Merge remote-tracking branch 'spi/for-next'

17 files changed, 808 insertions, 269 deletions

diff --git a/Documentation/devicetree/bindings/spi/spi-ath79.txt b/Documentation/devicetree/bindings/spi/spi-ath79.txt
new file mode 100644
index 000000000000..f1ad9c367532
--- /dev/null
+++ b/Documentation/devicetree/bindings/spi/spi-ath79.txt
@@ -0,0 +1,24 @@
+Binding for Qualcomm Atheros AR7xxx/AR9xxx SPI controller
+
+Required properties:
+- compatible: has to be "qca,<soc-type>-spi", "qca,ar7100-spi" as fallback.
+- reg: Base address and size of the controllers memory area
+- clocks: phandle to the AHB clock.
+- clock-names: has to be "ahb".
+- #address-cells: <1>, as required by generic SPI binding.
+- #size-cells: <0>, also as required by generic SPI binding.
+
+Child nodes as per the generic SPI binding.
+
+Example:
+
+	spi@1F000000 {
+		compatible = "qca,ar9132-spi", "qca,ar7100-spi";
+		reg = <0x1F000000 0x10>;
+
+		clocks = <&pll 2>;
+		clock-names = "ahb";
+
+		#address-cells = <1>;
+		#size-cells = <0>;
+	};
diff --git a/Documentation/devicetree/bindings/spi/spi_pl022.txt b/Documentation/devicetree/bindings/spi/spi_pl022.txt
index 22ed6797216d..4d1673ca8cf8 100644
--- a/Documentation/devicetree/bindings/spi/spi_pl022.txt
+++ b/Documentation/devicetree/bindings/spi/spi_pl022.txt
@@ -4,9 +4,9 @@ Required properties:
 - compatible : "arm,pl022", "arm,primecell"
 - reg : Offset and length of the register set for the device
 - interrupts : Should contain SPI controller interrupt
+- num-cs : total number of chipselects
 
 Optional properties:
-- num-cs : total number of chipselects
 - cs-gpios : should specify GPIOs used for chipselects.
   The gpios will be referred to as reg = <index> in the SPI child nodes.
   If unspecified, a single SPI device without a chip select can be used.
diff --git a/arch/mips/include/asm/mach-ath79/ath79_spi_platform.h b/arch/mips/include/asm/mach-ath79/ath79_spi_platform.h
index aa2283e602fc..aa71216edf99 100644
--- a/arch/mips/include/asm/mach-ath79/ath79_spi_platform.h
+++ b/arch/mips/include/asm/mach-ath79/ath79_spi_platform.h
@@ -16,8 +16,4 @@ struct ath79_spi_platform_data {
 	unsigned	num_chipselect;
 };
 
-struct ath79_spi_controller_data {
-	unsigned	gpio;
-};
-
 #endif /* _ATH79_SPI_PLATFORM_H */
diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig
index 72b059081559..5bd7811844ff 100644
--- a/drivers/spi/Kconfig
+++ b/drivers/spi/Kconfig
@@ -77,6 +77,7 @@ config SPI_ATMEL
 
 config SPI_BCM2835
 	tristate "BCM2835 SPI controller"
+	depends on GPIOLIB
 	depends on ARCH_BCM2835 || COMPILE_TEST
 	depends on GPIOLIB
 	help
@@ -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
@@ -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.
@@ -429,6 +430,12 @@ 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
diff --git a/drivers/spi/Makefile b/drivers/spi/Makefile
index d8cbf654976b..0218f39de7d2 100644
--- a/drivers/spi/Makefile
+++ b/drivers/spi/Makefile
@@ -66,6 +66,7 @@ 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
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-bcm2835.c b/drivers/spi/spi-bcm2835.c
index 37875cf942f7..59705ab23577 100644
--- a/drivers/spi/spi-bcm2835.c
+++ b/drivers/spi/spi-bcm2835.c
@@ -20,18 +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_gpio.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 */
@@ -69,7 +73,8 @@
 #define BCM2835_SPI_CS_CS_01		0x00000001
 
 #define BCM2835_SPI_POLLING_LIMIT_US	30
-#define BCM2835_SPI_TIMEOUT_MS		30000
+#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)
 
@@ -83,6 +88,7 @@ struct bcm2835_spi {
 	u8 *rx_buf;
 	int tx_len;
 	int rx_len;
+	bool dma_pending;
 };
 
 static inline u32 bcm2835_rd(struct bcm2835_spi *bs, unsigned reg)
@@ -128,12 +134,15 @@ static void bcm2835_spi_reset_hw(struct spi_master *master)
 	/* 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)
@@ -157,42 +166,6 @@ static irqreturn_t bcm2835_spi_interrupt(int irq, void *dev_id)
 	return IRQ_HANDLED;
 }
 
-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(master);
-	/* set timeout to 1 second of maximum polling */
-	unsigned long timeout = jiffies + HZ;
-
-	/* enable HW block without interrupts */
-	bcm2835_wr(bs, BCM2835_SPI_CS, cs | BCM2835_SPI_CS_TA);
-
-	/* loop until finished the transfer */
-	while (bs->rx_len) {
-		/* read from fifo as much as possible */
-		bcm2835_rd_fifo(bs);
-		/* fill in tx fifo as much as possible */
-		bcm2835_wr_fifo(bs);
-		/* if we still expect some data after the read,
-		 * check for a possible timeout
-		 */
-		if (bs->rx_len && time_after(jiffies, timeout)) {
-			/* Transfer complete - reset SPI HW */
-			bcm2835_spi_reset_hw(master);
-			/* and return timeout */
-			return -ETIMEDOUT;
-		}
-	}
-
-	/* Transfer complete - reset SPI HW */
-	bcm2835_spi_reset_hw(master);
-	/* and return without waiting for completion */
-	return 0;
-}
-
 static int bcm2835_spi_transfer_one_irq(struct spi_master *master,
 					struct spi_device *spi,
 					struct spi_transfer *tfr,
@@ -229,6 +202,329 @@ static int bcm2835_spi_transfer_one_irq(struct spi_master *master,
 	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 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;
+	}
+
+	/* 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_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(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)
@@ -288,12 +584,26 @@ static int bcm2835_spi_transfer_one(struct spi_master *master,
 		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 void bcm2835_spi_handle_err(struct spi_master *master,
 				   struct spi_message *msg)
 {
+	struct bcm2835_spi *bs = spi_master_get_devdata(master);
+
+	/* 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;
+	}
+	/* and reset */
 	bcm2835_spi_reset_hw(master);
 }
 
@@ -463,6 +773,8 @@ static int bcm2835_spi_probe(struct platform_device *pdev)
 		goto out_clk_disable;
 	}
 
+	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);
@@ -493,6 +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-fsl-dspi.c b/drivers/spi/spi-fsl-dspi.c
index 5fe54cda309f..ed0475d83e50 100644
--- a/drivers/spi/spi-fsl-dspi.c
+++ b/drivers/spi/spi-fsl-dspi.c
@@ -566,7 +566,7 @@ 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-imx.c b/drivers/spi/spi-imx.c
index f08e812b2984..eb7d3a6fb14c 100644
--- a/drivers/spi/spi-imx.c
+++ b/drivers/spi/spi-imx.c
@@ -674,7 +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,
diff --git a/drivers/spi/spi-omap2-mcspi.c b/drivers/spi/spi-omap2-mcspi.c
index d1a5b9fc3eba..a7d85c5ab2fa 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 +243,20 @@ 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;
+	if (spi->controller_state) {
+		l = mcspi_cached_chconf0(spi);
 
-	mcspi_write_chconf0(spi, l);
+		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 +1015,12 @@ static int omap2_mcspi_setup(struct spi_device *spi)
 			return ret;
 	}
 
+	if (gpio_is_valid(spi->cs_gpio)) {
+		if (gpio_request(spi->cs_gpio, dev_name(&spi->dev)) == 0)
+			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 +1060,13 @@ 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 +1076,14 @@ 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 +1100,84 @@ 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 (!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);
+	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;
@@ -1200,75 +1191,58 @@ static void omap2_mcspi_work(struct omap2_mcspi *mcspi, struct spi_message *m)
 	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;
-	int status;
+	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);
-			status = -EINVAL;
-			goto out;
-		}
 
-		if (m->is_dma_mapped || len < DMA_MIN_BYTES)
-			continue;
-
-		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);
-				status = -EINVAL;
-				goto out;
-			}
+	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 (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);
-				status = -EINVAL;
-				goto out;
-			}
+	}
+	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() changes the status inside the
-	 * spi_message, save the status here. */
-	status = m->status;
-out:
-	spi_finalize_current_message(master);
-	return status;
+skip_dma_map:
+	return omap2_mcspi_work_one(mcspi, spi, t);
 }
 
 static int omap2_mcspi_master_setup(struct omap2_mcspi *mcspi)
@@ -1347,7 +1321,8 @@ 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-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-rspi.c b/drivers/spi/spi-rspi.c
index f6bac9e77d06..568ea41625f5 100644
--- a/drivers/spi/spi-rspi.c
+++ b/drivers/spi/spi-rspi.c
@@ -665,15 +665,12 @@ static bool rspi_can_dma(struct spi_master *master, struct spi_device *spi,
 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)) {
-		/* rx_buf can be NULL on RSPI on SH in TX-only Mode */
-		int ret = rspi_dma_transfer(rspi, &xfer->tx_sg,
-					xfer->rx_buf ? &xfer->rx_sg : NULL);
-		if (ret != -EAGAIN)
-			return 0;
-	}
+	if (!rspi->master->can_dma || !__rspi_can_dma(rspi, xfer))
+		return -EAGAIN;
 
-	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,
@@ -1300,7 +1297,7 @@ 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 b1c6731fbf27..2a8c513c4d07 100644
--- a/drivers/spi/spi-s3c64xx.c
+++ b/drivers/spi/spi-s3c64xx.c
@@ -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-sh-msiof.c b/drivers/spi/spi-sh-msiof.c
index bcc7c635d8e7..d3370a612d84 100644
--- a/drivers/spi/spi-sh-msiof.c
+++ b/drivers/spi/spi-sh-msiof.c
@@ -1263,7 +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..ae4bd1154141 100644
--- a/drivers/spi/spi-sirf.c
+++ b/drivers/spi/spi-sirf.c
@@ -559,7 +559,8 @@ 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);
@@ -630,14 +631,47 @@ spi_sirfsoc_setup_transfer(struct spi_device *spi, struct spi_transfer *t)
 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);
+		}
+	}
+	writel(readl(sspi->base + SIRFSOC_SPI_CTRL) | SIRFSOC_SPI_CS_IO_MODE,
+			sspi->base + SIRFSOC_SPI_CTRL);
+	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 int spi_sirfsoc_probe(struct platform_device *pdev)
@@ -646,7 +680,7 @@ static int spi_sirfsoc_probe(struct platform_device *pdev)
 	struct spi_master *master;
 	struct resource *mem_res;
 	int irq;
-	int i, ret;
+	int ret;
 
 	ret = device_reset(&pdev->dev);
 	if (ret) {
@@ -684,6 +718,7 @@ 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) |
@@ -723,7 +758,7 @@ static int spi_sirfsoc_probe(struct platform_device *pdev)
 	/* 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);
+	sspi->dummypage = devm_kzalloc(&pdev->dev, 2 * PAGE_SIZE, GFP_KERNEL);
 	if (!sspi->dummypage) {
 		ret = -ENOMEM;
 		goto free_clk;
@@ -731,27 +766,10 @@ static int spi_sirfsoc_probe(struct platform_device *pdev)
 
 	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);
@@ -774,7 +792,6 @@ static int  spi_sirfsoc_remove(struct platform_device *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);
diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c
index 50910d85df5a..d35c1a13217c 100644
--- a/drivers/spi/spi.c
+++ b/drivers/spi/spi.c
@@ -988,9 +988,6 @@ 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);
@@ -1003,9 +1000,13 @@ 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)
diff --git a/drivers/spi/spidev.c b/drivers/spi/spidev.c
index 92c909eed6b5..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;
 }
 
@@ -647,7 +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;
@@ -676,7 +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 = {