Initial import

87 files changed, 63000 insertions, 0 deletions

diff --git a/drivers/mmc/host/Kconfig b/drivers/mmc/host/Kconfig
new file mode 100644
index 000000000..5fd5824c3
--- /dev/null
+++ b/drivers/mmc/host/Kconfig
@@ -0,0 +1,771 @@
+#
+# MMC/SD host controller drivers
+#
+
+comment "MMC/SD/SDIO Host Controller Drivers"
+
+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
+	tristate "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_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 Freescale eSDHC and Nintendo Wii SDHCI.
+
+config MMC_SDHCI_PCI
+	tristate "SDHCI support on PCI bus"
+	depends on MMC_SDHCI && PCI
+	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
+	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
+	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_ESDHC
+	tristate "SDHCI OF support for the Freescale eSDHC controller"
+	depends on MMC_SDHCI_PLTFM
+	depends on PPC
+	select MMC_SDHCI_BIG_ENDIAN_32BIT_BYTE_SWAPPER
+	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_CNS3XXX
+	tristate "SDHCI support on the Cavium Networks CNS3xxx SoC"
+	depends on ARCH_CNS3XXX
+	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_IMX
+	tristate "SDHCI support for the Freescale eSDHC/uSDHC i.MX controller"
+	depends on ARCH_MXC
+	depends on MMC_SDHCI_PLTFM
+	select MMC_SDHCI_IO_ACCESSORS
+	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
+	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
+	depends on MMC_SDHCI_PLTFM
+	select MMC_SDHCI_IO_ACCESSORS
+	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 && PLAT_SAMSUNG
+	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
+	depends on MMC_SDHCI_PLTFM
+	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 && PLAT_SPEAR
+	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
+	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_BCM2835
+	tristate "SDHCI platform support for the BCM2835 SD/MMC Controller"
+	depends on ARCH_BCM2835
+	depends on MMC_SDHCI_PLTFM
+	select MMC_SDHCI_IO_ACCESSORS
+	help
+	  This selects the BCM2835 SD/MMC controller. If you have a BCM2835
+	  platform with SD or MMC devices, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_SDHCI_F_SDH30
+	tristate "SDHCI support for Fujitsu Semiconductor F_SDH30"
+	depends on MMC_SDHCI_PLTFM
+	depends on OF
+	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_IPROC
+	tristate "SDHCI platform support for the iProc SD/MMC Controller"
+	depends on ARCH_BCM_IPROC || COMPILE_TEST
+	depends on MMC_SDHCI_PLTFM
+	default ARCH_BCM_IPROC
+	select MMC_SDHCI_IO_ACCESSORS
+	help
+	  This selects the iProc SD/MMC controller.
+
+	  If you have an IPROC platform with SD or MMC devices,
+	  say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_MOXART
+	tristate "MOXART SD/MMC Host Controller support"
+	depends on ARCH_MOXART && MMC
+	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
+	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 HAS_DMA
+	depends on ARCH_OMAP2PLUS || 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
+	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_AU1X
+	tristate "Alchemy AU1XX0 MMC Card Interface support"
+	depends on MIPS_ALCHEMY
+	help
+	  This selects the AMD Alchemy(R) Multimedia card interface.
+	  If you have a Alchemy platform with a MMC slot, say Y or M here.
+
+	  If unsure, say N.
+
+config MMC_ATMELMCI
+	tristate "Atmel SD/MMC Driver (Multimedia Card Interface)"
+	depends on AVR32 || ARCH_AT91
+	help
+	  This selects the Atmel Multimedia Card Interface driver. If
+	  you have an AT32 (AVR32) or 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 || (ARM && COMPILE_TEST)
+	depends on MMC_SDHCI_PLTFM
+	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
+	---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
+        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
+	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 && !HIGHMEM && HAS_DMA
+	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_TMIO_CORE
+	tristate
+
+config MMC_TMIO
+	tristate "Toshiba Mobile IO Controller (TMIO) MMC/SD function support"
+	depends on MFD_TMIO || MFD_ASIC3
+	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 "SH-Mobile SDHI SD/SDIO controller support"
+	depends on SUPERH || ARM
+	depends on SUPERH || ARCH_SHMOBILE || COMPILE_TEST
+	select MMC_TMIO_CORE
+	help
+	  This provides support for the SDHI SD/SDIO controller found in
+	  SuperH and ARM SH-Mobile 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 SDH_BFIN
+	tristate "Blackfin Secure Digital Host support"
+	depends on (BF54x && !BF544) || (BF51x && !BF512)
+	help
+	  If you say yes here you will get support for the Blackfin on-chip
+	  Secure Digital Host interface.  This includes support for MMC and
+	  SD cards.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called bfin_sdh.
+
+	  If unsure, say N.
+
+config SDH_BFIN_MISSING_CMD_PULLUP_WORKAROUND
+	bool "Blackfin EZkit Missing SDH_CMD Pull Up Resistor Workaround"
+	depends on SDH_BFIN
+	help
+	  If you say yes here SD-Cards may work on the EZkit.
+
+config MMC_DW
+	tristate "Synopsys DesignWare Memory Card Interface"
+	depends on HAS_DMA
+	depends on ARC || ARM || ARM64 || MIPS || 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 and external DMA modes.
+
+config MMC_DW_IDMAC
+	bool "Internal DMAC interface"
+	depends on MMC_DW
+	help
+	  This selects support for the internal DMAC block within the Synopsys
+	  Designware Mobile Storage IP block. This disables the external DMA
+	  interface.
+
+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_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_K3
+	tristate "K3 specific extensions for Synopsys DW Memory Card Interface"
+	depends on MMC_DW
+	select MMC_DW_PLTFM
+	select MMC_DW_IDMAC
+	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_SH_MMCIF
+	tristate "SuperH Internal MMCIF support"
+	depends on MMC_BLOCK && HAS_DMA
+	depends on SUPERH || ARCH_SHMOBILE || COMPILE_TEST
+	help
+	  This selects the MMC Host Interface controller (MMCIF).
+
+	  This driver supports MMCIF in sh7724/sh7757/sh7372.
+
+config MMC_JZ4740
+	tristate "JZ4740 SD/Multimedia Card Interface support"
+	depends on MACH_JZ4740
+	help
+	  This selects support for the SD/MMC controller on Ingenic JZ4740
+	  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.
+
+	  /*(DEBLOBBED)*/
+
+	  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 MFD_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 MFD_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
+	help
+	  This selects support for the SD/MMC Host Controller on
+	  Allwinner sunxi SoCs.
+
+config MMC_TOSHIBA_PCI
+	tristate "Toshiba Type A SD/MMC Card Interface Driver"
+	depends on PCI
+	help
diff --git a/drivers/mmc/host/Makefile b/drivers/mmc/host/Makefile
new file mode 100644
index 000000000..e3ab5b968
--- /dev/null
+++ b/drivers/mmc/host/Makefile
@@ -0,0 +1,79 @@
+#
+# Makefile for MMC/SD host controller drivers
+#
+
+obj-$(CONFIG_MMC_ARMMMCI)	+= mmci.o
+obj-$(CONFIG_MMC_QCOM_DML)	+= mmci_qcom_dml.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-$(CONFIG_MMC_SDHCI_PCI)	+= sdhci-pci.o
+obj-$(subst m,y,$(CONFIG_MMC_SDHCI_PCI))	+= sdhci-pci-data.o
+obj-$(subst m,y,$(CONFIG_MMC_SDHCI_PCI))	+= sdhci-pci-o2micro.o
+obj-$(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_SPEAR)	+= sdhci-spear.o
+obj-$(CONFIG_MMC_WBSD)		+= wbsd.o
+obj-$(CONFIG_MMC_AU1X)		+= au1xmmc.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
+tmio_mmc_core-y			:= tmio_mmc_pio.o
+tmio_mmc_core-$(subst m,y,$(CONFIG_MMC_SDHI))	+= tmio_mmc_dma.o
+obj-$(CONFIG_MMC_SDHI)		+= sh_mobile_sdhi.o
+obj-$(CONFIG_MMC_CB710)		+= cb710-mmc.o
+obj-$(CONFIG_MMC_VIA_SDMMC)	+= via-sdmmc.o
+obj-$(CONFIG_SDH_BFIN)		+= bfin_sdh.o
+obj-$(CONFIG_MMC_DW)		+= dw_mmc.o
+obj-$(CONFIG_MMC_DW_PLTFM)	+= dw_mmc-pltfm.o
+obj-$(CONFIG_MMC_DW_EXYNOS)	+= dw_mmc-exynos.o
+obj-$(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-$(CONFIG_MMC_SH_MMCIF)	+= sh_mmcif.o
+obj-$(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_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_REALTEK_PCI)	+= rtsx_pci_sdmmc.o
+obj-$(CONFIG_MMC_REALTEK_USB)	+= rtsx_usb_sdmmc.o
+
+obj-$(CONFIG_MMC_SDHCI_PLTFM)		+= sdhci-pltfm.o
+obj-$(CONFIG_MMC_SDHCI_CNS3XXX)		+= sdhci-cns3xxx.o
+obj-$(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_ESDHC)	+= sdhci-of-esdhc.o
+obj-$(CONFIG_MMC_SDHCI_OF_HLWD)		+= sdhci-of-hlwd.o
+obj-$(CONFIG_MMC_SDHCI_BCM_KONA)	+= sdhci-bcm-kona.o
+obj-$(CONFIG_MMC_SDHCI_BCM2835)		+= sdhci-bcm2835.o
+obj-$(CONFIG_MMC_SDHCI_IPROC)		+= sdhci-iproc.o
+obj-$(CONFIG_MMC_SDHCI_MSM)		+= sdhci-msm.o
+obj-$(CONFIG_MMC_SDHCI_ST)		+= sdhci-st.o
+
+ifeq ($(CONFIG_CB710_DEBUG),y)
+	CFLAGS-cb710-mmc	+= -DDEBUG
+endif
diff --git a/drivers/mmc/host/android-goldfish.c b/drivers/mmc/host/android-goldfish.c
new file mode 100644
index 000000000..8b4e20a3f
--- /dev/null
+++ b/drivers/mmc/host/android-goldfish.c
@@ -0,0 +1,568 @@
+/*
+ *  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
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#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/sdio.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/scatterlist.h>
+
+#include <asm/types.h>
+#include <asm/io.h>
+#include <asm/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 mmc_host		*mmc;
+	struct device		*dev;
+	unsigned char		id; /* 16xx chips have 2 MMC blocks */
+	void __iomem		*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(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->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;
+
+		if (data->flags & MMC_DATA_WRITE)
+			dma_data_dir = DMA_TO_DEVICE;
+		else
+			dma_data_dir = DMA_FROM_DEVICE;
+
+		if (dma_data_dir == DMA_FROM_DEVICE) {
+			/*
+			 * We don't really have DMA, so we need
+			 * to copy from our platform driver buffer
+			 */
+			uint8_t *dest = (uint8_t *)sg_virt(data->sg);
+			memcpy(dest, host->virt_base, data->sg->length);
+		}
+		host->data->bytes_xfered += data->sg->length;
+		dma_unmap_sg(mmc_dev(host->mmc), 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(host->mmc, 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(host->mmc, 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 transfer_error = 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(host->mmc, mrq);
+	}
+
+	if (end_command)
+		goldfish_mmc_cmd_done(host, host->cmd);
+
+	if (transfer_error)
+		goldfish_mmc_xfer_done(host, host->data);
+	else 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(host->mmc, 0);
+	}
+
+	if (!end_command && !end_transfer &&
+	    !transfer_error && !state_changed && !cmd_timeout) {
+		status = GOLDFISH_MMC_READ(host, MMC_INT_STATUS);
+		dev_info(mmc_dev(host->mmc),"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;
+
+	if (data->flags & MMC_DATA_WRITE)
+		dma_data_dir = DMA_TO_DEVICE;
+	else
+		dma_data_dir = DMA_FROM_DEVICE;
+
+	host->sg_len = dma_map_sg(mmc_dev(host->mmc), 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
+		 */
+		const uint8_t *src = (uint8_t *)sg_virt(data->sg);
+		memcpy(host->virt_base, src, 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);
+
+	/*
+	 * This is to avoid accidentally being detected as an SDIO card
+	 * in mmc_attach_sdio().
+	 */
+	if (req->cmd->opcode == SD_IO_SEND_OP_COND &&
+	    req->cmd->flags == (MMC_RSP_SPI_R4 | MMC_RSP_R4 | MMC_CMD_BCR))
+		req->cmd->error = -EINVAL;
+}
+
+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);
+	host->mmc = 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;
+
+	/* 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(host->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(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);
+
+	BUG_ON(host == NULL);
+
+	mmc_remove_host(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(host->mmc);
+	return 0;
+}
+
+static struct platform_driver goldfish_mmc_driver = {
+	.probe		= goldfish_mmc_probe,
+	.remove		= goldfish_mmc_remove,
+	.driver		= {
+		.name	= DRIVER_NAME,
+	},
+};
+
+module_platform_driver(goldfish_mmc_driver);
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/atmel-mci-regs.h b/drivers/mmc/host/atmel-mci-regs.h
new file mode 100644
index 000000000..0aa44e679
--- /dev/null
+++ b/drivers/mmc/host/atmel-mci-regs.h
@@ -0,0 +1,171 @@
+/*
+ * Atmel MultiMedia Card Interface driver
+ *
+ * Copyright (C) 2004-2006 Atmel Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+/*
+ * Superset of MCI IP registers integrated in Atmel AVR32 and AT91 Processors
+ * Registers and bitfields marked with [2] are only available in MCI2
+ */
+
+#ifndef __DRIVERS_MMC_ATMEL_MCI_H__
+#define __DRIVERS_MMC_ATMEL_MCI_H__
+
+/* MCI Register Definitions */
+#define ATMCI_CR			0x0000	/* Control */
+# define ATMCI_CR_MCIEN			(  1 <<  0)	/* MCI Enable */
+# define ATMCI_CR_MCIDIS		(  1 <<  1)	/* MCI Disable */
+# define ATMCI_CR_PWSEN			(  1 <<  2)	/* Power Save Enable */
+# define ATMCI_CR_PWSDIS		(  1 <<  3)	/* Power Save Disable */
+# define ATMCI_CR_SWRST			(  1 <<  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		(  1 << 11)	/* Read Proof */
+# define ATMCI_MR_WRPROOF		(  1 << 12)	/* Write Proof */
+# define ATMCI_MR_PDCFBYTE		(  1 << 13)	/* Force Byte Transfer */
+# define ATMCI_MR_PDCPADV		(  1 << 14)	/* Padding Value */
+# define ATMCI_MR_PDCMODE		(  1 << 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			(  1 <<   0)	/* Command Ready */
+# define ATMCI_RXRDY			(  1 <<   1)	/* Receiver Ready */
+# define ATMCI_TXRDY			(  1 <<   2)	/* Transmitter Ready */
+# define ATMCI_BLKE			(  1 <<   3)	/* Data Block Ended */
+# define ATMCI_DTIP			(  1 <<   4)	/* Data Transfer In Progress */
+# define ATMCI_NOTBUSY			(  1 <<   5)	/* Data Not Busy */
+# define ATMCI_ENDRX			(  1 <<   6)    /* End of RX Buffer */
+# define ATMCI_ENDTX			(  1 <<   7)    /* End of TX Buffer */
+# define ATMCI_SDIOIRQA			(  1 <<   8)	/* SDIO IRQ in slot A */
+# define ATMCI_SDIOIRQB			(  1 <<   9)	/* SDIO IRQ in slot B */
+# define ATMCI_SDIOWAIT			(  1 <<  12)    /* SDIO Read Wait Operation Status */
+# define ATMCI_CSRCV			(  1 <<  13)    /* CE-ATA Completion Signal Received */
+# define ATMCI_RXBUFF			(  1 <<  14)    /* RX Buffer Full */
+# define ATMCI_TXBUFE			(  1 <<  15)    /* TX Buffer Empty */
+# define ATMCI_RINDE			(  1 <<  16)	/* Response Index Error */
+# define ATMCI_RDIRE			(  1 <<  17)	/* Response Direction Error */
+# define ATMCI_RCRCE			(  1 <<  18)	/* Response CRC Error */
+# define ATMCI_RENDE			(  1 <<  19)	/* Response End Bit Error */
+# define ATMCI_RTOE			(  1 <<  20)	/* Response Time-Out Error */
+# define ATMCI_DCRCE			(  1 <<  21)	/* Data CRC Error */
+# define ATMCI_DTOE			(  1 <<  22)	/* Data Time-Out Error */
+# define ATMCI_CSTOE			(  1 <<  23)    /* Completion Signal Time-out Error */
+# define ATMCI_BLKOVRE			(  1 <<  24)    /* DMA Block Overrun Error */
+# define ATMCI_DMADONE			(  1 <<  25)    /* DMA Transfer Done */
+# define ATMCI_FIFOEMPTY		(  1 <<  26)    /* FIFO Empty Flag */
+# define ATMCI_XFRDONE			(  1 <<  27)    /* Transfer Done Flag */
+# define ATMCI_ACKRCV			(  1 <<  28)    /* Boot Operation Acknowledge Received */
+# define ATMCI_ACKRCVE			(  1 <<  29)    /* Boot Operation Acknowledge Error */
+# define ATMCI_OVRE			(  1 <<  30)	/* RX Overrun Error */
+# define ATMCI_UNRE			(  1 <<  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			(  1 <<  8)	/* DMA Hardware Handshaking Enable */
+#define ATMCI_CFG			0x0054	/* Configuration[2] */
+# define ATMCI_CFG_FIFOMODE_1DATA	(  1 <<  0)	/* MCI Internal FIFO control mode */
+# define ATMCI_CFG_FERRCTRL_COR		(  1 <<  4)	/* Flow Error flag reset control mode */
+# define ATMCI_CFG_HSMODE		(  1 <<  8)	/* High Speed Mode */
+# define ATMCI_CFG_LSYNC		(  1 << 12)	/* Synchronize on the last block */
+#define ATMCI_WPMR			0x00e4	/* Write Protection Mode[2] */
+# define ATMCI_WP_EN			(  1 <<  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 */
+#ifdef CONFIG_AVR32
+#define atmci_readl(port, reg)			\
+	__raw_readl((port)->regs + reg)
+#define atmci_writel(port, reg, value)			\
+	__raw_writel((value), (port)->regs + reg)
+#else
+#define atmci_readl(port, reg)			\
+	readl_relaxed((port)->regs + reg)
+#define atmci_writel(port, reg, value)			\
+	writel_relaxed((value), (port)->regs + reg)
+#endif
+
+/* On AVR chips the Peripheral DMA Controller is not connected to MCI. */
+#ifdef CONFIG_AVR32
+#	define ATMCI_PDC_CONNECTED	0
+#else
+#	define ATMCI_PDC_CONNECTED	1
+#endif
+
+/*
+ * Fix sconfig's burst size according to atmel MCI. We need to convert them as:
+ * 1 -> 0, 4 -> 1, 8 -> 2, 16 -> 3.
+ *
+ * This can be done by finding most significant bit set.
+ */
+static inline unsigned int atmci_convert_chksize(unsigned int maxburst)
+{
+	if (maxburst > 1)
+		return fls(maxburst) - 2;
+	else
+		return 0;
+}
+
+#endif /* __DRIVERS_MMC_ATMEL_MCI_H__ */
diff --git a/drivers/mmc/host/atmel-mci.c b/drivers/mmc/host/atmel-mci.c
new file mode 100644
index 000000000..9a39e0b7e
--- /dev/null
+++ b/drivers/mmc/host/atmel-mci.c
@@ -0,0 +1,2585 @@
+/*
+ * Atmel MultiMedia Card Interface driver
+ *
+ * Copyright (C) 2004-2008 Atmel Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#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/platform_data/atmel.h>
+#include <linux/platform_data/mmc-atmel-mci.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>
+
+#include "atmel-mci-regs.h"
+
+#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.
+ * @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.
+ * @phys_buf_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.
+ * @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;
+}
+
+static int atmci_req_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, atmci_req_show, inode->i_private);
+}
+
+static const struct file_operations atmci_req_fops = {
+	.owner		= THIS_MODULE,
+	.open		= atmci_req_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+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;
+}
+
+static int atmci_regs_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, atmci_regs_show, inode->i_private);
+}
+
+static const struct file_operations atmci_regs_fops = {
+	.owner		= THIS_MODULE,
+	.open		= atmci_regs_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+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;
+	struct dentry		*node;
+
+	root = mmc->debugfs_root;
+	if (!root)
+		return;
+
+	node = debugfs_create_file("regs", S_IRUSR, root, host,
+			&atmci_regs_fops);
+	if (IS_ERR(node))
+		return;
+	if (!node)
+		goto err;
+
+	node = debugfs_create_file("req", S_IRUSR, root, slot, &atmci_req_fops);
+	if (!node)
+		goto err;
+
+	node = debugfs_create_u32("state", S_IRUSR, root, (u32 *)&host->state);
+	if (!node)
+		goto err;
+
+	node = debugfs_create_x32("pending_events", S_IRUSR, root,
+				     (u32 *)&host->pending_events);
+	if (!node)
+		goto err;
+
+	node = debugfs_create_x32("completed_events", S_IRUSR, root,
+				     (u32 *)&host->completed_events);
+	if (!node)
+		goto err;
+
+	return;
+
+err:
+	dev_err(&mmc->class_dev, "failed to initialize debugfs for slot\n");
+}
+
+#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) {
+		dev_err(&pdev->dev, "could not allocate memory for pdata\n");
+		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 %s\n",
+				 cnp->full_name);
+			continue;
+		}
+
+		if (slot_id >= ATMCI_MAX_NR_SLOTS) {
+			dev_warn(&pdev->dev, "can't have more than %d slots\n",
+			         ATMCI_MAX_NR_SLOTS);
+			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;
+}
+
+static void atmci_timeout_timer(unsigned long data)
+{
+	struct atmel_mci *host;
+
+	host = (struct atmel_mci *)data;
+
+	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->flags & MMC_DATA_STREAM)
+				cmdr |= ATMCI_CMDR_STREAM;
+			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)
+{
+	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);
+}
+
+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,
+				((data->flags & MMC_DATA_WRITE)
+				 ? DMA_TO_DEVICE : DMA_FROM_DEVICE));
+}
+
+/*
+ * 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,
+				((data->flags & MMC_DATA_WRITE)
+				 ? DMA_TO_DEVICE : DMA_FROM_DEVICE));
+}
+
+/*
+ * 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;
+	unsigned int sg_len;
+	enum dma_data_direction dir;
+	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) {
+		dir = DMA_FROM_DEVICE;
+		iflags |= ATMCI_ENDRX | ATMCI_RXBUFF;
+	} else {
+		dir = DMA_TO_DEVICE;
+		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;
+	sg_len = dma_map_sg(&host->pdev->dev, data->sg, data->sg_len, dir);
+
+	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,
+			((dir == DMA_FROM_DEVICE) ? 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_data_direction		direction;
+	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) {
+		direction = DMA_FROM_DEVICE;
+		host->dma_conf.direction = slave_dirn = DMA_DEV_TO_MEM;
+		maxburst = atmci_convert_chksize(host->dma_conf.src_maxburst);
+	} else {
+		direction = DMA_TO_DEVICE;
+		host->dma_conf.direction = slave_dirn = DMA_MEM_TO_DEV;
+		maxburst = atmci_convert_chksize(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, direction);
+
+	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, direction);
+	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;
+		if (data->flags & MMC_DATA_STREAM)
+			host->stop_cmdr |= ATMCI_CMDR_STREAM;
+		else
+			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);
+
+	mod_timer(&host->timer, jiffies +  msecs_to_jiffies(2000));
+}
+
+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);
+
+	pm_runtime_get_sync(&host->pdev->dev);
+
+	/*
+	 * 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;
+
+	pm_runtime_get_sync(&host->pdev->dev);
+
+	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;
+	}
+
+	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:
+		/*
+		 * TODO: None of the currently available AVR32-based
+		 * boards allow MMC power to be turned off. Implement
+		 * power control when this can be tested properly.
+		 *
+		 * We also need to hook this into the clock management
+		 * somehow so that newly inserted cards aren't
+		 * subjected to a fast clock before we have a chance
+		 * to figure out what the maximum rate is. Currently,
+		 * there's no way to avoid this, and there never will
+		 * be for boards that don't support power control.
+		 */
+		break;
+	}
+
+	pm_runtime_mark_last_busy(&host->pdev->dev);
+	pm_runtime_put_autosuspend(&host->pdev->dev);
+}
+
+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);
+
+	/*
+	 * 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;
+	}
+
+	del_timer(&host->timer);
+
+	spin_unlock(&host->lock);
+	mmc_request_done(prev_mmc, mrq);
+	spin_lock(&host->lock);
+
+	pm_runtime_mark_last_busy(&host->pdev->dev);
+	pm_runtime_put_autosuspend(&host->pdev->dev);
+}
+
+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(unsigned long data)
+{
+	struct atmel_mci_slot	*slot = (struct atmel_mci_slot *)data;
+	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_writel(host, ATMCI_IER, ATMCI_CMDRDY);
+				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_writel(host, ATMCI_IER,
+					             ATMCI_CMDRDY);
+					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);
+			state = STATE_IDLE;
+			break;
+		}
+	} while (state != prev_state);
+
+	host->state = state;
+
+	spin_unlock(&host->lock);
+}
+
+static void atmci_read_data_pio(struct atmel_mci *host)
+{
+	struct scatterlist	*sg = host->sg;
+	void			*buf = sg_virt(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)) {
+			put_unaligned(value, (u32 *)(buf + 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;
+				buf = sg_virt(sg);
+			}
+		} else {
+			unsigned int remaining = sg->length - offset;
+			memcpy(buf + offset, &value, remaining);
+			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;
+			buf = sg_virt(sg);
+			memcpy(buf, (u8 *)&value + remaining, offset);
+			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;
+	void			*buf = sg_virt(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)) {
+			value = get_unaligned((u32 *)(buf + 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;
+				buf = sg_virt(sg);
+			}
+		} else {
+			unsigned int remaining = sg->length - offset;
+
+			value = 0;
+			memcpy(&value, buf + offset, remaining);
+			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;
+			buf = sg_virt(sg);
+			memcpy((u8 *)&value + remaining, buf, offset);
+			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;
+
+	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 (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);
+	mmc_add_host(mmc);
+
+	if (gpio_is_valid(slot->detect_pin)) {
+		int ret;
+
+		setup_timer(&slot->detect_timer, atmci_detect_change,
+				(unsigned long)slot);
+
+		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_slave_channel_reason(&host->pdev->dev,
+							"rxtx");
+	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 = ATMCI_PDC_CONNECTED;
+	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;
+	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;
+	case 0x200:
+		host->caps.has_rwproof = 1;
+		host->caps.need_blksz_mul_4 = 0;
+		host->caps.need_notbusy_for_read_ops = 1;
+	case 0x100:
+		host->caps.has_bad_data_ordering = 0;
+		host->caps.need_reset_after_xfer = 0;
+	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);
+
+	setup_timer(&host->timer, atmci_timeout_timer, (unsigned long)host);
+
+	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_pm_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",
+		.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..ed77fbfa4
--- /dev/null
+++ b/drivers/mmc/host/au1xmmc.c
@@ -0,0 +1,1240 @@
+/*
+ * 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.
+ *
+
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+/* 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/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, int wait,
+				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 */
+
+	if (wait)
+		IRQ_OFF(host, SD_CONFIG_CR);
+
+	__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 */;
+
+	/* Wait for the command to come back */
+	if (wait) {
+		u32 status = __raw_readl(HOST_STATUS(host));
+
+		while (!(status & SD_STATUS_CR))
+			status = __raw_readl(HOST_STATUS(host));
+
+		/* Clear the CR status */
+		__raw_writel(SD_STATUS_CR, HOST_STATUS(host));
+
+		IRQ_ON(host, SD_CONFIG_CR);
+	}
+
+	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 = sg_virt(sg) + 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++;
+
+		__raw_writel((unsigned long)val, HOST_TXPORT(host));
+		wmb(); /* drain writebuffer */
+	}
+
+	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 = sg_virt(sg) + 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++ = (unsigned char)(val & 0xFF);
+	}
+
+	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, 0, 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;
+	}
+
+	r = platform_get_resource(pdev, IORESOURCE_IRQ, 0);
+	if (!r) {
+		dev_err(&pdev->dev, "no IRQ defined\n");
+		goto out3;
+	}
+	host->irq = r->start;
+
+	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);
+out_clk:
+	clk_disable_unprepare(host->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,
+	},
+};
+
+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/bfin_sdh.c b/drivers/mmc/host/bfin_sdh.c
new file mode 100644
index 000000000..2b7f37e82
--- /dev/null
+++ b/drivers/mmc/host/bfin_sdh.c
@@ -0,0 +1,682 @@
+/*
+ * bfin_sdh.c - Analog Devices Blackfin SDH Controller
+ *
+ * Copyright (C) 2007-2009 Analog Device Inc.
+ *
+ * Licensed under the GPL-2 or later.
+ */
+
+#define DRIVER_NAME	"bfin-sdh"
+
+#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/dma-mapping.h>
+#include <linux/mmc/host.h>
+#include <linux/proc_fs.h>
+#include <linux/gfp.h>
+
+#include <asm/cacheflush.h>
+#include <asm/dma.h>
+#include <asm/portmux.h>
+#include <asm/bfin_sdh.h>
+
+#if defined(CONFIG_BF51x) || defined(__ADSPBF60x__)
+#define bfin_read_SDH_CLK_CTL		bfin_read_RSI_CLK_CTL
+#define bfin_write_SDH_CLK_CTL		bfin_write_RSI_CLK_CTL
+#define bfin_write_SDH_ARGUMENT		bfin_write_RSI_ARGUMENT
+#define bfin_write_SDH_COMMAND		bfin_write_RSI_COMMAND
+#define bfin_write_SDH_DATA_TIMER	bfin_write_RSI_DATA_TIMER
+#define bfin_read_SDH_RESPONSE0		bfin_read_RSI_RESPONSE0
+#define bfin_read_SDH_RESPONSE1		bfin_read_RSI_RESPONSE1
+#define bfin_read_SDH_RESPONSE2		bfin_read_RSI_RESPONSE2
+#define bfin_read_SDH_RESPONSE3		bfin_read_RSI_RESPONSE3
+#define bfin_write_SDH_DATA_LGTH	bfin_write_RSI_DATA_LGTH
+#define bfin_read_SDH_DATA_CTL		bfin_read_RSI_DATA_CTL
+#define bfin_write_SDH_DATA_CTL		bfin_write_RSI_DATA_CTL
+#define bfin_read_SDH_DATA_CNT		bfin_read_RSI_DATA_CNT
+#define bfin_write_SDH_STATUS_CLR	bfin_write_RSI_STATUS_CLR
+#define bfin_read_SDH_E_STATUS		bfin_read_RSI_E_STATUS
+#define bfin_write_SDH_E_STATUS		bfin_write_RSI_E_STATUS
+#define bfin_read_SDH_STATUS		bfin_read_RSI_STATUS
+#define bfin_write_SDH_MASK0		bfin_write_RSI_MASK0
+#define bfin_write_SDH_E_MASK		bfin_write_RSI_E_MASK
+#define bfin_read_SDH_CFG		bfin_read_RSI_CFG
+#define bfin_write_SDH_CFG		bfin_write_RSI_CFG
+# if defined(__ADSPBF60x__)
+#  define bfin_read_SDH_BLK_SIZE	bfin_read_RSI_BLKSZ
+#  define bfin_write_SDH_BLK_SIZE	bfin_write_RSI_BLKSZ
+# else
+#  define bfin_read_SDH_PWR_CTL		bfin_read_RSI_PWR_CTL
+#  define bfin_write_SDH_PWR_CTL	bfin_write_RSI_PWR_CTL
+# endif
+#endif
+
+struct sdh_host {
+	struct mmc_host		*mmc;
+	spinlock_t		lock;
+	struct resource		*res;
+	void __iomem		*base;
+	int			irq;
+	int			stat_irq;
+	int			dma_ch;
+	int			dma_dir;
+	struct dma_desc_array	*sg_cpu;
+	dma_addr_t		sg_dma;
+	int			dma_len;
+
+	unsigned long		sclk;
+	unsigned int		imask;
+	unsigned int		power_mode;
+	unsigned int		clk_div;
+
+	struct mmc_request	*mrq;
+	struct mmc_command	*cmd;
+	struct mmc_data		*data;
+};
+
+static struct bfin_sd_host *get_sdh_data(struct platform_device *pdev)
+{
+	return pdev->dev.platform_data;
+}
+
+static void sdh_stop_clock(struct sdh_host *host)
+{
+	bfin_write_SDH_CLK_CTL(bfin_read_SDH_CLK_CTL() & ~CLK_E);
+	SSYNC();
+}
+
+static void sdh_enable_stat_irq(struct sdh_host *host, unsigned int mask)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	host->imask |= mask;
+	bfin_write_SDH_MASK0(mask);
+	SSYNC();
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void sdh_disable_stat_irq(struct sdh_host *host, unsigned int mask)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	host->imask &= ~mask;
+	bfin_write_SDH_MASK0(host->imask);
+	SSYNC();
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static int sdh_setup_data(struct sdh_host *host, struct mmc_data *data)
+{
+	unsigned int length;
+	unsigned int data_ctl;
+	unsigned int dma_cfg;
+	unsigned int cycle_ns, timeout;
+
+	dev_dbg(mmc_dev(host->mmc), "%s enter flags: 0x%x\n", __func__, data->flags);
+	host->data = data;
+	data_ctl = 0;
+	dma_cfg = 0;
+
+	length = data->blksz * data->blocks;
+	bfin_write_SDH_DATA_LGTH(length);
+
+	if (data->flags & MMC_DATA_STREAM)
+		data_ctl |= DTX_MODE;
+
+	if (data->flags & MMC_DATA_READ)
+		data_ctl |= DTX_DIR;
+	/* Only supports power-of-2 block size */
+	if (data->blksz & (data->blksz - 1))
+		return -EINVAL;
+#ifndef RSI_BLKSZ
+	data_ctl |= ((ffs(data->blksz) - 1) << 4);
+#else
+        bfin_write_SDH_BLK_SIZE(data->blksz);
+#endif
+
+	bfin_write_SDH_DATA_CTL(data_ctl);
+	/* the time of a host clock period in ns */
+	cycle_ns = 1000000000 / (host->sclk / (2 * (host->clk_div + 1)));
+	timeout = data->timeout_ns / cycle_ns;
+	timeout += data->timeout_clks;
+	bfin_write_SDH_DATA_TIMER(timeout);
+	SSYNC();
+
+	if (data->flags & MMC_DATA_READ) {
+		host->dma_dir = DMA_FROM_DEVICE;
+		dma_cfg |= WNR;
+	} else
+		host->dma_dir = DMA_TO_DEVICE;
+
+	sdh_enable_stat_irq(host, (DAT_CRC_FAIL | DAT_TIME_OUT | DAT_END));
+	host->dma_len = dma_map_sg(mmc_dev(host->mmc), data->sg, data->sg_len, host->dma_dir);
+#if defined(CONFIG_BF54x) || defined(CONFIG_BF60x)
+	dma_cfg |= DMAFLOW_ARRAY | RESTART | WDSIZE_32 | DMAEN;
+# ifdef RSI_BLKSZ
+	dma_cfg |= PSIZE_32 | NDSIZE_3;
+# else
+	dma_cfg |= NDSIZE_5;
+# endif
+	{
+		struct scatterlist *sg;
+		int i;
+		for_each_sg(data->sg, sg, host->dma_len, i) {
+			host->sg_cpu[i].start_addr = sg_dma_address(sg);
+			host->sg_cpu[i].cfg = dma_cfg;
+			host->sg_cpu[i].x_count = sg_dma_len(sg) / 4;
+			host->sg_cpu[i].x_modify = 4;
+			dev_dbg(mmc_dev(host->mmc), "%d: start_addr:0x%lx, "
+				"cfg:0x%lx, x_count:0x%lx, x_modify:0x%lx\n",
+				i, host->sg_cpu[i].start_addr,
+				host->sg_cpu[i].cfg, host->sg_cpu[i].x_count,
+				host->sg_cpu[i].x_modify);
+		}
+	}
+	flush_dcache_range((unsigned int)host->sg_cpu,
+		(unsigned int)host->sg_cpu +
+			host->dma_len * sizeof(struct dma_desc_array));
+	/* Set the last descriptor to stop mode */
+	host->sg_cpu[host->dma_len - 1].cfg &= ~(DMAFLOW | NDSIZE);
+	host->sg_cpu[host->dma_len - 1].cfg |= DI_EN;
+
+	set_dma_curr_desc_addr(host->dma_ch, (unsigned long *)host->sg_dma);
+	set_dma_x_count(host->dma_ch, 0);
+	set_dma_x_modify(host->dma_ch, 0);
+	SSYNC();
+	set_dma_config(host->dma_ch, dma_cfg);
+#elif defined(CONFIG_BF51x)
+	/* RSI DMA doesn't work in array mode */
+	dma_cfg |= WDSIZE_32 | DMAEN;
+	set_dma_start_addr(host->dma_ch, sg_dma_address(&data->sg[0]));
+	set_dma_x_count(host->dma_ch, length / 4);
+	set_dma_x_modify(host->dma_ch, 4);
+	SSYNC();
+	set_dma_config(host->dma_ch, dma_cfg);
+#endif
+	bfin_write_SDH_DATA_CTL(bfin_read_SDH_DATA_CTL() | DTX_DMA_E | DTX_E);
+
+	SSYNC();
+
+	dev_dbg(mmc_dev(host->mmc), "%s exit\n", __func__);
+	return 0;
+}
+
+static void sdh_start_cmd(struct sdh_host *host, struct mmc_command *cmd)
+{
+	unsigned int sdh_cmd;
+	unsigned int stat_mask;
+
+	dev_dbg(mmc_dev(host->mmc), "%s enter cmd: 0x%p\n", __func__, cmd);
+	WARN_ON(host->cmd != NULL);
+	host->cmd = cmd;
+
+	sdh_cmd = 0;
+	stat_mask = 0;
+
+	sdh_cmd |= cmd->opcode;
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		sdh_cmd |= CMD_RSP;
+		stat_mask |= CMD_RESP_END;
+	} else {
+		stat_mask |= CMD_SENT;
+	}
+
+	if (cmd->flags & MMC_RSP_136)
+		sdh_cmd |= CMD_L_RSP;
+
+	stat_mask |= CMD_CRC_FAIL | CMD_TIME_OUT;
+
+	sdh_enable_stat_irq(host, stat_mask);
+
+	bfin_write_SDH_ARGUMENT(cmd->arg);
+	bfin_write_SDH_COMMAND(sdh_cmd | CMD_E);
+	bfin_write_SDH_CLK_CTL(bfin_read_SDH_CLK_CTL() | CLK_E);
+	SSYNC();
+}
+
+static void sdh_finish_request(struct sdh_host *host, struct mmc_request *mrq)
+{
+	dev_dbg(mmc_dev(host->mmc), "%s enter\n", __func__);
+	host->mrq = NULL;
+	host->cmd = NULL;
+	host->data = NULL;
+	mmc_request_done(host->mmc, mrq);
+}
+
+static int sdh_cmd_done(struct sdh_host *host, unsigned int stat)
+{
+	struct mmc_command *cmd = host->cmd;
+	int ret = 0;
+
+	dev_dbg(mmc_dev(host->mmc), "%s enter cmd: %p\n", __func__, cmd);
+	if (!cmd)
+		return 0;
+
+	host->cmd = NULL;
+
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		cmd->resp[0] = bfin_read_SDH_RESPONSE0();
+		if (cmd->flags & MMC_RSP_136) {
+			cmd->resp[1] = bfin_read_SDH_RESPONSE1();
+			cmd->resp[2] = bfin_read_SDH_RESPONSE2();
+			cmd->resp[3] = bfin_read_SDH_RESPONSE3();
+		}
+	}
+	if (stat & CMD_TIME_OUT)
+		cmd->error = -ETIMEDOUT;
+	else if (stat & CMD_CRC_FAIL && cmd->flags & MMC_RSP_CRC)
+		cmd->error = -EILSEQ;
+
+	sdh_disable_stat_irq(host, (CMD_SENT | CMD_RESP_END | CMD_TIME_OUT | CMD_CRC_FAIL));
+
+	if (host->data && !cmd->error) {
+		if (host->data->flags & MMC_DATA_WRITE) {
+			ret = sdh_setup_data(host, host->data);
+			if (ret)
+				return 0;
+		}
+
+		sdh_enable_stat_irq(host, DAT_END | RX_OVERRUN | TX_UNDERRUN | DAT_TIME_OUT);
+	} else
+		sdh_finish_request(host, host->mrq);
+
+	return 1;
+}
+
+static int sdh_data_done(struct sdh_host *host, unsigned int stat)
+{
+	struct mmc_data *data = host->data;
+
+	dev_dbg(mmc_dev(host->mmc), "%s enter stat: 0x%x\n", __func__, stat);
+	if (!data)
+		return 0;
+
+	disable_dma(host->dma_ch);
+	dma_unmap_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+		     host->dma_dir);
+
+	if (stat & DAT_TIME_OUT)
+		data->error = -ETIMEDOUT;
+	else if (stat & DAT_CRC_FAIL)
+		data->error = -EILSEQ;
+	else if (stat & (RX_OVERRUN | TX_UNDERRUN))
+		data->error = -EIO;
+
+	if (!data->error)
+		data->bytes_xfered = data->blocks * data->blksz;
+	else
+		data->bytes_xfered = 0;
+
+	bfin_write_SDH_STATUS_CLR(DAT_END_STAT | DAT_TIMEOUT_STAT | \
+			DAT_CRC_FAIL_STAT | DAT_BLK_END_STAT | RX_OVERRUN | TX_UNDERRUN);
+	bfin_write_SDH_DATA_CTL(0);
+	SSYNC();
+
+	host->data = NULL;
+	if (host->mrq->stop) {
+		sdh_stop_clock(host);
+		sdh_start_cmd(host, host->mrq->stop);
+	} else {
+		sdh_finish_request(host, host->mrq);
+	}
+
+	return 1;
+}
+
+static void sdh_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct sdh_host *host = mmc_priv(mmc);
+	int ret = 0;
+
+	dev_dbg(mmc_dev(host->mmc), "%s enter, mrp:%p, cmd:%p\n", __func__, mrq, mrq->cmd);
+	WARN_ON(host->mrq != NULL);
+
+	spin_lock(&host->lock);
+	host->mrq = mrq;
+	host->data = mrq->data;
+
+	if (mrq->data && mrq->data->flags & MMC_DATA_READ) {
+		ret = sdh_setup_data(host, mrq->data);
+		if (ret)
+			goto data_err;
+	}
+
+	sdh_start_cmd(host, mrq->cmd);
+data_err:
+	spin_unlock(&host->lock);
+}
+
+static void sdh_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sdh_host *host;
+	u16 clk_ctl = 0;
+#ifndef RSI_BLKSZ
+	u16 pwr_ctl = 0;
+#endif
+	u16 cfg;
+	host = mmc_priv(mmc);
+
+	spin_lock(&host->lock);
+
+	cfg = bfin_read_SDH_CFG();
+	cfg |= MWE;
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_4:
+#ifndef RSI_BLKSZ
+		cfg &= ~PD_SDDAT3;
+#endif
+		cfg |= PUP_SDDAT3;
+		/* Enable 4 bit SDIO */
+		cfg |= SD4E;
+		clk_ctl |= WIDE_BUS_4;
+		break;
+	case MMC_BUS_WIDTH_8:
+#ifndef RSI_BLKSZ
+		cfg &= ~PD_SDDAT3;
+#endif
+		cfg |= PUP_SDDAT3;
+		/* Disable 4 bit SDIO */
+		cfg &= ~SD4E;
+		clk_ctl |= BYTE_BUS_8;
+		break;
+	default:
+		cfg &= ~PUP_SDDAT3;
+		/* Disable 4 bit SDIO */
+		cfg &= ~SD4E;
+	}
+	bfin_write_SDH_CFG(cfg);
+
+	host->power_mode = ios->power_mode;
+#ifndef RSI_BLKSZ
+	if (ios->bus_mode == MMC_BUSMODE_OPENDRAIN) {
+		pwr_ctl |= ROD_CTL;
+# ifndef CONFIG_SDH_BFIN_MISSING_CMD_PULLUP_WORKAROUND
+		pwr_ctl |= SD_CMD_OD;
+# endif
+	}
+
+	if (ios->power_mode != MMC_POWER_OFF)
+		pwr_ctl |= PWR_ON;
+	else
+		pwr_ctl &= ~PWR_ON;
+
+	bfin_write_SDH_PWR_CTL(pwr_ctl);
+#else
+# ifndef CONFIG_SDH_BFIN_MISSING_CMD_PULLUP_WORKAROUND
+	if (ios->bus_mode == MMC_BUSMODE_OPENDRAIN)
+		cfg |= SD_CMD_OD;
+	else
+		cfg &= ~SD_CMD_OD;
+# endif
+
+	if (ios->power_mode != MMC_POWER_OFF)
+		cfg |= PWR_ON;
+	else
+		cfg &= ~PWR_ON;
+
+	bfin_write_SDH_CFG(cfg);
+#endif
+	SSYNC();
+
+	if (ios->power_mode == MMC_POWER_ON && ios->clock) {
+		unsigned char clk_div;
+		clk_div = (get_sclk() / ios->clock - 1) / 2;
+		clk_div = min_t(unsigned char, clk_div, 0xFF);
+		clk_ctl |= clk_div;
+		clk_ctl |= CLK_E;
+		host->clk_div = clk_div;
+		bfin_write_SDH_CLK_CTL(clk_ctl);
+	} else
+		sdh_stop_clock(host);
+
+	/* set up sdh interrupt mask*/
+	if (ios->power_mode == MMC_POWER_ON)
+		bfin_write_SDH_MASK0(DAT_END | DAT_TIME_OUT | DAT_CRC_FAIL |
+			RX_OVERRUN | TX_UNDERRUN | CMD_SENT | CMD_RESP_END |
+			CMD_TIME_OUT | CMD_CRC_FAIL);
+	else
+		bfin_write_SDH_MASK0(0);
+	SSYNC();
+
+	spin_unlock(&host->lock);
+
+	dev_dbg(mmc_dev(host->mmc), "SDH: clk_div = 0x%x actual clock:%ld expected clock:%d\n",
+		host->clk_div,
+		host->clk_div ? get_sclk() / (2 * (host->clk_div + 1)) : 0,
+		ios->clock);
+}
+
+static const struct mmc_host_ops sdh_ops = {
+	.request	= sdh_request,
+	.set_ios	= sdh_set_ios,
+};
+
+static irqreturn_t sdh_dma_irq(int irq, void *devid)
+{
+	struct sdh_host *host = devid;
+
+	dev_dbg(mmc_dev(host->mmc), "%s enter, irq_stat: 0x%04lx\n", __func__,
+		get_dma_curr_irqstat(host->dma_ch));
+	clear_dma_irqstat(host->dma_ch);
+	SSYNC();
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t sdh_stat_irq(int irq, void *devid)
+{
+	struct sdh_host *host = devid;
+	unsigned int status;
+	int handled = 0;
+
+	dev_dbg(mmc_dev(host->mmc), "%s enter\n", __func__);
+
+	spin_lock(&host->lock);
+
+	status = bfin_read_SDH_E_STATUS();
+	if (status & SD_CARD_DET) {
+		mmc_detect_change(host->mmc, 0);
+		bfin_write_SDH_E_STATUS(SD_CARD_DET);
+	}
+	status = bfin_read_SDH_STATUS();
+	if (status & (CMD_SENT | CMD_RESP_END | CMD_TIME_OUT | CMD_CRC_FAIL)) {
+		handled |= sdh_cmd_done(host, status);
+		bfin_write_SDH_STATUS_CLR(CMD_SENT_STAT | CMD_RESP_END_STAT | \
+				CMD_TIMEOUT_STAT | CMD_CRC_FAIL_STAT);
+		SSYNC();
+	}
+
+	status = bfin_read_SDH_STATUS();
+	if (status & (DAT_END | DAT_TIME_OUT | DAT_CRC_FAIL | RX_OVERRUN | TX_UNDERRUN))
+		handled |= sdh_data_done(host, status);
+
+	spin_unlock(&host->lock);
+
+	dev_dbg(mmc_dev(host->mmc), "%s exit\n\n", __func__);
+
+	return IRQ_RETVAL(handled);
+}
+
+static void sdh_reset(void)
+{
+#if defined(CONFIG_BF54x)
+	/* Secure Digital Host shares DMA with Nand controller */
+	bfin_write_DMAC1_PERIMUX(bfin_read_DMAC1_PERIMUX() | 0x1);
+#endif
+
+	bfin_write_SDH_CFG(bfin_read_SDH_CFG() | CLKS_EN);
+	SSYNC();
+
+	/* Disable card inserting detection pin. set MMC_CAP_NEEDS_POLL, and
+	 * mmc stack will do the detection.
+	 */
+	bfin_write_SDH_CFG((bfin_read_SDH_CFG() & 0x1F) | (PUP_SDDAT | PUP_SDDAT3));
+	SSYNC();
+}
+
+static int sdh_probe(struct platform_device *pdev)
+{
+	struct mmc_host *mmc;
+	struct sdh_host *host;
+	struct bfin_sd_host *drv_data = get_sdh_data(pdev);
+	int ret;
+
+	if (!drv_data) {
+		dev_err(&pdev->dev, "missing platform driver data\n");
+		ret = -EINVAL;
+		goto out;
+	}
+
+	mmc = mmc_alloc_host(sizeof(struct sdh_host), &pdev->dev);
+	if (!mmc) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	mmc->ops = &sdh_ops;
+#if defined(CONFIG_BF51x)
+	mmc->max_segs = 1;
+#else
+	mmc->max_segs = PAGE_SIZE / sizeof(struct dma_desc_array);
+#endif
+#ifdef RSI_BLKSZ
+	mmc->max_seg_size = -1;
+#else
+	mmc->max_seg_size = 1 << 16;
+#endif
+	mmc->max_blk_size = 1 << 11;
+	mmc->max_blk_count = 1 << 11;
+	mmc->max_req_size = PAGE_SIZE;
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+	mmc->f_max = get_sclk();
+	mmc->f_min = mmc->f_max >> 9;
+	mmc->caps = MMC_CAP_4_BIT_DATA | MMC_CAP_NEEDS_POLL;
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+	host->sclk = get_sclk();
+
+	spin_lock_init(&host->lock);
+	host->irq = drv_data->irq_int0;
+	host->dma_ch = drv_data->dma_chan;
+
+	ret = request_dma(host->dma_ch, DRIVER_NAME "DMA");
+	if (ret) {
+		dev_err(&pdev->dev, "unable to request DMA channel\n");
+		goto out1;
+	}
+
+	ret = set_dma_callback(host->dma_ch, sdh_dma_irq, host);
+	if (ret) {
+		dev_err(&pdev->dev, "unable to request DMA irq\n");
+		goto out2;
+	}
+
+	host->sg_cpu = dma_alloc_coherent(&pdev->dev, PAGE_SIZE, &host->sg_dma, GFP_KERNEL);
+	if (host->sg_cpu == NULL) {
+		ret = -ENOMEM;
+		goto out2;
+	}
+
+	platform_set_drvdata(pdev, mmc);
+
+	ret = request_irq(host->irq, sdh_stat_irq, 0, "SDH Status IRQ", host);
+	if (ret) {
+		dev_err(&pdev->dev, "unable to request status irq\n");
+		goto out3;
+	}
+
+	ret = peripheral_request_list(drv_data->pin_req, DRIVER_NAME);
+	if (ret) {
+		dev_err(&pdev->dev, "unable to request peripheral pins\n");
+		goto out4;
+	}
+
+	sdh_reset();
+
+	mmc_add_host(mmc);
+	return 0;
+
+out4:
+	free_irq(host->irq, host);
+out3:
+	mmc_remove_host(mmc);
+	dma_free_coherent(&pdev->dev, PAGE_SIZE, host->sg_cpu, host->sg_dma);
+out2:
+	free_dma(host->dma_ch);
+out1:
+	mmc_free_host(mmc);
+ out:
+	return ret;
+}
+
+static int sdh_remove(struct platform_device *pdev)
+{
+	struct mmc_host *mmc = platform_get_drvdata(pdev);
+
+	if (mmc) {
+		struct sdh_host *host = mmc_priv(mmc);
+
+		mmc_remove_host(mmc);
+
+		sdh_stop_clock(host);
+		free_irq(host->irq, host);
+		free_dma(host->dma_ch);
+		dma_free_coherent(&pdev->dev, PAGE_SIZE, host->sg_cpu, host->sg_dma);
+
+		mmc_free_host(mmc);
+	}
+
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static int sdh_suspend(struct platform_device *dev, pm_message_t state)
+{
+	struct bfin_sd_host *drv_data = get_sdh_data(dev);
+
+	peripheral_free_list(drv_data->pin_req);
+
+	return 0;
+}
+
+static int sdh_resume(struct platform_device *dev)
+{
+	struct bfin_sd_host *drv_data = get_sdh_data(dev);
+	int ret = 0;
+
+	ret = peripheral_request_list(drv_data->pin_req, DRIVER_NAME);
+	if (ret) {
+		dev_err(&dev->dev, "unable to request peripheral pins\n");
+		return ret;
+	}
+
+	sdh_reset();
+	return ret;
+}
+#else
+# define sdh_suspend NULL
+# define sdh_resume  NULL
+#endif
+
+static struct platform_driver sdh_driver = {
+	.probe   = sdh_probe,
+	.remove  = sdh_remove,
+	.suspend = sdh_suspend,
+	.resume  = sdh_resume,
+	.driver  = {
+		.name = DRIVER_NAME,
+	},
+};
+
+module_platform_driver(sdh_driver);
+
+MODULE_DESCRIPTION("Blackfin Secure Digital Host Driver");
+MODULE_AUTHOR("Cliff Cai, Roy Huang");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mmc/host/cb710-mmc.c b/drivers/mmc/host/cb710-mmc.c
new file mode 100644
index 000000000..1087b4c79
--- /dev/null
+++ b/drivers/mmc/host/cb710-mmc.c
@@ -0,0 +1,780 @@
+/*
+ *  cb710/mmc.c
+ *
+ *  Copyright by Michał Mirosław, 2008-2009
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/delay.h>
+#include "cb710-mmc.h"
+
+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 */;
+	}
+
+	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;
+
+	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..8984ec878
--- /dev/null
+++ b/drivers/mmc/host/cb710-mmc.h
@@ -0,0 +1,104 @@
+/*
+ *  cb710/cb710-mmc.h
+ *
+ *  Copyright by Michał Mirosław, 2008-2009
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#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 = container_of(mmc_dev(mmc),
+		struct platform_device, dev);
+	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/davinci_mmc.c b/drivers/mmc/host/davinci_mmc.c
new file mode 100644
index 000000000..1625f908d
--- /dev/null
+++ b/drivers/mmc/host/davinci_mmc.c
@@ -0,0 +1,1482 @@
+/*
+ * davinci_mmc.c - TI DaVinci MMC/SD/SDIO driver
+ *
+ * Copyright (C) 2006 Texas Instruments.
+ *       Original author: Purushotam Kumar
+ * Copyright (C) 2009 David Brownell
+ *
+ * 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; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#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/edma.h>
+#include <linux/mmc/mmc.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+
+#include <linux/platform_data/edma.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 __initdata 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;
+
+	u32 rxdma, txdma;
+	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 stream or block transfer */
+	if (cmd->flags & MMC_DATA_STREAM)
+		cmd_reg |= MMCCMD_STRMTP;
+
+	/* 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,
+				((data->flags & MMC_DATA_WRITE)
+				? DMA_TO_DEVICE
+				: DMA_FROM_DEVICE));
+
+	/* 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,
+					(data->flags & MMC_DATA_WRITE)
+					? DMA_TO_DEVICE
+					: DMA_FROM_DEVICE);
+			return -1;
+		}
+	}
+
+	host->do_dma = 1;
+	ret = mmc_davinci_send_dma_request(host, data);
+
+	return ret;
+}
+
+static void __init_or_module
+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 __init davinci_acquire_dma_channels(struct mmc_davinci_host *host)
+{
+	int r;
+	dma_cap_mask_t mask;
+
+	dma_cap_zero(mask);
+	dma_cap_set(DMA_SLAVE, mask);
+
+	host->dma_tx =
+		dma_request_slave_channel_compat(mask, edma_filter_fn,
+				&host->txdma, mmc_dev(host->mmc), "tx");
+	if (!host->dma_tx) {
+		dev_err(mmc_dev(host->mmc), "Can't get dma_tx channel\n");
+		return -ENODEV;
+	}
+
+	host->dma_rx =
+		dma_request_slave_channel_compat(mask, edma_filter_fn,
+				&host->rxdma, mmc_dev(host->mmc), "rx");
+	if (!host->dma_rx) {
+		dev_err(mmc_dev(host->mmc), "Can't get dma_rx channel\n");
+		r = -ENODEV;
+		goto free_master_write;
+	}
+
+	return 0;
+
+free_master_write:
+	dma_release_channel(host->dma_tx);
+
+	return r;
+}
+
+/*----------------------------------------------------------------------*/
+
+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 %s, %d blocks of %d bytes\n",
+		(data->flags & MMC_DATA_STREAM) ? "stream" : "block",
+		(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 */
+	switch (data->flags & MMC_DATA_WRITE) {
+	case 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);
+		break;
+
+	default:
+		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);
+		break;
+	}
+
+	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,
+			     (data->flags & MMC_DATA_WRITE)
+			     ? DMA_TO_DEVICE
+			     : DMA_FROM_DEVICE);
+		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 = (int) host->cmd;
+	}
+
+	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 -ENOSYS;
+	return config->get_cd(pdev->id);
+}
+
+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 -ENOSYS;
+	return config->get_ro(pdev->id);
+}
+
+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 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 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 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 struct davinci_mmc_config
+	*mmc_parse_pdata(struct platform_device *pdev)
+{
+	struct device_node *np;
+	struct davinci_mmc_config *pdata = pdev->dev.platform_data;
+	const struct of_device_id *match =
+		of_match_device(davinci_mmc_dt_ids, &pdev->dev);
+	u32 data;
+
+	np = pdev->dev.of_node;
+	if (!np)
+		return pdata;
+
+	pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata) {
+		dev_err(&pdev->dev, "Failed to allocate memory for struct davinci_mmc_config\n");
+		goto nodata;
+	}
+
+	if (match)
+		pdev->id_entry = match->data;
+
+	if (of_property_read_u32(np, "max-frequency", &pdata->max_freq))
+		dev_info(&pdev->dev, "'max-frequency' property not specified, defaulting to 25MHz\n");
+
+	of_property_read_u32(np, "bus-width", &data);
+	switch (data) {
+	case 1:
+	case 4:
+	case 8:
+		pdata->wires = data;
+		break;
+	default:
+		pdata->wires = 1;
+		dev_info(&pdev->dev, "Unsupported buswidth, defaulting to 1 bit\n");
+	}
+nodata:
+	return pdata;
+}
+
+static int __init davinci_mmcsd_probe(struct platform_device *pdev)
+{
+	struct davinci_mmc_config *pdata = NULL;
+	struct mmc_davinci_host *host = NULL;
+	struct mmc_host *mmc = NULL;
+	struct resource *r, *mem = NULL;
+	int ret = 0, irq = 0;
+	size_t mem_size;
+	const struct platform_device_id *id_entry;
+
+	pdata = mmc_parse_pdata(pdev);
+	if (pdata == NULL) {
+		dev_err(&pdev->dev, "Couldn't get platform data\n");
+		return -ENOENT;
+	}
+
+	ret = -ENODEV;
+	r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	irq = platform_get_irq(pdev, 0);
+	if (!r || irq == NO_IRQ)
+		goto out;
+
+	ret = -EBUSY;
+	mem_size = resource_size(r);
+	mem = request_mem_region(r->start, mem_size, pdev->name);
+	if (!mem)
+		goto out;
+
+	ret = -ENOMEM;
+	mmc = mmc_alloc_host(sizeof(struct mmc_davinci_host), &pdev->dev);
+	if (!mmc)
+		goto out;
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;	/* Important */
+
+	r = platform_get_resource(pdev, IORESOURCE_DMA, 0);
+	if (!r)
+		dev_warn(&pdev->dev, "RX DMA resource not specified\n");
+	else
+		host->rxdma = r->start;
+
+	r = platform_get_resource(pdev, IORESOURCE_DMA, 1);
+	if (!r)
+		dev_warn(&pdev->dev, "TX DMA resource not specified\n");
+	else
+		host->txdma = r->start;
+
+	host->mem_res = mem;
+	host->base = ioremap(mem->start, mem_size);
+	if (!host->base)
+		goto out;
+
+	ret = -ENXIO;
+	host->clk = clk_get(&pdev->dev, "MMCSDCLK");
+	if (IS_ERR(host->clk)) {
+		ret = PTR_ERR(host->clk);
+		goto out;
+	}
+	clk_enable(host->clk);
+	host->mmc_input_clk = clk_get_rate(host->clk);
+
+	init_mmcsd_host(host);
+
+	if (pdata->nr_sg)
+		host->nr_sg = pdata->nr_sg - 1;
+
+	if (host->nr_sg > MAX_NR_SG || !host->nr_sg)
+		host->nr_sg = MAX_NR_SG;
+
+	host->use_dma = use_dma;
+	host->mmc_irq = irq;
+	host->sdio_irq = platform_get_irq(pdev, 1);
+
+	if (host->use_dma && davinci_acquire_dma_channels(host) != 0)
+		host->use_dma = 0;
+
+	/* REVISIT:  someday, support IRQ-driven card detection.  */
+	mmc->caps |= MMC_CAP_NEEDS_POLL;
+	mmc->caps |= MMC_CAP_WAIT_WHILE_BUSY;
+
+	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);
+
+	id_entry = platform_get_device_id(pdev);
+	if (id_entry)
+		host->version = id_entry->driver_data;
+
+	mmc->ops = &mmc_davinci_ops;
+	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;
+	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 out;
+
+	ret = request_irq(irq, mmc_davinci_irq, 0, mmc_hostname(mmc), host);
+	if (ret)
+		goto out;
+
+	if (host->sdio_irq >= 0) {
+		ret = request_irq(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;
+
+out:
+	mmc_davinci_cpufreq_deregister(host);
+cpu_freq_fail:
+	if (host) {
+		davinci_release_dma_channels(host);
+
+		if (host->clk) {
+			clk_disable(host->clk);
+			clk_put(host->clk);
+		}
+
+		if (host->base)
+			iounmap(host->base);
+	}
+
+	if (mmc)
+		mmc_free_host(mmc);
+
+	if (mem)
+		release_resource(mem);
+
+	dev_dbg(&pdev->dev, "probe err %d\n", ret);
+
+	return ret;
+}
+
+static int __exit davinci_mmcsd_remove(struct platform_device *pdev)
+{
+	struct mmc_davinci_host *host = platform_get_drvdata(pdev);
+
+	if (host) {
+		mmc_davinci_cpufreq_deregister(host);
+
+		mmc_remove_host(host->mmc);
+		free_irq(host->mmc_irq, host);
+		if (host->mmc->caps & MMC_CAP_SDIO_IRQ)
+			free_irq(host->sdio_irq, host);
+
+		davinci_release_dma_channels(host);
+
+		clk_disable(host->clk);
+		clk_put(host->clk);
+
+		iounmap(host->base);
+
+		release_resource(host->mem_res);
+
+		mmc_free_host(host->mmc);
+	}
+
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static int davinci_mmcsd_suspend(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct mmc_davinci_host *host = platform_get_drvdata(pdev);
+
+	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 platform_device *pdev = to_platform_device(dev);
+	struct mmc_davinci_host *host = platform_get_drvdata(pdev);
+
+	clk_enable(host->clk);
+	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",
+		.pm	= davinci_mmcsd_pm_ops,
+		.of_match_table = davinci_mmc_dt_ids,
+	},
+	.remove		= __exit_p(davinci_mmcsd_remove),
+	.id_table	= davinci_mmc_devtype,
+};
+
+module_platform_driver_probe(davinci_mmcsd_driver, davinci_mmcsd_probe);
+
+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-exynos.c b/drivers/mmc/host/dw_mmc-exynos.c
new file mode 100644
index 000000000..e761eb1b1
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-exynos.c
@@ -0,0 +1,559 @@
+/*
+ * Exynos Specific Extensions for Synopsys DW Multimedia Card Interface driver
+ *
+ * Copyright (C) 2012, Samsung Electronics Co., Ltd.
+ *
+ * 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; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/clk.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/dw_mmc.h>
+#include <linux/mmc/mmc.h>
+#include <linux/of.h>
+#include <linux/of_gpio.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 int dw_mci_exynos_priv_init(struct dw_mci *host)
+{
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+
+	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);
+	}
+
+	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);
+	}
+
+	return 0;
+}
+
+static int dw_mci_exynos_setup_clock(struct dw_mci *host)
+{
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+
+	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);
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int dw_mci_exynos_suspend(struct device *dev)
+{
+	struct dw_mci *host = dev_get_drvdata(dev);
+
+	return dw_mci_suspend(host);
+}
+
+static int dw_mci_exynos_resume(struct device *dev)
+{
+	struct dw_mci *host = dev_get_drvdata(dev);
+
+	dw_mci_exynos_priv_init(host);
+	return dw_mci_resume(host);
+}
+
+/**
+ * dw_mci_exynos_resume_noirq - Exynos-specific resume code
+ *
+ * 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;
+
+	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);
+	}
+
+	return 0;
+}
+#else
+#define dw_mci_exynos_suspend		NULL
+#define dw_mci_exynos_resume		NULL
+#define dw_mci_exynos_resume_noirq	NULL
+#endif /* CONFIG_PM_SLEEP */
+
+static void dw_mci_exynos_prepare_command(struct dw_mci *host, u32 *cmdr)
+{
+	struct dw_mci_exynos_priv_data *priv = host->priv;
+	/*
+	 * 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 (priv->ctrl_type == DW_MCI_TYPE_EXYNOS7 ||
+		priv->ctrl_type == DW_MCI_TYPE_EXYNOS7_SMU) {
+		if (SDMMC_CLKSEL_GET_DRV_WD3(mci_readl(host, CLKSEL64)))
+			*cmdr |= SDMMC_CMD_USE_HOLD_REG;
+	 } else {
+		if (SDMMC_CLKSEL_GET_DRV_WD3(mci_readl(host, CLKSEL)))
+			*cmdr |= SDMMC_CMD_USE_HOLD_REG;
+	}
+}
+
+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)
+		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 {
+		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;
+	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;
+		}
+	}
+
+out:
+	return loc;
+}
+
+static int dw_mci_exynos_execute_tuning(struct dw_mci_slot *slot)
+{
+	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 = -1;
+	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))
+			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;
+	}
+
+	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,
+	.init			= dw_mci_exynos_priv_init,
+	.setup_clock		= dw_mci_exynos_setup_clock,
+	.prepare_command	= dw_mci_exynos_prepare_command,
+	.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;
+
+	match = of_match_node(dw_mci_exynos_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_exynos_pmops = {
+	SET_SYSTEM_SLEEP_PM_OPS(dw_mci_exynos_suspend, dw_mci_exynos_resume)
+	.resume_noirq = dw_mci_exynos_resume_noirq,
+	.thaw_noirq = dw_mci_exynos_resume_noirq,
+	.restore_noirq = dw_mci_exynos_resume_noirq,
+};
+
+static struct platform_driver dw_mci_exynos_pltfm_driver = {
+	.probe		= dw_mci_exynos_probe,
+	.remove		= dw_mci_pltfm_remove,
+	.driver		= {
+		.name		= "dwmmc_exynos",
+		.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..595c934e6
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-exynos.h
@@ -0,0 +1,73 @@
+/*
+ * Exynos Specific Extensions for Synopsys DW Multimedia Card Interface driver
+ *
+ * Copyright (C) 2012-2014 Samsung Electronics Co., Ltd.
+ *
+ * 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; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#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-k3.c b/drivers/mmc/host/dw_mmc-k3.c
new file mode 100644
index 000000000..650f9cc3f
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-k3.c
@@ -0,0 +1,97 @@
+/*
+ * Copyright (c) 2013 Linaro Ltd.
+ * Copyright (c) 2013 Hisilicon Limited.
+ *
+ * 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; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/clk.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/dw_mmc.h>
+#include <linux/of_address.h>
+
+#include "dw_mmc.h"
+#include "dw_mmc-pltfm.h"
+
+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 const struct of_device_id dw_mci_k3_match[] = {
+	{ .compatible = "hisilicon,hi4511-dw-mshc", .data = &k3_drv_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);
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int dw_mci_k3_suspend(struct device *dev)
+{
+	struct dw_mci *host = dev_get_drvdata(dev);
+	int ret;
+
+	ret = dw_mci_suspend(host);
+	if (!ret)
+		clk_disable_unprepare(host->ciu_clk);
+
+	return ret;
+}
+
+static int dw_mci_k3_resume(struct device *dev)
+{
+	struct dw_mci *host = dev_get_drvdata(dev);
+	int ret;
+
+	ret = clk_prepare_enable(host->ciu_clk);
+	if (ret) {
+		dev_err(host->dev, "failed to enable ciu clock\n");
+		return ret;
+	}
+
+	return dw_mci_resume(host);
+}
+#endif /* CONFIG_PM_SLEEP */
+
+static SIMPLE_DEV_PM_OPS(dw_mci_k3_pmops, dw_mci_k3_suspend, dw_mci_k3_resume);
+
+static struct platform_driver dw_mci_k3_pltfm_driver = {
+	.probe		= dw_mci_k3_probe,
+	.remove		= dw_mci_pltfm_remove,
+	.driver		= {
+		.name		= "dwmmc_k3",
+		.of_match_table	= dw_mci_k3_match,
+		.pm		= &dw_mci_k3_pmops,
+	},
+};
+
+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..4c69fbd29
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-pci.c
@@ -0,0 +1,122 @@
+/*
+ * Synopsys DesignWare Multimedia Card PCI Interface driver
+ *
+ * Copyright (C) 2012 Vayavya Labs Pvt. Ltd.
+ *
+ * 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; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/io.h>
+#include <linux/irq.h>
+#include <linux/pci.h>
+#include <linux/slab.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/dw_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 = {
+	.num_slots			= 1,
+	.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);
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int dw_mci_pci_suspend(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct dw_mci *host = pci_get_drvdata(pdev);
+
+	return dw_mci_suspend(host);
+}
+
+static int dw_mci_pci_resume(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct dw_mci *host = pci_get_drvdata(pdev);
+
+	return dw_mci_resume(host);
+}
+#endif /* CONFIG_PM_SLEEP */
+
+static SIMPLE_DEV_PM_OPS(dw_mci_pci_pmops, dw_mci_pci_suspend, dw_mci_pci_resume);
+
+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_pmops
+	},
+};
+
+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..ec6dbcdec
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-pltfm.c
@@ -0,0 +1,140 @@
+/*
+ * Synopsys DesignWare Multimedia Card Interface driver
+ *
+ * Copyright (C) 2009 NXP Semiconductors
+ * Copyright (C) 2009, 2010 Imagination Technologies Ltd.
+ *
+ * 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; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#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/slab.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/dw_mmc.h>
+#include <linux/of.h>
+#include <linux/clk.h>
+
+#include "dw_mmc.h"
+#include "dw_mmc-pltfm.h"
+
+static void dw_mci_pltfm_prepare_command(struct dw_mci *host, u32 *cmdr)
+{
+	*cmdr |= SDMMC_CMD_USE_HOLD_REG;
+}
+
+static const struct dw_mci_drv_data socfpga_drv_data = {
+	.prepare_command	= dw_mci_pltfm_prepare_command,
+};
+
+static const struct dw_mci_drv_data pistachio_drv_data = {
+	.prepare_command	= dw_mci_pltfm_prepare_command,
+};
+
+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);
+
+	platform_set_drvdata(pdev, host);
+	return dw_mci_probe(host);
+}
+EXPORT_SYMBOL_GPL(dw_mci_pltfm_register);
+
+#ifdef CONFIG_PM_SLEEP
+/*
+ * TODO: we should probably disable the clock to the card in the suspend path.
+ */
+static int dw_mci_pltfm_suspend(struct device *dev)
+{
+	struct dw_mci *host = dev_get_drvdata(dev);
+
+	return dw_mci_suspend(host);
+}
+
+static int dw_mci_pltfm_resume(struct device *dev)
+{
+	struct dw_mci *host = dev_get_drvdata(dev);
+
+	return dw_mci_resume(host);
+}
+#endif /* CONFIG_PM_SLEEP */
+
+SIMPLE_DEV_PM_OPS(dw_mci_pltfm_pmops, dw_mci_pltfm_suspend, dw_mci_pltfm_resume);
+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",
+		.data = &socfpga_drv_data },
+	{ .compatible = "img,pistachio-dw-mshc",
+		.data = &pistachio_drv_data },
+	{},
+};
+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",
+		.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..68e7fd2f6
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-pltfm.h
@@ -0,0 +1,20 @@
+/*
+ * Synopsys DesignWare Multimedia Card Interface Platform driver
+ *
+ * Copyright (C) 2012, Samsung Electronics Co., Ltd.
+ *
+ * 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; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#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..dbf166f94
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc-rockchip.c
@@ -0,0 +1,157 @@
+/*
+ * Copyright (c) 2014, Fuzhou Rockchip Electronics Co., Ltd
+ *
+ * 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; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/clk.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/dw_mmc.h>
+#include <linux/of_address.h>
+
+#include "dw_mmc.h"
+#include "dw_mmc-pltfm.h"
+
+#define RK3288_CLKGEN_DIV       2
+
+static void dw_mci_rockchip_prepare_command(struct dw_mci *host, u32 *cmdr)
+{
+	*cmdr |= SDMMC_CMD_USE_HOLD_REG;
+}
+
+static int dw_mci_rk3288_setup_clock(struct dw_mci *host)
+{
+	host->bus_hz /= RK3288_CLKGEN_DIV;
+
+	return 0;
+}
+
+static void dw_mci_rk3288_set_ios(struct dw_mci *host, struct mmc_ios *ios)
+{
+	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
+	 *       if DDR50 8bit mode(only emmc work in 8bit mode),
+	 *       div must be set 1
+	 */
+	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;
+	}
+}
+
+static int dw_mci_rockchip_init(struct dw_mci *host)
+{
+	/* It is slot 8 on Rockchip SoCs */
+	host->sdio_id0 = 8;
+
+	return 0;
+}
+
+/* Common capabilities of RK3288 SoC */
+static unsigned long dw_mci_rk3288_dwmmc_caps[4] = {
+	MMC_CAP_RUNTIME_RESUME, /* emmc */
+	MMC_CAP_RUNTIME_RESUME, /* sdmmc */
+	MMC_CAP_RUNTIME_RESUME, /* sdio0 */
+	MMC_CAP_RUNTIME_RESUME, /* sdio1 */
+};
+static const struct dw_mci_drv_data rk2928_drv_data = {
+	.prepare_command        = dw_mci_rockchip_prepare_command,
+	.init			= dw_mci_rockchip_init,
+};
+
+static const struct dw_mci_drv_data rk3288_drv_data = {
+	.caps			= dw_mci_rk3288_dwmmc_caps,
+	.prepare_command        = dw_mci_rockchip_prepare_command,
+	.set_ios		= dw_mci_rk3288_set_ios,
+	.setup_clock    = dw_mci_rk3288_setup_clock,
+	.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;
+
+	if (!pdev->dev.of_node)
+		return -ENODEV;
+
+	match = of_match_node(dw_mci_rockchip_match, pdev->dev.of_node);
+	drv_data = match->data;
+
+	return dw_mci_pltfm_register(pdev, drv_data);
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int dw_mci_rockchip_suspend(struct device *dev)
+{
+	struct dw_mci *host = dev_get_drvdata(dev);
+
+	return dw_mci_suspend(host);
+}
+
+static int dw_mci_rockchip_resume(struct device *dev)
+{
+	struct dw_mci *host = dev_get_drvdata(dev);
+
+	return dw_mci_resume(host);
+}
+#endif /* CONFIG_PM_SLEEP */
+
+static SIMPLE_DEV_PM_OPS(dw_mci_rockchip_pmops,
+			 dw_mci_rockchip_suspend,
+			 dw_mci_rockchip_resume);
+
+static struct platform_driver dw_mci_rockchip_pltfm_driver = {
+	.probe		= dw_mci_rockchip_probe,
+	.remove		= dw_mci_pltfm_remove,
+	.driver		= {
+		.name		= "dwmmc_rockchip",
+		.of_match_table	= dw_mci_rockchip_match,
+		.pm		= &dw_mci_rockchip_pmops,
+	},
+};
+
+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.c b/drivers/mmc/host/dw_mmc.c
new file mode 100644
index 000000000..5f5adafb2
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc.c
@@ -0,0 +1,3042 @@
+/*
+ * 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.
+ *
+ * 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; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#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/ioport.h>
+#include <linux/module.h>
+#include <linux/platform_device.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/mmc/dw_mmc.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)
+#define DW_MCI_CMD_ERROR_FLAGS	(SDMMC_INT_RTO | SDMMC_INT_RCRC | \
+				 SDMMC_INT_RESP_ERR)
+#define DW_MCI_ERROR_FLAGS	(DW_MCI_DATA_ERROR_FLAGS | \
+				 DW_MCI_CMD_ERROR_FLAGS  | SDMMC_INT_HLE)
+#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	400000		/* unit: HZ */
+
+#ifdef CONFIG_MMC_DW_IDMAC
+#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)
+
+struct idmac_desc_64addr {
+	u32		des0;	/* Control Descriptor */
+
+	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 & 0x03ffe000) | ((s) & 0x1fff))
+
+	__le32		des2;	/* buffer 1 physical address */
+
+	__le32		des3;	/* buffer 2 physical address */
+};
+#endif /* CONFIG_MMC_DW_IDMAC */
+
+static bool dw_mci_reset(struct dw_mci *host);
+static bool dw_mci_ctrl_reset(struct dw_mci *host, u32 reset);
+static int dw_mci_card_busy(struct mmc_host *mmc);
+
+#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;
+}
+
+static int dw_mci_req_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, dw_mci_req_show, inode->i_private);
+}
+
+static const struct file_operations dw_mci_req_fops = {
+	.owner		= THIS_MODULE,
+	.open		= dw_mci_req_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+static int dw_mci_regs_show(struct seq_file *s, void *v)
+{
+	seq_printf(s, "STATUS:\t0x%08x\n", SDMMC_STATUS);
+	seq_printf(s, "RINTSTS:\t0x%08x\n", SDMMC_RINTSTS);
+	seq_printf(s, "CMD:\t0x%08x\n", SDMMC_CMD);
+	seq_printf(s, "CTRL:\t0x%08x\n", SDMMC_CTRL);
+	seq_printf(s, "INTMASK:\t0x%08x\n", SDMMC_INTMASK);
+	seq_printf(s, "CLKENA:\t0x%08x\n", SDMMC_CLKENA);
+
+	return 0;
+}
+
+static int dw_mci_regs_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, dw_mci_regs_show, inode->i_private);
+}
+
+static const struct file_operations dw_mci_regs_fops = {
+	.owner		= THIS_MODULE,
+	.open		= dw_mci_regs_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+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;
+	struct dentry *node;
+
+	root = mmc->debugfs_root;
+	if (!root)
+		return;
+
+	node = debugfs_create_file("regs", S_IRUSR, root, host,
+				   &dw_mci_regs_fops);
+	if (!node)
+		goto err;
+
+	node = debugfs_create_file("req", S_IRUSR, root, slot,
+				   &dw_mci_req_fops);
+	if (!node)
+		goto err;
+
+	node = debugfs_create_u32("state", S_IRUSR, root, (u32 *)&host->state);
+	if (!node)
+		goto err;
+
+	node = debugfs_create_x32("pending_events", S_IRUSR, root,
+				  (u32 *)&host->pending_events);
+	if (!node)
+		goto err;
+
+	node = debugfs_create_x32("completed_events", S_IRUSR, root,
+				  (u32 *)&host->completed_events);
+	if (!node)
+		goto err;
+
+	return;
+
+err:
+	dev_err(&mmc->class_dev, "failed to initialize debugfs for slot\n");
+}
+#endif /* defined(CONFIG_DEBUG_FS) */
+
+static void mci_send_cmd(struct dw_mci_slot *slot, u32 cmd, u32 arg);
+
+static u32 dw_mci_prepare_command(struct mmc_host *mmc, struct mmc_command *cmd)
+{
+	struct mmc_data	*data;
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct dw_mci *host = slot->host;
+	const struct dw_mci_drv_data *drv_data = slot->host->drv_data;
+	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;
+
+	data = cmd->data;
+	if (data) {
+		cmdr |= SDMMC_CMD_DAT_EXP;
+		if (data->flags & MMC_DATA_STREAM)
+			cmdr |= SDMMC_CMD_STRM_MODE;
+		if (data->flags & MMC_DATA_WRITE)
+			cmdr |= SDMMC_CMD_DAT_WR;
+	}
+
+	if (drv_data && drv_data->prepare_command)
+		drv_data->prepare_command(slot->host, &cmdr);
+
+	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;
+
+	return cmdr;
+}
+
+static void dw_mci_wait_while_busy(struct dw_mci *host, u32 cmd_flags)
+{
+	unsigned long timeout = jiffies + msecs_to_jiffies(500);
+
+	/*
+	 * 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)) {
+		while (mci_readl(host, STATUS) & SDMMC_STATUS_BUSY) {
+			if (time_after(jiffies, timeout)) {
+				/* Command will fail; we'll pass error then */
+				dev_err(host->dev, "Busy; trying anyway\n");
+				break;
+			}
+			udelay(10);
+		}
+	}
+}
+
+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();
+	dw_mci_wait_while_busy(host, cmd_flags);
+
+	mci_writel(host, CMD, cmd_flags | SDMMC_CMD_START);
+}
+
+static inline void send_stop_abort(struct dw_mci *host, struct mmc_data *data)
+{
+	struct mmc_command *stop = data->stop ? data->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 int dw_mci_get_dma_dir(struct mmc_data *data)
+{
+	if (data->flags & MMC_DATA_WRITE)
+		return DMA_TO_DEVICE;
+	else
+		return DMA_FROM_DEVICE;
+}
+
+#ifdef CONFIG_MMC_DW_IDMAC
+static void dw_mci_dma_cleanup(struct dw_mci *host)
+{
+	struct mmc_data *data = host->data;
+
+	if (data)
+		if (!data->host_cookie)
+			dma_unmap_sg(host->dev,
+				     data->sg,
+				     data->sg_len,
+				     dw_mci_get_dma_dir(data));
+}
+
+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_idmac_complete_dma(struct dw_mci *host)
+{
+	struct mmc_data *data = host->data;
+
+	dev_vdbg(host->dev, "DMA complete\n");
+
+	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 void dw_mci_translate_sglist(struct dw_mci *host, struct mmc_data *data,
+				    unsigned int sg_len)
+{
+	int i;
+	if (host->dma_64bit_address == 1) {
+		struct idmac_desc_64addr *desc = host->sg_cpu;
+
+		for (i = 0; i < sg_len; i++, desc++) {
+			unsigned int length = sg_dma_len(&data->sg[i]);
+			u64 mem_addr = sg_dma_address(&data->sg[i]);
+
+			/*
+			 * 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, length);
+
+			/* Physical address to DMA to/from */
+			desc->des4 = mem_addr & 0xffffffff;
+			desc->des5 = mem_addr >> 32;
+		}
+
+		/* Set first descriptor */
+		desc = host->sg_cpu;
+		desc->des0 |= IDMAC_DES0_FD;
+
+		/* Set last descriptor */
+		desc = host->sg_cpu + (i - 1) *
+				sizeof(struct idmac_desc_64addr);
+		desc->des0 &= ~(IDMAC_DES0_CH | IDMAC_DES0_DIC);
+		desc->des0 |= IDMAC_DES0_LD;
+
+	} else {
+		struct idmac_desc *desc = host->sg_cpu;
+
+		for (i = 0; i < sg_len; i++, desc++) {
+			unsigned int length = sg_dma_len(&data->sg[i]);
+			u32 mem_addr = sg_dma_address(&data->sg[i]);
+
+			/*
+			 * 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, length);
+
+			/* Physical address to DMA to/from */
+			desc->des2 = cpu_to_le32(mem_addr);
+		}
+
+		/* Set first descriptor */
+		desc = host->sg_cpu;
+		desc->des0 |= cpu_to_le32(IDMAC_DES0_FD);
+
+		/* Set last descriptor */
+		desc = host->sg_cpu + (i - 1) * sizeof(struct idmac_desc);
+		desc->des0 &= cpu_to_le32(~(IDMAC_DES0_CH | IDMAC_DES0_DIC));
+		desc->des0 |= cpu_to_le32(IDMAC_DES0_LD);
+	}
+
+	wmb();
+}
+
+static void dw_mci_idmac_start_dma(struct dw_mci *host, unsigned int sg_len)
+{
+	u32 temp;
+
+	dw_mci_translate_sglist(host, host->data, sg_len);
+
+	/* 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);
+
+	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);
+}
+
+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 = PAGE_SIZE / 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->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 = PAGE_SIZE / 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->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 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_idmac_complete_dma,
+	.cleanup = dw_mci_dma_cleanup,
+};
+#endif /* CONFIG_MMC_DW_IDMAC */
+
+static int dw_mci_pre_dma_transfer(struct dw_mci *host,
+				   struct mmc_data *data,
+				   bool next)
+{
+	struct scatterlist *sg;
+	unsigned int i, sg_len;
+
+	if (!next && data->host_cookie)
+		return data->host_cookie;
+
+	/*
+	 * 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,
+			    dw_mci_get_dma_dir(data));
+	if (sg_len == 0)
+		return -EINVAL;
+
+	if (next)
+		data->host_cookie = sg_len;
+
+	return sg_len;
+}
+
+static void dw_mci_pre_req(struct mmc_host *mmc,
+			   struct mmc_request *mrq,
+			   bool is_first_req)
+{
+	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) {
+		data->host_cookie = 0;
+		return;
+	}
+
+	if (dw_mci_pre_dma_transfer(slot->host, mrq->data, 1) < 0)
+		data->host_cookie = 0;
+}
+
+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)
+		dma_unmap_sg(slot->host->dev,
+			     data->sg,
+			     data->sg_len,
+			     dw_mci_get_dma_dir(data));
+	data->host_cookie = 0;
+}
+
+static void dw_mci_adjust_fifoth(struct dw_mci *host, struct mmc_data *data)
+{
+#ifdef CONFIG_MMC_DW_IDMAC
+	unsigned int blksz = data->blksz;
+	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 = (sizeof(mszs) / sizeof(mszs[0])) - 1;
+
+	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) {
+		msize = 0;
+		rx_wmark = 1;
+		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);
+#endif
+}
+
+static void dw_mci_ctrl_rd_thld(struct dw_mci *host, struct mmc_data *data)
+{
+	unsigned int blksz = data->blksz;
+	u32 blksz_depth, fifo_depth;
+	u16 thld_size;
+
+	WARN_ON(!(data->flags & MMC_DATA_READ));
+
+	/*
+	 * 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)
+		return;
+
+	if (host->timing != MMC_TIMING_MMC_HS200 &&
+	    host->timing != MMC_TIMING_MMC_HS400 &&
+	    host->timing != MMC_TIMING_UHS_SDR104)
+		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_RD_THLD(thld_size, 1));
+	return;
+
+disable:
+	mci_writel(host, CDTHRCTL, SDMMC_SET_RD_THLD(0, 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, 0);
+	if (sg_len < 0) {
+		host->dma_ops->stop(host);
+		return sg_len;
+	}
+
+	host->using_dma = 1;
+
+	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);
+
+	host->dma_ops->start(host, sg_len);
+
+	return 0;
+}
+
+static void dw_mci_submit_data(struct dw_mci *host, struct mmc_data *data)
+{
+	unsigned long irqflags;
+	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;
+		dw_mci_ctrl_rd_thld(host, data);
+	} else {
+		host->dir_status = DW_MCI_SEND_STATUS;
+	}
+
+	if (dw_mci_submit_data_dma(host, data)) {
+		int 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->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 next issued data may be transfered by DMA mode,
+		 * prev_blksz should be invalidated.
+		 */
+		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 mci_send_cmd(struct dw_mci_slot *slot, u32 cmd, u32 arg)
+{
+	struct dw_mci *host = slot->host;
+	unsigned long timeout = jiffies + msecs_to_jiffies(500);
+	unsigned int cmd_status = 0;
+
+	mci_writel(host, CMDARG, arg);
+	wmb();
+	dw_mci_wait_while_busy(host, cmd);
+	mci_writel(host, CMD, SDMMC_CMD_START | cmd);
+
+	while (time_before(jiffies, timeout)) {
+		cmd_status = mci_readl(host, CMD);
+		if (!(cmd_status & SDMMC_CMD_START))
+			return;
+	}
+	dev_err(&slot->mmc->class_dev,
+		"Timeout sending command (cmd %#x arg %#x status %#x)\n",
+		cmd, arg, cmd_status);
+}
+
+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;
+
+	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 << div) != slot->__clk_old || 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);
+
+		/* 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 clock with reflecting clock dividor */
+		slot->__clk_old = clock << div;
+	}
+
+	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->cur_slot = slot;
+	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();
+	}
+
+	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);
+	}
+
+	if (mrq->stop)
+		host->stop_cmdr = dw_mci_prepare_command(slot->mmc, mrq->stop);
+	else
+		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.
+	 */
+	spin_lock_bh(&host->lock);
+
+	if (!test_bit(DW_MMC_CARD_PRESENT, &slot->flags)) {
+		spin_unlock_bh(&host->lock);
+		mrq->cmd->error = -ENOMEDIUM;
+		mmc_request_done(mmc, mrq);
+		return;
+	}
+
+	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_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;
+	u32 uhs;
+	u32 v18 = SDMMC_UHS_18V << slot->id;
+	int min_uv, max_uv;
+	int ret;
+
+	/*
+	 * 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) {
+		min_uv = 2700000;
+		max_uv = 3600000;
+		uhs &= ~v18;
+	} else {
+		min_uv = 1700000;
+		max_uv = 1950000;
+		uhs |= v18;
+	}
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+		ret = regulator_set_voltage(mmc->supply.vqmmc, min_uv, max_uv);
+
+		if (ret) {
+			dev_dbg(&mmc->class_dev,
+					 "Regulator set error %d: %d - %d\n",
+					 ret, min_uv, max_uv);
+			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 ((slot->quirks & DW_MCI_SLOT_QUIRK_NO_WRITE_PROTECT) ||
+			(slot->host->quirks & DW_MCI_QUIRK_NO_WRITE_PROTECT))
+		read_only = 0;
+	else if (!IS_ERR_VALUE(gpio_ro))
+		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 int dw_mci_get_cd(struct mmc_host *mmc)
+{
+	int present;
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	struct dw_mci_board *brd = slot->host->pdata;
+	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 ((brd->quirks & DW_MCI_QUIRK_BROKEN_CARD_DETECTION) ||
+	    (mmc->caps & MMC_CAP_NONREMOVABLE))
+		present = 1;
+	else if (!IS_ERR_VALUE(gpio_cd))
+		present = gpio_cd;
+	else
+		present = (mci_readl(slot->host, CDETECT) & (1 << slot->id))
+			== 0 ? 1 : 0;
+
+	spin_lock_bh(&host->lock);
+	if (present) {
+		set_bit(DW_MMC_CARD_PRESENT, &slot->flags);
+		dev_dbg(&mmc->class_dev, "card is present\n");
+	} else {
+		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_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 mmc_host *mmc, int enb)
+{
+	struct dw_mci_slot *slot = mmc_priv(mmc);
+	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 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 = -ENOSYS;
+
+	if (drv_data && drv_data->execute_tuning)
+		err = drv_data->execute_tuning(slot);
+	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 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,
+	.enable_sdio_irq	= dw_mci_enable_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->cur_slot->mmc;
+
+	WARN_ON(host->cmd || host->data);
+
+	host->cur_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;
+
+	if (cmd->error) {
+		/* newer ip versions need a delay between retries */
+		if (host->quirks & DW_MCI_QUIRK_RETRY_DELAY)
+			mdelay(20);
+	}
+
+	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 = -EIO;
+			}
+		} else {
+			/* SDMMC_INT_SBE is included */
+			data->error = -EIO;
+		}
+
+		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_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 (!test_and_clear_bit(EVENT_CMD_COMPLETE,
+						&host->pending_events))
+				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) {
+				prev_state = state = STATE_SENDING_CMD;
+				__dw_mci_start_request(host, host->cur_slot,
+						       mrq->cmd);
+				goto unlock;
+			}
+
+			if (cmd->data && err) {
+				dw_mci_stop_dma(host);
+				send_stop_abort(host, data);
+				state = STATE_SENDING_STOP;
+				break;
+			}
+
+			if (!cmd->data || err) {
+				dw_mci_request_end(host, mrq);
+				goto unlock;
+			}
+
+			prev_state = state = STATE_SENDING_DATA;
+			/* fall through */
+
+		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)) {
+				dw_mci_stop_dma(host);
+				if (data->stop ||
+				    !(host->data_status & (SDMMC_INT_DRTO |
+							   SDMMC_INT_EBE)))
+					send_stop_abort(host, data);
+				state = STATE_DATA_ERROR;
+				break;
+			}
+
+			if (!test_and_clear_bit(EVENT_XFER_COMPLETE,
+						&host->pending_events))
+				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)) {
+				dw_mci_stop_dma(host);
+				if (data->stop ||
+				    !(host->data_status & (SDMMC_INT_DRTO |
+							   SDMMC_INT_EBE)))
+					send_stop_abort(host, data);
+				state = STATE_DATA_ERROR;
+				break;
+			}
+			prev_state = state = STATE_DATA_BUSY;
+
+			/* fall through */
+
+		case STATE_DATA_BUSY:
+			if (!test_and_clear_bit(EVENT_DATA_COMPLETE,
+						&host->pending_events))
+				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;
+
+			/* fall through */
+
+		case STATE_SENDING_STOP:
+			if (!test_and_clear_bit(EVENT_CMD_COMPLETE,
+						&host->pending_events))
+				break;
+
+			/* CMD error in data command */
+			if (mrq->cmd->error && mrq->data)
+				dw_mci_reset(host);
+
+			host->cmd = NULL;
+			host->data = NULL;
+
+			if (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(cnt, (int)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();
+	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();
+	set_bit(EVENT_XFER_COMPLETE, &host->pending_events);
+}
+
+static void dw_mci_cmd_interrupt(struct dw_mci *host, u32 status)
+{
+	if (!host->cmd_status)
+		host->cmd_status = status;
+
+	smp_wmb();
+
+	set_bit(EVENT_CMD_COMPLETE, &host->pending_events);
+	tasklet_schedule(&host->tasklet);
+}
+
+static void dw_mci_handle_cd(struct dw_mci *host)
+{
+	int i;
+
+	for (i = 0; i < host->num_slots; i++) {
+		struct dw_mci_slot *slot = host->slot[i];
+
+		if (!slot)
+			continue;
+
+		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;
+	int i;
+
+	pending = mci_readl(host, MINTSTS); /* read-only mask reg */
+
+	/*
+	 * DTO fix - version 2.10a and below, and only if internal DMA
+	 * is configured.
+	 */
+	if (host->quirks & DW_MCI_QUIRK_IDMAC_DTO) {
+		if (!pending &&
+		    ((mci_readl(host, STATUS) >> 17) & 0x1fff))
+			pending |= SDMMC_INT_DATA_OVER;
+	}
+
+	if (pending) {
+		/* Check volt switch first, since it can look like an error */
+		if ((host->state == STATE_SENDING_CMD11) &&
+		    (pending & SDMMC_INT_VOLT_SWITCH)) {
+			unsigned long irqflags;
+
+			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) {
+			mci_writel(host, RINTSTS, DW_MCI_CMD_ERROR_FLAGS);
+			host->cmd_status = pending;
+			smp_wmb();
+			set_bit(EVENT_CMD_COMPLETE, &host->pending_events);
+		}
+
+		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();
+			set_bit(EVENT_DATA_ERROR, &host->pending_events);
+			tasklet_schedule(&host->tasklet);
+		}
+
+		if (pending & SDMMC_INT_DATA_OVER) {
+			mci_writel(host, RINTSTS, SDMMC_INT_DATA_OVER);
+			if (!host->data_status)
+				host->data_status = pending;
+			smp_wmb();
+			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);
+		}
+
+		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) {
+			mci_writel(host, RINTSTS, SDMMC_INT_CMD_DONE);
+			dw_mci_cmd_interrupt(host, pending);
+		}
+
+		if (pending & SDMMC_INT_CD) {
+			mci_writel(host, RINTSTS, SDMMC_INT_CD);
+			dw_mci_handle_cd(host);
+		}
+
+		/* Handle SDIO Interrupts */
+		for (i = 0; i < host->num_slots; i++) {
+			struct dw_mci_slot *slot = host->slot[i];
+
+			if (!slot)
+				continue;
+
+			if (pending & SDMMC_INT_SDIO(slot->sdio_id)) {
+				mci_writel(host, RINTSTS,
+					   SDMMC_INT_SDIO(slot->sdio_id));
+				mmc_signal_sdio_irq(slot->mmc);
+			}
+		}
+
+	}
+
+#ifdef CONFIG_MMC_DW_IDMAC
+	/* Handle DMA 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);
+			host->dma_ops->complete(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);
+			host->dma_ops->complete(host);
+		}
+	}
+#endif
+
+	return IRQ_HANDLED;
+}
+
+#ifdef CONFIG_OF
+/* given a slot id, find out the device node representing that slot */
+static struct device_node *dw_mci_of_find_slot_node(struct device *dev, u8 slot)
+{
+	struct device_node *np;
+	const __be32 *addr;
+	int len;
+
+	if (!dev || !dev->of_node)
+		return NULL;
+
+	for_each_child_of_node(dev->of_node, np) {
+		addr = of_get_property(np, "reg", &len);
+		if (!addr || (len < sizeof(int)))
+			continue;
+		if (be32_to_cpup(addr) == slot)
+			return np;
+	}
+	return NULL;
+}
+
+static struct dw_mci_of_slot_quirks {
+	char *quirk;
+	int id;
+} of_slot_quirks[] = {
+	{
+		.quirk	= "disable-wp",
+		.id	= DW_MCI_SLOT_QUIRK_NO_WRITE_PROTECT,
+	},
+};
+
+static int dw_mci_of_get_slot_quirks(struct device *dev, u8 slot)
+{
+	struct device_node *np = dw_mci_of_find_slot_node(dev, slot);
+	int quirks = 0;
+	int idx;
+
+	/* get quirks */
+	for (idx = 0; idx < ARRAY_SIZE(of_slot_quirks); idx++)
+		if (of_get_property(np, of_slot_quirks[idx].quirk, NULL)) {
+			dev_warn(dev, "Slot quirk %s is deprecated\n",
+					of_slot_quirks[idx].quirk);
+			quirks |= of_slot_quirks[idx].id;
+		}
+
+	return quirks;
+}
+#else /* CONFIG_OF */
+static int dw_mci_of_get_slot_quirks(struct device *dev, u8 slot)
+{
+	return 0;
+}
+#endif /* CONFIG_OF */
+
+static int dw_mci_init_slot(struct dw_mci *host, unsigned int id)
+{
+	struct mmc_host *mmc;
+	struct dw_mci_slot *slot;
+	const struct dw_mci_drv_data *drv_data = host->drv_data;
+	int ctrl_id, ret;
+	u32 freq[2];
+
+	mmc = mmc_alloc_host(sizeof(struct dw_mci_slot), host->dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	slot = mmc_priv(mmc);
+	slot->id = id;
+	slot->sdio_id = host->sdio_id0 + id;
+	slot->mmc = mmc;
+	slot->host = host;
+	host->slot[id] = slot;
+
+	slot->quirks = dw_mci_of_get_slot_quirks(host->dev, slot->id);
+
+	mmc->ops = &dw_mci_ops;
+	if (of_property_read_u32_array(host->dev->of_node,
+				       "clock-freq-min-max", freq, 2)) {
+		mmc->f_min = DW_MCI_FREQ_MIN;
+		mmc->f_max = DW_MCI_FREQ_MAX;
+	} else {
+		mmc->f_min = freq[0];
+		mmc->f_max = freq[1];
+	}
+
+	/*if there are external regulators, get them*/
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret == -EPROBE_DEFER)
+		goto err_host_allocated;
+
+	if (!mmc->ocr_avail)
+		mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+
+	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)
+		mmc->caps |= drv_data->caps[ctrl_id];
+
+	if (host->pdata->caps2)
+		mmc->caps2 = host->pdata->caps2;
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto err_host_allocated;
+
+	if (host->pdata->blk_settings) {
+		mmc->max_segs = host->pdata->blk_settings->max_segs;
+		mmc->max_blk_size = host->pdata->blk_settings->max_blk_size;
+		mmc->max_blk_count = host->pdata->blk_settings->max_blk_count;
+		mmc->max_req_size = host->pdata->blk_settings->max_req_size;
+		mmc->max_seg_size = host->pdata->blk_settings->max_seg_size;
+	} else {
+		/* Useful defaults if platform data is unset. */
+#ifdef CONFIG_MMC_DW_IDMAC
+		mmc->max_segs = host->ring_size;
+		mmc->max_blk_size = 65536;
+		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
+		mmc->max_segs = 64;
+		mmc->max_blk_size = 65536; /* 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;
+#endif /* CONFIG_MMC_DW_IDMAC */
+	}
+
+	if (dw_mci_get_cd(mmc))
+		set_bit(DW_MMC_CARD_PRESENT, &slot->flags);
+	else
+		clear_bit(DW_MMC_CARD_PRESENT, &slot->flags);
+
+	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, unsigned int id)
+{
+	/* Debugfs stuff is cleaned up by mmc core */
+	mmc_remove_host(slot->mmc);
+	slot->host->slot[id] = NULL;
+	mmc_free_host(slot->mmc);
+}
+
+static void dw_mci_init_dma(struct dw_mci *host)
+{
+	int addr_config;
+	/* Check ADDR_CONFIG bit in HCON to find IDMAC address bus width */
+	addr_config = (mci_readl(host, HCON) >> 27) & 0x01;
+
+	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, PAGE_SIZE,
+					  &host->sg_dma, GFP_KERNEL);
+	if (!host->sg_cpu) {
+		dev_err(host->dev, "%s: could not alloc DMA memory\n",
+			__func__);
+		goto no_dma;
+	}
+
+	/* Determine which DMA interface to use */
+#ifdef CONFIG_MMC_DW_IDMAC
+	host->dma_ops = &dw_mci_idmac_ops;
+	dev_info(host->dev, "Using internal DMA controller.\n");
+#endif
+
+	if (!host->dma_ops)
+		goto no_dma;
+
+	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;
+	}
+
+	host->use_dma = 1;
+	return;
+
+no_dma:
+	dev_info(host->dev, "Using PIO mode.\n");
+	host->use_dma = 0;
+	return;
+}
+
+static bool dw_mci_ctrl_reset(struct dw_mci *host, u32 reset)
+{
+	unsigned long timeout = jiffies + msecs_to_jiffies(500);
+	u32 ctrl;
+
+	ctrl = mci_readl(host, CTRL);
+	ctrl |= reset;
+	mci_writel(host, CTRL, ctrl);
+
+	/* wait till resets clear */
+	do {
+		ctrl = mci_readl(host, CTRL);
+		if (!(ctrl & reset))
+			return true;
+	} while (time_before(jiffies, timeout));
+
+	dev_err(host->dev,
+		"Timeout resetting block (ctrl reset %#x)\n",
+		ctrl & reset);
+
+	return false;
+}
+
+static bool dw_mci_reset(struct dw_mci *host)
+{
+	u32 flags = SDMMC_CTRL_RESET | SDMMC_CTRL_FIFO_RESET;
+	bool ret = false;
+
+	/*
+	 * Reseting 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 using dma we wait for dma_req to clear */
+		if (host->use_dma) {
+			unsigned long timeout = jiffies + msecs_to_jiffies(500);
+			u32 status;
+			do {
+				status = mci_readl(host, STATUS);
+				if (!(status & SDMMC_STATUS_DMA_REQ))
+					break;
+				cpu_relax();
+			} while (time_before(jiffies, timeout));
+
+			if (status & SDMMC_STATUS_DMA_REQ) {
+				dev_err(host->dev,
+					"%s: Timeout waiting for dma_req to "
+					"clear during reset\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 IS_ENABLED(CONFIG_MMC_DW_IDMAC)
+	/* It is also recommended that we reset and reprogram idmac */
+	dw_mci_idmac_reset(host);
+#endif
+
+	ret = true;
+
+ciu_out:
+	/* After a CTRL reset we need to have CIU set clock registers  */
+	mci_send_cmd(host->cur_slot, SDMMC_CMD_UPD_CLK, 0);
+
+	return ret;
+}
+
+static void dw_mci_cmd11_timer(unsigned long arg)
+{
+	struct dw_mci *host = (struct dw_mci *)arg;
+
+	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);
+}
+
+#ifdef CONFIG_OF
+static struct dw_mci_of_quirks {
+	char *quirk;
+	int id;
+} of_quirks[] = {
+	{
+		.quirk	= "broken-cd",
+		.id	= DW_MCI_QUIRK_BROKEN_CARD_DETECTION,
+	}, {
+		.quirk	= "disable-wp",
+		.id	= DW_MCI_QUIRK_NO_WRITE_PROTECT,
+	},
+};
+
+static struct dw_mci_board *dw_mci_parse_dt(struct dw_mci *host)
+{
+	struct dw_mci_board *pdata;
+	struct device *dev = host->dev;
+	struct device_node *np = dev->of_node;
+	const struct dw_mci_drv_data *drv_data = host->drv_data;
+	int idx, ret;
+	u32 clock_frequency;
+
+	pdata = devm_kzalloc(dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata)
+		return ERR_PTR(-ENOMEM);
+
+	/* find out number of slots supported */
+	if (of_property_read_u32(dev->of_node, "num-slots",
+				&pdata->num_slots)) {
+		dev_info(dev, "num-slots property not found, "
+				"assuming 1 slot is available\n");
+		pdata->num_slots = 1;
+	}
+
+	/* get quirks */
+	for (idx = 0; idx < ARRAY_SIZE(of_quirks); idx++)
+		if (of_get_property(np, of_quirks[idx].quirk, NULL))
+			pdata->quirks |= of_quirks[idx].id;
+
+	if (of_property_read_u32(np, "fifo-depth", &pdata->fifo_depth))
+		dev_info(dev, "fifo-depth property not found, using "
+				"value of FIFOTH register as default\n");
+
+	of_property_read_u32(np, "card-detect-delay", &pdata->detect_delay_ms);
+
+	if (!of_property_read_u32(np, "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);
+	}
+
+	if (of_find_property(np, "supports-highspeed", NULL))
+		pdata->caps |= MMC_CAP_SD_HIGHSPEED | MMC_CAP_MMC_HIGHSPEED;
+
+	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)
+{
+	struct dw_mci_board *brd = host->pdata;
+	unsigned long irqflags;
+	u32 temp;
+	int i;
+
+	/* No need for CD if broken card detection */
+	if (brd->quirks & DW_MCI_QUIRK_BROKEN_CARD_DETECTION)
+		return;
+
+	/* No need for CD if all slots have a non-error GPIO */
+	for (i = 0; i < host->num_slots; i++) {
+		struct dw_mci_slot *slot = host->slot[i];
+
+		if (IS_ERR_VALUE(mmc_gpio_get_cd(slot->mmc)))
+			break;
+	}
+	if (i == host->num_slots)
+		return;
+
+	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;
+	int init_slots = 0;
+
+	if (!host->pdata) {
+		host->pdata = dw_mci_parse_dt(host);
+		if (IS_ERR(host->pdata)) {
+			dev_err(host->dev, "platform data not available\n");
+			return -EINVAL;
+		}
+	}
+
+	if (host->pdata->num_slots > 1) {
+		dev_err(host->dev,
+			"Platform data must supply num_slots.\n");
+		return -ENODEV;
+	}
+
+	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 (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;
+		}
+	}
+
+	if (drv_data && drv_data->setup_clock) {
+		ret = drv_data->setup_clock(host);
+		if (ret) {
+			dev_err(host->dev,
+				"implementation specific clock setup failed\n");
+			goto err_clk_ciu;
+		}
+	}
+
+	setup_timer(&host->cmd11_timer,
+		    dw_mci_cmd11_timer, (unsigned long)host);
+
+	host->quirks = host->pdata->quirks;
+
+	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 = (mci_readl(host, HCON) >> 7) & 0x7;
+	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))
+		return -ENODEV;
+
+	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->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;
+
+	if (host->pdata->num_slots)
+		host->num_slots = host->pdata->num_slots;
+	else
+		host->num_slots = ((mci_readl(host, HCON) >> 1) & 0x1F) + 1;
+
+	/*
+	 * Enable interrupts for command done, data over, data empty,
+	 * receive ready and error such as transmit, receive timeout, crc error
+	 */
+	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); /* Enable mci interrupt */
+
+	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 */
+	for (i = 0; i < host->num_slots; i++) {
+		ret = dw_mci_init_slot(host, i);
+		if (ret)
+			dev_dbg(host->dev, "slot %d init failed\n", i);
+		else
+			init_slots++;
+	}
+
+	if (init_slots) {
+		dev_info(host->dev, "%d slots initialized\n", init_slots);
+	} else {
+		dev_dbg(host->dev, "attempted to initialize %d slots, "
+					"but failed on all\n", host->num_slots);
+		goto err_dmaunmap;
+	}
+
+	/* Now that slots are all setup, we can enable card detect */
+	dw_mci_enable_cd(host);
+
+	if (host->quirks & DW_MCI_QUIRK_IDMAC_DTO)
+		dev_info(host->dev, "Internal DMAC interrupt fix enabled.\n");
+
+	return 0;
+
+err_dmaunmap:
+	if (host->use_dma && host->dma_ops->exit)
+		host->dma_ops->exit(host);
+
+err_clk_ciu:
+	if (!IS_ERR(host->ciu_clk))
+		clk_disable_unprepare(host->ciu_clk);
+
+err_clk_biu:
+	if (!IS_ERR(host->biu_clk))
+		clk_disable_unprepare(host->biu_clk);
+
+	return ret;
+}
+EXPORT_SYMBOL(dw_mci_probe);
+
+void dw_mci_remove(struct dw_mci *host)
+{
+	int i;
+
+	mci_writel(host, RINTSTS, 0xFFFFFFFF);
+	mci_writel(host, INTMASK, 0); /* disable all mmc interrupt first */
+
+	for (i = 0; i < host->num_slots; i++) {
+		dev_dbg(host->dev, "remove slot %d\n", i);
+		if (host->slot[i])
+			dw_mci_cleanup_slot(host->slot[i], i);
+	}
+
+	/* 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->ciu_clk))
+		clk_disable_unprepare(host->ciu_clk);
+
+	if (!IS_ERR(host->biu_clk))
+		clk_disable_unprepare(host->biu_clk);
+}
+EXPORT_SYMBOL(dw_mci_remove);
+
+
+
+#ifdef CONFIG_PM_SLEEP
+/*
+ * TODO: we should probably disable the clock to the card in the suspend path.
+ */
+int dw_mci_suspend(struct dw_mci *host)
+{
+	return 0;
+}
+EXPORT_SYMBOL(dw_mci_suspend);
+
+int dw_mci_resume(struct dw_mci *host)
+{
+	int i, ret;
+
+	if (!dw_mci_ctrl_reset(host, SDMMC_CTRL_ALL_RESET_FLAGS)) {
+		ret = -ENODEV;
+		return ret;
+	}
+
+	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);
+
+	for (i = 0; i < host->num_slots; i++) {
+		struct dw_mci_slot *slot = host->slot[i];
+		if (!slot)
+			continue;
+		if (slot->mmc->pm_flags & MMC_PM_KEEP_POWER) {
+			dw_mci_set_ios(slot->mmc, &slot->mmc->ios);
+			dw_mci_setup_bus(slot, true);
+		}
+	}
+
+	/* Now that slots are all setup, we can enable card detect */
+	dw_mci_enable_cd(host);
+
+	return 0;
+}
+EXPORT_SYMBOL(dw_mci_resume);
+#endif /* CONFIG_PM_SLEEP */
+
+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..f45ab91de
--- /dev/null
+++ b/drivers/mmc/host/dw_mmc.h
@@ -0,0 +1,291 @@
+/*
+ * 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.
+ *
+ * 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; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef _DW_MMC_H_
+#define _DW_MMC_H_
+
+#define DW_MMC_240A		0x240a
+
+#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_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_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))
+/* 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)
+/* Version ID register define */
+#define SDMMC_GET_VERID(x)		((x) & 0xFFFF)
+/* Card read threshold */
+#define SDMMC_SET_RD_THLD(v, x)		(((v) & 0x1FFF) << 16 | (x))
+#define SDMMC_UHS_18V			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_SLEEP
+extern int dw_mci_suspend(struct dw_mci *host);
+extern int dw_mci_resume(struct dw_mci *host);
+#endif
+
+/**
+ * struct dw_mci_slot - MMC slot state
+ * @mmc: The mmc_host representing this slot.
+ * @host: The MMC controller this slot is using.
+ * @quirks: Slot-level quirks (DW_MCI_SLOT_QUIRK_XXX)
+ * @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 updated clock with reflecting clock divider.
+ *	Keeping track of this helps us to avoid spamming the console
+ *	with CONFIG_MMC_CLKGATE.
+ * @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;
+
+	int			quirks;
+
+	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
+	int			id;
+	int			sdio_id;
+};
+
+/**
+ * dw_mci driver data - dw-mshc implementation specific driver data.
+ * @caps: mmc subsystem specified capabilities of the controller(s).
+ * @init: early implementation specific initialization.
+ * @setup_clock: implementation specific clock configuration.
+ * @prepare_command: handle CMD register extensions.
+ * @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;
+	int		(*init)(struct dw_mci *host);
+	int		(*setup_clock)(struct dw_mci *host);
+	void		(*prepare_command)(struct dw_mci *host, u32 *cmdr);
+	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);
+	int		(*prepare_hs400_tuning)(struct dw_mci *host,
+						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..76e8bce6f
--- /dev/null
+++ b/drivers/mmc/host/jz4740_mmc.c
@@ -0,0 +1,1172 @@
+/*
+ *  Copyright (C) 2009-2010, Lars-Peter Clausen <lars@metafoo.de>
+ *  JZ4740 SD/MMC controller driver
+ *
+ *  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;  either version 2 of the  License, or (at your
+ *  option) any later version.
+ *
+ *  You should have received a copy of the  GNU General Public License along
+ *  with this program; if not, write  to the Free Software Foundation, Inc.,
+ *  675 Mass Ave, Cambridge, MA 02139, USA.
+ *
+ */
+
+#include <linux/mmc/host.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/err.h>
+#include <linux/io.h>
+#include <linux/irq.h>
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/delay.h>
+#include <linux/scatterlist.h>
+#include <linux/clk.h>
+
+#include <linux/bitops.h>
+#include <linux/gpio.h>
+#include <asm/mach-jz4740/gpio.h>
+#include <asm/cacheflush.h>
+#include <linux/dma-mapping.h>
+#include <linux/dmaengine.h>
+
+#include <asm/mach-jz4740/dma.h>
+#include <asm/mach-jz4740/jz4740_mmc.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_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_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_CLK_RATE 24000000
+
+enum jz4740_mmc_state {
+	JZ4740_MMC_STATE_READ_RESPONSE,
+	JZ4740_MMC_STATE_TRANSFER_DATA,
+	JZ4740_MMC_STATE_SEND_STOP,
+	JZ4740_MMC_STATE_DONE,
+};
+
+struct jz4740_mmc_host_next {
+	int sg_len;
+	s32 cookie;
+};
+
+struct jz4740_mmc_host {
+	struct mmc_host *mmc;
+	struct platform_device *pdev;
+	struct jz4740_mmc_platform_data *pdata;
+	struct clk *clk;
+
+	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;
+
+	uint16_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;
+	struct jz4740_mmc_host_next next_data;
+	bool use_dma;
+	int sg_len;
+
+/* 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
+};
+
+/*----------------------------------------------------------------------------*/
+/* 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)
+{
+	dma_cap_mask_t mask;
+
+	dma_cap_zero(mask);
+	dma_cap_set(DMA_SLAVE, mask);
+
+	host->dma_tx = dma_request_channel(mask, NULL, host);
+	if (!host->dma_tx) {
+		dev_err(mmc_dev(host->mmc), "Failed to get dma_tx channel\n");
+		return -ENODEV;
+	}
+
+	host->dma_rx = dma_request_channel(mask, NULL, host);
+	if (!host->dma_rx) {
+		dev_err(mmc_dev(host->mmc), "Failed to get dma_rx channel\n");
+		goto free_master_write;
+	}
+
+	/* Initialize DMA pre request cookie */
+	host->next_data.cookie = 1;
+
+	return 0;
+
+free_master_write:
+	dma_release_channel(host->dma_tx);
+	return -ENODEV;
+}
+
+static inline int jz4740_mmc_get_dma_dir(struct mmc_data *data)
+{
+	return (data->flags & MMC_DATA_READ) ? DMA_FROM_DEVICE : DMA_TO_DEVICE;
+}
+
+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 = jz4740_mmc_get_dma_dir(data);
+
+	dma_unmap_sg(chan->device->dev, data->sg, data->sg_len, dir);
+}
+
+/* Prepares DMA data for current/next transfer, returns non-zero on failure */
+static int jz4740_mmc_prepare_dma_data(struct jz4740_mmc_host *host,
+				       struct mmc_data *data,
+				       struct jz4740_mmc_host_next *next,
+				       struct dma_chan *chan)
+{
+	struct jz4740_mmc_host_next *next_data = &host->next_data;
+	enum dma_data_direction dir = jz4740_mmc_get_dma_dir(data);
+	int sg_len;
+
+	if (!next && data->host_cookie &&
+	    data->host_cookie != host->next_data.cookie) {
+		dev_warn(mmc_dev(host->mmc),
+			 "[%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) {
+		sg_len = dma_map_sg(chan->device->dev,
+				    data->sg,
+				    data->sg_len,
+				    dir);
+
+	} else {
+		sg_len = next_data->sg_len;
+		next_data->sg_len = 0;
+	}
+
+	if (sg_len <= 0) {
+		dev_err(mmc_dev(host->mmc),
+			"Failed to map scatterlist for DMA operation\n");
+		return -EINVAL;
+	}
+
+	if (next) {
+		next->sg_len = sg_len;
+		data->host_cookie = ++next->cookie < 0 ? 1 : next->cookie;
+	} else
+		host->sg_len = sg_len;
+
+	return 0;
+}
+
+static int jz4740_mmc_start_dma_transfer(struct jz4740_mmc_host *host,
+					 struct mmc_data *data)
+{
+	int ret;
+	struct dma_chan *chan;
+	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,
+	};
+
+	if (data->flags & MMC_DATA_WRITE) {
+		conf.direction = DMA_MEM_TO_DEV;
+		conf.dst_addr = host->mem_res->start + JZ_REG_MMC_TXFIFO;
+		conf.slave_id = JZ4740_DMA_TYPE_MMC_TRANSMIT;
+		chan = host->dma_tx;
+	} else {
+		conf.direction = DMA_DEV_TO_MEM;
+		conf.src_addr = host->mem_res->start + JZ_REG_MMC_RXFIFO;
+		conf.slave_id = JZ4740_DMA_TYPE_MMC_RECEIVE;
+		chan = host->dma_rx;
+	}
+
+	ret = jz4740_mmc_prepare_dma_data(host, data, NULL, chan);
+	if (ret)
+		return ret;
+
+	dmaengine_slave_config(chan, &conf);
+	desc = dmaengine_prep_slave_sg(chan,
+				       data->sg,
+				       host->sg_len,
+				       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:
+	jz4740_mmc_dma_unmap(host, data);
+	return -ENOMEM;
+}
+
+static void jz4740_mmc_pre_request(struct mmc_host *mmc,
+				   struct mmc_request *mrq,
+				   bool is_first_req)
+{
+	struct jz4740_mmc_host *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+	struct jz4740_mmc_host_next *next_data = &host->next_data;
+
+	BUG_ON(data->host_cookie);
+
+	if (host->use_dma) {
+		struct dma_chan *chan = jz4740_mmc_get_dma_chan(host, data);
+
+		if (jz4740_mmc_prepare_dma_data(host, data, next_data, chan))
+			data->host_cookie = 0;
+	}
+}
+
+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 (host->use_dma && data->host_cookie) {
+		jz4740_mmc_dma_unmap(host, data);
+		data->host_cookie = 0;
+	}
+
+	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;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	writew(host->irq_mask, host->base + JZ_REG_MMC_IMASK);
+}
+
+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;
+
+	req = host->req;
+	host->req = NULL;
+
+	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;
+	uint16_t status;
+
+	do {
+		status = readw(host->base + JZ_REG_MMC_IREG);
+	} while (!(status & irq) && --timeout);
+
+	if (timeout == 0) {
+		set_bit(0, &host->waiting);
+		mod_timer(&host->timeout_timer, jiffies + 5*HZ);
+		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;
+	uint16_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(unsigned long data)
+{
+	struct jz4740_mmc_host *host = (struct jz4740_mmc_host *)data;
+
+	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 (cmd->data->flags & MMC_DATA_STREAM)
+			cmdat |= JZ_MMC_CMDAT_STREAM;
+		if (host->use_dma)
+			cmdat |= JZ_MMC_CMDAT_DMA_EN;
+
+		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);
+
+	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;
+		}
+		writew(JZ_MMC_IRQ_DATA_TRAN_DONE, host->base + JZ_REG_MMC_IREG);
+
+	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;
+	uint16_t irq_reg, status, tmp;
+
+	irq_reg = readw(host->base + JZ_REG_MMC_IREG);
+
+	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)
+		writew(tmp & ~irq_reg, host->base + JZ_REG_MMC_IREG);
+
+	if (irq_reg & JZ_MMC_IRQ_SDIO) {
+		writew(JZ_MMC_IRQ_SDIO, host->base + JZ_REG_MMC_IREG);
+		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);
+
+			status = readl(host->base + JZ_REG_MMC_STATUS);
+
+			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);
+			writew(irq_reg, host->base + JZ_REG_MMC_IREG);
+
+			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, JZ_MMC_CLK_RATE);
+
+	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);
+	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;
+
+	writew(0xffff, host->base + JZ_REG_MMC_IREG);
+
+	writew(JZ_MMC_IRQ_END_CMD_RES, host->base + JZ_REG_MMC_IREG);
+	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 + 5*HZ);
+	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 (gpio_is_valid(host->pdata->gpio_power))
+			gpio_set_value(host->pdata->gpio_power,
+					!host->pdata->power_active_low);
+		host->cmdat |= JZ_MMC_CMDAT_INIT;
+		clk_prepare_enable(host->clk);
+		break;
+	case MMC_POWER_ON:
+		break;
+	default:
+		if (gpio_is_valid(host->pdata->gpio_power))
+			gpio_set_value(host->pdata->gpio_power,
+					host->pdata->power_active_low);
+		clk_disable_unprepare(host->clk);
+		break;
+	}
+
+	switch (ios->bus_width) {
+	case MMC_BUS_WIDTH_1:
+		host->cmdat &= ~JZ_MMC_CMDAT_BUS_WIDTH_4BIT;
+		break;
+	case MMC_BUS_WIDTH_4:
+		host->cmdat |= JZ_MMC_CMDAT_BUS_WIDTH_4BIT;
+		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 jz_gpio_bulk_request jz4740_mmc_pins[] = {
+	JZ_GPIO_BULK_PIN(MSC_CMD),
+	JZ_GPIO_BULK_PIN(MSC_CLK),
+	JZ_GPIO_BULK_PIN(MSC_DATA0),
+	JZ_GPIO_BULK_PIN(MSC_DATA1),
+	JZ_GPIO_BULK_PIN(MSC_DATA2),
+	JZ_GPIO_BULK_PIN(MSC_DATA3),
+};
+
+static int jz4740_mmc_request_gpio(struct device *dev, int gpio,
+	const char *name, bool output, int value)
+{
+	int ret;
+
+	if (!gpio_is_valid(gpio))
+		return 0;
+
+	ret = gpio_request(gpio, name);
+	if (ret) {
+		dev_err(dev, "Failed to request %s gpio: %d\n", name, ret);
+		return ret;
+	}
+
+	if (output)
+		gpio_direction_output(gpio, value);
+	else
+		gpio_direction_input(gpio);
+
+	return 0;
+}
+
+static int jz4740_mmc_request_gpios(struct mmc_host *mmc,
+	struct platform_device *pdev)
+{
+	struct jz4740_mmc_platform_data *pdata = pdev->dev.platform_data;
+	int ret = 0;
+
+	if (!pdata)
+		return 0;
+
+	if (!pdata->card_detect_active_low)
+		mmc->caps2 |= MMC_CAP2_CD_ACTIVE_HIGH;
+	if (!pdata->read_only_active_low)
+		mmc->caps2 |= MMC_CAP2_RO_ACTIVE_HIGH;
+
+	if (gpio_is_valid(pdata->gpio_card_detect)) {
+		ret = mmc_gpio_request_cd(mmc, pdata->gpio_card_detect, 0);
+		if (ret)
+			return ret;
+	}
+
+	if (gpio_is_valid(pdata->gpio_read_only)) {
+		ret = mmc_gpio_request_ro(mmc, pdata->gpio_read_only);
+		if (ret)
+			return ret;
+	}
+
+	return jz4740_mmc_request_gpio(&pdev->dev, pdata->gpio_power,
+			"MMC read only", true, pdata->power_active_low);
+}
+
+static void jz4740_mmc_free_gpios(struct platform_device *pdev)
+{
+	struct jz4740_mmc_platform_data *pdata = pdev->dev.platform_data;
+
+	if (!pdata)
+		return;
+
+	if (gpio_is_valid(pdata->gpio_power))
+		gpio_free(pdata->gpio_power);
+}
+
+static inline size_t jz4740_mmc_num_pins(struct jz4740_mmc_host *host)
+{
+	size_t num_pins = ARRAY_SIZE(jz4740_mmc_pins);
+	if (host->pdata && host->pdata->data_1bit)
+		num_pins -= 3;
+
+	return num_pins;
+}
+
+static int jz4740_mmc_probe(struct platform_device* pdev)
+{
+	int ret;
+	struct mmc_host *mmc;
+	struct jz4740_mmc_host *host;
+	struct jz4740_mmc_platform_data *pdata;
+
+	pdata = pdev->dev.platform_data;
+
+	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);
+	host->pdata = pdata;
+
+	host->irq = platform_get_irq(pdev, 0);
+	if (host->irq < 0) {
+		ret = host->irq;
+		dev_err(&pdev->dev, "Failed to get platform irq: %d\n", ret);
+		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;
+	}
+
+	ret = jz_gpio_bulk_request(jz4740_mmc_pins, jz4740_mmc_num_pins(host));
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to request mmc pins: %d\n", ret);
+		goto err_free_host;
+	}
+
+	ret = jz4740_mmc_request_gpios(mmc, pdev);
+	if (ret)
+		goto err_gpio_bulk_free;
+
+	mmc->ops = &jz4740_mmc_ops;
+	mmc->f_min = JZ_MMC_CLK_RATE / 128;
+	mmc->f_max = JZ_MMC_CLK_RATE;
+	mmc->ocr_avail = MMC_VDD_32_33 | MMC_VDD_33_34;
+	mmc->caps = (pdata && pdata->data_1bit) ? 0 : MMC_CAP_4_BIT_DATA;
+	mmc->caps |= MMC_CAP_SDIO_IRQ;
+
+	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 = 0xffff;
+
+	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_gpios;
+	}
+
+	jz4740_mmc_reset(host);
+	jz4740_mmc_clock_disable(host);
+	setup_timer(&host->timeout_timer, jz4740_mmc_timeout,
+			(unsigned long)host);
+	/* It is not important when it times out, it just needs to timeout. */
+	set_timer_slack(&host->timeout_timer, HZ);
+
+	host->use_dma = true;
+	if (host->use_dma && jz4740_mmc_acquire_dma_channels(host) != 0)
+		host->use_dma = false;
+
+	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_free_irq;
+	}
+	dev_info(&pdev->dev, "JZ 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_4_BIT_DATA) ? 4 : 1);
+
+	return 0;
+
+err_free_irq:
+	free_irq(host->irq, host);
+err_free_gpios:
+	jz4740_mmc_free_gpios(pdev);
+err_gpio_bulk_free:
+	if (host->use_dma)
+		jz4740_mmc_release_dma_channels(host);
+	jz_gpio_bulk_free(jz4740_mmc_pins, jz4740_mmc_num_pins(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);
+
+	jz4740_mmc_free_gpios(pdev);
+	jz_gpio_bulk_free(jz4740_mmc_pins, jz4740_mmc_num_pins(host));
+
+	if (host->use_dma)
+		jz4740_mmc_release_dma_channels(host);
+
+	mmc_free_host(host->mmc);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+
+static int jz4740_mmc_suspend(struct device *dev)
+{
+	struct jz4740_mmc_host *host = dev_get_drvdata(dev);
+
+	jz_gpio_bulk_suspend(jz4740_mmc_pins, jz4740_mmc_num_pins(host));
+
+	return 0;
+}
+
+static int jz4740_mmc_resume(struct device *dev)
+{
+	struct jz4740_mmc_host *host = dev_get_drvdata(dev);
+
+	jz_gpio_bulk_resume(jz4740_mmc_pins, jz4740_mmc_num_pins(host));
+
+	return 0;
+}
+
+static SIMPLE_DEV_PM_OPS(jz4740_mmc_pm_ops, jz4740_mmc_suspend,
+	jz4740_mmc_resume);
+#define JZ4740_MMC_PM_OPS (&jz4740_mmc_pm_ops)
+#else
+#define JZ4740_MMC_PM_OPS NULL
+#endif
+
+static struct platform_driver jz4740_mmc_driver = {
+	.probe = jz4740_mmc_probe,
+	.remove = jz4740_mmc_remove,
+	.driver = {
+		.name = "jz4740-mmc",
+		.pm = 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/mmc_spi.c b/drivers/mmc/host/mmc_spi.c
new file mode 100644
index 000000000..ae19d83bb
--- /dev/null
+++ b/drivers/mmc/host/mmc_spi.c
@@ -0,0 +1,1531 @@
+/*
+ * mmc_spi.c - 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>
+ *
+ *
+ * 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; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+#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
+
+
+/* These fixed timeouts come from the latest SD specs, which say to ignore
+ * the CSD values.  The R1B value is for card erase (e.g. the "I forgot the
+ * card's password" scenario); it's mostly applied to STOP_TRANSMISSION after
+ * reads which takes nowhere near that long.  Older cards may be able to use
+ * shorter timeouts ... but why bother?
+ */
+#define r1b_timeout		(HZ * 3)
+
+/* 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)
+{
+	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)
+			mmc_spi_wait_unbusy(host, r1b_timeout);
+		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 (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;
+
+	if (data->flags & MMC_DATA_READ)
+		direction = DMA_FROM_DEVICE;
+	else
+		direction = DMA_TO_DEVICE;
+	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 +
+		  data->timeout_clks * 1000000 / clock_rate;
+	timeout = usecs_to_jiffies((unsigned int)(timeout / 1000)) + 1;
+
+	/* Handle scatterlist segments one at a time, with synch for
+	 * each 512-byte block
+	 */
+	for (sg = data->sg, n_sg = data->sg_len; n_sg; n_sg--, 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 (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, r1b_timeout);
+	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.
+	 */
+	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;
+}
+
+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;
+
+	if (spi->master->dev.parent->dma_mask) {
+		struct device	*dev = spi->master->dev.parent;
+
+		host->dma_dev = dev;
+		host->ones_dma = dma_map_single(dev, ones,
+				MMC_SPI_BLOCKSIZE, DMA_TO_DEVICE);
+		host->data_dma = dma_map_single(dev, host->data,
+				sizeof(*host->data), DMA_BIDIRECTIONAL);
+
+		/* REVISIT in theory those map operations can fail... */
+
+		dma_sync_single_for_cpu(host->dma_dev,
+				host->data_dma, sizeof(*host->data),
+				DMA_BIDIRECTIONAL);
+	}
+
+	/* 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_add_host;
+
+	if (host->pdata && host->pdata->flags & MMC_SPI_USE_CD_GPIO) {
+		status = mmc_gpio_request_cd(mmc, host->pdata->cd_gpio,
+					     host->pdata->cd_debounce);
+		if (status != 0)
+			goto fail_add_host;
+		mmc_gpiod_request_cd_irq(mmc);
+	}
+
+	if (host->pdata && host->pdata->flags & MMC_SPI_USE_RO_GPIO) {
+		has_ro = true;
+		status = mmc_gpio_request_ro(mmc, host->pdata->ro_gpio);
+		if (status != 0)
+			goto fail_add_host;
+	}
+
+	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_add_host:
+	mmc_remove_host (mmc);
+fail_glue_init:
+	if (host->dma_dev)
+		dma_unmap_single(host->dma_dev, host->data_dma,
+				sizeof(*host->data), DMA_BIDIRECTIONAL);
+	kfree(host->data);
+
+fail_nobuf1:
+	mmc_free_host(mmc);
+	mmc_spi_put_pdata(spi);
+	dev_set_drvdata(&spi->dev, NULL);
+
+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;
+
+	if (mmc) {
+		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);
+
+		if (host->dma_dev) {
+			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);
+		}
+
+		kfree(host->data);
+		kfree(host->ones);
+
+		spi->max_speed_hz = mmc->f_max;
+		mmc_free_host(mmc);
+		mmc_spi_put_pdata(spi);
+		dev_set_drvdata(&spi->dev, NULL);
+	}
+	return 0;
+}
+
+static const struct of_device_id mmc_spi_of_match_table[] = {
+	{ .compatible = "mmc-spi-slot", },
+	{},
+};
+
+static struct spi_driver mmc_spi_driver = {
+	.driver = {
+		.name =		"mmc_spi",
+		.owner =	THIS_MODULE,
+		.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..fb266745f
--- /dev/null
+++ b/drivers/mmc/host/mmci.c
@@ -0,0 +1,1927 @@
+/*
+ *  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
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#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/pm.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/amba/bus.h>
+#include <linux/clk.h>
+#include <linux/scatterlist.h>
+#include <linux/gpio.h>
+#include <linux/of_gpio.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 <asm/div64.h>
+#include <asm/io.h>
+#include <asm/sizes.h>
+
+#include "mmci.h"
+#include "mmci_qcom_dml.h"
+
+#define DRIVER_NAME "mmci-pl18x"
+
+static unsigned int fmax = 515633;
+
+/**
+ * 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
+ * @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
+ * @datactrl_mask_ddrmode: ddr mode mask in datactrl register.
+ * @blksz_datactrl16: true if Block size is at b16..b30 position in datactrl register
+ * @blksz_datactrl4: true if Block size is at b4..b16 position in datactrl
+ *		     register
+ * @datactrl_mask_sdio: SDIO enable mask in datactrl register
+ * @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 busy detection on dat0 is supported
+ * @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.
+ */
+struct variant_data {
+	unsigned int		clkreg;
+	unsigned int		clkreg_enable;
+	unsigned int		clkreg_8bit_bus_enable;
+	unsigned int		clkreg_neg_edge_enable;
+	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;
+	bool			st_sdio;
+	bool			st_clkdiv;
+	bool			blksz_datactrl16;
+	bool			blksz_datactrl4;
+	u32			pwrreg_powerup;
+	u32			f_max;
+	bool			signal_direction;
+	bool			pwrreg_clkgate;
+	bool			busy_detect;
+	bool			pwrreg_nopower;
+	bool			explicit_mclk_control;
+	bool			qcom_fifo;
+	bool			qcom_dml;
+	bool			reversed_irq_handling;
+};
+
+static struct variant_data variant_arm = {
+	.fifosize		= 16 * 4,
+	.fifohalfsize		= 8 * 4,
+	.datalength_bits	= 16,
+	.pwrreg_powerup		= MCI_PWR_UP,
+	.f_max			= 100000000,
+	.reversed_irq_handling	= true,
+};
+
+static struct variant_data variant_arm_extended_fifo = {
+	.fifosize		= 128 * 4,
+	.fifohalfsize		= 64 * 4,
+	.datalength_bits	= 16,
+	.pwrreg_powerup		= MCI_PWR_UP,
+	.f_max			= 100000000,
+};
+
+static struct variant_data variant_arm_extended_fifo_hwfc = {
+	.fifosize		= 128 * 4,
+	.fifohalfsize		= 64 * 4,
+	.clkreg_enable		= MCI_ARM_HWFCEN,
+	.datalength_bits	= 16,
+	.pwrreg_powerup		= MCI_PWR_UP,
+	.f_max			= 100000000,
+};
+
+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,
+	.datalength_bits	= 16,
+	.datactrl_mask_sdio	= MCI_ST_DPSM_SDIOEN,
+	.st_sdio			= true,
+	.pwrreg_powerup		= MCI_PWR_ON,
+	.f_max			= 100000000,
+	.signal_direction	= true,
+	.pwrreg_clkgate		= true,
+	.pwrreg_nopower		= true,
+};
+
+static struct variant_data variant_nomadik = {
+	.fifosize		= 16 * 4,
+	.fifohalfsize		= 8 * 4,
+	.clkreg			= MCI_CLK_ENABLE,
+	.datalength_bits	= 24,
+	.datactrl_mask_sdio	= MCI_ST_DPSM_SDIOEN,
+	.st_sdio		= true,
+	.st_clkdiv		= true,
+	.pwrreg_powerup		= MCI_PWR_ON,
+	.f_max			= 100000000,
+	.signal_direction	= true,
+	.pwrreg_clkgate		= true,
+	.pwrreg_nopower		= true,
+};
+
+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,
+	.datalength_bits	= 24,
+	.datactrl_mask_sdio	= MCI_ST_DPSM_SDIOEN,
+	.st_sdio		= true,
+	.st_clkdiv		= true,
+	.pwrreg_powerup		= MCI_PWR_ON,
+	.f_max			= 100000000,
+	.signal_direction	= true,
+	.pwrreg_clkgate		= true,
+	.busy_detect		= true,
+	.pwrreg_nopower		= true,
+};
+
+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,
+	.datactrl_mask_ddrmode	= MCI_ST_DPSM_DDRMODE,
+	.datalength_bits	= 24,
+	.datactrl_mask_sdio	= MCI_ST_DPSM_SDIOEN,
+	.st_sdio		= true,
+	.st_clkdiv		= true,
+	.blksz_datactrl16	= true,
+	.pwrreg_powerup		= MCI_PWR_ON,
+	.f_max			= 100000000,
+	.signal_direction	= true,
+	.pwrreg_clkgate		= true,
+	.busy_detect		= true,
+	.pwrreg_nopower		= true,
+};
+
+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,
+	.data_cmd_enable	= MCI_QCOM_CSPM_DATCMD,
+	.blksz_datactrl4	= true,
+	.datalength_bits	= 24,
+	.pwrreg_powerup		= MCI_PWR_UP,
+	.f_max			= 208000000,
+	.explicit_mclk_control	= true,
+	.qcom_fifo		= true,
+	.qcom_dml		= true,
+};
+
+static int mmci_card_busy(struct mmc_host *mmc)
+{
+	struct mmci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+	int busy = 0;
+
+	pm_runtime_get_sync(mmc_dev(mmc));
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (readl(host->base + MMCISTATUS) & MCI_ST_CARDBUSY)
+		busy = 1;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	pm_runtime_mark_last_busy(mmc_dev(mmc));
+	pm_runtime_put_autosuspend(mmc_dev(mmc));
+
+	return busy;
+}
+
+/*
+ * Validate mmc prerequisites
+ */
+static int mmci_validate_data(struct mmci_host *host,
+			      struct mmc_data *data)
+{
+	if (!data)
+		return 0;
+
+	if (!is_power_of_2(data->blksz)) {
+		dev_err(mmc_dev(host->mmc),
+			"unsupported block size (%d bytes)\n", data->blksz);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+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
+ */
+static 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
+ */
+static 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 ST Micro busy mode if enabled */
+	datactrl |= host->datactrl_reg & MCI_ST_DPSM_BUSYMODE;
+
+	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_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);
+
+	pm_runtime_mark_last_busy(mmc_dev(host->mmc));
+	pm_runtime_put_autosuspend(mmc_dev(host->mmc));
+}
+
+static void mmci_set_mask1(struct mmci_host *host, unsigned int mask)
+{
+	void __iomem *base = host->base;
+
+	if (host->singleirq) {
+		unsigned int mask0 = readl(base + MMCIMASK0);
+
+		mask0 &= ~MCI_IRQ1MASK;
+		mask0 |= mask;
+
+		writel(mask0, base + MMCIMASK0);
+	}
+
+	writel(mask, base + MMCIMASK1);
+}
+
+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);
+}
+
+/*
+ * 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
+static void mmci_dma_setup(struct mmci_host *host)
+{
+	const char *rxname, *txname;
+	struct variant_data *variant = host->variant;
+
+	host->dma_rx_channel = dma_request_slave_channel(mmc_dev(host->mmc), "rx");
+	host->dma_tx_channel = dma_request_slave_channel(mmc_dev(host->mmc), "tx");
+
+	/* initialize pre request cookie */
+	host->next_data.cookie = 1;
+
+	/*
+	 * 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 (host->dma_rx_channel && !host->dma_tx_channel)
+		host->dma_tx_channel = host->dma_rx_channel;
+
+	if (host->dma_rx_channel)
+		rxname = dma_chan_name(host->dma_rx_channel);
+	else
+		rxname = "none";
+
+	if (host->dma_tx_channel)
+		txname = dma_chan_name(host->dma_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 (host->dma_tx_channel) {
+		struct device *dev = host->dma_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 (host->dma_rx_channel) {
+		struct device *dev = host->dma_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 (variant->qcom_dml && host->dma_rx_channel && host->dma_tx_channel)
+		if (dml_hw_init(host, host->mmc->parent->of_node))
+			variant->qcom_dml = false;
+}
+
+/*
+ * This is used in or so inline it
+ * so it can be discarded.
+ */
+static inline void mmci_dma_release(struct mmci_host *host)
+{
+	if (host->dma_rx_channel)
+		dma_release_channel(host->dma_rx_channel);
+	if (host->dma_tx_channel)
+		dma_release_channel(host->dma_tx_channel);
+	host->dma_rx_channel = host->dma_tx_channel = NULL;
+}
+
+static void mmci_dma_data_error(struct mmci_host *host)
+{
+	dev_err(mmc_dev(host->mmc), "error during DMA transfer!\n");
+	dmaengine_terminate_all(host->dma_current);
+	host->dma_current = NULL;
+	host->dma_desc_current = NULL;
+	host->data->host_cookie = 0;
+}
+
+static void mmci_dma_unmap(struct mmci_host *host, struct mmc_data *data)
+{
+	struct dma_chan *chan;
+	enum dma_data_direction dir;
+
+	if (data->flags & MMC_DATA_READ) {
+		dir = DMA_FROM_DEVICE;
+		chan = host->dma_rx_channel;
+	} else {
+		dir = DMA_TO_DEVICE;
+		chan = host->dma_tx_channel;
+	}
+
+	dma_unmap_sg(chan->device->dev, data->sg, data->sg_len, dir);
+}
+
+static void mmci_dma_finalize(struct mmci_host *host, struct mmc_data *data)
+{
+	u32 status;
+	int i;
+
+	/* 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_data_error(host);
+		if (!data->error)
+			data->error = -EIO;
+	}
+
+	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_current = NULL;
+	host->dma_desc_current = NULL;
+}
+
+/* prepares DMA channel and DMA descriptor, returns non-zero on failure */
+static int __mmci_dma_prep_data(struct mmci_host *host, struct mmc_data *data,
+				struct dma_chan **dma_chan,
+				struct dma_async_tx_descriptor **dma_desc)
+{
+	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;
+	enum dma_data_direction buffer_dirn;
+	int nr_sg;
+	unsigned long flags = DMA_CTRL_ACK;
+
+	if (data->flags & MMC_DATA_READ) {
+		conf.direction = DMA_DEV_TO_MEM;
+		buffer_dirn = DMA_FROM_DEVICE;
+		chan = host->dma_rx_channel;
+	} else {
+		conf.direction = DMA_MEM_TO_DEV;
+		buffer_dirn = DMA_TO_DEVICE;
+		chan = host->dma_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;
+
+	device = chan->device;
+	nr_sg = dma_map_sg(device->dev, data->sg, data->sg_len, buffer_dirn);
+	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, buffer_dirn);
+	return -ENOMEM;
+}
+
+static inline int mmci_dma_prep_data(struct mmci_host *host,
+				     struct mmc_data *data)
+{
+	/* Check if next job is already prepared. */
+	if (host->dma_current && host->dma_desc_current)
+		return 0;
+
+	/* No job were prepared thus do it now. */
+	return __mmci_dma_prep_data(host, data, &host->dma_current,
+				    &host->dma_desc_current);
+}
+
+static inline int mmci_dma_prep_next(struct mmci_host *host,
+				     struct mmc_data *data)
+{
+	struct mmci_host_next *nd = &host->next_data;
+	return __mmci_dma_prep_data(host, data, &nd->dma_chan, &nd->dma_desc);
+}
+
+static int mmci_dma_start_data(struct mmci_host *host, unsigned int datactrl)
+{
+	int ret;
+	struct mmc_data *data = host->data;
+
+	ret = mmci_dma_prep_data(host, host->data);
+	if (ret)
+		return ret;
+
+	/* 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);
+	dmaengine_submit(host->dma_desc_current);
+	dma_async_issue_pending(host->dma_current);
+
+	if (host->variant->qcom_dml)
+		dml_start_xfer(host, data);
+
+	datactrl |= MCI_DPSM_DMAENABLE;
+
+	/* 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_get_next_data(struct mmci_host *host, struct mmc_data *data)
+{
+	struct mmci_host_next *next = &host->next_data;
+
+	WARN_ON(data->host_cookie && data->host_cookie != next->cookie);
+	WARN_ON(!data->host_cookie && (next->dma_desc || next->dma_chan));
+
+	host->dma_desc_current = next->dma_desc;
+	host->dma_current = next->dma_chan;
+	next->dma_desc = NULL;
+	next->dma_chan = NULL;
+}
+
+static void mmci_pre_request(struct mmc_host *mmc, struct mmc_request *mrq,
+			     bool is_first_req)
+{
+	struct mmci_host *host = mmc_priv(mmc);
+	struct mmc_data *data = mrq->data;
+	struct mmci_host_next *nd = &host->next_data;
+
+	if (!data)
+		return;
+
+	BUG_ON(data->host_cookie);
+
+	if (mmci_validate_data(host, data))
+		return;
+
+	if (!mmci_dma_prep_next(host, data))
+		data->host_cookie = ++nd->cookie < 0 ? 1 : nd->cookie;
+}
+
+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_dma_unmap(host, data);
+
+	if (err) {
+		struct mmci_host_next *next = &host->next_data;
+		struct dma_chan *chan;
+		if (data->flags & MMC_DATA_READ)
+			chan = host->dma_rx_channel;
+		else
+			chan = host->dma_tx_channel;
+		dmaengine_terminate_all(chan);
+
+		if (host->dma_desc_current == next->dma_desc)
+			host->dma_desc_current = NULL;
+
+		if (host->dma_current == next->dma_chan)
+			host->dma_current = NULL;
+
+		next->dma_desc = NULL;
+		next->dma_chan = NULL;
+		data->host_cookie = 0;
+	}
+}
+
+#else
+/* Blank functions if the DMA engine is not available */
+static void mmci_get_next_data(struct mmci_host *host, struct mmc_data *data)
+{
+}
+static inline void mmci_dma_setup(struct mmci_host *host)
+{
+}
+
+static inline void mmci_dma_release(struct mmci_host *host)
+{
+}
+
+static inline void mmci_dma_unmap(struct mmci_host *host, struct mmc_data *data)
+{
+}
+
+static inline void mmci_dma_finalize(struct mmci_host *host,
+				     struct mmc_data *data)
+{
+}
+
+static inline void mmci_dma_data_error(struct mmci_host *host)
+{
+}
+
+static inline int mmci_dma_start_data(struct mmci_host *host, unsigned int datactrl)
+{
+	return -ENOSYS;
+}
+
+#define mmci_pre_request NULL
+#define mmci_post_request NULL
+
+#endif
+
+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;
+	int blksz_bits;
+
+	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);
+
+	blksz_bits = ffs(data->blksz) - 1;
+	BUG_ON(1 << blksz_bits != data->blksz);
+
+	if (variant->blksz_datactrl16)
+		datactrl = MCI_DPSM_ENABLE | (data->blksz << 16);
+	else if (variant->blksz_datactrl4)
+		datactrl = MCI_DPSM_ENABLE | (data->blksz << 4);
+	else
+		datactrl = MCI_DPSM_ENABLE | blksz_bits << 4;
+
+	if (data->flags & MMC_DATA_READ)
+		datactrl |= MCI_DPSM_DIRECTION;
+
+	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_data(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;
+
+	dev_dbg(mmc_dev(host->mmc), "op %02x arg %08x flags %08x\n",
+	    cmd->opcode, cmd->arg, cmd->flags);
+
+	if (readl(base + MMCICOMMAND) & MCI_CPSM_ENABLE) {
+		writel(0, base + MMCICOMMAND);
+		mmci_reg_delay(host);
+	}
+
+	c |= cmd->opcode | MCI_CPSM_ENABLE;
+	if (cmd->flags & MMC_RSP_PRESENT) {
+		if (cmd->flags & MMC_RSP_136)
+			c |= MCI_CPSM_LONGRSP;
+		c |= MCI_CPSM_RESPONSE;
+	}
+	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_data_irq(struct mmci_host *host, struct mmc_data *data,
+	      unsigned int status)
+{
+	/* Make sure we have data to handle */
+	if (!data)
+		return;
+
+	/* First check for errors */
+	if (status & (MCI_DATACRCFAIL|MCI_DATATIMEOUT|MCI_STARTBITERR|
+		      MCI_TXUNDERRUN|MCI_RXOVERRUN)) {
+		u32 remain, success;
+
+		/* Terminate the DMA transfer */
+		if (dma_inprogress(host)) {
+			mmci_dma_data_error(host);
+			mmci_dma_unmap(host, data);
+		}
+
+		/*
+		 * 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.
+		 */
+		remain = readl(host->base + MMCIDATACNT);
+		success = data->blksz * data->blocks - remain;
+
+		dev_dbg(mmc_dev(host->mmc), "MCI ERROR IRQ, status 0x%08x at 0x%08x\n",
+			status, success);
+		if (status & MCI_DATACRCFAIL) {
+			/* Last block was not successful */
+			success -= 1;
+			data->error = -EILSEQ;
+		} else if (status & MCI_DATATIMEOUT) {
+			data->error = -ETIMEDOUT;
+		} else if (status & MCI_STARTBITERR) {
+			data->error = -ECOMM;
+		} else if (status & MCI_TXUNDERRUN) {
+			data->error = -EIO;
+		} else if (status & 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) {
+		if (dma_inprogress(host))
+			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 || host->mrq->sbc) {
+			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)
+{
+	void __iomem *base = host->base;
+	bool sbc, busy_resp;
+
+	if (!cmd)
+		return;
+
+	sbc = (cmd == host->mrq->sbc);
+	busy_resp = host->variant->busy_detect && (cmd->flags & MMC_RSP_BUSY);
+
+	if (!((status|host->busy_status) & (MCI_CMDCRCFAIL|MCI_CMDTIMEOUT|
+		MCI_CMDSENT|MCI_CMDRESPEND)))
+		return;
+
+	/* Check if we need to wait for busy completion. */
+	if (host->busy_status && (status & MCI_ST_CARDBUSY))
+		return;
+
+	/* Enable busy completion if needed and supported. */
+	if (!host->busy_status && busy_resp &&
+		!(status & (MCI_CMDCRCFAIL|MCI_CMDTIMEOUT)) &&
+		(readl(base + MMCISTATUS) & MCI_ST_CARDBUSY)) {
+		writel(readl(base + MMCIMASK0) | MCI_ST_BUSYEND,
+			base + MMCIMASK0);
+		host->busy_status = status & (MCI_CMDSENT|MCI_CMDRESPEND);
+		return;
+	}
+
+	/* At busy completion, mask the IRQ and complete the request. */
+	if (host->busy_status) {
+		writel(readl(base + MMCIMASK0) & ~MCI_ST_BUSYEND,
+			base + MMCIMASK0);
+		host->busy_status = 0;
+	}
+
+	host->cmd = NULL;
+
+	if (status & MCI_CMDTIMEOUT) {
+		cmd->error = -ETIMEDOUT;
+	} else if (status & MCI_CMDCRCFAIL && cmd->flags & MMC_RSP_CRC) {
+		cmd->error = -EILSEQ;
+	} 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 */
+			if (dma_inprogress(host)) {
+				mmci_dma_data_error(host);
+				mmci_dma_unmap(host, host->data);
+			}
+			mmci_stop_data(host);
+		}
+		mmci_request_end(host, host->mrq);
+	} else if (sbc) {
+		mmci_start_command(host, host->mrq->cmd, 0);
+	} else if (!(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;
+	unsigned long flags;
+	u32 status;
+
+	status = readl(base + MMCISTATUS);
+
+	dev_dbg(mmc_dev(host->mmc), "irq1 (pio) %08x\n", status);
+
+	local_irq_save(flags);
+
+	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);
+
+	local_irq_restore(flags);
+
+	/*
+	 * 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;
+	int ret = 0;
+
+	spin_lock(&host->lock);
+
+	do {
+		status = readl(host->base + MMCISTATUS);
+
+		if (host->singleirq) {
+			if (status & readl(host->base + MMCIMASK1))
+				mmci_pio_irq(irq, dev_id);
+
+			status &= ~MCI_IRQ1MASK;
+		}
+
+		/*
+		 * We intentionally clear the MCI_ST_CARDBUSY IRQ here (if it's
+		 * enabled) since the HW seems to be triggering the IRQ on both
+		 * edges while monitoring DAT0 for busy completion.
+		 */
+		status &= readl(host->base + MMCIMASK0);
+		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);
+		}
+
+		/* Don't poll for busy completion in irq context. */
+		if (host->busy_status)
+			status &= ~MCI_ST_CARDBUSY;
+
+		ret = 1;
+	} while (status);
+
+	spin_unlock(&host->lock);
+
+	return IRQ_RETVAL(ret);
+}
+
+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;
+	}
+
+	pm_runtime_get_sync(mmc_dev(mmc));
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	host->mrq = mrq;
+
+	if (mrq->data)
+		mmci_get_next_data(host, mrq->data);
+
+	if (mrq->data && 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_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;
+
+	pm_runtime_get_sync(mmc_dev(mmc));
+
+	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 (ios->bus_mode == MMC_BUSMODE_OPENDRAIN) {
+		if (host->hw_designer != AMBA_VENDOR_ST)
+			pwr |= MCI_ROD;
+		else {
+			/*
+			 * The ST Micro variant use the ROD bit for something
+			 * else and only has OD (Open Drain).
+			 */
+			pwr |= MCI_OD;
+		}
+	}
+
+	/*
+	 * 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);
+
+	mmci_set_clkreg(host, ios->clock);
+	mmci_write_pwrreg(host, pwr);
+	mmci_reg_delay(host);
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	pm_runtime_mark_last_busy(mmc_dev(mmc));
+	pm_runtime_put_autosuspend(mmc_dev(mmc));
+}
+
+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)
+{
+	int ret = 0;
+
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+
+		pm_runtime_get_sync(mmc_dev(mmc));
+
+		switch (ios->signal_voltage) {
+		case MMC_SIGNAL_VOLTAGE_330:
+			ret = regulator_set_voltage(mmc->supply.vqmmc,
+						2700000, 3600000);
+			break;
+		case MMC_SIGNAL_VOLTAGE_180:
+			ret = regulator_set_voltage(mmc->supply.vqmmc,
+						1700000, 1950000);
+			break;
+		case MMC_SIGNAL_VOLTAGE_120:
+			ret = regulator_set_voltage(mmc->supply.vqmmc,
+						1100000, 1300000);
+			break;
+		}
+
+		if (ret)
+			dev_warn(mmc_dev(mmc), "Voltage switch failed\n");
+
+		pm_runtime_mark_last_busy(mmc_dev(mmc));
+		pm_runtime_put_autosuspend(mmc_dev(mmc));
+	}
+
+	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, "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->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;
+	}
+
+	/*
+	 * 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->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);
+
+	/* Get regulators and the supported OCR mask */
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret == -EPROBE_DEFER)
+		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");
+
+	/* DT takes precedence over platform data. */
+	if (!np) {
+		if (!plat->cd_invert)
+			mmc->caps2 |= MMC_CAP2_CD_ACTIVE_HIGH;
+		mmc->caps2 |= MMC_CAP2_RO_ACTIVE_HIGH;
+	}
+
+	/* We support these capabilities. */
+	mmc->caps |= MMC_CAP_CMD23;
+
+	if (variant->busy_detect) {
+		mmci_ops.card_busy = mmci_card_busy;
+		mmci_write_datactrlreg(host, MCI_ST_DPSM_BUSYMODE);
+		mmc->caps |= MMC_CAP_WAIT_WHILE_BUSY;
+		mmc->max_busy_timeout = 0;
+	}
+
+	mmc->ops = &mmci_ops;
+
+	/* 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 << 11;
+
+	/*
+	 * Limit the number of blocks transferred so that we don't overflow
+	 * the maximum request size.
+	 */
+	mmc->max_blk_count = mmc->max_req_size >> 11;
+
+	spin_lock_init(&host->lock);
+
+	writel(0, host->base + MMCIMASK0);
+	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 and proceed to try platform data
+	 */
+	if (!np) {
+		ret = mmc_gpiod_request_cd(mmc, "cd", 0, false, 0, NULL);
+		if (ret < 0) {
+			if (ret == -EPROBE_DEFER)
+				goto clk_disable;
+			else if (gpio_is_valid(plat->gpio_cd)) {
+				ret = mmc_gpio_request_cd(mmc, plat->gpio_cd, 0);
+				if (ret)
+					goto clk_disable;
+			}
+		}
+
+		ret = mmc_gpiod_request_ro(mmc, "wp", 0, false, 0, NULL);
+		if (ret < 0) {
+			if (ret == -EPROBE_DEFER)
+				goto clk_disable;
+			else if (gpio_is_valid(plat->gpio_wp)) {
+				ret = mmc_gpio_request_ro(mmc, plat->gpio_wp);
+				if (ret)
+					goto clk_disable;
+			}
+		}
+	}
+
+	ret = devm_request_irq(&dev->dev, dev->irq[0], mmci_irq, 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, 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);
+
+	mmc_add_host(mmc);
+
+	pm_runtime_put(&dev->dev);
+	return 0;
+
+ clk_disable:
+	clk_disable_unprepare(host->clk);
+ host_free:
+	mmc_free_host(mmc);
+	return ret;
+}
+
+static int mmci_remove(struct amba_device *dev)
+{
+	struct mmc_host *mmc = amba_get_drvdata(dev);
+
+	if (mmc) {
+		struct mmci_host *host = mmc_priv(mmc);
+
+		/*
+		 * 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);
+		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);
+	}
+
+	return 0;
+}
+
+#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->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_pm_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 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_u300,
+	},
+	{
+		.id     = 0x00480180,
+		.mask   = 0xf0ffffff,
+		.data	= &variant_ux500,
+	},
+	{
+		.id     = 0x10480180,
+		.mask   = 0xf0ffffff,
+		.data	= &variant_ux500v2,
+	},
+	/* 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		= 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..a1f5e4f49
--- /dev/null
+++ b/drivers/mmc/host/mmci.h
@@ -0,0 +1,247 @@
+/*
+ *  linux/drivers/mmc/host/mmci.h - ARM PrimeCell MMCI PL180/1 driver
+ *
+ *  Copyright (C) 2003 Deep Blue Solutions, Ltd, All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#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)
+
+#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))
+
+#define MMCIARGUMENT		0x008
+#define MMCICOMMAND		0x00c
+#define MCI_CPSM_RESPONSE	(1 << 6)
+#define MCI_CPSM_LONGRSP	(1 << 7)
+#define MCI_CPSM_INTERRUPT	(1 << 8)
+#define MCI_CPSM_PENDING	(1 << 9)
+#define MCI_CPSM_ENABLE		(1 << 10)
+/* Argument flag extenstions in the ST Micro versions */
+#define MCI_ST_SDIO_SUSP	(1 << 11)
+#define MCI_ST_ENCMD_COMPL	(1 << 12)
+#define MCI_ST_NIEN		(1 << 13)
+#define MCI_ST_CE_ATACMD	(1 << 14)
+
+/* Modified on Qualcomm Integrations */
+#define MCI_QCOM_CSPM_DATCMD		BIT(12)
+#define MCI_QCOM_CSPM_MCIABORT		BIT(13)
+#define MCI_QCOM_CSPM_CCSENABLE		BIT(14)
+#define MCI_QCOM_CSPM_CCSDISABLE	BIT(15)
+#define MCI_QCOM_CSPM_AUTO_CMD19	BIT(16)
+#define MCI_QCOM_CSPM_AUTO_CMD21	BIT(21)
+
+#define MMCIRESPCMD		0x010
+#define MMCIRESPONSE0		0x014
+#define MMCIRESPONSE1		0x018
+#define MMCIRESPONSE2		0x01c
+#define MMCIRESPONSE3		0x020
+#define MMCIDATATIMER		0x024
+#define MMCIDATALENGTH		0x028
+#define MMCIDATACTRL		0x02c
+#define MCI_DPSM_ENABLE		(1 << 0)
+#define MCI_DPSM_DIRECTION	(1 << 1)
+#define MCI_DPSM_MODE		(1 << 2)
+#define MCI_DPSM_DMAENABLE	(1 << 3)
+#define MCI_DPSM_BLOCKSIZE	(1 << 4)
+/* Control register extensions in the ST Micro U300 and Ux500 versions */
+#define MCI_ST_DPSM_RWSTART	(1 << 8)
+#define MCI_ST_DPSM_RWSTOP	(1 << 9)
+#define MCI_ST_DPSM_RWMOD	(1 << 10)
+#define MCI_ST_DPSM_SDIOEN	(1 << 11)
+/* Control register extensions in the ST Micro Ux500 versions */
+#define MCI_ST_DPSM_DMAREQCTL	(1 << 12)
+#define MCI_ST_DPSM_DBOOTMODEEN	(1 << 13)
+#define MCI_ST_DPSM_BUSYMODE	(1 << 14)
+#define MCI_ST_DPSM_DDRMODE	(1 << 15)
+
+#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)
+
+#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)
+
+#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_BUSYEND		(1 << 24)
+
+#define MMCIMASK1		0x040
+#define MMCIFIFOCNT		0x048
+#define MMCIFIFO		0x080 /* to 0x0bc */
+
+#define MCI_IRQENABLE	\
+	(MCI_CMDCRCFAILMASK|MCI_DATACRCFAILMASK|MCI_CMDTIMEOUTMASK|	\
+	MCI_DATATIMEOUTMASK|MCI_TXUNDERRUNMASK|MCI_RXOVERRUNMASK|	\
+	MCI_CMDRESPENDMASK|MCI_CMDSENTMASK|MCI_STARTBITERRMASK)
+
+/* These interrupts are directed to IRQ1 when two IRQ lines are available */
+#define MCI_IRQ1MASK \
+	(MCI_RXFIFOHALFFULLMASK | MCI_RXDATAAVLBLMASK | \
+	 MCI_TXFIFOHALFEMPTYMASK)
+
+#define NR_SG		128
+
+struct clk;
+struct variant_data;
+struct dma_chan;
+
+struct mmci_host_next {
+	struct dma_async_tx_descriptor	*dma_desc;
+	struct dma_chan			*dma_chan;
+	s32				cookie;
+};
+
+struct mmci_host {
+	phys_addr_t		phybase;
+	void __iomem		*base;
+	struct mmc_request	*mrq;
+	struct mmc_command	*cmd;
+	struct mmc_data		*data;
+	struct mmc_host		*mmc;
+	struct clk		*clk;
+	bool			singleirq;
+
+	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			datactrl_reg;
+	u32			busy_status;
+	bool			vqmmc_enabled;
+	struct mmci_platform_data *plat;
+	struct variant_data	*variant;
+
+	u8			hw_designer;
+	u8			hw_revision:4;
+
+	struct timer_list	timer;
+	unsigned int		oldstat;
+
+	/* pio stuff */
+	struct sg_mapping_iter	sg_miter;
+	unsigned int		size;
+	int (*get_rx_fifocnt)(struct mmci_host *h, u32 status, int remain);
+
+#ifdef CONFIG_DMA_ENGINE
+	/* DMA stuff */
+	struct dma_chan		*dma_current;
+	struct dma_chan		*dma_rx_channel;
+	struct dma_chan		*dma_tx_channel;
+	struct dma_async_tx_descriptor	*dma_desc_current;
+	struct mmci_host_next	next_data;
+
+#define dma_inprogress(host)	((host)->dma_current)
+#else
+#define dma_inprogress(host)	(0)
+#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..2b7fc3764
--- /dev/null
+++ b/drivers/mmc/host/mmci_qcom_dml.c
@@ -0,0 +1,177 @@
+/*
+ *
+ * Copyright (c) 2011, The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+#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
+
+void dml_start_xfer(struct mmci_host *host, struct mmc_data *data)
+{
+	u32 config;
+	void __iomem *base = host->base + DML_OFFSET;
+
+	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();
+}
+
+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 */
+int dml_hw_init(struct mmci_host *host, struct device_node *np)
+{
+	u32 config;
+	void __iomem *base;
+	int consumer_id, producer_id;
+
+	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)
+		return -ENODEV;
+
+	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 intialization is finished */
+	mb();
+
+	return 0;
+}
diff --git a/drivers/mmc/host/mmci_qcom_dml.h b/drivers/mmc/host/mmci_qcom_dml.h
new file mode 100644
index 000000000..6e405d09d
--- /dev/null
+++ b/drivers/mmc/host/mmci_qcom_dml.h
@@ -0,0 +1,31 @@
+/*
+ *
+ * Copyright (c) 2011, The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+#ifndef __MMC_QCOM_DML_H__
+#define __MMC_QCOM_DML_H__
+
+#ifdef CONFIG_MMC_QCOM_DML
+int dml_hw_init(struct mmci_host *host, struct device_node *np);
+void dml_start_xfer(struct mmci_host *host, struct mmc_data *data);
+#else
+static inline int dml_hw_init(struct mmci_host *host, struct device_node *np)
+{
+	return -ENOSYS;
+}
+static inline void dml_start_xfer(struct mmci_host *host, struct mmc_data *data)
+{
+}
+#endif /* CONFIG_MMC_QCOM_DML */
+
+#endif /* __MMC_QCOM_DML_H__ */
diff --git a/drivers/mmc/host/moxart-mmc.c b/drivers/mmc/host/moxart-mmc.c
new file mode 100644
index 000000000..006f18624
--- /dev/null
+++ b/drivers/mmc/host/moxart-mmc.c
@@ -0,0 +1,728 @@
+/*
+ * 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_8		BIT(2)
+#define BUS_WIDTH_4		BIT(1)
+#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_data, dir_slave;
+	unsigned 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_data = DMA_TO_DEVICE;
+		dir_slave = DMA_MEM_TO_DEV;
+	} else {
+		dma_chan = host->dma_chan_rx;
+		dir_data = DMA_FROM_DEVICE;
+		dir_slave = DMA_DEV_TO_MEM;
+	}
+
+	len = dma_map_sg(dma_chan->device->dev, data->sg,
+			 data->sg_len, 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,
+		     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;) {
+				/* SCR data must be read in big endian. */
+				if (data->mrq->cmd->opcode == SD_APP_SEND_SCR)
+					*sgp = ioread32be(host->base +
+							  REG_DATA_WINDOW);
+				else
+					*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);
+	unsigned long pio_time, 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;
+	case MMC_BUS_WIDTH_8:
+		writel(BUS_WIDTH_8, 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 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;
+	}
+
+	ret = of_address_to_resource(node, 0, &res_mmc);
+	if (ret) {
+		dev_err(dev, "of_address_to_resource failed\n");
+		goto out;
+	}
+
+	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;
+	}
+
+	clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(clk)) {
+		ret = PTR_ERR(clk);
+		goto out;
+	}
+
+	reg_mmc = devm_ioremap_resource(dev, &res_mmc);
+	if (IS_ERR(reg_mmc)) {
+		ret = PTR_ERR(reg_mmc);
+		goto out;
+	}
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto out;
+
+	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_slave_channel_reason(dev, "tx");
+	host->dma_chan_rx = dma_request_slave_channel_reason(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;
+		}
+		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;
+
+		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);
+	}
+
+	switch ((readl(host->base + REG_BUS_WIDTH) >> 3) & 3) {
+	case 1:
+		mmc->caps |= MMC_CAP_4_BIT_DATA;
+		break;
+	case 2:
+		mmc->caps |= MMC_CAP_4_BIT_DATA | MMC_CAP_8_BIT_DATA;
+		break;
+	default:
+		break;
+	}
+
+	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);
+	mmc_add_host(mmc);
+
+	dev_dbg(dev, "IRQ=%d, FIFO is %d bytes\n", irq, host->fifo_width);
+
+	return 0;
+
+out:
+	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 (mmc) {
+		if (!IS_ERR(host->dma_chan_tx))
+			dma_release_channel(host->dma_chan_tx);
+		if (!IS_ERR(host->dma_chan_rx))
+			dma_release_channel(host->dma_chan_rx);
+		mmc_remove_host(mmc);
+		mmc_free_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);
+	}
+	return 0;
+}
+
+static const struct of_device_id moxart_mmc_match[] = {
+	{ .compatible = "moxa,moxart-mmc" },
+	{ .compatible = "faraday,ftsdc010" },
+	{ }
+};
+
+static struct platform_driver moxart_mmc_driver = {
+	.probe      = moxart_probe,
+	.remove     = moxart_remove,
+	.driver     = {
+		.name		= "mmc-moxart",
+		.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/mvsdio.c b/drivers/mmc/host/mvsdio.c
new file mode 100644
index 000000000..a448498e3
--- /dev/null
+++ b/drivers/mmc/host/mvsdio.c
@@ -0,0 +1,876 @@
+/*
+ * Marvell MMC/SD/SDIO driver
+ *
+ * Authors: Maen Suleiman, Nicolas Pitre
+ * Copyright (C) 2008-2009 Marvell Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#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/gpio.h>
+#include <linux/of_gpio.h>
+#include <linux/of_irq.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/slot-gpio.h>
+
+#include <asm/sizes.h>
+#include <asm/unaligned.h>
+#include <linux/platform_data/mmc-mvsdio.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;
+		int dma_dir = (data->flags & MMC_DATA_READ) ?
+			DMA_FROM_DEVICE : DMA_TO_DEVICE;
+		host->sg_frags = dma_map_sg(mmc_dev(host->mmc), data->sg,
+					    data->sg_len, dma_dir);
+		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;
+
+	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);
+
+	mod_timer(&host->timer, jiffies + 5 * HZ);
+
+	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,
+			     (data->flags & MMC_DATA_READ) ?
+				DMA_FROM_DEVICE : DMA_TO_DEVICE);
+	}
+
+	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(unsigned long data)
+{
+	struct mvsd_host *host = (struct mvsd_host *)data;
+	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;
+	struct resource *r;
+	int ret, irq;
+
+	r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	irq = platform_get_irq(pdev, 0);
+	if (!r || irq < 0)
+		return -ENXIO;
+
+	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))
+		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;
+
+	if (np) {
+		if (IS_ERR(host->clk)) {
+			dev_err(&pdev->dev, "DT platforms must have a clock associated\n");
+			ret = -EINVAL;
+			goto out;
+		}
+
+		host->base_clock = clk_get_rate(host->clk) / 2;
+		ret = mmc_of_parse(mmc);
+		if (ret < 0)
+			goto out;
+	} else {
+		const struct mvsdio_platform_data *mvsd_data;
+
+		mvsd_data = pdev->dev.platform_data;
+		if (!mvsd_data) {
+			ret = -ENXIO;
+			goto out;
+		}
+		mmc->caps = MMC_CAP_4_BIT_DATA | MMC_CAP_SDIO_IRQ |
+			    MMC_CAP_SD_HIGHSPEED | MMC_CAP_MMC_HIGHSPEED;
+		host->base_clock = mvsd_data->clock / 2;
+		/* GPIO 0 regarded as invalid for backward compatibility */
+		if (mvsd_data->gpio_card_detect &&
+		    gpio_is_valid(mvsd_data->gpio_card_detect)) {
+			ret = mmc_gpio_request_cd(mmc,
+						  mvsd_data->gpio_card_detect,
+						  0);
+			if (ret)
+				goto out;
+		} else {
+			mmc->caps |= MMC_CAP_NEEDS_POLL;
+		}
+
+		if (mvsd_data->gpio_write_protect &&
+		    gpio_is_valid(mvsd_data->gpio_write_protect))
+			mmc_gpio_request_ro(mmc, mvsd_data->gpio_write_protect);
+	}
+
+	if (maxfreq)
+		mmc->f_max = maxfreq;
+
+	spin_lock_init(&host->lock);
+
+	host->base = devm_ioremap_resource(&pdev->dev, r);
+	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;
+	}
+
+	setup_timer(&host->timer, mvsd_timeout_timer, (unsigned long)host);
+	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,
+		.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..7d9727b9f
--- /dev/null
+++ b/drivers/mmc/host/mvsdio.h
@@ -0,0 +1,190 @@
+/*
+ *  Copyright (C) 2008 Marvell Semiconductors, All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#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..317d709f7
--- /dev/null
+++ b/drivers/mmc/host/mxcmmc.c
@@ -0,0 +1,1246 @@
+/*
+ *  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
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ */
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/ioport.h>
+#include <linux/platform_device.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/gpio.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/of_gpio.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;
+
+	if (data->flags & MMC_DATA_STREAM)
+		nob = 0xffff;
+
+	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);
+	}
+
+	stat = mxcmci_poll_status(host, STATUS_BUF_WRITE_RDY);
+	if (stat)
+		return stat;
+
+	return 0;
+}
+
+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);
+
+	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;
+	unsigned long flags;
+	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_irqsave(&host->lock, flags);
+	sdio_irq = (stat & STATUS_SDIO_INT_ACTIVE) && host->use_sdio;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	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(unsigned long data)
+{
+	struct mmc_host *mmc = (struct mmc_host *)data;
+	struct mxcmci_host *host = mmc_priv(mmc);
+	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 -EINVAL;
+
+	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->caps & MMC_CAP_NONREMOVABLE)
+			&& !of_property_read_bool(pdev->dev.of_node, "cd-gpios"))
+		dat3_card_detect = true;
+
+	ret = mmc_regulator_get_supply(mmc);
+	if (ret == -EPROBE_DEFER)
+		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;
+	}
+
+	clk_prepare_enable(host->clk_per);
+	clk_prepare_enable(host->clk_ipg);
+
+	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_slave_channel(&pdev->dev, "rx-tx");
+	} 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;
+	}
+
+	init_timer(&host->watchdog);
+	host->watchdog.function = &mxcmci_watchdog;
+	host->watchdog.data = (unsigned long)mmc;
+
+	mmc_add_host(mmc);
+
+	return 0;
+
+out_free_dma:
+	if (host->dma)
+		dma_release_channel(host->dma);
+
+out_clk_put:
+	clk_disable_unprepare(host->clk_per);
+	clk_disable_unprepare(host->clk_ipg);
+
+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;
+}
+
+static int __maybe_unused 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 __maybe_unused mxcmci_resume(struct device *dev)
+{
+	struct mmc_host *mmc = dev_get_drvdata(dev);
+	struct mxcmci_host *host = mmc_priv(mmc);
+
+	clk_prepare_enable(host->clk_per);
+	clk_prepare_enable(host->clk_ipg);
+	return 0;
+}
+
+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,
+		.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..a82411a2c
--- /dev/null
+++ b/drivers/mmc/host/mxs-mmc.c
@@ -0,0 +1,745 @@
+/*
+ * 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.
+ *
+ * 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; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/ioport.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/of_gpio.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/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/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 (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 (mrq->stop) {
+			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, DMA_CTRL_ACK);
+	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 (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, DMA_CTRL_ACK);
+	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 == MMC_STOP_TRANSMISSION) ||
+	    (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 | DMA_CTRL_ACK);
+	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;
+	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 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 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;
+	struct resource *iores;
+	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;
+	iores = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	ssp->base = devm_ioremap_resource(&pdev->dev, iores);
+	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;
+		}
+	}
+
+	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_slave_channel(&pdev->dev, "rx-tx");
+	if (!ssp->dmach) {
+		dev_err(mmc_dev(host->mmc),
+			"%s: failed to request dma\n", __func__);
+		ret = -ENODEV;
+		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;
+
+	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_clk_disable;
+
+	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);
+
+	ret = devm_request_irq(&pdev->dev, irq_err, mxs_mmc_irq_handler, 0,
+			       dev_name(&pdev->dev), host);
+	if (ret)
+		goto out_free_dma;
+
+	spin_lock_init(&host->lock);
+
+	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,
+		.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..6e218fb1a
--- /dev/null
+++ b/drivers/mmc/host/of_mmc_spi.c
@@ -0,0 +1,159 @@
+/*
+ * OpenFirmware bindings for the MMC-over-SPI driver
+ *
+ * Copyright (c) MontaVista Software, Inc. 2008.
+ *
+ * Author: Anton Vorontsov <avorontsov@ru.mvista.com>
+ *
+ * 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;  either version 2 of the  License, or (at your
+ * option) any later version.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/device.h>
+#include <linux/slab.h>
+#include <linux/irq.h>
+#include <linux/gpio.h>
+#include <linux/of.h>
+#include <linux/of_gpio.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");
+
+enum {
+	CD_GPIO = 0,
+	WP_GPIO,
+	NUM_GPIOS,
+};
+
+struct of_mmc_spi {
+	int gpios[NUM_GPIOS];
+	bool alow_gpios[NUM_GPIOS];
+	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, 0,
+				    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;
+	const u32 *voltage_ranges;
+	int num_ranges;
+	int i;
+	int ret = -EINVAL;
+
+	if (dev->platform_data || !np)
+		return dev->platform_data;
+
+	oms = kzalloc(sizeof(*oms), GFP_KERNEL);
+	if (!oms)
+		return NULL;
+
+	voltage_ranges = of_get_property(np, "voltage-ranges", &num_ranges);
+	num_ranges = num_ranges / sizeof(*voltage_ranges) / 2;
+	if (!voltage_ranges || !num_ranges) {
+		dev_err(dev, "OF: voltage-ranges unspecified\n");
+		goto err_ocr;
+	}
+
+	for (i = 0; i < num_ranges; i++) {
+		const int j = i * 2;
+		u32 mask;
+
+		mask = mmc_vddrange_to_ocrmask(be32_to_cpu(voltage_ranges[j]),
+					       be32_to_cpu(voltage_ranges[j + 1]));
+		if (!mask) {
+			ret = -EINVAL;
+			dev_err(dev, "OF: voltage-range #%d is invalid\n", i);
+			goto err_ocr;
+		}
+		oms->pdata.ocr_mask |= mask;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(oms->gpios); i++) {
+		enum of_gpio_flags gpio_flags;
+
+		oms->gpios[i] = of_get_gpio_flags(np, i, &gpio_flags);
+		if (!gpio_is_valid(oms->gpios[i]))
+			continue;
+
+		if (gpio_flags & OF_GPIO_ACTIVE_LOW)
+			oms->alow_gpios[i] = true;
+	}
+
+	if (gpio_is_valid(oms->gpios[CD_GPIO])) {
+		oms->pdata.cd_gpio = oms->gpios[CD_GPIO];
+		oms->pdata.flags |= MMC_SPI_USE_CD_GPIO;
+		if (!oms->alow_gpios[CD_GPIO])
+			oms->pdata.caps2 |= MMC_CAP2_CD_ACTIVE_HIGH;
+	}
+	if (gpio_is_valid(oms->gpios[WP_GPIO])) {
+		oms->pdata.ro_gpio = oms->gpios[WP_GPIO];
+		oms->pdata.flags |= MMC_SPI_USE_RO_GPIO;
+		if (!oms->alow_gpios[WP_GPIO])
+			oms->pdata.caps2 |= MMC_CAP2_RO_ACTIVE_HIGH;
+	}
+
+	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..68dd6c79c
--- /dev/null
+++ b/drivers/mmc/host/omap.c
@@ -0,0 +1,1505 @@
+/*
+ *  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
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#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/omap-dma.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;
+	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(unsigned long data)
+{
+	struct mmc_omap_host *host = (struct mmc_omap_host *) data;
+	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(unsigned long data)
+{
+	struct mmc_omap_host *host = (struct mmc_omap_host *) data;
+
+	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(unsigned long arg)
+{
+	struct mmc_omap_slot *slot = (struct mmc_omap_slot *) arg;
+	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 || !mmc_card_present(slot->mmc->card))
+		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, 0xff);
+}
+
+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;
+	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 (i = 0; i < sg_len; i++) {
+		if ((data->sg[i].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;
+
+			cfg.src_addr = host->phys_base + OMAP_MMC_REG(host, DATA);
+			cfg.dst_addr = host->phys_base + OMAP_MMC_REG(host, DATA);
+			cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES;
+			cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES;
+			cfg.src_maxburst = burst;
+			cfg.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;
+
+	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;
+	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);
+		goto exit;
+	case MMC_POWER_ON:
+		mmc_omap_fclk_enable(host, 1);
+		clk_enabled = 1;
+		dsor |= 1 << 11;
+		break;
+	}
+
+	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 (ios->power_mode == MMC_POWER_ON) {
+		/* 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->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_CAP_ERASE;
+
+	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) {
+		setup_timer(&slot->cover_timer, mmc_omap_cover_timer,
+			    (unsigned long)slot);
+		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;
+	dma_cap_mask_t mask;
+	unsigned sig = 0;
+	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 -ENXIO;
+
+	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);
+	setup_timer(&host->cmd_abort_timer, mmc_omap_cmd_timer,
+		    (unsigned long) host);
+
+	spin_lock_init(&host->clk_lock);
+	setup_timer(&host->clk_timer, mmc_omap_clk_timer, (unsigned long) host);
+
+	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;
+	}
+
+	dma_cap_zero(mask);
+	dma_cap_set(DMA_SLAVE, mask);
+
+	host->dma_tx_burst = -1;
+	host->dma_rx_burst = -1;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_DMA, "tx");
+	if (res)
+		sig = res->start;
+	host->dma_tx = dma_request_slave_channel_compat(mask,
+				omap_dma_filter_fn, &sig, &pdev->dev, "tx");
+	if (!host->dma_tx)
+		dev_warn(host->dev, "unable to obtain TX DMA engine channel %u\n",
+			sig);
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_DMA, "rx");
+	if (res)
+		sig = res->start;
+	host->dma_rx = dma_request_slave_channel_compat(mask,
+				omap_dma_filter_fn, &sig, &pdev->dev, "rx");
+	if (!host->dma_rx)
+		dev_warn(host->dev, "unable to obtain RX DMA engine channel %u\n",
+			sig);
+
+	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)
+		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", },
+	{ },
+};
+#endif
+
+static struct platform_driver mmc_omap_driver = {
+	.probe		= mmc_omap_probe,
+	.remove		= mmc_omap_remove,
+	.driver		= {
+		.name	= DRIVER_NAME,
+		.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..9df2b6801
--- /dev/null
+++ b/drivers/mmc/host/omap_hsmmc.c
@@ -0,0 +1,2344 @@
+/*
+ * 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_gpio.h>
+#include <linux/of_device.h>
+#include <linux/omap-dmaengine.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/gpio.h>
+#include <linux/regulator/consumer.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/pm_runtime.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"
+
+#define VDD_1V8			1800000		/* 180000 uV */
+#define VDD_3V0			3000000		/* 300000 uV */
+#define VDD_165_195		(ffs(MMC_VDD_165_195) - 1)
+
+/*
+ * 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;
+	/*
+	 * vcc == configured supply
+	 * vcc_aux == optional
+	 *   -	MMC1, supply for DAT4..DAT7
+	 *   -	MMC2/MMC2, external level shifter voltage supply, for
+	 *	chip (SDIO, eMMC, etc) or transceiver (MMC2 only)
+	 */
+	struct	regulator	*vcc;
+	struct	regulator	*vcc_aux;
+	struct	regulator	*pbias;
+	bool			pbias_enabled;
+	void	__iomem		*base;
+	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			protect_card;
+	int			reqs_blocked;
+	int			use_reg;
+	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 */
+#define HSMMC_WAKE_IRQ_ENABLED	(1 << 2)
+	struct omap_hsmmc_next	next_data;
+	struct	omap_hsmmc_platform_data	*pdata;
+
+	/* return MMC cover switch state, can be NULL if not supported.
+	 *
+	 * possible return values:
+	 *   0 - closed
+	 *   1 - open
+	 */
+	int (*get_cover_state)(struct device *dev);
+
+	int (*card_detect)(struct device *dev);
+};
+
+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_card_detect(struct device *dev)
+{
+	struct omap_hsmmc_host *host = dev_get_drvdata(dev);
+
+	return mmc_gpio_get_cd(host->mmc);
+}
+
+static int omap_hsmmc_get_cover_state(struct device *dev)
+{
+	struct omap_hsmmc_host *host = dev_get_drvdata(dev);
+
+	return mmc_gpio_get_cd(host->mmc);
+}
+
+#ifdef CONFIG_REGULATOR
+
+static int omap_hsmmc_set_power(struct device *dev, int power_on, int vdd)
+{
+	struct omap_hsmmc_host *host =
+		platform_get_drvdata(to_platform_device(dev));
+	int ret = 0;
+
+	/*
+	 * If we don't see a Vcc regulator, assume it's a fixed
+	 * voltage always-on regulator.
+	 */
+	if (!host->vcc)
+		return 0;
+
+	if (mmc_pdata(host)->before_set_reg)
+		mmc_pdata(host)->before_set_reg(dev, power_on, vdd);
+
+	if (host->pbias) {
+		if (host->pbias_enabled == 1) {
+			ret = regulator_disable(host->pbias);
+			if (!ret)
+				host->pbias_enabled = 0;
+		}
+		regulator_set_voltage(host->pbias, VDD_3V0, VDD_3V0);
+	}
+
+	/*
+	 * 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) {
+		if (host->vcc)
+			ret = mmc_regulator_set_ocr(host->mmc, host->vcc, vdd);
+		/* Enable interface voltage rail, if needed */
+		if (ret == 0 && host->vcc_aux) {
+			ret = regulator_enable(host->vcc_aux);
+			if (ret < 0 && host->vcc)
+				ret = mmc_regulator_set_ocr(host->mmc,
+							host->vcc, 0);
+		}
+	} else {
+		/* Shut down the rail */
+		if (host->vcc_aux)
+			ret = regulator_disable(host->vcc_aux);
+		if (host->vcc) {
+			/* Then proceed to shut down the local regulator */
+			ret = mmc_regulator_set_ocr(host->mmc,
+						host->vcc, 0);
+		}
+	}
+
+	if (host->pbias) {
+		if (vdd <= VDD_165_195)
+			ret = regulator_set_voltage(host->pbias, VDD_1V8,
+								VDD_1V8);
+		else
+			ret = regulator_set_voltage(host->pbias, VDD_3V0,
+								VDD_3V0);
+		if (ret < 0)
+			goto error_set_power;
+
+		if (host->pbias_enabled == 0) {
+			ret = regulator_enable(host->pbias);
+			if (!ret)
+				host->pbias_enabled = 1;
+		}
+	}
+
+	if (mmc_pdata(host)->after_set_reg)
+		mmc_pdata(host)->after_set_reg(dev, power_on, vdd);
+
+error_set_power:
+	return ret;
+}
+
+static int omap_hsmmc_reg_get(struct omap_hsmmc_host *host)
+{
+	struct regulator *reg;
+	int ocr_value = 0;
+
+	reg = devm_regulator_get(host->dev, "vmmc");
+	if (IS_ERR(reg)) {
+		dev_err(host->dev, "unable to get vmmc regulator %ld\n",
+			PTR_ERR(reg));
+		return PTR_ERR(reg);
+	} else {
+		host->vcc = reg;
+		ocr_value = mmc_regulator_get_ocrmask(reg);
+		if (!mmc_pdata(host)->ocr_mask) {
+			mmc_pdata(host)->ocr_mask = ocr_value;
+		} else {
+			if (!(mmc_pdata(host)->ocr_mask & ocr_value)) {
+				dev_err(host->dev, "ocrmask %x is not supported\n",
+					mmc_pdata(host)->ocr_mask);
+				mmc_pdata(host)->ocr_mask = 0;
+				return -EINVAL;
+			}
+		}
+	}
+	mmc_pdata(host)->set_power = omap_hsmmc_set_power;
+
+	/* Allow an aux regulator */
+	reg = devm_regulator_get_optional(host->dev, "vmmc_aux");
+	host->vcc_aux = IS_ERR(reg) ? NULL : reg;
+
+	reg = devm_regulator_get_optional(host->dev, "pbias");
+	host->pbias = IS_ERR(reg) ? NULL : reg;
+
+	/* For eMMC do not power off when not in sleep state */
+	if (mmc_pdata(host)->no_regulator_off_init)
+		return 0;
+	/*
+	 * To disable boot_on regulator, enable regulator
+	 * to increase usecount and then disable it.
+	 */
+	if ((host->vcc && regulator_is_enabled(host->vcc) > 0) ||
+	    (host->vcc_aux && regulator_is_enabled(host->vcc_aux))) {
+		int vdd = ffs(mmc_pdata(host)->ocr_mask) - 1;
+
+		mmc_pdata(host)->set_power(host->dev, 1, vdd);
+		mmc_pdata(host)->set_power(host->dev, 0, 0);
+	}
+
+	return 0;
+}
+
+static void omap_hsmmc_reg_put(struct omap_hsmmc_host *host)
+{
+	mmc_pdata(host)->set_power = NULL;
+}
+
+static inline int omap_hsmmc_have_reg(void)
+{
+	return 1;
+}
+
+#else
+
+static inline int omap_hsmmc_reg_get(struct omap_hsmmc_host *host)
+{
+	return -EINVAL;
+}
+
+static inline void omap_hsmmc_reg_put(struct omap_hsmmc_host *host)
+{
+}
+
+static inline int omap_hsmmc_have_reg(void)
+{
+	return 0;
+}
+
+#endif
+
+static irqreturn_t omap_hsmmc_cover_irq(int irq, void *dev_id);
+
+static int omap_hsmmc_gpio_init(struct mmc_host *mmc,
+				struct omap_hsmmc_host *host,
+				struct omap_hsmmc_platform_data *pdata)
+{
+	int ret;
+
+	if (gpio_is_valid(pdata->gpio_cod)) {
+		ret = mmc_gpio_request_cd(mmc, pdata->gpio_cod, 0);
+		if (ret)
+			return ret;
+
+		host->get_cover_state = omap_hsmmc_get_cover_state;
+		mmc_gpio_set_cd_isr(mmc, omap_hsmmc_cover_irq);
+	} else if (gpio_is_valid(pdata->gpio_cd)) {
+		ret = mmc_gpio_request_cd(mmc, pdata->gpio_cd, 0);
+		if (ret)
+			return ret;
+
+		host->card_detect = omap_hsmmc_card_detect;
+	}
+
+	if (gpio_is_valid(pdata->gpio_wp)) {
+		ret = mmc_gpio_request_ro(mmc, pdata->gpio_wp);
+		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;
+
+	if (host->protect_card)
+		return;
+
+	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 inline
+int omap_hsmmc_cover_is_closed(struct omap_hsmmc_host *host)
+{
+	int r = 1;
+
+	if (host->get_cover_state)
+		r = host->get_cover_state(host->dev);
+	return r;
+}
+
+static ssize_t
+omap_hsmmc_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 omap_hsmmc_host *host = mmc_priv(mmc);
+
+	return sprintf(buf, "%s\n",
+			omap_hsmmc_cover_is_closed(host) ? "closed" : "open");
+}
+
+static DEVICE_ATTR(cover_switch, S_IRUGO, omap_hsmmc_show_cover_switch, NULL);
+
+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 int
+omap_hsmmc_get_dma_dir(struct omap_hsmmc_host *host, struct mmc_data *data)
+{
+	if (data->flags & MMC_DATA_WRITE)
+		return DMA_TO_DEVICE;
+	else
+		return DMA_FROM_DEVICE;
+}
+
+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);
+	pm_runtime_mark_last_busy(host->dev);
+	pm_runtime_put_autosuspend(host->dev);
+}
+
+/*
+ * 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,
+			omap_hsmmc_get_dma_dir(host, 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))
+			end_cmd = 1;
+		if (status & (CTO_EN | DTO_EN))
+			hsmmc_command_incomplete(host, -ETIMEDOUT, end_cmd);
+		else if (status & (CCRC_EN | DCRC_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);
+		}
+		if (host->data || host->response_busy) {
+			end_trans = !end_cmd;
+			host->response_busy = 0;
+		}
+	}
+
+	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 irqreturn_t omap_hsmmc_wake_irq(int irq, void *dev_id)
+{
+	struct omap_hsmmc_host *host = dev_id;
+
+	/* cirq is level triggered, disable to avoid infinite loop */
+	spin_lock(&host->irq_lock);
+	if (host->flags & HSMMC_WAKE_IRQ_ENABLED) {
+		disable_irq_nosync(host->wake_irq);
+		host->flags &= ~HSMMC_WAKE_IRQ_ENABLED;
+	}
+	spin_unlock(&host->irq_lock);
+	pm_request_resume(host->dev); /* no use counter */
+
+	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 */
+	pm_runtime_put_sync(host->dev);
+	if (host->dbclk)
+		clk_disable_unprepare(host->dbclk);
+
+	/* Turn the power off */
+	ret = mmc_pdata(host)->set_power(host->dev, 0, 0);
+
+	/* Turn the power ON with given VDD 1.8 or 3.0v */
+	if (!ret)
+		ret = mmc_pdata(host)->set_power(host->dev, 1, vdd);
+	pm_runtime_get_sync(host->dev);
+	if (host->dbclk)
+		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;
+}
+
+/* Protect the card while the cover is open */
+static void omap_hsmmc_protect_card(struct omap_hsmmc_host *host)
+{
+	if (!host->get_cover_state)
+		return;
+
+	host->reqs_blocked = 0;
+	if (host->get_cover_state(host->dev)) {
+		if (host->protect_card) {
+			dev_info(host->dev, "%s: cover is closed, "
+					 "card is now accessible\n",
+					 mmc_hostname(host->mmc));
+			host->protect_card = 0;
+		}
+	} else {
+		if (!host->protect_card) {
+			dev_info(host->dev, "%s: cover is open, "
+					 "card is now inaccessible\n",
+					 mmc_hostname(host->mmc));
+			host->protect_card = 1;
+		}
+	}
+}
+
+/*
+ * irq handler when (cell-phone) cover is mounted/removed
+ */
+static irqreturn_t omap_hsmmc_cover_irq(int irq, void *dev_id)
+{
+	struct omap_hsmmc_host *host = dev_id;
+
+	sysfs_notify(&host->mmc->class_dev.kobj, NULL, "cover_switch");
+
+	omap_hsmmc_protect_card(host);
+	mmc_detect_change(host->mmc, (HZ * 200) / 1000);
+	return IRQ_HANDLED;
+}
+
+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,
+			     omap_hsmmc_get_dma_dir(host, 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);
+		pm_runtime_mark_last_busy(host->dev);
+		pm_runtime_put_autosuspend(host->dev);
+	}
+}
+
+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,
+				     omap_hsmmc_get_dma_dir(host, 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_slave_config cfg;
+	struct dma_async_tx_descriptor *tx;
+	int ret = 0, i;
+	struct mmc_data *data = req->data;
+	struct dma_chan *chan;
+
+	/* 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);
+
+	cfg.src_addr = host->mapbase + OMAP_HSMMC_DATA;
+	cfg.dst_addr = host->mapbase + OMAP_HSMMC_DATA;
+	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;
+
+	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 int timeout_ns,
+			     unsigned int timeout_clks)
+{
+	unsigned int timeout, 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);
+	timeout = timeout_ns / 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;
+	host->data = req->data;
+
+	if (req->data == NULL) {
+		OMAP_HSMMC_WRITE(host->base, BLK, 0);
+		/*
+		 * Set an arbitrary 100ms data timeout for commands with
+		 * busy signal.
+		 */
+		if (req->cmd->flags & MMC_RSP_BUSY)
+			set_data_timeout(host, 100000000U, 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,
+			     omap_hsmmc_get_dma_dir(host, data));
+		data->host_cookie = 0;
+	}
+}
+
+static void omap_hsmmc_pre_req(struct mmc_host *mmc, struct mmc_request *mrq,
+			       bool is_first_req)
+{
+	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);
+	pm_runtime_get_sync(host->dev);
+	if (host->protect_card) {
+		if (host->reqs_blocked < 3) {
+			/*
+			 * Ensure the controller is left in a consistent
+			 * state by resetting the command and data state
+			 * machines.
+			 */
+			omap_hsmmc_reset_controller_fsm(host, SRD);
+			omap_hsmmc_reset_controller_fsm(host, SRC);
+			host->reqs_blocked += 1;
+		}
+		req->cmd->error = -EBADF;
+		if (req->data)
+			req->data->error = -EBADF;
+		req->cmd->retries = 0;
+		mmc_request_done(mmc, req);
+		pm_runtime_mark_last_busy(host->dev);
+		pm_runtime_put_autosuspend(host->dev);
+		return;
+	} else 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);
+		pm_runtime_mark_last_busy(host->dev);
+		pm_runtime_put_autosuspend(host->dev);
+		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;
+
+	pm_runtime_get_sync(host->dev);
+
+	if (ios->power_mode != host->power_mode) {
+		switch (ios->power_mode) {
+		case MMC_POWER_OFF:
+			mmc_pdata(host)->set_power(host->dev, 0, 0);
+			break;
+		case MMC_POWER_UP:
+			mmc_pdata(host)->set_power(host->dev, 1, ios->vdd);
+			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);
+
+	pm_runtime_put_autosuspend(host->dev);
+}
+
+static int omap_hsmmc_get_cd(struct mmc_host *mmc)
+{
+	struct omap_hsmmc_host *host = mmc_priv(mmc);
+
+	if (!host->card_detect)
+		return -ENOSYS;
+	return host->card_detect(host->dev);
+}
+
+static void omap_hsmmc_init_card(struct mmc_host *mmc, struct mmc_card *card)
+{
+	struct omap_hsmmc_host *host = mmc_priv(mmc);
+
+	if (mmc_pdata(host)->init_card)
+		mmc_pdata(host)->init_card(card);
+}
+
+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)
+{
+	struct mmc_host *mmc = host->mmc;
+	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;
+
+	/* Prevent auto-enabling of IRQ */
+	irq_set_status_flags(host->wake_irq, IRQ_NOAUTOEN);
+	ret = devm_request_irq(host->dev, host->wake_irq, omap_hsmmc_wake_irq,
+			       IRQF_TRIGGER_LOW | IRQF_ONESHOT,
+			       mmc_hostname(mmc), host);
+	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 (!p) {
+			ret = -ENODEV;
+			goto err_free_irq;
+		}
+		if (IS_ERR(pinctrl_lookup_state(p, PINCTRL_STATE_DEFAULT))) {
+			dev_info(host->dev, "missing default pinctrl state\n");
+			devm_pinctrl_put(p);
+			ret = -EINVAL;
+			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:
+	devm_free_irq(host->dev, host->wake_irq, host);
+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 = omap_hsmmc_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 omap_hsmmc_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;
+}
+
+static int omap_hsmmc_regs_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, omap_hsmmc_regs_show, inode->i_private);
+}
+
+static const struct file_operations mmc_regs_fops = {
+	.open           = omap_hsmmc_regs_open,
+	.read           = seq_read,
+	.llseek         = seq_lseek,
+	.release        = single_release,
+};
+
+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;
+	struct device_node *np = dev->of_node;
+
+	pdata = devm_kzalloc(dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata)
+		return ERR_PTR(-ENOMEM); /* out of memory */
+
+	if (of_find_property(np, "ti,dual-volt", NULL))
+		pdata->controller_flags |= OMAP_HSMMC_SUPPORTS_DUAL_VOLT;
+
+	pdata->gpio_cd = -EINVAL;
+	pdata->gpio_cod = -EINVAL;
+	pdata->gpio_wp = -EINVAL;
+
+	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;
+	dma_cap_mask_t mask;
+	unsigned tx_req, rx_req;
+	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);
+	irq = platform_get_irq(pdev, 0);
+	if (res == NULL || irq < 0)
+		return -ENXIO;
+
+	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;
+
+	ret = omap_hsmmc_gpio_init(mmc, host, pdata);
+	if (ret)
+		goto err_gpio;
+
+	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;
+	}
+
+	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;
+	}
+
+	/* Since we do only SG emulation, we can have as many segs
+	 * as we want. */
+	mmc->max_segs = 1024;
+
+	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->max_seg_size = mmc->max_req_size;
+
+	mmc->caps |= MMC_CAP_MMC_HIGHSPEED | MMC_CAP_SD_HIGHSPEED |
+		     MMC_CAP_WAIT_WHILE_BUSY | MMC_CAP_ERASE;
+
+	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);
+
+	if (!pdev->dev.of_node) {
+		res = platform_get_resource_byname(pdev, IORESOURCE_DMA, "tx");
+		if (!res) {
+			dev_err(mmc_dev(host->mmc), "cannot get DMA TX channel\n");
+			ret = -ENXIO;
+			goto err_irq;
+		}
+		tx_req = res->start;
+
+		res = platform_get_resource_byname(pdev, IORESOURCE_DMA, "rx");
+		if (!res) {
+			dev_err(mmc_dev(host->mmc), "cannot get DMA RX channel\n");
+			ret = -ENXIO;
+			goto err_irq;
+		}
+		rx_req = res->start;
+	}
+
+	dma_cap_zero(mask);
+	dma_cap_set(DMA_SLAVE, mask);
+
+	host->rx_chan =
+		dma_request_slave_channel_compat(mask, omap_dma_filter_fn,
+						 &rx_req, &pdev->dev, "rx");
+
+	if (!host->rx_chan) {
+		dev_err(mmc_dev(host->mmc), "unable to obtain RX DMA engine channel %u\n", rx_req);
+		ret = -ENXIO;
+		goto err_irq;
+	}
+
+	host->tx_chan =
+		dma_request_slave_channel_compat(mask, omap_dma_filter_fn,
+						 &tx_req, &pdev->dev, "tx");
+
+	if (!host->tx_chan) {
+		dev_err(mmc_dev(host->mmc), "unable to obtain TX DMA engine channel %u\n", tx_req);
+		ret = -ENXIO;
+		goto err_irq;
+	}
+
+	/* 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;
+	}
+
+	if (omap_hsmmc_have_reg() && !mmc_pdata(host)->set_power) {
+		ret = omap_hsmmc_reg_get(host);
+		if (ret)
+			goto err_irq;
+		host->use_reg = 1;
+	}
+
+	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;
+
+	omap_hsmmc_protect_card(host);
+
+	mmc_add_host(mmc);
+
+	if (mmc_pdata(host)->name != NULL) {
+		ret = device_create_file(&mmc->class_dev, &dev_attr_slot_name);
+		if (ret < 0)
+			goto err_slot_name;
+	}
+	if (host->get_cover_state) {
+		ret = device_create_file(&mmc->class_dev,
+					 &dev_attr_cover_switch);
+		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);
+	if (host->use_reg)
+		omap_hsmmc_reg_put(host);
+err_irq:
+	if (host->tx_chan)
+		dma_release_channel(host->tx_chan);
+	if (host->rx_chan)
+		dma_release_channel(host->rx_chan);
+	pm_runtime_put_sync(host->dev);
+	pm_runtime_disable(host->dev);
+	if (host->dbclk)
+		clk_disable_unprepare(host->dbclk);
+err1:
+err_gpio:
+	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);
+	if (host->use_reg)
+		omap_hsmmc_reg_put(host);
+
+	if (host->tx_chan)
+		dma_release_channel(host->tx_chan);
+	if (host->rx_chan)
+		dma_release_channel(host->rx_chan);
+
+	pm_runtime_put_sync(host->dev);
+	pm_runtime_disable(host->dev);
+	if (host->dbclk)
+		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);
+	}
+
+	/* do not wake up due to sdio irq */
+	if ((host->mmc->caps & MMC_CAP_SDIO_IRQ) &&
+	    !(host->mmc->pm_flags & MMC_PM_WAKE_SDIO_IRQ))
+		disable_irq(host->wake_irq);
+
+	if (host->dbclk)
+		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);
+
+	if (host->dbclk)
+		clk_prepare_enable(host->dbclk);
+
+	if (!(host->mmc->pm_flags & MMC_PM_KEEP_POWER))
+		omap_hsmmc_conf_bus_power(host);
+
+	omap_hsmmc_protect_card(host);
+
+	if ((host->mmc->caps & MMC_CAP_SDIO_IRQ) &&
+	    !(host->mmc->pm_flags & MMC_PM_WAKE_SDIO_IRQ))
+		enable_irq(host->wake_irq);
+
+	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 = platform_get_drvdata(to_platform_device(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);
+
+		WARN_ON(host->flags & HSMMC_WAKE_IRQ_ENABLED);
+		enable_irq(host->wake_irq);
+		host->flags |= HSMMC_WAKE_IRQ_ENABLED;
+	} 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 = platform_get_drvdata(to_platform_device(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)) {
+		/* sdio irq flag can't change while in runtime suspend */
+		if (host->flags & HSMMC_WAKE_IRQ_ENABLED) {
+			disable_irq_nosync(host->wake_irq);
+			host->flags &= ~HSMMC_WAKE_IRQ_ENABLED;
+		}
+
+		pinctrl_pm_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_pm_select_default_state(host->dev);
+	}
+	spin_unlock_irqrestore(&host->irq_lock, flags);
+	return 0;
+}
+
+static 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,
+		.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/pxamci.c b/drivers/mmc/host/pxamci.c
new file mode 100644
index 000000000..1b6d0bfe3
--- /dev/null
+++ b/drivers/mmc/host/pxamci.c
@@ -0,0 +1,896 @@
+/*
+ *  linux/drivers/mmc/host/pxa.c - PXA MMCI driver
+ *
+ *  Copyright (C) 2003 Russell King, All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ *  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/dma-mapping.h>
+#include <linux/clk.h>
+#include <linux/err.h>
+#include <linux/mmc/host.h>
+#include <linux/io.h>
+#include <linux/regulator/consumer.h>
+#include <linux/gpio.h>
+#include <linux/gfp.h>
+#include <linux/of.h>
+#include <linux/of_gpio.h>
+#include <linux/of_device.h>
+
+#include <asm/sizes.h>
+
+#include <mach/hardware.h>
+#include <mach/dma.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;
+	int			irq;
+	int			dma;
+	unsigned int		clkrt;
+	unsigned int		cmdat;
+	unsigned int		imask;
+	unsigned int		power_mode;
+	struct pxamci_platform_data *pdata;
+
+	struct mmc_request	*mrq;
+	struct mmc_command	*cmd;
+	struct mmc_data		*data;
+
+	dma_addr_t		sg_dma;
+	struct pxa_dma_desc	*sg_cpu;
+	unsigned int		dma_len;
+
+	unsigned int		dma_dir;
+	unsigned int		dma_drcmrrx;
+	unsigned int		dma_drcmrtx;
+
+	struct regulator	*vcc;
+};
+
+static inline void pxamci_init_ocr(struct pxamci_host *host)
+{
+#ifdef CONFIG_REGULATOR
+	host->vcc = regulator_get_optional(mmc_dev(host->mmc), "vmmc");
+
+	if (IS_ERR(host->vcc))
+		host->vcc = NULL;
+	else {
+		host->mmc->ocr_avail = mmc_regulator_get_ocrmask(host->vcc);
+		if (host->pdata && host->pdata->ocr_mask)
+			dev_warn(mmc_dev(host->mmc),
+				"ocr_mask/setpower will not be used\n");
+	}
+#endif
+	if (host->vcc == NULL) {
+		/* fall-back to platform data */
+		host->mmc->ocr_avail = host->pdata ?
+			host->pdata->ocr_mask :
+			MMC_VDD_32_33 | MMC_VDD_33_34;
+	}
+}
+
+static inline int pxamci_set_power(struct pxamci_host *host,
+				    unsigned char power_mode,
+				    unsigned int vdd)
+{
+	int on;
+
+	if (host->vcc) {
+		int ret;
+
+		if (power_mode == MMC_POWER_UP) {
+			ret = mmc_regulator_set_ocr(host->mmc, host->vcc, vdd);
+			if (ret)
+				return ret;
+		} else if (power_mode == MMC_POWER_OFF) {
+			ret = mmc_regulator_set_ocr(host->mmc, host->vcc, 0);
+			if (ret)
+				return ret;
+		}
+	}
+	if (!host->vcc && host->pdata &&
+	    gpio_is_valid(host->pdata->gpio_power)) {
+		on = ((1 << vdd) & host->pdata->ocr_mask);
+		gpio_set_value(host->pdata->gpio_power,
+			       !!on ^ host->pdata->gpio_power_invert);
+	}
+	if (!host->vcc && 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_setup_data(struct pxamci_host *host, struct mmc_data *data)
+{
+	unsigned int nob = data->blocks;
+	unsigned long long clks;
+	unsigned int timeout;
+	bool dalgn = 0;
+	u32 dcmd;
+	int i;
+
+	host->data = data;
+
+	if (data->flags & MMC_DATA_STREAM)
+		nob = 0xffff;
+
+	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);
+
+	if (data->flags & MMC_DATA_READ) {
+		host->dma_dir = DMA_FROM_DEVICE;
+		dcmd = DCMD_INCTRGADDR | DCMD_FLOWSRC;
+		DRCMR(host->dma_drcmrtx) = 0;
+		DRCMR(host->dma_drcmrrx) = host->dma | DRCMR_MAPVLD;
+	} else {
+		host->dma_dir = DMA_TO_DEVICE;
+		dcmd = DCMD_INCSRCADDR | DCMD_FLOWTRG;
+		DRCMR(host->dma_drcmrrx) = 0;
+		DRCMR(host->dma_drcmrtx) = host->dma | DRCMR_MAPVLD;
+	}
+
+	dcmd |= DCMD_BURST32 | DCMD_WIDTH1;
+
+	host->dma_len = dma_map_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+				   host->dma_dir);
+
+	for (i = 0; i < host->dma_len; i++) {
+		unsigned int length = sg_dma_len(&data->sg[i]);
+		host->sg_cpu[i].dcmd = dcmd | length;
+		if (length & 31 && !(data->flags & MMC_DATA_READ))
+			host->sg_cpu[i].dcmd |= DCMD_ENDIRQEN;
+		/* Not aligned to 8-byte boundary? */
+		if (sg_dma_address(&data->sg[i]) & 0x7)
+			dalgn = 1;
+		if (data->flags & MMC_DATA_READ) {
+			host->sg_cpu[i].dsadr = host->res->start + MMC_RXFIFO;
+			host->sg_cpu[i].dtadr = sg_dma_address(&data->sg[i]);
+		} else {
+			host->sg_cpu[i].dsadr = sg_dma_address(&data->sg[i]);
+			host->sg_cpu[i].dtadr = host->res->start + MMC_TXFIFO;
+		}
+		host->sg_cpu[i].ddadr = host->sg_dma + (i + 1) *
+					sizeof(struct pxa_dma_desc);
+	}
+	host->sg_cpu[host->dma_len - 1].ddadr = DDADR_STOP;
+	wmb();
+
+	/*
+	 * The PXA27x DMA controller encounters overhead when working with
+	 * unaligned (to 8-byte boundaries) data, so switch on byte alignment
+	 * mode only if we have unaligned data.
+	 */
+	if (dalgn)
+		DALGN |= (1 << host->dma);
+	else
+		DALGN &= ~(1 << host->dma);
+	DDADR(host->dma) = host->sg_dma;
+
+	/*
+	 * 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)
+		DCSR(host->dma) = DCSR_RUN;
+}
+
+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)
+			DCSR(host->dma) = DCSR_RUN;
+	} 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;
+
+	if (!data)
+		return 0;
+
+	DCSR(host->dma) = 0;
+	dma_unmap_sg(mmc_dev(host->mmc), 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;
+
+		if (mrq->data->flags & MMC_DATA_STREAM)
+			cmdat |= CMDAT_STREAM;
+	}
+
+	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->pdata && gpio_is_valid(host->pdata->gpio_card_ro)) {
+		if (host->pdata->gpio_card_ro_invert)
+			return !gpio_get_value(host->pdata->gpio_card_ro);
+		else
+			return gpio_get_value(host->pdata->gpio_card_ro);
+	}
+	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_ro			= pxamci_get_ro,
+	.set_ios		= pxamci_set_ios,
+	.enable_sdio_irq	= pxamci_enable_sdio_irq,
+};
+
+static void pxamci_dma_irq(int dma, void *devid)
+{
+	struct pxamci_host *host = devid;
+	int dcsr = DCSR(dma);
+	DCSR(dma) = dcsr & ~DCSR_STOPIRQEN;
+
+	if (dcsr & DCSR_ENDINTR) {
+		writel(BUF_PART_FULL, host->base + MMC_PRTBUF);
+	} else {
+		pr_err("%s: DMA error on channel %d (DCSR=%#x)\n",
+		       mmc_hostname(host->mmc), dma, dcsr);
+		host->data->error = -EIO;
+		pxamci_data_done(host, 0);
+	}
+}
+
+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->pdata->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 device_node *np = pdev->dev.of_node;
+        struct pxamci_platform_data *pdata;
+        u32 tmp;
+
+        if (!np)
+                return 0;
+
+        pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
+        if (!pdata)
+                return -ENOMEM;
+
+	pdata->gpio_card_detect =
+		of_get_named_gpio(np, "cd-gpios", 0);
+	pdata->gpio_card_ro =
+		of_get_named_gpio(np, "wp-gpios", 0);
+
+	/* pxa-mmc specific */
+	pdata->gpio_power =
+		of_get_named_gpio(np, "pxa-mmc,gpio-power", 0);
+
+	if (of_property_read_u32(np, "pxa-mmc,detect-delay-ms", &tmp) == 0)
+		pdata->detect_delay_ms = tmp;
+
+        pdev->dev.platform_data = pdata;
+
+        return 0;
+}
+#else
+static int pxamci_of_init(struct platform_device *pdev)
+{
+        return 0;
+}
+#endif
+
+static int pxamci_probe(struct platform_device *pdev)
+{
+	struct mmc_host *mmc;
+	struct pxamci_host *host = NULL;
+	struct resource *r, *dmarx, *dmatx;
+	int ret, irq, gpio_cd = -1, gpio_ro = -1, gpio_power = -1;
+
+	ret = pxamci_of_init(pdev);
+	if (ret)
+		return ret;
+
+	r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	irq = platform_get_irq(pdev, 0);
+	if (!r || irq < 0)
+		return -ENXIO;
+
+	r = request_mem_region(r->start, SZ_4K, DRIVER_NAME);
+	if (!r)
+		return -EBUSY;
+
+	mmc = mmc_alloc_host(sizeof(struct pxamci_host), &pdev->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;
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+	host->dma = -1;
+	host->pdata = pdev->dev.platform_data;
+	host->clkrt = CLKRT_OFF;
+
+	host->clk = clk_get(&pdev->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;
+
+	pxamci_init_ocr(host);
+
+	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;
+	}
+
+	host->sg_cpu = dma_alloc_coherent(&pdev->dev, PAGE_SIZE, &host->sg_dma, GFP_KERNEL);
+	if (!host->sg_cpu) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	spin_lock_init(&host->lock);
+	host->res = r;
+	host->irq = irq;
+	host->imask = MMC_I_MASK_ALL;
+
+	host->base = ioremap(r->start, SZ_4K);
+	if (!host->base) {
+		ret = -ENOMEM;
+		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);
+
+	host->dma = pxa_request_dma(DRIVER_NAME, DMA_PRIO_LOW,
+				    pxamci_dma_irq, host);
+	if (host->dma < 0) {
+		ret = -EBUSY;
+		goto out;
+	}
+
+	ret = request_irq(host->irq, pxamci_irq, 0, DRIVER_NAME, host);
+	if (ret)
+		goto out;
+
+	platform_set_drvdata(pdev, mmc);
+
+	dmarx = platform_get_resource(pdev, IORESOURCE_DMA, 0);
+	if (!dmarx) {
+		ret = -ENXIO;
+		goto out;
+	}
+	host->dma_drcmrrx = dmarx->start;
+
+	dmatx = platform_get_resource(pdev, IORESOURCE_DMA, 1);
+	if (!dmatx) {
+		ret = -ENXIO;
+		goto out;
+	}
+	host->dma_drcmrtx = dmatx->start;
+
+	if (host->pdata) {
+		gpio_cd = host->pdata->gpio_card_detect;
+		gpio_ro = host->pdata->gpio_card_ro;
+		gpio_power = host->pdata->gpio_power;
+	}
+	if (gpio_is_valid(gpio_power)) {
+		ret = gpio_request(gpio_power, "mmc card power");
+		if (ret) {
+			dev_err(&pdev->dev, "Failed requesting gpio_power %d\n", gpio_power);
+			goto out;
+		}
+		gpio_direction_output(gpio_power,
+				      host->pdata->gpio_power_invert);
+	}
+	if (gpio_is_valid(gpio_ro)) {
+		ret = gpio_request(gpio_ro, "mmc card read only");
+		if (ret) {
+			dev_err(&pdev->dev, "Failed requesting gpio_ro %d\n", gpio_ro);
+			goto err_gpio_ro;
+		}
+		gpio_direction_input(gpio_ro);
+	}
+	if (gpio_is_valid(gpio_cd)) {
+		ret = gpio_request(gpio_cd, "mmc card detect");
+		if (ret) {
+			dev_err(&pdev->dev, "Failed requesting gpio_cd %d\n", gpio_cd);
+			goto err_gpio_cd;
+		}
+		gpio_direction_input(gpio_cd);
+
+		ret = request_irq(gpio_to_irq(gpio_cd), pxamci_detect_irq,
+				  IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING,
+				  "mmc card detect", mmc);
+		if (ret) {
+			dev_err(&pdev->dev, "failed to request card detect IRQ\n");
+			goto err_request_irq;
+		}
+	}
+
+	if (host->pdata && host->pdata->init)
+		host->pdata->init(&pdev->dev, pxamci_detect_irq, mmc);
+
+	if (gpio_is_valid(gpio_power) && host->pdata->setpower)
+		dev_warn(&pdev->dev, "gpio_power and setpower() both defined\n");
+	if (gpio_is_valid(gpio_ro) && host->pdata->get_ro)
+		dev_warn(&pdev->dev, "gpio_ro and get_ro() both defined\n");
+
+	mmc_add_host(mmc);
+
+	return 0;
+
+err_request_irq:
+	gpio_free(gpio_cd);
+err_gpio_cd:
+	gpio_free(gpio_ro);
+err_gpio_ro:
+	gpio_free(gpio_power);
+ out:
+	if (host) {
+		if (host->dma >= 0)
+			pxa_free_dma(host->dma);
+		if (host->base)
+			iounmap(host->base);
+		if (host->sg_cpu)
+			dma_free_coherent(&pdev->dev, PAGE_SIZE, host->sg_cpu, host->sg_dma);
+		if (host->clk)
+			clk_put(host->clk);
+	}
+	if (mmc)
+		mmc_free_host(mmc);
+	release_resource(r);
+	return ret;
+}
+
+static int pxamci_remove(struct platform_device *pdev)
+{
+	struct mmc_host *mmc = platform_get_drvdata(pdev);
+	int gpio_cd = -1, gpio_ro = -1, gpio_power = -1;
+
+	if (mmc) {
+		struct pxamci_host *host = mmc_priv(mmc);
+
+		mmc_remove_host(mmc);
+
+		if (host->pdata) {
+			gpio_cd = host->pdata->gpio_card_detect;
+			gpio_ro = host->pdata->gpio_card_ro;
+			gpio_power = host->pdata->gpio_power;
+		}
+		if (gpio_is_valid(gpio_cd)) {
+			free_irq(gpio_to_irq(gpio_cd), mmc);
+			gpio_free(gpio_cd);
+		}
+		if (gpio_is_valid(gpio_ro))
+			gpio_free(gpio_ro);
+		if (gpio_is_valid(gpio_power))
+			gpio_free(gpio_power);
+		if (host->vcc)
+			regulator_put(host->vcc);
+
+		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);
+
+		DRCMR(host->dma_drcmrrx) = 0;
+		DRCMR(host->dma_drcmrtx) = 0;
+
+		free_irq(host->irq, host);
+		pxa_free_dma(host->dma);
+		iounmap(host->base);
+		dma_free_coherent(&pdev->dev, PAGE_SIZE, host->sg_cpu, host->sg_dma);
+
+		clk_put(host->clk);
+
+		release_resource(host->res);
+
+		mmc_free_host(mmc);
+	}
+	return 0;
+}
+
+static struct platform_driver pxamci_driver = {
+	.probe		= pxamci_probe,
+	.remove		= pxamci_remove,
+	.driver		= {
+		.name	= DRIVER_NAME,
+		.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..f6c2e2fcc
--- /dev/null
+++ b/drivers/mmc/host/pxamci.h
@@ -0,0 +1,90 @@
+#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/rtsx_pci_sdmmc.c b/drivers/mmc/host/rtsx_pci_sdmmc.c
new file mode 100644
index 000000000..1d3d6c4bf
--- /dev/null
+++ b/drivers/mmc/host/rtsx_pci_sdmmc.c
@@ -0,0 +1,1498 @@
+/* Realtek PCI-Express SD/MMC Card Interface driver
+ *
+ * Copyright(c) 2009-2013 Realtek Semiconductor Corp. All rights reserved.
+ *
+ * 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; either version 2, or (at your option) any
+ * later version.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, see <http://www.gnu.org/licenses/>.
+ *
+ * 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/mfd/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;
+	struct workqueue_struct *workq;
+#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			power_state;
+#define SDMMC_POWER_ON		1
+#define SDMMC_POWER_OFF		0
+
+	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 & ~MMC_RSP_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,
+		bool is_first_req)
+{
+	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 = 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 int sd_rw_multi(struct realtek_pci_sdmmc *host, struct mmc_request *mrq)
+{
+	struct mmc_data *data = mrq->data;
+
+	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)
+		return sd_read_long_data(host, mrq);
+
+	return sd_write_long_data(host, mrq);
+}
+
+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 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;
+	int err;
+
+	dev_dbg(sdmmc_dev(host), "%s(%s): sample_point = %d\n",
+			__func__, rx ? "RX" : "TX", sample_point);
+
+	rtsx_pci_init_cmd(pcr);
+
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CLK_CTL, CHANGE_CLK, CHANGE_CLK);
+	if (rx)
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD,
+				SD_VPRX_CTL, 0x1F, sample_point);
+	else
+		rtsx_pci_add_cmd(pcr, WRITE_REG_CMD,
+				SD_VPTX_CTL, 0x1F, sample_point);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_VPCLK0_CTL, PHASE_NOT_RESET, 0);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_VPCLK0_CTL,
+			PHASE_NOT_RESET, PHASE_NOT_RESET);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, CLK_CTL, CHANGE_CLK, 0);
+	rtsx_pci_add_cmd(pcr, WRITE_REG_CMD, SD_CFG1, SD_ASYNC_FIFO_NOT_RST, 0);
+
+	err = rtsx_pci_send_cmd(pcr, 100);
+	if (err < 0)
+		return err;
+
+	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 err, i;
+	u8 val = 0;
+
+	for (i = 0; i < 100; i++) {
+		err = 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 = {0};
+
+	err = sd_change_phase(host, sample_point, true);
+	if (err < 0)
+		return err;
+
+	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);
+		return err;
+	}
+
+	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);
+
+	queue_work(host->workq, &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)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	int err;
+
+	if (host->power_state == SDMMC_POWER_ON)
+		return 0;
+
+	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;
+
+	err = rtsx_pci_write_register(pcr, CARD_OE, SD_OUTPUT_EN, SD_OUTPUT_EN);
+	if (err < 0)
+		return err;
+
+	host->power_state = SDMMC_POWER_ON;
+	return 0;
+}
+
+static int sd_power_off(struct realtek_pci_sdmmc *host)
+{
+	struct rtsx_pcr *pcr = host->pcr;
+	int err;
+
+	host->power_state = SDMMC_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);
+
+	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;
+}
+
+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->workq = create_singlethread_workqueue(SDMMC_WORKQ_NAME);
+	if (!host->workq) {
+		mmc_free_host(mmc);
+		return -ENOMEM;
+	}
+	host->pcr = pcr;
+	host->mmc = mmc;
+	host->pdev = pdev;
+	host->cookie = -1;
+	host->power_state = SDMMC_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_workqueue(host->workq);
+	destroy_workqueue(host->workq);
+	host->workq = NULL;
+
+	mmc_free_host(mmc);
+
+	dev_dbg(&(pdev->dev),
+		": Realtek PCI-E SDMMC controller has been removed\n");
+
+	return 0;
+}
+
+static 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,
+	},
+};
+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..88af827e0
--- /dev/null
+++ b/drivers/mmc/host/rtsx_usb_sdmmc.c
@@ -0,0 +1,1463 @@
+/* Realtek USB SD/MMC Card Interface driver
+ *
+ * Copyright(c) 2009-2013 Realtek Semiconductor Corp. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, see <http://www.gnu.org/licenses/>.
+ *
+ * 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/sdio.h>
+#include <linux/mmc/card.h>
+#include <linux/scatterlist.h>
+#include <linux/pm_runtime.h>
+
+#include <linux/mfd/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 & ~MMC_RSP_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 = 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;
+	int err;
+
+	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);
+
+	err = rtsx_usb_send_cmd(ucr, MODE_C, 100);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+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 err, i;
+	u8 val = 0;
+
+	for (i = 0; i < 100; i++) {
+		err = 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 = {0};
+
+	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;
+	}
+
+	/*
+	 * Reject SDIO CMDs to speed up card identification
+	 * since unsupported
+	 */
+	if (cmd->opcode == SD_IO_SEND_OP_COND ||
+	    cmd->opcode == SD_IO_RW_DIRECT ||
+	    cmd->opcode == SD_IO_RW_EXTENDED) {
+		cmd->error = -EINVAL;
+		goto finish;
+	}
+
+	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;
+	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(sdmmc_dev(host));
+	} else {
+		pm_runtime_get_sync(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);
+
+	if (rtsx_usb_card_exclusive_check(ucr, RTSX_USB_SD_CARD)) {
+		mutex_unlock(&ucr->dev_mutex);
+		return;
+	}
+
+	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;
+
+	mutex_lock(&ucr->dev_mutex);
+
+	if (host->led.brightness == LED_OFF)
+		rtsx_usb_turn_off_led(ucr);
+	else
+		rtsx_usb_turn_on_led(ucr);
+
+	mutex_unlock(&ucr->dev_mutex);
+}
+#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_NEEDS_POLL;
+	mmc->caps2 = MMC_CAP2_NO_PRESCAN_POWERUP | MMC_CAP2_FULL_PWR_CYCLE;
+
+	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
+
+	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
+	mmc_add_host(mmc);
+
+	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;
+}
+
+static 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",
+	},
+};
+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..94cddf381
--- /dev/null
+++ b/drivers/mmc/host/s3cmci.c
@@ -0,0 +1,1889 @@
+/*
+ *  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>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#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/irq.h>
+#include <linux/io.h>
+
+#include <plat/gpio-cfg.h>
+#include <mach/dma.h>
+#include <mach/gpio-samsung.h>
+
+#include <linux/platform_data/dma-s3c24xx.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, bsize;
+	u32 datcon, datcnt, datsta, fsta, imask;
+
+	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);
+	bsize 	= readl(host->base + S3C2410_SDIBSIZE);
+	datcon 	= readl(host->base + S3C2410_SDIDCON);
+	datcnt 	= readl(host->base + S3C2410_SDIDCNT);
+	datsta 	= readl(host->base + S3C2410_SDIDSTA);
+	fsta 	= readl(host->base + S3C2410_SDIFSTA);
+	imask   = readl(host->base + host->sdiimsk);
+
+	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 (gpio_get_value(S3C2410_GPE(8)) == 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;
+
+}
+
+/*
+ * ISR for the CardDetect Pin
+*/
+
+static irqreturn_t s3cmci_irq_cd(int irq, void *dev_id)
+{
+	struct s3cmci_host *host = (struct s3cmci_host *)dev_id;
+
+	dbg(host, dbg_irq, "card detect\n");
+
+	mmc_detect_change(host->mmc, msecs_to_jiffies(500));
+
+	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;
+
+	/* write DCON register */
+
+	if (!data) {
+		writel(0, host->base + S3C2410_SDIDCON);
+		return 0;
+	}
+
+	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;
+
+	if (!(data->flags & MMC_DATA_STREAM))
+		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,
+			     rw ? DMA_TO_DEVICE : DMA_FROM_DEVICE);
+
+	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,
+			     rw ? DMA_TO_DEVICE : DMA_FROM_DEVICE);
+	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 int s3cmci_card_present(struct mmc_host *mmc)
+{
+	struct s3cmci_host *host = mmc_priv(mmc);
+	struct s3c24xx_mci_pdata *pdata = host->pdata;
+	int ret;
+
+	if (pdata->no_detect)
+		return -ENOSYS;
+
+	ret = gpio_get_value(pdata->gpio_detect) ? 0 : 1;
+	return ret ^ pdata->detect_invert;
+}
+
+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 (s3cmci_card_present(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:
+		/* Configure GPE5...GPE10 pins in SD mode */
+		s3c_gpio_cfgall_range(S3C2410_GPE(5), 6, S3C_GPIO_SFN(2),
+				      S3C_GPIO_PULL_NONE);
+
+		if (host->pdata->set_power)
+			host->pdata->set_power(ios->power_mode, ios->vdd);
+
+		if (!host->is2440)
+			mci_con |= S3C2410_SDICON_FIFORESET;
+
+		break;
+
+	case MMC_POWER_OFF:
+	default:
+		gpio_direction_output(S3C2410_GPE(5), 0);
+
+		if (host->is2440)
+			mci_con |= S3C2440_SDICON_SDRESET;
+
+		if (host->pdata->set_power)
+			host->pdata->set_power(ios->power_mode, ios->vdd);
+
+		break;
+	}
+
+	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 int s3cmci_get_ro(struct mmc_host *mmc)
+{
+	struct s3cmci_host *host = mmc_priv(mmc);
+	struct s3c24xx_mci_pdata *pdata = host->pdata;
+	int ret;
+
+	if (pdata->no_wprotect)
+		return 0;
+
+	ret = gpio_get_value(pdata->gpio_wprotect) ? 1 : 0;
+	ret ^= pdata->wprotect_invert;
+
+	return ret;
+}
+
+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 struct mmc_host_ops s3cmci_ops = {
+	.request	= s3cmci_request,
+	.set_ios	= s3cmci_set_ios,
+	.get_ro		= s3cmci_get_ro,
+	.get_cd		= s3cmci_card_present,
+	.enable_sdio_irq = s3cmci_enable_sdio_irq,
+};
+
+static struct s3c24xx_mci_pdata s3cmci_def_pdata = {
+	/* This is currently here to avoid a number of if (host->pdata)
+	 * checks. Any zero fields to ensure reasonable defaults are picked. */
+	 .no_wprotect = 1,
+	 .no_detect = 1,
+};
+
+#ifdef CONFIG_CPU_FREQ
+
+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%08x\n", (u32)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;
+}
+
+static int s3cmci_state_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, s3cmci_state_show, inode->i_private);
+}
+
+static const struct file_operations s3cmci_fops_state = {
+	.owner		= THIS_MODULE,
+	.open		= s3cmci_state_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+#define DBG_REG(_r) { .addr = S3C2410_SDI##_r, .name = #_r }
+
+struct s3cmci_reg {
+	unsigned short	addr;
+	unsigned char	*name;
+} 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;
+	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;
+}
+
+static int s3cmci_regs_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, s3cmci_regs_show, inode->i_private);
+}
+
+static const struct file_operations s3cmci_fops_regs = {
+	.owner		= THIS_MODULE,
+	.open		= s3cmci_regs_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= single_release,
+};
+
+static void s3cmci_debugfs_attach(struct s3cmci_host *host)
+{
+	struct device *dev = &host->pdev->dev;
+
+	host->debug_root = debugfs_create_dir(dev_name(dev), NULL);
+	if (IS_ERR(host->debug_root)) {
+		dev_err(dev, "failed to create debugfs root\n");
+		return;
+	}
+
+	host->debug_state = debugfs_create_file("state", 0444,
+						host->debug_root, host,
+						&s3cmci_fops_state);
+
+	if (IS_ERR(host->debug_state))
+		dev_err(dev, "failed to create debug state file\n");
+
+	host->debug_regs = debugfs_create_file("regs", 0444,
+					       host->debug_root, host,
+					       &s3cmci_fops_regs);
+
+	if (IS_ERR(host->debug_regs))
+		dev_err(dev, "failed to create debug regs file\n");
+}
+
+static void s3cmci_debugfs_remove(struct s3cmci_host *host)
+{
+	debugfs_remove(host->debug_regs);
+	debugfs_remove(host->debug_state);
+	debugfs_remove(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(struct platform_device *pdev)
+{
+	struct s3cmci_host *host;
+	struct mmc_host	*mmc;
+	int ret;
+	int is2440;
+	int i;
+
+	is2440 = platform_get_device_id(pdev)->driver_data;
+
+	mmc = mmc_alloc_host(sizeof(struct s3cmci_host), &pdev->dev);
+	if (!mmc) {
+		ret = -ENOMEM;
+		goto probe_out;
+	}
+
+	for (i = S3C2410_GPE(5); i <= S3C2410_GPE(10); i++) {
+		ret = gpio_request(i, dev_name(&pdev->dev));
+		if (ret) {
+			dev_err(&pdev->dev, "failed to get gpio %d\n", i);
+
+			for (i--; i >= S3C2410_GPE(5); i--)
+				gpio_free(i);
+
+			goto probe_free_host;
+		}
+	}
+
+	host = mmc_priv(mmc);
+	host->mmc 	= mmc;
+	host->pdev	= pdev;
+	host->is2440	= is2440;
+
+	host->pdata = pdev->dev.platform_data;
+	if (!host->pdata) {
+		pdev->dev.platform_data = &s3cmci_def_pdata;
+		host->pdata = &s3cmci_def_pdata;
+	}
+
+	spin_lock_init(&host->complete_lock);
+	tasklet_init(&host->pio_tasklet, pio_tasklet, (unsigned long) host);
+
+	if (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_gpio;
+	}
+
+	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_gpio;
+	}
+
+	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) {
+		dev_err(&pdev->dev, "failed to get interrupt resource.\n");
+		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;
+
+	if (!host->pdata->no_detect) {
+		ret = gpio_request(host->pdata->gpio_detect, "s3cmci detect");
+		if (ret) {
+			dev_err(&pdev->dev, "failed to get detect gpio\n");
+			goto probe_free_irq;
+		}
+
+		host->irq_cd = gpio_to_irq(host->pdata->gpio_detect);
+
+		if (host->irq_cd >= 0) {
+			if (request_irq(host->irq_cd, s3cmci_irq_cd,
+					IRQF_TRIGGER_RISING |
+					IRQF_TRIGGER_FALLING,
+					DRIVER_NAME, host)) {
+				dev_err(&pdev->dev,
+					"can't get card detect irq.\n");
+				ret = -ENOENT;
+				goto probe_free_gpio_cd;
+			}
+		} else {
+			dev_warn(&pdev->dev,
+				 "host detect has no irq available\n");
+			gpio_direction_input(host->pdata->gpio_detect);
+		}
+	} else
+		host->irq_cd = -1;
+
+	if (!host->pdata->no_wprotect) {
+		ret = gpio_request(host->pdata->gpio_wprotect, "s3cmci wp");
+		if (ret) {
+			dev_err(&pdev->dev, "failed to get writeprotect\n");
+			goto probe_free_irq_cd;
+		}
+
+		gpio_direction_input(host->pdata->gpio_wprotect);
+	}
+
+	/* depending on the dma state, get a dma channel to use. */
+
+	if (s3cmci_host_usedma(host)) {
+		dma_cap_mask_t mask;
+
+		dma_cap_zero(mask);
+		dma_cap_set(DMA_SLAVE, mask);
+
+		host->dma = dma_request_slave_channel_compat(mask,
+			s3c24xx_dma_filter, (void *)DMACH_SDI, &pdev->dev, "rx-tx");
+		if (!host->dma) {
+			dev_err(&pdev->dev, "cannot get DMA channel.\n");
+			ret = -EBUSY;
+			goto probe_free_gpio_wp;
+		}
+	}
+
+	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_gpio_wp:
+	if (!host->pdata->no_wprotect)
+		gpio_free(host->pdata->gpio_wprotect);
+
+ probe_free_gpio_cd:
+	if (!host->pdata->no_detect)
+		gpio_free(host->pdata->gpio_detect);
+
+ probe_free_irq_cd:
+	if (host->irq_cd >= 0)
+		free_irq(host->irq_cd, host);
+
+ 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_gpio:
+	for (i = S3C2410_GPE(5); i <= S3C2410_GPE(10); i++)
+		gpio_free(i);
+
+ 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);
+	struct s3c24xx_mci_pdata *pd = host->pdata;
+	int i;
+
+	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);
+
+	if (!pd->no_wprotect)
+		gpio_free(pd->gpio_wprotect);
+
+	if (!pd->no_detect)
+		gpio_free(pd->gpio_detect);
+
+	for (i = S3C2410_GPE(5); i <= S3C2410_GPE(10); i++)
+		gpio_free(i);
+
+
+	iounmap(host->base);
+	release_mem_region(host->mem->start, resource_size(host->mem));
+
+	mmc_free_host(mmc);
+	return 0;
+}
+
+static 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",
+	},
+	.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..cc2e46cb5
--- /dev/null
+++ b/drivers/mmc/host/s3cmci.h
@@ -0,0 +1,80 @@
+/*
+ *  linux/drivers/mmc/s3cmci.h - Samsung S3C MCI driver
+ *
+ *  Copyright (C) 2004-2006 Thomas Kleffel, All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+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;
+	struct dentry		*debug_state;
+	struct dentry		*debug_regs;
+#endif
+
+#ifdef CONFIG_CPU_FREQ
+	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..22d929fa3
--- /dev/null
+++ b/drivers/mmc/host/sdhci-acpi.c
@@ -0,0 +1,473 @@
+/*
+ * Secure Digital Host Controller Interface ACPI driver.
+ *
+ * Copyright (c) 2012, Intel Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ */
+
+#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/mmc/host.h>
+#include <linux/mmc/pm.h>
+#include <linux/mmc/slot-gpio.h>
+
+#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;
+	int (*probe_slot)(struct platform_device *, const char *, const char *);
+	int (*remove_slot)(struct platform_device *);
+};
+
+struct sdhci_acpi_host {
+	struct sdhci_host		*host;
+	const struct sdhci_acpi_slot	*slot;
+	struct platform_device		*pdev;
+	bool				use_runtime_pm;
+	bool				dma_setup;
+};
+
+static inline bool sdhci_acpi_flag(struct sdhci_acpi_host *c, unsigned int flag)
+{
+	return c->slot && (c->slot->flags & flag);
+}
+
+static int sdhci_acpi_enable_dma(struct sdhci_host *host)
+{
+	struct sdhci_acpi_host *c = sdhci_priv(host);
+	struct device *dev = &c->pdev->dev;
+	int err = -1;
+
+	if (c->dma_setup)
+		return 0;
+
+	if (host->flags & SDHCI_USE_64_BIT_DMA) {
+		if (host->quirks2 & SDHCI_QUIRK2_BROKEN_64_BIT_DMA) {
+			host->flags &= ~SDHCI_USE_64_BIT_DMA;
+		} else {
+			err = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(64));
+			if (err)
+				dev_warn(dev, "Failed to set 64-bit DMA mask\n");
+		}
+	}
+
+	if (err)
+		err = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(32));
+
+	c->dma_setup = !err;
+
+	return err;
+}
+
+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,
+	.enable_dma = sdhci_acpi_enable_dma,
+	.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,
+	.enable_dma = sdhci_acpi_enable_dma,
+	.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,
+};
+
+static int sdhci_acpi_emmc_probe_slot(struct platform_device *pdev,
+				      const char *hid, const char *uid)
+{
+	struct sdhci_acpi_host *c = platform_get_drvdata(pdev);
+	struct sdhci_host *host;
+
+	if (!c || !c->host)
+		return 0;
+
+	host = c->host;
+
+	/* Platform specific code during emmc probe slot goes here */
+
+	if (hid && uid && !strcmp(hid, "80860F14") && !strcmp(uid, "1") &&
+	    sdhci_readl(host, SDHCI_CAPABILITIES) == 0x446cc8b2 &&
+	    sdhci_readl(host, SDHCI_CAPABILITIES_1) == 0x00000807)
+		host->timeout_clk = 1000; /* 1000 kHz i.e. 1 MHz */
+
+	return 0;
+}
+
+static int sdhci_acpi_sdio_probe_slot(struct platform_device *pdev,
+				      const char *hid, const char *uid)
+{
+	struct sdhci_acpi_host *c = platform_get_drvdata(pdev);
+	struct sdhci_host *host;
+
+	if (!c || !c->host)
+		return 0;
+
+	host = c->host;
+
+	/* Platform specific code during sdio probe slot goes here */
+
+	return 0;
+}
+
+static int sdhci_acpi_sd_probe_slot(struct platform_device *pdev,
+				    const char *hid, const char *uid)
+{
+	struct sdhci_acpi_host *c = platform_get_drvdata(pdev);
+	struct sdhci_host *host;
+
+	if (!c || !c->host || !c->slot)
+		return 0;
+
+	host = c->host;
+
+	/* Platform specific code during sd probe slot goes here */
+
+	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_BUS_WIDTH_TEST | MMC_CAP_WAIT_WHILE_BUSY,
+	.caps2   = MMC_CAP2_HC_ERASE_SZ,
+	.flags   = SDHCI_ACPI_RUNTIME_PM,
+	.quirks  = SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.quirks2 = SDHCI_QUIRK2_PRESET_VALUE_BROKEN | SDHCI_QUIRK2_STOP_WITH_TC,
+	.probe_slot	= sdhci_acpi_emmc_probe_slot,
+};
+
+static const struct sdhci_acpi_slot sdhci_acpi_slot_int_sdio = {
+	.quirks  = SDHCI_QUIRK_BROKEN_CARD_DETECTION |
+		   SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.quirks2 = SDHCI_QUIRK2_HOST_OFF_CARD_ON,
+	.caps    = MMC_CAP_NONREMOVABLE | MMC_CAP_POWER_OFF_CARD |
+		   MMC_CAP_BUS_WIDTH_TEST | MMC_CAP_WAIT_WHILE_BUSY,
+	.flags   = SDHCI_ACPI_RUNTIME_PM,
+	.pm_caps = MMC_PM_KEEP_POWER,
+	.probe_slot	= sdhci_acpi_sdio_probe_slot,
+};
+
+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,
+	.quirks2 = SDHCI_QUIRK2_CARD_ON_NEEDS_BUS_ON |
+		   SDHCI_QUIRK2_STOP_WITH_TC,
+	.caps    = MMC_CAP_BUS_WIDTH_TEST | MMC_CAP_WAIT_WHILE_BUSY,
+	.probe_slot	= sdhci_acpi_sd_probe_slot,
+};
+
+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[] = {
+	{ "80860F14" , "1" , &sdhci_acpi_slot_int_emmc },
+	{ "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 },
+	{ "PNP0D40"  },
+	{ },
+};
+
+static const struct acpi_device_id sdhci_acpi_ids[] = {
+	{ "80860F14" },
+	{ "80860F16" },
+	{ "INT33BB"  },
+	{ "INT33C6"  },
+	{ "INT3436"  },
+	{ "INT344D"  },
+	{ "PNP0D40"  },
+	{ },
+};
+MODULE_DEVICE_TABLE(acpi, sdhci_acpi_ids);
+
+static const struct sdhci_acpi_slot *sdhci_acpi_get_slot(const char *hid,
+							 const char *uid)
+{
+	const struct sdhci_acpi_uid_slot *u;
+
+	for (u = sdhci_acpi_uids; u->hid; u++) {
+		if (strcmp(u->hid, hid))
+			continue;
+		if (!u->uid)
+			return u->slot;
+		if (uid && !strcmp(u->uid, uid))
+			return u->slot;
+	}
+	return NULL;
+}
+
+static int sdhci_acpi_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	acpi_handle handle = ACPI_HANDLE(dev);
+	struct acpi_device *device;
+	struct sdhci_acpi_host *c;
+	struct sdhci_host *host;
+	struct resource *iomem;
+	resource_size_t len;
+	const char *hid;
+	const char *uid;
+	int err;
+
+	if (acpi_bus_get_device(handle, &device))
+		return -ENODEV;
+
+	if (acpi_bus_get_status(device) || !device->status.present)
+		return -ENODEV;
+
+	hid = acpi_device_hid(device);
+	uid = device->pnp.unique_id;
+
+	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;
+
+	host = sdhci_alloc_host(dev, sizeof(struct sdhci_acpi_host));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	c = sdhci_priv(host);
+	c->host = host;
+	c->slot = sdhci_acpi_get_slot(hid, uid);
+	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);
+
+	host->ioaddr = devm_ioremap_nocache(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, hid, uid);
+			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);
+
+		if (mmc_gpiod_request_cd(host->mmc, NULL, 0, v, 0, NULL)) {
+			dev_warn(dev, "failed to setup card detect gpio\n");
+			c->use_runtime_pm = false;
+		}
+	}
+
+	err = sdhci_add_host(host);
+	if (err)
+		goto err_free;
+
+	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);
+	}
+
+	return 0;
+
+err_free:
+	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);
+	sdhci_free_host(c->host);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+
+static int sdhci_acpi_suspend(struct device *dev)
+{
+	struct sdhci_acpi_host *c = dev_get_drvdata(dev);
+
+	return sdhci_suspend_host(c->host);
+}
+
+static int sdhci_acpi_resume(struct device *dev)
+{
+	struct sdhci_acpi_host *c = dev_get_drvdata(dev);
+
+	return sdhci_resume_host(c->host);
+}
+
+#else
+
+#define sdhci_acpi_suspend	NULL
+#define sdhci_acpi_resume	NULL
+
+#endif
+
+#ifdef CONFIG_PM
+
+static int sdhci_acpi_runtime_suspend(struct device *dev)
+{
+	struct sdhci_acpi_host *c = dev_get_drvdata(dev);
+
+	return sdhci_runtime_suspend_host(c->host);
+}
+
+static int sdhci_acpi_runtime_resume(struct device *dev)
+{
+	struct sdhci_acpi_host *c = dev_get_drvdata(dev);
+
+	return sdhci_runtime_resume_host(c->host);
+}
+
+#endif
+
+static const struct dev_pm_ops sdhci_acpi_pm_ops = {
+	.suspend		= sdhci_acpi_suspend,
+	.resume			= 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",
+		.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..2bd90fb35
--- /dev/null
+++ b/drivers/mmc/host/sdhci-bcm-kona.c
@@ -0,0 +1,339 @@
+/*
+ * 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/gpio.h>
+#include <linux/clk.h>
+#include <linux/regulator/consumer.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/of_gpio.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 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 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;
+	}
+
+	if (clk_set_rate(pltfm_priv->clk, host->mmc->f_max) != 0) {
+		dev_err(dev, "Failed to set rate core clock\n");
+		goto err_pltfm_free;
+	}
+
+	if (clk_prepare_enable(pltfm_priv->clk) != 0) {
+		dev_err(dev, "Failed to enable core clock\n");
+		goto err_pltfm_free;
+	}
+
+	dev_dbg(dev, "non-removable=%c\n",
+		(host->mmc->caps & MMC_CAP_NONREMOVABLE) ? 'Y' : 'N');
+	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 (host->mmc->caps & MMC_CAP_NONREMOVABLE)
+		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) {
+		dev_err(dev, "Failed sdhci_add_host\n");
+		goto err_reset;
+	}
+
+	/* if device is eMMC, emulate card insert right here */
+	if (host->mmc->caps & MMC_CAP_NONREMOVABLE) {
+		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",
+		.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-bcm2835.c b/drivers/mmc/host/sdhci-bcm2835.c
new file mode 100644
index 000000000..0ef0343c6
--- /dev/null
+++ b/drivers/mmc/host/sdhci-bcm2835.c
@@ -0,0 +1,202 @@
+/*
+ * BCM2835 SDHCI
+ * Copyright (C) 2012 Stephen Warren
+ * Based on U-Boot's MMC driver for the BCM2835 by Oleksandr Tymoshenko & me
+ * Portions of the code there were obviously based on the Linux kernel at:
+ * git://github.com/raspberrypi/linux.git rpi-3.6.y
+ * commit f5b930b "Main bcm2708 linux port" signed-off-by Dom Cobley.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/delay.h>
+#include <linux/module.h>
+#include <linux/mmc/host.h>
+#include "sdhci-pltfm.h"
+
+/*
+ * 400KHz is max freq for card ID etc. Use that as min card clock. We need to
+ * know the min to enable static calculation of max BCM2835_SDHCI_WRITE_DELAY.
+ */
+#define MIN_FREQ 400000
+
+/*
+ * The Arasan has a bugette whereby it may lose the content of successive
+ * writes to registers that are within two SD-card clock cycles of each other
+ * (a clock domain crossing problem). It seems, however, that the data
+ * register does not have this problem, which is just as well - otherwise we'd
+ * have to nobble the DMA engine too.
+ *
+ * This should probably be dynamically calculated based on the actual card
+ * frequency. However, this is the longest we'll have to wait, and doesn't
+ * seem to slow access down too much, so the added complexity doesn't seem
+ * worth it for now.
+ *
+ * 1/MIN_FREQ is (max) time per tick of eMMC clock.
+ * 2/MIN_FREQ is time for two ticks.
+ * Multiply by 1000000 to get uS per two ticks.
+ * *1000000 for uSecs.
+ * +1 for hack rounding.
+ */
+#define BCM2835_SDHCI_WRITE_DELAY	(((2 * 1000000) / MIN_FREQ) + 1)
+
+struct bcm2835_sdhci {
+	u32 shadow;
+};
+
+static void bcm2835_sdhci_writel(struct sdhci_host *host, u32 val, int reg)
+{
+	writel(val, host->ioaddr + reg);
+
+	udelay(BCM2835_SDHCI_WRITE_DELAY);
+}
+
+static inline u32 bcm2835_sdhci_readl(struct sdhci_host *host, int reg)
+{
+	u32 val = readl(host->ioaddr + reg);
+
+	if (reg == SDHCI_CAPABILITIES)
+		val |= SDHCI_CAN_VDD_330;
+
+	return val;
+}
+
+static void bcm2835_sdhci_writew(struct sdhci_host *host, u16 val, int reg)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct bcm2835_sdhci *bcm2835_host = pltfm_host->priv;
+	u32 oldval = (reg == SDHCI_COMMAND) ? bcm2835_host->shadow :
+		bcm2835_sdhci_readl(host, reg & ~3);
+	u32 word_num = (reg >> 1) & 1;
+	u32 word_shift = word_num * 16;
+	u32 mask = 0xffff << word_shift;
+	u32 newval = (oldval & ~mask) | (val << word_shift);
+
+	if (reg == SDHCI_TRANSFER_MODE)
+		bcm2835_host->shadow = newval;
+	else
+		bcm2835_sdhci_writel(host, newval, reg & ~3);
+}
+
+static u16 bcm2835_sdhci_readw(struct sdhci_host *host, int reg)
+{
+	u32 val = bcm2835_sdhci_readl(host, (reg & ~3));
+	u32 word_num = (reg >> 1) & 1;
+	u32 word_shift = word_num * 16;
+	u32 word = (val >> word_shift) & 0xffff;
+
+	return word;
+}
+
+static void bcm2835_sdhci_writeb(struct sdhci_host *host, u8 val, int reg)
+{
+	u32 oldval = bcm2835_sdhci_readl(host, reg & ~3);
+	u32 byte_num = reg & 3;
+	u32 byte_shift = byte_num * 8;
+	u32 mask = 0xff << byte_shift;
+	u32 newval = (oldval & ~mask) | (val << byte_shift);
+
+	bcm2835_sdhci_writel(host, newval, reg & ~3);
+}
+
+static u8 bcm2835_sdhci_readb(struct sdhci_host *host, int reg)
+{
+	u32 val = bcm2835_sdhci_readl(host, (reg & ~3));
+	u32 byte_num = reg & 3;
+	u32 byte_shift = byte_num * 8;
+	u32 byte = (val >> byte_shift) & 0xff;
+
+	return byte;
+}
+
+static unsigned int bcm2835_sdhci_get_min_clock(struct sdhci_host *host)
+{
+	return MIN_FREQ;
+}
+
+static const struct sdhci_ops bcm2835_sdhci_ops = {
+	.write_l = bcm2835_sdhci_writel,
+	.write_w = bcm2835_sdhci_writew,
+	.write_b = bcm2835_sdhci_writeb,
+	.read_l = bcm2835_sdhci_readl,
+	.read_w = bcm2835_sdhci_readw,
+	.read_b = bcm2835_sdhci_readb,
+	.set_clock = sdhci_set_clock,
+	.get_max_clock = sdhci_pltfm_clk_get_max_clock,
+	.get_min_clock = bcm2835_sdhci_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 bcm2835_sdhci_pdata = {
+	.quirks = SDHCI_QUIRK_BROKEN_CARD_DETECTION |
+		  SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK,
+	.ops = &bcm2835_sdhci_ops,
+};
+
+static int bcm2835_sdhci_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct bcm2835_sdhci *bcm2835_host;
+	struct sdhci_pltfm_host *pltfm_host;
+	int ret;
+
+	host = sdhci_pltfm_init(pdev, &bcm2835_sdhci_pdata, 0);
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	bcm2835_host = devm_kzalloc(&pdev->dev, sizeof(*bcm2835_host),
+					GFP_KERNEL);
+	if (!bcm2835_host) {
+		dev_err(mmc_dev(host->mmc),
+			"failed to allocate bcm2835_sdhci\n");
+		return -ENOMEM;
+	}
+
+	pltfm_host = sdhci_priv(host);
+	pltfm_host->priv = bcm2835_host;
+
+	pltfm_host->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(pltfm_host->clk)) {
+		ret = PTR_ERR(pltfm_host->clk);
+		goto err;
+	}
+
+	return sdhci_add_host(host);
+
+err:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static const struct of_device_id bcm2835_sdhci_of_match[] = {
+	{ .compatible = "brcm,bcm2835-sdhci" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, bcm2835_sdhci_of_match);
+
+static struct platform_driver bcm2835_sdhci_driver = {
+	.driver = {
+		.name = "sdhci-bcm2835",
+		.of_match_table = bcm2835_sdhci_of_match,
+		.pm = SDHCI_PLTFM_PMOPS,
+	},
+	.probe = bcm2835_sdhci_probe,
+	.remove = sdhci_pltfm_unregister,
+};
+module_platform_driver(bcm2835_sdhci_driver);
+
+MODULE_DESCRIPTION("BCM2835 SDHCI driver");
+MODULE_AUTHOR("Stephen Warren");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-cns3xxx.c b/drivers/mmc/host/sdhci-cns3xxx.c
new file mode 100644
index 000000000..59f2923f8
--- /dev/null
+++ b/drivers/mmc/host/sdhci-cns3xxx.c
@@ -0,0 +1,115 @@
+/*
+ * SDHCI support for CNS3xxx SoC
+ *
+ * Copyright 2008 Cavium Networks
+ * Copyright 2010 MontaVista Software, LLC.
+ *
+ * Authors: Scott Shu
+ *	    Anton Vorontsov <avorontsov@mvista.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#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",
+		.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-dove.c b/drivers/mmc/host/sdhci-dove.c
new file mode 100644
index 000000000..407c21f15
--- /dev/null
+++ b/drivers/mmc/host/sdhci-dove.c
@@ -0,0 +1,132 @@
+/*
+ * 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
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+#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",
+		.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..82f512d87
--- /dev/null
+++ b/drivers/mmc/host/sdhci-esdhc-imx.c
@@ -0,0 +1,1155 @@
+/*
+ * Freescale eSDHC i.MX controller driver for the platform bus.
+ *
+ * derived from the OF-version.
+ *
+ * Copyright (c) 2010 Pengutronix e.K.
+ *   Author: Wolfram Sang <w.sang@pengutronix.de>
+ *
+ * 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; either version 2 of the License.
+ */
+
+#include <linux/io.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/clk.h>
+#include <linux/gpio.h>
+#include <linux/module.h>
+#include <linux/slab.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/of_gpio.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/platform_data/mmc-esdhc-imx.h>
+#include <linux/pm_runtime.h>
+#include "sdhci-pltfm.h"
+#include "sdhci-esdhc.h"
+
+#define	ESDHC_CTRL_D3CD			0x08
+/* 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_WTMK_LVL			0x44
+#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_FBCLK_SEL	(1 << 25)
+/* 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)
+
+#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		0x1
+
+/* 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 mis-match 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 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
+ * exeception. Bit1 of Vendor Spec registor is used to fix it.
+ */
+#define ESDHC_FLAG_MULTIBLK_NO_INT	BIT(1)
+/*
+ * The flag enables the workaround for ESDHC errata ENGcm07207 which
+ * affects i.MX25 and i.MX35.
+ */
+#define ESDHC_FLAG_ENGCM07207		BIT(2)
+/*
+ * 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)
+
+struct esdhc_soc_data {
+	u32 flags;
+};
+
+static struct esdhc_soc_data esdhc_imx25_data = {
+	.flags = ESDHC_FLAG_ENGCM07207,
+};
+
+static struct esdhc_soc_data esdhc_imx35_data = {
+	.flags = ESDHC_FLAG_ENGCM07207,
+};
+
+static struct esdhc_soc_data esdhc_imx51_data = {
+	.flags = 0,
+};
+
+static struct esdhc_soc_data esdhc_imx53_data = {
+	.flags = ESDHC_FLAG_MULTIBLK_NO_INT,
+};
+
+static struct esdhc_soc_data usdhc_imx6q_data = {
+	.flags = ESDHC_FLAG_USDHC | ESDHC_FLAG_MAN_TUNING,
+};
+
+static struct esdhc_soc_data usdhc_imx6sl_data = {
+	.flags = ESDHC_FLAG_USDHC | ESDHC_FLAG_STD_TUNING
+			| ESDHC_FLAG_HAVE_CAP1,
+};
+
+struct pltfm_imx_data {
+	u32 scratchpad;
+	struct pinctrl *pinctrl;
+	struct pinctrl_state *pins_default;
+	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;
+	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;
+};
+
+static struct platform_device_id imx_esdhc_devtype[] = {
+	{
+		.name = "sdhci-esdhc-imx25",
+		.driver_data = (kernel_ulong_t) &esdhc_imx25_data,
+	}, {
+		.name = "sdhci-esdhc-imx35",
+		.driver_data = (kernel_ulong_t) &esdhc_imx35_data,
+	}, {
+		.name = "sdhci-esdhc-imx51",
+		.driver_data = (kernel_ulong_t) &esdhc_imx51_data,
+	}, {
+		/* sentinel */
+	}
+};
+MODULE_DEVICE_TABLE(platform, imx_esdhc_devtype);
+
+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,imx6sl-usdhc", .data = &usdhc_imx6sl_data, },
+	{ .compatible = "fsl,imx6q-usdhc", .data = &usdhc_imx6q_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);
+}
+
+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 = pltfm_host->priv;
+	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
+		 * uirk 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;
+		}
+	}
+
+	if (unlikely(reg == SDHCI_MAX_CURRENT) && esdhc_is_usdhc(imx_data)) {
+		val = 0;
+		val |= 0xFF << SDHCI_MAX_CURRENT_330_SHIFT;
+		val |= 0xFF << SDHCI_MAX_CURRENT_300_SHIFT;
+		val |= 0xFF << SDHCI_MAX_CURRENT_180_SHIFT;
+	}
+
+	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 = pltfm_host->priv;
+	u32 data;
+
+	if (unlikely(reg == SDHCI_INT_ENABLE || reg == SDHCI_SIGNAL_ENABLE)) {
+		if (val & SDHCI_INT_CARD_INT) {
+			/*
+			 * Clear and then set D3CD bit to avoid missing the
+			 * card interrupt.  This is a 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 (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;
+			}
+	}
+
+	if (unlikely(reg == SDHCI_INT_ENABLE || reg == SDHCI_SIGNAL_ENABLE)) {
+		if (val & SDHCI_INT_ADMA_ERROR) {
+			val &= ~SDHCI_INT_ADMA_ERROR;
+			val |= ESDHC_INT_VENDOR_SPEC_DMA_ERR;
+		}
+	}
+
+	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 = pltfm_host->priv;
+	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_ACMD12_ERR);
+		}
+
+		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 = pltfm_host->priv;
+	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);
+		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;
+			else
+				new_val &= ~ESDHC_MIX_CTRL_SMPCLK_SEL;
+			writel(new_val , host->ioaddr + ESDHC_MIX_CTRL);
+		} else if (imx_data->socdata->flags & ESDHC_FLAG_STD_TUNING) {
+			u32 v = readl(host->ioaddr + SDHCI_ACMD12_ERR);
+			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;
+			}
+
+			if (val & SDHCI_CTRL_EXEC_TUNING) {
+				v |= ESDHC_MIX_CTRL_EXE_TUNE;
+				m |= ESDHC_MIX_CTRL_FBCLK_SEL;
+			} else {
+				v &= ~ESDHC_MIX_CTRL_EXE_TUNE;
+			}
+
+			writel(v, host->ioaddr + SDHCI_ACMD12_ERR);
+			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 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);
+		} 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 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 = pltfm_host->priv;
+	u32 new_val;
+	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 errata workaround.
+		 */
+		mask = 0xffff & ~(ESDHC_CTRL_BUSWIDTH_MASK | ESDHC_CTRL_D3CD);
+
+		esdhc_clrset_le(host, mask, new_val, reg);
+		return;
+	}
+	esdhc_clrset_le(host, 0xff, val, reg);
+
+	/*
+	 * 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.
+	 */
+	if ((reg == SDHCI_SOFTWARE_RESET) && (val & 1)) {
+		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;
+		}
+	}
+}
+
+static unsigned int esdhc_pltfm_get_max_clock(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = pltfm_host->priv;
+	struct esdhc_platform_data *boarddata = &imx_data->boarddata;
+
+	if (boarddata->f_max && (boarddata->f_max < pltfm_host->clock))
+		return boarddata->f_max;
+	else
+		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 = pltfm_host->priv;
+	unsigned int host_clock = pltfm_host->clock;
+	int pre_div = 2;
+	int div = 1;
+	u32 temp, val;
+
+	if (clock == 0) {
+		host->mmc->actual_clock = 0;
+
+		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);
+		}
+		return;
+	}
+
+	if (esdhc_is_usdhc(imx_data) && !imx_data->is_ddr)
+		pre_div = 1;
+
+	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);
+
+	while (host_clock / pre_div / 16 > clock && pre_div < 256)
+		pre_div *= 2;
+
+	while (host_clock / pre_div / div > clock && div < 16)
+		div++;
+
+	host->mmc->actual_clock = host_clock / pre_div / div;
+	dev_dbg(mmc_dev(host->mmc), "desired SD clock: %d, actual: %d\n",
+		clock, host->mmc->actual_clock);
+
+	if (imx_data->is_ddr)
+		pre_div >>= 2;
+	else
+		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);
+
+	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);
+	}
+
+	mdelay(1);
+}
+
+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 = pltfm_host->priv;
+	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 void esdhc_prepare_tuning(struct sdhci_host *host, u32 val)
+{
+	u32 reg;
+
+	/* FIXME: delay a bit for card to be ready for next tuning due to errors */
+	mdelay(1);
+
+	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),
+		"tunning 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;
+	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))
+			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)) {
+			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);
+	esdhc_post_tuning(host);
+
+	dev_dbg(mmc_dev(host->mmc), "tunning %s at 0x%x ret %d\n",
+		ret ? "failed" : "passed", avg, ret);
+
+	return ret;
+}
+
+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 = pltfm_host->priv;
+	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_default) ||
+		IS_ERR(imx_data->pins_100mhz) ||
+		IS_ERR(imx_data->pins_200mhz))
+		return -EINVAL;
+
+	switch (uhs) {
+	case MMC_TIMING_UHS_SDR50:
+		pinctrl = imx_data->pins_100mhz;
+		break;
+	case MMC_TIMING_UHS_SDR104:
+	case MMC_TIMING_MMC_HS200:
+		pinctrl = imx_data->pins_200mhz;
+		break;
+	default:
+		/* back to default state for other legacy timing */
+		pinctrl = imx_data->pins_default;
+	}
+
+	return pinctrl_select_state(imx_data->pinctrl, pinctrl);
+}
+
+static void esdhc_set_uhs_signaling(struct sdhci_host *host, unsigned timing)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct pltfm_imx_data *imx_data = pltfm_host->priv;
+	struct esdhc_platform_data *boarddata = &imx_data->boarddata;
+
+	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_HS200:
+		break;
+	case MMC_TIMING_UHS_DDR50:
+	case MMC_TIMING_MMC_DDR52:
+		writel(readl(host->ioaddr + ESDHC_MIX_CTRL) |
+				ESDHC_MIX_CTRL_DDREN,
+				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;
+	}
+
+	esdhc_change_pinstate(host, timing);
+}
+
+static void esdhc_reset(struct sdhci_host *host, u8 mask)
+{
+	sdhci_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 = pltfm_host->priv;
+
+	return esdhc_is_usdhc(imx_data) ? 1 << 28 : 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 = pltfm_host->priv;
+
+	/* use maximum timeout counter */
+	sdhci_writeb(host, esdhc_is_usdhc(imx_data) ? 0xF : 0xE,
+			SDHCI_TIMEOUT_CONTROL);
+}
+
+static struct sdhci_ops sdhci_esdhc_ops = {
+	.read_l = esdhc_readl_le,
+	.read_w = esdhc_readw_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,
+};
+
+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,
+};
+
+#ifdef CONFIG_OF
+static int
+sdhci_esdhc_imx_probe_dt(struct platform_device *pdev,
+			 struct sdhci_host *host,
+			 struct esdhc_platform_data *boarddata)
+{
+	struct device_node *np = pdev->dev.of_node;
+
+	if (!np)
+		return -ENODEV;
+
+	if (of_get_property(np, "non-removable", NULL))
+		boarddata->cd_type = ESDHC_CD_PERMANENT;
+
+	if (of_get_property(np, "fsl,cd-controller", NULL))
+		boarddata->cd_type = ESDHC_CD_CONTROLLER;
+
+	if (of_get_property(np, "fsl,wp-controller", NULL))
+		boarddata->wp_type = ESDHC_WP_CONTROLLER;
+
+	boarddata->cd_gpio = of_get_named_gpio(np, "cd-gpios", 0);
+	if (gpio_is_valid(boarddata->cd_gpio))
+		boarddata->cd_type = ESDHC_CD_GPIO;
+
+	boarddata->wp_gpio = of_get_named_gpio(np, "wp-gpios", 0);
+	if (gpio_is_valid(boarddata->wp_gpio))
+		boarddata->wp_type = ESDHC_WP_GPIO;
+
+	of_property_read_u32(np, "bus-width", &boarddata->max_bus_width);
+
+	of_property_read_u32(np, "max-frequency", &boarddata->f_max);
+
+	if (of_find_property(np, "no-1-8-v", NULL))
+		boarddata->support_vsel = false;
+	else
+		boarddata->support_vsel = true;
+
+	if (of_property_read_u32(np, "fsl,delay-line", &boarddata->delay_line))
+		boarddata->delay_line = 0;
+
+	mmc_of_parse_voltage(np, &host->ocr_mask);
+
+	return 0;
+}
+#else
+static inline int
+sdhci_esdhc_imx_probe_dt(struct platform_device *pdev,
+			 struct sdhci_host *host,
+			 struct esdhc_platform_data *boarddata)
+{
+	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 esdhc_platform_data *boarddata;
+	int err;
+	struct pltfm_imx_data *imx_data;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_esdhc_imx_pdata, 0);
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+
+	imx_data = devm_kzalloc(&pdev->dev, sizeof(*imx_data), GFP_KERNEL);
+	if (!imx_data) {
+		err = -ENOMEM;
+		goto free_sdhci;
+	}
+
+	imx_data->socdata = of_id ? of_id->data : (struct esdhc_soc_data *)
+						  pdev->id_entry->driver_data;
+	pltfm_host->priv = imx_data;
+
+	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);
+	clk_prepare_enable(imx_data->clk_per);
+	clk_prepare_enable(imx_data->clk_ipg);
+	clk_prepare_enable(imx_data->clk_ahb);
+
+	imx_data->pinctrl = devm_pinctrl_get(&pdev->dev);
+	if (IS_ERR(imx_data->pinctrl)) {
+		err = PTR_ERR(imx_data->pinctrl);
+		goto disable_clk;
+	}
+
+	imx_data->pins_default = pinctrl_lookup_state(imx_data->pinctrl,
+						PINCTRL_STATE_DEFAULT);
+	if (IS_ERR(imx_data->pins_default))
+		dev_warn(mmc_dev(host->mmc), "could not get default state\n");
+
+	host->quirks |= SDHCI_QUIRK_BROKEN_TIMEOUT_VAL;
+
+	if (imx_data->socdata->flags & ESDHC_FLAG_ENGCM07207)
+		/* Fix errata ENGcm07207 present on i.MX25 and i.MX35 */
+		host->quirks |= SDHCI_QUIRK_NO_MULTIBLOCK
+			| SDHCI_QUIRK_BROKEN_ADMA;
+
+	/*
+	 * The imx6q ROM code will change the default watermark level setting
+	 * to something insane.  Change it back here.
+	 */
+	if (esdhc_is_usdhc(imx_data)) {
+		writel(0x08100810, host->ioaddr + ESDHC_WTMK_LVL);
+		host->quirks2 |= SDHCI_QUIRK2_PRESET_VALUE_BROKEN;
+		host->mmc->caps |= MMC_CAP_1_8V_DDR;
+	}
+
+	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_STD_TUNING)
+		writel(readl(host->ioaddr + ESDHC_TUNING_CTRL) |
+			ESDHC_STD_TUNING_EN | ESDHC_TUNING_START_TAP,
+			host->ioaddr + ESDHC_TUNING_CTRL);
+
+	boarddata = &imx_data->boarddata;
+	if (sdhci_esdhc_imx_probe_dt(pdev, host, boarddata) < 0) {
+		if (!host->mmc->parent->platform_data) {
+			dev_err(mmc_dev(host->mmc), "no board data!\n");
+			err = -EINVAL;
+			goto disable_clk;
+		}
+		imx_data->boarddata = *((struct esdhc_platform_data *)
+					host->mmc->parent->platform_data);
+	}
+
+	/* card_detect */
+	if (boarddata->cd_type == ESDHC_CD_CONTROLLER)
+		host->quirks &= ~SDHCI_QUIRK_BROKEN_CARD_DETECTION;
+
+	switch (boarddata->max_bus_width) {
+	case 8:
+		host->mmc->caps |= MMC_CAP_8_BIT_DATA | MMC_CAP_4_BIT_DATA;
+		break;
+	case 4:
+		host->mmc->caps |= MMC_CAP_4_BIT_DATA;
+		break;
+	case 1:
+	default:
+		host->quirks |= SDHCI_QUIRK_FORCE_1_BIT_DATA;
+		break;
+	}
+
+	/* sdr50 and sdr104 needs work on 1.8v signal voltage */
+	if ((boarddata->support_vsel) && esdhc_is_usdhc(imx_data) &&
+	    !IS_ERR(imx_data->pins_default)) {
+		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);
+		if (IS_ERR(imx_data->pins_100mhz) ||
+				IS_ERR(imx_data->pins_200mhz)) {
+			dev_warn(mmc_dev(host->mmc),
+				"could not get ultra high speed state, work on normal mode\n");
+			/* fall back to not support uhs by specify no 1.8v quirk */
+			host->quirks2 |= SDHCI_QUIRK2_NO_1_8_V;
+		}
+	} else {
+		host->quirks2 |= SDHCI_QUIRK2_NO_1_8_V;
+	}
+
+	/* call to generic mmc_of_parse to support additional capabilities */
+	err = mmc_of_parse(host->mmc);
+	if (err)
+		goto disable_clk;
+
+	err = sdhci_add_host(host);
+	if (err)
+		goto disable_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_clk:
+	clk_disable_unprepare(imx_data->clk_per);
+	clk_disable_unprepare(imx_data->clk_ipg);
+	clk_disable_unprepare(imx_data->clk_ahb);
+free_sdhci:
+	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 = pltfm_host->priv;
+	int dead = (readl(host->ioaddr + SDHCI_INT_STATUS) == 0xffffffff);
+
+	pm_runtime_get_sync(&pdev->dev);
+	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);
+
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+
+#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 = pltfm_host->priv;
+	int ret;
+
+	ret = sdhci_runtime_suspend_host(host);
+
+	if (!sdhci_sdio_irq_enabled(host)) {
+		clk_disable_unprepare(imx_data->clk_per);
+		clk_disable_unprepare(imx_data->clk_ipg);
+	}
+	clk_disable_unprepare(imx_data->clk_ahb);
+
+	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 = pltfm_host->priv;
+
+	if (!sdhci_sdio_irq_enabled(host)) {
+		clk_prepare_enable(imx_data->clk_per);
+		clk_prepare_enable(imx_data->clk_ipg);
+	}
+	clk_prepare_enable(imx_data->clk_ahb);
+
+	return sdhci_runtime_resume_host(host);
+}
+#endif
+
+static const struct dev_pm_ops sdhci_esdhc_pmops = {
+	SET_SYSTEM_SLEEP_PM_OPS(sdhci_pltfm_suspend, sdhci_pltfm_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",
+		.of_match_table = imx_esdhc_dt_ids,
+		.pm	= &sdhci_esdhc_pmops,
+	},
+	.id_table	= imx_esdhc_devtype,
+	.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 <w.sang@pengutronix.de>");
+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..3497cfaf6
--- /dev/null
+++ b/drivers/mmc/host/sdhci-esdhc.h
@@ -0,0 +1,50 @@
+/*
+ * 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.
+ *   Author: Wolfram Sang <w.sang@pengutronix.de>
+ *
+ * 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; either version 2 of the License.
+ */
+
+#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_NO_BUSY_IRQ | \
+				SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK | \
+				SDHCI_QUIRK_PIO_NEEDS_DELAY)
+
+#define ESDHC_SYSTEM_CONTROL	0x2c
+#define ESDHC_CLOCK_MASK	0x0000fff0
+#define ESDHC_PREDIV_SHIFT	8
+#define ESDHC_DIVIDER_SHIFT	4
+#define ESDHC_CLOCK_PEREN	0x00000004
+#define ESDHC_CLOCK_HCKEN	0x00000002
+#define ESDHC_CLOCK_IPGEN	0x00000001
+
+/* pltfm-specific */
+#define ESDHC_HOST_CONTROL_LE	0x20
+
+/*
+ * P2020 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)
+
+/* OF-specific */
+#define ESDHC_DMA_SYSCTL	0x40c
+#define ESDHC_DMA_SNOOP		0x00000040
+
+#define ESDHC_HOST_CONTROL_RES	0x05
+
+#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..3b423b0ad
--- /dev/null
+++ b/drivers/mmc/host/sdhci-iproc.c
@@ -0,0 +1,241 @@
+/*
+ * 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/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;
+};
+
+struct sdhci_iproc_host {
+	const struct sdhci_iproc_data *data;
+	u32 shadow_cmd;
+	u32 shadow_blk;
+};
+
+#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)
+{
+	u32 val = sdhci_iproc_readl(host, (reg & ~3));
+	u16 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->shadow_blk != 0) {
+			sdhci_iproc_writel(host, iproc_host->shadow_blk,
+				SDHCI_BLOCK_SIZE);
+			iproc_host->shadow_blk = 0;
+		}
+		oldval = iproc_host->shadow_cmd;
+	} else if (reg == SDHCI_BLOCK_SIZE || reg == SDHCI_BLOCK_COUNT) {
+		/* 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;
+	} else if (reg == SDHCI_BLOCK_SIZE || reg == SDHCI_BLOCK_COUNT) {
+		/* Save the block info until the command is issued */
+		iproc_host->shadow_blk = newval;
+	} 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 const struct sdhci_ops sdhci_iproc_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_pltfm_clk_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_pltfm_data = {
+	.quirks = SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK,
+	.quirks2 = SDHCI_QUIRK2_ACMD23_BROKEN,
+	.ops = &sdhci_iproc_ops,
+};
+
+static const struct sdhci_iproc_data iproc_data = {
+	.pdata = &sdhci_iproc_pltfm_data,
+	.caps = 0x05E90000,
+	.caps1 = 0x00000064,
+};
+
+static const struct of_device_id sdhci_iproc_of_match[] = {
+	{ .compatible = "brcm,sdhci-iproc-cygnus", .data = &iproc_data },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, sdhci_iproc_of_match);
+
+static int sdhci_iproc_probe(struct platform_device *pdev)
+{
+	const struct of_device_id *match;
+	const struct sdhci_iproc_data *iproc_data;
+	struct sdhci_host *host;
+	struct sdhci_iproc_host *iproc_host;
+	struct sdhci_pltfm_host *pltfm_host;
+	int ret;
+
+	match = of_match_device(sdhci_iproc_of_match, &pdev->dev);
+	if (!match)
+		return -EINVAL;
+	iproc_data = match->data;
+
+	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;
+
+	mmc_of_parse(host->mmc);
+	sdhci_get_of_property(pdev);
+
+	/* Enable EMMC 1/8V DDR capable */
+	host->mmc->caps |= MMC_CAP_1_8V_DDR;
+
+	pltfm_host->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(pltfm_host->clk)) {
+		ret = PTR_ERR(pltfm_host->clk);
+		goto err;
+	}
+
+	if (iproc_host->data->pdata->quirks & SDHCI_QUIRK_MISSING_CAPS) {
+		host->caps = iproc_host->data->caps;
+		host->caps1 = iproc_host->data->caps1;
+	}
+
+	return sdhci_add_host(host);
+
+err:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static int sdhci_iproc_remove(struct platform_device *pdev)
+{
+	return sdhci_pltfm_unregister(pdev);
+}
+
+static struct platform_driver sdhci_iproc_driver = {
+	.driver = {
+		.name = "sdhci-iproc",
+		.of_match_table = sdhci_iproc_of_match,
+		.pm = SDHCI_PLTFM_PMOPS,
+	},
+	.probe = sdhci_iproc_probe,
+	.remove = sdhci_iproc_remove,
+};
+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-msm.c b/drivers/mmc/host/sdhci-msm.c
new file mode 100644
index 000000000..4a09f7608
--- /dev/null
+++ b/drivers/mmc/host/sdhci-msm.c
@@ -0,0 +1,593 @@
+/*
+ * drivers/mmc/host/sdhci-msm.c - Qualcomm SDHCI Platform driver
+ *
+ * Copyright (c) 2013-2014, The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/delay.h>
+#include <linux/mmc/mmc.h>
+#include <linux/slab.h>
+
+#include "sdhci-pltfm.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_HC_MODE		0x78
+#define HC_MODE_EN		0x1
+#define CORE_POWER		0x0
+#define CORE_SW_RST		BIT(7)
+
+#define MAX_PHASES		16
+#define CORE_DLL_LOCK		BIT(7)
+#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_DLL_CONFIG		0x100
+#define CORE_DLL_STATUS		0x108
+
+#define CORE_VENDOR_SPEC	0x10c
+#define CORE_CLK_PWRSAVE	BIT(1)
+
+#define CORE_VENDOR_SPEC_CAPABILITIES0	0x11c
+
+#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)
+
+struct sdhci_msm_host {
+	struct platform_device *pdev;
+	void __iomem *core_mem;	/* MSM SDCC mapped address */
+	struct clk *clk;	/* main SD/MMC bus clock */
+	struct clk *pclk;	/* SDHC peripheral bus clock */
+	struct clk *bus_clk;	/* SDHC bus voter clock */
+	struct mmc_host *mmc;
+	struct sdhci_pltfm_data sdhci_msm_pdata;
+};
+
+/* 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;
+
+	/* Poll for CK_OUT_EN bit.  max. poll time = 50us */
+	ck_out_en = !!(readl_relaxed(host->ioaddr + 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 + 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;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	config = readl_relaxed(host->ioaddr + CORE_DLL_CONFIG);
+	config &= ~(CORE_CDR_EN | CORE_CK_OUT_EN);
+	config |= (CORE_CDR_EXT_EN | CORE_DLL_EN);
+	writel_relaxed(config, host->ioaddr + 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 + CORE_DLL_CONFIG);
+	config &= ~CDR_SELEXT_MASK;
+	config |= grey_coded_phase_table[phase] << CDR_SELEXT_SHIFT;
+	writel_relaxed(config, host->ioaddr + CORE_DLL_CONFIG);
+
+	/* Set CK_OUT_EN bit of DLL_CONFIG register to 1. */
+	writel_relaxed((readl_relaxed(host->ioaddr + CORE_DLL_CONFIG)
+			| CORE_CK_OUT_EN), host->ioaddr + 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 + CORE_DLL_CONFIG);
+	config |= CORE_CDR_EN;
+	config &= ~CORE_CDR_EXT_EN;
+	writel_relaxed(config, host->ioaddr + 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 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;
+
+	/* 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 + CORE_DLL_CONFIG);
+	config &= ~CMUX_SHIFT_PHASE_MASK;
+	config |= mclk_freq << CMUX_SHIFT_PHASE_SHIFT;
+	writel_relaxed(config, host->ioaddr + 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;
+	int wait_cnt = 50;
+	unsigned long flags;
+
+	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.
+	 */
+	writel_relaxed((readl_relaxed(host->ioaddr + CORE_VENDOR_SPEC)
+			& ~CORE_CLK_PWRSAVE), host->ioaddr + CORE_VENDOR_SPEC);
+
+	/* Write 1 to DLL_RST bit of DLL_CONFIG register */
+	writel_relaxed((readl_relaxed(host->ioaddr + CORE_DLL_CONFIG)
+			| CORE_DLL_RST), host->ioaddr + CORE_DLL_CONFIG);
+
+	/* Write 1 to DLL_PDN bit of DLL_CONFIG register */
+	writel_relaxed((readl_relaxed(host->ioaddr + CORE_DLL_CONFIG)
+			| CORE_DLL_PDN), host->ioaddr + CORE_DLL_CONFIG);
+	msm_cm_dll_set_freq(host);
+
+	/* Write 0 to DLL_RST bit of DLL_CONFIG register */
+	writel_relaxed((readl_relaxed(host->ioaddr + CORE_DLL_CONFIG)
+			& ~CORE_DLL_RST), host->ioaddr + CORE_DLL_CONFIG);
+
+	/* Write 0 to DLL_PDN bit of DLL_CONFIG register */
+	writel_relaxed((readl_relaxed(host->ioaddr + CORE_DLL_CONFIG)
+			& ~CORE_DLL_PDN), host->ioaddr + CORE_DLL_CONFIG);
+
+	/* Set DLL_EN bit to 1. */
+	writel_relaxed((readl_relaxed(host->ioaddr + CORE_DLL_CONFIG)
+			| CORE_DLL_EN), host->ioaddr + CORE_DLL_CONFIG);
+
+	/* Set CK_OUT_EN bit to 1. */
+	writel_relaxed((readl_relaxed(host->ioaddr + CORE_DLL_CONFIG)
+			| CORE_CK_OUT_EN), host->ioaddr + CORE_DLL_CONFIG);
+
+	/* Wait until DLL_LOCK bit of DLL_STATUS register becomes '1' */
+	while (!(readl_relaxed(host->ioaddr + 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 int sdhci_msm_execute_tuning(struct sdhci_host *host, u32 opcode)
+{
+	int tuning_seq_cnt = 3;
+	u8 phase, tuned_phases[16], tuned_phase_cnt = 0;
+	int rc;
+	struct mmc_host *mmc = host->mmc;
+	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 <= 100 * 1000 * 1000 ||
+	    !((ios.timing == MMC_TIMING_MMC_HS200) ||
+	      (ios.timing == MMC_TIMING_UHS_SDR104)))
+		return 0;
+
+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);
+		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) {
+		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;
+		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 of_device_id sdhci_msm_dt_match[] = {
+	{ .compatible = "qcom,sdhci-msm-v4" },
+	{},
+};
+
+MODULE_DEVICE_TABLE(of, sdhci_msm_dt_match);
+
+static struct sdhci_ops sdhci_msm_ops = {
+	.platform_execute_tuning = sdhci_msm_execute_tuning,
+	.reset = sdhci_reset,
+	.set_clock = sdhci_set_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+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 resource *core_memres;
+	int ret;
+	u16 host_version, core_minor;
+	u32 core_version, caps;
+	u8 core_major;
+
+	msm_host = devm_kzalloc(&pdev->dev, sizeof(*msm_host), GFP_KERNEL);
+	if (!msm_host)
+		return -ENOMEM;
+
+	msm_host->sdhci_msm_pdata.ops = &sdhci_msm_ops;
+	host = sdhci_pltfm_init(pdev, &msm_host->sdhci_msm_pdata, 0);
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	pltfm_host->priv = msm_host;
+	msm_host->mmc = host->mmc;
+	msm_host->pdev = pdev;
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto pltfm_free;
+
+	sdhci_get_of_property(pdev);
+
+	/* 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 */
+	msm_host->pclk = devm_clk_get(&pdev->dev, "iface");
+	if (IS_ERR(msm_host->pclk)) {
+		ret = PTR_ERR(msm_host->pclk);
+		dev_err(&pdev->dev, "Perpheral clk setup failed (%d)\n", ret);
+		goto bus_clk_disable;
+	}
+
+	ret = clk_prepare_enable(msm_host->pclk);
+	if (ret)
+		goto bus_clk_disable;
+
+	/* Setup SDC MMC clock */
+	msm_host->clk = devm_clk_get(&pdev->dev, "core");
+	if (IS_ERR(msm_host->clk)) {
+		ret = PTR_ERR(msm_host->clk);
+		dev_err(&pdev->dev, "SDC MMC clk setup failed (%d)\n", ret);
+		goto pclk_disable;
+	}
+
+	ret = clk_prepare_enable(msm_host->clk);
+	if (ret)
+		goto pclk_disable;
+
+	core_memres = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+	msm_host->core_mem = devm_ioremap_resource(&pdev->dev, core_memres);
+
+	if (IS_ERR(msm_host->core_mem)) {
+		dev_err(&pdev->dev, "Failed to remap registers\n");
+		ret = PTR_ERR(msm_host->core_mem);
+		goto clk_disable;
+	}
+
+	/* Reset the core and Enable SDHC mode */
+	writel_relaxed(readl_relaxed(msm_host->core_mem + CORE_POWER) |
+		       CORE_SW_RST, msm_host->core_mem + CORE_POWER);
+
+	/* SW reset can take upto 10HCLK + 15MCLK cycles. (min 40us) */
+	usleep_range(1000, 5000);
+	if (readl(msm_host->core_mem + CORE_POWER) & CORE_SW_RST) {
+		dev_err(&pdev->dev, "Stuck in reset\n");
+		ret = -ETIMEDOUT;
+		goto clk_disable;
+	}
+
+	/* Set HC_MODE_EN bit in HC_MODE register */
+	writel_relaxed(HC_MODE_EN, (msm_host->core_mem + CORE_HC_MODE));
+
+	host->quirks |= SDHCI_QUIRK_BROKEN_CARD_DETECTION;
+	host->quirks |= SDHCI_QUIRK_SINGLE_POWER_WRITE;
+
+	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 = readl_relaxed(msm_host->core_mem + 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);
+
+	/*
+	 * 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) {
+		caps = readl_relaxed(host->ioaddr + SDHCI_CAPABILITIES);
+		caps |= SDHCI_CAN_VDD_300 | SDHCI_CAN_DO_8BIT;
+		writel_relaxed(caps, host->ioaddr +
+			       CORE_VENDOR_SPEC_CAPABILITIES0);
+	}
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto clk_disable;
+
+	return 0;
+
+clk_disable:
+	clk_disable_unprepare(msm_host->clk);
+pclk_disable:
+	clk_disable_unprepare(msm_host->pclk);
+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 = pltfm_host->priv;
+	int dead = (readl_relaxed(host->ioaddr + SDHCI_INT_STATUS) ==
+		    0xffffffff);
+
+	sdhci_remove_host(host, dead);
+	sdhci_pltfm_free(pdev);
+	clk_disable_unprepare(msm_host->clk);
+	clk_disable_unprepare(msm_host->pclk);
+	if (!IS_ERR(msm_host->bus_clk))
+		clk_disable_unprepare(msm_host->bus_clk);
+	return 0;
+}
+
+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,
+	},
+};
+
+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..6287d426c
--- /dev/null
+++ b/drivers/mmc/host/sdhci-of-arasan.c
@@ -0,0 +1,229 @@
+/*
+ * 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>
+ *
+ * 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; either version 2 of the License, or (at
+ * your option) any later version.
+ */
+
+#include <linux/module.h>
+#include "sdhci-pltfm.h"
+
+#define SDHCI_ARASAN_CLK_CTRL_OFFSET	0x2c
+
+#define CLK_CTRL_TIMEOUT_SHIFT		16
+#define CLK_CTRL_TIMEOUT_MASK		(0xf << CLK_CTRL_TIMEOUT_SHIFT)
+#define CLK_CTRL_TIMEOUT_MIN_EXP	13
+
+/**
+ * struct sdhci_arasan_data
+ * @clk_ahb:	Pointer to the AHB clock
+ */
+struct sdhci_arasan_data {
+	struct clk	*clk_ahb;
+};
+
+static unsigned int sdhci_arasan_get_timeout_clock(struct sdhci_host *host)
+{
+	u32 div;
+	unsigned long freq;
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	div = readl(host->ioaddr + SDHCI_ARASAN_CLK_CTRL_OFFSET);
+	div = (div & CLK_CTRL_TIMEOUT_MASK) >> CLK_CTRL_TIMEOUT_SHIFT;
+
+	freq = clk_get_rate(pltfm_host->clk);
+	freq /= 1 << (CLK_CTRL_TIMEOUT_MIN_EXP + div);
+
+	return freq;
+}
+
+static struct sdhci_ops sdhci_arasan_ops = {
+	.set_clock = sdhci_set_clock,
+	.get_max_clock = sdhci_pltfm_clk_get_max_clock,
+	.get_timeout_clock = sdhci_arasan_get_timeout_clock,
+	.set_bus_width = sdhci_set_bus_width,
+	.reset = sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+static struct sdhci_pltfm_data sdhci_arasan_pdata = {
+	.ops = &sdhci_arasan_ops,
+};
+
+#ifdef CONFIG_PM_SLEEP
+/**
+ * sdhci_arasan_suspend - Suspend method for the driver
+ * @dev:	Address of the device structure
+ * Returns 0 on success and error value on error
+ *
+ * Put the device in a low power state.
+ */
+static int sdhci_arasan_suspend(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = pltfm_host->priv;
+	int ret;
+
+	ret = sdhci_suspend_host(host);
+	if (ret)
+		return ret;
+
+	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
+ * Returns 0 on success and error value on error
+ *
+ * Resume operation after suspend
+ */
+static int sdhci_arasan_resume(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = pltfm_host->priv;
+	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");
+		clk_disable(sdhci_arasan->clk_ahb);
+		return ret;
+	}
+
+	return sdhci_resume_host(host);
+}
+#endif /* ! CONFIG_PM_SLEEP */
+
+static SIMPLE_DEV_PM_OPS(sdhci_arasan_dev_pm_ops, sdhci_arasan_suspend,
+			 sdhci_arasan_resume);
+
+static int sdhci_arasan_probe(struct platform_device *pdev)
+{
+	int ret;
+	struct clk *clk_xin;
+	struct sdhci_host *host;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_arasan_data *sdhci_arasan;
+
+	sdhci_arasan = devm_kzalloc(&pdev->dev, sizeof(*sdhci_arasan),
+			GFP_KERNEL);
+	if (!sdhci_arasan)
+		return -ENOMEM;
+
+	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");
+		return PTR_ERR(sdhci_arasan->clk_ahb);
+	}
+
+	clk_xin = devm_clk_get(&pdev->dev, "clk_xin");
+	if (IS_ERR(clk_xin)) {
+		dev_err(&pdev->dev, "clk_xin clock not found.\n");
+		return PTR_ERR(clk_xin);
+	}
+
+	ret = clk_prepare_enable(sdhci_arasan->clk_ahb);
+	if (ret) {
+		dev_err(&pdev->dev, "Unable to enable AHB clock.\n");
+		return ret;
+	}
+
+	ret = clk_prepare_enable(clk_xin);
+	if (ret) {
+		dev_err(&pdev->dev, "Unable to enable SD clock.\n");
+		goto clk_dis_ahb;
+	}
+
+	host = sdhci_pltfm_init(pdev, &sdhci_arasan_pdata, 0);
+	if (IS_ERR(host)) {
+		ret = PTR_ERR(host);
+		goto clk_disable_all;
+	}
+
+	sdhci_get_of_property(pdev);
+	pltfm_host = sdhci_priv(host);
+	pltfm_host->priv = sdhci_arasan;
+	pltfm_host->clk = clk_xin;
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret) {
+		dev_err(&pdev->dev, "parsing dt failed (%u)\n", ret);
+		goto clk_disable_all;
+	}
+
+	ret = sdhci_add_host(host);
+	if (ret)
+		goto err_pltfm_free;
+
+	return 0;
+
+err_pltfm_free:
+	sdhci_pltfm_free(pdev);
+clk_disable_all:
+	clk_disable_unprepare(clk_xin);
+clk_dis_ahb:
+	clk_disable_unprepare(sdhci_arasan->clk_ahb);
+
+	return ret;
+}
+
+static int sdhci_arasan_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	struct sdhci_arasan_data *sdhci_arasan = pltfm_host->priv;
+
+	clk_disable_unprepare(sdhci_arasan->clk_ahb);
+
+	return sdhci_pltfm_unregister(pdev);
+}
+
+static const struct of_device_id sdhci_arasan_of_match[] = {
+	{ .compatible = "arasan,sdhci-8.9a" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, sdhci_arasan_of_match);
+
+static struct platform_driver sdhci_arasan_driver = {
+	.driver = {
+		.name = "sdhci-arasan",
+		.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-esdhc.c b/drivers/mmc/host/sdhci-of-esdhc.c
new file mode 100644
index 000000000..22e9111b1
--- /dev/null
+++ b/drivers/mmc/host/sdhci-of-esdhc.c
@@ -0,0 +1,412 @@
+/*
+ * Freescale eSDHC controller driver.
+ *
+ * Copyright (c) 2007, 2010, 2012 Freescale Semiconductor, Inc.
+ * Copyright (c) 2009 MontaVista Software, Inc.
+ *
+ * Authors: Xiaobo Xie <X.Xie@freescale.com>
+ *	    Anton Vorontsov <avorontsov@ru.mvista.com>
+ *
+ * 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; either version 2 of the License, or (at
+ * your option) any later version.
+ */
+
+#include <linux/err.h>
+#include <linux/io.h>
+#include <linux/of.h>
+#include <linux/delay.h>
+#include <linux/module.h>
+#include <linux/mmc/host.h>
+#include "sdhci-pltfm.h"
+#include "sdhci-esdhc.h"
+
+#define VENDOR_V_22	0x12
+#define VENDOR_V_23	0x13
+static u32 esdhc_readl(struct sdhci_host *host, int reg)
+{
+	u32 ret;
+
+	ret = in_be32(host->ioaddr + reg);
+	/*
+	 * 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 one, but some of them can't support ADMA,
+	 * only these vendor version is greater than 2.2/0x12 support ADMA.
+	 * For FSL eSDHC, must aligned 4-byte, so use 0xFC to read the
+	 * the verdor version number, oxFE is SDHCI_HOST_VERSION.
+	 */
+	if ((reg == SDHCI_CAPABILITIES) && (ret & SDHCI_CAN_DO_ADMA1)) {
+		u32 tmp = in_be32(host->ioaddr + SDHCI_SLOT_INT_STATUS);
+		tmp = (tmp & SDHCI_VENDOR_VER_MASK) >> SDHCI_VENDOR_VER_SHIFT;
+		if (tmp > VENDOR_V_22)
+			ret |= SDHCI_CAN_DO_ADMA2;
+	}
+
+	return ret;
+}
+
+static u16 esdhc_readw(struct sdhci_host *host, int reg)
+{
+	u16 ret;
+	int base = reg & ~0x3;
+	int shift = (reg & 0x2) * 8;
+
+	if (unlikely(reg == SDHCI_HOST_VERSION))
+		ret = in_be32(host->ioaddr + base) & 0xffff;
+	else
+		ret = (in_be32(host->ioaddr + base) >> shift) & 0xffff;
+	return ret;
+}
+
+static u8 esdhc_readb(struct sdhci_host *host, int reg)
+{
+	int base = reg & ~0x3;
+	int shift = (reg & 0x3) * 8;
+	u8 ret = (in_be32(host->ioaddr + base) >> shift) & 0xff;
+
+	/*
+	 * "DMA select" locates at offset 0x28 in SD specification, but on
+	 * P5020 or P3041, it locates at 0x29.
+	 */
+	if (reg == SDHCI_HOST_CONTROL) {
+		u32 dma_bits;
+
+		dma_bits = in_be32(host->ioaddr + reg);
+		/* DMA select is 22,23 bits in Protocol Control Register */
+		dma_bits = (dma_bits >> 5) & SDHCI_CTRL_DMA_MASK;
+
+		/* fixup the result */
+		ret &= ~SDHCI_CTRL_DMA_MASK;
+		ret |= dma_bits;
+	}
+
+	return ret;
+}
+
+static void esdhc_writel(struct sdhci_host *host, u32 val, int reg)
+{
+	/*
+	 * Enable IRQSTATEN[BGESEN] is just to set IRQSTAT[BGE]
+	 * when SYSCTL[RSTD]) is set for some special operations.
+	 * No any impact other operation.
+	 */
+	if (reg == SDHCI_INT_ENABLE)
+		val |= SDHCI_INT_BLK_GAP;
+	sdhci_be32bs_writel(host, val, reg);
+}
+
+static void esdhc_writew(struct sdhci_host *host, u16 val, int reg)
+{
+	if (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.
+		 */
+		val &= ~SDHCI_MAKE_BLKSZ(0x7, 0);
+	}
+	sdhci_be32bs_writew(host, val, reg);
+}
+
+static void esdhc_writeb(struct sdhci_host *host, u8 val, int reg)
+{
+	/*
+	 * "DMA select" location is offset 0x28 in SD specification, but on
+	 * P5020 or P3041, it's located at 0x29.
+	 */
+	if (reg == SDHCI_HOST_CONTROL) {
+		u32 dma_bits;
+
+		/*
+		 * If host control register is not standard, exit
+		 * this function
+		 */
+		if (host->quirks2 & SDHCI_QUIRK2_BROKEN_HOST_CONTROL)
+			return;
+
+		/* DMA select is 22,23 bits in Protocol Control Register */
+		dma_bits = (val & SDHCI_CTRL_DMA_MASK) << 5;
+		clrsetbits_be32(host->ioaddr + reg , SDHCI_CTRL_DMA_MASK << 5,
+			dma_bits);
+		val &= ~SDHCI_CTRL_DMA_MASK;
+		val |= in_be32(host->ioaddr + reg) & SDHCI_CTRL_DMA_MASK;
+	}
+
+	/* Prevent SDHCI core from writing reserved bits (e.g. HISPD). */
+	if (reg == SDHCI_HOST_CONTROL)
+		val &= ~ESDHC_HOST_CONTROL_RES;
+	sdhci_be32bs_writeb(host, val, reg);
+}
+
+/*
+ * 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 esdhci_of_adma_workaround(struct sdhci_host *host, u32 intmask)
+{
+	u32 tmp;
+	bool applicable;
+	dma_addr_t dmastart;
+	dma_addr_t dmanow;
+
+	tmp = in_be32(host->ioaddr + SDHCI_SLOT_INT_STATUS);
+	tmp = (tmp & SDHCI_VENDOR_VER_MASK) >> SDHCI_VENDOR_VER_SHIFT;
+
+	applicable = (intmask & SDHCI_INT_DATA_END) &&
+		(intmask & SDHCI_INT_BLK_GAP) &&
+		(tmp == 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)
+{
+	setbits32(host->ioaddr + ESDHC_DMA_SYSCTL, ESDHC_DMA_SNOOP);
+	return 0;
+}
+
+static unsigned int esdhc_of_get_max_clock(struct sdhci_host *host)
+{
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	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);
+
+	return pltfm_host->clock / 256 / 16;
+}
+
+static void esdhc_of_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	int pre_div = 2;
+	int div = 1;
+	u32 temp;
+
+	host->mmc->actual_clock = 0;
+
+	if (clock == 0)
+		return;
+
+	/* Workaround to reduce the clock frequency for p1010 esdhc */
+	if (of_find_compatible_node(NULL, NULL, "fsl,p1010-esdhc")) {
+		if (clock > 20000000)
+			clock -= 5000000;
+		if (clock > 40000000)
+			clock -= 5000000;
+	}
+
+	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);
+
+	while (host->max_clk / pre_div / 16 > clock && pre_div < 256)
+		pre_div *= 2;
+
+	while (host->max_clk / pre_div / div > clock && div < 16)
+		div++;
+
+	dev_dbg(mmc_dev(host->mmc), "desired SD clock: %d, actual: %d\n",
+		clock, host->max_clk / pre_div / div);
+
+	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);
+	mdelay(1);
+}
+
+static void esdhc_of_platform_init(struct sdhci_host *host)
+{
+	u32 vvn;
+
+	vvn = in_be32(host->ioaddr + SDHCI_SLOT_INT_STATUS);
+	vvn = (vvn & SDHCI_VENDOR_VER_MASK) >> SDHCI_VENDOR_VER_SHIFT;
+	if (vvn == VENDOR_V_22)
+		host->quirks2 |= SDHCI_QUIRK2_HOST_NO_CMD23;
+
+	if (vvn > VENDOR_V_22)
+		host->quirks &= ~SDHCI_QUIRK_NO_BUSY_IRQ;
+}
+
+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;
+	}
+
+	clrsetbits_be32(host->ioaddr + SDHCI_HOST_CONTROL,
+			ESDHC_CTRL_BUSWIDTH_MASK, ctrl);
+}
+
+static void esdhc_reset(struct sdhci_host *host, u8 mask)
+{
+	sdhci_reset(host, mask);
+
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+}
+
+static const struct sdhci_ops sdhci_esdhc_ops = {
+	.read_l = esdhc_readl,
+	.read_w = esdhc_readw,
+	.read_b = esdhc_readb,
+	.write_l = esdhc_writel,
+	.write_w = esdhc_writew,
+	.write_b = esdhc_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,
+	.platform_init = esdhc_of_platform_init,
+	.adma_workaround = esdhci_of_adma_workaround,
+	.set_bus_width = esdhc_pltfm_set_bus_width,
+	.reset = esdhc_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+};
+
+#ifdef CONFIG_PM
+
+static u32 esdhc_proctl;
+static int esdhc_of_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+
+	esdhc_proctl = sdhci_be32bs_readl(host, SDHCI_HOST_CONTROL);
+
+	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_be32bs_writel(host, esdhc_proctl, SDHCI_HOST_CONTROL);
+	}
+
+	return ret;
+}
+
+static const struct dev_pm_ops esdhc_pmops = {
+	.suspend	= esdhc_of_suspend,
+	.resume		= esdhc_of_resume,
+};
+#define ESDHC_PMOPS (&esdhc_pmops)
+#else
+#define ESDHC_PMOPS NULL
+#endif
+
+static const struct sdhci_pltfm_data sdhci_esdhc_pdata = {
+	/*
+	 * card detection could be handled via GPIO
+	 * eSDHC cannot support End Attribute in NOP ADMA descriptor
+	 */
+	.quirks = ESDHC_DEFAULT_QUIRKS | SDHCI_QUIRK_BROKEN_CARD_DETECTION
+		| SDHCI_QUIRK_NO_CARD_NO_RESET
+		| SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.ops = &sdhci_esdhc_ops,
+};
+
+static int sdhci_esdhc_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct device_node *np;
+	int ret;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_esdhc_pdata, 0);
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	sdhci_get_of_property(pdev);
+
+	np = pdev->dev.of_node;
+	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;
+	}
+
+	/* 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 const struct of_device_id sdhci_esdhc_of_match[] = {
+	{ .compatible = "fsl,mpc8379-esdhc" },
+	{ .compatible = "fsl,mpc8536-esdhc" },
+	{ .compatible = "fsl,esdhc" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, sdhci_esdhc_of_match);
+
+static struct platform_driver sdhci_esdhc_driver = {
+	.driver = {
+		.name = "sdhci-esdhc",
+		.of_match_table = sdhci_esdhc_of_match,
+		.pm = ESDHC_PMOPS,
+	},
+	.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..4079a96ad
--- /dev/null
+++ b/drivers/mmc/host/sdhci-of-hlwd.c
@@ -0,0 +1,98 @@
+/*
+ * 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>
+ *
+ * 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; either version 2 of the License, or (at
+ * your option) any later version.
+ */
+
+#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",
+		.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-pci-data.c b/drivers/mmc/host/sdhci-pci-data.c
new file mode 100644
index 000000000..a61121776
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pci-data.c
@@ -0,0 +1,5 @@
+#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-o2micro.c b/drivers/mmc/host/sdhci-pci-o2micro.c
new file mode 100644
index 000000000..e2ec108db
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pci-o2micro.c
@@ -0,0 +1,395 @@
+/*
+ * Copyright (C) 2013 BayHub Technology Ltd.
+ *
+ * Authors: Peter Guo <peter.guo@bayhubtech.com>
+ *          Adam Lee <adam.lee@canonical.com>
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#include <linux/pci.h>
+
+#include "sdhci.h"
+#include "sdhci-pci.h"
+#include "sdhci-pci-o2micro.h"
+
+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 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);
+
+}
+
+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);
+}
+EXPORT_SYMBOL_GPL(sdhci_pci_o2_fujin2_pci_init);
+
+int sdhci_pci_o2_probe_slot(struct sdhci_pci_slot *slot)
+{
+	struct sdhci_pci_chip *chip;
+	struct sdhci_host *host;
+	u32 reg;
+
+	chip = slot->chip;
+	host = slot->host;
+	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;
+
+		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;
+}
+EXPORT_SYMBOL_GPL(sdhci_pci_o2_probe_slot);
+
+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 = 0x2c280000;
+
+				/* Set Base Clock to 208MZ */
+				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 |= 0x2c280000;
+
+			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);
+		/* 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;
+}
+EXPORT_SYMBOL_GPL(sdhci_pci_o2_probe);
+
+int sdhci_pci_o2_resume(struct sdhci_pci_chip *chip)
+{
+	sdhci_pci_o2_probe(chip);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(sdhci_pci_o2_resume);
diff --git a/drivers/mmc/host/sdhci-pci-o2micro.h b/drivers/mmc/host/sdhci-pci-o2micro.h
new file mode 100644
index 000000000..f7ffc908d
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pci-o2micro.h
@@ -0,0 +1,75 @@
+/*
+ * Copyright (C) 2013 BayHub Technology Ltd.
+ *
+ * Authors: Peter Guo <peter.guo@bayhubtech.com>
+ *          Adam Lee <adam.lee@canonical.com>
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#ifndef __SDHCI_PCI_O2MICRO_H
+#define __SDHCI_PCI_O2MICRO_H
+
+#include "sdhci-pci.h"
+
+/*
+ * O2Micro device 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
+
+/*
+ * 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_INF_MOD		0xF1
+#define O2_SD_MISC_CTRL4	0xFC
+#define O2_SD_TUNING_CTRL	0x300
+#define O2_SD_PLL_SETTING	0x304
+#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
+
+extern void sdhci_pci_o2_fujin2_pci_init(struct sdhci_pci_chip *chip);
+
+extern int sdhci_pci_o2_probe_slot(struct sdhci_pci_slot *slot);
+
+extern int sdhci_pci_o2_probe(struct sdhci_pci_chip *chip);
+
+extern int sdhci_pci_o2_resume(struct sdhci_pci_chip *chip);
+
+#endif /* __SDHCI_PCI_O2MICRO_H */
diff --git a/drivers/mmc/host/sdhci-pci.c b/drivers/mmc/host/sdhci-pci.c
new file mode 100644
index 000000000..7a3fc16d0
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pci.c
@@ -0,0 +1,1708 @@
+/*  linux/drivers/mmc/host/sdhci-pci.c - SDHCI on PCI bus interface
+ *
+ *  Copyright (C) 2005-2008 Pierre Ossman, All Rights Reserved.
+ *
+ * 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; either version 2 of the License, or (at
+ * your option) any later version.
+ *
+ * Thanks to the following companies for their support:
+ *
+ *     - JMicron (hardware and technical support)
+ */
+
+#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/scatterlist.h>
+#include <linux/io.h>
+#include <linux/gpio.h>
+#include <linux/pm_runtime.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/mmc/sdhci-pci-data.h>
+
+#include "sdhci.h"
+#include "sdhci-pci.h"
+#include "sdhci-pci-o2micro.h"
+
+/*****************************************************************************\
+ *                                                                           *
+ * 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 =
+		((0x21 << SDHCI_TIMEOUT_CLK_SHIFT)
+			& SDHCI_TIMEOUT_CLK_MASK) |
+
+		((0x21 << SDHCI_CLOCK_BASE_SHIFT)
+			& SDHCI_CLOCK_BASE_MASK) |
+
+		SDHCI_TIMEOUT_CLK_UNIT |
+		SDHCI_CAN_VDD_330 |
+		SDHCI_CAN_DO_HISPD |
+		SDHCI_CAN_DO_SDMA;
+	return 0;
+}
+
+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 0;
+}
+
+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,
+	.resume		= ricoh_mmc_resume,
+	.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 = gpio_request(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:
+	gpio_free(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);
+	if (gpio_is_valid(slot->cd_gpio))
+		gpio_free(slot->cd_gpio);
+}
+
+#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 |
+				  MMC_CAP2_HC_ERASE_SZ;
+	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,
+};
+
+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 byt_emmc_probe_slot(struct sdhci_pci_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_BUS_WIDTH_TEST |
+				 MMC_CAP_WAIT_WHILE_BUSY;
+	slot->host->mmc->caps2 |= MMC_CAP2_HC_ERASE_SZ;
+	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 */
+	return 0;
+}
+
+static int byt_sdio_probe_slot(struct sdhci_pci_slot *slot)
+{
+	slot->host->mmc->caps |= MMC_CAP_POWER_OFF_CARD | MMC_CAP_NONREMOVABLE |
+				 MMC_CAP_BUS_WIDTH_TEST |
+				 MMC_CAP_WAIT_WHILE_BUSY;
+	return 0;
+}
+
+static int byt_sd_probe_slot(struct sdhci_pci_slot *slot)
+{
+	slot->host->mmc->caps |= MMC_CAP_BUS_WIDTH_TEST |
+				 MMC_CAP_WAIT_WHILE_BUSY;
+	slot->cd_con_id = NULL;
+	slot->cd_idx = 0;
+	slot->cd_override_level = true;
+	return 0;
+}
+
+static const struct sdhci_pci_fixes sdhci_intel_byt_emmc = {
+	.allow_runtime_pm = true,
+	.probe_slot	= byt_emmc_probe_slot,
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.quirks2	= SDHCI_QUIRK2_PRESET_VALUE_BROKEN |
+			  SDHCI_QUIRK2_STOP_WITH_TC,
+};
+
+static const struct sdhci_pci_fixes sdhci_intel_byt_sdio = {
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.quirks2	= SDHCI_QUIRK2_HOST_OFF_CARD_ON |
+			SDHCI_QUIRK2_PRESET_VALUE_BROKEN,
+	.allow_runtime_pm = true,
+	.probe_slot	= byt_sdio_probe_slot,
+};
+
+static const struct sdhci_pci_fixes sdhci_intel_byt_sd = {
+	.quirks		= SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.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,
+};
+
+/* Define Host controllers for Intel Merrifield platform */
+#define INTEL_MRFL_EMMC_0	0
+#define INTEL_MRFL_EMMC_1	1
+
+static int intel_mrfl_mmc_probe_slot(struct sdhci_pci_slot *slot)
+{
+	if ((PCI_FUNC(slot->chip->pdev->devfn) != INTEL_MRFL_EMMC_0) &&
+	    (PCI_FUNC(slot->chip->pdev->devfn) != INTEL_MRFL_EMMC_1))
+		/* SD support is not ready yet */
+		return -ENODEV;
+
+	slot->host->mmc->caps |= MMC_CAP_8_BIT_DATA | MMC_CAP_NONREMOVABLE |
+				 MMC_CAP_1_8V_DDR;
+
+	return 0;
+}
+
+static const struct sdhci_pci_fixes sdhci_intel_mrfl_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_mrfl_mmc_probe_slot,
+};
+
+/* O2Micro extra registers */
+#define O2_SD_LOCK_WP		0xD3
+#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_INF_MOD		0xF1
+
+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;
+
+	ret = pci_write_config_byte(chip->pdev, 0xAE, scratch);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+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);
+}
+
+static int jmicron_suspend(struct sdhci_pci_chip *chip)
+{
+	int 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, 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 0;
+}
+
+static const struct sdhci_pci_fixes sdhci_o2 = {
+	.probe = sdhci_pci_o2_probe,
+	.quirks = SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC,
+	.probe_slot = sdhci_pci_o2_probe_slot,
+	.resume = sdhci_pci_o2_resume,
+};
+
+static const struct sdhci_pci_fixes sdhci_jmicron = {
+	.probe		= jmicron_probe,
+
+	.probe_slot	= jmicron_probe_slot,
+	.remove_slot	= jmicron_remove_slot,
+
+	.suspend	= jmicron_suspend,
+	.resume		= jmicron_resume,
+};
+
+/* 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,
+};
+
+static int amd_probe(struct sdhci_pci_chip *chip)
+{
+	struct pci_dev	*smbus_dev;
+
+	smbus_dev = pci_get_device(PCI_VENDOR_ID_AMD,
+			PCI_DEVICE_ID_AMD_HUDSON2_SMBUS, NULL);
+
+	if (smbus_dev && (smbus_dev->revision < 0x51)) {
+		chip->quirks2 |= SDHCI_QUIRK2_CLEAR_TRANSFERMODE_REG_BEFORE_CMD;
+		chip->quirks2 |= SDHCI_QUIRK2_BROKEN_HS200;
+	}
+
+	return 0;
+}
+
+static const struct sdhci_pci_fixes sdhci_amd = {
+	.probe		= amd_probe,
+};
+
+static const struct pci_device_id pci_ids[] = {
+	{
+		.vendor		= PCI_VENDOR_ID_RICOH,
+		.device		= PCI_DEVICE_ID_RICOH_R5C822,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_ricoh,
+	},
+
+	{
+		.vendor         = PCI_VENDOR_ID_RICOH,
+		.device         = 0x843,
+		.subvendor      = PCI_ANY_ID,
+		.subdevice      = PCI_ANY_ID,
+		.driver_data    = (kernel_ulong_t)&sdhci_ricoh_mmc,
+	},
+
+	{
+		.vendor         = PCI_VENDOR_ID_RICOH,
+		.device         = 0xe822,
+		.subvendor      = PCI_ANY_ID,
+		.subdevice      = PCI_ANY_ID,
+		.driver_data    = (kernel_ulong_t)&sdhci_ricoh_mmc,
+	},
+
+	{
+		.vendor         = PCI_VENDOR_ID_RICOH,
+		.device         = 0xe823,
+		.subvendor      = PCI_ANY_ID,
+		.subdevice      = PCI_ANY_ID,
+		.driver_data    = (kernel_ulong_t)&sdhci_ricoh_mmc,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_ENE,
+		.device		= PCI_DEVICE_ID_ENE_CB712_SD,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_ene_712,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_ENE,
+		.device		= PCI_DEVICE_ID_ENE_CB712_SD_2,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_ene_712,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_ENE,
+		.device		= PCI_DEVICE_ID_ENE_CB714_SD,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_ene_714,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_ENE,
+		.device		= PCI_DEVICE_ID_ENE_CB714_SD_2,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_ene_714,
+	},
+
+	{
+		.vendor         = PCI_VENDOR_ID_MARVELL,
+		.device         = PCI_DEVICE_ID_MARVELL_88ALP01_SD,
+		.subvendor      = PCI_ANY_ID,
+		.subdevice      = PCI_ANY_ID,
+		.driver_data    = (kernel_ulong_t)&sdhci_cafe,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_JMICRON,
+		.device		= PCI_DEVICE_ID_JMICRON_JMB38X_SD,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_jmicron,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_JMICRON,
+		.device		= PCI_DEVICE_ID_JMICRON_JMB38X_MMC,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_jmicron,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_JMICRON,
+		.device		= PCI_DEVICE_ID_JMICRON_JMB388_SD,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_jmicron,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_JMICRON,
+		.device		= PCI_DEVICE_ID_JMICRON_JMB388_ESD,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_jmicron,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_SYSKONNECT,
+		.device		= 0x8000,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_syskt,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_VIA,
+		.device		= 0x95d0,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_via,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_REALTEK,
+		.device		= 0x5250,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_rtsx,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_QRK_SD,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_qrk,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_MRST_SD0,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mrst_hc0,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_MRST_SD1,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mrst_hc1_hc2,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_MRST_SD2,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mrst_hc1_hc2,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_MFD_SD,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mfd_sd,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_MFD_SDIO1,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mfd_sdio,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_MFD_SDIO2,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mfd_sdio,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_MFD_EMMC0,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mfd_emmc,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_MFD_EMMC1,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mfd_emmc,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_PCH_SDIO0,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_pch_sdio,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_PCH_SDIO1,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_pch_sdio,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_BYT_EMMC,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_byt_emmc,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_BYT_SDIO,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_byt_sdio,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_BYT_SD,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_byt_sd,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_BYT_EMMC2,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_byt_emmc,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_BSW_EMMC,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_byt_emmc,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_BSW_SDIO,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_byt_sdio,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_BSW_SD,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_byt_sd,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_CLV_SDIO0,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mfd_sd,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_CLV_SDIO1,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mfd_sdio,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_CLV_SDIO2,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mfd_sdio,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_CLV_EMMC0,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mfd_emmc,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_CLV_EMMC1,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mfd_emmc,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_MRFL_MMC,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_mrfl_mmc,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_SPT_EMMC,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_byt_emmc,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_SPT_SDIO,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_byt_sdio,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_INTEL,
+		.device		= PCI_DEVICE_ID_INTEL_SPT_SD,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_intel_byt_sd,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_O2,
+		.device		= PCI_DEVICE_ID_O2_8120,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_o2,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_O2,
+		.device		= PCI_DEVICE_ID_O2_8220,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_o2,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_O2,
+		.device		= PCI_DEVICE_ID_O2_8221,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_o2,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_O2,
+		.device		= PCI_DEVICE_ID_O2_8320,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_o2,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_O2,
+		.device		= PCI_DEVICE_ID_O2_8321,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_o2,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_O2,
+		.device		= PCI_DEVICE_ID_O2_FUJIN2,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_o2,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_O2,
+		.device		= PCI_DEVICE_ID_O2_SDS0,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_o2,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_O2,
+		.device		= PCI_DEVICE_ID_O2_SDS1,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_o2,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_O2,
+		.device		= PCI_DEVICE_ID_O2_SEABIRD0,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_o2,
+	},
+
+	{
+		.vendor		= PCI_VENDOR_ID_O2,
+		.device		= PCI_DEVICE_ID_O2_SEABIRD1,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_o2,
+	},
+	{
+		.vendor		= PCI_VENDOR_ID_AMD,
+		.device		= PCI_ANY_ID,
+		.class		= PCI_CLASS_SYSTEM_SDHCI << 8,
+		.class_mask	= 0xFFFF00,
+		.subvendor	= PCI_ANY_ID,
+		.subdevice	= PCI_ANY_ID,
+		.driver_data	= (kernel_ulong_t)&sdhci_amd,
+	},
+	{	/* Generic SD host controller */
+		PCI_DEVICE_CLASS((PCI_CLASS_SYSTEM_SDHCI << 8), 0xFFFF00)
+	},
+
+	{ /* end: all zeroes */ },
+};
+
+MODULE_DEVICE_TABLE(pci, pci_ids);
+
+/*****************************************************************************\
+ *                                                                           *
+ * SDHCI core callbacks                                                      *
+ *                                                                           *
+\*****************************************************************************/
+
+static int sdhci_pci_enable_dma(struct sdhci_host *host)
+{
+	struct sdhci_pci_slot *slot;
+	struct pci_dev *pdev;
+	int ret = -1;
+
+	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");
+	}
+
+	if (host->flags & SDHCI_USE_64_BIT_DMA) {
+		if (host->quirks2 & SDHCI_QUIRK2_BROKEN_64_BIT_DMA) {
+			host->flags &= ~SDHCI_USE_64_BIT_DMA;
+		} else {
+			ret = pci_set_dma_mask(pdev, DMA_BIT_MASK(64));
+			if (ret)
+				dev_warn(&pdev->dev, "Failed to set 64-bit DMA mask\n");
+		}
+	}
+	if (ret)
+		ret = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
+	if (ret)
+		return ret;
+
+	pci_set_master(pdev);
+
+	return 0;
+}
+
+static void sdhci_pci_set_bus_width(struct sdhci_host *host, int width)
+{
+	u8 ctrl;
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+
+	switch (width) {
+	case MMC_BUS_WIDTH_8:
+		ctrl |= SDHCI_CTRL_8BITBUS;
+		ctrl &= ~SDHCI_CTRL_4BITBUS;
+		break;
+	case MMC_BUS_WIDTH_4:
+		ctrl |= SDHCI_CTRL_4BITBUS;
+		ctrl &= ~SDHCI_CTRL_8BITBUS;
+		break;
+	default:
+		ctrl &= ~(SDHCI_CTRL_8BITBUS | SDHCI_CTRL_4BITBUS);
+		break;
+	}
+
+	sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+}
+
+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_pci_set_bus_width,
+	.reset		= sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+	.hw_reset		= sdhci_pci_hw_reset,
+};
+
+/*****************************************************************************\
+ *                                                                           *
+ * Suspend/resume                                                            *
+ *                                                                           *
+\*****************************************************************************/
+
+#ifdef CONFIG_PM
+
+static int sdhci_pci_suspend(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct sdhci_pci_chip *chip;
+	struct sdhci_pci_slot *slot;
+	mmc_pm_flag_t slot_pm_flags;
+	mmc_pm_flag_t pm_flags = 0;
+	int i, ret;
+
+	chip = pci_get_drvdata(pdev);
+	if (!chip)
+		return 0;
+
+	for (i = 0; i < chip->num_slots; i++) {
+		slot = chip->slots[i];
+		if (!slot)
+			continue;
+
+		ret = sdhci_suspend_host(slot->host);
+
+		if (ret)
+			goto err_pci_suspend;
+
+		slot_pm_flags = slot->host->mmc->pm_flags;
+		if (slot_pm_flags & MMC_PM_WAKE_SDIO_IRQ)
+			sdhci_enable_irq_wakeups(slot->host);
+
+		pm_flags |= slot_pm_flags;
+	}
+
+	if (chip->fixes && chip->fixes->suspend) {
+		ret = chip->fixes->suspend(chip);
+		if (ret)
+			goto err_pci_suspend;
+	}
+
+	if (pm_flags & MMC_PM_KEEP_POWER) {
+		if (pm_flags & MMC_PM_WAKE_SDIO_IRQ)
+			device_init_wakeup(dev, true);
+		else
+			device_init_wakeup(dev, false);
+	} else
+		device_init_wakeup(dev, false);
+
+	return 0;
+
+err_pci_suspend:
+	while (--i >= 0)
+		sdhci_resume_host(chip->slots[i]->host);
+	return ret;
+}
+
+static int sdhci_pci_resume(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct sdhci_pci_chip *chip;
+	struct sdhci_pci_slot *slot;
+	int i, ret;
+
+	chip = pci_get_drvdata(pdev);
+	if (!chip)
+		return 0;
+
+	if (chip->fixes && chip->fixes->resume) {
+		ret = chip->fixes->resume(chip);
+		if (ret)
+			return 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;
+	}
+
+	return 0;
+}
+
+static int sdhci_pci_runtime_suspend(struct device *dev)
+{
+	struct pci_dev *pdev = container_of(dev, struct pci_dev, dev);
+	struct sdhci_pci_chip *chip;
+	struct sdhci_pci_slot *slot;
+	int i, ret;
+
+	chip = pci_get_drvdata(pdev);
+	if (!chip)
+		return 0;
+
+	for (i = 0; i < chip->num_slots; i++) {
+		slot = chip->slots[i];
+		if (!slot)
+			continue;
+
+		ret = sdhci_runtime_suspend_host(slot->host);
+
+		if (ret)
+			goto err_pci_runtime_suspend;
+	}
+
+	if (chip->fixes && chip->fixes->suspend) {
+		ret = chip->fixes->suspend(chip);
+		if (ret)
+			goto err_pci_runtime_suspend;
+	}
+
+	return 0;
+
+err_pci_runtime_suspend:
+	while (--i >= 0)
+		sdhci_runtime_resume_host(chip->slots[i]->host);
+	return ret;
+}
+
+static int sdhci_pci_runtime_resume(struct device *dev)
+{
+	struct pci_dev *pdev = container_of(dev, struct pci_dev, dev);
+	struct sdhci_pci_chip *chip;
+	struct sdhci_pci_slot *slot;
+	int i, ret;
+
+	chip = pci_get_drvdata(pdev);
+	if (!chip)
+		return 0;
+
+	if (chip->fixes && chip->fixes->resume) {
+		ret = chip->fixes->resume(chip);
+		if (ret)
+			return ret;
+	}
+
+	for (i = 0; i < chip->num_slots; i++) {
+		slot = chip->slots[i];
+		if (!slot)
+			continue;
+
+		ret = sdhci_runtime_resume_host(slot->host);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+#else /* CONFIG_PM */
+
+#define sdhci_pci_suspend NULL
+#define sdhci_pci_resume NULL
+
+#endif /* CONFIG_PM */
+
+static const struct dev_pm_ops sdhci_pci_pm_ops = {
+	.suspend = sdhci_pci_suspend,
+	.resume = 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;
+
+	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(struct sdhci_pci_slot));
+	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->pci_bar = bar;
+	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 = &sdhci_pci_ops;
+	host->quirks = chip->quirks;
+	host->quirks2 = chip->quirks2;
+
+	host->irq = pdev->irq;
+
+	ret = pci_request_region(pdev, bar, mmc_hostname(host->mmc));
+	if (ret) {
+		dev_err(&pdev->dev, "cannot request region\n");
+		goto cleanup;
+	}
+
+	host->ioaddr = pci_ioremap_bar(pdev, bar);
+	if (!host->ioaddr) {
+		dev_err(&pdev->dev, "failed to remap registers\n");
+		ret = -ENOMEM;
+		goto release;
+	}
+
+	if (chip->fixes && chip->fixes->probe_slot) {
+		ret = chip->fixes->probe_slot(slot);
+		if (ret)
+			goto unmap;
+	}
+
+	if (gpio_is_valid(slot->rst_n_gpio)) {
+		if (!gpio_request(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 | MMC_PM_WAKE_SDIO_IRQ;
+	host->mmc->slotno = slotno;
+	host->mmc->caps2 |= MMC_CAP2_NO_PRESCAN_POWERUP;
+
+	if (slot->cd_idx >= 0 &&
+	    mmc_gpiod_request_cd(host->mmc, slot->cd_con_id, slot->cd_idx,
+				 slot->cd_override_level, 0, NULL)) {
+		dev_warn(&pdev->dev, "failed to setup card detect gpio\n");
+		slot->cd_idx = -1;
+	}
+
+	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 (gpio_is_valid(slot->rst_n_gpio))
+		gpio_free(slot->rst_n_gpio);
+
+	if (chip->fixes && chip->fixes->remove_slot)
+		chip->fixes->remove_slot(slot, 0);
+
+unmap:
+	iounmap(host->ioaddr);
+
+release:
+	pci_release_region(pdev, bar);
+
+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 (gpio_is_valid(slot->rst_n_gpio))
+		gpio_free(slot->rst_n_gpio);
+
+	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);
+
+	pci_release_region(slot->chip->pdev, slot->pci_bar);
+
+	sdhci_free_host(slot->host);
+}
+
+static void sdhci_pci_runtime_pm_allow(struct device *dev)
+{
+	pm_runtime_put_noidle(dev);
+	pm_runtime_allow(dev);
+	pm_runtime_set_autosuspend_delay(dev, 50);
+	pm_runtime_use_autosuspend(dev);
+	pm_suspend_ignore_children(dev, 1);
+}
+
+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);
+	if (slots == 0)
+		return -ENODEV;
+
+	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 = pci_enable_device(pdev);
+	if (ret)
+		return ret;
+
+	chip = kzalloc(sizeof(struct sdhci_pci_chip), GFP_KERNEL);
+	if (!chip) {
+		ret = -ENOMEM;
+		goto err;
+	}
+
+	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;
+
+	pci_set_drvdata(pdev, chip);
+
+	if (chip->fixes && chip->fixes->probe) {
+		ret = chip->fixes->probe(chip);
+		if (ret)
+			goto free;
+	}
+
+	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]);
+			ret = PTR_ERR(slot);
+			goto free;
+		}
+
+		chip->slots[i] = slot;
+	}
+
+	if (chip->allow_runtime_pm)
+		sdhci_pci_runtime_pm_allow(&pdev->dev);
+
+	return 0;
+
+free:
+	pci_set_drvdata(pdev, NULL);
+	kfree(chip);
+
+err:
+	pci_disable_device(pdev);
+	return ret;
+}
+
+static void sdhci_pci_remove(struct pci_dev *pdev)
+{
+	int i;
+	struct sdhci_pci_chip *chip;
+
+	chip = pci_get_drvdata(pdev);
+
+	if (chip) {
+		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]);
+
+		pci_set_drvdata(pdev, NULL);
+		kfree(chip);
+	}
+
+	pci_disable_device(pdev);
+}
+
+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.h b/drivers/mmc/host/sdhci-pci.h
new file mode 100644
index 000000000..1ec684d06
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pci.h
@@ -0,0 +1,89 @@
+#ifndef __SDHCI_PCI_H
+#define __SDHCI_PCI_H
+
+/*
+ * PCI device IDs
+ */
+
+#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_MRFL_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
+
+/*
+ * 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 *);
+	void			(*remove_slot) (struct sdhci_pci_slot *, int);
+
+	int			(*suspend) (struct sdhci_pci_chip *);
+	int			(*resume) (struct sdhci_pci_chip *);
+};
+
+struct sdhci_pci_slot {
+	struct sdhci_pci_chip	*chip;
+	struct sdhci_host	*host;
+	struct sdhci_pci_data	*data;
+
+	int			pci_bar;
+	int			rst_n_gpio;
+	int			cd_gpio;
+	int			cd_irq;
+
+	char			*cd_con_id;
+	int			cd_idx;
+	bool			cd_override_level;
+
+	void (*hw_reset)(struct sdhci_host *host);
+};
+
+struct sdhci_pci_chip {
+	struct pci_dev		*pdev;
+
+	unsigned int		quirks;
+	unsigned int		quirks2;
+	bool			allow_runtime_pm;
+	const struct sdhci_pci_fixes *fixes;
+
+	int			num_slots;	/* Slots on controller */
+	struct sdhci_pci_slot	*slots[MAX_SLOTS]; /* Pointers to host slots */
+};
+
+#endif /* __SDHCI_PCI_H */
diff --git a/drivers/mmc/host/sdhci-pltfm.c b/drivers/mmc/host/sdhci-pltfm.c
new file mode 100644
index 000000000..a207f5aaf
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pltfm.c
@@ -0,0 +1,276 @@
+/*
+ * 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>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+/* Supports:
+ * SDHCI platform devices
+ *
+ * Inspired by sdhci-pci.c, by Pierre Ossman
+ */
+
+#include <linux/err.h>
+#include <linux/module.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,
+};
+
+#ifdef CONFIG_OF
+static bool sdhci_of_wp_inverted(struct device_node *np)
+{
+	if (of_get_property(np, "sdhci,wp-inverted", NULL) ||
+	    of_get_property(np, "wp-inverted", NULL))
+		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 */
+}
+
+void sdhci_get_of_property(struct platform_device *pdev)
+{
+	struct device_node *np = pdev->dev.of_node;
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	const __be32 *clk;
+	u32 bus_width;
+	int size;
+
+	if (of_get_property(np, "sdhci,auto-cmd12", NULL))
+		host->quirks |= SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12;
+
+	if (of_get_property(np, "sdhci,1-bit-only", NULL) ||
+	    (of_property_read_u32(np, "bus-width", &bus_width) == 0 &&
+	    bus_width == 1))
+		host->quirks |= SDHCI_QUIRK_FORCE_1_BIT_DATA;
+
+	if (sdhci_of_wp_inverted(np))
+		host->quirks |= SDHCI_QUIRK_INVERTED_WRITE_PROTECT;
+
+	if (of_get_property(np, "broken-cd", NULL))
+		host->quirks |= SDHCI_QUIRK_BROKEN_CARD_DETECTION;
+
+	if (of_get_property(np, "no-1-8-v", NULL))
+		host->quirks2 |= SDHCI_QUIRK2_NO_1_8_V;
+
+	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;
+
+	clk = of_get_property(np, "clock-frequency", &size);
+	if (clk && size == sizeof(*clk) && *clk)
+		pltfm_host->clock = be32_to_cpup(clk);
+
+	if (of_find_property(np, "keep-power-in-suspend", NULL))
+		host->mmc->pm_caps |= MMC_PM_KEEP_POWER;
+
+	if (of_find_property(np, "enable-sdio-wakeup", NULL))
+		host->mmc->pm_caps |= MMC_PM_WAKE_SDIO_IRQ;
+}
+#else
+void sdhci_get_of_property(struct platform_device *pdev) {}
+#endif /* CONFIG_OF */
+EXPORT_SYMBOL_GPL(sdhci_get_of_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;
+	struct resource *iomem;
+	int ret;
+
+	iomem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!iomem) {
+		ret = -ENOMEM;
+		goto err;
+	}
+
+	if (resource_size(iomem) < 0x100)
+		dev_err(&pdev->dev, "Invalid iomem size!\n");
+
+	host = sdhci_alloc_host(&pdev->dev,
+		sizeof(struct sdhci_pltfm_host) + priv_size);
+
+	if (IS_ERR(host)) {
+		ret = PTR_ERR(host);
+		goto err;
+	}
+
+	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;
+	}
+
+	host->irq = platform_get_irq(pdev, 0);
+
+	if (!request_mem_region(iomem->start, resource_size(iomem),
+		mmc_hostname(host->mmc))) {
+		dev_err(&pdev->dev, "cannot request region\n");
+		ret = -EBUSY;
+		goto err_request;
+	}
+
+	host->ioaddr = ioremap(iomem->start, resource_size(iomem));
+	if (!host->ioaddr) {
+		dev_err(&pdev->dev, "failed to remap registers\n");
+		ret = -ENOMEM;
+		goto err_remap;
+	}
+
+	/*
+	 * Some platforms need to probe the controller to be able to
+	 * determine which caps should be used.
+	 */
+	if (host->ops && host->ops->platform_init)
+		host->ops->platform_init(host);
+
+	platform_set_drvdata(pdev, host);
+
+	return host;
+
+err_remap:
+	release_mem_region(iomem->start, resource_size(iomem));
+err_request:
+	sdhci_free_host(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);
+	struct resource *iomem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+
+	iounmap(host->ioaddr);
+	release_mem_region(iomem->start, resource_size(iomem));
+	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_of_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
+int sdhci_pltfm_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+
+	return sdhci_suspend_host(host);
+}
+EXPORT_SYMBOL_GPL(sdhci_pltfm_suspend);
+
+int sdhci_pltfm_resume(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+
+	return sdhci_resume_host(host);
+}
+EXPORT_SYMBOL_GPL(sdhci_pltfm_resume);
+
+const struct dev_pm_ops sdhci_pltfm_pmops = {
+	.suspend	= sdhci_pltfm_suspend,
+	.resume		= sdhci_pltfm_resume,
+};
+EXPORT_SYMBOL_GPL(sdhci_pltfm_pmops);
+#endif	/* CONFIG_PM */
+
+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..04bc2481e
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pltfm.h
@@ -0,0 +1,122 @@
+/*
+ * Copyright 2010 MontaVista Software, LLC.
+ *
+ * Author: Anton Vorontsov <avorontsov@ru.mvista.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#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;
+	void *priv; /* to handle quirks across io-accessor calls */
+
+	/* migrate from sdhci_of_host */
+	unsigned int clock;
+	u16 xfer_mode_shadow;
+
+	unsigned long private[0] ____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 */
+
+extern void sdhci_get_of_property(struct platform_device *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 (void *)host->private;
+}
+
+#ifdef CONFIG_PM
+extern int sdhci_pltfm_suspend(struct device *dev);
+extern int sdhci_pltfm_resume(struct device *dev);
+extern const struct dev_pm_ops sdhci_pltfm_pmops;
+#define SDHCI_PLTFM_PMOPS (&sdhci_pltfm_pmops)
+#else
+#define SDHCI_PLTFM_PMOPS NULL
+#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..f98008b5e
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pxav2.c
@@ -0,0 +1,269 @@
+/*
+ * 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>
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#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/gpio.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);
+	pltfm_host->priv = NULL;
+
+	clk = clk_get(dev, "PXA-SDHCLK");
+	if (IS_ERR(clk)) {
+		dev_err(dev, "failed to get io clock\n");
+		ret = PTR_ERR(clk);
+		goto err_clk_get;
+	}
+	pltfm_host->clk = clk;
+	clk_prepare_enable(clk);
+
+	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) {
+		dev_err(&pdev->dev, "failed to add host\n");
+		goto err_add_host;
+	}
+
+	platform_set_drvdata(pdev, host);
+
+	return 0;
+
+err_add_host:
+	clk_disable_unprepare(clk);
+	clk_put(clk);
+err_clk_get:
+	sdhci_pltfm_free(pdev);
+	return ret;
+}
+
+static int sdhci_pxav2_remove(struct platform_device *pdev)
+{
+	struct sdhci_host *host = platform_get_drvdata(pdev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+
+	sdhci_remove_host(host, 1);
+
+	clk_disable_unprepare(pltfm_host->clk);
+	clk_put(pltfm_host->clk);
+	sdhci_pltfm_free(pdev);
+
+	return 0;
+}
+
+static struct platform_driver sdhci_pxav2_driver = {
+	.driver		= {
+		.name	= "sdhci-pxav2",
+#ifdef CONFIG_OF
+		.of_match_table = sdhci_pxav2_of_match,
+#endif
+		.pm	= SDHCI_PLTFM_PMOPS,
+	},
+	.probe		= sdhci_pxav2_probe,
+	.remove		= sdhci_pxav2_remove,
+};
+
+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..b5103a247
--- /dev/null
+++ b/drivers/mmc/host/sdhci-pxav3.c
@@ -0,0 +1,595 @@
+/*
+ * 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>
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/platform_device.h>
+#include <linux/clk.h>
+#include <linux/io.h>
+#include <linux/gpio.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/slot-gpio.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/of_gpio.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 = pltfm_host->priv;
+	struct resource *res;
+
+	host->quirks |= SDHCI_QUIRK_MISSING_CAPS;
+	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_readl(host, SDHCI_CAPABILITIES_1);
+		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
+	 */
+	host->caps = sdhci_readl(host, SDHCI_CAPABILITIES);
+	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 = pltfm_host->priv;
+	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 = pltfm_host->priv;
+	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 {
+			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 const struct sdhci_ops pxav3_sdhci_ops = {
+	.set_clock = sdhci_set_clock,
+	.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 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;
+
+	pxa = devm_kzalloc(&pdev->dev, sizeof(struct sdhci_pxa), GFP_KERNEL);
+	if (!pxa)
+		return -ENOMEM;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_pxav3_pdata, 0);
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	pltfm_host->priv = pxa;
+
+	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_clk_get;
+		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);
+	} 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;
+
+		if (gpio_is_valid(pdata->ext_cd_gpio)) {
+			ret = mmc_gpio_request_cd(host->mmc, pdata->ext_cd_gpio,
+						  0);
+			if (ret) {
+				dev_err(mmc_dev(host->mmc),
+					"failed to allocate card detect gpio\n");
+				goto err_cd_req;
+			}
+		}
+	}
+
+	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) {
+		dev_err(&pdev->dev, "failed to add host\n");
+		goto err_add_host;
+	}
+
+	platform_set_drvdata(pdev, host);
+
+	if (host->mmc->pm_caps & MMC_PM_KEEP_POWER) {
+		device_init_wakeup(&pdev->dev, 1);
+		host->mmc->pm_flags |= MMC_PM_WAKE_SDIO_IRQ;
+	} else {
+		device_init_wakeup(&pdev->dev, 0);
+	}
+
+	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_cd_req:
+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 = pltfm_host->priv;
+
+	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);
+	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 = pltfm_host->priv;
+	int ret;
+
+	ret = sdhci_runtime_suspend_host(host);
+	if (ret)
+		return ret;
+
+	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 = pltfm_host->priv;
+
+	clk_prepare_enable(pxa->clk_io);
+	if (!IS_ERR(pxa->clk_core))
+		clk_prepare_enable(pxa->clk_core);
+
+	return sdhci_runtime_resume_host(host);
+}
+#endif
+
+#ifdef CONFIG_PM
+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)
+};
+
+#define SDHCI_PXAV3_PMOPS (&sdhci_pxav3_pmops)
+
+#else
+#define SDHCI_PXAV3_PMOPS NULL
+#endif
+
+static struct platform_driver sdhci_pxav3_driver = {
+	.driver		= {
+		.name	= "sdhci-pxav3",
+#ifdef CONFIG_OF
+		.of_match_table = sdhci_pxav3_of_match,
+#endif
+		.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-regs.h b/drivers/mmc/host/sdhci-s3c-regs.h
new file mode 100644
index 000000000..e34049ad4
--- /dev/null
+++ b/drivers/mmc/host/sdhci-s3c-regs.h
@@ -0,0 +1,87 @@
+/* linux/arch/arm/plat-s3c/include/plat/regs-sdhci.h
+ *
+ * Copyright 2008 Openmoko, Inc.
+ * Copyright 2008 Simtec Electronics
+ *	http://armlinux.simtec.co.uk/
+ *	Ben Dooks <ben@simtec.co.uk>
+ *
+ * S3C Platform - SDHCI (HSMMC) register definitions
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+*/
+
+#ifndef __PLAT_S3C_SDHCI_REGS_H
+#define __PLAT_S3C_SDHCI_REGS_H __FILE__
+
+#define S3C_SDHCI_CONTROL2			(0x80)
+#define S3C_SDHCI_CONTROL3			(0x84)
+#define S3C64XX_SDHCI_CONTROL4			(0x8C)
+
+#define S3C64XX_SDHCI_CTRL2_ENSTAASYNCCLR	(1 << 31)
+#define S3C64XX_SDHCI_CTRL2_ENCMDCNFMSK		(1 << 30)
+#define S3C_SDHCI_CTRL2_CDINVRXD3		(1 << 29)
+#define S3C_SDHCI_CTRL2_SLCARDOUT		(1 << 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		(1 << 15)
+#define S3C_SDHCI_CTRL2_ENFBCLKRX		(1 << 14)
+#define S3C_SDHCI_CTRL2_SDCDSEL			(1 << 13)
+#define S3C_SDHCI_CTRL2_SDSIGPC			(1 << 12)
+#define S3C_SDHCI_CTRL2_ENBUSYCHKTXSTART	(1 << 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		(1 << 8)
+#define S3C_SDHCI_CTRL2_RWAITMODE		(1 << 7)
+#define S3C_SDHCI_CTRL2_DISBUFRD		(1 << 6)
+#define S3C_SDHCI_CTRL2_SELBASECLK_MASK		(0x3 << 4)
+#define S3C_SDHCI_CTRL2_SELBASECLK_SHIFT	(4)
+#define S3C_SDHCI_CTRL2_PWRSYNC			(1 << 3)
+#define S3C_SDHCI_CTRL2_ENCLKOUTMSKCON		(1 << 1)
+#define S3C_SDHCI_CTRL2_HWINITFIN		(1 << 0)
+
+#define S3C_SDHCI_CTRL3_FCSEL3			(1 << 31)
+#define S3C_SDHCI_CTRL3_FCSEL2			(1 << 23)
+#define S3C_SDHCI_CTRL3_FCSEL1			(1 << 15)
+#define S3C_SDHCI_CTRL3_FCSEL0			(1 << 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)
+
+#endif /* __PLAT_S3C_SDHCI_REGS_H */
diff --git a/drivers/mmc/host/sdhci-s3c.c b/drivers/mmc/host/sdhci-s3c.c
new file mode 100644
index 000000000..c6d2dd731
--- /dev/null
+++ b/drivers/mmc/host/sdhci-s3c.c
@@ -0,0 +1,777 @@
+/* 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
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#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-s3c-regs.h"
+#include "sdhci.h"
+
+#define MAX_BUS_CLK	(4)
+
+/**
+ * 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.
+ * @clk_io: The clock for the internal bus interface.
+ * @clk_bus: The clocks that are available for the SD/MMC bus clock.
+ */
+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;
+	int			ext_cd_gpio;
+
+	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.
+ *
+ * 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);
+
+	spin_unlock_irq(&host->lock);
+	ret = clk_set_rate(ourhost->clk_bus[ourhost->cur_clk], clock);
+	spin_lock_irq(&host->lock);
+	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);
+}
+
+/**
+ * sdhci_s3c_set_bus_width - support 8bit buswidth
+ * @host: The SDHCI host being queried
+ * @width: MMC_BUS_WIDTH_ macro for the bus width being requested
+ *
+ * We have 8-bit width support but is not a v3 controller.
+ * So we add platform_bus_width() and support 8bit width.
+ */
+static void sdhci_s3c_set_bus_width(struct sdhci_host *host, int width)
+{
+	u8 ctrl;
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+
+	switch (width) {
+	case MMC_BUS_WIDTH_8:
+		ctrl |= SDHCI_CTRL_8BITBUS;
+		ctrl &= ~SDHCI_CTRL_4BITBUS;
+		break;
+	case MMC_BUS_WIDTH_4:
+		ctrl |= SDHCI_CTRL_4BITBUS;
+		ctrl &= ~SDHCI_CTRL_8BITBUS;
+		break;
+	default:
+		ctrl &= ~SDHCI_CTRL_4BITBUS;
+		ctrl &= ~SDHCI_CTRL_8BITBUS;
+		break;
+	}
+
+	sdhci_writeb(host, ctrl, SDHCI_HOST_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_s3c_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
+
+static const struct of_device_id sdhci_s3c_dt_match[];
+
+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;
+	struct resource *res;
+	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) {
+		dev_err(dev, "no irq specified\n");
+		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));
+		sc->ext_cd_gpio = -1; /* invalid gpio number */
+	}
+
+	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;
+	}
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	host->ioaddr = devm_ioremap_resource(&pdev->dev, res);
+	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;
+	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) {
+		dev_err(dev, "sdhci_add_host() failed\n");
+		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);
+
+	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 (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);
+	return ret;
+}
+#endif
+
+#ifdef CONFIG_PM
+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)
+};
+
+#define SDHCI_S3C_PMOPS (&sdhci_s3c_pmops)
+
+#else
+#define SDHCI_S3C_PMOPS NULL
+#endif
+
+#if defined(CONFIG_CPU_EXYNOS4210) || defined(CONFIG_SOC_EXYNOS4212)
+static struct sdhci_s3c_drv_data exynos4_sdhci_drv_data = {
+	.no_divider = true,
+};
+#define EXYNOS4_SDHCI_DRV_DATA ((kernel_ulong_t)&exynos4_sdhci_drv_data)
+#else
+#define EXYNOS4_SDHCI_DRV_DATA ((kernel_ulong_t)NULL)
+#endif
+
+static struct platform_device_id sdhci_s3c_driver_ids[] = {
+	{
+		.name		= "s3c-sdhci",
+		.driver_data	= (kernel_ulong_t)NULL,
+	}, {
+		.name		= "exynos4-sdhci",
+		.driver_data	= EXYNOS4_SDHCI_DRV_DATA,
+	},
+	{ }
+};
+MODULE_DEVICE_TABLE(platform, sdhci_s3c_driver_ids);
+
+#ifdef CONFIG_OF
+static const struct of_device_id sdhci_s3c_dt_match[] = {
+	{ .compatible = "samsung,s3c6410-sdhci", },
+	{ .compatible = "samsung,exynos4210-sdhci",
+		.data = (void *)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",
+		.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..32848eb7a
--- /dev/null
+++ b/drivers/mmc/host/sdhci-sirf.c
@@ -0,0 +1,252 @@
+/*
+ * SDHCI support for SiRF primaII and marco SoCs
+ *
+ * Copyright (c) 2011 Cambridge Silicon Radio Limited, a CSR plc group company.
+ *
+ * Licensed under GPLv2 or later.
+ */
+
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/mmc/host.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_gpio.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 128
+
+struct sdhci_sirf_priv {
+	int gpio_cd;
+};
+
+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 int sdhci_sirf_execute_tuning(struct sdhci_host *host, u32 opcode)
+{
+	int tuning_seq_cnt = 3;
+	u8 phase, tuned_phases[SIRF_TUNING_COUNT];
+	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;
+	do {
+		sdhci_writel(host,
+			clock_setting | phase,
+			SDHCI_CLK_DELAY_SETTING);
+
+		if (!mmc_send_tuning(mmc)) {
+			/* 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);
+			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 < ARRAY_SIZE(tuned_phases));
+
+	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 struct sdhci_ops sdhci_sirf_ops = {
+	.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 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_INVERTED_WRITE_PROTECT |
+		SDHCI_QUIRK_DELAY_AFTER_POWER,
+};
+
+static int sdhci_sirf_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct sdhci_pltfm_host *pltfm_host;
+	struct sdhci_sirf_priv *priv;
+	struct clk *clk;
+	int gpio_cd;
+	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);
+	}
+
+	if (pdev->dev.of_node)
+		gpio_cd = of_get_named_gpio(pdev->dev.of_node, "cd-gpios", 0);
+	else
+		gpio_cd = -EINVAL;
+
+	host = sdhci_pltfm_init(pdev, &sdhci_sirf_pdata, sizeof(struct sdhci_sirf_priv));
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+
+	pltfm_host = sdhci_priv(host);
+	pltfm_host->clk = clk;
+	priv = sdhci_pltfm_priv(pltfm_host);
+	priv->gpio_cd = gpio_cd;
+
+	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.
+	 */
+	if (gpio_is_valid(priv->gpio_cd)) {
+		ret = mmc_gpio_request_cd(host->mmc, priv->gpio_cd, 0);
+		if (ret) {
+			dev_err(&pdev->dev, "card detect irq request failed: %d\n",
+				ret);
+			goto err_request_cd;
+		}
+		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;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int sdhci_sirf_suspend(struct device *dev)
+{
+	struct sdhci_host *host = dev_get_drvdata(dev);
+	struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
+	int ret;
+
+	ret = sdhci_suspend_host(host);
+	if (ret)
+		return ret;
+
+	clk_disable(pltfm_host->clk);
+
+	return 0;
+}
+
+static int sdhci_sirf_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_enable(pltfm_host->clk);
+	if (ret) {
+		dev_dbg(dev, "Resume: Error enabling clock\n");
+		return ret;
+	}
+
+	return sdhci_resume_host(host);
+}
+
+static SIMPLE_DEV_PM_OPS(sdhci_sirf_pm_ops, sdhci_sirf_suspend, sdhci_sirf_resume);
+#endif
+
+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",
+		.of_match_table = sdhci_sirf_of_match,
+#ifdef CONFIG_PM_SLEEP
+		.pm	= &sdhci_sirf_pm_ops,
+#endif
+	},
+	.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..df088343d
--- /dev/null
+++ b/drivers/mmc/host/sdhci-spear.c
@@ -0,0 +1,215 @@
+/*
+ * drivers/mmc/host/sdhci-spear.c
+ *
+ * Support of SDHCI platform devices for spear soc family
+ *
+ * Copyright (C) 2010 ST Microelectronics
+ * Viresh Kumar <viresh.linux@gmail.com>
+ *
+ * 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/gpio.h>
+#include <linux/highmem.h>
+#include <linux/module.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/of.h>
+#include <linux/of_gpio.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;
+	int card_int_gpio;
+};
+
+/* 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 void sdhci_probe_config_dt(struct device_node *np,
+				struct spear_sdhci *host)
+{
+	int cd_gpio;
+
+	cd_gpio = of_get_named_gpio(np, "cd-gpios", 0);
+	if (!gpio_is_valid(cd_gpio))
+		cd_gpio = -1;
+
+	host->card_int_gpio = cd_gpio;
+}
+
+static int sdhci_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct resource *iomem;
+	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;
+	}
+
+	iomem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	host->ioaddr = devm_ioremap_resource(&pdev->dev, iomem);
+	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);
+	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));
+
+	sdhci_probe_config_dt(pdev->dev.of_node, sdhci);
+	/*
+	 * It is optional to use GPIOs for sdhci card detection. If
+	 * sdhci->card_int_gpio < 0, then use original sdhci lines otherwise
+	 * GPIO lines. We use the built-in GPIO support for this.
+	 */
+	if (sdhci->card_int_gpio >= 0) {
+		ret = mmc_gpio_request_cd(host->mmc, sdhci->card_int_gpio, 0);
+		if (ret < 0) {
+			dev_dbg(&pdev->dev,
+				"failed to request card-detect gpio%d\n",
+				sdhci->card_int_gpio);
+			goto disable_clk;
+		}
+	}
+
+	ret = sdhci_add_host(host);
+	if (ret) {
+		dev_dbg(&pdev->dev, "error adding host\n");
+		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;
+
+	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",
+		.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 <viresh.linux@gmail.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/sdhci-st.c b/drivers/mmc/host/sdhci-st.c
new file mode 100644
index 000000000..682f2bb0f
--- /dev/null
+++ b/drivers/mmc/host/sdhci-st.c
@@ -0,0 +1,512 @@
+/*
+ * 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
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#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;
+	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 (mhost->caps & MMC_CAP_NONREMOVABLE)
+		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 = pltfm_host->priv;
+
+	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 = pltfm_host->priv;
+	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;
+	int ret = 0;
+	u16 host_version;
+	struct resource *res;
+
+	pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
+	if (!pdata)
+		return -ENOMEM;
+
+	clk =  devm_clk_get(&pdev->dev, "mmc");
+	if (IS_ERR(clk)) {
+		dev_err(&pdev->dev, "Peripheral clk not found\n");
+		return PTR_ERR(clk);
+	}
+
+	pdata->rstc = devm_reset_control_get(&pdev->dev, NULL);
+	if (IS_ERR(pdata->rstc))
+		pdata->rstc = NULL;
+	else
+		reset_control_deassert(pdata->rstc);
+
+	host = sdhci_pltfm_init(pdev, &sdhci_st_pdata, 0);
+	if (IS_ERR(host)) {
+		dev_err(&pdev->dev, "Failed sdhci_pltfm_init\n");
+		ret = PTR_ERR(host);
+		goto err_pltfm_init;
+	}
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed mmc_of_parse\n");
+		goto err_of;
+	}
+
+	clk_prepare_enable(clk);
+
+	/* 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 = sdhci_priv(host);
+	pltfm_host->priv = pdata;
+	pltfm_host->clk = clk;
+
+	/* Configure the Arasan HC inside the flashSS */
+	st_mmcss_cconfig(np, host);
+
+	ret = sdhci_add_host(host);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed sdhci_add_host\n");
+		goto err_out;
+	}
+
+	platform_set_drvdata(pdev, host);
+
+	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(clk);
+err_of:
+	sdhci_pltfm_free(pdev);
+err_pltfm_init:
+	if (pdata->rstc)
+		reset_control_assert(pdata->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 = pltfm_host->priv;
+	int ret;
+
+	ret = sdhci_pltfm_unregister(pdev);
+
+	if (pdata->rstc)
+		reset_control_assert(pdata->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 = pltfm_host->priv;
+	int ret = sdhci_suspend_host(host);
+
+	if (ret)
+		goto out;
+
+	if (pdata->rstc)
+		reset_control_assert(pdata->rstc);
+
+	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 = pltfm_host->priv;
+	struct device_node *np = dev->of_node;
+
+	clk_prepare_enable(pltfm_host->clk);
+
+	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",
+		   .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:st-sdhci");
diff --git a/drivers/mmc/host/sdhci-tegra.c b/drivers/mmc/host/sdhci-tegra.c
new file mode 100644
index 000000000..ad28b49f0
--- /dev/null
+++ b/drivers/mmc/host/sdhci-tegra.c
@@ -0,0 +1,327 @@
+/*
+ * Copyright (C) 2010 Google, Inc.
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#include <linux/err.h>
+#include <linux/module.h>
+#include <linux/init.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/mmc/card.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/gpio/consumer.h>
+
+#include "sdhci-pltfm.h"
+
+/* Tegra SDHOST controller vendor register definitions */
+#define SDHCI_TEGRA_VENDOR_MISC_CTRL		0x120
+#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 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_DISABLE_SDR50		BIT(3)
+#define NVQUIRK_DISABLE_SDR104		BIT(4)
+#define NVQUIRK_DISABLE_DDR50		BIT(5)
+
+struct sdhci_tegra_soc_data {
+	const struct sdhci_pltfm_data *pdata;
+	u32 nvquirks;
+};
+
+struct sdhci_tegra {
+	const struct sdhci_tegra_soc_data *soc_data;
+	struct gpio_desc *power_gpio;
+};
+
+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 = pltfm_host->priv;
+	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 = pltfm_host->priv;
+	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 unsigned int tegra_sdhci_get_ro(struct sdhci_host *host)
+{
+	return mmc_gpio_get_ro(host->mmc);
+}
+
+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 = pltfm_host->priv;
+	const struct sdhci_tegra_soc_data *soc_data = tegra_host->soc_data;
+	u32 misc_ctrl;
+
+	sdhci_reset(host, mask);
+
+	if (!(mask & SDHCI_RESET_ALL))
+		return;
+
+	misc_ctrl = sdhci_readw(host, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+	/* 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;
+	/* Don't advertise UHS modes which aren't supported yet */
+	if (soc_data->nvquirks & NVQUIRK_DISABLE_SDR50)
+		misc_ctrl &= ~SDHCI_MISC_CTRL_ENABLE_SDR50;
+	if (soc_data->nvquirks & NVQUIRK_DISABLE_DDR50)
+		misc_ctrl &= ~SDHCI_MISC_CTRL_ENABLE_DDR50;
+	if (soc_data->nvquirks & NVQUIRK_DISABLE_SDR104)
+		misc_ctrl &= ~SDHCI_MISC_CTRL_ENABLE_SDR104;
+	sdhci_writew(host, misc_ctrl, SDHCI_TEGRA_VENDOR_MISC_CTRL);
+}
+
+static void tegra_sdhci_set_bus_width(struct sdhci_host *host, int bus_width)
+{
+	u32 ctrl;
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+	if ((host->mmc->caps & MMC_CAP_8_BIT_DATA) &&
+	    (bus_width == MMC_BUS_WIDTH_8)) {
+		ctrl &= ~SDHCI_CTRL_4BITBUS;
+		ctrl |= SDHCI_CTRL_8BITBUS;
+	} else {
+		ctrl &= ~SDHCI_CTRL_8BITBUS;
+		if (bus_width == MMC_BUS_WIDTH_4)
+			ctrl |= SDHCI_CTRL_4BITBUS;
+		else
+			ctrl &= ~SDHCI_CTRL_4BITBUS;
+	}
+	sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+}
+
+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  = sdhci_set_clock,
+	.set_bus_width = tegra_sdhci_set_bus_width,
+	.reset      = tegra_sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+	.get_max_clock = sdhci_pltfm_clk_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 struct sdhci_tegra_soc_data soc_data_tegra20 = {
+	.pdata = &sdhci_tegra20_pdata,
+	.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,
+	.ops  = &tegra_sdhci_ops,
+};
+
+static struct sdhci_tegra_soc_data soc_data_tegra30 = {
+	.pdata = &sdhci_tegra30_pdata,
+	.nvquirks = NVQUIRK_ENABLE_SDHCI_SPEC_300 |
+		    NVQUIRK_DISABLE_SDR50 |
+		    NVQUIRK_DISABLE_SDR104,
+};
+
+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  = sdhci_set_clock,
+	.set_bus_width = tegra_sdhci_set_bus_width,
+	.reset      = tegra_sdhci_reset,
+	.set_uhs_signaling = sdhci_set_uhs_signaling,
+	.get_max_clock = sdhci_pltfm_clk_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,
+	.ops  = &tegra114_sdhci_ops,
+};
+
+static struct sdhci_tegra_soc_data soc_data_tegra114 = {
+	.pdata = &sdhci_tegra114_pdata,
+	.nvquirks = NVQUIRK_DISABLE_SDR50 |
+		    NVQUIRK_DISABLE_DDR50 |
+		    NVQUIRK_DISABLE_SDR104,
+};
+
+static const struct of_device_id sdhci_tegra_dt_match[] = {
+	{ .compatible = "nvidia,tegra124-sdhci", .data = &soc_data_tegra114 },
+	{ .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_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, 0);
+	if (IS_ERR(host))
+		return PTR_ERR(host);
+	pltfm_host = sdhci_priv(host);
+
+	tegra_host = devm_kzalloc(&pdev->dev, sizeof(*tegra_host), GFP_KERNEL);
+	if (!tegra_host) {
+		dev_err(mmc_dev(host->mmc), "failed to allocate tegra_host\n");
+		rc = -ENOMEM;
+		goto err_alloc_tegra_host;
+	}
+	tegra_host->soc_data = soc_data;
+	pltfm_host->priv = tegra_host;
+
+	rc = mmc_of_parse(host->mmc);
+	if (rc)
+		goto err_parse_dt;
+
+	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;
+	}
+
+	clk = devm_clk_get(mmc_dev(host->mmc), NULL);
+	if (IS_ERR(clk)) {
+		dev_err(mmc_dev(host->mmc), "clk err\n");
+		rc = PTR_ERR(clk);
+		goto err_clk_get;
+	}
+	clk_prepare_enable(clk);
+	pltfm_host->clk = clk;
+
+	rc = sdhci_add_host(host);
+	if (rc)
+		goto err_add_host;
+
+	return 0;
+
+err_add_host:
+	clk_disable_unprepare(pltfm_host->clk);
+err_clk_get:
+err_power_req:
+err_parse_dt:
+err_alloc_tegra_host:
+	sdhci_pltfm_free(pdev);
+	return rc;
+}
+
+static struct platform_driver sdhci_tegra_driver = {
+	.driver		= {
+		.name	= "sdhci-tegra",
+		.of_match_table = sdhci_tegra_dt_match,
+		.pm	= SDHCI_PLTFM_PMOPS,
+	},
+	.probe		= sdhci_tegra_probe,
+	.remove		= sdhci_pltfm_unregister,
+};
+
+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.c b/drivers/mmc/host/sdhci.c
new file mode 100644
index 000000000..d3dbb2805
--- /dev/null
+++ b/drivers/mmc/host/sdhci.c
@@ -0,0 +1,3569 @@
+/*
+ *  linux/drivers/mmc/host/sdhci.c - Secure Digital Host Controller Interface driver
+ *
+ *  Copyright (C) 2005-2008 Pierre Ossman, All Rights Reserved.
+ *
+ * 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; either version 2 of the License, or (at
+ * your option) any later version.
+ *
+ * Thanks to the following companies for their support:
+ *
+ *     - JMicron (hardware and technical support)
+ */
+
+#include <linux/delay.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/regulator/consumer.h>
+#include <linux/pm_runtime.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(DRIVER_NAME " [%s()]: " f, __func__,## x)
+
+#if defined(CONFIG_LEDS_CLASS) || (defined(CONFIG_LEDS_CLASS_MODULE) && \
+	defined(CONFIG_MMC_SDHCI_MODULE))
+#define SDHCI_USE_LEDS_CLASS
+#endif
+
+#define MAX_TUNING_LOOP 40
+
+static unsigned int debug_quirks = 0;
+static unsigned int debug_quirks2;
+
+static void sdhci_finish_data(struct sdhci_host *);
+
+static void sdhci_finish_command(struct sdhci_host *);
+static int sdhci_execute_tuning(struct mmc_host *mmc, u32 opcode);
+static void sdhci_tuning_timer(unsigned long data);
+static void sdhci_enable_preset_value(struct sdhci_host *host, bool enable);
+static int sdhci_pre_dma_transfer(struct sdhci_host *host,
+					struct mmc_data *data,
+					struct sdhci_host_next *next);
+static int sdhci_do_get_cd(struct sdhci_host *host);
+
+#ifdef CONFIG_PM
+static int sdhci_runtime_pm_get(struct sdhci_host *host);
+static int sdhci_runtime_pm_put(struct sdhci_host *host);
+static void sdhci_runtime_pm_bus_on(struct sdhci_host *host);
+static void sdhci_runtime_pm_bus_off(struct sdhci_host *host);
+#else
+static inline int sdhci_runtime_pm_get(struct sdhci_host *host)
+{
+	return 0;
+}
+static inline int sdhci_runtime_pm_put(struct sdhci_host *host)
+{
+	return 0;
+}
+static void sdhci_runtime_pm_bus_on(struct sdhci_host *host)
+{
+}
+static void sdhci_runtime_pm_bus_off(struct sdhci_host *host)
+{
+}
+#endif
+
+static void sdhci_dumpregs(struct sdhci_host *host)
+{
+	pr_debug(DRIVER_NAME ": =========== REGISTER DUMP (%s)===========\n",
+		mmc_hostname(host->mmc));
+
+	pr_debug(DRIVER_NAME ": Sys addr: 0x%08x | Version:  0x%08x\n",
+		sdhci_readl(host, SDHCI_DMA_ADDRESS),
+		sdhci_readw(host, SDHCI_HOST_VERSION));
+	pr_debug(DRIVER_NAME ": Blk size: 0x%08x | Blk cnt:  0x%08x\n",
+		sdhci_readw(host, SDHCI_BLOCK_SIZE),
+		sdhci_readw(host, SDHCI_BLOCK_COUNT));
+	pr_debug(DRIVER_NAME ": Argument: 0x%08x | Trn mode: 0x%08x\n",
+		sdhci_readl(host, SDHCI_ARGUMENT),
+		sdhci_readw(host, SDHCI_TRANSFER_MODE));
+	pr_debug(DRIVER_NAME ": Present:  0x%08x | Host ctl: 0x%08x\n",
+		sdhci_readl(host, SDHCI_PRESENT_STATE),
+		sdhci_readb(host, SDHCI_HOST_CONTROL));
+	pr_debug(DRIVER_NAME ": Power:    0x%08x | Blk gap:  0x%08x\n",
+		sdhci_readb(host, SDHCI_POWER_CONTROL),
+		sdhci_readb(host, SDHCI_BLOCK_GAP_CONTROL));
+	pr_debug(DRIVER_NAME ": Wake-up:  0x%08x | Clock:    0x%08x\n",
+		sdhci_readb(host, SDHCI_WAKE_UP_CONTROL),
+		sdhci_readw(host, SDHCI_CLOCK_CONTROL));
+	pr_debug(DRIVER_NAME ": Timeout:  0x%08x | Int stat: 0x%08x\n",
+		sdhci_readb(host, SDHCI_TIMEOUT_CONTROL),
+		sdhci_readl(host, SDHCI_INT_STATUS));
+	pr_debug(DRIVER_NAME ": Int enab: 0x%08x | Sig enab: 0x%08x\n",
+		sdhci_readl(host, SDHCI_INT_ENABLE),
+		sdhci_readl(host, SDHCI_SIGNAL_ENABLE));
+	pr_debug(DRIVER_NAME ": AC12 err: 0x%08x | Slot int: 0x%08x\n",
+		sdhci_readw(host, SDHCI_ACMD12_ERR),
+		sdhci_readw(host, SDHCI_SLOT_INT_STATUS));
+	pr_debug(DRIVER_NAME ": Caps:     0x%08x | Caps_1:   0x%08x\n",
+		sdhci_readl(host, SDHCI_CAPABILITIES),
+		sdhci_readl(host, SDHCI_CAPABILITIES_1));
+	pr_debug(DRIVER_NAME ": Cmd:      0x%08x | Max curr: 0x%08x\n",
+		sdhci_readw(host, SDHCI_COMMAND),
+		sdhci_readl(host, SDHCI_MAX_CURRENT));
+	pr_debug(DRIVER_NAME ": 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)
+			pr_debug(DRIVER_NAME ": ADMA Err: 0x%08x | ADMA Ptr: 0x%08x%08x\n",
+				 readl(host->ioaddr + SDHCI_ADMA_ERROR),
+				 readl(host->ioaddr + SDHCI_ADMA_ADDRESS_HI),
+				 readl(host->ioaddr + SDHCI_ADMA_ADDRESS));
+		else
+			pr_debug(DRIVER_NAME ": ADMA Err: 0x%08x | ADMA Ptr: 0x%08x\n",
+				 readl(host->ioaddr + SDHCI_ADMA_ERROR),
+				 readl(host->ioaddr + SDHCI_ADMA_ADDRESS));
+	}
+
+	pr_debug(DRIVER_NAME ": ===========================================\n");
+}
+
+/*****************************************************************************\
+ *                                                                           *
+ * Low level functions                                                       *
+ *                                                                           *
+\*****************************************************************************/
+
+static void sdhci_set_card_detection(struct sdhci_host *host, bool enable)
+{
+	u32 present;
+
+	if ((host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) ||
+	    (host->mmc->caps & MMC_CAP_NONREMOVABLE))
+		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);
+}
+
+void sdhci_reset(struct sdhci_host *host, u8 mask)
+{
+	unsigned long 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 = 100;
+
+	/* hw clears the bit when it's done */
+	while (sdhci_readb(host, SDHCI_SOFTWARE_RESET) & mask) {
+		if (timeout == 0) {
+			pr_err("%s: Reset 0x%x never completed.\n",
+				mmc_hostname(host->mmc), (int)mask);
+			sdhci_dumpregs(host);
+			return;
+		}
+		timeout--;
+		mdelay(1);
+	}
+}
+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) {
+		if (!(sdhci_readl(host, SDHCI_PRESENT_STATE) &
+			SDHCI_CARD_PRESENT))
+			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_ios(struct mmc_host *mmc, struct mmc_ios *ios);
+
+static void sdhci_init(struct sdhci_host *host, int soft)
+{
+	if (soft)
+		sdhci_do_reset(host, SDHCI_RESET_CMD|SDHCI_RESET_DATA);
+	else
+		sdhci_do_reset(host, SDHCI_RESET_ALL);
+
+	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;
+
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+
+	if (soft) {
+		/* force clock reconfiguration */
+		host->clock = 0;
+		sdhci_set_ios(host->mmc, &host->mmc->ios);
+	}
+}
+
+static void sdhci_reinit(struct sdhci_host *host)
+{
+	sdhci_init(host, 0);
+	/*
+	 * Retuning stuffs are affected by different cards inserted and only
+	 * applicable to UHS-I cards. So reset these fields to their initial
+	 * value when card is removed.
+	 */
+	if (host->flags & SDHCI_USING_RETUNING_TIMER) {
+		host->flags &= ~SDHCI_USING_RETUNING_TIMER;
+
+		del_timer_sync(&host->tuning_timer);
+		host->flags &= ~SDHCI_NEEDS_RETUNING;
+	}
+	sdhci_enable_card_detection(host);
+}
+
+static void sdhci_activate_led(struct sdhci_host *host)
+{
+	u8 ctrl;
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+	ctrl |= SDHCI_CTRL_LED;
+	sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+}
+
+static void sdhci_deactivate_led(struct sdhci_host *host)
+{
+	u8 ctrl;
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+	ctrl &= ~SDHCI_CTRL_LED;
+	sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+}
+
+#ifdef SDHCI_USE_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_deactivate_led(host);
+	else
+		sdhci_activate_led(host);
+out:
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+#endif
+
+/*****************************************************************************\
+ *                                                                           *
+ * Core functions                                                            *
+ *                                                                           *
+\*****************************************************************************/
+
+static void sdhci_read_block_pio(struct sdhci_host *host)
+{
+	unsigned long flags;
+	size_t blksize, len, chunk;
+	u32 uninitialized_var(scratch);
+	u8 *buf;
+
+	DBG("PIO reading\n");
+
+	blksize = host->data->blksz;
+	chunk = 0;
+
+	local_irq_save(flags);
+
+	while (blksize) {
+		if (!sg_miter_next(&host->sg_miter))
+			BUG();
+
+		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) {
+		if (!sg_miter_next(&host->sg_miter))
+			BUG();
+
+		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;
+
+	BUG_ON(!host->data);
+
+	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 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);
+}
+
+static void sdhci_adma_write_desc(struct sdhci_host *host, void *desc,
+				  dma_addr_t addr, int len, unsigned 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((u32)addr);
+
+	if (host->flags & SDHCI_USE_64_BIT_DMA)
+		dma_desc->addr_hi = cpu_to_le32((u64)addr >> 32);
+}
+
+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 int sdhci_adma_table_pre(struct sdhci_host *host,
+	struct mmc_data *data)
+{
+	int direction;
+
+	void *desc;
+	void *align;
+	dma_addr_t addr;
+	dma_addr_t align_addr;
+	int len, offset;
+
+	struct scatterlist *sg;
+	int i;
+	char *buffer;
+	unsigned long flags;
+
+	/*
+	 * The spec does not specify endianness of descriptor table.
+	 * We currently guess that it is LE.
+	 */
+
+	if (data->flags & MMC_DATA_READ)
+		direction = DMA_FROM_DEVICE;
+	else
+		direction = DMA_TO_DEVICE;
+
+	host->align_addr = dma_map_single(mmc_dev(host->mmc),
+		host->align_buffer, host->align_buffer_sz, direction);
+	if (dma_mapping_error(mmc_dev(host->mmc), host->align_addr))
+		goto fail;
+	BUG_ON(host->align_addr & host->align_mask);
+
+	host->sg_count = sdhci_pre_dma_transfer(host, data, NULL);
+	if (host->sg_count < 0)
+		goto unmap_align;
+
+	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 = (host->align_sz - (addr & host->align_mask)) &
+			 host->align_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 += host->align_sz;
+			align_addr += host->align_sz;
+
+			desc += host->desc_sz;
+
+			addr += offset;
+			len -= offset;
+		}
+
+		BUG_ON(len > 65536);
+
+		/* tran, valid */
+		sdhci_adma_write_desc(host, desc, addr, len, ADMA2_TRAN_VALID);
+		desc += host->desc_sz;
+
+		/*
+		 * 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);
+	}
+
+	/*
+	 * Resync align buffer as we might have changed it.
+	 */
+	if (data->flags & MMC_DATA_WRITE) {
+		dma_sync_single_for_device(mmc_dev(host->mmc),
+			host->align_addr, host->align_buffer_sz, direction);
+	}
+
+	return 0;
+
+unmap_align:
+	dma_unmap_single(mmc_dev(host->mmc), host->align_addr,
+		host->align_buffer_sz, direction);
+fail:
+	return -EINVAL;
+}
+
+static void sdhci_adma_table_post(struct sdhci_host *host,
+	struct mmc_data *data)
+{
+	int direction;
+
+	struct scatterlist *sg;
+	int i, size;
+	void *align;
+	char *buffer;
+	unsigned long flags;
+	bool has_unaligned;
+
+	if (data->flags & MMC_DATA_READ)
+		direction = DMA_FROM_DEVICE;
+	else
+		direction = DMA_TO_DEVICE;
+
+	dma_unmap_single(mmc_dev(host->mmc), host->align_addr,
+		host->align_buffer_sz, direction);
+
+	/* Do a quick scan of the SG list for any unaligned mappings */
+	has_unaligned = false;
+	for_each_sg(data->sg, sg, host->sg_count, i)
+		if (sg_dma_address(sg) & host->align_mask) {
+			has_unaligned = true;
+			break;
+		}
+
+	if (has_unaligned && data->flags & MMC_DATA_READ) {
+		dma_sync_sg_for_cpu(mmc_dev(host->mmc), data->sg,
+			data->sg_len, direction);
+
+		align = host->align_buffer;
+
+		for_each_sg(data->sg, sg, host->sg_count, i) {
+			if (sg_dma_address(sg) & host->align_mask) {
+				size = host->align_sz -
+				       (sg_dma_address(sg) & host->align_mask);
+
+				buffer = sdhci_kmap_atomic(sg, &flags);
+				memcpy(buffer, align, size);
+				sdhci_kunmap_atomic(buffer, &flags);
+
+				align += host->align_sz;
+			}
+		}
+	}
+
+	if (!data->host_cookie)
+		dma_unmap_sg(mmc_dev(host->mmc), data->sg,
+			data->sg_len, direction);
+}
+
+static u8 sdhci_calc_timeout(struct sdhci_host *host, struct mmc_command *cmd)
+{
+	u8 count;
+	struct mmc_data *data = cmd->data;
+	unsigned target_timeout, current_timeout;
+
+	/*
+	 * 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 timeout, assume max */
+	if (!data && !cmd->busy_timeout)
+		return 0xE;
+
+	/* timeout in us */
+	if (!data)
+		target_timeout = cmd->busy_timeout * 1000;
+	else {
+		target_timeout = data->timeout_ns / 1000;
+		if (host->clock)
+			target_timeout += data->timeout_clks / host->clock;
+	}
+
+	/*
+	 * 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) {
+		DBG("%s: Too large timeout 0x%x requested for CMD%d!\n",
+		    mmc_hostname(host->mmc), count, cmd->opcode);
+		count = 0xE;
+	}
+
+	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;
+
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+}
+
+static void sdhci_set_timeout(struct sdhci_host *host, struct mmc_command *cmd)
+{
+	u8 count;
+
+	if (host->ops->set_timeout) {
+		host->ops->set_timeout(host, cmd);
+	} else {
+		count = sdhci_calc_timeout(host, cmd);
+		sdhci_writeb(host, count, SDHCI_TIMEOUT_CONTROL);
+	}
+}
+
+static void sdhci_prepare_data(struct sdhci_host *host, struct mmc_command *cmd)
+{
+	u8 ctrl;
+	struct mmc_data *data = cmd->data;
+	int ret;
+
+	WARN_ON(host->data);
+
+	if (data || (cmd->flags & MMC_RSP_BUSY))
+		sdhci_set_timeout(host, cmd);
+
+	if (!data)
+		return;
+
+	/* 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;
+
+	if (host->flags & (SDHCI_USE_SDMA | SDHCI_USE_ADMA))
+		host->flags |= SDHCI_REQ_USE_DMA;
+
+	/*
+	 * FIXME: This doesn't account for merging when mapping the
+	 * scatterlist.
+	 */
+	if (host->flags & SDHCI_REQ_USE_DMA) {
+		int broken, i;
+		struct scatterlist *sg;
+
+		broken = 0;
+		if (host->flags & SDHCI_USE_ADMA) {
+			if (host->quirks & SDHCI_QUIRK_32BIT_ADMA_SIZE)
+				broken = 1;
+		} else {
+			if (host->quirks & SDHCI_QUIRK_32BIT_DMA_SIZE)
+				broken = 1;
+		}
+
+		if (unlikely(broken)) {
+			for_each_sg(data->sg, sg, data->sg_len, i) {
+				if (sg->length & 0x3) {
+					DBG("Reverting to PIO because of "
+						"transfer size (%d)\n",
+						sg->length);
+					host->flags &= ~SDHCI_REQ_USE_DMA;
+					break;
+				}
+			}
+		}
+	}
+
+	/*
+	 * The assumption here being that alignment is the same after
+	 * translation to device address space.
+	 */
+	if (host->flags & SDHCI_REQ_USE_DMA) {
+		int broken, i;
+		struct scatterlist *sg;
+
+		broken = 0;
+		if (host->flags & SDHCI_USE_ADMA) {
+			/*
+			 * As we use 3 byte chunks to work around
+			 * alignment problems, we need to check this
+			 * quirk.
+			 */
+			if (host->quirks & SDHCI_QUIRK_32BIT_ADMA_SIZE)
+				broken = 1;
+		} else {
+			if (host->quirks & SDHCI_QUIRK_32BIT_DMA_ADDR)
+				broken = 1;
+		}
+
+		if (unlikely(broken)) {
+			for_each_sg(data->sg, sg, data->sg_len, i) {
+				if (sg->offset & 0x3) {
+					DBG("Reverting to PIO because of "
+						"bad alignment\n");
+					host->flags &= ~SDHCI_REQ_USE_DMA;
+					break;
+				}
+			}
+		}
+	}
+
+	if (host->flags & SDHCI_REQ_USE_DMA) {
+		if (host->flags & SDHCI_USE_ADMA) {
+			ret = sdhci_adma_table_pre(host, data);
+			if (ret) {
+				/*
+				 * This only happens when someone fed
+				 * us an invalid request.
+				 */
+				WARN_ON(1);
+				host->flags &= ~SDHCI_REQ_USE_DMA;
+			} else {
+				sdhci_writel(host, host->adma_addr,
+					SDHCI_ADMA_ADDRESS);
+				if (host->flags & SDHCI_USE_64_BIT_DMA)
+					sdhci_writel(host,
+						     (u64)host->adma_addr >> 32,
+						     SDHCI_ADMA_ADDRESS_HI);
+			}
+		} else {
+			int sg_cnt;
+
+			sg_cnt = sdhci_pre_dma_transfer(host, data, NULL);
+			if (sg_cnt <= 0) {
+				/*
+				 * This only happens when someone fed
+				 * us an invalid request.
+				 */
+				WARN_ON(1);
+				host->flags &= ~SDHCI_REQ_USE_DMA;
+			} else {
+				WARN_ON(sg_cnt != 1);
+				sdhci_writel(host, sg_dma_address(data->sg),
+					SDHCI_DMA_ADDRESS);
+			}
+		}
+	}
+
+	/*
+	 * Always adjust the DMA selection as some controllers
+	 * (e.g. JMicron) can't do PIO properly when the selection
+	 * is ADMA.
+	 */
+	if (host->version >= SDHCI_SPEC_200) {
+		ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+		ctrl &= ~SDHCI_CTRL_DMA_MASK;
+		if ((host->flags & SDHCI_REQ_USE_DMA) &&
+			(host->flags & SDHCI_USE_ADMA)) {
+			if (host->flags & SDHCI_USE_64_BIT_DMA)
+				ctrl |= SDHCI_CTRL_ADMA64;
+			else
+				ctrl |= SDHCI_CTRL_ADMA32;
+		} else {
+			ctrl |= SDHCI_CTRL_SDMA;
+		}
+		sdhci_writeb(host, ctrl, SDHCI_HOST_CONTROL);
+	}
+
+	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);
+
+	/* Set the DMA boundary value and block size */
+	sdhci_writew(host, SDHCI_MAKE_BLKSZ(SDHCI_DEFAULT_BOUNDARY_ARG,
+		data->blksz), SDHCI_BLOCK_SIZE);
+	sdhci_writew(host, data->blocks, SDHCI_BLOCK_COUNT);
+}
+
+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) {
+			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;
+		/*
+		 * If we are sending CMD23, CMD12 never gets sent
+		 * on successful completion (so no Auto-CMD12).
+		 */
+		if (!host->mrq->sbc && (host->flags & SDHCI_AUTO_CMD12) &&
+		    (cmd->opcode != SD_IO_RW_EXTENDED))
+			mode |= SDHCI_TRNS_AUTO_CMD12;
+		else if (host->mrq->sbc && (host->flags & SDHCI_AUTO_CMD23)) {
+			mode |= SDHCI_TRNS_AUTO_CMD23;
+			sdhci_writel(host, host->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 void sdhci_finish_data(struct sdhci_host *host)
+{
+	struct mmc_data *data;
+
+	BUG_ON(!host->data);
+
+	data = host->data;
+	host->data = NULL;
+
+	if (host->flags & SDHCI_REQ_USE_DMA) {
+		if (host->flags & SDHCI_USE_ADMA)
+			sdhci_adma_table_post(host, data);
+		else {
+			if (!data->host_cookie)
+				dma_unmap_sg(mmc_dev(host->mmc),
+					data->sg, data->sg_len,
+					(data->flags & MMC_DATA_READ) ?
+					DMA_FROM_DEVICE : DMA_TO_DEVICE);
+		}
+	}
+
+	/*
+	 * 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) {
+			sdhci_do_reset(host, SDHCI_RESET_CMD);
+			sdhci_do_reset(host, SDHCI_RESET_DATA);
+		}
+
+		sdhci_send_command(host, data->stop);
+	} else
+		tasklet_schedule(&host->finish_tasklet);
+}
+
+void sdhci_send_command(struct sdhci_host *host, struct mmc_command *cmd)
+{
+	int flags;
+	u32 mask;
+	unsigned long timeout;
+
+	WARN_ON(host->cmd);
+
+	/* Wait max 10 ms */
+	timeout = 10;
+
+	mask = SDHCI_CMD_INHIBIT;
+	if ((cmd->data != NULL) || (cmd->flags & MMC_RSP_BUSY))
+		mask |= SDHCI_DATA_INHIBIT;
+
+	/* We shouldn't wait for data inihibit for stop commands, even
+	   though they might use busy signaling */
+	if (host->mrq->data && (cmd == host->mrq->data->stop))
+		mask &= ~SDHCI_DATA_INHIBIT;
+
+	while (sdhci_readl(host, SDHCI_PRESENT_STATE) & mask) {
+		if (timeout == 0) {
+			pr_err("%s: Controller never released "
+				"inhibit bit(s).\n", mmc_hostname(host->mmc));
+			sdhci_dumpregs(host);
+			cmd->error = -EIO;
+			tasklet_schedule(&host->finish_tasklet);
+			return;
+		}
+		timeout--;
+		mdelay(1);
+	}
+
+	timeout = jiffies;
+	if (!cmd->data && cmd->busy_timeout > 9000)
+		timeout += DIV_ROUND_UP(cmd->busy_timeout, 1000) * HZ + HZ;
+	else
+		timeout += 10 * HZ;
+	mod_timer(&host->timer, timeout);
+
+	host->cmd = cmd;
+	host->busy_handle = 0;
+
+	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)) {
+		pr_err("%s: Unsupported response type!\n",
+			mmc_hostname(host->mmc));
+		cmd->error = -EINVAL;
+		tasklet_schedule(&host->finish_tasklet);
+		return;
+	}
+
+	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;
+
+	sdhci_writew(host, SDHCI_MAKE_CMD(cmd->opcode, flags), SDHCI_COMMAND);
+}
+EXPORT_SYMBOL_GPL(sdhci_send_command);
+
+static void sdhci_finish_command(struct sdhci_host *host)
+{
+	int i;
+
+	BUG_ON(host->cmd == NULL);
+
+	if (host->cmd->flags & MMC_RSP_PRESENT) {
+		if (host->cmd->flags & MMC_RSP_136) {
+			/* CRC is stripped so we need to do some shifting. */
+			for (i = 0;i < 4;i++) {
+				host->cmd->resp[i] = sdhci_readl(host,
+					SDHCI_RESPONSE + (3-i)*4) << 8;
+				if (i != 3)
+					host->cmd->resp[i] |=
+						sdhci_readb(host,
+						SDHCI_RESPONSE + (3-i)*4-1);
+			}
+		} else {
+			host->cmd->resp[0] = sdhci_readl(host, SDHCI_RESPONSE);
+		}
+	}
+
+	host->cmd->error = 0;
+
+	/* Finished CMD23, now send actual command. */
+	if (host->cmd == host->mrq->sbc) {
+		host->cmd = NULL;
+		sdhci_send_command(host, host->mrq->cmd);
+	} else {
+
+		/* Processed actual command. */
+		if (host->data && host->data_early)
+			sdhci_finish_data(host);
+
+		if (!host->cmd->data)
+			tasklet_schedule(&host->finish_tasklet);
+
+		host->cmd = NULL;
+	}
+}
+
+static u16 sdhci_get_preset_value(struct sdhci_host *host)
+{
+	u16 preset = 0;
+
+	switch (host->timing) {
+	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:
+		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;
+}
+
+void sdhci_set_clock(struct sdhci_host *host, unsigned int clock)
+{
+	int div = 0; /* Initialized for compiler warning */
+	int real_div = div, clk_mul = 1;
+	u16 clk = 0;
+	unsigned long timeout;
+
+	host->mmc->actual_clock = 0;
+
+	sdhci_writew(host, 0, SDHCI_CLOCK_CONTROL);
+
+	if (clock == 0)
+		return;
+
+	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 = (pre_val & SDHCI_PRESET_SDCLK_FREQ_MASK)
+				>> SDHCI_PRESET_SDCLK_FREQ_SHIFT;
+			if (host->clk_mul &&
+				(pre_val & SDHCI_PRESET_CLKGEN_SEL_MASK)) {
+				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;
+			}
+			/*
+			 * Set Programmable Clock Mode in the Clock
+			 * Control register.
+			 */
+			clk = SDHCI_PROG_CLOCK_MODE;
+			real_div = div;
+			clk_mul = host->clk_mul;
+			div--;
+		} else {
+			/* 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;
+		}
+	} 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)
+		host->mmc->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;
+	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) {
+			pr_err("%s: Internal clock never "
+				"stabilised.\n", mmc_hostname(host->mmc));
+			sdhci_dumpregs(host);
+			return;
+		}
+		timeout--;
+		mdelay(1);
+	}
+
+	clk |= SDHCI_CLOCK_CARD_EN;
+	sdhci_writew(host, clk, SDHCI_CLOCK_CONTROL);
+}
+EXPORT_SYMBOL_GPL(sdhci_set_clock);
+
+static void sdhci_set_power(struct sdhci_host *host, unsigned char mode,
+			    unsigned short vdd)
+{
+	struct mmc_host *mmc = host->mmc;
+	u8 pwr = 0;
+
+	if (!IS_ERR(mmc->supply.vmmc)) {
+		spin_unlock_irq(&host->lock);
+		mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, vdd);
+		spin_lock_irq(&host->lock);
+
+		if (mode != MMC_POWER_OFF)
+			sdhci_writeb(host, SDHCI_POWER_ON, SDHCI_POWER_CONTROL);
+		else
+			sdhci_writeb(host, 0, SDHCI_POWER_CONTROL);
+
+		return;
+	}
+
+	if (mode != MMC_POWER_OFF) {
+		switch (1 << vdd) {
+		case MMC_VDD_165_195:
+			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:
+			pwr = SDHCI_POWER_330;
+			break;
+		default:
+			BUG();
+		}
+	}
+
+	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);
+		vdd = 0;
+	} 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);
+	}
+}
+
+/*****************************************************************************\
+ *                                                                           *
+ * MMC callbacks                                                             *
+ *                                                                           *
+\*****************************************************************************/
+
+static void sdhci_request(struct mmc_host *mmc, struct mmc_request *mrq)
+{
+	struct sdhci_host *host;
+	int present;
+	unsigned long flags;
+	u32 tuning_opcode;
+
+	host = mmc_priv(mmc);
+
+	sdhci_runtime_pm_get(host);
+
+	/* Firstly check card presence */
+	present = sdhci_do_get_cd(host);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	WARN_ON(host->mrq != NULL);
+
+#ifndef SDHCI_USE_LEDS_CLASS
+	sdhci_activate_led(host);
+#endif
+
+	/*
+	 * Ensure we don't send the STOP for non-SET_BLOCK_COUNTED
+	 * requests if Auto-CMD12 is enabled.
+	 */
+	if (!mrq->sbc && (host->flags & SDHCI_AUTO_CMD12)) {
+		if (mrq->stop) {
+			mrq->data->stop = NULL;
+			mrq->stop = NULL;
+		}
+	}
+
+	host->mrq = mrq;
+
+	if (!present || host->flags & SDHCI_DEVICE_DEAD) {
+		host->mrq->cmd->error = -ENOMEDIUM;
+		tasklet_schedule(&host->finish_tasklet);
+	} else {
+		u32 present_state;
+
+		present_state = sdhci_readl(host, SDHCI_PRESENT_STATE);
+		/*
+		 * Check if the re-tuning timer has already expired and there
+		 * is no on-going data transfer and DAT0 is not busy. If so,
+		 * we need to execute tuning procedure before sending command.
+		 */
+		if ((host->flags & SDHCI_NEEDS_RETUNING) &&
+		    !(present_state & (SDHCI_DOING_WRITE | SDHCI_DOING_READ)) &&
+		    (present_state & SDHCI_DATA_0_LVL_MASK)) {
+			if (mmc->card) {
+				/* eMMC uses cmd21 but sd and sdio use cmd19 */
+				tuning_opcode =
+					mmc->card->type == MMC_TYPE_MMC ?
+					MMC_SEND_TUNING_BLOCK_HS200 :
+					MMC_SEND_TUNING_BLOCK;
+
+				/* Here we need to set the host->mrq to NULL,
+				 * in case the pending finish_tasklet
+				 * finishes it incorrectly.
+				 */
+				host->mrq = NULL;
+
+				spin_unlock_irqrestore(&host->lock, flags);
+				sdhci_execute_tuning(mmc, tuning_opcode);
+				spin_lock_irqsave(&host->lock, flags);
+
+				/* Restore original mmc_request structure */
+				host->mrq = mrq;
+			}
+		}
+
+		if (mrq->sbc && !(host->flags & SDHCI_AUTO_CMD23))
+			sdhci_send_command(host, mrq->sbc);
+		else
+			sdhci_send_command(host, mrq->cmd);
+	}
+
+	mmiowb();
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+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;
+		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;
+	}
+	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 void sdhci_do_set_ios(struct sdhci_host *host, struct mmc_ios *ios)
+{
+	unsigned long flags;
+	u8 ctrl;
+	struct mmc_host *mmc = host->mmc;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (host->flags & SDHCI_DEVICE_DEAD) {
+		spin_unlock_irqrestore(&host->lock, flags);
+		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) {
+		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;
+		}
+	}
+
+	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);
+
+	ctrl = sdhci_readb(host, SDHCI_HOST_CONTROL);
+
+	if ((ios->timing == MMC_TIMING_SD_HS ||
+	     ios->timing == MMC_TIMING_MMC_HS)
+	    && !(host->quirks & SDHCI_QUIRK_NO_HISPD_BIT))
+		ctrl |= SDHCI_CTRL_HISPD;
+	else
+		ctrl &= ~SDHCI_CTRL_HISPD;
+
+	if (host->version >= SDHCI_SPEC_300) {
+		u16 clk, ctrl_2;
+
+		/* In case of UHS-I modes, set High Speed Enable */
+		if ((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;
+
+		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_C)
+				ctrl_2 |= SDHCI_CTRL_DRV_TYPE_C;
+
+			sdhci_writew(host, ctrl_2, SDHCI_HOST_CONTROL2);
+		} 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 (!(host->quirks2 & SDHCI_QUIRK2_PRESET_VALUE_BROKEN) &&
+				((ios->timing == MMC_TIMING_UHS_SDR12) ||
+				 (ios->timing == MMC_TIMING_UHS_SDR25) ||
+				 (ios->timing == MMC_TIMING_UHS_SDR50) ||
+				 (ios->timing == MMC_TIMING_UHS_SDR104) ||
+				 (ios->timing == MMC_TIMING_UHS_DDR50))) {
+			u16 preset;
+
+			sdhci_enable_preset_value(host, true);
+			preset = sdhci_get_preset_value(host);
+			ios->drv_type = (preset & SDHCI_PRESET_DRV_MASK)
+				>> SDHCI_PRESET_DRV_SHIFT;
+		}
+
+		/* 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);
+
+	mmiowb();
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void sdhci_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	sdhci_runtime_pm_get(host);
+	sdhci_do_set_ios(host, ios);
+	sdhci_runtime_pm_put(host);
+}
+
+static int sdhci_do_get_cd(struct sdhci_host *host)
+{
+	int gpio_cd = mmc_gpio_get_cd(host->mmc);
+
+	if (host->flags & SDHCI_DEVICE_DEAD)
+		return 0;
+
+	/* If polling/nonremovable, assume that the card is always present. */
+	if ((host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) ||
+	    (host->mmc->caps & MMC_CAP_NONREMOVABLE))
+		return 1;
+
+	/* Try slot gpio detect */
+	if (!IS_ERR_VALUE(gpio_cd))
+		return !!gpio_cd;
+
+	/* Host native card detect */
+	return !!(sdhci_readl(host, SDHCI_PRESENT_STATE) & SDHCI_CARD_PRESENT);
+}
+
+static int sdhci_get_cd(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	int ret;
+
+	sdhci_runtime_pm_get(host);
+	ret = sdhci_do_get_cd(host);
+	sdhci_runtime_pm_put(host);
+	return ret;
+}
+
+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
+		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_do_get_ro(struct sdhci_host *host)
+{
+	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 int sdhci_get_ro(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	int ret;
+
+	sdhci_runtime_pm_get(host);
+	ret = sdhci_do_get_ro(host);
+	sdhci_runtime_pm_put(host);
+	return ret;
+}
+
+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);
+		mmiowb();
+	}
+}
+
+static void sdhci_enable_sdio_irq(struct mmc_host *mmc, int enable)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	unsigned long flags;
+
+	sdhci_runtime_pm_get(host);
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (enable)
+		host->flags |= SDHCI_SDIO_IRQ_ENABLED;
+	else
+		host->flags &= ~SDHCI_SDIO_IRQ_ENABLED;
+
+	sdhci_enable_sdio_irq_nolock(host, enable);
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	sdhci_runtime_pm_put(host);
+}
+
+static int sdhci_do_start_signal_voltage_switch(struct sdhci_host *host,
+						struct mmc_ios *ios)
+{
+	struct mmc_host *mmc = host->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:
+		/* 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 = regulator_set_voltage(mmc->supply.vqmmc, 2700000,
+						    3600000);
+			if (ret) {
+				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 became stable\n",
+			mmc_hostname(mmc));
+
+		return -EAGAIN;
+	case MMC_SIGNAL_VOLTAGE_180:
+		if (!IS_ERR(mmc->supply.vqmmc)) {
+			ret = regulator_set_voltage(mmc->supply.vqmmc,
+					1700000, 1950000);
+			if (ret) {
+				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 became stable\n",
+			mmc_hostname(mmc));
+
+		return -EAGAIN;
+	case MMC_SIGNAL_VOLTAGE_120:
+		if (!IS_ERR(mmc->supply.vqmmc)) {
+			ret = regulator_set_voltage(mmc->supply.vqmmc, 1100000,
+						    1300000);
+			if (ret) {
+				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;
+	}
+}
+
+static int sdhci_start_signal_voltage_switch(struct mmc_host *mmc,
+	struct mmc_ios *ios)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	int err;
+
+	if (host->version < SDHCI_SPEC_300)
+		return 0;
+	sdhci_runtime_pm_get(host);
+	err = sdhci_do_start_signal_voltage_switch(host, ios);
+	sdhci_runtime_pm_put(host);
+	return err;
+}
+
+static int sdhci_card_busy(struct mmc_host *mmc)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	u32 present_state;
+
+	sdhci_runtime_pm_get(host);
+	/* Check whether DAT[3:0] is 0000 */
+	present_state = sdhci_readl(host, SDHCI_PRESENT_STATE);
+	sdhci_runtime_pm_put(host);
+
+	return !(present_state & SDHCI_DATA_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;
+}
+
+static int sdhci_execute_tuning(struct mmc_host *mmc, u32 opcode)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+	u16 ctrl;
+	int tuning_loop_counter = MAX_TUNING_LOOP;
+	int err = 0;
+	unsigned long flags;
+	unsigned int tuning_count = 0;
+	bool hs400_tuning;
+
+	sdhci_runtime_pm_get(host);
+	spin_lock_irqsave(&host->lock, flags);
+
+	hs400_tuning = host->flags & SDHCI_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 only in case of SDR104 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_unlock;
+
+	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:
+		break;
+
+	case MMC_TIMING_UHS_SDR50:
+		if (host->flags & SDHCI_SDR50_NEEDS_TUNING ||
+		    host->flags & SDHCI_SDR104_NEEDS_TUNING)
+			break;
+		/* FALLTHROUGH */
+
+	default:
+		goto out_unlock;
+	}
+
+	if (host->ops->platform_execute_tuning) {
+		spin_unlock_irqrestore(&host->lock, flags);
+		err = host->ops->platform_execute_tuning(host, opcode);
+		sdhci_runtime_pm_put(host);
+		return err;
+	}
+
+	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);
+
+	/*
+	 * Issue CMD19 repeatedly till Execute Tuning is set to 0 or the number
+	 * of loops reaches 40 times or a timeout of 150ms occurs.
+	 */
+	do {
+		struct mmc_command cmd = {0};
+		struct mmc_request mrq = {NULL};
+
+		cmd.opcode = opcode;
+		cmd.arg = 0;
+		cmd.flags = MMC_RSP_R1 | MMC_CMD_ADTC;
+		cmd.retries = 0;
+		cmd.data = NULL;
+		cmd.error = 0;
+
+		if (tuning_loop_counter-- == 0)
+			break;
+
+		mrq.cmd = &cmd;
+		host->mrq = &mrq;
+
+		/*
+		 * 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) {
+			if (mmc->ios.bus_width == MMC_BUS_WIDTH_8)
+				sdhci_writew(host, SDHCI_MAKE_BLKSZ(7, 128),
+					     SDHCI_BLOCK_SIZE);
+			else if (mmc->ios.bus_width == MMC_BUS_WIDTH_4)
+				sdhci_writew(host, SDHCI_MAKE_BLKSZ(7, 64),
+					     SDHCI_BLOCK_SIZE);
+		} else {
+			sdhci_writew(host, SDHCI_MAKE_BLKSZ(7, 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);
+
+		sdhci_send_command(host, &cmd);
+
+		host->cmd = NULL;
+		host->mrq = NULL;
+
+		spin_unlock_irqrestore(&host->lock, flags);
+		/* Wait for Buffer Read Ready interrupt */
+		wait_event_interruptible_timeout(host->buf_ready_int,
+					(host->tuning_done == 1),
+					msecs_to_jiffies(50));
+		spin_lock_irqsave(&host->lock, flags);
+
+		if (!host->tuning_done) {
+			pr_info(DRIVER_NAME ": Timeout waiting for "
+				"Buffer Read Ready interrupt during tuning "
+				"procedure, falling back to fixed sampling "
+				"clock\n");
+			ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+			ctrl &= ~SDHCI_CTRL_TUNED_CLK;
+			ctrl &= ~SDHCI_CTRL_EXEC_TUNING;
+			sdhci_writew(host, ctrl, SDHCI_HOST_CONTROL2);
+
+			err = -EIO;
+			goto out;
+		}
+
+		host->tuning_done = 0;
+
+		ctrl = sdhci_readw(host, SDHCI_HOST_CONTROL2);
+
+		/* eMMC spec does not require a delay between tuning cycles */
+		if (opcode == MMC_SEND_TUNING_BLOCK)
+			mdelay(1);
+	} while (ctrl & SDHCI_CTRL_EXEC_TUNING);
+
+	/*
+	 * The Host Driver has exhausted the maximum number of loops allowed,
+	 * so use fixed sampling frequency.
+	 */
+	if (tuning_loop_counter < 0) {
+		ctrl &= ~SDHCI_CTRL_TUNED_CLK;
+		sdhci_writew(host, ctrl, SDHCI_HOST_CONTROL2);
+	}
+	if (!(ctrl & SDHCI_CTRL_TUNED_CLK)) {
+		pr_info(DRIVER_NAME ": Tuning procedure"
+			" failed, falling back to fixed sampling"
+			" clock\n");
+		err = -EIO;
+	}
+
+out:
+	host->flags &= ~SDHCI_NEEDS_RETUNING;
+
+	if (tuning_count) {
+		host->flags |= SDHCI_USING_RETUNING_TIMER;
+		mod_timer(&host->tuning_timer, jiffies + tuning_count * HZ);
+	}
+
+	/*
+	 * In case tuning fails, host controllers which support re-tuning can
+	 * try tuning again at a later time, when the re-tuning timer expires.
+	 * So for these controllers, we return 0. Since there might be other
+	 * controllers who do not have this capability, we return error for
+	 * them. SDHCI_USING_RETUNING_TIMER means the host is currently using
+	 * a retuning timer to do the retuning for the card.
+	 */
+	if (err && (host->flags & SDHCI_USING_RETUNING_TIMER))
+		err = 0;
+
+	sdhci_writel(host, host->ier, SDHCI_INT_ENABLE);
+	sdhci_writel(host, host->ier, SDHCI_SIGNAL_ENABLE);
+out_unlock:
+	spin_unlock_irqrestore(&host->lock, flags);
+	sdhci_runtime_pm_put(host);
+
+	return err;
+}
+
+
+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 (host->flags & SDHCI_REQ_USE_DMA) {
+		if (data->host_cookie)
+			dma_unmap_sg(mmc_dev(host->mmc), data->sg, data->sg_len,
+					 data->flags & MMC_DATA_WRITE ?
+					 DMA_TO_DEVICE : DMA_FROM_DEVICE);
+		mrq->data->host_cookie = 0;
+	}
+}
+
+static int sdhci_pre_dma_transfer(struct sdhci_host *host,
+				       struct mmc_data *data,
+				       struct sdhci_host_next *next)
+{
+	int sg_count;
+
+	if (!next && data->host_cookie &&
+	    data->host_cookie != host->next_data.cookie) {
+		pr_debug(DRIVER_NAME "[%s] invalid cookie: %d, next-cookie %d\n",
+			__func__, data->host_cookie, host->next_data.cookie);
+		data->host_cookie = 0;
+	}
+
+	/* Check if next job is already prepared */
+	if (next ||
+	    (!next && data->host_cookie != host->next_data.cookie)) {
+		sg_count = dma_map_sg(mmc_dev(host->mmc), data->sg,
+				     data->sg_len,
+				     data->flags & MMC_DATA_WRITE ?
+				     DMA_TO_DEVICE : DMA_FROM_DEVICE);
+
+	} else {
+		sg_count = host->next_data.sg_count;
+		host->next_data.sg_count = 0;
+	}
+
+
+	if (sg_count == 0)
+		return -EINVAL;
+
+	if (next) {
+		next->sg_count = sg_count;
+		data->host_cookie = ++next->cookie < 0 ? 1 : next->cookie;
+	} else
+		host->sg_count = sg_count;
+
+	return sg_count;
+}
+
+static void sdhci_pre_req(struct mmc_host *mmc, struct mmc_request *mrq,
+			       bool is_first_req)
+{
+	struct sdhci_host *host = mmc_priv(mmc);
+
+	if (mrq->data->host_cookie) {
+		mrq->data->host_cookie = 0;
+		return;
+	}
+
+	if (host->flags & SDHCI_REQ_USE_DMA)
+		if (sdhci_pre_dma_transfer(host,
+					mrq->data,
+					&host->next_data) < 0)
+			mrq->data->host_cookie = 0;
+}
+
+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 = sdhci_do_get_cd(host);
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	/* Check host->mrq first in case we are runtime suspended */
+	if (host->mrq && !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);
+
+		host->mrq->cmd->error = -ENOMEDIUM;
+		tasklet_schedule(&host->finish_tasklet);
+	}
+
+	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,
+	.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,
+};
+
+/*****************************************************************************\
+ *                                                                           *
+ * Tasklets                                                                  *
+ *                                                                           *
+\*****************************************************************************/
+
+static void sdhci_tasklet_finish(unsigned long param)
+{
+	struct sdhci_host *host;
+	unsigned long flags;
+	struct mmc_request *mrq;
+
+	host = (struct sdhci_host*)param;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+        /*
+         * If this tasklet gets rescheduled while running, it will
+         * be run again afterwards but without any active request.
+         */
+	if (!host->mrq) {
+		spin_unlock_irqrestore(&host->lock, flags);
+		return;
+	}
+
+	del_timer(&host->timer);
+
+	mrq = host->mrq;
+
+	/*
+	 * The controller needs a reset of internal state machines
+	 * upon error conditions.
+	 */
+	if (!(host->flags & SDHCI_DEVICE_DEAD) &&
+	    ((mrq->cmd && mrq->cmd->error) ||
+	     (mrq->sbc && mrq->sbc->error) ||
+	     (mrq->data && ((mrq->data->error && !mrq->data->stop) ||
+			    (mrq->data->stop && mrq->data->stop->error))) ||
+	     (host->quirks & SDHCI_QUIRK_RESET_AFTER_REQUEST))) {
+
+		/* 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->mrq = NULL;
+	host->cmd = NULL;
+	host->data = NULL;
+
+#ifndef SDHCI_USE_LEDS_CLASS
+	sdhci_deactivate_led(host);
+#endif
+
+	mmiowb();
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	mmc_request_done(host->mmc, mrq);
+	sdhci_runtime_pm_put(host);
+}
+
+static void sdhci_timeout_timer(unsigned long data)
+{
+	struct sdhci_host *host;
+	unsigned long flags;
+
+	host = (struct sdhci_host*)data;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	if (host->mrq) {
+		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);
+		} else {
+			if (host->cmd)
+				host->cmd->error = -ETIMEDOUT;
+			else
+				host->mrq->cmd->error = -ETIMEDOUT;
+
+			tasklet_schedule(&host->finish_tasklet);
+		}
+	}
+
+	mmiowb();
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+static void sdhci_tuning_timer(unsigned long data)
+{
+	struct sdhci_host *host;
+	unsigned long flags;
+
+	host = (struct sdhci_host *)data;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	host->flags |= SDHCI_NEEDS_RETUNING;
+
+	spin_unlock_irqrestore(&host->lock, flags);
+}
+
+/*****************************************************************************\
+ *                                                                           *
+ * Interrupt handling                                                        *
+ *                                                                           *
+\*****************************************************************************/
+
+static void sdhci_cmd_irq(struct sdhci_host *host, u32 intmask, u32 *mask)
+{
+	BUG_ON(intmask == 0);
+
+	if (!host->cmd) {
+		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)
+		host->cmd->error = -ETIMEDOUT;
+	else if (intmask & (SDHCI_INT_CRC | SDHCI_INT_END_BIT |
+			SDHCI_INT_INDEX))
+		host->cmd->error = -EILSEQ;
+
+	if (host->cmd->error) {
+		tasklet_schedule(&host->finish_tasklet);
+		return;
+	}
+
+	/*
+	 * The host can send and interrupt when the busy state has
+	 * ended, allowing us to wait without wasting CPU cycles.
+	 * Unfortunately this is overloaded on the "data complete"
+	 * interrupt, so we need to take some care when handling
+	 * it.
+	 *
+	 * Note: The 1.0 specification is a bit ambiguous about this
+	 *       feature so there might be some problems with older
+	 *       controllers.
+	 */
+	if (host->cmd->flags & MMC_RSP_BUSY) {
+		if (host->cmd->data)
+			DBG("Cannot wait for busy signal when also "
+				"doing a data transfer");
+		else if (!(host->quirks & SDHCI_QUIRK_NO_BUSY_IRQ)
+				&& !host->busy_handle) {
+			/* Mark that command complete before busy is ended */
+			host->busy_handle = 1;
+			return;
+		}
+
+		/* The controller does not support the end-of-busy IRQ,
+		 * fall through and take the SDHCI_INT_RESPONSE */
+	} else if ((host->quirks2 & SDHCI_QUIRK2_STOP_WITH_TC) &&
+		   host->cmd->opcode == MMC_STOP_TRANSMISSION && !host->data) {
+		*mask &= ~SDHCI_INT_DATA_END;
+	}
+
+	if (intmask & SDHCI_INT_RESPONSE)
+		sdhci_finish_command(host);
+}
+
+#ifdef CONFIG_MMC_DEBUG
+static void sdhci_adma_show_error(struct sdhci_host *host)
+{
+	const char *name = mmc_hostname(host->mmc);
+	void *desc = host->adma_table;
+
+	sdhci_dumpregs(host);
+
+	while (true) {
+		struct sdhci_adma2_64_desc *dma_desc = desc;
+
+		if (host->flags & SDHCI_USE_64_BIT_DMA)
+			DBG("%s: %p: DMA 0x%08x%08x, LEN 0x%04x, Attr=0x%02x\n",
+			    name, desc, 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
+			DBG("%s: %p: DMA 0x%08x, LEN 0x%04x, Attr=0x%02x\n",
+			    name, desc, le32_to_cpu(dma_desc->addr_lo),
+			    le16_to_cpu(dma_desc->len),
+			    le16_to_cpu(dma_desc->cmd));
+
+		desc += host->desc_sz;
+
+		if (dma_desc->cmd & cpu_to_le16(ADMA2_END))
+			break;
+	}
+}
+#else
+static void sdhci_adma_show_error(struct sdhci_host *host) { }
+#endif
+
+static void sdhci_data_irq(struct sdhci_host *host, u32 intmask)
+{
+	u32 command;
+	BUG_ON(intmask == 0);
+
+	/* 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) {
+		/*
+		 * The "data complete" interrupt is also used to
+		 * indicate that a busy state has ended. See comment
+		 * above in sdhci_cmd_irq().
+		 */
+		if (host->cmd && (host->cmd->flags & MMC_RSP_BUSY)) {
+			if (intmask & SDHCI_INT_DATA_TIMEOUT) {
+				host->cmd->error = -ETIMEDOUT;
+				tasklet_schedule(&host->finish_tasklet);
+				return;
+			}
+			if (intmask & SDHCI_INT_DATA_END) {
+				/*
+				 * Some cards handle busy-end interrupt
+				 * before the command completed, so make
+				 * sure we do things in the proper order.
+				 */
+				if (host->busy_handle)
+					sdhci_finish_command(host);
+				else
+					host->busy_handle = 1;
+				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\n", mmc_hostname(host->mmc));
+		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) {
+			u32 dmastart, dmanow;
+			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;
+			DBG("%s: DMA base 0x%08x, transferred 0x%06x bytes,"
+				" next 0x%08x\n",
+				mmc_hostname(host->mmc), dmastart,
+				host->data->bytes_xfered, dmanow);
+			sdhci_writel(host, dmanow, SDHCI_DMA_ADDRESS);
+		}
+
+		if (intmask & SDHCI_INT_DATA_END) {
+			if (host->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 irqreturn_t sdhci_irq(int irq, void *dev_id)
+{
+	irqreturn_t result = IRQ_NONE;
+	struct sdhci_host *host = dev_id;
+	u32 intmask, mask, unexpected = 0;
+	int max_loops = 16;
+
+	spin_lock(&host->lock);
+
+	if (host->runtime_suspended && !sdhci_sdio_irq_enabled(host)) {
+		spin_unlock(&host->lock);
+		return IRQ_NONE;
+	}
+
+	intmask = sdhci_readl(host, SDHCI_INT_STATUS);
+	if (!intmask || intmask == 0xffffffff) {
+		result = IRQ_NONE;
+		goto out;
+	}
+
+	do {
+		/* Clear selected interrupts. */
+		mask = intmask & (SDHCI_INT_CMD_MASK | SDHCI_INT_DATA_MASK |
+				  SDHCI_INT_BUS_POWER);
+		sdhci_writel(host, mask, SDHCI_INT_STATUS);
+
+		DBG("*** %s got interrupt: 0x%08x\n",
+			mmc_hostname(host->mmc), intmask);
+
+		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_CARD_INT) {
+			sdhci_enable_sdio_irq_nolock(host, false);
+			host->thread_isr |= SDHCI_INT_CARD_INT;
+			result = IRQ_WAKE_THREAD;
+		}
+
+		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_CARD_INT);
+
+		if (intmask) {
+			unexpected |= intmask;
+			sdhci_writel(host, intmask, SDHCI_INT_STATUS);
+		}
+
+		if (result == IRQ_NONE)
+			result = IRQ_HANDLED;
+
+		intmask = sdhci_readl(host, SDHCI_INT_STATUS);
+	} while (intmask && --max_loops);
+out:
+	spin_unlock(&host->lock);
+
+	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;
+	unsigned long flags;
+	u32 isr;
+
+	spin_lock_irqsave(&host->lock, flags);
+	isr = host->thread_isr;
+	host->thread_isr = 0;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (isr & (SDHCI_INT_CARD_INSERT | SDHCI_INT_CARD_REMOVE)) {
+		sdhci_card_event(host->mmc);
+		mmc_detect_change(host->mmc, msecs_to_jiffies(200));
+	}
+
+	if (isr & SDHCI_INT_CARD_INT) {
+		sdio_run_irqs(host->mmc);
+
+		spin_lock_irqsave(&host->lock, flags);
+		if (host->flags & SDHCI_SDIO_IRQ_ENABLED)
+			sdhci_enable_sdio_irq_nolock(host, true);
+		spin_unlock_irqrestore(&host->lock, flags);
+	}
+
+	return isr ? IRQ_HANDLED : IRQ_NONE;
+}
+
+/*****************************************************************************\
+ *                                                                           *
+ * Suspend/resume                                                            *
+ *                                                                           *
+\*****************************************************************************/
+
+#ifdef CONFIG_PM
+void sdhci_enable_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 ;
+	/* Avoid fake wake up */
+	if (host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION)
+		val &= ~(SDHCI_WAKE_ON_INSERT | SDHCI_WAKE_ON_REMOVE);
+	sdhci_writeb(host, val, SDHCI_WAKE_UP_CONTROL);
+}
+EXPORT_SYMBOL_GPL(sdhci_enable_irq_wakeups);
+
+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);
+}
+
+int sdhci_suspend_host(struct sdhci_host *host)
+{
+	sdhci_disable_card_detection(host);
+
+	/* Disable tuning since we are suspending */
+	if (host->flags & SDHCI_USING_RETUNING_TIMER) {
+		del_timer_sync(&host->tuning_timer);
+		host->flags &= ~SDHCI_NEEDS_RETUNING;
+	}
+
+	if (!device_may_wakeup(mmc_dev(host->mmc))) {
+		host->ier = 0;
+		sdhci_writel(host, 0, SDHCI_INT_ENABLE);
+		sdhci_writel(host, 0, SDHCI_SIGNAL_ENABLE);
+		free_irq(host->irq, host);
+	} else {
+		sdhci_enable_irq_wakeups(host);
+		enable_irq_wake(host->irq);
+	}
+	return 0;
+}
+
+EXPORT_SYMBOL_GPL(sdhci_suspend_host);
+
+int sdhci_resume_host(struct sdhci_host *host)
+{
+	int ret = 0;
+
+	if (host->flags & (SDHCI_USE_SDMA | SDHCI_USE_ADMA)) {
+		if (host->ops->enable_dma)
+			host->ops->enable_dma(host);
+	}
+
+	if (!device_may_wakeup(mmc_dev(host->mmc))) {
+		ret = request_threaded_irq(host->irq, sdhci_irq,
+					   sdhci_thread_irq, IRQF_SHARED,
+					   mmc_hostname(host->mmc), host);
+		if (ret)
+			return ret;
+	} else {
+		sdhci_disable_irq_wakeups(host);
+		disable_irq_wake(host->irq);
+	}
+
+	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;
+		sdhci_do_set_ios(host, &host->mmc->ios);
+	} else {
+		sdhci_init(host, (host->mmc->pm_flags & MMC_PM_KEEP_POWER));
+		mmiowb();
+	}
+
+	sdhci_enable_card_detection(host);
+
+	/* Set the re-tuning expiration flag */
+	if (host->flags & SDHCI_USING_RETUNING_TIMER)
+		host->flags |= SDHCI_NEEDS_RETUNING;
+
+	return ret;
+}
+
+EXPORT_SYMBOL_GPL(sdhci_resume_host);
+
+static int sdhci_runtime_pm_get(struct sdhci_host *host)
+{
+	return pm_runtime_get_sync(host->mmc->parent);
+}
+
+static int sdhci_runtime_pm_put(struct sdhci_host *host)
+{
+	pm_runtime_mark_last_busy(host->mmc->parent);
+	return pm_runtime_put_autosuspend(host->mmc->parent);
+}
+
+static void sdhci_runtime_pm_bus_on(struct sdhci_host *host)
+{
+	if (host->runtime_suspended || 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->runtime_suspended || !host->bus_on)
+		return;
+	host->bus_on = false;
+	pm_runtime_put_noidle(host->mmc->parent);
+}
+
+int sdhci_runtime_suspend_host(struct sdhci_host *host)
+{
+	unsigned long flags;
+
+	/* Disable tuning since we are suspending */
+	if (host->flags & SDHCI_USING_RETUNING_TIMER) {
+		del_timer_sync(&host->tuning_timer);
+		host->flags &= ~SDHCI_NEEDS_RETUNING;
+	}
+
+	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)
+{
+	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, 0);
+
+	/* Force clock and power re-program */
+	host->pwr = 0;
+	host->clock = 0;
+	sdhci_do_start_signal_voltage_switch(host, &host->mmc->ios);
+	sdhci_do_set_ios(host, &host->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);
+	}
+
+	/* Set the re-tuning expiration flag */
+	if (host->flags & SDHCI_USING_RETUNING_TIMER)
+		host->flags |= SDHCI_NEEDS_RETUNING;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	host->runtime_suspended = false;
+
+	/* Enable SDIO IRQ */
+	if (host->flags & SDHCI_SDIO_IRQ_ENABLED)
+		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 */
+
+/*****************************************************************************\
+ *                                                                           *
+ * 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;
+
+	return host;
+}
+
+EXPORT_SYMBOL_GPL(sdhci_alloc_host);
+
+int sdhci_add_host(struct sdhci_host *host)
+{
+	struct mmc_host *mmc;
+	u32 caps[2] = {0, 0};
+	u32 max_current_caps;
+	unsigned int ocr_avail;
+	unsigned int override_timeout_clk;
+	int ret;
+
+	WARN_ON(host == NULL);
+	if (host == NULL)
+		return -EINVAL;
+
+	mmc = host->mmc;
+
+	if (debug_quirks)
+		host->quirks = debug_quirks;
+	if (debug_quirks2)
+		host->quirks2 = debug_quirks2;
+
+	override_timeout_clk = host->timeout_clk;
+
+	sdhci_do_reset(host, SDHCI_RESET_ALL);
+
+	host->version = sdhci_readw(host, SDHCI_HOST_VERSION);
+	host->version = (host->version & SDHCI_SPEC_VER_MASK)
+				>> SDHCI_SPEC_VER_SHIFT;
+	if (host->version > SDHCI_SPEC_300) {
+		pr_err("%s: Unknown controller version (%d). "
+			"You may experience problems.\n", mmc_hostname(mmc),
+			host->version);
+	}
+
+	caps[0] = (host->quirks & SDHCI_QUIRK_MISSING_CAPS) ? host->caps :
+		sdhci_readl(host, SDHCI_CAPABILITIES);
+
+	if (host->version >= SDHCI_SPEC_300)
+		caps[1] = (host->quirks & SDHCI_QUIRK_MISSING_CAPS) ?
+			host->caps1 :
+			sdhci_readl(host, SDHCI_CAPABILITIES_1);
+
+	if (host->quirks & SDHCI_QUIRK_FORCE_DMA)
+		host->flags |= SDHCI_USE_SDMA;
+	else if (!(caps[0] & 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) &&
+		(caps[0] & 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;
+	}
+
+	/*
+	 * It is assumed that a 64-bit capable device has set a 64-bit DMA mask
+	 * and *must* do 64-bit DMA.  A driver has the opportunity to change
+	 * that during the first call to ->enable_dma().  Similarly
+	 * SDHCI_QUIRK2_BROKEN_64_BIT_DMA must be left to the drivers to
+	 * implement.
+	 */
+	if (sdhci_readl(host, SDHCI_CAPABILITIES) & SDHCI_CAN_64BIT)
+		host->flags |= SDHCI_USE_64_BIT_DMA;
+
+	if (host->flags & (SDHCI_USE_SDMA | SDHCI_USE_ADMA)) {
+		if (host->ops->enable_dma) {
+			if (host->ops->enable_dma(host)) {
+				pr_warn("%s: No suitable DMA available - falling back to PIO\n",
+					mmc_hostname(mmc));
+				host->flags &=
+					~(SDHCI_USE_SDMA | SDHCI_USE_ADMA);
+			}
+		}
+	}
+
+	/* SDMA does not support 64-bit DMA */
+	if (host->flags & SDHCI_USE_64_BIT_DMA)
+		host->flags &= ~SDHCI_USE_SDMA;
+
+	if (host->flags & SDHCI_USE_ADMA) {
+		/*
+		 * The DMA descriptor table size 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,
+		 * all multipled by the descriptor size.
+		 */
+		if (host->flags & SDHCI_USE_64_BIT_DMA) {
+			host->adma_table_sz = (SDHCI_MAX_SEGS * 2 + 1) *
+					      SDHCI_ADMA2_64_DESC_SZ;
+			host->align_buffer_sz = SDHCI_MAX_SEGS *
+						SDHCI_ADMA2_64_ALIGN;
+			host->desc_sz = SDHCI_ADMA2_64_DESC_SZ;
+			host->align_sz = SDHCI_ADMA2_64_ALIGN;
+			host->align_mask = SDHCI_ADMA2_64_ALIGN - 1;
+		} else {
+			host->adma_table_sz = (SDHCI_MAX_SEGS * 2 + 1) *
+					      SDHCI_ADMA2_32_DESC_SZ;
+			host->align_buffer_sz = SDHCI_MAX_SEGS *
+						SDHCI_ADMA2_32_ALIGN;
+			host->desc_sz = SDHCI_ADMA2_32_DESC_SZ;
+			host->align_sz = SDHCI_ADMA2_32_ALIGN;
+			host->align_mask = SDHCI_ADMA2_32_ALIGN - 1;
+		}
+		host->adma_table = dma_alloc_coherent(mmc_dev(mmc),
+						      host->adma_table_sz,
+						      &host->adma_addr,
+						      GFP_KERNEL);
+		host->align_buffer = kmalloc(host->align_buffer_sz, GFP_KERNEL);
+		if (!host->adma_table || !host->align_buffer) {
+			dma_free_coherent(mmc_dev(mmc), host->adma_table_sz,
+					  host->adma_table, host->adma_addr);
+			kfree(host->align_buffer);
+			pr_warn("%s: Unable to allocate ADMA buffers - falling back to standard DMA\n",
+				mmc_hostname(mmc));
+			host->flags &= ~SDHCI_USE_ADMA;
+			host->adma_table = NULL;
+			host->align_buffer = NULL;
+		} else if (host->adma_addr & host->align_mask) {
+			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->adma_table_sz,
+					  host->adma_table, host->adma_addr);
+			kfree(host->align_buffer);
+			host->adma_table = NULL;
+			host->align_buffer = NULL;
+		}
+	}
+
+	/*
+	 * 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 = (caps[0] & SDHCI_CLOCK_V3_BASE_MASK)
+			>> SDHCI_CLOCK_BASE_SHIFT;
+	else
+		host->max_clk = (caps[0] & SDHCI_CLOCK_BASE_MASK)
+			>> SDHCI_CLOCK_BASE_SHIFT;
+
+	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));
+			return -ENODEV;
+		}
+		host->max_clk = host->ops->get_max_clock(host);
+	}
+
+	host->next_data.cookie = 1;
+	/*
+	 * In case of Host Controller v3.00, find out whether clock
+	 * multiplier is supported.
+	 */
+	host->clk_mul = (caps[1] & SDHCI_CLOCK_MUL_MASK) >>
+			SDHCI_CLOCK_MUL_SHIFT;
+
+	/*
+	 * 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.
+	 */
+	mmc->ops = &sdhci_ops;
+	mmc->f_max = 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) {
+			mmc->f_min = (host->max_clk * host->clk_mul) / 1024;
+			mmc->f_max = host->max_clk * host->clk_mul;
+		} else
+			mmc->f_min = host->max_clk / SDHCI_MAX_DIV_SPEC_300;
+	} else
+		mmc->f_min = host->max_clk / SDHCI_MAX_DIV_SPEC_200;
+
+	if (!(host->quirks & SDHCI_QUIRK_DATA_TIMEOUT_USES_SDCLK)) {
+		host->timeout_clk = (caps[0] & SDHCI_TIMEOUT_CLK_MASK) >>
+					SDHCI_TIMEOUT_CLK_SHIFT;
+		if (host->timeout_clk == 0) {
+			if (host->ops->get_timeout_clock) {
+				host->timeout_clk =
+					host->ops->get_timeout_clock(host);
+			} else {
+				pr_err("%s: Hardware doesn't specify timeout clock frequency.\n",
+					mmc_hostname(mmc));
+				return -ENODEV;
+			}
+		}
+
+		if (caps[0] & SDHCI_TIMEOUT_CLK_UNIT)
+			host->timeout_clk *= 1000;
+
+		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 (override_timeout_clk)
+		host->timeout_clk = override_timeout_clk;
+
+	mmc->caps |= MMC_CAP_SDIO_IRQ | MMC_CAP_ERASE | MMC_CAP_CMD23;
+	mmc->caps2 |= MMC_CAP2_SDIO_IRQ_NOTHREAD;
+
+	if (host->quirks & SDHCI_QUIRK_MULTIBLOCK_READ_ACMD12)
+		host->flags |= SDHCI_AUTO_CMD12;
+
+	/* Auto-CMD23 stuff only works in ADMA or PIO. */
+	if ((host->version >= SDHCI_SPEC_300) &&
+	    ((host->flags & SDHCI_USE_ADMA) ||
+	     !(host->flags & SDHCI_USE_SDMA)) &&
+	     !(host->quirks2 & SDHCI_QUIRK2_ACMD23_BROKEN)) {
+		host->flags |= SDHCI_AUTO_CMD23;
+		DBG("%s: Auto-CMD23 available\n", mmc_hostname(mmc));
+	} else {
+		DBG("%s: Auto-CMD23 unavailable\n", mmc_hostname(mmc));
+	}
+
+	/*
+	 * 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 (caps[0] & SDHCI_CAN_DO_HISPD)
+		mmc->caps |= MMC_CAP_SD_HIGHSPEED | MMC_CAP_MMC_HIGHSPEED;
+
+	if ((host->quirks & SDHCI_QUIRK_BROKEN_CARD_DETECTION) &&
+	    !(mmc->caps & MMC_CAP_NONREMOVABLE))
+		mmc->caps |= MMC_CAP_NEEDS_POLL;
+
+	/* If there are external regulators, get them */
+	if (mmc_regulator_get_supply(mmc) == -EPROBE_DEFER)
+		return -EPROBE_DEFER;
+
+	/* If vqmmc regulator and no 1.8V signalling, then there's no UHS */
+	if (!IS_ERR(mmc->supply.vqmmc)) {
+		ret = regulator_enable(mmc->supply.vqmmc);
+		if (!regulator_is_supported_voltage(mmc->supply.vqmmc, 1700000,
+						    1950000))
+			caps[1] &= ~(SDHCI_SUPPORT_SDR104 |
+					SDHCI_SUPPORT_SDR50 |
+					SDHCI_SUPPORT_DDR50);
+		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)
+		caps[1] &= ~(SDHCI_SUPPORT_SDR104 | SDHCI_SUPPORT_SDR50 |
+		       SDHCI_SUPPORT_DDR50);
+
+	/* Any UHS-I mode in caps implies SDR12 and SDR25 support. */
+	if (caps[1] & (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 (caps[1] & 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 (caps[1] & SDHCI_SUPPORT_SDR50)
+		mmc->caps |= MMC_CAP_UHS_SDR50;
+
+	if (host->quirks2 & SDHCI_QUIRK2_CAPS_BIT63_FOR_HS400 &&
+	    (caps[1] & 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 ((caps[1] & SDHCI_SUPPORT_DDR50) &&
+		!(host->quirks2 & SDHCI_QUIRK2_BROKEN_DDR50))
+		mmc->caps |= MMC_CAP_UHS_DDR50;
+
+	/* Does the host need tuning for SDR50? */
+	if (caps[1] & SDHCI_USE_SDR50_TUNING)
+		host->flags |= SDHCI_SDR50_NEEDS_TUNING;
+
+	/* Does the host need tuning for SDR104 / HS200? */
+	if (mmc->caps2 & MMC_CAP2_HS200)
+		host->flags |= SDHCI_SDR104_NEEDS_TUNING;
+
+	/* Driver Type(s) (A, C, D) supported by the host */
+	if (caps[1] & SDHCI_DRIVER_TYPE_A)
+		mmc->caps |= MMC_CAP_DRIVER_TYPE_A;
+	if (caps[1] & SDHCI_DRIVER_TYPE_C)
+		mmc->caps |= MMC_CAP_DRIVER_TYPE_C;
+	if (caps[1] & SDHCI_DRIVER_TYPE_D)
+		mmc->caps |= MMC_CAP_DRIVER_TYPE_D;
+
+	/* Initial value for re-tuning timer count */
+	host->tuning_count = (caps[1] & SDHCI_RETUNING_TIMER_COUNT_MASK) >>
+			      SDHCI_RETUNING_TIMER_COUNT_SHIFT;
+
+	/*
+	 * 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 = (caps[1] & SDHCI_RETUNING_MODE_MASK) >>
+			     SDHCI_RETUNING_MODE_SHIFT;
+
+	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 =
+				(curr << SDHCI_MAX_CURRENT_330_SHIFT) |
+				(curr << SDHCI_MAX_CURRENT_300_SHIFT) |
+				(curr << SDHCI_MAX_CURRENT_180_SHIFT);
+		}
+	}
+
+	if (caps[0] & SDHCI_CAN_VDD_330) {
+		ocr_avail |= MMC_VDD_32_33 | MMC_VDD_33_34;
+
+		mmc->max_current_330 = ((max_current_caps &
+				   SDHCI_MAX_CURRENT_330_MASK) >>
+				   SDHCI_MAX_CURRENT_330_SHIFT) *
+				   SDHCI_MAX_CURRENT_MULTIPLIER;
+	}
+	if (caps[0] & SDHCI_CAN_VDD_300) {
+		ocr_avail |= MMC_VDD_29_30 | MMC_VDD_30_31;
+
+		mmc->max_current_300 = ((max_current_caps &
+				   SDHCI_MAX_CURRENT_300_MASK) >>
+				   SDHCI_MAX_CURRENT_300_SHIFT) *
+				   SDHCI_MAX_CURRENT_MULTIPLIER;
+	}
+	if (caps[0] & SDHCI_CAN_VDD_180) {
+		ocr_avail |= MMC_VDD_165_195;
+
+		mmc->max_current_180 = ((max_current_caps &
+				   SDHCI_MAX_CURRENT_180_MASK) >>
+				   SDHCI_MAX_CURRENT_180_SHIFT) *
+				   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));
+		return -ENODEV;
+	}
+
+	spin_lock_init(&host->lock);
+
+	/*
+	 * 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;
+	else /* PIO */
+		mmc->max_segs = SDHCI_MAX_SEGS;
+
+	/*
+	 * 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 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)
+			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 = (caps[0] & 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;
+
+	/*
+	 * Init tasklets.
+	 */
+	tasklet_init(&host->finish_tasklet,
+		sdhci_tasklet_finish, (unsigned long)host);
+
+	setup_timer(&host->timer, sdhci_timeout_timer, (unsigned long)host);
+
+	init_waitqueue_head(&host->buf_ready_int);
+
+	if (host->version >= SDHCI_SPEC_300) {
+		/* Initialize re-tuning timer */
+		init_timer(&host->tuning_timer);
+		host->tuning_timer.data = (unsigned long)host;
+		host->tuning_timer.function = sdhci_tuning_timer;
+	}
+
+	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 untasklet;
+	}
+
+#ifdef CONFIG_MMC_DEBUG
+	sdhci_dumpregs(host);
+#endif
+
+#ifdef SDHCI_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 = sdhci_led_control;
+
+	ret = led_classdev_register(mmc_dev(mmc), &host->led);
+	if (ret) {
+		pr_err("%s: Failed to register LED device: %d\n",
+		       mmc_hostname(mmc), ret);
+		goto reset;
+	}
+#endif
+
+	mmiowb();
+
+	mmc_add_host(mmc);
+
+	pr_info("%s: SDHCI controller on %s [%s] using %s\n",
+		mmc_hostname(mmc), host->hw_name, dev_name(mmc_dev(mmc)),
+		(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;
+
+#ifdef SDHCI_USE_LEDS_CLASS
+reset:
+	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);
+#endif
+untasklet:
+	tasklet_kill(&host->finish_tasklet);
+
+	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 (host->mrq) {
+			pr_err("%s: Controller removed during "
+				" transfer!\n", mmc_hostname(mmc));
+
+			host->mrq->cmd->error = -ENOMEDIUM;
+			tasklet_schedule(&host->finish_tasklet);
+		}
+
+		spin_unlock_irqrestore(&host->lock, flags);
+	}
+
+	sdhci_disable_card_detection(host);
+
+	mmc_remove_host(mmc);
+
+#ifdef SDHCI_USE_LEDS_CLASS
+	led_classdev_unregister(&host->led);
+#endif
+
+	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);
+
+	tasklet_kill(&host->finish_tasklet);
+
+	if (!IS_ERR(mmc->supply.vqmmc))
+		regulator_disable(mmc->supply.vqmmc);
+
+	if (host->adma_table)
+		dma_free_coherent(mmc_dev(mmc), host->adma_table_sz,
+				  host->adma_table, host->adma_addr);
+	kfree(host->align_buffer);
+
+	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..e639b7f43
--- /dev/null
+++ b/drivers/mmc/host/sdhci.h
@@ -0,0 +1,663 @@
+/*
+ *  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.
+ *
+ * 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; either version 2 of the License, or (at
+ * your option) any later version.
+ */
+#ifndef __SDHCI_HW_H
+#define __SDHCI_HW_H
+
+#include <linux/scatterlist.h>
+#include <linux/compiler.h>
+#include <linux/types.h>
+#include <linux/io.h>
+
+#include <linux/mmc/host.h>
+
+/*
+ * Controller registers
+ */
+
+#define SDHCI_DMA_ADDRESS	0x00
+#define SDHCI_ARGUMENT2		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_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_WRITE_PROTECT	0x00080000
+#define  SDHCI_DATA_LVL_MASK	0x00F00000
+#define   SDHCI_DATA_LVL_SHIFT	20
+#define   SDHCI_DATA_0_LVL_MASK	0x00100000
+
+#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_8BITBUS	0x20
+
+#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_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_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_ACMD12ERR	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)
+#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_ACMD12_ERR	0x3C
+
+#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_CTRL_PRESET_VAL_ENABLE	0x8000
+
+#define SDHCI_CAPABILITIES	0x40
+#define  SDHCI_TIMEOUT_CLK_MASK	0x0000003F
+#define  SDHCI_TIMEOUT_CLK_SHIFT 0
+#define  SDHCI_TIMEOUT_CLK_UNIT	0x00000080
+#define  SDHCI_CLOCK_BASE_MASK	0x00003F00
+#define  SDHCI_CLOCK_V3_BASE_MASK	0x0000FF00
+#define  SDHCI_CLOCK_BASE_SHIFT	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_VDD_330	0x01000000
+#define  SDHCI_CAN_VDD_300	0x02000000
+#define  SDHCI_CAN_VDD_180	0x04000000
+#define  SDHCI_CAN_64BIT	0x10000000
+
+#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	0x00000F00
+#define  SDHCI_RETUNING_TIMER_COUNT_SHIFT	8
+#define  SDHCI_USE_SDR50_TUNING			0x00002000
+#define  SDHCI_RETUNING_MODE_MASK		0x0000C000
+#define  SDHCI_RETUNING_MODE_SHIFT		14
+#define  SDHCI_CLOCK_MUL_MASK	0x00FF0000
+#define  SDHCI_CLOCK_MUL_SHIFT	16
+#define  SDHCI_SUPPORT_HS400	0x80000000 /* Non-standard */
+
+#define SDHCI_CAPABILITIES_1	0x44
+
+#define SDHCI_MAX_CURRENT		0x48
+#define  SDHCI_MAX_CURRENT_LIMIT	0xFF
+#define  SDHCI_MAX_CURRENT_330_MASK	0x0000FF
+#define  SDHCI_MAX_CURRENT_330_SHIFT	0
+#define  SDHCI_MAX_CURRENT_300_MASK	0x00FF00
+#define  SDHCI_MAX_CURRENT_300_SHIFT	8
+#define  SDHCI_MAX_CURRENT_180_MASK	0xFF0000
+#define  SDHCI_MAX_CURRENT_180_SHIFT	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_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  0xC000
+#define SDHCI_PRESET_DRV_SHIFT  14
+#define SDHCI_PRESET_CLKGEN_SEL_MASK   0x400
+#define SDHCI_PRESET_CLKGEN_SEL_SHIFT	10
+#define SDHCI_PRESET_SDCLK_FREQ_MASK   0x3FF
+#define SDHCI_PRESET_SDCLK_FREQ_SHIFT	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
+
+/*
+ * 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 DMA alignment */
+#define SDHCI_ADMA2_32_ALIGN	4
+
+/* ADMA2 32-bit descriptor */
+struct sdhci_adma2_32_desc {
+	__le16	cmd;
+	__le16	len;
+	__le32	addr;
+}  __packed __aligned(SDHCI_ADMA2_32_ALIGN);
+
+/* ADMA2 64-bit DMA descriptor size */
+#define SDHCI_ADMA2_64_DESC_SZ	12
+
+/* ADMA2 64-bit DMA alignment */
+#define SDHCI_ADMA2_64_ALIGN	8
+
+/*
+ * 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.
+ */
+#define SDHCI_MAX_SEGS		128
+
+struct sdhci_host_next {
+	unsigned int	sg_count;
+	s32		cookie;
+};
+
+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 does not like fast PIO transfers */
+#define SDHCI_QUIRK_PIO_NEEDS_DELAY			(1<<18)
+/* 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)
+
+	int irq;		/* Device IRQ */
+	void __iomem *ioaddr;	/* Mapped address */
+
+	const struct sdhci_ops *ops;	/* Low level hw interface */
+
+	/* Internal data */
+	struct mmc_host *mmc;	/* MMC structure */
+	u64 dma_mask;		/* custom DMA mask */
+
+#if defined(CONFIG_LEDS_CLASS) || defined(CONFIG_LEDS_CLASS_MODULE)
+	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_NEEDS_RETUNING	(1<<5)	/* Host needs retuning */
+#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_SDIO_IRQ_ENABLED	(1<<9)	/* SDIO irq enabled */
+#define SDHCI_SDR104_NEEDS_TUNING (1<<10)	/* SDR104/HS200 needs tuning */
+#define SDHCI_USING_RETUNING_TIMER (1<<11)	/* Host is using a retuning timer for the card */
+#define SDHCI_USE_64_BIT_DMA	(1<<12)	/* Use 64-bit DMA */
+#define SDHCI_HS400_TUNING	(1<<13)	/* Tuning for HS400 */
+
+	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 */
+
+	bool runtime_suspended;	/* Host is runtime suspended */
+	bool bus_on;		/* Bus power prevents runtime suspend */
+	bool preset_enabled;	/* Preset is enabled */
+
+	struct mmc_request *mrq;	/* Current request */
+	struct mmc_command *cmd;	/* Current command */
+	struct mmc_data *data;	/* Current data request */
+	unsigned int data_early:1;	/* Data finished before cmd */
+	unsigned int busy_handle:1;	/* Handling the order of Busy-end */
+
+	struct sg_mapping_iter sg_miter;	/* SG state for PIO */
+	unsigned int blocks;	/* remaining PIO blocks */
+
+	int sg_count;		/* Mapped sg entries */
+
+	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 descriptor size */
+	unsigned int align_sz;	/* ADMA alignment */
+	unsigned int align_mask;	/* ADMA alignment mask */
+
+	struct tasklet_struct finish_tasklet;	/* Tasklet structures */
+
+	struct timer_list timer;	/* Timer for timeouts */
+
+	u32 caps;		/* Alternative CAPABILITY_0 */
+	u32 caps1;		/* Alternative CAPABILITY_1 */
+
+	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;
+
+	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 */
+#define SDHCI_TUNING_MODE_1	0
+	struct timer_list	tuning_timer;	/* Timer for tuning */
+
+	struct sdhci_host_next	next_data;
+	unsigned long private[0] ____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);
+
+	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);
+	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	(*platform_init)(struct sdhci_host *host);
+	void    (*card_event)(struct sdhci_host *host);
+	void	(*voltage_switch)(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 */
+
+extern struct sdhci_host *sdhci_alloc_host(struct device *dev,
+	size_t priv_size);
+extern void sdhci_free_host(struct sdhci_host *host);
+
+static inline void *sdhci_priv(struct sdhci_host *host)
+{
+	return (void *)host->private;
+}
+
+extern void sdhci_card_detect(struct sdhci_host *host);
+extern int sdhci_add_host(struct sdhci_host *host);
+extern void sdhci_remove_host(struct sdhci_host *host, int dead);
+extern void sdhci_send_command(struct sdhci_host *host,
+				struct mmc_command *cmd);
+
+static inline bool sdhci_sdio_irq_enabled(struct sdhci_host *host)
+{
+	return !!(host->flags & SDHCI_SDIO_IRQ_ENABLED);
+}
+
+void sdhci_set_clock(struct sdhci_host *host, unsigned int clock);
+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);
+
+#ifdef CONFIG_PM
+extern int sdhci_suspend_host(struct sdhci_host *host);
+extern int sdhci_resume_host(struct sdhci_host *host);
+extern void sdhci_enable_irq_wakeups(struct sdhci_host *host);
+extern int sdhci_runtime_suspend_host(struct sdhci_host *host);
+extern int sdhci_runtime_resume_host(struct sdhci_host *host);
+#endif
+
+#endif /* __SDHCI_HW_H */
diff --git a/drivers/mmc/host/sdhci_f_sdh30.c b/drivers/mmc/host/sdhci_f_sdh30.c
new file mode 100644
index 000000000..2fe8b9148
--- /dev/null
+++ b/drivers/mmc/host/sdhci_f_sdh30.c
@@ -0,0 +1,237 @@
+/*
+ * 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>
+ *
+ * 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 of the License.
+ */
+
+#include <linux/err.h>
+#include <linux/delay.h>
+#include <linux/module.h>
+#include <linux/clk.h>
+
+#include "sdhci-pltfm.h"
+
+/* F_SDH30 extended Controller registers */
+#define F_SDH30_AHB_CONFIG		0x100
+#define  F_SDH30_AHB_BIGED		0x00000040
+#define  F_SDH30_BUSLOCK_DMA		0x00000020
+#define  F_SDH30_BUSLOCK_EN		0x00000010
+#define  F_SDH30_SIN			0x00000008
+#define  F_SDH30_AHB_INCR_16		0x00000004
+#define  F_SDH30_AHB_INCR_8		0x00000002
+#define  F_SDH30_AHB_INCR_4		0x00000001
+
+#define F_SDH30_TUNING_SETTING		0x108
+#define  F_SDH30_CMD_CHK_DIS		0x00010000
+
+#define F_SDH30_IO_CONTROL2		0x114
+#define  F_SDH30_CRES_O_DN		0x00080000
+#define  F_SDH30_MSEL_O_1_8		0x00040000
+
+#define F_SDH30_ESD_CONTROL		0x124
+#define  F_SDH30_EMMC_RST		0x00000002
+#define  F_SDH30_EMMC_HS200		0x01000000
+
+#define F_SDH30_CMD_DAT_DELAY		0x200
+
+#define F_SDH30_MIN_CLOCK		400000
+
+struct f_sdhost_priv {
+	struct clk *clk_iface;
+	struct clk *clk;
+	u32 vendor_hs200;
+	struct device *dev;
+};
+
+void sdhci_f_sdh30_soft_voltage_switch(struct sdhci_host *host)
+{
+	struct f_sdhost_priv *priv = sdhci_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);
+}
+
+unsigned int sdhci_f_sdh30_get_min_clock(struct sdhci_host *host)
+{
+	return F_SDH30_MIN_CLOCK;
+}
+
+void sdhci_f_sdh30_reset(struct sdhci_host *host, u8 mask)
+{
+	if (sdhci_readw(host, SDHCI_CLOCK_CONTROL) == 0)
+		sdhci_writew(host, 0xBC01, SDHCI_CLOCK_CONTROL);
+
+	sdhci_reset(host, mask);
+}
+
+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 int sdhci_f_sdh30_probe(struct platform_device *pdev)
+{
+	struct sdhci_host *host;
+	struct device *dev = &pdev->dev;
+	struct resource *res;
+	int irq, ctrl = 0, ret = 0;
+	struct f_sdhost_priv *priv;
+	u32 reg = 0;
+
+	irq = platform_get_irq(pdev, 0);
+	if (irq < 0) {
+		dev_err(dev, "%s: no irq specified\n", __func__);
+		return irq;
+	}
+
+	host = sdhci_alloc_host(dev, sizeof(struct sdhci_host) +
+						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;
+	host->quirks2 = SDHCI_QUIRK2_SUPPORT_SINGLE |
+			SDHCI_QUIRK2_TUNING_WORK_AROUND;
+
+	ret = mmc_of_parse(host->mmc);
+	if (ret)
+		goto err;
+
+	platform_set_drvdata(pdev, host);
+
+	sdhci_get_of_property(pdev);
+	host->hw_name = "f_sdh30";
+	host->ops = &sdhci_f_sdh30_ops;
+	host->irq = irq;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	host->ioaddr = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(host->ioaddr)) {
+		ret = PTR_ERR(host->ioaddr);
+		goto err;
+	}
+
+	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;
+
+	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_f_sdh30_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 const struct of_device_id f_sdh30_dt_ids[] = {
+	{ .compatible = "fujitsu,mb86s70-sdhci-3.0" },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, f_sdh30_dt_ids);
+
+static struct platform_driver sdhci_f_sdh30_driver = {
+	.driver = {
+		.name = "f_sdh30",
+		.of_match_table = f_sdh30_dt_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/sdricoh_cs.c b/drivers/mmc/host/sdricoh_cs.c
new file mode 100644
index 000000000..b7e305775
--- /dev/null
+++ b/drivers/mmc/host/sdricoh_cs.c
@@ -0,0 +1,552 @@
+/*
+ *  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>
+ *
+ * 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; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ *
+ */
+
+/*
+#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/scatterlist.h>
+
+#include <pcmcia/cistpl.h>
+#include <pcmcia/ds.h>
+#include <linux/io.h>
+
+#include <linux/mmc/host.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 INIT_TIMEOUT      100
+#define CMD_TIMEOUT       100000
+#define TRANSFER_TIMEOUT  100000
+#define BUSY_TIMEOUT      32767
+
+/* 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 int sdricoh_query_status(struct sdricoh_host *host, unsigned int wanted,
+				unsigned int timeout){
+	unsigned int loop;
+	unsigned int status = 0;
+	struct device *dev = host->dev;
+	for (loop = 0; loop < timeout; loop++) {
+		status = sdricoh_readl(host, R21C_STATUS);
+		sdricoh_writel(host, R2E4_STATUS_RESP, status);
+		if (status & wanted)
+			break;
+	}
+
+	if (loop == timeout) {
+		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, unsigned char opcode,
+			   unsigned int arg)
+{
+	unsigned int status;
+	int result = 0;
+	unsigned int loop = 0;
+	/* reset status reg? */
+	sdricoh_writel(host, R21C_STATUS, 0x18);
+	/* fill parameters */
+	sdricoh_writel(host, R204_CMD_ARG, arg);
+	sdricoh_writel(host, R200_CMD, (0x10000 << 8) | opcode);
+	/* wait for command completion */
+	if (opcode) {
+		for (loop = 0; loop < CMD_TIMEOUT; loop++) {
+			status = sdricoh_readl(host, R21C_STATUS);
+			sdricoh_writel(host, R2E4_STATUS_RESP, status);
+			if (status  & STATUS_CMD_FINISHED)
+				break;
+		}
+		/* don't check for timeout in the loop it is not always
+		   reset correctly
+		*/
+		if (loop == CMD_TIMEOUT || status & STATUS_CMD_TIMEOUT)
+			result = -ETIMEDOUT;
+
+	}
+
+	return result;
+
+}
+
+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,
+						TRANSFER_TIMEOUT))
+			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,
+						TRANSFER_TIMEOUT))
+			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);
+		}
+	}
+
+	if (len)
+		return -EIO;
+
+	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;
+	unsigned char opcode = cmd->opcode;
+	int i;
+
+	dev_dbg(dev, "=============================\n");
+	dev_dbg(dev, "sdricoh_request opcode=%i\n", opcode);
+
+	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 == 55)
+		host->app_cmd = 1;
+
+	/* 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, opcode, cmd->arg);
+
+	/* 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,
+					TRANSFER_TIMEOUT)) {
+			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 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 = 0;
+	void __iomem *iobase = NULL;
+	struct mmc_host *mmc = NULL;
+	struct sdricoh_host *host = NULL;
+	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 err;
+	}
+	/* 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 err;
+	}
+	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 err;
+
+	}
+
+	result = mmc_add_host(mmc);
+
+	if (!result) {
+		dev_dbg(dev, "mmc host registered\n");
+		return 0;
+	}
+
+err:
+	if (iobase)
+		pci_iounmap(pci_dev, iobase);
+	if (mmc)
+		mmc_free_host(mmc);
+
+	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..7eff087cf
--- /dev/null
+++ b/drivers/mmc/host/sh_mmcif.c
@@ -0,0 +1,1571 @@
+/*
+ * MMCIF eMMC driver.
+ *
+ * Copyright (C) 2010 Renesas Solutions Corp.
+ * Yusuke Goda <yusuke.goda.sx@renesas.com>
+ *
+ * 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; either version 2 of the License.
+ *
+ *
+ * TODO
+ *  1. DMA
+ *  2. Power management
+ *  3. Handle MMC errors better
+ *
+ */
+
+/*
+ * 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/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"
+#define DRIVER_VERSION	"2010-04-28"
+
+/* 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 /* 52MHz */
+#define CLKDEV_MMC_DATA		20000000 /* 20MHz */
+#define CLKDEV_INIT		400000   /* 400 KHz */
+
+enum mmcif_state {
+	STATE_IDLE,
+	STATE_REQUEST,
+	STATE_IOS,
+	STATE_TIMEOUT,
+};
+
+enum 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,
+};
+
+struct sh_mmcif_host {
+	struct mmc_host *mmc;
+	struct mmc_request *mrq;
+	struct platform_device *pd;
+	struct clk *hclk;
+	unsigned int 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 mmcif_state state;
+	enum mmcif_wait_for wait_for;
+	struct delayed_work timeout_work;
+	size_t blocksize;
+	int sg_idx;
+	int sg_blkidx;
+	bool power;
+	bool card_present;
+	bool ccs_enable;		/* Command Completion Signal support */
+	bool clk_ctrl2_enable;
+	struct mutex thread_lock;
+
+	/* DMA support */
+	struct dma_chan		*chan_rx;
+	struct dma_chan		*chan_tx;
+	struct completion	dma_complete;
+	bool			dma_active;
+};
+
+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 mmcif_dma_complete(void *arg)
+{
+	struct sh_mmcif_host *host = arg;
+	struct mmc_request *mrq = host->mrq;
+
+	dev_dbg(&host->pd->dev, "Command completed\n");
+
+	if (WARN(!mrq || !mrq->data, "%s: NULL data in DMA completion!\n",
+		 dev_name(&host->pd->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;
+	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 = 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(&host->pd->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(&host->pd->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(&host->pd->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;
+	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 = 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(&host->pd->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(&host->pd->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(&host->pd->dev, "%s(): desc %p, cookie %d\n", __func__,
+		desc, cookie);
+}
+
+static struct dma_chan *
+sh_mmcif_request_dma_one(struct sh_mmcif_host *host,
+			 struct sh_mmcif_plat_data *pdata,
+			 enum dma_transfer_direction direction)
+{
+	struct dma_slave_config cfg = { 0, };
+	struct dma_chan *chan;
+	void *slave_data = NULL;
+	struct resource *res;
+	dma_cap_mask_t mask;
+	int ret;
+
+	dma_cap_zero(mask);
+	dma_cap_set(DMA_SLAVE, mask);
+
+	if (pdata)
+		slave_data = direction == DMA_MEM_TO_DEV ?
+			(void *)pdata->slave_id_tx :
+			(void *)pdata->slave_id_rx;
+
+	chan = dma_request_slave_channel_compat(mask, shdma_chan_filter,
+				slave_data, &host->pd->dev,
+				direction == DMA_MEM_TO_DEV ? "tx" : "rx");
+
+	dev_dbg(&host->pd->dev, "%s: %s: got channel %p\n", __func__,
+		direction == DMA_MEM_TO_DEV ? "TX" : "RX", chan);
+
+	if (!chan)
+		return NULL;
+
+	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;
+	}
+
+	ret = dmaengine_slave_config(chan, &cfg);
+	if (ret < 0) {
+		dma_release_channel(chan);
+		return NULL;
+	}
+
+	return chan;
+}
+
+static void sh_mmcif_request_dma(struct sh_mmcif_host *host,
+				 struct sh_mmcif_plat_data *pdata)
+{
+	host->dma_active = false;
+
+	if (pdata) {
+		if (pdata->slave_id_tx <= 0 || pdata->slave_id_rx <= 0)
+			return;
+	} else if (!host->pd->dev.of_node) {
+		return;
+	}
+
+	/* We can only either use DMA for both Tx and Rx or not use it at all */
+	host->chan_tx = sh_mmcif_request_dma_one(host, pdata, DMA_MEM_TO_DEV);
+	if (!host->chan_tx)
+		return;
+
+	host->chan_rx = sh_mmcif_request_dma_one(host, pdata, DMA_DEV_TO_MEM);
+	if (!host->chan_rx) {
+		dma_release_channel(host->chan_tx);
+		host->chan_tx = 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 sh_mmcif_plat_data *p = host->pd->dev.platform_data;
+	bool sup_pclk = p ? p->sup_pclk : false;
+
+	sh_mmcif_bitclr(host, MMCIF_CE_CLK_CTRL, CLK_ENABLE);
+	sh_mmcif_bitclr(host, MMCIF_CE_CLK_CTRL, CLK_CLEAR);
+
+	if (!clk)
+		return;
+	if (sup_pclk && clk == host->clk)
+		sh_mmcif_bitset(host, MMCIF_CE_CLK_CTRL, CLK_SUP_PCLK);
+	else
+		sh_mmcif_bitset(host, MMCIF_CE_CLK_CTRL, CLK_CLEAR &
+				((fls(DIV_ROUND_UP(host->clk,
+						   clk) - 1) - 1) << 16));
+
+	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)
+{
+	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(&host->pd->dev, "ERR HOST_STS1 = %08x\n", state1);
+	dev_dbg(&host->pd->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 = 10000000; timeout; timeout--) {
+			if (!(sh_mmcif_readl(host->addr, MMCIF_CE_HOST_STS1)
+			      & STS1_CMDSEQ))
+				break;
+			mdelay(1);
+		}
+		if (!timeout) {
+			dev_err(&host->pd->dev,
+				"Forced end of command sequence timeout err\n");
+			return -EIO;
+		}
+		sh_mmcif_sync_reset(host);
+		dev_dbg(&host->pd->dev, "Forced end of command sequence\n");
+		return -EIO;
+	}
+
+	if (state2 & STS2_CRC_ERR) {
+		dev_err(&host->pd->dev, " CRC error: state %u, wait %u\n",
+			host->state, host->wait_for);
+		ret = -EIO;
+	} else if (state2 & STS2_TIMEOUT_ERR) {
+		dev_err(&host->pd->dev, " Timeout: state %u, wait %u\n",
+			host->state, host->wait_for);
+		ret = -ETIMEDOUT;
+	} else {
+		dev_dbg(&host->pd->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 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(&host->pd->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 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(&host->pd->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 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(&host->pd->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 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(&host->pd->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 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_R1B:
+	case MMC_RSP_R3:
+		tmp |= CMD_SET_RTYP_6B;
+		break;
+	case MMC_RSP_R2:
+		tmp |= CMD_SET_RTYP_17B;
+		break;
+	default:
+		dev_err(&host->pd->dev, "Unsupported response type.\n");
+		break;
+	}
+	switch (opc) {
+	/* RBSY */
+	case MMC_SLEEP_AWAKE:
+	case MMC_SWITCH:
+	case MMC_STOP_TRANSMISSION:
+	case MMC_SET_WRITE_PROT:
+	case MMC_CLR_WRITE_PROT:
+	case MMC_ERASE:
+		tmp |= CMD_SET_RBSY;
+		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(&host->pd->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)
+{
+	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(&host->pd->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 = cmd->opcode;
+	u32 mask;
+	unsigned long flags;
+
+	switch (opc) {
+	/* response busy check */
+	case MMC_SLEEP_AWAKE:
+	case MMC_SWITCH:
+	case MMC_STOP_TRANSMISSION:
+	case MMC_SET_WRITE_PROT:
+	case MMC_CLR_WRITE_PROT:
+	case MMC_ERASE:
+		mask = MASK_START_CMD | MASK_MRBSYE;
+		break;
+	default:
+		mask = MASK_START_CMD | MASK_MCRSPE;
+		break;
+	}
+
+	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)
+{
+	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(&host->pd->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);
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (host->state != STATE_IDLE) {
+		dev_dbg(&host->pd->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);
+
+	switch (mrq->cmd->opcode) {
+	/* MMCIF does not support SD/SDIO command */
+	case MMC_SLEEP_AWAKE: /* = SD_IO_SEND_OP_COND (5) */
+	case MMC_SEND_EXT_CSD: /* = SD_SEND_IF_COND (8) */
+		if ((mrq->cmd->flags & MMC_CMD_MASK) != MMC_CMD_BCR)
+			break;
+	case MMC_APP_CMD:
+	case SD_IO_RW_DIRECT:
+		host->state = STATE_IDLE;
+		mrq->cmd->error = -ETIMEDOUT;
+		mmc_request_done(mmc, mrq);
+		return;
+	default:
+		break;
+	}
+
+	host->mrq = mrq;
+
+	sh_mmcif_start_cmd(host, mrq);
+}
+
+static int sh_mmcif_clk_update(struct sh_mmcif_host *host)
+{
+	int ret = clk_prepare_enable(host->hclk);
+
+	if (!ret) {
+		host->clk = clk_get_rate(host->hclk);
+		host->mmc->f_max = host->clk / 2;
+		host->mmc->f_min = host->clk / 512;
+	}
+
+	return ret;
+}
+
+static void sh_mmcif_set_power(struct sh_mmcif_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 void sh_mmcif_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sh_mmcif_host *host = mmc_priv(mmc);
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+	if (host->state != STATE_IDLE) {
+		dev_dbg(&host->pd->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);
+
+	if (ios->power_mode == MMC_POWER_UP) {
+		if (!host->card_present) {
+			/* See if we also get DMA */
+			sh_mmcif_request_dma(host, host->pd->dev.platform_data);
+			host->card_present = true;
+		}
+		sh_mmcif_set_power(host, ios);
+	} else if (ios->power_mode == MMC_POWER_OFF || !ios->clock) {
+		/* clock stop */
+		sh_mmcif_clock_control(host, 0);
+		if (ios->power_mode == MMC_POWER_OFF) {
+			if (host->card_present) {
+				sh_mmcif_release_dma(host);
+				host->card_present = false;
+			}
+		}
+		if (host->power) {
+			pm_runtime_put_sync(&host->pd->dev);
+			clk_disable_unprepare(host->hclk);
+			host->power = false;
+			if (ios->power_mode == MMC_POWER_OFF)
+				sh_mmcif_set_power(host, ios);
+		}
+		host->state = STATE_IDLE;
+		return;
+	}
+
+	if (ios->clock) {
+		if (!host->power) {
+			sh_mmcif_clk_update(host);
+			pm_runtime_get_sync(&host->pd->dev);
+			host->power = true;
+			sh_mmcif_sync_reset(host);
+		}
+		sh_mmcif_clock_control(host, ios->clock);
+	}
+
+	host->timing = ios->timing;
+	host->bus_width = ios->bus_width;
+	host->state = STATE_IDLE;
+}
+
+static int sh_mmcif_get_cd(struct mmc_host *mmc)
+{
+	struct sh_mmcif_host *host = mmc_priv(mmc);
+	struct sh_mmcif_plat_data *p = host->pd->dev.platform_data;
+	int ret = mmc_gpio_get_cd(mmc);
+
+	if (ret >= 0)
+		return ret;
+
+	if (!p || !p->get_cd)
+		return -ENOSYS;
+	else
+		return p->get_cd(host->pd);
+}
+
+static struct mmc_host_ops sh_mmcif_ops = {
+	.request	= sh_mmcif_request,
+	.set_ios	= sh_mmcif_set_ios,
+	.get_cd		= sh_mmcif_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;
+	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(&host->pd->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;
+	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(&host->pd->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(&host->pd->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(&host->pd->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;
+	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(&host->pd->dev, "IRQ state = 0x%08x incompletely cleared\n",
+			state);
+
+	if (state & INT_ERR_STS || state & ~INT_ALL) {
+		host->sd_error = true;
+		dev_dbg(&host->pd->dev, "int err state = 0x%08x\n", state);
+	}
+	if (state & ~(INT_CMD12RBE | INT_CMD12CRE)) {
+		if (!host->mrq)
+			dev_dbg(&host->pd->dev, "NULL IRQ state = 0x%08x\n", state);
+		if (!host->dma_active)
+			return IRQ_WAKE_THREAD;
+		else if (host->sd_error)
+			mmcif_dma_complete(host);
+	} else {
+		dev_dbg(&host->pd->dev, "Unexpected IRQ 0x%x\n", state);
+	}
+
+	return IRQ_HANDLED;
+}
+
+static void mmcif_timeout_work(struct work_struct *work)
+{
+	struct delayed_work *d = container_of(work, struct delayed_work, work);
+	struct sh_mmcif_host *host = container_of(d, struct sh_mmcif_host, timeout_work);
+	struct mmc_request *mrq = host->mrq;
+	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(&host->pd->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 sh_mmcif_plat_data *pd = host->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 sh_mmcif_plat_data *pd = pdev->dev.platform_data;
+	struct resource *res;
+	void __iomem *reg;
+	const char *name;
+
+	irq[0] = platform_get_irq(pdev, 0);
+	irq[1] = platform_get_irq(pdev, 1);
+	if (irq[0] < 0) {
+		dev_err(&pdev->dev, "Get irq error\n");
+		return -ENXIO;
+	}
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	reg = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(reg))
+		return PTR_ERR(reg);
+
+	mmc = mmc_alloc_host(sizeof(struct sh_mmcif_host), &pdev->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 = !pd || !pd->ccs_unsupported;
+	host->clk_ctrl2_enable = pd && pd->clk_ctrl2_present;
+
+	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;
+	if (pd && pd->caps)
+		mmc->caps |= pd->caps;
+	mmc->max_segs = 32;
+	mmc->max_blk_size = 512;
+	mmc->max_req_size = PAGE_CACHE_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);
+
+	pm_runtime_enable(&pdev->dev);
+	host->power = false;
+
+	host->hclk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(host->hclk)) {
+		ret = PTR_ERR(host->hclk);
+		dev_err(&pdev->dev, "cannot get clock: %d\n", ret);
+		goto err_pm;
+	}
+	ret = sh_mmcif_clk_update(host);
+	if (ret < 0)
+		goto err_pm;
+
+	ret = pm_runtime_resume(&pdev->dev);
+	if (ret < 0)
+		goto err_clk;
+
+	INIT_DELAYED_WORK(&host->timeout_work, 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(&pdev->dev) : "sh_mmc:error";
+	ret = devm_request_threaded_irq(&pdev->dev, irq[0], sh_mmcif_intr,
+					sh_mmcif_irqt, 0, name, host);
+	if (ret) {
+		dev_err(&pdev->dev, "request_irq error (%s)\n", name);
+		goto err_clk;
+	}
+	if (irq[1] >= 0) {
+		ret = devm_request_threaded_irq(&pdev->dev, irq[1],
+						sh_mmcif_intr, sh_mmcif_irqt,
+						0, "sh_mmc:int", host);
+		if (ret) {
+			dev_err(&pdev->dev, "request_irq error (sh_mmc:int)\n");
+			goto err_clk;
+		}
+	}
+
+	if (pd && pd->use_cd_gpio) {
+		ret = mmc_gpio_request_cd(mmc, pd->cd_gpio, 0);
+		if (ret < 0)
+			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(&pdev->dev, 100);
+
+	dev_info(&pdev->dev, "Chip version 0x%04x, clock rate %luMHz\n",
+		 sh_mmcif_readl(host->addr, MMCIF_CE_VERSION) & 0xffff,
+		 clk_get_rate(host->hclk) / 1000000UL);
+
+	clk_disable_unprepare(host->hclk);
+	return ret;
+
+err_clk:
+	clk_disable_unprepare(host->hclk);
+err_pm:
+	pm_runtime_disable(&pdev->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->hclk);
+	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->hclk);
+	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);
+
+	sh_mmcif_writel(host->addr, MMCIF_CE_INT_MASK, MASK_ALL);
+
+	return 0;
+}
+
+static int sh_mmcif_resume(struct device *dev)
+{
+	return 0;
+}
+#endif
+
+static const struct of_device_id mmcif_of_match[] = {
+	{ .compatible = "renesas,sh-mmcif" },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, mmcif_of_match);
+
+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,
+		.pm	= &sh_mmcif_dev_pm_ops,
+		.of_match_table = mmcif_of_match,
+	},
+};
+
+module_platform_driver(sh_mmcif_driver);
+
+MODULE_DESCRIPTION("SuperH on-chip MMC/eMMC interface driver");
+MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:" DRIVER_NAME);
+MODULE_AUTHOR("Yusuke Goda <yusuke.goda.sx@renesas.com>");
diff --git a/drivers/mmc/host/sh_mobile_sdhi.c b/drivers/mmc/host/sh_mobile_sdhi.c
new file mode 100644
index 000000000..354f4f335
--- /dev/null
+++ b/drivers/mmc/host/sh_mobile_sdhi.c
@@ -0,0 +1,395 @@
+/*
+ * SuperH Mobile SDHI
+ *
+ * Copyright (C) 2009 Magnus Damm
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * 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/mod_devicetable.h>
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <linux/mmc/host.h>
+#include <linux/mmc/sh_mobile_sdhi.h>
+#include <linux/mfd/tmio.h>
+#include <linux/sh_dma.h>
+#include <linux/delay.h>
+
+#include "tmio_mmc.h"
+
+#define EXT_ACC           0xe4
+
+#define host_to_priv(host) container_of((host)->pdata, struct sh_mobile_sdhi, mmc_data)
+
+struct sh_mobile_sdhi_of_data {
+	unsigned long tmio_flags;
+	unsigned long capabilities;
+	unsigned long capabilities2;
+	enum dma_slave_buswidth dma_buswidth;
+	dma_addr_t dma_rx_offset;
+};
+
+static const struct sh_mobile_sdhi_of_data sh_mobile_sdhi_of_cfg[] = {
+	{
+		.tmio_flags = TMIO_MMC_HAS_IDLE_WAIT,
+	},
+};
+
+static const struct sh_mobile_sdhi_of_data of_rcar_gen1_compatible = {
+	.tmio_flags	= TMIO_MMC_HAS_IDLE_WAIT | TMIO_MMC_WRPROTECT_DISABLE |
+			  TMIO_MMC_CLK_ACTUAL,
+	.capabilities	= MMC_CAP_SD_HIGHSPEED | MMC_CAP_SDIO_IRQ,
+};
+
+static const struct sh_mobile_sdhi_of_data of_rcar_gen2_compatible = {
+	.tmio_flags	= TMIO_MMC_HAS_IDLE_WAIT | TMIO_MMC_WRPROTECT_DISABLE |
+			  TMIO_MMC_CLK_ACTUAL,
+	.capabilities	= MMC_CAP_SD_HIGHSPEED | MMC_CAP_SDIO_IRQ,
+	.dma_buswidth	= DMA_SLAVE_BUSWIDTH_4_BYTES,
+	.dma_rx_offset	= 0x2000,
+};
+
+static const struct of_device_id sh_mobile_sdhi_of_match[] = {
+	{ .compatible = "renesas,sdhi-shmobile" },
+	{ .compatible = "renesas,sdhi-sh7372" },
+	{ .compatible = "renesas,sdhi-sh73a0", .data = &sh_mobile_sdhi_of_cfg[0], },
+	{ .compatible = "renesas,sdhi-r8a73a4", .data = &sh_mobile_sdhi_of_cfg[0], },
+	{ .compatible = "renesas,sdhi-r8a7740", .data = &sh_mobile_sdhi_of_cfg[0], },
+	{ .compatible = "renesas,sdhi-r8a7778", .data = &of_rcar_gen1_compatible, },
+	{ .compatible = "renesas,sdhi-r8a7779", .data = &of_rcar_gen1_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, },
+	{},
+};
+MODULE_DEVICE_TABLE(of, sh_mobile_sdhi_of_match);
+
+struct sh_mobile_sdhi {
+	struct clk *clk;
+	struct tmio_mmc_data mmc_data;
+	struct tmio_mmc_dma dma_priv;
+};
+
+static void sh_mobile_sdhi_sdbuf_width(struct tmio_mmc_host *host, int width)
+{
+	u32 val;
+
+	/*
+	 * see also
+	 *	sh_mobile_sdhi_of_data :: dma_buswidth
+	 */
+	switch (sd_ctrl_read16(host, CTL_VERSION)) {
+	case 0x490C:
+		val = (width == 32) ? 0x0001 : 0x0000;
+		break;
+	case 0xCB0D:
+		val = (width == 32) ? 0x0000 : 0x0001;
+		break;
+	default:
+		/* nothing to do */
+		return;
+	}
+
+	sd_ctrl_write16(host, EXT_ACC, val);
+}
+
+static int sh_mobile_sdhi_clk_enable(struct platform_device *pdev, unsigned int *f)
+{
+	struct mmc_host *mmc = platform_get_drvdata(pdev);
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+	struct sh_mobile_sdhi *priv = host_to_priv(host);
+	int ret = clk_prepare_enable(priv->clk);
+	if (ret < 0)
+		return ret;
+
+	*f = clk_get_rate(priv->clk);
+
+	/* enable 16bit data access on SDBUF as default */
+	sh_mobile_sdhi_sdbuf_width(host, 16);
+
+	return 0;
+}
+
+static void sh_mobile_sdhi_clk_disable(struct platform_device *pdev)
+{
+	struct mmc_host *mmc = platform_get_drvdata(pdev);
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+	struct sh_mobile_sdhi *priv = host_to_priv(host);
+	clk_disable_unprepare(priv->clk);
+}
+
+static int sh_mobile_sdhi_wait_idle(struct tmio_mmc_host *host)
+{
+	int timeout = 1000;
+
+	while (--timeout && !(sd_ctrl_read16(host, CTL_STATUS2) & (1 << 13)))
+		udelay(1);
+
+	if (!timeout) {
+		dev_warn(&host->pdev->dev, "timeout waiting for SD bus idle\n");
+		return -EBUSY;
+	}
+
+	return 0;
+}
+
+static int sh_mobile_sdhi_write16_hook(struct tmio_mmc_host *host, int addr)
+{
+	switch (addr)
+	{
+	case CTL_SD_CMD:
+	case CTL_STOP_INTERNAL_ACTION:
+	case CTL_XFER_BLK_COUNT:
+	case CTL_SD_CARD_CLK_CTL:
+	case CTL_SD_XFER_LEN:
+	case CTL_SD_MEM_CARD_OPT:
+	case CTL_TRANSACTION_CTL:
+	case CTL_DMA_ENABLE:
+		return sh_mobile_sdhi_wait_idle(host);
+	}
+
+	return 0;
+}
+
+static int sh_mobile_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 sh_mobile_sdhi_enable_dma(struct tmio_mmc_host *host, bool enable)
+{
+	sd_ctrl_write16(host, CTL_DMA_ENABLE, enable ? 2 : 0);
+
+	/* enable 32bit access if DMA mode if possibile */
+	sh_mobile_sdhi_sdbuf_width(host, enable ? 32 : 16);
+}
+
+static int sh_mobile_sdhi_probe(struct platform_device *pdev)
+{
+	const struct of_device_id *of_id =
+		of_match_device(sh_mobile_sdhi_of_match, &pdev->dev);
+	struct sh_mobile_sdhi *priv;
+	struct tmio_mmc_data *mmc_data;
+	struct tmio_mmc_data *mmd = pdev->dev.platform_data;
+	struct tmio_mmc_host *host;
+	struct resource *res;
+	int irq, ret, i = 0;
+	bool multiplexed_isr = true;
+	struct tmio_mmc_dma *dma_priv;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res)
+		return -EINVAL;
+
+	priv = devm_kzalloc(&pdev->dev, sizeof(struct sh_mobile_sdhi), GFP_KERNEL);
+	if (priv == NULL) {
+		dev_err(&pdev->dev, "kzalloc failed\n");
+		return -ENOMEM;
+	}
+
+	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);
+		goto eprobe;
+	}
+
+	host = tmio_mmc_host_alloc(pdev);
+	if (!host) {
+		ret = -ENOMEM;
+		goto eprobe;
+	}
+
+	host->dma		= dma_priv;
+	host->write16_hook	= sh_mobile_sdhi_write16_hook;
+	host->clk_enable	= sh_mobile_sdhi_clk_enable;
+	host->clk_disable	= sh_mobile_sdhi_clk_disable;
+	host->multi_io_quirk	= sh_mobile_sdhi_multi_io_quirk;
+	/* SD control register space size is 0x100, 0x200 for bus_shift=1 */
+	if (resource_size(res) > 0x100)
+		host->bus_shift = 1;
+	else
+		host->bus_shift = 0;
+
+	if (mmd)
+		*mmc_data = *mmd;
+
+	dma_priv->filter = shdma_chan_filter;
+	dma_priv->enable = sh_mobile_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 controll bit
+	 */
+	mmc_data->flags |= TMIO_MMC_HAVE_CMD12_CTRL;
+
+	/*
+	 * All SDHI need SDIO_INFO1 reserved bit
+	 */
+	mmc_data->flags |= TMIO_MMC_SDIO_STATUS_QUIRK;
+
+	if (of_id && of_id->data) {
+		const struct sh_mobile_sdhi_of_data *of_data = of_id->data;
+		mmc_data->flags |= of_data->tmio_flags;
+		mmc_data->capabilities |= of_data->capabilities;
+		mmc_data->capabilities2 |= of_data->capabilities2;
+		mmc_data->dma_rx_offset = of_data->dma_rx_offset;
+		dma_priv->dma_buswidth = of_data->dma_buswidth;
+	}
+
+	ret = tmio_mmc_host_probe(host, mmc_data);
+	if (ret < 0)
+		goto efree;
+
+	/*
+	 * Allow one or more specific (named) ISRs or
+	 * one or more multiplexed (un-named) ISRs.
+	 */
+
+	irq = platform_get_irq_byname(pdev, SH_MOBILE_SDHI_IRQ_CARD_DETECT);
+	if (irq >= 0) {
+		multiplexed_isr = false;
+		ret = devm_request_irq(&pdev->dev, irq, tmio_mmc_card_detect_irq, 0,
+				  dev_name(&pdev->dev), host);
+		if (ret)
+			goto eirq;
+	}
+
+	irq = platform_get_irq_byname(pdev, SH_MOBILE_SDHI_IRQ_SDIO);
+	if (irq >= 0) {
+		multiplexed_isr = false;
+		ret = devm_request_irq(&pdev->dev, irq, tmio_mmc_sdio_irq, 0,
+				  dev_name(&pdev->dev), host);
+		if (ret)
+			goto eirq;
+	}
+
+	irq = platform_get_irq_byname(pdev, SH_MOBILE_SDHI_IRQ_SDCARD);
+	if (irq >= 0) {
+		multiplexed_isr = false;
+		ret = devm_request_irq(&pdev->dev, irq, tmio_mmc_sdcard_irq, 0,
+				  dev_name(&pdev->dev), host);
+		if (ret)
+			goto eirq;
+	} else if (!multiplexed_isr) {
+		dev_err(&pdev->dev,
+			"Principal SD-card IRQ is missing among named interrupts\n");
+		ret = irq;
+		goto eirq;
+	}
+
+	if (multiplexed_isr) {
+		while (1) {
+			irq = platform_get_irq(pdev, i);
+			if (irq < 0)
+				break;
+			i++;
+			ret = devm_request_irq(&pdev->dev, irq, tmio_mmc_irq, 0,
+					  dev_name(&pdev->dev), host);
+			if (ret)
+				goto eirq;
+		}
+
+		/* There must be at least one IRQ source */
+		if (!i) {
+			ret = irq;
+			goto eirq;
+		}
+	}
+
+	dev_info(&pdev->dev, "%s base at 0x%08lx clock rate %u MHz\n",
+		 mmc_hostname(host->mmc), (unsigned long)
+		 (platform_get_resource(pdev, IORESOURCE_MEM, 0)->start),
+		 host->mmc->f_max / 1000000);
+
+	return ret;
+
+eirq:
+	tmio_mmc_host_remove(host);
+efree:
+	tmio_mmc_host_free(host);
+eprobe:
+	return ret;
+}
+
+static int sh_mobile_sdhi_remove(struct platform_device *pdev)
+{
+	struct mmc_host *mmc = platform_get_drvdata(pdev);
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+
+	tmio_mmc_host_remove(host);
+
+	return 0;
+}
+
+static const struct dev_pm_ops tmio_mmc_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 sh_mobile_sdhi_driver = {
+	.driver		= {
+		.name	= "sh_mobile_sdhi",
+		.pm	= &tmio_mmc_dev_pm_ops,
+		.of_match_table = sh_mobile_sdhi_of_match,
+	},
+	.probe		= sh_mobile_sdhi_probe,
+	.remove		= sh_mobile_sdhi_remove,
+};
+
+module_platform_driver(sh_mobile_sdhi_driver);
+
+MODULE_DESCRIPTION("SuperH Mobile SDHI driver");
+MODULE_AUTHOR("Magnus Damm");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:sh_mobile_sdhi");
diff --git a/drivers/mmc/host/sunxi-mmc.c b/drivers/mmc/host/sunxi-mmc.c
new file mode 100644
index 000000000..4d3e1ffe5
--- /dev/null
+++ b/drivers/mmc/host/sunxi-mmc.c
@@ -0,0 +1,1089 @@
+/*
+ * 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 Lanzend�rfer <david.lanzendoerfer@o2s.ch>
+ * (C) Copyright 2013-2014 Hans de Goede <hdegoede@redhat.com>
+ *
+ * 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; either version 2 of
+ * the License, or (at your option) any later version.
+ */
+
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/io.h>
+#include <linux/device.h>
+#include <linux/interrupt.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+
+#include <linux/clk.h>
+#include <linux/gpio.h>
+#include <linux/platform_device.h>
+#include <linux/spinlock.h>
+#include <linux/scatterlist.h>
+#include <linux/dma-mapping.h>
+#include <linux/slab.h>
+#include <linux/reset.h>
+
+#include <linux/of_address.h>
+#include <linux/of_gpio.h>
+#include <linux/of_platform.h>
+
+#include <linux/mmc/host.h>
+#include <linux/mmc/sd.h>
+#include <linux/mmc/sdio.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/core.h>
+#include <linux/mmc/card.h>
+#include <linux/mmc/slot-gpio.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)
+
+#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_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 */
+
+struct sunxi_idma_des {
+	u32	config;
+	u32	buf_size;
+	u32	buf_addr_ptr1;
+	u32	buf_addr_ptr2;
+};
+
+struct sunxi_mmc_host {
+	struct mmc_host	*mmc;
+	struct reset_control *reset;
+
+	/* 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 */
+	u32		idma_des_size_bits;
+	dma_addr_t	sg_dma;
+	void		*sg_cpu;
+	bool		wait_dma;
+
+	struct mmc_request *mrq;
+	struct mmc_request *manual_stop_mrq;
+	int		ferror;
+};
+
+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 mmc_host *mmc)
+{
+	u32 rval;
+	struct sunxi_mmc_host *host = mmc_priv(mmc);
+
+	if (sunxi_mmc_reset_host(host))
+		return -EIO;
+
+	mmc_writel(host, REG_FTRGL, 0x20070008);
+	mmc_writel(host, REG_TMOUT, 0xffffffff);
+	mmc_writel(host, REG_IMASK, host->sdio_imask);
+	mmc_writel(host, REG_RINTR, 0xffffffff);
+	mmc_writel(host, REG_DBGC, 0xdeb);
+	mmc_writel(host, REG_FUNS, SDXC_CEATA_ON);
+	mmc_writel(host, REG_DLBA, host->sg_dma);
+
+	rval = mmc_readl(host, REG_GCTRL);
+	rval |= SDXC_INTERRUPT_ENABLE_BIT;
+	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->idma_des_size_bits);
+
+	for (i = 0; i < data->sg_len; i++) {
+		pdes[i].config = 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 = data->sg[i].length;
+
+		next_desc += sizeof(struct sunxi_idma_des);
+		pdes[i].buf_addr_ptr1 = sg_dma_address(&data->sg[i]);
+		pdes[i].buf_addr_ptr2 = (u32)next_desc;
+	}
+
+	pdes[0].config |= SDXC_IDMAC_DES0_FD;
+	pdes[i - 1].config |= SDXC_IDMAC_DES0_LD | SDXC_IDMAC_DES0_ER;
+	pdes[i - 1].config &= ~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 enum dma_data_direction sunxi_mmc_get_dma_dir(struct mmc_data *data)
+{
+	if (data->flags & MMC_DATA_WRITE)
+		return DMA_TO_DEVICE;
+	else
+		return DMA_FROM_DEVICE;
+}
+
+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,
+			     sunxi_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_err(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,
+				     sunxi_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(250);
+	u32 rval;
+
+	rval = mmc_readl(host, REG_CLKCR);
+	rval &= ~(SDXC_CARD_CLOCK_ON | SDXC_LOW_POWER_ON);
+
+	if (oclk_en)
+		rval |= SDXC_CARD_CLOCK_ON;
+
+	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;
+	}
+
+	return 0;
+}
+
+static int sunxi_mmc_clk_set_rate(struct sunxi_mmc_host *host,
+				  struct mmc_ios *ios)
+{
+	u32 rate, oclk_dly, rval, sclk_dly;
+	int ret;
+
+	rate = clk_round_rate(host->clk_mmc, ios->clock);
+	dev_dbg(mmc_dev(host->mmc), "setting clk to %d, rounded %d\n",
+		ios->clock, rate);
+
+	/* setting clock rate */
+	ret = clk_set_rate(host->clk_mmc, rate);
+	if (ret) {
+		dev_err(mmc_dev(host->mmc), "error setting clk to %d: %d\n",
+			rate, ret);
+		return ret;
+	}
+
+	ret = sunxi_mmc_oclk_onoff(host, 0);
+	if (ret)
+		return ret;
+
+	/* clear internal divider */
+	rval = mmc_readl(host, REG_CLKCR);
+	rval &= ~0xff;
+	mmc_writel(host, REG_CLKCR, rval);
+
+	/* determine delays */
+	if (rate <= 400000) {
+		oclk_dly = 180;
+		sclk_dly = 42;
+	} else if (rate <= 25000000) {
+		oclk_dly = 180;
+		sclk_dly = 75;
+	} else if (rate <= 50000000) {
+		if (ios->timing == MMC_TIMING_UHS_DDR50) {
+			oclk_dly = 60;
+			sclk_dly = 120;
+		} else {
+			oclk_dly = 90;
+			sclk_dly = 150;
+		}
+	} else if (rate <= 100000000) {
+		oclk_dly = 6;
+		sclk_dly = 24;
+	} else if (rate <= 200000000) {
+		oclk_dly = 3;
+		sclk_dly = 12;
+	} else {
+		return -EINVAL;
+	}
+
+	clk_set_phase(host->clk_sample, sclk_dly);
+	clk_set_phase(host->clk_output, oclk_dly);
+
+	return sunxi_mmc_oclk_onoff(host, 1);
+}
+
+static void sunxi_mmc_set_ios(struct mmc_host *mmc, struct mmc_ios *ios)
+{
+	struct sunxi_mmc_host *host = mmc_priv(mmc);
+	u32 rval;
+
+	/* Set the power state */
+	switch (ios->power_mode) {
+	case MMC_POWER_ON:
+		break;
+
+	case MMC_POWER_UP:
+		mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, ios->vdd);
+
+		host->ferror = sunxi_mmc_init_host(mmc);
+		if (host->ferror)
+			return;
+
+		dev_dbg(mmc_dev(mmc), "power on!\n");
+		break;
+
+	case MMC_POWER_OFF:
+		dev_dbg(mmc_dev(mmc), "power off!\n");
+		sunxi_mmc_reset_host(host);
+		mmc_regulator_set_ocr(mmc, mmc->supply.vmmc, 0);
+		break;
+	}
+
+	/* set bus width */
+	switch (ios->bus_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;
+	}
+
+	/* set ddr mode */
+	rval = mmc_readl(host, REG_GCTRL);
+	if (ios->timing == MMC_TIMING_UHS_DDR50)
+		rval |= SDXC_DDR_MODE;
+	else
+		rval &= ~SDXC_DDR_MODE;
+	mmc_writel(host, REG_GCTRL, rval);
+
+	/* set up clock */
+	if (ios->clock && ios->power_mode) {
+		host->ferror = sunxi_mmc_clk_set_rate(host, ios);
+		/* Android code had a usleep_range(50000, 55000); here */
+	}
+}
+
+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;
+
+	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);
+}
+
+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->flags & MMC_DATA_STREAM) {
+				imask |= SDXC_AUTO_COMMAND_DONE;
+				cmd_val |= SDXC_SEQUENCE_MODE |
+					   SDXC_SEND_AUTO_STOP;
+			}
+
+			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,
+				     sunxi_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 const struct of_device_id sunxi_mmc_of_match[] = {
+	{ .compatible = "allwinner,sun4i-a10-mmc", },
+	{ .compatible = "allwinner,sun5i-a13-mmc", },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, sunxi_mmc_of_match);
+
+static 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,
+	.hw_reset	 = sunxi_mmc_hw_reset,
+};
+
+static int sunxi_mmc_resource_request(struct sunxi_mmc_host *host,
+				      struct platform_device *pdev)
+{
+	struct device_node *np = pdev->dev.of_node;
+	int ret;
+
+	if (of_device_is_compatible(np, "allwinner,sun4i-a10-mmc"))
+		host->idma_des_size_bits = 13;
+	else
+		host->idma_des_size_bits = 16;
+
+	ret = mmc_regulator_get_supply(host->mmc);
+	if (ret) {
+		if (ret != -EPROBE_DEFER)
+			dev_err(&pdev->dev, "Could not get vmmc supply\n");
+		return ret;
+	}
+
+	host->reg_base = devm_ioremap_resource(&pdev->dev,
+			      platform_get_resource(pdev, IORESOURCE_MEM, 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);
+	}
+
+	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(&pdev->dev, "ahb");
+	if (PTR_ERR(host->reset) == -EPROBE_DEFER)
+		return PTR_ERR(host->reset);
+
+	ret = clk_prepare_enable(host->clk_ahb);
+	if (ret) {
+		dev_err(&pdev->dev, "Enable ahb clk err %d\n", ret);
+		return ret;
+	}
+
+	ret = clk_prepare_enable(host->clk_mmc);
+	if (ret) {
+		dev_err(&pdev->dev, "Enable mmc clk err %d\n", ret);
+		goto error_disable_clk_ahb;
+	}
+
+	ret = clk_prepare_enable(host->clk_output);
+	if (ret) {
+		dev_err(&pdev->dev, "Enable output clk err %d\n", ret);
+		goto error_disable_clk_mmc;
+	}
+
+	ret = clk_prepare_enable(host->clk_sample);
+	if (ret) {
+		dev_err(&pdev->dev, "Enable sample clk err %d\n", ret);
+		goto error_disable_clk_output;
+	}
+
+	if (!IS_ERR(host->reset)) {
+		ret = reset_control_deassert(host->reset);
+		if (ret) {
+			dev_err(&pdev->dev, "reset err %d\n", ret);
+			goto error_disable_clk_sample;
+		}
+	}
+
+	/*
+	 * 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_assert_reset;
+
+	host->irq = platform_get_irq(pdev, 0);
+	return devm_request_threaded_irq(&pdev->dev, host->irq, sunxi_mmc_irq,
+			sunxi_mmc_handle_manual_stop, 0, "sunxi-mmc", host);
+
+error_assert_reset:
+	if (!IS_ERR(host->reset))
+		reset_control_assert(host->reset);
+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);
+	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;
+	}
+
+	host = mmc_priv(mmc);
+	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;
+	}
+
+	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->idma_des_size_bits);
+	mmc->max_req_size	= mmc->max_seg_size * mmc->max_segs;
+	/* 400kHz ~ 50MHz */
+	mmc->f_min		=   400000;
+	mmc->f_max		= 50000000;
+	mmc->caps	       |= MMC_CAP_MMC_HIGHSPEED | MMC_CAP_SD_HIGHSPEED |
+				  MMC_CAP_ERASE | MMC_CAP_SDIO_IRQ;
+
+	ret = mmc_of_parse(mmc);
+	if (ret)
+		goto error_free_dma;
+
+	ret = mmc_add_host(mmc);
+	if (ret)
+		goto error_free_dma;
+
+	dev_info(&pdev->dev, "base:0x%p irq:%u\n", host->reg_base, host->irq);
+	platform_set_drvdata(pdev, mmc);
+	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);
+	disable_irq(host->irq);
+	sunxi_mmc_reset_host(host);
+
+	if (!IS_ERR(host->reset))
+		reset_control_assert(host->reset);
+
+	clk_disable_unprepare(host->clk_mmc);
+	clk_disable_unprepare(host->clk_ahb);
+
+	dma_free_coherent(&pdev->dev, PAGE_SIZE, host->sg_cpu, host->sg_dma);
+	mmc_free_host(mmc);
+
+	return 0;
+}
+
+static struct platform_driver sunxi_mmc_driver = {
+	.driver = {
+		.name	= "sunxi-mmc",
+		.of_match_table = of_match_ptr(sunxi_mmc_of_match),
+	},
+	.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 Lanzend�rfer <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..93c4b40df
--- /dev/null
+++ b/drivers/mmc/host/tifm_sd.c
@@ -0,0 +1,1091 @@
+/*
+ *  tifm_sd.c - TI FlashMedia driver
+ *
+ *  Copyright (C) 2006 Alex Dubov <oakad@yahoo.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * 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
+
+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; // deliberate fall-through
+	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(unsigned long data)
+{
+	struct tifm_sd *host = (struct tifm_sd*)data;
+
+	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);
+	mmiowb();
+	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);
+	mmiowb();
+
+	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(1000);
+
+	tasklet_init(&host->finish_tasklet, tifm_sd_end_cmd,
+		     (unsigned long)host);
+	setup_timer(&host->timer, tifm_sd_abort, (unsigned long)host);
+
+	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);
+	mmiowb();
+	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..f746df493
--- /dev/null
+++ b/drivers/mmc/host/tmio_mmc.c
@@ -0,0 +1,164 @@
+/*
+ * linux/drivers/mmc/host/tmio_mmc.c
+ *
+ * Copyright (C) 2007 Ian Molton
+ * Copyright (C) 2004 Ian Molton
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * Driver for the MMC / SD / SDIO cell found in:
+ *
+ * TC6393XB TC6391XB TC6387XB T7L66XB ASIC3
+ */
+
+#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"
+
+#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)
+		return -EINVAL;
+
+	pdata->flags |= TMIO_MMC_HAVE_HIGH_REG;
+
+	host = tmio_mmc_host_alloc(pdev);
+	if (!host)
+		goto cell_disable;
+
+	/* SD control register space size is 0x200, 0x400 for bus_shift=1 */
+	host->bus_shift = resource_size(res) >> 10;
+
+	ret = tmio_mmc_host_probe(host, pdata);
+	if (ret)
+		goto host_free;
+
+	ret = request_irq(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 mmc_host *mmc = platform_get_drvdata(pdev);
+
+	if (mmc) {
+		struct tmio_mmc_host *host = mmc_priv(mmc);
+		free_irq(platform_get_irq(pdev, 0), host);
+		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",
+		.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..4a597f5a5
--- /dev/null
+++ b/drivers/mmc/host/tmio_mmc.h
@@ -0,0 +1,209 @@
+/*
+ * linux/drivers/mmc/host/tmio_mmc.h
+ *
+ * Copyright (C) 2007 Ian Molton
+ * Copyright (C) 2004 Ian Molton
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * Driver for the MMC / SD / SDIO cell found in:
+ *
+ * TC6393XB TC6391XB TC6387XB T7L66XB ASIC3
+ */
+
+#ifndef TMIO_MMC_H
+#define TMIO_MMC_H
+
+#include <linux/dmaengine.h>
+#include <linux/highmem.h>
+#include <linux/mmc/tmio.h>
+#include <linux/mutex.h>
+#include <linux/pagemap.h>
+#include <linux/scatterlist.h>
+#include <linux/spinlock.h>
+
+/* Definitions for values the CTRL_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 some IRQ masks */
+/* This is the mask used at reset by the chip */
+#define TMIO_MASK_ALL           0x837f031d
+#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)
+
+struct tmio_mmc_data;
+struct tmio_mmc_host;
+
+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 tmio_mmc_host {
+	void __iomem *ctl;
+	struct mmc_command      *cmd;
+	struct mmc_request      *mrq;
+	struct mmc_data         *data;
+	struct mmc_host         *mmc;
+
+	/* Callbacks for clock / power control */
+	void (*set_pwr)(struct platform_device *host, int state);
+	void (*set_clk_div)(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 long		bus_shift;
+
+	struct platform_device *pdev;
+	struct tmio_mmc_data *pdata;
+	struct tmio_mmc_dma	*dma;
+
+	/* DMA support */
+	bool			force_pio;
+	struct dma_chan		*chan_rx;
+	struct dma_chan		*chan_tx;
+	struct tasklet_struct	dma_complete;
+	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;
+
+	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;
+
+	int (*write16_hook)(struct tmio_mmc_host *host, int addr);
+	int (*clk_enable)(struct platform_device *pdev, unsigned int *f);
+	void (*clk_disable)(struct platform_device *pdev);
+	int (*multi_io_quirk)(struct mmc_card *card,
+			      unsigned int direction, int blk_size);
+};
+
+struct tmio_mmc_host *tmio_mmc_host_alloc(struct platform_device *pdev);
+void tmio_mmc_host_free(struct tmio_mmc_host *host);
+int tmio_mmc_host_probe(struct tmio_mmc_host *host,
+			struct tmio_mmc_data *pdata);
+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);
+irqreturn_t tmio_mmc_sdcard_irq(int irq, void *devid);
+irqreturn_t tmio_mmc_card_detect_irq(int irq, void *devid);
+irqreturn_t tmio_mmc_sdio_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);
+}
+
+#if defined(CONFIG_MMC_SDHI) || defined(CONFIG_MMC_SDHI_MODULE)
+void tmio_mmc_start_dma(struct tmio_mmc_host *host, struct mmc_data *data);
+void tmio_mmc_enable_dma(struct tmio_mmc_host *host, bool enable);
+void tmio_mmc_request_dma(struct tmio_mmc_host *host, struct tmio_mmc_data *pdata);
+void tmio_mmc_release_dma(struct tmio_mmc_host *host);
+void tmio_mmc_abort_dma(struct tmio_mmc_host *host);
+#else
+static inline void tmio_mmc_start_dma(struct tmio_mmc_host *host,
+			       struct mmc_data *data)
+{
+}
+
+static inline void tmio_mmc_enable_dma(struct tmio_mmc_host *host, bool enable)
+{
+}
+
+static inline void tmio_mmc_request_dma(struct tmio_mmc_host *host,
+				 struct tmio_mmc_data *pdata)
+{
+	host->chan_tx = NULL;
+	host->chan_rx = NULL;
+}
+
+static inline void tmio_mmc_release_dma(struct tmio_mmc_host *host)
+{
+}
+
+static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
+{
+}
+#endif
+
+#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 readw(host->ctl + (addr << host->bus_shift));
+}
+
+static inline void sd_ctrl_read16_rep(struct tmio_mmc_host *host, int addr,
+		u16 *buf, int count)
+{
+	readsw(host->ctl + (addr << host->bus_shift), buf, count);
+}
+
+static inline u32 sd_ctrl_read32(struct tmio_mmc_host *host, int addr)
+{
+	return readw(host->ctl + (addr << host->bus_shift)) |
+	       readw(host->ctl + ((addr + 2) << host->bus_shift)) << 16;
+}
+
+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;
+	writew(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)
+{
+	writesw(host->ctl + (addr << host->bus_shift), buf, count);
+}
+
+static inline void sd_ctrl_write32(struct tmio_mmc_host *host, int addr, u32 val)
+{
+	writew(val, host->ctl + (addr << host->bus_shift));
+	writew(val >> 16, host->ctl + ((addr + 2) << host->bus_shift));
+}
+
+
+#endif
diff --git a/drivers/mmc/host/tmio_mmc_dma.c b/drivers/mmc/host/tmio_mmc_dma.c
new file mode 100644
index 000000000..e4b05dbb9
--- /dev/null
+++ b/drivers/mmc/host/tmio_mmc_dma.c
@@ -0,0 +1,356 @@
+/*
+ * linux/drivers/mmc/tmio_mmc_dma.c
+ *
+ * Copyright (C) 2010-2011 Guennadi Liakhovetski
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * DMA function for TMIO MMC implementations
+ */
+
+#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/mmc/tmio.h>
+#include <linux/pagemap.h>
+#include <linux/scatterlist.h>
+
+#include "tmio_mmc.h"
+
+#define TMIO_MMC_MIN_DMA_LEN 8
+
+void tmio_mmc_enable_dma(struct tmio_mmc_host *host, bool enable)
+{
+	if (!host->chan_tx || !host->chan_rx)
+		return;
+
+	if (host->dma->enable)
+		host->dma->enable(host, enable);
+}
+
+void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
+{
+	tmio_mmc_enable_dma(host, false);
+
+	if (host->chan_rx)
+		dmaengine_terminate_all(host->chan_rx);
+	if (host->chan_tx)
+		dmaengine_terminate_all(host->chan_tx);
+
+	tmio_mmc_enable_dma(host, true);
+}
+
+static void tmio_mmc_start_dma_rx(struct tmio_mmc_host *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_CACHE_SIZE ||
+			  (align & PAGE_MASK))) || !multiple) {
+		ret = -EINVAL;
+		goto pio;
+	}
+
+	if (sg->length < TMIO_MMC_MIN_DMA_LEN) {
+		host->force_pio = true;
+		return;
+	}
+
+	tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_RXRDY);
+
+	/* 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) {
+		cookie = dmaengine_submit(desc);
+		if (cookie < 0) {
+			desc = NULL;
+			ret = cookie;
+		}
+	}
+	dev_dbg(&host->pdev->dev, "%s(): mapped %d -> %d, cookie %d, rq %p\n",
+		__func__, host->sg_len, ret, cookie, host->mrq);
+
+pio:
+	if (!desc) {
+		/* DMA failed, fall back to PIO */
+		tmio_mmc_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);
+	}
+
+	dev_dbg(&host->pdev->dev, "%s(): desc %p, cookie %d, sg[%d]\n", __func__,
+		desc, cookie, host->sg_len);
+}
+
+static void tmio_mmc_start_dma_tx(struct tmio_mmc_host *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_CACHE_SIZE ||
+			  (align & PAGE_MASK))) || !multiple) {
+		ret = -EINVAL;
+		goto pio;
+	}
+
+	if (sg->length < TMIO_MMC_MIN_DMA_LEN) {
+		host->force_pio = true;
+		return;
+	}
+
+	tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_TXRQ);
+
+	/* 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) {
+		cookie = dmaengine_submit(desc);
+		if (cookie < 0) {
+			desc = NULL;
+			ret = cookie;
+		}
+	}
+	dev_dbg(&host->pdev->dev, "%s(): mapped %d -> %d, cookie %d, rq %p\n",
+		__func__, host->sg_len, ret, cookie, host->mrq);
+
+pio:
+	if (!desc) {
+		/* DMA failed, fall back to PIO */
+		tmio_mmc_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);
+	}
+
+	dev_dbg(&host->pdev->dev, "%s(): desc %p, cookie %d\n", __func__,
+		desc, cookie);
+}
+
+void tmio_mmc_start_dma(struct tmio_mmc_host *host,
+			       struct mmc_data *data)
+{
+	if (data->flags & MMC_DATA_READ) {
+		if (host->chan_rx)
+			tmio_mmc_start_dma_rx(host);
+	} else {
+		if (host->chan_tx)
+			tmio_mmc_start_dma_tx(host);
+	}
+}
+
+static void tmio_mmc_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 && 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 tmio_mmc_tasklet_fn(unsigned long arg)
+{
+	struct tmio_mmc_host *host = (struct tmio_mmc_host *)arg;
+
+	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);
+
+	tmio_mmc_do_data_irq(host);
+out:
+	spin_unlock_irq(&host->lock);
+}
+
+void tmio_mmc_request_dma(struct tmio_mmc_host *host, struct tmio_mmc_data *pdata)
+{
+	/* We can only either use DMA for both Tx and Rx or not use it at all */
+	if (!host->dma || (!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,
+					host->dma->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 = host->dma->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,
+					host->dma->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 = host->dma->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;
+
+		tasklet_init(&host->dma_complete, tmio_mmc_tasklet_fn, (unsigned long)host);
+		tasklet_init(&host->dma_issue, tmio_mmc_issue_tasklet_fn, (unsigned long)host);
+	}
+
+	tmio_mmc_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;
+}
+
+void tmio_mmc_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;
+	}
+}
diff --git a/drivers/mmc/host/tmio_mmc_pio.c b/drivers/mmc/host/tmio_mmc_pio.c
new file mode 100644
index 000000000..dba7e1c19
--- /dev/null
+++ b/drivers/mmc/host/tmio_mmc_pio.c
@@ -0,0 +1,1276 @@
+/*
+ * linux/drivers/mmc/host/tmio_mmc_pio.c
+ *
+ * Copyright (C) 2011 Guennadi Liakhovetski
+ * Copyright (C) 2007 Ian Molton
+ * Copyright (C) 2004 Ian Molton
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * Driver for the MMC / SD / SDIO IP found in:
+ *
+ * TC6393XB, TC6391XB, TC6387XB, T7L66XB, ASIC3, SH-Mobile SoCs
+ *
+ * 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
+ *   SDIO support
+ *   Better Power management
+ *   Handle MMC errors better
+ *   double buffer support
+ *
+ */
+
+#include <linux/delay.h>
+#include <linux/device.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/host.h>
+#include <linux/mmc/mmc.h>
+#include <linux/mmc/slot-gpio.h>
+#include <linux/mmc/tmio.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/spinlock.h>
+#include <linux/workqueue.h>
+
+#include "tmio_mmc.h"
+
+void tmio_mmc_enable_mmc_irqs(struct tmio_mmc_host *host, u32 i)
+{
+	host->sdcard_irq_mask &= ~(i & TMIO_MASK_IRQ);
+	sd_ctrl_write32(host, CTL_IRQ_MASK, host->sdcard_irq_mask);
+}
+
+void tmio_mmc_disable_mmc_irqs(struct tmio_mmc_host *host, u32 i)
+{
+	host->sdcard_irq_mask |= (i & TMIO_MASK_IRQ);
+	sd_ctrl_write32(host, CTL_IRQ_MASK, host->sdcard_irq_mask);
+}
+
+static void tmio_mmc_ack_mmc_irqs(struct tmio_mmc_host *host, u32 i)
+{
+	sd_ctrl_write32(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;
+}
+
+#ifdef CONFIG_MMC_DEBUG
+
+#define STATUS_TO_TEXT(a, status, i) \
+	do { \
+		if (status & TMIO_STAT_##a) { \
+			if (i++) \
+				printk(" | "); \
+			printk(#a); \
+		} \
+	} while (0)
+
+static void pr_debug_status(u32 status)
+{
+	int i = 0;
+	pr_debug("status: %08x = ", status);
+	STATUS_TO_TEXT(CARD_REMOVE, status, i);
+	STATUS_TO_TEXT(CARD_INSERT, status, i);
+	STATUS_TO_TEXT(SIGSTATE, status, i);
+	STATUS_TO_TEXT(WRPROTECT, status, i);
+	STATUS_TO_TEXT(CARD_REMOVE_A, status, i);
+	STATUS_TO_TEXT(CARD_INSERT_A, status, i);
+	STATUS_TO_TEXT(SIGSTATE_A, status, i);
+	STATUS_TO_TEXT(CMD_IDX_ERR, status, i);
+	STATUS_TO_TEXT(STOPBIT_ERR, status, i);
+	STATUS_TO_TEXT(ILL_FUNC, status, i);
+	STATUS_TO_TEXT(CMD_BUSY, status, i);
+	STATUS_TO_TEXT(CMDRESPEND, status, i);
+	STATUS_TO_TEXT(DATAEND, status, i);
+	STATUS_TO_TEXT(CRCFAIL, status, i);
+	STATUS_TO_TEXT(DATATIMEOUT, status, i);
+	STATUS_TO_TEXT(CMDTIMEOUT, status, i);
+	STATUS_TO_TEXT(RXOVERFLOW, status, i);
+	STATUS_TO_TEXT(TXUNDERRUN, status, i);
+	STATUS_TO_TEXT(RXRDY, status, i);
+	STATUS_TO_TEXT(TXRQ, status, i);
+	STATUS_TO_TEXT(ILL_ACCESS, status, i);
+	printk("\n");
+}
+
+#else
+#define pr_debug_status(s)  do { } while (0)
+#endif
+
+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) {
+		/* 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;
+		sd_ctrl_write16(host, CTL_TRANSACTION_CTL, 0x0001);
+		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);
+		sd_ctrl_write16(host, CTL_TRANSACTION_CTL, 0x0000);
+
+		host->sdio_irq_enabled = false;
+		pm_runtime_mark_last_busy(mmc_dev(mmc));
+		pm_runtime_put_autosuspend(mmc_dev(mmc));
+	}
+}
+
+static void tmio_mmc_set_clock(struct tmio_mmc_host *host,
+				unsigned int new_clock)
+{
+	u32 clk = 0, clock;
+
+	if (new_clock) {
+		for (clock = host->mmc->f_min, 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))
+			clk |= 0xff;
+	}
+
+	if (host->set_clk_div)
+		host->set_clk_div(host->pdev, (clk>>22) & 1);
+
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, clk & 0x1ff);
+	msleep(10);
+}
+
+static void tmio_mmc_clk_stop(struct tmio_mmc_host *host)
+{
+	/* implicit BUG_ON(!res) */
+	if (host->pdata->flags & TMIO_MMC_HAVE_HIGH_REG) {
+		sd_ctrl_write16(host, CTL_CLK_AND_WAIT_CTL, 0x0000);
+		msleep(10);
+	}
+
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, ~0x0100 &
+		sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+	msleep(10);
+}
+
+static void tmio_mmc_clk_start(struct tmio_mmc_host *host)
+{
+	sd_ctrl_write16(host, CTL_SD_CARD_CLK_CTL, 0x0100 |
+		sd_ctrl_read16(host, CTL_SD_CARD_CLK_CTL));
+	msleep(10);
+
+	/* implicit BUG_ON(!res) */
+	if (host->pdata->flags & TMIO_MMC_HAVE_HIGH_REG) {
+		sd_ctrl_write16(host, CTL_CLK_AND_WAIT_CTL, 0x0100);
+		msleep(10);
+	}
+}
+
+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);
+	/* implicit BUG_ON(!res) */
+	if (host->pdata->flags & TMIO_MMC_HAVE_HIGH_REG)
+		sd_ctrl_write16(host, CTL_RESET_SDIO, 0x0000);
+	msleep(10);
+	sd_ctrl_write16(host, CTL_RESET_SD, 0x0001);
+	if (host->pdata->flags & TMIO_MMC_HAVE_HIGH_REG)
+		sd_ctrl_write16(host, CTL_RESET_SDIO, 0x0001);
+	msleep(10);
+}
+
+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(2000))) {
+		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;
+	host->force_pio = false;
+
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	tmio_mmc_reset(host);
+
+	/* Ready for new calls */
+	host->mrq = NULL;
+
+	tmio_mmc_abort_dma(host);
+	mmc_request_done(host->mmc, mrq);
+
+	pm_runtime_mark_last_busy(mmc_dev(host->mmc));
+	pm_runtime_put_autosuspend(mmc_dev(host->mmc));
+}
+
+/* called with host->lock held, interrupts disabled */
+static void tmio_mmc_finish_request(struct tmio_mmc_host *host)
+{
+	struct mmc_request *mrq;
+	unsigned long flags;
+
+	spin_lock_irqsave(&host->lock, flags);
+
+	mrq = host->mrq;
+	if (IS_ERR_OR_NULL(mrq)) {
+		spin_unlock_irqrestore(&host->lock, flags);
+		return;
+	}
+
+	host->cmd = NULL;
+	host->data = NULL;
+	host->force_pio = false;
+
+	cancel_delayed_work(&host->delayed_reset_work);
+
+	host->mrq = NULL;
+	spin_unlock_irqrestore(&host->lock, flags);
+
+	if (mrq->cmd->error || (mrq->data && mrq->data->error))
+		tmio_mmc_abort_dma(host);
+
+	mmc_request_done(host->mmc, mrq);
+
+	pm_runtime_mark_last_busy(mmc_dev(host->mmc));
+	pm_runtime_put_autosuspend(mmc_dev(host->mmc));
+}
+
+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);
+}
+
+/* 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;
+	u32 irq_mask = TMIO_MASK_CMD;
+
+	/* CMD12 is handled by hardware */
+	if (cmd->opcode == MMC_STOP_TRANSMISSION && !cmd->arg) {
+		sd_ctrl_write16(host, CTL_STOP_INTERNAL_ACTION, 0x001);
+		return 0;
+	}
+
+	switch (mmc_resp_type(cmd)) {
+	case MMC_RSP_NONE: c |= RESP_NONE; break;
+	case MMC_RSP_R1:   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, 0x100);
+			c |= TRANSFER_MULTI;
+
+			/*
+			 * Disable auto CMD12 at IO_RW_EXTENDED when
+			 * multiple block transfer
+			 */
+			if ((host->pdata->flags & TMIO_MMC_HAVE_CMD12_CTRL) &&
+			    (cmd->opcode == SD_IO_RW_EXTENDED))
+				c |= NO_CMD12_ISSUE;
+		}
+		if (data->flags & MMC_DATA_READ)
+			c |= TRANSFER_READ;
+	}
+
+	if (!host->native_hotplug)
+		irq_mask &= ~(TMIO_STAT_CARD_REMOVE | TMIO_STAT_CARD_INSERT);
+	tmio_mmc_enable_mmc_irqs(host, irq_mask);
+
+	/* Fire off the command */
+	sd_ctrl_write32(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 (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->chan_tx || host->chan_rx) && !host->force_pio) {
+		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);
+
+	return;
+}
+
+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->chan_rx && !host->force_pio)
+			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) {
+		if (stop->opcode == MMC_STOP_TRANSMISSION && !stop->arg)
+			sd_ctrl_write16(host, CTL_STOP_INTERNAL_ACTION, 0x000);
+		else
+			BUG();
+	}
+
+	schedule_work(&host->done);
+}
+
+static void tmio_mmc_data_irq(struct tmio_mmc_host *host)
+{
+	struct mmc_data *data;
+	spin_lock(&host->lock);
+	data = host->data;
+
+	if (!data)
+		goto out;
+
+	if (host->chan_tx && (data->flags & MMC_DATA_WRITE) && !host->force_pio) {
+		u32 status = sd_ctrl_read32(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_ILL_FUNC)
+				done = true;
+		} else {
+			if (!(status & TMIO_STAT_CMD_BUSY))
+				done = true;
+		}
+
+		if (done) {
+			tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_DATAEND);
+			tasklet_schedule(&host->dma_complete);
+		}
+	} else if (host->chan_rx && (data->flags & MMC_DATA_READ) && !host->force_pio) {
+		tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_DATAEND);
+		tasklet_schedule(&host->dma_complete);
+	} 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;
+	}
+
+	host->cmd = NULL;
+
+	/* 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_read32(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)
+		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) {
+		if (host->data->flags & MMC_DATA_READ) {
+			if (host->force_pio || !host->chan_rx)
+				tmio_mmc_enable_mmc_irqs(host, TMIO_MASK_READOP);
+			else
+				tasklet_schedule(&host->dma_issue);
+		} else {
+			if (host->force_pio || !host->chan_tx)
+				tmio_mmc_enable_mmc_irqs(host, TMIO_MASK_WRITEOP);
+			else
+				tasklet_schedule(&host->dma_issue);
+		}
+	} else {
+		schedule_work(&host->done);
+	}
+
+out:
+	spin_unlock(&host->lock);
+}
+
+static void tmio_mmc_card_irq_status(struct tmio_mmc_host *host,
+				       int *ireg, int *status)
+{
+	*status = sd_ctrl_read32(host, CTL_STATUS);
+	*ireg = *status & TMIO_MASK_IRQ & ~host->sdcard_irq_mask;
+
+	pr_debug_status(*status);
+	pr_debug_status(*ireg);
+
+	/* Clear the status except the interrupt status */
+	sd_ctrl_write32(host, CTL_STATUS, TMIO_MASK_IRQ);
+}
+
+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;
+}
+
+irqreturn_t tmio_mmc_card_detect_irq(int irq, void *devid)
+{
+	unsigned int ireg, status;
+	struct tmio_mmc_host *host = devid;
+
+	tmio_mmc_card_irq_status(host, &ireg, &status);
+	__tmio_mmc_card_detect_irq(host, ireg, status);
+
+	return IRQ_HANDLED;
+}
+EXPORT_SYMBOL(tmio_mmc_card_detect_irq);
+
+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);
+		return true;
+	}
+
+	return false;
+}
+
+irqreturn_t tmio_mmc_sdcard_irq(int irq, void *devid)
+{
+	unsigned int ireg, status;
+	struct tmio_mmc_host *host = devid;
+
+	tmio_mmc_card_irq_status(host, &ireg, &status);
+	__tmio_mmc_sdcard_irq(host, ireg, status);
+
+	return IRQ_HANDLED;
+}
+EXPORT_SYMBOL(tmio_mmc_sdcard_irq);
+
+irqreturn_t tmio_mmc_sdio_irq(int irq, void *devid)
+{
+	struct tmio_mmc_host *host = devid;
+	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 IRQ_HANDLED;
+
+	status = sd_ctrl_read16(host, CTL_SDIO_STATUS);
+	ireg = status & TMIO_SDIO_MASK_ALL & ~host->sdcard_irq_mask;
+
+	sdio_status = status & ~TMIO_SDIO_MASK_ALL;
+	if (pdata->flags & TMIO_MMC_SDIO_STATUS_QUIRK)
+		sdio_status |= 6;
+
+	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 IRQ_HANDLED;
+}
+EXPORT_SYMBOL(tmio_mmc_sdio_irq);
+
+irqreturn_t tmio_mmc_irq(int irq, void *devid)
+{
+	struct tmio_mmc_host *host = devid;
+	unsigned int ireg, status;
+
+	pr_debug("MMC IRQ begin\n");
+
+	tmio_mmc_card_irq_status(host, &ireg, &status);
+	if (__tmio_mmc_card_detect_irq(host, ireg, status))
+		return IRQ_HANDLED;
+	if (__tmio_mmc_sdcard_irq(host, ireg, status))
+		return IRQ_HANDLED;
+
+	tmio_mmc_sdio_irq(irq, devid);
+
+	return IRQ_HANDLED;
+}
+EXPORT_SYMBOL(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 bit mode */
+	if (host->mmc->ios.bus_width == MMC_BUS_WIDTH_4) {
+		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 bit mode\n",
+			       mmc_hostname(host->mmc), data->blksz);
+			return -EINVAL;
+		}
+	}
+
+	tmio_mmc_init_sg(host, data);
+	host->data = data;
+
+	/* Set transfer length / blocksize */
+	sd_ctrl_write16(host, CTL_SD_XFER_LEN, data->blksz);
+	sd_ctrl_write16(host, CTL_XFER_BLK_COUNT, data->blocks);
+
+	tmio_mmc_start_dma(host, data);
+
+	return 0;
+}
+
+/* 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;
+	int ret;
+
+	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);
+
+	pm_runtime_get_sync(mmc_dev(mmc));
+
+	if (mrq->data) {
+		ret = tmio_mmc_start_data(host, mrq->data);
+		if (ret)
+			goto fail;
+	}
+
+	ret = tmio_mmc_start_command(host, mrq->cmd);
+	if (!ret) {
+		schedule_delayed_work(&host->delayed_reset_work,
+				      msecs_to_jiffies(2000));
+		return;
+	}
+
+fail:
+	host->force_pio = false;
+	host->mrq = NULL;
+	mrq->cmd->error = ret;
+	mmc_request_done(mmc, mrq);
+
+	pm_runtime_mark_last_busy(mmc_dev(mmc));
+	pm_runtime_put_autosuspend(mmc_dev(mmc));
+}
+
+static int tmio_mmc_clk_update(struct tmio_mmc_host *host)
+{
+	struct mmc_host *mmc = host->mmc;
+	int ret;
+
+	if (!host->clk_enable)
+		return -ENOTSUPP;
+
+	ret = host->clk_enable(host->pdev, &mmc->f_max);
+	if (!ret)
+		mmc->f_min = mmc->f_max / 512;
+
+	return ret;
+}
+
+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()?
+		 */
+		udelay(200);
+	}
+	/*
+	 * 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);
+		udelay(200);
+	}
+
+	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)
+{
+	switch (bus_width) {
+	case MMC_BUS_WIDTH_1:
+		sd_ctrl_write16(host, CTL_SD_MEM_CARD_OPT, 0x80e0);
+		break;
+	case MMC_BUS_WIDTH_4:
+		sd_ctrl_write16(host, CTL_SD_MEM_CARD_OPT, 0x00e0);
+		break;
+	}
+}
+
+/* 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;
+
+	pm_runtime_get_sync(mmc_dev(mmc));
+
+	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);
+		tmio_mmc_clk_stop(host);
+		break;
+	case MMC_POWER_UP:
+		tmio_mmc_set_clock(host, ios->clock);
+		tmio_mmc_power_on(host, ios->vdd);
+		tmio_mmc_clk_start(host);
+		tmio_mmc_set_bus_width(host, ios->bus_width);
+		break;
+	case MMC_POWER_ON:
+		tmio_mmc_set_clock(host, ios->clock);
+		tmio_mmc_clk_start(host);
+		tmio_mmc_set_bus_width(host, ios->bus_width);
+		break;
+	}
+
+	/* Let things settle. delay taken from winCE driver */
+	udelay(140);
+	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);
+
+	pm_runtime_mark_last_busy(mmc_dev(mmc));
+	pm_runtime_put_autosuspend(mmc_dev(mmc));
+}
+
+static int tmio_mmc_get_ro(struct mmc_host *mmc)
+{
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+	struct tmio_mmc_data *pdata = host->pdata;
+	int ret = mmc_gpio_get_ro(mmc);
+	if (ret >= 0)
+		return ret;
+
+	pm_runtime_get_sync(mmc_dev(mmc));
+	ret = !((pdata->flags & TMIO_MMC_WRPROTECT_DISABLE) ||
+		(sd_ctrl_read32(host, CTL_STATUS) & TMIO_STAT_WRPROTECT));
+	pm_runtime_mark_last_busy(mmc_dev(mmc));
+	pm_runtime_put_autosuspend(mmc_dev(mmc));
+
+	return ret;
+}
+
+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 const 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		= mmc_gpio_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;
+
+	mmc_regulator_get_supply(mmc);
+
+	/* 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 tmio_mmc_data *pdata)
+{
+	const struct device_node *np = pdev->dev.of_node;
+	if (!np)
+		return;
+
+	if (of_get_property(np, "toshiba,mmc-wrprotect-disable", NULL))
+		pdata->flags |= TMIO_MMC_WRPROTECT_DISABLE;
+}
+
+struct tmio_mmc_host*
+tmio_mmc_host_alloc(struct platform_device *pdev)
+{
+	struct tmio_mmc_host *host;
+	struct mmc_host *mmc;
+
+	mmc = mmc_alloc_host(sizeof(struct tmio_mmc_host), &pdev->dev);
+	if (!mmc)
+		return NULL;
+
+	host = mmc_priv(mmc);
+	host->mmc = mmc;
+	host->pdev = pdev;
+
+	return host;
+}
+EXPORT_SYMBOL(tmio_mmc_host_alloc);
+
+void tmio_mmc_host_free(struct tmio_mmc_host *host)
+{
+	mmc_free_host(host->mmc);
+}
+EXPORT_SYMBOL(tmio_mmc_host_free);
+
+int tmio_mmc_host_probe(struct tmio_mmc_host *_host,
+			struct tmio_mmc_data *pdata)
+{
+	struct platform_device *pdev = _host->pdev;
+	struct mmc_host *mmc = _host->mmc;
+	struct resource *res_ctl;
+	int ret;
+	u32 irq_mask = TMIO_MASK_CMD;
+
+	tmio_mmc_of_parse(pdev, pdata);
+
+	if (!(pdata->flags & TMIO_MMC_HAS_IDLE_WAIT))
+		_host->write16_hook = NULL;
+
+	res_ctl = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res_ctl)
+		return -EINVAL;
+
+	ret = mmc_of_parse(mmc);
+	if (ret < 0)
+		goto host_free;
+
+	_host->pdata = pdata;
+	platform_set_drvdata(pdev, mmc);
+
+	_host->set_pwr = pdata->set_pwr;
+	_host->set_clk_div = pdata->set_clk_div;
+
+	ret = tmio_mmc_init_ocr(_host);
+	if (ret < 0)
+		goto host_free;
+
+	_host->ctl = ioremap(res_ctl->start, resource_size(res_ctl));
+	if (!_host->ctl) {
+		ret = -ENOMEM;
+		goto host_free;
+	}
+
+	mmc->ops = &tmio_mmc_ops;
+	mmc->caps |= MMC_CAP_4_BIT_DATA | pdata->capabilities;
+	mmc->caps2 |= pdata->capabilities2;
+	mmc->max_segs = 32;
+	mmc->max_blk_size = 512;
+	mmc->max_blk_count = (PAGE_CACHE_SIZE / mmc->max_blk_size) *
+		mmc->max_segs;
+	mmc->max_req_size = mmc->max_blk_size * mmc->max_blk_count;
+	mmc->max_seg_size = mmc->max_req_size;
+
+	_host->native_hotplug = !(pdata->flags & TMIO_MMC_USE_GPIO_CD ||
+				  mmc->caps & MMC_CAP_NEEDS_POLL ||
+				  mmc->caps & MMC_CAP_NONREMOVABLE ||
+				  mmc->slot.cd_irq >= 0);
+
+	if (tmio_mmc_clk_update(_host) < 0) {
+		mmc->f_max = pdata->hclk;
+		mmc->f_min = mmc->f_max / 512;
+	}
+
+	/*
+	 * Check the sanity of mmc->f_min to prevent tmio_mmc_set_clock() from
+	 * looping forever...
+	 */
+	if (mmc->f_min == 0) {
+		ret = -EINVAL;
+		goto host_free;
+	}
+
+	/*
+	 * 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);
+
+	tmio_mmc_clk_stop(_host);
+	tmio_mmc_reset(_host);
+
+	_host->sdcard_irq_mask = sd_ctrl_read32(_host, CTL_IRQ_MASK);
+	tmio_mmc_disable_mmc_irqs(_host, TMIO_MASK_ALL);
+
+	/* Unmask the IRQs we want to know about */
+	if (!_host->chan_rx)
+		irq_mask |= TMIO_MASK_READOP;
+	if (!_host->chan_tx)
+		irq_mask |= TMIO_MASK_WRITEOP;
+	if (!_host->native_hotplug)
+		irq_mask &= ~(TMIO_STAT_CARD_REMOVE | TMIO_STAT_CARD_INSERT);
+
+	_host->sdcard_irq_mask &= ~irq_mask;
+
+	_host->sdio_irq_enabled = false;
+	if (pdata->flags & TMIO_MMC_SDIO_IRQ) {
+		_host->sdio_irq_mask = TMIO_SDIO_MASK_ALL;
+		sd_ctrl_write16(_host, CTL_SDIO_IRQ_MASK, _host->sdio_irq_mask);
+		sd_ctrl_write16(_host, CTL_TRANSACTION_CTL, 0x0000);
+	}
+
+	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_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 < 0) {
+		tmio_mmc_host_remove(_host);
+		return ret;
+	}
+
+	dev_pm_qos_expose_latency_limit(&pdev->dev, 100);
+
+	if (pdata->flags & TMIO_MMC_USE_GPIO_CD) {
+		ret = mmc_gpio_request_cd(mmc, pdata->cd_gpio, 0);
+		if (ret < 0) {
+			tmio_mmc_host_remove(_host);
+			return ret;
+		}
+		mmc_gpiod_request_cd_irq(mmc);
+	}
+
+	return 0;
+
+host_free:
+
+	return ret;
+}
+EXPORT_SYMBOL(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;
+
+	if (!host->native_hotplug)
+		pm_runtime_get_sync(&pdev->dev);
+
+	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);
+
+	pm_runtime_put_sync(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+
+	iounmap(host->ctl);
+}
+EXPORT_SYMBOL(tmio_mmc_host_remove);
+
+#ifdef CONFIG_PM
+int tmio_mmc_host_runtime_suspend(struct device *dev)
+{
+	struct mmc_host *mmc = dev_get_drvdata(dev);
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+
+	tmio_mmc_disable_mmc_irqs(host, TMIO_MASK_ALL);
+
+	if (host->clk_cache)
+		tmio_mmc_clk_stop(host);
+
+	if (host->clk_disable)
+		host->clk_disable(host->pdev);
+
+	return 0;
+}
+EXPORT_SYMBOL(tmio_mmc_host_runtime_suspend);
+
+int tmio_mmc_host_runtime_resume(struct device *dev)
+{
+	struct mmc_host *mmc = dev_get_drvdata(dev);
+	struct tmio_mmc_host *host = mmc_priv(mmc);
+
+	tmio_mmc_reset(host);
+	tmio_mmc_clk_update(host);
+
+	if (host->clk_cache) {
+		tmio_mmc_set_clock(host, host->clk_cache);
+		tmio_mmc_clk_start(host);
+	}
+
+	tmio_mmc_enable_dma(host, true);
+
+	return 0;
+}
+EXPORT_SYMBOL(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..e2cdd5fb1
--- /dev/null
+++ b/drivers/mmc/host/toshsd.c
@@ -0,0 +1,708 @@
+/*
+ *  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
+ *
+ * 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; either version 2 of the License, or (at
+ * your option) any later version.
+ */
+
+#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/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 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;
+
+	mmc_add_host(mmc);
+
+	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;
+
+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..b6c0d89e5
--- /dev/null
+++ b/drivers/mmc/host/toshsd.h
@@ -0,0 +1,176 @@
+/*
+ *  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
+ *
+ * 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; either version 2 of the License, or (at
+ * your option) any later version.
+ */
+
+#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/usdhi6rol0.c b/drivers/mmc/host/usdhi6rol0.c
new file mode 100644
index 000000000..54b082b18
--- /dev/null
+++ b/drivers/mmc/host/usdhi6rol0.c
@@ -0,0 +1,1846 @@
+/*
+ * Copyright (C) 2013-2014 Renesas Electronics Europe Ltd.
+ * Author: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License as
+ * published by the Free Software Foundation.
+ */
+
+#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/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
+
+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;
+};
+
+/*			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_slave_channel(mmc_dev(host->mmc), "tx");
+	dev_dbg(mmc_dev(host->mmc), "%s: TX: got channel %p\n", __func__,
+		host->chan_tx);
+
+	if (!host->chan_tx)
+		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_slave_channel(mmc_dev(host->mmc), "rx");
+	dev_dbg(mmc_dev(host->mmc), "%s: RX: got channel %p\n", __func__,
+		host->chan_rx);
+
+	if (!host->chan_rx)
+		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 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,
+};
+
+/*			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;
+		}
+		/* 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 = container_of(work, struct 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;
+
+	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);
+		/* mrq can be NULL in this actually impossible case */
+	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:
+		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(host->sg), host->sg->offset);
+		usdhi6_sg_unmap(host, true);
+		/*
+		 * If USDHI6_WAIT_FOR_DATA_END times out, we have already unmapped
+		 * the page
+		 */
+	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 || irq_sdio < 0)
+		return -ENODEV;
+
+	mmc = mmc_alloc_host(sizeof(struct usdhi6_host), dev);
+	if (!mmc)
+		return -ENOMEM;
+
+	ret = mmc_of_parse(mmc);
+	if (ret < 0)
+		goto e_free_mmc;
+
+	mmc_regulator_get_supply(mmc);
+
+	host		= mmc_priv(mmc);
+	host->mmc	= mmc;
+	host->wait	= USDHI6_WAIT_FOR_REQUEST;
+	host->timeout	= msecs_to_jiffies(4000);
+
+	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))
+		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) {
+		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_UHS_SDR50 | MMC_CAP_UHS_DDR50 | 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_CACHE_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_clk_off;
+
+	return 0;
+
+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",
+		.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..d2c386f09
--- /dev/null
+++ b/drivers/mmc/host/ushc.c
@@ -0,0 +1,569 @@
+/*
+ * USB SD Host Controller (USHC) controller driver.
+ *
+ * Copyright (C) 2010 Cambridge Silicon Radio Ltd.
+ *
+ * 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; either version 2 of the License, or (at
+ * your option) any later version.
+ *
+ * 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,
+				  sg_virt(data->sg), data->sg->length,
+				  data_callback, ushc);
+		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);
+	if (ret < 0)
+		goto out;
+
+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;
+
+	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..63fac78b3
--- /dev/null
+++ b/drivers/mmc/host/via-sdmmc.c
@@ -0,0 +1,1339 @@
+/*
+ *  drivers/mmc/host/via-sdmmc.c - VIA SD/MMC Card Reader driver
+ *  Copyright (c) 2008, VIA Technologies Inc. All Rights Reserved.
+ *
+ * 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; either version 2 of the License, or (at
+ * your option) any later version.
+ */
+
+#include <linux/pci.h>
+#include <linux/module.h>
+#include <linux/dma-mapping.h>
+#include <linux/highmem.h>
+#include <linux/delay.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
+
+static 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;
+	u32 cmdctrl = 0;
+
+	WARN_ON(host->cmd);
+
+	data = cmd->data;
+	mod_timer(&host->timer, jiffies + HZ);
+	host->cmd = cmd;
+
+	/*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);
+	}
+
+	mmiowb();
+	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);
+
+	mmiowb();
+	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);
+
+	mmiowb();
+	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);
+
+	mmiowb();
+	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 (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;
+
+	mmiowb();
+out:
+	spin_unlock(&sdhost->lock);
+
+	return result;
+}
+
+static void via_sdc_timeout(unsigned long ulongdata)
+{
+	struct via_crdr_mmc_host *sdhost;
+	unsigned long flags;
+
+	sdhost = (struct via_crdr_mmc_host *)ulongdata;
+
+	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);
+		}
+	}
+
+	mmiowb();
+	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);
+		}
+
+		mmiowb();
+		spin_unlock_irqrestore(&host->lock, flags);
+
+		via_reset_pcictrl(host);
+
+		spin_lock_irqsave(&host->lock, flags);
+	}
+
+	mmiowb();
+	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;
+
+	init_timer(&host->timer);
+	host->timer.data = (unsigned long)host;
+	host->timer.function = via_sdc_timeout;
+
+	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_nocache(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;
+
+	mmc_add_host(mmc);
+
+	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);
+	mmiowb();
+
+	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);
+		mmiowb();
+		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);
+}
+
+#ifdef CONFIG_PM
+
+static void 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 via_sd_suspend(struct pci_dev *pcidev, pm_message_t state)
+{
+	struct via_crdr_mmc_host *host;
+
+	host = pci_get_drvdata(pcidev);
+
+	via_save_pcictrlreg(host);
+	via_save_sdcreg(host);
+
+	pci_save_state(pcidev);
+	pci_enable_wake(pcidev, pci_choose_state(pcidev, state), 0);
+	pci_disable_device(pcidev);
+	pci_set_power_state(pcidev, pci_choose_state(pcidev, state));
+
+	return 0;
+}
+
+static int via_sd_resume(struct pci_dev *pcidev)
+{
+	struct via_crdr_mmc_host *sdhost;
+	int ret = 0;
+	u8 gatt;
+
+	sdhost = pci_get_drvdata(pcidev);
+
+	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);
+
+	pci_set_power_state(pcidev, PCI_D0);
+	pci_restore_state(pcidev);
+	ret = pci_enable_device(pcidev);
+	if (ret)
+		return ret;
+
+	via_restore_pcictrlreg(sdhost);
+	via_init_sdc_pm(sdhost);
+
+	return ret;
+}
+
+#else /* CONFIG_PM */
+
+#define via_sd_suspend NULL
+#define via_sd_resume NULL
+
+#endif /* CONFIG_PM */
+
+static struct pci_driver via_sd_driver = {
+	.name = DRV_NAME,
+	.id_table = via_ids,
+	.probe = via_sd_probe,
+	.remove = via_sd_remove,
+	.suspend = via_sd_suspend,
+	.resume = via_sd_resume,
+};
+
+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..c1b16619e
--- /dev/null
+++ b/drivers/mmc/host/vub300.c
@@ -0,0 +1,2488 @@
+/*
+ * 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)
+ *
+ * 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
+ *
+ * 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[0];
+} __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 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), HZ);
+	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 if (vub300->irqs_queued)
+			vub300->irqs_queued += 1;
+		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 &&
+		   mmc_card_present(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(unsigned long data)
+{				/* softirq */
+	struct vub300_mmc_host *vub300 = (struct vub300_mmc_host *)data;
+	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(unsigned long data)
+{
+	struct vub300_mmc_host *vub300 = (struct vub300_mmc_host *)data;
+	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, HZ);
+			kfree(xfer_buffer);
+			if (retval < 0) {
+				strncpy(vub300->vub_name,
+					"SDIO pseudocode download failed",
+					sizeof(vub300->vub_name));
+				return;
+			}
+		} 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, HZ);
+			kfree(xfer_buffer);
+			if (retval < 0) {
+				strncpy(vub300->vub_name,
+					"SDIO pseudocode download failed",
+					sizeof(vub300->vub_name));
+				return;
+			}
+		} 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;
+	}
+}
+
+/*
+ * 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),
+			 "/*(DEBLOBBED)*/", 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 += snprintf(vub300->vub_name + l,
+			      sizeof(vub300->vub_name) - l, "_%04X%04X",
+			      sf->vendor, sf->device);
+	};
+	/*(DEBLOBBED)*/
+	dev_info(&vub300->udev->dev, "requesting offload firmware %s\n",
+		 vub300->vub_name);
+	retval = reject_firmware(&fw, vub300->vub_name, &card->dev);
+	if (retval < 0) {
+		strncpy(vub300->vub_name, "/*(DEBLOBBED)*/",
+			sizeof(vub300->vub_name));
+		retval = reject_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 (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 ||
+		    !mmc_card_present(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, HZ);
+	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, HZ);
+		/* 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, HZ);
+		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) {
+		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);
+	} else {
+		vub300->irq_enabled = 0;
+	}
+	kref_put(&vub300->kref, vub300_delete);
+}
+
+static void vub300_init_card(struct mmc_host *mmc, struct mmc_card *card)
+{				/* NOT irq */
+	struct vub300_mmc_host *vub300 = mmc_priv(mmc);
+	dev_info(&vub300->udev->dev, "NO host QUIRKS for this card\n");
+}
+
+static 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,
+	.init_card = vub300_init_card,
+};
+
+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",
+		 udev->descriptor.idVendor, udev->descriptor.idProduct,
+		 manufacturer, product, serial_number);
+	command_out_urb = usb_alloc_urb(0, GFP_KERNEL);
+	if (!command_out_urb) {
+		retval = -ENOMEM;
+		dev_err(&udev->dev, "not enough memory for command_out_urb\n");
+		goto error0;
+	}
+	command_res_urb = usb_alloc_urb(0, GFP_KERNEL);
+	if (!command_res_urb) {
+		retval = -ENOMEM;
+		dev_err(&udev->dev, "not enough memory for command_res_urb\n");
+		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), HZ);
+	if (retval < 0)
+		goto error5;
+	retval =
+		usb_control_msg(vub300->udev, usb_rcvctrlpipe(vub300->udev, 0),
+				SET_ROM_WAIT_STATES,
+				USB_DIR_OUT | USB_TYPE_VENDOR | USB_RECIP_DEVICE,
+				firmware_rom_wait_states, 0x0000, NULL, 0, HZ);
+	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), HZ);
+	if (retval < 0) {
+		goto error4;
+	} 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 {
+		goto error4;
+	}
+	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);
+	init_timer(&vub300->sg_transfer_timer);
+	vub300->sg_transfer_timer.data = (unsigned long)vub300;
+	vub300->sg_transfer_timer.function = vub300_sg_timed_out;
+	kref_get(&vub300->kref);
+	init_timer(&vub300->inactivity_timer);
+	vub300->inactivity_timer.data = (unsigned long)vub300;
+	vub300->inactivity_timer.function = vub300_inactivity_timer_expired;
+	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));
+	mmc_add_host(mmc);
+	return 0;
+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..ca183ea76
--- /dev/null
+++ b/drivers/mmc/host/wbsd.c
@@ -0,0 +1,2013 @@
+/*
+ *  linux/drivers/mmc/host/wbsd.c - Winbond W83L51xD SD/MMC driver
+ *
+ *  Copyright (C) 2004-2007 Pierre Ossman, All Rights Reserved.
+ *
+ * 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; either version 2 of the License, or (at
+ * your option) any later version.
+ *
+ *
+ * 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/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_sg_to_buffer(struct wbsd_host *host)
+{
+	return sg_virt(host->cur_sg);
+}
+
+static inline void wbsd_sg_to_dma(struct wbsd_host *host, struct mmc_data *data)
+{
+	unsigned int len, i;
+	struct scatterlist *sg;
+	char *dmabuf = host->dma_buffer;
+	char *sgbuf;
+
+	sg = data->sg;
+	len = data->sg_len;
+
+	for (i = 0; i < len; i++) {
+		sgbuf = sg_virt(&sg[i]);
+		memcpy(dmabuf, sgbuf, sg[i].length);
+		dmabuf += sg[i].length;
+	}
+}
+
+static inline void wbsd_dma_to_sg(struct wbsd_host *host, struct mmc_data *data)
+{
+	unsigned int len, i;
+	struct scatterlist *sg;
+	char *dmabuf = host->dma_buffer;
+	char *sgbuf;
+
+	sg = data->sg;
+	len = data->sg_len;
+
+	for (i = 0; i < len; i++) {
+		sgbuf = sg_virt(&sg[i]);
+		memcpy(sgbuf, dmabuf, sg[i].length);
+		dmabuf += sg[i].length;
+	}
+}
+
+/*
+ * 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, fsr, fifo;
+
+	/*
+	 * Handle excessive data.
+	 */
+	if (host->num_sg == 0)
+		return;
+
+	buffer = wbsd_sg_to_buffer(host) + host->offset;
+
+	/*
+	 * 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 = inb(host->base + WBSD_DFR);
+			buffer++;
+			host->offset++;
+			host->remain--;
+
+			data->bytes_xfered++;
+
+			/*
+			 * End of scatter list entry?
+			 */
+			if (host->remain == 0) {
+				/*
+				 * Get next entry. Check if last.
+				 */
+				if (!wbsd_next_sg(host))
+					return;
+
+				buffer = wbsd_sg_to_buffer(host);
+			}
+		}
+	}
+
+	/*
+	 * 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, fsr, fifo;
+
+	/*
+	 * Check that we aren't being called after the
+	 * entire buffer has been transferred.
+	 */
+	if (host->num_sg == 0)
+		return;
+
+	buffer = wbsd_sg_to_buffer(host) + host->offset;
+
+	/*
+	 * 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, host->base + WBSD_DFR);
+			buffer++;
+			host->offset++;
+			host->remain--;
+
+			data->bytes_xfered++;
+
+			/*
+			 * End of scatter list entry?
+			 */
+			if (host->remain == 0) {
+				/*
+				 * Get next entry. Check if last.
+				 */
+				if (!wbsd_next_sg(host))
+					return;
+
+				buffer = wbsd_sg_to_buffer(host);
+			}
+		}
+	}
+
+	/*
+	 * 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 11:
+		case 17:
+		case 18:
+		case 20:
+		case 24:
+		case 25:
+		case 26:
+		case 27:
+		case 30:
+		case 42:
+		case 56:
+			break;
+
+		/* ACMDs. We don't keep track of state, so we just treat them
+		 * like any other command. */
+		case 51:
+			break;
+
+		default:
+#ifdef CONFIG_MMC_DEBUG
+			pr_warn("%s: Data command %d is not supported by this controller\n",
+				mmc_hostname(host->mmc), cmd->opcode);
+#endif
+			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(unsigned long data)
+{
+	struct wbsd_host *host = (struct wbsd_host *)data;
+
+	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
+	 */
+	init_timer(&host->ignore_timer);
+	host->ignore_timer.data = (unsigned long)host;
+	host->ignore_timer.function = wbsd_reset_ignore;
+
+	/*
+	 * 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_REPEAT | __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);
+
+	/*
+	 * ISA DMA must be aligned on a 64k basis.
+	 */
+	if ((host->dma_addr & 0xffff) != 0)
+		goto kfree;
+	/*
+	 * ISA cannot access memory above 16 MB.
+	 */
+	else if (host->dma_addr >= 0x1000000)
+		goto kfree;
+
+	host->dma = dma;
+
+	return;
+
+kfree:
+	/*
+	 * 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(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)
+{
+	if (host->dma_addr) {
+		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);
+
+	mmc_add_host(mmc);
+
+	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,
+	},
+};
+
+#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_named(nopnp, param_nopnp, uint, 0444);
+#endif
+module_param_named(io, param_io, uint, 0444);
+module_param_named(irq, param_irq, uint, 0444);
+module_param_named(dma, param_dma, int, 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..0877866f8
--- /dev/null
+++ b/drivers/mmc/host/wbsd.h
@@ -0,0 +1,185 @@
+/*
+ *  linux/drivers/mmc/host/wbsd.h - Winbond W83L51xD SD/MMC driver
+ *
+ *  Copyright (C) 2004-2007 Pierre Ossman, All Rights Reserved.
+ *
+ * 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; either version 2 of the License, or (at
+ * your option) any later version.
+ */
+
+#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..5af00559e
--- /dev/null
+++ b/drivers/mmc/host/wmt-sdmmc.c
@@ -0,0 +1,1004 @@
+/*
+ *  WM8505/WM8650 SD/MMC Host Controller
+ *
+ *  Copyright (C) 2010 Tony Prisk
+ *  Copyright (C) 2008 WonderMedia Technologies, Inc.
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License version 2 as
+ *  published by the Free Software Foundation
+ */
+
+#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/gpio.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 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;
+	}
+
+	clk_prepare_enable(priv->clk_sdmmc);
+
+	/* configure the controller to a known 'ready' state */
+	wmt_reset_hardware(mmc);
+
+	mmc_add_host(mmc);
+
+	dev_info(&pdev->dev, "WMT SDHC Controller initialized\n");
+
+	return 0;
+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 platform_device *pdev = to_platform_device(dev);
+	struct mmc_host *mmc = platform_get_drvdata(pdev);
+	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 platform_device *pdev = to_platform_device(dev);
+	struct mmc_host *mmc = platform_get_drvdata(pdev);
+	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,
+		.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);