Merge tag 'spi-v5.2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi

Pull spi updates from Mark Brown:
 "One small feature was added this release but the bulk of the diffstat
  and the changelog comes from the fact that several older drivers got
  some fairly hefty reworks and a couple of new drivers were added:

   - Support for detailed control of timing around chip selects from
     Sowjanya Komatineni.

   - A big set of fixes and imrovements for the Tegra114 driver from
     Sowjanya Komatineni.

   - A big simplification of the GPIO driver from Andrey Smirnov.

   - DMA support and fixes for the Freescale LPSPI driver from Clark
     Wang.

   - Fixes and optimizations for the bcm2835aux from Martin Sparl.

   - New drivers for Mediatek MT7621 (graduated from staging) and Zynq
     QSPI"

[ This is a so-called "evil merge" that additionally removes a warning
  due to an unused variable 'i' introduced by commit 1dfbf334f123 ("spi:
  ep93xx: Convert to use CS GPIO descriptors")     - Linus ]

* tag 'spi-v5.2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi: (127 commits)
  spi: rspi: Fix handling of QSPI code when transmit and receive
  spi: atmel-quadspi: fix crash while suspending
  spi: stm32: return the get_irq error
  spi: tegra114: fix PIO transfer
  spi: pxa2xx: fix SCR (divisor) calculation
  spi: Clear SPI_CS_HIGH flag from bad_bits for GPIO chip-select
  spi: ep93xx: Convert to use CS GPIO descriptors
  spi: AD ASoC: declare missing of table
  spi: spi-mem: zynq-qspi: Fix build error on architectures missing readsl/writesl
  spi: stm32-qspi: manage the get_irq error case
  spi/spi-bcm2835: Split transfers that exceed DLEN
  spi: expand mode support
  dt-bindings: spi: spi-mt65xx: add support for MT8516
  spi: pxa2xx: Add support for Intel Comet Lake
  spi/trace: Cap buffer contents at 64 bytes
  spi: Release spi_res after finalizing message
  spi: Remove warning in spi_split_transfers_maxsize()
  spi: Remove one needless transfer speed fall back case
  spi: sh-msiof: Document r8a77470 bindings
  spi: pxa2xx: use a module softdep for dw_dmac
  ...

33 files changed, 2824 insertions, 702 deletions

diff --git a/drivers/spi/Kconfig b/drivers/spi/Kconfig
index f761655e2a36..0fba8f400c59 100644
--- a/drivers/spi/Kconfig
+++ b/drivers/spi/Kconfig
@@ -426,6 +426,12 @@ config SPI_MT65XX
 	  say Y or M here.If you are not sure, say N.
 	  SPI drivers for Mediatek MT65XX and MT81XX series ARM SoCs.
 
+config SPI_MT7621
+	tristate "MediaTek MT7621 SPI Controller"
+	depends on RALINK || COMPILE_TEST
+	help
+	  This selects a driver for the MediaTek MT7621 SPI Controller.
+
 config SPI_NPCM_PSPI
 	tristate "Nuvoton NPCM PSPI Controller"
 	depends on ARCH_NPCM || COMPILE_TEST
@@ -842,9 +848,17 @@ config SPI_XTENSA_XTFPGA
 	  16 bit words in SPI mode 0, automatically asserting CS on transfer
 	  start and deasserting on end.
 
+config SPI_ZYNQ_QSPI
+	tristate "Xilinx Zynq QSPI controller"
+	depends on ARCH_ZYNQ || COMPILE_TEST
+	help
+	  This enables support for the Zynq Quad SPI controller
+	  in master mode.
+	  This controller only supports SPI memory interface.
+
 config SPI_ZYNQMP_GQSPI
 	tristate "Xilinx ZynqMP GQSPI controller"
-	depends on SPI_MASTER && HAS_DMA
+	depends on (SPI_MASTER && HAS_DMA) || COMPILE_TEST
 	help
 	  Enables Xilinx GQSPI controller driver for Zynq UltraScale+ MPSoC.
 
diff --git a/drivers/spi/Makefile b/drivers/spi/Makefile
index d8fc03c9faa2..f2f78d03dc28 100644
--- a/drivers/spi/Makefile
+++ b/drivers/spi/Makefile
@@ -60,6 +60,7 @@ obj-$(CONFIG_SPI_MPC512x_PSC)		+= spi-mpc512x-psc.o
 obj-$(CONFIG_SPI_MPC52xx_PSC)		+= spi-mpc52xx-psc.o
 obj-$(CONFIG_SPI_MPC52xx)		+= spi-mpc52xx.o
 obj-$(CONFIG_SPI_MT65XX)                += spi-mt65xx.o
+obj-$(CONFIG_SPI_MT7621)		+= spi-mt7621.o
 obj-$(CONFIG_SPI_MXIC)			+= spi-mxic.o
 obj-$(CONFIG_SPI_MXS)			+= spi-mxs.o
 obj-$(CONFIG_SPI_NPCM_PSPI)		+= spi-npcm-pspi.o
@@ -118,6 +119,7 @@ obj-$(CONFIG_SPI_XCOMM)		+= spi-xcomm.o
 obj-$(CONFIG_SPI_XILINX)		+= spi-xilinx.o
 obj-$(CONFIG_SPI_XLP)			+= spi-xlp.o
 obj-$(CONFIG_SPI_XTENSA_XTFPGA)		+= spi-xtensa-xtfpga.o
+obj-$(CONFIG_SPI_ZYNQ_QSPI)		+= spi-zynq-qspi.o
 obj-$(CONFIG_SPI_ZYNQMP_GQSPI)		+= spi-zynqmp-gqspi.o
 
 # SPI slave protocol handlers
diff --git a/drivers/spi/atmel-quadspi.c b/drivers/spi/atmel-quadspi.c
index fffc21cd5f79..9f24d5f0b431 100644
--- a/drivers/spi/atmel-quadspi.c
+++ b/drivers/spi/atmel-quadspi.c
@@ -366,7 +366,7 @@ static int atmel_qspi_exec_op(struct spi_mem *mem, const struct spi_mem_op *op)
 	return err;
 }
 
-const char *atmel_qspi_get_name(struct spi_mem *spimem)
+static const char *atmel_qspi_get_name(struct spi_mem *spimem)
 {
 	return dev_name(spimem->spi->dev.parent);
 }
@@ -570,7 +570,8 @@ static int atmel_qspi_remove(struct platform_device *pdev)
 
 static int __maybe_unused atmel_qspi_suspend(struct device *dev)
 {
-	struct atmel_qspi *aq = dev_get_drvdata(dev);
+	struct spi_controller *ctrl = dev_get_drvdata(dev);
+	struct atmel_qspi *aq = spi_controller_get_devdata(ctrl);
 
 	clk_disable_unprepare(aq->qspick);
 	clk_disable_unprepare(aq->pclk);
@@ -580,7 +581,8 @@ static int __maybe_unused atmel_qspi_suspend(struct device *dev)
 
 static int __maybe_unused atmel_qspi_resume(struct device *dev)
 {
-	struct atmel_qspi *aq = dev_get_drvdata(dev);
+	struct spi_controller *ctrl = dev_get_drvdata(dev);
+	struct atmel_qspi *aq = spi_controller_get_devdata(ctrl);
 
 	clk_prepare_enable(aq->pclk);
 	clk_prepare_enable(aq->qspick);
diff --git a/drivers/spi/spi-at91-usart.c b/drivers/spi/spi-at91-usart.c
index a694d702e574..f763e14bdf12 100644
--- a/drivers/spi/spi-at91-usart.c
+++ b/drivers/spi/spi-at91-usart.c
@@ -178,12 +178,6 @@ static int at91_usart_spi_setup(struct spi_device *spi)
 	struct at91_usart_spi *aus = spi_master_get_devdata(spi->controller);
 	u32 *ausd = spi->controller_state;
 	unsigned int mr = at91_usart_spi_readl(aus, MR);
-	u8 bits = spi->bits_per_word;
-
-	if (bits != 8) {
-		dev_dbg(&spi->dev, "Only 8 bits per word are supported\n");
-		return -EINVAL;
-	}
 
 	if (spi->mode & SPI_CPOL)
 		mr |= US_MR_CPOL;
@@ -212,7 +206,7 @@ static int at91_usart_spi_setup(struct spi_device *spi)
 
 	dev_dbg(&spi->dev,
 		"setup: bpw %u mode 0x%x -> mr %d %08x\n",
-		bits, spi->mode, spi->chip_select, mr);
+		spi->bits_per_word, spi->mode, spi->chip_select, mr);
 
 	return 0;
 }
diff --git a/drivers/spi/spi-bcm2835.c b/drivers/spi/spi-bcm2835.c
index 35aebdfd3b4e..8aa22713c483 100644
--- a/drivers/spi/spi-bcm2835.c
+++ b/drivers/spi/spi-bcm2835.c
@@ -335,20 +335,6 @@ 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
- *
- * 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...
- */
-
 /**
  * bcm2835_spi_transfer_prologue() - transfer first few bytes without DMA
  * @master: SPI master
@@ -630,19 +616,6 @@ static bool bcm2835_spi_can_dma(struct spi_master *master,
 	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;
-	}
-
 	/* return OK */
 	return true;
 }
@@ -707,7 +680,6 @@ static void bcm2835_dma_init(struct spi_master *master, struct device *dev)
 
 	/* 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;
 
@@ -844,6 +816,17 @@ static int bcm2835_spi_prepare_message(struct spi_master *master,
 	struct spi_device *spi = msg->spi;
 	struct bcm2835_spi *bs = spi_master_get_devdata(master);
 	u32 cs = bcm2835_rd(bs, BCM2835_SPI_CS);
+	int ret;
+
+	/*
+	 * DMA transfers are limited to 16 bit (0 to 65535 bytes) by the SPI HW
+	 * due to DLEN. Split up transfers (32-bit FIFO aligned) if the limit is
+	 * exceeded.
+	 */
+	ret = spi_split_transfers_maxsize(master, msg, 65532,
+					  GFP_KERNEL | GFP_DMA);
+	if (ret)
+		return ret;
 
 	cs &= ~(BCM2835_SPI_CS_CPOL | BCM2835_SPI_CS_CPHA);
 
diff --git a/drivers/spi/spi-bcm2835aux.c b/drivers/spi/spi-bcm2835aux.c
index f7e054848ca5..bbf87adb3ff8 100644
--- a/drivers/spi/spi-bcm2835aux.c
+++ b/drivers/spi/spi-bcm2835aux.c
@@ -21,6 +21,7 @@
 
 #include <linux/clk.h>
 #include <linux/completion.h>
+#include <linux/debugfs.h>
 #include <linux/delay.h>
 #include <linux/err.h>
 #include <linux/interrupt.h>
@@ -36,6 +37,12 @@
 #include <linux/spi/spi.h>
 #include <linux/spinlock.h>
 
+/* define polling limits */
+static unsigned int polling_limit_us = 30;
+module_param(polling_limit_us, uint, 0664);
+MODULE_PARM_DESC(polling_limit_us,
+		 "time in us to run a transfer in polling mode - if zero no polling is used\n");
+
 /*
  * spi register defines
  *
@@ -88,10 +95,6 @@
 #define BCM2835_AUX_SPI_STAT_BUSY	0x00000040
 #define BCM2835_AUX_SPI_STAT_BITCOUNT	0x0000003F
 
-/* timeout values */
-#define BCM2835_AUX_SPI_POLLING_LIMIT_US	30
-#define BCM2835_AUX_SPI_POLLING_JIFFIES		2
-
 struct bcm2835aux_spi {
 	void __iomem *regs;
 	struct clk *clk;
@@ -102,8 +105,53 @@ struct bcm2835aux_spi {
 	int tx_len;
 	int rx_len;
 	int pending;
+
+	u64 count_transfer_polling;
+	u64 count_transfer_irq;
+	u64 count_transfer_irq_after_poll;
+
+	struct dentry *debugfs_dir;
 };
 
+#if defined(CONFIG_DEBUG_FS)
+static void bcm2835aux_debugfs_create(struct bcm2835aux_spi *bs,
+				      const char *dname)
+{
+	char name[64];
+	struct dentry *dir;
+
+	/* get full name */
+	snprintf(name, sizeof(name), "spi-bcm2835aux-%s", dname);
+
+	/* the base directory */
+	dir = debugfs_create_dir(name, NULL);
+	bs->debugfs_dir = dir;
+
+	/* the counters */
+	debugfs_create_u64("count_transfer_polling", 0444, dir,
+			   &bs->count_transfer_polling);
+	debugfs_create_u64("count_transfer_irq", 0444, dir,
+			   &bs->count_transfer_irq);
+	debugfs_create_u64("count_transfer_irq_after_poll", 0444, dir,
+			   &bs->count_transfer_irq_after_poll);
+}
+
+static void bcm2835aux_debugfs_remove(struct bcm2835aux_spi *bs)
+{
+	debugfs_remove_recursive(bs->debugfs_dir);
+	bs->debugfs_dir = NULL;
+}
+#else
+static void bcm2835aux_debugfs_create(struct bcm2835aux_spi *bs,
+				      const char *dname)
+{
+}
+
+static void bcm2835aux_debugfs_remove(struct bcm2835aux_spi *bs)
+{
+}
+#endif /* CONFIG_DEBUG_FS */
+
 static inline u32 bcm2835aux_rd(struct bcm2835aux_spi *bs, unsigned reg)
 {
 	return readl(bs->regs + reg);
@@ -123,9 +171,6 @@ static inline void bcm2835aux_rd_fifo(struct bcm2835aux_spi *bs)
 	data = bcm2835aux_rd(bs, BCM2835_AUX_SPI_IO);
 	if (bs->rx_buf) {
 		switch (count) {
-		case 4:
-			*bs->rx_buf++ = (data >> 24) & 0xff;
-			/* fallthrough */
 		case 3:
 			*bs->rx_buf++ = (data >> 16) & 0xff;
 			/* fallthrough */
@@ -178,24 +223,14 @@ static void bcm2835aux_spi_reset_hw(struct bcm2835aux_spi *bs)
 		      BCM2835_AUX_SPI_CNTL0_CLEARFIFO);
 }
 
-static irqreturn_t bcm2835aux_spi_interrupt(int irq, void *dev_id)
+static void bcm2835aux_spi_transfer_helper(struct bcm2835aux_spi *bs)
 {
-	struct spi_master *master = dev_id;
-	struct bcm2835aux_spi *bs = spi_master_get_devdata(master);
-	irqreturn_t ret = IRQ_NONE;
-
-	/* IRQ may be shared, so return if our interrupts are disabled */
-	if (!(bcm2835aux_rd(bs, BCM2835_AUX_SPI_CNTL1) &
-	      (BCM2835_AUX_SPI_CNTL1_TXEMPTY | BCM2835_AUX_SPI_CNTL1_IDLE)))
-		return ret;
+	u32 stat = bcm2835aux_rd(bs, BCM2835_AUX_SPI_STAT);
 
 	/* check if we have data to read */
-	while (bs->rx_len &&
-	       (!(bcm2835aux_rd(bs, BCM2835_AUX_SPI_STAT) &
-		  BCM2835_AUX_SPI_STAT_RX_EMPTY))) {
+	for (; bs->rx_len && (stat & BCM2835_AUX_SPI_STAT_RX_LVL);
+	     stat = bcm2835aux_rd(bs, BCM2835_AUX_SPI_STAT))
 		bcm2835aux_rd_fifo(bs);
-		ret = IRQ_HANDLED;
-	}
 
 	/* check if we have data to write */
 	while (bs->tx_len &&
@@ -203,16 +238,21 @@ static irqreturn_t bcm2835aux_spi_interrupt(int irq, void *dev_id)
 	       (!(bcm2835aux_rd(bs, BCM2835_AUX_SPI_STAT) &
 		  BCM2835_AUX_SPI_STAT_TX_FULL))) {
 		bcm2835aux_wr_fifo(bs);
-		ret = IRQ_HANDLED;
 	}
+}
 
-	/* and check if we have reached "done" */
-	while (bs->rx_len &&
-	       (!(bcm2835aux_rd(bs, BCM2835_AUX_SPI_STAT) &
-		  BCM2835_AUX_SPI_STAT_BUSY))) {
-		bcm2835aux_rd_fifo(bs);
-		ret = IRQ_HANDLED;
-	}
+static irqreturn_t bcm2835aux_spi_interrupt(int irq, void *dev_id)
+{
+	struct spi_master *master = dev_id;
+	struct bcm2835aux_spi *bs = spi_master_get_devdata(master);
+
+	/* IRQ may be shared, so return if our interrupts are disabled */
+	if (!(bcm2835aux_rd(bs, BCM2835_AUX_SPI_CNTL1) &
+	      (BCM2835_AUX_SPI_CNTL1_TXEMPTY | BCM2835_AUX_SPI_CNTL1_IDLE)))
+		return IRQ_NONE;
+
+	/* do common fifo handling */
+	bcm2835aux_spi_transfer_helper(bs);
 
 	if (!bs->tx_len) {
 		/* disable tx fifo empty interrupt */
@@ -226,8 +266,7 @@ static irqreturn_t bcm2835aux_spi_interrupt(int irq, void *dev_id)
 		complete(&master->xfer_completion);
 	}
 
-	/* and return */
-	return ret;
+	return IRQ_HANDLED;
 }
 
 static int __bcm2835aux_spi_transfer_one_irq(struct spi_master *master,
@@ -251,6 +290,9 @@ static int bcm2835aux_spi_transfer_one_irq(struct spi_master *master,
 {
 	struct bcm2835aux_spi *bs = spi_master_get_devdata(master);
 
+	/* update statistics */
+	bs->count_transfer_irq++;
+
 	/* fill in registers and fifos before enabling interrupts */
 	bcm2835aux_wr(bs, BCM2835_AUX_SPI_CNTL1, bs->cntl[1]);
 	bcm2835aux_wr(bs, BCM2835_AUX_SPI_CNTL0, bs->cntl[0]);
@@ -273,35 +315,22 @@ static int bcm2835aux_spi_transfer_one_poll(struct spi_master *master,
 {
 	struct bcm2835aux_spi *bs = spi_master_get_devdata(master);
 	unsigned long timeout;
-	u32 stat;
+
+	/* update statistics */
+	bs->count_transfer_polling++;
 
 	/* configure spi */
 	bcm2835aux_wr(bs, BCM2835_AUX_SPI_CNTL1, bs->cntl[1]);
 	bcm2835aux_wr(bs, BCM2835_AUX_SPI_CNTL0, bs->cntl[0]);
 
-	/* set the timeout */
-	timeout = jiffies + BCM2835_AUX_SPI_POLLING_JIFFIES;
+	/* set the timeout to at least 2 jiffies */
+	timeout = jiffies + 2 + HZ * polling_limit_us / 1000000;
 
 	/* loop until finished the transfer */
 	while (bs->rx_len) {
-		/* read status */
-		stat = bcm2835aux_rd(bs, BCM2835_AUX_SPI_STAT);
 
-		/* fill in tx fifo with remaining data */
-		if ((bs->tx_len) && (!(stat & BCM2835_AUX_SPI_STAT_TX_FULL))) {
-			bcm2835aux_wr_fifo(bs);
-			continue;
-		}
-
-		/* read data from fifo for both cases */
-		if (!(stat & BCM2835_AUX_SPI_STAT_RX_EMPTY)) {
-			bcm2835aux_rd_fifo(bs);
-			continue;
-		}
-		if (!(stat & BCM2835_AUX_SPI_STAT_BUSY)) {
-			bcm2835aux_rd_fifo(bs);
-			continue;
-		}
+		/* do common fifo handling */
+		bcm2835aux_spi_transfer_helper(bs);
 
 		/* there is still data pending to read check the timeout */
 		if (bs->rx_len && time_after(jiffies, timeout)) {
@@ -310,6 +339,7 @@ static int bcm2835aux_spi_transfer_one_poll(struct spi_master *master,
 					    jiffies - timeout,
 					    bs->tx_len, bs->rx_len);
 			/* forward to interrupt handler */
+			bs->count_transfer_irq_after_poll++;
 			return __bcm2835aux_spi_transfer_one_irq(master,
 							       spi, tfr);
 		}
@@ -324,8 +354,8 @@ static int bcm2835aux_spi_transfer_one(struct spi_master *master,
 				       struct spi_transfer *tfr)
 {
 	struct bcm2835aux_spi *bs = spi_master_get_devdata(master);
-	unsigned long spi_hz, clk_hz, speed;
-	unsigned long spi_used_hz;
+	unsigned long spi_hz, clk_hz, speed, spi_used_hz;
+	unsigned long hz_per_byte, byte_limit;
 
 	/* calculate the registers to handle
 	 *
@@ -369,14 +399,15 @@ static int bcm2835aux_spi_transfer_one(struct spi_master *master,
 	 * of Hz per byte per polling limit.  E.g., we can transfer 1 byte in
 	 * 30 µs per 300,000 Hz of bus clock.
 	 */
-#define HZ_PER_BYTE ((9 * 1000000) / BCM2835_AUX_SPI_POLLING_LIMIT_US)
+	hz_per_byte = polling_limit_us ? (9 * 1000000) / polling_limit_us : 0;
+	byte_limit = hz_per_byte ? spi_used_hz / hz_per_byte : 1;
+
 	/* run in polling mode for short transfers */
-	if (tfr->len < spi_used_hz / HZ_PER_BYTE)
+	if (tfr->len < byte_limit)
 		return bcm2835aux_spi_transfer_one_poll(master, spi, tfr);
 
 	/* run in interrupt mode for all others */
 	return bcm2835aux_spi_transfer_one_irq(master, spi, tfr);
-#undef HZ_PER_BYTE
 }
 
 static int bcm2835aux_spi_prepare_message(struct spi_master *master,
@@ -421,6 +452,50 @@ static void bcm2835aux_spi_handle_err(struct spi_master *master,
 	bcm2835aux_spi_reset_hw(bs);
 }
 
+static int bcm2835aux_spi_setup(struct spi_device *spi)
+{
+	int ret;
+
+	/* sanity check for native cs */
+	if (spi->mode & SPI_NO_CS)
+		return 0;
+	if (gpio_is_valid(spi->cs_gpio)) {
+		/* with gpio-cs set the GPIO to the correct level
+		 * and as output (in case the dt has the gpio not configured
+		 * as output but native cs)
+		 */
+		ret = gpio_direction_output(spi->cs_gpio,
+					    (spi->mode & SPI_CS_HIGH) ? 0 : 1);
+		if (ret)
+			dev_err(&spi->dev,
+				"could not set gpio %i as output: %i\n",
+				spi->cs_gpio, ret);
+
+		return ret;
+	}
+
+	/* for dt-backwards compatibility: only support native on CS0
+	 * known things not supported with broken native CS:
+	 * * multiple chip-selects: cs0-cs2 are all
+	 *     simultaniously asserted whenever there is a transfer
+	 *     this even includes SPI_NO_CS
+	 * * SPI_CS_HIGH: cs are always asserted low
+	 * * cs_change: cs is deasserted after each spi_transfer
+	 * * cs_delay_usec: cs is always deasserted one SCK cycle
+	 *     after the last transfer
+	 * probably more...
+	 */
+	dev_warn(&spi->dev,
+		 "Native CS is not supported - please configure cs-gpio in device-tree\n");
+
+	if (spi->chip_select == 0)
+		return 0;
+
+	dev_warn(&spi->dev, "Native CS is not working for cs > 0\n");
+
+	return -EINVAL;
+}
+
 static int bcm2835aux_spi_probe(struct platform_device *pdev)
 {
 	struct spi_master *master;
@@ -438,7 +513,19 @@ static int bcm2835aux_spi_probe(struct platform_device *pdev)
 	platform_set_drvdata(pdev, master);
 	master->mode_bits = (SPI_CPOL | SPI_CS_HIGH | SPI_NO_CS);
 	master->bits_per_word_mask = SPI_BPW_MASK(8);
-	master->num_chipselect = -1;
+	/* even though the driver never officially supported native CS
+	 * allow a single native CS for legacy DT support purposes when
+	 * no cs-gpio is configured.
+	 * Known limitations for native cs are:
+	 * * multiple chip-selects: cs0-cs2 are all simultaniously asserted
+	 *     whenever there is a transfer -  this even includes SPI_NO_CS
+	 * * SPI_CS_HIGH: is ignores - cs are always asserted low
+	 * * cs_change: cs is deasserted after each spi_transfer
+	 * * cs_delay_usec: cs is always deasserted one SCK cycle after
+	 *     a spi_transfer
+	 */
+	master->num_chipselect = 1;
+	master->setup = bcm2835aux_spi_setup;
 	master->transfer_one = bcm2835aux_spi_transfer_one;
 	master->handle_err = bcm2835aux_spi_handle_err;
 	master->prepare_message = bcm2835aux_spi_prepare_message;
@@ -502,6 +589,8 @@ static int bcm2835aux_spi_probe(struct platform_device *pdev)
 		goto out_clk_disable;
 	}
 
+	bcm2835aux_debugfs_create(bs, dev_name(&pdev->dev));
+
 	return 0;
 
 out_clk_disable:
@@ -516,6 +605,8 @@ static int bcm2835aux_spi_remove(struct platform_device *pdev)
 	struct spi_master *master = platform_get_drvdata(pdev);
 	struct bcm2835aux_spi *bs = spi_master_get_devdata(master);
 
+	bcm2835aux_debugfs_remove(bs);
+
 	bcm2835aux_spi_reset_hw(bs);
 
 	/* disable the HW block by releasing the clock */
diff --git a/drivers/spi/spi-bitbang.c b/drivers/spi/spi-bitbang.c
index dd9a8c54a693..4243e53f9f7b 100644
--- a/drivers/spi/spi-bitbang.c
+++ b/drivers/spi/spi-bitbang.c
@@ -335,6 +335,42 @@ static void spi_bitbang_set_cs(struct spi_device *spi, bool enable)
 
 /*----------------------------------------------------------------------*/
 
+int spi_bitbang_init(struct spi_bitbang *bitbang)
+{
+	struct spi_master *master = bitbang->master;
+
+	if (!master || !bitbang->chipselect)
+		return -EINVAL;
+
+	mutex_init(&bitbang->lock);
+
+	if (!master->mode_bits)
+		master->mode_bits = SPI_CPOL | SPI_CPHA | bitbang->flags;
+
+	if (master->transfer || master->transfer_one_message)
+		return -EINVAL;
+
+	master->prepare_transfer_hardware = spi_bitbang_prepare_hardware;
+	master->unprepare_transfer_hardware = spi_bitbang_unprepare_hardware;
+	master->transfer_one = spi_bitbang_transfer_one;
+	master->set_cs = spi_bitbang_set_cs;
+
+	if (!bitbang->txrx_bufs) {
+		bitbang->use_dma = 0;
+		bitbang->txrx_bufs = spi_bitbang_bufs;
+		if (!master->setup) {
+			if (!bitbang->setup_transfer)
+				bitbang->setup_transfer =
+					 spi_bitbang_setup_transfer;
+			master->setup = spi_bitbang_setup;
+			master->cleanup = spi_bitbang_cleanup;
+		}
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(spi_bitbang_init);
+
 /**
  * spi_bitbang_start - start up a polled/bitbanging SPI master driver
  * @bitbang: driver handle
@@ -368,33 +404,9 @@ int spi_bitbang_start(struct spi_bitbang *bitbang)
 	struct spi_master *master = bitbang->master;
 	int ret;
 
-	if (!master || !bitbang->chipselect)
-		return -EINVAL;
-
-	mutex_init(&bitbang->lock);
-
-	if (!master->mode_bits)
-		master->mode_bits = SPI_CPOL | SPI_CPHA | bitbang->flags;
-
-	if (master->transfer || master->transfer_one_message)
-		return -EINVAL;
-
-	master->prepare_transfer_hardware = spi_bitbang_prepare_hardware;
-	master->unprepare_transfer_hardware = spi_bitbang_unprepare_hardware;
-	master->transfer_one = spi_bitbang_transfer_one;
-	master->set_cs = spi_bitbang_set_cs;
-
-	if (!bitbang->txrx_bufs) {
-		bitbang->use_dma = 0;
-		bitbang->txrx_bufs = spi_bitbang_bufs;
-		if (!master->setup) {
-			if (!bitbang->setup_transfer)
-				bitbang->setup_transfer =
-					 spi_bitbang_setup_transfer;
-			master->setup = spi_bitbang_setup;
-			master->cleanup = spi_bitbang_cleanup;
-		}
-	}
+	ret = spi_bitbang_init(bitbang);
+	if (ret)
+		return ret;
 
 	/* driver may get busy before register() returns, especially
 	 * if someone registered boardinfo for devices
diff --git a/drivers/spi/spi-dw-mmio.c b/drivers/spi/spi-dw-mmio.c
index 4bd59a93d988..de952b17bc10 100644
--- a/drivers/spi/spi-dw-mmio.c
+++ b/drivers/spi/spi-dw-mmio.c
@@ -30,6 +30,7 @@
 struct dw_spi_mmio {
 	struct dw_spi  dws;
 	struct clk     *clk;
+	struct clk     *pclk;
 	void           *priv;
 };
 
@@ -172,6 +173,14 @@ static int dw_spi_mmio_probe(struct platform_device *pdev)
 	if (ret)
 		return ret;
 
+	/* Optional clock needed to access the registers */
+	dwsmmio->pclk = devm_clk_get_optional(&pdev->dev, "pclk");
+	if (IS_ERR(dwsmmio->pclk))
+		return PTR_ERR(dwsmmio->pclk);
+	ret = clk_prepare_enable(dwsmmio->pclk);
+	if (ret)
+		goto out_clk;
+
 	dws->bus_num = pdev->id;
 
 	dws->max_freq = clk_get_rate(dwsmmio->clk);
@@ -199,6 +208,8 @@ static int dw_spi_mmio_probe(struct platform_device *pdev)
 	return 0;
 
 out:
+	clk_disable_unprepare(dwsmmio->pclk);
+out_clk:
 	clk_disable_unprepare(dwsmmio->clk);
 	return ret;
 }
@@ -208,6 +219,7 @@ static int dw_spi_mmio_remove(struct platform_device *pdev)
 	struct dw_spi_mmio *dwsmmio = platform_get_drvdata(pdev);
 
 	dw_spi_remove_host(&dwsmmio->dws);
+	clk_disable_unprepare(dwsmmio->pclk);
 	clk_disable_unprepare(dwsmmio->clk);
 
 	return 0;
diff --git a/drivers/spi/spi-ep93xx.c b/drivers/spi/spi-ep93xx.c
index 79fc3940245a..81889389280b 100644
--- a/drivers/spi/spi-ep93xx.c
+++ b/drivers/spi/spi-ep93xx.c
@@ -28,7 +28,6 @@
 #include <linux/platform_device.h>
 #include <linux/sched.h>
 #include <linux/scatterlist.h>
-#include <linux/gpio.h>
 #include <linux/spi/spi.h>
 
 #include <linux/platform_data/dma-ep93xx.h>
@@ -652,7 +651,6 @@ static int ep93xx_spi_probe(struct platform_device *pdev)
 	struct resource *res;
 	int irq;
 	int error;
-	int i;
 
 	info = dev_get_platdata(&pdev->dev);
 	if (!info) {
@@ -676,6 +674,7 @@ static int ep93xx_spi_probe(struct platform_device *pdev)
 	if (!master)
 		return -ENOMEM;
 
+	master->use_gpio_descriptors = true;
 	master->prepare_transfer_hardware = ep93xx_spi_prepare_hardware;
 	master->unprepare_transfer_hardware = ep93xx_spi_unprepare_hardware;
 	master->prepare_message = ep93xx_spi_prepare_message;
@@ -683,31 +682,11 @@ static int ep93xx_spi_probe(struct platform_device *pdev)
 	master->bus_num = pdev->id;
 	master->mode_bits = SPI_CPOL | SPI_CPHA | SPI_CS_HIGH;
 	master->bits_per_word_mask = SPI_BPW_RANGE_MASK(4, 16);
-
-	master->num_chipselect = info->num_chipselect;
-	master->cs_gpios = devm_kcalloc(&master->dev,
-					master->num_chipselect, sizeof(int),
-					GFP_KERNEL);
-	if (!master->cs_gpios) {
-		error = -ENOMEM;
-		goto fail_release_master;
-	}
-
-	for (i = 0; i < master->num_chipselect; i++) {
-		master->cs_gpios[i] = info->chipselect[i];
-
-		if (!gpio_is_valid(master->cs_gpios[i]))
-			continue;
-
-		error = devm_gpio_request_one(&pdev->dev, master->cs_gpios[i],
-					      GPIOF_OUT_INIT_HIGH,
-					      "ep93xx-spi");
-		if (error) {
-			dev_err(&pdev->dev, "could not request cs gpio %d\n",
-				master->cs_gpios[i]);
-			goto fail_release_master;
-		}
-	}
+	/*
+	 * The SPI core will count the number of GPIO descriptors to figure
+	 * out the number of chip selects available on the platform.
+	 */
+	master->num_chipselect = 0;
 
 	platform_set_drvdata(pdev, master);
 
diff --git a/drivers/spi/spi-fsl-lib.h b/drivers/spi/spi-fsl-lib.h
index f303f306b38e..483734bc1b1e 100644
--- a/drivers/spi/spi-fsl-lib.h
+++ b/drivers/spi/spi-fsl-lib.h
@@ -95,8 +95,10 @@ static inline u32 mpc8xxx_spi_read_reg(__be32 __iomem *reg)
 
 struct mpc8xxx_spi_probe_info {
 	struct fsl_spi_platform_data pdata;
+	int ngpios;
 	int *gpios;
 	bool *alow_flags;
+	__be32 __iomem *immr_spi_cs;
 };
 
 extern u32 mpc8xxx_spi_tx_buf_u8(struct mpc8xxx_spi *mpc8xxx_spi);
diff --git a/drivers/spi/spi-fsl-lpspi.c b/drivers/spi/spi-fsl-lpspi.c
index 391863914043..d08e9324140e 100644
--- a/drivers/spi/spi-fsl-lpspi.c
+++ b/drivers/spi/spi-fsl-lpspi.c
@@ -8,7 +8,10 @@
 #include <linux/clk.h>
 #include <linux/completion.h>
 #include <linux/delay.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
 #include <linux/err.h>
+#include <linux/gpio.h>
 #include <linux/interrupt.h>
 #include <linux/io.h>
 #include <linux/irq.h>
@@ -16,7 +19,12 @@
 #include <linux/module.h>
 #include <linux/of.h>
 #include <linux/of_device.h>
+#include <linux/of_gpio.h>
+#include <linux/pinctrl/consumer.h>
 #include <linux/platform_device.h>
+#include <linux/platform_data/dma-imx.h>
+#include <linux/platform_data/spi-imx.h>
+#include <linux/pm_runtime.h>
 #include <linux/slab.h>
 #include <linux/spi/spi.h>
 #include <linux/spi/spi_bitbang.h>
@@ -24,6 +32,11 @@
 
 #define DRIVER_NAME "fsl_lpspi"
 
+#define FSL_LPSPI_RPM_TIMEOUT 50 /* 50ms */
+
+/* The maximum bytes that edma can transfer once.*/
+#define FSL_LPSPI_MAX_EDMA_BYTES  ((1 << 15) - 1)
+
 /* i.MX7ULP LPSPI registers */
 #define IMX7ULP_VERID	0x0
 #define IMX7ULP_PARAM	0x4
@@ -57,12 +70,14 @@
 #define IER_FCIE	BIT(9)
 #define IER_RDIE	BIT(1)
 #define IER_TDIE	BIT(0)
+#define DER_RDDE	BIT(1)
+#define DER_TDDE	BIT(0)
 #define CFGR1_PCSCFG	BIT(27)
 #define CFGR1_PINCFG	(BIT(24)|BIT(25))
 #define CFGR1_PCSPOL	BIT(8)
 #define CFGR1_NOSTALL	BIT(3)
 #define CFGR1_MASTER	BIT(0)
-#define FSR_RXCOUNT	(BIT(16)|BIT(17)|BIT(18))
+#define FSR_TXCOUNT	(0xFF)
 #define RSR_RXEMPTY	BIT(1)
 #define TCR_CPOL	BIT(31)
 #define TCR_CPHA	BIT(30)
@@ -84,8 +99,11 @@ struct lpspi_config {
 struct fsl_lpspi_data {
 	struct device *dev;
 	void __iomem *base;
-	struct clk *clk;
+	unsigned long base_phys;
+	struct clk *clk_ipg;
+	struct clk *clk_per;
 	bool is_slave;
+	bool is_first_byte;
 
 	void *rx_buf;
 	const void *tx_buf;
@@ -101,6 +119,13 @@ struct fsl_lpspi_data {
 	struct completion xfer_done;
 
 	bool slave_aborted;
+
+	/* DMA */
+	bool usedma;
+	struct completion dma_rx_completion;
+	struct completion dma_tx_completion;
+
+	int chipselect[0];
 };
 
 static const struct of_device_id fsl_lpspi_dt_ids[] = {
@@ -147,12 +172,48 @@ static void fsl_lpspi_intctrl(struct fsl_lpspi_data *fsl_lpspi,
 	writel(enable, fsl_lpspi->base + IMX7ULP_IER);
 }
 
+static int fsl_lpspi_bytes_per_word(const int bpw)
+{
+	return DIV_ROUND_UP(bpw, BITS_PER_BYTE);
+}
+
+static bool fsl_lpspi_can_dma(struct spi_controller *controller,
+			      struct spi_device *spi,
+			      struct spi_transfer *transfer)
+{
+	unsigned int bytes_per_word;
+
+	if (!controller->dma_rx)
+		return false;
+
+	bytes_per_word = fsl_lpspi_bytes_per_word(transfer->bits_per_word);
+
+	switch (bytes_per_word)
+	{
+		case 1:
+		case 2:
+		case 4:
+			break;
+		default:
+			return false;
+	}
+
+	return true;
+}
+
 static int lpspi_prepare_xfer_hardware(struct spi_controller *controller)
 {
 	struct fsl_lpspi_data *fsl_lpspi =
 				spi_controller_get_devdata(controller);
+	int ret;
+
+	ret = pm_runtime_get_sync(fsl_lpspi->dev);
+	if (ret < 0) {
+		dev_err(fsl_lpspi->dev, "failed to enable clock\n");
+		return ret;
+	}
 
-	return clk_prepare_enable(fsl_lpspi->clk);
+	return 0;
 }
 
 static int lpspi_unprepare_xfer_hardware(struct spi_controller *controller)
@@ -160,7 +221,22 @@ static int lpspi_unprepare_xfer_hardware(struct spi_controller *controller)
 	struct fsl_lpspi_data *fsl_lpspi =
 				spi_controller_get_devdata(controller);
 
-	clk_disable_unprepare(fsl_lpspi->clk);
+	pm_runtime_mark_last_busy(fsl_lpspi->dev);
+	pm_runtime_put_autosuspend(fsl_lpspi->dev);
+
+	return 0;
+}
+
+static int fsl_lpspi_prepare_message(struct spi_controller *controller,
+				     struct spi_message *msg)
+{
+	struct fsl_lpspi_data *fsl_lpspi =
+					spi_controller_get_devdata(controller);
+	struct spi_device *spi = msg->spi;
+	int gpio = fsl_lpspi->chipselect[spi->chip_select];
+
+	if (gpio_is_valid(gpio))
+		gpio_direction_output(gpio, spi->mode & SPI_CS_HIGH ? 0 : 1);
 
 	return 0;
 }
@@ -197,8 +273,7 @@ static void fsl_lpspi_read_rx_fifo(struct fsl_lpspi_data *fsl_lpspi)
 		fsl_lpspi->rx(fsl_lpspi);
 }
 
-static void fsl_lpspi_set_cmd(struct fsl_lpspi_data *fsl_lpspi,
-			      bool is_first_xfer)
+static void fsl_lpspi_set_cmd(struct fsl_lpspi_data *fsl_lpspi)
 {
 	u32 temp = 0;
 
@@ -213,11 +288,13 @@ static void fsl_lpspi_set_cmd(struct fsl_lpspi_data *fsl_lpspi,
 		 * For the first transfer, clear TCR_CONTC to assert SS.
 		 * For subsequent transfer, set TCR_CONTC to keep SS asserted.
 		 */
-		temp |= TCR_CONT;
-		if (is_first_xfer)
-			temp &= ~TCR_CONTC;
-		else
-			temp |= TCR_CONTC;
+		if (!fsl_lpspi->usedma) {
+			temp |= TCR_CONT;
+			if (fsl_lpspi->is_first_byte)
+				temp &= ~TCR_CONTC;
+			else
+				temp |= TCR_CONTC;
+		}
 	}
 	writel(temp, fsl_lpspi->base + IMX7ULP_TCR);
 
@@ -228,7 +305,11 @@ static void fsl_lpspi_set_watermark(struct fsl_lpspi_data *fsl_lpspi)
 {
 	u32 temp;
 
-	temp = fsl_lpspi->watermark >> 1 | (fsl_lpspi->watermark >> 1) << 16;
+	if (!fsl_lpspi->usedma)
+		temp = fsl_lpspi->watermark >> 1 |
+		       (fsl_lpspi->watermark >> 1) << 16;
+	else
+		temp = fsl_lpspi->watermark >> 1;
 
 	writel(temp, fsl_lpspi->base + IMX7ULP_FCR);
 
@@ -241,7 +322,14 @@ static int fsl_lpspi_set_bitrate(struct fsl_lpspi_data *fsl_lpspi)
 	unsigned int perclk_rate, scldiv;
 	u8 prescale;
 
-	perclk_rate = clk_get_rate(fsl_lpspi->clk);
+	perclk_rate = clk_get_rate(fsl_lpspi->clk_per);
+
+	if (config.speed_hz > perclk_rate / 2) {
+		dev_err(fsl_lpspi->dev,
+		      "per-clk should be at least two times of transfer speed");
+		return -EINVAL;
+	}
+
 	for (prescale = 0; prescale < 8; prescale++) {
 		scldiv = perclk_rate /
 			 (clkdivs[prescale] * config.speed_hz) - 2;
@@ -257,12 +345,59 @@ static int fsl_lpspi_set_bitrate(struct fsl_lpspi_data *fsl_lpspi)
 	writel(scldiv | (scldiv << 8) | ((scldiv >> 1) << 16),
 					fsl_lpspi->base + IMX7ULP_CCR);
 
-	dev_dbg(fsl_lpspi->dev, "perclk=%d, speed=%d, prescale =%d, scldiv=%d\n",
+	dev_dbg(fsl_lpspi->dev, "perclk=%d, speed=%d, prescale=%d, scldiv=%d\n",
 		perclk_rate, config.speed_hz, prescale, scldiv);
 
 	return 0;
 }
 
+static int fsl_lpspi_dma_configure(struct spi_controller *controller)
+{
+	int ret;
+	enum dma_slave_buswidth buswidth;
+	struct dma_slave_config rx = {}, tx = {};
+	struct fsl_lpspi_data *fsl_lpspi =
+				spi_controller_get_devdata(controller);
+
+	switch (fsl_lpspi_bytes_per_word(fsl_lpspi->config.bpw)) {
+	case 4:
+		buswidth = DMA_SLAVE_BUSWIDTH_4_BYTES;
+		break;
+	case 2:
+		buswidth = DMA_SLAVE_BUSWIDTH_2_BYTES;
+		break;
+	case 1:
+		buswidth = DMA_SLAVE_BUSWIDTH_1_BYTE;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	tx.direction = DMA_MEM_TO_DEV;
+	tx.dst_addr = fsl_lpspi->base_phys + IMX7ULP_TDR;
+	tx.dst_addr_width = buswidth;
+	tx.dst_maxburst = 1;
+	ret = dmaengine_slave_config(controller->dma_tx, &tx);
+	if (ret) {
+		dev_err(fsl_lpspi->dev, "TX dma configuration failed with %d\n",
+			ret);
+		return ret;
+	}
+
+	rx.direction = DMA_DEV_TO_MEM;
+	rx.src_addr = fsl_lpspi->base_phys + IMX7ULP_RDR;
+	rx.src_addr_width = buswidth;
+	rx.src_maxburst = 1;
+	ret = dmaengine_slave_config(controller->dma_rx, &rx);
+	if (ret) {
+		dev_err(fsl_lpspi->dev, "RX dma configuration failed with %d\n",
+			ret);
+		return ret;
+	}
+
+	return 0;
+}
+
 static int fsl_lpspi_config(struct fsl_lpspi_data *fsl_lpspi)
 {
 	u32 temp;
@@ -288,18 +423,27 @@ static int fsl_lpspi_config(struct fsl_lpspi_data *fsl_lpspi)
 	temp |= CR_RRF | CR_RTF | CR_MEN;
 	writel(temp, fsl_lpspi->base + IMX7ULP_CR);
 
+	temp = 0;
+	if (fsl_lpspi->usedma)
+		temp = DER_TDDE | DER_RDDE;
+	writel(temp, fsl_lpspi->base + IMX7ULP_DER);
+
 	return 0;
 }
 
-static void fsl_lpspi_setup_transfer(struct spi_device *spi,
+static int fsl_lpspi_setup_transfer(struct spi_controller *controller,
+				     struct spi_device *spi,
 				     struct spi_transfer *t)
 {
 	struct fsl_lpspi_data *fsl_lpspi =
 				spi_controller_get_devdata(spi->controller);
 
+	if (t == NULL)
+		return -EINVAL;
+
 	fsl_lpspi->config.mode = spi->mode;
-	fsl_lpspi->config.bpw = t ? t->bits_per_word : spi->bits_per_word;
-	fsl_lpspi->config.speed_hz = t ? t->speed_hz : spi->max_speed_hz;
+	fsl_lpspi->config.bpw = t->bits_per_word;
+	fsl_lpspi->config.speed_hz = t->speed_hz;
 	fsl_lpspi->config.chip_select = spi->chip_select;
 
 	if (!fsl_lpspi->config.speed_hz)
@@ -324,7 +468,12 @@ static void fsl_lpspi_setup_transfer(struct spi_device *spi,
 	else
 		fsl_lpspi->watermark = fsl_lpspi->txfifosize;
 
-	fsl_lpspi_config(fsl_lpspi);
+	if (fsl_lpspi_can_dma(controller, spi, t))
+		fsl_lpspi->usedma = 1;
+	else
+		fsl_lpspi->usedma = 0;
+
+	return fsl_lpspi_config(fsl_lpspi);
 }
 
 static int fsl_lpspi_slave_abort(struct spi_controller *controller)
@@ -333,7 +482,13 @@ static int fsl_lpspi_slave_abort(struct spi_controller *controller)
 				spi_controller_get_devdata(controller);
 
 	fsl_lpspi->slave_aborted = true;
-	complete(&fsl_lpspi->xfer_done);
+	if (!fsl_lpspi->usedma)
+		complete(&fsl_lpspi->xfer_done);
+	else {
+		complete(&fsl_lpspi->dma_tx_completion);
+		complete(&fsl_lpspi->dma_rx_completion);
+	}
+
 	return 0;
 }
 
@@ -362,8 +517,10 @@ static int fsl_lpspi_reset(struct fsl_lpspi_data *fsl_lpspi)
 {
 	u32 temp;
 
-	/* Disable all interrupt */
-	fsl_lpspi_intctrl(fsl_lpspi, 0);
+	if (!fsl_lpspi->usedma) {
+		/* Disable all interrupt */
+		fsl_lpspi_intctrl(fsl_lpspi, 0);
+	}
 
 	/* W1C for all flags in SR */
 	temp = 0x3F << 8;
@@ -376,8 +533,177 @@ static int fsl_lpspi_reset(struct fsl_lpspi_data *fsl_lpspi)
 	return 0;
 }
 
-static int fsl_lpspi_transfer_one(struct spi_controller *controller,
-				  struct spi_device *spi,
+static void fsl_lpspi_dma_rx_callback(void *cookie)
+{
+	struct fsl_lpspi_data *fsl_lpspi = (struct fsl_lpspi_data *)cookie;
+
+	complete(&fsl_lpspi->dma_rx_completion);
+}
+
+static void fsl_lpspi_dma_tx_callback(void *cookie)
+{
+	struct fsl_lpspi_data *fsl_lpspi = (struct fsl_lpspi_data *)cookie;
+
+	complete(&fsl_lpspi->dma_tx_completion);
+}
+
+static int fsl_lpspi_calculate_timeout(struct fsl_lpspi_data *fsl_lpspi,
+				       int size)
+{
+	unsigned long timeout = 0;
+
+	/* Time with actual data transfer and CS change delay related to HW */
+	timeout = (8 + 4) * size / fsl_lpspi->config.speed_hz;
+
+	/* Add extra second for scheduler related activities */
+	timeout += 1;
+
+	/* Double calculated timeout */
+	return msecs_to_jiffies(2 * timeout * MSEC_PER_SEC);
+}
+
+static int fsl_lpspi_dma_transfer(struct spi_controller *controller,
+				struct fsl_lpspi_data *fsl_lpspi,
+				struct spi_transfer *transfer)
+{
+	struct dma_async_tx_descriptor *desc_tx, *desc_rx;
+	unsigned long transfer_timeout;
+	unsigned long timeout;
+	struct sg_table *tx = &transfer->tx_sg, *rx = &transfer->rx_sg;
+	int ret;
+
+	ret = fsl_lpspi_dma_configure(controller);
+	if (ret)
+		return ret;
+
+	desc_rx = dmaengine_prep_slave_sg(controller->dma_rx,
+				rx->sgl, rx->nents, DMA_DEV_TO_MEM,
+				DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc_rx)
+		return -EINVAL;
+
+	desc_rx->callback = fsl_lpspi_dma_rx_callback;
+	desc_rx->callback_param = (void *)fsl_lpspi;
+	dmaengine_submit(desc_rx);
+	reinit_completion(&fsl_lpspi->dma_rx_completion);
+	dma_async_issue_pending(controller->dma_rx);
+
+	desc_tx = dmaengine_prep_slave_sg(controller->dma_tx,
+				tx->sgl, tx->nents, DMA_MEM_TO_DEV,
+				DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc_tx) {
+		dmaengine_terminate_all(controller->dma_tx);
+		return -EINVAL;
+	}
+
+	desc_tx->callback = fsl_lpspi_dma_tx_callback;
+	desc_tx->callback_param = (void *)fsl_lpspi;
+	dmaengine_submit(desc_tx);
+	reinit_completion(&fsl_lpspi->dma_tx_completion);
+	dma_async_issue_pending(controller->dma_tx);
+
+	fsl_lpspi->slave_aborted = false;
+
+	if (!fsl_lpspi->is_slave) {
+		transfer_timeout = fsl_lpspi_calculate_timeout(fsl_lpspi,
+							       transfer->len);
+
+		/* Wait eDMA to finish the data transfer.*/
+		timeout = wait_for_completion_timeout(&fsl_lpspi->dma_tx_completion,
+						      transfer_timeout);
+		if (!timeout) {
+			dev_err(fsl_lpspi->dev, "I/O Error in DMA TX\n");
+			dmaengine_terminate_all(controller->dma_tx);
+			dmaengine_terminate_all(controller->dma_rx);
+			fsl_lpspi_reset(fsl_lpspi);
+			return -ETIMEDOUT;
+		}
+
+		timeout = wait_for_completion_timeout(&fsl_lpspi->dma_rx_completion,
+						      transfer_timeout);
+		if (!timeout) {
+			dev_err(fsl_lpspi->dev, "I/O Error in DMA RX\n");
+			dmaengine_terminate_all(controller->dma_tx);
+			dmaengine_terminate_all(controller->dma_rx);
+			fsl_lpspi_reset(fsl_lpspi);
+			return -ETIMEDOUT;
+		}
+	} else {
+		if (wait_for_completion_interruptible(&fsl_lpspi->dma_tx_completion) ||
+			fsl_lpspi->slave_aborted) {
+			dev_dbg(fsl_lpspi->dev,
+				"I/O Error in DMA TX interrupted\n");
+			dmaengine_terminate_all(controller->dma_tx);
+			dmaengine_terminate_all(controller->dma_rx);
+			fsl_lpspi_reset(fsl_lpspi);
+			return -EINTR;
+		}
+
+		if (wait_for_completion_interruptible(&fsl_lpspi->dma_rx_completion) ||
+			fsl_lpspi->slave_aborted) {
+			dev_dbg(fsl_lpspi->dev,
+				"I/O Error in DMA RX interrupted\n");
+			dmaengine_terminate_all(controller->dma_tx);
+			dmaengine_terminate_all(controller->dma_rx);
+			fsl_lpspi_reset(fsl_lpspi);
+			return -EINTR;
+		}
+	}
+
+	fsl_lpspi_reset(fsl_lpspi);
+
+	return 0;
+}
+
+static void fsl_lpspi_dma_exit(struct spi_controller *controller)
+{
+	if (controller->dma_rx) {
+		dma_release_channel(controller->dma_rx);
+		controller->dma_rx = NULL;
+	}
+
+	if (controller->dma_tx) {
+		dma_release_channel(controller->dma_tx);
+		controller->dma_tx = NULL;
+	}
+}
+
+static int fsl_lpspi_dma_init(struct device *dev,
+			      struct fsl_lpspi_data *fsl_lpspi,
+			      struct spi_controller *controller)
+{
+	int ret;
+
+	/* Prepare for TX DMA: */
+	controller->dma_tx = dma_request_slave_channel_reason(dev, "tx");
+	if (IS_ERR(controller->dma_tx)) {
+		ret = PTR_ERR(controller->dma_tx);
+		dev_dbg(dev, "can't get the TX DMA channel, error %d!\n", ret);
+		controller->dma_tx = NULL;
+		goto err;
+	}
+
+	/* Prepare for RX DMA: */
+	controller->dma_rx = dma_request_slave_channel_reason(dev, "rx");
+	if (IS_ERR(controller->dma_rx)) {
+		ret = PTR_ERR(controller->dma_rx);
+		dev_dbg(dev, "can't get the RX DMA channel, error %d\n", ret);
+		controller->dma_rx = NULL;
+		goto err;
+	}
+
+	init_completion(&fsl_lpspi->dma_rx_completion);
+	init_completion(&fsl_lpspi->dma_tx_completion);
+	controller->can_dma = fsl_lpspi_can_dma;
+	controller->max_dma_len = FSL_LPSPI_MAX_EDMA_BYTES;
+
+	return 0;
+err:
+	fsl_lpspi_dma_exit(controller);
+	return ret;
+}
+
+static int fsl_lpspi_pio_transfer(struct spi_controller *controller,
 				  struct spi_transfer *t)
 {
 	struct fsl_lpspi_data *fsl_lpspi =
@@ -402,37 +728,30 @@ static int fsl_lpspi_transfer_one(struct spi_controller *controller,
 	return 0;
 }
 
-static int fsl_lpspi_transfer_one_msg(struct spi_controller *controller,
-				      struct spi_message *msg)
+static int fsl_lpspi_transfer_one(struct spi_controller *controller,
+				  struct spi_device *spi,
+				  struct spi_transfer *t)
 {
 	struct fsl_lpspi_data *fsl_lpspi =
-				spi_controller_get_devdata(controller);
-	struct spi_device *spi = msg->spi;
-	struct spi_transfer *xfer;
-	bool is_first_xfer = true;
-	int ret = 0;
-
-	msg->status = 0;
-	msg->actual_length = 0;
-
-	list_for_each_entry(xfer, &msg->transfers, transfer_list) {
-		fsl_lpspi_setup_transfer(spi, xfer);
-		fsl_lpspi_set_cmd(fsl_lpspi, is_first_xfer);
-
-		is_first_xfer = false;
+					spi_controller_get_devdata(controller);
+	int ret;
 
-		ret = fsl_lpspi_transfer_one(controller, spi, xfer);
-		if (ret < 0)
-			goto complete;
+	fsl_lpspi->is_first_byte = true;
+	ret = fsl_lpspi_setup_transfer(controller, spi, t);
+	if (ret < 0)
+		return ret;
 
-		msg->actual_length += xfer->len;
-	}
+	fsl_lpspi_set_cmd(fsl_lpspi);
+	fsl_lpspi->is_first_byte = false;
 
-complete:
-	msg->status = ret;
-	spi_finalize_current_message(controller);
+	if (fsl_lpspi->usedma)
+		ret = fsl_lpspi_dma_transfer(controller, fsl_lpspi, t);
+	else
+		ret = fsl_lpspi_pio_transfer(controller, t);
+	if (ret < 0)
+		return ret;
 
-	return ret;
+	return 0;
 }
 
 static irqreturn_t fsl_lpspi_isr(int irq, void *dev_id)
@@ -452,7 +771,7 @@ static irqreturn_t fsl_lpspi_isr(int irq, void *dev_id)
 	}
 
 	if (temp_SR & SR_MBF ||
-	    readl(fsl_lpspi->base + IMX7ULP_FSR) & FSR_RXCOUNT) {
+	    readl(fsl_lpspi->base + IMX7ULP_FSR) & FSR_TXCOUNT) {
 		writel(SR_FCF, fsl_lpspi->base + IMX7ULP_SR);
 		fsl_lpspi_intctrl(fsl_lpspi, IER_FCIE);
 		return IRQ_HANDLED;
@@ -467,15 +786,67 @@ static irqreturn_t fsl_lpspi_isr(int irq, void *dev_id)
 	return IRQ_NONE;
 }
 
+#ifdef CONFIG_PM
+static int fsl_lpspi_runtime_resume(struct device *dev)
+{
+	struct spi_controller *controller = dev_get_drvdata(dev);
+	struct fsl_lpspi_data *fsl_lpspi;
+	int ret;
+
+	fsl_lpspi = spi_controller_get_devdata(controller);
+
+	ret = clk_prepare_enable(fsl_lpspi->clk_per);
+	if (ret)
+		return ret;
+
+	ret = clk_prepare_enable(fsl_lpspi->clk_ipg);
+	if (ret) {
+		clk_disable_unprepare(fsl_lpspi->clk_per);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int fsl_lpspi_runtime_suspend(struct device *dev)
+{
+	struct spi_controller *controller = dev_get_drvdata(dev);
+	struct fsl_lpspi_data *fsl_lpspi;
+
+	fsl_lpspi = spi_controller_get_devdata(controller);
+
+	clk_disable_unprepare(fsl_lpspi->clk_per);
+	clk_disable_unprepare(fsl_lpspi->clk_ipg);
+
+	return 0;
+}
+#endif
+
+static int fsl_lpspi_init_rpm(struct fsl_lpspi_data *fsl_lpspi)
+{
+	struct device *dev = fsl_lpspi->dev;
+
+	pm_runtime_enable(dev);
+	pm_runtime_set_autosuspend_delay(dev, FSL_LPSPI_RPM_TIMEOUT);
+	pm_runtime_use_autosuspend(dev);
+
+	return 0;
+}
+
 static int fsl_lpspi_probe(struct platform_device *pdev)
 {
+	struct device_node *np = pdev->dev.of_node;
 	struct fsl_lpspi_data *fsl_lpspi;
 	struct spi_controller *controller;
+	struct spi_imx_master *lpspi_platform_info =
+		dev_get_platdata(&pdev->dev);
 	struct resource *res;
-	int ret, irq;
+	int i, ret, irq;
 	u32 temp;
+	bool is_slave;
 
-	if (of_property_read_bool((&pdev->dev)->of_node, "spi-slave"))
+	is_slave = of_property_read_bool((&pdev->dev)->of_node, "spi-slave");
+	if (is_slave)
 		controller = spi_alloc_slave(&pdev->dev,
 					sizeof(struct fsl_lpspi_data));
 	else
@@ -487,15 +858,35 @@ static int fsl_lpspi_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, controller);
 
-	controller->bits_per_word_mask = SPI_BPW_RANGE_MASK(8, 32);
-	controller->bus_num = pdev->id;
-
 	fsl_lpspi = spi_controller_get_devdata(controller);
 	fsl_lpspi->dev = &pdev->dev;
-	fsl_lpspi->is_slave = of_property_read_bool((&pdev->dev)->of_node,
-						    "spi-slave");
+	fsl_lpspi->is_slave = is_slave;
 
-	controller->transfer_one_message = fsl_lpspi_transfer_one_msg;
+	if (!fsl_lpspi->is_slave) {
+		for (i = 0; i < controller->num_chipselect; i++) {
+			int cs_gpio = of_get_named_gpio(np, "cs-gpios", i);
+
+			if (!gpio_is_valid(cs_gpio) && lpspi_platform_info)
+				cs_gpio = lpspi_platform_info->chipselect[i];
+
+			fsl_lpspi->chipselect[i] = cs_gpio;
+			if (!gpio_is_valid(cs_gpio))
+				continue;
+
+			ret = devm_gpio_request(&pdev->dev,
+						fsl_lpspi->chipselect[i],
+						DRIVER_NAME);
+			if (ret) {
+				dev_err(&pdev->dev, "can't get cs gpios\n");
+				goto out_controller_put;
+			}
+		}
+		controller->cs_gpios = fsl_lpspi->chipselect;
+		controller->prepare_message = fsl_lpspi_prepare_message;
+	}
+
+	controller->bits_per_word_mask = SPI_BPW_RANGE_MASK(8, 32);
+	controller->transfer_one = fsl_lpspi_transfer_one;
 	controller->prepare_transfer_hardware = lpspi_prepare_xfer_hardware;
 	controller->unprepare_transfer_hardware = lpspi_unprepare_xfer_hardware;
 	controller->mode_bits = SPI_CPOL | SPI_CPHA | SPI_CS_HIGH;
@@ -512,6 +903,7 @@ static int fsl_lpspi_probe(struct platform_device *pdev)
 		ret = PTR_ERR(fsl_lpspi->base);
 		goto out_controller_put;
 	}
+	fsl_lpspi->base_phys = res->start;
 
 	irq = platform_get_irq(pdev, 0);
 	if (irq < 0) {
@@ -526,23 +918,39 @@ static int fsl_lpspi_probe(struct platform_device *pdev)
 		goto out_controller_put;
 	}
 
-	fsl_lpspi->clk = devm_clk_get(&pdev->dev, "ipg");
-	if (IS_ERR(fsl_lpspi->clk)) {
-		ret = PTR_ERR(fsl_lpspi->clk);
+	fsl_lpspi->clk_per = devm_clk_get(&pdev->dev, "per");
+	if (IS_ERR(fsl_lpspi->clk_per)) {
+		ret = PTR_ERR(fsl_lpspi->clk_per);
 		goto out_controller_put;
 	}
 
-	ret = clk_prepare_enable(fsl_lpspi->clk);
-	if (ret) {
-		dev_err(&pdev->dev, "can't enable lpspi clock, ret=%d\n", ret);
+	fsl_lpspi->clk_ipg = devm_clk_get(&pdev->dev, "ipg");
+	if (IS_ERR(fsl_lpspi->clk_ipg)) {
+		ret = PTR_ERR(fsl_lpspi->clk_ipg);
+		goto out_controller_put;
+	}
+
+	/* enable the clock */
+	ret = fsl_lpspi_init_rpm(fsl_lpspi);
+	if (ret)
 		goto out_controller_put;
+
+	ret = pm_runtime_get_sync(fsl_lpspi->dev);
+	if (ret < 0) {
+		dev_err(fsl_lpspi->dev, "failed to enable clock\n");
+		return ret;
 	}
 
 	temp = readl(fsl_lpspi->base + IMX7ULP_PARAM);
 	fsl_lpspi->txfifosize = 1 << (temp & 0x0f);
 	fsl_lpspi->rxfifosize = 1 << ((temp >> 8) & 0x0f);
 
-	clk_disable_unprepare(fsl_lpspi->clk);
+	ret = fsl_lpspi_dma_init(&pdev->dev, fsl_lpspi, controller);
+	if (ret == -EPROBE_DEFER)
+		goto out_controller_put;
+
+	if (ret < 0)
+		dev_err(&pdev->dev, "dma setup error %d, use pio\n", ret);
 
 	ret = devm_spi_register_controller(&pdev->dev, controller);
 	if (ret < 0) {
@@ -564,15 +972,50 @@ static int fsl_lpspi_remove(struct platform_device *pdev)
 	struct fsl_lpspi_data *fsl_lpspi =
 				spi_controller_get_devdata(controller);
 
-	clk_disable_unprepare(fsl_lpspi->clk);
+	pm_runtime_disable(fsl_lpspi->dev);
+
+	spi_master_put(controller);
 
 	return 0;
 }
 
+#ifdef CONFIG_PM_SLEEP
+static int fsl_lpspi_suspend(struct device *dev)
+{
+	int ret;
+
+	pinctrl_pm_select_sleep_state(dev);
+	ret = pm_runtime_force_suspend(dev);
+	return ret;
+}
+
+static int fsl_lpspi_resume(struct device *dev)
+{
+	int ret;
+
+	ret = pm_runtime_force_resume(dev);
+	if (ret) {
+		dev_err(dev, "Error in resume: %d\n", ret);
+		return ret;
+	}
+
+	pinctrl_pm_select_default_state(dev);
+
+	return 0;
+}
+#endif /* CONFIG_PM_SLEEP */
+
+static const struct dev_pm_ops fsl_lpspi_pm_ops = {
+	SET_RUNTIME_PM_OPS(fsl_lpspi_runtime_suspend,
+				fsl_lpspi_runtime_resume, NULL)
+	SET_SYSTEM_SLEEP_PM_OPS(fsl_lpspi_suspend, fsl_lpspi_resume)
+};
+
 static struct platform_driver fsl_lpspi_driver = {
 	.driver = {
 		.name = DRIVER_NAME,
 		.of_match_table = fsl_lpspi_dt_ids,
+		.pm = &fsl_lpspi_pm_ops,
 	},
 	.probe = fsl_lpspi_probe,
 	.remove = fsl_lpspi_remove,
diff --git a/drivers/spi/spi-fsl-qspi.c b/drivers/spi/spi-fsl-qspi.c
index 6a713f78a62e..41a49b93ca60 100644
--- a/drivers/spi/spi-fsl-qspi.c
+++ b/drivers/spi/spi-fsl-qspi.c
@@ -882,7 +882,7 @@ static int fsl_qspi_probe(struct platform_device *pdev)
 
 	ctlr->dev.of_node = np;
 
-	ret = spi_register_controller(ctlr);
+	ret = devm_spi_register_controller(dev, ctlr);
 	if (ret)
 		goto err_destroy_mutex;
 
diff --git a/drivers/spi/spi-fsl-spi.c b/drivers/spi/spi-fsl-spi.c
index 8f2e97857e8b..b36ac6aa3b1f 100644
--- a/drivers/spi/spi-fsl-spi.c
+++ b/drivers/spi/spi-fsl-spi.c
@@ -39,6 +39,14 @@
 #include <linux/spi/spi_bitbang.h>
 #include <linux/types.h>
 
+#ifdef CONFIG_FSL_SOC
+#include <sysdev/fsl_soc.h>
+#endif
+
+/* Specific to the MPC8306/MPC8309 */
+#define IMMR_SPI_CS_OFFSET 0x14c
+#define SPI_BOOT_SEL_BIT   0x80000000
+
 #include "spi-fsl-lib.h"
 #include "spi-fsl-cpm.h"
 #include "spi-fsl-spi.h"
@@ -355,33 +363,50 @@ static int fsl_spi_bufs(struct spi_device *spi, struct spi_transfer *t,
 static int fsl_spi_do_one_msg(struct spi_master *master,
 			      struct spi_message *m)
 {
+	struct mpc8xxx_spi *mpc8xxx_spi = spi_master_get_devdata(master);
 	struct spi_device *spi = m->spi;
 	struct spi_transfer *t, *first;
 	unsigned int cs_change;
 	const int nsecs = 50;
-	int status;
+	int status, last_bpw;
+
+	/*
+	 * In CPU mode, optimize large byte transfers to use larger
+	 * bits_per_word values to reduce number of interrupts taken.
+	 */
+	if (!(mpc8xxx_spi->flags & SPI_CPM_MODE)) {
+		list_for_each_entry(t, &m->transfers, transfer_list) {
+			if (t->len < 256 || t->bits_per_word != 8)
+				continue;
+			if ((t->len & 3) == 0)
+				t->bits_per_word = 32;
+			else if ((t->len & 1) == 0)
+				t->bits_per_word = 16;
+		}
+	}
 
 	/* Don't allow changes if CS is active */
-	first = list_first_entry(&m->transfers, struct spi_transfer,
-			transfer_list);
+	cs_change = 1;
 	list_for_each_entry(t, &m->transfers, transfer_list) {
-		if ((first->bits_per_word != t->bits_per_word) ||
-			(first->speed_hz != t->speed_hz)) {
+		if (cs_change)
+			first = t;
+		cs_change = t->cs_change;
+		if (first->speed_hz != t->speed_hz) {
 			dev_err(&spi->dev,
-				"bits_per_word/speed_hz should be same for the same SPI transfer\n");
+				"speed_hz cannot change while CS is active\n");
 			return -EINVAL;
 		}
 	}
 
+	last_bpw = -1;
 	cs_change = 1;
 	status = -EINVAL;
 	list_for_each_entry(t, &m->transfers, transfer_list) {
-		if (t->bits_per_word || t->speed_hz) {
-			if (cs_change)
-				status = fsl_spi_setup_transfer(spi, t);
-			if (status < 0)
-				break;
-		}
+		if (cs_change || last_bpw != t->bits_per_word)
+			status = fsl_spi_setup_transfer(spi, t);
+		if (status < 0)
+			break;
+		last_bpw = t->bits_per_word;
 
 		if (cs_change) {
 			fsl_spi_chipselect(spi, BITBANG_CS_ACTIVE);
@@ -701,10 +726,17 @@ static void fsl_spi_cs_control(struct spi_device *spi, bool on)
 	struct fsl_spi_platform_data *pdata = dev_get_platdata(dev);
 	struct mpc8xxx_spi_probe_info *pinfo = to_of_pinfo(pdata);
 	u16 cs = spi->chip_select;
-	int gpio = pinfo->gpios[cs];
-	bool alow = pinfo->alow_flags[cs];
 
-	gpio_set_value(gpio, on ^ alow);
+	if (cs < pinfo->ngpios) {
+		int gpio = pinfo->gpios[cs];
+		bool alow = pinfo->alow_flags[cs];
+
+		gpio_set_value(gpio, on ^ alow);
+	} else {
+		if (WARN_ON_ONCE(cs > pinfo->ngpios || !pinfo->immr_spi_cs))
+			return;
+		iowrite32be(on ? SPI_BOOT_SEL_BIT : 0, pinfo->immr_spi_cs);
+	}
 }
 
 static int of_fsl_spi_get_chipselects(struct device *dev)
@@ -712,12 +744,15 @@ static int of_fsl_spi_get_chipselects(struct device *dev)
 	struct device_node *np = dev->of_node;
 	struct fsl_spi_platform_data *pdata = dev_get_platdata(dev);
 	struct mpc8xxx_spi_probe_info *pinfo = to_of_pinfo(pdata);
+	bool spisel_boot = IS_ENABLED(CONFIG_FSL_SOC) &&
+		of_property_read_bool(np, "fsl,spisel_boot");
 	int ngpios;
 	int i = 0;
 	int ret;
 
 	ngpios = of_gpio_count(np);
-	if (ngpios <= 0) {
+	ngpios = max(ngpios, 0);
+	if (ngpios == 0 && !spisel_boot) {
 		/*
 		 * SPI w/o chip-select line. One SPI device is still permitted
 		 * though.
@@ -726,6 +761,7 @@ static int of_fsl_spi_get_chipselects(struct device *dev)
 		return 0;
 	}
 
+	pinfo->ngpios = ngpios;
 	pinfo->gpios = kmalloc_array(ngpios, sizeof(*pinfo->gpios),
 				     GFP_KERNEL);
 	if (!pinfo->gpios)
@@ -769,7 +805,18 @@ static int of_fsl_spi_get_chipselects(struct device *dev)
 		}
 	}
 
-	pdata->max_chipselect = ngpios;
+#if IS_ENABLED(CONFIG_FSL_SOC)
+	if (spisel_boot) {
+		pinfo->immr_spi_cs = ioremap(get_immrbase() + IMMR_SPI_CS_OFFSET, 4);
+		if (!pinfo->immr_spi_cs) {
+			ret = -ENOMEM;
+			i = ngpios - 1;
+			goto err_loop;
+		}
+	}
+#endif
+
+	pdata->max_chipselect = ngpios + spisel_boot;
 	pdata->cs_control = fsl_spi_cs_control;
 
 	return 0;
diff --git a/drivers/spi/spi-gpio.c b/drivers/spi/spi-gpio.c
index 53b35c56a557..487ee55d26f7 100644
--- a/drivers/spi/spi-gpio.c
+++ b/drivers/spi/spi-gpio.c
@@ -35,20 +35,16 @@
  * platform_device->driver_data ... points to spi_gpio
  *
  * spi->controller_state ... reserved for bitbang framework code
- * spi->controller_data ... holds chipselect GPIO
  *
  * spi->master->dev.driver_data ... points to spi_gpio->bitbang
  */
 
 struct spi_gpio {
 	struct spi_bitbang		bitbang;
-	struct spi_gpio_platform_data	pdata;
-	struct platform_device		*pdev;
 	struct gpio_desc		*sck;
 	struct gpio_desc		*miso;
 	struct gpio_desc		*mosi;
 	struct gpio_desc		**cs_gpios;
-	bool				has_cs;
 };
 
 /*----------------------------------------------------------------------*/
@@ -96,12 +92,6 @@ spi_to_spi_gpio(const struct spi_device *spi)
 	return spi_gpio;
 }
 
-static inline struct spi_gpio_platform_data *__pure
-spi_to_pdata(const struct spi_device *spi)
-{
-	return &spi_to_spi_gpio(spi)->pdata;
-}
-
 /* These helpers are in turn called by the bitbang inlines */
 static inline void setsck(const struct spi_device *spi, int is_on)
 {
@@ -224,7 +214,7 @@ static void spi_gpio_chipselect(struct spi_device *spi, int is_active)
 		gpiod_set_value_cansleep(spi_gpio->sck, spi->mode & SPI_CPOL);
 
 	/* Drive chip select line, if we have one */
-	if (spi_gpio->has_cs) {
+	if (spi_gpio->cs_gpios) {
 		struct gpio_desc *cs = spi_gpio->cs_gpios[spi->chip_select];
 
 		/* SPI chip selects are normally active-low */
@@ -242,10 +232,12 @@ static int spi_gpio_setup(struct spi_device *spi)
 	 * The CS GPIOs have already been
 	 * initialized from the descriptor lookup.
 	 */
-	cs = spi_gpio->cs_gpios[spi->chip_select];
-	if (!spi->controller_state && cs)
-		status = gpiod_direction_output(cs,
-						!(spi->mode & SPI_CS_HIGH));
+	if (spi_gpio->cs_gpios) {
+		cs = spi_gpio->cs_gpios[spi->chip_select];
+		if (!spi->controller_state && cs)
+			status = gpiod_direction_output(cs,
+						  !(spi->mode & SPI_CS_HIGH));
+	}
 
 	if (!status)
 		status = spi_bitbang_setup(spi);
@@ -296,40 +288,20 @@ static void spi_gpio_cleanup(struct spi_device *spi)
  * floating signals.  (A weak pulldown would save power too, but many
  * drivers expect to see all-ones data as the no slave "response".)
  */
-static int spi_gpio_request(struct device *dev,
-			    struct spi_gpio *spi_gpio,
-			    unsigned int num_chipselects,
-			    u16 *mflags)
+static int spi_gpio_request(struct device *dev, struct spi_gpio *spi_gpio)
 {
-	int i;
-
 	spi_gpio->mosi = devm_gpiod_get_optional(dev, "mosi", GPIOD_OUT_LOW);
 	if (IS_ERR(spi_gpio->mosi))
 		return PTR_ERR(spi_gpio->mosi);
-	if (!spi_gpio->mosi)
-		/* HW configuration without MOSI pin */
-		*mflags |= SPI_MASTER_NO_TX;
 
 	spi_gpio->miso = devm_gpiod_get_optional(dev, "miso", GPIOD_IN);
 	if (IS_ERR(spi_gpio->miso))
 		return PTR_ERR(spi_gpio->miso);
-	/*
-	 * No setting SPI_MASTER_NO_RX here - if there is only a MOSI
-	 * pin connected the host can still do RX by changing the
-	 * direction of the line.
-	 */
 
 	spi_gpio->sck = devm_gpiod_get(dev, "sck", GPIOD_OUT_LOW);
 	if (IS_ERR(spi_gpio->sck))
 		return PTR_ERR(spi_gpio->sck);
 
-	for (i = 0; i < num_chipselects; i++) {
-		spi_gpio->cs_gpios[i] = devm_gpiod_get_index(dev, "cs",
-							     i, GPIOD_OUT_HIGH);
-		if (IS_ERR(spi_gpio->cs_gpios[i]))
-			return PTR_ERR(spi_gpio->cs_gpios[i]);
-	}
-
 	return 0;
 }
 
@@ -340,142 +312,134 @@ static const struct of_device_id spi_gpio_dt_ids[] = {
 };
 MODULE_DEVICE_TABLE(of, spi_gpio_dt_ids);
 
-static int spi_gpio_probe_dt(struct platform_device *pdev)
+static int spi_gpio_probe_dt(struct platform_device *pdev,
+			     struct spi_master *master)
 {
-	int ret;
-	u32 tmp;
-	struct spi_gpio_platform_data	*pdata;
-	struct device_node *np = pdev->dev.of_node;
-	const struct of_device_id *of_id =
-			of_match_device(spi_gpio_dt_ids, &pdev->dev);
-
-	if (!of_id)
-		return 0;
+	master->dev.of_node = pdev->dev.of_node;
+	master->use_gpio_descriptors = true;
 
-	pdata = devm_kzalloc(&pdev->dev, sizeof(*pdata), GFP_KERNEL);
-	if (!pdata)
-		return -ENOMEM;
+	return 0;
+}
+#else
+static inline int spi_gpio_probe_dt(struct platform_device *pdev,
+				    struct spi_master *master)
+{
+	return 0;
+}
+#endif
 
+static int spi_gpio_probe_pdata(struct platform_device *pdev,
+				struct spi_master *master)
+{
+	struct device *dev = &pdev->dev;
+	struct spi_gpio_platform_data *pdata = dev_get_platdata(dev);
+	struct spi_gpio *spi_gpio = spi_master_get_devdata(master);
+	int i;
 
-	ret = of_property_read_u32(np, "num-chipselects", &tmp);
-	if (ret < 0) {
-		dev_err(&pdev->dev, "num-chipselects property not found\n");
-		goto error_free;
-	}
+#ifdef GENERIC_BITBANG
+	if (!pdata || !pdata->num_chipselect)
+		return -ENODEV;
+#endif
+	/*
+	 * The master needs to think there is a chipselect even if not
+	 * connected
+	 */
+	master->num_chipselect = pdata->num_chipselect ?: 1;
 
-	pdata->num_chipselect = tmp;
-	pdev->dev.platform_data = pdata;
+	spi_gpio->cs_gpios = devm_kcalloc(dev, master->num_chipselect,
+					  sizeof(*spi_gpio->cs_gpios),
+					  GFP_KERNEL);
+	if (!spi_gpio->cs_gpios)
+		return -ENOMEM;
 
-	return 1;
+	for (i = 0; i < master->num_chipselect; i++) {
+		spi_gpio->cs_gpios[i] = devm_gpiod_get_index(dev, "cs", i,
+							     GPIOD_OUT_HIGH);
+		if (IS_ERR(spi_gpio->cs_gpios[i]))
+			return PTR_ERR(spi_gpio->cs_gpios[i]);
+	}
 
-error_free:
-	devm_kfree(&pdev->dev, pdata);
-	return ret;
+	return 0;
 }
-#else
-static inline int spi_gpio_probe_dt(struct platform_device *pdev)
+
+static void spi_gpio_put(void *data)
 {
-	return 0;
+	spi_master_put(data);
 }
-#endif
 
 static int spi_gpio_probe(struct platform_device *pdev)
 {
 	int				status;
 	struct spi_master		*master;
 	struct spi_gpio			*spi_gpio;
-	struct spi_gpio_platform_data	*pdata;
-	u16 master_flags = 0;
-	bool use_of = 0;
+	struct device			*dev = &pdev->dev;
+	struct spi_bitbang		*bb;
+	const struct of_device_id	*of_id;
 
-	status = spi_gpio_probe_dt(pdev);
-	if (status < 0)
-		return status;
-	if (status > 0)
-		use_of = 1;
-
-	pdata = dev_get_platdata(&pdev->dev);
-#ifdef GENERIC_BITBANG
-	if (!pdata || (!use_of && !pdata->num_chipselect))
-		return -ENODEV;
-#endif
+	of_id = of_match_device(spi_gpio_dt_ids, &pdev->dev);
 
-	master = spi_alloc_master(&pdev->dev, sizeof(*spi_gpio));
+	master = spi_alloc_master(dev, sizeof(*spi_gpio));
 	if (!master)
 		return -ENOMEM;
 
-	spi_gpio = spi_master_get_devdata(master);
-
-	spi_gpio->cs_gpios = devm_kcalloc(&pdev->dev,
-				pdata->num_chipselect,
-				sizeof(*spi_gpio->cs_gpios),
-				GFP_KERNEL);
-	if (!spi_gpio->cs_gpios)
-		return -ENOMEM;
+	status = devm_add_action_or_reset(&pdev->dev, spi_gpio_put, master);
+	if (status)
+		return status;
 
-	platform_set_drvdata(pdev, spi_gpio);
+	if (of_id)
+		status = spi_gpio_probe_dt(pdev, master);
+	else
+		status = spi_gpio_probe_pdata(pdev, master);
 
-	/* Determine if we have chip selects connected */
-	spi_gpio->has_cs = !!pdata->num_chipselect;
+	if (status)
+		return status;
 
-	spi_gpio->pdev = pdev;
-	if (pdata)
-		spi_gpio->pdata = *pdata;
+	spi_gpio = spi_master_get_devdata(master);
 
-	status = spi_gpio_request(&pdev->dev, spi_gpio,
-				  pdata->num_chipselect, &master_flags);
+	status = spi_gpio_request(dev, spi_gpio);
 	if (status)
 		return status;
 
 	master->bits_per_word_mask = SPI_BPW_RANGE_MASK(1, 32);
 	master->mode_bits = SPI_3WIRE | SPI_3WIRE_HIZ | SPI_CPHA | SPI_CPOL |
 			    SPI_CS_HIGH;
-	master->flags = master_flags;
+	if (!spi_gpio->mosi) {
+		/* HW configuration without MOSI pin
+		 *
+		 * No setting SPI_MASTER_NO_RX here - if there is only
+		 * a MOSI pin connected the host can still do RX by
+		 * changing the direction of the line.
+		 */
+		master->flags = SPI_MASTER_NO_TX;
+	}
+
 	master->bus_num = pdev->id;
-	/* The master needs to think there is a chipselect even if not connected */
-	master->num_chipselect = spi_gpio->has_cs ? pdata->num_chipselect : 1;
 	master->setup = spi_gpio_setup;
 	master->cleanup = spi_gpio_cleanup;
-#ifdef CONFIG_OF
-	master->dev.of_node = pdev->dev.of_node;
-#endif
 
-	spi_gpio->bitbang.master = master;
-	spi_gpio->bitbang.chipselect = spi_gpio_chipselect;
-	spi_gpio->bitbang.set_line_direction = spi_gpio_set_direction;
+	bb = &spi_gpio->bitbang;
+	bb->master = master;
+	bb->chipselect = spi_gpio_chipselect;
+	bb->set_line_direction = spi_gpio_set_direction;
 
-	if ((master_flags & SPI_MASTER_NO_TX) == 0) {
-		spi_gpio->bitbang.txrx_word[SPI_MODE_0] = spi_gpio_txrx_word_mode0;
-		spi_gpio->bitbang.txrx_word[SPI_MODE_1] = spi_gpio_txrx_word_mode1;
-		spi_gpio->bitbang.txrx_word[SPI_MODE_2] = spi_gpio_txrx_word_mode2;
-		spi_gpio->bitbang.txrx_word[SPI_MODE_3] = spi_gpio_txrx_word_mode3;
+	if (master->flags & SPI_MASTER_NO_TX) {
+		bb->txrx_word[SPI_MODE_0] = spi_gpio_spec_txrx_word_mode0;
+		bb->txrx_word[SPI_MODE_1] = spi_gpio_spec_txrx_word_mode1;
+		bb->txrx_word[SPI_MODE_2] = spi_gpio_spec_txrx_word_mode2;
+		bb->txrx_word[SPI_MODE_3] = spi_gpio_spec_txrx_word_mode3;
 	} else {
-		spi_gpio->bitbang.txrx_word[SPI_MODE_0] = spi_gpio_spec_txrx_word_mode0;
-		spi_gpio->bitbang.txrx_word[SPI_MODE_1] = spi_gpio_spec_txrx_word_mode1;
-		spi_gpio->bitbang.txrx_word[SPI_MODE_2] = spi_gpio_spec_txrx_word_mode2;
-		spi_gpio->bitbang.txrx_word[SPI_MODE_3] = spi_gpio_spec_txrx_word_mode3;
+		bb->txrx_word[SPI_MODE_0] = spi_gpio_txrx_word_mode0;
+		bb->txrx_word[SPI_MODE_1] = spi_gpio_txrx_word_mode1;
+		bb->txrx_word[SPI_MODE_2] = spi_gpio_txrx_word_mode2;
+		bb->txrx_word[SPI_MODE_3] = spi_gpio_txrx_word_mode3;
 	}
-	spi_gpio->bitbang.setup_transfer = spi_bitbang_setup_transfer;
+	bb->setup_transfer = spi_bitbang_setup_transfer;
 
-	status = spi_bitbang_start(&spi_gpio->bitbang);
+	status = spi_bitbang_init(&spi_gpio->bitbang);
 	if (status)
-		spi_master_put(master);
-
-	return status;
-}
-
-static int spi_gpio_remove(struct platform_device *pdev)
-{
-	struct spi_gpio			*spi_gpio;
-
-	spi_gpio = platform_get_drvdata(pdev);
-
-	/* stop() unregisters child devices too */
-	spi_bitbang_stop(&spi_gpio->bitbang);
-
-	spi_master_put(spi_gpio->bitbang.master);
+		return status;
 
-	return 0;
+	return devm_spi_register_master(&pdev->dev, spi_master_get(master));
 }
 
 MODULE_ALIAS("platform:" DRIVER_NAME);
@@ -486,7 +450,6 @@ static struct platform_driver spi_gpio_driver = {
 		.of_match_table = of_match_ptr(spi_gpio_dt_ids),
 	},
 	.probe		= spi_gpio_probe,
-	.remove		= spi_gpio_remove,
 };
 module_platform_driver(spi_gpio_driver);
 
diff --git a/drivers/spi/spi-imx.c b/drivers/spi/spi-imx.c
index 6ec647bbba77..09c9a1edb2c6 100644
--- a/drivers/spi/spi-imx.c
+++ b/drivers/spi/spi-imx.c
@@ -28,6 +28,10 @@
 
 #define DRIVER_NAME "spi_imx"
 
+static bool use_dma = true;
+module_param(use_dma, bool, 0644);
+MODULE_PARM_DESC(use_dma, "Enable usage of DMA when available (default)");
+
 #define MXC_CSPIRXDATA		0x00
 #define MXC_CSPITXDATA		0x04
 #define MXC_CSPICTRL		0x08
@@ -219,6 +223,9 @@ static bool spi_imx_can_dma(struct spi_master *master, struct spi_device *spi,
 {
 	struct spi_imx_data *spi_imx = spi_master_get_devdata(master);
 
+	if (!use_dma)
+		return false;
+
 	if (!master->dma_rx)
 		return false;
 
@@ -1494,7 +1501,7 @@ static int spi_imx_transfer(struct spi_device *spi,
 
 	/* flush rxfifo before transfer */
 	while (spi_imx->devtype_data->rx_available(spi_imx))
-		spi_imx->rx(spi_imx);
+		readl(spi_imx->base + MXC_CSPIRXDATA);
 
 	if (spi_imx->slave_mode)
 		return spi_imx_pio_transfer_slave(spi, transfer);
diff --git a/drivers/spi/spi-mem.c b/drivers/spi/spi-mem.c
index a4d8d19ecff9..9f0fa9f3116d 100644
--- a/drivers/spi/spi-mem.c
+++ b/drivers/spi/spi-mem.c
@@ -135,8 +135,8 @@ static int spi_check_buswidth_req(struct spi_mem *mem, u8 buswidth, bool tx)
 	return -ENOTSUPP;
 }
 
-static bool spi_mem_default_supports_op(struct spi_mem *mem,
-					const struct spi_mem_op *op)
+bool spi_mem_default_supports_op(struct spi_mem *mem,
+				 const struct spi_mem_op *op)
 {
 	if (spi_check_buswidth_req(mem, op->cmd.buswidth, true))
 		return false;
@@ -622,7 +622,7 @@ void devm_spi_mem_dirmap_destroy(struct device *dev,
 EXPORT_SYMBOL_GPL(devm_spi_mem_dirmap_destroy);
 
 /**
- * spi_mem_dirmap_dirmap_read() - Read data through a direct mapping
+ * spi_mem_dirmap_read() - Read data through a direct mapping
  * @desc: direct mapping descriptor
  * @offs: offset to start reading from. Note that this is not an absolute
  *	  offset, but the offset within the direct mapping which already has
@@ -668,7 +668,7 @@ ssize_t spi_mem_dirmap_read(struct spi_mem_dirmap_desc *desc,
 EXPORT_SYMBOL_GPL(spi_mem_dirmap_read);
 
 /**
- * spi_mem_dirmap_dirmap_write() - Write data through a direct mapping
+ * spi_mem_dirmap_write() - Write data through a direct mapping
  * @desc: direct mapping descriptor
  * @offs: offset to start writing from. Note that this is not an absolute
  *	  offset, but the offset within the direct mapping which already has
diff --git a/drivers/spi/spi-mt7621.c b/drivers/spi/spi-mt7621.c
new file mode 100644
index 000000000000..ae836114ee3d
--- /dev/null
+++ b/drivers/spi/spi-mt7621.c
@@ -0,0 +1,421 @@
+// SPDX-License-Identifier: GPL-2.0
+//
+// spi-mt7621.c -- MediaTek MT7621 SPI controller driver
+//
+// Copyright (C) 2011 Sergiy <piratfm@gmail.com>
+// Copyright (C) 2011-2013 Gabor Juhos <juhosg@openwrt.org>
+// Copyright (C) 2014-2015 Felix Fietkau <nbd@nbd.name>
+//
+// Some parts are based on spi-orion.c:
+//   Author: Shadi Ammouri <shadi@marvell.com>
+//   Copyright (C) 2007-2008 Marvell Ltd.
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/of_device.h>
+#include <linux/reset.h>
+#include <linux/spi/spi.h>
+
+#define DRIVER_NAME		"spi-mt7621"
+
+/* in usec */
+#define RALINK_SPI_WAIT_MAX_LOOP 2000
+
+/* SPISTAT register bit field */
+#define SPISTAT_BUSY		BIT(0)
+
+#define MT7621_SPI_TRANS	0x00
+#define SPITRANS_BUSY		BIT(16)
+
+#define MT7621_SPI_OPCODE	0x04
+#define MT7621_SPI_DATA0	0x08
+#define MT7621_SPI_DATA4	0x18
+#define SPI_CTL_TX_RX_CNT_MASK	0xff
+#define SPI_CTL_START		BIT(8)
+
+#define MT7621_SPI_MASTER	0x28
+#define MASTER_MORE_BUFMODE	BIT(2)
+#define MASTER_FULL_DUPLEX	BIT(10)
+#define MASTER_RS_CLK_SEL	GENMASK(27, 16)
+#define MASTER_RS_CLK_SEL_SHIFT	16
+#define MASTER_RS_SLAVE_SEL	GENMASK(31, 29)
+
+#define MT7621_SPI_MOREBUF	0x2c
+#define MT7621_SPI_POLAR	0x38
+#define MT7621_SPI_SPACE	0x3c
+
+#define MT7621_CPHA		BIT(5)
+#define MT7621_CPOL		BIT(4)
+#define MT7621_LSB_FIRST	BIT(3)
+
+struct mt7621_spi {
+	struct spi_controller	*master;
+	void __iomem		*base;
+	unsigned int		sys_freq;
+	unsigned int		speed;
+	struct clk		*clk;
+	int			pending_write;
+
+	struct mt7621_spi_ops	*ops;
+};
+
+static inline struct mt7621_spi *spidev_to_mt7621_spi(struct spi_device *spi)
+{
+	return spi_controller_get_devdata(spi->master);
+}
+
+static inline u32 mt7621_spi_read(struct mt7621_spi *rs, u32 reg)
+{
+	return ioread32(rs->base + reg);
+}
+
+static inline void mt7621_spi_write(struct mt7621_spi *rs, u32 reg, u32 val)
+{
+	iowrite32(val, rs->base + reg);
+}
+
+static void mt7621_spi_set_cs(struct spi_device *spi, int enable)
+{
+	struct mt7621_spi *rs = spidev_to_mt7621_spi(spi);
+	int cs = spi->chip_select;
+	u32 polar = 0;
+	u32 master;
+
+	/*
+	 * Select SPI device 7, enable "more buffer mode" and disable
+	 * full-duplex (only half-duplex really works on this chip
+	 * reliably)
+	 */
+	master = mt7621_spi_read(rs, MT7621_SPI_MASTER);
+	master |= MASTER_RS_SLAVE_SEL | MASTER_MORE_BUFMODE;
+	master &= ~MASTER_FULL_DUPLEX;
+	mt7621_spi_write(rs, MT7621_SPI_MASTER, master);
+
+	rs->pending_write = 0;
+
+	if (enable)
+		polar = BIT(cs);
+	mt7621_spi_write(rs, MT7621_SPI_POLAR, polar);
+}
+
+static int mt7621_spi_prepare(struct spi_device *spi, unsigned int speed)
+{
+	struct mt7621_spi *rs = spidev_to_mt7621_spi(spi);
+	u32 rate;
+	u32 reg;
+
+	dev_dbg(&spi->dev, "speed:%u\n", speed);
+
+	rate = DIV_ROUND_UP(rs->sys_freq, speed);
+	dev_dbg(&spi->dev, "rate-1:%u\n", rate);
+
+	if (rate > 4097)
+		return -EINVAL;
+
+	if (rate < 2)
+		rate = 2;
+
+	reg = mt7621_spi_read(rs, MT7621_SPI_MASTER);
+	reg &= ~MASTER_RS_CLK_SEL;
+	reg |= (rate - 2) << MASTER_RS_CLK_SEL_SHIFT;
+	rs->speed = speed;
+
+	reg &= ~MT7621_LSB_FIRST;
+	if (spi->mode & SPI_LSB_FIRST)
+		reg |= MT7621_LSB_FIRST;
+
+	/*
+	 * This SPI controller seems to be tested on SPI flash only and some
+	 * bits are swizzled under other SPI modes probably due to incorrect
+	 * wiring inside the silicon. Only mode 0 works correctly.
+	 */
+	reg &= ~(MT7621_CPHA | MT7621_CPOL);
+
+	mt7621_spi_write(rs, MT7621_SPI_MASTER, reg);
+
+	return 0;
+}
+
+static inline int mt7621_spi_wait_till_ready(struct mt7621_spi *rs)
+{
+	int i;
+
+	for (i = 0; i < RALINK_SPI_WAIT_MAX_LOOP; i++) {
+		u32 status;
+
+		status = mt7621_spi_read(rs, MT7621_SPI_TRANS);
+		if ((status & SPITRANS_BUSY) == 0)
+			return 0;
+		cpu_relax();
+		udelay(1);
+	}
+
+	return -ETIMEDOUT;
+}
+
+static void mt7621_spi_read_half_duplex(struct mt7621_spi *rs,
+					int rx_len, u8 *buf)
+{
+	int tx_len;
+
+	/*
+	 * Combine with any pending write, and perform one or more half-duplex
+	 * transactions reading 'len' bytes. Data to be written is already in
+	 * MT7621_SPI_DATA.
+	 */
+	tx_len = rs->pending_write;
+	rs->pending_write = 0;
+
+	while (rx_len || tx_len) {
+		int i;
+		u32 val = (min(tx_len, 4) * 8) << 24;
+		int rx = min(rx_len, 32);
+
+		if (tx_len > 4)
+			val |= (tx_len - 4) * 8;
+		val |= (rx * 8) << 12;
+		mt7621_spi_write(rs, MT7621_SPI_MOREBUF, val);
+
+		tx_len = 0;
+
+		val = mt7621_spi_read(rs, MT7621_SPI_TRANS);
+		val |= SPI_CTL_START;
+		mt7621_spi_write(rs, MT7621_SPI_TRANS, val);
+
+		mt7621_spi_wait_till_ready(rs);
+
+		for (i = 0; i < rx; i++) {
+			if ((i % 4) == 0)
+				val = mt7621_spi_read(rs, MT7621_SPI_DATA0 + i);
+			*buf++ = val & 0xff;
+			val >>= 8;
+		}
+
+		rx_len -= i;
+	}
+}
+
+static inline void mt7621_spi_flush(struct mt7621_spi *rs)
+{
+	mt7621_spi_read_half_duplex(rs, 0, NULL);
+}
+
+static void mt7621_spi_write_half_duplex(struct mt7621_spi *rs,
+					 int tx_len, const u8 *buf)
+{
+	int len = rs->pending_write;
+	int val = 0;
+
+	if (len & 3) {
+		val = mt7621_spi_read(rs, MT7621_SPI_OPCODE + (len & ~3));
+		if (len < 4) {
+			val <<= (4 - len) * 8;
+			val = swab32(val);
+		}
+	}
+
+	while (tx_len > 0) {
+		if (len >= 36) {
+			rs->pending_write = len;
+			mt7621_spi_flush(rs);
+			len = 0;
+		}
+
+		val |= *buf++ << (8 * (len & 3));
+		len++;
+		if ((len & 3) == 0) {
+			if (len == 4)
+				/* The byte-order of the opcode is weird! */
+				val = swab32(val);
+			mt7621_spi_write(rs, MT7621_SPI_OPCODE + len - 4, val);
+			val = 0;
+		}
+		tx_len -= 1;
+	}
+
+	if (len & 3) {
+		if (len < 4) {
+			val = swab32(val);
+			val >>= (4 - len) * 8;
+		}
+		mt7621_spi_write(rs, MT7621_SPI_OPCODE + (len & ~3), val);
+	}
+
+	rs->pending_write = len;
+}
+
+static int mt7621_spi_transfer_one_message(struct spi_controller *master,
+					   struct spi_message *m)
+{
+	struct mt7621_spi *rs = spi_controller_get_devdata(master);
+	struct spi_device *spi = m->spi;
+	unsigned int speed = spi->max_speed_hz;
+	struct spi_transfer *t = NULL;
+	int status = 0;
+
+	mt7621_spi_wait_till_ready(rs);
+
+	list_for_each_entry(t, &m->transfers, transfer_list)
+		if (t->speed_hz < speed)
+			speed = t->speed_hz;
+
+	if (mt7621_spi_prepare(spi, speed)) {
+		status = -EIO;
+		goto msg_done;
+	}
+
+	/* Assert CS */
+	mt7621_spi_set_cs(spi, 1);
+
+	m->actual_length = 0;
+	list_for_each_entry(t, &m->transfers, transfer_list) {
+		if ((t->rx_buf) && (t->tx_buf)) {
+			/*
+			 * This controller will shift some extra data out
+			 * of spi_opcode if (mosi_bit_cnt > 0) &&
+			 * (cmd_bit_cnt == 0). So the claimed full-duplex
+			 * support is broken since we have no way to read
+			 * the MISO value during that bit.
+			 */
+			status = -EIO;
+			goto msg_done;
+		} else if (t->rx_buf) {
+			mt7621_spi_read_half_duplex(rs, t->len, t->rx_buf);
+		} else if (t->tx_buf) {
+			mt7621_spi_write_half_duplex(rs, t->len, t->tx_buf);
+		}
+		m->actual_length += t->len;
+	}
+
+	/* Flush data and deassert CS */
+	mt7621_spi_flush(rs);
+	mt7621_spi_set_cs(spi, 0);
+
+msg_done:
+	m->status = status;
+	spi_finalize_current_message(master);
+
+	return 0;
+}
+
+static int mt7621_spi_setup(struct spi_device *spi)
+{
+	struct mt7621_spi *rs = spidev_to_mt7621_spi(spi);
+
+	if ((spi->max_speed_hz == 0) ||
+	    (spi->max_speed_hz > (rs->sys_freq / 2)))
+		spi->max_speed_hz = (rs->sys_freq / 2);
+
+	if (spi->max_speed_hz < (rs->sys_freq / 4097)) {
+		dev_err(&spi->dev, "setup: requested speed is too low %d Hz\n",
+			spi->max_speed_hz);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static const struct of_device_id mt7621_spi_match[] = {
+	{ .compatible = "ralink,mt7621-spi" },
+	{},
+};
+MODULE_DEVICE_TABLE(of, mt7621_spi_match);
+
+static int mt7621_spi_probe(struct platform_device *pdev)
+{
+	const struct of_device_id *match;
+	struct spi_controller *master;
+	struct mt7621_spi *rs;
+	void __iomem *base;
+	struct resource *r;
+	int status = 0;
+	struct clk *clk;
+	struct mt7621_spi_ops *ops;
+	int ret;
+
+	match = of_match_device(mt7621_spi_match, &pdev->dev);
+	if (!match)
+		return -EINVAL;
+	ops = (struct mt7621_spi_ops *)match->data;
+
+	r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	base = devm_ioremap_resource(&pdev->dev, r);
+	if (IS_ERR(base))
+		return PTR_ERR(base);
+
+	clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(clk)) {
+		dev_err(&pdev->dev, "unable to get SYS clock, err=%d\n",
+			status);
+		return PTR_ERR(clk);
+	}
+
+	status = clk_prepare_enable(clk);
+	if (status)
+		return status;
+
+	master = spi_alloc_master(&pdev->dev, sizeof(*rs));
+	if (!master) {
+		dev_info(&pdev->dev, "master allocation failed\n");
+		return -ENOMEM;
+	}
+
+	master->mode_bits = SPI_LSB_FIRST;
+	master->flags = SPI_CONTROLLER_HALF_DUPLEX;
+	master->setup = mt7621_spi_setup;
+	master->transfer_one_message = mt7621_spi_transfer_one_message;
+	master->bits_per_word_mask = SPI_BPW_MASK(8);
+	master->dev.of_node = pdev->dev.of_node;
+	master->num_chipselect = 2;
+
+	dev_set_drvdata(&pdev->dev, master);
+
+	rs = spi_controller_get_devdata(master);
+	rs->base = base;
+	rs->clk = clk;
+	rs->master = master;
+	rs->sys_freq = clk_get_rate(rs->clk);
+	rs->ops = ops;
+	rs->pending_write = 0;
+	dev_info(&pdev->dev, "sys_freq: %u\n", rs->sys_freq);
+
+	ret = device_reset(&pdev->dev);
+	if (ret) {
+		dev_err(&pdev->dev, "SPI reset failed!\n");
+		return ret;
+	}
+
+	return devm_spi_register_controller(&pdev->dev, master);
+}
+
+static int mt7621_spi_remove(struct platform_device *pdev)
+{
+	struct spi_controller *master;
+	struct mt7621_spi *rs;
+
+	master = dev_get_drvdata(&pdev->dev);
+	rs = spi_controller_get_devdata(master);
+
+	clk_disable_unprepare(rs->clk);
+
+	return 0;
+}
+
+MODULE_ALIAS("platform:" DRIVER_NAME);
+
+static struct platform_driver mt7621_spi_driver = {
+	.driver = {
+		.name = DRIVER_NAME,
+		.of_match_table = mt7621_spi_match,
+	},
+	.probe = mt7621_spi_probe,
+	.remove = mt7621_spi_remove,
+};
+
+module_platform_driver(mt7621_spi_driver);
+
+MODULE_DESCRIPTION("MT7621 SPI driver");
+MODULE_AUTHOR("Felix Fietkau <nbd@nbd.name>");
+MODULE_LICENSE("GPL");
diff --git a/drivers/spi/spi-mxic.c b/drivers/spi/spi-mxic.c
index e41ae6ef0f8a..f48563c09b97 100644
--- a/drivers/spi/spi-mxic.c
+++ b/drivers/spi/spi-mxic.c
@@ -492,8 +492,7 @@ static int mxic_spi_transfer_one(struct spi_master *master,
 
 static int __maybe_unused mxic_spi_runtime_suspend(struct device *dev)
 {
-	struct platform_device *pdev = to_platform_device(dev);
-	struct spi_master *master = platform_get_drvdata(pdev);
+	struct spi_master *master = dev_get_drvdata(dev);
 	struct mxic_spi *mxic = spi_master_get_devdata(master);
 
 	mxic_spi_clk_disable(mxic);
@@ -504,8 +503,7 @@ static int __maybe_unused mxic_spi_runtime_suspend(struct device *dev)
 
 static int __maybe_unused mxic_spi_runtime_resume(struct device *dev)
 {
-	struct platform_device *pdev = to_platform_device(dev);
-	struct spi_master *master = platform_get_drvdata(pdev);
+	struct spi_master *master = dev_get_drvdata(dev);
 	struct mxic_spi *mxic = spi_master_get_devdata(master);
 	int ret;
 
diff --git a/drivers/spi/spi-orion.c b/drivers/spi/spi-orion.c
index 7f280567093e..25ea4a9e0dbc 100644
--- a/drivers/spi/spi-orion.c
+++ b/drivers/spi/spi-orion.c
@@ -470,6 +470,8 @@ orion_spi_write_read(struct spi_device *spi, struct spi_transfer *xfer)
 			if (orion_spi_write_read_8bit(spi, &tx, &rx) < 0)
 				goto out;
 			count--;
+			if (xfer->word_delay_usecs)
+				udelay(xfer->word_delay_usecs);
 		} while (count);
 	} else if (word_len == 16) {
 		const u16 *tx = xfer->tx_buf;
@@ -479,6 +481,8 @@ orion_spi_write_read(struct spi_device *spi, struct spi_transfer *xfer)
 			if (orion_spi_write_read_16bit(spi, &tx, &rx) < 0)
 				goto out;
 			count -= 2;
+			if (xfer->word_delay_usecs)
+				udelay(xfer->word_delay_usecs);
 		} while (count);
 	}
 
diff --git a/drivers/spi/spi-pic32.c b/drivers/spi/spi-pic32.c
index 131849adc570..d9f374c8b709 100644
--- a/drivers/spi/spi-pic32.c
+++ b/drivers/spi/spi-pic32.c
@@ -559,7 +559,7 @@ static int pic32_spi_one_transfer(struct spi_master *master,
 		dev_err(&spi->dev, "wait error/timedout\n");
 		if (dma_issued) {
 			dmaengine_terminate_all(master->dma_rx);
-			dmaengine_terminate_all(master->dma_rx);
+			dmaengine_terminate_all(master->dma_tx);
 		}
 		ret = -ETIMEDOUT;
 	} else {
diff --git a/drivers/spi/spi-pxa2xx-dma.c b/drivers/spi/spi-pxa2xx-dma.c
index 15592598273e..e5c26c1779ab 100644
--- a/drivers/spi/spi-pxa2xx-dma.c
+++ b/drivers/spi/spi-pxa2xx-dma.c
@@ -239,13 +239,15 @@ int pxa2xx_spi_set_dma_burst_and_threshold(struct chip_data *chip,
 					   u32 *threshold)
 {
 	struct pxa2xx_spi_chip *chip_info = spi->controller_data;
+	struct driver_data *drv_data = spi_controller_get_devdata(spi->controller);
+	u32 dma_burst_size = drv_data->controller_info->dma_burst_size;
 
 	/*
 	 * If the DMA burst size is given in chip_info we use that,
 	 * otherwise we use the default. Also we use the default FIFO
 	 * thresholds for now.
 	 */
-	*burst_code = chip_info ? chip_info->dma_burst_size : 1;
+	*burst_code = chip_info ? chip_info->dma_burst_size : dma_burst_size;
 	*threshold = SSCR1_RxTresh(RX_THRESH_DFLT)
 		   | SSCR1_TxTresh(TX_THRESH_DFLT);
 
diff --git a/drivers/spi/spi-pxa2xx-pci.c b/drivers/spi/spi-pxa2xx-pci.c
index 1727fdfbac28..d456c5251b5d 100644
--- a/drivers/spi/spi-pxa2xx-pci.c
+++ b/drivers/spi/spi-pxa2xx-pci.c
@@ -5,7 +5,6 @@
  */
 #include <linux/clk-provider.h>
 #include <linux/module.h>
-#include <linux/of_device.h>
 #include <linux/pci.h>
 #include <linux/platform_device.h>
 #include <linux/spi/pxa2xx_spi.h>
@@ -35,6 +34,8 @@ struct pxa_spi_info {
 	void *tx_param;
 	void *rx_param;
 
+	int dma_burst_size;
+
 	int (*setup)(struct pci_dev *pdev, struct pxa_spi_info *c);
 };
 
@@ -133,6 +134,7 @@ static int mrfld_spi_setup(struct pci_dev *dev, struct pxa_spi_info *c)
 	rx->dma_dev = &dma_dev->dev;
 
 	c->dma_filter = lpss_dma_filter;
+	c->dma_burst_size = 8;
 	return 0;
 }
 
@@ -223,6 +225,7 @@ static int pxa2xx_spi_pci_probe(struct pci_dev *dev,
 	spi_pdata.tx_param = c->tx_param;
 	spi_pdata.rx_param = c->rx_param;
 	spi_pdata.enable_dma = c->rx_param && c->tx_param;
+	spi_pdata.dma_burst_size = c->dma_burst_size ? c->dma_burst_size : 1;
 
 	ssp = &spi_pdata.ssp;
 	ssp->phys_base = pci_resource_start(dev, 0);
diff --git a/drivers/spi/spi-pxa2xx.c b/drivers/spi/spi-pxa2xx.c
index b6ddba833d02..e59c8b27b155 100644
--- a/drivers/spi/spi-pxa2xx.c
+++ b/drivers/spi/spi-pxa2xx.c
@@ -884,10 +884,14 @@ static unsigned int ssp_get_clk_div(struct driver_data *drv_data, int rate)
 
 	rate = min_t(int, ssp_clk, rate);
 
+	/*
+	 * Calculate the divisor for the SCR (Serial Clock Rate), avoiding
+	 * that the SSP transmission rate can be greater than the device rate
+	 */
 	if (ssp->type == PXA25x_SSP || ssp->type == CE4100_SSP)
-		return (ssp_clk / (2 * rate) - 1) & 0xff;
+		return (DIV_ROUND_UP(ssp_clk, 2 * rate) - 1) & 0xff;
 	else
-		return (ssp_clk / rate - 1) & 0xfff;
+		return (DIV_ROUND_UP(ssp_clk, rate) - 1)  & 0xfff;
 }
 
 static unsigned int pxa2xx_ssp_get_clk_div(struct driver_data *drv_data,
@@ -925,7 +929,7 @@ static int pxa2xx_spi_transfer_one(struct spi_controller *controller,
 {
 	struct driver_data *drv_data = spi_controller_get_devdata(controller);
 	struct spi_message *message = controller->cur_msg;
-	struct chip_data *chip = spi_get_ctldata(message->spi);
+	struct chip_data *chip = spi_get_ctldata(spi);
 	u32 dma_thresh = chip->dma_threshold;
 	u32 dma_burst = chip->dma_burst_size;
 	u32 change_mask = pxa2xx_spi_get_ssrc1_change_mask(drv_data);
@@ -943,21 +947,21 @@ static int pxa2xx_spi_transfer_one(struct spi_controller *controller,
 		/* reject already-mapped transfers; PIO won't always work */
 		if (message->is_dma_mapped
 				|| transfer->rx_dma || transfer->tx_dma) {
-			dev_err(&drv_data->pdev->dev,
+			dev_err(&spi->dev,
 				"Mapped transfer length of %u is greater than %d\n",
 				transfer->len, MAX_DMA_LEN);
 			return -EINVAL;
 		}
 
 		/* warn ... we force this to PIO mode */
-		dev_warn_ratelimited(&message->spi->dev,
+		dev_warn_ratelimited(&spi->dev,
 				     "DMA disabled for transfer length %ld greater than %d\n",
 				     (long)transfer->len, MAX_DMA_LEN);
 	}
 
 	/* Setup the transfer state based on the type of transfer */
 	if (pxa2xx_spi_flush(drv_data) == 0) {
-		dev_err(&drv_data->pdev->dev, "Flush failed\n");
+		dev_err(&spi->dev, "Flush failed\n");
 		return -EIO;
 	}
 	drv_data->n_bytes = chip->n_bytes;
@@ -999,15 +1003,15 @@ static int pxa2xx_spi_transfer_one(struct spi_controller *controller,
 	 */
 	if (chip->enable_dma) {
 		if (pxa2xx_spi_set_dma_burst_and_threshold(chip,
-						message->spi,
+						spi,
 						bits, &dma_burst,
 						&dma_thresh))
-			dev_warn_ratelimited(&message->spi->dev,
+			dev_warn_ratelimited(&spi->dev,
 					     "DMA burst size reduced to match bits_per_word\n");
 	}
 
 	dma_mapped = controller->can_dma &&
-		     controller->can_dma(controller, message->spi, transfer) &&
+		     controller->can_dma(controller, spi, transfer) &&
 		     controller->cur_msg_mapped;
 	if (dma_mapped) {
 
@@ -1035,12 +1039,12 @@ static int pxa2xx_spi_transfer_one(struct spi_controller *controller,
 	/* NOTE:  PXA25x_SSP _could_ use external clocking ... */
 	cr0 = pxa2xx_configure_sscr0(drv_data, clk_div, bits);
 	if (!pxa25x_ssp_comp(drv_data))
-		dev_dbg(&message->spi->dev, "%u Hz actual, %s\n",
+		dev_dbg(&spi->dev, "%u Hz actual, %s\n",
 			controller->max_speed_hz
 				/ (1 + ((cr0 & SSCR0_SCR(0xfff)) >> 8)),
 			dma_mapped ? "DMA" : "PIO");
 	else
-		dev_dbg(&message->spi->dev, "%u Hz actual, %s\n",
+		dev_dbg(&spi->dev, "%u Hz actual, %s\n",
 			controller->max_speed_hz / 2
 				/ (1 + ((cr0 & SSCR0_SCR(0x0ff)) >> 8)),
 			dma_mapped ? "DMA" : "PIO");
@@ -1333,6 +1337,9 @@ static int setup(struct spi_device *spi)
 			dev_warn(&spi->dev,
 				 "in setup: DMA burst size reduced to match bits_per_word\n");
 		}
+		dev_dbg(&spi->dev,
+			"in setup: DMA burst size set to %u\n",
+			chip->dma_burst_size);
 	}
 
 	switch (drv_data->ssp_type) {
@@ -1451,6 +1458,10 @@ static const struct pci_device_id pxa2xx_spi_pci_compound_match[] = {
 	{ PCI_VDEVICE(INTEL, 0xa32a), LPSS_CNL_SSP },
 	{ PCI_VDEVICE(INTEL, 0xa32b), LPSS_CNL_SSP },
 	{ PCI_VDEVICE(INTEL, 0xa37b), LPSS_CNL_SSP },
+	/* CML-LP */
+	{ PCI_VDEVICE(INTEL, 0x02aa), LPSS_CNL_SSP },
+	{ PCI_VDEVICE(INTEL, 0x02ab), LPSS_CNL_SSP },
+	{ PCI_VDEVICE(INTEL, 0x02fb), LPSS_CNL_SSP },
 	{ },
 };
 
@@ -1564,6 +1575,7 @@ pxa2xx_spi_init_pdata(struct platform_device *pdev)
 	pdata->is_slave = of_property_read_bool(pdev->dev.of_node, "spi-slave");
 	pdata->num_chipselect = 1;
 	pdata->enable_dma = true;
+	pdata->dma_burst_size = 1;
 
 	return pdata;
 }
@@ -1692,7 +1704,7 @@ static int pxa2xx_spi_probe(struct platform_device *pdev)
 	if (platform_info->enable_dma) {
 		status = pxa2xx_spi_dma_setup(drv_data);
 		if (status) {
-			dev_dbg(dev, "no DMA channels available, using PIO\n");
+			dev_warn(dev, "no DMA channels available, using PIO\n");
 			platform_info->enable_dma = false;
 		} else {
 			controller->can_dma = pxa2xx_spi_can_dma;
@@ -1953,3 +1965,5 @@ static void __exit pxa2xx_spi_exit(void)
 	platform_driver_unregister(&driver);
 }
 module_exit(pxa2xx_spi_exit);
+
+MODULE_SOFTDEP("pre: dw_dmac");
diff --git a/drivers/spi/spi-rspi.c b/drivers/spi/spi-rspi.c
index 556870dcdf79..15f5723d9f95 100644
--- a/drivers/spi/spi-rspi.c
+++ b/drivers/spi/spi-rspi.c
@@ -271,7 +271,8 @@ static int rspi_set_config_register(struct rspi_data *rspi, int access_size)
 	/* Sets parity, interrupt mask */
 	rspi_write8(rspi, 0x00, RSPI_SPCR2);
 
-	/* Sets SPCMD */
+	/* Resets sequencer */
+	rspi_write8(rspi, 0, RSPI_SPSCR);
 	rspi->spcmd |= SPCMD_SPB_8_TO_16(access_size);
 	rspi_write16(rspi, rspi->spcmd, RSPI_SPCMD0);
 
@@ -315,7 +316,8 @@ static int rspi_rz_set_config_register(struct rspi_data *rspi, int access_size)
 	rspi_write8(rspi, 0x00, RSPI_SSLND);
 	rspi_write8(rspi, 0x00, RSPI_SPND);
 
-	/* Sets SPCMD */
+	/* Resets sequencer */
+	rspi_write8(rspi, 0, RSPI_SPSCR);
 	rspi->spcmd |= SPCMD_SPB_8_TO_16(access_size);
 	rspi_write16(rspi, rspi->spcmd, RSPI_SPCMD0);
 
@@ -366,7 +368,8 @@ static int qspi_set_config_register(struct rspi_data *rspi, int access_size)
 	/* Sets buffer to allow normal operation */
 	rspi_write8(rspi, 0x00, QSPI_SPBFCR);
 
-	/* Sets SPCMD */
+	/* Resets sequencer */
+	rspi_write8(rspi, 0, RSPI_SPSCR);
 	rspi_write16(rspi, rspi->spcmd, RSPI_SPCMD0);
 
 	/* Sets RSPI mode */
@@ -736,27 +739,22 @@ static int qspi_trigger_transfer_out_in(struct rspi_data *rspi, const u8 *tx,
 	while (len > 0) {
 		n = qspi_set_send_trigger(rspi, len);
 		qspi_set_receive_trigger(rspi, len);
-		if (n == QSPI_BUFFER_SIZE) {
-			ret = rspi_wait_for_tx_empty(rspi);
-			if (ret < 0) {
-				dev_err(&rspi->ctlr->dev, "transmit timeout\n");
-				return ret;
-			}
-			for (i = 0; i < n; i++)
-				rspi_write_data(rspi, *tx++);
+		ret = rspi_wait_for_tx_empty(rspi);
+		if (ret < 0) {
+			dev_err(&rspi->ctlr->dev, "transmit timeout\n");
+			return ret;
+		}
+		for (i = 0; i < n; i++)
+			rspi_write_data(rspi, *tx++);
 
-			ret = rspi_wait_for_rx_full(rspi);
-			if (ret < 0) {
-				dev_err(&rspi->ctlr->dev, "receive timeout\n");
-				return ret;
-			}
-			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;
+		ret = rspi_wait_for_rx_full(rspi);
+		if (ret < 0) {
+			dev_err(&rspi->ctlr->dev, "receive timeout\n");
+			return ret;
 		}
+		for (i = 0; i < n; i++)
+			*rx++ = rspi_read_data(rspi);
+
 		len -= n;
 	}
 
@@ -793,19 +791,14 @@ static int qspi_transfer_out(struct rspi_data *rspi, struct spi_transfer *xfer)
 
 	while (n > 0) {
 		len = qspi_set_send_trigger(rspi, n);
-		if (len == QSPI_BUFFER_SIZE) {
-			ret = rspi_wait_for_tx_empty(rspi);
-			if (ret < 0) {
-				dev_err(&rspi->ctlr->dev, "transmit timeout\n");
-				return ret;
-			}
-			for (i = 0; i < len; i++)
-				rspi_write_data(rspi, *tx++);
-		} else {
-			ret = rspi_pio_transfer(rspi, tx, NULL, len);
-			if (ret < 0)
-				return ret;
+		ret = rspi_wait_for_tx_empty(rspi);
+		if (ret < 0) {
+			dev_err(&rspi->ctlr->dev, "transmit timeout\n");
+			return ret;
 		}
+		for (i = 0; i < len; i++)
+			rspi_write_data(rspi, *tx++);
+
 		n -= len;
 	}
 
@@ -830,19 +823,14 @@ static int qspi_transfer_in(struct rspi_data *rspi, struct spi_transfer *xfer)
 
 	while (n > 0) {
 		len = qspi_set_receive_trigger(rspi, n);
-		if (len == QSPI_BUFFER_SIZE) {
-			ret = rspi_wait_for_rx_full(rspi);
-			if (ret < 0) {
-				dev_err(&rspi->ctlr->dev, "receive timeout\n");
-				return ret;
-			}
-			for (i = 0; i < len; i++)
-				*rx++ = rspi_read_data(rspi);
-		} else {
-			ret = rspi_pio_transfer(rspi, NULL, rx, len);
-			if (ret < 0)
-				return ret;
+		ret = rspi_wait_for_rx_full(rspi);
+		if (ret < 0) {
+			dev_err(&rspi->ctlr->dev, "receive timeout\n");
+			return ret;
 		}
+		for (i = 0; i < len; i++)
+			*rx++ = rspi_read_data(rspi);
+
 		n -= len;
 	}
 
@@ -868,28 +856,6 @@ static int qspi_transfer_one(struct spi_controller *ctlr,
 	}
 }
 
-static int rspi_setup(struct spi_device *spi)
-{
-	struct rspi_data *rspi = spi_controller_get_devdata(spi->controller);
-
-	rspi->max_speed_hz = spi->max_speed_hz;
-
-	rspi->spcmd = SPCMD_SSLKP;
-	if (spi->mode & SPI_CPOL)
-		rspi->spcmd |= SPCMD_CPOL;
-	if (spi->mode & SPI_CPHA)
-		rspi->spcmd |= SPCMD_CPHA;
-
-	/* CMOS output mode and MOSI signal from previous transfer */
-	rspi->sppcr = 0;
-	if (spi->mode & SPI_LOOP)
-		rspi->sppcr |= SPPCR_SPLP;
-
-	set_config_register(rspi, 8);
-
-	return 0;
-}
-
 static u16 qspi_transfer_mode(const struct spi_transfer *xfer)
 {
 	if (xfer->tx_buf)
@@ -959,8 +925,24 @@ static int rspi_prepare_message(struct spi_controller *ctlr,
 				struct spi_message *msg)
 {
 	struct rspi_data *rspi = spi_controller_get_devdata(ctlr);
+	struct spi_device *spi = msg->spi;
 	int ret;
 
+	rspi->max_speed_hz = spi->max_speed_hz;
+
+	rspi->spcmd = SPCMD_SSLKP;
+	if (spi->mode & SPI_CPOL)
+		rspi->spcmd |= SPCMD_CPOL;
+	if (spi->mode & SPI_CPHA)
+		rspi->spcmd |= SPCMD_CPHA;
+
+	/* CMOS output mode and MOSI signal from previous transfer */
+	rspi->sppcr = 0;
+	if (spi->mode & SPI_LOOP)
+		rspi->sppcr |= SPPCR_SPLP;
+
+	set_config_register(rspi, 8);
+
 	if (msg->spi->mode &
 	    (SPI_TX_DUAL | SPI_TX_QUAD | SPI_RX_DUAL | SPI_RX_QUAD)) {
 		/* Setup sequencer for messages with multiple transfer modes */
@@ -1267,7 +1249,6 @@ static int rspi_probe(struct platform_device *pdev)
 	init_waitqueue_head(&rspi->wait);
 
 	ctlr->bus_num = pdev->id;
-	ctlr->setup = rspi_setup;
 	ctlr->auto_runtime_pm = true;
 	ctlr->transfer_one = ops->transfer_one;
 	ctlr->prepare_message = rspi_prepare_message;
diff --git a/drivers/spi/spi-sh-msiof.c b/drivers/spi/spi-sh-msiof.c
index e2eb466db10a..6aab7b2136db 100644
--- a/drivers/spi/spi-sh-msiof.c
+++ b/drivers/spi/spi-sh-msiof.c
@@ -18,6 +18,7 @@
 #include <linux/gpio/consumer.h>
 #include <linux/interrupt.h>
 #include <linux/io.h>
+#include <linux/iopoll.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/of.h>
@@ -82,111 +83,113 @@ struct sh_msiof_spi_priv {
 #define RFDR	0x60	/* Receive FIFO Data Register */
 
 /* TMDR1 and RMDR1 */
-#define MDR1_TRMD	 0x80000000 /* Transfer Mode (1 = Master mode) */
-#define MDR1_SYNCMD_MASK 0x30000000 /* SYNC Mode */
-#define MDR1_SYNCMD_SPI	 0x20000000 /*   Level mode/SPI */
-#define MDR1_SYNCMD_LR	 0x30000000 /*   L/R mode */
-#define MDR1_SYNCAC_SHIFT	 25 /* Sync Polarity (1 = Active-low) */
-#define MDR1_BITLSB_SHIFT	 24 /* MSB/LSB First (1 = LSB first) */
-#define MDR1_DTDL_SHIFT		 20 /* Data Pin Bit Delay for MSIOF_SYNC */
-#define MDR1_SYNCDL_SHIFT	 16 /* Frame Sync Signal Timing Delay */
-#define MDR1_FLD_MASK	 0x0000000c /* Frame Sync Signal Interval (0-3) */
-#define MDR1_FLD_SHIFT		  2
-#define MDR1_XXSTP	 0x00000001 /* Transmission/Reception Stop on FIFO */
+#define MDR1_TRMD	   BIT(31)  /* Transfer Mode (1 = Master mode) */
+#define MDR1_SYNCMD_MASK   GENMASK(29, 28) /* SYNC Mode */
+#define MDR1_SYNCMD_SPI	   (2 << 28)/*   Level mode/SPI */
+#define MDR1_SYNCMD_LR	   (3 << 28)/*   L/R mode */
+#define MDR1_SYNCAC_SHIFT  25       /* Sync Polarity (1 = Active-low) */
+#define MDR1_BITLSB_SHIFT  24       /* MSB/LSB First (1 = LSB first) */
+#define MDR1_DTDL_SHIFT	   20       /* Data Pin Bit Delay for MSIOF_SYNC */
+#define MDR1_SYNCDL_SHIFT  16       /* Frame Sync Signal Timing Delay */
+#define MDR1_FLD_MASK	   GENMASK(3, 2) /* Frame Sync Signal Interval (0-3) */
+#define MDR1_FLD_SHIFT	   2
+#define MDR1_XXSTP	   BIT(0)   /* Transmission/Reception Stop on FIFO */
 /* TMDR1 */
-#define TMDR1_PCON	 0x40000000 /* Transfer Signal Connection */
-#define TMDR1_SYNCCH_MASK 0xc000000 /* Synchronization Signal Channel Select */
-#define TMDR1_SYNCCH_SHIFT	 26 /* 0=MSIOF_SYNC, 1=MSIOF_SS1, 2=MSIOF_SS2 */
+#define TMDR1_PCON	   BIT(30)  /* Transfer Signal Connection */
+#define TMDR1_SYNCCH_MASK  GENMASK(27, 26) /* Sync Signal Channel Select */
+#define TMDR1_SYNCCH_SHIFT 26       /* 0=MSIOF_SYNC, 1=MSIOF_SS1, 2=MSIOF_SS2 */
 
 /* TMDR2 and RMDR2 */
 #define MDR2_BITLEN1(i)	(((i) - 1) << 24) /* Data Size (8-32 bits) */
 #define MDR2_WDLEN1(i)	(((i) - 1) << 16) /* Word Count (1-64/256 (SH, A1))) */
-#define MDR2_GRPMASK1	0x00000001 /* Group Output Mask 1 (SH, A1) */
+#define MDR2_GRPMASK1	BIT(0)      /* Group Output Mask 1 (SH, A1) */
 
 /* TSCR and RSCR */
-#define SCR_BRPS_MASK	    0x1f00 /* Prescaler Setting (1-32) */
+#define SCR_BRPS_MASK	GENMASK(12, 8) /* Prescaler Setting (1-32) */
 #define SCR_BRPS(i)	(((i) - 1) << 8)
-#define SCR_BRDV_MASK	    0x0007 /* Baud Rate Generator's Division Ratio */
-#define SCR_BRDV_DIV_2	    0x0000
-#define SCR_BRDV_DIV_4	    0x0001
-#define SCR_BRDV_DIV_8	    0x0002
-#define SCR_BRDV_DIV_16	    0x0003
-#define SCR_BRDV_DIV_32	    0x0004
-#define SCR_BRDV_DIV_1	    0x0007
+#define SCR_BRDV_MASK	GENMASK(2, 0) /* Baud Rate Generator's Division Ratio */
+#define SCR_BRDV_DIV_2	0
+#define SCR_BRDV_DIV_4	1
+#define SCR_BRDV_DIV_8	2
+#define SCR_BRDV_DIV_16	3
+#define SCR_BRDV_DIV_32	4
+#define SCR_BRDV_DIV_1	7
 
 /* CTR */
-#define CTR_TSCKIZ_MASK	0xc0000000 /* Transmit Clock I/O Polarity Select */
-#define CTR_TSCKIZ_SCK	0x80000000 /*   Disable SCK when TX disabled */
-#define CTR_TSCKIZ_POL_SHIFT	30 /*   Transmit Clock Polarity */
-#define CTR_RSCKIZ_MASK	0x30000000 /* Receive Clock Polarity Select */
-#define CTR_RSCKIZ_SCK	0x20000000 /*   Must match CTR_TSCKIZ_SCK */
-#define CTR_RSCKIZ_POL_SHIFT	28 /*   Receive Clock Polarity */
-#define CTR_TEDG_SHIFT		27 /* Transmit Timing (1 = falling edge) */
-#define CTR_REDG_SHIFT		26 /* Receive Timing (1 = falling edge) */
-#define CTR_TXDIZ_MASK	0x00c00000 /* Pin Output When TX is Disabled */
-#define CTR_TXDIZ_LOW	0x00000000 /*   0 */
-#define CTR_TXDIZ_HIGH	0x00400000 /*   1 */
-#define CTR_TXDIZ_HIZ	0x00800000 /*   High-impedance */
-#define CTR_TSCKE	0x00008000 /* Transmit Serial Clock Output Enable */
-#define CTR_TFSE	0x00004000 /* Transmit Frame Sync Signal Output Enable */
-#define CTR_TXE		0x00000200 /* Transmit Enable */
-#define CTR_RXE		0x00000100 /* Receive Enable */
+#define CTR_TSCKIZ_MASK	GENMASK(31, 30) /* Transmit Clock I/O Polarity Select */
+#define CTR_TSCKIZ_SCK	BIT(31)   /*   Disable SCK when TX disabled */
+#define CTR_TSCKIZ_POL_SHIFT 30   /*   Transmit Clock Polarity */
+#define CTR_RSCKIZ_MASK	GENMASK(29, 28) /* Receive Clock Polarity Select */
+#define CTR_RSCKIZ_SCK	BIT(29)   /*   Must match CTR_TSCKIZ_SCK */
+#define CTR_RSCKIZ_POL_SHIFT 28   /*   Receive Clock Polarity */
+#define CTR_TEDG_SHIFT	     27   /* Transmit Timing (1 = falling edge) */
+#define CTR_REDG_SHIFT	     26   /* Receive Timing (1 = falling edge) */
+#define CTR_TXDIZ_MASK	GENMASK(23, 22) /* Pin Output When TX is Disabled */
+#define CTR_TXDIZ_LOW	(0 << 22) /*   0 */
+#define CTR_TXDIZ_HIGH	(1 << 22) /*   1 */
+#define CTR_TXDIZ_HIZ	(2 << 22) /*   High-impedance */
+#define CTR_TSCKE	BIT(15)   /* Transmit Serial Clock Output Enable */
+#define CTR_TFSE	BIT(14)   /* Transmit Frame Sync Signal Output Enable */
+#define CTR_TXE		BIT(9)    /* Transmit Enable */
+#define CTR_RXE		BIT(8)    /* Receive Enable */
+#define CTR_TXRST	BIT(1)    /* Transmit Reset */
+#define CTR_RXRST	BIT(0)    /* Receive Reset */
 
 /* FCTR */
-#define FCTR_TFWM_MASK	0xe0000000 /* Transmit FIFO Watermark */
-#define FCTR_TFWM_64	0x00000000 /*  Transfer Request when 64 empty stages */
-#define FCTR_TFWM_32	0x20000000 /*  Transfer Request when 32 empty stages */
-#define FCTR_TFWM_24	0x40000000 /*  Transfer Request when 24 empty stages */
-#define FCTR_TFWM_16	0x60000000 /*  Transfer Request when 16 empty stages */
-#define FCTR_TFWM_12	0x80000000 /*  Transfer Request when 12 empty stages */
-#define FCTR_TFWM_8	0xa0000000 /*  Transfer Request when 8 empty stages */
-#define FCTR_TFWM_4	0xc0000000 /*  Transfer Request when 4 empty stages */
-#define FCTR_TFWM_1	0xe0000000 /*  Transfer Request when 1 empty stage */
-#define FCTR_TFUA_MASK	0x07f00000 /* Transmit FIFO Usable Area */
-#define FCTR_TFUA_SHIFT		20
+#define FCTR_TFWM_MASK	GENMASK(31, 29) /* Transmit FIFO Watermark */
+#define FCTR_TFWM_64	(0 << 29) /*  Transfer Request when 64 empty stages */
+#define FCTR_TFWM_32	(1 << 29) /*  Transfer Request when 32 empty stages */
+#define FCTR_TFWM_24	(2 << 29) /*  Transfer Request when 24 empty stages */
+#define FCTR_TFWM_16	(3 << 29) /*  Transfer Request when 16 empty stages */
+#define FCTR_TFWM_12	(4 << 29) /*  Transfer Request when 12 empty stages */
+#define FCTR_TFWM_8	(5 << 29) /*  Transfer Request when 8 empty stages */
+#define FCTR_TFWM_4	(6 << 29) /*  Transfer Request when 4 empty stages */
+#define FCTR_TFWM_1	(7 << 29) /*  Transfer Request when 1 empty stage */
+#define FCTR_TFUA_MASK	GENMASK(26, 20) /* Transmit FIFO Usable Area */
+#define FCTR_TFUA_SHIFT	20
 #define FCTR_TFUA(i)	((i) << FCTR_TFUA_SHIFT)
-#define FCTR_RFWM_MASK	0x0000e000 /* Receive FIFO Watermark */
-#define FCTR_RFWM_1	0x00000000 /*  Transfer Request when 1 valid stages */
-#define FCTR_RFWM_4	0x00002000 /*  Transfer Request when 4 valid stages */
-#define FCTR_RFWM_8	0x00004000 /*  Transfer Request when 8 valid stages */
-#define FCTR_RFWM_16	0x00006000 /*  Transfer Request when 16 valid stages */
-#define FCTR_RFWM_32	0x00008000 /*  Transfer Request when 32 valid stages */
-#define FCTR_RFWM_64	0x0000a000 /*  Transfer Request when 64 valid stages */
-#define FCTR_RFWM_128	0x0000c000 /*  Transfer Request when 128 valid stages */
-#define FCTR_RFWM_256	0x0000e000 /*  Transfer Request when 256 valid stages */
-#define FCTR_RFUA_MASK	0x00001ff0 /* Receive FIFO Usable Area (0x40 = full) */
-#define FCTR_RFUA_SHIFT		 4
+#define FCTR_RFWM_MASK	GENMASK(15, 13) /* Receive FIFO Watermark */
+#define FCTR_RFWM_1	(0 << 13) /*  Transfer Request when 1 valid stages */
+#define FCTR_RFWM_4	(1 << 13) /*  Transfer Request when 4 valid stages */
+#define FCTR_RFWM_8	(2 << 13) /*  Transfer Request when 8 valid stages */
+#define FCTR_RFWM_16	(3 << 13) /*  Transfer Request when 16 valid stages */
+#define FCTR_RFWM_32	(4 << 13) /*  Transfer Request when 32 valid stages */
+#define FCTR_RFWM_64	(5 << 13) /*  Transfer Request when 64 valid stages */
+#define FCTR_RFWM_128	(6 << 13) /*  Transfer Request when 128 valid stages */
+#define FCTR_RFWM_256	(7 << 13) /*  Transfer Request when 256 valid stages */
+#define FCTR_RFUA_MASK	GENMASK(12, 4) /* Receive FIFO Usable Area (0x40 = full) */
+#define FCTR_RFUA_SHIFT	4
 #define FCTR_RFUA(i)	((i) << FCTR_RFUA_SHIFT)
 
 /* STR */
-#define STR_TFEMP	0x20000000 /* Transmit FIFO Empty */
-#define STR_TDREQ	0x10000000 /* Transmit Data Transfer Request */
-#define STR_TEOF	0x00800000 /* Frame Transmission End */
-#define STR_TFSERR	0x00200000 /* Transmit Frame Synchronization Error */
-#define STR_TFOVF	0x00100000 /* Transmit FIFO Overflow */
-#define STR_TFUDF	0x00080000 /* Transmit FIFO Underflow */
-#define STR_RFFUL	0x00002000 /* Receive FIFO Full */
-#define STR_RDREQ	0x00001000 /* Receive Data Transfer Request */
-#define STR_REOF	0x00000080 /* Frame Reception End */
-#define STR_RFSERR	0x00000020 /* Receive Frame Synchronization Error */
-#define STR_RFUDF	0x00000010 /* Receive FIFO Underflow */
-#define STR_RFOVF	0x00000008 /* Receive FIFO Overflow */
+#define STR_TFEMP	BIT(29) /* Transmit FIFO Empty */
+#define STR_TDREQ	BIT(28) /* Transmit Data Transfer Request */
+#define STR_TEOF	BIT(23) /* Frame Transmission End */
+#define STR_TFSERR	BIT(21) /* Transmit Frame Synchronization Error */
+#define STR_TFOVF	BIT(20) /* Transmit FIFO Overflow */
+#define STR_TFUDF	BIT(19) /* Transmit FIFO Underflow */
+#define STR_RFFUL	BIT(13) /* Receive FIFO Full */
+#define STR_RDREQ	BIT(12) /* Receive Data Transfer Request */
+#define STR_REOF	BIT(7)  /* Frame Reception End */
+#define STR_RFSERR	BIT(5)  /* Receive Frame Synchronization Error */
+#define STR_RFUDF	BIT(4)  /* Receive FIFO Underflow */
+#define STR_RFOVF	BIT(3)  /* Receive FIFO Overflow */
 
 /* IER */
-#define IER_TDMAE	0x80000000 /* Transmit Data DMA Transfer Req. Enable */
-#define IER_TFEMPE	0x20000000 /* Transmit FIFO Empty Enable */
-#define IER_TDREQE	0x10000000 /* Transmit Data Transfer Request Enable */
-#define IER_TEOFE	0x00800000 /* Frame Transmission End Enable */
-#define IER_TFSERRE	0x00200000 /* Transmit Frame Sync Error Enable */
-#define IER_TFOVFE	0x00100000 /* Transmit FIFO Overflow Enable */
-#define IER_TFUDFE	0x00080000 /* Transmit FIFO Underflow Enable */
-#define IER_RDMAE	0x00008000 /* Receive Data DMA Transfer Req. Enable */
-#define IER_RFFULE	0x00002000 /* Receive FIFO Full Enable */
-#define IER_RDREQE	0x00001000 /* Receive Data Transfer Request Enable */
-#define IER_REOFE	0x00000080 /* Frame Reception End Enable */
-#define IER_RFSERRE	0x00000020 /* Receive Frame Sync Error Enable */
-#define IER_RFUDFE	0x00000010 /* Receive FIFO Underflow Enable */
-#define IER_RFOVFE	0x00000008 /* Receive FIFO Overflow Enable */
+#define IER_TDMAE	BIT(31) /* Transmit Data DMA Transfer Req. Enable */
+#define IER_TFEMPE	BIT(29) /* Transmit FIFO Empty Enable */
+#define IER_TDREQE	BIT(28) /* Transmit Data Transfer Request Enable */
+#define IER_TEOFE	BIT(23) /* Frame Transmission End Enable */
+#define IER_TFSERRE	BIT(21) /* Transmit Frame Sync Error Enable */
+#define IER_TFOVFE	BIT(20) /* Transmit FIFO Overflow Enable */
+#define IER_TFUDFE	BIT(19) /* Transmit FIFO Underflow Enable */
+#define IER_RDMAE	BIT(15) /* Receive Data DMA Transfer Req. Enable */
+#define IER_RFFULE	BIT(13) /* Receive FIFO Full Enable */
+#define IER_RDREQE	BIT(12) /* Receive Data Transfer Request Enable */
+#define IER_REOFE	BIT(7)  /* Frame Reception End Enable */
+#define IER_RFSERRE	BIT(5)  /* Receive Frame Sync Error Enable */
+#define IER_RFUDFE	BIT(4)  /* Receive FIFO Underflow Enable */
+#define IER_RFOVFE	BIT(3)  /* Receive FIFO Overflow Enable */
 
 
 static u32 sh_msiof_read(struct sh_msiof_spi_priv *p, int reg_offs)
@@ -219,21 +222,14 @@ static int sh_msiof_modify_ctr_wait(struct sh_msiof_spi_priv *p,
 {
 	u32 mask = clr | set;
 	u32 data;
-	int k;
 
 	data = sh_msiof_read(p, CTR);
 	data &= ~clr;
 	data |= set;
 	sh_msiof_write(p, CTR, data);
 
-	for (k = 100; k > 0; k--) {
-		if ((sh_msiof_read(p, CTR) & mask) == set)
-			break;
-
-		udelay(10);
-	}
-
-	return k > 0 ? 0 : -ETIMEDOUT;
+	return readl_poll_timeout_atomic(p->mapbase + CTR, data,
+					 (data & mask) == set, 10, 1000);
 }
 
 static irqreturn_t sh_msiof_spi_irq(int irq, void *data)
@@ -247,6 +243,19 @@ static irqreturn_t sh_msiof_spi_irq(int irq, void *data)
 	return IRQ_HANDLED;
 }
 
+static void sh_msiof_spi_reset_regs(struct sh_msiof_spi_priv *p)
+{
+	u32 mask = CTR_TXRST | CTR_RXRST;
+	u32 data;
+
+	data = sh_msiof_read(p, CTR);
+	data |= mask;
+	sh_msiof_write(p, CTR, data);
+
+	readl_poll_timeout_atomic(p->mapbase + CTR, data, !(data & mask), 1,
+				  100);
+}
+
 static const u32 sh_msiof_spi_div_array[] = {
 	SCR_BRDV_DIV_1, SCR_BRDV_DIV_2,	 SCR_BRDV_DIV_4,
 	SCR_BRDV_DIV_8,	SCR_BRDV_DIV_16, SCR_BRDV_DIV_32,
@@ -540,25 +549,11 @@ static void sh_msiof_spi_read_fifo_s32u(struct sh_msiof_spi_priv *p,
 
 static int sh_msiof_spi_setup(struct spi_device *spi)
 {
-	struct device_node *np = spi->controller->dev.of_node;
 	struct sh_msiof_spi_priv *p =
 		spi_controller_get_devdata(spi->controller);
 	u32 clr, set, tmp;
 
-	if (!np) {
-		/*
-		 * Use spi->controller_data for CS (same strategy as spi_gpio),
-		 * if any. otherwise let HW control CS
-		 */
-		spi->cs_gpio = (uintptr_t)spi->controller_data;
-	}
-
-	if (gpio_is_valid(spi->cs_gpio)) {
-		gpio_direction_output(spi->cs_gpio, !(spi->mode & SPI_CS_HIGH));
-		return 0;
-	}
-
-	if (spi_controller_is_slave(p->ctlr))
+	if (spi->cs_gpiod || spi_controller_is_slave(p->ctlr))
 		return 0;
 
 	if (p->native_cs_inited &&
@@ -591,7 +586,7 @@ static int sh_msiof_prepare_message(struct spi_controller *ctlr,
 	u32 ss, cs_high;
 
 	/* Configure pins before asserting CS */
-	if (gpio_is_valid(spi->cs_gpio)) {
+	if (spi->cs_gpiod) {
 		ss = p->unused_ss;
 		cs_high = p->native_cs_high;
 	} else {
@@ -926,6 +921,9 @@ static int sh_msiof_transfer_one(struct spi_controller *ctlr,
 	bool swab;
 	int ret;
 
+	/* reset registers */
+	sh_msiof_spi_reset_regs(p);
+
 	/* setup clocks (clock already enabled in chipselect()) */
 	if (!spi_controller_is_slave(p->ctlr))
 		sh_msiof_spi_set_clk_regs(p, clk_get_rate(p->clk), t->speed_hz);
@@ -1144,6 +1142,7 @@ static int sh_msiof_get_cs_gpios(struct sh_msiof_spi_priv *p)
 
 		gpiod = devm_gpiod_get_index(dev, "cs", i, GPIOD_ASIS);
 		if (!IS_ERR(gpiod)) {
+			devm_gpiod_put(dev, gpiod);
 			cs_gpios++;
 			continue;
 		}
@@ -1395,6 +1394,7 @@ static int sh_msiof_spi_probe(struct platform_device *pdev)
 	ctlr->bits_per_word_mask = chipdata->bits_per_word_mask;
 	ctlr->auto_runtime_pm = true;
 	ctlr->transfer_one = sh_msiof_transfer_one;
+	ctlr->use_gpio_descriptors = true;
 
 	ret = sh_msiof_request_dma(p);
 	if (ret < 0)
diff --git a/drivers/spi/spi-stm32-qspi.c b/drivers/spi/spi-stm32-qspi.c
index 3b2a9a6b990d..42f8e3c6aa1f 100644
--- a/drivers/spi/spi-stm32-qspi.c
+++ b/drivers/spi/spi-stm32-qspi.c
@@ -5,6 +5,8 @@
  */
 #include <linux/bitfield.h>
 #include <linux/clk.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
 #include <linux/errno.h>
 #include <linux/io.h>
 #include <linux/iopoll.h>
@@ -13,6 +15,7 @@
 #include <linux/mutex.h>
 #include <linux/of.h>
 #include <linux/of_device.h>
+#include <linux/pinctrl/consumer.h>
 #include <linux/platform_device.h>
 #include <linux/reset.h>
 #include <linux/sizes.h>
@@ -76,7 +79,6 @@
 #define QSPI_PSMAR		0x28
 #define QSPI_PIR		0x2c
 #define QSPI_LPTR		0x30
-#define LPTR_DFT_TIMEOUT	0x10
 
 #define STM32_QSPI_MAX_MMAP_SZ	SZ_256M
 #define STM32_QSPI_MAX_NORCHIP	2
@@ -84,6 +86,7 @@
 #define STM32_FIFO_TIMEOUT_US 30000
 #define STM32_BUSY_TIMEOUT_US 100000
 #define STM32_ABT_TIMEOUT_US 100000
+#define STM32_COMP_TIMEOUT_MS 1000
 
 struct stm32_qspi_flash {
 	struct stm32_qspi *qspi;
@@ -93,6 +96,8 @@ struct stm32_qspi_flash {
 
 struct stm32_qspi {
 	struct device *dev;
+	struct spi_controller *ctrl;
+	phys_addr_t phys_base;
 	void __iomem *io_base;
 	void __iomem *mm_base;
 	resource_size_t mm_size;
@@ -102,6 +107,13 @@ struct stm32_qspi {
 	struct completion data_completion;
 	u32 fmode;
 
+	struct dma_chan *dma_chtx;
+	struct dma_chan *dma_chrx;
+	struct completion dma_completion;
+
+	u32 cr_reg;
+	u32 dcr_reg;
+
 	/*
 	 * to protect device configuration, could be different between
 	 * 2 flash access (bk1, bk2)
@@ -177,6 +189,81 @@ static int stm32_qspi_tx_mm(struct stm32_qspi *qspi,
 	return 0;
 }
 
+static void stm32_qspi_dma_callback(void *arg)
+{
+	struct completion *dma_completion = arg;
+
+	complete(dma_completion);
+}
+
+static int stm32_qspi_tx_dma(struct stm32_qspi *qspi,
+			     const struct spi_mem_op *op)
+{
+	struct dma_async_tx_descriptor *desc;
+	enum dma_transfer_direction dma_dir;
+	struct dma_chan *dma_ch;
+	struct sg_table sgt;
+	dma_cookie_t cookie;
+	u32 cr, t_out;
+	int err;
+
+	if (op->data.dir == SPI_MEM_DATA_IN) {
+		dma_dir = DMA_DEV_TO_MEM;
+		dma_ch = qspi->dma_chrx;
+	} else {
+		dma_dir = DMA_MEM_TO_DEV;
+		dma_ch = qspi->dma_chtx;
+	}
+
+	/*
+	 * spi_map_buf return -EINVAL if the buffer is not DMA-able
+	 * (DMA-able: in vmalloc | kmap | virt_addr_valid)
+	 */
+	err = spi_controller_dma_map_mem_op_data(qspi->ctrl, op, &sgt);
+	if (err)
+		return err;
+
+	desc = dmaengine_prep_slave_sg(dma_ch, sgt.sgl, sgt.nents,
+				       dma_dir, DMA_PREP_INTERRUPT);
+	if (!desc) {
+		err = -ENOMEM;
+		goto out_unmap;
+	}
+
+	cr = readl_relaxed(qspi->io_base + QSPI_CR);
+
+	reinit_completion(&qspi->dma_completion);
+	desc->callback = stm32_qspi_dma_callback;
+	desc->callback_param = &qspi->dma_completion;
+	cookie = dmaengine_submit(desc);
+	err = dma_submit_error(cookie);
+	if (err)
+		goto out;
+
+	dma_async_issue_pending(dma_ch);
+
+	writel_relaxed(cr | CR_DMAEN, qspi->io_base + QSPI_CR);
+
+	t_out = sgt.nents * STM32_COMP_TIMEOUT_MS;
+	if (!wait_for_completion_interruptible_timeout(&qspi->dma_completion,
+						       msecs_to_jiffies(t_out)))
+		err = -ETIMEDOUT;
+
+	if (dma_async_is_tx_complete(dma_ch, cookie,
+				     NULL, NULL) != DMA_COMPLETE)
+		err = -ETIMEDOUT;
+
+	if (err)
+		dmaengine_terminate_all(dma_ch);
+
+out:
+	writel_relaxed(cr & ~CR_DMAEN, qspi->io_base + QSPI_CR);
+out_unmap:
+	spi_controller_dma_unmap_mem_op_data(qspi->ctrl, op, &sgt);
+
+	return err;
+}
+
 static int stm32_qspi_tx(struct stm32_qspi *qspi, const struct spi_mem_op *op)
 {
 	if (!op->data.nbytes)
@@ -184,6 +271,10 @@ static int stm32_qspi_tx(struct stm32_qspi *qspi, const struct spi_mem_op *op)
 
 	if (qspi->fmode == CCR_FMODE_MM)
 		return stm32_qspi_tx_mm(qspi, op);
+	else if ((op->data.dir == SPI_MEM_DATA_IN && qspi->dma_chrx) ||
+		 (op->data.dir == SPI_MEM_DATA_OUT && qspi->dma_chtx))
+		if (!stm32_qspi_tx_dma(qspi, op))
+			return 0;
 
 	return stm32_qspi_tx_poll(qspi, op);
 }
@@ -214,7 +305,7 @@ static int stm32_qspi_wait_cmd(struct stm32_qspi *qspi,
 	writel_relaxed(cr | CR_TCIE | CR_TEIE, qspi->io_base + QSPI_CR);
 
 	if (!wait_for_completion_interruptible_timeout(&qspi->data_completion,
-						msecs_to_jiffies(1000))) {
+				msecs_to_jiffies(STM32_COMP_TIMEOUT_MS))) {
 		err = -ETIMEDOUT;
 	} else {
 		sr = readl_relaxed(qspi->io_base + QSPI_SR);
@@ -356,7 +447,7 @@ static int stm32_qspi_setup(struct spi_device *spi)
 	struct spi_controller *ctrl = spi->master;
 	struct stm32_qspi *qspi = spi_controller_get_devdata(ctrl);
 	struct stm32_qspi_flash *flash;
-	u32 cr, presc;
+	u32 presc;
 
 	if (ctrl->busy)
 		return -EBUSY;
@@ -372,17 +463,60 @@ static int stm32_qspi_setup(struct spi_device *spi)
 	flash->presc = presc;
 
 	mutex_lock(&qspi->lock);
-	writel_relaxed(LPTR_DFT_TIMEOUT, qspi->io_base + QSPI_LPTR);
-	cr = FIELD_PREP(CR_FTHRES_MASK, 3) | CR_TCEN | CR_SSHIFT | CR_EN;
-	writel_relaxed(cr, qspi->io_base + QSPI_CR);
+	qspi->cr_reg = FIELD_PREP(CR_FTHRES_MASK, 3) | CR_SSHIFT | CR_EN;
+	writel_relaxed(qspi->cr_reg, qspi->io_base + QSPI_CR);
 
 	/* set dcr fsize to max address */
-	writel_relaxed(DCR_FSIZE_MASK, qspi->io_base + QSPI_DCR);
+	qspi->dcr_reg = DCR_FSIZE_MASK;
+	writel_relaxed(qspi->dcr_reg, qspi->io_base + QSPI_DCR);
 	mutex_unlock(&qspi->lock);
 
 	return 0;
 }
 
+static void stm32_qspi_dma_setup(struct stm32_qspi *qspi)
+{
+	struct dma_slave_config dma_cfg;
+	struct device *dev = qspi->dev;
+
+	memset(&dma_cfg, 0, sizeof(dma_cfg));
+
+	dma_cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_1_BYTE;
+	dma_cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_1_BYTE;
+	dma_cfg.src_addr = qspi->phys_base + QSPI_DR;
+	dma_cfg.dst_addr = qspi->phys_base + QSPI_DR;
+	dma_cfg.src_maxburst = 4;
+	dma_cfg.dst_maxburst = 4;
+
+	qspi->dma_chrx = dma_request_slave_channel(dev, "rx");
+	if (qspi->dma_chrx) {
+		if (dmaengine_slave_config(qspi->dma_chrx, &dma_cfg)) {
+			dev_err(dev, "dma rx config failed\n");
+			dma_release_channel(qspi->dma_chrx);
+			qspi->dma_chrx = NULL;
+		}
+	}
+
+	qspi->dma_chtx = dma_request_slave_channel(dev, "tx");
+	if (qspi->dma_chtx) {
+		if (dmaengine_slave_config(qspi->dma_chtx, &dma_cfg)) {
+			dev_err(dev, "dma tx config failed\n");
+			dma_release_channel(qspi->dma_chtx);
+			qspi->dma_chtx = NULL;
+		}
+	}
+
+	init_completion(&qspi->dma_completion);
+}
+
+static void stm32_qspi_dma_free(struct stm32_qspi *qspi)
+{
+	if (qspi->dma_chtx)
+		dma_release_channel(qspi->dma_chtx);
+	if (qspi->dma_chrx)
+		dma_release_channel(qspi->dma_chrx);
+}
+
 /*
  * no special host constraint, so use default spi_mem_default_supports_op
  * to check supported mode.
@@ -395,8 +529,10 @@ static void stm32_qspi_release(struct stm32_qspi *qspi)
 {
 	/* disable qspi */
 	writel_relaxed(0, qspi->io_base + QSPI_CR);
+	stm32_qspi_dma_free(qspi);
 	mutex_destroy(&qspi->lock);
 	clk_disable_unprepare(qspi->clk);
+	spi_master_put(qspi->ctrl);
 }
 
 static int stm32_qspi_probe(struct platform_device *pdev)
@@ -413,43 +549,62 @@ static int stm32_qspi_probe(struct platform_device *pdev)
 		return -ENOMEM;
 
 	qspi = spi_controller_get_devdata(ctrl);
+	qspi->ctrl = ctrl;
 
 	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "qspi");
 	qspi->io_base = devm_ioremap_resource(dev, res);
-	if (IS_ERR(qspi->io_base))
-		return PTR_ERR(qspi->io_base);
+	if (IS_ERR(qspi->io_base)) {
+		ret = PTR_ERR(qspi->io_base);
+		goto err;
+	}
+
+	qspi->phys_base = res->start;
 
 	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "qspi_mm");
 	qspi->mm_base = devm_ioremap_resource(dev, res);
-	if (IS_ERR(qspi->mm_base))
-		return PTR_ERR(qspi->mm_base);
+	if (IS_ERR(qspi->mm_base)) {
+		ret = PTR_ERR(qspi->mm_base);
+		goto err;
+	}
 
 	qspi->mm_size = resource_size(res);
-	if (qspi->mm_size > STM32_QSPI_MAX_MMAP_SZ)
-		return -EINVAL;
+	if (qspi->mm_size > STM32_QSPI_MAX_MMAP_SZ) {
+		ret = -EINVAL;
+		goto err;
+	}
 
 	irq = platform_get_irq(pdev, 0);
+	if (irq < 0) {
+		if (irq != -EPROBE_DEFER)
+			dev_err(dev, "IRQ error missing or invalid\n");
+		return irq;
+	}
+
 	ret = devm_request_irq(dev, irq, stm32_qspi_irq, 0,
 			       dev_name(dev), qspi);
 	if (ret) {
 		dev_err(dev, "failed to request irq\n");
-		return ret;
+		goto err;
 	}
 
 	init_completion(&qspi->data_completion);
 
 	qspi->clk = devm_clk_get(dev, NULL);
-	if (IS_ERR(qspi->clk))
-		return PTR_ERR(qspi->clk);
+	if (IS_ERR(qspi->clk)) {
+		ret = PTR_ERR(qspi->clk);
+		goto err;
+	}
 
 	qspi->clk_rate = clk_get_rate(qspi->clk);
-	if (!qspi->clk_rate)
-		return -EINVAL;
+	if (!qspi->clk_rate) {
+		ret = -EINVAL;
+		goto err;
+	}
 
 	ret = clk_prepare_enable(qspi->clk);
 	if (ret) {
 		dev_err(dev, "can not enable the clock\n");
-		return ret;
+		goto err;
 	}
 
 	rstc = devm_reset_control_get_exclusive(dev, NULL);
@@ -461,6 +616,7 @@ static int stm32_qspi_probe(struct platform_device *pdev)
 
 	qspi->dev = dev;
 	platform_set_drvdata(pdev, qspi);
+	stm32_qspi_dma_setup(qspi);
 	mutex_init(&qspi->lock);
 
 	ctrl->mode_bits = SPI_RX_DUAL | SPI_RX_QUAD
@@ -472,14 +628,11 @@ static int stm32_qspi_probe(struct platform_device *pdev)
 	ctrl->dev.of_node = dev->of_node;
 
 	ret = devm_spi_register_master(dev, ctrl);
-	if (ret)
-		goto err_spi_register;
-
-	return 0;
+	if (!ret)
+		return 0;
 
-err_spi_register:
+err:
 	stm32_qspi_release(qspi);
-
 	return ret;
 }
 
@@ -491,6 +644,31 @@ static int stm32_qspi_remove(struct platform_device *pdev)
 	return 0;
 }
 
+static int __maybe_unused stm32_qspi_suspend(struct device *dev)
+{
+	struct stm32_qspi *qspi = dev_get_drvdata(dev);
+
+	clk_disable_unprepare(qspi->clk);
+	pinctrl_pm_select_sleep_state(dev);
+
+	return 0;
+}
+
+static int __maybe_unused stm32_qspi_resume(struct device *dev)
+{
+	struct stm32_qspi *qspi = dev_get_drvdata(dev);
+
+	pinctrl_pm_select_default_state(dev);
+	clk_prepare_enable(qspi->clk);
+
+	writel_relaxed(qspi->cr_reg, qspi->io_base + QSPI_CR);
+	writel_relaxed(qspi->dcr_reg, qspi->io_base + QSPI_DCR);
+
+	return 0;
+}
+
+static SIMPLE_DEV_PM_OPS(stm32_qspi_pm_ops, stm32_qspi_suspend, stm32_qspi_resume);
+
 static const struct of_device_id stm32_qspi_match[] = {
 	{.compatible = "st,stm32f469-qspi"},
 	{}
@@ -503,6 +681,7 @@ static struct platform_driver stm32_qspi_driver = {
 	.driver	= {
 		.name = "stm32-qspi",
 		.of_match_table = stm32_qspi_match,
+		.pm = &stm32_qspi_pm_ops,
 	},
 };
 module_platform_driver(stm32_qspi_driver);
diff --git a/drivers/spi/spi-stm32.c b/drivers/spi/spi-stm32.c
index 4186ed20d796..b222ce8d083e 100644
--- a/drivers/spi/spi-stm32.c
+++ b/drivers/spi/spi-stm32.c
@@ -1839,8 +1839,9 @@ static int stm32_spi_probe(struct platform_device *pdev)
 
 	spi->irq = platform_get_irq(pdev, 0);
 	if (spi->irq <= 0) {
-		dev_err(&pdev->dev, "no irq: %d\n", spi->irq);
-		ret = -ENOENT;
+		ret = spi->irq;
+		if (ret != -EPROBE_DEFER)
+			dev_err(&pdev->dev, "failed to get irq: %d\n", ret);
 		goto err_master_put;
 	}
 	ret = devm_request_threaded_irq(&pdev->dev, spi->irq,
diff --git a/drivers/spi/spi-tegra114.c b/drivers/spi/spi-tegra114.c
index a76acedd7e2f..b1f31bb16659 100644
--- a/drivers/spi/spi-tegra114.c
+++ b/drivers/spi/spi-tegra114.c
@@ -149,6 +149,8 @@
 
 #define SPI_TX_FIFO				0x108
 #define SPI_RX_FIFO				0x188
+#define SPI_INTR_MASK				0x18c
+#define SPI_INTR_ALL_MASK			(0x1fUL << 25)
 #define MAX_CHIP_SELECT				4
 #define SPI_FIFO_DEPTH				64
 #define DATA_DIR_TX				(1 << 0)
@@ -161,6 +163,10 @@
 #define MAX_HOLD_CYCLES				16
 #define SPI_DEFAULT_SPEED			25000000
 
+struct tegra_spi_soc_data {
+	bool has_intr_mask_reg;
+};
+
 struct tegra_spi_data {
 	struct device				*dev;
 	struct spi_master			*master;
@@ -211,6 +217,7 @@ struct tegra_spi_data {
 	u32					*tx_dma_buf;
 	dma_addr_t				tx_dma_phys;
 	struct dma_async_tx_descriptor		*tx_dma_desc;
+	const struct tegra_spi_soc_data		*soc_data;
 };
 
 static int tegra_spi_runtime_suspend(struct device *dev);
@@ -259,7 +266,8 @@ static unsigned tegra_spi_calculate_curr_xfer_param(
 
 	tspi->bytes_per_word = DIV_ROUND_UP(bits_per_word, 8);
 
-	if (bits_per_word == 8 || bits_per_word == 16) {
+	if ((bits_per_word == 8 || bits_per_word == 16 ||
+	     bits_per_word == 32) && t->len > 3) {
 		tspi->is_packed = 1;
 		tspi->words_per_32bit = 32/bits_per_word;
 	} else {
@@ -307,10 +315,16 @@ static unsigned tegra_spi_fill_tx_fifo_from_client_txbuf(
 				x |= (u32)(*tx_buf++) << (i * 8);
 			tegra_spi_writel(tspi, x, SPI_TX_FIFO);
 		}
+
+		tspi->cur_tx_pos += written_words * tspi->bytes_per_word;
 	} else {
+		unsigned int write_bytes;
 		max_n_32bit = min(tspi->curr_dma_words,  tx_empty_count);
 		written_words = max_n_32bit;
 		nbytes = written_words * tspi->bytes_per_word;
+		if (nbytes > t->len - tspi->cur_pos)
+			nbytes = t->len - tspi->cur_pos;
+		write_bytes = nbytes;
 		for (count = 0; count < max_n_32bit; count++) {
 			u32 x = 0;
 
@@ -319,8 +333,10 @@ static unsigned tegra_spi_fill_tx_fifo_from_client_txbuf(
 				x |= (u32)(*tx_buf++) << (i * 8);
 			tegra_spi_writel(tspi, x, SPI_TX_FIFO);
 		}
+
+		tspi->cur_tx_pos += write_bytes;
 	}
-	tspi->cur_tx_pos += written_words * tspi->bytes_per_word;
+
 	return written_words;
 }
 
@@ -344,20 +360,27 @@ static unsigned int tegra_spi_read_rx_fifo_to_client_rxbuf(
 			for (i = 0; len && (i < 4); i++, len--)
 				*rx_buf++ = (x >> i*8) & 0xFF;
 		}
-		tspi->cur_rx_pos += tspi->curr_dma_words * tspi->bytes_per_word;
 		read_words += tspi->curr_dma_words;
+		tspi->cur_rx_pos += tspi->curr_dma_words * tspi->bytes_per_word;
 	} else {
 		u32 rx_mask = ((u32)1 << t->bits_per_word) - 1;
+		u8 bytes_per_word = tspi->bytes_per_word;
+		unsigned int read_bytes;
 
+		len = rx_full_count * bytes_per_word;
+		if (len > t->len - tspi->cur_pos)
+			len = t->len - tspi->cur_pos;
+		read_bytes = len;
 		for (count = 0; count < rx_full_count; count++) {
 			u32 x = tegra_spi_readl(tspi, SPI_RX_FIFO) & rx_mask;
 
-			for (i = 0; (i < tspi->bytes_per_word); i++)
+			for (i = 0; len && (i < bytes_per_word); i++, len--)
 				*rx_buf++ = (x >> (i*8)) & 0xFF;
 		}
-		tspi->cur_rx_pos += rx_full_count * tspi->bytes_per_word;
 		read_words += rx_full_count;
+		tspi->cur_rx_pos += read_bytes;
 	}
+
 	return read_words;
 }
 
@@ -372,12 +395,17 @@ static void tegra_spi_copy_client_txbuf_to_spi_txbuf(
 		unsigned len = tspi->curr_dma_words * tspi->bytes_per_word;
 
 		memcpy(tspi->tx_dma_buf, t->tx_buf + tspi->cur_pos, len);
+		tspi->cur_tx_pos += tspi->curr_dma_words * tspi->bytes_per_word;
 	} else {
 		unsigned int i;
 		unsigned int count;
 		u8 *tx_buf = (u8 *)t->tx_buf + tspi->cur_tx_pos;
 		unsigned consume = tspi->curr_dma_words * tspi->bytes_per_word;
+		unsigned int write_bytes;
 
+		if (consume > t->len - tspi->cur_pos)
+			consume = t->len - tspi->cur_pos;
+		write_bytes = consume;
 		for (count = 0; count < tspi->curr_dma_words; count++) {
 			u32 x = 0;
 
@@ -386,8 +414,9 @@ static void tegra_spi_copy_client_txbuf_to_spi_txbuf(
 				x |= (u32)(*tx_buf++) << (i * 8);
 			tspi->tx_dma_buf[count] = x;
 		}
+
+		tspi->cur_tx_pos += write_bytes;
 	}
-	tspi->cur_tx_pos += tspi->curr_dma_words * tspi->bytes_per_word;
 
 	/* Make the dma buffer to read by dma */
 	dma_sync_single_for_device(tspi->dev, tspi->tx_dma_phys,
@@ -405,20 +434,28 @@ static void tegra_spi_copy_spi_rxbuf_to_client_rxbuf(
 		unsigned len = tspi->curr_dma_words * tspi->bytes_per_word;
 
 		memcpy(t->rx_buf + tspi->cur_rx_pos, tspi->rx_dma_buf, len);
+		tspi->cur_rx_pos += tspi->curr_dma_words * tspi->bytes_per_word;
 	} else {
 		unsigned int i;
 		unsigned int count;
 		unsigned char *rx_buf = t->rx_buf + tspi->cur_rx_pos;
 		u32 rx_mask = ((u32)1 << t->bits_per_word) - 1;
+		unsigned consume = tspi->curr_dma_words * tspi->bytes_per_word;
+		unsigned int read_bytes;
 
+		if (consume > t->len - tspi->cur_pos)
+			consume = t->len - tspi->cur_pos;
+		read_bytes = consume;
 		for (count = 0; count < tspi->curr_dma_words; count++) {
 			u32 x = tspi->rx_dma_buf[count] & rx_mask;
 
-			for (i = 0; (i < tspi->bytes_per_word); i++)
+			for (i = 0; consume && (i < tspi->bytes_per_word);
+							i++, consume--)
 				*rx_buf++ = (x >> (i*8)) & 0xFF;
 		}
+
+		tspi->cur_rx_pos += read_bytes;
 	}
-	tspi->cur_rx_pos += tspi->curr_dma_words * tspi->bytes_per_word;
 
 	/* Make the dma buffer to read by dma */
 	dma_sync_single_for_device(tspi->dev, tspi->rx_dma_phys,
@@ -470,22 +507,39 @@ static int tegra_spi_start_rx_dma(struct tegra_spi_data *tspi, int len)
 	return 0;
 }
 
-static int tegra_spi_start_dma_based_transfer(
-		struct tegra_spi_data *tspi, struct spi_transfer *t)
+static int tegra_spi_flush_fifos(struct tegra_spi_data *tspi)
 {
-	u32 val;
-	unsigned int len;
-	int ret = 0;
+	unsigned long timeout = jiffies + HZ;
 	u32 status;
 
-	/* Make sure that Rx and Tx fifo are empty */
 	status = tegra_spi_readl(tspi, SPI_FIFO_STATUS);
 	if ((status & SPI_FIFO_EMPTY) != SPI_FIFO_EMPTY) {
-		dev_err(tspi->dev, "Rx/Tx fifo are not empty status 0x%08x\n",
-			(unsigned)status);
-		return -EIO;
+		status |= SPI_RX_FIFO_FLUSH | SPI_TX_FIFO_FLUSH;
+		tegra_spi_writel(tspi, status, SPI_FIFO_STATUS);
+		while ((status & SPI_FIFO_EMPTY) != SPI_FIFO_EMPTY) {
+			status = tegra_spi_readl(tspi, SPI_FIFO_STATUS);
+			if (time_after(jiffies, timeout)) {
+				dev_err(tspi->dev,
+					"timeout waiting for fifo flush\n");
+				return -EIO;
+			}
+
+			udelay(1);
+		}
 	}
 
+	return 0;
+}
+
+static int tegra_spi_start_dma_based_transfer(
+		struct tegra_spi_data *tspi, struct spi_transfer *t)
+{
+	u32 val;
+	unsigned int len;
+	int ret = 0;
+	u8 dma_burst;
+	struct dma_slave_config dma_sconfig = {0};
+
 	val = SPI_DMA_BLK_SET(tspi->curr_dma_words - 1);
 	tegra_spi_writel(tspi, val, SPI_DMA_BLK);
 
@@ -496,23 +550,40 @@ static int tegra_spi_start_dma_based_transfer(
 		len = tspi->curr_dma_words * 4;
 
 	/* Set attention level based on length of transfer */
-	if (len & 0xF)
+	if (len & 0xF) {
 		val |= SPI_TX_TRIG_1 | SPI_RX_TRIG_1;
-	else if (((len) >> 4) & 0x1)
+		dma_burst = 1;
+	} else if (((len) >> 4) & 0x1) {
 		val |= SPI_TX_TRIG_4 | SPI_RX_TRIG_4;
-	else
+		dma_burst = 4;
+	} else {
 		val |= SPI_TX_TRIG_8 | SPI_RX_TRIG_8;
+		dma_burst = 8;
+	}
 
-	if (tspi->cur_direction & DATA_DIR_TX)
-		val |= SPI_IE_TX;
+	if (!tspi->soc_data->has_intr_mask_reg) {
+		if (tspi->cur_direction & DATA_DIR_TX)
+			val |= SPI_IE_TX;
 
-	if (tspi->cur_direction & DATA_DIR_RX)
-		val |= SPI_IE_RX;
+		if (tspi->cur_direction & DATA_DIR_RX)
+			val |= SPI_IE_RX;
+	}
 
 	tegra_spi_writel(tspi, val, SPI_DMA_CTL);
 	tspi->dma_control_reg = val;
 
+	dma_sconfig.device_fc = true;
 	if (tspi->cur_direction & DATA_DIR_TX) {
+		dma_sconfig.dst_addr = tspi->phys + SPI_TX_FIFO;
+		dma_sconfig.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+		dma_sconfig.dst_maxburst = dma_burst;
+		ret = dmaengine_slave_config(tspi->tx_dma_chan, &dma_sconfig);
+		if (ret < 0) {
+			dev_err(tspi->dev,
+				"DMA slave config failed: %d\n", ret);
+			return ret;
+		}
+
 		tegra_spi_copy_client_txbuf_to_spi_txbuf(tspi, t);
 		ret = tegra_spi_start_tx_dma(tspi, len);
 		if (ret < 0) {
@@ -523,6 +594,16 @@ static int tegra_spi_start_dma_based_transfer(
 	}
 
 	if (tspi->cur_direction & DATA_DIR_RX) {
+		dma_sconfig.src_addr = tspi->phys + SPI_RX_FIFO;
+		dma_sconfig.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+		dma_sconfig.src_maxburst = dma_burst;
+		ret = dmaengine_slave_config(tspi->rx_dma_chan, &dma_sconfig);
+		if (ret < 0) {
+			dev_err(tspi->dev,
+				"DMA slave config failed: %d\n", ret);
+			return ret;
+		}
+
 		/* Make the dma buffer to read by dma */
 		dma_sync_single_for_device(tspi->dev, tspi->rx_dma_phys,
 				tspi->dma_buf_size, DMA_FROM_DEVICE);
@@ -570,8 +651,9 @@ static int tegra_spi_start_cpu_based_transfer(
 
 	tspi->is_curr_dma_xfer = false;
 
-	val |= SPI_DMA_EN;
-	tegra_spi_writel(tspi, val, SPI_DMA_CTL);
+	val = tspi->command1_reg;
+	val |= SPI_PIO;
+	tegra_spi_writel(tspi, val, SPI_COMMAND1);
 	return 0;
 }
 
@@ -582,7 +664,6 @@ static int tegra_spi_init_dma_param(struct tegra_spi_data *tspi,
 	u32 *dma_buf;
 	dma_addr_t dma_phys;
 	int ret;
-	struct dma_slave_config dma_sconfig;
 
 	dma_chan = dma_request_slave_channel_reason(tspi->dev,
 					dma_to_memory ? "rx" : "tx");
@@ -603,19 +684,6 @@ static int tegra_spi_init_dma_param(struct tegra_spi_data *tspi,
 	}
 
 	if (dma_to_memory) {
-		dma_sconfig.src_addr = tspi->phys + SPI_RX_FIFO;
-		dma_sconfig.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
-		dma_sconfig.src_maxburst = 0;
-	} else {
-		dma_sconfig.dst_addr = tspi->phys + SPI_TX_FIFO;
-		dma_sconfig.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
-		dma_sconfig.dst_maxburst = 0;
-	}
-
-	ret = dmaengine_slave_config(dma_chan, &dma_sconfig);
-	if (ret)
-		goto scrub;
-	if (dma_to_memory) {
 		tspi->rx_dma_chan = dma_chan;
 		tspi->rx_dma_buf = dma_buf;
 		tspi->rx_dma_phys = dma_phys;
@@ -625,11 +693,6 @@ static int tegra_spi_init_dma_param(struct tegra_spi_data *tspi,
 		tspi->tx_dma_phys = dma_phys;
 	}
 	return 0;
-
-scrub:
-	dma_free_coherent(tspi->dev, tspi->dma_buf_size, dma_buf, dma_phys);
-	dma_release_channel(dma_chan);
-	return ret;
 }
 
 static void tegra_spi_deinit_dma_param(struct tegra_spi_data *tspi,
@@ -696,6 +759,16 @@ static u32 tegra_spi_setup_transfer_one(struct spi_device *spi,
 		else if (req_mode == SPI_MODE_3)
 			command1 |= SPI_CONTROL_MODE_3;
 
+		if (spi->mode & SPI_LSB_FIRST)
+			command1 |= SPI_LSBIT_FE;
+		else
+			command1 &= ~SPI_LSBIT_FE;
+
+		if (spi->mode & SPI_3WIRE)
+			command1 |= SPI_BIDIROE;
+		else
+			command1 &= ~SPI_BIDIROE;
+
 		if (tspi->cs_control) {
 			if (tspi->cs_control != spi)
 				tegra_spi_writel(tspi, command1, SPI_COMMAND1);
@@ -728,8 +801,15 @@ static int tegra_spi_start_transfer_one(struct spi_device *spi,
 
 	total_fifo_words = tegra_spi_calculate_curr_xfer_param(spi, tspi, t);
 
+	if (t->rx_nbits == SPI_NBITS_DUAL || t->tx_nbits == SPI_NBITS_DUAL)
+		command1 |= SPI_BOTH_EN_BIT;
+	else
+		command1 &= ~SPI_BOTH_EN_BIT;
+
 	if (tspi->is_packed)
 		command1 |= SPI_PACKED;
+	else
+		command1 &= ~SPI_PACKED;
 
 	command1 &= ~(SPI_CS_SEL_MASK | SPI_TX_EN | SPI_RX_EN);
 	tspi->cur_direction = 0;
@@ -748,6 +828,9 @@ static int tegra_spi_start_transfer_one(struct spi_device *spi,
 	dev_dbg(tspi->dev, "The def 0x%x and written 0x%x\n",
 		tspi->def_command1_reg, (unsigned)command1);
 
+	ret = tegra_spi_flush_fifos(tspi);
+	if (ret < 0)
+		return ret;
 	if (total_fifo_words > SPI_FIFO_DEPTH)
 		ret = tegra_spi_start_dma_based_transfer(tspi, t);
 	else
@@ -774,6 +857,12 @@ static int tegra_spi_setup(struct spi_device *spi)
 		return ret;
 	}
 
+	if (tspi->soc_data->has_intr_mask_reg) {
+		val = tegra_spi_readl(tspi, SPI_INTR_MASK);
+		val &= ~SPI_INTR_ALL_MASK;
+		tegra_spi_writel(tspi, val, SPI_INTR_MASK);
+	}
+
 	spin_lock_irqsave(&tspi->lock, flags);
 	val = tspi->def_command1_reg;
 	if (spi->mode & SPI_CS_HIGH)
@@ -799,6 +888,33 @@ static void tegra_spi_transfer_delay(int delay)
 	udelay(delay % 1000);
 }
 
+static void tegra_spi_transfer_end(struct spi_device *spi)
+{
+	struct tegra_spi_data *tspi = spi_master_get_devdata(spi->master);
+	int cs_val = (spi->mode & SPI_CS_HIGH) ? 0 : 1;
+
+	if (cs_val)
+		tspi->command1_reg |= SPI_CS_SW_VAL;
+	else
+		tspi->command1_reg &= ~SPI_CS_SW_VAL;
+	tegra_spi_writel(tspi, tspi->command1_reg, SPI_COMMAND1);
+	tegra_spi_writel(tspi, tspi->def_command1_reg, SPI_COMMAND1);
+}
+
+static void tegra_spi_dump_regs(struct tegra_spi_data *tspi)
+{
+	dev_dbg(tspi->dev, "============ SPI REGISTER DUMP ============\n");
+	dev_dbg(tspi->dev, "Command1:    0x%08x | Command2:    0x%08x\n",
+		tegra_spi_readl(tspi, SPI_COMMAND1),
+		tegra_spi_readl(tspi, SPI_COMMAND2));
+	dev_dbg(tspi->dev, "DMA_CTL:     0x%08x | DMA_BLK:     0x%08x\n",
+		tegra_spi_readl(tspi, SPI_DMA_CTL),
+		tegra_spi_readl(tspi, SPI_DMA_BLK));
+	dev_dbg(tspi->dev, "TRANS_STAT:  0x%08x | FIFO_STATUS: 0x%08x\n",
+		tegra_spi_readl(tspi, SPI_TRANS_STATUS),
+		tegra_spi_readl(tspi, SPI_FIFO_STATUS));
+}
+
 static int tegra_spi_transfer_one_message(struct spi_master *master,
 			struct spi_message *msg)
 {
@@ -838,21 +954,32 @@ static int tegra_spi_transfer_one_message(struct spi_master *master,
 		if (WARN_ON(ret == 0)) {
 			dev_err(tspi->dev,
 				"spi transfer timeout, err %d\n", ret);
+			if (tspi->is_curr_dma_xfer &&
+			    (tspi->cur_direction & DATA_DIR_TX))
+				dmaengine_terminate_all(tspi->tx_dma_chan);
+			if (tspi->is_curr_dma_xfer &&
+			    (tspi->cur_direction & DATA_DIR_RX))
+				dmaengine_terminate_all(tspi->rx_dma_chan);
 			ret = -EIO;
+			tegra_spi_dump_regs(tspi);
+			tegra_spi_flush_fifos(tspi);
+			reset_control_assert(tspi->rst);
+			udelay(2);
+			reset_control_deassert(tspi->rst);
 			goto complete_xfer;
 		}
 
 		if (tspi->tx_status ||  tspi->rx_status) {
 			dev_err(tspi->dev, "Error in Transfer\n");
 			ret = -EIO;
+			tegra_spi_dump_regs(tspi);
 			goto complete_xfer;
 		}
 		msg->actual_length += xfer->len;
 
 complete_xfer:
 		if (ret < 0 || skip) {
-			tegra_spi_writel(tspi, tspi->def_command1_reg,
-					SPI_COMMAND1);
+			tegra_spi_transfer_end(spi);
 			tegra_spi_transfer_delay(xfer->delay_usecs);
 			goto exit;
 		} else if (list_is_last(&xfer->transfer_list,
@@ -860,13 +987,11 @@ complete_xfer:
 			if (xfer->cs_change)
 				tspi->cs_control = spi;
 			else {
-				tegra_spi_writel(tspi, tspi->def_command1_reg,
-						SPI_COMMAND1);
+				tegra_spi_transfer_end(spi);
 				tegra_spi_transfer_delay(xfer->delay_usecs);
 			}
 		} else if (xfer->cs_change) {
-			tegra_spi_writel(tspi, tspi->def_command1_reg,
-					SPI_COMMAND1);
+			tegra_spi_transfer_end(spi);
 			tegra_spi_transfer_delay(xfer->delay_usecs);
 		}
 
@@ -889,11 +1014,14 @@ static irqreturn_t handle_cpu_based_xfer(struct tegra_spi_data *tspi)
 			tspi->status_reg);
 		dev_err(tspi->dev, "CpuXfer 0x%08x:0x%08x\n",
 			tspi->command1_reg, tspi->dma_control_reg);
+		tegra_spi_dump_regs(tspi);
+		tegra_spi_flush_fifos(tspi);
+		complete(&tspi->xfer_completion);
+		spin_unlock_irqrestore(&tspi->lock, flags);
 		reset_control_assert(tspi->rst);
 		udelay(2);
 		reset_control_deassert(tspi->rst);
-		complete(&tspi->xfer_completion);
-		goto exit;
+		return IRQ_HANDLED;
 	}
 
 	if (tspi->cur_direction & DATA_DIR_RX)
@@ -961,11 +1089,13 @@ static irqreturn_t handle_dma_based_xfer(struct tegra_spi_data *tspi)
 			tspi->status_reg);
 		dev_err(tspi->dev, "DmaXfer 0x%08x:0x%08x\n",
 			tspi->command1_reg, tspi->dma_control_reg);
+		tegra_spi_dump_regs(tspi);
+		tegra_spi_flush_fifos(tspi);
+		complete(&tspi->xfer_completion);
+		spin_unlock_irqrestore(&tspi->lock, flags);
 		reset_control_assert(tspi->rst);
 		udelay(2);
 		reset_control_deassert(tspi->rst);
-		complete(&tspi->xfer_completion);
-		spin_unlock_irqrestore(&tspi->lock, flags);
 		return IRQ_HANDLED;
 	}
 
@@ -1021,8 +1151,29 @@ static irqreturn_t tegra_spi_isr(int irq, void *context_data)
 	return IRQ_WAKE_THREAD;
 }
 
+static struct tegra_spi_soc_data tegra114_spi_soc_data = {
+	.has_intr_mask_reg = false,
+};
+
+static struct tegra_spi_soc_data tegra124_spi_soc_data = {
+	.has_intr_mask_reg = false,
+};
+
+static struct tegra_spi_soc_data tegra210_spi_soc_data = {
+	.has_intr_mask_reg = true,
+};
+
 static const struct of_device_id tegra_spi_of_match[] = {
-	{ .compatible = "nvidia,tegra114-spi", },
+	{
+		.compatible = "nvidia,tegra114-spi",
+		.data	    = &tegra114_spi_soc_data,
+	}, {
+		.compatible = "nvidia,tegra124-spi",
+		.data	    = &tegra124_spi_soc_data,
+	}, {
+		.compatible = "nvidia,tegra210-spi",
+		.data	    = &tegra210_spi_soc_data,
+	},
 	{}
 };
 MODULE_DEVICE_TABLE(of, tegra_spi_of_match);
@@ -1033,6 +1184,7 @@ static int tegra_spi_probe(struct platform_device *pdev)
 	struct tegra_spi_data	*tspi;
 	struct resource		*r;
 	int ret, spi_irq;
+	int bus_num;
 
 	master = spi_alloc_master(&pdev->dev, sizeof(*tspi));
 	if (!master) {
@@ -1047,16 +1199,28 @@ static int tegra_spi_probe(struct platform_device *pdev)
 		master->max_speed_hz = 25000000; /* 25MHz */
 
 	/* the spi->mode bits understood by this driver: */
-	master->mode_bits = SPI_CPOL | SPI_CPHA | SPI_CS_HIGH;
+	master->mode_bits = SPI_CPOL | SPI_CPHA | SPI_CS_HIGH | SPI_LSB_FIRST |
+			    SPI_TX_DUAL | SPI_RX_DUAL | SPI_3WIRE;
+	master->bits_per_word_mask = SPI_BPW_RANGE_MASK(4, 32);
 	master->setup = tegra_spi_setup;
 	master->transfer_one_message = tegra_spi_transfer_one_message;
 	master->num_chipselect = MAX_CHIP_SELECT;
 	master->auto_runtime_pm = true;
+	bus_num = of_alias_get_id(pdev->dev.of_node, "spi");
+	if (bus_num >= 0)
+		master->bus_num = bus_num;
 
 	tspi->master = master;
 	tspi->dev = &pdev->dev;
 	spin_lock_init(&tspi->lock);
 
+	tspi->soc_data = of_device_get_match_data(&pdev->dev);
+	if (!tspi->soc_data) {
+		dev_err(&pdev->dev, "unsupported tegra\n");
+		ret = -ENODEV;
+		goto exit_free_master;
+	}
+
 	r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	tspi->base = devm_ioremap_resource(&pdev->dev, r);
 	if (IS_ERR(tspi->base)) {
@@ -1067,27 +1231,19 @@ static int tegra_spi_probe(struct platform_device *pdev)
 
 	spi_irq = platform_get_irq(pdev, 0);
 	tspi->irq = spi_irq;
-	ret = request_threaded_irq(tspi->irq, tegra_spi_isr,
-			tegra_spi_isr_thread, IRQF_ONESHOT,
-			dev_name(&pdev->dev), tspi);
-	if (ret < 0) {
-		dev_err(&pdev->dev, "Failed to register ISR for IRQ %d\n",
-					tspi->irq);
-		goto exit_free_master;
-	}
 
 	tspi->clk = devm_clk_get(&pdev->dev, "spi");
 	if (IS_ERR(tspi->clk)) {
 		dev_err(&pdev->dev, "can not get clock\n");
 		ret = PTR_ERR(tspi->clk);
-		goto exit_free_irq;
+		goto exit_free_master;
 	}
 
 	tspi->rst = devm_reset_control_get_exclusive(&pdev->dev, "spi");
 	if (IS_ERR(tspi->rst)) {
 		dev_err(&pdev->dev, "can not get reset\n");
 		ret = PTR_ERR(tspi->rst);
-		goto exit_free_irq;
+		goto exit_free_master;
 	}
 
 	tspi->max_buf_size = SPI_FIFO_DEPTH << 2;
@@ -1095,7 +1251,7 @@ static int tegra_spi_probe(struct platform_device *pdev)
 
 	ret = tegra_spi_init_dma_param(tspi, true);
 	if (ret < 0)
-		goto exit_free_irq;
+		goto exit_free_master;
 	ret = tegra_spi_init_dma_param(tspi, false);
 	if (ret < 0)
 		goto exit_rx_dma_free;
@@ -1117,18 +1273,32 @@ static int tegra_spi_probe(struct platform_device *pdev)
 		dev_err(&pdev->dev, "pm runtime get failed, e = %d\n", ret);
 		goto exit_pm_disable;
 	}
+
+	reset_control_assert(tspi->rst);
+	udelay(2);
+	reset_control_deassert(tspi->rst);
 	tspi->def_command1_reg  = SPI_M_S;
 	tegra_spi_writel(tspi, tspi->def_command1_reg, SPI_COMMAND1);
 	pm_runtime_put(&pdev->dev);
+	ret = request_threaded_irq(tspi->irq, tegra_spi_isr,
+				   tegra_spi_isr_thread, IRQF_ONESHOT,
+				   dev_name(&pdev->dev), tspi);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "Failed to register ISR for IRQ %d\n",
+			tspi->irq);
+		goto exit_pm_disable;
+	}
 
 	master->dev.of_node = pdev->dev.of_node;
 	ret = devm_spi_register_master(&pdev->dev, master);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "can not register to master err %d\n", ret);
-		goto exit_pm_disable;
+		goto exit_free_irq;
 	}
 	return ret;
 
+exit_free_irq:
+	free_irq(spi_irq, tspi);
 exit_pm_disable:
 	pm_runtime_disable(&pdev->dev);
 	if (!pm_runtime_status_suspended(&pdev->dev))
@@ -1136,8 +1306,6 @@ exit_pm_disable:
 	tegra_spi_deinit_dma_param(tspi, false);
 exit_rx_dma_free:
 	tegra_spi_deinit_dma_param(tspi, true);
-exit_free_irq:
-	free_irq(spi_irq, tspi);
 exit_free_master:
 	spi_master_put(master);
 	return ret;
diff --git a/drivers/spi/spi-tegra20-slink.c b/drivers/spi/spi-tegra20-slink.c
index 1427f343b39a..6d4679126213 100644
--- a/drivers/spi/spi-tegra20-slink.c
+++ b/drivers/spi/spi-tegra20-slink.c
@@ -717,9 +717,6 @@ static int tegra_slink_start_transfer_one(struct spi_device *spi,
 	command2 = tspi->command2_reg;
 	command2 &= ~(SLINK_RXEN | SLINK_TXEN);
 
-	tegra_slink_writel(tspi, command, SLINK_COMMAND);
-	tspi->command_reg = command;
-
 	tspi->cur_direction = 0;
 	if (t->rx_buf) {
 		command2 |= SLINK_RXEN;
@@ -729,9 +726,18 @@ static int tegra_slink_start_transfer_one(struct spi_device *spi,
 		command2 |= SLINK_TXEN;
 		tspi->cur_direction |= DATA_DIR_TX;
 	}
+
+	/*
+	 * Writing to the command2 register bevore the command register prevents
+	 * a spike in chip_select line 0. This selects the chip_select line
+	 * before changing the chip_select value.
+	 */
 	tegra_slink_writel(tspi, command2, SLINK_COMMAND2);
 	tspi->command2_reg = command2;
 
+	tegra_slink_writel(tspi, command, SLINK_COMMAND);
+	tspi->command_reg = command;
+
 	if (total_fifo_words > SLINK_FIFO_DEPTH)
 		ret = tegra_slink_start_dma_based_transfer(tspi, t);
 	else
diff --git a/drivers/spi/spi-topcliff-pch.c b/drivers/spi/spi-topcliff-pch.c
index fba3f180f233..8a5966963834 100644
--- a/drivers/spi/spi-topcliff-pch.c
+++ b/drivers/spi/spi-topcliff-pch.c
@@ -1299,18 +1299,27 @@ static void pch_free_dma_buf(struct pch_spi_board_data *board_dat,
 				  dma->rx_buf_virt, dma->rx_buf_dma);
 }
 
-static void pch_alloc_dma_buf(struct pch_spi_board_data *board_dat,
+static int pch_alloc_dma_buf(struct pch_spi_board_data *board_dat,
 			      struct pch_spi_data *data)
 {
 	struct pch_spi_dma_ctrl *dma;
+	int ret;
 
 	dma = &data->dma;
+	ret = 0;
 	/* Get Consistent memory for Tx DMA */
 	dma->tx_buf_virt = dma_alloc_coherent(&board_dat->pdev->dev,
 				PCH_BUF_SIZE, &dma->tx_buf_dma, GFP_KERNEL);
+	if (!dma->tx_buf_virt)
+		ret = -ENOMEM;
+
 	/* Get Consistent memory for Rx DMA */
 	dma->rx_buf_virt = dma_alloc_coherent(&board_dat->pdev->dev,
 				PCH_BUF_SIZE, &dma->rx_buf_dma, GFP_KERNEL);
+	if (!dma->rx_buf_virt)
+		ret = -ENOMEM;
+
+	return ret;
 }
 
 static int pch_spi_pd_probe(struct platform_device *plat_dev)
@@ -1387,7 +1396,9 @@ static int pch_spi_pd_probe(struct platform_device *plat_dev)
 
 	if (use_dma) {
 		dev_info(&plat_dev->dev, "Use DMA for data transfers\n");
-		pch_alloc_dma_buf(board_dat, data);
+		ret = pch_alloc_dma_buf(board_dat, data);
+		if (ret)
+			goto err_spi_register_master;
 	}
 
 	ret = spi_register_master(master);
diff --git a/drivers/spi/spi-zynq-qspi.c b/drivers/spi/spi-zynq-qspi.c
new file mode 100644
index 000000000000..c6bee67decb5
--- /dev/null
+++ b/drivers/spi/spi-zynq-qspi.c
@@ -0,0 +1,761 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright (C) 2019 Xilinx, Inc.
+ *
+ * Author: Naga Sureshkumar Relli <nagasure@xilinx.com>
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/gpio.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/workqueue.h>
+#include <linux/spi/spi-mem.h>
+
+/* Register offset definitions */
+#define ZYNQ_QSPI_CONFIG_OFFSET		0x00 /* Configuration  Register, RW */
+#define ZYNQ_QSPI_STATUS_OFFSET		0x04 /* Interrupt Status Register, RO */
+#define ZYNQ_QSPI_IEN_OFFSET		0x08 /* Interrupt Enable Register, WO */
+#define ZYNQ_QSPI_IDIS_OFFSET		0x0C /* Interrupt Disable Reg, WO */
+#define ZYNQ_QSPI_IMASK_OFFSET		0x10 /* Interrupt Enabled Mask Reg,RO */
+#define ZYNQ_QSPI_ENABLE_OFFSET		0x14 /* Enable/Disable Register, RW */
+#define ZYNQ_QSPI_DELAY_OFFSET		0x18 /* Delay Register, RW */
+#define ZYNQ_QSPI_TXD_00_00_OFFSET	0x1C /* Transmit 4-byte inst, WO */
+#define ZYNQ_QSPI_TXD_00_01_OFFSET	0x80 /* Transmit 1-byte inst, WO */
+#define ZYNQ_QSPI_TXD_00_10_OFFSET	0x84 /* Transmit 2-byte inst, WO */
+#define ZYNQ_QSPI_TXD_00_11_OFFSET	0x88 /* Transmit 3-byte inst, WO */
+#define ZYNQ_QSPI_RXD_OFFSET		0x20 /* Data Receive Register, RO */
+#define ZYNQ_QSPI_SIC_OFFSET		0x24 /* Slave Idle Count Register, RW */
+#define ZYNQ_QSPI_TX_THRESH_OFFSET	0x28 /* TX FIFO Watermark Reg, RW */
+#define ZYNQ_QSPI_RX_THRESH_OFFSET	0x2C /* RX FIFO Watermark Reg, RW */
+#define ZYNQ_QSPI_GPIO_OFFSET		0x30 /* GPIO Register, RW */
+#define ZYNQ_QSPI_LINEAR_CFG_OFFSET	0xA0 /* Linear Adapter Config Ref, RW */
+#define ZYNQ_QSPI_MOD_ID_OFFSET		0xFC /* Module ID Register, RO */
+
+/*
+ * QSPI Configuration Register bit Masks
+ *
+ * This register contains various control bits that effect the operation
+ * of the QSPI controller
+ */
+#define ZYNQ_QSPI_CONFIG_IFMODE_MASK	BIT(31) /* Flash Memory Interface */
+#define ZYNQ_QSPI_CONFIG_MANSRT_MASK	BIT(16) /* Manual TX Start */
+#define ZYNQ_QSPI_CONFIG_MANSRTEN_MASK	BIT(15) /* Enable Manual TX Mode */
+#define ZYNQ_QSPI_CONFIG_SSFORCE_MASK	BIT(14) /* Manual Chip Select */
+#define ZYNQ_QSPI_CONFIG_BDRATE_MASK	GENMASK(5, 3) /* Baud Rate Mask */
+#define ZYNQ_QSPI_CONFIG_CPHA_MASK	BIT(2) /* Clock Phase Control */
+#define ZYNQ_QSPI_CONFIG_CPOL_MASK	BIT(1) /* Clock Polarity Control */
+#define ZYNQ_QSPI_CONFIG_SSCTRL_MASK	BIT(10) /* Slave Select Mask */
+#define ZYNQ_QSPI_CONFIG_FWIDTH_MASK	GENMASK(7, 6) /* FIFO width */
+#define ZYNQ_QSPI_CONFIG_MSTREN_MASK	BIT(0) /* Master Mode */
+
+/*
+ * QSPI Configuration Register - Baud rate and slave select
+ *
+ * These are the values used in the calculation of baud rate divisor and
+ * setting the slave select.
+ */
+#define ZYNQ_QSPI_BAUD_DIV_MAX		GENMASK(2, 0) /* Baud rate maximum */
+#define ZYNQ_QSPI_BAUD_DIV_SHIFT	3 /* Baud rate divisor shift in CR */
+#define ZYNQ_QSPI_SS_SHIFT		10 /* Slave Select field shift in CR */
+
+/*
+ * QSPI Interrupt Registers bit Masks
+ *
+ * All the four interrupt registers (Status/Mask/Enable/Disable) have the same
+ * bit definitions.
+ */
+#define ZYNQ_QSPI_IXR_RX_OVERFLOW_MASK	BIT(0) /* QSPI RX FIFO Overflow */
+#define ZYNQ_QSPI_IXR_TXNFULL_MASK	BIT(2) /* QSPI TX FIFO Overflow */
+#define ZYNQ_QSPI_IXR_TXFULL_MASK	BIT(3) /* QSPI TX FIFO is full */
+#define ZYNQ_QSPI_IXR_RXNEMTY_MASK	BIT(4) /* QSPI RX FIFO Not Empty */
+#define ZYNQ_QSPI_IXR_RXF_FULL_MASK	BIT(5) /* QSPI RX FIFO is full */
+#define ZYNQ_QSPI_IXR_TXF_UNDRFLOW_MASK	BIT(6) /* QSPI TX FIFO Underflow */
+#define ZYNQ_QSPI_IXR_ALL_MASK		(ZYNQ_QSPI_IXR_RX_OVERFLOW_MASK | \
+					ZYNQ_QSPI_IXR_TXNFULL_MASK | \
+					ZYNQ_QSPI_IXR_TXFULL_MASK | \
+					ZYNQ_QSPI_IXR_RXNEMTY_MASK | \
+					ZYNQ_QSPI_IXR_RXF_FULL_MASK | \
+					ZYNQ_QSPI_IXR_TXF_UNDRFLOW_MASK)
+#define ZYNQ_QSPI_IXR_RXTX_MASK		(ZYNQ_QSPI_IXR_TXNFULL_MASK | \
+					ZYNQ_QSPI_IXR_RXNEMTY_MASK)
+
+/*
+ * QSPI Enable Register bit Masks
+ *
+ * This register is used to enable or disable the QSPI controller
+ */
+#define ZYNQ_QSPI_ENABLE_ENABLE_MASK	BIT(0) /* QSPI Enable Bit Mask */
+
+/*
+ * QSPI Linear Configuration Register
+ *
+ * It is named Linear Configuration but it controls other modes when not in
+ * linear mode also.
+ */
+#define ZYNQ_QSPI_LCFG_TWO_MEM_MASK	BIT(30) /* LQSPI Two memories Mask */
+#define ZYNQ_QSPI_LCFG_SEP_BUS_MASK	BIT(29) /* LQSPI Separate bus Mask */
+#define ZYNQ_QSPI_LCFG_U_PAGE_MASK	BIT(28) /* LQSPI Upper Page Mask */
+
+#define ZYNQ_QSPI_LCFG_DUMMY_SHIFT	8
+
+#define ZYNQ_QSPI_FAST_READ_QOUT_CODE	0x6B /* read instruction code */
+#define ZYNQ_QSPI_FIFO_DEPTH		63 /* FIFO depth in words */
+#define ZYNQ_QSPI_RX_THRESHOLD		32 /* Rx FIFO threshold level */
+#define ZYNQ_QSPI_TX_THRESHOLD		1 /* Tx FIFO threshold level */
+
+/*
+ * The modebits configurable by the driver to make the SPI support different
+ * data formats
+ */
+#define ZYNQ_QSPI_MODEBITS			(SPI_CPOL | SPI_CPHA)
+
+/* Default number of chip selects */
+#define ZYNQ_QSPI_DEFAULT_NUM_CS	1
+
+/**
+ * struct zynq_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
+ * @txbuf:		Pointer to the TX buffer
+ * @rxbuf:		Pointer to the RX buffer
+ * @tx_bytes:		Number of bytes left to transfer
+ * @rx_bytes:		Number of bytes left to receive
+ * @data_completion:	completion structure
+ */
+struct zynq_qspi {
+	struct device *dev;
+	void __iomem *regs;
+	struct clk *refclk;
+	struct clk *pclk;
+	int irq;
+	u8 *txbuf;
+	u8 *rxbuf;
+	int tx_bytes;
+	int rx_bytes;
+	struct completion data_completion;
+};
+
+/*
+ * Inline functions for the QSPI controller read/write
+ */
+static inline u32 zynq_qspi_read(struct zynq_qspi *xqspi, u32 offset)
+{
+	return readl_relaxed(xqspi->regs + offset);
+}
+
+static inline void zynq_qspi_write(struct zynq_qspi *xqspi, u32 offset,
+				   u32 val)
+{
+	writel_relaxed(val, xqspi->regs + offset);
+}
+
+/**
+ * zynq_qspi_init_hw - Initialize the hardware
+ * @xqspi:	Pointer to the zynq_qspi structure
+ *
+ * The default settings of the QSPI controller's configurable parameters on
+ * reset are
+ *	- Master mode
+ *	- Baud rate divisor is set to 2
+ *	- Tx threshold set to 1l Rx threshold set to 32
+ *	- Flash memory interface mode enabled
+ *	- Size of the word to be transferred as 8 bit
+ * 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 size of the word to be transferred as 32 bit
+ *	- Set the little endian mode of TX FIFO and
+ *	- Enable the QSPI controller
+ */
+static void zynq_qspi_init_hw(struct zynq_qspi *xqspi)
+{
+	u32 config_reg;
+
+	zynq_qspi_write(xqspi, ZYNQ_QSPI_ENABLE_OFFSET, 0);
+	zynq_qspi_write(xqspi, ZYNQ_QSPI_IDIS_OFFSET, ZYNQ_QSPI_IXR_ALL_MASK);
+
+	/* Disable linear mode as the boot loader may have used it */
+	zynq_qspi_write(xqspi, ZYNQ_QSPI_LINEAR_CFG_OFFSET, 0);
+
+	/* Clear the RX FIFO */
+	while (zynq_qspi_read(xqspi, ZYNQ_QSPI_STATUS_OFFSET) &
+			      ZYNQ_QSPI_IXR_RXNEMTY_MASK)
+		zynq_qspi_read(xqspi, ZYNQ_QSPI_RXD_OFFSET);
+
+	zynq_qspi_write(xqspi, ZYNQ_QSPI_STATUS_OFFSET, ZYNQ_QSPI_IXR_ALL_MASK);
+	config_reg = zynq_qspi_read(xqspi, ZYNQ_QSPI_CONFIG_OFFSET);
+	config_reg &= ~(ZYNQ_QSPI_CONFIG_MSTREN_MASK |
+			ZYNQ_QSPI_CONFIG_CPOL_MASK |
+			ZYNQ_QSPI_CONFIG_CPHA_MASK |
+			ZYNQ_QSPI_CONFIG_BDRATE_MASK |
+			ZYNQ_QSPI_CONFIG_SSFORCE_MASK |
+			ZYNQ_QSPI_CONFIG_MANSRTEN_MASK |
+			ZYNQ_QSPI_CONFIG_MANSRT_MASK);
+	config_reg |= (ZYNQ_QSPI_CONFIG_MSTREN_MASK |
+		       ZYNQ_QSPI_CONFIG_SSFORCE_MASK |
+		       ZYNQ_QSPI_CONFIG_FWIDTH_MASK |
+		       ZYNQ_QSPI_CONFIG_IFMODE_MASK);
+	zynq_qspi_write(xqspi, ZYNQ_QSPI_CONFIG_OFFSET, config_reg);
+
+	zynq_qspi_write(xqspi, ZYNQ_QSPI_RX_THRESH_OFFSET,
+			ZYNQ_QSPI_RX_THRESHOLD);
+	zynq_qspi_write(xqspi, ZYNQ_QSPI_TX_THRESH_OFFSET,
+			ZYNQ_QSPI_TX_THRESHOLD);
+
+	zynq_qspi_write(xqspi, ZYNQ_QSPI_ENABLE_OFFSET,
+			ZYNQ_QSPI_ENABLE_ENABLE_MASK);
+}
+
+static bool zynq_qspi_supports_op(struct spi_mem *mem,
+				  const struct spi_mem_op *op)
+{
+	if (!spi_mem_default_supports_op(mem, op))
+		return false;
+
+	/*
+	 * The number of address bytes should be equal to or less than 3 bytes.
+	 */
+	if (op->addr.nbytes > 3)
+		return false;
+
+	return true;
+}
+
+/**
+ * zynq_qspi_rxfifo_op - Read 1..4 bytes from RxFIFO to RX buffer
+ * @xqspi:	Pointer to the zynq_qspi structure
+ * @size:	Number of bytes to be read (1..4)
+ */
+static void zynq_qspi_rxfifo_op(struct zynq_qspi *xqspi, unsigned int size)
+{
+	u32 data;
+
+	data = zynq_qspi_read(xqspi, ZYNQ_QSPI_RXD_OFFSET);
+
+	if (xqspi->rxbuf) {
+		memcpy(xqspi->rxbuf, ((u8 *)&data) + 4 - size, size);
+		xqspi->rxbuf += size;
+	}
+
+	xqspi->rx_bytes -= size;
+	if (xqspi->rx_bytes < 0)
+		xqspi->rx_bytes = 0;
+}
+
+/**
+ * zynq_qspi_txfifo_op - Write 1..4 bytes from TX buffer to TxFIFO
+ * @xqspi:	Pointer to the zynq_qspi structure
+ * @size:	Number of bytes to be written (1..4)
+ */
+static void zynq_qspi_txfifo_op(struct zynq_qspi *xqspi, unsigned int size)
+{
+	static const unsigned int offset[4] = {
+		ZYNQ_QSPI_TXD_00_01_OFFSET, ZYNQ_QSPI_TXD_00_10_OFFSET,
+		ZYNQ_QSPI_TXD_00_11_OFFSET, ZYNQ_QSPI_TXD_00_00_OFFSET };
+	u32 data;
+
+	if (xqspi->txbuf) {
+		data = 0xffffffff;
+		memcpy(&data, xqspi->txbuf, size);
+		xqspi->txbuf += size;
+	} else {
+		data = 0;
+	}
+
+	xqspi->tx_bytes -= size;
+	zynq_qspi_write(xqspi, offset[size - 1], data);
+}
+
+/**
+ * zynq_qspi_chipselect - Select or deselect the chip select line
+ * @spi:	Pointer to the spi_device structure
+ * @assert:	1 for select or 0 for deselect the chip select line
+ */
+static void zynq_qspi_chipselect(struct spi_device *spi, bool assert)
+{
+	struct spi_controller *ctrl = spi->master;
+	struct zynq_qspi *xqspi = spi_controller_get_devdata(ctrl);
+	u32 config_reg;
+
+	config_reg = zynq_qspi_read(xqspi, ZYNQ_QSPI_CONFIG_OFFSET);
+	if (assert) {
+		/* Select the slave */
+		config_reg &= ~ZYNQ_QSPI_CONFIG_SSCTRL_MASK;
+		config_reg |= (((~(BIT(spi->chip_select))) <<
+				ZYNQ_QSPI_SS_SHIFT) &
+				ZYNQ_QSPI_CONFIG_SSCTRL_MASK);
+	} else {
+		config_reg |= ZYNQ_QSPI_CONFIG_SSCTRL_MASK;
+	}
+
+	zynq_qspi_write(xqspi, ZYNQ_QSPI_CONFIG_OFFSET, config_reg);
+}
+
+/**
+ * zynq_qspi_config_op - Configure QSPI controller for specified transfer
+ * @xqspi:	Pointer to the zynq_qspi structure
+ * @qspi:	Pointer to the spi_device structure
+ *
+ * Sets the operational mode of QSPI controller for the next QSPI transfer and
+ * sets the requested clock frequency.
+ *
+ * Return:	0 on success and -EINVAL on invalid input parameter
+ *
+ * Note: If the requested frequency is not an exact match with what can be
+ * obtained using the prescalar 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 zynq_qspi_config_op(struct zynq_qspi *xqspi, struct spi_device *spi)
+{
+	u32 config_reg, baud_rate_val = 0;
+
+	/*
+	 * Set the clock frequency
+	 * The baud rate divisor is not a direct mapping to the value written
+	 * into the configuration register (config_reg[5:3])
+	 * i.e. 000 - divide by 2
+	 *      001 - divide by 4
+	 *      ----------------
+	 *      111 - divide by 256
+	 */
+	while ((baud_rate_val < ZYNQ_QSPI_BAUD_DIV_MAX)  &&
+	       (clk_get_rate(xqspi->refclk) / (2 << baud_rate_val)) >
+		spi->max_speed_hz)
+		baud_rate_val++;
+
+	config_reg = zynq_qspi_read(xqspi, ZYNQ_QSPI_CONFIG_OFFSET);
+
+	/* Set the QSPI clock phase and clock polarity */
+	config_reg &= (~ZYNQ_QSPI_CONFIG_CPHA_MASK) &
+		      (~ZYNQ_QSPI_CONFIG_CPOL_MASK);
+	if (spi->mode & SPI_CPHA)
+		config_reg |= ZYNQ_QSPI_CONFIG_CPHA_MASK;
+	if (spi->mode & SPI_CPOL)
+		config_reg |= ZYNQ_QSPI_CONFIG_CPOL_MASK;
+
+	config_reg &= ~ZYNQ_QSPI_CONFIG_BDRATE_MASK;
+	config_reg |= (baud_rate_val << ZYNQ_QSPI_BAUD_DIV_SHIFT);
+	zynq_qspi_write(xqspi, ZYNQ_QSPI_CONFIG_OFFSET, config_reg);
+
+	return 0;
+}
+
+/**
+ * zynq_qspi_setup - Configure the QSPI controller
+ * @spi:	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 and error value on failure
+ */
+static int zynq_qspi_setup_op(struct spi_device *spi)
+{
+	struct spi_controller *ctrl = spi->master;
+	struct zynq_qspi *qspi = spi_controller_get_devdata(ctrl);
+
+	if (ctrl->busy)
+		return -EBUSY;
+
+	clk_enable(qspi->refclk);
+	clk_enable(qspi->pclk);
+	zynq_qspi_write(qspi, ZYNQ_QSPI_ENABLE_OFFSET,
+			ZYNQ_QSPI_ENABLE_ENABLE_MASK);
+
+	return 0;
+}
+
+/**
+ * zynq_qspi_write_op - Fills the TX FIFO with as many bytes as possible
+ * @xqspi:	Pointer to the zynq_qspi structure
+ * @txcount:	Maximum number of words to write
+ * @txempty:	Indicates that TxFIFO is empty
+ */
+static void zynq_qspi_write_op(struct zynq_qspi *xqspi, int txcount,
+			       bool txempty)
+{
+	int count, len, k;
+
+	len = xqspi->tx_bytes;
+	if (len && len < 4) {
+		/*
+		 * We must empty the TxFIFO between accesses to TXD0,
+		 * TXD1, TXD2, TXD3.
+		 */
+		if (txempty)
+			zynq_qspi_txfifo_op(xqspi, len);
+
+		return;
+	}
+
+	count = len / 4;
+	if (count > txcount)
+		count = txcount;
+
+	if (xqspi->txbuf) {
+		iowrite32_rep(xqspi->regs + ZYNQ_QSPI_TXD_00_00_OFFSET,
+			      xqspi->txbuf, count);
+		xqspi->txbuf += count * 4;
+	} else {
+		for (k = 0; k < count; k++)
+			writel_relaxed(0, xqspi->regs +
+					  ZYNQ_QSPI_TXD_00_00_OFFSET);
+	}
+
+	xqspi->tx_bytes -= count * 4;
+}
+
+/**
+ * zynq_qspi_read_op - Drains the RX FIFO by as many bytes as possible
+ * @xqspi:	Pointer to the zynq_qspi structure
+ * @rxcount:	Maximum number of words to read
+ */
+static void zynq_qspi_read_op(struct zynq_qspi *xqspi, int rxcount)
+{
+	int count, len, k;
+
+	len = xqspi->rx_bytes - xqspi->tx_bytes;
+	count = len / 4;
+	if (count > rxcount)
+		count = rxcount;
+	if (xqspi->rxbuf) {
+		ioread32_rep(xqspi->regs + ZYNQ_QSPI_RXD_OFFSET,
+			     xqspi->rxbuf, count);
+		xqspi->rxbuf += count * 4;
+	} else {
+		for (k = 0; k < count; k++)
+			readl_relaxed(xqspi->regs + ZYNQ_QSPI_RXD_OFFSET);
+	}
+	xqspi->rx_bytes -= count * 4;
+	len -= count * 4;
+
+	if (len && len < 4 && count < rxcount)
+		zynq_qspi_rxfifo_op(xqspi, len);
+}
+
+/**
+ * zynq_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 zynq_qspi_irq(int irq, void *dev_id)
+{
+	u32 intr_status;
+	bool txempty;
+	struct zynq_qspi *xqspi = (struct zynq_qspi *)dev_id;
+
+	intr_status = zynq_qspi_read(xqspi, ZYNQ_QSPI_STATUS_OFFSET);
+	zynq_qspi_write(xqspi, ZYNQ_QSPI_STATUS_OFFSET, intr_status);
+
+	if ((intr_status & ZYNQ_QSPI_IXR_TXNFULL_MASK) ||
+	    (intr_status & ZYNQ_QSPI_IXR_RXNEMTY_MASK)) {
+		/*
+		 * This bit is set when Tx FIFO has < THRESHOLD entries.
+		 * We have the THRESHOLD value set to 1,
+		 * so this bit indicates Tx FIFO is empty.
+		 */
+		txempty = !!(intr_status & ZYNQ_QSPI_IXR_TXNFULL_MASK);
+		/* Read out the data from the RX FIFO */
+		zynq_qspi_read_op(xqspi, ZYNQ_QSPI_RX_THRESHOLD);
+		if (xqspi->tx_bytes) {
+			/* There is more data to send */
+			zynq_qspi_write_op(xqspi, ZYNQ_QSPI_RX_THRESHOLD,
+					   txempty);
+		} else {
+			/*
+			 * If transfer and receive is completed then only send
+			 * complete signal.
+			 */
+			if (!xqspi->rx_bytes) {
+				zynq_qspi_write(xqspi,
+						ZYNQ_QSPI_IDIS_OFFSET,
+						ZYNQ_QSPI_IXR_RXTX_MASK);
+				complete(&xqspi->data_completion);
+			}
+		}
+		return IRQ_HANDLED;
+	}
+
+	return IRQ_NONE;
+}
+
+/**
+ * zynq_qspi_exec_mem_op() - Initiates the QSPI transfer
+ * @mem: the SPI memory
+ * @op: the memory operation to execute
+ *
+ * Executes a memory operation.
+ *
+ * This function first selects the chip and starts the memory operation.
+ *
+ * Return: 0 in case of success, a negative error code otherwise.
+ */
+static int zynq_qspi_exec_mem_op(struct spi_mem *mem,
+				 const struct spi_mem_op *op)
+{
+	struct zynq_qspi *xqspi = spi_controller_get_devdata(mem->spi->master);
+	int err = 0, i;
+	u8 *tmpbuf;
+
+	dev_dbg(xqspi->dev, "cmd:%#x mode:%d.%d.%d.%d\n",
+		op->cmd.opcode, op->cmd.buswidth, op->addr.buswidth,
+		op->dummy.buswidth, op->data.buswidth);
+
+	zynq_qspi_chipselect(mem->spi, true);
+	zynq_qspi_config_op(xqspi, mem->spi);
+
+	if (op->cmd.opcode) {
+		reinit_completion(&xqspi->data_completion);
+		xqspi->txbuf = (u8 *)&op->cmd.opcode;
+		xqspi->rxbuf = NULL;
+		xqspi->tx_bytes = sizeof(op->cmd.opcode);
+		xqspi->rx_bytes = sizeof(op->cmd.opcode);
+		zynq_qspi_write_op(xqspi, ZYNQ_QSPI_FIFO_DEPTH, true);
+		zynq_qspi_write(xqspi, ZYNQ_QSPI_IEN_OFFSET,
+				ZYNQ_QSPI_IXR_RXTX_MASK);
+		if (!wait_for_completion_interruptible_timeout(&xqspi->data_completion,
+							       msecs_to_jiffies(1000)))
+			err = -ETIMEDOUT;
+	}
+
+	if (op->addr.nbytes) {
+		for (i = 0; i < op->addr.nbytes; i++) {
+			xqspi->txbuf[i] = op->addr.val >>
+					(8 * (op->addr.nbytes - i - 1));
+		}
+
+		reinit_completion(&xqspi->data_completion);
+		xqspi->rxbuf = NULL;
+		xqspi->tx_bytes = op->addr.nbytes;
+		xqspi->rx_bytes = op->addr.nbytes;
+		zynq_qspi_write_op(xqspi, ZYNQ_QSPI_FIFO_DEPTH, true);
+		zynq_qspi_write(xqspi, ZYNQ_QSPI_IEN_OFFSET,
+				ZYNQ_QSPI_IXR_RXTX_MASK);
+		if (!wait_for_completion_interruptible_timeout(&xqspi->data_completion,
+							       msecs_to_jiffies(1000)))
+			err = -ETIMEDOUT;
+	}
+
+	if (op->dummy.nbytes) {
+		tmpbuf = kzalloc(op->dummy.nbytes, GFP_KERNEL);
+		memset(tmpbuf, 0xff, op->dummy.nbytes);
+		reinit_completion(&xqspi->data_completion);
+		xqspi->txbuf = tmpbuf;
+		xqspi->rxbuf = NULL;
+		xqspi->tx_bytes = op->dummy.nbytes;
+		xqspi->rx_bytes = op->dummy.nbytes;
+		zynq_qspi_write_op(xqspi, ZYNQ_QSPI_FIFO_DEPTH, true);
+		zynq_qspi_write(xqspi, ZYNQ_QSPI_IEN_OFFSET,
+				ZYNQ_QSPI_IXR_RXTX_MASK);
+		if (!wait_for_completion_interruptible_timeout(&xqspi->data_completion,
+							       msecs_to_jiffies(1000)))
+			err = -ETIMEDOUT;
+
+		kfree(tmpbuf);
+	}
+
+	if (op->data.nbytes) {
+		reinit_completion(&xqspi->data_completion);
+		if (op->data.dir == SPI_MEM_DATA_OUT) {
+			xqspi->txbuf = (u8 *)op->data.buf.out;
+			xqspi->tx_bytes = op->data.nbytes;
+			xqspi->rxbuf = NULL;
+			xqspi->rx_bytes = op->data.nbytes;
+		} else {
+			xqspi->txbuf = NULL;
+			xqspi->rxbuf = (u8 *)op->data.buf.in;
+			xqspi->rx_bytes = op->data.nbytes;
+			xqspi->tx_bytes = op->data.nbytes;
+		}
+
+		zynq_qspi_write_op(xqspi, ZYNQ_QSPI_FIFO_DEPTH, true);
+		zynq_qspi_write(xqspi, ZYNQ_QSPI_IEN_OFFSET,
+				ZYNQ_QSPI_IXR_RXTX_MASK);
+		if (!wait_for_completion_interruptible_timeout(&xqspi->data_completion,
+							       msecs_to_jiffies(1000)))
+			err = -ETIMEDOUT;
+	}
+	zynq_qspi_chipselect(mem->spi, false);
+
+	return err;
+}
+
+static const struct spi_controller_mem_ops zynq_qspi_mem_ops = {
+	.supports_op = zynq_qspi_supports_op,
+	.exec_op = zynq_qspi_exec_mem_op,
+};
+
+/**
+ * zynq_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 and error value on failure
+ */
+static int zynq_qspi_probe(struct platform_device *pdev)
+{
+	int ret = 0;
+	struct spi_controller *ctlr;
+	struct device *dev = &pdev->dev;
+	struct device_node *np = dev->of_node;
+	struct zynq_qspi *xqspi;
+	struct resource *res;
+	u32 num_cs;
+
+	ctlr = spi_alloc_master(&pdev->dev, sizeof(*xqspi));
+	if (!ctlr)
+		return -ENOMEM;
+
+	xqspi = spi_controller_get_devdata(ctlr);
+	xqspi->dev = dev;
+	platform_set_drvdata(pdev, xqspi);
+	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->pclk = devm_clk_get(&pdev->dev, "pclk");
+	if (IS_ERR(xqspi->pclk)) {
+		dev_err(&pdev->dev, "pclk clock not found.\n");
+		ret = PTR_ERR(xqspi->pclk);
+		goto remove_master;
+	}
+
+	init_completion(&xqspi->data_completion);
+
+	xqspi->refclk = devm_clk_get(&pdev->dev, "ref_clk");
+	if (IS_ERR(xqspi->refclk)) {
+		dev_err(&pdev->dev, "ref_clk clock not found.\n");
+		ret = PTR_ERR(xqspi->refclk);
+		goto remove_master;
+	}
+
+	ret = clk_prepare_enable(xqspi->pclk);
+	if (ret) {
+		dev_err(&pdev->dev, "Unable to enable APB clock.\n");
+		goto remove_master;
+	}
+
+	ret = clk_prepare_enable(xqspi->refclk);
+	if (ret) {
+		dev_err(&pdev->dev, "Unable to enable device clock.\n");
+		goto clk_dis_pclk;
+	}
+
+	/* QSPI controller initializations */
+	zynq_qspi_init_hw(xqspi);
+
+	xqspi->irq = platform_get_irq(pdev, 0);
+	if (xqspi->irq <= 0) {
+		ret = -ENXIO;
+		dev_err(&pdev->dev, "irq resource not found\n");
+		goto remove_master;
+	}
+	ret = devm_request_irq(&pdev->dev, xqspi->irq, zynq_qspi_irq,
+			       0, pdev->name, xqspi);
+	if (ret != 0) {
+		ret = -ENXIO;
+		dev_err(&pdev->dev, "request_irq failed\n");
+		goto remove_master;
+	}
+
+	ret = of_property_read_u32(np, "num-cs",
+				   &num_cs);
+	if (ret < 0)
+		ctlr->num_chipselect = ZYNQ_QSPI_DEFAULT_NUM_CS;
+	else
+		ctlr->num_chipselect = num_cs;
+
+	ctlr->mode_bits =  SPI_RX_DUAL | SPI_RX_QUAD |
+			    SPI_TX_DUAL | SPI_TX_QUAD;
+	ctlr->mem_ops = &zynq_qspi_mem_ops;
+	ctlr->setup = zynq_qspi_setup_op;
+	ctlr->max_speed_hz = clk_get_rate(xqspi->refclk) / 2;
+	ctlr->dev.of_node = np;
+	ret = spi_register_controller(ctlr);
+	if (ret) {
+		dev_err(&pdev->dev, "spi_register_master failed\n");
+		goto clk_dis_all;
+	}
+
+	return ret;
+
+clk_dis_all:
+	clk_disable_unprepare(xqspi->refclk);
+clk_dis_pclk:
+	clk_disable_unprepare(xqspi->pclk);
+remove_master:
+	spi_controller_put(ctlr);
+
+	return ret;
+}
+
+/**
+ * zynq_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 on success and error value on failure
+ */
+static int zynq_qspi_remove(struct platform_device *pdev)
+{
+	struct zynq_qspi *xqspi = platform_get_drvdata(pdev);
+
+	zynq_qspi_write(xqspi, ZYNQ_QSPI_ENABLE_OFFSET, 0);
+
+	clk_disable_unprepare(xqspi->refclk);
+	clk_disable_unprepare(xqspi->pclk);
+
+	return 0;
+}
+
+static const struct of_device_id zynq_qspi_of_match[] = {
+	{ .compatible = "xlnx,zynq-qspi-1.0", },
+	{ /* end of table */ }
+};
+
+MODULE_DEVICE_TABLE(of, zynq_qspi_of_match);
+
+/*
+ * zynq_qspi_driver - This structure defines the QSPI platform driver
+ */
+static struct platform_driver zynq_qspi_driver = {
+	.probe = zynq_qspi_probe,
+	.remove = zynq_qspi_remove,
+	.driver = {
+		.name = "zynq-qspi",
+		.of_match_table = zynq_qspi_of_match,
+	},
+};
+
+module_platform_driver(zynq_qspi_driver);
+
+MODULE_AUTHOR("Xilinx, Inc.");
+MODULE_DESCRIPTION("Xilinx Zynq QSPI driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c
index 93986f879b09..5e75944ad5d1 100644
--- a/drivers/spi/spi.c
+++ b/drivers/spi/spi.c
@@ -36,6 +36,8 @@
 
 #define CREATE_TRACE_POINTS
 #include <trace/events/spi.h>
+EXPORT_TRACEPOINT_SYMBOL(spi_transfer_start);
+EXPORT_TRACEPOINT_SYMBOL(spi_transfer_stop);
 
 #include "internals.h"
 
@@ -1039,6 +1041,8 @@ static int spi_map_msg(struct spi_controller *ctlr, struct spi_message *msg)
 		if (max_tx || max_rx) {
 			list_for_each_entry(xfer, &msg->transfers,
 					    transfer_list) {
+				if (!xfer->len)
+					continue;
 				if (!xfer->tx_buf)
 					xfer->tx_buf = ctlr->dummy_tx;
 				if (!xfer->rx_buf)
@@ -1177,10 +1181,10 @@ out:
 	if (msg->status && ctlr->handle_err)
 		ctlr->handle_err(ctlr, msg);
 
-	spi_res_release(ctlr, msg);
-
 	spi_finalize_current_message(ctlr);
 
+	spi_res_release(ctlr, msg);
+
 	return ret;
 }
 
@@ -2195,6 +2199,8 @@ static int spi_get_gpio_descs(struct spi_controller *ctlr)
 		 */
 		cs[i] = devm_gpiod_get_index_optional(dev, "cs", i,
 						      GPIOD_OUT_LOW);
+		if (IS_ERR(cs[i]))
+			return PTR_ERR(cs[i]);
 
 		if (cs[i]) {
 			/*
@@ -2261,7 +2267,7 @@ int spi_register_controller(struct spi_controller *ctlr)
 {
 	struct device		*dev = ctlr->dev.parent;
 	struct boardinfo	*bi;
-	int			status = -ENODEV;
+	int			status;
 	int			id, first_dynamic;
 
 	if (!dev)
@@ -2275,24 +2281,6 @@ int spi_register_controller(struct spi_controller *ctlr)
 	if (status)
 		return status;
 
-	if (!spi_controller_is_slave(ctlr)) {
-		if (ctlr->use_gpio_descriptors) {
-			status = spi_get_gpio_descs(ctlr);
-			if (status)
-				return status;
-			/*
-			 * A controller using GPIO descriptors always
-			 * supports SPI_CS_HIGH if need be.
-			 */
-			ctlr->mode_bits |= SPI_CS_HIGH;
-		} else {
-			/* Legacy code path for GPIOs from DT */
-			status = of_spi_register_master(ctlr);
-			if (status)
-				return status;
-		}
-	}
-
 	/* even if it's just one always-selected device, there must
 	 * be at least one chipselect
 	 */
@@ -2349,6 +2337,25 @@ int spi_register_controller(struct spi_controller *ctlr)
 	 * registration fails if the bus ID is in use.
 	 */
 	dev_set_name(&ctlr->dev, "spi%u", ctlr->bus_num);
+
+	if (!spi_controller_is_slave(ctlr)) {
+		if (ctlr->use_gpio_descriptors) {
+			status = spi_get_gpio_descs(ctlr);
+			if (status)
+				return status;
+			/*
+			 * A controller using GPIO descriptors always
+			 * supports SPI_CS_HIGH if need be.
+			 */
+			ctlr->mode_bits |= SPI_CS_HIGH;
+		} else {
+			/* Legacy code path for GPIOs from DT */
+			status = of_spi_register_master(ctlr);
+			if (status)
+				return status;
+		}
+	}
+
 	status = device_add(&ctlr->dev);
 	if (status < 0) {
 		/* free bus id */
@@ -2781,11 +2788,6 @@ static int __spi_split_transfer_maxsize(struct spi_controller *ctlr,
 	size_t offset;
 	size_t count, i;
 
-	/* warn once about this fact that we are splitting a transfer */
-	dev_warn_once(&msg->spi->dev,
-		      "spi_transfer of length %i exceed max length of %zu - needed to split transfers\n",
-		      xfer->len, maxsize);
-
 	/* calculate how many we have to replace */
 	count = DIV_ROUND_UP(xfer->len, maxsize);
 
@@ -2943,6 +2945,11 @@ int spi_setup(struct spi_device *spi)
 	 * so it is ignored here.
 	 */
 	bad_bits = spi->mode & ~(spi->controller->mode_bits | SPI_CS_WORD);
+	/* nothing prevents from working with active-high CS in case if it
+	 * is driven by GPIO.
+	 */
+	if (gpio_is_valid(spi->cs_gpio))
+		bad_bits &= ~SPI_CS_HIGH;
 	ugly_bits = bad_bits &
 		    (SPI_TX_DUAL | SPI_TX_QUAD | SPI_TX_OCTAL |
 		     SPI_RX_DUAL | SPI_RX_QUAD | SPI_RX_OCTAL);
@@ -2988,6 +2995,21 @@ int spi_setup(struct spi_device *spi)
 }
 EXPORT_SYMBOL_GPL(spi_setup);
 
+/**
+ * spi_set_cs_timing - configure CS setup, hold, and inactive delays
+ * @spi: the device that requires specific CS timing configuration
+ * @setup: CS setup time in terms of clock count
+ * @hold: CS hold time in terms of clock count
+ * @inactive_dly: CS inactive delay between transfers in terms of clock count
+ */
+void spi_set_cs_timing(struct spi_device *spi, u8 setup, u8 hold,
+		       u8 inactive_dly)
+{
+	if (spi->controller->set_cs_timing)
+		spi->controller->set_cs_timing(spi, setup, hold, inactive_dly);
+}
+EXPORT_SYMBOL_GPL(spi_set_cs_timing);
+
 static int __spi_validate(struct spi_device *spi, struct spi_message *message)
 {
 	struct spi_controller *ctlr = spi->controller;
@@ -3062,8 +3084,6 @@ static int __spi_validate(struct spi_device *spi, struct spi_message *message)
 
 		if (!xfer->speed_hz)
 			xfer->speed_hz = spi->max_speed_hz;
-		if (!xfer->speed_hz)
-			xfer->speed_hz = ctlr->max_speed_hz;
 
 		if (ctlr->max_speed_hz && xfer->speed_hz > ctlr->max_speed_hz)
 			xfer->speed_hz = ctlr->max_speed_hz;
diff --git a/drivers/spi/spidev.c b/drivers/spi/spidev.c
index b0c76e2626ce..70966e10be7e 100644
--- a/drivers/spi/spidev.c
+++ b/drivers/spi/spidev.c
@@ -276,17 +276,19 @@ static int spidev_message(struct spidev_data *spidev,
 		k_tmp->bits_per_word = u_tmp->bits_per_word;
 		k_tmp->delay_usecs = u_tmp->delay_usecs;
 		k_tmp->speed_hz = u_tmp->speed_hz;
+		k_tmp->word_delay_usecs = u_tmp->word_delay_usecs;
 		if (!k_tmp->speed_hz)
 			k_tmp->speed_hz = spidev->speed_hz;
 #ifdef VERBOSE
 		dev_dbg(&spidev->spi->dev,
-			"  xfer len %u %s%s%s%dbits %u usec %uHz\n",
+			"  xfer len %u %s%s%s%dbits %u usec %u usec %uHz\n",
 			u_tmp->len,
 			u_tmp->rx_buf ? "rx " : "",
 			u_tmp->tx_buf ? "tx " : "",
 			u_tmp->cs_change ? "cs " : "",
 			u_tmp->bits_per_word ? : spidev->spi->bits_per_word,
 			u_tmp->delay_usecs,
+			u_tmp->word_delay_usecs,
 			u_tmp->speed_hz ? : spidev->spi->max_speed_hz);
 #endif
 		spi_message_add_tail(k_tmp, &msg);