Merge branch 'next' into for-linus

Prepare first round of input updates for 4.3 merge window.

41 files changed, 4617 insertions, 2010 deletions

diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig
index ab8dfbef6f1b..0cae1694014d 100644
--- a/drivers/spi/Kconfig
+++ b/drivers/spi/Kconfig
@@ -77,7 +77,9 @@ config SPI_ATMEL
 
 config SPI_BCM2835
 	tristate "BCM2835 SPI controller"
+	depends on GPIOLIB
 	depends on ARCH_BCM2835 || COMPILE_TEST
+	depends on GPIOLIB
 	help
 	  This selects a driver for the Broadcom BCM2835 SPI master.
 
@@ -159,10 +161,9 @@ config SPI_BUTTERFLY
 
 config SPI_CADENCE
 	tristate "Cadence SPI controller"
-	depends on ARM
 	help
 	  This selects the Cadence SPI controller master driver
-	  used by Xilinx Zynq.
+	  used by Xilinx Zynq and ZynqMP.
 
 config SPI_CLPS711X
 	tristate "CLPS711X host SPI controller"
@@ -221,7 +222,7 @@ config SPI_FALCON
 
 config SPI_GPIO
 	tristate "GPIO-based bitbanging SPI Master"
-	depends on GPIOLIB
+	depends on GPIOLIB || COMPILE_TEST
 	select SPI_BITBANG
 	help
 	  This simple GPIO bitbanging SPI master uses the arch-neutral GPIO
@@ -303,7 +304,7 @@ config SPI_FSL_SPI
 config SPI_FSL_DSPI
 	tristate "Freescale DSPI controller"
 	select REGMAP_MMIO
-	depends on SOC_VF610 || COMPILE_TEST
+	depends on SOC_VF610 || SOC_LS1021A || COMPILE_TEST
 	help
 	  This enables support for the Freescale DSPI controller in master
 	  mode. VF610 platform uses the controller.
@@ -327,7 +328,7 @@ config SPI_MESON_SPIFC
 
 config SPI_OC_TINY
 	tristate "OpenCores tiny SPI"
-	depends on GPIOLIB
+	depends on GPIOLIB || COMPILE_TEST
 	select SPI_BITBANG
 	help
 	  This is the driver for OpenCores tiny SPI master controller.
@@ -394,16 +395,9 @@ config SPI_PPC4xx
 	help
 	  This selects a driver for the PPC4xx SPI Controller.
 
-config SPI_PXA2XX_PXADMA
-	bool "PXA2xx SSP legacy PXA DMA API support"
-	depends on SPI_PXA2XX && ARCH_PXA
-	help
-	  Enable PXA private legacy DMA API support. Note that this is
-	  deprecated in favor of generic DMA engine API.
-
 config SPI_PXA2XX_DMA
 	def_bool y
-	depends on SPI_PXA2XX && !SPI_PXA2XX_PXADMA
+	depends on SPI_PXA2XX
 
 config SPI_PXA2XX
 	tristate "PXA2xx SSP SPI master"
@@ -429,6 +423,12 @@ config SPI_ROCKCHIP
 	  The main usecase of this controller is to use spi flash as boot
 	  device.
 
+config SPI_RB4XX
+	tristate "Mikrotik RB4XX SPI master"
+	depends on SPI_MASTER && ATH79
+	help
+	  SPI controller driver for the Mikrotik RB4xx series boards.
+
 config SPI_RSPI
 	tristate "Renesas RSPI/QSPI controller"
 	depends on SUPERH || ARCH_SHMOBILE || COMPILE_TEST
@@ -610,6 +610,12 @@ config SPI_XTENSA_XTFPGA
 	  16 bit words in SPI mode 0, automatically asserting CS on transfer
 	  start and deasserting on end.
 
+config SPI_ZYNQMP_GQSPI
+	tristate "Xilinx ZynqMP GQSPI controller"
+	depends on SPI_MASTER
+	help
+	  Enables Xilinx GQSPI controller driver for Zynq UltraScale+ MPSoC.
+
 config SPI_NUC900
 	tristate "Nuvoton NUC900 series SPI"
 	depends on ARCH_W90X900
@@ -632,7 +638,7 @@ config SPI_DW_PCI
 
 config SPI_DW_MID_DMA
 	bool "DMA support for DW SPI controller on Intel MID platform"
-	depends on SPI_DW_PCI && INTEL_MID_DMAC
+	depends on SPI_DW_PCI && DW_DMAC_PCI
 
 config SPI_DW_MMIO
 	tristate "Memory-mapped io interface driver for DW SPI core"
diff --git a/drivers/spi/Makefile b/drivers/spi/Makefile
index d8cbf654976b..1154dbac8f2c 100644
--- a/drivers/spi/Makefile
+++ b/drivers/spi/Makefile
@@ -60,12 +60,12 @@ obj-$(CONFIG_SPI_ORION)			+= spi-orion.o
 obj-$(CONFIG_SPI_PL022)			+= spi-pl022.o
 obj-$(CONFIG_SPI_PPC4xx)		+= spi-ppc4xx.o
 spi-pxa2xx-platform-objs		:= spi-pxa2xx.o
-spi-pxa2xx-platform-$(CONFIG_SPI_PXA2XX_PXADMA)	+= spi-pxa2xx-pxadma.o
 spi-pxa2xx-platform-$(CONFIG_SPI_PXA2XX_DMA)	+= spi-pxa2xx-dma.o
 obj-$(CONFIG_SPI_PXA2XX)		+= spi-pxa2xx-platform.o
 obj-$(CONFIG_SPI_PXA2XX_PCI)		+= spi-pxa2xx-pci.o
 obj-$(CONFIG_SPI_QUP)			+= spi-qup.o
 obj-$(CONFIG_SPI_ROCKCHIP)		+= spi-rockchip.o
+obj-$(CONFIG_SPI_RB4XX)			+= spi-rb4xx.o
 obj-$(CONFIG_SPI_RSPI)			+= spi-rspi.o
 obj-$(CONFIG_SPI_S3C24XX)		+= spi-s3c24xx-hw.o
 spi-s3c24xx-hw-y			:= spi-s3c24xx.o
@@ -89,3 +89,4 @@ obj-$(CONFIG_SPI_TXX9)			+= spi-txx9.o
 obj-$(CONFIG_SPI_XCOMM)		+= spi-xcomm.o
 obj-$(CONFIG_SPI_XILINX)		+= spi-xilinx.o
 obj-$(CONFIG_SPI_XTENSA_XTFPGA)		+= spi-xtensa-xtfpga.o
+obj-$(CONFIG_SPI_ZYNQMP_GQSPI)		+= spi-zynqmp-gqspi.o
diff --git a/drivers/spi/spi-ath79.c b/drivers/spi/spi-ath79.c
index b02eb4ac0218..bf1f9b32c597 100644
--- a/drivers/spi/spi-ath79.c
+++ b/drivers/spi/spi-ath79.c
@@ -79,10 +79,8 @@ static void ath79_spi_chipselect(struct spi_device *spi, int is_active)
 	}
 
 	if (spi->chip_select) {
-		struct ath79_spi_controller_data *cdata = spi->controller_data;
-
 		/* SPI is normally active-low */
-		gpio_set_value(cdata->gpio, cs_high);
+		gpio_set_value(spi->cs_gpio, cs_high);
 	} else {
 		if (cs_high)
 			sp->ioc_base |= AR71XX_SPI_IOC_CS0;
@@ -117,11 +115,10 @@ static void ath79_spi_disable(struct ath79_spi *sp)
 
 static int ath79_spi_setup_cs(struct spi_device *spi)
 {
-	struct ath79_spi_controller_data *cdata;
+	struct ath79_spi *sp = ath79_spidev_to_sp(spi);
 	int status;
 
-	cdata = spi->controller_data;
-	if (spi->chip_select && !cdata)
+	if (spi->chip_select && !gpio_is_valid(spi->cs_gpio))
 		return -EINVAL;
 
 	status = 0;
@@ -134,8 +131,15 @@ static int ath79_spi_setup_cs(struct spi_device *spi)
 		else
 			flags |= GPIOF_INIT_HIGH;
 
-		status = gpio_request_one(cdata->gpio, flags,
+		status = gpio_request_one(spi->cs_gpio, flags,
 					  dev_name(&spi->dev));
+	} else {
+		if (spi->mode & SPI_CS_HIGH)
+			sp->ioc_base &= ~AR71XX_SPI_IOC_CS0;
+		else
+			sp->ioc_base |= AR71XX_SPI_IOC_CS0;
+
+		ath79_spi_wr(sp, AR71XX_SPI_REG_IOC, sp->ioc_base);
 	}
 
 	return status;
@@ -144,8 +148,7 @@ static int ath79_spi_setup_cs(struct spi_device *spi)
 static void ath79_spi_cleanup_cs(struct spi_device *spi)
 {
 	if (spi->chip_select) {
-		struct ath79_spi_controller_data *cdata = spi->controller_data;
-		gpio_free(cdata->gpio);
+		gpio_free(spi->cs_gpio);
 	}
 }
 
@@ -217,6 +220,7 @@ static int ath79_spi_probe(struct platform_device *pdev)
 	}
 
 	sp = spi_master_get_devdata(master);
+	master->dev.of_node = pdev->dev.of_node;
 	platform_set_drvdata(pdev, sp);
 
 	pdata = dev_get_platdata(&pdev->dev);
@@ -253,7 +257,7 @@ static int ath79_spi_probe(struct platform_device *pdev)
 		goto err_put_master;
 	}
 
-	ret = clk_enable(sp->clk);
+	ret = clk_prepare_enable(sp->clk);
 	if (ret)
 		goto err_put_master;
 
@@ -277,7 +281,7 @@ static int ath79_spi_probe(struct platform_device *pdev)
 err_disable:
 	ath79_spi_disable(sp);
 err_clk_disable:
-	clk_disable(sp->clk);
+	clk_disable_unprepare(sp->clk);
 err_put_master:
 	spi_master_put(sp->bitbang.master);
 
@@ -290,7 +294,7 @@ static int ath79_spi_remove(struct platform_device *pdev)
 
 	spi_bitbang_stop(&sp->bitbang);
 	ath79_spi_disable(sp);
-	clk_disable(sp->clk);
+	clk_disable_unprepare(sp->clk);
 	spi_master_put(sp->bitbang.master);
 
 	return 0;
@@ -301,12 +305,18 @@ static void ath79_spi_shutdown(struct platform_device *pdev)
 	ath79_spi_remove(pdev);
 }
 
+static const struct of_device_id ath79_spi_of_match[] = {
+	{ .compatible = "qca,ar7100-spi", },
+	{ },
+};
+
 static struct platform_driver ath79_spi_driver = {
 	.probe		= ath79_spi_probe,
 	.remove		= ath79_spi_remove,
 	.shutdown	= ath79_spi_shutdown,
 	.driver		= {
 		.name	= DRV_NAME,
+		.of_match_table = ath79_spi_of_match,
 	},
 };
 module_platform_driver(ath79_spi_driver);
diff --git a/drivers/spi/spi-atmel.c b/drivers/spi/spi-atmel.c
index 06de34001c66..c9eca347787d 100644
--- a/drivers/spi/spi-atmel.c
+++ b/drivers/spi/spi-atmel.c
@@ -41,6 +41,8 @@
 #define SPI_CSR1				0x0034
 #define SPI_CSR2				0x0038
 #define SPI_CSR3				0x003c
+#define SPI_FMR					0x0040
+#define SPI_FLR					0x0044
 #define SPI_VERSION				0x00fc
 #define SPI_RPR					0x0100
 #define SPI_RCR					0x0104
@@ -62,6 +64,14 @@
 #define SPI_SWRST_SIZE				1
 #define SPI_LASTXFER_OFFSET			24
 #define SPI_LASTXFER_SIZE			1
+#define SPI_TXFCLR_OFFSET			16
+#define SPI_TXFCLR_SIZE				1
+#define SPI_RXFCLR_OFFSET			17
+#define SPI_RXFCLR_SIZE				1
+#define SPI_FIFOEN_OFFSET			30
+#define SPI_FIFOEN_SIZE				1
+#define SPI_FIFODIS_OFFSET			31
+#define SPI_FIFODIS_SIZE			1
 
 /* Bitfields in MR */
 #define SPI_MSTR_OFFSET				0
@@ -114,6 +124,22 @@
 #define SPI_TXEMPTY_SIZE			1
 #define SPI_SPIENS_OFFSET			16
 #define SPI_SPIENS_SIZE				1
+#define SPI_TXFEF_OFFSET			24
+#define SPI_TXFEF_SIZE				1
+#define SPI_TXFFF_OFFSET			25
+#define SPI_TXFFF_SIZE				1
+#define SPI_TXFTHF_OFFSET			26
+#define SPI_TXFTHF_SIZE				1
+#define SPI_RXFEF_OFFSET			27
+#define SPI_RXFEF_SIZE				1
+#define SPI_RXFFF_OFFSET			28
+#define SPI_RXFFF_SIZE				1
+#define SPI_RXFTHF_OFFSET			29
+#define SPI_RXFTHF_SIZE				1
+#define SPI_TXFPTEF_OFFSET			30
+#define SPI_TXFPTEF_SIZE			1
+#define SPI_RXFPTEF_OFFSET			31
+#define SPI_RXFPTEF_SIZE			1
 
 /* Bitfields in CSR0 */
 #define SPI_CPOL_OFFSET				0
@@ -157,6 +183,22 @@
 #define SPI_TXTDIS_OFFSET			9
 #define SPI_TXTDIS_SIZE				1
 
+/* Bitfields in FMR */
+#define SPI_TXRDYM_OFFSET			0
+#define SPI_TXRDYM_SIZE				2
+#define SPI_RXRDYM_OFFSET			4
+#define SPI_RXRDYM_SIZE				2
+#define SPI_TXFTHRES_OFFSET			16
+#define SPI_TXFTHRES_SIZE			6
+#define SPI_RXFTHRES_OFFSET			24
+#define SPI_RXFTHRES_SIZE			6
+
+/* Bitfields in FLR */
+#define SPI_TXFL_OFFSET				0
+#define SPI_TXFL_SIZE				6
+#define SPI_RXFL_OFFSET				16
+#define SPI_RXFL_SIZE				6
+
 /* Constants for BITS */
 #define SPI_BITS_8_BPT				0
 #define SPI_BITS_9_BPT				1
@@ -167,6 +209,9 @@
 #define SPI_BITS_14_BPT				6
 #define SPI_BITS_15_BPT				7
 #define SPI_BITS_16_BPT				8
+#define SPI_ONE_DATA				0
+#define SPI_TWO_DATA				1
+#define SPI_FOUR_DATA				2
 
 /* Bit manipulation macros */
 #define SPI_BIT(name) \
@@ -180,11 +225,37 @@
 	  | SPI_BF(name, value))
 
 /* Register access macros */
+#ifdef CONFIG_AVR32
 #define spi_readl(port, reg) \
 	__raw_readl((port)->regs + SPI_##reg)
 #define spi_writel(port, reg, value) \
 	__raw_writel((value), (port)->regs + SPI_##reg)
 
+#define spi_readw(port, reg) \
+	__raw_readw((port)->regs + SPI_##reg)
+#define spi_writew(port, reg, value) \
+	__raw_writew((value), (port)->regs + SPI_##reg)
+
+#define spi_readb(port, reg) \
+	__raw_readb((port)->regs + SPI_##reg)
+#define spi_writeb(port, reg, value) \
+	__raw_writeb((value), (port)->regs + SPI_##reg)
+#else
+#define spi_readl(port, reg) \
+	readl_relaxed((port)->regs + SPI_##reg)
+#define spi_writel(port, reg, value) \
+	writel_relaxed((value), (port)->regs + SPI_##reg)
+
+#define spi_readw(port, reg) \
+	readw_relaxed((port)->regs + SPI_##reg)
+#define spi_writew(port, reg, value) \
+	writew_relaxed((value), (port)->regs + SPI_##reg)
+
+#define spi_readb(port, reg) \
+	readb_relaxed((port)->regs + SPI_##reg)
+#define spi_writeb(port, reg, value) \
+	writeb_relaxed((value), (port)->regs + SPI_##reg)
+#endif
 /* use PIO for small transfers, avoiding DMA setup/teardown overhead and
  * cache operations; better heuristics consider wordsize and bitrate.
  */
@@ -240,11 +311,14 @@ struct atmel_spi {
 
 	bool			use_dma;
 	bool			use_pdc;
+	bool			use_cs_gpios;
 	/* dmaengine data */
 	struct atmel_spi_dma	dma;
 
 	bool			keep_cs;
 	bool			cs_active;
+
+	u32			fifo_size;
 };
 
 /* Controller-specific per-slave state */
@@ -315,7 +389,8 @@ static void cs_activate(struct atmel_spi *as, struct spi_device *spi)
 		}
 
 		mr = spi_readl(as, MR);
-		gpio_set_value(asd->npcs_pin, active);
+		if (as->use_cs_gpios)
+			gpio_set_value(asd->npcs_pin, active);
 	} else {
 		u32 cpol = (spi->mode & SPI_CPOL) ? SPI_BIT(CPOL) : 0;
 		int i;
@@ -331,7 +406,7 @@ static void cs_activate(struct atmel_spi *as, struct spi_device *spi)
 
 		mr = spi_readl(as, MR);
 		mr = SPI_BFINS(PCS, ~(1 << spi->chip_select), mr);
-		if (spi->chip_select != 0)
+		if (as->use_cs_gpios && spi->chip_select != 0)
 			gpio_set_value(asd->npcs_pin, active);
 		spi_writel(as, MR, mr);
 	}
@@ -360,7 +435,9 @@ static void cs_deactivate(struct atmel_spi *as, struct spi_device *spi)
 			asd->npcs_pin, active ? " (low)" : "",
 			mr);
 
-	if (atmel_spi_is_v2(as) || spi->chip_select != 0)
+	if (!as->use_cs_gpios)
+		spi_writel(as, CR, SPI_BIT(LASTXFER));
+	else if (atmel_spi_is_v2(as) || spi->chip_select != 0)
 		gpio_set_value(asd->npcs_pin, !active);
 }
 
@@ -400,6 +477,20 @@ static int atmel_spi_dma_slave_config(struct atmel_spi *as,
 	slave_config->dst_maxburst = 1;
 	slave_config->device_fc = false;
 
+	/*
+	 * This driver uses fixed peripheral select mode (PS bit set to '0' in
+	 * the Mode Register).
+	 * So according to the datasheet, when FIFOs are available (and
+	 * enabled), the Transmit FIFO operates in Multiple Data Mode.
+	 * In this mode, up to 2 data, not 4, can be written into the Transmit
+	 * Data Register in a single access.
+	 * However, the first data has to be written into the lowest 16 bits and
+	 * the second data into the highest 16 bits of the Transmit
+	 * Data Register. For 8bit data (the most frequent case), it would
+	 * require to rework tx_buf so each data would actualy fit 16 bits.
+	 * So we'd rather write only one data at the time. Hence the transmit
+	 * path works the same whether FIFOs are available (and enabled) or not.
+	 */
 	slave_config->direction = DMA_MEM_TO_DEV;
 	if (dmaengine_slave_config(as->dma.chan_tx, slave_config)) {
 		dev_err(&as->pdev->dev,
@@ -407,6 +498,14 @@ static int atmel_spi_dma_slave_config(struct atmel_spi *as,
 		err = -EINVAL;
 	}
 
+	/*
+	 * This driver configures the spi controller for master mode (MSTR bit
+	 * set to '1' in the Mode Register).
+	 * So according to the datasheet, when FIFOs are available (and
+	 * enabled), the Receive FIFO operates in Single Data Mode.
+	 * So the receive path works the same whether FIFOs are available (and
+	 * enabled) or not.
+	 */
 	slave_config->direction = DMA_DEV_TO_MEM;
 	if (dmaengine_slave_config(as->dma.chan_rx, slave_config)) {
 		dev_err(&as->pdev->dev,
@@ -496,10 +595,10 @@ static void dma_callback(void *data)
 }
 
 /*
- * Next transfer using PIO.
+ * Next transfer using PIO without FIFO.
  */
-static void atmel_spi_next_xfer_pio(struct spi_master *master,
-				struct spi_transfer *xfer)
+static void atmel_spi_next_xfer_single(struct spi_master *master,
+				       struct spi_transfer *xfer)
 {
 	struct atmel_spi	*as = spi_master_get_devdata(master);
 	unsigned long xfer_pos = xfer->len - as->current_remaining_bytes;
@@ -532,6 +631,99 @@ static void atmel_spi_next_xfer_pio(struct spi_master *master,
 }
 
 /*
+ * Next transfer using PIO with FIFO.
+ */
+static void atmel_spi_next_xfer_fifo(struct spi_master *master,
+				     struct spi_transfer *xfer)
+{
+	struct atmel_spi *as = spi_master_get_devdata(master);
+	u32 current_remaining_data, num_data;
+	u32 offset = xfer->len - as->current_remaining_bytes;
+	const u16 *words = (const u16 *)((u8 *)xfer->tx_buf + offset);
+	const u8  *bytes = (const u8  *)((u8 *)xfer->tx_buf + offset);
+	u16 td0, td1;
+	u32 fifomr;
+
+	dev_vdbg(master->dev.parent, "atmel_spi_next_xfer_fifo\n");
+
+	/* Compute the number of data to transfer in the current iteration */
+	current_remaining_data = ((xfer->bits_per_word > 8) ?
+				  ((u32)as->current_remaining_bytes >> 1) :
+				  (u32)as->current_remaining_bytes);
+	num_data = min(current_remaining_data, as->fifo_size);
+
+	/* Flush RX and TX FIFOs */
+	spi_writel(as, CR, SPI_BIT(RXFCLR) | SPI_BIT(TXFCLR));
+	while (spi_readl(as, FLR))
+		cpu_relax();
+
+	/* Set RX FIFO Threshold to the number of data to transfer */
+	fifomr = spi_readl(as, FMR);
+	spi_writel(as, FMR, SPI_BFINS(RXFTHRES, num_data, fifomr));
+
+	/* Clear FIFO flags in the Status Register, especially RXFTHF */
+	(void)spi_readl(as, SR);
+
+	/* Fill TX FIFO */
+	while (num_data >= 2) {
+		if (xfer->tx_buf) {
+			if (xfer->bits_per_word > 8) {
+				td0 = *words++;
+				td1 = *words++;
+			} else {
+				td0 = *bytes++;
+				td1 = *bytes++;
+			}
+		} else {
+			td0 = 0;
+			td1 = 0;
+		}
+
+		spi_writel(as, TDR, (td1 << 16) | td0);
+		num_data -= 2;
+	}
+
+	if (num_data) {
+		if (xfer->tx_buf) {
+			if (xfer->bits_per_word > 8)
+				td0 = *words++;
+			else
+				td0 = *bytes++;
+		} else {
+			td0 = 0;
+		}
+
+		spi_writew(as, TDR, td0);
+		num_data--;
+	}
+
+	dev_dbg(master->dev.parent,
+		"  start fifo xfer %p: len %u tx %p rx %p bitpw %d\n",
+		xfer, xfer->len, xfer->tx_buf, xfer->rx_buf,
+		xfer->bits_per_word);
+
+	/*
+	 * Enable RX FIFO Threshold Flag interrupt to be notified about
+	 * transfer completion.
+	 */
+	spi_writel(as, IER, SPI_BIT(RXFTHF) | SPI_BIT(OVRES));
+}
+
+/*
+ * Next transfer using PIO.
+ */
+static void atmel_spi_next_xfer_pio(struct spi_master *master,
+				    struct spi_transfer *xfer)
+{
+	struct atmel_spi *as = spi_master_get_devdata(master);
+
+	if (as->fifo_size)
+		atmel_spi_next_xfer_fifo(master, xfer);
+	else
+		atmel_spi_next_xfer_single(master, xfer);
+}
+
+/*
  * Submit next transfer for DMA.
  */
 static int atmel_spi_next_xfer_dma_submit(struct spi_master *master,
@@ -833,13 +1025,8 @@ static void atmel_spi_disable_pdc_transfer(struct atmel_spi *as)
 	spi_writel(as, PTCR, SPI_BIT(RXTDIS) | SPI_BIT(TXTDIS));
 }
 
-/* Called from IRQ
- *
- * Must update "current_remaining_bytes" to keep track of data
- * to transfer.
- */
 static void
-atmel_spi_pump_pio_data(struct atmel_spi *as, struct spi_transfer *xfer)
+atmel_spi_pump_single_data(struct atmel_spi *as, struct spi_transfer *xfer)
 {
 	u8		*rxp;
 	u16		*rxp16;
@@ -866,6 +1053,57 @@ atmel_spi_pump_pio_data(struct atmel_spi *as, struct spi_transfer *xfer)
 	}
 }
 
+static void
+atmel_spi_pump_fifo_data(struct atmel_spi *as, struct spi_transfer *xfer)
+{
+	u32 fifolr = spi_readl(as, FLR);
+	u32 num_bytes, num_data = SPI_BFEXT(RXFL, fifolr);
+	u32 offset = xfer->len - as->current_remaining_bytes;
+	u16 *words = (u16 *)((u8 *)xfer->rx_buf + offset);
+	u8  *bytes = (u8  *)((u8 *)xfer->rx_buf + offset);
+	u16 rd; /* RD field is the lowest 16 bits of RDR */
+
+	/* Update the number of remaining bytes to transfer */
+	num_bytes = ((xfer->bits_per_word > 8) ?
+		     (num_data << 1) :
+		     num_data);
+
+	if (as->current_remaining_bytes > num_bytes)
+		as->current_remaining_bytes -= num_bytes;
+	else
+		as->current_remaining_bytes = 0;
+
+	/* Handle odd number of bytes when data are more than 8bit width */
+	if (xfer->bits_per_word > 8)
+		as->current_remaining_bytes &= ~0x1;
+
+	/* Read data */
+	while (num_data) {
+		rd = spi_readl(as, RDR);
+		if (xfer->rx_buf) {
+			if (xfer->bits_per_word > 8)
+				*words++ = rd;
+			else
+				*bytes++ = rd;
+		}
+		num_data--;
+	}
+}
+
+/* Called from IRQ
+ *
+ * Must update "current_remaining_bytes" to keep track of data
+ * to transfer.
+ */
+static void
+atmel_spi_pump_pio_data(struct atmel_spi *as, struct spi_transfer *xfer)
+{
+	if (as->fifo_size)
+		atmel_spi_pump_fifo_data(as, xfer);
+	else
+		atmel_spi_pump_single_data(as, xfer);
+}
+
 /* Interrupt
  *
  * No need for locking in this Interrupt handler: done_status is the
@@ -906,7 +1144,7 @@ atmel_spi_pio_interrupt(int irq, void *dev_id)
 
 		complete(&as->xfer_completion);
 
-	} else if (pending & SPI_BIT(RDRF)) {
+	} else if (pending & (SPI_BIT(RDRF) | SPI_BIT(RXFTHF))) {
 		atmel_spi_lock(as);
 
 		if (as->current_remaining_bytes) {
@@ -990,6 +1228,8 @@ static int atmel_spi_setup(struct spi_device *spi)
 		csr |= SPI_BIT(CPOL);
 	if (!(spi->mode & SPI_CPHA))
 		csr |= SPI_BIT(NCPHA);
+	if (!as->use_cs_gpios)
+		csr |= SPI_BIT(CSAAT);
 
 	/* DLYBS is mostly irrelevant since we manage chipselect using GPIOs.
 	 *
@@ -1003,7 +1243,9 @@ static int atmel_spi_setup(struct spi_device *spi)
 	/* chipselect must have been muxed as GPIO (e.g. in board setup) */
 	npcs_pin = (unsigned long)spi->controller_data;
 
-	if (gpio_is_valid(spi->cs_gpio))
+	if (!as->use_cs_gpios)
+		npcs_pin = spi->chip_select;
+	else if (gpio_is_valid(spi->cs_gpio))
 		npcs_pin = spi->cs_gpio;
 
 	asd = spi->controller_state;
@@ -1012,15 +1254,19 @@ static int atmel_spi_setup(struct spi_device *spi)
 		if (!asd)
 			return -ENOMEM;
 
-		ret = gpio_request(npcs_pin, dev_name(&spi->dev));
-		if (ret) {
-			kfree(asd);
-			return ret;
+		if (as->use_cs_gpios) {
+			ret = gpio_request(npcs_pin, dev_name(&spi->dev));
+			if (ret) {
+				kfree(asd);
+				return ret;
+			}
+
+			gpio_direction_output(npcs_pin,
+					      !(spi->mode & SPI_CS_HIGH));
 		}
 
 		asd->npcs_pin = npcs_pin;
 		spi->controller_state = asd;
-		gpio_direction_output(npcs_pin, !(spi->mode & SPI_CS_HIGH));
 	}
 
 	asd->csr = csr;
@@ -1332,6 +1578,13 @@ static int atmel_spi_probe(struct platform_device *pdev)
 
 	atmel_get_caps(as);
 
+	as->use_cs_gpios = true;
+	if (atmel_spi_is_v2(as) &&
+	    !of_get_property(pdev->dev.of_node, "cs-gpios", NULL)) {
+		as->use_cs_gpios = false;
+		master->num_chipselect = 4;
+	}
+
 	as->use_dma = false;
 	as->use_pdc = false;
 	if (as->caps.has_dma_support) {
@@ -1374,6 +1627,13 @@ static int atmel_spi_probe(struct platform_device *pdev)
 		spi_writel(as, PTCR, SPI_BIT(RXTDIS) | SPI_BIT(TXTDIS));
 	spi_writel(as, CR, SPI_BIT(SPIEN));
 
+	as->fifo_size = 0;
+	if (!of_property_read_u32(pdev->dev.of_node, "atmel,fifo-size",
+				  &as->fifo_size)) {
+		dev_info(&pdev->dev, "Using FIFO (%u data)\n", as->fifo_size);
+		spi_writel(as, CR, SPI_BIT(FIFOEN));
+	}
+
 	/* go! */
 	dev_info(&pdev->dev, "Atmel SPI Controller at 0x%08lx (irq %d)\n",
 			(unsigned long)regs->start, irq);
diff --git a/drivers/spi/spi-bcm2835.c b/drivers/spi/spi-bcm2835.c
index 419a782ab6d5..59705ab23577 100644
--- a/drivers/spi/spi-bcm2835.c
+++ b/drivers/spi/spi-bcm2835.c
@@ -3,6 +3,7 @@
  *
  * Copyright (C) 2012 Chris Boot
  * Copyright (C) 2013 Stephen Warren
+ * Copyright (C) 2015 Martin Sperl
  *
  * This driver is inspired by:
  * spi-ath79.c, Copyright (C) 2009-2011 Gabor Juhos <juhosg@openwrt.org>
@@ -19,17 +20,22 @@
  * GNU General Public License for more details.
  */
 
+#include <asm/page.h>
 #include <linux/clk.h>
 #include <linux/completion.h>
 #include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/dmaengine.h>
 #include <linux/err.h>
 #include <linux/interrupt.h>
 #include <linux/io.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/of.h>
-#include <linux/of_irq.h>
+#include <linux/of_address.h>
 #include <linux/of_device.h>
+#include <linux/of_gpio.h>
+#include <linux/of_irq.h>
 #include <linux/spi/spi.h>
 
 /* SPI register offsets */
@@ -66,8 +72,11 @@
 #define BCM2835_SPI_CS_CS_10		0x00000002
 #define BCM2835_SPI_CS_CS_01		0x00000001
 
-#define BCM2835_SPI_TIMEOUT_MS	30000
-#define BCM2835_SPI_MODE_BITS	(SPI_CPOL | SPI_CPHA | SPI_CS_HIGH | SPI_NO_CS)
+#define BCM2835_SPI_POLLING_LIMIT_US	30
+#define BCM2835_SPI_POLLING_JIFFIES	2
+#define BCM2835_SPI_DMA_MIN_LENGTH	96
+#define BCM2835_SPI_MODE_BITS	(SPI_CPOL | SPI_CPHA | SPI_CS_HIGH \
+				| SPI_NO_CS | SPI_3WIRE)
 
 #define DRV_NAME	"spi-bcm2835"
 
@@ -75,10 +84,11 @@ struct bcm2835_spi {
 	void __iomem *regs;
 	struct clk *clk;
 	int irq;
-	struct completion done;
 	const u8 *tx_buf;
 	u8 *rx_buf;
-	int len;
+	int tx_len;
+	int rx_len;
+	bool dma_pending;
 };
 
 static inline u32 bcm2835_rd(struct bcm2835_spi *bs, unsigned reg)
@@ -91,205 +101,618 @@ static inline void bcm2835_wr(struct bcm2835_spi *bs, unsigned reg, u32 val)
 	writel(val, bs->regs + reg);
 }
 
-static inline void bcm2835_rd_fifo(struct bcm2835_spi *bs, int len)
+static inline void bcm2835_rd_fifo(struct bcm2835_spi *bs)
 {
 	u8 byte;
 
-	while (len--) {
+	while ((bs->rx_len) &&
+	       (bcm2835_rd(bs, BCM2835_SPI_CS) & BCM2835_SPI_CS_RXD)) {
 		byte = bcm2835_rd(bs, BCM2835_SPI_FIFO);
 		if (bs->rx_buf)
 			*bs->rx_buf++ = byte;
+		bs->rx_len--;
 	}
 }
 
-static inline void bcm2835_wr_fifo(struct bcm2835_spi *bs, int len)
+static inline void bcm2835_wr_fifo(struct bcm2835_spi *bs)
 {
 	u8 byte;
 
-	if (len > bs->len)
-		len = bs->len;
-
-	while (len--) {
+	while ((bs->tx_len) &&
+	       (bcm2835_rd(bs, BCM2835_SPI_CS) & BCM2835_SPI_CS_TXD)) {
 		byte = bs->tx_buf ? *bs->tx_buf++ : 0;
 		bcm2835_wr(bs, BCM2835_SPI_FIFO, byte);
-		bs->len--;
+		bs->tx_len--;
 	}
 }
 
+static void bcm2835_spi_reset_hw(struct spi_master *master)
+{
+	struct bcm2835_spi *bs = spi_master_get_devdata(master);
+	u32 cs = bcm2835_rd(bs, BCM2835_SPI_CS);
+
+	/* Disable SPI interrupts and transfer */
+	cs &= ~(BCM2835_SPI_CS_INTR |
+		BCM2835_SPI_CS_INTD |
+		BCM2835_SPI_CS_DMAEN |
+		BCM2835_SPI_CS_TA);
+	/* and reset RX/TX FIFOS */
+	cs |= BCM2835_SPI_CS_CLEAR_RX | BCM2835_SPI_CS_CLEAR_TX;
+
+	/* and reset the SPI_HW */
+	bcm2835_wr(bs, BCM2835_SPI_CS, cs);
+	/* as well as DLEN */
+	bcm2835_wr(bs, BCM2835_SPI_DLEN, 0);
+}
+
 static irqreturn_t bcm2835_spi_interrupt(int irq, void *dev_id)
 {
 	struct spi_master *master = dev_id;
 	struct bcm2835_spi *bs = spi_master_get_devdata(master);
-	u32 cs = bcm2835_rd(bs, BCM2835_SPI_CS);
 
-	/*
-	 * RXR - RX needs Reading. This means 12 (or more) bytes have been
-	 * transmitted and hence 12 (or more) bytes have been received.
-	 *
-	 * The FIFO is 16-bytes deep. We check for this interrupt to keep the
-	 * FIFO full; we have a 4-byte-time buffer for IRQ latency. We check
-	 * this before DONE (TX empty) just in case we delayed processing this
-	 * interrupt for some reason.
-	 *
-	 * We only check for this case if we have more bytes to TX; at the end
-	 * of the transfer, we ignore this pipelining optimization, and let
-	 * bcm2835_spi_finish_transfer() drain the RX FIFO.
+	/* Read as many bytes as possible from FIFO */
+	bcm2835_rd_fifo(bs);
+	/* Write as many bytes as possible to FIFO */
+	bcm2835_wr_fifo(bs);
+
+	/* based on flags decide if we can finish the transfer */
+	if (bcm2835_rd(bs, BCM2835_SPI_CS) & BCM2835_SPI_CS_DONE) {
+		/* Transfer complete - reset SPI HW */
+		bcm2835_spi_reset_hw(master);
+		/* wake up the framework */
+		complete(&master->xfer_completion);
+	}
+
+	return IRQ_HANDLED;
+}
+
+static int bcm2835_spi_transfer_one_irq(struct spi_master *master,
+					struct spi_device *spi,
+					struct spi_transfer *tfr,
+					u32 cs)
+{
+	struct bcm2835_spi *bs = spi_master_get_devdata(master);
+
+	/* fill in fifo if we have gpio-cs
+	 * note that there have been rare events where the native-CS
+	 * flapped for <1us which may change the behaviour
+	 * with gpio-cs this does not happen, so it is implemented
+	 * only for this case
 	 */
-	if (bs->len && (cs & BCM2835_SPI_CS_RXR)) {
-		/* Read 12 bytes of data */
-		bcm2835_rd_fifo(bs, 12);
-
-		/* Write up to 12 bytes */
-		bcm2835_wr_fifo(bs, 12);
-
-		/*
-		 * We must have written something to the TX FIFO due to the
-		 * bs->len check above, so cannot be DONE. Hence, return
-		 * early. Note that DONE could also be set if we serviced an
-		 * RXR interrupt really late.
+	if (gpio_is_valid(spi->cs_gpio)) {
+		/* enable HW block, but without interrupts enabled
+		 * this would triggern an immediate interrupt
 		 */
-		return IRQ_HANDLED;
+		bcm2835_wr(bs, BCM2835_SPI_CS,
+			   cs | BCM2835_SPI_CS_TA);
+		/* fill in tx fifo as much as possible */
+		bcm2835_wr_fifo(bs);
 	}
 
 	/*
-	 * DONE - TX empty. This occurs when we first enable the transfer
-	 * since we do not pre-fill the TX FIFO. At any other time, given that
-	 * we refill the TX FIFO above based on RXR, and hence ignore DONE if
-	 * RXR is set, DONE really does mean end-of-transfer.
+	 * Enable the HW block. This will immediately trigger a DONE (TX
+	 * empty) interrupt, upon which we will fill the TX FIFO with the
+	 * first TX bytes. Pre-filling the TX FIFO here to avoid the
+	 * interrupt doesn't work:-(
 	 */
-	if (cs & BCM2835_SPI_CS_DONE) {
-		if (bs->len) { /* First interrupt in a transfer */
-			bcm2835_wr_fifo(bs, 16);
-		} else { /* Transfer complete */
-			/* Disable SPI interrupts */
-			cs &= ~(BCM2835_SPI_CS_INTR | BCM2835_SPI_CS_INTD);
-			bcm2835_wr(bs, BCM2835_SPI_CS, cs);
-
-			/*
-			 * Wake up bcm2835_spi_transfer_one(), which will call
-			 * bcm2835_spi_finish_transfer(), to drain the RX FIFO.
-			 */
-			complete(&bs->done);
-		}
+	cs |= BCM2835_SPI_CS_INTR | BCM2835_SPI_CS_INTD | BCM2835_SPI_CS_TA;
+	bcm2835_wr(bs, BCM2835_SPI_CS, cs);
+
+	/* signal that we need to wait for completion */
+	return 1;
+}
+
+/*
+ * DMA support
+ *
+ * this implementation has currently a few issues in so far as it does
+ * not work arrount limitations of the HW.
+ *
+ * the main one being that DMA transfers are limited to 16 bit
+ * (so 0 to 65535 bytes) by the SPI HW due to BCM2835_SPI_DLEN
+ *
+ * also we currently assume that the scatter-gather fragments are
+ * all multiple of 4 (except the last) - otherwise we would need
+ * to reset the FIFO before subsequent transfers...
+ * this also means that tx/rx transfers sg's need to be of equal size!
+ *
+ * there may be a few more border-cases we may need to address as well
+ * but unfortunately this would mean splitting up the scatter-gather
+ * list making it slightly unpractical...
+ */
+static void bcm2835_spi_dma_done(void *data)
+{
+	struct spi_master *master = data;
+	struct bcm2835_spi *bs = spi_master_get_devdata(master);
+
+	/* reset fifo and HW */
+	bcm2835_spi_reset_hw(master);
+
+	/* and terminate tx-dma as we do not have an irq for it
+	 * because when the rx dma will terminate and this callback
+	 * is called the tx-dma must have finished - can't get to this
+	 * situation otherwise...
+	 */
+	dmaengine_terminate_all(master->dma_tx);
+
+	/* mark as no longer pending */
+	bs->dma_pending = 0;
+
+	/* and mark as completed */;
+	complete(&master->xfer_completion);
+}
+
+static int bcm2835_spi_prepare_sg(struct spi_master *master,
+				  struct spi_transfer *tfr,
+				  bool is_tx)
+{
+	struct dma_chan *chan;
+	struct scatterlist *sgl;
+	unsigned int nents;
+	enum dma_transfer_direction dir;
+	unsigned long flags;
+
+	struct dma_async_tx_descriptor *desc;
+	dma_cookie_t cookie;
+
+	if (is_tx) {
+		dir   = DMA_MEM_TO_DEV;
+		chan  = master->dma_tx;
+		nents = tfr->tx_sg.nents;
+		sgl   = tfr->tx_sg.sgl;
+		flags = 0 /* no  tx interrupt */;
+
+	} else {
+		dir   = DMA_DEV_TO_MEM;
+		chan  = master->dma_rx;
+		nents = tfr->rx_sg.nents;
+		sgl   = tfr->rx_sg.sgl;
+		flags = DMA_PREP_INTERRUPT;
+	}
+	/* prepare the channel */
+	desc = dmaengine_prep_slave_sg(chan, sgl, nents, dir, flags);
+	if (!desc)
+		return -EINVAL;
+
+	/* set callback for rx */
+	if (!is_tx) {
+		desc->callback = bcm2835_spi_dma_done;
+		desc->callback_param = master;
+	}
+
+	/* submit it to DMA-engine */
+	cookie = dmaengine_submit(desc);
+
+	return dma_submit_error(cookie);
+}
+
+static inline int bcm2835_check_sg_length(struct sg_table *sgt)
+{
+	int i;
+	struct scatterlist *sgl;
+
+	/* check that the sg entries are word-sized (except for last) */
+	for_each_sg(sgt->sgl, sgl, (int)sgt->nents - 1, i) {
+		if (sg_dma_len(sgl) % 4)
+			return -EFAULT;
+	}
+
+	return 0;
+}
+
+static int bcm2835_spi_transfer_one_dma(struct spi_master *master,
+					struct spi_device *spi,
+					struct spi_transfer *tfr,
+					u32 cs)
+{
+	struct bcm2835_spi *bs = spi_master_get_devdata(master);
+	int ret;
+
+	/* check that the scatter gather segments are all a multiple of 4 */
+	if (bcm2835_check_sg_length(&tfr->tx_sg) ||
+	    bcm2835_check_sg_length(&tfr->rx_sg)) {
+		dev_warn_once(&spi->dev,
+			      "scatter gather segment length is not a multiple of 4 - falling back to interrupt mode\n");
+		return bcm2835_spi_transfer_one_irq(master, spi, tfr, cs);
+	}
+
+	/* setup tx-DMA */
+	ret = bcm2835_spi_prepare_sg(master, tfr, true);
+	if (ret)
+		return ret;
+
+	/* start TX early */
+	dma_async_issue_pending(master->dma_tx);
+
+	/* mark as dma pending */
+	bs->dma_pending = 1;
+
+	/* set the DMA length */
+	bcm2835_wr(bs, BCM2835_SPI_DLEN, tfr->len);
+
+	/* start the HW */
+	bcm2835_wr(bs, BCM2835_SPI_CS,
+		   cs | BCM2835_SPI_CS_TA | BCM2835_SPI_CS_DMAEN);
+
+	/* setup rx-DMA late - to run transfers while
+	 * mapping of the rx buffers still takes place
+	 * this saves 10us or more.
+	 */
+	ret = bcm2835_spi_prepare_sg(master, tfr, false);
+	if (ret) {
+		/* need to reset on errors */
+		dmaengine_terminate_all(master->dma_tx);
+		bcm2835_spi_reset_hw(master);
+		return ret;
+	}
+
+	/* start rx dma late */
+	dma_async_issue_pending(master->dma_rx);
+
+	/* wait for wakeup in framework */
+	return 1;
+}
+
+static bool bcm2835_spi_can_dma(struct spi_master *master,
+				struct spi_device *spi,
+				struct spi_transfer *tfr)
+{
+	/* only run for gpio_cs */
+	if (!gpio_is_valid(spi->cs_gpio))
+		return false;
+
+	/* we start DMA efforts only on bigger transfers */
+	if (tfr->len < BCM2835_SPI_DMA_MIN_LENGTH)
+		return false;
+
+	/* BCM2835_SPI_DLEN has defined a max transfer size as
+	 * 16 bit, so max is 65535
+	 * we can revisit this by using an alternative transfer
+	 * method - ideally this would get done without any more
+	 * interaction...
+	 */
+	if (tfr->len > 65535) {
+		dev_warn_once(&spi->dev,
+			      "transfer size of %d too big for dma-transfer\n",
+			      tfr->len);
+		return false;
+	}
+
+	/* if we run rx/tx_buf with word aligned addresses then we are OK */
+	if ((((size_t)tfr->rx_buf & 3) == 0) &&
+	    (((size_t)tfr->tx_buf & 3) == 0))
+		return true;
+
+	/* otherwise we only allow transfers within the same page
+	 * to avoid wasting time on dma_mapping when it is not practical
+	 */
+	if (((size_t)tfr->tx_buf & PAGE_MASK) + tfr->len > PAGE_SIZE) {
+		dev_warn_once(&spi->dev,
+			      "Unaligned spi tx-transfer bridging page\n");
+		return false;
+	}
+	if (((size_t)tfr->rx_buf & PAGE_MASK) + tfr->len > PAGE_SIZE) {
+		dev_warn_once(&spi->dev,
+			      "Unaligned spi tx-transfer bridging page\n");
+		return false;
+	}
 
-		return IRQ_HANDLED;
+	/* return OK */
+	return true;
+}
+
+static void bcm2835_dma_release(struct spi_master *master)
+{
+	if (master->dma_tx) {
+		dmaengine_terminate_all(master->dma_tx);
+		dma_release_channel(master->dma_tx);
+		master->dma_tx = NULL;
+	}
+	if (master->dma_rx) {
+		dmaengine_terminate_all(master->dma_rx);
+		dma_release_channel(master->dma_rx);
+		master->dma_rx = NULL;
+	}
+}
+
+static void bcm2835_dma_init(struct spi_master *master, struct device *dev)
+{
+	struct dma_slave_config slave_config;
+	const __be32 *addr;
+	dma_addr_t dma_reg_base;
+	int ret;
+
+	/* base address in dma-space */
+	addr = of_get_address(master->dev.of_node, 0, NULL, NULL);
+	if (!addr) {
+		dev_err(dev, "could not get DMA-register address - not using dma mode\n");
+		goto err;
+	}
+	dma_reg_base = be32_to_cpup(addr);
+
+	/* get tx/rx dma */
+	master->dma_tx = dma_request_slave_channel(dev, "tx");
+	if (!master->dma_tx) {
+		dev_err(dev, "no tx-dma configuration found - not using dma mode\n");
+		goto err;
+	}
+	master->dma_rx = dma_request_slave_channel(dev, "rx");
+	if (!master->dma_rx) {
+		dev_err(dev, "no rx-dma configuration found - not using dma mode\n");
+		goto err_release;
 	}
 
-	return IRQ_NONE;
+	/* configure DMAs */
+	slave_config.direction = DMA_MEM_TO_DEV;
+	slave_config.dst_addr = (u32)(dma_reg_base + BCM2835_SPI_FIFO);
+	slave_config.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+
+	ret = dmaengine_slave_config(master->dma_tx, &slave_config);
+	if (ret)
+		goto err_config;
+
+	slave_config.direction = DMA_DEV_TO_MEM;
+	slave_config.src_addr = (u32)(dma_reg_base + BCM2835_SPI_FIFO);
+	slave_config.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+
+	ret = dmaengine_slave_config(master->dma_rx, &slave_config);
+	if (ret)
+		goto err_config;
+
+	/* all went well, so set can_dma */
+	master->can_dma = bcm2835_spi_can_dma;
+	master->max_dma_len = 65535; /* limitation by BCM2835_SPI_DLEN */
+	/* need to do TX AND RX DMA, so we need dummy buffers */
+	master->flags = SPI_MASTER_MUST_RX | SPI_MASTER_MUST_TX;
+
+	return;
+
+err_config:
+	dev_err(dev, "issue configuring dma: %d - not using DMA mode\n",
+		ret);
+err_release:
+	bcm2835_dma_release(master);
+err:
+	return;
 }
 
-static int bcm2835_spi_start_transfer(struct spi_device *spi,
-		struct spi_transfer *tfr)
+static int bcm2835_spi_transfer_one_poll(struct spi_master *master,
+					 struct spi_device *spi,
+					 struct spi_transfer *tfr,
+					 u32 cs,
+					 unsigned long xfer_time_us)
 {
-	struct bcm2835_spi *bs = spi_master_get_devdata(spi->master);
+	struct bcm2835_spi *bs = spi_master_get_devdata(master);
+	unsigned long timeout;
+
+	/* enable HW block without interrupts */
+	bcm2835_wr(bs, BCM2835_SPI_CS, cs | BCM2835_SPI_CS_TA);
+
+	/* fill in the fifo before timeout calculations
+	 * if we are interrupted here, then the data is
+	 * getting transferred by the HW while we are interrupted
+	 */
+	bcm2835_wr_fifo(bs);
+
+	/* set the timeout */
+	timeout = jiffies + BCM2835_SPI_POLLING_JIFFIES;
+
+	/* loop until finished the transfer */
+	while (bs->rx_len) {
+		/* fill in tx fifo with remaining data */
+		bcm2835_wr_fifo(bs);
+
+		/* read from fifo as much as possible */
+		bcm2835_rd_fifo(bs);
+
+		/* if there is still data pending to read
+		 * then check the timeout
+		 */
+		if (bs->rx_len && time_after(jiffies, timeout)) {
+			dev_dbg_ratelimited(&spi->dev,
+					    "timeout period reached: jiffies: %lu remaining tx/rx: %d/%d - falling back to interrupt mode\n",
+					    jiffies - timeout,
+					    bs->tx_len, bs->rx_len);
+			/* fall back to interrupt mode */
+			return bcm2835_spi_transfer_one_irq(master, spi,
+							    tfr, cs);
+		}
+	}
+
+	/* Transfer complete - reset SPI HW */
+	bcm2835_spi_reset_hw(master);
+	/* and return without waiting for completion */
+	return 0;
+}
+
+static int bcm2835_spi_transfer_one(struct spi_master *master,
+				    struct spi_device *spi,
+				    struct spi_transfer *tfr)
+{
+	struct bcm2835_spi *bs = spi_master_get_devdata(master);
 	unsigned long spi_hz, clk_hz, cdiv;
-	u32 cs = BCM2835_SPI_CS_INTR | BCM2835_SPI_CS_INTD | BCM2835_SPI_CS_TA;
+	unsigned long spi_used_hz, xfer_time_us;
+	u32 cs = bcm2835_rd(bs, BCM2835_SPI_CS);
 
+	/* set clock */
 	spi_hz = tfr->speed_hz;
 	clk_hz = clk_get_rate(bs->clk);
 
 	if (spi_hz >= clk_hz / 2) {
 		cdiv = 2; /* clk_hz/2 is the fastest we can go */
 	} else if (spi_hz) {
-		/* CDIV must be a power of two */
-		cdiv = roundup_pow_of_two(DIV_ROUND_UP(clk_hz, spi_hz));
+		/* CDIV must be a multiple of two */
+		cdiv = DIV_ROUND_UP(clk_hz, spi_hz);
+		cdiv += (cdiv % 2);
 
 		if (cdiv >= 65536)
 			cdiv = 0; /* 0 is the slowest we can go */
-	} else
+	} else {
 		cdiv = 0; /* 0 is the slowest we can go */
+	}
+	spi_used_hz = cdiv ? (clk_hz / cdiv) : (clk_hz / 65536);
+	bcm2835_wr(bs, BCM2835_SPI_CLK, cdiv);
 
+	/* handle all the modes */
+	if ((spi->mode & SPI_3WIRE) && (tfr->rx_buf))
+		cs |= BCM2835_SPI_CS_REN;
 	if (spi->mode & SPI_CPOL)
 		cs |= BCM2835_SPI_CS_CPOL;
 	if (spi->mode & SPI_CPHA)
 		cs |= BCM2835_SPI_CS_CPHA;
 
-	if (!(spi->mode & SPI_NO_CS)) {
-		if (spi->mode & SPI_CS_HIGH) {
-			cs |= BCM2835_SPI_CS_CSPOL;
-			cs |= BCM2835_SPI_CS_CSPOL0 << spi->chip_select;
-		}
-
-		cs |= spi->chip_select;
-	}
+	/* for gpio_cs set dummy CS so that no HW-CS get changed
+	 * we can not run this in bcm2835_spi_set_cs, as it does
+	 * not get called for cs_gpio cases, so we need to do it here
+	 */
+	if (gpio_is_valid(spi->cs_gpio) || (spi->mode & SPI_NO_CS))
+		cs |= BCM2835_SPI_CS_CS_10 | BCM2835_SPI_CS_CS_01;
 
-	reinit_completion(&bs->done);
+	/* set transmit buffers and length */
 	bs->tx_buf = tfr->tx_buf;
 	bs->rx_buf = tfr->rx_buf;
-	bs->len = tfr->len;
+	bs->tx_len = tfr->len;
+	bs->rx_len = tfr->len;
 
-	bcm2835_wr(bs, BCM2835_SPI_CLK, cdiv);
-	/*
-	 * Enable the HW block. This will immediately trigger a DONE (TX
-	 * empty) interrupt, upon which we will fill the TX FIFO with the
-	 * first TX bytes. Pre-filling the TX FIFO here to avoid the
-	 * interrupt doesn't work:-(
-	 */
-	bcm2835_wr(bs, BCM2835_SPI_CS, cs);
+	/* calculate the estimated time in us the transfer runs */
+	xfer_time_us = tfr->len
+		* 9 /* clocks/byte - SPI-HW waits 1 clock after each byte */
+		* 1000000 / spi_used_hz;
 
-	return 0;
+	/* for short requests run polling*/
+	if (xfer_time_us <= BCM2835_SPI_POLLING_LIMIT_US)
+		return bcm2835_spi_transfer_one_poll(master, spi, tfr,
+						     cs, xfer_time_us);
+
+	/* run in dma mode if conditions are right */
+	if (master->can_dma && bcm2835_spi_can_dma(master, spi, tfr))
+		return bcm2835_spi_transfer_one_dma(master, spi, tfr, cs);
+
+	/* run in interrupt-mode */
+	return bcm2835_spi_transfer_one_irq(master, spi, tfr, cs);
 }
 
-static int bcm2835_spi_finish_transfer(struct spi_device *spi,
-		struct spi_transfer *tfr, bool cs_change)
+static void bcm2835_spi_handle_err(struct spi_master *master,
+				   struct spi_message *msg)
 {
-	struct bcm2835_spi *bs = spi_master_get_devdata(spi->master);
-	u32 cs = bcm2835_rd(bs, BCM2835_SPI_CS);
+	struct bcm2835_spi *bs = spi_master_get_devdata(master);
 
-	/* Drain RX FIFO */
-	while (cs & BCM2835_SPI_CS_RXD) {
-		bcm2835_rd_fifo(bs, 1);
-		cs = bcm2835_rd(bs, BCM2835_SPI_CS);
+	/* if an error occurred and we have an active dma, then terminate */
+	if (bs->dma_pending) {
+		dmaengine_terminate_all(master->dma_tx);
+		dmaengine_terminate_all(master->dma_rx);
+		bs->dma_pending = 0;
 	}
-
-	if (tfr->delay_usecs)
-		udelay(tfr->delay_usecs);
-
-	if (cs_change)
-		/* Clear TA flag */
-		bcm2835_wr(bs, BCM2835_SPI_CS, cs & ~BCM2835_SPI_CS_TA);
-
-	return 0;
+	/* and reset */
+	bcm2835_spi_reset_hw(master);
 }
 
-static int bcm2835_spi_transfer_one(struct spi_master *master,
-		struct spi_message *mesg)
+static void bcm2835_spi_set_cs(struct spi_device *spi, bool gpio_level)
 {
+	/*
+	 * we can assume that we are "native" as per spi_set_cs
+	 *   calling us ONLY when cs_gpio is not set
+	 * we can also assume that we are CS < 3 as per bcm2835_spi_setup
+	 *   we would not get called because of error handling there.
+	 * the level passed is the electrical level not enabled/disabled
+	 *   so it has to get translated back to enable/disable
+	 *   see spi_set_cs in spi.c for the implementation
+	 */
+
+	struct spi_master *master = spi->master;
 	struct bcm2835_spi *bs = spi_master_get_devdata(master);
-	struct spi_transfer *tfr;
-	struct spi_device *spi = mesg->spi;
-	int err = 0;
-	unsigned int timeout;
-	bool cs_change;
-
-	list_for_each_entry(tfr, &mesg->transfers, transfer_list) {
-		err = bcm2835_spi_start_transfer(spi, tfr);
-		if (err)
-			goto out;
-
-		timeout = wait_for_completion_timeout(&bs->done,
-				msecs_to_jiffies(BCM2835_SPI_TIMEOUT_MS));
-		if (!timeout) {
-			err = -ETIMEDOUT;
-			goto out;
+	u32 cs = bcm2835_rd(bs, BCM2835_SPI_CS);
+	bool enable;
+
+	/* calculate the enable flag from the passed gpio_level */
+	enable = (spi->mode & SPI_CS_HIGH) ? gpio_level : !gpio_level;
+
+	/* set flags for "reverse" polarity in the registers */
+	if (spi->mode & SPI_CS_HIGH) {
+		/* set the correct CS-bits */
+		cs |= BCM2835_SPI_CS_CSPOL;
+		cs |= BCM2835_SPI_CS_CSPOL0 << spi->chip_select;
+	} else {
+		/* clean the CS-bits */
+		cs &= ~BCM2835_SPI_CS_CSPOL;
+		cs &= ~(BCM2835_SPI_CS_CSPOL0 << spi->chip_select);
+	}
+
+	/* select the correct chip_select depending on disabled/enabled */
+	if (enable) {
+		/* set cs correctly */
+		if (spi->mode & SPI_NO_CS) {
+			/* use the "undefined" chip-select */
+			cs |= BCM2835_SPI_CS_CS_10 | BCM2835_SPI_CS_CS_01;
+		} else {
+			/* set the chip select */
+			cs &= ~(BCM2835_SPI_CS_CS_10 | BCM2835_SPI_CS_CS_01);
+			cs |= spi->chip_select;
 		}
+	} else {
+		/* disable CSPOL which puts HW-CS into deselected state */
+		cs &= ~BCM2835_SPI_CS_CSPOL;
+		/* use the "undefined" chip-select as precaution */
+		cs |= BCM2835_SPI_CS_CS_10 | BCM2835_SPI_CS_CS_01;
+	}
 
-		cs_change = tfr->cs_change ||
-			list_is_last(&tfr->transfer_list, &mesg->transfers);
+	/* finally set the calculated flags in SPI_CS */
+	bcm2835_wr(bs, BCM2835_SPI_CS, cs);
+}
 
-		err = bcm2835_spi_finish_transfer(spi, tfr, cs_change);
-		if (err)
-			goto out;
+static int chip_match_name(struct gpio_chip *chip, void *data)
+{
+	return !strcmp(chip->label, data);
+}
 
-		mesg->actual_length += (tfr->len - bs->len);
+static int bcm2835_spi_setup(struct spi_device *spi)
+{
+	int err;
+	struct gpio_chip *chip;
+	/*
+	 * sanity checking the native-chipselects
+	 */
+	if (spi->mode & SPI_NO_CS)
+		return 0;
+	if (gpio_is_valid(spi->cs_gpio))
+		return 0;
+	if (spi->chip_select > 1) {
+		/* error in the case of native CS requested with CS > 1
+		 * officially there is a CS2, but it is not documented
+		 * which GPIO is connected with that...
+		 */
+		dev_err(&spi->dev,
+			"setup: only two native chip-selects are supported\n");
+		return -EINVAL;
 	}
+	/* now translate native cs to GPIO */
 
-out:
-	/* Clear FIFOs, and disable the HW block */
-	bcm2835_wr(bs, BCM2835_SPI_CS,
-		   BCM2835_SPI_CS_CLEAR_RX | BCM2835_SPI_CS_CLEAR_TX);
-	mesg->status = err;
-	spi_finalize_current_message(master);
+	/* get the gpio chip for the base */
+	chip = gpiochip_find("pinctrl-bcm2835", chip_match_name);
+	if (!chip)
+		return 0;
+
+	/* and calculate the real CS */
+	spi->cs_gpio = chip->base + 8 - spi->chip_select;
+
+	/* and set up the "mode" and level */
+	dev_info(&spi->dev, "setting up native-CS%i as GPIO %i\n",
+		 spi->chip_select, spi->cs_gpio);
+
+	/* set up GPIO as output and pull to the correct level */
+	err = gpio_direction_output(spi->cs_gpio,
+				    (spi->mode & SPI_CS_HIGH) ? 0 : 1);
+	if (err) {
+		dev_err(&spi->dev,
+			"could not set CS%i gpio %i as output: %i",
+			spi->chip_select, spi->cs_gpio, err);
+		return err;
+	}
+	/* the implementation of pinctrl-bcm2835 currently does not
+	 * set the GPIO value when using gpio_direction_output
+	 * so we are setting it here explicitly
+	 */
+	gpio_set_value(spi->cs_gpio, (spi->mode & SPI_CS_HIGH) ? 0 : 1);
 
 	return 0;
 }
@@ -312,13 +735,14 @@ static int bcm2835_spi_probe(struct platform_device *pdev)
 	master->mode_bits = BCM2835_SPI_MODE_BITS;
 	master->bits_per_word_mask = SPI_BPW_MASK(8);
 	master->num_chipselect = 3;
-	master->transfer_one_message = bcm2835_spi_transfer_one;
+	master->setup = bcm2835_spi_setup;
+	master->set_cs = bcm2835_spi_set_cs;
+	master->transfer_one = bcm2835_spi_transfer_one;
+	master->handle_err = bcm2835_spi_handle_err;
 	master->dev.of_node = pdev->dev.of_node;
 
 	bs = spi_master_get_devdata(master);
 
-	init_completion(&bs->done);
-
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	bs->regs = devm_ioremap_resource(&pdev->dev, res);
 	if (IS_ERR(bs->regs)) {
@@ -343,13 +767,15 @@ static int bcm2835_spi_probe(struct platform_device *pdev)
 	clk_prepare_enable(bs->clk);
 
 	err = devm_request_irq(&pdev->dev, bs->irq, bcm2835_spi_interrupt, 0,
-				dev_name(&pdev->dev), master);
+			       dev_name(&pdev->dev), master);
 	if (err) {
 		dev_err(&pdev->dev, "could not request IRQ: %d\n", err);
 		goto out_clk_disable;
 	}
 
-	/* initialise the hardware */
+	bcm2835_dma_init(master, &pdev->dev);
+
+	/* initialise the hardware with the default polarities */
 	bcm2835_wr(bs, BCM2835_SPI_CS,
 		   BCM2835_SPI_CS_CLEAR_RX | BCM2835_SPI_CS_CLEAR_TX);
 
@@ -379,6 +805,8 @@ static int bcm2835_spi_remove(struct platform_device *pdev)
 
 	clk_disable_unprepare(bs->clk);
 
+	bcm2835_dma_release(master);
+
 	return 0;
 }
 
diff --git a/drivers/spi/spi-bcm53xx.c b/drivers/spi/spi-bcm53xx.c
index 3fb91c81015a..1520554978a3 100644
--- a/drivers/spi/spi-bcm53xx.c
+++ b/drivers/spi/spi-bcm53xx.c
@@ -44,7 +44,7 @@ static int bcm53xxspi_wait(struct bcm53xxspi *b53spi, unsigned int timeout_ms)
 	u32 tmp;
 
 	/* SPE bit has to be 0 before we read MSPI STATUS */
-	deadline = jiffies + BCM53XXSPI_SPE_TIMEOUT_MS * HZ / 1000;
+	deadline = jiffies + msecs_to_jiffies(BCM53XXSPI_SPE_TIMEOUT_MS);
 	do {
 		tmp = bcm53xxspi_read(b53spi, B53SPI_MSPI_SPCR2);
 		if (!(tmp & B53SPI_MSPI_SPCR2_SPE))
@@ -56,7 +56,7 @@ static int bcm53xxspi_wait(struct bcm53xxspi *b53spi, unsigned int timeout_ms)
 		goto spi_timeout;
 
 	/* Check status */
-	deadline = jiffies + timeout_ms * HZ / 1000;
+	deadline = jiffies + msecs_to_jiffies(timeout_ms);
 	do {
 		tmp = bcm53xxspi_read(b53spi, B53SPI_MSPI_MSPI_STATUS);
 		if (tmp & B53SPI_MSPI_MSPI_STATUS_SPIF) {
diff --git a/drivers/spi/spi-bfin5xx.c b/drivers/spi/spi-bfin5xx.c
index 37079937d2f7..a3d65b4f4944 100644
--- a/drivers/spi/spi-bfin5xx.c
+++ b/drivers/spi/spi-bfin5xx.c
@@ -559,7 +559,7 @@ static void bfin_spi_pump_transfers(unsigned long data)
 	struct spi_transfer *previous = NULL;
 	struct bfin_spi_slave_data *chip = NULL;
 	unsigned int bits_per_word;
-	u16 cr, cr_width, dma_width, dma_config;
+	u16 cr, cr_width = 0, dma_width, dma_config;
 	u32 tranf_success = 1;
 	u8 full_duplex = 0;
 
@@ -648,7 +648,6 @@ static void bfin_spi_pump_transfers(unsigned long data)
 	} else if (bits_per_word == 8) {
 		drv_data->n_bytes = bits_per_word/8;
 		drv_data->len = transfer->len;
-		cr_width = 0;
 		drv_data->ops = &bfin_bfin_spi_transfer_ops_u8;
 	}
 	cr = bfin_read(&drv_data->regs->ctl) & ~(BIT_CTL_TIMOD | BIT_CTL_WORDSIZE);
diff --git a/drivers/spi/spi-bitbang-txrx.h b/drivers/spi/spi-bitbang-txrx.h
index c616e41521be..06b34e5bcfa3 100644
--- a/drivers/spi/spi-bitbang-txrx.h
+++ b/drivers/spi/spi-bitbang-txrx.h
@@ -49,12 +49,17 @@ bitbang_txrx_be_cpha0(struct spi_device *spi,
 {
 	/* if (cpol == 0) this is SPI_MODE_0; else this is SPI_MODE_2 */
 
+	bool oldbit = !(word & 1);
 	/* clock starts at inactive polarity */
 	for (word <<= (32 - bits); likely(bits); bits--) {
 
 		/* setup MSB (to slave) on trailing edge */
-		if ((flags & SPI_MASTER_NO_TX) == 0)
-			setmosi(spi, word & (1 << 31));
+		if ((flags & SPI_MASTER_NO_TX) == 0) {
+			if ((word & (1 << 31)) != oldbit) {
+				setmosi(spi, word & (1 << 31));
+				oldbit = word & (1 << 31);
+			}
+		}
 		spidelay(nsecs);	/* T(setup) */
 
 		setsck(spi, !cpol);
@@ -76,13 +81,18 @@ bitbang_txrx_be_cpha1(struct spi_device *spi,
 {
 	/* if (cpol == 0) this is SPI_MODE_1; else this is SPI_MODE_3 */
 
+	bool oldbit = !(word & (1 << 31));
 	/* clock starts at inactive polarity */
 	for (word <<= (32 - bits); likely(bits); bits--) {
 
 		/* setup MSB (to slave) on leading edge */
 		setsck(spi, !cpol);
-		if ((flags & SPI_MASTER_NO_TX) == 0)
-			setmosi(spi, word & (1 << 31));
+		if ((flags & SPI_MASTER_NO_TX) == 0) {
+			if ((word & (1 << 31)) != oldbit) {
+				setmosi(spi, word & (1 << 31));
+				oldbit = word & (1 << 31);
+			}
+		}
 		spidelay(nsecs); /* T(setup) */
 
 		setsck(spi, cpol);
diff --git a/drivers/spi/spi-bitbang.c b/drivers/spi/spi-bitbang.c
index 5ef6638d5e8a..840a4984d365 100644
--- a/drivers/spi/spi-bitbang.c
+++ b/drivers/spi/spi-bitbang.c
@@ -180,7 +180,6 @@ int spi_bitbang_setup(struct spi_device *spi)
 {
 	struct spi_bitbang_cs	*cs = spi->controller_state;
 	struct spi_bitbang	*bitbang;
-	int			retval;
 	unsigned long		flags;
 
 	bitbang = spi_master_get_devdata(spi->master);
@@ -197,9 +196,11 @@ int spi_bitbang_setup(struct spi_device *spi)
 	if (!cs->txrx_word)
 		return -EINVAL;
 
-	retval = bitbang->setup_transfer(spi, NULL);
-	if (retval < 0)
-		return retval;
+	if (bitbang->setup_transfer) {
+		int retval = bitbang->setup_transfer(spi, NULL);
+		if (retval < 0)
+			return retval;
+	}
 
 	dev_dbg(&spi->dev, "%s, %u nsec/bit\n", __func__, 2 * cs->nsecs);
 
@@ -295,9 +296,11 @@ static int spi_bitbang_transfer_one(struct spi_master *master,
 
 		/* init (-1) or override (1) transfer params */
 		if (do_setup != 0) {
-			status = bitbang->setup_transfer(spi, t);
-			if (status < 0)
-				break;
+			if (bitbang->setup_transfer) {
+				status = bitbang->setup_transfer(spi, t);
+				if (status < 0)
+					break;
+			}
 			if (do_setup == -1)
 				do_setup = 0;
 		}
diff --git a/drivers/spi/spi-davinci.c b/drivers/spi/spi-davinci.c
index 5e991065f5b0..987afebea093 100644
--- a/drivers/spi/spi-davinci.c
+++ b/drivers/spi/spi-davinci.c
@@ -265,7 +265,7 @@ static inline int davinci_spi_get_prescale(struct davinci_spi *dspi,
 
 	ret = DIV_ROUND_UP(clk_get_rate(dspi->clk), max_speed_hz);
 
-	if (ret < 3 || ret > 256)
+	if (ret < 1 || ret > 256)
 		return -EINVAL;
 
 	return ret - 1;
diff --git a/drivers/spi/spi-dw-mid.c b/drivers/spi/spi-dw-mid.c
index 4f8c798e0633..bb1052e748f2 100644
--- a/drivers/spi/spi-dw-mid.c
+++ b/drivers/spi/spi-dw-mid.c
@@ -23,29 +23,31 @@
 #include "spi-dw.h"
 
 #ifdef CONFIG_SPI_DW_MID_DMA
-#include <linux/intel_mid_dma.h>
 #include <linux/pci.h>
+#include <linux/platform_data/dma-dw.h>
 
 #define RX_BUSY		0
 #define TX_BUSY		1
 
-struct mid_dma {
-	struct intel_mid_dma_slave	dmas_tx;
-	struct intel_mid_dma_slave	dmas_rx;
-};
+static struct dw_dma_slave mid_dma_tx = { .dst_id = 1 };
+static struct dw_dma_slave mid_dma_rx = { .src_id = 0 };
 
 static bool mid_spi_dma_chan_filter(struct dma_chan *chan, void *param)
 {
-	struct dw_spi *dws = param;
+	struct dw_dma_slave *s = param;
+
+	if (s->dma_dev != chan->device->dev)
+		return false;
 
-	return dws->dma_dev == chan->device->dev;
+	chan->private = s;
+	return true;
 }
 
 static int mid_spi_dma_init(struct dw_spi *dws)
 {
-	struct mid_dma *dw_dma = dws->dma_priv;
 	struct pci_dev *dma_dev;
-	struct intel_mid_dma_slave *rxs, *txs;
+	struct dw_dma_slave *tx = dws->dma_tx;
+	struct dw_dma_slave *rx = dws->dma_rx;
 	dma_cap_mask_t mask;
 
 	/*
@@ -56,28 +58,22 @@ static int mid_spi_dma_init(struct dw_spi *dws)
 	if (!dma_dev)
 		return -ENODEV;
 
-	dws->dma_dev = &dma_dev->dev;
-
 	dma_cap_zero(mask);
 	dma_cap_set(DMA_SLAVE, mask);
 
 	/* 1. Init rx channel */
-	dws->rxchan = dma_request_channel(mask, mid_spi_dma_chan_filter, dws);
+	rx->dma_dev = &dma_dev->dev;
+	dws->rxchan = dma_request_channel(mask, mid_spi_dma_chan_filter, rx);
 	if (!dws->rxchan)
 		goto err_exit;
-	rxs = &dw_dma->dmas_rx;
-	rxs->hs_mode = LNW_DMA_HW_HS;
-	rxs->cfg_mode = LNW_DMA_PER_TO_MEM;
-	dws->rxchan->private = rxs;
+	dws->master->dma_rx = dws->rxchan;
 
 	/* 2. Init tx channel */
-	dws->txchan = dma_request_channel(mask, mid_spi_dma_chan_filter, dws);
+	tx->dma_dev = &dma_dev->dev;
+	dws->txchan = dma_request_channel(mask, mid_spi_dma_chan_filter, tx);
 	if (!dws->txchan)
 		goto free_rxchan;
-	txs = &dw_dma->dmas_tx;
-	txs->hs_mode = LNW_DMA_HW_HS;
-	txs->cfg_mode = LNW_DMA_MEM_TO_PER;
-	dws->txchan->private = txs;
+	dws->master->dma_tx = dws->txchan;
 
 	dws->dma_inited = 1;
 	return 0;
@@ -100,6 +96,42 @@ static void mid_spi_dma_exit(struct dw_spi *dws)
 	dma_release_channel(dws->rxchan);
 }
 
+static irqreturn_t dma_transfer(struct dw_spi *dws)
+{
+	u16 irq_status = dw_readl(dws, DW_SPI_ISR);
+
+	if (!irq_status)
+		return IRQ_NONE;
+
+	dw_readl(dws, DW_SPI_ICR);
+	spi_reset_chip(dws);
+
+	dev_err(&dws->master->dev, "%s: FIFO overrun/underrun\n", __func__);
+	dws->master->cur_msg->status = -EIO;
+	spi_finalize_current_transfer(dws->master);
+	return IRQ_HANDLED;
+}
+
+static bool mid_spi_can_dma(struct spi_master *master, struct spi_device *spi,
+		struct spi_transfer *xfer)
+{
+	struct dw_spi *dws = spi_master_get_devdata(master);
+
+	if (!dws->dma_inited)
+		return false;
+
+	return xfer->len > dws->fifo_len;
+}
+
+static enum dma_slave_buswidth convert_dma_width(u32 dma_width) {
+	if (dma_width == 1)
+		return DMA_SLAVE_BUSWIDTH_1_BYTE;
+	else if (dma_width == 2)
+		return DMA_SLAVE_BUSWIDTH_2_BYTES;
+
+	return DMA_SLAVE_BUSWIDTH_UNDEFINED;
+}
+
 /*
  * dws->dma_chan_busy is set before the dma transfer starts, callback for tx
  * channel will clear a corresponding bit.
@@ -111,33 +143,30 @@ static void dw_spi_dma_tx_done(void *arg)
 	clear_bit(TX_BUSY, &dws->dma_chan_busy);
 	if (test_bit(RX_BUSY, &dws->dma_chan_busy))
 		return;
-	dw_spi_xfer_done(dws);
+	spi_finalize_current_transfer(dws->master);
 }
 
-static struct dma_async_tx_descriptor *dw_spi_dma_prepare_tx(struct dw_spi *dws)
+static struct dma_async_tx_descriptor *dw_spi_dma_prepare_tx(struct dw_spi *dws,
+		struct spi_transfer *xfer)
 {
 	struct dma_slave_config txconf;
 	struct dma_async_tx_descriptor *txdesc;
 
-	if (!dws->tx_dma)
+	if (!xfer->tx_buf)
 		return NULL;
 
 	txconf.direction = DMA_MEM_TO_DEV;
 	txconf.dst_addr = dws->dma_addr;
-	txconf.dst_maxburst = LNW_DMA_MSIZE_16;
+	txconf.dst_maxburst = 16;
 	txconf.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
-	txconf.dst_addr_width = dws->dma_width;
+	txconf.dst_addr_width = convert_dma_width(dws->dma_width);
 	txconf.device_fc = false;
 
 	dmaengine_slave_config(dws->txchan, &txconf);
 
-	memset(&dws->tx_sgl, 0, sizeof(dws->tx_sgl));
-	dws->tx_sgl.dma_address = dws->tx_dma;
-	dws->tx_sgl.length = dws->len;
-
 	txdesc = dmaengine_prep_slave_sg(dws->txchan,
-				&dws->tx_sgl,
-				1,
+				xfer->tx_sg.sgl,
+				xfer->tx_sg.nents,
 				DMA_MEM_TO_DEV,
 				DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
 	if (!txdesc)
@@ -160,33 +189,30 @@ static void dw_spi_dma_rx_done(void *arg)
 	clear_bit(RX_BUSY, &dws->dma_chan_busy);
 	if (test_bit(TX_BUSY, &dws->dma_chan_busy))
 		return;
-	dw_spi_xfer_done(dws);
+	spi_finalize_current_transfer(dws->master);
 }
 
-static struct dma_async_tx_descriptor *dw_spi_dma_prepare_rx(struct dw_spi *dws)
+static struct dma_async_tx_descriptor *dw_spi_dma_prepare_rx(struct dw_spi *dws,
+		struct spi_transfer *xfer)
 {
 	struct dma_slave_config rxconf;
 	struct dma_async_tx_descriptor *rxdesc;
 
-	if (!dws->rx_dma)
+	if (!xfer->rx_buf)
 		return NULL;
 
 	rxconf.direction = DMA_DEV_TO_MEM;
 	rxconf.src_addr = dws->dma_addr;
-	rxconf.src_maxburst = LNW_DMA_MSIZE_16;
+	rxconf.src_maxburst = 16;
 	rxconf.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
-	rxconf.src_addr_width = dws->dma_width;
+	rxconf.src_addr_width = convert_dma_width(dws->dma_width);
 	rxconf.device_fc = false;
 
 	dmaengine_slave_config(dws->rxchan, &rxconf);
 
-	memset(&dws->rx_sgl, 0, sizeof(dws->rx_sgl));
-	dws->rx_sgl.dma_address = dws->rx_dma;
-	dws->rx_sgl.length = dws->len;
-
 	rxdesc = dmaengine_prep_slave_sg(dws->rxchan,
-				&dws->rx_sgl,
-				1,
+				xfer->rx_sg.sgl,
+				xfer->rx_sg.nents,
 				DMA_DEV_TO_MEM,
 				DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
 	if (!rxdesc)
@@ -198,37 +224,36 @@ static struct dma_async_tx_descriptor *dw_spi_dma_prepare_rx(struct dw_spi *dws)
 	return rxdesc;
 }
 
-static void dw_spi_dma_setup(struct dw_spi *dws)
+static int mid_spi_dma_setup(struct dw_spi *dws, struct spi_transfer *xfer)
 {
 	u16 dma_ctrl = 0;
 
-	spi_enable_chip(dws, 0);
+	dw_writel(dws, DW_SPI_DMARDLR, 0xf);
+	dw_writel(dws, DW_SPI_DMATDLR, 0x10);
 
-	dw_writew(dws, DW_SPI_DMARDLR, 0xf);
-	dw_writew(dws, DW_SPI_DMATDLR, 0x10);
-
-	if (dws->tx_dma)
+	if (xfer->tx_buf)
 		dma_ctrl |= SPI_DMA_TDMAE;
-	if (dws->rx_dma)
+	if (xfer->rx_buf)
 		dma_ctrl |= SPI_DMA_RDMAE;
-	dw_writew(dws, DW_SPI_DMACR, dma_ctrl);
+	dw_writel(dws, DW_SPI_DMACR, dma_ctrl);
+
+	/* Set the interrupt mask */
+	spi_umask_intr(dws, SPI_INT_TXOI | SPI_INT_RXUI | SPI_INT_RXOI);
 
-	spi_enable_chip(dws, 1);
+	dws->transfer_handler = dma_transfer;
+
+	return 0;
 }
 
-static int mid_spi_dma_transfer(struct dw_spi *dws, int cs_change)
+static int mid_spi_dma_transfer(struct dw_spi *dws, struct spi_transfer *xfer)
 {
 	struct dma_async_tx_descriptor *txdesc, *rxdesc;
 
-	/* 1. setup DMA related registers */
-	if (cs_change)
-		dw_spi_dma_setup(dws);
+	/* Prepare the TX dma transfer */
+	txdesc = dw_spi_dma_prepare_tx(dws, xfer);
 
-	/* 2. Prepare the TX dma transfer */
-	txdesc = dw_spi_dma_prepare_tx(dws);
-
-	/* 3. Prepare the RX dma transfer */
-	rxdesc = dw_spi_dma_prepare_rx(dws);
+	/* Prepare the RX dma transfer */
+	rxdesc = dw_spi_dma_prepare_rx(dws, xfer);
 
 	/* rx must be started before tx due to spi instinct */
 	if (rxdesc) {
@@ -246,10 +271,25 @@ static int mid_spi_dma_transfer(struct dw_spi *dws, int cs_change)
 	return 0;
 }
 
+static void mid_spi_dma_stop(struct dw_spi *dws)
+{
+	if (test_bit(TX_BUSY, &dws->dma_chan_busy)) {
+		dmaengine_terminate_all(dws->txchan);
+		clear_bit(TX_BUSY, &dws->dma_chan_busy);
+	}
+	if (test_bit(RX_BUSY, &dws->dma_chan_busy)) {
+		dmaengine_terminate_all(dws->rxchan);
+		clear_bit(RX_BUSY, &dws->dma_chan_busy);
+	}
+}
+
 static struct dw_spi_dma_ops mid_dma_ops = {
 	.dma_init	= mid_spi_dma_init,
 	.dma_exit	= mid_spi_dma_exit,
+	.dma_setup	= mid_spi_dma_setup,
+	.can_dma	= mid_spi_can_dma,
 	.dma_transfer	= mid_spi_dma_transfer,
+	.dma_stop	= mid_spi_dma_stop,
 };
 #endif
 
@@ -282,9 +322,8 @@ int dw_spi_mid_init(struct dw_spi *dws)
 	iounmap(clk_reg);
 
 #ifdef CONFIG_SPI_DW_MID_DMA
-	dws->dma_priv = kzalloc(sizeof(struct mid_dma), GFP_KERNEL);
-	if (!dws->dma_priv)
-		return -ENOMEM;
+	dws->dma_tx = &mid_dma_tx;
+	dws->dma_rx = &mid_dma_rx;
 	dws->dma_ops = &mid_dma_ops;
 #endif
 	return 0;
diff --git a/drivers/spi/spi-dw.c b/drivers/spi/spi-dw.c
index 4847afba89f4..8d67d03c71eb 100644
--- a/drivers/spi/spi-dw.c
+++ b/drivers/spi/spi-dw.c
@@ -28,11 +28,6 @@
 #include <linux/debugfs.h>
 #endif
 
-#define START_STATE	((void *)0)
-#define RUNNING_STATE	((void *)1)
-#define DONE_STATE	((void *)2)
-#define ERROR_STATE	((void *)-1)
-
 /* Slave spi_dev related */
 struct chip_data {
 	u16 cr0;
@@ -143,13 +138,26 @@ static inline void dw_spi_debugfs_remove(struct dw_spi *dws)
 }
 #endif /* CONFIG_DEBUG_FS */
 
+static void dw_spi_set_cs(struct spi_device *spi, bool enable)
+{
+	struct dw_spi *dws = spi_master_get_devdata(spi->master);
+	struct chip_data *chip = spi_get_ctldata(spi);
+
+	/* Chip select logic is inverted from spi_set_cs() */
+	if (chip && chip->cs_control)
+		chip->cs_control(!enable);
+
+	if (!enable)
+		dw_writel(dws, DW_SPI_SER, BIT(spi->chip_select));
+}
+
 /* Return the max entries we can fill into tx fifo */
 static inline u32 tx_max(struct dw_spi *dws)
 {
 	u32 tx_left, tx_room, rxtx_gap;
 
 	tx_left = (dws->tx_end - dws->tx) / dws->n_bytes;
-	tx_room = dws->fifo_len - dw_readw(dws, DW_SPI_TXFLR);
+	tx_room = dws->fifo_len - dw_readl(dws, DW_SPI_TXFLR);
 
 	/*
 	 * Another concern is about the tx/rx mismatch, we
@@ -170,7 +178,7 @@ static inline u32 rx_max(struct dw_spi *dws)
 {
 	u32 rx_left = (dws->rx_end - dws->rx) / dws->n_bytes;
 
-	return min_t(u32, rx_left, dw_readw(dws, DW_SPI_RXFLR));
+	return min_t(u32, rx_left, dw_readl(dws, DW_SPI_RXFLR));
 }
 
 static void dw_writer(struct dw_spi *dws)
@@ -186,7 +194,7 @@ static void dw_writer(struct dw_spi *dws)
 			else
 				txw = *(u16 *)(dws->tx);
 		}
-		dw_writew(dws, DW_SPI_DR, txw);
+		dw_writel(dws, DW_SPI_DR, txw);
 		dws->tx += dws->n_bytes;
 	}
 }
@@ -197,7 +205,7 @@ static void dw_reader(struct dw_spi *dws)
 	u16 rxw;
 
 	while (max--) {
-		rxw = dw_readw(dws, DW_SPI_DR);
+		rxw = dw_readl(dws, DW_SPI_DR);
 		/* Care rx only if the transfer's original "rx" is not null */
 		if (dws->rx_end - dws->len) {
 			if (dws->n_bytes == 1)
@@ -209,103 +217,22 @@ static void dw_reader(struct dw_spi *dws)
 	}
 }
 
-static void *next_transfer(struct dw_spi *dws)
-{
-	struct spi_message *msg = dws->cur_msg;
-	struct spi_transfer *trans = dws->cur_transfer;
-
-	/* Move to next transfer */
-	if (trans->transfer_list.next != &msg->transfers) {
-		dws->cur_transfer =
-			list_entry(trans->transfer_list.next,
-					struct spi_transfer,
-					transfer_list);
-		return RUNNING_STATE;
-	}
-
-	return DONE_STATE;
-}
-
-/*
- * Note: first step is the protocol driver prepares
- * a dma-capable memory, and this func just need translate
- * the virt addr to physical
- */
-static int map_dma_buffers(struct dw_spi *dws)
-{
-	if (!dws->cur_msg->is_dma_mapped
-		|| !dws->dma_inited
-		|| !dws->cur_chip->enable_dma
-		|| !dws->dma_ops)
-		return 0;
-
-	if (dws->cur_transfer->tx_dma)
-		dws->tx_dma = dws->cur_transfer->tx_dma;
-
-	if (dws->cur_transfer->rx_dma)
-		dws->rx_dma = dws->cur_transfer->rx_dma;
-
-	return 1;
-}
-
-/* Caller already set message->status; dma and pio irqs are blocked */
-static void giveback(struct dw_spi *dws)
-{
-	struct spi_transfer *last_transfer;
-	struct spi_message *msg;
-
-	msg = dws->cur_msg;
-	dws->cur_msg = NULL;
-	dws->cur_transfer = NULL;
-	dws->prev_chip = dws->cur_chip;
-	dws->cur_chip = NULL;
-	dws->dma_mapped = 0;
-
-	last_transfer = list_last_entry(&msg->transfers, struct spi_transfer,
-					transfer_list);
-
-	if (!last_transfer->cs_change)
-		spi_chip_sel(dws, msg->spi, 0);
-
-	spi_finalize_current_message(dws->master);
-}
-
 static void int_error_stop(struct dw_spi *dws, const char *msg)
 {
-	/* Stop the hw */
-	spi_enable_chip(dws, 0);
+	spi_reset_chip(dws);
 
 	dev_err(&dws->master->dev, "%s\n", msg);
-	dws->cur_msg->state = ERROR_STATE;
-	tasklet_schedule(&dws->pump_transfers);
+	dws->master->cur_msg->status = -EIO;
+	spi_finalize_current_transfer(dws->master);
 }
 
-void dw_spi_xfer_done(struct dw_spi *dws)
-{
-	/* Update total byte transferred return count actual bytes read */
-	dws->cur_msg->actual_length += dws->len;
-
-	/* Move to next transfer */
-	dws->cur_msg->state = next_transfer(dws);
-
-	/* Handle end of message */
-	if (dws->cur_msg->state == DONE_STATE) {
-		dws->cur_msg->status = 0;
-		giveback(dws);
-	} else
-		tasklet_schedule(&dws->pump_transfers);
-}
-EXPORT_SYMBOL_GPL(dw_spi_xfer_done);
-
 static irqreturn_t interrupt_transfer(struct dw_spi *dws)
 {
-	u16 irq_status = dw_readw(dws, DW_SPI_ISR);
+	u16 irq_status = dw_readl(dws, DW_SPI_ISR);
 
 	/* Error handling */
 	if (irq_status & (SPI_INT_TXOI | SPI_INT_RXOI | SPI_INT_RXUI)) {
-		dw_readw(dws, DW_SPI_TXOICR);
-		dw_readw(dws, DW_SPI_RXOICR);
-		dw_readw(dws, DW_SPI_RXUICR);
+		dw_readl(dws, DW_SPI_ICR);
 		int_error_stop(dws, "interrupt_transfer: fifo overrun/underrun");
 		return IRQ_HANDLED;
 	}
@@ -313,7 +240,7 @@ static irqreturn_t interrupt_transfer(struct dw_spi *dws)
 	dw_reader(dws);
 	if (dws->rx_end == dws->rx) {
 		spi_mask_intr(dws, SPI_INT_TXEI);
-		dw_spi_xfer_done(dws);
+		spi_finalize_current_transfer(dws->master);
 		return IRQ_HANDLED;
 	}
 	if (irq_status & SPI_INT_TXEI) {
@@ -328,13 +255,14 @@ static irqreturn_t interrupt_transfer(struct dw_spi *dws)
 
 static irqreturn_t dw_spi_irq(int irq, void *dev_id)
 {
-	struct dw_spi *dws = dev_id;
-	u16 irq_status = dw_readw(dws, DW_SPI_ISR) & 0x3f;
+	struct spi_master *master = dev_id;
+	struct dw_spi *dws = spi_master_get_devdata(master);
+	u16 irq_status = dw_readl(dws, DW_SPI_ISR) & 0x3f;
 
 	if (!irq_status)
 		return IRQ_NONE;
 
-	if (!dws->cur_msg) {
+	if (!master->cur_msg) {
 		spi_mask_intr(dws, SPI_INT_TXEI);
 		return IRQ_HANDLED;
 	}
@@ -343,7 +271,7 @@ static irqreturn_t dw_spi_irq(int irq, void *dev_id)
 }
 
 /* Must be called inside pump_transfers() */
-static void poll_transfer(struct dw_spi *dws)
+static int poll_transfer(struct dw_spi *dws)
 {
 	do {
 		dw_writer(dws);
@@ -351,64 +279,32 @@ static void poll_transfer(struct dw_spi *dws)
 		cpu_relax();
 	} while (dws->rx_end > dws->rx);
 
-	dw_spi_xfer_done(dws);
+	return 0;
 }
 
-static void pump_transfers(unsigned long data)
+static int dw_spi_transfer_one(struct spi_master *master,
+		struct spi_device *spi, struct spi_transfer *transfer)
 {
-	struct dw_spi *dws = (struct dw_spi *)data;
-	struct spi_message *message = NULL;
-	struct spi_transfer *transfer = NULL;
-	struct spi_transfer *previous = NULL;
-	struct spi_device *spi = NULL;
-	struct chip_data *chip = NULL;
-	u8 bits = 0;
+	struct dw_spi *dws = spi_master_get_devdata(master);
+	struct chip_data *chip = spi_get_ctldata(spi);
 	u8 imask = 0;
-	u8 cs_change = 0;
-	u16 txint_level = 0;
+	u16 txlevel = 0;
 	u16 clk_div = 0;
 	u32 speed = 0;
 	u32 cr0 = 0;
+	int ret;
 
-	/* Get current state information */
-	message = dws->cur_msg;
-	transfer = dws->cur_transfer;
-	chip = dws->cur_chip;
-	spi = message->spi;
-
-	if (message->state == ERROR_STATE) {
-		message->status = -EIO;
-		goto early_exit;
-	}
-
-	/* Handle end of message */
-	if (message->state == DONE_STATE) {
-		message->status = 0;
-		goto early_exit;
-	}
-
-	/* Delay if requested at end of transfer */
-	if (message->state == RUNNING_STATE) {
-		previous = list_entry(transfer->transfer_list.prev,
-					struct spi_transfer,
-					transfer_list);
-		if (previous->delay_usecs)
-			udelay(previous->delay_usecs);
-	}
-
+	dws->dma_mapped = 0;
 	dws->n_bytes = chip->n_bytes;
 	dws->dma_width = chip->dma_width;
-	dws->cs_control = chip->cs_control;
 
-	dws->rx_dma = transfer->rx_dma;
-	dws->tx_dma = transfer->tx_dma;
 	dws->tx = (void *)transfer->tx_buf;
 	dws->tx_end = dws->tx + transfer->len;
 	dws->rx = transfer->rx_buf;
 	dws->rx_end = dws->rx + transfer->len;
-	dws->len = dws->cur_transfer->len;
-	if (chip != dws->prev_chip)
-		cs_change = 1;
+	dws->len = transfer->len;
+
+	spi_enable_chip(dws, 0);
 
 	cr0 = chip->cr0;
 
@@ -416,32 +312,37 @@ static void pump_transfers(unsigned long data)
 	if (transfer->speed_hz) {
 		speed = chip->speed_hz;
 
-		if ((transfer->speed_hz != speed) || (!chip->clk_div)) {
+		if ((transfer->speed_hz != speed) || !chip->clk_div) {
 			speed = transfer->speed_hz;
 
 			/* clk_div doesn't support odd number */
-			clk_div = dws->max_freq / speed;
-			clk_div = (clk_div + 1) & 0xfffe;
+			clk_div = (dws->max_freq / speed + 1) & 0xfffe;
 
 			chip->speed_hz = speed;
 			chip->clk_div = clk_div;
+
+			spi_set_clk(dws, chip->clk_div);
 		}
 	}
 	if (transfer->bits_per_word) {
-		bits = transfer->bits_per_word;
-		dws->n_bytes = dws->dma_width = bits >> 3;
-		cr0 = (bits - 1)
+		if (transfer->bits_per_word == 8) {
+			dws->n_bytes = 1;
+			dws->dma_width = 1;
+		} else if (transfer->bits_per_word == 16) {
+			dws->n_bytes = 2;
+			dws->dma_width = 2;
+		}
+		cr0 = (transfer->bits_per_word - 1)
 			| (chip->type << SPI_FRF_OFFSET)
 			| (spi->mode << SPI_MODE_OFFSET)
 			| (chip->tmode << SPI_TMOD_OFFSET);
 	}
-	message->state = RUNNING_STATE;
 
 	/*
 	 * Adjust transfer mode if necessary. Requires platform dependent
 	 * chipselect mechanism.
 	 */
-	if (dws->cs_control) {
+	if (chip->cs_control) {
 		if (dws->rx && dws->tx)
 			chip->tmode = SPI_TMOD_TR;
 		else if (dws->rx)
@@ -453,80 +354,60 @@ static void pump_transfers(unsigned long data)
 		cr0 |= (chip->tmode << SPI_TMOD_OFFSET);
 	}
 
+	dw_writel(dws, DW_SPI_CTRL0, cr0);
+
 	/* Check if current transfer is a DMA transaction */
-	dws->dma_mapped = map_dma_buffers(dws);
+	if (master->can_dma && master->can_dma(master, spi, transfer))
+		dws->dma_mapped = master->cur_msg_mapped;
+
+	/* For poll mode just disable all interrupts */
+	spi_mask_intr(dws, 0xff);
 
 	/*
 	 * Interrupt mode
 	 * we only need set the TXEI IRQ, as TX/RX always happen syncronizely
 	 */
-	if (!dws->dma_mapped && !chip->poll_mode) {
-		int templen = dws->len / dws->n_bytes;
-
-		txint_level = dws->fifo_len / 2;
-		txint_level = (templen > txint_level) ? txint_level : templen;
+	if (dws->dma_mapped) {
+		ret = dws->dma_ops->dma_setup(dws, transfer);
+		if (ret < 0) {
+			spi_enable_chip(dws, 1);
+			return ret;
+		}
+	} else if (!chip->poll_mode) {
+		txlevel = min_t(u16, dws->fifo_len / 2, dws->len / dws->n_bytes);
+		dw_writel(dws, DW_SPI_TXFLTR, txlevel);
 
+		/* Set the interrupt mask */
 		imask |= SPI_INT_TXEI | SPI_INT_TXOI |
 			 SPI_INT_RXUI | SPI_INT_RXOI;
+		spi_umask_intr(dws, imask);
+
 		dws->transfer_handler = interrupt_transfer;
 	}
 
-	/*
-	 * Reprogram registers only if
-	 *	1. chip select changes
-	 *	2. clk_div is changed
-	 *	3. control value changes
-	 */
-	if (dw_readw(dws, DW_SPI_CTRL0) != cr0 || cs_change || clk_div || imask) {
-		spi_enable_chip(dws, 0);
-
-		if (dw_readw(dws, DW_SPI_CTRL0) != cr0)
-			dw_writew(dws, DW_SPI_CTRL0, cr0);
-
-		spi_set_clk(dws, clk_div ? clk_div : chip->clk_div);
-		spi_chip_sel(dws, spi, 1);
-
-		/* Set the interrupt mask, for poll mode just disable all int */
-		spi_mask_intr(dws, 0xff);
-		if (imask)
-			spi_umask_intr(dws, imask);
-		if (txint_level)
-			dw_writew(dws, DW_SPI_TXFLTR, txint_level);
+	spi_enable_chip(dws, 1);
 
-		spi_enable_chip(dws, 1);
-		if (cs_change)
-			dws->prev_chip = chip;
+	if (dws->dma_mapped) {
+		ret = dws->dma_ops->dma_transfer(dws, transfer);
+		if (ret < 0)
+			return ret;
 	}
 
-	if (dws->dma_mapped)
-		dws->dma_ops->dma_transfer(dws, cs_change);
-
 	if (chip->poll_mode)
-		poll_transfer(dws);
-
-	return;
+		return poll_transfer(dws);
 
-early_exit:
-	giveback(dws);
+	return 1;
 }
 
-static int dw_spi_transfer_one_message(struct spi_master *master,
+static void dw_spi_handle_err(struct spi_master *master,
 		struct spi_message *msg)
 {
 	struct dw_spi *dws = spi_master_get_devdata(master);
 
-	dws->cur_msg = msg;
-	/* Initial message state */
-	dws->cur_msg->state = START_STATE;
-	dws->cur_transfer = list_entry(dws->cur_msg->transfers.next,
-						struct spi_transfer,
-						transfer_list);
-	dws->cur_chip = spi_get_ctldata(dws->cur_msg->spi);
-
-	/* Launch transfers */
-	tasklet_schedule(&dws->pump_transfers);
+	if (dws->dma_mapped)
+		dws->dma_ops->dma_stop(dws);
 
-	return 0;
+	spi_reset_chip(dws);
 }
 
 /* This may be called twice for each spi dev */
@@ -561,8 +442,6 @@ static int dw_spi_setup(struct spi_device *spi)
 
 		chip->rx_threshold = 0;
 		chip->tx_threshold = 0;
-
-		chip->enable_dma = chip_info->enable_dma;
 	}
 
 	if (spi->bits_per_word == 8) {
@@ -610,9 +489,7 @@ static void dw_spi_cleanup(struct spi_device *spi)
 /* Restart the controller, disable all interrupts, clean rx fifo */
 static void spi_hw_init(struct device *dev, struct dw_spi *dws)
 {
-	spi_enable_chip(dws, 0);
-	spi_mask_intr(dws, 0xff);
-	spi_enable_chip(dws, 1);
+	spi_reset_chip(dws);
 
 	/*
 	 * Try to detect the FIFO depth if not set by interface driver,
@@ -622,11 +499,11 @@ static void spi_hw_init(struct device *dev, struct dw_spi *dws)
 		u32 fifo;
 
 		for (fifo = 1; fifo < 256; fifo++) {
-			dw_writew(dws, DW_SPI_TXFLTR, fifo);
-			if (fifo != dw_readw(dws, DW_SPI_TXFLTR))
+			dw_writel(dws, DW_SPI_TXFLTR, fifo);
+			if (fifo != dw_readl(dws, DW_SPI_TXFLTR))
 				break;
 		}
-		dw_writew(dws, DW_SPI_TXFLTR, 0);
+		dw_writel(dws, DW_SPI_TXFLTR, 0);
 
 		dws->fifo_len = (fifo == 1) ? 0 : fifo;
 		dev_dbg(dev, "Detected FIFO size: %u bytes\n", dws->fifo_len);
@@ -646,13 +523,12 @@ int dw_spi_add_host(struct device *dev, struct dw_spi *dws)
 
 	dws->master = master;
 	dws->type = SSI_MOTO_SPI;
-	dws->prev_chip = NULL;
 	dws->dma_inited = 0;
 	dws->dma_addr = (dma_addr_t)(dws->paddr + 0x60);
 	snprintf(dws->name, sizeof(dws->name), "dw_spi%d", dws->bus_num);
 
 	ret = devm_request_irq(dev, dws->irq, dw_spi_irq, IRQF_SHARED,
-			dws->name, dws);
+			dws->name, master);
 	if (ret < 0) {
 		dev_err(&master->dev, "can not get IRQ\n");
 		goto err_free_master;
@@ -664,7 +540,9 @@ int dw_spi_add_host(struct device *dev, struct dw_spi *dws)
 	master->num_chipselect = dws->num_cs;
 	master->setup = dw_spi_setup;
 	master->cleanup = dw_spi_cleanup;
-	master->transfer_one_message = dw_spi_transfer_one_message;
+	master->set_cs = dw_spi_set_cs;
+	master->transfer_one = dw_spi_transfer_one;
+	master->handle_err = dw_spi_handle_err;
 	master->max_speed_hz = dws->max_freq;
 	master->dev.of_node = dev->of_node;
 
@@ -676,11 +554,11 @@ int dw_spi_add_host(struct device *dev, struct dw_spi *dws)
 		if (ret) {
 			dev_warn(dev, "DMA init failed\n");
 			dws->dma_inited = 0;
+		} else {
+			master->can_dma = dws->dma_ops->can_dma;
 		}
 	}
 
-	tasklet_init(&dws->pump_transfers, pump_transfers, (unsigned long)dws);
-
 	spi_master_set_devdata(master, dws);
 	ret = devm_spi_register_master(dev, master);
 	if (ret) {
diff --git a/drivers/spi/spi-dw.h b/drivers/spi/spi-dw.h
index 3d32be68c142..6c91391c1a4f 100644
--- a/drivers/spi/spi-dw.h
+++ b/drivers/spi/spi-dw.h
@@ -91,12 +91,15 @@ struct dw_spi;
 struct dw_spi_dma_ops {
 	int (*dma_init)(struct dw_spi *dws);
 	void (*dma_exit)(struct dw_spi *dws);
-	int (*dma_transfer)(struct dw_spi *dws, int cs_change);
+	int (*dma_setup)(struct dw_spi *dws, struct spi_transfer *xfer);
+	bool (*can_dma)(struct spi_master *master, struct spi_device *spi,
+			struct spi_transfer *xfer);
+	int (*dma_transfer)(struct dw_spi *dws, struct spi_transfer *xfer);
+	void (*dma_stop)(struct dw_spi *dws);
 };
 
 struct dw_spi {
 	struct spi_master	*master;
-	struct spi_device	*cur_dev;
 	enum dw_ssi_type	type;
 	char			name[16];
 
@@ -109,41 +112,26 @@ struct dw_spi {
 	u16			bus_num;
 	u16			num_cs;		/* supported slave numbers */
 
-	/* Message Transfer pump */
-	struct tasklet_struct	pump_transfers;
-
 	/* Current message transfer state info */
-	struct spi_message	*cur_msg;
-	struct spi_transfer	*cur_transfer;
-	struct chip_data	*cur_chip;
-	struct chip_data	*prev_chip;
 	size_t			len;
 	void			*tx;
 	void			*tx_end;
 	void			*rx;
 	void			*rx_end;
 	int			dma_mapped;
-	dma_addr_t		rx_dma;
-	dma_addr_t		tx_dma;
-	size_t			rx_map_len;
-	size_t			tx_map_len;
 	u8			n_bytes;	/* current is a 1/2 bytes op */
-	u8			max_bits_per_word;	/* maxim is 16b */
 	u32			dma_width;
 	irqreturn_t		(*transfer_handler)(struct dw_spi *dws);
-	void			(*cs_control)(u32 command);
 
-	/* Dma info */
+	/* DMA info */
 	int			dma_inited;
 	struct dma_chan		*txchan;
-	struct scatterlist	tx_sgl;
 	struct dma_chan		*rxchan;
-	struct scatterlist	rx_sgl;
 	unsigned long		dma_chan_busy;
-	struct device		*dma_dev;
 	dma_addr_t		dma_addr; /* phy address of the Data register */
 	struct dw_spi_dma_ops	*dma_ops;
-	void			*dma_priv; /* platform relate info */
+	void			*dma_tx;
+	void			*dma_rx;
 
 	/* Bus interface info */
 	void			*priv;
@@ -162,16 +150,6 @@ static inline void dw_writel(struct dw_spi *dws, u32 offset, u32 val)
 	__raw_writel(val, dws->regs + offset);
 }
 
-static inline u16 dw_readw(struct dw_spi *dws, u32 offset)
-{
-	return __raw_readw(dws->regs + offset);
-}
-
-static inline void dw_writew(struct dw_spi *dws, u32 offset, u16 val)
-{
-	__raw_writew(val, dws->regs + offset);
-}
-
 static inline void spi_enable_chip(struct dw_spi *dws, int enable)
 {
 	dw_writel(dws, DW_SPI_SSIENR, (enable ? 1 : 0));
@@ -182,22 +160,6 @@ static inline void spi_set_clk(struct dw_spi *dws, u16 div)
 	dw_writel(dws, DW_SPI_BAUDR, div);
 }
 
-static inline void spi_chip_sel(struct dw_spi *dws, struct spi_device *spi,
-		int active)
-{
-	u16 cs = spi->chip_select;
-	int gpio_val = active ? (spi->mode & SPI_CS_HIGH) :
-		!(spi->mode & SPI_CS_HIGH);
-
-	if (dws->cs_control)
-		dws->cs_control(active);
-	if (gpio_is_valid(spi->cs_gpio))
-		gpio_set_value(spi->cs_gpio, gpio_val);
-
-	if (active)
-		dw_writel(dws, DW_SPI_SER, 1 << cs);
-}
-
 /* Disable IRQ bits */
 static inline void spi_mask_intr(struct dw_spi *dws, u32 mask)
 {
@@ -217,15 +179,26 @@ static inline void spi_umask_intr(struct dw_spi *dws, u32 mask)
 }
 
 /*
+ * This does disable the SPI controller, interrupts, and re-enable the
+ * controller back. Transmit and receive FIFO buffers are cleared when the
+ * device is disabled.
+ */
+static inline void spi_reset_chip(struct dw_spi *dws)
+{
+	spi_enable_chip(dws, 0);
+	spi_mask_intr(dws, 0xff);
+	spi_enable_chip(dws, 1);
+}
+
+/*
  * Each SPI slave device to work with dw_api controller should
- * has such a structure claiming its working mode (PIO/DMA etc),
+ * has such a structure claiming its working mode (poll or PIO/DMA),
  * which can be save in the "controller_data" member of the
  * struct spi_device.
  */
 struct dw_spi_chip {
 	u8 poll_mode;	/* 1 for controller polling mode */
 	u8 type;	/* SPI/SSP/MicroWire */
-	u8 enable_dma;
 	void (*cs_control)(u32 command);
 };
 
@@ -233,7 +206,6 @@ extern int dw_spi_add_host(struct device *dev, struct dw_spi *dws);
 extern void dw_spi_remove_host(struct dw_spi *dws);
 extern int dw_spi_suspend_host(struct dw_spi *dws);
 extern int dw_spi_resume_host(struct dw_spi *dws);
-extern void dw_spi_xfer_done(struct dw_spi *dws);
 
 /* platform related setup */
 extern int dw_spi_mid_init(struct dw_spi *dws); /* Intel MID platforms */
diff --git a/drivers/spi/spi-fsl-cpm.c b/drivers/spi/spi-fsl-cpm.c
index 9c46a3058743..896add8cfd3b 100644
--- a/drivers/spi/spi-fsl-cpm.c
+++ b/drivers/spi/spi-fsl-cpm.c
@@ -24,6 +24,7 @@
 #include <linux/of_address.h>
 #include <linux/spi/spi.h>
 #include <linux/types.h>
+#include <linux/platform_device.h>
 
 #include "spi-fsl-cpm.h"
 #include "spi-fsl-lib.h"
@@ -269,17 +270,6 @@ static unsigned long fsl_spi_cpm_get_pram(struct mpc8xxx_spi *mspi)
 	if (mspi->flags & SPI_CPM2) {
 		pram_ofs = cpm_muram_alloc(SPI_PRAM_SIZE, 64);
 		out_be16(spi_base, pram_ofs);
-	} else {
-		struct spi_pram __iomem *pram = spi_base;
-		u16 rpbase = in_be16(&pram->rpbase);
-
-		/* Microcode relocation patch applied? */
-		if (rpbase) {
-			pram_ofs = rpbase;
-		} else {
-			pram_ofs = cpm_muram_alloc(SPI_PRAM_SIZE, 64);
-			out_be16(spi_base, pram_ofs);
-		}
 	}
 
 	iounmap(spi_base);
@@ -292,7 +282,6 @@ int fsl_spi_cpm_init(struct mpc8xxx_spi *mspi)
 	struct device_node *np = dev->of_node;
 	const u32 *iprop;
 	int size;
-	unsigned long pram_ofs;
 	unsigned long bds_ofs;
 
 	if (!(mspi->flags & SPI_CPM_MODE))
@@ -319,8 +308,26 @@ int fsl_spi_cpm_init(struct mpc8xxx_spi *mspi)
 		}
 	}
 
-	pram_ofs = fsl_spi_cpm_get_pram(mspi);
-	if (IS_ERR_VALUE(pram_ofs)) {
+	if (mspi->flags & SPI_CPM1) {
+		struct resource *res;
+		void *pram;
+
+		res = platform_get_resource(to_platform_device(dev),
+					    IORESOURCE_MEM, 1);
+		pram = devm_ioremap_resource(dev, res);
+		if (IS_ERR(pram))
+			mspi->pram = NULL;
+		else
+			mspi->pram = pram;
+	} else {
+		unsigned long pram_ofs = fsl_spi_cpm_get_pram(mspi);
+
+		if (IS_ERR_VALUE(pram_ofs))
+			mspi->pram = NULL;
+		else
+			mspi->pram = cpm_muram_addr(pram_ofs);
+	}
+	if (mspi->pram == NULL) {
 		dev_err(dev, "can't allocate spi parameter ram\n");
 		goto err_pram;
 	}
@@ -346,8 +353,6 @@ int fsl_spi_cpm_init(struct mpc8xxx_spi *mspi)
 		goto err_dummy_rx;
 	}
 
-	mspi->pram = cpm_muram_addr(pram_ofs);
-
 	mspi->tx_bd = cpm_muram_addr(bds_ofs);
 	mspi->rx_bd = cpm_muram_addr(bds_ofs + sizeof(*mspi->tx_bd));
 
@@ -375,7 +380,8 @@ err_dummy_rx:
 err_dummy_tx:
 	cpm_muram_free(bds_ofs);
 err_bds:
-	cpm_muram_free(pram_ofs);
+	if (!(mspi->flags & SPI_CPM1))
+		cpm_muram_free(cpm_muram_offset(mspi->pram));
 err_pram:
 	fsl_spi_free_dummy_rx();
 	return -ENOMEM;
diff --git a/drivers/spi/spi-fsl-dspi.c b/drivers/spi/spi-fsl-dspi.c
index d1a39249704a..86bcdd68c1fe 100644
--- a/drivers/spi/spi-fsl-dspi.c
+++ b/drivers/spi/spi-fsl-dspi.c
@@ -20,15 +20,18 @@
 #include <linux/interrupt.h>
 #include <linux/io.h>
 #include <linux/kernel.h>
+#include <linux/math64.h>
 #include <linux/module.h>
 #include <linux/of.h>
 #include <linux/of_device.h>
+#include <linux/pinctrl/consumer.h>
 #include <linux/platform_device.h>
 #include <linux/pm_runtime.h>
 #include <linux/regmap.h>
 #include <linux/sched.h>
 #include <linux/spi/spi.h>
 #include <linux/spi/spi_bitbang.h>
+#include <linux/time.h>
 
 #define DRIVER_NAME "fsl-dspi"
 
@@ -45,13 +48,14 @@
 #define SPI_MCR_CLR_RXF	(1 << 10)
 
 #define SPI_TCR			0x08
+#define SPI_TCR_GET_TCNT(x)	(((x) & 0xffff0000) >> 16)
 
 #define SPI_CTAR(x)		(0x0c + (((x) & 0x3) * 4))
 #define SPI_CTAR_FMSZ(x)	(((x) & 0x0000000f) << 27)
 #define SPI_CTAR_CPOL(x)	((x) << 26)
 #define SPI_CTAR_CPHA(x)	((x) << 25)
 #define SPI_CTAR_LSBFE(x)	((x) << 24)
-#define SPI_CTAR_PCSSCR(x)	(((x) & 0x00000003) << 22)
+#define SPI_CTAR_PCSSCK(x)	(((x) & 0x00000003) << 22)
 #define SPI_CTAR_PASC(x)	(((x) & 0x00000003) << 20)
 #define SPI_CTAR_PDT(x)	(((x) & 0x00000003) << 18)
 #define SPI_CTAR_PBR(x)	(((x) & 0x00000003) << 16)
@@ -59,14 +63,17 @@
 #define SPI_CTAR_ASC(x)	(((x) & 0x0000000f) << 8)
 #define SPI_CTAR_DT(x)		(((x) & 0x0000000f) << 4)
 #define SPI_CTAR_BR(x)		((x) & 0x0000000f)
+#define SPI_CTAR_SCALE_BITS	0xf
 
 #define SPI_CTAR0_SLAVE	0x0c
 
 #define SPI_SR			0x2c
 #define SPI_SR_EOQF		0x10000000
+#define SPI_SR_TCFQF		0x80000000
 
 #define SPI_RSER		0x30
 #define SPI_RSER_EOQFE		0x10000000
+#define SPI_RSER_TCFQE		0x80000000
 
 #define SPI_PUSHR		0x34
 #define SPI_PUSHR_CONT		(1 << 31)
@@ -99,12 +106,35 @@
 #define SPI_CS_ASSERT		0x02
 #define SPI_CS_DROP		0x04
 
+#define SPI_TCR_TCNT_MAX	0x10000
+
 struct chip_data {
 	u32 mcr_val;
 	u32 ctar_val;
 	u16 void_write_data;
 };
 
+enum dspi_trans_mode {
+	DSPI_EOQ_MODE = 0,
+	DSPI_TCFQ_MODE,
+};
+
+struct fsl_dspi_devtype_data {
+	enum dspi_trans_mode trans_mode;
+};
+
+static const struct fsl_dspi_devtype_data vf610_data = {
+	.trans_mode = DSPI_EOQ_MODE,
+};
+
+static const struct fsl_dspi_devtype_data ls1021a_v1_data = {
+	.trans_mode = DSPI_TCFQ_MODE,
+};
+
+static const struct fsl_dspi_devtype_data ls2085a_data = {
+	.trans_mode = DSPI_TCFQ_MODE,
+};
+
 struct fsl_dspi {
 	struct spi_master	*master;
 	struct platform_device	*pdev;
@@ -125,9 +155,12 @@ struct fsl_dspi {
 	u8			cs;
 	u16			void_write_data;
 	u32			cs_change;
+	struct fsl_dspi_devtype_data *devtype_data;
 
 	wait_queue_head_t	waitq;
 	u32			waitflags;
+
+	u32			spi_tcnt;
 };
 
 static inline int is_double_byte_mode(struct fsl_dspi *dspi)
@@ -148,82 +181,122 @@ static void hz_to_spi_baud(char *pbr, char *br, int speed_hz,
 		16,	32,	64,	128,
 		256,	512,	1024,	2048,
 		4096,	8192,	16384,	32768 };
-	int temp, i = 0, j = 0;
+	int scale_needed, scale, minscale = INT_MAX;
+	int i, j;
+
+	scale_needed = clkrate / speed_hz;
+	if (clkrate % speed_hz)
+		scale_needed++;
+
+	for (i = 0; i < ARRAY_SIZE(brs); i++)
+		for (j = 0; j < ARRAY_SIZE(pbr_tbl); j++) {
+			scale = brs[i] * pbr_tbl[j];
+			if (scale >= scale_needed) {
+				if (scale < minscale) {
+					minscale = scale;
+					*br = i;
+					*pbr = j;
+				}
+				break;
+			}
+		}
 
-	temp = clkrate / 2 / speed_hz;
+	if (minscale == INT_MAX) {
+		pr_warn("Can not find valid baud rate,speed_hz is %d,clkrate is %ld, we use the max prescaler value.\n",
+			speed_hz, clkrate);
+		*pbr = ARRAY_SIZE(pbr_tbl) - 1;
+		*br =  ARRAY_SIZE(brs) - 1;
+	}
+}
 
-	for (i = 0; i < ARRAY_SIZE(pbr_tbl); i++)
-		for (j = 0; j < ARRAY_SIZE(brs); j++) {
-			if (pbr_tbl[i] * brs[j] >= temp) {
-				*pbr = i;
-				*br = j;
-				return;
+static void ns_delay_scale(char *psc, char *sc, int delay_ns,
+		unsigned long clkrate)
+{
+	int pscale_tbl[4] = {1, 3, 5, 7};
+	int scale_needed, scale, minscale = INT_MAX;
+	int i, j;
+	u32 remainder;
+
+	scale_needed = div_u64_rem((u64)delay_ns * clkrate, NSEC_PER_SEC,
+			&remainder);
+	if (remainder)
+		scale_needed++;
+
+	for (i = 0; i < ARRAY_SIZE(pscale_tbl); i++)
+		for (j = 0; j <= SPI_CTAR_SCALE_BITS; j++) {
+			scale = pscale_tbl[i] * (2 << j);
+			if (scale >= scale_needed) {
+				if (scale < minscale) {
+					minscale = scale;
+					*psc = i;
+					*sc = j;
+				}
+				break;
 			}
 		}
 
-	pr_warn("Can not find valid baud rate,speed_hz is %d,clkrate is %ld\
-		,we use the max prescaler value.\n", speed_hz, clkrate);
-	*pbr = ARRAY_SIZE(pbr_tbl) - 1;
-	*br =  ARRAY_SIZE(brs) - 1;
+	if (minscale == INT_MAX) {
+		pr_warn("Cannot find correct scale values for %dns delay at clkrate %ld, using max prescaler value",
+			delay_ns, clkrate);
+		*psc = ARRAY_SIZE(pscale_tbl) - 1;
+		*sc = SPI_CTAR_SCALE_BITS;
+	}
 }
 
-static int dspi_transfer_write(struct fsl_dspi *dspi)
+static u32 dspi_data_to_pushr(struct fsl_dspi *dspi, int tx_word)
 {
-	int tx_count = 0;
-	int tx_word;
 	u16 d16;
-	u8  d8;
-	u32 dspi_pushr = 0;
-	int first = 1;
 
-	tx_word = is_double_byte_mode(dspi);
+	if (!(dspi->dataflags & TRAN_STATE_TX_VOID))
+		d16 = tx_word ? *(u16 *)dspi->tx : *(u8 *)dspi->tx;
+	else
+		d16 = dspi->void_write_data;
 
-	/* If we are in word mode, but only have a single byte to transfer
-	 * then switch to byte mode temporarily.  Will switch back at the
-	 * end of the transfer.
-	 */
-	if (tx_word && (dspi->len == 1)) {
-		dspi->dataflags |= TRAN_STATE_WORD_ODD_NUM;
-		regmap_update_bits(dspi->regmap, SPI_CTAR(dspi->cs),
-				SPI_FRAME_BITS_MASK, SPI_FRAME_BITS(8));
-		tx_word = 0;
-	}
+	dspi->tx += tx_word + 1;
+	dspi->len -= tx_word + 1;
 
-	while (dspi->len && (tx_count < DSPI_FIFO_SIZE)) {
-		if (tx_word) {
-			if (dspi->len == 1)
-				break;
+	return	SPI_PUSHR_TXDATA(d16) |
+		SPI_PUSHR_PCS(dspi->cs) |
+		SPI_PUSHR_CTAS(dspi->cs) |
+		SPI_PUSHR_CONT;
+}
 
-			if (!(dspi->dataflags & TRAN_STATE_TX_VOID)) {
-				d16 = *(u16 *)dspi->tx;
-				dspi->tx += 2;
-			} else {
-				d16 = dspi->void_write_data;
-			}
+static void dspi_data_from_popr(struct fsl_dspi *dspi, int rx_word)
+{
+	u16 d;
+	unsigned int val;
 
-			dspi_pushr = SPI_PUSHR_TXDATA(d16) |
-				SPI_PUSHR_PCS(dspi->cs) |
-				SPI_PUSHR_CTAS(dspi->cs) |
-				SPI_PUSHR_CONT;
+	regmap_read(dspi->regmap, SPI_POPR, &val);
+	d = SPI_POPR_RXDATA(val);
 
-			dspi->len -= 2;
-		} else {
-			if (!(dspi->dataflags & TRAN_STATE_TX_VOID)) {
+	if (!(dspi->dataflags & TRAN_STATE_RX_VOID))
+		rx_word ? (*(u16 *)dspi->rx = d) : (*(u8 *)dspi->rx = d);
 
-				d8 = *(u8 *)dspi->tx;
-				dspi->tx++;
-			} else {
-				d8 = (u8)dspi->void_write_data;
-			}
+	dspi->rx += rx_word + 1;
+}
 
-			dspi_pushr = SPI_PUSHR_TXDATA(d8) |
-				SPI_PUSHR_PCS(dspi->cs) |
-				SPI_PUSHR_CTAS(dspi->cs) |
-				SPI_PUSHR_CONT;
+static int dspi_eoq_write(struct fsl_dspi *dspi)
+{
+	int tx_count = 0;
+	int tx_word;
+	u32 dspi_pushr = 0;
+
+	tx_word = is_double_byte_mode(dspi);
 
-			dspi->len--;
+	while (dspi->len && (tx_count < DSPI_FIFO_SIZE)) {
+		/* If we are in word mode, only have a single byte to transfer
+		 * switch to byte mode temporarily.  Will switch back at the
+		 * end of the transfer.
+		 */
+		if (tx_word && (dspi->len == 1)) {
+			dspi->dataflags |= TRAN_STATE_WORD_ODD_NUM;
+			regmap_update_bits(dspi->regmap, SPI_CTAR(dspi->cs),
+					SPI_FRAME_BITS_MASK, SPI_FRAME_BITS(8));
+			tx_word = 0;
 		}
 
+		dspi_pushr = dspi_data_to_pushr(dspi, tx_word);
+
 		if (dspi->len == 0 || tx_count == DSPI_FIFO_SIZE - 1) {
 			/* last transfer in the transfer */
 			dspi_pushr |= SPI_PUSHR_EOQ;
@@ -232,11 +305,6 @@ static int dspi_transfer_write(struct fsl_dspi *dspi)
 		} else if (tx_word && (dspi->len == 1))
 			dspi_pushr |= SPI_PUSHR_EOQ;
 
-		if (first) {
-			first = 0;
-			dspi_pushr |= SPI_PUSHR_CTCNT; /* clear counter */
-		}
-
 		regmap_write(dspi->regmap, SPI_PUSHR, dspi_pushr);
 
 		tx_count++;
@@ -245,40 +313,55 @@ static int dspi_transfer_write(struct fsl_dspi *dspi)
 	return tx_count * (tx_word + 1);
 }
 
-static int dspi_transfer_read(struct fsl_dspi *dspi)
+static int dspi_eoq_read(struct fsl_dspi *dspi)
 {
 	int rx_count = 0;
 	int rx_word = is_double_byte_mode(dspi);
-	u16 d;
 
 	while ((dspi->rx < dspi->rx_end)
 			&& (rx_count < DSPI_FIFO_SIZE)) {
-		if (rx_word) {
-			unsigned int val;
+		if (rx_word && (dspi->rx_end - dspi->rx) == 1)
+			rx_word = 0;
 
-			if ((dspi->rx_end - dspi->rx) == 1)
-				break;
+		dspi_data_from_popr(dspi, rx_word);
+		rx_count++;
+	}
 
-			regmap_read(dspi->regmap, SPI_POPR, &val);
-			d = SPI_POPR_RXDATA(val);
+	return rx_count;
+}
 
-			if (!(dspi->dataflags & TRAN_STATE_RX_VOID))
-				*(u16 *)dspi->rx = d;
-			dspi->rx += 2;
+static int dspi_tcfq_write(struct fsl_dspi *dspi)
+{
+	int tx_word;
+	u32 dspi_pushr = 0;
 
-		} else {
-			unsigned int val;
+	tx_word = is_double_byte_mode(dspi);
 
-			regmap_read(dspi->regmap, SPI_POPR, &val);
-			d = SPI_POPR_RXDATA(val);
-			if (!(dspi->dataflags & TRAN_STATE_RX_VOID))
-				*(u8 *)dspi->rx = d;
-			dspi->rx++;
-		}
-		rx_count++;
+	if (tx_word && (dspi->len == 1)) {
+		dspi->dataflags |= TRAN_STATE_WORD_ODD_NUM;
+		regmap_update_bits(dspi->regmap, SPI_CTAR(dspi->cs),
+				SPI_FRAME_BITS_MASK, SPI_FRAME_BITS(8));
+		tx_word = 0;
 	}
 
-	return rx_count;
+	dspi_pushr = dspi_data_to_pushr(dspi, tx_word);
+
+	if ((dspi->cs_change) && (!dspi->len))
+		dspi_pushr &= ~SPI_PUSHR_CONT;
+
+	regmap_write(dspi->regmap, SPI_PUSHR, dspi_pushr);
+
+	return tx_word + 1;
+}
+
+static void dspi_tcfq_read(struct fsl_dspi *dspi)
+{
+	int rx_word = is_double_byte_mode(dspi);
+
+	if (rx_word && (dspi->rx_end - dspi->rx) == 1)
+		rx_word = 0;
+
+	dspi_data_from_popr(dspi, rx_word);
 }
 
 static int dspi_transfer_one_message(struct spi_master *master,
@@ -288,6 +371,12 @@ static int dspi_transfer_one_message(struct spi_master *master,
 	struct spi_device *spi = message->spi;
 	struct spi_transfer *transfer;
 	int status = 0;
+	enum dspi_trans_mode trans_mode;
+	u32 spi_tcr;
+
+	regmap_read(dspi->regmap, SPI_TCR, &spi_tcr);
+	dspi->spi_tcnt = SPI_TCR_GET_TCNT(spi_tcr);
+
 	message->actual_length = 0;
 
 	list_for_each_entry(transfer, &message->transfers, transfer_list) {
@@ -295,10 +384,10 @@ static int dspi_transfer_one_message(struct spi_master *master,
 		dspi->cur_msg = message;
 		dspi->cur_chip = spi_get_ctldata(spi);
 		dspi->cs = spi->chip_select;
+		dspi->cs_change = 0;
 		if (dspi->cur_transfer->transfer_list.next
 				== &dspi->cur_msg->transfers)
-			transfer->cs_change = 1;
-		dspi->cs_change = transfer->cs_change;
+			dspi->cs_change = 1;
 		dspi->void_write_data = dspi->cur_chip->void_write_data;
 
 		dspi->dataflags = 0;
@@ -324,8 +413,22 @@ static int dspi_transfer_one_message(struct spi_master *master,
 			regmap_write(dspi->regmap, SPI_CTAR(dspi->cs),
 					dspi->cur_chip->ctar_val);
 
-		regmap_write(dspi->regmap, SPI_RSER, SPI_RSER_EOQFE);
-		message->actual_length += dspi_transfer_write(dspi);
+		trans_mode = dspi->devtype_data->trans_mode;
+		switch (trans_mode) {
+		case DSPI_EOQ_MODE:
+			regmap_write(dspi->regmap, SPI_RSER, SPI_RSER_EOQFE);
+			dspi_eoq_write(dspi);
+			break;
+		case DSPI_TCFQ_MODE:
+			regmap_write(dspi->regmap, SPI_RSER, SPI_RSER_TCFQE);
+			dspi_tcfq_write(dspi);
+			break;
+		default:
+			dev_err(&dspi->pdev->dev, "unsupported trans_mode %u\n",
+				trans_mode);
+			status = -EINVAL;
+			goto out;
+		}
 
 		if (wait_event_interruptible(dspi->waitq, dspi->waitflags))
 			dev_err(&dspi->pdev->dev, "wait transfer complete fail!\n");
@@ -335,6 +438,7 @@ static int dspi_transfer_one_message(struct spi_master *master,
 			udelay(transfer->delay_usecs);
 	}
 
+out:
 	message->status = status;
 	spi_finalize_current_message(master);
 
@@ -345,7 +449,10 @@ static int dspi_setup(struct spi_device *spi)
 {
 	struct chip_data *chip;
 	struct fsl_dspi *dspi = spi_master_get_devdata(spi->master);
-	unsigned char br = 0, pbr = 0, fmsz = 0;
+	u32 cs_sck_delay = 0, sck_cs_delay = 0;
+	unsigned char br = 0, pbr = 0, pcssck = 0, cssck = 0;
+	unsigned char pasc = 0, asc = 0, fmsz = 0;
+	unsigned long clkrate;
 
 	if ((spi->bits_per_word >= 4) && (spi->bits_per_word <= 16)) {
 		fmsz = spi->bits_per_word - 1;
@@ -362,18 +469,34 @@ static int dspi_setup(struct spi_device *spi)
 			return -ENOMEM;
 	}
 
+	of_property_read_u32(spi->dev.of_node, "fsl,spi-cs-sck-delay",
+			&cs_sck_delay);
+
+	of_property_read_u32(spi->dev.of_node, "fsl,spi-sck-cs-delay",
+			&sck_cs_delay);
+
 	chip->mcr_val = SPI_MCR_MASTER | SPI_MCR_PCSIS |
 		SPI_MCR_CLR_TXF | SPI_MCR_CLR_RXF;
 
 	chip->void_write_data = 0;
 
-	hz_to_spi_baud(&pbr, &br,
-			spi->max_speed_hz, clk_get_rate(dspi->clk));
+	clkrate = clk_get_rate(dspi->clk);
+	hz_to_spi_baud(&pbr, &br, spi->max_speed_hz, clkrate);
+
+	/* Set PCS to SCK delay scale values */
+	ns_delay_scale(&pcssck, &cssck, cs_sck_delay, clkrate);
+
+	/* Set After SCK delay scale values */
+	ns_delay_scale(&pasc, &asc, sck_cs_delay, clkrate);
 
 	chip->ctar_val =  SPI_CTAR_FMSZ(fmsz)
 		| SPI_CTAR_CPOL(spi->mode & SPI_CPOL ? 1 : 0)
 		| SPI_CTAR_CPHA(spi->mode & SPI_CPHA ? 1 : 0)
 		| SPI_CTAR_LSBFE(spi->mode & SPI_LSB_FIRST ? 1 : 0)
+		| SPI_CTAR_PCSSCK(pcssck)
+		| SPI_CTAR_CSSCK(cssck)
+		| SPI_CTAR_PASC(pasc)
+		| SPI_CTAR_ASC(asc)
 		| SPI_CTAR_PBR(pbr)
 		| SPI_CTAR_BR(br);
 
@@ -395,27 +518,89 @@ static void dspi_cleanup(struct spi_device *spi)
 static irqreturn_t dspi_interrupt(int irq, void *dev_id)
 {
 	struct fsl_dspi *dspi = (struct fsl_dspi *)dev_id;
-
 	struct spi_message *msg = dspi->cur_msg;
+	enum dspi_trans_mode trans_mode;
+	u32 spi_sr, spi_tcr;
+	u32 spi_tcnt, tcnt_diff;
+	int tx_word;
 
-	regmap_write(dspi->regmap, SPI_SR, SPI_SR_EOQF);
-	dspi_transfer_read(dspi);
-
-	if (!dspi->len) {
+	regmap_read(dspi->regmap, SPI_SR, &spi_sr);
+	regmap_write(dspi->regmap, SPI_SR, spi_sr);
+
+
+	if (spi_sr & (SPI_SR_EOQF | SPI_SR_TCFQF)) {
+		tx_word = is_double_byte_mode(dspi);
+
+		regmap_read(dspi->regmap, SPI_TCR, &spi_tcr);
+		spi_tcnt = SPI_TCR_GET_TCNT(spi_tcr);
+		/*
+		 * The width of SPI Transfer Counter in SPI_TCR is 16bits,
+		 * so the max couner is 65535. When the counter reach 65535,
+		 * it will wrap around, counter reset to zero.
+		 * spi_tcnt my be less than dspi->spi_tcnt, it means the
+		 * counter already wrapped around.
+		 * SPI Transfer Counter is a counter of transmitted frames.
+		 * The size of frame maybe two bytes.
+		 */
+		tcnt_diff = ((spi_tcnt + SPI_TCR_TCNT_MAX) - dspi->spi_tcnt)
+			% SPI_TCR_TCNT_MAX;
+		tcnt_diff *= (tx_word + 1);
 		if (dspi->dataflags & TRAN_STATE_WORD_ODD_NUM)
-			regmap_update_bits(dspi->regmap, SPI_CTAR(dspi->cs),
-			SPI_FRAME_BITS_MASK, SPI_FRAME_BITS(16));
+			tcnt_diff--;
+
+		msg->actual_length += tcnt_diff;
+
+		dspi->spi_tcnt = spi_tcnt;
+
+		trans_mode = dspi->devtype_data->trans_mode;
+		switch (trans_mode) {
+		case DSPI_EOQ_MODE:
+			dspi_eoq_read(dspi);
+			break;
+		case DSPI_TCFQ_MODE:
+			dspi_tcfq_read(dspi);
+			break;
+		default:
+			dev_err(&dspi->pdev->dev, "unsupported trans_mode %u\n",
+				trans_mode);
+				return IRQ_HANDLED;
+		}
 
-		dspi->waitflags = 1;
-		wake_up_interruptible(&dspi->waitq);
-	} else
-		msg->actual_length += dspi_transfer_write(dspi);
+		if (!dspi->len) {
+			if (dspi->dataflags & TRAN_STATE_WORD_ODD_NUM) {
+				regmap_update_bits(dspi->regmap,
+						   SPI_CTAR(dspi->cs),
+						   SPI_FRAME_BITS_MASK,
+						   SPI_FRAME_BITS(16));
+				dspi->dataflags &= ~TRAN_STATE_WORD_ODD_NUM;
+			}
+
+			dspi->waitflags = 1;
+			wake_up_interruptible(&dspi->waitq);
+		} else {
+			switch (trans_mode) {
+			case DSPI_EOQ_MODE:
+				dspi_eoq_write(dspi);
+				break;
+			case DSPI_TCFQ_MODE:
+				dspi_tcfq_write(dspi);
+				break;
+			default:
+				dev_err(&dspi->pdev->dev,
+					"unsupported trans_mode %u\n",
+					trans_mode);
+			}
+		}
+	}
 
 	return IRQ_HANDLED;
 }
 
 static const struct of_device_id fsl_dspi_dt_ids[] = {
-	{ .compatible = "fsl,vf610-dspi", .data = NULL, },
+	{ .compatible = "fsl,vf610-dspi", .data = (void *)&vf610_data, },
+	{ .compatible = "fsl,ls1021a-v1.0-dspi",
+		.data = (void *)&ls1021a_v1_data, },
+	{ .compatible = "fsl,ls2085a-dspi", .data = (void *)&ls2085a_data, },
 	{ /* sentinel */ }
 };
 MODULE_DEVICE_TABLE(of, fsl_dspi_dt_ids);
@@ -429,6 +614,8 @@ static int dspi_suspend(struct device *dev)
 	spi_master_suspend(master);
 	clk_disable_unprepare(dspi->clk);
 
+	pinctrl_pm_select_sleep_state(dev);
+
 	return 0;
 }
 
@@ -437,6 +624,8 @@ static int dspi_resume(struct device *dev)
 	struct spi_master *master = dev_get_drvdata(dev);
 	struct fsl_dspi *dspi = spi_master_get_devdata(master);
 
+	pinctrl_pm_select_default_state(dev);
+
 	clk_prepare_enable(dspi->clk);
 	spi_master_resume(master);
 
@@ -461,6 +650,8 @@ static int dspi_probe(struct platform_device *pdev)
 	struct resource *res;
 	void __iomem *base;
 	int ret = 0, cs_num, bus_num;
+	const struct of_device_id *of_id =
+			of_match_device(fsl_dspi_dt_ids, &pdev->dev);
 
 	master = spi_alloc_master(&pdev->dev, sizeof(struct fsl_dspi));
 	if (!master)
@@ -494,6 +685,13 @@ static int dspi_probe(struct platform_device *pdev)
 	}
 	master->bus_num = bus_num;
 
+	dspi->devtype_data = (struct fsl_dspi_devtype_data *)of_id->data;
+	if (!dspi->devtype_data) {
+		dev_err(&pdev->dev, "can't get devtype_data\n");
+		ret = -EFAULT;
+		goto out_master_put;
+	}
+
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	base = devm_ioremap_resource(&pdev->dev, res);
 	if (IS_ERR(base)) {
@@ -501,7 +699,7 @@ static int dspi_probe(struct platform_device *pdev)
 		goto out_master_put;
 	}
 
-	dspi->regmap = devm_regmap_init_mmio_clk(&pdev->dev, "dspi", base,
+	dspi->regmap = devm_regmap_init_mmio_clk(&pdev->dev, NULL, base,
 						&dspi_regmap_config);
 	if (IS_ERR(dspi->regmap)) {
 		dev_err(&pdev->dev, "failed to init regmap: %ld\n",
diff --git a/drivers/spi/spi-fsl-espi.c b/drivers/spi/spi-fsl-espi.c
index d0a73a09a9bd..d3f05a0525a4 100644
--- a/drivers/spi/spi-fsl-espi.c
+++ b/drivers/spi/spi-fsl-espi.c
@@ -359,14 +359,16 @@ static void fsl_espi_rw_trans(struct spi_message *m,
 				struct fsl_espi_transfer *trans, u8 *rx_buff)
 {
 	struct fsl_espi_transfer *espi_trans = trans;
-	unsigned int n_tx = espi_trans->n_tx;
-	unsigned int n_rx = espi_trans->n_rx;
+	unsigned int total_len = espi_trans->len;
 	struct spi_transfer *t;
 	u8 *local_buf;
 	u8 *rx_buf = rx_buff;
 	unsigned int trans_len;
 	unsigned int addr;
-	int i, pos, loop;
+	unsigned int tx_only;
+	unsigned int rx_pos = 0;
+	unsigned int pos;
+	int i, loop;
 
 	local_buf = kzalloc(SPCOM_TRANLEN_MAX, GFP_KERNEL);
 	if (!local_buf) {
@@ -374,36 +376,48 @@ static void fsl_espi_rw_trans(struct spi_message *m,
 		return;
 	}
 
-	for (pos = 0, loop = 0; pos < n_rx; pos += trans_len, loop++) {
-		trans_len = n_rx - pos;
-		if (trans_len > SPCOM_TRANLEN_MAX - n_tx)
-			trans_len = SPCOM_TRANLEN_MAX - n_tx;
+	for (pos = 0, loop = 0; pos < total_len; pos += trans_len, loop++) {
+		trans_len = total_len - pos;
 
 		i = 0;
+		tx_only = 0;
 		list_for_each_entry(t, &m->transfers, transfer_list) {
 			if (t->tx_buf) {
 				memcpy(local_buf + i, t->tx_buf, t->len);
 				i += t->len;
+				if (!t->rx_buf)
+					tx_only += t->len;
 			}
 		}
 
+		/* Add additional TX bytes to compensate SPCOM_TRANLEN_MAX */
+		if (loop > 0)
+			trans_len += tx_only;
+
+		if (trans_len > SPCOM_TRANLEN_MAX)
+			trans_len = SPCOM_TRANLEN_MAX;
+
+		/* Update device offset */
 		if (pos > 0) {
 			addr = fsl_espi_cmd2addr(local_buf);
-			addr += pos;
+			addr += rx_pos;
 			fsl_espi_addr2cmd(addr, local_buf);
 		}
 
-		espi_trans->n_tx = n_tx;
-		espi_trans->n_rx = trans_len;
-		espi_trans->len = trans_len + n_tx;
+		espi_trans->len = trans_len;
 		espi_trans->tx_buf = local_buf;
 		espi_trans->rx_buf = local_buf;
 		fsl_espi_do_trans(m, espi_trans);
 
-		memcpy(rx_buf + pos, espi_trans->rx_buf + n_tx, trans_len);
+		/* If there is at least one RX byte then copy it to rx_buf */
+		if (tx_only < SPCOM_TRANLEN_MAX)
+			memcpy(rx_buf + rx_pos, espi_trans->rx_buf + tx_only,
+					trans_len - tx_only);
+
+		rx_pos += trans_len - tx_only;
 
 		if (loop > 0)
-			espi_trans->actual_length += espi_trans->len - n_tx;
+			espi_trans->actual_length += espi_trans->len - tx_only;
 		else
 			espi_trans->actual_length += espi_trans->len;
 	}
@@ -418,6 +432,7 @@ static int fsl_espi_do_one_msg(struct spi_master *master,
 	u8 *rx_buf = NULL;
 	unsigned int n_tx = 0;
 	unsigned int n_rx = 0;
+	unsigned int xfer_len = 0;
 	struct fsl_espi_transfer espi_trans;
 
 	list_for_each_entry(t, &m->transfers, transfer_list) {
@@ -427,11 +442,13 @@ static int fsl_espi_do_one_msg(struct spi_master *master,
 			n_rx += t->len;
 			rx_buf = t->rx_buf;
 		}
+		if ((t->tx_buf) || (t->rx_buf))
+			xfer_len += t->len;
 	}
 
 	espi_trans.n_tx = n_tx;
 	espi_trans.n_rx = n_rx;
-	espi_trans.len = n_tx + n_rx;
+	espi_trans.len = xfer_len;
 	espi_trans.actual_length = 0;
 	espi_trans.status = 0;
 
@@ -544,9 +561,13 @@ void fsl_espi_cpu_irq(struct mpc8xxx_spi *mspi, u32 events)
 
 		/* spin until TX is done */
 		ret = spin_event_timeout(((events = mpc8xxx_spi_read_reg(
-				&reg_base->event)) & SPIE_NF) == 0, 1000, 0);
+				&reg_base->event)) & SPIE_NF), 1000, 0);
 		if (!ret) {
 			dev_err(mspi->dev, "tired waiting for SPIE_NF\n");
+
+			/* Clear the SPIE bits */
+			mpc8xxx_spi_write_reg(&reg_base->event, events);
+			complete(&mspi->done);
 			return;
 		}
 	}
diff --git a/drivers/spi/spi-img-spfi.c b/drivers/spi/spi-img-spfi.c
index e649bc7d4c08..788e2b176a4f 100644
--- a/drivers/spi/spi-img-spfi.c
+++ b/drivers/spi/spi-img-spfi.c
@@ -12,6 +12,7 @@
 #include <linux/clk.h>
 #include <linux/delay.h>
 #include <linux/dmaengine.h>
+#include <linux/gpio.h>
 #include <linux/interrupt.h>
 #include <linux/io.h>
 #include <linux/irq.h>
@@ -122,36 +123,31 @@ static inline void spfi_start(struct img_spfi *spfi)
 	spfi_writel(spfi, val, SPFI_CONTROL);
 }
 
-static inline void spfi_stop(struct img_spfi *spfi)
-{
-	u32 val;
-
-	val = spfi_readl(spfi, SPFI_CONTROL);
-	val &= ~SPFI_CONTROL_SPFI_EN;
-	spfi_writel(spfi, val, SPFI_CONTROL);
-}
-
 static inline void spfi_reset(struct img_spfi *spfi)
 {
 	spfi_writel(spfi, SPFI_CONTROL_SOFT_RESET, SPFI_CONTROL);
-	udelay(1);
 	spfi_writel(spfi, 0, SPFI_CONTROL);
 }
 
-static void spfi_flush_tx_fifo(struct img_spfi *spfi)
+static int spfi_wait_all_done(struct img_spfi *spfi)
 {
-	unsigned long timeout = jiffies + msecs_to_jiffies(10);
+	unsigned long timeout = jiffies + msecs_to_jiffies(50);
 
-	spfi_writel(spfi, SPFI_INTERRUPT_SDE, SPFI_INTERRUPT_CLEAR);
 	while (time_before(jiffies, timeout)) {
-		if (spfi_readl(spfi, SPFI_INTERRUPT_STATUS) &
-		    SPFI_INTERRUPT_SDE)
-			return;
+		u32 status = spfi_readl(spfi, SPFI_INTERRUPT_STATUS);
+
+		if (status & SPFI_INTERRUPT_ALLDONETRIG) {
+			spfi_writel(spfi, SPFI_INTERRUPT_ALLDONETRIG,
+				    SPFI_INTERRUPT_CLEAR);
+			return 0;
+		}
 		cpu_relax();
 	}
 
-	dev_err(spfi->dev, "Timed out waiting for FIFO to drain\n");
+	dev_err(spfi->dev, "Timed out waiting for transaction to complete\n");
 	spfi_reset(spfi);
+
+	return -ETIMEDOUT;
 }
 
 static unsigned int spfi_pio_write32(struct img_spfi *spfi, const u32 *buf,
@@ -237,6 +233,7 @@ static int img_spfi_start_pio(struct spi_master *master,
 	const void *tx_buf = xfer->tx_buf;
 	void *rx_buf = xfer->rx_buf;
 	unsigned long timeout;
+	int ret;
 
 	if (tx_buf)
 		tx_bytes = xfer->len;
@@ -269,16 +266,15 @@ static int img_spfi_start_pio(struct spi_master *master,
 		cpu_relax();
 	}
 
+	ret = spfi_wait_all_done(spfi);
+	if (ret < 0)
+		return ret;
+
 	if (rx_bytes > 0 || tx_bytes > 0) {
 		dev_err(spfi->dev, "PIO transfer timed out\n");
-		spfi_reset(spfi);
 		return -ETIMEDOUT;
 	}
 
-	if (tx_buf)
-		spfi_flush_tx_fifo(spfi);
-	spfi_stop(spfi);
-
 	return 0;
 }
 
@@ -287,14 +283,12 @@ static void img_spfi_dma_rx_cb(void *data)
 	struct img_spfi *spfi = data;
 	unsigned long flags;
 
-	spin_lock_irqsave(&spfi->lock, flags);
+	spfi_wait_all_done(spfi);
 
+	spin_lock_irqsave(&spfi->lock, flags);
 	spfi->rx_dma_busy = false;
-	if (!spfi->tx_dma_busy) {
-		spfi_stop(spfi);
+	if (!spfi->tx_dma_busy)
 		spi_finalize_current_transfer(spfi->master);
-	}
-
 	spin_unlock_irqrestore(&spfi->lock, flags);
 }
 
@@ -303,16 +297,12 @@ static void img_spfi_dma_tx_cb(void *data)
 	struct img_spfi *spfi = data;
 	unsigned long flags;
 
-	spfi_flush_tx_fifo(spfi);
+	spfi_wait_all_done(spfi);
 
 	spin_lock_irqsave(&spfi->lock, flags);
-
 	spfi->tx_dma_busy = false;
-	if (!spfi->rx_dma_busy) {
-		spfi_stop(spfi);
+	if (!spfi->rx_dma_busy)
 		spi_finalize_current_transfer(spfi->master);
-	}
-
 	spin_unlock_irqrestore(&spfi->lock, flags);
 }
 
@@ -397,6 +387,75 @@ stop_dma:
 	return -EIO;
 }
 
+static void img_spfi_handle_err(struct spi_master *master,
+				struct spi_message *msg)
+{
+	struct img_spfi *spfi = spi_master_get_devdata(master);
+	unsigned long flags;
+
+	/*
+	 * Stop all DMA and reset the controller if the previous transaction
+	 * timed-out and never completed it's DMA.
+	 */
+	spin_lock_irqsave(&spfi->lock, flags);
+	if (spfi->tx_dma_busy || spfi->rx_dma_busy) {
+		spfi->tx_dma_busy = false;
+		spfi->rx_dma_busy = false;
+
+		dmaengine_terminate_all(spfi->tx_ch);
+		dmaengine_terminate_all(spfi->rx_ch);
+	}
+	spin_unlock_irqrestore(&spfi->lock, flags);
+}
+
+static int img_spfi_prepare(struct spi_master *master, struct spi_message *msg)
+{
+	struct img_spfi *spfi = spi_master_get_devdata(master);
+	u32 val;
+
+	val = spfi_readl(spfi, SPFI_PORT_STATE);
+	if (msg->spi->mode & SPI_CPHA)
+		val |= SPFI_PORT_STATE_CK_PHASE(msg->spi->chip_select);
+	else
+		val &= ~SPFI_PORT_STATE_CK_PHASE(msg->spi->chip_select);
+	if (msg->spi->mode & SPI_CPOL)
+		val |= SPFI_PORT_STATE_CK_POL(msg->spi->chip_select);
+	else
+		val &= ~SPFI_PORT_STATE_CK_POL(msg->spi->chip_select);
+	spfi_writel(spfi, val, SPFI_PORT_STATE);
+
+	return 0;
+}
+
+static int img_spfi_unprepare(struct spi_master *master,
+			      struct spi_message *msg)
+{
+	struct img_spfi *spfi = spi_master_get_devdata(master);
+
+	spfi_reset(spfi);
+
+	return 0;
+}
+
+static int img_spfi_setup(struct spi_device *spi)
+{
+	int ret;
+
+	ret = gpio_request_one(spi->cs_gpio, (spi->mode & SPI_CS_HIGH) ?
+			       GPIOF_OUT_INIT_LOW : GPIOF_OUT_INIT_HIGH,
+			       dev_name(&spi->dev));
+	if (ret)
+		dev_err(&spi->dev, "can't request chipselect gpio %d\n",
+				spi->cs_gpio);
+
+	return ret;
+}
+
+static void img_spfi_cleanup(struct spi_device *spi)
+{
+	gpio_free(spi->cs_gpio);
+}
+
 static void img_spfi_config(struct spi_master *master, struct spi_device *spi,
 			    struct spi_transfer *xfer)
 {
@@ -405,10 +464,10 @@ static void img_spfi_config(struct spi_master *master, struct spi_device *spi,
 
 	/*
 	 * output = spfi_clk * (BITCLK / 512), where BITCLK must be a
-	 * power of 2 up to 256 (where 255 == 256 since BITCLK is 8 bits)
+	 * power of 2 up to 128
 	 */
-	div = DIV_ROUND_UP(master->max_speed_hz, xfer->speed_hz);
-	div = clamp(512 / (1 << get_count_order(div)), 1, 255);
+	div = DIV_ROUND_UP(clk_get_rate(spfi->spfi_clk), xfer->speed_hz);
+	div = clamp(512 / (1 << get_count_order(div)), 1, 128);
 
 	val = spfi_readl(spfi, SPFI_DEVICE_PARAMETER(spi->chip_select));
 	val &= ~(SPFI_DEVICE_PARAMETER_BITCLK_MASK <<
@@ -416,6 +475,9 @@ static void img_spfi_config(struct spi_master *master, struct spi_device *spi,
 	val |= div << SPFI_DEVICE_PARAMETER_BITCLK_SHIFT;
 	spfi_writel(spfi, val, SPFI_DEVICE_PARAMETER(spi->chip_select));
 
+	spfi_writel(spfi, xfer->len << SPFI_TRANSACTION_TSIZE_SHIFT,
+		    SPFI_TRANSACTION);
+
 	val = spfi_readl(spfi, SPFI_CONTROL);
 	val &= ~(SPFI_CONTROL_SEND_DMA | SPFI_CONTROL_GET_DMA);
 	if (xfer->tx_buf)
@@ -429,25 +491,7 @@ static void img_spfi_config(struct spi_master *master, struct spi_device *spi,
 	else if (xfer->tx_nbits == SPI_NBITS_QUAD &&
 		 xfer->rx_nbits == SPI_NBITS_QUAD)
 		val |= SPFI_CONTROL_TMODE_QUAD << SPFI_CONTROL_TMODE_SHIFT;
-	val &= ~SPFI_CONTROL_CONTINUE;
-	if (!xfer->cs_change && !list_is_last(&xfer->transfer_list,
-					      &master->cur_msg->transfers))
-		val |= SPFI_CONTROL_CONTINUE;
 	spfi_writel(spfi, val, SPFI_CONTROL);
-
-	val = spfi_readl(spfi, SPFI_PORT_STATE);
-	if (spi->mode & SPI_CPHA)
-		val |= SPFI_PORT_STATE_CK_PHASE(spi->chip_select);
-	else
-		val &= ~SPFI_PORT_STATE_CK_PHASE(spi->chip_select);
-	if (spi->mode & SPI_CPOL)
-		val |= SPFI_PORT_STATE_CK_POL(spi->chip_select);
-	else
-		val &= ~SPFI_PORT_STATE_CK_POL(spi->chip_select);
-	spfi_writel(spfi, val, SPFI_PORT_STATE);
-
-	spfi_writel(spfi, xfer->len << SPFI_TRANSACTION_TSIZE_SHIFT,
-		    SPFI_TRANSACTION);
 }
 
 static int img_spfi_transfer_one(struct spi_master *master,
@@ -455,8 +499,6 @@ static int img_spfi_transfer_one(struct spi_master *master,
 				 struct spi_transfer *xfer)
 {
 	struct img_spfi *spfi = spi_master_get_devdata(spi->master);
-	bool dma_reset = false;
-	unsigned long flags;
 	int ret;
 
 	if (xfer->len > SPFI_TRANSACTION_TSIZE_MASK) {
@@ -466,23 +508,6 @@ static int img_spfi_transfer_one(struct spi_master *master,
 		return -EINVAL;
 	}
 
-	/*
-	 * Stop all DMA and reset the controller if the previous transaction
-	 * timed-out and never completed it's DMA.
-	 */
-	spin_lock_irqsave(&spfi->lock, flags);
-	if (spfi->tx_dma_busy || spfi->rx_dma_busy) {
-		dev_err(spfi->dev, "SPI DMA still busy\n");
-		dma_reset = true;
-	}
-	spin_unlock_irqrestore(&spfi->lock, flags);
-
-	if (dma_reset) {
-		dmaengine_terminate_all(spfi->tx_ch);
-		dmaengine_terminate_all(spfi->rx_ch);
-		spfi_reset(spfi);
-	}
-
 	img_spfi_config(master, spi, xfer);
 	if (master->can_dma && master->can_dma(master, spi, xfer))
 		ret = img_spfi_start_dma(master, spi, xfer);
@@ -492,17 +517,6 @@ static int img_spfi_transfer_one(struct spi_master *master,
 	return ret;
 }
 
-static void img_spfi_set_cs(struct spi_device *spi, bool enable)
-{
-	struct img_spfi *spfi = spi_master_get_devdata(spi->master);
-	u32 val;
-
-	val = spfi_readl(spfi, SPFI_PORT_STATE);
-	val &= ~(SPFI_PORT_STATE_DEV_SEL_MASK << SPFI_PORT_STATE_DEV_SEL_SHIFT);
-	val |= spi->chip_select << SPFI_PORT_STATE_DEV_SEL_SHIFT;
-	spfi_writel(spfi, val, SPFI_PORT_STATE);
-}
-
 static bool img_spfi_can_dma(struct spi_master *master, struct spi_device *spi,
 			     struct spi_transfer *xfer)
 {
@@ -591,14 +605,17 @@ static int img_spfi_probe(struct platform_device *pdev)
 	master->mode_bits = SPI_CPOL | SPI_CPHA | SPI_TX_DUAL | SPI_RX_DUAL;
 	if (of_property_read_bool(spfi->dev->of_node, "img,supports-quad-mode"))
 		master->mode_bits |= SPI_TX_QUAD | SPI_RX_QUAD;
-	master->num_chipselect = 5;
 	master->dev.of_node = pdev->dev.of_node;
 	master->bits_per_word_mask = SPI_BPW_MASK(32) | SPI_BPW_MASK(8);
-	master->max_speed_hz = clk_get_rate(spfi->spfi_clk);
-	master->min_speed_hz = master->max_speed_hz / 512;
+	master->max_speed_hz = clk_get_rate(spfi->spfi_clk) / 4;
+	master->min_speed_hz = clk_get_rate(spfi->spfi_clk) / 512;
 
-	master->set_cs = img_spfi_set_cs;
+	master->setup = img_spfi_setup;
+	master->cleanup = img_spfi_cleanup;
 	master->transfer_one = img_spfi_transfer_one;
+	master->prepare_message = img_spfi_prepare;
+	master->unprepare_message = img_spfi_unprepare;
+	master->handle_err = img_spfi_handle_err;
 
 	spfi->tx_ch = dma_request_slave_channel(spfi->dev, "tx");
 	spfi->rx_ch = dma_request_slave_channel(spfi->dev, "rx");
diff --git a/drivers/spi/spi-imx.c b/drivers/spi/spi-imx.c
index 6fea4af51c41..eb7d3a6fb14c 100644
--- a/drivers/spi/spi-imx.c
+++ b/drivers/spi/spi-imx.c
@@ -370,8 +370,6 @@ static int __maybe_unused mx51_ecspi_config(struct spi_imx_data *spi_imx,
 	if (spi_imx->dma_is_inited) {
 		dma = readl(spi_imx->base + MX51_ECSPI_DMA);
 
-		spi_imx->tx_wml = spi_imx_get_fifosize(spi_imx) / 2;
-		spi_imx->rx_wml = spi_imx_get_fifosize(spi_imx) / 2;
 		spi_imx->rxt_wml = spi_imx_get_fifosize(spi_imx) / 2;
 		rx_wml_cfg = spi_imx->rx_wml << MX51_ECSPI_DMA_RX_WML_OFFSET;
 		tx_wml_cfg = spi_imx->tx_wml << MX51_ECSPI_DMA_TX_WML_OFFSET;
@@ -676,7 +674,7 @@ static struct spi_imx_devtype_data imx51_ecspi_devtype_data = {
 	.devtype = IMX51_ECSPI,
 };
 
-static struct platform_device_id spi_imx_devtype[] = {
+static const struct platform_device_id spi_imx_devtype[] = {
 	{
 		.name = "imx1-cspi",
 		.driver_data = (kernel_ulong_t) &imx1_cspi_devtype_data,
@@ -868,6 +866,8 @@ static int spi_imx_sdma_init(struct device *dev, struct spi_imx_data *spi_imx,
 	master->max_dma_len = MAX_SDMA_BD_BYTES;
 	spi_imx->bitbang.master->flags = SPI_MASTER_MUST_RX |
 					 SPI_MASTER_MUST_TX;
+	spi_imx->tx_wml = spi_imx_get_fifosize(spi_imx) / 2;
+	spi_imx->rx_wml = spi_imx_get_fifosize(spi_imx) / 2;
 	spi_imx->dma_is_inited = 1;
 
 	return 0;
@@ -903,7 +903,7 @@ static int spi_imx_dma_transfer(struct spi_imx_data *spi_imx,
 
 	if (tx) {
 		desc_tx = dmaengine_prep_slave_sg(master->dma_tx,
-					tx->sgl, tx->nents, DMA_TO_DEVICE,
+					tx->sgl, tx->nents, DMA_MEM_TO_DEV,
 					DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
 		if (!desc_tx)
 			goto no_dma;
@@ -915,7 +915,7 @@ static int spi_imx_dma_transfer(struct spi_imx_data *spi_imx,
 
 	if (rx) {
 		desc_rx = dmaengine_prep_slave_sg(master->dma_rx,
-					rx->sgl, rx->nents, DMA_FROM_DEVICE,
+					rx->sgl, rx->nents, DMA_DEV_TO_MEM,
 					DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
 		if (!desc_rx)
 			goto no_dma;
diff --git a/drivers/spi/spi-mpc512x-psc.c b/drivers/spi/spi-mpc512x-psc.c
index ecae0d4e2945..965d2bdcfdcc 100644
--- a/drivers/spi/spi-mpc512x-psc.c
+++ b/drivers/spi/spi-mpc512x-psc.c
@@ -588,7 +588,7 @@ static int mpc512x_psc_spi_of_remove(struct platform_device *op)
 	return mpc512x_psc_spi_do_remove(&op->dev);
 }
 
-static struct of_device_id mpc512x_psc_spi_of_match[] = {
+static const struct of_device_id mpc512x_psc_spi_of_match[] = {
 	{ .compatible = "fsl,mpc5121-psc-spi", },
 	{},
 };
diff --git a/drivers/spi/spi-octeon.c b/drivers/spi/spi-octeon.c
index b283d537d16a..e99d6a93d394 100644
--- a/drivers/spi/spi-octeon.c
+++ b/drivers/spi/spi-octeon.c
@@ -238,7 +238,7 @@ static int octeon_spi_remove(struct platform_device *pdev)
 	return 0;
 }
 
-static struct of_device_id octeon_spi_match[] = {
+static const struct of_device_id octeon_spi_match[] = {
 	{ .compatible = "cavium,octeon-3010-spi", },
 	{},
 };
diff --git a/drivers/spi/spi-omap-100k.c b/drivers/spi/spi-omap-100k.c
index d890d309dff9..35b332dacb13 100644
--- a/drivers/spi/spi-omap-100k.c
+++ b/drivers/spi/spi-omap-100k.c
@@ -24,6 +24,7 @@
 #include <linux/device.h>
 #include <linux/delay.h>
 #include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
 #include <linux/err.h>
 #include <linux/clk.h>
 #include <linux/io.h>
@@ -294,16 +295,6 @@ static int omap1_spi100k_setup(struct spi_device *spi)
 	return ret;
 }
 
-static int omap1_spi100k_prepare_hardware(struct spi_master *master)
-{
-	struct omap1_spi100k *spi100k = spi_master_get_devdata(master);
-
-	clk_prepare_enable(spi100k->ick);
-	clk_prepare_enable(spi100k->fck);
-
-	return 0;
-}
-
 static int omap1_spi100k_transfer_one_message(struct spi_master *master,
 					      struct spi_message *m)
 {
@@ -372,16 +363,6 @@ static int omap1_spi100k_transfer_one_message(struct spi_master *master,
 	return status;
 }
 
-static int omap1_spi100k_unprepare_hardware(struct spi_master *master)
-{
-	struct omap1_spi100k *spi100k = spi_master_get_devdata(master);
-
-	clk_disable_unprepare(spi100k->ick);
-	clk_disable_unprepare(spi100k->fck);
-
-	return 0;
-}
-
 static int omap1_spi100k_probe(struct platform_device *pdev)
 {
 	struct spi_master       *master;
@@ -402,14 +383,12 @@ static int omap1_spi100k_probe(struct platform_device *pdev)
 
 	master->setup = omap1_spi100k_setup;
 	master->transfer_one_message = omap1_spi100k_transfer_one_message;
-	master->prepare_transfer_hardware = omap1_spi100k_prepare_hardware;
-	master->unprepare_transfer_hardware = omap1_spi100k_unprepare_hardware;
-	master->cleanup = NULL;
 	master->num_chipselect = 2;
 	master->mode_bits = MODEBITS;
 	master->bits_per_word_mask = SPI_BPW_RANGE_MASK(4, 32);
 	master->min_speed_hz = OMAP1_SPI100K_MAX_FREQ/(1<<16);
 	master->max_speed_hz = OMAP1_SPI100K_MAX_FREQ;
+	master->auto_runtime_pm = true;
 
 	spi100k = spi_master_get_devdata(master);
 
@@ -434,22 +413,96 @@ static int omap1_spi100k_probe(struct platform_device *pdev)
 		goto err;
 	}
 
+	status = clk_prepare_enable(spi100k->ick);
+	if (status != 0) {
+		dev_err(&pdev->dev, "failed to enable ick: %d\n", status);
+		goto err;
+	}
+
+	status = clk_prepare_enable(spi100k->fck);
+	if (status != 0) {
+		dev_err(&pdev->dev, "failed to enable fck: %d\n", status);
+		goto err_ick;
+	}
+
+	pm_runtime_enable(&pdev->dev);
+	pm_runtime_set_active(&pdev->dev);
+
 	status = devm_spi_register_master(&pdev->dev, master);
 	if (status < 0)
-		goto err;
+		goto err_fck;
 
 	return status;
 
+err_fck:
+	clk_disable_unprepare(spi100k->fck);
+err_ick:
+	clk_disable_unprepare(spi100k->ick);
 err:
 	spi_master_put(master);
 	return status;
 }
 
+static int omap1_spi100k_remove(struct platform_device *pdev)
+{
+	struct spi_master *master = spi_master_get(platform_get_drvdata(pdev));
+	struct omap1_spi100k *spi100k = spi_master_get_devdata(master);
+
+	pm_runtime_disable(&pdev->dev);
+
+	clk_disable_unprepare(spi100k->fck);
+	clk_disable_unprepare(spi100k->ick);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM
+static int omap1_spi100k_runtime_suspend(struct device *dev)
+{
+	struct spi_master *master = spi_master_get(dev_get_drvdata(dev));
+	struct omap1_spi100k *spi100k = spi_master_get_devdata(master);
+
+	clk_disable_unprepare(spi100k->ick);
+	clk_disable_unprepare(spi100k->fck);
+
+	return 0;
+}
+
+static int omap1_spi100k_runtime_resume(struct device *dev)
+{
+	struct spi_master *master = spi_master_get(dev_get_drvdata(dev));
+	struct omap1_spi100k *spi100k = spi_master_get_devdata(master);
+	int ret;
+
+	ret = clk_prepare_enable(spi100k->ick);
+	if (ret != 0) {
+		dev_err(dev, "Failed to enable ick: %d\n", ret);
+		return ret;
+	}
+
+	ret = clk_prepare_enable(spi100k->fck);
+	if (ret != 0) {
+		dev_err(dev, "Failed to enable fck: %d\n", ret);
+		clk_disable_unprepare(spi100k->ick);
+		return ret;
+	}
+
+	return 0;
+}
+#endif
+
+static const struct dev_pm_ops omap1_spi100k_pm = {
+	SET_RUNTIME_PM_OPS(omap1_spi100k_runtime_suspend,
+			   omap1_spi100k_runtime_resume, NULL)
+};
+
 static struct platform_driver omap1_spi100k_driver = {
 	.driver = {
 		.name		= "omap1_spi100k",
+		.pm		= &omap1_spi100k_pm,
 	},
 	.probe		= omap1_spi100k_probe,
+	.remove		= omap1_spi100k_remove,
 };
 
 module_platform_driver(omap1_spi100k_driver);
diff --git a/drivers/spi/spi-omap-uwire.c b/drivers/spi/spi-omap-uwire.c
index 3c0844457c07..55576db31549 100644
--- a/drivers/spi/spi-omap-uwire.c
+++ b/drivers/spi/spi-omap-uwire.c
@@ -44,7 +44,6 @@
 #include <linux/module.h>
 #include <linux/io.h>
 
-#include <asm/irq.h>
 #include <mach/hardware.h>
 #include <asm/mach-types.h>
 
diff --git a/drivers/spi/spi-omap2-mcspi.c b/drivers/spi/spi-omap2-mcspi.c
index 4df8942058de..58673841286c 100644
--- a/drivers/spi/spi-omap2-mcspi.c
+++ b/drivers/spi/spi-omap2-mcspi.c
@@ -35,6 +35,7 @@
 #include <linux/gcd.h>
 
 #include <linux/spi/spi.h>
+#include <linux/gpio.h>
 
 #include <linux/platform_data/spi-omap2-mcspi.h>
 
@@ -242,17 +243,27 @@ static void omap2_mcspi_set_enable(const struct spi_device *spi, int enable)
 	mcspi_read_cs_reg(spi, OMAP2_MCSPI_CHCTRL0);
 }
 
-static void omap2_mcspi_force_cs(struct spi_device *spi, int cs_active)
+static void omap2_mcspi_set_cs(struct spi_device *spi, bool enable)
 {
 	u32 l;
 
-	l = mcspi_cached_chconf0(spi);
-	if (cs_active)
-		l |= OMAP2_MCSPI_CHCONF_FORCE;
-	else
-		l &= ~OMAP2_MCSPI_CHCONF_FORCE;
+	/* The controller handles the inverted chip selects
+	 * using the OMAP2_MCSPI_CHCONF_EPOL bit so revert
+	 * the inversion from the core spi_set_cs function.
+	 */
+	if (spi->mode & SPI_CS_HIGH)
+		enable = !enable;
 
-	mcspi_write_chconf0(spi, l);
+	if (spi->controller_state) {
+		l = mcspi_cached_chconf0(spi);
+
+		if (enable)
+			l &= ~OMAP2_MCSPI_CHCONF_FORCE;
+		else
+			l |= OMAP2_MCSPI_CHCONF_FORCE;
+
+		mcspi_write_chconf0(spi, l);
+	}
 }
 
 static void omap2_mcspi_set_master_mode(struct spi_master *master)
@@ -1011,6 +1022,15 @@ static int omap2_mcspi_setup(struct spi_device *spi)
 			return ret;
 	}
 
+	if (gpio_is_valid(spi->cs_gpio)) {
+		ret = gpio_request(spi->cs_gpio, dev_name(&spi->dev));
+		if (ret) {
+			dev_err(&spi->dev, "failed to request gpio\n");
+			return ret;
+		}
+		gpio_direction_output(spi->cs_gpio, !(spi->mode & SPI_CS_HIGH));
+	}
+
 	ret = pm_runtime_get_sync(mcspi->dev);
 	if (ret < 0)
 		return ret;
@@ -1050,9 +1070,13 @@ static void omap2_mcspi_cleanup(struct spi_device *spi)
 			mcspi_dma->dma_tx = NULL;
 		}
 	}
+
+	if (gpio_is_valid(spi->cs_gpio))
+		gpio_free(spi->cs_gpio);
 }
 
-static void omap2_mcspi_work(struct omap2_mcspi *mcspi, struct spi_message *m)
+static int omap2_mcspi_work_one(struct omap2_mcspi *mcspi,
+		struct spi_device *spi, struct spi_transfer *t)
 {
 
 	/* We only enable one channel at a time -- the one whose message is
@@ -1062,18 +1086,14 @@ static void omap2_mcspi_work(struct omap2_mcspi *mcspi, struct spi_message *m)
 	 * chipselect with the FORCE bit ... CS != channel enable.
 	 */
 
-	struct spi_device		*spi;
-	struct spi_transfer		*t = NULL;
 	struct spi_master		*master;
 	struct omap2_mcspi_dma		*mcspi_dma;
-	int				cs_active = 0;
 	struct omap2_mcspi_cs		*cs;
 	struct omap2_mcspi_device_config *cd;
 	int				par_override = 0;
 	int				status = 0;
 	u32				chconf;
 
-	spi = m->spi;
 	master = spi->master;
 	mcspi_dma = mcspi->dma_channels + spi->chip_select;
 	cs = spi->controller_state;
@@ -1090,103 +1110,84 @@ static void omap2_mcspi_work(struct omap2_mcspi *mcspi, struct spi_message *m)
 		par_override = 1;
 
 	omap2_mcspi_set_enable(spi, 0);
-	list_for_each_entry(t, &m->transfers, transfer_list) {
-		if (t->tx_buf == NULL && t->rx_buf == NULL && t->len) {
-			status = -EINVAL;
-			break;
-		}
-		if (par_override ||
-		    (t->speed_hz != spi->max_speed_hz) ||
-		    (t->bits_per_word != spi->bits_per_word)) {
-			par_override = 1;
-			status = omap2_mcspi_setup_transfer(spi, t);
-			if (status < 0)
-				break;
-			if (t->speed_hz == spi->max_speed_hz &&
-			    t->bits_per_word == spi->bits_per_word)
-				par_override = 0;
-		}
-		if (cd && cd->cs_per_word) {
-			chconf = mcspi->ctx.modulctrl;
-			chconf &= ~OMAP2_MCSPI_MODULCTRL_SINGLE;
-			mcspi_write_reg(master, OMAP2_MCSPI_MODULCTRL, chconf);
-			mcspi->ctx.modulctrl =
-				mcspi_read_cs_reg(spi, OMAP2_MCSPI_MODULCTRL);
-		}
 
+	if (gpio_is_valid(spi->cs_gpio))
+		omap2_mcspi_set_cs(spi, spi->mode & SPI_CS_HIGH);
 
-		if (!cs_active) {
-			omap2_mcspi_force_cs(spi, 1);
-			cs_active = 1;
-		}
-
-		chconf = mcspi_cached_chconf0(spi);
-		chconf &= ~OMAP2_MCSPI_CHCONF_TRM_MASK;
-		chconf &= ~OMAP2_MCSPI_CHCONF_TURBO;
+	if (par_override ||
+	    (t->speed_hz != spi->max_speed_hz) ||
+	    (t->bits_per_word != spi->bits_per_word)) {
+		par_override = 1;
+		status = omap2_mcspi_setup_transfer(spi, t);
+		if (status < 0)
+			goto out;
+		if (t->speed_hz == spi->max_speed_hz &&
+		    t->bits_per_word == spi->bits_per_word)
+			par_override = 0;
+	}
+	if (cd && cd->cs_per_word) {
+		chconf = mcspi->ctx.modulctrl;
+		chconf &= ~OMAP2_MCSPI_MODULCTRL_SINGLE;
+		mcspi_write_reg(master, OMAP2_MCSPI_MODULCTRL, chconf);
+		mcspi->ctx.modulctrl =
+			mcspi_read_cs_reg(spi, OMAP2_MCSPI_MODULCTRL);
+	}
 
-		if (t->tx_buf == NULL)
-			chconf |= OMAP2_MCSPI_CHCONF_TRM_RX_ONLY;
-		else if (t->rx_buf == NULL)
-			chconf |= OMAP2_MCSPI_CHCONF_TRM_TX_ONLY;
-
-		if (cd && cd->turbo_mode && t->tx_buf == NULL) {
-			/* Turbo mode is for more than one word */
-			if (t->len > ((cs->word_len + 7) >> 3))
-				chconf |= OMAP2_MCSPI_CHCONF_TURBO;
-		}
+	chconf = mcspi_cached_chconf0(spi);
+	chconf &= ~OMAP2_MCSPI_CHCONF_TRM_MASK;
+	chconf &= ~OMAP2_MCSPI_CHCONF_TURBO;
+
+	if (t->tx_buf == NULL)
+		chconf |= OMAP2_MCSPI_CHCONF_TRM_RX_ONLY;
+	else if (t->rx_buf == NULL)
+		chconf |= OMAP2_MCSPI_CHCONF_TRM_TX_ONLY;
+
+	if (cd && cd->turbo_mode && t->tx_buf == NULL) {
+		/* Turbo mode is for more than one word */
+		if (t->len > ((cs->word_len + 7) >> 3))
+			chconf |= OMAP2_MCSPI_CHCONF_TURBO;
+	}
 
-		mcspi_write_chconf0(spi, chconf);
+	mcspi_write_chconf0(spi, chconf);
 
-		if (t->len) {
-			unsigned	count;
+	if (t->len) {
+		unsigned	count;
 
-			if ((mcspi_dma->dma_rx && mcspi_dma->dma_tx) &&
-			    (m->is_dma_mapped || t->len >= DMA_MIN_BYTES))
-				omap2_mcspi_set_fifo(spi, t, 1);
+		if ((mcspi_dma->dma_rx && mcspi_dma->dma_tx) &&
+		    (t->len >= DMA_MIN_BYTES))
+			omap2_mcspi_set_fifo(spi, t, 1);
 
-			omap2_mcspi_set_enable(spi, 1);
+		omap2_mcspi_set_enable(spi, 1);
 
-			/* RX_ONLY mode needs dummy data in TX reg */
-			if (t->tx_buf == NULL)
-				writel_relaxed(0, cs->base
-						+ OMAP2_MCSPI_TX0);
+		/* RX_ONLY mode needs dummy data in TX reg */
+		if (t->tx_buf == NULL)
+			writel_relaxed(0, cs->base
+					+ OMAP2_MCSPI_TX0);
 
-			if ((mcspi_dma->dma_rx && mcspi_dma->dma_tx) &&
-			    (m->is_dma_mapped || t->len >= DMA_MIN_BYTES))
-				count = omap2_mcspi_txrx_dma(spi, t);
-			else
-				count = omap2_mcspi_txrx_pio(spi, t);
-			m->actual_length += count;
+		if ((mcspi_dma->dma_rx && mcspi_dma->dma_tx) &&
+		    (t->len >= DMA_MIN_BYTES))
+			count = omap2_mcspi_txrx_dma(spi, t);
+		else
+			count = omap2_mcspi_txrx_pio(spi, t);
 
-			if (count != t->len) {
-				status = -EIO;
-				break;
-			}
+		if (count != t->len) {
+			status = -EIO;
+			goto out;
 		}
+	}
 
-		if (t->delay_usecs)
-			udelay(t->delay_usecs);
-
-		/* ignore the "leave it on after last xfer" hint */
-		if (t->cs_change) {
-			omap2_mcspi_force_cs(spi, 0);
-			cs_active = 0;
-		}
+	omap2_mcspi_set_enable(spi, 0);
 
-		omap2_mcspi_set_enable(spi, 0);
+	if (mcspi->fifo_depth > 0)
+		omap2_mcspi_set_fifo(spi, t, 0);
 
-		if (mcspi->fifo_depth > 0)
-			omap2_mcspi_set_fifo(spi, t, 0);
-	}
+out:
 	/* Restore defaults if they were overriden */
 	if (par_override) {
 		par_override = 0;
 		status = omap2_mcspi_setup_transfer(spi, NULL);
 	}
 
-	if (cs_active)
-		omap2_mcspi_force_cs(spi, 0);
-
 	if (cd && cd->cs_per_word) {
 		chconf = mcspi->ctx.modulctrl;
 		chconf |= OMAP2_MCSPI_MODULCTRL_SINGLE;
@@ -1197,70 +1198,64 @@ static void omap2_mcspi_work(struct omap2_mcspi *mcspi, struct spi_message *m)
 
 	omap2_mcspi_set_enable(spi, 0);
 
+	if (gpio_is_valid(spi->cs_gpio))
+		omap2_mcspi_set_cs(spi, !(spi->mode & SPI_CS_HIGH));
+
 	if (mcspi->fifo_depth > 0 && t)
 		omap2_mcspi_set_fifo(spi, t, 0);
 
-	m->status = status;
+	return status;
 }
 
-static int omap2_mcspi_transfer_one_message(struct spi_master *master,
-		struct spi_message *m)
+static int omap2_mcspi_transfer_one(struct spi_master *master,
+		struct spi_device *spi, struct spi_transfer *t)
 {
-	struct spi_device	*spi;
 	struct omap2_mcspi	*mcspi;
 	struct omap2_mcspi_dma	*mcspi_dma;
-	struct spi_transfer	*t;
+	const void	*tx_buf = t->tx_buf;
+	void		*rx_buf = t->rx_buf;
+	unsigned	len = t->len;
 
-	spi = m->spi;
 	mcspi = spi_master_get_devdata(master);
 	mcspi_dma = mcspi->dma_channels + spi->chip_select;
-	m->actual_length = 0;
-	m->status = 0;
-
-	list_for_each_entry(t, &m->transfers, transfer_list) {
-		const void	*tx_buf = t->tx_buf;
-		void		*rx_buf = t->rx_buf;
-		unsigned	len = t->len;
-
-		if ((len && !(rx_buf || tx_buf))) {
-			dev_dbg(mcspi->dev, "transfer: %d Hz, %d %s%s, %d bpw\n",
-					t->speed_hz,
-					len,
-					tx_buf ? "tx" : "",
-					rx_buf ? "rx" : "",
-					t->bits_per_word);
-			return -EINVAL;
-		}
 
-		if (m->is_dma_mapped || len < DMA_MIN_BYTES)
-			continue;
+	if ((len && !(rx_buf || tx_buf))) {
+		dev_dbg(mcspi->dev, "transfer: %d Hz, %d %s%s, %d bpw\n",
+				t->speed_hz,
+				len,
+				tx_buf ? "tx" : "",
+				rx_buf ? "rx" : "",
+				t->bits_per_word);
+		return -EINVAL;
+	}
 
-		if (mcspi_dma->dma_tx && tx_buf != NULL) {
-			t->tx_dma = dma_map_single(mcspi->dev, (void *) tx_buf,
-					len, DMA_TO_DEVICE);
-			if (dma_mapping_error(mcspi->dev, t->tx_dma)) {
-				dev_dbg(mcspi->dev, "dma %cX %d bytes error\n",
-						'T', len);
-				return -EINVAL;
-			}
+	if (len < DMA_MIN_BYTES)
+		goto skip_dma_map;
+
+	if (mcspi_dma->dma_tx && tx_buf != NULL) {
+		t->tx_dma = dma_map_single(mcspi->dev, (void *) tx_buf,
+				len, DMA_TO_DEVICE);
+		if (dma_mapping_error(mcspi->dev, t->tx_dma)) {
+			dev_dbg(mcspi->dev, "dma %cX %d bytes error\n",
+					'T', len);
+			return -EINVAL;
 		}
-		if (mcspi_dma->dma_rx && rx_buf != NULL) {
-			t->rx_dma = dma_map_single(mcspi->dev, rx_buf, t->len,
-					DMA_FROM_DEVICE);
-			if (dma_mapping_error(mcspi->dev, t->rx_dma)) {
-				dev_dbg(mcspi->dev, "dma %cX %d bytes error\n",
-						'R', len);
-				if (tx_buf != NULL)
-					dma_unmap_single(mcspi->dev, t->tx_dma,
-							len, DMA_TO_DEVICE);
-				return -EINVAL;
-			}
+	}
+	if (mcspi_dma->dma_rx && rx_buf != NULL) {
+		t->rx_dma = dma_map_single(mcspi->dev, rx_buf, t->len,
+				DMA_FROM_DEVICE);
+		if (dma_mapping_error(mcspi->dev, t->rx_dma)) {
+			dev_dbg(mcspi->dev, "dma %cX %d bytes error\n",
+					'R', len);
+			if (tx_buf != NULL)
+				dma_unmap_single(mcspi->dev, t->tx_dma,
+						len, DMA_TO_DEVICE);
+			return -EINVAL;
 		}
 	}
 
-	omap2_mcspi_work(mcspi, m);
-	spi_finalize_current_message(master);
-	return 0;
+skip_dma_map:
+	return omap2_mcspi_work_one(mcspi, spi, t);
 }
 
 static int omap2_mcspi_master_setup(struct omap2_mcspi *mcspi)
@@ -1339,7 +1334,8 @@ static int omap2_mcspi_probe(struct platform_device *pdev)
 	master->bits_per_word_mask = SPI_BPW_RANGE_MASK(4, 32);
 	master->setup = omap2_mcspi_setup;
 	master->auto_runtime_pm = true;
-	master->transfer_one_message = omap2_mcspi_transfer_one_message;
+	master->transfer_one = omap2_mcspi_transfer_one;
+	master->set_cs = omap2_mcspi_set_cs;
 	master->cleanup = omap2_mcspi_cleanup;
 	master->dev.of_node = node;
 	master->max_speed_hz = OMAP2_MCSPI_MAX_FREQ;
diff --git a/drivers/spi/spi-orion.c b/drivers/spi/spi-orion.c
index 861664776672..8cad107a5b3f 100644
--- a/drivers/spi/spi-orion.c
+++ b/drivers/spi/spi-orion.c
@@ -61,6 +61,12 @@ enum orion_spi_type {
 
 struct orion_spi_dev {
 	enum orion_spi_type	typ;
+	/*
+	 * min_divisor and max_hz should be exclusive, the only we can
+	 * have both is for managing the armada-370-spi case with old
+	 * device tree
+	 */
+	unsigned long		max_hz;
 	unsigned int		min_divisor;
 	unsigned int		max_divisor;
 	u32			prescale_mask;
@@ -385,16 +391,54 @@ static const struct orion_spi_dev orion_spi_dev_data = {
 	.prescale_mask = ORION_SPI_CLK_PRESCALE_MASK,
 };
 
-static const struct orion_spi_dev armada_spi_dev_data = {
+static const struct orion_spi_dev armada_370_spi_dev_data = {
 	.typ = ARMADA_SPI,
-	.min_divisor = 1,
+	.min_divisor = 4,
+	.max_divisor = 1920,
+	.max_hz = 50000000,
+	.prescale_mask = ARMADA_SPI_CLK_PRESCALE_MASK,
+};
+
+static const struct orion_spi_dev armada_xp_spi_dev_data = {
+	.typ = ARMADA_SPI,
+	.max_hz = 50000000,
+	.max_divisor = 1920,
+	.prescale_mask = ARMADA_SPI_CLK_PRESCALE_MASK,
+};
+
+static const struct orion_spi_dev armada_375_spi_dev_data = {
+	.typ = ARMADA_SPI,
+	.min_divisor = 15,
 	.max_divisor = 1920,
 	.prescale_mask = ARMADA_SPI_CLK_PRESCALE_MASK,
 };
 
 static const struct of_device_id orion_spi_of_match_table[] = {
-	{ .compatible = "marvell,orion-spi", .data = &orion_spi_dev_data, },
-	{ .compatible = "marvell,armada-370-spi", .data = &armada_spi_dev_data, },
+	{
+		.compatible = "marvell,orion-spi",
+		.data = &orion_spi_dev_data,
+	},
+	{
+		.compatible = "marvell,armada-370-spi",
+		.data = &armada_370_spi_dev_data,
+	},
+	{
+		.compatible = "marvell,armada-375-spi",
+		.data = &armada_375_spi_dev_data,
+	},
+	{
+		.compatible = "marvell,armada-380-spi",
+		.data = &armada_xp_spi_dev_data,
+	},
+	{
+		.compatible = "marvell,armada-390-spi",
+		.data = &armada_xp_spi_dev_data,
+	},
+	{
+		.compatible = "marvell,armada-xp-spi",
+		.data = &armada_xp_spi_dev_data,
+	},
+
 	{}
 };
 MODULE_DEVICE_TABLE(of, orion_spi_of_match_table);
@@ -454,7 +498,23 @@ static int orion_spi_probe(struct platform_device *pdev)
 		goto out;
 
 	tclk_hz = clk_get_rate(spi->clk);
-	master->max_speed_hz = DIV_ROUND_UP(tclk_hz, devdata->min_divisor);
+
+	/*
+	 * With old device tree, armada-370-spi could be used with
+	 * Armada XP, however for this SoC the maximum frequency is
+	 * 50MHz instead of tclk/4. On Armada 370, tclk cannot be
+	 * higher than 200MHz. So, in order to be able to handle both
+	 * SoCs, we can take the minimum of 50MHz and tclk/4.
+	 */
+	if (of_device_is_compatible(pdev->dev.of_node,
+					"marvell,armada-370-spi"))
+		master->max_speed_hz = min(devdata->max_hz,
+				DIV_ROUND_UP(tclk_hz, devdata->min_divisor));
+	else if (devdata->min_divisor)
+		master->max_speed_hz =
+			DIV_ROUND_UP(tclk_hz, devdata->min_divisor);
+	else
+		master->max_speed_hz = devdata->max_hz;
 	master->min_speed_hz = DIV_ROUND_UP(tclk_hz, devdata->max_divisor);
 
 	r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
diff --git a/drivers/spi/spi-pl022.c b/drivers/spi/spi-pl022.c
index ee513a85296b..94af80676684 100644
--- a/drivers/spi/spi-pl022.c
+++ b/drivers/spi/spi-pl022.c
@@ -285,7 +285,12 @@
  */
 #define DEFAULT_SSP_REG_IMSC  0x0UL
 #define DISABLE_ALL_INTERRUPTS DEFAULT_SSP_REG_IMSC
-#define ENABLE_ALL_INTERRUPTS (~DEFAULT_SSP_REG_IMSC)
+#define ENABLE_ALL_INTERRUPTS ( \
+	SSP_IMSC_MASK_RORIM | \
+	SSP_IMSC_MASK_RTIM | \
+	SSP_IMSC_MASK_RXIM | \
+	SSP_IMSC_MASK_TXIM \
+)
 
 #define CLEAR_ALL_INTERRUPTS  0x3
 
@@ -1251,7 +1256,6 @@ static irqreturn_t pl022_interrupt_handler(int irq, void *dev_id)
 	struct pl022 *pl022 = dev_id;
 	struct spi_message *msg = pl022->cur_msg;
 	u16 irq_status = 0;
-	u16 flag = 0;
 
 	if (unlikely(!msg)) {
 		dev_err(&pl022->adev->dev,
@@ -1280,9 +1284,6 @@ static irqreturn_t pl022_interrupt_handler(int irq, void *dev_id)
 		if (readw(SSP_SR(pl022->virtbase)) & SSP_SR_MASK_RFF)
 			dev_err(&pl022->adev->dev,
 				"RXFIFO is full\n");
-		if (readw(SSP_SR(pl022->virtbase)) & SSP_SR_MASK_TNF)
-			dev_err(&pl022->adev->dev,
-				"TXFIFO is full\n");
 
 		/*
 		 * Disable and clear interrupts, disable SSP,
@@ -1303,8 +1304,7 @@ static irqreturn_t pl022_interrupt_handler(int irq, void *dev_id)
 
 	readwriter(pl022);
 
-	if ((pl022->tx == pl022->tx_end) && (flag == 0)) {
-		flag = 1;
+	if (pl022->tx == pl022->tx_end) {
 		/* Disable Transmit interrupt, enable receive interrupt */
 		writew((readw(SSP_IMSC(pl022->virtbase)) &
 		       ~SSP_IMSC_MASK_TXIM) | SSP_IMSC_MASK_RXIM,
diff --git a/drivers/spi/spi-pxa2xx-pci.c b/drivers/spi/spi-pxa2xx-pci.c
index fa7399e84bbb..3cfd4357489a 100644
--- a/drivers/spi/spi-pxa2xx-pci.c
+++ b/drivers/spi/spi-pxa2xx-pci.c
@@ -62,7 +62,7 @@ static struct pxa_spi_info spi_info_configs[] = {
 		.max_clk_rate = 3686400,
 	},
 	[PORT_BYT] = {
-		.type = LPSS_SSP,
+		.type = LPSS_BYT_SSP,
 		.port_id = 0,
 		.num_chipselect = 1,
 		.max_clk_rate = 50000000,
@@ -70,7 +70,7 @@ static struct pxa_spi_info spi_info_configs[] = {
 		.rx_param = &byt_rx_param,
 	},
 	[PORT_BSW0] = {
-		.type = LPSS_SSP,
+		.type = LPSS_BYT_SSP,
 		.port_id = 0,
 		.num_chipselect = 1,
 		.max_clk_rate = 50000000,
@@ -78,7 +78,7 @@ static struct pxa_spi_info spi_info_configs[] = {
 		.rx_param = &bsw0_rx_param,
 	},
 	[PORT_BSW1] = {
-		.type = LPSS_SSP,
+		.type = LPSS_BYT_SSP,
 		.port_id = 1,
 		.num_chipselect = 1,
 		.max_clk_rate = 50000000,
@@ -86,7 +86,7 @@ static struct pxa_spi_info spi_info_configs[] = {
 		.rx_param = &bsw1_rx_param,
 	},
 	[PORT_BSW2] = {
-		.type = LPSS_SSP,
+		.type = LPSS_BYT_SSP,
 		.port_id = 2,
 		.num_chipselect = 1,
 		.max_clk_rate = 50000000,
diff --git a/drivers/spi/spi-pxa2xx-pxadma.c b/drivers/spi/spi-pxa2xx-pxadma.c
deleted file mode 100644
index 2e0796a0003f..000000000000
--- a/drivers/spi/spi-pxa2xx-pxadma.c
+++ /dev/null
@@ -1,487 +0,0 @@
-/*
- * PXA2xx SPI private DMA support.
- *
- * Copyright (C) 2005 Stephen Street / StreetFire Sound Labs
- *
- * 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.
- */
-
-#include <linux/delay.h>
-#include <linux/device.h>
-#include <linux/dma-mapping.h>
-#include <linux/pxa2xx_ssp.h>
-#include <linux/spi/spi.h>
-#include <linux/spi/pxa2xx_spi.h>
-
-#include <mach/dma.h>
-#include "spi-pxa2xx.h"
-
-#define DMA_INT_MASK		(DCSR_ENDINTR | DCSR_STARTINTR | DCSR_BUSERR)
-#define RESET_DMA_CHANNEL	(DCSR_NODESC | DMA_INT_MASK)
-
-bool pxa2xx_spi_dma_is_possible(size_t len)
-{
-	/* Try to map dma buffer and do a dma transfer if successful, but
-	 * only if the length is non-zero and less than MAX_DMA_LEN.
-	 *
-	 * Zero-length non-descriptor DMA is illegal on PXA2xx; force use
-	 * of PIO instead.  Care is needed above because the transfer may
-	 * have have been passed with buffers that are already dma mapped.
-	 * A zero-length transfer in PIO mode will not try to write/read
-	 * to/from the buffers
-	 *
-	 * REVISIT large transfers are exactly where we most want to be
-	 * using DMA.  If this happens much, split those transfers into
-	 * multiple DMA segments rather than forcing PIO.
-	 */
-	return len > 0 && len <= MAX_DMA_LEN;
-}
-
-int pxa2xx_spi_map_dma_buffers(struct driver_data *drv_data)
-{
-	struct spi_message *msg = drv_data->cur_msg;
-	struct device *dev = &msg->spi->dev;
-
-	if (!drv_data->cur_chip->enable_dma)
-		return 0;
-
-	if (msg->is_dma_mapped)
-		return  drv_data->rx_dma && drv_data->tx_dma;
-
-	if (!IS_DMA_ALIGNED(drv_data->rx) || !IS_DMA_ALIGNED(drv_data->tx))
-		return 0;
-
-	/* Modify setup if rx buffer is null */
-	if (drv_data->rx == NULL) {
-		*drv_data->null_dma_buf = 0;
-		drv_data->rx = drv_data->null_dma_buf;
-		drv_data->rx_map_len = 4;
-	} else
-		drv_data->rx_map_len = drv_data->len;
-
-
-	/* Modify setup if tx buffer is null */
-	if (drv_data->tx == NULL) {
-		*drv_data->null_dma_buf = 0;
-		drv_data->tx = drv_data->null_dma_buf;
-		drv_data->tx_map_len = 4;
-	} else
-		drv_data->tx_map_len = drv_data->len;
-
-	/* Stream map the tx buffer. Always do DMA_TO_DEVICE first
-	 * so we flush the cache *before* invalidating it, in case
-	 * the tx and rx buffers overlap.
-	 */
-	drv_data->tx_dma = dma_map_single(dev, drv_data->tx,
-					drv_data->tx_map_len, DMA_TO_DEVICE);
-	if (dma_mapping_error(dev, drv_data->tx_dma))
-		return 0;
-
-	/* Stream map the rx buffer */
-	drv_data->rx_dma = dma_map_single(dev, drv_data->rx,
-					drv_data->rx_map_len, DMA_FROM_DEVICE);
-	if (dma_mapping_error(dev, drv_data->rx_dma)) {
-		dma_unmap_single(dev, drv_data->tx_dma,
-					drv_data->tx_map_len, DMA_TO_DEVICE);
-		return 0;
-	}
-
-	return 1;
-}
-
-static void pxa2xx_spi_unmap_dma_buffers(struct driver_data *drv_data)
-{
-	struct device *dev;
-
-	if (!drv_data->dma_mapped)
-		return;
-
-	if (!drv_data->cur_msg->is_dma_mapped) {
-		dev = &drv_data->cur_msg->spi->dev;
-		dma_unmap_single(dev, drv_data->rx_dma,
-					drv_data->rx_map_len, DMA_FROM_DEVICE);
-		dma_unmap_single(dev, drv_data->tx_dma,
-					drv_data->tx_map_len, DMA_TO_DEVICE);
-	}
-
-	drv_data->dma_mapped = 0;
-}
-
-static int wait_ssp_rx_stall(struct driver_data *drv_data)
-{
-	unsigned long limit = loops_per_jiffy << 1;
-
-	while ((pxa2xx_spi_read(drv_data, SSSR) & SSSR_BSY) && --limit)
-		cpu_relax();
-
-	return limit;
-}
-
-static int wait_dma_channel_stop(int channel)
-{
-	unsigned long limit = loops_per_jiffy << 1;
-
-	while (!(DCSR(channel) & DCSR_STOPSTATE) && --limit)
-		cpu_relax();
-
-	return limit;
-}
-
-static void pxa2xx_spi_dma_error_stop(struct driver_data *drv_data,
-				      const char *msg)
-{
-	/* Stop and reset */
-	DCSR(drv_data->rx_channel) = RESET_DMA_CHANNEL;
-	DCSR(drv_data->tx_channel) = RESET_DMA_CHANNEL;
-	write_SSSR_CS(drv_data, drv_data->clear_sr);
-	pxa2xx_spi_write(drv_data, SSCR1,
-			 pxa2xx_spi_read(drv_data, SSCR1)
-			 & ~drv_data->dma_cr1);
-	if (!pxa25x_ssp_comp(drv_data))
-		pxa2xx_spi_write(drv_data, SSTO, 0);
-	pxa2xx_spi_flush(drv_data);
-	pxa2xx_spi_write(drv_data, SSCR0,
-			 pxa2xx_spi_read(drv_data, SSCR0) & ~SSCR0_SSE);
-
-	pxa2xx_spi_unmap_dma_buffers(drv_data);
-
-	dev_err(&drv_data->pdev->dev, "%s\n", msg);
-
-	drv_data->cur_msg->state = ERROR_STATE;
-	tasklet_schedule(&drv_data->pump_transfers);
-}
-
-static void pxa2xx_spi_dma_transfer_complete(struct driver_data *drv_data)
-{
-	struct spi_message *msg = drv_data->cur_msg;
-
-	/* Clear and disable interrupts on SSP and DMA channels*/
-	pxa2xx_spi_write(drv_data, SSCR1,
-			 pxa2xx_spi_read(drv_data, SSCR1)
-			 & ~drv_data->dma_cr1);
-	write_SSSR_CS(drv_data, drv_data->clear_sr);
-	DCSR(drv_data->tx_channel) = RESET_DMA_CHANNEL;
-	DCSR(drv_data->rx_channel) = RESET_DMA_CHANNEL;
-
-	if (wait_dma_channel_stop(drv_data->rx_channel) == 0)
-		dev_err(&drv_data->pdev->dev,
-			"dma_handler: dma rx channel stop failed\n");
-
-	if (wait_ssp_rx_stall(drv_data->ioaddr) == 0)
-		dev_err(&drv_data->pdev->dev,
-			"dma_transfer: ssp rx stall failed\n");
-
-	pxa2xx_spi_unmap_dma_buffers(drv_data);
-
-	/* update the buffer pointer for the amount completed in dma */
-	drv_data->rx += drv_data->len -
-			(DCMD(drv_data->rx_channel) & DCMD_LENGTH);
-
-	/* read trailing data from fifo, it does not matter how many
-	 * bytes are in the fifo just read until buffer is full
-	 * or fifo is empty, which ever occurs first */
-	drv_data->read(drv_data);
-
-	/* return count of what was actually read */
-	msg->actual_length += drv_data->len -
-				(drv_data->rx_end - drv_data->rx);
-
-	/* Transfer delays and chip select release are
-	 * handled in pump_transfers or giveback
-	 */
-
-	/* Move to next transfer */
-	msg->state = pxa2xx_spi_next_transfer(drv_data);
-
-	/* Schedule transfer tasklet */
-	tasklet_schedule(&drv_data->pump_transfers);
-}
-
-void pxa2xx_spi_dma_handler(int channel, void *data)
-{
-	struct driver_data *drv_data = data;
-	u32 irq_status = DCSR(channel) & DMA_INT_MASK;
-
-	if (irq_status & DCSR_BUSERR) {
-
-		if (channel == drv_data->tx_channel)
-			pxa2xx_spi_dma_error_stop(drv_data,
-				"dma_handler: bad bus address on tx channel");
-		else
-			pxa2xx_spi_dma_error_stop(drv_data,
-				"dma_handler: bad bus address on rx channel");
-		return;
-	}
-
-	/* PXA255x_SSP has no timeout interrupt, wait for tailing bytes */
-	if ((channel == drv_data->tx_channel)
-		&& (irq_status & DCSR_ENDINTR)
-		&& (drv_data->ssp_type == PXA25x_SSP)) {
-
-		/* Wait for rx to stall */
-		if (wait_ssp_rx_stall(drv_data) == 0)
-			dev_err(&drv_data->pdev->dev,
-				"dma_handler: ssp rx stall failed\n");
-
-		/* finish this transfer, start the next */
-		pxa2xx_spi_dma_transfer_complete(drv_data);
-	}
-}
-
-irqreturn_t pxa2xx_spi_dma_transfer(struct driver_data *drv_data)
-{
-	u32 irq_status;
-
-	irq_status = pxa2xx_spi_read(drv_data, SSSR) & drv_data->mask_sr;
-	if (irq_status & SSSR_ROR) {
-		pxa2xx_spi_dma_error_stop(drv_data,
-					  "dma_transfer: fifo overrun");
-		return IRQ_HANDLED;
-	}
-
-	/* Check for false positive timeout */
-	if ((irq_status & SSSR_TINT)
-		&& (DCSR(drv_data->tx_channel) & DCSR_RUN)) {
-		pxa2xx_spi_write(drv_data, SSSR, SSSR_TINT);
-		return IRQ_HANDLED;
-	}
-
-	if (irq_status & SSSR_TINT || drv_data->rx == drv_data->rx_end) {
-
-		/* Clear and disable timeout interrupt, do the rest in
-		 * dma_transfer_complete */
-		if (!pxa25x_ssp_comp(drv_data))
-			pxa2xx_spi_write(drv_data, SSTO, 0);
-
-		/* finish this transfer, start the next */
-		pxa2xx_spi_dma_transfer_complete(drv_data);
-
-		return IRQ_HANDLED;
-	}
-
-	/* Opps problem detected */
-	return IRQ_NONE;
-}
-
-int pxa2xx_spi_dma_prepare(struct driver_data *drv_data, u32 dma_burst)
-{
-	u32 dma_width;
-
-	switch (drv_data->n_bytes) {
-	case 1:
-		dma_width = DCMD_WIDTH1;
-		break;
-	case 2:
-		dma_width = DCMD_WIDTH2;
-		break;
-	default:
-		dma_width = DCMD_WIDTH4;
-		break;
-	}
-
-	/* Setup rx DMA Channel */
-	DCSR(drv_data->rx_channel) = RESET_DMA_CHANNEL;
-	DSADR(drv_data->rx_channel) = drv_data->ssdr_physical;
-	DTADR(drv_data->rx_channel) = drv_data->rx_dma;
-	if (drv_data->rx == drv_data->null_dma_buf)
-		/* No target address increment */
-		DCMD(drv_data->rx_channel) = DCMD_FLOWSRC
-						| dma_width
-						| dma_burst
-						| drv_data->len;
-	else
-		DCMD(drv_data->rx_channel) = DCMD_INCTRGADDR
-						| DCMD_FLOWSRC
-						| dma_width
-						| dma_burst
-						| drv_data->len;
-
-	/* Setup tx DMA Channel */
-	DCSR(drv_data->tx_channel) = RESET_DMA_CHANNEL;
-	DSADR(drv_data->tx_channel) = drv_data->tx_dma;
-	DTADR(drv_data->tx_channel) = drv_data->ssdr_physical;
-	if (drv_data->tx == drv_data->null_dma_buf)
-		/* No source address increment */
-		DCMD(drv_data->tx_channel) = DCMD_FLOWTRG
-						| dma_width
-						| dma_burst
-						| drv_data->len;
-	else
-		DCMD(drv_data->tx_channel) = DCMD_INCSRCADDR
-						| DCMD_FLOWTRG
-						| dma_width
-						| dma_burst
-						| drv_data->len;
-
-	/* Enable dma end irqs on SSP to detect end of transfer */
-	if (drv_data->ssp_type == PXA25x_SSP)
-		DCMD(drv_data->tx_channel) |= DCMD_ENDIRQEN;
-
-	return 0;
-}
-
-void pxa2xx_spi_dma_start(struct driver_data *drv_data)
-{
-	DCSR(drv_data->rx_channel) |= DCSR_RUN;
-	DCSR(drv_data->tx_channel) |= DCSR_RUN;
-}
-
-int pxa2xx_spi_dma_setup(struct driver_data *drv_data)
-{
-	struct device *dev = &drv_data->pdev->dev;
-	struct ssp_device *ssp = drv_data->ssp;
-
-	/* Get two DMA channels	(rx and tx) */
-	drv_data->rx_channel = pxa_request_dma("pxa2xx_spi_ssp_rx",
-						DMA_PRIO_HIGH,
-						pxa2xx_spi_dma_handler,
-						drv_data);
-	if (drv_data->rx_channel < 0) {
-		dev_err(dev, "problem (%d) requesting rx channel\n",
-			drv_data->rx_channel);
-		return -ENODEV;
-	}
-	drv_data->tx_channel = pxa_request_dma("pxa2xx_spi_ssp_tx",
-						DMA_PRIO_MEDIUM,
-						pxa2xx_spi_dma_handler,
-						drv_data);
-	if (drv_data->tx_channel < 0) {
-		dev_err(dev, "problem (%d) requesting tx channel\n",
-			drv_data->tx_channel);
-		pxa_free_dma(drv_data->rx_channel);
-		return -ENODEV;
-	}
-
-	DRCMR(ssp->drcmr_rx) = DRCMR_MAPVLD | drv_data->rx_channel;
-	DRCMR(ssp->drcmr_tx) = DRCMR_MAPVLD | drv_data->tx_channel;
-
-	return 0;
-}
-
-void pxa2xx_spi_dma_release(struct driver_data *drv_data)
-{
-	struct ssp_device *ssp = drv_data->ssp;
-
-	DRCMR(ssp->drcmr_rx) = 0;
-	DRCMR(ssp->drcmr_tx) = 0;
-
-	if (drv_data->tx_channel != 0)
-		pxa_free_dma(drv_data->tx_channel);
-	if (drv_data->rx_channel != 0)
-		pxa_free_dma(drv_data->rx_channel);
-}
-
-void pxa2xx_spi_dma_resume(struct driver_data *drv_data)
-{
-	if (drv_data->rx_channel != -1)
-		DRCMR(drv_data->ssp->drcmr_rx) =
-			DRCMR_MAPVLD | drv_data->rx_channel;
-	if (drv_data->tx_channel != -1)
-		DRCMR(drv_data->ssp->drcmr_tx) =
-			DRCMR_MAPVLD | drv_data->tx_channel;
-}
-
-int pxa2xx_spi_set_dma_burst_and_threshold(struct chip_data *chip,
-					   struct spi_device *spi,
-					   u8 bits_per_word, u32 *burst_code,
-					   u32 *threshold)
-{
-	struct pxa2xx_spi_chip *chip_info =
-			(struct pxa2xx_spi_chip *)spi->controller_data;
-	int bytes_per_word;
-	int burst_bytes;
-	int thresh_words;
-	int req_burst_size;
-	int retval = 0;
-
-	/* Set the threshold (in registers) to equal the same amount of data
-	 * as represented by burst size (in bytes).  The computation below
-	 * is (burst_size rounded up to nearest 8 byte, word or long word)
-	 * divided by (bytes/register); the tx threshold is the inverse of
-	 * the rx, so that there will always be enough data in the rx fifo
-	 * to satisfy a burst, and there will always be enough space in the
-	 * tx fifo to accept a burst (a tx burst will overwrite the fifo if
-	 * there is not enough space), there must always remain enough empty
-	 * space in the rx fifo for any data loaded to the tx fifo.
-	 * Whenever burst_size (in bytes) equals bits/word, the fifo threshold
-	 * will be 8, or half the fifo;
-	 * The threshold can only be set to 2, 4 or 8, but not 16, because
-	 * to burst 16 to the tx fifo, the fifo would have to be empty;
-	 * however, the minimum fifo trigger level is 1, and the tx will
-	 * request service when the fifo is at this level, with only 15 spaces.
-	 */
-
-	/* find bytes/word */
-	if (bits_per_word <= 8)
-		bytes_per_word = 1;
-	else if (bits_per_word <= 16)
-		bytes_per_word = 2;
-	else
-		bytes_per_word = 4;
-
-	/* use struct pxa2xx_spi_chip->dma_burst_size if available */
-	if (chip_info)
-		req_burst_size = chip_info->dma_burst_size;
-	else {
-		switch (chip->dma_burst_size) {
-		default:
-			/* if the default burst size is not set,
-			 * do it now */
-			chip->dma_burst_size = DCMD_BURST8;
-		case DCMD_BURST8:
-			req_burst_size = 8;
-			break;
-		case DCMD_BURST16:
-			req_burst_size = 16;
-			break;
-		case DCMD_BURST32:
-			req_burst_size = 32;
-			break;
-		}
-	}
-	if (req_burst_size <= 8) {
-		*burst_code = DCMD_BURST8;
-		burst_bytes = 8;
-	} else if (req_burst_size <= 16) {
-		if (bytes_per_word == 1) {
-			/* don't burst more than 1/2 the fifo */
-			*burst_code = DCMD_BURST8;
-			burst_bytes = 8;
-			retval = 1;
-		} else {
-			*burst_code = DCMD_BURST16;
-			burst_bytes = 16;
-		}
-	} else {
-		if (bytes_per_word == 1) {
-			/* don't burst more than 1/2 the fifo */
-			*burst_code = DCMD_BURST8;
-			burst_bytes = 8;
-			retval = 1;
-		} else if (bytes_per_word == 2) {
-			/* don't burst more than 1/2 the fifo */
-			*burst_code = DCMD_BURST16;
-			burst_bytes = 16;
-			retval = 1;
-		} else {
-			*burst_code = DCMD_BURST32;
-			burst_bytes = 32;
-		}
-	}
-
-	thresh_words = burst_bytes / bytes_per_word;
-
-	/* thresh_words will be between 2 and 8 */
-	*threshold = (SSCR1_RxTresh(thresh_words) & SSCR1_RFT)
-			| (SSCR1_TxTresh(16-thresh_words) & SSCR1_TFT);
-
-	return retval;
-}
diff --git a/drivers/spi/spi-pxa2xx.c b/drivers/spi/spi-pxa2xx.c
index 6f72ad01e041..7293d6d875c5 100644
--- a/drivers/spi/spi-pxa2xx.c
+++ b/drivers/spi/spi-pxa2xx.c
@@ -20,6 +20,7 @@
 #include <linux/errno.h>
 #include <linux/err.h>
 #include <linux/interrupt.h>
+#include <linux/kernel.h>
 #include <linux/platform_device.h>
 #include <linux/spi/pxa2xx_spi.h>
 #include <linux/spi/spi.h>
@@ -30,10 +31,6 @@
 #include <linux/pm_runtime.h>
 #include <linux/acpi.h>
 
-#include <asm/io.h>
-#include <asm/irq.h>
-#include <asm/delay.h>
-
 #include "spi-pxa2xx.h"
 
 MODULE_AUTHOR("Stephen Street");
@@ -63,69 +60,60 @@ MODULE_ALIAS("platform:pxa2xx-spi");
 				| QUARK_X1000_SSCR1_TFT		\
 				| SSCR1_SPH | SSCR1_SPO | SSCR1_LBM)
 
-#define LPSS_RX_THRESH_DFLT	64
-#define LPSS_TX_LOTHRESH_DFLT	160
-#define LPSS_TX_HITHRESH_DFLT	224
+#define GENERAL_REG_RXTO_HOLDOFF_DISABLE BIT(24)
+#define SPI_CS_CONTROL_SW_MODE	BIT(0)
+#define SPI_CS_CONTROL_CS_HIGH	BIT(1)
 
-struct quark_spi_rate {
-	u32 bitrate;
-	u32 dds_clk_rate;
-	u32 clk_div;
+struct lpss_config {
+	/* LPSS offset from drv_data->ioaddr */
+	unsigned offset;
+	/* Register offsets from drv_data->lpss_base or -1 */
+	int reg_general;
+	int reg_ssp;
+	int reg_cs_ctrl;
+	/* FIFO thresholds */
+	u32 rx_threshold;
+	u32 tx_threshold_lo;
+	u32 tx_threshold_hi;
 };
 
-/*
- * 'rate', 'dds', 'clk_div' lookup table, which is defined in
- * the Quark SPI datasheet.
- */
-static const struct quark_spi_rate quark_spi_rate_table[] = {
-/*	bitrate,	dds_clk_rate,	clk_div */
-	{50000000,	0x800000,	0},
-	{40000000,	0x666666,	0},
-	{25000000,	0x400000,	0},
-	{20000000,	0x666666,	1},
-	{16667000,	0x800000,	2},
-	{13333000,	0x666666,	2},
-	{12500000,	0x200000,	0},
-	{10000000,	0x800000,	4},
-	{8000000,	0x666666,	4},
-	{6250000,	0x400000,	3},
-	{5000000,	0x400000,	4},
-	{4000000,	0x666666,	9},
-	{3125000,	0x80000,	0},
-	{2500000,	0x400000,	9},
-	{2000000,	0x666666,	19},
-	{1563000,	0x40000,	0},
-	{1250000,	0x200000,	9},
-	{1000000,	0x400000,	24},
-	{800000,	0x666666,	49},
-	{781250,	0x20000,	0},
-	{625000,	0x200000,	19},
-	{500000,	0x400000,	49},
-	{400000,	0x666666,	99},
-	{390625,	0x10000,	0},
-	{250000,	0x400000,	99},
-	{200000,	0x666666,	199},
-	{195313,	0x8000,		0},
-	{125000,	0x100000,	49},
-	{100000,	0x200000,	124},
-	{50000,		0x100000,	124},
-	{25000,		0x80000,	124},
-	{10016,		0x20000,	77},
-	{5040,		0x20000,	154},
-	{1002,		0x8000,		194},
+/* Keep these sorted with enum pxa_ssp_type */
+static const struct lpss_config lpss_platforms[] = {
+	{	/* LPSS_LPT_SSP */
+		.offset = 0x800,
+		.reg_general = 0x08,
+		.reg_ssp = 0x0c,
+		.reg_cs_ctrl = 0x18,
+		.rx_threshold = 64,
+		.tx_threshold_lo = 160,
+		.tx_threshold_hi = 224,
+	},
+	{	/* LPSS_BYT_SSP */
+		.offset = 0x400,
+		.reg_general = 0x08,
+		.reg_ssp = 0x0c,
+		.reg_cs_ctrl = 0x18,
+		.rx_threshold = 64,
+		.tx_threshold_lo = 160,
+		.tx_threshold_hi = 224,
+	},
 };
 
-/* Offset from drv_data->lpss_base */
-#define GENERAL_REG		0x08
-#define GENERAL_REG_RXTO_HOLDOFF_DISABLE BIT(24)
-#define SSP_REG			0x0c
-#define SPI_CS_CONTROL		0x18
-#define SPI_CS_CONTROL_SW_MODE	BIT(0)
-#define SPI_CS_CONTROL_CS_HIGH	BIT(1)
+static inline const struct lpss_config
+*lpss_get_config(const struct driver_data *drv_data)
+{
+	return &lpss_platforms[drv_data->ssp_type - LPSS_LPT_SSP];
+}
 
 static bool is_lpss_ssp(const struct driver_data *drv_data)
 {
-	return drv_data->ssp_type == LPSS_SSP;
+	switch (drv_data->ssp_type) {
+	case LPSS_LPT_SSP:
+	case LPSS_BYT_SSP:
+		return true;
+	default:
+		return false;
+	}
 }
 
 static bool is_quark_x1000_ssp(const struct driver_data *drv_data)
@@ -243,63 +231,43 @@ static void __lpss_ssp_write_priv(struct driver_data *drv_data,
  */
 static void lpss_ssp_setup(struct driver_data *drv_data)
 {
-	unsigned offset = 0x400;
-	u32 value, orig;
-
-	/*
-	 * Perform auto-detection of the LPSS SSP private registers. They
-	 * can be either at 1k or 2k offset from the base address.
-	 */
-	orig = readl(drv_data->ioaddr + offset + SPI_CS_CONTROL);
-
-	/* Test SPI_CS_CONTROL_SW_MODE bit enabling */
-	value = orig | SPI_CS_CONTROL_SW_MODE;
-	writel(value, drv_data->ioaddr + offset + SPI_CS_CONTROL);
-	value = readl(drv_data->ioaddr + offset + SPI_CS_CONTROL);
-	if (value != (orig | SPI_CS_CONTROL_SW_MODE)) {
-		offset = 0x800;
-		goto detection_done;
-	}
-
-	orig = readl(drv_data->ioaddr + offset + SPI_CS_CONTROL);
-
-	/* Test SPI_CS_CONTROL_SW_MODE bit disabling */
-	value = orig & ~SPI_CS_CONTROL_SW_MODE;
-	writel(value, drv_data->ioaddr + offset + SPI_CS_CONTROL);
-	value = readl(drv_data->ioaddr + offset + SPI_CS_CONTROL);
-	if (value != (orig & ~SPI_CS_CONTROL_SW_MODE)) {
-		offset = 0x800;
-		goto detection_done;
-	}
+	const struct lpss_config *config;
+	u32 value;
 
-detection_done:
-	/* Now set the LPSS base */
-	drv_data->lpss_base = drv_data->ioaddr + offset;
+	config = lpss_get_config(drv_data);
+	drv_data->lpss_base = drv_data->ioaddr + config->offset;
 
 	/* Enable software chip select control */
 	value = SPI_CS_CONTROL_SW_MODE | SPI_CS_CONTROL_CS_HIGH;
-	__lpss_ssp_write_priv(drv_data, SPI_CS_CONTROL, value);
+	__lpss_ssp_write_priv(drv_data, config->reg_cs_ctrl, value);
 
 	/* Enable multiblock DMA transfers */
 	if (drv_data->master_info->enable_dma) {
-		__lpss_ssp_write_priv(drv_data, SSP_REG, 1);
-
-		value = __lpss_ssp_read_priv(drv_data, GENERAL_REG);
-		value |= GENERAL_REG_RXTO_HOLDOFF_DISABLE;
-		__lpss_ssp_write_priv(drv_data, GENERAL_REG, value);
+		__lpss_ssp_write_priv(drv_data, config->reg_ssp, 1);
+
+		if (config->reg_general >= 0) {
+			value = __lpss_ssp_read_priv(drv_data,
+						     config->reg_general);
+			value |= GENERAL_REG_RXTO_HOLDOFF_DISABLE;
+			__lpss_ssp_write_priv(drv_data,
+					      config->reg_general, value);
+		}
 	}
 }
 
 static void lpss_ssp_cs_control(struct driver_data *drv_data, bool enable)
 {
+	const struct lpss_config *config;
 	u32 value;
 
-	value = __lpss_ssp_read_priv(drv_data, SPI_CS_CONTROL);
+	config = lpss_get_config(drv_data);
+
+	value = __lpss_ssp_read_priv(drv_data, config->reg_cs_ctrl);
 	if (enable)
 		value &= ~SPI_CS_CONTROL_CS_HIGH;
 	else
 		value |= SPI_CS_CONTROL_CS_HIGH;
-	__lpss_ssp_write_priv(drv_data, SPI_CS_CONTROL, value);
+	__lpss_ssp_write_priv(drv_data, config->reg_cs_ctrl, value);
 }
 
 static void cs_assert(struct driver_data *drv_data)
@@ -701,25 +669,124 @@ static irqreturn_t ssp_int(int irq, void *dev_id)
 }
 
 /*
- * The Quark SPI data sheet gives a table, and for the given 'rate',
- * the 'dds' and 'clk_div' can be found in the table.
+ * The Quark SPI has an additional 24 bit register (DDS_CLK_RATE) to multiply
+ * input frequency by fractions of 2^24. It also has a divider by 5.
+ *
+ * There are formulas to get baud rate value for given input frequency and
+ * divider parameters, such as DDS_CLK_RATE and SCR:
+ *
+ * Fsys = 200MHz
+ *
+ * Fssp = Fsys * DDS_CLK_RATE / 2^24			(1)
+ * Baud rate = Fsclk = Fssp / (2 * (SCR + 1))		(2)
+ *
+ * DDS_CLK_RATE either 2^n or 2^n / 5.
+ * SCR is in range 0 .. 255
+ *
+ * Divisor = 5^i * 2^j * 2 * k
+ *       i = [0, 1]      i = 1 iff j = 0 or j > 3
+ *       j = [0, 23]     j = 0 iff i = 1
+ *       k = [1, 256]
+ * Special case: j = 0, i = 1: Divisor = 2 / 5
+ *
+ * Accordingly to the specification the recommended values for DDS_CLK_RATE
+ * are:
+ *	Case 1:		2^n, n = [0, 23]
+ *	Case 2:		2^24 * 2 / 5 (0x666666)
+ *	Case 3:		less than or equal to 2^24 / 5 / 16 (0x33333)
+ *
+ * In all cases the lowest possible value is better.
+ *
+ * The function calculates parameters for all cases and chooses the one closest
+ * to the asked baud rate.
  */
-static u32 quark_x1000_set_clk_regvals(u32 rate, u32 *dds, u32 *clk_div)
+static unsigned int quark_x1000_get_clk_div(int rate, u32 *dds)
 {
-	unsigned int i;
-
-	for (i = 0; i < ARRAY_SIZE(quark_spi_rate_table); i++) {
-		if (rate >= quark_spi_rate_table[i].bitrate) {
-			*dds = quark_spi_rate_table[i].dds_clk_rate;
-			*clk_div = quark_spi_rate_table[i].clk_div;
-			return quark_spi_rate_table[i].bitrate;
+	unsigned long xtal = 200000000;
+	unsigned long fref = xtal / 2;		/* mandatory division by 2,
+						   see (2) */
+						/* case 3 */
+	unsigned long fref1 = fref / 2;		/* case 1 */
+	unsigned long fref2 = fref * 2 / 5;	/* case 2 */
+	unsigned long scale;
+	unsigned long q, q1, q2;
+	long r, r1, r2;
+	u32 mul;
+
+	/* Case 1 */
+
+	/* Set initial value for DDS_CLK_RATE */
+	mul = (1 << 24) >> 1;
+
+	/* Calculate initial quot */
+	q1 = DIV_ROUND_CLOSEST(fref1, rate);
+
+	/* Scale q1 if it's too big */
+	if (q1 > 256) {
+		/* Scale q1 to range [1, 512] */
+		scale = fls_long(q1 - 1);
+		if (scale > 9) {
+			q1 >>= scale - 9;
+			mul >>= scale - 9;
 		}
+
+		/* Round the result if we have a remainder */
+		q1 += q1 & 1;
 	}
 
-	*dds = quark_spi_rate_table[i-1].dds_clk_rate;
-	*clk_div = quark_spi_rate_table[i-1].clk_div;
+	/* Decrease DDS_CLK_RATE as much as we can without loss in precision */
+	scale = __ffs(q1);
+	q1 >>= scale;
+	mul >>= scale;
+
+	/* Get the remainder */
+	r1 = abs(fref1 / (1 << (24 - fls_long(mul))) / q1 - rate);
+
+	/* Case 2 */
 
-	return quark_spi_rate_table[i-1].bitrate;
+	q2 = DIV_ROUND_CLOSEST(fref2, rate);
+	r2 = abs(fref2 / q2 - rate);
+
+	/*
+	 * Choose the best between two: less remainder we have the better. We
+	 * can't go case 2 if q2 is greater than 256 since SCR register can
+	 * hold only values 0 .. 255.
+	 */
+	if (r2 >= r1 || q2 > 256) {
+		/* case 1 is better */
+		r = r1;
+		q = q1;
+	} else {
+		/* case 2 is better */
+		r = r2;
+		q = q2;
+		mul = (1 << 24) * 2 / 5;
+	}
+
+	/* Check case 3 only If the divisor is big enough */
+	if (fref / rate >= 80) {
+		u64 fssp;
+		u32 m;
+
+		/* Calculate initial quot */
+		q1 = DIV_ROUND_CLOSEST(fref, rate);
+		m = (1 << 24) / q1;
+
+		/* Get the remainder */
+		fssp = (u64)fref * m;
+		do_div(fssp, 1 << 24);
+		r1 = abs(fssp - rate);
+
+		/* Choose this one if it suits better */
+		if (r1 < r) {
+			/* case 3 is better */
+			q = 1;
+			mul = m;
+		}
+	}
+
+	*dds = mul;
+	return q - 1;
 }
 
 static unsigned int ssp_get_clk_div(struct driver_data *drv_data, int rate)
@@ -730,23 +797,25 @@ static unsigned int ssp_get_clk_div(struct driver_data *drv_data, int rate)
 	rate = min_t(int, ssp_clk, rate);
 
 	if (ssp->type == PXA25x_SSP || ssp->type == CE4100_SSP)
-		return ((ssp_clk / (2 * rate) - 1) & 0xff) << 8;
+		return (ssp_clk / (2 * rate) - 1) & 0xff;
 	else
-		return ((ssp_clk / rate - 1) & 0xfff) << 8;
+		return (ssp_clk / rate - 1) & 0xfff;
 }
 
 static unsigned int pxa2xx_ssp_get_clk_div(struct driver_data *drv_data,
 					   struct chip_data *chip, int rate)
 {
-	u32 clk_div;
+	unsigned int clk_div;
 
 	switch (drv_data->ssp_type) {
 	case QUARK_X1000_SSP:
-		quark_x1000_set_clk_regvals(rate, &chip->dds_rate, &clk_div);
-		return clk_div << 8;
+		clk_div = quark_x1000_get_clk_div(rate, &chip->dds_rate);
+		break;
 	default:
-		return ssp_get_clk_div(drv_data, rate);
+		clk_div = ssp_get_clk_div(drv_data, rate);
+		break;
 	}
+	return clk_div << 8;
 }
 
 static void pump_transfers(unsigned long data)
@@ -1025,6 +1094,7 @@ static int setup(struct spi_device *spi)
 {
 	struct pxa2xx_spi_chip *chip_info = NULL;
 	struct chip_data *chip;
+	const struct lpss_config *config;
 	struct driver_data *drv_data = spi_master_get_devdata(spi->master);
 	unsigned int clk_div;
 	uint tx_thres, tx_hi_thres, rx_thres;
@@ -1035,10 +1105,12 @@ static int setup(struct spi_device *spi)
 		tx_hi_thres = 0;
 		rx_thres = RX_THRESH_QUARK_X1000_DFLT;
 		break;
-	case LPSS_SSP:
-		tx_thres = LPSS_TX_LOTHRESH_DFLT;
-		tx_hi_thres = LPSS_TX_HITHRESH_DFLT;
-		rx_thres = LPSS_RX_THRESH_DFLT;
+	case LPSS_LPT_SSP:
+	case LPSS_BYT_SSP:
+		config = lpss_get_config(drv_data);
+		tx_thres = config->tx_threshold_lo;
+		tx_hi_thres = config->tx_threshold_hi;
+		rx_thres = config->rx_threshold;
 		break;
 	default:
 		tx_thres = TX_THRESH_DFLT;
@@ -1192,6 +1264,18 @@ static void cleanup(struct spi_device *spi)
 }
 
 #ifdef CONFIG_ACPI
+
+static const struct acpi_device_id pxa2xx_spi_acpi_match[] = {
+	{ "INT33C0", LPSS_LPT_SSP },
+	{ "INT33C1", LPSS_LPT_SSP },
+	{ "INT3430", LPSS_LPT_SSP },
+	{ "INT3431", LPSS_LPT_SSP },
+	{ "80860F0E", LPSS_BYT_SSP },
+	{ "8086228E", LPSS_BYT_SSP },
+	{ },
+};
+MODULE_DEVICE_TABLE(acpi, pxa2xx_spi_acpi_match);
+
 static struct pxa2xx_spi_master *
 pxa2xx_spi_acpi_get_pdata(struct platform_device *pdev)
 {
@@ -1199,12 +1283,19 @@ pxa2xx_spi_acpi_get_pdata(struct platform_device *pdev)
 	struct acpi_device *adev;
 	struct ssp_device *ssp;
 	struct resource *res;
-	int devid;
+	const struct acpi_device_id *id;
+	int devid, type;
 
 	if (!ACPI_HANDLE(&pdev->dev) ||
 	    acpi_bus_get_device(ACPI_HANDLE(&pdev->dev), &adev))
 		return NULL;
 
+	id = acpi_match_device(pdev->dev.driver->acpi_match_table, &pdev->dev);
+	if (id)
+		type = (int)id->driver_data;
+	else
+		return NULL;
+
 	pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
 	if (!pdata)
 		return NULL;
@@ -1222,7 +1313,7 @@ pxa2xx_spi_acpi_get_pdata(struct platform_device *pdev)
 
 	ssp->clk = devm_clk_get(&pdev->dev, NULL);
 	ssp->irq = platform_get_irq(pdev, 0);
-	ssp->type = LPSS_SSP;
+	ssp->type = type;
 	ssp->pdev = pdev;
 
 	ssp->port_id = -1;
@@ -1235,16 +1326,6 @@ pxa2xx_spi_acpi_get_pdata(struct platform_device *pdev)
 	return pdata;
 }
 
-static struct acpi_device_id pxa2xx_spi_acpi_match[] = {
-	{ "INT33C0", 0 },
-	{ "INT33C1", 0 },
-	{ "INT3430", 0 },
-	{ "INT3431", 0 },
-	{ "80860F0E", 0 },
-	{ "8086228E", 0 },
-	{ },
-};
-MODULE_DEVICE_TABLE(acpi, pxa2xx_spi_acpi_match);
 #else
 static inline struct pxa2xx_spi_master *
 pxa2xx_spi_acpi_get_pdata(struct platform_device *pdev)
diff --git a/drivers/spi/spi-pxa2xx.h b/drivers/spi/spi-pxa2xx.h
index 85a58c906869..9f01e9c9aa75 100644
--- a/drivers/spi/spi-pxa2xx.h
+++ b/drivers/spi/spi-pxa2xx.h
@@ -162,11 +162,7 @@ extern void *pxa2xx_spi_next_transfer(struct driver_data *drv_data);
 /*
  * Select the right DMA implementation.
  */
-#if defined(CONFIG_SPI_PXA2XX_PXADMA)
-#define SPI_PXA2XX_USE_DMA	1
-#define MAX_DMA_LEN		8191
-#define DEFAULT_DMA_CR1		(SSCR1_TSRE | SSCR1_RSRE | SSCR1_TINTE)
-#elif defined(CONFIG_SPI_PXA2XX_DMA)
+#if defined(CONFIG_SPI_PXA2XX_DMA)
 #define SPI_PXA2XX_USE_DMA	1
 #define MAX_DMA_LEN		SZ_64K
 #define DEFAULT_DMA_CR1		(SSCR1_TSRE | SSCR1_RSRE | SSCR1_TRAIL)
diff --git a/drivers/spi/spi-qup.c b/drivers/spi/spi-qup.c
index 2b2c359f5a50..810a7fae3479 100644
--- a/drivers/spi/spi-qup.c
+++ b/drivers/spi/spi-qup.c
@@ -22,6 +22,8 @@
 #include <linux/platform_device.h>
 #include <linux/pm_runtime.h>
 #include <linux/spi/spi.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
 
 #define QUP_CONFIG			0x0000
 #define QUP_STATE			0x0004
@@ -116,6 +118,8 @@
 
 #define SPI_NUM_CHIPSELECTS		4
 
+#define SPI_MAX_DMA_XFER		(SZ_64K - 64)
+
 /* high speed mode is when bus rate is greater then 26MHz */
 #define SPI_HS_MIN_RATE			26000000
 #define SPI_MAX_RATE			50000000
@@ -140,9 +144,14 @@ struct spi_qup {
 	struct completion	done;
 	int			error;
 	int			w_size;	/* bytes per SPI word */
+	int			n_words;
 	int			tx_bytes;
 	int			rx_bytes;
 	int			qup_v1;
+
+	int			use_dma;
+	struct dma_slave_config	rx_conf;
+	struct dma_slave_config	tx_conf;
 };
 
 
@@ -198,7 +207,6 @@ static int spi_qup_set_state(struct spi_qup *controller, u32 state)
 	return 0;
 }
 
-
 static void spi_qup_fifo_read(struct spi_qup *controller,
 			    struct spi_transfer *xfer)
 {
@@ -266,6 +274,107 @@ static void spi_qup_fifo_write(struct spi_qup *controller,
 	}
 }
 
+static void spi_qup_dma_done(void *data)
+{
+	struct spi_qup *qup = data;
+
+	complete(&qup->done);
+}
+
+static int spi_qup_prep_sg(struct spi_master *master, struct spi_transfer *xfer,
+			   enum dma_transfer_direction dir,
+			   dma_async_tx_callback callback)
+{
+	struct spi_qup *qup = spi_master_get_devdata(master);
+	unsigned long flags = DMA_PREP_INTERRUPT | DMA_PREP_FENCE;
+	struct dma_async_tx_descriptor *desc;
+	struct scatterlist *sgl;
+	struct dma_chan *chan;
+	dma_cookie_t cookie;
+	unsigned int nents;
+
+	if (dir == DMA_MEM_TO_DEV) {
+		chan = master->dma_tx;
+		nents = xfer->tx_sg.nents;
+		sgl = xfer->tx_sg.sgl;
+	} else {
+		chan = master->dma_rx;
+		nents = xfer->rx_sg.nents;
+		sgl = xfer->rx_sg.sgl;
+	}
+
+	desc = dmaengine_prep_slave_sg(chan, sgl, nents, dir, flags);
+	if (!desc)
+		return -EINVAL;
+
+	desc->callback = callback;
+	desc->callback_param = qup;
+
+	cookie = dmaengine_submit(desc);
+
+	return dma_submit_error(cookie);
+}
+
+static void spi_qup_dma_terminate(struct spi_master *master,
+				  struct spi_transfer *xfer)
+{
+	if (xfer->tx_buf)
+		dmaengine_terminate_all(master->dma_tx);
+	if (xfer->rx_buf)
+		dmaengine_terminate_all(master->dma_rx);
+}
+
+static int spi_qup_do_dma(struct spi_master *master, struct spi_transfer *xfer)
+{
+	dma_async_tx_callback rx_done = NULL, tx_done = NULL;
+	int ret;
+
+	if (xfer->rx_buf)
+		rx_done = spi_qup_dma_done;
+	else if (xfer->tx_buf)
+		tx_done = spi_qup_dma_done;
+
+	if (xfer->rx_buf) {
+		ret = spi_qup_prep_sg(master, xfer, DMA_DEV_TO_MEM, rx_done);
+		if (ret)
+			return ret;
+
+		dma_async_issue_pending(master->dma_rx);
+	}
+
+	if (xfer->tx_buf) {
+		ret = spi_qup_prep_sg(master, xfer, DMA_MEM_TO_DEV, tx_done);
+		if (ret)
+			return ret;
+
+		dma_async_issue_pending(master->dma_tx);
+	}
+
+	return 0;
+}
+
+static int spi_qup_do_pio(struct spi_master *master, struct spi_transfer *xfer)
+{
+	struct spi_qup *qup = spi_master_get_devdata(master);
+	int ret;
+
+	ret = spi_qup_set_state(qup, QUP_STATE_RUN);
+	if (ret) {
+		dev_warn(qup->dev, "cannot set RUN state\n");
+		return ret;
+	}
+
+	ret = spi_qup_set_state(qup, QUP_STATE_PAUSE);
+	if (ret) {
+		dev_warn(qup->dev, "cannot set PAUSE state\n");
+		return ret;
+	}
+
+	spi_qup_fifo_write(qup, xfer);
+
+	return 0;
+}
+
 static irqreturn_t spi_qup_qup_irq(int irq, void *dev_id)
 {
 	struct spi_qup *controller = dev_id;
@@ -315,11 +424,13 @@ static irqreturn_t spi_qup_qup_irq(int irq, void *dev_id)
 		error = -EIO;
 	}
 
-	if (opflags & QUP_OP_IN_SERVICE_FLAG)
-		spi_qup_fifo_read(controller, xfer);
+	if (!controller->use_dma) {
+		if (opflags & QUP_OP_IN_SERVICE_FLAG)
+			spi_qup_fifo_read(controller, xfer);
 
-	if (opflags & QUP_OP_OUT_SERVICE_FLAG)
-		spi_qup_fifo_write(controller, xfer);
+		if (opflags & QUP_OP_OUT_SERVICE_FLAG)
+			spi_qup_fifo_write(controller, xfer);
+	}
 
 	spin_lock_irqsave(&controller->lock, flags);
 	controller->error = error;
@@ -332,13 +443,35 @@ static irqreturn_t spi_qup_qup_irq(int irq, void *dev_id)
 	return IRQ_HANDLED;
 }
 
+static u32
+spi_qup_get_mode(struct spi_master *master, struct spi_transfer *xfer)
+{
+	struct spi_qup *qup = spi_master_get_devdata(master);
+	u32 mode;
+
+	qup->w_size = 4;
+
+	if (xfer->bits_per_word <= 8)
+		qup->w_size = 1;
+	else if (xfer->bits_per_word <= 16)
+		qup->w_size = 2;
+
+	qup->n_words = xfer->len / qup->w_size;
+
+	if (qup->n_words <= (qup->in_fifo_sz / sizeof(u32)))
+		mode = QUP_IO_M_MODE_FIFO;
+	else
+		mode = QUP_IO_M_MODE_BLOCK;
+
+	return mode;
+}
 
 /* set clock freq ... bits per word */
 static int spi_qup_io_config(struct spi_device *spi, struct spi_transfer *xfer)
 {
 	struct spi_qup *controller = spi_master_get_devdata(spi->master);
 	u32 config, iomode, mode, control;
-	int ret, n_words, w_size;
+	int ret, n_words;
 
 	if (spi->mode & SPI_LOOP && xfer->len > controller->in_fifo_sz) {
 		dev_err(controller->dev, "too big size for loopback %d > %d\n",
@@ -358,35 +491,54 @@ static int spi_qup_io_config(struct spi_device *spi, struct spi_transfer *xfer)
 		return -EIO;
 	}
 
-	w_size = 4;
-	if (xfer->bits_per_word <= 8)
-		w_size = 1;
-	else if (xfer->bits_per_word <= 16)
-		w_size = 2;
-
-	n_words = xfer->len / w_size;
-	controller->w_size = w_size;
+	mode = spi_qup_get_mode(spi->master, xfer);
+	n_words = controller->n_words;
 
-	if (n_words <= (controller->in_fifo_sz / sizeof(u32))) {
-		mode = QUP_IO_M_MODE_FIFO;
+	if (mode == QUP_IO_M_MODE_FIFO) {
 		writel_relaxed(n_words, controller->base + QUP_MX_READ_CNT);
 		writel_relaxed(n_words, controller->base + QUP_MX_WRITE_CNT);
 		/* must be zero for FIFO */
 		writel_relaxed(0, controller->base + QUP_MX_INPUT_CNT);
 		writel_relaxed(0, controller->base + QUP_MX_OUTPUT_CNT);
-	} else {
-		mode = QUP_IO_M_MODE_BLOCK;
+	} else if (!controller->use_dma) {
 		writel_relaxed(n_words, controller->base + QUP_MX_INPUT_CNT);
 		writel_relaxed(n_words, controller->base + QUP_MX_OUTPUT_CNT);
 		/* must be zero for BLOCK and BAM */
 		writel_relaxed(0, controller->base + QUP_MX_READ_CNT);
 		writel_relaxed(0, controller->base + QUP_MX_WRITE_CNT);
+	} else {
+		mode = QUP_IO_M_MODE_BAM;
+		writel_relaxed(0, controller->base + QUP_MX_READ_CNT);
+		writel_relaxed(0, controller->base + QUP_MX_WRITE_CNT);
+
+		if (!controller->qup_v1) {
+			void __iomem *input_cnt;
+
+			input_cnt = controller->base + QUP_MX_INPUT_CNT;
+			/*
+			 * for DMA transfers, both QUP_MX_INPUT_CNT and
+			 * QUP_MX_OUTPUT_CNT must be zero to all cases but one.
+			 * That case is a non-balanced transfer when there is
+			 * only a rx_buf.
+			 */
+			if (xfer->tx_buf)
+				writel_relaxed(0, input_cnt);
+			else
+				writel_relaxed(n_words, input_cnt);
+
+			writel_relaxed(0, controller->base + QUP_MX_OUTPUT_CNT);
+		}
 	}
 
 	iomode = readl_relaxed(controller->base + QUP_IO_M_MODES);
 	/* Set input and output transfer mode */
 	iomode &= ~(QUP_IO_M_INPUT_MODE_MASK | QUP_IO_M_OUTPUT_MODE_MASK);
-	iomode &= ~(QUP_IO_M_PACK_EN | QUP_IO_M_UNPACK_EN);
+
+	if (!controller->use_dma)
+		iomode &= ~(QUP_IO_M_PACK_EN | QUP_IO_M_UNPACK_EN);
+	else
+		iomode |= QUP_IO_M_PACK_EN | QUP_IO_M_UNPACK_EN;
+
 	iomode |= (mode << QUP_IO_M_OUTPUT_MODE_MASK_SHIFT);
 	iomode |= (mode << QUP_IO_M_INPUT_MODE_MASK_SHIFT);
 
@@ -428,11 +580,31 @@ static int spi_qup_io_config(struct spi_device *spi, struct spi_transfer *xfer)
 	config &= ~(QUP_CONFIG_NO_INPUT | QUP_CONFIG_NO_OUTPUT | QUP_CONFIG_N);
 	config |= xfer->bits_per_word - 1;
 	config |= QUP_CONFIG_SPI_MODE;
+
+	if (controller->use_dma) {
+		if (!xfer->tx_buf)
+			config |= QUP_CONFIG_NO_OUTPUT;
+		if (!xfer->rx_buf)
+			config |= QUP_CONFIG_NO_INPUT;
+	}
+
 	writel_relaxed(config, controller->base + QUP_CONFIG);
 
 	/* only write to OPERATIONAL_MASK when register is present */
-	if (!controller->qup_v1)
-		writel_relaxed(0, controller->base + QUP_OPERATIONAL_MASK);
+	if (!controller->qup_v1) {
+		u32 mask = 0;
+
+		/*
+		 * mask INPUT and OUTPUT service flags to prevent IRQs on FIFO
+		 * status change in BAM mode
+		 */
+
+		if (mode == QUP_IO_M_MODE_BAM)
+			mask = QUP_OP_IN_SERVICE_FLAG | QUP_OP_OUT_SERVICE_FLAG;
+
+		writel_relaxed(mask, controller->base + QUP_OPERATIONAL_MASK);
+	}
+
 	return 0;
 }
 
@@ -461,17 +633,13 @@ static int spi_qup_transfer_one(struct spi_master *master,
 	controller->tx_bytes = 0;
 	spin_unlock_irqrestore(&controller->lock, flags);
 
-	if (spi_qup_set_state(controller, QUP_STATE_RUN)) {
-		dev_warn(controller->dev, "cannot set RUN state\n");
-		goto exit;
-	}
+	if (controller->use_dma)
+		ret = spi_qup_do_dma(master, xfer);
+	else
+		ret = spi_qup_do_pio(master, xfer);
 
-	if (spi_qup_set_state(controller, QUP_STATE_PAUSE)) {
-		dev_warn(controller->dev, "cannot set PAUSE state\n");
+	if (ret)
 		goto exit;
-	}
-
-	spi_qup_fifo_write(controller, xfer);
 
 	if (spi_qup_set_state(controller, QUP_STATE_RUN)) {
 		dev_warn(controller->dev, "cannot set EXECUTE state\n");
@@ -480,6 +648,7 @@ static int spi_qup_transfer_one(struct spi_master *master,
 
 	if (!wait_for_completion_timeout(&controller->done, timeout))
 		ret = -ETIMEDOUT;
+
 exit:
 	spi_qup_set_state(controller, QUP_STATE_RESET);
 	spin_lock_irqsave(&controller->lock, flags);
@@ -487,6 +656,97 @@ exit:
 	if (!ret)
 		ret = controller->error;
 	spin_unlock_irqrestore(&controller->lock, flags);
+
+	if (ret && controller->use_dma)
+		spi_qup_dma_terminate(master, xfer);
+
+	return ret;
+}
+
+static bool spi_qup_can_dma(struct spi_master *master, struct spi_device *spi,
+			    struct spi_transfer *xfer)
+{
+	struct spi_qup *qup = spi_master_get_devdata(master);
+	size_t dma_align = dma_get_cache_alignment();
+	u32 mode;
+
+	qup->use_dma = 0;
+
+	if (xfer->rx_buf && (xfer->len % qup->in_blk_sz ||
+	    IS_ERR_OR_NULL(master->dma_rx) ||
+	    !IS_ALIGNED((size_t)xfer->rx_buf, dma_align)))
+		return false;
+
+	if (xfer->tx_buf && (xfer->len % qup->out_blk_sz ||
+	    IS_ERR_OR_NULL(master->dma_tx) ||
+	    !IS_ALIGNED((size_t)xfer->tx_buf, dma_align)))
+		return false;
+
+	mode = spi_qup_get_mode(master, xfer);
+	if (mode == QUP_IO_M_MODE_FIFO)
+		return false;
+
+	qup->use_dma = 1;
+
+	return true;
+}
+
+static void spi_qup_release_dma(struct spi_master *master)
+{
+	if (!IS_ERR_OR_NULL(master->dma_rx))
+		dma_release_channel(master->dma_rx);
+	if (!IS_ERR_OR_NULL(master->dma_tx))
+		dma_release_channel(master->dma_tx);
+}
+
+static int spi_qup_init_dma(struct spi_master *master, resource_size_t base)
+{
+	struct spi_qup *spi = spi_master_get_devdata(master);
+	struct dma_slave_config *rx_conf = &spi->rx_conf,
+				*tx_conf = &spi->tx_conf;
+	struct device *dev = spi->dev;
+	int ret;
+
+	/* allocate dma resources, if available */
+	master->dma_rx = dma_request_slave_channel_reason(dev, "rx");
+	if (IS_ERR(master->dma_rx))
+		return PTR_ERR(master->dma_rx);
+
+	master->dma_tx = dma_request_slave_channel_reason(dev, "tx");
+	if (IS_ERR(master->dma_tx)) {
+		ret = PTR_ERR(master->dma_tx);
+		goto err_tx;
+	}
+
+	/* set DMA parameters */
+	rx_conf->direction = DMA_DEV_TO_MEM;
+	rx_conf->device_fc = 1;
+	rx_conf->src_addr = base + QUP_INPUT_FIFO;
+	rx_conf->src_maxburst = spi->in_blk_sz;
+
+	tx_conf->direction = DMA_MEM_TO_DEV;
+	tx_conf->device_fc = 1;
+	tx_conf->dst_addr = base + QUP_OUTPUT_FIFO;
+	tx_conf->dst_maxburst = spi->out_blk_sz;
+
+	ret = dmaengine_slave_config(master->dma_rx, rx_conf);
+	if (ret) {
+		dev_err(dev, "failed to configure RX channel\n");
+		goto err;
+	}
+
+	ret = dmaengine_slave_config(master->dma_tx, tx_conf);
+	if (ret) {
+		dev_err(dev, "failed to configure TX channel\n");
+		goto err;
+	}
+
+	return 0;
+
+err:
+	dma_release_channel(master->dma_tx);
+err_tx:
+	dma_release_channel(master->dma_rx);
 	return ret;
 }
 
@@ -563,6 +823,8 @@ static int spi_qup_probe(struct platform_device *pdev)
 	master->transfer_one = spi_qup_transfer_one;
 	master->dev.of_node = pdev->dev.of_node;
 	master->auto_runtime_pm = true;
+	master->dma_alignment = dma_get_cache_alignment();
+	master->max_dma_len = SPI_MAX_DMA_XFER;
 
 	platform_set_drvdata(pdev, master);
 
@@ -574,6 +836,12 @@ static int spi_qup_probe(struct platform_device *pdev)
 	controller->cclk = cclk;
 	controller->irq = irq;
 
+	ret = spi_qup_init_dma(master, res->start);
+	if (ret == -EPROBE_DEFER)
+		goto error;
+	else if (!ret)
+		master->can_dma = spi_qup_can_dma;
+
 	/* set v1 flag if device is version 1 */
 	if (of_device_is_compatible(dev->of_node, "qcom,spi-qup-v1.1.1"))
 		controller->qup_v1 = 1;
@@ -610,7 +878,7 @@ static int spi_qup_probe(struct platform_device *pdev)
 	ret = spi_qup_set_state(controller, QUP_STATE_RESET);
 	if (ret) {
 		dev_err(dev, "cannot set RESET state\n");
-		goto error;
+		goto error_dma;
 	}
 
 	writel_relaxed(0, base + QUP_OPERATIONAL);
@@ -634,7 +902,7 @@ static int spi_qup_probe(struct platform_device *pdev)
 	ret = devm_request_irq(dev, irq, spi_qup_qup_irq,
 			       IRQF_TRIGGER_HIGH, pdev->name, controller);
 	if (ret)
-		goto error;
+		goto error_dma;
 
 	pm_runtime_set_autosuspend_delay(dev, MSEC_PER_SEC);
 	pm_runtime_use_autosuspend(dev);
@@ -649,6 +917,8 @@ static int spi_qup_probe(struct platform_device *pdev)
 
 disable_pm:
 	pm_runtime_disable(&pdev->dev);
+error_dma:
+	spi_qup_release_dma(master);
 error:
 	clk_disable_unprepare(cclk);
 	clk_disable_unprepare(iclk);
@@ -740,6 +1010,8 @@ static int spi_qup_remove(struct platform_device *pdev)
 	if (ret)
 		return ret;
 
+	spi_qup_release_dma(master);
+
 	clk_disable_unprepare(controller->cclk);
 	clk_disable_unprepare(controller->iclk);
 
diff --git a/drivers/spi/spi-rb4xx.c b/drivers/spi/spi-rb4xx.c
new file mode 100644
index 000000000000..3641d0e20135
--- /dev/null
+++ b/drivers/spi/spi-rb4xx.c
@@ -0,0 +1,210 @@
+/*
+ * SPI controller driver for the Mikrotik RB4xx boards
+ *
+ * Copyright (C) 2010 Gabor Juhos <juhosg@openwrt.org>
+ * Copyright (C) 2015 Bert Vermeulen <bert@biot.com>
+ *
+ * This file was based on the patches for Linux 2.6.27.39 published by
+ * MikroTik for their RouterBoard 4xx series devices.
+ *
+ * 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/platform_device.h>
+#include <linux/clk.h>
+#include <linux/spi/spi.h>
+
+#include <asm/mach-ath79/ar71xx_regs.h>
+
+struct rb4xx_spi {
+	void __iomem *base;
+	struct clk *clk;
+};
+
+static inline u32 rb4xx_read(struct rb4xx_spi *rbspi, u32 reg)
+{
+	return __raw_readl(rbspi->base + reg);
+}
+
+static inline void rb4xx_write(struct rb4xx_spi *rbspi, u32 reg, u32 value)
+{
+	__raw_writel(value, rbspi->base + reg);
+}
+
+static inline void do_spi_clk(struct rb4xx_spi *rbspi, u32 spi_ioc, int value)
+{
+	u32 regval;
+
+	regval = spi_ioc;
+	if (value & BIT(0))
+		regval |= AR71XX_SPI_IOC_DO;
+
+	rb4xx_write(rbspi, AR71XX_SPI_REG_IOC, regval);
+	rb4xx_write(rbspi, AR71XX_SPI_REG_IOC, regval | AR71XX_SPI_IOC_CLK);
+}
+
+static void do_spi_byte(struct rb4xx_spi *rbspi, u32 spi_ioc, u8 byte)
+{
+	int i;
+
+	for (i = 7; i >= 0; i--)
+		do_spi_clk(rbspi, spi_ioc, byte >> i);
+}
+
+/* The CS2 pin is used to clock in a second bit per clock cycle. */
+static inline void do_spi_clk_two(struct rb4xx_spi *rbspi, u32 spi_ioc,
+				   u8 value)
+{
+	u32 regval;
+
+	regval = spi_ioc;
+	if (value & BIT(1))
+		regval |= AR71XX_SPI_IOC_DO;
+	if (value & BIT(0))
+		regval |= AR71XX_SPI_IOC_CS2;
+
+	rb4xx_write(rbspi, AR71XX_SPI_REG_IOC, regval);
+	rb4xx_write(rbspi, AR71XX_SPI_REG_IOC, regval | AR71XX_SPI_IOC_CLK);
+}
+
+/* Two bits at a time, msb first */
+static void do_spi_byte_two(struct rb4xx_spi *rbspi, u32 spi_ioc, u8 byte)
+{
+	do_spi_clk_two(rbspi, spi_ioc, byte >> 6);
+	do_spi_clk_two(rbspi, spi_ioc, byte >> 4);
+	do_spi_clk_two(rbspi, spi_ioc, byte >> 2);
+	do_spi_clk_two(rbspi, spi_ioc, byte >> 0);
+}
+
+static void rb4xx_set_cs(struct spi_device *spi, bool enable)
+{
+	struct rb4xx_spi *rbspi = spi_master_get_devdata(spi->master);
+
+	/*
+	 * Setting CS is done along with bitbanging the actual values,
+	 * since it's all on the same hardware register. However the
+	 * CPLD needs CS deselected after every command.
+	 */
+	if (enable)
+		rb4xx_write(rbspi, AR71XX_SPI_REG_IOC,
+			    AR71XX_SPI_IOC_CS0 | AR71XX_SPI_IOC_CS1);
+}
+
+static int rb4xx_transfer_one(struct spi_master *master,
+			      struct spi_device *spi, struct spi_transfer *t)
+{
+	struct rb4xx_spi *rbspi = spi_master_get_devdata(master);
+	int i;
+	u32 spi_ioc;
+	u8 *rx_buf;
+	const u8 *tx_buf;
+
+	/*
+	 * Prime the SPI register with the SPI device selected. The m25p80 boot
+	 * flash and CPLD share the CS0 pin. This works because the CPLD's
+	 * command set was designed to almost not clash with that of the
+	 * boot flash.
+	 */
+	if (spi->chip_select == 2)
+		/* MMC */
+		spi_ioc = AR71XX_SPI_IOC_CS0;
+	else
+		/* Boot flash and CPLD */
+		spi_ioc = AR71XX_SPI_IOC_CS1;
+
+	tx_buf = t->tx_buf;
+	rx_buf = t->rx_buf;
+	for (i = 0; i < t->len; ++i) {
+		if (t->tx_nbits == SPI_NBITS_DUAL)
+			/* CPLD can use two-wire transfers */
+			do_spi_byte_two(rbspi, spi_ioc, tx_buf[i]);
+		else
+			do_spi_byte(rbspi, spi_ioc, tx_buf[i]);
+		if (!rx_buf)
+			continue;
+		rx_buf[i] = rb4xx_read(rbspi, AR71XX_SPI_REG_RDS);
+	}
+	spi_finalize_current_transfer(master);
+
+	return 0;
+}
+
+static int rb4xx_spi_probe(struct platform_device *pdev)
+{
+	struct spi_master *master;
+	struct clk *ahb_clk;
+	struct rb4xx_spi *rbspi;
+	struct resource *r;
+	int err;
+	void __iomem *spi_base;
+
+	r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	spi_base = devm_ioremap_resource(&pdev->dev, r);
+	if (IS_ERR(spi_base))
+		return PTR_ERR(spi_base);
+
+	master = spi_alloc_master(&pdev->dev, sizeof(*rbspi));
+	if (!master)
+		return -ENOMEM;
+
+	ahb_clk = devm_clk_get(&pdev->dev, "ahb");
+	if (IS_ERR(ahb_clk))
+		return PTR_ERR(ahb_clk);
+
+	master->bus_num = 0;
+	master->num_chipselect = 3;
+	master->mode_bits = SPI_TX_DUAL;
+	master->bits_per_word_mask = BIT(7);
+	master->flags = SPI_MASTER_MUST_TX;
+	master->transfer_one = rb4xx_transfer_one;
+	master->set_cs = rb4xx_set_cs;
+
+	err = devm_spi_register_master(&pdev->dev, master);
+	if (err) {
+		dev_err(&pdev->dev, "failed to register SPI master\n");
+		return err;
+	}
+
+	err = clk_prepare_enable(ahb_clk);
+	if (err)
+		return err;
+
+	rbspi = spi_master_get_devdata(master);
+	rbspi->base = spi_base;
+	rbspi->clk = ahb_clk;
+	platform_set_drvdata(pdev, rbspi);
+
+	/* Enable SPI */
+	rb4xx_write(rbspi, AR71XX_SPI_REG_FS, AR71XX_SPI_FS_GPIO);
+
+	return 0;
+}
+
+static int rb4xx_spi_remove(struct platform_device *pdev)
+{
+	struct rb4xx_spi *rbspi = platform_get_drvdata(pdev);
+
+	clk_disable_unprepare(rbspi->clk);
+
+	return 0;
+}
+
+static struct platform_driver rb4xx_spi_drv = {
+	.probe = rb4xx_spi_probe,
+	.remove = rb4xx_spi_remove,
+	.driver = {
+		.name = "rb4xx-spi",
+	},
+};
+
+module_platform_driver(rb4xx_spi_drv);
+
+MODULE_DESCRIPTION("Mikrotik RB4xx SPI controller driver");
+MODULE_AUTHOR("Gabor Juhos <juhosg@openwrt.org>");
+MODULE_AUTHOR("Bert Vermeulen <bert@biot.com>");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/spi/spi-rockchip.c b/drivers/spi/spi-rockchip.c
index 1a777dc261d6..68e7efeb9a27 100644
--- a/drivers/spi/spi-rockchip.c
+++ b/drivers/spi/spi-rockchip.c
@@ -179,6 +179,7 @@ struct rockchip_spi {
 	u8 tmode;
 	u8 bpw;
 	u8 n_bytes;
+	u8 rsd_nsecs;
 	unsigned len;
 	u32 speed;
 
@@ -302,8 +303,8 @@ static int rockchip_spi_prepare_message(struct spi_master *master,
 	return 0;
 }
 
-static int rockchip_spi_unprepare_message(struct spi_master *master,
-					  struct spi_message *msg)
+static void rockchip_spi_handle_err(struct spi_master *master,
+				    struct spi_message *msg)
 {
 	unsigned long flags;
 	struct rockchip_spi *rs = spi_master_get_devdata(master);
@@ -313,8 +314,8 @@ static int rockchip_spi_unprepare_message(struct spi_master *master,
 	/*
 	 * For DMA mode, we need terminate DMA channel and flush
 	 * fifo for the next transfer if DMA thansfer timeout.
-	 * unprepare_message() was called by core if transfer complete
-	 * or timeout. Maybe it is reasonable for error handling here.
+	 * handle_err() was called by core if transfer failed.
+	 * Maybe it is reasonable for error handling here.
 	 */
 	if (rs->use_dma) {
 		if (rs->state & RXBUSY) {
@@ -327,6 +328,12 @@ static int rockchip_spi_unprepare_message(struct spi_master *master,
 	}
 
 	spin_unlock_irqrestore(&rs->lock, flags);
+}
+
+static int rockchip_spi_unprepare_message(struct spi_master *master,
+					  struct spi_message *msg)
+{
+	struct rockchip_spi *rs = spi_master_get_devdata(master);
 
 	spi_enable_chip(rs, 0);
 
@@ -493,6 +500,7 @@ static void rockchip_spi_config(struct rockchip_spi *rs)
 {
 	u32 div = 0;
 	u32 dmacr = 0;
+	int rsd = 0;
 
 	u32 cr0 = (CR0_BHT_8BIT << CR0_BHT_OFFSET)
 		| (CR0_SSD_ONE << CR0_SSD_OFFSET);
@@ -519,9 +527,23 @@ static void rockchip_spi_config(struct rockchip_spi *rs)
 	}
 
 	/* div doesn't support odd number */
-	div = max_t(u32, rs->max_freq / rs->speed, 1);
+	div = DIV_ROUND_UP(rs->max_freq, rs->speed);
 	div = (div + 1) & 0xfffe;
 
+	/* Rx sample delay is expressed in parent clock cycles (max 3) */
+	rsd = DIV_ROUND_CLOSEST(rs->rsd_nsecs * (rs->max_freq >> 8),
+				1000000000 >> 8);
+	if (!rsd && rs->rsd_nsecs) {
+		pr_warn_once("rockchip-spi: %u Hz are too slow to express %u ns delay\n",
+			     rs->max_freq, rs->rsd_nsecs);
+	} else if (rsd > 3) {
+		rsd = 3;
+		pr_warn_once("rockchip-spi: %u Hz are too fast to express %u ns delay, clamping at %u ns\n",
+			     rs->max_freq, rs->rsd_nsecs,
+			     rsd * 1000000000U / rs->max_freq);
+	}
+	cr0 |= rsd << CR0_RSD_OFFSET;
+
 	writel_relaxed(cr0, rs->regs + ROCKCHIP_SPI_CTRLR0);
 
 	writel_relaxed(rs->len - 1, rs->regs + ROCKCHIP_SPI_CTRLR1);
@@ -614,6 +636,7 @@ static int rockchip_spi_probe(struct platform_device *pdev)
 	struct rockchip_spi *rs;
 	struct spi_master *master;
 	struct resource *mem;
+	u32 rsd_nsecs;
 
 	master = spi_alloc_master(&pdev->dev, sizeof(struct rockchip_spi));
 	if (!master)
@@ -665,6 +688,10 @@ static int rockchip_spi_probe(struct platform_device *pdev)
 	rs->dev = &pdev->dev;
 	rs->max_freq = clk_get_rate(rs->spiclk);
 
+	if (!of_property_read_u32(pdev->dev.of_node, "rx-sample-delay-ns",
+				  &rsd_nsecs))
+		rs->rsd_nsecs = rsd_nsecs;
+
 	rs->fifo_len = get_fifo_len(rs);
 	if (!rs->fifo_len) {
 		dev_err(&pdev->dev, "Failed to get fifo length\n");
@@ -688,6 +715,7 @@ static int rockchip_spi_probe(struct platform_device *pdev)
 	master->prepare_message = rockchip_spi_prepare_message;
 	master->unprepare_message = rockchip_spi_unprepare_message;
 	master->transfer_one = rockchip_spi_transfer_one;
+	master->handle_err = rockchip_spi_handle_err;
 
 	rs->dma_tx.ch = dma_request_slave_channel(rs->dev, "tx");
 	if (!rs->dma_tx.ch)
diff --git a/drivers/spi/spi-rspi.c b/drivers/spi/spi-rspi.c
index 46ce47076e63..f9189a0c8cec 100644
--- a/drivers/spi/spi-rspi.c
+++ b/drivers/spi/spi-rspi.c
@@ -177,6 +177,13 @@
 #define SPBFCR_RXRST		0x40	/* Receive Buffer Data Reset */
 #define SPBFCR_TXTRG_MASK	0x30	/* Transmit Buffer Data Triggering Number */
 #define SPBFCR_RXTRG_MASK	0x07	/* Receive Buffer Data Triggering Number */
+/* QSPI on R-Car Gen2 */
+#define SPBFCR_TXTRG_1B		0x00	/* 31 bytes (1 byte available) */
+#define SPBFCR_TXTRG_32B	0x30	/* 0 byte (32 bytes available) */
+#define SPBFCR_RXTRG_1B		0x00	/* 1 byte (31 bytes available) */
+#define SPBFCR_RXTRG_32B	0x07	/* 32 bytes (0 byte available) */
+
+#define QSPI_BUFFER_SIZE        32u
 
 struct rspi_data {
 	void __iomem *addr;
@@ -366,6 +373,52 @@ static int qspi_set_config_register(struct rspi_data *rspi, int access_size)
 	return 0;
 }
 
+static void qspi_update(const struct rspi_data *rspi, u8 mask, u8 val, u8 reg)
+{
+	u8 data;
+
+	data = rspi_read8(rspi, reg);
+	data &= ~mask;
+	data |= (val & mask);
+	rspi_write8(rspi, data, reg);
+}
+
+static int qspi_set_send_trigger(struct rspi_data *rspi, unsigned int len)
+{
+	unsigned int n;
+
+	n = min(len, QSPI_BUFFER_SIZE);
+
+	if (len >= QSPI_BUFFER_SIZE) {
+		/* sets triggering number to 32 bytes */
+		qspi_update(rspi, SPBFCR_TXTRG_MASK,
+			     SPBFCR_TXTRG_32B, QSPI_SPBFCR);
+	} else {
+		/* sets triggering number to 1 byte */
+		qspi_update(rspi, SPBFCR_TXTRG_MASK,
+			     SPBFCR_TXTRG_1B, QSPI_SPBFCR);
+	}
+
+	return n;
+}
+
+static void qspi_set_receive_trigger(struct rspi_data *rspi, unsigned int len)
+{
+	unsigned int n;
+
+	n = min(len, QSPI_BUFFER_SIZE);
+
+	if (len >= QSPI_BUFFER_SIZE) {
+		/* sets triggering number to 32 bytes */
+		qspi_update(rspi, SPBFCR_RXTRG_MASK,
+			     SPBFCR_RXTRG_32B, QSPI_SPBFCR);
+	} else {
+		/* sets triggering number to 1 byte */
+		qspi_update(rspi, SPBFCR_RXTRG_MASK,
+			     SPBFCR_RXTRG_1B, QSPI_SPBFCR);
+	}
+}
+
 #define set_config_register(spi, n) spi->ops->set_config_register(spi, n)
 
 static void rspi_enable_irq(const struct rspi_data *rspi, u8 enable)
@@ -609,18 +662,25 @@ static bool rspi_can_dma(struct spi_master *master, struct spi_device *spi,
 	return __rspi_can_dma(rspi, xfer);
 }
 
+static int rspi_dma_check_then_transfer(struct rspi_data *rspi,
+					 struct spi_transfer *xfer)
+{
+	if (!rspi->master->can_dma || !__rspi_can_dma(rspi, xfer))
+		return -EAGAIN;
+
+	/* rx_buf can be NULL on RSPI on SH in TX-only Mode */
+	return rspi_dma_transfer(rspi, &xfer->tx_sg,
+				xfer->rx_buf ? &xfer->rx_sg : NULL);
+}
+
 static int rspi_common_transfer(struct rspi_data *rspi,
 				struct spi_transfer *xfer)
 {
 	int ret;
 
-	if (rspi->master->can_dma && __rspi_can_dma(rspi, xfer)) {
-		/* rx_buf can be NULL on RSPI on SH in TX-only Mode */
-		ret = rspi_dma_transfer(rspi, &xfer->tx_sg,
-					xfer->rx_buf ? &xfer->rx_sg : NULL);
-		if (ret != -EAGAIN)
-			return ret;
-	}
+	ret = rspi_dma_check_then_transfer(rspi, xfer);
+	if (ret != -EAGAIN)
+		return ret;
 
 	ret = rspi_pio_transfer(rspi, xfer->tx_buf, xfer->rx_buf, xfer->len);
 	if (ret < 0)
@@ -661,12 +721,55 @@ static int rspi_rz_transfer_one(struct spi_master *master,
 	return rspi_common_transfer(rspi, xfer);
 }
 
+static int qspi_trigger_transfer_out_in(struct rspi_data *rspi, const u8 *tx,
+					u8 *rx, unsigned int len)
+{
+	int i, n, ret;
+	int error;
+
+	while (len > 0) {
+		n = qspi_set_send_trigger(rspi, len);
+		qspi_set_receive_trigger(rspi, len);
+		if (n == QSPI_BUFFER_SIZE) {
+			error = rspi_wait_for_tx_empty(rspi);
+			if (error < 0) {
+				dev_err(&rspi->master->dev, "transmit timeout\n");
+				return error;
+			}
+			for (i = 0; i < n; i++)
+				rspi_write_data(rspi, *tx++);
+
+			error = rspi_wait_for_rx_full(rspi);
+			if (error < 0) {
+				dev_err(&rspi->master->dev, "receive timeout\n");
+				return error;
+			}
+			for (i = 0; i < n; i++)
+				*rx++ = rspi_read_data(rspi);
+		} else {
+			ret = rspi_pio_transfer(rspi, tx, rx, n);
+			if (ret < 0)
+				return ret;
+		}
+		len -= n;
+	}
+
+	return 0;
+}
+
 static int qspi_transfer_out_in(struct rspi_data *rspi,
 				struct spi_transfer *xfer)
 {
+	int ret;
+
 	qspi_receive_init(rspi);
 
-	return rspi_common_transfer(rspi, xfer);
+	ret = rspi_dma_check_then_transfer(rspi, xfer);
+	if (ret != -EAGAIN)
+		return ret;
+
+	return qspi_trigger_transfer_out_in(rspi, xfer->tx_buf,
+					    xfer->rx_buf, xfer->len);
 }
 
 static int qspi_transfer_out(struct rspi_data *rspi, struct spi_transfer *xfer)
@@ -913,7 +1016,6 @@ static struct dma_chan *rspi_request_dma_chan(struct device *dev,
 	}
 
 	memset(&cfg, 0, sizeof(cfg));
-	cfg.slave_id = id;
 	cfg.direction = dir;
 	if (dir == DMA_MEM_TO_DEV) {
 		cfg.dst_addr = port_addr;
@@ -1191,7 +1293,7 @@ error1:
 	return ret;
 }
 
-static struct platform_device_id spi_driver_ids[] = {
+static const struct platform_device_id spi_driver_ids[] = {
 	{ "rspi",	(kernel_ulong_t)&rspi_ops },
 	{ "rspi-rz",	(kernel_ulong_t)&rspi_rz_ops },
 	{ "qspi",	(kernel_ulong_t)&qspi_ops },
diff --git a/drivers/spi/spi-s3c64xx.c b/drivers/spi/spi-s3c64xx.c
index 9231c34b5a5c..2a8c513c4d07 100644
--- a/drivers/spi/spi-s3c64xx.c
+++ b/drivers/spi/spi-s3c64xx.c
@@ -324,7 +324,7 @@ static int s3c64xx_spi_prepare_transfer(struct spi_master *spi)
 
 		/* Acquire DMA channels */
 		sdd->rx_dma.ch = dma_request_slave_channel_compat(mask, filter,
-				   (void *)sdd->rx_dma.dmach, dev, "rx");
+				   (void *)(long)sdd->rx_dma.dmach, dev, "rx");
 		if (!sdd->rx_dma.ch) {
 			dev_err(dev, "Failed to get RX DMA channel\n");
 			ret = -EBUSY;
@@ -333,7 +333,7 @@ static int s3c64xx_spi_prepare_transfer(struct spi_master *spi)
 		spi->dma_rx = sdd->rx_dma.ch;
 
 		sdd->tx_dma.ch = dma_request_slave_channel_compat(mask, filter,
-				   (void *)sdd->tx_dma.dmach, dev, "tx");
+				   (void *)(long)sdd->tx_dma.dmach, dev, "tx");
 		if (!sdd->tx_dma.ch) {
 			dev_err(dev, "Failed to get TX DMA channel\n");
 			ret = -EBUSY;
@@ -1347,7 +1347,7 @@ static struct s3c64xx_spi_port_config exynos7_spi_port_config = {
 	.quirks		= S3C64XX_SPI_QUIRK_CS_AUTO,
 };
 
-static struct platform_device_id s3c64xx_spi_driver_ids[] = {
+static const struct platform_device_id s3c64xx_spi_driver_ids[] = {
 	{
 		.name		= "s3c2443-spi",
 		.driver_data	= (kernel_ulong_t)&s3c2443_spi_port_config,
diff --git a/drivers/spi/spi-sc18is602.c b/drivers/spi/spi-sc18is602.c
index 5a56acf8a43e..36af4d48a700 100644
--- a/drivers/spi/spi-sc18is602.c
+++ b/drivers/spi/spi-sc18is602.c
@@ -286,7 +286,7 @@ static int sc18is602_probe(struct i2c_client *client,
 			hw->freq = SC18IS602_CLOCK;
 		break;
 	}
-	master->bus_num = client->adapter->nr;
+	master->bus_num = np ? -1 : client->adapter->nr;
 	master->mode_bits = SPI_CPHA | SPI_CPOL | SPI_LSB_FIRST;
 	master->bits_per_word_mask = SPI_BPW_MASK(8);
 	master->setup = sc18is602_setup;
diff --git a/drivers/spi/spi-sh-msiof.c b/drivers/spi/spi-sh-msiof.c
index e57eec0b2f46..d3370a612d84 100644
--- a/drivers/spi/spi-sh-msiof.c
+++ b/drivers/spi/spi-sh-msiof.c
@@ -1030,7 +1030,6 @@ static struct dma_chan *sh_msiof_request_dma_chan(struct device *dev,
 	}
 
 	memset(&cfg, 0, sizeof(cfg));
-	cfg.slave_id = id;
 	cfg.direction = dir;
 	if (dir == DMA_MEM_TO_DEV) {
 		cfg.dst_addr = port_addr;
@@ -1264,7 +1263,7 @@ static int sh_msiof_spi_remove(struct platform_device *pdev)
 	return 0;
 }
 
-static struct platform_device_id spi_driver_ids[] = {
+static const struct platform_device_id spi_driver_ids[] = {
 	{ "spi_sh_msiof",	(kernel_ulong_t)&sh_data },
 	{ "spi_r8a7790_msiof",	(kernel_ulong_t)&r8a779x_data },
 	{ "spi_r8a7791_msiof",	(kernel_ulong_t)&r8a779x_data },
diff --git a/drivers/spi/spi-sirf.c b/drivers/spi/spi-sirf.c
index f5715c9f68b0..7072276ad354 100644
--- a/drivers/spi/spi-sirf.c
+++ b/drivers/spi/spi-sirf.c
@@ -26,28 +26,6 @@
 #include <linux/reset.h>
 
 #define DRIVER_NAME "sirfsoc_spi"
-
-#define SIRFSOC_SPI_CTRL		0x0000
-#define SIRFSOC_SPI_CMD			0x0004
-#define SIRFSOC_SPI_TX_RX_EN		0x0008
-#define SIRFSOC_SPI_INT_EN		0x000C
-#define SIRFSOC_SPI_INT_STATUS		0x0010
-#define SIRFSOC_SPI_TX_DMA_IO_CTRL	0x0100
-#define SIRFSOC_SPI_TX_DMA_IO_LEN	0x0104
-#define SIRFSOC_SPI_TXFIFO_CTRL		0x0108
-#define SIRFSOC_SPI_TXFIFO_LEVEL_CHK	0x010C
-#define SIRFSOC_SPI_TXFIFO_OP		0x0110
-#define SIRFSOC_SPI_TXFIFO_STATUS	0x0114
-#define SIRFSOC_SPI_TXFIFO_DATA		0x0118
-#define SIRFSOC_SPI_RX_DMA_IO_CTRL	0x0120
-#define SIRFSOC_SPI_RX_DMA_IO_LEN	0x0124
-#define SIRFSOC_SPI_RXFIFO_CTRL		0x0128
-#define SIRFSOC_SPI_RXFIFO_LEVEL_CHK	0x012C
-#define SIRFSOC_SPI_RXFIFO_OP		0x0130
-#define SIRFSOC_SPI_RXFIFO_STATUS	0x0134
-#define SIRFSOC_SPI_RXFIFO_DATA		0x0138
-#define SIRFSOC_SPI_DUMMY_DELAY_CTL	0x0144
-
 /* SPI CTRL register defines */
 #define SIRFSOC_SPI_SLV_MODE		BIT(16)
 #define SIRFSOC_SPI_CMD_MODE		BIT(17)
@@ -80,8 +58,6 @@
 #define SIRFSOC_SPI_TXFIFO_THD_INT_EN	BIT(9)
 #define SIRFSOC_SPI_FRM_END_INT_EN	BIT(10)
 
-#define SIRFSOC_SPI_INT_MASK_ALL	0x1FFF
-
 /* Interrupt status */
 #define SIRFSOC_SPI_RX_DONE		BIT(0)
 #define SIRFSOC_SPI_TX_DONE		BIT(1)
@@ -110,20 +86,66 @@
 #define SIRFSOC_SPI_FIFO_WIDTH_BYTE	(0 << 0)
 #define SIRFSOC_SPI_FIFO_WIDTH_WORD	(1 << 0)
 #define SIRFSOC_SPI_FIFO_WIDTH_DWORD	(2 << 0)
-
-/* FIFO Status */
-#define	SIRFSOC_SPI_FIFO_LEVEL_MASK	0xFF
-#define SIRFSOC_SPI_FIFO_FULL		BIT(8)
-#define SIRFSOC_SPI_FIFO_EMPTY		BIT(9)
-
-/* 256 bytes rx/tx FIFO */
-#define SIRFSOC_SPI_FIFO_SIZE		256
-#define SIRFSOC_SPI_DAT_FRM_LEN_MAX	(64 * 1024)
-
-#define SIRFSOC_SPI_FIFO_SC(x)		((x) & 0x3F)
-#define SIRFSOC_SPI_FIFO_LC(x)		(((x) & 0x3F) << 10)
-#define SIRFSOC_SPI_FIFO_HC(x)		(((x) & 0x3F) << 20)
-#define SIRFSOC_SPI_FIFO_THD(x)		(((x) & 0xFF) << 2)
+/* USP related */
+#define SIRFSOC_USP_SYNC_MODE		BIT(0)
+#define SIRFSOC_USP_SLV_MODE		BIT(1)
+#define SIRFSOC_USP_LSB			BIT(4)
+#define SIRFSOC_USP_EN			BIT(5)
+#define SIRFSOC_USP_RXD_FALLING_EDGE	BIT(6)
+#define SIRFSOC_USP_TXD_FALLING_EDGE	BIT(7)
+#define SIRFSOC_USP_CS_HIGH_VALID	BIT(9)
+#define SIRFSOC_USP_SCLK_IDLE_STAT	BIT(11)
+#define SIRFSOC_USP_TFS_IO_MODE		BIT(14)
+#define SIRFSOC_USP_TFS_IO_INPUT	BIT(19)
+
+#define SIRFSOC_USP_RXD_DELAY_LEN_MASK	0xFF
+#define SIRFSOC_USP_TXD_DELAY_LEN_MASK	0xFF
+#define SIRFSOC_USP_RXD_DELAY_OFFSET	0
+#define SIRFSOC_USP_TXD_DELAY_OFFSET	8
+#define SIRFSOC_USP_RXD_DELAY_LEN	1
+#define SIRFSOC_USP_TXD_DELAY_LEN	1
+#define SIRFSOC_USP_CLK_DIVISOR_OFFSET	21
+#define SIRFSOC_USP_CLK_DIVISOR_MASK	0x3FF
+#define SIRFSOC_USP_CLK_10_11_MASK	0x3
+#define SIRFSOC_USP_CLK_10_11_OFFSET	30
+#define SIRFSOC_USP_CLK_12_15_MASK	0xF
+#define SIRFSOC_USP_CLK_12_15_OFFSET	24
+
+#define SIRFSOC_USP_TX_DATA_OFFSET	0
+#define SIRFSOC_USP_TX_SYNC_OFFSET	8
+#define SIRFSOC_USP_TX_FRAME_OFFSET	16
+#define SIRFSOC_USP_TX_SHIFTER_OFFSET	24
+
+#define SIRFSOC_USP_TX_DATA_MASK	0xFF
+#define SIRFSOC_USP_TX_SYNC_MASK	0xFF
+#define SIRFSOC_USP_TX_FRAME_MASK	0xFF
+#define SIRFSOC_USP_TX_SHIFTER_MASK	0x1F
+
+#define SIRFSOC_USP_RX_DATA_OFFSET	0
+#define SIRFSOC_USP_RX_FRAME_OFFSET	8
+#define SIRFSOC_USP_RX_SHIFTER_OFFSET	16
+
+#define SIRFSOC_USP_RX_DATA_MASK	0xFF
+#define SIRFSOC_USP_RX_FRAME_MASK	0xFF
+#define SIRFSOC_USP_RX_SHIFTER_MASK	0x1F
+#define SIRFSOC_USP_CS_HIGH_VALUE	BIT(1)
+
+#define SIRFSOC_SPI_FIFO_SC_OFFSET	0
+#define SIRFSOC_SPI_FIFO_LC_OFFSET	10
+#define SIRFSOC_SPI_FIFO_HC_OFFSET	20
+
+#define SIRFSOC_SPI_FIFO_FULL_MASK(s)	(1 << ((s)->fifo_full_offset))
+#define SIRFSOC_SPI_FIFO_EMPTY_MASK(s)	(1 << ((s)->fifo_full_offset + 1))
+#define SIRFSOC_SPI_FIFO_THD_MASK(s)	((s)->fifo_size - 1)
+#define SIRFSOC_SPI_FIFO_THD_OFFSET	2
+#define SIRFSOC_SPI_FIFO_LEVEL_CHK_MASK(s, val)	\
+	((val) & (s)->fifo_level_chk_mask)
+
+enum sirf_spi_type {
+	SIRF_REAL_SPI,
+	SIRF_USP_SPI_P2,
+	SIRF_USP_SPI_A7,
+};
 
 /*
  * only if the rx/tx buffer and transfer size are 4-bytes aligned, we use dma
@@ -137,6 +159,95 @@
 #define SIRFSOC_MAX_CMD_BYTES	4
 #define SIRFSOC_SPI_DEFAULT_FRQ 1000000
 
+struct sirf_spi_register {
+	/*SPI and USP-SPI common*/
+	u32 tx_rx_en;
+	u32 int_en;
+	u32 int_st;
+	u32 tx_dma_io_ctrl;
+	u32 tx_dma_io_len;
+	u32 txfifo_ctrl;
+	u32 txfifo_level_chk;
+	u32 txfifo_op;
+	u32 txfifo_st;
+	u32 txfifo_data;
+	u32 rx_dma_io_ctrl;
+	u32 rx_dma_io_len;
+	u32 rxfifo_ctrl;
+	u32 rxfifo_level_chk;
+	u32 rxfifo_op;
+	u32 rxfifo_st;
+	u32 rxfifo_data;
+	/*SPI self*/
+	u32 spi_ctrl;
+	u32 spi_cmd;
+	u32 spi_dummy_delay_ctrl;
+	/*USP-SPI self*/
+	u32 usp_mode1;
+	u32 usp_mode2;
+	u32 usp_tx_frame_ctrl;
+	u32 usp_rx_frame_ctrl;
+	u32 usp_pin_io_data;
+	u32 usp_risc_dsp_mode;
+	u32 usp_async_param_reg;
+	u32 usp_irda_x_mode_div;
+	u32 usp_sm_cfg;
+	u32 usp_int_en_clr;
+};
+
+static const struct sirf_spi_register real_spi_register = {
+	.tx_rx_en		= 0x8,
+	.int_en		= 0xc,
+	.int_st		= 0x10,
+	.tx_dma_io_ctrl	= 0x100,
+	.tx_dma_io_len	= 0x104,
+	.txfifo_ctrl	= 0x108,
+	.txfifo_level_chk	= 0x10c,
+	.txfifo_op		= 0x110,
+	.txfifo_st		= 0x114,
+	.txfifo_data	= 0x118,
+	.rx_dma_io_ctrl	= 0x120,
+	.rx_dma_io_len	= 0x124,
+	.rxfifo_ctrl	= 0x128,
+	.rxfifo_level_chk	= 0x12c,
+	.rxfifo_op		= 0x130,
+	.rxfifo_st		= 0x134,
+	.rxfifo_data	= 0x138,
+	.spi_ctrl		= 0x0,
+	.spi_cmd		= 0x4,
+	.spi_dummy_delay_ctrl	= 0x144,
+};
+
+static const struct sirf_spi_register usp_spi_register = {
+	.tx_rx_en		= 0x10,
+	.int_en		= 0x14,
+	.int_st		= 0x18,
+	.tx_dma_io_ctrl	= 0x100,
+	.tx_dma_io_len	= 0x104,
+	.txfifo_ctrl	= 0x108,
+	.txfifo_level_chk	= 0x10c,
+	.txfifo_op		= 0x110,
+	.txfifo_st		= 0x114,
+	.txfifo_data	= 0x118,
+	.rx_dma_io_ctrl	= 0x120,
+	.rx_dma_io_len	= 0x124,
+	.rxfifo_ctrl	= 0x128,
+	.rxfifo_level_chk	= 0x12c,
+	.rxfifo_op		= 0x130,
+	.rxfifo_st		= 0x134,
+	.rxfifo_data	= 0x138,
+	.usp_mode1		= 0x0,
+	.usp_mode2		= 0x4,
+	.usp_tx_frame_ctrl	= 0x8,
+	.usp_rx_frame_ctrl	= 0xc,
+	.usp_pin_io_data	= 0x1c,
+	.usp_risc_dsp_mode	= 0x20,
+	.usp_async_param_reg	= 0x24,
+	.usp_irda_x_mode_div	= 0x28,
+	.usp_sm_cfg		= 0x2c,
+	.usp_int_en_clr		= 0x140,
+};
+
 struct sirfsoc_spi {
 	struct spi_bitbang bitbang;
 	struct completion rx_done;
@@ -164,7 +275,6 @@ struct sirfsoc_spi {
 	struct dma_chan *tx_chan;
 	dma_addr_t src_start;
 	dma_addr_t dst_start;
-	void *dummypage;
 	int word_width; /* in bytes */
 
 	/*
@@ -173,14 +283,39 @@ struct sirfsoc_spi {
 	 */
 	bool	tx_by_cmd;
 	bool	hw_cs;
+	enum sirf_spi_type type;
+	const struct sirf_spi_register *regs;
+	unsigned int fifo_size;
+	/* fifo empty offset is (fifo full offset + 1)*/
+	unsigned int fifo_full_offset;
+	/* fifo_level_chk_mask is (fifo_size/4 - 1) */
+	unsigned int fifo_level_chk_mask;
+	unsigned int dat_max_frm_len;
+};
+
+struct sirf_spi_comp_data {
+	const struct sirf_spi_register *regs;
+	enum sirf_spi_type type;
+	unsigned int dat_max_frm_len;
+	unsigned int fifo_size;
+	void (*hwinit)(struct sirfsoc_spi *sspi);
 };
 
+static void sirfsoc_usp_hwinit(struct sirfsoc_spi *sspi)
+{
+	/* reset USP and let USP can operate */
+	writel(readl(sspi->base + sspi->regs->usp_mode1) &
+		~SIRFSOC_USP_EN, sspi->base + sspi->regs->usp_mode1);
+	writel(readl(sspi->base + sspi->regs->usp_mode1) |
+		SIRFSOC_USP_EN, sspi->base + sspi->regs->usp_mode1);
+}
+
 static void spi_sirfsoc_rx_word_u8(struct sirfsoc_spi *sspi)
 {
 	u32 data;
 	u8 *rx = sspi->rx;
 
-	data = readl(sspi->base + SIRFSOC_SPI_RXFIFO_DATA);
+	data = readl(sspi->base + sspi->regs->rxfifo_data);
 
 	if (rx) {
 		*rx++ = (u8) data;
@@ -199,8 +334,7 @@ static void spi_sirfsoc_tx_word_u8(struct sirfsoc_spi *sspi)
 		data = *tx++;
 		sspi->tx = tx;
 	}
-
-	writel(data, sspi->base + SIRFSOC_SPI_TXFIFO_DATA);
+	writel(data, sspi->base + sspi->regs->txfifo_data);
 	sspi->left_tx_word--;
 }
 
@@ -209,7 +343,7 @@ static void spi_sirfsoc_rx_word_u16(struct sirfsoc_spi *sspi)
 	u32 data;
 	u16 *rx = sspi->rx;
 
-	data = readl(sspi->base + SIRFSOC_SPI_RXFIFO_DATA);
+	data = readl(sspi->base + sspi->regs->rxfifo_data);
 
 	if (rx) {
 		*rx++ = (u16) data;
@@ -229,7 +363,7 @@ static void spi_sirfsoc_tx_word_u16(struct sirfsoc_spi *sspi)
 		sspi->tx = tx;
 	}
 
-	writel(data, sspi->base + SIRFSOC_SPI_TXFIFO_DATA);
+	writel(data, sspi->base + sspi->regs->txfifo_data);
 	sspi->left_tx_word--;
 }
 
@@ -238,7 +372,7 @@ static void spi_sirfsoc_rx_word_u32(struct sirfsoc_spi *sspi)
 	u32 data;
 	u32 *rx = sspi->rx;
 
-	data = readl(sspi->base + SIRFSOC_SPI_RXFIFO_DATA);
+	data = readl(sspi->base + sspi->regs->rxfifo_data);
 
 	if (rx) {
 		*rx++ = (u32) data;
@@ -259,41 +393,59 @@ static void spi_sirfsoc_tx_word_u32(struct sirfsoc_spi *sspi)
 		sspi->tx = tx;
 	}
 
-	writel(data, sspi->base + SIRFSOC_SPI_TXFIFO_DATA);
+	writel(data, sspi->base + sspi->regs->txfifo_data);
 	sspi->left_tx_word--;
 }
 
 static irqreturn_t spi_sirfsoc_irq(int irq, void *dev_id)
 {
 	struct sirfsoc_spi *sspi = dev_id;
-	u32 spi_stat = readl(sspi->base + SIRFSOC_SPI_INT_STATUS);
-	if (sspi->tx_by_cmd && (spi_stat & SIRFSOC_SPI_FRM_END)) {
+	u32 spi_stat;
+
+	spi_stat = readl(sspi->base + sspi->regs->int_st);
+	if (sspi->tx_by_cmd && sspi->type == SIRF_REAL_SPI
+		&& (spi_stat & SIRFSOC_SPI_FRM_END)) {
 		complete(&sspi->tx_done);
-		writel(0x0, sspi->base + SIRFSOC_SPI_INT_EN);
-		writel(SIRFSOC_SPI_INT_MASK_ALL,
-				sspi->base + SIRFSOC_SPI_INT_STATUS);
+		writel(0x0, sspi->base + sspi->regs->int_en);
+		writel(readl(sspi->base + sspi->regs->int_st),
+				sspi->base + sspi->regs->int_st);
 		return IRQ_HANDLED;
 	}
-
 	/* Error Conditions */
 	if (spi_stat & SIRFSOC_SPI_RX_OFLOW ||
 			spi_stat & SIRFSOC_SPI_TX_UFLOW) {
 		complete(&sspi->tx_done);
 		complete(&sspi->rx_done);
-		writel(0x0, sspi->base + SIRFSOC_SPI_INT_EN);
-		writel(SIRFSOC_SPI_INT_MASK_ALL,
-				sspi->base + SIRFSOC_SPI_INT_STATUS);
+		switch (sspi->type) {
+		case SIRF_REAL_SPI:
+		case SIRF_USP_SPI_P2:
+			writel(0x0, sspi->base + sspi->regs->int_en);
+			break;
+		case SIRF_USP_SPI_A7:
+			writel(~0UL, sspi->base + sspi->regs->usp_int_en_clr);
+			break;
+		}
+		writel(readl(sspi->base + sspi->regs->int_st),
+				sspi->base + sspi->regs->int_st);
 		return IRQ_HANDLED;
 	}
 	if (spi_stat & SIRFSOC_SPI_TXFIFO_EMPTY)
 		complete(&sspi->tx_done);
-	while (!(readl(sspi->base + SIRFSOC_SPI_INT_STATUS) &
+	while (!(readl(sspi->base + sspi->regs->int_st) &
 		SIRFSOC_SPI_RX_IO_DMA))
 		cpu_relax();
 	complete(&sspi->rx_done);
-	writel(0x0, sspi->base + SIRFSOC_SPI_INT_EN);
-	writel(SIRFSOC_SPI_INT_MASK_ALL,
-			sspi->base + SIRFSOC_SPI_INT_STATUS);
+	switch (sspi->type) {
+	case SIRF_REAL_SPI:
+	case SIRF_USP_SPI_P2:
+		writel(0x0, sspi->base + sspi->regs->int_en);
+		break;
+	case SIRF_USP_SPI_A7:
+		writel(~0UL, sspi->base + sspi->regs->usp_int_en_clr);
+		break;
+	}
+	writel(readl(sspi->base + sspi->regs->int_st),
+			sspi->base + sspi->regs->int_st);
 
 	return IRQ_HANDLED;
 }
@@ -313,8 +465,8 @@ static void spi_sirfsoc_cmd_transfer(struct spi_device *spi,
 	u32 cmd;
 
 	sspi = spi_master_get_devdata(spi->master);
-	writel(SIRFSOC_SPI_FIFO_RESET, sspi->base + SIRFSOC_SPI_TXFIFO_OP);
-	writel(SIRFSOC_SPI_FIFO_START, sspi->base + SIRFSOC_SPI_TXFIFO_OP);
+	writel(SIRFSOC_SPI_FIFO_RESET, sspi->base + sspi->regs->txfifo_op);
+	writel(SIRFSOC_SPI_FIFO_START, sspi->base + sspi->regs->txfifo_op);
 	memcpy(&cmd, sspi->tx, t->len);
 	if (sspi->word_width == 1 && !(spi->mode & SPI_LSB_FIRST))
 		cmd = cpu_to_be32(cmd) >>
@@ -322,11 +474,11 @@ static void spi_sirfsoc_cmd_transfer(struct spi_device *spi,
 	if (sspi->word_width == 2 && t->len == 4 &&
 			(!(spi->mode & SPI_LSB_FIRST)))
 		cmd = ((cmd & 0xffff) << 16) | (cmd >> 16);
-	writel(cmd, sspi->base + SIRFSOC_SPI_CMD);
+	writel(cmd, sspi->base + sspi->regs->spi_cmd);
 	writel(SIRFSOC_SPI_FRM_END_INT_EN,
-		sspi->base + SIRFSOC_SPI_INT_EN);
+		sspi->base + sspi->regs->int_en);
 	writel(SIRFSOC_SPI_CMD_TX_EN,
-		sspi->base + SIRFSOC_SPI_TX_RX_EN);
+		sspi->base + sspi->regs->tx_rx_en);
 	if (wait_for_completion_timeout(&sspi->tx_done, timeout) == 0) {
 		dev_err(&spi->dev, "cmd transfer timeout\n");
 		return;
@@ -342,25 +494,56 @@ static void spi_sirfsoc_dma_transfer(struct spi_device *spi,
 	int timeout = t->len * 10;
 
 	sspi = spi_master_get_devdata(spi->master);
-	writel(SIRFSOC_SPI_FIFO_RESET, sspi->base + SIRFSOC_SPI_RXFIFO_OP);
-	writel(SIRFSOC_SPI_FIFO_RESET, sspi->base + SIRFSOC_SPI_TXFIFO_OP);
-	writel(SIRFSOC_SPI_FIFO_START, sspi->base + SIRFSOC_SPI_RXFIFO_OP);
-	writel(SIRFSOC_SPI_FIFO_START, sspi->base + SIRFSOC_SPI_TXFIFO_OP);
-	writel(0, sspi->base + SIRFSOC_SPI_INT_EN);
-	writel(SIRFSOC_SPI_INT_MASK_ALL, sspi->base + SIRFSOC_SPI_INT_STATUS);
-	if (sspi->left_tx_word < SIRFSOC_SPI_DAT_FRM_LEN_MAX) {
-		writel(readl(sspi->base + SIRFSOC_SPI_CTRL) |
-			SIRFSOC_SPI_ENA_AUTO_CLR | SIRFSOC_SPI_MUL_DAT_MODE,
-			sspi->base + SIRFSOC_SPI_CTRL);
-		writel(sspi->left_tx_word - 1,
-				sspi->base + SIRFSOC_SPI_TX_DMA_IO_LEN);
-		writel(sspi->left_tx_word - 1,
-				sspi->base + SIRFSOC_SPI_RX_DMA_IO_LEN);
+	writel(SIRFSOC_SPI_FIFO_RESET, sspi->base + sspi->regs->rxfifo_op);
+	writel(SIRFSOC_SPI_FIFO_RESET, sspi->base + sspi->regs->txfifo_op);
+	switch (sspi->type) {
+	case SIRF_REAL_SPI:
+		writel(SIRFSOC_SPI_FIFO_START,
+			sspi->base + sspi->regs->rxfifo_op);
+		writel(SIRFSOC_SPI_FIFO_START,
+			sspi->base + sspi->regs->txfifo_op);
+		writel(0, sspi->base + sspi->regs->int_en);
+		break;
+	case SIRF_USP_SPI_P2:
+		writel(0x0, sspi->base + sspi->regs->rxfifo_op);
+		writel(0x0, sspi->base + sspi->regs->txfifo_op);
+		writel(0, sspi->base + sspi->regs->int_en);
+		break;
+	case SIRF_USP_SPI_A7:
+		writel(0x0, sspi->base + sspi->regs->rxfifo_op);
+		writel(0x0, sspi->base + sspi->regs->txfifo_op);
+		writel(~0UL, sspi->base + sspi->regs->usp_int_en_clr);
+		break;
+	}
+	writel(readl(sspi->base + sspi->regs->int_st),
+		sspi->base + sspi->regs->int_st);
+	if (sspi->left_tx_word < sspi->dat_max_frm_len) {
+		switch (sspi->type) {
+		case SIRF_REAL_SPI:
+			writel(readl(sspi->base + sspi->regs->spi_ctrl) |
+				SIRFSOC_SPI_ENA_AUTO_CLR |
+				SIRFSOC_SPI_MUL_DAT_MODE,
+				sspi->base + sspi->regs->spi_ctrl);
+			writel(sspi->left_tx_word - 1,
+				sspi->base + sspi->regs->tx_dma_io_len);
+			writel(sspi->left_tx_word - 1,
+				sspi->base + sspi->regs->rx_dma_io_len);
+			break;
+		case SIRF_USP_SPI_P2:
+		case SIRF_USP_SPI_A7:
+			/*USP simulate SPI, tx/rx_dma_io_len indicates bytes*/
+			writel(sspi->left_tx_word * sspi->word_width,
+				sspi->base + sspi->regs->tx_dma_io_len);
+			writel(sspi->left_tx_word * sspi->word_width,
+				sspi->base + sspi->regs->rx_dma_io_len);
+			break;
+		}
 	} else {
-		writel(readl(sspi->base + SIRFSOC_SPI_CTRL),
-			sspi->base + SIRFSOC_SPI_CTRL);
-		writel(0, sspi->base + SIRFSOC_SPI_TX_DMA_IO_LEN);
-		writel(0, sspi->base + SIRFSOC_SPI_RX_DMA_IO_LEN);
+		if (sspi->type == SIRF_REAL_SPI)
+			writel(readl(sspi->base + sspi->regs->spi_ctrl),
+				sspi->base + sspi->regs->spi_ctrl);
+		writel(0, sspi->base + sspi->regs->tx_dma_io_len);
+		writel(0, sspi->base + sspi->regs->rx_dma_io_len);
 	}
 	sspi->dst_start = dma_map_single(&spi->dev, sspi->rx, t->len,
 					(t->tx_buf != t->rx_buf) ?
@@ -385,7 +568,14 @@ static void spi_sirfsoc_dma_transfer(struct spi_device *spi,
 	dma_async_issue_pending(sspi->tx_chan);
 	dma_async_issue_pending(sspi->rx_chan);
 	writel(SIRFSOC_SPI_RX_EN | SIRFSOC_SPI_TX_EN,
-			sspi->base + SIRFSOC_SPI_TX_RX_EN);
+			sspi->base + sspi->regs->tx_rx_en);
+	if (sspi->type == SIRF_USP_SPI_P2 ||
+		sspi->type == SIRF_USP_SPI_A7) {
+		writel(SIRFSOC_SPI_FIFO_START,
+			sspi->base + sspi->regs->rxfifo_op);
+		writel(SIRFSOC_SPI_FIFO_START,
+			sspi->base + sspi->regs->txfifo_op);
+	}
 	if (wait_for_completion_timeout(&sspi->rx_done, timeout) == 0) {
 		dev_err(&spi->dev, "transfer timeout\n");
 		dmaengine_terminate_all(sspi->rx_chan);
@@ -398,15 +588,21 @@ static void spi_sirfsoc_dma_transfer(struct spi_device *spi,
 	 */
 	if (wait_for_completion_timeout(&sspi->tx_done, timeout) == 0) {
 		dev_err(&spi->dev, "transfer timeout\n");
+		if (sspi->type == SIRF_USP_SPI_P2 ||
+			sspi->type == SIRF_USP_SPI_A7)
+			writel(0, sspi->base + sspi->regs->tx_rx_en);
 		dmaengine_terminate_all(sspi->tx_chan);
 	}
 	dma_unmap_single(&spi->dev, sspi->src_start, t->len, DMA_TO_DEVICE);
 	dma_unmap_single(&spi->dev, sspi->dst_start, t->len, DMA_FROM_DEVICE);
 	/* TX, RX FIFO stop */
-	writel(0, sspi->base + SIRFSOC_SPI_RXFIFO_OP);
-	writel(0, sspi->base + SIRFSOC_SPI_TXFIFO_OP);
-	if (sspi->left_tx_word >= SIRFSOC_SPI_DAT_FRM_LEN_MAX)
-		writel(0, sspi->base + SIRFSOC_SPI_TX_RX_EN);
+	writel(0, sspi->base + sspi->regs->rxfifo_op);
+	writel(0, sspi->base + sspi->regs->txfifo_op);
+	if (sspi->left_tx_word >= sspi->dat_max_frm_len)
+		writel(0, sspi->base + sspi->regs->tx_rx_en);
+	if (sspi->type == SIRF_USP_SPI_P2 ||
+		sspi->type == SIRF_USP_SPI_A7)
+		writel(0, sspi->base + sspi->regs->tx_rx_en);
 }
 
 static void spi_sirfsoc_pio_transfer(struct spi_device *spi,
@@ -414,57 +610,105 @@ static void spi_sirfsoc_pio_transfer(struct spi_device *spi,
 {
 	struct sirfsoc_spi *sspi;
 	int timeout = t->len * 10;
+	unsigned int data_units;
 
 	sspi = spi_master_get_devdata(spi->master);
 	do {
 		writel(SIRFSOC_SPI_FIFO_RESET,
-			sspi->base + SIRFSOC_SPI_RXFIFO_OP);
+			sspi->base + sspi->regs->rxfifo_op);
 		writel(SIRFSOC_SPI_FIFO_RESET,
-			sspi->base + SIRFSOC_SPI_TXFIFO_OP);
-		writel(SIRFSOC_SPI_FIFO_START,
-			sspi->base + SIRFSOC_SPI_RXFIFO_OP);
-		writel(SIRFSOC_SPI_FIFO_START,
-			sspi->base + SIRFSOC_SPI_TXFIFO_OP);
-		writel(0, sspi->base + SIRFSOC_SPI_INT_EN);
-		writel(SIRFSOC_SPI_INT_MASK_ALL,
-			sspi->base + SIRFSOC_SPI_INT_STATUS);
-		writel(readl(sspi->base + SIRFSOC_SPI_CTRL) |
-			SIRFSOC_SPI_MUL_DAT_MODE | SIRFSOC_SPI_ENA_AUTO_CLR,
-			sspi->base + SIRFSOC_SPI_CTRL);
-		writel(min(sspi->left_tx_word, (u32)(256 / sspi->word_width))
-				- 1, sspi->base + SIRFSOC_SPI_TX_DMA_IO_LEN);
-		writel(min(sspi->left_rx_word, (u32)(256 / sspi->word_width))
-				- 1, sspi->base + SIRFSOC_SPI_RX_DMA_IO_LEN);
-		while (!((readl(sspi->base + SIRFSOC_SPI_TXFIFO_STATUS)
-			& SIRFSOC_SPI_FIFO_FULL)) && sspi->left_tx_word)
+			sspi->base + sspi->regs->txfifo_op);
+		switch (sspi->type) {
+		case SIRF_USP_SPI_P2:
+			writel(0x0, sspi->base + sspi->regs->rxfifo_op);
+			writel(0x0, sspi->base + sspi->regs->txfifo_op);
+			writel(0, sspi->base + sspi->regs->int_en);
+			writel(readl(sspi->base + sspi->regs->int_st),
+				sspi->base + sspi->regs->int_st);
+			writel(min((sspi->left_tx_word * sspi->word_width),
+				sspi->fifo_size),
+				sspi->base + sspi->regs->tx_dma_io_len);
+			writel(min((sspi->left_rx_word * sspi->word_width),
+				sspi->fifo_size),
+				sspi->base + sspi->regs->rx_dma_io_len);
+			break;
+		case SIRF_USP_SPI_A7:
+			writel(0x0, sspi->base + sspi->regs->rxfifo_op);
+			writel(0x0, sspi->base + sspi->regs->txfifo_op);
+			writel(~0UL, sspi->base + sspi->regs->usp_int_en_clr);
+			writel(readl(sspi->base + sspi->regs->int_st),
+				sspi->base + sspi->regs->int_st);
+			writel(min((sspi->left_tx_word * sspi->word_width),
+				sspi->fifo_size),
+				sspi->base + sspi->regs->tx_dma_io_len);
+			writel(min((sspi->left_rx_word * sspi->word_width),
+				sspi->fifo_size),
+				sspi->base + sspi->regs->rx_dma_io_len);
+			break;
+		case SIRF_REAL_SPI:
+			writel(SIRFSOC_SPI_FIFO_START,
+				sspi->base + sspi->regs->rxfifo_op);
+			writel(SIRFSOC_SPI_FIFO_START,
+				sspi->base + sspi->regs->txfifo_op);
+			writel(0, sspi->base + sspi->regs->int_en);
+			writel(readl(sspi->base + sspi->regs->int_st),
+				sspi->base + sspi->regs->int_st);
+			writel(readl(sspi->base + sspi->regs->spi_ctrl) |
+				SIRFSOC_SPI_MUL_DAT_MODE |
+				SIRFSOC_SPI_ENA_AUTO_CLR,
+				sspi->base + sspi->regs->spi_ctrl);
+			data_units = sspi->fifo_size / sspi->word_width;
+			writel(min(sspi->left_tx_word, data_units) - 1,
+				sspi->base + sspi->regs->tx_dma_io_len);
+			writel(min(sspi->left_rx_word, data_units) - 1,
+				sspi->base + sspi->regs->rx_dma_io_len);
+			break;
+		}
+		while (!((readl(sspi->base + sspi->regs->txfifo_st)
+			& SIRFSOC_SPI_FIFO_FULL_MASK(sspi))) &&
+			sspi->left_tx_word)
 			sspi->tx_word(sspi);
 		writel(SIRFSOC_SPI_TXFIFO_EMPTY_INT_EN |
 			SIRFSOC_SPI_TX_UFLOW_INT_EN |
 			SIRFSOC_SPI_RX_OFLOW_INT_EN |
 			SIRFSOC_SPI_RX_IO_DMA_INT_EN,
-			sspi->base + SIRFSOC_SPI_INT_EN);
+			sspi->base + sspi->regs->int_en);
 		writel(SIRFSOC_SPI_RX_EN | SIRFSOC_SPI_TX_EN,
-			sspi->base + SIRFSOC_SPI_TX_RX_EN);
+			sspi->base + sspi->regs->tx_rx_en);
+		if (sspi->type == SIRF_USP_SPI_P2 ||
+			sspi->type == SIRF_USP_SPI_A7) {
+			writel(SIRFSOC_SPI_FIFO_START,
+				sspi->base + sspi->regs->rxfifo_op);
+			writel(SIRFSOC_SPI_FIFO_START,
+				sspi->base + sspi->regs->txfifo_op);
+		}
 		if (!wait_for_completion_timeout(&sspi->tx_done, timeout) ||
 			!wait_for_completion_timeout(&sspi->rx_done, timeout)) {
 			dev_err(&spi->dev, "transfer timeout\n");
+			if (sspi->type == SIRF_USP_SPI_P2 ||
+				sspi->type == SIRF_USP_SPI_A7)
+				writel(0, sspi->base + sspi->regs->tx_rx_en);
 			break;
 		}
-		while (!((readl(sspi->base + SIRFSOC_SPI_RXFIFO_STATUS)
-			& SIRFSOC_SPI_FIFO_EMPTY)) && sspi->left_rx_word)
+		while (!((readl(sspi->base + sspi->regs->rxfifo_st)
+			& SIRFSOC_SPI_FIFO_EMPTY_MASK(sspi))) &&
+			sspi->left_rx_word)
 			sspi->rx_word(sspi);
-		writel(0, sspi->base + SIRFSOC_SPI_RXFIFO_OP);
-		writel(0, sspi->base + SIRFSOC_SPI_TXFIFO_OP);
+		if (sspi->type == SIRF_USP_SPI_P2 ||
+			sspi->type == SIRF_USP_SPI_A7)
+			writel(0, sspi->base + sspi->regs->tx_rx_en);
+		writel(0, sspi->base + sspi->regs->rxfifo_op);
+		writel(0, sspi->base + sspi->regs->txfifo_op);
 	} while (sspi->left_tx_word != 0 || sspi->left_rx_word != 0);
 }
 
 static int spi_sirfsoc_transfer(struct spi_device *spi, struct spi_transfer *t)
 {
 	struct sirfsoc_spi *sspi;
-	sspi = spi_master_get_devdata(spi->master);
 
-	sspi->tx = t->tx_buf ? t->tx_buf : sspi->dummypage;
-	sspi->rx = t->rx_buf ? t->rx_buf : sspi->dummypage;
+	sspi = spi_master_get_devdata(spi->master);
+	sspi->tx = t->tx_buf;
+	sspi->rx = t->rx_buf;
 	sspi->left_tx_word = sspi->left_rx_word = t->len / sspi->word_width;
 	reinit_completion(&sspi->rx_done);
 	reinit_completion(&sspi->tx_done);
@@ -473,7 +717,7 @@ static int spi_sirfsoc_transfer(struct spi_device *spi, struct spi_transfer *t)
 	 * null, just fill command data into command register and wait for its
 	 * completion.
 	 */
-	if (sspi->tx_by_cmd)
+	if (sspi->type == SIRF_REAL_SPI && sspi->tx_by_cmd)
 		spi_sirfsoc_cmd_transfer(spi, t);
 	else if (IS_DMA_VALID(t))
 		spi_sirfsoc_dma_transfer(spi, t);
@@ -488,22 +732,49 @@ static void spi_sirfsoc_chipselect(struct spi_device *spi, int value)
 	struct sirfsoc_spi *sspi = spi_master_get_devdata(spi->master);
 
 	if (sspi->hw_cs) {
-		u32 regval = readl(sspi->base + SIRFSOC_SPI_CTRL);
-		switch (value) {
-		case BITBANG_CS_ACTIVE:
-			if (spi->mode & SPI_CS_HIGH)
-				regval |= SIRFSOC_SPI_CS_IO_OUT;
-			else
-				regval &= ~SIRFSOC_SPI_CS_IO_OUT;
+		u32 regval;
+
+		switch (sspi->type) {
+		case SIRF_REAL_SPI:
+			regval = readl(sspi->base + sspi->regs->spi_ctrl);
+			switch (value) {
+			case BITBANG_CS_ACTIVE:
+				if (spi->mode & SPI_CS_HIGH)
+					regval |= SIRFSOC_SPI_CS_IO_OUT;
+				else
+					regval &= ~SIRFSOC_SPI_CS_IO_OUT;
+				break;
+			case BITBANG_CS_INACTIVE:
+				if (spi->mode & SPI_CS_HIGH)
+					regval &= ~SIRFSOC_SPI_CS_IO_OUT;
+				else
+					regval |= SIRFSOC_SPI_CS_IO_OUT;
+				break;
+			}
+			writel(regval, sspi->base + sspi->regs->spi_ctrl);
 			break;
-		case BITBANG_CS_INACTIVE:
-			if (spi->mode & SPI_CS_HIGH)
-				regval &= ~SIRFSOC_SPI_CS_IO_OUT;
-			else
-				regval |= SIRFSOC_SPI_CS_IO_OUT;
+		case SIRF_USP_SPI_P2:
+		case SIRF_USP_SPI_A7:
+			regval = readl(sspi->base +
+					sspi->regs->usp_pin_io_data);
+			switch (value) {
+			case BITBANG_CS_ACTIVE:
+				if (spi->mode & SPI_CS_HIGH)
+					regval |= SIRFSOC_USP_CS_HIGH_VALUE;
+				else
+					regval &= ~(SIRFSOC_USP_CS_HIGH_VALUE);
+				break;
+			case BITBANG_CS_INACTIVE:
+				if (spi->mode & SPI_CS_HIGH)
+					regval &= ~(SIRFSOC_USP_CS_HIGH_VALUE);
+				else
+					regval |= SIRFSOC_USP_CS_HIGH_VALUE;
+				break;
+			}
+			writel(regval,
+				sspi->base + sspi->regs->usp_pin_io_data);
 			break;
 		}
-		writel(regval, sspi->base + SIRFSOC_SPI_CTRL);
 	} else {
 		switch (value) {
 		case BITBANG_CS_ACTIVE:
@@ -518,27 +789,102 @@ static void spi_sirfsoc_chipselect(struct spi_device *spi, int value)
 	}
 }
 
+static int spi_sirfsoc_config_mode(struct spi_device *spi)
+{
+	struct sirfsoc_spi *sspi;
+	u32 regval, usp_mode1;
+
+	sspi = spi_master_get_devdata(spi->master);
+	regval = readl(sspi->base + sspi->regs->spi_ctrl);
+	usp_mode1 = readl(sspi->base + sspi->regs->usp_mode1);
+	if (!(spi->mode & SPI_CS_HIGH)) {
+		regval |= SIRFSOC_SPI_CS_IDLE_STAT;
+		usp_mode1 &= ~SIRFSOC_USP_CS_HIGH_VALID;
+	} else {
+		regval &= ~SIRFSOC_SPI_CS_IDLE_STAT;
+		usp_mode1 |= SIRFSOC_USP_CS_HIGH_VALID;
+	}
+	if (!(spi->mode & SPI_LSB_FIRST)) {
+		regval |= SIRFSOC_SPI_TRAN_MSB;
+		usp_mode1 &= ~SIRFSOC_USP_LSB;
+	} else {
+		regval &= ~SIRFSOC_SPI_TRAN_MSB;
+		usp_mode1 |= SIRFSOC_USP_LSB;
+	}
+	if (spi->mode & SPI_CPOL) {
+		regval |= SIRFSOC_SPI_CLK_IDLE_STAT;
+		usp_mode1 |= SIRFSOC_USP_SCLK_IDLE_STAT;
+	} else {
+		regval &= ~SIRFSOC_SPI_CLK_IDLE_STAT;
+		usp_mode1 &= ~SIRFSOC_USP_SCLK_IDLE_STAT;
+	}
+	/*
+	 * Data should be driven at least 1/2 cycle before the fetch edge
+	 * to make sure that data gets stable at the fetch edge.
+	 */
+	if (((spi->mode & SPI_CPOL) && (spi->mode & SPI_CPHA)) ||
+	    (!(spi->mode & SPI_CPOL) && !(spi->mode & SPI_CPHA))) {
+		regval &= ~SIRFSOC_SPI_DRV_POS_EDGE;
+		usp_mode1 |= (SIRFSOC_USP_TXD_FALLING_EDGE |
+				SIRFSOC_USP_RXD_FALLING_EDGE);
+	} else {
+		regval |= SIRFSOC_SPI_DRV_POS_EDGE;
+		usp_mode1 &= ~(SIRFSOC_USP_RXD_FALLING_EDGE |
+				SIRFSOC_USP_TXD_FALLING_EDGE);
+	}
+	writel((SIRFSOC_SPI_FIFO_LEVEL_CHK_MASK(sspi, sspi->fifo_size - 2) <<
+		SIRFSOC_SPI_FIFO_SC_OFFSET) |
+		(SIRFSOC_SPI_FIFO_LEVEL_CHK_MASK(sspi, sspi->fifo_size / 2) <<
+		SIRFSOC_SPI_FIFO_LC_OFFSET) |
+		(SIRFSOC_SPI_FIFO_LEVEL_CHK_MASK(sspi, 2) <<
+		SIRFSOC_SPI_FIFO_HC_OFFSET),
+		sspi->base + sspi->regs->txfifo_level_chk);
+	writel((SIRFSOC_SPI_FIFO_LEVEL_CHK_MASK(sspi, 2) <<
+		SIRFSOC_SPI_FIFO_SC_OFFSET) |
+		(SIRFSOC_SPI_FIFO_LEVEL_CHK_MASK(sspi, sspi->fifo_size / 2) <<
+		SIRFSOC_SPI_FIFO_LC_OFFSET) |
+		(SIRFSOC_SPI_FIFO_LEVEL_CHK_MASK(sspi, sspi->fifo_size - 2) <<
+		SIRFSOC_SPI_FIFO_HC_OFFSET),
+		sspi->base + sspi->regs->rxfifo_level_chk);
+	/*
+	 * it should never set to hardware cs mode because in hardware cs mode,
+	 * cs signal can't controlled by driver.
+	 */
+	switch (sspi->type) {
+	case SIRF_REAL_SPI:
+		regval |= SIRFSOC_SPI_CS_IO_MODE;
+		writel(regval, sspi->base + sspi->regs->spi_ctrl);
+		break;
+	case SIRF_USP_SPI_P2:
+	case SIRF_USP_SPI_A7:
+		usp_mode1 |= SIRFSOC_USP_SYNC_MODE;
+		usp_mode1 |= SIRFSOC_USP_TFS_IO_MODE;
+		usp_mode1 &= ~SIRFSOC_USP_TFS_IO_INPUT;
+		writel(usp_mode1, sspi->base + sspi->regs->usp_mode1);
+		break;
+	}
+
+	return 0;
+}
+
 static int
 spi_sirfsoc_setup_transfer(struct spi_device *spi, struct spi_transfer *t)
 {
 	struct sirfsoc_spi *sspi;
 	u8 bits_per_word = 0;
 	int hz = 0;
-	u32 regval;
-	u32 txfifo_ctrl, rxfifo_ctrl;
-	u32 fifo_size = SIRFSOC_SPI_FIFO_SIZE / 4;
+	u32 regval, txfifo_ctrl, rxfifo_ctrl, tx_frm_ctl, rx_frm_ctl, usp_mode2;
 
 	sspi = spi_master_get_devdata(spi->master);
 
 	bits_per_word = (t) ? t->bits_per_word : spi->bits_per_word;
 	hz = t && t->speed_hz ? t->speed_hz : spi->max_speed_hz;
 
-	regval = (sspi->ctrl_freq / (2 * hz)) - 1;
+	usp_mode2 = regval = (sspi->ctrl_freq / (2 * hz)) - 1;
 	if (regval > 0xFFFF || regval < 0) {
 		dev_err(&spi->dev, "Speed %d not supported\n", hz);
 		return -EINVAL;
 	}
-
 	switch (bits_per_word) {
 	case 8:
 		regval |= SIRFSOC_SPI_TRAN_DAT_FORMAT_8;
@@ -559,94 +905,177 @@ spi_sirfsoc_setup_transfer(struct spi_device *spi, struct spi_transfer *t)
 		sspi->tx_word = spi_sirfsoc_tx_word_u32;
 		break;
 	default:
-		BUG();
+		dev_err(&spi->dev, "bpw %d not supported\n", bits_per_word);
+		return -EINVAL;
 	}
-
 	sspi->word_width = DIV_ROUND_UP(bits_per_word, 8);
-	txfifo_ctrl = SIRFSOC_SPI_FIFO_THD(SIRFSOC_SPI_FIFO_SIZE / 2) |
-					   (sspi->word_width >> 1);
-	rxfifo_ctrl = SIRFSOC_SPI_FIFO_THD(SIRFSOC_SPI_FIFO_SIZE / 2) |
-					   (sspi->word_width >> 1);
-
-	if (!(spi->mode & SPI_CS_HIGH))
-		regval |= SIRFSOC_SPI_CS_IDLE_STAT;
-	if (!(spi->mode & SPI_LSB_FIRST))
-		regval |= SIRFSOC_SPI_TRAN_MSB;
-	if (spi->mode & SPI_CPOL)
-		regval |= SIRFSOC_SPI_CLK_IDLE_STAT;
-
-	/*
-	 * Data should be driven at least 1/2 cycle before the fetch edge
-	 * to make sure that data gets stable at the fetch edge.
-	 */
-	if (((spi->mode & SPI_CPOL) && (spi->mode & SPI_CPHA)) ||
-	    (!(spi->mode & SPI_CPOL) && !(spi->mode & SPI_CPHA)))
-		regval &= ~SIRFSOC_SPI_DRV_POS_EDGE;
-	else
-		regval |= SIRFSOC_SPI_DRV_POS_EDGE;
-
-	writel(SIRFSOC_SPI_FIFO_SC(fifo_size - 2) |
-			SIRFSOC_SPI_FIFO_LC(fifo_size / 2) |
-			SIRFSOC_SPI_FIFO_HC(2),
-		sspi->base + SIRFSOC_SPI_TXFIFO_LEVEL_CHK);
-	writel(SIRFSOC_SPI_FIFO_SC(2) |
-			SIRFSOC_SPI_FIFO_LC(fifo_size / 2) |
-			SIRFSOC_SPI_FIFO_HC(fifo_size - 2),
-		sspi->base + SIRFSOC_SPI_RXFIFO_LEVEL_CHK);
-	writel(txfifo_ctrl, sspi->base + SIRFSOC_SPI_TXFIFO_CTRL);
-	writel(rxfifo_ctrl, sspi->base + SIRFSOC_SPI_RXFIFO_CTRL);
-
-	if (t && t->tx_buf && !t->rx_buf && (t->len <= SIRFSOC_MAX_CMD_BYTES)) {
-		regval |= (SIRFSOC_SPI_CMD_BYTE_NUM((t->len - 1)) |
-				SIRFSOC_SPI_CMD_MODE);
-		sspi->tx_by_cmd = true;
-	} else {
-		regval &= ~SIRFSOC_SPI_CMD_MODE;
-		sspi->tx_by_cmd = false;
+	txfifo_ctrl = (((sspi->fifo_size / 2) &
+			SIRFSOC_SPI_FIFO_THD_MASK(sspi))
+			<< SIRFSOC_SPI_FIFO_THD_OFFSET) |
+			(sspi->word_width >> 1);
+	rxfifo_ctrl = (((sspi->fifo_size / 2) &
+			SIRFSOC_SPI_FIFO_THD_MASK(sspi))
+			<< SIRFSOC_SPI_FIFO_THD_OFFSET) |
+			(sspi->word_width >> 1);
+	writel(txfifo_ctrl, sspi->base + sspi->regs->txfifo_ctrl);
+	writel(rxfifo_ctrl, sspi->base + sspi->regs->rxfifo_ctrl);
+	if (sspi->type == SIRF_USP_SPI_P2 ||
+		sspi->type == SIRF_USP_SPI_A7) {
+		tx_frm_ctl = 0;
+		tx_frm_ctl |= ((bits_per_word - 1) & SIRFSOC_USP_TX_DATA_MASK)
+				<< SIRFSOC_USP_TX_DATA_OFFSET;
+		tx_frm_ctl |= ((bits_per_word + 1 + SIRFSOC_USP_TXD_DELAY_LEN
+				- 1) & SIRFSOC_USP_TX_SYNC_MASK) <<
+				SIRFSOC_USP_TX_SYNC_OFFSET;
+		tx_frm_ctl |= ((bits_per_word + 1 + SIRFSOC_USP_TXD_DELAY_LEN
+				+ 2 - 1) & SIRFSOC_USP_TX_FRAME_MASK) <<
+				SIRFSOC_USP_TX_FRAME_OFFSET;
+		tx_frm_ctl |= ((bits_per_word - 1) &
+				SIRFSOC_USP_TX_SHIFTER_MASK) <<
+				SIRFSOC_USP_TX_SHIFTER_OFFSET;
+		rx_frm_ctl = 0;
+		rx_frm_ctl |= ((bits_per_word - 1) & SIRFSOC_USP_RX_DATA_MASK)
+				<< SIRFSOC_USP_RX_DATA_OFFSET;
+		rx_frm_ctl |= ((bits_per_word + 1 + SIRFSOC_USP_RXD_DELAY_LEN
+				+ 2 - 1) & SIRFSOC_USP_RX_FRAME_MASK) <<
+				SIRFSOC_USP_RX_FRAME_OFFSET;
+		rx_frm_ctl |= ((bits_per_word - 1)
+				& SIRFSOC_USP_RX_SHIFTER_MASK) <<
+				SIRFSOC_USP_RX_SHIFTER_OFFSET;
+		writel(tx_frm_ctl | (((usp_mode2 >> 10) &
+			SIRFSOC_USP_CLK_10_11_MASK) <<
+			SIRFSOC_USP_CLK_10_11_OFFSET),
+			sspi->base + sspi->regs->usp_tx_frame_ctrl);
+		writel(rx_frm_ctl | (((usp_mode2 >> 12) &
+			SIRFSOC_USP_CLK_12_15_MASK) <<
+			SIRFSOC_USP_CLK_12_15_OFFSET),
+			sspi->base + sspi->regs->usp_rx_frame_ctrl);
+		writel(readl(sspi->base + sspi->regs->usp_mode2) |
+			((usp_mode2 & SIRFSOC_USP_CLK_DIVISOR_MASK) <<
+			SIRFSOC_USP_CLK_DIVISOR_OFFSET) |
+			(SIRFSOC_USP_RXD_DELAY_LEN <<
+			 SIRFSOC_USP_RXD_DELAY_OFFSET) |
+			(SIRFSOC_USP_TXD_DELAY_LEN <<
+			 SIRFSOC_USP_TXD_DELAY_OFFSET),
+			sspi->base + sspi->regs->usp_mode2);
+	}
+	if (sspi->type == SIRF_REAL_SPI)
+		writel(regval, sspi->base + sspi->regs->spi_ctrl);
+	spi_sirfsoc_config_mode(spi);
+	if (sspi->type == SIRF_REAL_SPI) {
+		if (t && t->tx_buf && !t->rx_buf &&
+			(t->len <= SIRFSOC_MAX_CMD_BYTES)) {
+			sspi->tx_by_cmd = true;
+			writel(readl(sspi->base + sspi->regs->spi_ctrl) |
+				(SIRFSOC_SPI_CMD_BYTE_NUM((t->len - 1)) |
+				SIRFSOC_SPI_CMD_MODE),
+				sspi->base + sspi->regs->spi_ctrl);
+		} else {
+			sspi->tx_by_cmd = false;
+			writel(readl(sspi->base + sspi->regs->spi_ctrl) &
+				~SIRFSOC_SPI_CMD_MODE,
+				sspi->base + sspi->regs->spi_ctrl);
+		}
 	}
-	/*
-	 * it should never set to hardware cs mode because in hardware cs mode,
-	 * cs signal can't controlled by driver.
-	 */
-	regval |= SIRFSOC_SPI_CS_IO_MODE;
-	writel(regval, sspi->base + SIRFSOC_SPI_CTRL);
-
 	if (IS_DMA_VALID(t)) {
 		/* Enable DMA mode for RX, TX */
-		writel(0, sspi->base + SIRFSOC_SPI_TX_DMA_IO_CTRL);
+		writel(0, sspi->base + sspi->regs->tx_dma_io_ctrl);
 		writel(SIRFSOC_SPI_RX_DMA_FLUSH,
-			sspi->base + SIRFSOC_SPI_RX_DMA_IO_CTRL);
+			sspi->base + sspi->regs->rx_dma_io_ctrl);
 	} else {
 		/* Enable IO mode for RX, TX */
 		writel(SIRFSOC_SPI_IO_MODE_SEL,
-			sspi->base + SIRFSOC_SPI_TX_DMA_IO_CTRL);
+			sspi->base + sspi->regs->tx_dma_io_ctrl);
 		writel(SIRFSOC_SPI_IO_MODE_SEL,
-			sspi->base + SIRFSOC_SPI_RX_DMA_IO_CTRL);
+			sspi->base + sspi->regs->rx_dma_io_ctrl);
 	}
-
 	return 0;
 }
 
 static int spi_sirfsoc_setup(struct spi_device *spi)
 {
 	struct sirfsoc_spi *sspi;
+	int ret = 0;
 
 	sspi = spi_master_get_devdata(spi->master);
-
 	if (spi->cs_gpio == -ENOENT)
 		sspi->hw_cs = true;
-	else
+	else {
 		sspi->hw_cs = false;
-	return spi_sirfsoc_setup_transfer(spi, NULL);
+		if (!spi_get_ctldata(spi)) {
+			void *cs = kmalloc(sizeof(int), GFP_KERNEL);
+			if (!cs) {
+				ret = -ENOMEM;
+				goto exit;
+			}
+			ret = gpio_is_valid(spi->cs_gpio);
+			if (!ret) {
+				dev_err(&spi->dev, "no valid gpio\n");
+				ret = -ENOENT;
+				goto exit;
+			}
+			ret = gpio_request(spi->cs_gpio, DRIVER_NAME);
+			if (ret) {
+				dev_err(&spi->dev, "failed to request gpio\n");
+				goto exit;
+			}
+			spi_set_ctldata(spi, cs);
+		}
+	}
+	spi_sirfsoc_config_mode(spi);
+	spi_sirfsoc_chipselect(spi, BITBANG_CS_INACTIVE);
+exit:
+	return ret;
+}
+
+static void spi_sirfsoc_cleanup(struct spi_device *spi)
+{
+	if (spi_get_ctldata(spi)) {
+		gpio_free(spi->cs_gpio);
+		kfree(spi_get_ctldata(spi));
+	}
 }
 
+static const struct sirf_spi_comp_data sirf_real_spi = {
+	.regs = &real_spi_register,
+	.type = SIRF_REAL_SPI,
+	.dat_max_frm_len = 64 * 1024,
+	.fifo_size = 256,
+};
+
+static const struct sirf_spi_comp_data sirf_usp_spi_p2 = {
+	.regs = &usp_spi_register,
+	.type = SIRF_USP_SPI_P2,
+	.dat_max_frm_len = 1024 * 1024,
+	.fifo_size = 128,
+	.hwinit = sirfsoc_usp_hwinit,
+};
+
+static const struct sirf_spi_comp_data sirf_usp_spi_a7 = {
+	.regs = &usp_spi_register,
+	.type = SIRF_USP_SPI_A7,
+	.dat_max_frm_len = 1024 * 1024,
+	.fifo_size = 512,
+	.hwinit = sirfsoc_usp_hwinit,
+};
+
+static const struct of_device_id spi_sirfsoc_of_match[] = {
+	{ .compatible = "sirf,prima2-spi", .data = &sirf_real_spi},
+	{ .compatible = "sirf,prima2-usp-spi", .data = &sirf_usp_spi_p2},
+	{ .compatible = "sirf,atlas7-usp-spi", .data = &sirf_usp_spi_a7},
+	{}
+};
+MODULE_DEVICE_TABLE(of, spi_sirfsoc_of_match);
+
 static int spi_sirfsoc_probe(struct platform_device *pdev)
 {
 	struct sirfsoc_spi *sspi;
 	struct spi_master *master;
 	struct resource *mem_res;
+	struct sirf_spi_comp_data *spi_comp_data;
 	int irq;
-	int i, ret;
+	int ret;
+	const struct of_device_id *match;
 
 	ret = device_reset(&pdev->dev);
 	if (ret) {
@@ -659,16 +1088,22 @@ static int spi_sirfsoc_probe(struct platform_device *pdev)
 		dev_err(&pdev->dev, "Unable to allocate SPI master\n");
 		return -ENOMEM;
 	}
+	match = of_match_node(spi_sirfsoc_of_match, pdev->dev.of_node);
 	platform_set_drvdata(pdev, master);
 	sspi = spi_master_get_devdata(master);
-
+	sspi->fifo_full_offset = ilog2(sspi->fifo_size);
+	spi_comp_data = (struct sirf_spi_comp_data *)match->data;
+	sspi->regs = spi_comp_data->regs;
+	sspi->type = spi_comp_data->type;
+	sspi->fifo_level_chk_mask = (sspi->fifo_size / 4) - 1;
+	sspi->dat_max_frm_len = spi_comp_data->dat_max_frm_len;
+	sspi->fifo_size = spi_comp_data->fifo_size;
 	mem_res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	sspi->base = devm_ioremap_resource(&pdev->dev, mem_res);
 	if (IS_ERR(sspi->base)) {
 		ret = PTR_ERR(sspi->base);
 		goto free_master;
 	}
-
 	irq = platform_get_irq(pdev, 0);
 	if (irq < 0) {
 		ret = -ENXIO;
@@ -684,11 +1119,13 @@ static int spi_sirfsoc_probe(struct platform_device *pdev)
 	sspi->bitbang.setup_transfer = spi_sirfsoc_setup_transfer;
 	sspi->bitbang.txrx_bufs = spi_sirfsoc_transfer;
 	sspi->bitbang.master->setup = spi_sirfsoc_setup;
+	sspi->bitbang.master->cleanup = spi_sirfsoc_cleanup;
 	master->bus_num = pdev->id;
 	master->mode_bits = SPI_CPOL | SPI_CPHA | SPI_LSB_FIRST | SPI_CS_HIGH;
 	master->bits_per_word_mask = SPI_BPW_MASK(8) | SPI_BPW_MASK(12) |
 					SPI_BPW_MASK(16) | SPI_BPW_MASK(32);
 	master->max_speed_hz = SIRFSOC_SPI_DEFAULT_FRQ;
+	master->flags = SPI_MASTER_MUST_RX | SPI_MASTER_MUST_TX;
 	sspi->bitbang.master->dev.of_node = pdev->dev.of_node;
 
 	/* request DMA channels */
@@ -711,47 +1148,19 @@ static int spi_sirfsoc_probe(struct platform_device *pdev)
 		goto free_tx_dma;
 	}
 	clk_prepare_enable(sspi->clk);
+	if (spi_comp_data->hwinit)
+		spi_comp_data->hwinit(sspi);
 	sspi->ctrl_freq = clk_get_rate(sspi->clk);
 
 	init_completion(&sspi->rx_done);
 	init_completion(&sspi->tx_done);
 
-	writel(SIRFSOC_SPI_FIFO_RESET, sspi->base + SIRFSOC_SPI_RXFIFO_OP);
-	writel(SIRFSOC_SPI_FIFO_RESET, sspi->base + SIRFSOC_SPI_TXFIFO_OP);
-	writel(SIRFSOC_SPI_FIFO_START, sspi->base + SIRFSOC_SPI_RXFIFO_OP);
-	writel(SIRFSOC_SPI_FIFO_START, sspi->base + SIRFSOC_SPI_TXFIFO_OP);
-	/* We are not using dummy delay between command and data */
-	writel(0, sspi->base + SIRFSOC_SPI_DUMMY_DELAY_CTL);
-
-	sspi->dummypage = kmalloc(2 * PAGE_SIZE, GFP_KERNEL);
-	if (!sspi->dummypage) {
-		ret = -ENOMEM;
-		goto free_clk;
-	}
-
 	ret = spi_bitbang_start(&sspi->bitbang);
 	if (ret)
-		goto free_dummypage;
-	for (i = 0; master->cs_gpios && i < master->num_chipselect; i++) {
-		if (master->cs_gpios[i] == -ENOENT)
-			continue;
-		if (!gpio_is_valid(master->cs_gpios[i])) {
-			dev_err(&pdev->dev, "no valid gpio\n");
-			ret = -EINVAL;
-			goto free_dummypage;
-		}
-		ret = devm_gpio_request(&pdev->dev,
-				master->cs_gpios[i], DRIVER_NAME);
-		if (ret) {
-			dev_err(&pdev->dev, "failed to request gpio\n");
-			goto free_dummypage;
-		}
-	}
+		goto free_clk;
 	dev_info(&pdev->dev, "registerred, bus number = %d\n", master->bus_num);
 
 	return 0;
-free_dummypage:
-	kfree(sspi->dummypage);
 free_clk:
 	clk_disable_unprepare(sspi->clk);
 	clk_put(sspi->clk);
@@ -772,9 +1181,7 @@ static int  spi_sirfsoc_remove(struct platform_device *pdev)
 
 	master = platform_get_drvdata(pdev);
 	sspi = spi_master_get_devdata(master);
-
 	spi_bitbang_stop(&sspi->bitbang);
-	kfree(sspi->dummypage);
 	clk_disable_unprepare(sspi->clk);
 	clk_put(sspi->clk);
 	dma_release_channel(sspi->rx_chan);
@@ -804,24 +1211,17 @@ static int spi_sirfsoc_resume(struct device *dev)
 	struct sirfsoc_spi *sspi = spi_master_get_devdata(master);
 
 	clk_enable(sspi->clk);
-	writel(SIRFSOC_SPI_FIFO_RESET, sspi->base + SIRFSOC_SPI_RXFIFO_OP);
-	writel(SIRFSOC_SPI_FIFO_RESET, sspi->base + SIRFSOC_SPI_TXFIFO_OP);
-	writel(SIRFSOC_SPI_FIFO_START, sspi->base + SIRFSOC_SPI_RXFIFO_OP);
-	writel(SIRFSOC_SPI_FIFO_START, sspi->base + SIRFSOC_SPI_TXFIFO_OP);
-
-	return spi_master_resume(master);
+	writel(SIRFSOC_SPI_FIFO_RESET, sspi->base + sspi->regs->txfifo_op);
+	writel(SIRFSOC_SPI_FIFO_RESET, sspi->base + sspi->regs->rxfifo_op);
+	writel(SIRFSOC_SPI_FIFO_START, sspi->base + sspi->regs->txfifo_op);
+	writel(SIRFSOC_SPI_FIFO_START, sspi->base + sspi->regs->rxfifo_op);
+	return 0;
 }
 #endif
 
 static SIMPLE_DEV_PM_OPS(spi_sirfsoc_pm_ops, spi_sirfsoc_suspend,
 			 spi_sirfsoc_resume);
 
-static const struct of_device_id spi_sirfsoc_of_match[] = {
-	{ .compatible = "sirf,prima2-spi", },
-	{}
-};
-MODULE_DEVICE_TABLE(of, spi_sirfsoc_of_match);
-
 static struct platform_driver spi_sirfsoc_driver = {
 	.driver = {
 		.name = DRIVER_NAME,
@@ -835,4 +1235,5 @@ module_platform_driver(spi_sirfsoc_driver);
 MODULE_DESCRIPTION("SiRF SoC SPI master driver");
 MODULE_AUTHOR("Zhiwu Song <Zhiwu.Song@csr.com>");
 MODULE_AUTHOR("Barry Song <Baohua.Song@csr.com>");
+MODULE_AUTHOR("Qipan Li <Qipan.Li@csr.com>");
 MODULE_LICENSE("GPL v2");
diff --git a/drivers/spi/spi-st-ssc4.c b/drivers/spi/spi-st-ssc4.c
index 2faeaa7b57a8..f17c0abe299f 100644
--- a/drivers/spi/spi-st-ssc4.c
+++ b/drivers/spi/spi-st-ssc4.c
@@ -482,7 +482,7 @@ static const struct dev_pm_ops spi_st_pm = {
 	SET_RUNTIME_PM_OPS(spi_st_runtime_suspend, spi_st_runtime_resume, NULL)
 };
 
-static struct of_device_id stm_spi_match[] = {
+static const struct of_device_id stm_spi_match[] = {
 	{ .compatible = "st,comms-ssc4-spi", },
 	{},
 };
diff --git a/drivers/spi/spi-zynqmp-gqspi.c b/drivers/spi/spi-zynqmp-gqspi.c
new file mode 100644
index 000000000000..87b20a511a6b
--- /dev/null
+++ b/drivers/spi/spi-zynqmp-gqspi.c
@@ -0,0 +1,1122 @@
+/*
+ * Xilinx Zynq UltraScale+ MPSoC Quad-SPI (QSPI) controller driver
+ * (master mode only)
+ *
+ * Copyright (C) 2009 - 2015 Xilinx, 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; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/dmaengine.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/of_irq.h>
+#include <linux/of_address.h>
+#include <linux/platform_device.h>
+#include <linux/spi/spi.h>
+#include <linux/spinlock.h>
+#include <linux/workqueue.h>
+
+/* Generic QSPI register offsets */
+#define GQSPI_CONFIG_OFST		0x00000100
+#define GQSPI_ISR_OFST			0x00000104
+#define GQSPI_IDR_OFST			0x0000010C
+#define GQSPI_IER_OFST			0x00000108
+#define GQSPI_IMASK_OFST		0x00000110
+#define GQSPI_EN_OFST			0x00000114
+#define GQSPI_TXD_OFST			0x0000011C
+#define GQSPI_RXD_OFST			0x00000120
+#define GQSPI_TX_THRESHOLD_OFST		0x00000128
+#define GQSPI_RX_THRESHOLD_OFST		0x0000012C
+#define GQSPI_LPBK_DLY_ADJ_OFST		0x00000138
+#define GQSPI_GEN_FIFO_OFST		0x00000140
+#define GQSPI_SEL_OFST			0x00000144
+#define GQSPI_GF_THRESHOLD_OFST		0x00000150
+#define GQSPI_FIFO_CTRL_OFST		0x0000014C
+#define GQSPI_QSPIDMA_DST_CTRL_OFST	0x0000080C
+#define GQSPI_QSPIDMA_DST_SIZE_OFST	0x00000804
+#define GQSPI_QSPIDMA_DST_STS_OFST	0x00000808
+#define GQSPI_QSPIDMA_DST_I_STS_OFST	0x00000814
+#define GQSPI_QSPIDMA_DST_I_EN_OFST	0x00000818
+#define GQSPI_QSPIDMA_DST_I_DIS_OFST	0x0000081C
+#define GQSPI_QSPIDMA_DST_I_MASK_OFST	0x00000820
+#define GQSPI_QSPIDMA_DST_ADDR_OFST	0x00000800
+#define GQSPI_QSPIDMA_DST_ADDR_MSB_OFST 0x00000828
+
+/* GQSPI register bit masks */
+#define GQSPI_SEL_MASK				0x00000001
+#define GQSPI_EN_MASK				0x00000001
+#define GQSPI_LPBK_DLY_ADJ_USE_LPBK_MASK	0x00000020
+#define GQSPI_ISR_WR_TO_CLR_MASK		0x00000002
+#define GQSPI_IDR_ALL_MASK			0x00000FBE
+#define GQSPI_CFG_MODE_EN_MASK			0xC0000000
+#define GQSPI_CFG_GEN_FIFO_START_MODE_MASK	0x20000000
+#define GQSPI_CFG_ENDIAN_MASK			0x04000000
+#define GQSPI_CFG_EN_POLL_TO_MASK		0x00100000
+#define GQSPI_CFG_WP_HOLD_MASK			0x00080000
+#define GQSPI_CFG_BAUD_RATE_DIV_MASK		0x00000038
+#define GQSPI_CFG_CLK_PHA_MASK			0x00000004
+#define GQSPI_CFG_CLK_POL_MASK			0x00000002
+#define GQSPI_CFG_START_GEN_FIFO_MASK		0x10000000
+#define GQSPI_GENFIFO_IMM_DATA_MASK		0x000000FF
+#define GQSPI_GENFIFO_DATA_XFER			0x00000100
+#define GQSPI_GENFIFO_EXP			0x00000200
+#define GQSPI_GENFIFO_MODE_SPI			0x00000400
+#define GQSPI_GENFIFO_MODE_DUALSPI		0x00000800
+#define GQSPI_GENFIFO_MODE_QUADSPI		0x00000C00
+#define GQSPI_GENFIFO_MODE_MASK			0x00000C00
+#define GQSPI_GENFIFO_CS_LOWER			0x00001000
+#define GQSPI_GENFIFO_CS_UPPER			0x00002000
+#define GQSPI_GENFIFO_BUS_LOWER			0x00004000
+#define GQSPI_GENFIFO_BUS_UPPER			0x00008000
+#define GQSPI_GENFIFO_BUS_BOTH			0x0000C000
+#define GQSPI_GENFIFO_BUS_MASK			0x0000C000
+#define GQSPI_GENFIFO_TX			0x00010000
+#define GQSPI_GENFIFO_RX			0x00020000
+#define GQSPI_GENFIFO_STRIPE			0x00040000
+#define GQSPI_GENFIFO_POLL			0x00080000
+#define GQSPI_GENFIFO_EXP_START			0x00000100
+#define GQSPI_FIFO_CTRL_RST_RX_FIFO_MASK	0x00000004
+#define GQSPI_FIFO_CTRL_RST_TX_FIFO_MASK	0x00000002
+#define GQSPI_FIFO_CTRL_RST_GEN_FIFO_MASK	0x00000001
+#define GQSPI_ISR_RXEMPTY_MASK			0x00000800
+#define GQSPI_ISR_GENFIFOFULL_MASK		0x00000400
+#define GQSPI_ISR_GENFIFONOT_FULL_MASK		0x00000200
+#define GQSPI_ISR_TXEMPTY_MASK			0x00000100
+#define GQSPI_ISR_GENFIFOEMPTY_MASK		0x00000080
+#define GQSPI_ISR_RXFULL_MASK			0x00000020
+#define GQSPI_ISR_RXNEMPTY_MASK			0x00000010
+#define GQSPI_ISR_TXFULL_MASK			0x00000008
+#define GQSPI_ISR_TXNOT_FULL_MASK		0x00000004
+#define GQSPI_ISR_POLL_TIME_EXPIRE_MASK		0x00000002
+#define GQSPI_IER_TXNOT_FULL_MASK		0x00000004
+#define GQSPI_IER_RXEMPTY_MASK			0x00000800
+#define GQSPI_IER_POLL_TIME_EXPIRE_MASK		0x00000002
+#define GQSPI_IER_RXNEMPTY_MASK			0x00000010
+#define GQSPI_IER_GENFIFOEMPTY_MASK		0x00000080
+#define GQSPI_IER_TXEMPTY_MASK			0x00000100
+#define GQSPI_QSPIDMA_DST_INTR_ALL_MASK		0x000000FE
+#define GQSPI_QSPIDMA_DST_STS_WTC		0x0000E000
+#define GQSPI_CFG_MODE_EN_DMA_MASK		0x80000000
+#define GQSPI_ISR_IDR_MASK			0x00000994
+#define GQSPI_QSPIDMA_DST_I_EN_DONE_MASK	0x00000002
+#define GQSPI_QSPIDMA_DST_I_STS_DONE_MASK	0x00000002
+#define GQSPI_IRQ_MASK				0x00000980
+
+#define GQSPI_CFG_BAUD_RATE_DIV_SHIFT		3
+#define GQSPI_GENFIFO_CS_SETUP			0x4
+#define GQSPI_GENFIFO_CS_HOLD			0x3
+#define GQSPI_TXD_DEPTH				64
+#define GQSPI_RX_FIFO_THRESHOLD			32
+#define GQSPI_RX_FIFO_FILL	(GQSPI_RX_FIFO_THRESHOLD * 4)
+#define GQSPI_TX_FIFO_THRESHOLD_RESET_VAL	32
+#define GQSPI_TX_FIFO_FILL	(GQSPI_TXD_DEPTH -\
+				GQSPI_TX_FIFO_THRESHOLD_RESET_VAL)
+#define GQSPI_GEN_FIFO_THRESHOLD_RESET_VAL	0X10
+#define GQSPI_QSPIDMA_DST_CTRL_RESET_VAL	0x803FFA00
+#define GQSPI_SELECT_FLASH_CS_LOWER		0x1
+#define GQSPI_SELECT_FLASH_CS_UPPER		0x2
+#define GQSPI_SELECT_FLASH_CS_BOTH		0x3
+#define GQSPI_SELECT_FLASH_BUS_LOWER		0x1
+#define GQSPI_SELECT_FLASH_BUS_UPPER		0x2
+#define GQSPI_SELECT_FLASH_BUS_BOTH		0x3
+#define GQSPI_BAUD_DIV_MAX	7	/* Baud rate divisor maximum */
+#define GQSPI_BAUD_DIV_SHIFT	2	/* Baud rate divisor shift */
+#define GQSPI_SELECT_MODE_SPI		0x1
+#define GQSPI_SELECT_MODE_DUALSPI	0x2
+#define GQSPI_SELECT_MODE_QUADSPI	0x4
+#define GQSPI_DMA_UNALIGN		0x3
+#define GQSPI_DEFAULT_NUM_CS	1	/* Default number of chip selects */
+
+enum mode_type {GQSPI_MODE_IO, GQSPI_MODE_DMA};
+
+/**
+ * struct zynqmp_qspi - Defines qspi driver instance
+ * @regs:		Virtual address of the QSPI controller registers
+ * @refclk:		Pointer to the peripheral clock
+ * @pclk:		Pointer to the APB clock
+ * @irq:		IRQ number
+ * @dev:		Pointer to struct device
+ * @txbuf:		Pointer to the TX buffer
+ * @rxbuf:		Pointer to the RX buffer
+ * @bytes_to_transfer:	Number of bytes left to transfer
+ * @bytes_to_receive:	Number of bytes left to receive
+ * @genfifocs:		Used for chip select
+ * @genfifobus:		Used to select the upper or lower bus
+ * @dma_rx_bytes:	Remaining bytes to receive by DMA mode
+ * @dma_addr:		DMA address after mapping the kernel buffer
+ * @genfifoentry:	Used for storing the genfifoentry instruction.
+ * @mode:		Defines the mode in which QSPI is operating
+ */
+struct zynqmp_qspi {
+	void __iomem *regs;
+	struct clk *refclk;
+	struct clk *pclk;
+	int irq;
+	struct device *dev;
+	const void *txbuf;
+	void *rxbuf;
+	int bytes_to_transfer;
+	int bytes_to_receive;
+	u32 genfifocs;
+	u32 genfifobus;
+	u32 dma_rx_bytes;
+	dma_addr_t dma_addr;
+	u32 genfifoentry;
+	enum mode_type mode;
+};
+
+/**
+ * zynqmp_gqspi_read:	For GQSPI controller read operation
+ * @xqspi:	Pointer to the zynqmp_qspi structure
+ * @offset:	Offset from where to read
+ */
+static u32 zynqmp_gqspi_read(struct zynqmp_qspi *xqspi, u32 offset)
+{
+	return readl_relaxed(xqspi->regs + offset);
+}
+
+/**
+ * zynqmp_gqspi_write:	For GQSPI controller write operation
+ * @xqspi:	Pointer to the zynqmp_qspi structure
+ * @offset:	Offset where to write
+ * @val:	Value to be written
+ */
+static inline void zynqmp_gqspi_write(struct zynqmp_qspi *xqspi, u32 offset,
+				      u32 val)
+{
+	writel_relaxed(val, (xqspi->regs + offset));
+}
+
+/**
+ * zynqmp_gqspi_selectslave:	For selection of slave device
+ * @instanceptr:	Pointer to the zynqmp_qspi structure
+ * @flashcs:	For chip select
+ * @flashbus:	To check which bus is selected- upper or lower
+ */
+static void zynqmp_gqspi_selectslave(struct zynqmp_qspi *instanceptr,
+				     u8 slavecs, u8 slavebus)
+{
+	/*
+	 * Bus and CS lines selected here will be updated in the instance and
+	 * used for subsequent GENFIFO entries during transfer.
+	 */
+
+	/* Choose slave select line */
+	switch (slavecs) {
+	case GQSPI_SELECT_FLASH_CS_BOTH:
+		instanceptr->genfifocs = GQSPI_GENFIFO_CS_LOWER |
+			GQSPI_GENFIFO_CS_UPPER;
+	case GQSPI_SELECT_FLASH_CS_UPPER:
+		instanceptr->genfifocs = GQSPI_GENFIFO_CS_UPPER;
+		break;
+	case GQSPI_SELECT_FLASH_CS_LOWER:
+		instanceptr->genfifocs = GQSPI_GENFIFO_CS_LOWER;
+		break;
+	default:
+		dev_warn(instanceptr->dev, "Invalid slave select\n");
+	}
+
+	/* Choose the bus */
+	switch (slavebus) {
+	case GQSPI_SELECT_FLASH_BUS_BOTH:
+		instanceptr->genfifobus = GQSPI_GENFIFO_BUS_LOWER |
+			GQSPI_GENFIFO_BUS_UPPER;
+		break;
+	case GQSPI_SELECT_FLASH_BUS_UPPER:
+		instanceptr->genfifobus = GQSPI_GENFIFO_BUS_UPPER;
+		break;
+	case GQSPI_SELECT_FLASH_BUS_LOWER:
+		instanceptr->genfifobus = GQSPI_GENFIFO_BUS_LOWER;
+		break;
+	default:
+		dev_warn(instanceptr->dev, "Invalid slave bus\n");
+	}
+}
+
+/**
+ * zynqmp_qspi_init_hw:	Initialize the hardware
+ * @xqspi:	Pointer to the zynqmp_qspi structure
+ *
+ * The default settings of the QSPI controller's configurable parameters on
+ * reset are
+ *	- Master mode
+ *	- TX threshold set to 1
+ *	- RX threshold set to 1
+ *	- Flash memory interface mode enabled
+ * This function performs the following actions
+ *	- Disable and clear all the interrupts
+ *	- Enable manual slave select
+ *	- Enable manual start
+ *	- Deselect all the chip select lines
+ *	- Set the little endian mode of TX FIFO and
+ *	- Enable the QSPI controller
+ */
+static void zynqmp_qspi_init_hw(struct zynqmp_qspi *xqspi)
+{
+	u32 config_reg;
+
+	/* Select the GQSPI mode */
+	zynqmp_gqspi_write(xqspi, GQSPI_SEL_OFST, GQSPI_SEL_MASK);
+	/* Clear and disable interrupts */
+	zynqmp_gqspi_write(xqspi, GQSPI_ISR_OFST,
+			   zynqmp_gqspi_read(xqspi, GQSPI_ISR_OFST) |
+			   GQSPI_ISR_WR_TO_CLR_MASK);
+	/* Clear the DMA STS */
+	zynqmp_gqspi_write(xqspi, GQSPI_QSPIDMA_DST_I_STS_OFST,
+			   zynqmp_gqspi_read(xqspi,
+					     GQSPI_QSPIDMA_DST_I_STS_OFST));
+	zynqmp_gqspi_write(xqspi, GQSPI_QSPIDMA_DST_STS_OFST,
+			   zynqmp_gqspi_read(xqspi,
+					     GQSPI_QSPIDMA_DST_STS_OFST) |
+					     GQSPI_QSPIDMA_DST_STS_WTC);
+	zynqmp_gqspi_write(xqspi, GQSPI_IDR_OFST, GQSPI_IDR_ALL_MASK);
+	zynqmp_gqspi_write(xqspi,
+			   GQSPI_QSPIDMA_DST_I_DIS_OFST,
+			   GQSPI_QSPIDMA_DST_INTR_ALL_MASK);
+	/* Disable the GQSPI */
+	zynqmp_gqspi_write(xqspi, GQSPI_EN_OFST, 0x0);
+	config_reg = zynqmp_gqspi_read(xqspi, GQSPI_CONFIG_OFST);
+	config_reg &= ~GQSPI_CFG_MODE_EN_MASK;
+	/* Manual start */
+	config_reg |= GQSPI_CFG_GEN_FIFO_START_MODE_MASK;
+	/* Little endian by default */
+	config_reg &= ~GQSPI_CFG_ENDIAN_MASK;
+	/* Disable poll time out */
+	config_reg &= ~GQSPI_CFG_EN_POLL_TO_MASK;
+	/* Set hold bit */
+	config_reg |= GQSPI_CFG_WP_HOLD_MASK;
+	/* Clear pre-scalar by default */
+	config_reg &= ~GQSPI_CFG_BAUD_RATE_DIV_MASK;
+	/* CPHA 0 */
+	config_reg &= ~GQSPI_CFG_CLK_PHA_MASK;
+	/* CPOL 0 */
+	config_reg &= ~GQSPI_CFG_CLK_POL_MASK;
+	zynqmp_gqspi_write(xqspi, GQSPI_CONFIG_OFST, config_reg);
+
+	/* Clear the TX and RX FIFO */
+	zynqmp_gqspi_write(xqspi, GQSPI_FIFO_CTRL_OFST,
+			   GQSPI_FIFO_CTRL_RST_RX_FIFO_MASK |
+			   GQSPI_FIFO_CTRL_RST_TX_FIFO_MASK |
+			   GQSPI_FIFO_CTRL_RST_GEN_FIFO_MASK);
+	/* Set by default to allow for high frequencies */
+	zynqmp_gqspi_write(xqspi, GQSPI_LPBK_DLY_ADJ_OFST,
+			   zynqmp_gqspi_read(xqspi, GQSPI_LPBK_DLY_ADJ_OFST) |
+			   GQSPI_LPBK_DLY_ADJ_USE_LPBK_MASK);
+	/* Reset thresholds */
+	zynqmp_gqspi_write(xqspi, GQSPI_TX_THRESHOLD_OFST,
+			   GQSPI_TX_FIFO_THRESHOLD_RESET_VAL);
+	zynqmp_gqspi_write(xqspi, GQSPI_RX_THRESHOLD_OFST,
+			   GQSPI_RX_FIFO_THRESHOLD);
+	zynqmp_gqspi_write(xqspi, GQSPI_GF_THRESHOLD_OFST,
+			   GQSPI_GEN_FIFO_THRESHOLD_RESET_VAL);
+	zynqmp_gqspi_selectslave(xqspi,
+				 GQSPI_SELECT_FLASH_CS_LOWER,
+				 GQSPI_SELECT_FLASH_BUS_LOWER);
+	/* Initialize DMA */
+	zynqmp_gqspi_write(xqspi,
+			GQSPI_QSPIDMA_DST_CTRL_OFST,
+			GQSPI_QSPIDMA_DST_CTRL_RESET_VAL);
+
+	/* Enable the GQSPI */
+	zynqmp_gqspi_write(xqspi, GQSPI_EN_OFST, GQSPI_EN_MASK);
+}
+
+/**
+ * zynqmp_qspi_copy_read_data:	Copy data to RX buffer
+ * @xqspi:	Pointer to the zynqmp_qspi structure
+ * @data:	The variable where data is stored
+ * @size:	Number of bytes to be copied from data to RX buffer
+ */
+static void zynqmp_qspi_copy_read_data(struct zynqmp_qspi *xqspi,
+				       ulong data, u8 size)
+{
+	memcpy(xqspi->rxbuf, &data, size);
+	xqspi->rxbuf += size;
+	xqspi->bytes_to_receive -= size;
+}
+
+/**
+ * zynqmp_prepare_transfer_hardware:	Prepares hardware for transfer.
+ * @master:	Pointer to the spi_master structure which provides
+ *		information about the controller.
+ *
+ * This function enables SPI master controller.
+ *
+ * Return:	0 on success; error value otherwise
+ */
+static int zynqmp_prepare_transfer_hardware(struct spi_master *master)
+{
+	struct zynqmp_qspi *xqspi = spi_master_get_devdata(master);
+	int ret;
+
+	ret = clk_enable(xqspi->refclk);
+	if (ret)
+		goto clk_err;
+
+	ret = clk_enable(xqspi->pclk);
+	if (ret)
+		goto clk_err;
+
+	zynqmp_gqspi_write(xqspi, GQSPI_EN_OFST, GQSPI_EN_MASK);
+	return 0;
+clk_err:
+	return ret;
+}
+
+/**
+ * zynqmp_unprepare_transfer_hardware:	Relaxes hardware after transfer
+ * @master:	Pointer to the spi_master structure which provides
+ *		information about the controller.
+ *
+ * This function disables the SPI master controller.
+ *
+ * Return:	Always 0
+ */
+static int zynqmp_unprepare_transfer_hardware(struct spi_master *master)
+{
+	struct zynqmp_qspi *xqspi = spi_master_get_devdata(master);
+
+	zynqmp_gqspi_write(xqspi, GQSPI_EN_OFST, 0x0);
+	clk_disable(xqspi->refclk);
+	clk_disable(xqspi->pclk);
+	return 0;
+}
+
+/**
+ * zynqmp_qspi_chipselect:	Select or deselect the chip select line
+ * @qspi:	Pointer to the spi_device structure
+ * @is_high:	Select(0) or deselect (1) the chip select line
+ */
+static void zynqmp_qspi_chipselect(struct spi_device *qspi, bool is_high)
+{
+	struct zynqmp_qspi *xqspi = spi_master_get_devdata(qspi->master);
+	ulong timeout;
+	u32 genfifoentry = 0x0, statusreg;
+
+	genfifoentry |= GQSPI_GENFIFO_MODE_SPI;
+	genfifoentry |= xqspi->genfifobus;
+
+	if (!is_high) {
+		genfifoentry |= xqspi->genfifocs;
+		genfifoentry |= GQSPI_GENFIFO_CS_SETUP;
+	} else {
+		genfifoentry |= GQSPI_GENFIFO_CS_HOLD;
+	}
+
+	zynqmp_gqspi_write(xqspi, GQSPI_GEN_FIFO_OFST, genfifoentry);
+
+	/* Dummy generic FIFO entry */
+	zynqmp_gqspi_write(xqspi, GQSPI_GEN_FIFO_OFST, 0x0);
+
+	/* Manually start the generic FIFO command */
+	zynqmp_gqspi_write(xqspi, GQSPI_CONFIG_OFST,
+			zynqmp_gqspi_read(xqspi, GQSPI_CONFIG_OFST) |
+			GQSPI_CFG_START_GEN_FIFO_MASK);
+
+	timeout = jiffies + msecs_to_jiffies(1000);
+
+	/* Wait until the generic FIFO command is empty */
+	do {
+		statusreg = zynqmp_gqspi_read(xqspi, GQSPI_ISR_OFST);
+
+		if ((statusreg & GQSPI_ISR_GENFIFOEMPTY_MASK) &&
+			(statusreg & GQSPI_ISR_TXEMPTY_MASK))
+			break;
+		else
+			cpu_relax();
+	} while (!time_after_eq(jiffies, timeout));
+
+	if (time_after_eq(jiffies, timeout))
+		dev_err(xqspi->dev, "Chip select timed out\n");
+}
+
+/**
+ * zynqmp_qspi_setup_transfer:	Configure QSPI controller for specified
+ *				transfer
+ * @qspi:	Pointer to the spi_device structure
+ * @transfer:	Pointer to the spi_transfer structure which provides
+ *		information about next transfer setup parameters
+ *
+ * Sets the operational mode of QSPI controller for the next QSPI transfer and
+ * sets the requested clock frequency.
+ *
+ * Return:	Always 0
+ *
+ * Note:
+ *	If the requested frequency is not an exact match with what can be
+ *	obtained using the pre-scalar value, the driver sets the clock
+ *	frequency which is lower than the requested frequency (maximum lower)
+ *	for the transfer.
+ *
+ *	If the requested frequency is higher or lower than that is supported
+ *	by the QSPI controller the driver will set the highest or lowest
+ *	frequency supported by controller.
+ */
+static int zynqmp_qspi_setup_transfer(struct spi_device *qspi,
+				      struct spi_transfer *transfer)
+{
+	struct zynqmp_qspi *xqspi = spi_master_get_devdata(qspi->master);
+	ulong clk_rate;
+	u32 config_reg, req_hz, baud_rate_val = 0;
+
+	if (transfer)
+		req_hz = transfer->speed_hz;
+	else
+		req_hz = qspi->max_speed_hz;
+
+	/* Set the clock frequency */
+	/* If req_hz == 0, default to lowest speed */
+	clk_rate = clk_get_rate(xqspi->refclk);
+
+	while ((baud_rate_val < GQSPI_BAUD_DIV_MAX) &&
+	       (clk_rate /
+		(GQSPI_BAUD_DIV_SHIFT << baud_rate_val)) > req_hz)
+		baud_rate_val++;
+
+	config_reg = zynqmp_gqspi_read(xqspi, GQSPI_CONFIG_OFST);
+
+	/* Set the QSPI clock phase and clock polarity */
+	config_reg &= (~GQSPI_CFG_CLK_PHA_MASK) & (~GQSPI_CFG_CLK_POL_MASK);
+
+	if (qspi->mode & SPI_CPHA)
+		config_reg |= GQSPI_CFG_CLK_PHA_MASK;
+	if (qspi->mode & SPI_CPOL)
+		config_reg |= GQSPI_CFG_CLK_POL_MASK;
+
+	config_reg &= ~GQSPI_CFG_BAUD_RATE_DIV_MASK;
+	config_reg |= (baud_rate_val << GQSPI_CFG_BAUD_RATE_DIV_SHIFT);
+	zynqmp_gqspi_write(xqspi, GQSPI_CONFIG_OFST, config_reg);
+	return 0;
+}
+
+/**
+ * zynqmp_qspi_setup:	Configure the QSPI controller
+ * @qspi:	Pointer to the spi_device structure
+ *
+ * Sets the operational mode of QSPI controller for the next QSPI transfer,
+ * baud rate and divisor value to setup the requested qspi clock.
+ *
+ * Return:	0 on success; error value otherwise.
+ */
+static int zynqmp_qspi_setup(struct spi_device *qspi)
+{
+	if (qspi->master->busy)
+		return -EBUSY;
+	return 0;
+}
+
+/**
+ * zynqmp_qspi_filltxfifo:	Fills the TX FIFO as long as there is room in
+ *				the FIFO or the bytes required to be
+ *				transmitted.
+ * @xqspi:	Pointer to the zynqmp_qspi structure
+ * @size:	Number of bytes to be copied from TX buffer to TX FIFO
+ */
+static void zynqmp_qspi_filltxfifo(struct zynqmp_qspi *xqspi, int size)
+{
+	u32 count = 0, intermediate;
+
+	while ((xqspi->bytes_to_transfer > 0) && (count < size)) {
+		memcpy(&intermediate, xqspi->txbuf, 4);
+		zynqmp_gqspi_write(xqspi, GQSPI_TXD_OFST, intermediate);
+
+		if (xqspi->bytes_to_transfer >= 4) {
+			xqspi->txbuf += 4;
+			xqspi->bytes_to_transfer -= 4;
+		} else {
+			xqspi->txbuf += xqspi->bytes_to_transfer;
+			xqspi->bytes_to_transfer = 0;
+		}
+		count++;
+	}
+}
+
+/**
+ * zynqmp_qspi_readrxfifo:	Fills the RX FIFO as long as there is room in
+ *				the FIFO.
+ * @xqspi:	Pointer to the zynqmp_qspi structure
+ * @size:	Number of bytes to be copied from RX buffer to RX FIFO
+ */
+static void zynqmp_qspi_readrxfifo(struct zynqmp_qspi *xqspi, u32 size)
+{
+	ulong data;
+	int count = 0;
+
+	while ((count < size) && (xqspi->bytes_to_receive > 0)) {
+		if (xqspi->bytes_to_receive >= 4) {
+			(*(u32 *) xqspi->rxbuf) =
+			zynqmp_gqspi_read(xqspi, GQSPI_RXD_OFST);
+			xqspi->rxbuf += 4;
+			xqspi->bytes_to_receive -= 4;
+			count += 4;
+		} else {
+			data = zynqmp_gqspi_read(xqspi, GQSPI_RXD_OFST);
+			count += xqspi->bytes_to_receive;
+			zynqmp_qspi_copy_read_data(xqspi, data,
+						   xqspi->bytes_to_receive);
+			xqspi->bytes_to_receive = 0;
+		}
+	}
+}
+
+/**
+ * zynqmp_process_dma_irq:	Handler for DMA done interrupt of QSPI
+ *				controller
+ * @xqspi:	zynqmp_qspi instance pointer
+ *
+ * This function handles DMA interrupt only.
+ */
+static void zynqmp_process_dma_irq(struct zynqmp_qspi *xqspi)
+{
+	u32 config_reg, genfifoentry;
+
+	dma_unmap_single(xqspi->dev, xqspi->dma_addr,
+				xqspi->dma_rx_bytes, DMA_FROM_DEVICE);
+	xqspi->rxbuf += xqspi->dma_rx_bytes;
+	xqspi->bytes_to_receive -= xqspi->dma_rx_bytes;
+	xqspi->dma_rx_bytes = 0;
+
+	/* Disabling the DMA interrupts */
+	zynqmp_gqspi_write(xqspi, GQSPI_QSPIDMA_DST_I_DIS_OFST,
+					GQSPI_QSPIDMA_DST_I_EN_DONE_MASK);
+
+	if (xqspi->bytes_to_receive > 0) {
+		/* Switch to IO mode,for remaining bytes to receive */
+		config_reg = zynqmp_gqspi_read(xqspi, GQSPI_CONFIG_OFST);
+		config_reg &= ~GQSPI_CFG_MODE_EN_MASK;
+		zynqmp_gqspi_write(xqspi, GQSPI_CONFIG_OFST, config_reg);
+
+		/* Initiate the transfer of remaining bytes */
+		genfifoentry = xqspi->genfifoentry;
+		genfifoentry |= xqspi->bytes_to_receive;
+		zynqmp_gqspi_write(xqspi, GQSPI_GEN_FIFO_OFST, genfifoentry);
+
+		/* Dummy generic FIFO entry */
+		zynqmp_gqspi_write(xqspi, GQSPI_GEN_FIFO_OFST, 0x0);
+
+		/* Manual start */
+		zynqmp_gqspi_write(xqspi, GQSPI_CONFIG_OFST,
+			(zynqmp_gqspi_read(xqspi, GQSPI_CONFIG_OFST) |
+			GQSPI_CFG_START_GEN_FIFO_MASK));
+
+		/* Enable the RX interrupts for IO mode */
+		zynqmp_gqspi_write(xqspi, GQSPI_IER_OFST,
+				GQSPI_IER_GENFIFOEMPTY_MASK |
+				GQSPI_IER_RXNEMPTY_MASK |
+				GQSPI_IER_RXEMPTY_MASK);
+	}
+}
+
+/**
+ * zynqmp_qspi_irq:	Interrupt service routine of the QSPI controller
+ * @irq:	IRQ number
+ * @dev_id:	Pointer to the xqspi structure
+ *
+ * This function handles TX empty only.
+ * On TX empty interrupt this function reads the received data from RX FIFO
+ * and fills the TX FIFO if there is any data remaining to be transferred.
+ *
+ * Return:	IRQ_HANDLED when interrupt is handled
+ *		IRQ_NONE otherwise.
+ */
+static irqreturn_t zynqmp_qspi_irq(int irq, void *dev_id)
+{
+	struct spi_master *master = dev_id;
+	struct zynqmp_qspi *xqspi = spi_master_get_devdata(master);
+	int ret = IRQ_NONE;
+	u32 status, mask, dma_status = 0;
+
+	status = zynqmp_gqspi_read(xqspi, GQSPI_ISR_OFST);
+	zynqmp_gqspi_write(xqspi, GQSPI_ISR_OFST, status);
+	mask = (status & ~(zynqmp_gqspi_read(xqspi, GQSPI_IMASK_OFST)));
+
+	/* Read and clear DMA status */
+	if (xqspi->mode == GQSPI_MODE_DMA) {
+		dma_status =
+			zynqmp_gqspi_read(xqspi, GQSPI_QSPIDMA_DST_I_STS_OFST);
+		zynqmp_gqspi_write(xqspi, GQSPI_QSPIDMA_DST_I_STS_OFST,
+								dma_status);
+	}
+
+	if (mask & GQSPI_ISR_TXNOT_FULL_MASK) {
+		zynqmp_qspi_filltxfifo(xqspi, GQSPI_TX_FIFO_FILL);
+		ret = IRQ_HANDLED;
+	}
+
+	if (dma_status & GQSPI_QSPIDMA_DST_I_STS_DONE_MASK) {
+		zynqmp_process_dma_irq(xqspi);
+		ret = IRQ_HANDLED;
+	} else if (!(mask & GQSPI_IER_RXEMPTY_MASK) &&
+			(mask & GQSPI_IER_GENFIFOEMPTY_MASK)) {
+		zynqmp_qspi_readrxfifo(xqspi, GQSPI_RX_FIFO_FILL);
+		ret = IRQ_HANDLED;
+	}
+
+	if ((xqspi->bytes_to_receive == 0) && (xqspi->bytes_to_transfer == 0)
+			&& ((status & GQSPI_IRQ_MASK) == GQSPI_IRQ_MASK)) {
+		zynqmp_gqspi_write(xqspi, GQSPI_IDR_OFST, GQSPI_ISR_IDR_MASK);
+		spi_finalize_current_transfer(master);
+		ret = IRQ_HANDLED;
+	}
+	return ret;
+}
+
+/**
+ * zynqmp_qspi_selectspimode:	Selects SPI mode - x1 or x2 or x4.
+ * @xqspi:	xqspi is a pointer to the GQSPI instance
+ * @spimode:	spimode - SPI or DUAL or QUAD.
+ * Return:	Mask to set desired SPI mode in GENFIFO entry.
+ */
+static inline u32 zynqmp_qspi_selectspimode(struct zynqmp_qspi *xqspi,
+						u8 spimode)
+{
+	u32 mask = 0;
+
+	switch (spimode) {
+	case GQSPI_SELECT_MODE_DUALSPI:
+		mask = GQSPI_GENFIFO_MODE_DUALSPI;
+		break;
+	case GQSPI_SELECT_MODE_QUADSPI:
+		mask = GQSPI_GENFIFO_MODE_QUADSPI;
+		break;
+	case GQSPI_SELECT_MODE_SPI:
+		mask = GQSPI_GENFIFO_MODE_SPI;
+		break;
+	default:
+		dev_warn(xqspi->dev, "Invalid SPI mode\n");
+	}
+
+	return mask;
+}
+
+/**
+ * zynq_qspi_setuprxdma:	This function sets up the RX DMA operation
+ * @xqspi:	xqspi is a pointer to the GQSPI instance.
+ */
+static void zynq_qspi_setuprxdma(struct zynqmp_qspi *xqspi)
+{
+	u32 rx_bytes, rx_rem, config_reg;
+	dma_addr_t addr;
+	u64 dma_align =  (u64)(uintptr_t)xqspi->rxbuf;
+
+	if ((xqspi->bytes_to_receive < 8) ||
+		((dma_align & GQSPI_DMA_UNALIGN) != 0x0)) {
+		/* Setting to IO mode */
+		config_reg = zynqmp_gqspi_read(xqspi, GQSPI_CONFIG_OFST);
+		config_reg &= ~GQSPI_CFG_MODE_EN_MASK;
+		zynqmp_gqspi_write(xqspi, GQSPI_CONFIG_OFST, config_reg);
+		xqspi->mode = GQSPI_MODE_IO;
+		xqspi->dma_rx_bytes = 0;
+		return;
+	}
+
+	rx_rem = xqspi->bytes_to_receive % 4;
+	rx_bytes = (xqspi->bytes_to_receive - rx_rem);
+
+	addr = dma_map_single(xqspi->dev, (void *)xqspi->rxbuf,
+						rx_bytes, DMA_FROM_DEVICE);
+	if (dma_mapping_error(xqspi->dev, addr))
+		dev_err(xqspi->dev, "ERR:rxdma:memory not mapped\n");
+
+	xqspi->dma_rx_bytes = rx_bytes;
+	xqspi->dma_addr = addr;
+	zynqmp_gqspi_write(xqspi, GQSPI_QSPIDMA_DST_ADDR_OFST,
+				(u32)(addr & 0xffffffff));
+	addr = ((addr >> 16) >> 16);
+	zynqmp_gqspi_write(xqspi, GQSPI_QSPIDMA_DST_ADDR_MSB_OFST,
+				((u32)addr) & 0xfff);
+
+	/* Enabling the DMA mode */
+	config_reg = zynqmp_gqspi_read(xqspi, GQSPI_CONFIG_OFST);
+	config_reg &= ~GQSPI_CFG_MODE_EN_MASK;
+	config_reg |= GQSPI_CFG_MODE_EN_DMA_MASK;
+	zynqmp_gqspi_write(xqspi, GQSPI_CONFIG_OFST, config_reg);
+
+	/* Switch to DMA mode */
+	xqspi->mode = GQSPI_MODE_DMA;
+
+	/* Write the number of bytes to transfer */
+	zynqmp_gqspi_write(xqspi, GQSPI_QSPIDMA_DST_SIZE_OFST, rx_bytes);
+}
+
+/**
+ * zynqmp_qspi_txrxsetup:	This function checks the TX/RX buffers in
+ *				the transfer and sets up the GENFIFO entries,
+ *				TX FIFO as required.
+ * @xqspi:	xqspi is a pointer to the GQSPI instance.
+ * @transfer:	It is a pointer to the structure containing transfer data.
+ * @genfifoentry:	genfifoentry is pointer to the variable in which
+ *			GENFIFO	mask is returned to calling function
+ */
+static void zynqmp_qspi_txrxsetup(struct zynqmp_qspi *xqspi,
+				  struct spi_transfer *transfer,
+				  u32 *genfifoentry)
+{
+	u32 config_reg;
+
+	/* Transmit */
+	if ((xqspi->txbuf != NULL) && (xqspi->rxbuf == NULL)) {
+		/* Setup data to be TXed */
+		*genfifoentry &= ~GQSPI_GENFIFO_RX;
+		*genfifoentry |= GQSPI_GENFIFO_DATA_XFER;
+		*genfifoentry |= GQSPI_GENFIFO_TX;
+		*genfifoentry |=
+			zynqmp_qspi_selectspimode(xqspi, transfer->tx_nbits);
+		xqspi->bytes_to_transfer = transfer->len;
+		if (xqspi->mode == GQSPI_MODE_DMA) {
+			config_reg = zynqmp_gqspi_read(xqspi,
+							GQSPI_CONFIG_OFST);
+			config_reg &= ~GQSPI_CFG_MODE_EN_MASK;
+			zynqmp_gqspi_write(xqspi, GQSPI_CONFIG_OFST,
+								config_reg);
+			xqspi->mode = GQSPI_MODE_IO;
+		}
+		zynqmp_qspi_filltxfifo(xqspi, GQSPI_TXD_DEPTH);
+		/* Discard RX data */
+		xqspi->bytes_to_receive = 0;
+	} else if ((xqspi->txbuf == NULL) && (xqspi->rxbuf != NULL)) {
+		/* Receive */
+
+		/* TX auto fill */
+		*genfifoentry &= ~GQSPI_GENFIFO_TX;
+		/* Setup RX */
+		*genfifoentry |= GQSPI_GENFIFO_DATA_XFER;
+		*genfifoentry |= GQSPI_GENFIFO_RX;
+		*genfifoentry |=
+			zynqmp_qspi_selectspimode(xqspi, transfer->rx_nbits);
+		xqspi->bytes_to_transfer = 0;
+		xqspi->bytes_to_receive = transfer->len;
+		zynq_qspi_setuprxdma(xqspi);
+	}
+}
+
+/**
+ * zynqmp_qspi_start_transfer:	Initiates the QSPI transfer
+ * @master:	Pointer to the spi_master structure which provides
+ *		information about the controller.
+ * @qspi:	Pointer to the spi_device structure
+ * @transfer:	Pointer to the spi_transfer structure which provide information
+ *		about next transfer parameters
+ *
+ * This function fills the TX FIFO, starts the QSPI transfer, and waits for the
+ * transfer to be completed.
+ *
+ * Return:	Number of bytes transferred in the last transfer
+ */
+static int zynqmp_qspi_start_transfer(struct spi_master *master,
+				      struct spi_device *qspi,
+				      struct spi_transfer *transfer)
+{
+	struct zynqmp_qspi *xqspi = spi_master_get_devdata(master);
+	u32 genfifoentry = 0x0, transfer_len;
+
+	xqspi->txbuf = transfer->tx_buf;
+	xqspi->rxbuf = transfer->rx_buf;
+
+	zynqmp_qspi_setup_transfer(qspi, transfer);
+
+	genfifoentry |= xqspi->genfifocs;
+	genfifoentry |= xqspi->genfifobus;
+
+	zynqmp_qspi_txrxsetup(xqspi, transfer, &genfifoentry);
+
+	if (xqspi->mode == GQSPI_MODE_DMA)
+		transfer_len = xqspi->dma_rx_bytes;
+	else
+		transfer_len = transfer->len;
+
+	xqspi->genfifoentry = genfifoentry;
+	if ((transfer_len) < GQSPI_GENFIFO_IMM_DATA_MASK) {
+		genfifoentry &= ~GQSPI_GENFIFO_IMM_DATA_MASK;
+		genfifoentry |= transfer_len;
+		zynqmp_gqspi_write(xqspi, GQSPI_GEN_FIFO_OFST, genfifoentry);
+	} else {
+		int tempcount = transfer_len;
+		u32 exponent = 8;	/* 2^8 = 256 */
+		u8 imm_data = tempcount & 0xFF;
+
+		tempcount &= ~(tempcount & 0xFF);
+		/* Immediate entry */
+		if (tempcount != 0) {
+			/* Exponent entries */
+			genfifoentry |= GQSPI_GENFIFO_EXP;
+			while (tempcount != 0) {
+				if (tempcount & GQSPI_GENFIFO_EXP_START) {
+					genfifoentry &=
+					    ~GQSPI_GENFIFO_IMM_DATA_MASK;
+					genfifoentry |= exponent;
+					zynqmp_gqspi_write(xqspi,
+							   GQSPI_GEN_FIFO_OFST,
+							   genfifoentry);
+				}
+				tempcount = tempcount >> 1;
+				exponent++;
+			}
+		}
+		if (imm_data != 0) {
+			genfifoentry &= ~GQSPI_GENFIFO_EXP;
+			genfifoentry &= ~GQSPI_GENFIFO_IMM_DATA_MASK;
+			genfifoentry |= (u8) (imm_data & 0xFF);
+			zynqmp_gqspi_write(xqspi,
+					   GQSPI_GEN_FIFO_OFST, genfifoentry);
+		}
+	}
+
+	if ((xqspi->mode == GQSPI_MODE_IO) &&
+			(xqspi->rxbuf != NULL)) {
+		/* Dummy generic FIFO entry */
+		zynqmp_gqspi_write(xqspi, GQSPI_GEN_FIFO_OFST, 0x0);
+	}
+
+	/* Since we are using manual mode */
+	zynqmp_gqspi_write(xqspi, GQSPI_CONFIG_OFST,
+			   zynqmp_gqspi_read(xqspi, GQSPI_CONFIG_OFST) |
+			   GQSPI_CFG_START_GEN_FIFO_MASK);
+
+	if (xqspi->txbuf != NULL)
+		/* Enable interrupts for TX */
+		zynqmp_gqspi_write(xqspi, GQSPI_IER_OFST,
+				   GQSPI_IER_TXEMPTY_MASK |
+					GQSPI_IER_GENFIFOEMPTY_MASK |
+					GQSPI_IER_TXNOT_FULL_MASK);
+
+	if (xqspi->rxbuf != NULL) {
+		/* Enable interrupts for RX */
+		if (xqspi->mode == GQSPI_MODE_DMA) {
+			/* Enable DMA interrupts */
+			zynqmp_gqspi_write(xqspi,
+					GQSPI_QSPIDMA_DST_I_EN_OFST,
+					GQSPI_QSPIDMA_DST_I_EN_DONE_MASK);
+		} else {
+			zynqmp_gqspi_write(xqspi, GQSPI_IER_OFST,
+					GQSPI_IER_GENFIFOEMPTY_MASK |
+					GQSPI_IER_RXNEMPTY_MASK |
+					GQSPI_IER_RXEMPTY_MASK);
+		}
+	}
+
+	return transfer->len;
+}
+
+/**
+ * zynqmp_qspi_suspend:	Suspend method for the QSPI driver
+ * @_dev:	Address of the platform_device structure
+ *
+ * This function stops the QSPI driver queue and disables the QSPI controller
+ *
+ * Return:	Always 0
+ */
+static int __maybe_unused zynqmp_qspi_suspend(struct device *dev)
+{
+	struct platform_device *pdev = container_of(dev,
+						    struct platform_device,
+						    dev);
+	struct spi_master *master = platform_get_drvdata(pdev);
+
+	spi_master_suspend(master);
+
+	zynqmp_unprepare_transfer_hardware(master);
+
+	return 0;
+}
+
+/**
+ * zynqmp_qspi_resume:	Resume method for the QSPI driver
+ * @dev:	Address of the platform_device structure
+ *
+ * The function starts the QSPI driver queue and initializes the QSPI
+ * controller
+ *
+ * Return:	0 on success; error value otherwise
+ */
+static int __maybe_unused zynqmp_qspi_resume(struct device *dev)
+{
+	struct platform_device *pdev = container_of(dev,
+						    struct platform_device,
+						    dev);
+	struct spi_master *master = platform_get_drvdata(pdev);
+	struct zynqmp_qspi *xqspi = spi_master_get_devdata(master);
+	int ret = 0;
+
+	ret = clk_enable(xqspi->pclk);
+	if (ret) {
+		dev_err(dev, "Cannot enable APB clock.\n");
+		return ret;
+	}
+
+	ret = clk_enable(xqspi->refclk);
+	if (ret) {
+		dev_err(dev, "Cannot enable device clock.\n");
+		clk_disable(xqspi->pclk);
+		return ret;
+	}
+
+	spi_master_resume(master);
+
+	return 0;
+}
+
+static SIMPLE_DEV_PM_OPS(zynqmp_qspi_dev_pm_ops, zynqmp_qspi_suspend,
+			 zynqmp_qspi_resume);
+
+/**
+ * zynqmp_qspi_probe:	Probe method for the QSPI driver
+ * @pdev:	Pointer to the platform_device structure
+ *
+ * This function initializes the driver data structures and the hardware.
+ *
+ * Return:	0 on success; error value otherwise
+ */
+static int zynqmp_qspi_probe(struct platform_device *pdev)
+{
+	int ret = 0;
+	struct spi_master *master;
+	struct zynqmp_qspi *xqspi;
+	struct resource *res;
+	struct device *dev = &pdev->dev;
+
+	master = spi_alloc_master(&pdev->dev, sizeof(*xqspi));
+	if (!master)
+		return -ENOMEM;
+
+	xqspi = spi_master_get_devdata(master);
+	master->dev.of_node = pdev->dev.of_node;
+	platform_set_drvdata(pdev, master);
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	xqspi->regs = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(xqspi->regs)) {
+		ret = PTR_ERR(xqspi->regs);
+		goto remove_master;
+	}
+
+	xqspi->dev = dev;
+	xqspi->pclk = devm_clk_get(&pdev->dev, "pclk");
+	if (IS_ERR(xqspi->pclk)) {
+		dev_err(dev, "pclk clock not found.\n");
+		ret = PTR_ERR(xqspi->pclk);
+		goto remove_master;
+	}
+
+	ret = clk_prepare_enable(xqspi->pclk);
+	if (ret) {
+		dev_err(dev, "Unable to enable APB clock.\n");
+		goto remove_master;
+	}
+
+	xqspi->refclk = devm_clk_get(&pdev->dev, "ref_clk");
+	if (IS_ERR(xqspi->refclk)) {
+		dev_err(dev, "ref_clk clock not found.\n");
+		ret = PTR_ERR(xqspi->refclk);
+		goto clk_dis_pclk;
+	}
+
+	ret = clk_prepare_enable(xqspi->refclk);
+	if (ret) {
+		dev_err(dev, "Unable to enable device clock.\n");
+		goto clk_dis_pclk;
+	}
+
+	/* QSPI controller initializations */
+	zynqmp_qspi_init_hw(xqspi);
+
+	xqspi->irq = platform_get_irq(pdev, 0);
+	if (xqspi->irq <= 0) {
+		ret = -ENXIO;
+		dev_err(dev, "irq resource not found\n");
+		goto clk_dis_all;
+	}
+	ret = devm_request_irq(&pdev->dev, xqspi->irq, zynqmp_qspi_irq,
+			       0, pdev->name, master);
+	if (ret != 0) {
+		ret = -ENXIO;
+		dev_err(dev, "request_irq failed\n");
+		goto clk_dis_all;
+	}
+
+	master->num_chipselect = GQSPI_DEFAULT_NUM_CS;
+
+	master->setup = zynqmp_qspi_setup;
+	master->set_cs = zynqmp_qspi_chipselect;
+	master->transfer_one = zynqmp_qspi_start_transfer;
+	master->prepare_transfer_hardware = zynqmp_prepare_transfer_hardware;
+	master->unprepare_transfer_hardware =
+					zynqmp_unprepare_transfer_hardware;
+	master->max_speed_hz = clk_get_rate(xqspi->refclk) / 2;
+	master->bits_per_word_mask = SPI_BPW_MASK(8);
+	master->mode_bits = SPI_CPOL | SPI_CPHA | SPI_RX_DUAL | SPI_RX_QUAD |
+			    SPI_TX_DUAL | SPI_TX_QUAD;
+
+	if (master->dev.parent == NULL)
+		master->dev.parent = &master->dev;
+
+	ret = spi_register_master(master);
+	if (ret)
+		goto clk_dis_all;
+
+	return 0;
+
+clk_dis_all:
+	clk_disable_unprepare(xqspi->refclk);
+clk_dis_pclk:
+	clk_disable_unprepare(xqspi->pclk);
+remove_master:
+	spi_master_put(master);
+
+	return ret;
+}
+
+/**
+ * zynqmp_qspi_remove:	Remove method for the QSPI driver
+ * @pdev:	Pointer to the platform_device structure
+ *
+ * This function is called if a device is physically removed from the system or
+ * if the driver module is being unloaded. It frees all resources allocated to
+ * the device.
+ *
+ * Return:	0 Always
+ */
+static int zynqmp_qspi_remove(struct platform_device *pdev)
+{
+	struct spi_master *master = platform_get_drvdata(pdev);
+	struct zynqmp_qspi *xqspi = spi_master_get_devdata(master);
+
+	zynqmp_gqspi_write(xqspi, GQSPI_EN_OFST, 0x0);
+	clk_disable_unprepare(xqspi->refclk);
+	clk_disable_unprepare(xqspi->pclk);
+
+	spi_unregister_master(master);
+
+	return 0;
+}
+
+static const struct of_device_id zynqmp_qspi_of_match[] = {
+	{ .compatible = "xlnx,zynqmp-qspi-1.0", },
+	{ /* End of table */ }
+};
+
+MODULE_DEVICE_TABLE(of, zynqmp_qspi_of_match);
+
+static struct platform_driver zynqmp_qspi_driver = {
+	.probe = zynqmp_qspi_probe,
+	.remove = zynqmp_qspi_remove,
+	.driver = {
+		.name = "zynqmp-qspi",
+		.of_match_table = zynqmp_qspi_of_match,
+		.pm = &zynqmp_qspi_dev_pm_ops,
+	},
+};
+
+module_platform_driver(zynqmp_qspi_driver);
+
+MODULE_AUTHOR("Xilinx, Inc.");
+MODULE_DESCRIPTION("Xilinx Zynqmp QSPI driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c
index 57a195041dc7..cf8b91b23a76 100644
--- a/drivers/spi/spi.c
+++ b/drivers/spi/spi.c
@@ -16,7 +16,6 @@
  */
 
 #include <linux/kernel.h>
-#include <linux/kmod.h>
 #include <linux/device.h>
 #include <linux/init.h>
 #include <linux/cache.h>
@@ -129,125 +128,11 @@ static int spi_uevent(struct device *dev, struct kobj_uevent_env *env)
 	return 0;
 }
 
-#ifdef CONFIG_PM_SLEEP
-static int spi_legacy_suspend(struct device *dev, pm_message_t message)
-{
-	int			value = 0;
-	struct spi_driver	*drv = to_spi_driver(dev->driver);
-
-	/* suspend will stop irqs and dma; no more i/o */
-	if (drv) {
-		if (drv->suspend)
-			value = drv->suspend(to_spi_device(dev), message);
-		else
-			dev_dbg(dev, "... can't suspend\n");
-	}
-	return value;
-}
-
-static int spi_legacy_resume(struct device *dev)
-{
-	int			value = 0;
-	struct spi_driver	*drv = to_spi_driver(dev->driver);
-
-	/* resume may restart the i/o queue */
-	if (drv) {
-		if (drv->resume)
-			value = drv->resume(to_spi_device(dev));
-		else
-			dev_dbg(dev, "... can't resume\n");
-	}
-	return value;
-}
-
-static int spi_pm_suspend(struct device *dev)
-{
-	const struct dev_pm_ops *pm = dev->driver ? dev->driver->pm : NULL;
-
-	if (pm)
-		return pm_generic_suspend(dev);
-	else
-		return spi_legacy_suspend(dev, PMSG_SUSPEND);
-}
-
-static int spi_pm_resume(struct device *dev)
-{
-	const struct dev_pm_ops *pm = dev->driver ? dev->driver->pm : NULL;
-
-	if (pm)
-		return pm_generic_resume(dev);
-	else
-		return spi_legacy_resume(dev);
-}
-
-static int spi_pm_freeze(struct device *dev)
-{
-	const struct dev_pm_ops *pm = dev->driver ? dev->driver->pm : NULL;
-
-	if (pm)
-		return pm_generic_freeze(dev);
-	else
-		return spi_legacy_suspend(dev, PMSG_FREEZE);
-}
-
-static int spi_pm_thaw(struct device *dev)
-{
-	const struct dev_pm_ops *pm = dev->driver ? dev->driver->pm : NULL;
-
-	if (pm)
-		return pm_generic_thaw(dev);
-	else
-		return spi_legacy_resume(dev);
-}
-
-static int spi_pm_poweroff(struct device *dev)
-{
-	const struct dev_pm_ops *pm = dev->driver ? dev->driver->pm : NULL;
-
-	if (pm)
-		return pm_generic_poweroff(dev);
-	else
-		return spi_legacy_suspend(dev, PMSG_HIBERNATE);
-}
-
-static int spi_pm_restore(struct device *dev)
-{
-	const struct dev_pm_ops *pm = dev->driver ? dev->driver->pm : NULL;
-
-	if (pm)
-		return pm_generic_restore(dev);
-	else
-		return spi_legacy_resume(dev);
-}
-#else
-#define spi_pm_suspend	NULL
-#define spi_pm_resume	NULL
-#define spi_pm_freeze	NULL
-#define spi_pm_thaw	NULL
-#define spi_pm_poweroff	NULL
-#define spi_pm_restore	NULL
-#endif
-
-static const struct dev_pm_ops spi_pm = {
-	.suspend = spi_pm_suspend,
-	.resume = spi_pm_resume,
-	.freeze = spi_pm_freeze,
-	.thaw = spi_pm_thaw,
-	.poweroff = spi_pm_poweroff,
-	.restore = spi_pm_restore,
-	SET_RUNTIME_PM_OPS(
-		pm_generic_runtime_suspend,
-		pm_generic_runtime_resume,
-		NULL
-	)
-};
-
 struct bus_type spi_bus_type = {
 	.name		= "spi",
 	.dev_groups	= spi_dev_groups,
 	.match		= spi_match_device,
 	.uevent		= spi_uevent,
-	.pm		= &spi_pm,
 };
 EXPORT_SYMBOL_GPL(spi_bus_type);
 
@@ -686,7 +571,7 @@ static int __spi_map_msg(struct spi_master *master, struct spi_message *msg)
 	return 0;
 }
 
-static int spi_unmap_msg(struct spi_master *master, struct spi_message *msg)
+static int __spi_unmap_msg(struct spi_master *master, struct spi_message *msg)
 {
 	struct spi_transfer *xfer;
 	struct device *tx_dev, *rx_dev;
@@ -714,13 +599,32 @@ static inline int __spi_map_msg(struct spi_master *master,
 	return 0;
 }
 
-static inline int spi_unmap_msg(struct spi_master *master,
-				struct spi_message *msg)
+static inline int __spi_unmap_msg(struct spi_master *master,
+				  struct spi_message *msg)
 {
 	return 0;
 }
 #endif /* !CONFIG_HAS_DMA */
 
+static inline int spi_unmap_msg(struct spi_master *master,
+				struct spi_message *msg)
+{
+	struct spi_transfer *xfer;
+
+	list_for_each_entry(xfer, &msg->transfers, transfer_list) {
+		/*
+		 * Restore the original value of tx_buf or rx_buf if they are
+		 * NULL.
+		 */
+		if (xfer->tx_buf == master->dummy_tx)
+			xfer->tx_buf = NULL;
+		if (xfer->rx_buf == master->dummy_rx)
+			xfer->rx_buf = NULL;
+	}
+
+	return __spi_unmap_msg(master, msg);
+}
+
 static int spi_map_msg(struct spi_master *master, struct spi_message *msg)
 {
 	struct spi_transfer *xfer;
@@ -851,6 +755,9 @@ out:
 	if (msg->status == -EINPROGRESS)
 		msg->status = ret;
 
+	if (msg->status && master->handle_err)
+		master->handle_err(master, msg);
+
 	spi_finalize_current_message(master);
 
 	return ret;
@@ -1091,9 +998,6 @@ void spi_finalize_current_message(struct spi_master *master)
 
 	spin_lock_irqsave(&master->queue_lock, flags);
 	mesg = master->cur_msg;
-	master->cur_msg = NULL;
-
-	queue_kthread_work(&master->kworker, &master->pump_messages);
 	spin_unlock_irqrestore(&master->queue_lock, flags);
 
 	spi_unmap_msg(master, mesg);
@@ -1106,9 +1010,13 @@ void spi_finalize_current_message(struct spi_master *master)
 		}
 	}
 
-	trace_spi_message_done(mesg);
-
+	spin_lock_irqsave(&master->queue_lock, flags);
+	master->cur_msg = NULL;
 	master->cur_msg_prepared = false;
+	queue_kthread_work(&master->kworker, &master->pump_messages);
+	spin_unlock_irqrestore(&master->queue_lock, flags);
+
+	trace_spi_message_done(mesg);
 
 	mesg->state = NULL;
 	if (mesg->complete)
@@ -1360,7 +1268,6 @@ of_register_spi_device(struct spi_master *master, struct device_node *nc)
 	spi->dev.of_node = nc;
 
 	/* Register the new device */
-	request_module("%s%s", SPI_MODULE_PREFIX, spi->modalias);
 	rc = spi_add_device(spi);
 	if (rc) {
 		dev_err(&master->dev, "spi_device register error %s\n",
@@ -1894,6 +1801,8 @@ int spi_setup(struct spi_device *spi)
 	if (!spi->max_speed_hz)
 		spi->max_speed_hz = spi->master->max_speed_hz;
 
+	spi_set_cs(spi, false);
+
 	if (spi->master->setup)
 		status = spi->master->setup(spi);
 
diff --git a/drivers/spi/spidev.c b/drivers/spi/spidev.c
index 4eb7a980e670..dd616ff0ffc5 100644
--- a/drivers/spi/spidev.c
+++ b/drivers/spi/spidev.c
@@ -95,37 +95,25 @@ MODULE_PARM_DESC(bufsiz, "data bytes in biggest supported SPI message");
 
 /*-------------------------------------------------------------------------*/
 
-/*
- * We can't use the standard synchronous wrappers for file I/O; we
- * need to protect against async removal of the underlying spi_device.
- */
-static void spidev_complete(void *arg)
-{
-	complete(arg);
-}
-
 static ssize_t
 spidev_sync(struct spidev_data *spidev, struct spi_message *message)
 {
 	DECLARE_COMPLETION_ONSTACK(done);
 	int status;
-
-	message->complete = spidev_complete;
-	message->context = &done;
+	struct spi_device *spi;
 
 	spin_lock_irq(&spidev->spi_lock);
-	if (spidev->spi == NULL)
+	spi = spidev->spi;
+	spin_unlock_irq(&spidev->spi_lock);
+
+	if (spi == NULL)
 		status = -ESHUTDOWN;
 	else
-		status = spi_async(spidev->spi, message);
-	spin_unlock_irq(&spidev->spi_lock);
+		status = spi_sync(spi, message);
+
+	if (status == 0)
+		status = message->actual_length;
 
-	if (status == 0) {
-		wait_for_completion(&done);
-		status = message->status;
-		if (status == 0)
-			status = message->actual_length;
-	}
 	return status;
 }
 
@@ -223,7 +211,7 @@ static int spidev_message(struct spidev_data *spidev,
 	struct spi_transfer	*k_xfers;
 	struct spi_transfer	*k_tmp;
 	struct spi_ioc_transfer *u_tmp;
-	unsigned		n, total;
+	unsigned		n, total, tx_total, rx_total;
 	u8			*tx_buf, *rx_buf;
 	int			status = -EFAULT;
 
@@ -239,33 +227,52 @@ static int spidev_message(struct spidev_data *spidev,
 	tx_buf = spidev->tx_buffer;
 	rx_buf = spidev->rx_buffer;
 	total = 0;
+	tx_total = 0;
+	rx_total = 0;
 	for (n = n_xfers, k_tmp = k_xfers, u_tmp = u_xfers;
 			n;
 			n--, k_tmp++, u_tmp++) {
 		k_tmp->len = u_tmp->len;
 
 		total += k_tmp->len;
-		if (total > bufsiz) {
+		/* Since the function returns the total length of transfers
+		 * on success, restrict the total to positive int values to
+		 * avoid the return value looking like an error.  Also check
+		 * each transfer length to avoid arithmetic overflow.
+		 */
+		if (total > INT_MAX || k_tmp->len > INT_MAX) {
 			status = -EMSGSIZE;
 			goto done;
 		}
 
 		if (u_tmp->rx_buf) {
+			/* this transfer needs space in RX bounce buffer */
+			rx_total += k_tmp->len;
+			if (rx_total > bufsiz) {
+				status = -EMSGSIZE;
+				goto done;
+			}
 			k_tmp->rx_buf = rx_buf;
 			if (!access_ok(VERIFY_WRITE, (u8 __user *)
 						(uintptr_t) u_tmp->rx_buf,
 						u_tmp->len))
 				goto done;
+			rx_buf += k_tmp->len;
 		}
 		if (u_tmp->tx_buf) {
+			/* this transfer needs space in TX bounce buffer */
+			tx_total += k_tmp->len;
+			if (tx_total > bufsiz) {
+				status = -EMSGSIZE;
+				goto done;
+			}
 			k_tmp->tx_buf = tx_buf;
 			if (copy_from_user(tx_buf, (const u8 __user *)
 						(uintptr_t) u_tmp->tx_buf,
 					u_tmp->len))
 				goto done;
+			tx_buf += k_tmp->len;
 		}
-		tx_buf += k_tmp->len;
-		rx_buf += k_tmp->len;
 
 		k_tmp->cs_change = !!u_tmp->cs_change;
 		k_tmp->tx_nbits = u_tmp->tx_nbits;
@@ -303,8 +310,8 @@ static int spidev_message(struct spidev_data *spidev,
 				status = -EFAULT;
 				goto done;
 			}
+			rx_buf += u_tmp->len;
 		}
-		rx_buf += u_tmp->len;
 	}
 	status = total;
 
@@ -628,7 +635,6 @@ err_find_dev:
 static int spidev_release(struct inode *inode, struct file *filp)
 {
 	struct spidev_data	*spidev;
-	int			status = 0;
 
 	mutex_lock(&device_list_lock);
 	spidev = filp->private_data;
@@ -657,7 +663,7 @@ static int spidev_release(struct inode *inode, struct file *filp)
 	}
 	mutex_unlock(&device_list_lock);
 
-	return status;
+	return 0;
 }
 
 static const struct file_operations spidev_fops = {
@@ -684,6 +690,14 @@ static const struct file_operations spidev_fops = {
 
 static struct class *spidev_class;
 
+#ifdef CONFIG_OF
+static const struct of_device_id spidev_dt_ids[] = {
+	{ .compatible = "rohm,dh2228fv" },
+	{},
+};
+MODULE_DEVICE_TABLE(of, spidev_dt_ids);
+#endif
+
 /*-------------------------------------------------------------------------*/
 
 static int spidev_probe(struct spi_device *spi)
@@ -692,6 +706,17 @@ static int spidev_probe(struct spi_device *spi)
 	int			status;
 	unsigned long		minor;
 
+	/*
+	 * spidev should never be referenced in DT without a specific
+	 * compatbile string, it is a Linux implementation thing
+	 * rather than a description of the hardware.
+	 */
+	if (spi->dev.of_node && !of_match_device(spidev_dt_ids, &spi->dev)) {
+		dev_err(&spi->dev, "buggy DT: spidev listed directly in DT\n");
+		WARN_ON(spi->dev.of_node &&
+			!of_match_device(spidev_dt_ids, &spi->dev));
+	}
+
 	/* Allocate driver data */
 	spidev = kzalloc(sizeof(*spidev), GFP_KERNEL);
 	if (!spidev)
@@ -758,13 +783,6 @@ static int spidev_remove(struct spi_device *spi)
 	return 0;
 }
 
-static const struct of_device_id spidev_dt_ids[] = {
-	{ .compatible = "rohm,dh2228fv" },
-	{},
-};
-
-MODULE_DEVICE_TABLE(of, spidev_dt_ids);
-
 static struct spi_driver spidev_spi_driver = {
 	.driver = {
 		.name =		"spidev",