			ctrl = readl(host->ioaddr + SDHCI_INT_STATUS);
+			ctrl |= SDHCI_INT_DATA_AVAIL;
+			writel(ctrl, host->ioaddr + SDHCI_INT_STATUS);
+		}
+	}
+}
+
+static void esdhc_set_uhs_signaling(struct sdhci_host *host, unsigned timing)
+{
+	u32 m;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	struct esdhc_platform_data *boarddata = &imx_data->boarddata;
+
+	/* disable ddr mode and disable HS400 mode */
+	m = readl(host->ioaddr + ESDHC_MIX_CTRL);
+	m &= ~(ESDHC_MIX_CTRL_DDREN | ESDHC_MIX_CTRL_HS400_EN);
+	imx_data->is_ddr = 0;
+
+	switch (timing) {
+	case MMC_TIMING_UHS_SDR12:
+	case MMC_TIMING_UHS_SDR25:
+	case MMC_TIMING_UHS_SDR50:
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_MMC_HS:
+	case MMC_TIMING_MMC_HS200:
+		writel(m, host->ioaddr + ESDHC_MIX_CTRL);
+		break;
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_MMC_DDR52:
+		m |= ESDHC_MIX_CTRL_DDREN;
+		writel(m, host->ioaddr + ESDHC_MIX_CTRL);
+		imx_data->is_ddr = 1;
+		if (boarddata->delay_line) {
+			u32 v;
+			v = boarddata->delay_line <<
+				ESDHC_DLL_OVERRIDE_VAL_SHIFT |
+				(1 << ESDHC_DLL_OVERRIDE_EN_SHIFT);
+			if (is_imx53_esdhc(imx_data))
+				v <<= 1;
+			writel(v, host->ioaddr + ESDHC_DLL_CTRL);
+		}
+		break;
+	case MMC_TIMING_MMC_HS400:
+		m |= ESDHC_MIX_CTRL_DDREN | ESDHC_MIX_CTRL_HS400_EN;
+		writel(m, host->ioaddr + ESDHC_MIX_CTRL);
+		imx_data->is_ddr = 1;
+		/* update clock after enable DDR for strobe DLL lock */
+		host->ops->set_clock(host, host->clock);
+		esdhc_set_strobe_dll(host);
+		break;
+	case MMC_TIMING_LEGACY:
+	default:
+		esdhc_reset_tuning(host);
+		break;
+	}
+
+	esdhc_change_pinstate(host, timing);
+}
+
+static void esdhc_reset(struct sdhci_host *host, u8 mask)
+{
+	sdhci_and_cqhci_reset(host, mask);
+
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+}
+
+static unsigned int esdhc_get_max_timeout_count(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+
+	/* Doc Erratum: the uSDHC actual maximum timeout count is 1 << 29 */
+	return esdhc_is_usdhc(imx_data) ? 1 << 29 : 1 << 27;
+}
+
+static void esdhc_set_timeout(struct sdhci_host *host, struct mmc_command *cmd)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+
+	/* use maximum timeout counter */
+	esdhc_clrset_le(host, ESDHC_SYS_CTRL_DTOCV_MASK,
+			esdhc_is_usdhc(imx_data) ? 0xF : 0xE,
+			SDHCI_TIMEOUT_CONTROL);
+}
+
+static u32 esdhc_cqhci_irq(struct sdhci_host *host, u32 intmask)
+{
+	int cmd_error = 0;
+	int data_error = 0;
+
+	if (!sdhci_cqe_irq(host, intmask, &cmd_error, &data_error))
+		return intmask;
+
+	cqhci_irq(host->mmc, intmask, cmd_error, data_error);
+
+	return 0;
+}
+
+static struct sdhci_ops sdhci_esdhc_ops = {
+	.read_l = esdhc_readl_le,
+	.read_w = esdhc_readw_le,
+	.read_b = esdhc_readb_le,
+	.write_l = esdhc_writel_le,
+	.write_w = esdhc_writew_le,
+	.write_b = esdhc_writeb_le,
+	.set_clock = esdhc_pltfm_set_clock,
+	.get_max_clock = esdhc_pltfm_get_max_clock,
+	.get_min_clock = esdhc_pltfm_get_min_clock,
+	.get_max_timeout_count = esdhc_get_max_timeout_count,
+	.get_ro = esdhc_pltfm_get_ro,
+	.set_timeout = esdhc_set_timeout,
+	.set_bus_width = esdhc_pltfm_set_bus_width,
+	.set_uhs_signaling = esdhc_set_uhs_signaling,
+	.reset = esdhc_reset,
+	.irq = esdhc_cqhci_irq,
+	.dump_vendor_regs = esdhc_dump_debug_regs,
+};
+
+static const struct sdhci_pltfm_data sdhci_esdhc_imx_pdata = {
+	.quirks = ESDHC_DEFAULT_QUIRKS | SDHCI_QUIRK_NO_HISPD_BIT
+			| SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC
+			| SDHCI_QUIRK_BROKEN_ADMA_ZEROLEN_DESC
+			| SDHCI_QUIRK_BROKEN_CARD_DETECTION,
+	.ops = &sdhci_esdhc_ops,
+};
+
+static void sdhci_esdhc_imx_hwinit(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	struct cqhci_host *cq_host = host->mmc->cqe_private;
+	u32 tmp;
+
+	if (esdhc_is_usdhc(imx_data)) {
+		/*
+		 * The imx6q ROM code will change the default watermark
+		 * level setting to something insane.  Change it back here.
+		 */
+		writel(ESDHC_WTMK_DEFAULT_VAL, host->ioaddr + ESDHC_WTMK_LVL);
+
+		/*
+		 * ROM code will change the bit burst_length_enable setting
+		 * to zero if this usdhc is chosen to boot system. Change
+		 * it back here, otherwise it will impact the performance a
+		 * lot. This bit is used to enable/disable the burst length
+		 * for the external AHB2AXI bridge. It's useful especially
+		 * for INCR transfer because without burst length indicator,
+		 * the AHB2AXI bridge does not know the burst length in
+		 * advance. And without burst length indicator, AHB INCR
+		 * transfer can only be converted to singles on the AXI side.
+		 */
+		writel(readl(host->ioaddr + SDHCI_HOST_CONTROL)
+			| ESDHC_BURST_LEN_EN_INCR,
+			host->ioaddr + SDHCI_HOST_CONTROL);
+
+		/*
+		 * erratum ESDHC_FLAG_ERR004536 fix for MX6Q TO1.2 and MX6DL
+		 * TO1.1, it's harmless for MX6SL
+		 */
+		writel(readl(host->ioaddr + 0x6c) & ~BIT(7),
+			host->ioaddr + 0x6c);
+
+		/* disable DLL_CTRL delay line settings */
+		writel(0x0, host->ioaddr + ESDHC_DLL_CTRL);
+
+		/*
+		 * For the case of command with busy, if set the bit
+		 * ESDHC_VEND_SPEC2_EN_BUSY_IRQ, USDHC will generate a
+		 * transfer complete interrupt when busy is deasserted.
+		 * When CQHCI use DCMD to send a CMD need R1b respons,
+		 * CQHCI require to set ESDHC_VEND_SPEC2_EN_BUSY_IRQ,
+		 * otherwise DCMD will always meet timeout waiting for
+		 * hardware interrupt issue.
+		 */
+		if (imx_data->socdata->flags & ESDHC_FLAG_CQHCI) {
+			tmp = readl(host->ioaddr + ESDHC_VEND_SPEC2);
+			tmp |= ESDHC_VEND_SPEC2_EN_BUSY_IRQ;
+			writel(tmp, host->ioaddr + ESDHC_VEND_SPEC2);
+
+			host->quirks &= ~SDHCI_QUIRK_NO_BUSY_IRQ;
+		}
+
+		if (imx_data->socdata->flags & ESDHC_FLAG_STD_TUNING) {
+			tmp = readl(host->ioaddr + ESDHC_TUNING_CTRL);
+			tmp |= ESDHC_STD_TUNING_EN;
+
+			/*
+			 * ROM code or bootloader may config the start tap
+			 * and step, unmask them first.
+			 */
+			tmp &= ~(ESDHC_TUNING_START_TAP_MASK | ESDHC_TUNING_STEP_MASK);
+			if (imx_data->boarddata.tuning_start_tap)
+				tmp |= imx_data->boarddata.tuning_start_tap;
+			else
+				tmp |= ESDHC_TUNING_START_TAP_DEFAULT;
+
+			if (imx_data->boarddata.tuning_step) {
+				tmp |= imx_data->boarddata.tuning_step
+					<< ESDHC_TUNING_STEP_SHIFT;
+			} else {
+				tmp |= ESDHC_TUNING_STEP_DEFAULT
+					<< ESDHC_TUNING_STEP_SHIFT;
+			}
+
+			/* Disable the CMD CRC check for tuning, if not, need to
+			 * add some delay after every tuning command, because
+			 * hardware standard tuning logic will directly go to next
+			 * step once it detect the CMD CRC error, will not wait for
+			 * the card side to finally send out the tuning data, trigger
+			 * the buffer read ready interrupt immediately. If usdhc send
+			 * the next tuning command some eMMC card will stuck, can't
+			 * response, block the tuning procedure or the first command
+			 * after the whole tuning procedure always can't get any response.
+			 */
+			tmp |= ESDHC_TUNING_CMD_CRC_CHECK_DISABLE;
+			writel(tmp, host->ioaddr + ESDHC_TUNING_CTRL);
+		} else if (imx_data->socdata->flags & ESDHC_FLAG_MAN_TUNING) {
+			/*
+			 * ESDHC_STD_TUNING_EN may be configed in bootloader
+			 * or ROM code, so clear this bit here to make sure
+			 * the manual tuning can work.
+			 */
+			tmp = readl(host->ioaddr + ESDHC_TUNING_CTRL);
+			tmp &= ~ESDHC_STD_TUNING_EN;
+			writel(tmp, host->ioaddr + ESDHC_TUNING_CTRL);
+		}
+
+		/*
+		 * On i.MX8MM, we are running Dual Linux OS, with 1st Linux using SD Card
+		 * as rootfs storage, 2nd Linux using eMMC as rootfs storage. We let the
+		 * the 1st linux configure power/clock for the 2nd Linux.
+		 *
+		 * When the 2nd Linux is booting into rootfs stage, we let the 1st Linux
+		 * to destroy the 2nd linux, then restart the 2nd linux, we met SDHCI dump.
+		 * After we clear the pending interrupt and halt CQCTL, issue gone.
+		 */
+		if (cq_host) {
+			tmp = cqhci_readl(cq_host, CQHCI_IS);
+			cqhci_writel(cq_host, tmp, CQHCI_IS);
+			cqhci_writel(cq_host, CQHCI_HALT, CQHCI_CTL);
+		}
+	}
+}
+
+static void esdhc_cqe_enable(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	u32 reg;
+	u16 mode;
+	int count = 10;
+
+	/*
+	 * CQE gets stuck if it sees Buffer Read Enable bit set, which can be
+	 * the case after tuning, so ensure the buffer is drained.
+	 */
+	reg = sdhci_readl(host, SDHCI_PRESENT_STATE);
+	while (reg & SDHCI_DATA_AVAILABLE) {
+		sdhci_readl(host, SDHCI_BUFFER);
+		reg = sdhci_readl(host, SDHCI_PRESENT_STATE);
+		if (count-- == 0) {
+			dev_warn(mmc_dev(host->mmc),
+				"CQE may get stuck because the Buffer Read Enable bit is set\n");
+			break;
+		}
+		mdelay(1);
+	}
+
+	/*
+	 * Runtime resume will reset the entire host controller, which
+	 * will also clear the DMAEN/BCEN of register ESDHC_MIX_CTRL.
+	 * Here set DMAEN and BCEN when enable CMDQ.
+	 */
+	mode = sdhci_readw(host, SDHCI_TRANSFER_MODE);
+	if (host->flags & SDHCI_REQ_USE_DMA)
+		mode |= SDHCI_TRNS_DMA;
+	if (!(host->quirks2 & SDHCI_QUIRK2_SUPPORT_SINGLE))
+		mode |= SDHCI_TRNS_BLK_CNT_EN;
+	sdhci_writew(host, mode, SDHCI_TRANSFER_MODE);
+
+	/*
+	 * Though Runtime resume reset the entire host controller,
+	 * but do not impact the CQHCI side, need to clear the
+	 * HALT bit, avoid CQHCI stuck in the first request when
+	 * system resume back.
+	 */
+	cqhci_writel(cq_host, 0, CQHCI_CTL);
+	if (cqhci_readl(cq_host, CQHCI_CTL) & CQHCI_HALT)
+		dev_err(mmc_dev(host->mmc),
+			"failed to exit halt state when enable CQE\n");
+
+
+	sdhci_cqe_enable(mmc);
+}
+
+static void esdhc_sdhci_dumpregs(struct mmc_host *mmc)
+{
+	sdhci_dumpregs(mmc_priv(mmc));
+}
+
+static const struct cqhci_host_ops esdhc_cqhci_ops = {
+	.enable		= esdhc_cqe_enable,
+	.disable	= sdhci_cqe_disable,
+	.dumpregs	= esdhc_sdhci_dumpregs,
+};
+
+#ifdef CONFIG_OF
+static int
+sdhci_esdhc_imx_probe_dt(struct platform_device *pdev,
+			 struct sdhci_host *host,
+			 struct pltfm_imx_data *imx_data)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct esdhc_platform_data *boarddata = &imx_data->boarddata;
+	int ret;
+
+	if (of_get_property(np, "fsl,wp-controller", NULL))
+		boarddata->wp_type = ESDHC_WP_CONTROLLER;
+
+	/*
+	 * If we have this property, then activate WP check.
+	 * Retrieveing and requesting the actual WP GPIO will happen
+	 * in the call to mmc_of_parse().
+	 */
+	if (of_property_read_bool(np, "wp-gpios"))
+		boarddata->wp_type = ESDHC_WP_GPIO;
+
+	of_property_read_u32(np, "fsl,tuning-step", &boarddata->tuning_step);
+	of_property_read_u32(np, "fsl,tuning-start-tap",
+			     &boarddata->tuning_start_tap);
+
+	of_property_read_u32(np, "fsl,strobe-dll-delay-target",
+				&boarddata->strobe_dll_delay_target);
+	if (of_find_property(np, "no-1-8-v", NULL))
+		host->quirks2 |= SDHCI_QUIRK2_NO_1_8_V;
+
+	if (of_property_read_u32(np, "fsl,delay-line", &boarddata->delay_line))
+		boarddata->delay_line = 0;
+
+	mmc_of_parse_voltage(np, &host->ocr_mask);
+
+	if (esdhc_is_usdhc(imx_data) && !IS_ERR(imx_data->pinctrl)) {
+		imx_data->pins_100mhz = pinctrl_lookup_state(imx_data->pinctrl,
+						ESDHC_PINCTRL_STATE_100MHZ);
+		imx_data->pins_200mhz = pinctrl_lookup_state(imx_data->pinctrl,
+						ESDHC_PINCTRL_STATE_200MHZ);
+	}
+
+	/* call to generic mmc_of_parse to support additional capabilities */
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		return ret;
+
+	if (mmc_gpio_get_cd(host->mmc) >= 0)
+		host->quirks &= ~SDHCI_QUIRK_BROKEN_CARD_DETECTION;
+
+	return 0;
+}
+#else
+static inline int
+sdhci_esdhc_imx_probe_dt(struct platform_device *pdev,
+			 struct sdhci_host *host,
+			 struct pltfm_imx_data *imx_data)
+{
+	return -ENODEV;
+}
+#endif
+
+static int sdhci_esdhc_imx_probe(struct platform_device *pdev)
+{
+	const struct of_device_id *of_id =
+			of_match_device(imx_esdhc_dt_ids, &pdev->dev);
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_host *host;
+	struct cqhci_host *cq_host;
+	int err;
+	struct pltfm_imx_data *imx_data;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_esdhc_imx_pdata,
+				sizeof(*imx_data));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+
+	imx_data = sdhci_pltfm_priv(pltfm_host);
+
+	imx_data->socdata = of_id->data;
+
+	if (imx_data->socdata->flags & ESDHC_FLAG_PMQOS)
+		cpu_latency_qos_add_request(&imx_data->pm_qos_req, 0);
+
+	imx_data->clk_ipg = devm_clk_get(&pdev->dev, "ipg");
+	if (IS_ERR(imx_data->clk_ipg)) {
+		err = PTR_ERR(imx_data->clk_ipg);
+		goto free_sdhci;
+	}
+
+	imx_data->clk_ahb = devm_clk_get(&pdev->dev, "ahb");
+	if (IS_ERR(imx_data->clk_ahb)) {
+		err = PTR_ERR(imx_data->clk_ahb);
+		goto free_sdhci;
+	}
+
+	imx_data->clk_per = devm_clk_get(&pdev->dev, "per");
+	if (IS_ERR(imx_data->clk_per)) {
+		err = PTR_ERR(imx_data->clk_per);
+		goto free_sdhci;
+	}
+
+	pltfm_host->clk = imx_data->clk_per;
+	pltfm_host->clock = clk_get_rate(pltfm_host->clk);
+	err = clk_prepare_enable(imx_data->clk_per);
+	if (err)
+		goto free_sdhci;
+	err = clk_prepare_enable(imx_data->clk_ipg);
+	if (err)
+		goto disable_per_clk;
+	err = clk_prepare_enable(imx_data->clk_ahb);
+	if (err)
+		goto disable_ipg_clk;
+
+	imx_data->pinctrl = devm_pinctrl_get(&pdev->dev);
+	if (IS_ERR(imx_data->pinctrl))
+		dev_warn(mmc_dev(host->mmc), "could not get pinctrl\n");
+
+	if (esdhc_is_usdhc(imx_data)) {
+		host->quirks2 |= SDHCI_QUIRK2_PRESET_VALUE_BROKEN;
+		host->mmc->caps |= MMC_CAP_1_8V_DDR | MMC_CAP_3_3V_DDR;
+
+		/* GPIO CD can be set as a wakeup source */
+		host->mmc->caps |= MMC_CAP_CD_WAKE;
+
+		if (!(imx_data->socdata->flags & ESDHC_FLAG_HS200))
+			host->quirks2 |= SDHCI_QUIRK2_BROKEN_HS200;
+
+		/* clear tuning bits in case ROM has set it already */
+		writel(0x0, host->ioaddr + ESDHC_MIX_CTRL);
+		writel(0x0, host->ioaddr + SDHCI_AUTO_CMD_STATUS);
+		writel(0x0, host->ioaddr + ESDHC_TUNE_CTRL_STATUS);
+
+		/*
+		 * Link usdhc specific mmc_host_ops execute_tuning function,
+		 * to replace the standard one in sdhci_ops.
+		 */
+		host->mmc_host_ops.execute_tuning = usdhc_execute_tuning;
+	}
+
+	err = sdhci_esdhc_imx_probe_dt(pdev, host, imx_data);
+	if (err)
+		goto disable_ahb_clk;
+
+	if (imx_data->socdata->flags & ESDHC_FLAG_MAN_TUNING)
+		sdhci_esdhc_ops.platform_execute_tuning =
+					esdhc_executing_tuning;
+
+	if (imx_data->socdata->flags & ESDHC_FLAG_ERR004536)
+		host->quirks |= SDHCI_QUIRK_BROKEN_ADMA;
+
+	if (host->mmc->caps & MMC_CAP_8_BIT_DATA &&
+	    imx_data->socdata->flags & ESDHC_FLAG_HS400)
+		host->quirks2 |= SDHCI_QUIRK2_CAPS_BIT63_FOR_HS400;
+
+	if (imx_data->socdata->flags & ESDHC_FLAG_BROKEN_AUTO_CMD23)
+		host->quirks2 |= SDHCI_QUIRK2_ACMD23_BROKEN;
+
+	if (host->mmc->caps & MMC_CAP_8_BIT_DATA &&
+	    imx_data->socdata->flags & ESDHC_FLAG_HS400_ES) {
+		host->mmc->caps2 |= MMC_CAP2_HS400_ES;
+		host->mmc_host_ops.hs400_enhanced_strobe =
+					esdhc_hs400_enhanced_strobe;
+	}
+
+	if (imx_data->socdata->flags & ESDHC_FLAG_CQHCI) {
+		host->mmc->caps2 |= MMC_CAP2_CQE | MMC_CAP2_CQE_DCMD;
+		cq_host = devm_kzalloc(&pdev->dev, sizeof(*cq_host), GFP_KERNEL);
+		if (!cq_host) {
+			err = -ENOMEM;
+			goto disable_ahb_clk;
+		}
+
+		cq_host->mmio = host->ioaddr + ESDHC_CQHCI_ADDR_OFFSET;
+		cq_host->ops = &esdhc_cqhci_ops;
+
+		err = cqhci_init(cq_host, host->mmc, false);
+		if (err)
+			goto disable_ahb_clk;
+	}
+
+	sdhci_esdhc_imx_hwinit(host);
+
+	err = sdhci_add_host(host);
+	if (err)
+		goto disable_ahb_clk;
+
+	pm_runtime_set_active(&pdev->dev);
+	pm_runtime_set_autosuspend_delay(&pdev->dev, 50);
+	pm_runtime_use_autosuspend(&pdev->dev);
+	pm_suspend_ignore_children(&pdev->dev, 1);
+	pm_runtime_enable(&pdev->dev);
+
+	return 0;
+
+disable_ahb_clk:
+	clk_disable_unprepare(imx_data->clk_ahb);
+disable_ipg_clk:
+	clk_disable_unprepare(imx_data->clk_ipg);
+disable_per_clk:
+	clk_disable_unprepare(imx_data->clk_per);
+free_sdhci:
+	if (imx_data->socdata->flags & ESDHC_FLAG_PMQOS)
+		cpu_latency_qos_remove_request(&imx_data->pm_qos_req);
+	sdhci_pltfm_free(pdev);
+	return err;
+}
+
+static int sdhci_esdhc_imx_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	int dead;
+
+	pm_runtime_get_sync(&pdev->dev);
+	dead = (readl(host->ioaddr + SDHCI_INT_STATUS) == 0xffffffff);
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+
+	sdhci_remove_host(host, dead);
+
+	clk_disable_unprepare(imx_data->clk_per);
+	clk_disable_unprepare(imx_data->clk_ipg);
+	clk_disable_unprepare(imx_data->clk_ahb);
+
+	if (imx_data->socdata->flags & ESDHC_FLAG_PMQOS)
+		cpu_latency_qos_remove_request(&imx_data->pm_qos_req);
+
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int sdhci_esdhc_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	if (host->mmc->caps2 & MMC_CAP2_CQE) {
+		ret = cqhci_suspend(host->mmc);
+		if (ret)
+			return ret;
+	}
+
+	if ((imx_data->socdata->flags & ESDHC_FLAG_STATE_LOST_IN_LPMODE) &&
+		(host->tuning_mode != SDHCI_TUNING_MODE_1)) {
+		mmc_retune_timer_stop(host->mmc);
+		mmc_retune_needed(host->mmc);
+	}
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	ret = sdhci_suspend_host(host);
+	if (ret)
+		return ret;
+
+	ret = pinctrl_pm_select_sleep_state(dev);
+	if (ret)
+		return ret;
+
+	ret = mmc_gpio_set_cd_wake(host->mmc, true);
+
+	return ret;
+}
+
+static int sdhci_esdhc_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	int ret;
+
+	ret = pinctrl_pm_select_default_state(dev);
+	if (ret)
+		return ret;
+
+	/* re-initialize hw state in case it's lost in low power mode */
+	sdhci_esdhc_imx_hwinit(host);
+
+	ret = sdhci_resume_host(host);
+	if (ret)
+		return ret;
+
+	if (host->mmc->caps2 & MMC_CAP2_CQE)
+		ret = cqhci_resume(host->mmc);
+
+	if (!ret)
+		ret = mmc_gpio_set_cd_wake(host->mmc, false);
+
+	return ret;
+}
+#endif
+
+#ifdef CONFIG_PM
+static int sdhci_esdhc_runtime_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	if (host->mmc->caps2 & MMC_CAP2_CQE) {
+		ret = cqhci_suspend(host->mmc);
+		if (ret)
+			return ret;
+	}
+
+	ret = sdhci_runtime_suspend_host(host);
+	if (ret)
+		return ret;
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	imx_data->actual_clock = host->mmc->actual_clock;
+	esdhc_pltfm_set_clock(host, 0);
+	clk_disable_unprepare(imx_data->clk_per);
+	clk_disable_unprepare(imx_data->clk_ipg);
+	clk_disable_unprepare(imx_data->clk_ahb);
+
+	if (imx_data->socdata->flags & ESDHC_FLAG_PMQOS)
+		cpu_latency_qos_remove_request(&imx_data->pm_qos_req);
+
+	return ret;
+}
+
+static int sdhci_esdhc_runtime_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = sdhci_pltfm_priv(pltfm_host);
+	int err;
+
+	if (imx_data->socdata->flags & ESDHC_FLAG_PMQOS)
+		cpu_latency_qos_add_request(&imx_data->pm_qos_req, 0);
+
+	if (imx_data->socdata->flags & ESDHC_FLAG_CLK_RATE_LOST_IN_PM_RUNTIME)
+		clk_set_rate(imx_data->clk_per, pltfm_host->clock);
+
+	err = clk_prepare_enable(imx_data->clk_ahb);
+	if (err)
+		goto remove_pm_qos_request;
+
+	err = clk_prepare_enable(imx_data->clk_per);
+	if (err)
+		goto disable_ahb_clk;
+
+	err = clk_prepare_enable(imx_data->clk_ipg);
+	if (err)
+		goto disable_per_clk;
+
+	esdhc_pltfm_set_clock(host, imx_data->actual_clock);
+
+	err = sdhci_runtime_resume_host(host, 0);
+	if (err)
+		goto disable_ipg_clk;
+
+	if (host->mmc->caps2 & MMC_CAP2_CQE)
+		err = cqhci_resume(host->mmc);
+
+	return err;
+
+disable_ipg_clk:
+	clk_disable_unprepare(imx_data->clk_ipg);
+disable_per_clk:
+	clk_disable_unprepare(imx_data->clk_per);
+disable_ahb_clk:
+	clk_disable_unprepare(imx_data->clk_ahb);
+remove_pm_qos_request:
+	if (imx_data->socdata->flags & ESDHC_FLAG_PMQOS)
+		cpu_latency_qos_remove_request(&imx_data->pm_qos_req);
+	return err;
+}
+#endif
+
+static const struct dev_pm_ops sdhci_esdhc_pmops = {
+	SET_SYSTEM_SLEEP_PM_OPS(sdhci_esdhc_suspend, sdhci_esdhc_resume)
+	SET_RUNTIME_PM_OPS(sdhci_esdhc_runtime_suspend,
+				sdhci_esdhc_runtime_resume, NULL)
+};
+
+static struct platform_driver sdhci_esdhc_imx_driver = {
+	.driver		= {
+		.name	= "sdhci-esdhc-imx",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = imx_esdhc_dt_ids,
+		.pm	= &sdhci_esdhc_pmops,
+	},
+	.probe		= sdhci_esdhc_imx_probe,
+	.remove		= sdhci_esdhc_imx_remove,
+};
+
+module_platform_driver(sdhci_esdhc_imx_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for Freescale i.MX eSDHC");
+MODULE_AUTHOR("Wolfram Sang <kernel@pengutronix.de>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-esdhc-mcf.c b/drivers/mmc/host/sdhci-esdhc-mcf.c
new file mode 100644
index 000000000..ca7a1690b
--- /dev/null
+++ b/drivers/mmc/host/sdhci-esdhc-mcf.c
@@ -0,0 +1,522 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Freescale eSDHC ColdFire family controller driver, platform bus.
+ *
+ * Copyright (c) 2020 Timesys Corporation
+ *   Author: Angelo Dureghello <angelo.dureghello@timesys.it>
+ */
+
+#include <linux/module.h>
+#include <linux/delay.h>
+#include <linux/platform_data/mmc-esdhc-mcf.h>
+#include <linux/mmc/mmc.h>
+#include "sdhci-pltfm.h"
+#include "sdhci-esdhc.h"
+
+#define	ESDHC_PROCTL_D3CD		0x08
+#define ESDHC_SYS_CTRL_DTOCV_MASK	0x0f
+#define ESDHC_DEFAULT_HOST_CONTROL	0x28
+
+/*
+ * Freescale eSDHC has DMA ERR flag at bit 28, not as std spec says, bit 25.
+ */
+#define ESDHC_INT_VENDOR_SPEC_DMA_ERR	BIT(28)
+
+struct pltfm_mcf_data {
+	struct clk *clk_ipg;
+	struct clk *clk_ahb;
+	struct clk *clk_per;
+	int aside;
+	int current_bus_width;
+};
+
+static inline void esdhc_mcf_buffer_swap32(u32 *buf, int len)
+{
+	int i;
+	u32 temp;
+
+	len = (len + 3) >> 2;
+
+	for (i = 0; i < len;  i++) {
+		temp = swab32(*buf);
+		*buf++ = temp;
+	}
+}
+
+static inline void esdhc_clrset_be(struct sdhci_host *host,
+				   u32 mask, u32 val, int reg)
+{
+	void __iomem *base = host->ioaddr + (reg & ~3);
+	u8 shift = (reg & 3) << 3;
+
+	mask <<= shift;
+	val <<= shift;
+
+	if (reg == SDHCI_HOST_CONTROL)
+		val |= ESDHC_PROCTL_D3CD;
+
+	writel((readl(base) & ~mask) | val, base);
+}
+
+/*
+ * Note: mcf is big-endian, single bytes need to be accessed at big endian
+ * offsets.
+ */
+static void esdhc_mcf_writeb_be(struct sdhci_host *host, u8 val, int reg)
+{
+	void __iomem *base = host->ioaddr + (reg & ~3);
+	u8 shift = (reg & 3) << 3;
+	u32 mask = ~(0xff << shift);
+
+	if (reg == SDHCI_HOST_CONTROL) {
+		u32 host_ctrl = ESDHC_DEFAULT_HOST_CONTROL;
+		u8 dma_bits = (val & SDHCI_CTRL_DMA_MASK) >> 3;
+		u8 tmp = readb(host->ioaddr + SDHCI_HOST_CONTROL + 1);
+
+		tmp &= ~0x03;
+		tmp |= dma_bits;
+
+		/*
+		 * Recomposition needed, restore always endianness and
+		 * keep D3CD and AI, just setting bus width.
+		 */
+		host_ctrl |= val;
+		host_ctrl |= (dma_bits << 8);
+		writel(host_ctrl, host->ioaddr + SDHCI_HOST_CONTROL);
+
+		return;
+	}
+
+	writel((readl(base) & mask) | (val << shift), base);
+}
+
+static void esdhc_mcf_writew_be(struct sdhci_host *host, u16 val, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_mcf_data *mcf_data = sdhci_pltfm_priv(pltfm_host);
+	void __iomem *base = host->ioaddr + (reg & ~3);
+	u8 shift = (reg & 3) << 3;
+	u32 mask = ~(0xffff << shift);
+
+	switch (reg) {
+	case SDHCI_TRANSFER_MODE:
+		mcf_data->aside = val;
+		return;
+	case SDHCI_COMMAND:
+		if (host->cmd->opcode == MMC_STOP_TRANSMISSION)
+			val |= SDHCI_CMD_ABORTCMD;
+
+		/*
+		 * As for the fsl driver,
+		 * we have to set the mode in a single write here.
+		 */
+		writel(val << 16 | mcf_data->aside,
+		       host->ioaddr + SDHCI_TRANSFER_MODE);
+		return;
+	}
+
+	writel((readl(base) & mask) | (val << shift), base);
+}
+
+static void esdhc_mcf_writel_be(struct sdhci_host *host, u32 val, int reg)
+{
+	writel(val, host->ioaddr + reg);
+}
+
+static u8 esdhc_mcf_readb_be(struct sdhci_host *host, int reg)
+{
+	if (reg == SDHCI_HOST_CONTROL) {
+		u8 __iomem *base = host->ioaddr + (reg & ~3);
+		u16 val = readw(base + 2);
+		u8 dma_bits = (val >> 5) & SDHCI_CTRL_DMA_MASK;
+		u8 host_ctrl = val & 0xff;
+
+		host_ctrl &= ~SDHCI_CTRL_DMA_MASK;
+		host_ctrl |= dma_bits;
+
+		return host_ctrl;
+	}
+
+	return readb(host->ioaddr + (reg ^ 0x3));
+}
+
+static u16 esdhc_mcf_readw_be(struct sdhci_host *host, int reg)
+{
+	/*
+	 * For SDHCI_HOST_VERSION, sdhci specs defines 0xFE,
+	 * a wrong offset for us, we are at 0xFC.
+	 */
+	if (reg == SDHCI_HOST_VERSION)
+		reg -= 2;
+
+	return readw(host->ioaddr + (reg ^ 0x2));
+}
+
+static u32 esdhc_mcf_readl_be(struct sdhci_host *host, int reg)
+{
+	u32 val;
+
+	val = readl(host->ioaddr + reg);
+
+	/*
+	 * RM (25.3.9) sd pin clock must never exceed 25Mhz.
+	 * So forcing legacy mode at 25Mhz.
+	 */
+	if (unlikely(reg == SDHCI_CAPABILITIES))
+		val &= ~SDHCI_CAN_DO_HISPD;
+
+	if (unlikely(reg == SDHCI_INT_STATUS)) {
+		if (val & ESDHC_INT_VENDOR_SPEC_DMA_ERR) {
+			val &= ~ESDHC_INT_VENDOR_SPEC_DMA_ERR;
+			val |= SDHCI_INT_ADMA_ERROR;
+		}
+	}
+
+	return val;
+}
+
+static unsigned int esdhc_mcf_get_max_timeout_count(struct sdhci_host *host)
+{
+	return 1 << 27;
+}
+
+static void esdhc_mcf_set_timeout(struct sdhci_host *host,
+				  struct mmc_command *cmd)
+{
+	/* Use maximum timeout counter */
+	esdhc_clrset_be(host, ESDHC_SYS_CTRL_DTOCV_MASK, 0xE,
+			SDHCI_TIMEOUT_CONTROL);
+}
+
+static void esdhc_mcf_reset(struct sdhci_host *host, u8 mask)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_mcf_data *mcf_data = sdhci_pltfm_priv(pltfm_host);
+
+	sdhci_reset(host, mask);
+
+	esdhc_clrset_be(host, ESDHC_CTRL_BUSWIDTH_MASK,
+			mcf_data->current_bus_width, SDHCI_HOST_CONTROL);
+
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+}
+
+static unsigned int esdhc_mcf_pltfm_get_max_clock(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	return pltfm_host->clock;
+}
+
+static unsigned int esdhc_mcf_pltfm_get_min_clock(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	return pltfm_host->clock / 256 / 16;
+}
+
+static void esdhc_mcf_pltfm_set_clock(struct sdhci_host *host,
+				      unsigned int clock)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	unsigned long *pll_dr = (unsigned long *)MCF_PLL_DR;
+	u32 fvco, fsys, fesdhc, temp;
+	const int sdclkfs[] = {2, 4, 8, 16, 32, 64, 128, 256};
+	int delta, old_delta = clock;
+	int i, q, ri, rq;
+
+	if (clock == 0) {
+		host->mmc->actual_clock = 0;
+		return;
+	}
+
+	/*
+	 * ColdFire eSDHC clock.s
+	 *
+	 * pll -+-> / outdiv1 --> fsys
+	 *      +-> / outdiv3 --> eSDHC clock ---> / SDCCLKFS / DVS
+	 *
+	 * mcf5441x datasheet says:
+	 * (8.1.2) eSDHC should be 40 MHz max
+	 * (25.3.9) eSDHC input is, as example, 96 Mhz ...
+	 * (25.3.9) sd pin clock must never exceed 25Mhz
+	 *
+	 * fvco = fsys * outdvi1 + 1
+	 * fshdc = fvco / outdiv3 + 1
+	 */
+	temp = readl(pll_dr);
+	fsys = pltfm_host->clock;
+	fvco = fsys * ((temp & 0x1f) + 1);
+	fesdhc = fvco / (((temp >> 10) & 0x1f) + 1);
+
+	for (i = 0; i < 8; ++i) {
+		int result = fesdhc / sdclkfs[i];
+
+		for (q = 1; q < 17; ++q) {
+			int finale = result / q;
+
+			delta = abs(clock - finale);
+
+			if (delta < old_delta) {
+				old_delta = delta;
+				ri = i;
+				rq = q;
+			}
+		}
+	}
+
+	/*
+	 * Apply divisors and re-enable all the clocks
+	 */
+	temp = ((sdclkfs[ri] >> 1) << 8) | ((rq - 1) << 4) |
+		(ESDHC_CLOCK_IPGEN | ESDHC_CLOCK_HCKEN | ESDHC_CLOCK_PEREN);
+	esdhc_clrset_be(host, 0x0000fff7, temp, SDHCI_CLOCK_CONTROL);
+
+	host->mmc->actual_clock = clock;
+
+	mdelay(1);
+}
+
+static void esdhc_mcf_pltfm_set_bus_width(struct sdhci_host *host, int width)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_mcf_data *mcf_data = sdhci_pltfm_priv(pltfm_host);
+
+	switch (width) {
+	case MMC_BUS_WIDTH_4:
+		mcf_data->current_bus_width = ESDHC_CTRL_4BITBUS;
+		break;
+	default:
+		mcf_data->current_bus_width = 0;
+		break;
+	}
+
+	esdhc_clrset_be(host, ESDHC_CTRL_BUSWIDTH_MASK,
+			mcf_data->current_bus_width, SDHCI_HOST_CONTROL);
+}
+
+static void esdhc_mcf_request_done(struct sdhci_host *host,
+				   struct mmc_request *mrq)
+{
+	struct scatterlist *sg;
+	u32 *buffer;
+	int i;
+
+	if (!mrq->data || !mrq->data->bytes_xfered)
+		goto exit_done;
+
+	if (mmc_get_dma_dir(mrq->data) != DMA_FROM_DEVICE)
+		goto exit_done;
+
+	/*
+	 * On mcf5441x there is no hw sdma option/flag to select the dma
+	 * transfer endiannes. A swap after the transfer is needed.
+	 */
+	for_each_sg(mrq->data->sg, sg, mrq->data->sg_len, i) {
+		buffer = (u32 *)sg_virt(sg);
+		esdhc_mcf_buffer_swap32(buffer, sg->length);
+	}
+
+exit_done:
+	mmc_request_done(host->mmc, mrq);
+}
+
+static void esdhc_mcf_copy_to_bounce_buffer(struct sdhci_host *host,
+					    struct mmc_data *data,
+					    unsigned int length)
+{
+	sg_copy_to_buffer(data->sg, data->sg_len,
+			  host->bounce_buffer, length);
+
+	esdhc_mcf_buffer_swap32((u32 *)host->bounce_buffer,
+				data->blksz * data->blocks);
+}
+
+static struct sdhci_ops sdhci_esdhc_ops = {
+	.reset = esdhc_mcf_reset,
+	.set_clock = esdhc_mcf_pltfm_set_clock,
+	.get_max_clock = esdhc_mcf_pltfm_get_max_clock,
+	.get_min_clock = esdhc_mcf_pltfm_get_min_clock,
+	.set_bus_width = esdhc_mcf_pltfm_set_bus_width,
+	.get_max_timeout_count = esdhc_mcf_get_max_timeout_count,
+	.set_timeout = esdhc_mcf_set_timeout,
+	.write_b = esdhc_mcf_writeb_be,
+	.write_w = esdhc_mcf_writew_be,
+	.write_l = esdhc_mcf_writel_be,
+	.read_b = esdhc_mcf_readb_be,
+	.read_w = esdhc_mcf_readw_be,
+	.read_l = esdhc_mcf_readl_be,
+	.copy_to_bounce_buffer = esdhc_mcf_copy_to_bounce_buffer,
+	.request_done = esdhc_mcf_request_done,
+};
+
+static const struct sdhci_pltfm_data sdhci_esdhc_mcf_pdata = {
+	.ops = &sdhci_esdhc_ops,
+	.quirks = ESDHC_DEFAULT_QUIRKS | SDHCI_QUIRK_FORCE_DMA,
+		 /*
+		  * Mandatory quirk,
+		  * controller does not support cmd23,
+		  * without, on > 8G cards cmd23 is used, and
+		  * driver times out.
+		  */
+		  SDHCI_QUIRK2_HOST_NO_CMD23,
+};
+
+static int esdhc_mcf_plat_init(struct sdhci_host *host,
+			       struct pltfm_mcf_data *mcf_data)
+{
+	struct mcf_esdhc_platform_data *plat_data;
+
+	if (!host->mmc->parent->platform_data) {
+		dev_err(mmc_dev(host->mmc), "no platform data!\n");
+		return -EINVAL;
+	}
+
+	plat_data = (struct mcf_esdhc_platform_data *)
+			host->mmc->parent->platform_data;
+
+	/* Card_detect */
+	switch (plat_data->cd_type) {
+	default:
+	case ESDHC_CD_CONTROLLER:
+		/* We have a working card_detect back */
+		host->quirks &= ~SDHCI_QUIRK_BROKEN_CARD_DETECTION;
+		break;
+	case ESDHC_CD_PERMANENT:
+		host->mmc->caps |= MMC_CAP_NONREMOVABLE;
+		break;
+	case ESDHC_CD_NONE:
+		break;
+	}
+
+	switch (plat_data->max_bus_width) {
+	case 4:
+		host->mmc->caps |= MMC_CAP_4_BIT_DATA;
+		break;
+	case 1:
+	default:
+		host->quirks |= SDHCI_QUIRK_FORCE_1_BIT_DATA;
+		break;
+	}
+
+	return 0;
+}
+
+static int sdhci_esdhc_mcf_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct pltfm_mcf_data *mcf_data;
+	int err;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_esdhc_mcf_pdata,
+				sizeof(*mcf_data));
+
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	mcf_data = sdhci_pltfm_priv(pltfm_host);
+
+	host->sdma_boundary = 0;
+
+	host->flags |= SDHCI_AUTO_CMD12;
+
+	mcf_data->clk_ipg = devm_clk_get(&pdev->dev, "ipg");
+	if (IS_ERR(mcf_data->clk_ipg)) {
+		err = PTR_ERR(mcf_data->clk_ipg);
+		goto err_exit;
+	}
+
+	mcf_data->clk_ahb = devm_clk_get(&pdev->dev, "ahb");
+	if (IS_ERR(mcf_data->clk_ahb)) {
+		err = PTR_ERR(mcf_data->clk_ahb);
+		goto err_exit;
+	}
+
+	mcf_data->clk_per = devm_clk_get(&pdev->dev, "per");
+	if (IS_ERR(mcf_data->clk_per)) {
+		err = PTR_ERR(mcf_data->clk_per);
+		goto err_exit;
+	}
+
+	pltfm_host->clk = mcf_data->clk_per;
+	pltfm_host->clock = clk_get_rate(pltfm_host->clk);
+	err = clk_prepare_enable(mcf_data->clk_per);
+	if (err)
+		goto err_exit;
+
+	err = clk_prepare_enable(mcf_data->clk_ipg);
+	if (err)
+		goto unprep_per;
+
+	err = clk_prepare_enable(mcf_data->clk_ahb);
+	if (err)
+		goto unprep_ipg;
+
+	err = esdhc_mcf_plat_init(host, mcf_data);
+	if (err)
+		goto unprep_ahb;
+
+	err = sdhci_setup_host(host);
+	if (err)
+		goto unprep_ahb;
+
+	if (!host->bounce_buffer) {
+		dev_err(&pdev->dev, "bounce buffer not allocated");
+		err = -ENOMEM;
+		goto cleanup;
+	}
+
+	err = __sdhci_add_host(host);
+	if (err)
+		goto cleanup;
+
+	return 0;
+
+cleanup:
+	sdhci_cleanup_host(host);
+unprep_ahb:
+	clk_disable_unprepare(mcf_data->clk_ahb);
+unprep_ipg:
+	clk_disable_unprepare(mcf_data->clk_ipg);
+unprep_per:
+	clk_disable_unprepare(mcf_data->clk_per);
+err_exit:
+	sdhci_pltfm_free(pdev);
+
+	return err;
+}
+
+static int sdhci_esdhc_mcf_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_mcf_data *mcf_data = sdhci_pltfm_priv(pltfm_host);
+
+	sdhci_remove_host(host, 0);
+
+	clk_disable_unprepare(mcf_data->clk_ipg);
+	clk_disable_unprepare(mcf_data->clk_ahb);
+	clk_disable_unprepare(mcf_data->clk_per);
+
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+
+static struct platform_driver sdhci_esdhc_mcf_driver = {
+	.driver	= {
+		.name = "sdhci-esdhc-mcf",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+	},
+	.probe = sdhci_esdhc_mcf_probe,
+	.remove = sdhci_esdhc_mcf_remove,
+};
+
+module_platform_driver(sdhci_esdhc_mcf_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for Freescale ColdFire eSDHC");
+MODULE_AUTHOR("Angelo Dureghello <angelo.dureghello@timesys.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-esdhc.h b/drivers/mmc/host/sdhci-esdhc.h
new file mode 100644
index 000000000..6de02f09c
--- /dev/null
+++ b/drivers/mmc/host/sdhci-esdhc.h
@@ -0,0 +1,109 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Freescale eSDHC controller driver generics for OF and pltfm.
+ *
+ * Copyright (c) 2007 Freescale Semiconductor, Inc.
+ * Copyright (c) 2009 MontaVista Software, Inc.
+ * Copyright (c) 2010 Pengutronix e.K.
+ * Copyright 2020 NXP
+ *   Author: Wolfram Sang <kernel@pengutronix.de>
+ */
+
+#ifndef _DRIVERS_MMC_SDHCI_ESDHC_H
+#define _DRIVERS_MMC_SDHCI_ESDHC_H
+
+/*
+ * Ops and quirks for the Freescale eSDHC controller.
+ */
+
+#define ESDHC_DEFAULT_QUIRKS	(SDHCI_QUIRK_FORCE_BLK_SZ_2048 | \
+				SDHCI_QUIRK_32BIT_DMA_ADDR | \
+				SDHCI_QUIRK_NO_BUSY_IRQ | \
+				SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK | \
+				SDHCI_QUIRK_PIO_NEEDS_DELAY | \
+				SDHCI_QUIRK_NO_HISPD_BIT)
+
+/* pltfm-specific */
+#define ESDHC_HOST_CONTROL_LE	0x20
+
+/*
+ * eSDHC register definition
+ */
+
+/* Present State Register */
+#define ESDHC_PRSSTAT			0x24
+#define ESDHC_CLOCK_GATE_OFF		0x00000080
+#define ESDHC_CLOCK_STABLE		0x00000008
+
+/* Protocol Control Register */
+#define ESDHC_PROCTL			0x28
+#define ESDHC_VOLT_SEL			0x00000400
+#define ESDHC_CTRL_4BITBUS		(0x1 << 1)
+#define ESDHC_CTRL_8BITBUS		(0x2 << 1)
+#define ESDHC_CTRL_BUSWIDTH_MASK	(0x3 << 1)
+#define ESDHC_HOST_CONTROL_RES		0x01
+
+/* System Control Register */
+#define ESDHC_SYSTEM_CONTROL		0x2c
+#define ESDHC_CLOCK_MASK		0x0000fff0
+#define ESDHC_PREDIV_SHIFT		8
+#define ESDHC_DIVIDER_SHIFT		4
+#define ESDHC_CLOCK_SDCLKEN		0x00000008
+#define ESDHC_CLOCK_PEREN		0x00000004
+#define ESDHC_CLOCK_HCKEN		0x00000002
+#define ESDHC_CLOCK_IPGEN		0x00000001
+
+/* System Control 2 Register */
+#define ESDHC_SYSTEM_CONTROL_2		0x3c
+#define ESDHC_SMPCLKSEL			0x00800000
+#define ESDHC_EXTN			0x00400000
+
+/* Host Controller Capabilities Register 2 */
+#define ESDHC_CAPABILITIES_1		0x114
+
+/* Tuning Block Control Register */
+#define ESDHC_TBCTL			0x120
+#define ESDHC_HS400_WNDW_ADJUST		0x00000040
+#define ESDHC_HS400_MODE		0x00000010
+#define ESDHC_TB_EN			0x00000004
+#define ESDHC_TB_MODE_MASK		0x00000003
+#define ESDHC_TB_MODE_SW		0x00000003
+#define ESDHC_TB_MODE_3			0x00000002
+
+#define ESDHC_TBSTAT			0x124
+
+#define ESDHC_TBPTR			0x128
+#define ESDHC_WNDW_STRT_PTR_SHIFT	8
+#define ESDHC_WNDW_STRT_PTR_MASK	(0x7f << 8)
+#define ESDHC_WNDW_END_PTR_MASK		0x7f
+
+/* SD Clock Control Register */
+#define ESDHC_SDCLKCTL			0x144
+#define ESDHC_LPBK_CLK_SEL		0x80000000
+#define ESDHC_CMD_CLK_CTL		0x00008000
+
+/* SD Timing Control Register */
+#define ESDHC_SDTIMNGCTL		0x148
+#define ESDHC_FLW_CTL_BG		0x00008000
+
+/* DLL Config 0 Register */
+#define ESDHC_DLLCFG0			0x160
+#define ESDHC_DLL_ENABLE		0x80000000
+#define ESDHC_DLL_RESET			0x40000000
+#define ESDHC_DLL_FREQ_SEL		0x08000000
+
+/* DLL Config 1 Register */
+#define ESDHC_DLLCFG1			0x164
+#define ESDHC_DLL_PD_PULSE_STRETCH_SEL	0x80000000
+
+/* DLL Status 0 Register */
+#define ESDHC_DLLSTAT0			0x170
+#define ESDHC_DLL_STS_SLV_LOCK		0x08000000
+
+/* Control Register for DMA transfer */
+#define ESDHC_DMA_SYSCTL		0x40c
+#define ESDHC_PERIPHERAL_CLK_SEL	0x00080000
+#define ESDHC_FLUSH_ASYNC_FIFO		0x00040000
+#define ESDHC_DMA_SNOOP			0x00000040
+
+#endif /* _DRIVERS_MMC_SDHCI_ESDHC_H */
diff --git a/drivers/mmc/host/sdhci-iproc.c b/drivers/mmc/host/sdhci-iproc.c
new file mode 100644
index 000000000..b9eb2ec61
--- /dev/null
+++ b/drivers/mmc/host/sdhci-iproc.c
@@ -0,0 +1,420 @@
+/*
+ * Copyright (C) 2014 Broadcom Corporation
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+/*
+ * iProc SDHCI platform driver
+ */
+
+#include <linux/acpi.h>
+#include <linux/delay.h>
+#include <linux/module.h>
+#include <linux/mmc/host.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include "sdhci-pltfm.h"
+
+struct sdhci_iproc_data {
+	const struct sdhci_pltfm_data *pdata;
+	u32 caps;
+	u32 caps1;
+	u32 mmc_caps;
+};
+
+struct sdhci_iproc_host {
+	const struct sdhci_iproc_data *data;
+	u32 shadow_cmd;
+	u32 shadow_blk;
+	bool is_cmd_shadowed;
+	bool is_blk_shadowed;
+};
+
+#define REG_OFFSET_IN_BITS(reg) ((reg) << 3 & 0x18)
+
+static inline u32 sdhci_iproc_readl(struct sdhci_host *host, int reg)
+{
+	u32 val = readl(host->ioaddr + reg);
+
+	pr_debug("%s: readl [0x%02x] 0x%08x\n",
+		 mmc_hostname(host->mmc), reg, val);
+	return val;
+}
+
+static u16 sdhci_iproc_readw(struct sdhci_host *host, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_iproc_host *iproc_host = sdhci_pltfm_priv(pltfm_host);
+	u32 val;
+	u16 word;
+
+	if ((reg == SDHCI_TRANSFER_MODE) && iproc_host->is_cmd_shadowed) {
+		/* Get the saved transfer mode */
+		val = iproc_host->shadow_cmd;
+	} else if ((reg == SDHCI_BLOCK_SIZE || reg == SDHCI_BLOCK_COUNT) &&
+		   iproc_host->is_blk_shadowed) {
+		/* Get the saved block info */
+		val = iproc_host->shadow_blk;
+	} else {
+		val = sdhci_iproc_readl(host, (reg & ~3));
+	}
+	word = val >> REG_OFFSET_IN_BITS(reg) & 0xffff;
+	return word;
+}
+
+static u8 sdhci_iproc_readb(struct sdhci_host *host, int reg)
+{
+	u32 val = sdhci_iproc_readl(host, (reg & ~3));
+	u8 byte = val >> REG_OFFSET_IN_BITS(reg) & 0xff;
+	return byte;
+}
+
+static inline void sdhci_iproc_writel(struct sdhci_host *host, u32 val, int reg)
+{
+	pr_debug("%s: writel [0x%02x] 0x%08x\n",
+		 mmc_hostname(host->mmc), reg, val);
+
+	writel(val, host->ioaddr + reg);
+
+	if (host->clock <= 400000) {
+		/* Round up to micro-second four SD clock delay */
+		if (host->clock)
+			udelay((4 * 1000000 + host->clock - 1) / host->clock);
+		else
+			udelay(10);
+	}
+}
+
+/*
+ * The Arasan has a bugette whereby it may lose the content of successive
+ * writes to the same register that are within two SD-card clock cycles of
+ * each other (a clock domain crossing problem). The data
+ * register does not have this problem, which is just as well - otherwise we'd
+ * have to nobble the DMA engine too.
+ *
+ * This wouldn't be a problem with the code except that we can only write the
+ * controller with 32-bit writes.  So two different 16-bit registers are
+ * written back to back creates the problem.
+ *
+ * In reality, this only happens when SDHCI_BLOCK_SIZE and SDHCI_BLOCK_COUNT
+ * are written followed by SDHCI_TRANSFER_MODE and SDHCI_COMMAND.
+ * The BLOCK_SIZE and BLOCK_COUNT are meaningless until a command issued so
+ * the work around can be further optimized. We can keep shadow values of
+ * BLOCK_SIZE, BLOCK_COUNT, and TRANSFER_MODE until a COMMAND is issued.
+ * Then, write the BLOCK_SIZE+BLOCK_COUNT in a single 32-bit write followed
+ * by the TRANSFER+COMMAND in another 32-bit write.
+ */
+static void sdhci_iproc_writew(struct sdhci_host *host, u16 val, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_iproc_host *iproc_host = sdhci_pltfm_priv(pltfm_host);
+	u32 word_shift = REG_OFFSET_IN_BITS(reg);
+	u32 mask = 0xffff << word_shift;
+	u32 oldval, newval;
+
+	if (reg == SDHCI_COMMAND) {
+		/* Write the block now as we are issuing a command */
+		if (iproc_host->is_blk_shadowed) {
+			sdhci_iproc_writel(host, iproc_host->shadow_blk,
+				SDHCI_BLOCK_SIZE);
+			iproc_host->is_blk_shadowed = false;
+		}
+		oldval = iproc_host->shadow_cmd;
+		iproc_host->is_cmd_shadowed = false;
+	} else if ((reg == SDHCI_BLOCK_SIZE || reg == SDHCI_BLOCK_COUNT) &&
+		   iproc_host->is_blk_shadowed) {
+		/* Block size and count are stored in shadow reg */
+		oldval = iproc_host->shadow_blk;
+	} else {
+		/* Read reg, all other registers are not shadowed */
+		oldval = sdhci_iproc_readl(host, (reg & ~3));
+	}
+	newval = (oldval & ~mask) | (val << word_shift);
+
+	if (reg == SDHCI_TRANSFER_MODE) {
+		/* Save the transfer mode until the command is issued */
+		iproc_host->shadow_cmd = newval;
+		iproc_host->is_cmd_shadowed = true;
+	} else if (reg == SDHCI_BLOCK_SIZE || reg == SDHCI_BLOCK_COUNT) {
+		/* Save the block info until the command is issued */
+		iproc_host->shadow_blk = newval;
+		iproc_host->is_blk_shadowed = true;
+	} else {
+		/* Command or other regular 32-bit write */
+		sdhci_iproc_writel(host, newval, reg & ~3);
+	}
+}
+
+static void sdhci_iproc_writeb(struct sdhci_host *host, u8 val, int reg)
+{
+	u32 oldval = sdhci_iproc_readl(host, (reg & ~3));
+	u32 byte_shift = REG_OFFSET_IN_BITS(reg);
+	u32 mask = 0xff << byte_shift;
+	u32 newval = (oldval & ~mask) | (val << byte_shift);
+
+	sdhci_iproc_writel(host, newval, reg & ~3);
+}
+
+static unsigned int sdhci_iproc_get_max_clock(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	if (pltfm_host->clk)
+		return sdhci_pltfm_clk_get_max_clock(host);
+	else
+		return pltfm_host->clock;
+}
+
+/*
+ * There is a known bug on BCM2711's SDHCI core integration where the
+ * controller will hang when the difference between the core clock and the bus
+ * clock is too great. Specifically this can be reproduced under the following
+ * conditions:
+ *
+ *  - No SD card plugged in, polling thread is running, probing cards at
+ *    100 kHz.
+ *  - BCM2711's core clock configured at 500MHz or more
+ *
+ * So we set 200kHz as the minimum clock frequency available for that SoC.
+ */
+static unsigned int sdhci_iproc_bcm2711_get_min_clock(struct sdhci_host *host)
+{
+	return 200000;
+}
+
+static const struct sdhci_ops sdhci_iproc_ops = {
+	.set_clock = sdhci_set_clock,
+	.get_max_clock = sdhci_iproc_get_max_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static const struct sdhci_ops sdhci_iproc_32only_ops = {
+	.read_l = sdhci_iproc_readl,
+	.read_w = sdhci_iproc_readw,
+	.read_b = sdhci_iproc_readb,
+	.write_l = sdhci_iproc_writel,
+	.write_w = sdhci_iproc_writew,
+	.write_b = sdhci_iproc_writeb,
+	.set_clock = sdhci_set_clock,
+	.get_max_clock = sdhci_iproc_get_max_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static const struct sdhci_pltfm_data sdhci_iproc_cygnus_pltfm_data = {
+	.quirks = SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK |
+		  SDHCI_QUIRK_NO_HISPD_BIT,
+	.quirks2 = SDHCI_QUIRK2_ACMD23_BROKEN | SDHCI_QUIRK2_HOST_OFF_CARD_ON,
+	.ops = &sdhci_iproc_32only_ops,
+};
+
+static const struct sdhci_iproc_data iproc_cygnus_data = {
+	.pdata = &sdhci_iproc_cygnus_pltfm_data,
+	.caps = ((0x1 << SDHCI_MAX_BLOCK_SHIFT)
+			& SDHCI_MAX_BLOCK_MASK) |
+		SDHCI_CAN_VDD_330 |
+		SDHCI_CAN_VDD_180 |
+		SDHCI_CAN_DO_SUSPEND |
+		SDHCI_CAN_DO_HISPD |
+		SDHCI_CAN_DO_ADMA2 |
+		SDHCI_CAN_DO_SDMA,
+	.caps1 = SDHCI_DRIVER_TYPE_C |
+		 SDHCI_DRIVER_TYPE_D |
+		 SDHCI_SUPPORT_DDR50,
+	.mmc_caps = MMC_CAP_1_8V_DDR,
+};
+
+static const struct sdhci_pltfm_data sdhci_iproc_pltfm_data = {
+	.quirks = SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK |
+		  SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12 |
+		  SDHCI_QUIRK_NO_HISPD_BIT,
+	.quirks2 = SDHCI_QUIRK2_ACMD23_BROKEN,
+	.ops = &sdhci_iproc_ops,
+};
+
+static const struct sdhci_iproc_data iproc_data = {
+	.pdata = &sdhci_iproc_pltfm_data,
+	.caps = ((0x1 << SDHCI_MAX_BLOCK_SHIFT)
+			& SDHCI_MAX_BLOCK_MASK) |
+		SDHCI_CAN_VDD_330 |
+		SDHCI_CAN_VDD_180 |
+		SDHCI_CAN_DO_SUSPEND |
+		SDHCI_CAN_DO_HISPD |
+		SDHCI_CAN_DO_ADMA2 |
+		SDHCI_CAN_DO_SDMA,
+	.caps1 = SDHCI_DRIVER_TYPE_C |
+		 SDHCI_DRIVER_TYPE_D |
+		 SDHCI_SUPPORT_DDR50,
+};
+
+static const struct sdhci_pltfm_data sdhci_bcm2835_pltfm_data = {
+	.quirks = SDHCI_QUIRK_BROKEN_CARD_DETECTION |
+		  SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK |
+		  SDHCI_QUIRK_MISSING_CAPS |
+		  SDHCI_QUIRK_NO_HISPD_BIT,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+	.ops = &sdhci_iproc_32only_ops,
+};
+
+static const struct sdhci_iproc_data bcm2835_data = {
+	.pdata = &sdhci_bcm2835_pltfm_data,
+	.caps = ((0x1 << SDHCI_MAX_BLOCK_SHIFT)
+			& SDHCI_MAX_BLOCK_MASK) |
+		SDHCI_CAN_VDD_330 |
+		SDHCI_CAN_DO_HISPD,
+	.caps1 = SDHCI_DRIVER_TYPE_A |
+		 SDHCI_DRIVER_TYPE_C,
+	.mmc_caps = 0x00000000,
+};
+
+static const struct sdhci_ops sdhci_iproc_bcm2711_ops = {
+	.read_l = sdhci_iproc_readl,
+	.read_w = sdhci_iproc_readw,
+	.read_b = sdhci_iproc_readb,
+	.write_l = sdhci_iproc_writel,
+	.write_w = sdhci_iproc_writew,
+	.write_b = sdhci_iproc_writeb,
+	.set_clock = sdhci_set_clock,
+	.set_power = sdhci_set_power_and_bus_voltage,
+	.get_max_clock = sdhci_iproc_get_max_clock,
+	.get_min_clock = sdhci_iproc_bcm2711_get_min_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static const struct sdhci_pltfm_data sdhci_bcm2711_pltfm_data = {
+	.quirks = SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12,
+	.ops = &sdhci_iproc_bcm2711_ops,
+};
+
+static const struct sdhci_iproc_data bcm2711_data = {
+	.pdata = &sdhci_bcm2711_pltfm_data,
+	.mmc_caps = MMC_CAP_3_3V_DDR,
+};
+
+static const struct of_device_id sdhci_iproc_of_match[] = {
+	{ .compatible = "brcm,bcm2835-sdhci", .data = &bcm2835_data },
+	{ .compatible = "brcm,bcm2711-emmc2", .data = &bcm2711_data },
+	{ .compatible = "brcm,sdhci-iproc-cygnus", .data = &iproc_cygnus_data},
+	{ .compatible = "brcm,sdhci-iproc", .data = &iproc_data },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, sdhci_iproc_of_match);
+
+#ifdef CONFIG_ACPI
+/*
+ * This is a duplicate of bcm2835_(pltfrm_)data without caps quirks
+ * which are provided by the ACPI table.
+ */
+static const struct sdhci_pltfm_data sdhci_bcm_arasan_data = {
+	.quirks = SDHCI_QUIRK_BROKEN_CARD_DETECTION |
+		  SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK |
+		  SDHCI_QUIRK_NO_HISPD_BIT,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+	.ops = &sdhci_iproc_32only_ops,
+};
+
+static const struct sdhci_iproc_data bcm_arasan_data = {
+	.pdata = &sdhci_bcm_arasan_data,
+};
+
+static const struct acpi_device_id sdhci_iproc_acpi_ids[] = {
+	{ .id = "BRCM5871", .driver_data = (kernel_ulong_t)&iproc_cygnus_data },
+	{ .id = "BRCM5872", .driver_data = (kernel_ulong_t)&iproc_data },
+	{ .id = "BCM2847",  .driver_data = (kernel_ulong_t)&bcm_arasan_data },
+	{ .id = "BRCME88C", .driver_data = (kernel_ulong_t)&bcm2711_data },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(acpi, sdhci_iproc_acpi_ids);
+#endif
+
+static int sdhci_iproc_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	const struct sdhci_iproc_data *iproc_data = NULL;
+	struct sdhci_host *host;
+	struct sdhci_iproc_host *iproc_host;
+	struct sdhci_pltfm_host *pltfm_host;
+	int ret;
+
+	iproc_data = device_get_match_data(dev);
+	if (!iproc_data)
+		return -ENODEV;
+
+	host = sdhci_pltfm_init(pdev, iproc_data->pdata, sizeof(*iproc_host));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	iproc_host = sdhci_pltfm_priv(pltfm_host);
+
+	iproc_host->data = iproc_data;
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto err;
+
+	sdhci_get_property(pdev);
+
+	host->mmc->caps |= iproc_host->data->mmc_caps;
+
+	if (dev->of_node) {
+		pltfm_host->clk = devm_clk_get(dev, NULL);
+		if (IS_ERR(pltfm_host->clk)) {
+			ret = PTR_ERR(pltfm_host->clk);
+			goto err;
+		}
+		ret = clk_prepare_enable(pltfm_host->clk);
+		if (ret) {
+			dev_err(dev, "failed to enable host clk\n");
+			goto err;
+		}
+	}
+
+	if (iproc_host->data->pdata->quirks & SDHCI_QUIRK_MISSING_CAPS) {
+		host->caps = iproc_host->data->caps;
+		host->caps1 = iproc_host->data->caps1;
+	}
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err_clk;
+
+	return 0;
+
+err_clk:
+	if (dev->of_node)
+		clk_disable_unprepare(pltfm_host->clk);
+err:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static struct platform_driver sdhci_iproc_driver = {
+	.driver = {
+		.name = "sdhci-iproc",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = sdhci_iproc_of_match,
+		.acpi_match_table = ACPI_PTR(sdhci_iproc_acpi_ids),
+		.pm = &sdhci_pltfm_pmops,
+	},
+	.probe = sdhci_iproc_probe,
+	.remove = sdhci_pltfm_unregister,
+};
+module_platform_driver(sdhci_iproc_driver);
+
+MODULE_AUTHOR("Broadcom");
+MODULE_DESCRIPTION("IPROC SDHCI driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-milbeaut.c b/drivers/mmc/host/sdhci-milbeaut.c
new file mode 100644
index 000000000..148b37ac6
--- /dev/null
+++ b/drivers/mmc/host/sdhci-milbeaut.c
@@ -0,0 +1,348 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2013 - 2015 Fujitsu Semiconductor, Ltd
+ *              Vincent Yang <vincent.yang@tw.fujitsu.com>
+ * Copyright (C) 2015 Linaro Ltd  Andy Green <andy.green@linaro.org>
+ * Copyright (C) 2019 Socionext Inc.
+ *              Takao Orito <orito.takao@socionext.com>
+ */
+
+#include <linux/bits.h>
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/gpio/consumer.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/property.h>
+
+#include "sdhci-pltfm.h"
+#include "sdhci_f_sdh30.h"
+
+/* milbeaut bridge controller register */
+#define MLB_SOFT_RESET		0x0200
+#define  MLB_SOFT_RESET_RSTX		BIT(0)
+
+#define MLB_WP_CD_LED_SET	0x0210
+#define  MLB_WP_CD_LED_SET_LED_INV  BIT(2)
+
+#define MLB_CR_SET			0x0220
+#define  MLB_CR_SET_CR_TOCLKUNIT       BIT(24)
+#define  MLB_CR_SET_CR_TOCLKFREQ_SFT   (16)
+#define  MLB_CR_SET_CR_TOCLKFREQ_MASK  (0x3F << MLB_CR_SET_CR_TOCLKFREQ_SFT)
+#define  MLB_CR_SET_CR_BCLKFREQ_SFT    (8)
+#define  MLB_CR_SET_CR_BCLKFREQ_MASK   (0xFF << MLB_CR_SET_CR_BCLKFREQ_SFT)
+#define  MLB_CR_SET_CR_RTUNTIMER_SFT   (4)
+#define  MLB_CR_SET_CR_RTUNTIMER_MASK  (0xF << MLB_CR_SET_CR_RTUNTIMER_SFT)
+
+#define MLB_SD_TOCLK_I_DIV  16
+#define MLB_TOCLKFREQ_UNIT_THRES    16000000
+#define MLB_CAL_TOCLKFREQ_MHZ(rate) (rate / MLB_SD_TOCLK_I_DIV / 1000000)
+#define MLB_CAL_TOCLKFREQ_KHZ(rate) (rate / MLB_SD_TOCLK_I_DIV / 1000)
+#define MLB_TOCLKFREQ_MAX   63
+#define MLB_TOCLKFREQ_MIN    1
+
+#define MLB_SD_BCLK_I_DIV   4
+#define MLB_CAL_BCLKFREQ(rate)  (rate / MLB_SD_BCLK_I_DIV / 1000000)
+#define MLB_BCLKFREQ_MAX        255
+#define MLB_BCLKFREQ_MIN          1
+
+#define MLB_CDR_SET			0x0230
+#define MLB_CDR_SET_CLK2POW16	3
+
+struct f_sdhost_priv {
+	struct clk *clk_iface;
+	struct clk *clk;
+	struct device *dev;
+	bool enable_cmd_dat_delay;
+};
+
+static void sdhci_milbeaut_soft_voltage_switch(struct sdhci_host *host)
+{
+	u32 ctrl = 0;
+
+	usleep_range(2500, 3000);
+	ctrl = sdhci_readl(host, F_SDH30_IO_CONTROL2);
+	ctrl |= F_SDH30_CRES_O_DN;
+	sdhci_writel(host, ctrl, F_SDH30_IO_CONTROL2);
+	ctrl |= F_SDH30_MSEL_O_1_8;
+	sdhci_writel(host, ctrl, F_SDH30_IO_CONTROL2);
+
+	ctrl &= ~F_SDH30_CRES_O_DN;
+	sdhci_writel(host, ctrl, F_SDH30_IO_CONTROL2);
+	usleep_range(2500, 3000);
+
+	ctrl = sdhci_readl(host, F_SDH30_TUNING_SETTING);
+	ctrl |= F_SDH30_CMD_CHK_DIS;
+	sdhci_writel(host, ctrl, F_SDH30_TUNING_SETTING);
+}
+
+static unsigned int sdhci_milbeaut_get_min_clock(struct sdhci_host *host)
+{
+	return F_SDH30_MIN_CLOCK;
+}
+
+static void sdhci_milbeaut_reset(struct sdhci_host *host, u8 mask)
+{
+	struct f_sdhost_priv *priv = sdhci_priv(host);
+	u16 clk;
+	u32 ctl;
+	ktime_t timeout;
+
+	clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+	clk = (clk & ~SDHCI_CLOCK_CARD_EN) | SDHCI_CLOCK_INT_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+	sdhci_reset(host, mask);
+
+	clk |= SDHCI_CLOCK_CARD_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+	timeout = ktime_add_ms(ktime_get(), 10);
+	while (1) {
+		bool timedout = ktime_after(ktime_get(), timeout);
+
+		clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+		if (clk & SDHCI_CLOCK_INT_STABLE)
+			break;
+		if (timedout) {
+			pr_err("%s: Internal clock never stabilised.\n",
+				mmc_hostname(host->mmc));
+			sdhci_dumpregs(host);
+			return;
+		}
+		udelay(10);
+	}
+
+	if (priv->enable_cmd_dat_delay) {
+		ctl = sdhci_readl(host, F_SDH30_ESD_CONTROL);
+		ctl |= F_SDH30_CMD_DAT_DELAY;
+		sdhci_writel(host, ctl, F_SDH30_ESD_CONTROL);
+	}
+}
+
+static const struct sdhci_ops sdhci_milbeaut_ops = {
+	.voltage_switch = sdhci_milbeaut_soft_voltage_switch,
+	.get_min_clock = sdhci_milbeaut_get_min_clock,
+	.reset = sdhci_milbeaut_reset,
+	.set_clock = sdhci_set_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+	.set_power = sdhci_set_power_and_bus_voltage,
+};
+
+static void sdhci_milbeaut_bridge_reset(struct sdhci_host *host,
+						int reset_flag)
+{
+	if (reset_flag)
+		sdhci_writel(host, 0, MLB_SOFT_RESET);
+	else
+		sdhci_writel(host, MLB_SOFT_RESET_RSTX, MLB_SOFT_RESET);
+}
+
+static void sdhci_milbeaut_bridge_init(struct sdhci_host *host,
+						int rate)
+{
+	u32 val, clk;
+
+	/* IO_SDIO_CR_SET should be set while reset */
+	val = sdhci_readl(host, MLB_CR_SET);
+	val &= ~(MLB_CR_SET_CR_TOCLKFREQ_MASK | MLB_CR_SET_CR_TOCLKUNIT |
+			MLB_CR_SET_CR_BCLKFREQ_MASK);
+	if (rate >= MLB_TOCLKFREQ_UNIT_THRES) {
+		clk = MLB_CAL_TOCLKFREQ_MHZ(rate);
+		clk = min_t(u32, MLB_TOCLKFREQ_MAX, clk);
+		val |= MLB_CR_SET_CR_TOCLKUNIT |
+			(clk << MLB_CR_SET_CR_TOCLKFREQ_SFT);
+	} else {
+		clk = MLB_CAL_TOCLKFREQ_KHZ(rate);
+		clk = min_t(u32, MLB_TOCLKFREQ_MAX, clk);
+		clk = max_t(u32, MLB_TOCLKFREQ_MIN, clk);
+		val |= clk << MLB_CR_SET_CR_TOCLKFREQ_SFT;
+	}
+
+	clk = MLB_CAL_BCLKFREQ(rate);
+	clk = min_t(u32, MLB_BCLKFREQ_MAX, clk);
+	clk = max_t(u32, MLB_BCLKFREQ_MIN, clk);
+	val |=  clk << MLB_CR_SET_CR_BCLKFREQ_SFT;
+	val &= ~MLB_CR_SET_CR_RTUNTIMER_MASK;
+	sdhci_writel(host, val, MLB_CR_SET);
+
+	sdhci_writel(host, MLB_CDR_SET_CLK2POW16, MLB_CDR_SET);
+
+	sdhci_writel(host, MLB_WP_CD_LED_SET_LED_INV, MLB_WP_CD_LED_SET);
+}
+
+static void sdhci_milbeaut_vendor_init(struct sdhci_host *host)
+{
+	struct f_sdhost_priv *priv = sdhci_priv(host);
+	u32 ctl;
+
+	ctl = sdhci_readl(host, F_SDH30_IO_CONTROL2);
+	ctl |= F_SDH30_CRES_O_DN;
+	sdhci_writel(host, ctl, F_SDH30_IO_CONTROL2);
+	ctl &= ~F_SDH30_MSEL_O_1_8;
+	sdhci_writel(host, ctl, F_SDH30_IO_CONTROL2);
+	ctl &= ~F_SDH30_CRES_O_DN;
+	sdhci_writel(host, ctl, F_SDH30_IO_CONTROL2);
+
+	ctl = sdhci_readw(host, F_SDH30_AHB_CONFIG);
+	ctl |= F_SDH30_SIN | F_SDH30_AHB_INCR_16 | F_SDH30_AHB_INCR_8 |
+	       F_SDH30_AHB_INCR_4;
+	ctl &= ~(F_SDH30_AHB_BIGED | F_SDH30_BUSLOCK_EN);
+	sdhci_writew(host, ctl, F_SDH30_AHB_CONFIG);
+
+	if (priv->enable_cmd_dat_delay) {
+		ctl = sdhci_readl(host, F_SDH30_ESD_CONTROL);
+		ctl |= F_SDH30_CMD_DAT_DELAY;
+		sdhci_writel(host, ctl, F_SDH30_ESD_CONTROL);
+	}
+}
+
+static const struct of_device_id mlb_dt_ids[] = {
+	{
+		.compatible = "socionext,milbeaut-m10v-sdhci-3.0",
+	},
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, mlb_dt_ids);
+
+static void sdhci_milbeaut_init(struct sdhci_host *host)
+{
+	struct f_sdhost_priv *priv = sdhci_priv(host);
+	int rate = clk_get_rate(priv->clk);
+	u16 ctl;
+
+	sdhci_milbeaut_bridge_reset(host, 0);
+
+	ctl = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+	ctl &= ~(SDHCI_CLOCK_CARD_EN | SDHCI_CLOCK_INT_EN);
+	sdhci_writew(host, ctl, SDHCI_CLOCK_CONTROL);
+
+	sdhci_milbeaut_bridge_reset(host, 1);
+
+	sdhci_milbeaut_bridge_init(host, rate);
+	sdhci_milbeaut_bridge_reset(host, 0);
+
+	sdhci_milbeaut_vendor_init(host);
+}
+
+static int sdhci_milbeaut_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct device *dev = &pdev->dev;
+	int irq, ret = 0;
+	struct f_sdhost_priv *priv;
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return irq;
+
+	host = sdhci_alloc_host(dev, sizeof(struct f_sdhost_priv));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	priv = sdhci_priv(host);
+	priv->dev = dev;
+
+	host->quirks = SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+			   SDHCI_QUIRK_INVERTED_WRITE_PROTECT |
+			   SDHCI_QUIRK_CLOCK_BEFORE_RESET |
+			   SDHCI_QUIRK_DELAY_AFTER_POWER;
+	host->quirks2 = SDHCI_QUIRK2_SUPPORT_SINGLE |
+			SDHCI_QUIRK2_TUNING_WORK_AROUND |
+			SDHCI_QUIRK2_PRESET_VALUE_BROKEN;
+
+	priv->enable_cmd_dat_delay = device_property_read_bool(dev,
+						"fujitsu,cmd-dat-delay-select");
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto err;
+
+	platform_set_drvdata(pdev, host);
+
+	host->hw_name = "f_sdh30";
+	host->ops = &sdhci_milbeaut_ops;
+	host->irq = irq;
+
+	host->ioaddr = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(host->ioaddr)) {
+		ret = PTR_ERR(host->ioaddr);
+		goto err;
+	}
+
+	if (dev_of_node(dev)) {
+		sdhci_get_of_property(pdev);
+
+		priv->clk_iface = devm_clk_get(&pdev->dev, "iface");
+		if (IS_ERR(priv->clk_iface)) {
+			ret = PTR_ERR(priv->clk_iface);
+			goto err;
+		}
+
+		ret = clk_prepare_enable(priv->clk_iface);
+		if (ret)
+			goto err;
+
+		priv->clk = devm_clk_get(&pdev->dev, "core");
+		if (IS_ERR(priv->clk)) {
+			ret = PTR_ERR(priv->clk);
+			goto err_clk;
+		}
+
+		ret = clk_prepare_enable(priv->clk);
+		if (ret)
+			goto err_clk;
+	}
+
+	sdhci_milbeaut_init(host);
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err_add_host;
+
+	return 0;
+
+err_add_host:
+	clk_disable_unprepare(priv->clk);
+err_clk:
+	clk_disable_unprepare(priv->clk_iface);
+err:
+	sdhci_free_host(host);
+	return ret;
+}
+
+static int sdhci_milbeaut_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct f_sdhost_priv *priv = sdhci_priv(host);
+
+	sdhci_remove_host(host, readl(host->ioaddr + SDHCI_INT_STATUS) ==
+			  0xffffffff);
+
+	clk_disable_unprepare(priv->clk_iface);
+	clk_disable_unprepare(priv->clk);
+
+	sdhci_free_host(host);
+	platform_set_drvdata(pdev, NULL);
+
+	return 0;
+}
+
+static struct platform_driver sdhci_milbeaut_driver = {
+	.driver = {
+		.name = "sdhci-milbeaut",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(mlb_dt_ids),
+	},
+	.probe	= sdhci_milbeaut_probe,
+	.remove	= sdhci_milbeaut_remove,
+};
+
+module_platform_driver(sdhci_milbeaut_driver);
+
+MODULE_DESCRIPTION("MILBEAUT SD Card Controller driver");
+MODULE_AUTHOR("Takao Orito <orito.takao@socionext.com>");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:sdhci-milbeaut");
diff --git a/drivers/mmc/host/sdhci-msm.c b/drivers/mmc/host/sdhci-msm.c
new file mode 100644
index 000000000..3366956a4
--- /dev/null
+++ b/drivers/mmc/host/sdhci-msm.c
@@ -0,0 +1,2628 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * drivers/mmc/host/sdhci-msm.c - Qualcomm SDHCI Platform driver
+ *
+ * Copyright (c) 2013-2014, The Linux Foundation. All rights reserved.
+ */
+
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/delay.h>
+#include <linux/mmc/mmc.h>
+#include <linux/pm_runtime.h>
+#include <linux/pm_opp.h>
+#include <linux/slab.h>
+#include <linux/iopoll.h>
+#include <linux/regulator/consumer.h>
+#include <linux/interconnect.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/reset.h>
+
+#include "sdhci-pltfm.h"
+#include "cqhci.h"
+
+#define CORE_MCI_VERSION		0x50
+#define CORE_VERSION_MAJOR_SHIFT	28
+#define CORE_VERSION_MAJOR_MASK		(0xf << CORE_VERSION_MAJOR_SHIFT)
+#define CORE_VERSION_MINOR_MASK		0xff
+
+#define CORE_MCI_GENERICS		0x70
+#define SWITCHABLE_SIGNALING_VOLTAGE	BIT(29)
+
+#define HC_MODE_EN		0x1
+#define CORE_POWER		0x0
+#define CORE_SW_RST		BIT(7)
+#define FF_CLK_SW_RST_DIS	BIT(13)
+
+#define CORE_PWRCTL_BUS_OFF	BIT(0)
+#define CORE_PWRCTL_BUS_ON	BIT(1)
+#define CORE_PWRCTL_IO_LOW	BIT(2)
+#define CORE_PWRCTL_IO_HIGH	BIT(3)
+#define CORE_PWRCTL_BUS_SUCCESS BIT(0)
+#define CORE_PWRCTL_BUS_FAIL    BIT(1)
+#define CORE_PWRCTL_IO_SUCCESS	BIT(2)
+#define CORE_PWRCTL_IO_FAIL     BIT(3)
+#define REQ_BUS_OFF		BIT(0)
+#define REQ_BUS_ON		BIT(1)
+#define REQ_IO_LOW		BIT(2)
+#define REQ_IO_HIGH		BIT(3)
+#define INT_MASK		0xf
+#define MAX_PHASES		16
+#define CORE_DLL_LOCK		BIT(7)
+#define CORE_DDR_DLL_LOCK	BIT(11)
+#define CORE_DLL_EN		BIT(16)
+#define CORE_CDR_EN		BIT(17)
+#define CORE_CK_OUT_EN		BIT(18)
+#define CORE_CDR_EXT_EN		BIT(19)
+#define CORE_DLL_PDN		BIT(29)
+#define CORE_DLL_RST		BIT(30)
+#define CORE_CMD_DAT_TRACK_SEL	BIT(0)
+
+#define CORE_DDR_CAL_EN		BIT(0)
+#define CORE_FLL_CYCLE_CNT	BIT(18)
+#define CORE_DLL_CLOCK_DISABLE	BIT(21)
+
+#define DLL_USR_CTL_POR_VAL	0x10800
+#define ENABLE_DLL_LOCK_STATUS	BIT(26)
+#define FINE_TUNE_MODE_EN	BIT(27)
+#define BIAS_OK_SIGNAL		BIT(29)
+
+#define DLL_CONFIG_3_LOW_FREQ_VAL	0x08
+#define DLL_CONFIG_3_HIGH_FREQ_VAL	0x10
+
+#define CORE_VENDOR_SPEC_POR_VAL 0xa9c
+#define CORE_CLK_PWRSAVE	BIT(1)
+#define CORE_HC_MCLK_SEL_DFLT	(2 << 8)
+#define CORE_HC_MCLK_SEL_HS400	(3 << 8)
+#define CORE_HC_MCLK_SEL_MASK	(3 << 8)
+#define CORE_IO_PAD_PWR_SWITCH_EN	BIT(15)
+#define CORE_IO_PAD_PWR_SWITCH	BIT(16)
+#define CORE_HC_SELECT_IN_EN	BIT(18)
+#define CORE_HC_SELECT_IN_HS400	(6 << 19)
+#define CORE_HC_SELECT_IN_MASK	(7 << 19)
+
+#define CORE_3_0V_SUPPORT	BIT(25)
+#define CORE_1_8V_SUPPORT	BIT(26)
+#define CORE_VOLT_SUPPORT	(CORE_3_0V_SUPPORT | CORE_1_8V_SUPPORT)
+
+#define CORE_CSR_CDC_CTLR_CFG0		0x130
+#define CORE_SW_TRIG_FULL_CALIB		BIT(16)
+#define CORE_HW_AUTOCAL_ENA		BIT(17)
+
+#define CORE_CSR_CDC_CTLR_CFG1		0x134
+#define CORE_CSR_CDC_CAL_TIMER_CFG0	0x138
+#define CORE_TIMER_ENA			BIT(16)
+
+#define CORE_CSR_CDC_CAL_TIMER_CFG1	0x13C
+#define CORE_CSR_CDC_REFCOUNT_CFG	0x140
+#define CORE_CSR_CDC_COARSE_CAL_CFG	0x144
+#define CORE_CDC_OFFSET_CFG		0x14C
+#define CORE_CSR_CDC_DELAY_CFG		0x150
+#define CORE_CDC_SLAVE_DDA_CFG		0x160
+#define CORE_CSR_CDC_STATUS0		0x164
+#define CORE_CALIBRATION_DONE		BIT(0)
+
+#define CORE_CDC_ERROR_CODE_MASK	0x7000000
+
+#define CORE_CSR_CDC_GEN_CFG		0x178
+#define CORE_CDC_SWITCH_BYPASS_OFF	BIT(0)
+#define CORE_CDC_SWITCH_RC_EN		BIT(1)
+
+#define CORE_CDC_T4_DLY_SEL		BIT(0)
+#define CORE_CMDIN_RCLK_EN		BIT(1)
+#define CORE_START_CDC_TRAFFIC		BIT(6)
+
+#define CORE_PWRSAVE_DLL	BIT(3)
+
+#define DDR_CONFIG_POR_VAL	0x80040873
+
+
+#define INVALID_TUNING_PHASE	-1
+#define SDHCI_MSM_MIN_CLOCK	400000
+#define CORE_FREQ_100MHZ	(100 * 1000 * 1000)
+
+#define CDR_SELEXT_SHIFT	20
+#define CDR_SELEXT_MASK		(0xf << CDR_SELEXT_SHIFT)
+#define CMUX_SHIFT_PHASE_SHIFT	24
+#define CMUX_SHIFT_PHASE_MASK	(7 << CMUX_SHIFT_PHASE_SHIFT)
+
+#define MSM_MMC_AUTOSUSPEND_DELAY_MS	50
+
+/* Timeout value to avoid infinite waiting for pwr_irq */
+#define MSM_PWR_IRQ_TIMEOUT_MS 5000
+
+/* Max load for eMMC Vdd-io supply */
+#define MMC_VQMMC_MAX_LOAD_UA	325000
+
+#define msm_host_readl(msm_host, host, offset) \
+	msm_host->var_ops->msm_readl_relaxed(host, offset)
+
+#define msm_host_writel(msm_host, val, host, offset) \
+	msm_host->var_ops->msm_writel_relaxed(val, host, offset)
+
+/* CQHCI vendor specific registers */
+#define CQHCI_VENDOR_CFG1	0xA00
+#define CQHCI_VENDOR_DIS_RST_ON_CQ_EN	(0x3 << 13)
+
+struct sdhci_msm_offset {
+	u32 core_hc_mode;
+	u32 core_mci_data_cnt;
+	u32 core_mci_status;
+	u32 core_mci_fifo_cnt;
+	u32 core_mci_version;
+	u32 core_generics;
+	u32 core_testbus_config;
+	u32 core_testbus_sel2_bit;
+	u32 core_testbus_ena;
+	u32 core_testbus_sel2;
+	u32 core_pwrctl_status;
+	u32 core_pwrctl_mask;
+	u32 core_pwrctl_clear;
+	u32 core_pwrctl_ctl;
+	u32 core_sdcc_debug_reg;
+	u32 core_dll_config;
+	u32 core_dll_status;
+	u32 core_vendor_spec;
+	u32 core_vendor_spec_adma_err_addr0;
+	u32 core_vendor_spec_adma_err_addr1;
+	u32 core_vendor_spec_func2;
+	u32 core_vendor_spec_capabilities0;
+	u32 core_ddr_200_cfg;
+	u32 core_vendor_spec3;
+	u32 core_dll_config_2;
+	u32 core_dll_config_3;
+	u32 core_ddr_config_old; /* Applicable to sdcc minor ver < 0x49 */
+	u32 core_ddr_config;
+	u32 core_dll_usr_ctl; /* Present on SDCC5.1 onwards */
+};
+
+static const struct sdhci_msm_offset sdhci_msm_v5_offset = {
+	.core_mci_data_cnt = 0x35c,
+	.core_mci_status = 0x324,
+	.core_mci_fifo_cnt = 0x308,
+	.core_mci_version = 0x318,
+	.core_generics = 0x320,
+	.core_testbus_config = 0x32c,
+	.core_testbus_sel2_bit = 3,
+	.core_testbus_ena = (1 << 31),
+	.core_testbus_sel2 = (1 << 3),
+	.core_pwrctl_status = 0x240,
+	.core_pwrctl_mask = 0x244,
+	.core_pwrctl_clear = 0x248,
+	.core_pwrctl_ctl = 0x24c,
+	.core_sdcc_debug_reg = 0x358,
+	.core_dll_config = 0x200,
+	.core_dll_status = 0x208,
+	.core_vendor_spec = 0x20c,
+	.core_vendor_spec_adma_err_addr0 = 0x214,
+	.core_vendor_spec_adma_err_addr1 = 0x218,
+	.core_vendor_spec_func2 = 0x210,
+	.core_vendor_spec_capabilities0 = 0x21c,
+	.core_ddr_200_cfg = 0x224,
+	.core_vendor_spec3 = 0x250,
+	.core_dll_config_2 = 0x254,
+	.core_dll_config_3 = 0x258,
+	.core_ddr_config = 0x25c,
+	.core_dll_usr_ctl = 0x388,
+};
+
+static const struct sdhci_msm_offset sdhci_msm_mci_offset = {
+	.core_hc_mode = 0x78,
+	.core_mci_data_cnt = 0x30,
+	.core_mci_status = 0x34,
+	.core_mci_fifo_cnt = 0x44,
+	.core_mci_version = 0x050,
+	.core_generics = 0x70,
+	.core_testbus_config = 0x0cc,
+	.core_testbus_sel2_bit = 4,
+	.core_testbus_ena = (1 << 3),
+	.core_testbus_sel2 = (1 << 4),
+	.core_pwrctl_status = 0xdc,
+	.core_pwrctl_mask = 0xe0,
+	.core_pwrctl_clear = 0xe4,
+	.core_pwrctl_ctl = 0xe8,
+	.core_sdcc_debug_reg = 0x124,
+	.core_dll_config = 0x100,
+	.core_dll_status = 0x108,
+	.core_vendor_spec = 0x10c,
+	.core_vendor_spec_adma_err_addr0 = 0x114,
+	.core_vendor_spec_adma_err_addr1 = 0x118,
+	.core_vendor_spec_func2 = 0x110,
+	.core_vendor_spec_capabilities0 = 0x11c,
+	.core_ddr_200_cfg = 0x184,
+	.core_vendor_spec3 = 0x1b0,
+	.core_dll_config_2 = 0x1b4,
+	.core_ddr_config_old = 0x1b8,
+	.core_ddr_config = 0x1bc,
+};
+
+struct sdhci_msm_variant_ops {
+	u32 (*msm_readl_relaxed)(struct sdhci_host *host, u32 offset);
+	void (*msm_writel_relaxed)(u32 val, struct sdhci_host *host,
+			u32 offset);
+};
+
+/*
+ * From V5, register spaces have changed. Wrap this info in a structure
+ * and choose the data_structure based on version info mentioned in DT.
+ */
+struct sdhci_msm_variant_info {
+	bool mci_removed;
+	bool restore_dll_config;
+	bool uses_tassadar_dll;
+	const struct sdhci_msm_variant_ops *var_ops;
+	const struct sdhci_msm_offset *offset;
+};
+
+struct sdhci_msm_host {
+	struct platform_device *pdev;
+	void __iomem *core_mem;	/* MSM SDCC mapped address */
+	int pwr_irq;		/* power irq */
+	struct clk *bus_clk;	/* SDHC bus voter clock */
+	struct clk *xo_clk;	/* TCXO clk needed for FLL feature of cm_dll*/
+	struct clk_bulk_data bulk_clks[4]; /* core, iface, cal, sleep clocks */
+	unsigned long clk_rate;
+	struct mmc_host *mmc;
+	struct opp_table *opp_table;
+	bool use_14lpp_dll_reset;
+	bool tuning_done;
+	bool calibration_done;
+	u8 saved_tuning_phase;
+	bool use_cdclp533;
+	u32 curr_pwr_state;
+	u32 curr_io_level;
+	wait_queue_head_t pwr_irq_wait;
+	bool pwr_irq_flag;
+	u32 caps_0;
+	bool mci_removed;
+	bool restore_dll_config;
+	const struct sdhci_msm_variant_ops *var_ops;
+	const struct sdhci_msm_offset *offset;
+	bool use_cdr;
+	u32 transfer_mode;
+	bool updated_ddr_cfg;
+	bool uses_tassadar_dll;
+	u32 dll_config;
+	u32 ddr_config;
+	bool vqmmc_enabled;
+};
+
+static const struct sdhci_msm_offset *sdhci_priv_msm_offset(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+
+	return msm_host->offset;
+}
+
+/*
+ * APIs to read/write to vendor specific registers which were there in the
+ * core_mem region before MCI was removed.
+ */
+static u32 sdhci_msm_mci_variant_readl_relaxed(struct sdhci_host *host,
+		u32 offset)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+
+	return readl_relaxed(msm_host->core_mem + offset);
+}
+
+static u32 sdhci_msm_v5_variant_readl_relaxed(struct sdhci_host *host,
+		u32 offset)
+{
+	return readl_relaxed(host->ioaddr + offset);
+}
+
+static void sdhci_msm_mci_variant_writel_relaxed(u32 val,
+		struct sdhci_host *host, u32 offset)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+
+	writel_relaxed(val, msm_host->core_mem + offset);
+}
+
+static void sdhci_msm_v5_variant_writel_relaxed(u32 val,
+		struct sdhci_host *host, u32 offset)
+{
+	writel_relaxed(val, host->ioaddr + offset);
+}
+
+static unsigned int msm_get_clock_rate_for_bus_mode(struct sdhci_host *host,
+						    unsigned int clock)
+{
+	struct mmc_ios ios = host->mmc->ios;
+	/*
+	 * The SDHC requires internal clock frequency to be double the
+	 * actual clock that will be set for DDR mode. The controller
+	 * uses the faster clock(100/400MHz) for some of its parts and
+	 * send the actual required clock (50/200MHz) to the card.
+	 */
+	if (ios.timing == MMC_TIMING_UHS_DDR50 ||
+	    ios.timing == MMC_TIMING_MMC_DDR52 ||
+	    ios.timing == MMC_TIMING_MMC_HS400 ||
+	    host->flags & SDHCI_HS400_TUNING)
+		clock *= 2;
+	return clock;
+}
+
+static void msm_set_clock_rate_for_bus_mode(struct sdhci_host *host,
+					    unsigned int clock)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	struct mmc_ios curr_ios = host->mmc->ios;
+	struct clk *core_clk = msm_host->bulk_clks[0].clk;
+	int rc;
+
+	clock = msm_get_clock_rate_for_bus_mode(host, clock);
+	rc = dev_pm_opp_set_rate(mmc_dev(host->mmc), clock);
+	if (rc) {
+		pr_err("%s: Failed to set clock at rate %u at timing %d\n",
+		       mmc_hostname(host->mmc), clock,
+		       curr_ios.timing);
+		return;
+	}
+	msm_host->clk_rate = clock;
+	pr_debug("%s: Setting clock at rate %lu at timing %d\n",
+		 mmc_hostname(host->mmc), clk_get_rate(core_clk),
+		 curr_ios.timing);
+}
+
+/* Platform specific tuning */
+static inline int msm_dll_poll_ck_out_en(struct sdhci_host *host, u8 poll)
+{
+	u32 wait_cnt = 50;
+	u8 ck_out_en;
+	struct mmc_host *mmc = host->mmc;
+	const struct sdhci_msm_offset *msm_offset =
+					sdhci_priv_msm_offset(host);
+
+	/* Poll for CK_OUT_EN bit.  max. poll time = 50us */
+	ck_out_en = !!(readl_relaxed(host->ioaddr +
+			msm_offset->core_dll_config) & CORE_CK_OUT_EN);
+
+	while (ck_out_en != poll) {
+		if (--wait_cnt == 0) {
+			dev_err(mmc_dev(mmc), "%s: CK_OUT_EN bit is not %d\n",
+			       mmc_hostname(mmc), poll);
+			return -ETIMEDOUT;
+		}
+		udelay(1);
+
+		ck_out_en = !!(readl_relaxed(host->ioaddr +
+			msm_offset->core_dll_config) & CORE_CK_OUT_EN);
+	}
+
+	return 0;
+}
+
+static int msm_config_cm_dll_phase(struct sdhci_host *host, u8 phase)
+{
+	int rc;
+	static const u8 grey_coded_phase_table[] = {
+		0x0, 0x1, 0x3, 0x2, 0x6, 0x7, 0x5, 0x4,
+		0xc, 0xd, 0xf, 0xe, 0xa, 0xb, 0x9, 0x8
+	};
+	unsigned long flags;
+	u32 config;
+	struct mmc_host *mmc = host->mmc;
+	const struct sdhci_msm_offset *msm_offset =
+					sdhci_priv_msm_offset(host);
+
+	if (phase > 0xf)
+		return -EINVAL;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	config = readl_relaxed(host->ioaddr + msm_offset->core_dll_config);
+	config &= ~(CORE_CDR_EN | CORE_CK_OUT_EN);
+	config |= (CORE_CDR_EXT_EN | CORE_DLL_EN);
+	writel_relaxed(config, host->ioaddr + msm_offset->core_dll_config);
+
+	/* Wait until CK_OUT_EN bit of DLL_CONFIG register becomes '0' */
+	rc = msm_dll_poll_ck_out_en(host, 0);
+	if (rc)
+		goto err_out;
+
+	/*
+	 * Write the selected DLL clock output phase (0 ... 15)
+	 * to CDR_SELEXT bit field of DLL_CONFIG register.
+	 */
+	config = readl_relaxed(host->ioaddr + msm_offset->core_dll_config);
+	config &= ~CDR_SELEXT_MASK;
+	config |= grey_coded_phase_table[phase] << CDR_SELEXT_SHIFT;
+	writel_relaxed(config, host->ioaddr + msm_offset->core_dll_config);
+
+	config = readl_relaxed(host->ioaddr + msm_offset->core_dll_config);
+	config |= CORE_CK_OUT_EN;
+	writel_relaxed(config, host->ioaddr + msm_offset->core_dll_config);
+
+	/* Wait until CK_OUT_EN bit of DLL_CONFIG register becomes '1' */
+	rc = msm_dll_poll_ck_out_en(host, 1);
+	if (rc)
+		goto err_out;
+
+	config = readl_relaxed(host->ioaddr + msm_offset->core_dll_config);
+	config |= CORE_CDR_EN;
+	config &= ~CORE_CDR_EXT_EN;
+	writel_relaxed(config, host->ioaddr + msm_offset->core_dll_config);
+	goto out;
+
+err_out:
+	dev_err(mmc_dev(mmc), "%s: Failed to set DLL phase: %d\n",
+	       mmc_hostname(mmc), phase);
+out:
+	spin_unlock_irqrestore(&host->lock, flags);
+	return rc;
+}
+
+/*
+ * Find out the greatest range of consecuitive selected
+ * DLL clock output phases that can be used as sampling
+ * setting for SD3.0 UHS-I card read operation (in SDR104
+ * timing mode) or for eMMC4.5 card read operation (in
+ * HS400/HS200 timing mode).
+ * Select the 3/4 of the range and configure the DLL with the
+ * selected DLL clock output phase.
+ */
+
+static int msm_find_most_appropriate_phase(struct sdhci_host *host,
+					   u8 *phase_table, u8 total_phases)
+{
+	int ret;
+	u8 ranges[MAX_PHASES][MAX_PHASES] = { {0}, {0} };
+	u8 phases_per_row[MAX_PHASES] = { 0 };
+	int row_index = 0, col_index = 0, selected_row_index = 0, curr_max = 0;
+	int i, cnt, phase_0_raw_index = 0, phase_15_raw_index = 0;
+	bool phase_0_found = false, phase_15_found = false;
+	struct mmc_host *mmc = host->mmc;
+
+	if (!total_phases || (total_phases > MAX_PHASES)) {
+		dev_err(mmc_dev(mmc), "%s: Invalid argument: total_phases=%d\n",
+		       mmc_hostname(mmc), total_phases);
+		return -EINVAL;
+	}
+
+	for (cnt = 0; cnt < total_phases; cnt++) {
+		ranges[row_index][col_index] = phase_table[cnt];
+		phases_per_row[row_index] += 1;
+		col_index++;
+
+		if ((cnt + 1) == total_phases) {
+			continue;
+		/* check if next phase in phase_table is consecutive or not */
+		} else if ((phase_table[cnt] + 1) != phase_table[cnt + 1]) {
+			row_index++;
+			col_index = 0;
+		}
+	}
+
+	if (row_index >= MAX_PHASES)
+		return -EINVAL;
+
+	/* Check if phase-0 is present in first valid window? */
+	if (!ranges[0][0]) {
+		phase_0_found = true;
+		phase_0_raw_index = 0;
+		/* Check if cycle exist between 2 valid windows */
+		for (cnt = 1; cnt <= row_index; cnt++) {
+			if (phases_per_row[cnt]) {
+				for (i = 0; i < phases_per_row[cnt]; i++) {
+					if (ranges[cnt][i] == 15) {
+						phase_15_found = true;
+						phase_15_raw_index = cnt;
+						break;
+					}
+				}
+			}
+		}
+	}
+
+	/* If 2 valid windows form cycle then merge them as single window */
+	if (phase_0_found && phase_15_found) {
+		/* number of phases in raw where phase 0 is present */
+		u8 phases_0 = phases_per_row[phase_0_raw_index];
+		/* number of phases in raw where phase 15 is present */
+		u8 phases_15 = phases_per_row[phase_15_raw_index];
+
+		if (phases_0 + phases_15 >= MAX_PHASES)
+			/*
+			 * If there are more than 1 phase windows then total
+			 * number of phases in both the windows should not be
+			 * more than or equal to MAX_PHASES.
+			 */
+			return -EINVAL;
+
+		/* Merge 2 cyclic windows */
+		i = phases_15;
+		for (cnt = 0; cnt < phases_0; cnt++) {
+			ranges[phase_15_raw_index][i] =
+			    ranges[phase_0_raw_index][cnt];
+			if (++i >= MAX_PHASES)
+				break;
+		}
+
+		phases_per_row[phase_0_raw_index] = 0;
+		phases_per_row[phase_15_raw_index] = phases_15 + phases_0;
+	}
+
+	for (cnt = 0; cnt <= row_index; cnt++) {
+		if (phases_per_row[cnt] > curr_max) {
+			curr_max = phases_per_row[cnt];
+			selected_row_index = cnt;
+		}
+	}
+
+	i = (curr_max * 3) / 4;
+	if (i)
+		i--;
+
+	ret = ranges[selected_row_index][i];
+
+	if (ret >= MAX_PHASES) {
+		ret = -EINVAL;
+		dev_err(mmc_dev(mmc), "%s: Invalid phase selected=%d\n",
+		       mmc_hostname(mmc), ret);
+	}
+
+	return ret;
+}
+
+static inline void msm_cm_dll_set_freq(struct sdhci_host *host)
+{
+	u32 mclk_freq = 0, config;
+	const struct sdhci_msm_offset *msm_offset =
+					sdhci_priv_msm_offset(host);
+
+	/* Program the MCLK value to MCLK_FREQ bit field */
+	if (host->clock <= 112000000)
+		mclk_freq = 0;
+	else if (host->clock <= 125000000)
+		mclk_freq = 1;
+	else if (host->clock <= 137000000)
+		mclk_freq = 2;
+	else if (host->clock <= 150000000)
+		mclk_freq = 3;
+	else if (host->clock <= 162000000)
+		mclk_freq = 4;
+	else if (host->clock <= 175000000)
+		mclk_freq = 5;
+	else if (host->clock <= 187000000)
+		mclk_freq = 6;
+	else if (host->clock <= 200000000)
+		mclk_freq = 7;
+
+	config = readl_relaxed(host->ioaddr + msm_offset->core_dll_config);
+	config &= ~CMUX_SHIFT_PHASE_MASK;
+	config |= mclk_freq << CMUX_SHIFT_PHASE_SHIFT;
+	writel_relaxed(config, host->ioaddr + msm_offset->core_dll_config);
+}
+
+/* Initialize the DLL (Programmable Delay Line) */
+static int msm_init_cm_dll(struct sdhci_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	int wait_cnt = 50;
+	unsigned long flags, xo_clk = 0;
+	u32 config;
+	const struct sdhci_msm_offset *msm_offset =
+					msm_host->offset;
+
+	if (msm_host->use_14lpp_dll_reset && !IS_ERR_OR_NULL(msm_host->xo_clk))
+		xo_clk = clk_get_rate(msm_host->xo_clk);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	/*
+	 * Make sure that clock is always enabled when DLL
+	 * tuning is in progress. Keeping PWRSAVE ON may
+	 * turn off the clock.
+	 */
+	config = readl_relaxed(host->ioaddr + msm_offset->core_vendor_spec);
+	config &= ~CORE_CLK_PWRSAVE;
+	writel_relaxed(config, host->ioaddr + msm_offset->core_vendor_spec);
+
+	if (msm_host->dll_config)
+		writel_relaxed(msm_host->dll_config,
+				host->ioaddr + msm_offset->core_dll_config);
+
+	if (msm_host->use_14lpp_dll_reset) {
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_dll_config);
+		config &= ~CORE_CK_OUT_EN;
+		writel_relaxed(config, host->ioaddr +
+				msm_offset->core_dll_config);
+
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_dll_config_2);
+		config |= CORE_DLL_CLOCK_DISABLE;
+		writel_relaxed(config, host->ioaddr +
+				msm_offset->core_dll_config_2);
+	}
+
+	config = readl_relaxed(host->ioaddr +
+			msm_offset->core_dll_config);
+	config |= CORE_DLL_RST;
+	writel_relaxed(config, host->ioaddr +
+			msm_offset->core_dll_config);
+
+	config = readl_relaxed(host->ioaddr +
+			msm_offset->core_dll_config);
+	config |= CORE_DLL_PDN;
+	writel_relaxed(config, host->ioaddr +
+			msm_offset->core_dll_config);
+
+	if (!msm_host->dll_config)
+		msm_cm_dll_set_freq(host);
+
+	if (msm_host->use_14lpp_dll_reset &&
+	    !IS_ERR_OR_NULL(msm_host->xo_clk)) {
+		u32 mclk_freq = 0;
+
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_dll_config_2);
+		config &= CORE_FLL_CYCLE_CNT;
+		if (config)
+			mclk_freq = DIV_ROUND_CLOSEST_ULL((host->clock * 8),
+					xo_clk);
+		else
+			mclk_freq = DIV_ROUND_CLOSEST_ULL((host->clock * 4),
+					xo_clk);
+
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_dll_config_2);
+		config &= ~(0xFF << 10);
+		config |= mclk_freq << 10;
+
+		writel_relaxed(config, host->ioaddr +
+				msm_offset->core_dll_config_2);
+		/* wait for 5us before enabling DLL clock */
+		udelay(5);
+	}
+
+	config = readl_relaxed(host->ioaddr +
+			msm_offset->core_dll_config);
+	config &= ~CORE_DLL_RST;
+	writel_relaxed(config, host->ioaddr +
+			msm_offset->core_dll_config);
+
+	config = readl_relaxed(host->ioaddr +
+			msm_offset->core_dll_config);
+	config &= ~CORE_DLL_PDN;
+	writel_relaxed(config, host->ioaddr +
+			msm_offset->core_dll_config);
+
+	if (msm_host->use_14lpp_dll_reset) {
+		if (!msm_host->dll_config)
+			msm_cm_dll_set_freq(host);
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_dll_config_2);
+		config &= ~CORE_DLL_CLOCK_DISABLE;
+		writel_relaxed(config, host->ioaddr +
+				msm_offset->core_dll_config_2);
+	}
+
+	/*
+	 * Configure DLL user control register to enable DLL status.
+	 * This setting is applicable to SDCC v5.1 onwards only.
+	 */
+	if (msm_host->uses_tassadar_dll) {
+		config = DLL_USR_CTL_POR_VAL | FINE_TUNE_MODE_EN |
+			ENABLE_DLL_LOCK_STATUS | BIAS_OK_SIGNAL;
+		writel_relaxed(config, host->ioaddr +
+				msm_offset->core_dll_usr_ctl);
+
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_dll_config_3);
+		config &= ~0xFF;
+		if (msm_host->clk_rate < 150000000)
+			config |= DLL_CONFIG_3_LOW_FREQ_VAL;
+		else
+			config |= DLL_CONFIG_3_HIGH_FREQ_VAL;
+		writel_relaxed(config, host->ioaddr +
+			msm_offset->core_dll_config_3);
+	}
+
+	config = readl_relaxed(host->ioaddr +
+			msm_offset->core_dll_config);
+	config |= CORE_DLL_EN;
+	writel_relaxed(config, host->ioaddr +
+			msm_offset->core_dll_config);
+
+	config = readl_relaxed(host->ioaddr +
+			msm_offset->core_dll_config);
+	config |= CORE_CK_OUT_EN;
+	writel_relaxed(config, host->ioaddr +
+			msm_offset->core_dll_config);
+
+	/* Wait until DLL_LOCK bit of DLL_STATUS register becomes '1' */
+	while (!(readl_relaxed(host->ioaddr + msm_offset->core_dll_status) &
+		 CORE_DLL_LOCK)) {
+		/* max. wait for 50us sec for LOCK bit to be set */
+		if (--wait_cnt == 0) {
+			dev_err(mmc_dev(mmc), "%s: DLL failed to LOCK\n",
+			       mmc_hostname(mmc));
+			spin_unlock_irqrestore(&host->lock, flags);
+			return -ETIMEDOUT;
+		}
+		udelay(1);
+	}
+
+	spin_unlock_irqrestore(&host->lock, flags);
+	return 0;
+}
+
+static void msm_hc_select_default(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	u32 config;
+	const struct sdhci_msm_offset *msm_offset =
+					msm_host->offset;
+
+	if (!msm_host->use_cdclp533) {
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_vendor_spec3);
+		config &= ~CORE_PWRSAVE_DLL;
+		writel_relaxed(config, host->ioaddr +
+				msm_offset->core_vendor_spec3);
+	}
+
+	config = readl_relaxed(host->ioaddr + msm_offset->core_vendor_spec);
+	config &= ~CORE_HC_MCLK_SEL_MASK;
+	config |= CORE_HC_MCLK_SEL_DFLT;
+	writel_relaxed(config, host->ioaddr + msm_offset->core_vendor_spec);
+
+	/*
+	 * Disable HC_SELECT_IN to be able to use the UHS mode select
+	 * configuration from Host Control2 register for all other
+	 * modes.
+	 * Write 0 to HC_SELECT_IN and HC_SELECT_IN_EN field
+	 * in VENDOR_SPEC_FUNC
+	 */
+	config = readl_relaxed(host->ioaddr + msm_offset->core_vendor_spec);
+	config &= ~CORE_HC_SELECT_IN_EN;
+	config &= ~CORE_HC_SELECT_IN_MASK;
+	writel_relaxed(config, host->ioaddr + msm_offset->core_vendor_spec);
+
+	/*
+	 * Make sure above writes impacting free running MCLK are completed
+	 * before changing the clk_rate at GCC.
+	 */
+	wmb();
+}
+
+static void msm_hc_select_hs400(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	struct mmc_ios ios = host->mmc->ios;
+	u32 config, dll_lock;
+	int rc;
+	const struct sdhci_msm_offset *msm_offset =
+					msm_host->offset;
+
+	/* Select the divided clock (free running MCLK/2) */
+	config = readl_relaxed(host->ioaddr + msm_offset->core_vendor_spec);
+	config &= ~CORE_HC_MCLK_SEL_MASK;
+	config |= CORE_HC_MCLK_SEL_HS400;
+
+	writel_relaxed(config, host->ioaddr + msm_offset->core_vendor_spec);
+	/*
+	 * Select HS400 mode using the HC_SELECT_IN from VENDOR SPEC
+	 * register
+	 */
+	if ((msm_host->tuning_done || ios.enhanced_strobe) &&
+	    !msm_host->calibration_done) {
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_vendor_spec);
+		config |= CORE_HC_SELECT_IN_HS400;
+		config |= CORE_HC_SELECT_IN_EN;
+		writel_relaxed(config, host->ioaddr +
+				msm_offset->core_vendor_spec);
+	}
+	if (!msm_host->clk_rate && !msm_host->use_cdclp533) {
+		/*
+		 * Poll on DLL_LOCK or DDR_DLL_LOCK bits in
+		 * core_dll_status to be set. This should get set
+		 * within 15 us at 200 MHz.
+		 */
+		rc = readl_relaxed_poll_timeout(host->ioaddr +
+						msm_offset->core_dll_status,
+						dll_lock,
+						(dll_lock &
+						(CORE_DLL_LOCK |
+						CORE_DDR_DLL_LOCK)), 10,
+						1000);
+		if (rc == -ETIMEDOUT)
+			pr_err("%s: Unable to get DLL_LOCK/DDR_DLL_LOCK, dll_status: 0x%08x\n",
+			       mmc_hostname(host->mmc), dll_lock);
+	}
+	/*
+	 * Make sure above writes impacting free running MCLK are completed
+	 * before changing the clk_rate at GCC.
+	 */
+	wmb();
+}
+
+/*
+ * sdhci_msm_hc_select_mode :- In general all timing modes are
+ * controlled via UHS mode select in Host Control2 register.
+ * eMMC specific HS200/HS400 doesn't have their respective modes
+ * defined here, hence we use these values.
+ *
+ * HS200 - SDR104 (Since they both are equivalent in functionality)
+ * HS400 - This involves multiple configurations
+ *		Initially SDR104 - when tuning is required as HS200
+ *		Then when switching to DDR @ 400MHz (HS400) we use
+ *		the vendor specific HC_SELECT_IN to control the mode.
+ *
+ * In addition to controlling the modes we also need to select the
+ * correct input clock for DLL depending on the mode.
+ *
+ * HS400 - divided clock (free running MCLK/2)
+ * All other modes - default (free running MCLK)
+ */
+static void sdhci_msm_hc_select_mode(struct sdhci_host *host)
+{
+	struct mmc_ios ios = host->mmc->ios;
+
+	if (ios.timing == MMC_TIMING_MMC_HS400 ||
+	    host->flags & SDHCI_HS400_TUNING)
+		msm_hc_select_hs400(host);
+	else
+		msm_hc_select_default(host);
+}
+
+static int sdhci_msm_cdclp533_calibration(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	u32 config, calib_done;
+	int ret;
+	const struct sdhci_msm_offset *msm_offset =
+					msm_host->offset;
+
+	pr_debug("%s: %s: Enter\n", mmc_hostname(host->mmc), __func__);
+
+	/*
+	 * Retuning in HS400 (DDR mode) will fail, just reset the
+	 * tuning block and restore the saved tuning phase.
+	 */
+	ret = msm_init_cm_dll(host);
+	if (ret)
+		goto out;
+
+	/* Set the selected phase in delay line hw block */
+	ret = msm_config_cm_dll_phase(host, msm_host->saved_tuning_phase);
+	if (ret)
+		goto out;
+
+	config = readl_relaxed(host->ioaddr + msm_offset->core_dll_config);
+	config |= CORE_CMD_DAT_TRACK_SEL;
+	writel_relaxed(config, host->ioaddr + msm_offset->core_dll_config);
+
+	config = readl_relaxed(host->ioaddr + msm_offset->core_ddr_200_cfg);
+	config &= ~CORE_CDC_T4_DLY_SEL;
+	writel_relaxed(config, host->ioaddr + msm_offset->core_ddr_200_cfg);
+
+	config = readl_relaxed(host->ioaddr + CORE_CSR_CDC_GEN_CFG);
+	config &= ~CORE_CDC_SWITCH_BYPASS_OFF;
+	writel_relaxed(config, host->ioaddr + CORE_CSR_CDC_GEN_CFG);
+
+	config = readl_relaxed(host->ioaddr + CORE_CSR_CDC_GEN_CFG);
+	config |= CORE_CDC_SWITCH_RC_EN;
+	writel_relaxed(config, host->ioaddr + CORE_CSR_CDC_GEN_CFG);
+
+	config = readl_relaxed(host->ioaddr + msm_offset->core_ddr_200_cfg);
+	config &= ~CORE_START_CDC_TRAFFIC;
+	writel_relaxed(config, host->ioaddr + msm_offset->core_ddr_200_cfg);
+
+	/* Perform CDC Register Initialization Sequence */
+
+	writel_relaxed(0x11800EC, host->ioaddr + CORE_CSR_CDC_CTLR_CFG0);
+	writel_relaxed(0x3011111, host->ioaddr + CORE_CSR_CDC_CTLR_CFG1);
+	writel_relaxed(0x1201000, host->ioaddr + CORE_CSR_CDC_CAL_TIMER_CFG0);
+	writel_relaxed(0x4, host->ioaddr + CORE_CSR_CDC_CAL_TIMER_CFG1);
+	writel_relaxed(0xCB732020, host->ioaddr + CORE_CSR_CDC_REFCOUNT_CFG);
+	writel_relaxed(0xB19, host->ioaddr + CORE_CSR_CDC_COARSE_CAL_CFG);
+	writel_relaxed(0x4E2, host->ioaddr + CORE_CSR_CDC_DELAY_CFG);
+	writel_relaxed(0x0, host->ioaddr + CORE_CDC_OFFSET_CFG);
+	writel_relaxed(0x16334, host->ioaddr + CORE_CDC_SLAVE_DDA_CFG);
+
+	/* CDC HW Calibration */
+
+	config = readl_relaxed(host->ioaddr + CORE_CSR_CDC_CTLR_CFG0);
+	config |= CORE_SW_TRIG_FULL_CALIB;
+	writel_relaxed(config, host->ioaddr + CORE_CSR_CDC_CTLR_CFG0);
+
+	config = readl_relaxed(host->ioaddr + CORE_CSR_CDC_CTLR_CFG0);
+	config &= ~CORE_SW_TRIG_FULL_CALIB;
+	writel_relaxed(config, host->ioaddr + CORE_CSR_CDC_CTLR_CFG0);
+
+	config = readl_relaxed(host->ioaddr + CORE_CSR_CDC_CTLR_CFG0);
+	config |= CORE_HW_AUTOCAL_ENA;
+	writel_relaxed(config, host->ioaddr + CORE_CSR_CDC_CTLR_CFG0);
+
+	config = readl_relaxed(host->ioaddr + CORE_CSR_CDC_CAL_TIMER_CFG0);
+	config |= CORE_TIMER_ENA;
+	writel_relaxed(config, host->ioaddr + CORE_CSR_CDC_CAL_TIMER_CFG0);
+
+	ret = readl_relaxed_poll_timeout(host->ioaddr + CORE_CSR_CDC_STATUS0,
+					 calib_done,
+					 (calib_done & CORE_CALIBRATION_DONE),
+					 1, 50);
+
+	if (ret == -ETIMEDOUT) {
+		pr_err("%s: %s: CDC calibration was not completed\n",
+		       mmc_hostname(host->mmc), __func__);
+		goto out;
+	}
+
+	ret = readl_relaxed(host->ioaddr + CORE_CSR_CDC_STATUS0)
+			& CORE_CDC_ERROR_CODE_MASK;
+	if (ret) {
+		pr_err("%s: %s: CDC error code %d\n",
+		       mmc_hostname(host->mmc), __func__, ret);
+		ret = -EINVAL;
+		goto out;
+	}
+
+	config = readl_relaxed(host->ioaddr + msm_offset->core_ddr_200_cfg);
+	config |= CORE_START_CDC_TRAFFIC;
+	writel_relaxed(config, host->ioaddr + msm_offset->core_ddr_200_cfg);
+out:
+	pr_debug("%s: %s: Exit, ret %d\n", mmc_hostname(host->mmc),
+		 __func__, ret);
+	return ret;
+}
+
+static int sdhci_msm_cm_dll_sdc4_calibration(struct sdhci_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	u32 dll_status, config, ddr_cfg_offset;
+	int ret;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	const struct sdhci_msm_offset *msm_offset =
+					sdhci_priv_msm_offset(host);
+
+	pr_debug("%s: %s: Enter\n", mmc_hostname(host->mmc), __func__);
+
+	/*
+	 * Currently the core_ddr_config register defaults to desired
+	 * configuration on reset. Currently reprogramming the power on
+	 * reset (POR) value in case it might have been modified by
+	 * bootloaders. In the future, if this changes, then the desired
+	 * values will need to be programmed appropriately.
+	 */
+	if (msm_host->updated_ddr_cfg)
+		ddr_cfg_offset = msm_offset->core_ddr_config;
+	else
+		ddr_cfg_offset = msm_offset->core_ddr_config_old;
+	writel_relaxed(msm_host->ddr_config, host->ioaddr + ddr_cfg_offset);
+
+	if (mmc->ios.enhanced_strobe) {
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_ddr_200_cfg);
+		config |= CORE_CMDIN_RCLK_EN;
+		writel_relaxed(config, host->ioaddr +
+				msm_offset->core_ddr_200_cfg);
+	}
+
+	config = readl_relaxed(host->ioaddr + msm_offset->core_dll_config_2);
+	config |= CORE_DDR_CAL_EN;
+	writel_relaxed(config, host->ioaddr + msm_offset->core_dll_config_2);
+
+	ret = readl_relaxed_poll_timeout(host->ioaddr +
+					msm_offset->core_dll_status,
+					dll_status,
+					(dll_status & CORE_DDR_DLL_LOCK),
+					10, 1000);
+
+	if (ret == -ETIMEDOUT) {
+		pr_err("%s: %s: CM_DLL_SDC4 calibration was not completed\n",
+		       mmc_hostname(host->mmc), __func__);
+		goto out;
+	}
+
+	/*
+	 * Set CORE_PWRSAVE_DLL bit in CORE_VENDOR_SPEC3.
+	 * When MCLK is gated OFF, it is not gated for less than 0.5us
+	 * and MCLK must be switched on for at-least 1us before DATA
+	 * starts coming. Controllers with 14lpp and later tech DLL cannot
+	 * guarantee above requirement. So PWRSAVE_DLL should not be
+	 * turned on for host controllers using this DLL.
+	 */
+	if (!msm_host->use_14lpp_dll_reset) {
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_vendor_spec3);
+		config |= CORE_PWRSAVE_DLL;
+		writel_relaxed(config, host->ioaddr +
+				msm_offset->core_vendor_spec3);
+	}
+
+	/*
+	 * Drain writebuffer to ensure above DLL calibration
+	 * and PWRSAVE DLL is enabled.
+	 */
+	wmb();
+out:
+	pr_debug("%s: %s: Exit, ret %d\n", mmc_hostname(host->mmc),
+		 __func__, ret);
+	return ret;
+}
+
+static int sdhci_msm_hs400_dll_calibration(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	struct mmc_host *mmc = host->mmc;
+	int ret;
+	u32 config;
+	const struct sdhci_msm_offset *msm_offset =
+					msm_host->offset;
+
+	pr_debug("%s: %s: Enter\n", mmc_hostname(host->mmc), __func__);
+
+	/*
+	 * Retuning in HS400 (DDR mode) will fail, just reset the
+	 * tuning block and restore the saved tuning phase.
+	 */
+	ret = msm_init_cm_dll(host);
+	if (ret)
+		goto out;
+
+	if (!mmc->ios.enhanced_strobe) {
+		/* Set the selected phase in delay line hw block */
+		ret = msm_config_cm_dll_phase(host,
+					      msm_host->saved_tuning_phase);
+		if (ret)
+			goto out;
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_dll_config);
+		config |= CORE_CMD_DAT_TRACK_SEL;
+		writel_relaxed(config, host->ioaddr +
+				msm_offset->core_dll_config);
+	}
+
+	if (msm_host->use_cdclp533)
+		ret = sdhci_msm_cdclp533_calibration(host);
+	else
+		ret = sdhci_msm_cm_dll_sdc4_calibration(host);
+out:
+	pr_debug("%s: %s: Exit, ret %d\n", mmc_hostname(host->mmc),
+		 __func__, ret);
+	return ret;
+}
+
+static bool sdhci_msm_is_tuning_needed(struct sdhci_host *host)
+{
+	struct mmc_ios *ios = &host->mmc->ios;
+
+	/*
+	 * Tuning is required for SDR104, HS200 and HS400 cards and
+	 * if clock frequency is greater than 100MHz in these modes.
+	 */
+	if (host->clock <= CORE_FREQ_100MHZ ||
+	    !(ios->timing == MMC_TIMING_MMC_HS400 ||
+	    ios->timing == MMC_TIMING_MMC_HS200 ||
+	    ios->timing == MMC_TIMING_UHS_SDR104) ||
+	    ios->enhanced_strobe)
+		return false;
+
+	return true;
+}
+
+static int sdhci_msm_restore_sdr_dll_config(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	/*
+	 * SDR DLL comes into picture only for timing modes which needs
+	 * tuning.
+	 */
+	if (!sdhci_msm_is_tuning_needed(host))
+		return 0;
+
+	/* Reset the tuning block */
+	ret = msm_init_cm_dll(host);
+	if (ret)
+		return ret;
+
+	/* Restore the tuning block */
+	ret = msm_config_cm_dll_phase(host, msm_host->saved_tuning_phase);
+
+	return ret;
+}
+
+static void sdhci_msm_set_cdr(struct sdhci_host *host, bool enable)
+{
+	const struct sdhci_msm_offset *msm_offset = sdhci_priv_msm_offset(host);
+	u32 config, oldconfig = readl_relaxed(host->ioaddr +
+					      msm_offset->core_dll_config);
+
+	config = oldconfig;
+	if (enable) {
+		config |= CORE_CDR_EN;
+		config &= ~CORE_CDR_EXT_EN;
+	} else {
+		config &= ~CORE_CDR_EN;
+		config |= CORE_CDR_EXT_EN;
+	}
+
+	if (config != oldconfig) {
+		writel_relaxed(config, host->ioaddr +
+			       msm_offset->core_dll_config);
+	}
+}
+
+static int sdhci_msm_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	int tuning_seq_cnt = 10;
+	u8 phase, tuned_phases[16], tuned_phase_cnt = 0;
+	int rc;
+	struct mmc_ios ios = host->mmc->ios;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+
+	if (!sdhci_msm_is_tuning_needed(host)) {
+		msm_host->use_cdr = false;
+		sdhci_msm_set_cdr(host, false);
+		return 0;
+	}
+
+	/* Clock-Data-Recovery used to dynamically adjust RX sampling point */
+	msm_host->use_cdr = true;
+
+	/*
+	 * Clear tuning_done flag before tuning to ensure proper
+	 * HS400 settings.
+	 */
+	msm_host->tuning_done = 0;
+
+	/*
+	 * For HS400 tuning in HS200 timing requires:
+	 * - select MCLK/2 in VENDOR_SPEC
+	 * - program MCLK to 400MHz (or nearest supported) in GCC
+	 */
+	if (host->flags & SDHCI_HS400_TUNING) {
+		sdhci_msm_hc_select_mode(host);
+		msm_set_clock_rate_for_bus_mode(host, ios.clock);
+		host->flags &= ~SDHCI_HS400_TUNING;
+	}
+
+retry:
+	/* First of all reset the tuning block */
+	rc = msm_init_cm_dll(host);
+	if (rc)
+		return rc;
+
+	phase = 0;
+	do {
+		/* Set the phase in delay line hw block */
+		rc = msm_config_cm_dll_phase(host, phase);
+		if (rc)
+			return rc;
+
+		rc = mmc_send_tuning(mmc, opcode, NULL);
+		if (!rc) {
+			/* Tuning is successful at this tuning point */
+			tuned_phases[tuned_phase_cnt++] = phase;
+			dev_dbg(mmc_dev(mmc), "%s: Found good phase = %d\n",
+				 mmc_hostname(mmc), phase);
+		}
+	} while (++phase < ARRAY_SIZE(tuned_phases));
+
+	if (tuned_phase_cnt) {
+		if (tuned_phase_cnt == ARRAY_SIZE(tuned_phases)) {
+			/*
+			 * All phases valid is _almost_ as bad as no phases
+			 * valid.  Probably all phases are not really reliable
+			 * but we didn't detect where the unreliable place is.
+			 * That means we'll essentially be guessing and hoping
+			 * we get a good phase.  Better to try a few times.
+			 */
+			dev_dbg(mmc_dev(mmc), "%s: All phases valid; try again\n",
+				mmc_hostname(mmc));
+			if (--tuning_seq_cnt) {
+				tuned_phase_cnt = 0;
+				goto retry;
+			}
+		}
+
+		rc = msm_find_most_appropriate_phase(host, tuned_phases,
+						     tuned_phase_cnt);
+		if (rc < 0)
+			return rc;
+		else
+			phase = rc;
+
+		/*
+		 * Finally set the selected phase in delay
+		 * line hw block.
+		 */
+		rc = msm_config_cm_dll_phase(host, phase);
+		if (rc)
+			return rc;
+		msm_host->saved_tuning_phase = phase;
+		dev_dbg(mmc_dev(mmc), "%s: Setting the tuning phase to %d\n",
+			 mmc_hostname(mmc), phase);
+	} else {
+		if (--tuning_seq_cnt)
+			goto retry;
+		/* Tuning failed */
+		dev_dbg(mmc_dev(mmc), "%s: No tuning point found\n",
+		       mmc_hostname(mmc));
+		rc = -EIO;
+	}
+
+	if (!rc)
+		msm_host->tuning_done = true;
+	return rc;
+}
+
+/*
+ * sdhci_msm_hs400 - Calibrate the DLL for HS400 bus speed mode operation.
+ * This needs to be done for both tuning and enhanced_strobe mode.
+ * DLL operation is only needed for clock > 100MHz. For clock <= 100MHz
+ * fixed feedback clock is used.
+ */
+static void sdhci_msm_hs400(struct sdhci_host *host, struct mmc_ios *ios)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	if (host->clock > CORE_FREQ_100MHZ &&
+	    (msm_host->tuning_done || ios->enhanced_strobe) &&
+	    !msm_host->calibration_done) {
+		ret = sdhci_msm_hs400_dll_calibration(host);
+		if (!ret)
+			msm_host->calibration_done = true;
+		else
+			pr_err("%s: Failed to calibrate DLL for hs400 mode (%d)\n",
+			       mmc_hostname(host->mmc), ret);
+	}
+}
+
+static void sdhci_msm_set_uhs_signaling(struct sdhci_host *host,
+					unsigned int uhs)
+{
+	struct mmc_host *mmc = host->mmc;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	u16 ctrl_2;
+	u32 config;
+	const struct sdhci_msm_offset *msm_offset =
+					msm_host->offset;
+
+	ctrl_2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	/* Select Bus Speed Mode for host */
+	ctrl_2 &= ~SDHCI_CTRL_UHS_MASK;
+	switch (uhs) {
+	case MMC_TIMING_UHS_SDR12:
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR12;
+		break;
+	case MMC_TIMING_UHS_SDR25:
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR25;
+		break;
+	case MMC_TIMING_UHS_SDR50:
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR50;
+		break;
+	case MMC_TIMING_MMC_HS400:
+	case MMC_TIMING_MMC_HS200:
+	case MMC_TIMING_UHS_SDR104:
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR104;
+		break;
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_MMC_DDR52:
+		ctrl_2 |= SDHCI_CTRL_UHS_DDR50;
+		break;
+	}
+
+	/*
+	 * When clock frequency is less than 100MHz, the feedback clock must be
+	 * provided and DLL must not be used so that tuning can be skipped. To
+	 * provide feedback clock, the mode selection can be any value less
+	 * than 3'b011 in bits [2:0] of HOST CONTROL2 register.
+	 */
+	if (host->clock <= CORE_FREQ_100MHZ) {
+		if (uhs == MMC_TIMING_MMC_HS400 ||
+		    uhs == MMC_TIMING_MMC_HS200 ||
+		    uhs == MMC_TIMING_UHS_SDR104)
+			ctrl_2 &= ~SDHCI_CTRL_UHS_MASK;
+		/*
+		 * DLL is not required for clock <= 100MHz
+		 * Thus, make sure DLL it is disabled when not required
+		 */
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_dll_config);
+		config |= CORE_DLL_RST;
+		writel_relaxed(config, host->ioaddr +
+				msm_offset->core_dll_config);
+
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_dll_config);
+		config |= CORE_DLL_PDN;
+		writel_relaxed(config, host->ioaddr +
+				msm_offset->core_dll_config);
+
+		/*
+		 * The DLL needs to be restored and CDCLP533 recalibrated
+		 * when the clock frequency is set back to 400MHz.
+		 */
+		msm_host->calibration_done = false;
+	}
+
+	dev_dbg(mmc_dev(mmc), "%s: clock=%u uhs=%u ctrl_2=0x%x\n",
+		mmc_hostname(host->mmc), host->clock, uhs, ctrl_2);
+	sdhci_writew(host, ctrl_2, SDHCI_HOST_CONTROL2);
+
+	if (mmc->ios.timing == MMC_TIMING_MMC_HS400)
+		sdhci_msm_hs400(host, &mmc->ios);
+}
+
+static int sdhci_msm_set_pincfg(struct sdhci_msm_host *msm_host, bool level)
+{
+	struct platform_device *pdev = msm_host->pdev;
+	int ret;
+
+	if (level)
+		ret = pinctrl_pm_select_default_state(&pdev->dev);
+	else
+		ret = pinctrl_pm_select_sleep_state(&pdev->dev);
+
+	return ret;
+}
+
+static int sdhci_msm_set_vmmc(struct mmc_host *mmc)
+{
+	if (IS_ERR(mmc->supply.vmmc))
+		return 0;
+
+	return mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, mmc->ios.vdd);
+}
+
+static int msm_toggle_vqmmc(struct sdhci_msm_host *msm_host,
+			      struct mmc_host *mmc, bool level)
+{
+	int ret;
+	struct mmc_ios ios;
+
+	if (msm_host->vqmmc_enabled == level)
+		return 0;
+
+	if (level) {
+		/* Set the IO voltage regulator to default voltage level */
+		if (msm_host->caps_0 & CORE_3_0V_SUPPORT)
+			ios.signal_voltage = MMC_SIGNAL_VOLTAGE_330;
+		else if (msm_host->caps_0 & CORE_1_8V_SUPPORT)
+			ios.signal_voltage = MMC_SIGNAL_VOLTAGE_180;
+
+		if (msm_host->caps_0 & CORE_VOLT_SUPPORT) {
+			ret = mmc_regulator_set_vqmmc(mmc, &ios);
+			if (ret < 0) {
+				dev_err(mmc_dev(mmc), "%s: vqmmc set volgate failed: %d\n",
+					mmc_hostname(mmc), ret);
+				goto out;
+			}
+		}
+		ret = regulator_enable(mmc->supply.vqmmc);
+	} else {
+		ret = regulator_disable(mmc->supply.vqmmc);
+	}
+
+	if (ret)
+		dev_err(mmc_dev(mmc), "%s: vqmm %sable failed: %d\n",
+			mmc_hostname(mmc), level ? "en":"dis", ret);
+	else
+		msm_host->vqmmc_enabled = level;
+out:
+	return ret;
+}
+
+static int msm_config_vqmmc_mode(struct sdhci_msm_host *msm_host,
+			      struct mmc_host *mmc, bool hpm)
+{
+	int load, ret;
+
+	load = hpm ? MMC_VQMMC_MAX_LOAD_UA : 0;
+	ret = regulator_set_load(mmc->supply.vqmmc, load);
+	if (ret)
+		dev_err(mmc_dev(mmc), "%s: vqmmc set load failed: %d\n",
+			mmc_hostname(mmc), ret);
+	return ret;
+}
+
+static int sdhci_msm_set_vqmmc(struct sdhci_msm_host *msm_host,
+			      struct mmc_host *mmc, bool level)
+{
+	int ret;
+	bool always_on;
+
+	if (IS_ERR(mmc->supply.vqmmc) ||
+			(mmc->ios.power_mode == MMC_POWER_UNDEFINED))
+		return 0;
+	/*
+	 * For eMMC don't turn off Vqmmc, Instead just configure it in LPM
+	 * and HPM modes by setting the corresponding load.
+	 *
+	 * Till eMMC is initialized (i.e. always_on == 0), just turn on/off
+	 * Vqmmc. Vqmmc gets turned off only if init fails and mmc_power_off
+	 * gets invoked. Once eMMC is initialized (i.e. always_on == 1),
+	 * Vqmmc should remain ON, So just set the load instead of turning it
+	 * off/on.
+	 */
+	always_on = !mmc_card_is_removable(mmc) &&
+			mmc->card && mmc_card_mmc(mmc->card);
+
+	if (always_on)
+		ret = msm_config_vqmmc_mode(msm_host, mmc, level);
+	else
+		ret = msm_toggle_vqmmc(msm_host, mmc, level);
+
+	return ret;
+}
+
+static inline void sdhci_msm_init_pwr_irq_wait(struct sdhci_msm_host *msm_host)
+{
+	init_waitqueue_head(&msm_host->pwr_irq_wait);
+}
+
+static inline void sdhci_msm_complete_pwr_irq_wait(
+		struct sdhci_msm_host *msm_host)
+{
+	wake_up(&msm_host->pwr_irq_wait);
+}
+
+/*
+ * sdhci_msm_check_power_status API should be called when registers writes
+ * which can toggle sdhci IO bus ON/OFF or change IO lines HIGH/LOW happens.
+ * To what state the register writes will change the IO lines should be passed
+ * as the argument req_type. This API will check whether the IO line's state
+ * is already the expected state and will wait for power irq only if
+ * power irq is expected to be triggered based on the current IO line state
+ * and expected IO line state.
+ */
+static void sdhci_msm_check_power_status(struct sdhci_host *host, u32 req_type)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	bool done = false;
+	u32 val = SWITCHABLE_SIGNALING_VOLTAGE;
+	const struct sdhci_msm_offset *msm_offset =
+					msm_host->offset;
+
+	pr_debug("%s: %s: request %d curr_pwr_state %x curr_io_level %x\n",
+			mmc_hostname(host->mmc), __func__, req_type,
+			msm_host->curr_pwr_state, msm_host->curr_io_level);
+
+	/*
+	 * The power interrupt will not be generated for signal voltage
+	 * switches if SWITCHABLE_SIGNALING_VOLTAGE in MCI_GENERICS is not set.
+	 * Since sdhci-msm-v5, this bit has been removed and SW must consider
+	 * it as always set.
+	 */
+	if (!msm_host->mci_removed)
+		val = msm_host_readl(msm_host, host,
+				msm_offset->core_generics);
+	if ((req_type & REQ_IO_HIGH || req_type & REQ_IO_LOW) &&
+	    !(val & SWITCHABLE_SIGNALING_VOLTAGE)) {
+		return;
+	}
+
+	/*
+	 * The IRQ for request type IO High/LOW will be generated when -
+	 * there is a state change in 1.8V enable bit (bit 3) of
+	 * SDHCI_HOST_CONTROL2 register. The reset state of that bit is 0
+	 * which indicates 3.3V IO voltage. So, when MMC core layer tries
+	 * to set it to 3.3V before card detection happens, the
+	 * IRQ doesn't get triggered as there is no state change in this bit.
+	 * The driver already handles this case by changing the IO voltage
+	 * level to high as part of controller power up sequence. Hence, check
+	 * for host->pwr to handle a case where IO voltage high request is
+	 * issued even before controller power up.
+	 */
+	if ((req_type & REQ_IO_HIGH) && !host->pwr) {
+		pr_debug("%s: do not wait for power IRQ that never comes, req_type: %d\n",
+				mmc_hostname(host->mmc), req_type);
+		return;
+	}
+	if ((req_type & msm_host->curr_pwr_state) ||
+			(req_type & msm_host->curr_io_level))
+		done = true;
+	/*
+	 * This is needed here to handle cases where register writes will
+	 * not change the current bus state or io level of the controller.
+	 * In this case, no power irq will be triggerred and we should
+	 * not wait.
+	 */
+	if (!done) {
+		if (!wait_event_timeout(msm_host->pwr_irq_wait,
+				msm_host->pwr_irq_flag,
+				msecs_to_jiffies(MSM_PWR_IRQ_TIMEOUT_MS)))
+			dev_warn(&msm_host->pdev->dev,
+				 "%s: pwr_irq for req: (%d) timed out\n",
+				 mmc_hostname(host->mmc), req_type);
+	}
+	pr_debug("%s: %s: request %d done\n", mmc_hostname(host->mmc),
+			__func__, req_type);
+}
+
+static void sdhci_msm_dump_pwr_ctrl_regs(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	const struct sdhci_msm_offset *msm_offset =
+					msm_host->offset;
+
+	pr_err("%s: PWRCTL_STATUS: 0x%08x | PWRCTL_MASK: 0x%08x | PWRCTL_CTL: 0x%08x\n",
+		mmc_hostname(host->mmc),
+		msm_host_readl(msm_host, host, msm_offset->core_pwrctl_status),
+		msm_host_readl(msm_host, host, msm_offset->core_pwrctl_mask),
+		msm_host_readl(msm_host, host, msm_offset->core_pwrctl_ctl));
+}
+
+static void sdhci_msm_handle_pwr_irq(struct sdhci_host *host, int irq)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	struct mmc_host *mmc = host->mmc;
+	u32 irq_status, irq_ack = 0;
+	int retry = 10, ret;
+	u32 pwr_state = 0, io_level = 0;
+	u32 config;
+	const struct sdhci_msm_offset *msm_offset = msm_host->offset;
+
+	irq_status = msm_host_readl(msm_host, host,
+			msm_offset->core_pwrctl_status);
+	irq_status &= INT_MASK;
+
+	msm_host_writel(msm_host, irq_status, host,
+			msm_offset->core_pwrctl_clear);
+
+	/*
+	 * There is a rare HW scenario where the first clear pulse could be
+	 * lost when actual reset and clear/read of status register is
+	 * happening at a time. Hence, retry for at least 10 times to make
+	 * sure status register is cleared. Otherwise, this will result in
+	 * a spurious power IRQ resulting in system instability.
+	 */
+	while (irq_status & msm_host_readl(msm_host, host,
+				msm_offset->core_pwrctl_status)) {
+		if (retry == 0) {
+			pr_err("%s: Timedout clearing (0x%x) pwrctl status register\n",
+					mmc_hostname(host->mmc), irq_status);
+			sdhci_msm_dump_pwr_ctrl_regs(host);
+			WARN_ON(1);
+			break;
+		}
+		msm_host_writel(msm_host, irq_status, host,
+			msm_offset->core_pwrctl_clear);
+		retry--;
+		udelay(10);
+	}
+
+	/* Handle BUS ON/OFF*/
+	if (irq_status & CORE_PWRCTL_BUS_ON) {
+		pwr_state = REQ_BUS_ON;
+		io_level = REQ_IO_HIGH;
+	}
+	if (irq_status & CORE_PWRCTL_BUS_OFF) {
+		pwr_state = REQ_BUS_OFF;
+		io_level = REQ_IO_LOW;
+	}
+
+	if (pwr_state) {
+		ret = sdhci_msm_set_vmmc(mmc);
+		if (!ret)
+			ret = sdhci_msm_set_vqmmc(msm_host, mmc,
+					pwr_state & REQ_BUS_ON);
+		if (!ret)
+			ret = sdhci_msm_set_pincfg(msm_host,
+					pwr_state & REQ_BUS_ON);
+		if (!ret)
+			irq_ack |= CORE_PWRCTL_BUS_SUCCESS;
+		else
+			irq_ack |= CORE_PWRCTL_BUS_FAIL;
+	}
+
+	/* Handle IO LOW/HIGH */
+	if (irq_status & CORE_PWRCTL_IO_LOW)
+		io_level = REQ_IO_LOW;
+
+	if (irq_status & CORE_PWRCTL_IO_HIGH)
+		io_level = REQ_IO_HIGH;
+
+	if (io_level)
+		irq_ack |= CORE_PWRCTL_IO_SUCCESS;
+
+	if (io_level && !IS_ERR(mmc->supply.vqmmc) && !pwr_state) {
+		ret = mmc_regulator_set_vqmmc(mmc, &mmc->ios);
+		if (ret < 0) {
+			dev_err(mmc_dev(mmc), "%s: IO_level setting failed(%d). signal_voltage: %d, vdd: %d irq_status: 0x%08x\n",
+					mmc_hostname(mmc), ret,
+					mmc->ios.signal_voltage, mmc->ios.vdd,
+					irq_status);
+			irq_ack |= CORE_PWRCTL_IO_FAIL;
+		}
+	}
+
+	/*
+	 * The driver has to acknowledge the interrupt, switch voltages and
+	 * report back if it succeded or not to this register. The voltage
+	 * switches are handled by the sdhci core, so just report success.
+	 */
+	msm_host_writel(msm_host, irq_ack, host,
+			msm_offset->core_pwrctl_ctl);
+
+	/*
+	 * If we don't have info regarding the voltage levels supported by
+	 * regulators, don't change the IO PAD PWR SWITCH.
+	 */
+	if (msm_host->caps_0 & CORE_VOLT_SUPPORT) {
+		u32 new_config;
+		/*
+		 * We should unset IO PAD PWR switch only if the register write
+		 * can set IO lines high and the regulator also switches to 3 V.
+		 * Else, we should keep the IO PAD PWR switch set.
+		 * This is applicable to certain targets where eMMC vccq supply
+		 * is only 1.8V. In such targets, even during REQ_IO_HIGH, the
+		 * IO PAD PWR switch must be kept set to reflect actual
+		 * regulator voltage. This way, during initialization of
+		 * controllers with only 1.8V, we will set the IO PAD bit
+		 * without waiting for a REQ_IO_LOW.
+		 */
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_vendor_spec);
+		new_config = config;
+
+		if ((io_level & REQ_IO_HIGH) &&
+				(msm_host->caps_0 & CORE_3_0V_SUPPORT))
+			new_config &= ~CORE_IO_PAD_PWR_SWITCH;
+		else if ((io_level & REQ_IO_LOW) ||
+				(msm_host->caps_0 & CORE_1_8V_SUPPORT))
+			new_config |= CORE_IO_PAD_PWR_SWITCH;
+
+		if (config ^ new_config)
+			writel_relaxed(new_config, host->ioaddr +
+					msm_offset->core_vendor_spec);
+	}
+
+	if (pwr_state)
+		msm_host->curr_pwr_state = pwr_state;
+	if (io_level)
+		msm_host->curr_io_level = io_level;
+
+	dev_dbg(mmc_dev(mmc), "%s: %s: Handled IRQ(%d), irq_status=0x%x, ack=0x%x\n",
+		mmc_hostname(msm_host->mmc), __func__, irq, irq_status,
+		irq_ack);
+}
+
+static irqreturn_t sdhci_msm_pwr_irq(int irq, void *data)
+{
+	struct sdhci_host *host = (struct sdhci_host *)data;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+
+	sdhci_msm_handle_pwr_irq(host, irq);
+	msm_host->pwr_irq_flag = 1;
+	sdhci_msm_complete_pwr_irq_wait(msm_host);
+
+
+	return IRQ_HANDLED;
+}
+
+static unsigned int sdhci_msm_get_max_clock(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	struct clk *core_clk = msm_host->bulk_clks[0].clk;
+
+	return clk_round_rate(core_clk, ULONG_MAX);
+}
+
+static unsigned int sdhci_msm_get_min_clock(struct sdhci_host *host)
+{
+	return SDHCI_MSM_MIN_CLOCK;
+}
+
+/*
+ * __sdhci_msm_set_clock - sdhci_msm clock control.
+ *
+ * Description:
+ * MSM controller does not use internal divider and
+ * instead directly control the GCC clock as per
+ * HW recommendation.
+ **/
+static void __sdhci_msm_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	u16 clk;
+	/*
+	 * Keep actual_clock as zero -
+	 * - since there is no divider used so no need of having actual_clock.
+	 * - MSM controller uses SDCLK for data timeout calculation. If
+	 *   actual_clock is zero, host->clock is taken for calculation.
+	 */
+	host->mmc->actual_clock = 0;
+
+	sdhci_writew(host, 0, SDHCI_CLOCK_CONTROL);
+
+	if (clock == 0)
+		return;
+
+	/*
+	 * MSM controller do not use clock divider.
+	 * Thus read SDHCI_CLOCK_CONTROL and only enable
+	 * clock with no divider value programmed.
+	 */
+	clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+	sdhci_enable_clk(host, clk);
+}
+
+/* sdhci_msm_set_clock - Called with (host->lock) spinlock held. */
+static void sdhci_msm_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+
+	if (!clock) {
+		msm_host->clk_rate = clock;
+		goto out;
+	}
+
+	sdhci_msm_hc_select_mode(host);
+
+	msm_set_clock_rate_for_bus_mode(host, clock);
+out:
+	__sdhci_msm_set_clock(host, clock);
+}
+
+/*****************************************************************************\
+ *                                                                           *
+ * MSM Command Queue Engine (CQE)                                            *
+ *                                                                           *
+\*****************************************************************************/
+
+static u32 sdhci_msm_cqe_irq(struct sdhci_host *host, u32 intmask)
+{
+	int cmd_error = 0;
+	int data_error = 0;
+
+	if (!sdhci_cqe_irq(host, intmask, &cmd_error, &data_error))
+		return intmask;
+
+	cqhci_irq(host->mmc, intmask, cmd_error, data_error);
+	return 0;
+}
+
+static void sdhci_msm_cqe_disable(struct mmc_host *mmc, bool recovery)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+	u32 ctrl;
+
+	/*
+	 * When CQE is halted, the legacy SDHCI path operates only
+	 * on 16-byte descriptors in 64bit mode.
+	 */
+	if (host->flags & SDHCI_USE_64_BIT_DMA)
+		host->desc_sz = 16;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	/*
+	 * During CQE command transfers, command complete bit gets latched.
+	 * So s/w should clear command complete interrupt status when CQE is
+	 * either halted or disabled. Otherwise unexpected SDCHI legacy
+	 * interrupt gets triggered when CQE is halted/disabled.
+	 */
+	ctrl = sdhci_readl(host, SDHCI_INT_ENABLE);
+	ctrl |= SDHCI_INT_RESPONSE;
+	sdhci_writel(host,  ctrl, SDHCI_INT_ENABLE);
+	sdhci_writel(host, SDHCI_INT_RESPONSE, SDHCI_INT_STATUS);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	sdhci_cqe_disable(mmc, recovery);
+}
+
+static void sdhci_msm_set_timeout(struct sdhci_host *host, struct mmc_command *cmd)
+{
+	u32 count, start = 15;
+
+	__sdhci_set_timeout(host, cmd);
+	count = sdhci_readb(host, SDHCI_TIMEOUT_CONTROL);
+	/*
+	 * Update software timeout value if its value is less than hardware data
+	 * timeout value. Qcom SoC hardware data timeout value was calculated
+	 * using 4 * MCLK * 2^(count + 13). where MCLK = 1 / host->clock.
+	 */
+	if (cmd && cmd->data && host->clock > 400000 &&
+	    host->clock <= 50000000 &&
+	    ((1 << (count + start)) > (10 * host->clock)))
+		host->data_timeout = 22LL * NSEC_PER_SEC;
+}
+
+static const struct cqhci_host_ops sdhci_msm_cqhci_ops = {
+	.enable		= sdhci_cqe_enable,
+	.disable	= sdhci_msm_cqe_disable,
+};
+
+static int sdhci_msm_cqe_add_host(struct sdhci_host *host,
+				struct platform_device *pdev)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	struct cqhci_host *cq_host;
+	bool dma64;
+	u32 cqcfg;
+	int ret;
+
+	/*
+	 * When CQE is halted, SDHC operates only on 16byte ADMA descriptors.
+	 * So ensure ADMA table is allocated for 16byte descriptors.
+	 */
+	if (host->caps & SDHCI_CAN_64BIT)
+		host->alloc_desc_sz = 16;
+
+	ret = sdhci_setup_host(host);
+	if (ret)
+		return ret;
+
+	cq_host = cqhci_pltfm_init(pdev);
+	if (IS_ERR(cq_host)) {
+		ret = PTR_ERR(cq_host);
+		dev_err(&pdev->dev, "cqhci-pltfm init: failed: %d\n", ret);
+		goto cleanup;
+	}
+
+	msm_host->mmc->caps2 |= MMC_CAP2_CQE | MMC_CAP2_CQE_DCMD;
+	cq_host->ops = &sdhci_msm_cqhci_ops;
+
+	dma64 = host->flags & SDHCI_USE_64_BIT_DMA;
+
+	ret = cqhci_init(cq_host, host->mmc, dma64);
+	if (ret) {
+		dev_err(&pdev->dev, "%s: CQE init: failed (%d)\n",
+				mmc_hostname(host->mmc), ret);
+		goto cleanup;
+	}
+
+	/* Disable cqe reset due to cqe enable signal */
+	cqcfg = cqhci_readl(cq_host, CQHCI_VENDOR_CFG1);
+	cqcfg |= CQHCI_VENDOR_DIS_RST_ON_CQ_EN;
+	cqhci_writel(cq_host, cqcfg, CQHCI_VENDOR_CFG1);
+
+	/*
+	 * SDHC expects 12byte ADMA descriptors till CQE is enabled.
+	 * So limit desc_sz to 12 so that the data commands that are sent
+	 * during card initialization (before CQE gets enabled) would
+	 * get executed without any issues.
+	 */
+	if (host->flags & SDHCI_USE_64_BIT_DMA)
+		host->desc_sz = 12;
+
+	ret = __sdhci_add_host(host);
+	if (ret)
+		goto cleanup;
+
+	dev_info(&pdev->dev, "%s: CQE init: success\n",
+			mmc_hostname(host->mmc));
+	return ret;
+
+cleanup:
+	sdhci_cleanup_host(host);
+	return ret;
+}
+
+/*
+ * Platform specific register write functions. This is so that, if any
+ * register write needs to be followed up by platform specific actions,
+ * they can be added here. These functions can go to sleep when writes
+ * to certain registers are done.
+ * These functions are relying on sdhci_set_ios not using spinlock.
+ */
+static int __sdhci_msm_check_write(struct sdhci_host *host, u16 val, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	u32 req_type = 0;
+
+	switch (reg) {
+	case SDHCI_HOST_CONTROL2:
+		req_type = (val & SDHCI_CTRL_VDD_180) ? REQ_IO_LOW :
+			REQ_IO_HIGH;
+		break;
+	case SDHCI_SOFTWARE_RESET:
+		if (host->pwr && (val & SDHCI_RESET_ALL))
+			req_type = REQ_BUS_OFF;
+		break;
+	case SDHCI_POWER_CONTROL:
+		req_type = !val ? REQ_BUS_OFF : REQ_BUS_ON;
+		break;
+	case SDHCI_TRANSFER_MODE:
+		msm_host->transfer_mode = val;
+		break;
+	case SDHCI_COMMAND:
+		if (!msm_host->use_cdr)
+			break;
+		if ((msm_host->transfer_mode & SDHCI_TRNS_READ) &&
+		    SDHCI_GET_CMD(val) != MMC_SEND_TUNING_BLOCK_HS200 &&
+		    SDHCI_GET_CMD(val) != MMC_SEND_TUNING_BLOCK)
+			sdhci_msm_set_cdr(host, true);
+		else
+			sdhci_msm_set_cdr(host, false);
+		break;
+	}
+
+	if (req_type) {
+		msm_host->pwr_irq_flag = 0;
+		/*
+		 * Since this register write may trigger a power irq, ensure
+		 * all previous register writes are complete by this point.
+		 */
+		mb();
+	}
+	return req_type;
+}
+
+/* This function may sleep*/
+static void sdhci_msm_writew(struct sdhci_host *host, u16 val, int reg)
+{
+	u32 req_type = 0;
+
+	req_type = __sdhci_msm_check_write(host, val, reg);
+	writew_relaxed(val, host->ioaddr + reg);
+
+	if (req_type)
+		sdhci_msm_check_power_status(host, req_type);
+}
+
+/* This function may sleep*/
+static void sdhci_msm_writeb(struct sdhci_host *host, u8 val, int reg)
+{
+	u32 req_type = 0;
+
+	req_type = __sdhci_msm_check_write(host, val, reg);
+
+	writeb_relaxed(val, host->ioaddr + reg);
+
+	if (req_type)
+		sdhci_msm_check_power_status(host, req_type);
+}
+
+static void sdhci_msm_set_regulator_caps(struct sdhci_msm_host *msm_host)
+{
+	struct mmc_host *mmc = msm_host->mmc;
+	struct regulator *supply = mmc->supply.vqmmc;
+	u32 caps = 0, config;
+	struct sdhci_host *host = mmc_priv(mmc);
+	const struct sdhci_msm_offset *msm_offset = msm_host->offset;
+
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+		if (regulator_is_supported_voltage(supply, 1700000, 1950000))
+			caps |= CORE_1_8V_SUPPORT;
+		if (regulator_is_supported_voltage(supply, 2700000, 3600000))
+			caps |= CORE_3_0V_SUPPORT;
+
+		if (!caps)
+			pr_warn("%s: 1.8/3V not supported for vqmmc\n",
+					mmc_hostname(mmc));
+	}
+
+	if (caps) {
+		/*
+		 * Set the PAD_PWR_SWITCH_EN bit so that the PAD_PWR_SWITCH
+		 * bit can be used as required later on.
+		 */
+		u32 io_level = msm_host->curr_io_level;
+
+		config = readl_relaxed(host->ioaddr +
+				msm_offset->core_vendor_spec);
+		config |= CORE_IO_PAD_PWR_SWITCH_EN;
+
+		if ((io_level & REQ_IO_HIGH) && (caps &	CORE_3_0V_SUPPORT))
+			config &= ~CORE_IO_PAD_PWR_SWITCH;
+		else if ((io_level & REQ_IO_LOW) || (caps & CORE_1_8V_SUPPORT))
+			config |= CORE_IO_PAD_PWR_SWITCH;
+
+		writel_relaxed(config,
+				host->ioaddr + msm_offset->core_vendor_spec);
+	}
+	msm_host->caps_0 |= caps;
+	pr_debug("%s: supported caps: 0x%08x\n", mmc_hostname(mmc), caps);
+}
+
+static void sdhci_msm_reset(struct sdhci_host *host, u8 mask)
+{
+	if ((host->mmc->caps2 & MMC_CAP2_CQE) && (mask & SDHCI_RESET_ALL))
+		cqhci_deactivate(host->mmc);
+	sdhci_reset(host, mask);
+}
+
+static int sdhci_msm_register_vreg(struct sdhci_msm_host *msm_host)
+{
+	int ret;
+
+	ret = mmc_regulator_get_supply(msm_host->mmc);
+	if (ret)
+		return ret;
+
+	sdhci_msm_set_regulator_caps(msm_host);
+
+	return 0;
+}
+
+static int sdhci_msm_start_signal_voltage_switch(struct mmc_host *mmc,
+				      struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	u16 ctrl, status;
+
+	/*
+	 * Signal Voltage Switching is only applicable for Host Controllers
+	 * v3.00 and above.
+	 */
+	if (host->version < SDHCI_SPEC_300)
+		return 0;
+
+	ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+
+	switch (ios->signal_voltage) {
+	case MMC_SIGNAL_VOLTAGE_330:
+		if (!(host->flags & SDHCI_SIGNALING_330))
+			return -EINVAL;
+
+		/* Set 1.8V Signal Enable in the Host Control2 register to 0 */
+		ctrl &= ~SDHCI_CTRL_VDD_180;
+		break;
+	case MMC_SIGNAL_VOLTAGE_180:
+		if (!(host->flags & SDHCI_SIGNALING_180))
+			return -EINVAL;
+
+		/* Enable 1.8V Signal Enable in the Host Control2 register */
+		ctrl |= SDHCI_CTRL_VDD_180;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	sdhci_writew(host, ctrl, SDHCI_HOST_CONTROL2);
+
+	/* Wait for 5ms */
+	usleep_range(5000, 5500);
+
+	/* regulator output should be stable within 5 ms */
+	status = ctrl & SDHCI_CTRL_VDD_180;
+	ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	if ((ctrl & SDHCI_CTRL_VDD_180) == status)
+		return 0;
+
+	dev_warn(mmc_dev(mmc), "%s: Regulator output did not became stable\n",
+		mmc_hostname(mmc));
+
+	return -EAGAIN;
+}
+
+#define DRIVER_NAME "sdhci_msm"
+#define SDHCI_MSM_DUMP(f, x...) \
+	pr_err("%s: " DRIVER_NAME ": " f, mmc_hostname(host->mmc), ## x)
+
+static void sdhci_msm_dump_vendor_regs(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	const struct sdhci_msm_offset *msm_offset = msm_host->offset;
+
+	SDHCI_MSM_DUMP("----------- VENDOR REGISTER DUMP -----------\n");
+
+	SDHCI_MSM_DUMP(
+			"DLL sts: 0x%08x | DLL cfg:  0x%08x | DLL cfg2: 0x%08x\n",
+		readl_relaxed(host->ioaddr + msm_offset->core_dll_status),
+		readl_relaxed(host->ioaddr + msm_offset->core_dll_config),
+		readl_relaxed(host->ioaddr + msm_offset->core_dll_config_2));
+	SDHCI_MSM_DUMP(
+			"DLL cfg3: 0x%08x | DLL usr ctl:  0x%08x | DDR cfg: 0x%08x\n",
+		readl_relaxed(host->ioaddr + msm_offset->core_dll_config_3),
+		readl_relaxed(host->ioaddr + msm_offset->core_dll_usr_ctl),
+		readl_relaxed(host->ioaddr + msm_offset->core_ddr_config));
+	SDHCI_MSM_DUMP(
+			"Vndr func: 0x%08x | Vndr func2 : 0x%08x Vndr func3: 0x%08x\n",
+		readl_relaxed(host->ioaddr + msm_offset->core_vendor_spec),
+		readl_relaxed(host->ioaddr +
+			msm_offset->core_vendor_spec_func2),
+		readl_relaxed(host->ioaddr + msm_offset->core_vendor_spec3));
+}
+
+static const struct sdhci_msm_variant_ops mci_var_ops = {
+	.msm_readl_relaxed = sdhci_msm_mci_variant_readl_relaxed,
+	.msm_writel_relaxed = sdhci_msm_mci_variant_writel_relaxed,
+};
+
+static const struct sdhci_msm_variant_ops v5_var_ops = {
+	.msm_readl_relaxed = sdhci_msm_v5_variant_readl_relaxed,
+	.msm_writel_relaxed = sdhci_msm_v5_variant_writel_relaxed,
+};
+
+static const struct sdhci_msm_variant_info sdhci_msm_mci_var = {
+	.var_ops = &mci_var_ops,
+	.offset = &sdhci_msm_mci_offset,
+};
+
+static const struct sdhci_msm_variant_info sdhci_msm_v5_var = {
+	.mci_removed = true,
+	.var_ops = &v5_var_ops,
+	.offset = &sdhci_msm_v5_offset,
+};
+
+static const struct sdhci_msm_variant_info sdm845_sdhci_var = {
+	.mci_removed = true,
+	.restore_dll_config = true,
+	.var_ops = &v5_var_ops,
+	.offset = &sdhci_msm_v5_offset,
+};
+
+static const struct sdhci_msm_variant_info sm8250_sdhci_var = {
+	.mci_removed = true,
+	.uses_tassadar_dll = true,
+	.var_ops = &v5_var_ops,
+	.offset = &sdhci_msm_v5_offset,
+};
+
+static const struct of_device_id sdhci_msm_dt_match[] = {
+	{.compatible = "qcom,sdhci-msm-v4", .data = &sdhci_msm_mci_var},
+	{.compatible = "qcom,sdhci-msm-v5", .data = &sdhci_msm_v5_var},
+	{.compatible = "qcom,sdm670-sdhci", .data = &sdm845_sdhci_var},
+	{.compatible = "qcom,sdm845-sdhci", .data = &sdm845_sdhci_var},
+	{.compatible = "qcom,sm8250-sdhci", .data = &sm8250_sdhci_var},
+	{.compatible = "qcom,sc7180-sdhci", .data = &sdm845_sdhci_var},
+	{},
+};
+
+MODULE_DEVICE_TABLE(of, sdhci_msm_dt_match);
+
+static const struct sdhci_ops sdhci_msm_ops = {
+	.reset = sdhci_msm_reset,
+	.set_clock = sdhci_msm_set_clock,
+	.get_min_clock = sdhci_msm_get_min_clock,
+	.get_max_clock = sdhci_msm_get_max_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.set_uhs_signaling = sdhci_msm_set_uhs_signaling,
+	.write_w = sdhci_msm_writew,
+	.write_b = sdhci_msm_writeb,
+	.irq	= sdhci_msm_cqe_irq,
+	.dump_vendor_regs = sdhci_msm_dump_vendor_regs,
+	.set_power = sdhci_set_power_noreg,
+	.set_timeout = sdhci_msm_set_timeout,
+};
+
+static const struct sdhci_pltfm_data sdhci_msm_pdata = {
+	.quirks = SDHCI_QUIRK_BROKEN_CARD_DETECTION |
+		  SDHCI_QUIRK_SINGLE_POWER_WRITE |
+		  SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN |
+		  SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12,
+
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+	.ops = &sdhci_msm_ops,
+};
+
+static inline void sdhci_msm_get_of_property(struct platform_device *pdev,
+		struct sdhci_host *host)
+{
+	struct device_node *node = pdev->dev.of_node;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+
+	if (of_property_read_u32(node, "qcom,ddr-config",
+				&msm_host->ddr_config))
+		msm_host->ddr_config = DDR_CONFIG_POR_VAL;
+
+	of_property_read_u32(node, "qcom,dll-config", &msm_host->dll_config);
+
+	if (of_device_is_compatible(node, "qcom,msm8916-sdhci"))
+		host->quirks2 |= SDHCI_QUIRK2_BROKEN_64_BIT_DMA;
+}
+
+static int sdhci_msm_gcc_reset(struct device *dev, struct sdhci_host *host)
+{
+	struct reset_control *reset;
+	int ret = 0;
+
+	reset = reset_control_get_optional_exclusive(dev, NULL);
+	if (IS_ERR(reset))
+		return dev_err_probe(dev, PTR_ERR(reset),
+				"unable to acquire core_reset\n");
+
+	if (!reset)
+		return ret;
+
+	ret = reset_control_assert(reset);
+	if (ret) {
+		reset_control_put(reset);
+		return dev_err_probe(dev, ret, "core_reset assert failed\n");
+	}
+
+	/*
+	 * The hardware requirement for delay between assert/deassert
+	 * is at least 3-4 sleep clock (32.7KHz) cycles, which comes to
+	 * ~125us (4/32768). To be on the safe side add 200us delay.
+	 */
+	usleep_range(200, 210);
+
+	ret = reset_control_deassert(reset);
+	if (ret) {
+		reset_control_put(reset);
+		return dev_err_probe(dev, ret, "core_reset deassert failed\n");
+	}
+
+	usleep_range(200, 210);
+	reset_control_put(reset);
+
+	return ret;
+}
+
+static int sdhci_msm_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_msm_host *msm_host;
+	struct clk *clk;
+	int ret;
+	u16 host_version, core_minor;
+	u32 core_version, config;
+	u8 core_major;
+	const struct sdhci_msm_offset *msm_offset;
+	const struct sdhci_msm_variant_info *var_info;
+	struct device_node *node = pdev->dev.of_node;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_msm_pdata, sizeof(*msm_host));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	host->sdma_boundary = 0;
+	pltfm_host = sdhci_priv(host);
+	msm_host = sdhci_pltfm_priv(pltfm_host);
+	msm_host->mmc = host->mmc;
+	msm_host->pdev = pdev;
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto pltfm_free;
+
+	/*
+	 * Based on the compatible string, load the required msm host info from
+	 * the data associated with the version info.
+	 */
+	var_info = of_device_get_match_data(&pdev->dev);
+
+	msm_host->mci_removed = var_info->mci_removed;
+	msm_host->restore_dll_config = var_info->restore_dll_config;
+	msm_host->var_ops = var_info->var_ops;
+	msm_host->offset = var_info->offset;
+	msm_host->uses_tassadar_dll = var_info->uses_tassadar_dll;
+
+	msm_offset = msm_host->offset;
+
+	sdhci_get_of_property(pdev);
+	sdhci_msm_get_of_property(pdev, host);
+
+	msm_host->saved_tuning_phase = INVALID_TUNING_PHASE;
+
+	ret = sdhci_msm_gcc_reset(&pdev->dev, host);
+	if (ret)
+		goto pltfm_free;
+
+	/* Setup SDCC bus voter clock. */
+	msm_host->bus_clk = devm_clk_get(&pdev->dev, "bus");
+	if (!IS_ERR(msm_host->bus_clk)) {
+		/* Vote for max. clk rate for max. performance */
+		ret = clk_set_rate(msm_host->bus_clk, INT_MAX);
+		if (ret)
+			goto pltfm_free;
+		ret = clk_prepare_enable(msm_host->bus_clk);
+		if (ret)
+			goto pltfm_free;
+	}
+
+	/* Setup main peripheral bus clock */
+	clk = devm_clk_get(&pdev->dev, "iface");
+	if (IS_ERR(clk)) {
+		ret = PTR_ERR(clk);
+		dev_err(&pdev->dev, "Peripheral clk setup failed (%d)\n", ret);
+		goto bus_clk_disable;
+	}
+	msm_host->bulk_clks[1].clk = clk;
+
+	/* Setup SDC MMC clock */
+	clk = devm_clk_get(&pdev->dev, "core");
+	if (IS_ERR(clk)) {
+		ret = PTR_ERR(clk);
+		dev_err(&pdev->dev, "SDC MMC clk setup failed (%d)\n", ret);
+		goto bus_clk_disable;
+	}
+	msm_host->bulk_clks[0].clk = clk;
+
+	 /* Check for optional interconnect paths */
+	ret = dev_pm_opp_of_find_icc_paths(&pdev->dev, NULL);
+	if (ret)
+		goto bus_clk_disable;
+
+	msm_host->opp_table = dev_pm_opp_set_clkname(&pdev->dev, "core");
+	if (IS_ERR(msm_host->opp_table)) {
+		ret = PTR_ERR(msm_host->opp_table);
+		goto bus_clk_disable;
+	}
+
+	/* OPP table is optional */
+	ret = dev_pm_opp_of_add_table(&pdev->dev);
+	if (ret && ret != -ENODEV) {
+		dev_err(&pdev->dev, "Invalid OPP table in Device tree\n");
+		goto opp_put_clkname;
+	}
+
+	/* Vote for maximum clock rate for maximum performance */
+	ret = dev_pm_opp_set_rate(&pdev->dev, INT_MAX);
+	if (ret)
+		dev_warn(&pdev->dev, "core clock boost failed\n");
+
+	clk = devm_clk_get(&pdev->dev, "cal");
+	if (IS_ERR(clk))
+		clk = NULL;
+	msm_host->bulk_clks[2].clk = clk;
+
+	clk = devm_clk_get(&pdev->dev, "sleep");
+	if (IS_ERR(clk))
+		clk = NULL;
+	msm_host->bulk_clks[3].clk = clk;
+
+	ret = clk_bulk_prepare_enable(ARRAY_SIZE(msm_host->bulk_clks),
+				      msm_host->bulk_clks);
+	if (ret)
+		goto opp_cleanup;
+
+	/*
+	 * xo clock is needed for FLL feature of cm_dll.
+	 * In case if xo clock is not mentioned in DT, warn and proceed.
+	 */
+	msm_host->xo_clk = devm_clk_get(&pdev->dev, "xo");
+	if (IS_ERR(msm_host->xo_clk)) {
+		ret = PTR_ERR(msm_host->xo_clk);
+		dev_warn(&pdev->dev, "TCXO clk not present (%d)\n", ret);
+	}
+
+	if (!msm_host->mci_removed) {
+		msm_host->core_mem = devm_platform_ioremap_resource(pdev, 1);
+		if (IS_ERR(msm_host->core_mem)) {
+			ret = PTR_ERR(msm_host->core_mem);
+			goto clk_disable;
+		}
+	}
+
+	/* Reset the vendor spec register to power on reset state */
+	writel_relaxed(CORE_VENDOR_SPEC_POR_VAL,
+			host->ioaddr + msm_offset->core_vendor_spec);
+
+	if (!msm_host->mci_removed) {
+		/* Set HC_MODE_EN bit in HC_MODE register */
+		msm_host_writel(msm_host, HC_MODE_EN, host,
+				msm_offset->core_hc_mode);
+		config = msm_host_readl(msm_host, host,
+				msm_offset->core_hc_mode);
+		config |= FF_CLK_SW_RST_DIS;
+		msm_host_writel(msm_host, config, host,
+				msm_offset->core_hc_mode);
+	}
+
+	host_version = readw_relaxed((host->ioaddr + SDHCI_HOST_VERSION));
+	dev_dbg(&pdev->dev, "Host Version: 0x%x Vendor Version 0x%x\n",
+		host_version, ((host_version & SDHCI_VENDOR_VER_MASK) >>
+			       SDHCI_VENDOR_VER_SHIFT));
+
+	core_version = msm_host_readl(msm_host, host,
+			msm_offset->core_mci_version);
+	core_major = (core_version & CORE_VERSION_MAJOR_MASK) >>
+		      CORE_VERSION_MAJOR_SHIFT;
+	core_minor = core_version & CORE_VERSION_MINOR_MASK;
+	dev_dbg(&pdev->dev, "MCI Version: 0x%08x, major: 0x%04x, minor: 0x%02x\n",
+		core_version, core_major, core_minor);
+
+	if (core_major == 1 && core_minor >= 0x42)
+		msm_host->use_14lpp_dll_reset = true;
+
+	/*
+	 * SDCC 5 controller with major version 1, minor version 0x34 and later
+	 * with HS 400 mode support will use CM DLL instead of CDC LP 533 DLL.
+	 */
+	if (core_major == 1 && core_minor < 0x34)
+		msm_host->use_cdclp533 = true;
+
+	/*
+	 * Support for some capabilities is not advertised by newer
+	 * controller versions and must be explicitly enabled.
+	 */
+	if (core_major >= 1 && core_minor != 0x11 && core_minor != 0x12) {
+		config = readl_relaxed(host->ioaddr + SDHCI_CAPABILITIES);
+		config |= SDHCI_CAN_VDD_300 | SDHCI_CAN_DO_8BIT;
+		writel_relaxed(config, host->ioaddr +
+				msm_offset->core_vendor_spec_capabilities0);
+	}
+
+	if (core_major == 1 && core_minor >= 0x49)
+		msm_host->updated_ddr_cfg = true;
+
+	ret = sdhci_msm_register_vreg(msm_host);
+	if (ret)
+		goto clk_disable;
+
+	/*
+	 * Power on reset state may trigger power irq if previous status of
+	 * PWRCTL was either BUS_ON or IO_HIGH_V. So before enabling pwr irq
+	 * interrupt in GIC, any pending power irq interrupt should be
+	 * acknowledged. Otherwise power irq interrupt handler would be
+	 * fired prematurely.
+	 */
+	sdhci_msm_handle_pwr_irq(host, 0);
+
+	/*
+	 * Ensure that above writes are propogated before interrupt enablement
+	 * in GIC.
+	 */
+	mb();
+
+	/* Setup IRQ for handling power/voltage tasks with PMIC */
+	msm_host->pwr_irq = platform_get_irq_byname(pdev, "pwr_irq");
+	if (msm_host->pwr_irq < 0) {
+		ret = msm_host->pwr_irq;
+		goto clk_disable;
+	}
+
+	sdhci_msm_init_pwr_irq_wait(msm_host);
+	/* Enable pwr irq interrupts */
+	msm_host_writel(msm_host, INT_MASK, host,
+		msm_offset->core_pwrctl_mask);
+
+	ret = devm_request_threaded_irq(&pdev->dev, msm_host->pwr_irq, NULL,
+					sdhci_msm_pwr_irq, IRQF_ONESHOT,
+					dev_name(&pdev->dev), host);
+	if (ret) {
+		dev_err(&pdev->dev, "Request IRQ failed (%d)\n", ret);
+		goto clk_disable;
+	}
+
+	msm_host->mmc->caps |= MMC_CAP_WAIT_WHILE_BUSY | MMC_CAP_NEED_RSP_BUSY;
+
+	pm_runtime_get_noresume(&pdev->dev);
+	pm_runtime_set_active(&pdev->dev);
+	pm_runtime_enable(&pdev->dev);
+	pm_runtime_set_autosuspend_delay(&pdev->dev,
+					 MSM_MMC_AUTOSUSPEND_DELAY_MS);
+	pm_runtime_use_autosuspend(&pdev->dev);
+
+	host->mmc_host_ops.start_signal_voltage_switch =
+		sdhci_msm_start_signal_voltage_switch;
+	host->mmc_host_ops.execute_tuning = sdhci_msm_execute_tuning;
+	if (of_property_read_bool(node, "supports-cqe"))
+		ret = sdhci_msm_cqe_add_host(host, pdev);
+	else
+		ret = sdhci_add_host(host);
+	if (ret)
+		goto pm_runtime_disable;
+
+	pm_runtime_mark_last_busy(&pdev->dev);
+	pm_runtime_put_autosuspend(&pdev->dev);
+
+	return 0;
+
+pm_runtime_disable:
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_set_suspended(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+clk_disable:
+	clk_bulk_disable_unprepare(ARRAY_SIZE(msm_host->bulk_clks),
+				   msm_host->bulk_clks);
+opp_cleanup:
+	dev_pm_opp_of_remove_table(&pdev->dev);
+opp_put_clkname:
+	dev_pm_opp_put_clkname(msm_host->opp_table);
+bus_clk_disable:
+	if (!IS_ERR(msm_host->bus_clk))
+		clk_disable_unprepare(msm_host->bus_clk);
+pltfm_free:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static int sdhci_msm_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	int dead = (readl_relaxed(host->ioaddr + SDHCI_INT_STATUS) ==
+		    0xffffffff);
+
+	sdhci_remove_host(host, dead);
+
+	dev_pm_opp_of_remove_table(&pdev->dev);
+	dev_pm_opp_put_clkname(msm_host->opp_table);
+	pm_runtime_get_sync(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+
+	clk_bulk_disable_unprepare(ARRAY_SIZE(msm_host->bulk_clks),
+				   msm_host->bulk_clks);
+	if (!IS_ERR(msm_host->bus_clk))
+		clk_disable_unprepare(msm_host->bus_clk);
+	sdhci_pltfm_free(pdev);
+	return 0;
+}
+
+static __maybe_unused int sdhci_msm_runtime_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+
+	/* Drop the performance vote */
+	dev_pm_opp_set_rate(dev, 0);
+	clk_bulk_disable_unprepare(ARRAY_SIZE(msm_host->bulk_clks),
+				   msm_host->bulk_clks);
+
+	return 0;
+}
+
+static __maybe_unused int sdhci_msm_runtime_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	ret = clk_bulk_prepare_enable(ARRAY_SIZE(msm_host->bulk_clks),
+				       msm_host->bulk_clks);
+	if (ret)
+		return ret;
+	/*
+	 * Whenever core-clock is gated dynamically, it's needed to
+	 * restore the SDR DLL settings when the clock is ungated.
+	 */
+	if (msm_host->restore_dll_config && msm_host->clk_rate)
+		ret = sdhci_msm_restore_sdr_dll_config(host);
+
+	dev_pm_opp_set_rate(dev, msm_host->clk_rate);
+
+	return ret;
+}
+
+static const struct dev_pm_ops sdhci_msm_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(sdhci_msm_runtime_suspend,
+			   sdhci_msm_runtime_resume,
+			   NULL)
+};
+
+static struct platform_driver sdhci_msm_driver = {
+	.probe = sdhci_msm_probe,
+	.remove = sdhci_msm_remove,
+	.driver = {
+		   .name = "sdhci_msm",
+		   .of_match_table = sdhci_msm_dt_match,
+		   .pm = &sdhci_msm_pm_ops,
+		   .probe_type = PROBE_PREFER_ASYNCHRONOUS,
+	},
+};
+
+module_platform_driver(sdhci_msm_driver);
+
+MODULE_DESCRIPTION("Qualcomm Secure Digital Host Controller Interface driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-of-arasan.c b/drivers/mmc/host/sdhci-of-arasan.c
new file mode 100644
index 000000000..9da49dc15
--- /dev/null
+++ b/drivers/mmc/host/sdhci-of-arasan.c
@@ -0,0 +1,1729 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Arasan Secure Digital Host Controller Interface.
+ * Copyright (C) 2011 - 2012 Michal Simek <monstr@monstr.eu>
+ * Copyright (c) 2012 Wind River Systems, Inc.
+ * Copyright (C) 2013 Pengutronix e.K.
+ * Copyright (C) 2013 Xilinx Inc.
+ *
+ * Based on sdhci-of-esdhc.c
+ *
+ * Copyright (c) 2007 Freescale Semiconductor, Inc.
+ * Copyright (c) 2009 MontaVista Software, Inc.
+ *
+ * Authors: Xiaobo Xie <X.Xie@freescale.com>
+ *	    Anton Vorontsov <avorontsov@ru.mvista.com>
+ */
+
+#include <linux/clk-provider.h>
+#include <linux/mfd/syscon.h>
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/phy/phy.h>
+#include <linux/regmap.h>
+#include <linux/of.h>
+#include <linux/firmware/xlnx-zynqmp.h>
+
+#include "cqhci.h"
+#include "sdhci-cqhci.h"
+#include "sdhci-pltfm.h"
+
+#define SDHCI_ARASAN_VENDOR_REGISTER	0x78
+
+#define SDHCI_ARASAN_ITAPDLY_REGISTER	0xF0F8
+#define SDHCI_ARASAN_ITAPDLY_SEL_MASK	0xFF
+
+#define SDHCI_ARASAN_OTAPDLY_REGISTER	0xF0FC
+#define SDHCI_ARASAN_OTAPDLY_SEL_MASK	0x3F
+
+#define SDHCI_ARASAN_CQE_BASE_ADDR	0x200
+#define VENDOR_ENHANCED_STROBE		BIT(0)
+
+#define PHY_CLK_TOO_SLOW_HZ		400000
+
+#define SDHCI_ITAPDLY_CHGWIN		0x200
+#define SDHCI_ITAPDLY_ENABLE		0x100
+#define SDHCI_OTAPDLY_ENABLE		0x40
+
+/* Default settings for ZynqMP Clock Phases */
+#define ZYNQMP_ICLK_PHASE {0, 63, 63, 0, 63,  0,   0, 183, 54,  0, 0}
+#define ZYNQMP_OCLK_PHASE {0, 72, 60, 0, 60, 72, 135, 48, 72, 135, 0}
+
+#define VERSAL_ICLK_PHASE {0, 132, 132, 0, 132, 0, 0, 162, 90, 0, 0}
+#define VERSAL_OCLK_PHASE {0,  60, 48, 0, 48, 72, 90, 36, 60, 90, 0}
+
+/*
+ * On some SoCs the syscon area has a feature where the upper 16-bits of
+ * each 32-bit register act as a write mask for the lower 16-bits.  This allows
+ * atomic updates of the register without locking.  This macro is used on SoCs
+ * that have that feature.
+ */
+#define HIWORD_UPDATE(val, mask, shift) \
+		((val) << (shift) | (mask) << ((shift) + 16))
+
+/**
+ * struct sdhci_arasan_soc_ctl_field - Field used in sdhci_arasan_soc_ctl_map
+ *
+ * @reg:	Offset within the syscon of the register containing this field
+ * @width:	Number of bits for this field
+ * @shift:	Bit offset within @reg of this field (or -1 if not avail)
+ */
+struct sdhci_arasan_soc_ctl_field {
+	u32 reg;
+	u16 width;
+	s16 shift;
+};
+
+/**
+ * struct sdhci_arasan_soc_ctl_map - Map in syscon to corecfg registers
+ *
+ * @baseclkfreq:	Where to find corecfg_baseclkfreq
+ * @clockmultiplier:	Where to find corecfg_clockmultiplier
+ * @support64b:		Where to find SUPPORT64B bit
+ * @hiword_update:	If true, use HIWORD_UPDATE to access the syscon
+ *
+ * It's up to the licensee of the Arsan IP block to make these available
+ * somewhere if needed.  Presumably these will be scattered somewhere that's
+ * accessible via the syscon API.
+ */
+struct sdhci_arasan_soc_ctl_map {
+	struct sdhci_arasan_soc_ctl_field	baseclkfreq;
+	struct sdhci_arasan_soc_ctl_field	clockmultiplier;
+	struct sdhci_arasan_soc_ctl_field	support64b;
+	bool					hiword_update;
+};
+
+/**
+ * struct sdhci_arasan_clk_ops - Clock Operations for Arasan SD controller
+ *
+ * @sdcardclk_ops:	The output clock related operations
+ * @sampleclk_ops:	The sample clock related operations
+ */
+struct sdhci_arasan_clk_ops {
+	const struct clk_ops *sdcardclk_ops;
+	const struct clk_ops *sampleclk_ops;
+};
+
+/**
+ * struct sdhci_arasan_clk_data - Arasan Controller Clock Data.
+ *
+ * @sdcardclk_hw:	Struct for the clock we might provide to a PHY.
+ * @sdcardclk:		Pointer to normal 'struct clock' for sdcardclk_hw.
+ * @sampleclk_hw:	Struct for the clock we might provide to a PHY.
+ * @sampleclk:		Pointer to normal 'struct clock' for sampleclk_hw.
+ * @clk_phase_in:	Array of Input Clock Phase Delays for all speed modes
+ * @clk_phase_out:	Array of Output Clock Phase Delays for all speed modes
+ * @set_clk_delays:	Function pointer for setting Clock Delays
+ * @clk_of_data:	Platform specific runtime clock data storage pointer
+ */
+struct sdhci_arasan_clk_data {
+	struct clk_hw	sdcardclk_hw;
+	struct clk      *sdcardclk;
+	struct clk_hw	sampleclk_hw;
+	struct clk      *sampleclk;
+	int		clk_phase_in[MMC_TIMING_MMC_HS400 + 1];
+	int		clk_phase_out[MMC_TIMING_MMC_HS400 + 1];
+	void		(*set_clk_delays)(struct sdhci_host *host);
+	void		*clk_of_data;
+};
+
+/**
+ * struct sdhci_arasan_data - Arasan Controller Data
+ *
+ * @host:		Pointer to the main SDHCI host structure.
+ * @clk_ahb:		Pointer to the AHB clock
+ * @phy:		Pointer to the generic phy
+ * @is_phy_on:		True if the PHY is on; false if not.
+ * @has_cqe:		True if controller has command queuing engine.
+ * @clk_data:		Struct for the Arasan Controller Clock Data.
+ * @clk_ops:		Struct for the Arasan Controller Clock Operations.
+ * @soc_ctl_base:	Pointer to regmap for syscon for soc_ctl registers.
+ * @soc_ctl_map:	Map to get offsets into soc_ctl registers.
+ * @quirks:		Arasan deviations from spec.
+ */
+struct sdhci_arasan_data {
+	struct sdhci_host *host;
+	struct clk	*clk_ahb;
+	struct phy	*phy;
+	bool		is_phy_on;
+
+	bool		has_cqe;
+	struct sdhci_arasan_clk_data clk_data;
+	const struct sdhci_arasan_clk_ops *clk_ops;
+
+	struct regmap	*soc_ctl_base;
+	const struct sdhci_arasan_soc_ctl_map *soc_ctl_map;
+	unsigned int	quirks;
+
+/* Controller does not have CD wired and will not function normally without */
+#define SDHCI_ARASAN_QUIRK_FORCE_CDTEST	BIT(0)
+/* Controller immediately reports SDHCI_CLOCK_INT_STABLE after enabling the
+ * internal clock even when the clock isn't stable */
+#define SDHCI_ARASAN_QUIRK_CLOCK_UNSTABLE BIT(1)
+/*
+ * Some of the Arasan variations might not have timing requirements
+ * met at 25MHz for Default Speed mode, those controllers work at
+ * 19MHz instead
+ */
+#define SDHCI_ARASAN_QUIRK_CLOCK_25_BROKEN BIT(2)
+};
+
+struct sdhci_arasan_of_data {
+	const struct sdhci_arasan_soc_ctl_map *soc_ctl_map;
+	const struct sdhci_pltfm_data *pdata;
+	const struct sdhci_arasan_clk_ops *clk_ops;
+};
+
+static const struct sdhci_arasan_soc_ctl_map rk3399_soc_ctl_map = {
+	.baseclkfreq = { .reg = 0xf000, .width = 8, .shift = 8 },
+	.clockmultiplier = { .reg = 0xf02c, .width = 8, .shift = 0},
+	.hiword_update = true,
+};
+
+static const struct sdhci_arasan_soc_ctl_map intel_lgm_emmc_soc_ctl_map = {
+	.baseclkfreq = { .reg = 0xa0, .width = 8, .shift = 2 },
+	.clockmultiplier = { .reg = 0, .width = -1, .shift = -1 },
+	.hiword_update = false,
+};
+
+static const struct sdhci_arasan_soc_ctl_map intel_lgm_sdxc_soc_ctl_map = {
+	.baseclkfreq = { .reg = 0x80, .width = 8, .shift = 2 },
+	.clockmultiplier = { .reg = 0, .width = -1, .shift = -1 },
+	.hiword_update = false,
+};
+
+static const struct sdhci_arasan_soc_ctl_map intel_keembay_soc_ctl_map = {
+	.baseclkfreq = { .reg = 0x0, .width = 8, .shift = 14 },
+	.clockmultiplier = { .reg = 0x4, .width = 8, .shift = 14 },
+	.support64b = { .reg = 0x4, .width = 1, .shift = 24 },
+	.hiword_update = false,
+};
+
+/**
+ * sdhci_arasan_syscon_write - Write to a field in soc_ctl registers
+ *
+ * @host:	The sdhci_host
+ * @fld:	The field to write to
+ * @val:	The value to write
+ *
+ * This function allows writing to fields in sdhci_arasan_soc_ctl_map.
+ * Note that if a field is specified as not available (shift < 0) then
+ * this function will silently return an error code.  It will be noisy
+ * and print errors for any other (unexpected) errors.
+ *
+ * Return: 0 on success and error value on error
+ */
+static int sdhci_arasan_syscon_write(struct sdhci_host *host,
+				   const struct sdhci_arasan_soc_ctl_field *fld,
+				   u32 val)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
+	struct regmap *soc_ctl_base = sdhci_arasan->soc_ctl_base;
+	u32 reg = fld->reg;
+	u16 width = fld->width;
+	s16 shift = fld->shift;
+	int ret;
+
+	/*
+	 * Silently return errors for shift < 0 so caller doesn't have
+	 * to check for fields which are optional.  For fields that
+	 * are required then caller needs to do something special
+	 * anyway.
+	 */
+	if (shift < 0)
+		return -EINVAL;
+
+	if (sdhci_arasan->soc_ctl_map->hiword_update)
+		ret = regmap_write(soc_ctl_base, reg,
+				   HIWORD_UPDATE(val, GENMASK(width, 0),
+						 shift));
+	else
+		ret = regmap_update_bits(soc_ctl_base, reg,
+					 GENMASK(shift + width, shift),
+					 val << shift);
+
+	/* Yell about (unexpected) regmap errors */
+	if (ret)
+		pr_warn("%s: Regmap write fail: %d\n",
+			 mmc_hostname(host->mmc), ret);
+
+	return ret;
+}
+
+static void sdhci_arasan_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
+	struct sdhci_arasan_clk_data *clk_data = &sdhci_arasan->clk_data;
+	bool ctrl_phy = false;
+
+	if (!IS_ERR(sdhci_arasan->phy)) {
+		if (!sdhci_arasan->is_phy_on && clock <= PHY_CLK_TOO_SLOW_HZ) {
+			/*
+			 * If PHY off, set clock to max speed and power PHY on.
+			 *
+			 * Although PHY docs apparently suggest power cycling
+			 * when changing the clock the PHY doesn't like to be
+			 * powered on while at low speeds like those used in ID
+			 * mode.  Even worse is powering the PHY on while the
+			 * clock is off.
+			 *
+			 * To workaround the PHY limitations, the best we can
+			 * do is to power it on at a faster speed and then slam
+			 * through low speeds without power cycling.
+			 */
+			sdhci_set_clock(host, host->max_clk);
+			if (phy_power_on(sdhci_arasan->phy)) {
+				pr_err("%s: Cannot power on phy.\n",
+				       mmc_hostname(host->mmc));
+				return;
+			}
+
+			sdhci_arasan->is_phy_on = true;
+
+			/*
+			 * We'll now fall through to the below case with
+			 * ctrl_phy = false (so we won't turn off/on).  The
+			 * sdhci_set_clock() will set the real clock.
+			 */
+		} else if (clock > PHY_CLK_TOO_SLOW_HZ) {
+			/*
+			 * At higher clock speeds the PHY is fine being power
+			 * cycled and docs say you _should_ power cycle when
+			 * changing clock speeds.
+			 */
+			ctrl_phy = true;
+		}
+	}
+
+	if (ctrl_phy && sdhci_arasan->is_phy_on) {
+		phy_power_off(sdhci_arasan->phy);
+		sdhci_arasan->is_phy_on = false;
+	}
+
+	if (sdhci_arasan->quirks & SDHCI_ARASAN_QUIRK_CLOCK_25_BROKEN) {
+		/*
+		 * Some of the Arasan variations might not have timing
+		 * requirements met at 25MHz for Default Speed mode,
+		 * those controllers work at 19MHz instead.
+		 */
+		if (clock == DEFAULT_SPEED_MAX_DTR)
+			clock = (DEFAULT_SPEED_MAX_DTR * 19) / 25;
+	}
+
+	/* Set the Input and Output Clock Phase Delays */
+	if (clk_data->set_clk_delays)
+		clk_data->set_clk_delays(host);
+
+	sdhci_set_clock(host, clock);
+
+	if (sdhci_arasan->quirks & SDHCI_ARASAN_QUIRK_CLOCK_UNSTABLE)
+		/*
+		 * Some controllers immediately report SDHCI_CLOCK_INT_STABLE
+		 * after enabling the clock even though the clock is not
+		 * stable. Trying to use a clock without waiting here results
+		 * in EILSEQ while detecting some older/slower cards. The
+		 * chosen delay is the maximum delay from sdhci_set_clock.
+		 */
+		msleep(20);
+
+	if (ctrl_phy) {
+		if (phy_power_on(sdhci_arasan->phy)) {
+			pr_err("%s: Cannot power on phy.\n",
+			       mmc_hostname(host->mmc));
+			return;
+		}
+
+		sdhci_arasan->is_phy_on = true;
+	}
+}
+
+static void sdhci_arasan_hs400_enhanced_strobe(struct mmc_host *mmc,
+					struct mmc_ios *ios)
+{
+	u32 vendor;
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	vendor = sdhci_readl(host, SDHCI_ARASAN_VENDOR_REGISTER);
+	if (ios->enhanced_strobe)
+		vendor |= VENDOR_ENHANCED_STROBE;
+	else
+		vendor &= ~VENDOR_ENHANCED_STROBE;
+
+	sdhci_writel(host, vendor, SDHCI_ARASAN_VENDOR_REGISTER);
+}
+
+static void sdhci_arasan_reset(struct sdhci_host *host, u8 mask)
+{
+	u8 ctrl;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
+
+	sdhci_and_cqhci_reset(host, mask);
+
+	if (sdhci_arasan->quirks & SDHCI_ARASAN_QUIRK_FORCE_CDTEST) {
+		ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+		ctrl |= SDHCI_CTRL_CDTEST_INS | SDHCI_CTRL_CDTEST_EN;
+		sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+	}
+}
+
+static int sdhci_arasan_voltage_switch(struct mmc_host *mmc,
+				       struct mmc_ios *ios)
+{
+	switch (ios->signal_voltage) {
+	case MMC_SIGNAL_VOLTAGE_180:
+		/*
+		 * Plese don't switch to 1V8 as arasan,5.1 doesn't
+		 * actually refer to this setting to indicate the
+		 * signal voltage and the state machine will be broken
+		 * actually if we force to enable 1V8. That's something
+		 * like broken quirk but we could work around here.
+		 */
+		return 0;
+	case MMC_SIGNAL_VOLTAGE_330:
+	case MMC_SIGNAL_VOLTAGE_120:
+		/* We don't support 3V3 and 1V2 */
+		break;
+	}
+
+	return -EINVAL;
+}
+
+static const struct sdhci_ops sdhci_arasan_ops = {
+	.set_clock = sdhci_arasan_set_clock,
+	.get_max_clock = sdhci_pltfm_clk_get_max_clock,
+	.get_timeout_clock = sdhci_pltfm_clk_get_max_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_arasan_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+	.set_power = sdhci_set_power_and_bus_voltage,
+};
+
+static u32 sdhci_arasan_cqhci_irq(struct sdhci_host *host, u32 intmask)
+{
+	int cmd_error = 0;
+	int data_error = 0;
+
+	if (!sdhci_cqe_irq(host, intmask, &cmd_error, &data_error))
+		return intmask;
+
+	cqhci_irq(host->mmc, intmask, cmd_error, data_error);
+
+	return 0;
+}
+
+static void sdhci_arasan_dumpregs(struct mmc_host *mmc)
+{
+	sdhci_dumpregs(mmc_priv(mmc));
+}
+
+static void sdhci_arasan_cqe_enable(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	u32 reg;
+
+	reg = sdhci_readl(host, SDHCI_PRESENT_STATE);
+	while (reg & SDHCI_DATA_AVAILABLE) {
+		sdhci_readl(host, SDHCI_BUFFER);
+		reg = sdhci_readl(host, SDHCI_PRESENT_STATE);
+	}
+
+	sdhci_cqe_enable(mmc);
+}
+
+static const struct cqhci_host_ops sdhci_arasan_cqhci_ops = {
+	.enable         = sdhci_arasan_cqe_enable,
+	.disable        = sdhci_cqe_disable,
+	.dumpregs       = sdhci_arasan_dumpregs,
+};
+
+static const struct sdhci_ops sdhci_arasan_cqe_ops = {
+	.set_clock = sdhci_arasan_set_clock,
+	.get_max_clock = sdhci_pltfm_clk_get_max_clock,
+	.get_timeout_clock = sdhci_pltfm_clk_get_max_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_arasan_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+	.set_power = sdhci_set_power_and_bus_voltage,
+	.irq = sdhci_arasan_cqhci_irq,
+};
+
+static const struct sdhci_pltfm_data sdhci_arasan_cqe_pdata = {
+	.ops = &sdhci_arasan_cqe_ops,
+	.quirks = SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+			SDHCI_QUIRK2_CLOCK_DIV_ZERO_BROKEN,
+};
+
+#ifdef CONFIG_PM_SLEEP
+/**
+ * sdhci_arasan_suspend - Suspend method for the driver
+ * @dev:	Address of the device structure
+ *
+ * Put the device in a low power state.
+ *
+ * Return: 0 on success and error value on error
+ */
+static int sdhci_arasan_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	if (sdhci_arasan->has_cqe) {
+		ret = cqhci_suspend(host->mmc);
+		if (ret)
+			return ret;
+	}
+
+	ret = sdhci_suspend_host(host);
+	if (ret)
+		return ret;
+
+	if (!IS_ERR(sdhci_arasan->phy) && sdhci_arasan->is_phy_on) {
+		ret = phy_power_off(sdhci_arasan->phy);
+		if (ret) {
+			dev_err(dev, "Cannot power off phy.\n");
+			if (sdhci_resume_host(host))
+				dev_err(dev, "Cannot resume host.\n");
+
+			return ret;
+		}
+		sdhci_arasan->is_phy_on = false;
+	}
+
+	clk_disable(pltfm_host->clk);
+	clk_disable(sdhci_arasan->clk_ahb);
+
+	return 0;
+}
+
+/**
+ * sdhci_arasan_resume - Resume method for the driver
+ * @dev:	Address of the device structure
+ *
+ * Resume operation after suspend
+ *
+ * Return: 0 on success and error value on error
+ */
+static int sdhci_arasan_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	ret = clk_enable(sdhci_arasan->clk_ahb);
+	if (ret) {
+		dev_err(dev, "Cannot enable AHB clock.\n");
+		return ret;
+	}
+
+	ret = clk_enable(pltfm_host->clk);
+	if (ret) {
+		dev_err(dev, "Cannot enable SD clock.\n");
+		return ret;
+	}
+
+	if (!IS_ERR(sdhci_arasan->phy) && host->mmc->actual_clock) {
+		ret = phy_power_on(sdhci_arasan->phy);
+		if (ret) {
+			dev_err(dev, "Cannot power on phy.\n");
+			return ret;
+		}
+		sdhci_arasan->is_phy_on = true;
+	}
+
+	ret = sdhci_resume_host(host);
+	if (ret) {
+		dev_err(dev, "Cannot resume host.\n");
+		return ret;
+	}
+
+	if (sdhci_arasan->has_cqe)
+		return cqhci_resume(host->mmc);
+
+	return 0;
+}
+#endif /* ! CONFIG_PM_SLEEP */
+
+static SIMPLE_DEV_PM_OPS(sdhci_arasan_dev_pm_ops, sdhci_arasan_suspend,
+			 sdhci_arasan_resume);
+
+/**
+ * sdhci_arasan_sdcardclk_recalc_rate - Return the card clock rate
+ *
+ * @hw:			Pointer to the hardware clock structure.
+ * @parent_rate:		The parent rate (should be rate of clk_xin).
+ *
+ * Return the current actual rate of the SD card clock.  This can be used
+ * to communicate with out PHY.
+ *
+ * Return: The card clock rate.
+ */
+static unsigned long sdhci_arasan_sdcardclk_recalc_rate(struct clk_hw *hw,
+						      unsigned long parent_rate)
+{
+	struct sdhci_arasan_clk_data *clk_data =
+		container_of(hw, struct sdhci_arasan_clk_data, sdcardclk_hw);
+	struct sdhci_arasan_data *sdhci_arasan =
+		container_of(clk_data, struct sdhci_arasan_data, clk_data);
+	struct sdhci_host *host = sdhci_arasan->host;
+
+	return host->mmc->actual_clock;
+}
+
+static const struct clk_ops arasan_sdcardclk_ops = {
+	.recalc_rate = sdhci_arasan_sdcardclk_recalc_rate,
+};
+
+/**
+ * sdhci_arasan_sampleclk_recalc_rate - Return the sampling clock rate
+ *
+ * @hw:			Pointer to the hardware clock structure.
+ * @parent_rate:		The parent rate (should be rate of clk_xin).
+ *
+ * Return the current actual rate of the sampling clock.  This can be used
+ * to communicate with out PHY.
+ *
+ * Return: The sample clock rate.
+ */
+static unsigned long sdhci_arasan_sampleclk_recalc_rate(struct clk_hw *hw,
+						      unsigned long parent_rate)
+{
+	struct sdhci_arasan_clk_data *clk_data =
+		container_of(hw, struct sdhci_arasan_clk_data, sampleclk_hw);
+	struct sdhci_arasan_data *sdhci_arasan =
+		container_of(clk_data, struct sdhci_arasan_data, clk_data);
+	struct sdhci_host *host = sdhci_arasan->host;
+
+	return host->mmc->actual_clock;
+}
+
+static const struct clk_ops arasan_sampleclk_ops = {
+	.recalc_rate = sdhci_arasan_sampleclk_recalc_rate,
+};
+
+/**
+ * sdhci_zynqmp_sdcardclk_set_phase - Set the SD Output Clock Tap Delays
+ *
+ * @hw:			Pointer to the hardware clock structure.
+ * @degrees:		The clock phase shift between 0 - 359.
+ *
+ * Set the SD Output Clock Tap Delays for Output path
+ *
+ * Return: 0 on success and error value on error
+ */
+static int sdhci_zynqmp_sdcardclk_set_phase(struct clk_hw *hw, int degrees)
+{
+	struct sdhci_arasan_clk_data *clk_data =
+		container_of(hw, struct sdhci_arasan_clk_data, sdcardclk_hw);
+	struct sdhci_arasan_data *sdhci_arasan =
+		container_of(clk_data, struct sdhci_arasan_data, clk_data);
+	struct sdhci_host *host = sdhci_arasan->host;
+	const char *clk_name = clk_hw_get_name(hw);
+	u32 node_id = !strcmp(clk_name, "clk_out_sd0") ? NODE_SD_0 : NODE_SD_1;
+	u8 tap_delay, tap_max = 0;
+	int ret;
+
+	/* This is applicable for SDHCI_SPEC_300 and above */
+	if (host->version < SDHCI_SPEC_300)
+		return 0;
+
+	switch (host->timing) {
+	case MMC_TIMING_MMC_HS:
+	case MMC_TIMING_SD_HS:
+	case MMC_TIMING_UHS_SDR25:
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_MMC_DDR52:
+		/* For 50MHz clock, 30 Taps are available */
+		tap_max = 30;
+		break;
+	case MMC_TIMING_UHS_SDR50:
+		/* For 100MHz clock, 15 Taps are available */
+		tap_max = 15;
+		break;
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_MMC_HS200:
+		/* For 200MHz clock, 8 Taps are available */
+		tap_max = 8;
+	default:
+		break;
+	}
+
+	tap_delay = (degrees * tap_max) / 360;
+
+	/* Set the Clock Phase */
+	ret = zynqmp_pm_set_sd_tapdelay(node_id, PM_TAPDELAY_OUTPUT, tap_delay);
+	if (ret)
+		pr_err("Error setting Output Tap Delay\n");
+
+	/* Release DLL Reset */
+	zynqmp_pm_sd_dll_reset(node_id, PM_DLL_RESET_RELEASE);
+
+	return ret;
+}
+
+static const struct clk_ops zynqmp_sdcardclk_ops = {
+	.recalc_rate = sdhci_arasan_sdcardclk_recalc_rate,
+	.set_phase = sdhci_zynqmp_sdcardclk_set_phase,
+};
+
+/**
+ * sdhci_zynqmp_sampleclk_set_phase - Set the SD Input Clock Tap Delays
+ *
+ * @hw:			Pointer to the hardware clock structure.
+ * @degrees:		The clock phase shift between 0 - 359.
+ *
+ * Set the SD Input Clock Tap Delays for Input path
+ *
+ * Return: 0 on success and error value on error
+ */
+static int sdhci_zynqmp_sampleclk_set_phase(struct clk_hw *hw, int degrees)
+{
+	struct sdhci_arasan_clk_data *clk_data =
+		container_of(hw, struct sdhci_arasan_clk_data, sampleclk_hw);
+	struct sdhci_arasan_data *sdhci_arasan =
+		container_of(clk_data, struct sdhci_arasan_data, clk_data);
+	struct sdhci_host *host = sdhci_arasan->host;
+	const char *clk_name = clk_hw_get_name(hw);
+	u32 node_id = !strcmp(clk_name, "clk_in_sd0") ? NODE_SD_0 : NODE_SD_1;
+	u8 tap_delay, tap_max = 0;
+	int ret;
+
+	/* This is applicable for SDHCI_SPEC_300 and above */
+	if (host->version < SDHCI_SPEC_300)
+		return 0;
+
+	/* Assert DLL Reset */
+	zynqmp_pm_sd_dll_reset(node_id, PM_DLL_RESET_ASSERT);
+
+	switch (host->timing) {
+	case MMC_TIMING_MMC_HS:
+	case MMC_TIMING_SD_HS:
+	case MMC_TIMING_UHS_SDR25:
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_MMC_DDR52:
+		/* For 50MHz clock, 120 Taps are available */
+		tap_max = 120;
+		break;
+	case MMC_TIMING_UHS_SDR50:
+		/* For 100MHz clock, 60 Taps are available */
+		tap_max = 60;
+		break;
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_MMC_HS200:
+		/* For 200MHz clock, 30 Taps are available */
+		tap_max = 30;
+	default:
+		break;
+	}
+
+	tap_delay = (degrees * tap_max) / 360;
+
+	/* Set the Clock Phase */
+	ret = zynqmp_pm_set_sd_tapdelay(node_id, PM_TAPDELAY_INPUT, tap_delay);
+	if (ret)
+		pr_err("Error setting Input Tap Delay\n");
+
+	return ret;
+}
+
+static const struct clk_ops zynqmp_sampleclk_ops = {
+	.recalc_rate = sdhci_arasan_sampleclk_recalc_rate,
+	.set_phase = sdhci_zynqmp_sampleclk_set_phase,
+};
+
+/**
+ * sdhci_versal_sdcardclk_set_phase - Set the SD Output Clock Tap Delays
+ *
+ * @hw:			Pointer to the hardware clock structure.
+ * @degrees:		The clock phase shift between 0 - 359.
+ *
+ * Set the SD Output Clock Tap Delays for Output path
+ *
+ * Return: 0 on success and error value on error
+ */
+static int sdhci_versal_sdcardclk_set_phase(struct clk_hw *hw, int degrees)
+{
+	struct sdhci_arasan_clk_data *clk_data =
+		container_of(hw, struct sdhci_arasan_clk_data, sdcardclk_hw);
+	struct sdhci_arasan_data *sdhci_arasan =
+		container_of(clk_data, struct sdhci_arasan_data, clk_data);
+	struct sdhci_host *host = sdhci_arasan->host;
+	u8 tap_delay, tap_max = 0;
+
+	/* This is applicable for SDHCI_SPEC_300 and above */
+	if (host->version < SDHCI_SPEC_300)
+		return 0;
+
+	switch (host->timing) {
+	case MMC_TIMING_MMC_HS:
+	case MMC_TIMING_SD_HS:
+	case MMC_TIMING_UHS_SDR25:
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_MMC_DDR52:
+		/* For 50MHz clock, 30 Taps are available */
+		tap_max = 30;
+		break;
+	case MMC_TIMING_UHS_SDR50:
+		/* For 100MHz clock, 15 Taps are available */
+		tap_max = 15;
+		break;
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_MMC_HS200:
+		/* For 200MHz clock, 8 Taps are available */
+		tap_max = 8;
+	default:
+		break;
+	}
+
+	tap_delay = (degrees * tap_max) / 360;
+
+	/* Set the Clock Phase */
+	if (tap_delay) {
+		u32 regval;
+
+		regval = sdhci_readl(host, SDHCI_ARASAN_OTAPDLY_REGISTER);
+		regval |= SDHCI_OTAPDLY_ENABLE;
+		sdhci_writel(host, regval, SDHCI_ARASAN_OTAPDLY_REGISTER);
+		regval &= ~SDHCI_ARASAN_OTAPDLY_SEL_MASK;
+		regval |= tap_delay;
+		sdhci_writel(host, regval, SDHCI_ARASAN_OTAPDLY_REGISTER);
+	}
+
+	return 0;
+}
+
+static const struct clk_ops versal_sdcardclk_ops = {
+	.recalc_rate = sdhci_arasan_sdcardclk_recalc_rate,
+	.set_phase = sdhci_versal_sdcardclk_set_phase,
+};
+
+/**
+ * sdhci_versal_sampleclk_set_phase - Set the SD Input Clock Tap Delays
+ *
+ * @hw:			Pointer to the hardware clock structure.
+ * @degrees:		The clock phase shift between 0 - 359.
+ *
+ * Set the SD Input Clock Tap Delays for Input path
+ *
+ * Return: 0 on success and error value on error
+ */
+static int sdhci_versal_sampleclk_set_phase(struct clk_hw *hw, int degrees)
+{
+	struct sdhci_arasan_clk_data *clk_data =
+		container_of(hw, struct sdhci_arasan_clk_data, sampleclk_hw);
+	struct sdhci_arasan_data *sdhci_arasan =
+		container_of(clk_data, struct sdhci_arasan_data, clk_data);
+	struct sdhci_host *host = sdhci_arasan->host;
+	u8 tap_delay, tap_max = 0;
+
+	/* This is applicable for SDHCI_SPEC_300 and above */
+	if (host->version < SDHCI_SPEC_300)
+		return 0;
+
+	switch (host->timing) {
+	case MMC_TIMING_MMC_HS:
+	case MMC_TIMING_SD_HS:
+	case MMC_TIMING_UHS_SDR25:
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_MMC_DDR52:
+		/* For 50MHz clock, 120 Taps are available */
+		tap_max = 120;
+		break;
+	case MMC_TIMING_UHS_SDR50:
+		/* For 100MHz clock, 60 Taps are available */
+		tap_max = 60;
+		break;
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_MMC_HS200:
+		/* For 200MHz clock, 30 Taps are available */
+		tap_max = 30;
+	default:
+		break;
+	}
+
+	tap_delay = (degrees * tap_max) / 360;
+
+	/* Set the Clock Phase */
+	if (tap_delay) {
+		u32 regval;
+
+		regval = sdhci_readl(host, SDHCI_ARASAN_ITAPDLY_REGISTER);
+		regval |= SDHCI_ITAPDLY_CHGWIN;
+		sdhci_writel(host, regval, SDHCI_ARASAN_ITAPDLY_REGISTER);
+		regval |= SDHCI_ITAPDLY_ENABLE;
+		sdhci_writel(host, regval, SDHCI_ARASAN_ITAPDLY_REGISTER);
+		regval &= ~SDHCI_ARASAN_ITAPDLY_SEL_MASK;
+		regval |= tap_delay;
+		sdhci_writel(host, regval, SDHCI_ARASAN_ITAPDLY_REGISTER);
+		regval &= ~SDHCI_ITAPDLY_CHGWIN;
+		sdhci_writel(host, regval, SDHCI_ARASAN_ITAPDLY_REGISTER);
+	}
+
+	return 0;
+}
+
+static const struct clk_ops versal_sampleclk_ops = {
+	.recalc_rate = sdhci_arasan_sampleclk_recalc_rate,
+	.set_phase = sdhci_versal_sampleclk_set_phase,
+};
+
+static void arasan_zynqmp_dll_reset(struct sdhci_host *host, u32 deviceid)
+{
+	u16 clk;
+
+	clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+	clk &= ~(SDHCI_CLOCK_CARD_EN | SDHCI_CLOCK_INT_EN);
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+	/* Issue DLL Reset */
+	zynqmp_pm_sd_dll_reset(deviceid, PM_DLL_RESET_PULSE);
+
+	clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+
+	sdhci_enable_clk(host, clk);
+}
+
+static int arasan_zynqmp_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
+	struct clk_hw *hw = &sdhci_arasan->clk_data.sdcardclk_hw;
+	const char *clk_name = clk_hw_get_name(hw);
+	u32 device_id = !strcmp(clk_name, "clk_out_sd0") ? NODE_SD_0 :
+							   NODE_SD_1;
+	int err;
+
+	arasan_zynqmp_dll_reset(host, device_id);
+
+	err = sdhci_execute_tuning(mmc, opcode);
+	if (err)
+		return err;
+
+	arasan_zynqmp_dll_reset(host, device_id);
+
+	return 0;
+}
+
+/**
+ * sdhci_arasan_update_clockmultiplier - Set corecfg_clockmultiplier
+ *
+ * @host:		The sdhci_host
+ * @value:		The value to write
+ *
+ * The corecfg_clockmultiplier is supposed to contain clock multiplier
+ * value of programmable clock generator.
+ *
+ * NOTES:
+ * - Many existing devices don't seem to do this and work fine.  To keep
+ *   compatibility for old hardware where the device tree doesn't provide a
+ *   register map, this function is a noop if a soc_ctl_map hasn't been provided
+ *   for this platform.
+ * - The value of corecfg_clockmultiplier should sync with that of corresponding
+ *   value reading from sdhci_capability_register. So this function is called
+ *   once at probe time and never called again.
+ */
+static void sdhci_arasan_update_clockmultiplier(struct sdhci_host *host,
+						u32 value)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
+	const struct sdhci_arasan_soc_ctl_map *soc_ctl_map =
+		sdhci_arasan->soc_ctl_map;
+
+	/* Having a map is optional */
+	if (!soc_ctl_map)
+		return;
+
+	/* If we have a map, we expect to have a syscon */
+	if (!sdhci_arasan->soc_ctl_base) {
+		pr_warn("%s: Have regmap, but no soc-ctl-syscon\n",
+			mmc_hostname(host->mmc));
+		return;
+	}
+
+	sdhci_arasan_syscon_write(host, &soc_ctl_map->clockmultiplier, value);
+}
+
+/**
+ * sdhci_arasan_update_baseclkfreq - Set corecfg_baseclkfreq
+ *
+ * @host:		The sdhci_host
+ *
+ * The corecfg_baseclkfreq is supposed to contain the MHz of clk_xin.  This
+ * function can be used to make that happen.
+ *
+ * NOTES:
+ * - Many existing devices don't seem to do this and work fine.  To keep
+ *   compatibility for old hardware where the device tree doesn't provide a
+ *   register map, this function is a noop if a soc_ctl_map hasn't been provided
+ *   for this platform.
+ * - It's assumed that clk_xin is not dynamic and that we use the SDHCI divider
+ *   to achieve lower clock rates.  That means that this function is called once
+ *   at probe time and never called again.
+ */
+static void sdhci_arasan_update_baseclkfreq(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
+	const struct sdhci_arasan_soc_ctl_map *soc_ctl_map =
+		sdhci_arasan->soc_ctl_map;
+	u32 mhz = DIV_ROUND_CLOSEST(clk_get_rate(pltfm_host->clk), 1000000);
+
+	/* Having a map is optional */
+	if (!soc_ctl_map)
+		return;
+
+	/* If we have a map, we expect to have a syscon */
+	if (!sdhci_arasan->soc_ctl_base) {
+		pr_warn("%s: Have regmap, but no soc-ctl-syscon\n",
+			mmc_hostname(host->mmc));
+		return;
+	}
+
+	sdhci_arasan_syscon_write(host, &soc_ctl_map->baseclkfreq, mhz);
+}
+
+static void sdhci_arasan_set_clk_delays(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
+	struct sdhci_arasan_clk_data *clk_data = &sdhci_arasan->clk_data;
+
+	clk_set_phase(clk_data->sampleclk,
+		      clk_data->clk_phase_in[host->timing]);
+	clk_set_phase(clk_data->sdcardclk,
+		      clk_data->clk_phase_out[host->timing]);
+}
+
+static void arasan_dt_read_clk_phase(struct device *dev,
+				     struct sdhci_arasan_clk_data *clk_data,
+				     unsigned int timing, const char *prop)
+{
+	struct device_node *np = dev->of_node;
+
+	int clk_phase[2] = {0};
+
+	/*
+	 * Read Tap Delay values from DT, if the DT does not contain the
+	 * Tap Values then use the pre-defined values.
+	 */
+	if (of_property_read_variable_u32_array(np, prop, &clk_phase[0],
+						2, 0)) {
+		dev_dbg(dev, "Using predefined clock phase for %s = %d %d\n",
+			prop, clk_data->clk_phase_in[timing],
+			clk_data->clk_phase_out[timing]);
+		return;
+	}
+
+	/* The values read are Input and Output Clock Delays in order */
+	clk_data->clk_phase_in[timing] = clk_phase[0];
+	clk_data->clk_phase_out[timing] = clk_phase[1];
+}
+
+/**
+ * arasan_dt_parse_clk_phases - Read Clock Delay values from DT
+ *
+ * @dev:		Pointer to our struct device.
+ * @clk_data:		Pointer to the Clock Data structure
+ *
+ * Called at initialization to parse the values of Clock Delays.
+ */
+static void arasan_dt_parse_clk_phases(struct device *dev,
+				       struct sdhci_arasan_clk_data *clk_data)
+{
+	u32 mio_bank = 0;
+	int i;
+
+	/*
+	 * This has been kept as a pointer and is assigned a function here.
+	 * So that different controller variants can assign their own handling
+	 * function.
+	 */
+	clk_data->set_clk_delays = sdhci_arasan_set_clk_delays;
+
+	if (of_device_is_compatible(dev->of_node, "xlnx,zynqmp-8.9a")) {
+		u32 zynqmp_iclk_phase[MMC_TIMING_MMC_HS400 + 1] =
+			ZYNQMP_ICLK_PHASE;
+		u32 zynqmp_oclk_phase[MMC_TIMING_MMC_HS400 + 1] =
+			ZYNQMP_OCLK_PHASE;
+
+		of_property_read_u32(dev->of_node, "xlnx,mio-bank", &mio_bank);
+		if (mio_bank == 2) {
+			zynqmp_oclk_phase[MMC_TIMING_UHS_SDR104] = 90;
+			zynqmp_oclk_phase[MMC_TIMING_MMC_HS200] = 90;
+		}
+
+		for (i = 0; i <= MMC_TIMING_MMC_HS400; i++) {
+			clk_data->clk_phase_in[i] = zynqmp_iclk_phase[i];
+			clk_data->clk_phase_out[i] = zynqmp_oclk_phase[i];
+		}
+	}
+
+	if (of_device_is_compatible(dev->of_node, "xlnx,versal-8.9a")) {
+		u32 versal_iclk_phase[MMC_TIMING_MMC_HS400 + 1] =
+			VERSAL_ICLK_PHASE;
+		u32 versal_oclk_phase[MMC_TIMING_MMC_HS400 + 1] =
+			VERSAL_OCLK_PHASE;
+
+		for (i = 0; i <= MMC_TIMING_MMC_HS400; i++) {
+			clk_data->clk_phase_in[i] = versal_iclk_phase[i];
+			clk_data->clk_phase_out[i] = versal_oclk_phase[i];
+		}
+	}
+
+	arasan_dt_read_clk_phase(dev, clk_data, MMC_TIMING_LEGACY,
+				 "clk-phase-legacy");
+	arasan_dt_read_clk_phase(dev, clk_data, MMC_TIMING_MMC_HS,
+				 "clk-phase-mmc-hs");
+	arasan_dt_read_clk_phase(dev, clk_data, MMC_TIMING_SD_HS,
+				 "clk-phase-sd-hs");
+	arasan_dt_read_clk_phase(dev, clk_data, MMC_TIMING_UHS_SDR12,
+				 "clk-phase-uhs-sdr12");
+	arasan_dt_read_clk_phase(dev, clk_data, MMC_TIMING_UHS_SDR25,
+				 "clk-phase-uhs-sdr25");
+	arasan_dt_read_clk_phase(dev, clk_data, MMC_TIMING_UHS_SDR50,
+				 "clk-phase-uhs-sdr50");
+	arasan_dt_read_clk_phase(dev, clk_data, MMC_TIMING_UHS_SDR104,
+				 "clk-phase-uhs-sdr104");
+	arasan_dt_read_clk_phase(dev, clk_data, MMC_TIMING_UHS_DDR50,
+				 "clk-phase-uhs-ddr50");
+	arasan_dt_read_clk_phase(dev, clk_data, MMC_TIMING_MMC_DDR52,
+				 "clk-phase-mmc-ddr52");
+	arasan_dt_read_clk_phase(dev, clk_data, MMC_TIMING_MMC_HS200,
+				 "clk-phase-mmc-hs200");
+	arasan_dt_read_clk_phase(dev, clk_data, MMC_TIMING_MMC_HS400,
+				 "clk-phase-mmc-hs400");
+}
+
+static const struct sdhci_pltfm_data sdhci_arasan_pdata = {
+	.ops = &sdhci_arasan_ops,
+	.quirks = SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+			SDHCI_QUIRK2_CLOCK_DIV_ZERO_BROKEN |
+			SDHCI_QUIRK2_STOP_WITH_TC,
+};
+
+static const struct sdhci_arasan_clk_ops arasan_clk_ops = {
+	.sdcardclk_ops = &arasan_sdcardclk_ops,
+	.sampleclk_ops = &arasan_sampleclk_ops,
+};
+
+static struct sdhci_arasan_of_data sdhci_arasan_generic_data = {
+	.pdata = &sdhci_arasan_pdata,
+	.clk_ops = &arasan_clk_ops,
+};
+
+static const struct sdhci_pltfm_data sdhci_keembay_emmc_pdata = {
+	.ops = &sdhci_arasan_cqe_ops,
+	.quirks = SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN |
+		SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+		SDHCI_QUIRK_NO_LED |
+		SDHCI_QUIRK_32BIT_DMA_ADDR |
+		SDHCI_QUIRK_32BIT_DMA_SIZE |
+		SDHCI_QUIRK_32BIT_ADMA_SIZE,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+		SDHCI_QUIRK2_CLOCK_DIV_ZERO_BROKEN |
+		SDHCI_QUIRK2_CAPS_BIT63_FOR_HS400 |
+		SDHCI_QUIRK2_STOP_WITH_TC |
+		SDHCI_QUIRK2_BROKEN_64_BIT_DMA,
+};
+
+static const struct sdhci_pltfm_data sdhci_keembay_sd_pdata = {
+	.ops = &sdhci_arasan_ops,
+	.quirks = SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN |
+		SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+		SDHCI_QUIRK_NO_LED |
+		SDHCI_QUIRK_32BIT_DMA_ADDR |
+		SDHCI_QUIRK_32BIT_DMA_SIZE |
+		SDHCI_QUIRK_32BIT_ADMA_SIZE,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+		SDHCI_QUIRK2_CLOCK_DIV_ZERO_BROKEN |
+		SDHCI_QUIRK2_CARD_ON_NEEDS_BUS_ON |
+		SDHCI_QUIRK2_STOP_WITH_TC |
+		SDHCI_QUIRK2_BROKEN_64_BIT_DMA,
+};
+
+static const struct sdhci_pltfm_data sdhci_keembay_sdio_pdata = {
+	.ops = &sdhci_arasan_ops,
+	.quirks = SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN |
+		SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+		SDHCI_QUIRK_NO_LED |
+		SDHCI_QUIRK_32BIT_DMA_ADDR |
+		SDHCI_QUIRK_32BIT_DMA_SIZE |
+		SDHCI_QUIRK_32BIT_ADMA_SIZE,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+		SDHCI_QUIRK2_CLOCK_DIV_ZERO_BROKEN |
+		SDHCI_QUIRK2_HOST_OFF_CARD_ON |
+		SDHCI_QUIRK2_BROKEN_64_BIT_DMA,
+};
+
+static struct sdhci_arasan_of_data sdhci_arasan_rk3399_data = {
+	.soc_ctl_map = &rk3399_soc_ctl_map,
+	.pdata = &sdhci_arasan_cqe_pdata,
+	.clk_ops = &arasan_clk_ops,
+};
+
+static struct sdhci_arasan_of_data intel_lgm_emmc_data = {
+	.soc_ctl_map = &intel_lgm_emmc_soc_ctl_map,
+	.pdata = &sdhci_arasan_cqe_pdata,
+	.clk_ops = &arasan_clk_ops,
+};
+
+static struct sdhci_arasan_of_data intel_lgm_sdxc_data = {
+	.soc_ctl_map = &intel_lgm_sdxc_soc_ctl_map,
+	.pdata = &sdhci_arasan_cqe_pdata,
+	.clk_ops = &arasan_clk_ops,
+};
+
+static const struct sdhci_pltfm_data sdhci_arasan_zynqmp_pdata = {
+	.ops = &sdhci_arasan_ops,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+			SDHCI_QUIRK2_CLOCK_DIV_ZERO_BROKEN |
+			SDHCI_QUIRK2_STOP_WITH_TC,
+};
+
+static const struct sdhci_arasan_clk_ops zynqmp_clk_ops = {
+	.sdcardclk_ops = &zynqmp_sdcardclk_ops,
+	.sampleclk_ops = &zynqmp_sampleclk_ops,
+};
+
+static struct sdhci_arasan_of_data sdhci_arasan_zynqmp_data = {
+	.pdata = &sdhci_arasan_zynqmp_pdata,
+	.clk_ops = &zynqmp_clk_ops,
+};
+
+static const struct sdhci_arasan_clk_ops versal_clk_ops = {
+	.sdcardclk_ops = &versal_sdcardclk_ops,
+	.sampleclk_ops = &versal_sampleclk_ops,
+};
+
+static struct sdhci_arasan_of_data sdhci_arasan_versal_data = {
+	.pdata = &sdhci_arasan_zynqmp_pdata,
+	.clk_ops = &versal_clk_ops,
+};
+
+static struct sdhci_arasan_of_data intel_keembay_emmc_data = {
+	.soc_ctl_map = &intel_keembay_soc_ctl_map,
+	.pdata = &sdhci_keembay_emmc_pdata,
+	.clk_ops = &arasan_clk_ops,
+};
+
+static struct sdhci_arasan_of_data intel_keembay_sd_data = {
+	.soc_ctl_map = &intel_keembay_soc_ctl_map,
+	.pdata = &sdhci_keembay_sd_pdata,
+	.clk_ops = &arasan_clk_ops,
+};
+
+static struct sdhci_arasan_of_data intel_keembay_sdio_data = {
+	.soc_ctl_map = &intel_keembay_soc_ctl_map,
+	.pdata = &sdhci_keembay_sdio_pdata,
+	.clk_ops = &arasan_clk_ops,
+};
+
+static const struct of_device_id sdhci_arasan_of_match[] = {
+	/* SoC-specific compatible strings w/ soc_ctl_map */
+	{
+		.compatible = "rockchip,rk3399-sdhci-5.1",
+		.data = &sdhci_arasan_rk3399_data,
+	},
+	{
+		.compatible = "intel,lgm-sdhci-5.1-emmc",
+		.data = &intel_lgm_emmc_data,
+	},
+	{
+		.compatible = "intel,lgm-sdhci-5.1-sdxc",
+		.data = &intel_lgm_sdxc_data,
+	},
+	{
+		.compatible = "intel,keembay-sdhci-5.1-emmc",
+		.data = &intel_keembay_emmc_data,
+	},
+	{
+		.compatible = "intel,keembay-sdhci-5.1-sd",
+		.data = &intel_keembay_sd_data,
+	},
+	{
+		.compatible = "intel,keembay-sdhci-5.1-sdio",
+		.data = &intel_keembay_sdio_data,
+	},
+	/* Generic compatible below here */
+	{
+		.compatible = "arasan,sdhci-8.9a",
+		.data = &sdhci_arasan_generic_data,
+	},
+	{
+		.compatible = "arasan,sdhci-5.1",
+		.data = &sdhci_arasan_generic_data,
+	},
+	{
+		.compatible = "arasan,sdhci-4.9a",
+		.data = &sdhci_arasan_generic_data,
+	},
+	{
+		.compatible = "xlnx,zynqmp-8.9a",
+		.data = &sdhci_arasan_zynqmp_data,
+	},
+	{
+		.compatible = "xlnx,versal-8.9a",
+		.data = &sdhci_arasan_versal_data,
+	},
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, sdhci_arasan_of_match);
+
+/**
+ * sdhci_arasan_register_sdcardclk - Register the sdcardclk for a PHY to use
+ *
+ * @sdhci_arasan:	Our private data structure.
+ * @clk_xin:		Pointer to the functional clock
+ * @dev:		Pointer to our struct device.
+ *
+ * Some PHY devices need to know what the actual card clock is.  In order for
+ * them to find out, we'll provide a clock through the common clock framework
+ * for them to query.
+ *
+ * Return: 0 on success and error value on error
+ */
+static int
+sdhci_arasan_register_sdcardclk(struct sdhci_arasan_data *sdhci_arasan,
+				struct clk *clk_xin,
+				struct device *dev)
+{
+	struct sdhci_arasan_clk_data *clk_data = &sdhci_arasan->clk_data;
+	struct device_node *np = dev->of_node;
+	struct clk_init_data sdcardclk_init;
+	const char *parent_clk_name;
+	int ret;
+
+	ret = of_property_read_string_index(np, "clock-output-names", 0,
+					    &sdcardclk_init.name);
+	if (ret) {
+		dev_err(dev, "DT has #clock-cells but no clock-output-names\n");
+		return ret;
+	}
+
+	parent_clk_name = __clk_get_name(clk_xin);
+	sdcardclk_init.parent_names = &parent_clk_name;
+	sdcardclk_init.num_parents = 1;
+	sdcardclk_init.flags = CLK_GET_RATE_NOCACHE;
+	sdcardclk_init.ops = sdhci_arasan->clk_ops->sdcardclk_ops;
+
+	clk_data->sdcardclk_hw.init = &sdcardclk_init;
+	clk_data->sdcardclk =
+		devm_clk_register(dev, &clk_data->sdcardclk_hw);
+	if (IS_ERR(clk_data->sdcardclk))
+		return PTR_ERR(clk_data->sdcardclk);
+	clk_data->sdcardclk_hw.init = NULL;
+
+	ret = of_clk_add_provider(np, of_clk_src_simple_get,
+				  clk_data->sdcardclk);
+	if (ret)
+		dev_err(dev, "Failed to add sdcard clock provider\n");
+
+	return ret;
+}
+
+/**
+ * sdhci_arasan_register_sampleclk - Register the sampleclk for a PHY to use
+ *
+ * @sdhci_arasan:	Our private data structure.
+ * @clk_xin:		Pointer to the functional clock
+ * @dev:		Pointer to our struct device.
+ *
+ * Some PHY devices need to know what the actual card clock is.  In order for
+ * them to find out, we'll provide a clock through the common clock framework
+ * for them to query.
+ *
+ * Return: 0 on success and error value on error
+ */
+static int
+sdhci_arasan_register_sampleclk(struct sdhci_arasan_data *sdhci_arasan,
+				struct clk *clk_xin,
+				struct device *dev)
+{
+	struct sdhci_arasan_clk_data *clk_data = &sdhci_arasan->clk_data;
+	struct device_node *np = dev->of_node;
+	struct clk_init_data sampleclk_init;
+	const char *parent_clk_name;
+	int ret;
+
+	ret = of_property_read_string_index(np, "clock-output-names", 1,
+					    &sampleclk_init.name);
+	if (ret) {
+		dev_err(dev, "DT has #clock-cells but no clock-output-names\n");
+		return ret;
+	}
+
+	parent_clk_name = __clk_get_name(clk_xin);
+	sampleclk_init.parent_names = &parent_clk_name;
+	sampleclk_init.num_parents = 1;
+	sampleclk_init.flags = CLK_GET_RATE_NOCACHE;
+	sampleclk_init.ops = sdhci_arasan->clk_ops->sampleclk_ops;
+
+	clk_data->sampleclk_hw.init = &sampleclk_init;
+	clk_data->sampleclk =
+		devm_clk_register(dev, &clk_data->sampleclk_hw);
+	if (IS_ERR(clk_data->sampleclk))
+		return PTR_ERR(clk_data->sampleclk);
+	clk_data->sampleclk_hw.init = NULL;
+
+	ret = of_clk_add_provider(np, of_clk_src_simple_get,
+				  clk_data->sampleclk);
+	if (ret)
+		dev_err(dev, "Failed to add sample clock provider\n");
+
+	return ret;
+}
+
+/**
+ * sdhci_arasan_unregister_sdclk - Undoes sdhci_arasan_register_sdclk()
+ *
+ * @dev:		Pointer to our struct device.
+ *
+ * Should be called any time we're exiting and sdhci_arasan_register_sdclk()
+ * returned success.
+ */
+static void sdhci_arasan_unregister_sdclk(struct device *dev)
+{
+	struct device_node *np = dev->of_node;
+
+	if (!of_find_property(np, "#clock-cells", NULL))
+		return;
+
+	of_clk_del_provider(dev->of_node);
+}
+
+/**
+ * sdhci_arasan_update_support64b - Set SUPPORT_64B (64-bit System Bus Support)
+ *
+ * This should be set based on the System Address Bus.
+ * 0: the Core supports only 32-bit System Address Bus.
+ * 1: the Core supports 64-bit System Address Bus.
+ *
+ * NOTES:
+ * - For Keem Bay, it is required to clear this bit. Its default value is 1'b1.
+ *   Keem Bay does not support 64-bit access.
+ *
+ * @host:		The sdhci_host
+ * @value:		The value to write
+ */
+static void sdhci_arasan_update_support64b(struct sdhci_host *host, u32 value)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
+	const struct sdhci_arasan_soc_ctl_map *soc_ctl_map =
+		sdhci_arasan->soc_ctl_map;
+
+	/* Having a map is optional */
+	if (!soc_ctl_map)
+		return;
+
+	/* If we have a map, we expect to have a syscon */
+	if (!sdhci_arasan->soc_ctl_base) {
+		pr_warn("%s: Have regmap, but no soc-ctl-syscon\n",
+			mmc_hostname(host->mmc));
+		return;
+	}
+
+	sdhci_arasan_syscon_write(host, &soc_ctl_map->support64b, value);
+}
+
+/**
+ * sdhci_arasan_register_sdclk - Register the sdcardclk for a PHY to use
+ *
+ * @sdhci_arasan:	Our private data structure.
+ * @clk_xin:		Pointer to the functional clock
+ * @dev:		Pointer to our struct device.
+ *
+ * Some PHY devices need to know what the actual card clock is.  In order for
+ * them to find out, we'll provide a clock through the common clock framework
+ * for them to query.
+ *
+ * Note: without seriously re-architecting SDHCI's clock code and testing on
+ * all platforms, there's no way to create a totally beautiful clock here
+ * with all clock ops implemented.  Instead, we'll just create a clock that can
+ * be queried and set the CLK_GET_RATE_NOCACHE attribute to tell common clock
+ * framework that we're doing things behind its back.  This should be sufficient
+ * to create nice clean device tree bindings and later (if needed) we can try
+ * re-architecting SDHCI if we see some benefit to it.
+ *
+ * Return: 0 on success and error value on error
+ */
+static int sdhci_arasan_register_sdclk(struct sdhci_arasan_data *sdhci_arasan,
+				       struct clk *clk_xin,
+				       struct device *dev)
+{
+	struct device_node *np = dev->of_node;
+	u32 num_clks = 0;
+	int ret;
+
+	/* Providing a clock to the PHY is optional; no error if missing */
+	if (of_property_read_u32(np, "#clock-cells", &num_clks) < 0)
+		return 0;
+
+	ret = sdhci_arasan_register_sdcardclk(sdhci_arasan, clk_xin, dev);
+	if (ret)
+		return ret;
+
+	if (num_clks) {
+		ret = sdhci_arasan_register_sampleclk(sdhci_arasan, clk_xin,
+						      dev);
+		if (ret) {
+			sdhci_arasan_unregister_sdclk(dev);
+			return ret;
+		}
+	}
+
+	return 0;
+}
+
+static int sdhci_arasan_add_host(struct sdhci_arasan_data *sdhci_arasan)
+{
+	struct sdhci_host *host = sdhci_arasan->host;
+	struct cqhci_host *cq_host;
+	bool dma64;
+	int ret;
+
+	if (!sdhci_arasan->has_cqe)
+		return sdhci_add_host(host);
+
+	ret = sdhci_setup_host(host);
+	if (ret)
+		return ret;
+
+	cq_host = devm_kzalloc(host->mmc->parent,
+			       sizeof(*cq_host), GFP_KERNEL);
+	if (!cq_host) {
+		ret = -ENOMEM;
+		goto cleanup;
+	}
+
+	cq_host->mmio = host->ioaddr + SDHCI_ARASAN_CQE_BASE_ADDR;
+	cq_host->ops = &sdhci_arasan_cqhci_ops;
+
+	dma64 = host->flags & SDHCI_USE_64_BIT_DMA;
+	if (dma64)
+		cq_host->caps |= CQHCI_TASK_DESC_SZ_128;
+
+	ret = cqhci_init(cq_host, host->mmc, dma64);
+	if (ret)
+		goto cleanup;
+
+	ret = __sdhci_add_host(host);
+	if (ret)
+		goto cleanup;
+
+	return 0;
+
+cleanup:
+	sdhci_cleanup_host(host);
+	return ret;
+}
+
+static int sdhci_arasan_probe(struct platform_device *pdev)
+{
+	int ret;
+	const struct of_device_id *match;
+	struct device_node *node;
+	struct clk *clk_xin;
+	struct sdhci_host *host;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_arasan_data *sdhci_arasan;
+	struct device_node *np = pdev->dev.of_node;
+	const struct sdhci_arasan_of_data *data;
+
+	match = of_match_node(sdhci_arasan_of_match, pdev->dev.of_node);
+	data = match->data;
+	host = sdhci_pltfm_init(pdev, data->pdata, sizeof(*sdhci_arasan));
+
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
+	sdhci_arasan->host = host;
+
+	sdhci_arasan->soc_ctl_map = data->soc_ctl_map;
+	sdhci_arasan->clk_ops = data->clk_ops;
+
+	node = of_parse_phandle(pdev->dev.of_node, "arasan,soc-ctl-syscon", 0);
+	if (node) {
+		sdhci_arasan->soc_ctl_base = syscon_node_to_regmap(node);
+		of_node_put(node);
+
+		if (IS_ERR(sdhci_arasan->soc_ctl_base)) {
+			ret = dev_err_probe(&pdev->dev,
+					    PTR_ERR(sdhci_arasan->soc_ctl_base),
+					    "Can't get syscon\n");
+			goto err_pltfm_free;
+		}
+	}
+
+	sdhci_arasan->clk_ahb = devm_clk_get(&pdev->dev, "clk_ahb");
+	if (IS_ERR(sdhci_arasan->clk_ahb)) {
+		dev_err(&pdev->dev, "clk_ahb clock not found.\n");
+		ret = PTR_ERR(sdhci_arasan->clk_ahb);
+		goto err_pltfm_free;
+	}
+
+	clk_xin = devm_clk_get(&pdev->dev, "clk_xin");
+	if (IS_ERR(clk_xin)) {
+		dev_err(&pdev->dev, "clk_xin clock not found.\n");
+		ret = PTR_ERR(clk_xin);
+		goto err_pltfm_free;
+	}
+
+	ret = clk_prepare_enable(sdhci_arasan->clk_ahb);
+	if (ret) {
+		dev_err(&pdev->dev, "Unable to enable AHB clock.\n");
+		goto err_pltfm_free;
+	}
+
+	ret = clk_prepare_enable(clk_xin);
+	if (ret) {
+		dev_err(&pdev->dev, "Unable to enable SD clock.\n");
+		goto clk_dis_ahb;
+	}
+
+	sdhci_get_of_property(pdev);
+
+	if (of_property_read_bool(np, "xlnx,fails-without-test-cd"))
+		sdhci_arasan->quirks |= SDHCI_ARASAN_QUIRK_FORCE_CDTEST;
+
+	if (of_property_read_bool(np, "xlnx,int-clock-stable-broken"))
+		sdhci_arasan->quirks |= SDHCI_ARASAN_QUIRK_CLOCK_UNSTABLE;
+
+	pltfm_host->clk = clk_xin;
+
+	if (of_device_is_compatible(pdev->dev.of_node,
+				    "rockchip,rk3399-sdhci-5.1"))
+		sdhci_arasan_update_clockmultiplier(host, 0x0);
+
+	if (of_device_is_compatible(np, "intel,keembay-sdhci-5.1-emmc") ||
+	    of_device_is_compatible(np, "intel,keembay-sdhci-5.1-sd") ||
+	    of_device_is_compatible(np, "intel,keembay-sdhci-5.1-sdio")) {
+		sdhci_arasan_update_clockmultiplier(host, 0x0);
+		sdhci_arasan_update_support64b(host, 0x0);
+
+		host->mmc->caps |= MMC_CAP_WAIT_WHILE_BUSY;
+	}
+
+	sdhci_arasan_update_baseclkfreq(host);
+
+	ret = sdhci_arasan_register_sdclk(sdhci_arasan, clk_xin, &pdev->dev);
+	if (ret)
+		goto clk_disable_all;
+
+	if (of_device_is_compatible(np, "xlnx,zynqmp-8.9a")) {
+		host->mmc_host_ops.execute_tuning =
+			arasan_zynqmp_execute_tuning;
+
+		sdhci_arasan->quirks |= SDHCI_ARASAN_QUIRK_CLOCK_25_BROKEN;
+	}
+
+	arasan_dt_parse_clk_phases(&pdev->dev, &sdhci_arasan->clk_data);
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret) {
+		if (ret != -EPROBE_DEFER)
+			dev_err(&pdev->dev, "parsing dt failed (%d)\n", ret);
+		goto unreg_clk;
+	}
+
+	sdhci_arasan->phy = ERR_PTR(-ENODEV);
+	if (of_device_is_compatible(pdev->dev.of_node,
+				    "arasan,sdhci-5.1")) {
+		sdhci_arasan->phy = devm_phy_get(&pdev->dev,
+						 "phy_arasan");
+		if (IS_ERR(sdhci_arasan->phy)) {
+			ret = PTR_ERR(sdhci_arasan->phy);
+			dev_err(&pdev->dev, "No phy for arasan,sdhci-5.1.\n");
+			goto unreg_clk;
+		}
+
+		ret = phy_init(sdhci_arasan->phy);
+		if (ret < 0) {
+			dev_err(&pdev->dev, "phy_init err.\n");
+			goto unreg_clk;
+		}
+
+		host->mmc_host_ops.hs400_enhanced_strobe =
+					sdhci_arasan_hs400_enhanced_strobe;
+		host->mmc_host_ops.start_signal_voltage_switch =
+					sdhci_arasan_voltage_switch;
+		sdhci_arasan->has_cqe = true;
+		host->mmc->caps2 |= MMC_CAP2_CQE;
+
+		if (!of_property_read_bool(np, "disable-cqe-dcmd"))
+			host->mmc->caps2 |= MMC_CAP2_CQE_DCMD;
+	}
+
+	ret = sdhci_arasan_add_host(sdhci_arasan);
+	if (ret)
+		goto err_add_host;
+
+	return 0;
+
+err_add_host:
+	if (!IS_ERR(sdhci_arasan->phy))
+		phy_exit(sdhci_arasan->phy);
+unreg_clk:
+	sdhci_arasan_unregister_sdclk(&pdev->dev);
+clk_disable_all:
+	clk_disable_unprepare(clk_xin);
+clk_dis_ahb:
+	clk_disable_unprepare(sdhci_arasan->clk_ahb);
+err_pltfm_free:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static int sdhci_arasan_remove(struct platform_device *pdev)
+{
+	int ret;
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
+	struct clk *clk_ahb = sdhci_arasan->clk_ahb;
+
+	if (!IS_ERR(sdhci_arasan->phy)) {
+		if (sdhci_arasan->is_phy_on)
+			phy_power_off(sdhci_arasan->phy);
+		phy_exit(sdhci_arasan->phy);
+	}
+
+	sdhci_arasan_unregister_sdclk(&pdev->dev);
+
+	ret = sdhci_pltfm_unregister(pdev);
+
+	clk_disable_unprepare(clk_ahb);
+
+	return ret;
+}
+
+static struct platform_driver sdhci_arasan_driver = {
+	.driver = {
+		.name = "sdhci-arasan",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = sdhci_arasan_of_match,
+		.pm = &sdhci_arasan_dev_pm_ops,
+	},
+	.probe = sdhci_arasan_probe,
+	.remove = sdhci_arasan_remove,
+};
+
+module_platform_driver(sdhci_arasan_driver);
+
+MODULE_DESCRIPTION("Driver for the Arasan SDHCI Controller");
+MODULE_AUTHOR("Soeren Brinkmann <soren.brinkmann@xilinx.com>");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mmc/host/sdhci-of-aspeed.c b/drivers/mmc/host/sdhci-of-aspeed.c
new file mode 100644
index 000000000..4f008ba32
--- /dev/null
+++ b/drivers/mmc/host/sdhci-of-aspeed.c
@@ -0,0 +1,356 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Copyright (C) 2019 ASPEED Technology Inc. */
+/* Copyright (C) 2019 IBM Corp. */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/io.h>
+#include <linux/mmc/host.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/of.h>
+#include <linux/of_platform.h>
+#include <linux/platform_device.h>
+#include <linux/spinlock.h>
+
+#include "sdhci-pltfm.h"
+
+#define ASPEED_SDC_INFO		0x00
+#define   ASPEED_SDC_S1MMC8	BIT(25)
+#define   ASPEED_SDC_S0MMC8	BIT(24)
+
+struct aspeed_sdc {
+	struct clk *clk;
+	struct resource *res;
+
+	spinlock_t lock;
+	void __iomem *regs;
+};
+
+struct aspeed_sdhci {
+	struct aspeed_sdc *parent;
+	u32 width_mask;
+};
+
+static void aspeed_sdc_configure_8bit_mode(struct aspeed_sdc *sdc,
+					   struct aspeed_sdhci *sdhci,
+					   bool bus8)
+{
+	u32 info;
+
+	/* Set/clear 8 bit mode */
+	spin_lock(&sdc->lock);
+	info = readl(sdc->regs + ASPEED_SDC_INFO);
+	if (bus8)
+		info |= sdhci->width_mask;
+	else
+		info &= ~sdhci->width_mask;
+	writel(info, sdc->regs + ASPEED_SDC_INFO);
+	spin_unlock(&sdc->lock);
+}
+
+static void aspeed_sdhci_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	struct sdhci_pltfm_host *pltfm_host;
+	unsigned long parent;
+	int div;
+	u16 clk;
+
+	pltfm_host = sdhci_priv(host);
+	parent = clk_get_rate(pltfm_host->clk);
+
+	sdhci_writew(host, 0, SDHCI_CLOCK_CONTROL);
+
+	if (clock == 0)
+		return;
+
+	if (WARN_ON(clock > host->max_clk))
+		clock = host->max_clk;
+
+	for (div = 2; div < 256; div *= 2) {
+		if ((parent / div) <= clock)
+			break;
+	}
+	div >>= 1;
+
+	clk = div << SDHCI_DIVIDER_SHIFT;
+
+	sdhci_enable_clk(host, clk);
+}
+
+static unsigned int aspeed_sdhci_get_max_clock(struct sdhci_host *host)
+{
+	if (host->mmc->f_max)
+		return host->mmc->f_max;
+
+	return sdhci_pltfm_clk_get_max_clock(host);
+}
+
+static void aspeed_sdhci_set_bus_width(struct sdhci_host *host, int width)
+{
+	struct sdhci_pltfm_host *pltfm_priv;
+	struct aspeed_sdhci *aspeed_sdhci;
+	struct aspeed_sdc *aspeed_sdc;
+	u8 ctrl;
+
+	pltfm_priv = sdhci_priv(host);
+	aspeed_sdhci = sdhci_pltfm_priv(pltfm_priv);
+	aspeed_sdc = aspeed_sdhci->parent;
+
+	/* Set/clear 8-bit mode */
+	aspeed_sdc_configure_8bit_mode(aspeed_sdc, aspeed_sdhci,
+				       width == MMC_BUS_WIDTH_8);
+
+	/* Set/clear 1 or 4 bit mode */
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+	if (width == MMC_BUS_WIDTH_4)
+		ctrl |= SDHCI_CTRL_4BITBUS;
+	else
+		ctrl &= ~SDHCI_CTRL_4BITBUS;
+	sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+}
+
+static u32 aspeed_sdhci_readl(struct sdhci_host *host, int reg)
+{
+	u32 val = readl(host->ioaddr + reg);
+
+	if (unlikely(reg == SDHCI_PRESENT_STATE) &&
+	    (host->mmc->caps2 & MMC_CAP2_CD_ACTIVE_HIGH))
+		val ^= SDHCI_CARD_PRESENT;
+
+	return val;
+}
+
+static const struct sdhci_ops aspeed_sdhci_ops = {
+	.read_l = aspeed_sdhci_readl,
+	.set_clock = aspeed_sdhci_set_clock,
+	.get_max_clock = aspeed_sdhci_get_max_clock,
+	.set_bus_width = aspeed_sdhci_set_bus_width,
+	.get_timeout_clock = sdhci_pltfm_clk_get_max_clock,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static const struct sdhci_pltfm_data aspeed_sdhci_pdata = {
+	.ops = &aspeed_sdhci_ops,
+	.quirks = SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+};
+
+static inline int aspeed_sdhci_calculate_slot(struct aspeed_sdhci *dev,
+					      struct resource *res)
+{
+	resource_size_t delta;
+
+	if (!res || resource_type(res) != IORESOURCE_MEM)
+		return -EINVAL;
+
+	if (res->start < dev->parent->res->start)
+		return -EINVAL;
+
+	delta = res->start - dev->parent->res->start;
+	if (delta & (0x100 - 1))
+		return -EINVAL;
+
+	return (delta / 0x100) - 1;
+}
+
+static int aspeed_sdhci_probe(struct platform_device *pdev)
+{
+	struct sdhci_pltfm_host *pltfm_host;
+	struct aspeed_sdhci *dev;
+	struct sdhci_host *host;
+	struct resource *res;
+	int slot;
+	int ret;
+
+	host = sdhci_pltfm_init(pdev, &aspeed_sdhci_pdata, sizeof(*dev));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	dev = sdhci_pltfm_priv(pltfm_host);
+	dev->parent = dev_get_drvdata(pdev->dev.parent);
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	slot = aspeed_sdhci_calculate_slot(dev, res);
+
+	if (slot < 0)
+		return slot;
+	else if (slot >= 2)
+		return -EINVAL;
+
+	dev_info(&pdev->dev, "Configuring for slot %d\n", slot);
+	dev->width_mask = !slot ? ASPEED_SDC_S0MMC8 : ASPEED_SDC_S1MMC8;
+
+	sdhci_get_of_property(pdev);
+
+	pltfm_host->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(pltfm_host->clk))
+		return PTR_ERR(pltfm_host->clk);
+
+	ret = clk_prepare_enable(pltfm_host->clk);
+	if (ret) {
+		dev_err(&pdev->dev, "Unable to enable SDIO clock\n");
+		goto err_pltfm_free;
+	}
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto err_sdhci_add;
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err_sdhci_add;
+
+	return 0;
+
+err_sdhci_add:
+	clk_disable_unprepare(pltfm_host->clk);
+err_pltfm_free:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static int aspeed_sdhci_remove(struct platform_device *pdev)
+{
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_host *host;
+	int dead = 0;
+
+	host = platform_get_drvdata(pdev);
+	pltfm_host = sdhci_priv(host);
+
+	sdhci_remove_host(host, dead);
+
+	clk_disable_unprepare(pltfm_host->clk);
+
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+
+static const struct of_device_id aspeed_sdhci_of_match[] = {
+	{ .compatible = "aspeed,ast2400-sdhci", },
+	{ .compatible = "aspeed,ast2500-sdhci", },
+	{ .compatible = "aspeed,ast2600-sdhci", },
+	{ }
+};
+
+static struct platform_driver aspeed_sdhci_driver = {
+	.driver		= {
+		.name	= "sdhci-aspeed",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = aspeed_sdhci_of_match,
+	},
+	.probe		= aspeed_sdhci_probe,
+	.remove		= aspeed_sdhci_remove,
+};
+
+static int aspeed_sdc_probe(struct platform_device *pdev)
+
+{
+	struct device_node *parent, *child;
+	struct aspeed_sdc *sdc;
+	int ret;
+
+	sdc = devm_kzalloc(&pdev->dev, sizeof(*sdc), GFP_KERNEL);
+	if (!sdc)
+		return -ENOMEM;
+
+	spin_lock_init(&sdc->lock);
+
+	sdc->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(sdc->clk))
+		return PTR_ERR(sdc->clk);
+
+	ret = clk_prepare_enable(sdc->clk);
+	if (ret) {
+		dev_err(&pdev->dev, "Unable to enable SDCLK\n");
+		return ret;
+	}
+
+	sdc->res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	sdc->regs = devm_ioremap_resource(&pdev->dev, sdc->res);
+	if (IS_ERR(sdc->regs)) {
+		ret = PTR_ERR(sdc->regs);
+		goto err_clk;
+	}
+
+	dev_set_drvdata(&pdev->dev, sdc);
+
+	parent = pdev->dev.of_node;
+	for_each_available_child_of_node(parent, child) {
+		struct platform_device *cpdev;
+
+		cpdev = of_platform_device_create(child, NULL, &pdev->dev);
+		if (!cpdev) {
+			of_node_put(child);
+			ret = -ENODEV;
+			goto err_clk;
+		}
+	}
+
+	return 0;
+
+err_clk:
+	clk_disable_unprepare(sdc->clk);
+	return ret;
+}
+
+static int aspeed_sdc_remove(struct platform_device *pdev)
+{
+	struct aspeed_sdc *sdc = dev_get_drvdata(&pdev->dev);
+
+	clk_disable_unprepare(sdc->clk);
+
+	return 0;
+}
+
+static const struct of_device_id aspeed_sdc_of_match[] = {
+	{ .compatible = "aspeed,ast2400-sd-controller", },
+	{ .compatible = "aspeed,ast2500-sd-controller", },
+	{ .compatible = "aspeed,ast2600-sd-controller", },
+	{ }
+};
+
+MODULE_DEVICE_TABLE(of, aspeed_sdc_of_match);
+
+static struct platform_driver aspeed_sdc_driver = {
+	.driver		= {
+		.name	= "sd-controller-aspeed",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= &sdhci_pltfm_pmops,
+		.of_match_table = aspeed_sdc_of_match,
+	},
+	.probe		= aspeed_sdc_probe,
+	.remove		= aspeed_sdc_remove,
+};
+
+static int __init aspeed_sdc_init(void)
+{
+	int rc;
+
+	rc = platform_driver_register(&aspeed_sdhci_driver);
+	if (rc < 0)
+		return rc;
+
+	rc = platform_driver_register(&aspeed_sdc_driver);
+	if (rc < 0)
+		platform_driver_unregister(&aspeed_sdhci_driver);
+
+	return rc;
+}
+module_init(aspeed_sdc_init);
+
+static void __exit aspeed_sdc_exit(void)
+{
+	platform_driver_unregister(&aspeed_sdc_driver);
+	platform_driver_unregister(&aspeed_sdhci_driver);
+}
+module_exit(aspeed_sdc_exit);
+
+MODULE_DESCRIPTION("Driver for the ASPEED SD/SDIO/SDHCI Controllers");
+MODULE_AUTHOR("Ryan Chen <ryan_chen@aspeedtech.com>");
+MODULE_AUTHOR("Andrew Jeffery <andrew@aj.id.au>");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mmc/host/sdhci-of-at91.c b/drivers/mmc/host/sdhci-of-at91.c
new file mode 100644
index 000000000..0452c312b
--- /dev/null
+++ b/drivers/mmc/host/sdhci-of-at91.c
@@ -0,0 +1,485 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Atmel SDMMC controller driver.
+ *
+ * Copyright (C) 2015 Atmel,
+ *		 2015 Ludovic Desroches <ludovic.desroches@atmel.com>
+ */
+
+#include <linux/bitfield.h>
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/io.h>
+#include <linux/iopoll.h>
+#include <linux/kernel.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+
+#include "sdhci-pltfm.h"
+
+#define SDMMC_MC1R	0x204
+#define		SDMMC_MC1R_DDR		BIT(3)
+#define		SDMMC_MC1R_FCD		BIT(7)
+#define SDMMC_CACR	0x230
+#define		SDMMC_CACR_CAPWREN	BIT(0)
+#define		SDMMC_CACR_KEY		(0x46 << 8)
+#define SDMMC_CALCR	0x240
+#define		SDMMC_CALCR_EN		BIT(0)
+#define		SDMMC_CALCR_ALWYSON	BIT(4)
+
+#define SDHCI_AT91_PRESET_COMMON_CONF	0x400 /* drv type B, programmable clock mode */
+
+struct sdhci_at91_soc_data {
+	const struct sdhci_pltfm_data *pdata;
+	bool baseclk_is_generated_internally;
+	unsigned int divider_for_baseclk;
+};
+
+struct sdhci_at91_priv {
+	const struct sdhci_at91_soc_data *soc_data;
+	struct clk *hclock;
+	struct clk *gck;
+	struct clk *mainck;
+	bool restore_needed;
+	bool cal_always_on;
+};
+
+static void sdhci_at91_set_force_card_detect(struct sdhci_host *host)
+{
+	u8 mc1r;
+
+	mc1r = readb(host->ioaddr + SDMMC_MC1R);
+	mc1r |= SDMMC_MC1R_FCD;
+	writeb(mc1r, host->ioaddr + SDMMC_MC1R);
+}
+
+static void sdhci_at91_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	u16 clk;
+
+	host->mmc->actual_clock = 0;
+
+	/*
+	 * There is no requirement to disable the internal clock before
+	 * changing the SD clock configuration. Moreover, disabling the
+	 * internal clock, changing the configuration and re-enabling the
+	 * internal clock causes some bugs. It can prevent to get the internal
+	 * clock stable flag ready and an unexpected switch to the base clock
+	 * when using presets.
+	 */
+	clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+	clk &= SDHCI_CLOCK_INT_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+	if (clock == 0)
+		return;
+
+	clk = sdhci_calc_clk(host, clock, &host->mmc->actual_clock);
+
+	clk |= SDHCI_CLOCK_INT_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+	/* Wait max 20 ms */
+	if (read_poll_timeout(sdhci_readw, clk, (clk & SDHCI_CLOCK_INT_STABLE),
+			      1000, 20000, false, host, SDHCI_CLOCK_CONTROL)) {
+		pr_err("%s: Internal clock never stabilised.\n",
+		       mmc_hostname(host->mmc));
+		return;
+	}
+
+	clk |= SDHCI_CLOCK_CARD_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+}
+
+static void sdhci_at91_set_uhs_signaling(struct sdhci_host *host,
+					 unsigned int timing)
+{
+	u8 mc1r;
+
+	if (timing == MMC_TIMING_MMC_DDR52) {
+		mc1r = sdhci_readb(host, SDMMC_MC1R);
+		mc1r |= SDMMC_MC1R_DDR;
+		sdhci_writeb(host, mc1r, SDMMC_MC1R);
+	}
+	sdhci_set_uhs_signaling(host, timing);
+}
+
+static void sdhci_at91_reset(struct sdhci_host *host, u8 mask)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_at91_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	unsigned int tmp;
+
+	sdhci_reset(host, mask);
+
+	if ((host->mmc->caps & MMC_CAP_NONREMOVABLE)
+	    || mmc_gpio_get_cd(host->mmc) >= 0)
+		sdhci_at91_set_force_card_detect(host);
+
+	if (priv->cal_always_on && (mask & SDHCI_RESET_ALL)) {
+		u32 calcr = sdhci_readl(host, SDMMC_CALCR);
+
+		sdhci_writel(host, calcr | SDMMC_CALCR_ALWYSON | SDMMC_CALCR_EN,
+			     SDMMC_CALCR);
+
+		if (read_poll_timeout(sdhci_readl, tmp, !(tmp & SDMMC_CALCR_EN),
+				      10, 20000, false, host, SDMMC_CALCR))
+			dev_err(mmc_dev(host->mmc), "Failed to calibrate\n");
+	}
+}
+
+static const struct sdhci_ops sdhci_at91_sama5d2_ops = {
+	.set_clock		= sdhci_at91_set_clock,
+	.set_bus_width		= sdhci_set_bus_width,
+	.reset			= sdhci_at91_reset,
+	.set_uhs_signaling	= sdhci_at91_set_uhs_signaling,
+	.set_power		= sdhci_set_power_and_bus_voltage,
+};
+
+static const struct sdhci_pltfm_data sdhci_sama5d2_pdata = {
+	.ops = &sdhci_at91_sama5d2_ops,
+};
+
+static const struct sdhci_at91_soc_data soc_data_sama5d2 = {
+	.pdata = &sdhci_sama5d2_pdata,
+	.baseclk_is_generated_internally = false,
+};
+
+static const struct sdhci_at91_soc_data soc_data_sam9x60 = {
+	.pdata = &sdhci_sama5d2_pdata,
+	.baseclk_is_generated_internally = true,
+	.divider_for_baseclk = 2,
+};
+
+static const struct of_device_id sdhci_at91_dt_match[] = {
+	{ .compatible = "atmel,sama5d2-sdhci", .data = &soc_data_sama5d2 },
+	{ .compatible = "microchip,sam9x60-sdhci", .data = &soc_data_sam9x60 },
+	{}
+};
+MODULE_DEVICE_TABLE(of, sdhci_at91_dt_match);
+
+static int sdhci_at91_set_clks_presets(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_at91_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	unsigned int			caps0, caps1;
+	unsigned int			clk_base, clk_mul;
+	unsigned int			gck_rate, clk_base_rate;
+	unsigned int			preset_div;
+
+	clk_prepare_enable(priv->hclock);
+	caps0 = readl(host->ioaddr + SDHCI_CAPABILITIES);
+	caps1 = readl(host->ioaddr + SDHCI_CAPABILITIES_1);
+
+	gck_rate = clk_get_rate(priv->gck);
+	if (priv->soc_data->baseclk_is_generated_internally)
+		clk_base_rate = gck_rate / priv->soc_data->divider_for_baseclk;
+	else
+		clk_base_rate = clk_get_rate(priv->mainck);
+
+	clk_base = clk_base_rate / 1000000;
+	clk_mul = gck_rate / clk_base_rate - 1;
+
+	caps0 &= ~SDHCI_CLOCK_V3_BASE_MASK;
+	caps0 |= FIELD_PREP(SDHCI_CLOCK_V3_BASE_MASK, clk_base);
+	caps1 &= ~SDHCI_CLOCK_MUL_MASK;
+	caps1 |= FIELD_PREP(SDHCI_CLOCK_MUL_MASK, clk_mul);
+	/* Set capabilities in r/w mode. */
+	writel(SDMMC_CACR_KEY | SDMMC_CACR_CAPWREN, host->ioaddr + SDMMC_CACR);
+	writel(caps0, host->ioaddr + SDHCI_CAPABILITIES);
+	writel(caps1, host->ioaddr + SDHCI_CAPABILITIES_1);
+	/* Set capabilities in ro mode. */
+	writel(0, host->ioaddr + SDMMC_CACR);
+
+	dev_dbg(dev, "update clk mul to %u as gck rate is %u Hz and clk base is %u Hz\n",
+		clk_mul, gck_rate, clk_base_rate);
+
+	/*
+	 * We have to set preset values because it depends on the clk_mul
+	 * value. Moreover, SDR104 is supported in a degraded mode since the
+	 * maximum sd clock value is 120 MHz instead of 208 MHz. For that
+	 * reason, we need to use presets to support SDR104.
+	 */
+	preset_div = DIV_ROUND_UP(gck_rate, 24000000) - 1;
+	writew(SDHCI_AT91_PRESET_COMMON_CONF | preset_div,
+	       host->ioaddr + SDHCI_PRESET_FOR_SDR12);
+	preset_div = DIV_ROUND_UP(gck_rate, 50000000) - 1;
+	writew(SDHCI_AT91_PRESET_COMMON_CONF | preset_div,
+	       host->ioaddr + SDHCI_PRESET_FOR_SDR25);
+	preset_div = DIV_ROUND_UP(gck_rate, 100000000) - 1;
+	writew(SDHCI_AT91_PRESET_COMMON_CONF | preset_div,
+	       host->ioaddr + SDHCI_PRESET_FOR_SDR50);
+	preset_div = DIV_ROUND_UP(gck_rate, 120000000) - 1;
+	writew(SDHCI_AT91_PRESET_COMMON_CONF | preset_div,
+	       host->ioaddr + SDHCI_PRESET_FOR_SDR104);
+	preset_div = DIV_ROUND_UP(gck_rate, 50000000) - 1;
+	writew(SDHCI_AT91_PRESET_COMMON_CONF | preset_div,
+	       host->ioaddr + SDHCI_PRESET_FOR_DDR50);
+
+	clk_prepare_enable(priv->mainck);
+	clk_prepare_enable(priv->gck);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int sdhci_at91_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_at91_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	ret = pm_runtime_force_suspend(dev);
+
+	priv->restore_needed = true;
+
+	return ret;
+}
+#endif /* CONFIG_PM_SLEEP */
+
+#ifdef CONFIG_PM
+static int sdhci_at91_runtime_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_at91_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	ret = sdhci_runtime_suspend_host(host);
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	clk_disable_unprepare(priv->gck);
+	clk_disable_unprepare(priv->hclock);
+	clk_disable_unprepare(priv->mainck);
+
+	return ret;
+}
+
+static int sdhci_at91_runtime_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_at91_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	if (priv->restore_needed) {
+		ret = sdhci_at91_set_clks_presets(dev);
+		if (ret)
+			return ret;
+
+		priv->restore_needed = false;
+		goto out;
+	}
+
+	ret = clk_prepare_enable(priv->mainck);
+	if (ret) {
+		dev_err(dev, "can't enable mainck\n");
+		return ret;
+	}
+
+	ret = clk_prepare_enable(priv->hclock);
+	if (ret) {
+		dev_err(dev, "can't enable hclock\n");
+		return ret;
+	}
+
+	ret = clk_prepare_enable(priv->gck);
+	if (ret) {
+		dev_err(dev, "can't enable gck\n");
+		return ret;
+	}
+
+out:
+	return sdhci_runtime_resume_host(host, 0);
+}
+#endif /* CONFIG_PM */
+
+static const struct dev_pm_ops sdhci_at91_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(sdhci_at91_suspend, pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(sdhci_at91_runtime_suspend,
+			   sdhci_at91_runtime_resume,
+			   NULL)
+};
+
+static int sdhci_at91_probe(struct platform_device *pdev)
+{
+	const struct of_device_id	*match;
+	const struct sdhci_at91_soc_data	*soc_data;
+	struct sdhci_host		*host;
+	struct sdhci_pltfm_host		*pltfm_host;
+	struct sdhci_at91_priv		*priv;
+	int				ret;
+
+	match = of_match_device(sdhci_at91_dt_match, &pdev->dev);
+	if (!match)
+		return -EINVAL;
+	soc_data = match->data;
+
+	host = sdhci_pltfm_init(pdev, soc_data->pdata, sizeof(*priv));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	priv = sdhci_pltfm_priv(pltfm_host);
+	priv->soc_data = soc_data;
+
+	priv->mainck = devm_clk_get(&pdev->dev, "baseclk");
+	if (IS_ERR(priv->mainck)) {
+		if (soc_data->baseclk_is_generated_internally) {
+			priv->mainck = NULL;
+		} else {
+			dev_err(&pdev->dev, "failed to get baseclk\n");
+			ret = PTR_ERR(priv->mainck);
+			goto sdhci_pltfm_free;
+		}
+	}
+
+	priv->hclock = devm_clk_get(&pdev->dev, "hclock");
+	if (IS_ERR(priv->hclock)) {
+		dev_err(&pdev->dev, "failed to get hclock\n");
+		ret = PTR_ERR(priv->hclock);
+		goto sdhci_pltfm_free;
+	}
+
+	priv->gck = devm_clk_get(&pdev->dev, "multclk");
+	if (IS_ERR(priv->gck)) {
+		dev_err(&pdev->dev, "failed to get multclk\n");
+		ret = PTR_ERR(priv->gck);
+		goto sdhci_pltfm_free;
+	}
+
+	ret = sdhci_at91_set_clks_presets(&pdev->dev);
+	if (ret)
+		goto sdhci_pltfm_free;
+
+	priv->restore_needed = false;
+
+	/*
+	 * if SDCAL pin is wrongly connected, we must enable
+	 * the analog calibration cell permanently.
+	 */
+	priv->cal_always_on =
+		device_property_read_bool(&pdev->dev,
+					  "microchip,sdcal-inverted");
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto clocks_disable_unprepare;
+
+	sdhci_get_of_property(pdev);
+
+	pm_runtime_get_noresume(&pdev->dev);
+	pm_runtime_set_active(&pdev->dev);
+	pm_runtime_enable(&pdev->dev);
+	pm_runtime_set_autosuspend_delay(&pdev->dev, 50);
+	pm_runtime_use_autosuspend(&pdev->dev);
+
+	/* HS200 is broken at this moment */
+	host->quirks2 |= SDHCI_QUIRK2_BROKEN_HS200;
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto pm_runtime_disable;
+
+	/*
+	 * When calling sdhci_runtime_suspend_host(), the sdhci layer makes
+	 * the assumption that all the clocks of the controller are disabled.
+	 * It means we can't get irq from it when it is runtime suspended.
+	 * For that reason, it is not planned to wake-up on a card detect irq
+	 * from the controller.
+	 * If we want to use runtime PM and to be able to wake-up on card
+	 * insertion, we have to use a GPIO for the card detection or we can
+	 * use polling. Be aware that using polling will resume/suspend the
+	 * controller between each attempt.
+	 * Disable SDHCI_QUIRK_BROKEN_CARD_DETECTION to be sure nobody tries
+	 * to enable polling via device tree with broken-cd property.
+	 */
+	if (mmc_card_is_removable(host->mmc) &&
+	    mmc_gpio_get_cd(host->mmc) < 0) {
+		host->mmc->caps |= MMC_CAP_NEEDS_POLL;
+		host->quirks &= ~SDHCI_QUIRK_BROKEN_CARD_DETECTION;
+	}
+
+	/*
+	 * If the device attached to the MMC bus is not removable, it is safer
+	 * to set the Force Card Detect bit. People often don't connect the
+	 * card detect signal and use this pin for another purpose. If the card
+	 * detect pin is not muxed to SDHCI controller, a default value is
+	 * used. This value can be different from a SoC revision to another
+	 * one. Problems come when this default value is not card present. To
+	 * avoid this case, if the device is non removable then the card
+	 * detection procedure using the SDMCC_CD signal is bypassed.
+	 * This bit is reset when a software reset for all command is performed
+	 * so we need to implement our own reset function to set back this bit.
+	 *
+	 * WA: SAMA5D2 doesn't drive CMD if using CD GPIO line.
+	 */
+	if ((host->mmc->caps & MMC_CAP_NONREMOVABLE)
+	    || mmc_gpio_get_cd(host->mmc) >= 0)
+		sdhci_at91_set_force_card_detect(host);
+
+	pm_runtime_put_autosuspend(&pdev->dev);
+
+	return 0;
+
+pm_runtime_disable:
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_set_suspended(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+clocks_disable_unprepare:
+	clk_disable_unprepare(priv->gck);
+	clk_disable_unprepare(priv->mainck);
+	clk_disable_unprepare(priv->hclock);
+sdhci_pltfm_free:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static int sdhci_at91_remove(struct platform_device *pdev)
+{
+	struct sdhci_host	*host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host	*pltfm_host = sdhci_priv(host);
+	struct sdhci_at91_priv	*priv = sdhci_pltfm_priv(pltfm_host);
+	struct clk *gck = priv->gck;
+	struct clk *hclock = priv->hclock;
+	struct clk *mainck = priv->mainck;
+
+	pm_runtime_get_sync(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+
+	sdhci_pltfm_unregister(pdev);
+
+	clk_disable_unprepare(gck);
+	clk_disable_unprepare(hclock);
+	clk_disable_unprepare(mainck);
+
+	return 0;
+}
+
+static struct platform_driver sdhci_at91_driver = {
+	.driver		= {
+		.name	= "sdhci-at91",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = sdhci_at91_dt_match,
+		.pm	= &sdhci_at91_dev_pm_ops,
+	},
+	.probe		= sdhci_at91_probe,
+	.remove		= sdhci_at91_remove,
+};
+
+module_platform_driver(sdhci_at91_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for at91");
+MODULE_AUTHOR("Ludovic Desroches <ludovic.desroches@atmel.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-of-dwcmshc.c b/drivers/mmc/host/sdhci-of-dwcmshc.c
new file mode 100644
index 000000000..59d8d96ce
--- /dev/null
+++ b/drivers/mmc/host/sdhci-of-dwcmshc.c
@@ -0,0 +1,256 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Driver for Synopsys DesignWare Cores Mobile Storage Host Controller
+ *
+ * Copyright (C) 2018 Synaptics Incorporated
+ *
+ * Author: Jisheng Zhang <jszhang@kernel.org>
+ */
+
+#include <linux/clk.h>
+#include <linux/dma-mapping.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/sizes.h>
+
+#include "sdhci-pltfm.h"
+
+#define SDHCI_DWCMSHC_ARG2_STUFF	GENMASK(31, 16)
+
+/* DWCMSHC specific Mode Select value */
+#define DWCMSHC_CTRL_HS400		0x7
+
+#define BOUNDARY_OK(addr, len) \
+	((addr | (SZ_128M - 1)) == ((addr + len - 1) | (SZ_128M - 1)))
+
+struct dwcmshc_priv {
+	struct clk	*bus_clk;
+};
+
+/*
+ * If DMA addr spans 128MB boundary, we split the DMA transfer into two
+ * so that each DMA transfer doesn't exceed the boundary.
+ */
+static void dwcmshc_adma_write_desc(struct sdhci_host *host, void **desc,
+				    dma_addr_t addr, int len, unsigned int cmd)
+{
+	int tmplen, offset;
+
+	if (likely(!len || BOUNDARY_OK(addr, len))) {
+		sdhci_adma_write_desc(host, desc, addr, len, cmd);
+		return;
+	}
+
+	offset = addr & (SZ_128M - 1);
+	tmplen = SZ_128M - offset;
+	sdhci_adma_write_desc(host, desc, addr, tmplen, cmd);
+
+	addr += tmplen;
+	len -= tmplen;
+	sdhci_adma_write_desc(host, desc, addr, len, cmd);
+}
+
+static void dwcmshc_check_auto_cmd23(struct mmc_host *mmc,
+				     struct mmc_request *mrq)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	/*
+	 * No matter V4 is enabled or not, ARGUMENT2 register is 32-bit
+	 * block count register which doesn't support stuff bits of
+	 * CMD23 argument on dwcmsch host controller.
+	 */
+	if (mrq->sbc && (mrq->sbc->arg & SDHCI_DWCMSHC_ARG2_STUFF))
+		host->flags &= ~SDHCI_AUTO_CMD23;
+	else
+		host->flags |= SDHCI_AUTO_CMD23;
+}
+
+static void dwcmshc_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	dwcmshc_check_auto_cmd23(mmc, mrq);
+
+	sdhci_request(mmc, mrq);
+}
+
+static void dwcmshc_set_uhs_signaling(struct sdhci_host *host,
+				      unsigned int timing)
+{
+	u16 ctrl_2;
+
+	ctrl_2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	/* Select Bus Speed Mode for host */
+	ctrl_2 &= ~SDHCI_CTRL_UHS_MASK;
+	if ((timing == MMC_TIMING_MMC_HS200) ||
+	    (timing == MMC_TIMING_UHS_SDR104))
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR104;
+	else if (timing == MMC_TIMING_UHS_SDR12)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR12;
+	else if ((timing == MMC_TIMING_UHS_SDR25) ||
+		 (timing == MMC_TIMING_MMC_HS))
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR25;
+	else if (timing == MMC_TIMING_UHS_SDR50)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR50;
+	else if ((timing == MMC_TIMING_UHS_DDR50) ||
+		 (timing == MMC_TIMING_MMC_DDR52))
+		ctrl_2 |= SDHCI_CTRL_UHS_DDR50;
+	else if (timing == MMC_TIMING_MMC_HS400)
+		ctrl_2 |= DWCMSHC_CTRL_HS400;
+	sdhci_writew(host, ctrl_2, SDHCI_HOST_CONTROL2);
+}
+
+static const struct sdhci_ops sdhci_dwcmshc_ops = {
+	.set_clock		= sdhci_set_clock,
+	.set_bus_width		= sdhci_set_bus_width,
+	.set_uhs_signaling	= dwcmshc_set_uhs_signaling,
+	.get_max_clock		= sdhci_pltfm_clk_get_max_clock,
+	.reset			= sdhci_reset,
+	.adma_write_desc	= dwcmshc_adma_write_desc,
+};
+
+static const struct sdhci_pltfm_data sdhci_dwcmshc_pdata = {
+	.ops = &sdhci_dwcmshc_ops,
+	.quirks = SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+};
+
+static int dwcmshc_probe(struct platform_device *pdev)
+{
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_host *host;
+	struct dwcmshc_priv *priv;
+	int err;
+	u32 extra;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_dwcmshc_pdata,
+				sizeof(struct dwcmshc_priv));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	/*
+	 * extra adma table cnt for cross 128M boundary handling.
+	 */
+	extra = DIV_ROUND_UP_ULL(dma_get_required_mask(&pdev->dev), SZ_128M);
+	if (extra > SDHCI_MAX_SEGS)
+		extra = SDHCI_MAX_SEGS;
+	host->adma_table_cnt += extra;
+
+	pltfm_host = sdhci_priv(host);
+	priv = sdhci_pltfm_priv(pltfm_host);
+
+	pltfm_host->clk = devm_clk_get(&pdev->dev, "core");
+	if (IS_ERR(pltfm_host->clk)) {
+		err = PTR_ERR(pltfm_host->clk);
+		dev_err(&pdev->dev, "failed to get core clk: %d\n", err);
+		goto free_pltfm;
+	}
+	err = clk_prepare_enable(pltfm_host->clk);
+	if (err)
+		goto free_pltfm;
+
+	priv->bus_clk = devm_clk_get(&pdev->dev, "bus");
+	if (!IS_ERR(priv->bus_clk))
+		clk_prepare_enable(priv->bus_clk);
+
+	err = mmc_of_parse(host->mmc);
+	if (err)
+		goto err_clk;
+
+	sdhci_get_of_property(pdev);
+
+	host->mmc_host_ops.request = dwcmshc_request;
+
+	err = sdhci_add_host(host);
+	if (err)
+		goto err_clk;
+
+	return 0;
+
+err_clk:
+	clk_disable_unprepare(pltfm_host->clk);
+	clk_disable_unprepare(priv->bus_clk);
+free_pltfm:
+	sdhci_pltfm_free(pdev);
+	return err;
+}
+
+static int dwcmshc_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct dwcmshc_priv *priv = sdhci_pltfm_priv(pltfm_host);
+
+	sdhci_remove_host(host, 0);
+
+	clk_disable_unprepare(pltfm_host->clk);
+	clk_disable_unprepare(priv->bus_clk);
+
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int dwcmshc_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct dwcmshc_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	ret = sdhci_suspend_host(host);
+	if (ret)
+		return ret;
+
+	clk_disable_unprepare(pltfm_host->clk);
+	if (!IS_ERR(priv->bus_clk))
+		clk_disable_unprepare(priv->bus_clk);
+
+	return ret;
+}
+
+static int dwcmshc_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct dwcmshc_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	ret = clk_prepare_enable(pltfm_host->clk);
+	if (ret)
+		return ret;
+
+	if (!IS_ERR(priv->bus_clk)) {
+		ret = clk_prepare_enable(priv->bus_clk);
+		if (ret)
+			return ret;
+	}
+
+	return sdhci_resume_host(host);
+}
+#endif
+
+static SIMPLE_DEV_PM_OPS(dwcmshc_pmops, dwcmshc_suspend, dwcmshc_resume);
+
+static const struct of_device_id sdhci_dwcmshc_dt_ids[] = {
+	{ .compatible = "snps,dwcmshc-sdhci" },
+	{}
+};
+MODULE_DEVICE_TABLE(of, sdhci_dwcmshc_dt_ids);
+
+static struct platform_driver sdhci_dwcmshc_driver = {
+	.driver	= {
+		.name	= "sdhci-dwcmshc",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = sdhci_dwcmshc_dt_ids,
+		.pm = &dwcmshc_pmops,
+	},
+	.probe	= dwcmshc_probe,
+	.remove	= dwcmshc_remove,
+};
+module_platform_driver(sdhci_dwcmshc_driver);
+
+MODULE_DESCRIPTION("SDHCI platform driver for Synopsys DWC MSHC");
+MODULE_AUTHOR("Jisheng Zhang <jszhang@kernel.org>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-of-esdhc.c b/drivers/mmc/host/sdhci-of-esdhc.c
new file mode 100644
index 000000000..5b853f651
--- /dev/null
+++ b/drivers/mmc/host/sdhci-of-esdhc.c
@@ -0,0 +1,1523 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Freescale eSDHC controller driver.
+ *
+ * Copyright (c) 2007, 2010, 2012 Freescale Semiconductor, Inc.
+ * Copyright (c) 2009 MontaVista Software, Inc.
+ * Copyright 2020 NXP
+ *
+ * Authors: Xiaobo Xie <X.Xie@freescale.com>
+ *	    Anton Vorontsov <avorontsov@ru.mvista.com>
+ */
+
+#include <linux/err.h>
+#include <linux/io.h>
+#include <linux/of.h>
+#include <linux/of_address.h>
+#include <linux/delay.h>
+#include <linux/module.h>
+#include <linux/sys_soc.h>
+#include <linux/clk.h>
+#include <linux/ktime.h>
+#include <linux/dma-mapping.h>
+#include <linux/iopoll.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include "sdhci-pltfm.h"
+#include "sdhci-esdhc.h"
+
+#define VENDOR_V_22	0x12
+#define VENDOR_V_23	0x13
+
+#define MMC_TIMING_NUM (MMC_TIMING_MMC_HS400 + 1)
+
+struct esdhc_clk_fixup {
+	const unsigned int sd_dflt_max_clk;
+	const unsigned int max_clk[MMC_TIMING_NUM];
+};
+
+static const struct esdhc_clk_fixup ls1021a_esdhc_clk = {
+	.sd_dflt_max_clk = 25000000,
+	.max_clk[MMC_TIMING_MMC_HS] = 46500000,
+	.max_clk[MMC_TIMING_SD_HS] = 46500000,
+};
+
+static const struct esdhc_clk_fixup ls1046a_esdhc_clk = {
+	.sd_dflt_max_clk = 25000000,
+	.max_clk[MMC_TIMING_UHS_SDR104] = 167000000,
+	.max_clk[MMC_TIMING_MMC_HS200] = 167000000,
+};
+
+static const struct esdhc_clk_fixup ls1012a_esdhc_clk = {
+	.sd_dflt_max_clk = 25000000,
+	.max_clk[MMC_TIMING_UHS_SDR104] = 125000000,
+	.max_clk[MMC_TIMING_MMC_HS200] = 125000000,
+};
+
+static const struct esdhc_clk_fixup p1010_esdhc_clk = {
+	.sd_dflt_max_clk = 20000000,
+	.max_clk[MMC_TIMING_LEGACY] = 20000000,
+	.max_clk[MMC_TIMING_MMC_HS] = 42000000,
+	.max_clk[MMC_TIMING_SD_HS] = 40000000,
+};
+
+static const struct of_device_id sdhci_esdhc_of_match[] = {
+	{ .compatible = "fsl,ls1021a-esdhc", .data = &ls1021a_esdhc_clk},
+	{ .compatible = "fsl,ls1046a-esdhc", .data = &ls1046a_esdhc_clk},
+	{ .compatible = "fsl,ls1012a-esdhc", .data = &ls1012a_esdhc_clk},
+	{ .compatible = "fsl,p1010-esdhc",   .data = &p1010_esdhc_clk},
+	{ .compatible = "fsl,mpc8379-esdhc" },
+	{ .compatible = "fsl,mpc8536-esdhc" },
+	{ .compatible = "fsl,esdhc" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, sdhci_esdhc_of_match);
+
+struct sdhci_esdhc {
+	u8 vendor_ver;
+	u8 spec_ver;
+	bool quirk_incorrect_hostver;
+	bool quirk_limited_clk_division;
+	bool quirk_unreliable_pulse_detection;
+	bool quirk_tuning_erratum_type1;
+	bool quirk_tuning_erratum_type2;
+	bool quirk_ignore_data_inhibit;
+	bool quirk_delay_before_data_reset;
+	bool quirk_trans_complete_erratum;
+	bool in_sw_tuning;
+	unsigned int peripheral_clock;
+	const struct esdhc_clk_fixup *clk_fixup;
+	u32 div_ratio;
+};
+
+/**
+ * esdhc_read*_fixup - Fixup the value read from incompatible eSDHC register
+ *		       to make it compatible with SD spec.
+ *
+ * @host: pointer to sdhci_host
+ * @spec_reg: SD spec register address
+ * @value: 32bit eSDHC register value on spec_reg address
+ *
+ * In SD spec, there are 8/16/32/64 bits registers, while all of eSDHC
+ * registers are 32 bits. There are differences in register size, register
+ * address, register function, bit position and function between eSDHC spec
+ * and SD spec.
+ *
+ * Return a fixed up register value
+ */
+static u32 esdhc_readl_fixup(struct sdhci_host *host,
+				     int spec_reg, u32 value)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	u32 ret;
+
+	/*
+	 * The bit of ADMA flag in eSDHC is not compatible with standard
+	 * SDHC register, so set fake flag SDHCI_CAN_DO_ADMA2 when ADMA is
+	 * supported by eSDHC.
+	 * And for many FSL eSDHC controller, the reset value of field
+	 * SDHCI_CAN_DO_ADMA1 is 1, but some of them can't support ADMA,
+	 * only these vendor version is greater than 2.2/0x12 support ADMA.
+	 */
+	if ((spec_reg == SDHCI_CAPABILITIES) && (value & SDHCI_CAN_DO_ADMA1)) {
+		if (esdhc->vendor_ver > VENDOR_V_22) {
+			ret = value | SDHCI_CAN_DO_ADMA2;
+			return ret;
+		}
+	}
+
+	/*
+	 * The DAT[3:0] line signal levels and the CMD line signal level are
+	 * not compatible with standard SDHC register. The line signal levels
+	 * DAT[7:0] are at bits 31:24 and the command line signal level is at
+	 * bit 23. All other bits are the same as in the standard SDHC
+	 * register.
+	 */
+	if (spec_reg == SDHCI_PRESENT_STATE) {
+		ret = value & 0x000fffff;
+		ret |= (value >> 4) & SDHCI_DATA_LVL_MASK;
+		ret |= (value << 1) & SDHCI_CMD_LVL;
+
+		/*
+		 * Some controllers have unreliable Data Line Active
+		 * bit for commands with busy signal. This affects
+		 * Command Inhibit (data) bit. Just ignore it since
+		 * MMC core driver has already polled card status
+		 * with CMD13 after any command with busy siganl.
+		 */
+		if (esdhc->quirk_ignore_data_inhibit)
+			ret &= ~SDHCI_DATA_INHIBIT;
+		return ret;
+	}
+
+	/*
+	 * DTS properties of mmc host are used to enable each speed mode
+	 * according to soc and board capability. So clean up
+	 * SDR50/SDR104/DDR50 support bits here.
+	 */
+	if (spec_reg == SDHCI_CAPABILITIES_1) {
+		ret = value & ~(SDHCI_SUPPORT_SDR50 | SDHCI_SUPPORT_SDR104 |
+				SDHCI_SUPPORT_DDR50);
+		return ret;
+	}
+
+	ret = value;
+	return ret;
+}
+
+static u16 esdhc_readw_fixup(struct sdhci_host *host,
+				     int spec_reg, u32 value)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	u16 ret;
+	int shift = (spec_reg & 0x2) * 8;
+
+	if (spec_reg == SDHCI_TRANSFER_MODE)
+		return pltfm_host->xfer_mode_shadow;
+
+	if (spec_reg == SDHCI_HOST_VERSION)
+		ret = value & 0xffff;
+	else
+		ret = (value >> shift) & 0xffff;
+	/* Workaround for T4240-R1.0-R2.0 eSDHC which has incorrect
+	 * vendor version and spec version information.
+	 */
+	if ((spec_reg == SDHCI_HOST_VERSION) &&
+	    (esdhc->quirk_incorrect_hostver))
+		ret = (VENDOR_V_23 << SDHCI_VENDOR_VER_SHIFT) | SDHCI_SPEC_200;
+	return ret;
+}
+
+static u8 esdhc_readb_fixup(struct sdhci_host *host,
+				     int spec_reg, u32 value)
+{
+	u8 ret;
+	u8 dma_bits;
+	int shift = (spec_reg & 0x3) * 8;
+
+	ret = (value >> shift) & 0xff;
+
+	/*
+	 * "DMA select" locates at offset 0x28 in SD specification, but on
+	 * P5020 or P3041, it locates at 0x29.
+	 */
+	if (spec_reg == SDHCI_HOST_CONTROL) {
+		/* DMA select is 22,23 bits in Protocol Control Register */
+		dma_bits = (value >> 5) & SDHCI_CTRL_DMA_MASK;
+		/* fixup the result */
+		ret &= ~SDHCI_CTRL_DMA_MASK;
+		ret |= dma_bits;
+	}
+	return ret;
+}
+
+/**
+ * esdhc_write*_fixup - Fixup the SD spec register value so that it could be
+ *			written into eSDHC register.
+ *
+ * @host: pointer to sdhci_host
+ * @spec_reg: SD spec register address
+ * @value: 8/16/32bit SD spec register value that would be written
+ * @old_value: 32bit eSDHC register value on spec_reg address
+ *
+ * In SD spec, there are 8/16/32/64 bits registers, while all of eSDHC
+ * registers are 32 bits. There are differences in register size, register
+ * address, register function, bit position and function between eSDHC spec
+ * and SD spec.
+ *
+ * Return a fixed up register value
+ */
+static u32 esdhc_writel_fixup(struct sdhci_host *host,
+				     int spec_reg, u32 value, u32 old_value)
+{
+	u32 ret;
+
+	/*
+	 * Enabling IRQSTATEN[BGESEN] is just to set IRQSTAT[BGE]
+	 * when SYSCTL[RSTD] is set for some special operations.
+	 * No any impact on other operation.
+	 */
+	if (spec_reg == SDHCI_INT_ENABLE)
+		ret = value | SDHCI_INT_BLK_GAP;
+	else
+		ret = value;
+
+	return ret;
+}
+
+static u32 esdhc_writew_fixup(struct sdhci_host *host,
+				     int spec_reg, u16 value, u32 old_value)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	int shift = (spec_reg & 0x2) * 8;
+	u32 ret;
+
+	switch (spec_reg) {
+	case SDHCI_TRANSFER_MODE:
+		/*
+		 * Postpone this write, we must do it together with a
+		 * command write that is down below. Return old value.
+		 */
+		pltfm_host->xfer_mode_shadow = value;
+		return old_value;
+	case SDHCI_COMMAND:
+		ret = (value << 16) | pltfm_host->xfer_mode_shadow;
+		return ret;
+	}
+
+	ret = old_value & (~(0xffff << shift));
+	ret |= (value << shift);
+
+	if (spec_reg == SDHCI_BLOCK_SIZE) {
+		/*
+		 * Two last DMA bits are reserved, and first one is used for
+		 * non-standard blksz of 4096 bytes that we don't support
+		 * yet. So clear the DMA boundary bits.
+		 */
+		ret &= (~SDHCI_MAKE_BLKSZ(0x7, 0));
+	}
+	return ret;
+}
+
+static u32 esdhc_writeb_fixup(struct sdhci_host *host,
+				     int spec_reg, u8 value, u32 old_value)
+{
+	u32 ret;
+	u32 dma_bits;
+	u8 tmp;
+	int shift = (spec_reg & 0x3) * 8;
+
+	/*
+	 * eSDHC doesn't have a standard power control register, so we do
+	 * nothing here to avoid incorrect operation.
+	 */
+	if (spec_reg == SDHCI_POWER_CONTROL)
+		return old_value;
+	/*
+	 * "DMA select" location is offset 0x28 in SD specification, but on
+	 * P5020 or P3041, it's located at 0x29.
+	 */
+	if (spec_reg == SDHCI_HOST_CONTROL) {
+		/*
+		 * If host control register is not standard, exit
+		 * this function
+		 */
+		if (host->quirks2 & SDHCI_QUIRK2_BROKEN_HOST_CONTROL)
+			return old_value;
+
+		/* DMA select is 22,23 bits in Protocol Control Register */
+		dma_bits = (value & SDHCI_CTRL_DMA_MASK) << 5;
+		ret = (old_value & (~(SDHCI_CTRL_DMA_MASK << 5))) | dma_bits;
+		tmp = (value & (~SDHCI_CTRL_DMA_MASK)) |
+		      (old_value & SDHCI_CTRL_DMA_MASK);
+		ret = (ret & (~0xff)) | tmp;
+
+		/* Prevent SDHCI core from writing reserved bits (e.g. HISPD) */
+		ret &= ~ESDHC_HOST_CONTROL_RES;
+		return ret;
+	}
+
+	ret = (old_value & (~(0xff << shift))) | (value << shift);
+	return ret;
+}
+
+static u32 esdhc_be_readl(struct sdhci_host *host, int reg)
+{
+	u32 ret;
+	u32 value;
+
+	if (reg == SDHCI_CAPABILITIES_1)
+		value = ioread32be(host->ioaddr + ESDHC_CAPABILITIES_1);
+	else
+		value = ioread32be(host->ioaddr + reg);
+
+	ret = esdhc_readl_fixup(host, reg, value);
+
+	return ret;
+}
+
+static u32 esdhc_le_readl(struct sdhci_host *host, int reg)
+{
+	u32 ret;
+	u32 value;
+
+	if (reg == SDHCI_CAPABILITIES_1)
+		value = ioread32(host->ioaddr + ESDHC_CAPABILITIES_1);
+	else
+		value = ioread32(host->ioaddr + reg);
+
+	ret = esdhc_readl_fixup(host, reg, value);
+
+	return ret;
+}
+
+static u16 esdhc_be_readw(struct sdhci_host *host, int reg)
+{
+	u16 ret;
+	u32 value;
+	int base = reg & ~0x3;
+
+	value = ioread32be(host->ioaddr + base);
+	ret = esdhc_readw_fixup(host, reg, value);
+	return ret;
+}
+
+static u16 esdhc_le_readw(struct sdhci_host *host, int reg)
+{
+	u16 ret;
+	u32 value;
+	int base = reg & ~0x3;
+
+	value = ioread32(host->ioaddr + base);
+	ret = esdhc_readw_fixup(host, reg, value);
+	return ret;
+}
+
+static u8 esdhc_be_readb(struct sdhci_host *host, int reg)
+{
+	u8 ret;
+	u32 value;
+	int base = reg & ~0x3;
+
+	value = ioread32be(host->ioaddr + base);
+	ret = esdhc_readb_fixup(host, reg, value);
+	return ret;
+}
+
+static u8 esdhc_le_readb(struct sdhci_host *host, int reg)
+{
+	u8 ret;
+	u32 value;
+	int base = reg & ~0x3;
+
+	value = ioread32(host->ioaddr + base);
+	ret = esdhc_readb_fixup(host, reg, value);
+	return ret;
+}
+
+static void esdhc_be_writel(struct sdhci_host *host, u32 val, int reg)
+{
+	u32 value;
+
+	value = esdhc_writel_fixup(host, reg, val, 0);
+	iowrite32be(value, host->ioaddr + reg);
+}
+
+static void esdhc_le_writel(struct sdhci_host *host, u32 val, int reg)
+{
+	u32 value;
+
+	value = esdhc_writel_fixup(host, reg, val, 0);
+	iowrite32(value, host->ioaddr + reg);
+}
+
+static void esdhc_be_writew(struct sdhci_host *host, u16 val, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	int base = reg & ~0x3;
+	u32 value;
+	u32 ret;
+
+	value = ioread32be(host->ioaddr + base);
+	ret = esdhc_writew_fixup(host, reg, val, value);
+	if (reg != SDHCI_TRANSFER_MODE)
+		iowrite32be(ret, host->ioaddr + base);
+
+	/* Starting SW tuning requires ESDHC_SMPCLKSEL to be set
+	 * 1us later after ESDHC_EXTN is set.
+	 */
+	if (base == ESDHC_SYSTEM_CONTROL_2) {
+		if (!(value & ESDHC_EXTN) && (ret & ESDHC_EXTN) &&
+		    esdhc->in_sw_tuning) {
+			udelay(1);
+			ret |= ESDHC_SMPCLKSEL;
+			iowrite32be(ret, host->ioaddr + base);
+		}
+	}
+}
+
+static void esdhc_le_writew(struct sdhci_host *host, u16 val, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	int base = reg & ~0x3;
+	u32 value;
+	u32 ret;
+
+	value = ioread32(host->ioaddr + base);
+	ret = esdhc_writew_fixup(host, reg, val, value);
+	if (reg != SDHCI_TRANSFER_MODE)
+		iowrite32(ret, host->ioaddr + base);
+
+	/* Starting SW tuning requires ESDHC_SMPCLKSEL to be set
+	 * 1us later after ESDHC_EXTN is set.
+	 */
+	if (base == ESDHC_SYSTEM_CONTROL_2) {
+		if (!(value & ESDHC_EXTN) && (ret & ESDHC_EXTN) &&
+		    esdhc->in_sw_tuning) {
+			udelay(1);
+			ret |= ESDHC_SMPCLKSEL;
+			iowrite32(ret, host->ioaddr + base);
+		}
+	}
+}
+
+static void esdhc_be_writeb(struct sdhci_host *host, u8 val, int reg)
+{
+	int base = reg & ~0x3;
+	u32 value;
+	u32 ret;
+
+	value = ioread32be(host->ioaddr + base);
+	ret = esdhc_writeb_fixup(host, reg, val, value);
+	iowrite32be(ret, host->ioaddr + base);
+}
+
+static void esdhc_le_writeb(struct sdhci_host *host, u8 val, int reg)
+{
+	int base = reg & ~0x3;
+	u32 value;
+	u32 ret;
+
+	value = ioread32(host->ioaddr + base);
+	ret = esdhc_writeb_fixup(host, reg, val, value);
+	iowrite32(ret, host->ioaddr + base);
+}
+
+/*
+ * For Abort or Suspend after Stop at Block Gap, ignore the ADMA
+ * error(IRQSTAT[ADMAE]) if both Transfer Complete(IRQSTAT[TC])
+ * and Block Gap Event(IRQSTAT[BGE]) are also set.
+ * For Continue, apply soft reset for data(SYSCTL[RSTD]);
+ * and re-issue the entire read transaction from beginning.
+ */
+static void esdhc_of_adma_workaround(struct sdhci_host *host, u32 intmask)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	bool applicable;
+	dma_addr_t dmastart;
+	dma_addr_t dmanow;
+
+	applicable = (intmask & SDHCI_INT_DATA_END) &&
+		     (intmask & SDHCI_INT_BLK_GAP) &&
+		     (esdhc->vendor_ver == VENDOR_V_23);
+	if (!applicable)
+		return;
+
+	host->data->error = 0;
+	dmastart = sg_dma_address(host->data->sg);
+	dmanow = dmastart + host->data->bytes_xfered;
+	/*
+	 * Force update to the next DMA block boundary.
+	 */
+	dmanow = (dmanow & ~(SDHCI_DEFAULT_BOUNDARY_SIZE - 1)) +
+		SDHCI_DEFAULT_BOUNDARY_SIZE;
+	host->data->bytes_xfered = dmanow - dmastart;
+	sdhci_writel(host, dmanow, SDHCI_DMA_ADDRESS);
+}
+
+static int esdhc_of_enable_dma(struct sdhci_host *host)
+{
+	int ret;
+	u32 value;
+	struct device *dev = mmc_dev(host->mmc);
+
+	if (of_device_is_compatible(dev->of_node, "fsl,ls1043a-esdhc") ||
+	    of_device_is_compatible(dev->of_node, "fsl,ls1046a-esdhc")) {
+		ret = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(40));
+		if (ret)
+			return ret;
+	}
+
+	value = sdhci_readl(host, ESDHC_DMA_SYSCTL);
+
+	if (of_dma_is_coherent(dev->of_node))
+		value |= ESDHC_DMA_SNOOP;
+	else
+		value &= ~ESDHC_DMA_SNOOP;
+
+	sdhci_writel(host, value, ESDHC_DMA_SYSCTL);
+	return 0;
+}
+
+static unsigned int esdhc_of_get_max_clock(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+
+	if (esdhc->peripheral_clock)
+		return esdhc->peripheral_clock;
+	else
+		return pltfm_host->clock;
+}
+
+static unsigned int esdhc_of_get_min_clock(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	unsigned int clock;
+
+	if (esdhc->peripheral_clock)
+		clock = esdhc->peripheral_clock;
+	else
+		clock = pltfm_host->clock;
+	return clock / 256 / 16;
+}
+
+static void esdhc_clock_enable(struct sdhci_host *host, bool enable)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	ktime_t timeout;
+	u32 val, clk_en;
+
+	clk_en = ESDHC_CLOCK_SDCLKEN;
+
+	/*
+	 * IPGEN/HCKEN/PEREN bits exist on eSDHC whose vendor version
+	 * is 2.2 or lower.
+	 */
+	if (esdhc->vendor_ver <= VENDOR_V_22)
+		clk_en |= (ESDHC_CLOCK_IPGEN | ESDHC_CLOCK_HCKEN |
+			   ESDHC_CLOCK_PEREN);
+
+	val = sdhci_readl(host, ESDHC_SYSTEM_CONTROL);
+
+	if (enable)
+		val |= clk_en;
+	else
+		val &= ~clk_en;
+
+	sdhci_writel(host, val, ESDHC_SYSTEM_CONTROL);
+
+	/*
+	 * Wait max 20 ms. If vendor version is 2.2 or lower, do not
+	 * wait clock stable bit which does not exist.
+	 */
+	timeout = ktime_add_ms(ktime_get(), 20);
+	while (esdhc->vendor_ver > VENDOR_V_22) {
+		bool timedout = ktime_after(ktime_get(), timeout);
+
+		if (sdhci_readl(host, ESDHC_PRSSTAT) & ESDHC_CLOCK_STABLE)
+			break;
+		if (timedout) {
+			pr_err("%s: Internal clock never stabilised.\n",
+				mmc_hostname(host->mmc));
+			break;
+		}
+		usleep_range(10, 20);
+	}
+}
+
+static void esdhc_flush_async_fifo(struct sdhci_host *host)
+{
+	ktime_t timeout;
+	u32 val;
+
+	val = sdhci_readl(host, ESDHC_DMA_SYSCTL);
+	val |= ESDHC_FLUSH_ASYNC_FIFO;
+	sdhci_writel(host, val, ESDHC_DMA_SYSCTL);
+
+	/* Wait max 20 ms */
+	timeout = ktime_add_ms(ktime_get(), 20);
+	while (1) {
+		bool timedout = ktime_after(ktime_get(), timeout);
+
+		if (!(sdhci_readl(host, ESDHC_DMA_SYSCTL) &
+		      ESDHC_FLUSH_ASYNC_FIFO))
+			break;
+		if (timedout) {
+			pr_err("%s: flushing asynchronous FIFO timeout.\n",
+				mmc_hostname(host->mmc));
+			break;
+		}
+		usleep_range(10, 20);
+	}
+}
+
+static void esdhc_of_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	unsigned int pre_div = 1, div = 1;
+	unsigned int clock_fixup = 0;
+	ktime_t timeout;
+	u32 temp;
+
+	if (clock == 0) {
+		host->mmc->actual_clock = 0;
+		esdhc_clock_enable(host, false);
+		return;
+	}
+
+	/* Start pre_div at 2 for vendor version < 2.3. */
+	if (esdhc->vendor_ver < VENDOR_V_23)
+		pre_div = 2;
+
+	/* Fix clock value. */
+	if (host->mmc->card && mmc_card_sd(host->mmc->card) &&
+	    esdhc->clk_fixup && host->mmc->ios.timing == MMC_TIMING_LEGACY)
+		clock_fixup = esdhc->clk_fixup->sd_dflt_max_clk;
+	else if (esdhc->clk_fixup)
+		clock_fixup = esdhc->clk_fixup->max_clk[host->mmc->ios.timing];
+
+	if (clock_fixup == 0 || clock < clock_fixup)
+		clock_fixup = clock;
+
+	/* Calculate pre_div and div. */
+	while (host->max_clk / pre_div / 16 > clock_fixup && pre_div < 256)
+		pre_div *= 2;
+
+	while (host->max_clk / pre_div / div > clock_fixup && div < 16)
+		div++;
+
+	esdhc->div_ratio = pre_div * div;
+
+	/* Limit clock division for HS400 200MHz clock for quirk. */
+	if (esdhc->quirk_limited_clk_division &&
+	    clock == MMC_HS200_MAX_DTR &&
+	    (host->mmc->ios.timing == MMC_TIMING_MMC_HS400 ||
+	     host->flags & SDHCI_HS400_TUNING)) {
+		if (esdhc->div_ratio <= 4) {
+			pre_div = 4;
+			div = 1;
+		} else if (esdhc->div_ratio <= 8) {
+			pre_div = 4;
+			div = 2;
+		} else if (esdhc->div_ratio <= 12) {
+			pre_div = 4;
+			div = 3;
+		} else {
+			pr_warn("%s: using unsupported clock division.\n",
+				mmc_hostname(host->mmc));
+		}
+		esdhc->div_ratio = pre_div * div;
+	}
+
+	host->mmc->actual_clock = host->max_clk / esdhc->div_ratio;
+
+	dev_dbg(mmc_dev(host->mmc), "desired SD clock: %d, actual: %d\n",
+		clock, host->mmc->actual_clock);
+
+	/* Set clock division into register. */
+	pre_div >>= 1;
+	div--;
+
+	esdhc_clock_enable(host, false);
+
+	temp = sdhci_readl(host, ESDHC_SYSTEM_CONTROL);
+	temp &= ~ESDHC_CLOCK_MASK;
+	temp |= ((div << ESDHC_DIVIDER_SHIFT) |
+		(pre_div << ESDHC_PREDIV_SHIFT));
+	sdhci_writel(host, temp, ESDHC_SYSTEM_CONTROL);
+
+	/*
+	 * Wait max 20 ms. If vendor version is 2.2 or lower, do not
+	 * wait clock stable bit which does not exist.
+	 */
+	timeout = ktime_add_ms(ktime_get(), 20);
+	while (esdhc->vendor_ver > VENDOR_V_22) {
+		bool timedout = ktime_after(ktime_get(), timeout);
+
+		if (sdhci_readl(host, ESDHC_PRSSTAT) & ESDHC_CLOCK_STABLE)
+			break;
+		if (timedout) {
+			pr_err("%s: Internal clock never stabilised.\n",
+				mmc_hostname(host->mmc));
+			break;
+		}
+		usleep_range(10, 20);
+	}
+
+	/* Additional setting for HS400. */
+	if (host->mmc->ios.timing == MMC_TIMING_MMC_HS400 &&
+	    clock == MMC_HS200_MAX_DTR) {
+		temp = sdhci_readl(host, ESDHC_TBCTL);
+		sdhci_writel(host, temp | ESDHC_HS400_MODE, ESDHC_TBCTL);
+		temp = sdhci_readl(host, ESDHC_SDCLKCTL);
+		sdhci_writel(host, temp | ESDHC_CMD_CLK_CTL, ESDHC_SDCLKCTL);
+		esdhc_clock_enable(host, true);
+
+		temp = sdhci_readl(host, ESDHC_DLLCFG0);
+		temp |= ESDHC_DLL_ENABLE;
+		if (host->mmc->actual_clock == MMC_HS200_MAX_DTR)
+			temp |= ESDHC_DLL_FREQ_SEL;
+		sdhci_writel(host, temp, ESDHC_DLLCFG0);
+
+		temp |= ESDHC_DLL_RESET;
+		sdhci_writel(host, temp, ESDHC_DLLCFG0);
+		udelay(1);
+		temp &= ~ESDHC_DLL_RESET;
+		sdhci_writel(host, temp, ESDHC_DLLCFG0);
+
+		/* Wait max 20 ms */
+		if (read_poll_timeout(sdhci_readl, temp,
+				      temp & ESDHC_DLL_STS_SLV_LOCK,
+				      10, 20000, false,
+				      host, ESDHC_DLLSTAT0))
+			pr_err("%s: timeout for delay chain lock.\n",
+			       mmc_hostname(host->mmc));
+
+		temp = sdhci_readl(host, ESDHC_TBCTL);
+		sdhci_writel(host, temp | ESDHC_HS400_WNDW_ADJUST, ESDHC_TBCTL);
+
+		esdhc_clock_enable(host, false);
+		esdhc_flush_async_fifo(host);
+	}
+	esdhc_clock_enable(host, true);
+}
+
+static void esdhc_pltfm_set_bus_width(struct sdhci_host *host, int width)
+{
+	u32 ctrl;
+
+	ctrl = sdhci_readl(host, ESDHC_PROCTL);
+	ctrl &= (~ESDHC_CTRL_BUSWIDTH_MASK);
+	switch (width) {
+	case MMC_BUS_WIDTH_8:
+		ctrl |= ESDHC_CTRL_8BITBUS;
+		break;
+
+	case MMC_BUS_WIDTH_4:
+		ctrl |= ESDHC_CTRL_4BITBUS;
+		break;
+
+	default:
+		break;
+	}
+
+	sdhci_writel(host, ctrl, ESDHC_PROCTL);
+}
+
+static void esdhc_reset(struct sdhci_host *host, u8 mask)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	u32 val, bus_width = 0;
+
+	/*
+	 * Add delay to make sure all the DMA transfers are finished
+	 * for quirk.
+	 */
+	if (esdhc->quirk_delay_before_data_reset &&
+	    (mask & SDHCI_RESET_DATA) &&
+	    (host->flags & SDHCI_REQ_USE_DMA))
+		mdelay(5);
+
+	/*
+	 * Save bus-width for eSDHC whose vendor version is 2.2
+	 * or lower for data reset.
+	 */
+	if ((mask & SDHCI_RESET_DATA) &&
+	    (esdhc->vendor_ver <= VENDOR_V_22)) {
+		val = sdhci_readl(host, ESDHC_PROCTL);
+		bus_width = val & ESDHC_CTRL_BUSWIDTH_MASK;
+	}
+
+	sdhci_reset(host, mask);
+
+	/*
+	 * Restore bus-width setting and interrupt registers for eSDHC
+	 * whose vendor version is 2.2 or lower for data reset.
+	 */
+	if ((mask & SDHCI_RESET_DATA) &&
+	    (esdhc->vendor_ver <= VENDOR_V_22)) {
+		val = sdhci_readl(host, ESDHC_PROCTL);
+		val &= ~ESDHC_CTRL_BUSWIDTH_MASK;
+		val |= bus_width;
+		sdhci_writel(host, val, ESDHC_PROCTL);
+
+		sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+		sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+	}
+
+	/*
+	 * Some bits have to be cleaned manually for eSDHC whose spec
+	 * version is higher than 3.0 for all reset.
+	 */
+	if ((mask & SDHCI_RESET_ALL) &&
+	    (esdhc->spec_ver >= SDHCI_SPEC_300)) {
+		val = sdhci_readl(host, ESDHC_TBCTL);
+		val &= ~ESDHC_TB_EN;
+		sdhci_writel(host, val, ESDHC_TBCTL);
+
+		/*
+		 * Initialize eSDHC_DLLCFG1[DLL_PD_PULSE_STRETCH_SEL] to
+		 * 0 for quirk.
+		 */
+		if (esdhc->quirk_unreliable_pulse_detection) {
+			val = sdhci_readl(host, ESDHC_DLLCFG1);
+			val &= ~ESDHC_DLL_PD_PULSE_STRETCH_SEL;
+			sdhci_writel(host, val, ESDHC_DLLCFG1);
+		}
+	}
+}
+
+/* The SCFG, Supplemental Configuration Unit, provides SoC specific
+ * configuration and status registers for the device. There is a
+ * SDHC IO VSEL control register on SCFG for some platforms. It's
+ * used to support SDHC IO voltage switching.
+ */
+static const struct of_device_id scfg_device_ids[] = {
+	{ .compatible = "fsl,t1040-scfg", },
+	{ .compatible = "fsl,ls1012a-scfg", },
+	{ .compatible = "fsl,ls1046a-scfg", },
+	{}
+};
+
+/* SDHC IO VSEL control register definition */
+#define SCFG_SDHCIOVSELCR	0x408
+#define SDHCIOVSELCR_TGLEN	0x80000000
+#define SDHCIOVSELCR_VSELVAL	0x60000000
+#define SDHCIOVSELCR_SDHC_VS	0x00000001
+
+static int esdhc_signal_voltage_switch(struct mmc_host *mmc,
+				       struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct device_node *scfg_node;
+	void __iomem *scfg_base = NULL;
+	u32 sdhciovselcr;
+	u32 val;
+
+	/*
+	 * Signal Voltage Switching is only applicable for Host Controllers
+	 * v3.00 and above.
+	 */
+	if (host->version < SDHCI_SPEC_300)
+		return 0;
+
+	val = sdhci_readl(host, ESDHC_PROCTL);
+
+	switch (ios->signal_voltage) {
+	case MMC_SIGNAL_VOLTAGE_330:
+		val &= ~ESDHC_VOLT_SEL;
+		sdhci_writel(host, val, ESDHC_PROCTL);
+		return 0;
+	case MMC_SIGNAL_VOLTAGE_180:
+		scfg_node = of_find_matching_node(NULL, scfg_device_ids);
+		if (scfg_node)
+			scfg_base = of_iomap(scfg_node, 0);
+		of_node_put(scfg_node);
+		if (scfg_base) {
+			sdhciovselcr = SDHCIOVSELCR_TGLEN |
+				       SDHCIOVSELCR_VSELVAL;
+			iowrite32be(sdhciovselcr,
+				scfg_base + SCFG_SDHCIOVSELCR);
+
+			val |= ESDHC_VOLT_SEL;
+			sdhci_writel(host, val, ESDHC_PROCTL);
+			mdelay(5);
+
+			sdhciovselcr = SDHCIOVSELCR_TGLEN |
+				       SDHCIOVSELCR_SDHC_VS;
+			iowrite32be(sdhciovselcr,
+				scfg_base + SCFG_SDHCIOVSELCR);
+			iounmap(scfg_base);
+		} else {
+			val |= ESDHC_VOLT_SEL;
+			sdhci_writel(host, val, ESDHC_PROCTL);
+		}
+		return 0;
+	default:
+		return 0;
+	}
+}
+
+static struct soc_device_attribute soc_tuning_erratum_type1[] = {
+	{ .family = "QorIQ T1023", },
+	{ .family = "QorIQ T1040", },
+	{ .family = "QorIQ T2080", },
+	{ .family = "QorIQ LS1021A", },
+	{ },
+};
+
+static struct soc_device_attribute soc_tuning_erratum_type2[] = {
+	{ .family = "QorIQ LS1012A", },
+	{ .family = "QorIQ LS1043A", },
+	{ .family = "QorIQ LS1046A", },
+	{ .family = "QorIQ LS1080A", },
+	{ .family = "QorIQ LS2080A", },
+	{ .family = "QorIQ LA1575A", },
+	{ },
+};
+
+static void esdhc_tuning_block_enable(struct sdhci_host *host, bool enable)
+{
+	u32 val;
+
+	esdhc_clock_enable(host, false);
+	esdhc_flush_async_fifo(host);
+
+	val = sdhci_readl(host, ESDHC_TBCTL);
+	if (enable)
+		val |= ESDHC_TB_EN;
+	else
+		val &= ~ESDHC_TB_EN;
+	sdhci_writel(host, val, ESDHC_TBCTL);
+
+	esdhc_clock_enable(host, true);
+}
+
+static void esdhc_tuning_window_ptr(struct sdhci_host *host, u8 *window_start,
+				    u8 *window_end)
+{
+	u32 val;
+
+	/* Write TBCTL[11:8]=4'h8 */
+	val = sdhci_readl(host, ESDHC_TBCTL);
+	val &= ~(0xf << 8);
+	val |= 8 << 8;
+	sdhci_writel(host, val, ESDHC_TBCTL);
+
+	mdelay(1);
+
+	/* Read TBCTL[31:0] register and rewrite again */
+	val = sdhci_readl(host, ESDHC_TBCTL);
+	sdhci_writel(host, val, ESDHC_TBCTL);
+
+	mdelay(1);
+
+	/* Read the TBSTAT[31:0] register twice */
+	val = sdhci_readl(host, ESDHC_TBSTAT);
+	val = sdhci_readl(host, ESDHC_TBSTAT);
+
+	*window_end = val & 0xff;
+	*window_start = (val >> 8) & 0xff;
+}
+
+static void esdhc_prepare_sw_tuning(struct sdhci_host *host, u8 *window_start,
+				    u8 *window_end)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	u8 start_ptr, end_ptr;
+
+	if (esdhc->quirk_tuning_erratum_type1) {
+		*window_start = 5 * esdhc->div_ratio;
+		*window_end = 3 * esdhc->div_ratio;
+		return;
+	}
+
+	esdhc_tuning_window_ptr(host, &start_ptr, &end_ptr);
+
+	/* Reset data lines by setting ESDHCCTL[RSTD] */
+	sdhci_reset(host, SDHCI_RESET_DATA);
+	/* Write 32'hFFFF_FFFF to IRQSTAT register */
+	sdhci_writel(host, 0xFFFFFFFF, SDHCI_INT_STATUS);
+
+	/* If TBSTAT[15:8]-TBSTAT[7:0] > (4 * div_ratio) + 2
+	 * or TBSTAT[7:0]-TBSTAT[15:8] > (4 * div_ratio) + 2,
+	 * then program TBPTR[TB_WNDW_END_PTR] = 4 * div_ratio
+	 * and program TBPTR[TB_WNDW_START_PTR] = 8 * div_ratio.
+	 */
+
+	if (abs(start_ptr - end_ptr) > (4 * esdhc->div_ratio + 2)) {
+		*window_start = 8 * esdhc->div_ratio;
+		*window_end = 4 * esdhc->div_ratio;
+	} else {
+		*window_start = 5 * esdhc->div_ratio;
+		*window_end = 3 * esdhc->div_ratio;
+	}
+}
+
+static int esdhc_execute_sw_tuning(struct mmc_host *mmc, u32 opcode,
+				   u8 window_start, u8 window_end)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	u32 val;
+	int ret;
+
+	/* Program TBPTR[TB_WNDW_END_PTR] and TBPTR[TB_WNDW_START_PTR] */
+	val = ((u32)window_start << ESDHC_WNDW_STRT_PTR_SHIFT) &
+	      ESDHC_WNDW_STRT_PTR_MASK;
+	val |= window_end & ESDHC_WNDW_END_PTR_MASK;
+	sdhci_writel(host, val, ESDHC_TBPTR);
+
+	/* Program the software tuning mode by setting TBCTL[TB_MODE]=2'h3 */
+	val = sdhci_readl(host, ESDHC_TBCTL);
+	val &= ~ESDHC_TB_MODE_MASK;
+	val |= ESDHC_TB_MODE_SW;
+	sdhci_writel(host, val, ESDHC_TBCTL);
+
+	esdhc->in_sw_tuning = true;
+	ret = sdhci_execute_tuning(mmc, opcode);
+	esdhc->in_sw_tuning = false;
+	return ret;
+}
+
+static int esdhc_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	u8 window_start, window_end;
+	int ret, retries = 1;
+	bool hs400_tuning;
+	unsigned int clk;
+	u32 val;
+
+	/* For tuning mode, the sd clock divisor value
+	 * must be larger than 3 according to reference manual.
+	 */
+	clk = esdhc->peripheral_clock / 3;
+	if (host->clock > clk)
+		esdhc_of_set_clock(host, clk);
+
+	esdhc_tuning_block_enable(host, true);
+
+	/*
+	 * The eSDHC controller takes the data timeout value into account
+	 * during tuning. If the SD card is too slow sending the response, the
+	 * timer will expire and a "Buffer Read Ready" interrupt without data
+	 * is triggered. This leads to tuning errors.
+	 *
+	 * Just set the timeout to the maximum value because the core will
+	 * already take care of it in sdhci_send_tuning().
+	 */
+	sdhci_writeb(host, 0xe, SDHCI_TIMEOUT_CONTROL);
+
+	hs400_tuning = host->flags & SDHCI_HS400_TUNING;
+
+	do {
+		if (esdhc->quirk_limited_clk_division &&
+		    hs400_tuning)
+			esdhc_of_set_clock(host, host->clock);
+
+		/* Do HW tuning */
+		val = sdhci_readl(host, ESDHC_TBCTL);
+		val &= ~ESDHC_TB_MODE_MASK;
+		val |= ESDHC_TB_MODE_3;
+		sdhci_writel(host, val, ESDHC_TBCTL);
+
+		ret = sdhci_execute_tuning(mmc, opcode);
+		if (ret)
+			break;
+
+		/* For type2 affected platforms of the tuning erratum,
+		 * tuning may succeed although eSDHC might not have
+		 * tuned properly. Need to check tuning window.
+		 */
+		if (esdhc->quirk_tuning_erratum_type2 &&
+		    !host->tuning_err) {
+			esdhc_tuning_window_ptr(host, &window_start,
+						&window_end);
+			if (abs(window_start - window_end) >
+			    (4 * esdhc->div_ratio + 2))
+				host->tuning_err = -EAGAIN;
+		}
+
+		/* If HW tuning fails and triggers erratum,
+		 * try workaround.
+		 */
+		ret = host->tuning_err;
+		if (ret == -EAGAIN &&
+		    (esdhc->quirk_tuning_erratum_type1 ||
+		     esdhc->quirk_tuning_erratum_type2)) {
+			/* Recover HS400 tuning flag */
+			if (hs400_tuning)
+				host->flags |= SDHCI_HS400_TUNING;
+			pr_info("%s: Hold on to use fixed sampling clock. Try SW tuning!\n",
+				mmc_hostname(mmc));
+			/* Do SW tuning */
+			esdhc_prepare_sw_tuning(host, &window_start,
+						&window_end);
+			ret = esdhc_execute_sw_tuning(mmc, opcode,
+						      window_start,
+						      window_end);
+			if (ret)
+				break;
+
+			/* Retry both HW/SW tuning with reduced clock. */
+			ret = host->tuning_err;
+			if (ret == -EAGAIN && retries) {
+				/* Recover HS400 tuning flag */
+				if (hs400_tuning)
+					host->flags |= SDHCI_HS400_TUNING;
+
+				clk = host->max_clk / (esdhc->div_ratio + 1);
+				esdhc_of_set_clock(host, clk);
+				pr_info("%s: Hold on to use fixed sampling clock. Try tuning with reduced clock!\n",
+					mmc_hostname(mmc));
+			} else {
+				break;
+			}
+		} else {
+			break;
+		}
+	} while (retries--);
+
+	if (ret) {
+		esdhc_tuning_block_enable(host, false);
+	} else if (hs400_tuning) {
+		val = sdhci_readl(host, ESDHC_SDTIMNGCTL);
+		val |= ESDHC_FLW_CTL_BG;
+		sdhci_writel(host, val, ESDHC_SDTIMNGCTL);
+	}
+
+	return ret;
+}
+
+static void esdhc_set_uhs_signaling(struct sdhci_host *host,
+				   unsigned int timing)
+{
+	u32 val;
+
+	/*
+	 * There are specific registers setting for HS400 mode.
+	 * Clean all of them if controller is in HS400 mode to
+	 * exit HS400 mode before re-setting any speed mode.
+	 */
+	val = sdhci_readl(host, ESDHC_TBCTL);
+	if (val & ESDHC_HS400_MODE) {
+		val = sdhci_readl(host, ESDHC_SDTIMNGCTL);
+		val &= ~ESDHC_FLW_CTL_BG;
+		sdhci_writel(host, val, ESDHC_SDTIMNGCTL);
+
+		val = sdhci_readl(host, ESDHC_SDCLKCTL);
+		val &= ~ESDHC_CMD_CLK_CTL;
+		sdhci_writel(host, val, ESDHC_SDCLKCTL);
+
+		esdhc_clock_enable(host, false);
+		val = sdhci_readl(host, ESDHC_TBCTL);
+		val &= ~ESDHC_HS400_MODE;
+		sdhci_writel(host, val, ESDHC_TBCTL);
+		esdhc_clock_enable(host, true);
+
+		val = sdhci_readl(host, ESDHC_DLLCFG0);
+		val &= ~(ESDHC_DLL_ENABLE | ESDHC_DLL_FREQ_SEL);
+		sdhci_writel(host, val, ESDHC_DLLCFG0);
+
+		val = sdhci_readl(host, ESDHC_TBCTL);
+		val &= ~ESDHC_HS400_WNDW_ADJUST;
+		sdhci_writel(host, val, ESDHC_TBCTL);
+
+		esdhc_tuning_block_enable(host, false);
+	}
+
+	if (timing == MMC_TIMING_MMC_HS400)
+		esdhc_tuning_block_enable(host, true);
+	else
+		sdhci_set_uhs_signaling(host, timing);
+}
+
+static u32 esdhc_irq(struct sdhci_host *host, u32 intmask)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_esdhc *esdhc = sdhci_pltfm_priv(pltfm_host);
+	u32 command;
+
+	if (esdhc->quirk_trans_complete_erratum) {
+		command = SDHCI_GET_CMD(sdhci_readw(host,
+					SDHCI_COMMAND));
+		if (command == MMC_WRITE_MULTIPLE_BLOCK &&
+				sdhci_readw(host, SDHCI_BLOCK_COUNT) &&
+				intmask & SDHCI_INT_DATA_END) {
+			intmask &= ~SDHCI_INT_DATA_END;
+			sdhci_writel(host, SDHCI_INT_DATA_END,
+					SDHCI_INT_STATUS);
+		}
+	}
+	return intmask;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static u32 esdhc_proctl;
+static int esdhc_of_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+
+	esdhc_proctl = sdhci_readl(host, SDHCI_HOST_CONTROL);
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	return sdhci_suspend_host(host);
+}
+
+static int esdhc_of_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	int ret = sdhci_resume_host(host);
+
+	if (ret == 0) {
+		/* Isn't this already done by sdhci_resume_host() ? --rmk */
+		esdhc_of_enable_dma(host);
+		sdhci_writel(host, esdhc_proctl, SDHCI_HOST_CONTROL);
+	}
+	return ret;
+}
+#endif
+
+static SIMPLE_DEV_PM_OPS(esdhc_of_dev_pm_ops,
+			esdhc_of_suspend,
+			esdhc_of_resume);
+
+static const struct sdhci_ops sdhci_esdhc_be_ops = {
+	.read_l = esdhc_be_readl,
+	.read_w = esdhc_be_readw,
+	.read_b = esdhc_be_readb,
+	.write_l = esdhc_be_writel,
+	.write_w = esdhc_be_writew,
+	.write_b = esdhc_be_writeb,
+	.set_clock = esdhc_of_set_clock,
+	.enable_dma = esdhc_of_enable_dma,
+	.get_max_clock = esdhc_of_get_max_clock,
+	.get_min_clock = esdhc_of_get_min_clock,
+	.adma_workaround = esdhc_of_adma_workaround,
+	.set_bus_width = esdhc_pltfm_set_bus_width,
+	.reset = esdhc_reset,
+	.set_uhs_signaling = esdhc_set_uhs_signaling,
+	.irq = esdhc_irq,
+};
+
+static const struct sdhci_ops sdhci_esdhc_le_ops = {
+	.read_l = esdhc_le_readl,
+	.read_w = esdhc_le_readw,
+	.read_b = esdhc_le_readb,
+	.write_l = esdhc_le_writel,
+	.write_w = esdhc_le_writew,
+	.write_b = esdhc_le_writeb,
+	.set_clock = esdhc_of_set_clock,
+	.enable_dma = esdhc_of_enable_dma,
+	.get_max_clock = esdhc_of_get_max_clock,
+	.get_min_clock = esdhc_of_get_min_clock,
+	.adma_workaround = esdhc_of_adma_workaround,
+	.set_bus_width = esdhc_pltfm_set_bus_width,
+	.reset = esdhc_reset,
+	.set_uhs_signaling = esdhc_set_uhs_signaling,
+	.irq = esdhc_irq,
+};
+
+static const struct sdhci_pltfm_data sdhci_esdhc_be_pdata = {
+	.quirks = ESDHC_DEFAULT_QUIRKS |
+#ifdef CONFIG_PPC
+		  SDHCI_QUIRK_BROKEN_CARD_DETECTION |
+#endif
+		  SDHCI_QUIRK_NO_CARD_NO_RESET |
+		  SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.ops = &sdhci_esdhc_be_ops,
+};
+
+static const struct sdhci_pltfm_data sdhci_esdhc_le_pdata = {
+	.quirks = ESDHC_DEFAULT_QUIRKS |
+		  SDHCI_QUIRK_NO_CARD_NO_RESET |
+		  SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.ops = &sdhci_esdhc_le_ops,
+};
+
+static struct soc_device_attribute soc_incorrect_hostver[] = {
+	{ .family = "QorIQ T4240", .revision = "1.0", },
+	{ .family = "QorIQ T4240", .revision = "2.0", },
+	{ },
+};
+
+static struct soc_device_attribute soc_fixup_sdhc_clkdivs[] = {
+	{ .family = "QorIQ LX2160A", .revision = "1.0", },
+	{ .family = "QorIQ LX2160A", .revision = "2.0", },
+	{ .family = "QorIQ LS1028A", .revision = "1.0", },
+	{ },
+};
+
+static struct soc_device_attribute soc_unreliable_pulse_detection[] = {
+	{ .family = "QorIQ LX2160A", .revision = "1.0", },
+	{ .family = "QorIQ LX2160A", .revision = "2.0", },
+	{ .family = "QorIQ LS1028A", .revision = "1.0", },
+	{ },
+};
+
+static void esdhc_init(struct platform_device *pdev, struct sdhci_host *host)
+{
+	const struct of_device_id *match;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_esdhc *esdhc;
+	struct device_node *np;
+	struct clk *clk;
+	u32 val;
+	u16 host_ver;
+
+	pltfm_host = sdhci_priv(host);
+	esdhc = sdhci_pltfm_priv(pltfm_host);
+
+	host_ver = sdhci_readw(host, SDHCI_HOST_VERSION);
+	esdhc->vendor_ver = (host_ver & SDHCI_VENDOR_VER_MASK) >>
+			     SDHCI_VENDOR_VER_SHIFT;
+	esdhc->spec_ver = host_ver & SDHCI_SPEC_VER_MASK;
+	if (soc_device_match(soc_incorrect_hostver))
+		esdhc->quirk_incorrect_hostver = true;
+	else
+		esdhc->quirk_incorrect_hostver = false;
+
+	if (soc_device_match(soc_fixup_sdhc_clkdivs))
+		esdhc->quirk_limited_clk_division = true;
+	else
+		esdhc->quirk_limited_clk_division = false;
+
+	if (soc_device_match(soc_unreliable_pulse_detection))
+		esdhc->quirk_unreliable_pulse_detection = true;
+	else
+		esdhc->quirk_unreliable_pulse_detection = false;
+
+	match = of_match_node(sdhci_esdhc_of_match, pdev->dev.of_node);
+	if (match)
+		esdhc->clk_fixup = match->data;
+	np = pdev->dev.of_node;
+
+	if (of_device_is_compatible(np, "fsl,p2020-esdhc")) {
+		esdhc->quirk_delay_before_data_reset = true;
+		esdhc->quirk_trans_complete_erratum = true;
+	}
+
+	clk = of_clk_get(np, 0);
+	if (!IS_ERR(clk)) {
+		/*
+		 * esdhc->peripheral_clock would be assigned with a value
+		 * which is eSDHC base clock when use periperal clock.
+		 * For some platforms, the clock value got by common clk
+		 * API is peripheral clock while the eSDHC base clock is
+		 * 1/2 peripheral clock.
+		 */
+		if (of_device_is_compatible(np, "fsl,ls1046a-esdhc") ||
+		    of_device_is_compatible(np, "fsl,ls1028a-esdhc") ||
+		    of_device_is_compatible(np, "fsl,ls1088a-esdhc"))
+			esdhc->peripheral_clock = clk_get_rate(clk) / 2;
+		else
+			esdhc->peripheral_clock = clk_get_rate(clk);
+
+		clk_put(clk);
+	}
+
+	esdhc_clock_enable(host, false);
+	val = sdhci_readl(host, ESDHC_DMA_SYSCTL);
+	/*
+	 * This bit is not able to be reset by SDHCI_RESET_ALL. Need to
+	 * initialize it as 1 or 0 once, to override the different value
+	 * which may be configured in bootloader.
+	 */
+	if (esdhc->peripheral_clock)
+		val |= ESDHC_PERIPHERAL_CLK_SEL;
+	else
+		val &= ~ESDHC_PERIPHERAL_CLK_SEL;
+	sdhci_writel(host, val, ESDHC_DMA_SYSCTL);
+	esdhc_clock_enable(host, true);
+}
+
+static int esdhc_hs400_prepare_ddr(struct mmc_host *mmc)
+{
+	esdhc_tuning_block_enable(mmc_priv(mmc), false);
+	return 0;
+}
+
+static int sdhci_esdhc_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct device_node *np;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_esdhc *esdhc;
+	int ret;
+
+	np = pdev->dev.of_node;
+
+	if (of_property_read_bool(np, "little-endian"))
+		host = sdhci_pltfm_init(pdev, &sdhci_esdhc_le_pdata,
+					sizeof(struct sdhci_esdhc));
+	else
+		host = sdhci_pltfm_init(pdev, &sdhci_esdhc_be_pdata,
+					sizeof(struct sdhci_esdhc));
+
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	host->mmc_host_ops.start_signal_voltage_switch =
+		esdhc_signal_voltage_switch;
+	host->mmc_host_ops.execute_tuning = esdhc_execute_tuning;
+	host->mmc_host_ops.hs400_prepare_ddr = esdhc_hs400_prepare_ddr;
+	host->tuning_delay = 1;
+
+	esdhc_init(pdev, host);
+
+	sdhci_get_of_property(pdev);
+
+	pltfm_host = sdhci_priv(host);
+	esdhc = sdhci_pltfm_priv(pltfm_host);
+	if (soc_device_match(soc_tuning_erratum_type1))
+		esdhc->quirk_tuning_erratum_type1 = true;
+	else
+		esdhc->quirk_tuning_erratum_type1 = false;
+
+	if (soc_device_match(soc_tuning_erratum_type2))
+		esdhc->quirk_tuning_erratum_type2 = true;
+	else
+		esdhc->quirk_tuning_erratum_type2 = false;
+
+	if (esdhc->vendor_ver == VENDOR_V_22)
+		host->quirks2 |= SDHCI_QUIRK2_HOST_NO_CMD23;
+
+	if (esdhc->vendor_ver > VENDOR_V_22)
+		host->quirks &= ~SDHCI_QUIRK_NO_BUSY_IRQ;
+
+	if (of_find_compatible_node(NULL, NULL, "fsl,p2020-esdhc")) {
+		host->quirks |= SDHCI_QUIRK_RESET_AFTER_REQUEST;
+		host->quirks |= SDHCI_QUIRK_BROKEN_TIMEOUT_VAL;
+	}
+
+	if (of_device_is_compatible(np, "fsl,p5040-esdhc") ||
+	    of_device_is_compatible(np, "fsl,p5020-esdhc") ||
+	    of_device_is_compatible(np, "fsl,p4080-esdhc") ||
+	    of_device_is_compatible(np, "fsl,p1020-esdhc") ||
+	    of_device_is_compatible(np, "fsl,t1040-esdhc"))
+		host->quirks &= ~SDHCI_QUIRK_BROKEN_CARD_DETECTION;
+
+	if (of_device_is_compatible(np, "fsl,ls1021a-esdhc"))
+		host->quirks |= SDHCI_QUIRK_BROKEN_TIMEOUT_VAL;
+
+	esdhc->quirk_ignore_data_inhibit = false;
+	if (of_device_is_compatible(np, "fsl,p2020-esdhc")) {
+		/*
+		 * Freescale messed up with P2020 as it has a non-standard
+		 * host control register
+		 */
+		host->quirks2 |= SDHCI_QUIRK2_BROKEN_HOST_CONTROL;
+		esdhc->quirk_ignore_data_inhibit = true;
+	}
+
+	/* call to generic mmc_of_parse to support additional capabilities */
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto err;
+
+	mmc_of_parse_voltage(np, &host->ocr_mask);
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err;
+
+	return 0;
+ err:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static struct platform_driver sdhci_esdhc_driver = {
+	.driver = {
+		.name = "sdhci-esdhc",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = sdhci_esdhc_of_match,
+		.pm = &esdhc_of_dev_pm_ops,
+	},
+	.probe = sdhci_esdhc_probe,
+	.remove = sdhci_pltfm_unregister,
+};
+
+module_platform_driver(sdhci_esdhc_driver);
+
+MODULE_DESCRIPTION("SDHCI OF driver for Freescale MPC eSDHC");
+MODULE_AUTHOR("Xiaobo Xie <X.Xie@freescale.com>, "
+	      "Anton Vorontsov <avorontsov@ru.mvista.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-of-hlwd.c b/drivers/mmc/host/sdhci-of-hlwd.c
new file mode 100644
index 000000000..12675797b
--- /dev/null
+++ b/drivers/mmc/host/sdhci-of-hlwd.c
@@ -0,0 +1,95 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * drivers/mmc/host/sdhci-of-hlwd.c
+ *
+ * Nintendo Wii Secure Digital Host Controller Interface.
+ * Copyright (C) 2009 The GameCube Linux Team
+ * Copyright (C) 2009 Albert Herranz
+ *
+ * Based on sdhci-of-esdhc.c
+ *
+ * Copyright (c) 2007 Freescale Semiconductor, Inc.
+ * Copyright (c) 2009 MontaVista Software, Inc.
+ *
+ * Authors: Xiaobo Xie <X.Xie@freescale.com>
+ *	    Anton Vorontsov <avorontsov@ru.mvista.com>
+ */
+
+#include <linux/delay.h>
+#include <linux/module.h>
+#include <linux/mmc/host.h>
+#include "sdhci-pltfm.h"
+
+/*
+ * Ops and quirks for the Nintendo Wii SDHCI controllers.
+ */
+
+/*
+ * We need a small delay after each write, or things go horribly wrong.
+ */
+#define SDHCI_HLWD_WRITE_DELAY	5 /* usecs */
+
+static void sdhci_hlwd_writel(struct sdhci_host *host, u32 val, int reg)
+{
+	sdhci_be32bs_writel(host, val, reg);
+	udelay(SDHCI_HLWD_WRITE_DELAY);
+}
+
+static void sdhci_hlwd_writew(struct sdhci_host *host, u16 val, int reg)
+{
+	sdhci_be32bs_writew(host, val, reg);
+	udelay(SDHCI_HLWD_WRITE_DELAY);
+}
+
+static void sdhci_hlwd_writeb(struct sdhci_host *host, u8 val, int reg)
+{
+	sdhci_be32bs_writeb(host, val, reg);
+	udelay(SDHCI_HLWD_WRITE_DELAY);
+}
+
+static const struct sdhci_ops sdhci_hlwd_ops = {
+	.read_l = sdhci_be32bs_readl,
+	.read_w = sdhci_be32bs_readw,
+	.read_b = sdhci_be32bs_readb,
+	.write_l = sdhci_hlwd_writel,
+	.write_w = sdhci_hlwd_writew,
+	.write_b = sdhci_hlwd_writeb,
+	.set_clock = sdhci_set_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static const struct sdhci_pltfm_data sdhci_hlwd_pdata = {
+	.quirks = SDHCI_QUIRK_32BIT_DMA_ADDR |
+		  SDHCI_QUIRK_32BIT_DMA_SIZE,
+	.ops = &sdhci_hlwd_ops,
+};
+
+static int sdhci_hlwd_probe(struct platform_device *pdev)
+{
+	return sdhci_pltfm_register(pdev, &sdhci_hlwd_pdata, 0);
+}
+
+static const struct of_device_id sdhci_hlwd_of_match[] = {
+	{ .compatible = "nintendo,hollywood-sdhci" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, sdhci_hlwd_of_match);
+
+static struct platform_driver sdhci_hlwd_driver = {
+	.driver = {
+		.name = "sdhci-hlwd",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = sdhci_hlwd_of_match,
+		.pm = &sdhci_pltfm_pmops,
+	},
+	.probe = sdhci_hlwd_probe,
+	.remove = sdhci_pltfm_unregister,
+};
+
+module_platform_driver(sdhci_hlwd_driver);
+
+MODULE_DESCRIPTION("Nintendo Wii SDHCI OF driver");
+MODULE_AUTHOR("The GameCube Linux Team, Albert Herranz");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-of-sparx5.c b/drivers/mmc/host/sdhci-of-sparx5.c
new file mode 100644
index 000000000..28e4ee69e
--- /dev/null
+++ b/drivers/mmc/host/sdhci-of-sparx5.c
@@ -0,0 +1,270 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * drivers/mmc/host/sdhci-of-sparx5.c
+ *
+ * MCHP Sparx5 SoC Secure Digital Host Controller Interface.
+ *
+ * Copyright (c) 2019 Microchip Inc.
+ *
+ * Author: Lars Povlsen <lars.povlsen@microchip.com>
+ */
+
+#include <linux/sizes.h>
+#include <linux/delay.h>
+#include <linux/module.h>
+#include <linux/regmap.h>
+#include <linux/of_device.h>
+#include <linux/mfd/syscon.h>
+#include <linux/dma-mapping.h>
+
+#include "sdhci-pltfm.h"
+
+#define CPU_REGS_GENERAL_CTRL	(0x22 * 4)
+#define  MSHC_DLY_CC_MASK	GENMASK(16, 13)
+#define  MSHC_DLY_CC_SHIFT	13
+#define  MSHC_DLY_CC_MAX	15
+
+#define CPU_REGS_PROC_CTRL	(0x2C * 4)
+#define  ACP_CACHE_FORCE_ENA	BIT(4)
+#define  ACP_AWCACHE		BIT(3)
+#define  ACP_ARCACHE		BIT(2)
+#define  ACP_CACHE_MASK		(ACP_CACHE_FORCE_ENA|ACP_AWCACHE|ACP_ARCACHE)
+
+#define MSHC2_VERSION			0x500	/* Off 0x140, reg 0x0 */
+#define MSHC2_TYPE			0x504	/* Off 0x140, reg 0x1 */
+#define MSHC2_EMMC_CTRL			0x52c	/* Off 0x140, reg 0xB */
+#define  MSHC2_EMMC_CTRL_EMMC_RST_N	BIT(2)
+#define  MSHC2_EMMC_CTRL_IS_EMMC	BIT(0)
+
+struct sdhci_sparx5_data {
+	struct sdhci_host *host;
+	struct regmap *cpu_ctrl;
+	int delay_clock;
+};
+
+#define BOUNDARY_OK(addr, len) \
+	((addr | (SZ_128M - 1)) == ((addr + len - 1) | (SZ_128M - 1)))
+
+/*
+ * If DMA addr spans 128MB boundary, we split the DMA transfer into two
+ * so that each DMA transfer doesn't exceed the boundary.
+ */
+static void sdhci_sparx5_adma_write_desc(struct sdhci_host *host, void **desc,
+					  dma_addr_t addr, int len,
+					  unsigned int cmd)
+{
+	int tmplen, offset;
+
+	if (likely(!len || BOUNDARY_OK(addr, len))) {
+		sdhci_adma_write_desc(host, desc, addr, len, cmd);
+		return;
+	}
+
+	pr_debug("%s: write_desc: splitting dma len %d, offset %pad\n",
+		 mmc_hostname(host->mmc), len, &addr);
+
+	offset = addr & (SZ_128M - 1);
+	tmplen = SZ_128M - offset;
+	sdhci_adma_write_desc(host, desc, addr, tmplen, cmd);
+
+	addr += tmplen;
+	len -= tmplen;
+	sdhci_adma_write_desc(host, desc, addr, len, cmd);
+}
+
+static void sparx5_set_cacheable(struct sdhci_host *host, u32 value)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_sparx5_data *sdhci_sparx5 = sdhci_pltfm_priv(pltfm_host);
+
+	pr_debug("%s: Set Cacheable = 0x%x\n", mmc_hostname(host->mmc), value);
+
+	/* Update ACP caching attributes in HW */
+	regmap_update_bits(sdhci_sparx5->cpu_ctrl,
+			   CPU_REGS_PROC_CTRL, ACP_CACHE_MASK, value);
+}
+
+static void sparx5_set_delay(struct sdhci_host *host, u8 value)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_sparx5_data *sdhci_sparx5 = sdhci_pltfm_priv(pltfm_host);
+
+	pr_debug("%s: Set DLY_CC = %u\n", mmc_hostname(host->mmc), value);
+
+	/* Update DLY_CC in HW */
+	regmap_update_bits(sdhci_sparx5->cpu_ctrl,
+			   CPU_REGS_GENERAL_CTRL,
+			   MSHC_DLY_CC_MASK,
+			   (value << MSHC_DLY_CC_SHIFT));
+}
+
+static void sdhci_sparx5_set_emmc(struct sdhci_host *host)
+{
+	if (!mmc_card_is_removable(host->mmc)) {
+		u8 value;
+
+		value = sdhci_readb(host, MSHC2_EMMC_CTRL);
+		if (!(value & MSHC2_EMMC_CTRL_IS_EMMC)) {
+			value |= MSHC2_EMMC_CTRL_IS_EMMC;
+			pr_debug("%s: Set EMMC_CTRL: 0x%08x\n",
+				 mmc_hostname(host->mmc), value);
+			sdhci_writeb(host, value, MSHC2_EMMC_CTRL);
+		}
+	}
+}
+
+static void sdhci_sparx5_reset_emmc(struct sdhci_host *host)
+{
+	u8 value;
+
+	pr_debug("%s: Toggle EMMC_CTRL.EMMC_RST_N\n", mmc_hostname(host->mmc));
+	value = sdhci_readb(host, MSHC2_EMMC_CTRL) &
+		~MSHC2_EMMC_CTRL_EMMC_RST_N;
+	sdhci_writeb(host, value, MSHC2_EMMC_CTRL);
+	/* For eMMC, minimum is 1us but give it 10us for good measure */
+	usleep_range(10, 20);
+	sdhci_writeb(host, value | MSHC2_EMMC_CTRL_EMMC_RST_N,
+		     MSHC2_EMMC_CTRL);
+	/* For eMMC, minimum is 200us but give it 300us for good measure */
+	usleep_range(300, 400);
+}
+
+static void sdhci_sparx5_reset(struct sdhci_host *host, u8 mask)
+{
+	pr_debug("%s: *** RESET: mask %d\n", mmc_hostname(host->mmc), mask);
+
+	sdhci_reset(host, mask);
+
+	/* Be sure CARD_IS_EMMC stays set */
+	sdhci_sparx5_set_emmc(host);
+}
+
+static const struct sdhci_ops sdhci_sparx5_ops = {
+	.set_clock		= sdhci_set_clock,
+	.set_bus_width		= sdhci_set_bus_width,
+	.set_uhs_signaling	= sdhci_set_uhs_signaling,
+	.get_max_clock		= sdhci_pltfm_clk_get_max_clock,
+	.reset			= sdhci_sparx5_reset,
+	.adma_write_desc	= sdhci_sparx5_adma_write_desc,
+};
+
+static const struct sdhci_pltfm_data sdhci_sparx5_pdata = {
+	.quirks  = 0,
+	.quirks2 = SDHCI_QUIRK2_HOST_NO_CMD23 | /* Controller issue */
+		   SDHCI_QUIRK2_NO_1_8_V, /* No sdr104, ddr50, etc */
+	.ops = &sdhci_sparx5_ops,
+};
+
+static int sdhci_sparx5_probe(struct platform_device *pdev)
+{
+	int ret;
+	const char *syscon = "microchip,sparx5-cpu-syscon";
+	struct sdhci_host *host;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_sparx5_data *sdhci_sparx5;
+	struct device_node *np = pdev->dev.of_node;
+	u32 value;
+	u32 extra;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_sparx5_pdata,
+				sizeof(*sdhci_sparx5));
+
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	/*
+	 * extra adma table cnt for cross 128M boundary handling.
+	 */
+	extra = DIV_ROUND_UP_ULL(dma_get_required_mask(&pdev->dev), SZ_128M);
+	if (extra > SDHCI_MAX_SEGS)
+		extra = SDHCI_MAX_SEGS;
+	host->adma_table_cnt += extra;
+
+	pltfm_host = sdhci_priv(host);
+	sdhci_sparx5 = sdhci_pltfm_priv(pltfm_host);
+	sdhci_sparx5->host = host;
+
+	pltfm_host->clk = devm_clk_get(&pdev->dev, "core");
+	if (IS_ERR(pltfm_host->clk)) {
+		ret = PTR_ERR(pltfm_host->clk);
+		dev_err(&pdev->dev, "failed to get core clk: %d\n", ret);
+		goto free_pltfm;
+	}
+	ret = clk_prepare_enable(pltfm_host->clk);
+	if (ret)
+		goto free_pltfm;
+
+	if (!of_property_read_u32(np, "microchip,clock-delay", &value) &&
+	    (value > 0 && value <= MSHC_DLY_CC_MAX))
+		sdhci_sparx5->delay_clock = value;
+
+	sdhci_get_of_property(pdev);
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto err_clk;
+
+	sdhci_sparx5->cpu_ctrl = syscon_regmap_lookup_by_compatible(syscon);
+	if (IS_ERR(sdhci_sparx5->cpu_ctrl)) {
+		dev_err(&pdev->dev, "No CPU syscon regmap !\n");
+		ret = PTR_ERR(sdhci_sparx5->cpu_ctrl);
+		goto err_clk;
+	}
+
+	if (sdhci_sparx5->delay_clock >= 0)
+		sparx5_set_delay(host, sdhci_sparx5->delay_clock);
+
+	if (!mmc_card_is_removable(host->mmc)) {
+		/* Do a HW reset of eMMC card */
+		sdhci_sparx5_reset_emmc(host);
+		/* Update EMMC_CTRL */
+		sdhci_sparx5_set_emmc(host);
+		/* If eMMC, disable SD and SDIO */
+		host->mmc->caps2 |= (MMC_CAP2_NO_SDIO|MMC_CAP2_NO_SD);
+	}
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err_clk;
+
+	/* Set AXI bus master to use un-cached access (for DMA) */
+	if (host->flags & (SDHCI_USE_SDMA | SDHCI_USE_ADMA) &&
+	    IS_ENABLED(CONFIG_DMA_DECLARE_COHERENT))
+		sparx5_set_cacheable(host, ACP_CACHE_FORCE_ENA);
+
+	pr_debug("%s: SDHC version: 0x%08x\n",
+		 mmc_hostname(host->mmc), sdhci_readl(host, MSHC2_VERSION));
+	pr_debug("%s: SDHC type:    0x%08x\n",
+		 mmc_hostname(host->mmc), sdhci_readl(host, MSHC2_TYPE));
+
+	return ret;
+
+err_clk:
+	clk_disable_unprepare(pltfm_host->clk);
+free_pltfm:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static const struct of_device_id sdhci_sparx5_of_match[] = {
+	{ .compatible = "microchip,dw-sparx5-sdhci" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, sdhci_sparx5_of_match);
+
+static struct platform_driver sdhci_sparx5_driver = {
+	.driver = {
+		.name = "sdhci-sparx5",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = sdhci_sparx5_of_match,
+		.pm = &sdhci_pltfm_pmops,
+	},
+	.probe = sdhci_sparx5_probe,
+	.remove = sdhci_pltfm_unregister,
+};
+
+module_platform_driver(sdhci_sparx5_driver);
+
+MODULE_DESCRIPTION("Sparx5 SDHCI OF driver");
+MODULE_AUTHOR("Lars Povlsen <lars.povlsen@microchip.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-omap.c b/drivers/mmc/host/sdhci-omap.c
new file mode 100644
index 000000000..53c362bb2
--- /dev/null
+++ b/drivers/mmc/host/sdhci-omap.c
@@ -0,0 +1,1325 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/**
+ * SDHCI Controller driver for TI's OMAP SoCs
+ *
+ * Copyright (C) 2017 Texas Instruments
+ * Author: Kishon Vijay Abraham I <kishon@ti.com>
+ */
+
+#include <linux/delay.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/regulator/consumer.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/sys_soc.h>
+#include <linux/thermal.h>
+
+#include "sdhci-pltfm.h"
+
+#define SDHCI_OMAP_CON		0x12c
+#define CON_DW8			BIT(5)
+#define CON_DMA_MASTER		BIT(20)
+#define CON_DDR			BIT(19)
+#define CON_CLKEXTFREE		BIT(16)
+#define CON_PADEN		BIT(15)
+#define CON_CTPL		BIT(11)
+#define CON_INIT		BIT(1)
+#define CON_OD			BIT(0)
+
+#define SDHCI_OMAP_DLL		0x0134
+#define DLL_SWT			BIT(20)
+#define DLL_FORCE_SR_C_SHIFT	13
+#define DLL_FORCE_SR_C_MASK	(0x7f << DLL_FORCE_SR_C_SHIFT)
+#define DLL_FORCE_VALUE		BIT(12)
+#define DLL_CALIB		BIT(1)
+
+#define SDHCI_OMAP_CMD		0x20c
+
+#define SDHCI_OMAP_PSTATE	0x0224
+#define PSTATE_DLEV_DAT0	BIT(20)
+#define PSTATE_DATI		BIT(1)
+
+#define SDHCI_OMAP_HCTL		0x228
+#define HCTL_SDBP		BIT(8)
+#define HCTL_SDVS_SHIFT		9
+#define HCTL_SDVS_MASK		(0x7 << HCTL_SDVS_SHIFT)
+#define HCTL_SDVS_33		(0x7 << HCTL_SDVS_SHIFT)
+#define HCTL_SDVS_30		(0x6 << HCTL_SDVS_SHIFT)
+#define HCTL_SDVS_18		(0x5 << HCTL_SDVS_SHIFT)
+
+#define SDHCI_OMAP_SYSCTL	0x22c
+#define SYSCTL_CEN		BIT(2)
+#define SYSCTL_CLKD_SHIFT	6
+#define SYSCTL_CLKD_MASK	0x3ff
+
+#define SDHCI_OMAP_STAT		0x230
+
+#define SDHCI_OMAP_IE		0x234
+#define INT_CC_EN		BIT(0)
+
+#define SDHCI_OMAP_ISE		0x238
+
+#define SDHCI_OMAP_AC12		0x23c
+#define AC12_V1V8_SIGEN		BIT(19)
+#define AC12_SCLK_SEL		BIT(23)
+
+#define SDHCI_OMAP_CAPA		0x240
+#define CAPA_VS33		BIT(24)
+#define CAPA_VS30		BIT(25)
+#define CAPA_VS18		BIT(26)
+
+#define SDHCI_OMAP_CAPA2	0x0244
+#define CAPA2_TSDR50		BIT(13)
+
+#define SDHCI_OMAP_TIMEOUT	1		/* 1 msec */
+
+#define SYSCTL_CLKD_MAX		0x3FF
+
+#define IOV_1V8			1800000		/* 180000 uV */
+#define IOV_3V0			3000000		/* 300000 uV */
+#define IOV_3V3			3300000		/* 330000 uV */
+
+#define MAX_PHASE_DELAY		0x7C
+
+/* sdhci-omap controller flags */
+#define SDHCI_OMAP_REQUIRE_IODELAY	BIT(0)
+#define SDHCI_OMAP_SPECIAL_RESET	BIT(1)
+
+struct sdhci_omap_data {
+	u32 offset;
+	u8 flags;
+};
+
+struct sdhci_omap_host {
+	char			*version;
+	void __iomem		*base;
+	struct device		*dev;
+	struct	regulator	*pbias;
+	bool			pbias_enabled;
+	struct sdhci_host	*host;
+	u8			bus_mode;
+	u8			power_mode;
+	u8			timing;
+	u8			flags;
+
+	struct pinctrl		*pinctrl;
+	struct pinctrl_state	**pinctrl_state;
+	bool			is_tuning;
+	/* Omap specific context save */
+	u32			con;
+	u32			hctl;
+	u32			sysctl;
+	u32			capa;
+	u32			ie;
+	u32			ise;
+};
+
+static void sdhci_omap_start_clock(struct sdhci_omap_host *omap_host);
+static void sdhci_omap_stop_clock(struct sdhci_omap_host *omap_host);
+
+static inline u32 sdhci_omap_readl(struct sdhci_omap_host *host,
+				   unsigned int offset)
+{
+	return readl(host->base + offset);
+}
+
+static inline void sdhci_omap_writel(struct sdhci_omap_host *host,
+				     unsigned int offset, u32 data)
+{
+	writel(data, host->base + offset);
+}
+
+static int sdhci_omap_set_pbias(struct sdhci_omap_host *omap_host,
+				bool power_on, unsigned int iov)
+{
+	int ret;
+	struct device *dev = omap_host->dev;
+
+	if (IS_ERR(omap_host->pbias))
+		return 0;
+
+	if (power_on) {
+		ret = regulator_set_voltage(omap_host->pbias, iov, iov);
+		if (ret) {
+			dev_err(dev, "pbias set voltage failed\n");
+			return ret;
+		}
+
+		if (omap_host->pbias_enabled)
+			return 0;
+
+		ret = regulator_enable(omap_host->pbias);
+		if (ret) {
+			dev_err(dev, "pbias reg enable fail\n");
+			return ret;
+		}
+
+		omap_host->pbias_enabled = true;
+	} else {
+		if (!omap_host->pbias_enabled)
+			return 0;
+
+		ret = regulator_disable(omap_host->pbias);
+		if (ret) {
+			dev_err(dev, "pbias reg disable fail\n");
+			return ret;
+		}
+		omap_host->pbias_enabled = false;
+	}
+
+	return 0;
+}
+
+static int sdhci_omap_enable_iov(struct sdhci_omap_host *omap_host,
+				 unsigned int iov)
+{
+	int ret;
+	struct sdhci_host *host = omap_host->host;
+	struct mmc_host *mmc = host->mmc;
+
+	ret = sdhci_omap_set_pbias(omap_host, false, 0);
+	if (ret)
+		return ret;
+
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+		ret = regulator_set_voltage(mmc->supply.vqmmc, iov, iov);
+		if (ret) {
+			dev_err(mmc_dev(mmc), "vqmmc set voltage failed\n");
+			return ret;
+		}
+	}
+
+	ret = sdhci_omap_set_pbias(omap_host, true, iov);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static void sdhci_omap_conf_bus_power(struct sdhci_omap_host *omap_host,
+				      unsigned char signal_voltage)
+{
+	u32 reg;
+	ktime_t timeout;
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_HCTL);
+	reg &= ~HCTL_SDVS_MASK;
+
+	if (signal_voltage == MMC_SIGNAL_VOLTAGE_330)
+		reg |= HCTL_SDVS_33;
+	else
+		reg |= HCTL_SDVS_18;
+
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_HCTL, reg);
+
+	reg |= HCTL_SDBP;
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_HCTL, reg);
+
+	/* wait 1ms */
+	timeout = ktime_add_ms(ktime_get(), SDHCI_OMAP_TIMEOUT);
+	while (1) {
+		bool timedout = ktime_after(ktime_get(), timeout);
+
+		if (sdhci_omap_readl(omap_host, SDHCI_OMAP_HCTL) & HCTL_SDBP)
+			break;
+		if (WARN_ON(timedout))
+			return;
+		usleep_range(5, 10);
+	}
+}
+
+static void sdhci_omap_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_omap_host *omap_host = sdhci_pltfm_priv(pltfm_host);
+	u32 reg;
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_CON);
+	if (enable)
+		reg |= (CON_CTPL | CON_CLKEXTFREE);
+	else
+		reg &= ~(CON_CTPL | CON_CLKEXTFREE);
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_CON, reg);
+
+	sdhci_enable_sdio_irq(mmc, enable);
+}
+
+static inline void sdhci_omap_set_dll(struct sdhci_omap_host *omap_host,
+				      int count)
+{
+	int i;
+	u32 reg;
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_DLL);
+	reg |= DLL_FORCE_VALUE;
+	reg &= ~DLL_FORCE_SR_C_MASK;
+	reg |= (count << DLL_FORCE_SR_C_SHIFT);
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_DLL, reg);
+
+	reg |= DLL_CALIB;
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_DLL, reg);
+	for (i = 0; i < 1000; i++) {
+		reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_DLL);
+		if (reg & DLL_CALIB)
+			break;
+	}
+	reg &= ~DLL_CALIB;
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_DLL, reg);
+}
+
+static void sdhci_omap_disable_tuning(struct sdhci_omap_host *omap_host)
+{
+	u32 reg;
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_AC12);
+	reg &= ~AC12_SCLK_SEL;
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_AC12, reg);
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_DLL);
+	reg &= ~(DLL_FORCE_VALUE | DLL_SWT);
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_DLL, reg);
+}
+
+static int sdhci_omap_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_omap_host *omap_host = sdhci_pltfm_priv(pltfm_host);
+	struct thermal_zone_device *thermal_dev;
+	struct device *dev = omap_host->dev;
+	struct mmc_ios *ios = &mmc->ios;
+	u32 start_window = 0, max_window = 0;
+	bool single_point_failure = false;
+	bool dcrc_was_enabled = false;
+	u8 cur_match, prev_match = 0;
+	u32 length = 0, max_len = 0;
+	u32 phase_delay = 0;
+	int temperature;
+	int ret = 0;
+	u32 reg;
+	int i;
+
+	/* clock tuning is not needed for upto 52MHz */
+	if (ios->clock <= 52000000)
+		return 0;
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_CAPA2);
+	if (ios->timing == MMC_TIMING_UHS_SDR50 && !(reg & CAPA2_TSDR50))
+		return 0;
+
+	thermal_dev = thermal_zone_get_zone_by_name("cpu_thermal");
+	if (IS_ERR(thermal_dev)) {
+		dev_err(dev, "Unable to get thermal zone for tuning\n");
+		return PTR_ERR(thermal_dev);
+	}
+
+	ret = thermal_zone_get_temp(thermal_dev, &temperature);
+	if (ret)
+		return ret;
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_DLL);
+	reg |= DLL_SWT;
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_DLL, reg);
+
+	/*
+	 * OMAP5/DRA74X/DRA72x Errata i802:
+	 * DCRC error interrupts (MMCHS_STAT[21] DCRC=0x1) can occur
+	 * during the tuning procedure. So disable it during the
+	 * tuning procedure.
+	 */
+	if (host->ier & SDHCI_INT_DATA_CRC) {
+		host->ier &= ~SDHCI_INT_DATA_CRC;
+		dcrc_was_enabled = true;
+	}
+
+	omap_host->is_tuning = true;
+
+	/*
+	 * Stage 1: Search for a maximum pass window ignoring any
+	 * any single point failures. If the tuning value ends up
+	 * near it, move away from it in stage 2 below
+	 */
+	while (phase_delay <= MAX_PHASE_DELAY) {
+		sdhci_omap_set_dll(omap_host, phase_delay);
+
+		cur_match = !mmc_send_tuning(mmc, opcode, NULL);
+		if (cur_match) {
+			if (prev_match) {
+				length++;
+			} else if (single_point_failure) {
+				/* ignore single point failure */
+				length++;
+			} else {
+				start_window = phase_delay;
+				length = 1;
+			}
+		} else {
+			single_point_failure = prev_match;
+		}
+
+		if (length > max_len) {
+			max_window = start_window;
+			max_len = length;
+		}
+
+		prev_match = cur_match;
+		phase_delay += 4;
+	}
+
+	if (!max_len) {
+		dev_err(dev, "Unable to find match\n");
+		ret = -EIO;
+		goto tuning_error;
+	}
+
+	/*
+	 * Assign tuning value as a ratio of maximum pass window based
+	 * on temperature
+	 */
+	if (temperature < -20000)
+		phase_delay = min(max_window + 4 * (max_len - 1) - 24,
+				  max_window +
+				  DIV_ROUND_UP(13 * max_len, 16) * 4);
+	else if (temperature < 20000)
+		phase_delay = max_window + DIV_ROUND_UP(9 * max_len, 16) * 4;
+	else if (temperature < 40000)
+		phase_delay = max_window + DIV_ROUND_UP(8 * max_len, 16) * 4;
+	else if (temperature < 70000)
+		phase_delay = max_window + DIV_ROUND_UP(7 * max_len, 16) * 4;
+	else if (temperature < 90000)
+		phase_delay = max_window + DIV_ROUND_UP(5 * max_len, 16) * 4;
+	else if (temperature < 120000)
+		phase_delay = max_window + DIV_ROUND_UP(4 * max_len, 16) * 4;
+	else
+		phase_delay = max_window + DIV_ROUND_UP(3 * max_len, 16) * 4;
+
+	/*
+	 * Stage 2: Search for a single point failure near the chosen tuning
+	 * value in two steps. First in the +3 to +10 range and then in the
+	 * +2 to -10 range. If found, move away from it in the appropriate
+	 * direction by the appropriate amount depending on the temperature.
+	 */
+	for (i = 3; i <= 10; i++) {
+		sdhci_omap_set_dll(omap_host, phase_delay + i);
+
+		if (mmc_send_tuning(mmc, opcode, NULL)) {
+			if (temperature < 10000)
+				phase_delay += i + 6;
+			else if (temperature < 20000)
+				phase_delay += i - 12;
+			else if (temperature < 70000)
+				phase_delay += i - 8;
+			else
+				phase_delay += i - 6;
+
+			goto single_failure_found;
+		}
+	}
+
+	for (i = 2; i >= -10; i--) {
+		sdhci_omap_set_dll(omap_host, phase_delay + i);
+
+		if (mmc_send_tuning(mmc, opcode, NULL)) {
+			if (temperature < 10000)
+				phase_delay += i + 12;
+			else if (temperature < 20000)
+				phase_delay += i + 8;
+			else if (temperature < 70000)
+				phase_delay += i + 8;
+			else if (temperature < 90000)
+				phase_delay += i + 10;
+			else
+				phase_delay += i + 12;
+
+			goto single_failure_found;
+		}
+	}
+
+single_failure_found:
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_AC12);
+	if (!(reg & AC12_SCLK_SEL)) {
+		ret = -EIO;
+		goto tuning_error;
+	}
+
+	sdhci_omap_set_dll(omap_host, phase_delay);
+
+	omap_host->is_tuning = false;
+
+	goto ret;
+
+tuning_error:
+	omap_host->is_tuning = false;
+	dev_err(dev, "Tuning failed\n");
+	sdhci_omap_disable_tuning(omap_host);
+
+ret:
+	sdhci_reset(host, SDHCI_RESET_CMD | SDHCI_RESET_DATA);
+	/* Reenable forbidden interrupt */
+	if (dcrc_was_enabled)
+		host->ier |= SDHCI_INT_DATA_CRC;
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+	return ret;
+}
+
+static int sdhci_omap_card_busy(struct mmc_host *mmc)
+{
+	u32 reg, ac12;
+	int ret = false;
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_omap_host *omap_host;
+	u32 ier = host->ier;
+
+	pltfm_host = sdhci_priv(host);
+	omap_host = sdhci_pltfm_priv(pltfm_host);
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_CON);
+	ac12 = sdhci_omap_readl(omap_host, SDHCI_OMAP_AC12);
+	reg &= ~CON_CLKEXTFREE;
+	if (ac12 & AC12_V1V8_SIGEN)
+		reg |= CON_CLKEXTFREE;
+	reg |= CON_PADEN;
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_CON, reg);
+
+	disable_irq(host->irq);
+	ier |= SDHCI_INT_CARD_INT;
+	sdhci_writel(host, ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, ier, SDHCI_SIGNAL_ENABLE);
+
+	/*
+	 * Delay is required for PSTATE to correctly reflect
+	 * DLEV/CLEV values after PADEN is set.
+	 */
+	usleep_range(50, 100);
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_PSTATE);
+	if ((reg & PSTATE_DATI) || !(reg & PSTATE_DLEV_DAT0))
+		ret = true;
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_CON);
+	reg &= ~(CON_CLKEXTFREE | CON_PADEN);
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_CON, reg);
+
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+	enable_irq(host->irq);
+
+	return ret;
+}
+
+static int sdhci_omap_start_signal_voltage_switch(struct mmc_host *mmc,
+						  struct mmc_ios *ios)
+{
+	u32 reg;
+	int ret;
+	unsigned int iov;
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_omap_host *omap_host;
+	struct device *dev;
+
+	pltfm_host = sdhci_priv(host);
+	omap_host = sdhci_pltfm_priv(pltfm_host);
+	dev = omap_host->dev;
+
+	if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_330) {
+		reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_CAPA);
+		if (!(reg & CAPA_VS33))
+			return -EOPNOTSUPP;
+
+		sdhci_omap_conf_bus_power(omap_host, ios->signal_voltage);
+
+		reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_AC12);
+		reg &= ~AC12_V1V8_SIGEN;
+		sdhci_omap_writel(omap_host, SDHCI_OMAP_AC12, reg);
+
+		iov = IOV_3V3;
+	} else if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_180) {
+		reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_CAPA);
+		if (!(reg & CAPA_VS18))
+			return -EOPNOTSUPP;
+
+		sdhci_omap_conf_bus_power(omap_host, ios->signal_voltage);
+
+		reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_AC12);
+		reg |= AC12_V1V8_SIGEN;
+		sdhci_omap_writel(omap_host, SDHCI_OMAP_AC12, reg);
+
+		iov = IOV_1V8;
+	} else {
+		return -EOPNOTSUPP;
+	}
+
+	ret = sdhci_omap_enable_iov(omap_host, iov);
+	if (ret) {
+		dev_err(dev, "failed to switch IO voltage to %dmV\n", iov);
+		return ret;
+	}
+
+	dev_dbg(dev, "IO voltage switched to %dmV\n", iov);
+	return 0;
+}
+
+static void sdhci_omap_set_timing(struct sdhci_omap_host *omap_host, u8 timing)
+{
+	int ret;
+	struct pinctrl_state *pinctrl_state;
+	struct device *dev = omap_host->dev;
+
+	if (!(omap_host->flags & SDHCI_OMAP_REQUIRE_IODELAY))
+		return;
+
+	if (omap_host->timing == timing)
+		return;
+
+	sdhci_omap_stop_clock(omap_host);
+
+	pinctrl_state = omap_host->pinctrl_state[timing];
+	ret = pinctrl_select_state(omap_host->pinctrl, pinctrl_state);
+	if (ret) {
+		dev_err(dev, "failed to select pinctrl state\n");
+		return;
+	}
+
+	sdhci_omap_start_clock(omap_host);
+	omap_host->timing = timing;
+}
+
+static void sdhci_omap_set_power_mode(struct sdhci_omap_host *omap_host,
+				      u8 power_mode)
+{
+	if (omap_host->bus_mode == MMC_POWER_OFF)
+		sdhci_omap_disable_tuning(omap_host);
+	omap_host->power_mode = power_mode;
+}
+
+static void sdhci_omap_set_bus_mode(struct sdhci_omap_host *omap_host,
+				    unsigned int mode)
+{
+	u32 reg;
+
+	if (omap_host->bus_mode == mode)
+		return;
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_CON);
+	if (mode == MMC_BUSMODE_OPENDRAIN)
+		reg |= CON_OD;
+	else
+		reg &= ~CON_OD;
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_CON, reg);
+
+	omap_host->bus_mode = mode;
+}
+
+static void sdhci_omap_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_omap_host *omap_host;
+
+	pltfm_host = sdhci_priv(host);
+	omap_host = sdhci_pltfm_priv(pltfm_host);
+
+	sdhci_omap_set_bus_mode(omap_host, ios->bus_mode);
+	sdhci_omap_set_timing(omap_host, ios->timing);
+	sdhci_set_ios(mmc, ios);
+	sdhci_omap_set_power_mode(omap_host, ios->power_mode);
+}
+
+static u16 sdhci_omap_calc_divisor(struct sdhci_pltfm_host *host,
+				   unsigned int clock)
+{
+	u16 dsor;
+
+	dsor = DIV_ROUND_UP(clk_get_rate(host->clk), clock);
+	if (dsor > SYSCTL_CLKD_MAX)
+		dsor = SYSCTL_CLKD_MAX;
+
+	return dsor;
+}
+
+static void sdhci_omap_start_clock(struct sdhci_omap_host *omap_host)
+{
+	u32 reg;
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_SYSCTL);
+	reg |= SYSCTL_CEN;
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_SYSCTL, reg);
+}
+
+static void sdhci_omap_stop_clock(struct sdhci_omap_host *omap_host)
+{
+	u32 reg;
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_SYSCTL);
+	reg &= ~SYSCTL_CEN;
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_SYSCTL, reg);
+}
+
+static void sdhci_omap_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_omap_host *omap_host = sdhci_pltfm_priv(pltfm_host);
+	unsigned long clkdiv;
+
+	sdhci_omap_stop_clock(omap_host);
+
+	if (!clock)
+		return;
+
+	clkdiv = sdhci_omap_calc_divisor(pltfm_host, clock);
+	clkdiv = (clkdiv & SYSCTL_CLKD_MASK) << SYSCTL_CLKD_SHIFT;
+	sdhci_enable_clk(host, clkdiv);
+
+	sdhci_omap_start_clock(omap_host);
+}
+
+static void sdhci_omap_set_power(struct sdhci_host *host, unsigned char mode,
+			  unsigned short vdd)
+{
+	struct mmc_host *mmc = host->mmc;
+
+	if (!IS_ERR(mmc->supply.vmmc))
+		mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, vdd);
+}
+
+static int sdhci_omap_enable_dma(struct sdhci_host *host)
+{
+	u32 reg;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_omap_host *omap_host = sdhci_pltfm_priv(pltfm_host);
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_CON);
+	reg &= ~CON_DMA_MASTER;
+	/* Switch to DMA slave mode when using external DMA */
+	if (!host->use_external_dma)
+		reg |= CON_DMA_MASTER;
+
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_CON, reg);
+
+	return 0;
+}
+
+static unsigned int sdhci_omap_get_min_clock(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	return clk_get_rate(pltfm_host->clk) / SYSCTL_CLKD_MAX;
+}
+
+static void sdhci_omap_set_bus_width(struct sdhci_host *host, int width)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_omap_host *omap_host = sdhci_pltfm_priv(pltfm_host);
+	u32 reg;
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_CON);
+	if (width == MMC_BUS_WIDTH_8)
+		reg |= CON_DW8;
+	else
+		reg &= ~CON_DW8;
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_CON, reg);
+
+	sdhci_set_bus_width(host, width);
+}
+
+static void sdhci_omap_init_74_clocks(struct sdhci_host *host, u8 power_mode)
+{
+	u32 reg;
+	ktime_t timeout;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_omap_host *omap_host = sdhci_pltfm_priv(pltfm_host);
+
+	if (omap_host->power_mode == power_mode)
+		return;
+
+	if (power_mode != MMC_POWER_ON)
+		return;
+
+	disable_irq(host->irq);
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_CON);
+	reg |= CON_INIT;
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_CON, reg);
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_CMD, 0x0);
+
+	/* wait 1ms */
+	timeout = ktime_add_ms(ktime_get(), SDHCI_OMAP_TIMEOUT);
+	while (1) {
+		bool timedout = ktime_after(ktime_get(), timeout);
+
+		if (sdhci_omap_readl(omap_host, SDHCI_OMAP_STAT) & INT_CC_EN)
+			break;
+		if (WARN_ON(timedout))
+			return;
+		usleep_range(5, 10);
+	}
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_CON);
+	reg &= ~CON_INIT;
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_CON, reg);
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_STAT, INT_CC_EN);
+
+	enable_irq(host->irq);
+}
+
+static void sdhci_omap_set_uhs_signaling(struct sdhci_host *host,
+					 unsigned int timing)
+{
+	u32 reg;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_omap_host *omap_host = sdhci_pltfm_priv(pltfm_host);
+
+	sdhci_omap_stop_clock(omap_host);
+
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_CON);
+	if (timing == MMC_TIMING_UHS_DDR50 || timing == MMC_TIMING_MMC_DDR52)
+		reg |= CON_DDR;
+	else
+		reg &= ~CON_DDR;
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_CON, reg);
+
+	sdhci_set_uhs_signaling(host, timing);
+	sdhci_omap_start_clock(omap_host);
+}
+
+#define MMC_TIMEOUT_US		20000		/* 20000 micro Sec */
+static void sdhci_omap_reset(struct sdhci_host *host, u8 mask)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_omap_host *omap_host = sdhci_pltfm_priv(pltfm_host);
+	unsigned long limit = MMC_TIMEOUT_US;
+	unsigned long i = 0;
+
+	/* Don't reset data lines during tuning operation */
+	if (omap_host->is_tuning)
+		mask &= ~SDHCI_RESET_DATA;
+
+	if (omap_host->flags & SDHCI_OMAP_SPECIAL_RESET) {
+		sdhci_writeb(host, mask, SDHCI_SOFTWARE_RESET);
+		while ((!(sdhci_readb(host, SDHCI_SOFTWARE_RESET) & mask)) &&
+		       (i++ < limit))
+			udelay(1);
+		i = 0;
+		while ((sdhci_readb(host, SDHCI_SOFTWARE_RESET) & mask) &&
+		       (i++ < limit))
+			udelay(1);
+
+		if (sdhci_readb(host, SDHCI_SOFTWARE_RESET) & mask)
+			dev_err(mmc_dev(host->mmc),
+				"Timeout waiting on controller reset in %s\n",
+				__func__);
+		return;
+	}
+
+	sdhci_reset(host, mask);
+}
+
+#define CMD_ERR_MASK (SDHCI_INT_CRC | SDHCI_INT_END_BIT | SDHCI_INT_INDEX |\
+		      SDHCI_INT_TIMEOUT)
+#define CMD_MASK (CMD_ERR_MASK | SDHCI_INT_RESPONSE)
+
+static u32 sdhci_omap_irq(struct sdhci_host *host, u32 intmask)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_omap_host *omap_host = sdhci_pltfm_priv(pltfm_host);
+
+	if (omap_host->is_tuning && host->cmd && !host->data_early &&
+	    (intmask & CMD_ERR_MASK)) {
+
+		/*
+		 * Since we are not resetting data lines during tuning
+		 * operation, data error or data complete interrupts
+		 * might still arrive. Mark this request as a failure
+		 * but still wait for the data interrupt
+		 */
+		if (intmask & SDHCI_INT_TIMEOUT)
+			host->cmd->error = -ETIMEDOUT;
+		else
+			host->cmd->error = -EILSEQ;
+
+		host->cmd = NULL;
+
+		/*
+		 * Sometimes command error interrupts and command complete
+		 * interrupt will arrive together. Clear all command related
+		 * interrupts here.
+		 */
+		sdhci_writel(host, intmask & CMD_MASK, SDHCI_INT_STATUS);
+		intmask &= ~CMD_MASK;
+	}
+
+	return intmask;
+}
+
+static void sdhci_omap_set_timeout(struct sdhci_host *host,
+				   struct mmc_command *cmd)
+{
+	if (cmd->opcode == MMC_ERASE)
+		sdhci_set_data_timeout_irq(host, false);
+
+	__sdhci_set_timeout(host, cmd);
+}
+
+static struct sdhci_ops sdhci_omap_ops = {
+	.set_clock = sdhci_omap_set_clock,
+	.set_power = sdhci_omap_set_power,
+	.enable_dma = sdhci_omap_enable_dma,
+	.get_max_clock = sdhci_pltfm_clk_get_max_clock,
+	.get_min_clock = sdhci_omap_get_min_clock,
+	.set_bus_width = sdhci_omap_set_bus_width,
+	.platform_send_init_74_clocks = sdhci_omap_init_74_clocks,
+	.reset = sdhci_omap_reset,
+	.set_uhs_signaling = sdhci_omap_set_uhs_signaling,
+	.irq = sdhci_omap_irq,
+	.set_timeout = sdhci_omap_set_timeout,
+};
+
+static int sdhci_omap_set_capabilities(struct sdhci_omap_host *omap_host)
+{
+	u32 reg;
+	int ret = 0;
+	struct device *dev = omap_host->dev;
+	struct regulator *vqmmc;
+
+	vqmmc = regulator_get(dev, "vqmmc");
+	if (IS_ERR(vqmmc)) {
+		ret = PTR_ERR(vqmmc);
+		goto reg_put;
+	}
+
+	/* voltage capabilities might be set by boot loader, clear it */
+	reg = sdhci_omap_readl(omap_host, SDHCI_OMAP_CAPA);
+	reg &= ~(CAPA_VS18 | CAPA_VS30 | CAPA_VS33);
+
+	if (regulator_is_supported_voltage(vqmmc, IOV_3V3, IOV_3V3))
+		reg |= CAPA_VS33;
+	if (regulator_is_supported_voltage(vqmmc, IOV_1V8, IOV_1V8))
+		reg |= CAPA_VS18;
+
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_CAPA, reg);
+
+reg_put:
+	regulator_put(vqmmc);
+
+	return ret;
+}
+
+static const struct sdhci_pltfm_data sdhci_omap_pdata = {
+	.quirks = SDHCI_QUIRK_BROKEN_CARD_DETECTION |
+		  SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK |
+		  SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN |
+		  SDHCI_QUIRK_NO_HISPD_BIT |
+		  SDHCI_QUIRK_BROKEN_ADMA_ZEROLEN_DESC,
+	.quirks2 = SDHCI_QUIRK2_ACMD23_BROKEN |
+		   SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+		   SDHCI_QUIRK2_RSP_136_HAS_CRC |
+		   SDHCI_QUIRK2_DISABLE_HW_TIMEOUT,
+	.ops = &sdhci_omap_ops,
+};
+
+static const struct sdhci_omap_data k2g_data = {
+	.offset = 0x200,
+};
+
+static const struct sdhci_omap_data am335_data = {
+	.offset = 0x200,
+	.flags = SDHCI_OMAP_SPECIAL_RESET,
+};
+
+static const struct sdhci_omap_data am437_data = {
+	.offset = 0x200,
+	.flags = SDHCI_OMAP_SPECIAL_RESET,
+};
+
+static const struct sdhci_omap_data dra7_data = {
+	.offset = 0x200,
+	.flags	= SDHCI_OMAP_REQUIRE_IODELAY,
+};
+
+static const struct of_device_id omap_sdhci_match[] = {
+	{ .compatible = "ti,dra7-sdhci", .data = &dra7_data },
+	{ .compatible = "ti,k2g-sdhci", .data = &k2g_data },
+	{ .compatible = "ti,am335-sdhci", .data = &am335_data },
+	{ .compatible = "ti,am437-sdhci", .data = &am437_data },
+	{},
+};
+MODULE_DEVICE_TABLE(of, omap_sdhci_match);
+
+static struct pinctrl_state
+*sdhci_omap_iodelay_pinctrl_state(struct sdhci_omap_host *omap_host, char *mode,
+				  u32 *caps, u32 capmask)
+{
+	struct device *dev = omap_host->dev;
+	char *version = omap_host->version;
+	struct pinctrl_state *pinctrl_state = ERR_PTR(-ENODEV);
+	char str[20];
+
+	if (!(*caps & capmask))
+		goto ret;
+
+	if (version) {
+		snprintf(str, 20, "%s-%s", mode, version);
+		pinctrl_state = pinctrl_lookup_state(omap_host->pinctrl, str);
+	}
+
+	if (IS_ERR(pinctrl_state))
+		pinctrl_state = pinctrl_lookup_state(omap_host->pinctrl, mode);
+
+	if (IS_ERR(pinctrl_state)) {
+		dev_err(dev, "no pinctrl state for %s mode", mode);
+		*caps &= ~capmask;
+	}
+
+ret:
+	return pinctrl_state;
+}
+
+static int sdhci_omap_config_iodelay_pinctrl_state(struct sdhci_omap_host
+						   *omap_host)
+{
+	struct device *dev = omap_host->dev;
+	struct sdhci_host *host = omap_host->host;
+	struct mmc_host *mmc = host->mmc;
+	u32 *caps = &mmc->caps;
+	u32 *caps2 = &mmc->caps2;
+	struct pinctrl_state *state;
+	struct pinctrl_state **pinctrl_state;
+
+	if (!(omap_host->flags & SDHCI_OMAP_REQUIRE_IODELAY))
+		return 0;
+
+	pinctrl_state = devm_kcalloc(dev,
+				     MMC_TIMING_MMC_HS200 + 1,
+				     sizeof(*pinctrl_state),
+				     GFP_KERNEL);
+	if (!pinctrl_state)
+		return -ENOMEM;
+
+	omap_host->pinctrl = devm_pinctrl_get(omap_host->dev);
+	if (IS_ERR(omap_host->pinctrl)) {
+		dev_err(dev, "Cannot get pinctrl\n");
+		return PTR_ERR(omap_host->pinctrl);
+	}
+
+	state = pinctrl_lookup_state(omap_host->pinctrl, "default");
+	if (IS_ERR(state)) {
+		dev_err(dev, "no pinctrl state for default mode\n");
+		return PTR_ERR(state);
+	}
+	pinctrl_state[MMC_TIMING_LEGACY] = state;
+
+	state = sdhci_omap_iodelay_pinctrl_state(omap_host, "sdr104", caps,
+						 MMC_CAP_UHS_SDR104);
+	if (!IS_ERR(state))
+		pinctrl_state[MMC_TIMING_UHS_SDR104] = state;
+
+	state = sdhci_omap_iodelay_pinctrl_state(omap_host, "ddr50", caps,
+						 MMC_CAP_UHS_DDR50);
+	if (!IS_ERR(state))
+		pinctrl_state[MMC_TIMING_UHS_DDR50] = state;
+
+	state = sdhci_omap_iodelay_pinctrl_state(omap_host, "sdr50", caps,
+						 MMC_CAP_UHS_SDR50);
+	if (!IS_ERR(state))
+		pinctrl_state[MMC_TIMING_UHS_SDR50] = state;
+
+	state = sdhci_omap_iodelay_pinctrl_state(omap_host, "sdr25", caps,
+						 MMC_CAP_UHS_SDR25);
+	if (!IS_ERR(state))
+		pinctrl_state[MMC_TIMING_UHS_SDR25] = state;
+
+	state = sdhci_omap_iodelay_pinctrl_state(omap_host, "sdr12", caps,
+						 MMC_CAP_UHS_SDR12);
+	if (!IS_ERR(state))
+		pinctrl_state[MMC_TIMING_UHS_SDR12] = state;
+
+	state = sdhci_omap_iodelay_pinctrl_state(omap_host, "ddr_1_8v", caps,
+						 MMC_CAP_1_8V_DDR);
+	if (!IS_ERR(state)) {
+		pinctrl_state[MMC_TIMING_MMC_DDR52] = state;
+	} else {
+		state = sdhci_omap_iodelay_pinctrl_state(omap_host, "ddr_3_3v",
+							 caps,
+							 MMC_CAP_3_3V_DDR);
+		if (!IS_ERR(state))
+			pinctrl_state[MMC_TIMING_MMC_DDR52] = state;
+	}
+
+	state = sdhci_omap_iodelay_pinctrl_state(omap_host, "hs", caps,
+						 MMC_CAP_SD_HIGHSPEED);
+	if (!IS_ERR(state))
+		pinctrl_state[MMC_TIMING_SD_HS] = state;
+
+	state = sdhci_omap_iodelay_pinctrl_state(omap_host, "hs", caps,
+						 MMC_CAP_MMC_HIGHSPEED);
+	if (!IS_ERR(state))
+		pinctrl_state[MMC_TIMING_MMC_HS] = state;
+
+	state = sdhci_omap_iodelay_pinctrl_state(omap_host, "hs200_1_8v", caps2,
+						 MMC_CAP2_HS200_1_8V_SDR);
+	if (!IS_ERR(state))
+		pinctrl_state[MMC_TIMING_MMC_HS200] = state;
+
+	omap_host->pinctrl_state = pinctrl_state;
+
+	return 0;
+}
+
+static const struct soc_device_attribute sdhci_omap_soc_devices[] = {
+	{
+		.machine = "DRA7[45]*",
+		.revision = "ES1.[01]",
+	},
+	{
+		/* sentinel */
+	}
+};
+
+static int sdhci_omap_probe(struct platform_device *pdev)
+{
+	int ret;
+	u32 offset;
+	struct device *dev = &pdev->dev;
+	struct sdhci_host *host;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_omap_host *omap_host;
+	struct mmc_host *mmc;
+	const struct of_device_id *match;
+	struct sdhci_omap_data *data;
+	const struct soc_device_attribute *soc;
+	struct resource *regs;
+
+	match = of_match_device(omap_sdhci_match, dev);
+	if (!match)
+		return -EINVAL;
+
+	data = (struct sdhci_omap_data *)match->data;
+	if (!data) {
+		dev_err(dev, "no sdhci omap data\n");
+		return -EINVAL;
+	}
+	offset = data->offset;
+
+	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!regs)
+		return -ENXIO;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_omap_pdata,
+				sizeof(*omap_host));
+	if (IS_ERR(host)) {
+		dev_err(dev, "Failed sdhci_pltfm_init\n");
+		return PTR_ERR(host);
+	}
+
+	pltfm_host = sdhci_priv(host);
+	omap_host = sdhci_pltfm_priv(pltfm_host);
+	omap_host->host = host;
+	omap_host->base = host->ioaddr;
+	omap_host->dev = dev;
+	omap_host->power_mode = MMC_POWER_UNDEFINED;
+	omap_host->timing = MMC_TIMING_LEGACY;
+	omap_host->flags = data->flags;
+	host->ioaddr += offset;
+	host->mapbase = regs->start + offset;
+
+	mmc = host->mmc;
+	sdhci_get_of_property(pdev);
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto err_pltfm_free;
+
+	soc = soc_device_match(sdhci_omap_soc_devices);
+	if (soc) {
+		omap_host->version = "rev11";
+		if (!strcmp(dev_name(dev), "4809c000.mmc"))
+			mmc->f_max = 96000000;
+		if (!strcmp(dev_name(dev), "480b4000.mmc"))
+			mmc->f_max = 48000000;
+		if (!strcmp(dev_name(dev), "480ad000.mmc"))
+			mmc->f_max = 48000000;
+	}
+
+	if (!mmc_can_gpio_ro(mmc))
+		mmc->caps2 |= MMC_CAP2_NO_WRITE_PROTECT;
+
+	pltfm_host->clk = devm_clk_get(dev, "fck");
+	if (IS_ERR(pltfm_host->clk)) {
+		ret = PTR_ERR(pltfm_host->clk);
+		goto err_pltfm_free;
+	}
+
+	ret = clk_set_rate(pltfm_host->clk, mmc->f_max);
+	if (ret) {
+		dev_err(dev, "failed to set clock to %d\n", mmc->f_max);
+		goto err_pltfm_free;
+	}
+
+	omap_host->pbias = devm_regulator_get_optional(dev, "pbias");
+	if (IS_ERR(omap_host->pbias)) {
+		ret = PTR_ERR(omap_host->pbias);
+		if (ret != -ENODEV)
+			goto err_pltfm_free;
+		dev_dbg(dev, "unable to get pbias regulator %d\n", ret);
+	}
+	omap_host->pbias_enabled = false;
+
+	/*
+	 * omap_device_pm_domain has callbacks to enable the main
+	 * functional clock, interface clock and also configure the
+	 * SYSCONFIG register of omap devices. The callback will be invoked
+	 * as part of pm_runtime_get_sync.
+	 */
+	pm_runtime_enable(dev);
+	ret = pm_runtime_get_sync(dev);
+	if (ret < 0) {
+		dev_err(dev, "pm_runtime_get_sync failed\n");
+		pm_runtime_put_noidle(dev);
+		goto err_rpm_disable;
+	}
+
+	ret = sdhci_omap_set_capabilities(omap_host);
+	if (ret) {
+		dev_err(dev, "failed to set system capabilities\n");
+		goto err_put_sync;
+	}
+
+	host->mmc_host_ops.start_signal_voltage_switch =
+					sdhci_omap_start_signal_voltage_switch;
+	host->mmc_host_ops.set_ios = sdhci_omap_set_ios;
+	host->mmc_host_ops.card_busy = sdhci_omap_card_busy;
+	host->mmc_host_ops.execute_tuning = sdhci_omap_execute_tuning;
+	host->mmc_host_ops.enable_sdio_irq = sdhci_omap_enable_sdio_irq;
+
+	/* Switch to external DMA only if there is the "dmas" property */
+	if (of_find_property(dev->of_node, "dmas", NULL))
+		sdhci_switch_external_dma(host, true);
+
+	/* R1B responses is required to properly manage HW busy detection. */
+	mmc->caps |= MMC_CAP_NEED_RSP_BUSY;
+
+	ret = sdhci_setup_host(host);
+	if (ret)
+		goto err_put_sync;
+
+	ret = sdhci_omap_config_iodelay_pinctrl_state(omap_host);
+	if (ret)
+		goto err_cleanup_host;
+
+	ret = __sdhci_add_host(host);
+	if (ret)
+		goto err_cleanup_host;
+
+	return 0;
+
+err_cleanup_host:
+	sdhci_cleanup_host(host);
+
+err_put_sync:
+	pm_runtime_put_sync(dev);
+
+err_rpm_disable:
+	pm_runtime_disable(dev);
+
+err_pltfm_free:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static int sdhci_omap_remove(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+
+	sdhci_remove_host(host, true);
+	pm_runtime_put_sync(dev);
+	pm_runtime_disable(dev);
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+#ifdef CONFIG_PM_SLEEP
+static void sdhci_omap_context_save(struct sdhci_omap_host *omap_host)
+{
+	omap_host->con = sdhci_omap_readl(omap_host, SDHCI_OMAP_CON);
+	omap_host->hctl = sdhci_omap_readl(omap_host, SDHCI_OMAP_HCTL);
+	omap_host->sysctl = sdhci_omap_readl(omap_host, SDHCI_OMAP_SYSCTL);
+	omap_host->capa = sdhci_omap_readl(omap_host, SDHCI_OMAP_CAPA);
+	omap_host->ie = sdhci_omap_readl(omap_host, SDHCI_OMAP_IE);
+	omap_host->ise = sdhci_omap_readl(omap_host, SDHCI_OMAP_ISE);
+}
+
+/* Order matters here, HCTL must be restored in two phases */
+static void sdhci_omap_context_restore(struct sdhci_omap_host *omap_host)
+{
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_HCTL, omap_host->hctl);
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_CAPA, omap_host->capa);
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_HCTL, omap_host->hctl);
+
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_SYSCTL, omap_host->sysctl);
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_CON, omap_host->con);
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_IE, omap_host->ie);
+	sdhci_omap_writel(omap_host, SDHCI_OMAP_ISE, omap_host->ise);
+}
+
+static int __maybe_unused sdhci_omap_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_omap_host *omap_host = sdhci_pltfm_priv(pltfm_host);
+
+	sdhci_suspend_host(host);
+
+	sdhci_omap_context_save(omap_host);
+
+	pinctrl_pm_select_idle_state(dev);
+
+	pm_runtime_force_suspend(dev);
+
+	return 0;
+}
+
+static int __maybe_unused sdhci_omap_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_omap_host *omap_host = sdhci_pltfm_priv(pltfm_host);
+
+	pm_runtime_force_resume(dev);
+
+	pinctrl_pm_select_default_state(dev);
+
+	sdhci_omap_context_restore(omap_host);
+
+	sdhci_resume_host(host);
+
+	return 0;
+}
+#endif
+static SIMPLE_DEV_PM_OPS(sdhci_omap_dev_pm_ops, sdhci_omap_suspend,
+			 sdhci_omap_resume);
+
+static struct platform_driver sdhci_omap_driver = {
+	.probe = sdhci_omap_probe,
+	.remove = sdhci_omap_remove,
+	.driver = {
+		   .name = "sdhci-omap",
+		   .probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		   .pm = &sdhci_omap_dev_pm_ops,
+		   .of_match_table = omap_sdhci_match,
+		  },
+};
+
+module_platform_driver(sdhci_omap_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for OMAP SoCs");
+MODULE_AUTHOR("Texas Instruments Inc.");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:sdhci_omap");
diff --git a/drivers/mmc/host/sdhci-pci-arasan.c b/drivers/mmc/host/sdhci-pci-arasan.c
new file mode 100644
index 000000000..499f3205e
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pci-arasan.c
@@ -0,0 +1,331 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * sdhci-pci-arasan.c - Driver for Arasan PCI Controller with
+ * integrated phy.
+ *
+ * Copyright (C) 2017 Arasan Chip Systems Inc.
+ *
+ * Author: Atul Garg <agarg@arasan.com>
+ */
+
+#include <linux/pci.h>
+#include <linux/delay.h>
+
+#include "sdhci.h"
+#include "sdhci-pci.h"
+
+/* Extra registers for Arasan SD/SDIO/MMC Host Controller with PHY */
+#define PHY_ADDR_REG	0x300
+#define PHY_DAT_REG	0x304
+
+#define PHY_WRITE	BIT(8)
+#define PHY_BUSY	BIT(9)
+#define DATA_MASK	0xFF
+
+/* PHY Specific Registers */
+#define DLL_STATUS	0x00
+#define IPAD_CTRL1	0x01
+#define IPAD_CTRL2	0x02
+#define IPAD_STS	0x03
+#define IOREN_CTRL1	0x06
+#define IOREN_CTRL2	0x07
+#define IOPU_CTRL1	0x08
+#define IOPU_CTRL2	0x09
+#define ITAP_DELAY	0x0C
+#define OTAP_DELAY	0x0D
+#define STRB_SEL	0x0E
+#define CLKBUF_SEL	0x0F
+#define MODE_CTRL	0x11
+#define DLL_TRIM	0x12
+#define CMD_CTRL	0x20
+#define DATA_CTRL	0x21
+#define STRB_CTRL	0x22
+#define CLK_CTRL	0x23
+#define PHY_CTRL	0x24
+
+#define DLL_ENBL	BIT(3)
+#define RTRIM_EN	BIT(1)
+#define PDB_ENBL	BIT(1)
+#define RETB_ENBL	BIT(6)
+#define ODEN_CMD	BIT(1)
+#define ODEN_DAT	0xFF
+#define REN_STRB	BIT(0)
+#define REN_CMND	BIT(1)
+#define REN_DATA	0xFF
+#define PU_CMD		BIT(1)
+#define PU_DAT		0xFF
+#define ITAPDLY_EN	BIT(0)
+#define OTAPDLY_EN	BIT(0)
+#define OD_REL_CMD	BIT(1)
+#define OD_REL_DAT	0xFF
+#define DLLTRM_ICP	0x8
+#define PDB_CMND	BIT(0)
+#define PDB_DATA	0xFF
+#define PDB_STRB	BIT(0)
+#define PDB_CLOCK	BIT(0)
+#define CALDONE_MASK	0x10
+#define DLL_RDY_MASK	0x10
+#define MAX_CLK_BUF	0x7
+
+/* Mode Controls */
+#define ENHSTRB_MODE	BIT(0)
+#define HS400_MODE	BIT(1)
+#define LEGACY_MODE	BIT(2)
+#define DDR50_MODE	BIT(3)
+
+/*
+ * Controller has no specific bits for HS200/HS.
+ * Used BIT(4), BIT(5) for software programming.
+ */
+#define HS200_MODE	BIT(4)
+#define HISPD_MODE	BIT(5)
+
+#define OTAPDLY(x)	(((x) << 1) | OTAPDLY_EN)
+#define ITAPDLY(x)	(((x) << 1) | ITAPDLY_EN)
+#define FREQSEL(x)	(((x) << 5) | DLL_ENBL)
+#define IOPAD(x, y)	((x) | ((y) << 2))
+
+/* Arasan private data */
+struct arasan_host {
+	u32 chg_clk;
+};
+
+static int arasan_phy_addr_poll(struct sdhci_host *host, u32 offset, u32 mask)
+{
+	ktime_t timeout = ktime_add_us(ktime_get(), 100);
+	bool failed;
+	u8 val = 0;
+
+	while (1) {
+		failed = ktime_after(ktime_get(), timeout);
+		val = sdhci_readw(host, PHY_ADDR_REG);
+		if (!(val & mask))
+			return 0;
+		if (failed)
+			return -EBUSY;
+	}
+}
+
+static int arasan_phy_write(struct sdhci_host *host, u8 data, u8 offset)
+{
+	sdhci_writew(host, data, PHY_DAT_REG);
+	sdhci_writew(host, (PHY_WRITE | offset), PHY_ADDR_REG);
+	return arasan_phy_addr_poll(host, PHY_ADDR_REG, PHY_BUSY);
+}
+
+static int arasan_phy_read(struct sdhci_host *host, u8 offset, u8 *data)
+{
+	int ret;
+
+	sdhci_writew(host, 0, PHY_DAT_REG);
+	sdhci_writew(host, offset, PHY_ADDR_REG);
+	ret = arasan_phy_addr_poll(host, PHY_ADDR_REG, PHY_BUSY);
+
+	/* Masking valid data bits */
+	*data = sdhci_readw(host, PHY_DAT_REG) & DATA_MASK;
+	return ret;
+}
+
+static int arasan_phy_sts_poll(struct sdhci_host *host, u32 offset, u32 mask)
+{
+	int ret;
+	ktime_t timeout = ktime_add_us(ktime_get(), 100);
+	bool failed;
+	u8 val = 0;
+
+	while (1) {
+		failed = ktime_after(ktime_get(), timeout);
+		ret = arasan_phy_read(host, offset, &val);
+		if (ret)
+			return -EBUSY;
+		else if (val & mask)
+			return 0;
+		if (failed)
+			return -EBUSY;
+	}
+}
+
+/* Initialize the Arasan PHY */
+static int arasan_phy_init(struct sdhci_host *host)
+{
+	int ret;
+	u8 val;
+
+	/* Program IOPADs and wait for calibration to be done */
+	if (arasan_phy_read(host, IPAD_CTRL1, &val) ||
+	    arasan_phy_write(host, val | RETB_ENBL | PDB_ENBL, IPAD_CTRL1) ||
+	    arasan_phy_read(host, IPAD_CTRL2, &val) ||
+	    arasan_phy_write(host, val | RTRIM_EN, IPAD_CTRL2))
+		return -EBUSY;
+	ret = arasan_phy_sts_poll(host, IPAD_STS, CALDONE_MASK);
+	if (ret)
+		return -EBUSY;
+
+	/* Program CMD/Data lines */
+	if (arasan_phy_read(host, IOREN_CTRL1, &val) ||
+	    arasan_phy_write(host, val | REN_CMND | REN_STRB, IOREN_CTRL1) ||
+	    arasan_phy_read(host, IOPU_CTRL1, &val) ||
+	    arasan_phy_write(host, val | PU_CMD, IOPU_CTRL1) ||
+	    arasan_phy_read(host, CMD_CTRL, &val) ||
+	    arasan_phy_write(host, val | PDB_CMND, CMD_CTRL) ||
+	    arasan_phy_read(host, IOREN_CTRL2, &val) ||
+	    arasan_phy_write(host, val | REN_DATA, IOREN_CTRL2) ||
+	    arasan_phy_read(host, IOPU_CTRL2, &val) ||
+	    arasan_phy_write(host, val | PU_DAT, IOPU_CTRL2) ||
+	    arasan_phy_read(host, DATA_CTRL, &val) ||
+	    arasan_phy_write(host, val | PDB_DATA, DATA_CTRL) ||
+	    arasan_phy_read(host, STRB_CTRL, &val) ||
+	    arasan_phy_write(host, val | PDB_STRB, STRB_CTRL) ||
+	    arasan_phy_read(host, CLK_CTRL, &val) ||
+	    arasan_phy_write(host, val | PDB_CLOCK, CLK_CTRL) ||
+	    arasan_phy_read(host, CLKBUF_SEL, &val) ||
+	    arasan_phy_write(host, val | MAX_CLK_BUF, CLKBUF_SEL) ||
+	    arasan_phy_write(host, LEGACY_MODE, MODE_CTRL))
+		return -EBUSY;
+	return 0;
+}
+
+/* Set Arasan PHY for different modes */
+static int arasan_phy_set(struct sdhci_host *host, u8 mode, u8 otap,
+			  u8 drv_type, u8 itap, u8 trim, u8 clk)
+{
+	u8 val;
+	int ret;
+
+	if (mode == HISPD_MODE || mode == HS200_MODE)
+		ret = arasan_phy_write(host, 0x0, MODE_CTRL);
+	else
+		ret = arasan_phy_write(host, mode, MODE_CTRL);
+	if (ret)
+		return ret;
+	if (mode == HS400_MODE || mode == HS200_MODE) {
+		ret = arasan_phy_read(host, IPAD_CTRL1, &val);
+		if (ret)
+			return ret;
+		ret = arasan_phy_write(host, IOPAD(val, drv_type), IPAD_CTRL1);
+		if (ret)
+			return ret;
+	}
+	if (mode == LEGACY_MODE) {
+		ret = arasan_phy_write(host, 0x0, OTAP_DELAY);
+		if (ret)
+			return ret;
+		ret = arasan_phy_write(host, 0x0, ITAP_DELAY);
+	} else {
+		ret = arasan_phy_write(host, OTAPDLY(otap), OTAP_DELAY);
+		if (ret)
+			return ret;
+		if (mode != HS200_MODE)
+			ret = arasan_phy_write(host, ITAPDLY(itap), ITAP_DELAY);
+		else
+			ret = arasan_phy_write(host, 0x0, ITAP_DELAY);
+	}
+	if (ret)
+		return ret;
+	if (mode != LEGACY_MODE) {
+		ret = arasan_phy_write(host, trim, DLL_TRIM);
+		if (ret)
+			return ret;
+	}
+	ret = arasan_phy_write(host, 0, DLL_STATUS);
+	if (ret)
+		return ret;
+	if (mode != LEGACY_MODE) {
+		ret = arasan_phy_write(host, FREQSEL(clk), DLL_STATUS);
+		if (ret)
+			return ret;
+		ret = arasan_phy_sts_poll(host, DLL_STATUS, DLL_RDY_MASK);
+		if (ret)
+			return -EBUSY;
+	}
+	return 0;
+}
+
+static int arasan_select_phy_clock(struct sdhci_host *host)
+{
+	struct sdhci_pci_slot *slot = sdhci_priv(host);
+	struct arasan_host *arasan_host = sdhci_pci_priv(slot);
+	u8 clk;
+
+	if (arasan_host->chg_clk == host->mmc->ios.clock)
+		return 0;
+
+	arasan_host->chg_clk = host->mmc->ios.clock;
+	if (host->mmc->ios.clock == 200000000)
+		clk = 0x0;
+	else if (host->mmc->ios.clock == 100000000)
+		clk = 0x2;
+	else if (host->mmc->ios.clock == 50000000)
+		clk = 0x1;
+	else
+		clk = 0x0;
+
+	if (host->mmc_host_ops.hs400_enhanced_strobe) {
+		arasan_phy_set(host, ENHSTRB_MODE, 1, 0x0, 0x0,
+			       DLLTRM_ICP, clk);
+	} else {
+		switch (host->mmc->ios.timing) {
+		case MMC_TIMING_LEGACY:
+			arasan_phy_set(host, LEGACY_MODE, 0x0, 0x0, 0x0,
+				       0x0, 0x0);
+			break;
+		case MMC_TIMING_MMC_HS:
+		case MMC_TIMING_SD_HS:
+			arasan_phy_set(host, HISPD_MODE, 0x3, 0x0, 0x2,
+				       DLLTRM_ICP, clk);
+			break;
+		case MMC_TIMING_MMC_HS200:
+		case MMC_TIMING_UHS_SDR104:
+			arasan_phy_set(host, HS200_MODE, 0x2,
+				       host->mmc->ios.drv_type, 0x0,
+				       DLLTRM_ICP, clk);
+			break;
+		case MMC_TIMING_MMC_DDR52:
+		case MMC_TIMING_UHS_DDR50:
+			arasan_phy_set(host, DDR50_MODE, 0x1, 0x0,
+				       0x0, DLLTRM_ICP, clk);
+			break;
+		case MMC_TIMING_MMC_HS400:
+			arasan_phy_set(host, HS400_MODE, 0x1,
+				       host->mmc->ios.drv_type, 0xa,
+				       DLLTRM_ICP, clk);
+			break;
+		default:
+			break;
+		}
+	}
+	return 0;
+}
+
+static int arasan_pci_probe_slot(struct sdhci_pci_slot *slot)
+{
+	int err;
+
+	slot->host->mmc->caps |= MMC_CAP_NONREMOVABLE | MMC_CAP_8_BIT_DATA;
+	err = arasan_phy_init(slot->host);
+	if (err)
+		return -ENODEV;
+	return 0;
+}
+
+static void arasan_sdhci_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	sdhci_set_clock(host, clock);
+
+	/* Change phy settings for the new clock */
+	arasan_select_phy_clock(host);
+}
+
+static const struct sdhci_ops arasan_sdhci_pci_ops = {
+	.set_clock	= arasan_sdhci_set_clock,
+	.enable_dma	= sdhci_pci_enable_dma,
+	.set_bus_width	= sdhci_set_bus_width,
+	.reset		= sdhci_reset,
+	.set_uhs_signaling	= sdhci_set_uhs_signaling,
+};
+
+const struct sdhci_pci_fixes sdhci_arasan = {
+	.probe_slot = arasan_pci_probe_slot,
+	.ops        = &arasan_sdhci_pci_ops,
+	.priv_size  = sizeof(struct arasan_host),
+};
diff --git a/drivers/mmc/host/sdhci-pci-core.c b/drivers/mmc/host/sdhci-pci-core.c
new file mode 100644
index 000000000..8b02fe391
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pci-core.c
@@ -0,0 +1,2401 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*  linux/drivers/mmc/host/sdhci-pci.c - SDHCI on PCI bus interface
+ *
+ *  Copyright (C) 2005-2008 Pierre Ossman, All Rights Reserved.
+ *
+ * Thanks to the following companies for their support:
+ *
+ *     - JMicron (hardware and technical support)
+ */
+
+#include <linux/bitfield.h>
+#include <linux/string.h>
+#include <linux/delay.h>
+#include <linux/highmem.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/dma-mapping.h>
+#include <linux/slab.h>
+#include <linux/device.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/scatterlist.h>
+#include <linux/io.h>
+#include <linux/iopoll.h>
+#include <linux/gpio.h>
+#include <linux/pm_runtime.h>
+#include <linux/pm_qos.h>
+#include <linux/debugfs.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/mmc/sdhci-pci-data.h>
+#include <linux/acpi.h>
+#include <linux/dmi.h>
+
+#ifdef CONFIG_X86
+#include <asm/iosf_mbi.h>
+#endif
+
+#include "cqhci.h"
+
+#include "sdhci.h"
+#include "sdhci-pci.h"
+
+static void sdhci_pci_hw_reset(struct sdhci_host *host);
+
+#ifdef CONFIG_PM_SLEEP
+static int sdhci_pci_init_wakeup(struct sdhci_pci_chip *chip)
+{
+	mmc_pm_flag_t pm_flags = 0;
+	bool cap_cd_wake = false;
+	int i;
+
+	for (i = 0; i < chip->num_slots; i++) {
+		struct sdhci_pci_slot *slot = chip->slots[i];
+
+		if (slot) {
+			pm_flags |= slot->host->mmc->pm_flags;
+			if (slot->host->mmc->caps & MMC_CAP_CD_WAKE)
+				cap_cd_wake = true;
+		}
+	}
+
+	if ((pm_flags & MMC_PM_KEEP_POWER) && (pm_flags & MMC_PM_WAKE_SDIO_IRQ))
+		return device_wakeup_enable(&chip->pdev->dev);
+	else if (!cap_cd_wake)
+		return device_wakeup_disable(&chip->pdev->dev);
+
+	return 0;
+}
+
+static int sdhci_pci_suspend_host(struct sdhci_pci_chip *chip)
+{
+	int i, ret;
+
+	sdhci_pci_init_wakeup(chip);
+
+	for (i = 0; i < chip->num_slots; i++) {
+		struct sdhci_pci_slot *slot = chip->slots[i];
+		struct sdhci_host *host;
+
+		if (!slot)
+			continue;
+
+		host = slot->host;
+
+		if (chip->pm_retune && host->tuning_mode != SDHCI_TUNING_MODE_3)
+			mmc_retune_needed(host->mmc);
+
+		ret = sdhci_suspend_host(host);
+		if (ret)
+			goto err_pci_suspend;
+
+		if (device_may_wakeup(&chip->pdev->dev))
+			mmc_gpio_set_cd_wake(host->mmc, true);
+	}
+
+	return 0;
+
+err_pci_suspend:
+	while (--i >= 0)
+		sdhci_resume_host(chip->slots[i]->host);
+	return ret;
+}
+
+int sdhci_pci_resume_host(struct sdhci_pci_chip *chip)
+{
+	struct sdhci_pci_slot *slot;
+	int i, ret;
+
+	for (i = 0; i < chip->num_slots; i++) {
+		slot = chip->slots[i];
+		if (!slot)
+			continue;
+
+		ret = sdhci_resume_host(slot->host);
+		if (ret)
+			return ret;
+
+		mmc_gpio_set_cd_wake(slot->host->mmc, false);
+	}
+
+	return 0;
+}
+
+static int sdhci_cqhci_suspend(struct sdhci_pci_chip *chip)
+{
+	int ret;
+
+	ret = cqhci_suspend(chip->slots[0]->host->mmc);
+	if (ret)
+		return ret;
+
+	return sdhci_pci_suspend_host(chip);
+}
+
+static int sdhci_cqhci_resume(struct sdhci_pci_chip *chip)
+{
+	int ret;
+
+	ret = sdhci_pci_resume_host(chip);
+	if (ret)
+		return ret;
+
+	return cqhci_resume(chip->slots[0]->host->mmc);
+}
+#endif
+
+#ifdef CONFIG_PM
+static int sdhci_pci_runtime_suspend_host(struct sdhci_pci_chip *chip)
+{
+	struct sdhci_pci_slot *slot;
+	struct sdhci_host *host;
+	int i, ret;
+
+	for (i = 0; i < chip->num_slots; i++) {
+		slot = chip->slots[i];
+		if (!slot)
+			continue;
+
+		host = slot->host;
+
+		ret = sdhci_runtime_suspend_host(host);
+		if (ret)
+			goto err_pci_runtime_suspend;
+
+		if (chip->rpm_retune &&
+		    host->tuning_mode != SDHCI_TUNING_MODE_3)
+			mmc_retune_needed(host->mmc);
+	}
+
+	return 0;
+
+err_pci_runtime_suspend:
+	while (--i >= 0)
+		sdhci_runtime_resume_host(chip->slots[i]->host, 0);
+	return ret;
+}
+
+static int sdhci_pci_runtime_resume_host(struct sdhci_pci_chip *chip)
+{
+	struct sdhci_pci_slot *slot;
+	int i, ret;
+
+	for (i = 0; i < chip->num_slots; i++) {
+		slot = chip->slots[i];
+		if (!slot)
+			continue;
+
+		ret = sdhci_runtime_resume_host(slot->host, 0);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int sdhci_cqhci_runtime_suspend(struct sdhci_pci_chip *chip)
+{
+	int ret;
+
+	ret = cqhci_suspend(chip->slots[0]->host->mmc);
+	if (ret)
+		return ret;
+
+	return sdhci_pci_runtime_suspend_host(chip);
+}
+
+static int sdhci_cqhci_runtime_resume(struct sdhci_pci_chip *chip)
+{
+	int ret;
+
+	ret = sdhci_pci_runtime_resume_host(chip);
+	if (ret)
+		return ret;
+
+	return cqhci_resume(chip->slots[0]->host->mmc);
+}
+#endif
+
+static u32 sdhci_cqhci_irq(struct sdhci_host *host, u32 intmask)
+{
+	int cmd_error = 0;
+	int data_error = 0;
+
+	if (!sdhci_cqe_irq(host, intmask, &cmd_error, &data_error))
+		return intmask;
+
+	cqhci_irq(host->mmc, intmask, cmd_error, data_error);
+
+	return 0;
+}
+
+static void sdhci_pci_dumpregs(struct mmc_host *mmc)
+{
+	sdhci_dumpregs(mmc_priv(mmc));
+}
+
+static void sdhci_cqhci_reset(struct sdhci_host *host, u8 mask)
+{
+	if ((host->mmc->caps2 & MMC_CAP2_CQE) && (mask & SDHCI_RESET_ALL) &&
+	    host->mmc->cqe_private)
+		cqhci_deactivate(host->mmc);
+	sdhci_reset(host, mask);
+}
+
+/*****************************************************************************\
+ *                                                                           *
+ * Hardware specific quirk handling                                          *
+ *                                                                           *
+\*****************************************************************************/
+
+static int ricoh_probe(struct sdhci_pci_chip *chip)
+{
+	if (chip->pdev->subsystem_vendor == PCI_VENDOR_ID_SAMSUNG ||
+	    chip->pdev->subsystem_vendor == PCI_VENDOR_ID_SONY)
+		chip->quirks |= SDHCI_QUIRK_NO_CARD_NO_RESET;
+	return 0;
+}
+
+static int ricoh_mmc_probe_slot(struct sdhci_pci_slot *slot)
+{
+	slot->host->caps =
+		FIELD_PREP(SDHCI_TIMEOUT_CLK_MASK, 0x21) |
+		FIELD_PREP(SDHCI_CLOCK_BASE_MASK, 0x21) |
+		SDHCI_TIMEOUT_CLK_UNIT |
+		SDHCI_CAN_VDD_330 |
+		SDHCI_CAN_DO_HISPD |
+		SDHCI_CAN_DO_SDMA;
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int ricoh_mmc_resume(struct sdhci_pci_chip *chip)
+{
+	/* Apply a delay to allow controller to settle */
+	/* Otherwise it becomes confused if card state changed
+		during suspend */
+	msleep(500);
+	return sdhci_pci_resume_host(chip);
+}
+#endif
+
+static const struct sdhci_pci_fixes sdhci_ricoh = {
+	.probe		= ricoh_probe,
+	.quirks		= SDHCI_QUIRK_32BIT_DMA_ADDR |
+			  SDHCI_QUIRK_FORCE_DMA |
+			  SDHCI_QUIRK_CLOCK_BEFORE_RESET,
+};
+
+static const struct sdhci_pci_fixes sdhci_ricoh_mmc = {
+	.probe_slot	= ricoh_mmc_probe_slot,
+#ifdef CONFIG_PM_SLEEP
+	.resume		= ricoh_mmc_resume,
+#endif
+	.quirks		= SDHCI_QUIRK_32BIT_DMA_ADDR |
+			  SDHCI_QUIRK_CLOCK_BEFORE_RESET |
+			  SDHCI_QUIRK_NO_CARD_NO_RESET |
+			  SDHCI_QUIRK_MISSING_CAPS
+};
+
+static const struct sdhci_pci_fixes sdhci_ene_712 = {
+	.quirks		= SDHCI_QUIRK_SINGLE_POWER_WRITE |
+			  SDHCI_QUIRK_BROKEN_DMA,
+};
+
+static const struct sdhci_pci_fixes sdhci_ene_714 = {
+	.quirks		= SDHCI_QUIRK_SINGLE_POWER_WRITE |
+			  SDHCI_QUIRK_RESET_CMD_DATA_ON_IOS |
+			  SDHCI_QUIRK_BROKEN_DMA,
+};
+
+static const struct sdhci_pci_fixes sdhci_cafe = {
+	.quirks		= SDHCI_QUIRK_NO_SIMULT_VDD_AND_POWER |
+			  SDHCI_QUIRK_NO_BUSY_IRQ |
+			  SDHCI_QUIRK_BROKEN_CARD_DETECTION |
+			  SDHCI_QUIRK_BROKEN_TIMEOUT_VAL,
+};
+
+static const struct sdhci_pci_fixes sdhci_intel_qrk = {
+	.quirks		= SDHCI_QUIRK_NO_HISPD_BIT,
+};
+
+static int mrst_hc_probe_slot(struct sdhci_pci_slot *slot)
+{
+	slot->host->mmc->caps |= MMC_CAP_8_BIT_DATA;
+	return 0;
+}
+
+/*
+ * ADMA operation is disabled for Moorestown platform due to
+ * hardware bugs.
+ */
+static int mrst_hc_probe(struct sdhci_pci_chip *chip)
+{
+	/*
+	 * slots number is fixed here for MRST as SDIO3/5 are never used and
+	 * have hardware bugs.
+	 */
+	chip->num_slots = 1;
+	return 0;
+}
+
+static int pch_hc_probe_slot(struct sdhci_pci_slot *slot)
+{
+	slot->host->mmc->caps |= MMC_CAP_8_BIT_DATA;
+	return 0;
+}
+
+#ifdef CONFIG_PM
+
+static irqreturn_t sdhci_pci_sd_cd(int irq, void *dev_id)
+{
+	struct sdhci_pci_slot *slot = dev_id;
+	struct sdhci_host *host = slot->host;
+
+	mmc_detect_change(host->mmc, msecs_to_jiffies(200));
+	return IRQ_HANDLED;
+}
+
+static void sdhci_pci_add_own_cd(struct sdhci_pci_slot *slot)
+{
+	int err, irq, gpio = slot->cd_gpio;
+
+	slot->cd_gpio = -EINVAL;
+	slot->cd_irq = -EINVAL;
+
+	if (!gpio_is_valid(gpio))
+		return;
+
+	err = devm_gpio_request(&slot->chip->pdev->dev, gpio, "sd_cd");
+	if (err < 0)
+		goto out;
+
+	err = gpio_direction_input(gpio);
+	if (err < 0)
+		goto out_free;
+
+	irq = gpio_to_irq(gpio);
+	if (irq < 0)
+		goto out_free;
+
+	err = request_irq(irq, sdhci_pci_sd_cd, IRQF_TRIGGER_RISING |
+			  IRQF_TRIGGER_FALLING, "sd_cd", slot);
+	if (err)
+		goto out_free;
+
+	slot->cd_gpio = gpio;
+	slot->cd_irq = irq;
+
+	return;
+
+out_free:
+	devm_gpio_free(&slot->chip->pdev->dev, gpio);
+out:
+	dev_warn(&slot->chip->pdev->dev, "failed to setup card detect wake up\n");
+}
+
+static void sdhci_pci_remove_own_cd(struct sdhci_pci_slot *slot)
+{
+	if (slot->cd_irq >= 0)
+		free_irq(slot->cd_irq, slot);
+}
+
+#else
+
+static inline void sdhci_pci_add_own_cd(struct sdhci_pci_slot *slot)
+{
+}
+
+static inline void sdhci_pci_remove_own_cd(struct sdhci_pci_slot *slot)
+{
+}
+
+#endif
+
+static int mfd_emmc_probe_slot(struct sdhci_pci_slot *slot)
+{
+	slot->host->mmc->caps |= MMC_CAP_8_BIT_DATA | MMC_CAP_NONREMOVABLE;
+	slot->host->mmc->caps2 |= MMC_CAP2_BOOTPART_NOACC;
+	return 0;
+}
+
+static int mfd_sdio_probe_slot(struct sdhci_pci_slot *slot)
+{
+	slot->host->mmc->caps |= MMC_CAP_POWER_OFF_CARD | MMC_CAP_NONREMOVABLE;
+	return 0;
+}
+
+static const struct sdhci_pci_fixes sdhci_intel_mrst_hc0 = {
+	.quirks		= SDHCI_QUIRK_BROKEN_ADMA | SDHCI_QUIRK_NO_HISPD_BIT,
+	.probe_slot	= mrst_hc_probe_slot,
+};
+
+static const struct sdhci_pci_fixes sdhci_intel_mrst_hc1_hc2 = {
+	.quirks		= SDHCI_QUIRK_BROKEN_ADMA | SDHCI_QUIRK_NO_HISPD_BIT,
+	.probe		= mrst_hc_probe,
+};
+
+static const struct sdhci_pci_fixes sdhci_intel_mfd_sd = {
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.allow_runtime_pm = true,
+	.own_cd_for_runtime_pm = true,
+};
+
+static const struct sdhci_pci_fixes sdhci_intel_mfd_sdio = {
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.quirks2	= SDHCI_QUIRK2_HOST_OFF_CARD_ON,
+	.allow_runtime_pm = true,
+	.probe_slot	= mfd_sdio_probe_slot,
+};
+
+static const struct sdhci_pci_fixes sdhci_intel_mfd_emmc = {
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.allow_runtime_pm = true,
+	.probe_slot	= mfd_emmc_probe_slot,
+};
+
+static const struct sdhci_pci_fixes sdhci_intel_pch_sdio = {
+	.quirks		= SDHCI_QUIRK_BROKEN_ADMA,
+	.probe_slot	= pch_hc_probe_slot,
+};
+
+#ifdef CONFIG_X86
+
+#define BYT_IOSF_SCCEP			0x63
+#define BYT_IOSF_OCP_NETCTRL0		0x1078
+#define BYT_IOSF_OCP_TIMEOUT_BASE	GENMASK(10, 8)
+
+static void byt_ocp_setting(struct pci_dev *pdev)
+{
+	u32 val = 0;
+
+	if (pdev->device != PCI_DEVICE_ID_INTEL_BYT_EMMC &&
+	    pdev->device != PCI_DEVICE_ID_INTEL_BYT_SDIO &&
+	    pdev->device != PCI_DEVICE_ID_INTEL_BYT_SD &&
+	    pdev->device != PCI_DEVICE_ID_INTEL_BYT_EMMC2)
+		return;
+
+	if (iosf_mbi_read(BYT_IOSF_SCCEP, MBI_CR_READ, BYT_IOSF_OCP_NETCTRL0,
+			  &val)) {
+		dev_err(&pdev->dev, "%s read error\n", __func__);
+		return;
+	}
+
+	if (!(val & BYT_IOSF_OCP_TIMEOUT_BASE))
+		return;
+
+	val &= ~BYT_IOSF_OCP_TIMEOUT_BASE;
+
+	if (iosf_mbi_write(BYT_IOSF_SCCEP, MBI_CR_WRITE, BYT_IOSF_OCP_NETCTRL0,
+			   val)) {
+		dev_err(&pdev->dev, "%s write error\n", __func__);
+		return;
+	}
+
+	dev_dbg(&pdev->dev, "%s completed\n", __func__);
+}
+
+#else
+
+static inline void byt_ocp_setting(struct pci_dev *pdev)
+{
+}
+
+#endif
+
+enum {
+	INTEL_DSM_FNS		=  0,
+	INTEL_DSM_V18_SWITCH	=  3,
+	INTEL_DSM_V33_SWITCH	=  4,
+	INTEL_DSM_DRV_STRENGTH	=  9,
+	INTEL_DSM_D3_RETUNE	= 10,
+};
+
+struct intel_host {
+	u32	dsm_fns;
+	int	drv_strength;
+	bool	d3_retune;
+	bool	rpm_retune_ok;
+	bool	needs_pwr_off;
+	u32	glk_rx_ctrl1;
+	u32	glk_tun_val;
+	u32	active_ltr;
+	u32	idle_ltr;
+};
+
+static const guid_t intel_dsm_guid =
+	GUID_INIT(0xF6C13EA5, 0x65CD, 0x461F,
+		  0xAB, 0x7A, 0x29, 0xF7, 0xE8, 0xD5, 0xBD, 0x61);
+
+static int __intel_dsm(struct intel_host *intel_host, struct device *dev,
+		       unsigned int fn, u32 *result)
+{
+	union acpi_object *obj;
+	int err = 0;
+	size_t len;
+
+	obj = acpi_evaluate_dsm(ACPI_HANDLE(dev), &intel_dsm_guid, 0, fn, NULL);
+	if (!obj)
+		return -EOPNOTSUPP;
+
+	if (obj->type != ACPI_TYPE_BUFFER || obj->buffer.length < 1) {
+		err = -EINVAL;
+		goto out;
+	}
+
+	len = min_t(size_t, obj->buffer.length, 4);
+
+	*result = 0;
+	memcpy(result, obj->buffer.pointer, len);
+out:
+	ACPI_FREE(obj);
+
+	return err;
+}
+
+static int intel_dsm(struct intel_host *intel_host, struct device *dev,
+		     unsigned int fn, u32 *result)
+{
+	if (fn > 31 || !(intel_host->dsm_fns & (1 << fn)))
+		return -EOPNOTSUPP;
+
+	return __intel_dsm(intel_host, dev, fn, result);
+}
+
+static void intel_dsm_init(struct intel_host *intel_host, struct device *dev,
+			   struct mmc_host *mmc)
+{
+	int err;
+	u32 val;
+
+	intel_host->d3_retune = true;
+
+	err = __intel_dsm(intel_host, dev, INTEL_DSM_FNS, &intel_host->dsm_fns);
+	if (err) {
+		pr_debug("%s: DSM not supported, error %d\n",
+			 mmc_hostname(mmc), err);
+		return;
+	}
+
+	pr_debug("%s: DSM function mask %#x\n",
+		 mmc_hostname(mmc), intel_host->dsm_fns);
+
+	err = intel_dsm(intel_host, dev, INTEL_DSM_DRV_STRENGTH, &val);
+	intel_host->drv_strength = err ? 0 : val;
+
+	err = intel_dsm(intel_host, dev, INTEL_DSM_D3_RETUNE, &val);
+	intel_host->d3_retune = err ? true : !!val;
+}
+
+static void sdhci_pci_int_hw_reset(struct sdhci_host *host)
+{
+	u8 reg;
+
+	reg = sdhci_readb(host, SDHCI_POWER_CONTROL);
+	reg |= 0x10;
+	sdhci_writeb(host, reg, SDHCI_POWER_CONTROL);
+	/* For eMMC, minimum is 1us but give it 9us for good measure */
+	udelay(9);
+	reg &= ~0x10;
+	sdhci_writeb(host, reg, SDHCI_POWER_CONTROL);
+	/* For eMMC, minimum is 200us but give it 300us for good measure */
+	usleep_range(300, 1000);
+}
+
+static int intel_select_drive_strength(struct mmc_card *card,
+				       unsigned int max_dtr, int host_drv,
+				       int card_drv, int *drv_type)
+{
+	struct sdhci_host *host = mmc_priv(card->host);
+	struct sdhci_pci_slot *slot = sdhci_priv(host);
+	struct intel_host *intel_host = sdhci_pci_priv(slot);
+
+	if (!(mmc_driver_type_mask(intel_host->drv_strength) & card_drv))
+		return 0;
+
+	return intel_host->drv_strength;
+}
+
+static int bxt_get_cd(struct mmc_host *mmc)
+{
+	int gpio_cd = mmc_gpio_get_cd(mmc);
+	struct sdhci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+	int ret = 0;
+
+	if (!gpio_cd)
+		return 0;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (host->flags & SDHCI_DEVICE_DEAD)
+		goto out;
+
+	ret = !!(sdhci_readl(host, SDHCI_PRESENT_STATE) & SDHCI_CARD_PRESENT);
+out:
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	return ret;
+}
+
+#define SDHCI_INTEL_PWR_TIMEOUT_CNT	20
+#define SDHCI_INTEL_PWR_TIMEOUT_UDELAY	100
+
+static void sdhci_intel_set_power(struct sdhci_host *host, unsigned char mode,
+				  unsigned short vdd)
+{
+	struct sdhci_pci_slot *slot = sdhci_priv(host);
+	struct intel_host *intel_host = sdhci_pci_priv(slot);
+	int cntr;
+	u8 reg;
+
+	/*
+	 * Bus power may control card power, but a full reset still may not
+	 * reset the power, whereas a direct write to SDHCI_POWER_CONTROL can.
+	 * That might be needed to initialize correctly, if the card was left
+	 * powered on previously.
+	 */
+	if (intel_host->needs_pwr_off) {
+		intel_host->needs_pwr_off = false;
+		if (mode != MMC_POWER_OFF) {
+			sdhci_writeb(host, 0, SDHCI_POWER_CONTROL);
+			usleep_range(10000, 12500);
+		}
+	}
+
+	sdhci_set_power(host, mode, vdd);
+
+	if (mode == MMC_POWER_OFF)
+		return;
+
+	/*
+	 * Bus power might not enable after D3 -> D0 transition due to the
+	 * present state not yet having propagated. Retry for up to 2ms.
+	 */
+	for (cntr = 0; cntr < SDHCI_INTEL_PWR_TIMEOUT_CNT; cntr++) {
+		reg = sdhci_readb(host, SDHCI_POWER_CONTROL);
+		if (reg & SDHCI_POWER_ON)
+			break;
+		udelay(SDHCI_INTEL_PWR_TIMEOUT_UDELAY);
+		reg |= SDHCI_POWER_ON;
+		sdhci_writeb(host, reg, SDHCI_POWER_CONTROL);
+	}
+}
+
+static void sdhci_intel_set_uhs_signaling(struct sdhci_host *host,
+					  unsigned int timing)
+{
+	/* Set UHS timing to SDR25 for High Speed mode */
+	if (timing == MMC_TIMING_MMC_HS || timing == MMC_TIMING_SD_HS)
+		timing = MMC_TIMING_UHS_SDR25;
+	sdhci_set_uhs_signaling(host, timing);
+}
+
+#define INTEL_HS400_ES_REG 0x78
+#define INTEL_HS400_ES_BIT BIT(0)
+
+static void intel_hs400_enhanced_strobe(struct mmc_host *mmc,
+					struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	u32 val;
+
+	val = sdhci_readl(host, INTEL_HS400_ES_REG);
+	if (ios->enhanced_strobe)
+		val |= INTEL_HS400_ES_BIT;
+	else
+		val &= ~INTEL_HS400_ES_BIT;
+	sdhci_writel(host, val, INTEL_HS400_ES_REG);
+}
+
+static int intel_start_signal_voltage_switch(struct mmc_host *mmc,
+					     struct mmc_ios *ios)
+{
+	struct device *dev = mmc_dev(mmc);
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pci_slot *slot = sdhci_priv(host);
+	struct intel_host *intel_host = sdhci_pci_priv(slot);
+	unsigned int fn;
+	u32 result = 0;
+	int err;
+
+	err = sdhci_start_signal_voltage_switch(mmc, ios);
+	if (err)
+		return err;
+
+	switch (ios->signal_voltage) {
+	case MMC_SIGNAL_VOLTAGE_330:
+		fn = INTEL_DSM_V33_SWITCH;
+		break;
+	case MMC_SIGNAL_VOLTAGE_180:
+		fn = INTEL_DSM_V18_SWITCH;
+		break;
+	default:
+		return 0;
+	}
+
+	err = intel_dsm(intel_host, dev, fn, &result);
+	pr_debug("%s: %s DSM fn %u error %d result %u\n",
+		 mmc_hostname(mmc), __func__, fn, err, result);
+
+	return 0;
+}
+
+static const struct sdhci_ops sdhci_intel_byt_ops = {
+	.set_clock		= sdhci_set_clock,
+	.set_power		= sdhci_intel_set_power,
+	.enable_dma		= sdhci_pci_enable_dma,
+	.set_bus_width		= sdhci_set_bus_width,
+	.reset			= sdhci_reset,
+	.set_uhs_signaling	= sdhci_intel_set_uhs_signaling,
+	.hw_reset		= sdhci_pci_hw_reset,
+};
+
+static const struct sdhci_ops sdhci_intel_glk_ops = {
+	.set_clock		= sdhci_set_clock,
+	.set_power		= sdhci_intel_set_power,
+	.enable_dma		= sdhci_pci_enable_dma,
+	.set_bus_width		= sdhci_set_bus_width,
+	.reset			= sdhci_cqhci_reset,
+	.set_uhs_signaling	= sdhci_intel_set_uhs_signaling,
+	.hw_reset		= sdhci_pci_hw_reset,
+	.irq			= sdhci_cqhci_irq,
+};
+
+static void byt_read_dsm(struct sdhci_pci_slot *slot)
+{
+	struct intel_host *intel_host = sdhci_pci_priv(slot);
+	struct device *dev = &slot->chip->pdev->dev;
+	struct mmc_host *mmc = slot->host->mmc;
+
+	intel_dsm_init(intel_host, dev, mmc);
+	slot->chip->rpm_retune = intel_host->d3_retune;
+}
+
+static int intel_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	int err = sdhci_execute_tuning(mmc, opcode);
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	if (err)
+		return err;
+
+	/*
+	 * Tuning can leave the IP in an active state (Buffer Read Enable bit
+	 * set) which prevents the entry to low power states (i.e. S0i3). Data
+	 * reset will clear it.
+	 */
+	sdhci_reset(host, SDHCI_RESET_DATA);
+
+	return 0;
+}
+
+#define INTEL_ACTIVELTR		0x804
+#define INTEL_IDLELTR		0x808
+
+#define INTEL_LTR_REQ		BIT(15)
+#define INTEL_LTR_SCALE_MASK	GENMASK(11, 10)
+#define INTEL_LTR_SCALE_1US	(2 << 10)
+#define INTEL_LTR_SCALE_32US	(3 << 10)
+#define INTEL_LTR_VALUE_MASK	GENMASK(9, 0)
+
+static void intel_cache_ltr(struct sdhci_pci_slot *slot)
+{
+	struct intel_host *intel_host = sdhci_pci_priv(slot);
+	struct sdhci_host *host = slot->host;
+
+	intel_host->active_ltr = readl(host->ioaddr + INTEL_ACTIVELTR);
+	intel_host->idle_ltr = readl(host->ioaddr + INTEL_IDLELTR);
+}
+
+static void intel_ltr_set(struct device *dev, s32 val)
+{
+	struct sdhci_pci_chip *chip = dev_get_drvdata(dev);
+	struct sdhci_pci_slot *slot = chip->slots[0];
+	struct intel_host *intel_host = sdhci_pci_priv(slot);
+	struct sdhci_host *host = slot->host;
+	u32 ltr;
+
+	pm_runtime_get_sync(dev);
+
+	/*
+	 * Program latency tolerance (LTR) accordingly what has been asked
+	 * by the PM QoS layer or disable it in case we were passed
+	 * negative value or PM_QOS_LATENCY_ANY.
+	 */
+	ltr = readl(host->ioaddr + INTEL_ACTIVELTR);
+
+	if (val == PM_QOS_LATENCY_ANY || val < 0) {
+		ltr &= ~INTEL_LTR_REQ;
+	} else {
+		ltr |= INTEL_LTR_REQ;
+		ltr &= ~INTEL_LTR_SCALE_MASK;
+		ltr &= ~INTEL_LTR_VALUE_MASK;
+
+		if (val > INTEL_LTR_VALUE_MASK) {
+			val >>= 5;
+			if (val > INTEL_LTR_VALUE_MASK)
+				val = INTEL_LTR_VALUE_MASK;
+			ltr |= INTEL_LTR_SCALE_32US | val;
+		} else {
+			ltr |= INTEL_LTR_SCALE_1US | val;
+		}
+	}
+
+	if (ltr == intel_host->active_ltr)
+		goto out;
+
+	writel(ltr, host->ioaddr + INTEL_ACTIVELTR);
+	writel(ltr, host->ioaddr + INTEL_IDLELTR);
+
+	/* Cache the values into lpss structure */
+	intel_cache_ltr(slot);
+out:
+	pm_runtime_put_autosuspend(dev);
+}
+
+static bool intel_use_ltr(struct sdhci_pci_chip *chip)
+{
+	switch (chip->pdev->device) {
+	case PCI_DEVICE_ID_INTEL_BYT_EMMC:
+	case PCI_DEVICE_ID_INTEL_BYT_EMMC2:
+	case PCI_DEVICE_ID_INTEL_BYT_SDIO:
+	case PCI_DEVICE_ID_INTEL_BYT_SD:
+	case PCI_DEVICE_ID_INTEL_BSW_EMMC:
+	case PCI_DEVICE_ID_INTEL_BSW_SDIO:
+	case PCI_DEVICE_ID_INTEL_BSW_SD:
+		return false;
+	default:
+		return true;
+	}
+}
+
+static void intel_ltr_expose(struct sdhci_pci_chip *chip)
+{
+	struct device *dev = &chip->pdev->dev;
+
+	if (!intel_use_ltr(chip))
+		return;
+
+	dev->power.set_latency_tolerance = intel_ltr_set;
+	dev_pm_qos_expose_latency_tolerance(dev);
+}
+
+static void intel_ltr_hide(struct sdhci_pci_chip *chip)
+{
+	struct device *dev = &chip->pdev->dev;
+
+	if (!intel_use_ltr(chip))
+		return;
+
+	dev_pm_qos_hide_latency_tolerance(dev);
+	dev->power.set_latency_tolerance = NULL;
+}
+
+static void byt_probe_slot(struct sdhci_pci_slot *slot)
+{
+	struct mmc_host_ops *ops = &slot->host->mmc_host_ops;
+	struct device *dev = &slot->chip->pdev->dev;
+	struct mmc_host *mmc = slot->host->mmc;
+
+	byt_read_dsm(slot);
+
+	byt_ocp_setting(slot->chip->pdev);
+
+	ops->execute_tuning = intel_execute_tuning;
+	ops->start_signal_voltage_switch = intel_start_signal_voltage_switch;
+
+	device_property_read_u32(dev, "max-frequency", &mmc->f_max);
+
+	if (!mmc->slotno) {
+		slot->chip->slots[mmc->slotno] = slot;
+		intel_ltr_expose(slot->chip);
+	}
+}
+
+static void byt_add_debugfs(struct sdhci_pci_slot *slot)
+{
+	struct intel_host *intel_host = sdhci_pci_priv(slot);
+	struct mmc_host *mmc = slot->host->mmc;
+	struct dentry *dir = mmc->debugfs_root;
+
+	if (!intel_use_ltr(slot->chip))
+		return;
+
+	debugfs_create_x32("active_ltr", 0444, dir, &intel_host->active_ltr);
+	debugfs_create_x32("idle_ltr", 0444, dir, &intel_host->idle_ltr);
+
+	intel_cache_ltr(slot);
+}
+
+static int byt_add_host(struct sdhci_pci_slot *slot)
+{
+	int ret = sdhci_add_host(slot->host);
+
+	if (!ret)
+		byt_add_debugfs(slot);
+	return ret;
+}
+
+static void byt_remove_slot(struct sdhci_pci_slot *slot, int dead)
+{
+	struct mmc_host *mmc = slot->host->mmc;
+
+	if (!mmc->slotno)
+		intel_ltr_hide(slot->chip);
+}
+
+static int byt_emmc_probe_slot(struct sdhci_pci_slot *slot)
+{
+	byt_probe_slot(slot);
+	slot->host->mmc->caps |= MMC_CAP_8_BIT_DATA | MMC_CAP_NONREMOVABLE |
+				 MMC_CAP_HW_RESET | MMC_CAP_1_8V_DDR |
+				 MMC_CAP_CMD_DURING_TFR |
+				 MMC_CAP_WAIT_WHILE_BUSY;
+	slot->hw_reset = sdhci_pci_int_hw_reset;
+	if (slot->chip->pdev->device == PCI_DEVICE_ID_INTEL_BSW_EMMC)
+		slot->host->timeout_clk = 1000; /* 1000 kHz i.e. 1 MHz */
+	slot->host->mmc_host_ops.select_drive_strength =
+						intel_select_drive_strength;
+	return 0;
+}
+
+static bool glk_broken_cqhci(struct sdhci_pci_slot *slot)
+{
+	return slot->chip->pdev->device == PCI_DEVICE_ID_INTEL_GLK_EMMC &&
+	       (dmi_match(DMI_BIOS_VENDOR, "LENOVO") ||
+		dmi_match(DMI_SYS_VENDOR, "IRBIS"));
+}
+
+static bool jsl_broken_hs400es(struct sdhci_pci_slot *slot)
+{
+	return slot->chip->pdev->device == PCI_DEVICE_ID_INTEL_JSL_EMMC &&
+			dmi_match(DMI_BIOS_VENDOR, "ASUSTeK COMPUTER INC.");
+}
+
+static int glk_emmc_probe_slot(struct sdhci_pci_slot *slot)
+{
+	int ret = byt_emmc_probe_slot(slot);
+
+	if (!glk_broken_cqhci(slot))
+		slot->host->mmc->caps2 |= MMC_CAP2_CQE;
+
+	if (slot->chip->pdev->device != PCI_DEVICE_ID_INTEL_GLK_EMMC) {
+		if (!jsl_broken_hs400es(slot)) {
+			slot->host->mmc->caps2 |= MMC_CAP2_HS400_ES;
+			slot->host->mmc_host_ops.hs400_enhanced_strobe =
+							intel_hs400_enhanced_strobe;
+		}
+		slot->host->mmc->caps2 |= MMC_CAP2_CQE_DCMD;
+	}
+
+	return ret;
+}
+
+static const struct cqhci_host_ops glk_cqhci_ops = {
+	.enable		= sdhci_cqe_enable,
+	.disable	= sdhci_cqe_disable,
+	.dumpregs	= sdhci_pci_dumpregs,
+};
+
+static int glk_emmc_add_host(struct sdhci_pci_slot *slot)
+{
+	struct device *dev = &slot->chip->pdev->dev;
+	struct sdhci_host *host = slot->host;
+	struct cqhci_host *cq_host;
+	bool dma64;
+	int ret;
+
+	ret = sdhci_setup_host(host);
+	if (ret)
+		return ret;
+
+	cq_host = devm_kzalloc(dev, sizeof(*cq_host), GFP_KERNEL);
+	if (!cq_host) {
+		ret = -ENOMEM;
+		goto cleanup;
+	}
+
+	cq_host->mmio = host->ioaddr + 0x200;
+	cq_host->quirks |= CQHCI_QUIRK_SHORT_TXFR_DESC_SZ;
+	cq_host->ops = &glk_cqhci_ops;
+
+	dma64 = host->flags & SDHCI_USE_64_BIT_DMA;
+	if (dma64)
+		cq_host->caps |= CQHCI_TASK_DESC_SZ_128;
+
+	ret = cqhci_init(cq_host, host->mmc, dma64);
+	if (ret)
+		goto cleanup;
+
+	ret = __sdhci_add_host(host);
+	if (ret)
+		goto cleanup;
+
+	byt_add_debugfs(slot);
+
+	return 0;
+
+cleanup:
+	sdhci_cleanup_host(host);
+	return ret;
+}
+
+#ifdef CONFIG_PM
+#define GLK_RX_CTRL1	0x834
+#define GLK_TUN_VAL	0x840
+#define GLK_PATH_PLL	GENMASK(13, 8)
+#define GLK_DLY		GENMASK(6, 0)
+/* Workaround firmware failing to restore the tuning value */
+static void glk_rpm_retune_wa(struct sdhci_pci_chip *chip, bool susp)
+{
+	struct sdhci_pci_slot *slot = chip->slots[0];
+	struct intel_host *intel_host = sdhci_pci_priv(slot);
+	struct sdhci_host *host = slot->host;
+	u32 glk_rx_ctrl1;
+	u32 glk_tun_val;
+	u32 dly;
+
+	if (intel_host->rpm_retune_ok || !mmc_can_retune(host->mmc))
+		return;
+
+	glk_rx_ctrl1 = sdhci_readl(host, GLK_RX_CTRL1);
+	glk_tun_val = sdhci_readl(host, GLK_TUN_VAL);
+
+	if (susp) {
+		intel_host->glk_rx_ctrl1 = glk_rx_ctrl1;
+		intel_host->glk_tun_val = glk_tun_val;
+		return;
+	}
+
+	if (!intel_host->glk_tun_val)
+		return;
+
+	if (glk_rx_ctrl1 != intel_host->glk_rx_ctrl1) {
+		intel_host->rpm_retune_ok = true;
+		return;
+	}
+
+	dly = FIELD_PREP(GLK_DLY, FIELD_GET(GLK_PATH_PLL, glk_rx_ctrl1) +
+				  (intel_host->glk_tun_val << 1));
+	if (dly == FIELD_GET(GLK_DLY, glk_rx_ctrl1))
+		return;
+
+	glk_rx_ctrl1 = (glk_rx_ctrl1 & ~GLK_DLY) | dly;
+	sdhci_writel(host, glk_rx_ctrl1, GLK_RX_CTRL1);
+
+	intel_host->rpm_retune_ok = true;
+	chip->rpm_retune = true;
+	mmc_retune_needed(host->mmc);
+	pr_info("%s: Requiring re-tune after rpm resume", mmc_hostname(host->mmc));
+}
+
+static void glk_rpm_retune_chk(struct sdhci_pci_chip *chip, bool susp)
+{
+	if (chip->pdev->device == PCI_DEVICE_ID_INTEL_GLK_EMMC &&
+	    !chip->rpm_retune)
+		glk_rpm_retune_wa(chip, susp);
+}
+
+static int glk_runtime_suspend(struct sdhci_pci_chip *chip)
+{
+	glk_rpm_retune_chk(chip, true);
+
+	return sdhci_cqhci_runtime_suspend(chip);
+}
+
+static int glk_runtime_resume(struct sdhci_pci_chip *chip)
+{
+	glk_rpm_retune_chk(chip, false);
+
+	return sdhci_cqhci_runtime_resume(chip);
+}
+#endif
+
+#ifdef CONFIG_ACPI
+static int ni_set_max_freq(struct sdhci_pci_slot *slot)
+{
+	acpi_status status;
+	unsigned long long max_freq;
+
+	status = acpi_evaluate_integer(ACPI_HANDLE(&slot->chip->pdev->dev),
+				       "MXFQ", NULL, &max_freq);
+	if (ACPI_FAILURE(status)) {
+		dev_err(&slot->chip->pdev->dev,
+			"MXFQ not found in acpi table\n");
+		return -EINVAL;
+	}
+
+	slot->host->mmc->f_max = max_freq * 1000000;
+
+	return 0;
+}
+#else
+static inline int ni_set_max_freq(struct sdhci_pci_slot *slot)
+{
+	return 0;
+}
+#endif
+
+static int ni_byt_sdio_probe_slot(struct sdhci_pci_slot *slot)
+{
+	int err;
+
+	byt_probe_slot(slot);
+
+	err = ni_set_max_freq(slot);
+	if (err)
+		return err;
+
+	slot->host->mmc->caps |= MMC_CAP_POWER_OFF_CARD | MMC_CAP_NONREMOVABLE |
+				 MMC_CAP_WAIT_WHILE_BUSY;
+	return 0;
+}
+
+static int byt_sdio_probe_slot(struct sdhci_pci_slot *slot)
+{
+	byt_probe_slot(slot);
+	slot->host->mmc->caps |= MMC_CAP_POWER_OFF_CARD | MMC_CAP_NONREMOVABLE |
+				 MMC_CAP_WAIT_WHILE_BUSY;
+	return 0;
+}
+
+static void byt_needs_pwr_off(struct sdhci_pci_slot *slot)
+{
+	struct intel_host *intel_host = sdhci_pci_priv(slot);
+	u8 reg = sdhci_readb(slot->host, SDHCI_POWER_CONTROL);
+
+	intel_host->needs_pwr_off = reg  & SDHCI_POWER_ON;
+}
+
+static int byt_sd_probe_slot(struct sdhci_pci_slot *slot)
+{
+	byt_probe_slot(slot);
+	slot->host->mmc->caps |= MMC_CAP_WAIT_WHILE_BUSY |
+				 MMC_CAP_AGGRESSIVE_PM | MMC_CAP_CD_WAKE;
+	slot->cd_idx = 0;
+	slot->cd_override_level = true;
+	if (slot->chip->pdev->device == PCI_DEVICE_ID_INTEL_BXT_SD ||
+	    slot->chip->pdev->device == PCI_DEVICE_ID_INTEL_BXTM_SD ||
+	    slot->chip->pdev->device == PCI_DEVICE_ID_INTEL_APL_SD ||
+	    slot->chip->pdev->device == PCI_DEVICE_ID_INTEL_GLK_SD)
+		slot->host->mmc_host_ops.get_cd = bxt_get_cd;
+
+	if (slot->chip->pdev->subsystem_vendor == PCI_VENDOR_ID_NI &&
+	    slot->chip->pdev->subsystem_device == PCI_SUBDEVICE_ID_NI_78E3)
+		slot->host->mmc->caps2 |= MMC_CAP2_AVOID_3_3V;
+
+	byt_needs_pwr_off(slot);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+
+static int byt_resume(struct sdhci_pci_chip *chip)
+{
+	byt_ocp_setting(chip->pdev);
+
+	return sdhci_pci_resume_host(chip);
+}
+
+#endif
+
+#ifdef CONFIG_PM
+
+static int byt_runtime_resume(struct sdhci_pci_chip *chip)
+{
+	byt_ocp_setting(chip->pdev);
+
+	return sdhci_pci_runtime_resume_host(chip);
+}
+
+#endif
+
+static const struct sdhci_pci_fixes sdhci_intel_byt_emmc = {
+#ifdef CONFIG_PM_SLEEP
+	.resume		= byt_resume,
+#endif
+#ifdef CONFIG_PM
+	.runtime_resume	= byt_runtime_resume,
+#endif
+	.allow_runtime_pm = true,
+	.probe_slot	= byt_emmc_probe_slot,
+	.add_host	= byt_add_host,
+	.remove_slot	= byt_remove_slot,
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+			  SDHCI_QUIRK_NO_LED,
+	.quirks2	= SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+			  SDHCI_QUIRK2_CAPS_BIT63_FOR_HS400 |
+			  SDHCI_QUIRK2_STOP_WITH_TC,
+	.ops		= &sdhci_intel_byt_ops,
+	.priv_size	= sizeof(struct intel_host),
+};
+
+static const struct sdhci_pci_fixes sdhci_intel_glk_emmc = {
+	.allow_runtime_pm	= true,
+	.probe_slot		= glk_emmc_probe_slot,
+	.add_host		= glk_emmc_add_host,
+	.remove_slot		= byt_remove_slot,
+#ifdef CONFIG_PM_SLEEP
+	.suspend		= sdhci_cqhci_suspend,
+	.resume			= sdhci_cqhci_resume,
+#endif
+#ifdef CONFIG_PM
+	.runtime_suspend	= glk_runtime_suspend,
+	.runtime_resume		= glk_runtime_resume,
+#endif
+	.quirks			= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+				  SDHCI_QUIRK_NO_LED,
+	.quirks2		= SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+				  SDHCI_QUIRK2_CAPS_BIT63_FOR_HS400 |
+				  SDHCI_QUIRK2_STOP_WITH_TC,
+	.ops			= &sdhci_intel_glk_ops,
+	.priv_size		= sizeof(struct intel_host),
+};
+
+static const struct sdhci_pci_fixes sdhci_ni_byt_sdio = {
+#ifdef CONFIG_PM_SLEEP
+	.resume		= byt_resume,
+#endif
+#ifdef CONFIG_PM
+	.runtime_resume	= byt_runtime_resume,
+#endif
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+			  SDHCI_QUIRK_NO_LED,
+	.quirks2	= SDHCI_QUIRK2_HOST_OFF_CARD_ON |
+			  SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+	.allow_runtime_pm = true,
+	.probe_slot	= ni_byt_sdio_probe_slot,
+	.add_host	= byt_add_host,
+	.remove_slot	= byt_remove_slot,
+	.ops		= &sdhci_intel_byt_ops,
+	.priv_size	= sizeof(struct intel_host),
+};
+
+static const struct sdhci_pci_fixes sdhci_intel_byt_sdio = {
+#ifdef CONFIG_PM_SLEEP
+	.resume		= byt_resume,
+#endif
+#ifdef CONFIG_PM
+	.runtime_resume	= byt_runtime_resume,
+#endif
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+			  SDHCI_QUIRK_NO_LED,
+	.quirks2	= SDHCI_QUIRK2_HOST_OFF_CARD_ON |
+			SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+	.allow_runtime_pm = true,
+	.probe_slot	= byt_sdio_probe_slot,
+	.add_host	= byt_add_host,
+	.remove_slot	= byt_remove_slot,
+	.ops		= &sdhci_intel_byt_ops,
+	.priv_size	= sizeof(struct intel_host),
+};
+
+static const struct sdhci_pci_fixes sdhci_intel_byt_sd = {
+#ifdef CONFIG_PM_SLEEP
+	.resume		= byt_resume,
+#endif
+#ifdef CONFIG_PM
+	.runtime_resume	= byt_runtime_resume,
+#endif
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+			  SDHCI_QUIRK_NO_LED,
+	.quirks2	= SDHCI_QUIRK2_CARD_ON_NEEDS_BUS_ON |
+			  SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+			  SDHCI_QUIRK2_STOP_WITH_TC,
+	.allow_runtime_pm = true,
+	.own_cd_for_runtime_pm = true,
+	.probe_slot	= byt_sd_probe_slot,
+	.add_host	= byt_add_host,
+	.remove_slot	= byt_remove_slot,
+	.ops		= &sdhci_intel_byt_ops,
+	.priv_size	= sizeof(struct intel_host),
+};
+
+/* Define Host controllers for Intel Merrifield platform */
+#define INTEL_MRFLD_EMMC_0	0
+#define INTEL_MRFLD_EMMC_1	1
+#define INTEL_MRFLD_SD		2
+#define INTEL_MRFLD_SDIO	3
+
+#ifdef CONFIG_ACPI
+static void intel_mrfld_mmc_fix_up_power_slot(struct sdhci_pci_slot *slot)
+{
+	struct acpi_device *device, *child;
+
+	device = ACPI_COMPANION(&slot->chip->pdev->dev);
+	if (!device)
+		return;
+
+	acpi_device_fix_up_power(device);
+	list_for_each_entry(child, &device->children, node)
+		if (child->status.present && child->status.enabled)
+			acpi_device_fix_up_power(child);
+}
+#else
+static inline void intel_mrfld_mmc_fix_up_power_slot(struct sdhci_pci_slot *slot) {}
+#endif
+
+static int intel_mrfld_mmc_probe_slot(struct sdhci_pci_slot *slot)
+{
+	unsigned int func = PCI_FUNC(slot->chip->pdev->devfn);
+
+	switch (func) {
+	case INTEL_MRFLD_EMMC_0:
+	case INTEL_MRFLD_EMMC_1:
+		slot->host->mmc->caps |= MMC_CAP_NONREMOVABLE |
+					 MMC_CAP_8_BIT_DATA |
+					 MMC_CAP_1_8V_DDR;
+		break;
+	case INTEL_MRFLD_SD:
+		slot->host->quirks2 |= SDHCI_QUIRK2_NO_1_8_V;
+		break;
+	case INTEL_MRFLD_SDIO:
+		/* Advertise 2.0v for compatibility with the SDIO card's OCR */
+		slot->host->ocr_mask = MMC_VDD_20_21 | MMC_VDD_165_195;
+		slot->host->mmc->caps |= MMC_CAP_NONREMOVABLE |
+					 MMC_CAP_POWER_OFF_CARD;
+		break;
+	default:
+		return -ENODEV;
+	}
+
+	intel_mrfld_mmc_fix_up_power_slot(slot);
+	return 0;
+}
+
+static const struct sdhci_pci_fixes sdhci_intel_mrfld_mmc = {
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.quirks2	= SDHCI_QUIRK2_BROKEN_HS200 |
+			SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+	.allow_runtime_pm = true,
+	.probe_slot	= intel_mrfld_mmc_probe_slot,
+};
+
+static int jmicron_pmos(struct sdhci_pci_chip *chip, int on)
+{
+	u8 scratch;
+	int ret;
+
+	ret = pci_read_config_byte(chip->pdev, 0xAE, &scratch);
+	if (ret)
+		return ret;
+
+	/*
+	 * Turn PMOS on [bit 0], set over current detection to 2.4 V
+	 * [bit 1:2] and enable over current debouncing [bit 6].
+	 */
+	if (on)
+		scratch |= 0x47;
+	else
+		scratch &= ~0x47;
+
+	return pci_write_config_byte(chip->pdev, 0xAE, scratch);
+}
+
+static int jmicron_probe(struct sdhci_pci_chip *chip)
+{
+	int ret;
+	u16 mmcdev = 0;
+
+	if (chip->pdev->revision == 0) {
+		chip->quirks |= SDHCI_QUIRK_32BIT_DMA_ADDR |
+			  SDHCI_QUIRK_32BIT_DMA_SIZE |
+			  SDHCI_QUIRK_32BIT_ADMA_SIZE |
+			  SDHCI_QUIRK_RESET_AFTER_REQUEST |
+			  SDHCI_QUIRK_BROKEN_SMALL_PIO;
+	}
+
+	/*
+	 * JMicron chips can have two interfaces to the same hardware
+	 * in order to work around limitations in Microsoft's driver.
+	 * We need to make sure we only bind to one of them.
+	 *
+	 * This code assumes two things:
+	 *
+	 * 1. The PCI code adds subfunctions in order.
+	 *
+	 * 2. The MMC interface has a lower subfunction number
+	 *    than the SD interface.
+	 */
+	if (chip->pdev->device == PCI_DEVICE_ID_JMICRON_JMB38X_SD)
+		mmcdev = PCI_DEVICE_ID_JMICRON_JMB38X_MMC;
+	else if (chip->pdev->device == PCI_DEVICE_ID_JMICRON_JMB388_SD)
+		mmcdev = PCI_DEVICE_ID_JMICRON_JMB388_ESD;
+
+	if (mmcdev) {
+		struct pci_dev *sd_dev;
+
+		sd_dev = NULL;
+		while ((sd_dev = pci_get_device(PCI_VENDOR_ID_JMICRON,
+						mmcdev, sd_dev)) != NULL) {
+			if ((PCI_SLOT(chip->pdev->devfn) ==
+				PCI_SLOT(sd_dev->devfn)) &&
+				(chip->pdev->bus == sd_dev->bus))
+				break;
+		}
+
+		if (sd_dev) {
+			pci_dev_put(sd_dev);
+			dev_info(&chip->pdev->dev, "Refusing to bind to "
+				"secondary interface.\n");
+			return -ENODEV;
+		}
+	}
+
+	/*
+	 * JMicron chips need a bit of a nudge to enable the power
+	 * output pins.
+	 */
+	ret = jmicron_pmos(chip, 1);
+	if (ret) {
+		dev_err(&chip->pdev->dev, "Failure enabling card power\n");
+		return ret;
+	}
+
+	/* quirk for unsable RO-detection on JM388 chips */
+	if (chip->pdev->device == PCI_DEVICE_ID_JMICRON_JMB388_SD ||
+	    chip->pdev->device == PCI_DEVICE_ID_JMICRON_JMB388_ESD)
+		chip->quirks |= SDHCI_QUIRK_UNSTABLE_RO_DETECT;
+
+	return 0;
+}
+
+static void jmicron_enable_mmc(struct sdhci_host *host, int on)
+{
+	u8 scratch;
+
+	scratch = readb(host->ioaddr + 0xC0);
+
+	if (on)
+		scratch |= 0x01;
+	else
+		scratch &= ~0x01;
+
+	writeb(scratch, host->ioaddr + 0xC0);
+}
+
+static int jmicron_probe_slot(struct sdhci_pci_slot *slot)
+{
+	if (slot->chip->pdev->revision == 0) {
+		u16 version;
+
+		version = readl(slot->host->ioaddr + SDHCI_HOST_VERSION);
+		version = (version & SDHCI_VENDOR_VER_MASK) >>
+			SDHCI_VENDOR_VER_SHIFT;
+
+		/*
+		 * Older versions of the chip have lots of nasty glitches
+		 * in the ADMA engine. It's best just to avoid it
+		 * completely.
+		 */
+		if (version < 0xAC)
+			slot->host->quirks |= SDHCI_QUIRK_BROKEN_ADMA;
+	}
+
+	/* JM388 MMC doesn't support 1.8V while SD supports it */
+	if (slot->chip->pdev->device == PCI_DEVICE_ID_JMICRON_JMB388_ESD) {
+		slot->host->ocr_avail_sd = MMC_VDD_32_33 | MMC_VDD_33_34 |
+			MMC_VDD_29_30 | MMC_VDD_30_31 |
+			MMC_VDD_165_195; /* allow 1.8V */
+		slot->host->ocr_avail_mmc = MMC_VDD_32_33 | MMC_VDD_33_34 |
+			MMC_VDD_29_30 | MMC_VDD_30_31; /* no 1.8V for MMC */
+	}
+
+	/*
+	 * The secondary interface requires a bit set to get the
+	 * interrupts.
+	 */
+	if (slot->chip->pdev->device == PCI_DEVICE_ID_JMICRON_JMB38X_MMC ||
+	    slot->chip->pdev->device == PCI_DEVICE_ID_JMICRON_JMB388_ESD)
+		jmicron_enable_mmc(slot->host, 1);
+
+	slot->host->mmc->caps |= MMC_CAP_BUS_WIDTH_TEST;
+
+	return 0;
+}
+
+static void jmicron_remove_slot(struct sdhci_pci_slot *slot, int dead)
+{
+	if (dead)
+		return;
+
+	if (slot->chip->pdev->device == PCI_DEVICE_ID_JMICRON_JMB38X_MMC ||
+	    slot->chip->pdev->device == PCI_DEVICE_ID_JMICRON_JMB388_ESD)
+		jmicron_enable_mmc(slot->host, 0);
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int jmicron_suspend(struct sdhci_pci_chip *chip)
+{
+	int i, ret;
+
+	ret = sdhci_pci_suspend_host(chip);
+	if (ret)
+		return ret;
+
+	if (chip->pdev->device == PCI_DEVICE_ID_JMICRON_JMB38X_MMC ||
+	    chip->pdev->device == PCI_DEVICE_ID_JMICRON_JMB388_ESD) {
+		for (i = 0; i < chip->num_slots; i++)
+			jmicron_enable_mmc(chip->slots[i]->host, 0);
+	}
+
+	return 0;
+}
+
+static int jmicron_resume(struct sdhci_pci_chip *chip)
+{
+	int ret, i;
+
+	if (chip->pdev->device == PCI_DEVICE_ID_JMICRON_JMB38X_MMC ||
+	    chip->pdev->device == PCI_DEVICE_ID_JMICRON_JMB388_ESD) {
+		for (i = 0; i < chip->num_slots; i++)
+			jmicron_enable_mmc(chip->slots[i]->host, 1);
+	}
+
+	ret = jmicron_pmos(chip, 1);
+	if (ret) {
+		dev_err(&chip->pdev->dev, "Failure enabling card power\n");
+		return ret;
+	}
+
+	return sdhci_pci_resume_host(chip);
+}
+#endif
+
+static const struct sdhci_pci_fixes sdhci_jmicron = {
+	.probe		= jmicron_probe,
+
+	.probe_slot	= jmicron_probe_slot,
+	.remove_slot	= jmicron_remove_slot,
+
+#ifdef CONFIG_PM_SLEEP
+	.suspend	= jmicron_suspend,
+	.resume		= jmicron_resume,
+#endif
+};
+
+/* SysKonnect CardBus2SDIO extra registers */
+#define SYSKT_CTRL		0x200
+#define SYSKT_RDFIFO_STAT	0x204
+#define SYSKT_WRFIFO_STAT	0x208
+#define SYSKT_POWER_DATA	0x20c
+#define   SYSKT_POWER_330	0xef
+#define   SYSKT_POWER_300	0xf8
+#define   SYSKT_POWER_184	0xcc
+#define SYSKT_POWER_CMD		0x20d
+#define   SYSKT_POWER_START	(1 << 7)
+#define SYSKT_POWER_STATUS	0x20e
+#define   SYSKT_POWER_STATUS_OK	(1 << 0)
+#define SYSKT_BOARD_REV		0x210
+#define SYSKT_CHIP_REV		0x211
+#define SYSKT_CONF_DATA		0x212
+#define   SYSKT_CONF_DATA_1V8	(1 << 2)
+#define   SYSKT_CONF_DATA_2V5	(1 << 1)
+#define   SYSKT_CONF_DATA_3V3	(1 << 0)
+
+static int syskt_probe(struct sdhci_pci_chip *chip)
+{
+	if ((chip->pdev->class & 0x0000FF) == PCI_SDHCI_IFVENDOR) {
+		chip->pdev->class &= ~0x0000FF;
+		chip->pdev->class |= PCI_SDHCI_IFDMA;
+	}
+	return 0;
+}
+
+static int syskt_probe_slot(struct sdhci_pci_slot *slot)
+{
+	int tm, ps;
+
+	u8 board_rev = readb(slot->host->ioaddr + SYSKT_BOARD_REV);
+	u8  chip_rev = readb(slot->host->ioaddr + SYSKT_CHIP_REV);
+	dev_info(&slot->chip->pdev->dev, "SysKonnect CardBus2SDIO, "
+					 "board rev %d.%d, chip rev %d.%d\n",
+					 board_rev >> 4, board_rev & 0xf,
+					 chip_rev >> 4,  chip_rev & 0xf);
+	if (chip_rev >= 0x20)
+		slot->host->quirks |= SDHCI_QUIRK_FORCE_DMA;
+
+	writeb(SYSKT_POWER_330, slot->host->ioaddr + SYSKT_POWER_DATA);
+	writeb(SYSKT_POWER_START, slot->host->ioaddr + SYSKT_POWER_CMD);
+	udelay(50);
+	tm = 10;  /* Wait max 1 ms */
+	do {
+		ps = readw(slot->host->ioaddr + SYSKT_POWER_STATUS);
+		if (ps & SYSKT_POWER_STATUS_OK)
+			break;
+		udelay(100);
+	} while (--tm);
+	if (!tm) {
+		dev_err(&slot->chip->pdev->dev,
+			"power regulator never stabilized");
+		writeb(0, slot->host->ioaddr + SYSKT_POWER_CMD);
+		return -ENODEV;
+	}
+
+	return 0;
+}
+
+static const struct sdhci_pci_fixes sdhci_syskt = {
+	.quirks		= SDHCI_QUIRK_NO_SIMULT_VDD_AND_POWER,
+	.probe		= syskt_probe,
+	.probe_slot	= syskt_probe_slot,
+};
+
+static int via_probe(struct sdhci_pci_chip *chip)
+{
+	if (chip->pdev->revision == 0x10)
+		chip->quirks |= SDHCI_QUIRK_DELAY_AFTER_POWER;
+
+	return 0;
+}
+
+static const struct sdhci_pci_fixes sdhci_via = {
+	.probe		= via_probe,
+};
+
+static int rtsx_probe_slot(struct sdhci_pci_slot *slot)
+{
+	slot->host->mmc->caps2 |= MMC_CAP2_HS200;
+	return 0;
+}
+
+static const struct sdhci_pci_fixes sdhci_rtsx = {
+	.quirks2	= SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+			SDHCI_QUIRK2_BROKEN_64_BIT_DMA |
+			SDHCI_QUIRK2_BROKEN_DDR50,
+	.probe_slot	= rtsx_probe_slot,
+};
+
+/*AMD chipset generation*/
+enum amd_chipset_gen {
+	AMD_CHIPSET_BEFORE_ML,
+	AMD_CHIPSET_CZ,
+	AMD_CHIPSET_NL,
+	AMD_CHIPSET_UNKNOWN,
+};
+
+/* AMD registers */
+#define AMD_SD_AUTO_PATTERN		0xB8
+#define AMD_MSLEEP_DURATION		4
+#define AMD_SD_MISC_CONTROL		0xD0
+#define AMD_MAX_TUNE_VALUE		0x0B
+#define AMD_AUTO_TUNE_SEL		0x10800
+#define AMD_FIFO_PTR			0x30
+#define AMD_BIT_MASK			0x1F
+
+static void amd_tuning_reset(struct sdhci_host *host)
+{
+	unsigned int val;
+
+	val = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	val |= SDHCI_CTRL_PRESET_VAL_ENABLE | SDHCI_CTRL_EXEC_TUNING;
+	sdhci_writew(host, val, SDHCI_HOST_CONTROL2);
+
+	val = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	val &= ~SDHCI_CTRL_EXEC_TUNING;
+	sdhci_writew(host, val, SDHCI_HOST_CONTROL2);
+}
+
+static void amd_config_tuning_phase(struct pci_dev *pdev, u8 phase)
+{
+	unsigned int val;
+
+	pci_read_config_dword(pdev, AMD_SD_AUTO_PATTERN, &val);
+	val &= ~AMD_BIT_MASK;
+	val |= (AMD_AUTO_TUNE_SEL | (phase << 1));
+	pci_write_config_dword(pdev, AMD_SD_AUTO_PATTERN, val);
+}
+
+static void amd_enable_manual_tuning(struct pci_dev *pdev)
+{
+	unsigned int val;
+
+	pci_read_config_dword(pdev, AMD_SD_MISC_CONTROL, &val);
+	val |= AMD_FIFO_PTR;
+	pci_write_config_dword(pdev, AMD_SD_MISC_CONTROL, val);
+}
+
+static int amd_execute_tuning_hs200(struct sdhci_host *host, u32 opcode)
+{
+	struct sdhci_pci_slot *slot = sdhci_priv(host);
+	struct pci_dev *pdev = slot->chip->pdev;
+	u8 valid_win = 0;
+	u8 valid_win_max = 0;
+	u8 valid_win_end = 0;
+	u8 ctrl, tune_around;
+
+	amd_tuning_reset(host);
+
+	for (tune_around = 0; tune_around < 12; tune_around++) {
+		amd_config_tuning_phase(pdev, tune_around);
+
+		if (mmc_send_tuning(host->mmc, opcode, NULL)) {
+			valid_win = 0;
+			msleep(AMD_MSLEEP_DURATION);
+			ctrl = SDHCI_RESET_CMD | SDHCI_RESET_DATA;
+			sdhci_writeb(host, ctrl, SDHCI_SOFTWARE_RESET);
+		} else if (++valid_win > valid_win_max) {
+			valid_win_max = valid_win;
+			valid_win_end = tune_around;
+		}
+	}
+
+	if (!valid_win_max) {
+		dev_err(&pdev->dev, "no tuning point found\n");
+		return -EIO;
+	}
+
+	amd_config_tuning_phase(pdev, valid_win_end - valid_win_max / 2);
+
+	amd_enable_manual_tuning(pdev);
+
+	host->mmc->retune_period = 0;
+
+	return 0;
+}
+
+static int amd_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	/* AMD requires custom HS200 tuning */
+	if (host->timing == MMC_TIMING_MMC_HS200)
+		return amd_execute_tuning_hs200(host, opcode);
+
+	/* Otherwise perform standard SDHCI tuning */
+	return sdhci_execute_tuning(mmc, opcode);
+}
+
+static int amd_probe_slot(struct sdhci_pci_slot *slot)
+{
+	struct mmc_host_ops *ops = &slot->host->mmc_host_ops;
+
+	ops->execute_tuning = amd_execute_tuning;
+
+	return 0;
+}
+
+static int amd_probe(struct sdhci_pci_chip *chip)
+{
+	struct pci_dev	*smbus_dev;
+	enum amd_chipset_gen gen;
+
+	smbus_dev = pci_get_device(PCI_VENDOR_ID_AMD,
+			PCI_DEVICE_ID_AMD_HUDSON2_SMBUS, NULL);
+	if (smbus_dev) {
+		gen = AMD_CHIPSET_BEFORE_ML;
+	} else {
+		smbus_dev = pci_get_device(PCI_VENDOR_ID_AMD,
+				PCI_DEVICE_ID_AMD_KERNCZ_SMBUS, NULL);
+		if (smbus_dev) {
+			if (smbus_dev->revision < 0x51)
+				gen = AMD_CHIPSET_CZ;
+			else
+				gen = AMD_CHIPSET_NL;
+		} else {
+			gen = AMD_CHIPSET_UNKNOWN;
+		}
+	}
+
+	pci_dev_put(smbus_dev);
+
+	if (gen == AMD_CHIPSET_BEFORE_ML || gen == AMD_CHIPSET_CZ)
+		chip->quirks2 |= SDHCI_QUIRK2_CLEAR_TRANSFERMODE_REG_BEFORE_CMD;
+
+	return 0;
+}
+
+static u32 sdhci_read_present_state(struct sdhci_host *host)
+{
+	return sdhci_readl(host, SDHCI_PRESENT_STATE);
+}
+
+static void amd_sdhci_reset(struct sdhci_host *host, u8 mask)
+{
+	struct sdhci_pci_slot *slot = sdhci_priv(host);
+	struct pci_dev *pdev = slot->chip->pdev;
+	u32 present_state;
+
+	/*
+	 * SDHC 0x7906 requires a hard reset to clear all internal state.
+	 * Otherwise it can get into a bad state where the DATA lines are always
+	 * read as zeros.
+	 */
+	if (pdev->device == 0x7906 && (mask & SDHCI_RESET_ALL)) {
+		pci_clear_master(pdev);
+
+		pci_save_state(pdev);
+
+		pci_set_power_state(pdev, PCI_D3cold);
+		pr_debug("%s: power_state=%u\n", mmc_hostname(host->mmc),
+			pdev->current_state);
+		pci_set_power_state(pdev, PCI_D0);
+
+		pci_restore_state(pdev);
+
+		/*
+		 * SDHCI_RESET_ALL says the card detect logic should not be
+		 * reset, but since we need to reset the entire controller
+		 * we should wait until the card detect logic has stabilized.
+		 *
+		 * This normally takes about 40ms.
+		 */
+		readx_poll_timeout(
+			sdhci_read_present_state,
+			host,
+			present_state,
+			present_state & SDHCI_CD_STABLE,
+			10000,
+			100000
+		);
+	}
+
+	return sdhci_reset(host, mask);
+}
+
+static const struct sdhci_ops amd_sdhci_pci_ops = {
+	.set_clock			= sdhci_set_clock,
+	.enable_dma			= sdhci_pci_enable_dma,
+	.set_bus_width			= sdhci_set_bus_width,
+	.reset				= amd_sdhci_reset,
+	.set_uhs_signaling		= sdhci_set_uhs_signaling,
+};
+
+static const struct sdhci_pci_fixes sdhci_amd = {
+	.probe		= amd_probe,
+	.ops		= &amd_sdhci_pci_ops,
+	.probe_slot	= amd_probe_slot,
+};
+
+static const struct pci_device_id pci_ids[] = {
+	SDHCI_PCI_DEVICE(RICOH, R5C822,  ricoh),
+	SDHCI_PCI_DEVICE(RICOH, R5C843,  ricoh_mmc),
+	SDHCI_PCI_DEVICE(RICOH, R5CE822, ricoh_mmc),
+	SDHCI_PCI_DEVICE(RICOH, R5CE823, ricoh_mmc),
+	SDHCI_PCI_DEVICE(ENE, CB712_SD,   ene_712),
+	SDHCI_PCI_DEVICE(ENE, CB712_SD_2, ene_712),
+	SDHCI_PCI_DEVICE(ENE, CB714_SD,   ene_714),
+	SDHCI_PCI_DEVICE(ENE, CB714_SD_2, ene_714),
+	SDHCI_PCI_DEVICE(MARVELL, 88ALP01_SD, cafe),
+	SDHCI_PCI_DEVICE(JMICRON, JMB38X_SD,  jmicron),
+	SDHCI_PCI_DEVICE(JMICRON, JMB38X_MMC, jmicron),
+	SDHCI_PCI_DEVICE(JMICRON, JMB388_SD,  jmicron),
+	SDHCI_PCI_DEVICE(JMICRON, JMB388_ESD, jmicron),
+	SDHCI_PCI_DEVICE(SYSKONNECT, 8000, syskt),
+	SDHCI_PCI_DEVICE(VIA, 95D0, via),
+	SDHCI_PCI_DEVICE(REALTEK, 5250, rtsx),
+	SDHCI_PCI_DEVICE(INTEL, QRK_SD,    intel_qrk),
+	SDHCI_PCI_DEVICE(INTEL, MRST_SD0,  intel_mrst_hc0),
+	SDHCI_PCI_DEVICE(INTEL, MRST_SD1,  intel_mrst_hc1_hc2),
+	SDHCI_PCI_DEVICE(INTEL, MRST_SD2,  intel_mrst_hc1_hc2),
+	SDHCI_PCI_DEVICE(INTEL, MFD_SD,    intel_mfd_sd),
+	SDHCI_PCI_DEVICE(INTEL, MFD_SDIO1, intel_mfd_sdio),
+	SDHCI_PCI_DEVICE(INTEL, MFD_SDIO2, intel_mfd_sdio),
+	SDHCI_PCI_DEVICE(INTEL, MFD_EMMC0, intel_mfd_emmc),
+	SDHCI_PCI_DEVICE(INTEL, MFD_EMMC1, intel_mfd_emmc),
+	SDHCI_PCI_DEVICE(INTEL, PCH_SDIO0, intel_pch_sdio),
+	SDHCI_PCI_DEVICE(INTEL, PCH_SDIO1, intel_pch_sdio),
+	SDHCI_PCI_DEVICE(INTEL, BYT_EMMC,  intel_byt_emmc),
+	SDHCI_PCI_SUBDEVICE(INTEL, BYT_SDIO, NI, 7884, ni_byt_sdio),
+	SDHCI_PCI_DEVICE(INTEL, BYT_SDIO,  intel_byt_sdio),
+	SDHCI_PCI_DEVICE(INTEL, BYT_SD,    intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, BYT_EMMC2, intel_byt_emmc),
+	SDHCI_PCI_DEVICE(INTEL, BSW_EMMC,  intel_byt_emmc),
+	SDHCI_PCI_DEVICE(INTEL, BSW_SDIO,  intel_byt_sdio),
+	SDHCI_PCI_DEVICE(INTEL, BSW_SD,    intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, CLV_SDIO0, intel_mfd_sd),
+	SDHCI_PCI_DEVICE(INTEL, CLV_SDIO1, intel_mfd_sdio),
+	SDHCI_PCI_DEVICE(INTEL, CLV_SDIO2, intel_mfd_sdio),
+	SDHCI_PCI_DEVICE(INTEL, CLV_EMMC0, intel_mfd_emmc),
+	SDHCI_PCI_DEVICE(INTEL, CLV_EMMC1, intel_mfd_emmc),
+	SDHCI_PCI_DEVICE(INTEL, MRFLD_MMC, intel_mrfld_mmc),
+	SDHCI_PCI_DEVICE(INTEL, SPT_EMMC,  intel_byt_emmc),
+	SDHCI_PCI_DEVICE(INTEL, SPT_SDIO,  intel_byt_sdio),
+	SDHCI_PCI_DEVICE(INTEL, SPT_SD,    intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, DNV_EMMC,  intel_byt_emmc),
+	SDHCI_PCI_DEVICE(INTEL, CDF_EMMC,  intel_glk_emmc),
+	SDHCI_PCI_DEVICE(INTEL, BXT_EMMC,  intel_byt_emmc),
+	SDHCI_PCI_DEVICE(INTEL, BXT_SDIO,  intel_byt_sdio),
+	SDHCI_PCI_DEVICE(INTEL, BXT_SD,    intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, BXTM_EMMC, intel_byt_emmc),
+	SDHCI_PCI_DEVICE(INTEL, BXTM_SDIO, intel_byt_sdio),
+	SDHCI_PCI_DEVICE(INTEL, BXTM_SD,   intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, APL_EMMC,  intel_byt_emmc),
+	SDHCI_PCI_DEVICE(INTEL, APL_SDIO,  intel_byt_sdio),
+	SDHCI_PCI_DEVICE(INTEL, APL_SD,    intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, GLK_EMMC,  intel_glk_emmc),
+	SDHCI_PCI_DEVICE(INTEL, GLK_SDIO,  intel_byt_sdio),
+	SDHCI_PCI_DEVICE(INTEL, GLK_SD,    intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, CNP_EMMC,  intel_glk_emmc),
+	SDHCI_PCI_DEVICE(INTEL, CNP_SD,    intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, CNPH_SD,   intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, ICP_EMMC,  intel_glk_emmc),
+	SDHCI_PCI_DEVICE(INTEL, ICP_SD,    intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, EHL_EMMC,  intel_glk_emmc),
+	SDHCI_PCI_DEVICE(INTEL, EHL_SD,    intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, CML_EMMC,  intel_glk_emmc),
+	SDHCI_PCI_DEVICE(INTEL, CML_SD,    intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, CMLH_SD,   intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, JSL_EMMC,  intel_glk_emmc),
+	SDHCI_PCI_DEVICE(INTEL, JSL_SD,    intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, LKF_EMMC,  intel_glk_emmc),
+	SDHCI_PCI_DEVICE(INTEL, LKF_SD,    intel_byt_sd),
+	SDHCI_PCI_DEVICE(INTEL, ADL_EMMC,  intel_glk_emmc),
+	SDHCI_PCI_DEVICE(O2, 8120,     o2),
+	SDHCI_PCI_DEVICE(O2, 8220,     o2),
+	SDHCI_PCI_DEVICE(O2, 8221,     o2),
+	SDHCI_PCI_DEVICE(O2, 8320,     o2),
+	SDHCI_PCI_DEVICE(O2, 8321,     o2),
+	SDHCI_PCI_DEVICE(O2, FUJIN2,   o2),
+	SDHCI_PCI_DEVICE(O2, SDS0,     o2),
+	SDHCI_PCI_DEVICE(O2, SDS1,     o2),
+	SDHCI_PCI_DEVICE(O2, SEABIRD0, o2),
+	SDHCI_PCI_DEVICE(O2, SEABIRD1, o2),
+	SDHCI_PCI_DEVICE(ARASAN, PHY_EMMC, arasan),
+	SDHCI_PCI_DEVICE(SYNOPSYS, DWC_MSHC, snps),
+	SDHCI_PCI_DEVICE(GLI, 9750, gl9750),
+	SDHCI_PCI_DEVICE(GLI, 9755, gl9755),
+	SDHCI_PCI_DEVICE(GLI, 9763E, gl9763e),
+	SDHCI_PCI_DEVICE_CLASS(AMD, SYSTEM_SDHCI, PCI_CLASS_MASK, amd),
+	/* Generic SD host controller */
+	{PCI_DEVICE_CLASS(SYSTEM_SDHCI, PCI_CLASS_MASK)},
+	{ /* end: all zeroes */ },
+};
+
+MODULE_DEVICE_TABLE(pci, pci_ids);
+
+/*****************************************************************************\
+ *                                                                           *
+ * SDHCI core callbacks                                                      *
+ *                                                                           *
+\*****************************************************************************/
+
+int sdhci_pci_enable_dma(struct sdhci_host *host)
+{
+	struct sdhci_pci_slot *slot;
+	struct pci_dev *pdev;
+
+	slot = sdhci_priv(host);
+	pdev = slot->chip->pdev;
+
+	if (((pdev->class & 0xFFFF00) == (PCI_CLASS_SYSTEM_SDHCI << 8)) &&
+		((pdev->class & 0x0000FF) != PCI_SDHCI_IFDMA) &&
+		(host->flags & SDHCI_USE_SDMA)) {
+		dev_warn(&pdev->dev, "Will use DMA mode even though HW "
+			"doesn't fully claim to support it.\n");
+	}
+
+	pci_set_master(pdev);
+
+	return 0;
+}
+
+static void sdhci_pci_gpio_hw_reset(struct sdhci_host *host)
+{
+	struct sdhci_pci_slot *slot = sdhci_priv(host);
+	int rst_n_gpio = slot->rst_n_gpio;
+
+	if (!gpio_is_valid(rst_n_gpio))
+		return;
+	gpio_set_value_cansleep(rst_n_gpio, 0);
+	/* For eMMC, minimum is 1us but give it 10us for good measure */
+	udelay(10);
+	gpio_set_value_cansleep(rst_n_gpio, 1);
+	/* For eMMC, minimum is 200us but give it 300us for good measure */
+	usleep_range(300, 1000);
+}
+
+static void sdhci_pci_hw_reset(struct sdhci_host *host)
+{
+	struct sdhci_pci_slot *slot = sdhci_priv(host);
+
+	if (slot->hw_reset)
+		slot->hw_reset(host);
+}
+
+static const struct sdhci_ops sdhci_pci_ops = {
+	.set_clock	= sdhci_set_clock,
+	.enable_dma	= sdhci_pci_enable_dma,
+	.set_bus_width	= sdhci_set_bus_width,
+	.reset		= sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+	.hw_reset		= sdhci_pci_hw_reset,
+};
+
+/*****************************************************************************\
+ *                                                                           *
+ * Suspend/resume                                                            *
+ *                                                                           *
+\*****************************************************************************/
+
+#ifdef CONFIG_PM_SLEEP
+static int sdhci_pci_suspend(struct device *dev)
+{
+	struct sdhci_pci_chip *chip = dev_get_drvdata(dev);
+
+	if (!chip)
+		return 0;
+
+	if (chip->fixes && chip->fixes->suspend)
+		return chip->fixes->suspend(chip);
+
+	return sdhci_pci_suspend_host(chip);
+}
+
+static int sdhci_pci_resume(struct device *dev)
+{
+	struct sdhci_pci_chip *chip = dev_get_drvdata(dev);
+
+	if (!chip)
+		return 0;
+
+	if (chip->fixes && chip->fixes->resume)
+		return chip->fixes->resume(chip);
+
+	return sdhci_pci_resume_host(chip);
+}
+#endif
+
+#ifdef CONFIG_PM
+static int sdhci_pci_runtime_suspend(struct device *dev)
+{
+	struct sdhci_pci_chip *chip = dev_get_drvdata(dev);
+
+	if (!chip)
+		return 0;
+
+	if (chip->fixes && chip->fixes->runtime_suspend)
+		return chip->fixes->runtime_suspend(chip);
+
+	return sdhci_pci_runtime_suspend_host(chip);
+}
+
+static int sdhci_pci_runtime_resume(struct device *dev)
+{
+	struct sdhci_pci_chip *chip = dev_get_drvdata(dev);
+
+	if (!chip)
+		return 0;
+
+	if (chip->fixes && chip->fixes->runtime_resume)
+		return chip->fixes->runtime_resume(chip);
+
+	return sdhci_pci_runtime_resume_host(chip);
+}
+#endif
+
+static const struct dev_pm_ops sdhci_pci_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(sdhci_pci_suspend, sdhci_pci_resume)
+	SET_RUNTIME_PM_OPS(sdhci_pci_runtime_suspend,
+			sdhci_pci_runtime_resume, NULL)
+};
+
+/*****************************************************************************\
+ *                                                                           *
+ * Device probing/removal                                                    *
+ *                                                                           *
+\*****************************************************************************/
+
+static struct sdhci_pci_slot *sdhci_pci_probe_slot(
+	struct pci_dev *pdev, struct sdhci_pci_chip *chip, int first_bar,
+	int slotno)
+{
+	struct sdhci_pci_slot *slot;
+	struct sdhci_host *host;
+	int ret, bar = first_bar + slotno;
+	size_t priv_size = chip->fixes ? chip->fixes->priv_size : 0;
+
+	if (!(pci_resource_flags(pdev, bar) & IORESOURCE_MEM)) {
+		dev_err(&pdev->dev, "BAR %d is not iomem. Aborting.\n", bar);
+		return ERR_PTR(-ENODEV);
+	}
+
+	if (pci_resource_len(pdev, bar) < 0x100) {
+		dev_err(&pdev->dev, "Invalid iomem size. You may "
+			"experience problems.\n");
+	}
+
+	if ((pdev->class & 0x0000FF) == PCI_SDHCI_IFVENDOR) {
+		dev_err(&pdev->dev, "Vendor specific interface. Aborting.\n");
+		return ERR_PTR(-ENODEV);
+	}
+
+	if ((pdev->class & 0x0000FF) > PCI_SDHCI_IFVENDOR) {
+		dev_err(&pdev->dev, "Unknown interface. Aborting.\n");
+		return ERR_PTR(-ENODEV);
+	}
+
+	host = sdhci_alloc_host(&pdev->dev, sizeof(*slot) + priv_size);
+	if (IS_ERR(host)) {
+		dev_err(&pdev->dev, "cannot allocate host\n");
+		return ERR_CAST(host);
+	}
+
+	slot = sdhci_priv(host);
+
+	slot->chip = chip;
+	slot->host = host;
+	slot->rst_n_gpio = -EINVAL;
+	slot->cd_gpio = -EINVAL;
+	slot->cd_idx = -1;
+
+	/* Retrieve platform data if there is any */
+	if (*sdhci_pci_get_data)
+		slot->data = sdhci_pci_get_data(pdev, slotno);
+
+	if (slot->data) {
+		if (slot->data->setup) {
+			ret = slot->data->setup(slot->data);
+			if (ret) {
+				dev_err(&pdev->dev, "platform setup failed\n");
+				goto free;
+			}
+		}
+		slot->rst_n_gpio = slot->data->rst_n_gpio;
+		slot->cd_gpio = slot->data->cd_gpio;
+	}
+
+	host->hw_name = "PCI";
+	host->ops = chip->fixes && chip->fixes->ops ?
+		    chip->fixes->ops :
+		    &sdhci_pci_ops;
+	host->quirks = chip->quirks;
+	host->quirks2 = chip->quirks2;
+
+	host->irq = pdev->irq;
+
+	ret = pcim_iomap_regions(pdev, BIT(bar), mmc_hostname(host->mmc));
+	if (ret) {
+		dev_err(&pdev->dev, "cannot request region\n");
+		goto cleanup;
+	}
+
+	host->ioaddr = pcim_iomap_table(pdev)[bar];
+
+	if (chip->fixes && chip->fixes->probe_slot) {
+		ret = chip->fixes->probe_slot(slot);
+		if (ret)
+			goto cleanup;
+	}
+
+	if (gpio_is_valid(slot->rst_n_gpio)) {
+		if (!devm_gpio_request(&pdev->dev, slot->rst_n_gpio, "eMMC_reset")) {
+			gpio_direction_output(slot->rst_n_gpio, 1);
+			slot->host->mmc->caps |= MMC_CAP_HW_RESET;
+			slot->hw_reset = sdhci_pci_gpio_hw_reset;
+		} else {
+			dev_warn(&pdev->dev, "failed to request rst_n_gpio\n");
+			slot->rst_n_gpio = -EINVAL;
+		}
+	}
+
+	host->mmc->pm_caps = MMC_PM_KEEP_POWER;
+	host->mmc->slotno = slotno;
+	host->mmc->caps2 |= MMC_CAP2_NO_PRESCAN_POWERUP;
+
+	if (device_can_wakeup(&pdev->dev))
+		host->mmc->pm_caps |= MMC_PM_WAKE_SDIO_IRQ;
+
+	if (host->mmc->caps & MMC_CAP_CD_WAKE)
+		device_init_wakeup(&pdev->dev, true);
+
+	if (slot->cd_idx >= 0) {
+		ret = mmc_gpiod_request_cd(host->mmc, "cd", slot->cd_idx,
+					   slot->cd_override_level, 0);
+		if (ret && ret != -EPROBE_DEFER)
+			ret = mmc_gpiod_request_cd(host->mmc, NULL,
+						   slot->cd_idx,
+						   slot->cd_override_level,
+						   0);
+		if (ret == -EPROBE_DEFER)
+			goto remove;
+
+		if (ret) {
+			dev_warn(&pdev->dev, "failed to setup card detect gpio\n");
+			slot->cd_idx = -1;
+		}
+	}
+
+	if (chip->fixes && chip->fixes->add_host)
+		ret = chip->fixes->add_host(slot);
+	else
+		ret = sdhci_add_host(host);
+	if (ret)
+		goto remove;
+
+	sdhci_pci_add_own_cd(slot);
+
+	/*
+	 * Check if the chip needs a separate GPIO for card detect to wake up
+	 * from runtime suspend.  If it is not there, don't allow runtime PM.
+	 * Note sdhci_pci_add_own_cd() sets slot->cd_gpio to -EINVAL on failure.
+	 */
+	if (chip->fixes && chip->fixes->own_cd_for_runtime_pm &&
+	    !gpio_is_valid(slot->cd_gpio) && slot->cd_idx < 0)
+		chip->allow_runtime_pm = false;
+
+	return slot;
+
+remove:
+	if (chip->fixes && chip->fixes->remove_slot)
+		chip->fixes->remove_slot(slot, 0);
+
+cleanup:
+	if (slot->data && slot->data->cleanup)
+		slot->data->cleanup(slot->data);
+
+free:
+	sdhci_free_host(host);
+
+	return ERR_PTR(ret);
+}
+
+static void sdhci_pci_remove_slot(struct sdhci_pci_slot *slot)
+{
+	int dead;
+	u32 scratch;
+
+	sdhci_pci_remove_own_cd(slot);
+
+	dead = 0;
+	scratch = readl(slot->host->ioaddr + SDHCI_INT_STATUS);
+	if (scratch == (u32)-1)
+		dead = 1;
+
+	sdhci_remove_host(slot->host, dead);
+
+	if (slot->chip->fixes && slot->chip->fixes->remove_slot)
+		slot->chip->fixes->remove_slot(slot, dead);
+
+	if (slot->data && slot->data->cleanup)
+		slot->data->cleanup(slot->data);
+
+	sdhci_free_host(slot->host);
+}
+
+static void sdhci_pci_runtime_pm_allow(struct device *dev)
+{
+	pm_suspend_ignore_children(dev, 1);
+	pm_runtime_set_autosuspend_delay(dev, 50);
+	pm_runtime_use_autosuspend(dev);
+	pm_runtime_allow(dev);
+	/* Stay active until mmc core scans for a card */
+	pm_runtime_put_noidle(dev);
+}
+
+static void sdhci_pci_runtime_pm_forbid(struct device *dev)
+{
+	pm_runtime_forbid(dev);
+	pm_runtime_get_noresume(dev);
+}
+
+static int sdhci_pci_probe(struct pci_dev *pdev,
+				     const struct pci_device_id *ent)
+{
+	struct sdhci_pci_chip *chip;
+	struct sdhci_pci_slot *slot;
+
+	u8 slots, first_bar;
+	int ret, i;
+
+	BUG_ON(pdev == NULL);
+	BUG_ON(ent == NULL);
+
+	dev_info(&pdev->dev, "SDHCI controller found [%04x:%04x] (rev %x)\n",
+		 (int)pdev->vendor, (int)pdev->device, (int)pdev->revision);
+
+	ret = pci_read_config_byte(pdev, PCI_SLOT_INFO, &slots);
+	if (ret)
+		return ret;
+
+	slots = PCI_SLOT_INFO_SLOTS(slots) + 1;
+	dev_dbg(&pdev->dev, "found %d slot(s)\n", slots);
+
+	BUG_ON(slots > MAX_SLOTS);
+
+	ret = pci_read_config_byte(pdev, PCI_SLOT_INFO, &first_bar);
+	if (ret)
+		return ret;
+
+	first_bar &= PCI_SLOT_INFO_FIRST_BAR_MASK;
+
+	if (first_bar > 5) {
+		dev_err(&pdev->dev, "Invalid first BAR. Aborting.\n");
+		return -ENODEV;
+	}
+
+	ret = pcim_enable_device(pdev);
+	if (ret)
+		return ret;
+
+	chip = devm_kzalloc(&pdev->dev, sizeof(*chip), GFP_KERNEL);
+	if (!chip)
+		return -ENOMEM;
+
+	chip->pdev = pdev;
+	chip->fixes = (const struct sdhci_pci_fixes *)ent->driver_data;
+	if (chip->fixes) {
+		chip->quirks = chip->fixes->quirks;
+		chip->quirks2 = chip->fixes->quirks2;
+		chip->allow_runtime_pm = chip->fixes->allow_runtime_pm;
+	}
+	chip->num_slots = slots;
+	chip->pm_retune = true;
+	chip->rpm_retune = true;
+
+	pci_set_drvdata(pdev, chip);
+
+	if (chip->fixes && chip->fixes->probe) {
+		ret = chip->fixes->probe(chip);
+		if (ret)
+			return ret;
+	}
+
+	slots = chip->num_slots;	/* Quirk may have changed this */
+
+	for (i = 0; i < slots; i++) {
+		slot = sdhci_pci_probe_slot(pdev, chip, first_bar, i);
+		if (IS_ERR(slot)) {
+			for (i--; i >= 0; i--)
+				sdhci_pci_remove_slot(chip->slots[i]);
+			return PTR_ERR(slot);
+		}
+
+		chip->slots[i] = slot;
+	}
+
+	if (chip->allow_runtime_pm)
+		sdhci_pci_runtime_pm_allow(&pdev->dev);
+
+	return 0;
+}
+
+static void sdhci_pci_remove(struct pci_dev *pdev)
+{
+	int i;
+	struct sdhci_pci_chip *chip = pci_get_drvdata(pdev);
+
+	if (chip->allow_runtime_pm)
+		sdhci_pci_runtime_pm_forbid(&pdev->dev);
+
+	for (i = 0; i < chip->num_slots; i++)
+		sdhci_pci_remove_slot(chip->slots[i]);
+}
+
+static struct pci_driver sdhci_driver = {
+	.name =		"sdhci-pci",
+	.id_table =	pci_ids,
+	.probe =	sdhci_pci_probe,
+	.remove =	sdhci_pci_remove,
+	.driver =	{
+		.pm =   &sdhci_pci_pm_ops
+	},
+};
+
+module_pci_driver(sdhci_driver);
+
+MODULE_AUTHOR("Pierre Ossman <pierre@ossman.eu>");
+MODULE_DESCRIPTION("Secure Digital Host Controller Interface PCI driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mmc/host/sdhci-pci-data.c b/drivers/mmc/host/sdhci-pci-data.c
new file mode 100644
index 000000000..18638fb36
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pci-data.c
@@ -0,0 +1,6 @@
+// SPDX-License-Identifier: GPL-2.0-only
+#include <linux/module.h>
+#include <linux/mmc/sdhci-pci-data.h>
+
+struct sdhci_pci_data *(*sdhci_pci_get_data)(struct pci_dev *pdev, int slotno);
+EXPORT_SYMBOL_GPL(sdhci_pci_get_data);
diff --git a/drivers/mmc/host/sdhci-pci-dwc-mshc.c b/drivers/mmc/host/sdhci-pci-dwc-mshc.c
new file mode 100644
index 000000000..f78d65448
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pci-dwc-mshc.c
@@ -0,0 +1,84 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * SDHCI driver for Synopsys DWC_MSHC controller
+ *
+ * Copyright (C) 2018 Synopsys, Inc. (www.synopsys.com)
+ *
+ * Authors:
+ *	Prabu Thangamuthu <prabu.t@synopsys.com>
+ *	Manjunath M B <manjumb@synopsys.com>
+ */
+
+#include "sdhci.h"
+#include "sdhci-pci.h"
+
+#define SDHCI_VENDOR_PTR_R	0xE8
+
+/* Synopsys vendor specific registers */
+#define SDHC_GPIO_OUT		0x34
+#define SDHC_AT_CTRL_R		0x40
+#define SDHC_SW_TUNE_EN		0x00000010
+
+/* MMCM DRP */
+#define SDHC_MMCM_DIV_REG	0x1020
+#define DIV_REG_100_MHZ		0x1145
+#define DIV_REG_200_MHZ		0x1083
+#define SDHC_MMCM_CLKFBOUT	0x1024
+#define CLKFBOUT_100_MHZ	0x0000
+#define CLKFBOUT_200_MHZ	0x0080
+#define SDHC_CCLK_MMCM_RST	0x00000001
+
+static void sdhci_snps_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	u16 clk;
+	u32 reg, vendor_ptr;
+
+	vendor_ptr = sdhci_readw(host, SDHCI_VENDOR_PTR_R);
+
+	/* Disable software managed rx tuning */
+	reg = sdhci_readl(host, (SDHC_AT_CTRL_R + vendor_ptr));
+	reg &= ~SDHC_SW_TUNE_EN;
+	sdhci_writel(host, reg, (SDHC_AT_CTRL_R + vendor_ptr));
+
+	if (clock <= 52000000) {
+		sdhci_set_clock(host, clock);
+	} else {
+		/* Assert reset to MMCM */
+		reg = sdhci_readl(host, (SDHC_GPIO_OUT + vendor_ptr));
+		reg |= SDHC_CCLK_MMCM_RST;
+		sdhci_writel(host, reg, (SDHC_GPIO_OUT + vendor_ptr));
+
+		/* Configure MMCM */
+		if (clock == 100000000) {
+			sdhci_writel(host, DIV_REG_100_MHZ, SDHC_MMCM_DIV_REG);
+			sdhci_writel(host, CLKFBOUT_100_MHZ,
+					SDHC_MMCM_CLKFBOUT);
+		} else {
+			sdhci_writel(host, DIV_REG_200_MHZ, SDHC_MMCM_DIV_REG);
+			sdhci_writel(host, CLKFBOUT_200_MHZ,
+					SDHC_MMCM_CLKFBOUT);
+		}
+
+		/* De-assert reset to MMCM */
+		reg = sdhci_readl(host, (SDHC_GPIO_OUT + vendor_ptr));
+		reg &= ~SDHC_CCLK_MMCM_RST;
+		sdhci_writel(host, reg, (SDHC_GPIO_OUT + vendor_ptr));
+
+		/* Enable clock */
+		clk = SDHCI_PROG_CLOCK_MODE | SDHCI_CLOCK_INT_EN |
+			SDHCI_CLOCK_CARD_EN;
+		sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+	}
+}
+
+static const struct sdhci_ops sdhci_snps_ops = {
+	.set_clock	= sdhci_snps_set_clock,
+	.enable_dma	= sdhci_pci_enable_dma,
+	.set_bus_width	= sdhci_set_bus_width,
+	.reset		= sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+const struct sdhci_pci_fixes sdhci_snps = {
+	.ops		= &sdhci_snps_ops,
+};
diff --git a/drivers/mmc/host/sdhci-pci-gli.c b/drivers/mmc/host/sdhci-pci-gli.c
new file mode 100644
index 000000000..23b89b4ca
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pci-gli.c
@@ -0,0 +1,865 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright (C) 2019 Genesys Logic, Inc.
+ *
+ * Authors: Ben Chuang <ben.chuang@genesyslogic.com.tw>
+ *
+ * Version: v0.9.0 (2019-08-08)
+ */
+
+#include <linux/bitfield.h>
+#include <linux/bits.h>
+#include <linux/pci.h>
+#include <linux/mmc/mmc.h>
+#include <linux/delay.h>
+#include "sdhci.h"
+#include "sdhci-pci.h"
+#include "cqhci.h"
+
+/*  Genesys Logic extra registers */
+#define SDHCI_GLI_9750_WT         0x800
+#define   SDHCI_GLI_9750_WT_EN      BIT(0)
+#define   GLI_9750_WT_EN_ON	    0x1
+#define   GLI_9750_WT_EN_OFF	    0x0
+
+#define SDHCI_GLI_9750_DRIVING      0x860
+#define   SDHCI_GLI_9750_DRIVING_1    GENMASK(11, 0)
+#define   SDHCI_GLI_9750_DRIVING_2    GENMASK(27, 26)
+#define   GLI_9750_DRIVING_1_VALUE    0xFFF
+#define   GLI_9750_DRIVING_2_VALUE    0x3
+#define   SDHCI_GLI_9750_SEL_1        BIT(29)
+#define   SDHCI_GLI_9750_SEL_2        BIT(31)
+#define   SDHCI_GLI_9750_ALL_RST      (BIT(24)|BIT(25)|BIT(28)|BIT(30))
+
+#define SDHCI_GLI_9750_PLL	      0x864
+#define   SDHCI_GLI_9750_PLL_LDIV       GENMASK(9, 0)
+#define   SDHCI_GLI_9750_PLL_PDIV       GENMASK(14, 12)
+#define   SDHCI_GLI_9750_PLL_DIR        BIT(15)
+#define   SDHCI_GLI_9750_PLL_TX2_INV    BIT(23)
+#define   SDHCI_GLI_9750_PLL_TX2_DLY    GENMASK(22, 20)
+#define   GLI_9750_PLL_TX2_INV_VALUE    0x1
+#define   GLI_9750_PLL_TX2_DLY_VALUE    0x0
+#define   SDHCI_GLI_9750_PLLSSC_STEP    GENMASK(28, 24)
+#define   SDHCI_GLI_9750_PLLSSC_EN      BIT(31)
+
+#define SDHCI_GLI_9750_PLLSSC        0x86C
+#define   SDHCI_GLI_9750_PLLSSC_PPM    GENMASK(31, 16)
+
+#define SDHCI_GLI_9750_SW_CTRL      0x874
+#define   SDHCI_GLI_9750_SW_CTRL_4    GENMASK(7, 6)
+#define   GLI_9750_SW_CTRL_4_VALUE    0x3
+
+#define SDHCI_GLI_9750_MISC            0x878
+#define   SDHCI_GLI_9750_MISC_TX1_INV    BIT(2)
+#define   SDHCI_GLI_9750_MISC_RX_INV     BIT(3)
+#define   SDHCI_GLI_9750_MISC_TX1_DLY    GENMASK(6, 4)
+#define   GLI_9750_MISC_TX1_INV_VALUE    0x0
+#define   GLI_9750_MISC_RX_INV_ON        0x1
+#define   GLI_9750_MISC_RX_INV_OFF       0x0
+#define   GLI_9750_MISC_RX_INV_VALUE     GLI_9750_MISC_RX_INV_OFF
+#define   GLI_9750_MISC_TX1_DLY_VALUE    0x5
+
+#define SDHCI_GLI_9750_TUNING_CONTROL	          0x540
+#define   SDHCI_GLI_9750_TUNING_CONTROL_EN          BIT(4)
+#define   GLI_9750_TUNING_CONTROL_EN_ON             0x1
+#define   GLI_9750_TUNING_CONTROL_EN_OFF            0x0
+#define   SDHCI_GLI_9750_TUNING_CONTROL_GLITCH_1    BIT(16)
+#define   SDHCI_GLI_9750_TUNING_CONTROL_GLITCH_2    GENMASK(20, 19)
+#define   GLI_9750_TUNING_CONTROL_GLITCH_1_VALUE    0x1
+#define   GLI_9750_TUNING_CONTROL_GLITCH_2_VALUE    0x2
+
+#define SDHCI_GLI_9750_TUNING_PARAMETERS           0x544
+#define   SDHCI_GLI_9750_TUNING_PARAMETERS_RX_DLY    GENMASK(2, 0)
+#define   GLI_9750_TUNING_PARAMETERS_RX_DLY_VALUE    0x1
+
+#define SDHCI_GLI_9763E_CTRL_HS400  0x7
+
+#define SDHCI_GLI_9763E_HS400_ES_REG      0x52C
+#define   SDHCI_GLI_9763E_HS400_ES_BIT      BIT(8)
+
+#define PCIE_GLI_9763E_VHS	 0x884
+#define   GLI_9763E_VHS_REV	   GENMASK(19, 16)
+#define   GLI_9763E_VHS_REV_R      0x0
+#define   GLI_9763E_VHS_REV_M      0x1
+#define   GLI_9763E_VHS_REV_W      0x2
+#define PCIE_GLI_9763E_MB	 0x888
+#define   GLI_9763E_MB_CMDQ_OFF	   BIT(19)
+#define PCIE_GLI_9763E_SCR	 0x8E0
+#define   GLI_9763E_SCR_AXI_REQ	   BIT(9)
+
+#define SDHCI_GLI_9763E_CQE_BASE_ADDR	 0x200
+#define GLI_9763E_CQE_TRNS_MODE	   (SDHCI_TRNS_MULTI | \
+				    SDHCI_TRNS_BLK_CNT_EN | \
+				    SDHCI_TRNS_DMA)
+
+#define PCI_GLI_9755_WT       0x800
+#define   PCI_GLI_9755_WT_EN    BIT(0)
+#define   GLI_9755_WT_EN_ON     0x1
+#define   GLI_9755_WT_EN_OFF    0x0
+
+#define PCI_GLI_9755_PLL            0x64
+#define   PCI_GLI_9755_PLL_LDIV       GENMASK(9, 0)
+#define   PCI_GLI_9755_PLL_PDIV       GENMASK(14, 12)
+#define   PCI_GLI_9755_PLL_DIR        BIT(15)
+#define   PCI_GLI_9755_PLLSSC_STEP    GENMASK(28, 24)
+#define   PCI_GLI_9755_PLLSSC_EN      BIT(31)
+
+#define PCI_GLI_9755_PLLSSC        0x68
+#define   PCI_GLI_9755_PLLSSC_PPM    GENMASK(15, 0)
+
+#define GLI_MAX_TUNING_LOOP 40
+
+/* Genesys Logic chipset */
+static inline void gl9750_wt_on(struct sdhci_host *host)
+{
+	u32 wt_value;
+	u32 wt_enable;
+
+	wt_value = sdhci_readl(host, SDHCI_GLI_9750_WT);
+	wt_enable = FIELD_GET(SDHCI_GLI_9750_WT_EN, wt_value);
+
+	if (wt_enable == GLI_9750_WT_EN_ON)
+		return;
+
+	wt_value &= ~SDHCI_GLI_9750_WT_EN;
+	wt_value |= FIELD_PREP(SDHCI_GLI_9750_WT_EN, GLI_9750_WT_EN_ON);
+
+	sdhci_writel(host, wt_value, SDHCI_GLI_9750_WT);
+}
+
+static inline void gl9750_wt_off(struct sdhci_host *host)
+{
+	u32 wt_value;
+	u32 wt_enable;
+
+	wt_value = sdhci_readl(host, SDHCI_GLI_9750_WT);
+	wt_enable = FIELD_GET(SDHCI_GLI_9750_WT_EN, wt_value);
+
+	if (wt_enable == GLI_9750_WT_EN_OFF)
+		return;
+
+	wt_value &= ~SDHCI_GLI_9750_WT_EN;
+	wt_value |= FIELD_PREP(SDHCI_GLI_9750_WT_EN, GLI_9750_WT_EN_OFF);
+
+	sdhci_writel(host, wt_value, SDHCI_GLI_9750_WT);
+}
+
+static void gli_set_9750(struct sdhci_host *host)
+{
+	u32 driving_value;
+	u32 pll_value;
+	u32 sw_ctrl_value;
+	u32 misc_value;
+	u32 parameter_value;
+	u32 control_value;
+	u16 ctrl2;
+
+	gl9750_wt_on(host);
+
+	driving_value = sdhci_readl(host, SDHCI_GLI_9750_DRIVING);
+	pll_value = sdhci_readl(host, SDHCI_GLI_9750_PLL);
+	sw_ctrl_value = sdhci_readl(host, SDHCI_GLI_9750_SW_CTRL);
+	misc_value = sdhci_readl(host, SDHCI_GLI_9750_MISC);
+	parameter_value = sdhci_readl(host, SDHCI_GLI_9750_TUNING_PARAMETERS);
+	control_value = sdhci_readl(host, SDHCI_GLI_9750_TUNING_CONTROL);
+
+	driving_value &= ~(SDHCI_GLI_9750_DRIVING_1);
+	driving_value &= ~(SDHCI_GLI_9750_DRIVING_2);
+	driving_value |= FIELD_PREP(SDHCI_GLI_9750_DRIVING_1,
+				    GLI_9750_DRIVING_1_VALUE);
+	driving_value |= FIELD_PREP(SDHCI_GLI_9750_DRIVING_2,
+				    GLI_9750_DRIVING_2_VALUE);
+	driving_value &= ~(SDHCI_GLI_9750_SEL_1|SDHCI_GLI_9750_SEL_2|SDHCI_GLI_9750_ALL_RST);
+	driving_value |= SDHCI_GLI_9750_SEL_2;
+	sdhci_writel(host, driving_value, SDHCI_GLI_9750_DRIVING);
+
+	sw_ctrl_value &= ~SDHCI_GLI_9750_SW_CTRL_4;
+	sw_ctrl_value |= FIELD_PREP(SDHCI_GLI_9750_SW_CTRL_4,
+				    GLI_9750_SW_CTRL_4_VALUE);
+	sdhci_writel(host, sw_ctrl_value, SDHCI_GLI_9750_SW_CTRL);
+
+	/* reset the tuning flow after reinit and before starting tuning */
+	pll_value &= ~SDHCI_GLI_9750_PLL_TX2_INV;
+	pll_value &= ~SDHCI_GLI_9750_PLL_TX2_DLY;
+	pll_value |= FIELD_PREP(SDHCI_GLI_9750_PLL_TX2_INV,
+				GLI_9750_PLL_TX2_INV_VALUE);
+	pll_value |= FIELD_PREP(SDHCI_GLI_9750_PLL_TX2_DLY,
+				GLI_9750_PLL_TX2_DLY_VALUE);
+
+	misc_value &= ~SDHCI_GLI_9750_MISC_TX1_INV;
+	misc_value &= ~SDHCI_GLI_9750_MISC_RX_INV;
+	misc_value &= ~SDHCI_GLI_9750_MISC_TX1_DLY;
+	misc_value |= FIELD_PREP(SDHCI_GLI_9750_MISC_TX1_INV,
+				 GLI_9750_MISC_TX1_INV_VALUE);
+	misc_value |= FIELD_PREP(SDHCI_GLI_9750_MISC_RX_INV,
+				 GLI_9750_MISC_RX_INV_VALUE);
+	misc_value |= FIELD_PREP(SDHCI_GLI_9750_MISC_TX1_DLY,
+				 GLI_9750_MISC_TX1_DLY_VALUE);
+
+	parameter_value &= ~SDHCI_GLI_9750_TUNING_PARAMETERS_RX_DLY;
+	parameter_value |= FIELD_PREP(SDHCI_GLI_9750_TUNING_PARAMETERS_RX_DLY,
+				      GLI_9750_TUNING_PARAMETERS_RX_DLY_VALUE);
+
+	control_value &= ~SDHCI_GLI_9750_TUNING_CONTROL_GLITCH_1;
+	control_value &= ~SDHCI_GLI_9750_TUNING_CONTROL_GLITCH_2;
+	control_value |= FIELD_PREP(SDHCI_GLI_9750_TUNING_CONTROL_GLITCH_1,
+				    GLI_9750_TUNING_CONTROL_GLITCH_1_VALUE);
+	control_value |= FIELD_PREP(SDHCI_GLI_9750_TUNING_CONTROL_GLITCH_2,
+				    GLI_9750_TUNING_CONTROL_GLITCH_2_VALUE);
+
+	sdhci_writel(host, pll_value, SDHCI_GLI_9750_PLL);
+	sdhci_writel(host, misc_value, SDHCI_GLI_9750_MISC);
+
+	/* disable tuned clk */
+	ctrl2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	ctrl2 &= ~SDHCI_CTRL_TUNED_CLK;
+	sdhci_writew(host, ctrl2, SDHCI_HOST_CONTROL2);
+
+	/* enable tuning parameters control */
+	control_value &= ~SDHCI_GLI_9750_TUNING_CONTROL_EN;
+	control_value |= FIELD_PREP(SDHCI_GLI_9750_TUNING_CONTROL_EN,
+				    GLI_9750_TUNING_CONTROL_EN_ON);
+	sdhci_writel(host, control_value, SDHCI_GLI_9750_TUNING_CONTROL);
+
+	/* write tuning parameters */
+	sdhci_writel(host, parameter_value, SDHCI_GLI_9750_TUNING_PARAMETERS);
+
+	/* disable tuning parameters control */
+	control_value &= ~SDHCI_GLI_9750_TUNING_CONTROL_EN;
+	control_value |= FIELD_PREP(SDHCI_GLI_9750_TUNING_CONTROL_EN,
+				    GLI_9750_TUNING_CONTROL_EN_OFF);
+	sdhci_writel(host, control_value, SDHCI_GLI_9750_TUNING_CONTROL);
+
+	/* clear tuned clk */
+	ctrl2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	ctrl2 &= ~SDHCI_CTRL_TUNED_CLK;
+	sdhci_writew(host, ctrl2, SDHCI_HOST_CONTROL2);
+
+	gl9750_wt_off(host);
+}
+
+static void gli_set_9750_rx_inv(struct sdhci_host *host, bool b)
+{
+	u32 misc_value;
+
+	gl9750_wt_on(host);
+
+	misc_value = sdhci_readl(host, SDHCI_GLI_9750_MISC);
+	misc_value &= ~SDHCI_GLI_9750_MISC_RX_INV;
+	if (b) {
+		misc_value |= FIELD_PREP(SDHCI_GLI_9750_MISC_RX_INV,
+					 GLI_9750_MISC_RX_INV_ON);
+	} else {
+		misc_value |= FIELD_PREP(SDHCI_GLI_9750_MISC_RX_INV,
+					 GLI_9750_MISC_RX_INV_OFF);
+	}
+	sdhci_writel(host, misc_value, SDHCI_GLI_9750_MISC);
+
+	gl9750_wt_off(host);
+}
+
+static int __sdhci_execute_tuning_9750(struct sdhci_host *host, u32 opcode)
+{
+	int i;
+	int rx_inv;
+
+	for (rx_inv = 0; rx_inv < 2; rx_inv++) {
+		gli_set_9750_rx_inv(host, !!rx_inv);
+		sdhci_start_tuning(host);
+
+		for (i = 0; i < GLI_MAX_TUNING_LOOP; i++) {
+			u16 ctrl;
+
+			sdhci_send_tuning(host, opcode);
+
+			if (!host->tuning_done) {
+				sdhci_abort_tuning(host, opcode);
+				break;
+			}
+
+			ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+			if (!(ctrl & SDHCI_CTRL_EXEC_TUNING)) {
+				if (ctrl & SDHCI_CTRL_TUNED_CLK)
+					return 0; /* Success! */
+				break;
+			}
+		}
+	}
+	if (!host->tuning_done) {
+		pr_info("%s: Tuning timeout, falling back to fixed sampling clock\n",
+			mmc_hostname(host->mmc));
+		return -ETIMEDOUT;
+	}
+
+	pr_info("%s: Tuning failed, falling back to fixed sampling clock\n",
+		mmc_hostname(host->mmc));
+	sdhci_reset_tuning(host);
+
+	return -EAGAIN;
+}
+
+static int gl9750_execute_tuning(struct sdhci_host *host, u32 opcode)
+{
+	host->mmc->retune_period = 0;
+	if (host->tuning_mode == SDHCI_TUNING_MODE_1)
+		host->mmc->retune_period = host->tuning_count;
+
+	gli_set_9750(host);
+	host->tuning_err = __sdhci_execute_tuning_9750(host, opcode);
+	sdhci_end_tuning(host);
+
+	return 0;
+}
+
+static void gl9750_disable_ssc_pll(struct sdhci_host *host)
+{
+	u32 pll;
+
+	gl9750_wt_on(host);
+	pll = sdhci_readl(host, SDHCI_GLI_9750_PLL);
+	pll &= ~(SDHCI_GLI_9750_PLL_DIR | SDHCI_GLI_9750_PLLSSC_EN);
+	sdhci_writel(host, pll, SDHCI_GLI_9750_PLL);
+	gl9750_wt_off(host);
+}
+
+static void gl9750_set_pll(struct sdhci_host *host, u8 dir, u16 ldiv, u8 pdiv)
+{
+	u32 pll;
+
+	gl9750_wt_on(host);
+	pll = sdhci_readl(host, SDHCI_GLI_9750_PLL);
+	pll &= ~(SDHCI_GLI_9750_PLL_LDIV |
+		 SDHCI_GLI_9750_PLL_PDIV |
+		 SDHCI_GLI_9750_PLL_DIR);
+	pll |= FIELD_PREP(SDHCI_GLI_9750_PLL_LDIV, ldiv) |
+	       FIELD_PREP(SDHCI_GLI_9750_PLL_PDIV, pdiv) |
+	       FIELD_PREP(SDHCI_GLI_9750_PLL_DIR, dir);
+	sdhci_writel(host, pll, SDHCI_GLI_9750_PLL);
+	gl9750_wt_off(host);
+
+	/* wait for pll stable */
+	mdelay(1);
+}
+
+static void gl9750_set_ssc(struct sdhci_host *host, u8 enable, u8 step, u16 ppm)
+{
+	u32 pll;
+	u32 ssc;
+
+	gl9750_wt_on(host);
+	pll = sdhci_readl(host, SDHCI_GLI_9750_PLL);
+	ssc = sdhci_readl(host, SDHCI_GLI_9750_PLLSSC);
+	pll &= ~(SDHCI_GLI_9750_PLLSSC_STEP |
+		 SDHCI_GLI_9750_PLLSSC_EN);
+	ssc &= ~SDHCI_GLI_9750_PLLSSC_PPM;
+	pll |= FIELD_PREP(SDHCI_GLI_9750_PLLSSC_STEP, step) |
+	       FIELD_PREP(SDHCI_GLI_9750_PLLSSC_EN, enable);
+	ssc |= FIELD_PREP(SDHCI_GLI_9750_PLLSSC_PPM, ppm);
+	sdhci_writel(host, ssc, SDHCI_GLI_9750_PLLSSC);
+	sdhci_writel(host, pll, SDHCI_GLI_9750_PLL);
+	gl9750_wt_off(host);
+}
+
+static void gl9750_set_ssc_pll_205mhz(struct sdhci_host *host)
+{
+	/* set pll to 205MHz and enable ssc */
+	gl9750_set_ssc(host, 0x1, 0x1F, 0xFFE7);
+	gl9750_set_pll(host, 0x1, 0x246, 0x0);
+}
+
+static void sdhci_gl9750_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	struct mmc_ios *ios = &host->mmc->ios;
+	u16 clk;
+
+	host->mmc->actual_clock = 0;
+
+	gl9750_disable_ssc_pll(host);
+	sdhci_writew(host, 0, SDHCI_CLOCK_CONTROL);
+
+	if (clock == 0)
+		return;
+
+	clk = sdhci_calc_clk(host, clock, &host->mmc->actual_clock);
+	if (clock == 200000000 && ios->timing == MMC_TIMING_UHS_SDR104) {
+		host->mmc->actual_clock = 205000000;
+		gl9750_set_ssc_pll_205mhz(host);
+	}
+
+	sdhci_enable_clk(host, clk);
+}
+
+static void gli_pcie_enable_msi(struct sdhci_pci_slot *slot)
+{
+	int ret;
+
+	ret = pci_alloc_irq_vectors(slot->chip->pdev, 1, 1,
+				    PCI_IRQ_MSI | PCI_IRQ_MSIX);
+	if (ret < 0) {
+		pr_warn("%s: enable PCI MSI failed, error=%d\n",
+		       mmc_hostname(slot->host->mmc), ret);
+		return;
+	}
+
+	slot->host->irq = pci_irq_vector(slot->chip->pdev, 0);
+}
+
+static inline void gl9755_wt_on(struct pci_dev *pdev)
+{
+	u32 wt_value;
+	u32 wt_enable;
+
+	pci_read_config_dword(pdev, PCI_GLI_9755_WT, &wt_value);
+	wt_enable = FIELD_GET(PCI_GLI_9755_WT_EN, wt_value);
+
+	if (wt_enable == GLI_9755_WT_EN_ON)
+		return;
+
+	wt_value &= ~PCI_GLI_9755_WT_EN;
+	wt_value |= FIELD_PREP(PCI_GLI_9755_WT_EN, GLI_9755_WT_EN_ON);
+
+	pci_write_config_dword(pdev, PCI_GLI_9755_WT, wt_value);
+}
+
+static inline void gl9755_wt_off(struct pci_dev *pdev)
+{
+	u32 wt_value;
+	u32 wt_enable;
+
+	pci_read_config_dword(pdev, PCI_GLI_9755_WT, &wt_value);
+	wt_enable = FIELD_GET(PCI_GLI_9755_WT_EN, wt_value);
+
+	if (wt_enable == GLI_9755_WT_EN_OFF)
+		return;
+
+	wt_value &= ~PCI_GLI_9755_WT_EN;
+	wt_value |= FIELD_PREP(PCI_GLI_9755_WT_EN, GLI_9755_WT_EN_OFF);
+
+	pci_write_config_dword(pdev, PCI_GLI_9755_WT, wt_value);
+}
+
+static void gl9755_disable_ssc_pll(struct pci_dev *pdev)
+{
+	u32 pll;
+
+	gl9755_wt_on(pdev);
+	pci_read_config_dword(pdev, PCI_GLI_9755_PLL, &pll);
+	pll &= ~(PCI_GLI_9755_PLL_DIR | PCI_GLI_9755_PLLSSC_EN);
+	pci_write_config_dword(pdev, PCI_GLI_9755_PLL, pll);
+	gl9755_wt_off(pdev);
+}
+
+static void gl9755_set_pll(struct pci_dev *pdev, u8 dir, u16 ldiv, u8 pdiv)
+{
+	u32 pll;
+
+	gl9755_wt_on(pdev);
+	pci_read_config_dword(pdev, PCI_GLI_9755_PLL, &pll);
+	pll &= ~(PCI_GLI_9755_PLL_LDIV |
+		 PCI_GLI_9755_PLL_PDIV |
+		 PCI_GLI_9755_PLL_DIR);
+	pll |= FIELD_PREP(PCI_GLI_9755_PLL_LDIV, ldiv) |
+	       FIELD_PREP(PCI_GLI_9755_PLL_PDIV, pdiv) |
+	       FIELD_PREP(PCI_GLI_9755_PLL_DIR, dir);
+	pci_write_config_dword(pdev, PCI_GLI_9755_PLL, pll);
+	gl9755_wt_off(pdev);
+
+	/* wait for pll stable */
+	mdelay(1);
+}
+
+static void gl9755_set_ssc(struct pci_dev *pdev, u8 enable, u8 step, u16 ppm)
+{
+	u32 pll;
+	u32 ssc;
+
+	gl9755_wt_on(pdev);
+	pci_read_config_dword(pdev, PCI_GLI_9755_PLL, &pll);
+	pci_read_config_dword(pdev, PCI_GLI_9755_PLLSSC, &ssc);
+	pll &= ~(PCI_GLI_9755_PLLSSC_STEP |
+		 PCI_GLI_9755_PLLSSC_EN);
+	ssc &= ~PCI_GLI_9755_PLLSSC_PPM;
+	pll |= FIELD_PREP(PCI_GLI_9755_PLLSSC_STEP, step) |
+	       FIELD_PREP(PCI_GLI_9755_PLLSSC_EN, enable);
+	ssc |= FIELD_PREP(PCI_GLI_9755_PLLSSC_PPM, ppm);
+	pci_write_config_dword(pdev, PCI_GLI_9755_PLLSSC, ssc);
+	pci_write_config_dword(pdev, PCI_GLI_9755_PLL, pll);
+	gl9755_wt_off(pdev);
+}
+
+static void gl9755_set_ssc_pll_205mhz(struct pci_dev *pdev)
+{
+	/* set pll to 205MHz and enable ssc */
+	gl9755_set_ssc(pdev, 0x1, 0x1F, 0xFFE7);
+	gl9755_set_pll(pdev, 0x1, 0x246, 0x0);
+}
+
+static void sdhci_gl9755_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	struct sdhci_pci_slot *slot = sdhci_priv(host);
+	struct mmc_ios *ios = &host->mmc->ios;
+	struct pci_dev *pdev;
+	u16 clk;
+
+	pdev = slot->chip->pdev;
+	host->mmc->actual_clock = 0;
+
+	gl9755_disable_ssc_pll(pdev);
+	sdhci_writew(host, 0, SDHCI_CLOCK_CONTROL);
+
+	if (clock == 0)
+		return;
+
+	clk = sdhci_calc_clk(host, clock, &host->mmc->actual_clock);
+	if (clock == 200000000 && ios->timing == MMC_TIMING_UHS_SDR104) {
+		host->mmc->actual_clock = 205000000;
+		gl9755_set_ssc_pll_205mhz(pdev);
+	}
+
+	sdhci_enable_clk(host, clk);
+}
+
+static int gli_probe_slot_gl9750(struct sdhci_pci_slot *slot)
+{
+	struct sdhci_host *host = slot->host;
+
+	gli_pcie_enable_msi(slot);
+	slot->host->mmc->caps2 |= MMC_CAP2_NO_SDIO;
+	sdhci_enable_v4_mode(host);
+
+	return 0;
+}
+
+static int gli_probe_slot_gl9755(struct sdhci_pci_slot *slot)
+{
+	struct sdhci_host *host = slot->host;
+
+	gli_pcie_enable_msi(slot);
+	slot->host->mmc->caps2 |= MMC_CAP2_NO_SDIO;
+	sdhci_enable_v4_mode(host);
+
+	return 0;
+}
+
+static void sdhci_gli_voltage_switch(struct sdhci_host *host)
+{
+	/*
+	 * According to Section 3.6.1 signal voltage switch procedure in
+	 * SD Host Controller Simplified Spec. 4.20, steps 6~8 are as
+	 * follows:
+	 * (6) Set 1.8V Signal Enable in the Host Control 2 register.
+	 * (7) Wait 5ms. 1.8V voltage regulator shall be stable within this
+	 *     period.
+	 * (8) If 1.8V Signal Enable is cleared by Host Controller, go to
+	 *     step (12).
+	 *
+	 * Wait 5ms after set 1.8V signal enable in Host Control 2 register
+	 * to ensure 1.8V signal enable bit is set by GL9750/GL9755.
+	 *
+	 * ...however, the controller in the NUC10i3FNK4 (a 9755) requires
+	 * slightly longer than 5ms before the control register reports that
+	 * 1.8V is ready, and far longer still before the card will actually
+	 * work reliably.
+	 */
+	usleep_range(100000, 110000);
+}
+
+static void sdhci_gl9750_reset(struct sdhci_host *host, u8 mask)
+{
+	sdhci_reset(host, mask);
+	gli_set_9750(host);
+}
+
+static u32 sdhci_gl9750_readl(struct sdhci_host *host, int reg)
+{
+	u32 value;
+
+	value = readl(host->ioaddr + reg);
+	if (unlikely(reg == SDHCI_MAX_CURRENT && !(value & 0xff)))
+		value |= 0xc8;
+
+	return value;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int sdhci_pci_gli_resume(struct sdhci_pci_chip *chip)
+{
+	struct sdhci_pci_slot *slot = chip->slots[0];
+
+	pci_free_irq_vectors(slot->chip->pdev);
+	gli_pcie_enable_msi(slot);
+
+	return sdhci_pci_resume_host(chip);
+}
+
+static int sdhci_cqhci_gli_resume(struct sdhci_pci_chip *chip)
+{
+	struct sdhci_pci_slot *slot = chip->slots[0];
+	int ret;
+
+	ret = sdhci_pci_gli_resume(chip);
+	if (ret)
+		return ret;
+
+	return cqhci_resume(slot->host->mmc);
+}
+
+static int sdhci_cqhci_gli_suspend(struct sdhci_pci_chip *chip)
+{
+	struct sdhci_pci_slot *slot = chip->slots[0];
+	int ret;
+
+	ret = cqhci_suspend(slot->host->mmc);
+	if (ret)
+		return ret;
+
+	return sdhci_suspend_host(slot->host);
+}
+#endif
+
+static void gl9763e_hs400_enhanced_strobe(struct mmc_host *mmc,
+					  struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	u32 val;
+
+	val = sdhci_readl(host, SDHCI_GLI_9763E_HS400_ES_REG);
+	if (ios->enhanced_strobe)
+		val |= SDHCI_GLI_9763E_HS400_ES_BIT;
+	else
+		val &= ~SDHCI_GLI_9763E_HS400_ES_BIT;
+
+	sdhci_writel(host, val, SDHCI_GLI_9763E_HS400_ES_REG);
+}
+
+static void sdhci_set_gl9763e_signaling(struct sdhci_host *host,
+					unsigned int timing)
+{
+	u16 ctrl_2;
+
+	ctrl_2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	ctrl_2 &= ~SDHCI_CTRL_UHS_MASK;
+	if (timing == MMC_TIMING_MMC_HS200)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR104;
+	else if (timing == MMC_TIMING_MMC_HS)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR25;
+	else if (timing == MMC_TIMING_MMC_DDR52)
+		ctrl_2 |= SDHCI_CTRL_UHS_DDR50;
+	else if (timing == MMC_TIMING_MMC_HS400)
+		ctrl_2 |= SDHCI_GLI_9763E_CTRL_HS400;
+
+	sdhci_writew(host, ctrl_2, SDHCI_HOST_CONTROL2);
+}
+
+static void sdhci_gl9763e_dumpregs(struct mmc_host *mmc)
+{
+	sdhci_dumpregs(mmc_priv(mmc));
+}
+
+static void sdhci_gl9763e_cqe_pre_enable(struct mmc_host *mmc)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	u32 value;
+
+	value = cqhci_readl(cq_host, CQHCI_CFG);
+	value |= CQHCI_ENABLE;
+	cqhci_writel(cq_host, value, CQHCI_CFG);
+}
+
+static void sdhci_gl9763e_cqe_enable(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	sdhci_writew(host, GLI_9763E_CQE_TRNS_MODE, SDHCI_TRANSFER_MODE);
+	sdhci_cqe_enable(mmc);
+}
+
+static u32 sdhci_gl9763e_cqhci_irq(struct sdhci_host *host, u32 intmask)
+{
+	int cmd_error = 0;
+	int data_error = 0;
+
+	if (!sdhci_cqe_irq(host, intmask, &cmd_error, &data_error))
+		return intmask;
+
+	cqhci_irq(host->mmc, intmask, cmd_error, data_error);
+
+	return 0;
+}
+
+static void sdhci_gl9763e_cqe_post_disable(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	u32 value;
+
+	value = cqhci_readl(cq_host, CQHCI_CFG);
+	value &= ~CQHCI_ENABLE;
+	cqhci_writel(cq_host, value, CQHCI_CFG);
+	sdhci_writew(host, 0x0, SDHCI_TRANSFER_MODE);
+}
+
+static const struct cqhci_host_ops sdhci_gl9763e_cqhci_ops = {
+	.enable         = sdhci_gl9763e_cqe_enable,
+	.disable        = sdhci_cqe_disable,
+	.dumpregs       = sdhci_gl9763e_dumpregs,
+	.pre_enable     = sdhci_gl9763e_cqe_pre_enable,
+	.post_disable   = sdhci_gl9763e_cqe_post_disable,
+};
+
+static int gl9763e_add_host(struct sdhci_pci_slot *slot)
+{
+	struct device *dev = &slot->chip->pdev->dev;
+	struct sdhci_host *host = slot->host;
+	struct cqhci_host *cq_host;
+	bool dma64;
+	int ret;
+
+	ret = sdhci_setup_host(host);
+	if (ret)
+		return ret;
+
+	cq_host = devm_kzalloc(dev, sizeof(*cq_host), GFP_KERNEL);
+	if (!cq_host) {
+		ret = -ENOMEM;
+		goto cleanup;
+	}
+
+	cq_host->mmio = host->ioaddr + SDHCI_GLI_9763E_CQE_BASE_ADDR;
+	cq_host->ops = &sdhci_gl9763e_cqhci_ops;
+
+	dma64 = host->flags & SDHCI_USE_64_BIT_DMA;
+	if (dma64)
+		cq_host->caps |= CQHCI_TASK_DESC_SZ_128;
+
+	ret = cqhci_init(cq_host, host->mmc, dma64);
+	if (ret)
+		goto cleanup;
+
+	ret = __sdhci_add_host(host);
+	if (ret)
+		goto cleanup;
+
+	return 0;
+
+cleanup:
+	sdhci_cleanup_host(host);
+	return ret;
+}
+
+static void sdhci_gl9763e_reset(struct sdhci_host *host, u8 mask)
+{
+	if ((host->mmc->caps2 & MMC_CAP2_CQE) && (mask & SDHCI_RESET_ALL) &&
+	    host->mmc->cqe_private)
+		cqhci_deactivate(host->mmc);
+	sdhci_reset(host, mask);
+}
+
+static void gli_set_gl9763e(struct sdhci_pci_slot *slot)
+{
+	struct pci_dev *pdev = slot->chip->pdev;
+	u32 value;
+
+	pci_read_config_dword(pdev, PCIE_GLI_9763E_VHS, &value);
+	value &= ~GLI_9763E_VHS_REV;
+	value |= FIELD_PREP(GLI_9763E_VHS_REV, GLI_9763E_VHS_REV_W);
+	pci_write_config_dword(pdev, PCIE_GLI_9763E_VHS, value);
+
+	pci_read_config_dword(pdev, PCIE_GLI_9763E_SCR, &value);
+	value |= GLI_9763E_SCR_AXI_REQ;
+	pci_write_config_dword(pdev, PCIE_GLI_9763E_SCR, value);
+
+	pci_read_config_dword(pdev, PCIE_GLI_9763E_VHS, &value);
+	value &= ~GLI_9763E_VHS_REV;
+	value |= FIELD_PREP(GLI_9763E_VHS_REV, GLI_9763E_VHS_REV_R);
+	pci_write_config_dword(pdev, PCIE_GLI_9763E_VHS, value);
+}
+
+static int gli_probe_slot_gl9763e(struct sdhci_pci_slot *slot)
+{
+	struct pci_dev *pdev = slot->chip->pdev;
+	struct sdhci_host *host = slot->host;
+	u32 value;
+
+	host->mmc->caps |= MMC_CAP_8_BIT_DATA |
+			   MMC_CAP_1_8V_DDR |
+			   MMC_CAP_NONREMOVABLE;
+	host->mmc->caps2 |= MMC_CAP2_HS200_1_8V_SDR |
+			    MMC_CAP2_HS400_1_8V |
+			    MMC_CAP2_HS400_ES |
+			    MMC_CAP2_NO_SDIO |
+			    MMC_CAP2_NO_SD;
+
+	pci_read_config_dword(pdev, PCIE_GLI_9763E_MB, &value);
+	if (!(value & GLI_9763E_MB_CMDQ_OFF))
+		host->mmc->caps2 |= MMC_CAP2_CQE | MMC_CAP2_CQE_DCMD;
+
+	gli_pcie_enable_msi(slot);
+	host->mmc_host_ops.hs400_enhanced_strobe =
+					gl9763e_hs400_enhanced_strobe;
+	gli_set_gl9763e(slot);
+	sdhci_enable_v4_mode(host);
+
+	return 0;
+}
+
+static const struct sdhci_ops sdhci_gl9755_ops = {
+	.set_clock		= sdhci_gl9755_set_clock,
+	.enable_dma		= sdhci_pci_enable_dma,
+	.set_bus_width		= sdhci_set_bus_width,
+	.reset			= sdhci_reset,
+	.set_uhs_signaling	= sdhci_set_uhs_signaling,
+	.voltage_switch		= sdhci_gli_voltage_switch,
+};
+
+const struct sdhci_pci_fixes sdhci_gl9755 = {
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.quirks2	= SDHCI_QUIRK2_BROKEN_DDR50,
+	.probe_slot	= gli_probe_slot_gl9755,
+	.ops            = &sdhci_gl9755_ops,
+#ifdef CONFIG_PM_SLEEP
+	.resume         = sdhci_pci_gli_resume,
+#endif
+};
+
+static const struct sdhci_ops sdhci_gl9750_ops = {
+	.read_l                 = sdhci_gl9750_readl,
+	.set_clock		= sdhci_gl9750_set_clock,
+	.enable_dma		= sdhci_pci_enable_dma,
+	.set_bus_width		= sdhci_set_bus_width,
+	.reset			= sdhci_gl9750_reset,
+	.set_uhs_signaling	= sdhci_set_uhs_signaling,
+	.voltage_switch		= sdhci_gli_voltage_switch,
+	.platform_execute_tuning = gl9750_execute_tuning,
+};
+
+const struct sdhci_pci_fixes sdhci_gl9750 = {
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.quirks2	= SDHCI_QUIRK2_BROKEN_DDR50,
+	.probe_slot	= gli_probe_slot_gl9750,
+	.ops            = &sdhci_gl9750_ops,
+#ifdef CONFIG_PM_SLEEP
+	.resume         = sdhci_pci_gli_resume,
+#endif
+};
+
+static const struct sdhci_ops sdhci_gl9763e_ops = {
+	.set_clock		= sdhci_set_clock,
+	.enable_dma		= sdhci_pci_enable_dma,
+	.set_bus_width		= sdhci_set_bus_width,
+	.reset			= sdhci_gl9763e_reset,
+	.set_uhs_signaling	= sdhci_set_gl9763e_signaling,
+	.voltage_switch		= sdhci_gli_voltage_switch,
+	.irq                    = sdhci_gl9763e_cqhci_irq,
+};
+
+const struct sdhci_pci_fixes sdhci_gl9763e = {
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.probe_slot	= gli_probe_slot_gl9763e,
+	.ops            = &sdhci_gl9763e_ops,
+#ifdef CONFIG_PM_SLEEP
+	.resume		= sdhci_cqhci_gli_resume,
+	.suspend	= sdhci_cqhci_gli_suspend,
+#endif
+	.add_host       = gl9763e_add_host,
+};
diff --git a/drivers/mmc/host/sdhci-pci-o2micro.c b/drivers/mmc/host/sdhci-pci-o2micro.c
new file mode 100644
index 000000000..72234790a
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pci-o2micro.c
@@ -0,0 +1,871 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2013 BayHub Technology Ltd.
+ *
+ * Authors: Peter Guo <peter.guo@bayhubtech.com>
+ *          Adam Lee <adam.lee@canonical.com>
+ *          Ernest Zhang <ernest.zhang@bayhubtech.com>
+ */
+
+#include <linux/pci.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/delay.h>
+#include <linux/iopoll.h>
+
+#include "sdhci.h"
+#include "sdhci-pci.h"
+
+/*
+ * O2Micro device registers
+ */
+
+#define O2_SD_MISC_REG5		0x64
+#define O2_SD_LD0_CTRL		0x68
+#define O2_SD_DEV_CTRL		0x88
+#define O2_SD_LOCK_WP		0xD3
+#define O2_SD_TEST_REG		0xD4
+#define O2_SD_FUNC_REG0		0xDC
+#define O2_SD_MULTI_VCC3V	0xEE
+#define O2_SD_CLKREQ		0xEC
+#define O2_SD_CAPS		0xE0
+#define O2_SD_ADMA1		0xE2
+#define O2_SD_ADMA2		0xE7
+#define O2_SD_MISC_CTRL2	0xF0
+#define O2_SD_INF_MOD		0xF1
+#define O2_SD_MISC_CTRL4	0xFC
+#define O2_SD_MISC_CTRL		0x1C0
+#define O2_SD_PWR_FORCE_L0	0x0002
+#define O2_SD_TUNING_CTRL	0x300
+#define O2_SD_PLL_SETTING	0x304
+#define O2_SD_MISC_SETTING	0x308
+#define O2_SD_CLK_SETTING	0x328
+#define O2_SD_CAP_REG2		0x330
+#define O2_SD_CAP_REG0		0x334
+#define O2_SD_UHS1_CAP_SETTING	0x33C
+#define O2_SD_DELAY_CTRL	0x350
+#define O2_SD_UHS2_L1_CTRL	0x35C
+#define O2_SD_FUNC_REG3		0x3E0
+#define O2_SD_FUNC_REG4		0x3E4
+#define O2_SD_LED_ENABLE	BIT(6)
+#define O2_SD_FREG0_LEDOFF	BIT(13)
+#define O2_SD_FREG4_ENABLE_CLK_SET	BIT(22)
+
+#define O2_SD_VENDOR_SETTING	0x110
+#define O2_SD_VENDOR_SETTING2	0x1C8
+#define O2_SD_HW_TUNING_DISABLE	BIT(4)
+
+#define O2_PLL_DLL_WDT_CONTROL1	0x1CC
+#define  O2_PLL_FORCE_ACTIVE	BIT(18)
+#define  O2_PLL_LOCK_STATUS	BIT(14)
+#define  O2_PLL_SOFT_RESET	BIT(12)
+#define  O2_DLL_LOCK_STATUS	BIT(11)
+
+#define O2_SD_DETECT_SETTING 0x324
+
+static const u32 dmdn_table[] = {0x2B1C0000,
+	0x2C1A0000, 0x371B0000, 0x35100000};
+#define DMDN_SZ ARRAY_SIZE(dmdn_table)
+
+struct o2_host {
+	u8 dll_adjust_count;
+};
+
+static void sdhci_o2_wait_card_detect_stable(struct sdhci_host *host)
+{
+	ktime_t timeout;
+	u32 scratch32;
+
+	/* Wait max 50 ms */
+	timeout = ktime_add_ms(ktime_get(), 50);
+	while (1) {
+		bool timedout = ktime_after(ktime_get(), timeout);
+
+		scratch32 = sdhci_readl(host, SDHCI_PRESENT_STATE);
+		if ((scratch32 & SDHCI_CARD_PRESENT) >> SDHCI_CARD_PRES_SHIFT
+		    == (scratch32 & SDHCI_CD_LVL) >> SDHCI_CD_LVL_SHIFT)
+			break;
+
+		if (timedout) {
+			pr_err("%s: Card Detect debounce never finished.\n",
+			       mmc_hostname(host->mmc));
+			sdhci_dumpregs(host);
+			return;
+		}
+		udelay(10);
+	}
+}
+
+static void sdhci_o2_enable_internal_clock(struct sdhci_host *host)
+{
+	ktime_t timeout;
+	u16 scratch;
+	u32 scratch32;
+
+	/* PLL software reset */
+	scratch32 = sdhci_readl(host, O2_PLL_DLL_WDT_CONTROL1);
+	scratch32 |= O2_PLL_SOFT_RESET;
+	sdhci_writel(host, scratch32, O2_PLL_DLL_WDT_CONTROL1);
+	udelay(1);
+	scratch32 &= ~(O2_PLL_SOFT_RESET);
+	sdhci_writel(host, scratch32, O2_PLL_DLL_WDT_CONTROL1);
+
+	/* PLL force active */
+	scratch32 |= O2_PLL_FORCE_ACTIVE;
+	sdhci_writel(host, scratch32, O2_PLL_DLL_WDT_CONTROL1);
+
+	/* Wait max 20 ms */
+	timeout = ktime_add_ms(ktime_get(), 20);
+	while (1) {
+		bool timedout = ktime_after(ktime_get(), timeout);
+
+		scratch = sdhci_readw(host, O2_PLL_DLL_WDT_CONTROL1);
+		if (scratch & O2_PLL_LOCK_STATUS)
+			break;
+		if (timedout) {
+			pr_err("%s: Internal clock never stabilised.\n",
+			       mmc_hostname(host->mmc));
+			sdhci_dumpregs(host);
+			goto out;
+		}
+		udelay(10);
+	}
+
+	/* Wait for card detect finish */
+	udelay(1);
+	sdhci_o2_wait_card_detect_stable(host);
+
+out:
+	/* Cancel PLL force active */
+	scratch32 = sdhci_readl(host, O2_PLL_DLL_WDT_CONTROL1);
+	scratch32 &= ~O2_PLL_FORCE_ACTIVE;
+	sdhci_writel(host, scratch32, O2_PLL_DLL_WDT_CONTROL1);
+}
+
+static int sdhci_o2_get_cd(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	if (!(sdhci_readw(host, O2_PLL_DLL_WDT_CONTROL1) & O2_PLL_LOCK_STATUS))
+		sdhci_o2_enable_internal_clock(host);
+	else
+		sdhci_o2_wait_card_detect_stable(host);
+
+	return !!(sdhci_readl(host, SDHCI_PRESENT_STATE) & SDHCI_CARD_PRESENT);
+}
+
+static void o2_pci_set_baseclk(struct sdhci_pci_chip *chip, u32 value)
+{
+	u32 scratch_32;
+
+	pci_read_config_dword(chip->pdev,
+			      O2_SD_PLL_SETTING, &scratch_32);
+
+	scratch_32 &= 0x0000FFFF;
+	scratch_32 |= value;
+
+	pci_write_config_dword(chip->pdev,
+			       O2_SD_PLL_SETTING, scratch_32);
+}
+
+static u32 sdhci_o2_pll_dll_wdt_control(struct sdhci_host *host)
+{
+	return sdhci_readl(host, O2_PLL_DLL_WDT_CONTROL1);
+}
+
+/*
+ * This function is used to detect dll lock status.
+ * Since the dll lock status bit will toggle randomly
+ * with very short interval which needs to be polled
+ * as fast as possible. Set sleep_us as 1 microsecond.
+ */
+static int sdhci_o2_wait_dll_detect_lock(struct sdhci_host *host)
+{
+	u32	scratch32 = 0;
+
+	return readx_poll_timeout(sdhci_o2_pll_dll_wdt_control, host,
+		scratch32, !(scratch32 & O2_DLL_LOCK_STATUS), 1, 1000000);
+}
+
+static void sdhci_o2_set_tuning_mode(struct sdhci_host *host)
+{
+	u16 reg;
+
+	/* enable hardware tuning */
+	reg = sdhci_readw(host, O2_SD_VENDOR_SETTING);
+	reg &= ~O2_SD_HW_TUNING_DISABLE;
+	sdhci_writew(host, reg, O2_SD_VENDOR_SETTING);
+}
+
+static void __sdhci_o2_execute_tuning(struct sdhci_host *host, u32 opcode)
+{
+	int i;
+
+	sdhci_send_tuning(host, opcode);
+
+	for (i = 0; i < 150; i++) {
+		u16 ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+
+		if (!(ctrl & SDHCI_CTRL_EXEC_TUNING)) {
+			if (ctrl & SDHCI_CTRL_TUNED_CLK) {
+				host->tuning_done = true;
+				return;
+			}
+			pr_warn("%s: HW tuning failed !\n",
+				mmc_hostname(host->mmc));
+			break;
+		}
+
+		mdelay(1);
+	}
+
+	pr_info("%s: Tuning failed, falling back to fixed sampling clock\n",
+		mmc_hostname(host->mmc));
+	sdhci_reset_tuning(host);
+}
+
+/*
+ * This function is used to fix o2 dll shift issue.
+ * It isn't necessary to detect card present before recovery.
+ * Firstly, it is used by bht emmc card, which is embedded.
+ * Second, before call recovery card present will be detected
+ * outside of the execute tuning function.
+ */
+static int sdhci_o2_dll_recovery(struct sdhci_host *host)
+{
+	int ret = 0;
+	u8 scratch_8 = 0;
+	u32 scratch_32 = 0;
+	struct sdhci_pci_slot *slot = sdhci_priv(host);
+	struct sdhci_pci_chip *chip = slot->chip;
+	struct o2_host *o2_host = sdhci_pci_priv(slot);
+
+	/* UnLock WP */
+	pci_read_config_byte(chip->pdev,
+			O2_SD_LOCK_WP, &scratch_8);
+	scratch_8 &= 0x7f;
+	pci_write_config_byte(chip->pdev, O2_SD_LOCK_WP, scratch_8);
+	while (o2_host->dll_adjust_count < DMDN_SZ && !ret) {
+		/* Disable clock */
+		sdhci_writeb(host, 0, SDHCI_CLOCK_CONTROL);
+
+		/* PLL software reset */
+		scratch_32 = sdhci_readl(host, O2_PLL_DLL_WDT_CONTROL1);
+		scratch_32 |= O2_PLL_SOFT_RESET;
+		sdhci_writel(host, scratch_32, O2_PLL_DLL_WDT_CONTROL1);
+
+		pci_read_config_dword(chip->pdev,
+					    O2_SD_FUNC_REG4,
+					    &scratch_32);
+		/* Enable Base Clk setting change */
+		scratch_32 |= O2_SD_FREG4_ENABLE_CLK_SET;
+		pci_write_config_dword(chip->pdev, O2_SD_FUNC_REG4, scratch_32);
+		o2_pci_set_baseclk(chip, dmdn_table[o2_host->dll_adjust_count]);
+
+		/* Enable internal clock */
+		scratch_8 = SDHCI_CLOCK_INT_EN;
+		sdhci_writeb(host, scratch_8, SDHCI_CLOCK_CONTROL);
+
+		if (sdhci_o2_get_cd(host->mmc)) {
+			/*
+			 * need wait at least 5ms for dll status stable,
+			 * after enable internal clock
+			 */
+			usleep_range(5000, 6000);
+			if (sdhci_o2_wait_dll_detect_lock(host)) {
+				scratch_8 |= SDHCI_CLOCK_CARD_EN;
+				sdhci_writeb(host, scratch_8,
+					SDHCI_CLOCK_CONTROL);
+				ret = 1;
+			} else {
+				pr_warn("%s: DLL unlocked when dll_adjust_count is %d.\n",
+					mmc_hostname(host->mmc),
+					o2_host->dll_adjust_count);
+			}
+		} else {
+			pr_err("%s: card present detect failed.\n",
+				mmc_hostname(host->mmc));
+			break;
+		}
+
+		o2_host->dll_adjust_count++;
+	}
+	if (!ret && o2_host->dll_adjust_count == DMDN_SZ)
+		pr_err("%s: DLL adjust over max times\n",
+		mmc_hostname(host->mmc));
+	/* Lock WP */
+	pci_read_config_byte(chip->pdev,
+				   O2_SD_LOCK_WP, &scratch_8);
+	scratch_8 |= 0x80;
+	pci_write_config_byte(chip->pdev, O2_SD_LOCK_WP, scratch_8);
+	return ret;
+}
+
+static int sdhci_o2_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	int current_bus_width = 0;
+	u32 scratch32 = 0;
+	u16 scratch = 0;
+
+	/*
+	 * This handler only implements the eMMC tuning that is specific to
+	 * this controller.  Fall back to the standard method for other TIMING.
+	 */
+	if ((host->timing != MMC_TIMING_MMC_HS200) &&
+		(host->timing != MMC_TIMING_UHS_SDR104))
+		return sdhci_execute_tuning(mmc, opcode);
+
+	if (WARN_ON((opcode != MMC_SEND_TUNING_BLOCK_HS200) &&
+			(opcode != MMC_SEND_TUNING_BLOCK)))
+		return -EINVAL;
+
+	/* Force power mode enter L0 */
+	scratch = sdhci_readw(host, O2_SD_MISC_CTRL);
+	scratch |= O2_SD_PWR_FORCE_L0;
+	sdhci_writew(host, scratch, O2_SD_MISC_CTRL);
+
+	/* wait DLL lock, timeout value 5ms */
+	if (readx_poll_timeout(sdhci_o2_pll_dll_wdt_control, host,
+		scratch32, (scratch32 & O2_DLL_LOCK_STATUS), 1, 5000))
+		pr_warn("%s: DLL can't lock in 5ms after force L0 during tuning.\n",
+				mmc_hostname(host->mmc));
+	/*
+	 * Judge the tuning reason, whether caused by dll shift
+	 * If cause by dll shift, should call sdhci_o2_dll_recovery
+	 */
+	if (!sdhci_o2_wait_dll_detect_lock(host))
+		if (!sdhci_o2_dll_recovery(host)) {
+			pr_err("%s: o2 dll recovery failed\n",
+				mmc_hostname(host->mmc));
+			return -EINVAL;
+		}
+	/*
+	 * o2 sdhci host didn't support 8bit emmc tuning
+	 */
+	if (mmc->ios.bus_width == MMC_BUS_WIDTH_8) {
+		current_bus_width = mmc->ios.bus_width;
+		mmc->ios.bus_width = MMC_BUS_WIDTH_4;
+		sdhci_set_bus_width(host, MMC_BUS_WIDTH_4);
+	}
+
+	sdhci_o2_set_tuning_mode(host);
+
+	sdhci_start_tuning(host);
+
+	__sdhci_o2_execute_tuning(host, opcode);
+
+	sdhci_end_tuning(host);
+
+	if (current_bus_width == MMC_BUS_WIDTH_8) {
+		mmc->ios.bus_width = MMC_BUS_WIDTH_8;
+		sdhci_set_bus_width(host, current_bus_width);
+	}
+
+	/* Cancel force power mode enter L0 */
+	scratch = sdhci_readw(host, O2_SD_MISC_CTRL);
+	scratch &= ~(O2_SD_PWR_FORCE_L0);
+	sdhci_writew(host, scratch, O2_SD_MISC_CTRL);
+
+	sdhci_reset(host, SDHCI_RESET_CMD);
+	sdhci_reset(host, SDHCI_RESET_DATA);
+
+	host->flags &= ~SDHCI_HS400_TUNING;
+	return 0;
+}
+
+static void o2_pci_led_enable(struct sdhci_pci_chip *chip)
+{
+	int ret;
+	u32 scratch_32;
+
+	/* Set led of SD host function enable */
+	ret = pci_read_config_dword(chip->pdev,
+				    O2_SD_FUNC_REG0, &scratch_32);
+	if (ret)
+		return;
+
+	scratch_32 &= ~O2_SD_FREG0_LEDOFF;
+	pci_write_config_dword(chip->pdev,
+			       O2_SD_FUNC_REG0, scratch_32);
+
+	ret = pci_read_config_dword(chip->pdev,
+				    O2_SD_TEST_REG, &scratch_32);
+	if (ret)
+		return;
+
+	scratch_32 |= O2_SD_LED_ENABLE;
+	pci_write_config_dword(chip->pdev,
+			       O2_SD_TEST_REG, scratch_32);
+}
+
+static void sdhci_pci_o2_fujin2_pci_init(struct sdhci_pci_chip *chip)
+{
+	u32 scratch_32;
+	int ret;
+	/* Improve write performance for SD3.0 */
+	ret = pci_read_config_dword(chip->pdev, O2_SD_DEV_CTRL, &scratch_32);
+	if (ret)
+		return;
+	scratch_32 &= ~((1 << 12) | (1 << 13) | (1 << 14));
+	pci_write_config_dword(chip->pdev, O2_SD_DEV_CTRL, scratch_32);
+
+	/* Enable Link abnormal reset generating Reset */
+	ret = pci_read_config_dword(chip->pdev, O2_SD_MISC_REG5, &scratch_32);
+	if (ret)
+		return;
+	scratch_32 &= ~((1 << 19) | (1 << 11));
+	scratch_32 |= (1 << 10);
+	pci_write_config_dword(chip->pdev, O2_SD_MISC_REG5, scratch_32);
+
+	/* set card power over current protection */
+	ret = pci_read_config_dword(chip->pdev, O2_SD_TEST_REG, &scratch_32);
+	if (ret)
+		return;
+	scratch_32 |= (1 << 4);
+	pci_write_config_dword(chip->pdev, O2_SD_TEST_REG, scratch_32);
+
+	/* adjust the output delay for SD mode */
+	pci_write_config_dword(chip->pdev, O2_SD_DELAY_CTRL, 0x00002492);
+
+	/* Set the output voltage setting of Aux 1.2v LDO */
+	ret = pci_read_config_dword(chip->pdev, O2_SD_LD0_CTRL, &scratch_32);
+	if (ret)
+		return;
+	scratch_32 &= ~(3 << 12);
+	pci_write_config_dword(chip->pdev, O2_SD_LD0_CTRL, scratch_32);
+
+	/* Set Max power supply capability of SD host */
+	ret = pci_read_config_dword(chip->pdev, O2_SD_CAP_REG0, &scratch_32);
+	if (ret)
+		return;
+	scratch_32 &= ~(0x01FE);
+	scratch_32 |= 0x00CC;
+	pci_write_config_dword(chip->pdev, O2_SD_CAP_REG0, scratch_32);
+	/* Set DLL Tuning Window */
+	ret = pci_read_config_dword(chip->pdev,
+				    O2_SD_TUNING_CTRL, &scratch_32);
+	if (ret)
+		return;
+	scratch_32 &= ~(0x000000FF);
+	scratch_32 |= 0x00000066;
+	pci_write_config_dword(chip->pdev, O2_SD_TUNING_CTRL, scratch_32);
+
+	/* Set UHS2 T_EIDLE */
+	ret = pci_read_config_dword(chip->pdev,
+				    O2_SD_UHS2_L1_CTRL, &scratch_32);
+	if (ret)
+		return;
+	scratch_32 &= ~(0x000000FC);
+	scratch_32 |= 0x00000084;
+	pci_write_config_dword(chip->pdev, O2_SD_UHS2_L1_CTRL, scratch_32);
+
+	/* Set UHS2 Termination */
+	ret = pci_read_config_dword(chip->pdev, O2_SD_FUNC_REG3, &scratch_32);
+	if (ret)
+		return;
+	scratch_32 &= ~((1 << 21) | (1 << 30));
+
+	pci_write_config_dword(chip->pdev, O2_SD_FUNC_REG3, scratch_32);
+
+	/* Set L1 Entrance Timer */
+	ret = pci_read_config_dword(chip->pdev, O2_SD_CAPS, &scratch_32);
+	if (ret)
+		return;
+	scratch_32 &= ~(0xf0000000);
+	scratch_32 |= 0x30000000;
+	pci_write_config_dword(chip->pdev, O2_SD_CAPS, scratch_32);
+
+	ret = pci_read_config_dword(chip->pdev,
+				    O2_SD_MISC_CTRL4, &scratch_32);
+	if (ret)
+		return;
+	scratch_32 &= ~(0x000f0000);
+	scratch_32 |= 0x00080000;
+	pci_write_config_dword(chip->pdev, O2_SD_MISC_CTRL4, scratch_32);
+}
+
+static void sdhci_pci_o2_enable_msi(struct sdhci_pci_chip *chip,
+				    struct sdhci_host *host)
+{
+	int ret;
+
+	ret = pci_find_capability(chip->pdev, PCI_CAP_ID_MSI);
+	if (!ret) {
+		pr_info("%s: unsupport msi, use INTx irq\n",
+			mmc_hostname(host->mmc));
+		return;
+	}
+
+	ret = pci_alloc_irq_vectors(chip->pdev, 1, 1,
+				    PCI_IRQ_MSI | PCI_IRQ_MSIX);
+	if (ret < 0) {
+		pr_err("%s: enable PCI MSI failed, err=%d\n",
+		       mmc_hostname(host->mmc), ret);
+		return;
+	}
+
+	host->irq = pci_irq_vector(chip->pdev, 0);
+}
+
+static void sdhci_o2_enable_clk(struct sdhci_host *host, u16 clk)
+{
+	/* Enable internal clock */
+	clk |= SDHCI_CLOCK_INT_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+	sdhci_o2_enable_internal_clock(host);
+	if (sdhci_o2_get_cd(host->mmc)) {
+		clk |= SDHCI_CLOCK_CARD_EN;
+		sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+	}
+}
+
+static void sdhci_pci_o2_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	u16 clk;
+	u8 scratch;
+	u32 scratch_32;
+	struct sdhci_pci_slot *slot = sdhci_priv(host);
+	struct sdhci_pci_chip *chip = slot->chip;
+
+	host->mmc->actual_clock = 0;
+
+	sdhci_writew(host, 0, SDHCI_CLOCK_CONTROL);
+
+	if (clock == 0)
+		return;
+
+	if ((host->timing == MMC_TIMING_UHS_SDR104) && (clock == 200000000)) {
+		pci_read_config_byte(chip->pdev, O2_SD_LOCK_WP, &scratch);
+
+		scratch &= 0x7f;
+		pci_write_config_byte(chip->pdev, O2_SD_LOCK_WP, scratch);
+
+		pci_read_config_dword(chip->pdev, O2_SD_PLL_SETTING, &scratch_32);
+
+		if ((scratch_32 & 0xFFFF0000) != 0x2c280000)
+			o2_pci_set_baseclk(chip, 0x2c280000);
+
+		pci_read_config_byte(chip->pdev, O2_SD_LOCK_WP, &scratch);
+
+		scratch |= 0x80;
+		pci_write_config_byte(chip->pdev, O2_SD_LOCK_WP, scratch);
+	}
+
+	clk = sdhci_calc_clk(host, clock, &host->mmc->actual_clock);
+	sdhci_o2_enable_clk(host, clk);
+}
+
+static int sdhci_pci_o2_probe_slot(struct sdhci_pci_slot *slot)
+{
+	struct sdhci_pci_chip *chip;
+	struct sdhci_host *host;
+	struct o2_host *o2_host = sdhci_pci_priv(slot);
+	u32 reg, caps;
+	int ret;
+
+	chip = slot->chip;
+	host = slot->host;
+
+	o2_host->dll_adjust_count = 0;
+	caps = sdhci_readl(host, SDHCI_CAPABILITIES);
+
+	/*
+	 * mmc_select_bus_width() will test the bus to determine the actual bus
+	 * width.
+	 */
+	if (caps & SDHCI_CAN_DO_8BIT)
+		host->mmc->caps |= MMC_CAP_8_BIT_DATA;
+
+	switch (chip->pdev->device) {
+	case PCI_DEVICE_ID_O2_SDS0:
+	case PCI_DEVICE_ID_O2_SEABIRD0:
+	case PCI_DEVICE_ID_O2_SEABIRD1:
+	case PCI_DEVICE_ID_O2_SDS1:
+	case PCI_DEVICE_ID_O2_FUJIN2:
+		reg = sdhci_readl(host, O2_SD_VENDOR_SETTING);
+		if (reg & 0x1)
+			host->quirks |= SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12;
+
+		sdhci_pci_o2_enable_msi(chip, host);
+
+		if (chip->pdev->device == PCI_DEVICE_ID_O2_SEABIRD0) {
+			ret = pci_read_config_dword(chip->pdev,
+						    O2_SD_MISC_SETTING, &reg);
+			if (ret)
+				return -EIO;
+			if (reg & (1 << 4)) {
+				pr_info("%s: emmc 1.8v flag is set, force 1.8v signaling voltage\n",
+					mmc_hostname(host->mmc));
+				host->flags &= ~SDHCI_SIGNALING_330;
+				host->flags |= SDHCI_SIGNALING_180;
+				host->mmc->caps2 |= MMC_CAP2_NO_SD;
+				host->mmc->caps2 |= MMC_CAP2_NO_SDIO;
+				pci_write_config_dword(chip->pdev,
+						       O2_SD_DETECT_SETTING, 3);
+			}
+
+			slot->host->mmc_host_ops.get_cd = sdhci_o2_get_cd;
+		}
+
+		if (chip->pdev->device == PCI_DEVICE_ID_O2_SEABIRD1) {
+			slot->host->mmc_host_ops.get_cd = sdhci_o2_get_cd;
+			host->mmc->caps2 |= MMC_CAP2_NO_SDIO;
+			host->quirks2 |= SDHCI_QUIRK2_PRESET_VALUE_BROKEN;
+		}
+
+		host->mmc_host_ops.execute_tuning = sdhci_o2_execute_tuning;
+
+		if (chip->pdev->device != PCI_DEVICE_ID_O2_FUJIN2)
+			break;
+		/* set dll watch dog timer */
+		reg = sdhci_readl(host, O2_SD_VENDOR_SETTING2);
+		reg |= (1 << 12);
+		sdhci_writel(host, reg, O2_SD_VENDOR_SETTING2);
+
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int sdhci_pci_o2_probe(struct sdhci_pci_chip *chip)
+{
+	int ret;
+	u8 scratch;
+	u32 scratch_32;
+
+	switch (chip->pdev->device) {
+	case PCI_DEVICE_ID_O2_8220:
+	case PCI_DEVICE_ID_O2_8221:
+	case PCI_DEVICE_ID_O2_8320:
+	case PCI_DEVICE_ID_O2_8321:
+		/* This extra setup is required due to broken ADMA. */
+		ret = pci_read_config_byte(chip->pdev,
+				O2_SD_LOCK_WP, &scratch);
+		if (ret)
+			return ret;
+		scratch &= 0x7f;
+		pci_write_config_byte(chip->pdev, O2_SD_LOCK_WP, scratch);
+
+		/* Set Multi 3 to VCC3V# */
+		pci_write_config_byte(chip->pdev, O2_SD_MULTI_VCC3V, 0x08);
+
+		/* Disable CLK_REQ# support after media DET */
+		ret = pci_read_config_byte(chip->pdev,
+				O2_SD_CLKREQ, &scratch);
+		if (ret)
+			return ret;
+		scratch |= 0x20;
+		pci_write_config_byte(chip->pdev, O2_SD_CLKREQ, scratch);
+
+		/* Choose capabilities, enable SDMA.  We have to write 0x01
+		 * to the capabilities register first to unlock it.
+		 */
+		ret = pci_read_config_byte(chip->pdev, O2_SD_CAPS, &scratch);
+		if (ret)
+			return ret;
+		scratch |= 0x01;
+		pci_write_config_byte(chip->pdev, O2_SD_CAPS, scratch);
+		pci_write_config_byte(chip->pdev, O2_SD_CAPS, 0x73);
+
+		/* Disable ADMA1/2 */
+		pci_write_config_byte(chip->pdev, O2_SD_ADMA1, 0x39);
+		pci_write_config_byte(chip->pdev, O2_SD_ADMA2, 0x08);
+
+		/* Disable the infinite transfer mode */
+		ret = pci_read_config_byte(chip->pdev,
+				O2_SD_INF_MOD, &scratch);
+		if (ret)
+			return ret;
+		scratch |= 0x08;
+		pci_write_config_byte(chip->pdev, O2_SD_INF_MOD, scratch);
+
+		/* Lock WP */
+		ret = pci_read_config_byte(chip->pdev,
+				O2_SD_LOCK_WP, &scratch);
+		if (ret)
+			return ret;
+		scratch |= 0x80;
+		pci_write_config_byte(chip->pdev, O2_SD_LOCK_WP, scratch);
+		break;
+	case PCI_DEVICE_ID_O2_SDS0:
+	case PCI_DEVICE_ID_O2_SDS1:
+	case PCI_DEVICE_ID_O2_FUJIN2:
+		/* UnLock WP */
+		ret = pci_read_config_byte(chip->pdev,
+				O2_SD_LOCK_WP, &scratch);
+		if (ret)
+			return ret;
+
+		scratch &= 0x7f;
+		pci_write_config_byte(chip->pdev, O2_SD_LOCK_WP, scratch);
+
+		/* DevId=8520 subId= 0x11 or 0x12  Type Chip support */
+		if (chip->pdev->device == PCI_DEVICE_ID_O2_FUJIN2) {
+			ret = pci_read_config_dword(chip->pdev,
+						    O2_SD_FUNC_REG0,
+						    &scratch_32);
+			scratch_32 = ((scratch_32 & 0xFF000000) >> 24);
+
+			/* Check Whether subId is 0x11 or 0x12 */
+			if ((scratch_32 == 0x11) || (scratch_32 == 0x12)) {
+				scratch_32 = 0x25100000;
+
+				o2_pci_set_baseclk(chip, scratch_32);
+				ret = pci_read_config_dword(chip->pdev,
+							    O2_SD_FUNC_REG4,
+							    &scratch_32);
+
+				/* Enable Base Clk setting change */
+				scratch_32 |= O2_SD_FREG4_ENABLE_CLK_SET;
+				pci_write_config_dword(chip->pdev,
+						       O2_SD_FUNC_REG4,
+						       scratch_32);
+
+				/* Set Tuning Window to 4 */
+				pci_write_config_byte(chip->pdev,
+						      O2_SD_TUNING_CTRL, 0x44);
+
+				break;
+			}
+		}
+
+		/* Enable 8520 led function */
+		o2_pci_led_enable(chip);
+
+		/* Set timeout CLK */
+		ret = pci_read_config_dword(chip->pdev,
+					    O2_SD_CLK_SETTING, &scratch_32);
+		if (ret)
+			return ret;
+
+		scratch_32 &= ~(0xFF00);
+		scratch_32 |= 0x07E0C800;
+		pci_write_config_dword(chip->pdev,
+				       O2_SD_CLK_SETTING, scratch_32);
+
+		ret = pci_read_config_dword(chip->pdev,
+					    O2_SD_CLKREQ, &scratch_32);
+		if (ret)
+			return ret;
+		scratch_32 |= 0x3;
+		pci_write_config_dword(chip->pdev, O2_SD_CLKREQ, scratch_32);
+
+		ret = pci_read_config_dword(chip->pdev,
+					    O2_SD_PLL_SETTING, &scratch_32);
+		if (ret)
+			return ret;
+
+		scratch_32 &= ~(0x1F3F070E);
+		scratch_32 |= 0x18270106;
+		pci_write_config_dword(chip->pdev,
+				       O2_SD_PLL_SETTING, scratch_32);
+
+		/* Disable UHS1 funciton */
+		ret = pci_read_config_dword(chip->pdev,
+					    O2_SD_CAP_REG2, &scratch_32);
+		if (ret)
+			return ret;
+		scratch_32 &= ~(0xE0);
+		pci_write_config_dword(chip->pdev,
+				       O2_SD_CAP_REG2, scratch_32);
+
+		if (chip->pdev->device == PCI_DEVICE_ID_O2_FUJIN2)
+			sdhci_pci_o2_fujin2_pci_init(chip);
+
+		/* Lock WP */
+		ret = pci_read_config_byte(chip->pdev,
+					   O2_SD_LOCK_WP, &scratch);
+		if (ret)
+			return ret;
+		scratch |= 0x80;
+		pci_write_config_byte(chip->pdev, O2_SD_LOCK_WP, scratch);
+		break;
+	case PCI_DEVICE_ID_O2_SEABIRD0:
+	case PCI_DEVICE_ID_O2_SEABIRD1:
+		/* UnLock WP */
+		ret = pci_read_config_byte(chip->pdev,
+				O2_SD_LOCK_WP, &scratch);
+		if (ret)
+			return ret;
+
+		scratch &= 0x7f;
+		pci_write_config_byte(chip->pdev, O2_SD_LOCK_WP, scratch);
+
+		ret = pci_read_config_dword(chip->pdev,
+					    O2_SD_PLL_SETTING, &scratch_32);
+
+		if ((scratch_32 & 0xff000000) == 0x01000000) {
+			scratch_32 &= 0x0000FFFF;
+			scratch_32 |= 0x1F340000;
+
+			pci_write_config_dword(chip->pdev,
+					       O2_SD_PLL_SETTING, scratch_32);
+		} else {
+			scratch_32 &= 0x0000FFFF;
+			scratch_32 |= 0x25100000;
+
+			pci_write_config_dword(chip->pdev,
+					       O2_SD_PLL_SETTING, scratch_32);
+
+			ret = pci_read_config_dword(chip->pdev,
+						    O2_SD_FUNC_REG4,
+						    &scratch_32);
+			scratch_32 |= (1 << 22);
+			pci_write_config_dword(chip->pdev,
+					       O2_SD_FUNC_REG4, scratch_32);
+		}
+
+		/* Set Tuning Windows to 5 */
+		pci_write_config_byte(chip->pdev,
+				O2_SD_TUNING_CTRL, 0x55);
+		//Adjust 1st and 2nd CD debounce time
+		pci_read_config_dword(chip->pdev, O2_SD_MISC_CTRL2, &scratch_32);
+		scratch_32 &= 0xFFE7FFFF;
+		scratch_32 |= 0x00180000;
+		pci_write_config_dword(chip->pdev, O2_SD_MISC_CTRL2, scratch_32);
+		pci_write_config_dword(chip->pdev, O2_SD_DETECT_SETTING, 1);
+		/* Lock WP */
+		ret = pci_read_config_byte(chip->pdev,
+					   O2_SD_LOCK_WP, &scratch);
+		if (ret)
+			return ret;
+		scratch |= 0x80;
+		pci_write_config_byte(chip->pdev, O2_SD_LOCK_WP, scratch);
+		break;
+	}
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int sdhci_pci_o2_resume(struct sdhci_pci_chip *chip)
+{
+	sdhci_pci_o2_probe(chip);
+	return sdhci_pci_resume_host(chip);
+}
+#endif
+
+static const struct sdhci_ops sdhci_pci_o2_ops = {
+	.set_clock = sdhci_pci_o2_set_clock,
+	.enable_dma = sdhci_pci_enable_dma,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+const struct sdhci_pci_fixes sdhci_o2 = {
+	.probe = sdhci_pci_o2_probe,
+	.quirks = SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.quirks2 = SDHCI_QUIRK2_CLEAR_TRANSFERMODE_REG_BEFORE_CMD,
+	.probe_slot = sdhci_pci_o2_probe_slot,
+#ifdef CONFIG_PM_SLEEP
+	.resume = sdhci_pci_o2_resume,
+#endif
+	.ops = &sdhci_pci_o2_ops,
+	.priv_size = sizeof(struct o2_host),
+};
diff --git a/drivers/mmc/host/sdhci-pci.h b/drivers/mmc/host/sdhci-pci.h
new file mode 100644
index 000000000..dcd99d505
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pci.h
@@ -0,0 +1,204 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __SDHCI_PCI_H
+#define __SDHCI_PCI_H
+
+/*
+ * PCI device IDs, sub IDs
+ */
+
+#define PCI_DEVICE_ID_O2_SDS0		0x8420
+#define PCI_DEVICE_ID_O2_SDS1		0x8421
+#define PCI_DEVICE_ID_O2_FUJIN2		0x8520
+#define PCI_DEVICE_ID_O2_SEABIRD0	0x8620
+#define PCI_DEVICE_ID_O2_SEABIRD1	0x8621
+
+#define PCI_DEVICE_ID_INTEL_PCH_SDIO0	0x8809
+#define PCI_DEVICE_ID_INTEL_PCH_SDIO1	0x880a
+#define PCI_DEVICE_ID_INTEL_BYT_EMMC	0x0f14
+#define PCI_DEVICE_ID_INTEL_BYT_SDIO	0x0f15
+#define PCI_DEVICE_ID_INTEL_BYT_SD	0x0f16
+#define PCI_DEVICE_ID_INTEL_BYT_EMMC2	0x0f50
+#define PCI_DEVICE_ID_INTEL_BSW_EMMC	0x2294
+#define PCI_DEVICE_ID_INTEL_BSW_SDIO	0x2295
+#define PCI_DEVICE_ID_INTEL_BSW_SD	0x2296
+#define PCI_DEVICE_ID_INTEL_MRFLD_MMC	0x1190
+#define PCI_DEVICE_ID_INTEL_CLV_SDIO0	0x08f9
+#define PCI_DEVICE_ID_INTEL_CLV_SDIO1	0x08fa
+#define PCI_DEVICE_ID_INTEL_CLV_SDIO2	0x08fb
+#define PCI_DEVICE_ID_INTEL_CLV_EMMC0	0x08e5
+#define PCI_DEVICE_ID_INTEL_CLV_EMMC1	0x08e6
+#define PCI_DEVICE_ID_INTEL_QRK_SD	0x08A7
+#define PCI_DEVICE_ID_INTEL_SPT_EMMC	0x9d2b
+#define PCI_DEVICE_ID_INTEL_SPT_SDIO	0x9d2c
+#define PCI_DEVICE_ID_INTEL_SPT_SD	0x9d2d
+#define PCI_DEVICE_ID_INTEL_DNV_EMMC	0x19db
+#define PCI_DEVICE_ID_INTEL_CDF_EMMC	0x18db
+#define PCI_DEVICE_ID_INTEL_BXT_SD	0x0aca
+#define PCI_DEVICE_ID_INTEL_BXT_EMMC	0x0acc
+#define PCI_DEVICE_ID_INTEL_BXT_SDIO	0x0ad0
+#define PCI_DEVICE_ID_INTEL_BXTM_SD	0x1aca
+#define PCI_DEVICE_ID_INTEL_BXTM_EMMC	0x1acc
+#define PCI_DEVICE_ID_INTEL_BXTM_SDIO	0x1ad0
+#define PCI_DEVICE_ID_INTEL_APL_SD	0x5aca
+#define PCI_DEVICE_ID_INTEL_APL_EMMC	0x5acc
+#define PCI_DEVICE_ID_INTEL_APL_SDIO	0x5ad0
+#define PCI_DEVICE_ID_INTEL_GLK_SD	0x31ca
+#define PCI_DEVICE_ID_INTEL_GLK_EMMC	0x31cc
+#define PCI_DEVICE_ID_INTEL_GLK_SDIO	0x31d0
+#define PCI_DEVICE_ID_INTEL_CNP_EMMC	0x9dc4
+#define PCI_DEVICE_ID_INTEL_CNP_SD	0x9df5
+#define PCI_DEVICE_ID_INTEL_CNPH_SD	0xa375
+#define PCI_DEVICE_ID_INTEL_ICP_EMMC	0x34c4
+#define PCI_DEVICE_ID_INTEL_ICP_SD	0x34f8
+#define PCI_DEVICE_ID_INTEL_EHL_EMMC	0x4b47
+#define PCI_DEVICE_ID_INTEL_EHL_SD	0x4b48
+#define PCI_DEVICE_ID_INTEL_CML_EMMC	0x02c4
+#define PCI_DEVICE_ID_INTEL_CML_SD	0x02f5
+#define PCI_DEVICE_ID_INTEL_CMLH_SD	0x06f5
+#define PCI_DEVICE_ID_INTEL_JSL_EMMC	0x4dc4
+#define PCI_DEVICE_ID_INTEL_JSL_SD	0x4df8
+#define PCI_DEVICE_ID_INTEL_LKF_EMMC	0x98c4
+#define PCI_DEVICE_ID_INTEL_LKF_SD	0x98f8
+#define PCI_DEVICE_ID_INTEL_ADL_EMMC	0x54c4
+
+#define PCI_DEVICE_ID_SYSKONNECT_8000	0x8000
+#define PCI_DEVICE_ID_VIA_95D0		0x95d0
+#define PCI_DEVICE_ID_REALTEK_5250	0x5250
+
+#define PCI_SUBDEVICE_ID_NI_7884	0x7884
+#define PCI_SUBDEVICE_ID_NI_78E3	0x78e3
+
+#define PCI_VENDOR_ID_ARASAN		0x16e6
+#define PCI_DEVICE_ID_ARASAN_PHY_EMMC	0x0670
+
+#define PCI_DEVICE_ID_SYNOPSYS_DWC_MSHC 0xc202
+
+#define PCI_DEVICE_ID_GLI_9755		0x9755
+#define PCI_DEVICE_ID_GLI_9750		0x9750
+#define PCI_DEVICE_ID_GLI_9763E		0xe763
+
+/*
+ * PCI device class and mask
+ */
+
+#define SYSTEM_SDHCI			(PCI_CLASS_SYSTEM_SDHCI << 8)
+#define PCI_CLASS_MASK			0xFFFF00
+
+/*
+ * Macros for PCI device-description
+ */
+
+#define _PCI_VEND(vend) PCI_VENDOR_ID_##vend
+#define _PCI_DEV(vend, dev) PCI_DEVICE_ID_##vend##_##dev
+#define _PCI_SUBDEV(subvend, subdev) PCI_SUBDEVICE_ID_##subvend##_##subdev
+
+#define SDHCI_PCI_DEVICE(vend, dev, cfg) { \
+	.vendor = _PCI_VEND(vend), .device = _PCI_DEV(vend, dev), \
+	.subvendor = PCI_ANY_ID, .subdevice = PCI_ANY_ID, \
+	.driver_data = (kernel_ulong_t)&(sdhci_##cfg) \
+}
+
+#define SDHCI_PCI_SUBDEVICE(vend, dev, subvend, subdev, cfg) { \
+	.vendor = _PCI_VEND(vend), .device = _PCI_DEV(vend, dev), \
+	.subvendor = _PCI_VEND(subvend), \
+	.subdevice = _PCI_SUBDEV(subvend, subdev), \
+	.driver_data = (kernel_ulong_t)&(sdhci_##cfg) \
+}
+
+#define SDHCI_PCI_DEVICE_CLASS(vend, cl, cl_msk, cfg) { \
+	.vendor = _PCI_VEND(vend), .device = PCI_ANY_ID, \
+	.subvendor = PCI_ANY_ID, .subdevice = PCI_ANY_ID, \
+	.class = (cl), .class_mask = (cl_msk), \
+	.driver_data = (kernel_ulong_t)&(sdhci_##cfg) \
+}
+
+/*
+ * PCI registers
+ */
+
+#define PCI_SDHCI_IFPIO			0x00
+#define PCI_SDHCI_IFDMA			0x01
+#define PCI_SDHCI_IFVENDOR		0x02
+
+#define PCI_SLOT_INFO			0x40	/* 8 bits */
+#define  PCI_SLOT_INFO_SLOTS(x)		((x >> 4) & 7)
+#define  PCI_SLOT_INFO_FIRST_BAR_MASK	0x07
+
+#define MAX_SLOTS			8
+
+struct sdhci_pci_chip;
+struct sdhci_pci_slot;
+
+struct sdhci_pci_fixes {
+	unsigned int		quirks;
+	unsigned int		quirks2;
+	bool			allow_runtime_pm;
+	bool			own_cd_for_runtime_pm;
+
+	int			(*probe) (struct sdhci_pci_chip *);
+
+	int			(*probe_slot) (struct sdhci_pci_slot *);
+	int			(*add_host) (struct sdhci_pci_slot *);
+	void			(*remove_slot) (struct sdhci_pci_slot *, int);
+
+#ifdef CONFIG_PM_SLEEP
+	int			(*suspend) (struct sdhci_pci_chip *);
+	int			(*resume) (struct sdhci_pci_chip *);
+#endif
+#ifdef CONFIG_PM
+	int			(*runtime_suspend) (struct sdhci_pci_chip *);
+	int			(*runtime_resume) (struct sdhci_pci_chip *);
+#endif
+
+	const struct sdhci_ops	*ops;
+	size_t			priv_size;
+};
+
+struct sdhci_pci_slot {
+	struct sdhci_pci_chip	*chip;
+	struct sdhci_host	*host;
+	struct sdhci_pci_data	*data;
+
+	int			rst_n_gpio;
+	int			cd_gpio;
+	int			cd_irq;
+
+	int			cd_idx;
+	bool			cd_override_level;
+
+	void (*hw_reset)(struct sdhci_host *host);
+	unsigned long		private[] ____cacheline_aligned;
+};
+
+struct sdhci_pci_chip {
+	struct pci_dev		*pdev;
+
+	unsigned int		quirks;
+	unsigned int		quirks2;
+	bool			allow_runtime_pm;
+	bool			pm_retune;
+	bool			rpm_retune;
+	const struct sdhci_pci_fixes *fixes;
+
+	int			num_slots;	/* Slots on controller */
+	struct sdhci_pci_slot	*slots[MAX_SLOTS]; /* Pointers to host slots */
+};
+
+static inline void *sdhci_pci_priv(struct sdhci_pci_slot *slot)
+{
+	return (void *)slot->private;
+}
+
+#ifdef CONFIG_PM_SLEEP
+int sdhci_pci_resume_host(struct sdhci_pci_chip *chip);
+#endif
+int sdhci_pci_enable_dma(struct sdhci_host *host);
+
+extern const struct sdhci_pci_fixes sdhci_arasan;
+extern const struct sdhci_pci_fixes sdhci_snps;
+extern const struct sdhci_pci_fixes sdhci_o2;
+extern const struct sdhci_pci_fixes sdhci_gl9750;
+extern const struct sdhci_pci_fixes sdhci_gl9755;
+extern const struct sdhci_pci_fixes sdhci_gl9763e;
+
+#endif /* __SDHCI_PCI_H */
diff --git a/drivers/mmc/host/sdhci-pic32.c b/drivers/mmc/host/sdhci-pic32.c
new file mode 100644
index 000000000..6ce1519ae
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pic32.c
@@ -0,0 +1,255 @@
+/*
+ * Support of SDHCI platform devices for Microchip PIC32.
+ *
+ * Copyright (C) 2015 Microchip
+ * Andrei Pistirica, Paul Thacker
+ *
+ * Inspired by sdhci-pltfm.c
+ *
+ * This file is licensed under the terms of the GNU General Public
+ * License version 2. This program is licensed "as is" without any
+ * warranty of any kind, whether express or implied.
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/highmem.h>
+#include <linux/module.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm.h>
+#include <linux/slab.h>
+#include <linux/mmc/host.h>
+#include <linux/io.h>
+#include "sdhci.h"
+#include "sdhci-pltfm.h"
+#include <linux/platform_data/sdhci-pic32.h>
+
+#define SDH_SHARED_BUS_CTRL		0x000000E0
+#define SDH_SHARED_BUS_NR_CLK_PINS_MASK	0x7
+#define SDH_SHARED_BUS_NR_IRQ_PINS_MASK	0x30
+#define SDH_SHARED_BUS_CLK_PINS		0x10
+#define SDH_SHARED_BUS_IRQ_PINS		0x14
+#define SDH_CAPS_SDH_SLOT_TYPE_MASK	0xC0000000
+#define SDH_SLOT_TYPE_REMOVABLE		0x0
+#define SDH_SLOT_TYPE_EMBEDDED		0x1
+#define SDH_SLOT_TYPE_SHARED_BUS	0x2
+#define SDHCI_CTRL_CDSSEL		0x80
+#define SDHCI_CTRL_CDTLVL		0x40
+
+#define ADMA_FIFO_RD_THSHLD	512
+#define ADMA_FIFO_WR_THSHLD	512
+
+struct pic32_sdhci_priv {
+	struct platform_device	*pdev;
+	struct clk *sys_clk;
+	struct clk *base_clk;
+};
+
+static unsigned int pic32_sdhci_get_max_clock(struct sdhci_host *host)
+{
+	struct pic32_sdhci_priv *sdhci_pdata = sdhci_priv(host);
+
+	return clk_get_rate(sdhci_pdata->base_clk);
+}
+
+static void pic32_sdhci_set_bus_width(struct sdhci_host *host, int width)
+{
+	u8 ctrl;
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+	if (width == MMC_BUS_WIDTH_8) {
+		ctrl &= ~SDHCI_CTRL_4BITBUS;
+		if (host->version >= SDHCI_SPEC_300)
+			ctrl |= SDHCI_CTRL_8BITBUS;
+	} else {
+		if (host->version >= SDHCI_SPEC_300)
+			ctrl &= ~SDHCI_CTRL_8BITBUS;
+		if (width == MMC_BUS_WIDTH_4)
+			ctrl |= SDHCI_CTRL_4BITBUS;
+		else
+			ctrl &= ~SDHCI_CTRL_4BITBUS;
+	}
+
+	/* CD select and test bits must be set for errata workaround. */
+	ctrl &= ~SDHCI_CTRL_CDTLVL;
+	ctrl |= SDHCI_CTRL_CDSSEL;
+	sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+}
+
+static unsigned int pic32_sdhci_get_ro(struct sdhci_host *host)
+{
+	/*
+	 * The SDHCI_WRITE_PROTECT bit is unstable on current hardware so we
+	 * can't depend on its value in any way.
+	 */
+	return 0;
+}
+
+static const struct sdhci_ops pic32_sdhci_ops = {
+	.get_max_clock = pic32_sdhci_get_max_clock,
+	.set_clock = sdhci_set_clock,
+	.set_bus_width = pic32_sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+	.get_ro = pic32_sdhci_get_ro,
+};
+
+static const struct sdhci_pltfm_data sdhci_pic32_pdata = {
+	.ops = &pic32_sdhci_ops,
+	.quirks = SDHCI_QUIRK_NO_HISPD_BIT,
+	.quirks2 = SDHCI_QUIRK2_NO_1_8_V,
+};
+
+static void pic32_sdhci_shared_bus(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	u32 bus = readl(host->ioaddr + SDH_SHARED_BUS_CTRL);
+	u32 clk_pins = (bus & SDH_SHARED_BUS_NR_CLK_PINS_MASK) >> 0;
+	u32 irq_pins = (bus & SDH_SHARED_BUS_NR_IRQ_PINS_MASK) >> 4;
+
+	/* select first clock */
+	if (clk_pins & 1)
+		bus |= (1 << SDH_SHARED_BUS_CLK_PINS);
+
+	/* select first interrupt */
+	if (irq_pins & 1)
+		bus |= (1 << SDH_SHARED_BUS_IRQ_PINS);
+
+	writel(bus, host->ioaddr + SDH_SHARED_BUS_CTRL);
+}
+
+static int pic32_sdhci_probe_platform(struct platform_device *pdev,
+				      struct pic32_sdhci_priv *pdata)
+{
+	int ret = 0;
+	u32 caps_slot_type;
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+
+	/* Check card slot connected on shared bus. */
+	host->caps = readl(host->ioaddr + SDHCI_CAPABILITIES);
+	caps_slot_type = (host->caps & SDH_CAPS_SDH_SLOT_TYPE_MASK) >> 30;
+	if (caps_slot_type == SDH_SLOT_TYPE_SHARED_BUS)
+		pic32_sdhci_shared_bus(pdev);
+
+	return ret;
+}
+
+static int pic32_sdhci_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct pic32_sdhci_priv *sdhci_pdata;
+	struct pic32_sdhci_platform_data *plat_data;
+	int ret;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_pic32_pdata,
+				sizeof(struct pic32_sdhci_priv));
+	if (IS_ERR(host)) {
+		ret = PTR_ERR(host);
+		goto err;
+	}
+
+	pltfm_host = sdhci_priv(host);
+	sdhci_pdata = sdhci_pltfm_priv(pltfm_host);
+
+	plat_data = pdev->dev.platform_data;
+	if (plat_data && plat_data->setup_dma) {
+		ret = plat_data->setup_dma(ADMA_FIFO_RD_THSHLD,
+					   ADMA_FIFO_WR_THSHLD);
+		if (ret)
+			goto err_host;
+	}
+
+	sdhci_pdata->sys_clk = devm_clk_get(&pdev->dev, "sys_clk");
+	if (IS_ERR(sdhci_pdata->sys_clk)) {
+		ret = PTR_ERR(sdhci_pdata->sys_clk);
+		dev_err(&pdev->dev, "Error getting clock\n");
+		goto err_host;
+	}
+
+	ret = clk_prepare_enable(sdhci_pdata->sys_clk);
+	if (ret) {
+		dev_err(&pdev->dev, "Error enabling clock\n");
+		goto err_host;
+	}
+
+	sdhci_pdata->base_clk = devm_clk_get(&pdev->dev, "base_clk");
+	if (IS_ERR(sdhci_pdata->base_clk)) {
+		ret = PTR_ERR(sdhci_pdata->base_clk);
+		dev_err(&pdev->dev, "Error getting clock\n");
+		goto err_sys_clk;
+	}
+
+	ret = clk_prepare_enable(sdhci_pdata->base_clk);
+	if (ret) {
+		dev_err(&pdev->dev, "Error enabling clock\n");
+		goto err_base_clk;
+	}
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto err_base_clk;
+
+	ret = pic32_sdhci_probe_platform(pdev, sdhci_pdata);
+	if (ret) {
+		dev_err(&pdev->dev, "failed to probe platform!\n");
+		goto err_base_clk;
+	}
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err_base_clk;
+
+	dev_info(&pdev->dev, "Successfully added sdhci host\n");
+	return 0;
+
+err_base_clk:
+	clk_disable_unprepare(sdhci_pdata->base_clk);
+err_sys_clk:
+	clk_disable_unprepare(sdhci_pdata->sys_clk);
+err_host:
+	sdhci_pltfm_free(pdev);
+err:
+	dev_err(&pdev->dev, "pic32-sdhci probe failed: %d\n", ret);
+	return ret;
+}
+
+static int pic32_sdhci_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct pic32_sdhci_priv *sdhci_pdata = sdhci_priv(host);
+	u32 scratch;
+
+	scratch = readl(host->ioaddr + SDHCI_INT_STATUS);
+	sdhci_remove_host(host, scratch == (u32)~0);
+	clk_disable_unprepare(sdhci_pdata->base_clk);
+	clk_disable_unprepare(sdhci_pdata->sys_clk);
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+
+static const struct of_device_id pic32_sdhci_id_table[] = {
+	{ .compatible = "microchip,pic32mzda-sdhci" },
+	{}
+};
+MODULE_DEVICE_TABLE(of, pic32_sdhci_id_table);
+
+static struct platform_driver pic32_sdhci_driver = {
+	.driver = {
+		.name	= "pic32-sdhci",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(pic32_sdhci_id_table),
+	},
+	.probe		= pic32_sdhci_probe,
+	.remove		= pic32_sdhci_remove,
+};
+
+module_platform_driver(pic32_sdhci_driver);
+
+MODULE_DESCRIPTION("Microchip PIC32 SDHCI driver");
+MODULE_AUTHOR("Pistirica Sorin Andrei & Sandeep Sheriker");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-pltfm.c b/drivers/mmc/host/sdhci-pltfm.c
new file mode 100644
index 000000000..328b132bb
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pltfm.c
@@ -0,0 +1,267 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * sdhci-pltfm.c Support for SDHCI platform devices
+ * Copyright (c) 2009 Intel Corporation
+ *
+ * Copyright (c) 2007, 2011 Freescale Semiconductor, Inc.
+ * Copyright (c) 2009 MontaVista Software, Inc.
+ *
+ * Authors: Xiaobo Xie <X.Xie@freescale.com>
+ *	    Anton Vorontsov <avorontsov@ru.mvista.com>
+ */
+
+/* Supports:
+ * SDHCI platform devices
+ *
+ * Inspired by sdhci-pci.c, by Pierre Ossman
+ */
+
+#include <linux/err.h>
+#include <linux/module.h>
+#include <linux/property.h>
+#include <linux/of.h>
+#ifdef CONFIG_PPC
+#include <asm/machdep.h>
+#endif
+#include "sdhci-pltfm.h"
+
+unsigned int sdhci_pltfm_clk_get_max_clock(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	return clk_get_rate(pltfm_host->clk);
+}
+EXPORT_SYMBOL_GPL(sdhci_pltfm_clk_get_max_clock);
+
+static const struct sdhci_ops sdhci_pltfm_ops = {
+	.set_clock = sdhci_set_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static bool sdhci_wp_inverted(struct device *dev)
+{
+	if (device_property_present(dev, "sdhci,wp-inverted") ||
+	    device_property_present(dev, "wp-inverted"))
+		return true;
+
+	/* Old device trees don't have the wp-inverted property. */
+#ifdef CONFIG_PPC
+	return machine_is(mpc837x_rdb) || machine_is(mpc837x_mds);
+#else
+	return false;
+#endif /* CONFIG_PPC */
+}
+
+#ifdef CONFIG_OF
+static void sdhci_get_compatibility(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct device_node *np = pdev->dev.of_node;
+
+	if (!np)
+		return;
+
+	if (of_device_is_compatible(np, "fsl,p2020-rev1-esdhc"))
+		host->quirks |= SDHCI_QUIRK_BROKEN_DMA;
+
+	if (of_device_is_compatible(np, "fsl,p2020-esdhc") ||
+	    of_device_is_compatible(np, "fsl,p1010-esdhc") ||
+	    of_device_is_compatible(np, "fsl,t4240-esdhc") ||
+	    of_device_is_compatible(np, "fsl,mpc8536-esdhc"))
+		host->quirks |= SDHCI_QUIRK_BROKEN_TIMEOUT_VAL;
+}
+#else
+void sdhci_get_compatibility(struct platform_device *pdev) {}
+#endif /* CONFIG_OF */
+
+void sdhci_get_property(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	u32 bus_width;
+
+	if (device_property_present(dev, "sdhci,auto-cmd12"))
+		host->quirks |= SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12;
+
+	if (device_property_present(dev, "sdhci,1-bit-only") ||
+	    (device_property_read_u32(dev, "bus-width", &bus_width) == 0 &&
+	    bus_width == 1))
+		host->quirks |= SDHCI_QUIRK_FORCE_1_BIT_DATA;
+
+	if (sdhci_wp_inverted(dev))
+		host->quirks |= SDHCI_QUIRK_INVERTED_WRITE_PROTECT;
+
+	if (device_property_present(dev, "broken-cd"))
+		host->quirks |= SDHCI_QUIRK_BROKEN_CARD_DETECTION;
+
+	if (device_property_present(dev, "no-1-8-v"))
+		host->quirks2 |= SDHCI_QUIRK2_NO_1_8_V;
+
+	sdhci_get_compatibility(pdev);
+
+	device_property_read_u32(dev, "clock-frequency", &pltfm_host->clock);
+
+	if (device_property_present(dev, "keep-power-in-suspend"))
+		host->mmc->pm_caps |= MMC_PM_KEEP_POWER;
+
+	if (device_property_read_bool(dev, "wakeup-source") ||
+	    device_property_read_bool(dev, "enable-sdio-wakeup")) /* legacy */
+		host->mmc->pm_caps |= MMC_PM_WAKE_SDIO_IRQ;
+}
+EXPORT_SYMBOL_GPL(sdhci_get_property);
+
+struct sdhci_host *sdhci_pltfm_init(struct platform_device *pdev,
+				    const struct sdhci_pltfm_data *pdata,
+				    size_t priv_size)
+{
+	struct sdhci_host *host;
+	void __iomem *ioaddr;
+	int irq, ret;
+
+	ioaddr = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(ioaddr)) {
+		ret = PTR_ERR(ioaddr);
+		goto err;
+	}
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0) {
+		ret = irq;
+		goto err;
+	}
+
+	host = sdhci_alloc_host(&pdev->dev,
+		sizeof(struct sdhci_pltfm_host) + priv_size);
+
+	if (IS_ERR(host)) {
+		ret = PTR_ERR(host);
+		goto err;
+	}
+
+	host->ioaddr = ioaddr;
+	host->irq = irq;
+	host->hw_name = dev_name(&pdev->dev);
+	if (pdata && pdata->ops)
+		host->ops = pdata->ops;
+	else
+		host->ops = &sdhci_pltfm_ops;
+	if (pdata) {
+		host->quirks = pdata->quirks;
+		host->quirks2 = pdata->quirks2;
+	}
+
+	platform_set_drvdata(pdev, host);
+
+	return host;
+err:
+	dev_err(&pdev->dev, "%s failed %d\n", __func__, ret);
+	return ERR_PTR(ret);
+}
+EXPORT_SYMBOL_GPL(sdhci_pltfm_init);
+
+void sdhci_pltfm_free(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+
+	sdhci_free_host(host);
+}
+EXPORT_SYMBOL_GPL(sdhci_pltfm_free);
+
+int sdhci_pltfm_register(struct platform_device *pdev,
+			const struct sdhci_pltfm_data *pdata,
+			size_t priv_size)
+{
+	struct sdhci_host *host;
+	int ret = 0;
+
+	host = sdhci_pltfm_init(pdev, pdata, priv_size);
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	sdhci_get_property(pdev);
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		sdhci_pltfm_free(pdev);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(sdhci_pltfm_register);
+
+int sdhci_pltfm_unregister(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	int dead = (readl(host->ioaddr + SDHCI_INT_STATUS) == 0xffffffff);
+
+	sdhci_remove_host(host, dead);
+	clk_disable_unprepare(pltfm_host->clk);
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(sdhci_pltfm_unregister);
+
+#ifdef CONFIG_PM_SLEEP
+int sdhci_pltfm_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	int ret;
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	ret = sdhci_suspend_host(host);
+	if (ret)
+		return ret;
+
+	clk_disable_unprepare(pltfm_host->clk);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(sdhci_pltfm_suspend);
+
+int sdhci_pltfm_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	int ret;
+
+	ret = clk_prepare_enable(pltfm_host->clk);
+	if (ret)
+		return ret;
+
+	ret = sdhci_resume_host(host);
+	if (ret)
+		clk_disable_unprepare(pltfm_host->clk);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(sdhci_pltfm_resume);
+#endif
+
+const struct dev_pm_ops sdhci_pltfm_pmops = {
+	SET_SYSTEM_SLEEP_PM_OPS(sdhci_pltfm_suspend, sdhci_pltfm_resume)
+};
+EXPORT_SYMBOL_GPL(sdhci_pltfm_pmops);
+
+static int __init sdhci_pltfm_drv_init(void)
+{
+	pr_info("sdhci-pltfm: SDHCI platform and OF driver helper\n");
+
+	return 0;
+}
+module_init(sdhci_pltfm_drv_init);
+
+static void __exit sdhci_pltfm_drv_exit(void)
+{
+}
+module_exit(sdhci_pltfm_drv_exit);
+
+MODULE_DESCRIPTION("SDHCI platform and OF driver helper");
+MODULE_AUTHOR("Intel Corporation");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-pltfm.h b/drivers/mmc/host/sdhci-pltfm.h
new file mode 100644
index 000000000..9bd717ff7
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pltfm.h
@@ -0,0 +1,123 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright 2010 MontaVista Software, LLC.
+ *
+ * Author: Anton Vorontsov <avorontsov@ru.mvista.com>
+ */
+
+#ifndef _DRIVERS_MMC_SDHCI_PLTFM_H
+#define _DRIVERS_MMC_SDHCI_PLTFM_H
+
+#include <linux/clk.h>
+#include <linux/platform_device.h>
+#include "sdhci.h"
+
+struct sdhci_pltfm_data {
+	const struct sdhci_ops *ops;
+	unsigned int quirks;
+	unsigned int quirks2;
+};
+
+struct sdhci_pltfm_host {
+	struct clk *clk;
+
+	/* migrate from sdhci_of_host */
+	unsigned int clock;
+	u16 xfer_mode_shadow;
+
+	unsigned long private[] ____cacheline_aligned;
+};
+
+#ifdef CONFIG_MMC_SDHCI_BIG_ENDIAN_32BIT_BYTE_SWAPPER
+/*
+ * These accessors are designed for big endian hosts doing I/O to
+ * little endian controllers incorporating a 32-bit hardware byte swapper.
+ */
+static inline u32 sdhci_be32bs_readl(struct sdhci_host *host, int reg)
+{
+	return in_be32(host->ioaddr + reg);
+}
+
+static inline u16 sdhci_be32bs_readw(struct sdhci_host *host, int reg)
+{
+	return in_be16(host->ioaddr + (reg ^ 0x2));
+}
+
+static inline u8 sdhci_be32bs_readb(struct sdhci_host *host, int reg)
+{
+	return in_8(host->ioaddr + (reg ^ 0x3));
+}
+
+static inline void sdhci_be32bs_writel(struct sdhci_host *host,
+				       u32 val, int reg)
+{
+	out_be32(host->ioaddr + reg, val);
+}
+
+static inline void sdhci_be32bs_writew(struct sdhci_host *host,
+				       u16 val, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	int base = reg & ~0x3;
+	int shift = (reg & 0x2) * 8;
+
+	switch (reg) {
+	case SDHCI_TRANSFER_MODE:
+		/*
+		 * Postpone this write, we must do it together with a
+		 * command write that is down below.
+		 */
+		pltfm_host->xfer_mode_shadow = val;
+		return;
+	case SDHCI_COMMAND:
+		sdhci_be32bs_writel(host,
+				    val << 16 | pltfm_host->xfer_mode_shadow,
+				    SDHCI_TRANSFER_MODE);
+		return;
+	}
+	clrsetbits_be32(host->ioaddr + base, 0xffff << shift, val << shift);
+}
+
+static inline void sdhci_be32bs_writeb(struct sdhci_host *host, u8 val, int reg)
+{
+	int base = reg & ~0x3;
+	int shift = (reg & 0x3) * 8;
+
+	clrsetbits_be32(host->ioaddr + base , 0xff << shift, val << shift);
+}
+#endif /* CONFIG_MMC_SDHCI_BIG_ENDIAN_32BIT_BYTE_SWAPPER */
+
+void sdhci_get_property(struct platform_device *pdev);
+
+static inline void sdhci_get_of_property(struct platform_device *pdev)
+{
+	return sdhci_get_property(pdev);
+}
+
+extern struct sdhci_host *sdhci_pltfm_init(struct platform_device *pdev,
+					  const struct sdhci_pltfm_data *pdata,
+					  size_t priv_size);
+extern void sdhci_pltfm_free(struct platform_device *pdev);
+
+extern int sdhci_pltfm_register(struct platform_device *pdev,
+				const struct sdhci_pltfm_data *pdata,
+				size_t priv_size);
+extern int sdhci_pltfm_unregister(struct platform_device *pdev);
+
+extern unsigned int sdhci_pltfm_clk_get_max_clock(struct sdhci_host *host);
+
+static inline void *sdhci_pltfm_priv(struct sdhci_pltfm_host *host)
+{
+	return host->private;
+}
+
+extern const struct dev_pm_ops sdhci_pltfm_pmops;
+#ifdef CONFIG_PM_SLEEP
+int sdhci_pltfm_suspend(struct device *dev);
+int sdhci_pltfm_resume(struct device *dev);
+#else
+static inline int sdhci_pltfm_suspend(struct device *dev) { return 0; }
+static inline int sdhci_pltfm_resume(struct device *dev) { return 0; }
+#endif
+
+#endif /* _DRIVERS_MMC_SDHCI_PLTFM_H */
diff --git a/drivers/mmc/host/sdhci-pxav2.c b/drivers/mmc/host/sdhci-pxav2.c
new file mode 100644
index 000000000..f18906b55
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pxav2.c
@@ -0,0 +1,242 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2010 Marvell International Ltd.
+ *		Zhangfei Gao <zhangfei.gao@marvell.com>
+ *		Kevin Wang <dwang4@marvell.com>
+ *		Jun Nie <njun@marvell.com>
+ *		Qiming Wu <wuqm@marvell.com>
+ *		Philip Rakity <prakity@marvell.com>
+ */
+
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/platform_device.h>
+#include <linux/clk.h>
+#include <linux/module.h>
+#include <linux/io.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/host.h>
+#include <linux/platform_data/pxa_sdhci.h>
+#include <linux/slab.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+
+#include "sdhci.h"
+#include "sdhci-pltfm.h"
+
+#define SD_FIFO_PARAM		0xe0
+#define DIS_PAD_SD_CLK_GATE	0x0400 /* Turn on/off Dynamic SD Clock Gating */
+#define CLK_GATE_ON		0x0200 /* Disable/enable Clock Gate */
+#define CLK_GATE_CTL		0x0100 /* Clock Gate Control */
+#define CLK_GATE_SETTING_BITS	(DIS_PAD_SD_CLK_GATE | \
+		CLK_GATE_ON | CLK_GATE_CTL)
+
+#define SD_CLOCK_BURST_SIZE_SETUP	0xe6
+#define SDCLK_SEL_SHIFT		8
+#define SDCLK_SEL_MASK		0x3
+#define SDCLK_DELAY_SHIFT	10
+#define SDCLK_DELAY_MASK	0x3c
+
+#define SD_CE_ATA_2		0xea
+#define MMC_CARD		0x1000
+#define MMC_WIDTH		0x0100
+
+static void pxav2_reset(struct sdhci_host *host, u8 mask)
+{
+	struct platform_device *pdev = to_platform_device(mmc_dev(host->mmc));
+	struct sdhci_pxa_platdata *pdata = pdev->dev.platform_data;
+
+	sdhci_reset(host, mask);
+
+	if (mask == SDHCI_RESET_ALL) {
+		u16 tmp = 0;
+
+		/*
+		 * tune timing of read data/command when crc error happen
+		 * no performance impact
+		 */
+		if (pdata && pdata->clk_delay_sel == 1) {
+			tmp = readw(host->ioaddr + SD_CLOCK_BURST_SIZE_SETUP);
+
+			tmp &= ~(SDCLK_DELAY_MASK << SDCLK_DELAY_SHIFT);
+			tmp |= (pdata->clk_delay_cycles & SDCLK_DELAY_MASK)
+				<< SDCLK_DELAY_SHIFT;
+			tmp &= ~(SDCLK_SEL_MASK << SDCLK_SEL_SHIFT);
+			tmp |= (1 & SDCLK_SEL_MASK) << SDCLK_SEL_SHIFT;
+
+			writew(tmp, host->ioaddr + SD_CLOCK_BURST_SIZE_SETUP);
+		}
+
+		if (pdata && (pdata->flags & PXA_FLAG_ENABLE_CLOCK_GATING)) {
+			tmp = readw(host->ioaddr + SD_FIFO_PARAM);
+			tmp &= ~CLK_GATE_SETTING_BITS;
+			writew(tmp, host->ioaddr + SD_FIFO_PARAM);
+		} else {
+			tmp = readw(host->ioaddr + SD_FIFO_PARAM);
+			tmp &= ~CLK_GATE_SETTING_BITS;
+			tmp |= CLK_GATE_SETTING_BITS;
+			writew(tmp, host->ioaddr + SD_FIFO_PARAM);
+		}
+	}
+}
+
+static void pxav2_mmc_set_bus_width(struct sdhci_host *host, int width)
+{
+	u8 ctrl;
+	u16 tmp;
+
+	ctrl = readb(host->ioaddr + SDHCI_HOST_CONTROL);
+	tmp = readw(host->ioaddr + SD_CE_ATA_2);
+	if (width == MMC_BUS_WIDTH_8) {
+		ctrl &= ~SDHCI_CTRL_4BITBUS;
+		tmp |= MMC_CARD | MMC_WIDTH;
+	} else {
+		tmp &= ~(MMC_CARD | MMC_WIDTH);
+		if (width == MMC_BUS_WIDTH_4)
+			ctrl |= SDHCI_CTRL_4BITBUS;
+		else
+			ctrl &= ~SDHCI_CTRL_4BITBUS;
+	}
+	writew(tmp, host->ioaddr + SD_CE_ATA_2);
+	writeb(ctrl, host->ioaddr + SDHCI_HOST_CONTROL);
+}
+
+static const struct sdhci_ops pxav2_sdhci_ops = {
+	.set_clock     = sdhci_set_clock,
+	.get_max_clock = sdhci_pltfm_clk_get_max_clock,
+	.set_bus_width = pxav2_mmc_set_bus_width,
+	.reset         = pxav2_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+#ifdef CONFIG_OF
+static const struct of_device_id sdhci_pxav2_of_match[] = {
+	{
+		.compatible = "mrvl,pxav2-mmc",
+	},
+	{},
+};
+MODULE_DEVICE_TABLE(of, sdhci_pxav2_of_match);
+
+static struct sdhci_pxa_platdata *pxav2_get_mmc_pdata(struct device *dev)
+{
+	struct sdhci_pxa_platdata *pdata;
+	struct device_node *np = dev->of_node;
+	u32 bus_width;
+	u32 clk_delay_cycles;
+
+	pdata = devm_kzalloc(dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata)
+		return NULL;
+
+	if (of_find_property(np, "non-removable", NULL))
+		pdata->flags |= PXA_FLAG_CARD_PERMANENT;
+
+	of_property_read_u32(np, "bus-width", &bus_width);
+	if (bus_width == 8)
+		pdata->flags |= PXA_FLAG_SD_8_BIT_CAPABLE_SLOT;
+
+	of_property_read_u32(np, "mrvl,clk-delay-cycles", &clk_delay_cycles);
+	if (clk_delay_cycles > 0) {
+		pdata->clk_delay_sel = 1;
+		pdata->clk_delay_cycles = clk_delay_cycles;
+	}
+
+	return pdata;
+}
+#else
+static inline struct sdhci_pxa_platdata *pxav2_get_mmc_pdata(struct device *dev)
+{
+	return NULL;
+}
+#endif
+
+static int sdhci_pxav2_probe(struct platform_device *pdev)
+{
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_pxa_platdata *pdata = pdev->dev.platform_data;
+	struct device *dev = &pdev->dev;
+	struct sdhci_host *host = NULL;
+	const struct of_device_id *match;
+
+	int ret;
+	struct clk *clk;
+
+	host = sdhci_pltfm_init(pdev, NULL, 0);
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+
+	clk = devm_clk_get(dev, "PXA-SDHCLK");
+	if (IS_ERR(clk)) {
+		dev_err(dev, "failed to get io clock\n");
+		ret = PTR_ERR(clk);
+		goto free;
+	}
+	pltfm_host->clk = clk;
+	ret = clk_prepare_enable(clk);
+	if (ret) {
+		dev_err(&pdev->dev, "failed to enable io clock\n");
+		goto free;
+	}
+
+	host->quirks = SDHCI_QUIRK_BROKEN_ADMA
+		| SDHCI_QUIRK_BROKEN_TIMEOUT_VAL
+		| SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN;
+
+	match = of_match_device(of_match_ptr(sdhci_pxav2_of_match), &pdev->dev);
+	if (match) {
+		pdata = pxav2_get_mmc_pdata(dev);
+	}
+	if (pdata) {
+		if (pdata->flags & PXA_FLAG_CARD_PERMANENT) {
+			/* on-chip device */
+			host->quirks |= SDHCI_QUIRK_BROKEN_CARD_DETECTION;
+			host->mmc->caps |= MMC_CAP_NONREMOVABLE;
+		}
+
+		/* If slot design supports 8 bit data, indicate this to MMC. */
+		if (pdata->flags & PXA_FLAG_SD_8_BIT_CAPABLE_SLOT)
+			host->mmc->caps |= MMC_CAP_8_BIT_DATA;
+
+		if (pdata->quirks)
+			host->quirks |= pdata->quirks;
+		if (pdata->host_caps)
+			host->mmc->caps |= pdata->host_caps;
+		if (pdata->pm_caps)
+			host->mmc->pm_caps |= pdata->pm_caps;
+	}
+
+	host->ops = &pxav2_sdhci_ops;
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto disable_clk;
+
+	return 0;
+
+disable_clk:
+	clk_disable_unprepare(clk);
+free:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static struct platform_driver sdhci_pxav2_driver = {
+	.driver		= {
+		.name	= "sdhci-pxav2",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(sdhci_pxav2_of_match),
+		.pm	= &sdhci_pltfm_pmops,
+	},
+	.probe		= sdhci_pxav2_probe,
+	.remove		= sdhci_pltfm_unregister,
+};
+
+module_platform_driver(sdhci_pxav2_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for pxav2");
+MODULE_AUTHOR("Marvell International Ltd.");
+MODULE_LICENSE("GPL v2");
+
diff --git a/drivers/mmc/host/sdhci-pxav3.c b/drivers/mmc/host/sdhci-pxav3.c
new file mode 100644
index 000000000..a6d89a3f1
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pxav3.c
@@ -0,0 +1,583 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2010 Marvell International Ltd.
+ *		Zhangfei Gao <zhangfei.gao@marvell.com>
+ *		Kevin Wang <dwang4@marvell.com>
+ *		Mingwei Wang <mwwang@marvell.com>
+ *		Philip Rakity <prakity@marvell.com>
+ *		Mark Brown <markb@marvell.com>
+ */
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/platform_device.h>
+#include <linux/clk.h>
+#include <linux/io.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/host.h>
+#include <linux/platform_data/pxa_sdhci.h>
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+#include <linux/mbus.h>
+
+#include "sdhci.h"
+#include "sdhci-pltfm.h"
+
+#define PXAV3_RPM_DELAY_MS     50
+
+#define SD_CLOCK_BURST_SIZE_SETUP		0x10A
+#define SDCLK_SEL	0x100
+#define SDCLK_DELAY_SHIFT	9
+#define SDCLK_DELAY_MASK	0x1f
+
+#define SD_CFG_FIFO_PARAM       0x100
+#define SDCFG_GEN_PAD_CLK_ON	(1<<6)
+#define SDCFG_GEN_PAD_CLK_CNT_MASK	0xFF
+#define SDCFG_GEN_PAD_CLK_CNT_SHIFT	24
+
+#define SD_SPI_MODE          0x108
+#define SD_CE_ATA_1          0x10C
+
+#define SD_CE_ATA_2          0x10E
+#define SDCE_MISC_INT		(1<<2)
+#define SDCE_MISC_INT_EN	(1<<1)
+
+struct sdhci_pxa {
+	struct clk *clk_core;
+	struct clk *clk_io;
+	u8	power_mode;
+	void __iomem *sdio3_conf_reg;
+};
+
+/*
+ * These registers are relative to the second register region, for the
+ * MBus bridge.
+ */
+#define SDHCI_WINDOW_CTRL(i)	(0x80 + ((i) << 3))
+#define SDHCI_WINDOW_BASE(i)	(0x84 + ((i) << 3))
+#define SDHCI_MAX_WIN_NUM	8
+
+/*
+ * Fields below belong to SDIO3 Configuration Register (third register
+ * region for the Armada 38x flavor)
+ */
+
+#define SDIO3_CONF_CLK_INV	BIT(0)
+#define SDIO3_CONF_SD_FB_CLK	BIT(2)
+
+static int mv_conf_mbus_windows(struct platform_device *pdev,
+				const struct mbus_dram_target_info *dram)
+{
+	int i;
+	void __iomem *regs;
+	struct resource *res;
+
+	if (!dram) {
+		dev_err(&pdev->dev, "no mbus dram info\n");
+		return -EINVAL;
+	}
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+	if (!res) {
+		dev_err(&pdev->dev, "cannot get mbus registers\n");
+		return -EINVAL;
+	}
+
+	regs = ioremap(res->start, resource_size(res));
+	if (!regs) {
+		dev_err(&pdev->dev, "cannot map mbus registers\n");
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < SDHCI_MAX_WIN_NUM; i++) {
+		writel(0, regs + SDHCI_WINDOW_CTRL(i));
+		writel(0, regs + SDHCI_WINDOW_BASE(i));
+	}
+
+	for (i = 0; i < dram->num_cs; i++) {
+		const struct mbus_dram_window *cs = dram->cs + i;
+
+		/* Write size, attributes and target id to control register */
+		writel(((cs->size - 1) & 0xffff0000) |
+			(cs->mbus_attr << 8) |
+			(dram->mbus_dram_target_id << 4) | 1,
+			regs + SDHCI_WINDOW_CTRL(i));
+		/* Write base address to base register */
+		writel(cs->base, regs + SDHCI_WINDOW_BASE(i));
+	}
+
+	iounmap(regs);
+
+	return 0;
+}
+
+static int armada_38x_quirks(struct platform_device *pdev,
+			     struct sdhci_host *host)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_pxa *pxa = sdhci_pltfm_priv(pltfm_host);
+	struct resource *res;
+
+	host->quirks &= ~SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN;
+	host->quirks |= SDHCI_QUIRK_MISSING_CAPS;
+
+	host->caps = sdhci_readl(host, SDHCI_CAPABILITIES);
+	host->caps1 = sdhci_readl(host, SDHCI_CAPABILITIES_1);
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
+					   "conf-sdio3");
+	if (res) {
+		pxa->sdio3_conf_reg = devm_ioremap_resource(&pdev->dev, res);
+		if (IS_ERR(pxa->sdio3_conf_reg))
+			return PTR_ERR(pxa->sdio3_conf_reg);
+	} else {
+		/*
+		 * According to erratum 'FE-2946959' both SDR50 and DDR50
+		 * modes require specific clock adjustments in SDIO3
+		 * Configuration register, if the adjustment is not done,
+		 * remove them from the capabilities.
+		 */
+		host->caps1 &= ~(SDHCI_SUPPORT_SDR50 | SDHCI_SUPPORT_DDR50);
+
+		dev_warn(&pdev->dev, "conf-sdio3 register not found: disabling SDR50 and DDR50 modes.\nConsider updating your dtb\n");
+	}
+
+	/*
+	 * According to erratum 'ERR-7878951' Armada 38x SDHCI
+	 * controller has different capabilities than the ones shown
+	 * in its registers
+	 */
+	if (of_property_read_bool(np, "no-1-8-v")) {
+		host->caps &= ~SDHCI_CAN_VDD_180;
+		host->mmc->caps &= ~MMC_CAP_1_8V_DDR;
+	} else {
+		host->caps &= ~SDHCI_CAN_VDD_330;
+	}
+	host->caps1 &= ~(SDHCI_SUPPORT_SDR104 | SDHCI_USE_SDR50_TUNING);
+
+	return 0;
+}
+
+static void pxav3_reset(struct sdhci_host *host, u8 mask)
+{
+	struct platform_device *pdev = to_platform_device(mmc_dev(host->mmc));
+	struct sdhci_pxa_platdata *pdata = pdev->dev.platform_data;
+
+	sdhci_reset(host, mask);
+
+	if (mask == SDHCI_RESET_ALL) {
+		/*
+		 * tune timing of read data/command when crc error happen
+		 * no performance impact
+		 */
+		if (pdata && 0 != pdata->clk_delay_cycles) {
+			u16 tmp;
+
+			tmp = readw(host->ioaddr + SD_CLOCK_BURST_SIZE_SETUP);
+			tmp |= (pdata->clk_delay_cycles & SDCLK_DELAY_MASK)
+				<< SDCLK_DELAY_SHIFT;
+			tmp |= SDCLK_SEL;
+			writew(tmp, host->ioaddr + SD_CLOCK_BURST_SIZE_SETUP);
+		}
+	}
+}
+
+#define MAX_WAIT_COUNT 5
+static void pxav3_gen_init_74_clocks(struct sdhci_host *host, u8 power_mode)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_pxa *pxa = sdhci_pltfm_priv(pltfm_host);
+	u16 tmp;
+	int count;
+
+	if (pxa->power_mode == MMC_POWER_UP
+			&& power_mode == MMC_POWER_ON) {
+
+		dev_dbg(mmc_dev(host->mmc),
+				"%s: slot->power_mode = %d,"
+				"ios->power_mode = %d\n",
+				__func__,
+				pxa->power_mode,
+				power_mode);
+
+		/* set we want notice of when 74 clocks are sent */
+		tmp = readw(host->ioaddr + SD_CE_ATA_2);
+		tmp |= SDCE_MISC_INT_EN;
+		writew(tmp, host->ioaddr + SD_CE_ATA_2);
+
+		/* start sending the 74 clocks */
+		tmp = readw(host->ioaddr + SD_CFG_FIFO_PARAM);
+		tmp |= SDCFG_GEN_PAD_CLK_ON;
+		writew(tmp, host->ioaddr + SD_CFG_FIFO_PARAM);
+
+		/* slowest speed is about 100KHz or 10usec per clock */
+		udelay(740);
+		count = 0;
+
+		while (count++ < MAX_WAIT_COUNT) {
+			if ((readw(host->ioaddr + SD_CE_ATA_2)
+						& SDCE_MISC_INT) == 0)
+				break;
+			udelay(10);
+		}
+
+		if (count == MAX_WAIT_COUNT)
+			dev_warn(mmc_dev(host->mmc), "74 clock interrupt not cleared\n");
+
+		/* clear the interrupt bit if posted */
+		tmp = readw(host->ioaddr + SD_CE_ATA_2);
+		tmp |= SDCE_MISC_INT;
+		writew(tmp, host->ioaddr + SD_CE_ATA_2);
+	}
+	pxa->power_mode = power_mode;
+}
+
+static void pxav3_set_uhs_signaling(struct sdhci_host *host, unsigned int uhs)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_pxa *pxa = sdhci_pltfm_priv(pltfm_host);
+	u16 ctrl_2;
+
+	/*
+	 * Set V18_EN -- UHS modes do not work without this.
+	 * does not change signaling voltage
+	 */
+	ctrl_2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+
+	/* Select Bus Speed Mode for host */
+	ctrl_2 &= ~SDHCI_CTRL_UHS_MASK;
+	switch (uhs) {
+	case MMC_TIMING_UHS_SDR12:
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR12;
+		break;
+	case MMC_TIMING_UHS_SDR25:
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR25;
+		break;
+	case MMC_TIMING_UHS_SDR50:
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR50 | SDHCI_CTRL_VDD_180;
+		break;
+	case MMC_TIMING_UHS_SDR104:
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR104 | SDHCI_CTRL_VDD_180;
+		break;
+	case MMC_TIMING_MMC_DDR52:
+	case MMC_TIMING_UHS_DDR50:
+		ctrl_2 |= SDHCI_CTRL_UHS_DDR50 | SDHCI_CTRL_VDD_180;
+		break;
+	}
+
+	/*
+	 * Update SDIO3 Configuration register according to erratum
+	 * FE-2946959
+	 */
+	if (pxa->sdio3_conf_reg) {
+		u8 reg_val  = readb(pxa->sdio3_conf_reg);
+
+		if (uhs == MMC_TIMING_UHS_SDR50 ||
+		    uhs == MMC_TIMING_UHS_DDR50) {
+			reg_val &= ~SDIO3_CONF_CLK_INV;
+			reg_val |= SDIO3_CONF_SD_FB_CLK;
+		} else if (uhs == MMC_TIMING_MMC_HS) {
+			reg_val &= ~SDIO3_CONF_CLK_INV;
+			reg_val &= ~SDIO3_CONF_SD_FB_CLK;
+		} else {
+			reg_val |= SDIO3_CONF_CLK_INV;
+			reg_val &= ~SDIO3_CONF_SD_FB_CLK;
+		}
+		writeb(reg_val, pxa->sdio3_conf_reg);
+	}
+
+	sdhci_writew(host, ctrl_2, SDHCI_HOST_CONTROL2);
+	dev_dbg(mmc_dev(host->mmc),
+		"%s uhs = %d, ctrl_2 = %04X\n",
+		__func__, uhs, ctrl_2);
+}
+
+static void pxav3_set_power(struct sdhci_host *host, unsigned char mode,
+			    unsigned short vdd)
+{
+	struct mmc_host *mmc = host->mmc;
+	u8 pwr = host->pwr;
+
+	sdhci_set_power_noreg(host, mode, vdd);
+
+	if (host->pwr == pwr)
+		return;
+
+	if (host->pwr == 0)
+		vdd = 0;
+
+	if (!IS_ERR(mmc->supply.vmmc))
+		mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, vdd);
+}
+
+static const struct sdhci_ops pxav3_sdhci_ops = {
+	.set_clock = sdhci_set_clock,
+	.set_power = pxav3_set_power,
+	.platform_send_init_74_clocks = pxav3_gen_init_74_clocks,
+	.get_max_clock = sdhci_pltfm_clk_get_max_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = pxav3_reset,
+	.set_uhs_signaling = pxav3_set_uhs_signaling,
+};
+
+static const struct sdhci_pltfm_data sdhci_pxav3_pdata = {
+	.quirks = SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK
+		| SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC
+		| SDHCI_QUIRK_32BIT_ADMA_SIZE
+		| SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+	.ops = &pxav3_sdhci_ops,
+};
+
+#ifdef CONFIG_OF
+static const struct of_device_id sdhci_pxav3_of_match[] = {
+	{
+		.compatible = "mrvl,pxav3-mmc",
+	},
+	{
+		.compatible = "marvell,armada-380-sdhci",
+	},
+	{},
+};
+MODULE_DEVICE_TABLE(of, sdhci_pxav3_of_match);
+
+static struct sdhci_pxa_platdata *pxav3_get_mmc_pdata(struct device *dev)
+{
+	struct sdhci_pxa_platdata *pdata;
+	struct device_node *np = dev->of_node;
+	u32 clk_delay_cycles;
+
+	pdata = devm_kzalloc(dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata)
+		return NULL;
+
+	if (!of_property_read_u32(np, "mrvl,clk-delay-cycles",
+				  &clk_delay_cycles))
+		pdata->clk_delay_cycles = clk_delay_cycles;
+
+	return pdata;
+}
+#else
+static inline struct sdhci_pxa_platdata *pxav3_get_mmc_pdata(struct device *dev)
+{
+	return NULL;
+}
+#endif
+
+static int sdhci_pxav3_probe(struct platform_device *pdev)
+{
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_pxa_platdata *pdata = pdev->dev.platform_data;
+	struct device *dev = &pdev->dev;
+	struct device_node *np = pdev->dev.of_node;
+	struct sdhci_host *host = NULL;
+	struct sdhci_pxa *pxa = NULL;
+	const struct of_device_id *match;
+	int ret;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_pxav3_pdata, sizeof(*pxa));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	pxa = sdhci_pltfm_priv(pltfm_host);
+
+	pxa->clk_io = devm_clk_get(dev, "io");
+	if (IS_ERR(pxa->clk_io))
+		pxa->clk_io = devm_clk_get(dev, NULL);
+	if (IS_ERR(pxa->clk_io)) {
+		dev_err(dev, "failed to get io clock\n");
+		ret = PTR_ERR(pxa->clk_io);
+		goto err_clk_get;
+	}
+	pltfm_host->clk = pxa->clk_io;
+	clk_prepare_enable(pxa->clk_io);
+
+	pxa->clk_core = devm_clk_get(dev, "core");
+	if (!IS_ERR(pxa->clk_core))
+		clk_prepare_enable(pxa->clk_core);
+
+	/* enable 1/8V DDR capable */
+	host->mmc->caps |= MMC_CAP_1_8V_DDR;
+
+	if (of_device_is_compatible(np, "marvell,armada-380-sdhci")) {
+		ret = armada_38x_quirks(pdev, host);
+		if (ret < 0)
+			goto err_mbus_win;
+		ret = mv_conf_mbus_windows(pdev, mv_mbus_dram_info());
+		if (ret < 0)
+			goto err_mbus_win;
+	}
+
+	match = of_match_device(of_match_ptr(sdhci_pxav3_of_match), &pdev->dev);
+	if (match) {
+		ret = mmc_of_parse(host->mmc);
+		if (ret)
+			goto err_of_parse;
+		sdhci_get_of_property(pdev);
+		pdata = pxav3_get_mmc_pdata(dev);
+		pdev->dev.platform_data = pdata;
+	} else if (pdata) {
+		/* on-chip device */
+		if (pdata->flags & PXA_FLAG_CARD_PERMANENT)
+			host->mmc->caps |= MMC_CAP_NONREMOVABLE;
+
+		/* If slot design supports 8 bit data, indicate this to MMC. */
+		if (pdata->flags & PXA_FLAG_SD_8_BIT_CAPABLE_SLOT)
+			host->mmc->caps |= MMC_CAP_8_BIT_DATA;
+
+		if (pdata->quirks)
+			host->quirks |= pdata->quirks;
+		if (pdata->quirks2)
+			host->quirks2 |= pdata->quirks2;
+		if (pdata->host_caps)
+			host->mmc->caps |= pdata->host_caps;
+		if (pdata->host_caps2)
+			host->mmc->caps2 |= pdata->host_caps2;
+		if (pdata->pm_caps)
+			host->mmc->pm_caps |= pdata->pm_caps;
+	}
+
+	pm_runtime_get_noresume(&pdev->dev);
+	pm_runtime_set_active(&pdev->dev);
+	pm_runtime_set_autosuspend_delay(&pdev->dev, PXAV3_RPM_DELAY_MS);
+	pm_runtime_use_autosuspend(&pdev->dev);
+	pm_runtime_enable(&pdev->dev);
+	pm_suspend_ignore_children(&pdev->dev, 1);
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err_add_host;
+
+	if (host->mmc->pm_caps & MMC_PM_WAKE_SDIO_IRQ)
+		device_init_wakeup(&pdev->dev, 1);
+
+	pm_runtime_put_autosuspend(&pdev->dev);
+
+	return 0;
+
+err_add_host:
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+err_of_parse:
+err_mbus_win:
+	clk_disable_unprepare(pxa->clk_io);
+	clk_disable_unprepare(pxa->clk_core);
+err_clk_get:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static int sdhci_pxav3_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_pxa *pxa = sdhci_pltfm_priv(pltfm_host);
+
+	pm_runtime_get_sync(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+
+	sdhci_remove_host(host, 1);
+
+	clk_disable_unprepare(pxa->clk_io);
+	clk_disable_unprepare(pxa->clk_core);
+
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int sdhci_pxav3_suspend(struct device *dev)
+{
+	int ret;
+	struct sdhci_host *host = dev_get_drvdata(dev);
+
+	pm_runtime_get_sync(dev);
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+	ret = sdhci_suspend_host(host);
+	pm_runtime_mark_last_busy(dev);
+	pm_runtime_put_autosuspend(dev);
+
+	return ret;
+}
+
+static int sdhci_pxav3_resume(struct device *dev)
+{
+	int ret;
+	struct sdhci_host *host = dev_get_drvdata(dev);
+
+	pm_runtime_get_sync(dev);
+	ret = sdhci_resume_host(host);
+	pm_runtime_mark_last_busy(dev);
+	pm_runtime_put_autosuspend(dev);
+
+	return ret;
+}
+#endif
+
+#ifdef CONFIG_PM
+static int sdhci_pxav3_runtime_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_pxa *pxa = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	ret = sdhci_runtime_suspend_host(host);
+	if (ret)
+		return ret;
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	clk_disable_unprepare(pxa->clk_io);
+	if (!IS_ERR(pxa->clk_core))
+		clk_disable_unprepare(pxa->clk_core);
+
+	return 0;
+}
+
+static int sdhci_pxav3_runtime_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_pxa *pxa = sdhci_pltfm_priv(pltfm_host);
+
+	clk_prepare_enable(pxa->clk_io);
+	if (!IS_ERR(pxa->clk_core))
+		clk_prepare_enable(pxa->clk_core);
+
+	return sdhci_runtime_resume_host(host, 0);
+}
+#endif
+
+static const struct dev_pm_ops sdhci_pxav3_pmops = {
+	SET_SYSTEM_SLEEP_PM_OPS(sdhci_pxav3_suspend, sdhci_pxav3_resume)
+	SET_RUNTIME_PM_OPS(sdhci_pxav3_runtime_suspend,
+		sdhci_pxav3_runtime_resume, NULL)
+};
+
+static struct platform_driver sdhci_pxav3_driver = {
+	.driver		= {
+		.name	= "sdhci-pxav3",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(sdhci_pxav3_of_match),
+		.pm	= &sdhci_pxav3_pmops,
+	},
+	.probe		= sdhci_pxav3_probe,
+	.remove		= sdhci_pxav3_remove,
+};
+
+module_platform_driver(sdhci_pxav3_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for pxav3");
+MODULE_AUTHOR("Marvell International Ltd.");
+MODULE_LICENSE("GPL v2");
+
diff --git a/drivers/mmc/host/sdhci-s3c.c b/drivers/mmc/host/sdhci-s3c.c
new file mode 100644
index 000000000..f48a788a9
--- /dev/null
+++ b/drivers/mmc/host/sdhci-s3c.c
@@ -0,0 +1,800 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* linux/drivers/mmc/host/sdhci-s3c.c
+ *
+ * Copyright 2008 Openmoko Inc.
+ * Copyright 2008 Simtec Electronics
+ *      Ben Dooks <ben@simtec.co.uk>
+ *      http://armlinux.simtec.co.uk/
+ *
+ * SDHCI (HSMMC) support for Samsung SoC
+ */
+
+#include <linux/spinlock.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/platform_device.h>
+#include <linux/platform_data/mmc-sdhci-s3c.h>
+#include <linux/slab.h>
+#include <linux/clk.h>
+#include <linux/io.h>
+#include <linux/gpio.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_gpio.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+
+#include <linux/mmc/host.h>
+
+#include "sdhci.h"
+
+#define MAX_BUS_CLK	(4)
+
+#define S3C_SDHCI_CONTROL2			(0x80)
+#define S3C_SDHCI_CONTROL3			(0x84)
+#define S3C64XX_SDHCI_CONTROL4			(0x8C)
+
+#define S3C64XX_SDHCI_CTRL2_ENSTAASYNCCLR	BIT(31)
+#define S3C64XX_SDHCI_CTRL2_ENCMDCNFMSK		BIT(30)
+#define S3C_SDHCI_CTRL2_CDINVRXD3		BIT(29)
+#define S3C_SDHCI_CTRL2_SLCARDOUT		BIT(28)
+
+#define S3C_SDHCI_CTRL2_FLTCLKSEL_MASK		(0xf << 24)
+#define S3C_SDHCI_CTRL2_FLTCLKSEL_SHIFT		(24)
+#define S3C_SDHCI_CTRL2_FLTCLKSEL(_x)		((_x) << 24)
+
+#define S3C_SDHCI_CTRL2_LVLDAT_MASK		(0xff << 16)
+#define S3C_SDHCI_CTRL2_LVLDAT_SHIFT		(16)
+#define S3C_SDHCI_CTRL2_LVLDAT(_x)		((_x) << 16)
+
+#define S3C_SDHCI_CTRL2_ENFBCLKTX		BIT(15)
+#define S3C_SDHCI_CTRL2_ENFBCLKRX		BIT(14)
+#define S3C_SDHCI_CTRL2_SDCDSEL			BIT(13)
+#define S3C_SDHCI_CTRL2_SDSIGPC			BIT(12)
+#define S3C_SDHCI_CTRL2_ENBUSYCHKTXSTART	BIT(11)
+
+#define S3C_SDHCI_CTRL2_DFCNT_MASK		(0x3 << 9)
+#define S3C_SDHCI_CTRL2_DFCNT_SHIFT		(9)
+#define S3C_SDHCI_CTRL2_DFCNT_NONE		(0x0 << 9)
+#define S3C_SDHCI_CTRL2_DFCNT_4SDCLK		(0x1 << 9)
+#define S3C_SDHCI_CTRL2_DFCNT_16SDCLK		(0x2 << 9)
+#define S3C_SDHCI_CTRL2_DFCNT_64SDCLK		(0x3 << 9)
+
+#define S3C_SDHCI_CTRL2_ENCLKOUTHOLD		BIT(8)
+#define S3C_SDHCI_CTRL2_RWAITMODE		BIT(7)
+#define S3C_SDHCI_CTRL2_DISBUFRD		BIT(6)
+
+#define S3C_SDHCI_CTRL2_SELBASECLK_MASK		(0x3 << 4)
+#define S3C_SDHCI_CTRL2_SELBASECLK_SHIFT	(4)
+#define S3C_SDHCI_CTRL2_PWRSYNC			BIT(3)
+#define S3C_SDHCI_CTRL2_ENCLKOUTMSKCON		BIT(1)
+#define S3C_SDHCI_CTRL2_HWINITFIN		BIT(0)
+
+#define S3C_SDHCI_CTRL3_FCSEL3			BIT(31)
+#define S3C_SDHCI_CTRL3_FCSEL2			BIT(23)
+#define S3C_SDHCI_CTRL3_FCSEL1			BIT(15)
+#define S3C_SDHCI_CTRL3_FCSEL0			BIT(7)
+
+#define S3C_SDHCI_CTRL3_FIA3_MASK		(0x7f << 24)
+#define S3C_SDHCI_CTRL3_FIA3_SHIFT		(24)
+#define S3C_SDHCI_CTRL3_FIA3(_x)		((_x) << 24)
+
+#define S3C_SDHCI_CTRL3_FIA2_MASK		(0x7f << 16)
+#define S3C_SDHCI_CTRL3_FIA2_SHIFT		(16)
+#define S3C_SDHCI_CTRL3_FIA2(_x)		((_x) << 16)
+
+#define S3C_SDHCI_CTRL3_FIA1_MASK		(0x7f << 8)
+#define S3C_SDHCI_CTRL3_FIA1_SHIFT		(8)
+#define S3C_SDHCI_CTRL3_FIA1(_x)		((_x) << 8)
+
+#define S3C_SDHCI_CTRL3_FIA0_MASK		(0x7f << 0)
+#define S3C_SDHCI_CTRL3_FIA0_SHIFT		(0)
+#define S3C_SDHCI_CTRL3_FIA0(_x)		((_x) << 0)
+
+#define S3C64XX_SDHCI_CONTROL4_DRIVE_MASK	(0x3 << 16)
+#define S3C64XX_SDHCI_CONTROL4_DRIVE_SHIFT	(16)
+#define S3C64XX_SDHCI_CONTROL4_DRIVE_2mA	(0x0 << 16)
+#define S3C64XX_SDHCI_CONTROL4_DRIVE_4mA	(0x1 << 16)
+#define S3C64XX_SDHCI_CONTROL4_DRIVE_7mA	(0x2 << 16)
+#define S3C64XX_SDHCI_CONTROL4_DRIVE_9mA	(0x3 << 16)
+
+#define S3C64XX_SDHCI_CONTROL4_BUSY		(1)
+
+/**
+ * struct sdhci_s3c - S3C SDHCI instance
+ * @host: The SDHCI host created
+ * @pdev: The platform device we where created from.
+ * @ioarea: The resource created when we claimed the IO area.
+ * @pdata: The platform data for this controller.
+ * @cur_clk: The index of the current bus clock.
+ * @ext_cd_irq: External card detect interrupt.
+ * @clk_io: The clock for the internal bus interface.
+ * @clk_rates: Clock frequencies.
+ * @clk_bus: The clocks that are available for the SD/MMC bus clock.
+ * @no_divider: No or non-standard internal clock divider.
+ */
+struct sdhci_s3c {
+	struct sdhci_host	*host;
+	struct platform_device	*pdev;
+	struct resource		*ioarea;
+	struct s3c_sdhci_platdata *pdata;
+	int			cur_clk;
+	int			ext_cd_irq;
+
+	struct clk		*clk_io;
+	struct clk		*clk_bus[MAX_BUS_CLK];
+	unsigned long		clk_rates[MAX_BUS_CLK];
+
+	bool			no_divider;
+};
+
+/**
+ * struct sdhci_s3c_driver_data - S3C SDHCI platform specific driver data
+ * @sdhci_quirks: sdhci host specific quirks.
+ * @no_divider: no or non-standard internal clock divider.
+ *
+ * Specifies platform specific configuration of sdhci controller.
+ * Note: A structure for driver specific platform data is used for future
+ * expansion of its usage.
+ */
+struct sdhci_s3c_drv_data {
+	unsigned int	sdhci_quirks;
+	bool		no_divider;
+};
+
+static inline struct sdhci_s3c *to_s3c(struct sdhci_host *host)
+{
+	return sdhci_priv(host);
+}
+
+/**
+ * sdhci_s3c_get_max_clk - callback to get maximum clock frequency.
+ * @host: The SDHCI host instance.
+ *
+ * Callback to return the maximum clock rate acheivable by the controller.
+*/
+static unsigned int sdhci_s3c_get_max_clk(struct sdhci_host *host)
+{
+	struct sdhci_s3c *ourhost = to_s3c(host);
+	unsigned long rate, max = 0;
+	int src;
+
+	for (src = 0; src < MAX_BUS_CLK; src++) {
+		rate = ourhost->clk_rates[src];
+		if (rate > max)
+			max = rate;
+	}
+
+	return max;
+}
+
+/**
+ * sdhci_s3c_consider_clock - consider one the bus clocks for current setting
+ * @ourhost: Our SDHCI instance.
+ * @src: The source clock index.
+ * @wanted: The clock frequency wanted.
+ */
+static unsigned int sdhci_s3c_consider_clock(struct sdhci_s3c *ourhost,
+					     unsigned int src,
+					     unsigned int wanted)
+{
+	unsigned long rate;
+	struct clk *clksrc = ourhost->clk_bus[src];
+	int shift;
+
+	if (IS_ERR(clksrc))
+		return UINT_MAX;
+
+	/*
+	 * If controller uses a non-standard clock division, find the best clock
+	 * speed possible with selected clock source and skip the division.
+	 */
+	if (ourhost->no_divider) {
+		rate = clk_round_rate(clksrc, wanted);
+		return wanted - rate;
+	}
+
+	rate = ourhost->clk_rates[src];
+
+	for (shift = 0; shift <= 8; ++shift) {
+		if ((rate >> shift) <= wanted)
+			break;
+	}
+
+	if (shift > 8) {
+		dev_dbg(&ourhost->pdev->dev,
+			"clk %d: rate %ld, min rate %lu > wanted %u\n",
+			src, rate, rate / 256, wanted);
+		return UINT_MAX;
+	}
+
+	dev_dbg(&ourhost->pdev->dev, "clk %d: rate %ld, want %d, got %ld\n",
+		src, rate, wanted, rate >> shift);
+
+	return wanted - (rate >> shift);
+}
+
+/**
+ * sdhci_s3c_set_clock - callback on clock change
+ * @host: The SDHCI host being changed
+ * @clock: The clock rate being requested.
+ *
+ * When the card's clock is going to be changed, look at the new frequency
+ * and find the best clock source to go with it.
+*/
+static void sdhci_s3c_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	struct sdhci_s3c *ourhost = to_s3c(host);
+	unsigned int best = UINT_MAX;
+	unsigned int delta;
+	int best_src = 0;
+	int src;
+	u32 ctrl;
+
+	host->mmc->actual_clock = 0;
+
+	/* don't bother if the clock is going off. */
+	if (clock == 0) {
+		sdhci_set_clock(host, clock);
+		return;
+	}
+
+	for (src = 0; src < MAX_BUS_CLK; src++) {
+		delta = sdhci_s3c_consider_clock(ourhost, src, clock);
+		if (delta < best) {
+			best = delta;
+			best_src = src;
+		}
+	}
+
+	dev_dbg(&ourhost->pdev->dev,
+		"selected source %d, clock %d, delta %d\n",
+		 best_src, clock, best);
+
+	/* select the new clock source */
+	if (ourhost->cur_clk != best_src) {
+		struct clk *clk = ourhost->clk_bus[best_src];
+
+		clk_prepare_enable(clk);
+		if (ourhost->cur_clk >= 0)
+			clk_disable_unprepare(
+					ourhost->clk_bus[ourhost->cur_clk]);
+
+		ourhost->cur_clk = best_src;
+		host->max_clk = ourhost->clk_rates[best_src];
+	}
+
+	/* turn clock off to card before changing clock source */
+	writew(0, host->ioaddr + SDHCI_CLOCK_CONTROL);
+
+	ctrl = readl(host->ioaddr + S3C_SDHCI_CONTROL2);
+	ctrl &= ~S3C_SDHCI_CTRL2_SELBASECLK_MASK;
+	ctrl |= best_src << S3C_SDHCI_CTRL2_SELBASECLK_SHIFT;
+	writel(ctrl, host->ioaddr + S3C_SDHCI_CONTROL2);
+
+	/* reprogram default hardware configuration */
+	writel(S3C64XX_SDHCI_CONTROL4_DRIVE_9mA,
+		host->ioaddr + S3C64XX_SDHCI_CONTROL4);
+
+	ctrl = readl(host->ioaddr + S3C_SDHCI_CONTROL2);
+	ctrl |= (S3C64XX_SDHCI_CTRL2_ENSTAASYNCCLR |
+		  S3C64XX_SDHCI_CTRL2_ENCMDCNFMSK |
+		  S3C_SDHCI_CTRL2_ENFBCLKRX |
+		  S3C_SDHCI_CTRL2_DFCNT_NONE |
+		  S3C_SDHCI_CTRL2_ENCLKOUTHOLD);
+	writel(ctrl, host->ioaddr + S3C_SDHCI_CONTROL2);
+
+	/* reconfigure the controller for new clock rate */
+	ctrl = (S3C_SDHCI_CTRL3_FCSEL1 | S3C_SDHCI_CTRL3_FCSEL0);
+	if (clock < 25 * 1000000)
+		ctrl |= (S3C_SDHCI_CTRL3_FCSEL3 | S3C_SDHCI_CTRL3_FCSEL2);
+	writel(ctrl, host->ioaddr + S3C_SDHCI_CONTROL3);
+
+	sdhci_set_clock(host, clock);
+}
+
+/**
+ * sdhci_s3c_get_min_clock - callback to get minimal supported clock value
+ * @host: The SDHCI host being queried
+ *
+ * To init mmc host properly a minimal clock value is needed. For high system
+ * bus clock's values the standard formula gives values out of allowed range.
+ * The clock still can be set to lower values, if clock source other then
+ * system bus is selected.
+*/
+static unsigned int sdhci_s3c_get_min_clock(struct sdhci_host *host)
+{
+	struct sdhci_s3c *ourhost = to_s3c(host);
+	unsigned long rate, min = ULONG_MAX;
+	int src;
+
+	for (src = 0; src < MAX_BUS_CLK; src++) {
+		rate = ourhost->clk_rates[src] / 256;
+		if (!rate)
+			continue;
+		if (rate < min)
+			min = rate;
+	}
+
+	return min;
+}
+
+/* sdhci_cmu_get_max_clk - callback to get maximum clock frequency.*/
+static unsigned int sdhci_cmu_get_max_clock(struct sdhci_host *host)
+{
+	struct sdhci_s3c *ourhost = to_s3c(host);
+	unsigned long rate, max = 0;
+	int src;
+
+	for (src = 0; src < MAX_BUS_CLK; src++) {
+		struct clk *clk;
+
+		clk = ourhost->clk_bus[src];
+		if (IS_ERR(clk))
+			continue;
+
+		rate = clk_round_rate(clk, ULONG_MAX);
+		if (rate > max)
+			max = rate;
+	}
+
+	return max;
+}
+
+/* sdhci_cmu_get_min_clock - callback to get minimal supported clock value. */
+static unsigned int sdhci_cmu_get_min_clock(struct sdhci_host *host)
+{
+	struct sdhci_s3c *ourhost = to_s3c(host);
+	unsigned long rate, min = ULONG_MAX;
+	int src;
+
+	for (src = 0; src < MAX_BUS_CLK; src++) {
+		struct clk *clk;
+
+		clk = ourhost->clk_bus[src];
+		if (IS_ERR(clk))
+			continue;
+
+		rate = clk_round_rate(clk, 0);
+		if (rate < min)
+			min = rate;
+	}
+
+	return min;
+}
+
+/* sdhci_cmu_set_clock - callback on clock change.*/
+static void sdhci_cmu_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	struct sdhci_s3c *ourhost = to_s3c(host);
+	struct device *dev = &ourhost->pdev->dev;
+	unsigned long timeout;
+	u16 clk = 0;
+	int ret;
+
+	host->mmc->actual_clock = 0;
+
+	/* If the clock is going off, set to 0 at clock control register */
+	if (clock == 0) {
+		sdhci_writew(host, 0, SDHCI_CLOCK_CONTROL);
+		return;
+	}
+
+	sdhci_s3c_set_clock(host, clock);
+
+	/* Reset SD Clock Enable */
+	clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+	clk &= ~SDHCI_CLOCK_CARD_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+	ret = clk_set_rate(ourhost->clk_bus[ourhost->cur_clk], clock);
+	if (ret != 0) {
+		dev_err(dev, "%s: failed to set clock rate %uHz\n",
+			mmc_hostname(host->mmc), clock);
+		return;
+	}
+
+	clk = SDHCI_CLOCK_INT_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+	/* Wait max 20 ms */
+	timeout = 20;
+	while (!((clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL))
+		& SDHCI_CLOCK_INT_STABLE)) {
+		if (timeout == 0) {
+			dev_err(dev, "%s: Internal clock never stabilised.\n",
+				mmc_hostname(host->mmc));
+			return;
+		}
+		timeout--;
+		mdelay(1);
+	}
+
+	clk |= SDHCI_CLOCK_CARD_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+}
+
+static struct sdhci_ops sdhci_s3c_ops = {
+	.get_max_clock		= sdhci_s3c_get_max_clk,
+	.set_clock		= sdhci_s3c_set_clock,
+	.get_min_clock		= sdhci_s3c_get_min_clock,
+	.set_bus_width		= sdhci_set_bus_width,
+	.reset			= sdhci_reset,
+	.set_uhs_signaling	= sdhci_set_uhs_signaling,
+};
+
+#ifdef CONFIG_OF
+static int sdhci_s3c_parse_dt(struct device *dev,
+		struct sdhci_host *host, struct s3c_sdhci_platdata *pdata)
+{
+	struct device_node *node = dev->of_node;
+	u32 max_width;
+
+	/* if the bus-width property is not specified, assume width as 1 */
+	if (of_property_read_u32(node, "bus-width", &max_width))
+		max_width = 1;
+	pdata->max_width = max_width;
+
+	/* get the card detection method */
+	if (of_get_property(node, "broken-cd", NULL)) {
+		pdata->cd_type = S3C_SDHCI_CD_NONE;
+		return 0;
+	}
+
+	if (of_get_property(node, "non-removable", NULL)) {
+		pdata->cd_type = S3C_SDHCI_CD_PERMANENT;
+		return 0;
+	}
+
+	if (of_get_named_gpio(node, "cd-gpios", 0))
+		return 0;
+
+	/* assuming internal card detect that will be configured by pinctrl */
+	pdata->cd_type = S3C_SDHCI_CD_INTERNAL;
+	return 0;
+}
+#else
+static int sdhci_s3c_parse_dt(struct device *dev,
+		struct sdhci_host *host, struct s3c_sdhci_platdata *pdata)
+{
+	return -EINVAL;
+}
+#endif
+
+#ifdef CONFIG_OF
+static const struct of_device_id sdhci_s3c_dt_match[];
+#endif
+
+static inline struct sdhci_s3c_drv_data *sdhci_s3c_get_driver_data(
+			struct platform_device *pdev)
+{
+#ifdef CONFIG_OF
+	if (pdev->dev.of_node) {
+		const struct of_device_id *match;
+		match = of_match_node(sdhci_s3c_dt_match, pdev->dev.of_node);
+		return (struct sdhci_s3c_drv_data *)match->data;
+	}
+#endif
+	return (struct sdhci_s3c_drv_data *)
+			platform_get_device_id(pdev)->driver_data;
+}
+
+static int sdhci_s3c_probe(struct platform_device *pdev)
+{
+	struct s3c_sdhci_platdata *pdata;
+	struct sdhci_s3c_drv_data *drv_data;
+	struct device *dev = &pdev->dev;
+	struct sdhci_host *host;
+	struct sdhci_s3c *sc;
+	int ret, irq, ptr, clks;
+
+	if (!pdev->dev.platform_data && !pdev->dev.of_node) {
+		dev_err(dev, "no device data specified\n");
+		return -ENOENT;
+	}
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return irq;
+
+	host = sdhci_alloc_host(dev, sizeof(struct sdhci_s3c));
+	if (IS_ERR(host)) {
+		dev_err(dev, "sdhci_alloc_host() failed\n");
+		return PTR_ERR(host);
+	}
+	sc = sdhci_priv(host);
+
+	pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata) {
+		ret = -ENOMEM;
+		goto err_pdata_io_clk;
+	}
+
+	if (pdev->dev.of_node) {
+		ret = sdhci_s3c_parse_dt(&pdev->dev, host, pdata);
+		if (ret)
+			goto err_pdata_io_clk;
+	} else {
+		memcpy(pdata, pdev->dev.platform_data, sizeof(*pdata));
+	}
+
+	drv_data = sdhci_s3c_get_driver_data(pdev);
+
+	sc->host = host;
+	sc->pdev = pdev;
+	sc->pdata = pdata;
+	sc->cur_clk = -1;
+
+	platform_set_drvdata(pdev, host);
+
+	sc->clk_io = devm_clk_get(dev, "hsmmc");
+	if (IS_ERR(sc->clk_io)) {
+		dev_err(dev, "failed to get io clock\n");
+		ret = PTR_ERR(sc->clk_io);
+		goto err_pdata_io_clk;
+	}
+
+	/* enable the local io clock and keep it running for the moment. */
+	clk_prepare_enable(sc->clk_io);
+
+	for (clks = 0, ptr = 0; ptr < MAX_BUS_CLK; ptr++) {
+		char name[14];
+
+		snprintf(name, 14, "mmc_busclk.%d", ptr);
+		sc->clk_bus[ptr] = devm_clk_get(dev, name);
+		if (IS_ERR(sc->clk_bus[ptr]))
+			continue;
+
+		clks++;
+		sc->clk_rates[ptr] = clk_get_rate(sc->clk_bus[ptr]);
+
+		dev_info(dev, "clock source %d: %s (%ld Hz)\n",
+				ptr, name, sc->clk_rates[ptr]);
+	}
+
+	if (clks == 0) {
+		dev_err(dev, "failed to find any bus clocks\n");
+		ret = -ENOENT;
+		goto err_no_busclks;
+	}
+
+	host->ioaddr = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(host->ioaddr)) {
+		ret = PTR_ERR(host->ioaddr);
+		goto err_req_regs;
+	}
+
+	/* Ensure we have minimal gpio selected CMD/CLK/Detect */
+	if (pdata->cfg_gpio)
+		pdata->cfg_gpio(pdev, pdata->max_width);
+
+	host->hw_name = "samsung-hsmmc";
+	host->ops = &sdhci_s3c_ops;
+	host->quirks = 0;
+	host->quirks2 = 0;
+	host->irq = irq;
+
+	/* Setup quirks for the controller */
+	host->quirks |= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC;
+	host->quirks |= SDHCI_QUIRK_NO_HISPD_BIT;
+	if (drv_data) {
+		host->quirks |= drv_data->sdhci_quirks;
+		sc->no_divider = drv_data->no_divider;
+	}
+
+#ifndef CONFIG_MMC_SDHCI_S3C_DMA
+
+	/* we currently see overruns on errors, so disable the SDMA
+	 * support as well. */
+	host->quirks |= SDHCI_QUIRK_BROKEN_DMA;
+
+#endif /* CONFIG_MMC_SDHCI_S3C_DMA */
+
+	/* It seems we do not get an DATA transfer complete on non-busy
+	 * transfers, not sure if this is a problem with this specific
+	 * SDHCI block, or a missing configuration that needs to be set. */
+	host->quirks |= SDHCI_QUIRK_NO_BUSY_IRQ;
+
+	/* This host supports the Auto CMD12 */
+	host->quirks |= SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12;
+
+	/* Samsung SoCs need BROKEN_ADMA_ZEROLEN_DESC */
+	host->quirks |= SDHCI_QUIRK_BROKEN_ADMA_ZEROLEN_DESC;
+
+	if (pdata->cd_type == S3C_SDHCI_CD_NONE ||
+	    pdata->cd_type == S3C_SDHCI_CD_PERMANENT)
+		host->quirks |= SDHCI_QUIRK_BROKEN_CARD_DETECTION;
+
+	if (pdata->cd_type == S3C_SDHCI_CD_PERMANENT)
+		host->mmc->caps = MMC_CAP_NONREMOVABLE;
+
+	switch (pdata->max_width) {
+	case 8:
+		host->mmc->caps |= MMC_CAP_8_BIT_DATA;
+		fallthrough;
+	case 4:
+		host->mmc->caps |= MMC_CAP_4_BIT_DATA;
+		break;
+	}
+
+	if (pdata->pm_caps)
+		host->mmc->pm_caps |= pdata->pm_caps;
+
+	host->quirks |= (SDHCI_QUIRK_32BIT_DMA_ADDR |
+			 SDHCI_QUIRK_32BIT_DMA_SIZE);
+
+	/* HSMMC on Samsung SoCs uses SDCLK as timeout clock */
+	host->quirks |= SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK;
+
+	/*
+	 * If controller does not have internal clock divider,
+	 * we can use overriding functions instead of default.
+	 */
+	if (sc->no_divider) {
+		sdhci_s3c_ops.set_clock = sdhci_cmu_set_clock;
+		sdhci_s3c_ops.get_min_clock = sdhci_cmu_get_min_clock;
+		sdhci_s3c_ops.get_max_clock = sdhci_cmu_get_max_clock;
+	}
+
+	/* It supports additional host capabilities if needed */
+	if (pdata->host_caps)
+		host->mmc->caps |= pdata->host_caps;
+
+	if (pdata->host_caps2)
+		host->mmc->caps2 |= pdata->host_caps2;
+
+	pm_runtime_enable(&pdev->dev);
+	pm_runtime_set_autosuspend_delay(&pdev->dev, 50);
+	pm_runtime_use_autosuspend(&pdev->dev);
+	pm_suspend_ignore_children(&pdev->dev, 1);
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto err_req_regs;
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err_req_regs;
+
+#ifdef CONFIG_PM
+	if (pdata->cd_type != S3C_SDHCI_CD_INTERNAL)
+		clk_disable_unprepare(sc->clk_io);
+#endif
+	return 0;
+
+ err_req_regs:
+	pm_runtime_disable(&pdev->dev);
+
+ err_no_busclks:
+	clk_disable_unprepare(sc->clk_io);
+
+ err_pdata_io_clk:
+	sdhci_free_host(host);
+
+	return ret;
+}
+
+static int sdhci_s3c_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host =  platform_get_drvdata(pdev);
+	struct sdhci_s3c *sc = sdhci_priv(host);
+
+	if (sc->ext_cd_irq)
+		free_irq(sc->ext_cd_irq, sc);
+
+#ifdef CONFIG_PM
+	if (sc->pdata->cd_type != S3C_SDHCI_CD_INTERNAL)
+		clk_prepare_enable(sc->clk_io);
+#endif
+	sdhci_remove_host(host, 1);
+
+	pm_runtime_dont_use_autosuspend(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+
+	clk_disable_unprepare(sc->clk_io);
+
+	sdhci_free_host(host);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int sdhci_s3c_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	return sdhci_suspend_host(host);
+}
+
+static int sdhci_s3c_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+
+	return sdhci_resume_host(host);
+}
+#endif
+
+#ifdef CONFIG_PM
+static int sdhci_s3c_runtime_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_s3c *ourhost = to_s3c(host);
+	struct clk *busclk = ourhost->clk_io;
+	int ret;
+
+	ret = sdhci_runtime_suspend_host(host);
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	if (ourhost->cur_clk >= 0)
+		clk_disable_unprepare(ourhost->clk_bus[ourhost->cur_clk]);
+	clk_disable_unprepare(busclk);
+	return ret;
+}
+
+static int sdhci_s3c_runtime_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_s3c *ourhost = to_s3c(host);
+	struct clk *busclk = ourhost->clk_io;
+	int ret;
+
+	clk_prepare_enable(busclk);
+	if (ourhost->cur_clk >= 0)
+		clk_prepare_enable(ourhost->clk_bus[ourhost->cur_clk]);
+	ret = sdhci_runtime_resume_host(host, 0);
+	return ret;
+}
+#endif
+
+static const struct dev_pm_ops sdhci_s3c_pmops = {
+	SET_SYSTEM_SLEEP_PM_OPS(sdhci_s3c_suspend, sdhci_s3c_resume)
+	SET_RUNTIME_PM_OPS(sdhci_s3c_runtime_suspend, sdhci_s3c_runtime_resume,
+			   NULL)
+};
+
+static const struct platform_device_id sdhci_s3c_driver_ids[] = {
+	{
+		.name		= "s3c-sdhci",
+		.driver_data	= (kernel_ulong_t)NULL,
+	},
+	{ }
+};
+MODULE_DEVICE_TABLE(platform, sdhci_s3c_driver_ids);
+
+#ifdef CONFIG_OF
+static struct sdhci_s3c_drv_data exynos4_sdhci_drv_data = {
+	.no_divider = true,
+};
+
+static const struct of_device_id sdhci_s3c_dt_match[] = {
+	{ .compatible = "samsung,s3c6410-sdhci", },
+	{ .compatible = "samsung,exynos4210-sdhci",
+		.data = &exynos4_sdhci_drv_data },
+	{},
+};
+MODULE_DEVICE_TABLE(of, sdhci_s3c_dt_match);
+#endif
+
+static struct platform_driver sdhci_s3c_driver = {
+	.probe		= sdhci_s3c_probe,
+	.remove		= sdhci_s3c_remove,
+	.id_table	= sdhci_s3c_driver_ids,
+	.driver		= {
+		.name	= "s3c-sdhci",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(sdhci_s3c_dt_match),
+		.pm	= &sdhci_s3c_pmops,
+	},
+};
+
+module_platform_driver(sdhci_s3c_driver);
+
+MODULE_DESCRIPTION("Samsung SDHCI (HSMMC) glue");
+MODULE_AUTHOR("Ben Dooks, <ben@simtec.co.uk>");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:s3c-sdhci");
diff --git a/drivers/mmc/host/sdhci-sirf.c b/drivers/mmc/host/sdhci-sirf.c
new file mode 100644
index 000000000..e9b347b3a
--- /dev/null
+++ b/drivers/mmc/host/sdhci-sirf.c
@@ -0,0 +1,235 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * SDHCI support for SiRF primaII and marco SoCs
+ *
+ * Copyright (c) 2011 Cambridge Silicon Radio Limited, a CSR plc group company.
+ */
+
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/mmc/host.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/mmc/slot-gpio.h>
+#include "sdhci-pltfm.h"
+
+#define SDHCI_CLK_DELAY_SETTING 0x4C
+#define SDHCI_SIRF_8BITBUS BIT(3)
+#define SIRF_TUNING_COUNT 16384
+
+static void sdhci_sirf_set_bus_width(struct sdhci_host *host, int width)
+{
+	u8 ctrl;
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+	ctrl &= ~(SDHCI_CTRL_4BITBUS | SDHCI_SIRF_8BITBUS);
+
+	/*
+	 * CSR atlas7 and prima2 SD host version is not 3.0
+	 * 8bit-width enable bit of CSR SD hosts is 3,
+	 * while stardard hosts use bit 5
+	 */
+	if (width == MMC_BUS_WIDTH_8)
+		ctrl |= SDHCI_SIRF_8BITBUS;
+	else if (width == MMC_BUS_WIDTH_4)
+		ctrl |= SDHCI_CTRL_4BITBUS;
+
+	sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+}
+
+static u32 sdhci_sirf_readl_le(struct sdhci_host *host, int reg)
+{
+	u32 val = readl(host->ioaddr + reg);
+
+	if (unlikely((reg == SDHCI_CAPABILITIES_1) &&
+			(host->mmc->caps & MMC_CAP_UHS_SDR50))) {
+		/* fake CAP_1 register */
+		val = SDHCI_SUPPORT_DDR50 |
+			SDHCI_SUPPORT_SDR50 | SDHCI_USE_SDR50_TUNING;
+	}
+
+	if (unlikely(reg == SDHCI_SLOT_INT_STATUS)) {
+		u32 prss = val;
+		/* fake chips as V3.0 host conreoller */
+		prss &= ~(0xFF << 16);
+		val = prss | (SDHCI_SPEC_300 << 16);
+	}
+	return val;
+}
+
+static u16 sdhci_sirf_readw_le(struct sdhci_host *host, int reg)
+{
+	u16 ret = 0;
+
+	ret = readw(host->ioaddr + reg);
+
+	if (unlikely(reg == SDHCI_HOST_VERSION)) {
+		ret = readw(host->ioaddr + SDHCI_HOST_VERSION);
+		ret |= SDHCI_SPEC_300;
+	}
+
+	return ret;
+}
+
+static int sdhci_sirf_execute_tuning(struct sdhci_host *host, u32 opcode)
+{
+	int tuning_seq_cnt = 3;
+	int phase;
+	u8 tuned_phase_cnt = 0;
+	int rc = 0, longest_range = 0;
+	int start = -1, end = 0, tuning_value = -1, range = 0;
+	u16 clock_setting;
+	struct mmc_host *mmc = host->mmc;
+
+	clock_setting = sdhci_readw(host, SDHCI_CLK_DELAY_SETTING);
+	clock_setting &= ~0x3fff;
+
+retry:
+	phase = 0;
+	tuned_phase_cnt = 0;
+	do {
+		sdhci_writel(host,
+			clock_setting | phase,
+			SDHCI_CLK_DELAY_SETTING);
+
+		if (!mmc_send_tuning(mmc, opcode, NULL)) {
+			/* Tuning is successful at this tuning point */
+			tuned_phase_cnt++;
+			dev_dbg(mmc_dev(mmc), "%s: Found good phase = %d\n",
+				 mmc_hostname(mmc), phase);
+			if (start == -1)
+				start = phase;
+			end = phase;
+			range++;
+			if (phase == (SIRF_TUNING_COUNT - 1)
+				&& range > longest_range)
+				tuning_value = (start + end) / 2;
+		} else {
+			dev_dbg(mmc_dev(mmc), "%s: Found bad phase = %d\n",
+				 mmc_hostname(mmc), phase);
+			if (range > longest_range) {
+				tuning_value = (start + end) / 2;
+				longest_range = range;
+			}
+			start = -1;
+			end = range = 0;
+		}
+	} while (++phase < SIRF_TUNING_COUNT);
+
+	if (tuned_phase_cnt && tuning_value > 0) {
+		/*
+		 * Finally set the selected phase in delay
+		 * line hw block.
+		 */
+		phase = tuning_value;
+		sdhci_writel(host,
+			clock_setting | phase,
+			SDHCI_CLK_DELAY_SETTING);
+
+		dev_dbg(mmc_dev(mmc), "%s: Setting the tuning phase to %d\n",
+			 mmc_hostname(mmc), phase);
+	} else {
+		if (--tuning_seq_cnt)
+			goto retry;
+		/* Tuning failed */
+		dev_dbg(mmc_dev(mmc), "%s: No tuning point found\n",
+		       mmc_hostname(mmc));
+		rc = -EIO;
+	}
+
+	return rc;
+}
+
+static const struct sdhci_ops sdhci_sirf_ops = {
+	.read_l = sdhci_sirf_readl_le,
+	.read_w = sdhci_sirf_readw_le,
+	.platform_execute_tuning = sdhci_sirf_execute_tuning,
+	.set_clock = sdhci_set_clock,
+	.get_max_clock	= sdhci_pltfm_clk_get_max_clock,
+	.set_bus_width = sdhci_sirf_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static const struct sdhci_pltfm_data sdhci_sirf_pdata = {
+	.ops = &sdhci_sirf_ops,
+	.quirks = SDHCI_QUIRK_BROKEN_TIMEOUT_VAL |
+		SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK |
+		SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN |
+		SDHCI_QUIRK_RESET_CMD_DATA_ON_IOS,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+};
+
+static int sdhci_sirf_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct clk *clk;
+	int ret;
+
+	clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(clk)) {
+		dev_err(&pdev->dev, "unable to get clock");
+		return PTR_ERR(clk);
+	}
+
+	host = sdhci_pltfm_init(pdev, &sdhci_sirf_pdata, 0);
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	pltfm_host->clk = clk;
+
+	sdhci_get_of_property(pdev);
+
+	ret = clk_prepare_enable(pltfm_host->clk);
+	if (ret)
+		goto err_clk_prepare;
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err_sdhci_add;
+
+	/*
+	 * We must request the IRQ after sdhci_add_host(), as the tasklet only
+	 * gets setup in sdhci_add_host() and we oops.
+	 */
+	ret = mmc_gpiod_request_cd(host->mmc, "cd", 0, false, 0);
+	if (ret == -EPROBE_DEFER)
+		goto err_request_cd;
+	if (!ret)
+		mmc_gpiod_request_cd_irq(host->mmc);
+
+	return 0;
+
+err_request_cd:
+	sdhci_remove_host(host, 0);
+err_sdhci_add:
+	clk_disable_unprepare(pltfm_host->clk);
+err_clk_prepare:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static const struct of_device_id sdhci_sirf_of_match[] = {
+	{ .compatible = "sirf,prima2-sdhc" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, sdhci_sirf_of_match);
+
+static struct platform_driver sdhci_sirf_driver = {
+	.driver		= {
+		.name	= "sdhci-sirf",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = sdhci_sirf_of_match,
+		.pm	= &sdhci_pltfm_pmops,
+	},
+	.probe		= sdhci_sirf_probe,
+	.remove		= sdhci_pltfm_unregister,
+};
+
+module_platform_driver(sdhci_sirf_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for SiRFprimaII/SiRFmarco");
+MODULE_AUTHOR("Barry Song <21cnbao@gmail.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-spear.c b/drivers/mmc/host/sdhci-spear.c
new file mode 100644
index 000000000..d463e2fd5
--- /dev/null
+++ b/drivers/mmc/host/sdhci-spear.c
@@ -0,0 +1,196 @@
+/*
+ * drivers/mmc/host/sdhci-spear.c
+ *
+ * Support of SDHCI platform devices for spear soc family
+ *
+ * Copyright (C) 2010 ST Microelectronics
+ * Viresh Kumar <vireshk@kernel.org>
+ *
+ * Inspired by sdhci-pltfm.c
+ *
+ * This file is licensed under the terms of the GNU General Public
+ * License version 2. This program is licensed "as is" without any
+ * warranty of any kind, whether express or implied.
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/highmem.h>
+#include <linux/module.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm.h>
+#include <linux/slab.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/io.h>
+#include "sdhci.h"
+
+struct spear_sdhci {
+	struct clk *clk;
+};
+
+/* sdhci ops */
+static const struct sdhci_ops sdhci_pltfm_ops = {
+	.set_clock = sdhci_set_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static int sdhci_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct spear_sdhci *sdhci;
+	struct device *dev;
+	int ret;
+
+	dev = pdev->dev.parent ? pdev->dev.parent : &pdev->dev;
+	host = sdhci_alloc_host(dev, sizeof(*sdhci));
+	if (IS_ERR(host)) {
+		ret = PTR_ERR(host);
+		dev_dbg(&pdev->dev, "cannot allocate memory for sdhci\n");
+		goto err;
+	}
+
+	host->ioaddr = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(host->ioaddr)) {
+		ret = PTR_ERR(host->ioaddr);
+		dev_dbg(&pdev->dev, "unable to map iomem: %d\n", ret);
+		goto err_host;
+	}
+
+	host->hw_name = "sdhci";
+	host->ops = &sdhci_pltfm_ops;
+	host->irq = platform_get_irq(pdev, 0);
+	if (host->irq <= 0) {
+		ret = -EINVAL;
+		goto err_host;
+	}
+	host->quirks = SDHCI_QUIRK_BROKEN_ADMA;
+
+	sdhci = sdhci_priv(host);
+
+	/* clk enable */
+	sdhci->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(sdhci->clk)) {
+		ret = PTR_ERR(sdhci->clk);
+		dev_dbg(&pdev->dev, "Error getting clock\n");
+		goto err_host;
+	}
+
+	ret = clk_prepare_enable(sdhci->clk);
+	if (ret) {
+		dev_dbg(&pdev->dev, "Error enabling clock\n");
+		goto err_host;
+	}
+
+	ret = clk_set_rate(sdhci->clk, 50000000);
+	if (ret)
+		dev_dbg(&pdev->dev, "Error setting desired clk, clk=%lu\n",
+				clk_get_rate(sdhci->clk));
+
+	/*
+	 * It is optional to use GPIOs for sdhci card detection. If we
+	 * find a descriptor using slot GPIO, we use it.
+	 */
+	ret = mmc_gpiod_request_cd(host->mmc, "cd", 0, false, 0);
+	if (ret == -EPROBE_DEFER)
+		goto disable_clk;
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto disable_clk;
+
+	platform_set_drvdata(pdev, host);
+
+	return 0;
+
+disable_clk:
+	clk_disable_unprepare(sdhci->clk);
+err_host:
+	sdhci_free_host(host);
+err:
+	dev_err(&pdev->dev, "spear-sdhci probe failed: %d\n", ret);
+	return ret;
+}
+
+static int sdhci_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct spear_sdhci *sdhci = sdhci_priv(host);
+	int dead = 0;
+	u32 scratch;
+
+	scratch = readl(host->ioaddr + SDHCI_INT_STATUS);
+	if (scratch == (u32)-1)
+		dead = 1;
+
+	sdhci_remove_host(host, dead);
+	clk_disable_unprepare(sdhci->clk);
+	sdhci_free_host(host);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int sdhci_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct spear_sdhci *sdhci = sdhci_priv(host);
+	int ret;
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	ret = sdhci_suspend_host(host);
+	if (!ret)
+		clk_disable(sdhci->clk);
+
+	return ret;
+}
+
+static int sdhci_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct spear_sdhci *sdhci = sdhci_priv(host);
+	int ret;
+
+	ret = clk_enable(sdhci->clk);
+	if (ret) {
+		dev_dbg(dev, "Resume: Error enabling clock\n");
+		return ret;
+	}
+
+	return sdhci_resume_host(host);
+}
+#endif
+
+static SIMPLE_DEV_PM_OPS(sdhci_pm_ops, sdhci_suspend, sdhci_resume);
+
+#ifdef CONFIG_OF
+static const struct of_device_id sdhci_spear_id_table[] = {
+	{ .compatible = "st,spear300-sdhci" },
+	{}
+};
+MODULE_DEVICE_TABLE(of, sdhci_spear_id_table);
+#endif
+
+static struct platform_driver sdhci_driver = {
+	.driver = {
+		.name	= "sdhci",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= &sdhci_pm_ops,
+		.of_match_table = of_match_ptr(sdhci_spear_id_table),
+	},
+	.probe		= sdhci_probe,
+	.remove		= sdhci_remove,
+};
+
+module_platform_driver(sdhci_driver);
+
+MODULE_DESCRIPTION("SPEAr Secure Digital Host Controller Interface driver");
+MODULE_AUTHOR("Viresh Kumar <vireshk@kernel.org>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-sprd.c b/drivers/mmc/host/sdhci-sprd.c
new file mode 100644
index 000000000..52bfe356f
--- /dev/null
+++ b/drivers/mmc/host/sdhci-sprd.c
@@ -0,0 +1,833 @@
+// SPDX-License-Identifier: GPL-2.0
+//
+// Secure Digital Host Controller
+//
+// Copyright (C) 2018 Spreadtrum, Inc.
+// Author: Chunyan Zhang <chunyan.zhang@unisoc.com>
+
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/highmem.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/of_gpio.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/regulator/consumer.h>
+#include <linux/slab.h>
+
+#include "sdhci-pltfm.h"
+#include "mmc_hsq.h"
+
+/* SDHCI_ARGUMENT2 register high 16bit */
+#define SDHCI_SPRD_ARG2_STUFF		GENMASK(31, 16)
+
+#define SDHCI_SPRD_REG_32_DLL_CFG	0x200
+#define  SDHCI_SPRD_DLL_ALL_CPST_EN	(BIT(18) | BIT(24) | BIT(25) | BIT(26) | BIT(27))
+#define  SDHCI_SPRD_DLL_EN		BIT(21)
+#define  SDHCI_SPRD_DLL_SEARCH_MODE	BIT(16)
+#define  SDHCI_SPRD_DLL_INIT_COUNT	0xc00
+#define  SDHCI_SPRD_DLL_PHASE_INTERNAL	0x3
+
+#define SDHCI_SPRD_REG_32_DLL_DLY	0x204
+
+#define SDHCI_SPRD_REG_32_DLL_DLY_OFFSET	0x208
+#define  SDHCIBSPRD_IT_WR_DLY_INV		BIT(5)
+#define  SDHCI_SPRD_BIT_CMD_DLY_INV		BIT(13)
+#define  SDHCI_SPRD_BIT_POSRD_DLY_INV		BIT(21)
+#define  SDHCI_SPRD_BIT_NEGRD_DLY_INV		BIT(29)
+
+#define SDHCI_SPRD_REG_32_BUSY_POSI		0x250
+#define  SDHCI_SPRD_BIT_OUTR_CLK_AUTO_EN	BIT(25)
+#define  SDHCI_SPRD_BIT_INNR_CLK_AUTO_EN	BIT(24)
+
+#define SDHCI_SPRD_REG_DEBOUNCE		0x28C
+#define  SDHCI_SPRD_BIT_DLL_BAK		BIT(0)
+#define  SDHCI_SPRD_BIT_DLL_VAL		BIT(1)
+
+#define  SDHCI_SPRD_INT_SIGNAL_MASK	0x1B7F410B
+
+/* SDHCI_HOST_CONTROL2 */
+#define  SDHCI_SPRD_CTRL_HS200		0x0005
+#define  SDHCI_SPRD_CTRL_HS400		0x0006
+#define  SDHCI_SPRD_CTRL_HS400ES	0x0007
+
+/*
+ * According to the standard specification, BIT(3) of SDHCI_SOFTWARE_RESET is
+ * reserved, and only used on Spreadtrum's design, the hardware cannot work
+ * if this bit is cleared.
+ * 1 : normal work
+ * 0 : hardware reset
+ */
+#define  SDHCI_HW_RESET_CARD		BIT(3)
+
+#define SDHCI_SPRD_MAX_CUR		0xFFFFFF
+#define SDHCI_SPRD_CLK_MAX_DIV		1023
+
+#define SDHCI_SPRD_CLK_DEF_RATE		26000000
+#define SDHCI_SPRD_PHY_DLL_CLK		52000000
+
+struct sdhci_sprd_host {
+	u32 version;
+	struct clk *clk_sdio;
+	struct clk *clk_enable;
+	struct clk *clk_2x_enable;
+	struct pinctrl *pinctrl;
+	struct pinctrl_state *pins_uhs;
+	struct pinctrl_state *pins_default;
+	u32 base_rate;
+	int flags; /* backup of host attribute */
+	u32 phy_delay[MMC_TIMING_MMC_HS400 + 2];
+};
+
+struct sdhci_sprd_phy_cfg {
+	const char *property;
+	u8 timing;
+};
+
+static const struct sdhci_sprd_phy_cfg sdhci_sprd_phy_cfgs[] = {
+	{ "sprd,phy-delay-legacy", MMC_TIMING_LEGACY, },
+	{ "sprd,phy-delay-sd-highspeed", MMC_TIMING_SD_HS, },
+	{ "sprd,phy-delay-sd-uhs-sdr50", MMC_TIMING_UHS_SDR50, },
+	{ "sprd,phy-delay-sd-uhs-sdr104", MMC_TIMING_UHS_SDR104, },
+	{ "sprd,phy-delay-mmc-highspeed", MMC_TIMING_MMC_HS, },
+	{ "sprd,phy-delay-mmc-ddr52", MMC_TIMING_MMC_DDR52, },
+	{ "sprd,phy-delay-mmc-hs200", MMC_TIMING_MMC_HS200, },
+	{ "sprd,phy-delay-mmc-hs400", MMC_TIMING_MMC_HS400, },
+	{ "sprd,phy-delay-mmc-hs400es", MMC_TIMING_MMC_HS400 + 1, },
+};
+
+#define TO_SPRD_HOST(host) sdhci_pltfm_priv(sdhci_priv(host))
+
+static void sdhci_sprd_init_config(struct sdhci_host *host)
+{
+	u16 val;
+
+	/* set dll backup mode */
+	val = sdhci_readl(host, SDHCI_SPRD_REG_DEBOUNCE);
+	val |= SDHCI_SPRD_BIT_DLL_BAK | SDHCI_SPRD_BIT_DLL_VAL;
+	sdhci_writel(host, val, SDHCI_SPRD_REG_DEBOUNCE);
+}
+
+static inline u32 sdhci_sprd_readl(struct sdhci_host *host, int reg)
+{
+	if (unlikely(reg == SDHCI_MAX_CURRENT))
+		return SDHCI_SPRD_MAX_CUR;
+
+	return readl_relaxed(host->ioaddr + reg);
+}
+
+static inline void sdhci_sprd_writel(struct sdhci_host *host, u32 val, int reg)
+{
+	/* SDHCI_MAX_CURRENT is reserved on Spreadtrum's platform */
+	if (unlikely(reg == SDHCI_MAX_CURRENT))
+		return;
+
+	if (unlikely(reg == SDHCI_SIGNAL_ENABLE || reg == SDHCI_INT_ENABLE))
+		val = val & SDHCI_SPRD_INT_SIGNAL_MASK;
+
+	writel_relaxed(val, host->ioaddr + reg);
+}
+
+static inline void sdhci_sprd_writew(struct sdhci_host *host, u16 val, int reg)
+{
+	/* SDHCI_BLOCK_COUNT is Read Only on Spreadtrum's platform */
+	if (unlikely(reg == SDHCI_BLOCK_COUNT))
+		return;
+
+	writew_relaxed(val, host->ioaddr + reg);
+}
+
+static inline void sdhci_sprd_writeb(struct sdhci_host *host, u8 val, int reg)
+{
+	/*
+	 * Since BIT(3) of SDHCI_SOFTWARE_RESET is reserved according to the
+	 * standard specification, sdhci_reset() write this register directly
+	 * without checking other reserved bits, that will clear BIT(3) which
+	 * is defined as hardware reset on Spreadtrum's platform and clearing
+	 * it by mistake will lead the card not work. So here we need to work
+	 * around it.
+	 */
+	if (unlikely(reg == SDHCI_SOFTWARE_RESET)) {
+		if (readb_relaxed(host->ioaddr + reg) & SDHCI_HW_RESET_CARD)
+			val |= SDHCI_HW_RESET_CARD;
+	}
+
+	writeb_relaxed(val, host->ioaddr + reg);
+}
+
+static inline void sdhci_sprd_sd_clk_off(struct sdhci_host *host)
+{
+	u16 ctrl = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+
+	ctrl &= ~SDHCI_CLOCK_CARD_EN;
+	sdhci_writew(host, ctrl, SDHCI_CLOCK_CONTROL);
+}
+
+static inline void sdhci_sprd_sd_clk_on(struct sdhci_host *host)
+{
+	u16 ctrl;
+
+	ctrl = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+	ctrl |= SDHCI_CLOCK_CARD_EN;
+	sdhci_writew(host, ctrl, SDHCI_CLOCK_CONTROL);
+}
+
+static inline void
+sdhci_sprd_set_dll_invert(struct sdhci_host *host, u32 mask, bool en)
+{
+	u32 dll_dly_offset;
+
+	dll_dly_offset = sdhci_readl(host, SDHCI_SPRD_REG_32_DLL_DLY_OFFSET);
+	if (en)
+		dll_dly_offset |= mask;
+	else
+		dll_dly_offset &= ~mask;
+	sdhci_writel(host, dll_dly_offset, SDHCI_SPRD_REG_32_DLL_DLY_OFFSET);
+}
+
+static inline u32 sdhci_sprd_calc_div(u32 base_clk, u32 clk)
+{
+	u32 div;
+
+	/* select 2x clock source */
+	if (base_clk <= clk * 2)
+		return 0;
+
+	div = (u32) (base_clk / (clk * 2));
+
+	if ((base_clk / div) > (clk * 2))
+		div++;
+
+	if (div > SDHCI_SPRD_CLK_MAX_DIV)
+		div = SDHCI_SPRD_CLK_MAX_DIV;
+
+	if (div % 2)
+		div = (div + 1) / 2;
+	else
+		div = div / 2;
+
+	return div;
+}
+
+static inline void _sdhci_sprd_set_clock(struct sdhci_host *host,
+					unsigned int clk)
+{
+	struct sdhci_sprd_host *sprd_host = TO_SPRD_HOST(host);
+	u32 div, val, mask;
+
+	sdhci_writew(host, 0, SDHCI_CLOCK_CONTROL);
+
+	div = sdhci_sprd_calc_div(sprd_host->base_rate, clk);
+	div = ((div & 0x300) >> 2) | ((div & 0xFF) << 8);
+	sdhci_enable_clk(host, div);
+
+	val = sdhci_readl(host, SDHCI_SPRD_REG_32_BUSY_POSI);
+	mask = SDHCI_SPRD_BIT_OUTR_CLK_AUTO_EN | SDHCI_SPRD_BIT_INNR_CLK_AUTO_EN;
+	/* Enable CLK_AUTO when the clock is greater than 400K. */
+	if (clk > 400000) {
+		if (mask != (val & mask)) {
+			val |= mask;
+			sdhci_writel(host, val, SDHCI_SPRD_REG_32_BUSY_POSI);
+		}
+	} else {
+		if (val & mask) {
+			val &= ~mask;
+			sdhci_writel(host, val, SDHCI_SPRD_REG_32_BUSY_POSI);
+		}
+	}
+}
+
+static void sdhci_sprd_enable_phy_dll(struct sdhci_host *host)
+{
+	u32 tmp;
+
+	tmp = sdhci_readl(host, SDHCI_SPRD_REG_32_DLL_CFG);
+	tmp &= ~(SDHCI_SPRD_DLL_EN | SDHCI_SPRD_DLL_ALL_CPST_EN);
+	sdhci_writel(host, tmp, SDHCI_SPRD_REG_32_DLL_CFG);
+	/* wait 1ms */
+	usleep_range(1000, 1250);
+
+	tmp = sdhci_readl(host, SDHCI_SPRD_REG_32_DLL_CFG);
+	tmp |= SDHCI_SPRD_DLL_ALL_CPST_EN | SDHCI_SPRD_DLL_SEARCH_MODE |
+		SDHCI_SPRD_DLL_INIT_COUNT | SDHCI_SPRD_DLL_PHASE_INTERNAL;
+	sdhci_writel(host, tmp, SDHCI_SPRD_REG_32_DLL_CFG);
+	/* wait 1ms */
+	usleep_range(1000, 1250);
+
+	tmp = sdhci_readl(host, SDHCI_SPRD_REG_32_DLL_CFG);
+	tmp |= SDHCI_SPRD_DLL_EN;
+	sdhci_writel(host, tmp, SDHCI_SPRD_REG_32_DLL_CFG);
+	/* wait 1ms */
+	usleep_range(1000, 1250);
+}
+
+static void sdhci_sprd_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	bool en = false, clk_changed = false;
+
+	if (clock == 0) {
+		sdhci_writew(host, 0, SDHCI_CLOCK_CONTROL);
+	} else if (clock != host->clock) {
+		sdhci_sprd_sd_clk_off(host);
+		_sdhci_sprd_set_clock(host, clock);
+
+		if (clock <= 400000)
+			en = true;
+		sdhci_sprd_set_dll_invert(host, SDHCI_SPRD_BIT_CMD_DLY_INV |
+					  SDHCI_SPRD_BIT_POSRD_DLY_INV, en);
+		clk_changed = true;
+	} else {
+		_sdhci_sprd_set_clock(host, clock);
+	}
+
+	/*
+	 * According to the Spreadtrum SD host specification, when we changed
+	 * the clock to be more than 52M, we should enable the PHY DLL which
+	 * is used to track the clock frequency to make the clock work more
+	 * stable. Otherwise deviation may occur of the higher clock.
+	 */
+	if (clk_changed && clock > SDHCI_SPRD_PHY_DLL_CLK)
+		sdhci_sprd_enable_phy_dll(host);
+}
+
+static unsigned int sdhci_sprd_get_max_clock(struct sdhci_host *host)
+{
+	struct sdhci_sprd_host *sprd_host = TO_SPRD_HOST(host);
+
+	return clk_round_rate(sprd_host->clk_sdio, ULONG_MAX);
+}
+
+static unsigned int sdhci_sprd_get_min_clock(struct sdhci_host *host)
+{
+	return 100000;
+}
+
+static void sdhci_sprd_set_uhs_signaling(struct sdhci_host *host,
+					 unsigned int timing)
+{
+	struct sdhci_sprd_host *sprd_host = TO_SPRD_HOST(host);
+	struct mmc_host *mmc = host->mmc;
+	u32 *p = sprd_host->phy_delay;
+	u16 ctrl_2;
+
+	if (timing == host->timing)
+		return;
+
+	ctrl_2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	/* Select Bus Speed Mode for host */
+	ctrl_2 &= ~SDHCI_CTRL_UHS_MASK;
+	switch (timing) {
+	case MMC_TIMING_UHS_SDR12:
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR12;
+		break;
+	case MMC_TIMING_MMC_HS:
+	case MMC_TIMING_SD_HS:
+	case MMC_TIMING_UHS_SDR25:
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR25;
+		break;
+	case MMC_TIMING_UHS_SDR50:
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR50;
+		break;
+	case MMC_TIMING_UHS_SDR104:
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR104;
+		break;
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_MMC_DDR52:
+		ctrl_2 |= SDHCI_CTRL_UHS_DDR50;
+		break;
+	case MMC_TIMING_MMC_HS200:
+		ctrl_2 |= SDHCI_SPRD_CTRL_HS200;
+		break;
+	case MMC_TIMING_MMC_HS400:
+		ctrl_2 |= SDHCI_SPRD_CTRL_HS400;
+		break;
+	default:
+		break;
+	}
+
+	sdhci_writew(host, ctrl_2, SDHCI_HOST_CONTROL2);
+
+	if (!mmc->ios.enhanced_strobe)
+		sdhci_writel(host, p[timing], SDHCI_SPRD_REG_32_DLL_DLY);
+}
+
+static void sdhci_sprd_hw_reset(struct sdhci_host *host)
+{
+	int val;
+
+	/*
+	 * Note: don't use sdhci_writeb() API here since it is redirected to
+	 * sdhci_sprd_writeb() in which we have a workaround for
+	 * SDHCI_SOFTWARE_RESET which would make bit SDHCI_HW_RESET_CARD can
+	 * not be cleared.
+	 */
+	val = readb_relaxed(host->ioaddr + SDHCI_SOFTWARE_RESET);
+	val &= ~SDHCI_HW_RESET_CARD;
+	writeb_relaxed(val, host->ioaddr + SDHCI_SOFTWARE_RESET);
+	/* wait for 10 us */
+	usleep_range(10, 20);
+
+	val |= SDHCI_HW_RESET_CARD;
+	writeb_relaxed(val, host->ioaddr + SDHCI_SOFTWARE_RESET);
+	usleep_range(300, 500);
+}
+
+static unsigned int sdhci_sprd_get_max_timeout_count(struct sdhci_host *host)
+{
+	/* The Spredtrum controller actual maximum timeout count is 1 << 31 */
+	return 1 << 31;
+}
+
+static unsigned int sdhci_sprd_get_ro(struct sdhci_host *host)
+{
+	return 0;
+}
+
+static void sdhci_sprd_request_done(struct sdhci_host *host,
+				    struct mmc_request *mrq)
+{
+	/* Validate if the request was from software queue firstly. */
+	if (mmc_hsq_finalize_request(host->mmc, mrq))
+		return;
+
+	mmc_request_done(host->mmc, mrq);
+}
+
+static void sdhci_sprd_set_power(struct sdhci_host *host, unsigned char mode,
+				 unsigned short vdd)
+{
+	struct mmc_host *mmc = host->mmc;
+
+	switch (mode) {
+	case MMC_POWER_OFF:
+		mmc_regulator_set_ocr(host->mmc, mmc->supply.vmmc, 0);
+
+		mmc_regulator_disable_vqmmc(mmc);
+		break;
+	case MMC_POWER_ON:
+		mmc_regulator_enable_vqmmc(mmc);
+		break;
+	case MMC_POWER_UP:
+		mmc_regulator_set_ocr(host->mmc, mmc->supply.vmmc, vdd);
+		break;
+	}
+}
+
+static struct sdhci_ops sdhci_sprd_ops = {
+	.read_l = sdhci_sprd_readl,
+	.write_l = sdhci_sprd_writel,
+	.write_w = sdhci_sprd_writew,
+	.write_b = sdhci_sprd_writeb,
+	.set_clock = sdhci_sprd_set_clock,
+	.set_power = sdhci_sprd_set_power,
+	.get_max_clock = sdhci_sprd_get_max_clock,
+	.get_min_clock = sdhci_sprd_get_min_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_sprd_set_uhs_signaling,
+	.hw_reset = sdhci_sprd_hw_reset,
+	.get_max_timeout_count = sdhci_sprd_get_max_timeout_count,
+	.get_ro = sdhci_sprd_get_ro,
+	.request_done = sdhci_sprd_request_done,
+};
+
+static void sdhci_sprd_check_auto_cmd23(struct mmc_host *mmc,
+					struct mmc_request *mrq)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_sprd_host *sprd_host = TO_SPRD_HOST(host);
+
+	host->flags |= sprd_host->flags & SDHCI_AUTO_CMD23;
+
+	/*
+	 * From version 4.10 onward, ARGUMENT2 register is also as 32-bit
+	 * block count register which doesn't support stuff bits of
+	 * CMD23 argument on Spreadtrum's sd host controller.
+	 */
+	if (host->version >= SDHCI_SPEC_410 &&
+	    mrq->sbc && (mrq->sbc->arg & SDHCI_SPRD_ARG2_STUFF) &&
+	    (host->flags & SDHCI_AUTO_CMD23))
+		host->flags &= ~SDHCI_AUTO_CMD23;
+}
+
+static void sdhci_sprd_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	sdhci_sprd_check_auto_cmd23(mmc, mrq);
+
+	sdhci_request(mmc, mrq);
+}
+
+static int sdhci_sprd_request_atomic(struct mmc_host *mmc,
+				     struct mmc_request *mrq)
+{
+	sdhci_sprd_check_auto_cmd23(mmc, mrq);
+
+	return sdhci_request_atomic(mmc, mrq);
+}
+
+static int sdhci_sprd_voltage_switch(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_sprd_host *sprd_host = TO_SPRD_HOST(host);
+	int ret;
+
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+		ret = mmc_regulator_set_vqmmc(mmc, ios);
+		if (ret < 0) {
+			pr_err("%s: Switching signalling voltage failed\n",
+			       mmc_hostname(mmc));
+			return ret;
+		}
+	}
+
+	if (IS_ERR(sprd_host->pinctrl))
+		goto reset;
+
+	switch (ios->signal_voltage) {
+	case MMC_SIGNAL_VOLTAGE_180:
+		ret = pinctrl_select_state(sprd_host->pinctrl,
+					   sprd_host->pins_uhs);
+		if (ret) {
+			pr_err("%s: failed to select uhs pin state\n",
+			       mmc_hostname(mmc));
+			return ret;
+		}
+		break;
+
+	default:
+		fallthrough;
+	case MMC_SIGNAL_VOLTAGE_330:
+		ret = pinctrl_select_state(sprd_host->pinctrl,
+					   sprd_host->pins_default);
+		if (ret) {
+			pr_err("%s: failed to select default pin state\n",
+			       mmc_hostname(mmc));
+			return ret;
+		}
+		break;
+	}
+
+	/* Wait for 300 ~ 500 us for pin state stable */
+	usleep_range(300, 500);
+
+reset:
+	sdhci_reset(host, SDHCI_RESET_CMD | SDHCI_RESET_DATA);
+
+	return 0;
+}
+
+static void sdhci_sprd_hs400_enhanced_strobe(struct mmc_host *mmc,
+					     struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_sprd_host *sprd_host = TO_SPRD_HOST(host);
+	u32 *p = sprd_host->phy_delay;
+	u16 ctrl_2;
+
+	if (!ios->enhanced_strobe)
+		return;
+
+	sdhci_sprd_sd_clk_off(host);
+
+	/* Set HS400 enhanced strobe mode */
+	ctrl_2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	ctrl_2 &= ~SDHCI_CTRL_UHS_MASK;
+	ctrl_2 |= SDHCI_SPRD_CTRL_HS400ES;
+	sdhci_writew(host, ctrl_2, SDHCI_HOST_CONTROL2);
+
+	sdhci_sprd_sd_clk_on(host);
+
+	/* Set the PHY DLL delay value for HS400 enhanced strobe mode */
+	sdhci_writel(host, p[MMC_TIMING_MMC_HS400 + 1],
+		     SDHCI_SPRD_REG_32_DLL_DLY);
+}
+
+static void sdhci_sprd_phy_param_parse(struct sdhci_sprd_host *sprd_host,
+				       struct device_node *np)
+{
+	u32 *p = sprd_host->phy_delay;
+	int ret, i, index;
+	u32 val[4];
+
+	for (i = 0; i < ARRAY_SIZE(sdhci_sprd_phy_cfgs); i++) {
+		ret = of_property_read_u32_array(np,
+				sdhci_sprd_phy_cfgs[i].property, val, 4);
+		if (ret)
+			continue;
+
+		index = sdhci_sprd_phy_cfgs[i].timing;
+		p[index] = val[0] | (val[1] << 8) | (val[2] << 16) | (val[3] << 24);
+	}
+}
+
+static const struct sdhci_pltfm_data sdhci_sprd_pdata = {
+	.quirks = SDHCI_QUIRK_BROKEN_CARD_DETECTION |
+		  SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK |
+		  SDHCI_QUIRK_MISSING_CAPS,
+	.quirks2 = SDHCI_QUIRK2_BROKEN_HS200 |
+		   SDHCI_QUIRK2_USE_32BIT_BLK_CNT |
+		   SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+	.ops = &sdhci_sprd_ops,
+};
+
+static int sdhci_sprd_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct sdhci_sprd_host *sprd_host;
+	struct mmc_hsq *hsq;
+	struct clk *clk;
+	int ret = 0;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_sprd_pdata, sizeof(*sprd_host));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	host->dma_mask = DMA_BIT_MASK(64);
+	pdev->dev.dma_mask = &host->dma_mask;
+	host->mmc_host_ops.request = sdhci_sprd_request;
+	host->mmc_host_ops.hs400_enhanced_strobe =
+		sdhci_sprd_hs400_enhanced_strobe;
+	/*
+	 * We can not use the standard ops to change and detect the voltage
+	 * signal for Spreadtrum SD host controller, since our voltage regulator
+	 * for I/O is fixed in hardware, that means we do not need control
+	 * the standard SD host controller to change the I/O voltage.
+	 */
+	host->mmc_host_ops.start_signal_voltage_switch =
+		sdhci_sprd_voltage_switch;
+
+	host->mmc->caps = MMC_CAP_SD_HIGHSPEED | MMC_CAP_MMC_HIGHSPEED |
+		MMC_CAP_WAIT_WHILE_BUSY;
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto pltfm_free;
+
+	if (!mmc_card_is_removable(host->mmc))
+		host->mmc_host_ops.request_atomic = sdhci_sprd_request_atomic;
+	else
+		host->always_defer_done = true;
+
+	sprd_host = TO_SPRD_HOST(host);
+	sdhci_sprd_phy_param_parse(sprd_host, pdev->dev.of_node);
+
+	sprd_host->pinctrl = devm_pinctrl_get(&pdev->dev);
+	if (!IS_ERR(sprd_host->pinctrl)) {
+		sprd_host->pins_uhs =
+			pinctrl_lookup_state(sprd_host->pinctrl, "state_uhs");
+		if (IS_ERR(sprd_host->pins_uhs)) {
+			ret = PTR_ERR(sprd_host->pins_uhs);
+			goto pltfm_free;
+		}
+
+		sprd_host->pins_default =
+			pinctrl_lookup_state(sprd_host->pinctrl, "default");
+		if (IS_ERR(sprd_host->pins_default)) {
+			ret = PTR_ERR(sprd_host->pins_default);
+			goto pltfm_free;
+		}
+	}
+
+	clk = devm_clk_get(&pdev->dev, "sdio");
+	if (IS_ERR(clk)) {
+		ret = PTR_ERR(clk);
+		goto pltfm_free;
+	}
+	sprd_host->clk_sdio = clk;
+	sprd_host->base_rate = clk_get_rate(sprd_host->clk_sdio);
+	if (!sprd_host->base_rate)
+		sprd_host->base_rate = SDHCI_SPRD_CLK_DEF_RATE;
+
+	clk = devm_clk_get(&pdev->dev, "enable");
+	if (IS_ERR(clk)) {
+		ret = PTR_ERR(clk);
+		goto pltfm_free;
+	}
+	sprd_host->clk_enable = clk;
+
+	clk = devm_clk_get(&pdev->dev, "2x_enable");
+	if (!IS_ERR(clk))
+		sprd_host->clk_2x_enable = clk;
+
+	ret = clk_prepare_enable(sprd_host->clk_sdio);
+	if (ret)
+		goto pltfm_free;
+
+	ret = clk_prepare_enable(sprd_host->clk_enable);
+	if (ret)
+		goto clk_disable;
+
+	ret = clk_prepare_enable(sprd_host->clk_2x_enable);
+	if (ret)
+		goto clk_disable2;
+
+	sdhci_sprd_init_config(host);
+	host->version = sdhci_readw(host, SDHCI_HOST_VERSION);
+	sprd_host->version = ((host->version & SDHCI_VENDOR_VER_MASK) >>
+			       SDHCI_VENDOR_VER_SHIFT);
+
+	pm_runtime_get_noresume(&pdev->dev);
+	pm_runtime_set_active(&pdev->dev);
+	pm_runtime_enable(&pdev->dev);
+	pm_runtime_set_autosuspend_delay(&pdev->dev, 50);
+	pm_runtime_use_autosuspend(&pdev->dev);
+	pm_suspend_ignore_children(&pdev->dev, 1);
+
+	sdhci_enable_v4_mode(host);
+
+	/*
+	 * Supply the existing CAPS, but clear the UHS-I modes. This
+	 * will allow these modes to be specified only by device
+	 * tree properties through mmc_of_parse().
+	 */
+	host->caps = sdhci_readl(host, SDHCI_CAPABILITIES);
+	host->caps1 = sdhci_readl(host, SDHCI_CAPABILITIES_1);
+	host->caps1 &= ~(SDHCI_SUPPORT_SDR50 | SDHCI_SUPPORT_SDR104 |
+			 SDHCI_SUPPORT_DDR50);
+
+	ret = mmc_regulator_get_supply(host->mmc);
+	if (ret)
+		goto pm_runtime_disable;
+
+	ret = sdhci_setup_host(host);
+	if (ret)
+		goto pm_runtime_disable;
+
+	sprd_host->flags = host->flags;
+
+	hsq = devm_kzalloc(&pdev->dev, sizeof(*hsq), GFP_KERNEL);
+	if (!hsq) {
+		ret = -ENOMEM;
+		goto err_cleanup_host;
+	}
+
+	ret = mmc_hsq_init(hsq, host->mmc);
+	if (ret)
+		goto err_cleanup_host;
+
+	ret = __sdhci_add_host(host);
+	if (ret)
+		goto err_cleanup_host;
+
+	pm_runtime_mark_last_busy(&pdev->dev);
+	pm_runtime_put_autosuspend(&pdev->dev);
+
+	return 0;
+
+err_cleanup_host:
+	sdhci_cleanup_host(host);
+
+pm_runtime_disable:
+	pm_runtime_put_noidle(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_set_suspended(&pdev->dev);
+
+	clk_disable_unprepare(sprd_host->clk_2x_enable);
+
+clk_disable2:
+	clk_disable_unprepare(sprd_host->clk_enable);
+
+clk_disable:
+	clk_disable_unprepare(sprd_host->clk_sdio);
+
+pltfm_free:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static int sdhci_sprd_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_sprd_host *sprd_host = TO_SPRD_HOST(host);
+
+	sdhci_remove_host(host, 0);
+
+	clk_disable_unprepare(sprd_host->clk_sdio);
+	clk_disable_unprepare(sprd_host->clk_enable);
+	clk_disable_unprepare(sprd_host->clk_2x_enable);
+
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+
+static const struct of_device_id sdhci_sprd_of_match[] = {
+	{ .compatible = "sprd,sdhci-r11", },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, sdhci_sprd_of_match);
+
+#ifdef CONFIG_PM
+static int sdhci_sprd_runtime_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_sprd_host *sprd_host = TO_SPRD_HOST(host);
+
+	mmc_hsq_suspend(host->mmc);
+	sdhci_runtime_suspend_host(host);
+
+	clk_disable_unprepare(sprd_host->clk_sdio);
+	clk_disable_unprepare(sprd_host->clk_enable);
+	clk_disable_unprepare(sprd_host->clk_2x_enable);
+
+	return 0;
+}
+
+static int sdhci_sprd_runtime_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_sprd_host *sprd_host = TO_SPRD_HOST(host);
+	int ret;
+
+	ret = clk_prepare_enable(sprd_host->clk_2x_enable);
+	if (ret)
+		return ret;
+
+	ret = clk_prepare_enable(sprd_host->clk_enable);
+	if (ret)
+		goto clk_2x_disable;
+
+	ret = clk_prepare_enable(sprd_host->clk_sdio);
+	if (ret)
+		goto clk_disable;
+
+	sdhci_runtime_resume_host(host, 1);
+	mmc_hsq_resume(host->mmc);
+
+	return 0;
+
+clk_disable:
+	clk_disable_unprepare(sprd_host->clk_enable);
+
+clk_2x_disable:
+	clk_disable_unprepare(sprd_host->clk_2x_enable);
+
+	return ret;
+}
+#endif
+
+static const struct dev_pm_ops sdhci_sprd_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(sdhci_sprd_runtime_suspend,
+			   sdhci_sprd_runtime_resume, NULL)
+};
+
+static struct platform_driver sdhci_sprd_driver = {
+	.probe = sdhci_sprd_probe,
+	.remove = sdhci_sprd_remove,
+	.driver = {
+		.name = "sdhci_sprd_r11",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(sdhci_sprd_of_match),
+		.pm = &sdhci_sprd_pm_ops,
+	},
+};
+module_platform_driver(sdhci_sprd_driver);
+
+MODULE_DESCRIPTION("Spreadtrum sdio host controller r11 driver");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:sdhci-sprd-r11");
diff --git a/drivers/mmc/host/sdhci-st.c b/drivers/mmc/host/sdhci-st.c
new file mode 100644
index 000000000..4e9ff3e82
--- /dev/null
+++ b/drivers/mmc/host/sdhci-st.c
@@ -0,0 +1,535 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Support for SDHCI on STMicroelectronics SoCs
+ *
+ * Copyright (C) 2014 STMicroelectronics Ltd
+ * Author: Giuseppe Cavallaro <peppe.cavallaro@st.com>
+ * Contributors: Peter Griffin <peter.griffin@linaro.org>
+ *
+ * Based on sdhci-cns3xxx.c
+ */
+
+#include <linux/io.h>
+#include <linux/of.h>
+#include <linux/module.h>
+#include <linux/err.h>
+#include <linux/mmc/host.h>
+#include <linux/reset.h>
+#include "sdhci-pltfm.h"
+
+struct st_mmc_platform_data {
+	struct  reset_control *rstc;
+	struct  clk *icnclk;
+	void __iomem *top_ioaddr;
+};
+
+/* MMCSS glue logic to setup the HC on some ST SoCs (e.g. STiH407 family) */
+
+#define ST_MMC_CCONFIG_REG_1		0x400
+#define ST_MMC_CCONFIG_TIMEOUT_CLK_UNIT	BIT(24)
+#define ST_MMC_CCONFIG_TIMEOUT_CLK_FREQ	BIT(12)
+#define ST_MMC_CCONFIG_TUNING_COUNT_DEFAULT	BIT(8)
+#define ST_MMC_CCONFIG_ASYNC_WAKEUP	BIT(0)
+#define ST_MMC_CCONFIG_1_DEFAULT	\
+				((ST_MMC_CCONFIG_TIMEOUT_CLK_UNIT) | \
+				 (ST_MMC_CCONFIG_TIMEOUT_CLK_FREQ) | \
+				 (ST_MMC_CCONFIG_TUNING_COUNT_DEFAULT))
+
+#define ST_MMC_CCONFIG_REG_2		0x404
+#define ST_MMC_CCONFIG_HIGH_SPEED	BIT(28)
+#define ST_MMC_CCONFIG_ADMA2		BIT(24)
+#define ST_MMC_CCONFIG_8BIT		BIT(20)
+#define ST_MMC_CCONFIG_MAX_BLK_LEN	16
+#define  MAX_BLK_LEN_1024		1
+#define  MAX_BLK_LEN_2048		2
+#define BASE_CLK_FREQ_200		0xc8
+#define BASE_CLK_FREQ_100		0x64
+#define BASE_CLK_FREQ_50		0x32
+#define ST_MMC_CCONFIG_2_DEFAULT \
+	(ST_MMC_CCONFIG_HIGH_SPEED | ST_MMC_CCONFIG_ADMA2 | \
+	 ST_MMC_CCONFIG_8BIT | \
+	 (MAX_BLK_LEN_1024 << ST_MMC_CCONFIG_MAX_BLK_LEN))
+
+#define ST_MMC_CCONFIG_REG_3			0x408
+#define ST_MMC_CCONFIG_EMMC_SLOT_TYPE		BIT(28)
+#define ST_MMC_CCONFIG_64BIT			BIT(24)
+#define ST_MMC_CCONFIG_ASYNCH_INTR_SUPPORT	BIT(20)
+#define ST_MMC_CCONFIG_1P8_VOLT			BIT(16)
+#define ST_MMC_CCONFIG_3P0_VOLT			BIT(12)
+#define ST_MMC_CCONFIG_3P3_VOLT			BIT(8)
+#define ST_MMC_CCONFIG_SUSP_RES_SUPPORT		BIT(4)
+#define ST_MMC_CCONFIG_SDMA			BIT(0)
+#define ST_MMC_CCONFIG_3_DEFAULT	\
+			 (ST_MMC_CCONFIG_ASYNCH_INTR_SUPPORT	| \
+			  ST_MMC_CCONFIG_3P3_VOLT		| \
+			  ST_MMC_CCONFIG_SUSP_RES_SUPPORT	| \
+			  ST_MMC_CCONFIG_SDMA)
+
+#define ST_MMC_CCONFIG_REG_4	0x40c
+#define ST_MMC_CCONFIG_D_DRIVER	BIT(20)
+#define ST_MMC_CCONFIG_C_DRIVER	BIT(16)
+#define ST_MMC_CCONFIG_A_DRIVER	BIT(12)
+#define ST_MMC_CCONFIG_DDR50	BIT(8)
+#define ST_MMC_CCONFIG_SDR104	BIT(4)
+#define ST_MMC_CCONFIG_SDR50	BIT(0)
+#define ST_MMC_CCONFIG_4_DEFAULT	0
+
+#define ST_MMC_CCONFIG_REG_5		0x410
+#define ST_MMC_CCONFIG_TUNING_FOR_SDR50	BIT(8)
+#define RETUNING_TIMER_CNT_MAX		0xf
+#define ST_MMC_CCONFIG_5_DEFAULT	0
+
+/* I/O configuration for Arasan IP */
+#define ST_MMC_GP_OUTPUT	0x450
+#define ST_MMC_GP_OUTPUT_CD	BIT(12)
+
+#define ST_MMC_STATUS_R		0x460
+
+#define ST_TOP_MMC_DLY_FIX_OFF(x)	(x - 0x8)
+
+/* TOP config registers to manage static and dynamic delay */
+#define ST_TOP_MMC_TX_CLK_DLY			ST_TOP_MMC_DLY_FIX_OFF(0x8)
+#define ST_TOP_MMC_RX_CLK_DLY			ST_TOP_MMC_DLY_FIX_OFF(0xc)
+/* MMC delay control register */
+#define ST_TOP_MMC_DLY_CTRL			ST_TOP_MMC_DLY_FIX_OFF(0x18)
+#define ST_TOP_MMC_DLY_CTRL_DLL_BYPASS_CMD	BIT(0)
+#define ST_TOP_MMC_DLY_CTRL_DLL_BYPASS_PH_SEL	BIT(1)
+#define ST_TOP_MMC_DLY_CTRL_TX_DLL_ENABLE	BIT(8)
+#define ST_TOP_MMC_DLY_CTRL_RX_DLL_ENABLE	BIT(9)
+#define ST_TOP_MMC_DLY_CTRL_ATUNE_NOT_CFG_DLY	BIT(10)
+#define ST_TOP_MMC_START_DLL_LOCK		BIT(11)
+
+/* register to provide the phase-shift value for DLL */
+#define ST_TOP_MMC_TX_DLL_STEP_DLY		ST_TOP_MMC_DLY_FIX_OFF(0x1c)
+#define ST_TOP_MMC_RX_DLL_STEP_DLY		ST_TOP_MMC_DLY_FIX_OFF(0x20)
+#define ST_TOP_MMC_RX_CMD_STEP_DLY		ST_TOP_MMC_DLY_FIX_OFF(0x24)
+
+/* phase shift delay on the tx clk 2.188ns */
+#define ST_TOP_MMC_TX_DLL_STEP_DLY_VALID	0x6
+
+#define ST_TOP_MMC_DLY_MAX			0xf
+
+#define ST_TOP_MMC_DYN_DLY_CONF	\
+		(ST_TOP_MMC_DLY_CTRL_TX_DLL_ENABLE | \
+		 ST_TOP_MMC_DLY_CTRL_ATUNE_NOT_CFG_DLY | \
+		 ST_TOP_MMC_START_DLL_LOCK)
+
+/*
+ * For clock speeds greater than 90MHz, we need to check that the
+ * DLL procedure has finished before switching to ultra-speed modes.
+ */
+#define	CLK_TO_CHECK_DLL_LOCK	90000000
+
+static inline void st_mmcss_set_static_delay(void __iomem *ioaddr)
+{
+	if (!ioaddr)
+		return;
+
+	writel_relaxed(0x0, ioaddr + ST_TOP_MMC_DLY_CTRL);
+	writel_relaxed(ST_TOP_MMC_DLY_MAX,
+			ioaddr + ST_TOP_MMC_TX_CLK_DLY);
+}
+
+/**
+ * st_mmcss_cconfig: configure the Arasan HC inside the flashSS.
+ * @np: dt device node.
+ * @host: sdhci host
+ * Description: this function is to configure the Arasan host controller.
+ * On some ST SoCs, i.e. STiH407 family, the MMC devices inside a dedicated
+ * flashSS sub-system which needs to be configured to be compliant to eMMC 4.5
+ * or eMMC4.3.  This has to be done before registering the sdhci host.
+ */
+static void st_mmcss_cconfig(struct device_node *np, struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct mmc_host *mhost = host->mmc;
+	u32 cconf2, cconf3, cconf4, cconf5;
+
+	if (!of_device_is_compatible(np, "st,sdhci-stih407"))
+		return;
+
+	cconf2 = ST_MMC_CCONFIG_2_DEFAULT;
+	cconf3 = ST_MMC_CCONFIG_3_DEFAULT;
+	cconf4 = ST_MMC_CCONFIG_4_DEFAULT;
+	cconf5 = ST_MMC_CCONFIG_5_DEFAULT;
+
+	writel_relaxed(ST_MMC_CCONFIG_1_DEFAULT,
+			host->ioaddr + ST_MMC_CCONFIG_REG_1);
+
+	/* Set clock frequency, default to 50MHz if max-frequency is not
+	 * provided */
+
+	switch (mhost->f_max) {
+	case 200000000:
+		clk_set_rate(pltfm_host->clk, mhost->f_max);
+		cconf2 |= BASE_CLK_FREQ_200;
+		break;
+	case 100000000:
+		clk_set_rate(pltfm_host->clk, mhost->f_max);
+		cconf2 |= BASE_CLK_FREQ_100;
+		break;
+	default:
+		clk_set_rate(pltfm_host->clk, 50000000);
+		cconf2 |= BASE_CLK_FREQ_50;
+	}
+
+	writel_relaxed(cconf2, host->ioaddr + ST_MMC_CCONFIG_REG_2);
+
+	if (!mmc_card_is_removable(mhost))
+		cconf3 |= ST_MMC_CCONFIG_EMMC_SLOT_TYPE;
+	else
+		/* CARD _D ET_CTRL */
+		writel_relaxed(ST_MMC_GP_OUTPUT_CD,
+				host->ioaddr + ST_MMC_GP_OUTPUT);
+
+	if (mhost->caps & MMC_CAP_UHS_SDR50) {
+		/* use 1.8V */
+		cconf3 |= ST_MMC_CCONFIG_1P8_VOLT;
+		cconf4 |= ST_MMC_CCONFIG_SDR50;
+		/* Use tuning */
+		cconf5 |= ST_MMC_CCONFIG_TUNING_FOR_SDR50;
+		/* Max timeout for retuning */
+		cconf5 |= RETUNING_TIMER_CNT_MAX;
+	}
+
+	if (mhost->caps & MMC_CAP_UHS_SDR104) {
+		/*
+		 * SDR104 implies the HC can support HS200 mode, so
+		 * it's mandatory to use 1.8V
+		 */
+		cconf3 |= ST_MMC_CCONFIG_1P8_VOLT;
+		cconf4 |= ST_MMC_CCONFIG_SDR104;
+		/* Max timeout for retuning */
+		cconf5 |= RETUNING_TIMER_CNT_MAX;
+	}
+
+	if (mhost->caps & MMC_CAP_UHS_DDR50)
+		cconf4 |= ST_MMC_CCONFIG_DDR50;
+
+	writel_relaxed(cconf3, host->ioaddr + ST_MMC_CCONFIG_REG_3);
+	writel_relaxed(cconf4, host->ioaddr + ST_MMC_CCONFIG_REG_4);
+	writel_relaxed(cconf5, host->ioaddr + ST_MMC_CCONFIG_REG_5);
+}
+
+static inline void st_mmcss_set_dll(void __iomem *ioaddr)
+{
+	if (!ioaddr)
+		return;
+
+	writel_relaxed(ST_TOP_MMC_DYN_DLY_CONF,	ioaddr + ST_TOP_MMC_DLY_CTRL);
+	writel_relaxed(ST_TOP_MMC_TX_DLL_STEP_DLY_VALID,
+			ioaddr + ST_TOP_MMC_TX_DLL_STEP_DLY);
+}
+
+static int st_mmcss_lock_dll(void __iomem *ioaddr)
+{
+	unsigned long curr, value;
+	unsigned long finish = jiffies + HZ;
+
+	/* Checks if the DLL procedure is finished */
+	do {
+		curr = jiffies;
+		value = readl(ioaddr + ST_MMC_STATUS_R);
+		if (value & 0x1)
+			return 0;
+
+		cpu_relax();
+	} while (!time_after_eq(curr, finish));
+
+	return -EBUSY;
+}
+
+static int sdhci_st_set_dll_for_clock(struct sdhci_host *host)
+{
+	int ret = 0;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct st_mmc_platform_data *pdata = sdhci_pltfm_priv(pltfm_host);
+
+	if (host->clock > CLK_TO_CHECK_DLL_LOCK) {
+		st_mmcss_set_dll(pdata->top_ioaddr);
+		ret = st_mmcss_lock_dll(host->ioaddr);
+	}
+
+	return ret;
+}
+
+static void sdhci_st_set_uhs_signaling(struct sdhci_host *host,
+					unsigned int uhs)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct st_mmc_platform_data *pdata = sdhci_pltfm_priv(pltfm_host);
+	u16 ctrl_2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	int ret = 0;
+
+	/* Select Bus Speed Mode for host */
+	ctrl_2 &= ~SDHCI_CTRL_UHS_MASK;
+	switch (uhs) {
+	/*
+	 * Set V18_EN -- UHS modes do not work without this.
+	 * does not change signaling voltage
+	 */
+
+	case MMC_TIMING_UHS_SDR12:
+		st_mmcss_set_static_delay(pdata->top_ioaddr);
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR12 | SDHCI_CTRL_VDD_180;
+		break;
+	case MMC_TIMING_UHS_SDR25:
+		st_mmcss_set_static_delay(pdata->top_ioaddr);
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR25 | SDHCI_CTRL_VDD_180;
+		break;
+	case MMC_TIMING_UHS_SDR50:
+		st_mmcss_set_static_delay(pdata->top_ioaddr);
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR50 | SDHCI_CTRL_VDD_180;
+		ret = sdhci_st_set_dll_for_clock(host);
+		break;
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_MMC_HS200:
+		st_mmcss_set_static_delay(pdata->top_ioaddr);
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR104 | SDHCI_CTRL_VDD_180;
+		ret =  sdhci_st_set_dll_for_clock(host);
+		break;
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_MMC_DDR52:
+		st_mmcss_set_static_delay(pdata->top_ioaddr);
+		ctrl_2 |= SDHCI_CTRL_UHS_DDR50 | SDHCI_CTRL_VDD_180;
+		break;
+	}
+
+	if (ret)
+		dev_warn(mmc_dev(host->mmc), "Error setting dll for clock "
+						"(uhs %d)\n", uhs);
+
+	dev_dbg(mmc_dev(host->mmc), "uhs %d, ctrl_2 %04X\n", uhs, ctrl_2);
+
+	sdhci_writew(host, ctrl_2, SDHCI_HOST_CONTROL2);
+}
+
+static u32 sdhci_st_readl(struct sdhci_host *host, int reg)
+{
+	u32 ret;
+
+	switch (reg) {
+	case SDHCI_CAPABILITIES:
+		ret = readl_relaxed(host->ioaddr + reg);
+		/* Support 3.3V and 1.8V */
+		ret &= ~SDHCI_CAN_VDD_300;
+		break;
+	default:
+		ret = readl_relaxed(host->ioaddr + reg);
+	}
+	return ret;
+}
+
+static const struct sdhci_ops sdhci_st_ops = {
+	.get_max_clock = sdhci_pltfm_clk_get_max_clock,
+	.set_clock = sdhci_set_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.read_l = sdhci_st_readl,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_st_set_uhs_signaling,
+};
+
+static const struct sdhci_pltfm_data sdhci_st_pdata = {
+	.ops = &sdhci_st_ops,
+	.quirks = SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+		SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN |
+		SDHCI_QUIRK_NO_HISPD_BIT,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+		SDHCI_QUIRK2_STOP_WITH_TC,
+};
+
+
+static int sdhci_st_probe(struct platform_device *pdev)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct sdhci_host *host;
+	struct st_mmc_platform_data *pdata;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct clk *clk, *icnclk;
+	int ret = 0;
+	u16 host_version;
+	struct resource *res;
+	struct reset_control *rstc;
+
+	clk =  devm_clk_get(&pdev->dev, "mmc");
+	if (IS_ERR(clk)) {
+		dev_err(&pdev->dev, "Peripheral clk not found\n");
+		return PTR_ERR(clk);
+	}
+
+	/* ICN clock isn't compulsory, but use it if it's provided. */
+	icnclk = devm_clk_get(&pdev->dev, "icn");
+	if (IS_ERR(icnclk))
+		icnclk = NULL;
+
+	rstc = devm_reset_control_get_exclusive(&pdev->dev, NULL);
+	if (IS_ERR(rstc))
+		rstc = NULL;
+	else
+		reset_control_deassert(rstc);
+
+	host = sdhci_pltfm_init(pdev, &sdhci_st_pdata, sizeof(*pdata));
+	if (IS_ERR(host)) {
+		dev_err(&pdev->dev, "Failed sdhci_pltfm_init\n");
+		ret = PTR_ERR(host);
+		goto err_pltfm_init;
+	}
+
+	pltfm_host = sdhci_priv(host);
+	pdata = sdhci_pltfm_priv(pltfm_host);
+	pdata->rstc = rstc;
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed mmc_of_parse\n");
+		goto err_of;
+	}
+
+	ret = clk_prepare_enable(clk);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to prepare clock\n");
+		goto err_of;
+	}
+
+	ret = clk_prepare_enable(icnclk);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to prepare icn clock\n");
+		goto err_icnclk;
+	}
+
+	/* Configure the FlashSS Top registers for setting eMMC TX/RX delay */
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
+					   "top-mmc-delay");
+	pdata->top_ioaddr = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(pdata->top_ioaddr)) {
+		dev_warn(&pdev->dev, "FlashSS Top Dly registers not available");
+		pdata->top_ioaddr = NULL;
+	}
+
+	pltfm_host->clk = clk;
+	pdata->icnclk = icnclk;
+
+	/* Configure the Arasan HC inside the flashSS */
+	st_mmcss_cconfig(np, host);
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err_out;
+
+	host_version = readw_relaxed((host->ioaddr + SDHCI_HOST_VERSION));
+
+	dev_info(&pdev->dev, "SDHCI ST Initialised: Host Version: 0x%x Vendor Version 0x%x\n",
+		((host_version & SDHCI_SPEC_VER_MASK) >> SDHCI_SPEC_VER_SHIFT),
+		((host_version & SDHCI_VENDOR_VER_MASK) >>
+		SDHCI_VENDOR_VER_SHIFT));
+
+	return 0;
+
+err_out:
+	clk_disable_unprepare(icnclk);
+err_icnclk:
+	clk_disable_unprepare(clk);
+err_of:
+	sdhci_pltfm_free(pdev);
+err_pltfm_init:
+	if (rstc)
+		reset_control_assert(rstc);
+
+	return ret;
+}
+
+static int sdhci_st_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct st_mmc_platform_data *pdata = sdhci_pltfm_priv(pltfm_host);
+	struct reset_control *rstc = pdata->rstc;
+	int ret;
+
+	ret = sdhci_pltfm_unregister(pdev);
+
+	clk_disable_unprepare(pdata->icnclk);
+
+	if (rstc)
+		reset_control_assert(rstc);
+
+	return ret;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int sdhci_st_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct st_mmc_platform_data *pdata = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	ret = sdhci_suspend_host(host);
+	if (ret)
+		goto out;
+
+	if (pdata->rstc)
+		reset_control_assert(pdata->rstc);
+
+	clk_disable_unprepare(pdata->icnclk);
+	clk_disable_unprepare(pltfm_host->clk);
+out:
+	return ret;
+}
+
+static int sdhci_st_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct st_mmc_platform_data *pdata = sdhci_pltfm_priv(pltfm_host);
+	struct device_node *np = dev->of_node;
+	int ret;
+
+	ret = clk_prepare_enable(pltfm_host->clk);
+	if (ret)
+		return ret;
+
+	ret = clk_prepare_enable(pdata->icnclk);
+	if (ret) {
+		clk_disable_unprepare(pltfm_host->clk);
+		return ret;
+	}
+
+	if (pdata->rstc)
+		reset_control_deassert(pdata->rstc);
+
+	st_mmcss_cconfig(np, host);
+
+	return sdhci_resume_host(host);
+}
+#endif
+
+static SIMPLE_DEV_PM_OPS(sdhci_st_pmops, sdhci_st_suspend, sdhci_st_resume);
+
+static const struct of_device_id st_sdhci_match[] = {
+	{ .compatible = "st,sdhci" },
+	{},
+};
+
+MODULE_DEVICE_TABLE(of, st_sdhci_match);
+
+static struct platform_driver sdhci_st_driver = {
+	.probe = sdhci_st_probe,
+	.remove = sdhci_st_remove,
+	.driver = {
+		   .name = "sdhci-st",
+		   .probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		   .pm = &sdhci_st_pmops,
+		   .of_match_table = of_match_ptr(st_sdhci_match),
+		  },
+};
+
+module_platform_driver(sdhci_st_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for STMicroelectronics SoCs");
+MODULE_AUTHOR("Giuseppe Cavallaro <peppe.cavallaro@st.com>");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:sdhci-st");
diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
new file mode 100644
index 000000000..d8fd2b5ef
--- /dev/null
+++ b/drivers/mmc/host/sdhci-tegra.c
@@ -0,0 +1,1839 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2010 Google, Inc.
+ */
+
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/err.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/iopoll.h>
+#include <linux/platform_device.h>
+#include <linux/clk.h>
+#include <linux/io.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/regulator/consumer.h>
+#include <linux/reset.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/gpio/consumer.h>
+#include <linux/ktime.h>
+
+#include "sdhci-cqhci.h"
+#include "sdhci-pltfm.h"
+#include "cqhci.h"
+
+/* Tegra SDHOST controller vendor register definitions */
+#define SDHCI_TEGRA_VENDOR_CLOCK_CTRL			0x100
+#define SDHCI_CLOCK_CTRL_TAP_MASK			0x00ff0000
+#define SDHCI_CLOCK_CTRL_TAP_SHIFT			16
+#define SDHCI_CLOCK_CTRL_TRIM_MASK			0x1f000000
+#define SDHCI_CLOCK_CTRL_TRIM_SHIFT			24
+#define SDHCI_CLOCK_CTRL_SDR50_TUNING_OVERRIDE		BIT(5)
+#define SDHCI_CLOCK_CTRL_PADPIPE_CLKEN_OVERRIDE		BIT(3)
+#define SDHCI_CLOCK_CTRL_SPI_MODE_CLKEN_OVERRIDE	BIT(2)
+
+#define SDHCI_TEGRA_VENDOR_SYS_SW_CTRL			0x104
+#define SDHCI_TEGRA_SYS_SW_CTRL_ENHANCED_STROBE		BIT(31)
+
+#define SDHCI_TEGRA_VENDOR_CAP_OVERRIDES		0x10c
+#define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_MASK		0x00003f00
+#define SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT	8
+
+#define SDHCI_TEGRA_VENDOR_MISC_CTRL			0x120
+#define SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT		BIT(0)
+#define SDHCI_MISC_CTRL_ENABLE_SDR104			0x8
+#define SDHCI_MISC_CTRL_ENABLE_SDR50			0x10
+#define SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300		0x20
+#define SDHCI_MISC_CTRL_ENABLE_DDR50			0x200
+
+#define SDHCI_TEGRA_VENDOR_DLLCAL_CFG			0x1b0
+#define SDHCI_TEGRA_DLLCAL_CALIBRATE			BIT(31)
+
+#define SDHCI_TEGRA_VENDOR_DLLCAL_STA			0x1bc
+#define SDHCI_TEGRA_DLLCAL_STA_ACTIVE			BIT(31)
+
+#define SDHCI_VNDR_TUN_CTRL0_0				0x1c0
+#define SDHCI_VNDR_TUN_CTRL0_TUN_HW_TAP			0x20000
+#define SDHCI_VNDR_TUN_CTRL0_START_TAP_VAL_MASK		0x03fc0000
+#define SDHCI_VNDR_TUN_CTRL0_START_TAP_VAL_SHIFT	18
+#define SDHCI_VNDR_TUN_CTRL0_MUL_M_MASK			0x00001fc0
+#define SDHCI_VNDR_TUN_CTRL0_MUL_M_SHIFT		6
+#define SDHCI_VNDR_TUN_CTRL0_TUN_ITER_MASK		0x000e000
+#define SDHCI_VNDR_TUN_CTRL0_TUN_ITER_SHIFT		13
+#define TRIES_128					2
+#define TRIES_256					4
+#define SDHCI_VNDR_TUN_CTRL0_TUN_WORD_SEL_MASK		0x7
+
+#define SDHCI_TEGRA_VNDR_TUN_CTRL1_0			0x1c4
+#define SDHCI_TEGRA_VNDR_TUN_STATUS0			0x1C8
+#define SDHCI_TEGRA_VNDR_TUN_STATUS1			0x1CC
+#define SDHCI_TEGRA_VNDR_TUN_STATUS1_TAP_MASK		0xFF
+#define SDHCI_TEGRA_VNDR_TUN_STATUS1_END_TAP_SHIFT	0x8
+#define TUNING_WORD_BIT_SIZE				32
+
+#define SDHCI_TEGRA_AUTO_CAL_CONFIG			0x1e4
+#define SDHCI_AUTO_CAL_START				BIT(31)
+#define SDHCI_AUTO_CAL_ENABLE				BIT(29)
+#define SDHCI_AUTO_CAL_PDPU_OFFSET_MASK			0x0000ffff
+
+#define SDHCI_TEGRA_SDMEM_COMP_PADCTRL			0x1e0
+#define SDHCI_TEGRA_SDMEM_COMP_PADCTRL_VREF_SEL_MASK	0x0000000f
+#define SDHCI_TEGRA_SDMEM_COMP_PADCTRL_VREF_SEL_VAL	0x7
+#define SDHCI_TEGRA_SDMEM_COMP_PADCTRL_E_INPUT_E_PWRD	BIT(31)
+#define SDHCI_COMP_PADCTRL_DRVUPDN_OFFSET_MASK		0x07FFF000
+
+#define SDHCI_TEGRA_AUTO_CAL_STATUS			0x1ec
+#define SDHCI_TEGRA_AUTO_CAL_ACTIVE			BIT(31)
+
+#define NVQUIRK_FORCE_SDHCI_SPEC_200			BIT(0)
+#define NVQUIRK_ENABLE_BLOCK_GAP_DET			BIT(1)
+#define NVQUIRK_ENABLE_SDHCI_SPEC_300			BIT(2)
+#define NVQUIRK_ENABLE_SDR50				BIT(3)
+#define NVQUIRK_ENABLE_SDR104				BIT(4)
+#define NVQUIRK_ENABLE_DDR50				BIT(5)
+/*
+ * HAS_PADCALIB NVQUIRK is for SoC's supporting auto calibration of pads
+ * drive strength.
+ */
+#define NVQUIRK_HAS_PADCALIB				BIT(6)
+/*
+ * NEEDS_PAD_CONTROL NVQUIRK is for SoC's having separate 3V3 and 1V8 pads.
+ * 3V3/1V8 pad selection happens through pinctrl state selection depending
+ * on the signaling mode.
+ */
+#define NVQUIRK_NEEDS_PAD_CONTROL			BIT(7)
+#define NVQUIRK_DIS_CARD_CLK_CONFIG_TAP			BIT(8)
+#define NVQUIRK_CQHCI_DCMD_R1B_CMD_TIMING		BIT(9)
+
+/*
+ * NVQUIRK_HAS_TMCLK is for SoC's having separate timeout clock for Tegra
+ * SDMMC hardware data timeout.
+ */
+#define NVQUIRK_HAS_TMCLK				BIT(10)
+
+/* SDMMC CQE Base Address for Tegra Host Ver 4.1 and Higher */
+#define SDHCI_TEGRA_CQE_BASE_ADDR			0xF000
+
+#define SDHCI_TEGRA_CQE_TRNS_MODE	(SDHCI_TRNS_MULTI | \
+					 SDHCI_TRNS_BLK_CNT_EN | \
+					 SDHCI_TRNS_DMA)
+
+struct sdhci_tegra_soc_data {
+	const struct sdhci_pltfm_data *pdata;
+	u64 dma_mask;
+	u32 nvquirks;
+	u8 min_tap_delay;
+	u8 max_tap_delay;
+};
+
+/* Magic pull up and pull down pad calibration offsets */
+struct sdhci_tegra_autocal_offsets {
+	u32 pull_up_3v3;
+	u32 pull_down_3v3;
+	u32 pull_up_3v3_timeout;
+	u32 pull_down_3v3_timeout;
+	u32 pull_up_1v8;
+	u32 pull_down_1v8;
+	u32 pull_up_1v8_timeout;
+	u32 pull_down_1v8_timeout;
+	u32 pull_up_sdr104;
+	u32 pull_down_sdr104;
+	u32 pull_up_hs400;
+	u32 pull_down_hs400;
+};
+
+struct sdhci_tegra {
+	const struct sdhci_tegra_soc_data *soc_data;
+	struct gpio_desc *power_gpio;
+	struct clk *tmclk;
+	bool ddr_signaling;
+	bool pad_calib_required;
+	bool pad_control_available;
+
+	struct reset_control *rst;
+	struct pinctrl *pinctrl_sdmmc;
+	struct pinctrl_state *pinctrl_state_3v3;
+	struct pinctrl_state *pinctrl_state_1v8;
+	struct pinctrl_state *pinctrl_state_3v3_drv;
+	struct pinctrl_state *pinctrl_state_1v8_drv;
+
+	struct sdhci_tegra_autocal_offsets autocal_offsets;
+	ktime_t last_calib;
+
+	u32 default_tap;
+	u32 default_trim;
+	u32 dqs_trim;
+	bool enable_hwcq;
+	unsigned long curr_clk_rate;
+	u8 tuned_tap_delay;
+};
+
+static u16 tegra_sdhci_readw(struct sdhci_host *host, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	const struct sdhci_tegra_soc_data *soc_data = tegra_host->soc_data;
+
+	if (unlikely((soc_data->nvquirks & NVQUIRK_FORCE_SDHCI_SPEC_200) &&
+			(reg == SDHCI_HOST_VERSION))) {
+		/* Erratum: Version register is invalid in HW. */
+		return SDHCI_SPEC_200;
+	}
+
+	return readw(host->ioaddr + reg);
+}
+
+static void tegra_sdhci_writew(struct sdhci_host *host, u16 val, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	switch (reg) {
+	case SDHCI_TRANSFER_MODE:
+		/*
+		 * Postpone this write, we must do it together with a
+		 * command write that is down below.
+		 */
+		pltfm_host->xfer_mode_shadow = val;
+		return;
+	case SDHCI_COMMAND:
+		writel((val << 16) | pltfm_host->xfer_mode_shadow,
+			host->ioaddr + SDHCI_TRANSFER_MODE);
+		return;
+	}
+
+	writew(val, host->ioaddr + reg);
+}
+
+static void tegra_sdhci_writel(struct sdhci_host *host, u32 val, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	const struct sdhci_tegra_soc_data *soc_data = tegra_host->soc_data;
+
+	/* Seems like we're getting spurious timeout and crc errors, so
+	 * disable signalling of them. In case of real errors software
+	 * timers should take care of eventually detecting them.
+	 */
+	if (unlikely(reg == SDHCI_SIGNAL_ENABLE))
+		val &= ~(SDHCI_INT_TIMEOUT|SDHCI_INT_CRC);
+
+	writel(val, host->ioaddr + reg);
+
+	if (unlikely((soc_data->nvquirks & NVQUIRK_ENABLE_BLOCK_GAP_DET) &&
+			(reg == SDHCI_INT_ENABLE))) {
+		/* Erratum: Must enable block gap interrupt detection */
+		u8 gap_ctrl = readb(host->ioaddr + SDHCI_BLOCK_GAP_CONTROL);
+		if (val & SDHCI_INT_CARD_INT)
+			gap_ctrl |= 0x8;
+		else
+			gap_ctrl &= ~0x8;
+		writeb(gap_ctrl, host->ioaddr + SDHCI_BLOCK_GAP_CONTROL);
+	}
+}
+
+static bool tegra_sdhci_configure_card_clk(struct sdhci_host *host, bool enable)
+{
+	bool status;
+	u32 reg;
+
+	reg = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+	status = !!(reg & SDHCI_CLOCK_CARD_EN);
+
+	if (status == enable)
+		return status;
+
+	if (enable)
+		reg |= SDHCI_CLOCK_CARD_EN;
+	else
+		reg &= ~SDHCI_CLOCK_CARD_EN;
+
+	sdhci_writew(host, reg, SDHCI_CLOCK_CONTROL);
+
+	return status;
+}
+
+static void tegra210_sdhci_writew(struct sdhci_host *host, u16 val, int reg)
+{
+	bool is_tuning_cmd = 0;
+	bool clk_enabled;
+	u8 cmd;
+
+	if (reg == SDHCI_COMMAND) {
+		cmd = SDHCI_GET_CMD(val);
+		is_tuning_cmd = cmd == MMC_SEND_TUNING_BLOCK ||
+				cmd == MMC_SEND_TUNING_BLOCK_HS200;
+	}
+
+	if (is_tuning_cmd)
+		clk_enabled = tegra_sdhci_configure_card_clk(host, 0);
+
+	writew(val, host->ioaddr + reg);
+
+	if (is_tuning_cmd) {
+		udelay(1);
+		sdhci_reset(host, SDHCI_RESET_CMD | SDHCI_RESET_DATA);
+		tegra_sdhci_configure_card_clk(host, clk_enabled);
+	}
+}
+
+static unsigned int tegra_sdhci_get_ro(struct sdhci_host *host)
+{
+	/*
+	 * Write-enable shall be assumed if GPIO is missing in a board's
+	 * device-tree because SDHCI's WRITE_PROTECT bit doesn't work on
+	 * Tegra.
+	 */
+	return mmc_gpio_get_ro(host->mmc);
+}
+
+static bool tegra_sdhci_is_pad_and_regulator_valid(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	int has_1v8, has_3v3;
+
+	/*
+	 * The SoCs which have NVQUIRK_NEEDS_PAD_CONTROL require software pad
+	 * voltage configuration in order to perform voltage switching. This
+	 * means that valid pinctrl info is required on SDHCI instances capable
+	 * of performing voltage switching. Whether or not an SDHCI instance is
+	 * capable of voltage switching is determined based on the regulator.
+	 */
+
+	if (!(tegra_host->soc_data->nvquirks & NVQUIRK_NEEDS_PAD_CONTROL))
+		return true;
+
+	if (IS_ERR(host->mmc->supply.vqmmc))
+		return false;
+
+	has_1v8 = regulator_is_supported_voltage(host->mmc->supply.vqmmc,
+						 1700000, 1950000);
+
+	has_3v3 = regulator_is_supported_voltage(host->mmc->supply.vqmmc,
+						 2700000, 3600000);
+
+	if (has_1v8 == 1 && has_3v3 == 1)
+		return tegra_host->pad_control_available;
+
+	/* Fixed voltage, no pad control required. */
+	return true;
+}
+
+static void tegra_sdhci_set_tap(struct sdhci_host *host, unsigned int tap)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	const struct sdhci_tegra_soc_data *soc_data = tegra_host->soc_data;
+	bool card_clk_enabled = false;
+	u32 reg;
+
+	/*
+	 * Touching the tap values is a bit tricky on some SoC generations.
+	 * The quirk enables a workaround for a glitch that sometimes occurs if
+	 * the tap values are changed.
+	 */
+
+	if (soc_data->nvquirks & NVQUIRK_DIS_CARD_CLK_CONFIG_TAP)
+		card_clk_enabled = tegra_sdhci_configure_card_clk(host, false);
+
+	reg = sdhci_readl(host, SDHCI_TEGRA_VENDOR_CLOCK_CTRL);
+	reg &= ~SDHCI_CLOCK_CTRL_TAP_MASK;
+	reg |= tap << SDHCI_CLOCK_CTRL_TAP_SHIFT;
+	sdhci_writel(host, reg, SDHCI_TEGRA_VENDOR_CLOCK_CTRL);
+
+	if (soc_data->nvquirks & NVQUIRK_DIS_CARD_CLK_CONFIG_TAP &&
+	    card_clk_enabled) {
+		udelay(1);
+		sdhci_reset(host, SDHCI_RESET_CMD | SDHCI_RESET_DATA);
+		tegra_sdhci_configure_card_clk(host, card_clk_enabled);
+	}
+}
+
+static void tegra_sdhci_reset(struct sdhci_host *host, u8 mask)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	const struct sdhci_tegra_soc_data *soc_data = tegra_host->soc_data;
+	u32 misc_ctrl, clk_ctrl, pad_ctrl;
+
+	sdhci_and_cqhci_reset(host, mask);
+
+	if (!(mask & SDHCI_RESET_ALL))
+		return;
+
+	tegra_sdhci_set_tap(host, tegra_host->default_tap);
+
+	misc_ctrl = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+	clk_ctrl = sdhci_readl(host, SDHCI_TEGRA_VENDOR_CLOCK_CTRL);
+
+	misc_ctrl &= ~(SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300 |
+		       SDHCI_MISC_CTRL_ENABLE_SDR50 |
+		       SDHCI_MISC_CTRL_ENABLE_DDR50 |
+		       SDHCI_MISC_CTRL_ENABLE_SDR104);
+
+	clk_ctrl &= ~(SDHCI_CLOCK_CTRL_TRIM_MASK |
+		      SDHCI_CLOCK_CTRL_SPI_MODE_CLKEN_OVERRIDE);
+
+	if (tegra_sdhci_is_pad_and_regulator_valid(host)) {
+		/* Erratum: Enable SDHCI spec v3.00 support */
+		if (soc_data->nvquirks & NVQUIRK_ENABLE_SDHCI_SPEC_300)
+			misc_ctrl |= SDHCI_MISC_CTRL_ENABLE_SDHCI_SPEC_300;
+		/* Advertise UHS modes as supported by host */
+		if (soc_data->nvquirks & NVQUIRK_ENABLE_SDR50)
+			misc_ctrl |= SDHCI_MISC_CTRL_ENABLE_SDR50;
+		if (soc_data->nvquirks & NVQUIRK_ENABLE_DDR50)
+			misc_ctrl |= SDHCI_MISC_CTRL_ENABLE_DDR50;
+		if (soc_data->nvquirks & NVQUIRK_ENABLE_SDR104)
+			misc_ctrl |= SDHCI_MISC_CTRL_ENABLE_SDR104;
+		if (soc_data->nvquirks & NVQUIRK_ENABLE_SDR50)
+			clk_ctrl |= SDHCI_CLOCK_CTRL_SDR50_TUNING_OVERRIDE;
+	}
+
+	clk_ctrl |= tegra_host->default_trim << SDHCI_CLOCK_CTRL_TRIM_SHIFT;
+
+	sdhci_writel(host, misc_ctrl, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+	sdhci_writel(host, clk_ctrl, SDHCI_TEGRA_VENDOR_CLOCK_CTRL);
+
+	if (soc_data->nvquirks & NVQUIRK_HAS_PADCALIB) {
+		pad_ctrl = sdhci_readl(host, SDHCI_TEGRA_SDMEM_COMP_PADCTRL);
+		pad_ctrl &= ~SDHCI_TEGRA_SDMEM_COMP_PADCTRL_VREF_SEL_MASK;
+		pad_ctrl |= SDHCI_TEGRA_SDMEM_COMP_PADCTRL_VREF_SEL_VAL;
+		sdhci_writel(host, pad_ctrl, SDHCI_TEGRA_SDMEM_COMP_PADCTRL);
+
+		tegra_host->pad_calib_required = true;
+	}
+
+	tegra_host->ddr_signaling = false;
+}
+
+static void tegra_sdhci_configure_cal_pad(struct sdhci_host *host, bool enable)
+{
+	u32 val;
+
+	/*
+	 * Enable or disable the additional I/O pad used by the drive strength
+	 * calibration process.
+	 */
+	val = sdhci_readl(host, SDHCI_TEGRA_SDMEM_COMP_PADCTRL);
+
+	if (enable)
+		val |= SDHCI_TEGRA_SDMEM_COMP_PADCTRL_E_INPUT_E_PWRD;
+	else
+		val &= ~SDHCI_TEGRA_SDMEM_COMP_PADCTRL_E_INPUT_E_PWRD;
+
+	sdhci_writel(host, val, SDHCI_TEGRA_SDMEM_COMP_PADCTRL);
+
+	if (enable)
+		usleep_range(1, 2);
+}
+
+static void tegra_sdhci_set_pad_autocal_offset(struct sdhci_host *host,
+					       u16 pdpu)
+{
+	u32 reg;
+
+	reg = sdhci_readl(host, SDHCI_TEGRA_AUTO_CAL_CONFIG);
+	reg &= ~SDHCI_AUTO_CAL_PDPU_OFFSET_MASK;
+	reg |= pdpu;
+	sdhci_writel(host, reg, SDHCI_TEGRA_AUTO_CAL_CONFIG);
+}
+
+static int tegra_sdhci_set_padctrl(struct sdhci_host *host, int voltage,
+				   bool state_drvupdn)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	struct sdhci_tegra_autocal_offsets *offsets =
+						&tegra_host->autocal_offsets;
+	struct pinctrl_state *pinctrl_drvupdn = NULL;
+	int ret = 0;
+	u8 drvup = 0, drvdn = 0;
+	u32 reg;
+
+	if (!state_drvupdn) {
+		/* PADS Drive Strength */
+		if (voltage == MMC_SIGNAL_VOLTAGE_180) {
+			if (tegra_host->pinctrl_state_1v8_drv) {
+				pinctrl_drvupdn =
+					tegra_host->pinctrl_state_1v8_drv;
+			} else {
+				drvup = offsets->pull_up_1v8_timeout;
+				drvdn = offsets->pull_down_1v8_timeout;
+			}
+		} else {
+			if (tegra_host->pinctrl_state_3v3_drv) {
+				pinctrl_drvupdn =
+					tegra_host->pinctrl_state_3v3_drv;
+			} else {
+				drvup = offsets->pull_up_3v3_timeout;
+				drvdn = offsets->pull_down_3v3_timeout;
+			}
+		}
+
+		if (pinctrl_drvupdn != NULL) {
+			ret = pinctrl_select_state(tegra_host->pinctrl_sdmmc,
+							pinctrl_drvupdn);
+			if (ret < 0)
+				dev_err(mmc_dev(host->mmc),
+					"failed pads drvupdn, ret: %d\n", ret);
+		} else if ((drvup) || (drvdn)) {
+			reg = sdhci_readl(host,
+					SDHCI_TEGRA_SDMEM_COMP_PADCTRL);
+			reg &= ~SDHCI_COMP_PADCTRL_DRVUPDN_OFFSET_MASK;
+			reg |= (drvup << 20) | (drvdn << 12);
+			sdhci_writel(host, reg,
+					SDHCI_TEGRA_SDMEM_COMP_PADCTRL);
+		}
+
+	} else {
+		/* Dual Voltage PADS Voltage selection */
+		if (!tegra_host->pad_control_available)
+			return 0;
+
+		if (voltage == MMC_SIGNAL_VOLTAGE_180) {
+			ret = pinctrl_select_state(tegra_host->pinctrl_sdmmc,
+						tegra_host->pinctrl_state_1v8);
+			if (ret < 0)
+				dev_err(mmc_dev(host->mmc),
+					"setting 1.8V failed, ret: %d\n", ret);
+		} else {
+			ret = pinctrl_select_state(tegra_host->pinctrl_sdmmc,
+						tegra_host->pinctrl_state_3v3);
+			if (ret < 0)
+				dev_err(mmc_dev(host->mmc),
+					"setting 3.3V failed, ret: %d\n", ret);
+		}
+	}
+
+	return ret;
+}
+
+static void tegra_sdhci_pad_autocalib(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	struct sdhci_tegra_autocal_offsets offsets =
+			tegra_host->autocal_offsets;
+	struct mmc_ios *ios = &host->mmc->ios;
+	bool card_clk_enabled;
+	u16 pdpu;
+	u32 reg;
+	int ret;
+
+	switch (ios->timing) {
+	case MMC_TIMING_UHS_SDR104:
+		pdpu = offsets.pull_down_sdr104 << 8 | offsets.pull_up_sdr104;
+		break;
+	case MMC_TIMING_MMC_HS400:
+		pdpu = offsets.pull_down_hs400 << 8 | offsets.pull_up_hs400;
+		break;
+	default:
+		if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_180)
+			pdpu = offsets.pull_down_1v8 << 8 | offsets.pull_up_1v8;
+		else
+			pdpu = offsets.pull_down_3v3 << 8 | offsets.pull_up_3v3;
+	}
+
+	/* Set initial offset before auto-calibration */
+	tegra_sdhci_set_pad_autocal_offset(host, pdpu);
+
+	card_clk_enabled = tegra_sdhci_configure_card_clk(host, false);
+
+	tegra_sdhci_configure_cal_pad(host, true);
+
+	reg = sdhci_readl(host, SDHCI_TEGRA_AUTO_CAL_CONFIG);
+	reg |= SDHCI_AUTO_CAL_ENABLE | SDHCI_AUTO_CAL_START;
+	sdhci_writel(host, reg, SDHCI_TEGRA_AUTO_CAL_CONFIG);
+
+	usleep_range(1, 2);
+	/* 10 ms timeout */
+	ret = readl_poll_timeout(host->ioaddr + SDHCI_TEGRA_AUTO_CAL_STATUS,
+				 reg, !(reg & SDHCI_TEGRA_AUTO_CAL_ACTIVE),
+				 1000, 10000);
+
+	tegra_sdhci_configure_cal_pad(host, false);
+
+	tegra_sdhci_configure_card_clk(host, card_clk_enabled);
+
+	if (ret) {
+		dev_err(mmc_dev(host->mmc), "Pad autocal timed out\n");
+
+		/* Disable automatic cal and use fixed Drive Strengths */
+		reg = sdhci_readl(host, SDHCI_TEGRA_AUTO_CAL_CONFIG);
+		reg &= ~SDHCI_AUTO_CAL_ENABLE;
+		sdhci_writel(host, reg, SDHCI_TEGRA_AUTO_CAL_CONFIG);
+
+		ret = tegra_sdhci_set_padctrl(host, ios->signal_voltage, false);
+		if (ret < 0)
+			dev_err(mmc_dev(host->mmc),
+				"Setting drive strengths failed: %d\n", ret);
+	}
+}
+
+static void tegra_sdhci_parse_pad_autocal_dt(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	struct sdhci_tegra_autocal_offsets *autocal =
+			&tegra_host->autocal_offsets;
+	int err;
+
+	err = device_property_read_u32(host->mmc->parent,
+			"nvidia,pad-autocal-pull-up-offset-3v3",
+			&autocal->pull_up_3v3);
+	if (err)
+		autocal->pull_up_3v3 = 0;
+
+	err = device_property_read_u32(host->mmc->parent,
+			"nvidia,pad-autocal-pull-down-offset-3v3",
+			&autocal->pull_down_3v3);
+	if (err)
+		autocal->pull_down_3v3 = 0;
+
+	err = device_property_read_u32(host->mmc->parent,
+			"nvidia,pad-autocal-pull-up-offset-1v8",
+			&autocal->pull_up_1v8);
+	if (err)
+		autocal->pull_up_1v8 = 0;
+
+	err = device_property_read_u32(host->mmc->parent,
+			"nvidia,pad-autocal-pull-down-offset-1v8",
+			&autocal->pull_down_1v8);
+	if (err)
+		autocal->pull_down_1v8 = 0;
+
+	err = device_property_read_u32(host->mmc->parent,
+			"nvidia,pad-autocal-pull-up-offset-sdr104",
+			&autocal->pull_up_sdr104);
+	if (err)
+		autocal->pull_up_sdr104 = autocal->pull_up_1v8;
+
+	err = device_property_read_u32(host->mmc->parent,
+			"nvidia,pad-autocal-pull-down-offset-sdr104",
+			&autocal->pull_down_sdr104);
+	if (err)
+		autocal->pull_down_sdr104 = autocal->pull_down_1v8;
+
+	err = device_property_read_u32(host->mmc->parent,
+			"nvidia,pad-autocal-pull-up-offset-hs400",
+			&autocal->pull_up_hs400);
+	if (err)
+		autocal->pull_up_hs400 = autocal->pull_up_1v8;
+
+	err = device_property_read_u32(host->mmc->parent,
+			"nvidia,pad-autocal-pull-down-offset-hs400",
+			&autocal->pull_down_hs400);
+	if (err)
+		autocal->pull_down_hs400 = autocal->pull_down_1v8;
+
+	/*
+	 * Different fail-safe drive strength values based on the signaling
+	 * voltage are applicable for SoCs supporting 3V3 and 1V8 pad controls.
+	 * So, avoid reading below device tree properties for SoCs that don't
+	 * have NVQUIRK_NEEDS_PAD_CONTROL.
+	 */
+	if (!(tegra_host->soc_data->nvquirks & NVQUIRK_NEEDS_PAD_CONTROL))
+		return;
+
+	err = device_property_read_u32(host->mmc->parent,
+			"nvidia,pad-autocal-pull-up-offset-3v3-timeout",
+			&autocal->pull_up_3v3_timeout);
+	if (err) {
+		if (!IS_ERR(tegra_host->pinctrl_state_3v3) &&
+			(tegra_host->pinctrl_state_3v3_drv == NULL))
+			pr_warn("%s: Missing autocal timeout 3v3-pad drvs\n",
+				mmc_hostname(host->mmc));
+		autocal->pull_up_3v3_timeout = 0;
+	}
+
+	err = device_property_read_u32(host->mmc->parent,
+			"nvidia,pad-autocal-pull-down-offset-3v3-timeout",
+			&autocal->pull_down_3v3_timeout);
+	if (err) {
+		if (!IS_ERR(tegra_host->pinctrl_state_3v3) &&
+			(tegra_host->pinctrl_state_3v3_drv == NULL))
+			pr_warn("%s: Missing autocal timeout 3v3-pad drvs\n",
+				mmc_hostname(host->mmc));
+		autocal->pull_down_3v3_timeout = 0;
+	}
+
+	err = device_property_read_u32(host->mmc->parent,
+			"nvidia,pad-autocal-pull-up-offset-1v8-timeout",
+			&autocal->pull_up_1v8_timeout);
+	if (err) {
+		if (!IS_ERR(tegra_host->pinctrl_state_1v8) &&
+			(tegra_host->pinctrl_state_1v8_drv == NULL))
+			pr_warn("%s: Missing autocal timeout 1v8-pad drvs\n",
+				mmc_hostname(host->mmc));
+		autocal->pull_up_1v8_timeout = 0;
+	}
+
+	err = device_property_read_u32(host->mmc->parent,
+			"nvidia,pad-autocal-pull-down-offset-1v8-timeout",
+			&autocal->pull_down_1v8_timeout);
+	if (err) {
+		if (!IS_ERR(tegra_host->pinctrl_state_1v8) &&
+			(tegra_host->pinctrl_state_1v8_drv == NULL))
+			pr_warn("%s: Missing autocal timeout 1v8-pad drvs\n",
+				mmc_hostname(host->mmc));
+		autocal->pull_down_1v8_timeout = 0;
+	}
+}
+
+static void tegra_sdhci_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	ktime_t since_calib = ktime_sub(ktime_get(), tegra_host->last_calib);
+
+	/* 100 ms calibration interval is specified in the TRM */
+	if (ktime_to_ms(since_calib) > 100) {
+		tegra_sdhci_pad_autocalib(host);
+		tegra_host->last_calib = ktime_get();
+	}
+
+	sdhci_request(mmc, mrq);
+}
+
+static void tegra_sdhci_parse_tap_and_trim(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	int err;
+
+	err = device_property_read_u32(host->mmc->parent, "nvidia,default-tap",
+				       &tegra_host->default_tap);
+	if (err)
+		tegra_host->default_tap = 0;
+
+	err = device_property_read_u32(host->mmc->parent, "nvidia,default-trim",
+				       &tegra_host->default_trim);
+	if (err)
+		tegra_host->default_trim = 0;
+
+	err = device_property_read_u32(host->mmc->parent, "nvidia,dqs-trim",
+				       &tegra_host->dqs_trim);
+	if (err)
+		tegra_host->dqs_trim = 0x11;
+}
+
+static void tegra_sdhci_parse_dt(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+
+	if (device_property_read_bool(host->mmc->parent, "supports-cqe"))
+		tegra_host->enable_hwcq = true;
+	else
+		tegra_host->enable_hwcq = false;
+
+	tegra_sdhci_parse_pad_autocal_dt(host);
+	tegra_sdhci_parse_tap_and_trim(host);
+}
+
+static void tegra_sdhci_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	unsigned long host_clk;
+
+	if (!clock)
+		return sdhci_set_clock(host, clock);
+
+	/*
+	 * In DDR50/52 modes the Tegra SDHCI controllers require the SDHCI
+	 * divider to be configured to divided the host clock by two. The SDHCI
+	 * clock divider is calculated as part of sdhci_set_clock() by
+	 * sdhci_calc_clk(). The divider is calculated from host->max_clk and
+	 * the requested clock rate.
+	 *
+	 * By setting the host->max_clk to clock * 2 the divider calculation
+	 * will always result in the correct value for DDR50/52 modes,
+	 * regardless of clock rate rounding, which may happen if the value
+	 * from clk_get_rate() is used.
+	 */
+	host_clk = tegra_host->ddr_signaling ? clock * 2 : clock;
+	clk_set_rate(pltfm_host->clk, host_clk);
+	tegra_host->curr_clk_rate = clk_get_rate(pltfm_host->clk);
+	if (tegra_host->ddr_signaling)
+		host->max_clk = host_clk;
+	else
+		host->max_clk = clk_get_rate(pltfm_host->clk);
+
+	sdhci_set_clock(host, clock);
+
+	if (tegra_host->pad_calib_required) {
+		tegra_sdhci_pad_autocalib(host);
+		tegra_host->pad_calib_required = false;
+	}
+}
+
+static void tegra_sdhci_hs400_enhanced_strobe(struct mmc_host *mmc,
+					      struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	u32 val;
+
+	val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_SYS_SW_CTRL);
+
+	if (ios->enhanced_strobe) {
+		val |= SDHCI_TEGRA_SYS_SW_CTRL_ENHANCED_STROBE;
+		/*
+		 * When CMD13 is sent from mmc_select_hs400es() after
+		 * switching to HS400ES mode, the bus is operating at
+		 * either MMC_HIGH_26_MAX_DTR or MMC_HIGH_52_MAX_DTR.
+		 * To meet Tegra SDHCI requirement at HS400ES mode, force SDHCI
+		 * interface clock to MMC_HS200_MAX_DTR (200 MHz) so that host
+		 * controller CAR clock and the interface clock are rate matched.
+		 */
+		tegra_sdhci_set_clock(host, MMC_HS200_MAX_DTR);
+	} else {
+		val &= ~SDHCI_TEGRA_SYS_SW_CTRL_ENHANCED_STROBE;
+	}
+
+	sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_SYS_SW_CTRL);
+}
+
+static unsigned int tegra_sdhci_get_max_clock(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	return clk_round_rate(pltfm_host->clk, UINT_MAX);
+}
+
+static void tegra_sdhci_set_dqs_trim(struct sdhci_host *host, u8 trim)
+{
+	u32 val;
+
+	val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_CAP_OVERRIDES);
+	val &= ~SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_MASK;
+	val |= trim << SDHCI_TEGRA_CAP_OVERRIDES_DQS_TRIM_SHIFT;
+	sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_CAP_OVERRIDES);
+}
+
+static void tegra_sdhci_hs400_dll_cal(struct sdhci_host *host)
+{
+	u32 reg;
+	int err;
+
+	reg = sdhci_readl(host, SDHCI_TEGRA_VENDOR_DLLCAL_CFG);
+	reg |= SDHCI_TEGRA_DLLCAL_CALIBRATE;
+	sdhci_writel(host, reg, SDHCI_TEGRA_VENDOR_DLLCAL_CFG);
+
+	/* 1 ms sleep, 5 ms timeout */
+	err = readl_poll_timeout(host->ioaddr + SDHCI_TEGRA_VENDOR_DLLCAL_STA,
+				 reg, !(reg & SDHCI_TEGRA_DLLCAL_STA_ACTIVE),
+				 1000, 5000);
+	if (err)
+		dev_err(mmc_dev(host->mmc),
+			"HS400 delay line calibration timed out\n");
+}
+
+static void tegra_sdhci_tap_correction(struct sdhci_host *host, u8 thd_up,
+				       u8 thd_low, u8 fixed_tap)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	u32 val, tun_status;
+	u8 word, bit, edge1, tap, window;
+	bool tap_result;
+	bool start_fail = false;
+	bool start_pass = false;
+	bool end_pass = false;
+	bool first_fail = false;
+	bool first_pass = false;
+	u8 start_pass_tap = 0;
+	u8 end_pass_tap = 0;
+	u8 first_fail_tap = 0;
+	u8 first_pass_tap = 0;
+	u8 total_tuning_words = host->tuning_loop_count / TUNING_WORD_BIT_SIZE;
+
+	/*
+	 * Read auto-tuned results and extract good valid passing window by
+	 * filtering out un-wanted bubble/partial/merged windows.
+	 */
+	for (word = 0; word < total_tuning_words; word++) {
+		val = sdhci_readl(host, SDHCI_VNDR_TUN_CTRL0_0);
+		val &= ~SDHCI_VNDR_TUN_CTRL0_TUN_WORD_SEL_MASK;
+		val |= word;
+		sdhci_writel(host, val, SDHCI_VNDR_TUN_CTRL0_0);
+		tun_status = sdhci_readl(host, SDHCI_TEGRA_VNDR_TUN_STATUS0);
+		bit = 0;
+		while (bit < TUNING_WORD_BIT_SIZE) {
+			tap = word * TUNING_WORD_BIT_SIZE + bit;
+			tap_result = tun_status & (1 << bit);
+			if (!tap_result && !start_fail) {
+				start_fail = true;
+				if (!first_fail) {
+					first_fail_tap = tap;
+					first_fail = true;
+				}
+
+			} else if (tap_result && start_fail && !start_pass) {
+				start_pass_tap = tap;
+				start_pass = true;
+				if (!first_pass) {
+					first_pass_tap = tap;
+					first_pass = true;
+				}
+
+			} else if (!tap_result && start_fail && start_pass &&
+				   !end_pass) {
+				end_pass_tap = tap - 1;
+				end_pass = true;
+			} else if (tap_result && start_pass && start_fail &&
+				   end_pass) {
+				window = end_pass_tap - start_pass_tap;
+				/* discard merged window and bubble window */
+				if (window >= thd_up || window < thd_low) {
+					start_pass_tap = tap;
+					end_pass = false;
+				} else {
+					/* set tap at middle of valid window */
+					tap = start_pass_tap + window / 2;
+					tegra_host->tuned_tap_delay = tap;
+					return;
+				}
+			}
+
+			bit++;
+		}
+	}
+
+	if (!first_fail) {
+		WARN(1, "no edge detected, continue with hw tuned delay.\n");
+	} else if (first_pass) {
+		/* set tap location at fixed tap relative to the first edge */
+		edge1 = first_fail_tap + (first_pass_tap - first_fail_tap) / 2;
+		if (edge1 - 1 > fixed_tap)
+			tegra_host->tuned_tap_delay = edge1 - fixed_tap;
+		else
+			tegra_host->tuned_tap_delay = edge1 + fixed_tap;
+	}
+}
+
+static void tegra_sdhci_post_tuning(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	const struct sdhci_tegra_soc_data *soc_data = tegra_host->soc_data;
+	u32 avg_tap_dly, val, min_tap_dly, max_tap_dly;
+	u8 fixed_tap, start_tap, end_tap, window_width;
+	u8 thdupper, thdlower;
+	u8 num_iter;
+	u32 clk_rate_mhz, period_ps, bestcase, worstcase;
+
+	/* retain HW tuned tap to use incase if no correction is needed */
+	val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_CLOCK_CTRL);
+	tegra_host->tuned_tap_delay = (val & SDHCI_CLOCK_CTRL_TAP_MASK) >>
+				      SDHCI_CLOCK_CTRL_TAP_SHIFT;
+	if (soc_data->min_tap_delay && soc_data->max_tap_delay) {
+		min_tap_dly = soc_data->min_tap_delay;
+		max_tap_dly = soc_data->max_tap_delay;
+		clk_rate_mhz = tegra_host->curr_clk_rate / USEC_PER_SEC;
+		period_ps = USEC_PER_SEC / clk_rate_mhz;
+		bestcase = period_ps / min_tap_dly;
+		worstcase = period_ps / max_tap_dly;
+		/*
+		 * Upper and Lower bound thresholds used to detect merged and
+		 * bubble windows
+		 */
+		thdupper = (2 * worstcase + bestcase) / 2;
+		thdlower = worstcase / 4;
+		/*
+		 * fixed tap is used when HW tuning result contains single edge
+		 * and tap is set at fixed tap delay relative to the first edge
+		 */
+		avg_tap_dly = (period_ps * 2) / (min_tap_dly + max_tap_dly);
+		fixed_tap = avg_tap_dly / 2;
+
+		val = sdhci_readl(host, SDHCI_TEGRA_VNDR_TUN_STATUS1);
+		start_tap = val & SDHCI_TEGRA_VNDR_TUN_STATUS1_TAP_MASK;
+		end_tap = (val >> SDHCI_TEGRA_VNDR_TUN_STATUS1_END_TAP_SHIFT) &
+			  SDHCI_TEGRA_VNDR_TUN_STATUS1_TAP_MASK;
+		window_width = end_tap - start_tap;
+		num_iter = host->tuning_loop_count;
+		/*
+		 * partial window includes edges of the tuning range.
+		 * merged window includes more taps so window width is higher
+		 * than upper threshold.
+		 */
+		if (start_tap == 0 || (end_tap == (num_iter - 1)) ||
+		    (end_tap == num_iter - 2) || window_width >= thdupper) {
+			pr_debug("%s: Apply tuning correction\n",
+				 mmc_hostname(host->mmc));
+			tegra_sdhci_tap_correction(host, thdupper, thdlower,
+						   fixed_tap);
+		}
+	}
+
+	tegra_sdhci_set_tap(host, tegra_host->tuned_tap_delay);
+}
+
+static int tegra_sdhci_execute_hw_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	int err;
+
+	err = sdhci_execute_tuning(mmc, opcode);
+	if (!err && !host->tuning_err)
+		tegra_sdhci_post_tuning(host);
+
+	return err;
+}
+
+static void tegra_sdhci_set_uhs_signaling(struct sdhci_host *host,
+					  unsigned timing)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	bool set_default_tap = false;
+	bool set_dqs_trim = false;
+	bool do_hs400_dll_cal = false;
+	u8 iter = TRIES_256;
+	u32 val;
+
+	tegra_host->ddr_signaling = false;
+	switch (timing) {
+	case MMC_TIMING_UHS_SDR50:
+		break;
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_MMC_HS200:
+		/* Don't set default tap on tunable modes. */
+		iter = TRIES_128;
+		break;
+	case MMC_TIMING_MMC_HS400:
+		set_dqs_trim = true;
+		do_hs400_dll_cal = true;
+		iter = TRIES_128;
+		break;
+	case MMC_TIMING_MMC_DDR52:
+	case MMC_TIMING_UHS_DDR50:
+		tegra_host->ddr_signaling = true;
+		set_default_tap = true;
+		break;
+	default:
+		set_default_tap = true;
+		break;
+	}
+
+	val = sdhci_readl(host, SDHCI_VNDR_TUN_CTRL0_0);
+	val &= ~(SDHCI_VNDR_TUN_CTRL0_TUN_ITER_MASK |
+		 SDHCI_VNDR_TUN_CTRL0_START_TAP_VAL_MASK |
+		 SDHCI_VNDR_TUN_CTRL0_MUL_M_MASK);
+	val |= (iter << SDHCI_VNDR_TUN_CTRL0_TUN_ITER_SHIFT |
+		0 << SDHCI_VNDR_TUN_CTRL0_START_TAP_VAL_SHIFT |
+		1 << SDHCI_VNDR_TUN_CTRL0_MUL_M_SHIFT);
+	sdhci_writel(host, val, SDHCI_VNDR_TUN_CTRL0_0);
+	sdhci_writel(host, 0, SDHCI_TEGRA_VNDR_TUN_CTRL1_0);
+
+	host->tuning_loop_count = (iter == TRIES_128) ? 128 : 256;
+
+	sdhci_set_uhs_signaling(host, timing);
+
+	tegra_sdhci_pad_autocalib(host);
+
+	if (tegra_host->tuned_tap_delay && !set_default_tap)
+		tegra_sdhci_set_tap(host, tegra_host->tuned_tap_delay);
+	else
+		tegra_sdhci_set_tap(host, tegra_host->default_tap);
+
+	if (set_dqs_trim)
+		tegra_sdhci_set_dqs_trim(host, tegra_host->dqs_trim);
+
+	if (do_hs400_dll_cal)
+		tegra_sdhci_hs400_dll_cal(host);
+}
+
+static int tegra_sdhci_execute_tuning(struct sdhci_host *host, u32 opcode)
+{
+	unsigned int min, max;
+
+	/*
+	 * Start search for minimum tap value at 10, as smaller values are
+	 * may wrongly be reported as working but fail at higher speeds,
+	 * according to the TRM.
+	 */
+	min = 10;
+	while (min < 255) {
+		tegra_sdhci_set_tap(host, min);
+		if (!mmc_send_tuning(host->mmc, opcode, NULL))
+			break;
+		min++;
+	}
+
+	/* Find the maximum tap value that still passes. */
+	max = min + 1;
+	while (max < 255) {
+		tegra_sdhci_set_tap(host, max);
+		if (mmc_send_tuning(host->mmc, opcode, NULL)) {
+			max--;
+			break;
+		}
+		max++;
+	}
+
+	/* The TRM states the ideal tap value is at 75% in the passing range. */
+	tegra_sdhci_set_tap(host, min + ((max - min) * 3 / 4));
+
+	return mmc_send_tuning(host->mmc, opcode, NULL);
+}
+
+static int sdhci_tegra_start_signal_voltage_switch(struct mmc_host *mmc,
+						   struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	int ret = 0;
+
+	if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_330) {
+		ret = tegra_sdhci_set_padctrl(host, ios->signal_voltage, true);
+		if (ret < 0)
+			return ret;
+		ret = sdhci_start_signal_voltage_switch(mmc, ios);
+	} else if (ios->signal_voltage == MMC_SIGNAL_VOLTAGE_180) {
+		ret = sdhci_start_signal_voltage_switch(mmc, ios);
+		if (ret < 0)
+			return ret;
+		ret = tegra_sdhci_set_padctrl(host, ios->signal_voltage, true);
+	}
+
+	if (tegra_host->pad_calib_required)
+		tegra_sdhci_pad_autocalib(host);
+
+	return ret;
+}
+
+static int tegra_sdhci_init_pinctrl_info(struct device *dev,
+					 struct sdhci_tegra *tegra_host)
+{
+	tegra_host->pinctrl_sdmmc = devm_pinctrl_get(dev);
+	if (IS_ERR(tegra_host->pinctrl_sdmmc)) {
+		dev_dbg(dev, "No pinctrl info, err: %ld\n",
+			PTR_ERR(tegra_host->pinctrl_sdmmc));
+		return -1;
+	}
+
+	tegra_host->pinctrl_state_1v8_drv = pinctrl_lookup_state(
+				tegra_host->pinctrl_sdmmc, "sdmmc-1v8-drv");
+	if (IS_ERR(tegra_host->pinctrl_state_1v8_drv)) {
+		if (PTR_ERR(tegra_host->pinctrl_state_1v8_drv) == -ENODEV)
+			tegra_host->pinctrl_state_1v8_drv = NULL;
+	}
+
+	tegra_host->pinctrl_state_3v3_drv = pinctrl_lookup_state(
+				tegra_host->pinctrl_sdmmc, "sdmmc-3v3-drv");
+	if (IS_ERR(tegra_host->pinctrl_state_3v3_drv)) {
+		if (PTR_ERR(tegra_host->pinctrl_state_3v3_drv) == -ENODEV)
+			tegra_host->pinctrl_state_3v3_drv = NULL;
+	}
+
+	tegra_host->pinctrl_state_3v3 =
+		pinctrl_lookup_state(tegra_host->pinctrl_sdmmc, "sdmmc-3v3");
+	if (IS_ERR(tegra_host->pinctrl_state_3v3)) {
+		dev_warn(dev, "Missing 3.3V pad state, err: %ld\n",
+			 PTR_ERR(tegra_host->pinctrl_state_3v3));
+		return -1;
+	}
+
+	tegra_host->pinctrl_state_1v8 =
+		pinctrl_lookup_state(tegra_host->pinctrl_sdmmc, "sdmmc-1v8");
+	if (IS_ERR(tegra_host->pinctrl_state_1v8)) {
+		dev_warn(dev, "Missing 1.8V pad state, err: %ld\n",
+			 PTR_ERR(tegra_host->pinctrl_state_1v8));
+		return -1;
+	}
+
+	tegra_host->pad_control_available = true;
+
+	return 0;
+}
+
+static void tegra_sdhci_voltage_switch(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	const struct sdhci_tegra_soc_data *soc_data = tegra_host->soc_data;
+
+	if (soc_data->nvquirks & NVQUIRK_HAS_PADCALIB)
+		tegra_host->pad_calib_required = true;
+}
+
+static void tegra_cqhci_writel(struct cqhci_host *cq_host, u32 val, int reg)
+{
+	struct mmc_host *mmc = cq_host->mmc;
+	struct sdhci_host *host = mmc_priv(mmc);
+	u8 ctrl;
+	ktime_t timeout;
+	bool timed_out;
+
+	/*
+	 * During CQE resume/unhalt, CQHCI driver unhalts CQE prior to
+	 * cqhci_host_ops enable where SDHCI DMA and BLOCK_SIZE registers need
+	 * to be re-configured.
+	 * Tegra CQHCI/SDHCI prevents write access to block size register when
+	 * CQE is unhalted. So handling CQE resume sequence here to configure
+	 * SDHCI block registers prior to exiting CQE halt state.
+	 */
+	if (reg == CQHCI_CTL && !(val & CQHCI_HALT) &&
+	    cqhci_readl(cq_host, CQHCI_CTL) & CQHCI_HALT) {
+		sdhci_writew(host, SDHCI_TEGRA_CQE_TRNS_MODE, SDHCI_TRANSFER_MODE);
+		sdhci_cqe_enable(mmc);
+		writel(val, cq_host->mmio + reg);
+		timeout = ktime_add_us(ktime_get(), 50);
+		while (1) {
+			timed_out = ktime_compare(ktime_get(), timeout) > 0;
+			ctrl = cqhci_readl(cq_host, CQHCI_CTL);
+			if (!(ctrl & CQHCI_HALT) || timed_out)
+				break;
+		}
+		/*
+		 * CQE usually resumes very quick, but incase if Tegra CQE
+		 * doesn't resume retry unhalt.
+		 */
+		if (timed_out)
+			writel(val, cq_host->mmio + reg);
+	} else {
+		writel(val, cq_host->mmio + reg);
+	}
+}
+
+static void sdhci_tegra_update_dcmd_desc(struct mmc_host *mmc,
+					 struct mmc_request *mrq, u64 *data)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(mmc_priv(mmc));
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	const struct sdhci_tegra_soc_data *soc_data = tegra_host->soc_data;
+
+	if (soc_data->nvquirks & NVQUIRK_CQHCI_DCMD_R1B_CMD_TIMING &&
+	    mrq->cmd->flags & MMC_RSP_R1B)
+		*data |= CQHCI_CMD_TIMING(1);
+}
+
+static void sdhci_tegra_cqe_enable(struct mmc_host *mmc)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	struct sdhci_host *host = mmc_priv(mmc);
+	u32 val;
+
+	/*
+	 * Tegra CQHCI/SDMMC design prevents write access to sdhci block size
+	 * register when CQE is enabled and unhalted.
+	 * CQHCI driver enables CQE prior to activation, so disable CQE before
+	 * programming block size in sdhci controller and enable it back.
+	 */
+	if (!cq_host->activated) {
+		val = cqhci_readl(cq_host, CQHCI_CFG);
+		if (val & CQHCI_ENABLE)
+			cqhci_writel(cq_host, (val & ~CQHCI_ENABLE),
+				     CQHCI_CFG);
+		sdhci_writew(host, SDHCI_TEGRA_CQE_TRNS_MODE, SDHCI_TRANSFER_MODE);
+		sdhci_cqe_enable(mmc);
+		if (val & CQHCI_ENABLE)
+			cqhci_writel(cq_host, val, CQHCI_CFG);
+	}
+
+	/*
+	 * CMD CRC errors are seen sometimes with some eMMC devices when status
+	 * command is sent during transfer of last data block which is the
+	 * default case as send status command block counter (CBC) is 1.
+	 * Recommended fix to set CBC to 0 allowing send status command only
+	 * when data lines are idle.
+	 */
+	val = cqhci_readl(cq_host, CQHCI_SSC1);
+	val &= ~CQHCI_SSC1_CBC_MASK;
+	cqhci_writel(cq_host, val, CQHCI_SSC1);
+}
+
+static void sdhci_tegra_dumpregs(struct mmc_host *mmc)
+{
+	sdhci_dumpregs(mmc_priv(mmc));
+}
+
+static u32 sdhci_tegra_cqhci_irq(struct sdhci_host *host, u32 intmask)
+{
+	int cmd_error = 0;
+	int data_error = 0;
+
+	if (!sdhci_cqe_irq(host, intmask, &cmd_error, &data_error))
+		return intmask;
+
+	cqhci_irq(host->mmc, intmask, cmd_error, data_error);
+
+	return 0;
+}
+
+static void tegra_sdhci_set_timeout(struct sdhci_host *host,
+				    struct mmc_command *cmd)
+{
+	u32 val;
+
+	/*
+	 * HW busy detection timeout is based on programmed data timeout
+	 * counter and maximum supported timeout is 11s which may not be
+	 * enough for long operations like cache flush, sleep awake, erase.
+	 *
+	 * ERASE_TIMEOUT_LIMIT bit of VENDOR_MISC_CTRL register allows
+	 * host controller to wait for busy state until the card is busy
+	 * without HW timeout.
+	 *
+	 * So, use infinite busy wait mode for operations that may take
+	 * more than maximum HW busy timeout of 11s otherwise use finite
+	 * busy wait mode.
+	 */
+	val = sdhci_readl(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+	if (cmd && cmd->busy_timeout >= 11 * MSEC_PER_SEC)
+		val |= SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
+	else
+		val &= ~SDHCI_MISC_CTRL_ERASE_TIMEOUT_LIMIT;
+	sdhci_writel(host, val, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+
+	__sdhci_set_timeout(host, cmd);
+}
+
+static void sdhci_tegra_cqe_pre_enable(struct mmc_host *mmc)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	u32 reg;
+
+	reg = cqhci_readl(cq_host, CQHCI_CFG);
+	reg |= CQHCI_ENABLE;
+	cqhci_writel(cq_host, reg, CQHCI_CFG);
+}
+
+static void sdhci_tegra_cqe_post_disable(struct mmc_host *mmc)
+{
+	struct cqhci_host *cq_host = mmc->cqe_private;
+	struct sdhci_host *host = mmc_priv(mmc);
+	u32 reg;
+
+	reg = cqhci_readl(cq_host, CQHCI_CFG);
+	reg &= ~CQHCI_ENABLE;
+	cqhci_writel(cq_host, reg, CQHCI_CFG);
+	sdhci_writew(host, 0x0, SDHCI_TRANSFER_MODE);
+}
+
+static const struct cqhci_host_ops sdhci_tegra_cqhci_ops = {
+	.write_l    = tegra_cqhci_writel,
+	.enable	= sdhci_tegra_cqe_enable,
+	.disable = sdhci_cqe_disable,
+	.dumpregs = sdhci_tegra_dumpregs,
+	.update_dcmd_desc = sdhci_tegra_update_dcmd_desc,
+	.pre_enable = sdhci_tegra_cqe_pre_enable,
+	.post_disable = sdhci_tegra_cqe_post_disable,
+};
+
+static int tegra_sdhci_set_dma_mask(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *platform = sdhci_priv(host);
+	struct sdhci_tegra *tegra = sdhci_pltfm_priv(platform);
+	const struct sdhci_tegra_soc_data *soc = tegra->soc_data;
+	struct device *dev = mmc_dev(host->mmc);
+
+	if (soc->dma_mask)
+		return dma_set_mask_and_coherent(dev, soc->dma_mask);
+
+	return 0;
+}
+
+static const struct sdhci_ops tegra_sdhci_ops = {
+	.get_ro     = tegra_sdhci_get_ro,
+	.read_w     = tegra_sdhci_readw,
+	.write_l    = tegra_sdhci_writel,
+	.set_clock  = tegra_sdhci_set_clock,
+	.set_dma_mask = tegra_sdhci_set_dma_mask,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset      = tegra_sdhci_reset,
+	.platform_execute_tuning = tegra_sdhci_execute_tuning,
+	.set_uhs_signaling = tegra_sdhci_set_uhs_signaling,
+	.voltage_switch = tegra_sdhci_voltage_switch,
+	.get_max_clock = tegra_sdhci_get_max_clock,
+};
+
+static const struct sdhci_pltfm_data sdhci_tegra20_pdata = {
+	.quirks = SDHCI_QUIRK_BROKEN_TIMEOUT_VAL |
+		  SDHCI_QUIRK_SINGLE_POWER_WRITE |
+		  SDHCI_QUIRK_NO_HISPD_BIT |
+		  SDHCI_QUIRK_BROKEN_ADMA_ZEROLEN_DESC |
+		  SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+	.ops  = &tegra_sdhci_ops,
+};
+
+static const struct sdhci_tegra_soc_data soc_data_tegra20 = {
+	.pdata = &sdhci_tegra20_pdata,
+	.dma_mask = DMA_BIT_MASK(32),
+	.nvquirks = NVQUIRK_FORCE_SDHCI_SPEC_200 |
+		    NVQUIRK_ENABLE_BLOCK_GAP_DET,
+};
+
+static const struct sdhci_pltfm_data sdhci_tegra30_pdata = {
+	.quirks = SDHCI_QUIRK_BROKEN_TIMEOUT_VAL |
+		  SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK |
+		  SDHCI_QUIRK_SINGLE_POWER_WRITE |
+		  SDHCI_QUIRK_NO_HISPD_BIT |
+		  SDHCI_QUIRK_BROKEN_ADMA_ZEROLEN_DESC |
+		  SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+		   SDHCI_QUIRK2_BROKEN_HS200 |
+		   /*
+		    * Auto-CMD23 leads to "Got command interrupt 0x00010000 even
+		    * though no command operation was in progress."
+		    *
+		    * The exact reason is unknown, as the same hardware seems
+		    * to support Auto CMD23 on a downstream 3.1 kernel.
+		    */
+		   SDHCI_QUIRK2_ACMD23_BROKEN,
+	.ops  = &tegra_sdhci_ops,
+};
+
+static const struct sdhci_tegra_soc_data soc_data_tegra30 = {
+	.pdata = &sdhci_tegra30_pdata,
+	.dma_mask = DMA_BIT_MASK(32),
+	.nvquirks = NVQUIRK_ENABLE_SDHCI_SPEC_300 |
+		    NVQUIRK_ENABLE_SDR50 |
+		    NVQUIRK_ENABLE_SDR104 |
+		    NVQUIRK_HAS_PADCALIB,
+};
+
+static const struct sdhci_ops tegra114_sdhci_ops = {
+	.get_ro     = tegra_sdhci_get_ro,
+	.read_w     = tegra_sdhci_readw,
+	.write_w    = tegra_sdhci_writew,
+	.write_l    = tegra_sdhci_writel,
+	.set_clock  = tegra_sdhci_set_clock,
+	.set_dma_mask = tegra_sdhci_set_dma_mask,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset      = tegra_sdhci_reset,
+	.platform_execute_tuning = tegra_sdhci_execute_tuning,
+	.set_uhs_signaling = tegra_sdhci_set_uhs_signaling,
+	.voltage_switch = tegra_sdhci_voltage_switch,
+	.get_max_clock = tegra_sdhci_get_max_clock,
+};
+
+static const struct sdhci_pltfm_data sdhci_tegra114_pdata = {
+	.quirks = SDHCI_QUIRK_BROKEN_TIMEOUT_VAL |
+		  SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK |
+		  SDHCI_QUIRK_SINGLE_POWER_WRITE |
+		  SDHCI_QUIRK_NO_HISPD_BIT |
+		  SDHCI_QUIRK_BROKEN_ADMA_ZEROLEN_DESC |
+		  SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+	.ops  = &tegra114_sdhci_ops,
+};
+
+static const struct sdhci_tegra_soc_data soc_data_tegra114 = {
+	.pdata = &sdhci_tegra114_pdata,
+	.dma_mask = DMA_BIT_MASK(32),
+};
+
+static const struct sdhci_pltfm_data sdhci_tegra124_pdata = {
+	.quirks = SDHCI_QUIRK_BROKEN_TIMEOUT_VAL |
+		  SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK |
+		  SDHCI_QUIRK_SINGLE_POWER_WRITE |
+		  SDHCI_QUIRK_NO_HISPD_BIT |
+		  SDHCI_QUIRK_BROKEN_ADMA_ZEROLEN_DESC |
+		  SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+	.ops  = &tegra114_sdhci_ops,
+};
+
+static const struct sdhci_tegra_soc_data soc_data_tegra124 = {
+	.pdata = &sdhci_tegra124_pdata,
+	.dma_mask = DMA_BIT_MASK(34),
+};
+
+static const struct sdhci_ops tegra210_sdhci_ops = {
+	.get_ro     = tegra_sdhci_get_ro,
+	.read_w     = tegra_sdhci_readw,
+	.write_w    = tegra210_sdhci_writew,
+	.write_l    = tegra_sdhci_writel,
+	.set_clock  = tegra_sdhci_set_clock,
+	.set_dma_mask = tegra_sdhci_set_dma_mask,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset      = tegra_sdhci_reset,
+	.set_uhs_signaling = tegra_sdhci_set_uhs_signaling,
+	.voltage_switch = tegra_sdhci_voltage_switch,
+	.get_max_clock = tegra_sdhci_get_max_clock,
+	.set_timeout = tegra_sdhci_set_timeout,
+};
+
+static const struct sdhci_pltfm_data sdhci_tegra210_pdata = {
+	.quirks = SDHCI_QUIRK_BROKEN_TIMEOUT_VAL |
+		  SDHCI_QUIRK_SINGLE_POWER_WRITE |
+		  SDHCI_QUIRK_NO_HISPD_BIT |
+		  SDHCI_QUIRK_BROKEN_ADMA_ZEROLEN_DESC |
+		  SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+	.ops  = &tegra210_sdhci_ops,
+};
+
+static const struct sdhci_tegra_soc_data soc_data_tegra210 = {
+	.pdata = &sdhci_tegra210_pdata,
+	.dma_mask = DMA_BIT_MASK(34),
+	.nvquirks = NVQUIRK_NEEDS_PAD_CONTROL |
+		    NVQUIRK_HAS_PADCALIB |
+		    NVQUIRK_DIS_CARD_CLK_CONFIG_TAP |
+		    NVQUIRK_ENABLE_SDR50 |
+		    NVQUIRK_ENABLE_SDR104 |
+		    NVQUIRK_HAS_TMCLK,
+	.min_tap_delay = 106,
+	.max_tap_delay = 185,
+};
+
+static const struct sdhci_ops tegra186_sdhci_ops = {
+	.get_ro     = tegra_sdhci_get_ro,
+	.read_w     = tegra_sdhci_readw,
+	.write_l    = tegra_sdhci_writel,
+	.set_clock  = tegra_sdhci_set_clock,
+	.set_dma_mask = tegra_sdhci_set_dma_mask,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset      = tegra_sdhci_reset,
+	.set_uhs_signaling = tegra_sdhci_set_uhs_signaling,
+	.voltage_switch = tegra_sdhci_voltage_switch,
+	.get_max_clock = tegra_sdhci_get_max_clock,
+	.irq = sdhci_tegra_cqhci_irq,
+	.set_timeout = tegra_sdhci_set_timeout,
+};
+
+static const struct sdhci_pltfm_data sdhci_tegra186_pdata = {
+	.quirks = SDHCI_QUIRK_BROKEN_TIMEOUT_VAL |
+		  SDHCI_QUIRK_SINGLE_POWER_WRITE |
+		  SDHCI_QUIRK_NO_HISPD_BIT |
+		  SDHCI_QUIRK_BROKEN_ADMA_ZEROLEN_DESC |
+		  SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+	.ops  = &tegra186_sdhci_ops,
+};
+
+static const struct sdhci_tegra_soc_data soc_data_tegra186 = {
+	.pdata = &sdhci_tegra186_pdata,
+	.dma_mask = DMA_BIT_MASK(40),
+	.nvquirks = NVQUIRK_NEEDS_PAD_CONTROL |
+		    NVQUIRK_HAS_PADCALIB |
+		    NVQUIRK_DIS_CARD_CLK_CONFIG_TAP |
+		    NVQUIRK_ENABLE_SDR50 |
+		    NVQUIRK_ENABLE_SDR104 |
+		    NVQUIRK_HAS_TMCLK |
+		    NVQUIRK_CQHCI_DCMD_R1B_CMD_TIMING,
+	.min_tap_delay = 84,
+	.max_tap_delay = 136,
+};
+
+static const struct sdhci_tegra_soc_data soc_data_tegra194 = {
+	.pdata = &sdhci_tegra186_pdata,
+	.dma_mask = DMA_BIT_MASK(39),
+	.nvquirks = NVQUIRK_NEEDS_PAD_CONTROL |
+		    NVQUIRK_HAS_PADCALIB |
+		    NVQUIRK_DIS_CARD_CLK_CONFIG_TAP |
+		    NVQUIRK_ENABLE_SDR50 |
+		    NVQUIRK_ENABLE_SDR104 |
+		    NVQUIRK_HAS_TMCLK,
+	.min_tap_delay = 96,
+	.max_tap_delay = 139,
+};
+
+static const struct of_device_id sdhci_tegra_dt_match[] = {
+	{ .compatible = "nvidia,tegra194-sdhci", .data = &soc_data_tegra194 },
+	{ .compatible = "nvidia,tegra186-sdhci", .data = &soc_data_tegra186 },
+	{ .compatible = "nvidia,tegra210-sdhci", .data = &soc_data_tegra210 },
+	{ .compatible = "nvidia,tegra124-sdhci", .data = &soc_data_tegra124 },
+	{ .compatible = "nvidia,tegra114-sdhci", .data = &soc_data_tegra114 },
+	{ .compatible = "nvidia,tegra30-sdhci", .data = &soc_data_tegra30 },
+	{ .compatible = "nvidia,tegra20-sdhci", .data = &soc_data_tegra20 },
+	{}
+};
+MODULE_DEVICE_TABLE(of, sdhci_tegra_dt_match);
+
+static int sdhci_tegra_add_host(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+	struct cqhci_host *cq_host;
+	bool dma64;
+	int ret;
+
+	if (!tegra_host->enable_hwcq)
+		return sdhci_add_host(host);
+
+	sdhci_enable_v4_mode(host);
+
+	ret = sdhci_setup_host(host);
+	if (ret)
+		return ret;
+
+	host->mmc->caps2 |= MMC_CAP2_CQE | MMC_CAP2_CQE_DCMD;
+
+	cq_host = devm_kzalloc(host->mmc->parent,
+				sizeof(*cq_host), GFP_KERNEL);
+	if (!cq_host) {
+		ret = -ENOMEM;
+		goto cleanup;
+	}
+
+	cq_host->mmio = host->ioaddr + SDHCI_TEGRA_CQE_BASE_ADDR;
+	cq_host->ops = &sdhci_tegra_cqhci_ops;
+
+	dma64 = host->flags & SDHCI_USE_64_BIT_DMA;
+	if (dma64)
+		cq_host->caps |= CQHCI_TASK_DESC_SZ_128;
+
+	ret = cqhci_init(cq_host, host->mmc, dma64);
+	if (ret)
+		goto cleanup;
+
+	ret = __sdhci_add_host(host);
+	if (ret)
+		goto cleanup;
+
+	return 0;
+
+cleanup:
+	sdhci_cleanup_host(host);
+	return ret;
+}
+
+static int sdhci_tegra_probe(struct platform_device *pdev)
+{
+	const struct of_device_id *match;
+	const struct sdhci_tegra_soc_data *soc_data;
+	struct sdhci_host *host;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_tegra *tegra_host;
+	struct clk *clk;
+	int rc;
+
+	match = of_match_device(sdhci_tegra_dt_match, &pdev->dev);
+	if (!match)
+		return -EINVAL;
+	soc_data = match->data;
+
+	host = sdhci_pltfm_init(pdev, soc_data->pdata, sizeof(*tegra_host));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+	pltfm_host = sdhci_priv(host);
+
+	tegra_host = sdhci_pltfm_priv(pltfm_host);
+	tegra_host->ddr_signaling = false;
+	tegra_host->pad_calib_required = false;
+	tegra_host->pad_control_available = false;
+	tegra_host->soc_data = soc_data;
+
+	if (soc_data->nvquirks & NVQUIRK_NEEDS_PAD_CONTROL) {
+		rc = tegra_sdhci_init_pinctrl_info(&pdev->dev, tegra_host);
+		if (rc == 0)
+			host->mmc_host_ops.start_signal_voltage_switch =
+				sdhci_tegra_start_signal_voltage_switch;
+	}
+
+	/* Hook to periodically rerun pad calibration */
+	if (soc_data->nvquirks & NVQUIRK_HAS_PADCALIB)
+		host->mmc_host_ops.request = tegra_sdhci_request;
+
+	host->mmc_host_ops.hs400_enhanced_strobe =
+			tegra_sdhci_hs400_enhanced_strobe;
+
+	if (!host->ops->platform_execute_tuning)
+		host->mmc_host_ops.execute_tuning =
+				tegra_sdhci_execute_hw_tuning;
+
+	rc = mmc_of_parse(host->mmc);
+	if (rc)
+		goto err_parse_dt;
+
+	if (tegra_host->soc_data->nvquirks & NVQUIRK_ENABLE_DDR50)
+		host->mmc->caps |= MMC_CAP_1_8V_DDR;
+
+	/* HW busy detection is supported, but R1B responses are required. */
+	host->mmc->caps |= MMC_CAP_WAIT_WHILE_BUSY | MMC_CAP_NEED_RSP_BUSY;
+
+	tegra_sdhci_parse_dt(host);
+
+	tegra_host->power_gpio = devm_gpiod_get_optional(&pdev->dev, "power",
+							 GPIOD_OUT_HIGH);
+	if (IS_ERR(tegra_host->power_gpio)) {
+		rc = PTR_ERR(tegra_host->power_gpio);
+		goto err_power_req;
+	}
+
+	/*
+	 * Tegra210 has a separate SDMMC_LEGACY_TM clock used for host
+	 * timeout clock and SW can choose TMCLK or SDCLK for hardware
+	 * data timeout through the bit USE_TMCLK_FOR_DATA_TIMEOUT of
+	 * the register SDHCI_TEGRA_VENDOR_SYS_SW_CTRL.
+	 *
+	 * USE_TMCLK_FOR_DATA_TIMEOUT bit default is set to 1 and SDMMC uses
+	 * 12Mhz TMCLK which is advertised in host capability register.
+	 * With TMCLK of 12Mhz provides maximum data timeout period that can
+	 * be achieved is 11s better than using SDCLK for data timeout.
+	 *
+	 * So, TMCLK is set to 12Mhz and kept enabled all the time on SoC's
+	 * supporting separate TMCLK.
+	 */
+
+	if (soc_data->nvquirks & NVQUIRK_HAS_TMCLK) {
+		clk = devm_clk_get(&pdev->dev, "tmclk");
+		if (IS_ERR(clk)) {
+			rc = PTR_ERR(clk);
+			if (rc == -EPROBE_DEFER)
+				goto err_power_req;
+
+			dev_warn(&pdev->dev, "failed to get tmclk: %d\n", rc);
+			clk = NULL;
+		}
+
+		clk_set_rate(clk, 12000000);
+		rc = clk_prepare_enable(clk);
+		if (rc) {
+			dev_err(&pdev->dev,
+				"failed to enable tmclk: %d\n", rc);
+			goto err_power_req;
+		}
+
+		tegra_host->tmclk = clk;
+	}
+
+	clk = devm_clk_get(mmc_dev(host->mmc), NULL);
+	if (IS_ERR(clk)) {
+		rc = dev_err_probe(&pdev->dev, PTR_ERR(clk),
+				   "failed to get clock\n");
+		goto err_clk_get;
+	}
+	clk_prepare_enable(clk);
+	pltfm_host->clk = clk;
+
+	tegra_host->rst = devm_reset_control_get_exclusive(&pdev->dev,
+							   "sdhci");
+	if (IS_ERR(tegra_host->rst)) {
+		rc = PTR_ERR(tegra_host->rst);
+		dev_err(&pdev->dev, "failed to get reset control: %d\n", rc);
+		goto err_rst_get;
+	}
+
+	rc = reset_control_assert(tegra_host->rst);
+	if (rc)
+		goto err_rst_get;
+
+	usleep_range(2000, 4000);
+
+	rc = reset_control_deassert(tegra_host->rst);
+	if (rc)
+		goto err_rst_get;
+
+	usleep_range(2000, 4000);
+
+	rc = sdhci_tegra_add_host(host);
+	if (rc)
+		goto err_add_host;
+
+	return 0;
+
+err_add_host:
+	reset_control_assert(tegra_host->rst);
+err_rst_get:
+	clk_disable_unprepare(pltfm_host->clk);
+err_clk_get:
+	clk_disable_unprepare(tegra_host->tmclk);
+err_power_req:
+err_parse_dt:
+	sdhci_pltfm_free(pdev);
+	return rc;
+}
+
+static int sdhci_tegra_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_tegra *tegra_host = sdhci_pltfm_priv(pltfm_host);
+
+	sdhci_remove_host(host, 0);
+
+	reset_control_assert(tegra_host->rst);
+	usleep_range(2000, 4000);
+	clk_disable_unprepare(pltfm_host->clk);
+	clk_disable_unprepare(tegra_host->tmclk);
+
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int __maybe_unused sdhci_tegra_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	int ret;
+
+	if (host->mmc->caps2 & MMC_CAP2_CQE) {
+		ret = cqhci_suspend(host->mmc);
+		if (ret)
+			return ret;
+	}
+
+	ret = sdhci_suspend_host(host);
+	if (ret) {
+		cqhci_resume(host->mmc);
+		return ret;
+	}
+
+	clk_disable_unprepare(pltfm_host->clk);
+	return 0;
+}
+
+static int __maybe_unused sdhci_tegra_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	int ret;
+
+	ret = clk_prepare_enable(pltfm_host->clk);
+	if (ret)
+		return ret;
+
+	ret = sdhci_resume_host(host);
+	if (ret)
+		goto disable_clk;
+
+	if (host->mmc->caps2 & MMC_CAP2_CQE) {
+		ret = cqhci_resume(host->mmc);
+		if (ret)
+			goto suspend_host;
+	}
+
+	return 0;
+
+suspend_host:
+	sdhci_suspend_host(host);
+disable_clk:
+	clk_disable_unprepare(pltfm_host->clk);
+	return ret;
+}
+#endif
+
+static SIMPLE_DEV_PM_OPS(sdhci_tegra_dev_pm_ops, sdhci_tegra_suspend,
+			 sdhci_tegra_resume);
+
+static struct platform_driver sdhci_tegra_driver = {
+	.driver		= {
+		.name	= "sdhci-tegra",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = sdhci_tegra_dt_match,
+		.pm	= &sdhci_tegra_dev_pm_ops,
+	},
+	.probe		= sdhci_tegra_probe,
+	.remove		= sdhci_tegra_remove,
+};
+
+module_platform_driver(sdhci_tegra_driver);
+
+MODULE_DESCRIPTION("SDHCI driver for Tegra");
+MODULE_AUTHOR("Google, Inc.");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-xenon-phy.c b/drivers/mmc/host/sdhci-xenon-phy.c
new file mode 100644
index 000000000..03ce57ef4
--- /dev/null
+++ b/drivers/mmc/host/sdhci-xenon-phy.c
@@ -0,0 +1,844 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * PHY support for Xenon SDHC
+ *
+ * Copyright (C) 2016 Marvell, All Rights Reserved.
+ *
+ * Author:	Hu Ziji <huziji@marvell.com>
+ * Date:	2016-8-24
+ */
+
+#include <linux/slab.h>
+#include <linux/delay.h>
+#include <linux/ktime.h>
+#include <linux/of_address.h>
+
+#include "sdhci-pltfm.h"
+#include "sdhci-xenon.h"
+
+/* Register base for eMMC PHY 5.0 Version */
+#define XENON_EMMC_5_0_PHY_REG_BASE		0x0160
+/* Register base for eMMC PHY 5.1 Version */
+#define XENON_EMMC_PHY_REG_BASE			0x0170
+
+#define XENON_EMMC_PHY_TIMING_ADJUST		XENON_EMMC_PHY_REG_BASE
+#define XENON_EMMC_5_0_PHY_TIMING_ADJUST	XENON_EMMC_5_0_PHY_REG_BASE
+#define XENON_TIMING_ADJUST_SLOW_MODE		BIT(29)
+#define XENON_TIMING_ADJUST_SDIO_MODE		BIT(28)
+#define XENON_SAMPL_INV_QSP_PHASE_SELECT	BIT(18)
+#define XENON_SAMPL_INV_QSP_PHASE_SELECT_SHIFT	18
+#define XENON_PHY_INITIALIZAION			BIT(31)
+#define XENON_WAIT_CYCLE_BEFORE_USING_MASK	0xF
+#define XENON_WAIT_CYCLE_BEFORE_USING_SHIFT	12
+#define XENON_FC_SYNC_EN_DURATION_MASK		0xF
+#define XENON_FC_SYNC_EN_DURATION_SHIFT		8
+#define XENON_FC_SYNC_RST_EN_DURATION_MASK	0xF
+#define XENON_FC_SYNC_RST_EN_DURATION_SHIFT	4
+#define XENON_FC_SYNC_RST_DURATION_MASK		0xF
+#define XENON_FC_SYNC_RST_DURATION_SHIFT	0
+
+#define XENON_EMMC_PHY_FUNC_CONTROL		(XENON_EMMC_PHY_REG_BASE + 0x4)
+#define XENON_EMMC_5_0_PHY_FUNC_CONTROL		\
+	(XENON_EMMC_5_0_PHY_REG_BASE + 0x4)
+#define XENON_ASYNC_DDRMODE_MASK		BIT(23)
+#define XENON_ASYNC_DDRMODE_SHIFT		23
+#define XENON_CMD_DDR_MODE			BIT(16)
+#define XENON_DQ_DDR_MODE_SHIFT			8
+#define XENON_DQ_DDR_MODE_MASK			0xFF
+#define XENON_DQ_ASYNC_MODE			BIT(4)
+
+#define XENON_EMMC_PHY_PAD_CONTROL		(XENON_EMMC_PHY_REG_BASE + 0x8)
+#define XENON_EMMC_5_0_PHY_PAD_CONTROL		\
+	(XENON_EMMC_5_0_PHY_REG_BASE + 0x8)
+#define XENON_REC_EN_SHIFT			24
+#define XENON_REC_EN_MASK			0xF
+#define XENON_FC_DQ_RECEN			BIT(24)
+#define XENON_FC_CMD_RECEN			BIT(25)
+#define XENON_FC_QSP_RECEN			BIT(26)
+#define XENON_FC_QSN_RECEN			BIT(27)
+#define XENON_OEN_QSN				BIT(28)
+#define XENON_AUTO_RECEN_CTRL			BIT(30)
+#define XENON_FC_ALL_CMOS_RECEIVER		0xF000
+
+#define XENON_EMMC5_FC_QSP_PD			BIT(18)
+#define XENON_EMMC5_FC_QSP_PU			BIT(22)
+#define XENON_EMMC5_FC_CMD_PD			BIT(17)
+#define XENON_EMMC5_FC_CMD_PU			BIT(21)
+#define XENON_EMMC5_FC_DQ_PD			BIT(16)
+#define XENON_EMMC5_FC_DQ_PU			BIT(20)
+
+#define XENON_EMMC_PHY_PAD_CONTROL1		(XENON_EMMC_PHY_REG_BASE + 0xC)
+#define XENON_EMMC5_1_FC_QSP_PD			BIT(9)
+#define XENON_EMMC5_1_FC_QSP_PU			BIT(25)
+#define XENON_EMMC5_1_FC_CMD_PD			BIT(8)
+#define XENON_EMMC5_1_FC_CMD_PU			BIT(24)
+#define XENON_EMMC5_1_FC_DQ_PD			0xFF
+#define XENON_EMMC5_1_FC_DQ_PU			(0xFF << 16)
+
+#define XENON_EMMC_PHY_PAD_CONTROL2		(XENON_EMMC_PHY_REG_BASE + 0x10)
+#define XENON_EMMC_5_0_PHY_PAD_CONTROL2		\
+	(XENON_EMMC_5_0_PHY_REG_BASE + 0xC)
+#define XENON_ZNR_MASK				0x1F
+#define XENON_ZNR_SHIFT				8
+#define XENON_ZPR_MASK				0x1F
+/* Preferred ZNR and ZPR value vary between different boards.
+ * The specific ZNR and ZPR value should be defined here
+ * according to board actual timing.
+ */
+#define XENON_ZNR_DEF_VALUE			0xF
+#define XENON_ZPR_DEF_VALUE			0xF
+
+#define XENON_EMMC_PHY_DLL_CONTROL		(XENON_EMMC_PHY_REG_BASE + 0x14)
+#define XENON_EMMC_5_0_PHY_DLL_CONTROL		\
+	(XENON_EMMC_5_0_PHY_REG_BASE + 0x10)
+#define XENON_DLL_ENABLE			BIT(31)
+#define XENON_DLL_UPDATE_STROBE_5_0		BIT(30)
+#define XENON_DLL_REFCLK_SEL			BIT(30)
+#define XENON_DLL_UPDATE			BIT(23)
+#define XENON_DLL_PHSEL1_SHIFT			24
+#define XENON_DLL_PHSEL0_SHIFT			16
+#define XENON_DLL_PHASE_MASK			0x3F
+#define XENON_DLL_PHASE_90_DEGREE		0x1F
+#define XENON_DLL_FAST_LOCK			BIT(5)
+#define XENON_DLL_GAIN2X			BIT(3)
+#define XENON_DLL_BYPASS_EN			BIT(0)
+
+#define XENON_EMMC_5_0_PHY_LOGIC_TIMING_ADJUST	\
+	(XENON_EMMC_5_0_PHY_REG_BASE + 0x14)
+#define XENON_EMMC_5_0_PHY_LOGIC_TIMING_VALUE	0x5A54
+#define XENON_EMMC_PHY_LOGIC_TIMING_ADJUST	(XENON_EMMC_PHY_REG_BASE + 0x18)
+#define XENON_LOGIC_TIMING_VALUE		0x00AA8977
+
+/*
+ * List offset of PHY registers and some special register values
+ * in eMMC PHY 5.0 or eMMC PHY 5.1
+ */
+struct xenon_emmc_phy_regs {
+	/* Offset of Timing Adjust register */
+	u16 timing_adj;
+	/* Offset of Func Control register */
+	u16 func_ctrl;
+	/* Offset of Pad Control register */
+	u16 pad_ctrl;
+	/* Offset of Pad Control register 2 */
+	u16 pad_ctrl2;
+	/* Offset of DLL Control register */
+	u16 dll_ctrl;
+	/* Offset of Logic Timing Adjust register */
+	u16 logic_timing_adj;
+	/* DLL Update Enable bit */
+	u32 dll_update;
+	/* value in Logic Timing Adjustment register */
+	u32 logic_timing_val;
+};
+
+static const char * const phy_types[] = {
+	"emmc 5.0 phy",
+	"emmc 5.1 phy"
+};
+
+enum xenon_phy_type_enum {
+	EMMC_5_0_PHY,
+	EMMC_5_1_PHY,
+	NR_PHY_TYPES
+};
+
+enum soc_pad_ctrl_type {
+	SOC_PAD_SD,
+	SOC_PAD_FIXED_1_8V,
+};
+
+struct soc_pad_ctrl {
+	/* Register address of SoC PHY PAD ctrl */
+	void __iomem	*reg;
+	/* SoC PHY PAD ctrl type */
+	enum soc_pad_ctrl_type pad_type;
+	/* SoC specific operation to set SoC PHY PAD */
+	void (*set_soc_pad)(struct sdhci_host *host,
+			    unsigned char signal_voltage);
+};
+
+static struct xenon_emmc_phy_regs xenon_emmc_5_0_phy_regs = {
+	.timing_adj	= XENON_EMMC_5_0_PHY_TIMING_ADJUST,
+	.func_ctrl	= XENON_EMMC_5_0_PHY_FUNC_CONTROL,
+	.pad_ctrl	= XENON_EMMC_5_0_PHY_PAD_CONTROL,
+	.pad_ctrl2	= XENON_EMMC_5_0_PHY_PAD_CONTROL2,
+	.dll_ctrl	= XENON_EMMC_5_0_PHY_DLL_CONTROL,
+	.logic_timing_adj = XENON_EMMC_5_0_PHY_LOGIC_TIMING_ADJUST,
+	.dll_update	= XENON_DLL_UPDATE_STROBE_5_0,
+	.logic_timing_val = XENON_EMMC_5_0_PHY_LOGIC_TIMING_VALUE,
+};
+
+static struct xenon_emmc_phy_regs xenon_emmc_5_1_phy_regs = {
+	.timing_adj	= XENON_EMMC_PHY_TIMING_ADJUST,
+	.func_ctrl	= XENON_EMMC_PHY_FUNC_CONTROL,
+	.pad_ctrl	= XENON_EMMC_PHY_PAD_CONTROL,
+	.pad_ctrl2	= XENON_EMMC_PHY_PAD_CONTROL2,
+	.dll_ctrl	= XENON_EMMC_PHY_DLL_CONTROL,
+	.logic_timing_adj = XENON_EMMC_PHY_LOGIC_TIMING_ADJUST,
+	.dll_update	= XENON_DLL_UPDATE,
+	.logic_timing_val = XENON_LOGIC_TIMING_VALUE,
+};
+
+/*
+ * eMMC PHY configuration and operations
+ */
+struct xenon_emmc_phy_params {
+	bool	slow_mode;
+
+	u8	znr;
+	u8	zpr;
+
+	/* Nr of consecutive Sampling Points of a Valid Sampling Window */
+	u8	nr_tun_times;
+	/* Divider for calculating Tuning Step */
+	u8	tun_step_divider;
+
+	struct soc_pad_ctrl pad_ctrl;
+};
+
+static int xenon_alloc_emmc_phy(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	struct xenon_emmc_phy_params *params;
+
+	params = devm_kzalloc(mmc_dev(host->mmc), sizeof(*params), GFP_KERNEL);
+	if (!params)
+		return -ENOMEM;
+
+	priv->phy_params = params;
+	if (priv->phy_type == EMMC_5_0_PHY)
+		priv->emmc_phy_regs = &xenon_emmc_5_0_phy_regs;
+	else
+		priv->emmc_phy_regs = &xenon_emmc_5_1_phy_regs;
+
+	return 0;
+}
+
+/*
+ * eMMC 5.0/5.1 PHY init/re-init.
+ * eMMC PHY init should be executed after:
+ * 1. SDCLK frequency changes.
+ * 2. SDCLK is stopped and re-enabled.
+ * 3. config in emmc_phy_regs->timing_adj and emmc_phy_regs->func_ctrl
+ * are changed
+ */
+static int xenon_emmc_phy_init(struct sdhci_host *host)
+{
+	u32 reg;
+	u32 wait, clock;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	struct xenon_emmc_phy_regs *phy_regs = priv->emmc_phy_regs;
+
+	reg = sdhci_readl(host, phy_regs->timing_adj);
+	reg |= XENON_PHY_INITIALIZAION;
+	sdhci_writel(host, reg, phy_regs->timing_adj);
+
+	/* Add duration of FC_SYNC_RST */
+	wait = ((reg >> XENON_FC_SYNC_RST_DURATION_SHIFT) &
+			XENON_FC_SYNC_RST_DURATION_MASK);
+	/* Add interval between FC_SYNC_EN and FC_SYNC_RST */
+	wait += ((reg >> XENON_FC_SYNC_RST_EN_DURATION_SHIFT) &
+			XENON_FC_SYNC_RST_EN_DURATION_MASK);
+	/* Add duration of asserting FC_SYNC_EN */
+	wait += ((reg >> XENON_FC_SYNC_EN_DURATION_SHIFT) &
+			XENON_FC_SYNC_EN_DURATION_MASK);
+	/* Add duration of waiting for PHY */
+	wait += ((reg >> XENON_WAIT_CYCLE_BEFORE_USING_SHIFT) &
+			XENON_WAIT_CYCLE_BEFORE_USING_MASK);
+	/* 4 additional bus clock and 4 AXI bus clock are required */
+	wait += 8;
+	wait <<= 20;
+
+	clock = host->clock;
+	if (!clock)
+		/* Use the possibly slowest bus frequency value */
+		clock = XENON_LOWEST_SDCLK_FREQ;
+	/* get the wait time */
+	wait /= clock;
+	wait++;
+	/* wait for host eMMC PHY init completes */
+	udelay(wait);
+
+	reg = sdhci_readl(host, phy_regs->timing_adj);
+	reg &= XENON_PHY_INITIALIZAION;
+	if (reg) {
+		dev_err(mmc_dev(host->mmc), "eMMC PHY init cannot complete after %d us\n",
+			wait);
+		return -ETIMEDOUT;
+	}
+
+	return 0;
+}
+
+#define ARMADA_3700_SOC_PAD_1_8V	0x1
+#define ARMADA_3700_SOC_PAD_3_3V	0x0
+
+static void armada_3700_soc_pad_voltage_set(struct sdhci_host *host,
+					    unsigned char signal_voltage)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	struct xenon_emmc_phy_params *params = priv->phy_params;
+
+	if (params->pad_ctrl.pad_type == SOC_PAD_FIXED_1_8V) {
+		writel(ARMADA_3700_SOC_PAD_1_8V, params->pad_ctrl.reg);
+	} else if (params->pad_ctrl.pad_type == SOC_PAD_SD) {
+		if (signal_voltage == MMC_SIGNAL_VOLTAGE_180)
+			writel(ARMADA_3700_SOC_PAD_1_8V, params->pad_ctrl.reg);
+		else if (signal_voltage == MMC_SIGNAL_VOLTAGE_330)
+			writel(ARMADA_3700_SOC_PAD_3_3V, params->pad_ctrl.reg);
+	}
+}
+
+/*
+ * Set SoC PHY voltage PAD control register,
+ * according to the operation voltage on PAD.
+ * The detailed operation depends on SoC implementation.
+ */
+static void xenon_emmc_phy_set_soc_pad(struct sdhci_host *host,
+				       unsigned char signal_voltage)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	struct xenon_emmc_phy_params *params = priv->phy_params;
+
+	if (!params->pad_ctrl.reg)
+		return;
+
+	if (params->pad_ctrl.set_soc_pad)
+		params->pad_ctrl.set_soc_pad(host, signal_voltage);
+}
+
+/*
+ * Enable eMMC PHY HW DLL
+ * DLL should be enabled and stable before HS200/SDR104 tuning,
+ * and before HS400 data strobe setting.
+ */
+static int xenon_emmc_phy_enable_dll(struct sdhci_host *host)
+{
+	u32 reg;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	struct xenon_emmc_phy_regs *phy_regs = priv->emmc_phy_regs;
+	ktime_t timeout;
+
+	if (WARN_ON(host->clock <= MMC_HIGH_52_MAX_DTR))
+		return -EINVAL;
+
+	reg = sdhci_readl(host, phy_regs->dll_ctrl);
+	if (reg & XENON_DLL_ENABLE)
+		return 0;
+
+	/* Enable DLL */
+	reg = sdhci_readl(host, phy_regs->dll_ctrl);
+	reg |= (XENON_DLL_ENABLE | XENON_DLL_FAST_LOCK);
+
+	/*
+	 * Set Phase as 90 degree, which is most common value.
+	 * Might set another value if necessary.
+	 * The granularity is 1 degree.
+	 */
+	reg &= ~((XENON_DLL_PHASE_MASK << XENON_DLL_PHSEL0_SHIFT) |
+		 (XENON_DLL_PHASE_MASK << XENON_DLL_PHSEL1_SHIFT));
+	reg |= ((XENON_DLL_PHASE_90_DEGREE << XENON_DLL_PHSEL0_SHIFT) |
+		(XENON_DLL_PHASE_90_DEGREE << XENON_DLL_PHSEL1_SHIFT));
+
+	reg &= ~XENON_DLL_BYPASS_EN;
+	reg |= phy_regs->dll_update;
+	if (priv->phy_type == EMMC_5_1_PHY)
+		reg &= ~XENON_DLL_REFCLK_SEL;
+	sdhci_writel(host, reg, phy_regs->dll_ctrl);
+
+	/* Wait max 32 ms */
+	timeout = ktime_add_ms(ktime_get(), 32);
+	while (1) {
+		bool timedout = ktime_after(ktime_get(), timeout);
+
+		if (sdhci_readw(host, XENON_SLOT_EXT_PRESENT_STATE) &
+		    XENON_DLL_LOCK_STATE)
+			break;
+		if (timedout) {
+			dev_err(mmc_dev(host->mmc), "Wait for DLL Lock time-out\n");
+			return -ETIMEDOUT;
+		}
+		udelay(100);
+	}
+	return 0;
+}
+
+/*
+ * Config to eMMC PHY to prepare for tuning.
+ * Enable HW DLL and set the TUNING_STEP
+ */
+static int xenon_emmc_phy_config_tuning(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	struct xenon_emmc_phy_params *params = priv->phy_params;
+	u32 reg, tuning_step;
+	int ret;
+
+	if (host->clock <= MMC_HIGH_52_MAX_DTR)
+		return -EINVAL;
+
+	ret = xenon_emmc_phy_enable_dll(host);
+	if (ret)
+		return ret;
+
+	/* Achieve TUNING_STEP with HW DLL help */
+	reg = sdhci_readl(host, XENON_SLOT_DLL_CUR_DLY_VAL);
+	tuning_step = reg / params->tun_step_divider;
+	if (unlikely(tuning_step > XENON_TUNING_STEP_MASK)) {
+		dev_warn(mmc_dev(host->mmc),
+			 "HS200 TUNING_STEP %d is larger than MAX value\n",
+			 tuning_step);
+		tuning_step = XENON_TUNING_STEP_MASK;
+	}
+
+	/* Set TUNING_STEP for later tuning */
+	reg = sdhci_readl(host, XENON_SLOT_OP_STATUS_CTRL);
+	reg &= ~(XENON_TUN_CONSECUTIVE_TIMES_MASK <<
+		 XENON_TUN_CONSECUTIVE_TIMES_SHIFT);
+	reg |= (params->nr_tun_times << XENON_TUN_CONSECUTIVE_TIMES_SHIFT);
+	reg &= ~(XENON_TUNING_STEP_MASK << XENON_TUNING_STEP_SHIFT);
+	reg |= (tuning_step << XENON_TUNING_STEP_SHIFT);
+	sdhci_writel(host, reg, XENON_SLOT_OP_STATUS_CTRL);
+
+	return 0;
+}
+
+static void xenon_emmc_phy_disable_strobe(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	u32 reg;
+
+	/* Disable both SDHC Data Strobe and Enhanced Strobe */
+	reg = sdhci_readl(host, XENON_SLOT_EMMC_CTRL);
+	reg &= ~(XENON_ENABLE_DATA_STROBE | XENON_ENABLE_RESP_STROBE);
+	sdhci_writel(host, reg, XENON_SLOT_EMMC_CTRL);
+
+	/* Clear Strobe line Pull down or Pull up */
+	if (priv->phy_type == EMMC_5_0_PHY) {
+		reg = sdhci_readl(host, XENON_EMMC_5_0_PHY_PAD_CONTROL);
+		reg &= ~(XENON_EMMC5_FC_QSP_PD | XENON_EMMC5_FC_QSP_PU);
+		sdhci_writel(host, reg, XENON_EMMC_5_0_PHY_PAD_CONTROL);
+	} else {
+		reg = sdhci_readl(host, XENON_EMMC_PHY_PAD_CONTROL1);
+		reg &= ~(XENON_EMMC5_1_FC_QSP_PD | XENON_EMMC5_1_FC_QSP_PU);
+		sdhci_writel(host, reg, XENON_EMMC_PHY_PAD_CONTROL1);
+	}
+}
+
+/* Set HS400 Data Strobe and Enhanced Strobe */
+static void xenon_emmc_phy_strobe_delay_adj(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	u32 reg;
+
+	if (WARN_ON(host->timing != MMC_TIMING_MMC_HS400))
+		return;
+
+	if (host->clock <= MMC_HIGH_52_MAX_DTR)
+		return;
+
+	dev_dbg(mmc_dev(host->mmc), "starts HS400 strobe delay adjustment\n");
+
+	xenon_emmc_phy_enable_dll(host);
+
+	/* Enable SDHC Data Strobe */
+	reg = sdhci_readl(host, XENON_SLOT_EMMC_CTRL);
+	reg |= XENON_ENABLE_DATA_STROBE;
+	/*
+	 * Enable SDHC Enhanced Strobe if supported
+	 * Xenon Enhanced Strobe should be enabled only when
+	 * 1. card is in HS400 mode and
+	 * 2. SDCLK is higher than 52MHz
+	 * 3. DLL is enabled
+	 */
+	if (host->mmc->ios.enhanced_strobe)
+		reg |= XENON_ENABLE_RESP_STROBE;
+	sdhci_writel(host, reg, XENON_SLOT_EMMC_CTRL);
+
+	/* Set Data Strobe Pull down */
+	if (priv->phy_type == EMMC_5_0_PHY) {
+		reg = sdhci_readl(host, XENON_EMMC_5_0_PHY_PAD_CONTROL);
+		reg |= XENON_EMMC5_FC_QSP_PD;
+		reg &= ~XENON_EMMC5_FC_QSP_PU;
+		sdhci_writel(host, reg, XENON_EMMC_5_0_PHY_PAD_CONTROL);
+	} else {
+		reg = sdhci_readl(host, XENON_EMMC_PHY_PAD_CONTROL1);
+		reg |= XENON_EMMC5_1_FC_QSP_PD;
+		reg &= ~XENON_EMMC5_1_FC_QSP_PU;
+		sdhci_writel(host, reg, XENON_EMMC_PHY_PAD_CONTROL1);
+	}
+}
+
+/*
+ * If eMMC PHY Slow Mode is required in lower speed mode (SDCLK < 55MHz)
+ * in SDR mode, enable Slow Mode to bypass eMMC PHY.
+ * SDIO slower SDR mode also requires Slow Mode.
+ *
+ * If Slow Mode is enabled, return true.
+ * Otherwise, return false.
+ */
+static bool xenon_emmc_phy_slow_mode(struct sdhci_host *host,
+				     unsigned char timing)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	struct xenon_emmc_phy_params *params = priv->phy_params;
+	struct xenon_emmc_phy_regs *phy_regs = priv->emmc_phy_regs;
+	u32 reg;
+	int ret;
+
+	if (host->clock > MMC_HIGH_52_MAX_DTR)
+		return false;
+
+	reg = sdhci_readl(host, phy_regs->timing_adj);
+	/* When in slower SDR mode, enable Slow Mode for SDIO
+	 * or when Slow Mode flag is set
+	 */
+	switch (timing) {
+	case MMC_TIMING_LEGACY:
+		/*
+		 * If Slow Mode is required, enable Slow Mode by default
+		 * in early init phase to avoid any potential issue.
+		 */
+		if (params->slow_mode) {
+			reg |= XENON_TIMING_ADJUST_SLOW_MODE;
+			ret = true;
+		} else {
+			reg &= ~XENON_TIMING_ADJUST_SLOW_MODE;
+			ret = false;
+		}
+		break;
+	case MMC_TIMING_UHS_SDR25:
+	case MMC_TIMING_UHS_SDR12:
+	case MMC_TIMING_SD_HS:
+	case MMC_TIMING_MMC_HS:
+		if ((priv->init_card_type == MMC_TYPE_SDIO) ||
+		    params->slow_mode) {
+			reg |= XENON_TIMING_ADJUST_SLOW_MODE;
+			ret = true;
+			break;
+		}
+		fallthrough;
+	default:
+		reg &= ~XENON_TIMING_ADJUST_SLOW_MODE;
+		ret = false;
+	}
+
+	sdhci_writel(host, reg, phy_regs->timing_adj);
+	return ret;
+}
+
+/*
+ * Set-up eMMC 5.0/5.1 PHY.
+ * Specific configuration depends on the current speed mode in use.
+ */
+static void xenon_emmc_phy_set(struct sdhci_host *host,
+			       unsigned char timing)
+{
+	u32 reg;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	struct xenon_emmc_phy_params *params = priv->phy_params;
+	struct xenon_emmc_phy_regs *phy_regs = priv->emmc_phy_regs;
+
+	dev_dbg(mmc_dev(host->mmc), "eMMC PHY setting starts\n");
+
+	/* Setup pad, set bit[28] and bits[26:24] */
+	reg = sdhci_readl(host, phy_regs->pad_ctrl);
+	reg |= (XENON_FC_DQ_RECEN | XENON_FC_CMD_RECEN |
+		XENON_FC_QSP_RECEN | XENON_OEN_QSN);
+	/* All FC_XX_RECEIVCE should be set as CMOS Type */
+	reg |= XENON_FC_ALL_CMOS_RECEIVER;
+	sdhci_writel(host, reg, phy_regs->pad_ctrl);
+
+	/* Set CMD and DQ Pull Up */
+	if (priv->phy_type == EMMC_5_0_PHY) {
+		reg = sdhci_readl(host, XENON_EMMC_5_0_PHY_PAD_CONTROL);
+		reg |= (XENON_EMMC5_FC_CMD_PU | XENON_EMMC5_FC_DQ_PU);
+		reg &= ~(XENON_EMMC5_FC_CMD_PD | XENON_EMMC5_FC_DQ_PD);
+		sdhci_writel(host, reg, XENON_EMMC_5_0_PHY_PAD_CONTROL);
+	} else {
+		reg = sdhci_readl(host, XENON_EMMC_PHY_PAD_CONTROL1);
+		reg |= (XENON_EMMC5_1_FC_CMD_PU | XENON_EMMC5_1_FC_DQ_PU);
+		reg &= ~(XENON_EMMC5_1_FC_CMD_PD | XENON_EMMC5_1_FC_DQ_PD);
+		sdhci_writel(host, reg, XENON_EMMC_PHY_PAD_CONTROL1);
+	}
+
+	if (timing == MMC_TIMING_LEGACY) {
+		xenon_emmc_phy_slow_mode(host, timing);
+		goto phy_init;
+	}
+
+	/*
+	 * If SDIO card, set SDIO Mode
+	 * Otherwise, clear SDIO Mode
+	 */
+	reg = sdhci_readl(host, phy_regs->timing_adj);
+	if (priv->init_card_type == MMC_TYPE_SDIO)
+		reg |= XENON_TIMING_ADJUST_SDIO_MODE;
+	else
+		reg &= ~XENON_TIMING_ADJUST_SDIO_MODE;
+	sdhci_writel(host, reg, phy_regs->timing_adj);
+
+	if (xenon_emmc_phy_slow_mode(host, timing))
+		goto phy_init;
+
+	/*
+	 * Set preferred ZNR and ZPR value
+	 * The ZNR and ZPR value vary between different boards.
+	 * Define them both in sdhci-xenon-emmc-phy.h.
+	 */
+	reg = sdhci_readl(host, phy_regs->pad_ctrl2);
+	reg &= ~((XENON_ZNR_MASK << XENON_ZNR_SHIFT) | XENON_ZPR_MASK);
+	reg |= ((params->znr << XENON_ZNR_SHIFT) | params->zpr);
+	sdhci_writel(host, reg, phy_regs->pad_ctrl2);
+
+	/*
+	 * When setting EMMC_PHY_FUNC_CONTROL register,
+	 * SD clock should be disabled
+	 */
+	reg = sdhci_readl(host, SDHCI_CLOCK_CONTROL);
+	reg &= ~SDHCI_CLOCK_CARD_EN;
+	sdhci_writew(host, reg, SDHCI_CLOCK_CONTROL);
+
+	reg = sdhci_readl(host, phy_regs->func_ctrl);
+	switch (timing) {
+	case MMC_TIMING_MMC_HS400:
+		reg |= (XENON_DQ_DDR_MODE_MASK << XENON_DQ_DDR_MODE_SHIFT) |
+		       XENON_CMD_DDR_MODE;
+		reg &= ~XENON_DQ_ASYNC_MODE;
+		break;
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_MMC_DDR52:
+		reg |= (XENON_DQ_DDR_MODE_MASK << XENON_DQ_DDR_MODE_SHIFT) |
+		       XENON_CMD_DDR_MODE | XENON_DQ_ASYNC_MODE;
+		break;
+	default:
+		reg &= ~((XENON_DQ_DDR_MODE_MASK << XENON_DQ_DDR_MODE_SHIFT) |
+			 XENON_CMD_DDR_MODE);
+		reg |= XENON_DQ_ASYNC_MODE;
+	}
+	sdhci_writel(host, reg, phy_regs->func_ctrl);
+
+	/* Enable bus clock */
+	reg = sdhci_readl(host, SDHCI_CLOCK_CONTROL);
+	reg |= SDHCI_CLOCK_CARD_EN;
+	sdhci_writew(host, reg, SDHCI_CLOCK_CONTROL);
+
+	if (timing == MMC_TIMING_MMC_HS400)
+		/* Hardware team recommend a value for HS400 */
+		sdhci_writel(host, phy_regs->logic_timing_val,
+			     phy_regs->logic_timing_adj);
+	else
+		xenon_emmc_phy_disable_strobe(host);
+
+phy_init:
+	xenon_emmc_phy_init(host);
+
+	dev_dbg(mmc_dev(host->mmc), "eMMC PHY setting completes\n");
+}
+
+static int get_dt_pad_ctrl_data(struct sdhci_host *host,
+				struct device_node *np,
+				struct xenon_emmc_phy_params *params)
+{
+	int ret = 0;
+	const char *name;
+	struct resource iomem;
+
+	if (of_device_is_compatible(np, "marvell,armada-3700-sdhci"))
+		params->pad_ctrl.set_soc_pad = armada_3700_soc_pad_voltage_set;
+	else
+		return 0;
+
+	if (of_address_to_resource(np, 1, &iomem)) {
+		dev_err(mmc_dev(host->mmc), "Unable to find SoC PAD ctrl register address for %pOFn\n",
+			np);
+		return -EINVAL;
+	}
+
+	params->pad_ctrl.reg = devm_ioremap_resource(mmc_dev(host->mmc),
+						     &iomem);
+	if (IS_ERR(params->pad_ctrl.reg))
+		return PTR_ERR(params->pad_ctrl.reg);
+
+	ret = of_property_read_string(np, "marvell,pad-type", &name);
+	if (ret) {
+		dev_err(mmc_dev(host->mmc), "Unable to determine SoC PHY PAD ctrl type\n");
+		return ret;
+	}
+	if (!strcmp(name, "sd")) {
+		params->pad_ctrl.pad_type = SOC_PAD_SD;
+	} else if (!strcmp(name, "fixed-1-8v")) {
+		params->pad_ctrl.pad_type = SOC_PAD_FIXED_1_8V;
+	} else {
+		dev_err(mmc_dev(host->mmc), "Unsupported SoC PHY PAD ctrl type %s\n",
+			name);
+		return -EINVAL;
+	}
+
+	return ret;
+}
+
+static int xenon_emmc_phy_parse_param_dt(struct sdhci_host *host,
+					 struct device_node *np,
+					 struct xenon_emmc_phy_params *params)
+{
+	u32 value;
+
+	params->slow_mode = false;
+	if (of_property_read_bool(np, "marvell,xenon-phy-slow-mode"))
+		params->slow_mode = true;
+
+	params->znr = XENON_ZNR_DEF_VALUE;
+	if (!of_property_read_u32(np, "marvell,xenon-phy-znr", &value))
+		params->znr = value & XENON_ZNR_MASK;
+
+	params->zpr = XENON_ZPR_DEF_VALUE;
+	if (!of_property_read_u32(np, "marvell,xenon-phy-zpr", &value))
+		params->zpr = value & XENON_ZPR_MASK;
+
+	params->nr_tun_times = XENON_TUN_CONSECUTIVE_TIMES;
+	if (!of_property_read_u32(np, "marvell,xenon-phy-nr-success-tun",
+				  &value))
+		params->nr_tun_times = value & XENON_TUN_CONSECUTIVE_TIMES_MASK;
+
+	params->tun_step_divider = XENON_TUNING_STEP_DIVIDER;
+	if (!of_property_read_u32(np, "marvell,xenon-phy-tun-step-divider",
+				  &value))
+		params->tun_step_divider = value & 0xFF;
+
+	return get_dt_pad_ctrl_data(host, np, params);
+}
+
+/* Set SoC PHY Voltage PAD */
+void xenon_soc_pad_ctrl(struct sdhci_host *host,
+			unsigned char signal_voltage)
+{
+	xenon_emmc_phy_set_soc_pad(host, signal_voltage);
+}
+
+/*
+ * Setting PHY when card is working in High Speed Mode.
+ * HS400 set Data Strobe and Enhanced Strobe if it is supported.
+ * HS200/SDR104 set tuning config to prepare for tuning.
+ */
+static int xenon_hs_delay_adj(struct sdhci_host *host)
+{
+	int ret = 0;
+
+	if (WARN_ON(host->clock <= XENON_DEFAULT_SDCLK_FREQ))
+		return -EINVAL;
+
+	switch (host->timing) {
+	case MMC_TIMING_MMC_HS400:
+		xenon_emmc_phy_strobe_delay_adj(host);
+		return 0;
+	case MMC_TIMING_MMC_HS200:
+	case MMC_TIMING_UHS_SDR104:
+		return xenon_emmc_phy_config_tuning(host);
+	case MMC_TIMING_MMC_DDR52:
+	case MMC_TIMING_UHS_DDR50:
+		/*
+		 * DDR Mode requires driver to scan Sampling Fixed Delay Line,
+		 * to find out a perfect operation sampling point.
+		 * It is hard to implement such a scan in host driver
+		 * since initiating commands by host driver is not safe.
+		 * Thus so far just keep PHY Sampling Fixed Delay in
+		 * default value of DDR mode.
+		 *
+		 * If any timing issue occurs in DDR mode on Marvell products,
+		 * please contact maintainer for internal support in Marvell.
+		 */
+		dev_warn_once(mmc_dev(host->mmc), "Timing issue might occur in DDR mode\n");
+		return 0;
+	}
+
+	return ret;
+}
+
+/*
+ * Adjust PHY setting.
+ * PHY setting should be adjusted when SDCLK frequency, Bus Width
+ * or Speed Mode is changed.
+ * Additional config are required when card is working in High Speed mode,
+ * after leaving Legacy Mode.
+ */
+int xenon_phy_adj(struct sdhci_host *host, struct mmc_ios *ios)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	int ret = 0;
+
+	if (!host->clock) {
+		priv->clock = 0;
+		return 0;
+	}
+
+	/*
+	 * The timing, frequency or bus width is changed,
+	 * better to set eMMC PHY based on current setting
+	 * and adjust Xenon SDHC delay.
+	 */
+	if ((host->clock == priv->clock) &&
+	    (ios->bus_width == priv->bus_width) &&
+	    (ios->timing == priv->timing))
+		return 0;
+
+	xenon_emmc_phy_set(host, ios->timing);
+
+	/* Update the record */
+	priv->bus_width = ios->bus_width;
+
+	priv->timing = ios->timing;
+	priv->clock = host->clock;
+
+	/* Legacy mode is a special case */
+	if (ios->timing == MMC_TIMING_LEGACY)
+		return 0;
+
+	if (host->clock > XENON_DEFAULT_SDCLK_FREQ)
+		ret = xenon_hs_delay_adj(host);
+	return ret;
+}
+
+static int xenon_add_phy(struct device_node *np, struct sdhci_host *host,
+			 const char *phy_name)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	priv->phy_type = match_string(phy_types, NR_PHY_TYPES, phy_name);
+	if (priv->phy_type < 0) {
+		dev_err(mmc_dev(host->mmc),
+			"Unable to determine PHY name %s. Use default eMMC 5.1 PHY\n",
+			phy_name);
+		priv->phy_type = EMMC_5_1_PHY;
+	}
+
+	ret = xenon_alloc_emmc_phy(host);
+	if (ret)
+		return ret;
+
+	return xenon_emmc_phy_parse_param_dt(host, np, priv->phy_params);
+}
+
+int xenon_phy_parse_dt(struct device_node *np, struct sdhci_host *host)
+{
+	const char *phy_type = NULL;
+
+	if (!of_property_read_string(np, "marvell,xenon-phy-type", &phy_type))
+		return xenon_add_phy(np, host, phy_type);
+
+	return xenon_add_phy(np, host, "emmc 5.1 phy");
+}
diff --git a/drivers/mmc/host/sdhci-xenon.c b/drivers/mmc/host/sdhci-xenon.c
new file mode 100644
index 000000000..d509198c0
--- /dev/null
+++ b/drivers/mmc/host/sdhci-xenon.c
@@ -0,0 +1,687 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Driver for Marvell Xenon SDHC as a platform device
+ *
+ * Copyright (C) 2016 Marvell, All Rights Reserved.
+ *
+ * Author:	Hu Ziji <huziji@marvell.com>
+ * Date:	2016-8-24
+ *
+ * Inspired by Jisheng Zhang <jszhang@marvell.com>
+ * Special thanks to Video BG4 project team.
+ */
+
+#include <linux/delay.h>
+#include <linux/ktime.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+
+#include "sdhci-pltfm.h"
+#include "sdhci-xenon.h"
+
+static int xenon_enable_internal_clk(struct sdhci_host *host)
+{
+	u32 reg;
+	ktime_t timeout;
+
+	reg = sdhci_readl(host, SDHCI_CLOCK_CONTROL);
+	reg |= SDHCI_CLOCK_INT_EN;
+	sdhci_writel(host, reg, SDHCI_CLOCK_CONTROL);
+	/* Wait max 20 ms */
+	timeout = ktime_add_ms(ktime_get(), 20);
+	while (1) {
+		bool timedout = ktime_after(ktime_get(), timeout);
+
+		reg = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+		if (reg & SDHCI_CLOCK_INT_STABLE)
+			break;
+		if (timedout) {
+			dev_err(mmc_dev(host->mmc), "Internal clock never stabilised.\n");
+			return -ETIMEDOUT;
+		}
+		usleep_range(900, 1100);
+	}
+
+	return 0;
+}
+
+/* Set SDCLK-off-while-idle */
+static void xenon_set_sdclk_off_idle(struct sdhci_host *host,
+				     unsigned char sdhc_id, bool enable)
+{
+	u32 reg;
+	u32 mask;
+
+	reg = sdhci_readl(host, XENON_SYS_OP_CTRL);
+	/* Get the bit shift basing on the SDHC index */
+	mask = (0x1 << (XENON_SDCLK_IDLEOFF_ENABLE_SHIFT + sdhc_id));
+	if (enable)
+		reg |= mask;
+	else
+		reg &= ~mask;
+
+	sdhci_writel(host, reg, XENON_SYS_OP_CTRL);
+}
+
+/* Enable/Disable the Auto Clock Gating function */
+static void xenon_set_acg(struct sdhci_host *host, bool enable)
+{
+	u32 reg;
+
+	reg = sdhci_readl(host, XENON_SYS_OP_CTRL);
+	if (enable)
+		reg &= ~XENON_AUTO_CLKGATE_DISABLE_MASK;
+	else
+		reg |= XENON_AUTO_CLKGATE_DISABLE_MASK;
+	sdhci_writel(host, reg, XENON_SYS_OP_CTRL);
+}
+
+/* Enable this SDHC */
+static void xenon_enable_sdhc(struct sdhci_host *host,
+			      unsigned char sdhc_id)
+{
+	u32 reg;
+
+	reg = sdhci_readl(host, XENON_SYS_OP_CTRL);
+	reg |= (BIT(sdhc_id) << XENON_SLOT_ENABLE_SHIFT);
+	sdhci_writel(host, reg, XENON_SYS_OP_CTRL);
+
+	host->mmc->caps |= MMC_CAP_WAIT_WHILE_BUSY;
+	/*
+	 * Force to clear BUS_TEST to
+	 * skip bus_test_pre and bus_test_post
+	 */
+	host->mmc->caps &= ~MMC_CAP_BUS_WIDTH_TEST;
+}
+
+/* Disable this SDHC */
+static void xenon_disable_sdhc(struct sdhci_host *host,
+			       unsigned char sdhc_id)
+{
+	u32 reg;
+
+	reg = sdhci_readl(host, XENON_SYS_OP_CTRL);
+	reg &= ~(BIT(sdhc_id) << XENON_SLOT_ENABLE_SHIFT);
+	sdhci_writel(host, reg, XENON_SYS_OP_CTRL);
+}
+
+/* Enable Parallel Transfer Mode */
+static void xenon_enable_sdhc_parallel_tran(struct sdhci_host *host,
+					    unsigned char sdhc_id)
+{
+	u32 reg;
+
+	reg = sdhci_readl(host, XENON_SYS_EXT_OP_CTRL);
+	reg |= BIT(sdhc_id);
+	sdhci_writel(host, reg, XENON_SYS_EXT_OP_CTRL);
+}
+
+/* Mask command conflict error */
+static void xenon_mask_cmd_conflict_err(struct sdhci_host *host)
+{
+	u32  reg;
+
+	reg = sdhci_readl(host, XENON_SYS_EXT_OP_CTRL);
+	reg |= XENON_MASK_CMD_CONFLICT_ERR;
+	sdhci_writel(host, reg, XENON_SYS_EXT_OP_CTRL);
+}
+
+static void xenon_retune_setup(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	u32 reg;
+
+	/* Disable the Re-Tuning Request functionality */
+	reg = sdhci_readl(host, XENON_SLOT_RETUNING_REQ_CTRL);
+	reg &= ~XENON_RETUNING_COMPATIBLE;
+	sdhci_writel(host, reg, XENON_SLOT_RETUNING_REQ_CTRL);
+
+	/* Disable the Re-tuning Interrupt */
+	reg = sdhci_readl(host, SDHCI_SIGNAL_ENABLE);
+	reg &= ~SDHCI_INT_RETUNE;
+	sdhci_writel(host, reg, SDHCI_SIGNAL_ENABLE);
+	reg = sdhci_readl(host, SDHCI_INT_ENABLE);
+	reg &= ~SDHCI_INT_RETUNE;
+	sdhci_writel(host, reg, SDHCI_INT_ENABLE);
+
+	/* Force to use Tuning Mode 1 */
+	host->tuning_mode = SDHCI_TUNING_MODE_1;
+	/* Set re-tuning period */
+	host->tuning_count = 1 << (priv->tuning_count - 1);
+}
+
+/*
+ * Operations inside struct sdhci_ops
+ */
+/* Recover the Register Setting cleared during SOFTWARE_RESET_ALL */
+static void xenon_reset_exit(struct sdhci_host *host,
+			     unsigned char sdhc_id, u8 mask)
+{
+	/* Only SOFTWARE RESET ALL will clear the register setting */
+	if (!(mask & SDHCI_RESET_ALL))
+		return;
+
+	/* Disable tuning request and auto-retuning again */
+	xenon_retune_setup(host);
+
+	/*
+	 * The ACG should be turned off at the early init time, in order
+	 * to solve a possible issues with the 1.8V regulator stabilization.
+	 * The feature is enabled in later stage.
+	 */
+	xenon_set_acg(host, false);
+
+	xenon_set_sdclk_off_idle(host, sdhc_id, false);
+
+	xenon_mask_cmd_conflict_err(host);
+}
+
+static void xenon_reset(struct sdhci_host *host, u8 mask)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+
+	sdhci_reset(host, mask);
+	xenon_reset_exit(host, priv->sdhc_id, mask);
+}
+
+/*
+ * Xenon defines different values for HS200 and HS400
+ * in Host_Control_2
+ */
+static void xenon_set_uhs_signaling(struct sdhci_host *host,
+				    unsigned int timing)
+{
+	u16 ctrl_2;
+
+	ctrl_2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	/* Select Bus Speed Mode for host */
+	ctrl_2 &= ~SDHCI_CTRL_UHS_MASK;
+	if (timing == MMC_TIMING_MMC_HS200)
+		ctrl_2 |= XENON_CTRL_HS200;
+	else if (timing == MMC_TIMING_UHS_SDR104)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR104;
+	else if (timing == MMC_TIMING_UHS_SDR12)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR12;
+	else if (timing == MMC_TIMING_UHS_SDR25)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR25;
+	else if (timing == MMC_TIMING_UHS_SDR50)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR50;
+	else if ((timing == MMC_TIMING_UHS_DDR50) ||
+		 (timing == MMC_TIMING_MMC_DDR52))
+		ctrl_2 |= SDHCI_CTRL_UHS_DDR50;
+	else if (timing == MMC_TIMING_MMC_HS400)
+		ctrl_2 |= XENON_CTRL_HS400;
+	sdhci_writew(host, ctrl_2, SDHCI_HOST_CONTROL2);
+}
+
+static void xenon_set_power(struct sdhci_host *host, unsigned char mode,
+		unsigned short vdd)
+{
+	struct mmc_host *mmc = host->mmc;
+	u8 pwr = host->pwr;
+
+	sdhci_set_power_noreg(host, mode, vdd);
+
+	if (host->pwr == pwr)
+		return;
+
+	if (host->pwr == 0)
+		vdd = 0;
+
+	if (!IS_ERR(mmc->supply.vmmc))
+		mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, vdd);
+}
+
+static void xenon_voltage_switch(struct sdhci_host *host)
+{
+	/* Wait for 5ms after set 1.8V signal enable bit */
+	usleep_range(5000, 5500);
+}
+
+static const struct sdhci_ops sdhci_xenon_ops = {
+	.voltage_switch		= xenon_voltage_switch,
+	.set_clock		= sdhci_set_clock,
+	.set_power		= xenon_set_power,
+	.set_bus_width		= sdhci_set_bus_width,
+	.reset			= xenon_reset,
+	.set_uhs_signaling	= xenon_set_uhs_signaling,
+	.get_max_clock		= sdhci_pltfm_clk_get_max_clock,
+};
+
+static const struct sdhci_pltfm_data sdhci_xenon_pdata = {
+	.ops = &sdhci_xenon_ops,
+	.quirks = SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC |
+		  SDHCI_QUIRK_NO_SIMULT_VDD_AND_POWER |
+		  SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN,
+};
+
+/*
+ * Xenon Specific Operations in mmc_host_ops
+ */
+static void xenon_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	u32 reg;
+
+	/*
+	 * HS400/HS200/eMMC HS doesn't have Preset Value register.
+	 * However, sdhci_set_ios will read HS400/HS200 Preset register.
+	 * Disable Preset Value register for HS400/HS200.
+	 * eMMC HS with preset_enabled set will trigger a bug in
+	 * get_preset_value().
+	 */
+	if ((ios->timing == MMC_TIMING_MMC_HS400) ||
+	    (ios->timing == MMC_TIMING_MMC_HS200) ||
+	    (ios->timing == MMC_TIMING_MMC_HS)) {
+		host->preset_enabled = false;
+		host->quirks2 |= SDHCI_QUIRK2_PRESET_VALUE_BROKEN;
+		host->flags &= ~SDHCI_PV_ENABLED;
+
+		reg = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+		reg &= ~SDHCI_CTRL_PRESET_VAL_ENABLE;
+		sdhci_writew(host, reg, SDHCI_HOST_CONTROL2);
+	} else {
+		host->quirks2 &= ~SDHCI_QUIRK2_PRESET_VALUE_BROKEN;
+	}
+
+	sdhci_set_ios(mmc, ios);
+	xenon_phy_adj(host, ios);
+
+	if (host->clock > XENON_DEFAULT_SDCLK_FREQ)
+		xenon_set_sdclk_off_idle(host, priv->sdhc_id, true);
+}
+
+static int xenon_start_signal_voltage_switch(struct mmc_host *mmc,
+					     struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	/*
+	 * Before SD/SDIO set signal voltage, SD bus clock should be
+	 * disabled. However, sdhci_set_clock will also disable the Internal
+	 * clock in mmc_set_signal_voltage().
+	 * If Internal clock is disabled, the 3.3V/1.8V bit can not be updated.
+	 * Thus here manually enable internal clock.
+	 *
+	 * After switch completes, it is unnecessary to disable internal clock,
+	 * since keeping internal clock active obeys SD spec.
+	 */
+	xenon_enable_internal_clk(host);
+
+	xenon_soc_pad_ctrl(host, ios->signal_voltage);
+
+	/*
+	 * If Vqmmc is fixed on platform, vqmmc regulator should be unavailable.
+	 * Thus SDHCI_CTRL_VDD_180 bit might not work then.
+	 * Skip the standard voltage switch to avoid any issue.
+	 */
+	if (PTR_ERR(mmc->supply.vqmmc) == -ENODEV)
+		return 0;
+
+	return sdhci_start_signal_voltage_switch(mmc, ios);
+}
+
+/*
+ * Update card type.
+ * priv->init_card_type will be used in PHY timing adjustment.
+ */
+static void xenon_init_card(struct mmc_host *mmc, struct mmc_card *card)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+
+	/* Update card type*/
+	priv->init_card_type = card->type;
+}
+
+static int xenon_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	if (host->timing == MMC_TIMING_UHS_DDR50 ||
+		host->timing == MMC_TIMING_MMC_DDR52)
+		return 0;
+
+	/*
+	 * Currently force Xenon driver back to support mode 1 only,
+	 * even though Xenon might claim to support mode 2 or mode 3.
+	 * It requires more time to test mode 2/mode 3 on more platforms.
+	 */
+	if (host->tuning_mode != SDHCI_TUNING_MODE_1)
+		xenon_retune_setup(host);
+
+	return sdhci_execute_tuning(mmc, opcode);
+}
+
+static void xenon_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	u32 reg;
+	u8 sdhc_id = priv->sdhc_id;
+
+	sdhci_enable_sdio_irq(mmc, enable);
+
+	if (enable) {
+		/*
+		 * Set SDIO Card Inserted indication
+		 * to enable detecting SDIO async irq.
+		 */
+		reg = sdhci_readl(host, XENON_SYS_CFG_INFO);
+		reg |= (1 << (sdhc_id + XENON_SLOT_TYPE_SDIO_SHIFT));
+		sdhci_writel(host, reg, XENON_SYS_CFG_INFO);
+	} else {
+		/* Clear SDIO Card Inserted indication */
+		reg = sdhci_readl(host, XENON_SYS_CFG_INFO);
+		reg &= ~(1 << (sdhc_id + XENON_SLOT_TYPE_SDIO_SHIFT));
+		sdhci_writel(host, reg, XENON_SYS_CFG_INFO);
+	}
+}
+
+static void xenon_replace_mmc_host_ops(struct sdhci_host *host)
+{
+	host->mmc_host_ops.set_ios = xenon_set_ios;
+	host->mmc_host_ops.start_signal_voltage_switch =
+			xenon_start_signal_voltage_switch;
+	host->mmc_host_ops.init_card = xenon_init_card;
+	host->mmc_host_ops.execute_tuning = xenon_execute_tuning;
+	host->mmc_host_ops.enable_sdio_irq = xenon_enable_sdio_irq;
+}
+
+/*
+ * Parse Xenon specific DT properties:
+ * sdhc-id: the index of current SDHC.
+ *	    Refer to XENON_SYS_CFG_INFO register
+ * tun-count: the interval between re-tuning
+ */
+static int xenon_probe_dt(struct platform_device *pdev)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct mmc_host *mmc = host->mmc;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	u32 sdhc_id, nr_sdhc;
+	u32 tuning_count;
+
+	/* Disable HS200 on Armada AP806 */
+	if (of_device_is_compatible(np, "marvell,armada-ap806-sdhci"))
+		host->quirks2 |= SDHCI_QUIRK2_BROKEN_HS200;
+
+	sdhc_id = 0x0;
+	if (!of_property_read_u32(np, "marvell,xenon-sdhc-id", &sdhc_id)) {
+		nr_sdhc = sdhci_readl(host, XENON_SYS_CFG_INFO);
+		nr_sdhc &= XENON_NR_SUPPORTED_SLOT_MASK;
+		if (unlikely(sdhc_id > nr_sdhc)) {
+			dev_err(mmc_dev(mmc), "SDHC Index %d exceeds Number of SDHCs %d\n",
+				sdhc_id, nr_sdhc);
+			return -EINVAL;
+		}
+	}
+	priv->sdhc_id = sdhc_id;
+
+	tuning_count = XENON_DEF_TUNING_COUNT;
+	if (!of_property_read_u32(np, "marvell,xenon-tun-count",
+				  &tuning_count)) {
+		if (unlikely(tuning_count >= XENON_TMR_RETUN_NO_PRESENT)) {
+			dev_err(mmc_dev(mmc), "Wrong Re-tuning Count. Set default value %d\n",
+				XENON_DEF_TUNING_COUNT);
+			tuning_count = XENON_DEF_TUNING_COUNT;
+		}
+	}
+	priv->tuning_count = tuning_count;
+
+	return xenon_phy_parse_dt(np, host);
+}
+
+static int xenon_sdhc_prepare(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	u8 sdhc_id = priv->sdhc_id;
+
+	/* Enable SDHC */
+	xenon_enable_sdhc(host, sdhc_id);
+
+	/* Enable ACG */
+	xenon_set_acg(host, true);
+
+	/* Enable Parallel Transfer Mode */
+	xenon_enable_sdhc_parallel_tran(host, sdhc_id);
+
+	/* Disable SDCLK-Off-While-Idle before card init */
+	xenon_set_sdclk_off_idle(host, sdhc_id, false);
+
+	xenon_mask_cmd_conflict_err(host);
+
+	return 0;
+}
+
+static void xenon_sdhc_unprepare(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	u8 sdhc_id = priv->sdhc_id;
+
+	/* disable SDHC */
+	xenon_disable_sdhc(host, sdhc_id);
+}
+
+static int xenon_probe(struct platform_device *pdev)
+{
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_host *host;
+	struct xenon_priv *priv;
+	int err;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_xenon_pdata,
+				sizeof(struct xenon_priv));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	priv = sdhci_pltfm_priv(pltfm_host);
+
+	/*
+	 * Link Xenon specific mmc_host_ops function,
+	 * to replace standard ones in sdhci_ops.
+	 */
+	xenon_replace_mmc_host_ops(host);
+
+	pltfm_host->clk = devm_clk_get(&pdev->dev, "core");
+	if (IS_ERR(pltfm_host->clk)) {
+		err = PTR_ERR(pltfm_host->clk);
+		dev_err(&pdev->dev, "Failed to setup input clk: %d\n", err);
+		goto free_pltfm;
+	}
+	err = clk_prepare_enable(pltfm_host->clk);
+	if (err)
+		goto free_pltfm;
+
+	priv->axi_clk = devm_clk_get(&pdev->dev, "axi");
+	if (IS_ERR(priv->axi_clk)) {
+		err = PTR_ERR(priv->axi_clk);
+		if (err == -EPROBE_DEFER)
+			goto err_clk;
+	} else {
+		err = clk_prepare_enable(priv->axi_clk);
+		if (err)
+			goto err_clk;
+	}
+
+	err = mmc_of_parse(host->mmc);
+	if (err)
+		goto err_clk_axi;
+
+	sdhci_get_of_property(pdev);
+
+	xenon_set_acg(host, false);
+
+	/* Xenon specific dt parse */
+	err = xenon_probe_dt(pdev);
+	if (err)
+		goto err_clk_axi;
+
+	err = xenon_sdhc_prepare(host);
+	if (err)
+		goto err_clk_axi;
+
+	pm_runtime_get_noresume(&pdev->dev);
+	pm_runtime_set_active(&pdev->dev);
+	pm_runtime_set_autosuspend_delay(&pdev->dev, 50);
+	pm_runtime_use_autosuspend(&pdev->dev);
+	pm_runtime_enable(&pdev->dev);
+	pm_suspend_ignore_children(&pdev->dev, 1);
+
+	err = sdhci_add_host(host);
+	if (err)
+		goto remove_sdhc;
+
+	pm_runtime_put_autosuspend(&pdev->dev);
+
+	return 0;
+
+remove_sdhc:
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+	xenon_sdhc_unprepare(host);
+err_clk_axi:
+	clk_disable_unprepare(priv->axi_clk);
+err_clk:
+	clk_disable_unprepare(pltfm_host->clk);
+free_pltfm:
+	sdhci_pltfm_free(pdev);
+	return err;
+}
+
+static int xenon_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+
+	pm_runtime_get_sync(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+
+	sdhci_remove_host(host, 0);
+
+	xenon_sdhc_unprepare(host);
+	clk_disable_unprepare(priv->axi_clk);
+	clk_disable_unprepare(pltfm_host->clk);
+
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int xenon_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	ret = pm_runtime_force_suspend(dev);
+
+	priv->restore_needed = true;
+	return ret;
+}
+#endif
+
+#ifdef CONFIG_PM
+static int xenon_runtime_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	ret = sdhci_runtime_suspend_host(host);
+	if (ret)
+		return ret;
+
+	if (host->tuning_mode != SDHCI_TUNING_MODE_3)
+		mmc_retune_needed(host->mmc);
+
+	clk_disable_unprepare(pltfm_host->clk);
+	/*
+	 * Need to update the priv->clock here, or when runtime resume
+	 * back, phy don't aware the clock change and won't adjust phy
+	 * which will cause cmd err
+	 */
+	priv->clock = 0;
+	return 0;
+}
+
+static int xenon_runtime_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct xenon_priv *priv = sdhci_pltfm_priv(pltfm_host);
+	int ret;
+
+	ret = clk_prepare_enable(pltfm_host->clk);
+	if (ret) {
+		dev_err(dev, "can't enable mainck\n");
+		return ret;
+	}
+
+	if (priv->restore_needed) {
+		ret = xenon_sdhc_prepare(host);
+		if (ret)
+			goto out;
+		priv->restore_needed = false;
+	}
+
+	ret = sdhci_runtime_resume_host(host, 0);
+	if (ret)
+		goto out;
+	return 0;
+out:
+	clk_disable_unprepare(pltfm_host->clk);
+	return ret;
+}
+#endif /* CONFIG_PM */
+
+static const struct dev_pm_ops sdhci_xenon_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(xenon_suspend,
+				pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(xenon_runtime_suspend,
+			   xenon_runtime_resume,
+			   NULL)
+};
+
+static const struct of_device_id sdhci_xenon_dt_ids[] = {
+	{ .compatible = "marvell,armada-ap806-sdhci",},
+	{ .compatible = "marvell,armada-cp110-sdhci",},
+	{ .compatible = "marvell,armada-3700-sdhci",},
+	{}
+};
+MODULE_DEVICE_TABLE(of, sdhci_xenon_dt_ids);
+
+static struct platform_driver sdhci_xenon_driver = {
+	.driver	= {
+		.name	= "xenon-sdhci",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = sdhci_xenon_dt_ids,
+		.pm = &sdhci_xenon_dev_pm_ops,
+	},
+	.probe	= xenon_probe,
+	.remove	= xenon_remove,
+};
+
+module_platform_driver(sdhci_xenon_driver);
+
+MODULE_DESCRIPTION("SDHCI platform driver for Marvell Xenon SDHC");
+MODULE_AUTHOR("Hu Ziji <huziji@marvell.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-xenon.h b/drivers/mmc/host/sdhci-xenon.h
new file mode 100644
index 000000000..593b82d7b
--- /dev/null
+++ b/drivers/mmc/host/sdhci-xenon.h
@@ -0,0 +1,100 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2016 Marvell, All Rights Reserved.
+ *
+ * Author:	Hu Ziji <huziji@marvell.com>
+ * Date:	2016-8-24
+ */
+#ifndef SDHCI_XENON_H_
+#define SDHCI_XENON_H_
+
+/* Register Offset of Xenon SDHC self-defined register */
+#define XENON_SYS_CFG_INFO			0x0104
+#define XENON_SLOT_TYPE_SDIO_SHIFT		24
+#define XENON_NR_SUPPORTED_SLOT_MASK		0x7
+
+#define XENON_SYS_OP_CTRL			0x0108
+#define XENON_AUTO_CLKGATE_DISABLE_MASK		BIT(20)
+#define XENON_SDCLK_IDLEOFF_ENABLE_SHIFT	8
+#define XENON_SLOT_ENABLE_SHIFT			0
+
+#define XENON_SYS_EXT_OP_CTRL			0x010C
+#define XENON_MASK_CMD_CONFLICT_ERR		BIT(8)
+
+#define XENON_SLOT_OP_STATUS_CTRL		0x0128
+#define XENON_TUN_CONSECUTIVE_TIMES_SHIFT	16
+#define XENON_TUN_CONSECUTIVE_TIMES_MASK	0x7
+#define XENON_TUN_CONSECUTIVE_TIMES		0x4
+#define XENON_TUNING_STEP_SHIFT			12
+#define XENON_TUNING_STEP_MASK			0xF
+#define XENON_TUNING_STEP_DIVIDER		BIT(6)
+
+#define XENON_SLOT_EMMC_CTRL			0x0130
+#define XENON_ENABLE_RESP_STROBE		BIT(25)
+#define XENON_ENABLE_DATA_STROBE		BIT(24)
+
+#define XENON_SLOT_RETUNING_REQ_CTRL		0x0144
+/* retuning compatible */
+#define XENON_RETUNING_COMPATIBLE		0x1
+
+#define XENON_SLOT_EXT_PRESENT_STATE		0x014C
+#define XENON_DLL_LOCK_STATE			0x1
+
+#define XENON_SLOT_DLL_CUR_DLY_VAL		0x0150
+
+/* Tuning Parameter */
+#define XENON_TMR_RETUN_NO_PRESENT		0xF
+#define XENON_DEF_TUNING_COUNT			0x9
+
+#define XENON_DEFAULT_SDCLK_FREQ		400000
+#define XENON_LOWEST_SDCLK_FREQ			100000
+
+/* Xenon specific Mode Select value */
+#define XENON_CTRL_HS200			0x5
+#define XENON_CTRL_HS400			0x6
+
+struct xenon_priv {
+	unsigned char	tuning_count;
+	/* idx of SDHC */
+	u8		sdhc_id;
+
+	/*
+	 * eMMC/SD/SDIO require different register settings.
+	 * Xenon driver has to recognize card type
+	 * before mmc_host->card is not available.
+	 * This field records the card type during init.
+	 * It is updated in xenon_init_card().
+	 *
+	 * It is only valid during initialization after it is updated.
+	 * Do not access this variable in normal transfers after
+	 * initialization completes.
+	 */
+	unsigned int	init_card_type;
+
+	/*
+	 * The bus_width, timing, and clock fields in below
+	 * record the current ios setting of Xenon SDHC.
+	 * Driver will adjust PHY setting if any change to
+	 * ios affects PHY timing.
+	 */
+	unsigned char	bus_width;
+	unsigned char	timing;
+	unsigned int	clock;
+	struct clk      *axi_clk;
+
+	int		phy_type;
+	/*
+	 * Contains board-specific PHY parameters
+	 * passed from device tree.
+	 */
+	void		*phy_params;
+	struct xenon_emmc_phy_regs *emmc_phy_regs;
+	bool restore_needed;
+};
+
+int xenon_phy_adj(struct sdhci_host *host, struct mmc_ios *ios);
+int xenon_phy_parse_dt(struct device_node *np,
+		       struct sdhci_host *host);
+void xenon_soc_pad_ctrl(struct sdhci_host *host,
+			unsigned char signal_voltage);
+#endif
diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
new file mode 100644
index 000000000..bad01cc68
--- /dev/null
+++ b/drivers/mmc/host/sdhci.c
@@ -0,0 +1,4929 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ *  linux/drivers/mmc/host/sdhci.c - Secure Digital Host Controller Interface driver
+ *
+ *  Copyright (C) 2005-2008 Pierre Ossman, All Rights Reserved.
+ *
+ * Thanks to the following companies for their support:
+ *
+ *     - JMicron (hardware and technical support)
+ */
+
+#include <linux/bitfield.h>
+#include <linux/delay.h>
+#include <linux/dmaengine.h>
+#include <linux/ktime.h>
+#include <linux/highmem.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/dma-mapping.h>
+#include <linux/slab.h>
+#include <linux/scatterlist.h>
+#include <linux/sizes.h>
+#include <linux/swiotlb.h>
+#include <linux/regulator/consumer.h>
+#include <linux/pm_runtime.h>
+#include <linux/of.h>
+
+#include <linux/leds.h>
+
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/sdio.h>
+#include <linux/mmc/slot-gpio.h>
+
+#include "sdhci.h"
+
+#define DRIVER_NAME "sdhci"
+
+#define DBG(f, x...) \
+	pr_debug("%s: " DRIVER_NAME ": " f, mmc_hostname(host->mmc), ## x)
+
+#define SDHCI_DUMP(f, x...) \
+	pr_err("%s: " DRIVER_NAME ": " f, mmc_hostname(host->mmc), ## x)
+
+#define MAX_TUNING_LOOP 40
+
+static unsigned int debug_quirks = 0;
+static unsigned int debug_quirks2;
+
+static void sdhci_enable_preset_value(struct sdhci_host *host, bool enable);
+
+static bool sdhci_send_command(struct sdhci_host *host, struct mmc_command *cmd);
+
+void sdhci_dumpregs(struct sdhci_host *host)
+{
+	SDHCI_DUMP("============ SDHCI REGISTER DUMP ===========\n");
+
+	SDHCI_DUMP("Sys addr:  0x%08x | Version:  0x%08x\n",
+		   sdhci_readl(host, SDHCI_DMA_ADDRESS),
+		   sdhci_readw(host, SDHCI_HOST_VERSION));
+	SDHCI_DUMP("Blk size:  0x%08x | Blk cnt:  0x%08x\n",
+		   sdhci_readw(host, SDHCI_BLOCK_SIZE),
+		   sdhci_readw(host, SDHCI_BLOCK_COUNT));
+	SDHCI_DUMP("Argument:  0x%08x | Trn mode: 0x%08x\n",
+		   sdhci_readl(host, SDHCI_ARGUMENT),
+		   sdhci_readw(host, SDHCI_TRANSFER_MODE));
+	SDHCI_DUMP("Present:   0x%08x | Host ctl: 0x%08x\n",
+		   sdhci_readl(host, SDHCI_PRESENT_STATE),
+		   sdhci_readb(host, SDHCI_HOST_CONTROL));
+	SDHCI_DUMP("Power:     0x%08x | Blk gap:  0x%08x\n",
+		   sdhci_readb(host, SDHCI_POWER_CONTROL),
+		   sdhci_readb(host, SDHCI_BLOCK_GAP_CONTROL));
+	SDHCI_DUMP("Wake-up:   0x%08x | Clock:    0x%08x\n",
+		   sdhci_readb(host, SDHCI_WAKE_UP_CONTROL),
+		   sdhci_readw(host, SDHCI_CLOCK_CONTROL));
+	SDHCI_DUMP("Timeout:   0x%08x | Int stat: 0x%08x\n",
+		   sdhci_readb(host, SDHCI_TIMEOUT_CONTROL),
+		   sdhci_readl(host, SDHCI_INT_STATUS));
+	SDHCI_DUMP("Int enab:  0x%08x | Sig enab: 0x%08x\n",
+		   sdhci_readl(host, SDHCI_INT_ENABLE),
+		   sdhci_readl(host, SDHCI_SIGNAL_ENABLE));
+	SDHCI_DUMP("ACmd stat: 0x%08x | Slot int: 0x%08x\n",
+		   sdhci_readw(host, SDHCI_AUTO_CMD_STATUS),
+		   sdhci_readw(host, SDHCI_SLOT_INT_STATUS));
+	SDHCI_DUMP("Caps:      0x%08x | Caps_1:   0x%08x\n",
+		   sdhci_readl(host, SDHCI_CAPABILITIES),
+		   sdhci_readl(host, SDHCI_CAPABILITIES_1));
+	SDHCI_DUMP("Cmd:       0x%08x | Max curr: 0x%08x\n",
+		   sdhci_readw(host, SDHCI_COMMAND),
+		   sdhci_readl(host, SDHCI_MAX_CURRENT));
+	SDHCI_DUMP("Resp[0]:   0x%08x | Resp[1]:  0x%08x\n",
+		   sdhci_readl(host, SDHCI_RESPONSE),
+		   sdhci_readl(host, SDHCI_RESPONSE + 4));
+	SDHCI_DUMP("Resp[2]:   0x%08x | Resp[3]:  0x%08x\n",
+		   sdhci_readl(host, SDHCI_RESPONSE + 8),
+		   sdhci_readl(host, SDHCI_RESPONSE + 12));
+	SDHCI_DUMP("Host ctl2: 0x%08x\n",
+		   sdhci_readw(host, SDHCI_HOST_CONTROL2));
+
+	if (host->flags & SDHCI_USE_ADMA) {
+		if (host->flags & SDHCI_USE_64_BIT_DMA) {
+			SDHCI_DUMP("ADMA Err:  0x%08x | ADMA Ptr: 0x%08x%08x\n",
+				   sdhci_readl(host, SDHCI_ADMA_ERROR),
+				   sdhci_readl(host, SDHCI_ADMA_ADDRESS_HI),
+				   sdhci_readl(host, SDHCI_ADMA_ADDRESS));
+		} else {
+			SDHCI_DUMP("ADMA Err:  0x%08x | ADMA Ptr: 0x%08x\n",
+				   sdhci_readl(host, SDHCI_ADMA_ERROR),
+				   sdhci_readl(host, SDHCI_ADMA_ADDRESS));
+		}
+	}
+
+	if (host->ops->dump_vendor_regs)
+		host->ops->dump_vendor_regs(host);
+
+	SDHCI_DUMP("============================================\n");
+}
+EXPORT_SYMBOL_GPL(sdhci_dumpregs);
+
+/*****************************************************************************\
+ *                                                                           *
+ * Low level functions                                                       *
+ *                                                                           *
+\*****************************************************************************/
+
+static void sdhci_do_enable_v4_mode(struct sdhci_host *host)
+{
+	u16 ctrl2;
+
+	ctrl2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	if (ctrl2 & SDHCI_CTRL_V4_MODE)
+		return;
+
+	ctrl2 |= SDHCI_CTRL_V4_MODE;
+	sdhci_writew(host, ctrl2, SDHCI_HOST_CONTROL2);
+}
+
+/*
+ * This can be called before sdhci_add_host() by Vendor's host controller
+ * driver to enable v4 mode if supported.
+ */
+void sdhci_enable_v4_mode(struct sdhci_host *host)
+{
+	host->v4_mode = true;
+	sdhci_do_enable_v4_mode(host);
+}
+EXPORT_SYMBOL_GPL(sdhci_enable_v4_mode);
+
+static inline bool sdhci_data_line_cmd(struct mmc_command *cmd)
+{
+	return cmd->data || cmd->flags & MMC_RSP_BUSY;
+}
+
+static void sdhci_set_card_detection(struct sdhci_host *host, bool enable)
+{
+	u32 present;
+
+	if ((host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) ||
+	    !mmc_card_is_removable(host->mmc) || mmc_can_gpio_cd(host->mmc))
+		return;
+
+	if (enable) {
+		present = sdhci_readl(host, SDHCI_PRESENT_STATE) &
+				      SDHCI_CARD_PRESENT;
+
+		host->ier |= present ? SDHCI_INT_CARD_REMOVE :
+				       SDHCI_INT_CARD_INSERT;
+	} else {
+		host->ier &= ~(SDHCI_INT_CARD_REMOVE | SDHCI_INT_CARD_INSERT);
+	}
+
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+}
+
+static void sdhci_enable_card_detection(struct sdhci_host *host)
+{
+	sdhci_set_card_detection(host, true);
+}
+
+static void sdhci_disable_card_detection(struct sdhci_host *host)
+{
+	sdhci_set_card_detection(host, false);
+}
+
+static void sdhci_runtime_pm_bus_on(struct sdhci_host *host)
+{
+	if (host->bus_on)
+		return;
+	host->bus_on = true;
+	pm_runtime_get_noresume(host->mmc->parent);
+}
+
+static void sdhci_runtime_pm_bus_off(struct sdhci_host *host)
+{
+	if (!host->bus_on)
+		return;
+	host->bus_on = false;
+	pm_runtime_put_noidle(host->mmc->parent);
+}
+
+void sdhci_reset(struct sdhci_host *host, u8 mask)
+{
+	ktime_t timeout;
+
+	sdhci_writeb(host, mask, SDHCI_SOFTWARE_RESET);
+
+	if (mask & SDHCI_RESET_ALL) {
+		host->clock = 0;
+		/* Reset-all turns off SD Bus Power */
+		if (host->quirks2 & SDHCI_QUIRK2_CARD_ON_NEEDS_BUS_ON)
+			sdhci_runtime_pm_bus_off(host);
+	}
+
+	/* Wait max 100 ms */
+	timeout = ktime_add_ms(ktime_get(), 100);
+
+	/* hw clears the bit when it's done */
+	while (1) {
+		bool timedout = ktime_after(ktime_get(), timeout);
+
+		if (!(sdhci_readb(host, SDHCI_SOFTWARE_RESET) & mask))
+			break;
+		if (timedout) {
+			pr_err("%s: Reset 0x%x never completed.\n",
+				mmc_hostname(host->mmc), (int)mask);
+			sdhci_dumpregs(host);
+			return;
+		}
+		udelay(10);
+	}
+}
+EXPORT_SYMBOL_GPL(sdhci_reset);
+
+static void sdhci_do_reset(struct sdhci_host *host, u8 mask)
+{
+	if (host->quirks & SDHCI_QUIRK_NO_CARD_NO_RESET) {
+		struct mmc_host *mmc = host->mmc;
+
+		if (!mmc->ops->get_cd(mmc))
+			return;
+	}
+
+	host->ops->reset(host, mask);
+
+	if (mask & SDHCI_RESET_ALL) {
+		if (host->flags & (SDHCI_USE_SDMA | SDHCI_USE_ADMA)) {
+			if (host->ops->enable_dma)
+				host->ops->enable_dma(host);
+		}
+
+		/* Resetting the controller clears many */
+		host->preset_enabled = false;
+	}
+}
+
+static void sdhci_set_default_irqs(struct sdhci_host *host)
+{
+	host->ier = SDHCI_INT_BUS_POWER | SDHCI_INT_DATA_END_BIT |
+		    SDHCI_INT_DATA_CRC | SDHCI_INT_DATA_TIMEOUT |
+		    SDHCI_INT_INDEX | SDHCI_INT_END_BIT | SDHCI_INT_CRC |
+		    SDHCI_INT_TIMEOUT | SDHCI_INT_DATA_END |
+		    SDHCI_INT_RESPONSE;
+
+	if (host->tuning_mode == SDHCI_TUNING_MODE_2 ||
+	    host->tuning_mode == SDHCI_TUNING_MODE_3)
+		host->ier |= SDHCI_INT_RETUNE;
+
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+}
+
+static void sdhci_config_dma(struct sdhci_host *host)
+{
+	u8 ctrl;
+	u16 ctrl2;
+
+	if (host->version < SDHCI_SPEC_200)
+		return;
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+
+	/*
+	 * Always adjust the DMA selection as some controllers
+	 * (e.g. JMicron) can't do PIO properly when the selection
+	 * is ADMA.
+	 */
+	ctrl &= ~SDHCI_CTRL_DMA_MASK;
+	if (!(host->flags & SDHCI_REQ_USE_DMA))
+		goto out;
+
+	/* Note if DMA Select is zero then SDMA is selected */
+	if (host->flags & SDHCI_USE_ADMA)
+		ctrl |= SDHCI_CTRL_ADMA32;
+
+	if (host->flags & SDHCI_USE_64_BIT_DMA) {
+		/*
+		 * If v4 mode, all supported DMA can be 64-bit addressing if
+		 * controller supports 64-bit system address, otherwise only
+		 * ADMA can support 64-bit addressing.
+		 */
+		if (host->v4_mode) {
+			ctrl2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+			ctrl2 |= SDHCI_CTRL_64BIT_ADDR;
+			sdhci_writew(host, ctrl2, SDHCI_HOST_CONTROL2);
+		} else if (host->flags & SDHCI_USE_ADMA) {
+			/*
+			 * Don't need to undo SDHCI_CTRL_ADMA32 in order to
+			 * set SDHCI_CTRL_ADMA64.
+			 */
+			ctrl |= SDHCI_CTRL_ADMA64;
+		}
+	}
+
+out:
+	sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+}
+
+static void sdhci_init(struct sdhci_host *host, int soft)
+{
+	struct mmc_host *mmc = host->mmc;
+	unsigned long flags;
+
+	if (soft)
+		sdhci_do_reset(host, SDHCI_RESET_CMD | SDHCI_RESET_DATA);
+	else
+		sdhci_do_reset(host, SDHCI_RESET_ALL);
+
+	if (host->v4_mode)
+		sdhci_do_enable_v4_mode(host);
+
+	spin_lock_irqsave(&host->lock, flags);
+	sdhci_set_default_irqs(host);
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	host->cqe_on = false;
+
+	if (soft) {
+		/* force clock reconfiguration */
+		host->clock = 0;
+		host->reinit_uhs = true;
+		mmc->ops->set_ios(mmc, &mmc->ios);
+	}
+}
+
+static void sdhci_reinit(struct sdhci_host *host)
+{
+	u32 cd = host->ier & (SDHCI_INT_CARD_REMOVE | SDHCI_INT_CARD_INSERT);
+
+	sdhci_init(host, 0);
+	sdhci_enable_card_detection(host);
+
+	/*
+	 * A change to the card detect bits indicates a change in present state,
+	 * refer sdhci_set_card_detection(). A card detect interrupt might have
+	 * been missed while the host controller was being reset, so trigger a
+	 * rescan to check.
+	 */
+	if (cd != (host->ier & (SDHCI_INT_CARD_REMOVE | SDHCI_INT_CARD_INSERT)))
+		mmc_detect_change(host->mmc, msecs_to_jiffies(200));
+}
+
+static void __sdhci_led_activate(struct sdhci_host *host)
+{
+	u8 ctrl;
+
+	if (host->quirks & SDHCI_QUIRK_NO_LED)
+		return;
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+	ctrl |= SDHCI_CTRL_LED;
+	sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+}
+
+static void __sdhci_led_deactivate(struct sdhci_host *host)
+{
+	u8 ctrl;
+
+	if (host->quirks & SDHCI_QUIRK_NO_LED)
+		return;
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+	ctrl &= ~SDHCI_CTRL_LED;
+	sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+}
+
+#if IS_REACHABLE(CONFIG_LEDS_CLASS)
+static void sdhci_led_control(struct led_classdev *led,
+			      enum led_brightness brightness)
+{
+	struct sdhci_host *host = container_of(led, struct sdhci_host, led);
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (host->runtime_suspended)
+		goto out;
+
+	if (brightness == LED_OFF)
+		__sdhci_led_deactivate(host);
+	else
+		__sdhci_led_activate(host);
+out:
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static int sdhci_led_register(struct sdhci_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+
+	if (host->quirks & SDHCI_QUIRK_NO_LED)
+		return 0;
+
+	snprintf(host->led_name, sizeof(host->led_name),
+		 "%s::", mmc_hostname(mmc));
+
+	host->led.name = host->led_name;
+	host->led.brightness = LED_OFF;
+	host->led.default_trigger = mmc_hostname(mmc);
+	host->led.brightness_set = sdhci_led_control;
+
+	return led_classdev_register(mmc_dev(mmc), &host->led);
+}
+
+static void sdhci_led_unregister(struct sdhci_host *host)
+{
+	if (host->quirks & SDHCI_QUIRK_NO_LED)
+		return;
+
+	led_classdev_unregister(&host->led);
+}
+
+static inline void sdhci_led_activate(struct sdhci_host *host)
+{
+}
+
+static inline void sdhci_led_deactivate(struct sdhci_host *host)
+{
+}
+
+#else
+
+static inline int sdhci_led_register(struct sdhci_host *host)
+{
+	return 0;
+}
+
+static inline void sdhci_led_unregister(struct sdhci_host *host)
+{
+}
+
+static inline void sdhci_led_activate(struct sdhci_host *host)
+{
+	__sdhci_led_activate(host);
+}
+
+static inline void sdhci_led_deactivate(struct sdhci_host *host)
+{
+	__sdhci_led_deactivate(host);
+}
+
+#endif
+
+static void sdhci_mod_timer(struct sdhci_host *host, struct mmc_request *mrq,
+			    unsigned long timeout)
+{
+	if (sdhci_data_line_cmd(mrq->cmd))
+		mod_timer(&host->data_timer, timeout);
+	else
+		mod_timer(&host->timer, timeout);
+}
+
+static void sdhci_del_timer(struct sdhci_host *host, struct mmc_request *mrq)
+{
+	if (sdhci_data_line_cmd(mrq->cmd))
+		del_timer(&host->data_timer);
+	else
+		del_timer(&host->timer);
+}
+
+static inline bool sdhci_has_requests(struct sdhci_host *host)
+{
+	return host->cmd || host->data_cmd;
+}
+
+/*****************************************************************************\
+ *                                                                           *
+ * Core functions                                                            *
+ *                                                                           *
+\*****************************************************************************/
+
+static void sdhci_read_block_pio(struct sdhci_host *host)
+{
+	unsigned long flags;
+	size_t blksize, len, chunk;
+	u32 scratch;
+	u8 *buf;
+
+	DBG("PIO reading\n");
+
+	blksize = host->data->blksz;
+	chunk = 0;
+
+	local_irq_save(flags);
+
+	while (blksize) {
+		BUG_ON(!sg_miter_next(&host->sg_miter));
+
+		len = min(host->sg_miter.length, blksize);
+
+		blksize -= len;
+		host->sg_miter.consumed = len;
+
+		buf = host->sg_miter.addr;
+
+		while (len) {
+			if (chunk == 0) {
+				scratch = sdhci_readl(host, SDHCI_BUFFER);
+				chunk = 4;
+			}
+
+			*buf = scratch & 0xFF;
+
+			buf++;
+			scratch >>= 8;
+			chunk--;
+			len--;
+		}
+	}
+
+	sg_miter_stop(&host->sg_miter);
+
+	local_irq_restore(flags);
+}
+
+static void sdhci_write_block_pio(struct sdhci_host *host)
+{
+	unsigned long flags;
+	size_t blksize, len, chunk;
+	u32 scratch;
+	u8 *buf;
+
+	DBG("PIO writing\n");
+
+	blksize = host->data->blksz;
+	chunk = 0;
+	scratch = 0;
+
+	local_irq_save(flags);
+
+	while (blksize) {
+		BUG_ON(!sg_miter_next(&host->sg_miter));
+
+		len = min(host->sg_miter.length, blksize);
+
+		blksize -= len;
+		host->sg_miter.consumed = len;
+
+		buf = host->sg_miter.addr;
+
+		while (len) {
+			scratch |= (u32)*buf << (chunk * 8);
+
+			buf++;
+			chunk++;
+			len--;
+
+			if ((chunk == 4) || ((len == 0) && (blksize == 0))) {
+				sdhci_writel(host, scratch, SDHCI_BUFFER);
+				chunk = 0;
+				scratch = 0;
+			}
+		}
+	}
+
+	sg_miter_stop(&host->sg_miter);
+
+	local_irq_restore(flags);
+}
+
+static void sdhci_transfer_pio(struct sdhci_host *host)
+{
+	u32 mask;
+
+	if (host->blocks == 0)
+		return;
+
+	if (host->data->flags & MMC_DATA_READ)
+		mask = SDHCI_DATA_AVAILABLE;
+	else
+		mask = SDHCI_SPACE_AVAILABLE;
+
+	/*
+	 * Some controllers (JMicron JMB38x) mess up the buffer bits
+	 * for transfers < 4 bytes. As long as it is just one block,
+	 * we can ignore the bits.
+	 */
+	if ((host->quirks & SDHCI_QUIRK_BROKEN_SMALL_PIO) &&
+		(host->data->blocks == 1))
+		mask = ~0;
+
+	while (sdhci_readl(host, SDHCI_PRESENT_STATE) & mask) {
+		if (host->quirks & SDHCI_QUIRK_PIO_NEEDS_DELAY)
+			udelay(100);
+
+		if (host->data->flags & MMC_DATA_READ)
+			sdhci_read_block_pio(host);
+		else
+			sdhci_write_block_pio(host);
+
+		host->blocks--;
+		if (host->blocks == 0)
+			break;
+	}
+
+	DBG("PIO transfer complete.\n");
+}
+
+static int sdhci_pre_dma_transfer(struct sdhci_host *host,
+				  struct mmc_data *data, int cookie)
+{
+	int sg_count;
+
+	/*
+	 * If the data buffers are already mapped, return the previous
+	 * dma_map_sg() result.
+	 */
+	if (data->host_cookie == COOKIE_PRE_MAPPED)
+		return data->sg_count;
+
+	/* Bounce write requests to the bounce buffer */
+	if (host->bounce_buffer) {
+		unsigned int length = data->blksz * data->blocks;
+
+		if (length > host->bounce_buffer_size) {
+			pr_err("%s: asked for transfer of %u bytes exceeds bounce buffer %u bytes\n",
+			       mmc_hostname(host->mmc), length,
+			       host->bounce_buffer_size);
+			return -EIO;
+		}
+		if (mmc_get_dma_dir(data) == DMA_TO_DEVICE) {
+			/* Copy the data to the bounce buffer */
+			if (host->ops->copy_to_bounce_buffer) {
+				host->ops->copy_to_bounce_buffer(host,
+								 data, length);
+			} else {
+				sg_copy_to_buffer(data->sg, data->sg_len,
+						  host->bounce_buffer, length);
+			}
+		}
+		/* Switch ownership to the DMA */
+		dma_sync_single_for_device(host->mmc->parent,
+					   host->bounce_addr,
+					   host->bounce_buffer_size,
+					   mmc_get_dma_dir(data));
+		/* Just a dummy value */
+		sg_count = 1;
+	} else {
+		/* Just access the data directly from memory */
+		sg_count = dma_map_sg(mmc_dev(host->mmc),
+				      data->sg, data->sg_len,
+				      mmc_get_dma_dir(data));
+	}
+
+	if (sg_count == 0)
+		return -ENOSPC;
+
+	data->sg_count = sg_count;
+	data->host_cookie = cookie;
+
+	return sg_count;
+}
+
+static char *sdhci_kmap_atomic(struct scatterlist *sg, unsigned long *flags)
+{
+	local_irq_save(*flags);
+	return kmap_atomic(sg_page(sg)) + sg->offset;
+}
+
+static void sdhci_kunmap_atomic(void *buffer, unsigned long *flags)
+{
+	kunmap_atomic(buffer);
+	local_irq_restore(*flags);
+}
+
+void sdhci_adma_write_desc(struct sdhci_host *host, void **desc,
+			   dma_addr_t addr, int len, unsigned int cmd)
+{
+	struct sdhci_adma2_64_desc *dma_desc = *desc;
+
+	/* 32-bit and 64-bit descriptors have these members in same position */
+	dma_desc->cmd = cpu_to_le16(cmd);
+	dma_desc->len = cpu_to_le16(len);
+	dma_desc->addr_lo = cpu_to_le32(lower_32_bits(addr));
+
+	if (host->flags & SDHCI_USE_64_BIT_DMA)
+		dma_desc->addr_hi = cpu_to_le32(upper_32_bits(addr));
+
+	*desc += host->desc_sz;
+}
+EXPORT_SYMBOL_GPL(sdhci_adma_write_desc);
+
+static inline void __sdhci_adma_write_desc(struct sdhci_host *host,
+					   void **desc, dma_addr_t addr,
+					   int len, unsigned int cmd)
+{
+	if (host->ops->adma_write_desc)
+		host->ops->adma_write_desc(host, desc, addr, len, cmd);
+	else
+		sdhci_adma_write_desc(host, desc, addr, len, cmd);
+}
+
+static void sdhci_adma_mark_end(void *desc)
+{
+	struct sdhci_adma2_64_desc *dma_desc = desc;
+
+	/* 32-bit and 64-bit descriptors have 'cmd' in same position */
+	dma_desc->cmd |= cpu_to_le16(ADMA2_END);
+}
+
+static void sdhci_adma_table_pre(struct sdhci_host *host,
+	struct mmc_data *data, int sg_count)
+{
+	struct scatterlist *sg;
+	unsigned long flags;
+	dma_addr_t addr, align_addr;
+	void *desc, *align;
+	char *buffer;
+	int len, offset, i;
+
+	/*
+	 * The spec does not specify endianness of descriptor table.
+	 * We currently guess that it is LE.
+	 */
+
+	host->sg_count = sg_count;
+
+	desc = host->adma_table;
+	align = host->align_buffer;
+
+	align_addr = host->align_addr;
+
+	for_each_sg(data->sg, sg, host->sg_count, i) {
+		addr = sg_dma_address(sg);
+		len = sg_dma_len(sg);
+
+		/*
+		 * The SDHCI specification states that ADMA addresses must
+		 * be 32-bit aligned. If they aren't, then we use a bounce
+		 * buffer for the (up to three) bytes that screw up the
+		 * alignment.
+		 */
+		offset = (SDHCI_ADMA2_ALIGN - (addr & SDHCI_ADMA2_MASK)) &
+			 SDHCI_ADMA2_MASK;
+		if (offset) {
+			if (data->flags & MMC_DATA_WRITE) {
+				buffer = sdhci_kmap_atomic(sg, &flags);
+				memcpy(align, buffer, offset);
+				sdhci_kunmap_atomic(buffer, &flags);
+			}
+
+			/* tran, valid */
+			__sdhci_adma_write_desc(host, &desc, align_addr,
+						offset, ADMA2_TRAN_VALID);
+
+			BUG_ON(offset > 65536);
+
+			align += SDHCI_ADMA2_ALIGN;
+			align_addr += SDHCI_ADMA2_ALIGN;
+
+			addr += offset;
+			len -= offset;
+		}
+
+		/*
+		 * The block layer forces a minimum segment size of PAGE_SIZE,
+		 * so 'len' can be too big here if PAGE_SIZE >= 64KiB. Write
+		 * multiple descriptors, noting that the ADMA table is sized
+		 * for 4KiB chunks anyway, so it will be big enough.
+		 */
+		while (len > host->max_adma) {
+			int n = 32 * 1024; /* 32KiB*/
+
+			__sdhci_adma_write_desc(host, &desc, addr, n, ADMA2_TRAN_VALID);
+			addr += n;
+			len -= n;
+		}
+
+		/* tran, valid */
+		if (len)
+			__sdhci_adma_write_desc(host, &desc, addr, len,
+						ADMA2_TRAN_VALID);
+
+		/*
+		 * If this triggers then we have a calculation bug
+		 * somewhere. :/
+		 */
+		WARN_ON((desc - host->adma_table) >= host->adma_table_sz);
+	}
+
+	if (host->quirks & SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC) {
+		/* Mark the last descriptor as the terminating descriptor */
+		if (desc != host->adma_table) {
+			desc -= host->desc_sz;
+			sdhci_adma_mark_end(desc);
+		}
+	} else {
+		/* Add a terminating entry - nop, end, valid */
+		__sdhci_adma_write_desc(host, &desc, 0, 0, ADMA2_NOP_END_VALID);
+	}
+}
+
+static void sdhci_adma_table_post(struct sdhci_host *host,
+	struct mmc_data *data)
+{
+	struct scatterlist *sg;
+	int i, size;
+	void *align;
+	char *buffer;
+	unsigned long flags;
+
+	if (data->flags & MMC_DATA_READ) {
+		bool has_unaligned = false;
+
+		/* Do a quick scan of the SG list for any unaligned mappings */
+		for_each_sg(data->sg, sg, host->sg_count, i)
+			if (sg_dma_address(sg) & SDHCI_ADMA2_MASK) {
+				has_unaligned = true;
+				break;
+			}
+
+		if (has_unaligned) {
+			dma_sync_sg_for_cpu(mmc_dev(host->mmc), data->sg,
+					    data->sg_len, DMA_FROM_DEVICE);
+
+			align = host->align_buffer;
+
+			for_each_sg(data->sg, sg, host->sg_count, i) {
+				if (sg_dma_address(sg) & SDHCI_ADMA2_MASK) {
+					size = SDHCI_ADMA2_ALIGN -
+					       (sg_dma_address(sg) & SDHCI_ADMA2_MASK);
+
+					buffer = sdhci_kmap_atomic(sg, &flags);
+					memcpy(buffer, align, size);
+					sdhci_kunmap_atomic(buffer, &flags);
+
+					align += SDHCI_ADMA2_ALIGN;
+				}
+			}
+		}
+	}
+}
+
+static void sdhci_set_adma_addr(struct sdhci_host *host, dma_addr_t addr)
+{
+	sdhci_writel(host, lower_32_bits(addr), SDHCI_ADMA_ADDRESS);
+	if (host->flags & SDHCI_USE_64_BIT_DMA)
+		sdhci_writel(host, upper_32_bits(addr), SDHCI_ADMA_ADDRESS_HI);
+}
+
+static dma_addr_t sdhci_sdma_address(struct sdhci_host *host)
+{
+	if (host->bounce_buffer)
+		return host->bounce_addr;
+	else
+		return sg_dma_address(host->data->sg);
+}
+
+static void sdhci_set_sdma_addr(struct sdhci_host *host, dma_addr_t addr)
+{
+	if (host->v4_mode)
+		sdhci_set_adma_addr(host, addr);
+	else
+		sdhci_writel(host, addr, SDHCI_DMA_ADDRESS);
+}
+
+static unsigned int sdhci_target_timeout(struct sdhci_host *host,
+					 struct mmc_command *cmd,
+					 struct mmc_data *data)
+{
+	unsigned int target_timeout;
+
+	/* timeout in us */
+	if (!data) {
+		target_timeout = cmd->busy_timeout * 1000;
+	} else {
+		target_timeout = DIV_ROUND_UP(data->timeout_ns, 1000);
+		if (host->clock && data->timeout_clks) {
+			unsigned long long val;
+
+			/*
+			 * data->timeout_clks is in units of clock cycles.
+			 * host->clock is in Hz.  target_timeout is in us.
+			 * Hence, us = 1000000 * cycles / Hz.  Round up.
+			 */
+			val = 1000000ULL * data->timeout_clks;
+			if (do_div(val, host->clock))
+				target_timeout++;
+			target_timeout += val;
+		}
+	}
+
+	return target_timeout;
+}
+
+static void sdhci_calc_sw_timeout(struct sdhci_host *host,
+				  struct mmc_command *cmd)
+{
+	struct mmc_data *data = cmd->data;
+	struct mmc_host *mmc = host->mmc;
+	struct mmc_ios *ios = &mmc->ios;
+	unsigned char bus_width = 1 << ios->bus_width;
+	unsigned int blksz;
+	unsigned int freq;
+	u64 target_timeout;
+	u64 transfer_time;
+
+	target_timeout = sdhci_target_timeout(host, cmd, data);
+	target_timeout *= NSEC_PER_USEC;
+
+	if (data) {
+		blksz = data->blksz;
+		freq = host->mmc->actual_clock ? : host->clock;
+		transfer_time = (u64)blksz * NSEC_PER_SEC * (8 / bus_width);
+		do_div(transfer_time, freq);
+		/* multiply by '2' to account for any unknowns */
+		transfer_time = transfer_time * 2;
+		/* calculate timeout for the entire data */
+		host->data_timeout = data->blocks * target_timeout +
+				     transfer_time;
+	} else {
+		host->data_timeout = target_timeout;
+	}
+
+	if (host->data_timeout)
+		host->data_timeout += MMC_CMD_TRANSFER_TIME;
+}
+
+static u8 sdhci_calc_timeout(struct sdhci_host *host, struct mmc_command *cmd,
+			     bool *too_big)
+{
+	u8 count;
+	struct mmc_data *data;
+	unsigned target_timeout, current_timeout;
+
+	*too_big = true;
+
+	/*
+	 * If the host controller provides us with an incorrect timeout
+	 * value, just skip the check and use 0xE.  The hardware may take
+	 * longer to time out, but that's much better than having a too-short
+	 * timeout value.
+	 */
+	if (host->quirks & SDHCI_QUIRK_BROKEN_TIMEOUT_VAL)
+		return 0xE;
+
+	/* Unspecified command, asume max */
+	if (cmd == NULL)
+		return 0xE;
+
+	data = cmd->data;
+	/* Unspecified timeout, assume max */
+	if (!data && !cmd->busy_timeout)
+		return 0xE;
+
+	/* timeout in us */
+	target_timeout = sdhci_target_timeout(host, cmd, data);
+
+	/*
+	 * Figure out needed cycles.
+	 * We do this in steps in order to fit inside a 32 bit int.
+	 * The first step is the minimum timeout, which will have a
+	 * minimum resolution of 6 bits:
+	 * (1) 2^13*1000 > 2^22,
+	 * (2) host->timeout_clk < 2^16
+	 *     =>
+	 *     (1) / (2) > 2^6
+	 */
+	count = 0;
+	current_timeout = (1 << 13) * 1000 / host->timeout_clk;
+	while (current_timeout < target_timeout) {
+		count++;
+		current_timeout <<= 1;
+		if (count >= 0xF)
+			break;
+	}
+
+	if (count >= 0xF) {
+		if (!(host->quirks2 & SDHCI_QUIRK2_DISABLE_HW_TIMEOUT))
+			DBG("Too large timeout 0x%x requested for CMD%d!\n",
+			    count, cmd->opcode);
+		count = 0xE;
+	} else {
+		*too_big = false;
+	}
+
+	return count;
+}
+
+static void sdhci_set_transfer_irqs(struct sdhci_host *host)
+{
+	u32 pio_irqs = SDHCI_INT_DATA_AVAIL | SDHCI_INT_SPACE_AVAIL;
+	u32 dma_irqs = SDHCI_INT_DMA_END | SDHCI_INT_ADMA_ERROR;
+
+	if (host->flags & SDHCI_REQ_USE_DMA)
+		host->ier = (host->ier & ~pio_irqs) | dma_irqs;
+	else
+		host->ier = (host->ier & ~dma_irqs) | pio_irqs;
+
+	if (host->flags & (SDHCI_AUTO_CMD23 | SDHCI_AUTO_CMD12))
+		host->ier |= SDHCI_INT_AUTO_CMD_ERR;
+	else
+		host->ier &= ~SDHCI_INT_AUTO_CMD_ERR;
+
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+}
+
+void sdhci_set_data_timeout_irq(struct sdhci_host *host, bool enable)
+{
+	if (enable)
+		host->ier |= SDHCI_INT_DATA_TIMEOUT;
+	else
+		host->ier &= ~SDHCI_INT_DATA_TIMEOUT;
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+}
+EXPORT_SYMBOL_GPL(sdhci_set_data_timeout_irq);
+
+void __sdhci_set_timeout(struct sdhci_host *host, struct mmc_command *cmd)
+{
+	bool too_big = false;
+	u8 count = sdhci_calc_timeout(host, cmd, &too_big);
+
+	if (too_big &&
+	    host->quirks2 & SDHCI_QUIRK2_DISABLE_HW_TIMEOUT) {
+		sdhci_calc_sw_timeout(host, cmd);
+		sdhci_set_data_timeout_irq(host, false);
+	} else if (!(host->ier & SDHCI_INT_DATA_TIMEOUT)) {
+		sdhci_set_data_timeout_irq(host, true);
+	}
+
+	sdhci_writeb(host, count, SDHCI_TIMEOUT_CONTROL);
+}
+EXPORT_SYMBOL_GPL(__sdhci_set_timeout);
+
+static void sdhci_set_timeout(struct sdhci_host *host, struct mmc_command *cmd)
+{
+	if (host->ops->set_timeout)
+		host->ops->set_timeout(host, cmd);
+	else
+		__sdhci_set_timeout(host, cmd);
+}
+
+static void sdhci_initialize_data(struct sdhci_host *host,
+				  struct mmc_data *data)
+{
+	WARN_ON(host->data);
+
+	/* Sanity checks */
+	BUG_ON(data->blksz * data->blocks > 524288);
+	BUG_ON(data->blksz > host->mmc->max_blk_size);
+	BUG_ON(data->blocks > 65535);
+
+	host->data = data;
+	host->data_early = 0;
+	host->data->bytes_xfered = 0;
+}
+
+static inline void sdhci_set_block_info(struct sdhci_host *host,
+					struct mmc_data *data)
+{
+	/* Set the DMA boundary value and block size */
+	sdhci_writew(host,
+		     SDHCI_MAKE_BLKSZ(host->sdma_boundary, data->blksz),
+		     SDHCI_BLOCK_SIZE);
+	/*
+	 * For Version 4.10 onwards, if v4 mode is enabled, 32-bit Block Count
+	 * can be supported, in that case 16-bit block count register must be 0.
+	 */
+	if (host->version >= SDHCI_SPEC_410 && host->v4_mode &&
+	    (host->quirks2 & SDHCI_QUIRK2_USE_32BIT_BLK_CNT)) {
+		if (sdhci_readw(host, SDHCI_BLOCK_COUNT))
+			sdhci_writew(host, 0, SDHCI_BLOCK_COUNT);
+		sdhci_writew(host, data->blocks, SDHCI_32BIT_BLK_CNT);
+	} else {
+		sdhci_writew(host, data->blocks, SDHCI_BLOCK_COUNT);
+	}
+}
+
+static void sdhci_prepare_data(struct sdhci_host *host, struct mmc_command *cmd)
+{
+	struct mmc_data *data = cmd->data;
+
+	sdhci_initialize_data(host, data);
+
+	if (host->flags & (SDHCI_USE_SDMA | SDHCI_USE_ADMA)) {
+		struct scatterlist *sg;
+		unsigned int length_mask, offset_mask;
+		int i;
+
+		host->flags |= SDHCI_REQ_USE_DMA;
+
+		/*
+		 * FIXME: This doesn't account for merging when mapping the
+		 * scatterlist.
+		 *
+		 * The assumption here being that alignment and lengths are
+		 * the same after DMA mapping to device address space.
+		 */
+		length_mask = 0;
+		offset_mask = 0;
+		if (host->flags & SDHCI_USE_ADMA) {
+			if (host->quirks & SDHCI_QUIRK_32BIT_ADMA_SIZE) {
+				length_mask = 3;
+				/*
+				 * As we use up to 3 byte chunks to work
+				 * around alignment problems, we need to
+				 * check the offset as well.
+				 */
+				offset_mask = 3;
+			}
+		} else {
+			if (host->quirks & SDHCI_QUIRK_32BIT_DMA_SIZE)
+				length_mask = 3;
+			if (host->quirks & SDHCI_QUIRK_32BIT_DMA_ADDR)
+				offset_mask = 3;
+		}
+
+		if (unlikely(length_mask | offset_mask)) {
+			for_each_sg(data->sg, sg, data->sg_len, i) {
+				if (sg->length & length_mask) {
+					DBG("Reverting to PIO because of transfer size (%d)\n",
+					    sg->length);
+					host->flags &= ~SDHCI_REQ_USE_DMA;
+					break;
+				}
+				if (sg->offset & offset_mask) {
+					DBG("Reverting to PIO because of bad alignment\n");
+					host->flags &= ~SDHCI_REQ_USE_DMA;
+					break;
+				}
+			}
+		}
+	}
+
+	sdhci_config_dma(host);
+
+	if (host->flags & SDHCI_REQ_USE_DMA) {
+		int sg_cnt = sdhci_pre_dma_transfer(host, data, COOKIE_MAPPED);
+
+		if (sg_cnt <= 0) {
+			/*
+			 * This only happens when someone fed
+			 * us an invalid request.
+			 */
+			WARN_ON(1);
+			host->flags &= ~SDHCI_REQ_USE_DMA;
+		} else if (host->flags & SDHCI_USE_ADMA) {
+			sdhci_adma_table_pre(host, data, sg_cnt);
+			sdhci_set_adma_addr(host, host->adma_addr);
+		} else {
+			WARN_ON(sg_cnt != 1);
+			sdhci_set_sdma_addr(host, sdhci_sdma_address(host));
+		}
+	}
+
+	if (!(host->flags & SDHCI_REQ_USE_DMA)) {
+		int flags;
+
+		flags = SG_MITER_ATOMIC;
+		if (host->data->flags & MMC_DATA_READ)
+			flags |= SG_MITER_TO_SG;
+		else
+			flags |= SG_MITER_FROM_SG;
+		sg_miter_start(&host->sg_miter, data->sg, data->sg_len, flags);
+		host->blocks = data->blocks;
+	}
+
+	sdhci_set_transfer_irqs(host);
+
+	sdhci_set_block_info(host, data);
+}
+
+#if IS_ENABLED(CONFIG_MMC_SDHCI_EXTERNAL_DMA)
+
+static int sdhci_external_dma_init(struct sdhci_host *host)
+{
+	int ret = 0;
+	struct mmc_host *mmc = host->mmc;
+
+	host->tx_chan = dma_request_chan(mmc->parent, "tx");
+	if (IS_ERR(host->tx_chan)) {
+		ret = PTR_ERR(host->tx_chan);
+		if (ret != -EPROBE_DEFER)
+			pr_warn("Failed to request TX DMA channel.\n");
+		host->tx_chan = NULL;
+		return ret;
+	}
+
+	host->rx_chan = dma_request_chan(mmc->parent, "rx");
+	if (IS_ERR(host->rx_chan)) {
+		if (host->tx_chan) {
+			dma_release_channel(host->tx_chan);
+			host->tx_chan = NULL;
+		}
+
+		ret = PTR_ERR(host->rx_chan);
+		if (ret != -EPROBE_DEFER)
+			pr_warn("Failed to request RX DMA channel.\n");
+		host->rx_chan = NULL;
+	}
+
+	return ret;
+}
+
+static struct dma_chan *sdhci_external_dma_channel(struct sdhci_host *host,
+						   struct mmc_data *data)
+{
+	return data->flags & MMC_DATA_WRITE ? host->tx_chan : host->rx_chan;
+}
+
+static int sdhci_external_dma_setup(struct sdhci_host *host,
+				    struct mmc_command *cmd)
+{
+	int ret, i;
+	enum dma_transfer_direction dir;
+	struct dma_async_tx_descriptor *desc;
+	struct mmc_data *data = cmd->data;
+	struct dma_chan *chan;
+	struct dma_slave_config cfg;
+	dma_cookie_t cookie;
+	int sg_cnt;
+
+	if (!host->mapbase)
+		return -EINVAL;
+
+	memset(&cfg, 0, sizeof(cfg));
+	cfg.src_addr = host->mapbase + SDHCI_BUFFER;
+	cfg.dst_addr = host->mapbase + SDHCI_BUFFER;
+	cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	cfg.src_maxburst = data->blksz / 4;
+	cfg.dst_maxburst = data->blksz / 4;
+
+	/* Sanity check: all the SG entries must be aligned by block size. */
+	for (i = 0; i < data->sg_len; i++) {
+		if ((data->sg + i)->length % data->blksz)
+			return -EINVAL;
+	}
+
+	chan = sdhci_external_dma_channel(host, data);
+
+	ret = dmaengine_slave_config(chan, &cfg);
+	if (ret)
+		return ret;
+
+	sg_cnt = sdhci_pre_dma_transfer(host, data, COOKIE_MAPPED);
+	if (sg_cnt <= 0)
+		return -EINVAL;
+
+	dir = data->flags & MMC_DATA_WRITE ? DMA_MEM_TO_DEV : DMA_DEV_TO_MEM;
+	desc = dmaengine_prep_slave_sg(chan, data->sg, data->sg_len, dir,
+				       DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc)
+		return -EINVAL;
+
+	desc->callback = NULL;
+	desc->callback_param = NULL;
+
+	cookie = dmaengine_submit(desc);
+	if (dma_submit_error(cookie))
+		ret = cookie;
+
+	return ret;
+}
+
+static void sdhci_external_dma_release(struct sdhci_host *host)
+{
+	if (host->tx_chan) {
+		dma_release_channel(host->tx_chan);
+		host->tx_chan = NULL;
+	}
+
+	if (host->rx_chan) {
+		dma_release_channel(host->rx_chan);
+		host->rx_chan = NULL;
+	}
+
+	sdhci_switch_external_dma(host, false);
+}
+
+static void __sdhci_external_dma_prepare_data(struct sdhci_host *host,
+					      struct mmc_command *cmd)
+{
+	struct mmc_data *data = cmd->data;
+
+	sdhci_initialize_data(host, data);
+
+	host->flags |= SDHCI_REQ_USE_DMA;
+	sdhci_set_transfer_irqs(host);
+
+	sdhci_set_block_info(host, data);
+}
+
+static void sdhci_external_dma_prepare_data(struct sdhci_host *host,
+					    struct mmc_command *cmd)
+{
+	if (!sdhci_external_dma_setup(host, cmd)) {
+		__sdhci_external_dma_prepare_data(host, cmd);
+	} else {
+		sdhci_external_dma_release(host);
+		pr_err("%s: Cannot use external DMA, switch to the DMA/PIO which standard SDHCI provides.\n",
+		       mmc_hostname(host->mmc));
+		sdhci_prepare_data(host, cmd);
+	}
+}
+
+static void sdhci_external_dma_pre_transfer(struct sdhci_host *host,
+					    struct mmc_command *cmd)
+{
+	struct dma_chan *chan;
+
+	if (!cmd->data)
+		return;
+
+	chan = sdhci_external_dma_channel(host, cmd->data);
+	if (chan)
+		dma_async_issue_pending(chan);
+}
+
+#else
+
+static inline int sdhci_external_dma_init(struct sdhci_host *host)
+{
+	return -EOPNOTSUPP;
+}
+
+static inline void sdhci_external_dma_release(struct sdhci_host *host)
+{
+}
+
+static inline void sdhci_external_dma_prepare_data(struct sdhci_host *host,
+						   struct mmc_command *cmd)
+{
+	/* This should never happen */
+	WARN_ON_ONCE(1);
+}
+
+static inline void sdhci_external_dma_pre_transfer(struct sdhci_host *host,
+						   struct mmc_command *cmd)
+{
+}
+
+static inline struct dma_chan *sdhci_external_dma_channel(struct sdhci_host *host,
+							  struct mmc_data *data)
+{
+	return NULL;
+}
+
+#endif
+
+void sdhci_switch_external_dma(struct sdhci_host *host, bool en)
+{
+	host->use_external_dma = en;
+}
+EXPORT_SYMBOL_GPL(sdhci_switch_external_dma);
+
+static inline bool sdhci_auto_cmd12(struct sdhci_host *host,
+				    struct mmc_request *mrq)
+{
+	return !mrq->sbc && (host->flags & SDHCI_AUTO_CMD12) &&
+	       !mrq->cap_cmd_during_tfr;
+}
+
+static inline bool sdhci_auto_cmd23(struct sdhci_host *host,
+				    struct mmc_request *mrq)
+{
+	return mrq->sbc && (host->flags & SDHCI_AUTO_CMD23);
+}
+
+static inline bool sdhci_manual_cmd23(struct sdhci_host *host,
+				      struct mmc_request *mrq)
+{
+	return mrq->sbc && !(host->flags & SDHCI_AUTO_CMD23);
+}
+
+static inline void sdhci_auto_cmd_select(struct sdhci_host *host,
+					 struct mmc_command *cmd,
+					 u16 *mode)
+{
+	bool use_cmd12 = sdhci_auto_cmd12(host, cmd->mrq) &&
+			 (cmd->opcode != SD_IO_RW_EXTENDED);
+	bool use_cmd23 = sdhci_auto_cmd23(host, cmd->mrq);
+	u16 ctrl2;
+
+	/*
+	 * In case of Version 4.10 or later, use of 'Auto CMD Auto
+	 * Select' is recommended rather than use of 'Auto CMD12
+	 * Enable' or 'Auto CMD23 Enable'. We require Version 4 Mode
+	 * here because some controllers (e.g sdhci-of-dwmshc) expect it.
+	 */
+	if (host->version >= SDHCI_SPEC_410 && host->v4_mode &&
+	    (use_cmd12 || use_cmd23)) {
+		*mode |= SDHCI_TRNS_AUTO_SEL;
+
+		ctrl2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+		if (use_cmd23)
+			ctrl2 |= SDHCI_CMD23_ENABLE;
+		else
+			ctrl2 &= ~SDHCI_CMD23_ENABLE;
+		sdhci_writew(host, ctrl2, SDHCI_HOST_CONTROL2);
+
+		return;
+	}
+
+	/*
+	 * If we are sending CMD23, CMD12 never gets sent
+	 * on successful completion (so no Auto-CMD12).
+	 */
+	if (use_cmd12)
+		*mode |= SDHCI_TRNS_AUTO_CMD12;
+	else if (use_cmd23)
+		*mode |= SDHCI_TRNS_AUTO_CMD23;
+}
+
+static void sdhci_set_transfer_mode(struct sdhci_host *host,
+	struct mmc_command *cmd)
+{
+	u16 mode = 0;
+	struct mmc_data *data = cmd->data;
+
+	if (data == NULL) {
+		if (host->quirks2 &
+			SDHCI_QUIRK2_CLEAR_TRANSFERMODE_REG_BEFORE_CMD) {
+			/* must not clear SDHCI_TRANSFER_MODE when tuning */
+			if (cmd->opcode != MMC_SEND_TUNING_BLOCK_HS200)
+				sdhci_writew(host, 0x0, SDHCI_TRANSFER_MODE);
+		} else {
+		/* clear Auto CMD settings for no data CMDs */
+			mode = sdhci_readw(host, SDHCI_TRANSFER_MODE);
+			sdhci_writew(host, mode & ~(SDHCI_TRNS_AUTO_CMD12 |
+				SDHCI_TRNS_AUTO_CMD23), SDHCI_TRANSFER_MODE);
+		}
+		return;
+	}
+
+	WARN_ON(!host->data);
+
+	if (!(host->quirks2 & SDHCI_QUIRK2_SUPPORT_SINGLE))
+		mode = SDHCI_TRNS_BLK_CNT_EN;
+
+	if (mmc_op_multi(cmd->opcode) || data->blocks > 1) {
+		mode = SDHCI_TRNS_BLK_CNT_EN | SDHCI_TRNS_MULTI;
+		sdhci_auto_cmd_select(host, cmd, &mode);
+		if (sdhci_auto_cmd23(host, cmd->mrq))
+			sdhci_writel(host, cmd->mrq->sbc->arg, SDHCI_ARGUMENT2);
+	}
+
+	if (data->flags & MMC_DATA_READ)
+		mode |= SDHCI_TRNS_READ;
+	if (host->flags & SDHCI_REQ_USE_DMA)
+		mode |= SDHCI_TRNS_DMA;
+
+	sdhci_writew(host, mode, SDHCI_TRANSFER_MODE);
+}
+
+static bool sdhci_needs_reset(struct sdhci_host *host, struct mmc_request *mrq)
+{
+	return (!(host->flags & SDHCI_DEVICE_DEAD) &&
+		((mrq->cmd && mrq->cmd->error) ||
+		 (mrq->sbc && mrq->sbc->error) ||
+		 (mrq->data && mrq->data->stop && mrq->data->stop->error) ||
+		 (host->quirks & SDHCI_QUIRK_RESET_AFTER_REQUEST)));
+}
+
+static void sdhci_set_mrq_done(struct sdhci_host *host, struct mmc_request *mrq)
+{
+	int i;
+
+	for (i = 0; i < SDHCI_MAX_MRQS; i++) {
+		if (host->mrqs_done[i] == mrq) {
+			WARN_ON(1);
+			return;
+		}
+	}
+
+	for (i = 0; i < SDHCI_MAX_MRQS; i++) {
+		if (!host->mrqs_done[i]) {
+			host->mrqs_done[i] = mrq;
+			break;
+		}
+	}
+
+	WARN_ON(i >= SDHCI_MAX_MRQS);
+}
+
+static void __sdhci_finish_mrq(struct sdhci_host *host, struct mmc_request *mrq)
+{
+	if (host->cmd && host->cmd->mrq == mrq)
+		host->cmd = NULL;
+
+	if (host->data_cmd && host->data_cmd->mrq == mrq)
+		host->data_cmd = NULL;
+
+	if (host->deferred_cmd && host->deferred_cmd->mrq == mrq)
+		host->deferred_cmd = NULL;
+
+	if (host->data && host->data->mrq == mrq)
+		host->data = NULL;
+
+	if (sdhci_needs_reset(host, mrq))
+		host->pending_reset = true;
+
+	sdhci_set_mrq_done(host, mrq);
+
+	sdhci_del_timer(host, mrq);
+
+	if (!sdhci_has_requests(host))
+		sdhci_led_deactivate(host);
+}
+
+static void sdhci_finish_mrq(struct sdhci_host *host, struct mmc_request *mrq)
+{
+	__sdhci_finish_mrq(host, mrq);
+
+	queue_work(host->complete_wq, &host->complete_work);
+}
+
+static void __sdhci_finish_data(struct sdhci_host *host, bool sw_data_timeout)
+{
+	struct mmc_command *data_cmd = host->data_cmd;
+	struct mmc_data *data = host->data;
+
+	host->data = NULL;
+	host->data_cmd = NULL;
+
+	/*
+	 * The controller needs a reset of internal state machines upon error
+	 * conditions.
+	 */
+	if (data->error) {
+		if (!host->cmd || host->cmd == data_cmd)
+			sdhci_do_reset(host, SDHCI_RESET_CMD);
+		sdhci_do_reset(host, SDHCI_RESET_DATA);
+	}
+
+	if ((host->flags & (SDHCI_REQ_USE_DMA | SDHCI_USE_ADMA)) ==
+	    (SDHCI_REQ_USE_DMA | SDHCI_USE_ADMA))
+		sdhci_adma_table_post(host, data);
+
+	/*
+	 * The specification states that the block count register must
+	 * be updated, but it does not specify at what point in the
+	 * data flow. That makes the register entirely useless to read
+	 * back so we have to assume that nothing made it to the card
+	 * in the event of an error.
+	 */
+	if (data->error)
+		data->bytes_xfered = 0;
+	else
+		data->bytes_xfered = data->blksz * data->blocks;
+
+	/*
+	 * Need to send CMD12 if -
+	 * a) open-ended multiblock transfer not using auto CMD12 (no CMD23)
+	 * b) error in multiblock transfer
+	 */
+	if (data->stop &&
+	    ((!data->mrq->sbc && !sdhci_auto_cmd12(host, data->mrq)) ||
+	     data->error)) {
+		/*
+		 * 'cap_cmd_during_tfr' request must not use the command line
+		 * after mmc_command_done() has been called. It is upper layer's
+		 * responsibility to send the stop command if required.
+		 */
+		if (data->mrq->cap_cmd_during_tfr) {
+			__sdhci_finish_mrq(host, data->mrq);
+		} else {
+			/* Avoid triggering warning in sdhci_send_command() */
+			host->cmd = NULL;
+			if (!sdhci_send_command(host, data->stop)) {
+				if (sw_data_timeout) {
+					/*
+					 * This is anyway a sw data timeout, so
+					 * give up now.
+					 */
+					data->stop->error = -EIO;
+					__sdhci_finish_mrq(host, data->mrq);
+				} else {
+					WARN_ON(host->deferred_cmd);
+					host->deferred_cmd = data->stop;
+				}
+			}
+		}
+	} else {
+		__sdhci_finish_mrq(host, data->mrq);
+	}
+}
+
+static void sdhci_finish_data(struct sdhci_host *host)
+{
+	__sdhci_finish_data(host, false);
+}
+
+static bool sdhci_send_command(struct sdhci_host *host, struct mmc_command *cmd)
+{
+	int flags;
+	u32 mask;
+	unsigned long timeout;
+
+	WARN_ON(host->cmd);
+
+	/* Initially, a command has no error */
+	cmd->error = 0;
+
+	if ((host->quirks2 & SDHCI_QUIRK2_STOP_WITH_TC) &&
+	    cmd->opcode == MMC_STOP_TRANSMISSION)
+		cmd->flags |= MMC_RSP_BUSY;
+
+	mask = SDHCI_CMD_INHIBIT;
+	if (sdhci_data_line_cmd(cmd))
+		mask |= SDHCI_DATA_INHIBIT;
+
+	/* We shouldn't wait for data inihibit for stop commands, even
+	   though they might use busy signaling */
+	if (cmd->mrq->data && (cmd == cmd->mrq->data->stop))
+		mask &= ~SDHCI_DATA_INHIBIT;
+
+	if (sdhci_readl(host, SDHCI_PRESENT_STATE) & mask)
+		return false;
+
+	host->cmd = cmd;
+	host->data_timeout = 0;
+	if (sdhci_data_line_cmd(cmd)) {
+		WARN_ON(host->data_cmd);
+		host->data_cmd = cmd;
+		sdhci_set_timeout(host, cmd);
+	}
+
+	if (cmd->data) {
+		if (host->use_external_dma)
+			sdhci_external_dma_prepare_data(host, cmd);
+		else
+			sdhci_prepare_data(host, cmd);
+	}
+
+	sdhci_writel(host, cmd->arg, SDHCI_ARGUMENT);
+
+	sdhci_set_transfer_mode(host, cmd);
+
+	if ((cmd->flags & MMC_RSP_136) && (cmd->flags & MMC_RSP_BUSY)) {
+		WARN_ONCE(1, "Unsupported response type!\n");
+		/*
+		 * This does not happen in practice because 136-bit response
+		 * commands never have busy waiting, so rather than complicate
+		 * the error path, just remove busy waiting and continue.
+		 */
+		cmd->flags &= ~MMC_RSP_BUSY;
+	}
+
+	if (!(cmd->flags & MMC_RSP_PRESENT))
+		flags = SDHCI_CMD_RESP_NONE;
+	else if (cmd->flags & MMC_RSP_136)
+		flags = SDHCI_CMD_RESP_LONG;
+	else if (cmd->flags & MMC_RSP_BUSY)
+		flags = SDHCI_CMD_RESP_SHORT_BUSY;
+	else
+		flags = SDHCI_CMD_RESP_SHORT;
+
+	if (cmd->flags & MMC_RSP_CRC)
+		flags |= SDHCI_CMD_CRC;
+	if (cmd->flags & MMC_RSP_OPCODE)
+		flags |= SDHCI_CMD_INDEX;
+
+	/* CMD19 is special in that the Data Present Select should be set */
+	if (cmd->data || cmd->opcode == MMC_SEND_TUNING_BLOCK ||
+	    cmd->opcode == MMC_SEND_TUNING_BLOCK_HS200)
+		flags |= SDHCI_CMD_DATA;
+
+	timeout = jiffies;
+	if (host->data_timeout)
+		timeout += nsecs_to_jiffies(host->data_timeout);
+	else if (!cmd->data && cmd->busy_timeout > 9000)
+		timeout += DIV_ROUND_UP(cmd->busy_timeout, 1000) * HZ + HZ;
+	else
+		timeout += 10 * HZ;
+	sdhci_mod_timer(host, cmd->mrq, timeout);
+
+	if (host->use_external_dma)
+		sdhci_external_dma_pre_transfer(host, cmd);
+
+	sdhci_writew(host, SDHCI_MAKE_CMD(cmd->opcode, flags), SDHCI_COMMAND);
+
+	return true;
+}
+
+static bool sdhci_present_error(struct sdhci_host *host,
+				struct mmc_command *cmd, bool present)
+{
+	if (!present || host->flags & SDHCI_DEVICE_DEAD) {
+		cmd->error = -ENOMEDIUM;
+		return true;
+	}
+
+	return false;
+}
+
+static bool sdhci_send_command_retry(struct sdhci_host *host,
+				     struct mmc_command *cmd,
+				     unsigned long flags)
+	__releases(host->lock)
+	__acquires(host->lock)
+{
+	struct mmc_command *deferred_cmd = host->deferred_cmd;
+	int timeout = 10; /* Approx. 10 ms */
+	bool present;
+
+	while (!sdhci_send_command(host, cmd)) {
+		if (!timeout--) {
+			pr_err("%s: Controller never released inhibit bit(s).\n",
+			       mmc_hostname(host->mmc));
+			sdhci_dumpregs(host);
+			cmd->error = -EIO;
+			return false;
+		}
+
+		spin_unlock_irqrestore(&host->lock, flags);
+
+		usleep_range(1000, 1250);
+
+		present = host->mmc->ops->get_cd(host->mmc);
+
+		spin_lock_irqsave(&host->lock, flags);
+
+		/* A deferred command might disappear, handle that */
+		if (cmd == deferred_cmd && cmd != host->deferred_cmd)
+			return true;
+
+		if (sdhci_present_error(host, cmd, present))
+			return false;
+	}
+
+	if (cmd == host->deferred_cmd)
+		host->deferred_cmd = NULL;
+
+	return true;
+}
+
+static void sdhci_read_rsp_136(struct sdhci_host *host, struct mmc_command *cmd)
+{
+	int i, reg;
+
+	for (i = 0; i < 4; i++) {
+		reg = SDHCI_RESPONSE + (3 - i) * 4;
+		cmd->resp[i] = sdhci_readl(host, reg);
+	}
+
+	if (host->quirks2 & SDHCI_QUIRK2_RSP_136_HAS_CRC)
+		return;
+
+	/* CRC is stripped so we need to do some shifting */
+	for (i = 0; i < 4; i++) {
+		cmd->resp[i] <<= 8;
+		if (i != 3)
+			cmd->resp[i] |= cmd->resp[i + 1] >> 24;
+	}
+}
+
+static void sdhci_finish_command(struct sdhci_host *host)
+{
+	struct mmc_command *cmd = host->cmd;
+
+	host->cmd = NULL;
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136) {
+			sdhci_read_rsp_136(host, cmd);
+		} else {
+			cmd->resp[0] = sdhci_readl(host, SDHCI_RESPONSE);
+		}
+	}
+
+	if (cmd->mrq->cap_cmd_during_tfr && cmd == cmd->mrq->cmd)
+		mmc_command_done(host->mmc, cmd->mrq);
+
+	/*
+	 * The host can send and interrupt when the busy state has
+	 * ended, allowing us to wait without wasting CPU cycles.
+	 * The busy signal uses DAT0 so this is similar to waiting
+	 * for data to complete.
+	 *
+	 * Note: The 1.0 specification is a bit ambiguous about this
+	 *       feature so there might be some problems with older
+	 *       controllers.
+	 */
+	if (cmd->flags & MMC_RSP_BUSY) {
+		if (cmd->data) {
+			DBG("Cannot wait for busy signal when also doing a data transfer");
+		} else if (!(host->quirks & SDHCI_QUIRK_NO_BUSY_IRQ) &&
+			   cmd == host->data_cmd) {
+			/* Command complete before busy is ended */
+			return;
+		}
+	}
+
+	/* Finished CMD23, now send actual command. */
+	if (cmd == cmd->mrq->sbc) {
+		if (!sdhci_send_command(host, cmd->mrq->cmd)) {
+			WARN_ON(host->deferred_cmd);
+			host->deferred_cmd = cmd->mrq->cmd;
+		}
+	} else {
+
+		/* Processed actual command. */
+		if (host->data && host->data_early)
+			sdhci_finish_data(host);
+
+		if (!cmd->data)
+			__sdhci_finish_mrq(host, cmd->mrq);
+	}
+}
+
+static u16 sdhci_get_preset_value(struct sdhci_host *host)
+{
+	u16 preset = 0;
+
+	switch (host->timing) {
+	case MMC_TIMING_MMC_HS:
+	case MMC_TIMING_SD_HS:
+		preset = sdhci_readw(host, SDHCI_PRESET_FOR_HIGH_SPEED);
+		break;
+	case MMC_TIMING_UHS_SDR12:
+		preset = sdhci_readw(host, SDHCI_PRESET_FOR_SDR12);
+		break;
+	case MMC_TIMING_UHS_SDR25:
+		preset = sdhci_readw(host, SDHCI_PRESET_FOR_SDR25);
+		break;
+	case MMC_TIMING_UHS_SDR50:
+		preset = sdhci_readw(host, SDHCI_PRESET_FOR_SDR50);
+		break;
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_MMC_HS200:
+		preset = sdhci_readw(host, SDHCI_PRESET_FOR_SDR104);
+		break;
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_MMC_DDR52:
+		preset = sdhci_readw(host, SDHCI_PRESET_FOR_DDR50);
+		break;
+	case MMC_TIMING_MMC_HS400:
+		preset = sdhci_readw(host, SDHCI_PRESET_FOR_HS400);
+		break;
+	default:
+		pr_warn("%s: Invalid UHS-I mode selected\n",
+			mmc_hostname(host->mmc));
+		preset = sdhci_readw(host, SDHCI_PRESET_FOR_SDR12);
+		break;
+	}
+	return preset;
+}
+
+u16 sdhci_calc_clk(struct sdhci_host *host, unsigned int clock,
+		   unsigned int *actual_clock)
+{
+	int div = 0; /* Initialized for compiler warning */
+	int real_div = div, clk_mul = 1;
+	u16 clk = 0;
+	bool switch_base_clk = false;
+
+	if (host->version >= SDHCI_SPEC_300) {
+		if (host->preset_enabled) {
+			u16 pre_val;
+
+			clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+			pre_val = sdhci_get_preset_value(host);
+			div = FIELD_GET(SDHCI_PRESET_SDCLK_FREQ_MASK, pre_val);
+			if (host->clk_mul &&
+				(pre_val & SDHCI_PRESET_CLKGEN_SEL)) {
+				clk = SDHCI_PROG_CLOCK_MODE;
+				real_div = div + 1;
+				clk_mul = host->clk_mul;
+			} else {
+				real_div = max_t(int, 1, div << 1);
+			}
+			goto clock_set;
+		}
+
+		/*
+		 * Check if the Host Controller supports Programmable Clock
+		 * Mode.
+		 */
+		if (host->clk_mul) {
+			for (div = 1; div <= 1024; div++) {
+				if ((host->max_clk * host->clk_mul / div)
+					<= clock)
+					break;
+			}
+			if ((host->max_clk * host->clk_mul / div) <= clock) {
+				/*
+				 * Set Programmable Clock Mode in the Clock
+				 * Control register.
+				 */
+				clk = SDHCI_PROG_CLOCK_MODE;
+				real_div = div;
+				clk_mul = host->clk_mul;
+				div--;
+			} else {
+				/*
+				 * Divisor can be too small to reach clock
+				 * speed requirement. Then use the base clock.
+				 */
+				switch_base_clk = true;
+			}
+		}
+
+		if (!host->clk_mul || switch_base_clk) {
+			/* Version 3.00 divisors must be a multiple of 2. */
+			if (host->max_clk <= clock)
+				div = 1;
+			else {
+				for (div = 2; div < SDHCI_MAX_DIV_SPEC_300;
+				     div += 2) {
+					if ((host->max_clk / div) <= clock)
+						break;
+				}
+			}
+			real_div = div;
+			div >>= 1;
+			if ((host->quirks2 & SDHCI_QUIRK2_CLOCK_DIV_ZERO_BROKEN)
+				&& !div && host->max_clk <= 25000000)
+				div = 1;
+		}
+	} else {
+		/* Version 2.00 divisors must be a power of 2. */
+		for (div = 1; div < SDHCI_MAX_DIV_SPEC_200; div *= 2) {
+			if ((host->max_clk / div) <= clock)
+				break;
+		}
+		real_div = div;
+		div >>= 1;
+	}
+
+clock_set:
+	if (real_div)
+		*actual_clock = (host->max_clk * clk_mul) / real_div;
+	clk |= (div & SDHCI_DIV_MASK) << SDHCI_DIVIDER_SHIFT;
+	clk |= ((div & SDHCI_DIV_HI_MASK) >> SDHCI_DIV_MASK_LEN)
+		<< SDHCI_DIVIDER_HI_SHIFT;
+
+	return clk;
+}
+EXPORT_SYMBOL_GPL(sdhci_calc_clk);
+
+void sdhci_enable_clk(struct sdhci_host *host, u16 clk)
+{
+	ktime_t timeout;
+
+	clk |= SDHCI_CLOCK_INT_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+	/* Wait max 150 ms */
+	timeout = ktime_add_ms(ktime_get(), 150);
+	while (1) {
+		bool timedout = ktime_after(ktime_get(), timeout);
+
+		clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+		if (clk & SDHCI_CLOCK_INT_STABLE)
+			break;
+		if (timedout) {
+			pr_err("%s: Internal clock never stabilised.\n",
+			       mmc_hostname(host->mmc));
+			sdhci_dumpregs(host);
+			return;
+		}
+		udelay(10);
+	}
+
+	if (host->version >= SDHCI_SPEC_410 && host->v4_mode) {
+		clk |= SDHCI_CLOCK_PLL_EN;
+		clk &= ~SDHCI_CLOCK_INT_STABLE;
+		sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+		/* Wait max 150 ms */
+		timeout = ktime_add_ms(ktime_get(), 150);
+		while (1) {
+			bool timedout = ktime_after(ktime_get(), timeout);
+
+			clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+			if (clk & SDHCI_CLOCK_INT_STABLE)
+				break;
+			if (timedout) {
+				pr_err("%s: PLL clock never stabilised.\n",
+				       mmc_hostname(host->mmc));
+				sdhci_dumpregs(host);
+				return;
+			}
+			udelay(10);
+		}
+	}
+
+	clk |= SDHCI_CLOCK_CARD_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+}
+EXPORT_SYMBOL_GPL(sdhci_enable_clk);
+
+void sdhci_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	u16 clk;
+
+	host->mmc->actual_clock = 0;
+
+	sdhci_writew(host, 0, SDHCI_CLOCK_CONTROL);
+
+	if (clock == 0)
+		return;
+
+	clk = sdhci_calc_clk(host, clock, &host->mmc->actual_clock);
+	sdhci_enable_clk(host, clk);
+}
+EXPORT_SYMBOL_GPL(sdhci_set_clock);
+
+static void sdhci_set_power_reg(struct sdhci_host *host, unsigned char mode,
+				unsigned short vdd)
+{
+	struct mmc_host *mmc = host->mmc;
+
+	mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, vdd);
+
+	if (mode != MMC_POWER_OFF)
+		sdhci_writeb(host, SDHCI_POWER_ON, SDHCI_POWER_CONTROL);
+	else
+		sdhci_writeb(host, 0, SDHCI_POWER_CONTROL);
+}
+
+void sdhci_set_power_noreg(struct sdhci_host *host, unsigned char mode,
+			   unsigned short vdd)
+{
+	u8 pwr = 0;
+
+	if (mode != MMC_POWER_OFF) {
+		switch (1 << vdd) {
+		case MMC_VDD_165_195:
+		/*
+		 * Without a regulator, SDHCI does not support 2.0v
+		 * so we only get here if the driver deliberately
+		 * added the 2.0v range to ocr_avail. Map it to 1.8v
+		 * for the purpose of turning on the power.
+		 */
+		case MMC_VDD_20_21:
+			pwr = SDHCI_POWER_180;
+			break;
+		case MMC_VDD_29_30:
+		case MMC_VDD_30_31:
+			pwr = SDHCI_POWER_300;
+			break;
+		case MMC_VDD_32_33:
+		case MMC_VDD_33_34:
+		/*
+		 * 3.4 ~ 3.6V are valid only for those platforms where it's
+		 * known that the voltage range is supported by hardware.
+		 */
+		case MMC_VDD_34_35:
+		case MMC_VDD_35_36:
+			pwr = SDHCI_POWER_330;
+			break;
+		default:
+			WARN(1, "%s: Invalid vdd %#x\n",
+			     mmc_hostname(host->mmc), vdd);
+			break;
+		}
+	}
+
+	if (host->pwr == pwr)
+		return;
+
+	host->pwr = pwr;
+
+	if (pwr == 0) {
+		sdhci_writeb(host, 0, SDHCI_POWER_CONTROL);
+		if (host->quirks2 & SDHCI_QUIRK2_CARD_ON_NEEDS_BUS_ON)
+			sdhci_runtime_pm_bus_off(host);
+	} else {
+		/*
+		 * Spec says that we should clear the power reg before setting
+		 * a new value. Some controllers don't seem to like this though.
+		 */
+		if (!(host->quirks & SDHCI_QUIRK_SINGLE_POWER_WRITE))
+			sdhci_writeb(host, 0, SDHCI_POWER_CONTROL);
+
+		/*
+		 * At least the Marvell CaFe chip gets confused if we set the
+		 * voltage and set turn on power at the same time, so set the
+		 * voltage first.
+		 */
+		if (host->quirks & SDHCI_QUIRK_NO_SIMULT_VDD_AND_POWER)
+			sdhci_writeb(host, pwr, SDHCI_POWER_CONTROL);
+
+		pwr |= SDHCI_POWER_ON;
+
+		sdhci_writeb(host, pwr, SDHCI_POWER_CONTROL);
+
+		if (host->quirks2 & SDHCI_QUIRK2_CARD_ON_NEEDS_BUS_ON)
+			sdhci_runtime_pm_bus_on(host);
+
+		/*
+		 * Some controllers need an extra 10ms delay of 10ms before
+		 * they can apply clock after applying power
+		 */
+		if (host->quirks & SDHCI_QUIRK_DELAY_AFTER_POWER)
+			mdelay(10);
+	}
+}
+EXPORT_SYMBOL_GPL(sdhci_set_power_noreg);
+
+void sdhci_set_power(struct sdhci_host *host, unsigned char mode,
+		     unsigned short vdd)
+{
+	if (IS_ERR(host->mmc->supply.vmmc))
+		sdhci_set_power_noreg(host, mode, vdd);
+	else
+		sdhci_set_power_reg(host, mode, vdd);
+}
+EXPORT_SYMBOL_GPL(sdhci_set_power);
+
+/*
+ * Some controllers need to configure a valid bus voltage on their power
+ * register regardless of whether an external regulator is taking care of power
+ * supply. This helper function takes care of it if set as the controller's
+ * sdhci_ops.set_power callback.
+ */
+void sdhci_set_power_and_bus_voltage(struct sdhci_host *host,
+				     unsigned char mode,
+				     unsigned short vdd)
+{
+	if (!IS_ERR(host->mmc->supply.vmmc)) {
+		struct mmc_host *mmc = host->mmc;
+
+		mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, vdd);
+	}
+	sdhci_set_power_noreg(host, mode, vdd);
+}
+EXPORT_SYMBOL_GPL(sdhci_set_power_and_bus_voltage);
+
+/*****************************************************************************\
+ *                                                                           *
+ * MMC callbacks                                                             *
+ *                                                                           *
+\*****************************************************************************/
+
+void sdhci_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct mmc_command *cmd;
+	unsigned long flags;
+	bool present;
+
+	/* Firstly check card presence */
+	present = mmc->ops->get_cd(mmc);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	sdhci_led_activate(host);
+
+	if (sdhci_present_error(host, mrq->cmd, present))
+		goto out_finish;
+
+	cmd = sdhci_manual_cmd23(host, mrq) ? mrq->sbc : mrq->cmd;
+
+	if (!sdhci_send_command_retry(host, cmd, flags))
+		goto out_finish;
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	return;
+
+out_finish:
+	sdhci_finish_mrq(host, mrq);
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+EXPORT_SYMBOL_GPL(sdhci_request);
+
+int sdhci_request_atomic(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct mmc_command *cmd;
+	unsigned long flags;
+	int ret = 0;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (sdhci_present_error(host, mrq->cmd, true)) {
+		sdhci_finish_mrq(host, mrq);
+		goto out_finish;
+	}
+
+	cmd = sdhci_manual_cmd23(host, mrq) ? mrq->sbc : mrq->cmd;
+
+	/*
+	 * The HSQ may send a command in interrupt context without polling
+	 * the busy signaling, which means we should return BUSY if controller
+	 * has not released inhibit bits to allow HSQ trying to send request
+	 * again in non-atomic context. So we should not finish this request
+	 * here.
+	 */
+	if (!sdhci_send_command(host, cmd))
+		ret = -EBUSY;
+	else
+		sdhci_led_activate(host);
+
+out_finish:
+	spin_unlock_irqrestore(&host->lock, flags);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(sdhci_request_atomic);
+
+void sdhci_set_bus_width(struct sdhci_host *host, int width)
+{
+	u8 ctrl;
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+	if (width == MMC_BUS_WIDTH_8) {
+		ctrl &= ~SDHCI_CTRL_4BITBUS;
+		ctrl |= SDHCI_CTRL_8BITBUS;
+	} else {
+		if (host->mmc->caps & MMC_CAP_8_BIT_DATA)
+			ctrl &= ~SDHCI_CTRL_8BITBUS;
+		if (width == MMC_BUS_WIDTH_4)
+			ctrl |= SDHCI_CTRL_4BITBUS;
+		else
+			ctrl &= ~SDHCI_CTRL_4BITBUS;
+	}
+	sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+}
+EXPORT_SYMBOL_GPL(sdhci_set_bus_width);
+
+void sdhci_set_uhs_signaling(struct sdhci_host *host, unsigned timing)
+{
+	u16 ctrl_2;
+
+	ctrl_2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	/* Select Bus Speed Mode for host */
+	ctrl_2 &= ~SDHCI_CTRL_UHS_MASK;
+	if ((timing == MMC_TIMING_MMC_HS200) ||
+	    (timing == MMC_TIMING_UHS_SDR104))
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR104;
+	else if (timing == MMC_TIMING_UHS_SDR12)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR12;
+	else if (timing == MMC_TIMING_UHS_SDR25)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR25;
+	else if (timing == MMC_TIMING_UHS_SDR50)
+		ctrl_2 |= SDHCI_CTRL_UHS_SDR50;
+	else if ((timing == MMC_TIMING_UHS_DDR50) ||
+		 (timing == MMC_TIMING_MMC_DDR52))
+		ctrl_2 |= SDHCI_CTRL_UHS_DDR50;
+	else if (timing == MMC_TIMING_MMC_HS400)
+		ctrl_2 |= SDHCI_CTRL_HS400; /* Non-standard */
+	sdhci_writew(host, ctrl_2, SDHCI_HOST_CONTROL2);
+}
+EXPORT_SYMBOL_GPL(sdhci_set_uhs_signaling);
+
+static bool sdhci_timing_has_preset(unsigned char timing)
+{
+	switch (timing) {
+	case MMC_TIMING_UHS_SDR12:
+	case MMC_TIMING_UHS_SDR25:
+	case MMC_TIMING_UHS_SDR50:
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_MMC_DDR52:
+		return true;
+	};
+	return false;
+}
+
+static bool sdhci_preset_needed(struct sdhci_host *host, unsigned char timing)
+{
+	return !(host->quirks2 & SDHCI_QUIRK2_PRESET_VALUE_BROKEN) &&
+	       sdhci_timing_has_preset(timing);
+}
+
+static bool sdhci_presetable_values_change(struct sdhci_host *host, struct mmc_ios *ios)
+{
+	/*
+	 * Preset Values are: Driver Strength, Clock Generator and SDCLK/RCLK
+	 * Frequency. Check if preset values need to be enabled, or the Driver
+	 * Strength needs updating. Note, clock changes are handled separately.
+	 */
+	return !host->preset_enabled &&
+	       (sdhci_preset_needed(host, ios->timing) || host->drv_type != ios->drv_type);
+}
+
+void sdhci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	bool reinit_uhs = host->reinit_uhs;
+	bool turning_on_clk = false;
+	u8 ctrl;
+
+	host->reinit_uhs = false;
+
+	if (ios->power_mode == MMC_POWER_UNDEFINED)
+		return;
+
+	if (host->flags & SDHCI_DEVICE_DEAD) {
+		if (!IS_ERR(mmc->supply.vmmc) &&
+		    ios->power_mode == MMC_POWER_OFF)
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+		return;
+	}
+
+	/*
+	 * Reset the chip on each power off.
+	 * Should clear out any weird states.
+	 */
+	if (ios->power_mode == MMC_POWER_OFF) {
+		sdhci_writel(host, 0, SDHCI_SIGNAL_ENABLE);
+		sdhci_reinit(host);
+	}
+
+	if (host->version >= SDHCI_SPEC_300 &&
+		(ios->power_mode == MMC_POWER_UP) &&
+		!(host->quirks2 & SDHCI_QUIRK2_PRESET_VALUE_BROKEN))
+		sdhci_enable_preset_value(host, false);
+
+	if (!ios->clock || ios->clock != host->clock) {
+		turning_on_clk = ios->clock && !host->clock;
+
+		host->ops->set_clock(host, ios->clock);
+		host->clock = ios->clock;
+
+		if (host->quirks & SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK &&
+		    host->clock) {
+			host->timeout_clk = host->mmc->actual_clock ?
+						host->mmc->actual_clock / 1000 :
+						host->clock / 1000;
+			host->mmc->max_busy_timeout =
+				host->ops->get_max_timeout_count ?
+				host->ops->get_max_timeout_count(host) :
+				1 << 27;
+			host->mmc->max_busy_timeout /= host->timeout_clk;
+		}
+	}
+
+	if (host->ops->set_power)
+		host->ops->set_power(host, ios->power_mode, ios->vdd);
+	else
+		sdhci_set_power(host, ios->power_mode, ios->vdd);
+
+	if (host->ops->platform_send_init_74_clocks)
+		host->ops->platform_send_init_74_clocks(host, ios->power_mode);
+
+	host->ops->set_bus_width(host, ios->bus_width);
+
+	/*
+	 * Special case to avoid multiple clock changes during voltage
+	 * switching.
+	 */
+	if (!reinit_uhs &&
+	    turning_on_clk &&
+	    host->timing == ios->timing &&
+	    host->version >= SDHCI_SPEC_300 &&
+	    !sdhci_presetable_values_change(host, ios))
+		return;
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+
+	if (!(host->quirks & SDHCI_QUIRK_NO_HISPD_BIT)) {
+		if (ios->timing == MMC_TIMING_SD_HS ||
+		     ios->timing == MMC_TIMING_MMC_HS ||
+		     ios->timing == MMC_TIMING_MMC_HS400 ||
+		     ios->timing == MMC_TIMING_MMC_HS200 ||
+		     ios->timing == MMC_TIMING_MMC_DDR52 ||
+		     ios->timing == MMC_TIMING_UHS_SDR50 ||
+		     ios->timing == MMC_TIMING_UHS_SDR104 ||
+		     ios->timing == MMC_TIMING_UHS_DDR50 ||
+		     ios->timing == MMC_TIMING_UHS_SDR25)
+			ctrl |= SDHCI_CTRL_HISPD;
+		else
+			ctrl &= ~SDHCI_CTRL_HISPD;
+	}
+
+	if (host->version >= SDHCI_SPEC_300) {
+		u16 clk, ctrl_2;
+
+		if (!host->preset_enabled) {
+			sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+			/*
+			 * We only need to set Driver Strength if the
+			 * preset value enable is not set.
+			 */
+			ctrl_2 = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+			ctrl_2 &= ~SDHCI_CTRL_DRV_TYPE_MASK;
+			if (ios->drv_type == MMC_SET_DRIVER_TYPE_A)
+				ctrl_2 |= SDHCI_CTRL_DRV_TYPE_A;
+			else if (ios->drv_type == MMC_SET_DRIVER_TYPE_B)
+				ctrl_2 |= SDHCI_CTRL_DRV_TYPE_B;
+			else if (ios->drv_type == MMC_SET_DRIVER_TYPE_C)
+				ctrl_2 |= SDHCI_CTRL_DRV_TYPE_C;
+			else if (ios->drv_type == MMC_SET_DRIVER_TYPE_D)
+				ctrl_2 |= SDHCI_CTRL_DRV_TYPE_D;
+			else {
+				pr_warn("%s: invalid driver type, default to driver type B\n",
+					mmc_hostname(mmc));
+				ctrl_2 |= SDHCI_CTRL_DRV_TYPE_B;
+			}
+
+			sdhci_writew(host, ctrl_2, SDHCI_HOST_CONTROL2);
+			host->drv_type = ios->drv_type;
+		} else {
+			/*
+			 * According to SDHC Spec v3.00, if the Preset Value
+			 * Enable in the Host Control 2 register is set, we
+			 * need to reset SD Clock Enable before changing High
+			 * Speed Enable to avoid generating clock gliches.
+			 */
+
+			/* Reset SD Clock Enable */
+			clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+			clk &= ~SDHCI_CLOCK_CARD_EN;
+			sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+			sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+
+			/* Re-enable SD Clock */
+			host->ops->set_clock(host, host->clock);
+		}
+
+		/* Reset SD Clock Enable */
+		clk = sdhci_readw(host, SDHCI_CLOCK_CONTROL);
+		clk &= ~SDHCI_CLOCK_CARD_EN;
+		sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+
+		host->ops->set_uhs_signaling(host, ios->timing);
+		host->timing = ios->timing;
+
+		if (sdhci_preset_needed(host, ios->timing)) {
+			u16 preset;
+
+			sdhci_enable_preset_value(host, true);
+			preset = sdhci_get_preset_value(host);
+			ios->drv_type = FIELD_GET(SDHCI_PRESET_DRV_MASK,
+						  preset);
+			host->drv_type = ios->drv_type;
+		}
+
+		/* Re-enable SD Clock */
+		host->ops->set_clock(host, host->clock);
+	} else
+		sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+
+	/*
+	 * Some (ENE) controllers go apeshit on some ios operation,
+	 * signalling timeout and CRC errors even on CMD0. Resetting
+	 * it on each ios seems to solve the problem.
+	 */
+	if (host->quirks & SDHCI_QUIRK_RESET_CMD_DATA_ON_IOS)
+		sdhci_do_reset(host, SDHCI_RESET_CMD | SDHCI_RESET_DATA);
+}
+EXPORT_SYMBOL_GPL(sdhci_set_ios);
+
+static int sdhci_get_cd(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	int gpio_cd = mmc_gpio_get_cd(mmc);
+
+	if (host->flags & SDHCI_DEVICE_DEAD)
+		return 0;
+
+	/* If nonremovable, assume that the card is always present. */
+	if (!mmc_card_is_removable(host->mmc))
+		return 1;
+
+	/*
+	 * Try slot gpio detect, if defined it take precedence
+	 * over build in controller functionality
+	 */
+	if (gpio_cd >= 0)
+		return !!gpio_cd;
+
+	/* If polling, assume that the card is always present. */
+	if (host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION)
+		return 1;
+
+	/* Host native card detect */
+	return !!(sdhci_readl(host, SDHCI_PRESENT_STATE) & SDHCI_CARD_PRESENT);
+}
+
+static int sdhci_check_ro(struct sdhci_host *host)
+{
+	unsigned long flags;
+	int is_readonly;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (host->flags & SDHCI_DEVICE_DEAD)
+		is_readonly = 0;
+	else if (host->ops->get_ro)
+		is_readonly = host->ops->get_ro(host);
+	else if (mmc_can_gpio_ro(host->mmc))
+		is_readonly = mmc_gpio_get_ro(host->mmc);
+	else
+		is_readonly = !(sdhci_readl(host, SDHCI_PRESENT_STATE)
+				& SDHCI_WRITE_PROTECT);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	/* This quirk needs to be replaced by a callback-function later */
+	return host->quirks & SDHCI_QUIRK_INVERTED_WRITE_PROTECT ?
+		!is_readonly : is_readonly;
+}
+
+#define SAMPLE_COUNT	5
+
+static int sdhci_get_ro(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	int i, ro_count;
+
+	if (!(host->quirks & SDHCI_QUIRK_UNSTABLE_RO_DETECT))
+		return sdhci_check_ro(host);
+
+	ro_count = 0;
+	for (i = 0; i < SAMPLE_COUNT; i++) {
+		if (sdhci_check_ro(host)) {
+			if (++ro_count > SAMPLE_COUNT / 2)
+				return 1;
+		}
+		msleep(30);
+	}
+	return 0;
+}
+
+static void sdhci_hw_reset(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	if (host->ops && host->ops->hw_reset)
+		host->ops->hw_reset(host);
+}
+
+static void sdhci_enable_sdio_irq_nolock(struct sdhci_host *host, int enable)
+{
+	if (!(host->flags & SDHCI_DEVICE_DEAD)) {
+		if (enable)
+			host->ier |= SDHCI_INT_CARD_INT;
+		else
+			host->ier &= ~SDHCI_INT_CARD_INT;
+
+		sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+		sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+	}
+}
+
+void sdhci_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+
+	if (enable)
+		pm_runtime_get_noresume(host->mmc->parent);
+
+	spin_lock_irqsave(&host->lock, flags);
+	sdhci_enable_sdio_irq_nolock(host, enable);
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (!enable)
+		pm_runtime_put_noidle(host->mmc->parent);
+}
+EXPORT_SYMBOL_GPL(sdhci_enable_sdio_irq);
+
+static void sdhci_ack_sdio_irq(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	sdhci_enable_sdio_irq_nolock(host, true);
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+int sdhci_start_signal_voltage_switch(struct mmc_host *mmc,
+				      struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	u16 ctrl;
+	int ret;
+
+	/*
+	 * Signal Voltage Switching is only applicable for Host Controllers
+	 * v3.00 and above.
+	 */
+	if (host->version < SDHCI_SPEC_300)
+		return 0;
+
+	ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+
+	switch (ios->signal_voltage) {
+	case MMC_SIGNAL_VOLTAGE_330:
+		if (!(host->flags & SDHCI_SIGNALING_330))
+			return -EINVAL;
+		/* Set 1.8V Signal Enable in the Host Control2 register to 0 */
+		ctrl &= ~SDHCI_CTRL_VDD_180;
+		sdhci_writew(host, ctrl, SDHCI_HOST_CONTROL2);
+
+		if (!IS_ERR(mmc->supply.vqmmc)) {
+			ret = mmc_regulator_set_vqmmc(mmc, ios);
+			if (ret < 0) {
+				pr_warn("%s: Switching to 3.3V signalling voltage failed\n",
+					mmc_hostname(mmc));
+				return -EIO;
+			}
+		}
+		/* Wait for 5ms */
+		usleep_range(5000, 5500);
+
+		/* 3.3V regulator output should be stable within 5 ms */
+		ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+		if (!(ctrl & SDHCI_CTRL_VDD_180))
+			return 0;
+
+		pr_warn("%s: 3.3V regulator output did not become stable\n",
+			mmc_hostname(mmc));
+
+		return -EAGAIN;
+	case MMC_SIGNAL_VOLTAGE_180:
+		if (!(host->flags & SDHCI_SIGNALING_180))
+			return -EINVAL;
+		if (!IS_ERR(mmc->supply.vqmmc)) {
+			ret = mmc_regulator_set_vqmmc(mmc, ios);
+			if (ret < 0) {
+				pr_warn("%s: Switching to 1.8V signalling voltage failed\n",
+					mmc_hostname(mmc));
+				return -EIO;
+			}
+		}
+
+		/*
+		 * Enable 1.8V Signal Enable in the Host Control2
+		 * register
+		 */
+		ctrl |= SDHCI_CTRL_VDD_180;
+		sdhci_writew(host, ctrl, SDHCI_HOST_CONTROL2);
+
+		/* Some controller need to do more when switching */
+		if (host->ops->voltage_switch)
+			host->ops->voltage_switch(host);
+
+		/* 1.8V regulator output should be stable within 5 ms */
+		ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+		if (ctrl & SDHCI_CTRL_VDD_180)
+			return 0;
+
+		pr_warn("%s: 1.8V regulator output did not become stable\n",
+			mmc_hostname(mmc));
+
+		return -EAGAIN;
+	case MMC_SIGNAL_VOLTAGE_120:
+		if (!(host->flags & SDHCI_SIGNALING_120))
+			return -EINVAL;
+		if (!IS_ERR(mmc->supply.vqmmc)) {
+			ret = mmc_regulator_set_vqmmc(mmc, ios);
+			if (ret < 0) {
+				pr_warn("%s: Switching to 1.2V signalling voltage failed\n",
+					mmc_hostname(mmc));
+				return -EIO;
+			}
+		}
+		return 0;
+	default:
+		/* No signal voltage switch required */
+		return 0;
+	}
+}
+EXPORT_SYMBOL_GPL(sdhci_start_signal_voltage_switch);
+
+static int sdhci_card_busy(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	u32 present_state;
+
+	/* Check whether DAT[0] is 0 */
+	present_state = sdhci_readl(host, SDHCI_PRESENT_STATE);
+
+	return !(present_state & SDHCI_DATA_0_LVL_MASK);
+}
+
+static int sdhci_prepare_hs400_tuning(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	host->flags |= SDHCI_HS400_TUNING;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	return 0;
+}
+
+void sdhci_start_tuning(struct sdhci_host *host)
+{
+	u16 ctrl;
+
+	ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	ctrl |= SDHCI_CTRL_EXEC_TUNING;
+	if (host->quirks2 & SDHCI_QUIRK2_TUNING_WORK_AROUND)
+		ctrl |= SDHCI_CTRL_TUNED_CLK;
+	sdhci_writew(host, ctrl, SDHCI_HOST_CONTROL2);
+
+	/*
+	 * As per the Host Controller spec v3.00, tuning command
+	 * generates Buffer Read Ready interrupt, so enable that.
+	 *
+	 * Note: The spec clearly says that when tuning sequence
+	 * is being performed, the controller does not generate
+	 * interrupts other than Buffer Read Ready interrupt. But
+	 * to make sure we don't hit a controller bug, we _only_
+	 * enable Buffer Read Ready interrupt here.
+	 */
+	sdhci_writel(host, SDHCI_INT_DATA_AVAIL, SDHCI_INT_ENABLE);
+	sdhci_writel(host, SDHCI_INT_DATA_AVAIL, SDHCI_SIGNAL_ENABLE);
+}
+EXPORT_SYMBOL_GPL(sdhci_start_tuning);
+
+void sdhci_end_tuning(struct sdhci_host *host)
+{
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+}
+EXPORT_SYMBOL_GPL(sdhci_end_tuning);
+
+void sdhci_reset_tuning(struct sdhci_host *host)
+{
+	u16 ctrl;
+
+	ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+	ctrl &= ~SDHCI_CTRL_TUNED_CLK;
+	ctrl &= ~SDHCI_CTRL_EXEC_TUNING;
+	sdhci_writew(host, ctrl, SDHCI_HOST_CONTROL2);
+}
+EXPORT_SYMBOL_GPL(sdhci_reset_tuning);
+
+void sdhci_abort_tuning(struct sdhci_host *host, u32 opcode)
+{
+	sdhci_reset_tuning(host);
+
+	sdhci_do_reset(host, SDHCI_RESET_CMD);
+	sdhci_do_reset(host, SDHCI_RESET_DATA);
+
+	sdhci_end_tuning(host);
+
+	mmc_abort_tuning(host->mmc, opcode);
+}
+EXPORT_SYMBOL_GPL(sdhci_abort_tuning);
+
+/*
+ * We use sdhci_send_tuning() because mmc_send_tuning() is not a good fit. SDHCI
+ * tuning command does not have a data payload (or rather the hardware does it
+ * automatically) so mmc_send_tuning() will return -EIO. Also the tuning command
+ * interrupt setup is different to other commands and there is no timeout
+ * interrupt so special handling is needed.
+ */
+void sdhci_send_tuning(struct sdhci_host *host, u32 opcode)
+{
+	struct mmc_host *mmc = host->mmc;
+	struct mmc_command cmd = {};
+	struct mmc_request mrq = {};
+	unsigned long flags;
+	u32 b = host->sdma_boundary;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	cmd.opcode = opcode;
+	cmd.flags = MMC_RSP_R1 | MMC_CMD_ADTC;
+	cmd.mrq = &mrq;
+
+	mrq.cmd = &cmd;
+	/*
+	 * In response to CMD19, the card sends 64 bytes of tuning
+	 * block to the Host Controller. So we set the block size
+	 * to 64 here.
+	 */
+	if (cmd.opcode == MMC_SEND_TUNING_BLOCK_HS200 &&
+	    mmc->ios.bus_width == MMC_BUS_WIDTH_8)
+		sdhci_writew(host, SDHCI_MAKE_BLKSZ(b, 128), SDHCI_BLOCK_SIZE);
+	else
+		sdhci_writew(host, SDHCI_MAKE_BLKSZ(b, 64), SDHCI_BLOCK_SIZE);
+
+	/*
+	 * The tuning block is sent by the card to the host controller.
+	 * So we set the TRNS_READ bit in the Transfer Mode register.
+	 * This also takes care of setting DMA Enable and Multi Block
+	 * Select in the same register to 0.
+	 */
+	sdhci_writew(host, SDHCI_TRNS_READ, SDHCI_TRANSFER_MODE);
+
+	if (!sdhci_send_command_retry(host, &cmd, flags)) {
+		spin_unlock_irqrestore(&host->lock, flags);
+		host->tuning_done = 0;
+		return;
+	}
+
+	host->cmd = NULL;
+
+	sdhci_del_timer(host, &mrq);
+
+	host->tuning_done = 0;
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	/* Wait for Buffer Read Ready interrupt */
+	wait_event_timeout(host->buf_ready_int, (host->tuning_done == 1),
+			   msecs_to_jiffies(50));
+
+}
+EXPORT_SYMBOL_GPL(sdhci_send_tuning);
+
+static int __sdhci_execute_tuning(struct sdhci_host *host, u32 opcode)
+{
+	int i;
+
+	/*
+	 * Issue opcode repeatedly till Execute Tuning is set to 0 or the number
+	 * of loops reaches tuning loop count.
+	 */
+	for (i = 0; i < host->tuning_loop_count; i++) {
+		u16 ctrl;
+
+		sdhci_send_tuning(host, opcode);
+
+		if (!host->tuning_done) {
+			pr_debug("%s: Tuning timeout, falling back to fixed sampling clock\n",
+				 mmc_hostname(host->mmc));
+			sdhci_abort_tuning(host, opcode);
+			return -ETIMEDOUT;
+		}
+
+		/* Spec does not require a delay between tuning cycles */
+		if (host->tuning_delay > 0)
+			mdelay(host->tuning_delay);
+
+		ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+		if (!(ctrl & SDHCI_CTRL_EXEC_TUNING)) {
+			if (ctrl & SDHCI_CTRL_TUNED_CLK)
+				return 0; /* Success! */
+			break;
+		}
+
+	}
+
+	pr_info("%s: Tuning failed, falling back to fixed sampling clock\n",
+		mmc_hostname(host->mmc));
+	sdhci_reset_tuning(host);
+	return -EAGAIN;
+}
+
+int sdhci_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	int err = 0;
+	unsigned int tuning_count = 0;
+	bool hs400_tuning;
+
+	hs400_tuning = host->flags & SDHCI_HS400_TUNING;
+
+	if (host->tuning_mode == SDHCI_TUNING_MODE_1)
+		tuning_count = host->tuning_count;
+
+	/*
+	 * The Host Controller needs tuning in case of SDR104 and DDR50
+	 * mode, and for SDR50 mode when Use Tuning for SDR50 is set in
+	 * the Capabilities register.
+	 * If the Host Controller supports the HS200 mode then the
+	 * tuning function has to be executed.
+	 */
+	switch (host->timing) {
+	/* HS400 tuning is done in HS200 mode */
+	case MMC_TIMING_MMC_HS400:
+		err = -EINVAL;
+		goto out;
+
+	case MMC_TIMING_MMC_HS200:
+		/*
+		 * Periodic re-tuning for HS400 is not expected to be needed, so
+		 * disable it here.
+		 */
+		if (hs400_tuning)
+			tuning_count = 0;
+		break;
+
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_UHS_DDR50:
+		break;
+
+	case MMC_TIMING_UHS_SDR50:
+		if (host->flags & SDHCI_SDR50_NEEDS_TUNING)
+			break;
+		fallthrough;
+
+	default:
+		goto out;
+	}
+
+	if (host->ops->platform_execute_tuning) {
+		err = host->ops->platform_execute_tuning(host, opcode);
+		goto out;
+	}
+
+	host->mmc->retune_period = tuning_count;
+
+	if (host->tuning_delay < 0)
+		host->tuning_delay = opcode == MMC_SEND_TUNING_BLOCK;
+
+	sdhci_start_tuning(host);
+
+	host->tuning_err = __sdhci_execute_tuning(host, opcode);
+
+	sdhci_end_tuning(host);
+out:
+	host->flags &= ~SDHCI_HS400_TUNING;
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(sdhci_execute_tuning);
+
+static void sdhci_enable_preset_value(struct sdhci_host *host, bool enable)
+{
+	/* Host Controller v3.00 defines preset value registers */
+	if (host->version < SDHCI_SPEC_300)
+		return;
+
+	/*
+	 * We only enable or disable Preset Value if they are not already
+	 * enabled or disabled respectively. Otherwise, we bail out.
+	 */
+	if (host->preset_enabled != enable) {
+		u16 ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+
+		if (enable)
+			ctrl |= SDHCI_CTRL_PRESET_VAL_ENABLE;
+		else
+			ctrl &= ~SDHCI_CTRL_PRESET_VAL_ENABLE;
+
+		sdhci_writew(host, ctrl, SDHCI_HOST_CONTROL2);
+
+		if (enable)
+			host->flags |= SDHCI_PV_ENABLED;
+		else
+			host->flags &= ~SDHCI_PV_ENABLED;
+
+		host->preset_enabled = enable;
+	}
+}
+
+static void sdhci_post_req(struct mmc_host *mmc, struct mmc_request *mrq,
+				int err)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+
+	if (data->host_cookie != COOKIE_UNMAPPED)
+		dma_unmap_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+			     mmc_get_dma_dir(data));
+
+	data->host_cookie = COOKIE_UNMAPPED;
+}
+
+static void sdhci_pre_req(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	mrq->data->host_cookie = COOKIE_UNMAPPED;
+
+	/*
+	 * No pre-mapping in the pre hook if we're using the bounce buffer,
+	 * for that we would need two bounce buffers since one buffer is
+	 * in flight when this is getting called.
+	 */
+	if (host->flags & SDHCI_REQ_USE_DMA && !host->bounce_buffer)
+		sdhci_pre_dma_transfer(host, mrq->data, COOKIE_PRE_MAPPED);
+}
+
+static void sdhci_error_out_mrqs(struct sdhci_host *host, int err)
+{
+	if (host->data_cmd) {
+		host->data_cmd->error = err;
+		sdhci_finish_mrq(host, host->data_cmd->mrq);
+	}
+
+	if (host->cmd) {
+		host->cmd->error = err;
+		sdhci_finish_mrq(host, host->cmd->mrq);
+	}
+}
+
+static void sdhci_card_event(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+	int present;
+
+	/* First check if client has provided their own card event */
+	if (host->ops->card_event)
+		host->ops->card_event(host);
+
+	present = mmc->ops->get_cd(mmc);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	/* Check sdhci_has_requests() first in case we are runtime suspended */
+	if (sdhci_has_requests(host) && !present) {
+		pr_err("%s: Card removed during transfer!\n",
+			mmc_hostname(host->mmc));
+		pr_err("%s: Resetting controller.\n",
+			mmc_hostname(host->mmc));
+
+		sdhci_do_reset(host, SDHCI_RESET_CMD);
+		sdhci_do_reset(host, SDHCI_RESET_DATA);
+
+		sdhci_error_out_mrqs(host, -ENOMEDIUM);
+	}
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static const struct mmc_host_ops sdhci_ops = {
+	.request	= sdhci_request,
+	.post_req	= sdhci_post_req,
+	.pre_req	= sdhci_pre_req,
+	.set_ios	= sdhci_set_ios,
+	.get_cd		= sdhci_get_cd,
+	.get_ro		= sdhci_get_ro,
+	.hw_reset	= sdhci_hw_reset,
+	.enable_sdio_irq = sdhci_enable_sdio_irq,
+	.ack_sdio_irq    = sdhci_ack_sdio_irq,
+	.start_signal_voltage_switch	= sdhci_start_signal_voltage_switch,
+	.prepare_hs400_tuning		= sdhci_prepare_hs400_tuning,
+	.execute_tuning			= sdhci_execute_tuning,
+	.card_event			= sdhci_card_event,
+	.card_busy	= sdhci_card_busy,
+};
+
+/*****************************************************************************\
+ *                                                                           *
+ * Request done                                                              *
+ *                                                                           *
+\*****************************************************************************/
+
+static bool sdhci_request_done(struct sdhci_host *host)
+{
+	unsigned long flags;
+	struct mmc_request *mrq;
+	int i;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	for (i = 0; i < SDHCI_MAX_MRQS; i++) {
+		mrq = host->mrqs_done[i];
+		if (mrq)
+			break;
+	}
+
+	if (!mrq) {
+		spin_unlock_irqrestore(&host->lock, flags);
+		return true;
+	}
+
+	/*
+	 * The controller needs a reset of internal state machines
+	 * upon error conditions.
+	 */
+	if (sdhci_needs_reset(host, mrq)) {
+		/*
+		 * Do not finish until command and data lines are available for
+		 * reset. Note there can only be one other mrq, so it cannot
+		 * also be in mrqs_done, otherwise host->cmd and host->data_cmd
+		 * would both be null.
+		 */
+		if (host->cmd || host->data_cmd) {
+			spin_unlock_irqrestore(&host->lock, flags);
+			return true;
+		}
+
+		/* Some controllers need this kick or reset won't work here */
+		if (host->quirks & SDHCI_QUIRK_CLOCK_BEFORE_RESET)
+			/* This is to force an update */
+			host->ops->set_clock(host, host->clock);
+
+		/*
+		 * Spec says we should do both at the same time, but Ricoh
+		 * controllers do not like that.
+		 */
+		sdhci_do_reset(host, SDHCI_RESET_CMD);
+		sdhci_do_reset(host, SDHCI_RESET_DATA);
+
+		host->pending_reset = false;
+	}
+
+	/*
+	 * Always unmap the data buffers if they were mapped by
+	 * sdhci_prepare_data() whenever we finish with a request.
+	 * This avoids leaking DMA mappings on error.
+	 */
+	if (host->flags & SDHCI_REQ_USE_DMA) {
+		struct mmc_data *data = mrq->data;
+
+		if (host->use_external_dma && data &&
+		    (mrq->cmd->error || data->error)) {
+			struct dma_chan *chan = sdhci_external_dma_channel(host, data);
+
+			host->mrqs_done[i] = NULL;
+			spin_unlock_irqrestore(&host->lock, flags);
+			dmaengine_terminate_sync(chan);
+			spin_lock_irqsave(&host->lock, flags);
+			sdhci_set_mrq_done(host, mrq);
+		}
+
+		if (data && data->host_cookie == COOKIE_MAPPED) {
+			if (host->bounce_buffer) {
+				/*
+				 * On reads, copy the bounced data into the
+				 * sglist
+				 */
+				if (mmc_get_dma_dir(data) == DMA_FROM_DEVICE) {
+					unsigned int length = data->bytes_xfered;
+
+					if (length > host->bounce_buffer_size) {
+						pr_err("%s: bounce buffer is %u bytes but DMA claims to have transferred %u bytes\n",
+						       mmc_hostname(host->mmc),
+						       host->bounce_buffer_size,
+						       data->bytes_xfered);
+						/* Cap it down and continue */
+						length = host->bounce_buffer_size;
+					}
+					dma_sync_single_for_cpu(
+						host->mmc->parent,
+						host->bounce_addr,
+						host->bounce_buffer_size,
+						DMA_FROM_DEVICE);
+					sg_copy_from_buffer(data->sg,
+						data->sg_len,
+						host->bounce_buffer,
+						length);
+				} else {
+					/* No copying, just switch ownership */
+					dma_sync_single_for_cpu(
+						host->mmc->parent,
+						host->bounce_addr,
+						host->bounce_buffer_size,
+						mmc_get_dma_dir(data));
+				}
+			} else {
+				/* Unmap the raw data */
+				dma_unmap_sg(mmc_dev(host->mmc), data->sg,
+					     data->sg_len,
+					     mmc_get_dma_dir(data));
+			}
+			data->host_cookie = COOKIE_UNMAPPED;
+		}
+	}
+
+	host->mrqs_done[i] = NULL;
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (host->ops->request_done)
+		host->ops->request_done(host, mrq);
+	else
+		mmc_request_done(host->mmc, mrq);
+
+	return false;
+}
+
+static void sdhci_complete_work(struct work_struct *work)
+{
+	struct sdhci_host *host = container_of(work, struct sdhci_host,
+					       complete_work);
+
+	while (!sdhci_request_done(host))
+		;
+}
+
+static void sdhci_timeout_timer(struct timer_list *t)
+{
+	struct sdhci_host *host;
+	unsigned long flags;
+
+	host = from_timer(host, t, timer);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (host->cmd && !sdhci_data_line_cmd(host->cmd)) {
+		pr_err("%s: Timeout waiting for hardware cmd interrupt.\n",
+		       mmc_hostname(host->mmc));
+		sdhci_dumpregs(host);
+
+		host->cmd->error = -ETIMEDOUT;
+		sdhci_finish_mrq(host, host->cmd->mrq);
+	}
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void sdhci_timeout_data_timer(struct timer_list *t)
+{
+	struct sdhci_host *host;
+	unsigned long flags;
+
+	host = from_timer(host, t, data_timer);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (host->data || host->data_cmd ||
+	    (host->cmd && sdhci_data_line_cmd(host->cmd))) {
+		pr_err("%s: Timeout waiting for hardware interrupt.\n",
+		       mmc_hostname(host->mmc));
+		sdhci_dumpregs(host);
+
+		if (host->data) {
+			host->data->error = -ETIMEDOUT;
+			__sdhci_finish_data(host, true);
+			queue_work(host->complete_wq, &host->complete_work);
+		} else if (host->data_cmd) {
+			host->data_cmd->error = -ETIMEDOUT;
+			sdhci_finish_mrq(host, host->data_cmd->mrq);
+		} else {
+			host->cmd->error = -ETIMEDOUT;
+			sdhci_finish_mrq(host, host->cmd->mrq);
+		}
+	}
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+/*****************************************************************************\
+ *                                                                           *
+ * Interrupt handling                                                        *
+ *                                                                           *
+\*****************************************************************************/
+
+static void sdhci_cmd_irq(struct sdhci_host *host, u32 intmask, u32 *intmask_p)
+{
+	/* Handle auto-CMD12 error */
+	if (intmask & SDHCI_INT_AUTO_CMD_ERR && host->data_cmd) {
+		struct mmc_request *mrq = host->data_cmd->mrq;
+		u16 auto_cmd_status = sdhci_readw(host, SDHCI_AUTO_CMD_STATUS);
+		int data_err_bit = (auto_cmd_status & SDHCI_AUTO_CMD_TIMEOUT) ?
+				   SDHCI_INT_DATA_TIMEOUT :
+				   SDHCI_INT_DATA_CRC;
+
+		/* Treat auto-CMD12 error the same as data error */
+		if (!mrq->sbc && (host->flags & SDHCI_AUTO_CMD12)) {
+			*intmask_p |= data_err_bit;
+			return;
+		}
+	}
+
+	if (!host->cmd) {
+		/*
+		 * SDHCI recovers from errors by resetting the cmd and data
+		 * circuits.  Until that is done, there very well might be more
+		 * interrupts, so ignore them in that case.
+		 */
+		if (host->pending_reset)
+			return;
+		pr_err("%s: Got command interrupt 0x%08x even though no command operation was in progress.\n",
+		       mmc_hostname(host->mmc), (unsigned)intmask);
+		sdhci_dumpregs(host);
+		return;
+	}
+
+	if (intmask & (SDHCI_INT_TIMEOUT | SDHCI_INT_CRC |
+		       SDHCI_INT_END_BIT | SDHCI_INT_INDEX)) {
+		if (intmask & SDHCI_INT_TIMEOUT)
+			host->cmd->error = -ETIMEDOUT;
+		else
+			host->cmd->error = -EILSEQ;
+
+		/* Treat data command CRC error the same as data CRC error */
+		if (host->cmd->data &&
+		    (intmask & (SDHCI_INT_CRC | SDHCI_INT_TIMEOUT)) ==
+		     SDHCI_INT_CRC) {
+			host->cmd = NULL;
+			*intmask_p |= SDHCI_INT_DATA_CRC;
+			return;
+		}
+
+		__sdhci_finish_mrq(host, host->cmd->mrq);
+		return;
+	}
+
+	/* Handle auto-CMD23 error */
+	if (intmask & SDHCI_INT_AUTO_CMD_ERR) {
+		struct mmc_request *mrq = host->cmd->mrq;
+		u16 auto_cmd_status = sdhci_readw(host, SDHCI_AUTO_CMD_STATUS);
+		int err = (auto_cmd_status & SDHCI_AUTO_CMD_TIMEOUT) ?
+			  -ETIMEDOUT :
+			  -EILSEQ;
+
+		if (mrq->sbc && (host->flags & SDHCI_AUTO_CMD23)) {
+			mrq->sbc->error = err;
+			__sdhci_finish_mrq(host, mrq);
+			return;
+		}
+	}
+
+	if (intmask & SDHCI_INT_RESPONSE)
+		sdhci_finish_command(host);
+}
+
+static void sdhci_adma_show_error(struct sdhci_host *host)
+{
+	void *desc = host->adma_table;
+	dma_addr_t dma = host->adma_addr;
+
+	sdhci_dumpregs(host);
+
+	while (true) {
+		struct sdhci_adma2_64_desc *dma_desc = desc;
+
+		if (host->flags & SDHCI_USE_64_BIT_DMA)
+			SDHCI_DUMP("%08llx: DMA 0x%08x%08x, LEN 0x%04x, Attr=0x%02x\n",
+			    (unsigned long long)dma,
+			    le32_to_cpu(dma_desc->addr_hi),
+			    le32_to_cpu(dma_desc->addr_lo),
+			    le16_to_cpu(dma_desc->len),
+			    le16_to_cpu(dma_desc->cmd));
+		else
+			SDHCI_DUMP("%08llx: DMA 0x%08x, LEN 0x%04x, Attr=0x%02x\n",
+			    (unsigned long long)dma,
+			    le32_to_cpu(dma_desc->addr_lo),
+			    le16_to_cpu(dma_desc->len),
+			    le16_to_cpu(dma_desc->cmd));
+
+		desc += host->desc_sz;
+		dma += host->desc_sz;
+
+		if (dma_desc->cmd & cpu_to_le16(ADMA2_END))
+			break;
+	}
+}
+
+static void sdhci_data_irq(struct sdhci_host *host, u32 intmask)
+{
+	u32 command;
+
+	/* CMD19 generates _only_ Buffer Read Ready interrupt */
+	if (intmask & SDHCI_INT_DATA_AVAIL) {
+		command = SDHCI_GET_CMD(sdhci_readw(host, SDHCI_COMMAND));
+		if (command == MMC_SEND_TUNING_BLOCK ||
+		    command == MMC_SEND_TUNING_BLOCK_HS200) {
+			host->tuning_done = 1;
+			wake_up(&host->buf_ready_int);
+			return;
+		}
+	}
+
+	if (!host->data) {
+		struct mmc_command *data_cmd = host->data_cmd;
+
+		/*
+		 * The "data complete" interrupt is also used to
+		 * indicate that a busy state has ended. See comment
+		 * above in sdhci_cmd_irq().
+		 */
+		if (data_cmd && (data_cmd->flags & MMC_RSP_BUSY)) {
+			if (intmask & SDHCI_INT_DATA_TIMEOUT) {
+				host->data_cmd = NULL;
+				data_cmd->error = -ETIMEDOUT;
+				__sdhci_finish_mrq(host, data_cmd->mrq);
+				return;
+			}
+			if (intmask & SDHCI_INT_DATA_END) {
+				host->data_cmd = NULL;
+				/*
+				 * Some cards handle busy-end interrupt
+				 * before the command completed, so make
+				 * sure we do things in the proper order.
+				 */
+				if (host->cmd == data_cmd)
+					return;
+
+				__sdhci_finish_mrq(host, data_cmd->mrq);
+				return;
+			}
+		}
+
+		/*
+		 * SDHCI recovers from errors by resetting the cmd and data
+		 * circuits. Until that is done, there very well might be more
+		 * interrupts, so ignore them in that case.
+		 */
+		if (host->pending_reset)
+			return;
+
+		pr_err("%s: Got data interrupt 0x%08x even though no data operation was in progress.\n",
+		       mmc_hostname(host->mmc), (unsigned)intmask);
+		sdhci_dumpregs(host);
+
+		return;
+	}
+
+	if (intmask & SDHCI_INT_DATA_TIMEOUT)
+		host->data->error = -ETIMEDOUT;
+	else if (intmask & SDHCI_INT_DATA_END_BIT)
+		host->data->error = -EILSEQ;
+	else if ((intmask & SDHCI_INT_DATA_CRC) &&
+		SDHCI_GET_CMD(sdhci_readw(host, SDHCI_COMMAND))
+			!= MMC_BUS_TEST_R)
+		host->data->error = -EILSEQ;
+	else if (intmask & SDHCI_INT_ADMA_ERROR) {
+		pr_err("%s: ADMA error: 0x%08x\n", mmc_hostname(host->mmc),
+		       intmask);
+		sdhci_adma_show_error(host);
+		host->data->error = -EIO;
+		if (host->ops->adma_workaround)
+			host->ops->adma_workaround(host, intmask);
+	}
+
+	if (host->data->error)
+		sdhci_finish_data(host);
+	else {
+		if (intmask & (SDHCI_INT_DATA_AVAIL | SDHCI_INT_SPACE_AVAIL))
+			sdhci_transfer_pio(host);
+
+		/*
+		 * We currently don't do anything fancy with DMA
+		 * boundaries, but as we can't disable the feature
+		 * we need to at least restart the transfer.
+		 *
+		 * According to the spec sdhci_readl(host, SDHCI_DMA_ADDRESS)
+		 * should return a valid address to continue from, but as
+		 * some controllers are faulty, don't trust them.
+		 */
+		if (intmask & SDHCI_INT_DMA_END) {
+			dma_addr_t dmastart, dmanow;
+
+			dmastart = sdhci_sdma_address(host);
+			dmanow = dmastart + host->data->bytes_xfered;
+			/*
+			 * Force update to the next DMA block boundary.
+			 */
+			dmanow = (dmanow &
+				~((dma_addr_t)SDHCI_DEFAULT_BOUNDARY_SIZE - 1)) +
+				SDHCI_DEFAULT_BOUNDARY_SIZE;
+			host->data->bytes_xfered = dmanow - dmastart;
+			DBG("DMA base %pad, transferred 0x%06x bytes, next %pad\n",
+			    &dmastart, host->data->bytes_xfered, &dmanow);
+			sdhci_set_sdma_addr(host, dmanow);
+		}
+
+		if (intmask & SDHCI_INT_DATA_END) {
+			if (host->cmd == host->data_cmd) {
+				/*
+				 * Data managed to finish before the
+				 * command completed. Make sure we do
+				 * things in the proper order.
+				 */
+				host->data_early = 1;
+			} else {
+				sdhci_finish_data(host);
+			}
+		}
+	}
+}
+
+static inline bool sdhci_defer_done(struct sdhci_host *host,
+				    struct mmc_request *mrq)
+{
+	struct mmc_data *data = mrq->data;
+
+	return host->pending_reset || host->always_defer_done ||
+	       ((host->flags & SDHCI_REQ_USE_DMA) && data &&
+		data->host_cookie == COOKIE_MAPPED);
+}
+
+static irqreturn_t sdhci_irq(int irq, void *dev_id)
+{
+	struct mmc_request *mrqs_done[SDHCI_MAX_MRQS] = {0};
+	irqreturn_t result = IRQ_NONE;
+	struct sdhci_host *host = dev_id;
+	u32 intmask, mask, unexpected = 0;
+	int max_loops = 16;
+	int i;
+
+	spin_lock(&host->lock);
+
+	if (host->runtime_suspended) {
+		spin_unlock(&host->lock);
+		return IRQ_NONE;
+	}
+
+	intmask = sdhci_readl(host, SDHCI_INT_STATUS);
+	if (!intmask || intmask == 0xffffffff) {
+		result = IRQ_NONE;
+		goto out;
+	}
+
+	do {
+		DBG("IRQ status 0x%08x\n", intmask);
+
+		if (host->ops->irq) {
+			intmask = host->ops->irq(host, intmask);
+			if (!intmask)
+				goto cont;
+		}
+
+		/* Clear selected interrupts. */
+		mask = intmask & (SDHCI_INT_CMD_MASK | SDHCI_INT_DATA_MASK |
+				  SDHCI_INT_BUS_POWER);
+		sdhci_writel(host, mask, SDHCI_INT_STATUS);
+
+		if (intmask & (SDHCI_INT_CARD_INSERT | SDHCI_INT_CARD_REMOVE)) {
+			u32 present = sdhci_readl(host, SDHCI_PRESENT_STATE) &
+				      SDHCI_CARD_PRESENT;
+
+			/*
+			 * There is a observation on i.mx esdhc.  INSERT
+			 * bit will be immediately set again when it gets
+			 * cleared, if a card is inserted.  We have to mask
+			 * the irq to prevent interrupt storm which will
+			 * freeze the system.  And the REMOVE gets the
+			 * same situation.
+			 *
+			 * More testing are needed here to ensure it works
+			 * for other platforms though.
+			 */
+			host->ier &= ~(SDHCI_INT_CARD_INSERT |
+				       SDHCI_INT_CARD_REMOVE);
+			host->ier |= present ? SDHCI_INT_CARD_REMOVE :
+					       SDHCI_INT_CARD_INSERT;
+			sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+			sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+
+			sdhci_writel(host, intmask & (SDHCI_INT_CARD_INSERT |
+				     SDHCI_INT_CARD_REMOVE), SDHCI_INT_STATUS);
+
+			host->thread_isr |= intmask & (SDHCI_INT_CARD_INSERT |
+						       SDHCI_INT_CARD_REMOVE);
+			result = IRQ_WAKE_THREAD;
+		}
+
+		if (intmask & SDHCI_INT_CMD_MASK)
+			sdhci_cmd_irq(host, intmask & SDHCI_INT_CMD_MASK, &intmask);
+
+		if (intmask & SDHCI_INT_DATA_MASK)
+			sdhci_data_irq(host, intmask & SDHCI_INT_DATA_MASK);
+
+		if (intmask & SDHCI_INT_BUS_POWER)
+			pr_err("%s: Card is consuming too much power!\n",
+				mmc_hostname(host->mmc));
+
+		if (intmask & SDHCI_INT_RETUNE)
+			mmc_retune_needed(host->mmc);
+
+		if ((intmask & SDHCI_INT_CARD_INT) &&
+		    (host->ier & SDHCI_INT_CARD_INT)) {
+			sdhci_enable_sdio_irq_nolock(host, false);
+			sdio_signal_irq(host->mmc);
+		}
+
+		intmask &= ~(SDHCI_INT_CARD_INSERT | SDHCI_INT_CARD_REMOVE |
+			     SDHCI_INT_CMD_MASK | SDHCI_INT_DATA_MASK |
+			     SDHCI_INT_ERROR | SDHCI_INT_BUS_POWER |
+			     SDHCI_INT_RETUNE | SDHCI_INT_CARD_INT);
+
+		if (intmask) {
+			unexpected |= intmask;
+			sdhci_writel(host, intmask, SDHCI_INT_STATUS);
+		}
+cont:
+		if (result == IRQ_NONE)
+			result = IRQ_HANDLED;
+
+		intmask = sdhci_readl(host, SDHCI_INT_STATUS);
+	} while (intmask && --max_loops);
+
+	/* Determine if mrqs can be completed immediately */
+	for (i = 0; i < SDHCI_MAX_MRQS; i++) {
+		struct mmc_request *mrq = host->mrqs_done[i];
+
+		if (!mrq)
+			continue;
+
+		if (sdhci_defer_done(host, mrq)) {
+			result = IRQ_WAKE_THREAD;
+		} else {
+			mrqs_done[i] = mrq;
+			host->mrqs_done[i] = NULL;
+		}
+	}
+out:
+	if (host->deferred_cmd)
+		result = IRQ_WAKE_THREAD;
+
+	spin_unlock(&host->lock);
+
+	/* Process mrqs ready for immediate completion */
+	for (i = 0; i < SDHCI_MAX_MRQS; i++) {
+		if (!mrqs_done[i])
+			continue;
+
+		if (host->ops->request_done)
+			host->ops->request_done(host, mrqs_done[i]);
+		else
+			mmc_request_done(host->mmc, mrqs_done[i]);
+	}
+
+	if (unexpected) {
+		pr_err("%s: Unexpected interrupt 0x%08x.\n",
+			   mmc_hostname(host->mmc), unexpected);
+		sdhci_dumpregs(host);
+	}
+
+	return result;
+}
+
+static irqreturn_t sdhci_thread_irq(int irq, void *dev_id)
+{
+	struct sdhci_host *host = dev_id;
+	struct mmc_command *cmd;
+	unsigned long flags;
+	u32 isr;
+
+	while (!sdhci_request_done(host))
+		;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	isr = host->thread_isr;
+	host->thread_isr = 0;
+
+	cmd = host->deferred_cmd;
+	if (cmd && !sdhci_send_command_retry(host, cmd, flags))
+		sdhci_finish_mrq(host, cmd->mrq);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (isr & (SDHCI_INT_CARD_INSERT | SDHCI_INT_CARD_REMOVE)) {
+		struct mmc_host *mmc = host->mmc;
+
+		mmc->ops->card_event(mmc);
+		mmc_detect_change(mmc, msecs_to_jiffies(200));
+	}
+
+	return IRQ_HANDLED;
+}
+
+/*****************************************************************************\
+ *                                                                           *
+ * Suspend/resume                                                            *
+ *                                                                           *
+\*****************************************************************************/
+
+#ifdef CONFIG_PM
+
+static bool sdhci_cd_irq_can_wakeup(struct sdhci_host *host)
+{
+	return mmc_card_is_removable(host->mmc) &&
+	       !(host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) &&
+	       !mmc_can_gpio_cd(host->mmc);
+}
+
+/*
+ * To enable wakeup events, the corresponding events have to be enabled in
+ * the Interrupt Status Enable register too. See 'Table 1-6: Wakeup Signal
+ * Table' in the SD Host Controller Standard Specification.
+ * It is useless to restore SDHCI_INT_ENABLE state in
+ * sdhci_disable_irq_wakeups() since it will be set by
+ * sdhci_enable_card_detection() or sdhci_init().
+ */
+static bool sdhci_enable_irq_wakeups(struct sdhci_host *host)
+{
+	u8 mask = SDHCI_WAKE_ON_INSERT | SDHCI_WAKE_ON_REMOVE |
+		  SDHCI_WAKE_ON_INT;
+	u32 irq_val = 0;
+	u8 wake_val = 0;
+	u8 val;
+
+	if (sdhci_cd_irq_can_wakeup(host)) {
+		wake_val |= SDHCI_WAKE_ON_INSERT | SDHCI_WAKE_ON_REMOVE;
+		irq_val |= SDHCI_INT_CARD_INSERT | SDHCI_INT_CARD_REMOVE;
+	}
+
+	if (mmc_card_wake_sdio_irq(host->mmc)) {
+		wake_val |= SDHCI_WAKE_ON_INT;
+		irq_val |= SDHCI_INT_CARD_INT;
+	}
+
+	if (!irq_val)
+		return false;
+
+	val = sdhci_readb(host, SDHCI_WAKE_UP_CONTROL);
+	val &= ~mask;
+	val |= wake_val;
+	sdhci_writeb(host, val, SDHCI_WAKE_UP_CONTROL);
+
+	sdhci_writel(host, irq_val, SDHCI_INT_ENABLE);
+
+	host->irq_wake_enabled = !enable_irq_wake(host->irq);
+
+	return host->irq_wake_enabled;
+}
+
+static void sdhci_disable_irq_wakeups(struct sdhci_host *host)
+{
+	u8 val;
+	u8 mask = SDHCI_WAKE_ON_INSERT | SDHCI_WAKE_ON_REMOVE
+			| SDHCI_WAKE_ON_INT;
+
+	val = sdhci_readb(host, SDHCI_WAKE_UP_CONTROL);
+	val &= ~mask;
+	sdhci_writeb(host, val, SDHCI_WAKE_UP_CONTROL);
+
+	disable_irq_wake(host->irq);
+
+	host->irq_wake_enabled = false;
+}
+
+int sdhci_suspend_host(struct sdhci_host *host)
+{
+	sdhci_disable_card_detection(host);
+
+	mmc_retune_timer_stop(host->mmc);
+
+	if (!device_may_wakeup(mmc_dev(host->mmc)) ||
+	    !sdhci_enable_irq_wakeups(host)) {
+		host->ier = 0;
+		sdhci_writel(host, 0, SDHCI_INT_ENABLE);
+		sdhci_writel(host, 0, SDHCI_SIGNAL_ENABLE);
+		free_irq(host->irq, host);
+	}
+
+	return 0;
+}
+
+EXPORT_SYMBOL_GPL(sdhci_suspend_host);
+
+int sdhci_resume_host(struct sdhci_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	int ret = 0;
+
+	if (host->flags & (SDHCI_USE_SDMA | SDHCI_USE_ADMA)) {
+		if (host->ops->enable_dma)
+			host->ops->enable_dma(host);
+	}
+
+	if ((host->mmc->pm_flags & MMC_PM_KEEP_POWER) &&
+	    (host->quirks2 & SDHCI_QUIRK2_HOST_OFF_CARD_ON)) {
+		/* Card keeps power but host controller does not */
+		sdhci_init(host, 0);
+		host->pwr = 0;
+		host->clock = 0;
+		host->reinit_uhs = true;
+		mmc->ops->set_ios(mmc, &mmc->ios);
+	} else {
+		sdhci_init(host, (host->mmc->pm_flags & MMC_PM_KEEP_POWER));
+	}
+
+	if (host->irq_wake_enabled) {
+		sdhci_disable_irq_wakeups(host);
+	} else {
+		ret = request_threaded_irq(host->irq, sdhci_irq,
+					   sdhci_thread_irq, IRQF_SHARED,
+					   mmc_hostname(host->mmc), host);
+		if (ret)
+			return ret;
+	}
+
+	sdhci_enable_card_detection(host);
+
+	return ret;
+}
+
+EXPORT_SYMBOL_GPL(sdhci_resume_host);
+
+int sdhci_runtime_suspend_host(struct sdhci_host *host)
+{
+	unsigned long flags;
+
+	mmc_retune_timer_stop(host->mmc);
+
+	spin_lock_irqsave(&host->lock, flags);
+	host->ier &= SDHCI_INT_CARD_INT;
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	synchronize_hardirq(host->irq);
+
+	spin_lock_irqsave(&host->lock, flags);
+	host->runtime_suspended = true;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(sdhci_runtime_suspend_host);
+
+int sdhci_runtime_resume_host(struct sdhci_host *host, int soft_reset)
+{
+	struct mmc_host *mmc = host->mmc;
+	unsigned long flags;
+	int host_flags = host->flags;
+
+	if (host_flags & (SDHCI_USE_SDMA | SDHCI_USE_ADMA)) {
+		if (host->ops->enable_dma)
+			host->ops->enable_dma(host);
+	}
+
+	sdhci_init(host, soft_reset);
+
+	if (mmc->ios.power_mode != MMC_POWER_UNDEFINED &&
+	    mmc->ios.power_mode != MMC_POWER_OFF) {
+		/* Force clock and power re-program */
+		host->pwr = 0;
+		host->clock = 0;
+		host->reinit_uhs = true;
+		mmc->ops->start_signal_voltage_switch(mmc, &mmc->ios);
+		mmc->ops->set_ios(mmc, &mmc->ios);
+
+		if ((host_flags & SDHCI_PV_ENABLED) &&
+		    !(host->quirks2 & SDHCI_QUIRK2_PRESET_VALUE_BROKEN)) {
+			spin_lock_irqsave(&host->lock, flags);
+			sdhci_enable_preset_value(host, true);
+			spin_unlock_irqrestore(&host->lock, flags);
+		}
+
+		if ((mmc->caps2 & MMC_CAP2_HS400_ES) &&
+		    mmc->ops->hs400_enhanced_strobe)
+			mmc->ops->hs400_enhanced_strobe(mmc, &mmc->ios);
+	}
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	host->runtime_suspended = false;
+
+	/* Enable SDIO IRQ */
+	if (sdio_irq_claimed(mmc))
+		sdhci_enable_sdio_irq_nolock(host, true);
+
+	/* Enable Card Detection */
+	sdhci_enable_card_detection(host);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(sdhci_runtime_resume_host);
+
+#endif /* CONFIG_PM */
+
+/*****************************************************************************\
+ *                                                                           *
+ * Command Queue Engine (CQE) helpers                                        *
+ *                                                                           *
+\*****************************************************************************/
+
+void sdhci_cqe_enable(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+	u8 ctrl;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+	ctrl &= ~SDHCI_CTRL_DMA_MASK;
+	/*
+	 * Host from V4.10 supports ADMA3 DMA type.
+	 * ADMA3 performs integrated descriptor which is more suitable
+	 * for cmd queuing to fetch both command and transfer descriptors.
+	 */
+	if (host->v4_mode && (host->caps1 & SDHCI_CAN_DO_ADMA3))
+		ctrl |= SDHCI_CTRL_ADMA3;
+	else if (host->flags & SDHCI_USE_64_BIT_DMA)
+		ctrl |= SDHCI_CTRL_ADMA64;
+	else
+		ctrl |= SDHCI_CTRL_ADMA32;
+	sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+
+	sdhci_writew(host, SDHCI_MAKE_BLKSZ(host->sdma_boundary, 512),
+		     SDHCI_BLOCK_SIZE);
+
+	/* Set maximum timeout */
+	sdhci_set_timeout(host, NULL);
+
+	host->ier = host->cqe_ier;
+
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+
+	host->cqe_on = true;
+
+	pr_debug("%s: sdhci: CQE on, IRQ mask %#x, IRQ status %#x\n",
+		 mmc_hostname(mmc), host->ier,
+		 sdhci_readl(host, SDHCI_INT_STATUS));
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+EXPORT_SYMBOL_GPL(sdhci_cqe_enable);
+
+void sdhci_cqe_disable(struct mmc_host *mmc, bool recovery)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	sdhci_set_default_irqs(host);
+
+	host->cqe_on = false;
+
+	if (recovery) {
+		sdhci_do_reset(host, SDHCI_RESET_CMD);
+		sdhci_do_reset(host, SDHCI_RESET_DATA);
+	}
+
+	pr_debug("%s: sdhci: CQE off, IRQ mask %#x, IRQ status %#x\n",
+		 mmc_hostname(mmc), host->ier,
+		 sdhci_readl(host, SDHCI_INT_STATUS));
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+EXPORT_SYMBOL_GPL(sdhci_cqe_disable);
+
+bool sdhci_cqe_irq(struct sdhci_host *host, u32 intmask, int *cmd_error,
+		   int *data_error)
+{
+	u32 mask;
+
+	if (!host->cqe_on)
+		return false;
+
+	if (intmask & (SDHCI_INT_INDEX | SDHCI_INT_END_BIT | SDHCI_INT_CRC))
+		*cmd_error = -EILSEQ;
+	else if (intmask & SDHCI_INT_TIMEOUT)
+		*cmd_error = -ETIMEDOUT;
+	else
+		*cmd_error = 0;
+
+	if (intmask & (SDHCI_INT_DATA_END_BIT | SDHCI_INT_DATA_CRC))
+		*data_error = -EILSEQ;
+	else if (intmask & SDHCI_INT_DATA_TIMEOUT)
+		*data_error = -ETIMEDOUT;
+	else if (intmask & SDHCI_INT_ADMA_ERROR)
+		*data_error = -EIO;
+	else
+		*data_error = 0;
+
+	/* Clear selected interrupts. */
+	mask = intmask & host->cqe_ier;
+	sdhci_writel(host, mask, SDHCI_INT_STATUS);
+
+	if (intmask & SDHCI_INT_BUS_POWER)
+		pr_err("%s: Card is consuming too much power!\n",
+		       mmc_hostname(host->mmc));
+
+	intmask &= ~(host->cqe_ier | SDHCI_INT_ERROR);
+	if (intmask) {
+		sdhci_writel(host, intmask, SDHCI_INT_STATUS);
+		pr_err("%s: CQE: Unexpected interrupt 0x%08x.\n",
+		       mmc_hostname(host->mmc), intmask);
+		sdhci_dumpregs(host);
+	}
+
+	return true;
+}
+EXPORT_SYMBOL_GPL(sdhci_cqe_irq);
+
+/*****************************************************************************\
+ *                                                                           *
+ * Device allocation/registration                                            *
+ *                                                                           *
+\*****************************************************************************/
+
+struct sdhci_host *sdhci_alloc_host(struct device *dev,
+	size_t priv_size)
+{
+	struct mmc_host *mmc;
+	struct sdhci_host *host;
+
+	WARN_ON(dev == NULL);
+
+	mmc = mmc_alloc_host(sizeof(struct sdhci_host) + priv_size, dev);
+	if (!mmc)
+		return ERR_PTR(-ENOMEM);
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+	host->mmc_host_ops = sdhci_ops;
+	mmc->ops = &host->mmc_host_ops;
+
+	host->flags = SDHCI_SIGNALING_330;
+
+	host->cqe_ier     = SDHCI_CQE_INT_MASK;
+	host->cqe_err_ier = SDHCI_CQE_INT_ERR_MASK;
+
+	host->tuning_delay = -1;
+	host->tuning_loop_count = MAX_TUNING_LOOP;
+
+	host->sdma_boundary = SDHCI_DEFAULT_BOUNDARY_ARG;
+
+	/*
+	 * The DMA table descriptor count is calculated as the maximum
+	 * number of segments times 2, to allow for an alignment
+	 * descriptor for each segment, plus 1 for a nop end descriptor.
+	 */
+	host->adma_table_cnt = SDHCI_MAX_SEGS * 2 + 1;
+	host->max_adma = 65536;
+
+	return host;
+}
+
+EXPORT_SYMBOL_GPL(sdhci_alloc_host);
+
+static int sdhci_set_dma_mask(struct sdhci_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	struct device *dev = mmc_dev(mmc);
+	int ret = -EINVAL;
+
+	if (host->quirks2 & SDHCI_QUIRK2_BROKEN_64_BIT_DMA)
+		host->flags &= ~SDHCI_USE_64_BIT_DMA;
+
+	/* Try 64-bit mask if hardware is capable  of it */
+	if (host->flags & SDHCI_USE_64_BIT_DMA) {
+		ret = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(64));
+		if (ret) {
+			pr_warn("%s: Failed to set 64-bit DMA mask.\n",
+				mmc_hostname(mmc));
+			host->flags &= ~SDHCI_USE_64_BIT_DMA;
+		}
+	}
+
+	/* 32-bit mask as default & fallback */
+	if (ret) {
+		ret = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(32));
+		if (ret)
+			pr_warn("%s: Failed to set 32-bit DMA mask.\n",
+				mmc_hostname(mmc));
+	}
+
+	return ret;
+}
+
+void __sdhci_read_caps(struct sdhci_host *host, const u16 *ver,
+		       const u32 *caps, const u32 *caps1)
+{
+	u16 v;
+	u64 dt_caps_mask = 0;
+	u64 dt_caps = 0;
+
+	if (host->read_caps)
+		return;
+
+	host->read_caps = true;
+
+	if (debug_quirks)
+		host->quirks = debug_quirks;
+
+	if (debug_quirks2)
+		host->quirks2 = debug_quirks2;
+
+	sdhci_do_reset(host, SDHCI_RESET_ALL);
+
+	if (host->v4_mode)
+		sdhci_do_enable_v4_mode(host);
+
+	device_property_read_u64_array(mmc_dev(host->mmc),
+				       "sdhci-caps-mask", &dt_caps_mask, 1);
+	device_property_read_u64_array(mmc_dev(host->mmc),
+				       "sdhci-caps", &dt_caps, 1);
+
+	v = ver ? *ver : sdhci_readw(host, SDHCI_HOST_VERSION);
+	host->version = (v & SDHCI_SPEC_VER_MASK) >> SDHCI_SPEC_VER_SHIFT;
+
+	if (host->quirks & SDHCI_QUIRK_MISSING_CAPS)
+		return;
+
+	if (caps) {
+		host->caps = *caps;
+	} else {
+		host->caps = sdhci_readl(host, SDHCI_CAPABILITIES);
+		host->caps &= ~lower_32_bits(dt_caps_mask);
+		host->caps |= lower_32_bits(dt_caps);
+	}
+
+	if (host->version < SDHCI_SPEC_300)
+		return;
+
+	if (caps1) {
+		host->caps1 = *caps1;
+	} else {
+		host->caps1 = sdhci_readl(host, SDHCI_CAPABILITIES_1);
+		host->caps1 &= ~upper_32_bits(dt_caps_mask);
+		host->caps1 |= upper_32_bits(dt_caps);
+	}
+}
+EXPORT_SYMBOL_GPL(__sdhci_read_caps);
+
+static void sdhci_allocate_bounce_buffer(struct sdhci_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	unsigned int max_blocks;
+	unsigned int bounce_size;
+	int ret;
+
+	/*
+	 * Cap the bounce buffer at 64KB. Using a bigger bounce buffer
+	 * has diminishing returns, this is probably because SD/MMC
+	 * cards are usually optimized to handle this size of requests.
+	 */
+	bounce_size = SZ_64K;
+	/*
+	 * Adjust downwards to maximum request size if this is less
+	 * than our segment size, else hammer down the maximum
+	 * request size to the maximum buffer size.
+	 */
+	if (mmc->max_req_size < bounce_size)
+		bounce_size = mmc->max_req_size;
+	max_blocks = bounce_size / 512;
+
+	/*
+	 * When we just support one segment, we can get significant
+	 * speedups by the help of a bounce buffer to group scattered
+	 * reads/writes together.
+	 */
+	host->bounce_buffer = devm_kmalloc(mmc->parent,
+					   bounce_size,
+					   GFP_KERNEL);
+	if (!host->bounce_buffer) {
+		pr_err("%s: failed to allocate %u bytes for bounce buffer, falling back to single segments\n",
+		       mmc_hostname(mmc),
+		       bounce_size);
+		/*
+		 * Exiting with zero here makes sure we proceed with
+		 * mmc->max_segs == 1.
+		 */
+		return;
+	}
+
+	host->bounce_addr = dma_map_single(mmc->parent,
+					   host->bounce_buffer,
+					   bounce_size,
+					   DMA_BIDIRECTIONAL);
+	ret = dma_mapping_error(mmc->parent, host->bounce_addr);
+	if (ret)
+		/* Again fall back to max_segs == 1 */
+		return;
+	host->bounce_buffer_size = bounce_size;
+
+	/* Lie about this since we're bouncing */
+	mmc->max_segs = max_blocks;
+	mmc->max_seg_size = bounce_size;
+	mmc->max_req_size = bounce_size;
+
+	pr_info("%s bounce up to %u segments into one, max segment size %u bytes\n",
+		mmc_hostname(mmc), max_blocks, bounce_size);
+}
+
+static inline bool sdhci_can_64bit_dma(struct sdhci_host *host)
+{
+	/*
+	 * According to SD Host Controller spec v4.10, bit[27] added from
+	 * version 4.10 in Capabilities Register is used as 64-bit System
+	 * Address support for V4 mode.
+	 */
+	if (host->version >= SDHCI_SPEC_410 && host->v4_mode)
+		return host->caps & SDHCI_CAN_64BIT_V4;
+
+	return host->caps & SDHCI_CAN_64BIT;
+}
+
+int sdhci_setup_host(struct sdhci_host *host)
+{
+	struct mmc_host *mmc;
+	u32 max_current_caps;
+	unsigned int ocr_avail;
+	unsigned int override_timeout_clk;
+	u32 max_clk;
+	int ret = 0;
+	bool enable_vqmmc = false;
+
+	WARN_ON(host == NULL);
+	if (host == NULL)
+		return -EINVAL;
+
+	mmc = host->mmc;
+
+	/*
+	 * If there are external regulators, get them. Note this must be done
+	 * early before resetting the host and reading the capabilities so that
+	 * the host can take the appropriate action if regulators are not
+	 * available.
+	 */
+	if (!mmc->supply.vqmmc) {
+		ret = mmc_regulator_get_supply(mmc);
+		if (ret)
+			return ret;
+		enable_vqmmc  = true;
+	}
+
+	DBG("Version:   0x%08x | Present:  0x%08x\n",
+	    sdhci_readw(host, SDHCI_HOST_VERSION),
+	    sdhci_readl(host, SDHCI_PRESENT_STATE));
+	DBG("Caps:      0x%08x | Caps_1:   0x%08x\n",
+	    sdhci_readl(host, SDHCI_CAPABILITIES),
+	    sdhci_readl(host, SDHCI_CAPABILITIES_1));
+
+	sdhci_read_caps(host);
+
+	override_timeout_clk = host->timeout_clk;
+
+	if (host->version > SDHCI_SPEC_420) {
+		pr_err("%s: Unknown controller version (%d). You may experience problems.\n",
+		       mmc_hostname(mmc), host->version);
+	}
+
+	if (host->quirks & SDHCI_QUIRK_FORCE_DMA)
+		host->flags |= SDHCI_USE_SDMA;
+	else if (!(host->caps & SDHCI_CAN_DO_SDMA))
+		DBG("Controller doesn't have SDMA capability\n");
+	else
+		host->flags |= SDHCI_USE_SDMA;
+
+	if ((host->quirks & SDHCI_QUIRK_BROKEN_DMA) &&
+		(host->flags & SDHCI_USE_SDMA)) {
+		DBG("Disabling DMA as it is marked broken\n");
+		host->flags &= ~SDHCI_USE_SDMA;
+	}
+
+	if ((host->version >= SDHCI_SPEC_200) &&
+		(host->caps & SDHCI_CAN_DO_ADMA2))
+		host->flags |= SDHCI_USE_ADMA;
+
+	if ((host->quirks & SDHCI_QUIRK_BROKEN_ADMA) &&
+		(host->flags & SDHCI_USE_ADMA)) {
+		DBG("Disabling ADMA as it is marked broken\n");
+		host->flags &= ~SDHCI_USE_ADMA;
+	}
+
+	if (sdhci_can_64bit_dma(host))
+		host->flags |= SDHCI_USE_64_BIT_DMA;
+
+	if (host->use_external_dma) {
+		ret = sdhci_external_dma_init(host);
+		if (ret == -EPROBE_DEFER)
+			goto unreg;
+		/*
+		 * Fall back to use the DMA/PIO integrated in standard SDHCI
+		 * instead of external DMA devices.
+		 */
+		else if (ret)
+			sdhci_switch_external_dma(host, false);
+		/* Disable internal DMA sources */
+		else
+			host->flags &= ~(SDHCI_USE_SDMA | SDHCI_USE_ADMA);
+	}
+
+	if (host->flags & (SDHCI_USE_SDMA | SDHCI_USE_ADMA)) {
+		if (host->ops->set_dma_mask)
+			ret = host->ops->set_dma_mask(host);
+		else
+			ret = sdhci_set_dma_mask(host);
+
+		if (!ret && host->ops->enable_dma)
+			ret = host->ops->enable_dma(host);
+
+		if (ret) {
+			pr_warn("%s: No suitable DMA available - falling back to PIO\n",
+				mmc_hostname(mmc));
+			host->flags &= ~(SDHCI_USE_SDMA | SDHCI_USE_ADMA);
+
+			ret = 0;
+		}
+	}
+
+	/* SDMA does not support 64-bit DMA if v4 mode not set */
+	if ((host->flags & SDHCI_USE_64_BIT_DMA) && !host->v4_mode)
+		host->flags &= ~SDHCI_USE_SDMA;
+
+	if (host->flags & SDHCI_USE_ADMA) {
+		dma_addr_t dma;
+		void *buf;
+
+		if (!(host->flags & SDHCI_USE_64_BIT_DMA))
+			host->alloc_desc_sz = SDHCI_ADMA2_32_DESC_SZ;
+		else if (!host->alloc_desc_sz)
+			host->alloc_desc_sz = SDHCI_ADMA2_64_DESC_SZ(host);
+
+		host->desc_sz = host->alloc_desc_sz;
+		host->adma_table_sz = host->adma_table_cnt * host->desc_sz;
+
+		host->align_buffer_sz = SDHCI_MAX_SEGS * SDHCI_ADMA2_ALIGN;
+		/*
+		 * Use zalloc to zero the reserved high 32-bits of 128-bit
+		 * descriptors so that they never need to be written.
+		 */
+		buf = dma_alloc_coherent(mmc_dev(mmc),
+					 host->align_buffer_sz + host->adma_table_sz,
+					 &dma, GFP_KERNEL);
+		if (!buf) {
+			pr_warn("%s: Unable to allocate ADMA buffers - falling back to standard DMA\n",
+				mmc_hostname(mmc));
+			host->flags &= ~SDHCI_USE_ADMA;
+		} else if ((dma + host->align_buffer_sz) &
+			   (SDHCI_ADMA2_DESC_ALIGN - 1)) {
+			pr_warn("%s: unable to allocate aligned ADMA descriptor\n",
+				mmc_hostname(mmc));
+			host->flags &= ~SDHCI_USE_ADMA;
+			dma_free_coherent(mmc_dev(mmc), host->align_buffer_sz +
+					  host->adma_table_sz, buf, dma);
+		} else {
+			host->align_buffer = buf;
+			host->align_addr = dma;
+
+			host->adma_table = buf + host->align_buffer_sz;
+			host->adma_addr = dma + host->align_buffer_sz;
+		}
+	}
+
+	/*
+	 * If we use DMA, then it's up to the caller to set the DMA
+	 * mask, but PIO does not need the hw shim so we set a new
+	 * mask here in that case.
+	 */
+	if (!(host->flags & (SDHCI_USE_SDMA | SDHCI_USE_ADMA))) {
+		host->dma_mask = DMA_BIT_MASK(64);
+		mmc_dev(mmc)->dma_mask = &host->dma_mask;
+	}
+
+	if (host->version >= SDHCI_SPEC_300)
+		host->max_clk = FIELD_GET(SDHCI_CLOCK_V3_BASE_MASK, host->caps);
+	else
+		host->max_clk = FIELD_GET(SDHCI_CLOCK_BASE_MASK, host->caps);
+
+	host->max_clk *= 1000000;
+	if (host->max_clk == 0 || host->quirks &
+			SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN) {
+		if (!host->ops->get_max_clock) {
+			pr_err("%s: Hardware doesn't specify base clock frequency.\n",
+			       mmc_hostname(mmc));
+			ret = -ENODEV;
+			goto undma;
+		}
+		host->max_clk = host->ops->get_max_clock(host);
+	}
+
+	/*
+	 * In case of Host Controller v3.00, find out whether clock
+	 * multiplier is supported.
+	 */
+	host->clk_mul = FIELD_GET(SDHCI_CLOCK_MUL_MASK, host->caps1);
+
+	/*
+	 * In case the value in Clock Multiplier is 0, then programmable
+	 * clock mode is not supported, otherwise the actual clock
+	 * multiplier is one more than the value of Clock Multiplier
+	 * in the Capabilities Register.
+	 */
+	if (host->clk_mul)
+		host->clk_mul += 1;
+
+	/*
+	 * Set host parameters.
+	 */
+	max_clk = host->max_clk;
+
+	if (host->ops->get_min_clock)
+		mmc->f_min = host->ops->get_min_clock(host);
+	else if (host->version >= SDHCI_SPEC_300) {
+		if (host->clk_mul)
+			max_clk = host->max_clk * host->clk_mul;
+		/*
+		 * Divided Clock Mode minimum clock rate is always less than
+		 * Programmable Clock Mode minimum clock rate.
+		 */
+		mmc->f_min = host->max_clk / SDHCI_MAX_DIV_SPEC_300;
+	} else
+		mmc->f_min = host->max_clk / SDHCI_MAX_DIV_SPEC_200;
+
+	if (!mmc->f_max || mmc->f_max > max_clk)
+		mmc->f_max = max_clk;
+
+	if (!(host->quirks & SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK)) {
+		host->timeout_clk = FIELD_GET(SDHCI_TIMEOUT_CLK_MASK, host->caps);
+
+		if (host->caps & SDHCI_TIMEOUT_CLK_UNIT)
+			host->timeout_clk *= 1000;
+
+		if (host->timeout_clk == 0) {
+			if (!host->ops->get_timeout_clock) {
+				pr_err("%s: Hardware doesn't specify timeout clock frequency.\n",
+					mmc_hostname(mmc));
+				ret = -ENODEV;
+				goto undma;
+			}
+
+			host->timeout_clk =
+				DIV_ROUND_UP(host->ops->get_timeout_clock(host),
+					     1000);
+		}
+
+		if (override_timeout_clk)
+			host->timeout_clk = override_timeout_clk;
+
+		mmc->max_busy_timeout = host->ops->get_max_timeout_count ?
+			host->ops->get_max_timeout_count(host) : 1 << 27;
+		mmc->max_busy_timeout /= host->timeout_clk;
+	}
+
+	if (host->quirks2 & SDHCI_QUIRK2_DISABLE_HW_TIMEOUT &&
+	    !host->ops->get_max_timeout_count)
+		mmc->max_busy_timeout = 0;
+
+	mmc->caps |= MMC_CAP_SDIO_IRQ | MMC_CAP_CMD23;
+	mmc->caps2 |= MMC_CAP2_SDIO_IRQ_NOTHREAD;
+
+	if (host->quirks & SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12)
+		host->flags |= SDHCI_AUTO_CMD12;
+
+	/*
+	 * For v3 mode, Auto-CMD23 stuff only works in ADMA or PIO.
+	 * For v4 mode, SDMA may use Auto-CMD23 as well.
+	 */
+	if ((host->version >= SDHCI_SPEC_300) &&
+	    ((host->flags & SDHCI_USE_ADMA) ||
+	     !(host->flags & SDHCI_USE_SDMA) || host->v4_mode) &&
+	     !(host->quirks2 & SDHCI_QUIRK2_ACMD23_BROKEN)) {
+		host->flags |= SDHCI_AUTO_CMD23;
+		DBG("Auto-CMD23 available\n");
+	} else {
+		DBG("Auto-CMD23 unavailable\n");
+	}
+
+	/*
+	 * A controller may support 8-bit width, but the board itself
+	 * might not have the pins brought out.  Boards that support
+	 * 8-bit width must set "mmc->caps |= MMC_CAP_8_BIT_DATA;" in
+	 * their platform code before calling sdhci_add_host(), and we
+	 * won't assume 8-bit width for hosts without that CAP.
+	 */
+	if (!(host->quirks & SDHCI_QUIRK_FORCE_1_BIT_DATA))
+		mmc->caps |= MMC_CAP_4_BIT_DATA;
+
+	if (host->quirks2 & SDHCI_QUIRK2_HOST_NO_CMD23)
+		mmc->caps &= ~MMC_CAP_CMD23;
+
+	if (host->caps & SDHCI_CAN_DO_HISPD)
+		mmc->caps |= MMC_CAP_SD_HIGHSPEED | MMC_CAP_MMC_HIGHSPEED;
+
+	if ((host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) &&
+	    mmc_card_is_removable(mmc) &&
+	    mmc_gpio_get_cd(host->mmc) < 0)
+		mmc->caps |= MMC_CAP_NEEDS_POLL;
+
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+		if (enable_vqmmc) {
+			ret = regulator_enable(mmc->supply.vqmmc);
+			host->sdhci_core_to_disable_vqmmc = !ret;
+		}
+
+		/* If vqmmc provides no 1.8V signalling, then there's no UHS */
+		if (!regulator_is_supported_voltage(mmc->supply.vqmmc, 1700000,
+						    1950000))
+			host->caps1 &= ~(SDHCI_SUPPORT_SDR104 |
+					 SDHCI_SUPPORT_SDR50 |
+					 SDHCI_SUPPORT_DDR50);
+
+		/* In eMMC case vqmmc might be a fixed 1.8V regulator */
+		if (!regulator_is_supported_voltage(mmc->supply.vqmmc, 2700000,
+						    3600000))
+			host->flags &= ~SDHCI_SIGNALING_330;
+
+		if (ret) {
+			pr_warn("%s: Failed to enable vqmmc regulator: %d\n",
+				mmc_hostname(mmc), ret);
+			mmc->supply.vqmmc = ERR_PTR(-EINVAL);
+		}
+
+	}
+
+	if (host->quirks2 & SDHCI_QUIRK2_NO_1_8_V) {
+		host->caps1 &= ~(SDHCI_SUPPORT_SDR104 | SDHCI_SUPPORT_SDR50 |
+				 SDHCI_SUPPORT_DDR50);
+		/*
+		 * The SDHCI controller in a SoC might support HS200/HS400
+		 * (indicated using mmc-hs200-1_8v/mmc-hs400-1_8v dt property),
+		 * but if the board is modeled such that the IO lines are not
+		 * connected to 1.8v then HS200/HS400 cannot be supported.
+		 * Disable HS200/HS400 if the board does not have 1.8v connected
+		 * to the IO lines. (Applicable for other modes in 1.8v)
+		 */
+		mmc->caps2 &= ~(MMC_CAP2_HSX00_1_8V | MMC_CAP2_HS400_ES);
+		mmc->caps &= ~(MMC_CAP_1_8V_DDR | MMC_CAP_UHS);
+	}
+
+	/* Any UHS-I mode in caps implies SDR12 and SDR25 support. */
+	if (host->caps1 & (SDHCI_SUPPORT_SDR104 | SDHCI_SUPPORT_SDR50 |
+			   SDHCI_SUPPORT_DDR50))
+		mmc->caps |= MMC_CAP_UHS_SDR12 | MMC_CAP_UHS_SDR25;
+
+	/* SDR104 supports also implies SDR50 support */
+	if (host->caps1 & SDHCI_SUPPORT_SDR104) {
+		mmc->caps |= MMC_CAP_UHS_SDR104 | MMC_CAP_UHS_SDR50;
+		/* SD3.0: SDR104 is supported so (for eMMC) the caps2
+		 * field can be promoted to support HS200.
+		 */
+		if (!(host->quirks2 & SDHCI_QUIRK2_BROKEN_HS200))
+			mmc->caps2 |= MMC_CAP2_HS200;
+	} else if (host->caps1 & SDHCI_SUPPORT_SDR50) {
+		mmc->caps |= MMC_CAP_UHS_SDR50;
+	}
+
+	if (host->quirks2 & SDHCI_QUIRK2_CAPS_BIT63_FOR_HS400 &&
+	    (host->caps1 & SDHCI_SUPPORT_HS400))
+		mmc->caps2 |= MMC_CAP2_HS400;
+
+	if ((mmc->caps2 & MMC_CAP2_HSX00_1_2V) &&
+	    (IS_ERR(mmc->supply.vqmmc) ||
+	     !regulator_is_supported_voltage(mmc->supply.vqmmc, 1100000,
+					     1300000)))
+		mmc->caps2 &= ~MMC_CAP2_HSX00_1_2V;
+
+	if ((host->caps1 & SDHCI_SUPPORT_DDR50) &&
+	    !(host->quirks2 & SDHCI_QUIRK2_BROKEN_DDR50))
+		mmc->caps |= MMC_CAP_UHS_DDR50;
+
+	/* Does the host need tuning for SDR50? */
+	if (host->caps1 & SDHCI_USE_SDR50_TUNING)
+		host->flags |= SDHCI_SDR50_NEEDS_TUNING;
+
+	/* Driver Type(s) (A, C, D) supported by the host */
+	if (host->caps1 & SDHCI_DRIVER_TYPE_A)
+		mmc->caps |= MMC_CAP_DRIVER_TYPE_A;
+	if (host->caps1 & SDHCI_DRIVER_TYPE_C)
+		mmc->caps |= MMC_CAP_DRIVER_TYPE_C;
+	if (host->caps1 & SDHCI_DRIVER_TYPE_D)
+		mmc->caps |= MMC_CAP_DRIVER_TYPE_D;
+
+	/* Initial value for re-tuning timer count */
+	host->tuning_count = FIELD_GET(SDHCI_RETUNING_TIMER_COUNT_MASK,
+				       host->caps1);
+
+	/*
+	 * In case Re-tuning Timer is not disabled, the actual value of
+	 * re-tuning timer will be 2 ^ (n - 1).
+	 */
+	if (host->tuning_count)
+		host->tuning_count = 1 << (host->tuning_count - 1);
+
+	/* Re-tuning mode supported by the Host Controller */
+	host->tuning_mode = FIELD_GET(SDHCI_RETUNING_MODE_MASK, host->caps1);
+
+	ocr_avail = 0;
+
+	/*
+	 * According to SD Host Controller spec v3.00, if the Host System
+	 * can afford more than 150mA, Host Driver should set XPC to 1. Also
+	 * the value is meaningful only if Voltage Support in the Capabilities
+	 * register is set. The actual current value is 4 times the register
+	 * value.
+	 */
+	max_current_caps = sdhci_readl(host, SDHCI_MAX_CURRENT);
+	if (!max_current_caps && !IS_ERR(mmc->supply.vmmc)) {
+		int curr = regulator_get_current_limit(mmc->supply.vmmc);
+		if (curr > 0) {
+
+			/* convert to SDHCI_MAX_CURRENT format */
+			curr = curr/1000;  /* convert to mA */
+			curr = curr/SDHCI_MAX_CURRENT_MULTIPLIER;
+
+			curr = min_t(u32, curr, SDHCI_MAX_CURRENT_LIMIT);
+			max_current_caps =
+				FIELD_PREP(SDHCI_MAX_CURRENT_330_MASK, curr) |
+				FIELD_PREP(SDHCI_MAX_CURRENT_300_MASK, curr) |
+				FIELD_PREP(SDHCI_MAX_CURRENT_180_MASK, curr);
+		}
+	}
+
+	if (host->caps & SDHCI_CAN_VDD_330) {
+		ocr_avail |= MMC_VDD_32_33 | MMC_VDD_33_34;
+
+		mmc->max_current_330 = FIELD_GET(SDHCI_MAX_CURRENT_330_MASK,
+						 max_current_caps) *
+						SDHCI_MAX_CURRENT_MULTIPLIER;
+	}
+	if (host->caps & SDHCI_CAN_VDD_300) {
+		ocr_avail |= MMC_VDD_29_30 | MMC_VDD_30_31;
+
+		mmc->max_current_300 = FIELD_GET(SDHCI_MAX_CURRENT_300_MASK,
+						 max_current_caps) *
+						SDHCI_MAX_CURRENT_MULTIPLIER;
+	}
+	if (host->caps & SDHCI_CAN_VDD_180) {
+		ocr_avail |= MMC_VDD_165_195;
+
+		mmc->max_current_180 = FIELD_GET(SDHCI_MAX_CURRENT_180_MASK,
+						 max_current_caps) *
+						SDHCI_MAX_CURRENT_MULTIPLIER;
+	}
+
+	/* If OCR set by host, use it instead. */
+	if (host->ocr_mask)
+		ocr_avail = host->ocr_mask;
+
+	/* If OCR set by external regulators, give it highest prio. */
+	if (mmc->ocr_avail)
+		ocr_avail = mmc->ocr_avail;
+
+	mmc->ocr_avail = ocr_avail;
+	mmc->ocr_avail_sdio = ocr_avail;
+	if (host->ocr_avail_sdio)
+		mmc->ocr_avail_sdio &= host->ocr_avail_sdio;
+	mmc->ocr_avail_sd = ocr_avail;
+	if (host->ocr_avail_sd)
+		mmc->ocr_avail_sd &= host->ocr_avail_sd;
+	else /* normal SD controllers don't support 1.8V */
+		mmc->ocr_avail_sd &= ~MMC_VDD_165_195;
+	mmc->ocr_avail_mmc = ocr_avail;
+	if (host->ocr_avail_mmc)
+		mmc->ocr_avail_mmc &= host->ocr_avail_mmc;
+
+	if (mmc->ocr_avail == 0) {
+		pr_err("%s: Hardware doesn't report any support voltages.\n",
+		       mmc_hostname(mmc));
+		ret = -ENODEV;
+		goto unreg;
+	}
+
+	if ((mmc->caps & (MMC_CAP_UHS_SDR12 | MMC_CAP_UHS_SDR25 |
+			  MMC_CAP_UHS_SDR50 | MMC_CAP_UHS_SDR104 |
+			  MMC_CAP_UHS_DDR50 | MMC_CAP_1_8V_DDR)) ||
+	    (mmc->caps2 & (MMC_CAP2_HS200_1_8V_SDR | MMC_CAP2_HS400_1_8V)))
+		host->flags |= SDHCI_SIGNALING_180;
+
+	if (mmc->caps2 & MMC_CAP2_HSX00_1_2V)
+		host->flags |= SDHCI_SIGNALING_120;
+
+	spin_lock_init(&host->lock);
+
+	/*
+	 * Maximum number of sectors in one transfer. Limited by SDMA boundary
+	 * size (512KiB). Note some tuning modes impose a 4MiB limit, but this
+	 * is less anyway.
+	 */
+	mmc->max_req_size = 524288;
+
+	/*
+	 * Maximum number of segments. Depends on if the hardware
+	 * can do scatter/gather or not.
+	 */
+	if (host->flags & SDHCI_USE_ADMA) {
+		mmc->max_segs = SDHCI_MAX_SEGS;
+	} else if (host->flags & SDHCI_USE_SDMA) {
+		mmc->max_segs = 1;
+		if (swiotlb_max_segment()) {
+			unsigned int max_req_size = (1 << IO_TLB_SHIFT) *
+						IO_TLB_SEGSIZE;
+			mmc->max_req_size = min(mmc->max_req_size,
+						max_req_size);
+		}
+	} else { /* PIO */
+		mmc->max_segs = SDHCI_MAX_SEGS;
+	}
+
+	/*
+	 * Maximum segment size. Could be one segment with the maximum number
+	 * of bytes. When doing hardware scatter/gather, each entry cannot
+	 * be larger than 64 KiB though.
+	 */
+	if (host->flags & SDHCI_USE_ADMA) {
+		if (host->quirks & SDHCI_QUIRK_BROKEN_ADMA_ZEROLEN_DESC) {
+			host->max_adma = 65532; /* 32-bit alignment */
+			mmc->max_seg_size = 65535;
+		} else {
+			mmc->max_seg_size = 65536;
+		}
+	} else {
+		mmc->max_seg_size = mmc->max_req_size;
+	}
+
+	/*
+	 * Maximum block size. This varies from controller to controller and
+	 * is specified in the capabilities register.
+	 */
+	if (host->quirks & SDHCI_QUIRK_FORCE_BLK_SZ_2048) {
+		mmc->max_blk_size = 2;
+	} else {
+		mmc->max_blk_size = (host->caps & SDHCI_MAX_BLOCK_MASK) >>
+				SDHCI_MAX_BLOCK_SHIFT;
+		if (mmc->max_blk_size >= 3) {
+			pr_warn("%s: Invalid maximum block size, assuming 512 bytes\n",
+				mmc_hostname(mmc));
+			mmc->max_blk_size = 0;
+		}
+	}
+
+	mmc->max_blk_size = 512 << mmc->max_blk_size;
+
+	/*
+	 * Maximum block count.
+	 */
+	mmc->max_blk_count = (host->quirks & SDHCI_QUIRK_NO_MULTIBLOCK) ? 1 : 65535;
+
+	if (mmc->max_segs == 1)
+		/* This may alter mmc->*_blk_* parameters */
+		sdhci_allocate_bounce_buffer(host);
+
+	return 0;
+
+unreg:
+	if (host->sdhci_core_to_disable_vqmmc)
+		regulator_disable(mmc->supply.vqmmc);
+undma:
+	if (host->align_buffer)
+		dma_free_coherent(mmc_dev(mmc), host->align_buffer_sz +
+				  host->adma_table_sz, host->align_buffer,
+				  host->align_addr);
+	host->adma_table = NULL;
+	host->align_buffer = NULL;
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(sdhci_setup_host);
+
+void sdhci_cleanup_host(struct sdhci_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+
+	if (host->sdhci_core_to_disable_vqmmc)
+		regulator_disable(mmc->supply.vqmmc);
+
+	if (host->align_buffer)
+		dma_free_coherent(mmc_dev(mmc), host->align_buffer_sz +
+				  host->adma_table_sz, host->align_buffer,
+				  host->align_addr);
+
+	if (host->use_external_dma)
+		sdhci_external_dma_release(host);
+
+	host->adma_table = NULL;
+	host->align_buffer = NULL;
+}
+EXPORT_SYMBOL_GPL(sdhci_cleanup_host);
+
+int __sdhci_add_host(struct sdhci_host *host)
+{
+	unsigned int flags = WQ_UNBOUND | WQ_MEM_RECLAIM | WQ_HIGHPRI;
+	struct mmc_host *mmc = host->mmc;
+	int ret;
+
+	if ((mmc->caps2 & MMC_CAP2_CQE) &&
+	    (host->quirks & SDHCI_QUIRK_BROKEN_CQE)) {
+		mmc->caps2 &= ~MMC_CAP2_CQE;
+		mmc->cqe_ops = NULL;
+	}
+
+	host->complete_wq = alloc_workqueue("sdhci", flags, 0);
+	if (!host->complete_wq)
+		return -ENOMEM;
+
+	INIT_WORK(&host->complete_work, sdhci_complete_work);
+
+	timer_setup(&host->timer, sdhci_timeout_timer, 0);
+	timer_setup(&host->data_timer, sdhci_timeout_data_timer, 0);
+
+	init_waitqueue_head(&host->buf_ready_int);
+
+	sdhci_init(host, 0);
+
+	ret = request_threaded_irq(host->irq, sdhci_irq, sdhci_thread_irq,
+				   IRQF_SHARED,	mmc_hostname(mmc), host);
+	if (ret) {
+		pr_err("%s: Failed to request IRQ %d: %d\n",
+		       mmc_hostname(mmc), host->irq, ret);
+		goto unwq;
+	}
+
+	ret = sdhci_led_register(host);
+	if (ret) {
+		pr_err("%s: Failed to register LED device: %d\n",
+		       mmc_hostname(mmc), ret);
+		goto unirq;
+	}
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto unled;
+
+	pr_info("%s: SDHCI controller on %s [%s] using %s\n",
+		mmc_hostname(mmc), host->hw_name, dev_name(mmc_dev(mmc)),
+		host->use_external_dma ? "External DMA" :
+		(host->flags & SDHCI_USE_ADMA) ?
+		(host->flags & SDHCI_USE_64_BIT_DMA) ? "ADMA 64-bit" : "ADMA" :
+		(host->flags & SDHCI_USE_SDMA) ? "DMA" : "PIO");
+
+	sdhci_enable_card_detection(host);
+
+	return 0;
+
+unled:
+	sdhci_led_unregister(host);
+unirq:
+	sdhci_do_reset(host, SDHCI_RESET_ALL);
+	sdhci_writel(host, 0, SDHCI_INT_ENABLE);
+	sdhci_writel(host, 0, SDHCI_SIGNAL_ENABLE);
+	free_irq(host->irq, host);
+unwq:
+	destroy_workqueue(host->complete_wq);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(__sdhci_add_host);
+
+int sdhci_add_host(struct sdhci_host *host)
+{
+	int ret;
+
+	ret = sdhci_setup_host(host);
+	if (ret)
+		return ret;
+
+	ret = __sdhci_add_host(host);
+	if (ret)
+		goto cleanup;
+
+	return 0;
+
+cleanup:
+	sdhci_cleanup_host(host);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(sdhci_add_host);
+
+void sdhci_remove_host(struct sdhci_host *host, int dead)
+{
+	struct mmc_host *mmc = host->mmc;
+	unsigned long flags;
+
+	if (dead) {
+		spin_lock_irqsave(&host->lock, flags);
+
+		host->flags |= SDHCI_DEVICE_DEAD;
+
+		if (sdhci_has_requests(host)) {
+			pr_err("%s: Controller removed during "
+				" transfer!\n", mmc_hostname(mmc));
+			sdhci_error_out_mrqs(host, -ENOMEDIUM);
+		}
+
+		spin_unlock_irqrestore(&host->lock, flags);
+	}
+
+	sdhci_disable_card_detection(host);
+
+	mmc_remove_host(mmc);
+
+	sdhci_led_unregister(host);
+
+	if (!dead)
+		sdhci_do_reset(host, SDHCI_RESET_ALL);
+
+	sdhci_writel(host, 0, SDHCI_INT_ENABLE);
+	sdhci_writel(host, 0, SDHCI_SIGNAL_ENABLE);
+	free_irq(host->irq, host);
+
+	del_timer_sync(&host->timer);
+	del_timer_sync(&host->data_timer);
+
+	destroy_workqueue(host->complete_wq);
+
+	if (host->sdhci_core_to_disable_vqmmc)
+		regulator_disable(mmc->supply.vqmmc);
+
+	if (host->align_buffer)
+		dma_free_coherent(mmc_dev(mmc), host->align_buffer_sz +
+				  host->adma_table_sz, host->align_buffer,
+				  host->align_addr);
+
+	if (host->use_external_dma)
+		sdhci_external_dma_release(host);
+
+	host->adma_table = NULL;
+	host->align_buffer = NULL;
+}
+
+EXPORT_SYMBOL_GPL(sdhci_remove_host);
+
+void sdhci_free_host(struct sdhci_host *host)
+{
+	mmc_free_host(host->mmc);
+}
+
+EXPORT_SYMBOL_GPL(sdhci_free_host);
+
+/*****************************************************************************\
+ *                                                                           *
+ * Driver init/exit                                                          *
+ *                                                                           *
+\*****************************************************************************/
+
+static int __init sdhci_drv_init(void)
+{
+	pr_info(DRIVER_NAME
+		": Secure Digital Host Controller Interface driver\n");
+	pr_info(DRIVER_NAME ": Copyright(c) Pierre Ossman\n");
+
+	return 0;
+}
+
+static void __exit sdhci_drv_exit(void)
+{
+}
+
+module_init(sdhci_drv_init);
+module_exit(sdhci_drv_exit);
+
+module_param(debug_quirks, uint, 0444);
+module_param(debug_quirks2, uint, 0444);
+
+MODULE_AUTHOR("Pierre Ossman <pierre@ossman.eu>");
+MODULE_DESCRIPTION("Secure Digital Host Controller Interface core driver");
+MODULE_LICENSE("GPL");
+
+MODULE_PARM_DESC(debug_quirks, "Force certain quirks.");
+MODULE_PARM_DESC(debug_quirks2, "Force certain other quirks.");
diff --git a/drivers/mmc/host/sdhci.h b/drivers/mmc/host/sdhci.h
new file mode 100644
index 000000000..4db57c3a8
--- /dev/null
+++ b/drivers/mmc/host/sdhci.h
@@ -0,0 +1,816 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ *  linux/drivers/mmc/host/sdhci.h - Secure Digital Host Controller Interface driver
+ *
+ * Header file for Host Controller registers and I/O accessors.
+ *
+ *  Copyright (C) 2005-2008 Pierre Ossman, All Rights Reserved.
+ */
+#ifndef __SDHCI_HW_H
+#define __SDHCI_HW_H
+
+#include <linux/bits.h>
+#include <linux/scatterlist.h>
+#include <linux/compiler.h>
+#include <linux/types.h>
+#include <linux/io.h>
+#include <linux/leds.h>
+#include <linux/interrupt.h>
+
+#include <linux/mmc/host.h>
+
+/*
+ * Controller registers
+ */
+
+#define SDHCI_DMA_ADDRESS	0x00
+#define SDHCI_ARGUMENT2		SDHCI_DMA_ADDRESS
+#define SDHCI_32BIT_BLK_CNT	SDHCI_DMA_ADDRESS
+
+#define SDHCI_BLOCK_SIZE	0x04
+#define  SDHCI_MAKE_BLKSZ(dma, blksz) (((dma & 0x7) << 12) | (blksz & 0xFFF))
+
+#define SDHCI_BLOCK_COUNT	0x06
+
+#define SDHCI_ARGUMENT		0x08
+
+#define SDHCI_TRANSFER_MODE	0x0C
+#define  SDHCI_TRNS_DMA		0x01
+#define  SDHCI_TRNS_BLK_CNT_EN	0x02
+#define  SDHCI_TRNS_AUTO_CMD12	0x04
+#define  SDHCI_TRNS_AUTO_CMD23	0x08
+#define  SDHCI_TRNS_AUTO_SEL	0x0C
+#define  SDHCI_TRNS_READ	0x10
+#define  SDHCI_TRNS_MULTI	0x20
+
+#define SDHCI_COMMAND		0x0E
+#define  SDHCI_CMD_RESP_MASK	0x03
+#define  SDHCI_CMD_CRC		0x08
+#define  SDHCI_CMD_INDEX	0x10
+#define  SDHCI_CMD_DATA		0x20
+#define  SDHCI_CMD_ABORTCMD	0xC0
+
+#define  SDHCI_CMD_RESP_NONE	0x00
+#define  SDHCI_CMD_RESP_LONG	0x01
+#define  SDHCI_CMD_RESP_SHORT	0x02
+#define  SDHCI_CMD_RESP_SHORT_BUSY 0x03
+
+#define SDHCI_MAKE_CMD(c, f) (((c & 0xff) << 8) | (f & 0xff))
+#define SDHCI_GET_CMD(c) ((c>>8) & 0x3f)
+
+#define SDHCI_RESPONSE		0x10
+
+#define SDHCI_BUFFER		0x20
+
+#define SDHCI_PRESENT_STATE	0x24
+#define  SDHCI_CMD_INHIBIT	0x00000001
+#define  SDHCI_DATA_INHIBIT	0x00000002
+#define  SDHCI_DOING_WRITE	0x00000100
+#define  SDHCI_DOING_READ	0x00000200
+#define  SDHCI_SPACE_AVAILABLE	0x00000400
+#define  SDHCI_DATA_AVAILABLE	0x00000800
+#define  SDHCI_CARD_PRESENT	0x00010000
+#define   SDHCI_CARD_PRES_SHIFT	16
+#define  SDHCI_CD_STABLE	0x00020000
+#define  SDHCI_CD_LVL		0x00040000
+#define   SDHCI_CD_LVL_SHIFT	18
+#define  SDHCI_WRITE_PROTECT	0x00080000
+#define  SDHCI_DATA_LVL_MASK	0x00F00000
+#define   SDHCI_DATA_LVL_SHIFT	20
+#define   SDHCI_DATA_0_LVL_MASK	0x00100000
+#define  SDHCI_CMD_LVL		0x01000000
+
+#define SDHCI_HOST_CONTROL	0x28
+#define  SDHCI_CTRL_LED		0x01
+#define  SDHCI_CTRL_4BITBUS	0x02
+#define  SDHCI_CTRL_HISPD	0x04
+#define  SDHCI_CTRL_DMA_MASK	0x18
+#define   SDHCI_CTRL_SDMA	0x00
+#define   SDHCI_CTRL_ADMA1	0x08
+#define   SDHCI_CTRL_ADMA32	0x10
+#define   SDHCI_CTRL_ADMA64	0x18
+#define   SDHCI_CTRL_ADMA3	0x18
+#define  SDHCI_CTRL_8BITBUS	0x20
+#define  SDHCI_CTRL_CDTEST_INS	0x40
+#define  SDHCI_CTRL_CDTEST_EN	0x80
+
+#define SDHCI_POWER_CONTROL	0x29
+#define  SDHCI_POWER_ON		0x01
+#define  SDHCI_POWER_180	0x0A
+#define  SDHCI_POWER_300	0x0C
+#define  SDHCI_POWER_330	0x0E
+
+#define SDHCI_BLOCK_GAP_CONTROL	0x2A
+
+#define SDHCI_WAKE_UP_CONTROL	0x2B
+#define  SDHCI_WAKE_ON_INT	0x01
+#define  SDHCI_WAKE_ON_INSERT	0x02
+#define  SDHCI_WAKE_ON_REMOVE	0x04
+
+#define SDHCI_CLOCK_CONTROL	0x2C
+#define  SDHCI_DIVIDER_SHIFT	8
+#define  SDHCI_DIVIDER_HI_SHIFT	6
+#define  SDHCI_DIV_MASK	0xFF
+#define  SDHCI_DIV_MASK_LEN	8
+#define  SDHCI_DIV_HI_MASK	0x300
+#define  SDHCI_PROG_CLOCK_MODE	0x0020
+#define  SDHCI_CLOCK_CARD_EN	0x0004
+#define  SDHCI_CLOCK_PLL_EN	0x0008
+#define  SDHCI_CLOCK_INT_STABLE	0x0002
+#define  SDHCI_CLOCK_INT_EN	0x0001
+
+#define SDHCI_TIMEOUT_CONTROL	0x2E
+
+#define SDHCI_SOFTWARE_RESET	0x2F
+#define  SDHCI_RESET_ALL	0x01
+#define  SDHCI_RESET_CMD	0x02
+#define  SDHCI_RESET_DATA	0x04
+
+#define SDHCI_INT_STATUS	0x30
+#define SDHCI_INT_ENABLE	0x34
+#define SDHCI_SIGNAL_ENABLE	0x38
+#define  SDHCI_INT_RESPONSE	0x00000001
+#define  SDHCI_INT_DATA_END	0x00000002
+#define  SDHCI_INT_BLK_GAP	0x00000004
+#define  SDHCI_INT_DMA_END	0x00000008
+#define  SDHCI_INT_SPACE_AVAIL	0x00000010
+#define  SDHCI_INT_DATA_AVAIL	0x00000020
+#define  SDHCI_INT_CARD_INSERT	0x00000040
+#define  SDHCI_INT_CARD_REMOVE	0x00000080
+#define  SDHCI_INT_CARD_INT	0x00000100
+#define  SDHCI_INT_RETUNE	0x00001000
+#define  SDHCI_INT_CQE		0x00004000
+#define  SDHCI_INT_ERROR	0x00008000
+#define  SDHCI_INT_TIMEOUT	0x00010000
+#define  SDHCI_INT_CRC		0x00020000
+#define  SDHCI_INT_END_BIT	0x00040000
+#define  SDHCI_INT_INDEX	0x00080000
+#define  SDHCI_INT_DATA_TIMEOUT	0x00100000
+#define  SDHCI_INT_DATA_CRC	0x00200000
+#define  SDHCI_INT_DATA_END_BIT	0x00400000
+#define  SDHCI_INT_BUS_POWER	0x00800000
+#define  SDHCI_INT_AUTO_CMD_ERR	0x01000000
+#define  SDHCI_INT_ADMA_ERROR	0x02000000
+
+#define  SDHCI_INT_NORMAL_MASK	0x00007FFF
+#define  SDHCI_INT_ERROR_MASK	0xFFFF8000
+
+#define  SDHCI_INT_CMD_MASK	(SDHCI_INT_RESPONSE | SDHCI_INT_TIMEOUT | \
+		SDHCI_INT_CRC | SDHCI_INT_END_BIT | SDHCI_INT_INDEX | \
+		SDHCI_INT_AUTO_CMD_ERR)
+#define  SDHCI_INT_DATA_MASK	(SDHCI_INT_DATA_END | SDHCI_INT_DMA_END | \
+		SDHCI_INT_DATA_AVAIL | SDHCI_INT_SPACE_AVAIL | \
+		SDHCI_INT_DATA_TIMEOUT | SDHCI_INT_DATA_CRC | \
+		SDHCI_INT_DATA_END_BIT | SDHCI_INT_ADMA_ERROR | \
+		SDHCI_INT_BLK_GAP)
+#define SDHCI_INT_ALL_MASK	((unsigned int)-1)
+
+#define SDHCI_CQE_INT_ERR_MASK ( \
+	SDHCI_INT_ADMA_ERROR | SDHCI_INT_BUS_POWER | SDHCI_INT_DATA_END_BIT | \
+	SDHCI_INT_DATA_CRC | SDHCI_INT_DATA_TIMEOUT | SDHCI_INT_INDEX | \
+	SDHCI_INT_END_BIT | SDHCI_INT_CRC | SDHCI_INT_TIMEOUT)
+
+#define SDHCI_CQE_INT_MASK (SDHCI_CQE_INT_ERR_MASK | SDHCI_INT_CQE)
+
+#define SDHCI_AUTO_CMD_STATUS	0x3C
+#define  SDHCI_AUTO_CMD_TIMEOUT	0x00000002
+#define  SDHCI_AUTO_CMD_CRC	0x00000004
+#define  SDHCI_AUTO_CMD_END_BIT	0x00000008
+#define  SDHCI_AUTO_CMD_INDEX	0x00000010
+
+#define SDHCI_HOST_CONTROL2		0x3E
+#define  SDHCI_CTRL_UHS_MASK		0x0007
+#define   SDHCI_CTRL_UHS_SDR12		0x0000
+#define   SDHCI_CTRL_UHS_SDR25		0x0001
+#define   SDHCI_CTRL_UHS_SDR50		0x0002
+#define   SDHCI_CTRL_UHS_SDR104		0x0003
+#define   SDHCI_CTRL_UHS_DDR50		0x0004
+#define   SDHCI_CTRL_HS400		0x0005 /* Non-standard */
+#define  SDHCI_CTRL_VDD_180		0x0008
+#define  SDHCI_CTRL_DRV_TYPE_MASK	0x0030
+#define   SDHCI_CTRL_DRV_TYPE_B		0x0000
+#define   SDHCI_CTRL_DRV_TYPE_A		0x0010
+#define   SDHCI_CTRL_DRV_TYPE_C		0x0020
+#define   SDHCI_CTRL_DRV_TYPE_D		0x0030
+#define  SDHCI_CTRL_EXEC_TUNING		0x0040
+#define  SDHCI_CTRL_TUNED_CLK		0x0080
+#define  SDHCI_CMD23_ENABLE		0x0800
+#define  SDHCI_CTRL_V4_MODE		0x1000
+#define  SDHCI_CTRL_64BIT_ADDR		0x2000
+#define  SDHCI_CTRL_PRESET_VAL_ENABLE	0x8000
+
+#define SDHCI_CAPABILITIES	0x40
+#define  SDHCI_TIMEOUT_CLK_MASK		GENMASK(5, 0)
+#define  SDHCI_TIMEOUT_CLK_UNIT	0x00000080
+#define  SDHCI_CLOCK_BASE_MASK		GENMASK(13, 8)
+#define  SDHCI_CLOCK_V3_BASE_MASK	GENMASK(15, 8)
+#define  SDHCI_MAX_BLOCK_MASK	0x00030000
+#define  SDHCI_MAX_BLOCK_SHIFT  16
+#define  SDHCI_CAN_DO_8BIT	0x00040000
+#define  SDHCI_CAN_DO_ADMA2	0x00080000
+#define  SDHCI_CAN_DO_ADMA1	0x00100000
+#define  SDHCI_CAN_DO_HISPD	0x00200000
+#define  SDHCI_CAN_DO_SDMA	0x00400000
+#define  SDHCI_CAN_DO_SUSPEND	0x00800000
+#define  SDHCI_CAN_VDD_330	0x01000000
+#define  SDHCI_CAN_VDD_300	0x02000000
+#define  SDHCI_CAN_VDD_180	0x04000000
+#define  SDHCI_CAN_64BIT_V4	0x08000000
+#define  SDHCI_CAN_64BIT	0x10000000
+
+#define SDHCI_CAPABILITIES_1	0x44
+#define  SDHCI_SUPPORT_SDR50	0x00000001
+#define  SDHCI_SUPPORT_SDR104	0x00000002
+#define  SDHCI_SUPPORT_DDR50	0x00000004
+#define  SDHCI_DRIVER_TYPE_A	0x00000010
+#define  SDHCI_DRIVER_TYPE_C	0x00000020
+#define  SDHCI_DRIVER_TYPE_D	0x00000040
+#define  SDHCI_RETUNING_TIMER_COUNT_MASK	GENMASK(11, 8)
+#define  SDHCI_USE_SDR50_TUNING			0x00002000
+#define  SDHCI_RETUNING_MODE_MASK		GENMASK(15, 14)
+#define  SDHCI_CLOCK_MUL_MASK			GENMASK(23, 16)
+#define  SDHCI_CAN_DO_ADMA3	0x08000000
+#define  SDHCI_SUPPORT_HS400	0x80000000 /* Non-standard */
+
+#define SDHCI_MAX_CURRENT		0x48
+#define  SDHCI_MAX_CURRENT_LIMIT	GENMASK(7, 0)
+#define  SDHCI_MAX_CURRENT_330_MASK	GENMASK(7, 0)
+#define  SDHCI_MAX_CURRENT_300_MASK	GENMASK(15, 8)
+#define  SDHCI_MAX_CURRENT_180_MASK	GENMASK(23, 16)
+#define   SDHCI_MAX_CURRENT_MULTIPLIER	4
+
+/* 4C-4F reserved for more max current */
+
+#define SDHCI_SET_ACMD12_ERROR	0x50
+#define SDHCI_SET_INT_ERROR	0x52
+
+#define SDHCI_ADMA_ERROR	0x54
+
+/* 55-57 reserved */
+
+#define SDHCI_ADMA_ADDRESS	0x58
+#define SDHCI_ADMA_ADDRESS_HI	0x5C
+
+/* 60-FB reserved */
+
+#define SDHCI_PRESET_FOR_HIGH_SPEED	0x64
+#define SDHCI_PRESET_FOR_SDR12 0x66
+#define SDHCI_PRESET_FOR_SDR25 0x68
+#define SDHCI_PRESET_FOR_SDR50 0x6A
+#define SDHCI_PRESET_FOR_SDR104        0x6C
+#define SDHCI_PRESET_FOR_DDR50 0x6E
+#define SDHCI_PRESET_FOR_HS400 0x74 /* Non-standard */
+#define SDHCI_PRESET_DRV_MASK		GENMASK(15, 14)
+#define SDHCI_PRESET_CLKGEN_SEL		BIT(10)
+#define SDHCI_PRESET_SDCLK_FREQ_MASK	GENMASK(9, 0)
+
+#define SDHCI_SLOT_INT_STATUS	0xFC
+
+#define SDHCI_HOST_VERSION	0xFE
+#define  SDHCI_VENDOR_VER_MASK	0xFF00
+#define  SDHCI_VENDOR_VER_SHIFT	8
+#define  SDHCI_SPEC_VER_MASK	0x00FF
+#define  SDHCI_SPEC_VER_SHIFT	0
+#define   SDHCI_SPEC_100	0
+#define   SDHCI_SPEC_200	1
+#define   SDHCI_SPEC_300	2
+#define   SDHCI_SPEC_400	3
+#define   SDHCI_SPEC_410	4
+#define   SDHCI_SPEC_420	5
+
+/*
+ * End of controller registers.
+ */
+
+#define SDHCI_MAX_DIV_SPEC_200	256
+#define SDHCI_MAX_DIV_SPEC_300	2046
+
+/*
+ * Host SDMA buffer boundary. Valid values from 4K to 512K in powers of 2.
+ */
+#define SDHCI_DEFAULT_BOUNDARY_SIZE  (512 * 1024)
+#define SDHCI_DEFAULT_BOUNDARY_ARG   (ilog2(SDHCI_DEFAULT_BOUNDARY_SIZE) - 12)
+
+/* ADMA2 32-bit DMA descriptor size */
+#define SDHCI_ADMA2_32_DESC_SZ	8
+
+/* ADMA2 32-bit descriptor */
+struct sdhci_adma2_32_desc {
+	__le16	cmd;
+	__le16	len;
+	__le32	addr;
+}  __packed __aligned(4);
+
+/* ADMA2 data alignment */
+#define SDHCI_ADMA2_ALIGN	4
+#define SDHCI_ADMA2_MASK	(SDHCI_ADMA2_ALIGN - 1)
+
+/*
+ * ADMA2 descriptor alignment.  Some controllers (e.g. Intel) require 8 byte
+ * alignment for the descriptor table even in 32-bit DMA mode.  Memory
+ * allocation is at least 8 byte aligned anyway, so just stipulate 8 always.
+ */
+#define SDHCI_ADMA2_DESC_ALIGN	8
+
+/*
+ * ADMA2 64-bit DMA descriptor size
+ * According to SD Host Controller spec v4.10, there are two kinds of
+ * descriptors for 64-bit addressing mode: 96-bit Descriptor and 128-bit
+ * Descriptor, if Host Version 4 Enable is set in the Host Control 2
+ * register, 128-bit Descriptor will be selected.
+ */
+#define SDHCI_ADMA2_64_DESC_SZ(host)	((host)->v4_mode ? 16 : 12)
+
+/*
+ * ADMA2 64-bit descriptor. Note 12-byte descriptor can't always be 8-byte
+ * aligned.
+ */
+struct sdhci_adma2_64_desc {
+	__le16	cmd;
+	__le16	len;
+	__le32	addr_lo;
+	__le32	addr_hi;
+}  __packed __aligned(4);
+
+#define ADMA2_TRAN_VALID	0x21
+#define ADMA2_NOP_END_VALID	0x3
+#define ADMA2_END		0x2
+
+/*
+ * Maximum segments assuming a 512KiB maximum requisition size and a minimum
+ * 4KiB page size. Note this also allows enough for multiple descriptors in
+ * case of PAGE_SIZE >= 64KiB.
+ */
+#define SDHCI_MAX_SEGS		128
+
+/* Allow for a a command request and a data request at the same time */
+#define SDHCI_MAX_MRQS		2
+
+/*
+ * 48bit command and 136 bit response in 100KHz clock could take upto 2.48ms.
+ * However since the start time of the command, the time between
+ * command and response, and the time between response and start of data is
+ * not known, set the command transfer time to 10ms.
+ */
+#define MMC_CMD_TRANSFER_TIME	(10 * NSEC_PER_MSEC) /* max 10 ms */
+
+enum sdhci_cookie {
+	COOKIE_UNMAPPED,
+	COOKIE_PRE_MAPPED,	/* mapped by sdhci_pre_req() */
+	COOKIE_MAPPED,		/* mapped by sdhci_prepare_data() */
+};
+
+struct sdhci_host {
+	/* Data set by hardware interface driver */
+	const char *hw_name;	/* Hardware bus name */
+
+	unsigned int quirks;	/* Deviations from spec. */
+
+/* Controller doesn't honor resets unless we touch the clock register */
+#define SDHCI_QUIRK_CLOCK_BEFORE_RESET			(1<<0)
+/* Controller has bad caps bits, but really supports DMA */
+#define SDHCI_QUIRK_FORCE_DMA				(1<<1)
+/* Controller doesn't like to be reset when there is no card inserted. */
+#define SDHCI_QUIRK_NO_CARD_NO_RESET			(1<<2)
+/* Controller doesn't like clearing the power reg before a change */
+#define SDHCI_QUIRK_SINGLE_POWER_WRITE			(1<<3)
+/* Controller has flaky internal state so reset it on each ios change */
+#define SDHCI_QUIRK_RESET_CMD_DATA_ON_IOS		(1<<4)
+/* Controller has an unusable DMA engine */
+#define SDHCI_QUIRK_BROKEN_DMA				(1<<5)
+/* Controller has an unusable ADMA engine */
+#define SDHCI_QUIRK_BROKEN_ADMA				(1<<6)
+/* Controller can only DMA from 32-bit aligned addresses */
+#define SDHCI_QUIRK_32BIT_DMA_ADDR			(1<<7)
+/* Controller can only DMA chunk sizes that are a multiple of 32 bits */
+#define SDHCI_QUIRK_32BIT_DMA_SIZE			(1<<8)
+/* Controller can only ADMA chunks that are a multiple of 32 bits */
+#define SDHCI_QUIRK_32BIT_ADMA_SIZE			(1<<9)
+/* Controller needs to be reset after each request to stay stable */
+#define SDHCI_QUIRK_RESET_AFTER_REQUEST			(1<<10)
+/* Controller needs voltage and power writes to happen separately */
+#define SDHCI_QUIRK_NO_SIMULT_VDD_AND_POWER		(1<<11)
+/* Controller provides an incorrect timeout value for transfers */
+#define SDHCI_QUIRK_BROKEN_TIMEOUT_VAL			(1<<12)
+/* Controller has an issue with buffer bits for small transfers */
+#define SDHCI_QUIRK_BROKEN_SMALL_PIO			(1<<13)
+/* Controller does not provide transfer-complete interrupt when not busy */
+#define SDHCI_QUIRK_NO_BUSY_IRQ				(1<<14)
+/* Controller has unreliable card detection */
+#define SDHCI_QUIRK_BROKEN_CARD_DETECTION		(1<<15)
+/* Controller reports inverted write-protect state */
+#define SDHCI_QUIRK_INVERTED_WRITE_PROTECT		(1<<16)
+/* Controller has unusable command queue engine */
+#define SDHCI_QUIRK_BROKEN_CQE				(1<<17)
+/* Controller does not like fast PIO transfers */
+#define SDHCI_QUIRK_PIO_NEEDS_DELAY			(1<<18)
+/* Controller does not have a LED */
+#define SDHCI_QUIRK_NO_LED				(1<<19)
+/* Controller has to be forced to use block size of 2048 bytes */
+#define SDHCI_QUIRK_FORCE_BLK_SZ_2048			(1<<20)
+/* Controller cannot do multi-block transfers */
+#define SDHCI_QUIRK_NO_MULTIBLOCK			(1<<21)
+/* Controller can only handle 1-bit data transfers */
+#define SDHCI_QUIRK_FORCE_1_BIT_DATA			(1<<22)
+/* Controller needs 10ms delay between applying power and clock */
+#define SDHCI_QUIRK_DELAY_AFTER_POWER			(1<<23)
+/* Controller uses SDCLK instead of TMCLK for data timeouts */
+#define SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK		(1<<24)
+/* Controller reports wrong base clock capability */
+#define SDHCI_QUIRK_CAP_CLOCK_BASE_BROKEN		(1<<25)
+/* Controller cannot support End Attribute in NOP ADMA descriptor */
+#define SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC		(1<<26)
+/* Controller is missing device caps. Use caps provided by host */
+#define SDHCI_QUIRK_MISSING_CAPS			(1<<27)
+/* Controller uses Auto CMD12 command to stop the transfer */
+#define SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12		(1<<28)
+/* Controller doesn't have HISPD bit field in HI-SPEED SD card */
+#define SDHCI_QUIRK_NO_HISPD_BIT			(1<<29)
+/* Controller treats ADMA descriptors with length 0000h incorrectly */
+#define SDHCI_QUIRK_BROKEN_ADMA_ZEROLEN_DESC		(1<<30)
+/* The read-only detection via SDHCI_PRESENT_STATE register is unstable */
+#define SDHCI_QUIRK_UNSTABLE_RO_DETECT			(1<<31)
+
+	unsigned int quirks2;	/* More deviations from spec. */
+
+#define SDHCI_QUIRK2_HOST_OFF_CARD_ON			(1<<0)
+#define SDHCI_QUIRK2_HOST_NO_CMD23			(1<<1)
+/* The system physically doesn't support 1.8v, even if the host does */
+#define SDHCI_QUIRK2_NO_1_8_V				(1<<2)
+#define SDHCI_QUIRK2_PRESET_VALUE_BROKEN		(1<<3)
+#define SDHCI_QUIRK2_CARD_ON_NEEDS_BUS_ON		(1<<4)
+/* Controller has a non-standard host control register */
+#define SDHCI_QUIRK2_BROKEN_HOST_CONTROL		(1<<5)
+/* Controller does not support HS200 */
+#define SDHCI_QUIRK2_BROKEN_HS200			(1<<6)
+/* Controller does not support DDR50 */
+#define SDHCI_QUIRK2_BROKEN_DDR50			(1<<7)
+/* Stop command (CMD12) can set Transfer Complete when not using MMC_RSP_BUSY */
+#define SDHCI_QUIRK2_STOP_WITH_TC			(1<<8)
+/* Controller does not support 64-bit DMA */
+#define SDHCI_QUIRK2_BROKEN_64_BIT_DMA			(1<<9)
+/* need clear transfer mode register before send cmd */
+#define SDHCI_QUIRK2_CLEAR_TRANSFERMODE_REG_BEFORE_CMD	(1<<10)
+/* Capability register bit-63 indicates HS400 support */
+#define SDHCI_QUIRK2_CAPS_BIT63_FOR_HS400		(1<<11)
+/* forced tuned clock */
+#define SDHCI_QUIRK2_TUNING_WORK_AROUND			(1<<12)
+/* disable the block count for single block transactions */
+#define SDHCI_QUIRK2_SUPPORT_SINGLE			(1<<13)
+/* Controller broken with using ACMD23 */
+#define SDHCI_QUIRK2_ACMD23_BROKEN			(1<<14)
+/* Broken Clock divider zero in controller */
+#define SDHCI_QUIRK2_CLOCK_DIV_ZERO_BROKEN		(1<<15)
+/* Controller has CRC in 136 bit Command Response */
+#define SDHCI_QUIRK2_RSP_136_HAS_CRC			(1<<16)
+/*
+ * Disable HW timeout if the requested timeout is more than the maximum
+ * obtainable timeout.
+ */
+#define SDHCI_QUIRK2_DISABLE_HW_TIMEOUT			(1<<17)
+/*
+ * 32-bit block count may not support eMMC where upper bits of CMD23 are used
+ * for other purposes.  Consequently we support 16-bit block count by default.
+ * Otherwise, SDHCI_QUIRK2_USE_32BIT_BLK_CNT can be selected to use 32-bit
+ * block count.
+ */
+#define SDHCI_QUIRK2_USE_32BIT_BLK_CNT			(1<<18)
+
+	int irq;		/* Device IRQ */
+	void __iomem *ioaddr;	/* Mapped address */
+	phys_addr_t mapbase;	/* physical address base */
+	char *bounce_buffer;	/* For packing SDMA reads/writes */
+	dma_addr_t bounce_addr;
+	unsigned int bounce_buffer_size;
+
+	const struct sdhci_ops *ops;	/* Low level hw interface */
+
+	/* Internal data */
+	struct mmc_host *mmc;	/* MMC structure */
+	struct mmc_host_ops mmc_host_ops;	/* MMC host ops */
+	u64 dma_mask;		/* custom DMA mask */
+
+#if IS_ENABLED(CONFIG_LEDS_CLASS)
+	struct led_classdev led;	/* LED control */
+	char led_name[32];
+#endif
+
+	spinlock_t lock;	/* Mutex */
+
+	int flags;		/* Host attributes */
+#define SDHCI_USE_SDMA		(1<<0)	/* Host is SDMA capable */
+#define SDHCI_USE_ADMA		(1<<1)	/* Host is ADMA capable */
+#define SDHCI_REQ_USE_DMA	(1<<2)	/* Use DMA for this req. */
+#define SDHCI_DEVICE_DEAD	(1<<3)	/* Device unresponsive */
+#define SDHCI_SDR50_NEEDS_TUNING (1<<4)	/* SDR50 needs tuning */
+#define SDHCI_AUTO_CMD12	(1<<6)	/* Auto CMD12 support */
+#define SDHCI_AUTO_CMD23	(1<<7)	/* Auto CMD23 support */
+#define SDHCI_PV_ENABLED	(1<<8)	/* Preset value enabled */
+#define SDHCI_USE_64_BIT_DMA	(1<<12)	/* Use 64-bit DMA */
+#define SDHCI_HS400_TUNING	(1<<13)	/* Tuning for HS400 */
+#define SDHCI_SIGNALING_330	(1<<14)	/* Host is capable of 3.3V signaling */
+#define SDHCI_SIGNALING_180	(1<<15)	/* Host is capable of 1.8V signaling */
+#define SDHCI_SIGNALING_120	(1<<16)	/* Host is capable of 1.2V signaling */
+
+	unsigned int version;	/* SDHCI spec. version */
+
+	unsigned int max_clk;	/* Max possible freq (MHz) */
+	unsigned int timeout_clk;	/* Timeout freq (KHz) */
+	unsigned int clk_mul;	/* Clock Muliplier value */
+
+	unsigned int clock;	/* Current clock (MHz) */
+	u8 pwr;			/* Current voltage */
+	u8 drv_type;		/* Current UHS-I driver type */
+	bool reinit_uhs;	/* Force UHS-related re-initialization */
+
+	bool runtime_suspended;	/* Host is runtime suspended */
+	bool bus_on;		/* Bus power prevents runtime suspend */
+	bool preset_enabled;	/* Preset is enabled */
+	bool pending_reset;	/* Cmd/data reset is pending */
+	bool irq_wake_enabled;	/* IRQ wakeup is enabled */
+	bool v4_mode;		/* Host Version 4 Enable */
+	bool use_external_dma;	/* Host selects to use external DMA */
+	bool always_defer_done;	/* Always defer to complete requests */
+
+	struct mmc_request *mrqs_done[SDHCI_MAX_MRQS];	/* Requests done */
+	struct mmc_command *cmd;	/* Current command */
+	struct mmc_command *data_cmd;	/* Current data command */
+	struct mmc_command *deferred_cmd;	/* Deferred command */
+	struct mmc_data *data;	/* Current data request */
+	unsigned int data_early:1;	/* Data finished before cmd */
+
+	struct sg_mapping_iter sg_miter;	/* SG state for PIO */
+	unsigned int blocks;	/* remaining PIO blocks */
+
+	int sg_count;		/* Mapped sg entries */
+	int max_adma;		/* Max. length in ADMA descriptor */
+
+	void *adma_table;	/* ADMA descriptor table */
+	void *align_buffer;	/* Bounce buffer */
+
+	size_t adma_table_sz;	/* ADMA descriptor table size */
+	size_t align_buffer_sz;	/* Bounce buffer size */
+
+	dma_addr_t adma_addr;	/* Mapped ADMA descr. table */
+	dma_addr_t align_addr;	/* Mapped bounce buffer */
+
+	unsigned int desc_sz;	/* ADMA current descriptor size */
+	unsigned int alloc_desc_sz;	/* ADMA descr. max size host supports */
+
+	struct workqueue_struct *complete_wq;	/* Request completion wq */
+	struct work_struct	complete_work;	/* Request completion work */
+
+	struct timer_list timer;	/* Timer for timeouts */
+	struct timer_list data_timer;	/* Timer for data timeouts */
+
+#if IS_ENABLED(CONFIG_MMC_SDHCI_EXTERNAL_DMA)
+	struct dma_chan *rx_chan;
+	struct dma_chan *tx_chan;
+#endif
+
+	u32 caps;		/* CAPABILITY_0 */
+	u32 caps1;		/* CAPABILITY_1 */
+	bool read_caps;		/* Capability flags have been read */
+
+	bool sdhci_core_to_disable_vqmmc;  /* sdhci core can disable vqmmc */
+	unsigned int            ocr_avail_sdio;	/* OCR bit masks */
+	unsigned int            ocr_avail_sd;
+	unsigned int            ocr_avail_mmc;
+	u32 ocr_mask;		/* available voltages */
+
+	unsigned		timing;		/* Current timing */
+
+	u32			thread_isr;
+
+	/* cached registers */
+	u32			ier;
+
+	bool			cqe_on;		/* CQE is operating */
+	u32			cqe_ier;	/* CQE interrupt mask */
+	u32			cqe_err_ier;	/* CQE error interrupt mask */
+
+	wait_queue_head_t	buf_ready_int;	/* Waitqueue for Buffer Read Ready interrupt */
+	unsigned int		tuning_done;	/* Condition flag set when CMD19 succeeds */
+
+	unsigned int		tuning_count;	/* Timer count for re-tuning */
+	unsigned int		tuning_mode;	/* Re-tuning mode supported by host */
+	unsigned int		tuning_err;	/* Error code for re-tuning */
+#define SDHCI_TUNING_MODE_1	0
+#define SDHCI_TUNING_MODE_2	1
+#define SDHCI_TUNING_MODE_3	2
+	/* Delay (ms) between tuning commands */
+	int			tuning_delay;
+	int			tuning_loop_count;
+
+	/* Host SDMA buffer boundary. */
+	u32			sdma_boundary;
+
+	/* Host ADMA table count */
+	u32			adma_table_cnt;
+
+	u64			data_timeout;
+
+	unsigned long private[] ____cacheline_aligned;
+};
+
+struct sdhci_ops {
+#ifdef CONFIG_MMC_SDHCI_IO_ACCESSORS
+	u32		(*read_l)(struct sdhci_host *host, int reg);
+	u16		(*read_w)(struct sdhci_host *host, int reg);
+	u8		(*read_b)(struct sdhci_host *host, int reg);
+	void		(*write_l)(struct sdhci_host *host, u32 val, int reg);
+	void		(*write_w)(struct sdhci_host *host, u16 val, int reg);
+	void		(*write_b)(struct sdhci_host *host, u8 val, int reg);
+#endif
+
+	void	(*set_clock)(struct sdhci_host *host, unsigned int clock);
+	void	(*set_power)(struct sdhci_host *host, unsigned char mode,
+			     unsigned short vdd);
+
+	u32		(*irq)(struct sdhci_host *host, u32 intmask);
+
+	int		(*set_dma_mask)(struct sdhci_host *host);
+	int		(*enable_dma)(struct sdhci_host *host);
+	unsigned int	(*get_max_clock)(struct sdhci_host *host);
+	unsigned int	(*get_min_clock)(struct sdhci_host *host);
+	/* get_timeout_clock should return clk rate in unit of Hz */
+	unsigned int	(*get_timeout_clock)(struct sdhci_host *host);
+	unsigned int	(*get_max_timeout_count)(struct sdhci_host *host);
+	void		(*set_timeout)(struct sdhci_host *host,
+				       struct mmc_command *cmd);
+	void		(*set_bus_width)(struct sdhci_host *host, int width);
+	void (*platform_send_init_74_clocks)(struct sdhci_host *host,
+					     u8 power_mode);
+	unsigned int    (*get_ro)(struct sdhci_host *host);
+	void		(*reset)(struct sdhci_host *host, u8 mask);
+	int	(*platform_execute_tuning)(struct sdhci_host *host, u32 opcode);
+	void	(*set_uhs_signaling)(struct sdhci_host *host, unsigned int uhs);
+	void	(*hw_reset)(struct sdhci_host *host);
+	void    (*adma_workaround)(struct sdhci_host *host, u32 intmask);
+	void    (*card_event)(struct sdhci_host *host);
+	void	(*voltage_switch)(struct sdhci_host *host);
+	void	(*adma_write_desc)(struct sdhci_host *host, void **desc,
+				   dma_addr_t addr, int len, unsigned int cmd);
+	void	(*copy_to_bounce_buffer)(struct sdhci_host *host,
+					 struct mmc_data *data,
+					 unsigned int length);
+	void	(*request_done)(struct sdhci_host *host,
+				struct mmc_request *mrq);
+	void    (*dump_vendor_regs)(struct sdhci_host *host);
+};
+
+#ifdef CONFIG_MMC_SDHCI_IO_ACCESSORS
+
+static inline void sdhci_writel(struct sdhci_host *host, u32 val, int reg)
+{
+	if (unlikely(host->ops->write_l))
+		host->ops->write_l(host, val, reg);
+	else
+		writel(val, host->ioaddr + reg);
+}
+
+static inline void sdhci_writew(struct sdhci_host *host, u16 val, int reg)
+{
+	if (unlikely(host->ops->write_w))
+		host->ops->write_w(host, val, reg);
+	else
+		writew(val, host->ioaddr + reg);
+}
+
+static inline void sdhci_writeb(struct sdhci_host *host, u8 val, int reg)
+{
+	if (unlikely(host->ops->write_b))
+		host->ops->write_b(host, val, reg);
+	else
+		writeb(val, host->ioaddr + reg);
+}
+
+static inline u32 sdhci_readl(struct sdhci_host *host, int reg)
+{
+	if (unlikely(host->ops->read_l))
+		return host->ops->read_l(host, reg);
+	else
+		return readl(host->ioaddr + reg);
+}
+
+static inline u16 sdhci_readw(struct sdhci_host *host, int reg)
+{
+	if (unlikely(host->ops->read_w))
+		return host->ops->read_w(host, reg);
+	else
+		return readw(host->ioaddr + reg);
+}
+
+static inline u8 sdhci_readb(struct sdhci_host *host, int reg)
+{
+	if (unlikely(host->ops->read_b))
+		return host->ops->read_b(host, reg);
+	else
+		return readb(host->ioaddr + reg);
+}
+
+#else
+
+static inline void sdhci_writel(struct sdhci_host *host, u32 val, int reg)
+{
+	writel(val, host->ioaddr + reg);
+}
+
+static inline void sdhci_writew(struct sdhci_host *host, u16 val, int reg)
+{
+	writew(val, host->ioaddr + reg);
+}
+
+static inline void sdhci_writeb(struct sdhci_host *host, u8 val, int reg)
+{
+	writeb(val, host->ioaddr + reg);
+}
+
+static inline u32 sdhci_readl(struct sdhci_host *host, int reg)
+{
+	return readl(host->ioaddr + reg);
+}
+
+static inline u16 sdhci_readw(struct sdhci_host *host, int reg)
+{
+	return readw(host->ioaddr + reg);
+}
+
+static inline u8 sdhci_readb(struct sdhci_host *host, int reg)
+{
+	return readb(host->ioaddr + reg);
+}
+
+#endif /* CONFIG_MMC_SDHCI_IO_ACCESSORS */
+
+struct sdhci_host *sdhci_alloc_host(struct device *dev, size_t priv_size);
+void sdhci_free_host(struct sdhci_host *host);
+
+static inline void *sdhci_priv(struct sdhci_host *host)
+{
+	return host->private;
+}
+
+void sdhci_card_detect(struct sdhci_host *host);
+void __sdhci_read_caps(struct sdhci_host *host, const u16 *ver,
+		       const u32 *caps, const u32 *caps1);
+int sdhci_setup_host(struct sdhci_host *host);
+void sdhci_cleanup_host(struct sdhci_host *host);
+int __sdhci_add_host(struct sdhci_host *host);
+int sdhci_add_host(struct sdhci_host *host);
+void sdhci_remove_host(struct sdhci_host *host, int dead);
+
+static inline void sdhci_read_caps(struct sdhci_host *host)
+{
+	__sdhci_read_caps(host, NULL, NULL, NULL);
+}
+
+u16 sdhci_calc_clk(struct sdhci_host *host, unsigned int clock,
+		   unsigned int *actual_clock);
+void sdhci_set_clock(struct sdhci_host *host, unsigned int clock);
+void sdhci_enable_clk(struct sdhci_host *host, u16 clk);
+void sdhci_set_power(struct sdhci_host *host, unsigned char mode,
+		     unsigned short vdd);
+void sdhci_set_power_and_bus_voltage(struct sdhci_host *host,
+				     unsigned char mode,
+				     unsigned short vdd);
+void sdhci_set_power_noreg(struct sdhci_host *host, unsigned char mode,
+			   unsigned short vdd);
+void sdhci_request(struct mmc_host *mmc, struct mmc_request *mrq);
+int sdhci_request_atomic(struct mmc_host *mmc, struct mmc_request *mrq);
+void sdhci_set_bus_width(struct sdhci_host *host, int width);
+void sdhci_reset(struct sdhci_host *host, u8 mask);
+void sdhci_set_uhs_signaling(struct sdhci_host *host, unsigned timing);
+int sdhci_execute_tuning(struct mmc_host *mmc, u32 opcode);
+void sdhci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios);
+int sdhci_start_signal_voltage_switch(struct mmc_host *mmc,
+				      struct mmc_ios *ios);
+void sdhci_enable_sdio_irq(struct mmc_host *mmc, int enable);
+void sdhci_adma_write_desc(struct sdhci_host *host, void **desc,
+			   dma_addr_t addr, int len, unsigned int cmd);
+
+#ifdef CONFIG_PM
+int sdhci_suspend_host(struct sdhci_host *host);
+int sdhci_resume_host(struct sdhci_host *host);
+int sdhci_runtime_suspend_host(struct sdhci_host *host);
+int sdhci_runtime_resume_host(struct sdhci_host *host, int soft_reset);
+#endif
+
+void sdhci_cqe_enable(struct mmc_host *mmc);
+void sdhci_cqe_disable(struct mmc_host *mmc, bool recovery);
+bool sdhci_cqe_irq(struct sdhci_host *host, u32 intmask, int *cmd_error,
+		   int *data_error);
+
+void sdhci_dumpregs(struct sdhci_host *host);
+void sdhci_enable_v4_mode(struct sdhci_host *host);
+
+void sdhci_start_tuning(struct sdhci_host *host);
+void sdhci_end_tuning(struct sdhci_host *host);
+void sdhci_reset_tuning(struct sdhci_host *host);
+void sdhci_send_tuning(struct sdhci_host *host, u32 opcode);
+void sdhci_abort_tuning(struct sdhci_host *host, u32 opcode);
+void sdhci_switch_external_dma(struct sdhci_host *host, bool en);
+void sdhci_set_data_timeout_irq(struct sdhci_host *host, bool enable);
+void __sdhci_set_timeout(struct sdhci_host *host, struct mmc_command *cmd);
+
+#endif /* __SDHCI_HW_H */
diff --git a/drivers/mmc/host/sdhci_am654.c b/drivers/mmc/host/sdhci_am654.c
new file mode 100644
index 000000000..eb52e0c5a
--- /dev/null
+++ b/drivers/mmc/host/sdhci_am654.c
@@ -0,0 +1,881 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * sdhci_am654.c - SDHCI driver for TI's AM654 SOCs
+ *
+ * Copyright (C) 2018 Texas Instruments Incorporated - https://www.ti.com
+ *
+ */
+#include <linux/clk.h>
+#include <linux/iopoll.h>
+#include <linux/of.h>
+#include <linux/module.h>
+#include <linux/pm_runtime.h>
+#include <linux/property.h>
+#include <linux/regmap.h>
+#include <linux/sys_soc.h>
+
+#include "cqhci.h"
+#include "sdhci-cqhci.h"
+#include "sdhci-pltfm.h"
+
+/* CTL_CFG Registers */
+#define CTL_CFG_2		0x14
+#define CTL_CFG_3		0x18
+
+#define SLOTTYPE_MASK		GENMASK(31, 30)
+#define SLOTTYPE_EMBEDDED	BIT(30)
+#define TUNINGFORSDR50_MASK	BIT(13)
+
+/* PHY Registers */
+#define PHY_CTRL1	0x100
+#define PHY_CTRL2	0x104
+#define PHY_CTRL3	0x108
+#define PHY_CTRL4	0x10C
+#define PHY_CTRL5	0x110
+#define PHY_CTRL6	0x114
+#define PHY_STAT1	0x130
+#define PHY_STAT2	0x134
+
+#define IOMUX_ENABLE_SHIFT	31
+#define IOMUX_ENABLE_MASK	BIT(IOMUX_ENABLE_SHIFT)
+#define OTAPDLYENA_SHIFT	20
+#define OTAPDLYENA_MASK		BIT(OTAPDLYENA_SHIFT)
+#define OTAPDLYSEL_SHIFT	12
+#define OTAPDLYSEL_MASK		GENMASK(15, 12)
+#define STRBSEL_SHIFT		24
+#define STRBSEL_4BIT_MASK	GENMASK(27, 24)
+#define STRBSEL_8BIT_MASK	GENMASK(31, 24)
+#define SEL50_SHIFT		8
+#define SEL50_MASK		BIT(SEL50_SHIFT)
+#define SEL100_SHIFT		9
+#define SEL100_MASK		BIT(SEL100_SHIFT)
+#define FREQSEL_SHIFT		8
+#define FREQSEL_MASK		GENMASK(10, 8)
+#define CLKBUFSEL_SHIFT		0
+#define CLKBUFSEL_MASK		GENMASK(2, 0)
+#define DLL_TRIM_ICP_SHIFT	4
+#define DLL_TRIM_ICP_MASK	GENMASK(7, 4)
+#define DR_TY_SHIFT		20
+#define DR_TY_MASK		GENMASK(22, 20)
+#define ENDLL_SHIFT		1
+#define ENDLL_MASK		BIT(ENDLL_SHIFT)
+#define DLLRDY_SHIFT		0
+#define DLLRDY_MASK		BIT(DLLRDY_SHIFT)
+#define PDB_SHIFT		0
+#define PDB_MASK		BIT(PDB_SHIFT)
+#define CALDONE_SHIFT		1
+#define CALDONE_MASK		BIT(CALDONE_SHIFT)
+#define RETRIM_SHIFT		17
+#define RETRIM_MASK		BIT(RETRIM_SHIFT)
+#define SELDLYTXCLK_SHIFT	17
+#define SELDLYTXCLK_MASK	BIT(SELDLYTXCLK_SHIFT)
+#define SELDLYRXCLK_SHIFT	16
+#define SELDLYRXCLK_MASK	BIT(SELDLYRXCLK_SHIFT)
+#define ITAPDLYSEL_SHIFT	0
+#define ITAPDLYSEL_MASK		GENMASK(4, 0)
+#define ITAPDLYENA_SHIFT	8
+#define ITAPDLYENA_MASK		BIT(ITAPDLYENA_SHIFT)
+#define ITAPCHGWIN_SHIFT	9
+#define ITAPCHGWIN_MASK		BIT(ITAPCHGWIN_SHIFT)
+
+#define DRIVER_STRENGTH_50_OHM	0x0
+#define DRIVER_STRENGTH_33_OHM	0x1
+#define DRIVER_STRENGTH_66_OHM	0x2
+#define DRIVER_STRENGTH_100_OHM	0x3
+#define DRIVER_STRENGTH_40_OHM	0x4
+
+#define CLOCK_TOO_SLOW_HZ	50000000
+
+/* Command Queue Host Controller Interface Base address */
+#define SDHCI_AM654_CQE_BASE_ADDR 0x200
+
+static struct regmap_config sdhci_am654_regmap_config = {
+	.reg_bits = 32,
+	.val_bits = 32,
+	.reg_stride = 4,
+	.fast_io = true,
+};
+
+struct timing_data {
+	const char *otap_binding;
+	const char *itap_binding;
+	u32 capability;
+};
+
+static const struct timing_data td[] = {
+	[MMC_TIMING_LEGACY]	= {"ti,otap-del-sel-legacy",
+				   "ti,itap-del-sel-legacy",
+				   0},
+	[MMC_TIMING_MMC_HS]	= {"ti,otap-del-sel-mmc-hs",
+				   "ti,itap-del-sel-mmc-hs",
+				   MMC_CAP_MMC_HIGHSPEED},
+	[MMC_TIMING_SD_HS]	= {"ti,otap-del-sel-sd-hs",
+				   "ti,itap-del-sel-sd-hs",
+				   MMC_CAP_SD_HIGHSPEED},
+	[MMC_TIMING_UHS_SDR12]	= {"ti,otap-del-sel-sdr12",
+				   "ti,itap-del-sel-sdr12",
+				   MMC_CAP_UHS_SDR12},
+	[MMC_TIMING_UHS_SDR25]	= {"ti,otap-del-sel-sdr25",
+				   "ti,itap-del-sel-sdr25",
+				   MMC_CAP_UHS_SDR25},
+	[MMC_TIMING_UHS_SDR50]	= {"ti,otap-del-sel-sdr50",
+				   NULL,
+				   MMC_CAP_UHS_SDR50},
+	[MMC_TIMING_UHS_SDR104]	= {"ti,otap-del-sel-sdr104",
+				   NULL,
+				   MMC_CAP_UHS_SDR104},
+	[MMC_TIMING_UHS_DDR50]	= {"ti,otap-del-sel-ddr50",
+				   NULL,
+				   MMC_CAP_UHS_DDR50},
+	[MMC_TIMING_MMC_DDR52]	= {"ti,otap-del-sel-ddr52",
+				   "ti,itap-del-sel-ddr52",
+				   MMC_CAP_DDR},
+	[MMC_TIMING_MMC_HS200]	= {"ti,otap-del-sel-hs200",
+				   NULL,
+				   MMC_CAP2_HS200},
+	[MMC_TIMING_MMC_HS400]	= {"ti,otap-del-sel-hs400",
+				   NULL,
+				   MMC_CAP2_HS400},
+};
+
+struct sdhci_am654_data {
+	struct regmap *base;
+	bool legacy_otapdly;
+	int otap_del_sel[ARRAY_SIZE(td)];
+	int itap_del_sel[ARRAY_SIZE(td)];
+	int clkbuf_sel;
+	int trm_icp;
+	int drv_strength;
+	int strb_sel;
+	u32 flags;
+	u32 quirks;
+
+#define SDHCI_AM654_QUIRK_FORCE_CDTEST BIT(0)
+};
+
+struct sdhci_am654_driver_data {
+	const struct sdhci_pltfm_data *pdata;
+	u32 flags;
+#define IOMUX_PRESENT	(1 << 0)
+#define FREQSEL_2_BIT	(1 << 1)
+#define STRBSEL_4_BIT	(1 << 2)
+#define DLL_PRESENT	(1 << 3)
+#define DLL_CALIB	(1 << 4)
+};
+
+static void sdhci_am654_setup_dll(struct sdhci_host *host, unsigned int clock)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_am654_data *sdhci_am654 = sdhci_pltfm_priv(pltfm_host);
+	int sel50, sel100, freqsel;
+	u32 mask, val;
+	int ret;
+
+	/* Disable delay chain mode */
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL5,
+			   SELDLYTXCLK_MASK | SELDLYRXCLK_MASK, 0);
+
+	if (sdhci_am654->flags & FREQSEL_2_BIT) {
+		switch (clock) {
+		case 200000000:
+			sel50 = 0;
+			sel100 = 0;
+			break;
+		case 100000000:
+			sel50 = 0;
+			sel100 = 1;
+			break;
+		default:
+			sel50 = 1;
+			sel100 = 0;
+		}
+
+		/* Configure PHY DLL frequency */
+		mask = SEL50_MASK | SEL100_MASK;
+		val = (sel50 << SEL50_SHIFT) | (sel100 << SEL100_SHIFT);
+		regmap_update_bits(sdhci_am654->base, PHY_CTRL5, mask, val);
+
+	} else {
+		switch (clock) {
+		case 200000000:
+			freqsel = 0x0;
+			break;
+		default:
+			freqsel = 0x4;
+		}
+
+		regmap_update_bits(sdhci_am654->base, PHY_CTRL5, FREQSEL_MASK,
+				   freqsel << FREQSEL_SHIFT);
+	}
+	/* Configure DLL TRIM */
+	mask = DLL_TRIM_ICP_MASK;
+	val = sdhci_am654->trm_icp << DLL_TRIM_ICP_SHIFT;
+
+	/* Configure DLL driver strength */
+	mask |= DR_TY_MASK;
+	val |= sdhci_am654->drv_strength << DR_TY_SHIFT;
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL1, mask, val);
+
+	/* Enable DLL */
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL1, ENDLL_MASK,
+			   0x1 << ENDLL_SHIFT);
+	/*
+	 * Poll for DLL ready. Use a one second timeout.
+	 * Works in all experiments done so far
+	 */
+	ret = regmap_read_poll_timeout(sdhci_am654->base, PHY_STAT1, val,
+				       val & DLLRDY_MASK, 1000, 1000000);
+	if (ret) {
+		dev_err(mmc_dev(host->mmc), "DLL failed to relock\n");
+		return;
+	}
+}
+
+static void sdhci_am654_write_itapdly(struct sdhci_am654_data *sdhci_am654,
+				      u32 itapdly)
+{
+	/* Set ITAPCHGWIN before writing to ITAPDLY */
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL4, ITAPCHGWIN_MASK,
+			   1 << ITAPCHGWIN_SHIFT);
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL4, ITAPDLYSEL_MASK,
+			   itapdly << ITAPDLYSEL_SHIFT);
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL4, ITAPCHGWIN_MASK, 0);
+}
+
+static void sdhci_am654_setup_delay_chain(struct sdhci_am654_data *sdhci_am654,
+					  unsigned char timing)
+{
+	u32 mask, val;
+
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL1, ENDLL_MASK, 0);
+
+	val = 1 << SELDLYTXCLK_SHIFT | 1 << SELDLYRXCLK_SHIFT;
+	mask = SELDLYTXCLK_MASK | SELDLYRXCLK_MASK;
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL5, mask, val);
+
+	sdhci_am654_write_itapdly(sdhci_am654,
+				  sdhci_am654->itap_del_sel[timing]);
+}
+
+static void sdhci_am654_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_am654_data *sdhci_am654 = sdhci_pltfm_priv(pltfm_host);
+	unsigned char timing = host->mmc->ios.timing;
+	u32 otap_del_sel;
+	u32 otap_del_ena;
+	u32 mask, val;
+
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL1, ENDLL_MASK, 0);
+
+	sdhci_set_clock(host, clock);
+
+	/* Setup DLL Output TAP delay */
+	if (sdhci_am654->legacy_otapdly)
+		otap_del_sel = sdhci_am654->otap_del_sel[0];
+	else
+		otap_del_sel = sdhci_am654->otap_del_sel[timing];
+
+	otap_del_ena = (timing > MMC_TIMING_UHS_SDR25) ? 1 : 0;
+
+	mask = OTAPDLYENA_MASK | OTAPDLYSEL_MASK;
+	val = (otap_del_ena << OTAPDLYENA_SHIFT) |
+	      (otap_del_sel << OTAPDLYSEL_SHIFT);
+
+	/* Write to STRBSEL for HS400 speed mode */
+	if (timing == MMC_TIMING_MMC_HS400) {
+		if (sdhci_am654->flags & STRBSEL_4_BIT)
+			mask |= STRBSEL_4BIT_MASK;
+		else
+			mask |= STRBSEL_8BIT_MASK;
+
+		val |= sdhci_am654->strb_sel << STRBSEL_SHIFT;
+	}
+
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL4, mask, val);
+
+	if (timing > MMC_TIMING_UHS_SDR25 && clock >= CLOCK_TOO_SLOW_HZ)
+		sdhci_am654_setup_dll(host, clock);
+	else
+		sdhci_am654_setup_delay_chain(sdhci_am654, timing);
+
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL5, CLKBUFSEL_MASK,
+			   sdhci_am654->clkbuf_sel);
+}
+
+static void sdhci_j721e_4bit_set_clock(struct sdhci_host *host,
+				       unsigned int clock)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_am654_data *sdhci_am654 = sdhci_pltfm_priv(pltfm_host);
+	unsigned char timing = host->mmc->ios.timing;
+	u32 otap_del_sel;
+	u32 mask, val;
+
+	/* Setup DLL Output TAP delay */
+	if (sdhci_am654->legacy_otapdly)
+		otap_del_sel = sdhci_am654->otap_del_sel[0];
+	else
+		otap_del_sel = sdhci_am654->otap_del_sel[timing];
+
+	mask = OTAPDLYENA_MASK | OTAPDLYSEL_MASK;
+	val = (0x1 << OTAPDLYENA_SHIFT) |
+	      (otap_del_sel << OTAPDLYSEL_SHIFT);
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL4, mask, val);
+
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL5, CLKBUFSEL_MASK,
+			   sdhci_am654->clkbuf_sel);
+
+	sdhci_set_clock(host, clock);
+}
+
+static u8 sdhci_am654_write_power_on(struct sdhci_host *host, u8 val, int reg)
+{
+	writeb(val, host->ioaddr + reg);
+	usleep_range(1000, 10000);
+	return readb(host->ioaddr + reg);
+}
+
+#define MAX_POWER_ON_TIMEOUT	1500000 /* us */
+static void sdhci_am654_write_b(struct sdhci_host *host, u8 val, int reg)
+{
+	unsigned char timing = host->mmc->ios.timing;
+	u8 pwr;
+	int ret;
+
+	if (reg == SDHCI_HOST_CONTROL) {
+		switch (timing) {
+		/*
+		 * According to the data manual, HISPD bit
+		 * should not be set in these speed modes.
+		 */
+		case MMC_TIMING_SD_HS:
+		case MMC_TIMING_MMC_HS:
+			val &= ~SDHCI_CTRL_HISPD;
+		}
+	}
+
+	writeb(val, host->ioaddr + reg);
+	if (reg == SDHCI_POWER_CONTROL && (val & SDHCI_POWER_ON)) {
+		/*
+		 * Power on will not happen until the card detect debounce
+		 * timer expires. Wait at least 1.5 seconds for the power on
+		 * bit to be set
+		 */
+		ret = read_poll_timeout(sdhci_am654_write_power_on, pwr,
+					pwr & SDHCI_POWER_ON, 0,
+					MAX_POWER_ON_TIMEOUT, false, host, val,
+					reg);
+		if (ret)
+			dev_info(mmc_dev(host->mmc), "Power on failed\n");
+	}
+}
+
+static void sdhci_am654_reset(struct sdhci_host *host, u8 mask)
+{
+	u8 ctrl;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_am654_data *sdhci_am654 = sdhci_pltfm_priv(pltfm_host);
+
+	sdhci_and_cqhci_reset(host, mask);
+
+	if (sdhci_am654->quirks & SDHCI_AM654_QUIRK_FORCE_CDTEST) {
+		ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+		ctrl |= SDHCI_CTRL_CDTEST_INS | SDHCI_CTRL_CDTEST_EN;
+		sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+	}
+}
+
+static int sdhci_am654_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	int err = sdhci_execute_tuning(mmc, opcode);
+
+	if (err)
+		return err;
+	/*
+	 * Tuning data remains in the buffer after tuning.
+	 * Do a command and data reset to get rid of it
+	 */
+	sdhci_reset(host, SDHCI_RESET_CMD | SDHCI_RESET_DATA);
+
+	return 0;
+}
+
+static u32 sdhci_am654_cqhci_irq(struct sdhci_host *host, u32 intmask)
+{
+	int cmd_error = 0;
+	int data_error = 0;
+
+	if (!sdhci_cqe_irq(host, intmask, &cmd_error, &data_error))
+		return intmask;
+
+	cqhci_irq(host->mmc, intmask, cmd_error, data_error);
+
+	return 0;
+}
+
+#define ITAP_MAX	32
+static int sdhci_am654_platform_execute_tuning(struct sdhci_host *host,
+					       u32 opcode)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_am654_data *sdhci_am654 = sdhci_pltfm_priv(pltfm_host);
+	int cur_val, prev_val = 1, fail_len = 0, pass_window = 0, pass_len;
+	u32 itap;
+
+	/* Enable ITAPDLY */
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL4, ITAPDLYENA_MASK,
+			   1 << ITAPDLYENA_SHIFT);
+
+	for (itap = 0; itap < ITAP_MAX; itap++) {
+		sdhci_am654_write_itapdly(sdhci_am654, itap);
+
+		cur_val = !mmc_send_tuning(host->mmc, opcode, NULL);
+		if (cur_val && !prev_val)
+			pass_window = itap;
+
+		if (!cur_val)
+			fail_len++;
+
+		prev_val = cur_val;
+	}
+	/*
+	 * Having determined the length of the failing window and start of
+	 * the passing window calculate the length of the passing window and
+	 * set the final value halfway through it considering the range as a
+	 * circular buffer
+	 */
+	pass_len = ITAP_MAX - fail_len;
+	itap = (pass_window + (pass_len >> 1)) % ITAP_MAX;
+	sdhci_am654_write_itapdly(sdhci_am654, itap);
+
+	return 0;
+}
+
+static struct sdhci_ops sdhci_am654_ops = {
+	.platform_execute_tuning = sdhci_am654_platform_execute_tuning,
+	.get_max_clock = sdhci_pltfm_clk_get_max_clock,
+	.get_timeout_clock = sdhci_pltfm_clk_get_max_clock,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+	.set_bus_width = sdhci_set_bus_width,
+	.set_power = sdhci_set_power_and_bus_voltage,
+	.set_clock = sdhci_am654_set_clock,
+	.write_b = sdhci_am654_write_b,
+	.irq = sdhci_am654_cqhci_irq,
+	.reset = sdhci_and_cqhci_reset,
+};
+
+static const struct sdhci_pltfm_data sdhci_am654_pdata = {
+	.ops = &sdhci_am654_ops,
+	.quirks = SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+};
+
+static const struct sdhci_am654_driver_data sdhci_am654_sr1_drvdata = {
+	.pdata = &sdhci_am654_pdata,
+	.flags = IOMUX_PRESENT | FREQSEL_2_BIT | STRBSEL_4_BIT | DLL_PRESENT |
+		 DLL_CALIB,
+};
+
+static const struct sdhci_am654_driver_data sdhci_am654_drvdata = {
+	.pdata = &sdhci_am654_pdata,
+	.flags = IOMUX_PRESENT | FREQSEL_2_BIT | STRBSEL_4_BIT | DLL_PRESENT,
+};
+
+static struct sdhci_ops sdhci_j721e_8bit_ops = {
+	.platform_execute_tuning = sdhci_am654_platform_execute_tuning,
+	.get_max_clock = sdhci_pltfm_clk_get_max_clock,
+	.get_timeout_clock = sdhci_pltfm_clk_get_max_clock,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+	.set_bus_width = sdhci_set_bus_width,
+	.set_power = sdhci_set_power_and_bus_voltage,
+	.set_clock = sdhci_am654_set_clock,
+	.write_b = sdhci_am654_write_b,
+	.irq = sdhci_am654_cqhci_irq,
+	.reset = sdhci_and_cqhci_reset,
+};
+
+static const struct sdhci_pltfm_data sdhci_j721e_8bit_pdata = {
+	.ops = &sdhci_j721e_8bit_ops,
+	.quirks = SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+};
+
+static const struct sdhci_am654_driver_data sdhci_j721e_8bit_drvdata = {
+	.pdata = &sdhci_j721e_8bit_pdata,
+	.flags = DLL_PRESENT | DLL_CALIB,
+};
+
+static struct sdhci_ops sdhci_j721e_4bit_ops = {
+	.platform_execute_tuning = sdhci_am654_platform_execute_tuning,
+	.get_max_clock = sdhci_pltfm_clk_get_max_clock,
+	.get_timeout_clock = sdhci_pltfm_clk_get_max_clock,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+	.set_bus_width = sdhci_set_bus_width,
+	.set_power = sdhci_set_power_and_bus_voltage,
+	.set_clock = sdhci_j721e_4bit_set_clock,
+	.write_b = sdhci_am654_write_b,
+	.irq = sdhci_am654_cqhci_irq,
+	.reset = sdhci_am654_reset,
+};
+
+static const struct sdhci_pltfm_data sdhci_j721e_4bit_pdata = {
+	.ops = &sdhci_j721e_4bit_ops,
+	.quirks = SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+};
+
+static const struct sdhci_am654_driver_data sdhci_j721e_4bit_drvdata = {
+	.pdata = &sdhci_j721e_4bit_pdata,
+	.flags = IOMUX_PRESENT,
+};
+
+static const struct soc_device_attribute sdhci_am654_devices[] = {
+	{ .family = "AM65X",
+	  .revision = "SR1.0",
+	  .data = &sdhci_am654_sr1_drvdata
+	},
+	{/* sentinel */}
+};
+
+static void sdhci_am654_dumpregs(struct mmc_host *mmc)
+{
+	sdhci_dumpregs(mmc_priv(mmc));
+}
+
+static const struct cqhci_host_ops sdhci_am654_cqhci_ops = {
+	.enable		= sdhci_cqe_enable,
+	.disable	= sdhci_cqe_disable,
+	.dumpregs	= sdhci_am654_dumpregs,
+};
+
+static int sdhci_am654_cqe_add_host(struct sdhci_host *host)
+{
+	struct cqhci_host *cq_host;
+	int ret;
+
+	cq_host = devm_kzalloc(host->mmc->parent, sizeof(struct cqhci_host),
+			       GFP_KERNEL);
+	if (!cq_host)
+		return -ENOMEM;
+
+	cq_host->mmio = host->ioaddr + SDHCI_AM654_CQE_BASE_ADDR;
+	cq_host->quirks |= CQHCI_QUIRK_SHORT_TXFR_DESC_SZ;
+	cq_host->caps |= CQHCI_TASK_DESC_SZ_128;
+	cq_host->ops = &sdhci_am654_cqhci_ops;
+
+	host->mmc->caps2 |= MMC_CAP2_CQE;
+
+	ret = cqhci_init(cq_host, host->mmc, 1);
+
+	return ret;
+}
+
+static int sdhci_am654_get_otap_delay(struct sdhci_host *host,
+				      struct sdhci_am654_data *sdhci_am654)
+{
+	struct device *dev = mmc_dev(host->mmc);
+	int i;
+	int ret;
+
+	ret = device_property_read_u32(dev, td[MMC_TIMING_LEGACY].otap_binding,
+				 &sdhci_am654->otap_del_sel[MMC_TIMING_LEGACY]);
+	if (ret) {
+		/*
+		 * ti,otap-del-sel-legacy is mandatory, look for old binding
+		 * if not found.
+		 */
+		ret = device_property_read_u32(dev, "ti,otap-del-sel",
+					       &sdhci_am654->otap_del_sel[0]);
+		if (ret) {
+			dev_err(dev, "Couldn't find otap-del-sel\n");
+
+			return ret;
+		}
+
+		dev_info(dev, "Using legacy binding ti,otap-del-sel\n");
+		sdhci_am654->legacy_otapdly = true;
+
+		return 0;
+	}
+
+	for (i = MMC_TIMING_LEGACY; i <= MMC_TIMING_MMC_HS400; i++) {
+
+		ret = device_property_read_u32(dev, td[i].otap_binding,
+					       &sdhci_am654->otap_del_sel[i]);
+		if (ret) {
+			dev_dbg(dev, "Couldn't find %s\n",
+				td[i].otap_binding);
+			/*
+			 * Remove the corresponding capability
+			 * if an otap-del-sel value is not found
+			 */
+			if (i <= MMC_TIMING_MMC_DDR52)
+				host->mmc->caps &= ~td[i].capability;
+			else
+				host->mmc->caps2 &= ~td[i].capability;
+		}
+
+		if (td[i].itap_binding)
+			device_property_read_u32(dev, td[i].itap_binding,
+						 &sdhci_am654->itap_del_sel[i]);
+	}
+
+	return 0;
+}
+
+static int sdhci_am654_init(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_am654_data *sdhci_am654 = sdhci_pltfm_priv(pltfm_host);
+	u32 ctl_cfg_2 = 0;
+	u32 mask;
+	u32 val;
+	int ret;
+
+	/* Reset OTAP to default value */
+	mask = OTAPDLYENA_MASK | OTAPDLYSEL_MASK;
+	regmap_update_bits(sdhci_am654->base, PHY_CTRL4, mask, 0x0);
+
+	if (sdhci_am654->flags & DLL_CALIB) {
+		regmap_read(sdhci_am654->base, PHY_STAT1, &val);
+		if (~val & CALDONE_MASK) {
+			/* Calibrate IO lines */
+			regmap_update_bits(sdhci_am654->base, PHY_CTRL1,
+					   PDB_MASK, PDB_MASK);
+			ret = regmap_read_poll_timeout(sdhci_am654->base,
+						       PHY_STAT1, val,
+						       val & CALDONE_MASK,
+						       1, 20);
+			if (ret)
+				return ret;
+		}
+	}
+
+	/* Enable pins by setting IO mux to 0 */
+	if (sdhci_am654->flags & IOMUX_PRESENT)
+		regmap_update_bits(sdhci_am654->base, PHY_CTRL1,
+				   IOMUX_ENABLE_MASK, 0);
+
+	/* Set slot type based on SD or eMMC */
+	if (host->mmc->caps & MMC_CAP_NONREMOVABLE)
+		ctl_cfg_2 = SLOTTYPE_EMBEDDED;
+
+	regmap_update_bits(sdhci_am654->base, CTL_CFG_2, SLOTTYPE_MASK,
+			   ctl_cfg_2);
+
+	/* Enable tuning for SDR50 */
+	regmap_update_bits(sdhci_am654->base, CTL_CFG_3, TUNINGFORSDR50_MASK,
+			   TUNINGFORSDR50_MASK);
+
+	ret = sdhci_setup_host(host);
+	if (ret)
+		return ret;
+
+	ret = sdhci_am654_cqe_add_host(host);
+	if (ret)
+		goto err_cleanup_host;
+
+	ret = sdhci_am654_get_otap_delay(host, sdhci_am654);
+	if (ret)
+		goto err_cleanup_host;
+
+	ret = __sdhci_add_host(host);
+	if (ret)
+		goto err_cleanup_host;
+
+	return 0;
+
+err_cleanup_host:
+	sdhci_cleanup_host(host);
+	return ret;
+}
+
+static int sdhci_am654_get_of_property(struct platform_device *pdev,
+					struct sdhci_am654_data *sdhci_am654)
+{
+	struct device *dev = &pdev->dev;
+	int drv_strength;
+	int ret;
+
+	if (sdhci_am654->flags & DLL_PRESENT) {
+		ret = device_property_read_u32(dev, "ti,trm-icp",
+					       &sdhci_am654->trm_icp);
+		if (ret)
+			return ret;
+
+		ret = device_property_read_u32(dev, "ti,driver-strength-ohm",
+					       &drv_strength);
+		if (ret)
+			return ret;
+
+		switch (drv_strength) {
+		case 50:
+			sdhci_am654->drv_strength = DRIVER_STRENGTH_50_OHM;
+			break;
+		case 33:
+			sdhci_am654->drv_strength = DRIVER_STRENGTH_33_OHM;
+			break;
+		case 66:
+			sdhci_am654->drv_strength = DRIVER_STRENGTH_66_OHM;
+			break;
+		case 100:
+			sdhci_am654->drv_strength = DRIVER_STRENGTH_100_OHM;
+			break;
+		case 40:
+			sdhci_am654->drv_strength = DRIVER_STRENGTH_40_OHM;
+			break;
+		default:
+			dev_err(dev, "Invalid driver strength\n");
+			return -EINVAL;
+		}
+	}
+
+	device_property_read_u32(dev, "ti,strobe-sel", &sdhci_am654->strb_sel);
+	device_property_read_u32(dev, "ti,clkbuf-sel",
+				 &sdhci_am654->clkbuf_sel);
+
+	if (device_property_read_bool(dev, "ti,fails-without-test-cd"))
+		sdhci_am654->quirks |= SDHCI_AM654_QUIRK_FORCE_CDTEST;
+
+	sdhci_get_of_property(pdev);
+
+	return 0;
+}
+
+static const struct of_device_id sdhci_am654_of_match[] = {
+	{
+		.compatible = "ti,am654-sdhci-5.1",
+		.data = &sdhci_am654_drvdata,
+	},
+	{
+		.compatible = "ti,j721e-sdhci-8bit",
+		.data = &sdhci_j721e_8bit_drvdata,
+	},
+	{
+		.compatible = "ti,j721e-sdhci-4bit",
+		.data = &sdhci_j721e_4bit_drvdata,
+	},
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, sdhci_am654_of_match);
+
+static int sdhci_am654_probe(struct platform_device *pdev)
+{
+	const struct sdhci_am654_driver_data *drvdata;
+	const struct soc_device_attribute *soc;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_am654_data *sdhci_am654;
+	const struct of_device_id *match;
+	struct sdhci_host *host;
+	struct clk *clk_xin;
+	struct device *dev = &pdev->dev;
+	void __iomem *base;
+	int ret;
+
+	match = of_match_node(sdhci_am654_of_match, pdev->dev.of_node);
+	drvdata = match->data;
+
+	/* Update drvdata based on SoC revision */
+	soc = soc_device_match(sdhci_am654_devices);
+	if (soc && soc->data)
+		drvdata = soc->data;
+
+	host = sdhci_pltfm_init(pdev, drvdata->pdata, sizeof(*sdhci_am654));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	sdhci_am654 = sdhci_pltfm_priv(pltfm_host);
+	sdhci_am654->flags = drvdata->flags;
+
+	clk_xin = devm_clk_get(dev, "clk_xin");
+	if (IS_ERR(clk_xin)) {
+		dev_err(dev, "clk_xin clock not found.\n");
+		ret = PTR_ERR(clk_xin);
+		goto err_pltfm_free;
+	}
+
+	pltfm_host->clk = clk_xin;
+
+	/* Clocks are enabled using pm_runtime */
+	pm_runtime_enable(dev);
+	ret = pm_runtime_get_sync(dev);
+	if (ret < 0) {
+		pm_runtime_put_noidle(dev);
+		goto pm_runtime_disable;
+	}
+
+	base = devm_platform_ioremap_resource(pdev, 1);
+	if (IS_ERR(base)) {
+		ret = PTR_ERR(base);
+		goto pm_runtime_put;
+	}
+
+	sdhci_am654->base = devm_regmap_init_mmio(dev, base,
+						  &sdhci_am654_regmap_config);
+	if (IS_ERR(sdhci_am654->base)) {
+		dev_err(dev, "Failed to initialize regmap\n");
+		ret = PTR_ERR(sdhci_am654->base);
+		goto pm_runtime_put;
+	}
+
+	ret = sdhci_am654_get_of_property(pdev, sdhci_am654);
+	if (ret)
+		goto pm_runtime_put;
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret) {
+		dev_err(dev, "parsing dt failed (%d)\n", ret);
+		goto pm_runtime_put;
+	}
+
+	host->mmc_host_ops.execute_tuning = sdhci_am654_execute_tuning;
+
+	ret = sdhci_am654_init(host);
+	if (ret)
+		goto pm_runtime_put;
+
+	return 0;
+
+pm_runtime_put:
+	pm_runtime_put_sync(dev);
+pm_runtime_disable:
+	pm_runtime_disable(dev);
+err_pltfm_free:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static int sdhci_am654_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	int ret;
+
+	sdhci_remove_host(host, true);
+	ret = pm_runtime_put_sync(&pdev->dev);
+	if (ret < 0)
+		return ret;
+
+	pm_runtime_disable(&pdev->dev);
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+
+static struct platform_driver sdhci_am654_driver = {
+	.driver = {
+		.name = "sdhci-am654",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = sdhci_am654_of_match,
+	},
+	.probe = sdhci_am654_probe,
+	.remove = sdhci_am654_remove,
+};
+
+module_platform_driver(sdhci_am654_driver);
+
+MODULE_DESCRIPTION("Driver for SDHCI Controller on TI's AM654 devices");
+MODULE_AUTHOR("Faiz Abbas <faiz_abbas@ti.com>");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mmc/host/sdhci_f_sdh30.c b/drivers/mmc/host/sdhci_f_sdh30.c
new file mode 100644
index 000000000..7ede74bf3
--- /dev/null
+++ b/drivers/mmc/host/sdhci_f_sdh30.c
@@ -0,0 +1,235 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * linux/drivers/mmc/host/sdhci_f_sdh30.c
+ *
+ * Copyright (C) 2013 - 2015 Fujitsu Semiconductor, Ltd
+ *              Vincent Yang <vincent.yang@tw.fujitsu.com>
+ * Copyright (C) 2015 Linaro Ltd  Andy Green <andy.green@linaro.org>
+ */
+
+#include <linux/acpi.h>
+#include <linux/err.h>
+#include <linux/delay.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/property.h>
+#include <linux/clk.h>
+
+#include "sdhci-pltfm.h"
+#include "sdhci_f_sdh30.h"
+
+struct f_sdhost_priv {
+	struct clk *clk_iface;
+	struct clk *clk;
+	u32 vendor_hs200;
+	struct device *dev;
+	bool enable_cmd_dat_delay;
+};
+
+static void *sdhci_f_sdhost_priv(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	return sdhci_pltfm_priv(pltfm_host);
+}
+
+static void sdhci_f_sdh30_soft_voltage_switch(struct sdhci_host *host)
+{
+	struct f_sdhost_priv *priv = sdhci_f_sdhost_priv(host);
+	u32 ctrl = 0;
+
+	usleep_range(2500, 3000);
+	ctrl = sdhci_readl(host, F_SDH30_IO_CONTROL2);
+	ctrl |= F_SDH30_CRES_O_DN;
+	sdhci_writel(host, ctrl, F_SDH30_IO_CONTROL2);
+	ctrl |= F_SDH30_MSEL_O_1_8;
+	sdhci_writel(host, ctrl, F_SDH30_IO_CONTROL2);
+
+	ctrl &= ~F_SDH30_CRES_O_DN;
+	sdhci_writel(host, ctrl, F_SDH30_IO_CONTROL2);
+	usleep_range(2500, 3000);
+
+	if (priv->vendor_hs200) {
+		dev_info(priv->dev, "%s: setting hs200\n", __func__);
+		ctrl = sdhci_readl(host, F_SDH30_ESD_CONTROL);
+		ctrl |= priv->vendor_hs200;
+		sdhci_writel(host, ctrl, F_SDH30_ESD_CONTROL);
+	}
+
+	ctrl = sdhci_readl(host, F_SDH30_TUNING_SETTING);
+	ctrl |= F_SDH30_CMD_CHK_DIS;
+	sdhci_writel(host, ctrl, F_SDH30_TUNING_SETTING);
+}
+
+static unsigned int sdhci_f_sdh30_get_min_clock(struct sdhci_host *host)
+{
+	return F_SDH30_MIN_CLOCK;
+}
+
+static void sdhci_f_sdh30_reset(struct sdhci_host *host, u8 mask)
+{
+	struct f_sdhost_priv *priv = sdhci_f_sdhost_priv(host);
+	u32 ctl;
+
+	if (sdhci_readw(host, SDHCI_CLOCK_CONTROL) == 0)
+		sdhci_writew(host, 0xBC01, SDHCI_CLOCK_CONTROL);
+
+	sdhci_reset(host, mask);
+
+	if (priv->enable_cmd_dat_delay) {
+		ctl = sdhci_readl(host, F_SDH30_ESD_CONTROL);
+		ctl |= F_SDH30_CMD_DAT_DELAY;
+		sdhci_writel(host, ctl, F_SDH30_ESD_CONTROL);
+	}
+}
+
+static const struct sdhci_ops sdhci_f_sdh30_ops = {
+	.voltage_switch = sdhci_f_sdh30_soft_voltage_switch,
+	.get_min_clock = sdhci_f_sdh30_get_min_clock,
+	.reset = sdhci_f_sdh30_reset,
+	.set_clock = sdhci_set_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static const struct sdhci_pltfm_data sdhci_f_sdh30_pltfm_data = {
+	.ops = &sdhci_f_sdh30_ops,
+	.quirks = SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC
+		| SDHCI_QUIRK_INVERTED_WRITE_PROTECT,
+	.quirks2 = SDHCI_QUIRK2_SUPPORT_SINGLE
+		|  SDHCI_QUIRK2_TUNING_WORK_AROUND,
+};
+
+static int sdhci_f_sdh30_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct device *dev = &pdev->dev;
+	int ctrl = 0, ret = 0;
+	struct f_sdhost_priv *priv;
+	struct sdhci_pltfm_host *pltfm_host;
+	u32 reg = 0;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_f_sdh30_pltfm_data,
+				sizeof(struct f_sdhost_priv));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	priv = sdhci_pltfm_priv(pltfm_host);
+	priv->dev = dev;
+
+	priv->enable_cmd_dat_delay = device_property_read_bool(dev,
+						"fujitsu,cmd-dat-delay-select");
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto err;
+
+	if (dev_of_node(dev)) {
+		sdhci_get_of_property(pdev);
+
+		priv->clk_iface = devm_clk_get(&pdev->dev, "iface");
+		if (IS_ERR(priv->clk_iface)) {
+			ret = PTR_ERR(priv->clk_iface);
+			goto err;
+		}
+
+		ret = clk_prepare_enable(priv->clk_iface);
+		if (ret)
+			goto err;
+
+		priv->clk = devm_clk_get(&pdev->dev, "core");
+		if (IS_ERR(priv->clk)) {
+			ret = PTR_ERR(priv->clk);
+			goto err_clk;
+		}
+
+		ret = clk_prepare_enable(priv->clk);
+		if (ret)
+			goto err_clk;
+	}
+
+	/* init vendor specific regs */
+	ctrl = sdhci_readw(host, F_SDH30_AHB_CONFIG);
+	ctrl |= F_SDH30_SIN | F_SDH30_AHB_INCR_16 | F_SDH30_AHB_INCR_8 |
+		F_SDH30_AHB_INCR_4;
+	ctrl &= ~(F_SDH30_AHB_BIGED | F_SDH30_BUSLOCK_EN);
+	sdhci_writew(host, ctrl, F_SDH30_AHB_CONFIG);
+
+	reg = sdhci_readl(host, F_SDH30_ESD_CONTROL);
+	sdhci_writel(host, reg & ~F_SDH30_EMMC_RST, F_SDH30_ESD_CONTROL);
+	msleep(20);
+	sdhci_writel(host, reg | F_SDH30_EMMC_RST, F_SDH30_ESD_CONTROL);
+
+	reg = sdhci_readl(host, SDHCI_CAPABILITIES);
+	if (reg & SDHCI_CAN_DO_8BIT)
+		priv->vendor_hs200 = F_SDH30_EMMC_HS200;
+
+	if (!(reg & SDHCI_TIMEOUT_CLK_MASK))
+		host->quirks |= SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK;
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err_add_host;
+
+	return 0;
+
+err_add_host:
+	clk_disable_unprepare(priv->clk);
+err_clk:
+	clk_disable_unprepare(priv->clk_iface);
+err:
+	sdhci_pltfm_free(pdev);
+
+	return ret;
+}
+
+static int sdhci_f_sdh30_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct f_sdhost_priv *priv = sdhci_f_sdhost_priv(host);
+	struct clk *clk_iface = priv->clk_iface;
+	struct clk *clk = priv->clk;
+
+	sdhci_pltfm_unregister(pdev);
+
+	clk_disable_unprepare(clk_iface);
+	clk_disable_unprepare(clk);
+
+	return 0;
+}
+
+#ifdef CONFIG_OF
+static const struct of_device_id f_sdh30_dt_ids[] = {
+	{ .compatible = "fujitsu,mb86s70-sdhci-3.0" },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, f_sdh30_dt_ids);
+#endif
+
+#ifdef CONFIG_ACPI
+static const struct acpi_device_id f_sdh30_acpi_ids[] = {
+	{ "SCX0002" },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(acpi, f_sdh30_acpi_ids);
+#endif
+
+static struct platform_driver sdhci_f_sdh30_driver = {
+	.driver = {
+		.name = "f_sdh30",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(f_sdh30_dt_ids),
+		.acpi_match_table = ACPI_PTR(f_sdh30_acpi_ids),
+		.pm	= &sdhci_pltfm_pmops,
+	},
+	.probe	= sdhci_f_sdh30_probe,
+	.remove	= sdhci_f_sdh30_remove,
+};
+
+module_platform_driver(sdhci_f_sdh30_driver);
+
+MODULE_DESCRIPTION("F_SDH30 SD Card Controller driver");
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("FUJITSU SEMICONDUCTOR LTD.");
+MODULE_ALIAS("platform:f_sdh30");
diff --git a/drivers/mmc/host/sdhci_f_sdh30.h b/drivers/mmc/host/sdhci_f_sdh30.h
new file mode 100644
index 000000000..fc1ad28f7
--- /dev/null
+++ b/drivers/mmc/host/sdhci_f_sdh30.h
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2013 - 2015 Fujitsu Semiconductor, Ltd
+ *              Vincent Yang <vincent.yang@tw.fujitsu.com>
+ * Copyright (C) 2015 Linaro Ltd  Andy Green <andy.green@linaro.org>
+ * Copyright (C) 2019 Socionext Inc.
+ *
+ */
+
+/* F_SDH30 extended Controller registers */
+#define F_SDH30_AHB_CONFIG      0x100
+#define  F_SDH30_AHB_BIGED      BIT(6)
+#define  F_SDH30_BUSLOCK_DMA    BIT(5)
+#define  F_SDH30_BUSLOCK_EN     BIT(4)
+#define  F_SDH30_SIN            BIT(3)
+#define  F_SDH30_AHB_INCR_16    BIT(2)
+#define  F_SDH30_AHB_INCR_8     BIT(1)
+#define  F_SDH30_AHB_INCR_4     BIT(0)
+
+#define F_SDH30_TUNING_SETTING  0x108
+#define  F_SDH30_CMD_CHK_DIS    BIT(16)
+
+#define F_SDH30_IO_CONTROL2     0x114
+#define  F_SDH30_CRES_O_DN      BIT(19)
+#define  F_SDH30_MSEL_O_1_8     BIT(18)
+
+#define F_SDH30_ESD_CONTROL     0x124
+#define	 F_SDH30_EMMC_RST		BIT(1)
+#define  F_SDH30_CMD_DAT_DELAY	BIT(9)
+#define	 F_SDH30_EMMC_HS200		BIT(24)
+
+#define F_SDH30_MIN_CLOCK		400000
diff --git a/drivers/mmc/host/sdricoh_cs.c b/drivers/mmc/host/sdricoh_cs.c
new file mode 100644
index 000000000..76a8cd3a1
--- /dev/null
+++ b/drivers/mmc/host/sdricoh_cs.c
@@ -0,0 +1,536 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ *  sdricoh_cs.c - driver for Ricoh Secure Digital Card Readers that can be
+ *     found on some Ricoh RL5c476 II cardbus bridge
+ *
+ *  Copyright (C) 2006 - 2008 Sascha Sommer <saschasommer@freenet.de>
+ */
+
+/*
+#define DEBUG
+#define VERBOSE_DEBUG
+*/
+#include <linux/delay.h>
+#include <linux/highmem.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/ioport.h>
+#include <linux/iopoll.h>
+#include <linux/scatterlist.h>
+
+#include <pcmcia/cistpl.h>
+#include <pcmcia/ds.h>
+#include <linux/io.h>
+
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+
+#define DRIVER_NAME "sdricoh_cs"
+
+static unsigned int switchlocked;
+
+/* i/o region */
+#define SDRICOH_PCI_REGION 0
+#define SDRICOH_PCI_REGION_SIZE 0x1000
+
+/* registers */
+#define R104_VERSION     0x104
+#define R200_CMD         0x200
+#define R204_CMD_ARG     0x204
+#define R208_DATAIO      0x208
+#define R20C_RESP        0x20c
+#define R21C_STATUS      0x21c
+#define R2E0_INIT        0x2e0
+#define R2E4_STATUS_RESP 0x2e4
+#define R2F0_RESET       0x2f0
+#define R224_MODE        0x224
+#define R226_BLOCKSIZE   0x226
+#define R228_POWER       0x228
+#define R230_DATA        0x230
+
+/* flags for the R21C_STATUS register */
+#define STATUS_CMD_FINISHED      0x00000001
+#define STATUS_TRANSFER_FINISHED 0x00000004
+#define STATUS_CARD_INSERTED     0x00000020
+#define STATUS_CARD_LOCKED       0x00000080
+#define STATUS_CMD_TIMEOUT       0x00400000
+#define STATUS_READY_TO_READ     0x01000000
+#define STATUS_READY_TO_WRITE    0x02000000
+#define STATUS_BUSY              0x40000000
+
+/* timeouts */
+#define SDRICOH_CMD_TIMEOUT_US	1000000
+#define SDRICOH_DATA_TIMEOUT_US	1000000
+
+/* list of supported pcmcia devices */
+static const struct pcmcia_device_id pcmcia_ids[] = {
+	/* vendor and device strings followed by their crc32 hashes */
+	PCMCIA_DEVICE_PROD_ID12("RICOH", "Bay1Controller", 0xd9f522ed,
+				0xc3901202),
+	PCMCIA_DEVICE_PROD_ID12("RICOH", "Bay Controller", 0xd9f522ed,
+				0xace80909),
+	PCMCIA_DEVICE_NULL,
+};
+
+MODULE_DEVICE_TABLE(pcmcia, pcmcia_ids);
+
+/* mmc privdata */
+struct sdricoh_host {
+	struct device *dev;
+	struct mmc_host *mmc;	/* MMC structure */
+	unsigned char __iomem *iobase;
+	struct pci_dev *pci_dev;
+	int app_cmd;
+};
+
+/***************** register i/o helper functions *****************************/
+
+static inline unsigned int sdricoh_readl(struct sdricoh_host *host,
+					 unsigned int reg)
+{
+	unsigned int value = readl(host->iobase + reg);
+	dev_vdbg(host->dev, "rl %x 0x%x\n", reg, value);
+	return value;
+}
+
+static inline void sdricoh_writel(struct sdricoh_host *host, unsigned int reg,
+				  unsigned int value)
+{
+	writel(value, host->iobase + reg);
+	dev_vdbg(host->dev, "wl %x 0x%x\n", reg, value);
+
+}
+
+static inline unsigned int sdricoh_readw(struct sdricoh_host *host,
+					 unsigned int reg)
+{
+	unsigned int value = readw(host->iobase + reg);
+	dev_vdbg(host->dev, "rb %x 0x%x\n", reg, value);
+	return value;
+}
+
+static inline void sdricoh_writew(struct sdricoh_host *host, unsigned int reg,
+					 unsigned short value)
+{
+	writew(value, host->iobase + reg);
+	dev_vdbg(host->dev, "ww %x 0x%x\n", reg, value);
+}
+
+static inline unsigned int sdricoh_readb(struct sdricoh_host *host,
+					 unsigned int reg)
+{
+	unsigned int value = readb(host->iobase + reg);
+	dev_vdbg(host->dev, "rb %x 0x%x\n", reg, value);
+	return value;
+}
+
+static bool sdricoh_status_ok(struct sdricoh_host *host, unsigned int status,
+			      unsigned int wanted)
+{
+	sdricoh_writel(host, R2E4_STATUS_RESP, status);
+	return status & wanted;
+}
+
+static int sdricoh_query_status(struct sdricoh_host *host, unsigned int wanted)
+{
+	int ret;
+	unsigned int status = 0;
+	struct device *dev = host->dev;
+
+	ret = read_poll_timeout(sdricoh_readl, status,
+				sdricoh_status_ok(host, status, wanted),
+				32, SDRICOH_DATA_TIMEOUT_US, false,
+				host, R21C_STATUS);
+	if (ret) {
+		dev_err(dev, "query_status: timeout waiting for %x\n", wanted);
+		return -ETIMEDOUT;
+	}
+
+	/* do not do this check in the loop as some commands fail otherwise */
+	if (status & 0x7F0000) {
+		dev_err(dev, "waiting for status bit %x failed\n", wanted);
+		return -EINVAL;
+	}
+	return 0;
+
+}
+
+static int sdricoh_mmc_cmd(struct sdricoh_host *host, struct mmc_command *cmd)
+{
+	unsigned int status, timeout_us;
+	int ret;
+	unsigned char opcode = cmd->opcode;
+
+	/* reset status reg? */
+	sdricoh_writel(host, R21C_STATUS, 0x18);
+
+	/* MMC_APP_CMDs need some special handling */
+	if (host->app_cmd) {
+		opcode |= 64;
+		host->app_cmd = 0;
+	} else if (opcode == MMC_APP_CMD)
+		host->app_cmd = 1;
+
+	/* fill parameters */
+	sdricoh_writel(host, R204_CMD_ARG, cmd->arg);
+	sdricoh_writel(host, R200_CMD, (0x10000 << 8) | opcode);
+
+	/* wait for command completion */
+	if (!opcode)
+		return 0;
+
+	timeout_us = cmd->busy_timeout ? cmd->busy_timeout * 1000 :
+		SDRICOH_CMD_TIMEOUT_US;
+
+	ret = read_poll_timeout(sdricoh_readl, status,
+			sdricoh_status_ok(host, status, STATUS_CMD_FINISHED),
+			32, timeout_us, false,
+			host, R21C_STATUS);
+
+	/*
+	 * Don't check for timeout status in the loop, as it's not always reset
+	 * correctly.
+	 */
+	if (ret || status & STATUS_CMD_TIMEOUT)
+		return -ETIMEDOUT;
+
+	return 0;
+}
+
+static int sdricoh_reset(struct sdricoh_host *host)
+{
+	dev_dbg(host->dev, "reset\n");
+	sdricoh_writel(host, R2F0_RESET, 0x10001);
+	sdricoh_writel(host, R2E0_INIT, 0x10000);
+	if (sdricoh_readl(host, R2E0_INIT) != 0x10000)
+		return -EIO;
+	sdricoh_writel(host, R2E0_INIT, 0x10007);
+
+	sdricoh_writel(host, R224_MODE, 0x2000000);
+	sdricoh_writel(host, R228_POWER, 0xe0);
+
+
+	/* status register ? */
+	sdricoh_writel(host, R21C_STATUS, 0x18);
+
+	return 0;
+}
+
+static int sdricoh_blockio(struct sdricoh_host *host, int read,
+				u8 *buf, int len)
+{
+	int size;
+	u32 data = 0;
+	/* wait until the data is available */
+	if (read) {
+		if (sdricoh_query_status(host, STATUS_READY_TO_READ))
+			return -ETIMEDOUT;
+		sdricoh_writel(host, R21C_STATUS, 0x18);
+		/* read data */
+		while (len) {
+			data = sdricoh_readl(host, R230_DATA);
+			size = min(len, 4);
+			len -= size;
+			while (size) {
+				*buf = data & 0xFF;
+				buf++;
+				data >>= 8;
+				size--;
+			}
+		}
+	} else {
+		if (sdricoh_query_status(host, STATUS_READY_TO_WRITE))
+			return -ETIMEDOUT;
+		sdricoh_writel(host, R21C_STATUS, 0x18);
+		/* write data */
+		while (len) {
+			size = min(len, 4);
+			len -= size;
+			while (size) {
+				data >>= 8;
+				data |= (u32)*buf << 24;
+				buf++;
+				size--;
+			}
+			sdricoh_writel(host, R230_DATA, data);
+		}
+	}
+
+	return 0;
+}
+
+static void sdricoh_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct sdricoh_host *host = mmc_priv(mmc);
+	struct mmc_command *cmd = mrq->cmd;
+	struct mmc_data *data = cmd->data;
+	struct device *dev = host->dev;
+	int i;
+
+	dev_dbg(dev, "=============================\n");
+	dev_dbg(dev, "sdricoh_request opcode=%i\n", cmd->opcode);
+
+	sdricoh_writel(host, R21C_STATUS, 0x18);
+
+	/* read/write commands seem to require this */
+	if (data) {
+		sdricoh_writew(host, R226_BLOCKSIZE, data->blksz);
+		sdricoh_writel(host, R208_DATAIO, 0);
+	}
+
+	cmd->error = sdricoh_mmc_cmd(host, cmd);
+
+	/* read response buffer */
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136) {
+			/* CRC is stripped so we need to do some shifting. */
+			for (i = 0; i < 4; i++) {
+				cmd->resp[i] =
+				    sdricoh_readl(host,
+						  R20C_RESP + (3 - i) * 4) << 8;
+				if (i != 3)
+					cmd->resp[i] |=
+					    sdricoh_readb(host, R20C_RESP +
+							  (3 - i) * 4 - 1);
+			}
+		} else
+			cmd->resp[0] = sdricoh_readl(host, R20C_RESP);
+	}
+
+	/* transfer data */
+	if (data && cmd->error == 0) {
+		dev_dbg(dev, "transfer: blksz %i blocks %i sg_len %i "
+			"sg length %i\n", data->blksz, data->blocks,
+			data->sg_len, data->sg->length);
+
+		/* enter data reading mode */
+		sdricoh_writel(host, R21C_STATUS, 0x837f031e);
+		for (i = 0; i < data->blocks; i++) {
+			size_t len = data->blksz;
+			u8 *buf;
+			struct page *page;
+			int result;
+			page = sg_page(data->sg);
+
+			buf = kmap(page) + data->sg->offset + (len * i);
+			result =
+				sdricoh_blockio(host,
+					data->flags & MMC_DATA_READ, buf, len);
+			kunmap(page);
+			flush_dcache_page(page);
+			if (result) {
+				dev_err(dev, "sdricoh_request: cmd %i "
+					"block transfer failed\n", cmd->opcode);
+				cmd->error = result;
+				break;
+			} else
+				data->bytes_xfered += len;
+		}
+
+		sdricoh_writel(host, R208_DATAIO, 1);
+
+		if (sdricoh_query_status(host, STATUS_TRANSFER_FINISHED)) {
+			dev_err(dev, "sdricoh_request: transfer end error\n");
+			cmd->error = -EINVAL;
+		}
+	}
+	/* FIXME check busy flag */
+
+	mmc_request_done(mmc, mrq);
+	dev_dbg(dev, "=============================\n");
+}
+
+static void sdricoh_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sdricoh_host *host = mmc_priv(mmc);
+	dev_dbg(host->dev, "set_ios\n");
+
+	if (ios->power_mode == MMC_POWER_ON) {
+		sdricoh_writel(host, R228_POWER, 0xc0e0);
+
+		if (ios->bus_width == MMC_BUS_WIDTH_4) {
+			sdricoh_writel(host, R224_MODE, 0x2000300);
+			sdricoh_writel(host, R228_POWER, 0x40e0);
+		} else {
+			sdricoh_writel(host, R224_MODE, 0x2000340);
+		}
+
+	} else if (ios->power_mode == MMC_POWER_UP) {
+		sdricoh_writel(host, R224_MODE, 0x2000320);
+		sdricoh_writel(host, R228_POWER, 0xe0);
+	}
+}
+
+static int sdricoh_get_ro(struct mmc_host *mmc)
+{
+	struct sdricoh_host *host = mmc_priv(mmc);
+	unsigned int status;
+
+	status = sdricoh_readl(host, R21C_STATUS);
+	sdricoh_writel(host, R2E4_STATUS_RESP, status);
+
+	/* some notebooks seem to have the locked flag switched */
+	if (switchlocked)
+		return !(status & STATUS_CARD_LOCKED);
+
+	return (status & STATUS_CARD_LOCKED);
+}
+
+static const struct mmc_host_ops sdricoh_ops = {
+	.request = sdricoh_request,
+	.set_ios = sdricoh_set_ios,
+	.get_ro = sdricoh_get_ro,
+};
+
+/* initialize the control and register it to the mmc framework */
+static int sdricoh_init_mmc(struct pci_dev *pci_dev,
+			    struct pcmcia_device *pcmcia_dev)
+{
+	int result;
+	void __iomem *iobase;
+	struct mmc_host *mmc;
+	struct sdricoh_host *host;
+	struct device *dev = &pcmcia_dev->dev;
+	/* map iomem */
+	if (pci_resource_len(pci_dev, SDRICOH_PCI_REGION) !=
+	    SDRICOH_PCI_REGION_SIZE) {
+		dev_dbg(dev, "unexpected pci resource len\n");
+		return -ENODEV;
+	}
+	iobase =
+	    pci_iomap(pci_dev, SDRICOH_PCI_REGION, SDRICOH_PCI_REGION_SIZE);
+	if (!iobase) {
+		dev_err(dev, "unable to map iobase\n");
+		return -ENODEV;
+	}
+	/* check version? */
+	if (readl(iobase + R104_VERSION) != 0x4000) {
+		dev_dbg(dev, "no supported mmc controller found\n");
+		result = -ENODEV;
+		goto unmap_io;
+	}
+	/* allocate privdata */
+	mmc = pcmcia_dev->priv =
+	    mmc_alloc_host(sizeof(struct sdricoh_host), &pcmcia_dev->dev);
+	if (!mmc) {
+		dev_err(dev, "mmc_alloc_host failed\n");
+		result = -ENOMEM;
+		goto unmap_io;
+	}
+	host = mmc_priv(mmc);
+
+	host->iobase = iobase;
+	host->dev = dev;
+	host->pci_dev = pci_dev;
+
+	mmc->ops = &sdricoh_ops;
+
+	/* FIXME: frequency and voltage handling is done by the controller
+	 */
+	mmc->f_min = 450000;
+	mmc->f_max = 24000000;
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+	mmc->caps |= MMC_CAP_4_BIT_DATA;
+
+	mmc->max_seg_size = 1024 * 512;
+	mmc->max_blk_size = 512;
+
+	/* reset the controller */
+	if (sdricoh_reset(host)) {
+		dev_dbg(dev, "could not reset\n");
+		result = -EIO;
+		goto free_host;
+	}
+
+	result = mmc_add_host(mmc);
+
+	if (!result) {
+		dev_dbg(dev, "mmc host registered\n");
+		return 0;
+	}
+free_host:
+	mmc_free_host(mmc);
+unmap_io:
+	pci_iounmap(pci_dev, iobase);
+	return result;
+}
+
+/* search for supported mmc controllers */
+static int sdricoh_pcmcia_probe(struct pcmcia_device *pcmcia_dev)
+{
+	struct pci_dev *pci_dev = NULL;
+
+	dev_info(&pcmcia_dev->dev, "Searching MMC controller for pcmcia device"
+		" %s %s ...\n", pcmcia_dev->prod_id[0], pcmcia_dev->prod_id[1]);
+
+	/* search pci cardbus bridge that contains the mmc controller */
+	/* the io region is already claimed by yenta_socket... */
+	while ((pci_dev =
+		pci_get_device(PCI_VENDOR_ID_RICOH, PCI_DEVICE_ID_RICOH_RL5C476,
+			       pci_dev))) {
+		/* try to init the device */
+		if (!sdricoh_init_mmc(pci_dev, pcmcia_dev)) {
+			dev_info(&pcmcia_dev->dev, "MMC controller found\n");
+			return 0;
+		}
+
+	}
+	dev_err(&pcmcia_dev->dev, "No MMC controller was found.\n");
+	return -ENODEV;
+}
+
+static void sdricoh_pcmcia_detach(struct pcmcia_device *link)
+{
+	struct mmc_host *mmc = link->priv;
+
+	dev_dbg(&link->dev, "detach\n");
+
+	/* remove mmc host */
+	if (mmc) {
+		struct sdricoh_host *host = mmc_priv(mmc);
+		mmc_remove_host(mmc);
+		pci_iounmap(host->pci_dev, host->iobase);
+		pci_dev_put(host->pci_dev);
+		mmc_free_host(mmc);
+	}
+	pcmcia_disable_device(link);
+
+}
+
+#ifdef CONFIG_PM
+static int sdricoh_pcmcia_suspend(struct pcmcia_device *link)
+{
+	dev_dbg(&link->dev, "suspend\n");
+	return 0;
+}
+
+static int sdricoh_pcmcia_resume(struct pcmcia_device *link)
+{
+	struct mmc_host *mmc = link->priv;
+	dev_dbg(&link->dev, "resume\n");
+	sdricoh_reset(mmc_priv(mmc));
+	return 0;
+}
+#else
+#define sdricoh_pcmcia_suspend NULL
+#define sdricoh_pcmcia_resume NULL
+#endif
+
+static struct pcmcia_driver sdricoh_driver = {
+	.name = DRIVER_NAME,
+	.probe = sdricoh_pcmcia_probe,
+	.remove = sdricoh_pcmcia_detach,
+	.id_table = pcmcia_ids,
+	.suspend = sdricoh_pcmcia_suspend,
+	.resume = sdricoh_pcmcia_resume,
+};
+module_pcmcia_driver(sdricoh_driver);
+
+module_param(switchlocked, uint, 0444);
+
+MODULE_AUTHOR("Sascha Sommer <saschasommer@freenet.de>");
+MODULE_DESCRIPTION("Ricoh PCMCIA Secure Digital Interface driver");
+MODULE_LICENSE("GPL");
+
+MODULE_PARM_DESC(switchlocked, "Switch the cards locked status."
+		"Use this when unlocked cards are shown readonly (default 0)");
diff --git a/drivers/mmc/host/sh_mmcif.c b/drivers/mmc/host/sh_mmcif.c
new file mode 100644
index 000000000..5dec9e239
--- /dev/null
+++ b/drivers/mmc/host/sh_mmcif.c
@@ -0,0 +1,1576 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * MMCIF eMMC driver.
+ *
+ * Copyright (C) 2010 Renesas Solutions Corp.
+ * Yusuke Goda <yusuke.goda.sx@renesas.com>
+ */
+
+/*
+ * The MMCIF driver is now processing MMC requests asynchronously, according
+ * to the Linux MMC API requirement.
+ *
+ * The MMCIF driver processes MMC requests in up to 3 stages: command, optional
+ * data, and optional stop. To achieve asynchronous processing each of these
+ * stages is split into two halves: a top and a bottom half. The top half
+ * initialises the hardware, installs a timeout handler to handle completion
+ * timeouts, and returns. In case of the command stage this immediately returns
+ * control to the caller, leaving all further processing to run asynchronously.
+ * All further request processing is performed by the bottom halves.
+ *
+ * The bottom half further consists of a "hard" IRQ handler, an IRQ handler
+ * thread, a DMA completion callback, if DMA is used, a timeout work, and
+ * request- and stage-specific handler methods.
+ *
+ * Each bottom half run begins with either a hardware interrupt, a DMA callback
+ * invocation, or a timeout work run. In case of an error or a successful
+ * processing completion, the MMC core is informed and the request processing is
+ * finished. In case processing has to continue, i.e., if data has to be read
+ * from or written to the card, or if a stop command has to be sent, the next
+ * top half is called, which performs the necessary hardware handling and
+ * reschedules the timeout work. This returns the driver state machine into the
+ * bottom half waiting state.
+ */
+
+#include <linux/bitops.h>
+#include <linux/clk.h>
+#include <linux/completion.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/dmaengine.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/core.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sdio.h>
+#include <linux/mmc/sh_mmcif.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/mod_devicetable.h>
+#include <linux/mutex.h>
+#include <linux/of_device.h>
+#include <linux/pagemap.h>
+#include <linux/platform_device.h>
+#include <linux/pm_qos.h>
+#include <linux/pm_runtime.h>
+#include <linux/sh_dma.h>
+#include <linux/spinlock.h>
+#include <linux/module.h>
+
+#define DRIVER_NAME	"sh_mmcif"
+
+/* CE_CMD_SET */
+#define CMD_MASK		0x3f000000
+#define CMD_SET_RTYP_NO		((0 << 23) | (0 << 22))
+#define CMD_SET_RTYP_6B		((0 << 23) | (1 << 22)) /* R1/R1b/R3/R4/R5 */
+#define CMD_SET_RTYP_17B	((1 << 23) | (0 << 22)) /* R2 */
+#define CMD_SET_RBSY		(1 << 21) /* R1b */
+#define CMD_SET_CCSEN		(1 << 20)
+#define CMD_SET_WDAT		(1 << 19) /* 1: on data, 0: no data */
+#define CMD_SET_DWEN		(1 << 18) /* 1: write, 0: read */
+#define CMD_SET_CMLTE		(1 << 17) /* 1: multi block trans, 0: single */
+#define CMD_SET_CMD12EN		(1 << 16) /* 1: CMD12 auto issue */
+#define CMD_SET_RIDXC_INDEX	((0 << 15) | (0 << 14)) /* index check */
+#define CMD_SET_RIDXC_BITS	((0 << 15) | (1 << 14)) /* check bits check */
+#define CMD_SET_RIDXC_NO	((1 << 15) | (0 << 14)) /* no check */
+#define CMD_SET_CRC7C		((0 << 13) | (0 << 12)) /* CRC7 check*/
+#define CMD_SET_CRC7C_BITS	((0 << 13) | (1 << 12)) /* check bits check*/
+#define CMD_SET_CRC7C_INTERNAL	((1 << 13) | (0 << 12)) /* internal CRC7 check*/
+#define CMD_SET_CRC16C		(1 << 10) /* 0: CRC16 check*/
+#define CMD_SET_CRCSTE		(1 << 8) /* 1: not receive CRC status */
+#define CMD_SET_TBIT		(1 << 7) /* 1: tran mission bit "Low" */
+#define CMD_SET_OPDM		(1 << 6) /* 1: open/drain */
+#define CMD_SET_CCSH		(1 << 5)
+#define CMD_SET_DARS		(1 << 2) /* Dual Data Rate */
+#define CMD_SET_DATW_1		((0 << 1) | (0 << 0)) /* 1bit */
+#define CMD_SET_DATW_4		((0 << 1) | (1 << 0)) /* 4bit */
+#define CMD_SET_DATW_8		((1 << 1) | (0 << 0)) /* 8bit */
+
+/* CE_CMD_CTRL */
+#define CMD_CTRL_BREAK		(1 << 0)
+
+/* CE_BLOCK_SET */
+#define BLOCK_SIZE_MASK		0x0000ffff
+
+/* CE_INT */
+#define INT_CCSDE		(1 << 29)
+#define INT_CMD12DRE		(1 << 26)
+#define INT_CMD12RBE		(1 << 25)
+#define INT_CMD12CRE		(1 << 24)
+#define INT_DTRANE		(1 << 23)
+#define INT_BUFRE		(1 << 22)
+#define INT_BUFWEN		(1 << 21)
+#define INT_BUFREN		(1 << 20)
+#define INT_CCSRCV		(1 << 19)
+#define INT_RBSYE		(1 << 17)
+#define INT_CRSPE		(1 << 16)
+#define INT_CMDVIO		(1 << 15)
+#define INT_BUFVIO		(1 << 14)
+#define INT_WDATERR		(1 << 11)
+#define INT_RDATERR		(1 << 10)
+#define INT_RIDXERR		(1 << 9)
+#define INT_RSPERR		(1 << 8)
+#define INT_CCSTO		(1 << 5)
+#define INT_CRCSTO		(1 << 4)
+#define INT_WDATTO		(1 << 3)
+#define INT_RDATTO		(1 << 2)
+#define INT_RBSYTO		(1 << 1)
+#define INT_RSPTO		(1 << 0)
+#define INT_ERR_STS		(INT_CMDVIO | INT_BUFVIO | INT_WDATERR |  \
+				 INT_RDATERR | INT_RIDXERR | INT_RSPERR | \
+				 INT_CCSTO | INT_CRCSTO | INT_WDATTO |	  \
+				 INT_RDATTO | INT_RBSYTO | INT_RSPTO)
+
+#define INT_ALL			(INT_RBSYE | INT_CRSPE | INT_BUFREN |	 \
+				 INT_BUFWEN | INT_CMD12DRE | INT_BUFRE | \
+				 INT_DTRANE | INT_CMD12RBE | INT_CMD12CRE)
+
+#define INT_CCS			(INT_CCSTO | INT_CCSRCV | INT_CCSDE)
+
+/* CE_INT_MASK */
+#define MASK_ALL		0x00000000
+#define MASK_MCCSDE		(1 << 29)
+#define MASK_MCMD12DRE		(1 << 26)
+#define MASK_MCMD12RBE		(1 << 25)
+#define MASK_MCMD12CRE		(1 << 24)
+#define MASK_MDTRANE		(1 << 23)
+#define MASK_MBUFRE		(1 << 22)
+#define MASK_MBUFWEN		(1 << 21)
+#define MASK_MBUFREN		(1 << 20)
+#define MASK_MCCSRCV		(1 << 19)
+#define MASK_MRBSYE		(1 << 17)
+#define MASK_MCRSPE		(1 << 16)
+#define MASK_MCMDVIO		(1 << 15)
+#define MASK_MBUFVIO		(1 << 14)
+#define MASK_MWDATERR		(1 << 11)
+#define MASK_MRDATERR		(1 << 10)
+#define MASK_MRIDXERR		(1 << 9)
+#define MASK_MRSPERR		(1 << 8)
+#define MASK_MCCSTO		(1 << 5)
+#define MASK_MCRCSTO		(1 << 4)
+#define MASK_MWDATTO		(1 << 3)
+#define MASK_MRDATTO		(1 << 2)
+#define MASK_MRBSYTO		(1 << 1)
+#define MASK_MRSPTO		(1 << 0)
+
+#define MASK_START_CMD		(MASK_MCMDVIO | MASK_MBUFVIO | MASK_MWDATERR | \
+				 MASK_MRDATERR | MASK_MRIDXERR | MASK_MRSPERR | \
+				 MASK_MCRCSTO | MASK_MWDATTO | \
+				 MASK_MRDATTO | MASK_MRBSYTO | MASK_MRSPTO)
+
+#define MASK_CLEAN		(INT_ERR_STS | MASK_MRBSYE | MASK_MCRSPE |	\
+				 MASK_MBUFREN | MASK_MBUFWEN |			\
+				 MASK_MCMD12DRE | MASK_MBUFRE | MASK_MDTRANE |	\
+				 MASK_MCMD12RBE | MASK_MCMD12CRE)
+
+/* CE_HOST_STS1 */
+#define STS1_CMDSEQ		(1 << 31)
+
+/* CE_HOST_STS2 */
+#define STS2_CRCSTE		(1 << 31)
+#define STS2_CRC16E		(1 << 30)
+#define STS2_AC12CRCE		(1 << 29)
+#define STS2_RSPCRC7E		(1 << 28)
+#define STS2_CRCSTEBE		(1 << 27)
+#define STS2_RDATEBE		(1 << 26)
+#define STS2_AC12REBE		(1 << 25)
+#define STS2_RSPEBE		(1 << 24)
+#define STS2_AC12IDXE		(1 << 23)
+#define STS2_RSPIDXE		(1 << 22)
+#define STS2_CCSTO		(1 << 15)
+#define STS2_RDATTO		(1 << 14)
+#define STS2_DATBSYTO		(1 << 13)
+#define STS2_CRCSTTO		(1 << 12)
+#define STS2_AC12BSYTO		(1 << 11)
+#define STS2_RSPBSYTO		(1 << 10)
+#define STS2_AC12RSPTO		(1 << 9)
+#define STS2_RSPTO		(1 << 8)
+#define STS2_CRC_ERR		(STS2_CRCSTE | STS2_CRC16E |		\
+				 STS2_AC12CRCE | STS2_RSPCRC7E | STS2_CRCSTEBE)
+#define STS2_TIMEOUT_ERR	(STS2_CCSTO | STS2_RDATTO |		\
+				 STS2_DATBSYTO | STS2_CRCSTTO |		\
+				 STS2_AC12BSYTO | STS2_RSPBSYTO |	\
+				 STS2_AC12RSPTO | STS2_RSPTO)
+
+#define CLKDEV_EMMC_DATA	52000000 /* 52 MHz */
+#define CLKDEV_MMC_DATA		20000000 /* 20 MHz */
+#define CLKDEV_INIT		400000   /* 400 kHz */
+
+enum sh_mmcif_state {
+	STATE_IDLE,
+	STATE_REQUEST,
+	STATE_IOS,
+	STATE_TIMEOUT,
+};
+
+enum sh_mmcif_wait_for {
+	MMCIF_WAIT_FOR_REQUEST,
+	MMCIF_WAIT_FOR_CMD,
+	MMCIF_WAIT_FOR_MREAD,
+	MMCIF_WAIT_FOR_MWRITE,
+	MMCIF_WAIT_FOR_READ,
+	MMCIF_WAIT_FOR_WRITE,
+	MMCIF_WAIT_FOR_READ_END,
+	MMCIF_WAIT_FOR_WRITE_END,
+	MMCIF_WAIT_FOR_STOP,
+};
+
+/*
+ * difference for each SoC
+ */
+struct sh_mmcif_host {
+	struct mmc_host *mmc;
+	struct mmc_request *mrq;
+	struct platform_device *pd;
+	struct clk *clk;
+	int bus_width;
+	unsigned char timing;
+	bool sd_error;
+	bool dying;
+	long timeout;
+	void __iomem *addr;
+	u32 *pio_ptr;
+	spinlock_t lock;		/* protect sh_mmcif_host::state */
+	enum sh_mmcif_state state;
+	enum sh_mmcif_wait_for wait_for;
+	struct delayed_work timeout_work;
+	size_t blocksize;
+	int sg_idx;
+	int sg_blkidx;
+	bool power;
+	bool ccs_enable;		/* Command Completion Signal support */
+	bool clk_ctrl2_enable;
+	struct mutex thread_lock;
+	u32 clkdiv_map;         /* see CE_CLK_CTRL::CLKDIV */
+
+	/* DMA support */
+	struct dma_chan		*chan_rx;
+	struct dma_chan		*chan_tx;
+	struct completion	dma_complete;
+	bool			dma_active;
+};
+
+static const struct of_device_id sh_mmcif_of_match[] = {
+	{ .compatible = "renesas,sh-mmcif" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, sh_mmcif_of_match);
+
+#define sh_mmcif_host_to_dev(host) (&host->pd->dev)
+
+static inline void sh_mmcif_bitset(struct sh_mmcif_host *host,
+					unsigned int reg, u32 val)
+{
+	writel(val | readl(host->addr + reg), host->addr + reg);
+}
+
+static inline void sh_mmcif_bitclr(struct sh_mmcif_host *host,
+					unsigned int reg, u32 val)
+{
+	writel(~val & readl(host->addr + reg), host->addr + reg);
+}
+
+static void sh_mmcif_dma_complete(void *arg)
+{
+	struct sh_mmcif_host *host = arg;
+	struct mmc_request *mrq = host->mrq;
+	struct device *dev = sh_mmcif_host_to_dev(host);
+
+	dev_dbg(dev, "Command completed\n");
+
+	if (WARN(!mrq || !mrq->data, "%s: NULL data in DMA completion!\n",
+		 dev_name(dev)))
+		return;
+
+	complete(&host->dma_complete);
+}
+
+static void sh_mmcif_start_dma_rx(struct sh_mmcif_host *host)
+{
+	struct mmc_data *data = host->mrq->data;
+	struct scatterlist *sg = data->sg;
+	struct dma_async_tx_descriptor *desc = NULL;
+	struct dma_chan *chan = host->chan_rx;
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	dma_cookie_t cookie = -EINVAL;
+	int ret;
+
+	ret = dma_map_sg(chan->device->dev, sg, data->sg_len,
+			 DMA_FROM_DEVICE);
+	if (ret > 0) {
+		host->dma_active = true;
+		desc = dmaengine_prep_slave_sg(chan, sg, ret,
+			DMA_DEV_TO_MEM, DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	}
+
+	if (desc) {
+		desc->callback = sh_mmcif_dma_complete;
+		desc->callback_param = host;
+		cookie = dmaengine_submit(desc);
+		sh_mmcif_bitset(host, MMCIF_CE_BUF_ACC, BUF_ACC_DMAREN);
+		dma_async_issue_pending(chan);
+	}
+	dev_dbg(dev, "%s(): mapped %d -> %d, cookie %d\n",
+		__func__, data->sg_len, ret, cookie);
+
+	if (!desc) {
+		/* DMA failed, fall back to PIO */
+		if (ret >= 0)
+			ret = -EIO;
+		host->chan_rx = NULL;
+		host->dma_active = false;
+		dma_release_channel(chan);
+		/* Free the Tx channel too */
+		chan = host->chan_tx;
+		if (chan) {
+			host->chan_tx = NULL;
+			dma_release_channel(chan);
+		}
+		dev_warn(dev,
+			 "DMA failed: %d, falling back to PIO\n", ret);
+		sh_mmcif_bitclr(host, MMCIF_CE_BUF_ACC, BUF_ACC_DMAREN | BUF_ACC_DMAWEN);
+	}
+
+	dev_dbg(dev, "%s(): desc %p, cookie %d, sg[%d]\n", __func__,
+		desc, cookie, data->sg_len);
+}
+
+static void sh_mmcif_start_dma_tx(struct sh_mmcif_host *host)
+{
+	struct mmc_data *data = host->mrq->data;
+	struct scatterlist *sg = data->sg;
+	struct dma_async_tx_descriptor *desc = NULL;
+	struct dma_chan *chan = host->chan_tx;
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	dma_cookie_t cookie = -EINVAL;
+	int ret;
+
+	ret = dma_map_sg(chan->device->dev, sg, data->sg_len,
+			 DMA_TO_DEVICE);
+	if (ret > 0) {
+		host->dma_active = true;
+		desc = dmaengine_prep_slave_sg(chan, sg, ret,
+			DMA_MEM_TO_DEV, DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	}
+
+	if (desc) {
+		desc->callback = sh_mmcif_dma_complete;
+		desc->callback_param = host;
+		cookie = dmaengine_submit(desc);
+		sh_mmcif_bitset(host, MMCIF_CE_BUF_ACC, BUF_ACC_DMAWEN);
+		dma_async_issue_pending(chan);
+	}
+	dev_dbg(dev, "%s(): mapped %d -> %d, cookie %d\n",
+		__func__, data->sg_len, ret, cookie);
+
+	if (!desc) {
+		/* DMA failed, fall back to PIO */
+		if (ret >= 0)
+			ret = -EIO;
+		host->chan_tx = NULL;
+		host->dma_active = false;
+		dma_release_channel(chan);
+		/* Free the Rx channel too */
+		chan = host->chan_rx;
+		if (chan) {
+			host->chan_rx = NULL;
+			dma_release_channel(chan);
+		}
+		dev_warn(dev,
+			 "DMA failed: %d, falling back to PIO\n", ret);
+		sh_mmcif_bitclr(host, MMCIF_CE_BUF_ACC, BUF_ACC_DMAREN | BUF_ACC_DMAWEN);
+	}
+
+	dev_dbg(dev, "%s(): desc %p, cookie %d\n", __func__,
+		desc, cookie);
+}
+
+static struct dma_chan *
+sh_mmcif_request_dma_pdata(struct sh_mmcif_host *host, uintptr_t slave_id)
+{
+	dma_cap_mask_t mask;
+
+	dma_cap_zero(mask);
+	dma_cap_set(DMA_SLAVE, mask);
+	if (slave_id <= 0)
+		return NULL;
+
+	return dma_request_channel(mask, shdma_chan_filter, (void *)slave_id);
+}
+
+static int sh_mmcif_dma_slave_config(struct sh_mmcif_host *host,
+				     struct dma_chan *chan,
+				     enum dma_transfer_direction direction)
+{
+	struct resource *res;
+	struct dma_slave_config cfg = { 0, };
+
+	res = platform_get_resource(host->pd, IORESOURCE_MEM, 0);
+	cfg.direction = direction;
+
+	if (direction == DMA_DEV_TO_MEM) {
+		cfg.src_addr = res->start + MMCIF_CE_DATA;
+		cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	} else {
+		cfg.dst_addr = res->start + MMCIF_CE_DATA;
+		cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	}
+
+	return dmaengine_slave_config(chan, &cfg);
+}
+
+static void sh_mmcif_request_dma(struct sh_mmcif_host *host)
+{
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	host->dma_active = false;
+
+	/* We can only either use DMA for both Tx and Rx or not use it at all */
+	if (IS_ENABLED(CONFIG_SUPERH) && dev->platform_data) {
+		struct sh_mmcif_plat_data *pdata = dev->platform_data;
+
+		host->chan_tx = sh_mmcif_request_dma_pdata(host,
+							pdata->slave_id_tx);
+		host->chan_rx = sh_mmcif_request_dma_pdata(host,
+							pdata->slave_id_rx);
+	} else {
+		host->chan_tx = dma_request_chan(dev, "tx");
+		if (IS_ERR(host->chan_tx))
+			host->chan_tx = NULL;
+		host->chan_rx = dma_request_chan(dev, "rx");
+		if (IS_ERR(host->chan_rx))
+			host->chan_rx = NULL;
+	}
+	dev_dbg(dev, "%s: got channel TX %p RX %p\n", __func__, host->chan_tx,
+		host->chan_rx);
+
+	if (!host->chan_tx || !host->chan_rx ||
+	    sh_mmcif_dma_slave_config(host, host->chan_tx, DMA_MEM_TO_DEV) ||
+	    sh_mmcif_dma_slave_config(host, host->chan_rx, DMA_DEV_TO_MEM))
+		goto error;
+
+	return;
+
+error:
+	if (host->chan_tx)
+		dma_release_channel(host->chan_tx);
+	if (host->chan_rx)
+		dma_release_channel(host->chan_rx);
+	host->chan_tx = host->chan_rx = NULL;
+}
+
+static void sh_mmcif_release_dma(struct sh_mmcif_host *host)
+{
+	sh_mmcif_bitclr(host, MMCIF_CE_BUF_ACC, BUF_ACC_DMAREN | BUF_ACC_DMAWEN);
+	/* Descriptors are freed automatically */
+	if (host->chan_tx) {
+		struct dma_chan *chan = host->chan_tx;
+		host->chan_tx = NULL;
+		dma_release_channel(chan);
+	}
+	if (host->chan_rx) {
+		struct dma_chan *chan = host->chan_rx;
+		host->chan_rx = NULL;
+		dma_release_channel(chan);
+	}
+
+	host->dma_active = false;
+}
+
+static void sh_mmcif_clock_control(struct sh_mmcif_host *host, unsigned int clk)
+{
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	struct sh_mmcif_plat_data *p = dev->platform_data;
+	bool sup_pclk = p ? p->sup_pclk : false;
+	unsigned int current_clk = clk_get_rate(host->clk);
+	unsigned int clkdiv;
+
+	sh_mmcif_bitclr(host, MMCIF_CE_CLK_CTRL, CLK_ENABLE);
+	sh_mmcif_bitclr(host, MMCIF_CE_CLK_CTRL, CLK_CLEAR);
+
+	if (!clk)
+		return;
+
+	if (host->clkdiv_map) {
+		unsigned int freq, best_freq, myclk, div, diff_min, diff;
+		int i;
+
+		clkdiv = 0;
+		diff_min = ~0;
+		best_freq = 0;
+		for (i = 31; i >= 0; i--) {
+			if (!((1 << i) & host->clkdiv_map))
+				continue;
+
+			/*
+			 * clk = parent_freq / div
+			 * -> parent_freq = clk x div
+			 */
+
+			div = 1 << (i + 1);
+			freq = clk_round_rate(host->clk, clk * div);
+			myclk = freq / div;
+			diff = (myclk > clk) ? myclk - clk : clk - myclk;
+
+			if (diff <= diff_min) {
+				best_freq = freq;
+				clkdiv = i;
+				diff_min = diff;
+			}
+		}
+
+		dev_dbg(dev, "clk %u/%u (%u, 0x%x)\n",
+			(best_freq / (1 << (clkdiv + 1))), clk,
+			best_freq, clkdiv);
+
+		clk_set_rate(host->clk, best_freq);
+		clkdiv = clkdiv << 16;
+	} else if (sup_pclk && clk == current_clk) {
+		clkdiv = CLK_SUP_PCLK;
+	} else {
+		clkdiv = (fls(DIV_ROUND_UP(current_clk, clk) - 1) - 1) << 16;
+	}
+
+	sh_mmcif_bitset(host, MMCIF_CE_CLK_CTRL, CLK_CLEAR & clkdiv);
+	sh_mmcif_bitset(host, MMCIF_CE_CLK_CTRL, CLK_ENABLE);
+}
+
+static void sh_mmcif_sync_reset(struct sh_mmcif_host *host)
+{
+	u32 tmp;
+
+	tmp = 0x010f0000 & sh_mmcif_readl(host->addr, MMCIF_CE_CLK_CTRL);
+
+	sh_mmcif_writel(host->addr, MMCIF_CE_VERSION, SOFT_RST_ON);
+	sh_mmcif_writel(host->addr, MMCIF_CE_VERSION, SOFT_RST_OFF);
+	if (host->ccs_enable)
+		tmp |= SCCSTO_29;
+	if (host->clk_ctrl2_enable)
+		sh_mmcif_writel(host->addr, MMCIF_CE_CLK_CTRL2, 0x0F0F0000);
+	sh_mmcif_bitset(host, MMCIF_CE_CLK_CTRL, tmp |
+		SRSPTO_256 | SRBSYTO_29 | SRWDTO_29);
+	/* byte swap on */
+	sh_mmcif_bitset(host, MMCIF_CE_BUF_ACC, BUF_ACC_ATYP);
+}
+
+static int sh_mmcif_error_manage(struct sh_mmcif_host *host)
+{
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	u32 state1, state2;
+	int ret, timeout;
+
+	host->sd_error = false;
+
+	state1 = sh_mmcif_readl(host->addr, MMCIF_CE_HOST_STS1);
+	state2 = sh_mmcif_readl(host->addr, MMCIF_CE_HOST_STS2);
+	dev_dbg(dev, "ERR HOST_STS1 = %08x\n", state1);
+	dev_dbg(dev, "ERR HOST_STS2 = %08x\n", state2);
+
+	if (state1 & STS1_CMDSEQ) {
+		sh_mmcif_bitset(host, MMCIF_CE_CMD_CTRL, CMD_CTRL_BREAK);
+		sh_mmcif_bitset(host, MMCIF_CE_CMD_CTRL, ~CMD_CTRL_BREAK);
+		for (timeout = 10000; timeout; timeout--) {
+			if (!(sh_mmcif_readl(host->addr, MMCIF_CE_HOST_STS1)
+			      & STS1_CMDSEQ))
+				break;
+			mdelay(1);
+		}
+		if (!timeout) {
+			dev_err(dev,
+				"Forced end of command sequence timeout err\n");
+			return -EIO;
+		}
+		sh_mmcif_sync_reset(host);
+		dev_dbg(dev, "Forced end of command sequence\n");
+		return -EIO;
+	}
+
+	if (state2 & STS2_CRC_ERR) {
+		dev_err(dev, " CRC error: state %u, wait %u\n",
+			host->state, host->wait_for);
+		ret = -EIO;
+	} else if (state2 & STS2_TIMEOUT_ERR) {
+		dev_err(dev, " Timeout: state %u, wait %u\n",
+			host->state, host->wait_for);
+		ret = -ETIMEDOUT;
+	} else {
+		dev_dbg(dev, " End/Index error: state %u, wait %u\n",
+			host->state, host->wait_for);
+		ret = -EIO;
+	}
+	return ret;
+}
+
+static bool sh_mmcif_next_block(struct sh_mmcif_host *host, u32 *p)
+{
+	struct mmc_data *data = host->mrq->data;
+
+	host->sg_blkidx += host->blocksize;
+
+	/* data->sg->length must be a multiple of host->blocksize? */
+	BUG_ON(host->sg_blkidx > data->sg->length);
+
+	if (host->sg_blkidx == data->sg->length) {
+		host->sg_blkidx = 0;
+		if (++host->sg_idx < data->sg_len)
+			host->pio_ptr = sg_virt(++data->sg);
+	} else {
+		host->pio_ptr = p;
+	}
+
+	return host->sg_idx != data->sg_len;
+}
+
+static void sh_mmcif_single_read(struct sh_mmcif_host *host,
+				 struct mmc_request *mrq)
+{
+	host->blocksize = (sh_mmcif_readl(host->addr, MMCIF_CE_BLOCK_SET) &
+			   BLOCK_SIZE_MASK) + 3;
+
+	host->wait_for = MMCIF_WAIT_FOR_READ;
+
+	/* buf read enable */
+	sh_mmcif_bitset(host, MMCIF_CE_INT_MASK, MASK_MBUFREN);
+}
+
+static bool sh_mmcif_read_block(struct sh_mmcif_host *host)
+{
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	struct mmc_data *data = host->mrq->data;
+	u32 *p = sg_virt(data->sg);
+	int i;
+
+	if (host->sd_error) {
+		data->error = sh_mmcif_error_manage(host);
+		dev_dbg(dev, "%s(): %d\n", __func__, data->error);
+		return false;
+	}
+
+	for (i = 0; i < host->blocksize / 4; i++)
+		*p++ = sh_mmcif_readl(host->addr, MMCIF_CE_DATA);
+
+	/* buffer read end */
+	sh_mmcif_bitset(host, MMCIF_CE_INT_MASK, MASK_MBUFRE);
+	host->wait_for = MMCIF_WAIT_FOR_READ_END;
+
+	return true;
+}
+
+static void sh_mmcif_multi_read(struct sh_mmcif_host *host,
+				struct mmc_request *mrq)
+{
+	struct mmc_data *data = mrq->data;
+
+	if (!data->sg_len || !data->sg->length)
+		return;
+
+	host->blocksize = sh_mmcif_readl(host->addr, MMCIF_CE_BLOCK_SET) &
+		BLOCK_SIZE_MASK;
+
+	host->wait_for = MMCIF_WAIT_FOR_MREAD;
+	host->sg_idx = 0;
+	host->sg_blkidx = 0;
+	host->pio_ptr = sg_virt(data->sg);
+
+	sh_mmcif_bitset(host, MMCIF_CE_INT_MASK, MASK_MBUFREN);
+}
+
+static bool sh_mmcif_mread_block(struct sh_mmcif_host *host)
+{
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	struct mmc_data *data = host->mrq->data;
+	u32 *p = host->pio_ptr;
+	int i;
+
+	if (host->sd_error) {
+		data->error = sh_mmcif_error_manage(host);
+		dev_dbg(dev, "%s(): %d\n", __func__, data->error);
+		return false;
+	}
+
+	BUG_ON(!data->sg->length);
+
+	for (i = 0; i < host->blocksize / 4; i++)
+		*p++ = sh_mmcif_readl(host->addr, MMCIF_CE_DATA);
+
+	if (!sh_mmcif_next_block(host, p))
+		return false;
+
+	sh_mmcif_bitset(host, MMCIF_CE_INT_MASK, MASK_MBUFREN);
+
+	return true;
+}
+
+static void sh_mmcif_single_write(struct sh_mmcif_host *host,
+					struct mmc_request *mrq)
+{
+	host->blocksize = (sh_mmcif_readl(host->addr, MMCIF_CE_BLOCK_SET) &
+			   BLOCK_SIZE_MASK) + 3;
+
+	host->wait_for = MMCIF_WAIT_FOR_WRITE;
+
+	/* buf write enable */
+	sh_mmcif_bitset(host, MMCIF_CE_INT_MASK, MASK_MBUFWEN);
+}
+
+static bool sh_mmcif_write_block(struct sh_mmcif_host *host)
+{
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	struct mmc_data *data = host->mrq->data;
+	u32 *p = sg_virt(data->sg);
+	int i;
+
+	if (host->sd_error) {
+		data->error = sh_mmcif_error_manage(host);
+		dev_dbg(dev, "%s(): %d\n", __func__, data->error);
+		return false;
+	}
+
+	for (i = 0; i < host->blocksize / 4; i++)
+		sh_mmcif_writel(host->addr, MMCIF_CE_DATA, *p++);
+
+	/* buffer write end */
+	sh_mmcif_bitset(host, MMCIF_CE_INT_MASK, MASK_MDTRANE);
+	host->wait_for = MMCIF_WAIT_FOR_WRITE_END;
+
+	return true;
+}
+
+static void sh_mmcif_multi_write(struct sh_mmcif_host *host,
+				struct mmc_request *mrq)
+{
+	struct mmc_data *data = mrq->data;
+
+	if (!data->sg_len || !data->sg->length)
+		return;
+
+	host->blocksize = sh_mmcif_readl(host->addr, MMCIF_CE_BLOCK_SET) &
+		BLOCK_SIZE_MASK;
+
+	host->wait_for = MMCIF_WAIT_FOR_MWRITE;
+	host->sg_idx = 0;
+	host->sg_blkidx = 0;
+	host->pio_ptr = sg_virt(data->sg);
+
+	sh_mmcif_bitset(host, MMCIF_CE_INT_MASK, MASK_MBUFWEN);
+}
+
+static bool sh_mmcif_mwrite_block(struct sh_mmcif_host *host)
+{
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	struct mmc_data *data = host->mrq->data;
+	u32 *p = host->pio_ptr;
+	int i;
+
+	if (host->sd_error) {
+		data->error = sh_mmcif_error_manage(host);
+		dev_dbg(dev, "%s(): %d\n", __func__, data->error);
+		return false;
+	}
+
+	BUG_ON(!data->sg->length);
+
+	for (i = 0; i < host->blocksize / 4; i++)
+		sh_mmcif_writel(host->addr, MMCIF_CE_DATA, *p++);
+
+	if (!sh_mmcif_next_block(host, p))
+		return false;
+
+	sh_mmcif_bitset(host, MMCIF_CE_INT_MASK, MASK_MBUFWEN);
+
+	return true;
+}
+
+static void sh_mmcif_get_response(struct sh_mmcif_host *host,
+						struct mmc_command *cmd)
+{
+	if (cmd->flags & MMC_RSP_136) {
+		cmd->resp[0] = sh_mmcif_readl(host->addr, MMCIF_CE_RESP3);
+		cmd->resp[1] = sh_mmcif_readl(host->addr, MMCIF_CE_RESP2);
+		cmd->resp[2] = sh_mmcif_readl(host->addr, MMCIF_CE_RESP1);
+		cmd->resp[3] = sh_mmcif_readl(host->addr, MMCIF_CE_RESP0);
+	} else
+		cmd->resp[0] = sh_mmcif_readl(host->addr, MMCIF_CE_RESP0);
+}
+
+static void sh_mmcif_get_cmd12response(struct sh_mmcif_host *host,
+						struct mmc_command *cmd)
+{
+	cmd->resp[0] = sh_mmcif_readl(host->addr, MMCIF_CE_RESP_CMD12);
+}
+
+static u32 sh_mmcif_set_cmd(struct sh_mmcif_host *host,
+			    struct mmc_request *mrq)
+{
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	struct mmc_data *data = mrq->data;
+	struct mmc_command *cmd = mrq->cmd;
+	u32 opc = cmd->opcode;
+	u32 tmp = 0;
+
+	/* Response Type check */
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE:
+		tmp |= CMD_SET_RTYP_NO;
+		break;
+	case MMC_RSP_R1:
+	case MMC_RSP_R3:
+		tmp |= CMD_SET_RTYP_6B;
+		break;
+	case MMC_RSP_R1B:
+		tmp |= CMD_SET_RBSY | CMD_SET_RTYP_6B;
+		break;
+	case MMC_RSP_R2:
+		tmp |= CMD_SET_RTYP_17B;
+		break;
+	default:
+		dev_err(dev, "Unsupported response type.\n");
+		break;
+	}
+
+	/* WDAT / DATW */
+	if (data) {
+		tmp |= CMD_SET_WDAT;
+		switch (host->bus_width) {
+		case MMC_BUS_WIDTH_1:
+			tmp |= CMD_SET_DATW_1;
+			break;
+		case MMC_BUS_WIDTH_4:
+			tmp |= CMD_SET_DATW_4;
+			break;
+		case MMC_BUS_WIDTH_8:
+			tmp |= CMD_SET_DATW_8;
+			break;
+		default:
+			dev_err(dev, "Unsupported bus width.\n");
+			break;
+		}
+		switch (host->timing) {
+		case MMC_TIMING_MMC_DDR52:
+			/*
+			 * MMC core will only set this timing, if the host
+			 * advertises the MMC_CAP_1_8V_DDR/MMC_CAP_1_2V_DDR
+			 * capability. MMCIF implementations with this
+			 * capability, e.g. sh73a0, will have to set it
+			 * in their platform data.
+			 */
+			tmp |= CMD_SET_DARS;
+			break;
+		}
+	}
+	/* DWEN */
+	if (opc == MMC_WRITE_BLOCK || opc == MMC_WRITE_MULTIPLE_BLOCK)
+		tmp |= CMD_SET_DWEN;
+	/* CMLTE/CMD12EN */
+	if (opc == MMC_READ_MULTIPLE_BLOCK || opc == MMC_WRITE_MULTIPLE_BLOCK) {
+		tmp |= CMD_SET_CMLTE | CMD_SET_CMD12EN;
+		sh_mmcif_bitset(host, MMCIF_CE_BLOCK_SET,
+				data->blocks << 16);
+	}
+	/* RIDXC[1:0] check bits */
+	if (opc == MMC_SEND_OP_COND || opc == MMC_ALL_SEND_CID ||
+	    opc == MMC_SEND_CSD || opc == MMC_SEND_CID)
+		tmp |= CMD_SET_RIDXC_BITS;
+	/* RCRC7C[1:0] check bits */
+	if (opc == MMC_SEND_OP_COND)
+		tmp |= CMD_SET_CRC7C_BITS;
+	/* RCRC7C[1:0] internal CRC7 */
+	if (opc == MMC_ALL_SEND_CID ||
+		opc == MMC_SEND_CSD || opc == MMC_SEND_CID)
+		tmp |= CMD_SET_CRC7C_INTERNAL;
+
+	return (opc << 24) | tmp;
+}
+
+static int sh_mmcif_data_trans(struct sh_mmcif_host *host,
+			       struct mmc_request *mrq, u32 opc)
+{
+	struct device *dev = sh_mmcif_host_to_dev(host);
+
+	switch (opc) {
+	case MMC_READ_MULTIPLE_BLOCK:
+		sh_mmcif_multi_read(host, mrq);
+		return 0;
+	case MMC_WRITE_MULTIPLE_BLOCK:
+		sh_mmcif_multi_write(host, mrq);
+		return 0;
+	case MMC_WRITE_BLOCK:
+		sh_mmcif_single_write(host, mrq);
+		return 0;
+	case MMC_READ_SINGLE_BLOCK:
+	case MMC_SEND_EXT_CSD:
+		sh_mmcif_single_read(host, mrq);
+		return 0;
+	default:
+		dev_err(dev, "Unsupported CMD%d\n", opc);
+		return -EINVAL;
+	}
+}
+
+static void sh_mmcif_start_cmd(struct sh_mmcif_host *host,
+			       struct mmc_request *mrq)
+{
+	struct mmc_command *cmd = mrq->cmd;
+	u32 opc;
+	u32 mask = 0;
+	unsigned long flags;
+
+	if (cmd->flags & MMC_RSP_BUSY)
+		mask = MASK_START_CMD | MASK_MRBSYE;
+	else
+		mask = MASK_START_CMD | MASK_MCRSPE;
+
+	if (host->ccs_enable)
+		mask |= MASK_MCCSTO;
+
+	if (mrq->data) {
+		sh_mmcif_writel(host->addr, MMCIF_CE_BLOCK_SET, 0);
+		sh_mmcif_writel(host->addr, MMCIF_CE_BLOCK_SET,
+				mrq->data->blksz);
+	}
+	opc = sh_mmcif_set_cmd(host, mrq);
+
+	if (host->ccs_enable)
+		sh_mmcif_writel(host->addr, MMCIF_CE_INT, 0xD80430C0);
+	else
+		sh_mmcif_writel(host->addr, MMCIF_CE_INT, 0xD80430C0 | INT_CCS);
+	sh_mmcif_writel(host->addr, MMCIF_CE_INT_MASK, mask);
+	/* set arg */
+	sh_mmcif_writel(host->addr, MMCIF_CE_ARG, cmd->arg);
+	/* set cmd */
+	spin_lock_irqsave(&host->lock, flags);
+	sh_mmcif_writel(host->addr, MMCIF_CE_CMD_SET, opc);
+
+	host->wait_for = MMCIF_WAIT_FOR_CMD;
+	schedule_delayed_work(&host->timeout_work, host->timeout);
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void sh_mmcif_stop_cmd(struct sh_mmcif_host *host,
+			      struct mmc_request *mrq)
+{
+	struct device *dev = sh_mmcif_host_to_dev(host);
+
+	switch (mrq->cmd->opcode) {
+	case MMC_READ_MULTIPLE_BLOCK:
+		sh_mmcif_bitset(host, MMCIF_CE_INT_MASK, MASK_MCMD12DRE);
+		break;
+	case MMC_WRITE_MULTIPLE_BLOCK:
+		sh_mmcif_bitset(host, MMCIF_CE_INT_MASK, MASK_MCMD12RBE);
+		break;
+	default:
+		dev_err(dev, "unsupported stop cmd\n");
+		mrq->stop->error = sh_mmcif_error_manage(host);
+		return;
+	}
+
+	host->wait_for = MMCIF_WAIT_FOR_STOP;
+}
+
+static void sh_mmcif_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct sh_mmcif_host *host = mmc_priv(mmc);
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (host->state != STATE_IDLE) {
+		dev_dbg(dev, "%s() rejected, state %u\n",
+			__func__, host->state);
+		spin_unlock_irqrestore(&host->lock, flags);
+		mrq->cmd->error = -EAGAIN;
+		mmc_request_done(mmc, mrq);
+		return;
+	}
+
+	host->state = STATE_REQUEST;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	host->mrq = mrq;
+
+	sh_mmcif_start_cmd(host, mrq);
+}
+
+static void sh_mmcif_clk_setup(struct sh_mmcif_host *host)
+{
+	struct device *dev = sh_mmcif_host_to_dev(host);
+
+	if (host->mmc->f_max) {
+		unsigned int f_max, f_min = 0, f_min_old;
+
+		f_max = host->mmc->f_max;
+		for (f_min_old = f_max; f_min_old > 2;) {
+			f_min = clk_round_rate(host->clk, f_min_old / 2);
+			if (f_min == f_min_old)
+				break;
+			f_min_old = f_min;
+		}
+
+		/*
+		 * This driver assumes this SoC is R-Car Gen2 or later
+		 */
+		host->clkdiv_map = 0x3ff;
+
+		host->mmc->f_max = f_max / (1 << ffs(host->clkdiv_map));
+		host->mmc->f_min = f_min / (1 << fls(host->clkdiv_map));
+	} else {
+		unsigned int clk = clk_get_rate(host->clk);
+
+		host->mmc->f_max = clk / 2;
+		host->mmc->f_min = clk / 512;
+	}
+
+	dev_dbg(dev, "clk max/min = %d/%d\n",
+		host->mmc->f_max, host->mmc->f_min);
+}
+
+static void sh_mmcif_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sh_mmcif_host *host = mmc_priv(mmc);
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (host->state != STATE_IDLE) {
+		dev_dbg(dev, "%s() rejected, state %u\n",
+			__func__, host->state);
+		spin_unlock_irqrestore(&host->lock, flags);
+		return;
+	}
+
+	host->state = STATE_IOS;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	switch (ios->power_mode) {
+	case MMC_POWER_UP:
+		if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, ios->vdd);
+		if (!host->power) {
+			clk_prepare_enable(host->clk);
+			pm_runtime_get_sync(dev);
+			sh_mmcif_sync_reset(host);
+			sh_mmcif_request_dma(host);
+			host->power = true;
+		}
+		break;
+	case MMC_POWER_OFF:
+		if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+		if (host->power) {
+			sh_mmcif_clock_control(host, 0);
+			sh_mmcif_release_dma(host);
+			pm_runtime_put(dev);
+			clk_disable_unprepare(host->clk);
+			host->power = false;
+		}
+		break;
+	case MMC_POWER_ON:
+		sh_mmcif_clock_control(host, ios->clock);
+		break;
+	}
+
+	host->timing = ios->timing;
+	host->bus_width = ios->bus_width;
+	host->state = STATE_IDLE;
+}
+
+static const struct mmc_host_ops sh_mmcif_ops = {
+	.request	= sh_mmcif_request,
+	.set_ios	= sh_mmcif_set_ios,
+	.get_cd		= mmc_gpio_get_cd,
+};
+
+static bool sh_mmcif_end_cmd(struct sh_mmcif_host *host)
+{
+	struct mmc_command *cmd = host->mrq->cmd;
+	struct mmc_data *data = host->mrq->data;
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	long time;
+
+	if (host->sd_error) {
+		switch (cmd->opcode) {
+		case MMC_ALL_SEND_CID:
+		case MMC_SELECT_CARD:
+		case MMC_APP_CMD:
+			cmd->error = -ETIMEDOUT;
+			break;
+		default:
+			cmd->error = sh_mmcif_error_manage(host);
+			break;
+		}
+		dev_dbg(dev, "CMD%d error %d\n",
+			cmd->opcode, cmd->error);
+		host->sd_error = false;
+		return false;
+	}
+	if (!(cmd->flags & MMC_RSP_PRESENT)) {
+		cmd->error = 0;
+		return false;
+	}
+
+	sh_mmcif_get_response(host, cmd);
+
+	if (!data)
+		return false;
+
+	/*
+	 * Completion can be signalled from DMA callback and error, so, have to
+	 * reset here, before setting .dma_active
+	 */
+	init_completion(&host->dma_complete);
+
+	if (data->flags & MMC_DATA_READ) {
+		if (host->chan_rx)
+			sh_mmcif_start_dma_rx(host);
+	} else {
+		if (host->chan_tx)
+			sh_mmcif_start_dma_tx(host);
+	}
+
+	if (!host->dma_active) {
+		data->error = sh_mmcif_data_trans(host, host->mrq, cmd->opcode);
+		return !data->error;
+	}
+
+	/* Running in the IRQ thread, can sleep */
+	time = wait_for_completion_interruptible_timeout(&host->dma_complete,
+							 host->timeout);
+
+	if (data->flags & MMC_DATA_READ)
+		dma_unmap_sg(host->chan_rx->device->dev,
+			     data->sg, data->sg_len,
+			     DMA_FROM_DEVICE);
+	else
+		dma_unmap_sg(host->chan_tx->device->dev,
+			     data->sg, data->sg_len,
+			     DMA_TO_DEVICE);
+
+	if (host->sd_error) {
+		dev_err(host->mmc->parent,
+			"Error IRQ while waiting for DMA completion!\n");
+		/* Woken up by an error IRQ: abort DMA */
+		data->error = sh_mmcif_error_manage(host);
+	} else if (!time) {
+		dev_err(host->mmc->parent, "DMA timeout!\n");
+		data->error = -ETIMEDOUT;
+	} else if (time < 0) {
+		dev_err(host->mmc->parent,
+			"wait_for_completion_...() error %ld!\n", time);
+		data->error = time;
+	}
+	sh_mmcif_bitclr(host, MMCIF_CE_BUF_ACC,
+			BUF_ACC_DMAREN | BUF_ACC_DMAWEN);
+	host->dma_active = false;
+
+	if (data->error) {
+		data->bytes_xfered = 0;
+		/* Abort DMA */
+		if (data->flags & MMC_DATA_READ)
+			dmaengine_terminate_all(host->chan_rx);
+		else
+			dmaengine_terminate_all(host->chan_tx);
+	}
+
+	return false;
+}
+
+static irqreturn_t sh_mmcif_irqt(int irq, void *dev_id)
+{
+	struct sh_mmcif_host *host = dev_id;
+	struct mmc_request *mrq;
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	bool wait = false;
+	unsigned long flags;
+	int wait_work;
+
+	spin_lock_irqsave(&host->lock, flags);
+	wait_work = host->wait_for;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	cancel_delayed_work_sync(&host->timeout_work);
+
+	mutex_lock(&host->thread_lock);
+
+	mrq = host->mrq;
+	if (!mrq) {
+		dev_dbg(dev, "IRQ thread state %u, wait %u: NULL mrq!\n",
+			host->state, host->wait_for);
+		mutex_unlock(&host->thread_lock);
+		return IRQ_HANDLED;
+	}
+
+	/*
+	 * All handlers return true, if processing continues, and false, if the
+	 * request has to be completed - successfully or not
+	 */
+	switch (wait_work) {
+	case MMCIF_WAIT_FOR_REQUEST:
+		/* We're too late, the timeout has already kicked in */
+		mutex_unlock(&host->thread_lock);
+		return IRQ_HANDLED;
+	case MMCIF_WAIT_FOR_CMD:
+		/* Wait for data? */
+		wait = sh_mmcif_end_cmd(host);
+		break;
+	case MMCIF_WAIT_FOR_MREAD:
+		/* Wait for more data? */
+		wait = sh_mmcif_mread_block(host);
+		break;
+	case MMCIF_WAIT_FOR_READ:
+		/* Wait for data end? */
+		wait = sh_mmcif_read_block(host);
+		break;
+	case MMCIF_WAIT_FOR_MWRITE:
+		/* Wait data to write? */
+		wait = sh_mmcif_mwrite_block(host);
+		break;
+	case MMCIF_WAIT_FOR_WRITE:
+		/* Wait for data end? */
+		wait = sh_mmcif_write_block(host);
+		break;
+	case MMCIF_WAIT_FOR_STOP:
+		if (host->sd_error) {
+			mrq->stop->error = sh_mmcif_error_manage(host);
+			dev_dbg(dev, "%s(): %d\n", __func__, mrq->stop->error);
+			break;
+		}
+		sh_mmcif_get_cmd12response(host, mrq->stop);
+		mrq->stop->error = 0;
+		break;
+	case MMCIF_WAIT_FOR_READ_END:
+	case MMCIF_WAIT_FOR_WRITE_END:
+		if (host->sd_error) {
+			mrq->data->error = sh_mmcif_error_manage(host);
+			dev_dbg(dev, "%s(): %d\n", __func__, mrq->data->error);
+		}
+		break;
+	default:
+		BUG();
+	}
+
+	if (wait) {
+		schedule_delayed_work(&host->timeout_work, host->timeout);
+		/* Wait for more data */
+		mutex_unlock(&host->thread_lock);
+		return IRQ_HANDLED;
+	}
+
+	if (host->wait_for != MMCIF_WAIT_FOR_STOP) {
+		struct mmc_data *data = mrq->data;
+		if (!mrq->cmd->error && data && !data->error)
+			data->bytes_xfered =
+				data->blocks * data->blksz;
+
+		if (mrq->stop && !mrq->cmd->error && (!data || !data->error)) {
+			sh_mmcif_stop_cmd(host, mrq);
+			if (!mrq->stop->error) {
+				schedule_delayed_work(&host->timeout_work, host->timeout);
+				mutex_unlock(&host->thread_lock);
+				return IRQ_HANDLED;
+			}
+		}
+	}
+
+	host->wait_for = MMCIF_WAIT_FOR_REQUEST;
+	host->state = STATE_IDLE;
+	host->mrq = NULL;
+	mmc_request_done(host->mmc, mrq);
+
+	mutex_unlock(&host->thread_lock);
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t sh_mmcif_intr(int irq, void *dev_id)
+{
+	struct sh_mmcif_host *host = dev_id;
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	u32 state, mask;
+
+	state = sh_mmcif_readl(host->addr, MMCIF_CE_INT);
+	mask = sh_mmcif_readl(host->addr, MMCIF_CE_INT_MASK);
+	if (host->ccs_enable)
+		sh_mmcif_writel(host->addr, MMCIF_CE_INT, ~(state & mask));
+	else
+		sh_mmcif_writel(host->addr, MMCIF_CE_INT, INT_CCS | ~(state & mask));
+	sh_mmcif_bitclr(host, MMCIF_CE_INT_MASK, state & MASK_CLEAN);
+
+	if (state & ~MASK_CLEAN)
+		dev_dbg(dev, "IRQ state = 0x%08x incompletely cleared\n",
+			state);
+
+	if (state & INT_ERR_STS || state & ~INT_ALL) {
+		host->sd_error = true;
+		dev_dbg(dev, "int err state = 0x%08x\n", state);
+	}
+	if (state & ~(INT_CMD12RBE | INT_CMD12CRE)) {
+		if (!host->mrq)
+			dev_dbg(dev, "NULL IRQ state = 0x%08x\n", state);
+		if (!host->dma_active)
+			return IRQ_WAKE_THREAD;
+		else if (host->sd_error)
+			sh_mmcif_dma_complete(host);
+	} else {
+		dev_dbg(dev, "Unexpected IRQ 0x%x\n", state);
+	}
+
+	return IRQ_HANDLED;
+}
+
+static void sh_mmcif_timeout_work(struct work_struct *work)
+{
+	struct delayed_work *d = to_delayed_work(work);
+	struct sh_mmcif_host *host = container_of(d, struct sh_mmcif_host, timeout_work);
+	struct mmc_request *mrq = host->mrq;
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	unsigned long flags;
+
+	if (host->dying)
+		/* Don't run after mmc_remove_host() */
+		return;
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (host->state == STATE_IDLE) {
+		spin_unlock_irqrestore(&host->lock, flags);
+		return;
+	}
+
+	dev_err(dev, "Timeout waiting for %u on CMD%u\n",
+		host->wait_for, mrq->cmd->opcode);
+
+	host->state = STATE_TIMEOUT;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	/*
+	 * Handle races with cancel_delayed_work(), unless
+	 * cancel_delayed_work_sync() is used
+	 */
+	switch (host->wait_for) {
+	case MMCIF_WAIT_FOR_CMD:
+		mrq->cmd->error = sh_mmcif_error_manage(host);
+		break;
+	case MMCIF_WAIT_FOR_STOP:
+		mrq->stop->error = sh_mmcif_error_manage(host);
+		break;
+	case MMCIF_WAIT_FOR_MREAD:
+	case MMCIF_WAIT_FOR_MWRITE:
+	case MMCIF_WAIT_FOR_READ:
+	case MMCIF_WAIT_FOR_WRITE:
+	case MMCIF_WAIT_FOR_READ_END:
+	case MMCIF_WAIT_FOR_WRITE_END:
+		mrq->data->error = sh_mmcif_error_manage(host);
+		break;
+	default:
+		BUG();
+	}
+
+	host->state = STATE_IDLE;
+	host->wait_for = MMCIF_WAIT_FOR_REQUEST;
+	host->mrq = NULL;
+	mmc_request_done(host->mmc, mrq);
+}
+
+static void sh_mmcif_init_ocr(struct sh_mmcif_host *host)
+{
+	struct device *dev = sh_mmcif_host_to_dev(host);
+	struct sh_mmcif_plat_data *pd = dev->platform_data;
+	struct mmc_host *mmc = host->mmc;
+
+	mmc_regulator_get_supply(mmc);
+
+	if (!pd)
+		return;
+
+	if (!mmc->ocr_avail)
+		mmc->ocr_avail = pd->ocr;
+	else if (pd->ocr)
+		dev_warn(mmc_dev(mmc), "Platform OCR mask is ignored\n");
+}
+
+static int sh_mmcif_probe(struct platform_device *pdev)
+{
+	int ret = 0, irq[2];
+	struct mmc_host *mmc;
+	struct sh_mmcif_host *host;
+	struct device *dev = &pdev->dev;
+	struct sh_mmcif_plat_data *pd = dev->platform_data;
+	void __iomem *reg;
+	const char *name;
+
+	irq[0] = platform_get_irq(pdev, 0);
+	irq[1] = platform_get_irq_optional(pdev, 1);
+	if (irq[0] < 0)
+		return irq[0];
+
+	reg = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(reg))
+		return PTR_ERR(reg);
+
+	mmc = mmc_alloc_host(sizeof(struct sh_mmcif_host), dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	ret = mmc_of_parse(mmc);
+	if (ret < 0)
+		goto err_host;
+
+	host		= mmc_priv(mmc);
+	host->mmc	= mmc;
+	host->addr	= reg;
+	host->timeout	= msecs_to_jiffies(10000);
+	host->ccs_enable = true;
+	host->clk_ctrl2_enable = false;
+
+	host->pd = pdev;
+
+	spin_lock_init(&host->lock);
+
+	mmc->ops = &sh_mmcif_ops;
+	sh_mmcif_init_ocr(host);
+
+	mmc->caps |= MMC_CAP_MMC_HIGHSPEED | MMC_CAP_WAIT_WHILE_BUSY;
+	mmc->caps2 |= MMC_CAP2_NO_SD | MMC_CAP2_NO_SDIO;
+	mmc->max_busy_timeout = 10000;
+
+	if (pd && pd->caps)
+		mmc->caps |= pd->caps;
+	mmc->max_segs = 32;
+	mmc->max_blk_size = 512;
+	mmc->max_req_size = PAGE_SIZE * mmc->max_segs;
+	mmc->max_blk_count = mmc->max_req_size / mmc->max_blk_size;
+	mmc->max_seg_size = mmc->max_req_size;
+
+	platform_set_drvdata(pdev, host);
+
+	host->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(host->clk)) {
+		ret = PTR_ERR(host->clk);
+		dev_err(dev, "cannot get clock: %d\n", ret);
+		goto err_host;
+	}
+
+	ret = clk_prepare_enable(host->clk);
+	if (ret < 0)
+		goto err_host;
+
+	sh_mmcif_clk_setup(host);
+
+	pm_runtime_enable(dev);
+	host->power = false;
+
+	ret = pm_runtime_get_sync(dev);
+	if (ret < 0)
+		goto err_clk;
+
+	INIT_DELAYED_WORK(&host->timeout_work, sh_mmcif_timeout_work);
+
+	sh_mmcif_sync_reset(host);
+	sh_mmcif_writel(host->addr, MMCIF_CE_INT_MASK, MASK_ALL);
+
+	name = irq[1] < 0 ? dev_name(dev) : "sh_mmc:error";
+	ret = devm_request_threaded_irq(dev, irq[0], sh_mmcif_intr,
+					sh_mmcif_irqt, 0, name, host);
+	if (ret) {
+		dev_err(dev, "request_irq error (%s)\n", name);
+		goto err_clk;
+	}
+	if (irq[1] >= 0) {
+		ret = devm_request_threaded_irq(dev, irq[1],
+						sh_mmcif_intr, sh_mmcif_irqt,
+						0, "sh_mmc:int", host);
+		if (ret) {
+			dev_err(dev, "request_irq error (sh_mmc:int)\n");
+			goto err_clk;
+		}
+	}
+
+	mutex_init(&host->thread_lock);
+
+	ret = mmc_add_host(mmc);
+	if (ret < 0)
+		goto err_clk;
+
+	dev_pm_qos_expose_latency_limit(dev, 100);
+
+	dev_info(dev, "Chip version 0x%04x, clock rate %luMHz\n",
+		 sh_mmcif_readl(host->addr, MMCIF_CE_VERSION) & 0xffff,
+		 clk_get_rate(host->clk) / 1000000UL);
+
+	pm_runtime_put(dev);
+	clk_disable_unprepare(host->clk);
+	return ret;
+
+err_clk:
+	clk_disable_unprepare(host->clk);
+	pm_runtime_put_sync(dev);
+	pm_runtime_disable(dev);
+err_host:
+	mmc_free_host(mmc);
+	return ret;
+}
+
+static int sh_mmcif_remove(struct platform_device *pdev)
+{
+	struct sh_mmcif_host *host = platform_get_drvdata(pdev);
+
+	host->dying = true;
+	clk_prepare_enable(host->clk);
+	pm_runtime_get_sync(&pdev->dev);
+
+	dev_pm_qos_hide_latency_limit(&pdev->dev);
+
+	mmc_remove_host(host->mmc);
+	sh_mmcif_writel(host->addr, MMCIF_CE_INT_MASK, MASK_ALL);
+
+	/*
+	 * FIXME: cancel_delayed_work(_sync)() and free_irq() race with the
+	 * mmc_remove_host() call above. But swapping order doesn't help either
+	 * (a query on the linux-mmc mailing list didn't bring any replies).
+	 */
+	cancel_delayed_work_sync(&host->timeout_work);
+
+	clk_disable_unprepare(host->clk);
+	mmc_free_host(host->mmc);
+	pm_runtime_put_sync(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int sh_mmcif_suspend(struct device *dev)
+{
+	struct sh_mmcif_host *host = dev_get_drvdata(dev);
+
+	pm_runtime_get_sync(dev);
+	sh_mmcif_writel(host->addr, MMCIF_CE_INT_MASK, MASK_ALL);
+	pm_runtime_put(dev);
+
+	return 0;
+}
+
+static int sh_mmcif_resume(struct device *dev)
+{
+	return 0;
+}
+#endif
+
+static const struct dev_pm_ops sh_mmcif_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(sh_mmcif_suspend, sh_mmcif_resume)
+};
+
+static struct platform_driver sh_mmcif_driver = {
+	.probe		= sh_mmcif_probe,
+	.remove		= sh_mmcif_remove,
+	.driver		= {
+		.name	= DRIVER_NAME,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm	= &sh_mmcif_dev_pm_ops,
+		.of_match_table = sh_mmcif_of_match,
+	},
+};
+
+module_platform_driver(sh_mmcif_driver);
+
+MODULE_DESCRIPTION("SuperH on-chip MMC/eMMC interface driver");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:" DRIVER_NAME);
+MODULE_AUTHOR("Yusuke Goda <yusuke.goda.sx@renesas.com>");
diff --git a/drivers/mmc/host/sunxi-mmc.c b/drivers/mmc/host/sunxi-mmc.c
new file mode 100644
index 000000000..b834fde3f
--- /dev/null
+++ b/drivers/mmc/host/sunxi-mmc.c
@@ -0,0 +1,1531 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Driver for sunxi SD/MMC host controllers
+ * (C) Copyright 2007-2011 Reuuimlla Technology Co., Ltd.
+ * (C) Copyright 2007-2011 Aaron Maoye <leafy.myeh@reuuimllatech.com>
+ * (C) Copyright 2013-2014 O2S GmbH <www.o2s.ch>
+ * (C) Copyright 2013-2014 David Lanzendörfer <david.lanzendoerfer@o2s.ch>
+ * (C) Copyright 2013-2014 Hans de Goede <hdegoede@redhat.com>
+ * (C) Copyright 2017 Sootech SA
+ */
+
+#include <linux/clk.h>
+#include <linux/clk/sunxi-ng.h>
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/err.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/core.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sd.h>
+#include <linux/mmc/sdio.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/of_platform.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/regulator/consumer.h>
+#include <linux/reset.h>
+#include <linux/scatterlist.h>
+#include <linux/slab.h>
+#include <linux/spinlock.h>
+
+/* register offset definitions */
+#define SDXC_REG_GCTRL	(0x00) /* SMC Global Control Register */
+#define SDXC_REG_CLKCR	(0x04) /* SMC Clock Control Register */
+#define SDXC_REG_TMOUT	(0x08) /* SMC Time Out Register */
+#define SDXC_REG_WIDTH	(0x0C) /* SMC Bus Width Register */
+#define SDXC_REG_BLKSZ	(0x10) /* SMC Block Size Register */
+#define SDXC_REG_BCNTR	(0x14) /* SMC Byte Count Register */
+#define SDXC_REG_CMDR	(0x18) /* SMC Command Register */
+#define SDXC_REG_CARG	(0x1C) /* SMC Argument Register */
+#define SDXC_REG_RESP0	(0x20) /* SMC Response Register 0 */
+#define SDXC_REG_RESP1	(0x24) /* SMC Response Register 1 */
+#define SDXC_REG_RESP2	(0x28) /* SMC Response Register 2 */
+#define SDXC_REG_RESP3	(0x2C) /* SMC Response Register 3 */
+#define SDXC_REG_IMASK	(0x30) /* SMC Interrupt Mask Register */
+#define SDXC_REG_MISTA	(0x34) /* SMC Masked Interrupt Status Register */
+#define SDXC_REG_RINTR	(0x38) /* SMC Raw Interrupt Status Register */
+#define SDXC_REG_STAS	(0x3C) /* SMC Status Register */
+#define SDXC_REG_FTRGL	(0x40) /* SMC FIFO Threshold Watermark Registe */
+#define SDXC_REG_FUNS	(0x44) /* SMC Function Select Register */
+#define SDXC_REG_CBCR	(0x48) /* SMC CIU Byte Count Register */
+#define SDXC_REG_BBCR	(0x4C) /* SMC BIU Byte Count Register */
+#define SDXC_REG_DBGC	(0x50) /* SMC Debug Enable Register */
+#define SDXC_REG_HWRST	(0x78) /* SMC Card Hardware Reset for Register */
+#define SDXC_REG_DMAC	(0x80) /* SMC IDMAC Control Register */
+#define SDXC_REG_DLBA	(0x84) /* SMC IDMAC Descriptor List Base Addre */
+#define SDXC_REG_IDST	(0x88) /* SMC IDMAC Status Register */
+#define SDXC_REG_IDIE	(0x8C) /* SMC IDMAC Interrupt Enable Register */
+#define SDXC_REG_CHDA	(0x90)
+#define SDXC_REG_CBDA	(0x94)
+
+/* New registers introduced in A64 */
+#define SDXC_REG_A12A		0x058 /* SMC Auto Command 12 Register */
+#define SDXC_REG_SD_NTSR	0x05C /* SMC New Timing Set Register */
+#define SDXC_REG_DRV_DL		0x140 /* Drive Delay Control Register */
+#define SDXC_REG_SAMP_DL_REG	0x144 /* SMC sample delay control */
+#define SDXC_REG_DS_DL_REG	0x148 /* SMC data strobe delay control */
+
+#define mmc_readl(host, reg) \
+	readl((host)->reg_base + SDXC_##reg)
+#define mmc_writel(host, reg, value) \
+	writel((value), (host)->reg_base + SDXC_##reg)
+
+/* global control register bits */
+#define SDXC_SOFT_RESET			BIT(0)
+#define SDXC_FIFO_RESET			BIT(1)
+#define SDXC_DMA_RESET			BIT(2)
+#define SDXC_INTERRUPT_ENABLE_BIT	BIT(4)
+#define SDXC_DMA_ENABLE_BIT		BIT(5)
+#define SDXC_DEBOUNCE_ENABLE_BIT	BIT(8)
+#define SDXC_POSEDGE_LATCH_DATA		BIT(9)
+#define SDXC_DDR_MODE			BIT(10)
+#define SDXC_MEMORY_ACCESS_DONE		BIT(29)
+#define SDXC_ACCESS_DONE_DIRECT		BIT(30)
+#define SDXC_ACCESS_BY_AHB		BIT(31)
+#define SDXC_ACCESS_BY_DMA		(0 << 31)
+#define SDXC_HARDWARE_RESET \
+	(SDXC_SOFT_RESET | SDXC_FIFO_RESET | SDXC_DMA_RESET)
+
+/* clock control bits */
+#define SDXC_MASK_DATA0			BIT(31)
+#define SDXC_CARD_CLOCK_ON		BIT(16)
+#define SDXC_LOW_POWER_ON		BIT(17)
+
+/* bus width */
+#define SDXC_WIDTH1			0
+#define SDXC_WIDTH4			1
+#define SDXC_WIDTH8			2
+
+/* smc command bits */
+#define SDXC_RESP_EXPIRE		BIT(6)
+#define SDXC_LONG_RESPONSE		BIT(7)
+#define SDXC_CHECK_RESPONSE_CRC		BIT(8)
+#define SDXC_DATA_EXPIRE		BIT(9)
+#define SDXC_WRITE			BIT(10)
+#define SDXC_SEQUENCE_MODE		BIT(11)
+#define SDXC_SEND_AUTO_STOP		BIT(12)
+#define SDXC_WAIT_PRE_OVER		BIT(13)
+#define SDXC_STOP_ABORT_CMD		BIT(14)
+#define SDXC_SEND_INIT_SEQUENCE		BIT(15)
+#define SDXC_UPCLK_ONLY			BIT(21)
+#define SDXC_READ_CEATA_DEV		BIT(22)
+#define SDXC_CCS_EXPIRE			BIT(23)
+#define SDXC_ENABLE_BIT_BOOT		BIT(24)
+#define SDXC_ALT_BOOT_OPTIONS		BIT(25)
+#define SDXC_BOOT_ACK_EXPIRE		BIT(26)
+#define SDXC_BOOT_ABORT			BIT(27)
+#define SDXC_VOLTAGE_SWITCH	        BIT(28)
+#define SDXC_USE_HOLD_REGISTER	        BIT(29)
+#define SDXC_START			BIT(31)
+
+/* interrupt bits */
+#define SDXC_RESP_ERROR			BIT(1)
+#define SDXC_COMMAND_DONE		BIT(2)
+#define SDXC_DATA_OVER			BIT(3)
+#define SDXC_TX_DATA_REQUEST		BIT(4)
+#define SDXC_RX_DATA_REQUEST		BIT(5)
+#define SDXC_RESP_CRC_ERROR		BIT(6)
+#define SDXC_DATA_CRC_ERROR		BIT(7)
+#define SDXC_RESP_TIMEOUT		BIT(8)
+#define SDXC_DATA_TIMEOUT		BIT(9)
+#define SDXC_VOLTAGE_CHANGE_DONE	BIT(10)
+#define SDXC_FIFO_RUN_ERROR		BIT(11)
+#define SDXC_HARD_WARE_LOCKED		BIT(12)
+#define SDXC_START_BIT_ERROR		BIT(13)
+#define SDXC_AUTO_COMMAND_DONE		BIT(14)
+#define SDXC_END_BIT_ERROR		BIT(15)
+#define SDXC_SDIO_INTERRUPT		BIT(16)
+#define SDXC_CARD_INSERT		BIT(30)
+#define SDXC_CARD_REMOVE		BIT(31)
+#define SDXC_INTERRUPT_ERROR_BIT \
+	(SDXC_RESP_ERROR | SDXC_RESP_CRC_ERROR | SDXC_DATA_CRC_ERROR | \
+	 SDXC_RESP_TIMEOUT | SDXC_DATA_TIMEOUT | SDXC_FIFO_RUN_ERROR | \
+	 SDXC_HARD_WARE_LOCKED | SDXC_START_BIT_ERROR | SDXC_END_BIT_ERROR)
+#define SDXC_INTERRUPT_DONE_BIT \
+	(SDXC_AUTO_COMMAND_DONE | SDXC_DATA_OVER | \
+	 SDXC_COMMAND_DONE | SDXC_VOLTAGE_CHANGE_DONE)
+
+/* status */
+#define SDXC_RXWL_FLAG			BIT(0)
+#define SDXC_TXWL_FLAG			BIT(1)
+#define SDXC_FIFO_EMPTY			BIT(2)
+#define SDXC_FIFO_FULL			BIT(3)
+#define SDXC_CARD_PRESENT		BIT(8)
+#define SDXC_CARD_DATA_BUSY		BIT(9)
+#define SDXC_DATA_FSM_BUSY		BIT(10)
+#define SDXC_DMA_REQUEST		BIT(31)
+#define SDXC_FIFO_SIZE			16
+
+/* Function select */
+#define SDXC_CEATA_ON			(0xceaa << 16)
+#define SDXC_SEND_IRQ_RESPONSE		BIT(0)
+#define SDXC_SDIO_READ_WAIT		BIT(1)
+#define SDXC_ABORT_READ_DATA		BIT(2)
+#define SDXC_SEND_CCSD			BIT(8)
+#define SDXC_SEND_AUTO_STOPCCSD		BIT(9)
+#define SDXC_CEATA_DEV_IRQ_ENABLE	BIT(10)
+
+/* IDMA controller bus mod bit field */
+#define SDXC_IDMAC_SOFT_RESET		BIT(0)
+#define SDXC_IDMAC_FIX_BURST		BIT(1)
+#define SDXC_IDMAC_IDMA_ON		BIT(7)
+#define SDXC_IDMAC_REFETCH_DES		BIT(31)
+
+/* IDMA status bit field */
+#define SDXC_IDMAC_TRANSMIT_INTERRUPT		BIT(0)
+#define SDXC_IDMAC_RECEIVE_INTERRUPT		BIT(1)
+#define SDXC_IDMAC_FATAL_BUS_ERROR		BIT(2)
+#define SDXC_IDMAC_DESTINATION_INVALID		BIT(4)
+#define SDXC_IDMAC_CARD_ERROR_SUM		BIT(5)
+#define SDXC_IDMAC_NORMAL_INTERRUPT_SUM		BIT(8)
+#define SDXC_IDMAC_ABNORMAL_INTERRUPT_SUM	BIT(9)
+#define SDXC_IDMAC_HOST_ABORT_INTERRUPT		BIT(10)
+#define SDXC_IDMAC_IDLE				(0 << 13)
+#define SDXC_IDMAC_SUSPEND			(1 << 13)
+#define SDXC_IDMAC_DESC_READ			(2 << 13)
+#define SDXC_IDMAC_DESC_CHECK			(3 << 13)
+#define SDXC_IDMAC_READ_REQUEST_WAIT		(4 << 13)
+#define SDXC_IDMAC_WRITE_REQUEST_WAIT		(5 << 13)
+#define SDXC_IDMAC_READ				(6 << 13)
+#define SDXC_IDMAC_WRITE			(7 << 13)
+#define SDXC_IDMAC_DESC_CLOSE			(8 << 13)
+
+/*
+* If the idma-des-size-bits of property is ie 13, bufsize bits are:
+*  Bits  0-12: buf1 size
+*  Bits 13-25: buf2 size
+*  Bits 26-31: not used
+* Since we only ever set buf1 size, we can simply store it directly.
+*/
+#define SDXC_IDMAC_DES0_DIC	BIT(1)  /* disable interrupt on completion */
+#define SDXC_IDMAC_DES0_LD	BIT(2)  /* last descriptor */
+#define SDXC_IDMAC_DES0_FD	BIT(3)  /* first descriptor */
+#define SDXC_IDMAC_DES0_CH	BIT(4)  /* chain mode */
+#define SDXC_IDMAC_DES0_ER	BIT(5)  /* end of ring */
+#define SDXC_IDMAC_DES0_CES	BIT(30) /* card error summary */
+#define SDXC_IDMAC_DES0_OWN	BIT(31) /* 1-idma owns it, 0-host owns it */
+
+#define SDXC_CLK_400K		0
+#define SDXC_CLK_25M		1
+#define SDXC_CLK_50M		2
+#define SDXC_CLK_50M_DDR	3
+#define SDXC_CLK_50M_DDR_8BIT	4
+
+#define SDXC_2X_TIMING_MODE	BIT(31)
+
+#define SDXC_CAL_START		BIT(15)
+#define SDXC_CAL_DONE		BIT(14)
+#define SDXC_CAL_DL_SHIFT	8
+#define SDXC_CAL_DL_SW_EN	BIT(7)
+#define SDXC_CAL_DL_SW_SHIFT	0
+#define SDXC_CAL_DL_MASK	0x3f
+
+#define SDXC_CAL_TIMEOUT	3	/* in seconds, 3s is enough*/
+
+struct sunxi_mmc_clk_delay {
+	u32 output;
+	u32 sample;
+};
+
+struct sunxi_idma_des {
+	__le32 config;
+	__le32 buf_size;
+	__le32 buf_addr_ptr1;
+	__le32 buf_addr_ptr2;
+};
+
+struct sunxi_mmc_cfg {
+	u32 idma_des_size_bits;
+	const struct sunxi_mmc_clk_delay *clk_delays;
+
+	/* does the IP block support autocalibration? */
+	bool can_calibrate;
+
+	/* Does DATA0 needs to be masked while the clock is updated */
+	bool mask_data0;
+
+	/*
+	 * hardware only supports new timing mode, either due to lack of
+	 * a mode switch in the clock controller, or the mmc controller
+	 * is permanently configured in the new timing mode, without the
+	 * NTSR mode switch.
+	 */
+	bool needs_new_timings;
+
+	/* clock hardware can switch between old and new timing modes */
+	bool ccu_has_timings_switch;
+};
+
+struct sunxi_mmc_host {
+	struct device *dev;
+	struct mmc_host	*mmc;
+	struct reset_control *reset;
+	const struct sunxi_mmc_cfg *cfg;
+
+	/* IO mapping base */
+	void __iomem	*reg_base;
+
+	/* clock management */
+	struct clk	*clk_ahb;
+	struct clk	*clk_mmc;
+	struct clk	*clk_sample;
+	struct clk	*clk_output;
+
+	/* irq */
+	spinlock_t	lock;
+	int		irq;
+	u32		int_sum;
+	u32		sdio_imask;
+
+	/* dma */
+	dma_addr_t	sg_dma;
+	void		*sg_cpu;
+	bool		wait_dma;
+
+	struct mmc_request *mrq;
+	struct mmc_request *manual_stop_mrq;
+	int		ferror;
+
+	/* vqmmc */
+	bool		vqmmc_enabled;
+
+	/* timings */
+	bool		use_new_timings;
+};
+
+static int sunxi_mmc_reset_host(struct sunxi_mmc_host *host)
+{
+	unsigned long expire = jiffies + msecs_to_jiffies(250);
+	u32 rval;
+
+	mmc_writel(host, REG_GCTRL, SDXC_HARDWARE_RESET);
+	do {
+		rval = mmc_readl(host, REG_GCTRL);
+	} while (time_before(jiffies, expire) && (rval & SDXC_HARDWARE_RESET));
+
+	if (rval & SDXC_HARDWARE_RESET) {
+		dev_err(mmc_dev(host->mmc), "fatal err reset timeout\n");
+		return -EIO;
+	}
+
+	return 0;
+}
+
+static int sunxi_mmc_init_host(struct sunxi_mmc_host *host)
+{
+	u32 rval;
+
+	if (sunxi_mmc_reset_host(host))
+		return -EIO;
+
+	/*
+	 * Burst 8 transfers, RX trigger level: 7, TX trigger level: 8
+	 *
+	 * TODO: sun9i has a larger FIFO and supports higher trigger values
+	 */
+	mmc_writel(host, REG_FTRGL, 0x20070008);
+	/* Maximum timeout value */
+	mmc_writel(host, REG_TMOUT, 0xffffffff);
+	/* Unmask SDIO interrupt if needed */
+	mmc_writel(host, REG_IMASK, host->sdio_imask);
+	/* Clear all pending interrupts */
+	mmc_writel(host, REG_RINTR, 0xffffffff);
+	/* Debug register? undocumented */
+	mmc_writel(host, REG_DBGC, 0xdeb);
+	/* Enable CEATA support */
+	mmc_writel(host, REG_FUNS, SDXC_CEATA_ON);
+	/* Set DMA descriptor list base address */
+	mmc_writel(host, REG_DLBA, host->sg_dma);
+
+	rval = mmc_readl(host, REG_GCTRL);
+	rval |= SDXC_INTERRUPT_ENABLE_BIT;
+	/* Undocumented, but found in Allwinner code */
+	rval &= ~SDXC_ACCESS_DONE_DIRECT;
+	mmc_writel(host, REG_GCTRL, rval);
+
+	return 0;
+}
+
+static void sunxi_mmc_init_idma_des(struct sunxi_mmc_host *host,
+				    struct mmc_data *data)
+{
+	struct sunxi_idma_des *pdes = (struct sunxi_idma_des *)host->sg_cpu;
+	dma_addr_t next_desc = host->sg_dma;
+	int i, max_len = (1 << host->cfg->idma_des_size_bits);
+
+	for (i = 0; i < data->sg_len; i++) {
+		pdes[i].config = cpu_to_le32(SDXC_IDMAC_DES0_CH |
+					     SDXC_IDMAC_DES0_OWN |
+					     SDXC_IDMAC_DES0_DIC);
+
+		if (data->sg[i].length == max_len)
+			pdes[i].buf_size = 0; /* 0 == max_len */
+		else
+			pdes[i].buf_size = cpu_to_le32(data->sg[i].length);
+
+		next_desc += sizeof(struct sunxi_idma_des);
+		pdes[i].buf_addr_ptr1 =
+			cpu_to_le32(sg_dma_address(&data->sg[i]));
+		pdes[i].buf_addr_ptr2 = cpu_to_le32((u32)next_desc);
+	}
+
+	pdes[0].config |= cpu_to_le32(SDXC_IDMAC_DES0_FD);
+	pdes[i - 1].config |= cpu_to_le32(SDXC_IDMAC_DES0_LD |
+					  SDXC_IDMAC_DES0_ER);
+	pdes[i - 1].config &= cpu_to_le32(~SDXC_IDMAC_DES0_DIC);
+	pdes[i - 1].buf_addr_ptr2 = 0;
+
+	/*
+	 * Avoid the io-store starting the idmac hitting io-mem before the
+	 * descriptors hit the main-mem.
+	 */
+	wmb();
+}
+
+static int sunxi_mmc_map_dma(struct sunxi_mmc_host *host,
+			     struct mmc_data *data)
+{
+	u32 i, dma_len;
+	struct scatterlist *sg;
+
+	dma_len = dma_map_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+			     mmc_get_dma_dir(data));
+	if (dma_len == 0) {
+		dev_err(mmc_dev(host->mmc), "dma_map_sg failed\n");
+		return -ENOMEM;
+	}
+
+	for_each_sg(data->sg, sg, data->sg_len, i) {
+		if (sg->offset & 3 || sg->length & 3) {
+			dev_err(mmc_dev(host->mmc),
+				"unaligned scatterlist: os %x length %d\n",
+				sg->offset, sg->length);
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static void sunxi_mmc_start_dma(struct sunxi_mmc_host *host,
+				struct mmc_data *data)
+{
+	u32 rval;
+
+	sunxi_mmc_init_idma_des(host, data);
+
+	rval = mmc_readl(host, REG_GCTRL);
+	rval |= SDXC_DMA_ENABLE_BIT;
+	mmc_writel(host, REG_GCTRL, rval);
+	rval |= SDXC_DMA_RESET;
+	mmc_writel(host, REG_GCTRL, rval);
+
+	mmc_writel(host, REG_DMAC, SDXC_IDMAC_SOFT_RESET);
+
+	if (!(data->flags & MMC_DATA_WRITE))
+		mmc_writel(host, REG_IDIE, SDXC_IDMAC_RECEIVE_INTERRUPT);
+
+	mmc_writel(host, REG_DMAC,
+		   SDXC_IDMAC_FIX_BURST | SDXC_IDMAC_IDMA_ON);
+}
+
+static void sunxi_mmc_send_manual_stop(struct sunxi_mmc_host *host,
+				       struct mmc_request *req)
+{
+	u32 arg, cmd_val, ri;
+	unsigned long expire = jiffies + msecs_to_jiffies(1000);
+
+	cmd_val = SDXC_START | SDXC_RESP_EXPIRE |
+		  SDXC_STOP_ABORT_CMD | SDXC_CHECK_RESPONSE_CRC;
+
+	if (req->cmd->opcode == SD_IO_RW_EXTENDED) {
+		cmd_val |= SD_IO_RW_DIRECT;
+		arg = (1 << 31) | (0 << 28) | (SDIO_CCCR_ABORT << 9) |
+		      ((req->cmd->arg >> 28) & 0x7);
+	} else {
+		cmd_val |= MMC_STOP_TRANSMISSION;
+		arg = 0;
+	}
+
+	mmc_writel(host, REG_CARG, arg);
+	mmc_writel(host, REG_CMDR, cmd_val);
+
+	do {
+		ri = mmc_readl(host, REG_RINTR);
+	} while (!(ri & (SDXC_COMMAND_DONE | SDXC_INTERRUPT_ERROR_BIT)) &&
+		 time_before(jiffies, expire));
+
+	if (!(ri & SDXC_COMMAND_DONE) || (ri & SDXC_INTERRUPT_ERROR_BIT)) {
+		dev_err(mmc_dev(host->mmc), "send stop command failed\n");
+		if (req->stop)
+			req->stop->resp[0] = -ETIMEDOUT;
+	} else {
+		if (req->stop)
+			req->stop->resp[0] = mmc_readl(host, REG_RESP0);
+	}
+
+	mmc_writel(host, REG_RINTR, 0xffff);
+}
+
+static void sunxi_mmc_dump_errinfo(struct sunxi_mmc_host *host)
+{
+	struct mmc_command *cmd = host->mrq->cmd;
+	struct mmc_data *data = host->mrq->data;
+
+	/* For some cmds timeout is normal with sd/mmc cards */
+	if ((host->int_sum & SDXC_INTERRUPT_ERROR_BIT) ==
+		SDXC_RESP_TIMEOUT && (cmd->opcode == SD_IO_SEND_OP_COND ||
+				      cmd->opcode == SD_IO_RW_DIRECT))
+		return;
+
+	dev_dbg(mmc_dev(host->mmc),
+		"smc %d err, cmd %d,%s%s%s%s%s%s%s%s%s%s !!\n",
+		host->mmc->index, cmd->opcode,
+		data ? (data->flags & MMC_DATA_WRITE ? " WR" : " RD") : "",
+		host->int_sum & SDXC_RESP_ERROR     ? " RE"     : "",
+		host->int_sum & SDXC_RESP_CRC_ERROR  ? " RCE"    : "",
+		host->int_sum & SDXC_DATA_CRC_ERROR  ? " DCE"    : "",
+		host->int_sum & SDXC_RESP_TIMEOUT ? " RTO"    : "",
+		host->int_sum & SDXC_DATA_TIMEOUT ? " DTO"    : "",
+		host->int_sum & SDXC_FIFO_RUN_ERROR  ? " FE"     : "",
+		host->int_sum & SDXC_HARD_WARE_LOCKED ? " HL"     : "",
+		host->int_sum & SDXC_START_BIT_ERROR ? " SBE"    : "",
+		host->int_sum & SDXC_END_BIT_ERROR   ? " EBE"    : ""
+		);
+}
+
+/* Called in interrupt context! */
+static irqreturn_t sunxi_mmc_finalize_request(struct sunxi_mmc_host *host)
+{
+	struct mmc_request *mrq = host->mrq;
+	struct mmc_data *data = mrq->data;
+	u32 rval;
+
+	mmc_writel(host, REG_IMASK, host->sdio_imask);
+	mmc_writel(host, REG_IDIE, 0);
+
+	if (host->int_sum & SDXC_INTERRUPT_ERROR_BIT) {
+		sunxi_mmc_dump_errinfo(host);
+		mrq->cmd->error = -ETIMEDOUT;
+
+		if (data) {
+			data->error = -ETIMEDOUT;
+			host->manual_stop_mrq = mrq;
+		}
+
+		if (mrq->stop)
+			mrq->stop->error = -ETIMEDOUT;
+	} else {
+		if (mrq->cmd->flags & MMC_RSP_136) {
+			mrq->cmd->resp[0] = mmc_readl(host, REG_RESP3);
+			mrq->cmd->resp[1] = mmc_readl(host, REG_RESP2);
+			mrq->cmd->resp[2] = mmc_readl(host, REG_RESP1);
+			mrq->cmd->resp[3] = mmc_readl(host, REG_RESP0);
+		} else {
+			mrq->cmd->resp[0] = mmc_readl(host, REG_RESP0);
+		}
+
+		if (data)
+			data->bytes_xfered = data->blocks * data->blksz;
+	}
+
+	if (data) {
+		mmc_writel(host, REG_IDST, 0x337);
+		mmc_writel(host, REG_DMAC, 0);
+		rval = mmc_readl(host, REG_GCTRL);
+		rval |= SDXC_DMA_RESET;
+		mmc_writel(host, REG_GCTRL, rval);
+		rval &= ~SDXC_DMA_ENABLE_BIT;
+		mmc_writel(host, REG_GCTRL, rval);
+		rval |= SDXC_FIFO_RESET;
+		mmc_writel(host, REG_GCTRL, rval);
+		dma_unmap_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+			     mmc_get_dma_dir(data));
+	}
+
+	mmc_writel(host, REG_RINTR, 0xffff);
+
+	host->mrq = NULL;
+	host->int_sum = 0;
+	host->wait_dma = false;
+
+	return host->manual_stop_mrq ? IRQ_WAKE_THREAD : IRQ_HANDLED;
+}
+
+static irqreturn_t sunxi_mmc_irq(int irq, void *dev_id)
+{
+	struct sunxi_mmc_host *host = dev_id;
+	struct mmc_request *mrq;
+	u32 msk_int, idma_int;
+	bool finalize = false;
+	bool sdio_int = false;
+	irqreturn_t ret = IRQ_HANDLED;
+
+	spin_lock(&host->lock);
+
+	idma_int  = mmc_readl(host, REG_IDST);
+	msk_int   = mmc_readl(host, REG_MISTA);
+
+	dev_dbg(mmc_dev(host->mmc), "irq: rq %p mi %08x idi %08x\n",
+		host->mrq, msk_int, idma_int);
+
+	mrq = host->mrq;
+	if (mrq) {
+		if (idma_int & SDXC_IDMAC_RECEIVE_INTERRUPT)
+			host->wait_dma = false;
+
+		host->int_sum |= msk_int;
+
+		/* Wait for COMMAND_DONE on RESPONSE_TIMEOUT before finalize */
+		if ((host->int_sum & SDXC_RESP_TIMEOUT) &&
+				!(host->int_sum & SDXC_COMMAND_DONE))
+			mmc_writel(host, REG_IMASK,
+				   host->sdio_imask | SDXC_COMMAND_DONE);
+		/* Don't wait for dma on error */
+		else if (host->int_sum & SDXC_INTERRUPT_ERROR_BIT)
+			finalize = true;
+		else if ((host->int_sum & SDXC_INTERRUPT_DONE_BIT) &&
+				!host->wait_dma)
+			finalize = true;
+	}
+
+	if (msk_int & SDXC_SDIO_INTERRUPT)
+		sdio_int = true;
+
+	mmc_writel(host, REG_RINTR, msk_int);
+	mmc_writel(host, REG_IDST, idma_int);
+
+	if (finalize)
+		ret = sunxi_mmc_finalize_request(host);
+
+	spin_unlock(&host->lock);
+
+	if (finalize && ret == IRQ_HANDLED)
+		mmc_request_done(host->mmc, mrq);
+
+	if (sdio_int)
+		mmc_signal_sdio_irq(host->mmc);
+
+	return ret;
+}
+
+static irqreturn_t sunxi_mmc_handle_manual_stop(int irq, void *dev_id)
+{
+	struct sunxi_mmc_host *host = dev_id;
+	struct mmc_request *mrq;
+	unsigned long iflags;
+
+	spin_lock_irqsave(&host->lock, iflags);
+	mrq = host->manual_stop_mrq;
+	spin_unlock_irqrestore(&host->lock, iflags);
+
+	if (!mrq) {
+		dev_err(mmc_dev(host->mmc), "no request for manual stop\n");
+		return IRQ_HANDLED;
+	}
+
+	dev_err(mmc_dev(host->mmc), "data error, sending stop command\n");
+
+	/*
+	 * We will never have more than one outstanding request,
+	 * and we do not complete the request until after
+	 * we've cleared host->manual_stop_mrq so we do not need to
+	 * spin lock this function.
+	 * Additionally we have wait states within this function
+	 * so having it in a lock is a very bad idea.
+	 */
+	sunxi_mmc_send_manual_stop(host, mrq);
+
+	spin_lock_irqsave(&host->lock, iflags);
+	host->manual_stop_mrq = NULL;
+	spin_unlock_irqrestore(&host->lock, iflags);
+
+	mmc_request_done(host->mmc, mrq);
+
+	return IRQ_HANDLED;
+}
+
+static int sunxi_mmc_oclk_onoff(struct sunxi_mmc_host *host, u32 oclk_en)
+{
+	unsigned long expire = jiffies + msecs_to_jiffies(750);
+	u32 rval;
+
+	dev_dbg(mmc_dev(host->mmc), "%sabling the clock\n",
+		oclk_en ? "en" : "dis");
+
+	rval = mmc_readl(host, REG_CLKCR);
+	rval &= ~(SDXC_CARD_CLOCK_ON | SDXC_LOW_POWER_ON | SDXC_MASK_DATA0);
+
+	if (oclk_en)
+		rval |= SDXC_CARD_CLOCK_ON;
+	if (host->cfg->mask_data0)
+		rval |= SDXC_MASK_DATA0;
+
+	mmc_writel(host, REG_CLKCR, rval);
+
+	rval = SDXC_START | SDXC_UPCLK_ONLY | SDXC_WAIT_PRE_OVER;
+	mmc_writel(host, REG_CMDR, rval);
+
+	do {
+		rval = mmc_readl(host, REG_CMDR);
+	} while (time_before(jiffies, expire) && (rval & SDXC_START));
+
+	/* clear irq status bits set by the command */
+	mmc_writel(host, REG_RINTR,
+		   mmc_readl(host, REG_RINTR) & ~SDXC_SDIO_INTERRUPT);
+
+	if (rval & SDXC_START) {
+		dev_err(mmc_dev(host->mmc), "fatal err update clk timeout\n");
+		return -EIO;
+	}
+
+	if (host->cfg->mask_data0) {
+		rval = mmc_readl(host, REG_CLKCR);
+		mmc_writel(host, REG_CLKCR, rval & ~SDXC_MASK_DATA0);
+	}
+
+	return 0;
+}
+
+static int sunxi_mmc_calibrate(struct sunxi_mmc_host *host, int reg_off)
+{
+	if (!host->cfg->can_calibrate)
+		return 0;
+
+	/*
+	 * FIXME:
+	 * This is not clear how the calibration is supposed to work
+	 * yet. The best rate have been obtained by simply setting the
+	 * delay to 0, as Allwinner does in its BSP.
+	 *
+	 * The only mode that doesn't have such a delay is HS400, that
+	 * is in itself a TODO.
+	 */
+	writel(SDXC_CAL_DL_SW_EN, host->reg_base + reg_off);
+
+	return 0;
+}
+
+static int sunxi_mmc_clk_set_phase(struct sunxi_mmc_host *host,
+				   struct mmc_ios *ios, u32 rate)
+{
+	int index;
+
+	/* clk controller delays not used under new timings mode */
+	if (host->use_new_timings)
+		return 0;
+
+	/* some old controllers don't support delays */
+	if (!host->cfg->clk_delays)
+		return 0;
+
+	/* determine delays */
+	if (rate <= 400000) {
+		index = SDXC_CLK_400K;
+	} else if (rate <= 25000000) {
+		index = SDXC_CLK_25M;
+	} else if (rate <= 52000000) {
+		if (ios->timing != MMC_TIMING_UHS_DDR50 &&
+		    ios->timing != MMC_TIMING_MMC_DDR52) {
+			index = SDXC_CLK_50M;
+		} else if (ios->bus_width == MMC_BUS_WIDTH_8) {
+			index = SDXC_CLK_50M_DDR_8BIT;
+		} else {
+			index = SDXC_CLK_50M_DDR;
+		}
+	} else {
+		dev_dbg(mmc_dev(host->mmc), "Invalid clock... returning\n");
+		return -EINVAL;
+	}
+
+	clk_set_phase(host->clk_sample, host->cfg->clk_delays[index].sample);
+	clk_set_phase(host->clk_output, host->cfg->clk_delays[index].output);
+
+	return 0;
+}
+
+static int sunxi_mmc_clk_set_rate(struct sunxi_mmc_host *host,
+				  struct mmc_ios *ios)
+{
+	struct mmc_host *mmc = host->mmc;
+	long rate;
+	u32 rval, clock = ios->clock, div = 1;
+	int ret;
+
+	ret = sunxi_mmc_oclk_onoff(host, 0);
+	if (ret)
+		return ret;
+
+	/* Our clock is gated now */
+	mmc->actual_clock = 0;
+
+	if (!ios->clock)
+		return 0;
+
+	/*
+	 * Under the old timing mode, 8 bit DDR requires the module
+	 * clock to be double the card clock. Under the new timing
+	 * mode, all DDR modes require a doubled module clock.
+	 *
+	 * We currently only support the standard MMC DDR52 mode.
+	 * This block should be updated once support for other DDR
+	 * modes is added.
+	 */
+	if (ios->timing == MMC_TIMING_MMC_DDR52 &&
+	    (host->use_new_timings ||
+	     ios->bus_width == MMC_BUS_WIDTH_8)) {
+		div = 2;
+		clock <<= 1;
+	}
+
+	if (host->use_new_timings && host->cfg->ccu_has_timings_switch) {
+		ret = sunxi_ccu_set_mmc_timing_mode(host->clk_mmc, true);
+		if (ret) {
+			dev_err(mmc_dev(mmc),
+				"error setting new timing mode\n");
+			return ret;
+		}
+	}
+
+	rate = clk_round_rate(host->clk_mmc, clock);
+	if (rate < 0) {
+		dev_err(mmc_dev(mmc), "error rounding clk to %d: %ld\n",
+			clock, rate);
+		return rate;
+	}
+	dev_dbg(mmc_dev(mmc), "setting clk to %d, rounded %ld\n",
+		clock, rate);
+
+	/* setting clock rate */
+	ret = clk_set_rate(host->clk_mmc, rate);
+	if (ret) {
+		dev_err(mmc_dev(mmc), "error setting clk to %ld: %d\n",
+			rate, ret);
+		return ret;
+	}
+
+	/* set internal divider */
+	rval = mmc_readl(host, REG_CLKCR);
+	rval &= ~0xff;
+	rval |= div - 1;
+	mmc_writel(host, REG_CLKCR, rval);
+
+	/* update card clock rate to account for internal divider */
+	rate /= div;
+
+	/*
+	 * Configure the controller to use the new timing mode if needed.
+	 * On controllers that only support the new timing mode, such as
+	 * the eMMC controller on the A64, this register does not exist,
+	 * and any writes to it are ignored.
+	 */
+	if (host->use_new_timings) {
+		/* Don't touch the delay bits */
+		rval = mmc_readl(host, REG_SD_NTSR);
+		rval |= SDXC_2X_TIMING_MODE;
+		mmc_writel(host, REG_SD_NTSR, rval);
+	}
+
+	/* sunxi_mmc_clk_set_phase expects the actual card clock rate */
+	ret = sunxi_mmc_clk_set_phase(host, ios, rate);
+	if (ret)
+		return ret;
+
+	ret = sunxi_mmc_calibrate(host, SDXC_REG_SAMP_DL_REG);
+	if (ret)
+		return ret;
+
+	/*
+	 * FIXME:
+	 *
+	 * In HS400 we'll also need to calibrate the data strobe
+	 * signal. This should only happen on the MMC2 controller (at
+	 * least on the A64).
+	 */
+
+	ret = sunxi_mmc_oclk_onoff(host, 1);
+	if (ret)
+		return ret;
+
+	/* And we just enabled our clock back */
+	mmc->actual_clock = rate;
+
+	return 0;
+}
+
+static void sunxi_mmc_set_bus_width(struct sunxi_mmc_host *host,
+				   unsigned char width)
+{
+	switch (width) {
+	case MMC_BUS_WIDTH_1:
+		mmc_writel(host, REG_WIDTH, SDXC_WIDTH1);
+		break;
+	case MMC_BUS_WIDTH_4:
+		mmc_writel(host, REG_WIDTH, SDXC_WIDTH4);
+		break;
+	case MMC_BUS_WIDTH_8:
+		mmc_writel(host, REG_WIDTH, SDXC_WIDTH8);
+		break;
+	}
+}
+
+static void sunxi_mmc_set_clk(struct sunxi_mmc_host *host, struct mmc_ios *ios)
+{
+	u32 rval;
+
+	/* set ddr mode */
+	rval = mmc_readl(host, REG_GCTRL);
+	if (ios->timing == MMC_TIMING_UHS_DDR50 ||
+	    ios->timing == MMC_TIMING_MMC_DDR52)
+		rval |= SDXC_DDR_MODE;
+	else
+		rval &= ~SDXC_DDR_MODE;
+	mmc_writel(host, REG_GCTRL, rval);
+
+	host->ferror = sunxi_mmc_clk_set_rate(host, ios);
+	/* Android code had a usleep_range(50000, 55000); here */
+}
+
+static void sunxi_mmc_card_power(struct sunxi_mmc_host *host,
+				 struct mmc_ios *ios)
+{
+	struct mmc_host *mmc = host->mmc;
+
+	switch (ios->power_mode) {
+	case MMC_POWER_UP:
+		dev_dbg(mmc_dev(mmc), "Powering card up\n");
+
+		if (!IS_ERR(mmc->supply.vmmc)) {
+			host->ferror = mmc_regulator_set_ocr(mmc,
+							     mmc->supply.vmmc,
+							     ios->vdd);
+			if (host->ferror)
+				return;
+		}
+
+		if (!IS_ERR(mmc->supply.vqmmc)) {
+			host->ferror = regulator_enable(mmc->supply.vqmmc);
+			if (host->ferror) {
+				dev_err(mmc_dev(mmc),
+					"failed to enable vqmmc\n");
+				return;
+			}
+			host->vqmmc_enabled = true;
+		}
+		break;
+
+	case MMC_POWER_OFF:
+		dev_dbg(mmc_dev(mmc), "Powering card off\n");
+
+		if (!IS_ERR(mmc->supply.vmmc))
+			mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+
+		if (!IS_ERR(mmc->supply.vqmmc) && host->vqmmc_enabled)
+			regulator_disable(mmc->supply.vqmmc);
+
+		host->vqmmc_enabled = false;
+		break;
+
+	default:
+		dev_dbg(mmc_dev(mmc), "Ignoring unknown card power state\n");
+		break;
+	}
+}
+
+static void sunxi_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sunxi_mmc_host *host = mmc_priv(mmc);
+
+	sunxi_mmc_card_power(host, ios);
+	sunxi_mmc_set_bus_width(host, ios->bus_width);
+	sunxi_mmc_set_clk(host, ios);
+}
+
+static int sunxi_mmc_volt_switch(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	int ret;
+
+	/* vqmmc regulator is available */
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+		ret = mmc_regulator_set_vqmmc(mmc, ios);
+		return ret < 0 ? ret : 0;
+	}
+
+	/* no vqmmc regulator, assume fixed regulator at 3/3.3V */
+	if (mmc->ios.signal_voltage == MMC_SIGNAL_VOLTAGE_330)
+		return 0;
+
+	return -EINVAL;
+}
+
+static void sunxi_mmc_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct sunxi_mmc_host *host = mmc_priv(mmc);
+	unsigned long flags;
+	u32 imask;
+
+	if (enable)
+		pm_runtime_get_noresume(host->dev);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	imask = mmc_readl(host, REG_IMASK);
+	if (enable) {
+		host->sdio_imask = SDXC_SDIO_INTERRUPT;
+		imask |= SDXC_SDIO_INTERRUPT;
+	} else {
+		host->sdio_imask = 0;
+		imask &= ~SDXC_SDIO_INTERRUPT;
+	}
+	mmc_writel(host, REG_IMASK, imask);
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (!enable)
+		pm_runtime_put_noidle(host->mmc->parent);
+}
+
+static void sunxi_mmc_hw_reset(struct mmc_host *mmc)
+{
+	struct sunxi_mmc_host *host = mmc_priv(mmc);
+	mmc_writel(host, REG_HWRST, 0);
+	udelay(10);
+	mmc_writel(host, REG_HWRST, 1);
+	udelay(300);
+}
+
+static void sunxi_mmc_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct sunxi_mmc_host *host = mmc_priv(mmc);
+	struct mmc_command *cmd = mrq->cmd;
+	struct mmc_data *data = mrq->data;
+	unsigned long iflags;
+	u32 imask = SDXC_INTERRUPT_ERROR_BIT;
+	u32 cmd_val = SDXC_START | (cmd->opcode & 0x3f);
+	bool wait_dma = host->wait_dma;
+	int ret;
+
+	/* Check for set_ios errors (should never happen) */
+	if (host->ferror) {
+		mrq->cmd->error = host->ferror;
+		mmc_request_done(mmc, mrq);
+		return;
+	}
+
+	if (data) {
+		ret = sunxi_mmc_map_dma(host, data);
+		if (ret < 0) {
+			dev_err(mmc_dev(mmc), "map DMA failed\n");
+			cmd->error = ret;
+			data->error = ret;
+			mmc_request_done(mmc, mrq);
+			return;
+		}
+	}
+
+	if (cmd->opcode == MMC_GO_IDLE_STATE) {
+		cmd_val |= SDXC_SEND_INIT_SEQUENCE;
+		imask |= SDXC_COMMAND_DONE;
+	}
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		cmd_val |= SDXC_RESP_EXPIRE;
+		if (cmd->flags & MMC_RSP_136)
+			cmd_val |= SDXC_LONG_RESPONSE;
+		if (cmd->flags & MMC_RSP_CRC)
+			cmd_val |= SDXC_CHECK_RESPONSE_CRC;
+
+		if ((cmd->flags & MMC_CMD_MASK) == MMC_CMD_ADTC) {
+			cmd_val |= SDXC_DATA_EXPIRE | SDXC_WAIT_PRE_OVER;
+
+			if (cmd->data->stop) {
+				imask |= SDXC_AUTO_COMMAND_DONE;
+				cmd_val |= SDXC_SEND_AUTO_STOP;
+			} else {
+				imask |= SDXC_DATA_OVER;
+			}
+
+			if (cmd->data->flags & MMC_DATA_WRITE)
+				cmd_val |= SDXC_WRITE;
+			else
+				wait_dma = true;
+		} else {
+			imask |= SDXC_COMMAND_DONE;
+		}
+	} else {
+		imask |= SDXC_COMMAND_DONE;
+	}
+
+	dev_dbg(mmc_dev(mmc), "cmd %d(%08x) arg %x ie 0x%08x len %d\n",
+		cmd_val & 0x3f, cmd_val, cmd->arg, imask,
+		mrq->data ? mrq->data->blksz * mrq->data->blocks : 0);
+
+	spin_lock_irqsave(&host->lock, iflags);
+
+	if (host->mrq || host->manual_stop_mrq) {
+		spin_unlock_irqrestore(&host->lock, iflags);
+
+		if (data)
+			dma_unmap_sg(mmc_dev(mmc), data->sg, data->sg_len,
+				     mmc_get_dma_dir(data));
+
+		dev_err(mmc_dev(mmc), "request already pending\n");
+		mrq->cmd->error = -EBUSY;
+		mmc_request_done(mmc, mrq);
+		return;
+	}
+
+	if (data) {
+		mmc_writel(host, REG_BLKSZ, data->blksz);
+		mmc_writel(host, REG_BCNTR, data->blksz * data->blocks);
+		sunxi_mmc_start_dma(host, data);
+	}
+
+	host->mrq = mrq;
+	host->wait_dma = wait_dma;
+	mmc_writel(host, REG_IMASK, host->sdio_imask | imask);
+	mmc_writel(host, REG_CARG, cmd->arg);
+	mmc_writel(host, REG_CMDR, cmd_val);
+
+	spin_unlock_irqrestore(&host->lock, iflags);
+}
+
+static int sunxi_mmc_card_busy(struct mmc_host *mmc)
+{
+	struct sunxi_mmc_host *host = mmc_priv(mmc);
+
+	return !!(mmc_readl(host, REG_STAS) & SDXC_CARD_DATA_BUSY);
+}
+
+static const struct mmc_host_ops sunxi_mmc_ops = {
+	.request	 = sunxi_mmc_request,
+	.set_ios	 = sunxi_mmc_set_ios,
+	.get_ro		 = mmc_gpio_get_ro,
+	.get_cd		 = mmc_gpio_get_cd,
+	.enable_sdio_irq = sunxi_mmc_enable_sdio_irq,
+	.start_signal_voltage_switch = sunxi_mmc_volt_switch,
+	.hw_reset	 = sunxi_mmc_hw_reset,
+	.card_busy	 = sunxi_mmc_card_busy,
+};
+
+static const struct sunxi_mmc_clk_delay sunxi_mmc_clk_delays[] = {
+	[SDXC_CLK_400K]		= { .output = 180, .sample = 180 },
+	[SDXC_CLK_25M]		= { .output = 180, .sample =  75 },
+	[SDXC_CLK_50M]		= { .output =  90, .sample = 120 },
+	[SDXC_CLK_50M_DDR]	= { .output =  60, .sample = 120 },
+	/* Value from A83T "new timing mode". Works but might not be right. */
+	[SDXC_CLK_50M_DDR_8BIT]	= { .output =  90, .sample = 180 },
+};
+
+static const struct sunxi_mmc_clk_delay sun9i_mmc_clk_delays[] = {
+	[SDXC_CLK_400K]		= { .output = 180, .sample = 180 },
+	[SDXC_CLK_25M]		= { .output = 180, .sample =  75 },
+	[SDXC_CLK_50M]		= { .output = 150, .sample = 120 },
+	[SDXC_CLK_50M_DDR]	= { .output =  54, .sample =  36 },
+	[SDXC_CLK_50M_DDR_8BIT]	= { .output =  72, .sample =  72 },
+};
+
+static const struct sunxi_mmc_cfg sun4i_a10_cfg = {
+	.idma_des_size_bits = 13,
+	.clk_delays = NULL,
+	.can_calibrate = false,
+};
+
+static const struct sunxi_mmc_cfg sun5i_a13_cfg = {
+	.idma_des_size_bits = 16,
+	.clk_delays = NULL,
+	.can_calibrate = false,
+};
+
+static const struct sunxi_mmc_cfg sun7i_a20_cfg = {
+	.idma_des_size_bits = 16,
+	.clk_delays = sunxi_mmc_clk_delays,
+	.can_calibrate = false,
+};
+
+static const struct sunxi_mmc_cfg sun8i_a83t_emmc_cfg = {
+	.idma_des_size_bits = 16,
+	.clk_delays = sunxi_mmc_clk_delays,
+	.can_calibrate = false,
+	.ccu_has_timings_switch = true,
+};
+
+static const struct sunxi_mmc_cfg sun9i_a80_cfg = {
+	.idma_des_size_bits = 16,
+	.clk_delays = sun9i_mmc_clk_delays,
+	.can_calibrate = false,
+};
+
+static const struct sunxi_mmc_cfg sun50i_a64_cfg = {
+	.idma_des_size_bits = 16,
+	.clk_delays = NULL,
+	.can_calibrate = true,
+	.mask_data0 = true,
+	.needs_new_timings = true,
+};
+
+static const struct sunxi_mmc_cfg sun50i_a64_emmc_cfg = {
+	.idma_des_size_bits = 13,
+	.clk_delays = NULL,
+	.can_calibrate = true,
+	.needs_new_timings = true,
+};
+
+static const struct of_device_id sunxi_mmc_of_match[] = {
+	{ .compatible = "allwinner,sun4i-a10-mmc", .data = &sun4i_a10_cfg },
+	{ .compatible = "allwinner,sun5i-a13-mmc", .data = &sun5i_a13_cfg },
+	{ .compatible = "allwinner,sun7i-a20-mmc", .data = &sun7i_a20_cfg },
+	{ .compatible = "allwinner,sun8i-a83t-emmc", .data = &sun8i_a83t_emmc_cfg },
+	{ .compatible = "allwinner,sun9i-a80-mmc", .data = &sun9i_a80_cfg },
+	{ .compatible = "allwinner,sun50i-a64-mmc", .data = &sun50i_a64_cfg },
+	{ .compatible = "allwinner,sun50i-a64-emmc", .data = &sun50i_a64_emmc_cfg },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, sunxi_mmc_of_match);
+
+static int sunxi_mmc_enable(struct sunxi_mmc_host *host)
+{
+	int ret;
+
+	if (!IS_ERR(host->reset)) {
+		ret = reset_control_reset(host->reset);
+		if (ret) {
+			dev_err(host->dev, "Couldn't reset the MMC controller (%d)\n",
+				ret);
+			return ret;
+		}
+	}
+
+	ret = clk_prepare_enable(host->clk_ahb);
+	if (ret) {
+		dev_err(host->dev, "Couldn't enable the bus clocks (%d)\n", ret);
+		goto error_assert_reset;
+	}
+
+	ret = clk_prepare_enable(host->clk_mmc);
+	if (ret) {
+		dev_err(host->dev, "Enable mmc clk err %d\n", ret);
+		goto error_disable_clk_ahb;
+	}
+
+	ret = clk_prepare_enable(host->clk_output);
+	if (ret) {
+		dev_err(host->dev, "Enable output clk err %d\n", ret);
+		goto error_disable_clk_mmc;
+	}
+
+	ret = clk_prepare_enable(host->clk_sample);
+	if (ret) {
+		dev_err(host->dev, "Enable sample clk err %d\n", ret);
+		goto error_disable_clk_output;
+	}
+
+	/*
+	 * Sometimes the controller asserts the irq on boot for some reason,
+	 * make sure the controller is in a sane state before enabling irqs.
+	 */
+	ret = sunxi_mmc_reset_host(host);
+	if (ret)
+		goto error_disable_clk_sample;
+
+	return 0;
+
+error_disable_clk_sample:
+	clk_disable_unprepare(host->clk_sample);
+error_disable_clk_output:
+	clk_disable_unprepare(host->clk_output);
+error_disable_clk_mmc:
+	clk_disable_unprepare(host->clk_mmc);
+error_disable_clk_ahb:
+	clk_disable_unprepare(host->clk_ahb);
+error_assert_reset:
+	if (!IS_ERR(host->reset))
+		reset_control_assert(host->reset);
+	return ret;
+}
+
+static void sunxi_mmc_disable(struct sunxi_mmc_host *host)
+{
+	sunxi_mmc_reset_host(host);
+
+	clk_disable_unprepare(host->clk_sample);
+	clk_disable_unprepare(host->clk_output);
+	clk_disable_unprepare(host->clk_mmc);
+	clk_disable_unprepare(host->clk_ahb);
+
+	if (!IS_ERR(host->reset))
+		reset_control_assert(host->reset);
+}
+
+static int sunxi_mmc_resource_request(struct sunxi_mmc_host *host,
+				      struct platform_device *pdev)
+{
+	int ret;
+
+	host->cfg = of_device_get_match_data(&pdev->dev);
+	if (!host->cfg)
+		return -EINVAL;
+
+	ret = mmc_regulator_get_supply(host->mmc);
+	if (ret)
+		return ret;
+
+	host->reg_base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(host->reg_base))
+		return PTR_ERR(host->reg_base);
+
+	host->clk_ahb = devm_clk_get(&pdev->dev, "ahb");
+	if (IS_ERR(host->clk_ahb)) {
+		dev_err(&pdev->dev, "Could not get ahb clock\n");
+		return PTR_ERR(host->clk_ahb);
+	}
+
+	host->clk_mmc = devm_clk_get(&pdev->dev, "mmc");
+	if (IS_ERR(host->clk_mmc)) {
+		dev_err(&pdev->dev, "Could not get mmc clock\n");
+		return PTR_ERR(host->clk_mmc);
+	}
+
+	if (host->cfg->clk_delays) {
+		host->clk_output = devm_clk_get(&pdev->dev, "output");
+		if (IS_ERR(host->clk_output)) {
+			dev_err(&pdev->dev, "Could not get output clock\n");
+			return PTR_ERR(host->clk_output);
+		}
+
+		host->clk_sample = devm_clk_get(&pdev->dev, "sample");
+		if (IS_ERR(host->clk_sample)) {
+			dev_err(&pdev->dev, "Could not get sample clock\n");
+			return PTR_ERR(host->clk_sample);
+		}
+	}
+
+	host->reset = devm_reset_control_get_optional_exclusive(&pdev->dev,
+								"ahb");
+	if (PTR_ERR(host->reset) == -EPROBE_DEFER)
+		return PTR_ERR(host->reset);
+
+	ret = sunxi_mmc_enable(host);
+	if (ret)
+		return ret;
+
+	host->irq = platform_get_irq(pdev, 0);
+	if (host->irq < 0) {
+		ret = host->irq;
+		goto error_disable_mmc;
+	}
+
+	return devm_request_threaded_irq(&pdev->dev, host->irq, sunxi_mmc_irq,
+			sunxi_mmc_handle_manual_stop, 0, "sunxi-mmc", host);
+
+error_disable_mmc:
+	sunxi_mmc_disable(host);
+	return ret;
+}
+
+static int sunxi_mmc_probe(struct platform_device *pdev)
+{
+	struct sunxi_mmc_host *host;
+	struct mmc_host *mmc;
+	int ret;
+
+	mmc = mmc_alloc_host(sizeof(struct sunxi_mmc_host), &pdev->dev);
+	if (!mmc) {
+		dev_err(&pdev->dev, "mmc alloc host failed\n");
+		return -ENOMEM;
+	}
+	platform_set_drvdata(pdev, mmc);
+
+	host = mmc_priv(mmc);
+	host->dev = &pdev->dev;
+	host->mmc = mmc;
+	spin_lock_init(&host->lock);
+
+	ret = sunxi_mmc_resource_request(host, pdev);
+	if (ret)
+		goto error_free_host;
+
+	host->sg_cpu = dma_alloc_coherent(&pdev->dev, PAGE_SIZE,
+					  &host->sg_dma, GFP_KERNEL);
+	if (!host->sg_cpu) {
+		dev_err(&pdev->dev, "Failed to allocate DMA descriptor mem\n");
+		ret = -ENOMEM;
+		goto error_free_host;
+	}
+
+	if (host->cfg->ccu_has_timings_switch) {
+		/*
+		 * Supports both old and new timing modes.
+		 * Try setting the clk to new timing mode.
+		 */
+		sunxi_ccu_set_mmc_timing_mode(host->clk_mmc, true);
+
+		/* And check the result */
+		ret = sunxi_ccu_get_mmc_timing_mode(host->clk_mmc);
+		if (ret < 0) {
+			/*
+			 * For whatever reason we were not able to get
+			 * the current active mode. Default to old mode.
+			 */
+			dev_warn(&pdev->dev, "MMC clk timing mode unknown\n");
+			host->use_new_timings = false;
+		} else {
+			host->use_new_timings = !!ret;
+		}
+	} else if (host->cfg->needs_new_timings) {
+		/* Supports new timing mode only */
+		host->use_new_timings = true;
+	}
+
+	mmc->ops		= &sunxi_mmc_ops;
+	mmc->max_blk_count	= 8192;
+	mmc->max_blk_size	= 4096;
+	mmc->max_segs		= PAGE_SIZE / sizeof(struct sunxi_idma_des);
+	mmc->max_seg_size	= (1 << host->cfg->idma_des_size_bits);
+	mmc->max_req_size	= mmc->max_seg_size * mmc->max_segs;
+	/* 400kHz ~ 52MHz */
+	mmc->f_min		=   400000;
+	mmc->f_max		= 52000000;
+	mmc->caps	       |= MMC_CAP_MMC_HIGHSPEED | MMC_CAP_SD_HIGHSPEED |
+				  MMC_CAP_SDIO_IRQ;
+
+	/*
+	 * Some H5 devices do not have signal traces precise enough to
+	 * use HS DDR mode for their eMMC chips.
+	 *
+	 * We still enable HS DDR modes for all the other controller
+	 * variants that support them.
+	 */
+	if ((host->cfg->clk_delays || host->use_new_timings) &&
+	    !of_device_is_compatible(pdev->dev.of_node,
+				     "allwinner,sun50i-h5-emmc"))
+		mmc->caps      |= MMC_CAP_1_8V_DDR | MMC_CAP_3_3V_DDR;
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto error_free_dma;
+
+	/*
+	 * If we don't support delay chains in the SoC, we can't use any
+	 * of the higher speed modes. Mask them out in case the device
+	 * tree specifies the properties for them, which gets added to
+	 * the caps by mmc_of_parse() above.
+	 */
+	if (!(host->cfg->clk_delays || host->use_new_timings)) {
+		mmc->caps &= ~(MMC_CAP_3_3V_DDR | MMC_CAP_1_8V_DDR |
+			       MMC_CAP_1_2V_DDR | MMC_CAP_UHS);
+		mmc->caps2 &= ~MMC_CAP2_HS200;
+	}
+
+	/* TODO: This driver doesn't support HS400 mode yet */
+	mmc->caps2 &= ~MMC_CAP2_HS400;
+
+	ret = sunxi_mmc_init_host(host);
+	if (ret)
+		goto error_free_dma;
+
+	pm_runtime_set_active(&pdev->dev);
+	pm_runtime_set_autosuspend_delay(&pdev->dev, 50);
+	pm_runtime_use_autosuspend(&pdev->dev);
+	pm_runtime_enable(&pdev->dev);
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto error_free_dma;
+
+	dev_info(&pdev->dev, "initialized, max. request size: %u KB%s\n",
+		 mmc->max_req_size >> 10,
+		 host->use_new_timings ? ", uses new timings mode" : "");
+
+	return 0;
+
+error_free_dma:
+	dma_free_coherent(&pdev->dev, PAGE_SIZE, host->sg_cpu, host->sg_dma);
+error_free_host:
+	mmc_free_host(mmc);
+	return ret;
+}
+
+static int sunxi_mmc_remove(struct platform_device *pdev)
+{
+	struct mmc_host	*mmc = platform_get_drvdata(pdev);
+	struct sunxi_mmc_host *host = mmc_priv(mmc);
+
+	mmc_remove_host(mmc);
+	pm_runtime_disable(&pdev->dev);
+	if (!pm_runtime_status_suspended(&pdev->dev)) {
+		disable_irq(host->irq);
+		sunxi_mmc_disable(host);
+	}
+	dma_free_coherent(&pdev->dev, PAGE_SIZE, host->sg_cpu, host->sg_dma);
+	mmc_free_host(mmc);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static int sunxi_mmc_runtime_resume(struct device *dev)
+{
+	struct mmc_host	*mmc = dev_get_drvdata(dev);
+	struct sunxi_mmc_host *host = mmc_priv(mmc);
+	int ret;
+
+	ret = sunxi_mmc_enable(host);
+	if (ret)
+		return ret;
+
+	sunxi_mmc_init_host(host);
+	sunxi_mmc_set_bus_width(host, mmc->ios.bus_width);
+	sunxi_mmc_set_clk(host, &mmc->ios);
+	enable_irq(host->irq);
+
+	return 0;
+}
+
+static int sunxi_mmc_runtime_suspend(struct device *dev)
+{
+	struct mmc_host	*mmc = dev_get_drvdata(dev);
+	struct sunxi_mmc_host *host = mmc_priv(mmc);
+
+	/*
+	 * When clocks are off, it's possible receiving
+	 * fake interrupts, which will stall the system.
+	 * Disabling the irq  will prevent this.
+	 */
+	disable_irq(host->irq);
+	sunxi_mmc_reset_host(host);
+	sunxi_mmc_disable(host);
+
+	return 0;
+}
+#endif
+
+static const struct dev_pm_ops sunxi_mmc_pm_ops = {
+	SET_RUNTIME_PM_OPS(sunxi_mmc_runtime_suspend,
+			   sunxi_mmc_runtime_resume,
+			   NULL)
+};
+
+static struct platform_driver sunxi_mmc_driver = {
+	.driver = {
+		.name	= "sunxi-mmc",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = of_match_ptr(sunxi_mmc_of_match),
+		.pm = &sunxi_mmc_pm_ops,
+	},
+	.probe		= sunxi_mmc_probe,
+	.remove		= sunxi_mmc_remove,
+};
+module_platform_driver(sunxi_mmc_driver);
+
+MODULE_DESCRIPTION("Allwinner's SD/MMC Card Controller Driver");
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("David Lanzendörfer <david.lanzendoerfer@o2s.ch>");
+MODULE_ALIAS("platform:sunxi-mmc");
diff --git a/drivers/mmc/host/tifm_sd.c b/drivers/mmc/host/tifm_sd.c
new file mode 100644
index 000000000..fd8b72d3e
--- /dev/null
+++ b/drivers/mmc/host/tifm_sd.c
@@ -0,0 +1,1092 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ *  tifm_sd.c - TI FlashMedia driver
+ *
+ *  Copyright (C) 2006 Alex Dubov <oakad@yahoo.com>
+ *
+ * Special thanks to Brad Campbell for extensive testing of this driver.
+ */
+
+
+#include <linux/tifm.h>
+#include <linux/mmc/host.h>
+#include <linux/highmem.h>
+#include <linux/scatterlist.h>
+#include <linux/module.h>
+#include <asm/io.h>
+
+#define DRIVER_NAME "tifm_sd"
+#define DRIVER_VERSION "0.8"
+
+static bool no_dma = 0;
+static bool fixed_timeout = 0;
+module_param(no_dma, bool, 0644);
+module_param(fixed_timeout, bool, 0644);
+
+/* Constants here are mostly from OMAP5912 datasheet */
+#define TIFM_MMCSD_RESET      0x0002
+#define TIFM_MMCSD_CLKMASK    0x03ff
+#define TIFM_MMCSD_POWER      0x0800
+#define TIFM_MMCSD_4BBUS      0x8000
+#define TIFM_MMCSD_RXDE       0x8000   /* rx dma enable */
+#define TIFM_MMCSD_TXDE       0x0080   /* tx dma enable */
+#define TIFM_MMCSD_BUFINT     0x0c00   /* set bits: AE, AF */
+#define TIFM_MMCSD_DPE        0x0020   /* data timeout counted in kilocycles */
+#define TIFM_MMCSD_INAB       0x0080   /* abort / initialize command */
+#define TIFM_MMCSD_READ       0x8000
+
+#define TIFM_MMCSD_ERRMASK    0x01e0   /* set bits: CCRC, CTO, DCRC, DTO */
+#define TIFM_MMCSD_EOC        0x0001   /* end of command phase  */
+#define TIFM_MMCSD_CD         0x0002   /* card detect           */
+#define TIFM_MMCSD_CB         0x0004   /* card enter busy state */
+#define TIFM_MMCSD_BRS        0x0008   /* block received/sent   */
+#define TIFM_MMCSD_EOFB       0x0010   /* card exit busy state  */
+#define TIFM_MMCSD_DTO        0x0020   /* data time-out         */
+#define TIFM_MMCSD_DCRC       0x0040   /* data crc error        */
+#define TIFM_MMCSD_CTO        0x0080   /* command time-out      */
+#define TIFM_MMCSD_CCRC       0x0100   /* command crc error     */
+#define TIFM_MMCSD_AF         0x0400   /* fifo almost full      */
+#define TIFM_MMCSD_AE         0x0800   /* fifo almost empty     */
+#define TIFM_MMCSD_OCRB       0x1000   /* OCR busy              */
+#define TIFM_MMCSD_CIRQ       0x2000   /* card irq (cmd40/sdio) */
+#define TIFM_MMCSD_CERR       0x4000   /* card status error     */
+
+#define TIFM_MMCSD_ODTO       0x0040   /* open drain / extended timeout */
+#define TIFM_MMCSD_CARD_RO    0x0200   /* card is read-only     */
+
+#define TIFM_MMCSD_FIFO_SIZE  0x0020
+
+#define TIFM_MMCSD_RSP_R0     0x0000
+#define TIFM_MMCSD_RSP_R1     0x0100
+#define TIFM_MMCSD_RSP_R2     0x0200
+#define TIFM_MMCSD_RSP_R3     0x0300
+#define TIFM_MMCSD_RSP_R4     0x0400
+#define TIFM_MMCSD_RSP_R5     0x0500
+#define TIFM_MMCSD_RSP_R6     0x0600
+
+#define TIFM_MMCSD_RSP_BUSY   0x0800
+
+#define TIFM_MMCSD_CMD_BC     0x0000
+#define TIFM_MMCSD_CMD_BCR    0x1000
+#define TIFM_MMCSD_CMD_AC     0x2000
+#define TIFM_MMCSD_CMD_ADTC   0x3000
+
+#define TIFM_MMCSD_MAX_BLOCK_SIZE  0x0800UL
+
+#define TIFM_MMCSD_REQ_TIMEOUT_MS  1000
+
+enum {
+	CMD_READY    = 0x0001,
+	FIFO_READY   = 0x0002,
+	BRS_READY    = 0x0004,
+	SCMD_ACTIVE  = 0x0008,
+	SCMD_READY   = 0x0010,
+	CARD_BUSY    = 0x0020,
+	DATA_CARRY   = 0x0040
+};
+
+struct tifm_sd {
+	struct tifm_dev       *dev;
+
+	unsigned short        eject:1,
+			      open_drain:1,
+			      no_dma:1;
+	unsigned short        cmd_flags;
+
+	unsigned int          clk_freq;
+	unsigned int          clk_div;
+	unsigned long         timeout_jiffies;
+
+	struct tasklet_struct finish_tasklet;
+	struct timer_list     timer;
+	struct mmc_request    *req;
+
+	int                   sg_len;
+	int                   sg_pos;
+	unsigned int          block_pos;
+	struct scatterlist    bounce_buf;
+	unsigned char         bounce_buf_data[TIFM_MMCSD_MAX_BLOCK_SIZE];
+};
+
+/* for some reason, host won't respond correctly to readw/writew */
+static void tifm_sd_read_fifo(struct tifm_sd *host, struct page *pg,
+			      unsigned int off, unsigned int cnt)
+{
+	struct tifm_dev *sock = host->dev;
+	unsigned char *buf;
+	unsigned int pos = 0, val;
+
+	buf = kmap_atomic(pg) + off;
+	if (host->cmd_flags & DATA_CARRY) {
+		buf[pos++] = host->bounce_buf_data[0];
+		host->cmd_flags &= ~DATA_CARRY;
+	}
+
+	while (pos < cnt) {
+		val = readl(sock->addr + SOCK_MMCSD_DATA);
+		buf[pos++] = val & 0xff;
+		if (pos == cnt) {
+			host->bounce_buf_data[0] = (val >> 8) & 0xff;
+			host->cmd_flags |= DATA_CARRY;
+			break;
+		}
+		buf[pos++] = (val >> 8) & 0xff;
+	}
+	kunmap_atomic(buf - off);
+}
+
+static void tifm_sd_write_fifo(struct tifm_sd *host, struct page *pg,
+			       unsigned int off, unsigned int cnt)
+{
+	struct tifm_dev *sock = host->dev;
+	unsigned char *buf;
+	unsigned int pos = 0, val;
+
+	buf = kmap_atomic(pg) + off;
+	if (host->cmd_flags & DATA_CARRY) {
+		val = host->bounce_buf_data[0] | ((buf[pos++] << 8) & 0xff00);
+		writel(val, sock->addr + SOCK_MMCSD_DATA);
+		host->cmd_flags &= ~DATA_CARRY;
+	}
+
+	while (pos < cnt) {
+		val = buf[pos++];
+		if (pos == cnt) {
+			host->bounce_buf_data[0] = val & 0xff;
+			host->cmd_flags |= DATA_CARRY;
+			break;
+		}
+		val |= (buf[pos++] << 8) & 0xff00;
+		writel(val, sock->addr + SOCK_MMCSD_DATA);
+	}
+	kunmap_atomic(buf - off);
+}
+
+static void tifm_sd_transfer_data(struct tifm_sd *host)
+{
+	struct mmc_data *r_data = host->req->cmd->data;
+	struct scatterlist *sg = r_data->sg;
+	unsigned int off, cnt, t_size = TIFM_MMCSD_FIFO_SIZE * 2;
+	unsigned int p_off, p_cnt;
+	struct page *pg;
+
+	if (host->sg_pos == host->sg_len)
+		return;
+	while (t_size) {
+		cnt = sg[host->sg_pos].length - host->block_pos;
+		if (!cnt) {
+			host->block_pos = 0;
+			host->sg_pos++;
+			if (host->sg_pos == host->sg_len) {
+				if ((r_data->flags & MMC_DATA_WRITE)
+				    && (host->cmd_flags & DATA_CARRY))
+					writel(host->bounce_buf_data[0],
+					       host->dev->addr
+					       + SOCK_MMCSD_DATA);
+
+				return;
+			}
+			cnt = sg[host->sg_pos].length;
+		}
+		off = sg[host->sg_pos].offset + host->block_pos;
+
+		pg = nth_page(sg_page(&sg[host->sg_pos]), off >> PAGE_SHIFT);
+		p_off = offset_in_page(off);
+		p_cnt = PAGE_SIZE - p_off;
+		p_cnt = min(p_cnt, cnt);
+		p_cnt = min(p_cnt, t_size);
+
+		if (r_data->flags & MMC_DATA_READ)
+			tifm_sd_read_fifo(host, pg, p_off, p_cnt);
+		else if (r_data->flags & MMC_DATA_WRITE)
+			tifm_sd_write_fifo(host, pg, p_off, p_cnt);
+
+		t_size -= p_cnt;
+		host->block_pos += p_cnt;
+	}
+}
+
+static void tifm_sd_copy_page(struct page *dst, unsigned int dst_off,
+			      struct page *src, unsigned int src_off,
+			      unsigned int count)
+{
+	unsigned char *src_buf = kmap_atomic(src) + src_off;
+	unsigned char *dst_buf = kmap_atomic(dst) + dst_off;
+
+	memcpy(dst_buf, src_buf, count);
+
+	kunmap_atomic(dst_buf - dst_off);
+	kunmap_atomic(src_buf - src_off);
+}
+
+static void tifm_sd_bounce_block(struct tifm_sd *host, struct mmc_data *r_data)
+{
+	struct scatterlist *sg = r_data->sg;
+	unsigned int t_size = r_data->blksz;
+	unsigned int off, cnt;
+	unsigned int p_off, p_cnt;
+	struct page *pg;
+
+	dev_dbg(&host->dev->dev, "bouncing block\n");
+	while (t_size) {
+		cnt = sg[host->sg_pos].length - host->block_pos;
+		if (!cnt) {
+			host->block_pos = 0;
+			host->sg_pos++;
+			if (host->sg_pos == host->sg_len)
+				return;
+			cnt = sg[host->sg_pos].length;
+		}
+		off = sg[host->sg_pos].offset + host->block_pos;
+
+		pg = nth_page(sg_page(&sg[host->sg_pos]), off >> PAGE_SHIFT);
+		p_off = offset_in_page(off);
+		p_cnt = PAGE_SIZE - p_off;
+		p_cnt = min(p_cnt, cnt);
+		p_cnt = min(p_cnt, t_size);
+
+		if (r_data->flags & MMC_DATA_WRITE)
+			tifm_sd_copy_page(sg_page(&host->bounce_buf),
+					  r_data->blksz - t_size,
+					  pg, p_off, p_cnt);
+		else if (r_data->flags & MMC_DATA_READ)
+			tifm_sd_copy_page(pg, p_off, sg_page(&host->bounce_buf),
+					  r_data->blksz - t_size, p_cnt);
+
+		t_size -= p_cnt;
+		host->block_pos += p_cnt;
+	}
+}
+
+static int tifm_sd_set_dma_data(struct tifm_sd *host, struct mmc_data *r_data)
+{
+	struct tifm_dev *sock = host->dev;
+	unsigned int t_size = TIFM_DMA_TSIZE * r_data->blksz;
+	unsigned int dma_len, dma_blk_cnt, dma_off;
+	struct scatterlist *sg = NULL;
+	unsigned long flags;
+
+	if (host->sg_pos == host->sg_len)
+		return 1;
+
+	if (host->cmd_flags & DATA_CARRY) {
+		host->cmd_flags &= ~DATA_CARRY;
+		local_irq_save(flags);
+		tifm_sd_bounce_block(host, r_data);
+		local_irq_restore(flags);
+		if (host->sg_pos == host->sg_len)
+			return 1;
+	}
+
+	dma_len = sg_dma_len(&r_data->sg[host->sg_pos]) - host->block_pos;
+	if (!dma_len) {
+		host->block_pos = 0;
+		host->sg_pos++;
+		if (host->sg_pos == host->sg_len)
+			return 1;
+		dma_len = sg_dma_len(&r_data->sg[host->sg_pos]);
+	}
+
+	if (dma_len < t_size) {
+		dma_blk_cnt = dma_len / r_data->blksz;
+		dma_off = host->block_pos;
+		host->block_pos += dma_blk_cnt * r_data->blksz;
+	} else {
+		dma_blk_cnt = TIFM_DMA_TSIZE;
+		dma_off = host->block_pos;
+		host->block_pos += t_size;
+	}
+
+	if (dma_blk_cnt)
+		sg = &r_data->sg[host->sg_pos];
+	else if (dma_len) {
+		if (r_data->flags & MMC_DATA_WRITE) {
+			local_irq_save(flags);
+			tifm_sd_bounce_block(host, r_data);
+			local_irq_restore(flags);
+		} else
+			host->cmd_flags |= DATA_CARRY;
+
+		sg = &host->bounce_buf;
+		dma_off = 0;
+		dma_blk_cnt = 1;
+	} else
+		return 1;
+
+	dev_dbg(&sock->dev, "setting dma for %d blocks\n", dma_blk_cnt);
+	writel(sg_dma_address(sg) + dma_off, sock->addr + SOCK_DMA_ADDRESS);
+	if (r_data->flags & MMC_DATA_WRITE)
+		writel((dma_blk_cnt << 8) | TIFM_DMA_TX | TIFM_DMA_EN,
+		       sock->addr + SOCK_DMA_CONTROL);
+	else
+		writel((dma_blk_cnt << 8) | TIFM_DMA_EN,
+		       sock->addr + SOCK_DMA_CONTROL);
+
+	return 0;
+}
+
+static unsigned int tifm_sd_op_flags(struct mmc_command *cmd)
+{
+	unsigned int rc = 0;
+
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE:
+		rc |= TIFM_MMCSD_RSP_R0;
+		break;
+	case MMC_RSP_R1B:
+		rc |= TIFM_MMCSD_RSP_BUSY;
+		fallthrough;
+	case MMC_RSP_R1:
+		rc |= TIFM_MMCSD_RSP_R1;
+		break;
+	case MMC_RSP_R2:
+		rc |= TIFM_MMCSD_RSP_R2;
+		break;
+	case MMC_RSP_R3:
+		rc |= TIFM_MMCSD_RSP_R3;
+		break;
+	default:
+		BUG();
+	}
+
+	switch (mmc_cmd_type(cmd)) {
+	case MMC_CMD_BC:
+		rc |= TIFM_MMCSD_CMD_BC;
+		break;
+	case MMC_CMD_BCR:
+		rc |= TIFM_MMCSD_CMD_BCR;
+		break;
+	case MMC_CMD_AC:
+		rc |= TIFM_MMCSD_CMD_AC;
+		break;
+	case MMC_CMD_ADTC:
+		rc |= TIFM_MMCSD_CMD_ADTC;
+		break;
+	default:
+		BUG();
+	}
+	return rc;
+}
+
+static void tifm_sd_exec(struct tifm_sd *host, struct mmc_command *cmd)
+{
+	struct tifm_dev *sock = host->dev;
+	unsigned int cmd_mask = tifm_sd_op_flags(cmd);
+
+	if (host->open_drain)
+		cmd_mask |= TIFM_MMCSD_ODTO;
+
+	if (cmd->data && (cmd->data->flags & MMC_DATA_READ))
+		cmd_mask |= TIFM_MMCSD_READ;
+
+	dev_dbg(&sock->dev, "executing opcode 0x%x, arg: 0x%x, mask: 0x%x\n",
+		cmd->opcode, cmd->arg, cmd_mask);
+
+	writel((cmd->arg >> 16) & 0xffff, sock->addr + SOCK_MMCSD_ARG_HIGH);
+	writel(cmd->arg & 0xffff, sock->addr + SOCK_MMCSD_ARG_LOW);
+	writel(cmd->opcode | cmd_mask, sock->addr + SOCK_MMCSD_COMMAND);
+}
+
+static void tifm_sd_fetch_resp(struct mmc_command *cmd, struct tifm_dev *sock)
+{
+	cmd->resp[0] = (readl(sock->addr + SOCK_MMCSD_RESPONSE + 0x1c) << 16)
+		       | readl(sock->addr + SOCK_MMCSD_RESPONSE + 0x18);
+	cmd->resp[1] = (readl(sock->addr + SOCK_MMCSD_RESPONSE + 0x14) << 16)
+		       | readl(sock->addr + SOCK_MMCSD_RESPONSE + 0x10);
+	cmd->resp[2] = (readl(sock->addr + SOCK_MMCSD_RESPONSE + 0x0c) << 16)
+		       | readl(sock->addr + SOCK_MMCSD_RESPONSE + 0x08);
+	cmd->resp[3] = (readl(sock->addr + SOCK_MMCSD_RESPONSE + 0x04) << 16)
+		       | readl(sock->addr + SOCK_MMCSD_RESPONSE + 0x00);
+}
+
+static void tifm_sd_check_status(struct tifm_sd *host)
+{
+	struct tifm_dev *sock = host->dev;
+	struct mmc_command *cmd = host->req->cmd;
+
+	if (cmd->error)
+		goto finish_request;
+
+	if (!(host->cmd_flags & CMD_READY))
+		return;
+
+	if (cmd->data) {
+		if (cmd->data->error) {
+			if ((host->cmd_flags & SCMD_ACTIVE)
+			    && !(host->cmd_flags & SCMD_READY))
+				return;
+
+			goto finish_request;
+		}
+
+		if (!(host->cmd_flags & BRS_READY))
+			return;
+
+		if (!(host->no_dma || (host->cmd_flags & FIFO_READY)))
+			return;
+
+		if (cmd->data->flags & MMC_DATA_WRITE) {
+			if (host->req->stop) {
+				if (!(host->cmd_flags & SCMD_ACTIVE)) {
+					host->cmd_flags |= SCMD_ACTIVE;
+					writel(TIFM_MMCSD_EOFB
+					       | readl(sock->addr
+						       + SOCK_MMCSD_INT_ENABLE),
+					       sock->addr
+					       + SOCK_MMCSD_INT_ENABLE);
+					tifm_sd_exec(host, host->req->stop);
+					return;
+				} else {
+					if (!(host->cmd_flags & SCMD_READY)
+					    || (host->cmd_flags & CARD_BUSY))
+						return;
+					writel((~TIFM_MMCSD_EOFB)
+					       & readl(sock->addr
+						       + SOCK_MMCSD_INT_ENABLE),
+					       sock->addr
+					       + SOCK_MMCSD_INT_ENABLE);
+				}
+			} else {
+				if (host->cmd_flags & CARD_BUSY)
+					return;
+				writel((~TIFM_MMCSD_EOFB)
+				       & readl(sock->addr
+					       + SOCK_MMCSD_INT_ENABLE),
+				       sock->addr + SOCK_MMCSD_INT_ENABLE);
+			}
+		} else {
+			if (host->req->stop) {
+				if (!(host->cmd_flags & SCMD_ACTIVE)) {
+					host->cmd_flags |= SCMD_ACTIVE;
+					tifm_sd_exec(host, host->req->stop);
+					return;
+				} else {
+					if (!(host->cmd_flags & SCMD_READY))
+						return;
+				}
+			}
+		}
+	}
+finish_request:
+	tasklet_schedule(&host->finish_tasklet);
+}
+
+/* Called from interrupt handler */
+static void tifm_sd_data_event(struct tifm_dev *sock)
+{
+	struct tifm_sd *host;
+	unsigned int fifo_status = 0;
+	struct mmc_data *r_data = NULL;
+
+	spin_lock(&sock->lock);
+	host = mmc_priv((struct mmc_host*)tifm_get_drvdata(sock));
+	fifo_status = readl(sock->addr + SOCK_DMA_FIFO_STATUS);
+	dev_dbg(&sock->dev, "data event: fifo_status %x, flags %x\n",
+		fifo_status, host->cmd_flags);
+
+	if (host->req) {
+		r_data = host->req->cmd->data;
+
+		if (r_data && (fifo_status & TIFM_FIFO_READY)) {
+			if (tifm_sd_set_dma_data(host, r_data)) {
+				host->cmd_flags |= FIFO_READY;
+				tifm_sd_check_status(host);
+			}
+		}
+	}
+
+	writel(fifo_status, sock->addr + SOCK_DMA_FIFO_STATUS);
+	spin_unlock(&sock->lock);
+}
+
+/* Called from interrupt handler */
+static void tifm_sd_card_event(struct tifm_dev *sock)
+{
+	struct tifm_sd *host;
+	unsigned int host_status = 0;
+	int cmd_error = 0;
+	struct mmc_command *cmd = NULL;
+	unsigned long flags;
+
+	spin_lock(&sock->lock);
+	host = mmc_priv((struct mmc_host*)tifm_get_drvdata(sock));
+	host_status = readl(sock->addr + SOCK_MMCSD_STATUS);
+	dev_dbg(&sock->dev, "host event: host_status %x, flags %x\n",
+		host_status, host->cmd_flags);
+
+	if (host->req) {
+		cmd = host->req->cmd;
+
+		if (host_status & TIFM_MMCSD_ERRMASK) {
+			writel(host_status & TIFM_MMCSD_ERRMASK,
+			       sock->addr + SOCK_MMCSD_STATUS);
+			if (host_status & TIFM_MMCSD_CTO)
+				cmd_error = -ETIMEDOUT;
+			else if (host_status & TIFM_MMCSD_CCRC)
+				cmd_error = -EILSEQ;
+
+			if (cmd->data) {
+				if (host_status & TIFM_MMCSD_DTO)
+					cmd->data->error = -ETIMEDOUT;
+				else if (host_status & TIFM_MMCSD_DCRC)
+					cmd->data->error = -EILSEQ;
+			}
+
+			writel(TIFM_FIFO_INT_SETALL,
+			       sock->addr + SOCK_DMA_FIFO_INT_ENABLE_CLEAR);
+			writel(TIFM_DMA_RESET, sock->addr + SOCK_DMA_CONTROL);
+
+			if (host->req->stop) {
+				if (host->cmd_flags & SCMD_ACTIVE) {
+					host->req->stop->error = cmd_error;
+					host->cmd_flags |= SCMD_READY;
+				} else {
+					cmd->error = cmd_error;
+					host->cmd_flags |= SCMD_ACTIVE;
+					tifm_sd_exec(host, host->req->stop);
+					goto done;
+				}
+			} else
+				cmd->error = cmd_error;
+		} else {
+			if (host_status & (TIFM_MMCSD_EOC | TIFM_MMCSD_CERR)) {
+				if (!(host->cmd_flags & CMD_READY)) {
+					host->cmd_flags |= CMD_READY;
+					tifm_sd_fetch_resp(cmd, sock);
+				} else if (host->cmd_flags & SCMD_ACTIVE) {
+					host->cmd_flags |= SCMD_READY;
+					tifm_sd_fetch_resp(host->req->stop,
+							   sock);
+				}
+			}
+			if (host_status & TIFM_MMCSD_BRS)
+				host->cmd_flags |= BRS_READY;
+		}
+
+		if (host->no_dma && cmd->data) {
+			if (host_status & TIFM_MMCSD_AE)
+				writel(host_status & TIFM_MMCSD_AE,
+				       sock->addr + SOCK_MMCSD_STATUS);
+
+			if (host_status & (TIFM_MMCSD_AE | TIFM_MMCSD_AF
+					   | TIFM_MMCSD_BRS)) {
+				local_irq_save(flags);
+				tifm_sd_transfer_data(host);
+				local_irq_restore(flags);
+				host_status &= ~TIFM_MMCSD_AE;
+			}
+		}
+
+		if (host_status & TIFM_MMCSD_EOFB)
+			host->cmd_flags &= ~CARD_BUSY;
+		else if (host_status & TIFM_MMCSD_CB)
+			host->cmd_flags |= CARD_BUSY;
+
+		tifm_sd_check_status(host);
+	}
+done:
+	writel(host_status, sock->addr + SOCK_MMCSD_STATUS);
+	spin_unlock(&sock->lock);
+}
+
+static void tifm_sd_set_data_timeout(struct tifm_sd *host,
+				     struct mmc_data *data)
+{
+	struct tifm_dev *sock = host->dev;
+	unsigned int data_timeout = data->timeout_clks;
+
+	if (fixed_timeout)
+		return;
+
+	data_timeout += data->timeout_ns /
+			((1000000000UL / host->clk_freq) * host->clk_div);
+
+	if (data_timeout < 0xffff) {
+		writel(data_timeout, sock->addr + SOCK_MMCSD_DATA_TO);
+		writel((~TIFM_MMCSD_DPE)
+		       & readl(sock->addr + SOCK_MMCSD_SDIO_MODE_CONFIG),
+		       sock->addr + SOCK_MMCSD_SDIO_MODE_CONFIG);
+	} else {
+		data_timeout = (data_timeout >> 10) + 1;
+		if (data_timeout > 0xffff)
+			data_timeout = 0;	/* set to unlimited */
+		writel(data_timeout, sock->addr + SOCK_MMCSD_DATA_TO);
+		writel(TIFM_MMCSD_DPE
+		       | readl(sock->addr + SOCK_MMCSD_SDIO_MODE_CONFIG),
+		       sock->addr + SOCK_MMCSD_SDIO_MODE_CONFIG);
+	}
+}
+
+static void tifm_sd_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct tifm_sd *host = mmc_priv(mmc);
+	struct tifm_dev *sock = host->dev;
+	unsigned long flags;
+	struct mmc_data *r_data = mrq->cmd->data;
+
+	spin_lock_irqsave(&sock->lock, flags);
+	if (host->eject) {
+		mrq->cmd->error = -ENOMEDIUM;
+		goto err_out;
+	}
+
+	if (host->req) {
+		pr_err("%s : unfinished request detected\n",
+		       dev_name(&sock->dev));
+		mrq->cmd->error = -ETIMEDOUT;
+		goto err_out;
+	}
+
+	host->cmd_flags = 0;
+	host->block_pos = 0;
+	host->sg_pos = 0;
+
+	if (mrq->data && !is_power_of_2(mrq->data->blksz))
+		host->no_dma = 1;
+	else
+		host->no_dma = no_dma ? 1 : 0;
+
+	if (r_data) {
+		tifm_sd_set_data_timeout(host, r_data);
+
+		if ((r_data->flags & MMC_DATA_WRITE) && !mrq->stop)
+			 writel(TIFM_MMCSD_EOFB
+				| readl(sock->addr + SOCK_MMCSD_INT_ENABLE),
+				sock->addr + SOCK_MMCSD_INT_ENABLE);
+
+		if (host->no_dma) {
+			writel(TIFM_MMCSD_BUFINT
+			       | readl(sock->addr + SOCK_MMCSD_INT_ENABLE),
+			       sock->addr + SOCK_MMCSD_INT_ENABLE);
+			writel(((TIFM_MMCSD_FIFO_SIZE - 1) << 8)
+			       | (TIFM_MMCSD_FIFO_SIZE - 1),
+			       sock->addr + SOCK_MMCSD_BUFFER_CONFIG);
+
+			host->sg_len = r_data->sg_len;
+		} else {
+			sg_init_one(&host->bounce_buf, host->bounce_buf_data,
+				    r_data->blksz);
+
+			if(1 != tifm_map_sg(sock, &host->bounce_buf, 1,
+					    r_data->flags & MMC_DATA_WRITE
+					    ? PCI_DMA_TODEVICE
+					    : PCI_DMA_FROMDEVICE)) {
+				pr_err("%s : scatterlist map failed\n",
+				       dev_name(&sock->dev));
+				mrq->cmd->error = -ENOMEM;
+				goto err_out;
+			}
+			host->sg_len = tifm_map_sg(sock, r_data->sg,
+						   r_data->sg_len,
+						   r_data->flags
+						   & MMC_DATA_WRITE
+						   ? PCI_DMA_TODEVICE
+						   : PCI_DMA_FROMDEVICE);
+			if (host->sg_len < 1) {
+				pr_err("%s : scatterlist map failed\n",
+				       dev_name(&sock->dev));
+				tifm_unmap_sg(sock, &host->bounce_buf, 1,
+					      r_data->flags & MMC_DATA_WRITE
+					      ? PCI_DMA_TODEVICE
+					      : PCI_DMA_FROMDEVICE);
+				mrq->cmd->error = -ENOMEM;
+				goto err_out;
+			}
+
+			writel(TIFM_FIFO_INT_SETALL,
+			       sock->addr + SOCK_DMA_FIFO_INT_ENABLE_CLEAR);
+			writel(ilog2(r_data->blksz) - 2,
+			       sock->addr + SOCK_FIFO_PAGE_SIZE);
+			writel(TIFM_FIFO_ENABLE,
+			       sock->addr + SOCK_FIFO_CONTROL);
+			writel(TIFM_FIFO_INTMASK,
+			       sock->addr + SOCK_DMA_FIFO_INT_ENABLE_SET);
+
+			if (r_data->flags & MMC_DATA_WRITE)
+				writel(TIFM_MMCSD_TXDE,
+				       sock->addr + SOCK_MMCSD_BUFFER_CONFIG);
+			else
+				writel(TIFM_MMCSD_RXDE,
+				       sock->addr + SOCK_MMCSD_BUFFER_CONFIG);
+
+			tifm_sd_set_dma_data(host, r_data);
+		}
+
+		writel(r_data->blocks - 1,
+		       sock->addr + SOCK_MMCSD_NUM_BLOCKS);
+		writel(r_data->blksz - 1,
+		       sock->addr + SOCK_MMCSD_BLOCK_LEN);
+	}
+
+	host->req = mrq;
+	mod_timer(&host->timer, jiffies + host->timeout_jiffies);
+	writel(TIFM_CTRL_LED | readl(sock->addr + SOCK_CONTROL),
+	       sock->addr + SOCK_CONTROL);
+	tifm_sd_exec(host, mrq->cmd);
+	spin_unlock_irqrestore(&sock->lock, flags);
+	return;
+
+err_out:
+	spin_unlock_irqrestore(&sock->lock, flags);
+	mmc_request_done(mmc, mrq);
+}
+
+static void tifm_sd_end_cmd(unsigned long data)
+{
+	struct tifm_sd *host = (struct tifm_sd*)data;
+	struct tifm_dev *sock = host->dev;
+	struct mmc_host *mmc = tifm_get_drvdata(sock);
+	struct mmc_request *mrq;
+	struct mmc_data *r_data = NULL;
+	unsigned long flags;
+
+	spin_lock_irqsave(&sock->lock, flags);
+
+	del_timer(&host->timer);
+	mrq = host->req;
+	host->req = NULL;
+
+	if (!mrq) {
+		pr_err(" %s : no request to complete?\n",
+		       dev_name(&sock->dev));
+		spin_unlock_irqrestore(&sock->lock, flags);
+		return;
+	}
+
+	r_data = mrq->cmd->data;
+	if (r_data) {
+		if (host->no_dma) {
+			writel((~TIFM_MMCSD_BUFINT)
+			       & readl(sock->addr + SOCK_MMCSD_INT_ENABLE),
+			       sock->addr + SOCK_MMCSD_INT_ENABLE);
+		} else {
+			tifm_unmap_sg(sock, &host->bounce_buf, 1,
+				      (r_data->flags & MMC_DATA_WRITE)
+				      ? PCI_DMA_TODEVICE : PCI_DMA_FROMDEVICE);
+			tifm_unmap_sg(sock, r_data->sg, r_data->sg_len,
+				      (r_data->flags & MMC_DATA_WRITE)
+				      ? PCI_DMA_TODEVICE : PCI_DMA_FROMDEVICE);
+		}
+
+		r_data->bytes_xfered = r_data->blocks
+			- readl(sock->addr + SOCK_MMCSD_NUM_BLOCKS) - 1;
+		r_data->bytes_xfered *= r_data->blksz;
+		r_data->bytes_xfered += r_data->blksz
+			- readl(sock->addr + SOCK_MMCSD_BLOCK_LEN) + 1;
+	}
+
+	writel((~TIFM_CTRL_LED) & readl(sock->addr + SOCK_CONTROL),
+	       sock->addr + SOCK_CONTROL);
+
+	spin_unlock_irqrestore(&sock->lock, flags);
+	mmc_request_done(mmc, mrq);
+}
+
+static void tifm_sd_abort(struct timer_list *t)
+{
+	struct tifm_sd *host = from_timer(host, t, timer);
+
+	pr_err("%s : card failed to respond for a long period of time "
+	       "(%x, %x)\n",
+	       dev_name(&host->dev->dev), host->req->cmd->opcode, host->cmd_flags);
+
+	tifm_eject(host->dev);
+}
+
+static void tifm_sd_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct tifm_sd *host = mmc_priv(mmc);
+	struct tifm_dev *sock = host->dev;
+	unsigned int clk_div1, clk_div2;
+	unsigned long flags;
+
+	spin_lock_irqsave(&sock->lock, flags);
+
+	dev_dbg(&sock->dev, "ios: clock = %u, vdd = %x, bus_mode = %x, "
+		"chip_select = %x, power_mode = %x, bus_width = %x\n",
+		ios->clock, ios->vdd, ios->bus_mode, ios->chip_select,
+		ios->power_mode, ios->bus_width);
+
+	if (ios->bus_width == MMC_BUS_WIDTH_4) {
+		writel(TIFM_MMCSD_4BBUS | readl(sock->addr + SOCK_MMCSD_CONFIG),
+		       sock->addr + SOCK_MMCSD_CONFIG);
+	} else {
+		writel((~TIFM_MMCSD_4BBUS)
+		       & readl(sock->addr + SOCK_MMCSD_CONFIG),
+		       sock->addr + SOCK_MMCSD_CONFIG);
+	}
+
+	if (ios->clock) {
+		clk_div1 = 20000000 / ios->clock;
+		if (!clk_div1)
+			clk_div1 = 1;
+
+		clk_div2 = 24000000 / ios->clock;
+		if (!clk_div2)
+			clk_div2 = 1;
+
+		if ((20000000 / clk_div1) > ios->clock)
+			clk_div1++;
+		if ((24000000 / clk_div2) > ios->clock)
+			clk_div2++;
+		if ((20000000 / clk_div1) > (24000000 / clk_div2)) {
+			host->clk_freq = 20000000;
+			host->clk_div = clk_div1;
+			writel((~TIFM_CTRL_FAST_CLK)
+			       & readl(sock->addr + SOCK_CONTROL),
+			       sock->addr + SOCK_CONTROL);
+		} else {
+			host->clk_freq = 24000000;
+			host->clk_div = clk_div2;
+			writel(TIFM_CTRL_FAST_CLK
+			       | readl(sock->addr + SOCK_CONTROL),
+			       sock->addr + SOCK_CONTROL);
+		}
+	} else {
+		host->clk_div = 0;
+	}
+	host->clk_div &= TIFM_MMCSD_CLKMASK;
+	writel(host->clk_div
+	       | ((~TIFM_MMCSD_CLKMASK)
+		  & readl(sock->addr + SOCK_MMCSD_CONFIG)),
+	       sock->addr + SOCK_MMCSD_CONFIG);
+
+	host->open_drain = (ios->bus_mode == MMC_BUSMODE_OPENDRAIN);
+
+	/* chip_select : maybe later */
+	//vdd
+	//power is set before probe / after remove
+
+	spin_unlock_irqrestore(&sock->lock, flags);
+}
+
+static int tifm_sd_ro(struct mmc_host *mmc)
+{
+	int rc = 0;
+	struct tifm_sd *host = mmc_priv(mmc);
+	struct tifm_dev *sock = host->dev;
+	unsigned long flags;
+
+	spin_lock_irqsave(&sock->lock, flags);
+	if (TIFM_MMCSD_CARD_RO & readl(sock->addr + SOCK_PRESENT_STATE))
+		rc = 1;
+	spin_unlock_irqrestore(&sock->lock, flags);
+	return rc;
+}
+
+static const struct mmc_host_ops tifm_sd_ops = {
+	.request = tifm_sd_request,
+	.set_ios = tifm_sd_ios,
+	.get_ro  = tifm_sd_ro
+};
+
+static int tifm_sd_initialize_host(struct tifm_sd *host)
+{
+	int rc;
+	unsigned int host_status = 0;
+	struct tifm_dev *sock = host->dev;
+
+	writel(0, sock->addr + SOCK_MMCSD_INT_ENABLE);
+	host->clk_div = 61;
+	host->clk_freq = 20000000;
+	writel(TIFM_MMCSD_RESET, sock->addr + SOCK_MMCSD_SYSTEM_CONTROL);
+	writel(host->clk_div | TIFM_MMCSD_POWER,
+	       sock->addr + SOCK_MMCSD_CONFIG);
+
+	/* wait up to 0.51 sec for reset */
+	for (rc = 32; rc <= 256; rc <<= 1) {
+		if (1 & readl(sock->addr + SOCK_MMCSD_SYSTEM_STATUS)) {
+			rc = 0;
+			break;
+		}
+		msleep(rc);
+	}
+
+	if (rc) {
+		pr_err("%s : controller failed to reset\n",
+		       dev_name(&sock->dev));
+		return -ENODEV;
+	}
+
+	writel(0, sock->addr + SOCK_MMCSD_NUM_BLOCKS);
+	writel(host->clk_div | TIFM_MMCSD_POWER,
+	       sock->addr + SOCK_MMCSD_CONFIG);
+	writel(TIFM_MMCSD_RXDE, sock->addr + SOCK_MMCSD_BUFFER_CONFIG);
+
+	// command timeout fixed to 64 clocks for now
+	writel(64, sock->addr + SOCK_MMCSD_COMMAND_TO);
+	writel(TIFM_MMCSD_INAB, sock->addr + SOCK_MMCSD_COMMAND);
+
+	for (rc = 16; rc <= 64; rc <<= 1) {
+		host_status = readl(sock->addr + SOCK_MMCSD_STATUS);
+		writel(host_status, sock->addr + SOCK_MMCSD_STATUS);
+		if (!(host_status & TIFM_MMCSD_ERRMASK)
+		    && (host_status & TIFM_MMCSD_EOC)) {
+			rc = 0;
+			break;
+		}
+		msleep(rc);
+	}
+
+	if (rc) {
+		pr_err("%s : card not ready - probe failed on initialization\n",
+		       dev_name(&sock->dev));
+		return -ENODEV;
+	}
+
+	writel(TIFM_MMCSD_CERR | TIFM_MMCSD_BRS | TIFM_MMCSD_EOC
+	       | TIFM_MMCSD_ERRMASK,
+	       sock->addr + SOCK_MMCSD_INT_ENABLE);
+
+	return 0;
+}
+
+static int tifm_sd_probe(struct tifm_dev *sock)
+{
+	struct mmc_host *mmc;
+	struct tifm_sd *host;
+	int rc = -EIO;
+
+	if (!(TIFM_SOCK_STATE_OCCUPIED
+	      & readl(sock->addr + SOCK_PRESENT_STATE))) {
+		pr_warn("%s : card gone, unexpectedly\n",
+			dev_name(&sock->dev));
+		return rc;
+	}
+
+	mmc = mmc_alloc_host(sizeof(struct tifm_sd), &sock->dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	host = mmc_priv(mmc);
+	tifm_set_drvdata(sock, mmc);
+	host->dev = sock;
+	host->timeout_jiffies = msecs_to_jiffies(TIFM_MMCSD_REQ_TIMEOUT_MS);
+	/*
+	 * We use a fixed request timeout of 1s, hence inform the core about it.
+	 * A future improvement should instead respect the cmd->busy_timeout.
+	 */
+	mmc->max_busy_timeout = TIFM_MMCSD_REQ_TIMEOUT_MS;
+
+	tasklet_init(&host->finish_tasklet, tifm_sd_end_cmd,
+		     (unsigned long)host);
+	timer_setup(&host->timer, tifm_sd_abort, 0);
+
+	mmc->ops = &tifm_sd_ops;
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+	mmc->caps = MMC_CAP_4_BIT_DATA;
+	mmc->f_min = 20000000 / 60;
+	mmc->f_max = 24000000;
+
+	mmc->max_blk_count = 2048;
+	mmc->max_segs = mmc->max_blk_count;
+	mmc->max_blk_size = min(TIFM_MMCSD_MAX_BLOCK_SIZE, PAGE_SIZE);
+	mmc->max_seg_size = mmc->max_blk_count * mmc->max_blk_size;
+	mmc->max_req_size = mmc->max_seg_size;
+
+	sock->card_event = tifm_sd_card_event;
+	sock->data_event = tifm_sd_data_event;
+	rc = tifm_sd_initialize_host(host);
+
+	if (!rc)
+		rc = mmc_add_host(mmc);
+	if (!rc)
+		return 0;
+
+	mmc_free_host(mmc);
+	return rc;
+}
+
+static void tifm_sd_remove(struct tifm_dev *sock)
+{
+	struct mmc_host *mmc = tifm_get_drvdata(sock);
+	struct tifm_sd *host = mmc_priv(mmc);
+	unsigned long flags;
+
+	spin_lock_irqsave(&sock->lock, flags);
+	host->eject = 1;
+	writel(0, sock->addr + SOCK_MMCSD_INT_ENABLE);
+	spin_unlock_irqrestore(&sock->lock, flags);
+
+	tasklet_kill(&host->finish_tasklet);
+
+	spin_lock_irqsave(&sock->lock, flags);
+	if (host->req) {
+		writel(TIFM_FIFO_INT_SETALL,
+		       sock->addr + SOCK_DMA_FIFO_INT_ENABLE_CLEAR);
+		writel(0, sock->addr + SOCK_DMA_FIFO_INT_ENABLE_SET);
+		host->req->cmd->error = -ENOMEDIUM;
+		if (host->req->stop)
+			host->req->stop->error = -ENOMEDIUM;
+		tasklet_schedule(&host->finish_tasklet);
+	}
+	spin_unlock_irqrestore(&sock->lock, flags);
+	mmc_remove_host(mmc);
+	dev_dbg(&sock->dev, "after remove\n");
+
+	mmc_free_host(mmc);
+}
+
+#ifdef CONFIG_PM
+
+static int tifm_sd_suspend(struct tifm_dev *sock, pm_message_t state)
+{
+	return 0;
+}
+
+static int tifm_sd_resume(struct tifm_dev *sock)
+{
+	struct mmc_host *mmc = tifm_get_drvdata(sock);
+	struct tifm_sd *host = mmc_priv(mmc);
+	int rc;
+
+	rc = tifm_sd_initialize_host(host);
+	dev_dbg(&sock->dev, "resume initialize %d\n", rc);
+
+	if (rc)
+		host->eject = 1;
+
+	return rc;
+}
+
+#else
+
+#define tifm_sd_suspend NULL
+#define tifm_sd_resume NULL
+
+#endif /* CONFIG_PM */
+
+static struct tifm_device_id tifm_sd_id_tbl[] = {
+	{ TIFM_TYPE_SD }, { }
+};
+
+static struct tifm_driver tifm_sd_driver = {
+	.driver = {
+		.name  = DRIVER_NAME,
+		.owner = THIS_MODULE
+	},
+	.id_table = tifm_sd_id_tbl,
+	.probe    = tifm_sd_probe,
+	.remove   = tifm_sd_remove,
+	.suspend  = tifm_sd_suspend,
+	.resume   = tifm_sd_resume
+};
+
+static int __init tifm_sd_init(void)
+{
+	return tifm_register_driver(&tifm_sd_driver);
+}
+
+static void __exit tifm_sd_exit(void)
+{
+	tifm_unregister_driver(&tifm_sd_driver);
+}
+
+MODULE_AUTHOR("Alex Dubov");
+MODULE_DESCRIPTION("TI FlashMedia SD driver");
+MODULE_LICENSE("GPL");
+MODULE_DEVICE_TABLE(tifm, tifm_sd_id_tbl);
+MODULE_VERSION(DRIVER_VERSION);
+
+module_init(tifm_sd_init);
+module_exit(tifm_sd_exit);
diff --git a/drivers/mmc/host/tmio_mmc.c b/drivers/mmc/host/tmio_mmc.c
new file mode 100644
index 000000000..d2d3b8df1
--- /dev/null
+++ b/drivers/mmc/host/tmio_mmc.c
@@ -0,0 +1,228 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Driver for the MMC / SD / SDIO cell found in:
+ *
+ * TC6393XB TC6391XB TC6387XB T7L66XB ASIC3
+ *
+ * Copyright (C) 2017 Renesas Electronics Corporation
+ * Copyright (C) 2017 Horms Solutions, Simon Horman
+ * Copyright (C) 2007 Ian Molton
+ * Copyright (C) 2004 Ian Molton
+ */
+
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/mfd/core.h>
+#include <linux/mfd/tmio.h>
+#include <linux/mmc/host.h>
+#include <linux/module.h>
+#include <linux/pagemap.h>
+#include <linux/scatterlist.h>
+
+#include "tmio_mmc.h"
+
+/* Registers specific to this variant */
+#define CTL_SDIO_REGS		0x100
+#define CTL_CLK_AND_WAIT_CTL	0x138
+#define CTL_RESET_SDIO		0x1e0
+
+static void tmio_mmc_clk_start(struct tmio_mmc_host *host)
+{
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, CLK_CTL_SCLKEN |
+		sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+
+	usleep_range(10000, 11000);
+	sd_ctrl_write16(host, CTL_CLK_AND_WAIT_CTL, 0x0100);
+	usleep_range(10000, 11000);
+}
+
+static void tmio_mmc_clk_stop(struct tmio_mmc_host *host)
+{
+	sd_ctrl_write16(host, CTL_CLK_AND_WAIT_CTL, 0x0000);
+	usleep_range(10000, 11000);
+
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, ~CLK_CTL_SCLKEN &
+		sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+
+	usleep_range(10000, 11000);
+}
+
+static void tmio_mmc_set_clock(struct tmio_mmc_host *host,
+			       unsigned int new_clock)
+{
+	unsigned int divisor;
+	u32 clk = 0;
+	int clk_sel;
+
+	if (new_clock == 0) {
+		tmio_mmc_clk_stop(host);
+		return;
+	}
+
+	divisor = host->pdata->hclk / new_clock;
+
+	/* bit7 set: 1/512, ... bit0 set: 1/4, all bits clear: 1/2 */
+	clk_sel = (divisor <= 1);
+	clk = clk_sel ? 0 : (roundup_pow_of_two(divisor) >> 2);
+
+	host->pdata->set_clk_div(host->pdev, clk_sel);
+
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, ~CLK_CTL_SCLKEN &
+			sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, clk & CLK_CTL_DIV_MASK);
+	usleep_range(10000, 11000);
+
+	tmio_mmc_clk_start(host);
+}
+
+static void tmio_mmc_reset(struct tmio_mmc_host *host)
+{
+	sd_ctrl_write16(host, CTL_RESET_SDIO, 0x0000);
+	usleep_range(10000, 11000);
+	sd_ctrl_write16(host, CTL_RESET_SDIO, 0x0001);
+	usleep_range(10000, 11000);
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int tmio_mmc_suspend(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	const struct mfd_cell *cell = mfd_get_cell(pdev);
+	int ret;
+
+	ret = pm_runtime_force_suspend(dev);
+
+	/* Tell MFD core it can disable us now.*/
+	if (!ret && cell->disable)
+		cell->disable(pdev);
+
+	return ret;
+}
+
+static int tmio_mmc_resume(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	const struct mfd_cell *cell = mfd_get_cell(pdev);
+	int ret = 0;
+
+	/* Tell the MFD core we are ready to be enabled */
+	if (cell->resume)
+		ret = cell->resume(pdev);
+
+	if (!ret)
+		ret = pm_runtime_force_resume(dev);
+
+	return ret;
+}
+#endif
+
+static int tmio_mmc_probe(struct platform_device *pdev)
+{
+	const struct mfd_cell *cell = mfd_get_cell(pdev);
+	struct tmio_mmc_data *pdata;
+	struct tmio_mmc_host *host;
+	struct resource *res;
+	int ret = -EINVAL, irq;
+
+	if (pdev->num_resources != 2)
+		goto out;
+
+	pdata = pdev->dev.platform_data;
+	if (!pdata || !pdata->hclk)
+		goto out;
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0) {
+		ret = irq;
+		goto out;
+	}
+
+	/* Tell the MFD core we are ready to be enabled */
+	if (cell->enable) {
+		ret = cell->enable(pdev);
+		if (ret)
+			goto out;
+	}
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res) {
+		ret = -EINVAL;
+		goto cell_disable;
+	}
+
+	host = tmio_mmc_host_alloc(pdev, pdata);
+	if (IS_ERR(host)) {
+		ret = PTR_ERR(host);
+		goto cell_disable;
+	}
+
+	/* SD control register space size is 0x200, 0x400 for bus_shift=1 */
+	host->bus_shift = resource_size(res) >> 10;
+	host->set_clock = tmio_mmc_set_clock;
+	host->reset = tmio_mmc_reset;
+
+	host->mmc->f_max = pdata->hclk;
+	host->mmc->f_min = pdata->hclk / 512;
+
+	ret = tmio_mmc_host_probe(host);
+	if (ret)
+		goto host_free;
+
+	ret = devm_request_irq(&pdev->dev, irq, tmio_mmc_irq,
+			       IRQF_TRIGGER_FALLING,
+			       dev_name(&pdev->dev), host);
+	if (ret)
+		goto host_remove;
+
+	pr_info("%s at 0x%08lx irq %d\n", mmc_hostname(host->mmc),
+		(unsigned long)host->ctl, irq);
+
+	return 0;
+
+host_remove:
+	tmio_mmc_host_remove(host);
+host_free:
+	tmio_mmc_host_free(host);
+cell_disable:
+	if (cell->disable)
+		cell->disable(pdev);
+out:
+	return ret;
+}
+
+static int tmio_mmc_remove(struct platform_device *pdev)
+{
+	const struct mfd_cell *cell = mfd_get_cell(pdev);
+	struct tmio_mmc_host *host = platform_get_drvdata(pdev);
+
+	tmio_mmc_host_remove(host);
+	if (cell->disable)
+		cell->disable(pdev);
+
+	return 0;
+}
+
+/* ------------------- device registration ----------------------- */
+
+static const struct dev_pm_ops tmio_mmc_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(tmio_mmc_suspend, tmio_mmc_resume)
+	SET_RUNTIME_PM_OPS(tmio_mmc_host_runtime_suspend,
+			   tmio_mmc_host_runtime_resume, NULL)
+};
+
+static struct platform_driver tmio_mmc_driver = {
+	.driver = {
+		.name = "tmio-mmc",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm = &tmio_mmc_dev_pm_ops,
+	},
+	.probe = tmio_mmc_probe,
+	.remove = tmio_mmc_remove,
+};
+
+module_platform_driver(tmio_mmc_driver);
+
+MODULE_DESCRIPTION("Toshiba TMIO SD/MMC driver");
+MODULE_AUTHOR("Ian Molton <spyro@f2s.com>");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:tmio-mmc");
diff --git a/drivers/mmc/host/tmio_mmc.h b/drivers/mmc/host/tmio_mmc.h
new file mode 100644
index 000000000..330a17267
--- /dev/null
+++ b/drivers/mmc/host/tmio_mmc.h
@@ -0,0 +1,284 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Driver for the MMC / SD / SDIO cell found in:
+ *
+ * TC6393XB TC6391XB TC6387XB T7L66XB ASIC3
+ *
+ * Copyright (C) 2015-19 Renesas Electronics Corporation
+ * Copyright (C) 2016-19 Sang Engineering, Wolfram Sang
+ * Copyright (C) 2016-17 Horms Solutions, Simon Horman
+ * Copyright (C) 2007 Ian Molton
+ * Copyright (C) 2004 Ian Molton
+ */
+
+#ifndef TMIO_MMC_H
+#define TMIO_MMC_H
+
+#include <linux/dmaengine.h>
+#include <linux/highmem.h>
+#include <linux/mutex.h>
+#include <linux/pagemap.h>
+#include <linux/scatterlist.h>
+#include <linux/spinlock.h>
+#include <linux/interrupt.h>
+
+#define CTL_SD_CMD 0x00
+#define CTL_ARG_REG 0x04
+#define CTL_STOP_INTERNAL_ACTION 0x08
+#define CTL_XFER_BLK_COUNT 0xa
+#define CTL_RESPONSE 0x0c
+/* driver merges STATUS and following STATUS2 */
+#define CTL_STATUS 0x1c
+/* driver merges IRQ_MASK and following IRQ_MASK2 */
+#define CTL_IRQ_MASK 0x20
+#define CTL_SD_CARD_CLK_CTL 0x24
+#define CTL_SD_XFER_LEN 0x26
+#define CTL_SD_MEM_CARD_OPT 0x28
+#define CTL_SD_ERROR_DETAIL_STATUS 0x2c
+#define CTL_SD_DATA_PORT 0x30
+#define CTL_TRANSACTION_CTL 0x34
+#define CTL_SDIO_STATUS 0x36
+#define CTL_SDIO_IRQ_MASK 0x38
+#define CTL_DMA_ENABLE 0xd8
+#define CTL_RESET_SD 0xe0
+#define CTL_VERSION 0xe2
+#define CTL_SDIF_MODE 0xe6
+
+/* Definitions for values the CTL_STOP_INTERNAL_ACTION register can take */
+#define TMIO_STOP_STP		BIT(0)
+#define TMIO_STOP_SEC		BIT(8)
+
+/* Definitions for values the CTL_STATUS register can take */
+#define TMIO_STAT_CMDRESPEND    BIT(0)
+#define TMIO_STAT_DATAEND       BIT(2)
+#define TMIO_STAT_CARD_REMOVE   BIT(3)
+#define TMIO_STAT_CARD_INSERT   BIT(4)
+#define TMIO_STAT_SIGSTATE      BIT(5)
+#define TMIO_STAT_WRPROTECT     BIT(7)
+#define TMIO_STAT_CARD_REMOVE_A BIT(8)
+#define TMIO_STAT_CARD_INSERT_A BIT(9)
+#define TMIO_STAT_SIGSTATE_A    BIT(10)
+
+/* These belong technically to CTL_STATUS2, but the driver merges them */
+#define TMIO_STAT_CMD_IDX_ERR   BIT(16)
+#define TMIO_STAT_CRCFAIL       BIT(17)
+#define TMIO_STAT_STOPBIT_ERR   BIT(18)
+#define TMIO_STAT_DATATIMEOUT   BIT(19)
+#define TMIO_STAT_RXOVERFLOW    BIT(20)
+#define TMIO_STAT_TXUNDERRUN    BIT(21)
+#define TMIO_STAT_CMDTIMEOUT    BIT(22)
+#define TMIO_STAT_DAT0		BIT(23)	/* only known on R-Car so far */
+#define TMIO_STAT_RXRDY         BIT(24)
+#define TMIO_STAT_TXRQ          BIT(25)
+#define TMIO_STAT_ALWAYS_SET_27	BIT(27) /* only known on R-Car 2+ so far */
+#define TMIO_STAT_ILL_FUNC      BIT(29) /* only when !TMIO_MMC_HAS_IDLE_WAIT */
+#define TMIO_STAT_SCLKDIVEN     BIT(29) /* only when TMIO_MMC_HAS_IDLE_WAIT */
+#define TMIO_STAT_CMD_BUSY      BIT(30)
+#define TMIO_STAT_ILL_ACCESS    BIT(31)
+
+/* Definitions for values the CTL_SD_CARD_CLK_CTL register can take */
+#define	CLK_CTL_DIV_MASK	0xff
+#define	CLK_CTL_SCLKEN		BIT(8)
+
+/* Definitions for values the CTL_SD_MEM_CARD_OPT register can take */
+#define CARD_OPT_WIDTH8		BIT(13)
+#define CARD_OPT_WIDTH		BIT(15)
+
+/* Definitions for values the CTL_SDIO_STATUS register can take */
+#define TMIO_SDIO_STAT_IOIRQ	0x0001
+#define TMIO_SDIO_STAT_EXPUB52	0x4000
+#define TMIO_SDIO_STAT_EXWT	0x8000
+#define TMIO_SDIO_MASK_ALL	0xc007
+
+#define TMIO_SDIO_SETBITS_MASK	0x0006
+
+/* Definitions for values the CTL_DMA_ENABLE register can take */
+#define DMA_ENABLE_DMASDRW	BIT(1)
+
+/* Define some IRQ masks */
+/* This is the mask used at reset by the chip */
+#define TMIO_MASK_ALL           0x837f031d
+#define TMIO_MASK_ALL_RCAR2	0x8b7f031d
+#define TMIO_MASK_READOP  (TMIO_STAT_RXRDY | TMIO_STAT_DATAEND)
+#define TMIO_MASK_WRITEOP (TMIO_STAT_TXRQ | TMIO_STAT_DATAEND)
+#define TMIO_MASK_CMD     (TMIO_STAT_CMDRESPEND | TMIO_STAT_CMDTIMEOUT | \
+		TMIO_STAT_CARD_REMOVE | TMIO_STAT_CARD_INSERT)
+#define TMIO_MASK_IRQ     (TMIO_MASK_READOP | TMIO_MASK_WRITEOP | TMIO_MASK_CMD)
+
+#define TMIO_MAX_BLK_SIZE 512
+
+struct tmio_mmc_data;
+struct tmio_mmc_host;
+
+struct tmio_mmc_dma_ops {
+	void (*start)(struct tmio_mmc_host *host, struct mmc_data *data);
+	void (*enable)(struct tmio_mmc_host *host, bool enable);
+	void (*request)(struct tmio_mmc_host *host,
+			struct tmio_mmc_data *pdata);
+	void (*release)(struct tmio_mmc_host *host);
+	void (*abort)(struct tmio_mmc_host *host);
+	void (*dataend)(struct tmio_mmc_host *host);
+
+	/* optional */
+	void (*end)(struct tmio_mmc_host *host);	/* held host->lock */
+};
+
+struct tmio_mmc_host {
+	void __iomem *ctl;
+	struct mmc_command      *cmd;
+	struct mmc_request      *mrq;
+	struct mmc_data         *data;
+	struct mmc_host         *mmc;
+	struct mmc_host_ops     ops;
+
+	/* Callbacks for clock / power control */
+	void (*set_pwr)(struct platform_device *host, int state);
+
+	/* pio related stuff */
+	struct scatterlist      *sg_ptr;
+	struct scatterlist      *sg_orig;
+	unsigned int            sg_len;
+	unsigned int            sg_off;
+	unsigned int		bus_shift;
+
+	struct platform_device *pdev;
+	struct tmio_mmc_data *pdata;
+
+	/* DMA support */
+	bool			dma_on;
+	struct dma_chan		*chan_rx;
+	struct dma_chan		*chan_tx;
+	struct tasklet_struct	dma_issue;
+	struct scatterlist	bounce_sg;
+	u8			*bounce_buf;
+
+	/* Track lost interrupts */
+	struct delayed_work	delayed_reset_work;
+	struct work_struct	done;
+
+	/* Cache */
+	u32			sdcard_irq_mask;
+	u32			sdio_irq_mask;
+	unsigned int		clk_cache;
+	u32			sdcard_irq_setbit_mask;
+	u32			sdcard_irq_mask_all;
+
+	spinlock_t		lock;		/* protect host private data */
+	unsigned long		last_req_ts;
+	struct mutex		ios_lock;	/* protect set_ios() context */
+	bool			native_hotplug;
+	bool			sdio_irq_enabled;
+
+	/* Mandatory callback */
+	int (*clk_enable)(struct tmio_mmc_host *host);
+	void (*set_clock)(struct tmio_mmc_host *host, unsigned int clock);
+
+	/* Optional callbacks */
+	void (*clk_disable)(struct tmio_mmc_host *host);
+	int (*multi_io_quirk)(struct mmc_card *card,
+			      unsigned int direction, int blk_size);
+	int (*write16_hook)(struct tmio_mmc_host *host, int addr);
+	void (*reset)(struct tmio_mmc_host *host);
+	bool (*check_retune)(struct tmio_mmc_host *host);
+	void (*fixup_request)(struct tmio_mmc_host *host, struct mmc_request *mrq);
+
+	void (*prepare_hs400_tuning)(struct tmio_mmc_host *host);
+	void (*hs400_downgrade)(struct tmio_mmc_host *host);
+	void (*hs400_complete)(struct tmio_mmc_host *host);
+
+	const struct tmio_mmc_dma_ops *dma_ops;
+};
+
+struct tmio_mmc_host *tmio_mmc_host_alloc(struct platform_device *pdev,
+					  struct tmio_mmc_data *pdata);
+void tmio_mmc_host_free(struct tmio_mmc_host *host);
+int tmio_mmc_host_probe(struct tmio_mmc_host *host);
+void tmio_mmc_host_remove(struct tmio_mmc_host *host);
+void tmio_mmc_do_data_irq(struct tmio_mmc_host *host);
+
+void tmio_mmc_enable_mmc_irqs(struct tmio_mmc_host *host, u32 i);
+void tmio_mmc_disable_mmc_irqs(struct tmio_mmc_host *host, u32 i);
+irqreturn_t tmio_mmc_irq(int irq, void *devid);
+
+static inline char *tmio_mmc_kmap_atomic(struct scatterlist *sg,
+					 unsigned long *flags)
+{
+	local_irq_save(*flags);
+	return kmap_atomic(sg_page(sg)) + sg->offset;
+}
+
+static inline void tmio_mmc_kunmap_atomic(struct scatterlist *sg,
+					  unsigned long *flags, void *virt)
+{
+	kunmap_atomic(virt - sg->offset);
+	local_irq_restore(*flags);
+}
+
+#ifdef CONFIG_PM
+int tmio_mmc_host_runtime_suspend(struct device *dev);
+int tmio_mmc_host_runtime_resume(struct device *dev);
+#endif
+
+static inline u16 sd_ctrl_read16(struct tmio_mmc_host *host, int addr)
+{
+	return ioread16(host->ctl + (addr << host->bus_shift));
+}
+
+static inline void sd_ctrl_read16_rep(struct tmio_mmc_host *host, int addr,
+				      u16 *buf, int count)
+{
+	ioread16_rep(host->ctl + (addr << host->bus_shift), buf, count);
+}
+
+static inline u32 sd_ctrl_read16_and_16_as_32(struct tmio_mmc_host *host,
+					      int addr)
+{
+	return ioread16(host->ctl + (addr << host->bus_shift)) |
+	       ioread16(host->ctl + ((addr + 2) << host->bus_shift)) << 16;
+}
+
+static inline void sd_ctrl_read32_rep(struct tmio_mmc_host *host, int addr,
+				      u32 *buf, int count)
+{
+	ioread32_rep(host->ctl + (addr << host->bus_shift), buf, count);
+}
+
+static inline void sd_ctrl_write16(struct tmio_mmc_host *host, int addr,
+				   u16 val)
+{
+	/* If there is a hook and it returns non-zero then there
+	 * is an error and the write should be skipped
+	 */
+	if (host->write16_hook && host->write16_hook(host, addr))
+		return;
+	iowrite16(val, host->ctl + (addr << host->bus_shift));
+}
+
+static inline void sd_ctrl_write16_rep(struct tmio_mmc_host *host, int addr,
+				       u16 *buf, int count)
+{
+	iowrite16_rep(host->ctl + (addr << host->bus_shift), buf, count);
+}
+
+static inline void sd_ctrl_write32_as_16_and_16(struct tmio_mmc_host *host,
+						int addr, u32 val)
+{
+	if (addr == CTL_IRQ_MASK || addr == CTL_STATUS)
+		val |= host->sdcard_irq_setbit_mask;
+
+	iowrite16(val & 0xffff, host->ctl + (addr << host->bus_shift));
+	iowrite16(val >> 16, host->ctl + ((addr + 2) << host->bus_shift));
+}
+
+static inline void sd_ctrl_write32(struct tmio_mmc_host *host, int addr, u32 val)
+{
+	iowrite32(val, host->ctl + (addr << host->bus_shift));
+}
+
+static inline void sd_ctrl_write32_rep(struct tmio_mmc_host *host, int addr,
+				       const u32 *buf, int count)
+{
+	iowrite32_rep(host->ctl + (addr << host->bus_shift), buf, count);
+}
+
+#endif
diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
new file mode 100644
index 000000000..abf36acb2
--- /dev/null
+++ b/drivers/mmc/host/tmio_mmc_core.c
@@ -0,0 +1,1281 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Driver for the MMC / SD / SDIO IP found in:
+ *
+ * TC6393XB, TC6391XB, TC6387XB, T7L66XB, ASIC3, SH-Mobile SoCs
+ *
+ * Copyright (C) 2015-19 Renesas Electronics Corporation
+ * Copyright (C) 2016-19 Sang Engineering, Wolfram Sang
+ * Copyright (C) 2017 Horms Solutions, Simon Horman
+ * Copyright (C) 2011 Guennadi Liakhovetski
+ * Copyright (C) 2007 Ian Molton
+ * Copyright (C) 2004 Ian Molton
+ *
+ * This driver draws mainly on scattered spec sheets, Reverse engineering
+ * of the toshiba e800  SD driver and some parts of the 2.4 ASIC3 driver (4 bit
+ * support). (Further 4 bit support from a later datasheet).
+ *
+ * TODO:
+ *   Investigate using a workqueue for PIO transfers
+ *   Eliminate FIXMEs
+ *   Better Power management
+ *   Handle MMC errors better
+ *   double buffer support
+ *
+ */
+
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/highmem.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/irq.h>
+#include <linux/mfd/tmio.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/module.h>
+#include <linux/pagemap.h>
+#include <linux/platform_device.h>
+#include <linux/pm_qos.h>
+#include <linux/pm_runtime.h>
+#include <linux/regulator/consumer.h>
+#include <linux/mmc/sdio.h>
+#include <linux/scatterlist.h>
+#include <linux/sizes.h>
+#include <linux/spinlock.h>
+#include <linux/workqueue.h>
+
+#include "tmio_mmc.h"
+
+static inline void tmio_mmc_start_dma(struct tmio_mmc_host *host,
+				      struct mmc_data *data)
+{
+	if (host->dma_ops)
+		host->dma_ops->start(host, data);
+}
+
+static inline void tmio_mmc_end_dma(struct tmio_mmc_host *host)
+{
+	if (host->dma_ops && host->dma_ops->end)
+		host->dma_ops->end(host);
+}
+
+static inline void tmio_mmc_enable_dma(struct tmio_mmc_host *host, bool enable)
+{
+	if (host->dma_ops)
+		host->dma_ops->enable(host, enable);
+}
+
+static inline void tmio_mmc_request_dma(struct tmio_mmc_host *host,
+					struct tmio_mmc_data *pdata)
+{
+	if (host->dma_ops) {
+		host->dma_ops->request(host, pdata);
+	} else {
+		host->chan_tx = NULL;
+		host->chan_rx = NULL;
+	}
+}
+
+static inline void tmio_mmc_release_dma(struct tmio_mmc_host *host)
+{
+	if (host->dma_ops)
+		host->dma_ops->release(host);
+}
+
+static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
+{
+	if (host->dma_ops)
+		host->dma_ops->abort(host);
+}
+
+static inline void tmio_mmc_dataend_dma(struct tmio_mmc_host *host)
+{
+	if (host->dma_ops)
+		host->dma_ops->dataend(host);
+}
+
+void tmio_mmc_enable_mmc_irqs(struct tmio_mmc_host *host, u32 i)
+{
+	host->sdcard_irq_mask &= ~(i & TMIO_MASK_IRQ);
+	sd_ctrl_write32_as_16_and_16(host, CTL_IRQ_MASK, host->sdcard_irq_mask);
+}
+EXPORT_SYMBOL_GPL(tmio_mmc_enable_mmc_irqs);
+
+void tmio_mmc_disable_mmc_irqs(struct tmio_mmc_host *host, u32 i)
+{
+	host->sdcard_irq_mask |= (i & TMIO_MASK_IRQ);
+	sd_ctrl_write32_as_16_and_16(host, CTL_IRQ_MASK, host->sdcard_irq_mask);
+}
+EXPORT_SYMBOL_GPL(tmio_mmc_disable_mmc_irqs);
+
+static void tmio_mmc_ack_mmc_irqs(struct tmio_mmc_host *host, u32 i)
+{
+	sd_ctrl_write32_as_16_and_16(host, CTL_STATUS, ~i);
+}
+
+static void tmio_mmc_init_sg(struct tmio_mmc_host *host, struct mmc_data *data)
+{
+	host->sg_len = data->sg_len;
+	host->sg_ptr = data->sg;
+	host->sg_orig = data->sg;
+	host->sg_off = 0;
+}
+
+static int tmio_mmc_next_sg(struct tmio_mmc_host *host)
+{
+	host->sg_ptr = sg_next(host->sg_ptr);
+	host->sg_off = 0;
+	return --host->sg_len;
+}
+
+#define CMDREQ_TIMEOUT	5000
+
+static void tmio_mmc_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+
+	if (enable && !host->sdio_irq_enabled) {
+		u16 sdio_status;
+
+		/* Keep device active while SDIO irq is enabled */
+		pm_runtime_get_sync(mmc_dev(mmc));
+
+		host->sdio_irq_enabled = true;
+		host->sdio_irq_mask = TMIO_SDIO_MASK_ALL & ~TMIO_SDIO_STAT_IOIRQ;
+
+		/* Clear obsolete interrupts before enabling */
+		sdio_status = sd_ctrl_read16(host, CTL_SDIO_STATUS) & ~TMIO_SDIO_MASK_ALL;
+		if (host->pdata->flags & TMIO_MMC_SDIO_STATUS_SETBITS)
+			sdio_status |= TMIO_SDIO_SETBITS_MASK;
+		sd_ctrl_write16(host, CTL_SDIO_STATUS, sdio_status);
+
+		sd_ctrl_write16(host, CTL_SDIO_IRQ_MASK, host->sdio_irq_mask);
+	} else if (!enable && host->sdio_irq_enabled) {
+		host->sdio_irq_mask = TMIO_SDIO_MASK_ALL;
+		sd_ctrl_write16(host, CTL_SDIO_IRQ_MASK, host->sdio_irq_mask);
+
+		host->sdio_irq_enabled = false;
+		pm_runtime_mark_last_busy(mmc_dev(mmc));
+		pm_runtime_put_autosuspend(mmc_dev(mmc));
+	}
+}
+
+static void tmio_mmc_reset(struct tmio_mmc_host *host)
+{
+	/* FIXME - should we set stop clock reg here */
+	sd_ctrl_write16(host, CTL_RESET_SD, 0x0000);
+	usleep_range(10000, 11000);
+	sd_ctrl_write16(host, CTL_RESET_SD, 0x0001);
+	usleep_range(10000, 11000);
+
+	if (host->reset)
+		host->reset(host);
+
+	tmio_mmc_abort_dma(host);
+
+	if (host->pdata->flags & TMIO_MMC_SDIO_IRQ) {
+		sd_ctrl_write16(host, CTL_SDIO_IRQ_MASK, host->sdio_irq_mask);
+		sd_ctrl_write16(host, CTL_TRANSACTION_CTL, 0x0001);
+	}
+}
+
+static void tmio_mmc_reset_work(struct work_struct *work)
+{
+	struct tmio_mmc_host *host = container_of(work, struct tmio_mmc_host,
+						  delayed_reset_work.work);
+	struct mmc_request *mrq;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	mrq = host->mrq;
+
+	/*
+	 * is request already finished? Since we use a non-blocking
+	 * cancel_delayed_work(), it can happen, that a .set_ios() call preempts
+	 * us, so, have to check for IS_ERR(host->mrq)
+	 */
+	if (IS_ERR_OR_NULL(mrq) ||
+	    time_is_after_jiffies(host->last_req_ts +
+				  msecs_to_jiffies(CMDREQ_TIMEOUT))) {
+		spin_unlock_irqrestore(&host->lock, flags);
+		return;
+	}
+
+	dev_warn(&host->pdev->dev,
+		 "timeout waiting for hardware interrupt (CMD%u)\n",
+		 mrq->cmd->opcode);
+
+	if (host->data)
+		host->data->error = -ETIMEDOUT;
+	else if (host->cmd)
+		host->cmd->error = -ETIMEDOUT;
+	else
+		mrq->cmd->error = -ETIMEDOUT;
+
+	host->cmd = NULL;
+	host->data = NULL;
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	tmio_mmc_reset(host);
+
+	/* Ready for new calls */
+	host->mrq = NULL;
+	mmc_request_done(host->mmc, mrq);
+}
+
+/* These are the bitmasks the tmio chip requires to implement the MMC response
+ * types. Note that R1 and R6 are the same in this scheme. */
+#define APP_CMD        0x0040
+#define RESP_NONE      0x0300
+#define RESP_R1        0x0400
+#define RESP_R1B       0x0500
+#define RESP_R2        0x0600
+#define RESP_R3        0x0700
+#define DATA_PRESENT   0x0800
+#define TRANSFER_READ  0x1000
+#define TRANSFER_MULTI 0x2000
+#define SECURITY_CMD   0x4000
+#define NO_CMD12_ISSUE 0x4000 /* TMIO_MMC_HAVE_CMD12_CTRL */
+
+static int tmio_mmc_start_command(struct tmio_mmc_host *host,
+				  struct mmc_command *cmd)
+{
+	struct mmc_data *data = host->data;
+	int c = cmd->opcode;
+
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE: c |= RESP_NONE; break;
+	case MMC_RSP_R1:
+	case MMC_RSP_R1_NO_CRC:
+			   c |= RESP_R1;   break;
+	case MMC_RSP_R1B:  c |= RESP_R1B;  break;
+	case MMC_RSP_R2:   c |= RESP_R2;   break;
+	case MMC_RSP_R3:   c |= RESP_R3;   break;
+	default:
+		pr_debug("Unknown response type %d\n", mmc_resp_type(cmd));
+		return -EINVAL;
+	}
+
+	host->cmd = cmd;
+
+/* FIXME - this seems to be ok commented out but the spec suggest this bit
+ *         should be set when issuing app commands.
+ *	if(cmd->flags & MMC_FLAG_ACMD)
+ *		c |= APP_CMD;
+ */
+	if (data) {
+		c |= DATA_PRESENT;
+		if (data->blocks > 1) {
+			sd_ctrl_write16(host, CTL_STOP_INTERNAL_ACTION, TMIO_STOP_SEC);
+			c |= TRANSFER_MULTI;
+
+			/*
+			 * Disable auto CMD12 at IO_RW_EXTENDED and
+			 * SET_BLOCK_COUNT when doing multiple block transfer
+			 */
+			if ((host->pdata->flags & TMIO_MMC_HAVE_CMD12_CTRL) &&
+			    (cmd->opcode == SD_IO_RW_EXTENDED || host->mrq->sbc))
+				c |= NO_CMD12_ISSUE;
+		}
+		if (data->flags & MMC_DATA_READ)
+			c |= TRANSFER_READ;
+	}
+
+	tmio_mmc_enable_mmc_irqs(host, TMIO_MASK_CMD);
+
+	/* Fire off the command */
+	sd_ctrl_write32_as_16_and_16(host, CTL_ARG_REG, cmd->arg);
+	sd_ctrl_write16(host, CTL_SD_CMD, c);
+
+	return 0;
+}
+
+static void tmio_mmc_transfer_data(struct tmio_mmc_host *host,
+				   unsigned short *buf,
+				   unsigned int count)
+{
+	int is_read = host->data->flags & MMC_DATA_READ;
+	u8  *buf8;
+
+	/*
+	 * Transfer the data
+	 */
+	if (host->pdata->flags & TMIO_MMC_32BIT_DATA_PORT) {
+		u32 data = 0;
+		u32 *buf32 = (u32 *)buf;
+
+		if (is_read)
+			sd_ctrl_read32_rep(host, CTL_SD_DATA_PORT, buf32,
+					   count >> 2);
+		else
+			sd_ctrl_write32_rep(host, CTL_SD_DATA_PORT, buf32,
+					    count >> 2);
+
+		/* if count was multiple of 4 */
+		if (!(count & 0x3))
+			return;
+
+		buf32 += count >> 2;
+		count %= 4;
+
+		if (is_read) {
+			sd_ctrl_read32_rep(host, CTL_SD_DATA_PORT, &data, 1);
+			memcpy(buf32, &data, count);
+		} else {
+			memcpy(&data, buf32, count);
+			sd_ctrl_write32_rep(host, CTL_SD_DATA_PORT, &data, 1);
+		}
+
+		return;
+	}
+
+	if (is_read)
+		sd_ctrl_read16_rep(host, CTL_SD_DATA_PORT, buf, count >> 1);
+	else
+		sd_ctrl_write16_rep(host, CTL_SD_DATA_PORT, buf, count >> 1);
+
+	/* if count was even number */
+	if (!(count & 0x1))
+		return;
+
+	/* if count was odd number */
+	buf8 = (u8 *)(buf + (count >> 1));
+
+	/*
+	 * FIXME
+	 *
+	 * driver and this function are assuming that
+	 * it is used as little endian
+	 */
+	if (is_read)
+		*buf8 = sd_ctrl_read16(host, CTL_SD_DATA_PORT) & 0xff;
+	else
+		sd_ctrl_write16(host, CTL_SD_DATA_PORT, *buf8);
+}
+
+/*
+ * This chip always returns (at least?) as much data as you ask for.
+ * I'm unsure what happens if you ask for less than a block. This should be
+ * looked into to ensure that a funny length read doesn't hose the controller.
+ */
+static void tmio_mmc_pio_irq(struct tmio_mmc_host *host)
+{
+	struct mmc_data *data = host->data;
+	void *sg_virt;
+	unsigned short *buf;
+	unsigned int count;
+	unsigned long flags;
+
+	if (host->dma_on) {
+		pr_err("PIO IRQ in DMA mode!\n");
+		return;
+	} else if (!data) {
+		pr_debug("Spurious PIO IRQ\n");
+		return;
+	}
+
+	sg_virt = tmio_mmc_kmap_atomic(host->sg_ptr, &flags);
+	buf = (unsigned short *)(sg_virt + host->sg_off);
+
+	count = host->sg_ptr->length - host->sg_off;
+	if (count > data->blksz)
+		count = data->blksz;
+
+	pr_debug("count: %08x offset: %08x flags %08x\n",
+		 count, host->sg_off, data->flags);
+
+	/* Transfer the data */
+	tmio_mmc_transfer_data(host, buf, count);
+
+	host->sg_off += count;
+
+	tmio_mmc_kunmap_atomic(host->sg_ptr, &flags, sg_virt);
+
+	if (host->sg_off == host->sg_ptr->length)
+		tmio_mmc_next_sg(host);
+}
+
+static void tmio_mmc_check_bounce_buffer(struct tmio_mmc_host *host)
+{
+	if (host->sg_ptr == &host->bounce_sg) {
+		unsigned long flags;
+		void *sg_vaddr = tmio_mmc_kmap_atomic(host->sg_orig, &flags);
+
+		memcpy(sg_vaddr, host->bounce_buf, host->bounce_sg.length);
+		tmio_mmc_kunmap_atomic(host->sg_orig, &flags, sg_vaddr);
+	}
+}
+
+/* needs to be called with host->lock held */
+void tmio_mmc_do_data_irq(struct tmio_mmc_host *host)
+{
+	struct mmc_data *data = host->data;
+	struct mmc_command *stop;
+
+	host->data = NULL;
+
+	if (!data) {
+		dev_warn(&host->pdev->dev, "Spurious data end IRQ\n");
+		return;
+	}
+	stop = data->stop;
+
+	/* FIXME - return correct transfer count on errors */
+	if (!data->error)
+		data->bytes_xfered = data->blocks * data->blksz;
+	else
+		data->bytes_xfered = 0;
+
+	pr_debug("Completed data request\n");
+
+	/*
+	 * FIXME: other drivers allow an optional stop command of any given type
+	 *        which we dont do, as the chip can auto generate them.
+	 *        Perhaps we can be smarter about when to use auto CMD12 and
+	 *        only issue the auto request when we know this is the desired
+	 *        stop command, allowing fallback to the stop command the
+	 *        upper layers expect. For now, we do what works.
+	 */
+
+	if (data->flags & MMC_DATA_READ) {
+		if (host->dma_on)
+			tmio_mmc_check_bounce_buffer(host);
+		dev_dbg(&host->pdev->dev, "Complete Rx request %p\n",
+			host->mrq);
+	} else {
+		dev_dbg(&host->pdev->dev, "Complete Tx request %p\n",
+			host->mrq);
+	}
+
+	if (stop && !host->mrq->sbc) {
+		if (stop->opcode != MMC_STOP_TRANSMISSION || stop->arg)
+			dev_err(&host->pdev->dev, "unsupported stop: CMD%u,0x%x. We did CMD12,0\n",
+				stop->opcode, stop->arg);
+
+		/* fill in response from auto CMD12 */
+		stop->resp[0] = sd_ctrl_read16_and_16_as_32(host, CTL_RESPONSE);
+
+		sd_ctrl_write16(host, CTL_STOP_INTERNAL_ACTION, 0);
+	}
+
+	schedule_work(&host->done);
+}
+EXPORT_SYMBOL_GPL(tmio_mmc_do_data_irq);
+
+static void tmio_mmc_data_irq(struct tmio_mmc_host *host, unsigned int stat)
+{
+	struct mmc_data *data;
+
+	spin_lock(&host->lock);
+	data = host->data;
+
+	if (!data)
+		goto out;
+
+	if (stat & TMIO_STAT_CRCFAIL || stat & TMIO_STAT_STOPBIT_ERR ||
+	    stat & TMIO_STAT_TXUNDERRUN)
+		data->error = -EILSEQ;
+	if (host->dma_on && (data->flags & MMC_DATA_WRITE)) {
+		u32 status = sd_ctrl_read16_and_16_as_32(host, CTL_STATUS);
+		bool done = false;
+
+		/*
+		 * Has all data been written out yet? Testing on SuperH showed,
+		 * that in most cases the first interrupt comes already with the
+		 * BUSY status bit clear, but on some operations, like mount or
+		 * in the beginning of a write / sync / umount, there is one
+		 * DATAEND interrupt with the BUSY bit set, in this cases
+		 * waiting for one more interrupt fixes the problem.
+		 */
+		if (host->pdata->flags & TMIO_MMC_HAS_IDLE_WAIT) {
+			if (status & TMIO_STAT_SCLKDIVEN)
+				done = true;
+		} else {
+			if (!(status & TMIO_STAT_CMD_BUSY))
+				done = true;
+		}
+
+		if (done) {
+			tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_DATAEND);
+			tmio_mmc_dataend_dma(host);
+		}
+	} else if (host->dma_on && (data->flags & MMC_DATA_READ)) {
+		tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_DATAEND);
+		tmio_mmc_dataend_dma(host);
+	} else {
+		tmio_mmc_do_data_irq(host);
+		tmio_mmc_disable_mmc_irqs(host, TMIO_MASK_READOP | TMIO_MASK_WRITEOP);
+	}
+out:
+	spin_unlock(&host->lock);
+}
+
+static void tmio_mmc_cmd_irq(struct tmio_mmc_host *host, unsigned int stat)
+{
+	struct mmc_command *cmd = host->cmd;
+	int i, addr;
+
+	spin_lock(&host->lock);
+
+	if (!host->cmd) {
+		pr_debug("Spurious CMD irq\n");
+		goto out;
+	}
+
+	/* This controller is sicker than the PXA one. Not only do we need to
+	 * drop the top 8 bits of the first response word, we also need to
+	 * modify the order of the response for short response command types.
+	 */
+
+	for (i = 3, addr = CTL_RESPONSE ; i >= 0 ; i--, addr += 4)
+		cmd->resp[i] = sd_ctrl_read16_and_16_as_32(host, addr);
+
+	if (cmd->flags &  MMC_RSP_136) {
+		cmd->resp[0] = (cmd->resp[0] << 8) | (cmd->resp[1] >> 24);
+		cmd->resp[1] = (cmd->resp[1] << 8) | (cmd->resp[2] >> 24);
+		cmd->resp[2] = (cmd->resp[2] << 8) | (cmd->resp[3] >> 24);
+		cmd->resp[3] <<= 8;
+	} else if (cmd->flags & MMC_RSP_R3) {
+		cmd->resp[0] = cmd->resp[3];
+	}
+
+	if (stat & TMIO_STAT_CMDTIMEOUT)
+		cmd->error = -ETIMEDOUT;
+	else if ((stat & TMIO_STAT_CRCFAIL && cmd->flags & MMC_RSP_CRC) ||
+		 stat & TMIO_STAT_STOPBIT_ERR ||
+		 stat & TMIO_STAT_CMD_IDX_ERR)
+		cmd->error = -EILSEQ;
+
+	/* If there is data to handle we enable data IRQs here, and
+	 * we will ultimatley finish the request in the data_end handler.
+	 * If theres no data or we encountered an error, finish now.
+	 */
+	if (host->data && (!cmd->error || cmd->error == -EILSEQ)) {
+		if (host->data->flags & MMC_DATA_READ) {
+			if (!host->dma_on) {
+				tmio_mmc_enable_mmc_irqs(host, TMIO_MASK_READOP);
+			} else {
+				tmio_mmc_disable_mmc_irqs(host,
+							  TMIO_MASK_READOP);
+				tasklet_schedule(&host->dma_issue);
+			}
+		} else {
+			if (!host->dma_on) {
+				tmio_mmc_enable_mmc_irqs(host, TMIO_MASK_WRITEOP);
+			} else {
+				tmio_mmc_disable_mmc_irqs(host,
+							  TMIO_MASK_WRITEOP);
+				tasklet_schedule(&host->dma_issue);
+			}
+		}
+	} else {
+		schedule_work(&host->done);
+	}
+
+out:
+	spin_unlock(&host->lock);
+}
+
+static bool __tmio_mmc_card_detect_irq(struct tmio_mmc_host *host,
+				       int ireg, int status)
+{
+	struct mmc_host *mmc = host->mmc;
+
+	/* Card insert / remove attempts */
+	if (ireg & (TMIO_STAT_CARD_INSERT | TMIO_STAT_CARD_REMOVE)) {
+		tmio_mmc_ack_mmc_irqs(host, TMIO_STAT_CARD_INSERT |
+			TMIO_STAT_CARD_REMOVE);
+		if ((((ireg & TMIO_STAT_CARD_REMOVE) && mmc->card) ||
+		     ((ireg & TMIO_STAT_CARD_INSERT) && !mmc->card)) &&
+		    !work_pending(&mmc->detect.work))
+			mmc_detect_change(host->mmc, msecs_to_jiffies(100));
+		return true;
+	}
+
+	return false;
+}
+
+static bool __tmio_mmc_sdcard_irq(struct tmio_mmc_host *host, int ireg,
+				  int status)
+{
+	/* Command completion */
+	if (ireg & (TMIO_STAT_CMDRESPEND | TMIO_STAT_CMDTIMEOUT)) {
+		tmio_mmc_ack_mmc_irqs(host, TMIO_STAT_CMDRESPEND |
+				      TMIO_STAT_CMDTIMEOUT);
+		tmio_mmc_cmd_irq(host, status);
+		return true;
+	}
+
+	/* Data transfer */
+	if (ireg & (TMIO_STAT_RXRDY | TMIO_STAT_TXRQ)) {
+		tmio_mmc_ack_mmc_irqs(host, TMIO_STAT_RXRDY | TMIO_STAT_TXRQ);
+		tmio_mmc_pio_irq(host);
+		return true;
+	}
+
+	/* Data transfer completion */
+	if (ireg & TMIO_STAT_DATAEND) {
+		tmio_mmc_ack_mmc_irqs(host, TMIO_STAT_DATAEND);
+		tmio_mmc_data_irq(host, status);
+		return true;
+	}
+
+	return false;
+}
+
+static bool __tmio_mmc_sdio_irq(struct tmio_mmc_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	struct tmio_mmc_data *pdata = host->pdata;
+	unsigned int ireg, status;
+	unsigned int sdio_status;
+
+	if (!(pdata->flags & TMIO_MMC_SDIO_IRQ))
+		return false;
+
+	status = sd_ctrl_read16(host, CTL_SDIO_STATUS);
+	ireg = status & TMIO_SDIO_MASK_ALL & ~host->sdio_irq_mask;
+
+	sdio_status = status & ~TMIO_SDIO_MASK_ALL;
+	if (pdata->flags & TMIO_MMC_SDIO_STATUS_SETBITS)
+		sdio_status |= TMIO_SDIO_SETBITS_MASK;
+
+	sd_ctrl_write16(host, CTL_SDIO_STATUS, sdio_status);
+
+	if (mmc->caps & MMC_CAP_SDIO_IRQ && ireg & TMIO_SDIO_STAT_IOIRQ)
+		mmc_signal_sdio_irq(mmc);
+
+	return ireg;
+}
+
+irqreturn_t tmio_mmc_irq(int irq, void *devid)
+{
+	struct tmio_mmc_host *host = devid;
+	unsigned int ireg, status;
+
+	status = sd_ctrl_read16_and_16_as_32(host, CTL_STATUS);
+	ireg = status & TMIO_MASK_IRQ & ~host->sdcard_irq_mask;
+
+	/* Clear the status except the interrupt status */
+	sd_ctrl_write32_as_16_and_16(host, CTL_STATUS, TMIO_MASK_IRQ);
+
+	if (__tmio_mmc_card_detect_irq(host, ireg, status))
+		return IRQ_HANDLED;
+	if (__tmio_mmc_sdcard_irq(host, ireg, status))
+		return IRQ_HANDLED;
+
+	if (__tmio_mmc_sdio_irq(host))
+		return IRQ_HANDLED;
+
+	return IRQ_NONE;
+}
+EXPORT_SYMBOL_GPL(tmio_mmc_irq);
+
+static int tmio_mmc_start_data(struct tmio_mmc_host *host,
+			       struct mmc_data *data)
+{
+	struct tmio_mmc_data *pdata = host->pdata;
+
+	pr_debug("setup data transfer: blocksize %08x  nr_blocks %d\n",
+		 data->blksz, data->blocks);
+
+	/* Some hardware cannot perform 2 byte requests in 4/8 bit mode */
+	if (host->mmc->ios.bus_width == MMC_BUS_WIDTH_4 ||
+	    host->mmc->ios.bus_width == MMC_BUS_WIDTH_8) {
+		int blksz_2bytes = pdata->flags & TMIO_MMC_BLKSZ_2BYTES;
+
+		if (data->blksz < 2 || (data->blksz < 4 && !blksz_2bytes)) {
+			pr_err("%s: %d byte block unsupported in 4/8 bit mode\n",
+			       mmc_hostname(host->mmc), data->blksz);
+			return -EINVAL;
+		}
+	}
+
+	tmio_mmc_init_sg(host, data);
+	host->data = data;
+	host->dma_on = false;
+
+	/* Set transfer length / blocksize */
+	sd_ctrl_write16(host, CTL_SD_XFER_LEN, data->blksz);
+	if (host->mmc->max_blk_count >= SZ_64K)
+		sd_ctrl_write32(host, CTL_XFER_BLK_COUNT, data->blocks);
+	else
+		sd_ctrl_write16(host, CTL_XFER_BLK_COUNT, data->blocks);
+
+	tmio_mmc_start_dma(host, data);
+
+	return 0;
+}
+
+static void tmio_process_mrq(struct tmio_mmc_host *host,
+			     struct mmc_request *mrq)
+{
+	struct mmc_command *cmd;
+	int ret;
+
+	if (mrq->sbc && host->cmd != mrq->sbc) {
+		cmd = mrq->sbc;
+	} else {
+		cmd = mrq->cmd;
+		if (mrq->data) {
+			ret = tmio_mmc_start_data(host, mrq->data);
+			if (ret)
+				goto fail;
+		}
+	}
+
+	ret = tmio_mmc_start_command(host, cmd);
+	if (ret)
+		goto fail;
+
+	schedule_delayed_work(&host->delayed_reset_work,
+			      msecs_to_jiffies(CMDREQ_TIMEOUT));
+	return;
+
+fail:
+	host->mrq = NULL;
+	mrq->cmd->error = ret;
+	mmc_request_done(host->mmc, mrq);
+}
+
+/* Process requests from the MMC layer */
+static void tmio_mmc_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (host->mrq) {
+		pr_debug("request not null\n");
+		if (IS_ERR(host->mrq)) {
+			spin_unlock_irqrestore(&host->lock, flags);
+			mrq->cmd->error = -EAGAIN;
+			mmc_request_done(mmc, mrq);
+			return;
+		}
+	}
+
+	host->last_req_ts = jiffies;
+	wmb();
+	host->mrq = mrq;
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	tmio_process_mrq(host, mrq);
+}
+
+static void tmio_mmc_finish_request(struct tmio_mmc_host *host)
+{
+	struct mmc_request *mrq;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	tmio_mmc_end_dma(host);
+
+	mrq = host->mrq;
+	if (IS_ERR_OR_NULL(mrq)) {
+		spin_unlock_irqrestore(&host->lock, flags);
+		return;
+	}
+
+	/* If not SET_BLOCK_COUNT, clear old data */
+	if (host->cmd != mrq->sbc) {
+		host->cmd = NULL;
+		host->data = NULL;
+		host->mrq = NULL;
+	}
+
+	cancel_delayed_work(&host->delayed_reset_work);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (mrq->cmd->error || (mrq->data && mrq->data->error))
+		tmio_mmc_abort_dma(host);
+
+	/* Error means retune, but executed command was still successful */
+	if (host->check_retune && host->check_retune(host))
+		mmc_retune_needed(host->mmc);
+
+	/* If SET_BLOCK_COUNT, continue with main command */
+	if (host->mrq && !mrq->cmd->error) {
+		tmio_process_mrq(host, mrq);
+		return;
+	}
+
+	if (host->fixup_request)
+		host->fixup_request(host, mrq);
+
+	mmc_request_done(host->mmc, mrq);
+}
+
+static void tmio_mmc_done_work(struct work_struct *work)
+{
+	struct tmio_mmc_host *host = container_of(work, struct tmio_mmc_host,
+						  done);
+	tmio_mmc_finish_request(host);
+}
+
+static void tmio_mmc_power_on(struct tmio_mmc_host *host, unsigned short vdd)
+{
+	struct mmc_host *mmc = host->mmc;
+	int ret = 0;
+
+	/* .set_ios() is returning void, so, no chance to report an error */
+
+	if (host->set_pwr)
+		host->set_pwr(host->pdev, 1);
+
+	if (!IS_ERR(mmc->supply.vmmc)) {
+		ret = mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, vdd);
+		/*
+		 * Attention: empiric value. With a b43 WiFi SDIO card this
+		 * delay proved necessary for reliable card-insertion probing.
+		 * 100us were not enough. Is this the same 140us delay, as in
+		 * tmio_mmc_set_ios()?
+		 */
+		usleep_range(200, 300);
+	}
+	/*
+	 * It seems, VccQ should be switched on after Vcc, this is also what the
+	 * omap_hsmmc.c driver does.
+	 */
+	if (!IS_ERR(mmc->supply.vqmmc) && !ret) {
+		ret = regulator_enable(mmc->supply.vqmmc);
+		usleep_range(200, 300);
+	}
+
+	if (ret < 0)
+		dev_dbg(&host->pdev->dev, "Regulators failed to power up: %d\n",
+			ret);
+}
+
+static void tmio_mmc_power_off(struct tmio_mmc_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+
+	if (!IS_ERR(mmc->supply.vqmmc))
+		regulator_disable(mmc->supply.vqmmc);
+
+	if (!IS_ERR(mmc->supply.vmmc))
+		mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+
+	if (host->set_pwr)
+		host->set_pwr(host->pdev, 0);
+}
+
+static void tmio_mmc_set_bus_width(struct tmio_mmc_host *host,
+				   unsigned char bus_width)
+{
+	u16 reg = sd_ctrl_read16(host, CTL_SD_MEM_CARD_OPT)
+				& ~(CARD_OPT_WIDTH | CARD_OPT_WIDTH8);
+
+	/* reg now applies to MMC_BUS_WIDTH_4 */
+	if (bus_width == MMC_BUS_WIDTH_1)
+		reg |= CARD_OPT_WIDTH;
+	else if (bus_width == MMC_BUS_WIDTH_8)
+		reg |= CARD_OPT_WIDTH8;
+
+	sd_ctrl_write16(host, CTL_SD_MEM_CARD_OPT, reg);
+}
+
+/* Set MMC clock / power.
+ * Note: This controller uses a simple divider scheme therefore it cannot
+ * run a MMC card at full speed (20MHz). The max clock is 24MHz on SD, but as
+ * MMC wont run that fast, it has to be clocked at 12MHz which is the next
+ * slowest setting.
+ */
+static void tmio_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+	struct device *dev = &host->pdev->dev;
+	unsigned long flags;
+
+	mutex_lock(&host->ios_lock);
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (host->mrq) {
+		if (IS_ERR(host->mrq)) {
+			dev_dbg(dev,
+				"%s.%d: concurrent .set_ios(), clk %u, mode %u\n",
+				current->comm, task_pid_nr(current),
+				ios->clock, ios->power_mode);
+			host->mrq = ERR_PTR(-EINTR);
+		} else {
+			dev_dbg(dev,
+				"%s.%d: CMD%u active since %lu, now %lu!\n",
+				current->comm, task_pid_nr(current),
+				host->mrq->cmd->opcode, host->last_req_ts,
+				jiffies);
+		}
+		spin_unlock_irqrestore(&host->lock, flags);
+
+		mutex_unlock(&host->ios_lock);
+		return;
+	}
+
+	host->mrq = ERR_PTR(-EBUSY);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	switch (ios->power_mode) {
+	case MMC_POWER_OFF:
+		tmio_mmc_power_off(host);
+		/* For R-Car Gen2+, we need to reset SDHI specific SCC */
+		if (host->pdata->flags & TMIO_MMC_MIN_RCAR2)
+			host->reset(host);
+		host->set_clock(host, 0);
+		break;
+	case MMC_POWER_UP:
+		tmio_mmc_power_on(host, ios->vdd);
+		host->set_clock(host, ios->clock);
+		tmio_mmc_set_bus_width(host, ios->bus_width);
+		break;
+	case MMC_POWER_ON:
+		host->set_clock(host, ios->clock);
+		tmio_mmc_set_bus_width(host, ios->bus_width);
+		break;
+	}
+
+	/* Let things settle. delay taken from winCE driver */
+	usleep_range(140, 200);
+	if (PTR_ERR(host->mrq) == -EINTR)
+		dev_dbg(&host->pdev->dev,
+			"%s.%d: IOS interrupted: clk %u, mode %u",
+			current->comm, task_pid_nr(current),
+			ios->clock, ios->power_mode);
+	host->mrq = NULL;
+
+	host->clk_cache = ios->clock;
+
+	mutex_unlock(&host->ios_lock);
+}
+
+static int tmio_mmc_get_ro(struct mmc_host *mmc)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+
+	return !(sd_ctrl_read16_and_16_as_32(host, CTL_STATUS) &
+		 TMIO_STAT_WRPROTECT);
+}
+
+static int tmio_mmc_get_cd(struct mmc_host *mmc)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+
+	return !!(sd_ctrl_read16_and_16_as_32(host, CTL_STATUS) &
+		  TMIO_STAT_SIGSTATE);
+}
+
+static int tmio_multi_io_quirk(struct mmc_card *card,
+			       unsigned int direction, int blk_size)
+{
+	struct tmio_mmc_host *host = mmc_priv(card->host);
+
+	if (host->multi_io_quirk)
+		return host->multi_io_quirk(card, direction, blk_size);
+
+	return blk_size;
+}
+
+static struct mmc_host_ops tmio_mmc_ops = {
+	.request	= tmio_mmc_request,
+	.set_ios	= tmio_mmc_set_ios,
+	.get_ro         = tmio_mmc_get_ro,
+	.get_cd		= tmio_mmc_get_cd,
+	.enable_sdio_irq = tmio_mmc_enable_sdio_irq,
+	.multi_io_quirk	= tmio_multi_io_quirk,
+};
+
+static int tmio_mmc_init_ocr(struct tmio_mmc_host *host)
+{
+	struct tmio_mmc_data *pdata = host->pdata;
+	struct mmc_host *mmc = host->mmc;
+	int err;
+
+	err = mmc_regulator_get_supply(mmc);
+	if (err)
+		return err;
+
+	/* use ocr_mask if no regulator */
+	if (!mmc->ocr_avail)
+		mmc->ocr_avail = pdata->ocr_mask;
+
+	/*
+	 * try again.
+	 * There is possibility that regulator has not been probed
+	 */
+	if (!mmc->ocr_avail)
+		return -EPROBE_DEFER;
+
+	return 0;
+}
+
+static void tmio_mmc_of_parse(struct platform_device *pdev,
+			      struct mmc_host *mmc)
+{
+	const struct device_node *np = pdev->dev.of_node;
+
+	if (!np)
+		return;
+
+	/*
+	 * DEPRECATED:
+	 * For new platforms, please use "disable-wp" instead of
+	 * "toshiba,mmc-wrprotect-disable"
+	 */
+	if (of_get_property(np, "toshiba,mmc-wrprotect-disable", NULL))
+		mmc->caps2 |= MMC_CAP2_NO_WRITE_PROTECT;
+}
+
+struct tmio_mmc_host *tmio_mmc_host_alloc(struct platform_device *pdev,
+					  struct tmio_mmc_data *pdata)
+{
+	struct tmio_mmc_host *host;
+	struct mmc_host *mmc;
+	void __iomem *ctl;
+	int ret;
+
+	ctl = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(ctl))
+		return ERR_CAST(ctl);
+
+	mmc = mmc_alloc_host(sizeof(struct tmio_mmc_host), &pdev->dev);
+	if (!mmc)
+		return ERR_PTR(-ENOMEM);
+
+	host = mmc_priv(mmc);
+	host->ctl = ctl;
+	host->mmc = mmc;
+	host->pdev = pdev;
+	host->pdata = pdata;
+	host->ops = tmio_mmc_ops;
+	mmc->ops = &host->ops;
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret) {
+		host = ERR_PTR(ret);
+		goto free;
+	}
+
+	tmio_mmc_of_parse(pdev, mmc);
+
+	platform_set_drvdata(pdev, host);
+
+	return host;
+free:
+	mmc_free_host(mmc);
+
+	return host;
+}
+EXPORT_SYMBOL_GPL(tmio_mmc_host_alloc);
+
+void tmio_mmc_host_free(struct tmio_mmc_host *host)
+{
+	mmc_free_host(host->mmc);
+}
+EXPORT_SYMBOL_GPL(tmio_mmc_host_free);
+
+int tmio_mmc_host_probe(struct tmio_mmc_host *_host)
+{
+	struct platform_device *pdev = _host->pdev;
+	struct tmio_mmc_data *pdata = _host->pdata;
+	struct mmc_host *mmc = _host->mmc;
+	int ret;
+
+	/*
+	 * Check the sanity of mmc->f_min to prevent host->set_clock() from
+	 * looping forever...
+	 */
+	if (mmc->f_min == 0)
+		return -EINVAL;
+
+	if (!(pdata->flags & TMIO_MMC_HAS_IDLE_WAIT))
+		_host->write16_hook = NULL;
+
+	_host->set_pwr = pdata->set_pwr;
+
+	ret = tmio_mmc_init_ocr(_host);
+	if (ret < 0)
+		return ret;
+
+	/*
+	 * Look for a card detect GPIO, if it fails with anything
+	 * else than a probe deferral, just live without it.
+	 */
+	ret = mmc_gpiod_request_cd(mmc, "cd", 0, false, 0);
+	if (ret == -EPROBE_DEFER)
+		return ret;
+
+	mmc->caps |= MMC_CAP_4_BIT_DATA | pdata->capabilities;
+	mmc->caps2 |= pdata->capabilities2;
+	mmc->max_segs = pdata->max_segs ? : 32;
+	mmc->max_blk_size = TMIO_MAX_BLK_SIZE;
+	mmc->max_blk_count = pdata->max_blk_count ? :
+		(PAGE_SIZE / mmc->max_blk_size) * mmc->max_segs;
+	mmc->max_req_size = min_t(size_t,
+				  mmc->max_blk_size * mmc->max_blk_count,
+				  dma_max_mapping_size(&pdev->dev));
+	mmc->max_seg_size = mmc->max_req_size;
+
+	if (mmc_can_gpio_ro(mmc))
+		_host->ops.get_ro = mmc_gpio_get_ro;
+
+	if (mmc_can_gpio_cd(mmc))
+		_host->ops.get_cd = mmc_gpio_get_cd;
+
+	_host->native_hotplug = !(mmc_can_gpio_cd(mmc) ||
+				  mmc->caps & MMC_CAP_NEEDS_POLL ||
+				  !mmc_card_is_removable(mmc));
+
+	/*
+	 * On Gen2+, eMMC with NONREMOVABLE currently fails because native
+	 * hotplug gets disabled. It seems RuntimePM related yet we need further
+	 * research. Since we are planning a PM overhaul anyway, let's enforce
+	 * for now the device being active by enabling native hotplug always.
+	 */
+	if (pdata->flags & TMIO_MMC_MIN_RCAR2)
+		_host->native_hotplug = true;
+
+	/*
+	 * While using internal tmio hardware logic for card detection, we need
+	 * to ensure it stays powered for it to work.
+	 */
+	if (_host->native_hotplug)
+		pm_runtime_get_noresume(&pdev->dev);
+
+	_host->sdio_irq_enabled = false;
+	if (pdata->flags & TMIO_MMC_SDIO_IRQ)
+		_host->sdio_irq_mask = TMIO_SDIO_MASK_ALL;
+
+	_host->set_clock(_host, 0);
+	tmio_mmc_reset(_host);
+
+	_host->sdcard_irq_mask = sd_ctrl_read16_and_16_as_32(_host, CTL_IRQ_MASK);
+	if (!_host->sdcard_irq_mask_all)
+		_host->sdcard_irq_mask_all = TMIO_MASK_ALL;
+	tmio_mmc_disable_mmc_irqs(_host, _host->sdcard_irq_mask_all);
+
+	if (_host->native_hotplug)
+		tmio_mmc_enable_mmc_irqs(_host,
+				TMIO_STAT_CARD_REMOVE | TMIO_STAT_CARD_INSERT);
+
+	spin_lock_init(&_host->lock);
+	mutex_init(&_host->ios_lock);
+
+	/* Init delayed work for request timeouts */
+	INIT_DELAYED_WORK(&_host->delayed_reset_work, tmio_mmc_reset_work);
+	INIT_WORK(&_host->done, tmio_mmc_done_work);
+
+	/* See if we also get DMA */
+	tmio_mmc_request_dma(_host, pdata);
+
+	pm_runtime_get_noresume(&pdev->dev);
+	pm_runtime_set_active(&pdev->dev);
+	pm_runtime_set_autosuspend_delay(&pdev->dev, 50);
+	pm_runtime_use_autosuspend(&pdev->dev);
+	pm_runtime_enable(&pdev->dev);
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto remove_host;
+
+	dev_pm_qos_expose_latency_limit(&pdev->dev, 100);
+	pm_runtime_put(&pdev->dev);
+
+	return 0;
+
+remove_host:
+	pm_runtime_put_noidle(&pdev->dev);
+	tmio_mmc_host_remove(_host);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(tmio_mmc_host_probe);
+
+void tmio_mmc_host_remove(struct tmio_mmc_host *host)
+{
+	struct platform_device *pdev = host->pdev;
+	struct mmc_host *mmc = host->mmc;
+
+	pm_runtime_get_sync(&pdev->dev);
+
+	if (host->pdata->flags & TMIO_MMC_SDIO_IRQ)
+		sd_ctrl_write16(host, CTL_TRANSACTION_CTL, 0x0000);
+
+	dev_pm_qos_hide_latency_limit(&pdev->dev);
+
+	mmc_remove_host(mmc);
+	cancel_work_sync(&host->done);
+	cancel_delayed_work_sync(&host->delayed_reset_work);
+	tmio_mmc_release_dma(host);
+	tmio_mmc_disable_mmc_irqs(host, host->sdcard_irq_mask_all);
+
+	if (host->native_hotplug)
+		pm_runtime_put_noidle(&pdev->dev);
+
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_dont_use_autosuspend(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+}
+EXPORT_SYMBOL_GPL(tmio_mmc_host_remove);
+
+#ifdef CONFIG_PM
+static int tmio_mmc_clk_enable(struct tmio_mmc_host *host)
+{
+	if (!host->clk_enable)
+		return -ENOTSUPP;
+
+	return host->clk_enable(host);
+}
+
+static void tmio_mmc_clk_disable(struct tmio_mmc_host *host)
+{
+	if (host->clk_disable)
+		host->clk_disable(host);
+}
+
+int tmio_mmc_host_runtime_suspend(struct device *dev)
+{
+	struct tmio_mmc_host *host = dev_get_drvdata(dev);
+
+	tmio_mmc_disable_mmc_irqs(host, host->sdcard_irq_mask_all);
+
+	if (host->clk_cache)
+		host->set_clock(host, 0);
+
+	tmio_mmc_clk_disable(host);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(tmio_mmc_host_runtime_suspend);
+
+int tmio_mmc_host_runtime_resume(struct device *dev)
+{
+	struct tmio_mmc_host *host = dev_get_drvdata(dev);
+
+	tmio_mmc_clk_enable(host);
+	tmio_mmc_reset(host);
+
+	if (host->clk_cache)
+		host->set_clock(host, host->clk_cache);
+
+	if (host->native_hotplug)
+		tmio_mmc_enable_mmc_irqs(host,
+				TMIO_STAT_CARD_REMOVE | TMIO_STAT_CARD_INSERT);
+
+	tmio_mmc_enable_dma(host, true);
+
+	mmc_retune_needed(host->mmc);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(tmio_mmc_host_runtime_resume);
+#endif
+
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/toshsd.c b/drivers/mmc/host/toshsd.c
new file mode 100644
index 000000000..497791ffa
--- /dev/null
+++ b/drivers/mmc/host/toshsd.c
@@ -0,0 +1,709 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ *  Toshiba PCI Secure Digital Host Controller Interface driver
+ *
+ *  Copyright (C) 2014 Ondrej Zary
+ *  Copyright (C) 2007 Richard Betts, All Rights Reserved.
+ *
+ *	Based on asic3_mmc.c, copyright (c) 2005 SDG Systems, LLC and,
+ *	sdhci.c, copyright (C) 2005-2006 Pierre Ossman
+ */
+
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/scatterlist.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+
+#include "toshsd.h"
+
+#define DRIVER_NAME "toshsd"
+
+static const struct pci_device_id pci_ids[] = {
+	{ PCI_DEVICE(PCI_VENDOR_ID_TOSHIBA, 0x0805) },
+	{ /* end: all zeroes */ },
+};
+
+MODULE_DEVICE_TABLE(pci, pci_ids);
+
+static void toshsd_init(struct toshsd_host *host)
+{
+	/* enable clock */
+	pci_write_config_byte(host->pdev, SD_PCICFG_CLKSTOP,
+					SD_PCICFG_CLKSTOP_ENABLE_ALL);
+	pci_write_config_byte(host->pdev, SD_PCICFG_CARDDETECT, 2);
+
+	/* reset */
+	iowrite16(0, host->ioaddr + SD_SOFTWARERESET); /* assert */
+	mdelay(2);
+	iowrite16(1, host->ioaddr + SD_SOFTWARERESET); /* deassert */
+	mdelay(2);
+
+	/* Clear card registers */
+	iowrite16(0, host->ioaddr + SD_CARDCLOCKCTRL);
+	iowrite32(0, host->ioaddr + SD_CARDSTATUS);
+	iowrite32(0, host->ioaddr + SD_ERRORSTATUS0);
+	iowrite16(0, host->ioaddr + SD_STOPINTERNAL);
+
+	/* SDIO clock? */
+	iowrite16(0x100, host->ioaddr + SDIO_BASE + SDIO_CLOCKNWAITCTRL);
+
+	/* enable LED */
+	pci_write_config_byte(host->pdev, SD_PCICFG_SDLED_ENABLE1,
+					SD_PCICFG_LED_ENABLE1_START);
+	pci_write_config_byte(host->pdev, SD_PCICFG_SDLED_ENABLE2,
+					SD_PCICFG_LED_ENABLE2_START);
+
+	/* set interrupt masks */
+	iowrite32(~(u32)(SD_CARD_RESP_END | SD_CARD_RW_END
+			| SD_CARD_CARD_REMOVED_0 | SD_CARD_CARD_INSERTED_0
+			| SD_BUF_READ_ENABLE | SD_BUF_WRITE_ENABLE
+			| SD_BUF_CMD_TIMEOUT),
+			host->ioaddr + SD_INTMASKCARD);
+
+	iowrite16(0x1000, host->ioaddr + SD_TRANSACTIONCTRL);
+}
+
+/* Set MMC clock / power.
+ * Note: This controller uses a simple divider scheme therefore it cannot run
+ * SD/MMC cards at full speed (24/20MHz). HCLK (=33MHz PCI clock?) is too high
+ * and the next slowest is 16MHz (div=2).
+ */
+static void __toshsd_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct toshsd_host *host = mmc_priv(mmc);
+
+	if (ios->clock) {
+		u16 clk;
+		int div = 1;
+
+		while (ios->clock < HCLK / div)
+			div *= 2;
+
+		clk = div >> 2;
+
+		if (div == 1) { /* disable the divider */
+			pci_write_config_byte(host->pdev, SD_PCICFG_CLKMODE,
+					      SD_PCICFG_CLKMODE_DIV_DISABLE);
+			clk |= SD_CARDCLK_DIV_DISABLE;
+		} else
+			pci_write_config_byte(host->pdev, SD_PCICFG_CLKMODE, 0);
+
+		clk |= SD_CARDCLK_ENABLE_CLOCK;
+		iowrite16(clk, host->ioaddr + SD_CARDCLOCKCTRL);
+
+		mdelay(10);
+	} else
+		iowrite16(0, host->ioaddr + SD_CARDCLOCKCTRL);
+
+	switch (ios->power_mode) {
+	case MMC_POWER_OFF:
+		pci_write_config_byte(host->pdev, SD_PCICFG_POWER1,
+					SD_PCICFG_PWR1_OFF);
+		mdelay(1);
+		break;
+	case MMC_POWER_UP:
+		break;
+	case MMC_POWER_ON:
+		pci_write_config_byte(host->pdev, SD_PCICFG_POWER1,
+					SD_PCICFG_PWR1_33V);
+		pci_write_config_byte(host->pdev, SD_PCICFG_POWER2,
+					SD_PCICFG_PWR2_AUTO);
+		mdelay(20);
+		break;
+	}
+
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_1:
+		iowrite16(SD_CARDOPT_REQUIRED | SD_CARDOPT_DATA_RESP_TIMEOUT(14)
+				| SD_CARDOPT_C2_MODULE_ABSENT
+				| SD_CARDOPT_DATA_XFR_WIDTH_1,
+				host->ioaddr + SD_CARDOPTIONSETUP);
+		break;
+	case MMC_BUS_WIDTH_4:
+		iowrite16(SD_CARDOPT_REQUIRED | SD_CARDOPT_DATA_RESP_TIMEOUT(14)
+				| SD_CARDOPT_C2_MODULE_ABSENT
+				| SD_CARDOPT_DATA_XFR_WIDTH_4,
+				host->ioaddr + SD_CARDOPTIONSETUP);
+		break;
+	}
+}
+
+static void toshsd_set_led(struct toshsd_host *host, unsigned char state)
+{
+	iowrite16(state, host->ioaddr + SDIO_BASE + SDIO_LEDCTRL);
+}
+
+static void toshsd_finish_request(struct toshsd_host *host)
+{
+	struct mmc_request *mrq = host->mrq;
+
+	/* Write something to end the command */
+	host->mrq = NULL;
+	host->cmd = NULL;
+	host->data = NULL;
+
+	toshsd_set_led(host, 0);
+	mmc_request_done(host->mmc, mrq);
+}
+
+static irqreturn_t toshsd_thread_irq(int irq, void *dev_id)
+{
+	struct toshsd_host *host = dev_id;
+	struct mmc_data *data = host->data;
+	struct sg_mapping_iter *sg_miter = &host->sg_miter;
+	unsigned short *buf;
+	int count;
+	unsigned long flags;
+
+	if (!data) {
+		dev_warn(&host->pdev->dev, "Spurious Data IRQ\n");
+		if (host->cmd) {
+			host->cmd->error = -EIO;
+			toshsd_finish_request(host);
+		}
+		return IRQ_NONE;
+	}
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (!sg_miter_next(sg_miter))
+		goto done;
+
+	buf = sg_miter->addr;
+
+	/* Ensure we dont read more than one block. The chip will interrupt us
+	 * When the next block is available.
+	 */
+	count = sg_miter->length;
+	if (count > data->blksz)
+		count = data->blksz;
+
+	dev_dbg(&host->pdev->dev, "count: %08x, flags %08x\n", count,
+		data->flags);
+
+	/* Transfer the data */
+	if (data->flags & MMC_DATA_READ)
+		ioread32_rep(host->ioaddr + SD_DATAPORT, buf, count >> 2);
+	else
+		iowrite32_rep(host->ioaddr + SD_DATAPORT, buf, count >> 2);
+
+	sg_miter->consumed = count;
+	sg_miter_stop(sg_miter);
+
+done:
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	return IRQ_HANDLED;
+}
+
+static void toshsd_cmd_irq(struct toshsd_host *host)
+{
+	struct mmc_command *cmd = host->cmd;
+	u8 *buf;
+	u16 data;
+
+	if (!host->cmd) {
+		dev_warn(&host->pdev->dev, "Spurious CMD irq\n");
+		return;
+	}
+	buf = (u8 *)cmd->resp;
+	host->cmd = NULL;
+
+	if (cmd->flags & MMC_RSP_PRESENT && cmd->flags & MMC_RSP_136) {
+		/* R2 */
+		buf[12] = 0xff;
+		data = ioread16(host->ioaddr + SD_RESPONSE0);
+		buf[13] = data & 0xff;
+		buf[14] = data >> 8;
+		data = ioread16(host->ioaddr + SD_RESPONSE1);
+		buf[15] = data & 0xff;
+		buf[8] = data >> 8;
+		data = ioread16(host->ioaddr + SD_RESPONSE2);
+		buf[9] = data & 0xff;
+		buf[10] = data >> 8;
+		data = ioread16(host->ioaddr + SD_RESPONSE3);
+		buf[11] = data & 0xff;
+		buf[4] = data >> 8;
+		data = ioread16(host->ioaddr + SD_RESPONSE4);
+		buf[5] = data & 0xff;
+		buf[6] = data >> 8;
+		data = ioread16(host->ioaddr + SD_RESPONSE5);
+		buf[7] = data & 0xff;
+		buf[0] = data >> 8;
+		data = ioread16(host->ioaddr + SD_RESPONSE6);
+		buf[1] = data & 0xff;
+		buf[2] = data >> 8;
+		data = ioread16(host->ioaddr + SD_RESPONSE7);
+		buf[3] = data & 0xff;
+	} else if (cmd->flags & MMC_RSP_PRESENT) {
+		/* R1, R1B, R3, R6, R7 */
+		data = ioread16(host->ioaddr + SD_RESPONSE0);
+		buf[0] = data & 0xff;
+		buf[1] = data >> 8;
+		data = ioread16(host->ioaddr + SD_RESPONSE1);
+		buf[2] = data & 0xff;
+		buf[3] = data >> 8;
+	}
+
+	dev_dbg(&host->pdev->dev, "Command IRQ complete %d %d %x\n",
+		cmd->opcode, cmd->error, cmd->flags);
+
+	/* If there is data to handle we will
+	 * finish the request in the mmc_data_end_irq handler.*/
+	if (host->data)
+		return;
+
+	toshsd_finish_request(host);
+}
+
+static void toshsd_data_end_irq(struct toshsd_host *host)
+{
+	struct mmc_data *data = host->data;
+
+	host->data = NULL;
+
+	if (!data) {
+		dev_warn(&host->pdev->dev, "Spurious data end IRQ\n");
+		return;
+	}
+
+	if (data->error == 0)
+		data->bytes_xfered = data->blocks * data->blksz;
+	else
+		data->bytes_xfered = 0;
+
+	dev_dbg(&host->pdev->dev, "Completed data request xfr=%d\n",
+		data->bytes_xfered);
+
+	iowrite16(0, host->ioaddr + SD_STOPINTERNAL);
+
+	toshsd_finish_request(host);
+}
+
+static irqreturn_t toshsd_irq(int irq, void *dev_id)
+{
+	struct toshsd_host *host = dev_id;
+	u32 int_reg, int_mask, int_status, detail;
+	int error = 0, ret = IRQ_HANDLED;
+
+	spin_lock(&host->lock);
+	int_status = ioread32(host->ioaddr + SD_CARDSTATUS);
+	int_mask = ioread32(host->ioaddr + SD_INTMASKCARD);
+	int_reg = int_status & ~int_mask & ~IRQ_DONT_CARE_BITS;
+
+	dev_dbg(&host->pdev->dev, "IRQ status:%x mask:%x\n",
+		int_status, int_mask);
+
+	/* nothing to do: it's not our IRQ */
+	if (!int_reg) {
+		ret = IRQ_NONE;
+		goto irq_end;
+	}
+
+	if (int_reg & SD_BUF_CMD_TIMEOUT) {
+		error = -ETIMEDOUT;
+		dev_dbg(&host->pdev->dev, "Timeout\n");
+	} else if (int_reg & SD_BUF_CRC_ERR) {
+		error = -EILSEQ;
+		dev_err(&host->pdev->dev, "BadCRC\n");
+	} else if (int_reg & (SD_BUF_ILLEGAL_ACCESS
+				| SD_BUF_CMD_INDEX_ERR
+				| SD_BUF_STOP_BIT_END_ERR
+				| SD_BUF_OVERFLOW
+				| SD_BUF_UNDERFLOW
+				| SD_BUF_DATA_TIMEOUT)) {
+		dev_err(&host->pdev->dev, "Buffer status error: { %s%s%s%s%s%s}\n",
+			int_reg & SD_BUF_ILLEGAL_ACCESS ? "ILLEGAL_ACC " : "",
+			int_reg & SD_BUF_CMD_INDEX_ERR ? "CMD_INDEX " : "",
+			int_reg & SD_BUF_STOP_BIT_END_ERR ? "STOPBIT_END " : "",
+			int_reg & SD_BUF_OVERFLOW ? "OVERFLOW " : "",
+			int_reg & SD_BUF_UNDERFLOW ? "UNDERFLOW " : "",
+			int_reg & SD_BUF_DATA_TIMEOUT ? "DATA_TIMEOUT " : "");
+
+		detail = ioread32(host->ioaddr + SD_ERRORSTATUS0);
+		dev_err(&host->pdev->dev, "detail error status { %s%s%s%s%s%s%s%s%s%s%s%s%s}\n",
+			detail & SD_ERR0_RESP_CMD_ERR ? "RESP_CMD " : "",
+			detail & SD_ERR0_RESP_NON_CMD12_END_BIT_ERR ? "RESP_END_BIT " : "",
+			detail & SD_ERR0_RESP_CMD12_END_BIT_ERR ? "RESP_END_BIT " : "",
+			detail & SD_ERR0_READ_DATA_END_BIT_ERR ? "READ_DATA_END_BIT " : "",
+			detail & SD_ERR0_WRITE_CRC_STATUS_END_BIT_ERR ? "WRITE_CMD_END_BIT " : "",
+			detail & SD_ERR0_RESP_NON_CMD12_CRC_ERR ? "RESP_CRC " : "",
+			detail & SD_ERR0_RESP_CMD12_CRC_ERR ? "RESP_CRC " : "",
+			detail & SD_ERR0_READ_DATA_CRC_ERR ? "READ_DATA_CRC " : "",
+			detail & SD_ERR0_WRITE_CMD_CRC_ERR ? "WRITE_CMD_CRC " : "",
+			detail & SD_ERR1_NO_CMD_RESP ? "NO_CMD_RESP " : "",
+			detail & SD_ERR1_TIMEOUT_READ_DATA ? "READ_DATA_TIMEOUT " : "",
+			detail & SD_ERR1_TIMEOUT_CRS_STATUS ? "CRS_STATUS_TIMEOUT " : "",
+			detail & SD_ERR1_TIMEOUT_CRC_BUSY ? "CRC_BUSY_TIMEOUT " : "");
+		error = -EIO;
+	}
+
+	if (error) {
+		if (host->cmd)
+			host->cmd->error = error;
+
+		if (error == -ETIMEDOUT) {
+			iowrite32(int_status &
+				  ~(SD_BUF_CMD_TIMEOUT | SD_CARD_RESP_END),
+				  host->ioaddr + SD_CARDSTATUS);
+		} else {
+			toshsd_init(host);
+			__toshsd_set_ios(host->mmc, &host->mmc->ios);
+			goto irq_end;
+		}
+	}
+
+	/* Card insert/remove. The mmc controlling code is stateless. */
+	if (int_reg & (SD_CARD_CARD_INSERTED_0 | SD_CARD_CARD_REMOVED_0)) {
+		iowrite32(int_status &
+			  ~(SD_CARD_CARD_REMOVED_0 | SD_CARD_CARD_INSERTED_0),
+			  host->ioaddr + SD_CARDSTATUS);
+
+		if (int_reg & SD_CARD_CARD_INSERTED_0)
+			toshsd_init(host);
+
+		mmc_detect_change(host->mmc, 1);
+	}
+
+	/* Data transfer */
+	if (int_reg & (SD_BUF_READ_ENABLE | SD_BUF_WRITE_ENABLE)) {
+		iowrite32(int_status &
+			  ~(SD_BUF_WRITE_ENABLE | SD_BUF_READ_ENABLE),
+			  host->ioaddr + SD_CARDSTATUS);
+
+		ret = IRQ_WAKE_THREAD;
+		goto irq_end;
+	}
+
+	/* Command completion */
+	if (int_reg & SD_CARD_RESP_END) {
+		iowrite32(int_status & ~(SD_CARD_RESP_END),
+			  host->ioaddr + SD_CARDSTATUS);
+		toshsd_cmd_irq(host);
+	}
+
+	/* Data transfer completion */
+	if (int_reg & SD_CARD_RW_END) {
+		iowrite32(int_status & ~(SD_CARD_RW_END),
+			  host->ioaddr + SD_CARDSTATUS);
+		toshsd_data_end_irq(host);
+	}
+irq_end:
+	spin_unlock(&host->lock);
+	return ret;
+}
+
+static void toshsd_start_cmd(struct toshsd_host *host, struct mmc_command *cmd)
+{
+	struct mmc_data *data = host->data;
+	int c = cmd->opcode;
+
+	dev_dbg(&host->pdev->dev, "Command opcode: %d\n", cmd->opcode);
+
+	if (cmd->opcode == MMC_STOP_TRANSMISSION) {
+		iowrite16(SD_STOPINT_ISSUE_CMD12,
+			  host->ioaddr + SD_STOPINTERNAL);
+
+		cmd->resp[0] = cmd->opcode;
+		cmd->resp[1] = 0;
+		cmd->resp[2] = 0;
+		cmd->resp[3] = 0;
+
+		toshsd_finish_request(host);
+		return;
+	}
+
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE:
+		c |= SD_CMD_RESP_TYPE_NONE;
+		break;
+
+	case MMC_RSP_R1:
+		c |= SD_CMD_RESP_TYPE_EXT_R1;
+		break;
+	case MMC_RSP_R1B:
+		c |= SD_CMD_RESP_TYPE_EXT_R1B;
+		break;
+	case MMC_RSP_R2:
+		c |= SD_CMD_RESP_TYPE_EXT_R2;
+		break;
+	case MMC_RSP_R3:
+		c |= SD_CMD_RESP_TYPE_EXT_R3;
+		break;
+
+	default:
+		dev_err(&host->pdev->dev, "Unknown response type %d\n",
+			mmc_resp_type(cmd));
+		break;
+	}
+
+	host->cmd = cmd;
+
+	if (cmd->opcode == MMC_APP_CMD)
+		c |= SD_CMD_TYPE_ACMD;
+
+	if (cmd->opcode == MMC_GO_IDLE_STATE)
+		c |= (3 << 8);  /* removed from ipaq-asic3.h for some reason */
+
+	if (data) {
+		c |= SD_CMD_DATA_PRESENT;
+
+		if (data->blocks > 1) {
+			iowrite16(SD_STOPINT_AUTO_ISSUE_CMD12,
+				  host->ioaddr + SD_STOPINTERNAL);
+			c |= SD_CMD_MULTI_BLOCK;
+		}
+
+		if (data->flags & MMC_DATA_READ)
+			c |= SD_CMD_TRANSFER_READ;
+
+		/* MMC_DATA_WRITE does not require a bit to be set */
+	}
+
+	/* Send the command */
+	iowrite32(cmd->arg, host->ioaddr + SD_ARG0);
+	iowrite16(c, host->ioaddr + SD_CMD);
+}
+
+static void toshsd_start_data(struct toshsd_host *host, struct mmc_data *data)
+{
+	unsigned int flags = SG_MITER_ATOMIC;
+
+	dev_dbg(&host->pdev->dev, "setup data transfer: blocksize %08x  nr_blocks %d, offset: %08x\n",
+		data->blksz, data->blocks, data->sg->offset);
+
+	host->data = data;
+
+	if (data->flags & MMC_DATA_READ)
+		flags |= SG_MITER_TO_SG;
+	else
+		flags |= SG_MITER_FROM_SG;
+
+	sg_miter_start(&host->sg_miter, data->sg, data->sg_len, flags);
+
+	/* Set transfer length and blocksize */
+	iowrite16(data->blocks, host->ioaddr + SD_BLOCKCOUNT);
+	iowrite16(data->blksz, host->ioaddr + SD_CARDXFERDATALEN);
+}
+
+/* Process requests from the MMC layer */
+static void toshsd_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct toshsd_host *host = mmc_priv(mmc);
+	unsigned long flags;
+
+	/* abort if card not present */
+	if (!(ioread16(host->ioaddr + SD_CARDSTATUS) & SD_CARD_PRESENT_0)) {
+		mrq->cmd->error = -ENOMEDIUM;
+		mmc_request_done(mmc, mrq);
+		return;
+	}
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	WARN_ON(host->mrq != NULL);
+
+	host->mrq = mrq;
+
+	if (mrq->data)
+		toshsd_start_data(host, mrq->data);
+
+	toshsd_set_led(host, 1);
+
+	toshsd_start_cmd(host, mrq->cmd);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void toshsd_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct toshsd_host *host = mmc_priv(mmc);
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	__toshsd_set_ios(mmc, ios);
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static int toshsd_get_ro(struct mmc_host *mmc)
+{
+	struct toshsd_host *host = mmc_priv(mmc);
+
+	/* active low */
+	return !(ioread16(host->ioaddr + SD_CARDSTATUS) & SD_CARD_WRITE_PROTECT);
+}
+
+static int toshsd_get_cd(struct mmc_host *mmc)
+{
+	struct toshsd_host *host = mmc_priv(mmc);
+
+	return !!(ioread16(host->ioaddr + SD_CARDSTATUS) & SD_CARD_PRESENT_0);
+}
+
+static const struct mmc_host_ops toshsd_ops = {
+	.request = toshsd_request,
+	.set_ios = toshsd_set_ios,
+	.get_ro = toshsd_get_ro,
+	.get_cd = toshsd_get_cd,
+};
+
+
+static void toshsd_powerdown(struct toshsd_host *host)
+{
+	/* mask all interrupts */
+	iowrite32(0xffffffff, host->ioaddr + SD_INTMASKCARD);
+	/* disable card clock */
+	iowrite16(0x000, host->ioaddr + SDIO_BASE + SDIO_CLOCKNWAITCTRL);
+	iowrite16(0, host->ioaddr + SD_CARDCLOCKCTRL);
+	/* power down card */
+	pci_write_config_byte(host->pdev, SD_PCICFG_POWER1, SD_PCICFG_PWR1_OFF);
+	/* disable clock */
+	pci_write_config_byte(host->pdev, SD_PCICFG_CLKSTOP, 0);
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int toshsd_pm_suspend(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct toshsd_host *host = pci_get_drvdata(pdev);
+
+	toshsd_powerdown(host);
+
+	pci_save_state(pdev);
+	pci_enable_wake(pdev, PCI_D3hot, 0);
+	pci_disable_device(pdev);
+	pci_set_power_state(pdev, PCI_D3hot);
+
+	return 0;
+}
+
+static int toshsd_pm_resume(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct toshsd_host *host = pci_get_drvdata(pdev);
+	int ret;
+
+	pci_set_power_state(pdev, PCI_D0);
+	pci_restore_state(pdev);
+	ret = pci_enable_device(pdev);
+	if (ret)
+		return ret;
+
+	toshsd_init(host);
+
+	return 0;
+}
+#endif /* CONFIG_PM_SLEEP */
+
+static int toshsd_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
+{
+	int ret;
+	struct toshsd_host *host;
+	struct mmc_host *mmc;
+	resource_size_t base;
+
+	ret = pci_enable_device(pdev);
+	if (ret)
+		return ret;
+
+	mmc = mmc_alloc_host(sizeof(struct toshsd_host), &pdev->dev);
+	if (!mmc) {
+		ret = -ENOMEM;
+		goto err;
+	}
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+
+	host->pdev = pdev;
+	pci_set_drvdata(pdev, host);
+
+	ret = pci_request_regions(pdev, DRIVER_NAME);
+	if (ret)
+		goto free;
+
+	host->ioaddr = pci_iomap(pdev, 0, 0);
+	if (!host->ioaddr) {
+		ret = -ENOMEM;
+		goto release;
+	}
+
+	/* Set MMC host parameters */
+	mmc->ops = &toshsd_ops;
+	mmc->caps = MMC_CAP_4_BIT_DATA;
+	mmc->ocr_avail = MMC_VDD_32_33;
+
+	mmc->f_min = HCLK / 512;
+	mmc->f_max = HCLK;
+
+	spin_lock_init(&host->lock);
+
+	toshsd_init(host);
+
+	ret = request_threaded_irq(pdev->irq, toshsd_irq, toshsd_thread_irq,
+				   IRQF_SHARED, DRIVER_NAME, host);
+	if (ret)
+		goto unmap;
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto free_irq;
+
+	base = pci_resource_start(pdev, 0);
+	dev_dbg(&pdev->dev, "MMIO %pa, IRQ %d\n", &base, pdev->irq);
+
+	pm_suspend_ignore_children(&pdev->dev, 1);
+
+	return 0;
+
+free_irq:
+	free_irq(pdev->irq, host);
+unmap:
+	pci_iounmap(pdev, host->ioaddr);
+release:
+	pci_release_regions(pdev);
+free:
+	mmc_free_host(mmc);
+	pci_set_drvdata(pdev, NULL);
+err:
+	pci_disable_device(pdev);
+	return ret;
+}
+
+static void toshsd_remove(struct pci_dev *pdev)
+{
+	struct toshsd_host *host = pci_get_drvdata(pdev);
+
+	mmc_remove_host(host->mmc);
+	toshsd_powerdown(host);
+	free_irq(pdev->irq, host);
+	pci_iounmap(pdev, host->ioaddr);
+	pci_release_regions(pdev);
+	mmc_free_host(host->mmc);
+	pci_set_drvdata(pdev, NULL);
+	pci_disable_device(pdev);
+}
+
+static const struct dev_pm_ops toshsd_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(toshsd_pm_suspend, toshsd_pm_resume)
+};
+
+static struct pci_driver toshsd_driver = {
+	.name = DRIVER_NAME,
+	.id_table = pci_ids,
+	.probe = toshsd_probe,
+	.remove = toshsd_remove,
+	.driver.pm = &toshsd_pm_ops,
+};
+
+module_pci_driver(toshsd_driver);
+
+MODULE_AUTHOR("Ondrej Zary, Richard Betts");
+MODULE_DESCRIPTION("Toshiba PCI Secure Digital Host Controller Interface driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mmc/host/toshsd.h b/drivers/mmc/host/toshsd.h
new file mode 100644
index 000000000..3ba876eaa
--- /dev/null
+++ b/drivers/mmc/host/toshsd.h
@@ -0,0 +1,172 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ *  Toshiba PCI Secure Digital Host Controller Interface driver
+ *
+ *  Copyright (C) 2014 Ondrej Zary
+ *  Copyright (C) 2007 Richard Betts, All Rights Reserved.
+ *
+ *      Based on asic3_mmc.c Copyright (c) 2005 SDG Systems, LLC
+ */
+
+#define HCLK	33000000	/* 33 MHz (PCI clock) */
+
+#define SD_PCICFG_CLKSTOP	0x40	/* 0x1f = clock controller, 0 = stop */
+#define SD_PCICFG_GATEDCLK	0x41	/* Gated clock */
+#define SD_PCICFG_CLKMODE	0x42	/* Control clock of SD controller */
+#define SD_PCICFG_PINSTATUS	0x44	/* R/O: read status of SD pins */
+#define SD_PCICFG_POWER1	0x48
+#define SD_PCICFG_POWER2	0x49
+#define SD_PCICFG_POWER3	0x4a
+#define SD_PCICFG_CARDDETECT	0x4c
+#define SD_PCICFG_SLOTS		0x50	/* R/O: define support slot number */
+#define SD_PCICFG_EXTGATECLK1	0xf0	/* Could be used for gated clock */
+#define SD_PCICFG_EXTGATECLK2	0xf1	/* Could be used for gated clock */
+#define SD_PCICFG_EXTGATECLK3	0xf9	/* Bit 1: double buffer/single buffer */
+#define SD_PCICFG_SDLED_ENABLE1	0xfa
+#define SD_PCICFG_SDLED_ENABLE2	0xfe
+
+#define SD_PCICFG_CLKMODE_DIV_DISABLE	BIT(0)
+#define SD_PCICFG_CLKSTOP_ENABLE_ALL	0x1f
+#define SD_PCICFG_LED_ENABLE1_START	0x12
+#define SD_PCICFG_LED_ENABLE2_START	0x80
+
+#define SD_PCICFG_PWR1_33V	0x08	/* Set for 3.3 volts */
+#define SD_PCICFG_PWR1_OFF	0x00	/* Turn off power */
+#define SD_PCICFG_PWR2_AUTO	0x02
+
+#define SD_CMD			0x00	/* also for SDIO */
+#define SD_ARG0			0x04	/* also for SDIO */
+#define SD_ARG1			0x06	/* also for SDIO */
+#define SD_STOPINTERNAL		0x08
+#define SD_BLOCKCOUNT		0x0a	/* also for SDIO */
+#define SD_RESPONSE0		0x0c	/* also for SDIO */
+#define SD_RESPONSE1		0x0e	/* also for SDIO */
+#define SD_RESPONSE2		0x10	/* also for SDIO */
+#define SD_RESPONSE3		0x12	/* also for SDIO */
+#define SD_RESPONSE4		0x14	/* also for SDIO */
+#define SD_RESPONSE5		0x16	/* also for SDIO */
+#define SD_RESPONSE6		0x18	/* also for SDIO */
+#define SD_RESPONSE7		0x1a	/* also for SDIO */
+#define SD_CARDSTATUS		0x1c	/* also for SDIO */
+#define SD_BUFFERCTRL		0x1e	/* also for SDIO */
+#define SD_INTMASKCARD		0x20	/* also for SDIO */
+#define SD_INTMASKBUFFER	0x22	/* also for SDIO */
+#define SD_CARDCLOCKCTRL	0x24
+#define SD_CARDXFERDATALEN	0x26	/* also for SDIO */
+#define SD_CARDOPTIONSETUP	0x28	/* also for SDIO */
+#define SD_ERRORSTATUS0		0x2c	/* also for SDIO */
+#define SD_ERRORSTATUS1		0x2e	/* also for SDIO */
+#define SD_DATAPORT		0x30	/* also for SDIO */
+#define SD_TRANSACTIONCTRL	0x34	/* also for SDIO */
+#define SD_SOFTWARERESET	0xe0	/* also for SDIO */
+
+/* registers above marked "also for SDIO" and all SDIO registers below can be
+ * accessed at SDIO_BASE + reg address */
+#define SDIO_BASE	 0x100
+
+#define SDIO_CARDPORTSEL	0x02
+#define SDIO_CARDINTCTRL	0x36
+#define SDIO_CLOCKNWAITCTRL	0x38
+#define SDIO_HOSTINFORMATION	0x3a
+#define SDIO_ERRORCTRL		0x3c
+#define SDIO_LEDCTRL		0x3e
+
+#define SD_TRANSCTL_SET		BIT(8)
+
+#define SD_CARDCLK_DIV_DISABLE	BIT(15)
+#define SD_CARDCLK_ENABLE_CLOCK	BIT(8)
+#define SD_CARDCLK_CLK_DIV_512	BIT(7)
+#define SD_CARDCLK_CLK_DIV_256	BIT(6)
+#define SD_CARDCLK_CLK_DIV_128	BIT(5)
+#define SD_CARDCLK_CLK_DIV_64	BIT(4)
+#define SD_CARDCLK_CLK_DIV_32	BIT(3)
+#define SD_CARDCLK_CLK_DIV_16	BIT(2)
+#define SD_CARDCLK_CLK_DIV_8	BIT(1)
+#define SD_CARDCLK_CLK_DIV_4	BIT(0)
+#define SD_CARDCLK_CLK_DIV_2	0
+
+#define SD_CARDOPT_REQUIRED		0x000e
+#define SD_CARDOPT_DATA_RESP_TIMEOUT(x)	(((x) & 0x0f) << 4) /* 4 bits */
+#define SD_CARDOPT_C2_MODULE_ABSENT	BIT(14)
+#define SD_CARDOPT_DATA_XFR_WIDTH_1	(1 << 15)
+#define SD_CARDOPT_DATA_XFR_WIDTH_4	(0 << 15)
+
+#define SD_CMD_TYPE_CMD			(0 << 6)
+#define SD_CMD_TYPE_ACMD		(1 << 6)
+#define SD_CMD_TYPE_AUTHEN		(2 << 6)
+#define SD_CMD_RESP_TYPE_NONE		(3 << 8)
+#define SD_CMD_RESP_TYPE_EXT_R1		(4 << 8)
+#define SD_CMD_RESP_TYPE_EXT_R1B	(5 << 8)
+#define SD_CMD_RESP_TYPE_EXT_R2		(6 << 8)
+#define SD_CMD_RESP_TYPE_EXT_R3		(7 << 8)
+#define SD_CMD_RESP_TYPE_EXT_R6		(4 << 8)
+#define SD_CMD_RESP_TYPE_EXT_R7		(4 << 8)
+#define SD_CMD_DATA_PRESENT		BIT(11)
+#define SD_CMD_TRANSFER_READ		BIT(12)
+#define SD_CMD_MULTI_BLOCK		BIT(13)
+#define SD_CMD_SECURITY_CMD		BIT(14)
+
+#define SD_STOPINT_ISSUE_CMD12		BIT(0)
+#define SD_STOPINT_AUTO_ISSUE_CMD12	BIT(8)
+
+#define SD_CARD_RESP_END	BIT(0)
+#define SD_CARD_RW_END		BIT(2)
+#define SD_CARD_CARD_REMOVED_0	BIT(3)
+#define SD_CARD_CARD_INSERTED_0	BIT(4)
+#define SD_CARD_PRESENT_0	BIT(5)
+#define SD_CARD_UNK6		BIT(6)
+#define SD_CARD_WRITE_PROTECT	BIT(7)
+#define SD_CARD_CARD_REMOVED_3	BIT(8)
+#define SD_CARD_CARD_INSERTED_3	BIT(9)
+#define SD_CARD_PRESENT_3	BIT(10)
+
+#define SD_BUF_CMD_INDEX_ERR	BIT(16)
+#define SD_BUF_CRC_ERR		BIT(17)
+#define SD_BUF_STOP_BIT_END_ERR	BIT(18)
+#define SD_BUF_DATA_TIMEOUT	BIT(19)
+#define SD_BUF_OVERFLOW		BIT(20)
+#define SD_BUF_UNDERFLOW	BIT(21)
+#define SD_BUF_CMD_TIMEOUT	BIT(22)
+#define SD_BUF_UNK7		BIT(23)
+#define SD_BUF_READ_ENABLE	BIT(24)
+#define SD_BUF_WRITE_ENABLE	BIT(25)
+#define SD_BUF_ILLEGAL_FUNCTION	BIT(29)
+#define SD_BUF_CMD_BUSY		BIT(30)
+#define SD_BUF_ILLEGAL_ACCESS	BIT(31)
+
+#define SD_ERR0_RESP_CMD_ERR			BIT(0)
+#define SD_ERR0_RESP_NON_CMD12_END_BIT_ERR	BIT(2)
+#define SD_ERR0_RESP_CMD12_END_BIT_ERR		BIT(3)
+#define SD_ERR0_READ_DATA_END_BIT_ERR		BIT(4)
+#define SD_ERR0_WRITE_CRC_STATUS_END_BIT_ERR	BIT(5)
+#define SD_ERR0_RESP_NON_CMD12_CRC_ERR		BIT(8)
+#define SD_ERR0_RESP_CMD12_CRC_ERR		BIT(9)
+#define SD_ERR0_READ_DATA_CRC_ERR		BIT(10)
+#define SD_ERR0_WRITE_CMD_CRC_ERR		BIT(11)
+
+#define SD_ERR1_NO_CMD_RESP		BIT(16)
+#define SD_ERR1_TIMEOUT_READ_DATA	BIT(20)
+#define SD_ERR1_TIMEOUT_CRS_STATUS	BIT(21)
+#define SD_ERR1_TIMEOUT_CRC_BUSY	BIT(22)
+
+#define IRQ_DONT_CARE_BITS (SD_CARD_PRESENT_3 \
+	| SD_CARD_WRITE_PROTECT \
+	| SD_CARD_UNK6 \
+	| SD_CARD_PRESENT_0 \
+	| SD_BUF_UNK7 \
+	| SD_BUF_CMD_BUSY)
+
+struct toshsd_host {
+	struct pci_dev *pdev;
+	struct mmc_host *mmc;
+
+	spinlock_t lock;
+
+	struct mmc_request *mrq;/* Current request */
+	struct mmc_command *cmd;/* Current command */
+	struct mmc_data *data;	/* Current data request */
+
+	struct sg_mapping_iter sg_miter; /* for PIO */
+
+	void __iomem *ioaddr; /* mapped address */
+};
diff --git a/drivers/mmc/host/uniphier-sd.c b/drivers/mmc/host/uniphier-sd.c
new file mode 100644
index 000000000..196e94bf3
--- /dev/null
+++ b/drivers/mmc/host/uniphier-sd.c
@@ -0,0 +1,699 @@
+// SPDX-License-Identifier: GPL-2.0
+//
+// Copyright (C) 2017-2018 Socionext Inc.
+//   Author: Masahiro Yamada <yamada.masahiro@socionext.com>
+
+#include <linux/bitfield.h>
+#include <linux/bitops.h>
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/mfd/tmio.h>
+#include <linux/mmc/host.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/platform_device.h>
+#include <linux/reset.h>
+
+#include "tmio_mmc.h"
+
+#define   UNIPHIER_SD_CLK_CTL_DIV1024		BIT(16)
+#define   UNIPHIER_SD_CLK_CTL_DIV1		BIT(10)
+#define   UNIPHIER_SD_CLKCTL_OFFEN		BIT(9)  // auto SDCLK stop
+#define UNIPHIER_SD_CC_EXT_MODE		0x1b0
+#define   UNIPHIER_SD_CC_EXT_MODE_DMA		BIT(1)
+#define UNIPHIER_SD_HOST_MODE		0x1c8
+#define UNIPHIER_SD_VOLT		0x1e4
+#define   UNIPHIER_SD_VOLT_MASK			GENMASK(1, 0)
+#define   UNIPHIER_SD_VOLT_OFF			0
+#define   UNIPHIER_SD_VOLT_330			1	// 3.3V signal
+#define   UNIPHIER_SD_VOLT_180			2	// 1.8V signal
+#define UNIPHIER_SD_DMA_MODE		0x410
+#define   UNIPHIER_SD_DMA_MODE_DIR_MASK		GENMASK(17, 16)
+#define   UNIPHIER_SD_DMA_MODE_DIR_TO_DEV	0
+#define   UNIPHIER_SD_DMA_MODE_DIR_FROM_DEV	1
+#define   UNIPHIER_SD_DMA_MODE_WIDTH_MASK	GENMASK(5, 4)
+#define   UNIPHIER_SD_DMA_MODE_WIDTH_8		0
+#define   UNIPHIER_SD_DMA_MODE_WIDTH_16		1
+#define   UNIPHIER_SD_DMA_MODE_WIDTH_32		2
+#define   UNIPHIER_SD_DMA_MODE_WIDTH_64		3
+#define   UNIPHIER_SD_DMA_MODE_ADDR_INC		BIT(0)	// 1: inc, 0: fixed
+#define UNIPHIER_SD_DMA_CTL		0x414
+#define   UNIPHIER_SD_DMA_CTL_START	BIT(0)	// start DMA (auto cleared)
+#define UNIPHIER_SD_DMA_RST		0x418
+#define   UNIPHIER_SD_DMA_RST_CH1	BIT(9)
+#define   UNIPHIER_SD_DMA_RST_CH0	BIT(8)
+#define UNIPHIER_SD_DMA_ADDR_L		0x440
+#define UNIPHIER_SD_DMA_ADDR_H		0x444
+
+/*
+ * IP is extended to support various features: built-in DMA engine,
+ * 1/1024 divisor, etc.
+ */
+#define UNIPHIER_SD_CAP_EXTENDED_IP		BIT(0)
+/* RX channel of the built-in DMA controller is broken (Pro5) */
+#define UNIPHIER_SD_CAP_BROKEN_DMA_RX		BIT(1)
+
+struct uniphier_sd_priv {
+	struct tmio_mmc_data tmio_data;
+	struct pinctrl *pinctrl;
+	struct pinctrl_state *pinstate_uhs;
+	struct clk *clk;
+	struct reset_control *rst;
+	struct reset_control *rst_br;
+	struct reset_control *rst_hw;
+	struct dma_chan *chan;
+	enum dma_data_direction dma_dir;
+	unsigned long clk_rate;
+	unsigned long caps;
+};
+
+static void *uniphier_sd_priv(struct tmio_mmc_host *host)
+{
+	return container_of(host->pdata, struct uniphier_sd_priv, tmio_data);
+}
+
+static void uniphier_sd_dma_endisable(struct tmio_mmc_host *host, int enable)
+{
+	sd_ctrl_write16(host, CTL_DMA_ENABLE, enable ? DMA_ENABLE_DMASDRW : 0);
+}
+
+/* external DMA engine */
+static void uniphier_sd_external_dma_issue(unsigned long arg)
+{
+	struct tmio_mmc_host *host = (void *)arg;
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+
+	uniphier_sd_dma_endisable(host, 1);
+	dma_async_issue_pending(priv->chan);
+}
+
+static void uniphier_sd_external_dma_callback(void *param,
+					const struct dmaengine_result *result)
+{
+	struct tmio_mmc_host *host = param;
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+	unsigned long flags;
+
+	dma_unmap_sg(mmc_dev(host->mmc), host->sg_ptr, host->sg_len,
+		     priv->dma_dir);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (result->result == DMA_TRANS_NOERROR) {
+		/*
+		 * When the external DMA engine is enabled, strangely enough,
+		 * the DATAEND flag can be asserted even if the DMA engine has
+		 * not been kicked yet.  Enable the TMIO_STAT_DATAEND irq only
+		 * after we make sure the DMA engine finishes the transfer,
+		 * hence, in this callback.
+		 */
+		tmio_mmc_enable_mmc_irqs(host, TMIO_STAT_DATAEND);
+	} else {
+		host->data->error = -ETIMEDOUT;
+		tmio_mmc_do_data_irq(host);
+	}
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void uniphier_sd_external_dma_start(struct tmio_mmc_host *host,
+					   struct mmc_data *data)
+{
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+	enum dma_transfer_direction dma_tx_dir;
+	struct dma_async_tx_descriptor *desc;
+	dma_cookie_t cookie;
+	int sg_len;
+
+	if (!priv->chan)
+		goto force_pio;
+
+	if (data->flags & MMC_DATA_READ) {
+		priv->dma_dir = DMA_FROM_DEVICE;
+		dma_tx_dir = DMA_DEV_TO_MEM;
+	} else {
+		priv->dma_dir = DMA_TO_DEVICE;
+		dma_tx_dir = DMA_MEM_TO_DEV;
+	}
+
+	sg_len = dma_map_sg(mmc_dev(host->mmc), host->sg_ptr, host->sg_len,
+			    priv->dma_dir);
+	if (sg_len == 0)
+		goto force_pio;
+
+	desc = dmaengine_prep_slave_sg(priv->chan, host->sg_ptr, sg_len,
+				       dma_tx_dir, DMA_CTRL_ACK);
+	if (!desc)
+		goto unmap_sg;
+
+	desc->callback_result = uniphier_sd_external_dma_callback;
+	desc->callback_param = host;
+
+	cookie = dmaengine_submit(desc);
+	if (cookie < 0)
+		goto unmap_sg;
+
+	host->dma_on = true;
+
+	return;
+
+unmap_sg:
+	dma_unmap_sg(mmc_dev(host->mmc), host->sg_ptr, host->sg_len,
+		     priv->dma_dir);
+force_pio:
+	uniphier_sd_dma_endisable(host, 0);
+}
+
+static void uniphier_sd_external_dma_enable(struct tmio_mmc_host *host,
+					    bool enable)
+{
+}
+
+static void uniphier_sd_external_dma_request(struct tmio_mmc_host *host,
+					     struct tmio_mmc_data *pdata)
+{
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+	struct dma_chan *chan;
+
+	chan = dma_request_chan(mmc_dev(host->mmc), "rx-tx");
+	if (IS_ERR(chan)) {
+		dev_warn(mmc_dev(host->mmc),
+			 "failed to request DMA channel. falling back to PIO\n");
+		return;	/* just use PIO even for -EPROBE_DEFER */
+	}
+
+	/* this driver uses a single channel for both RX an TX */
+	priv->chan = chan;
+	host->chan_rx = chan;
+	host->chan_tx = chan;
+
+	tasklet_init(&host->dma_issue, uniphier_sd_external_dma_issue,
+		     (unsigned long)host);
+}
+
+static void uniphier_sd_external_dma_release(struct tmio_mmc_host *host)
+{
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+
+	if (priv->chan)
+		dma_release_channel(priv->chan);
+}
+
+static void uniphier_sd_external_dma_abort(struct tmio_mmc_host *host)
+{
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+
+	uniphier_sd_dma_endisable(host, 0);
+
+	if (priv->chan)
+		dmaengine_terminate_sync(priv->chan);
+}
+
+static void uniphier_sd_external_dma_dataend(struct tmio_mmc_host *host)
+{
+	uniphier_sd_dma_endisable(host, 0);
+
+	tmio_mmc_do_data_irq(host);
+}
+
+static const struct tmio_mmc_dma_ops uniphier_sd_external_dma_ops = {
+	.start = uniphier_sd_external_dma_start,
+	.enable = uniphier_sd_external_dma_enable,
+	.request = uniphier_sd_external_dma_request,
+	.release = uniphier_sd_external_dma_release,
+	.abort = uniphier_sd_external_dma_abort,
+	.dataend = uniphier_sd_external_dma_dataend,
+};
+
+static void uniphier_sd_internal_dma_issue(unsigned long arg)
+{
+	struct tmio_mmc_host *host = (void *)arg;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	tmio_mmc_enable_mmc_irqs(host, TMIO_STAT_DATAEND);
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	uniphier_sd_dma_endisable(host, 1);
+	writel(UNIPHIER_SD_DMA_CTL_START, host->ctl + UNIPHIER_SD_DMA_CTL);
+}
+
+static void uniphier_sd_internal_dma_start(struct tmio_mmc_host *host,
+					   struct mmc_data *data)
+{
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+	struct scatterlist *sg = host->sg_ptr;
+	dma_addr_t dma_addr;
+	unsigned int dma_mode_dir;
+	u32 dma_mode;
+	int sg_len;
+
+	if ((data->flags & MMC_DATA_READ) && !host->chan_rx)
+		goto force_pio;
+
+	if (WARN_ON(host->sg_len != 1))
+		goto force_pio;
+
+	if (!IS_ALIGNED(sg->offset, 8))
+		goto force_pio;
+
+	if (data->flags & MMC_DATA_READ) {
+		priv->dma_dir = DMA_FROM_DEVICE;
+		dma_mode_dir = UNIPHIER_SD_DMA_MODE_DIR_FROM_DEV;
+	} else {
+		priv->dma_dir = DMA_TO_DEVICE;
+		dma_mode_dir = UNIPHIER_SD_DMA_MODE_DIR_TO_DEV;
+	}
+
+	sg_len = dma_map_sg(mmc_dev(host->mmc), sg, 1, priv->dma_dir);
+	if (sg_len == 0)
+		goto force_pio;
+
+	dma_mode = FIELD_PREP(UNIPHIER_SD_DMA_MODE_DIR_MASK, dma_mode_dir);
+	dma_mode |= FIELD_PREP(UNIPHIER_SD_DMA_MODE_WIDTH_MASK,
+			       UNIPHIER_SD_DMA_MODE_WIDTH_64);
+	dma_mode |= UNIPHIER_SD_DMA_MODE_ADDR_INC;
+
+	writel(dma_mode, host->ctl + UNIPHIER_SD_DMA_MODE);
+
+	dma_addr = sg_dma_address(data->sg);
+	writel(lower_32_bits(dma_addr), host->ctl + UNIPHIER_SD_DMA_ADDR_L);
+	writel(upper_32_bits(dma_addr), host->ctl + UNIPHIER_SD_DMA_ADDR_H);
+
+	host->dma_on = true;
+
+	return;
+force_pio:
+	uniphier_sd_dma_endisable(host, 0);
+}
+
+static void uniphier_sd_internal_dma_enable(struct tmio_mmc_host *host,
+					    bool enable)
+{
+}
+
+static void uniphier_sd_internal_dma_request(struct tmio_mmc_host *host,
+					     struct tmio_mmc_data *pdata)
+{
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+
+	/*
+	 * Due to a hardware bug, Pro5 cannot use DMA for RX.
+	 * We can still use DMA for TX, but PIO for RX.
+	 */
+	if (!(priv->caps & UNIPHIER_SD_CAP_BROKEN_DMA_RX))
+		host->chan_rx = (void *)0xdeadbeaf;
+
+	host->chan_tx = (void *)0xdeadbeaf;
+
+	tasklet_init(&host->dma_issue, uniphier_sd_internal_dma_issue,
+		     (unsigned long)host);
+}
+
+static void uniphier_sd_internal_dma_release(struct tmio_mmc_host *host)
+{
+	/* Each value is set to zero to assume "disabling" each DMA */
+	host->chan_rx = NULL;
+	host->chan_tx = NULL;
+}
+
+static void uniphier_sd_internal_dma_abort(struct tmio_mmc_host *host)
+{
+	u32 tmp;
+
+	uniphier_sd_dma_endisable(host, 0);
+
+	tmp = readl(host->ctl + UNIPHIER_SD_DMA_RST);
+	tmp &= ~(UNIPHIER_SD_DMA_RST_CH1 | UNIPHIER_SD_DMA_RST_CH0);
+	writel(tmp, host->ctl + UNIPHIER_SD_DMA_RST);
+
+	tmp |= UNIPHIER_SD_DMA_RST_CH1 | UNIPHIER_SD_DMA_RST_CH0;
+	writel(tmp, host->ctl + UNIPHIER_SD_DMA_RST);
+}
+
+static void uniphier_sd_internal_dma_dataend(struct tmio_mmc_host *host)
+{
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+
+	uniphier_sd_dma_endisable(host, 0);
+	dma_unmap_sg(mmc_dev(host->mmc), host->sg_ptr, 1, priv->dma_dir);
+
+	tmio_mmc_do_data_irq(host);
+}
+
+static const struct tmio_mmc_dma_ops uniphier_sd_internal_dma_ops = {
+	.start = uniphier_sd_internal_dma_start,
+	.enable = uniphier_sd_internal_dma_enable,
+	.request = uniphier_sd_internal_dma_request,
+	.release = uniphier_sd_internal_dma_release,
+	.abort = uniphier_sd_internal_dma_abort,
+	.dataend = uniphier_sd_internal_dma_dataend,
+};
+
+static int uniphier_sd_clk_enable(struct tmio_mmc_host *host)
+{
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+	struct mmc_host *mmc = host->mmc;
+	int ret;
+
+	ret = clk_prepare_enable(priv->clk);
+	if (ret)
+		return ret;
+
+	ret = clk_set_rate(priv->clk, ULONG_MAX);
+	if (ret)
+		goto disable_clk;
+
+	priv->clk_rate = clk_get_rate(priv->clk);
+
+	/* If max-frequency property is set, use it. */
+	if (!mmc->f_max)
+		mmc->f_max = priv->clk_rate;
+
+	/*
+	 * 1/512 is the finest divisor in the original IP.  Newer versions
+	 * also supports 1/1024 divisor. (UniPhier-specific extension)
+	 */
+	if (priv->caps & UNIPHIER_SD_CAP_EXTENDED_IP)
+		mmc->f_min = priv->clk_rate / 1024;
+	else
+		mmc->f_min = priv->clk_rate / 512;
+
+	ret = reset_control_deassert(priv->rst);
+	if (ret)
+		goto disable_clk;
+
+	ret = reset_control_deassert(priv->rst_br);
+	if (ret)
+		goto assert_rst;
+
+	return 0;
+
+assert_rst:
+	reset_control_assert(priv->rst);
+disable_clk:
+	clk_disable_unprepare(priv->clk);
+
+	return ret;
+}
+
+static void uniphier_sd_clk_disable(struct tmio_mmc_host *host)
+{
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+
+	reset_control_assert(priv->rst_br);
+	reset_control_assert(priv->rst);
+	clk_disable_unprepare(priv->clk);
+}
+
+static void uniphier_sd_hw_reset(struct mmc_host *mmc)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+
+	reset_control_assert(priv->rst_hw);
+	/* For eMMC, minimum is 1us but give it 9us for good measure */
+	udelay(9);
+	reset_control_deassert(priv->rst_hw);
+	/* For eMMC, minimum is 200us but give it 300us for good measure */
+	usleep_range(300, 1000);
+}
+
+static void uniphier_sd_set_clock(struct tmio_mmc_host *host,
+				  unsigned int clock)
+{
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+	unsigned long divisor;
+	u32 tmp;
+
+	tmp = readl(host->ctl + (CTL_SD_CARD_CLK_CTL << 1));
+
+	/* stop the clock before changing its rate to avoid a glitch signal */
+	tmp &= ~CLK_CTL_SCLKEN;
+	writel(tmp, host->ctl + (CTL_SD_CARD_CLK_CTL << 1));
+
+	if (clock == 0)
+		return;
+
+	tmp &= ~UNIPHIER_SD_CLK_CTL_DIV1024;
+	tmp &= ~UNIPHIER_SD_CLK_CTL_DIV1;
+	tmp &= ~CLK_CTL_DIV_MASK;
+
+	divisor = priv->clk_rate / clock;
+
+	/*
+	 * In the original IP, bit[7:0] represents the divisor.
+	 * bit7 set: 1/512, ... bit0 set:1/4, all bits clear: 1/2
+	 *
+	 * The IP does not define a way to achieve 1/1.  For UniPhier variants,
+	 * bit10 is used for 1/1.  Newer versions of UniPhier variants use
+	 * bit16 for 1/1024.
+	 */
+	if (divisor <= 1)
+		tmp |= UNIPHIER_SD_CLK_CTL_DIV1;
+	else if (priv->caps & UNIPHIER_SD_CAP_EXTENDED_IP && divisor > 512)
+		tmp |= UNIPHIER_SD_CLK_CTL_DIV1024;
+	else
+		tmp |= roundup_pow_of_two(divisor) >> 2;
+
+	writel(tmp, host->ctl + (CTL_SD_CARD_CLK_CTL << 1));
+
+	tmp |= CLK_CTL_SCLKEN;
+	writel(tmp, host->ctl + (CTL_SD_CARD_CLK_CTL << 1));
+}
+
+static void uniphier_sd_host_init(struct tmio_mmc_host *host)
+{
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+	u32 val;
+
+	/*
+	 * Connected to 32bit AXI.
+	 * This register holds settings for SoC-specific internal bus
+	 * connection.  What is worse, the register spec was changed,
+	 * breaking the backward compatibility.  Write an appropriate
+	 * value depending on a flag associated with a compatible string.
+	 */
+	if (priv->caps & UNIPHIER_SD_CAP_EXTENDED_IP)
+		val = 0x00000101;
+	else
+		val = 0x00000000;
+
+	writel(val, host->ctl + UNIPHIER_SD_HOST_MODE);
+
+	val = 0;
+	/*
+	 * If supported, the controller can automatically
+	 * enable/disable the clock line to the card.
+	 */
+	if (priv->caps & UNIPHIER_SD_CAP_EXTENDED_IP)
+		val |= UNIPHIER_SD_CLKCTL_OFFEN;
+
+	writel(val, host->ctl + (CTL_SD_CARD_CLK_CTL << 1));
+}
+
+static int uniphier_sd_start_signal_voltage_switch(struct mmc_host *mmc,
+						   struct mmc_ios *ios)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+	struct uniphier_sd_priv *priv = uniphier_sd_priv(host);
+	struct pinctrl_state *pinstate = NULL;
+	u32 val, tmp;
+
+	switch (ios->signal_voltage) {
+	case MMC_SIGNAL_VOLTAGE_330:
+		val = UNIPHIER_SD_VOLT_330;
+		break;
+	case MMC_SIGNAL_VOLTAGE_180:
+		val = UNIPHIER_SD_VOLT_180;
+		pinstate = priv->pinstate_uhs;
+		break;
+	default:
+		return -ENOTSUPP;
+	}
+
+	tmp = readl(host->ctl + UNIPHIER_SD_VOLT);
+	tmp &= ~UNIPHIER_SD_VOLT_MASK;
+	tmp |= FIELD_PREP(UNIPHIER_SD_VOLT_MASK, val);
+	writel(tmp, host->ctl + UNIPHIER_SD_VOLT);
+
+	if (pinstate)
+		pinctrl_select_state(priv->pinctrl, pinstate);
+	else
+		pinctrl_select_default_state(mmc_dev(mmc));
+
+	return 0;
+}
+
+static int uniphier_sd_uhs_init(struct tmio_mmc_host *host,
+				struct uniphier_sd_priv *priv)
+{
+	priv->pinctrl = devm_pinctrl_get(mmc_dev(host->mmc));
+	if (IS_ERR(priv->pinctrl))
+		return PTR_ERR(priv->pinctrl);
+
+	priv->pinstate_uhs = pinctrl_lookup_state(priv->pinctrl, "uhs");
+	if (IS_ERR(priv->pinstate_uhs))
+		return PTR_ERR(priv->pinstate_uhs);
+
+	host->ops.start_signal_voltage_switch =
+					uniphier_sd_start_signal_voltage_switch;
+
+	return 0;
+}
+
+static int uniphier_sd_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct uniphier_sd_priv *priv;
+	struct tmio_mmc_data *tmio_data;
+	struct tmio_mmc_host *host;
+	int irq, ret;
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0)
+		return irq;
+
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->caps = (unsigned long)of_device_get_match_data(dev);
+
+	priv->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(priv->clk)) {
+		dev_err(dev, "failed to get clock\n");
+		return PTR_ERR(priv->clk);
+	}
+
+	priv->rst = devm_reset_control_get_shared(dev, "host");
+	if (IS_ERR(priv->rst)) {
+		dev_err(dev, "failed to get host reset\n");
+		return PTR_ERR(priv->rst);
+	}
+
+	/* old version has one more reset */
+	if (!(priv->caps & UNIPHIER_SD_CAP_EXTENDED_IP)) {
+		priv->rst_br = devm_reset_control_get_shared(dev, "bridge");
+		if (IS_ERR(priv->rst_br)) {
+			dev_err(dev, "failed to get bridge reset\n");
+			return PTR_ERR(priv->rst_br);
+		}
+	}
+
+	tmio_data = &priv->tmio_data;
+	tmio_data->flags |= TMIO_MMC_32BIT_DATA_PORT;
+
+	host = tmio_mmc_host_alloc(pdev, tmio_data);
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	if (host->mmc->caps & MMC_CAP_HW_RESET) {
+		priv->rst_hw = devm_reset_control_get_exclusive(dev, "hw");
+		if (IS_ERR(priv->rst_hw)) {
+			dev_err(dev, "failed to get hw reset\n");
+			ret = PTR_ERR(priv->rst_hw);
+			goto free_host;
+		}
+		host->ops.hw_reset = uniphier_sd_hw_reset;
+	}
+
+	if (host->mmc->caps & MMC_CAP_UHS) {
+		ret = uniphier_sd_uhs_init(host, priv);
+		if (ret) {
+			dev_warn(dev,
+				 "failed to setup UHS (error %d).  Disabling UHS.",
+				 ret);
+			host->mmc->caps &= ~MMC_CAP_UHS;
+		}
+	}
+
+	if (priv->caps & UNIPHIER_SD_CAP_EXTENDED_IP)
+		host->dma_ops = &uniphier_sd_internal_dma_ops;
+	else
+		host->dma_ops = &uniphier_sd_external_dma_ops;
+
+	host->bus_shift = 1;
+	host->clk_enable = uniphier_sd_clk_enable;
+	host->clk_disable = uniphier_sd_clk_disable;
+	host->set_clock = uniphier_sd_set_clock;
+
+	ret = uniphier_sd_clk_enable(host);
+	if (ret)
+		goto free_host;
+
+	uniphier_sd_host_init(host);
+
+	tmio_data->ocr_mask = MMC_VDD_32_33 | MMC_VDD_33_34;
+	if (host->mmc->caps & MMC_CAP_UHS)
+		tmio_data->ocr_mask |= MMC_VDD_165_195;
+
+	tmio_data->max_segs = 1;
+	tmio_data->max_blk_count = U16_MAX;
+
+	ret = tmio_mmc_host_probe(host);
+	if (ret)
+		goto disable_clk;
+
+	ret = devm_request_irq(dev, irq, tmio_mmc_irq, IRQF_SHARED,
+			       dev_name(dev), host);
+	if (ret)
+		goto remove_host;
+
+	return 0;
+
+remove_host:
+	tmio_mmc_host_remove(host);
+disable_clk:
+	uniphier_sd_clk_disable(host);
+free_host:
+	tmio_mmc_host_free(host);
+
+	return ret;
+}
+
+static int uniphier_sd_remove(struct platform_device *pdev)
+{
+	struct tmio_mmc_host *host = platform_get_drvdata(pdev);
+
+	tmio_mmc_host_remove(host);
+	uniphier_sd_clk_disable(host);
+	tmio_mmc_host_free(host);
+
+	return 0;
+}
+
+static const struct of_device_id uniphier_sd_match[] = {
+	{
+		.compatible = "socionext,uniphier-sd-v2.91",
+	},
+	{
+		.compatible = "socionext,uniphier-sd-v3.1",
+		.data = (void *)(UNIPHIER_SD_CAP_EXTENDED_IP |
+				 UNIPHIER_SD_CAP_BROKEN_DMA_RX),
+	},
+	{
+		.compatible = "socionext,uniphier-sd-v3.1.1",
+		.data = (void *)UNIPHIER_SD_CAP_EXTENDED_IP,
+	},
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, uniphier_sd_match);
+
+static struct platform_driver uniphier_sd_driver = {
+	.probe = uniphier_sd_probe,
+	.remove = uniphier_sd_remove,
+	.driver = {
+		.name = "uniphier-sd",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = uniphier_sd_match,
+	},
+};
+module_platform_driver(uniphier_sd_driver);
+
+MODULE_AUTHOR("Masahiro Yamada <yamada.masahiro@socionext.com>");
+MODULE_DESCRIPTION("UniPhier SD/eMMC host controller driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/usdhi6rol0.c b/drivers/mmc/host/usdhi6rol0.c
new file mode 100644
index 000000000..4f22ecef9
--- /dev/null
+++ b/drivers/mmc/host/usdhi6rol0.c
@@ -0,0 +1,1908 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2013-2014 Renesas Electronics Europe Ltd.
+ * Author: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/dmaengine.h>
+#include <linux/highmem.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/log2.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sd.h>
+#include <linux/mmc/sdio.h>
+#include <linux/module.h>
+#include <linux/pagemap.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/platform_device.h>
+#include <linux/scatterlist.h>
+#include <linux/string.h>
+#include <linux/time.h>
+#include <linux/virtio.h>
+#include <linux/workqueue.h>
+
+#define USDHI6_SD_CMD		0x0000
+#define USDHI6_SD_PORT_SEL	0x0004
+#define USDHI6_SD_ARG		0x0008
+#define USDHI6_SD_STOP		0x0010
+#define USDHI6_SD_SECCNT	0x0014
+#define USDHI6_SD_RSP10		0x0018
+#define USDHI6_SD_RSP32		0x0020
+#define USDHI6_SD_RSP54		0x0028
+#define USDHI6_SD_RSP76		0x0030
+#define USDHI6_SD_INFO1		0x0038
+#define USDHI6_SD_INFO2		0x003c
+#define USDHI6_SD_INFO1_MASK	0x0040
+#define USDHI6_SD_INFO2_MASK	0x0044
+#define USDHI6_SD_CLK_CTRL	0x0048
+#define USDHI6_SD_SIZE		0x004c
+#define USDHI6_SD_OPTION	0x0050
+#define USDHI6_SD_ERR_STS1	0x0058
+#define USDHI6_SD_ERR_STS2	0x005c
+#define USDHI6_SD_BUF0		0x0060
+#define USDHI6_SDIO_MODE	0x0068
+#define USDHI6_SDIO_INFO1	0x006c
+#define USDHI6_SDIO_INFO1_MASK	0x0070
+#define USDHI6_CC_EXT_MODE	0x01b0
+#define USDHI6_SOFT_RST		0x01c0
+#define USDHI6_VERSION		0x01c4
+#define USDHI6_HOST_MODE	0x01c8
+#define USDHI6_SDIF_MODE	0x01cc
+
+#define USDHI6_SD_CMD_APP		0x0040
+#define USDHI6_SD_CMD_MODE_RSP_AUTO	0x0000
+#define USDHI6_SD_CMD_MODE_RSP_NONE	0x0300
+#define USDHI6_SD_CMD_MODE_RSP_R1	0x0400	/* Also R5, R6, R7 */
+#define USDHI6_SD_CMD_MODE_RSP_R1B	0x0500	/* R1b */
+#define USDHI6_SD_CMD_MODE_RSP_R2	0x0600
+#define USDHI6_SD_CMD_MODE_RSP_R3	0x0700	/* Also R4 */
+#define USDHI6_SD_CMD_DATA		0x0800
+#define USDHI6_SD_CMD_READ		0x1000
+#define USDHI6_SD_CMD_MULTI		0x2000
+#define USDHI6_SD_CMD_CMD12_AUTO_OFF	0x4000
+
+#define USDHI6_CC_EXT_MODE_SDRW		BIT(1)
+
+#define USDHI6_SD_INFO1_RSP_END		BIT(0)
+#define USDHI6_SD_INFO1_ACCESS_END	BIT(2)
+#define USDHI6_SD_INFO1_CARD_OUT	BIT(3)
+#define USDHI6_SD_INFO1_CARD_IN		BIT(4)
+#define USDHI6_SD_INFO1_CD		BIT(5)
+#define USDHI6_SD_INFO1_WP		BIT(7)
+#define USDHI6_SD_INFO1_D3_CARD_OUT	BIT(8)
+#define USDHI6_SD_INFO1_D3_CARD_IN	BIT(9)
+
+#define USDHI6_SD_INFO2_CMD_ERR		BIT(0)
+#define USDHI6_SD_INFO2_CRC_ERR		BIT(1)
+#define USDHI6_SD_INFO2_END_ERR		BIT(2)
+#define USDHI6_SD_INFO2_TOUT		BIT(3)
+#define USDHI6_SD_INFO2_IWA_ERR		BIT(4)
+#define USDHI6_SD_INFO2_IRA_ERR		BIT(5)
+#define USDHI6_SD_INFO2_RSP_TOUT	BIT(6)
+#define USDHI6_SD_INFO2_SDDAT0		BIT(7)
+#define USDHI6_SD_INFO2_BRE		BIT(8)
+#define USDHI6_SD_INFO2_BWE		BIT(9)
+#define USDHI6_SD_INFO2_SCLKDIVEN	BIT(13)
+#define USDHI6_SD_INFO2_CBSY		BIT(14)
+#define USDHI6_SD_INFO2_ILA		BIT(15)
+
+#define USDHI6_SD_INFO1_CARD_INSERT (USDHI6_SD_INFO1_CARD_IN | USDHI6_SD_INFO1_D3_CARD_IN)
+#define USDHI6_SD_INFO1_CARD_EJECT (USDHI6_SD_INFO1_CARD_OUT | USDHI6_SD_INFO1_D3_CARD_OUT)
+#define USDHI6_SD_INFO1_CARD (USDHI6_SD_INFO1_CARD_INSERT | USDHI6_SD_INFO1_CARD_EJECT)
+#define USDHI6_SD_INFO1_CARD_CD (USDHI6_SD_INFO1_CARD_IN | USDHI6_SD_INFO1_CARD_OUT)
+
+#define USDHI6_SD_INFO2_ERR	(USDHI6_SD_INFO2_CMD_ERR |	\
+	USDHI6_SD_INFO2_CRC_ERR | USDHI6_SD_INFO2_END_ERR |	\
+	USDHI6_SD_INFO2_TOUT | USDHI6_SD_INFO2_IWA_ERR |	\
+	USDHI6_SD_INFO2_IRA_ERR | USDHI6_SD_INFO2_RSP_TOUT |	\
+	USDHI6_SD_INFO2_ILA)
+
+#define USDHI6_SD_INFO1_IRQ	(USDHI6_SD_INFO1_RSP_END | USDHI6_SD_INFO1_ACCESS_END | \
+				 USDHI6_SD_INFO1_CARD)
+
+#define USDHI6_SD_INFO2_IRQ	(USDHI6_SD_INFO2_ERR | USDHI6_SD_INFO2_BRE | \
+				 USDHI6_SD_INFO2_BWE | 0x0800 | USDHI6_SD_INFO2_ILA)
+
+#define USDHI6_SD_CLK_CTRL_SCLKEN	BIT(8)
+
+#define USDHI6_SD_STOP_STP		BIT(0)
+#define USDHI6_SD_STOP_SEC		BIT(8)
+
+#define USDHI6_SDIO_INFO1_IOIRQ		BIT(0)
+#define USDHI6_SDIO_INFO1_EXPUB52	BIT(14)
+#define USDHI6_SDIO_INFO1_EXWT		BIT(15)
+
+#define USDHI6_SD_ERR_STS1_CRC_NO_ERROR	BIT(13)
+
+#define USDHI6_SOFT_RST_RESERVED	(BIT(1) | BIT(2))
+#define USDHI6_SOFT_RST_RESET		BIT(0)
+
+#define USDHI6_SD_OPTION_TIMEOUT_SHIFT	4
+#define USDHI6_SD_OPTION_TIMEOUT_MASK	(0xf << USDHI6_SD_OPTION_TIMEOUT_SHIFT)
+#define USDHI6_SD_OPTION_WIDTH_1	BIT(15)
+
+#define USDHI6_SD_PORT_SEL_PORTS_SHIFT	8
+
+#define USDHI6_SD_CLK_CTRL_DIV_MASK	0xff
+
+#define USDHI6_SDIO_INFO1_IRQ	(USDHI6_SDIO_INFO1_IOIRQ | 3 | \
+				 USDHI6_SDIO_INFO1_EXPUB52 | USDHI6_SDIO_INFO1_EXWT)
+
+#define USDHI6_MIN_DMA 64
+
+#define USDHI6_REQ_TIMEOUT_MS 4000
+
+enum usdhi6_wait_for {
+	USDHI6_WAIT_FOR_REQUEST,
+	USDHI6_WAIT_FOR_CMD,
+	USDHI6_WAIT_FOR_MREAD,
+	USDHI6_WAIT_FOR_MWRITE,
+	USDHI6_WAIT_FOR_READ,
+	USDHI6_WAIT_FOR_WRITE,
+	USDHI6_WAIT_FOR_DATA_END,
+	USDHI6_WAIT_FOR_STOP,
+	USDHI6_WAIT_FOR_DMA,
+};
+
+struct usdhi6_page {
+	struct page *page;
+	void *mapped;		/* mapped page */
+};
+
+struct usdhi6_host {
+	struct mmc_host *mmc;
+	struct mmc_request *mrq;
+	void __iomem *base;
+	struct clk *clk;
+
+	/* SG memory handling */
+
+	/* Common for multiple and single block requests */
+	struct usdhi6_page pg;	/* current page from an SG */
+	void *blk_page;		/* either a mapped page, or the bounce buffer */
+	size_t offset;		/* offset within a page, including sg->offset */
+
+	/* Blocks, crossing a page boundary */
+	size_t head_len;
+	struct usdhi6_page head_pg;
+
+	/* A bounce buffer for unaligned blocks or blocks, crossing a page boundary */
+	struct scatterlist bounce_sg;
+	u8 bounce_buf[512];
+
+	/* Multiple block requests only */
+	struct scatterlist *sg;	/* current SG segment */
+	int page_idx;		/* page index within an SG segment */
+
+	enum usdhi6_wait_for wait;
+	u32 status_mask;
+	u32 status2_mask;
+	u32 sdio_mask;
+	u32 io_error;
+	u32 irq_status;
+	unsigned long imclk;
+	unsigned long rate;
+	bool app_cmd;
+
+	/* Timeout handling */
+	struct delayed_work timeout_work;
+	unsigned long timeout;
+
+	/* DMA support */
+	struct dma_chan *chan_rx;
+	struct dma_chan *chan_tx;
+	bool dma_active;
+
+	/* Pin control */
+	struct pinctrl *pinctrl;
+	struct pinctrl_state *pins_uhs;
+};
+
+/*			I/O primitives					*/
+
+static void usdhi6_write(struct usdhi6_host *host, u32 reg, u32 data)
+{
+	iowrite32(data, host->base + reg);
+	dev_vdbg(mmc_dev(host->mmc), "%s(0x%p + 0x%x) = 0x%x\n", __func__,
+		host->base, reg, data);
+}
+
+static void usdhi6_write16(struct usdhi6_host *host, u32 reg, u16 data)
+{
+	iowrite16(data, host->base + reg);
+	dev_vdbg(mmc_dev(host->mmc), "%s(0x%p + 0x%x) = 0x%x\n", __func__,
+		host->base, reg, data);
+}
+
+static u32 usdhi6_read(struct usdhi6_host *host, u32 reg)
+{
+	u32 data = ioread32(host->base + reg);
+	dev_vdbg(mmc_dev(host->mmc), "%s(0x%p + 0x%x) = 0x%x\n", __func__,
+		host->base, reg, data);
+	return data;
+}
+
+static u16 usdhi6_read16(struct usdhi6_host *host, u32 reg)
+{
+	u16 data = ioread16(host->base + reg);
+	dev_vdbg(mmc_dev(host->mmc), "%s(0x%p + 0x%x) = 0x%x\n", __func__,
+		host->base, reg, data);
+	return data;
+}
+
+static void usdhi6_irq_enable(struct usdhi6_host *host, u32 info1, u32 info2)
+{
+	host->status_mask = USDHI6_SD_INFO1_IRQ & ~info1;
+	host->status2_mask = USDHI6_SD_INFO2_IRQ & ~info2;
+	usdhi6_write(host, USDHI6_SD_INFO1_MASK, host->status_mask);
+	usdhi6_write(host, USDHI6_SD_INFO2_MASK, host->status2_mask);
+}
+
+static void usdhi6_wait_for_resp(struct usdhi6_host *host)
+{
+	usdhi6_irq_enable(host, USDHI6_SD_INFO1_RSP_END |
+			  USDHI6_SD_INFO1_ACCESS_END | USDHI6_SD_INFO1_CARD_CD,
+			  USDHI6_SD_INFO2_ERR);
+}
+
+static void usdhi6_wait_for_brwe(struct usdhi6_host *host, bool read)
+{
+	usdhi6_irq_enable(host, USDHI6_SD_INFO1_ACCESS_END |
+			  USDHI6_SD_INFO1_CARD_CD, USDHI6_SD_INFO2_ERR |
+			  (read ? USDHI6_SD_INFO2_BRE : USDHI6_SD_INFO2_BWE));
+}
+
+static void usdhi6_only_cd(struct usdhi6_host *host)
+{
+	/* Mask all except card hotplug */
+	usdhi6_irq_enable(host, USDHI6_SD_INFO1_CARD_CD, 0);
+}
+
+static void usdhi6_mask_all(struct usdhi6_host *host)
+{
+	usdhi6_irq_enable(host, 0, 0);
+}
+
+static int usdhi6_error_code(struct usdhi6_host *host)
+{
+	u32 err;
+
+	usdhi6_write(host, USDHI6_SD_STOP, USDHI6_SD_STOP_STP);
+
+	if (host->io_error &
+	    (USDHI6_SD_INFO2_RSP_TOUT | USDHI6_SD_INFO2_TOUT)) {
+		u32 rsp54 = usdhi6_read(host, USDHI6_SD_RSP54);
+		int opc = host->mrq ? host->mrq->cmd->opcode : -1;
+
+		err = usdhi6_read(host, USDHI6_SD_ERR_STS2);
+		/* Response timeout is often normal, don't spam the log */
+		if (host->wait == USDHI6_WAIT_FOR_CMD)
+			dev_dbg(mmc_dev(host->mmc),
+				"T-out sts 0x%x, resp 0x%x, state %u, CMD%d\n",
+				err, rsp54, host->wait, opc);
+		else
+			dev_warn(mmc_dev(host->mmc),
+				 "T-out sts 0x%x, resp 0x%x, state %u, CMD%d\n",
+				 err, rsp54, host->wait, opc);
+		return -ETIMEDOUT;
+	}
+
+	err = usdhi6_read(host, USDHI6_SD_ERR_STS1);
+	if (err != USDHI6_SD_ERR_STS1_CRC_NO_ERROR)
+		dev_warn(mmc_dev(host->mmc), "Err sts 0x%x, state %u, CMD%d\n",
+			 err, host->wait, host->mrq ? host->mrq->cmd->opcode : -1);
+	if (host->io_error & USDHI6_SD_INFO2_ILA)
+		return -EILSEQ;
+
+	return -EIO;
+}
+
+/*			Scatter-Gather management			*/
+
+/*
+ * In PIO mode we have to map each page separately, using kmap(). That way
+ * adjacent pages are mapped to non-adjacent virtual addresses. That's why we
+ * have to use a bounce buffer for blocks, crossing page boundaries. Such blocks
+ * have been observed with an SDIO WiFi card (b43 driver).
+ */
+static void usdhi6_blk_bounce(struct usdhi6_host *host,
+			      struct scatterlist *sg)
+{
+	struct mmc_data *data = host->mrq->data;
+	size_t blk_head = host->head_len;
+
+	dev_dbg(mmc_dev(host->mmc), "%s(): CMD%u of %u SG: %ux%u @ 0x%x\n",
+		__func__, host->mrq->cmd->opcode, data->sg_len,
+		data->blksz, data->blocks, sg->offset);
+
+	host->head_pg.page	= host->pg.page;
+	host->head_pg.mapped	= host->pg.mapped;
+	host->pg.page		= nth_page(host->pg.page, 1);
+	host->pg.mapped		= kmap(host->pg.page);
+
+	host->blk_page = host->bounce_buf;
+	host->offset = 0;
+
+	if (data->flags & MMC_DATA_READ)
+		return;
+
+	memcpy(host->bounce_buf, host->head_pg.mapped + PAGE_SIZE - blk_head,
+	       blk_head);
+	memcpy(host->bounce_buf + blk_head, host->pg.mapped,
+	       data->blksz - blk_head);
+}
+
+/* Only called for multiple block IO */
+static void usdhi6_sg_prep(struct usdhi6_host *host)
+{
+	struct mmc_request *mrq = host->mrq;
+	struct mmc_data *data = mrq->data;
+
+	usdhi6_write(host, USDHI6_SD_SECCNT, data->blocks);
+
+	host->sg = data->sg;
+	/* TODO: if we always map, this is redundant */
+	host->offset = host->sg->offset;
+}
+
+/* Map the first page in an SG segment: common for multiple and single block IO */
+static void *usdhi6_sg_map(struct usdhi6_host *host)
+{
+	struct mmc_data *data = host->mrq->data;
+	struct scatterlist *sg = data->sg_len > 1 ? host->sg : data->sg;
+	size_t head = PAGE_SIZE - sg->offset;
+	size_t blk_head = head % data->blksz;
+
+	WARN(host->pg.page, "%p not properly unmapped!\n", host->pg.page);
+	if (WARN(sg_dma_len(sg) % data->blksz,
+		 "SG size %u isn't a multiple of block size %u\n",
+		 sg_dma_len(sg), data->blksz))
+		return NULL;
+
+	host->pg.page = sg_page(sg);
+	host->pg.mapped = kmap(host->pg.page);
+	host->offset = sg->offset;
+
+	/*
+	 * Block size must be a power of 2 for multi-block transfers,
+	 * therefore blk_head is equal for all pages in this SG
+	 */
+	host->head_len = blk_head;
+
+	if (head < data->blksz)
+		/*
+		 * The first block in the SG crosses a page boundary.
+		 * Max blksz = 512, so blocks can only span 2 pages
+		 */
+		usdhi6_blk_bounce(host, sg);
+	else
+		host->blk_page = host->pg.mapped;
+
+	dev_dbg(mmc_dev(host->mmc), "Mapped %p (%lx) at %p + %u for CMD%u @ 0x%p\n",
+		host->pg.page, page_to_pfn(host->pg.page), host->pg.mapped,
+		sg->offset, host->mrq->cmd->opcode, host->mrq);
+
+	return host->blk_page + host->offset;
+}
+
+/* Unmap the current page: common for multiple and single block IO */
+static void usdhi6_sg_unmap(struct usdhi6_host *host, bool force)
+{
+	struct mmc_data *data = host->mrq->data;
+	struct page *page = host->head_pg.page;
+
+	if (page) {
+		/* Previous block was cross-page boundary */
+		struct scatterlist *sg = data->sg_len > 1 ?
+			host->sg : data->sg;
+		size_t blk_head = host->head_len;
+
+		if (!data->error && data->flags & MMC_DATA_READ) {
+			memcpy(host->head_pg.mapped + PAGE_SIZE - blk_head,
+			       host->bounce_buf, blk_head);
+			memcpy(host->pg.mapped, host->bounce_buf + blk_head,
+			       data->blksz - blk_head);
+		}
+
+		flush_dcache_page(page);
+		kunmap(page);
+
+		host->head_pg.page = NULL;
+
+		if (!force && sg_dma_len(sg) + sg->offset >
+		    (host->page_idx << PAGE_SHIFT) + data->blksz - blk_head)
+			/* More blocks in this SG, don't unmap the next page */
+			return;
+	}
+
+	page = host->pg.page;
+	if (!page)
+		return;
+
+	flush_dcache_page(page);
+	kunmap(page);
+
+	host->pg.page = NULL;
+}
+
+/* Called from MMC_WRITE_MULTIPLE_BLOCK or MMC_READ_MULTIPLE_BLOCK */
+static void usdhi6_sg_advance(struct usdhi6_host *host)
+{
+	struct mmc_data *data = host->mrq->data;
+	size_t done, total;
+
+	/* New offset: set at the end of the previous block */
+	if (host->head_pg.page) {
+		/* Finished a cross-page block, jump to the new page */
+		host->page_idx++;
+		host->offset = data->blksz - host->head_len;
+		host->blk_page = host->pg.mapped;
+		usdhi6_sg_unmap(host, false);
+	} else {
+		host->offset += data->blksz;
+		/* The completed block didn't cross a page boundary */
+		if (host->offset == PAGE_SIZE) {
+			/* If required, we'll map the page below */
+			host->offset = 0;
+			host->page_idx++;
+		}
+	}
+
+	/*
+	 * Now host->blk_page + host->offset point at the end of our last block
+	 * and host->page_idx is the index of the page, in which our new block
+	 * is located, if any
+	 */
+
+	done = (host->page_idx << PAGE_SHIFT) + host->offset;
+	total = host->sg->offset + sg_dma_len(host->sg);
+
+	dev_dbg(mmc_dev(host->mmc), "%s(): %zu of %zu @ %zu\n", __func__,
+		done, total, host->offset);
+
+	if (done < total && host->offset) {
+		/* More blocks in this page */
+		if (host->offset + data->blksz > PAGE_SIZE)
+			/* We approached at a block, that spans 2 pages */
+			usdhi6_blk_bounce(host, host->sg);
+
+		return;
+	}
+
+	/* Finished current page or an SG segment */
+	usdhi6_sg_unmap(host, false);
+
+	if (done == total) {
+		/*
+		 * End of an SG segment or the complete SG: jump to the next
+		 * segment, we'll map it later in usdhi6_blk_read() or
+		 * usdhi6_blk_write()
+		 */
+		struct scatterlist *next = sg_next(host->sg);
+
+		host->page_idx = 0;
+
+		if (!next)
+			host->wait = USDHI6_WAIT_FOR_DATA_END;
+		host->sg = next;
+
+		if (WARN(next && sg_dma_len(next) % data->blksz,
+			 "SG size %u isn't a multiple of block size %u\n",
+			 sg_dma_len(next), data->blksz))
+			data->error = -EINVAL;
+
+		return;
+	}
+
+	/* We cannot get here after crossing a page border */
+
+	/* Next page in the same SG */
+	host->pg.page = nth_page(sg_page(host->sg), host->page_idx);
+	host->pg.mapped = kmap(host->pg.page);
+	host->blk_page = host->pg.mapped;
+
+	dev_dbg(mmc_dev(host->mmc), "Mapped %p (%lx) at %p for CMD%u @ 0x%p\n",
+		host->pg.page, page_to_pfn(host->pg.page), host->pg.mapped,
+		host->mrq->cmd->opcode, host->mrq);
+}
+
+/*			DMA handling					*/
+
+static void usdhi6_dma_release(struct usdhi6_host *host)
+{
+	host->dma_active = false;
+	if (host->chan_tx) {
+		struct dma_chan *chan = host->chan_tx;
+		host->chan_tx = NULL;
+		dma_release_channel(chan);
+	}
+	if (host->chan_rx) {
+		struct dma_chan *chan = host->chan_rx;
+		host->chan_rx = NULL;
+		dma_release_channel(chan);
+	}
+}
+
+static void usdhi6_dma_stop_unmap(struct usdhi6_host *host)
+{
+	struct mmc_data *data = host->mrq->data;
+
+	if (!host->dma_active)
+		return;
+
+	usdhi6_write(host, USDHI6_CC_EXT_MODE, 0);
+	host->dma_active = false;
+
+	if (data->flags & MMC_DATA_READ)
+		dma_unmap_sg(host->chan_rx->device->dev, data->sg,
+			     data->sg_len, DMA_FROM_DEVICE);
+	else
+		dma_unmap_sg(host->chan_tx->device->dev, data->sg,
+			     data->sg_len, DMA_TO_DEVICE);
+}
+
+static void usdhi6_dma_complete(void *arg)
+{
+	struct usdhi6_host *host = arg;
+	struct mmc_request *mrq = host->mrq;
+
+	if (WARN(!mrq || !mrq->data, "%s: NULL data in DMA completion for %p!\n",
+		 dev_name(mmc_dev(host->mmc)), mrq))
+		return;
+
+	dev_dbg(mmc_dev(host->mmc), "%s(): CMD%u DMA completed\n", __func__,
+		mrq->cmd->opcode);
+
+	usdhi6_dma_stop_unmap(host);
+	usdhi6_wait_for_brwe(host, mrq->data->flags & MMC_DATA_READ);
+}
+
+static int usdhi6_dma_setup(struct usdhi6_host *host, struct dma_chan *chan,
+			    enum dma_transfer_direction dir)
+{
+	struct mmc_data *data = host->mrq->data;
+	struct scatterlist *sg = data->sg;
+	struct dma_async_tx_descriptor *desc = NULL;
+	dma_cookie_t cookie = -EINVAL;
+	enum dma_data_direction data_dir;
+	int ret;
+
+	switch (dir) {
+	case DMA_MEM_TO_DEV:
+		data_dir = DMA_TO_DEVICE;
+		break;
+	case DMA_DEV_TO_MEM:
+		data_dir = DMA_FROM_DEVICE;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	ret = dma_map_sg(chan->device->dev, sg, data->sg_len, data_dir);
+	if (ret > 0) {
+		host->dma_active = true;
+		desc = dmaengine_prep_slave_sg(chan, sg, ret, dir,
+					DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	}
+
+	if (desc) {
+		desc->callback = usdhi6_dma_complete;
+		desc->callback_param = host;
+		cookie = dmaengine_submit(desc);
+	}
+
+	dev_dbg(mmc_dev(host->mmc), "%s(): mapped %d -> %d, cookie %d @ %p\n",
+		__func__, data->sg_len, ret, cookie, desc);
+
+	if (cookie < 0) {
+		/* DMA failed, fall back to PIO */
+		if (ret >= 0)
+			ret = cookie;
+		usdhi6_dma_release(host);
+		dev_warn(mmc_dev(host->mmc),
+			 "DMA failed: %d, falling back to PIO\n", ret);
+	}
+
+	return cookie;
+}
+
+static int usdhi6_dma_start(struct usdhi6_host *host)
+{
+	if (!host->chan_rx || !host->chan_tx)
+		return -ENODEV;
+
+	if (host->mrq->data->flags & MMC_DATA_READ)
+		return usdhi6_dma_setup(host, host->chan_rx, DMA_DEV_TO_MEM);
+
+	return usdhi6_dma_setup(host, host->chan_tx, DMA_MEM_TO_DEV);
+}
+
+static void usdhi6_dma_kill(struct usdhi6_host *host)
+{
+	struct mmc_data *data = host->mrq->data;
+
+	dev_dbg(mmc_dev(host->mmc), "%s(): SG of %u: %ux%u\n",
+		__func__, data->sg_len, data->blocks, data->blksz);
+	/* Abort DMA */
+	if (data->flags & MMC_DATA_READ)
+		dmaengine_terminate_all(host->chan_rx);
+	else
+		dmaengine_terminate_all(host->chan_tx);
+}
+
+static void usdhi6_dma_check_error(struct usdhi6_host *host)
+{
+	struct mmc_data *data = host->mrq->data;
+
+	dev_dbg(mmc_dev(host->mmc), "%s(): IO error %d, status 0x%x\n",
+		__func__, host->io_error, usdhi6_read(host, USDHI6_SD_INFO1));
+
+	if (host->io_error) {
+		data->error = usdhi6_error_code(host);
+		data->bytes_xfered = 0;
+		usdhi6_dma_kill(host);
+		usdhi6_dma_release(host);
+		dev_warn(mmc_dev(host->mmc),
+			 "DMA failed: %d, falling back to PIO\n", data->error);
+		return;
+	}
+
+	/*
+	 * The datasheet tells us to check a response from the card, whereas
+	 * responses only come after the command phase, not after the data
+	 * phase. Let's check anyway.
+	 */
+	if (host->irq_status & USDHI6_SD_INFO1_RSP_END)
+		dev_warn(mmc_dev(host->mmc), "Unexpected response received!\n");
+}
+
+static void usdhi6_dma_kick(struct usdhi6_host *host)
+{
+	if (host->mrq->data->flags & MMC_DATA_READ)
+		dma_async_issue_pending(host->chan_rx);
+	else
+		dma_async_issue_pending(host->chan_tx);
+}
+
+static void usdhi6_dma_request(struct usdhi6_host *host, phys_addr_t start)
+{
+	struct dma_slave_config cfg = {
+		.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+		.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+	};
+	int ret;
+
+	host->chan_tx = dma_request_chan(mmc_dev(host->mmc), "tx");
+	dev_dbg(mmc_dev(host->mmc), "%s: TX: got channel %p\n", __func__,
+		host->chan_tx);
+
+	if (IS_ERR(host->chan_tx)) {
+		host->chan_tx = NULL;
+		return;
+	}
+
+	cfg.direction = DMA_MEM_TO_DEV;
+	cfg.dst_addr = start + USDHI6_SD_BUF0;
+	cfg.dst_maxburst = 128;	/* 128 words * 4 bytes = 512 bytes */
+	cfg.src_addr = 0;
+	ret = dmaengine_slave_config(host->chan_tx, &cfg);
+	if (ret < 0)
+		goto e_release_tx;
+
+	host->chan_rx = dma_request_chan(mmc_dev(host->mmc), "rx");
+	dev_dbg(mmc_dev(host->mmc), "%s: RX: got channel %p\n", __func__,
+		host->chan_rx);
+
+	if (IS_ERR(host->chan_rx)) {
+		host->chan_rx = NULL;
+		goto e_release_tx;
+	}
+
+	cfg.direction = DMA_DEV_TO_MEM;
+	cfg.src_addr = cfg.dst_addr;
+	cfg.src_maxburst = 128;	/* 128 words * 4 bytes = 512 bytes */
+	cfg.dst_addr = 0;
+	ret = dmaengine_slave_config(host->chan_rx, &cfg);
+	if (ret < 0)
+		goto e_release_rx;
+
+	return;
+
+e_release_rx:
+	dma_release_channel(host->chan_rx);
+	host->chan_rx = NULL;
+e_release_tx:
+	dma_release_channel(host->chan_tx);
+	host->chan_tx = NULL;
+}
+
+/*			API helpers					*/
+
+static void usdhi6_clk_set(struct usdhi6_host *host, struct mmc_ios *ios)
+{
+	unsigned long rate = ios->clock;
+	u32 val;
+	unsigned int i;
+
+	for (i = 1000; i; i--) {
+		if (usdhi6_read(host, USDHI6_SD_INFO2) & USDHI6_SD_INFO2_SCLKDIVEN)
+			break;
+		usleep_range(10, 100);
+	}
+
+	if (!i) {
+		dev_err(mmc_dev(host->mmc), "SD bus busy, clock set aborted\n");
+		return;
+	}
+
+	val = usdhi6_read(host, USDHI6_SD_CLK_CTRL) & ~USDHI6_SD_CLK_CTRL_DIV_MASK;
+
+	if (rate) {
+		unsigned long new_rate;
+
+		if (host->imclk <= rate) {
+			if (ios->timing != MMC_TIMING_UHS_DDR50) {
+				/* Cannot have 1-to-1 clock in DDR mode */
+				new_rate = host->imclk;
+				val |= 0xff;
+			} else {
+				new_rate = host->imclk / 2;
+			}
+		} else {
+			unsigned long div =
+				roundup_pow_of_two(DIV_ROUND_UP(host->imclk, rate));
+			val |= div >> 2;
+			new_rate = host->imclk / div;
+		}
+
+		if (host->rate == new_rate)
+			return;
+
+		host->rate = new_rate;
+
+		dev_dbg(mmc_dev(host->mmc), "target %lu, div %u, set %lu\n",
+			rate, (val & 0xff) << 2, new_rate);
+	}
+
+	/*
+	 * if old or new rate is equal to input rate, have to switch the clock
+	 * off before changing and on after
+	 */
+	if (host->imclk == rate || host->imclk == host->rate || !rate)
+		usdhi6_write(host, USDHI6_SD_CLK_CTRL,
+			     val & ~USDHI6_SD_CLK_CTRL_SCLKEN);
+
+	if (!rate) {
+		host->rate = 0;
+		return;
+	}
+
+	usdhi6_write(host, USDHI6_SD_CLK_CTRL, val);
+
+	if (host->imclk == rate || host->imclk == host->rate ||
+	    !(val & USDHI6_SD_CLK_CTRL_SCLKEN))
+		usdhi6_write(host, USDHI6_SD_CLK_CTRL,
+			     val | USDHI6_SD_CLK_CTRL_SCLKEN);
+}
+
+static void usdhi6_set_power(struct usdhi6_host *host, struct mmc_ios *ios)
+{
+	struct mmc_host *mmc = host->mmc;
+
+	if (!IS_ERR(mmc->supply.vmmc))
+		/* Errors ignored... */
+		mmc_regulator_set_ocr(mmc, mmc->supply.vmmc,
+				      ios->power_mode ? ios->vdd : 0);
+}
+
+static int usdhi6_reset(struct usdhi6_host *host)
+{
+	int i;
+
+	usdhi6_write(host, USDHI6_SOFT_RST, USDHI6_SOFT_RST_RESERVED);
+	cpu_relax();
+	usdhi6_write(host, USDHI6_SOFT_RST, USDHI6_SOFT_RST_RESERVED | USDHI6_SOFT_RST_RESET);
+	for (i = 1000; i; i--)
+		if (usdhi6_read(host, USDHI6_SOFT_RST) & USDHI6_SOFT_RST_RESET)
+			break;
+
+	return i ? 0 : -ETIMEDOUT;
+}
+
+static void usdhi6_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct usdhi6_host *host = mmc_priv(mmc);
+	u32 option, mode;
+	int ret;
+
+	dev_dbg(mmc_dev(mmc), "%uHz, OCR: %u, power %u, bus-width %u, timing %u\n",
+		ios->clock, ios->vdd, ios->power_mode, ios->bus_width, ios->timing);
+
+	switch (ios->power_mode) {
+	case MMC_POWER_OFF:
+		usdhi6_set_power(host, ios);
+		usdhi6_only_cd(host);
+		break;
+	case MMC_POWER_UP:
+		/*
+		 * We only also touch USDHI6_SD_OPTION from .request(), which
+		 * cannot race with MMC_POWER_UP
+		 */
+		ret = usdhi6_reset(host);
+		if (ret < 0) {
+			dev_err(mmc_dev(mmc), "Cannot reset the interface!\n");
+		} else {
+			usdhi6_set_power(host, ios);
+			usdhi6_only_cd(host);
+		}
+		break;
+	case MMC_POWER_ON:
+		option = usdhi6_read(host, USDHI6_SD_OPTION);
+		/*
+		 * The eMMC standard only allows 4 or 8 bits in the DDR mode,
+		 * the same probably holds for SD cards. We check here anyway,
+		 * since the datasheet explicitly requires 4 bits for DDR.
+		 */
+		if (ios->bus_width == MMC_BUS_WIDTH_1) {
+			if (ios->timing == MMC_TIMING_UHS_DDR50)
+				dev_err(mmc_dev(mmc),
+					"4 bits are required for DDR\n");
+			option |= USDHI6_SD_OPTION_WIDTH_1;
+			mode = 0;
+		} else {
+			option &= ~USDHI6_SD_OPTION_WIDTH_1;
+			mode = ios->timing == MMC_TIMING_UHS_DDR50;
+		}
+		usdhi6_write(host, USDHI6_SD_OPTION, option);
+		usdhi6_write(host, USDHI6_SDIF_MODE, mode);
+		break;
+	}
+
+	if (host->rate != ios->clock)
+		usdhi6_clk_set(host, ios);
+}
+
+/* This is data timeout. Response timeout is fixed to 640 clock cycles */
+static void usdhi6_timeout_set(struct usdhi6_host *host)
+{
+	struct mmc_request *mrq = host->mrq;
+	u32 val;
+	unsigned long ticks;
+
+	if (!mrq->data)
+		ticks = host->rate / 1000 * mrq->cmd->busy_timeout;
+	else
+		ticks = host->rate / 1000000 * (mrq->data->timeout_ns / 1000) +
+			mrq->data->timeout_clks;
+
+	if (!ticks || ticks > 1 << 27)
+		/* Max timeout */
+		val = 14;
+	else if (ticks < 1 << 13)
+		/* Min timeout */
+		val = 0;
+	else
+		val = order_base_2(ticks) - 13;
+
+	dev_dbg(mmc_dev(host->mmc), "Set %s timeout %lu ticks @ %lu Hz\n",
+		mrq->data ? "data" : "cmd", ticks, host->rate);
+
+	/* Timeout Counter mask: 0xf0 */
+	usdhi6_write(host, USDHI6_SD_OPTION, (val << USDHI6_SD_OPTION_TIMEOUT_SHIFT) |
+		     (usdhi6_read(host, USDHI6_SD_OPTION) & ~USDHI6_SD_OPTION_TIMEOUT_MASK));
+}
+
+static void usdhi6_request_done(struct usdhi6_host *host)
+{
+	struct mmc_request *mrq = host->mrq;
+	struct mmc_data *data = mrq->data;
+
+	if (WARN(host->pg.page || host->head_pg.page,
+		 "Page %p or %p not unmapped: wait %u, CMD%d(%c) @ +0x%zx %ux%u in SG%u!\n",
+		 host->pg.page, host->head_pg.page, host->wait, mrq->cmd->opcode,
+		 data ? (data->flags & MMC_DATA_READ ? 'R' : 'W') : '-',
+		 data ? host->offset : 0, data ? data->blocks : 0,
+		 data ? data->blksz : 0, data ? data->sg_len : 0))
+		usdhi6_sg_unmap(host, true);
+
+	if (mrq->cmd->error ||
+	    (data && data->error) ||
+	    (mrq->stop && mrq->stop->error))
+		dev_dbg(mmc_dev(host->mmc), "%s(CMD%d: %ux%u): err %d %d %d\n",
+			__func__, mrq->cmd->opcode, data ? data->blocks : 0,
+			data ? data->blksz : 0,
+			mrq->cmd->error,
+			data ? data->error : 1,
+			mrq->stop ? mrq->stop->error : 1);
+
+	/* Disable DMA */
+	usdhi6_write(host, USDHI6_CC_EXT_MODE, 0);
+	host->wait = USDHI6_WAIT_FOR_REQUEST;
+	host->mrq = NULL;
+
+	mmc_request_done(host->mmc, mrq);
+}
+
+static int usdhi6_cmd_flags(struct usdhi6_host *host)
+{
+	struct mmc_request *mrq = host->mrq;
+	struct mmc_command *cmd = mrq->cmd;
+	u16 opc = cmd->opcode;
+
+	if (host->app_cmd) {
+		host->app_cmd = false;
+		opc |= USDHI6_SD_CMD_APP;
+	}
+
+	if (mrq->data) {
+		opc |= USDHI6_SD_CMD_DATA;
+
+		if (mrq->data->flags & MMC_DATA_READ)
+			opc |= USDHI6_SD_CMD_READ;
+
+		if (cmd->opcode == MMC_READ_MULTIPLE_BLOCK ||
+		    cmd->opcode == MMC_WRITE_MULTIPLE_BLOCK ||
+		    (cmd->opcode == SD_IO_RW_EXTENDED &&
+		     mrq->data->blocks > 1)) {
+			opc |= USDHI6_SD_CMD_MULTI;
+			if (!mrq->stop)
+				opc |= USDHI6_SD_CMD_CMD12_AUTO_OFF;
+		}
+
+		switch (mmc_resp_type(cmd)) {
+		case MMC_RSP_NONE:
+			opc |= USDHI6_SD_CMD_MODE_RSP_NONE;
+			break;
+		case MMC_RSP_R1:
+			opc |= USDHI6_SD_CMD_MODE_RSP_R1;
+			break;
+		case MMC_RSP_R1B:
+			opc |= USDHI6_SD_CMD_MODE_RSP_R1B;
+			break;
+		case MMC_RSP_R2:
+			opc |= USDHI6_SD_CMD_MODE_RSP_R2;
+			break;
+		case MMC_RSP_R3:
+			opc |= USDHI6_SD_CMD_MODE_RSP_R3;
+			break;
+		default:
+			dev_warn(mmc_dev(host->mmc),
+				 "Unknown response type %d\n",
+				 mmc_resp_type(cmd));
+			return -EINVAL;
+		}
+	}
+
+	return opc;
+}
+
+static int usdhi6_rq_start(struct usdhi6_host *host)
+{
+	struct mmc_request *mrq = host->mrq;
+	struct mmc_command *cmd = mrq->cmd;
+	struct mmc_data *data = mrq->data;
+	int opc = usdhi6_cmd_flags(host);
+	int i;
+
+	if (opc < 0)
+		return opc;
+
+	for (i = 1000; i; i--) {
+		if (!(usdhi6_read(host, USDHI6_SD_INFO2) & USDHI6_SD_INFO2_CBSY))
+			break;
+		usleep_range(10, 100);
+	}
+
+	if (!i) {
+		dev_dbg(mmc_dev(host->mmc), "Command active, request aborted\n");
+		return -EAGAIN;
+	}
+
+	if (data) {
+		bool use_dma;
+		int ret = 0;
+
+		host->page_idx = 0;
+
+		if (cmd->opcode == SD_IO_RW_EXTENDED && data->blocks > 1) {
+			switch (data->blksz) {
+			case 512:
+				break;
+			case 32:
+			case 64:
+			case 128:
+			case 256:
+				if (mrq->stop)
+					ret = -EINVAL;
+				break;
+			default:
+				ret = -EINVAL;
+			}
+		} else if ((cmd->opcode == MMC_READ_MULTIPLE_BLOCK ||
+			    cmd->opcode == MMC_WRITE_MULTIPLE_BLOCK) &&
+			   data->blksz != 512) {
+			ret = -EINVAL;
+		}
+
+		if (ret < 0) {
+			dev_warn(mmc_dev(host->mmc), "%s(): %u blocks of %u bytes\n",
+				 __func__, data->blocks, data->blksz);
+			return -EINVAL;
+		}
+
+		if (cmd->opcode == MMC_READ_MULTIPLE_BLOCK ||
+		    cmd->opcode == MMC_WRITE_MULTIPLE_BLOCK ||
+		    (cmd->opcode == SD_IO_RW_EXTENDED &&
+		     data->blocks > 1))
+			usdhi6_sg_prep(host);
+
+		usdhi6_write(host, USDHI6_SD_SIZE, data->blksz);
+
+		if ((data->blksz >= USDHI6_MIN_DMA ||
+		     data->blocks > 1) &&
+		    (data->blksz % 4 ||
+		     data->sg->offset % 4))
+			dev_dbg(mmc_dev(host->mmc),
+				"Bad SG of %u: %ux%u @ %u\n", data->sg_len,
+				data->blksz, data->blocks, data->sg->offset);
+
+		/* Enable DMA for USDHI6_MIN_DMA bytes or more */
+		use_dma = data->blksz >= USDHI6_MIN_DMA &&
+			!(data->blksz % 4) &&
+			usdhi6_dma_start(host) >= DMA_MIN_COOKIE;
+
+		if (use_dma)
+			usdhi6_write(host, USDHI6_CC_EXT_MODE, USDHI6_CC_EXT_MODE_SDRW);
+
+		dev_dbg(mmc_dev(host->mmc),
+			"%s(): request opcode %u, %u blocks of %u bytes in %u segments, %s %s @+0x%x%s\n",
+			__func__, cmd->opcode, data->blocks, data->blksz,
+			data->sg_len, use_dma ? "DMA" : "PIO",
+			data->flags & MMC_DATA_READ ? "read" : "write",
+			data->sg->offset, mrq->stop ? " + stop" : "");
+	} else {
+		dev_dbg(mmc_dev(host->mmc), "%s(): request opcode %u\n",
+			__func__, cmd->opcode);
+	}
+
+	/* We have to get a command completion interrupt with DMA too */
+	usdhi6_wait_for_resp(host);
+
+	host->wait = USDHI6_WAIT_FOR_CMD;
+	schedule_delayed_work(&host->timeout_work, host->timeout);
+
+	/* SEC bit is required to enable block counting by the core */
+	usdhi6_write(host, USDHI6_SD_STOP,
+		     data && data->blocks > 1 ? USDHI6_SD_STOP_SEC : 0);
+	usdhi6_write(host, USDHI6_SD_ARG, cmd->arg);
+
+	/* Kick command execution */
+	usdhi6_write(host, USDHI6_SD_CMD, opc);
+
+	return 0;
+}
+
+static void usdhi6_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct usdhi6_host *host = mmc_priv(mmc);
+	int ret;
+
+	cancel_delayed_work_sync(&host->timeout_work);
+
+	host->mrq = mrq;
+	host->sg = NULL;
+
+	usdhi6_timeout_set(host);
+	ret = usdhi6_rq_start(host);
+	if (ret < 0) {
+		mrq->cmd->error = ret;
+		usdhi6_request_done(host);
+	}
+}
+
+static int usdhi6_get_cd(struct mmc_host *mmc)
+{
+	struct usdhi6_host *host = mmc_priv(mmc);
+	/* Read is atomic, no need to lock */
+	u32 status = usdhi6_read(host, USDHI6_SD_INFO1) & USDHI6_SD_INFO1_CD;
+
+/*
+ *	level	status.CD	CD_ACTIVE_HIGH	card present
+ *	1	0		0		0
+ *	1	0		1		1
+ *	0	1		0		1
+ *	0	1		1		0
+ */
+	return !status ^ !(mmc->caps2 & MMC_CAP2_CD_ACTIVE_HIGH);
+}
+
+static int usdhi6_get_ro(struct mmc_host *mmc)
+{
+	struct usdhi6_host *host = mmc_priv(mmc);
+	/* No locking as above */
+	u32 status = usdhi6_read(host, USDHI6_SD_INFO1) & USDHI6_SD_INFO1_WP;
+
+/*
+ *	level	status.WP	RO_ACTIVE_HIGH	card read-only
+ *	1	0		0		0
+ *	1	0		1		1
+ *	0	1		0		1
+ *	0	1		1		0
+ */
+	return !status ^ !(mmc->caps2 & MMC_CAP2_RO_ACTIVE_HIGH);
+}
+
+static void usdhi6_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct usdhi6_host *host = mmc_priv(mmc);
+
+	dev_dbg(mmc_dev(mmc), "%s(): %sable\n", __func__, enable ? "en" : "dis");
+
+	if (enable) {
+		host->sdio_mask = USDHI6_SDIO_INFO1_IRQ & ~USDHI6_SDIO_INFO1_IOIRQ;
+		usdhi6_write(host, USDHI6_SDIO_INFO1_MASK, host->sdio_mask);
+		usdhi6_write(host, USDHI6_SDIO_MODE, 1);
+	} else {
+		usdhi6_write(host, USDHI6_SDIO_MODE, 0);
+		usdhi6_write(host, USDHI6_SDIO_INFO1_MASK, USDHI6_SDIO_INFO1_IRQ);
+		host->sdio_mask = USDHI6_SDIO_INFO1_IRQ;
+	}
+}
+
+static int usdhi6_set_pinstates(struct usdhi6_host *host, int voltage)
+{
+	if (IS_ERR(host->pins_uhs))
+		return 0;
+
+	switch (voltage) {
+	case MMC_SIGNAL_VOLTAGE_180:
+	case MMC_SIGNAL_VOLTAGE_120:
+		return pinctrl_select_state(host->pinctrl,
+					    host->pins_uhs);
+
+	default:
+		return pinctrl_select_default_state(mmc_dev(host->mmc));
+	}
+}
+
+static int usdhi6_sig_volt_switch(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	int ret;
+
+	ret = mmc_regulator_set_vqmmc(mmc, ios);
+	if (ret < 0)
+		return ret;
+
+	ret = usdhi6_set_pinstates(mmc_priv(mmc), ios->signal_voltage);
+	if (ret)
+		dev_warn_once(mmc_dev(mmc),
+			      "Failed to set pinstate err=%d\n", ret);
+	return ret;
+}
+
+static const struct mmc_host_ops usdhi6_ops = {
+	.request	= usdhi6_request,
+	.set_ios	= usdhi6_set_ios,
+	.get_cd		= usdhi6_get_cd,
+	.get_ro		= usdhi6_get_ro,
+	.enable_sdio_irq = usdhi6_enable_sdio_irq,
+	.start_signal_voltage_switch = usdhi6_sig_volt_switch,
+};
+
+/*			State machine handlers				*/
+
+static void usdhi6_resp_cmd12(struct usdhi6_host *host)
+{
+	struct mmc_command *cmd = host->mrq->stop;
+	cmd->resp[0] = usdhi6_read(host, USDHI6_SD_RSP10);
+}
+
+static void usdhi6_resp_read(struct usdhi6_host *host)
+{
+	struct mmc_command *cmd = host->mrq->cmd;
+	u32 *rsp = cmd->resp, tmp = 0;
+	int i;
+
+/*
+ * RSP10	39-8
+ * RSP32	71-40
+ * RSP54	103-72
+ * RSP76	127-104
+ * R2-type response:
+ * resp[0]	= r[127..96]
+ * resp[1]	= r[95..64]
+ * resp[2]	= r[63..32]
+ * resp[3]	= r[31..0]
+ * Other responses:
+ * resp[0]	= r[39..8]
+ */
+
+	if (mmc_resp_type(cmd) == MMC_RSP_NONE)
+		return;
+
+	if (!(host->irq_status & USDHI6_SD_INFO1_RSP_END)) {
+		dev_err(mmc_dev(host->mmc),
+			"CMD%d: response expected but is missing!\n", cmd->opcode);
+		return;
+	}
+
+	if (mmc_resp_type(cmd) & MMC_RSP_136)
+		for (i = 0; i < 4; i++) {
+			if (i)
+				rsp[3 - i] = tmp >> 24;
+			tmp = usdhi6_read(host, USDHI6_SD_RSP10 + i * 8);
+			rsp[3 - i] |= tmp << 8;
+		}
+	else if (cmd->opcode == MMC_READ_MULTIPLE_BLOCK ||
+		 cmd->opcode == MMC_WRITE_MULTIPLE_BLOCK)
+		/* Read RSP54 to avoid conflict with auto CMD12 */
+		rsp[0] = usdhi6_read(host, USDHI6_SD_RSP54);
+	else
+		rsp[0] = usdhi6_read(host, USDHI6_SD_RSP10);
+
+	dev_dbg(mmc_dev(host->mmc), "Response 0x%x\n", rsp[0]);
+}
+
+static int usdhi6_blk_read(struct usdhi6_host *host)
+{
+	struct mmc_data *data = host->mrq->data;
+	u32 *p;
+	int i, rest;
+
+	if (host->io_error) {
+		data->error = usdhi6_error_code(host);
+		goto error;
+	}
+
+	if (host->pg.page) {
+		p = host->blk_page + host->offset;
+	} else {
+		p = usdhi6_sg_map(host);
+		if (!p) {
+			data->error = -ENOMEM;
+			goto error;
+		}
+	}
+
+	for (i = 0; i < data->blksz / 4; i++, p++)
+		*p = usdhi6_read(host, USDHI6_SD_BUF0);
+
+	rest = data->blksz % 4;
+	for (i = 0; i < (rest + 1) / 2; i++) {
+		u16 d = usdhi6_read16(host, USDHI6_SD_BUF0);
+		((u8 *)p)[2 * i] = ((u8 *)&d)[0];
+		if (rest > 1 && !i)
+			((u8 *)p)[2 * i + 1] = ((u8 *)&d)[1];
+	}
+
+	return 0;
+
+error:
+	dev_dbg(mmc_dev(host->mmc), "%s(): %d\n", __func__, data->error);
+	host->wait = USDHI6_WAIT_FOR_REQUEST;
+	return data->error;
+}
+
+static int usdhi6_blk_write(struct usdhi6_host *host)
+{
+	struct mmc_data *data = host->mrq->data;
+	u32 *p;
+	int i, rest;
+
+	if (host->io_error) {
+		data->error = usdhi6_error_code(host);
+		goto error;
+	}
+
+	if (host->pg.page) {
+		p = host->blk_page + host->offset;
+	} else {
+		p = usdhi6_sg_map(host);
+		if (!p) {
+			data->error = -ENOMEM;
+			goto error;
+		}
+	}
+
+	for (i = 0; i < data->blksz / 4; i++, p++)
+		usdhi6_write(host, USDHI6_SD_BUF0, *p);
+
+	rest = data->blksz % 4;
+	for (i = 0; i < (rest + 1) / 2; i++) {
+		u16 d;
+		((u8 *)&d)[0] = ((u8 *)p)[2 * i];
+		if (rest > 1 && !i)
+			((u8 *)&d)[1] = ((u8 *)p)[2 * i + 1];
+		else
+			((u8 *)&d)[1] = 0;
+		usdhi6_write16(host, USDHI6_SD_BUF0, d);
+	}
+
+	return 0;
+
+error:
+	dev_dbg(mmc_dev(host->mmc), "%s(): %d\n", __func__, data->error);
+	host->wait = USDHI6_WAIT_FOR_REQUEST;
+	return data->error;
+}
+
+static int usdhi6_stop_cmd(struct usdhi6_host *host)
+{
+	struct mmc_request *mrq = host->mrq;
+
+	switch (mrq->cmd->opcode) {
+	case MMC_READ_MULTIPLE_BLOCK:
+	case MMC_WRITE_MULTIPLE_BLOCK:
+		if (mrq->stop->opcode == MMC_STOP_TRANSMISSION) {
+			host->wait = USDHI6_WAIT_FOR_STOP;
+			return 0;
+		}
+		fallthrough;	/* Unsupported STOP command */
+	default:
+		dev_err(mmc_dev(host->mmc),
+			"unsupported stop CMD%d for CMD%d\n",
+			mrq->stop->opcode, mrq->cmd->opcode);
+		mrq->stop->error = -EOPNOTSUPP;
+	}
+
+	return -EOPNOTSUPP;
+}
+
+static bool usdhi6_end_cmd(struct usdhi6_host *host)
+{
+	struct mmc_request *mrq = host->mrq;
+	struct mmc_command *cmd = mrq->cmd;
+
+	if (host->io_error) {
+		cmd->error = usdhi6_error_code(host);
+		return false;
+	}
+
+	usdhi6_resp_read(host);
+
+	if (!mrq->data)
+		return false;
+
+	if (host->dma_active) {
+		usdhi6_dma_kick(host);
+		if (!mrq->stop)
+			host->wait = USDHI6_WAIT_FOR_DMA;
+		else if (usdhi6_stop_cmd(host) < 0)
+			return false;
+	} else if (mrq->data->flags & MMC_DATA_READ) {
+		if (cmd->opcode == MMC_READ_MULTIPLE_BLOCK ||
+		    (cmd->opcode == SD_IO_RW_EXTENDED &&
+		     mrq->data->blocks > 1))
+			host->wait = USDHI6_WAIT_FOR_MREAD;
+		else
+			host->wait = USDHI6_WAIT_FOR_READ;
+	} else {
+		if (cmd->opcode == MMC_WRITE_MULTIPLE_BLOCK ||
+		    (cmd->opcode == SD_IO_RW_EXTENDED &&
+		     mrq->data->blocks > 1))
+			host->wait = USDHI6_WAIT_FOR_MWRITE;
+		else
+			host->wait = USDHI6_WAIT_FOR_WRITE;
+	}
+
+	return true;
+}
+
+static bool usdhi6_read_block(struct usdhi6_host *host)
+{
+	/* ACCESS_END IRQ is already unmasked */
+	int ret = usdhi6_blk_read(host);
+
+	/*
+	 * Have to force unmapping both pages: the single block could have been
+	 * cross-page, in which case for single-block IO host->page_idx == 0.
+	 * So, if we don't force, the second page won't be unmapped.
+	 */
+	usdhi6_sg_unmap(host, true);
+
+	if (ret < 0)
+		return false;
+
+	host->wait = USDHI6_WAIT_FOR_DATA_END;
+	return true;
+}
+
+static bool usdhi6_mread_block(struct usdhi6_host *host)
+{
+	int ret = usdhi6_blk_read(host);
+
+	if (ret < 0)
+		return false;
+
+	usdhi6_sg_advance(host);
+
+	return !host->mrq->data->error &&
+		(host->wait != USDHI6_WAIT_FOR_DATA_END || !host->mrq->stop);
+}
+
+static bool usdhi6_write_block(struct usdhi6_host *host)
+{
+	int ret = usdhi6_blk_write(host);
+
+	/* See comment in usdhi6_read_block() */
+	usdhi6_sg_unmap(host, true);
+
+	if (ret < 0)
+		return false;
+
+	host->wait = USDHI6_WAIT_FOR_DATA_END;
+	return true;
+}
+
+static bool usdhi6_mwrite_block(struct usdhi6_host *host)
+{
+	int ret = usdhi6_blk_write(host);
+
+	if (ret < 0)
+		return false;
+
+	usdhi6_sg_advance(host);
+
+	return !host->mrq->data->error &&
+		(host->wait != USDHI6_WAIT_FOR_DATA_END || !host->mrq->stop);
+}
+
+/*			Interrupt & timeout handlers			*/
+
+static irqreturn_t usdhi6_sd_bh(int irq, void *dev_id)
+{
+	struct usdhi6_host *host = dev_id;
+	struct mmc_request *mrq;
+	struct mmc_command *cmd;
+	struct mmc_data *data;
+	bool io_wait = false;
+
+	cancel_delayed_work_sync(&host->timeout_work);
+
+	mrq = host->mrq;
+	if (!mrq)
+		return IRQ_HANDLED;
+
+	cmd = mrq->cmd;
+	data = mrq->data;
+
+	switch (host->wait) {
+	case USDHI6_WAIT_FOR_REQUEST:
+		/* We're too late, the timeout has already kicked in */
+		return IRQ_HANDLED;
+	case USDHI6_WAIT_FOR_CMD:
+		/* Wait for data? */
+		io_wait = usdhi6_end_cmd(host);
+		break;
+	case USDHI6_WAIT_FOR_MREAD:
+		/* Wait for more data? */
+		io_wait = usdhi6_mread_block(host);
+		break;
+	case USDHI6_WAIT_FOR_READ:
+		/* Wait for data end? */
+		io_wait = usdhi6_read_block(host);
+		break;
+	case USDHI6_WAIT_FOR_MWRITE:
+		/* Wait data to write? */
+		io_wait = usdhi6_mwrite_block(host);
+		break;
+	case USDHI6_WAIT_FOR_WRITE:
+		/* Wait for data end? */
+		io_wait = usdhi6_write_block(host);
+		break;
+	case USDHI6_WAIT_FOR_DMA:
+		usdhi6_dma_check_error(host);
+		break;
+	case USDHI6_WAIT_FOR_STOP:
+		usdhi6_write(host, USDHI6_SD_STOP, 0);
+		if (host->io_error) {
+			int ret = usdhi6_error_code(host);
+			if (mrq->stop)
+				mrq->stop->error = ret;
+			else
+				mrq->data->error = ret;
+			dev_warn(mmc_dev(host->mmc), "%s(): %d\n", __func__, ret);
+			break;
+		}
+		usdhi6_resp_cmd12(host);
+		mrq->stop->error = 0;
+		break;
+	case USDHI6_WAIT_FOR_DATA_END:
+		if (host->io_error) {
+			mrq->data->error = usdhi6_error_code(host);
+			dev_warn(mmc_dev(host->mmc), "%s(): %d\n", __func__,
+				 mrq->data->error);
+		}
+		break;
+	default:
+		cmd->error = -EFAULT;
+		dev_err(mmc_dev(host->mmc), "Invalid state %u\n", host->wait);
+		usdhi6_request_done(host);
+		return IRQ_HANDLED;
+	}
+
+	if (io_wait) {
+		schedule_delayed_work(&host->timeout_work, host->timeout);
+		/* Wait for more data or ACCESS_END */
+		if (!host->dma_active)
+			usdhi6_wait_for_brwe(host, mrq->data->flags & MMC_DATA_READ);
+		return IRQ_HANDLED;
+	}
+
+	if (!cmd->error) {
+		if (data) {
+			if (!data->error) {
+				if (host->wait != USDHI6_WAIT_FOR_STOP &&
+				    host->mrq->stop &&
+				    !host->mrq->stop->error &&
+				    !usdhi6_stop_cmd(host)) {
+					/* Sending STOP */
+					usdhi6_wait_for_resp(host);
+
+					schedule_delayed_work(&host->timeout_work,
+							      host->timeout);
+
+					return IRQ_HANDLED;
+				}
+
+				data->bytes_xfered = data->blocks * data->blksz;
+			} else {
+				/* Data error: might need to unmap the last page */
+				dev_warn(mmc_dev(host->mmc), "%s(): data error %d\n",
+					 __func__, data->error);
+				usdhi6_sg_unmap(host, true);
+			}
+		} else if (cmd->opcode == MMC_APP_CMD) {
+			host->app_cmd = true;
+		}
+	}
+
+	usdhi6_request_done(host);
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t usdhi6_sd(int irq, void *dev_id)
+{
+	struct usdhi6_host *host = dev_id;
+	u16 status, status2, error;
+
+	status = usdhi6_read(host, USDHI6_SD_INFO1) & ~host->status_mask &
+		~USDHI6_SD_INFO1_CARD;
+	status2 = usdhi6_read(host, USDHI6_SD_INFO2) & ~host->status2_mask;
+
+	usdhi6_only_cd(host);
+
+	dev_dbg(mmc_dev(host->mmc),
+		"IRQ status = 0x%08x, status2 = 0x%08x\n", status, status2);
+
+	if (!status && !status2)
+		return IRQ_NONE;
+
+	error = status2 & USDHI6_SD_INFO2_ERR;
+
+	/* Ack / clear interrupts */
+	if (USDHI6_SD_INFO1_IRQ & status)
+		usdhi6_write(host, USDHI6_SD_INFO1,
+			     0xffff & ~(USDHI6_SD_INFO1_IRQ & status));
+
+	if (USDHI6_SD_INFO2_IRQ & status2) {
+		if (error)
+			/* In error cases BWE and BRE aren't cleared automatically */
+			status2 |= USDHI6_SD_INFO2_BWE | USDHI6_SD_INFO2_BRE;
+
+		usdhi6_write(host, USDHI6_SD_INFO2,
+			     0xffff & ~(USDHI6_SD_INFO2_IRQ & status2));
+	}
+
+	host->io_error = error;
+	host->irq_status = status;
+
+	if (error) {
+		/* Don't pollute the log with unsupported command timeouts */
+		if (host->wait != USDHI6_WAIT_FOR_CMD ||
+		    error != USDHI6_SD_INFO2_RSP_TOUT)
+			dev_warn(mmc_dev(host->mmc),
+				 "%s(): INFO2 error bits 0x%08x\n",
+				 __func__, error);
+		else
+			dev_dbg(mmc_dev(host->mmc),
+				"%s(): INFO2 error bits 0x%08x\n",
+				__func__, error);
+	}
+
+	return IRQ_WAKE_THREAD;
+}
+
+static irqreturn_t usdhi6_sdio(int irq, void *dev_id)
+{
+	struct usdhi6_host *host = dev_id;
+	u32 status = usdhi6_read(host, USDHI6_SDIO_INFO1) & ~host->sdio_mask;
+
+	dev_dbg(mmc_dev(host->mmc), "%s(): status 0x%x\n", __func__, status);
+
+	if (!status)
+		return IRQ_NONE;
+
+	usdhi6_write(host, USDHI6_SDIO_INFO1, ~status);
+
+	mmc_signal_sdio_irq(host->mmc);
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t usdhi6_cd(int irq, void *dev_id)
+{
+	struct usdhi6_host *host = dev_id;
+	struct mmc_host *mmc = host->mmc;
+	u16 status;
+
+	/* We're only interested in hotplug events here */
+	status = usdhi6_read(host, USDHI6_SD_INFO1) & ~host->status_mask &
+		USDHI6_SD_INFO1_CARD;
+
+	if (!status)
+		return IRQ_NONE;
+
+	/* Ack */
+	usdhi6_write(host, USDHI6_SD_INFO1, ~status);
+
+	if (!work_pending(&mmc->detect.work) &&
+	    (((status & USDHI6_SD_INFO1_CARD_INSERT) &&
+	      !mmc->card) ||
+	     ((status & USDHI6_SD_INFO1_CARD_EJECT) &&
+	      mmc->card)))
+		mmc_detect_change(mmc, msecs_to_jiffies(100));
+
+	return IRQ_HANDLED;
+}
+
+/*
+ * Actually this should not be needed, if the built-in timeout works reliably in
+ * the both PIO cases and DMA never fails. But if DMA does fail, a timeout
+ * handler might be the only way to catch the error.
+ */
+static void usdhi6_timeout_work(struct work_struct *work)
+{
+	struct delayed_work *d = to_delayed_work(work);
+	struct usdhi6_host *host = container_of(d, struct usdhi6_host, timeout_work);
+	struct mmc_request *mrq = host->mrq;
+	struct mmc_data *data = mrq ? mrq->data : NULL;
+	struct scatterlist *sg;
+
+	dev_warn(mmc_dev(host->mmc),
+		 "%s timeout wait %u CMD%d: IRQ 0x%08x:0x%08x, last IRQ 0x%08x\n",
+		 host->dma_active ? "DMA" : "PIO",
+		 host->wait, mrq ? mrq->cmd->opcode : -1,
+		 usdhi6_read(host, USDHI6_SD_INFO1),
+		 usdhi6_read(host, USDHI6_SD_INFO2), host->irq_status);
+
+	if (host->dma_active) {
+		usdhi6_dma_kill(host);
+		usdhi6_dma_stop_unmap(host);
+	}
+
+	switch (host->wait) {
+	default:
+		dev_err(mmc_dev(host->mmc), "Invalid state %u\n", host->wait);
+		fallthrough;	/* mrq can be NULL, but is impossible */
+	case USDHI6_WAIT_FOR_CMD:
+		usdhi6_error_code(host);
+		if (mrq)
+			mrq->cmd->error = -ETIMEDOUT;
+		break;
+	case USDHI6_WAIT_FOR_STOP:
+		usdhi6_error_code(host);
+		mrq->stop->error = -ETIMEDOUT;
+		break;
+	case USDHI6_WAIT_FOR_DMA:
+	case USDHI6_WAIT_FOR_MREAD:
+	case USDHI6_WAIT_FOR_MWRITE:
+	case USDHI6_WAIT_FOR_READ:
+	case USDHI6_WAIT_FOR_WRITE:
+		sg = host->sg ?: data->sg;
+		dev_dbg(mmc_dev(host->mmc),
+			"%c: page #%u @ +0x%zx %ux%u in SG%u. Current SG %u bytes @ %u\n",
+			data->flags & MMC_DATA_READ ? 'R' : 'W', host->page_idx,
+			host->offset, data->blocks, data->blksz, data->sg_len,
+			sg_dma_len(sg), sg->offset);
+		usdhi6_sg_unmap(host, true);
+		fallthrough;	/* page unmapped in USDHI6_WAIT_FOR_DATA_END */
+	case USDHI6_WAIT_FOR_DATA_END:
+		usdhi6_error_code(host);
+		data->error = -ETIMEDOUT;
+	}
+
+	if (mrq)
+		usdhi6_request_done(host);
+}
+
+/*			 Probe / release				*/
+
+static const struct of_device_id usdhi6_of_match[] = {
+	{.compatible = "renesas,usdhi6rol0"},
+	{}
+};
+MODULE_DEVICE_TABLE(of, usdhi6_of_match);
+
+static int usdhi6_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct mmc_host *mmc;
+	struct usdhi6_host *host;
+	struct resource *res;
+	int irq_cd, irq_sd, irq_sdio;
+	u32 version;
+	int ret;
+
+	if (!dev->of_node)
+		return -ENODEV;
+
+	irq_cd = platform_get_irq_byname(pdev, "card detect");
+	irq_sd = platform_get_irq_byname(pdev, "data");
+	irq_sdio = platform_get_irq_byname(pdev, "SDIO");
+	if (irq_sd < 0)
+		return irq_sd;
+	if (irq_sdio < 0)
+		return irq_sdio;
+
+	mmc = mmc_alloc_host(sizeof(struct usdhi6_host), dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret)
+		goto e_free_mmc;
+
+	ret = mmc_of_parse(mmc);
+	if (ret < 0)
+		goto e_free_mmc;
+
+	host		= mmc_priv(mmc);
+	host->mmc	= mmc;
+	host->wait	= USDHI6_WAIT_FOR_REQUEST;
+	host->timeout	= msecs_to_jiffies(USDHI6_REQ_TIMEOUT_MS);
+	/*
+	 * We use a fixed timeout of 4s, hence inform the core about it. A
+	 * future improvement should instead respect the cmd->busy_timeout.
+	 */
+	mmc->max_busy_timeout = USDHI6_REQ_TIMEOUT_MS;
+
+	host->pinctrl = devm_pinctrl_get(&pdev->dev);
+	if (IS_ERR(host->pinctrl)) {
+		ret = PTR_ERR(host->pinctrl);
+		goto e_free_mmc;
+	}
+
+	host->pins_uhs = pinctrl_lookup_state(host->pinctrl, "state_uhs");
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	host->base = devm_ioremap_resource(dev, res);
+	if (IS_ERR(host->base)) {
+		ret = PTR_ERR(host->base);
+		goto e_free_mmc;
+	}
+
+	host->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(host->clk)) {
+		ret = PTR_ERR(host->clk);
+		goto e_free_mmc;
+	}
+
+	host->imclk = clk_get_rate(host->clk);
+
+	ret = clk_prepare_enable(host->clk);
+	if (ret < 0)
+		goto e_free_mmc;
+
+	version = usdhi6_read(host, USDHI6_VERSION);
+	if ((version & 0xfff) != 0xa0d) {
+		ret = -EPERM;
+		dev_err(dev, "Version not recognized %x\n", version);
+		goto e_clk_off;
+	}
+
+	dev_info(dev, "A USDHI6ROL0 SD host detected with %d ports\n",
+		 usdhi6_read(host, USDHI6_SD_PORT_SEL) >> USDHI6_SD_PORT_SEL_PORTS_SHIFT);
+
+	usdhi6_mask_all(host);
+
+	if (irq_cd >= 0) {
+		ret = devm_request_irq(dev, irq_cd, usdhi6_cd, 0,
+				       dev_name(dev), host);
+		if (ret < 0)
+			goto e_clk_off;
+	} else {
+		mmc->caps |= MMC_CAP_NEEDS_POLL;
+	}
+
+	ret = devm_request_threaded_irq(dev, irq_sd, usdhi6_sd, usdhi6_sd_bh, 0,
+			       dev_name(dev), host);
+	if (ret < 0)
+		goto e_clk_off;
+
+	ret = devm_request_irq(dev, irq_sdio, usdhi6_sdio, 0,
+			       dev_name(dev), host);
+	if (ret < 0)
+		goto e_clk_off;
+
+	INIT_DELAYED_WORK(&host->timeout_work, usdhi6_timeout_work);
+
+	usdhi6_dma_request(host, res->start);
+
+	mmc->ops = &usdhi6_ops;
+	mmc->caps |= MMC_CAP_SD_HIGHSPEED | MMC_CAP_MMC_HIGHSPEED |
+		     MMC_CAP_SDIO_IRQ;
+	/* Set .max_segs to some random number. Feel free to adjust. */
+	mmc->max_segs = 32;
+	mmc->max_blk_size = 512;
+	mmc->max_req_size = PAGE_SIZE * mmc->max_segs;
+	mmc->max_blk_count = mmc->max_req_size / mmc->max_blk_size;
+	/*
+	 * Setting .max_seg_size to 1 page would simplify our page-mapping code,
+	 * But OTOH, having large segments makes DMA more efficient. We could
+	 * check, whether we managed to get DMA and fall back to 1 page
+	 * segments, but if we do manage to obtain DMA and then it fails at
+	 * run-time and we fall back to PIO, we will continue getting large
+	 * segments. So, we wouldn't be able to get rid of the code anyway.
+	 */
+	mmc->max_seg_size = mmc->max_req_size;
+	if (!mmc->f_max)
+		mmc->f_max = host->imclk;
+	mmc->f_min = host->imclk / 512;
+
+	platform_set_drvdata(pdev, host);
+
+	ret = mmc_add_host(mmc);
+	if (ret < 0)
+		goto e_release_dma;
+
+	return 0;
+
+e_release_dma:
+	usdhi6_dma_release(host);
+e_clk_off:
+	clk_disable_unprepare(host->clk);
+e_free_mmc:
+	mmc_free_host(mmc);
+
+	return ret;
+}
+
+static int usdhi6_remove(struct platform_device *pdev)
+{
+	struct usdhi6_host *host = platform_get_drvdata(pdev);
+
+	mmc_remove_host(host->mmc);
+
+	usdhi6_mask_all(host);
+	cancel_delayed_work_sync(&host->timeout_work);
+	usdhi6_dma_release(host);
+	clk_disable_unprepare(host->clk);
+	mmc_free_host(host->mmc);
+
+	return 0;
+}
+
+static struct platform_driver usdhi6_driver = {
+	.probe		= usdhi6_probe,
+	.remove		= usdhi6_remove,
+	.driver		= {
+		.name	= "usdhi6rol0",
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.of_match_table = usdhi6_of_match,
+	},
+};
+
+module_platform_driver(usdhi6_driver);
+
+MODULE_DESCRIPTION("Renesas usdhi6rol0 SD/SDIO host driver");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:usdhi6rol0");
+MODULE_AUTHOR("Guennadi Liakhovetski <g.liakhovetski@gmx.de>");
diff --git a/drivers/mmc/host/ushc.c b/drivers/mmc/host/ushc.c
new file mode 100644
index 000000000..9a6358fd9
--- /dev/null
+++ b/drivers/mmc/host/ushc.c
@@ -0,0 +1,568 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * USB SD Host Controller (USHC) controller driver.
+ *
+ * Copyright (C) 2010 Cambridge Silicon Radio Ltd.
+ *
+ * Notes:
+ *   - Only version 2 devices are supported.
+ *   - Version 2 devices only support SDIO cards/devices (R2 response is
+ *     unsupported).
+ *
+ * References:
+ *   [USHC] USB SD Host Controller specification (CS-118793-SP)
+ */
+#include <linux/module.h>
+#include <linux/usb.h>
+#include <linux/kernel.h>
+#include <linux/slab.h>
+#include <linux/dma-mapping.h>
+#include <linux/mmc/host.h>
+
+enum ushc_request {
+	USHC_GET_CAPS  = 0x00,
+	USHC_HOST_CTRL = 0x01,
+	USHC_PWR_CTRL  = 0x02,
+	USHC_CLK_FREQ  = 0x03,
+	USHC_EXEC_CMD  = 0x04,
+	USHC_READ_RESP = 0x05,
+	USHC_RESET     = 0x06,
+};
+
+enum ushc_request_type {
+	USHC_GET_CAPS_TYPE  = USB_DIR_IN  | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+	USHC_HOST_CTRL_TYPE = USB_DIR_OUT | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+	USHC_PWR_CTRL_TYPE  = USB_DIR_OUT | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+	USHC_CLK_FREQ_TYPE  = USB_DIR_OUT | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+	USHC_EXEC_CMD_TYPE  = USB_DIR_OUT | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+	USHC_READ_RESP_TYPE = USB_DIR_IN  | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+	USHC_RESET_TYPE     = USB_DIR_OUT | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+};
+
+#define USHC_GET_CAPS_VERSION_MASK 0xff
+#define USHC_GET_CAPS_3V3      (1 << 8)
+#define USHC_GET_CAPS_3V0      (1 << 9)
+#define USHC_GET_CAPS_1V8      (1 << 10)
+#define USHC_GET_CAPS_HIGH_SPD (1 << 16)
+
+#define USHC_HOST_CTRL_4BIT     (1 << 1)
+#define USHC_HOST_CTRL_HIGH_SPD (1 << 0)
+
+#define USHC_PWR_CTRL_OFF 0x00
+#define USHC_PWR_CTRL_3V3 0x01
+#define USHC_PWR_CTRL_3V0 0x02
+#define USHC_PWR_CTRL_1V8 0x03
+
+#define USHC_READ_RESP_BUSY        (1 << 4)
+#define USHC_READ_RESP_ERR_TIMEOUT (1 << 3)
+#define USHC_READ_RESP_ERR_CRC     (1 << 2)
+#define USHC_READ_RESP_ERR_DAT     (1 << 1)
+#define USHC_READ_RESP_ERR_CMD     (1 << 0)
+#define USHC_READ_RESP_ERR_MASK    0x0f
+
+struct ushc_cbw {
+	__u8 signature;
+	__u8 cmd_idx;
+	__le16 block_size;
+	__le32 arg;
+} __attribute__((packed));
+
+#define USHC_CBW_SIGNATURE 'C'
+
+struct ushc_csw {
+	__u8 signature;
+	__u8 status;
+	__le32 response;
+} __attribute__((packed));
+
+#define USHC_CSW_SIGNATURE 'S'
+
+struct ushc_int_data {
+	u8 status;
+	u8 reserved[3];
+};
+
+#define USHC_INT_STATUS_SDIO_INT     (1 << 1)
+#define USHC_INT_STATUS_CARD_PRESENT (1 << 0)
+
+
+struct ushc_data {
+	struct usb_device *usb_dev;
+	struct mmc_host *mmc;
+
+	struct urb *int_urb;
+	struct ushc_int_data *int_data;
+
+	struct urb *cbw_urb;
+	struct ushc_cbw *cbw;
+
+	struct urb *data_urb;
+
+	struct urb *csw_urb;
+	struct ushc_csw *csw;
+
+	spinlock_t lock;
+	struct mmc_request *current_req;
+	u32 caps;
+	u16 host_ctrl;
+	unsigned long flags;
+	u8 last_status;
+	int clock_freq;
+};
+
+#define DISCONNECTED    0
+#define INT_EN          1
+#define IGNORE_NEXT_INT 2
+
+static void data_callback(struct urb *urb);
+
+static int ushc_hw_reset(struct ushc_data *ushc)
+{
+	return usb_control_msg(ushc->usb_dev, usb_sndctrlpipe(ushc->usb_dev, 0),
+			       USHC_RESET, USHC_RESET_TYPE,
+			       0, 0, NULL, 0, 100);
+}
+
+static int ushc_hw_get_caps(struct ushc_data *ushc)
+{
+	int ret;
+	int version;
+
+	ret = usb_control_msg(ushc->usb_dev, usb_rcvctrlpipe(ushc->usb_dev, 0),
+			      USHC_GET_CAPS, USHC_GET_CAPS_TYPE,
+			      0, 0, &ushc->caps, sizeof(ushc->caps), 100);
+	if (ret < 0)
+		return ret;
+
+	ushc->caps = le32_to_cpu(ushc->caps);
+
+	version = ushc->caps & USHC_GET_CAPS_VERSION_MASK;
+	if (version != 0x02) {
+		dev_err(&ushc->usb_dev->dev, "controller version %d is not supported\n", version);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int ushc_hw_set_host_ctrl(struct ushc_data *ushc, u16 mask, u16 val)
+{
+	u16 host_ctrl;
+	int ret;
+
+	host_ctrl = (ushc->host_ctrl & ~mask) | val;
+	ret = usb_control_msg(ushc->usb_dev, usb_sndctrlpipe(ushc->usb_dev, 0),
+			      USHC_HOST_CTRL, USHC_HOST_CTRL_TYPE,
+			      host_ctrl, 0, NULL, 0, 100);
+	if (ret < 0)
+		return ret;
+	ushc->host_ctrl = host_ctrl;
+	return 0;
+}
+
+static void int_callback(struct urb *urb)
+{
+	struct ushc_data *ushc = urb->context;
+	u8 status, last_status;
+
+	if (urb->status < 0)
+		return;
+
+	status = ushc->int_data->status;
+	last_status = ushc->last_status;
+	ushc->last_status = status;
+
+	/*
+	 * Ignore the card interrupt status on interrupt transfers that
+	 * were submitted while card interrupts where disabled.
+	 *
+	 * This avoid occasional spurious interrupts when enabling
+	 * interrupts immediately after clearing the source on the card.
+	 */
+
+	if (!test_and_clear_bit(IGNORE_NEXT_INT, &ushc->flags)
+	    && test_bit(INT_EN, &ushc->flags)
+	    && status & USHC_INT_STATUS_SDIO_INT) {
+		mmc_signal_sdio_irq(ushc->mmc);
+	}
+
+	if ((status ^ last_status) & USHC_INT_STATUS_CARD_PRESENT)
+		mmc_detect_change(ushc->mmc, msecs_to_jiffies(100));
+
+	if (!test_bit(INT_EN, &ushc->flags))
+		set_bit(IGNORE_NEXT_INT, &ushc->flags);
+	usb_submit_urb(ushc->int_urb, GFP_ATOMIC);
+}
+
+static void cbw_callback(struct urb *urb)
+{
+	struct ushc_data *ushc = urb->context;
+
+	if (urb->status != 0) {
+		usb_unlink_urb(ushc->data_urb);
+		usb_unlink_urb(ushc->csw_urb);
+	}
+}
+
+static void data_callback(struct urb *urb)
+{
+	struct ushc_data *ushc = urb->context;
+
+	if (urb->status != 0)
+		usb_unlink_urb(ushc->csw_urb);
+}
+
+static void csw_callback(struct urb *urb)
+{
+	struct ushc_data *ushc = urb->context;
+	struct mmc_request *req = ushc->current_req;
+	int status;
+
+	status = ushc->csw->status;
+
+	if (urb->status != 0) {
+		req->cmd->error = urb->status;
+	} else if (status & USHC_READ_RESP_ERR_CMD) {
+		if (status & USHC_READ_RESP_ERR_CRC)
+			req->cmd->error = -EIO;
+		else
+			req->cmd->error = -ETIMEDOUT;
+	}
+	if (req->data) {
+		if (status & USHC_READ_RESP_ERR_DAT) {
+			if (status & USHC_READ_RESP_ERR_CRC)
+				req->data->error = -EIO;
+			else
+				req->data->error = -ETIMEDOUT;
+			req->data->bytes_xfered = 0;
+		} else {
+			req->data->bytes_xfered = req->data->blksz * req->data->blocks;
+		}
+	}
+
+	req->cmd->resp[0] = le32_to_cpu(ushc->csw->response);
+
+	mmc_request_done(ushc->mmc, req);
+}
+
+static void ushc_request(struct mmc_host *mmc, struct mmc_request *req)
+{
+	struct ushc_data *ushc = mmc_priv(mmc);
+	int ret;
+	unsigned long flags;
+
+	spin_lock_irqsave(&ushc->lock, flags);
+
+	if (test_bit(DISCONNECTED, &ushc->flags)) {
+		ret = -ENODEV;
+		goto out;
+	}
+
+	/* Version 2 firmware doesn't support the R2 response format. */
+	if (req->cmd->flags & MMC_RSP_136) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	/* The Astoria's data FIFOs don't work with clock speeds < 5MHz so
+	   limit commands with data to 6MHz or more. */
+	if (req->data && ushc->clock_freq < 6000000) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	ushc->current_req = req;
+
+	/* Start cmd with CBW. */
+	ushc->cbw->cmd_idx = cpu_to_le16(req->cmd->opcode);
+	if (req->data)
+		ushc->cbw->block_size = cpu_to_le16(req->data->blksz);
+	else
+		ushc->cbw->block_size = 0;
+	ushc->cbw->arg = cpu_to_le32(req->cmd->arg);
+
+	ret = usb_submit_urb(ushc->cbw_urb, GFP_ATOMIC);
+	if (ret < 0)
+		goto out;
+
+	/* Submit data (if any). */
+	if (req->data) {
+		struct mmc_data *data = req->data;
+		int pipe;
+
+		if (data->flags & MMC_DATA_READ)
+			pipe = usb_rcvbulkpipe(ushc->usb_dev, 6);
+		else
+			pipe = usb_sndbulkpipe(ushc->usb_dev, 2);
+
+		usb_fill_bulk_urb(ushc->data_urb, ushc->usb_dev, pipe,
+				  NULL, data->sg->length,
+				  data_callback, ushc);
+		ushc->data_urb->num_sgs = 1;
+		ushc->data_urb->sg = data->sg;
+		ret = usb_submit_urb(ushc->data_urb, GFP_ATOMIC);
+		if (ret < 0)
+			goto out;
+	}
+
+	/* Submit CSW. */
+	ret = usb_submit_urb(ushc->csw_urb, GFP_ATOMIC);
+
+out:
+	spin_unlock_irqrestore(&ushc->lock, flags);
+	if (ret < 0) {
+		usb_unlink_urb(ushc->cbw_urb);
+		usb_unlink_urb(ushc->data_urb);
+		req->cmd->error = ret;
+		mmc_request_done(mmc, req);
+	}
+}
+
+static int ushc_set_power(struct ushc_data *ushc, unsigned char power_mode)
+{
+	u16 voltage;
+
+	switch (power_mode) {
+	case MMC_POWER_OFF:
+		voltage = USHC_PWR_CTRL_OFF;
+		break;
+	case MMC_POWER_UP:
+	case MMC_POWER_ON:
+		voltage = USHC_PWR_CTRL_3V3;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return usb_control_msg(ushc->usb_dev, usb_sndctrlpipe(ushc->usb_dev, 0),
+			       USHC_PWR_CTRL, USHC_PWR_CTRL_TYPE,
+			       voltage, 0, NULL, 0, 100);
+}
+
+static int ushc_set_bus_width(struct ushc_data *ushc, int bus_width)
+{
+	return ushc_hw_set_host_ctrl(ushc, USHC_HOST_CTRL_4BIT,
+				     bus_width == 4 ? USHC_HOST_CTRL_4BIT : 0);
+}
+
+static int ushc_set_bus_freq(struct ushc_data *ushc, int clk, bool enable_hs)
+{
+	int ret;
+
+	/* Hardware can't detect interrupts while the clock is off. */
+	if (clk == 0)
+		clk = 400000;
+
+	ret = ushc_hw_set_host_ctrl(ushc, USHC_HOST_CTRL_HIGH_SPD,
+				    enable_hs ? USHC_HOST_CTRL_HIGH_SPD : 0);
+	if (ret < 0)
+		return ret;
+
+	ret = usb_control_msg(ushc->usb_dev, usb_sndctrlpipe(ushc->usb_dev, 0),
+			      USHC_CLK_FREQ, USHC_CLK_FREQ_TYPE,
+			      clk & 0xffff, (clk >> 16) & 0xffff, NULL, 0, 100);
+	if (ret < 0)
+		return ret;
+
+	ushc->clock_freq = clk;
+	return 0;
+}
+
+static void ushc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct ushc_data *ushc = mmc_priv(mmc);
+
+	ushc_set_power(ushc, ios->power_mode);
+	ushc_set_bus_width(ushc, 1 << ios->bus_width);
+	ushc_set_bus_freq(ushc, ios->clock, ios->timing == MMC_TIMING_SD_HS);
+}
+
+static int ushc_get_cd(struct mmc_host *mmc)
+{
+	struct ushc_data *ushc = mmc_priv(mmc);
+
+	return !!(ushc->last_status & USHC_INT_STATUS_CARD_PRESENT);
+}
+
+static void ushc_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct ushc_data *ushc = mmc_priv(mmc);
+
+	if (enable)
+		set_bit(INT_EN, &ushc->flags);
+	else
+		clear_bit(INT_EN, &ushc->flags);
+}
+
+static void ushc_clean_up(struct ushc_data *ushc)
+{
+	usb_free_urb(ushc->int_urb);
+	usb_free_urb(ushc->csw_urb);
+	usb_free_urb(ushc->data_urb);
+	usb_free_urb(ushc->cbw_urb);
+
+	kfree(ushc->int_data);
+	kfree(ushc->cbw);
+	kfree(ushc->csw);
+
+	mmc_free_host(ushc->mmc);
+}
+
+static const struct mmc_host_ops ushc_ops = {
+	.request         = ushc_request,
+	.set_ios         = ushc_set_ios,
+	.get_cd          = ushc_get_cd,
+	.enable_sdio_irq = ushc_enable_sdio_irq,
+};
+
+static int ushc_probe(struct usb_interface *intf, const struct usb_device_id *id)
+{
+	struct usb_device *usb_dev = interface_to_usbdev(intf);
+	struct mmc_host *mmc;
+	struct ushc_data *ushc;
+	int ret;
+
+	if (intf->cur_altsetting->desc.bNumEndpoints < 1)
+		return -ENODEV;
+
+	mmc = mmc_alloc_host(sizeof(struct ushc_data), &intf->dev);
+	if (mmc == NULL)
+		return -ENOMEM;
+	ushc = mmc_priv(mmc);
+	usb_set_intfdata(intf, ushc);
+
+	ushc->usb_dev = usb_dev;
+	ushc->mmc = mmc;
+
+	spin_lock_init(&ushc->lock);
+
+	ret = ushc_hw_reset(ushc);
+	if (ret < 0)
+		goto err;
+
+	/* Read capabilities. */
+	ret = ushc_hw_get_caps(ushc);
+	if (ret < 0)
+		goto err;
+
+	mmc->ops = &ushc_ops;
+
+	mmc->f_min = 400000;
+	mmc->f_max = 50000000;
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+	mmc->caps = MMC_CAP_4_BIT_DATA | MMC_CAP_SDIO_IRQ;
+	mmc->caps |= (ushc->caps & USHC_GET_CAPS_HIGH_SPD) ? MMC_CAP_SD_HIGHSPEED : 0;
+
+	mmc->max_seg_size  = 512*511;
+	mmc->max_segs      = 1;
+	mmc->max_req_size  = 512*511;
+	mmc->max_blk_size  = 512;
+	mmc->max_blk_count = 511;
+
+	ushc->int_urb = usb_alloc_urb(0, GFP_KERNEL);
+	if (ushc->int_urb == NULL) {
+		ret = -ENOMEM;
+		goto err;
+	}
+	ushc->int_data = kzalloc(sizeof(struct ushc_int_data), GFP_KERNEL);
+	if (ushc->int_data == NULL) {
+		ret = -ENOMEM;
+		goto err;
+	}
+	usb_fill_int_urb(ushc->int_urb, ushc->usb_dev,
+			 usb_rcvintpipe(usb_dev,
+					intf->cur_altsetting->endpoint[0].desc.bEndpointAddress),
+			 ushc->int_data, sizeof(struct ushc_int_data),
+			 int_callback, ushc,
+			 intf->cur_altsetting->endpoint[0].desc.bInterval);
+
+	ushc->cbw_urb = usb_alloc_urb(0, GFP_KERNEL);
+	if (ushc->cbw_urb == NULL) {
+		ret = -ENOMEM;
+		goto err;
+	}
+	ushc->cbw = kzalloc(sizeof(struct ushc_cbw), GFP_KERNEL);
+	if (ushc->cbw == NULL) {
+		ret = -ENOMEM;
+		goto err;
+	}
+	ushc->cbw->signature = USHC_CBW_SIGNATURE;
+
+	usb_fill_bulk_urb(ushc->cbw_urb, ushc->usb_dev, usb_sndbulkpipe(usb_dev, 2),
+			  ushc->cbw, sizeof(struct ushc_cbw),
+			  cbw_callback, ushc);
+
+	ushc->data_urb = usb_alloc_urb(0, GFP_KERNEL);
+	if (ushc->data_urb == NULL) {
+		ret = -ENOMEM;
+		goto err;
+	}
+
+	ushc->csw_urb = usb_alloc_urb(0, GFP_KERNEL);
+	if (ushc->csw_urb == NULL) {
+		ret = -ENOMEM;
+		goto err;
+	}
+	ushc->csw = kzalloc(sizeof(struct ushc_csw), GFP_KERNEL);
+	if (ushc->csw == NULL) {
+		ret = -ENOMEM;
+		goto err;
+	}
+	usb_fill_bulk_urb(ushc->csw_urb, ushc->usb_dev, usb_rcvbulkpipe(usb_dev, 6),
+			  ushc->csw, sizeof(struct ushc_csw),
+			  csw_callback, ushc);
+
+	ret = mmc_add_host(ushc->mmc);
+	if (ret)
+		goto err;
+
+	ret = usb_submit_urb(ushc->int_urb, GFP_KERNEL);
+	if (ret < 0) {
+		mmc_remove_host(ushc->mmc);
+		goto err;
+	}
+
+	return 0;
+
+err:
+	ushc_clean_up(ushc);
+	return ret;
+}
+
+static void ushc_disconnect(struct usb_interface *intf)
+{
+	struct ushc_data *ushc = usb_get_intfdata(intf);
+
+	spin_lock_irq(&ushc->lock);
+	set_bit(DISCONNECTED, &ushc->flags);
+	spin_unlock_irq(&ushc->lock);
+
+	usb_kill_urb(ushc->int_urb);
+	usb_kill_urb(ushc->cbw_urb);
+	usb_kill_urb(ushc->data_urb);
+	usb_kill_urb(ushc->csw_urb);
+
+	mmc_remove_host(ushc->mmc);
+
+	ushc_clean_up(ushc);
+}
+
+static struct usb_device_id ushc_id_table[] = {
+	/* CSR USB SD Host Controller */
+	{ USB_DEVICE(0x0a12, 0x5d10) },
+	{ },
+};
+MODULE_DEVICE_TABLE(usb, ushc_id_table);
+
+static struct usb_driver ushc_driver = {
+	.name       = "ushc",
+	.id_table   = ushc_id_table,
+	.probe      = ushc_probe,
+	.disconnect = ushc_disconnect,
+};
+
+module_usb_driver(ushc_driver);
+
+MODULE_DESCRIPTION("USB SD Host Controller driver");
+MODULE_AUTHOR("David Vrabel <david.vrabel@csr.com>");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mmc/host/via-sdmmc.c b/drivers/mmc/host/via-sdmmc.c
new file mode 100644
index 000000000..f6b525fb5
--- /dev/null
+++ b/drivers/mmc/host/via-sdmmc.c
@@ -0,0 +1,1320 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ *  drivers/mmc/host/via-sdmmc.c - VIA SD/MMC Card Reader driver
+ *  Copyright (c) 2008, VIA Technologies Inc. All Rights Reserved.
+ */
+
+#include <linux/pci.h>
+#include <linux/module.h>
+#include <linux/dma-mapping.h>
+#include <linux/highmem.h>
+#include <linux/delay.h>
+#include <linux/interrupt.h>
+
+#include <linux/mmc/host.h>
+
+#define DRV_NAME	"via_sdmmc"
+
+#define PCI_DEVICE_ID_VIA_9530	0x9530
+
+#define VIA_CRDR_SDC_OFF	0x200
+#define VIA_CRDR_DDMA_OFF	0x400
+#define VIA_CRDR_PCICTRL_OFF	0x600
+
+#define VIA_CRDR_MIN_CLOCK	375000
+#define VIA_CRDR_MAX_CLOCK	48000000
+
+/*
+ * PCI registers
+ */
+
+#define VIA_CRDR_PCI_WORK_MODE	0x40
+#define VIA_CRDR_PCI_DBG_MODE	0x41
+
+/*
+ * SDC MMIO Registers
+ */
+
+#define VIA_CRDR_SDCTRL			0x0
+#define VIA_CRDR_SDCTRL_START		0x01
+#define VIA_CRDR_SDCTRL_WRITE		0x04
+#define VIA_CRDR_SDCTRL_SINGLE_WR	0x10
+#define VIA_CRDR_SDCTRL_SINGLE_RD	0x20
+#define VIA_CRDR_SDCTRL_MULTI_WR	0x30
+#define VIA_CRDR_SDCTRL_MULTI_RD	0x40
+#define VIA_CRDR_SDCTRL_STOP		0x70
+
+#define VIA_CRDR_SDCTRL_RSP_NONE	0x0
+#define VIA_CRDR_SDCTRL_RSP_R1		0x10000
+#define VIA_CRDR_SDCTRL_RSP_R2		0x20000
+#define VIA_CRDR_SDCTRL_RSP_R3		0x30000
+#define VIA_CRDR_SDCTRL_RSP_R1B		0x90000
+
+#define VIA_CRDR_SDCARG 	0x4
+
+#define VIA_CRDR_SDBUSMODE	0x8
+#define VIA_CRDR_SDMODE_4BIT	0x02
+#define VIA_CRDR_SDMODE_CLK_ON	0x40
+
+#define VIA_CRDR_SDBLKLEN	0xc
+/*
+ * Bit 0 -Bit 10 : Block length. So, the maximum block length should be 2048.
+ * Bit 11 - Bit 13 : Reserved.
+ * GPIDET : Select GPI pin to detect card, GPI means CR_CD# in top design.
+ * INTEN : Enable SD host interrupt.
+ * Bit 16 - Bit 31 : Block count. So, the maximun block count should be 65536.
+ */
+#define VIA_CRDR_SDBLKLEN_GPIDET	0x2000
+#define VIA_CRDR_SDBLKLEN_INTEN		0x8000
+#define VIA_CRDR_MAX_BLOCK_COUNT	65536
+#define VIA_CRDR_MAX_BLOCK_LENGTH	2048
+
+#define VIA_CRDR_SDRESP0	0x10
+#define VIA_CRDR_SDRESP1	0x14
+#define VIA_CRDR_SDRESP2	0x18
+#define VIA_CRDR_SDRESP3	0x1c
+
+#define VIA_CRDR_SDCURBLKCNT	0x20
+
+#define VIA_CRDR_SDINTMASK	0x24
+/*
+ * MBDIE : Multiple Blocks transfer Done Interrupt Enable
+ * BDDIE : Block Data transfer Done Interrupt Enable
+ * CIRIE : Card Insertion or Removal Interrupt Enable
+ * CRDIE : Command-Response transfer Done Interrupt Enable
+ * CRTOIE : Command-Response response TimeOut Interrupt Enable
+ * ASCRDIE : Auto Stop Command-Response transfer Done Interrupt Enable
+ * DTIE : Data access Timeout Interrupt Enable
+ * SCIE : reSponse CRC error Interrupt Enable
+ * RCIE : Read data CRC error Interrupt Enable
+ * WCIE : Write data CRC error Interrupt Enable
+ */
+#define VIA_CRDR_SDINTMASK_MBDIE	0x10
+#define VIA_CRDR_SDINTMASK_BDDIE	0x20
+#define VIA_CRDR_SDINTMASK_CIRIE	0x80
+#define VIA_CRDR_SDINTMASK_CRDIE	0x200
+#define VIA_CRDR_SDINTMASK_CRTOIE	0x400
+#define VIA_CRDR_SDINTMASK_ASCRDIE	0x800
+#define VIA_CRDR_SDINTMASK_DTIE		0x1000
+#define VIA_CRDR_SDINTMASK_SCIE		0x2000
+#define VIA_CRDR_SDINTMASK_RCIE		0x4000
+#define VIA_CRDR_SDINTMASK_WCIE		0x8000
+
+#define VIA_CRDR_SDACTIVE_INTMASK \
+	(VIA_CRDR_SDINTMASK_MBDIE | VIA_CRDR_SDINTMASK_CIRIE \
+	| VIA_CRDR_SDINTMASK_CRDIE | VIA_CRDR_SDINTMASK_CRTOIE \
+	| VIA_CRDR_SDINTMASK_DTIE | VIA_CRDR_SDINTMASK_SCIE \
+	| VIA_CRDR_SDINTMASK_RCIE | VIA_CRDR_SDINTMASK_WCIE)
+
+#define VIA_CRDR_SDSTATUS	0x28
+/*
+ * CECC : Reserved
+ * WP : SD card Write Protect status
+ * SLOTD : Reserved
+ * SLOTG : SD SLOT status(Gpi pin status)
+ * MBD : Multiple Blocks transfer Done interrupt status
+ * BDD : Block Data transfer Done interrupt status
+ * CD : Reserved
+ * CIR : Card Insertion or Removal interrupt detected on GPI pin
+ * IO : Reserved
+ * CRD : Command-Response transfer Done interrupt status
+ * CRTO : Command-Response response TimeOut interrupt status
+ * ASCRDIE : Auto Stop Command-Response transfer Done interrupt status
+ * DT : Data access Timeout interrupt status
+ * SC : reSponse CRC error interrupt status
+ * RC : Read data CRC error interrupt status
+ * WC : Write data CRC error interrupt status
+ */
+#define VIA_CRDR_SDSTS_CECC		0x01
+#define VIA_CRDR_SDSTS_WP		0x02
+#define VIA_CRDR_SDSTS_SLOTD		0x04
+#define VIA_CRDR_SDSTS_SLOTG		0x08
+#define VIA_CRDR_SDSTS_MBD		0x10
+#define VIA_CRDR_SDSTS_BDD		0x20
+#define VIA_CRDR_SDSTS_CD		0x40
+#define VIA_CRDR_SDSTS_CIR		0x80
+#define VIA_CRDR_SDSTS_IO		0x100
+#define VIA_CRDR_SDSTS_CRD		0x200
+#define VIA_CRDR_SDSTS_CRTO		0x400
+#define VIA_CRDR_SDSTS_ASCRDIE		0x800
+#define VIA_CRDR_SDSTS_DT		0x1000
+#define VIA_CRDR_SDSTS_SC		0x2000
+#define VIA_CRDR_SDSTS_RC		0x4000
+#define VIA_CRDR_SDSTS_WC		0x8000
+
+#define VIA_CRDR_SDSTS_IGN_MASK\
+	(VIA_CRDR_SDSTS_BDD | VIA_CRDR_SDSTS_ASCRDIE | VIA_CRDR_SDSTS_IO)
+#define VIA_CRDR_SDSTS_INT_MASK \
+	(VIA_CRDR_SDSTS_MBD | VIA_CRDR_SDSTS_BDD | VIA_CRDR_SDSTS_CD \
+	| VIA_CRDR_SDSTS_CIR | VIA_CRDR_SDSTS_IO | VIA_CRDR_SDSTS_CRD \
+	| VIA_CRDR_SDSTS_CRTO | VIA_CRDR_SDSTS_ASCRDIE | VIA_CRDR_SDSTS_DT \
+	| VIA_CRDR_SDSTS_SC | VIA_CRDR_SDSTS_RC | VIA_CRDR_SDSTS_WC)
+#define VIA_CRDR_SDSTS_W1C_MASK \
+	(VIA_CRDR_SDSTS_CECC | VIA_CRDR_SDSTS_MBD | VIA_CRDR_SDSTS_BDD \
+	| VIA_CRDR_SDSTS_CD | VIA_CRDR_SDSTS_CIR | VIA_CRDR_SDSTS_CRD \
+	| VIA_CRDR_SDSTS_CRTO | VIA_CRDR_SDSTS_ASCRDIE | VIA_CRDR_SDSTS_DT \
+	| VIA_CRDR_SDSTS_SC | VIA_CRDR_SDSTS_RC | VIA_CRDR_SDSTS_WC)
+#define  VIA_CRDR_SDSTS_CMD_MASK \
+	(VIA_CRDR_SDSTS_CRD | VIA_CRDR_SDSTS_CRTO | VIA_CRDR_SDSTS_SC)
+#define  VIA_CRDR_SDSTS_DATA_MASK\
+	(VIA_CRDR_SDSTS_MBD | VIA_CRDR_SDSTS_DT \
+	| VIA_CRDR_SDSTS_RC | VIA_CRDR_SDSTS_WC)
+
+#define VIA_CRDR_SDSTATUS2	0x2a
+/*
+ * CFE : Enable SD host automatic Clock FReezing
+ */
+#define VIA_CRDR_SDSTS_CFE		0x80
+
+#define VIA_CRDR_SDRSPTMO	0x2C
+
+#define VIA_CRDR_SDCLKSEL	0x30
+
+#define VIA_CRDR_SDEXTCTRL	0x34
+#define VIS_CRDR_SDEXTCTRL_AUTOSTOP_SD	0x01
+#define VIS_CRDR_SDEXTCTRL_SHIFT_9	0x02
+#define VIS_CRDR_SDEXTCTRL_MMC_8BIT	0x04
+#define VIS_CRDR_SDEXTCTRL_RELD_BLK	0x08
+#define VIS_CRDR_SDEXTCTRL_BAD_CMDA	0x10
+#define VIS_CRDR_SDEXTCTRL_BAD_DATA	0x20
+#define VIS_CRDR_SDEXTCTRL_AUTOSTOP_SPI	0x40
+#define VIA_CRDR_SDEXTCTRL_HISPD	0x80
+/* 0x38-0xFF reserved */
+
+/*
+ * Data DMA Control Registers
+ */
+
+#define VIA_CRDR_DMABASEADD	0x0
+#define VIA_CRDR_DMACOUNTER	0x4
+
+#define VIA_CRDR_DMACTRL	0x8
+/*
+ * DIR :Transaction Direction
+ * 0 : From card to memory
+ * 1 : From memory to card
+ */
+#define VIA_CRDR_DMACTRL_DIR		0x100
+#define VIA_CRDR_DMACTRL_ENIRQ		0x10000
+#define VIA_CRDR_DMACTRL_SFTRST		0x1000000
+
+#define VIA_CRDR_DMASTS		0xc
+
+#define VIA_CRDR_DMASTART	0x10
+/*0x14-0xFF reserved*/
+
+/*
+ * PCI Control Registers
+ */
+
+/*0x0 - 0x1 reserved*/
+#define VIA_CRDR_PCICLKGATT	0x2
+/*
+ * SFTRST :
+ * 0 : Soft reset all the controller and it will be de-asserted automatically
+ * 1 : Soft reset is de-asserted
+ */
+#define VIA_CRDR_PCICLKGATT_SFTRST	0x01
+/*
+ * 3V3 : Pad power select
+ * 0 : 1.8V
+ * 1 : 3.3V
+ * NOTE : No mater what the actual value should be, this bit always
+ * read as 0. This is a hardware bug.
+ */
+#define VIA_CRDR_PCICLKGATT_3V3	0x10
+/*
+ * PAD_PWRON : Pad Power on/off select
+ * 0 : Power off
+ * 1 : Power on
+  * NOTE : No mater what the actual value should be, this bit always
+ * read as 0. This is a hardware bug.
+ */
+#define VIA_CRDR_PCICLKGATT_PAD_PWRON	0x20
+
+#define VIA_CRDR_PCISDCCLK	0x5
+
+#define VIA_CRDR_PCIDMACLK	0x7
+#define VIA_CRDR_PCIDMACLK_SDC		0x2
+
+#define VIA_CRDR_PCIINTCTRL	0x8
+#define VIA_CRDR_PCIINTCTRL_SDCIRQEN	0x04
+
+#define VIA_CRDR_PCIINTSTATUS	0x9
+#define VIA_CRDR_PCIINTSTATUS_SDC	0x04
+
+#define  VIA_CRDR_PCITMOCTRL	0xa
+#define VIA_CRDR_PCITMOCTRL_NO		0x0
+#define VIA_CRDR_PCITMOCTRL_32US	0x1
+#define VIA_CRDR_PCITMOCTRL_256US	0x2
+#define VIA_CRDR_PCITMOCTRL_1024US	0x3
+#define VIA_CRDR_PCITMOCTRL_256MS	0x4
+#define VIA_CRDR_PCITMOCTRL_512MS	0x5
+#define VIA_CRDR_PCITMOCTRL_1024MS	0x6
+
+/*0xB-0xFF reserved*/
+
+enum PCI_HOST_CLK_CONTROL {
+	PCI_CLK_375K = 0x03,
+	PCI_CLK_8M = 0x04,
+	PCI_CLK_12M = 0x00,
+	PCI_CLK_16M = 0x05,
+	PCI_CLK_24M = 0x01,
+	PCI_CLK_33M = 0x06,
+	PCI_CLK_48M = 0x02
+};
+
+struct sdhcreg {
+	u32 sdcontrol_reg;
+	u32 sdcmdarg_reg;
+	u32 sdbusmode_reg;
+	u32 sdblklen_reg;
+	u32 sdresp_reg[4];
+	u32 sdcurblkcnt_reg;
+	u32 sdintmask_reg;
+	u32 sdstatus_reg;
+	u32 sdrsptmo_reg;
+	u32 sdclksel_reg;
+	u32 sdextctrl_reg;
+};
+
+struct pcictrlreg {
+	u8 reserve[2];
+	u8 pciclkgat_reg;
+	u8 pcinfcclk_reg;
+	u8 pcimscclk_reg;
+	u8 pcisdclk_reg;
+	u8 pcicaclk_reg;
+	u8 pcidmaclk_reg;
+	u8 pciintctrl_reg;
+	u8 pciintstatus_reg;
+	u8 pcitmoctrl_reg;
+	u8 Resv;
+};
+
+struct via_crdr_mmc_host {
+	struct mmc_host *mmc;
+	struct mmc_request *mrq;
+	struct mmc_command *cmd;
+	struct mmc_data *data;
+
+	void __iomem *mmiobase;
+	void __iomem *sdhc_mmiobase;
+	void __iomem *ddma_mmiobase;
+	void __iomem *pcictrl_mmiobase;
+
+	struct pcictrlreg pm_pcictrl_reg;
+	struct sdhcreg pm_sdhc_reg;
+
+	struct work_struct carddet_work;
+	struct tasklet_struct finish_tasklet;
+
+	struct timer_list timer;
+	spinlock_t lock;
+	u8 power;
+	int reject;
+	unsigned int quirks;
+};
+
+/* some devices need a very long delay for power to stabilize */
+#define VIA_CRDR_QUIRK_300MS_PWRDELAY	0x0001
+
+#define VIA_CMD_TIMEOUT_MS		1000
+
+static const struct pci_device_id via_ids[] = {
+	{PCI_VENDOR_ID_VIA, PCI_DEVICE_ID_VIA_9530,
+	  PCI_ANY_ID, PCI_ANY_ID, 0, 0, 0,},
+	{0,}
+};
+
+MODULE_DEVICE_TABLE(pci, via_ids);
+
+static void via_print_sdchc(struct via_crdr_mmc_host *host)
+{
+	void __iomem *addrbase = host->sdhc_mmiobase;
+
+	pr_debug("SDC MMIO Registers:\n");
+	pr_debug("SDCONTROL=%08x, SDCMDARG=%08x, SDBUSMODE=%08x\n",
+		 readl(addrbase + VIA_CRDR_SDCTRL),
+		 readl(addrbase + VIA_CRDR_SDCARG),
+		 readl(addrbase + VIA_CRDR_SDBUSMODE));
+	pr_debug("SDBLKLEN=%08x, SDCURBLKCNT=%08x, SDINTMASK=%08x\n",
+		 readl(addrbase + VIA_CRDR_SDBLKLEN),
+		 readl(addrbase + VIA_CRDR_SDCURBLKCNT),
+		 readl(addrbase + VIA_CRDR_SDINTMASK));
+	pr_debug("SDSTATUS=%08x, SDCLKSEL=%08x, SDEXTCTRL=%08x\n",
+		 readl(addrbase + VIA_CRDR_SDSTATUS),
+		 readl(addrbase + VIA_CRDR_SDCLKSEL),
+		 readl(addrbase + VIA_CRDR_SDEXTCTRL));
+}
+
+static void via_print_pcictrl(struct via_crdr_mmc_host *host)
+{
+	void __iomem *addrbase = host->pcictrl_mmiobase;
+
+	pr_debug("PCI Control Registers:\n");
+	pr_debug("PCICLKGATT=%02x, PCISDCCLK=%02x, PCIDMACLK=%02x\n",
+		 readb(addrbase + VIA_CRDR_PCICLKGATT),
+		 readb(addrbase + VIA_CRDR_PCISDCCLK),
+		 readb(addrbase + VIA_CRDR_PCIDMACLK));
+	pr_debug("PCIINTCTRL=%02x, PCIINTSTATUS=%02x\n",
+		 readb(addrbase + VIA_CRDR_PCIINTCTRL),
+		 readb(addrbase + VIA_CRDR_PCIINTSTATUS));
+}
+
+static void via_save_pcictrlreg(struct via_crdr_mmc_host *host)
+{
+	struct pcictrlreg *pm_pcictrl_reg;
+	void __iomem *addrbase;
+
+	pm_pcictrl_reg = &(host->pm_pcictrl_reg);
+	addrbase = host->pcictrl_mmiobase;
+
+	pm_pcictrl_reg->pciclkgat_reg = readb(addrbase + VIA_CRDR_PCICLKGATT);
+	pm_pcictrl_reg->pciclkgat_reg |=
+		VIA_CRDR_PCICLKGATT_3V3 | VIA_CRDR_PCICLKGATT_PAD_PWRON;
+	pm_pcictrl_reg->pcisdclk_reg = readb(addrbase + VIA_CRDR_PCISDCCLK);
+	pm_pcictrl_reg->pcidmaclk_reg = readb(addrbase + VIA_CRDR_PCIDMACLK);
+	pm_pcictrl_reg->pciintctrl_reg = readb(addrbase + VIA_CRDR_PCIINTCTRL);
+	pm_pcictrl_reg->pciintstatus_reg =
+		readb(addrbase + VIA_CRDR_PCIINTSTATUS);
+	pm_pcictrl_reg->pcitmoctrl_reg = readb(addrbase + VIA_CRDR_PCITMOCTRL);
+}
+
+static void via_restore_pcictrlreg(struct via_crdr_mmc_host *host)
+{
+	struct pcictrlreg *pm_pcictrl_reg;
+	void __iomem *addrbase;
+
+	pm_pcictrl_reg = &(host->pm_pcictrl_reg);
+	addrbase = host->pcictrl_mmiobase;
+
+	writeb(pm_pcictrl_reg->pciclkgat_reg, addrbase + VIA_CRDR_PCICLKGATT);
+	writeb(pm_pcictrl_reg->pcisdclk_reg, addrbase + VIA_CRDR_PCISDCCLK);
+	writeb(pm_pcictrl_reg->pcidmaclk_reg, addrbase + VIA_CRDR_PCIDMACLK);
+	writeb(pm_pcictrl_reg->pciintctrl_reg, addrbase + VIA_CRDR_PCIINTCTRL);
+	writeb(pm_pcictrl_reg->pciintstatus_reg,
+		addrbase + VIA_CRDR_PCIINTSTATUS);
+	writeb(pm_pcictrl_reg->pcitmoctrl_reg, addrbase + VIA_CRDR_PCITMOCTRL);
+}
+
+static void via_save_sdcreg(struct via_crdr_mmc_host *host)
+{
+	struct sdhcreg *pm_sdhc_reg;
+	void __iomem *addrbase;
+
+	pm_sdhc_reg = &(host->pm_sdhc_reg);
+	addrbase = host->sdhc_mmiobase;
+
+	pm_sdhc_reg->sdcontrol_reg = readl(addrbase + VIA_CRDR_SDCTRL);
+	pm_sdhc_reg->sdcmdarg_reg = readl(addrbase + VIA_CRDR_SDCARG);
+	pm_sdhc_reg->sdbusmode_reg = readl(addrbase + VIA_CRDR_SDBUSMODE);
+	pm_sdhc_reg->sdblklen_reg = readl(addrbase + VIA_CRDR_SDBLKLEN);
+	pm_sdhc_reg->sdcurblkcnt_reg = readl(addrbase + VIA_CRDR_SDCURBLKCNT);
+	pm_sdhc_reg->sdintmask_reg = readl(addrbase + VIA_CRDR_SDINTMASK);
+	pm_sdhc_reg->sdstatus_reg = readl(addrbase + VIA_CRDR_SDSTATUS);
+	pm_sdhc_reg->sdrsptmo_reg = readl(addrbase + VIA_CRDR_SDRSPTMO);
+	pm_sdhc_reg->sdclksel_reg = readl(addrbase + VIA_CRDR_SDCLKSEL);
+	pm_sdhc_reg->sdextctrl_reg = readl(addrbase + VIA_CRDR_SDEXTCTRL);
+}
+
+static void via_restore_sdcreg(struct via_crdr_mmc_host *host)
+{
+	struct sdhcreg *pm_sdhc_reg;
+	void __iomem *addrbase;
+
+	pm_sdhc_reg = &(host->pm_sdhc_reg);
+	addrbase = host->sdhc_mmiobase;
+
+	writel(pm_sdhc_reg->sdcontrol_reg, addrbase + VIA_CRDR_SDCTRL);
+	writel(pm_sdhc_reg->sdcmdarg_reg, addrbase + VIA_CRDR_SDCARG);
+	writel(pm_sdhc_reg->sdbusmode_reg, addrbase + VIA_CRDR_SDBUSMODE);
+	writel(pm_sdhc_reg->sdblklen_reg, addrbase + VIA_CRDR_SDBLKLEN);
+	writel(pm_sdhc_reg->sdcurblkcnt_reg, addrbase + VIA_CRDR_SDCURBLKCNT);
+	writel(pm_sdhc_reg->sdintmask_reg, addrbase + VIA_CRDR_SDINTMASK);
+	writel(pm_sdhc_reg->sdstatus_reg, addrbase + VIA_CRDR_SDSTATUS);
+	writel(pm_sdhc_reg->sdrsptmo_reg, addrbase + VIA_CRDR_SDRSPTMO);
+	writel(pm_sdhc_reg->sdclksel_reg, addrbase + VIA_CRDR_SDCLKSEL);
+	writel(pm_sdhc_reg->sdextctrl_reg, addrbase + VIA_CRDR_SDEXTCTRL);
+}
+
+static void via_pwron_sleep(struct via_crdr_mmc_host *sdhost)
+{
+	if (sdhost->quirks & VIA_CRDR_QUIRK_300MS_PWRDELAY)
+		msleep(300);
+	else
+		msleep(3);
+}
+
+static void via_set_ddma(struct via_crdr_mmc_host *host,
+			 dma_addr_t dmaaddr, u32 count, int dir, int enirq)
+{
+	void __iomem *addrbase;
+	u32 ctrl_data = 0;
+
+	if (enirq)
+		ctrl_data |= VIA_CRDR_DMACTRL_ENIRQ;
+
+	if (dir)
+		ctrl_data |= VIA_CRDR_DMACTRL_DIR;
+
+	addrbase = host->ddma_mmiobase;
+
+	writel(dmaaddr, addrbase + VIA_CRDR_DMABASEADD);
+	writel(count, addrbase + VIA_CRDR_DMACOUNTER);
+	writel(ctrl_data, addrbase + VIA_CRDR_DMACTRL);
+	writel(0x01, addrbase + VIA_CRDR_DMASTART);
+
+	/* It seems that our DMA can not work normally with 375kHz clock */
+	/* FIXME: don't brute-force 8MHz but use PIO at 375kHz !! */
+	addrbase = host->pcictrl_mmiobase;
+	if (readb(addrbase + VIA_CRDR_PCISDCCLK) == PCI_CLK_375K) {
+		dev_info(host->mmc->parent, "forcing card speed to 8MHz\n");
+		writeb(PCI_CLK_8M, addrbase + VIA_CRDR_PCISDCCLK);
+	}
+}
+
+static void via_sdc_preparedata(struct via_crdr_mmc_host *host,
+				struct mmc_data *data)
+{
+	void __iomem *addrbase;
+	u32 blk_reg;
+	int count;
+
+	WARN_ON(host->data);
+
+	/* Sanity checks */
+	BUG_ON(data->blksz > host->mmc->max_blk_size);
+	BUG_ON(data->blocks > host->mmc->max_blk_count);
+
+	host->data = data;
+
+	count = dma_map_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+		((data->flags & MMC_DATA_READ) ?
+		PCI_DMA_FROMDEVICE : PCI_DMA_TODEVICE));
+	BUG_ON(count != 1);
+
+	via_set_ddma(host, sg_dma_address(data->sg), sg_dma_len(data->sg),
+		(data->flags & MMC_DATA_WRITE) ? 1 : 0, 1);
+
+	addrbase = host->sdhc_mmiobase;
+
+	blk_reg = data->blksz - 1;
+	blk_reg |= VIA_CRDR_SDBLKLEN_GPIDET | VIA_CRDR_SDBLKLEN_INTEN;
+	blk_reg |= (data->blocks) << 16;
+
+	writel(blk_reg, addrbase + VIA_CRDR_SDBLKLEN);
+}
+
+static void via_sdc_get_response(struct via_crdr_mmc_host *host,
+				 struct mmc_command *cmd)
+{
+	void __iomem *addrbase = host->sdhc_mmiobase;
+	u32 dwdata0 = readl(addrbase + VIA_CRDR_SDRESP0);
+	u32 dwdata1 = readl(addrbase + VIA_CRDR_SDRESP1);
+	u32 dwdata2 = readl(addrbase + VIA_CRDR_SDRESP2);
+	u32 dwdata3 = readl(addrbase + VIA_CRDR_SDRESP3);
+
+	if (cmd->flags & MMC_RSP_136) {
+		cmd->resp[0] = ((u8) (dwdata1)) |
+		    (((u8) (dwdata0 >> 24)) << 8) |
+		    (((u8) (dwdata0 >> 16)) << 16) |
+		    (((u8) (dwdata0 >> 8)) << 24);
+
+		cmd->resp[1] = ((u8) (dwdata2)) |
+		    (((u8) (dwdata1 >> 24)) << 8) |
+		    (((u8) (dwdata1 >> 16)) << 16) |
+		    (((u8) (dwdata1 >> 8)) << 24);
+
+		cmd->resp[2] = ((u8) (dwdata3)) |
+		    (((u8) (dwdata2 >> 24)) << 8) |
+		    (((u8) (dwdata2 >> 16)) << 16) |
+		    (((u8) (dwdata2 >> 8)) << 24);
+
+		cmd->resp[3] = 0xff |
+		    ((((u8) (dwdata3 >> 24))) << 8) |
+		    (((u8) (dwdata3 >> 16)) << 16) |
+		    (((u8) (dwdata3 >> 8)) << 24);
+	} else {
+		dwdata0 >>= 8;
+		cmd->resp[0] = ((dwdata0 & 0xff) << 24) |
+		    (((dwdata0 >> 8) & 0xff) << 16) |
+		    (((dwdata0 >> 16) & 0xff) << 8) | (dwdata1 & 0xff);
+
+		dwdata1 >>= 8;
+		cmd->resp[1] = ((dwdata1 & 0xff) << 24) |
+		    (((dwdata1 >> 8) & 0xff) << 16) |
+		    (((dwdata1 >> 16) & 0xff) << 8);
+	}
+}
+
+static void via_sdc_send_command(struct via_crdr_mmc_host *host,
+				 struct mmc_command *cmd)
+{
+	void __iomem *addrbase;
+	struct mmc_data *data;
+	unsigned int timeout_ms;
+	u32 cmdctrl = 0;
+
+	WARN_ON(host->cmd);
+
+	data = cmd->data;
+	host->cmd = cmd;
+
+	timeout_ms = cmd->busy_timeout ? cmd->busy_timeout : VIA_CMD_TIMEOUT_MS;
+	mod_timer(&host->timer, jiffies + msecs_to_jiffies(timeout_ms));
+
+	/*Command index*/
+	cmdctrl = cmd->opcode << 8;
+
+	/*Response type*/
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE:
+		cmdctrl |= VIA_CRDR_SDCTRL_RSP_NONE;
+		break;
+	case MMC_RSP_R1:
+		cmdctrl |= VIA_CRDR_SDCTRL_RSP_R1;
+		break;
+	case MMC_RSP_R1B:
+		cmdctrl |= VIA_CRDR_SDCTRL_RSP_R1B;
+		break;
+	case MMC_RSP_R2:
+		cmdctrl |= VIA_CRDR_SDCTRL_RSP_R2;
+		break;
+	case MMC_RSP_R3:
+		cmdctrl |= VIA_CRDR_SDCTRL_RSP_R3;
+		break;
+	default:
+		pr_err("%s: cmd->flag is not valid\n", mmc_hostname(host->mmc));
+		break;
+	}
+
+	if (!(cmd->data))
+		goto nodata;
+
+	via_sdc_preparedata(host, data);
+
+	/*Command control*/
+	if (data->blocks > 1) {
+		if (data->flags & MMC_DATA_WRITE) {
+			cmdctrl |= VIA_CRDR_SDCTRL_WRITE;
+			cmdctrl |= VIA_CRDR_SDCTRL_MULTI_WR;
+		} else {
+			cmdctrl |= VIA_CRDR_SDCTRL_MULTI_RD;
+		}
+	} else {
+		if (data->flags & MMC_DATA_WRITE) {
+			cmdctrl |= VIA_CRDR_SDCTRL_WRITE;
+			cmdctrl |= VIA_CRDR_SDCTRL_SINGLE_WR;
+		} else {
+			cmdctrl |= VIA_CRDR_SDCTRL_SINGLE_RD;
+		}
+	}
+
+nodata:
+	if (cmd == host->mrq->stop)
+		cmdctrl |= VIA_CRDR_SDCTRL_STOP;
+
+	cmdctrl |= VIA_CRDR_SDCTRL_START;
+
+	addrbase = host->sdhc_mmiobase;
+	writel(cmd->arg, addrbase + VIA_CRDR_SDCARG);
+	writel(cmdctrl, addrbase + VIA_CRDR_SDCTRL);
+}
+
+static void via_sdc_finish_data(struct via_crdr_mmc_host *host)
+{
+	struct mmc_data *data;
+
+	BUG_ON(!host->data);
+
+	data = host->data;
+	host->data = NULL;
+
+	if (data->error)
+		data->bytes_xfered = 0;
+	else
+		data->bytes_xfered = data->blocks * data->blksz;
+
+	dma_unmap_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+		((data->flags & MMC_DATA_READ) ?
+		PCI_DMA_FROMDEVICE : PCI_DMA_TODEVICE));
+
+	if (data->stop)
+		via_sdc_send_command(host, data->stop);
+	else
+		tasklet_schedule(&host->finish_tasklet);
+}
+
+static void via_sdc_finish_command(struct via_crdr_mmc_host *host)
+{
+	via_sdc_get_response(host, host->cmd);
+
+	host->cmd->error = 0;
+
+	if (!host->cmd->data)
+		tasklet_schedule(&host->finish_tasklet);
+
+	host->cmd = NULL;
+}
+
+static void via_sdc_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	void __iomem *addrbase;
+	struct via_crdr_mmc_host *host;
+	unsigned long flags;
+	u16 status;
+
+	host = mmc_priv(mmc);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	addrbase = host->pcictrl_mmiobase;
+	writeb(VIA_CRDR_PCIDMACLK_SDC, addrbase + VIA_CRDR_PCIDMACLK);
+
+	status = readw(host->sdhc_mmiobase + VIA_CRDR_SDSTATUS);
+	status &= VIA_CRDR_SDSTS_W1C_MASK;
+	writew(status, host->sdhc_mmiobase + VIA_CRDR_SDSTATUS);
+
+	WARN_ON(host->mrq != NULL);
+	host->mrq = mrq;
+
+	status = readw(host->sdhc_mmiobase + VIA_CRDR_SDSTATUS);
+	if (!(status & VIA_CRDR_SDSTS_SLOTG) || host->reject) {
+		host->mrq->cmd->error = -ENOMEDIUM;
+		tasklet_schedule(&host->finish_tasklet);
+	} else {
+		via_sdc_send_command(host, mrq->cmd);
+	}
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void via_sdc_set_power(struct via_crdr_mmc_host *host,
+			      unsigned short power, unsigned int on)
+{
+	unsigned long flags;
+	u8 gatt;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	host->power = (1 << power);
+
+	gatt = readb(host->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT);
+	if (host->power == MMC_VDD_165_195)
+		gatt &= ~VIA_CRDR_PCICLKGATT_3V3;
+	else
+		gatt |= VIA_CRDR_PCICLKGATT_3V3;
+	if (on)
+		gatt |= VIA_CRDR_PCICLKGATT_PAD_PWRON;
+	else
+		gatt &= ~VIA_CRDR_PCICLKGATT_PAD_PWRON;
+	writeb(gatt, host->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	via_pwron_sleep(host);
+}
+
+static void via_sdc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct via_crdr_mmc_host *host;
+	unsigned long flags;
+	void __iomem *addrbase;
+	u32 org_data, sdextctrl;
+	u8 clock;
+
+	host = mmc_priv(mmc);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	addrbase = host->sdhc_mmiobase;
+	org_data = readl(addrbase + VIA_CRDR_SDBUSMODE);
+	sdextctrl = readl(addrbase + VIA_CRDR_SDEXTCTRL);
+
+	if (ios->bus_width == MMC_BUS_WIDTH_1)
+		org_data &= ~VIA_CRDR_SDMODE_4BIT;
+	else
+		org_data |= VIA_CRDR_SDMODE_4BIT;
+
+	if (ios->power_mode == MMC_POWER_OFF)
+		org_data &= ~VIA_CRDR_SDMODE_CLK_ON;
+	else
+		org_data |= VIA_CRDR_SDMODE_CLK_ON;
+
+	if (ios->timing == MMC_TIMING_SD_HS)
+		sdextctrl |= VIA_CRDR_SDEXTCTRL_HISPD;
+	else
+		sdextctrl &= ~VIA_CRDR_SDEXTCTRL_HISPD;
+
+	writel(org_data, addrbase + VIA_CRDR_SDBUSMODE);
+	writel(sdextctrl, addrbase + VIA_CRDR_SDEXTCTRL);
+
+	if (ios->clock >= 48000000)
+		clock = PCI_CLK_48M;
+	else if (ios->clock >= 33000000)
+		clock = PCI_CLK_33M;
+	else if (ios->clock >= 24000000)
+		clock = PCI_CLK_24M;
+	else if (ios->clock >= 16000000)
+		clock = PCI_CLK_16M;
+	else if (ios->clock >= 12000000)
+		clock = PCI_CLK_12M;
+	else if (ios->clock >=  8000000)
+		clock = PCI_CLK_8M;
+	else
+		clock = PCI_CLK_375K;
+
+	addrbase = host->pcictrl_mmiobase;
+	if (readb(addrbase + VIA_CRDR_PCISDCCLK) != clock)
+		writeb(clock, addrbase + VIA_CRDR_PCISDCCLK);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (ios->power_mode != MMC_POWER_OFF)
+		via_sdc_set_power(host, ios->vdd, 1);
+	else
+		via_sdc_set_power(host, ios->vdd, 0);
+}
+
+static int via_sdc_get_ro(struct mmc_host *mmc)
+{
+	struct via_crdr_mmc_host *host;
+	unsigned long flags;
+	u16 status;
+
+	host = mmc_priv(mmc);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	status = readw(host->sdhc_mmiobase + VIA_CRDR_SDSTATUS);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	return !(status & VIA_CRDR_SDSTS_WP);
+}
+
+static const struct mmc_host_ops via_sdc_ops = {
+	.request = via_sdc_request,
+	.set_ios = via_sdc_set_ios,
+	.get_ro = via_sdc_get_ro,
+};
+
+static void via_reset_pcictrl(struct via_crdr_mmc_host *host)
+{
+	unsigned long flags;
+	u8 gatt;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	via_save_pcictrlreg(host);
+	via_save_sdcreg(host);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	gatt = VIA_CRDR_PCICLKGATT_PAD_PWRON;
+	if (host->power == MMC_VDD_165_195)
+		gatt &= VIA_CRDR_PCICLKGATT_3V3;
+	else
+		gatt |= VIA_CRDR_PCICLKGATT_3V3;
+	writeb(gatt, host->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT);
+	via_pwron_sleep(host);
+	gatt |= VIA_CRDR_PCICLKGATT_SFTRST;
+	writeb(gatt, host->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT);
+	msleep(3);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	via_restore_pcictrlreg(host);
+	via_restore_sdcreg(host);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void via_sdc_cmd_isr(struct via_crdr_mmc_host *host, u16 intmask)
+{
+	BUG_ON(intmask == 0);
+
+	if (!host->cmd) {
+		pr_err("%s: Got command interrupt 0x%x even "
+		       "though no command operation was in progress.\n",
+		       mmc_hostname(host->mmc), intmask);
+		return;
+	}
+
+	if (intmask & VIA_CRDR_SDSTS_CRTO)
+		host->cmd->error = -ETIMEDOUT;
+	else if (intmask & VIA_CRDR_SDSTS_SC)
+		host->cmd->error = -EILSEQ;
+
+	if (host->cmd->error)
+		tasklet_schedule(&host->finish_tasklet);
+	else if (intmask & VIA_CRDR_SDSTS_CRD)
+		via_sdc_finish_command(host);
+}
+
+static void via_sdc_data_isr(struct via_crdr_mmc_host *host, u16 intmask)
+{
+	BUG_ON(intmask == 0);
+
+	if (!host->data)
+		return;
+
+	if (intmask & VIA_CRDR_SDSTS_DT)
+		host->data->error = -ETIMEDOUT;
+	else if (intmask & (VIA_CRDR_SDSTS_RC | VIA_CRDR_SDSTS_WC))
+		host->data->error = -EILSEQ;
+
+	via_sdc_finish_data(host);
+}
+
+static irqreturn_t via_sdc_isr(int irq, void *dev_id)
+{
+	struct via_crdr_mmc_host *sdhost = dev_id;
+	void __iomem *addrbase;
+	u8 pci_status;
+	u16 sd_status;
+	irqreturn_t result;
+
+	if (!sdhost)
+		return IRQ_NONE;
+
+	spin_lock(&sdhost->lock);
+
+	addrbase = sdhost->pcictrl_mmiobase;
+	pci_status = readb(addrbase + VIA_CRDR_PCIINTSTATUS);
+	if (!(pci_status & VIA_CRDR_PCIINTSTATUS_SDC)) {
+		result = IRQ_NONE;
+		goto out;
+	}
+
+	addrbase = sdhost->sdhc_mmiobase;
+	sd_status = readw(addrbase + VIA_CRDR_SDSTATUS);
+	sd_status &= VIA_CRDR_SDSTS_INT_MASK;
+	sd_status &= ~VIA_CRDR_SDSTS_IGN_MASK;
+	if (!sd_status) {
+		result = IRQ_NONE;
+		goto out;
+	}
+
+	if (sd_status & VIA_CRDR_SDSTS_CIR) {
+		writew(sd_status & VIA_CRDR_SDSTS_CIR,
+			addrbase + VIA_CRDR_SDSTATUS);
+
+		schedule_work(&sdhost->carddet_work);
+	}
+
+	sd_status &= ~VIA_CRDR_SDSTS_CIR;
+	if (sd_status & VIA_CRDR_SDSTS_CMD_MASK) {
+		writew(sd_status & VIA_CRDR_SDSTS_CMD_MASK,
+			addrbase + VIA_CRDR_SDSTATUS);
+		via_sdc_cmd_isr(sdhost, sd_status & VIA_CRDR_SDSTS_CMD_MASK);
+	}
+	if (sd_status & VIA_CRDR_SDSTS_DATA_MASK) {
+		writew(sd_status & VIA_CRDR_SDSTS_DATA_MASK,
+			addrbase + VIA_CRDR_SDSTATUS);
+		via_sdc_data_isr(sdhost, sd_status & VIA_CRDR_SDSTS_DATA_MASK);
+	}
+
+	sd_status &= ~(VIA_CRDR_SDSTS_CMD_MASK | VIA_CRDR_SDSTS_DATA_MASK);
+	if (sd_status) {
+		pr_err("%s: Unexpected interrupt 0x%x\n",
+		       mmc_hostname(sdhost->mmc), sd_status);
+		writew(sd_status, addrbase + VIA_CRDR_SDSTATUS);
+	}
+
+	result = IRQ_HANDLED;
+
+out:
+	spin_unlock(&sdhost->lock);
+
+	return result;
+}
+
+static void via_sdc_timeout(struct timer_list *t)
+{
+	struct via_crdr_mmc_host *sdhost;
+	unsigned long flags;
+
+	sdhost = from_timer(sdhost, t, timer);
+
+	spin_lock_irqsave(&sdhost->lock, flags);
+
+	if (sdhost->mrq) {
+		pr_err("%s: Timeout waiting for hardware interrupt."
+		       "cmd:0x%x\n", mmc_hostname(sdhost->mmc),
+		       sdhost->mrq->cmd->opcode);
+
+		if (sdhost->data) {
+			writel(VIA_CRDR_DMACTRL_SFTRST,
+				sdhost->ddma_mmiobase + VIA_CRDR_DMACTRL);
+			sdhost->data->error = -ETIMEDOUT;
+			via_sdc_finish_data(sdhost);
+		} else {
+			if (sdhost->cmd)
+				sdhost->cmd->error = -ETIMEDOUT;
+			else
+				sdhost->mrq->cmd->error = -ETIMEDOUT;
+			tasklet_schedule(&sdhost->finish_tasklet);
+		}
+	}
+
+	spin_unlock_irqrestore(&sdhost->lock, flags);
+}
+
+static void via_sdc_tasklet_finish(unsigned long param)
+{
+	struct via_crdr_mmc_host *host;
+	unsigned long flags;
+	struct mmc_request *mrq;
+
+	host = (struct via_crdr_mmc_host *)param;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	del_timer(&host->timer);
+	mrq = host->mrq;
+	host->mrq = NULL;
+	host->cmd = NULL;
+	host->data = NULL;
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	mmc_request_done(host->mmc, mrq);
+}
+
+static void via_sdc_card_detect(struct work_struct *work)
+{
+	struct via_crdr_mmc_host *host;
+	void __iomem *addrbase;
+	unsigned long flags;
+	u16 status;
+
+	host = container_of(work, struct via_crdr_mmc_host, carddet_work);
+
+	addrbase = host->ddma_mmiobase;
+	writel(VIA_CRDR_DMACTRL_SFTRST, addrbase + VIA_CRDR_DMACTRL);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	addrbase = host->pcictrl_mmiobase;
+	writeb(VIA_CRDR_PCIDMACLK_SDC, addrbase + VIA_CRDR_PCIDMACLK);
+
+	addrbase = host->sdhc_mmiobase;
+	status = readw(addrbase + VIA_CRDR_SDSTATUS);
+	if (!(status & VIA_CRDR_SDSTS_SLOTG)) {
+		if (host->mrq) {
+			pr_err("%s: Card removed during transfer!\n",
+			       mmc_hostname(host->mmc));
+			host->mrq->cmd->error = -ENOMEDIUM;
+			tasklet_schedule(&host->finish_tasklet);
+		}
+
+		spin_unlock_irqrestore(&host->lock, flags);
+
+		via_reset_pcictrl(host);
+
+		spin_lock_irqsave(&host->lock, flags);
+	}
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	via_print_pcictrl(host);
+	via_print_sdchc(host);
+
+	mmc_detect_change(host->mmc, msecs_to_jiffies(500));
+}
+
+static void via_init_mmc_host(struct via_crdr_mmc_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	void __iomem *addrbase;
+	u32 lenreg;
+	u32 status;
+
+	timer_setup(&host->timer, via_sdc_timeout, 0);
+
+	spin_lock_init(&host->lock);
+
+	mmc->f_min = VIA_CRDR_MIN_CLOCK;
+	mmc->f_max = VIA_CRDR_MAX_CLOCK;
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34 | MMC_VDD_165_195;
+	mmc->caps = MMC_CAP_4_BIT_DATA | MMC_CAP_SD_HIGHSPEED;
+	mmc->ops = &via_sdc_ops;
+
+	/*Hardware cannot do scatter lists*/
+	mmc->max_segs = 1;
+
+	mmc->max_blk_size = VIA_CRDR_MAX_BLOCK_LENGTH;
+	mmc->max_blk_count = VIA_CRDR_MAX_BLOCK_COUNT;
+
+	mmc->max_seg_size = mmc->max_blk_size * mmc->max_blk_count;
+	mmc->max_req_size = mmc->max_seg_size;
+
+	INIT_WORK(&host->carddet_work, via_sdc_card_detect);
+
+	tasklet_init(&host->finish_tasklet, via_sdc_tasklet_finish,
+		     (unsigned long)host);
+
+	addrbase = host->sdhc_mmiobase;
+	writel(0x0, addrbase + VIA_CRDR_SDINTMASK);
+	msleep(1);
+
+	lenreg = VIA_CRDR_SDBLKLEN_GPIDET | VIA_CRDR_SDBLKLEN_INTEN;
+	writel(lenreg, addrbase + VIA_CRDR_SDBLKLEN);
+
+	status = readw(addrbase + VIA_CRDR_SDSTATUS);
+	status &= VIA_CRDR_SDSTS_W1C_MASK;
+	writew(status, addrbase + VIA_CRDR_SDSTATUS);
+
+	status = readw(addrbase + VIA_CRDR_SDSTATUS2);
+	status |= VIA_CRDR_SDSTS_CFE;
+	writew(status, addrbase + VIA_CRDR_SDSTATUS2);
+
+	writeb(0x0, addrbase + VIA_CRDR_SDEXTCTRL);
+
+	writel(VIA_CRDR_SDACTIVE_INTMASK, addrbase + VIA_CRDR_SDINTMASK);
+	msleep(1);
+}
+
+static int via_sd_probe(struct pci_dev *pcidev,
+				    const struct pci_device_id *id)
+{
+	struct mmc_host *mmc;
+	struct via_crdr_mmc_host *sdhost;
+	u32 base, len;
+	u8  gatt;
+	int ret;
+
+	pr_info(DRV_NAME
+		": VIA SDMMC controller found at %s [%04x:%04x] (rev %x)\n",
+		pci_name(pcidev), (int)pcidev->vendor, (int)pcidev->device,
+		(int)pcidev->revision);
+
+	ret = pci_enable_device(pcidev);
+	if (ret)
+		return ret;
+
+	ret = pci_request_regions(pcidev, DRV_NAME);
+	if (ret)
+		goto disable;
+
+	pci_write_config_byte(pcidev, VIA_CRDR_PCI_WORK_MODE, 0);
+	pci_write_config_byte(pcidev, VIA_CRDR_PCI_DBG_MODE, 0);
+
+	mmc = mmc_alloc_host(sizeof(struct via_crdr_mmc_host), &pcidev->dev);
+	if (!mmc) {
+		ret = -ENOMEM;
+		goto release;
+	}
+
+	sdhost = mmc_priv(mmc);
+	sdhost->mmc = mmc;
+	dev_set_drvdata(&pcidev->dev, sdhost);
+
+	len = pci_resource_len(pcidev, 0);
+	base = pci_resource_start(pcidev, 0);
+	sdhost->mmiobase = ioremap(base, len);
+	if (!sdhost->mmiobase) {
+		ret = -ENOMEM;
+		goto free_mmc_host;
+	}
+
+	sdhost->sdhc_mmiobase =
+		sdhost->mmiobase + VIA_CRDR_SDC_OFF;
+	sdhost->ddma_mmiobase =
+		sdhost->mmiobase + VIA_CRDR_DDMA_OFF;
+	sdhost->pcictrl_mmiobase =
+		sdhost->mmiobase + VIA_CRDR_PCICTRL_OFF;
+
+	sdhost->power = MMC_VDD_165_195;
+
+	gatt = VIA_CRDR_PCICLKGATT_3V3 | VIA_CRDR_PCICLKGATT_PAD_PWRON;
+	writeb(gatt, sdhost->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT);
+	via_pwron_sleep(sdhost);
+	gatt |= VIA_CRDR_PCICLKGATT_SFTRST;
+	writeb(gatt, sdhost->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT);
+	msleep(3);
+
+	via_init_mmc_host(sdhost);
+
+	ret =
+	    request_irq(pcidev->irq, via_sdc_isr, IRQF_SHARED, DRV_NAME,
+			sdhost);
+	if (ret)
+		goto unmap;
+
+	writeb(VIA_CRDR_PCIINTCTRL_SDCIRQEN,
+	       sdhost->pcictrl_mmiobase + VIA_CRDR_PCIINTCTRL);
+	writeb(VIA_CRDR_PCITMOCTRL_1024MS,
+	       sdhost->pcictrl_mmiobase + VIA_CRDR_PCITMOCTRL);
+
+	/* device-specific quirks */
+	if (pcidev->subsystem_vendor == PCI_VENDOR_ID_LENOVO &&
+	    pcidev->subsystem_device == 0x3891)
+		sdhost->quirks = VIA_CRDR_QUIRK_300MS_PWRDELAY;
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto unmap;
+
+	return 0;
+
+unmap:
+	iounmap(sdhost->mmiobase);
+free_mmc_host:
+	dev_set_drvdata(&pcidev->dev, NULL);
+	mmc_free_host(mmc);
+release:
+	pci_release_regions(pcidev);
+disable:
+	pci_disable_device(pcidev);
+
+	return ret;
+}
+
+static void via_sd_remove(struct pci_dev *pcidev)
+{
+	struct via_crdr_mmc_host *sdhost = pci_get_drvdata(pcidev);
+	unsigned long flags;
+	u8 gatt;
+
+	spin_lock_irqsave(&sdhost->lock, flags);
+
+	/* Ensure we don't accept more commands from mmc layer */
+	sdhost->reject = 1;
+
+	/* Disable generating further interrupts */
+	writeb(0x0, sdhost->pcictrl_mmiobase + VIA_CRDR_PCIINTCTRL);
+
+	if (sdhost->mrq) {
+		pr_err("%s: Controller removed during "
+			"transfer\n", mmc_hostname(sdhost->mmc));
+
+		/* make sure all DMA is stopped */
+		writel(VIA_CRDR_DMACTRL_SFTRST,
+			sdhost->ddma_mmiobase + VIA_CRDR_DMACTRL);
+		sdhost->mrq->cmd->error = -ENOMEDIUM;
+		if (sdhost->mrq->stop)
+			sdhost->mrq->stop->error = -ENOMEDIUM;
+		tasklet_schedule(&sdhost->finish_tasklet);
+	}
+	spin_unlock_irqrestore(&sdhost->lock, flags);
+
+	mmc_remove_host(sdhost->mmc);
+
+	free_irq(pcidev->irq, sdhost);
+
+	del_timer_sync(&sdhost->timer);
+
+	tasklet_kill(&sdhost->finish_tasklet);
+
+	/* switch off power */
+	gatt = readb(sdhost->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT);
+	gatt &= ~VIA_CRDR_PCICLKGATT_PAD_PWRON;
+	writeb(gatt, sdhost->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT);
+
+	iounmap(sdhost->mmiobase);
+	dev_set_drvdata(&pcidev->dev, NULL);
+	mmc_free_host(sdhost->mmc);
+	pci_release_regions(pcidev);
+	pci_disable_device(pcidev);
+
+	pr_info(DRV_NAME
+		": VIA SDMMC controller at %s [%04x:%04x] has been removed\n",
+		pci_name(pcidev), (int)pcidev->vendor, (int)pcidev->device);
+}
+
+static void __maybe_unused via_init_sdc_pm(struct via_crdr_mmc_host *host)
+{
+	struct sdhcreg *pm_sdhcreg;
+	void __iomem *addrbase;
+	u32 lenreg;
+	u16 status;
+
+	pm_sdhcreg = &(host->pm_sdhc_reg);
+	addrbase = host->sdhc_mmiobase;
+
+	writel(0x0, addrbase + VIA_CRDR_SDINTMASK);
+
+	lenreg = VIA_CRDR_SDBLKLEN_GPIDET | VIA_CRDR_SDBLKLEN_INTEN;
+	writel(lenreg, addrbase + VIA_CRDR_SDBLKLEN);
+
+	status = readw(addrbase + VIA_CRDR_SDSTATUS);
+	status &= VIA_CRDR_SDSTS_W1C_MASK;
+	writew(status, addrbase + VIA_CRDR_SDSTATUS);
+
+	status = readw(addrbase + VIA_CRDR_SDSTATUS2);
+	status |= VIA_CRDR_SDSTS_CFE;
+	writew(status, addrbase + VIA_CRDR_SDSTATUS2);
+
+	writel(pm_sdhcreg->sdcontrol_reg, addrbase + VIA_CRDR_SDCTRL);
+	writel(pm_sdhcreg->sdcmdarg_reg, addrbase + VIA_CRDR_SDCARG);
+	writel(pm_sdhcreg->sdintmask_reg, addrbase + VIA_CRDR_SDINTMASK);
+	writel(pm_sdhcreg->sdrsptmo_reg, addrbase + VIA_CRDR_SDRSPTMO);
+	writel(pm_sdhcreg->sdclksel_reg, addrbase + VIA_CRDR_SDCLKSEL);
+	writel(pm_sdhcreg->sdextctrl_reg, addrbase + VIA_CRDR_SDEXTCTRL);
+
+	via_print_pcictrl(host);
+	via_print_sdchc(host);
+}
+
+static int __maybe_unused via_sd_suspend(struct device *dev)
+{
+	struct via_crdr_mmc_host *host;
+	unsigned long flags;
+
+	host = dev_get_drvdata(dev);
+
+	spin_lock_irqsave(&host->lock, flags);
+	via_save_pcictrlreg(host);
+	via_save_sdcreg(host);
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	device_wakeup_enable(dev);
+
+	return 0;
+}
+
+static int __maybe_unused via_sd_resume(struct device *dev)
+{
+	struct via_crdr_mmc_host *sdhost;
+	int ret = 0;
+	u8 gatt;
+
+	sdhost = dev_get_drvdata(dev);
+
+	gatt = VIA_CRDR_PCICLKGATT_PAD_PWRON;
+	if (sdhost->power == MMC_VDD_165_195)
+		gatt &= ~VIA_CRDR_PCICLKGATT_3V3;
+	else
+		gatt |= VIA_CRDR_PCICLKGATT_3V3;
+	writeb(gatt, sdhost->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT);
+	via_pwron_sleep(sdhost);
+	gatt |= VIA_CRDR_PCICLKGATT_SFTRST;
+	writeb(gatt, sdhost->pcictrl_mmiobase + VIA_CRDR_PCICLKGATT);
+	msleep(3);
+
+	msleep(100);
+
+	via_restore_pcictrlreg(sdhost);
+	via_init_sdc_pm(sdhost);
+
+	return ret;
+}
+
+static SIMPLE_DEV_PM_OPS(via_sd_pm_ops, via_sd_suspend, via_sd_resume);
+
+static struct pci_driver via_sd_driver = {
+	.name = DRV_NAME,
+	.id_table = via_ids,
+	.probe = via_sd_probe,
+	.remove = via_sd_remove,
+	.driver.pm = &via_sd_pm_ops,
+};
+
+module_pci_driver(via_sd_driver);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("VIA Technologies Inc.");
+MODULE_DESCRIPTION("VIA SD/MMC Card Interface driver");
diff --git a/drivers/mmc/host/vub300.c b/drivers/mmc/host/vub300.c
new file mode 100644
index 000000000..05ffd5bf5
--- /dev/null
+++ b/drivers/mmc/host/vub300.c
@@ -0,0 +1,2481 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Remote VUB300 SDIO/SDmem Host Controller Driver
+ *
+ * Copyright (C) 2010 Elan Digital Systems Limited
+ *
+ * based on USB Skeleton driver - 2.2
+ *
+ * Copyright (C) 2001-2004 Greg Kroah-Hartman (greg@kroah.com)
+ *
+ * VUB300: is a USB 2.0 client device with a single SDIO/SDmem/MMC slot
+ *         Any SDIO/SDmem/MMC device plugged into the VUB300 will appear,
+ *         by virtue of this driver, to have been plugged into a local
+ *         SDIO host controller, similar to, say, a PCI Ricoh controller
+ *         This is because this kernel device driver is both a USB 2.0
+ *         client device driver AND an MMC host controller driver. Thus
+ *         if there is an existing driver for the inserted SDIO/SDmem/MMC
+ *         device then that driver will be used by the kernel to manage
+ *         the device in exactly the same fashion as if it had been
+ *         directly plugged into, say, a local pci bus Ricoh controller
+ *
+ * RANT: this driver was written using a display 128x48 - converting it
+ *       to a line width of 80 makes it very difficult to support. In
+ *       particular functions have been broken down into sub functions
+ *       and the original meaningful names have been shortened into
+ *       cryptic ones.
+ *       The problem is that executing a fragment of code subject to
+ *       two conditions means an indentation of 24, thus leaving only
+ *       56 characters for a C statement. And that is quite ridiculous!
+ *
+ * Data types: data passed to/from the VUB300 is fixed to a number of
+ *             bits and driver data fields reflect that limit by using
+ *             u8, u16, u32
+ */
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <linux/kref.h>
+#include <linux/uaccess.h>
+#include <linux/usb.h>
+#include <linux/mutex.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/sdio_func.h>
+#include <linux/mmc/sdio_ids.h>
+#include <linux/workqueue.h>
+#include <linux/ctype.h>
+#include <linux/firmware.h>
+#include <linux/scatterlist.h>
+
+struct host_controller_info {
+	u8 info_size;
+	u16 firmware_version;
+	u8 number_of_ports;
+} __packed;
+
+#define FIRMWARE_BLOCK_BOUNDARY 1024
+struct sd_command_header {
+	u8 header_size;
+	u8 header_type;
+	u8 port_number;
+	u8 command_type; /* Bit7 - Rd/Wr */
+	u8 command_index;
+	u8 transfer_size[4]; /* ReadSize + ReadSize */
+	u8 response_type;
+	u8 arguments[4];
+	u8 block_count[2];
+	u8 block_size[2];
+	u8 block_boundary[2];
+	u8 reserved[44]; /* to pad out to 64 bytes */
+} __packed;
+
+struct sd_irqpoll_header {
+	u8 header_size;
+	u8 header_type;
+	u8 port_number;
+	u8 command_type; /* Bit7 - Rd/Wr */
+	u8 padding[16]; /* don't ask why !! */
+	u8 poll_timeout_msb;
+	u8 poll_timeout_lsb;
+	u8 reserved[42]; /* to pad out to 64 bytes */
+} __packed;
+
+struct sd_common_header {
+	u8 header_size;
+	u8 header_type;
+	u8 port_number;
+} __packed;
+
+struct sd_response_header {
+	u8 header_size;
+	u8 header_type;
+	u8 port_number;
+	u8 command_type;
+	u8 command_index;
+	u8 command_response[];
+} __packed;
+
+struct sd_status_header {
+	u8 header_size;
+	u8 header_type;
+	u8 port_number;
+	u16 port_flags;
+	u32 sdio_clock;
+	u16 host_header_size;
+	u16 func_header_size;
+	u16 ctrl_header_size;
+} __packed;
+
+struct sd_error_header {
+	u8 header_size;
+	u8 header_type;
+	u8 port_number;
+	u8 error_code;
+} __packed;
+
+struct sd_interrupt_header {
+	u8 header_size;
+	u8 header_type;
+	u8 port_number;
+} __packed;
+
+struct offload_registers_access {
+	u8 command_byte[4];
+	u8 Respond_Byte[4];
+} __packed;
+
+#define INTERRUPT_REGISTER_ACCESSES 15
+struct sd_offloaded_interrupt {
+	u8 header_size;
+	u8 header_type;
+	u8 port_number;
+	struct offload_registers_access reg[INTERRUPT_REGISTER_ACCESSES];
+} __packed;
+
+struct sd_register_header {
+	u8 header_size;
+	u8 header_type;
+	u8 port_number;
+	u8 command_type;
+	u8 command_index;
+	u8 command_response[6];
+} __packed;
+
+#define PIGGYBACK_REGISTER_ACCESSES 14
+struct sd_offloaded_piggyback {
+	struct sd_register_header sdio;
+	struct offload_registers_access reg[PIGGYBACK_REGISTER_ACCESSES];
+} __packed;
+
+union sd_response {
+	struct sd_common_header common;
+	struct sd_status_header status;
+	struct sd_error_header error;
+	struct sd_interrupt_header interrupt;
+	struct sd_response_header response;
+	struct sd_offloaded_interrupt irq;
+	struct sd_offloaded_piggyback pig;
+} __packed;
+
+union sd_command {
+	struct sd_command_header head;
+	struct sd_irqpoll_header poll;
+} __packed;
+
+enum SD_RESPONSE_TYPE {
+	SDRT_UNSPECIFIED = 0,
+	SDRT_NONE,
+	SDRT_1,
+	SDRT_1B,
+	SDRT_2,
+	SDRT_3,
+	SDRT_4,
+	SDRT_5,
+	SDRT_5B,
+	SDRT_6,
+	SDRT_7,
+};
+
+#define RESPONSE_INTERRUPT			0x01
+#define RESPONSE_ERROR				0x02
+#define RESPONSE_STATUS				0x03
+#define RESPONSE_IRQ_DISABLED			0x05
+#define RESPONSE_IRQ_ENABLED			0x06
+#define RESPONSE_PIGGYBACKED			0x07
+#define RESPONSE_NO_INTERRUPT			0x08
+#define RESPONSE_PIG_DISABLED			0x09
+#define RESPONSE_PIG_ENABLED			0x0A
+#define SD_ERROR_1BIT_TIMEOUT			0x01
+#define SD_ERROR_4BIT_TIMEOUT			0x02
+#define SD_ERROR_1BIT_CRC_WRONG			0x03
+#define SD_ERROR_4BIT_CRC_WRONG			0x04
+#define SD_ERROR_1BIT_CRC_ERROR			0x05
+#define SD_ERROR_4BIT_CRC_ERROR			0x06
+#define SD_ERROR_NO_CMD_ENDBIT			0x07
+#define SD_ERROR_NO_1BIT_DATEND			0x08
+#define SD_ERROR_NO_4BIT_DATEND			0x09
+#define SD_ERROR_1BIT_UNEXPECTED_TIMEOUT	0x0A
+#define SD_ERROR_4BIT_UNEXPECTED_TIMEOUT	0x0B
+#define SD_ERROR_ILLEGAL_COMMAND		0x0C
+#define SD_ERROR_NO_DEVICE			0x0D
+#define SD_ERROR_TRANSFER_LENGTH		0x0E
+#define SD_ERROR_1BIT_DATA_TIMEOUT		0x0F
+#define SD_ERROR_4BIT_DATA_TIMEOUT		0x10
+#define SD_ERROR_ILLEGAL_STATE			0x11
+#define SD_ERROR_UNKNOWN_ERROR			0x12
+#define SD_ERROR_RESERVED_ERROR			0x13
+#define SD_ERROR_INVALID_FUNCTION		0x14
+#define SD_ERROR_OUT_OF_RANGE			0x15
+#define SD_ERROR_STAT_CMD			0x16
+#define SD_ERROR_STAT_DATA			0x17
+#define SD_ERROR_STAT_CMD_TIMEOUT		0x18
+#define SD_ERROR_SDCRDY_STUCK			0x19
+#define SD_ERROR_UNHANDLED			0x1A
+#define SD_ERROR_OVERRUN			0x1B
+#define SD_ERROR_PIO_TIMEOUT			0x1C
+
+#define FUN(c) (0x000007 & (c->arg>>28))
+#define REG(c) (0x01FFFF & (c->arg>>9))
+
+static bool limit_speed_to_24_MHz;
+module_param(limit_speed_to_24_MHz, bool, 0644);
+MODULE_PARM_DESC(limit_speed_to_24_MHz, "Limit Max SDIO Clock Speed to 24 MHz");
+
+static bool pad_input_to_usb_pkt;
+module_param(pad_input_to_usb_pkt, bool, 0644);
+MODULE_PARM_DESC(pad_input_to_usb_pkt,
+		 "Pad USB data input transfers to whole USB Packet");
+
+static bool disable_offload_processing;
+module_param(disable_offload_processing, bool, 0644);
+MODULE_PARM_DESC(disable_offload_processing, "Disable Offload Processing");
+
+static bool force_1_bit_data_xfers;
+module_param(force_1_bit_data_xfers, bool, 0644);
+MODULE_PARM_DESC(force_1_bit_data_xfers,
+		 "Force SDIO Data Transfers to 1-bit Mode");
+
+static bool force_polling_for_irqs;
+module_param(force_polling_for_irqs, bool, 0644);
+MODULE_PARM_DESC(force_polling_for_irqs, "Force Polling for SDIO interrupts");
+
+static int firmware_irqpoll_timeout = 1024;
+module_param(firmware_irqpoll_timeout, int, 0644);
+MODULE_PARM_DESC(firmware_irqpoll_timeout, "VUB300 firmware irqpoll timeout");
+
+static int force_max_req_size = 128;
+module_param(force_max_req_size, int, 0644);
+MODULE_PARM_DESC(force_max_req_size, "set max request size in kBytes");
+
+#ifdef SMSC_DEVELOPMENT_BOARD
+static int firmware_rom_wait_states = 0x04;
+#else
+static int firmware_rom_wait_states = 0x1C;
+#endif
+
+module_param(firmware_rom_wait_states, int, 0644);
+MODULE_PARM_DESC(firmware_rom_wait_states,
+		 "ROM wait states byte=RRRIIEEE (Reserved Internal External)");
+
+#define ELAN_VENDOR_ID		0x2201
+#define VUB300_VENDOR_ID	0x0424
+#define VUB300_PRODUCT_ID	0x012C
+static const struct usb_device_id vub300_table[] = {
+	{USB_DEVICE(ELAN_VENDOR_ID, VUB300_PRODUCT_ID)},
+	{USB_DEVICE(VUB300_VENDOR_ID, VUB300_PRODUCT_ID)},
+	{} /* Terminating entry */
+};
+MODULE_DEVICE_TABLE(usb, vub300_table);
+
+static struct workqueue_struct *cmndworkqueue;
+static struct workqueue_struct *pollworkqueue;
+static struct workqueue_struct *deadworkqueue;
+
+static inline int interface_to_InterfaceNumber(struct usb_interface *interface)
+{
+	if (!interface)
+		return -1;
+	if (!interface->cur_altsetting)
+		return -1;
+	return interface->cur_altsetting->desc.bInterfaceNumber;
+}
+
+struct sdio_register {
+	unsigned func_num:3;
+	unsigned sdio_reg:17;
+	unsigned activate:1;
+	unsigned prepared:1;
+	unsigned regvalue:8;
+	unsigned response:8;
+	unsigned sparebit:26;
+};
+
+struct vub300_mmc_host {
+	struct usb_device *udev;
+	struct usb_interface *interface;
+	struct kref kref;
+	struct mutex cmd_mutex;
+	struct mutex irq_mutex;
+	char vub_name[3 + (9 * 8) + 4 + 1]; /* max of 7 sdio fn's */
+	u8 cmnd_out_ep; /* EndPoint for commands */
+	u8 cmnd_res_ep; /* EndPoint for responses */
+	u8 data_out_ep; /* EndPoint for out data */
+	u8 data_inp_ep; /* EndPoint for inp data */
+	bool card_powered;
+	bool card_present;
+	bool read_only;
+	bool large_usb_packets;
+	bool app_spec; /* ApplicationSpecific */
+	bool irq_enabled; /* by the MMC CORE */
+	bool irq_disabled; /* in the firmware */
+	unsigned bus_width:4;
+	u8 total_offload_count;
+	u8 dynamic_register_count;
+	u8 resp_len;
+	u32 datasize;
+	int errors;
+	int usb_transport_fail;
+	int usb_timed_out;
+	int irqs_queued;
+	struct sdio_register sdio_register[16];
+	struct offload_interrupt_function_register {
+#define MAXREGBITS 4
+#define MAXREGS (1<<MAXREGBITS)
+#define MAXREGMASK (MAXREGS-1)
+		u8 offload_count;
+		u32 offload_point;
+		struct offload_registers_access reg[MAXREGS];
+	} fn[8];
+	u16 fbs[8]; /* Function Block Size */
+	struct mmc_command *cmd;
+	struct mmc_request *req;
+	struct mmc_data *data;
+	struct mmc_host *mmc;
+	struct urb *urb;
+	struct urb *command_out_urb;
+	struct urb *command_res_urb;
+	struct completion command_complete;
+	struct completion irqpoll_complete;
+	union sd_command cmnd;
+	union sd_response resp;
+	struct timer_list sg_transfer_timer;
+	struct usb_sg_request sg_request;
+	struct timer_list inactivity_timer;
+	struct work_struct deadwork;
+	struct work_struct cmndwork;
+	struct delayed_work pollwork;
+	struct host_controller_info hc_info;
+	struct sd_status_header system_port_status;
+	u8 padded_buffer[64];
+};
+
+#define kref_to_vub300_mmc_host(d) container_of(d, struct vub300_mmc_host, kref)
+#define SET_TRANSFER_PSEUDOCODE		21
+#define SET_INTERRUPT_PSEUDOCODE	20
+#define SET_FAILURE_MODE		18
+#define SET_ROM_WAIT_STATES		16
+#define SET_IRQ_ENABLE			13
+#define SET_CLOCK_SPEED			11
+#define SET_FUNCTION_BLOCK_SIZE		9
+#define SET_SD_DATA_MODE		6
+#define SET_SD_POWER			4
+#define ENTER_DFU_MODE			3
+#define GET_HC_INF0			1
+#define GET_SYSTEM_PORT_STATUS		0
+
+static void vub300_delete(struct kref *kref)
+{				/* kref callback - softirq */
+	struct vub300_mmc_host *vub300 = kref_to_vub300_mmc_host(kref);
+	struct mmc_host *mmc = vub300->mmc;
+	usb_free_urb(vub300->command_out_urb);
+	vub300->command_out_urb = NULL;
+	usb_free_urb(vub300->command_res_urb);
+	vub300->command_res_urb = NULL;
+	usb_put_dev(vub300->udev);
+	mmc_free_host(mmc);
+	/*
+	 * and hence also frees vub300
+	 * which is contained at the end of struct mmc
+	 */
+}
+
+static void vub300_queue_cmnd_work(struct vub300_mmc_host *vub300)
+{
+	kref_get(&vub300->kref);
+	if (queue_work(cmndworkqueue, &vub300->cmndwork)) {
+		/*
+		 * then the cmndworkqueue was not previously
+		 * running and the above get ref is obvious
+		 * required and will be put when the thread
+		 * terminates by a specific call
+		 */
+	} else {
+		/*
+		 * the cmndworkqueue was already running from
+		 * a previous invocation and thus to keep the
+		 * kref counts correct we must undo the get
+		 */
+		kref_put(&vub300->kref, vub300_delete);
+	}
+}
+
+static void vub300_queue_poll_work(struct vub300_mmc_host *vub300, int delay)
+{
+	kref_get(&vub300->kref);
+	if (queue_delayed_work(pollworkqueue, &vub300->pollwork, delay)) {
+		/*
+		 * then the pollworkqueue was not previously
+		 * running and the above get ref is obvious
+		 * required and will be put when the thread
+		 * terminates by a specific call
+		 */
+	} else {
+		/*
+		 * the pollworkqueue was already running from
+		 * a previous invocation and thus to keep the
+		 * kref counts correct we must undo the get
+		 */
+		kref_put(&vub300->kref, vub300_delete);
+	}
+}
+
+static void vub300_queue_dead_work(struct vub300_mmc_host *vub300)
+{
+	kref_get(&vub300->kref);
+	if (queue_work(deadworkqueue, &vub300->deadwork)) {
+		/*
+		 * then the deadworkqueue was not previously
+		 * running and the above get ref is obvious
+		 * required and will be put when the thread
+		 * terminates by a specific call
+		 */
+	} else {
+		/*
+		 * the deadworkqueue was already running from
+		 * a previous invocation and thus to keep the
+		 * kref counts correct we must undo the get
+		 */
+		kref_put(&vub300->kref, vub300_delete);
+	}
+}
+
+static void irqpoll_res_completed(struct urb *urb)
+{				/* urb completion handler - hardirq */
+	struct vub300_mmc_host *vub300 = (struct vub300_mmc_host *)urb->context;
+	if (urb->status)
+		vub300->usb_transport_fail = urb->status;
+	complete(&vub300->irqpoll_complete);
+}
+
+static void irqpoll_out_completed(struct urb *urb)
+{				/* urb completion handler - hardirq */
+	struct vub300_mmc_host *vub300 = (struct vub300_mmc_host *)urb->context;
+	if (urb->status) {
+		vub300->usb_transport_fail = urb->status;
+		complete(&vub300->irqpoll_complete);
+		return;
+	} else {
+		int ret;
+		unsigned int pipe =
+			usb_rcvbulkpipe(vub300->udev, vub300->cmnd_res_ep);
+		usb_fill_bulk_urb(vub300->command_res_urb, vub300->udev, pipe,
+				  &vub300->resp, sizeof(vub300->resp),
+				  irqpoll_res_completed, vub300);
+		vub300->command_res_urb->actual_length = 0;
+		ret = usb_submit_urb(vub300->command_res_urb, GFP_ATOMIC);
+		if (ret) {
+			vub300->usb_transport_fail = ret;
+			complete(&vub300->irqpoll_complete);
+		}
+		return;
+	}
+}
+
+static void send_irqpoll(struct vub300_mmc_host *vub300)
+{
+	/* cmd_mutex is held by vub300_pollwork_thread */
+	int retval;
+	int timeout = 0xFFFF & (0x0001FFFF - firmware_irqpoll_timeout);
+	vub300->cmnd.poll.header_size = 22;
+	vub300->cmnd.poll.header_type = 1;
+	vub300->cmnd.poll.port_number = 0;
+	vub300->cmnd.poll.command_type = 2;
+	vub300->cmnd.poll.poll_timeout_lsb = 0xFF & (unsigned)timeout;
+	vub300->cmnd.poll.poll_timeout_msb = 0xFF & (unsigned)(timeout >> 8);
+	usb_fill_bulk_urb(vub300->command_out_urb, vub300->udev,
+			  usb_sndbulkpipe(vub300->udev, vub300->cmnd_out_ep)
+			  , &vub300->cmnd, sizeof(vub300->cmnd)
+			  , irqpoll_out_completed, vub300);
+	retval = usb_submit_urb(vub300->command_out_urb, GFP_KERNEL);
+	if (0 > retval) {
+		vub300->usb_transport_fail = retval;
+		vub300_queue_poll_work(vub300, 1);
+		complete(&vub300->irqpoll_complete);
+		return;
+	} else {
+		return;
+	}
+}
+
+static void new_system_port_status(struct vub300_mmc_host *vub300)
+{
+	int old_card_present = vub300->card_present;
+	int new_card_present =
+		(0x0001 & vub300->system_port_status.port_flags) ? 1 : 0;
+	vub300->read_only =
+		(0x0010 & vub300->system_port_status.port_flags) ? 1 : 0;
+	if (new_card_present && !old_card_present) {
+		dev_info(&vub300->udev->dev, "card just inserted\n");
+		vub300->card_present = 1;
+		vub300->bus_width = 0;
+		if (disable_offload_processing)
+			strncpy(vub300->vub_name, "EMPTY Processing Disabled",
+				sizeof(vub300->vub_name));
+		else
+			vub300->vub_name[0] = 0;
+		mmc_detect_change(vub300->mmc, 1);
+	} else if (!new_card_present && old_card_present) {
+		dev_info(&vub300->udev->dev, "card just ejected\n");
+		vub300->card_present = 0;
+		mmc_detect_change(vub300->mmc, 0);
+	} else {
+		/* no change */
+	}
+}
+
+static void __add_offloaded_reg_to_fifo(struct vub300_mmc_host *vub300,
+					struct offload_registers_access
+					*register_access, u8 func)
+{
+	u8 r = vub300->fn[func].offload_point + vub300->fn[func].offload_count;
+	memcpy(&vub300->fn[func].reg[MAXREGMASK & r], register_access,
+	       sizeof(struct offload_registers_access));
+	vub300->fn[func].offload_count += 1;
+	vub300->total_offload_count += 1;
+}
+
+static void add_offloaded_reg(struct vub300_mmc_host *vub300,
+			      struct offload_registers_access *register_access)
+{
+	u32 Register = ((0x03 & register_access->command_byte[0]) << 15)
+			| ((0xFF & register_access->command_byte[1]) << 7)
+			| ((0xFE & register_access->command_byte[2]) >> 1);
+	u8 func = ((0x70 & register_access->command_byte[0]) >> 4);
+	u8 regs = vub300->dynamic_register_count;
+	u8 i = 0;
+	while (0 < regs-- && 1 == vub300->sdio_register[i].activate) {
+		if (vub300->sdio_register[i].func_num == func &&
+		    vub300->sdio_register[i].sdio_reg == Register) {
+			if (vub300->sdio_register[i].prepared == 0)
+				vub300->sdio_register[i].prepared = 1;
+			vub300->sdio_register[i].response =
+				register_access->Respond_Byte[2];
+			vub300->sdio_register[i].regvalue =
+				register_access->Respond_Byte[3];
+			return;
+		} else {
+			i += 1;
+			continue;
+		}
+	}
+	__add_offloaded_reg_to_fifo(vub300, register_access, func);
+}
+
+static void check_vub300_port_status(struct vub300_mmc_host *vub300)
+{
+	/*
+	 * cmd_mutex is held by vub300_pollwork_thread,
+	 * vub300_deadwork_thread or vub300_cmndwork_thread
+	 */
+	int retval;
+	retval =
+		usb_control_msg(vub300->udev, usb_rcvctrlpipe(vub300->udev, 0),
+				GET_SYSTEM_PORT_STATUS,
+				USB_DIR_IN | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+				0x0000, 0x0000, &vub300->system_port_status,
+				sizeof(vub300->system_port_status), 1000);
+	if (sizeof(vub300->system_port_status) == retval)
+		new_system_port_status(vub300);
+}
+
+static void __vub300_irqpoll_response(struct vub300_mmc_host *vub300)
+{
+	/* cmd_mutex is held by vub300_pollwork_thread */
+	if (vub300->command_res_urb->actual_length == 0)
+		return;
+
+	switch (vub300->resp.common.header_type) {
+	case RESPONSE_INTERRUPT:
+		mutex_lock(&vub300->irq_mutex);
+		if (vub300->irq_enabled)
+			mmc_signal_sdio_irq(vub300->mmc);
+		else
+			vub300->irqs_queued += 1;
+		vub300->irq_disabled = 1;
+		mutex_unlock(&vub300->irq_mutex);
+		break;
+	case RESPONSE_ERROR:
+		if (vub300->resp.error.error_code == SD_ERROR_NO_DEVICE)
+			check_vub300_port_status(vub300);
+		break;
+	case RESPONSE_STATUS:
+		vub300->system_port_status = vub300->resp.status;
+		new_system_port_status(vub300);
+		if (!vub300->card_present)
+			vub300_queue_poll_work(vub300, HZ / 5);
+		break;
+	case RESPONSE_IRQ_DISABLED:
+	{
+		int offloaded_data_length = vub300->resp.common.header_size - 3;
+		int register_count = offloaded_data_length >> 3;
+		int ri = 0;
+		while (register_count--) {
+			add_offloaded_reg(vub300, &vub300->resp.irq.reg[ri]);
+			ri += 1;
+		}
+		mutex_lock(&vub300->irq_mutex);
+		if (vub300->irq_enabled)
+			mmc_signal_sdio_irq(vub300->mmc);
+		else
+			vub300->irqs_queued += 1;
+		vub300->irq_disabled = 1;
+		mutex_unlock(&vub300->irq_mutex);
+		break;
+	}
+	case RESPONSE_IRQ_ENABLED:
+	{
+		int offloaded_data_length = vub300->resp.common.header_size - 3;
+		int register_count = offloaded_data_length >> 3;
+		int ri = 0;
+		while (register_count--) {
+			add_offloaded_reg(vub300, &vub300->resp.irq.reg[ri]);
+			ri += 1;
+		}
+		mutex_lock(&vub300->irq_mutex);
+		if (vub300->irq_enabled)
+			mmc_signal_sdio_irq(vub300->mmc);
+		else
+			vub300->irqs_queued += 1;
+		vub300->irq_disabled = 0;
+		mutex_unlock(&vub300->irq_mutex);
+		break;
+	}
+	case RESPONSE_NO_INTERRUPT:
+		vub300_queue_poll_work(vub300, 1);
+		break;
+	default:
+		break;
+	}
+}
+
+static void __do_poll(struct vub300_mmc_host *vub300)
+{
+	/* cmd_mutex is held by vub300_pollwork_thread */
+	unsigned long commretval;
+	mod_timer(&vub300->inactivity_timer, jiffies + HZ);
+	init_completion(&vub300->irqpoll_complete);
+	send_irqpoll(vub300);
+	commretval = wait_for_completion_timeout(&vub300->irqpoll_complete,
+						 msecs_to_jiffies(500));
+	if (vub300->usb_transport_fail) {
+		/* no need to do anything */
+	} else if (commretval == 0) {
+		vub300->usb_timed_out = 1;
+		usb_kill_urb(vub300->command_out_urb);
+		usb_kill_urb(vub300->command_res_urb);
+	} else { /* commretval > 0 */
+		__vub300_irqpoll_response(vub300);
+	}
+}
+
+/* this thread runs only when the driver
+ * is trying to poll the device for an IRQ
+ */
+static void vub300_pollwork_thread(struct work_struct *work)
+{				/* NOT irq */
+	struct vub300_mmc_host *vub300 = container_of(work,
+			      struct vub300_mmc_host, pollwork.work);
+	if (!vub300->interface) {
+		kref_put(&vub300->kref, vub300_delete);
+		return;
+	}
+	mutex_lock(&vub300->cmd_mutex);
+	if (vub300->cmd) {
+		vub300_queue_poll_work(vub300, 1);
+	} else if (!vub300->card_present) {
+		/* no need to do anything */
+	} else { /* vub300->card_present */
+		mutex_lock(&vub300->irq_mutex);
+		if (!vub300->irq_enabled) {
+			mutex_unlock(&vub300->irq_mutex);
+		} else if (vub300->irqs_queued) {
+			vub300->irqs_queued -= 1;
+			mmc_signal_sdio_irq(vub300->mmc);
+			mod_timer(&vub300->inactivity_timer, jiffies + HZ);
+			mutex_unlock(&vub300->irq_mutex);
+		} else { /* NOT vub300->irqs_queued */
+			mutex_unlock(&vub300->irq_mutex);
+			__do_poll(vub300);
+		}
+	}
+	mutex_unlock(&vub300->cmd_mutex);
+	kref_put(&vub300->kref, vub300_delete);
+}
+
+static void vub300_deadwork_thread(struct work_struct *work)
+{				/* NOT irq */
+	struct vub300_mmc_host *vub300 =
+		container_of(work, struct vub300_mmc_host, deadwork);
+	if (!vub300->interface) {
+		kref_put(&vub300->kref, vub300_delete);
+		return;
+	}
+	mutex_lock(&vub300->cmd_mutex);
+	if (vub300->cmd) {
+		/*
+		 * a command got in as the inactivity
+		 * timer expired - so we just let the
+		 * processing of the command show if
+		 * the device is dead
+		 */
+	} else if (vub300->card_present) {
+		check_vub300_port_status(vub300);
+	} else if (vub300->mmc && vub300->mmc->card) {
+		/*
+		 * the MMC core must not have responded
+		 * to the previous indication - lets
+		 * hope that it eventually does so we
+		 * will just ignore this for now
+		 */
+	} else {
+		check_vub300_port_status(vub300);
+	}
+	mod_timer(&vub300->inactivity_timer, jiffies + HZ);
+	mutex_unlock(&vub300->cmd_mutex);
+	kref_put(&vub300->kref, vub300_delete);
+}
+
+static void vub300_inactivity_timer_expired(struct timer_list *t)
+{				/* softirq */
+	struct vub300_mmc_host *vub300 = from_timer(vub300, t,
+						    inactivity_timer);
+	if (!vub300->interface) {
+		kref_put(&vub300->kref, vub300_delete);
+	} else if (vub300->cmd) {
+		mod_timer(&vub300->inactivity_timer, jiffies + HZ);
+	} else {
+		vub300_queue_dead_work(vub300);
+		mod_timer(&vub300->inactivity_timer, jiffies + HZ);
+	}
+}
+
+static int vub300_response_error(u8 error_code)
+{
+	switch (error_code) {
+	case SD_ERROR_PIO_TIMEOUT:
+	case SD_ERROR_1BIT_TIMEOUT:
+	case SD_ERROR_4BIT_TIMEOUT:
+		return -ETIMEDOUT;
+	case SD_ERROR_STAT_DATA:
+	case SD_ERROR_OVERRUN:
+	case SD_ERROR_STAT_CMD:
+	case SD_ERROR_STAT_CMD_TIMEOUT:
+	case SD_ERROR_SDCRDY_STUCK:
+	case SD_ERROR_UNHANDLED:
+	case SD_ERROR_1BIT_CRC_WRONG:
+	case SD_ERROR_4BIT_CRC_WRONG:
+	case SD_ERROR_1BIT_CRC_ERROR:
+	case SD_ERROR_4BIT_CRC_ERROR:
+	case SD_ERROR_NO_CMD_ENDBIT:
+	case SD_ERROR_NO_1BIT_DATEND:
+	case SD_ERROR_NO_4BIT_DATEND:
+	case SD_ERROR_1BIT_DATA_TIMEOUT:
+	case SD_ERROR_4BIT_DATA_TIMEOUT:
+	case SD_ERROR_1BIT_UNEXPECTED_TIMEOUT:
+	case SD_ERROR_4BIT_UNEXPECTED_TIMEOUT:
+		return -EILSEQ;
+	case 33:
+		return -EILSEQ;
+	case SD_ERROR_ILLEGAL_COMMAND:
+		return -EINVAL;
+	case SD_ERROR_NO_DEVICE:
+		return -ENOMEDIUM;
+	default:
+		return -ENODEV;
+	}
+}
+
+static void command_res_completed(struct urb *urb)
+{				/* urb completion handler - hardirq */
+	struct vub300_mmc_host *vub300 = (struct vub300_mmc_host *)urb->context;
+	if (urb->status) {
+		/* we have to let the initiator handle the error */
+	} else if (vub300->command_res_urb->actual_length == 0) {
+		/*
+		 * we have seen this happen once or twice and
+		 * we suspect a buggy USB host controller
+		 */
+	} else if (!vub300->data) {
+		/* this means that the command (typically CMD52) succeeded */
+	} else if (vub300->resp.common.header_type != 0x02) {
+		/*
+		 * this is an error response from the VUB300 chip
+		 * and we let the initiator handle it
+		 */
+	} else if (vub300->urb) {
+		vub300->cmd->error =
+			vub300_response_error(vub300->resp.error.error_code);
+		usb_unlink_urb(vub300->urb);
+	} else {
+		vub300->cmd->error =
+			vub300_response_error(vub300->resp.error.error_code);
+		usb_sg_cancel(&vub300->sg_request);
+	}
+	complete(&vub300->command_complete);	/* got_response_in */
+}
+
+static void command_out_completed(struct urb *urb)
+{				/* urb completion handler - hardirq */
+	struct vub300_mmc_host *vub300 = (struct vub300_mmc_host *)urb->context;
+	if (urb->status) {
+		complete(&vub300->command_complete);
+	} else {
+		int ret;
+		unsigned int pipe =
+			usb_rcvbulkpipe(vub300->udev, vub300->cmnd_res_ep);
+		usb_fill_bulk_urb(vub300->command_res_urb, vub300->udev, pipe,
+				  &vub300->resp, sizeof(vub300->resp),
+				  command_res_completed, vub300);
+		vub300->command_res_urb->actual_length = 0;
+		ret = usb_submit_urb(vub300->command_res_urb, GFP_ATOMIC);
+		if (ret == 0) {
+			/*
+			 * the urb completion handler will call
+			 * our completion handler
+			 */
+		} else {
+			/*
+			 * and thus we only call it directly
+			 * when it will not be called
+			 */
+			complete(&vub300->command_complete);
+		}
+	}
+}
+
+/*
+ * the STUFF bits are masked out for the comparisons
+ */
+static void snoop_block_size_and_bus_width(struct vub300_mmc_host *vub300,
+					   u32 cmd_arg)
+{
+	if ((0xFBFFFE00 & cmd_arg) == 0x80022200)
+		vub300->fbs[1] = (cmd_arg << 8) | (0x00FF & vub300->fbs[1]);
+	else if ((0xFBFFFE00 & cmd_arg) == 0x80022000)
+		vub300->fbs[1] = (0xFF & cmd_arg) | (0xFF00 & vub300->fbs[1]);
+	else if ((0xFBFFFE00 & cmd_arg) == 0x80042200)
+		vub300->fbs[2] = (cmd_arg << 8) | (0x00FF & vub300->fbs[2]);
+	else if ((0xFBFFFE00 & cmd_arg) == 0x80042000)
+		vub300->fbs[2] = (0xFF & cmd_arg) | (0xFF00 & vub300->fbs[2]);
+	else if ((0xFBFFFE00 & cmd_arg) == 0x80062200)
+		vub300->fbs[3] = (cmd_arg << 8) | (0x00FF & vub300->fbs[3]);
+	else if ((0xFBFFFE00 & cmd_arg) == 0x80062000)
+		vub300->fbs[3] = (0xFF & cmd_arg) | (0xFF00 & vub300->fbs[3]);
+	else if ((0xFBFFFE00 & cmd_arg) == 0x80082200)
+		vub300->fbs[4] = (cmd_arg << 8) | (0x00FF & vub300->fbs[4]);
+	else if ((0xFBFFFE00 & cmd_arg) == 0x80082000)
+		vub300->fbs[4] = (0xFF & cmd_arg) | (0xFF00 & vub300->fbs[4]);
+	else if ((0xFBFFFE00 & cmd_arg) == 0x800A2200)
+		vub300->fbs[5] = (cmd_arg << 8) | (0x00FF & vub300->fbs[5]);
+	else if ((0xFBFFFE00 & cmd_arg) == 0x800A2000)
+		vub300->fbs[5] = (0xFF & cmd_arg) | (0xFF00 & vub300->fbs[5]);
+	else if ((0xFBFFFE00 & cmd_arg) == 0x800C2200)
+		vub300->fbs[6] = (cmd_arg << 8) | (0x00FF & vub300->fbs[6]);
+	else if ((0xFBFFFE00 & cmd_arg) == 0x800C2000)
+		vub300->fbs[6] = (0xFF & cmd_arg) | (0xFF00 & vub300->fbs[6]);
+	else if ((0xFBFFFE00 & cmd_arg) == 0x800E2200)
+		vub300->fbs[7] = (cmd_arg << 8) | (0x00FF & vub300->fbs[7]);
+	else if ((0xFBFFFE00 & cmd_arg) == 0x800E2000)
+		vub300->fbs[7] = (0xFF & cmd_arg) | (0xFF00 & vub300->fbs[7]);
+	else if ((0xFBFFFE03 & cmd_arg) == 0x80000E00)
+		vub300->bus_width = 1;
+	else if ((0xFBFFFE03 & cmd_arg) == 0x80000E02)
+		vub300->bus_width = 4;
+}
+
+static void send_command(struct vub300_mmc_host *vub300)
+{
+	/* cmd_mutex is held by vub300_cmndwork_thread */
+	struct mmc_command *cmd = vub300->cmd;
+	struct mmc_data *data = vub300->data;
+	int retval;
+	int i;
+	u8 response_type;
+	if (vub300->app_spec) {
+		switch (cmd->opcode) {
+		case 6:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			if (0x00000000 == (0x00000003 & cmd->arg))
+				vub300->bus_width = 1;
+			else if (0x00000002 == (0x00000003 & cmd->arg))
+				vub300->bus_width = 4;
+			else
+				dev_err(&vub300->udev->dev,
+					"unexpected ACMD6 bus_width=%d\n",
+					0x00000003 & cmd->arg);
+			break;
+		case 13:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			break;
+		case 22:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			break;
+		case 23:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			break;
+		case 41:
+			response_type = SDRT_3;
+			vub300->resp_len = 6;
+			break;
+		case 42:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			break;
+		case 51:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			break;
+		case 55:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			break;
+		default:
+			vub300->resp_len = 0;
+			cmd->error = -EINVAL;
+			complete(&vub300->command_complete);
+			return;
+		}
+		vub300->app_spec = 0;
+	} else {
+		switch (cmd->opcode) {
+		case 0:
+			response_type = SDRT_NONE;
+			vub300->resp_len = 0;
+			break;
+		case 1:
+			response_type = SDRT_3;
+			vub300->resp_len = 6;
+			break;
+		case 2:
+			response_type = SDRT_2;
+			vub300->resp_len = 17;
+			break;
+		case 3:
+			response_type = SDRT_6;
+			vub300->resp_len = 6;
+			break;
+		case 4:
+			response_type = SDRT_NONE;
+			vub300->resp_len = 0;
+			break;
+		case 5:
+			response_type = SDRT_4;
+			vub300->resp_len = 6;
+			break;
+		case 6:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			break;
+		case 7:
+			response_type = SDRT_1B;
+			vub300->resp_len = 6;
+			break;
+		case 8:
+			response_type = SDRT_7;
+			vub300->resp_len = 6;
+			break;
+		case 9:
+			response_type = SDRT_2;
+			vub300->resp_len = 17;
+			break;
+		case 10:
+			response_type = SDRT_2;
+			vub300->resp_len = 17;
+			break;
+		case 12:
+			response_type = SDRT_1B;
+			vub300->resp_len = 6;
+			break;
+		case 13:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			break;
+		case 15:
+			response_type = SDRT_NONE;
+			vub300->resp_len = 0;
+			break;
+		case 16:
+			for (i = 0; i < ARRAY_SIZE(vub300->fbs); i++)
+				vub300->fbs[i] = 0xFFFF & cmd->arg;
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			break;
+		case 17:
+		case 18:
+		case 24:
+		case 25:
+		case 27:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			break;
+		case 28:
+		case 29:
+			response_type = SDRT_1B;
+			vub300->resp_len = 6;
+			break;
+		case 30:
+		case 32:
+		case 33:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			break;
+		case 38:
+			response_type = SDRT_1B;
+			vub300->resp_len = 6;
+			break;
+		case 42:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			break;
+		case 52:
+			response_type = SDRT_5;
+			vub300->resp_len = 6;
+			snoop_block_size_and_bus_width(vub300, cmd->arg);
+			break;
+		case 53:
+			response_type = SDRT_5;
+			vub300->resp_len = 6;
+			break;
+		case 55:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			vub300->app_spec = 1;
+			break;
+		case 56:
+			response_type = SDRT_1;
+			vub300->resp_len = 6;
+			break;
+		default:
+			vub300->resp_len = 0;
+			cmd->error = -EINVAL;
+			complete(&vub300->command_complete);
+			return;
+		}
+	}
+	/*
+	 * it is a shame that we can not use "sizeof(struct sd_command_header)"
+	 * this is because the packet _must_ be padded to 64 bytes
+	 */
+	vub300->cmnd.head.header_size = 20;
+	vub300->cmnd.head.header_type = 0x00;
+	vub300->cmnd.head.port_number = 0; /* "0" means port 1 */
+	vub300->cmnd.head.command_type = 0x00; /* standard read command */
+	vub300->cmnd.head.response_type = response_type;
+	vub300->cmnd.head.command_index = cmd->opcode;
+	vub300->cmnd.head.arguments[0] = cmd->arg >> 24;
+	vub300->cmnd.head.arguments[1] = cmd->arg >> 16;
+	vub300->cmnd.head.arguments[2] = cmd->arg >> 8;
+	vub300->cmnd.head.arguments[3] = cmd->arg >> 0;
+	if (cmd->opcode == 52) {
+		int fn = 0x7 & (cmd->arg >> 28);
+		vub300->cmnd.head.block_count[0] = 0;
+		vub300->cmnd.head.block_count[1] = 0;
+		vub300->cmnd.head.block_size[0] = (vub300->fbs[fn] >> 8) & 0xFF;
+		vub300->cmnd.head.block_size[1] = (vub300->fbs[fn] >> 0) & 0xFF;
+		vub300->cmnd.head.command_type = 0x00;
+		vub300->cmnd.head.transfer_size[0] = 0;
+		vub300->cmnd.head.transfer_size[1] = 0;
+		vub300->cmnd.head.transfer_size[2] = 0;
+		vub300->cmnd.head.transfer_size[3] = 0;
+	} else if (!data) {
+		vub300->cmnd.head.block_count[0] = 0;
+		vub300->cmnd.head.block_count[1] = 0;
+		vub300->cmnd.head.block_size[0] = (vub300->fbs[0] >> 8) & 0xFF;
+		vub300->cmnd.head.block_size[1] = (vub300->fbs[0] >> 0) & 0xFF;
+		vub300->cmnd.head.command_type = 0x00;
+		vub300->cmnd.head.transfer_size[0] = 0;
+		vub300->cmnd.head.transfer_size[1] = 0;
+		vub300->cmnd.head.transfer_size[2] = 0;
+		vub300->cmnd.head.transfer_size[3] = 0;
+	} else if (cmd->opcode == 53) {
+		int fn = 0x7 & (cmd->arg >> 28);
+		if (0x08 & vub300->cmnd.head.arguments[0]) { /* BLOCK MODE */
+			vub300->cmnd.head.block_count[0] =
+				(data->blocks >> 8) & 0xFF;
+			vub300->cmnd.head.block_count[1] =
+				(data->blocks >> 0) & 0xFF;
+			vub300->cmnd.head.block_size[0] =
+				(data->blksz >> 8) & 0xFF;
+			vub300->cmnd.head.block_size[1] =
+				(data->blksz >> 0) & 0xFF;
+		} else {	/* BYTE MODE */
+			vub300->cmnd.head.block_count[0] = 0;
+			vub300->cmnd.head.block_count[1] = 0;
+			vub300->cmnd.head.block_size[0] =
+				(vub300->datasize >> 8) & 0xFF;
+			vub300->cmnd.head.block_size[1] =
+				(vub300->datasize >> 0) & 0xFF;
+		}
+		vub300->cmnd.head.command_type =
+			(MMC_DATA_READ & data->flags) ? 0x00 : 0x80;
+		vub300->cmnd.head.transfer_size[0] =
+			(vub300->datasize >> 24) & 0xFF;
+		vub300->cmnd.head.transfer_size[1] =
+			(vub300->datasize >> 16) & 0xFF;
+		vub300->cmnd.head.transfer_size[2] =
+			(vub300->datasize >> 8) & 0xFF;
+		vub300->cmnd.head.transfer_size[3] =
+			(vub300->datasize >> 0) & 0xFF;
+		if (vub300->datasize < vub300->fbs[fn]) {
+			vub300->cmnd.head.block_count[0] = 0;
+			vub300->cmnd.head.block_count[1] = 0;
+		}
+	} else {
+		vub300->cmnd.head.block_count[0] = (data->blocks >> 8) & 0xFF;
+		vub300->cmnd.head.block_count[1] = (data->blocks >> 0) & 0xFF;
+		vub300->cmnd.head.block_size[0] = (data->blksz >> 8) & 0xFF;
+		vub300->cmnd.head.block_size[1] = (data->blksz >> 0) & 0xFF;
+		vub300->cmnd.head.command_type =
+			(MMC_DATA_READ & data->flags) ? 0x00 : 0x80;
+		vub300->cmnd.head.transfer_size[0] =
+			(vub300->datasize >> 24) & 0xFF;
+		vub300->cmnd.head.transfer_size[1] =
+			(vub300->datasize >> 16) & 0xFF;
+		vub300->cmnd.head.transfer_size[2] =
+			(vub300->datasize >> 8) & 0xFF;
+		vub300->cmnd.head.transfer_size[3] =
+			(vub300->datasize >> 0) & 0xFF;
+		if (vub300->datasize < vub300->fbs[0]) {
+			vub300->cmnd.head.block_count[0] = 0;
+			vub300->cmnd.head.block_count[1] = 0;
+		}
+	}
+	if (vub300->cmnd.head.block_size[0] || vub300->cmnd.head.block_size[1]) {
+		u16 block_size = vub300->cmnd.head.block_size[1] |
+			(vub300->cmnd.head.block_size[0] << 8);
+		u16 block_boundary = FIRMWARE_BLOCK_BOUNDARY -
+			(FIRMWARE_BLOCK_BOUNDARY % block_size);
+		vub300->cmnd.head.block_boundary[0] =
+			(block_boundary >> 8) & 0xFF;
+		vub300->cmnd.head.block_boundary[1] =
+			(block_boundary >> 0) & 0xFF;
+	} else {
+		vub300->cmnd.head.block_boundary[0] = 0;
+		vub300->cmnd.head.block_boundary[1] = 0;
+	}
+	usb_fill_bulk_urb(vub300->command_out_urb, vub300->udev,
+			  usb_sndbulkpipe(vub300->udev, vub300->cmnd_out_ep),
+			  &vub300->cmnd, sizeof(vub300->cmnd),
+			  command_out_completed, vub300);
+	retval = usb_submit_urb(vub300->command_out_urb, GFP_KERNEL);
+	if (retval < 0) {
+		cmd->error = retval;
+		complete(&vub300->command_complete);
+		return;
+	} else {
+		return;
+	}
+}
+
+/*
+ * timer callback runs in atomic mode
+ *       so it cannot call usb_kill_urb()
+ */
+static void vub300_sg_timed_out(struct timer_list *t)
+{
+	struct vub300_mmc_host *vub300 = from_timer(vub300, t,
+						    sg_transfer_timer);
+	vub300->usb_timed_out = 1;
+	usb_sg_cancel(&vub300->sg_request);
+	usb_unlink_urb(vub300->command_out_urb);
+	usb_unlink_urb(vub300->command_res_urb);
+}
+
+static u16 roundup_to_multiple_of_64(u16 number)
+{
+	return 0xFFC0 & (0x3F + number);
+}
+
+/*
+ * this is a separate function to solve the 80 column width restriction
+ */
+static void __download_offload_pseudocode(struct vub300_mmc_host *vub300,
+					  const struct firmware *fw)
+{
+	u8 register_count = 0;
+	u16 ts = 0;
+	u16 interrupt_size = 0;
+	const u8 *data = fw->data;
+	int size = fw->size;
+	u8 c;
+	dev_info(&vub300->udev->dev, "using %s for SDIO offload processing\n",
+		 vub300->vub_name);
+	do {
+		c = *data++;
+	} while (size-- && c); /* skip comment */
+	dev_info(&vub300->udev->dev, "using offload firmware %s %s\n", fw->data,
+		 vub300->vub_name);
+	if (size < 4) {
+		dev_err(&vub300->udev->dev,
+			"corrupt offload pseudocode in firmware %s\n",
+			vub300->vub_name);
+		strncpy(vub300->vub_name, "corrupt offload pseudocode",
+			sizeof(vub300->vub_name));
+		return;
+	}
+	interrupt_size += *data++;
+	size -= 1;
+	interrupt_size <<= 8;
+	interrupt_size += *data++;
+	size -= 1;
+	if (interrupt_size < size) {
+		u16 xfer_length = roundup_to_multiple_of_64(interrupt_size);
+		u8 *xfer_buffer = kmalloc(xfer_length, GFP_KERNEL);
+		if (xfer_buffer) {
+			int retval;
+			memcpy(xfer_buffer, data, interrupt_size);
+			memset(xfer_buffer + interrupt_size, 0,
+			       xfer_length - interrupt_size);
+			size -= interrupt_size;
+			data += interrupt_size;
+			retval =
+				usb_control_msg(vub300->udev,
+						usb_sndctrlpipe(vub300->udev, 0),
+						SET_INTERRUPT_PSEUDOCODE,
+						USB_DIR_OUT | USB_TYPE_VENDOR |
+						USB_RECIP_DEVICE, 0x0000, 0x0000,
+						xfer_buffer, xfer_length, 1000);
+			kfree(xfer_buffer);
+			if (retval < 0)
+				goto copy_error_message;
+		} else {
+			dev_err(&vub300->udev->dev,
+				"not enough memory for xfer buffer to send"
+				" INTERRUPT_PSEUDOCODE for %s %s\n", fw->data,
+				vub300->vub_name);
+			strncpy(vub300->vub_name,
+				"SDIO interrupt pseudocode download failed",
+				sizeof(vub300->vub_name));
+			return;
+		}
+	} else {
+		dev_err(&vub300->udev->dev,
+			"corrupt interrupt pseudocode in firmware %s %s\n",
+			fw->data, vub300->vub_name);
+		strncpy(vub300->vub_name, "corrupt interrupt pseudocode",
+			sizeof(vub300->vub_name));
+		return;
+	}
+	ts += *data++;
+	size -= 1;
+	ts <<= 8;
+	ts += *data++;
+	size -= 1;
+	if (ts < size) {
+		u16 xfer_length = roundup_to_multiple_of_64(ts);
+		u8 *xfer_buffer = kmalloc(xfer_length, GFP_KERNEL);
+		if (xfer_buffer) {
+			int retval;
+			memcpy(xfer_buffer, data, ts);
+			memset(xfer_buffer + ts, 0,
+			       xfer_length - ts);
+			size -= ts;
+			data += ts;
+			retval =
+				usb_control_msg(vub300->udev,
+						usb_sndctrlpipe(vub300->udev, 0),
+						SET_TRANSFER_PSEUDOCODE,
+						USB_DIR_OUT | USB_TYPE_VENDOR |
+						USB_RECIP_DEVICE, 0x0000, 0x0000,
+						xfer_buffer, xfer_length, 1000);
+			kfree(xfer_buffer);
+			if (retval < 0)
+				goto copy_error_message;
+		} else {
+			dev_err(&vub300->udev->dev,
+				"not enough memory for xfer buffer to send"
+				" TRANSFER_PSEUDOCODE for %s %s\n", fw->data,
+				vub300->vub_name);
+			strncpy(vub300->vub_name,
+				"SDIO transfer pseudocode download failed",
+				sizeof(vub300->vub_name));
+			return;
+		}
+	} else {
+		dev_err(&vub300->udev->dev,
+			"corrupt transfer pseudocode in firmware %s %s\n",
+			fw->data, vub300->vub_name);
+		strncpy(vub300->vub_name, "corrupt transfer pseudocode",
+			sizeof(vub300->vub_name));
+		return;
+	}
+	register_count += *data++;
+	size -= 1;
+	if (register_count * 4 == size) {
+		int I = vub300->dynamic_register_count = register_count;
+		int i = 0;
+		while (I--) {
+			unsigned int func_num = 0;
+			vub300->sdio_register[i].func_num = *data++;
+			size -= 1;
+			func_num += *data++;
+			size -= 1;
+			func_num <<= 8;
+			func_num += *data++;
+			size -= 1;
+			func_num <<= 8;
+			func_num += *data++;
+			size -= 1;
+			vub300->sdio_register[i].sdio_reg = func_num;
+			vub300->sdio_register[i].activate = 1;
+			vub300->sdio_register[i].prepared = 0;
+			i += 1;
+		}
+		dev_info(&vub300->udev->dev,
+			 "initialized %d dynamic pseudocode registers\n",
+			 vub300->dynamic_register_count);
+		return;
+	} else {
+		dev_err(&vub300->udev->dev,
+			"corrupt dynamic registers in firmware %s\n",
+			vub300->vub_name);
+		strncpy(vub300->vub_name, "corrupt dynamic registers",
+			sizeof(vub300->vub_name));
+		return;
+	}
+
+	return;
+
+copy_error_message:
+	strncpy(vub300->vub_name, "SDIO pseudocode download failed",
+		sizeof(vub300->vub_name));
+}
+
+/*
+ * if the binary containing the EMPTY PseudoCode can not be found
+ * vub300->vub_name is set anyway in order to prevent an automatic retry
+ */
+static void download_offload_pseudocode(struct vub300_mmc_host *vub300)
+{
+	struct mmc_card *card = vub300->mmc->card;
+	int sdio_funcs = card->sdio_funcs;
+	const struct firmware *fw = NULL;
+	int l = snprintf(vub300->vub_name, sizeof(vub300->vub_name),
+			 "vub_%04X%04X", card->cis.vendor, card->cis.device);
+	int n = 0;
+	int retval;
+	for (n = 0; n < sdio_funcs; n++) {
+		struct sdio_func *sf = card->sdio_func[n];
+		l += scnprintf(vub300->vub_name + l,
+			      sizeof(vub300->vub_name) - l, "_%04X%04X",
+			      sf->vendor, sf->device);
+	}
+	snprintf(vub300->vub_name + l, sizeof(vub300->vub_name) - l, ".bin");
+	dev_info(&vub300->udev->dev, "requesting offload firmware %s\n",
+		 vub300->vub_name);
+	retval = request_firmware(&fw, vub300->vub_name, &card->dev);
+	if (retval < 0) {
+		strncpy(vub300->vub_name, "vub_default.bin",
+			sizeof(vub300->vub_name));
+		retval = request_firmware(&fw, vub300->vub_name, &card->dev);
+		if (retval < 0) {
+			strncpy(vub300->vub_name,
+				"no SDIO offload firmware found",
+				sizeof(vub300->vub_name));
+		} else {
+			__download_offload_pseudocode(vub300, fw);
+			release_firmware(fw);
+		}
+	} else {
+		__download_offload_pseudocode(vub300, fw);
+		release_firmware(fw);
+	}
+}
+
+static void vub300_usb_bulk_msg_completion(struct urb *urb)
+{				/* urb completion handler - hardirq */
+	complete((struct completion *)urb->context);
+}
+
+static int vub300_usb_bulk_msg(struct vub300_mmc_host *vub300,
+			       unsigned int pipe, void *data, int len,
+			       int *actual_length, int timeout_msecs)
+{
+	/* cmd_mutex is held by vub300_cmndwork_thread */
+	struct usb_device *usb_dev = vub300->udev;
+	struct completion done;
+	int retval;
+	vub300->urb = usb_alloc_urb(0, GFP_KERNEL);
+	if (!vub300->urb)
+		return -ENOMEM;
+	usb_fill_bulk_urb(vub300->urb, usb_dev, pipe, data, len,
+			  vub300_usb_bulk_msg_completion, NULL);
+	init_completion(&done);
+	vub300->urb->context = &done;
+	vub300->urb->actual_length = 0;
+	retval = usb_submit_urb(vub300->urb, GFP_KERNEL);
+	if (unlikely(retval))
+		goto out;
+	if (!wait_for_completion_timeout
+	    (&done, msecs_to_jiffies(timeout_msecs))) {
+		retval = -ETIMEDOUT;
+		usb_kill_urb(vub300->urb);
+	} else {
+		retval = vub300->urb->status;
+	}
+out:
+	*actual_length = vub300->urb->actual_length;
+	usb_free_urb(vub300->urb);
+	vub300->urb = NULL;
+	return retval;
+}
+
+static int __command_read_data(struct vub300_mmc_host *vub300,
+			       struct mmc_command *cmd, struct mmc_data *data)
+{
+	/* cmd_mutex is held by vub300_cmndwork_thread */
+	int linear_length = vub300->datasize;
+	int padded_length = vub300->large_usb_packets ?
+		((511 + linear_length) >> 9) << 9 :
+		((63 + linear_length) >> 6) << 6;
+	if ((padded_length == linear_length) || !pad_input_to_usb_pkt) {
+		int result;
+		unsigned pipe;
+		pipe = usb_rcvbulkpipe(vub300->udev, vub300->data_inp_ep);
+		result = usb_sg_init(&vub300->sg_request, vub300->udev,
+				     pipe, 0, data->sg,
+				     data->sg_len, 0, GFP_KERNEL);
+		if (result < 0) {
+			usb_unlink_urb(vub300->command_out_urb);
+			usb_unlink_urb(vub300->command_res_urb);
+			cmd->error = result;
+			data->bytes_xfered = 0;
+			return 0;
+		} else {
+			vub300->sg_transfer_timer.expires =
+				jiffies + msecs_to_jiffies(2000 +
+						  (linear_length / 16384));
+			add_timer(&vub300->sg_transfer_timer);
+			usb_sg_wait(&vub300->sg_request);
+			del_timer(&vub300->sg_transfer_timer);
+			if (vub300->sg_request.status < 0) {
+				cmd->error = vub300->sg_request.status;
+				data->bytes_xfered = 0;
+				return 0;
+			} else {
+				data->bytes_xfered = vub300->datasize;
+				return linear_length;
+			}
+		}
+	} else {
+		u8 *buf = kmalloc(padded_length, GFP_KERNEL);
+		if (buf) {
+			int result;
+			unsigned pipe = usb_rcvbulkpipe(vub300->udev,
+							vub300->data_inp_ep);
+			int actual_length = 0;
+			result = vub300_usb_bulk_msg(vub300, pipe, buf,
+					     padded_length, &actual_length,
+					     2000 + (padded_length / 16384));
+			if (result < 0) {
+				cmd->error = result;
+				data->bytes_xfered = 0;
+				kfree(buf);
+				return 0;
+			} else if (actual_length < linear_length) {
+				cmd->error = -EREMOTEIO;
+				data->bytes_xfered = 0;
+				kfree(buf);
+				return 0;
+			} else {
+				sg_copy_from_buffer(data->sg, data->sg_len, buf,
+						    linear_length);
+				kfree(buf);
+				data->bytes_xfered = vub300->datasize;
+				return linear_length;
+			}
+		} else {
+			cmd->error = -ENOMEM;
+			data->bytes_xfered = 0;
+			return 0;
+		}
+	}
+}
+
+static int __command_write_data(struct vub300_mmc_host *vub300,
+				struct mmc_command *cmd, struct mmc_data *data)
+{
+	/* cmd_mutex is held by vub300_cmndwork_thread */
+	unsigned pipe = usb_sndbulkpipe(vub300->udev, vub300->data_out_ep);
+	int linear_length = vub300->datasize;
+	int modulo_64_length = linear_length & 0x003F;
+	int modulo_512_length = linear_length & 0x01FF;
+	if (linear_length < 64) {
+		int result;
+		int actual_length;
+		sg_copy_to_buffer(data->sg, data->sg_len,
+				  vub300->padded_buffer,
+				  sizeof(vub300->padded_buffer));
+		memset(vub300->padded_buffer + linear_length, 0,
+		       sizeof(vub300->padded_buffer) - linear_length);
+		result = vub300_usb_bulk_msg(vub300, pipe, vub300->padded_buffer,
+					     sizeof(vub300->padded_buffer),
+					     &actual_length, 2000 +
+					     (sizeof(vub300->padded_buffer) /
+					      16384));
+		if (result < 0) {
+			cmd->error = result;
+			data->bytes_xfered = 0;
+		} else {
+			data->bytes_xfered = vub300->datasize;
+		}
+	} else if ((!vub300->large_usb_packets && (0 < modulo_64_length)) ||
+		    (vub300->large_usb_packets && (64 > modulo_512_length))
+		) {		/* don't you just love these work-rounds */
+		int padded_length = ((63 + linear_length) >> 6) << 6;
+		u8 *buf = kmalloc(padded_length, GFP_KERNEL);
+		if (buf) {
+			int result;
+			int actual_length;
+			sg_copy_to_buffer(data->sg, data->sg_len, buf,
+					  padded_length);
+			memset(buf + linear_length, 0,
+			       padded_length - linear_length);
+			result =
+				vub300_usb_bulk_msg(vub300, pipe, buf,
+						    padded_length, &actual_length,
+						    2000 + padded_length / 16384);
+			kfree(buf);
+			if (result < 0) {
+				cmd->error = result;
+				data->bytes_xfered = 0;
+			} else {
+				data->bytes_xfered = vub300->datasize;
+			}
+		} else {
+			cmd->error = -ENOMEM;
+			data->bytes_xfered = 0;
+		}
+	} else {		/* no data padding required */
+		int result;
+		unsigned char buf[64 * 4];
+		sg_copy_to_buffer(data->sg, data->sg_len, buf, sizeof(buf));
+		result = usb_sg_init(&vub300->sg_request, vub300->udev,
+				     pipe, 0, data->sg,
+				     data->sg_len, 0, GFP_KERNEL);
+		if (result < 0) {
+			usb_unlink_urb(vub300->command_out_urb);
+			usb_unlink_urb(vub300->command_res_urb);
+			cmd->error = result;
+			data->bytes_xfered = 0;
+		} else {
+			vub300->sg_transfer_timer.expires =
+				jiffies + msecs_to_jiffies(2000 +
+							   linear_length / 16384);
+			add_timer(&vub300->sg_transfer_timer);
+			usb_sg_wait(&vub300->sg_request);
+			if (cmd->error) {
+				data->bytes_xfered = 0;
+			} else {
+				del_timer(&vub300->sg_transfer_timer);
+				if (vub300->sg_request.status < 0) {
+					cmd->error = vub300->sg_request.status;
+					data->bytes_xfered = 0;
+				} else {
+					data->bytes_xfered = vub300->datasize;
+				}
+			}
+		}
+	}
+	return linear_length;
+}
+
+static void __vub300_command_response(struct vub300_mmc_host *vub300,
+				      struct mmc_command *cmd,
+				      struct mmc_data *data, int data_length)
+{
+	/* cmd_mutex is held by vub300_cmndwork_thread */
+	long respretval;
+	int msec_timeout = 1000 + data_length / 4;
+	respretval =
+		wait_for_completion_timeout(&vub300->command_complete,
+					    msecs_to_jiffies(msec_timeout));
+	if (respretval == 0) { /* TIMED OUT */
+		/* we don't know which of "out" and "res" if any failed */
+		int result;
+		vub300->usb_timed_out = 1;
+		usb_kill_urb(vub300->command_out_urb);
+		usb_kill_urb(vub300->command_res_urb);
+		cmd->error = -ETIMEDOUT;
+		result = usb_lock_device_for_reset(vub300->udev,
+						   vub300->interface);
+		if (result == 0) {
+			result = usb_reset_device(vub300->udev);
+			usb_unlock_device(vub300->udev);
+		}
+	} else if (respretval < 0) {
+		/* we don't know which of "out" and "res" if any failed */
+		usb_kill_urb(vub300->command_out_urb);
+		usb_kill_urb(vub300->command_res_urb);
+		cmd->error = respretval;
+	} else if (cmd->error) {
+		/*
+		 * the error occurred sending the command
+		 * or receiving the response
+		 */
+	} else if (vub300->command_out_urb->status) {
+		vub300->usb_transport_fail = vub300->command_out_urb->status;
+		cmd->error = -EPROTO == vub300->command_out_urb->status ?
+			-ESHUTDOWN : vub300->command_out_urb->status;
+	} else if (vub300->command_res_urb->status) {
+		vub300->usb_transport_fail = vub300->command_res_urb->status;
+		cmd->error = -EPROTO == vub300->command_res_urb->status ?
+			-ESHUTDOWN : vub300->command_res_urb->status;
+	} else if (vub300->resp.common.header_type == 0x00) {
+		/*
+		 * the command completed successfully
+		 * and there was no piggybacked data
+		 */
+	} else if (vub300->resp.common.header_type == RESPONSE_ERROR) {
+		cmd->error =
+			vub300_response_error(vub300->resp.error.error_code);
+		if (vub300->data)
+			usb_sg_cancel(&vub300->sg_request);
+	} else if (vub300->resp.common.header_type == RESPONSE_PIGGYBACKED) {
+		int offloaded_data_length =
+			vub300->resp.common.header_size -
+			sizeof(struct sd_register_header);
+		int register_count = offloaded_data_length >> 3;
+		int ri = 0;
+		while (register_count--) {
+			add_offloaded_reg(vub300, &vub300->resp.pig.reg[ri]);
+			ri += 1;
+		}
+		vub300->resp.common.header_size =
+			sizeof(struct sd_register_header);
+		vub300->resp.common.header_type = 0x00;
+		cmd->error = 0;
+	} else if (vub300->resp.common.header_type == RESPONSE_PIG_DISABLED) {
+		int offloaded_data_length =
+			vub300->resp.common.header_size -
+			sizeof(struct sd_register_header);
+		int register_count = offloaded_data_length >> 3;
+		int ri = 0;
+		while (register_count--) {
+			add_offloaded_reg(vub300, &vub300->resp.pig.reg[ri]);
+			ri += 1;
+		}
+		mutex_lock(&vub300->irq_mutex);
+		if (vub300->irqs_queued) {
+			vub300->irqs_queued += 1;
+		} else if (vub300->irq_enabled) {
+			vub300->irqs_queued += 1;
+			vub300_queue_poll_work(vub300, 0);
+		} else {
+			vub300->irqs_queued += 1;
+		}
+		vub300->irq_disabled = 1;
+		mutex_unlock(&vub300->irq_mutex);
+		vub300->resp.common.header_size =
+			sizeof(struct sd_register_header);
+		vub300->resp.common.header_type = 0x00;
+		cmd->error = 0;
+	} else if (vub300->resp.common.header_type == RESPONSE_PIG_ENABLED) {
+		int offloaded_data_length =
+			vub300->resp.common.header_size -
+			sizeof(struct sd_register_header);
+		int register_count = offloaded_data_length >> 3;
+		int ri = 0;
+		while (register_count--) {
+			add_offloaded_reg(vub300, &vub300->resp.pig.reg[ri]);
+			ri += 1;
+		}
+		mutex_lock(&vub300->irq_mutex);
+		if (vub300->irqs_queued) {
+			vub300->irqs_queued += 1;
+		} else if (vub300->irq_enabled) {
+			vub300->irqs_queued += 1;
+			vub300_queue_poll_work(vub300, 0);
+		} else {
+			vub300->irqs_queued += 1;
+		}
+		vub300->irq_disabled = 0;
+		mutex_unlock(&vub300->irq_mutex);
+		vub300->resp.common.header_size =
+			sizeof(struct sd_register_header);
+		vub300->resp.common.header_type = 0x00;
+		cmd->error = 0;
+	} else {
+		cmd->error = -EINVAL;
+	}
+}
+
+static void construct_request_response(struct vub300_mmc_host *vub300,
+				       struct mmc_command *cmd)
+{
+	int resp_len = vub300->resp_len;
+	int less_cmd = (17 == resp_len) ? resp_len : resp_len - 1;
+	int bytes = 3 & less_cmd;
+	int words = less_cmd >> 2;
+	u8 *r = vub300->resp.response.command_response;
+
+	if (!resp_len)
+		return;
+	if (bytes == 3) {
+		cmd->resp[words] = (r[1 + (words << 2)] << 24)
+			| (r[2 + (words << 2)] << 16)
+			| (r[3 + (words << 2)] << 8);
+	} else if (bytes == 2) {
+		cmd->resp[words] = (r[1 + (words << 2)] << 24)
+			| (r[2 + (words << 2)] << 16);
+	} else if (bytes == 1) {
+		cmd->resp[words] = (r[1 + (words << 2)] << 24);
+	}
+	while (words-- > 0) {
+		cmd->resp[words] = (r[1 + (words << 2)] << 24)
+			| (r[2 + (words << 2)] << 16)
+			| (r[3 + (words << 2)] << 8)
+			| (r[4 + (words << 2)] << 0);
+	}
+	if ((cmd->opcode == 53) && (0x000000FF & cmd->resp[0]))
+		cmd->resp[0] &= 0xFFFFFF00;
+}
+
+/* this thread runs only when there is an upper level command req outstanding */
+static void vub300_cmndwork_thread(struct work_struct *work)
+{
+	struct vub300_mmc_host *vub300 =
+		container_of(work, struct vub300_mmc_host, cmndwork);
+	if (!vub300->interface) {
+		kref_put(&vub300->kref, vub300_delete);
+		return;
+	} else {
+		struct mmc_request *req = vub300->req;
+		struct mmc_command *cmd = vub300->cmd;
+		struct mmc_data *data = vub300->data;
+		int data_length;
+		mutex_lock(&vub300->cmd_mutex);
+		init_completion(&vub300->command_complete);
+		if (likely(vub300->vub_name[0]) || !vub300->mmc->card) {
+			/*
+			 * the name of the EMPTY Pseudo firmware file
+			 * is used as a flag to indicate that the file
+			 * has been already downloaded to the VUB300 chip
+			 */
+		} else if (0 == vub300->mmc->card->sdio_funcs) {
+			strncpy(vub300->vub_name, "SD memory device",
+				sizeof(vub300->vub_name));
+		} else {
+			download_offload_pseudocode(vub300);
+		}
+		send_command(vub300);
+		if (!data)
+			data_length = 0;
+		else if (MMC_DATA_READ & data->flags)
+			data_length = __command_read_data(vub300, cmd, data);
+		else
+			data_length = __command_write_data(vub300, cmd, data);
+		__vub300_command_response(vub300, cmd, data, data_length);
+		vub300->req = NULL;
+		vub300->cmd = NULL;
+		vub300->data = NULL;
+		if (cmd->error) {
+			if (cmd->error == -ENOMEDIUM)
+				check_vub300_port_status(vub300);
+			mutex_unlock(&vub300->cmd_mutex);
+			mmc_request_done(vub300->mmc, req);
+			kref_put(&vub300->kref, vub300_delete);
+			return;
+		} else {
+			construct_request_response(vub300, cmd);
+			vub300->resp_len = 0;
+			mutex_unlock(&vub300->cmd_mutex);
+			kref_put(&vub300->kref, vub300_delete);
+			mmc_request_done(vub300->mmc, req);
+			return;
+		}
+	}
+}
+
+static int examine_cyclic_buffer(struct vub300_mmc_host *vub300,
+				 struct mmc_command *cmd, u8 Function)
+{
+	/* cmd_mutex is held by vub300_mmc_request */
+	u8 cmd0 = 0xFF & (cmd->arg >> 24);
+	u8 cmd1 = 0xFF & (cmd->arg >> 16);
+	u8 cmd2 = 0xFF & (cmd->arg >> 8);
+	u8 cmd3 = 0xFF & (cmd->arg >> 0);
+	int first = MAXREGMASK & vub300->fn[Function].offload_point;
+	struct offload_registers_access *rf = &vub300->fn[Function].reg[first];
+	if (cmd0 == rf->command_byte[0] &&
+	    cmd1 == rf->command_byte[1] &&
+	    cmd2 == rf->command_byte[2] &&
+	    cmd3 == rf->command_byte[3]) {
+		u8 checksum = 0x00;
+		cmd->resp[1] = checksum << 24;
+		cmd->resp[0] = (rf->Respond_Byte[0] << 24)
+			| (rf->Respond_Byte[1] << 16)
+			| (rf->Respond_Byte[2] << 8)
+			| (rf->Respond_Byte[3] << 0);
+		vub300->fn[Function].offload_point += 1;
+		vub300->fn[Function].offload_count -= 1;
+		vub300->total_offload_count -= 1;
+		return 1;
+	} else {
+		int delta = 1;	/* because it does not match the first one */
+		u8 register_count = vub300->fn[Function].offload_count - 1;
+		u32 register_point = vub300->fn[Function].offload_point + 1;
+		while (0 < register_count) {
+			int point = MAXREGMASK & register_point;
+			struct offload_registers_access *r =
+				&vub300->fn[Function].reg[point];
+			if (cmd0 == r->command_byte[0] &&
+			    cmd1 == r->command_byte[1] &&
+			    cmd2 == r->command_byte[2] &&
+			    cmd3 == r->command_byte[3]) {
+				u8 checksum = 0x00;
+				cmd->resp[1] = checksum << 24;
+				cmd->resp[0] = (r->Respond_Byte[0] << 24)
+					| (r->Respond_Byte[1] << 16)
+					| (r->Respond_Byte[2] << 8)
+					| (r->Respond_Byte[3] << 0);
+				vub300->fn[Function].offload_point += delta;
+				vub300->fn[Function].offload_count -= delta;
+				vub300->total_offload_count -= delta;
+				return 1;
+			} else {
+				register_point += 1;
+				register_count -= 1;
+				delta += 1;
+				continue;
+			}
+		}
+		return 0;
+	}
+}
+
+static int satisfy_request_from_offloaded_data(struct vub300_mmc_host *vub300,
+					       struct mmc_command *cmd)
+{
+	/* cmd_mutex is held by vub300_mmc_request */
+	u8 regs = vub300->dynamic_register_count;
+	u8 i = 0;
+	u8 func = FUN(cmd);
+	u32 reg = REG(cmd);
+	while (0 < regs--) {
+		if ((vub300->sdio_register[i].func_num == func) &&
+		    (vub300->sdio_register[i].sdio_reg == reg)) {
+			if (!vub300->sdio_register[i].prepared) {
+				return 0;
+			} else if ((0x80000000 & cmd->arg) == 0x80000000) {
+				/*
+				 * a write to a dynamic register
+				 * nullifies our offloaded value
+				 */
+				vub300->sdio_register[i].prepared = 0;
+				return 0;
+			} else {
+				u8 checksum = 0x00;
+				u8 rsp0 = 0x00;
+				u8 rsp1 = 0x00;
+				u8 rsp2 = vub300->sdio_register[i].response;
+				u8 rsp3 = vub300->sdio_register[i].regvalue;
+				vub300->sdio_register[i].prepared = 0;
+				cmd->resp[1] = checksum << 24;
+				cmd->resp[0] = (rsp0 << 24)
+					| (rsp1 << 16)
+					| (rsp2 << 8)
+					| (rsp3 << 0);
+				return 1;
+			}
+		} else {
+			i += 1;
+			continue;
+		}
+	}
+	if (vub300->total_offload_count == 0)
+		return 0;
+	else if (vub300->fn[func].offload_count == 0)
+		return 0;
+	else
+		return examine_cyclic_buffer(vub300, cmd, func);
+}
+
+static void vub300_mmc_request(struct mmc_host *mmc, struct mmc_request *req)
+{				/* NOT irq */
+	struct mmc_command *cmd = req->cmd;
+	struct vub300_mmc_host *vub300 = mmc_priv(mmc);
+	if (!vub300->interface) {
+		cmd->error = -ESHUTDOWN;
+		mmc_request_done(mmc, req);
+		return;
+	} else {
+		struct mmc_data *data = req->data;
+		if (!vub300->card_powered) {
+			cmd->error = -ENOMEDIUM;
+			mmc_request_done(mmc, req);
+			return;
+		}
+		if (!vub300->card_present) {
+			cmd->error = -ENOMEDIUM;
+			mmc_request_done(mmc, req);
+			return;
+		}
+		if (vub300->usb_transport_fail) {
+			cmd->error = vub300->usb_transport_fail;
+			mmc_request_done(mmc, req);
+			return;
+		}
+		if (!vub300->interface) {
+			cmd->error = -ENODEV;
+			mmc_request_done(mmc, req);
+			return;
+		}
+		kref_get(&vub300->kref);
+		mutex_lock(&vub300->cmd_mutex);
+		mod_timer(&vub300->inactivity_timer, jiffies + HZ);
+		/*
+		 * for performance we have to return immediately
+		 * if the requested data has been offloaded
+		 */
+		if (cmd->opcode == 52 &&
+		    satisfy_request_from_offloaded_data(vub300, cmd)) {
+			cmd->error = 0;
+			mutex_unlock(&vub300->cmd_mutex);
+			kref_put(&vub300->kref, vub300_delete);
+			mmc_request_done(mmc, req);
+			return;
+		} else {
+			vub300->cmd = cmd;
+			vub300->req = req;
+			vub300->data = data;
+			if (data)
+				vub300->datasize = data->blksz * data->blocks;
+			else
+				vub300->datasize = 0;
+			vub300_queue_cmnd_work(vub300);
+			mutex_unlock(&vub300->cmd_mutex);
+			kref_put(&vub300->kref, vub300_delete);
+			/*
+			 * the kernel lock diagnostics complain
+			 * if the cmd_mutex * is "passed on"
+			 * to the cmndwork thread,
+			 * so we must release it now
+			 * and re-acquire it in the cmndwork thread
+			 */
+		}
+	}
+}
+
+static void __set_clock_speed(struct vub300_mmc_host *vub300, u8 buf[8],
+			      struct mmc_ios *ios)
+{
+	int buf_array_size = 8; /* ARRAY_SIZE(buf) does not work !!! */
+	int retval;
+	u32 kHzClock;
+	if (ios->clock >= 48000000)
+		kHzClock = 48000;
+	else if (ios->clock >= 24000000)
+		kHzClock = 24000;
+	else if (ios->clock >= 20000000)
+		kHzClock = 20000;
+	else if (ios->clock >= 15000000)
+		kHzClock = 15000;
+	else if (ios->clock >= 200000)
+		kHzClock = 200;
+	else
+		kHzClock = 0;
+	{
+		int i;
+		u64 c = kHzClock;
+		for (i = 0; i < buf_array_size; i++) {
+			buf[i] = c;
+			c >>= 8;
+		}
+	}
+	retval =
+		usb_control_msg(vub300->udev, usb_sndctrlpipe(vub300->udev, 0),
+				SET_CLOCK_SPEED,
+				USB_DIR_OUT | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+				0x00, 0x00, buf, buf_array_size, 1000);
+	if (retval != 8) {
+		dev_err(&vub300->udev->dev, "SET_CLOCK_SPEED"
+			" %dkHz failed with retval=%d\n", kHzClock, retval);
+	} else {
+		dev_dbg(&vub300->udev->dev, "SET_CLOCK_SPEED"
+			" %dkHz\n", kHzClock);
+	}
+}
+
+static void vub300_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{				/* NOT irq */
+	struct vub300_mmc_host *vub300 = mmc_priv(mmc);
+	if (!vub300->interface)
+		return;
+	kref_get(&vub300->kref);
+	mutex_lock(&vub300->cmd_mutex);
+	if ((ios->power_mode == MMC_POWER_OFF) && vub300->card_powered) {
+		vub300->card_powered = 0;
+		usb_control_msg(vub300->udev, usb_sndctrlpipe(vub300->udev, 0),
+				SET_SD_POWER,
+				USB_DIR_OUT | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+				0x0000, 0x0000, NULL, 0, 1000);
+		/* must wait for the VUB300 u-proc to boot up */
+		msleep(600);
+	} else if ((ios->power_mode == MMC_POWER_UP) && !vub300->card_powered) {
+		usb_control_msg(vub300->udev, usb_sndctrlpipe(vub300->udev, 0),
+				SET_SD_POWER,
+				USB_DIR_OUT | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+				0x0001, 0x0000, NULL, 0, 1000);
+		msleep(600);
+		vub300->card_powered = 1;
+	} else if (ios->power_mode == MMC_POWER_ON) {
+		u8 *buf = kmalloc(8, GFP_KERNEL);
+		if (buf) {
+			__set_clock_speed(vub300, buf, ios);
+			kfree(buf);
+		}
+	} else {
+		/* this should mean no change of state */
+	}
+	mutex_unlock(&vub300->cmd_mutex);
+	kref_put(&vub300->kref, vub300_delete);
+}
+
+static int vub300_mmc_get_ro(struct mmc_host *mmc)
+{
+	struct vub300_mmc_host *vub300 = mmc_priv(mmc);
+	return vub300->read_only;
+}
+
+static void vub300_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{				/* NOT irq */
+	struct vub300_mmc_host *vub300 = mmc_priv(mmc);
+	if (!vub300->interface)
+		return;
+	kref_get(&vub300->kref);
+	if (enable) {
+		set_current_state(TASK_RUNNING);
+		mutex_lock(&vub300->irq_mutex);
+		if (vub300->irqs_queued) {
+			vub300->irqs_queued -= 1;
+			mmc_signal_sdio_irq(vub300->mmc);
+		} else if (vub300->irq_disabled) {
+			vub300->irq_disabled = 0;
+			vub300->irq_enabled = 1;
+			vub300_queue_poll_work(vub300, 0);
+		} else if (vub300->irq_enabled) {
+			/* this should not happen, so we will just ignore it */
+		} else {
+			vub300->irq_enabled = 1;
+			vub300_queue_poll_work(vub300, 0);
+		}
+		mutex_unlock(&vub300->irq_mutex);
+		set_current_state(TASK_INTERRUPTIBLE);
+	} else {
+		vub300->irq_enabled = 0;
+	}
+	kref_put(&vub300->kref, vub300_delete);
+}
+
+static const struct mmc_host_ops vub300_mmc_ops = {
+	.request = vub300_mmc_request,
+	.set_ios = vub300_mmc_set_ios,
+	.get_ro = vub300_mmc_get_ro,
+	.enable_sdio_irq = vub300_enable_sdio_irq,
+};
+
+static int vub300_probe(struct usb_interface *interface,
+			const struct usb_device_id *id)
+{				/* NOT irq */
+	struct vub300_mmc_host *vub300;
+	struct usb_host_interface *iface_desc;
+	struct usb_device *udev = usb_get_dev(interface_to_usbdev(interface));
+	int i;
+	int retval = -ENOMEM;
+	struct urb *command_out_urb;
+	struct urb *command_res_urb;
+	struct mmc_host *mmc;
+	char manufacturer[48];
+	char product[32];
+	char serial_number[32];
+	usb_string(udev, udev->descriptor.iManufacturer, manufacturer,
+		   sizeof(manufacturer));
+	usb_string(udev, udev->descriptor.iProduct, product, sizeof(product));
+	usb_string(udev, udev->descriptor.iSerialNumber, serial_number,
+		   sizeof(serial_number));
+	dev_info(&udev->dev, "probing VID:PID(%04X:%04X) %s %s %s\n",
+		 le16_to_cpu(udev->descriptor.idVendor),
+		 le16_to_cpu(udev->descriptor.idProduct),
+		 manufacturer, product, serial_number);
+	command_out_urb = usb_alloc_urb(0, GFP_KERNEL);
+	if (!command_out_urb) {
+		retval = -ENOMEM;
+		goto error0;
+	}
+	command_res_urb = usb_alloc_urb(0, GFP_KERNEL);
+	if (!command_res_urb) {
+		retval = -ENOMEM;
+		goto error1;
+	}
+	/* this also allocates memory for our VUB300 mmc host device */
+	mmc = mmc_alloc_host(sizeof(struct vub300_mmc_host), &udev->dev);
+	if (!mmc) {
+		retval = -ENOMEM;
+		dev_err(&udev->dev, "not enough memory for the mmc_host\n");
+		goto error4;
+	}
+	/* MMC core transfer sizes tunable parameters */
+	mmc->caps = 0;
+	if (!force_1_bit_data_xfers)
+		mmc->caps |= MMC_CAP_4_BIT_DATA;
+	if (!force_polling_for_irqs)
+		mmc->caps |= MMC_CAP_SDIO_IRQ;
+	mmc->caps &= ~MMC_CAP_NEEDS_POLL;
+	/*
+	 * MMC_CAP_NEEDS_POLL causes core.c:mmc_rescan() to poll
+	 * for devices which results in spurious CMD7's being
+	 * issued which stops some SDIO cards from working
+	 */
+	if (limit_speed_to_24_MHz) {
+		mmc->caps |= MMC_CAP_MMC_HIGHSPEED;
+		mmc->caps |= MMC_CAP_SD_HIGHSPEED;
+		mmc->f_max = 24000000;
+		dev_info(&udev->dev, "limiting SDIO speed to 24_MHz\n");
+	} else {
+		mmc->caps |= MMC_CAP_MMC_HIGHSPEED;
+		mmc->caps |= MMC_CAP_SD_HIGHSPEED;
+		mmc->f_max = 48000000;
+	}
+	mmc->f_min = 200000;
+	mmc->max_blk_count = 511;
+	mmc->max_blk_size = 512;
+	mmc->max_segs = 128;
+	if (force_max_req_size)
+		mmc->max_req_size = force_max_req_size * 1024;
+	else
+		mmc->max_req_size = 64 * 1024;
+	mmc->max_seg_size = mmc->max_req_size;
+	mmc->ocr_avail = 0;
+	mmc->ocr_avail |= MMC_VDD_165_195;
+	mmc->ocr_avail |= MMC_VDD_20_21;
+	mmc->ocr_avail |= MMC_VDD_21_22;
+	mmc->ocr_avail |= MMC_VDD_22_23;
+	mmc->ocr_avail |= MMC_VDD_23_24;
+	mmc->ocr_avail |= MMC_VDD_24_25;
+	mmc->ocr_avail |= MMC_VDD_25_26;
+	mmc->ocr_avail |= MMC_VDD_26_27;
+	mmc->ocr_avail |= MMC_VDD_27_28;
+	mmc->ocr_avail |= MMC_VDD_28_29;
+	mmc->ocr_avail |= MMC_VDD_29_30;
+	mmc->ocr_avail |= MMC_VDD_30_31;
+	mmc->ocr_avail |= MMC_VDD_31_32;
+	mmc->ocr_avail |= MMC_VDD_32_33;
+	mmc->ocr_avail |= MMC_VDD_33_34;
+	mmc->ocr_avail |= MMC_VDD_34_35;
+	mmc->ocr_avail |= MMC_VDD_35_36;
+	mmc->ops = &vub300_mmc_ops;
+	vub300 = mmc_priv(mmc);
+	vub300->mmc = mmc;
+	vub300->card_powered = 0;
+	vub300->bus_width = 0;
+	vub300->cmnd.head.block_size[0] = 0x00;
+	vub300->cmnd.head.block_size[1] = 0x00;
+	vub300->app_spec = 0;
+	mutex_init(&vub300->cmd_mutex);
+	mutex_init(&vub300->irq_mutex);
+	vub300->command_out_urb = command_out_urb;
+	vub300->command_res_urb = command_res_urb;
+	vub300->usb_timed_out = 0;
+	vub300->dynamic_register_count = 0;
+
+	for (i = 0; i < ARRAY_SIZE(vub300->fn); i++) {
+		vub300->fn[i].offload_point = 0;
+		vub300->fn[i].offload_count = 0;
+	}
+
+	vub300->total_offload_count = 0;
+	vub300->irq_enabled = 0;
+	vub300->irq_disabled = 0;
+	vub300->irqs_queued = 0;
+
+	for (i = 0; i < ARRAY_SIZE(vub300->sdio_register); i++)
+		vub300->sdio_register[i++].activate = 0;
+
+	vub300->udev = udev;
+	vub300->interface = interface;
+	vub300->cmnd_res_ep = 0;
+	vub300->cmnd_out_ep = 0;
+	vub300->data_inp_ep = 0;
+	vub300->data_out_ep = 0;
+
+	for (i = 0; i < ARRAY_SIZE(vub300->fbs); i++)
+		vub300->fbs[i] = 512;
+
+	/*
+	 *      set up the endpoint information
+	 *
+	 * use the first pair of bulk-in and bulk-out
+	 *     endpoints for Command/Response+Interrupt
+	 *
+	 * use the second pair of bulk-in and bulk-out
+	 *     endpoints for Data In/Out
+	 */
+	vub300->large_usb_packets = 0;
+	iface_desc = interface->cur_altsetting;
+	for (i = 0; i < iface_desc->desc.bNumEndpoints; ++i) {
+		struct usb_endpoint_descriptor *endpoint =
+			&iface_desc->endpoint[i].desc;
+		dev_info(&vub300->udev->dev,
+			 "vub300 testing %s EndPoint(%d) %02X\n",
+			 usb_endpoint_is_bulk_in(endpoint) ? "BULK IN" :
+			 usb_endpoint_is_bulk_out(endpoint) ? "BULK OUT" :
+			 "UNKNOWN", i, endpoint->bEndpointAddress);
+		if (endpoint->wMaxPacketSize > 64)
+			vub300->large_usb_packets = 1;
+		if (usb_endpoint_is_bulk_in(endpoint)) {
+			if (!vub300->cmnd_res_ep) {
+				vub300->cmnd_res_ep =
+					endpoint->bEndpointAddress;
+			} else if (!vub300->data_inp_ep) {
+				vub300->data_inp_ep =
+					endpoint->bEndpointAddress;
+			} else {
+				dev_warn(&vub300->udev->dev,
+					 "ignoring"
+					 " unexpected bulk_in endpoint");
+			}
+		} else if (usb_endpoint_is_bulk_out(endpoint)) {
+			if (!vub300->cmnd_out_ep) {
+				vub300->cmnd_out_ep =
+					endpoint->bEndpointAddress;
+			} else if (!vub300->data_out_ep) {
+				vub300->data_out_ep =
+					endpoint->bEndpointAddress;
+			} else {
+				dev_warn(&vub300->udev->dev,
+					 "ignoring"
+					 " unexpected bulk_out endpoint");
+			}
+		} else {
+			dev_warn(&vub300->udev->dev,
+				 "vub300 ignoring EndPoint(%d) %02X", i,
+				 endpoint->bEndpointAddress);
+		}
+	}
+	if (vub300->cmnd_res_ep && vub300->cmnd_out_ep &&
+	    vub300->data_inp_ep && vub300->data_out_ep) {
+		dev_info(&vub300->udev->dev,
+			 "vub300 %s packets"
+			 " using EndPoints %02X %02X %02X %02X\n",
+			 vub300->large_usb_packets ? "LARGE" : "SMALL",
+			 vub300->cmnd_out_ep, vub300->cmnd_res_ep,
+			 vub300->data_out_ep, vub300->data_inp_ep);
+		/* we have the expected EndPoints */
+	} else {
+		dev_err(&vub300->udev->dev,
+		    "Could not find two sets of bulk-in/out endpoint pairs\n");
+		retval = -EINVAL;
+		goto error5;
+	}
+	retval =
+		usb_control_msg(vub300->udev, usb_rcvctrlpipe(vub300->udev, 0),
+				GET_HC_INF0,
+				USB_DIR_IN | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+				0x0000, 0x0000, &vub300->hc_info,
+				sizeof(vub300->hc_info), 1000);
+	if (retval < 0)
+		goto error5;
+	retval =
+		usb_control_msg(vub300->udev, usb_sndctrlpipe(vub300->udev, 0),
+				SET_ROM_WAIT_STATES,
+				USB_DIR_OUT | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+				firmware_rom_wait_states, 0x0000, NULL, 0, 1000);
+	if (retval < 0)
+		goto error5;
+	dev_info(&vub300->udev->dev,
+		 "operating_mode = %s %s %d MHz %s %d byte USB packets\n",
+		 (mmc->caps & MMC_CAP_SDIO_IRQ) ? "IRQs" : "POLL",
+		 (mmc->caps & MMC_CAP_4_BIT_DATA) ? "4-bit" : "1-bit",
+		 mmc->f_max / 1000000,
+		 pad_input_to_usb_pkt ? "padding input data to" : "with",
+		 vub300->large_usb_packets ? 512 : 64);
+	retval =
+		usb_control_msg(vub300->udev, usb_rcvctrlpipe(vub300->udev, 0),
+				GET_SYSTEM_PORT_STATUS,
+				USB_DIR_IN | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+				0x0000, 0x0000, &vub300->system_port_status,
+				sizeof(vub300->system_port_status), 1000);
+	if (retval < 0) {
+		goto error5;
+	} else if (sizeof(vub300->system_port_status) == retval) {
+		vub300->card_present =
+			(0x0001 & vub300->system_port_status.port_flags) ? 1 : 0;
+		vub300->read_only =
+			(0x0010 & vub300->system_port_status.port_flags) ? 1 : 0;
+	} else {
+		retval = -EINVAL;
+		goto error5;
+	}
+	usb_set_intfdata(interface, vub300);
+	INIT_DELAYED_WORK(&vub300->pollwork, vub300_pollwork_thread);
+	INIT_WORK(&vub300->cmndwork, vub300_cmndwork_thread);
+	INIT_WORK(&vub300->deadwork, vub300_deadwork_thread);
+	kref_init(&vub300->kref);
+	timer_setup(&vub300->sg_transfer_timer, vub300_sg_timed_out, 0);
+	kref_get(&vub300->kref);
+	timer_setup(&vub300->inactivity_timer,
+		    vub300_inactivity_timer_expired, 0);
+	vub300->inactivity_timer.expires = jiffies + HZ;
+	add_timer(&vub300->inactivity_timer);
+	if (vub300->card_present)
+		dev_info(&vub300->udev->dev,
+			 "USB vub300 remote SDIO host controller[%d]"
+			 "connected with SD/SDIO card inserted\n",
+			 interface_to_InterfaceNumber(interface));
+	else
+		dev_info(&vub300->udev->dev,
+			 "USB vub300 remote SDIO host controller[%d]"
+			 "connected with no SD/SDIO card inserted\n",
+			 interface_to_InterfaceNumber(interface));
+	retval = mmc_add_host(mmc);
+	if (retval)
+		goto error6;
+
+	return 0;
+error6:
+	del_timer_sync(&vub300->inactivity_timer);
+error5:
+	mmc_free_host(mmc);
+	/*
+	 * and hence also frees vub300
+	 * which is contained at the end of struct mmc
+	 */
+error4:
+	usb_free_urb(command_res_urb);
+error1:
+	usb_free_urb(command_out_urb);
+error0:
+	usb_put_dev(udev);
+	return retval;
+}
+
+static void vub300_disconnect(struct usb_interface *interface)
+{				/* NOT irq */
+	struct vub300_mmc_host *vub300 = usb_get_intfdata(interface);
+	if (!vub300 || !vub300->mmc) {
+		return;
+	} else {
+		struct mmc_host *mmc = vub300->mmc;
+		if (!vub300->mmc) {
+			return;
+		} else {
+			int ifnum = interface_to_InterfaceNumber(interface);
+			usb_set_intfdata(interface, NULL);
+			/* prevent more I/O from starting */
+			vub300->interface = NULL;
+			kref_put(&vub300->kref, vub300_delete);
+			mmc_remove_host(mmc);
+			pr_info("USB vub300 remote SDIO host controller[%d]"
+				" now disconnected", ifnum);
+			return;
+		}
+	}
+}
+
+#ifdef CONFIG_PM
+static int vub300_suspend(struct usb_interface *intf, pm_message_t message)
+{
+	return 0;
+}
+
+static int vub300_resume(struct usb_interface *intf)
+{
+	return 0;
+}
+#else
+#define vub300_suspend NULL
+#define vub300_resume NULL
+#endif
+static int vub300_pre_reset(struct usb_interface *intf)
+{				/* NOT irq */
+	struct vub300_mmc_host *vub300 = usb_get_intfdata(intf);
+	mutex_lock(&vub300->cmd_mutex);
+	return 0;
+}
+
+static int vub300_post_reset(struct usb_interface *intf)
+{				/* NOT irq */
+	struct vub300_mmc_host *vub300 = usb_get_intfdata(intf);
+	/* we are sure no URBs are active - no locking needed */
+	vub300->errors = -EPIPE;
+	mutex_unlock(&vub300->cmd_mutex);
+	return 0;
+}
+
+static struct usb_driver vub300_driver = {
+	.name = "vub300",
+	.probe = vub300_probe,
+	.disconnect = vub300_disconnect,
+	.suspend = vub300_suspend,
+	.resume = vub300_resume,
+	.pre_reset = vub300_pre_reset,
+	.post_reset = vub300_post_reset,
+	.id_table = vub300_table,
+	.supports_autosuspend = 1,
+};
+
+static int __init vub300_init(void)
+{				/* NOT irq */
+	int result;
+
+	pr_info("VUB300 Driver rom wait states = %02X irqpoll timeout = %04X",
+		firmware_rom_wait_states, 0x0FFFF & firmware_irqpoll_timeout);
+	cmndworkqueue = create_singlethread_workqueue("kvub300c");
+	if (!cmndworkqueue) {
+		pr_err("not enough memory for the REQUEST workqueue");
+		result = -ENOMEM;
+		goto out1;
+	}
+	pollworkqueue = create_singlethread_workqueue("kvub300p");
+	if (!pollworkqueue) {
+		pr_err("not enough memory for the IRQPOLL workqueue");
+		result = -ENOMEM;
+		goto out2;
+	}
+	deadworkqueue = create_singlethread_workqueue("kvub300d");
+	if (!deadworkqueue) {
+		pr_err("not enough memory for the EXPIRED workqueue");
+		result = -ENOMEM;
+		goto out3;
+	}
+	result = usb_register(&vub300_driver);
+	if (result) {
+		pr_err("usb_register failed. Error number %d", result);
+		goto out4;
+	}
+	return 0;
+out4:
+	destroy_workqueue(deadworkqueue);
+out3:
+	destroy_workqueue(pollworkqueue);
+out2:
+	destroy_workqueue(cmndworkqueue);
+out1:
+	return result;
+}
+
+static void __exit vub300_exit(void)
+{
+	usb_deregister(&vub300_driver);
+	flush_workqueue(cmndworkqueue);
+	flush_workqueue(pollworkqueue);
+	flush_workqueue(deadworkqueue);
+	destroy_workqueue(cmndworkqueue);
+	destroy_workqueue(pollworkqueue);
+	destroy_workqueue(deadworkqueue);
+}
+
+module_init(vub300_init);
+module_exit(vub300_exit);
+
+MODULE_AUTHOR("Tony Olech <tony.olech@elandigitalsystems.com>");
+MODULE_DESCRIPTION("VUB300 USB to SD/MMC/SDIO adapter driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mmc/host/wbsd.c b/drivers/mmc/host/wbsd.c
new file mode 100644
index 000000000..6db08070b
--- /dev/null
+++ b/drivers/mmc/host/wbsd.c
@@ -0,0 +1,2013 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ *  linux/drivers/mmc/host/wbsd.c - Winbond W83L51xD SD/MMC driver
+ *
+ *  Copyright (C) 2004-2007 Pierre Ossman, All Rights Reserved.
+ *
+ * Warning!
+ *
+ * Changes to the FIFO system should be done with extreme care since
+ * the hardware is full of bugs related to the FIFO. Known issues are:
+ *
+ * - FIFO size field in FSR is always zero.
+ *
+ * - FIFO interrupts tend not to work as they should. Interrupts are
+ *   triggered only for full/empty events, not for threshold values.
+ *
+ * - On APIC systems the FIFO empty interrupt is sometimes lost.
+ */
+
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/init.h>
+#include <linux/ioport.h>
+#include <linux/platform_device.h>
+#include <linux/interrupt.h>
+#include <linux/dma-mapping.h>
+#include <linux/delay.h>
+#include <linux/pnp.h>
+#include <linux/highmem.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sd.h>
+#include <linux/scatterlist.h>
+#include <linux/slab.h>
+
+#include <asm/io.h>
+#include <asm/dma.h>
+
+#include "wbsd.h"
+
+#define DRIVER_NAME "wbsd"
+
+#define DBG(x...) \
+	pr_debug(DRIVER_NAME ": " x)
+#define DBGF(f, x...) \
+	pr_debug(DRIVER_NAME " [%s()]: " f, __func__ , ##x)
+
+/*
+ * Device resources
+ */
+
+#ifdef CONFIG_PNP
+
+static const struct pnp_device_id pnp_dev_table[] = {
+	{ "WEC0517", 0 },
+	{ "WEC0518", 0 },
+	{ "", 0 },
+};
+
+MODULE_DEVICE_TABLE(pnp, pnp_dev_table);
+
+#endif /* CONFIG_PNP */
+
+static const int config_ports[] = { 0x2E, 0x4E };
+static const int unlock_codes[] = { 0x83, 0x87 };
+
+static const int valid_ids[] = {
+	0x7112,
+};
+
+#ifdef CONFIG_PNP
+static unsigned int param_nopnp = 0;
+#else
+static const unsigned int param_nopnp = 1;
+#endif
+static unsigned int param_io = 0x248;
+static unsigned int param_irq = 6;
+static int param_dma = 2;
+
+/*
+ * Basic functions
+ */
+
+static inline void wbsd_unlock_config(struct wbsd_host *host)
+{
+	BUG_ON(host->config == 0);
+
+	outb(host->unlock_code, host->config);
+	outb(host->unlock_code, host->config);
+}
+
+static inline void wbsd_lock_config(struct wbsd_host *host)
+{
+	BUG_ON(host->config == 0);
+
+	outb(LOCK_CODE, host->config);
+}
+
+static inline void wbsd_write_config(struct wbsd_host *host, u8 reg, u8 value)
+{
+	BUG_ON(host->config == 0);
+
+	outb(reg, host->config);
+	outb(value, host->config + 1);
+}
+
+static inline u8 wbsd_read_config(struct wbsd_host *host, u8 reg)
+{
+	BUG_ON(host->config == 0);
+
+	outb(reg, host->config);
+	return inb(host->config + 1);
+}
+
+static inline void wbsd_write_index(struct wbsd_host *host, u8 index, u8 value)
+{
+	outb(index, host->base + WBSD_IDXR);
+	outb(value, host->base + WBSD_DATAR);
+}
+
+static inline u8 wbsd_read_index(struct wbsd_host *host, u8 index)
+{
+	outb(index, host->base + WBSD_IDXR);
+	return inb(host->base + WBSD_DATAR);
+}
+
+/*
+ * Common routines
+ */
+
+static void wbsd_init_device(struct wbsd_host *host)
+{
+	u8 setup, ier;
+
+	/*
+	 * Reset chip (SD/MMC part) and fifo.
+	 */
+	setup = wbsd_read_index(host, WBSD_IDX_SETUP);
+	setup |= WBSD_FIFO_RESET | WBSD_SOFT_RESET;
+	wbsd_write_index(host, WBSD_IDX_SETUP, setup);
+
+	/*
+	 * Set DAT3 to input
+	 */
+	setup &= ~WBSD_DAT3_H;
+	wbsd_write_index(host, WBSD_IDX_SETUP, setup);
+	host->flags &= ~WBSD_FIGNORE_DETECT;
+
+	/*
+	 * Read back default clock.
+	 */
+	host->clk = wbsd_read_index(host, WBSD_IDX_CLK);
+
+	/*
+	 * Power down port.
+	 */
+	outb(WBSD_POWER_N, host->base + WBSD_CSR);
+
+	/*
+	 * Set maximum timeout.
+	 */
+	wbsd_write_index(host, WBSD_IDX_TAAC, 0x7F);
+
+	/*
+	 * Test for card presence
+	 */
+	if (inb(host->base + WBSD_CSR) & WBSD_CARDPRESENT)
+		host->flags |= WBSD_FCARD_PRESENT;
+	else
+		host->flags &= ~WBSD_FCARD_PRESENT;
+
+	/*
+	 * Enable interesting interrupts.
+	 */
+	ier = 0;
+	ier |= WBSD_EINT_CARD;
+	ier |= WBSD_EINT_FIFO_THRE;
+	ier |= WBSD_EINT_CRC;
+	ier |= WBSD_EINT_TIMEOUT;
+	ier |= WBSD_EINT_TC;
+
+	outb(ier, host->base + WBSD_EIR);
+
+	/*
+	 * Clear interrupts.
+	 */
+	inb(host->base + WBSD_ISR);
+}
+
+static void wbsd_reset(struct wbsd_host *host)
+{
+	u8 setup;
+
+	pr_err("%s: Resetting chip\n", mmc_hostname(host->mmc));
+
+	/*
+	 * Soft reset of chip (SD/MMC part).
+	 */
+	setup = wbsd_read_index(host, WBSD_IDX_SETUP);
+	setup |= WBSD_SOFT_RESET;
+	wbsd_write_index(host, WBSD_IDX_SETUP, setup);
+}
+
+static void wbsd_request_end(struct wbsd_host *host, struct mmc_request *mrq)
+{
+	unsigned long dmaflags;
+
+	if (host->dma >= 0) {
+		/*
+		 * Release ISA DMA controller.
+		 */
+		dmaflags = claim_dma_lock();
+		disable_dma(host->dma);
+		clear_dma_ff(host->dma);
+		release_dma_lock(dmaflags);
+
+		/*
+		 * Disable DMA on host.
+		 */
+		wbsd_write_index(host, WBSD_IDX_DMA, 0);
+	}
+
+	host->mrq = NULL;
+
+	/*
+	 * MMC layer might call back into the driver so first unlock.
+	 */
+	spin_unlock(&host->lock);
+	mmc_request_done(host->mmc, mrq);
+	spin_lock(&host->lock);
+}
+
+/*
+ * Scatter/gather functions
+ */
+
+static inline void wbsd_init_sg(struct wbsd_host *host, struct mmc_data *data)
+{
+	/*
+	 * Get info. about SG list from data structure.
+	 */
+	host->cur_sg = data->sg;
+	host->num_sg = data->sg_len;
+
+	host->offset = 0;
+	host->remain = host->cur_sg->length;
+}
+
+static inline int wbsd_next_sg(struct wbsd_host *host)
+{
+	/*
+	 * Skip to next SG entry.
+	 */
+	host->cur_sg++;
+	host->num_sg--;
+
+	/*
+	 * Any entries left?
+	 */
+	if (host->num_sg > 0) {
+		host->offset = 0;
+		host->remain = host->cur_sg->length;
+	}
+
+	return host->num_sg;
+}
+
+static inline char *wbsd_map_sg(struct wbsd_host *host)
+{
+	return kmap_atomic(sg_page(host->cur_sg)) + host->cur_sg->offset;
+}
+
+static inline void wbsd_sg_to_dma(struct wbsd_host *host, struct mmc_data *data)
+{
+	size_t len = 0;
+	int i;
+
+	for (i = 0; i < data->sg_len; i++)
+		len += data->sg[i].length;
+	sg_copy_to_buffer(data->sg, data->sg_len, host->dma_buffer, len);
+}
+
+static inline void wbsd_dma_to_sg(struct wbsd_host *host, struct mmc_data *data)
+{
+	size_t len = 0;
+	int i;
+
+	for (i = 0; i < data->sg_len; i++)
+		len += data->sg[i].length;
+	sg_copy_from_buffer(data->sg, data->sg_len, host->dma_buffer, len);
+}
+
+/*
+ * Command handling
+ */
+
+static inline void wbsd_get_short_reply(struct wbsd_host *host,
+					struct mmc_command *cmd)
+{
+	/*
+	 * Correct response type?
+	 */
+	if (wbsd_read_index(host, WBSD_IDX_RSPLEN) != WBSD_RSP_SHORT) {
+		cmd->error = -EILSEQ;
+		return;
+	}
+
+	cmd->resp[0]  = wbsd_read_index(host, WBSD_IDX_RESP12) << 24;
+	cmd->resp[0] |= wbsd_read_index(host, WBSD_IDX_RESP13) << 16;
+	cmd->resp[0] |= wbsd_read_index(host, WBSD_IDX_RESP14) << 8;
+	cmd->resp[0] |= wbsd_read_index(host, WBSD_IDX_RESP15) << 0;
+	cmd->resp[1]  = wbsd_read_index(host, WBSD_IDX_RESP16) << 24;
+}
+
+static inline void wbsd_get_long_reply(struct wbsd_host *host,
+	struct mmc_command *cmd)
+{
+	int i;
+
+	/*
+	 * Correct response type?
+	 */
+	if (wbsd_read_index(host, WBSD_IDX_RSPLEN) != WBSD_RSP_LONG) {
+		cmd->error = -EILSEQ;
+		return;
+	}
+
+	for (i = 0; i < 4; i++) {
+		cmd->resp[i] =
+			wbsd_read_index(host, WBSD_IDX_RESP1 + i * 4) << 24;
+		cmd->resp[i] |=
+			wbsd_read_index(host, WBSD_IDX_RESP2 + i * 4) << 16;
+		cmd->resp[i] |=
+			wbsd_read_index(host, WBSD_IDX_RESP3 + i * 4) << 8;
+		cmd->resp[i] |=
+			wbsd_read_index(host, WBSD_IDX_RESP4 + i * 4) << 0;
+	}
+}
+
+static void wbsd_send_command(struct wbsd_host *host, struct mmc_command *cmd)
+{
+	int i;
+	u8 status, isr;
+
+	/*
+	 * Clear accumulated ISR. The interrupt routine
+	 * will fill this one with events that occur during
+	 * transfer.
+	 */
+	host->isr = 0;
+
+	/*
+	 * Send the command (CRC calculated by host).
+	 */
+	outb(cmd->opcode, host->base + WBSD_CMDR);
+	for (i = 3; i >= 0; i--)
+		outb((cmd->arg >> (i * 8)) & 0xff, host->base + WBSD_CMDR);
+
+	cmd->error = 0;
+
+	/*
+	 * Wait for the request to complete.
+	 */
+	do {
+		status = wbsd_read_index(host, WBSD_IDX_STATUS);
+	} while (status & WBSD_CARDTRAFFIC);
+
+	/*
+	 * Do we expect a reply?
+	 */
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		/*
+		 * Read back status.
+		 */
+		isr = host->isr;
+
+		/* Card removed? */
+		if (isr & WBSD_INT_CARD)
+			cmd->error = -ENOMEDIUM;
+		/* Timeout? */
+		else if (isr & WBSD_INT_TIMEOUT)
+			cmd->error = -ETIMEDOUT;
+		/* CRC? */
+		else if ((cmd->flags & MMC_RSP_CRC) && (isr & WBSD_INT_CRC))
+			cmd->error = -EILSEQ;
+		/* All ok */
+		else {
+			if (cmd->flags & MMC_RSP_136)
+				wbsd_get_long_reply(host, cmd);
+			else
+				wbsd_get_short_reply(host, cmd);
+		}
+	}
+}
+
+/*
+ * Data functions
+ */
+
+static void wbsd_empty_fifo(struct wbsd_host *host)
+{
+	struct mmc_data *data = host->mrq->cmd->data;
+	char *buffer;
+	int i, idx, fsr, fifo;
+
+	/*
+	 * Handle excessive data.
+	 */
+	if (host->num_sg == 0)
+		return;
+
+	buffer = wbsd_map_sg(host) + host->offset;
+	idx = 0;
+
+	/*
+	 * Drain the fifo. This has a tendency to loop longer
+	 * than the FIFO length (usually one block).
+	 */
+	while (!((fsr = inb(host->base + WBSD_FSR)) & WBSD_FIFO_EMPTY)) {
+		/*
+		 * The size field in the FSR is broken so we have to
+		 * do some guessing.
+		 */
+		if (fsr & WBSD_FIFO_FULL)
+			fifo = 16;
+		else if (fsr & WBSD_FIFO_FUTHRE)
+			fifo = 8;
+		else
+			fifo = 1;
+
+		for (i = 0; i < fifo; i++) {
+			buffer[idx++] = inb(host->base + WBSD_DFR);
+			host->offset++;
+			host->remain--;
+
+			data->bytes_xfered++;
+
+			/*
+			 * End of scatter list entry?
+			 */
+			if (host->remain == 0) {
+				kunmap_atomic(buffer);
+				/*
+				 * Get next entry. Check if last.
+				 */
+				if (!wbsd_next_sg(host))
+					return;
+
+				buffer = wbsd_map_sg(host);
+				idx = 0;
+			}
+		}
+	}
+	kunmap_atomic(buffer);
+
+	/*
+	 * This is a very dirty hack to solve a
+	 * hardware problem. The chip doesn't trigger
+	 * FIFO threshold interrupts properly.
+	 */
+	if ((data->blocks * data->blksz - data->bytes_xfered) < 16)
+		tasklet_schedule(&host->fifo_tasklet);
+}
+
+static void wbsd_fill_fifo(struct wbsd_host *host)
+{
+	struct mmc_data *data = host->mrq->cmd->data;
+	char *buffer;
+	int i, idx, fsr, fifo;
+
+	/*
+	 * Check that we aren't being called after the
+	 * entire buffer has been transferred.
+	 */
+	if (host->num_sg == 0)
+		return;
+
+	buffer = wbsd_map_sg(host) + host->offset;
+	idx = 0;
+
+	/*
+	 * Fill the fifo. This has a tendency to loop longer
+	 * than the FIFO length (usually one block).
+	 */
+	while (!((fsr = inb(host->base + WBSD_FSR)) & WBSD_FIFO_FULL)) {
+		/*
+		 * The size field in the FSR is broken so we have to
+		 * do some guessing.
+		 */
+		if (fsr & WBSD_FIFO_EMPTY)
+			fifo = 0;
+		else if (fsr & WBSD_FIFO_EMTHRE)
+			fifo = 8;
+		else
+			fifo = 15;
+
+		for (i = 16; i > fifo; i--) {
+			outb(buffer[idx], host->base + WBSD_DFR);
+			host->offset++;
+			host->remain--;
+
+			data->bytes_xfered++;
+
+			/*
+			 * End of scatter list entry?
+			 */
+			if (host->remain == 0) {
+				kunmap_atomic(buffer);
+				/*
+				 * Get next entry. Check if last.
+				 */
+				if (!wbsd_next_sg(host))
+					return;
+
+				buffer = wbsd_map_sg(host);
+				idx = 0;
+			}
+		}
+	}
+	kunmap_atomic(buffer);
+
+	/*
+	 * The controller stops sending interrupts for
+	 * 'FIFO empty' under certain conditions. So we
+	 * need to be a bit more pro-active.
+	 */
+	tasklet_schedule(&host->fifo_tasklet);
+}
+
+static void wbsd_prepare_data(struct wbsd_host *host, struct mmc_data *data)
+{
+	u16 blksize;
+	u8 setup;
+	unsigned long dmaflags;
+	unsigned int size;
+
+	/*
+	 * Calculate size.
+	 */
+	size = data->blocks * data->blksz;
+
+	/*
+	 * Check timeout values for overflow.
+	 * (Yes, some cards cause this value to overflow).
+	 */
+	if (data->timeout_ns > 127000000)
+		wbsd_write_index(host, WBSD_IDX_TAAC, 127);
+	else {
+		wbsd_write_index(host, WBSD_IDX_TAAC,
+			data->timeout_ns / 1000000);
+	}
+
+	if (data->timeout_clks > 255)
+		wbsd_write_index(host, WBSD_IDX_NSAC, 255);
+	else
+		wbsd_write_index(host, WBSD_IDX_NSAC, data->timeout_clks);
+
+	/*
+	 * Inform the chip of how large blocks will be
+	 * sent. It needs this to determine when to
+	 * calculate CRC.
+	 *
+	 * Space for CRC must be included in the size.
+	 * Two bytes are needed for each data line.
+	 */
+	if (host->bus_width == MMC_BUS_WIDTH_1) {
+		blksize = data->blksz + 2;
+
+		wbsd_write_index(host, WBSD_IDX_PBSMSB, (blksize >> 4) & 0xF0);
+		wbsd_write_index(host, WBSD_IDX_PBSLSB, blksize & 0xFF);
+	} else if (host->bus_width == MMC_BUS_WIDTH_4) {
+		blksize = data->blksz + 2 * 4;
+
+		wbsd_write_index(host, WBSD_IDX_PBSMSB,
+			((blksize >> 4) & 0xF0) | WBSD_DATA_WIDTH);
+		wbsd_write_index(host, WBSD_IDX_PBSLSB, blksize & 0xFF);
+	} else {
+		data->error = -EINVAL;
+		return;
+	}
+
+	/*
+	 * Clear the FIFO. This is needed even for DMA
+	 * transfers since the chip still uses the FIFO
+	 * internally.
+	 */
+	setup = wbsd_read_index(host, WBSD_IDX_SETUP);
+	setup |= WBSD_FIFO_RESET;
+	wbsd_write_index(host, WBSD_IDX_SETUP, setup);
+
+	/*
+	 * DMA transfer?
+	 */
+	if (host->dma >= 0) {
+		/*
+		 * The buffer for DMA is only 64 kB.
+		 */
+		BUG_ON(size > 0x10000);
+		if (size > 0x10000) {
+			data->error = -EINVAL;
+			return;
+		}
+
+		/*
+		 * Transfer data from the SG list to
+		 * the DMA buffer.
+		 */
+		if (data->flags & MMC_DATA_WRITE)
+			wbsd_sg_to_dma(host, data);
+
+		/*
+		 * Initialise the ISA DMA controller.
+		 */
+		dmaflags = claim_dma_lock();
+		disable_dma(host->dma);
+		clear_dma_ff(host->dma);
+		if (data->flags & MMC_DATA_READ)
+			set_dma_mode(host->dma, DMA_MODE_READ & ~0x40);
+		else
+			set_dma_mode(host->dma, DMA_MODE_WRITE & ~0x40);
+		set_dma_addr(host->dma, host->dma_addr);
+		set_dma_count(host->dma, size);
+
+		enable_dma(host->dma);
+		release_dma_lock(dmaflags);
+
+		/*
+		 * Enable DMA on the host.
+		 */
+		wbsd_write_index(host, WBSD_IDX_DMA, WBSD_DMA_ENABLE);
+	} else {
+		/*
+		 * This flag is used to keep printk
+		 * output to a minimum.
+		 */
+		host->firsterr = 1;
+
+		/*
+		 * Initialise the SG list.
+		 */
+		wbsd_init_sg(host, data);
+
+		/*
+		 * Turn off DMA.
+		 */
+		wbsd_write_index(host, WBSD_IDX_DMA, 0);
+
+		/*
+		 * Set up FIFO threshold levels (and fill
+		 * buffer if doing a write).
+		 */
+		if (data->flags & MMC_DATA_READ) {
+			wbsd_write_index(host, WBSD_IDX_FIFOEN,
+				WBSD_FIFOEN_FULL | 8);
+		} else {
+			wbsd_write_index(host, WBSD_IDX_FIFOEN,
+				WBSD_FIFOEN_EMPTY | 8);
+			wbsd_fill_fifo(host);
+		}
+	}
+
+	data->error = 0;
+}
+
+static void wbsd_finish_data(struct wbsd_host *host, struct mmc_data *data)
+{
+	unsigned long dmaflags;
+	int count;
+	u8 status;
+
+	WARN_ON(host->mrq == NULL);
+
+	/*
+	 * Send a stop command if needed.
+	 */
+	if (data->stop)
+		wbsd_send_command(host, data->stop);
+
+	/*
+	 * Wait for the controller to leave data
+	 * transfer state.
+	 */
+	do {
+		status = wbsd_read_index(host, WBSD_IDX_STATUS);
+	} while (status & (WBSD_BLOCK_READ | WBSD_BLOCK_WRITE));
+
+	/*
+	 * DMA transfer?
+	 */
+	if (host->dma >= 0) {
+		/*
+		 * Disable DMA on the host.
+		 */
+		wbsd_write_index(host, WBSD_IDX_DMA, 0);
+
+		/*
+		 * Turn of ISA DMA controller.
+		 */
+		dmaflags = claim_dma_lock();
+		disable_dma(host->dma);
+		clear_dma_ff(host->dma);
+		count = get_dma_residue(host->dma);
+		release_dma_lock(dmaflags);
+
+		data->bytes_xfered = host->mrq->data->blocks *
+			host->mrq->data->blksz - count;
+		data->bytes_xfered -= data->bytes_xfered % data->blksz;
+
+		/*
+		 * Any leftover data?
+		 */
+		if (count) {
+			pr_err("%s: Incomplete DMA transfer. "
+				"%d bytes left.\n",
+				mmc_hostname(host->mmc), count);
+
+			if (!data->error)
+				data->error = -EIO;
+		} else {
+			/*
+			 * Transfer data from DMA buffer to
+			 * SG list.
+			 */
+			if (data->flags & MMC_DATA_READ)
+				wbsd_dma_to_sg(host, data);
+		}
+
+		if (data->error) {
+			if (data->bytes_xfered)
+				data->bytes_xfered -= data->blksz;
+		}
+	}
+
+	wbsd_request_end(host, host->mrq);
+}
+
+/*****************************************************************************\
+ *                                                                           *
+ * MMC layer callbacks                                                       *
+ *                                                                           *
+\*****************************************************************************/
+
+static void wbsd_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct wbsd_host *host = mmc_priv(mmc);
+	struct mmc_command *cmd;
+
+	/*
+	 * Disable tasklets to avoid a deadlock.
+	 */
+	spin_lock_bh(&host->lock);
+
+	BUG_ON(host->mrq != NULL);
+
+	cmd = mrq->cmd;
+
+	host->mrq = mrq;
+
+	/*
+	 * Check that there is actually a card in the slot.
+	 */
+	if (!(host->flags & WBSD_FCARD_PRESENT)) {
+		cmd->error = -ENOMEDIUM;
+		goto done;
+	}
+
+	if (cmd->data) {
+		/*
+		 * The hardware is so delightfully stupid that it has a list
+		 * of "data" commands. If a command isn't on this list, it'll
+		 * just go back to the idle state and won't send any data
+		 * interrupts.
+		 */
+		switch (cmd->opcode) {
+		case SD_SWITCH_VOLTAGE:
+		case MMC_READ_SINGLE_BLOCK:
+		case MMC_READ_MULTIPLE_BLOCK:
+		case MMC_WRITE_DAT_UNTIL_STOP:
+		case MMC_WRITE_BLOCK:
+		case MMC_WRITE_MULTIPLE_BLOCK:
+		case MMC_PROGRAM_CID:
+		case MMC_PROGRAM_CSD:
+		case MMC_SEND_WRITE_PROT:
+		case MMC_LOCK_UNLOCK:
+		case MMC_GEN_CMD:
+			break;
+
+		/* ACMDs. We don't keep track of state, so we just treat them
+		 * like any other command. */
+		case SD_APP_SEND_SCR:
+			break;
+
+		default:
+			pr_warn("%s: Data command %d is not supported by this controller\n",
+				mmc_hostname(host->mmc), cmd->opcode);
+			cmd->error = -EINVAL;
+
+			goto done;
+		}
+	}
+
+	/*
+	 * Does the request include data?
+	 */
+	if (cmd->data) {
+		wbsd_prepare_data(host, cmd->data);
+
+		if (cmd->data->error)
+			goto done;
+	}
+
+	wbsd_send_command(host, cmd);
+
+	/*
+	 * If this is a data transfer the request
+	 * will be finished after the data has
+	 * transferred.
+	 */
+	if (cmd->data && !cmd->error) {
+		/*
+		 * Dirty fix for hardware bug.
+		 */
+		if (host->dma == -1)
+			tasklet_schedule(&host->fifo_tasklet);
+
+		spin_unlock_bh(&host->lock);
+
+		return;
+	}
+
+done:
+	wbsd_request_end(host, mrq);
+
+	spin_unlock_bh(&host->lock);
+}
+
+static void wbsd_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct wbsd_host *host = mmc_priv(mmc);
+	u8 clk, setup, pwr;
+
+	spin_lock_bh(&host->lock);
+
+	/*
+	 * Reset the chip on each power off.
+	 * Should clear out any weird states.
+	 */
+	if (ios->power_mode == MMC_POWER_OFF)
+		wbsd_init_device(host);
+
+	if (ios->clock >= 24000000)
+		clk = WBSD_CLK_24M;
+	else if (ios->clock >= 16000000)
+		clk = WBSD_CLK_16M;
+	else if (ios->clock >= 12000000)
+		clk = WBSD_CLK_12M;
+	else
+		clk = WBSD_CLK_375K;
+
+	/*
+	 * Only write to the clock register when
+	 * there is an actual change.
+	 */
+	if (clk != host->clk) {
+		wbsd_write_index(host, WBSD_IDX_CLK, clk);
+		host->clk = clk;
+	}
+
+	/*
+	 * Power up card.
+	 */
+	if (ios->power_mode != MMC_POWER_OFF) {
+		pwr = inb(host->base + WBSD_CSR);
+		pwr &= ~WBSD_POWER_N;
+		outb(pwr, host->base + WBSD_CSR);
+	}
+
+	/*
+	 * MMC cards need to have pin 1 high during init.
+	 * It wreaks havoc with the card detection though so
+	 * that needs to be disabled.
+	 */
+	setup = wbsd_read_index(host, WBSD_IDX_SETUP);
+	if (ios->chip_select == MMC_CS_HIGH) {
+		BUG_ON(ios->bus_width != MMC_BUS_WIDTH_1);
+		setup |= WBSD_DAT3_H;
+		host->flags |= WBSD_FIGNORE_DETECT;
+	} else {
+		if (setup & WBSD_DAT3_H) {
+			setup &= ~WBSD_DAT3_H;
+
+			/*
+			 * We cannot resume card detection immediately
+			 * because of capacitance and delays in the chip.
+			 */
+			mod_timer(&host->ignore_timer, jiffies + HZ / 100);
+		}
+	}
+	wbsd_write_index(host, WBSD_IDX_SETUP, setup);
+
+	/*
+	 * Store bus width for later. Will be used when
+	 * setting up the data transfer.
+	 */
+	host->bus_width = ios->bus_width;
+
+	spin_unlock_bh(&host->lock);
+}
+
+static int wbsd_get_ro(struct mmc_host *mmc)
+{
+	struct wbsd_host *host = mmc_priv(mmc);
+	u8 csr;
+
+	spin_lock_bh(&host->lock);
+
+	csr = inb(host->base + WBSD_CSR);
+	csr |= WBSD_MSLED;
+	outb(csr, host->base + WBSD_CSR);
+
+	mdelay(1);
+
+	csr = inb(host->base + WBSD_CSR);
+	csr &= ~WBSD_MSLED;
+	outb(csr, host->base + WBSD_CSR);
+
+	spin_unlock_bh(&host->lock);
+
+	return !!(csr & WBSD_WRPT);
+}
+
+static const struct mmc_host_ops wbsd_ops = {
+	.request	= wbsd_request,
+	.set_ios	= wbsd_set_ios,
+	.get_ro		= wbsd_get_ro,
+};
+
+/*****************************************************************************\
+ *                                                                           *
+ * Interrupt handling                                                        *
+ *                                                                           *
+\*****************************************************************************/
+
+/*
+ * Helper function to reset detection ignore
+ */
+
+static void wbsd_reset_ignore(struct timer_list *t)
+{
+	struct wbsd_host *host = from_timer(host, t, ignore_timer);
+
+	BUG_ON(host == NULL);
+
+	DBG("Resetting card detection ignore\n");
+
+	spin_lock_bh(&host->lock);
+
+	host->flags &= ~WBSD_FIGNORE_DETECT;
+
+	/*
+	 * Card status might have changed during the
+	 * blackout.
+	 */
+	tasklet_schedule(&host->card_tasklet);
+
+	spin_unlock_bh(&host->lock);
+}
+
+/*
+ * Tasklets
+ */
+
+static inline struct mmc_data *wbsd_get_data(struct wbsd_host *host)
+{
+	WARN_ON(!host->mrq);
+	if (!host->mrq)
+		return NULL;
+
+	WARN_ON(!host->mrq->cmd);
+	if (!host->mrq->cmd)
+		return NULL;
+
+	WARN_ON(!host->mrq->cmd->data);
+	if (!host->mrq->cmd->data)
+		return NULL;
+
+	return host->mrq->cmd->data;
+}
+
+static void wbsd_tasklet_card(unsigned long param)
+{
+	struct wbsd_host *host = (struct wbsd_host *)param;
+	u8 csr;
+	int delay = -1;
+
+	spin_lock(&host->lock);
+
+	if (host->flags & WBSD_FIGNORE_DETECT) {
+		spin_unlock(&host->lock);
+		return;
+	}
+
+	csr = inb(host->base + WBSD_CSR);
+	WARN_ON(csr == 0xff);
+
+	if (csr & WBSD_CARDPRESENT) {
+		if (!(host->flags & WBSD_FCARD_PRESENT)) {
+			DBG("Card inserted\n");
+			host->flags |= WBSD_FCARD_PRESENT;
+
+			delay = 500;
+		}
+	} else if (host->flags & WBSD_FCARD_PRESENT) {
+		DBG("Card removed\n");
+		host->flags &= ~WBSD_FCARD_PRESENT;
+
+		if (host->mrq) {
+			pr_err("%s: Card removed during transfer!\n",
+				mmc_hostname(host->mmc));
+			wbsd_reset(host);
+
+			host->mrq->cmd->error = -ENOMEDIUM;
+			tasklet_schedule(&host->finish_tasklet);
+		}
+
+		delay = 0;
+	}
+
+	/*
+	 * Unlock first since we might get a call back.
+	 */
+
+	spin_unlock(&host->lock);
+
+	if (delay != -1)
+		mmc_detect_change(host->mmc, msecs_to_jiffies(delay));
+}
+
+static void wbsd_tasklet_fifo(unsigned long param)
+{
+	struct wbsd_host *host = (struct wbsd_host *)param;
+	struct mmc_data *data;
+
+	spin_lock(&host->lock);
+
+	if (!host->mrq)
+		goto end;
+
+	data = wbsd_get_data(host);
+	if (!data)
+		goto end;
+
+	if (data->flags & MMC_DATA_WRITE)
+		wbsd_fill_fifo(host);
+	else
+		wbsd_empty_fifo(host);
+
+	/*
+	 * Done?
+	 */
+	if (host->num_sg == 0) {
+		wbsd_write_index(host, WBSD_IDX_FIFOEN, 0);
+		tasklet_schedule(&host->finish_tasklet);
+	}
+
+end:
+	spin_unlock(&host->lock);
+}
+
+static void wbsd_tasklet_crc(unsigned long param)
+{
+	struct wbsd_host *host = (struct wbsd_host *)param;
+	struct mmc_data *data;
+
+	spin_lock(&host->lock);
+
+	if (!host->mrq)
+		goto end;
+
+	data = wbsd_get_data(host);
+	if (!data)
+		goto end;
+
+	DBGF("CRC error\n");
+
+	data->error = -EILSEQ;
+
+	tasklet_schedule(&host->finish_tasklet);
+
+end:
+	spin_unlock(&host->lock);
+}
+
+static void wbsd_tasklet_timeout(unsigned long param)
+{
+	struct wbsd_host *host = (struct wbsd_host *)param;
+	struct mmc_data *data;
+
+	spin_lock(&host->lock);
+
+	if (!host->mrq)
+		goto end;
+
+	data = wbsd_get_data(host);
+	if (!data)
+		goto end;
+
+	DBGF("Timeout\n");
+
+	data->error = -ETIMEDOUT;
+
+	tasklet_schedule(&host->finish_tasklet);
+
+end:
+	spin_unlock(&host->lock);
+}
+
+static void wbsd_tasklet_finish(unsigned long param)
+{
+	struct wbsd_host *host = (struct wbsd_host *)param;
+	struct mmc_data *data;
+
+	spin_lock(&host->lock);
+
+	WARN_ON(!host->mrq);
+	if (!host->mrq)
+		goto end;
+
+	data = wbsd_get_data(host);
+	if (!data)
+		goto end;
+
+	wbsd_finish_data(host, data);
+
+end:
+	spin_unlock(&host->lock);
+}
+
+/*
+ * Interrupt handling
+ */
+
+static irqreturn_t wbsd_irq(int irq, void *dev_id)
+{
+	struct wbsd_host *host = dev_id;
+	int isr;
+
+	isr = inb(host->base + WBSD_ISR);
+
+	/*
+	 * Was it actually our hardware that caused the interrupt?
+	 */
+	if (isr == 0xff || isr == 0x00)
+		return IRQ_NONE;
+
+	host->isr |= isr;
+
+	/*
+	 * Schedule tasklets as needed.
+	 */
+	if (isr & WBSD_INT_CARD)
+		tasklet_schedule(&host->card_tasklet);
+	if (isr & WBSD_INT_FIFO_THRE)
+		tasklet_schedule(&host->fifo_tasklet);
+	if (isr & WBSD_INT_CRC)
+		tasklet_hi_schedule(&host->crc_tasklet);
+	if (isr & WBSD_INT_TIMEOUT)
+		tasklet_hi_schedule(&host->timeout_tasklet);
+	if (isr & WBSD_INT_TC)
+		tasklet_schedule(&host->finish_tasklet);
+
+	return IRQ_HANDLED;
+}
+
+/*****************************************************************************\
+ *                                                                           *
+ * Device initialisation and shutdown                                        *
+ *                                                                           *
+\*****************************************************************************/
+
+/*
+ * Allocate/free MMC structure.
+ */
+
+static int wbsd_alloc_mmc(struct device *dev)
+{
+	struct mmc_host *mmc;
+	struct wbsd_host *host;
+
+	/*
+	 * Allocate MMC structure.
+	 */
+	mmc = mmc_alloc_host(sizeof(struct wbsd_host), dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+
+	host->dma = -1;
+
+	/*
+	 * Set host parameters.
+	 */
+	mmc->ops = &wbsd_ops;
+	mmc->f_min = 375000;
+	mmc->f_max = 24000000;
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+	mmc->caps = MMC_CAP_4_BIT_DATA;
+
+	spin_lock_init(&host->lock);
+
+	/*
+	 * Set up timers
+	 */
+	timer_setup(&host->ignore_timer, wbsd_reset_ignore, 0);
+
+	/*
+	 * Maximum number of segments. Worst case is one sector per segment
+	 * so this will be 64kB/512.
+	 */
+	mmc->max_segs = 128;
+
+	/*
+	 * Maximum request size. Also limited by 64KiB buffer.
+	 */
+	mmc->max_req_size = 65536;
+
+	/*
+	 * Maximum segment size. Could be one segment with the maximum number
+	 * of bytes.
+	 */
+	mmc->max_seg_size = mmc->max_req_size;
+
+	/*
+	 * Maximum block size. We have 12 bits (= 4095) but have to subtract
+	 * space for CRC. So the maximum is 4095 - 4*2 = 4087.
+	 */
+	mmc->max_blk_size = 4087;
+
+	/*
+	 * Maximum block count. There is no real limit so the maximum
+	 * request size will be the only restriction.
+	 */
+	mmc->max_blk_count = mmc->max_req_size;
+
+	dev_set_drvdata(dev, mmc);
+
+	return 0;
+}
+
+static void wbsd_free_mmc(struct device *dev)
+{
+	struct mmc_host *mmc;
+	struct wbsd_host *host;
+
+	mmc = dev_get_drvdata(dev);
+	if (!mmc)
+		return;
+
+	host = mmc_priv(mmc);
+	BUG_ON(host == NULL);
+
+	del_timer_sync(&host->ignore_timer);
+
+	mmc_free_host(mmc);
+
+	dev_set_drvdata(dev, NULL);
+}
+
+/*
+ * Scan for known chip id:s
+ */
+
+static int wbsd_scan(struct wbsd_host *host)
+{
+	int i, j, k;
+	int id;
+
+	/*
+	 * Iterate through all ports, all codes to
+	 * find hardware that is in our known list.
+	 */
+	for (i = 0; i < ARRAY_SIZE(config_ports); i++) {
+		if (!request_region(config_ports[i], 2, DRIVER_NAME))
+			continue;
+
+		for (j = 0; j < ARRAY_SIZE(unlock_codes); j++) {
+			id = 0xFFFF;
+
+			host->config = config_ports[i];
+			host->unlock_code = unlock_codes[j];
+
+			wbsd_unlock_config(host);
+
+			outb(WBSD_CONF_ID_HI, config_ports[i]);
+			id = inb(config_ports[i] + 1) << 8;
+
+			outb(WBSD_CONF_ID_LO, config_ports[i]);
+			id |= inb(config_ports[i] + 1);
+
+			wbsd_lock_config(host);
+
+			for (k = 0; k < ARRAY_SIZE(valid_ids); k++) {
+				if (id == valid_ids[k]) {
+					host->chip_id = id;
+
+					return 0;
+				}
+			}
+
+			if (id != 0xFFFF) {
+				DBG("Unknown hardware (id %x) found at %x\n",
+					id, config_ports[i]);
+			}
+		}
+
+		release_region(config_ports[i], 2);
+	}
+
+	host->config = 0;
+	host->unlock_code = 0;
+
+	return -ENODEV;
+}
+
+/*
+ * Allocate/free io port ranges
+ */
+
+static int wbsd_request_region(struct wbsd_host *host, int base)
+{
+	if (base & 0x7)
+		return -EINVAL;
+
+	if (!request_region(base, 8, DRIVER_NAME))
+		return -EIO;
+
+	host->base = base;
+
+	return 0;
+}
+
+static void wbsd_release_regions(struct wbsd_host *host)
+{
+	if (host->base)
+		release_region(host->base, 8);
+
+	host->base = 0;
+
+	if (host->config)
+		release_region(host->config, 2);
+
+	host->config = 0;
+}
+
+/*
+ * Allocate/free DMA port and buffer
+ */
+
+static void wbsd_request_dma(struct wbsd_host *host, int dma)
+{
+	if (dma < 0)
+		return;
+
+	if (request_dma(dma, DRIVER_NAME))
+		goto err;
+
+	/*
+	 * We need to allocate a special buffer in
+	 * order for ISA to be able to DMA to it.
+	 */
+	host->dma_buffer = kmalloc(WBSD_DMA_SIZE,
+		GFP_NOIO | GFP_DMA | __GFP_RETRY_MAYFAIL | __GFP_NOWARN);
+	if (!host->dma_buffer)
+		goto free;
+
+	/*
+	 * Translate the address to a physical address.
+	 */
+	host->dma_addr = dma_map_single(mmc_dev(host->mmc), host->dma_buffer,
+		WBSD_DMA_SIZE, DMA_BIDIRECTIONAL);
+	if (dma_mapping_error(mmc_dev(host->mmc), host->dma_addr))
+		goto kfree;
+
+	/*
+	 * ISA DMA must be aligned on a 64k basis.
+	 */
+	if ((host->dma_addr & 0xffff) != 0)
+		goto unmap;
+	/*
+	 * ISA cannot access memory above 16 MB.
+	 */
+	else if (host->dma_addr >= 0x1000000)
+		goto unmap;
+
+	host->dma = dma;
+
+	return;
+
+unmap:
+	/*
+	 * If we've gotten here then there is some kind of alignment bug
+	 */
+	BUG_ON(1);
+
+	dma_unmap_single(mmc_dev(host->mmc), host->dma_addr,
+		WBSD_DMA_SIZE, DMA_BIDIRECTIONAL);
+	host->dma_addr = 0;
+
+kfree:
+	kfree(host->dma_buffer);
+	host->dma_buffer = NULL;
+
+free:
+	free_dma(dma);
+
+err:
+	pr_warn(DRIVER_NAME ": Unable to allocate DMA %d - falling back on FIFO\n",
+		dma);
+}
+
+static void wbsd_release_dma(struct wbsd_host *host)
+{
+	/*
+	 * host->dma_addr is valid here iff host->dma_buffer is not NULL.
+	 */
+	if (host->dma_buffer) {
+		dma_unmap_single(mmc_dev(host->mmc), host->dma_addr,
+			WBSD_DMA_SIZE, DMA_BIDIRECTIONAL);
+		kfree(host->dma_buffer);
+	}
+	if (host->dma >= 0)
+		free_dma(host->dma);
+
+	host->dma = -1;
+	host->dma_buffer = NULL;
+	host->dma_addr = 0;
+}
+
+/*
+ * Allocate/free IRQ.
+ */
+
+static int wbsd_request_irq(struct wbsd_host *host, int irq)
+{
+	int ret;
+
+	/*
+	 * Set up tasklets. Must be done before requesting interrupt.
+	 */
+	tasklet_init(&host->card_tasklet, wbsd_tasklet_card,
+			(unsigned long)host);
+	tasklet_init(&host->fifo_tasklet, wbsd_tasklet_fifo,
+			(unsigned long)host);
+	tasklet_init(&host->crc_tasklet, wbsd_tasklet_crc,
+			(unsigned long)host);
+	tasklet_init(&host->timeout_tasklet, wbsd_tasklet_timeout,
+			(unsigned long)host);
+	tasklet_init(&host->finish_tasklet, wbsd_tasklet_finish,
+			(unsigned long)host);
+
+	/*
+	 * Allocate interrupt.
+	 */
+	ret = request_irq(irq, wbsd_irq, IRQF_SHARED, DRIVER_NAME, host);
+	if (ret)
+		return ret;
+
+	host->irq = irq;
+
+	return 0;
+}
+
+static void  wbsd_release_irq(struct wbsd_host *host)
+{
+	if (!host->irq)
+		return;
+
+	free_irq(host->irq, host);
+
+	host->irq = 0;
+
+	tasklet_kill(&host->card_tasklet);
+	tasklet_kill(&host->fifo_tasklet);
+	tasklet_kill(&host->crc_tasklet);
+	tasklet_kill(&host->timeout_tasklet);
+	tasklet_kill(&host->finish_tasklet);
+}
+
+/*
+ * Allocate all resources for the host.
+ */
+
+static int wbsd_request_resources(struct wbsd_host *host,
+	int base, int irq, int dma)
+{
+	int ret;
+
+	/*
+	 * Allocate I/O ports.
+	 */
+	ret = wbsd_request_region(host, base);
+	if (ret)
+		return ret;
+
+	/*
+	 * Allocate interrupt.
+	 */
+	ret = wbsd_request_irq(host, irq);
+	if (ret)
+		return ret;
+
+	/*
+	 * Allocate DMA.
+	 */
+	wbsd_request_dma(host, dma);
+
+	return 0;
+}
+
+/*
+ * Release all resources for the host.
+ */
+
+static void wbsd_release_resources(struct wbsd_host *host)
+{
+	wbsd_release_dma(host);
+	wbsd_release_irq(host);
+	wbsd_release_regions(host);
+}
+
+/*
+ * Configure the resources the chip should use.
+ */
+
+static void wbsd_chip_config(struct wbsd_host *host)
+{
+	wbsd_unlock_config(host);
+
+	/*
+	 * Reset the chip.
+	 */
+	wbsd_write_config(host, WBSD_CONF_SWRST, 1);
+	wbsd_write_config(host, WBSD_CONF_SWRST, 0);
+
+	/*
+	 * Select SD/MMC function.
+	 */
+	wbsd_write_config(host, WBSD_CONF_DEVICE, DEVICE_SD);
+
+	/*
+	 * Set up card detection.
+	 */
+	wbsd_write_config(host, WBSD_CONF_PINS, WBSD_PINS_DETECT_GP11);
+
+	/*
+	 * Configure chip
+	 */
+	wbsd_write_config(host, WBSD_CONF_PORT_HI, host->base >> 8);
+	wbsd_write_config(host, WBSD_CONF_PORT_LO, host->base & 0xff);
+
+	wbsd_write_config(host, WBSD_CONF_IRQ, host->irq);
+
+	if (host->dma >= 0)
+		wbsd_write_config(host, WBSD_CONF_DRQ, host->dma);
+
+	/*
+	 * Enable and power up chip.
+	 */
+	wbsd_write_config(host, WBSD_CONF_ENABLE, 1);
+	wbsd_write_config(host, WBSD_CONF_POWER, 0x20);
+
+	wbsd_lock_config(host);
+}
+
+/*
+ * Check that configured resources are correct.
+ */
+
+static int wbsd_chip_validate(struct wbsd_host *host)
+{
+	int base, irq, dma;
+
+	wbsd_unlock_config(host);
+
+	/*
+	 * Select SD/MMC function.
+	 */
+	wbsd_write_config(host, WBSD_CONF_DEVICE, DEVICE_SD);
+
+	/*
+	 * Read configuration.
+	 */
+	base = wbsd_read_config(host, WBSD_CONF_PORT_HI) << 8;
+	base |= wbsd_read_config(host, WBSD_CONF_PORT_LO);
+
+	irq = wbsd_read_config(host, WBSD_CONF_IRQ);
+
+	dma = wbsd_read_config(host, WBSD_CONF_DRQ);
+
+	wbsd_lock_config(host);
+
+	/*
+	 * Validate against given configuration.
+	 */
+	if (base != host->base)
+		return 0;
+	if (irq != host->irq)
+		return 0;
+	if ((dma != host->dma) && (host->dma != -1))
+		return 0;
+
+	return 1;
+}
+
+/*
+ * Powers down the SD function
+ */
+
+static void wbsd_chip_poweroff(struct wbsd_host *host)
+{
+	wbsd_unlock_config(host);
+
+	wbsd_write_config(host, WBSD_CONF_DEVICE, DEVICE_SD);
+	wbsd_write_config(host, WBSD_CONF_ENABLE, 0);
+
+	wbsd_lock_config(host);
+}
+
+/*****************************************************************************\
+ *                                                                           *
+ * Devices setup and shutdown                                                *
+ *                                                                           *
+\*****************************************************************************/
+
+static int wbsd_init(struct device *dev, int base, int irq, int dma,
+	int pnp)
+{
+	struct wbsd_host *host = NULL;
+	struct mmc_host *mmc = NULL;
+	int ret;
+
+	ret = wbsd_alloc_mmc(dev);
+	if (ret)
+		return ret;
+
+	mmc = dev_get_drvdata(dev);
+	host = mmc_priv(mmc);
+
+	/*
+	 * Scan for hardware.
+	 */
+	ret = wbsd_scan(host);
+	if (ret) {
+		if (pnp && (ret == -ENODEV)) {
+			pr_warn(DRIVER_NAME ": Unable to confirm device presence - you may experience lock-ups\n");
+		} else {
+			wbsd_free_mmc(dev);
+			return ret;
+		}
+	}
+
+	/*
+	 * Request resources.
+	 */
+	ret = wbsd_request_resources(host, base, irq, dma);
+	if (ret) {
+		wbsd_release_resources(host);
+		wbsd_free_mmc(dev);
+		return ret;
+	}
+
+	/*
+	 * See if chip needs to be configured.
+	 */
+	if (pnp) {
+		if ((host->config != 0) && !wbsd_chip_validate(host)) {
+			pr_warn(DRIVER_NAME ": PnP active but chip not configured! You probably have a buggy BIOS. Configuring chip manually.\n");
+			wbsd_chip_config(host);
+		}
+	} else
+		wbsd_chip_config(host);
+
+	/*
+	 * Power Management stuff. No idea how this works.
+	 * Not tested.
+	 */
+#ifdef CONFIG_PM
+	if (host->config) {
+		wbsd_unlock_config(host);
+		wbsd_write_config(host, WBSD_CONF_PME, 0xA0);
+		wbsd_lock_config(host);
+	}
+#endif
+	/*
+	 * Allow device to initialise itself properly.
+	 */
+	mdelay(5);
+
+	/*
+	 * Reset the chip into a known state.
+	 */
+	wbsd_init_device(host);
+
+	ret = mmc_add_host(mmc);
+	if (ret) {
+		if (!pnp)
+			wbsd_chip_poweroff(host);
+
+		wbsd_release_resources(host);
+		wbsd_free_mmc(dev);
+		return ret;
+	}
+
+	pr_info("%s: W83L51xD", mmc_hostname(mmc));
+	if (host->chip_id != 0)
+		printk(" id %x", (int)host->chip_id);
+	printk(" at 0x%x irq %d", (int)host->base, (int)host->irq);
+	if (host->dma >= 0)
+		printk(" dma %d", (int)host->dma);
+	else
+		printk(" FIFO");
+	if (pnp)
+		printk(" PnP");
+	printk("\n");
+
+	return 0;
+}
+
+static void wbsd_shutdown(struct device *dev, int pnp)
+{
+	struct mmc_host *mmc = dev_get_drvdata(dev);
+	struct wbsd_host *host;
+
+	if (!mmc)
+		return;
+
+	host = mmc_priv(mmc);
+
+	mmc_remove_host(mmc);
+
+	/*
+	 * Power down the SD/MMC function.
+	 */
+	if (!pnp)
+		wbsd_chip_poweroff(host);
+
+	wbsd_release_resources(host);
+
+	wbsd_free_mmc(dev);
+}
+
+/*
+ * Non-PnP
+ */
+
+static int wbsd_probe(struct platform_device *dev)
+{
+	/* Use the module parameters for resources */
+	return wbsd_init(&dev->dev, param_io, param_irq, param_dma, 0);
+}
+
+static int wbsd_remove(struct platform_device *dev)
+{
+	wbsd_shutdown(&dev->dev, 0);
+
+	return 0;
+}
+
+/*
+ * PnP
+ */
+
+#ifdef CONFIG_PNP
+
+static int
+wbsd_pnp_probe(struct pnp_dev *pnpdev, const struct pnp_device_id *dev_id)
+{
+	int io, irq, dma;
+
+	/*
+	 * Get resources from PnP layer.
+	 */
+	io = pnp_port_start(pnpdev, 0);
+	irq = pnp_irq(pnpdev, 0);
+	if (pnp_dma_valid(pnpdev, 0))
+		dma = pnp_dma(pnpdev, 0);
+	else
+		dma = -1;
+
+	DBGF("PnP resources: port %3x irq %d dma %d\n", io, irq, dma);
+
+	return wbsd_init(&pnpdev->dev, io, irq, dma, 1);
+}
+
+static void wbsd_pnp_remove(struct pnp_dev *dev)
+{
+	wbsd_shutdown(&dev->dev, 1);
+}
+
+#endif /* CONFIG_PNP */
+
+/*
+ * Power management
+ */
+
+#ifdef CONFIG_PM
+
+static int wbsd_platform_suspend(struct platform_device *dev,
+				 pm_message_t state)
+{
+	struct mmc_host *mmc = platform_get_drvdata(dev);
+	struct wbsd_host *host;
+
+	if (mmc == NULL)
+		return 0;
+
+	DBGF("Suspending...\n");
+
+	host = mmc_priv(mmc);
+
+	wbsd_chip_poweroff(host);
+	return 0;
+}
+
+static int wbsd_platform_resume(struct platform_device *dev)
+{
+	struct mmc_host *mmc = platform_get_drvdata(dev);
+	struct wbsd_host *host;
+
+	if (mmc == NULL)
+		return 0;
+
+	DBGF("Resuming...\n");
+
+	host = mmc_priv(mmc);
+
+	wbsd_chip_config(host);
+
+	/*
+	 * Allow device to initialise itself properly.
+	 */
+	mdelay(5);
+
+	wbsd_init_device(host);
+	return 0;
+}
+
+#ifdef CONFIG_PNP
+
+static int wbsd_pnp_suspend(struct pnp_dev *pnp_dev, pm_message_t state)
+{
+	struct mmc_host *mmc = dev_get_drvdata(&pnp_dev->dev);
+
+	if (mmc == NULL)
+		return 0;
+
+	DBGF("Suspending...\n");
+	return 0;
+}
+
+static int wbsd_pnp_resume(struct pnp_dev *pnp_dev)
+{
+	struct mmc_host *mmc = dev_get_drvdata(&pnp_dev->dev);
+	struct wbsd_host *host;
+
+	if (mmc == NULL)
+		return 0;
+
+	DBGF("Resuming...\n");
+
+	host = mmc_priv(mmc);
+
+	/*
+	 * See if chip needs to be configured.
+	 */
+	if (host->config != 0) {
+		if (!wbsd_chip_validate(host)) {
+			pr_warn(DRIVER_NAME ": PnP active but chip not configured! You probably have a buggy BIOS. Configuring chip manually.\n");
+			wbsd_chip_config(host);
+		}
+	}
+
+	/*
+	 * Allow device to initialise itself properly.
+	 */
+	mdelay(5);
+
+	wbsd_init_device(host);
+	return 0;
+}
+
+#endif /* CONFIG_PNP */
+
+#else /* CONFIG_PM */
+
+#define wbsd_platform_suspend NULL
+#define wbsd_platform_resume NULL
+
+#define wbsd_pnp_suspend NULL
+#define wbsd_pnp_resume NULL
+
+#endif /* CONFIG_PM */
+
+static struct platform_device *wbsd_device;
+
+static struct platform_driver wbsd_driver = {
+	.probe		= wbsd_probe,
+	.remove		= wbsd_remove,
+
+	.suspend	= wbsd_platform_suspend,
+	.resume		= wbsd_platform_resume,
+	.driver		= {
+		.name	= DRIVER_NAME,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+	},
+};
+
+#ifdef CONFIG_PNP
+
+static struct pnp_driver wbsd_pnp_driver = {
+	.name		= DRIVER_NAME,
+	.id_table	= pnp_dev_table,
+	.probe		= wbsd_pnp_probe,
+	.remove		= wbsd_pnp_remove,
+
+	.suspend	= wbsd_pnp_suspend,
+	.resume		= wbsd_pnp_resume,
+};
+
+#endif /* CONFIG_PNP */
+
+/*
+ * Module loading/unloading
+ */
+
+static int __init wbsd_drv_init(void)
+{
+	int result;
+
+	pr_info(DRIVER_NAME
+		": Winbond W83L51xD SD/MMC card interface driver\n");
+	pr_info(DRIVER_NAME ": Copyright(c) Pierre Ossman\n");
+
+#ifdef CONFIG_PNP
+
+	if (!param_nopnp) {
+		result = pnp_register_driver(&wbsd_pnp_driver);
+		if (result < 0)
+			return result;
+	}
+#endif /* CONFIG_PNP */
+
+	if (param_nopnp) {
+		result = platform_driver_register(&wbsd_driver);
+		if (result < 0)
+			return result;
+
+		wbsd_device = platform_device_alloc(DRIVER_NAME, -1);
+		if (!wbsd_device) {
+			platform_driver_unregister(&wbsd_driver);
+			return -ENOMEM;
+		}
+
+		result = platform_device_add(wbsd_device);
+		if (result) {
+			platform_device_put(wbsd_device);
+			platform_driver_unregister(&wbsd_driver);
+			return result;
+		}
+	}
+
+	return 0;
+}
+
+static void __exit wbsd_drv_exit(void)
+{
+#ifdef CONFIG_PNP
+
+	if (!param_nopnp)
+		pnp_unregister_driver(&wbsd_pnp_driver);
+
+#endif /* CONFIG_PNP */
+
+	if (param_nopnp) {
+		platform_device_unregister(wbsd_device);
+
+		platform_driver_unregister(&wbsd_driver);
+	}
+
+	DBG("unloaded\n");
+}
+
+module_init(wbsd_drv_init);
+module_exit(wbsd_drv_exit);
+#ifdef CONFIG_PNP
+module_param_hw_named(nopnp, param_nopnp, uint, other, 0444);
+#endif
+module_param_hw_named(io, param_io, uint, ioport, 0444);
+module_param_hw_named(irq, param_irq, uint, irq, 0444);
+module_param_hw_named(dma, param_dma, int, dma, 0444);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Pierre Ossman <pierre@ossman.eu>");
+MODULE_DESCRIPTION("Winbond W83L51xD SD/MMC card interface driver");
+
+#ifdef CONFIG_PNP
+MODULE_PARM_DESC(nopnp, "Scan for device instead of relying on PNP. (default 0)");
+#endif
+MODULE_PARM_DESC(io, "I/O base to allocate. Must be 8 byte aligned. (default 0x248)");
+MODULE_PARM_DESC(irq, "IRQ to allocate. (default 6)");
+MODULE_PARM_DESC(dma, "DMA channel to allocate. -1 for no DMA. (default 2)");
diff --git a/drivers/mmc/host/wbsd.h b/drivers/mmc/host/wbsd.h
new file mode 100644
index 000000000..be30b4d8c
--- /dev/null
+++ b/drivers/mmc/host/wbsd.h
@@ -0,0 +1,181 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ *  linux/drivers/mmc/host/wbsd.h - Winbond W83L51xD SD/MMC driver
+ *
+ *  Copyright (C) 2004-2007 Pierre Ossman, All Rights Reserved.
+ */
+
+#define LOCK_CODE		0xAA
+
+#define WBSD_CONF_SWRST		0x02
+#define WBSD_CONF_DEVICE	0x07
+#define WBSD_CONF_ID_HI		0x20
+#define WBSD_CONF_ID_LO		0x21
+#define WBSD_CONF_POWER		0x22
+#define WBSD_CONF_PME		0x23
+#define WBSD_CONF_PMES		0x24
+
+#define WBSD_CONF_ENABLE	0x30
+#define WBSD_CONF_PORT_HI	0x60
+#define WBSD_CONF_PORT_LO	0x61
+#define WBSD_CONF_IRQ		0x70
+#define WBSD_CONF_DRQ		0x74
+
+#define WBSD_CONF_PINS		0xF0
+
+#define DEVICE_SD		0x03
+
+#define WBSD_PINS_DAT3_HI	0x20
+#define WBSD_PINS_DAT3_OUT	0x10
+#define WBSD_PINS_GP11_HI	0x04
+#define WBSD_PINS_DETECT_GP11	0x02
+#define WBSD_PINS_DETECT_DAT3	0x01
+
+#define WBSD_CMDR		0x00
+#define WBSD_DFR		0x01
+#define WBSD_EIR		0x02
+#define WBSD_ISR		0x03
+#define WBSD_FSR		0x04
+#define WBSD_IDXR		0x05
+#define WBSD_DATAR		0x06
+#define WBSD_CSR		0x07
+
+#define WBSD_EINT_CARD		0x40
+#define WBSD_EINT_FIFO_THRE	0x20
+#define WBSD_EINT_CRC		0x10
+#define WBSD_EINT_TIMEOUT	0x08
+#define WBSD_EINT_PROGEND	0x04
+#define WBSD_EINT_BUSYEND	0x02
+#define WBSD_EINT_TC		0x01
+
+#define WBSD_INT_PENDING	0x80
+#define WBSD_INT_CARD		0x40
+#define WBSD_INT_FIFO_THRE	0x20
+#define WBSD_INT_CRC		0x10
+#define WBSD_INT_TIMEOUT	0x08
+#define WBSD_INT_PROGEND	0x04
+#define WBSD_INT_BUSYEND	0x02
+#define WBSD_INT_TC		0x01
+
+#define WBSD_FIFO_EMPTY		0x80
+#define WBSD_FIFO_FULL		0x40
+#define WBSD_FIFO_EMTHRE	0x20
+#define WBSD_FIFO_FUTHRE	0x10
+#define WBSD_FIFO_SZMASK	0x0F
+
+#define WBSD_MSLED		0x20
+#define WBSD_POWER_N		0x10
+#define WBSD_WRPT		0x04
+#define WBSD_CARDPRESENT	0x01
+
+#define WBSD_IDX_CLK		0x01
+#define WBSD_IDX_PBSMSB		0x02
+#define WBSD_IDX_TAAC		0x03
+#define WBSD_IDX_NSAC		0x04
+#define WBSD_IDX_PBSLSB		0x05
+#define WBSD_IDX_SETUP		0x06
+#define WBSD_IDX_DMA		0x07
+#define WBSD_IDX_FIFOEN		0x08
+#define WBSD_IDX_STATUS		0x10
+#define WBSD_IDX_RSPLEN		0x1E
+#define WBSD_IDX_RESP0		0x1F
+#define WBSD_IDX_RESP1		0x20
+#define WBSD_IDX_RESP2		0x21
+#define WBSD_IDX_RESP3		0x22
+#define WBSD_IDX_RESP4		0x23
+#define WBSD_IDX_RESP5		0x24
+#define WBSD_IDX_RESP6		0x25
+#define WBSD_IDX_RESP7		0x26
+#define WBSD_IDX_RESP8		0x27
+#define WBSD_IDX_RESP9		0x28
+#define WBSD_IDX_RESP10		0x29
+#define WBSD_IDX_RESP11		0x2A
+#define WBSD_IDX_RESP12		0x2B
+#define WBSD_IDX_RESP13		0x2C
+#define WBSD_IDX_RESP14		0x2D
+#define WBSD_IDX_RESP15		0x2E
+#define WBSD_IDX_RESP16		0x2F
+#define WBSD_IDX_CRCSTATUS	0x30
+#define WBSD_IDX_ISR		0x3F
+
+#define WBSD_CLK_375K		0x00
+#define WBSD_CLK_12M		0x01
+#define WBSD_CLK_16M		0x02
+#define WBSD_CLK_24M		0x03
+
+#define WBSD_DATA_WIDTH		0x01
+
+#define WBSD_DAT3_H		0x08
+#define WBSD_FIFO_RESET		0x04
+#define WBSD_SOFT_RESET		0x02
+#define WBSD_INC_INDEX		0x01
+
+#define WBSD_DMA_SINGLE		0x02
+#define WBSD_DMA_ENABLE		0x01
+
+#define WBSD_FIFOEN_EMPTY	0x20
+#define WBSD_FIFOEN_FULL	0x10
+#define WBSD_FIFO_THREMASK	0x0F
+
+#define WBSD_BLOCK_READ		0x80
+#define WBSD_BLOCK_WRITE	0x40
+#define WBSD_BUSY		0x20
+#define WBSD_CARDTRAFFIC	0x04
+#define WBSD_SENDCMD		0x02
+#define WBSD_RECVRES		0x01
+
+#define WBSD_RSP_SHORT		0x00
+#define WBSD_RSP_LONG		0x01
+
+#define WBSD_CRC_MASK		0x1F
+#define WBSD_CRC_OK		0x05 /* S010E (00101) */
+#define WBSD_CRC_FAIL		0x0B /* S101E (01011) */
+
+#define WBSD_DMA_SIZE		65536
+
+struct wbsd_host
+{
+	struct mmc_host*	mmc;		/* MMC structure */
+
+	spinlock_t		lock;		/* Mutex */
+
+	int			flags;		/* Driver states */
+
+#define WBSD_FCARD_PRESENT	(1<<0)		/* Card is present */
+#define WBSD_FIGNORE_DETECT	(1<<1)		/* Ignore card detection */
+
+	struct mmc_request*	mrq;		/* Current request */
+
+	u8			isr;		/* Accumulated ISR */
+
+	struct scatterlist*	cur_sg;		/* Current SG entry */
+	unsigned int		num_sg;		/* Number of entries left */
+
+	unsigned int		offset;		/* Offset into current entry */
+	unsigned int		remain;		/* Data left in curren entry */
+
+	char*			dma_buffer;	/* ISA DMA buffer */
+	dma_addr_t		dma_addr;	/* Physical address for same */
+
+	int			firsterr;	/* See fifo functions */
+
+	u8			clk;		/* Current clock speed */
+	unsigned char		bus_width;	/* Current bus width */
+
+	int			config;		/* Config port */
+	u8			unlock_code;	/* Code to unlock config */
+
+	int			chip_id;	/* ID of controller */
+
+	int			base;		/* I/O port base */
+	int			irq;		/* Interrupt */
+	int			dma;		/* DMA channel */
+
+	struct tasklet_struct	card_tasklet;	/* Tasklet structures */
+	struct tasklet_struct	fifo_tasklet;
+	struct tasklet_struct	crc_tasklet;
+	struct tasklet_struct	timeout_tasklet;
+	struct tasklet_struct	finish_tasklet;
+
+	struct timer_list	ignore_timer;	/* Ignore detection timer */
+};
diff --git a/drivers/mmc/host/wmt-sdmmc.c b/drivers/mmc/host/wmt-sdmmc.c
new file mode 100644
index 000000000..393319548
--- /dev/null
+++ b/drivers/mmc/host/wmt-sdmmc.c
@@ -0,0 +1,1011 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ *  WM8505/WM8650 SD/MMC Host Controller
+ *
+ *  Copyright (C) 2010 Tony Prisk
+ *  Copyright (C) 2008 WonderMedia Technologies, Inc.
+ */
+
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/ioport.h>
+#include <linux/errno.h>
+#include <linux/dma-mapping.h>
+#include <linux/delay.h>
+#include <linux/io.h>
+#include <linux/irq.h>
+#include <linux/clk.h>
+#include <linux/interrupt.h>
+
+#include <linux/of.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_device.h>
+
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/sd.h>
+
+#include <asm/byteorder.h>
+
+
+#define DRIVER_NAME "wmt-sdhc"
+
+
+/* MMC/SD controller registers */
+#define SDMMC_CTLR			0x00
+#define SDMMC_CMD			0x01
+#define SDMMC_RSPTYPE			0x02
+#define SDMMC_ARG			0x04
+#define SDMMC_BUSMODE			0x08
+#define SDMMC_BLKLEN			0x0C
+#define SDMMC_BLKCNT			0x0E
+#define SDMMC_RSP			0x10
+#define SDMMC_CBCR			0x20
+#define SDMMC_INTMASK0			0x24
+#define SDMMC_INTMASK1			0x25
+#define SDMMC_STS0			0x28
+#define SDMMC_STS1			0x29
+#define SDMMC_STS2			0x2A
+#define SDMMC_STS3			0x2B
+#define SDMMC_RSPTIMEOUT		0x2C
+#define SDMMC_CLK			0x30	/* VT8500 only */
+#define SDMMC_EXTCTRL			0x34
+#define SDMMC_SBLKLEN			0x38
+#define SDMMC_DMATIMEOUT		0x3C
+
+
+/* SDMMC_CTLR bit fields */
+#define CTLR_CMD_START			0x01
+#define CTLR_CMD_WRITE			0x04
+#define CTLR_FIFO_RESET			0x08
+
+/* SDMMC_BUSMODE bit fields */
+#define BM_SPI_MODE			0x01
+#define BM_FOURBIT_MODE			0x02
+#define BM_EIGHTBIT_MODE		0x04
+#define BM_SD_OFF			0x10
+#define BM_SPI_CS			0x20
+#define BM_SD_POWER			0x40
+#define BM_SOFT_RESET			0x80
+
+/* SDMMC_BLKLEN bit fields */
+#define BLKL_CRCERR_ABORT		0x0800
+#define BLKL_CD_POL_HIGH		0x1000
+#define BLKL_GPI_CD			0x2000
+#define BLKL_DATA3_CD			0x4000
+#define BLKL_INT_ENABLE			0x8000
+
+/* SDMMC_INTMASK0 bit fields */
+#define INT0_MBLK_TRAN_DONE_INT_EN	0x10
+#define INT0_BLK_TRAN_DONE_INT_EN	0x20
+#define INT0_CD_INT_EN			0x40
+#define INT0_DI_INT_EN			0x80
+
+/* SDMMC_INTMASK1 bit fields */
+#define INT1_CMD_RES_TRAN_DONE_INT_EN	0x02
+#define INT1_CMD_RES_TOUT_INT_EN	0x04
+#define INT1_MBLK_AUTO_STOP_INT_EN	0x08
+#define INT1_DATA_TOUT_INT_EN		0x10
+#define INT1_RESCRC_ERR_INT_EN		0x20
+#define INT1_RCRC_ERR_INT_EN		0x40
+#define INT1_WCRC_ERR_INT_EN		0x80
+
+/* SDMMC_STS0 bit fields */
+#define STS0_WRITE_PROTECT		0x02
+#define STS0_CD_DATA3			0x04
+#define STS0_CD_GPI			0x08
+#define STS0_MBLK_DONE			0x10
+#define STS0_BLK_DONE			0x20
+#define STS0_CARD_DETECT		0x40
+#define STS0_DEVICE_INS			0x80
+
+/* SDMMC_STS1 bit fields */
+#define STS1_SDIO_INT			0x01
+#define STS1_CMDRSP_DONE		0x02
+#define STS1_RSP_TIMEOUT		0x04
+#define STS1_AUTOSTOP_DONE		0x08
+#define STS1_DATA_TIMEOUT		0x10
+#define STS1_RSP_CRC_ERR		0x20
+#define STS1_RCRC_ERR			0x40
+#define STS1_WCRC_ERR			0x80
+
+/* SDMMC_STS2 bit fields */
+#define STS2_CMD_RES_BUSY		0x10
+#define STS2_DATARSP_BUSY		0x20
+#define STS2_DIS_FORCECLK		0x80
+
+/* SDMMC_EXTCTRL bit fields */
+#define EXT_EIGHTBIT			0x04
+
+/* MMC/SD DMA Controller Registers */
+#define SDDMA_GCR			0x100
+#define SDDMA_IER			0x104
+#define SDDMA_ISR			0x108
+#define SDDMA_DESPR			0x10C
+#define SDDMA_RBR			0x110
+#define SDDMA_DAR			0x114
+#define SDDMA_BAR			0x118
+#define SDDMA_CPR			0x11C
+#define SDDMA_CCR			0x120
+
+
+/* SDDMA_GCR bit fields */
+#define DMA_GCR_DMA_EN			0x00000001
+#define DMA_GCR_SOFT_RESET		0x00000100
+
+/* SDDMA_IER bit fields */
+#define DMA_IER_INT_EN			0x00000001
+
+/* SDDMA_ISR bit fields */
+#define DMA_ISR_INT_STS			0x00000001
+
+/* SDDMA_RBR bit fields */
+#define DMA_RBR_FORMAT			0x40000000
+#define DMA_RBR_END			0x80000000
+
+/* SDDMA_CCR bit fields */
+#define DMA_CCR_RUN			0x00000080
+#define DMA_CCR_IF_TO_PERIPHERAL	0x00000000
+#define DMA_CCR_PERIPHERAL_TO_IF	0x00400000
+
+/* SDDMA_CCR event status */
+#define DMA_CCR_EVT_NO_STATUS		0x00000000
+#define DMA_CCR_EVT_UNDERRUN		0x00000001
+#define DMA_CCR_EVT_OVERRUN		0x00000002
+#define DMA_CCR_EVT_DESP_READ		0x00000003
+#define DMA_CCR_EVT_DATA_RW		0x00000004
+#define DMA_CCR_EVT_EARLY_END		0x00000005
+#define DMA_CCR_EVT_SUCCESS		0x0000000F
+
+#define PDMA_READ			0x00
+#define PDMA_WRITE			0x01
+
+#define WMT_SD_POWER_OFF		0
+#define WMT_SD_POWER_ON			1
+
+struct wmt_dma_descriptor {
+	u32 flags;
+	u32 data_buffer_addr;
+	u32 branch_addr;
+	u32 reserved1;
+};
+
+struct wmt_mci_caps {
+	unsigned int	f_min;
+	unsigned int	f_max;
+	u32		ocr_avail;
+	u32		caps;
+	u32		max_seg_size;
+	u32		max_segs;
+	u32		max_blk_size;
+};
+
+struct wmt_mci_priv {
+	struct mmc_host *mmc;
+	void __iomem *sdmmc_base;
+
+	int irq_regular;
+	int irq_dma;
+
+	void *dma_desc_buffer;
+	dma_addr_t dma_desc_device_addr;
+
+	struct completion cmdcomp;
+	struct completion datacomp;
+
+	struct completion *comp_cmd;
+	struct completion *comp_dma;
+
+	struct mmc_request *req;
+	struct mmc_command *cmd;
+
+	struct clk *clk_sdmmc;
+	struct device *dev;
+
+	u8 power_inverted;
+	u8 cd_inverted;
+};
+
+static void wmt_set_sd_power(struct wmt_mci_priv *priv, int enable)
+{
+	u32 reg_tmp = readb(priv->sdmmc_base + SDMMC_BUSMODE);
+
+	if (enable ^ priv->power_inverted)
+		reg_tmp &= ~BM_SD_OFF;
+	else
+		reg_tmp |= BM_SD_OFF;
+
+	writeb(reg_tmp, priv->sdmmc_base + SDMMC_BUSMODE);
+}
+
+static void wmt_mci_read_response(struct mmc_host *mmc)
+{
+	struct wmt_mci_priv *priv;
+	int idx1, idx2;
+	u8 tmp_resp;
+	u32 response;
+
+	priv = mmc_priv(mmc);
+
+	for (idx1 = 0; idx1 < 4; idx1++) {
+		response = 0;
+		for (idx2 = 0; idx2 < 4; idx2++) {
+			if ((idx1 == 3) && (idx2 == 3))
+				tmp_resp = readb(priv->sdmmc_base + SDMMC_RSP);
+			else
+				tmp_resp = readb(priv->sdmmc_base + SDMMC_RSP +
+						 (idx1*4) + idx2 + 1);
+			response |= (tmp_resp << (idx2 * 8));
+		}
+		priv->cmd->resp[idx1] = cpu_to_be32(response);
+	}
+}
+
+static void wmt_mci_start_command(struct wmt_mci_priv *priv)
+{
+	u32 reg_tmp;
+
+	reg_tmp = readb(priv->sdmmc_base + SDMMC_CTLR);
+	writeb(reg_tmp | CTLR_CMD_START, priv->sdmmc_base + SDMMC_CTLR);
+}
+
+static int wmt_mci_send_command(struct mmc_host *mmc, u8 command, u8 cmdtype,
+				u32 arg, u8 rsptype)
+{
+	struct wmt_mci_priv *priv;
+	u32 reg_tmp;
+
+	priv = mmc_priv(mmc);
+
+	/* write command, arg, resptype registers */
+	writeb(command, priv->sdmmc_base + SDMMC_CMD);
+	writel(arg, priv->sdmmc_base + SDMMC_ARG);
+	writeb(rsptype, priv->sdmmc_base + SDMMC_RSPTYPE);
+
+	/* reset response FIFO */
+	reg_tmp = readb(priv->sdmmc_base + SDMMC_CTLR);
+	writeb(reg_tmp | CTLR_FIFO_RESET, priv->sdmmc_base + SDMMC_CTLR);
+
+	/* ensure clock enabled - VT3465 */
+	wmt_set_sd_power(priv, WMT_SD_POWER_ON);
+
+	/* clear status bits */
+	writeb(0xFF, priv->sdmmc_base + SDMMC_STS0);
+	writeb(0xFF, priv->sdmmc_base + SDMMC_STS1);
+	writeb(0xFF, priv->sdmmc_base + SDMMC_STS2);
+	writeb(0xFF, priv->sdmmc_base + SDMMC_STS3);
+
+	/* set command type */
+	reg_tmp = readb(priv->sdmmc_base + SDMMC_CTLR);
+	writeb((reg_tmp & 0x0F) | (cmdtype << 4),
+	       priv->sdmmc_base + SDMMC_CTLR);
+
+	return 0;
+}
+
+static void wmt_mci_disable_dma(struct wmt_mci_priv *priv)
+{
+	writel(DMA_ISR_INT_STS, priv->sdmmc_base + SDDMA_ISR);
+	writel(0, priv->sdmmc_base + SDDMA_IER);
+}
+
+static void wmt_complete_data_request(struct wmt_mci_priv *priv)
+{
+	struct mmc_request *req;
+	req = priv->req;
+
+	req->data->bytes_xfered = req->data->blksz * req->data->blocks;
+
+	/* unmap the DMA pages used for write data */
+	if (req->data->flags & MMC_DATA_WRITE)
+		dma_unmap_sg(mmc_dev(priv->mmc), req->data->sg,
+			     req->data->sg_len, DMA_TO_DEVICE);
+	else
+		dma_unmap_sg(mmc_dev(priv->mmc), req->data->sg,
+			     req->data->sg_len, DMA_FROM_DEVICE);
+
+	/* Check if the DMA ISR returned a data error */
+	if ((req->cmd->error) || (req->data->error))
+		mmc_request_done(priv->mmc, req);
+	else {
+		wmt_mci_read_response(priv->mmc);
+		if (!req->data->stop) {
+			/* single-block read/write requests end here */
+			mmc_request_done(priv->mmc, req);
+		} else {
+			/*
+			 * we change the priv->cmd variable so the response is
+			 * stored in the stop struct rather than the original
+			 * calling command struct
+			 */
+			priv->comp_cmd = &priv->cmdcomp;
+			init_completion(priv->comp_cmd);
+			priv->cmd = req->data->stop;
+			wmt_mci_send_command(priv->mmc, req->data->stop->opcode,
+					     7, req->data->stop->arg, 9);
+			wmt_mci_start_command(priv);
+		}
+	}
+}
+
+static irqreturn_t wmt_mci_dma_isr(int irq_num, void *data)
+{
+	struct wmt_mci_priv *priv;
+
+	int status;
+
+	priv = (struct wmt_mci_priv *)data;
+
+	status = readl(priv->sdmmc_base + SDDMA_CCR) & 0x0F;
+
+	if (status != DMA_CCR_EVT_SUCCESS) {
+		dev_err(priv->dev, "DMA Error: Status = %d\n", status);
+		priv->req->data->error = -ETIMEDOUT;
+		complete(priv->comp_dma);
+		return IRQ_HANDLED;
+	}
+
+	priv->req->data->error = 0;
+
+	wmt_mci_disable_dma(priv);
+
+	complete(priv->comp_dma);
+
+	if (priv->comp_cmd) {
+		if (completion_done(priv->comp_cmd)) {
+			/*
+			 * if the command (regular) interrupt has already
+			 * completed, finish off the request otherwise we wait
+			 * for the command interrupt and finish from there.
+			 */
+			wmt_complete_data_request(priv);
+		}
+	}
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t wmt_mci_regular_isr(int irq_num, void *data)
+{
+	struct wmt_mci_priv *priv;
+	u32 status0;
+	u32 status1;
+	u32 status2;
+	u32 reg_tmp;
+	int cmd_done;
+
+	priv = (struct wmt_mci_priv *)data;
+	cmd_done = 0;
+	status0 = readb(priv->sdmmc_base + SDMMC_STS0);
+	status1 = readb(priv->sdmmc_base + SDMMC_STS1);
+	status2 = readb(priv->sdmmc_base + SDMMC_STS2);
+
+	/* Check for card insertion */
+	reg_tmp = readb(priv->sdmmc_base + SDMMC_INTMASK0);
+	if ((reg_tmp & INT0_DI_INT_EN) && (status0 & STS0_DEVICE_INS)) {
+		mmc_detect_change(priv->mmc, 0);
+		if (priv->cmd)
+			priv->cmd->error = -ETIMEDOUT;
+		if (priv->comp_cmd)
+			complete(priv->comp_cmd);
+		if (priv->comp_dma) {
+			wmt_mci_disable_dma(priv);
+			complete(priv->comp_dma);
+		}
+		writeb(STS0_DEVICE_INS, priv->sdmmc_base + SDMMC_STS0);
+		return IRQ_HANDLED;
+	}
+
+	if ((!priv->req->data) ||
+	    ((priv->req->data->stop) && (priv->cmd == priv->req->data->stop))) {
+		/* handle non-data & stop_transmission requests */
+		if (status1 & STS1_CMDRSP_DONE) {
+			priv->cmd->error = 0;
+			cmd_done = 1;
+		} else if ((status1 & STS1_RSP_TIMEOUT) ||
+			   (status1 & STS1_DATA_TIMEOUT)) {
+			priv->cmd->error = -ETIMEDOUT;
+			cmd_done = 1;
+		}
+
+		if (cmd_done) {
+			priv->comp_cmd = NULL;
+
+			if (!priv->cmd->error)
+				wmt_mci_read_response(priv->mmc);
+
+			priv->cmd = NULL;
+
+			mmc_request_done(priv->mmc, priv->req);
+		}
+	} else {
+		/* handle data requests */
+		if (status1 & STS1_CMDRSP_DONE) {
+			if (priv->cmd)
+				priv->cmd->error = 0;
+			if (priv->comp_cmd)
+				complete(priv->comp_cmd);
+		}
+
+		if ((status1 & STS1_RSP_TIMEOUT) ||
+		    (status1 & STS1_DATA_TIMEOUT)) {
+			if (priv->cmd)
+				priv->cmd->error = -ETIMEDOUT;
+			if (priv->comp_cmd)
+				complete(priv->comp_cmd);
+			if (priv->comp_dma) {
+				wmt_mci_disable_dma(priv);
+				complete(priv->comp_dma);
+			}
+		}
+
+		if (priv->comp_dma) {
+			/*
+			 * If the dma interrupt has already completed, finish
+			 * off the request; otherwise we wait for the DMA
+			 * interrupt and finish from there.
+			 */
+			if (completion_done(priv->comp_dma))
+				wmt_complete_data_request(priv);
+		}
+	}
+
+	writeb(status0, priv->sdmmc_base + SDMMC_STS0);
+	writeb(status1, priv->sdmmc_base + SDMMC_STS1);
+	writeb(status2, priv->sdmmc_base + SDMMC_STS2);
+
+	return IRQ_HANDLED;
+}
+
+static void wmt_reset_hardware(struct mmc_host *mmc)
+{
+	struct wmt_mci_priv *priv;
+	u32 reg_tmp;
+
+	priv = mmc_priv(mmc);
+
+	/* reset controller */
+	reg_tmp = readb(priv->sdmmc_base + SDMMC_BUSMODE);
+	writeb(reg_tmp | BM_SOFT_RESET, priv->sdmmc_base + SDMMC_BUSMODE);
+
+	/* reset response FIFO */
+	reg_tmp = readb(priv->sdmmc_base + SDMMC_CTLR);
+	writeb(reg_tmp | CTLR_FIFO_RESET, priv->sdmmc_base + SDMMC_CTLR);
+
+	/* enable GPI pin to detect card */
+	writew(BLKL_INT_ENABLE | BLKL_GPI_CD, priv->sdmmc_base + SDMMC_BLKLEN);
+
+	/* clear interrupt status */
+	writeb(0xFF, priv->sdmmc_base + SDMMC_STS0);
+	writeb(0xFF, priv->sdmmc_base + SDMMC_STS1);
+
+	/* setup interrupts */
+	writeb(INT0_CD_INT_EN | INT0_DI_INT_EN, priv->sdmmc_base +
+	       SDMMC_INTMASK0);
+	writeb(INT1_DATA_TOUT_INT_EN | INT1_CMD_RES_TRAN_DONE_INT_EN |
+	       INT1_CMD_RES_TOUT_INT_EN, priv->sdmmc_base + SDMMC_INTMASK1);
+
+	/* set the DMA timeout */
+	writew(8191, priv->sdmmc_base + SDMMC_DMATIMEOUT);
+
+	/* auto clock freezing enable */
+	reg_tmp = readb(priv->sdmmc_base + SDMMC_STS2);
+	writeb(reg_tmp | STS2_DIS_FORCECLK, priv->sdmmc_base + SDMMC_STS2);
+
+	/* set a default clock speed of 400Khz */
+	clk_set_rate(priv->clk_sdmmc, 400000);
+}
+
+static int wmt_dma_init(struct mmc_host *mmc)
+{
+	struct wmt_mci_priv *priv;
+
+	priv = mmc_priv(mmc);
+
+	writel(DMA_GCR_SOFT_RESET, priv->sdmmc_base + SDDMA_GCR);
+	writel(DMA_GCR_DMA_EN, priv->sdmmc_base + SDDMA_GCR);
+	if ((readl(priv->sdmmc_base + SDDMA_GCR) & DMA_GCR_DMA_EN) != 0)
+		return 0;
+	else
+		return 1;
+}
+
+static void wmt_dma_init_descriptor(struct wmt_dma_descriptor *desc,
+		u16 req_count, u32 buffer_addr, u32 branch_addr, int end)
+{
+	desc->flags = 0x40000000 | req_count;
+	if (end)
+		desc->flags |= 0x80000000;
+	desc->data_buffer_addr = buffer_addr;
+	desc->branch_addr = branch_addr;
+}
+
+static void wmt_dma_config(struct mmc_host *mmc, u32 descaddr, u8 dir)
+{
+	struct wmt_mci_priv *priv;
+	u32 reg_tmp;
+
+	priv = mmc_priv(mmc);
+
+	/* Enable DMA Interrupts */
+	writel(DMA_IER_INT_EN, priv->sdmmc_base + SDDMA_IER);
+
+	/* Write DMA Descriptor Pointer Register */
+	writel(descaddr, priv->sdmmc_base + SDDMA_DESPR);
+
+	writel(0x00, priv->sdmmc_base + SDDMA_CCR);
+
+	if (dir == PDMA_WRITE) {
+		reg_tmp = readl(priv->sdmmc_base + SDDMA_CCR);
+		writel(reg_tmp & DMA_CCR_IF_TO_PERIPHERAL, priv->sdmmc_base +
+		       SDDMA_CCR);
+	} else {
+		reg_tmp = readl(priv->sdmmc_base + SDDMA_CCR);
+		writel(reg_tmp | DMA_CCR_PERIPHERAL_TO_IF, priv->sdmmc_base +
+		       SDDMA_CCR);
+	}
+}
+
+static void wmt_dma_start(struct wmt_mci_priv *priv)
+{
+	u32 reg_tmp;
+
+	reg_tmp = readl(priv->sdmmc_base + SDDMA_CCR);
+	writel(reg_tmp | DMA_CCR_RUN, priv->sdmmc_base + SDDMA_CCR);
+}
+
+static void wmt_mci_request(struct mmc_host *mmc, struct mmc_request *req)
+{
+	struct wmt_mci_priv *priv;
+	struct wmt_dma_descriptor *desc;
+	u8 command;
+	u8 cmdtype;
+	u32 arg;
+	u8 rsptype;
+	u32 reg_tmp;
+
+	struct scatterlist *sg;
+	int i;
+	int sg_cnt;
+	int offset;
+	u32 dma_address;
+	int desc_cnt;
+
+	priv = mmc_priv(mmc);
+	priv->req = req;
+
+	/*
+	 * Use the cmd variable to pass a pointer to the resp[] structure
+	 * This is required on multi-block requests to pass the pointer to the
+	 * stop command
+	 */
+	priv->cmd = req->cmd;
+
+	command = req->cmd->opcode;
+	arg = req->cmd->arg;
+	rsptype = mmc_resp_type(req->cmd);
+	cmdtype = 0;
+
+	/* rsptype=7 only valid for SPI commands - should be =2 for SD */
+	if (rsptype == 7)
+		rsptype = 2;
+	/* rsptype=21 is R1B, convert for controller */
+	if (rsptype == 21)
+		rsptype = 9;
+
+	if (!req->data) {
+		wmt_mci_send_command(mmc, command, cmdtype, arg, rsptype);
+		wmt_mci_start_command(priv);
+		/* completion is now handled in the regular_isr() */
+	}
+	if (req->data) {
+		priv->comp_cmd = &priv->cmdcomp;
+		init_completion(priv->comp_cmd);
+
+		wmt_dma_init(mmc);
+
+		/* set controller data length */
+		reg_tmp = readw(priv->sdmmc_base + SDMMC_BLKLEN);
+		writew((reg_tmp & 0xF800) | (req->data->blksz - 1),
+		       priv->sdmmc_base + SDMMC_BLKLEN);
+
+		/* set controller block count */
+		writew(req->data->blocks, priv->sdmmc_base + SDMMC_BLKCNT);
+
+		desc = (struct wmt_dma_descriptor *)priv->dma_desc_buffer;
+
+		if (req->data->flags & MMC_DATA_WRITE) {
+			sg_cnt = dma_map_sg(mmc_dev(mmc), req->data->sg,
+					    req->data->sg_len, DMA_TO_DEVICE);
+			cmdtype = 1;
+			if (req->data->blocks > 1)
+				cmdtype = 3;
+		} else {
+			sg_cnt = dma_map_sg(mmc_dev(mmc), req->data->sg,
+					    req->data->sg_len, DMA_FROM_DEVICE);
+			cmdtype = 2;
+			if (req->data->blocks > 1)
+				cmdtype = 4;
+		}
+
+		dma_address = priv->dma_desc_device_addr + 16;
+		desc_cnt = 0;
+
+		for_each_sg(req->data->sg, sg, sg_cnt, i) {
+			offset = 0;
+			while (offset < sg_dma_len(sg)) {
+				wmt_dma_init_descriptor(desc, req->data->blksz,
+						sg_dma_address(sg)+offset,
+						dma_address, 0);
+				desc++;
+				desc_cnt++;
+				offset += req->data->blksz;
+				dma_address += 16;
+				if (desc_cnt == req->data->blocks)
+					break;
+			}
+		}
+		desc--;
+		desc->flags |= 0x80000000;
+
+		if (req->data->flags & MMC_DATA_WRITE)
+			wmt_dma_config(mmc, priv->dma_desc_device_addr,
+				       PDMA_WRITE);
+		else
+			wmt_dma_config(mmc, priv->dma_desc_device_addr,
+				       PDMA_READ);
+
+		wmt_mci_send_command(mmc, command, cmdtype, arg, rsptype);
+
+		priv->comp_dma = &priv->datacomp;
+		init_completion(priv->comp_dma);
+
+		wmt_dma_start(priv);
+		wmt_mci_start_command(priv);
+	}
+}
+
+static void wmt_mci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct wmt_mci_priv *priv;
+	u32 busmode, extctrl;
+
+	priv = mmc_priv(mmc);
+
+	if (ios->power_mode == MMC_POWER_UP) {
+		wmt_reset_hardware(mmc);
+
+		wmt_set_sd_power(priv, WMT_SD_POWER_ON);
+	}
+	if (ios->power_mode == MMC_POWER_OFF)
+		wmt_set_sd_power(priv, WMT_SD_POWER_OFF);
+
+	if (ios->clock != 0)
+		clk_set_rate(priv->clk_sdmmc, ios->clock);
+
+	busmode = readb(priv->sdmmc_base + SDMMC_BUSMODE);
+	extctrl = readb(priv->sdmmc_base + SDMMC_EXTCTRL);
+
+	busmode &= ~(BM_EIGHTBIT_MODE | BM_FOURBIT_MODE);
+	extctrl &= ~EXT_EIGHTBIT;
+
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_8:
+		busmode |= BM_EIGHTBIT_MODE;
+		extctrl |= EXT_EIGHTBIT;
+		break;
+	case MMC_BUS_WIDTH_4:
+		busmode |= BM_FOURBIT_MODE;
+		break;
+	case MMC_BUS_WIDTH_1:
+		break;
+	}
+
+	writeb(busmode, priv->sdmmc_base + SDMMC_BUSMODE);
+	writeb(extctrl, priv->sdmmc_base + SDMMC_EXTCTRL);
+}
+
+static int wmt_mci_get_ro(struct mmc_host *mmc)
+{
+	struct wmt_mci_priv *priv = mmc_priv(mmc);
+
+	return !(readb(priv->sdmmc_base + SDMMC_STS0) & STS0_WRITE_PROTECT);
+}
+
+static int wmt_mci_get_cd(struct mmc_host *mmc)
+{
+	struct wmt_mci_priv *priv = mmc_priv(mmc);
+	u32 cd = (readb(priv->sdmmc_base + SDMMC_STS0) & STS0_CD_GPI) >> 3;
+
+	return !(cd ^ priv->cd_inverted);
+}
+
+static const struct mmc_host_ops wmt_mci_ops = {
+	.request = wmt_mci_request,
+	.set_ios = wmt_mci_set_ios,
+	.get_ro = wmt_mci_get_ro,
+	.get_cd = wmt_mci_get_cd,
+};
+
+/* Controller capabilities */
+static struct wmt_mci_caps wm8505_caps = {
+	.f_min = 390425,
+	.f_max = 50000000,
+	.ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34,
+	.caps = MMC_CAP_4_BIT_DATA | MMC_CAP_MMC_HIGHSPEED |
+		MMC_CAP_SD_HIGHSPEED,
+	.max_seg_size = 65024,
+	.max_segs = 128,
+	.max_blk_size = 2048,
+};
+
+static const struct of_device_id wmt_mci_dt_ids[] = {
+	{ .compatible = "wm,wm8505-sdhc", .data = &wm8505_caps },
+	{ /* Sentinel */ },
+};
+
+static int wmt_mci_probe(struct platform_device *pdev)
+{
+	struct mmc_host *mmc;
+	struct wmt_mci_priv *priv;
+	struct device_node *np = pdev->dev.of_node;
+	const struct of_device_id *of_id =
+		of_match_device(wmt_mci_dt_ids, &pdev->dev);
+	const struct wmt_mci_caps *wmt_caps;
+	int ret;
+	int regular_irq, dma_irq;
+
+	if (!of_id || !of_id->data) {
+		dev_err(&pdev->dev, "Controller capabilities data missing\n");
+		return -EFAULT;
+	}
+
+	wmt_caps = of_id->data;
+
+	if (!np) {
+		dev_err(&pdev->dev, "Missing SDMMC description in devicetree\n");
+		return -EFAULT;
+	}
+
+	regular_irq = irq_of_parse_and_map(np, 0);
+	dma_irq = irq_of_parse_and_map(np, 1);
+
+	if (!regular_irq || !dma_irq) {
+		dev_err(&pdev->dev, "Getting IRQs failed!\n");
+		ret = -ENXIO;
+		goto fail1;
+	}
+
+	mmc = mmc_alloc_host(sizeof(struct wmt_mci_priv), &pdev->dev);
+	if (!mmc) {
+		dev_err(&pdev->dev, "Failed to allocate mmc_host\n");
+		ret = -ENOMEM;
+		goto fail1;
+	}
+
+	mmc->ops = &wmt_mci_ops;
+	mmc->f_min = wmt_caps->f_min;
+	mmc->f_max = wmt_caps->f_max;
+	mmc->ocr_avail = wmt_caps->ocr_avail;
+	mmc->caps = wmt_caps->caps;
+
+	mmc->max_seg_size = wmt_caps->max_seg_size;
+	mmc->max_segs = wmt_caps->max_segs;
+	mmc->max_blk_size = wmt_caps->max_blk_size;
+
+	mmc->max_req_size = (16*512*mmc->max_segs);
+	mmc->max_blk_count = mmc->max_req_size / 512;
+
+	priv = mmc_priv(mmc);
+	priv->mmc = mmc;
+	priv->dev = &pdev->dev;
+
+	priv->power_inverted = 0;
+	priv->cd_inverted = 0;
+
+	if (of_get_property(np, "sdon-inverted", NULL))
+		priv->power_inverted = 1;
+	if (of_get_property(np, "cd-inverted", NULL))
+		priv->cd_inverted = 1;
+
+	priv->sdmmc_base = of_iomap(np, 0);
+	if (!priv->sdmmc_base) {
+		dev_err(&pdev->dev, "Failed to map IO space\n");
+		ret = -ENOMEM;
+		goto fail2;
+	}
+
+	priv->irq_regular = regular_irq;
+	priv->irq_dma = dma_irq;
+
+	ret = request_irq(regular_irq, wmt_mci_regular_isr, 0, "sdmmc", priv);
+	if (ret) {
+		dev_err(&pdev->dev, "Register regular IRQ fail\n");
+		goto fail3;
+	}
+
+	ret = request_irq(dma_irq, wmt_mci_dma_isr, 0, "sdmmc", priv);
+	if (ret) {
+		dev_err(&pdev->dev, "Register DMA IRQ fail\n");
+		goto fail4;
+	}
+
+	/* alloc some DMA buffers for descriptors/transfers */
+	priv->dma_desc_buffer = dma_alloc_coherent(&pdev->dev,
+						   mmc->max_blk_count * 16,
+						   &priv->dma_desc_device_addr,
+						   GFP_KERNEL);
+	if (!priv->dma_desc_buffer) {
+		dev_err(&pdev->dev, "DMA alloc fail\n");
+		ret = -EPERM;
+		goto fail5;
+	}
+
+	platform_set_drvdata(pdev, mmc);
+
+	priv->clk_sdmmc = of_clk_get(np, 0);
+	if (IS_ERR(priv->clk_sdmmc)) {
+		dev_err(&pdev->dev, "Error getting clock\n");
+		ret = PTR_ERR(priv->clk_sdmmc);
+		goto fail5_and_a_half;
+	}
+
+	ret = clk_prepare_enable(priv->clk_sdmmc);
+	if (ret)
+		goto fail6;
+
+	/* configure the controller to a known 'ready' state */
+	wmt_reset_hardware(mmc);
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto fail7;
+
+	dev_info(&pdev->dev, "WMT SDHC Controller initialized\n");
+
+	return 0;
+fail7:
+	clk_disable_unprepare(priv->clk_sdmmc);
+fail6:
+	clk_put(priv->clk_sdmmc);
+fail5_and_a_half:
+	dma_free_coherent(&pdev->dev, mmc->max_blk_count * 16,
+			  priv->dma_desc_buffer, priv->dma_desc_device_addr);
+fail5:
+	free_irq(dma_irq, priv);
+fail4:
+	free_irq(regular_irq, priv);
+fail3:
+	iounmap(priv->sdmmc_base);
+fail2:
+	mmc_free_host(mmc);
+fail1:
+	return ret;
+}
+
+static int wmt_mci_remove(struct platform_device *pdev)
+{
+	struct mmc_host *mmc;
+	struct wmt_mci_priv *priv;
+	struct resource *res;
+	u32 reg_tmp;
+
+	mmc = platform_get_drvdata(pdev);
+	priv = mmc_priv(mmc);
+
+	/* reset SD controller */
+	reg_tmp = readb(priv->sdmmc_base + SDMMC_BUSMODE);
+	writel(reg_tmp | BM_SOFT_RESET, priv->sdmmc_base + SDMMC_BUSMODE);
+	reg_tmp = readw(priv->sdmmc_base + SDMMC_BLKLEN);
+	writew(reg_tmp & ~(0xA000), priv->sdmmc_base + SDMMC_BLKLEN);
+	writeb(0xFF, priv->sdmmc_base + SDMMC_STS0);
+	writeb(0xFF, priv->sdmmc_base + SDMMC_STS1);
+
+	/* release the dma buffers */
+	dma_free_coherent(&pdev->dev, priv->mmc->max_blk_count * 16,
+			  priv->dma_desc_buffer, priv->dma_desc_device_addr);
+
+	mmc_remove_host(mmc);
+
+	free_irq(priv->irq_regular, priv);
+	free_irq(priv->irq_dma, priv);
+
+	iounmap(priv->sdmmc_base);
+
+	clk_disable_unprepare(priv->clk_sdmmc);
+	clk_put(priv->clk_sdmmc);
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	release_mem_region(res->start, resource_size(res));
+
+	mmc_free_host(mmc);
+
+	dev_info(&pdev->dev, "WMT MCI device removed\n");
+
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static int wmt_mci_suspend(struct device *dev)
+{
+	u32 reg_tmp;
+	struct mmc_host *mmc = dev_get_drvdata(dev);
+	struct wmt_mci_priv *priv;
+
+	if (!mmc)
+		return 0;
+
+	priv = mmc_priv(mmc);
+	reg_tmp = readb(priv->sdmmc_base + SDMMC_BUSMODE);
+	writeb(reg_tmp | BM_SOFT_RESET, priv->sdmmc_base +
+	       SDMMC_BUSMODE);
+
+	reg_tmp = readw(priv->sdmmc_base + SDMMC_BLKLEN);
+	writew(reg_tmp & 0x5FFF, priv->sdmmc_base + SDMMC_BLKLEN);
+
+	writeb(0xFF, priv->sdmmc_base + SDMMC_STS0);
+	writeb(0xFF, priv->sdmmc_base + SDMMC_STS1);
+
+	clk_disable(priv->clk_sdmmc);
+	return 0;
+}
+
+static int wmt_mci_resume(struct device *dev)
+{
+	u32 reg_tmp;
+	struct mmc_host *mmc = dev_get_drvdata(dev);
+	struct wmt_mci_priv *priv;
+
+	if (mmc) {
+		priv = mmc_priv(mmc);
+		clk_enable(priv->clk_sdmmc);
+
+		reg_tmp = readb(priv->sdmmc_base + SDMMC_BUSMODE);
+		writeb(reg_tmp | BM_SOFT_RESET, priv->sdmmc_base +
+		       SDMMC_BUSMODE);
+
+		reg_tmp = readw(priv->sdmmc_base + SDMMC_BLKLEN);
+		writew(reg_tmp | (BLKL_GPI_CD | BLKL_INT_ENABLE),
+		       priv->sdmmc_base + SDMMC_BLKLEN);
+
+		reg_tmp = readb(priv->sdmmc_base + SDMMC_INTMASK0);
+		writeb(reg_tmp | INT0_DI_INT_EN, priv->sdmmc_base +
+		       SDMMC_INTMASK0);
+
+	}
+
+	return 0;
+}
+
+static const struct dev_pm_ops wmt_mci_pm = {
+	.suspend        = wmt_mci_suspend,
+	.resume         = wmt_mci_resume,
+};
+
+#define wmt_mci_pm_ops (&wmt_mci_pm)
+
+#else	/* !CONFIG_PM */
+
+#define wmt_mci_pm_ops NULL
+
+#endif
+
+static struct platform_driver wmt_mci_driver = {
+	.probe = wmt_mci_probe,
+	.remove = wmt_mci_remove,
+	.driver = {
+		.name = DRIVER_NAME,
+		.probe_type = PROBE_PREFER_ASYNCHRONOUS,
+		.pm = wmt_mci_pm_ops,
+		.of_match_table = wmt_mci_dt_ids,
+	},
+};
+
+module_platform_driver(wmt_mci_driver);
+
+MODULE_DESCRIPTION("Wondermedia MMC/SD Driver");
+MODULE_AUTHOR("Tony Prisk");
+MODULE_LICENSE("GPL v2");
+MODULE_DEVICE_TABLE(of, wmt_mci_dt_ids);