Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/hskinnemoen/atmel-mci-2.6.28

Conflicts:
	drivers/mmc/host/atmel-mci.c

9 files changed, 1126 insertions, 412 deletions

diff --git a/arch/avr32/boards/atngw100/setup.c b/arch/avr32/boards/atngw100/setup.c
index b8286f1ce85..f3085208959 100644
--- a/arch/avr32/boards/atngw100/setup.c
+++ b/arch/avr32/boards/atngw100/setup.c
@@ -53,8 +53,11 @@ static struct spi_board_info spi0_board_info[] __initdata = {
 };
 
 static struct mci_platform_data __initdata mci0_data = {
-	.detect_pin	= GPIO_PIN_PC(25),
-	.wp_pin		= GPIO_PIN_PE(0),
+	.slot[0] = {
+		.bus_width	= 4,
+		.detect_pin	= GPIO_PIN_PC(25),
+		.wp_pin		= GPIO_PIN_PE(0),
+	},
 };
 
 /*
diff --git a/arch/avr32/boards/atstk1000/atstk1002.c b/arch/avr32/boards/atstk1000/atstk1002.c
index dfc3443e23a..4fedbc4488d 100644
--- a/arch/avr32/boards/atstk1000/atstk1002.c
+++ b/arch/avr32/boards/atstk1000/atstk1002.c
@@ -264,16 +264,20 @@ void __init setup_board(void)
 
 #ifndef CONFIG_BOARD_ATSTK100X_SW2_CUSTOM
 
+static struct mci_platform_data __initdata mci0_data = {
+	.slot[0] = {
+		.bus_width	= 4,
+
 /* MMC card detect requires MACB0 *NOT* be used */
 #ifdef CONFIG_BOARD_ATSTK1002_SW6_CUSTOM
-static struct mci_platform_data __initdata mci0_data = {
-	.detect_pin	= GPIO_PIN_PC(14),	/* gpio30/sdcd */
-	.wp_pin		= GPIO_PIN_PC(15),	/* gpio31/sdwp */
-};
-#define MCI_PDATA	&mci0_data
+		.detect_pin	= GPIO_PIN_PC(14), /* gpio30/sdcd */
+		.wp_pin		= GPIO_PIN_PC(15), /* gpio31/sdwp */
 #else
-#define MCI_PDATA	NULL
+		.detect_pin	= -ENODEV,
+		.wp_pin		= -ENODEV,
 #endif	/* SW6 for sd{cd,wp} routing */
+	},
+};
 
 #endif	/* SW2 for MMC signal routing */
 
@@ -326,7 +330,7 @@ static int __init atstk1002_init(void)
 	at32_add_device_spi(1, spi1_board_info, ARRAY_SIZE(spi1_board_info));
 #endif
 #ifndef CONFIG_BOARD_ATSTK100X_SW2_CUSTOM
-	at32_add_device_mci(0, MCI_PDATA);
+	at32_add_device_mci(0, &mci0_pdata);
 #endif
 #ifdef CONFIG_BOARD_ATSTK1002_SW5_CUSTOM
 	set_hw_addr(at32_add_device_eth(1, &eth_data[1]));
diff --git a/arch/avr32/boards/atstk1000/atstk1003.c b/arch/avr32/boards/atstk1000/atstk1003.c
index 0cf664174c1..acc61235b89 100644
--- a/arch/avr32/boards/atstk1000/atstk1003.c
+++ b/arch/avr32/boards/atstk1000/atstk1003.c
@@ -66,6 +66,16 @@ static struct spi_board_info spi1_board_info[] __initdata = { {
 } };
 #endif
 
+#ifndef CONFIG_BOARD_ATSTK100X_SW2_CUSTOM
+static struct mci_platform_data __initdata mci0_data = {
+	.slot[0] = {
+		.bus_width	= 4,
+		.detect_pin	= -ENODEV,
+		.wp_pin		= -ENODEV,
+	},
+};
+#endif
+
 #ifdef CONFIG_BOARD_ATSTK1000_EXTDAC
 static void __init atstk1003_setup_extdac(void)
 {
@@ -154,7 +164,7 @@ static int __init atstk1003_init(void)
 	at32_add_device_spi(1, spi1_board_info, ARRAY_SIZE(spi1_board_info));
 #endif
 #ifndef CONFIG_BOARD_ATSTK100X_SW2_CUSTOM
-	at32_add_device_mci(0, NULL);
+	at32_add_device_mci(0, &mci0_data);
 #endif
 	at32_add_device_usba(0, NULL);
 #ifndef CONFIG_BOARD_ATSTK100X_SW3_CUSTOM
diff --git a/arch/avr32/boards/atstk1000/atstk1004.c b/arch/avr32/boards/atstk1000/atstk1004.c
index 50a5273e591..d6a2d02f032 100644
--- a/arch/avr32/boards/atstk1000/atstk1004.c
+++ b/arch/avr32/boards/atstk1000/atstk1004.c
@@ -71,6 +71,16 @@ static struct spi_board_info spi1_board_info[] __initdata = { {
 } };
 #endif
 
+#ifndef CONFIG_BOARD_ATSTK100X_SW2_CUSTOM
+static struct mci_platform_data __initdata mci0_data = {
+	.slot[0] = {
+		.bus_width	= 4,
+		.detect_pin	= -ENODEV,
+		.wp_pin		= -ENODEV,
+	},
+};
+#endif
+
 #ifdef CONFIG_BOARD_ATSTK1000_EXTDAC
 static void __init atstk1004_setup_extdac(void)
 {
@@ -137,7 +147,7 @@ static int __init atstk1004_init(void)
 	at32_add_device_spi(1, spi1_board_info, ARRAY_SIZE(spi1_board_info));
 #endif
 #ifndef CONFIG_BOARD_ATSTK100X_SW2_CUSTOM
-	at32_add_device_mci(0, NULL);
+	at32_add_device_mci(0, &mci0_data);
 #endif
 	at32_add_device_lcdc(0, &atstk1000_lcdc_data,
 			     fbmem_start, fbmem_size, 0);
diff --git a/arch/avr32/include/asm/atmel-mci.h b/arch/avr32/include/asm/atmel-mci.h
index c2ea6e1c9aa..59f3fadd0b6 100644
--- a/arch/avr32/include/asm/atmel-mci.h
+++ b/arch/avr32/include/asm/atmel-mci.h
@@ -1,9 +1,39 @@
 #ifndef __ASM_AVR32_ATMEL_MCI_H
 #define __ASM_AVR32_ATMEL_MCI_H
 
-struct mci_platform_data {
+#define ATMEL_MCI_MAX_NR_SLOTS	2
+
+struct dma_slave;
+
+/**
+ * struct mci_slot_pdata - board-specific per-slot configuration
+ * @bus_width: Number of data lines wired up the slot
+ * @detect_pin: GPIO pin wired to the card detect switch
+ * @wp_pin: GPIO pin wired to the write protect sensor
+ *
+ * If a given slot is not present on the board, @bus_width should be
+ * set to 0. The other fields are ignored in this case.
+ *
+ * Any pins that aren't available should be set to a negative value.
+ *
+ * Note that support for multiple slots is experimental -- some cards
+ * might get upset if we don't get the clock management exactly right.
+ * But in most cases, it should work just fine.
+ */
+struct mci_slot_pdata {
+	unsigned int		bus_width;
 	int			detect_pin;
 	int			wp_pin;
 };
 
+/**
+ * struct mci_platform_data - board-specific MMC/SDcard configuration
+ * @dma_slave: DMA slave interface to use in data transfers, or NULL.
+ * @slot: Per-slot configuration data.
+ */
+struct mci_platform_data {
+	struct dma_slave	*dma_slave;
+	struct mci_slot_pdata	slot[ATMEL_MCI_MAX_NR_SLOTS];
+};
+
 #endif /* __ASM_AVR32_ATMEL_MCI_H */
diff --git a/arch/avr32/mach-at32ap/at32ap700x.c b/arch/avr32/mach-at32ap/at32ap700x.c
index e01dbe4ebb4..f1b9a3ac273 100644
--- a/arch/avr32/mach-at32ap/at32ap700x.c
+++ b/arch/avr32/mach-at32ap/at32ap700x.c
@@ -1272,10 +1272,14 @@ static struct clk atmel_mci0_pclk = {
 struct platform_device *__init
 at32_add_device_mci(unsigned int id, struct mci_platform_data *data)
 {
-	struct mci_platform_data	_data;
 	struct platform_device		*pdev;
+	struct dw_dma_slave		*dws;
 
-	if (id != 0)
+	if (id != 0 || !data)
+		return NULL;
+
+	/* Must have at least one usable slot */
+	if (!data->slot[0].bus_width && !data->slot[1].bus_width)
 		return NULL;
 
 	pdev = platform_device_alloc("atmel_mci", id);
@@ -1286,28 +1290,76 @@ at32_add_device_mci(unsigned int id, struct mci_platform_data *data)
 				ARRAY_SIZE(atmel_mci0_resource)))
 		goto fail;
 
-	if (!data) {
-		data = &_data;
-		memset(data, -1, sizeof(struct mci_platform_data));
-		data->detect_pin = GPIO_PIN_NONE;
-		data->wp_pin = GPIO_PIN_NONE;
-	}
+	if (data->dma_slave)
+		dws = kmemdup(to_dw_dma_slave(data->dma_slave),
+				sizeof(struct dw_dma_slave), GFP_KERNEL);
+	else
+		dws = kzalloc(sizeof(struct dw_dma_slave), GFP_KERNEL);
+
+	dws->slave.dev = &pdev->dev;
+	dws->slave.dma_dev = &dw_dmac0_device.dev;
+	dws->slave.reg_width = DMA_SLAVE_WIDTH_32BIT;
+	dws->cfg_hi = (DWC_CFGH_SRC_PER(0)
+				| DWC_CFGH_DST_PER(1));
+	dws->cfg_lo &= ~(DWC_CFGL_HS_DST_POL
+				| DWC_CFGL_HS_SRC_POL);
+
+	data->dma_slave = &dws->slave;
 
 	if (platform_device_add_data(pdev, data,
 				sizeof(struct mci_platform_data)))
 		goto fail;
 
-	select_peripheral(PA(10), PERIPH_A, 0);	/* CLK	 */
-	select_peripheral(PA(11), PERIPH_A, 0);	/* CMD	 */
-	select_peripheral(PA(12), PERIPH_A, 0);	/* DATA0 */
-	select_peripheral(PA(13), PERIPH_A, 0);	/* DATA1 */
-	select_peripheral(PA(14), PERIPH_A, 0);	/* DATA2 */
-	select_peripheral(PA(15), PERIPH_A, 0);	/* DATA3 */
+	/* CLK line is common to both slots */
+	select_peripheral(PA(10), PERIPH_A, 0);
 
-	if (gpio_is_valid(data->detect_pin))
-		at32_select_gpio(data->detect_pin, 0);
-	if (gpio_is_valid(data->wp_pin))
-		at32_select_gpio(data->wp_pin, 0);
+	switch (data->slot[0].bus_width) {
+	case 4:
+		select_peripheral(PA(13), PERIPH_A, 0);	/* DATA1 */
+		select_peripheral(PA(14), PERIPH_A, 0);	/* DATA2 */
+		select_peripheral(PA(15), PERIPH_A, 0);	/* DATA3 */
+		/* fall through */
+	case 1:
+		select_peripheral(PA(11), PERIPH_A, 0);	/* CMD	 */
+		select_peripheral(PA(12), PERIPH_A, 0);	/* DATA0 */
+
+		if (gpio_is_valid(data->slot[0].detect_pin))
+			at32_select_gpio(data->slot[0].detect_pin, 0);
+		if (gpio_is_valid(data->slot[0].wp_pin))
+			at32_select_gpio(data->slot[0].wp_pin, 0);
+		break;
+	case 0:
+		/* Slot is unused */
+		break;
+	default:
+		goto fail;
+	}
+
+	switch (data->slot[1].bus_width) {
+	case 4:
+		select_peripheral(PB(8),  PERIPH_B, 0);	/* DATA1 */
+		select_peripheral(PB(9),  PERIPH_B, 0);	/* DATA2 */
+		select_peripheral(PB(10), PERIPH_B, 0);	/* DATA3 */
+		/* fall through */
+	case 1:
+		select_peripheral(PB(6),  PERIPH_B, 0); /* CMD	 */
+		select_peripheral(PB(7),  PERIPH_B, 0); /* DATA0 */
+
+		if (gpio_is_valid(data->slot[1].detect_pin))
+			at32_select_gpio(data->slot[1].detect_pin, 0);
+		if (gpio_is_valid(data->slot[1].wp_pin))
+			at32_select_gpio(data->slot[1].wp_pin, 0);
+		break;
+	case 0:
+		/* Slot is unused */
+		break;
+	default:
+		if (!data->slot[0].bus_width)
+			goto fail;
+
+		data->slot[1].bus_width = 0;
+		break;
+	}
 
 	atmel_mci0_pclk.dev = &pdev->dev;
 
diff --git a/drivers/mmc/host/Kconfig b/drivers/mmc/host/Kconfig
index 38468b2bd1a..dfa585f7fea 100644
--- a/drivers/mmc/host/Kconfig
+++ b/drivers/mmc/host/Kconfig
@@ -114,6 +114,17 @@ config MMC_ATMELMCI
 
 	  If unsure, say N.
 
+config MMC_ATMELMCI_DMA
+	bool "Atmel MCI DMA support (EXPERIMENTAL)"
+	depends on MMC_ATMELMCI && DMA_ENGINE && EXPERIMENTAL
+	help
+	  Say Y here to have the Atmel MCI driver use a DMA engine to
+	  do data transfers and thus increase the throughput and
+	  reduce the CPU utilization. Note that this is highly
+	  experimental and may cause the driver to lock up.
+
+	  If unsure, say N.
+
 config MMC_IMX
 	tristate "Motorola i.MX Multimedia Card Interface support"
 	depends on ARCH_IMX
diff --git a/drivers/mmc/host/atmel-mci-regs.h b/drivers/mmc/host/atmel-mci-regs.h
index 26bd80e6503..b58364ed6bb 100644
--- a/drivers/mmc/host/atmel-mci-regs.h
+++ b/drivers/mmc/host/atmel-mci-regs.h
@@ -25,8 +25,10 @@
 #define MCI_SDCR		0x000c	/* SD Card / SDIO */
 # define MCI_SDCSEL_SLOT_A	(  0 <<  0)	/* Select SD slot A */
 # define MCI_SDCSEL_SLOT_B	(  1 <<  0)	/* Select SD slot A */
-# define MCI_SDCBUS_1BIT	(  0 <<  7)	/* 1-bit data bus */
-# define MCI_SDCBUS_4BIT	(  1 <<  7)	/* 4-bit data bus */
+# define MCI_SDCSEL_MASK	(  3 <<  0)
+# define MCI_SDCBUS_1BIT	(  0 <<  6)	/* 1-bit data bus */
+# define MCI_SDCBUS_4BIT	(  2 <<  6)	/* 4-bit data bus */
+# define MCI_SDCBUS_MASK	(  3 <<  6)
 #define MCI_ARGR		0x0010	/* Command Argument */
 #define MCI_CMDR		0x0014	/* Command */
 # define MCI_CMDR_CMDNB(x)	((x) <<  0)	/* Command Opcode */
diff --git a/drivers/mmc/host/atmel-mci.c b/drivers/mmc/host/atmel-mci.c
index 00008967ef7..7a3f2436b01 100644
--- a/drivers/mmc/host/atmel-mci.c
+++ b/drivers/mmc/host/atmel-mci.c
@@ -11,6 +11,8 @@
 #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>
@@ -33,64 +35,178 @@
 #include "atmel-mci-regs.h"
 
 #define ATMCI_DATA_ERROR_FLAGS	(MCI_DCRCE | MCI_DTOE | MCI_OVRE | MCI_UNRE)
+#define ATMCI_DMA_THRESHOLD	16
 
 enum {
 	EVENT_CMD_COMPLETE = 0,
-	EVENT_DATA_ERROR,
-	EVENT_DATA_COMPLETE,
-	EVENT_STOP_SENT,
-	EVENT_STOP_COMPLETE,
 	EVENT_XFER_COMPLETE,
+	EVENT_DATA_COMPLETE,
+	EVENT_DATA_ERROR,
+};
+
+enum atmel_mci_state {
+	STATE_IDLE = 0,
+	STATE_SENDING_CMD,
+	STATE_SENDING_DATA,
+	STATE_DATA_BUSY,
+	STATE_SENDING_STOP,
+	STATE_DATA_ERROR,
+};
+
+struct atmel_mci_dma {
+#ifdef CONFIG_MMC_ATMELMCI_DMA
+	struct dma_client		client;
+	struct dma_chan			*chan;
+	struct dma_async_tx_descriptor	*data_desc;
+#endif
 };
 
+/**
+ * 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 code, if any.
+ * @pio_offset: Offset into the current scatterlist entry.
+ * @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.
+ * @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.
+ * @mode_reg: Value of the MR 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.
+ *
+ * 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 interupt 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 {
-	struct mmc_host		*mmc;
+	spinlock_t		lock;
 	void __iomem		*regs;
 
 	struct scatterlist	*sg;
 	unsigned int		pio_offset;
 
+	struct atmel_mci_slot	*cur_slot;
 	struct mmc_request	*mrq;
 	struct mmc_command	*cmd;
 	struct mmc_data		*data;
 
+	struct atmel_mci_dma	dma;
+	struct dma_chan		*data_chan;
+
 	u32			cmd_status;
 	u32			data_status;
-	u32			stop_status;
 	u32			stop_cmdr;
 
-	u32			mode_reg;
-	u32			sdc_reg;
-
 	struct tasklet_struct	tasklet;
 	unsigned long		pending_events;
 	unsigned long		completed_events;
+	enum atmel_mci_state	state;
+	struct list_head	queue;
 
-	int			present;
-	int			detect_pin;
-	int			wp_pin;
-
-	/* For detect pin debouncing */
-	struct timer_list	detect_timer;
-
+	bool			need_clock_update;
+	bool			need_reset;
+	u32			mode_reg;
 	unsigned long		bus_hz;
 	unsigned long		mapbase;
 	struct clk		*mck;
 	struct platform_device	*pdev;
+
+	struct atmel_mci_slot	*slot[ATMEL_MCI_MAX_NR_SLOTS];
+};
+
+/**
+ * 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.
+ * @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_timer: Timer used for debouncing @detect_pin interrupts.
+ */
+struct atmel_mci_slot {
+	struct mmc_host		*mmc;
+	struct atmel_mci	*host;
+
+	u32			sdc_reg;
+
+	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;
+
+	struct timer_list	detect_timer;
 };
 
-#define atmci_is_completed(host, event)				\
-	test_bit(event, &host->completed_events)
 #define atmci_test_and_clear_pending(host, event)		\
 	test_and_clear_bit(event, &host->pending_events)
-#define atmci_test_and_set_completed(host, event)		\
-	test_and_set_bit(event, &host->completed_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)
-#define atmci_clear_pending(host, event)			\
-	clear_bit(event, &host->pending_events)
 
 /*
  * The debugfs stuff below is mostly optimized away when
@@ -98,14 +214,15 @@ struct atmel_mci {
  */
 static int atmci_req_show(struct seq_file *s, void *v)
 {
-	struct atmel_mci	*host = s->private;
-	struct mmc_request	*mrq = host->mrq;
+	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_irq(&host->mmc->lock);
+	spin_lock_bh(&slot->host->lock);
+	mrq = slot->mrq;
 
 	if (mrq) {
 		cmd = mrq->cmd;
@@ -130,7 +247,7 @@ static int atmci_req_show(struct seq_file *s, void *v)
 				stop->resp[2], stop->error);
 	}
 
-	spin_unlock_irq(&host->mmc->lock);
+	spin_unlock_bh(&slot->host->lock);
 
 	return 0;
 }
@@ -193,12 +310,16 @@ static int atmci_regs_show(struct seq_file *s, void *v)
 	if (!buf)
 		return -ENOMEM;
 
-	/* Grab a more or less consistent snapshot */
-	spin_lock_irq(&host->mmc->lock);
+	/*
+	 * 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);
 	clk_enable(host->mck);
 	memcpy_fromio(buf, host->regs, MCI_REGS_SIZE);
 	clk_disable(host->mck);
-	spin_unlock_irq(&host->mmc->lock);
+	spin_unlock_bh(&host->lock);
 
 	seq_printf(s, "MR:\t0x%08x%s%s CLKDIV=%u\n",
 			buf[MCI_MR / 4],
@@ -236,13 +357,13 @@ static const struct file_operations atmci_regs_fops = {
 	.release	= single_release,
 };
 
-static void atmci_init_debugfs(struct atmel_mci *host)
+static void atmci_init_debugfs(struct atmel_mci_slot *slot)
 {
-	struct mmc_host	*mmc;
-	struct dentry	*root;
-	struct dentry	*node;
+	struct mmc_host		*mmc = slot->mmc;
+	struct atmel_mci	*host = slot->host;
+	struct dentry		*root;
+	struct dentry		*node;
 
-	mmc = host->mmc;
 	root = mmc->debugfs_root;
 	if (!root)
 		return;
@@ -254,7 +375,11 @@ static void atmci_init_debugfs(struct atmel_mci *host)
 	if (!node)
 		goto err;
 
-	node = debugfs_create_file("req", S_IRUSR, root, host, &atmci_req_fops);
+	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;
 
@@ -271,25 +396,7 @@ static void atmci_init_debugfs(struct atmel_mci *host)
 	return;
 
 err:
-	dev_err(&host->pdev->dev,
-		"failed to initialize debugfs for controller\n");
-}
-
-static void atmci_enable(struct atmel_mci *host)
-{
-	clk_enable(host->mck);
-	mci_writel(host, CR, MCI_CR_MCIEN);
-	mci_writel(host, MR, host->mode_reg);
-	mci_writel(host, SDCR, host->sdc_reg);
-}
-
-static void atmci_disable(struct atmel_mci *host)
-{
-	mci_writel(host, CR, MCI_CR_SWRST);
-
-	/* Stall until write is complete, then disable the bus clock */
-	mci_readl(host, SR);
-	clk_disable(host->mck);
+	dev_err(&mmc->class_dev, "failed to initialize debugfs for slot\n");
 }
 
 static inline unsigned int ns_to_clocks(struct atmel_mci *host,
@@ -299,7 +406,7 @@ static inline unsigned int ns_to_clocks(struct atmel_mci *host,
 }
 
 static void atmci_set_timeout(struct atmel_mci *host,
-			      struct mmc_data *data)
+		struct atmel_mci_slot *slot, struct mmc_data *data)
 {
 	static unsigned	dtomul_to_shift[] = {
 		0, 4, 7, 8, 10, 12, 16, 20
@@ -322,7 +429,7 @@ static void atmci_set_timeout(struct atmel_mci *host,
 		dtocyc = 15;
 	}
 
-	dev_vdbg(&host->mmc->class_dev, "setting timeout to %u cycles\n",
+	dev_vdbg(&slot->mmc->class_dev, "setting timeout to %u cycles\n",
 			dtocyc << dtomul_to_shift[dtomul]);
 	mci_writel(host, DTOR, (MCI_DTOMUL(dtomul) | MCI_DTOCYC(dtocyc)));
 }
@@ -375,15 +482,12 @@ static u32 atmci_prepare_command(struct mmc_host *mmc,
 }
 
 static void atmci_start_command(struct atmel_mci *host,
-				struct mmc_command *cmd,
-				u32 cmd_flags)
+		struct mmc_command *cmd, u32 cmd_flags)
 {
-	/* Must read host->cmd after testing event flags */
-	smp_rmb();
 	WARN_ON(host->cmd);
 	host->cmd = cmd;
 
-	dev_vdbg(&host->mmc->class_dev,
+	dev_vdbg(&host->pdev->dev,
 			"start command: ARGR=0x%08x CMDR=0x%08x\n",
 			cmd->arg, cmd_flags);
 
@@ -391,34 +495,157 @@ static void atmci_start_command(struct atmel_mci *host,
 	mci_writel(host, CMDR, cmd_flags);
 }
 
-static void send_stop_cmd(struct mmc_host *mmc, struct mmc_data *data)
+static void send_stop_cmd(struct atmel_mci *host, struct mmc_data *data)
 {
-	struct atmel_mci *host = mmc_priv(mmc);
-
 	atmci_start_command(host, data->stop, host->stop_cmdr);
 	mci_writel(host, IER, MCI_CMDRDY);
 }
 
-static void atmci_request_end(struct mmc_host *mmc, struct mmc_request *mrq)
+#ifdef CONFIG_MMC_ATMELMCI_DMA
+static void atmci_dma_cleanup(struct atmel_mci *host)
 {
-	struct atmel_mci *host = mmc_priv(mmc);
+	struct mmc_data			*data = host->data;
 
-	WARN_ON(host->cmd || host->data);
-	host->mrq = NULL;
+	dma_unmap_sg(&host->pdev->dev, data->sg, data->sg_len,
+		     ((data->flags & MMC_DATA_WRITE)
+		      ? DMA_TO_DEVICE : DMA_FROM_DEVICE));
+}
+
+static void atmci_stop_dma(struct atmel_mci *host)
+{
+	struct dma_chan *chan = host->data_chan;
+
+	if (chan) {
+		chan->device->device_terminate_all(chan);
+		atmci_dma_cleanup(host);
+	} else {
+		/* Data transfer was stopped by the interrupt handler */
+		atmci_set_pending(host, EVENT_XFER_COMPLETE);
+		mci_writel(host, IER, MCI_NOTBUSY);
+	}
+}
+
+/* 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");
+
+	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) {
+		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.
+		 */
+		mci_writel(host, IER, MCI_NOTBUSY);
+	}
+}
+
+static int
+atmci_submit_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;
+
+	/*
+	 * 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 -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;
+	}
+
+	/* If we don't have a channel, we can't do DMA */
+	chan = host->dma.chan;
+	if (chan) {
+		dma_chan_get(chan);
+		host->data_chan = chan;
+	}
+
+	if (!chan)
+		return -ENODEV;
+
+	if (data->flags & MMC_DATA_READ)
+		direction = DMA_FROM_DEVICE;
+	else
+		direction = DMA_TO_DEVICE;
+
+	desc = chan->device->device_prep_slave_sg(chan,
+			data->sg, data->sg_len, direction,
+			DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc)
+		return -ENOMEM;
 
-	atmci_disable(host);
+	host->dma.data_desc = desc;
+	desc->callback = atmci_dma_complete;
+	desc->callback_param = host;
+	desc->tx_submit(desc);
 
-	mmc_request_done(mmc, mrq);
+	/* Go! */
+	chan->device->device_issue_pending(chan);
+
+	return 0;
+}
+
+#else /* CONFIG_MMC_ATMELMCI_DMA */
+
+static int atmci_submit_data_dma(struct atmel_mci *host, struct mmc_data *data)
+{
+	return -ENOSYS;
 }
 
+static void atmci_stop_dma(struct atmel_mci *host)
+{
+	/* Data transfer was stopped by the interrupt handler */
+	atmci_set_pending(host, EVENT_XFER_COMPLETE);
+	mci_writel(host, IER, MCI_NOTBUSY);
+}
+
+#endif /* CONFIG_MMC_ATMELMCI_DMA */
+
 /*
  * Returns a mask of interrupt flags to be enabled after the whole
  * request has been prepared.
  */
-static u32 atmci_submit_data(struct mmc_host *mmc, struct mmc_data *data)
+static u32 atmci_submit_data(struct atmel_mci *host, struct mmc_data *data)
 {
-	struct atmel_mci	*host = mmc_priv(mmc);
-	u32			iflags;
+	u32 iflags;
 
 	data->error = -EINPROGRESS;
 
@@ -426,77 +653,89 @@ static u32 atmci_submit_data(struct mmc_host *mmc, struct mmc_data *data)
 	host->sg = NULL;
 	host->data = data;
 
-	dev_vdbg(&mmc->class_dev, "BLKR=0x%08x\n",
-			MCI_BCNT(data->blocks) | MCI_BLKLEN(data->blksz));
-
 	iflags = ATMCI_DATA_ERROR_FLAGS;
-	host->sg = data->sg;
-	host->pio_offset = 0;
-	if (data->flags & MMC_DATA_READ)
-		iflags |= MCI_RXRDY;
-	else
-		iflags |= MCI_TXRDY;
+	if (atmci_submit_data_dma(host, data)) {
+		host->data_chan = NULL;
+
+		/*
+		 * 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->sg = data->sg;
+		host->pio_offset = 0;
+		if (data->flags & MMC_DATA_READ)
+			iflags |= MCI_RXRDY;
+		else
+			iflags |= MCI_TXRDY;
+	}
 
 	return iflags;
 }
 
-static void atmci_request(struct mmc_host *mmc, struct mmc_request *mrq)
+static void atmci_start_request(struct atmel_mci *host,
+		struct atmel_mci_slot *slot)
 {
-	struct atmel_mci	*host = mmc_priv(mmc);
-	struct mmc_data		*data;
+	struct mmc_request	*mrq;
 	struct mmc_command	*cmd;
+	struct mmc_data		*data;
 	u32			iflags;
-	u32			cmdflags = 0;
-
-	iflags = mci_readl(host, IMR);
-	if (iflags)
-		dev_warn(&mmc->class_dev, "WARNING: IMR=0x%08x\n",
-				mci_readl(host, IMR));
-
-	WARN_ON(host->mrq != NULL);
-
-	/*
-	 * 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 (!host->present) {
-		mrq->cmd->error = -ENOMEDIUM;
-		mmc_request_done(mmc, mrq);
-		return;
-	}
+	u32			cmdflags;
 
+	mrq = slot->mrq;
+	host->cur_slot = slot;
 	host->mrq = mrq;
+
 	host->pending_events = 0;
 	host->completed_events = 0;
+	host->data_status = 0;
 
-	atmci_enable(host);
+	if (host->need_reset) {
+		mci_writel(host, CR, MCI_CR_SWRST);
+		mci_writel(host, CR, MCI_CR_MCIEN);
+		mci_writel(host, MR, host->mode_reg);
+		host->need_reset = false;
+	}
+	mci_writel(host, SDCR, slot->sdc_reg);
 
-	/* We don't support multiple blocks of weird lengths. */
+	iflags = mci_readl(host, IMR);
+	if (iflags)
+		dev_warn(&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) */
+		mci_writel(host, CMDR, MCI_CMDR_SPCMD_INIT);
+		while (!(mci_readl(host, SR) & MCI_CMDRDY))
+			cpu_relax();
+	}
 	data = mrq->data;
 	if (data) {
-		if (data->blocks > 1 && data->blksz & 3)
-			goto fail;
-		atmci_set_timeout(host, data);
+		atmci_set_timeout(host, slot, data);
 
 		/* Must set block count/size before sending command */
 		mci_writel(host, BLKR, MCI_BCNT(data->blocks)
 				| MCI_BLKLEN(data->blksz));
+		dev_vdbg(&slot->mmc->class_dev, "BLKR=0x%08x\n",
+			MCI_BCNT(data->blocks) | MCI_BLKLEN(data->blksz));
 	}
 
 	iflags = MCI_CMDRDY;
 	cmd = mrq->cmd;
-	cmdflags = atmci_prepare_command(mmc, cmd);
+	cmdflags = atmci_prepare_command(slot->mmc, cmd);
 	atmci_start_command(host, cmd, cmdflags);
 
 	if (data)
-		iflags |= atmci_submit_data(mmc, data);
+		iflags |= atmci_submit_data(host, data);
 
 	if (mrq->stop) {
-		host->stop_cmdr = atmci_prepare_command(mmc, mrq->stop);
+		host->stop_cmdr = atmci_prepare_command(slot->mmc, mrq->stop);
 		host->stop_cmdr |= MCI_CMDR_STOP_XFER;
 		if (!(data->flags & MMC_DATA_WRITE))
 			host->stop_cmdr |= MCI_CMDR_TRDIR_READ;
@@ -513,59 +752,156 @@ static void atmci_request(struct mmc_host *mmc, struct mmc_request *mrq)
 	 * prepared yet.)
 	 */
 	mci_writel(host, IER, iflags);
+}
 
-	return;
+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 {
+		list_add_tail(&slot->queue_node, &host->queue);
+	}
+	spin_unlock_bh(&host->lock);
+}
 
-fail:
-	atmci_disable(host);
-	host->mrq = NULL;
-	mrq->cmd->error = -EINVAL;
-	mmc_request_done(mmc, mrq);
+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);
+
+	/*
+	 * We may "know" the card is gone even though there's still an
+	 * electrical