Merge commit 'c104f1fa1ecf4ee0fc06e31b1f77630b2551be81' into stable/for-linus-3.4

* commit 'c104f1fa1ecf4ee0fc06e31b1f77630b2551be81': (14566 commits)
  cpufreq: OMAP: fix build errors: depends on ARCH_OMAP2PLUS
  sparc64: Eliminate obsolete __handle_softirq() function
  sparc64: Fix bootup crash on sun4v.
  kconfig: delete last traces of __enabled_ from autoconf.h
  Revert "kconfig: fix __enabled_ macros definition for invisible and un-selected symbols"
  kconfig: fix IS_ENABLED to not require all options to be defined
  irq_domain: fix type mismatch in debugfs output format
  staging: android: fix mem leaks in __persistent_ram_init()
  staging: vt6656: Don't leak memory in drivers/staging/vt6656/ioctl.c::private_ioctl()
  staging: iio: hmc5843: Fix crash in probe function.
  panic: fix stack dump print on direct call to panic()
  drivers/rtc/rtc-pl031.c: enable clock on all ST variants
  Revert "mm: vmscan: fix misused nr_reclaimed in shrink_mem_cgroup_zone()"
  hugetlb: fix race condition in hugetlb_fault()
  drivers/rtc/rtc-twl.c: use static register while reading time
  drivers/rtc/rtc-s3c.c: add placeholder for driver private data
  drivers/rtc/rtc-s3c.c: fix compilation error
  MAINTAINERS: add PCDP console maintainer
  memcg: do not open code accesses to res_counter members
  drivers/rtc/rtc-efi.c: fix section mismatch warning
  ...

44 files changed, 6266 insertions, 1347 deletions

diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig
index 5a99bb3f255a..cf9da362d64f 100644
--- a/drivers/dma/Kconfig
+++ b/drivers/dma/Kconfig
@@ -124,7 +124,7 @@ config MV_XOR
 
 config MX3_IPU
 	bool "MX3x Image Processing Unit support"
-	depends on SOC_IMX31 || SOC_IMX35
+	depends on ARCH_MXC
 	select DMA_ENGINE
 	default y
 	help
@@ -187,6 +187,13 @@ config TIMB_DMA
 	help
 	  Enable support for the Timberdale FPGA DMA engine.
 
+config SIRF_DMA
+	tristate "CSR SiRFprimaII DMA support"
+	depends on ARCH_PRIMA2
+	select DMA_ENGINE
+	help
+	  Enable support for the CSR SiRFprimaII DMA engine.
+
 config ARCH_HAS_ASYNC_TX_FIND_CHANNEL
 	bool
 
@@ -194,37 +201,36 @@ config PL330_DMA
 	tristate "DMA API Driver for PL330"
 	select DMA_ENGINE
 	depends on ARM_AMBA
-	select PL330
 	help
 	  Select if your platform has one or more PL330 DMACs.
 	  You need to provide platform specific settings via
 	  platform_data for a dma-pl330 device.
 
 config PCH_DMA
-	tristate "Intel EG20T PCH / OKI Semi IOH(ML7213/ML7223) DMA support"
+	tristate "Intel EG20T PCH / LAPIS Semicon IOH(ML7213/ML7223/ML7831) DMA"
 	depends on PCI && X86
 	select DMA_ENGINE
 	help
 	  Enable support for Intel EG20T PCH DMA engine.
 
-	  This driver also can be used for OKI SEMICONDUCTOR IOH(Input/
-	  Output Hub), ML7213 and ML7223.
-	  ML7213 IOH is for IVI(In-Vehicle Infotainment) use and ML7223 IOH is
-	  for MP(Media Phone) use.
-	  ML7213/ML7223 is companion chip for Intel Atom E6xx series.
-	  ML7213/ML7223 is completely compatible for Intel EG20T PCH.
+	  This driver also can be used for LAPIS Semiconductor IOH(Input/
+	  Output Hub), ML7213, ML7223 and ML7831.
+	  ML7213 IOH is for IVI(In-Vehicle Infotainment) use, ML7223 IOH is
+	  for MP(Media Phone) use and ML7831 IOH is for general purpose use.
+	  ML7213/ML7223/ML7831 is companion chip for Intel Atom E6xx series.
+	  ML7213/ML7223/ML7831 is completely compatible for Intel EG20T PCH.
 
 config IMX_SDMA
 	tristate "i.MX SDMA support"
-	depends on ARCH_MX25 || SOC_IMX31 || SOC_IMX35 || ARCH_MX5
+	depends on ARCH_MXC
 	select DMA_ENGINE
 	help
 	  Support the i.MX SDMA engine. This engine is integrated into
-	  Freescale i.MX25/31/35/51 chips.
+	  Freescale i.MX25/31/35/51/53 chips.
 
 config IMX_DMA
 	tristate "i.MX DMA support"
-	depends on IMX_HAVE_DMA_V1
+	depends on ARCH_MXC
 	select DMA_ENGINE
 	help
 	  Support the i.MX DMA engine. This engine is integrated into
@@ -245,6 +251,15 @@ config EP93XX_DMA
 	help
 	  Enable support for the Cirrus Logic EP93xx M2P/M2M DMA controller.
 
+config DMA_SA11X0
+	tristate "SA-11x0 DMA support"
+	depends on ARCH_SA1100
+	select DMA_ENGINE
+	help
+	  Support the DMA engine found on Intel StrongARM SA-1100 and
+	  SA-1110 SoCs.  This DMA engine can only be used with on-chip
+	  devices.
+
 config DMA_ENGINE
 	bool
 
diff --git a/drivers/dma/Makefile b/drivers/dma/Makefile
index 30cf3b1f0c5c..86b795baba98 100644
--- a/drivers/dma/Makefile
+++ b/drivers/dma/Makefile
@@ -21,8 +21,10 @@ obj-$(CONFIG_IMX_SDMA) += imx-sdma.o
 obj-$(CONFIG_IMX_DMA) += imx-dma.o
 obj-$(CONFIG_MXS_DMA) += mxs-dma.o
 obj-$(CONFIG_TIMB_DMA) += timb_dma.o
+obj-$(CONFIG_SIRF_DMA) += sirf-dma.o
 obj-$(CONFIG_STE_DMA40) += ste_dma40.o ste_dma40_ll.o
 obj-$(CONFIG_PL330_DMA) += pl330.o
 obj-$(CONFIG_PCH_DMA) += pch_dma.o
 obj-$(CONFIG_AMBA_PL08X) += amba-pl08x.o
 obj-$(CONFIG_EP93XX_DMA) += ep93xx_dma.o
+obj-$(CONFIG_DMA_SA11X0) += sa11x0-dma.o
diff --git a/drivers/dma/amba-pl08x.c b/drivers/dma/amba-pl08x.c
index 0698695e8bf9..c301a8ec31aa 100644
--- a/drivers/dma/amba-pl08x.c
+++ b/drivers/dma/amba-pl08x.c
@@ -85,6 +85,8 @@
 #include <linux/slab.h>
 #include <asm/hardware/pl080.h>
 
+#include "dmaengine.h"
+
 #define DRIVER_NAME	"pl08xdmac"
 
 static struct amba_driver pl08x_amba_driver;
@@ -649,7 +651,7 @@ static int pl08x_fill_llis_for_desc(struct pl08x_driver_data *pl08x,
 			}
 
 			if ((bd.srcbus.addr % bd.srcbus.buswidth) ||
-					(bd.srcbus.addr % bd.srcbus.buswidth)) {
+					(bd.dstbus.addr % bd.dstbus.buswidth)) {
 				dev_err(&pl08x->adev->dev,
 					"%s src & dst address must be aligned to src"
 					" & dst width if peripheral is flow controller",
@@ -854,8 +856,10 @@ static int prep_phy_channel(struct pl08x_dma_chan *plchan,
 	int ret;
 
 	/* Check if we already have a channel */
-	if (plchan->phychan)
-		return 0;
+	if (plchan->phychan) {
+		ch = plchan->phychan;
+		goto got_channel;
+	}
 
 	ch = pl08x_get_phy_channel(pl08x, plchan);
 	if (!ch) {
@@ -880,21 +884,22 @@ static int prep_phy_channel(struct pl08x_dma_chan *plchan,
 			return -EBUSY;
 		}
 		ch->signal = ret;
-
-		/* Assign the flow control signal to this channel */
-		if (txd->direction == DMA_TO_DEVICE)
-			txd->ccfg |= ch->signal << PL080_CONFIG_DST_SEL_SHIFT;
-		else if (txd->direction == DMA_FROM_DEVICE)
-			txd->ccfg |= ch->signal << PL080_CONFIG_SRC_SEL_SHIFT;
 	}
 
+	plchan->phychan = ch;
 	dev_dbg(&pl08x->adev->dev, "allocated physical channel %d and signal %d for xfer on %s\n",
 		 ch->id,
 		 ch->signal,
 		 plchan->name);
 
+got_channel:
+	/* Assign the flow control signal to this channel */
+	if (txd->direction == DMA_MEM_TO_DEV)
+		txd->ccfg |= ch->signal << PL080_CONFIG_DST_SEL_SHIFT;
+	else if (txd->direction == DMA_DEV_TO_MEM)
+		txd->ccfg |= ch->signal << PL080_CONFIG_SRC_SEL_SHIFT;
+
 	plchan->phychan_hold++;
-	plchan->phychan = ch;
 
 	return 0;
 }
@@ -916,13 +921,10 @@ static dma_cookie_t pl08x_tx_submit(struct dma_async_tx_descriptor *tx)
 	struct pl08x_dma_chan *plchan = to_pl08x_chan(tx->chan);
 	struct pl08x_txd *txd = to_pl08x_txd(tx);
 	unsigned long flags;
+	dma_cookie_t cookie;
 
 	spin_lock_irqsave(&plchan->lock, flags);
-
-	plchan->chan.cookie += 1;
-	if (plchan->chan.cookie < 0)
-		plchan->chan.cookie = 1;
-	tx->cookie = plchan->chan.cookie;
+	cookie = dma_cookie_assign(tx);
 
 	/* Put this onto the pending list */
 	list_add_tail(&txd->node, &plchan->pend_list);
@@ -942,7 +944,7 @@ static dma_cookie_t pl08x_tx_submit(struct dma_async_tx_descriptor *tx)
 
 	spin_unlock_irqrestore(&plchan->lock, flags);
 
-	return tx->cookie;
+	return cookie;
 }
 
 static struct dma_async_tx_descriptor *pl08x_prep_dma_interrupt(
@@ -962,31 +964,17 @@ static enum dma_status pl08x_dma_tx_status(struct dma_chan *chan,
 		dma_cookie_t cookie, struct dma_tx_state *txstate)
 {
 	struct pl08x_dma_chan *plchan = to_pl08x_chan(chan);
-	dma_cookie_t last_used;
-	dma_cookie_t last_complete;
 	enum dma_status ret;
-	u32 bytesleft = 0;
-
-	last_used = plchan->chan.cookie;
-	last_complete = plchan->lc;
 
-	ret = dma_async_is_complete(cookie, last_complete, last_used);
-	if (ret == DMA_SUCCESS) {
-		dma_set_tx_state(txstate, last_complete, last_used, 0);
+	ret = dma_cookie_status(chan, cookie, txstate);
+	if (ret == DMA_SUCCESS)
 		return ret;
-	}
 
 	/*
 	 * This cookie not complete yet
+	 * Get number of bytes left in the active transactions and queue
 	 */
-	last_used = plchan->chan.cookie;
-	last_complete = plchan->lc;
-
-	/* Get number of bytes left in the active transactions and queue */
-	bytesleft = pl08x_getbytes_chan(plchan);
-
-	dma_set_tx_state(txstate, last_complete, last_used,
-			 bytesleft);
+	dma_set_residue(txstate, pl08x_getbytes_chan(plchan));
 
 	if (plchan->state == PL08X_CHAN_PAUSED)
 		return DMA_PAUSED;
@@ -1102,10 +1090,10 @@ static int dma_set_runtime_config(struct dma_chan *chan,
 
 	/* Transfer direction */
 	plchan->runtime_direction = config->direction;
-	if (config->direction == DMA_TO_DEVICE) {
+	if (config->direction == DMA_MEM_TO_DEV) {
 		addr_width = config->dst_addr_width;
 		maxburst = config->dst_maxburst;
-	} else if (config->direction == DMA_FROM_DEVICE) {
+	} else if (config->direction == DMA_DEV_TO_MEM) {
 		addr_width = config->src_addr_width;
 		maxburst = config->src_maxburst;
 	} else {
@@ -1136,7 +1124,9 @@ static int dma_set_runtime_config(struct dma_chan *chan,
 	cctl |= burst << PL080_CONTROL_SB_SIZE_SHIFT;
 	cctl |= burst << PL080_CONTROL_DB_SIZE_SHIFT;
 
-	if (plchan->runtime_direction == DMA_FROM_DEVICE) {
+	plchan->device_fc = config->device_fc;
+
+	if (plchan->runtime_direction == DMA_DEV_TO_MEM) {
 		plchan->src_addr = config->src_addr;
 		plchan->src_cctl = pl08x_cctl(cctl) | PL080_CONTROL_DST_INCR |
 			pl08x_select_bus(plchan->cd->periph_buses,
@@ -1152,7 +1142,7 @@ static int dma_set_runtime_config(struct dma_chan *chan,
 		"configured channel %s (%s) for %s, data width %d, "
 		"maxburst %d words, LE, CCTL=0x%08x\n",
 		dma_chan_name(chan), plchan->name,
-		(config->direction == DMA_FROM_DEVICE) ? "RX" : "TX",
+		(config->direction == DMA_DEV_TO_MEM) ? "RX" : "TX",
 		addr_width,
 		maxburst,
 		cctl);
@@ -1322,8 +1312,8 @@ static struct dma_async_tx_descriptor *pl08x_prep_dma_memcpy(
 
 static struct dma_async_tx_descriptor *pl08x_prep_slave_sg(
 		struct dma_chan *chan, struct scatterlist *sgl,
-		unsigned int sg_len, enum dma_data_direction direction,
-		unsigned long flags)
+		unsigned int sg_len, enum dma_transfer_direction direction,
+		unsigned long flags, void *context)
 {
 	struct pl08x_dma_chan *plchan = to_pl08x_chan(chan);
 	struct pl08x_driver_data *pl08x = plchan->host;
@@ -1354,10 +1344,10 @@ static struct dma_async_tx_descriptor *pl08x_prep_slave_sg(
 	 */
 	txd->direction = direction;
 
-	if (direction == DMA_TO_DEVICE) {
+	if (direction == DMA_MEM_TO_DEV) {
 		txd->cctl = plchan->dst_cctl;
 		slave_addr = plchan->dst_addr;
-	} else if (direction == DMA_FROM_DEVICE) {
+	} else if (direction == DMA_DEV_TO_MEM) {
 		txd->cctl = plchan->src_cctl;
 		slave_addr = plchan->src_addr;
 	} else {
@@ -1367,11 +1357,11 @@ static struct dma_async_tx_descriptor *pl08x_prep_slave_sg(
 		return NULL;
 	}
 
-	if (plchan->cd->device_fc)
-		tmp = (direction == DMA_TO_DEVICE) ? PL080_FLOW_MEM2PER_PER :
+	if (plchan->device_fc)
+		tmp = (direction == DMA_MEM_TO_DEV) ? PL080_FLOW_MEM2PER_PER :
 			PL080_FLOW_PER2MEM_PER;
 	else
-		tmp = (direction == DMA_TO_DEVICE) ? PL080_FLOW_MEM2PER :
+		tmp = (direction == DMA_MEM_TO_DEV) ? PL080_FLOW_MEM2PER :
 			PL080_FLOW_PER2MEM;
 
 	txd->ccfg |= tmp << PL080_CONFIG_FLOW_CONTROL_SHIFT;
@@ -1387,7 +1377,7 @@ static struct dma_async_tx_descriptor *pl08x_prep_slave_sg(
 		list_add_tail(&dsg->node, &txd->dsg_list);
 
 		dsg->len = sg_dma_len(sg);
-		if (direction == DMA_TO_DEVICE) {
+		if (direction == DMA_MEM_TO_DEV) {
 			dsg->src_addr = sg_phys(sg);
 			dsg->dst_addr = slave_addr;
 		} else {
@@ -1538,7 +1528,7 @@ static void pl08x_tasklet(unsigned long data)
 
 	if (txd) {
 		/* Update last completed */
-		plchan->lc = txd->tx.cookie;
+		dma_cookie_complete(&txd->tx);
 	}
 
 	/* If a new descriptor is queued, set it up plchan->at is NULL here */
@@ -1719,8 +1709,7 @@ static int pl08x_dma_init_virtual_channels(struct pl08x_driver_data *pl08x,
 			 chan->name);
 
 		chan->chan.device = dmadev;
-		chan->chan.cookie = 0;
-		chan->lc = 0;
+		dma_cookie_init(&chan->chan);
 
 		spin_lock_init(&chan->lock);
 		INIT_LIST_HEAD(&chan->pend_list);
diff --git a/drivers/dma/at_hdmac.c b/drivers/dma/at_hdmac.c
index fcfa0a8b5c59..7aa58d204892 100644
--- a/drivers/dma/at_hdmac.c
+++ b/drivers/dma/at_hdmac.c
@@ -23,8 +23,11 @@
 #include <linux/module.h>
 #include <linux/platform_device.h>
 #include <linux/slab.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
 
 #include "at_hdmac_regs.h"
+#include "dmaengine.h"
 
 /*
  * Glossary
@@ -190,27 +193,6 @@ static void atc_desc_chain(struct at_desc **first, struct at_desc **prev,
 }
 
 /**
- * atc_assign_cookie - compute and assign new cookie
- * @atchan: channel we work on
- * @desc: descriptor to assign cookie for
- *
- * Called with atchan->lock held and bh disabled
- */
-static dma_cookie_t
-atc_assign_cookie(struct at_dma_chan *atchan, struct at_desc *desc)
-{
-	dma_cookie_t cookie = atchan->chan_common.cookie;
-
-	if (++cookie < 0)
-		cookie = 1;
-
-	atchan->chan_common.cookie = cookie;
-	desc->txd.cookie = cookie;
-
-	return cookie;
-}
-
-/**
  * atc_dostart - starts the DMA engine for real
  * @atchan: the channel we want to start
  * @first: first descriptor in the list we want to begin with
@@ -267,7 +249,7 @@ atc_chain_complete(struct at_dma_chan *atchan, struct at_desc *desc)
 	dev_vdbg(chan2dev(&atchan->chan_common),
 		"descriptor %u complete\n", txd->cookie);
 
-	atchan->completed_cookie = txd->cookie;
+	dma_cookie_complete(txd);
 
 	/* move children to free_list */
 	list_splice_init(&desc->tx_list, &atchan->free_list);
@@ -545,7 +527,7 @@ static dma_cookie_t atc_tx_submit(struct dma_async_tx_descriptor *tx)
 	unsigned long		flags;
 
 	spin_lock_irqsave(&atchan->lock, flags);
-	cookie = atc_assign_cookie(atchan, desc);
+	cookie = dma_cookie_assign(tx);
 
 	if (list_empty(&atchan->active_list)) {
 		dev_vdbg(chan2dev(tx->chan), "tx_submit: started %u\n",
@@ -657,14 +639,16 @@ err_desc_get:
  * @sg_len: number of entries in @scatterlist
  * @direction: DMA direction
  * @flags: tx descriptor status flags
+ * @context: transaction context (ignored)
  */
 static struct dma_async_tx_descriptor *
 atc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
-		unsigned int sg_len, enum dma_data_direction direction,
-		unsigned long flags)
+		unsigned int sg_len, enum dma_transfer_direction direction,
+		unsigned long flags, void *context)
 {
 	struct at_dma_chan	*atchan = to_at_dma_chan(chan);
 	struct at_dma_slave	*atslave = chan->private;
+	struct dma_slave_config	*sconfig = &atchan->dma_sconfig;
 	struct at_desc		*first = NULL;
 	struct at_desc		*prev = NULL;
 	u32			ctrla;
@@ -678,7 +662,7 @@ atc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 
 	dev_vdbg(chan2dev(chan), "prep_slave_sg (%d): %s f0x%lx\n",
 			sg_len,
-			direction == DMA_TO_DEVICE ? "TO DEVICE" : "FROM DEVICE",
+			direction == DMA_MEM_TO_DEV ? "TO DEVICE" : "FROM DEVICE",
 			flags);
 
 	if (unlikely(!atslave || !sg_len)) {
@@ -686,19 +670,18 @@ atc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 		return NULL;
 	}
 
-	reg_width = atslave->reg_width;
-
 	ctrla = ATC_DEFAULT_CTRLA | atslave->ctrla;
 	ctrlb = ATC_IEN;
 
 	switch (direction) {
-	case DMA_TO_DEVICE:
+	case DMA_MEM_TO_DEV:
+		reg_width = convert_buswidth(sconfig->dst_addr_width);
 		ctrla |=  ATC_DST_WIDTH(reg_width);
 		ctrlb |=  ATC_DST_ADDR_MODE_FIXED
 			| ATC_SRC_ADDR_MODE_INCR
 			| ATC_FC_MEM2PER
 			| ATC_SIF(AT_DMA_MEM_IF) | ATC_DIF(AT_DMA_PER_IF);
-		reg = atslave->tx_reg;
+		reg = sconfig->dst_addr;
 		for_each_sg(sgl, sg, sg_len, i) {
 			struct at_desc	*desc;
 			u32		len;
@@ -725,14 +708,15 @@ atc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 			total_len += len;
 		}
 		break;
-	case DMA_FROM_DEVICE:
+	case DMA_DEV_TO_MEM:
+		reg_width = convert_buswidth(sconfig->src_addr_width);
 		ctrla |=  ATC_SRC_WIDTH(reg_width);
 		ctrlb |=  ATC_DST_ADDR_MODE_INCR
 			| ATC_SRC_ADDR_MODE_FIXED
 			| ATC_FC_PER2MEM
 			| ATC_SIF(AT_DMA_PER_IF) | ATC_DIF(AT_DMA_MEM_IF);
 
-		reg = atslave->rx_reg;
+		reg = sconfig->src_addr;
 		for_each_sg(sgl, sg, sg_len, i) {
 			struct at_desc	*desc;
 			u32		len;
@@ -787,7 +771,7 @@ err_desc_get:
  */
 static int
 atc_dma_cyclic_check_values(unsigned int reg_width, dma_addr_t buf_addr,
-		size_t period_len, enum dma_data_direction direction)
+		size_t period_len, enum dma_transfer_direction direction)
 {
 	if (period_len > (ATC_BTSIZE_MAX << reg_width))
 		goto err_out;
@@ -795,7 +779,7 @@ atc_dma_cyclic_check_values(unsigned int reg_width, dma_addr_t buf_addr,
 		goto err_out;
 	if (unlikely(buf_addr & ((1 << reg_width) - 1)))
 		goto err_out;
-	if (unlikely(!(direction & (DMA_TO_DEVICE | DMA_FROM_DEVICE))))
+	if (unlikely(!(direction & (DMA_DEV_TO_MEM | DMA_MEM_TO_DEV))))
 		goto err_out;
 
 	return 0;
@@ -808,12 +792,15 @@ err_out:
  * atc_dma_cyclic_fill_desc - Fill one period decriptor
  */
 static int
-atc_dma_cyclic_fill_desc(struct at_dma_slave *atslave, struct at_desc *desc,
+atc_dma_cyclic_fill_desc(struct dma_chan *chan, struct at_desc *desc,
 		unsigned int period_index, dma_addr_t buf_addr,
-		size_t period_len, enum dma_data_direction direction)
+		unsigned int reg_width, size_t period_len,
+		enum dma_transfer_direction direction)
 {
-	u32		ctrla;
-	unsigned int	reg_width = atslave->reg_width;
+	struct at_dma_chan	*atchan = to_at_dma_chan(chan);
+	struct at_dma_slave	*atslave = chan->private;
+	struct dma_slave_config	*sconfig = &atchan->dma_sconfig;
+	u32			ctrla;
 
 	/* prepare common CRTLA value */
 	ctrla =   ATC_DEFAULT_CTRLA | atslave->ctrla
@@ -822,9 +809,9 @@ atc_dma_cyclic_fill_desc(struct at_dma_slave *atslave, struct at_desc *desc,
 		| period_len >> reg_width;
 
 	switch (direction) {
-	case DMA_TO_DEVICE:
+	case DMA_MEM_TO_DEV:
 		desc->lli.saddr = buf_addr + (period_len * period_index);
-		desc->lli.daddr = atslave->tx_reg;
+		desc->lli.daddr = sconfig->dst_addr;
 		desc->lli.ctrla = ctrla;
 		desc->lli.ctrlb = ATC_DST_ADDR_MODE_FIXED
 				| ATC_SRC_ADDR_MODE_INCR
@@ -833,8 +820,8 @@ atc_dma_cyclic_fill_desc(struct at_dma_slave *atslave, struct at_desc *desc,
 				| ATC_DIF(AT_DMA_PER_IF);
 		break;
 
-	case DMA_FROM_DEVICE:
-		desc->lli.saddr = atslave->rx_reg;
+	case DMA_DEV_TO_MEM:
+		desc->lli.saddr = sconfig->src_addr;
 		desc->lli.daddr = buf_addr + (period_len * period_index);
 		desc->lli.ctrla = ctrla;
 		desc->lli.ctrlb = ATC_DST_ADDR_MODE_INCR
@@ -858,21 +845,25 @@ atc_dma_cyclic_fill_desc(struct at_dma_slave *atslave, struct at_desc *desc,
  * @buf_len: total number of bytes for the entire buffer
  * @period_len: number of bytes for each period
  * @direction: transfer direction, to or from device
+ * @context: transfer context (ignored)
  */
 static struct dma_async_tx_descriptor *
 atc_prep_dma_cyclic(struct dma_chan *chan, dma_addr_t buf_addr, size_t buf_len,
-		size_t period_len, enum dma_data_direction direction)
+		size_t period_len, enum dma_transfer_direction direction,
+		void *context)
 {
 	struct at_dma_chan	*atchan = to_at_dma_chan(chan);
 	struct at_dma_slave	*atslave = chan->private;
+	struct dma_slave_config	*sconfig = &atchan->dma_sconfig;
 	struct at_desc		*first = NULL;
 	struct at_desc		*prev = NULL;
 	unsigned long		was_cyclic;
+	unsigned int		reg_width;
 	unsigned int		periods = buf_len / period_len;
 	unsigned int		i;
 
 	dev_vdbg(chan2dev(chan), "prep_dma_cyclic: %s buf@0x%08x - %d (%d/%d)\n",
-			direction == DMA_TO_DEVICE ? "TO DEVICE" : "FROM DEVICE",
+			direction == DMA_MEM_TO_DEV ? "TO DEVICE" : "FROM DEVICE",
 			buf_addr,
 			periods, buf_len, period_len);
 
@@ -887,8 +878,13 @@ atc_prep_dma_cyclic(struct dma_chan *chan, dma_addr_t buf_addr, size_t buf_len,
 		return NULL;
 	}
 
+	if (sconfig->direction == DMA_MEM_TO_DEV)
+		reg_width = convert_buswidth(sconfig->dst_addr_width);
+	else
+		reg_width = convert_buswidth(sconfig->src_addr_width);
+
 	/* Check for too big/unaligned periods and unaligned DMA buffer */
-	if (atc_dma_cyclic_check_values(atslave->reg_width, buf_addr,
+	if (atc_dma_cyclic_check_values(reg_width, buf_addr,
 					period_len, direction))
 		goto err_out;
 
@@ -900,8 +896,8 @@ atc_prep_dma_cyclic(struct dma_chan *chan, dma_addr_t buf_addr, size_t buf_len,
 		if (!desc)
 			goto err_desc_get;
 
-		if (atc_dma_cyclic_fill_desc(atslave, desc, i, buf_addr,
-						period_len, direction))
+		if (atc_dma_cyclic_fill_desc(chan, desc, i, buf_addr,
+					     reg_width, period_len, direction))
 			goto err_desc_get;
 
 		atc_desc_chain(&first, &prev, desc);
@@ -924,6 +920,23 @@ err_out:
 	return NULL;
 }
 
+static int set_runtime_config(struct dma_chan *chan,
+			      struct dma_slave_config *sconfig)
+{
+	struct at_dma_chan	*atchan = to_at_dma_chan(chan);
+
+	/* Check if it is chan is configured for slave transfers */
+	if (!chan->private)
+		return -EINVAL;
+
+	memcpy(&atchan->dma_sconfig, sconfig, sizeof(*sconfig));
+
+	convert_burst(&atchan->dma_sconfig.src_maxburst);
+	convert_burst(&atchan->dma_sconfig.dst_maxburst);
+
+	return 0;
+}
+
 
 static int atc_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
 		       unsigned long arg)
@@ -984,6 +997,8 @@ static int atc_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
 		clear_bit(ATC_IS_CYCLIC, &atchan->status);
 
 		spin_unlock_irqrestore(&atchan->lock, flags);
+	} else if (cmd == DMA_SLAVE_CONFIG) {
+		return set_runtime_config(chan, (struct dma_slave_config *)arg);
 	} else {
 		return -ENXIO;
 	}
@@ -1014,26 +1029,20 @@ atc_tx_status(struct dma_chan *chan,
 
 	spin_lock_irqsave(&atchan->lock, flags);
 
-	last_complete = atchan->completed_cookie;
-	last_used = chan->cookie;
-
-	ret = dma_async_is_complete(cookie, last_complete, last_used);
+	ret = dma_cookie_status(chan, cookie, txstate);
 	if (ret != DMA_SUCCESS) {
 		atc_cleanup_descriptors(atchan);
 
-		last_complete = atchan->completed_cookie;
-		last_used = chan->cookie;
-
-		ret = dma_async_is_complete(cookie, last_complete, last_used);
+		ret = dma_cookie_status(chan, cookie, txstate);
 	}
 
+	last_complete = chan->completed_cookie;
+	last_used = chan->cookie;
+
 	spin_unlock_irqrestore(&atchan->lock, flags);
 
 	if (ret != DMA_SUCCESS)
-		dma_set_tx_state(txstate, last_complete, last_used,
-			atc_first_active(atchan)->len);
-	else
-		dma_set_tx_state(txstate, last_complete, last_used, 0);
+		dma_set_residue(txstate, atc_first_active(atchan)->len);
 
 	if (atc_chan_is_paused(atchan))
 		ret = DMA_PAUSED;
@@ -1127,7 +1136,7 @@ static int atc_alloc_chan_resources(struct dma_chan *chan)
 	spin_lock_irqsave(&atchan->lock, flags);
 	atchan->descs_allocated = i;
 	list_splice(&tmp_list, &atchan->free_list);
-	atchan->completed_cookie = chan->cookie = 1;
+	dma_cookie_init(chan);
 	spin_unlock_irqrestore(&atchan->lock, flags);
 
 	/* channel parameters */
@@ -1175,6 +1184,56 @@ static void atc_free_chan_resources(struct dma_chan *chan)
 
 /*--  Module Management  -----------------------------------------------*/
 
+/* cap_mask is a multi-u32 bitfield, fill it with proper C code. */
+static struct at_dma_platform_data at91sam9rl_config = {
+	.nr_channels = 2,
+};
+static struct at_dma_platform_data at91sam9g45_config = {
+	.nr_channels = 8,
+};
+
+#if defined(CONFIG_OF)
+static const struct of_device_id atmel_dma_dt_ids[] = {
+	{
+		.compatible = "atmel,at91sam9rl-dma",
+		.data = &at91sam9rl_config,
+	}, {
+		.compatible = "atmel,at91sam9g45-dma",
+		.data = &at91sam9g45_config,
+	}, {
+		/* sentinel */
+	}
+};
+
+MODULE_DEVICE_TABLE(of, atmel_dma_dt_ids);
+#endif
+
+static const struct platform_device_id atdma_devtypes[] = {
+	{
+		.name = "at91sam9rl_dma",
+		.driver_data = (unsigned long) &at91sam9rl_config,
+	}, {
+		.name = "at91sam9g45_dma",
+		.driver_data = (unsigned long) &at91sam9g45_config,
+	}, {
+		/* sentinel */
+	}
+};
+
+static inline struct at_dma_platform_data * __init at_dma_get_driver_data(
+						struct platform_device *pdev)
+{
+	if (pdev->dev.of_node) {
+		const struct of_device_id *match;
+		match = of_match_node(atmel_dma_dt_ids, pdev->dev.of_node);
+		if (match == NULL)
+			return NULL;
+		return match->data;
+	}
+	return (struct at_dma_platform_data *)
+			platform_get_device_id(pdev)->driver_data;
+}
+
 /**
  * at_dma_off - disable DMA controller
  * @atdma: the Atmel HDAMC device
@@ -1193,18 +1252,23 @@ static void at_dma_off(struct at_dma *atdma)
 
 static int __init at_dma_probe(struct platform_device *pdev)
 {
-	struct at_dma_platform_data *pdata;
 	struct resource		*io;
 	struct at_dma		*atdma;
 	size_t			size;
 	int			irq;
 	int			err;
 	int			i;
+	struct at_dma_platform_data *plat_dat;
 
-	/* get DMA Controller parameters from platform */
-	pdata = pdev->dev.platform_data;
-	if (!pdata || pdata->nr_channels > AT_DMA_MAX_NR_CHANNELS)
-		return -EINVAL;
+	/* setup platform data for each SoC */
+	dma_cap_set(DMA_MEMCPY, at91sam9rl_config.cap_mask);
+	dma_cap_set(DMA_MEMCPY, at91sam9g45_config.cap_mask);
+	dma_cap_set(DMA_SLAVE, at91sam9g45_config.cap_mask);
+
+	/* get DMA parameters from controller type */
+	plat_dat = at_dma_get_driver_data(pdev);
+	if (!plat_dat)
+		return -ENODEV;
 
 	io = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	if (!io)
@@ -1215,14 +1279,14 @@ static int __init at_dma_probe(struct platform_device *pdev)
 		return irq;
 
 	size = sizeof(struct at_dma);
-	size += pdata->nr_channels * sizeof(struct at_dma_chan);
+	size += plat_dat->nr_channels * sizeof(struct at_dma_chan);
 	atdma = kzalloc(size, GFP_KERNEL);
 	if (!atdma)
 		return -ENOMEM;
 
-	/* discover transaction capabilites from the platform data */
-	atdma->dma_common.cap_mask = pdata->cap_mask;
-	atdma->all_chan_mask = (1 << pdata->nr_channels) - 1;
+	/* discover transaction capabilities */
+	atdma->dma_common.cap_mask = plat_dat->cap_mask;
+	atdma->all_chan_mask = (1 << plat_dat->nr_channels) - 1;
 
 	size = resource_size(io);
 	if (!request_mem_region(io->start, size, pdev->dev.driver->name)) {
@@ -1268,11 +1332,11 @@ static int __init at_dma_probe(struct platform_device *pdev)
 
 	/* initialize channels related values */
 	INIT_LIST_HEAD(&atdma->dma_common.channels);
-	for (i = 0; i < pdata->nr_channels; i++) {
+	for (i = 0; i < plat_dat->nr_channels; i++) {
 		struct at_dma_chan	*atchan = &atdma->chan[i];
 
 		atchan->chan_common.device = &atdma->dma_common;
-		atchan->chan_common.cookie = atchan->completed_cookie = 1;
+		dma_cookie_init(&atchan->chan_common);
 		list_add_tail(&atchan->chan_common.device_node,
 				&atdma->dma_common.channels);
 
@@ -1286,7 +1350,7 @@ static int __init at_dma_probe(struct platform_device *pdev)
 
 		tasklet_init(&atchan->tasklet, atc_tasklet,
 				(unsigned long)atchan);
-		atc_enable_irq(atchan);
+		atc_enable_chan_irq(atdma, i);
 	}
 
 	/* set base routines */
@@ -1313,7 +1377,7 @@ static int __init at_dma_probe(struct platform_device *pdev)
 	dev_info(&pdev->dev, "Atmel AHB DMA Controller ( %s%s), %d channels\n",
 	  dma_has_cap(DMA_MEMCPY, atdma->dma_common.cap_mask) ? "cpy " : "",
 	  dma_has_cap(DMA_SLAVE, atdma->dma_common.cap_mask)  ? "slave " : "",
-	  pdata->nr_channels);
+	  plat_dat->nr_channels);
 
 	dma_async_device_register(&atdma->dma_common);
 
@@ -1353,7 +1417,7 @@ static int __exit at_dma_remove(struct platform_device *pdev)
 		struct at_dma_chan	*atchan = to_at_dma_chan(chan);
 
 		/* Disable interrupts */
-		atc_disable_irq(atchan);
+		atc_disable_chan_irq(atdma, chan->chan_id);
 		tasklet_disable(&atchan->tasklet);
 
 		tasklet_kill(&atchan->tasklet);
@@ -1495,9 +1559,11 @@ static const struct dev_pm_ops at_dma_dev_pm_ops = {
 static struct platform_driver at_dma_driver = {
 	.remove		= __exit_p(at_dma_remove),
 	.shutdown	= at_dma_shutdown,
+	.id_table	= atdma_devtypes,
 	.driver = {
 		.name	= "at_hdmac",
 		.pm	= &at_dma_dev_pm_ops,
+		.of_match_table	= of_match_ptr(atmel_dma_dt_ids),
 	},
 };
 
diff --git a/drivers/dma/at_hdmac_regs.h b/drivers/dma/at_hdmac_regs.h
index aa4c9aebab7c..897a8bcaec90 100644
--- a/drivers/dma/at_hdmac_regs.h
+++ b/drivers/dma/at_hdmac_regs.h
@@ -207,8 +207,8 @@ enum atc_status {
  * @save_cfg: configuration register that is saved on suspend/resume cycle
  * @save_dscr: for cyclic operations, preserve next descriptor address in
  *             the cyclic list on suspend/resume cycle
+ * @dma_sconfig: configuration for slave transfers, passed via DMA_SLAVE_CONFIG
  * @lock: serializes enqueue/dequeue operations to descriptors lists
- * @completed_cookie: identifier for the most recently completed operation
  * @active_list: list of descriptors dmaengine is being running on
  * @queue: list of descriptors ready to be submitted to engine
  * @free_list: list of descriptors usable by the channel
@@ -223,11 +223,11 @@ struct at_dma_chan {
 	struct tasklet_struct	tasklet;
 	u32			save_cfg;
 	u32			save_dscr;
+	struct dma_slave_config dma_sconfig;
 
 	spinlock_t		lock;
 
 	/* these other elements are all protected by lock */
-	dma_cookie_t		completed_cookie;
 	struct list_head	active_list;
 	struct list_head	queue;
 	struct list_head	free_list;
@@ -245,12 +245,43 @@ static inline struct at_dma_chan *to_at_dma_chan(struct dma_chan *dchan)
 	return container_of(dchan, struct at_dma_chan, chan_common);
 }
 
+/*
+ * Fix sconfig's burst size according to at_hdmac. We need to convert them as:
+ * 1 -> 0, 4 -> 1, 8 -> 2, 16 -> 3, 32 -> 4, 64 -> 5, 128 -> 6, 256 -> 7.
+ *
+ * This can be done by finding most significant bit set.
+ */
+static inline void convert_burst(u32 *maxburst)
+{
+	if (*maxburst > 1)
+		*maxburst = fls(*maxburst) - 2;
+	else
+		*maxburst = 0;
+}
+
+/*
+ * Fix sconfig's bus width according to at_hdmac.
+ * 1 byte -> 0, 2 bytes -> 1, 4 bytes -> 2.
+ */
+static inline u8 convert_buswidth(enum dma_slave_buswidth addr_width)
+{
+	switch (addr_width) {
+	case DMA_SLAVE_BUSWIDTH_2_BYTES:
+		return 1;
+	case DMA_SLAVE_BUSWIDTH_4_BYTES:
+		return 2;
+	default:
+		/* For 1 byte width or fallback */
+		return 0;
+	}
+}
 
 /*--  Controller  ------------------------------------------------------*/
 
 /**
  * struct at_dma - internal representation of an Atmel HDMA Controller
  * @chan_common: common dmaengine dma_device object members
+ * @atdma_devtype: identifier of DMA controller compatibility
  * @ch_regs: memory mapped register base
  * @clk: dma controller clock
  * @save_imr: interrupt mask register that is saved on suspend/resume cycle
@@ -326,28 +357,27 @@ static void atc_dump_lli(struct at_dma_chan *atchan, struct at_lli *lli)
 }
 
 
-static void atc_setup_irq(struct at_dma_chan *atchan, int on)
+static void atc_setup_irq(struct at_dma *atdma, int chan_id, int on)
 {
-	struct at_dma	*atdma = to_at_dma(atchan->chan_common.device);
-	u32		ebci;
+	u32 ebci;
 
 	/* enable interrupts on buffer transfer completion & error */
-	ebci =    AT_DMA_BTC(atchan->chan_common.chan_id)
-		| AT_DMA_ERR(atchan->chan_common.chan_id);
+	ebci =    AT_DMA_BTC(chan_id)
+		| AT_DMA_ERR(chan_id);
 	if (on)
 		dma_writel(atdma, EBCIER, ebci);
 	else
 		dma_writel(atdma, EBCIDR, ebci);
 }
 
-static inline void atc_enable_irq(struct at_dma_chan *atchan)
+static void atc_enable_chan_irq(struct at_dma *atdma, int chan_id)
 {
-	atc_setup_irq(atchan, 1);
+	atc_setup_irq(atdma, chan_id, 1);
 }
 
-static inline void atc_disable_irq(struct at_dma_chan *atchan)
+static void atc_disable_chan_irq(struct at_dma *atdma, int chan_id)
 {
-	atc_setup_irq(atchan, 0);
+	atc_setup_irq(atdma, chan_id, 0);
 }
 
 
diff --git a/drivers/dma/coh901318.c b/drivers/dma/coh901318.c
index 4234f416ef11..750925f9638b 100644
--- a/drivers/dma/coh901318.c
+++ b/drivers/dma/coh901318.c
@@ -24,6 +24,7 @@
 #include <mach/coh901318.h>
 
 #include "coh901318_lli.h"
+#include "dmaengine.h"
 
 #define COHC_2_DEV(cohc) (&cohc->chan.dev->device)
 
@@ -39,7 +40,7 @@ struct coh901318_desc {
 	struct scatterlist *sg;
 	unsigned int sg_len;
 	struct coh901318_lli *lli;
-	enum dma_data_direction dir;
+	enum dma_transfer_direction dir;
 	unsigned long flags;
 	u32 head_config;
 	u32 head_ctrl;
@@ -59,7 +60,6 @@ struct coh901318_base {
 struct coh901318_chan {
 	spinlock_t lock;
 	int allocated;
-	int completed;
 	int id;
 	int stopped;
 
@@ -104,13 +104,6 @@ static void coh901318_list_print(struct coh901318_chan *cohc,
 static struct coh901318_base *debugfs_dma_base;
 static struct dentry *dma_dentry;
 
-static int coh901318_debugfs_open(struct inode *inode, struct file *file)
-{
-
-	file->private_data = inode->i_private;
-	return 0;
-}
-
 static int coh901318_debugfs_read(struct file *file, char __user *buf,
 				  size_t count, loff_t *f_pos)
 {
@@ -158,7 +151,7 @@ static int coh901318_debugfs_read(struct file *file, char __user *buf,
 
 static const struct file_operations coh901318_debugfs_status_operations = {
 	.owner		= THIS_MODULE,
-	.open		= coh901318_debugfs_open,
+	.open		= simple_open,
 	.read		= coh901318_debugfs_read,
 	.llseek		= default_llseek,
 };
@@ -318,20 +311,6 @@ static int coh901318_prep_linked_list(struct coh901318_chan *cohc,
 
 	return 0;
 }
-static dma_cookie_t
-coh901318_assign_cookie(struct coh901318_chan *cohc,
-			struct coh901318_desc *cohd)
-{
-	dma_cookie_t cookie = cohc->chan.cookie;
-
-	if (++cookie < 0)
-		cookie = 1;
-
-	cohc->chan.cookie = cookie;
-	cohd->desc.cookie = cookie;
-
-	return cookie;
-}
 
 static struct coh901318_desc *
 coh901318_desc_get(struct coh901318_chan *cohc)
@@ -705,7 +684,7 @@ static void dma_tasklet(unsigned long data)
 	callback_param = cohd_fin->desc.callback_param;
 
 	/* sign this job as completed on the channel */
-	cohc->completed = cohd_fin->desc.cookie;
+	dma_cookie_complete(&cohd_fin->desc);
 
 	/* release the lli allocation and remove the descriptor */
 	coh901318_lli_free(&cohc->base->pool, &cohd_fin->lli);
@@ -929,7 +908,7 @@ static int coh901318_alloc_chan_resources(struct dma_chan *chan)
 	coh901318_config(cohc, NULL);
 
 	cohc->allocated = 1;
-	cohc->completed = chan->cookie = 1;
+	dma_cookie_init(chan);
 
 	spin_unlock_irqrestore(&cohc->lock, flags);
 
@@ -966,16 +945,16 @@ coh901318_tx_submit(struct dma_async_tx_descriptor *tx)
 						   desc);
 	struct coh901318_chan *cohc = to_coh901318_chan(tx->chan);
 	unsigned long flags;
+	dma_cookie_t cookie;
 
 	spin_lock_irqsave(&cohc->lock, flags);
-
-	tx->cookie = coh901318_assign_cookie(cohc, cohd);
+	cookie = dma_cookie_assign(tx);
 
 	coh901318_desc_queue(cohc, cohd);
 
 	spin_unlock_irqrestore(&cohc->lock, flags);
 
-	return tx->cookie;
+	return cookie;
 }
 
 static struct dma_async_tx_descriptor *
@@ -1034,8 +1013,8 @@ coh901318_prep_memcpy(struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
 
 static struct dma_async_tx_descriptor *
 coh901318_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
-			unsigned int sg_len, enum dma_data_direction direction,
-			unsigned long flags)
+			unsigned int sg_len, enum dma_transfer_direction direction,
+			unsigned long flags, void *context)
 {
 	struct coh901318_chan *cohc = to_coh901318_chan(chan);
 	struct coh901318_lli *lli;
@@ -1077,7 +1056,7 @@ coh901318_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 	ctrl_last |= cohc->runtime_ctrl;
 	ctrl |= cohc->runtime_ctrl;
 
-	if (direction == DMA_TO_DEVICE) {
+	if (direction == DMA_MEM_TO_DEV) {
 		u32 tx_flags = COH901318_CX_CTRL_PRDD_SOURCE |
 			COH901318_CX_CTRL_SRC_ADDR_INC_ENABLE;
 
@@ -1085,7 +1064,7 @@ coh901318_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 		ctrl_chained |= tx_flags;
 		ctrl_last |= tx_flags;
 		ctrl |= tx_flags;
-	} else if (direction == DMA_FROM_DEVICE) {
+	} else if (direction == DMA_DEV_TO_MEM) {
 		u32 rx_flags = COH901318_CX_CTRL_PRDD_DEST |
 			COH901318_CX_CTRL_DST_ADDR_INC_ENABLE;
 
@@ -1165,17 +1144,12 @@ coh901318_tx_status(struct dma_chan *chan, dma_cookie_t cookie,
 		 struct dma_tx_state *txstate)
 {
 	struct coh901318_chan *cohc = to_coh901318_chan(chan);
-	dma_cookie_t last_used;
-	dma_cookie_t last_complete;
-	int ret;
-
-	last_complete = cohc->completed;
-	last_used = chan->cookie;
+	enum dma_status ret;
 
-	ret = dma_async_is_complete(cookie, last_complete, last_used);
+	ret = dma_cookie_status(chan, cookie, txstate);
+	/* FIXME: should be conditional on ret != DMA_SUCCESS? */
+	dma_set_residue(txstate, coh901318_get_bytes_left(chan));
 
-	dma_set_tx_state(txstate, last_complete, last_used,
-			 coh901318_get_bytes_left(chan));
 	if (ret == DMA_IN_PROGRESS && cohc->stopped)
 		ret = DMA_PAUSED;
 
@@ -1274,11 +1248,11 @@ static void coh901318_dma_set_runtimeconfig(struct dma_chan *chan,
 	int i = 0;
 
 	/* We only support mem to per or per to mem transfers */
-	if (config->direction == DMA_FROM_DEVICE) {
+	if (config->direction == DMA_DEV_TO_MEM) {
 		addr = config->src_addr;
 		addr_width = config->src_addr_width;
 		maxburst = config->src_maxburst;
-	} else if (config->direction == DMA_TO_DEVICE) {
+	} else if (config->direction == DMA_MEM_TO_DEV) {
 		addr = config->dst_addr;
 		addr_width = config->dst_addr_width;
 		maxburst = config->dst_maxburst;
diff --git a/drivers/dma/coh901318_lli.c b/drivers/dma/coh901318_lli.c
index 9f7e0e6a7eea..6c0e2d4c6682 100644
--- a/drivers/dma/coh901318_lli.c
+++ b/drivers/dma/coh901318_lli.c
@@ -7,11 +7,10 @@
  * Author: Per Friden <per.friden@stericsson.com>
  */
 
-#include <linux/dma-mapping.h>
 #include <linux/spinlock.h>
-#include <linux/dmapool.h>
 #include <linux/memory.h>
 #include <linux/gfp.h>
+#include <linux/dmapool.h>
 #include <mach/coh901318.h>
 
 #include "coh901318_lli.h"
@@ -177,18 +176,18 @@ coh901318_lli_fill_single(struct coh901318_pool *pool,
 			  struct coh901318_lli *lli,
 			  dma_addr_t buf, unsigned int size,
 			  dma_addr_t dev_addr, u32 ctrl_chained, u32 ctrl_eom,
-			  enum dma_data_direction dir)
+			  enum dma_transfer_direction dir)
 {
 	int s = size;
 	dma_addr_t src;
 	dma_addr_t dst;
 
 
-	if (dir == DMA_TO_DEVICE) {
+	if (dir == DMA_MEM_TO_DEV) {
 		src = buf;
 		dst = dev_addr;
 
-	} else if (dir == DMA_FROM_DEVICE) {
+	} else if (dir == DMA_DEV_TO_MEM) {
 
 		src = dev_addr;
 		dst = buf;
@@ -215,9 +214,9 @@ coh901318_lli_fill_single(struct coh901318_pool *pool,
 
 		lli = coh901318_lli_next(lli);
 
-		if (dir == DMA_TO_DEVICE)
+		if (dir == DMA_MEM_TO_DEV)
 			src += block_size;
-		else if (dir == DMA_FROM_DEVICE)
+		else if (dir == DMA_DEV_TO_MEM)
 			dst += block_size;
 	}
 
@@ -234,7 +233,7 @@ coh901318_lli_fill_sg(struct coh901318_pool *pool,
 		      struct scatterlist *sgl, unsigned int nents,
 		      dma_addr_t dev_addr, u32 ctrl_chained, u32 ctrl,
 		      u32 ctrl_last,
-		      enum dma_data_direction dir, u32 ctrl_irq_mask)
+		      enum dma_transfer_direction dir, u32 ctrl_irq_mask)
 {
 	int i;
 	struct scatterlist *sg;
@@ -249,9 +248,9 @@ coh901318_lli_fill_sg(struct coh901318_pool *pool,
 
 	spin_lock(&pool->lock);
 
-	if (dir == DMA_TO_DEVICE)
+	if (dir == DMA_MEM_TO_DEV)
 		dst = dev_addr;
-	else if (dir == DMA_FROM_DEVICE)
+	else if (dir == DMA_DEV_TO_MEM)
 		src = dev_addr;
 	else
 		goto err;
@@ -269,7 +268,7 @@ coh901318_lli_fill_sg(struct coh901318_pool *pool,
 			ctrl_sg = ctrl ? ctrl : ctrl_last;
 
 
-		if (dir == DMA_TO_DEVICE)
+		if (dir == DMA_MEM_TO_DEV)
 			/* increment source address */
 			src = sg_phys(sg);
 		else
@@ -293,7 +292,7 @@ coh901318_lli_fill_sg(struct coh901318_pool *pool,
 			lli->src_addr = src;
 			lli->dst_addr = dst;
 
-			if (dir == DMA_FROM_DEVICE)
+			if (dir == DMA_DEV_TO_MEM)
 				dst += elem_size;
 			else
 				src += elem_size;
diff --git a/drivers/dma/coh901318_lli.h b/drivers/dma/coh901318_lli.h
index 7a5c80990e9e..abff3714fdda 100644
--- a/drivers/dma/coh901318_lli.h
+++ b/drivers/dma/coh901318_lli.h
@@ -97,7 +97,7 @@ coh901318_lli_fill_single(struct coh901318_pool *pool,
 			  struct coh901318_lli *lli,
 			  dma_addr_t buf, unsigned int size,
 			  dma_addr_t dev_addr, u32 ctrl_chained, u32 ctrl_last,
-			  enum dma_data_direction dir);
+			  enum dma_transfer_direction dir);
 
 /**
  * coh901318_lli_fill_single() - Prepares the lli:s for dma scatter list transfer
@@ -119,6 +119,6 @@ coh901318_lli_fill_sg(struct coh901318_pool *pool,
 		      struct scatterlist *sg, unsigned int nents,
 		      dma_addr_t dev_addr, u32 ctrl_chained,
 		      u32 ctrl, u32 ctrl_last,
-		      enum dma_data_direction dir, u32 ctrl_irq_mask);
+		      enum dma_transfer_direction dir, u32 ctrl_irq_mask);
 
 #endif /* COH901318_LLI_H */
diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c
index b48967b499da..2397f6f451b1 100644
--- a/drivers/dma/dmaengine.c
+++ b/drivers/dma/dmaengine.c
@@ -332,6 +332,20 @@ struct dma_chan *dma_find_channel(enum dma_transaction_type tx_type)
 }
 EXPORT_SYMBOL(dma_find_channel);
 
+/*
+ * net_dma_find_channel - find a channel for net_dma
+ * net_dma has alignment requirements
+ */
+struct dma_chan *net_dma_find_channel(void)
+{
+	struct dma_chan *chan = dma_find_channel(DMA_MEMCPY);
+	if (chan && !is_dma_copy_aligned(chan->device, 1, 1, 1))
+		return NULL;
+
+	return chan;
+}
+EXPORT_SYMBOL(net_dma_find_channel);
+
 /**
  * dma_issue_pending_all - flush all pending operations across all channels
  */
@@ -510,8 +524,8 @@ struct dma_chan *__dma_request_channel(dma_cap_mask_t *mask, dma_filter_fn fn, v
 					 dma_chan_name(chan));
 				list_del_rcu(&device->global_node);
 			} else if (err)
-				pr_debug("dmaengine: failed to get %s: (%d)\n",
-					 dma_chan_name(chan), err);
+				pr_debug("%s: failed to get %s: (%d)\n",
+					__func__, dma_chan_name(chan), err);
 			else
 				break;
 			if (--device->privatecnt == 0)
@@ -564,8 +578,8 @@ void dmaengine_get(void)
 				list_del_rcu(&device->global_node);
 				break;
 			} else if (err)
-				pr_err("dmaengine: failed to get %s: (%d)\n",
-				       dma_chan_name(chan), err);
+				pr_err("%s: failed to get %s: (%d)\n",
+					__func__, dma_chan_name(chan), err);
 		}
 	}
 
@@ -693,12 +707,12 @@ int dma_async_device_register(struct dma_device *device)
 		!device->device_prep_dma_interrupt);
 	BUG_ON(dma_has_cap(DMA_SG, device->cap_mask) &&
 		!device->device_prep_dma_sg);
-	BUG_ON(dma_has_cap(DMA_SLAVE, device->cap_mask) &&
-		!device->device_prep_slave_sg);
 	BUG_ON(dma_has_cap(DMA_CYCLIC, device->cap_mask) &&
 		!device->device_prep_dma_cyclic);
 	BUG_ON(dma_has_cap(DMA_SLAVE, device->cap_mask) &&
 		!device->device_control);
+	BUG_ON(dma_has_cap(DMA_INTERLEAVE, device->cap_mask) &&
+		!device->device_prep_interleaved_dma);
 
 	BUG_ON(!device->device_alloc_chan_resources);
 	BUG_ON(!device->device_free_chan_resources);
diff --git a/drivers/dma/dmaengine.h b/drivers/dma/dmaengine.h
new file mode 100644
index 000000000000..17f983a4e9ba
--- /dev/null
+++ b/drivers/dma/dmaengine.h
@@ -0,0 +1,89 @@
+/*
+ * The contents of this file are private to DMA engine drivers, and is not
+ * part of the API to be used by DMA engine users.
+ */
+#ifndef DMAENGINE_H
+#define DMAENGINE_H
+
+#include <linux/bug.h>
+#include <linux/dmaengine.h>
+
+/**
+ * dma_cookie_init - initialize the cookies for a DMA channel
+ * @chan: dma channel to initialize
+ */
+static inline void dma_cookie_init(struct dma_chan *chan)
+{
+	chan->cookie = DMA_MIN_COOKIE;
+	chan->completed_cookie = DMA_MIN_COOKIE;
+}
+
+/**
+ * dma_cookie_assign - assign a DMA engine cookie to the descriptor
+ * @tx: descriptor needing cookie
+ *
+ * Assign a unique non-zero per-channel cookie to the descriptor.
+ * Note: caller is expected to hold a lock to prevent concurrency.
+ */
+static inline dma_cookie_t dma_cookie_assign(struct dma_async_tx_descriptor *tx)
+{
+	struct dma_chan *chan = tx->chan;
+	dma_cookie_t cookie;
+
+	cookie = chan->cookie + 1;
+	if (cookie < DMA_MIN_COOKIE)
+		cookie = DMA_MIN_COOKIE;
+	tx->cookie = chan->cookie = cookie;
+
+	return cookie;
+}
+
+/**
+ * dma_cookie_complete - complete a descriptor
+ * @tx: descriptor to complete
+ *
+ * Mark this descriptor complete by updating the channels completed
+ * cookie marker.  Zero the descriptors cookie to prevent accidental
+ * repeated completions.
+ *
+ * Note: caller is expected to hold a lock to prevent concurrency.
+ */
+static inline void dma_cookie_complete(struct dma_async_tx_descriptor *tx)
+{
+	BUG_ON(tx->cookie < DMA_MIN_COOKIE);
+	tx->chan->completed_cookie = tx->cookie;
+	tx->cookie = 0;
+}
+
+/**
+ * dma_cookie_status - report cookie status
+ * @chan: dma channel
+ * @cookie: cookie we are interested in
+ * @state: dma_tx_state structure to return last/used cookies
+ *
+ * Report the status of the cookie, filling in the state structure if
+ * non-NULL.  No locking is required.
+ */
+static inline enum dma_status dma_cookie_status(struct dma_chan *chan,
+	dma_cookie_t cookie, struct dma_tx_state *state)
+{
+	dma_cookie_t used, complete;
+
+	used = chan->cookie;
+	complete = chan->completed_cookie;
+	barrier();
+	if (state) {
+		state->last = complete;
+		state->used = used;
+		state->residue = 0;
+	}
+	return dma_async_is_complete(cookie, complete, used);
+}
+
+static inline void dma_set_residue(struct dma_tx_state *state, u32 residue)
+{
+	if (state)
+		state->residue = residue;
+}
+
+#endif
diff --git a/drivers/dma/dmatest.c b/drivers/dma/dmatest.c
index 2b8661b54eaf..24225f0fdcd8 100644
--- a/drivers/dma/dmatest.c
+++ b/drivers/dma/dmatest.c
@@ -599,7 +599,7 @@ static int dmatest_add_channel(struct dma_chan *chan)
 	}
 	if (dma_has_cap(DMA_PQ, dma_dev->cap_mask)) {
 		cnt = dmatest_add_threads(dtc, DMA_PQ);
-		thread_count += cnt > 0 ?: 0;
+		thread_count += cnt > 0 ? cnt : 0;
 	}
 
 	pr_info("dmatest: Started %u threads using %s\n",
diff --git a/drivers/dma/dw_dmac.c b/drivers/dma/dw_dmac.c
index 9bfd6d360718..7439079f5eed 100644
--- a/drivers/dma/dw_dmac.c
+++ b/drivers/dma/dw_dmac.c
@@ -9,6 +9,7 @@
  * it under the terms of the GNU General Public License version 2 as
  * published by the Free Software Foundation.
  */
+#include <linux/bitops.h>
 #include <linux/clk.h>
 #include <linux/delay.h>
 #include <linux/dmaengine.h>
@@ -22,6 +23,7 @@
 #include <linux/slab.h>
 
 #include "dw_dmac_regs.h"
+#include "dmaengine.h"
 
 /*
  * This supports the Synopsys "DesignWare AHB Central DMA Controller",
@@ -33,19 +35,23 @@
  * which does not support descriptor writeback.
  */
 
-#define DWC_DEFAULT_CTLLO(private) ({				\
-		struct dw_dma_slave *__slave = (private);	\
-		int dms = __slave ? __slave->dst_master : 0;	\
-		int sms = __slave ? __slave->src_master : 1;	\
-		u8 smsize = __slave ? __slave->src_msize : DW_DMA_MSIZE_16; \
-		u8 dmsize = __slave ? __slave->dst_msize : DW_DMA_MSIZE_16; \
+#define DWC_DEFAULT_CTLLO(_chan) ({				\
+		struct dw_dma_slave *__slave = (_chan->private);	\
+		struct dw_dma_chan *_dwc = to_dw_dma_chan(_chan);	\
+		struct dma_slave_config	*_sconfig = &_dwc->dma_sconfig;	\
+		int _dms = __slave ? __slave->dst_master : 0;	\
+		int _sms = __slave ? __slave->src_master : 1;	\
+		u8 _smsize = __slave ? _sconfig->src_maxburst :	\
+			DW_DMA_MSIZE_16;			\
+		u8 _dmsize = __slave ? _sconfig->dst_maxburst :	\
+			DW_DMA_MSIZE_16;			\
 								\
-		(DWC_CTLL_DST_MSIZE(dmsize)			\
-		 | DWC_CTLL_SRC_MSIZE(smsize)			\
+		(DWC_CTLL_DST_MSIZE(_dmsize)			\
+		 | DWC_CTLL_SRC_MSIZE(_smsize)			\
 		 | DWC_CTLL_LLP_D_EN				\
 		 | DWC_CTLL_LLP_S_EN				\
-		 | DWC_CTLL_DMS(dms)				\
-		 | DWC_CTLL_SMS(sms));				\
+		 | DWC_CTLL_DMS(_dms)				\
+		 | DWC_CTLL_SMS(_sms));				\
 	})
 
 /*
@@ -151,19 +157,35 @@ static void dwc_desc_put(struct dw_dma_chan *dwc, struct dw_desc *desc)
 	}
 }
 
-/* Called with dwc->lock held and bh disabled */
-static dma_cookie_t
-dwc_assign_cookie(struct dw_dma_chan *dwc, struct dw_desc *desc)
+static void dwc_initialize(struct dw_dma_chan *dwc)
 {
-	dma_cookie_t cookie = dwc->chan.cookie;
+	struct dw_dma *dw = to_dw_dma(dwc->chan.device);
+	struct dw_dma_slave *dws = dwc->chan.private;
+	u32 cfghi = DWC_CFGH_FIFO_MODE;
+	u32 cfglo = DWC_CFGL_CH_PRIOR(dwc->priority);
 
-	if (++cookie < 0)
-		cookie = 1;
+	if (dwc->initialized == true)
+		return;
 
-	dwc->chan.cookie = cookie;
-	desc->txd.cookie = cookie;
+	if (dws) {
+		/*
+		 * We need controller-specific data to set up slave
+		 * transfers.
+		 */
+		BUG_ON(!dws->dma_dev || dws->dma_dev != dw->dma.dev);
 
-	return cookie;
+		cfghi = dws->cfg_hi;
+		cfglo |= dws->cfg_lo & ~DWC_CFGL_CH_PRIOR_MASK;
+	}
+
+	channel_writel(dwc, CFG_LO, cfglo);
+	channel_writel(dwc, CFG_HI, cfghi);
+
+	/* Enable interrupts */
+	channel_set_bit(dw, MASK.XFER, dwc->mask);
+	channel_set_bit(dw, MASK.ERROR, dwc->mask);
+
+	dwc->initialized = true;
 }
 
 /*----------------------------------------------------------------------*/
@@ -189,6 +211,8 @@ static void dwc_dostart(struct dw_dma_chan *dwc, struct dw_desc *first)
 		return;
 	}
 
+	dwc_initialize(dwc);
+
 	channel_writel(dwc, LLP, first->txd.phys);
 	channel_writel(dwc, CTL_LO,
 			DWC_CTLL_LLP_D_EN | DWC_CTLL_LLP_S_EN);
@@ -211,7 +235,7 @@ dwc_descriptor_complete(struct dw_dma_chan *dwc, struct dw_desc *desc,
 	dev_vdbg(chan2dev(&dwc->chan), "descriptor %u complete\n", txd->cookie);
 
 	spin_lock_irqsave(&dwc->lock, flags);
-	dwc->completed = txd->cookie;
+	dma_cookie_complete(txd);
 	if (callback_required) {
 		callback = txd->callback;
 		param = txd->callback_param;
@@ -295,12 +319,6 @@ static void dwc_scan_descriptors(struct dw_dma *dw, struct dw_dma_chan *dwc)
 	unsigned long flags;
 
 	spin_lock_irqsave(&dwc->lock, flags);
-	/*
-	 * Clear block interrupt flag before scanning so that we don't
-	 * miss any, and read LLP before RAW_XFER to ensure it is
-	 * valid if we decide to scan the list.
-	 */
-	dma_writel(dw, CLEAR.BLOCK, dwc->mask);
 	llp = channel_readl(dwc, LLP);
 	status_xfer = dma_readl(dw, RAW.XFER);
 
@@ -436,17 +454,16 @@ EXPORT_SYMBOL(dw_dma_get_dst_addr);
 
 /* called with dwc->lock held and all DMAC interrupts disabled */
 static void dwc_handle_cyclic(struct dw_dma *dw, struct dw_dma_chan *dwc,
-		u32 status_block, u32 status_err, u32 status_xfer)
+		u32 status_err, u32 status_xfer)
 {
 	unsigned long flags;
 
-	if (status_block & dwc->mask) {
+	if (dwc->mask) {
 		void (*callback)(void *param);
 		void *callback_param;
 
 		dev_vdbg(chan2dev(&dwc->chan), "new cyclic period llp 0x%08x\n",
 				channel_readl(dwc, LLP));
-		dma_writel(dw, CLEAR.BLOCK, dwc->mask);
 
 		callback = dwc->cdesc->period_callback;
 		callback_param = dwc->cdesc->period_callback_param;
@@ -486,7 +503,6 @@ static void dwc_handle_cyclic(struct dw_dma *dw, struct dw_dma_chan *dwc,
 		channel_writel(dwc, CTL_LO, 0);
 		channel_writel(dwc, CTL_HI, 0);
 
-		dma_writel(dw, CLEAR.BLOCK, dwc->mask);
 		dma_writel(dw, CLEAR.ERROR, dwc->mask);
 		dma_writel(dw, CLEAR.XFER, dwc->mask);
 
@@ -503,36 +519,29 @@ static void dw_dma_tasklet(unsigned long data)
 {
 	struct dw_dma *dw = (struct dw_dma *)data;
 	struct dw_dma_chan *dwc;
-	u32 status_block;
 	u32 status_xfer;
 	u32 status_err;
 	int i;
 
-	status_block = dma_readl(dw, RAW.BLOCK);
 	status_xfer = dma_readl(dw, RAW.XFER);
 	status_err = dma_readl(dw, RAW.ERROR);
 
-	dev_vdbg(dw->dma.dev, "tasklet: status_block=%x status_err=%x\n",
-			status_block, status_err);
+	dev_vdbg(dw->dma.dev, "tasklet: status_err=%x\n", status_err);
 
 	for (i = 0; i < dw->dma.chancnt; i++) {
 		dwc = &dw->chan[i];
 		if (test_bit(DW_DMA_IS_CYCLIC, &dwc->flags))
-			dwc_handle_cyclic(dw, dwc, status_block, status_err,
-					status_xfer);
+			dwc_handle_cyclic(dw, dwc, status_err, status_xfer);
 		else if (status_err & (1 << i))
 			dwc_handle_error(dw, dwc);
-		else if ((status_block | status_xfer) & (1 << i))
+		else if (status_xfer & (1 << i))
 			dwc_scan_descriptors(dw, dwc);
 	}
 
 	/*
-	 * Re-enable interrupts. Block Complete interrupts are only
-	 * enabled if the INT_EN bit in the descriptor is set. This
-	 * will trigger a scan before the whole list is done.
+	 * Re-enable interrupts.
 	 */
 	channel_set_bit(dw, MASK.XFER, dw->all_chan_mask);
-	channel_set_bit(dw, MASK.BLOCK, dw->all_chan_mask);
 	channel_set_bit(dw, MASK.ERROR, dw->all_chan_mask);
 }
 
@@ -549,7 +558,6 @@ static irqreturn_t dw_dma_interrupt(int irq, void *dev_id)
 	 * softirq handler.
 	 */
 	channel_clear_bit(dw, MASK.XFER, dw->all_chan_mask);
-	channel_clear_bit(dw, MASK.BLOCK, dw->all_chan_mask);
 	channel_clear_bit(dw, MASK.ERROR, dw->all_chan_mask);
 
 	status = dma_readl(dw, STATUS_INT);
@@ -560,7 +568,6 @@ static irqreturn_t dw_dma_interrupt(int irq, void *dev_id)
 
 		/* Try to recover */
 		channel_clear_bit(dw, MASK.XFER, (1 << 8) - 1);
-		channel_clear_bit(dw, MASK.BLOCK, (1 << 8) - 1);
 		channel_clear_bit(dw, MASK.SRC_TRAN, (1 << 8) - 1);
 		channel_clear_bit(dw, MASK.DST_TRAN, (1 << 8) - 1);
 		channel_clear_bit(dw, MASK.ERROR, (1 << 8) - 1);
@@ -581,7 +588,7 @@ static dma_cookie_t dwc_tx_submit(struct dma_async_tx_descriptor *tx)
 	unsigned long		flags;
 
 	spin_lock_irqsave(&dwc->lock, flags);
-	cookie = dwc_assign_cookie(dwc, desc);
+	cookie = dma_cookie_assign(tx);
 
 	/*
 	 * REVISIT: We should attempt to chain as many descriptors as
@@ -640,7 +647,7 @@ dwc_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
 	else
 		src_width = dst_width = 0;
 
-	ctllo = DWC_DEFAULT_CTLLO(chan->private)
+	ctllo = DWC_DEFAULT_CTLLO(chan)
 			| DWC_CTLL_DST_WIDTH(dst_width)
 			| DWC_CTLL_SRC_WIDTH(src_width)
 			| DWC_CTLL_DST_INC
@@ -696,11 +703,12 @@ err_desc_get:
 
 static struct dma_async_tx_descriptor *
 dwc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
-		unsigned int sg_len, enum dma_data_direction direction,
-		unsigned long flags)
+		unsigned int sg_len, enum dma_transfer_direction direction,
+		unsigned long flags, void *context)
 {
 	struct dw_dma_chan	*dwc = to_dw_dma_chan(chan);
 	struct dw_dma_slave	*dws = chan->private;
+	struct dma_slave_config	*sconfig = &dwc->dma_sconfig;
 	struct dw_desc		*prev;
 	struct dw_desc		*first;
 	u32			ctllo;
@@ -716,25 +724,34 @@ dwc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 	if (unlikely(!dws || !sg_len))
 		return NULL;
 
-	reg_width = dws->reg_width;
 	prev = first = NULL;
 
 	switch (direction) {
-	case DMA_TO_DEVICE:
-		ctllo = (DWC_DEFAULT_CTLLO(chan->private)
+	case DMA_MEM_TO_DEV:
+		reg_width = __fls(sconfig->dst_addr_width);
+		reg = sconfig->dst_addr;
+		ctllo = (DWC_DEFAULT_CTLLO(chan)
 				| DWC_CTLL_DST_WIDTH(reg_width)
 				| DWC_CTLL_DST_FIX
-				| DWC_CTLL_SRC_INC
-				| DWC_CTLL_FC(dws->fc));
-		reg = dws->tx_reg;
+				| DWC_CTLL_SRC_INC);
+
+		ctllo |= sconfig->device_fc ? DWC_CTLL_FC(DW_DMA_FC_P_M2P) :
+			DWC_CTLL_FC(DW_DMA_FC_D_M2P);
+
 		for_each_sg(sgl, sg, sg_len, i) {
 			struct dw_desc	*desc;
 			u32		len, dlen, mem;
 
 			mem = sg_phys(sg);
 			len = sg_dma_len(sg);
-			mem_width = 2;
-			if (unlikely(mem & 3 || len & 3))
+
+			if (!((mem | len) & 7))
+				mem_width = 3;
+			else if (!((mem | len) & 3))
+				mem_width = 2;
+			else if (!((mem | len) & 1))
+				mem_width = 1;
+			else
 				mem_width = 0;
 
 slave_sg_todev_fill_desc:
@@ -777,22 +794,31 @@ slave_sg_todev_fill_desc:
 				goto slave_sg_todev_fill_desc;
 		}
 		break;
-	case DMA_FROM_DEVICE:
-		ctllo = (DWC_DEFAULT_CTLLO(chan->private)
+	case DMA_DEV_TO_MEM:
+		reg_width = __fls(sconfig->src_addr_width);
+		reg = sconfig->src_addr;
+		ctllo = (DWC_DEFAULT_CTLLO(chan)
 				| DWC_CTLL_SRC_WIDTH(reg_width)
 				| DWC_CTLL_DST_INC
-				| DWC_CTLL_SRC_FIX
-				| DWC_CTLL_FC(dws->fc));
+				| DWC_CTLL_SRC_FIX);
+
+		ctllo |= sconfig->device_fc ? DWC_CTLL_FC(DW_DMA_FC_P_P2M) :
+			DWC_CTLL_FC(DW_DMA_FC_D_P2M);
 
-		reg = dws->rx_reg;
 		for_each_sg(sgl, sg, sg_len, i) {
 			struct dw_desc	*desc;
 			u32		len, dlen, mem;
 
 			mem = sg_phys(sg);
 			len = sg_dma_len(sg);
-			mem_width = 2;
-			if (unlikely(mem & 3 || len & 3))
+
+			if (!((mem | len) & 7))
+				mem_width = 3;
+			else if (!((mem | len) & 3))
+				mem_width = 2;
+			else if (!((mem | len) & 1))
+				mem_width = 1;
+			else
 				mem_width = 0;
 
 slave_sg_fromdev_fill_desc:
@@ -856,6 +882,39 @@ err_desc_get:
 	return NULL;
 }
 
+/*
+ * Fix sconfig's burst size according to dw_dmac. We need to convert them as:
+ * 1 -> 0, 4 -> 1, 8 -> 2, 16 -> 3.
+ *
+ * NOTE: burst size 2 is not supported by controller.
+ *
+ * This can be done by finding least significant bit set: n & (n - 1)
+ */
+static inline void convert_burst(u32 *maxburst)
+{
+	if (*maxburst > 1)
+		*maxburst = fls(*maxburst) - 2;
+	else
+		*maxburst = 0;
+}
+
+static int
+set_runtime_config(struct dma_chan *chan, struct dma_slave_config *sconfig)
+{
+	struct dw_dma_chan *dwc = to_dw_dma_chan(chan);
+
+	/* Check if it is chan is configured for slave transfers */
+	if (!chan->private)
+		return -EINVAL;
+
+	memcpy(&dwc->dma_sconfig, sconfig, sizeof(*sconfig));
+
+	convert_burst(&dwc->dma_sconfig.src_maxburst);
+	convert_burst(&dwc->dma_sconfig.dst_maxburst);
+
+	return 0;
+}
+
 static int dwc_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
 		       unsigned long arg)
 {
@@ -905,8 +964,11 @@ static int dwc_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
 		/* Flush all pending and queued descriptors */
 		list_for_each_entry_safe(desc, _desc, &list, desc_node)
 			dwc_descriptor_complete(dwc, desc, false);
-	} else
+	} else if (cmd == DMA_SLAVE_CONFIG) {
+		return set_runtime_config(chan, (struct dma_slave_config *)arg);
+	} else {
 		return -ENXIO;
+	}
 
 	return 0;
 }
@@ -917,28 +979,17 @@ dwc_tx_status(struct dma_chan *chan,
 	      struct dma_tx_state *txstate)
 {
 	struct dw_dma_chan	*dwc = to_dw_dma_chan(chan);
-	dma_cookie_t		last_used;
-	dma_cookie_t		last_complete;
-	int			ret;
-
-	last_complete = dwc->completed;
-	last_used = chan->cookie;
+	enum dma_status		ret;
 
-	ret = dma_async_is_complete(cookie, last_complete, last_used);
+	ret = dma_cookie_status(chan, cookie, txstate);
 	if (ret != DMA_SUCCESS) {
 		dwc_scan_descriptors(to_dw_dma(chan->device), dwc);
 
-		last_complete = dwc->completed;
-		last_used = chan->cookie;
-
-		ret = dma_async_is_complete(cookie, last_complete, last_used);
+		ret = dma_cookie_status(chan, cookie, txstate);
 	}
 
 	if (ret != DMA_SUCCESS)
-		dma_set_tx_state(txstate, last_complete, last_used,
-				dwc_first_active(dwc)->len);
-	else
-		dma_set_tx_state(txstate, last_complete, last_used, 0);
+		dma_set_residue(txstate, dwc_first_active(dwc)->len);
 
 	if (dwc->paused)
 		return DMA_PAUSED;
@@ -959,10 +1010,7 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
 	struct dw_dma_chan	*dwc = to_dw_dma_chan(chan);
 	struct dw_dma		*dw = to_dw_dma(chan->device);
 	struct dw_desc		*desc;
-	struct dw_dma_slave	*dws;
 	int			i;
-	u32			cfghi;
-	u32			cfglo;
 	unsigned long		flags;
 
 	dev_vdbg(chan2dev(chan), "alloc_chan_resources\n");
@@ -973,27 +1021,7 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
 		return -EIO;
 	}
 
-	dwc->completed = chan->cookie = 1;
-
-	cfghi = DWC_CFGH_FIFO_MODE;
-	cfglo = 0;
-
-	dws = chan->private;
-	if (dws) {
-		/*
-		 * We need controller-specific data to set up slave
-		 * transfers.
-		 */
-		BUG_ON(!dws->dma_dev || dws->dma_dev != dw->dma.dev);
-
-		cfghi = dws->cfg_hi;
-		cfglo = dws->cfg_lo & ~DWC_CFGL_CH_PRIOR_MASK;
-	}
-
-	cfglo |= DWC_CFGL_CH_PRIOR(dwc->priority);
-
-	channel_writel(dwc, CFG_LO, cfglo);
-	channel_writel(dwc, CFG_HI, cfghi);
+	dma_cookie_init(chan);
 
 	/*
 	 * NOTE: some controllers may have additional features that we
@@ -1026,11 +1054,6 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
 		i = ++dwc->descs_allocated;
 	}
 
-	/* Enable interrupts */
-	channel_set_bit(dw, MASK.XFER, dwc->mask);
-	channel_set_bit(dw, MASK.BLOCK, dwc->mask);
-	channel_set_bit(dw, MASK.ERROR, dwc->mask);
-
 	spin_unlock_irqrestore(&dwc->lock, flags);
 
 	dev_dbg(chan2dev(chan),
@@ -1058,10 +1081,10 @@ static void dwc_free_chan_resources(struct dma_chan *chan)
 	spin_lock_irqsave(&dwc->lock, flags);
 	list_splice_init(&dwc->free_list, &list);
 	dwc->descs_allocated = 0;
+	dwc->initialized = false;
 
 	/* Disable interrupts */
 	channel_clear_bit(dw, MASK.XFER, dwc->mask);
-	channel_clear_bit(dw, MASK.BLOCK, dwc->mask);
 	channel_clear_bit(dw, MASK.ERROR, dwc->mask);
 
 	spin_unlock_irqrestore(&dwc->lock, flags);
@@ -1113,7 +1136,6 @@ int dw_dma_cyclic_start(struct dma_chan *chan)
 		return -EBUSY;
 	}
 
-	dma_writel(dw, CLEAR.BLOCK, dwc->mask);
 	dma_writel(dw, CLEAR.ERROR, dwc->mask);
 	dma_writel(dw, CLEAR.XFER, dwc->mask);
 
@@ -1165,14 +1187,14 @@ EXPORT_SYMBOL(dw_dma_cyclic_stop);
  */
 struct dw_cyclic_desc *dw_dma_cyclic_prep(struct dma_chan *chan,
 		dma_addr_t buf_addr, size_t buf_len, size_t period_len,
-		enum dma_data_direction direction)
+		enum dma_transfer_direction direction)
 {
 	struct dw_dma_chan		*dwc = to_dw_dma_chan(chan);
+	struct dma_slave_config		*sconfig = &dwc->dma_sconfig;
 	struct dw_cyclic_desc		*cdesc;
 	struct dw_cyclic_desc		*retval = NULL;
 	struct dw_desc			*desc;
 	struct dw_desc			*last = NULL;
-	struct dw_dma_slave		*dws = chan->private;
 	unsigned long			was_cyclic;
 	unsigned int			reg_width;
 	unsigned int			periods;
@@ -1196,7 +1218,12 @@ struct dw_cyclic_desc *dw_dma_cyclic_prep(struct dma_chan *chan,
 	}
 
 	retval = ERR_PTR(-EINVAL);
-	reg_width = dws->reg_width;
+
+	if (direction == DMA_MEM_TO_DEV)
+		reg_width = __ffs(sconfig->dst_addr_width);
+	else
+		reg_width = __ffs(sconfig->src_addr_width);
+
 	periods = buf_len / period_len;
 
 	/* Check for too big/unaligned periods and unaligned DMA buffer. */
@@ -1206,7 +1233,7 @@ struct dw_cyclic_desc *dw_dma_cyclic_prep(struct dma_chan *chan,
 		goto out_err;
 	if (unlikely(buf_addr & ((1 << reg_width) - 1)))
 		goto out_err;
-	if (unlikely(!(direction & (DMA_TO_DEVICE | DMA_FROM_DEVICE))))
+	if (unlikely(!(direction & (DMA_MEM_TO_DEV | DMA_DEV_TO_MEM))))
 		goto out_err;
 
 	retval = ERR_PTR(-ENOMEM);
@@ -1228,27 +1255,35 @@ struct dw_cyclic_desc *dw_dma_cyclic_prep(struct dma_chan *chan,
 			goto out_err_desc_get;
 
 		switch (direction) {
-		case DMA_TO_DEVICE:
-			desc->lli.dar = dws->tx_reg;
+		case DMA_MEM_TO_DEV:
+			desc->lli.dar = sconfig->dst_addr;
 			desc->lli.sar = buf_addr + (period_len * i);
-			desc->lli.ctllo = (DWC_DEFAULT_CTLLO(chan->private)
+			desc->lli.ctllo = (DWC_DEFAULT_CTLLO(chan)
 					| DWC_CTLL_DST_WIDTH(reg_width)
 					| DWC_CTLL_SRC_WIDTH(reg_width)
 					| DWC_CTLL_DST_FIX
 					| DWC_CTLL_SRC_INC
-					| DWC_CTLL_FC(dws->fc)
 					| DWC_CTLL_INT_EN);
+
+			desc->lli.ctllo |= sconfig->device_fc ?
+				DWC_CTLL_FC(DW_DMA_FC_P_M2P) :
+				DWC_CTLL_FC(DW_DMA_FC_D_M2P);
+
 			break;
-		case DMA_FROM_DEVICE:
+		case DMA_DEV_TO_MEM:
 			desc->lli.dar = buf_addr + (period_len * i);
-			desc->lli.sar = dws->rx_reg;
-			desc->lli.ctllo = (DWC_DEFAULT_CTLLO(chan->private)
+			desc->lli.sar = sconfig->src_addr;
+			desc->lli.ctllo = (DWC_DEFAULT_CTLLO(chan)
 					| DWC_CTLL_SRC_WIDTH(reg_width)
 					| DWC_CTLL_DST_WIDTH(reg_width)
 					| DWC_CTLL_DST_INC
 					| DWC_CTLL_SRC_FIX
-					| DWC_CTLL_FC(dws->fc)
 					| DWC_CTLL_INT_EN);
+
+			desc->lli.ctllo |= sconfig->device_fc ?
+				DWC_CTLL_FC(DW_DMA_FC_P_P2M) :
+				DWC_CTLL_FC(DW_DMA_FC_D_P2M);
+
 			break;
 		default:
 			break;
@@ -1315,7 +1350,6 @@ void dw_dma_cyclic_free(struct dma_chan *chan)
 	while (dma_readl(dw, CH_EN) & dwc->mask)
 		cpu_relax();
 
-	dma_writel(dw, CLEAR.BLOCK, dwc->mask);
 	dma_writel(dw, CLEAR.ERROR, dwc->mask);
 	dma_writel(dw, CLEAR.XFER, dwc->mask);
 
@@ -1335,16 +1369,20 @@ EXPORT_SYMBOL(dw_dma_cyclic_free);
 
 static void dw_dma_off(struct dw_dma *dw)
 {
+	int i;
+
 	dma_writel(dw, CFG, 0);
 
 	channel_clear_bit(dw, MASK.XFER, dw->all_chan_mask);
-	channel_clear_bit(dw, MASK.BLOCK, dw->all_chan_mask);
 	channel_clear_bit(dw, MASK.SRC_TRAN, dw->all_chan_mask);
 	channel_clear_bit(dw, MASK.DST_TRAN, dw->all_chan_mask);
 	channel_clear_bit(dw, MASK.ERROR, dw->all_chan_mask);
 
 	while (dma_readl(dw, CFG) & DW_CFG_DMA_EN)
 		cpu_relax();
+
+	for (i = 0; i < dw->dma.chancnt; i++)
+		dw->chan[i].initialized = false;
 }
 
 static int __init dw_probe(struct platform_device *pdev)
@@ -1357,7 +1395,7 @@ static int __init dw_probe(struct platform_device *pdev)
 	int			err;
 	int			i;
 
-	pdata = pdev->dev.platform_data;
+	pdata = dev_get_platdata(&pdev->dev);
 	if (!pdata || pdata->nr_channels > DW_DMA_MAX_NR_CHANNELS)
 		return -EINVAL;
 
@@ -1411,7 +1449,7 @@ static int __init dw_probe(struct platform_device *pdev)
 		struct dw_dma_chan	*dwc = &dw->chan[i];
 
 		dwc->chan.device = &dw->dma;
-		dwc->chan.cookie = dwc->completed = 1;
+		dma_cookie_init(&dwc->chan);
 		if (pdata->chan_allocation_order == CHAN_ALLOCATION_ASCENDING)
 			list_add_tail(&dwc->chan.device_node,
 					&dw->dma.channels);
@@ -1420,7 +1458,7 @@ static int __init dw_probe(struct platform_device *pdev)
 
 		/* 7 is highest priority & 0 is lowest. */
 		if (pdata->chan_priority == CHAN_PRIORITY_ASCENDING)
-			dwc->priority = 7 - i;
+			dwc->priority = pdata->nr_channels - i - 1;
 		else
 			dwc->priority = i;
 
@@ -1437,13 +1475,11 @@ static int __init dw_probe(struct platform_device *pdev)
 
 	/* Clear/disable all interrupts on all channels. */
 	dma_writel(dw, CLEAR.XFER, dw->all_chan_mask);
-	dma_writel(dw, CLEAR.BLOCK, dw->all_chan_mask);
 	dma_writel(dw, CLEAR.SRC_TRAN, dw->all_chan_mask);
 	dma_writel(dw, CLEAR.DST_TRAN, dw->all_chan_mask);
 	dma_writel(dw, CLEAR.ERROR, dw->all_chan_mask);
 
 	channel_clear_bit(dw, MASK.XFER, dw->all_chan_mask);
-	channel_clear_bit(dw, MASK.BLOCK, dw->all_chan_mask);
 	channel_clear_bit(dw, MASK.SRC_TRAN, dw->all_chan_mask);
 	channel_clear_bit(dw, MASK.DST_TRAN, dw->all_chan_mask);
 	channel_clear_bit(dw, MASK.ERROR, dw->all_chan_mask);
@@ -1533,6 +1569,7 @@ static int dw_suspend_noirq(struct device *dev)
 
 	dw_dma_off(platform_get_drvdata(pdev));
 	clk_disable(dw->clk);
+
 	return 0;
 }
 
@@ -1549,6 +1586,10 @@ static int dw_resume_noirq(struct device *dev)
 static const struct dev_pm_ops dw_dev_pm_ops = {
 	.suspend_noirq = dw_suspend_noirq,
 	.resume_noirq = dw_resume_noirq,
+	.freeze_noirq = dw_suspend_noirq,
+	.thaw_noirq = dw_resume_noirq,
+	.restore_noirq = dw_resume_noirq,
+	.poweroff_noirq = dw_suspend_noirq,
 };
 
 static struct platform_driver dw_driver = {
diff --git a/drivers/dma/dw_dmac_regs.h b/drivers/dma/dw_dmac_regs.h
index c3419518d701..f298f69ecbf9 100644
--- a/drivers/dma/dw_dmac_regs.h
+++ b/drivers/dma/dw_dmac_regs.h
@@ -13,6 +13,18 @@
 
 #define DW_DMA_MAX_NR_CHANNELS	8
 
+/* flow controller */
+enum dw_dma_fc {
+	DW_DMA_FC_D_M2M,
+	DW_DMA_FC_D_M2P,
+	DW_DMA_FC_D_P2M,
+	DW_DMA_FC_D_P2P,
+	DW_DMA_FC_P_P2M,
+	DW_DMA_FC_SP_P2P,
+	DW_DMA_FC_P_M2P,
+	DW_DMA_FC_DP_P2P,
+};
+
 /*
  * Redefine this macro to handle differences between 32- and 64-bit
  * addressing, big vs. little endian, etc.
@@ -140,18 +152,21 @@ struct dw_dma_chan {
 	u8			mask;
 	u8			priority;
 	bool			paused;
+	bool			initialized;
 
 	spinlock_t		lock;
 
 	/* these other elements are all protected by lock */
 	unsigned long		flags;
-	dma_cookie_t		completed;
 	struct list_head	active_list;
 	struct list_head	queue;
 	struct list_head	free_list;
 	struct dw_cyclic_desc	*cdesc;
 
 	unsigned int		descs_allocated;
+
+	/* configuration passed via DMA_SLAVE_CONFIG */
+	struct dma_slave_config dma_sconfig;
 };
 
 static inline struct dw_dma_chan_regs __iomem *
diff --git a/drivers/dma/ep93xx_dma.c b/drivers/dma/ep93xx_dma.c
index b47e2b803faf..e6f133b78dc2 100644
--- a/drivers/dma/ep93xx_dma.c
+++ b/drivers/dma/ep93xx_dma.c
@@ -28,6 +28,8 @@
 
 #include <mach/dma.h>
 
+#include "dmaengine.h"
+
 /* M2P registers */
 #define M2P_CONTROL			0x0000
 #define M2P_CONTROL_STALLINT		BIT(0)
@@ -122,7 +124,6 @@ struct ep93xx_dma_desc {
  * @lock: lock protecting the fields following
  * @flags: flags for the channel
  * @buffer: which buffer to use next (0/1)
- * @last_completed: last completed cookie value
  * @active: flattened chain of descriptors currently being processed
  * @queue: pending descriptors which are handled next
  * @free_list: list of free descriptors which can be used
@@ -157,7 +158,6 @@ struct ep93xx_dma_chan {
 #define EP93XX_DMA_IS_CYCLIC		0
 
 	int				buffer;
-	dma_cookie_t			last_completed;
 	struct list_head		active;
 	struct list_head		queue;
 	struct list_head		free_list;
@@ -246,6 +246,9 @@ static void ep93xx_dma_set_active(struct ep93xx_dma_chan *edmac,
 static struct ep93xx_dma_desc *
 ep93xx_dma_get_active(struct ep93xx_dma_chan *edmac)
 {
+	if (list_empty(&edmac->active))
+		return NULL;
+
 	return list_first_entry(&edmac->active, struct ep93xx_dma_desc, node);
 }
 
@@ -263,16 +266,22 @@ ep93xx_dma_get_active(struct ep93xx_dma_chan *edmac)
  */
 static bool ep93xx_dma_advance_active(struct ep93xx_dma_chan *edmac)
 {
+	struct ep93xx_dma_desc *desc;
+
 	list_rotate_left(&edmac->active);
 
 	if (test_bit(EP93XX_DMA_IS_CYCLIC, &edmac->flags))
 		return true;
 
+	desc = ep93xx_dma_get_active(edmac);
+	if (!desc)
+		return false;
+
 	/*
 	 * If txd.cookie is set it means that we are back in the first
 	 * descriptor in the chain and hence done with it.
 	 */
-	return !ep93xx_dma_get_active(edmac)->txd.cookie;
+	return !desc->txd.cookie;
 }
 
 /*
@@ -327,10 +336,16 @@ static void m2p_hw_shutdown(struct ep93xx_dma_chan *edmac)
 
 static void m2p_fill_desc(struct ep93xx_dma_chan *edmac)
 {
-	struct ep93xx_dma_desc *desc = ep93xx_dma_get_active(edmac);
+	struct ep93xx_dma_desc *desc;
 	u32 bus_addr;
 
-	if (ep93xx_dma_chan_direction(&edmac->chan) == DMA_TO_DEVICE)
+	desc = ep93xx_dma_get_active(edmac);
+	if (!desc) {
+		dev_warn(chan2dev(edmac), "M2P: empty descriptor list\n");
+		return;
+	}
+
+	if (ep93xx_dma_chan_direction(&edmac->chan) == DMA_MEM_TO_DEV)
 		bus_addr = desc->src_addr;
 	else
 		bus_addr = desc->dst_addr;
@@ -443,7 +458,7 @@ static int m2m_hw_setup(struct ep93xx_dma_chan *edmac)
 		control = (5 << M2M_CONTROL_PWSC_SHIFT);
 		control |= M2M_CONTROL_NO_HDSK;
 
-		if (data->direction == DMA_TO_DEVICE) {
+		if (data->direction == DMA_MEM_TO_DEV) {
 			control |= M2M_CONTROL_DAH;
 			control |= M2M_CONTROL_TM_TX;
 			control |= M2M_CONTROL_RSS_SSPTX;
@@ -459,11 +474,7 @@ static int m2m_hw_setup(struct ep93xx_dma_chan *edmac)
 		 * This IDE part is totally untested. Values below are taken
 		 * from the EP93xx Users's Guide and might not be correct.
 		 */
-		control |= M2M_CONTROL_NO_HDSK;
-		control |= M2M_CONTROL_RSS_IDE;
-		control |= M2M_CONTROL_PW_16;
-
-		if (data->direction == DMA_TO_DEVICE) {
+		if (data->direction == DMA_MEM_TO_DEV) {
 			/* Worst case from the UG */
 			control = (3 << M2M_CONTROL_PWSC_SHIFT);
 			control |= M2M_CONTROL_DAH;
@@ -473,6 +484,10 @@ static int m2m_hw_setup(struct ep93xx_dma_chan *edmac)
 			control |= M2M_CONTROL_SAH;
 			control |= M2M_CONTROL_TM_RX;
 		}
+
+		control |= M2M_CONTROL_NO_HDSK;
+		control |= M2M_CONTROL_RSS_IDE;
+		control |= M2M_CONTROL_PW_16;
 		break;
 
 	default:
@@ -491,7 +506,13 @@ static void m2m_hw_shutdown(struct ep93xx_dma_chan *edmac)
 
 static void m2m_fill_desc(struct ep93xx_dma_chan *edmac)
 {
-	struct ep93xx_dma_desc *desc = ep93xx_dma_get_active(edmac);
+	struct ep93xx_dma_desc *desc;
+
+	desc = ep93xx_dma_get_active(edmac);
+	if (!desc) {
+		dev_warn(chan2dev(edmac), "M2M: empty descriptor list\n");
+		return;
+	}
 
 	if (edmac->buffer == 0) {
 		writel(desc->src_addr, edmac->regs + M2M_SAR_BASE0);
@@ -669,24 +690,30 @@ static void ep93xx_dma_tasklet(unsigned long data)
 {
 	struct ep93xx_dma_chan *edmac = (struct ep93xx_dma_chan *)data;
 	struct ep93xx_dma_desc *desc, *d;
-	dma_async_tx_callback callback;
-	void *callback_param;
+	dma_async_tx_callback callback = NULL;
+	void *callback_param = NULL;
 	LIST_HEAD(list);
 
 	spin_lock_irq(&edmac->lock);
+	/*
+	 * If dma_terminate_all() was called before we get to run, the active
+	 * list has become empty. If that happens we aren't supposed to do
+	 * anything more than call ep93xx_dma_advance_work().
+	 */
 	desc = ep93xx_dma_get_active(edmac);
-	if (desc->complete) {
-		edmac->last_completed = desc->txd.cookie;
-		list_splice_init(&edmac->active, &list);
+	if (desc) {
+		if (desc->complete) {
+			dma_cookie_complete(&desc->txd);
+			list_splice_init(&edmac->active, &list);
+		}
+		callback = desc->txd.callback;
+		callback_param = desc->txd.callback_param;
 	}
 	spin_unlock_irq(&edmac->lock);
 
 	/* Pick up the next descriptor from the queue */
 	ep93xx_dma_advance_work(edmac);
 
-	callback = desc->txd.callback;
-	callback_param = desc->txd.callback_param;
-
 	/* Now we can release all the chained descriptors */
 	list_for_each_entry_safe(desc, d, &list, node) {
 		/*
@@ -706,13 +733,22 @@ static void ep93xx_dma_tasklet(unsigned long data)
 static irqreturn_t ep93xx_dma_interrupt(int irq, void *dev_id)
 {
 	struct ep93xx_dma_chan *edmac = dev_id;
+	struct ep93xx_dma_desc *desc;
 	irqreturn_t ret = IRQ_HANDLED;
 
 	spin_lock(&edmac->lock);
 
+	desc = ep93xx_dma_get_active(edmac);
+	if (!desc) {
+		dev_warn(chan2dev(edmac),
+			 "got interrupt while active list is empty\n");
+		spin_unlock(&edmac->lock);
+		return IRQ_NONE;
+	}
+
 	switch (edmac->edma->hw_interrupt(edmac)) {
 	case INTERRUPT_DONE:
-		ep93xx_dma_get_active(edmac)->complete = true;
+		desc->complete = true;
 		tasklet_schedule(&edmac->tasklet);
 		break;
 
@@ -747,17 +783,10 @@ static dma_cookie_t ep93xx_dma_tx_submit(struct dma_async_tx_descriptor *tx)
 	unsigned long flags;
 
 	spin_lock_irqsave(&edmac->lock, flags);
-
-	cookie = edmac->chan.cookie;
-
-	if (++cookie < 0)
-		cookie = 1;
+	cookie = dma_cookie_assign(tx);
 
 	desc = container_of(tx, struct ep93xx_dma_desc, txd);
 
-	edmac->chan.cookie = cookie;
-	desc->txd.cookie = cookie;
-
 	/*
 	 * If nothing is currently prosessed, we push this descriptor
 	 * directly to the hardware. Otherwise we put the descriptor
@@ -803,8 +832,8 @@ static int ep93xx_dma_alloc_chan_resources(struct dma_chan *chan)
 			switch (data->port) {
 			case EP93XX_DMA_SSP:
 			case EP93XX_DMA_IDE:
-				if (data->direction != DMA_TO_DEVICE &&
-				    data->direction != DMA_FROM_DEVICE)
+				if (data->direction != DMA_MEM_TO_DEV &&
+				    data->direction != DMA_DEV_TO_MEM)
 					return -EINVAL;
 				break;
 			default:
@@ -825,8 +854,7 @@ static int ep93xx_dma_alloc_chan_resources(struct dma_chan *chan)
 		goto fail_clk_disable;
 
 	spin_lock_irq(&edmac->lock);
-	edmac->last_completed = 1;
-	edmac->chan.cookie = 1;
+	dma_cookie_init(&edmac->chan);
 	ret = edmac->edma->hw_setup(edmac);
 	spin_unlock_irq(&edmac->lock);
 
@@ -947,13 +975,14 @@ fail:
  * @sg_len: number of entries in @sgl
  * @dir: direction of tha DMA transfer
  * @flags: flags for the descriptor
+ * @context: operation context (ignored)
  *
  * Returns a valid DMA descriptor or %NULL in case of failure.
  */
 static struct dma_async_tx_descriptor *
 ep93xx_dma_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
-			 unsigned int sg_len, enum dma_data_direction dir,
-			 unsigned long flags)
+			 unsigned int sg_len, enum dma_transfer_direction dir,
+			 unsigned long flags, void *context)
 {
 	struct ep93xx_dma_chan *edmac = to_ep93xx_dma_chan(chan);
 	struct ep93xx_dma_desc *desc, *first;
@@ -988,7 +1017,7 @@ ep93xx_dma_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 			goto fail;
 		}
 
-		if (dir == DMA_TO_DEVICE) {
+		if (dir == DMA_MEM_TO_DEV) {
 			desc->src_addr = sg_dma_address(sg);
 			desc->dst_addr = edmac->runtime_addr;
 		} else {
@@ -1020,6 +1049,7 @@ fail:
  * @buf_len: length of the buffer (in bytes)
  * @period_len: lenght of a single period
  * @dir: direction of the operation
+ * @context: operation context (ignored)
  *
  * Prepares a descriptor for cyclic DMA operation. This means that once the
  * descriptor is submitted, we will be submitting in a @period_len sized
@@ -1032,7 +1062,7 @@ fail:
 static struct dma_async_tx_descriptor *
 ep93xx_dma_prep_dma_cyclic(struct dma_chan *chan, dma_addr_t dma_addr,
 			   size_t buf_len, size_t period_len,
-			   enum dma_data_direction dir)
+			   enum dma_transfer_direction dir, void *context)
 {
 	struct ep93xx_dma_chan *edmac = to_ep93xx_dma_chan(chan);
 	struct ep93xx_dma_desc *desc, *first;
@@ -1065,7 +1095,7 @@ ep93xx_dma_prep_dma_cyclic(struct dma_chan *chan, dma_addr_t dma_addr,
 			goto fail;
 		}
 
-		if (dir == DMA_TO_DEVICE) {
+		if (dir == DMA_MEM_TO_DEV) {
 			desc->src_addr = dma_addr + offset;
 			desc->dst_addr = edmac->runtime_addr;
 		} else {
@@ -1133,12 +1163,12 @@ static int ep93xx_dma_slave_config(struct ep93xx_dma_chan *edmac,
 		return -EINVAL;
 
 	switch (config->direction) {
-	case DMA_FROM_DEVICE:
+	case DMA_DEV_TO_MEM:
 		width = config->src_addr_width;
 		addr = config->src_addr;
 		break;
 
-	case DMA_TO_DEVICE:
+	case DMA_MEM_TO_DEV:
 		width = config->dst_addr_width;
 		addr = config->dst_addr;
 		break;
@@ -1212,18 +1242,13 @@ static enum dma_status ep93xx_dma_tx_status(struct dma_chan *chan,
 					    struct dma_tx_state *state)
 {
 	struct ep93xx_dma_chan *edmac = to_ep93xx_dma_chan(chan);
-	dma_cookie_t last_used, last_completed;
 	enum dma_status ret;
 	unsigned long flags;
 
 	spin_lock_irqsave(&edmac->lock, flags);
-	last_used = chan->cookie;
-	last_completed = edmac->last_completed;
+	ret = dma_cookie_status(chan, cookie, state);
 	spin_unlock_irqrestore(&edmac->lock, flags);
 
-	ret = dma_async_is_complete(cookie, last_completed, last_used);
-	dma_set_tx_state(state, last_completed, last_used, 0);
-
 	return ret;
 }
 
diff --git a/drivers/dma/fsldma.c b/drivers/dma/fsldma.c
index 8a781540590c..8f84761f98ba 100644
--- a/drivers/dma/fsldma.c
+++ b/drivers/dma/fsldma.c
@@ -35,6 +35,7 @@
 #include <linux/dmapool.h>
 #include <linux/of_platform.h>
 
+#include "dmaengine.h"
 #include "fsldma.h"
 
 #define chan_dbg(chan, fmt, arg...)					\
@@ -413,17 +414,10 @@ static dma_cookie_t fsl_dma_tx_submit(struct dma_async_tx_descriptor *tx)
 	 * assign cookies to all of the software descriptors
 	 * that make up this transaction
 	 */
-	cookie = chan->common.cookie;
 	list_for_each_entry(child, &desc->tx_list, node) {
-		cookie++;
-		if (cookie < DMA_MIN_COOKIE)
-			cookie = DMA_MIN_COOKIE;
-
-		child->async_tx.cookie = cookie;
+		cookie = dma_cookie_assign(&child->async_tx);
 	}
 
-	chan->common.cookie = cookie;
-
 	/* put this transaction onto the tail of the pending queue */
 	append_ld_queue(chan, desc);
 
@@ -765,6 +759,7 @@ fail:
  * @sg_len: number of entries in @scatterlist
  * @direction: DMA direction
  * @flags: DMAEngine flags
+ * @context: transaction context (ignored)
  *
  * Prepare a set of descriptors for a DMA_SLAVE transaction. Following the
  * DMA_SLAVE API, this gets the device-specific information from the
@@ -772,7 +767,8 @@ fail:
  */
 static struct dma_async_tx_descriptor *fsl_dma_prep_slave_sg(
 	struct dma_chan *dchan, struct scatterlist *sgl, unsigned int sg_len,
-	enum dma_data_direction direction, unsigned long flags)
+	enum dma_transfer_direction direction, unsigned long flags,
+	void *context)
 {
 	/*
 	 * This operation is not supported on the Freescale DMA controller
@@ -819,7 +815,7 @@ static int fsl_dma_device_control(struct dma_chan *dchan,
 			return -ENXIO;
 
 		/* we set the controller burst size depending on direction */
-		if (config->direction == DMA_TO_DEVICE)
+		if (config->direction == DMA_MEM_TO_DEV)
 			size = config->dst_addr_width * config->dst_maxburst;
 		else
 			size = config->src_addr_width * config->src_maxburst;
@@ -984,19 +980,14 @@ static enum dma_status fsl_tx_status(struct dma_chan *dchan,
 					struct dma_tx_state *txstate)
 {
 	struct fsldma_chan *chan = to_fsl_chan(dchan);
-	dma_cookie_t last_complete;
-	dma_cookie_t last_used;
+	enum dma_status ret;
 	unsigned long flags;
 
 	spin_lock_irqsave(&chan->desc_lock, flags);
-
-	last_complete = chan->completed_cookie;
-	last_used = dchan->cookie;
-
+	ret = dma_cookie_status(dchan, cookie, txstate);
 	spin_unlock_irqrestore(&chan->desc_lock, flags);
 
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
-	return dma_async_is_complete(cookie, last_complete, last_used);
+	return ret;
 }
 
 /*----------------------------------------------------------------------------*/
@@ -1087,8 +1078,8 @@ static void dma_do_tasklet(unsigned long data)
 
 		desc = to_fsl_desc(chan->ld_running.prev);
 		cookie = desc->async_tx.cookie;
+		dma_cookie_complete(&desc->async_tx);
 
-		chan->completed_cookie = cookie;
 		chan_dbg(chan, "completed_cookie=%d\n", cookie);
 	}
 
@@ -1303,6 +1294,7 @@ static int __devinit fsl_dma_chan_probe(struct fsldma_device *fdev,
 	chan->idle = true;
 
 	chan->common.device = &fdev->common;
+	dma_cookie_init(&chan->common);
 
 	/* find the IRQ line, if it exists in the device tree */
 	chan->irq = irq_of_parse_and_map(node, 0);
diff --git a/drivers/dma/fsldma.h b/drivers/dma/fsldma.h
index 9cb5aa57c677..f5c38791fc74 100644
--- a/drivers/dma/fsldma.h
+++ b/drivers/dma/fsldma.h
@@ -137,7 +137,6 @@ struct fsldma_device {
 struct fsldma_chan {
 	char name[8];			/* Channel name */
 	struct fsldma_chan_regs __iomem *regs;
-	dma_cookie_t completed_cookie;	/* The maximum cookie completed */
 	spinlock_t desc_lock;		/* Descriptor operation lock */
 	struct list_head ld_pending;	/* Link descriptors queue */
 	struct list_head ld_running;	/* Link descriptors queue */
diff --git a/drivers/dma/imx-dma.c b/drivers/dma/imx-dma.c
index 4be55f9bb6c1..a45b5d2a5987 100644
--- a/drivers/dma/imx-dma.c
+++ b/drivers/dma/imx-dma.c
@@ -5,6 +5,7 @@
  * found on i.MX1/21/27
  *
  * Copyright 2010 Sascha Hauer, Pengutronix <s.hauer@pengutronix.de>
+ * Copyright 2012 Javier Martin, Vista Silicon <javier.martin@vista-silicon.com>
  *
  * The code contained herein is licensed under the GNU General Public
  * License. You may obtain a copy of the GNU General Public License
@@ -14,7 +15,6 @@
  * http://www.gnu.org/copyleft/gpl.html
  */
 #include <linux/init.h>
-#include <linux/module.h>
 #include <linux/types.h>
 #include <linux/mm.h>
 #include <linux/interrupt.h>
@@ -23,37 +23,159 @@
 #include <linux/dma-mapping.h>
 #include <linux/slab.h>
 #include <linux/platform_device.h>
+#include <linux/clk.h>
 #include <linux/dmaengine.h>
 #include <linux/module.h>
 
 #include <asm/irq.h>
-#include <mach/dma-v1.h>
+#include <mach/dma.h>
 #include <mach/hardware.h>
 
+#include "dmaengine.h"
+#define IMXDMA_MAX_CHAN_DESCRIPTORS	16
+#define IMX_DMA_CHANNELS  16
+
+#define IMX_DMA_2D_SLOTS	2
+#define IMX_DMA_2D_SLOT_A	0
+#define IMX_DMA_2D_SLOT_B	1
+
+#define IMX_DMA_LENGTH_LOOP	((unsigned int)-1)
+#define IMX_DMA_MEMSIZE_32	(0 << 4)
+#define IMX_DMA_MEMSIZE_8	(1 << 4)
+#define IMX_DMA_MEMSIZE_16	(2 << 4)
+#define IMX_DMA_TYPE_LINEAR	(0 << 10)
+#define IMX_DMA_TYPE_2D		(1 << 10)
+#define IMX_DMA_TYPE_FIFO	(2 << 10)
+
+#define IMX_DMA_ERR_BURST     (1 << 0)
+#define IMX_DMA_ERR_REQUEST   (1 << 1)
+#define IMX_DMA_ERR_TRANSFER  (1 << 2)
+#define IMX_DMA_ERR_BUFFER    (1 << 3)
+#define IMX_DMA_ERR_TIMEOUT   (1 << 4)
+
+#define DMA_DCR     0x00		/* Control Register */
+#define DMA_DISR    0x04		/* Interrupt status Register */
+#define DMA_DIMR    0x08		/* Interrupt mask Register */
+#define DMA_DBTOSR  0x0c		/* Burst timeout status Register */
+#define DMA_DRTOSR  0x10		/* Request timeout Register */
+#define DMA_DSESR   0x14		/* Transfer Error Status Register */
+#define DMA_DBOSR   0x18		/* Buffer overflow status Register */
+#define DMA_DBTOCR  0x1c		/* Burst timeout control Register */
+#define DMA_WSRA    0x40		/* W-Size Register A */
+#define DMA_XSRA    0x44		/* X-Size Register A */
+#define DMA_YSRA    0x48		/* Y-Size Register A */
+#define DMA_WSRB    0x4c		/* W-Size Register B */
+#define DMA_XSRB    0x50		/* X-Size Register B */
+#define DMA_YSRB    0x54		/* Y-Size Register B */
+#define DMA_SAR(x)  (0x80 + ((x) << 6))	/* Source Address Registers */
+#define DMA_DAR(x)  (0x84 + ((x) << 6))	/* Destination Address Registers */
+#define DMA_CNTR(x) (0x88 + ((x) << 6))	/* Count Registers */
+#define DMA_CCR(x)  (0x8c + ((x) << 6))	/* Control Registers */
+#define DMA_RSSR(x) (0x90 + ((x) << 6))	/* Request source select Registers */
+#define DMA_BLR(x)  (0x94 + ((x) << 6))	/* Burst length Registers */
+#define DMA_RTOR(x) (0x98 + ((x) << 6))	/* Request timeout Registers */
+#define DMA_BUCR(x) (0x98 + ((x) << 6))	/* Bus Utilization Registers */
+#define DMA_CCNR(x) (0x9C + ((x) << 6))	/* Channel counter Registers */
+
+#define DCR_DRST           (1<<1)
+#define DCR_DEN            (1<<0)
+#define DBTOCR_EN          (1<<15)
+#define DBTOCR_CNT(x)      ((x) & 0x7fff)
+#define CNTR_CNT(x)        ((x) & 0xffffff)
+#define CCR_ACRPT          (1<<14)
+#define CCR_DMOD_LINEAR    (0x0 << 12)
+#define CCR_DMOD_2D        (0x1 << 12)
+#define CCR_DMOD_FIFO      (0x2 << 12)
+#define CCR_DMOD_EOBFIFO   (0x3 << 12)
+#define CCR_SMOD_LINEAR    (0x0 << 10)
+#define CCR_SMOD_2D        (0x1 << 10)
+#define CCR_SMOD_FIFO      (0x2 << 10)
+#define CCR_SMOD_EOBFIFO   (0x3 << 10)
+#define CCR_MDIR_DEC       (1<<9)
+#define CCR_MSEL_B         (1<<8)
+#define CCR_DSIZ_32        (0x0 << 6)
+#define CCR_DSIZ_8         (0x1 << 6)
+#define CCR_DSIZ_16        (0x2 << 6)
+#define CCR_SSIZ_32        (0x0 << 4)
+#define CCR_SSIZ_8         (0x1 << 4)
+#define CCR_SSIZ_16        (0x2 << 4)
+#define CCR_REN            (1<<3)
+#define CCR_RPT            (1<<2)
+#define CCR_FRC            (1<<1)
+#define CCR_CEN            (1<<0)
+#define RTOR_EN            (1<<15)
+#define RTOR_CLK           (1<<14)
+#define RTOR_PSC           (1<<13)
+
+enum  imxdma_prep_type {
+	IMXDMA_DESC_MEMCPY,
+	IMXDMA_DESC_INTERLEAVED,
+	IMXDMA_DESC_SLAVE_SG,
+	IMXDMA_DESC_CYCLIC,
+};
+
+struct imx_dma_2d_config {
+	u16		xsr;
+	u16		ysr;
+	u16		wsr;
+	int		count;
+};
+
+struct imxdma_desc {
+	struct list_head		node;
+	struct dma_async_tx_descriptor	desc;
+	enum dma_status			status;
+	dma_addr_t			src;
+	dma_addr_t			dest;
+	size_t				len;
+	enum dma_transfer_direction	direction;
+	enum imxdma_prep_type		type;
+	/* For memcpy and interleaved */
+	unsigned int			config_port;
+	unsigned int			config_mem;
+	/* For interleaved transfers */
+	unsigned int			x;
+	unsigned int			y;
+	unsigned int			w;
+	/* For slave sg and cyclic */
+	struct scatterlist		*sg;
+	unsigned int			sgcount;
+};
+
 struct imxdma_channel {
+	int				hw_chaining;
+	struct timer_list		watchdog;
 	struct imxdma_engine		*imxdma;
 	unsigned int			channel;
-	unsigned int			imxdma_channel;
 
+	struct tasklet_struct		dma_tasklet;
+	struct list_head		ld_free;
+	struct list_head		ld_queue;
+	struct list_head		ld_active;
+	int				descs_allocated;
 	enum dma_slave_buswidth		word_size;
 	dma_addr_t			per_address;
 	u32				watermark_level;
 	struct dma_chan			chan;
-	spinlock_t			lock;
 	struct dma_async_tx_descriptor	desc;
-	dma_cookie_t			last_completed;
 	enum dma_status			status;
 	int				dma_request;
 	struct scatterlist		*sg_list;
+	u32				ccr_from_device;
+	u32				ccr_to_device;
+	bool				enabled_2d;
+	int				slot_2d;
 };
 
-#define MAX_DMA_CHANNELS 8
-
 struct imxdma_engine {
 	struct device			*dev;
 	struct device_dma_parameters	dma_parms;
 	struct dma_device		dma_device;
-	struct imxdma_channel		channel[MAX_DMA_CHANNELS];
+	void __iomem			*base;
+	struct clk			*dma_clk;
+	spinlock_t			lock;
+	struct imx_dma_2d_config	slots_2d[IMX_DMA_2D_SLOTS];
+	struct imxdma_channel		channel[IMX_DMA_CHANNELS];
 };
 
 static struct imxdma_channel *to_imxdma_chan(struct dma_chan *chan)
@@ -61,36 +183,418 @@ static struct imxdma_channel *to_imxdma_chan(struct dma_chan *chan)
 	return container_of(chan, struct imxdma_channel, chan);
 }
 
-static void imxdma_handle(struct imxdma_channel *imxdmac)
+static inline bool imxdma_chan_is_doing_cyclic(struct imxdma_channel *imxdmac)
+{
+	struct imxdma_desc *desc;
+
+	if (!list_empty(&imxdmac->ld_active)) {
+		desc = list_first_entry(&imxdmac->ld_active, struct imxdma_desc,
+					node);
+		if (desc->type == IMXDMA_DESC_CYCLIC)
+			return true;
+	}
+	return false;
+}
+
+
+
+static void imx_dmav1_writel(struct imxdma_engine *imxdma, unsigned val,
+			     unsigned offset)
+{
+	__raw_writel(val, imxdma->base + offset);
+}
+
+static unsigned imx_dmav1_readl(struct imxdma_engine *imxdma, unsigned offset)
+{
+	return __raw_readl(imxdma->base + offset);
+}
+
+static int imxdma_hw_chain(struct imxdma_channel *imxdmac)
+{
+	if (cpu_is_mx27())
+		return imxdmac->hw_chaining;
+	else
+		return 0;
+}
+
+/*
+ * imxdma_sg_next - prepare next chunk for scatter-gather DMA emulation
+ */
+static inline int imxdma_sg_next(struct imxdma_desc *d)
+{
+	struct imxdma_channel *imxdmac = to_imxdma_chan(d->desc.chan);
+	struct imxdma_engine *imxdma = imxdmac->imxdma;
+	struct scatterlist *sg = d->sg;
+	unsigned long now;
+
+	now = min(d->len, sg->length);
+	if (d->len != IMX_DMA_LENGTH_LOOP)
+		d->len -= now;
+
+	if (d->direction == DMA_DEV_TO_MEM)
+		imx_dmav1_writel(imxdma, sg->dma_address,
+				 DMA_DAR(imxdmac->channel));
+	else
+		imx_dmav1_writel(imxdma, sg->dma_address,
+				 DMA_SAR(imxdmac->channel));
+
+	imx_dmav1_writel(imxdma, now, DMA_CNTR(imxdmac->channel));
+
+	dev_dbg(imxdma->dev, " %s channel: %d dst 0x%08x, src 0x%08x, "
+		"size 0x%08x\n", __func__, imxdmac->channel,
+		 imx_dmav1_readl(imxdma, DMA_DAR(imxdmac->channel)),
+		 imx_dmav1_readl(imxdma, DMA_SAR(imxdmac->channel)),
+		 imx_dmav1_readl(imxdma, DMA_CNTR(imxdmac->channel)));
+
+	return now;
+}
+
+static void imxdma_enable_hw(struct imxdma_desc *d)
+{
+	struct imxdma_channel *imxdmac = to_imxdma_chan(d->desc.chan);
+	struct imxdma_engine *imxdma = imxdmac->imxdma;
+	int channel = imxdmac->channel;
+	unsigned long flags;
+
+	dev_dbg(imxdma->dev, "%s channel %d\n", __func__, channel);
+
+	local_irq_save(flags);
+
+	imx_dmav1_writel(imxdma, 1 << channel, DMA_DISR);
+	imx_dmav1_writel(imxdma, imx_dmav1_readl(imxdma, DMA_DIMR) &
+			 ~(1 << channel), DMA_DIMR);
+	imx_dmav1_writel(imxdma, imx_dmav1_readl(imxdma, DMA_CCR(channel)) |
+			 CCR_CEN | CCR_ACRPT, DMA_CCR(channel));
+
+	if ((cpu_is_mx21() || cpu_is_mx27()) &&
+			d->sg && imxdma_hw_chain(imxdmac)) {
+		d->sg = sg_next(d->sg);
+		if (d->sg) {
+			u32 tmp;
+			imxdma_sg_next(d);
+			tmp = imx_dmav1_readl(imxdma, DMA_CCR(channel));
+			imx_dmav1_writel(imxdma, tmp | CCR_RPT | CCR_ACRPT,
+					 DMA_CCR(channel));
+		}
+	}
+
+	local_irq_restore(flags);
+}
+
+static void imxdma_disable_hw(struct imxdma_channel *imxdmac)
+{
+	struct imxdma_engine *imxdma = imxdmac->imxdma;
+	int channel = imxdmac->channel;
+	unsigned long flags;
+
+	dev_dbg(imxdma->dev, "%s channel %d\n", __func__, channel);
+
+	if (imxdma_hw_chain(imxdmac))
+		del_timer(&imxdmac->watchdog);
+
+	local_irq_save(flags);
+	imx_dmav1_writel(imxdma, imx_dmav1_readl(imxdma, DMA_DIMR) |
+			 (1 << channel), DMA_DIMR);
+	imx_dmav1_writel(imxdma, imx_dmav1_readl(imxdma, DMA_CCR(channel)) &
+			 ~CCR_CEN, DMA_CCR(channel));
+	imx_dmav1_writel(imxdma, 1 << channel, DMA_DISR);
+	local_irq_restore(flags);
+}
+
+static void imxdma_watchdog(unsigned long data)
 {
-	if (imxdmac->desc.callback)
-		imxdmac->desc.callback(imxdmac->desc.callback_param);
-	imxdmac->last_completed = imxdmac->desc.cookie;
+	struct imxdma_channel *imxdmac = (struct imxdma_channel *)data;
+	struct imxdma_engine *imxdma = imxdmac->imxdma;
+	int channel = imxdmac->channel;
+
+	imx_dmav1_writel(imxdma, 0, DMA_CCR(channel));
+
+	/* Tasklet watchdog error handler */
+	tasklet_schedule(&imxdmac->dma_tasklet);
+	dev_dbg(imxdma->dev, "channel %d: watchdog timeout!\n",
+		imxdmac->channel);
 }
 
-static void imxdma_irq_handler(int channel, void *data)
+static irqreturn_t imxdma_err_handler(int irq, void *dev_id)
 {
-	struct imxdma_channel *imxdmac = data;
+	struct imxdma_engine *imxdma = dev_id;
+	unsigned int err_mask;
+	int i, disr;
+	int errcode;
+
+	disr = imx_dmav1_readl(imxdma, DMA_DISR);
+
+	err_mask = imx_dmav1_readl(imxdma, DMA_DBTOSR) |
+		   imx_dmav1_readl(imxdma, DMA_DRTOSR) |
+		   imx_dmav1_readl(imxdma, DMA_DSESR)  |
+		   imx_dmav1_readl(imxdma, DMA_DBOSR);
 
-	imxdmac->status = DMA_SUCCESS;
-	imxdma_handle(imxdmac);
+	if (!err_mask)
+		return IRQ_HANDLED;
+
+	imx_dmav1_writel(imxdma, disr & err_mask, DMA_DISR);
+
+	for (i = 0; i < IMX_DMA_CHANNELS; i++) {
+		if (!(err_mask & (1 << i)))
+			continue;
+		errcode = 0;
+
+		if (imx_dmav1_readl(imxdma, DMA_DBTOSR) & (1 << i)) {
+			imx_dmav1_writel(imxdma, 1 << i, DMA_DBTOSR);
+			errcode |= IMX_DMA_ERR_BURST;
+		}
+		if (imx_dmav1_readl(imxdma, DMA_DRTOSR) & (1 << i)) {
+			imx_dmav1_writel(imxdma, 1 << i, DMA_DRTOSR);
+			errcode |= IMX_DMA_ERR_REQUEST;
+		}
+		if (imx_dmav1_readl(imxdma, DMA_DSESR) & (1 << i)) {
+			imx_dmav1_writel(imxdma, 1 << i, DMA_DSESR);
+			errcode |= IMX_DMA_ERR_TRANSFER;
+		}
+		if (imx_dmav1_readl(imxdma, DMA_DBOSR) & (1 << i)) {
+			imx_dmav1_writel(imxdma, 1 << i, DMA_DBOSR);
+			errcode |= IMX_DMA_ERR_BUFFER;
+		}
+		/* Tasklet error handler */
+		tasklet_schedule(&imxdma->channel[i].dma_tasklet);
+
+		printk(KERN_WARNING
+		       "DMA timeout on channel %d -%s%s%s%s\n", i,
+		       errcode & IMX_DMA_ERR_BURST ?    " burst" : "",
+		       errcode & IMX_DMA_ERR_REQUEST ?  " request" : "",
+		       errcode & IMX_DMA_ERR_TRANSFER ? " transfer" : "",
+		       errcode & IMX_DMA_ERR_BUFFER ?   " buffer" : "");
+	}
+	return IRQ_HANDLED;
 }
 
-static void imxdma_err_handler(int channel, void *data, int error)
+static void dma_irq_handle_channel(struct imxdma_channel *imxdmac)
 {
-	struct imxdma_channel *imxdmac = data;
+	struct imxdma_engine *imxdma = imxdmac->imxdma;
+	int chno = imxdmac->channel;
+	struct imxdma_desc *desc;
+
+	spin_lock(&imxdma->lock);
+	if (list_empty(&imxdmac->ld_active)) {
+		spin_unlock(&imxdma->lock);
+		goto out;
+	}
+
+	desc = list_first_entry(&imxdmac->ld_active,
+				struct imxdma_desc,
+				node);
+	spin_unlock(&imxdma->lock);
+
+	if (desc->sg) {
+		u32 tmp;
+		desc->sg = sg_next(desc->sg);
+
+		if (desc->sg) {
+			imxdma_sg_next(desc);
+
+			tmp = imx_dmav1_readl(imxdma, DMA_CCR(chno));
+
+			if (imxdma_hw_chain(imxdmac)) {
+				/* FIXME: The timeout should probably be
+				 * configurable
+				 */
+				mod_timer(&imxdmac->watchdog,
+					jiffies + msecs_to_jiffies(500));
+
+				tmp |= CCR_CEN | CCR_RPT | CCR_ACRPT;
+				imx_dmav1_writel(imxdma, tmp, DMA_CCR(chno));
+			} else {
+				imx_dmav1_writel(imxdma, tmp & ~CCR_CEN,
+						 DMA_CCR(chno));
+				tmp |= CCR_CEN;
+			}
+
+			imx_dmav1_writel(imxdma, tmp, DMA_CCR(chno));
+
+			if (imxdma_chan_is_doing_cyclic(imxdmac))
+				/* Tasklet progression */
+				tasklet_schedule(&imxdmac->dma_tasklet);
+
+			return;
+		}
+
+		if (imxdma_hw_chain(imxdmac)) {
+			del_timer(&imxdmac->watchdog);
+			return;
+		}
+	}
 
-	imxdmac->status = DMA_ERROR;
-	imxdma_handle(imxdmac);
+out:
+	imx_dmav1_writel(imxdma, 0, DMA_CCR(chno));
+	/* Tasklet irq */
+	tasklet_schedule(&imxdmac->dma_tasklet);
 }
 
-static void imxdma_progression(int channel, void *data,
-		struct scatterlist *sg)
+static irqreturn_t dma_irq_handler(int irq, void *dev_id)
 {
-	struct imxdma_channel *imxdmac = data;
+	struct imxdma_engine *imxdma = dev_id;
+	int i, disr;
+
+	if (cpu_is_mx21() || cpu_is_mx27())
+		imxdma_err_handler(irq, dev_id);
+
+	disr = imx_dmav1_readl(imxdma, DMA_DISR);
+
+	dev_dbg(imxdma->dev, "%s called, disr=0x%08x\n", __func__, disr);
 
-	imxdmac->status = DMA_SUCCESS;
-	imxdma_handle(imxdmac);
+	imx_dmav1_writel(imxdma, disr, DMA_DISR);
+	for (i = 0; i < IMX_DMA_CHANNELS; i++) {
+		if (disr & (1 << i))
+			dma_irq_handle_channel(&imxdma->channel[i]);
+	}
+
+	return IRQ_HANDLED;
+}
+
+static int imxdma_xfer_desc(struct imxdma_desc *d)
+{
+	struct imxdma_channel *imxdmac = to_imxdma_chan(d->desc.chan);
+	struct imxdma_engine *imxdma = imxdmac->imxdma;
+	unsigned long flags;
+	int slot = -1;
+	int i;
+
+	/* Configure and enable */
+	switch (d->type) {
+	case IMXDMA_DESC_INTERLEAVED:
+		/* Try to get a free 2D slot */
+		spin_lock_irqsave(&imxdma->lock, flags);
+		for (i = 0; i < IMX_DMA_2D_SLOTS; i++) {
+			if ((imxdma->slots_2d[i].count > 0) &&
+			((imxdma->slots_2d[i].xsr != d->x) ||
+			(imxdma->slots_2d[i].ysr != d->y) ||
+			(imxdma->slots_2d[i].wsr != d->w)))
+				continue;
+			slot = i;
+			break;
+		}
+		if (slot < 0)
+			return -EBUSY;
+
+		imxdma->slots_2d[slot].xsr = d->x;
+		imxdma->slots_2d[slot].ysr = d->y;
+		imxdma->slots_2d[slot].wsr = d->w;
+		imxdma->slots_2d[slot].count++;
+
+		imxdmac->slot_2d = slot;
+		imxdmac->enabled_2d = true;
+		spin_unlock_irqrestore(&imxdma->lock, flags);
+
+		if (slot == IMX_DMA_2D_SLOT_A) {
+			d->config_mem &= ~CCR_MSEL_B;
+			d->config_port &= ~CCR_MSEL_B;
+			imx_dmav1_writel(imxdma, d->x, DMA_XSRA);
+			imx_dmav1_writel(imxdma, d->y, DMA_YSRA);
+			imx_dmav1_writel(imxdma, d->w, DMA_WSRA);
+		} else {
+			d->config_mem |= CCR_MSEL_B;
+			d->config_port |= CCR_MSEL_B;
+			imx_dmav1_writel(imxdma, d->x, DMA_XSRB);
+			imx_dmav1_writel(imxdma, d->y, DMA_YSRB);
+			imx_dmav1_writel(imxdma, d->w, DMA_WSRB);
+		}
+		/*
+		 * We fall-through here intentionally, since a 2D transfer is
+		 * similar to MEMCPY just adding the 2D slot configuration.
+		 */
+	case IMXDMA_DESC_MEMCPY:
+		imx_dmav1_writel(imxdma, d->src, DMA_SAR(imxdmac->channel));
+		imx_dmav1_writel(imxdma, d->dest, DMA_DAR(imxdmac->channel));
+		imx_dmav1_writel(imxdma, d->config_mem | (d->config_port << 2),
+			 DMA_CCR(imxdmac->channel));
+
+		imx_dmav1_writel(imxdma, d->len, DMA_CNTR(imxdmac->channel));
+
+		dev_dbg(imxdma->dev, "%s channel: %d dest=0x%08x src=0x%08x "
+			"dma_length=%d\n", __func__, imxdmac->channel,
+			d->dest, d->src, d->len);
+
+		break;
+	/* Cyclic transfer is the same as slave_sg with special sg configuration. */
+	case IMXDMA_DESC_CYCLIC:
+	case IMXDMA_DESC_SLAVE_SG:
+		if (d->direction == DMA_DEV_TO_MEM) {
+			imx_dmav1_writel(imxdma, imxdmac->per_address,
+					 DMA_SAR(imxdmac->channel));
+			imx_dmav1_writel(imxdma, imxdmac->ccr_from_device,
+					 DMA_CCR(imxdmac->channel));
+
+			dev_dbg(imxdma->dev, "%s channel: %d sg=%p sgcount=%d "
+				"total length=%d dev_addr=0x%08x (dev2mem)\n",
+				__func__, imxdmac->channel, d->sg, d->sgcount,
+				d->len, imxdmac->per_address);
+		} else if (d->direction == DMA_MEM_TO_DEV) {
+			imx_dmav1_writel(imxdma, imxdmac->per_address,
+					 DMA_DAR(imxdmac->channel));
+			imx_dmav1_writel(imxdma, imxdmac->ccr_to_device,
+					 DMA_CCR(imxdmac->channel));
+
+			dev_dbg(imxdma->dev, "%s channel: %d sg=%p sgcount=%d "
+				"total length=%d dev_addr=0x%08x (mem2dev)\n",
+				__func__, imxdmac->channel, d->sg, d->sgcount,
+				d->len, imxdmac->per_address);
+		} else {
+			dev_err(imxdma->dev, "%s channel: %d bad dma mode\n",
+				__func__, imxdmac->channel);
+			return -EINVAL;
+		}
+
+		imxdma_sg_next(d);
+
+		break;
+	default:
+		return -EINVAL;
+	}
+	imxdma_enable_hw(d);
+	return 0;
+}
+
+static void imxdma_tasklet(unsigned long data)
+{
+	struct imxdma_channel *imxdmac = (void *)data;
+	struct imxdma_engine *imxdma = imxdmac->imxdma;
+	struct imxdma_desc *desc;
+
+	spin_lock(&imxdma->lock);
+
+	if (list_empty(&imxdmac->ld_active)) {
+		/* Someone might have called terminate all */
+		goto out;
+	}
+	desc = list_first_entry(&imxdmac->ld_active, struct imxdma_desc, node);
+
+	if (desc->desc.callback)
+		desc->desc.callback(desc->desc.callback_param);
+
+	dma_cookie_complete(&desc->desc);
+
+	/* If we are dealing with a cyclic descriptor keep it on ld_active */
+	if (imxdma_chan_is_doing_cyclic(imxdmac))
+		goto out;
+
+	/* Free 2D slot if it was an interleaved transfer */
+	if (imxdmac->enabled_2d) {
+		imxdma->slots_2d[imxdmac->slot_2d].count--;
+		imxdmac->enabled_2d = false;
+	}
+
+	list_move_tail(imxdmac->ld_active.next, &imxdmac->ld_free);
+
+	if (!list_empty(&imxdmac->ld_queue)) {
+		desc = list_first_entry(&imxdmac->ld_queue, struct imxdma_desc,
+					node);
+		list_move_tail(imxdmac->ld_queue.next, &imxdmac->ld_active);
+		if (imxdma_xfer_desc(desc) < 0)
+			dev_warn(imxdma->dev, "%s: channel: %d couldn't xfer desc\n",
+				 __func__, imxdmac->channel);
+	}
+out:
+	spin_unlock(&imxdma->lock);
 }
 
 static int imxdma_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
@@ -98,16 +602,21 @@ static int imxdma_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
 {
 	struct imxdma_channel *imxdmac = to_imxdma_chan(chan);
 	struct dma_slave_config *dmaengine_cfg = (void *)arg;
-	int ret;
+	struct imxdma_engine *imxdma = imxdmac->imxdma;
+	unsigned long flags;
 	unsigned int mode = 0;
 
 	switch (cmd) {
 	case DMA_TERMINATE_ALL:
-		imxdmac->status = DMA_ERROR;
-		imx_dma_disable(imxdmac->imxdma_channel);
+		imxdma_disable_hw(imxdmac);
+
+		spin_lock_irqsave(&imxdma->lock, flags);
+		list_splice_tail_init(&imxdmac->ld_active, &imxdmac->ld_free);
+		list_splice_tail_init(&imxdmac->ld_queue, &imxdmac->ld_free);
+		spin_unlock_irqrestore(&imxdma->lock, flags);
 		return 0;
 	case DMA_SLAVE_CONFIG:
-		if (dmaengine_cfg->direction == DMA_FROM_DEVICE) {
+		if (dmaengine_cfg->direction == DMA_DEV_TO_MEM) {
 			imxdmac->per_address = dmaengine_cfg->src_addr;
 			imxdmac->watermark_level = dmaengine_cfg->src_maxburst;
 			imxdmac->word_size = dmaengine_cfg->src_addr_width;
@@ -129,16 +638,22 @@ static int imxdma_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
 			mode = IMX_DMA_MEMSIZE_32;
 			break;
 		}
-		ret = imx_dma_config_channel(imxdmac->imxdma_channel,
-				mode | IMX_DMA_TYPE_FIFO,
-				IMX_DMA_MEMSIZE_32 | IMX_DMA_TYPE_LINEAR,
-				imxdmac->dma_request, 1);
-
-		if (ret)
-			return ret;
 
-		imx_dma_config_burstlen(imxdmac->imxdma_channel,
-				imxdmac->watermark_level * imxdmac->word_size);
+		imxdmac->hw_chaining = 1;
+		if (!imxdma_hw_chain(imxdmac))
+			return -EINVAL;
+		imxdmac->ccr_from_device = (mode | IMX_DMA_TYPE_FIFO) |
+			((IMX_DMA_MEMSIZE_32 | IMX_DMA_TYPE_LINEAR) << 2) |
+			CCR_REN;
+		imxdmac->ccr_to_device =
+			(IMX_DMA_MEMSIZE_32 | IMX_DMA_TYPE_LINEAR) |
+			((mode | IMX_DMA_TYPE_FIFO) << 2) | CCR_REN;
+		imx_dmav1_writel(imxdma, imxdmac->dma_request,
+				 DMA_RSSR(imxdmac->channel));
+
+		/* Set burst length */
+		imx_dmav1_writel(imxdma, imxdmac->watermark_level *
+				imxdmac->word_size, DMA_BLR(imxdmac->channel));
 
 		return 0;
 	default:
@@ -152,43 +667,20 @@ static enum dma_status imxdma_tx_status(struct dma_chan *chan,
 					    dma_cookie_t cookie,
 					    struct dma_tx_state *txstate)
 {
-	struct imxdma_channel *imxdmac = to_imxdma_chan(chan);
-	dma_cookie_t last_used;
-	enum dma_status ret;
-
-	last_used = chan->cookie;
-
-	ret = dma_async_is_complete(cookie, imxdmac->last_completed, last_used);
-	dma_set_tx_state(txstate, imxdmac->last_completed, last_used, 0);
-
-	return ret;
-}
-
-static dma_cookie_t imxdma_assign_cookie(struct imxdma_channel *imxdma)
-{
-	dma_cookie_t cookie = imxdma->chan.cookie;
-
-	if (++cookie < 0)
-		cookie = 1;
-
-	imxdma->chan.cookie = cookie;
-	imxdma->desc.cookie = cookie;
-
-	return cookie;
+	return dma_cookie_status(chan, cookie, txstate);
 }
 
 static dma_cookie_t imxdma_tx_submit(struct dma_async_tx_descriptor *tx)
 {
 	struct imxdma_channel *imxdmac = to_imxdma_chan(tx->chan);
+	struct imxdma_engine *imxdma = imxdmac->imxdma;
 	dma_cookie_t cookie;
+	unsigned long flags;
 
-	spin_lock_irq(&imxdmac->lock);
-
-	cookie = imxdma_assign_cookie(imxdmac);
-
-	imx_dma_enable(imxdmac->imxdma_channel);
-
-	spin_unlock_irq(&imxdmac->lock);
+	spin_lock_irqsave(&imxdma->lock, flags);
+	list_move_tail(imxdmac->ld_free.next, &imxdmac->ld_queue);
+	cookie = dma_cookie_assign(tx);
+	spin_unlock_irqrestore(&imxdma->lock, flags);
 
 	return cookie;
 }
@@ -198,23 +690,52 @@ static int imxdma_alloc_chan_resources(struct dma_chan *chan)
 	struct imxdma_channel *imxdmac = to_imxdma_chan(chan);
 	struct imx_dma_data *data = chan->private;
 
-	imxdmac->dma_request = data->dma_request;
+	if (data != NULL)
+		imxdmac->dma_request = data->dma_request;
 
-	dma_async_tx_descriptor_init(&imxdmac->desc, chan);
-	imxdmac->desc.tx_submit = imxdma_tx_submit;
-	/* txd.flags will be overwritten in prep funcs */
-	imxdmac->desc.flags = DMA_CTRL_ACK;
+	while (imxdmac->descs_allocated < IMXDMA_MAX_CHAN_DESCRIPTORS) {
+		struct imxdma_desc *desc;
 
-	imxdmac->status = DMA_SUCCESS;
+		desc = kzalloc(sizeof(*desc), GFP_KERNEL);
+		if (!desc)
+			break;
+		__memzero(&desc->desc, sizeof(struct dma_async_tx_descriptor));
+		dma_async_tx_descriptor_init(&desc->desc, chan);
+		desc->desc.tx_submit = imxdma_tx_submit;
+		/* txd.flags will be overwritten in prep funcs */
+		desc->desc.flags = DMA_CTRL_ACK;
+		desc->status = DMA_SUCCESS;
+
+		list_add_tail(&desc->node, &imxdmac->ld_free);
+		imxdmac->descs_allocated++;
+	}
 
-	return 0;
+	if (!imxdmac->descs_allocated)
+		return -ENOMEM;
+
+	return imxdmac->descs_allocated;
 }
 
 static void imxdma_free_chan_resources(struct dma_chan *chan)
 {
 	struct imxdma_channel *imxdmac = to_imxdma_chan(chan);
+	struct imxdma_engine *imxdma = imxdmac->imxdma;
+	struct imxdma_desc *desc, *_desc;
+	unsigned long flags;
+
+	spin_lock_irqsave(&imxdma->lock, flags);
+
+	imxdma_disable_hw(imxdmac);
+	list_splice_tail_init(&imxdmac->ld_active, &imxdmac->ld_free);
+	list_splice_tail_init(&imxdmac->ld_queue, &imxdmac->ld_free);
 
-	imx_dma_disable(imxdmac->imxdma_channel);
+	spin_unlock_irqrestore(&imxdma->lock, flags);
+
+	list_for_each_entry_safe(desc, _desc, &imxdmac->ld_free, node) {
+		kfree(desc);
+		imxdmac->descs_allocated--;
+	}
+	INIT_LIST_HEAD(&imxdmac->ld_free);
 
 	if (imxdmac->sg_list) {
 		kfree(imxdmac->sg_list);
@@ -224,28 +745,24 @@ static void imxdma_free_chan_resources(struct dma_chan *chan)
 
 static struct dma_async_tx_descriptor *imxdma_prep_slave_sg(
 		struct dma_chan *chan, struct scatterlist *sgl,
-		unsigned int sg_len, enum dma_data_direction direction,
-		unsigned long flags)
+		unsigned int sg_len, enum dma_transfer_direction direction,
+		unsigned long flags, void *context)
 {
 	struct imxdma_channel *imxdmac = to_imxdma_chan(chan);
 	struct scatterlist *sg;
-	int i, ret, dma_length = 0;
-	unsigned int dmamode;
+	int i, dma_length = 0;
+	struct imxdma_desc *desc;
 
-	if (imxdmac->status == DMA_IN_PROGRESS)
+	if (list_empty(&imxdmac->ld_free) ||
+	    imxdma_chan_is_doing_cyclic(imxdmac))
 		return NULL;
 
-	imxdmac->status = DMA_IN_PROGRESS;
+	desc = list_first_entry(&imxdmac->ld_free, struct imxdma_desc, node);
 
 	for_each_sg(sgl, sg, sg_len, i) {
 		dma_length += sg->length;
 	}
 
-	if (direction == DMA_FROM_DEVICE)
-		dmamode = DMA_MODE_READ;
-	else
-		dmamode = DMA_MODE_WRITE;
-
 	switch (imxdmac->word_size) {
 	case DMA_SLAVE_BUSWIDTH_4_BYTES:
 		if (sgl->length & 3 || sgl->dma_address & 3)
@@ -261,37 +778,41 @@ static struct dma_async_tx_descriptor *imxdma_prep_slave_sg(
 		return NULL;
 	}
 
-	ret = imx_dma_setup_sg(imxdmac->imxdma_channel, sgl, sg_len,
-		 dma_length, imxdmac->per_address, dmamode);
-	if (ret)
-		return NULL;
+	desc->type = IMXDMA_DESC_SLAVE_SG;
+	desc->sg = sgl;
+	desc->sgcount = sg_len;
+	desc->len = dma_length;
+	desc->direction = direction;
+	if (direction == DMA_DEV_TO_MEM) {
+		desc->src = imxdmac->per_address;
+	} else {
+		desc->dest = imxdmac->per_address;
+	}
+	desc->desc.callback = NULL;
+	desc->desc.callback_param = NULL;
 
-	return &imxdmac->desc;
+	return &desc->desc;
 }
 
 static struct dma_async_tx_descriptor *imxdma_prep_dma_cyclic(
 		struct dma_chan *chan, dma_addr_t dma_addr, size_t buf_len,
-		size_t period_len, enum dma_data_direction direction)
+		size_t period_len, enum dma_transfer_direction direction,
+		void *context)
 {
 	struct imxdma_channel *imxdmac = to_imxdma_chan(chan);
 	struct imxdma_engine *imxdma = imxdmac->imxdma;
-	int i, ret;
+	struct imxdma_desc *desc;
+	int i;
 	unsigned int periods = buf_len / period_len;
-	unsigned int dmamode;
 
 	dev_dbg(imxdma->dev, "%s channel: %d buf_len=%d period_len=%d\n",
 			__func__, imxdmac->channel, buf_len, period_len);
 
-	if (imxdmac->status == DMA_IN_PROGRESS)
+	if (list_empty(&imxdmac->ld_free) ||
+	    imxdma_chan_is_doing_cyclic(imxdmac))
 		return NULL;
-	imxdmac->status = DMA_IN_PROGRESS;
 
-	ret = imx_dma_setup_progression_handler(imxdmac->imxdma_channel,
-			imxdma_progression);
-	if (ret) {
-		dev_err(imxdma->dev, "Failed to setup the DMA handler\n");
-		return NULL;
-	}
+	desc = list_first_entry(&imxdmac->ld_free, struct imxdma_desc, node);
 
 	if (imxdmac->sg_list)
 		kfree(imxdmac->sg_list);
@@ -317,62 +838,221 @@ static struct dma_async_tx_descriptor *imxdma_prep_dma_cyclic(
 	imxdmac->sg_list[periods].page_link =
 		((unsigned long)imxdmac->sg_list | 0x01) & ~0x02;
 
-	if (direction == DMA_FROM_DEVICE)
-		dmamode = DMA_MODE_READ;
-	else
-		dmamode = DMA_MODE_WRITE;
+	desc->type = IMXDMA_DESC_CYCLIC;
+	desc->sg = imxdmac->sg_list;
+	desc->sgcount = periods;
+	desc->len = IMX_DMA_LENGTH_LOOP;
+	desc->direction = direction;
+	if (direction == DMA_DEV_TO_MEM) {
+		desc->src = imxdmac->per_address;
+	} else {
+		desc->dest = imxdmac->per_address;
+	}
+	desc->desc.callback = NULL;
+	desc->desc.callback_param = NULL;
+
+	return &desc->desc;
+}
+
+static struct dma_async_tx_descriptor *imxdma_prep_dma_memcpy(
+	struct dma_chan *chan, dma_addr_t dest,
+	dma_addr_t src, size_t len, unsigned long flags)
+{
+	struct imxdma_channel *imxdmac = to_imxdma_chan(chan);
+	struct imxdma_engine *imxdma = imxdmac->imxdma;
+	struct imxdma_desc *desc;
 
-	ret = imx_dma_setup_sg(imxdmac->imxdma_channel, imxdmac->sg_list, periods,
-		 IMX_DMA_LENGTH_LOOP, imxdmac->per_address, dmamode);
-	if (ret)
+	dev_dbg(imxdma->dev, "%s channel: %d src=0x%x dst=0x%x len=%d\n",
+			__func__, imxdmac->channel, src, dest, len);
+
+	if (list_empty(&imxdmac->ld_free) ||
+	    imxdma_chan_is_doing_cyclic(imxdmac))
 		return NULL;
 
-	return &imxdmac->desc;
+	desc = list_first_entry(&imxdmac->ld_free, struct imxdma_desc, node);
+
+	desc->type = IMXDMA_DESC_MEMCPY;
+	desc->src = src;
+	desc->dest = dest;
+	desc->len = len;
+	desc->direction = DMA_MEM_TO_MEM;
+	desc->config_port = IMX_DMA_MEMSIZE_32 | IMX_DMA_TYPE_LINEAR;
+	desc->config_mem = IMX_DMA_MEMSIZE_32 | IMX_DMA_TYPE_LINEAR;
+	desc->desc.callback = NULL;
+	desc->desc.callback_param = NULL;
+
+	return &desc->desc;
+}
+
+static struct dma_async_tx_descriptor *imxdma_prep_dma_interleaved(
+	struct dma_chan *chan, struct dma_interleaved_template *xt,
+	unsigned long flags)
+{
+	struct imxdma_channel *imxdmac = to_imxdma_chan(chan);
+	struct imxdma_engine *imxdma = imxdmac->imxdma;
+	struct imxdma_desc *desc;
+
+	dev_dbg(imxdma->dev, "%s channel: %d src_start=0x%x dst_start=0x%x\n"
+		"   src_sgl=%s dst_sgl=%s numf=%d frame_size=%d\n", __func__,
+		imxdmac->channel, xt->src_start, xt->dst_start,
+		xt->src_sgl ? "true" : "false", xt->dst_sgl ? "true" : "false",
+		xt->numf, xt->frame_size);
+
+	if (list_empty(&imxdmac->ld_free) ||
+	    imxdma_chan_is_doing_cyclic(imxdmac))
+		return NULL;
+
+	if (xt->frame_size != 1 || xt->numf <= 0 || xt->dir != DMA_MEM_TO_MEM)
+		return NULL;
+
+	desc = list_first_entry(&imxdmac->ld_free, struct imxdma_desc, node);
+
+	desc->type = IMXDMA_DESC_INTERLEAVED;
+	desc->src = xt->src_start;
+	desc->dest = xt->dst_start;
+	desc->x = xt->sgl[0].size;
+	desc->y = xt->numf;
+	desc->w = xt->sgl[0].icg + desc->x;
+	desc->len = desc->x * desc->y;
+	desc->direction = DMA_MEM_TO_MEM;
+	desc->config_port = IMX_DMA_MEMSIZE_32;
+	desc->config_mem = IMX_DMA_MEMSIZE_32;
+	if (xt->src_sgl)
+		desc->config_mem |= IMX_DMA_TYPE_2D;
+	if (xt->dst_sgl)
+		desc->config_port |= IMX_DMA_TYPE_2D;
+	desc->desc.callback = NULL;
+	desc->desc.callback_param = NULL;
+
+	return &desc->desc;
 }
 
 static void imxdma_issue_pending(struct dma_chan *chan)
 {
-	/*
-	 * Nothing to do. We only have a single descriptor
-	 */
+	struct imxdma_channel *imxdmac = to_imxdma_chan(chan);
+	struct imxdma_engine *imxdma = imxdmac->imxdma;
+	struct imxdma_desc *desc;
+	unsigned long flags;
+
+	spin_lock_irqsave(&imxdma->lock, flags);
+	if (list_empty(&imxdmac->ld_active) &&
+	    !list_empty(&imxdmac->ld_queue)) {
+		desc = list_first_entry(&imxdmac->ld_queue,
+					struct imxdma_desc, node);
+
+		if (imxdma_xfer_desc(desc) < 0) {
+			dev_warn(imxdma->dev,
+				 "%s: channel: %d couldn't issue DMA xfer\n",
+				 __func__, imxdmac->channel);
+		} else {
+			list_move_tail(imxdmac->ld_queue.next,
+				       &imxdmac->ld_active);
+		}
+	}
+	spin_unlock_irqrestore(&imxdma->lock, flags);
 }
 
 static int __init imxdma_probe(struct platform_device *pdev)
-{
+	{
 	struct imxdma_engine *imxdma;
 	int ret, i;
 
+
 	imxdma = kzalloc(sizeof(*imxdma), GFP_KERNEL);
 	if (!imxdma)
 		return -ENOMEM;
 
+	if (cpu_is_mx1()) {
+		imxdma->base = MX1_IO_ADDRESS(MX1_DMA_BASE_ADDR);
+	} else if (cpu_is_mx21()) {
+		imxdma->base = MX21_IO_ADDRESS(MX21_DMA_BASE_ADDR);
+	} else if (cpu_is_mx27()) {
+		imxdma->base = MX27_IO_ADDRESS(MX27_DMA_BASE_ADDR);
+	} else {
+		kfree(imxdma);
+		return 0;
+	}
+
+	imxdma->dma_clk = clk_get(NULL, "dma");
+	if (IS_ERR(imxdma->dma_clk))
+		return PTR_ERR(imxdma->dma_clk);
+	clk_enable(imxdma->dma_clk);
+
+	/* reset DMA module */
+	imx_dmav1_writel(imxdma, DCR_DRST, DMA_DCR);
+
+	if (cpu_is_mx1()) {
+		ret = request_irq(MX1_DMA_INT, dma_irq_handler, 0, "DMA", imxdma);
+		if (ret) {
+			dev_warn(imxdma->dev, "Can't register IRQ for DMA\n");
+			kfree(imxdma);
+			return ret;
+		}
+
+		ret = request_irq(MX1_DMA_ERR, imxdma_err_handler, 0, "DMA", imxdma);
+		if (ret) {
+			dev_warn(imxdma->dev, "Can't register ERRIRQ for DMA\n");
+			free_irq(MX1_DMA_INT, NULL);
+			kfree(imxdma);
+			return ret;
+		}
+	}
+
+	/* enable DMA module */
+	imx_dmav1_writel(imxdma, DCR_DEN, DMA_DCR);
+
+	/* clear all interrupts */
+	imx_dmav1_writel(imxdma, (1 << IMX_DMA_CHANNELS) - 1, DMA_DISR);
+
+	/* disable interrupts */
+	imx_dmav1_writel(imxdma, (1 << IMX_DMA_CHANNELS) - 1, DMA_DIMR);
+
 	INIT_LIST_HEAD(&imxdma->dma_device.channels);
 
 	dma_cap_set(DMA_SLAVE, imxdma->dma_device.cap_mask);
 	dma_cap_set(DMA_CYCLIC, imxdma->dma_device.cap_mask);
+	dma_cap_set(DMA_MEMCPY, imxdma->dma_device.cap_mask);
+	dma_cap_set(DMA_INTERLEAVE, imxdma->dma_device.cap_mask);
+
+	/* Initialize 2D global parameters */
+	for (i = 0; i < IMX_DMA_2D_SLOTS; i++)
+		imxdma->slots_2d[i].count = 0;
+
+	spin_lock_init(&imxdma->lock);
 
 	/* Initialize channel parameters */
-	for (i = 0; i < MAX_DMA_CHANNELS; i++) {
+	for (i = 0; i < IMX_DMA_CHANNELS; i++) {
 		struct imxdma_channel *imxdmac = &imxdma->channel[i];
 
-		imxdmac->imxdma_channel = imx_dma_request_by_prio("dmaengine",
-				DMA_PRIO_MEDIUM);
-		if ((int)imxdmac->channel < 0) {
-			ret = -ENODEV;
-			goto err_init;
+		if (cpu_is_mx21() || cpu_is_mx27()) {
+			ret = request_irq(MX2x_INT_DMACH0 + i,
+					dma_irq_handler, 0, "DMA", imxdma);
+			if (ret) {
+				dev_warn(imxdma->dev, "Can't register IRQ %d "
+					 "for DMA channel %d\n",
+					 MX2x_INT_DMACH0 + i, i);
+				goto err_init;
+			}
+			init_timer(&imxdmac->watchdog);
+			imxdmac->watchdog.function = &imxdma_watchdog;
+			imxdmac->watchdog.data = (unsigned long)imxdmac;
 		}
 
-		imx_dma_setup_handlers(imxdmac->imxdma_channel,
-		       imxdma_irq_handler, imxdma_err_handler, imxdmac);
-
 		imxdmac->imxdma = imxdma;
-		spin_lock_init(&imxdmac->lock);
 
+		INIT_LIST_HEAD(&imxdmac->ld_queue);
+		INIT_LIST_HEAD(&imxdmac->ld_free);
+		INIT_LIST_HEAD(&imxdmac->ld_active);
+
+		tasklet_init(&imxdmac->dma_tasklet, imxdma_tasklet,
+			     (unsigned long)imxdmac);
 		imxdmac->chan.device = &imxdma->dma_device;
+		dma_cookie_init(&imxdmac->chan);
 		imxdmac->channel = i;
 
 		/* Add the channel to the DMAC list */
-		list_add_tail(&imxdmac->chan.device_node, &imxdma->dma_device.channels);
+		list_add_tail(&imxdmac->chan.device_node,
+			      &imxdma->dma_device.channels);
 	}
 
 	imxdma->dev = &pdev->dev;
@@ -383,11 +1063,14 @@ static int __init imxdma_probe(struct platform_device *pdev)
 	imxdma->dma_device.device_tx_status = imxdma_tx_status;
 	imxdma->dma_device.device_prep_slave_sg = imxdma_prep_slave_sg;
 	imxdma->dma_device.device_prep_dma_cyclic = imxdma_prep_dma_cyclic;
+	imxdma->dma_device.device_prep_dma_memcpy = imxdma_prep_dma_memcpy;
+	imxdma->dma_device.device_prep_interleaved_dma = imxdma_prep_dma_interleaved;
 	imxdma->dma_device.device_control = imxdma_control;
 	imxdma->dma_device.device_issue_pending = imxdma_issue_pending;
 
 	platform_set_drvdata(pdev, imxdma);
 
+	imxdma->dma_device.copy_align = 2; /* 2^2 = 4 bytes alignment */
 	imxdma->dma_device.dev->dma_parms = &imxdma->dma_parms;
 	dma_set_max_seg_size(imxdma->dma_device.dev, 0xffffff);
 
@@ -400,9 +1083,13 @@ static int __init imxdma_probe(struct platform_device *pdev)
 	return 0;
 
 err_init:
-	while (--i >= 0) {
-		struct imxdma_channel *imxdmac = &imxdma->channel[i];
-		imx_dma_free(imxdmac->imxdma_channel);
+
+	if (cpu_is_mx21() || cpu_is_mx27()) {
+		while (--i >= 0)
+			free_irq(MX2x_INT_DMACH0 + i, NULL);
+	} else if cpu_is_mx1() {
+		free_irq(MX1_DMA_INT, NULL);
+		free_irq(MX1_DMA_ERR, NULL);
 	}
 
 	kfree(imxdma);
@@ -416,10 +1103,12 @@ static int __exit imxdma_remove(struct platform_device *pdev)
 
         dma_async_device_unregister(&imxdma->dma_device);
 
-	for (i = 0; i < MAX_DMA_CHANNELS; i++) {
-		struct imxdma_channel *imxdmac = &imxdma->channel[i];
-
-		 imx_dma_free(imxdmac->imxdma_channel);
+	if (cpu_is_mx21() || cpu_is_mx27()) {
+		for (i = 0; i < IMX_DMA_CHANNELS; i++)
+			free_irq(MX2x_INT_DMACH0 + i, NULL);
+	} else if cpu_is_mx1() {
+		free_irq(MX1_DMA_INT, NULL);
+		free_irq(MX1_DMA_ERR, NULL);
 	}
 
         kfree(imxdma);
diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
index f993955a640c..d3e38e28bb6b 100644
--- a/drivers/dma/imx-sdma.c
+++ b/drivers/dma/imx-sdma.c
@@ -20,6 +20,7 @@
 #include <linux/init.h>
 #include <linux/module.h>
 #include <linux/types.h>
+#include <linux/bitops.h>
 #include <linux/mm.h>
 #include <linux/interrupt.h>
 #include <linux/clk.h>
@@ -35,13 +36,14 @@
 #include <linux/dmaengine.h>
 #include <linux/of.h>
 #include <linux/of_device.h>
-#include <linux/module.h>
 
 #include <asm/irq.h>
 #include <mach/sdma.h>
 #include <mach/dma.h>
 #include <mach/hardware.h>
 
+#include "dmaengine.h"
+
 /* SDMA registers */
 #define SDMA_H_C0PTR		0x000
 #define SDMA_H_INTR		0x004
@@ -247,7 +249,7 @@ struct sdma_engine;
 struct sdma_channel {
 	struct sdma_engine		*sdma;
 	unsigned int			channel;
-	enum dma_data_direction		direction;
+	enum dma_transfer_direction		direction;
 	enum sdma_peripheral_type	peripheral_type;
 	unsigned int			event_id0;
 	unsigned int			event_id1;
@@ -260,17 +262,18 @@ struct sdma_channel {
 	unsigned int			pc_from_device, pc_to_device;
 	unsigned long			flags;
 	dma_addr_t			per_address;
-	u32				event_mask0, event_mask1;
-	u32				watermark_level;
+	unsigned long			event_mask[2];
+	unsigned long			watermark_level;
 	u32				shp_addr, per_addr;
 	struct dma_chan			chan;
 	spinlock_t			lock;
 	struct dma_async_tx_descriptor	desc;
-	dma_cookie_t			last_completed;
 	enum dma_status			status;
+	unsigned int			chn_count;
+	unsigned int			chn_real_count;
 };
 
-#define IMX_DMA_SG_LOOP		(1 << 0)
+#define IMX_DMA_SG_LOOP		BIT(0)
 
 #define MAX_DMA_CHANNELS 32
 #define MXC_SDMA_DEFAULT_PRIORITY 1
@@ -344,9 +347,9 @@ static const struct of_device_id sdma_dt_ids[] = {
 };
 MODULE_DEVICE_TABLE(of, sdma_dt_ids);
 
-#define SDMA_H_CONFIG_DSPDMA	(1 << 12) /* indicates if the DSPDMA is used */
-#define SDMA_H_CONFIG_RTD_PINS	(1 << 11) /* indicates if Real-Time Debug pins are enabled */
-#define SDMA_H_CONFIG_ACR	(1 << 4)  /* indicates if AHB freq /core freq = 2 or 1 */
+#define SDMA_H_CONFIG_DSPDMA	BIT(12) /* indicates if the DSPDMA is used */
+#define SDMA_H_CONFIG_RTD_PINS	BIT(11) /* indicates if Real-Time Debug pins are enabled */
+#define SDMA_H_CONFIG_ACR	BIT(4)  /* indicates if AHB freq /core freq = 2 or 1 */
 #define SDMA_H_CONFIG_CSM	(3)       /* indicates which context switch mode is selected*/
 
 static inline u32 chnenbl_ofs(struct sdma_engine *sdma, unsigned int event)
@@ -361,37 +364,42 @@ static int sdma_config_ownership(struct sdma_channel *sdmac,
 {
 	struct sdma_engine *sdma = sdmac->sdma;
 	int channel = sdmac->channel;
-	u32 evt, mcu, dsp;
+	unsigned long evt, mcu, dsp;
 
 	if (event_override && mcu_override && dsp_override)
 		return -EINVAL;
 
-	evt = __raw_readl(sdma->regs + SDMA_H_EVTOVR);
-	mcu = __raw_readl(sdma->regs + SDMA_H_HOSTOVR);
-	dsp = __raw_readl(sdma->regs + SDMA_H_DSPOVR);
+	evt = readl_relaxed(sdma->regs + SDMA_H_EVTOVR);
+	mcu = readl_relaxed(sdma->regs + SDMA_H_HOSTOVR);
+	dsp = readl_relaxed(sdma->regs + SDMA_H_DSPOVR);
 
 	if (dsp_override)
-		dsp &= ~(1 << channel);
+		__clear_bit(channel, &dsp);
 	else
-		dsp |= (1 << channel);
+		__set_bit(channel, &dsp);
 
 	if (event_override)
-		evt &= ~(1 << channel);
+		__clear_bit(channel, &evt);
 	else
-		evt |= (1 << channel);
+		__set_bit(channel, &evt);
 
 	if (mcu_override)
-		mcu &= ~(1 << channel);
+		__clear_bit(channel, &mcu);
 	else
-		mcu |= (1 << channel);
+		__set_bit(channel, &mcu);
 
-	__raw_writel(evt, sdma->regs + SDMA_H_EVTOVR);
-	__raw_writel(mcu, sdma->regs + SDMA_H_HOSTOVR);
-	__raw_writel(dsp, sdma->regs + SDMA_H_DSPOVR);
+	writel_relaxed(evt, sdma->regs + SDMA_H_EVTOVR);
+	writel_relaxed(mcu, sdma->regs + SDMA_H_HOSTOVR);
+	writel_relaxed(dsp, sdma->regs + SDMA_H_DSPOVR);
 
 	return 0;
 }
 
+static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
+{
+	writel(BIT(channel), sdma->regs + SDMA_H_START);
+}
+
 /*
  * sdma_run_channel - run a channel and wait till it's done
  */
@@ -403,7 +411,7 @@ static int sdma_run_channel(struct sdma_channel *sdmac)
 
 	init_completion(&sdmac->done);
 
-	__raw_writel(1 << channel, sdma->regs + SDMA_H_START);
+	sdma_enable_channel(sdma, channel);
 
 	ret = wait_for_completion_timeout(&sdmac->done, HZ);
 
@@ -450,12 +458,12 @@ static void sdma_event_enable(struct sdma_channel *sdmac, unsigned int event)
 {
 	struct sdma_engine *sdma = sdmac->sdma;
 	int channel = sdmac->channel;
-	u32 val;
+	unsigned long val;
 	u32 chnenbl = chnenbl_ofs(sdma, event);
 
-	val = __raw_readl(sdma->regs + chnenbl);
-	val |= (1 << channel);
-	__raw_writel(val, sdma->regs + chnenbl);
+	val = readl_relaxed(sdma->regs + chnenbl);
+	__set_bit(channel, &val);
+	writel_relaxed(val, sdma->regs + chnenbl);
 }
 
 static void sdma_event_disable(struct sdma_channel *sdmac, unsigned int event)
@@ -463,11 +471,11 @@ static void sdma_event_disable(struct sdma_channel *sdmac, unsigned int event)
 	struct sdma_engine *sdma = sdmac->sdma;
 	int channel = sdmac->channel;
 	u32 chnenbl = chnenbl_ofs(sdma, event);
-	u32 val;
+	unsigned long val;
 
-	val = __raw_readl(sdma->regs + chnenbl);
-	val &= ~(1 << channel);
-	__raw_writel(val, sdma->regs + chnenbl);
+	val = readl_relaxed(sdma->regs + chnenbl);
+	__clear_bit(channel, &val);
+	writel_relaxed(val, sdma->regs + chnenbl);
 }
 
 static void sdma_handle_channel_loop(struct sdma_channel *sdmac)
@@ -503,6 +511,7 @@ static void mxc_sdma_handle_channel_normal(struct sdma_channel *sdmac)
 	struct sdma_buffer_descriptor *bd;
 	int i, error = 0;
 
+	sdmac->chn_real_count = 0;
 	/*
 	 * non loop mode. Iterate over all descriptors, collect
 	 * errors and call callback function
@@ -512,6 +521,7 @@ static void mxc_sdma_handle_channel_normal(struct sdma_channel *sdmac)
 
 		 if (bd->mode.status & (BD_DONE | BD_RROR))
 			error = -EIO;
+		 sdmac->chn_real_count += bd->mode.count;
 	}
 
 	if (error)
@@ -519,9 +529,9 @@ static void mxc_sdma_handle_channel_normal(struct sdma_channel *sdmac)
 	else
 		sdmac->status = DMA_SUCCESS;
 
+	dma_cookie_complete(&sdmac->desc);
 	if (sdmac->desc.callback)
 		sdmac->desc.callback(sdmac->desc.callback_param);
-	sdmac->last_completed = sdmac->desc.cookie;
 }
 
 static void mxc_sdma_handle_channel(struct sdma_channel *sdmac)
@@ -541,10 +551,10 @@ static void mxc_sdma_handle_channel(struct sdma_channel *sdmac)
 static irqreturn_t sdma_int_handler(int irq, void *dev_id)
 {
 	struct sdma_engine *sdma = dev_id;
-	u32 stat;
+	unsigned long stat;
 
-	stat = __raw_readl(sdma->regs + SDMA_H_INTR);
-	__raw_writel(stat, sdma->regs + SDMA_H_INTR);
+	stat = readl_relaxed(sdma->regs + SDMA_H_INTR);
+	writel_relaxed(stat, sdma->regs + SDMA_H_INTR);
 
 	while (stat) {
 		int channel = fls(stat) - 1;
@@ -552,7 +562,7 @@ static irqreturn_t sdma_int_handler(int irq, void *dev_id)
 
 		mxc_sdma_handle_channel(sdmac);
 
-		stat &= ~(1 << channel);
+		__clear_bit(channel, &stat);
 	}
 
 	return IRQ_HANDLED;
@@ -650,7 +660,7 @@ static int sdma_load_context(struct sdma_channel *sdmac)
 	struct sdma_buffer_descriptor *bd0 = sdma->channel[0].bd;
 	int ret;
 
-	if (sdmac->direction == DMA_FROM_DEVICE) {
+	if (sdmac->direction == DMA_DEV_TO_MEM) {
 		load_address = sdmac->pc_from_device;
 	} else {
 		load_address = sdmac->pc_to_device;
@@ -660,11 +670,11 @@ static int sdma_load_context(struct sdma_channel *sdmac)
 		return load_address;
 
 	dev_dbg(sdma->dev, "load_address = %d\n", load_address);
-	dev_dbg(sdma->dev, "wml = 0x%08x\n", sdmac->watermark_level);
+	dev_dbg(sdma->dev, "wml = 0x%08x\n", (u32)sdmac->watermark_level);
 	dev_dbg(sdma->dev, "shp_addr = 0x%08x\n", sdmac->shp_addr);
 	dev_dbg(sdma->dev, "per_addr = 0x%08x\n", sdmac->per_addr);
-	dev_dbg(sdma->dev, "event_mask0 = 0x%08x\n", sdmac->event_mask0);
-	dev_dbg(sdma->dev, "event_mask1 = 0x%08x\n", sdmac->event_mask1);
+	dev_dbg(sdma->dev, "event_mask0 = 0x%08x\n", (u32)sdmac->event_mask[0]);
+	dev_dbg(sdma->dev, "event_mask1 = 0x%08x\n", (u32)sdmac->event_mask[1]);
 
 	mutex_lock(&sdma->channel_0_lock);
 
@@ -674,8 +684,8 @@ static int sdma_load_context(struct sdma_channel *sdmac)
 	/* Send by context the event mask,base address for peripheral
 	 * and watermark level
 	 */
-	context->gReg[0] = sdmac->event_mask1;
-	context->gReg[1] = sdmac->event_mask0;
+	context->gReg[0] = sdmac->event_mask[1];
+	context->gReg[1] = sdmac->event_mask[0];
 	context->gReg[2] = sdmac->per_addr;
 	context->gReg[6] = sdmac->shp_addr;
 	context->gReg[7] = sdmac->watermark_level;
@@ -698,7 +708,7 @@ static void sdma_disable_channel(struct sdma_channel *sdmac)
 	struct sdma_engine *sdma = sdmac->sdma;
 	int channel = sdmac->channel;
 
-	__raw_writel(1 << channel, sdma->regs + SDMA_H_STATSTOP);
+	writel_relaxed(BIT(channel), sdma->regs + SDMA_H_STATSTOP);
 	sdmac->status = DMA_ERROR;
 }
 
@@ -708,13 +718,13 @@ static int sdma_config_channel(struct sdma_channel *sdmac)
 
 	sdma_disable_channel(sdmac);
 
-	sdmac->event_mask0 = 0;
-	sdmac->event_mask1 = 0;
+	sdmac->event_mask[0] = 0;
+	sdmac->event_mask[1] = 0;
 	sdmac->shp_addr = 0;
 	sdmac->per_addr = 0;
 
 	if (sdmac->event_id0) {
-		if (sdmac->event_id0 > 32)
+		if (sdmac->event_id0 >= sdmac->sdma->num_events)
 			return -EINVAL;
 		sdma_event_enable(sdmac, sdmac->event_id0);
 	}
@@ -737,15 +747,14 @@ static int sdma_config_channel(struct sdma_channel *sdmac)
 			(sdmac->peripheral_type != IMX_DMATYPE_DSP)) {
 		/* Handle multiple event channels differently */
 		if (sdmac->event_id1) {
-			sdmac->event_mask1 = 1 << (sdmac->event_id1 % 32);
+			sdmac->event_mask[1] = BIT(sdmac->event_id1 % 32);
 			if (sdmac->event_id1 > 31)
-				sdmac->watermark_level |= 1 << 31;
-			sdmac->event_mask0 = 1 << (sdmac->event_id0 % 32);
+				__set_bit(31, &sdmac->watermark_level);
+			sdmac->event_mask[0] = BIT(sdmac->event_id0 % 32);
 			if (sdmac->event_id0 > 31)
-				sdmac->watermark_level |= 1 << 30;
+				__set_bit(30, &sdmac->watermark_level);
 		} else {
-			sdmac->event_mask0 = 1 << sdmac->event_id0;
-			sdmac->event_mask1 = 1 << (sdmac->event_id0 - 32);
+			__set_bit(sdmac->event_id0, sdmac->event_mask);
 		}
 		/* Watermark Level */
 		sdmac->watermark_level |= sdmac->watermark_level;
@@ -771,7 +780,7 @@ static int sdma_set_channel_priority(struct sdma_channel *sdmac,
 		return -EINVAL;
 	}
 
-	__raw_writel(priority, sdma->regs + SDMA_CHNPRI_0 + 4 * channel);
+	writel_relaxed(priority, sdma->regs + SDMA_CHNPRI_0 + 4 * channel);
 
 	return 0;
 }
@@ -793,8 +802,6 @@ static int sdma_request_channel(struct sdma_channel *sdmac)
 	sdma->channel_control[channel].base_bd_ptr = sdmac->bd_phys;
 	sdma->channel_control[channel].current_bd_ptr = sdmac->bd_phys;
 
-	clk_enable(sdma->clk);
-
 	sdma_set_channel_priority(sdmac, MXC_SDMA_DEFAULT_PRIORITY);
 
 	init_completion(&sdmac->done);
@@ -807,24 +814,6 @@ out:
 	return ret;
 }
 
-static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
-{
-	__raw_writel(1 << channel, sdma->regs + SDMA_H_START);
-}
-
-static dma_cookie_t sdma_assign_cookie(struct sdma_channel *sdmac)
-{
-	dma_cookie_t cookie = sdmac->chan.cookie;
-
-	if (++cookie < 0)
-		cookie = 1;
-
-	sdmac->chan.cookie = cookie;
-	sdmac->desc.cookie = cookie;
-
-	return cookie;
-}
-
 static struct sdma_channel *to_sdma_chan(struct dma_chan *chan)
 {
 	return container_of(chan, struct sdma_channel, chan);
@@ -832,17 +821,15 @@ static struct sdma_channel *to_sdma_chan(struct dma_chan *chan)
 
 static dma_cookie_t sdma_tx_submit(struct dma_async_tx_descriptor *tx)
 {
+	unsigned long flags;
 	struct sdma_channel *sdmac = to_sdma_chan(tx->chan);
-	struct sdma_engine *sdma = sdmac->sdma;
 	dma_cookie_t cookie;
 
-	spin_lock_irq(&sdmac->lock);
-
-	cookie = sdma_assign_cookie(sdmac);
+	spin_lock_irqsave(&sdmac->lock, flags);
 
-	sdma_enable_channel(sdma, sdmac->channel);
+	cookie = dma_cookie_assign(tx);
 
-	spin_unlock_irq(&sdmac->lock);
+	spin_unlock_irqrestore(&sdmac->lock, flags);
 
 	return cookie;
 }
@@ -871,11 +858,14 @@ static int sdma_alloc_chan_resources(struct dma_chan *chan)
 
 	sdmac->peripheral_type = data->peripheral_type;
 	sdmac->event_id0 = data->dma_request;
-	ret = sdma_set_channel_priority(sdmac, prio);
+
+	clk_enable(sdmac->sdma->clk);
+
+	ret = sdma_request_channel(sdmac);
 	if (ret)
 		return ret;
 
-	ret = sdma_request_channel(sdmac);
+	ret = sdma_set_channel_priority(sdmac, prio);
 	if (ret)
 		return ret;
 
@@ -911,8 +901,8 @@ static void sdma_free_chan_resources(struct dma_chan *chan)
 
 static struct dma_async_tx_descriptor *sdma_prep_slave_sg(
 		struct dma_chan *chan, struct scatterlist *sgl,
-		unsigned int sg_len, enum dma_data_direction direction,
-		unsigned long flags)
+		unsigned int sg_len, enum dma_transfer_direction direction,
+		unsigned long flags, void *context)
 {
 	struct sdma_channel *sdmac = to_sdma_chan(chan);
 	struct sdma_engine *sdma = sdmac->sdma;
@@ -941,6 +931,7 @@ static struct dma_async_tx_descriptor *sdma_prep_slave_sg(
 		goto err_out;
 	}
 
+	sdmac->chn_count = 0;
 	for_each_sg(sgl, sg, sg_len, i) {
 		struct sdma_buffer_descriptor *bd = &sdmac->bd[i];
 		int param;
@@ -957,6 +948,7 @@ static struct dma_async_tx_descriptor *sdma_prep_slave_sg(
 		}
 
 		bd->mode.count = count;
+		sdmac->chn_count += count;
 
 		if (sdmac->word_size > DMA_SLAVE_BUSWIDTH_4_BYTES) {
 			ret =  -EINVAL;
@@ -1008,7 +1000,8 @@ err_out:
 
 static struct dma_async_tx_descriptor *sdma_prep_dma_cyclic(
 		struct dma_chan *chan, dma_addr_t dma_addr, size_t buf_len,
-		size_t period_len, enum dma_data_direction direction)
+		size_t period_len, enum dma_transfer_direction direction,
+		void *context)
 {
 	struct sdma_channel *sdmac = to_sdma_chan(chan);
 	struct sdma_engine *sdma = sdmac->sdma;
@@ -1093,15 +1086,18 @@ static int sdma_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
 		sdma_disable_channel(sdmac);
 		return 0;
 	case DMA_SLAVE_CONFIG:
-		if (dmaengine_cfg->direction == DMA_FROM_DEVICE) {
+		if (dmaengine_cfg->direction == DMA_DEV_TO_MEM) {
 			sdmac->per_address = dmaengine_cfg->src_addr;
-			sdmac->watermark_level = dmaengine_cfg->src_maxburst;
+			sdmac->watermark_level = dmaengine_cfg->src_maxburst *
+						dmaengine_cfg->src_addr_width;
 			sdmac->word_size = dmaengine_cfg->src_addr_width;
 		} else {
 			sdmac->per_address = dmaengine_cfg->dst_addr;
-			sdmac->watermark_level = dmaengine_cfg->dst_maxburst;
+			sdmac->watermark_level = dmaengine_cfg->dst_maxburst *
+						dmaengine_cfg->dst_addr_width;
 			sdmac->word_size = dmaengine_cfg->dst_addr_width;
 		}
+		sdmac->direction = dmaengine_cfg->direction;
 		return sdma_config_channel(sdmac);
 	default:
 		return -ENOSYS;
@@ -1119,16 +1115,19 @@ static enum dma_status sdma_tx_status(struct dma_chan *chan,
 
 	last_used = chan->cookie;
 
-	dma_set_tx_state(txstate, sdmac->last_completed, last_used, 0);
+	dma_set_tx_state(txstate, chan->completed_cookie, last_used,
+			sdmac->chn_count - sdmac->chn_real_count);
 
 	return sdmac->status;
 }
 
 static void sdma_issue_pending(struct dma_chan *chan)
 {
-	/*
-	 * Nothing to do. We only have a single descriptor
-	 */
+	struct sdma_channel *sdmac = to_sdma_chan(chan);
+	struct sdma_engine *sdma = sdmac->sdma;
+
+	if (sdmac->status == DMA_IN_PROGRESS)
+		sdma_enable_channel(sdma, sdmac->channel);
 }
 
 #define SDMA_SCRIPT_ADDRS_ARRAY_SIZE_V1	34
@@ -1220,7 +1219,7 @@ static int __init sdma_init(struct sdma_engine *sdma)
 	clk_enable(sdma->clk);
 
 	/* Be sure SDMA has not started yet */
-	__raw_writel(0, sdma->regs + SDMA_H_C0PTR);
+	writel_relaxed(0, sdma->regs + SDMA_H_C0PTR);
 
 	sdma->channel_control = dma_alloc_coherent(NULL,
 			MAX_DMA_CHANNELS * sizeof (struct sdma_channel_control) +
@@ -1243,11 +1242,11 @@ static int __init sdma_init(struct sdma_engine *sdma)
 
 	/* disable all channels */
 	for (i = 0; i < sdma->num_events; i++)
-		__raw_writel(0, sdma->regs + chnenbl_ofs(sdma, i));
+		writel_relaxed(0, sdma->regs + chnenbl_ofs(sdma, i));
 
 	/* All channels have priority 0 */
 	for (i = 0; i < MAX_DMA_CHANNELS; i++)
-		__raw_writel(0, sdma->regs + SDMA_CHNPRI_0 + i * 4);
+		writel_relaxed(0, sdma->regs + SDMA_CHNPRI_0 + i * 4);
 
 	ret = sdma_request_channel(&sdma->channel[0]);
 	if (ret)
@@ -1256,16 +1255,16 @@ static int __init sdma_init(struct sdma_engine *sdma)
 	sdma_config_ownership(&sdma->channel[0], false, true, false);
 
 	/* Set Command Channel (Channel Zero) */
-	__raw_writel(0x4050, sdma->regs + SDMA_CHN0ADDR);
+	writel_relaxed(0x4050, sdma->regs + SDMA_CHN0ADDR);
 
 	/* Set bits of CONFIG register but with static context switching */
 	/* FIXME: Check whether to set ACR bit depending on clock ratios */
-	__raw_writel(0, sdma->regs + SDMA_H_CONFIG);
+	writel_relaxed(0, sdma->regs + SDMA_H_CONFIG);
 
-	__raw_writel(ccb_phys, sdma->regs + SDMA_H_C0PTR);
+	writel_relaxed(ccb_phys, sdma->regs + SDMA_H_C0PTR);
 
 	/* Set bits of CONFIG register with given context switching mode */
-	__raw_writel(SDMA_H_CONFIG_CSM, sdma->regs + SDMA_H_CONFIG);
+	writel_relaxed(SDMA_H_CONFIG_CSM, sdma->regs + SDMA_H_CONFIG);
 
 	/* Initializes channel's priorities */
 	sdma_set_channel_priority(&sdma->channel[0], 7);
@@ -1357,6 +1356,7 @@ static int __init sdma_probe(struct platform_device *pdev)
 		spin_lock_init(&sdmac->lock);
 
 		sdmac->chan.device = &sdma->dma_device;
+		dma_cookie_init(&sdmac->chan);
 		sdmac->channel = i;
 
 		/*
@@ -1377,7 +1377,9 @@ static int __init sdma_probe(struct platform_device *pdev)
 		sdma_add_scripts(sdma, pdata->script_addrs);
 
 	if (pdata) {
-		sdma_get_firmware(sdma, pdata->fw_name);
+		ret = sdma_get_firmware(sdma, pdata->fw_name);
+		if (ret)
+			dev_warn(&pdev->dev, "failed to get firmware from platform data\n");
 	} else {
 		/*
 		 * Because that device tree does not encode ROM script address,
@@ -1386,15 +1388,12 @@ static int __init sdma_probe(struct platform_device *pdev)
 		 */
 		ret = of_property_read_string(np, "fsl,sdma-ram-script-name",
 					      &fw_name);
-		if (ret) {
-			dev_err(&pdev->dev, "failed to get firmware name\n");
-			goto err_init;
-		}
-
-		ret = sdma_get_firmware(sdma, fw_name);
-		if (ret) {
-			dev_err(&pdev->dev, "failed to get firmware\n");
-			goto err_init;
+		if (ret)
+			dev_warn(&pdev->dev, "failed to get firmware name\n");
+		else {
+			ret = sdma_get_firmware(sdma, fw_name);
+			if (ret)
+				dev_warn(&pdev->dev, "failed to get firmware from device tree\n");
 		}
 	}
 
diff --git a/drivers/dma/intel_mid_dma.c b/drivers/dma/intel_mid_dma.c
index 19a0c64d45d3..c900ca7aaec4 100644
--- a/drivers/dma/intel_mid_dma.c
+++ b/drivers/dma/intel_mid_dma.c
@@ -29,6 +29,8 @@
 #include <linux/intel_mid_dma.h>
 #include <linux/module.h>
 
+#include "dmaengine.h"
+
 #define MAX_CHAN	4 /*max ch across controllers*/
 #include "intel_mid_dma_regs.h"
 
@@ -280,14 +282,15 @@ static void midc_dostart(struct intel_mid_dma_chan *midc,
  * callbacks but must be called with the lock held.
  */
 static void midc_descriptor_complete(struct intel_mid_dma_chan *midc,
-	       struct intel_mid_dma_desc *desc)
+		struct intel_mid_dma_desc *desc)
+		__releases(&midc->lock) __acquires(&midc->lock)
 {
 	struct dma_async_tx_descriptor	*txd = &desc->txd;
 	dma_async_tx_callback callback_txd = NULL;
 	struct intel_mid_dma_lli	*llitem;
 	void *param_txd = NULL;
 
-	midc->completed = txd->cookie;
+	dma_cookie_complete(txd);
 	callback_txd = txd->callback;
 	param_txd = txd->callback_param;
 
@@ -311,6 +314,7 @@ static void midc_descriptor_complete(struct intel_mid_dma_chan *midc,
 			pci_pool_free(desc->lli_pool, desc->lli,
 						desc->lli_phys);
 			pci_pool_destroy(desc->lli_pool);
+			desc->lli = NULL;
 		}
 		list_move(&desc->desc_node, &midc->free_list);
 		midc->busy = false;
@@ -395,10 +399,10 @@ static int midc_lli_fill_sg(struct intel_mid_dma_chan *midc,
 							midc->dma->block_size);
 		/*Populate SAR and DAR values*/
 		sg_phy_addr = sg_phys(sg);
-		if (desc->dirn ==  DMA_TO_DEVICE) {
+		if (desc->dirn ==  DMA_MEM_TO_DEV) {
 			lli_bloc_desc->sar  = sg_phy_addr;
 			lli_bloc_desc->dar  = mids->dma_slave.dst_addr;
-		} else if (desc->dirn ==  DMA_FROM_DEVICE) {
+		} else if (desc->dirn ==  DMA_DEV_TO_MEM) {
 			lli_bloc_desc->sar  = mids->dma_slave.src_addr;
 			lli_bloc_desc->dar  = sg_phy_addr;
 		}
@@ -432,14 +436,7 @@ static dma_cookie_t intel_mid_dma_tx_submit(struct dma_async_tx_descriptor *tx)
 	dma_cookie_t		cookie;
 
 	spin_lock_bh(&midc->lock);
-	cookie = midc->chan.cookie;
-
-	if (++cookie < 0)
-		cookie = 1;
-
-	midc->chan.cookie = cookie;
-	desc->txd.cookie = cookie;
-
+	cookie = dma_cookie_assign(tx);
 
 	if (list_empty(&midc->active_list))
 		list_add_tail(&desc->desc_node, &midc->active_list);
@@ -480,29 +477,18 @@ static enum dma_status intel_mid_dma_tx_status(struct dma_chan *chan,
 						dma_cookie_t cookie,
 						struct dma_tx_state *txstate)
 {
-	struct intel_mid_dma_chan	*midc = to_intel_mid_dma_chan(chan);
-	dma_cookie_t		last_used;
-	dma_cookie_t		last_complete;
-	int				ret;
+	struct intel_mid_dma_chan *midc = to_intel_mid_dma_chan(chan);
+	enum dma_status ret;
 
-	last_complete = midc->completed;
-	last_used = chan->cookie;
-
-	ret = dma_async_is_complete(cookie, last_complete, last_used);
+	ret = dma_cookie_status(chan, cookie, txstate);
 	if (ret != DMA_SUCCESS) {
+		spin_lock_bh(&midc->lock);
 		midc_scan_descriptors(to_middma_device(chan->device), midc);
+		spin_unlock_bh(&midc->lock);
 
-		last_complete = midc->completed;
-		last_used = chan->cookie;
-
-		ret = dma_async_is_complete(cookie, last_complete, last_used);
+		ret = dma_cookie_status(chan, cookie, txstate);
 	}
 
-	if (txstate) {
-		txstate->last = last_complete;
-		txstate->used = last_used;
-		txstate->residue = 0;
-	}
 	return ret;
 }
 
@@ -566,6 +552,7 @@ static int intel_mid_dma_device_control(struct dma_chan *chan,
 			pci_pool_free(desc->lli_pool, desc->lli,
 						desc->lli_phys);
 			pci_pool_destroy(desc->lli_pool);
+			desc->lli = NULL;
 		}
 		list_move(&desc->desc_node, &midc->free_list);
 	}
@@ -632,13 +619,13 @@ static struct dma_async_tx_descriptor *intel_mid_dma_prep_memcpy(
 		if (midc->dma->pimr_mask) {
 			cfg_hi.cfgx.protctl = 0x0; /*default value*/
 			cfg_hi.cfgx.fifo_mode = 1;
-			if (mids->dma_slave.direction == DMA_TO_DEVICE) {
+			if (mids->dma_slave.direction == DMA_MEM_TO_DEV) {
 				cfg_hi.cfgx.src_per = 0;
 				if (mids->device_instance == 0)
 					cfg_hi.cfgx.dst_per = 3;
 				if (mids->device_instance == 1)
 					cfg_hi.cfgx.dst_per = 1;
-			} else if (mids->dma_slave.direction == DMA_FROM_DEVICE) {
+			} else if (mids->dma_slave.direction == DMA_DEV_TO_MEM) {
 				if (mids->device_instance == 0)
 					cfg_hi.cfgx.src_per = 2;
 				if (mids->device_instance == 1)
@@ -682,11 +669,11 @@ static struct dma_async_tx_descriptor *intel_mid_dma_prep_memcpy(
 		ctl_lo.ctlx.sinc = 0;
 		ctl_lo.ctlx.dinc = 0;
 	} else {
-		if (mids->dma_slave.direction == DMA_TO_DEVICE) {
+		if (mids->dma_slave.direction == DMA_MEM_TO_DEV) {
 			ctl_lo.ctlx.sinc = 0;
 			ctl_lo.ctlx.dinc = 2;
 			ctl_lo.ctlx.tt_fc = 1;
-		} else if (mids->dma_slave.direction == DMA_FROM_DEVICE) {
+		} else if (mids->dma_slave.direction == DMA_DEV_TO_MEM) {
 			ctl_lo.ctlx.sinc = 2;
 			ctl_lo.ctlx.dinc = 0;
 			ctl_lo.ctlx.tt_fc = 2;
@@ -727,13 +714,14 @@ err_desc_get:
  * @sg_len: length of sg txn
  * @direction: DMA transfer dirtn
  * @flags: DMA flags
+ * @context: transfer context (ignored)
  *
  * Prepares LLI based periphral transfer
  */
 static struct dma_async_tx_descriptor *intel_mid_dma_prep_slave_sg(
 			struct dma_chan *chan, struct scatterlist *sgl,
-			unsigned int sg_len, enum dma_data_direction direction,
-			unsigned long flags)
+			unsigned int sg_len, enum dma_transfer_direction direction,
+			unsigned long flags, void *context)
 {
 	struct intel_mid_dma_chan *midc = NULL;
 	struct intel_mid_dma_slave *mids = NULL;
@@ -827,7 +815,6 @@ static void intel_mid_dma_free_chan_resources(struct dma_chan *chan)
 		/*trying to free ch in use!!!!!*/
 		pr_err("ERR_MDMA: trying to free ch in use\n");
 	}
-	pm_runtime_put(&mid->pdev->dev);
 	spin_lock_bh(&midc->lock);
 	midc->descs_allocated = 0;
 	list_for_each_entry_safe(desc, _desc, &midc->active_list, desc_node) {
@@ -848,6 +835,7 @@ static void intel_mid_dma_free_chan_resources(struct dma_chan *chan)
 	/* Disable CH interrupts */
 	iowrite32(MASK_INTR_REG(midc->ch_id), mid->dma_base + MASK_BLOCK);
 	iowrite32(MASK_INTR_REG(midc->ch_id), mid->dma_base + MASK_ERR);
+	pm_runtime_put(&mid->pdev->dev);
 }
 
 /**
@@ -868,7 +856,7 @@ static int intel_mid_dma_alloc_chan_resources(struct dma_chan *chan)
 	pm_runtime_get_sync(&mid->pdev->dev);
 
 	if (mid->state == SUSPENDED) {
-		if (dma_resume(mid->pdev)) {
+		if (dma_resume(&mid->pdev->dev)) {
 			pr_err("ERR_MDMA: resume failed");
 			return -EFAULT;
 		}
@@ -881,7 +869,7 @@ static int intel_mid_dma_alloc_chan_resources(struct dma_chan *chan)
 		pm_runtime_put(&mid->pdev->dev);
 		return -EIO;
 	}
-	midc->completed = chan->cookie = 1;
+	dma_cookie_init(chan);
 
 	spin_lock_bh(&midc->lock);
 	while (midc->descs_allocated < DESCS_PER_CHANNEL) {
@@ -1051,7 +1039,8 @@ static irqreturn_t intel_mid_dma_interrupt(int irq, void *data)
 	}
 	err_status &= mid->intr_mask;
 	if (err_status) {
-		iowrite32(MASK_INTR_REG(err_status), mid->dma_base + MASK_ERR);
+		iowrite32((err_status << INT_MASK_WE),
+			  mid->dma_base + MASK_ERR);
 		call_tasklet = 1;
 	}
 	if (call_tasklet)
@@ -1099,7 +1088,8 @@ static int mid_setup_dma(struct pci_dev *pdev)
 					LNW_PERIPHRAL_MASK_SIZE);
 		if (dma->mask_reg == NULL) {
 			pr_err("ERR_MDMA:Can't map periphral intr space !!\n");
-			return -ENOMEM;
+			err = -ENOMEM;
+			goto err_ioremap;
 		}
 	} else
 		dma->mask_reg = NULL;
@@ -1112,7 +1102,7 @@ static int mid_setup_dma(struct pci_dev *pdev)
 		struct intel_mid_dma_chan *midch = &dma->ch[i];
 
 		midch->chan.device = &dma->common;
-		midch->chan.cookie =  1;
+		dma_cookie_init(&midch->chan);
 		midch->ch_id = dma->chan_base + i;
 		pr_debug("MDMA:Init CH %d, ID %d\n", i, midch->ch_id);
 
@@ -1196,6 +1186,9 @@ static int mid_setup_dma(struct pci_dev *pdev)
 err_engine:
 	free_irq(pdev->irq, dma);
 err_irq:
+	if (dma->mask_reg)
+		iounmap(dma->mask_reg);
+err_ioremap:
 	pci_pool_destroy(dma->dma_pool);
 err_dma_pool:
 	pr_err("ERR_MDMA:setup_dma failed: %d\n", err);
@@ -1337,8 +1330,9 @@ static void __devexit intel_mid_dma_remove(struct pci_dev *pdev)
 *
 * This function is called by OS when a power event occurs
 */
-int dma_suspend(struct pci_dev *pci, pm_message_t state)
+static int dma_suspend(struct device *dev)
 {
+	struct pci_dev *pci = to_pci_dev(dev);
 	int i;
 	struct middma_device *device = pci_get_drvdata(pci);
 	pr_debug("MDMA: dma_suspend called\n");
@@ -1362,8 +1356,9 @@ int dma_suspend(struct pci_dev *pci, pm_message_t state)
 *
 * This function is called by OS when a power event occurs
 */
-int dma_resume(struct pci_dev *pci)
+int dma_resume(struct device *dev)
 {
+	struct pci_dev *pci = to_pci_dev(dev);
 	int ret;
 	struct middma_device *device = pci_get_drvdata(pci);
 
@@ -1429,6 +1424,8 @@ static const struct dev_pm_ops intel_mid_dma_pm = {
 	.runtime_suspend = dma_runtime_suspend,
 	.runtime_resume = dma_runtime_resume,
 	.runtime_idle = dma_runtime_idle,
+	.suspend = dma_suspend,
+	.resume = dma_resume,
 };
 
 static struct pci_driver intel_mid_dma_pci_driver = {
@@ -1437,8 +1434,6 @@ static struct pci_driver intel_mid_dma_pci_driver = {
 	.probe		=	intel_mid_dma_probe,
 	.remove		=	__devexit_p(intel_mid_dma_remove),
 #ifdef CONFIG_PM
-	.suspend = dma_suspend,
-	.resume = dma_resume,
 	.driver = {
 		.pm = &intel_mid_dma_pm,
 	},
diff --git a/drivers/dma/intel_mid_dma_regs.h b/drivers/dma/intel_mid_dma_regs.h
index aea5ee88ce03..1bfa9268feaf 100644
--- a/drivers/dma/intel_mid_dma_regs.h
+++ b/drivers/dma/intel_mid_dma_regs.h
@@ -165,7 +165,6 @@ union intel_mid_dma_cfg_hi {
  * @dma_base: MMIO register space DMA engine base pointer
  * @ch_id: DMA channel id
  * @lock: channel spinlock
- * @completed: DMA cookie
  * @active_list: current active descriptors
  * @queue: current queued up descriptors
  * @free_list: current free descriptors
@@ -183,7 +182,6 @@ struct intel_mid_dma_chan {
 	void __iomem		*dma_base;
 	int			ch_id;
 	spinlock_t		lock;
-	dma_cookie_t		completed;
 	struct list_head	active_list;
 	struct list_head	queue;
 	struct list_head	free_list;
@@ -262,7 +260,7 @@ struct intel_mid_dma_desc {
 	unsigned int			lli_length;
 	unsigned int			current_lli;
 	dma_addr_t			next;
-	enum dma_data_direction		dirn;
+	enum dma_transfer_direction		dirn;
 	enum dma_status			status;
 	enum dma_slave_buswidth		width; /*width of DMA txn*/
 	enum intel_mid_dma_mode		cfg_mode; /*mode configuration*/
@@ -296,6 +294,6 @@ static inline struct intel_mid_dma_slave *to_intel_mid_dma_slave
 }
 
 
-int dma_resume(struct pci_dev *pci);
+int dma_resume(struct device *dev);
 
 #endif /*__INTEL_MID_DMAC_REGS_H__*/
diff --git a/drivers/dma/ioat/dma.c b/drivers/dma/ioat/dma.c
index a4d6cb0c0343..73b2b65cb1de 100644
--- a/drivers/dma/ioat/dma.c
+++ b/drivers/dma/ioat/dma.c
@@ -40,6 +40,8 @@
 #include "registers.h"
 #include "hw.h"
 
+#include "../dmaengine.h"
+
 int ioat_pending_level = 4;
 module_param(ioat_pending_level, int, 0644);
 MODULE_PARM_DESC(ioat_pending_level,
@@ -107,6 +109,7 @@ void ioat_init_channel(struct ioatdma_device *device, struct ioat_chan_common *c
 	chan->reg_base = device->reg_base + (0x80 * (idx + 1));
 	spin_lock_init(&chan->cleanup_lock);
 	chan->common.device = dma;
+	dma_cookie_init(&chan->common);
 	list_add_tail(&chan->common.device_node, &dma->channels);
 	device->idx[idx] = chan;
 	init_timer(&chan->timer);
@@ -235,12 +238,7 @@ static dma_cookie_t ioat1_tx_submit(struct dma_async_tx_descriptor *tx)
 
 	spin_lock_bh(&ioat->desc_lock);
 	/* cookie incr and addition to used_list must be atomic */
-	cookie = c->cookie;
-	cookie++;
-	if (cookie < 0)
-		cookie = 1;
-	c->cookie = cookie;
-	tx->cookie = cookie;
+	cookie = dma_cookie_assign(tx);
 	dev_dbg(to_dev(&ioat->base), "%s: cookie: %d\n", __func__, cookie);
 
 	/* write address into NextDescriptor field of last desc in chain */
@@ -548,9 +546,9 @@ void ioat_dma_unmap(struct ioat_chan_common *chan, enum dma_ctrl_flags flags,
 			   PCI_DMA_TODEVICE, flags, 0);
 }
 
-unsigned long ioat_get_current_completion(struct ioat_chan_common *chan)
+dma_addr_t ioat_get_current_completion(struct ioat_chan_common *chan)
 {
-	unsigned long phys_complete;
+	dma_addr_t phys_complete;
 	u64 completion;
 
 	completion = *chan->completion;
@@ -571,7 +569,7 @@ unsigned long ioat_get_current_completion(struct ioat_chan_common *chan)
 }
 
 bool ioat_cleanup_preamble(struct ioat_chan_common *chan,
-			   unsigned long *phys_complete)
+			   dma_addr_t *phys_complete)
 {
 	*phys_complete = ioat_get_current_completion(chan);
 	if (*phys_complete == chan->last_completion)
@@ -582,14 +580,14 @@ bool ioat_cleanup_preamble(struct ioat_chan_common *chan,
 	return true;
 }
 
-static void __cleanup(struct ioat_dma_chan *ioat, unsigned long phys_complete)
+static void __cleanup(struct ioat_dma_chan *ioat, dma_addr_t phys_complete)
 {
 	struct ioat_chan_common *chan = &ioat->base;
 	struct list_head *_desc, *n;
 	struct dma_async_tx_descriptor *tx;
 
-	dev_dbg(to_dev(chan), "%s: phys_complete: %lx\n",
-		 __func__, phys_complete);
+	dev_dbg(to_dev(chan), "%s: phys_complete: %llx\n",
+		 __func__, (unsigned long long) phys_complete);
 	list_for_each_safe(_desc, n, &ioat->used_desc) {
 		struct ioat_desc_sw *desc;
 
@@ -603,8 +601,7 @@ static void __cleanup(struct ioat_dma_chan *ioat, unsigned long phys_complete)
 		 */
 		dump_desc_dbg(ioat, desc);
 		if (tx->cookie) {
-			chan->completed_cookie = tx->cookie;
-			tx->cookie = 0;
+			dma_cookie_complete(tx);
 			ioat_dma_unmap(chan, tx->flags, desc->len, desc->hw);
 			ioat->active -= desc->hw->tx_cnt;
 			if (tx->callback) {
@@ -655,7 +652,7 @@ static void __cleanup(struct ioat_dma_chan *ioat, unsigned long phys_complete)
 static void ioat1_cleanup(struct ioat_dma_chan *ioat)
 {
 	struct ioat_chan_common *chan = &ioat->base;
-	unsigned long phys_complete;
+	dma_addr_t phys_complete;
 
 	prefetch(chan->completion);
 
@@ -701,7 +698,7 @@ static void ioat1_timer_event(unsigned long data)
 		mod_timer(&chan->timer, jiffies + COMPLETION_TIMEOUT);
 		spin_unlock_bh(&ioat->desc_lock);
 	} else if (test_bit(IOAT_COMPLETION_PENDING, &chan->state)) {
-		unsigned long phys_complete;
+		dma_addr_t phys_complete;
 
 		spin_lock_bh(&ioat->desc_lock);
 		/* if we haven't made progress and we have already
@@ -733,13 +730,15 @@ ioat_dma_tx_status(struct dma_chan *c, dma_cookie_t cookie,
 {
 	struct ioat_chan_common *chan = to_chan_common(c);
 	struct ioatdma_device *device = chan->device;
+	enum dma_status ret;
 
-	if (ioat_tx_status(c, cookie, txstate) == DMA_SUCCESS)
-		return DMA_SUCCESS;
+	ret = dma_cookie_status(c, cookie, txstate);
+	if (ret == DMA_SUCCESS)
+		return ret;
 
 	device->cleanup_fn((unsigned long) c);
 
-	return ioat_tx_status(c, cookie, txstate);
+	return dma_cookie_status(c, cookie, txstate);
 }
 
 static void ioat1_dma_start_null_desc(struct ioat_dma_chan *ioat)
diff --git a/drivers/dma/ioat/dma.h b/drivers/dma/ioat/dma.h
index 5216c8a92a21..5e8fe01ba69d 100644
--- a/drivers/dma/ioat/dma.h
+++ b/drivers/dma/ioat/dma.h
@@ -88,9 +88,8 @@ struct ioatdma_device {
 struct ioat_chan_common {
 	struct dma_chan common;
 	void __iomem *reg_base;
-	unsigned long last_completion;
+	dma_addr_t last_completion;
 	spinlock_t cleanup_lock;
-	dma_cookie_t completed_cookie;
 	unsigned long state;
 	#define IOAT_COMPLETION_PENDING 0
 	#define IOAT_COMPLETION_ACK 1
@@ -143,28 +142,6 @@ static inline struct ioat_dma_chan *to_ioat_chan(struct dma_chan *c)
 	return container_of(chan, struct ioat_dma_chan, base);
 }
 
-/**
- * ioat_tx_status - poll the status of an ioat transaction
- * @c: channel handle
- * @cookie: transaction identifier
- * @txstate: if set, updated with the transaction state
- */
-static inline enum dma_status
-ioat_tx_status(struct dma_chan *c, dma_cookie_t cookie,
-		 struct dma_tx_state *txstate)
-{
-	struct ioat_chan_common *chan = to_chan_common(c);
-	dma_cookie_t last_used;
-	dma_cookie_t last_complete;
-
-	last_used = c->cookie;
-	last_complete = chan->completed_cookie;
-
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
-
-	return dma_async_is_complete(cookie, last_complete, last_used);
-}
-
 /* wrapper around hardware descriptor format + additional software fields */
 
 /**
@@ -333,7 +310,7 @@ int __devinit ioat_dma_self_test(struct ioatdma_device *device);
 void __devexit ioat_dma_remove(struct ioatdma_device *device);
 struct dca_provider * __devinit ioat_dca_init(struct pci_dev *pdev,
 					      void __iomem *iobase);
-unsigned long ioat_get_current_completion(struct ioat_chan_common *chan);
+dma_addr_t ioat_get_current_completion(struct ioat_chan_common *chan);
 void ioat_init_channel(struct ioatdma_device *device,
 		       struct ioat_chan_common *chan, int idx);
 enum dma_status ioat_dma_tx_status(struct dma_chan *c, dma_cookie_t cookie,
@@ -341,7 +318,7 @@ enum dma_status ioat_dma_tx_status(struct dma_chan *c, dma_cookie_t cookie,
 void ioat_dma_unmap(struct ioat_chan_common *chan, enum dma_ctrl_flags flags,
 		    size_t len, struct ioat_dma_descriptor *hw);
 bool ioat_cleanup_preamble(struct ioat_chan_common *chan,
-			   unsigned long *phys_complete);
+			   dma_addr_t *phys_complete);
 void ioat_kobject_add(struct ioatdma_device *device, struct kobj_type *type);
 void ioat_kobject_del(struct ioatdma_device *device);
 extern const struct sysfs_ops ioat_sysfs_ops;
diff --git a/drivers/dma/ioat/dma_v2.c b/drivers/dma/ioat/dma_v2.c
index 5d65f8377971..86895760b598 100644
--- a/drivers/dma/ioat/dma_v2.c
+++ b/drivers/dma/ioat/dma_v2.c
@@ -41,6 +41,8 @@
 #include "registers.h"
 #include "hw.h"
 
+#include "../dmaengine.h"
+
 int ioat_ring_alloc_order = 8;
 module_param(ioat_ring_alloc_order, int, 0644);
 MODULE_PARM_DESC(ioat_ring_alloc_order,
@@ -126,7 +128,7 @@ static void ioat2_start_null_desc(struct ioat2_dma_chan *ioat)
 	spin_unlock_bh(&ioat->prep_lock);
 }
 
-static void __cleanup(struct ioat2_dma_chan *ioat, unsigned long phys_complete)
+static void __cleanup(struct ioat2_dma_chan *ioat, dma_addr_t phys_complete)
 {
 	struct ioat_chan_common *chan = &ioat->base;
 	struct dma_async_tx_descriptor *tx;
@@ -147,8 +149,7 @@ static void __cleanup(struct ioat2_dma_chan *ioat, unsigned long phys_complete)
 		dump_desc_dbg(ioat, desc);
 		if (tx->cookie) {
 			ioat_dma_unmap(chan, tx->flags, desc->len, desc->hw);
-			chan->completed_cookie = tx->cookie;
-			tx->cookie = 0;
+			dma_cookie_complete(tx);
 			if (tx->callback) {
 				tx->callback(tx->callback_param);
 				tx->callback = NULL;
@@ -178,7 +179,7 @@ static void __cleanup(struct ioat2_dma_chan *ioat, unsigned long phys_complete)
 static void ioat2_cleanup(struct ioat2_dma_chan *ioat)
 {
 	struct ioat_chan_common *chan = &ioat->base;
-	unsigned long phys_complete;
+	dma_addr_t phys_complete;
 
 	spin_lock_bh(&chan->cleanup_lock);
 	if (ioat_cleanup_preamble(chan, &phys_complete))
@@ -259,7 +260,7 @@ int ioat2_reset_sync(struct ioat_chan_common *chan, unsigned long tmo)
 static void ioat2_restart_channel(struct ioat2_dma_chan *ioat)
 {
 	struct ioat_chan_common *chan = &ioat->base;
-	unsigned long phys_complete;
+	dma_addr_t phys_complete;
 
 	ioat2_quiesce(chan, 0);
 	if (ioat_cleanup_preamble(chan, &phys_complete))
@@ -274,7 +275,7 @@ void ioat2_timer_event(unsigned long data)
 	struct ioat_chan_common *chan = &ioat->base;
 
 	if (test_bit(IOAT_COMPLETION_PENDING, &chan->state)) {
-		unsigned long phys_complete;
+		dma_addr_t phys_complete;
 		u64 status;
 
 		status = ioat_chansts(chan);
@@ -398,13 +399,9 @@ static dma_cookie_t ioat2_tx_submit_unlock(struct dma_async_tx_descriptor *tx)
 	struct dma_chan *c = tx->chan;
 	struct ioat2_dma_chan *ioat = to_ioat2_chan(c);
 	struct ioat_chan_common *chan = &ioat->base;
-	dma_cookie_t cookie = c->cookie;
+	dma_cookie_t cookie;
 
-	cookie++;
-	if (cookie < 0)
-		cookie = 1;
-	tx->cookie = cookie;
-	c->cookie = cookie;
+	cookie = dma_cookie_assign(tx);
 	dev_dbg(to_dev(&ioat->base), "%s: cookie: %d\n", __func__, cookie);
 
 	if (!test_and_set_bit(IOAT_COMPLETION_PENDING, &chan->state))
@@ -575,9 +572,9 @@ bool reshape_ring(struct ioat2_dma_chan *ioat, int order)
 	 */
 	struct ioat_chan_common *chan = &ioat->base;
 	struct dma_chan *c = &chan->common;
-	const u16 curr_size = ioat2_ring_size(ioat);
+	const u32 curr_size = ioat2_ring_size(ioat);
 	const u16 active = ioat2_ring_active(ioat);
-	const u16 new_size = 1 << order;
+	const u32 new_size = 1 << order;
 	struct ioat_ring_ent **ring;
 	u16 i;
 
diff --git a/drivers/dma/ioat/dma_v2.h b/drivers/dma/ioat/dma_v2.h
index a2c413b2b8d8..be2a55b95c23 100644
--- a/drivers/dma/ioat/dma_v2.h
+++ b/drivers/dma/ioat/dma_v2.h
@@ -74,7 +74,7 @@ static inline struct ioat2_dma_chan *to_ioat2_chan(struct dma_chan *c)
 	return container_of(chan, struct ioat2_dma_chan, base);
 }
 
-static inline u16 ioat2_ring_size(struct ioat2_dma_chan *ioat)
+static inline u32 ioat2_ring_size(struct ioat2_dma_chan *ioat)
 {
 	return 1 << ioat->alloc_order;
 }
@@ -91,7 +91,7 @@ static inline u16 ioat2_ring_pending(struct ioat2_dma_chan *ioat)
 	return CIRC_CNT(ioat->head, ioat->issued, ioat2_ring_size(ioat));
 }
 
-static inline u16 ioat2_ring_space(struct ioat2_dma_chan *ioat)
+static inline u32 ioat2_ring_space(struct ioat2_dma_chan *ioat)
 {
 	return ioat2_ring_size(ioat) - ioat2_ring_active(ioat);
 }
diff --git a/drivers/dma/ioat/dma_v3.c b/drivers/dma/ioat/dma_v3.c
index f519c93a61e7..f7f1dc62c15c 100644
--- a/drivers/dma/ioat/dma_v3.c
+++ b/drivers/dma/ioat/dma_v3.c
@@ -61,6 +61,7 @@
 #include <linux/dmaengine.h>
 #include <linux/dma-mapping.h>
 #include <linux/prefetch.h>
+#include "../dmaengine.h"
 #include "registers.h"
 #include "hw.h"
 #include "dma.h"
@@ -256,7 +257,7 @@ static bool desc_has_ext(struct ioat_ring_ent *desc)
  * The difference from the dma_v2.c __cleanup() is that this routine
  * handles extended descriptors and dma-unmapping raid operations.
  */
-static void __cleanup(struct ioat2_dma_chan *ioat, unsigned long phys_complete)
+static void __cleanup(struct ioat2_dma_chan *ioat, dma_addr_t phys_complete)
 {
 	struct ioat_chan_common *chan = &ioat->base;
 	struct ioat_ring_ent *desc;
@@ -277,9 +278,8 @@ static void __cleanup(struct ioat2_dma_chan *ioat, unsigned long phys_complete)
 		dump_desc_dbg(ioat, desc);
 		tx = &desc->txd;
 		if (tx->cookie) {
-			chan->completed_cookie = tx->cookie;
+			dma_cookie_complete(tx);
 			ioat3_dma_unmap(ioat, desc, idx + i);
-			tx->cookie = 0;
 			if (tx->callback) {
 				tx->callback(tx->callback_param);
 				tx->callback = NULL;
@@ -314,7 +314,7 @@ static void __cleanup(struct ioat2_dma_chan *ioat, unsigned long phys_complete)
 static void ioat3_cleanup(struct ioat2_dma_chan *ioat)
 {
 	struct ioat_chan_common *chan = &ioat->base;
-	unsigned long phys_complete;
+	dma_addr_t phys_complete;
 
 	spin_lock_bh(&chan->cleanup_lock);
 	if (ioat_cleanup_preamble(chan, &phys_complete))
@@ -333,7 +333,7 @@ static void ioat3_cleanup_event(unsigned long data)
 static void ioat3_restart_channel(struct ioat2_dma_chan *ioat)
 {
 	struct ioat_chan_common *chan = &ioat->base;
-	unsigned long phys_complete;
+	dma_addr_t phys_complete;
 
 	ioat2_quiesce(chan, 0);
 	if (ioat_cleanup_preamble(chan, &phys_complete))
@@ -348,7 +348,7 @@ static void ioat3_timer_event(unsigned long data)
 	struct ioat_chan_common *chan = &ioat->base;
 
 	if (test_bit(IOAT_COMPLETION_PENDING, &chan->state)) {
-		unsigned long phys_complete;
+		dma_addr_t phys_complete;
 		u64 status;
 
 		status = ioat_chansts(chan);
@@ -411,13 +411,15 @@ ioat3_tx_status(struct dma_chan *c, dma_cookie_t cookie,
 		struct dma_tx_state *txstate)
 {
 	struct ioat2_dma_chan *ioat = to_ioat2_chan(c);
+	enum dma_status ret;
 
-	if (ioat_tx_status(c, cookie, txstate) == DMA_SUCCESS)
-		return DMA_SUCCESS;
+	ret = dma_cookie_status(c, cookie, txstate);
+	if (ret == DMA_SUCCESS)
+		return ret;
 
 	ioat3_cleanup(ioat);
 
-	return ioat_tx_status(c, cookie, txstate);
+	return dma_cookie_status(c, cookie, txstate);
 }
 
 static struct dma_async_tx_descriptor *
@@ -1147,6 +1149,44 @@ static int ioat3_reset_hw(struct ioat_chan_common *chan)
 	return ioat2_reset_sync(chan, msecs_to_jiffies(200));
 }
 
+static bool is_jf_ioat(struct pci_dev *pdev)
+{
+	switch (pdev->device) {
+	case PCI_DEVICE_ID_INTEL_IOAT_JSF0:
+	case PCI_DEVICE_ID_INTEL_IOAT_JSF1:
+	case PCI_DEVICE_ID_INTEL_IOAT_JSF2:
+	case PCI_DEVICE_ID_INTEL_IOAT_JSF3:
+	case PCI_DEVICE_ID_INTEL_IOAT_JSF4:
+	case PCI_DEVICE_ID_INTEL_IOAT_JSF5:
+	case PCI_DEVICE_ID_INTEL_IOAT_JSF6:
+	case PCI_DEVICE_ID_INTEL_IOAT_JSF7:
+	case PCI_DEVICE_ID_INTEL_IOAT_JSF8:
+	case PCI_DEVICE_ID_INTEL_IOAT_JSF9:
+		return true;
+	default:
+		return false;
+	}
+}
+
+static bool is_snb_ioat(struct pci_dev *pdev)
+{
+	switch (pdev->device) {
+	case PCI_DEVICE_ID_INTEL_IOAT_SNB0:
+	case PCI_DEVICE_ID_INTEL_IOAT_SNB1:
+	case PCI_DEVICE_ID_INTEL_IOAT_SNB2:
+	case PCI_DEVICE_ID_INTEL_IOAT_SNB3:
+	case PCI_DEVICE_ID_INTEL_IOAT_SNB4:
+	case PCI_DEVICE_ID_INTEL_IOAT_SNB5:
+	case PCI_DEVICE_ID_INTEL_IOAT_SNB6:
+	case PCI_DEVICE_ID_INTEL_IOAT_SNB7:
+	case PCI_DEVICE_ID_INTEL_IOAT_SNB8:
+	case PCI_DEVICE_ID_INTEL_IOAT_SNB9:
+		return true;
+	default:
+		return false;
+	}
+}
+
 int __devinit ioat3_dma_probe(struct ioatdma_device *device, int dca)
 {
 	struct pci_dev *pdev = device->pdev;
@@ -1167,6 +1207,9 @@ int __devinit ioat3_dma_probe(struct ioatdma_device *device, int dca)
 	dma->device_alloc_chan_resources = ioat2_alloc_chan_resources;
 	dma->device_free_chan_resources = ioat2_free_chan_resources;
 
+	if (is_jf_ioat(pdev) || is_snb_ioat(pdev))
+		dma->copy_align = 6;
+
 	dma_cap_set(DMA_INTERRUPT, dma->cap_mask);
 	dma->device_prep_dma_interrupt = ioat3_prep_interrupt_lock;
 
diff --git a/drivers/dma/iop-adma.c b/drivers/dma/iop-adma.c
index e03f811a83dd..79e3eba29702 100644
--- a/drivers/dma/iop-adma.c
+++ b/drivers/dma/iop-adma.c
@@ -36,6 +36,8 @@
 
 #include <mach/adma.h>
 
+#include "dmaengine.h"
+
 #define to_iop_adma_chan(chan) container_of(chan, struct iop_adma_chan, common)
 #define to_iop_adma_device(dev) \
 	container_of(dev, struct iop_adma_device, common)
@@ -317,7 +319,7 @@ static void __iop_adma_slot_cleanup(struct iop_adma_chan *iop_chan)
 	}
 
 	if (cookie > 0) {
-		iop_chan->completed_cookie = cookie;
+		iop_chan->common.completed_cookie = cookie;
 		pr_debug("\tcompleted cookie %d\n", cookie);
 	}
 }
@@ -438,18 +440,6 @@ retry:
 	return NULL;
 }
 
-static dma_cookie_t
-iop_desc_assign_cookie(struct iop_adma_chan *iop_chan,
-	struct iop_adma_desc_slot *desc)
-{
-	dma_cookie_t cookie = iop_chan->common.cookie;
-	cookie++;
-	if (cookie < 0)
-		cookie = 1;
-	iop_chan->common.cookie = desc->async_tx.cookie = cookie;
-	return cookie;
-}
-
 static void iop_adma_check_threshold(struct iop_adma_chan *iop_chan)
 {
 	dev_dbg(iop_chan->device->common.dev, "pending: %d\n",
@@ -477,7 +467,7 @@ iop_adma_tx_submit(struct dma_async_tx_descriptor *tx)
 	slots_per_op = grp_start->slots_per_op;
 
 	spin_lock_bh(&iop_chan->lock);
-	cookie = iop_desc_assign_cookie(iop_chan, sw_desc);
+	cookie = dma_cookie_assign(tx);
 
 	old_chain_tail = list_entry(iop_chan->chain.prev,
 		struct iop_adma_desc_slot, chain_node);
@@ -904,24 +894,15 @@ static enum dma_status iop_adma_status(struct dma_chan *chan,
 					struct dma_tx_state *txstate)
 {
 	struct iop_adma_chan *iop_chan = to_iop_adma_chan(chan);
-	dma_cookie_t last_used;
-	dma_cookie_t last_complete;
-	enum dma_status ret;
-
-	last_used = chan->cookie;
-	last_complete = iop_chan->completed_cookie;
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
-	ret = dma_async_is_complete(cookie, last_complete, last_used);
+	int ret;
+
+	ret = dma_cookie_status(chan, cookie, txstate);
 	if (ret == DMA_SUCCESS)
 		return ret;
 
 	iop_adma_slot_cleanup(iop_chan);
 
-	last_used = chan->cookie;
-	last_complete = iop_chan->completed_cookie;
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
-
-	return dma_async_is_complete(cookie, last_complete, last_used);
+	return dma_cookie_status(chan, cookie, txstate);
 }
 
 static irqreturn_t iop_adma_eot_handler(int irq, void *data)
@@ -1271,8 +1252,8 @@ iop_adma_pq_zero_sum_self_test(struct iop_adma_device *device)
 	struct page **pq_hw = &pq[IOP_ADMA_NUM_SRC_TEST+2];
 	/* address conversion buffers (dma_map / page_address) */
 	void *pq_sw[IOP_ADMA_NUM_SRC_TEST+2];
-	dma_addr_t pq_src[IOP_ADMA_NUM_SRC_TEST];
-	dma_addr_t pq_dest[2];
+	dma_addr_t pq_src[IOP_ADMA_NUM_SRC_TEST+2];
+	dma_addr_t *pq_dest = &pq_src[IOP_ADMA_NUM_SRC_TEST];
 
 	int i;
 	struct dma_async_tx_descriptor *tx;
@@ -1482,7 +1463,7 @@ static int __devinit iop_adma_probe(struct platform_device *pdev)
 		goto err_free_adev;
 	}
 
-	dev_dbg(&pdev->dev, "%s: allocted descriptor pool virt %p phys %p\n",
+	dev_dbg(&pdev->dev, "%s: allocated descriptor pool virt %p phys %p\n",
 		__func__, adev->dma_desc_pool_virt,
 		(void *) adev->dma_desc_pool);
 
@@ -1565,6 +1546,7 @@ static int __devinit iop_adma_probe(struct platform_device *pdev)
 	INIT_LIST_HEAD(&iop_chan->chain);
 	INIT_LIST_HEAD(&iop_chan->all_slots);
 	iop_chan->common.device = dma_dev;
+	dma_cookie_init(&iop_chan->common);
 	list_add_tail(&iop_chan->common.device_node, &dma_dev->channels);
 
 	if (dma_has_cap(DMA_MEMCPY, dma_dev->cap_mask)) {
@@ -1642,16 +1624,12 @@ static void iop_chan_start_null_memcpy(struct iop_adma_chan *iop_chan)
 		iop_desc_set_dest_addr(grp_start, iop_chan, 0);
 		iop_desc_set_memcpy_src_addr(grp_start, 0);
 
-		cookie = iop_chan->common.cookie;
-		cookie++;
-		if (cookie <= 1)
-			cookie = 2;
+		cookie = dma_cookie_assign(&sw_desc->async_tx);
 
 		/* initialize the completed cookie to be less than
 		 * the most recently used cookie
 		 */
-		iop_chan->completed_cookie = cookie - 1;
-		iop_chan->common.cookie = sw_desc->async_tx.cookie = cookie;
+		iop_chan->common.completed_cookie = cookie - 1;
 
 		/* channel should not be busy */
 		BUG_ON(iop_chan_is_busy(iop_chan));
@@ -1699,16 +1677,12 @@ static void iop_chan_start_null_xor(struct iop_adma_chan *iop_chan)
 		iop_desc_set_xor_src_addr(grp_start, 0, 0);
 		iop_desc_set_xor_src_addr(grp_start, 1, 0);
 
-		cookie = iop_chan->common.cookie;
-		cookie++;
-		if (cookie <= 1)
-			cookie = 2;
+		cookie = dma_cookie_assign(&sw_desc->async_tx);
 
 		/* initialize the completed cookie to be less than
 		 * the most recently used cookie
 		 */
-		iop_chan->completed_cookie = cookie - 1;
-		iop_chan->common.cookie = sw_desc->async_tx.cookie = cookie;
+		iop_chan->common.completed_cookie = cookie - 1;
 
 		/* channel should not be busy */
 		BUG_ON(iop_chan_is_busy(iop_chan));
@@ -1735,8 +1709,6 @@ static void iop_chan_start_null_xor(struct iop_adma_chan *iop_chan)
 	spin_unlock_bh(&iop_chan->lock);
 }
 
-MODULE_ALIAS("platform:iop-adma");
-
 static struct platform_driver iop_adma_driver = {
 	.probe		= iop_adma_probe,
 	.remove		= __devexit_p(iop_adma_remove),
@@ -1746,19 +1718,9 @@ static struct platform_driver iop_adma_driver = {
 	},
 };
 
-static int __init iop_adma_init (void)
-{
-	return platform_driver_register(&iop_adma_driver);
-}
-
-static void __exit iop_adma_exit (void)
-{
-	platform_driver_unregister(&iop_adma_driver);
-	return;
-}
-module_exit(iop_adma_exit);
-module_init(iop_adma_init);
+module_platform_driver(iop_adma_driver);
 
 MODULE_AUTHOR("Intel Corporation");
 MODULE_DESCRIPTION("IOP ADMA Engine Driver");
 MODULE_LICENSE("GPL");
+MODULE_ALIAS("platform:iop-adma");
diff --git a/drivers/dma/ipu/ipu_idmac.c b/drivers/dma/ipu/ipu_idmac.c
index 0e5ef33f90a1..62e3f8ec2461 100644
--- a/drivers/dma/ipu/ipu_idmac.c
+++ b/drivers/dma/ipu/ipu_idmac.c
@@ -25,6 +25,7 @@
 
 #include <mach/ipu.h>
 
+#include "../dmaengine.h"
 #include "ipu_intern.h"
 
 #define FS_VF_IN_VALID	0x00000002
@@ -312,7 +313,7 @@ static void ipu_ch_param_set_size(union chan_param_mem *params,
 	case IPU_PIX_FMT_RGB565:
 		params->ip.bpp	= 2;
 		params->ip.pfs	= 4;
-		params->ip.npb	= 7;
+		params->ip.npb	= 15;
 		params->ip.sat	= 2;		/* SAT = 32-bit access */
 		params->ip.ofs0	= 0;		/* Red bit offset */
 		params->ip.ofs1	= 5;		/* Green bit offset */
@@ -422,12 +423,6 @@ static void ipu_ch_param_set_size(union chan_param_mem *params,
 	params->pp.nsb = 1;
 }
 
-static void ipu_ch_param_set_burst_size(union chan_param_mem *params,
-					uint16_t burst_pixels)
-{
-	params->pp.npb = burst_pixels - 1;
-}
-
 static void ipu_ch_param_set_buffer(union chan_param_mem *params,
 				    dma_addr_t buf0, dma_addr_t buf1)
 {
@@ -690,23 +685,6 @@ static int ipu_init_channel_buffer(struct idmac_channel *ichan,
 	ipu_ch_param_set_size(&params, pixel_fmt, width, height, stride_bytes);
 	ipu_ch_param_set_buffer(&params, phyaddr_0, phyaddr_1);
 	ipu_ch_param_set_rotation(&params, rot_mode);
-	/* Some channels (rotation) have restriction on burst length */
-	switch (channel) {
-	case IDMAC_IC_7:	/* Hangs with burst 8, 16, other values
-				   invalid - Table 44-30 */
-/*
-		ipu_ch_param_set_burst_size(&params, 8);
- */
-		break;
-	case IDMAC_SDC_0:
-	case IDMAC_SDC_1:
-		/* In original code only IPU_PIX_FMT_RGB565 was setting burst */
-		ipu_ch_param_set_burst_size(&params, 16);
-		break;
-	case IDMAC_IC_0:
-	default:
-		break;
-	}
 
 	spin_lock_irqsave(&ipu->lock, flags);
 
@@ -889,14 +867,7 @@ static dma_cookie_t idmac_tx_submit(struct dma_async_tx_descriptor *tx)
 
 	dev_dbg(dev, "Submitting sg %p\n", &desc->sg[0]);
 
-	cookie = ichan->dma_chan.cookie;
-
-	if (++cookie < 0)
-		cookie = 1;
-
-	/* from dmaengine.h: "last cookie value returned to client" */
-	ichan->dma_chan.cookie = cookie;
-	tx->cookie = cookie;
+	cookie = dma_cookie_assign(tx);
 
 	/* ipu->lock can be taken under ichan->lock, but not v.v. */
 	spin_lock_irqsave(&ichan->lock, flags);
@@ -1318,7 +1289,7 @@ static irqreturn_t idmac_interrupt(int irq, void *dev_id)
 	/* Flip the active buffer - even if update above failed */
 	ichan->active_buffer = !ichan->active_buffer;
 	if (done)
-		ichan->completed = desc->txd.cookie;
+		dma_cookie_complete(&desc->txd);
 
 	callback = desc->txd.callback;
 	callback_param = desc->txd.callback_param;
@@ -1364,7 +1335,8 @@ static void ipu_gc_tasklet(unsigned long arg)
 /* Allocate and initialise a transfer descriptor. */
 static struct dma_async_tx_descriptor *idmac_prep_slave_sg(struct dma_chan *chan,
 		struct scatterlist *sgl, unsigned int sg_len,
-		enum dma_data_direction direction, unsigned long tx_flags)
+		enum dma_transfer_direction direction, unsigned long tx_flags,
+		void *context)
 {
 	struct idmac_channel *ichan = to_idmac_chan(chan);
 	struct idmac_tx_desc *desc = NULL;
@@ -1376,7 +1348,7 @@ static struct dma_async_tx_descriptor *idmac_prep_slave_sg(struct dma_chan *chan
 	    chan->chan_id != IDMAC_IC_7)
 		return NULL;
 
-	if (direction != DMA_FROM_DEVICE && direction != DMA_TO_DEVICE) {
+	if (direction != DMA_DEV_TO_MEM && direction != DMA_MEM_TO_DEV) {
 		dev_err(chan->device->dev, "Invalid DMA direction %d!\n", direction);
 		return NULL;
 	}
@@ -1533,8 +1505,7 @@ static int idmac_alloc_chan_resources(struct dma_chan *chan)
 	BUG_ON(chan->client_count > 1);
 	WARN_ON(ichan->status != IPU_CHANNEL_FREE);
 
-	chan->cookie		= 1;
-	ichan->completed	= -ENXIO;
+	dma_cookie_init(chan);
 
 	ret = ipu_irq_map(chan->chan_id);
 	if (ret < 0)
@@ -1623,9 +1594,7 @@ static void idmac_free_chan_resources(struct dma_chan *chan)
 static enum dma_status idmac_tx_status(struct dma_chan *chan,
 		       dma_cookie_t cookie, struct dma_tx_state *txstate)
 {
-	struct idmac_channel *ichan = to_idmac_chan(chan);
-
-	dma_set_tx_state(txstate, ichan->completed, chan->cookie, 0);
+	dma_set_tx_state(txstate, chan->completed_cookie, chan->cookie, 0);
 	if (cookie != chan->cookie)
 		return DMA_ERROR;
 	return DMA_SUCCESS;
@@ -1661,11 +1630,10 @@ static int __init ipu_idmac_init(struct ipu *ipu)
 
 		ichan->status		= IPU_CHANNEL_FREE;
 		ichan->sec_chan_en	= false;
-		ichan->completed	= -ENXIO;
 		snprintf(ichan->eof_name, sizeof(ichan->eof_name), "IDMAC EOF %d", i);
 
 		dma_chan->device	= &idmac->dma;
-		dma_chan->cookie	= 1;
+		dma_cookie_init(dma_chan);
 		dma_chan->chan_id	= i;
 		list_add_tail(&dma_chan->device_node, &dma->channels);
 	}
diff --git a/drivers/dma/mpc512x_dma.c b/drivers/dma/mpc512x_dma.c
index 8ba4edc6185e..2ab0a3d0eed5 100644
--- a/drivers/dma/mpc512x_dma.c
+++ b/drivers/dma/mpc512x_dma.c
@@ -44,6 +44,8 @@
 
 #include <linux/random.h>
 
+#include "dmaengine.h"
+
 /* Number of DMA Transfer descriptors allocated per channel */
 #define MPC_DMA_DESCRIPTORS	64
 
@@ -188,7 +190,6 @@ struct mpc_dma_chan {
 	struct list_head		completed;
 	struct mpc_dma_tcd		*tcd;
 	dma_addr_t			tcd_paddr;
-	dma_cookie_t			completed_cookie;
 
 	/* Lock for this structure */
 	spinlock_t			lock;
@@ -365,7 +366,7 @@ static void mpc_dma_process_completed(struct mpc_dma *mdma)
 		/* Free descriptors */
 		spin_lock_irqsave(&mchan->lock, flags);
 		list_splice_tail_init(&list, &mchan->free);
-		mchan->completed_cookie = last_cookie;
+		mchan->chan.completed_cookie = last_cookie;
 		spin_unlock_irqrestore(&mchan->lock, flags);
 	}
 }
@@ -438,13 +439,7 @@ static dma_cookie_t mpc_dma_tx_submit(struct dma_async_tx_descriptor *txd)
 		mpc_dma_execute(mchan);
 
 	/* Update cookie */
-	cookie = mchan->chan.cookie + 1;
-	if (cookie <= 0)
-		cookie = 1;
-
-	mchan->chan.cookie = cookie;
-	mdesc->desc.cookie = cookie;
-
+	cookie = dma_cookie_assign(txd);
 	spin_unlock_irqrestore(&mchan->lock, flags);
 
 	return cookie;
@@ -562,17 +557,14 @@ mpc_dma_tx_status(struct dma_chan *chan, dma_cookie_t cookie,
 	       struct dma_tx_state *txstate)
 {
 	struct mpc_dma_chan *mchan = dma_chan_to_mpc_dma_chan(chan);
+	enum dma_status ret;
 	unsigned long flags;
-	dma_cookie_t last_used;
-	dma_cookie_t last_complete;
 
 	spin_lock_irqsave(&mchan->lock, flags);
-	last_used = mchan->chan.cookie;
-	last_complete = mchan->completed_cookie;
+	ret = dma_cookie_status(chan, cookie, txstate);
 	spin_unlock_irqrestore(&mchan->lock, flags);
 
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
-	return dma_async_is_complete(cookie, last_complete, last_used);
+	return ret;
 }
 
 /* Prepare descriptor for memory to memory copy */
@@ -741,8 +733,7 @@ static int __devinit mpc_dma_probe(struct platform_device *op)
 		mchan = &mdma->channels[i];
 
 		mchan->chan.device = dma;
-		mchan->chan.cookie = 1;
-		mchan->completed_cookie = mchan->chan.cookie;
+		dma_cookie_init(&mchan->chan);
 
 		INIT_LIST_HEAD(&mchan->free);
 		INIT_LIST_HEAD(&mchan->prepared);
@@ -835,17 +826,7 @@ static struct platform_driver mpc_dma_driver = {
 	},
 };
 
-static int __init mpc_dma_init(void)
-{
-	return platform_driver_register(&mpc_dma_driver);
-}
-module_init(mpc_dma_init);
-
-static void __exit mpc_dma_exit(void)
-{
-	platform_driver_unregister(&mpc_dma_driver);
-}
-module_exit(mpc_dma_exit);
+module_platform_driver(mpc_dma_driver);
 
 MODULE_LICENSE("GPL");
 MODULE_AUTHOR("Piotr Ziecik <kosmo@semihalf.com>");
diff --git a/drivers/dma/mv_xor.c b/drivers/dma/mv_xor.c
index e779b434af45..fa5d55fea46c 100644
--- a/drivers/dma/mv_xor.c
+++ b/drivers/dma/mv_xor.c
@@ -26,6 +26,8 @@
 #include <linux/platform_device.h>
 #include <linux/memory.h>
 #include <plat/mv_xor.h>
+
+#include "dmaengine.h"
 #include "mv_xor.h"
 
 static void mv_xor_issue_pending(struct dma_chan *chan);
@@ -435,7 +437,7 @@ static void __mv_xor_slot_cleanup(struct mv_xor_chan *mv_chan)
 	}
 
 	if (cookie > 0)
-		mv_chan->completed_cookie = cookie;
+		mv_chan->common.completed_cookie = cookie;
 }
 
 static void
@@ -534,18 +536,6 @@ retry:
 	return NULL;
 }
 
-static dma_cookie_t
-mv_desc_assign_cookie(struct mv_xor_chan *mv_chan,
-		      struct mv_xor_desc_slot *desc)
-{
-	dma_cookie_t cookie = mv_chan->common.cookie;
-
-	if (++cookie < 0)
-		cookie = 1;
-	mv_chan->common.cookie = desc->async_tx.cookie = cookie;
-	return cookie;
-}
-
 /************************ DMA engine API functions ****************************/
 static dma_cookie_t
 mv_xor_tx_submit(struct dma_async_tx_descriptor *tx)
@@ -563,7 +553,7 @@ mv_xor_tx_submit(struct dma_async_tx_descriptor *tx)
 	grp_start = sw_desc->group_head;
 
 	spin_lock_bh(&mv_chan->lock);
-	cookie = mv_desc_assign_cookie(mv_chan, sw_desc);
+	cookie = dma_cookie_assign(tx);
 
 	if (list_empty(&mv_chan->chain))
 		list_splice_init(&sw_desc->tx_list, &mv_chan->chain);
@@ -820,27 +810,16 @@ static enum dma_status mv_xor_status(struct dma_chan *chan,
 					  struct dma_tx_state *txstate)
 {
 	struct mv_xor_chan *mv_chan = to_mv_xor_chan(chan);
-	dma_cookie_t last_used;
-	dma_cookie_t last_complete;
 	enum dma_status ret;
 
-	last_used = chan->cookie;
-	last_complete = mv_chan->completed_cookie;
-	mv_chan->is_complete_cookie = cookie;
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
-
-	ret = dma_async_is_complete(cookie, last_complete, last_used);
+	ret = dma_cookie_status(chan, cookie, txstate);
 	if (ret == DMA_SUCCESS) {
 		mv_xor_clean_completed_slots(mv_chan);
 		return ret;
 	}
 	mv_xor_slot_cleanup(mv_chan);
 
-	last_used = chan->cookie;
-	last_complete = mv_chan->completed_cookie;
-
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
-	return dma_async_is_complete(cookie, last_complete, last_used);
+	return dma_cookie_status(chan, cookie, txstate);
 }
 
 static void mv_dump_xor_regs(struct mv_xor_chan *chan)
@@ -1214,6 +1193,7 @@ static int __devinit mv_xor_probe(struct platform_device *pdev)
 	INIT_LIST_HEAD(&mv_chan->completed_slots);
 	INIT_LIST_HEAD(&mv_chan->all_slots);
 	mv_chan->common.device = dma_dev;
+	dma_cookie_init(&mv_chan->common);
 
 	list_add_tail(&mv_chan->common.device_node, &dma_dev->channels);
 
diff --git a/drivers/dma/mv_xor.h b/drivers/dma/mv_xor.h
index 977b592e976b..654876b7ba1d 100644
--- a/drivers/dma/mv_xor.h
+++ b/drivers/dma/mv_xor.h
@@ -78,7 +78,6 @@ struct mv_xor_device {
 /**
  * struct mv_xor_chan - internal representation of a XOR channel
  * @pending: allows batching of hardware operations
- * @completed_cookie: identifier for the most recently completed operation
  * @lock: serializes enqueue/dequeue operations to the descriptors pool
  * @mmr_base: memory mapped register base
  * @idx: the index of the xor channel
@@ -93,7 +92,6 @@ struct mv_xor_device {
  */
 struct mv_xor_chan {
 	int			pending;
-	dma_cookie_t		completed_cookie;
 	spinlock_t		lock; /* protects the descriptor slot pool */
 	void __iomem		*mmr_base;
 	unsigned int		idx;
@@ -109,7 +107,6 @@ struct mv_xor_chan {
 #ifdef USE_TIMER
 	unsigned long		cleanup_time;
 	u32			current_on_last_cleanup;
-	dma_cookie_t		is_complete_cookie;
 #endif
 };
 
diff --git a/drivers/dma/mxs-dma.c b/drivers/dma/mxs-dma.c
index fc903c0ed234..c81ef7e10e08 100644
--- a/drivers/dma/mxs-dma.c
+++ b/drivers/dma/mxs-dma.c
@@ -22,12 +22,14 @@
 #include <linux/platform_device.h>
 #include <linux/dmaengine.h>
 #include <linux/delay.h>
+#include <linux/fsl/mxs-dma.h>
 
 #include <asm/irq.h>
 #include <mach/mxs.h>
-#include <mach/dma.h>
 #include <mach/common.h>
 
+#include "dmaengine.h"
+
 /*
  * NOTE: The term "PIO" throughout the mxs-dma implementation means
  * PIO mode of mxs apbh-dma and apbx-dma.  With this working mode,
@@ -44,7 +46,6 @@
 #define HW_APBHX_CTRL0				0x000
 #define BM_APBH_CTRL0_APB_BURST8_EN		(1 << 29)
 #define BM_APBH_CTRL0_APB_BURST_EN		(1 << 28)
-#define BP_APBH_CTRL0_CLKGATE_CHANNEL		8
 #define BP_APBH_CTRL0_RESET_CHANNEL		16
 #define HW_APBHX_CTRL1				0x010
 #define HW_APBHX_CTRL2				0x020
@@ -111,7 +112,7 @@ struct mxs_dma_chan {
 	int				chan_irq;
 	struct mxs_dma_ccw		*ccw;
 	dma_addr_t			ccw_phys;
-	dma_cookie_t			last_completed;
+	int				desc_count;
 	enum dma_status			status;
 	unsigned int			flags;
 #define MXS_DMA_SG_LOOP			(1 << 0)
@@ -130,23 +131,6 @@ struct mxs_dma_engine {
 	struct mxs_dma_chan		mxs_chans[MXS_DMA_CHANNELS];
 };
 
-static inline void mxs_dma_clkgate(struct mxs_dma_chan *mxs_chan, int enable)
-{
-	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
-	int chan_id = mxs_chan->chan.chan_id;
-	int set_clr = enable ? MXS_CLR_ADDR : MXS_SET_ADDR;
-
-	/* enable apbh channel clock */
-	if (dma_is_apbh()) {
-		if (apbh_is_old())
-			writel(1 << (chan_id + BP_APBH_CTRL0_CLKGATE_CHANNEL),
-				mxs_dma->base + HW_APBHX_CTRL0 + set_clr);
-		else
-			writel(1 << chan_id,
-				mxs_dma->base + HW_APBHX_CTRL0 + set_clr);
-	}
-}
-
 static void mxs_dma_reset_chan(struct mxs_dma_chan *mxs_chan)
 {
 	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
@@ -165,9 +149,6 @@ static void mxs_dma_enable_chan(struct mxs_dma_chan *mxs_chan)
 	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
 	int chan_id = mxs_chan->chan.chan_id;
 
-	/* clkgate needs to be enabled before writing other registers */
-	mxs_dma_clkgate(mxs_chan, 1);
-
 	/* set cmd_addr up */
 	writel(mxs_chan->ccw_phys,
 		mxs_dma->base + HW_APBHX_CHn_NXTCMDAR(chan_id));
@@ -178,9 +159,6 @@ static void mxs_dma_enable_chan(struct mxs_dma_chan *mxs_chan)
 
 static void mxs_dma_disable_chan(struct mxs_dma_chan *mxs_chan)
 {
-	/* disable apbh channel clock */
-	mxs_dma_clkgate(mxs_chan, 0);
-
 	mxs_chan->status = DMA_SUCCESS;
 }
 
@@ -216,19 +194,6 @@ static void mxs_dma_resume_chan(struct mxs_dma_chan *mxs_chan)
 	mxs_chan->status = DMA_IN_PROGRESS;
 }
 
-static dma_cookie_t mxs_dma_assign_cookie(struct mxs_dma_chan *mxs_chan)
-{
-	dma_cookie_t cookie = mxs_chan->chan.cookie;
-
-	if (++cookie < 0)
-		cookie = 1;
-
-	mxs_chan->chan.cookie = cookie;
-	mxs_chan->desc.cookie = cookie;
-
-	return cookie;
-}
-
 static struct mxs_dma_chan *to_mxs_dma_chan(struct dma_chan *chan)
 {
 	return container_of(chan, struct mxs_dma_chan, chan);
@@ -240,7 +205,7 @@ static dma_cookie_t mxs_dma_tx_submit(struct dma_async_tx_descriptor *tx)
 
 	mxs_dma_enable_chan(mxs_chan);
 
-	return mxs_dma_assign_cookie(mxs_chan);
+	return dma_cookie_assign(tx);
 }
 
 static void mxs_dma_tasklet(unsigned long data)
@@ -268,7 +233,7 @@ static irqreturn_t mxs_dma_int_handler(int irq, void *dev_id)
 	/*
 	 * When both completion and error of termination bits set at the
 	 * same time, we do not take it as an error.  IOW, it only becomes
-	 * an error we need to handler here in case of ether it's (1) an bus
+	 * an error we need to handle here in case of either it's (1) a bus
 	 * error or (2) a termination error with no completion.
 	 */
 	stat2 = ((stat2 >> MXS_DMA_CHANNELS) & stat2) | /* (1) */
@@ -297,7 +262,7 @@ static irqreturn_t mxs_dma_int_handler(int irq, void *dev_id)
 		stat1 &= ~(1 << channel);
 
 		if (mxs_chan->status == DMA_SUCCESS)
-			mxs_chan->last_completed = mxs_chan->desc.cookie;
+			dma_cookie_complete(&mxs_chan->desc);
 
 		/* schedule tasklet on this channel */
 		tasklet_schedule(&mxs_chan->tasklet);
@@ -338,10 +303,7 @@ static int mxs_dma_alloc_chan_resources(struct dma_chan *chan)
 	if (ret)
 		goto err_clk;
 
-	/* clkgate needs to be enabled for reset to finish */
-	mxs_dma_clkgate(mxs_chan, 1);
 	mxs_dma_reset_chan(mxs_chan);
-	mxs_dma_clkgate(mxs_chan, 0);
 
 	dma_async_tx_descriptor_init(&mxs_chan->desc, chan);
 	mxs_chan->desc.tx_submit = mxs_dma_tx_submit;
@@ -375,10 +337,32 @@ static void mxs_dma_free_chan_resources(struct dma_chan *chan)
 	clk_disable_unprepare(mxs_dma->clk);
 }
 
+/*
+ * How to use the flags for ->device_prep_slave_sg() :
+ *    [1] If there is only one DMA command in the DMA chain, the code should be:
+ *            ......
+ *            ->device_prep_slave_sg(DMA_CTRL_ACK);
+ *            ......
+ *    [2] If there are two DMA commands in the DMA chain, the code should be
+ *            ......
+ *            ->device_prep_slave_sg(0);
+ *            ......
+ *            ->device_prep_slave_sg(DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+ *            ......
+ *    [3] If there are more than two DMA commands in the DMA chain, the code
+ *        should be:
+ *            ......
+ *            ->device_prep_slave_sg(0);                                // First
+ *            ......
+ *            ->device_prep_slave_sg(DMA_PREP_INTERRUPT [| DMA_CTRL_ACK]);
+ *            ......
+ *            ->device_prep_slave_sg(DMA_PREP_INTERRUPT | DMA_CTRL_ACK); // Last
+ *            ......
+ */
 static struct dma_async_tx_descriptor *mxs_dma_prep_slave_sg(
 		struct dma_chan *chan, struct scatterlist *sgl,
-		unsigned int sg_len, enum dma_data_direction direction,
-		unsigned long append)
+		unsigned int sg_len, enum dma_transfer_direction direction,
+		unsigned long flags, void *context)
 {
 	struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan);
 	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
@@ -386,7 +370,8 @@ static struct dma_async_tx_descriptor *mxs_dma_prep_slave_sg(
 	struct scatterlist *sg;
 	int i, j;
 	u32 *pio;
-	static int idx;
+	bool append = flags & DMA_PREP_INTERRUPT;
+	int idx = append ? mxs_chan->desc_count : 0;
 
 	if (mxs_chan->status == DMA_IN_PROGRESS && !append)
 		return NULL;
@@ -412,12 +397,11 @@ static struct dma_async_tx_descriptor *mxs_dma_prep_slave_sg(
 		ccw->bits |= CCW_CHAIN;
 		ccw->bits &= ~CCW_IRQ;
 		ccw->bits &= ~CCW_DEC_SEM;
-		ccw->bits &= ~CCW_WAIT4END;
 	} else {
 		idx = 0;
 	}
 
-	if (direction == DMA_NONE) {
+	if (direction == DMA_TRANS_NONE) {
 		ccw = &mxs_chan->ccw[idx++];
 		pio = (u32 *) sgl;
 
@@ -427,7 +411,8 @@ static struct dma_async_tx_descriptor *mxs_dma_prep_slave_sg(
 		ccw->bits = 0;
 		ccw->bits |= CCW_IRQ;
 		ccw->bits |= CCW_DEC_SEM;
-		ccw->bits |= CCW_WAIT4END;
+		if (flags & DMA_CTRL_ACK)
+			ccw->bits |= CCW_WAIT4END;
 		ccw->bits |= CCW_HALT_ON_TERM;
 		ccw->bits |= CCW_TERM_FLUSH;
 		ccw->bits |= BF_CCW(sg_len, PIO_NUM);
@@ -450,7 +435,7 @@ static struct dma_async_tx_descriptor *mxs_dma_prep_slave_sg(
 			ccw->bits |= CCW_CHAIN;
 			ccw->bits |= CCW_HALT_ON_TERM;
 			ccw->bits |= CCW_TERM_FLUSH;
-			ccw->bits |= BF_CCW(direction == DMA_FROM_DEVICE ?
+			ccw->bits |= BF_CCW(direction == DMA_DEV_TO_MEM ?
 					MXS_DMA_CMD_WRITE : MXS_DMA_CMD_READ,
 					COMMAND);
 
@@ -458,10 +443,12 @@ static struct dma_async_tx_descriptor *mxs_dma_prep_slave_sg(
 				ccw->bits &= ~CCW_CHAIN;
 				ccw->bits |= CCW_IRQ;
 				ccw->bits |= CCW_DEC_SEM;
-				ccw->bits |= CCW_WAIT4END;
+				if (flags & DMA_CTRL_ACK)
+					ccw->bits |= CCW_WAIT4END;
 			}
 		}
 	}
+	mxs_chan->desc_count = idx;
 
 	return &mxs_chan->desc;
 
@@ -472,7 +459,8 @@ err_out:
 
 static struct dma_async_tx_descriptor *mxs_dma_prep_dma_cyclic(
 		struct dma_chan *chan, dma_addr_t dma_addr, size_t buf_len,
-		size_t period_len, enum dma_data_direction direction)
+		size_t period_len, enum dma_transfer_direction direction,
+		void *context)
 {
 	struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan);
 	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
@@ -515,7 +503,7 @@ static struct dma_async_tx_descriptor *mxs_dma_prep_dma_cyclic(
 		ccw->bits |= CCW_IRQ;
 		ccw->bits |= CCW_HALT_ON_TERM;
 		ccw->bits |= CCW_TERM_FLUSH;
-		ccw->bits |= BF_CCW(direction == DMA_FROM_DEVICE ?
+		ccw->bits |= BF_CCW(direction == DMA_DEV_TO_MEM ?
 				MXS_DMA_CMD_WRITE : MXS_DMA_CMD_READ, COMMAND);
 
 		dma_addr += period_len;
@@ -523,6 +511,7 @@ static struct dma_async_tx_descriptor *mxs_dma_prep_dma_cyclic(
 
 		i++;
 	}
+	mxs_chan->desc_count = i;
 
 	return &mxs_chan->desc;
 
@@ -539,8 +528,8 @@ static int mxs_dma_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
 
 	switch (cmd) {
 	case DMA_TERMINATE_ALL:
-		mxs_dma_disable_chan(mxs_chan);
 		mxs_dma_reset_chan(mxs_chan);
+		mxs_dma_disable_chan(mxs_chan);
 		break;
 	case DMA_PAUSE:
 		mxs_dma_pause_chan(mxs_chan);
@@ -562,7 +551,7 @@ static enum dma_status mxs_dma_tx_status(struct dma_chan *chan,
 	dma_cookie_t last_used;
 
 	last_used = chan->cookie;
-	dma_set_tx_state(txstate, mxs_chan->last_completed, last_used, 0);
+	dma_set_tx_state(txstate, chan->completed_cookie, last_used, 0);
 
 	return mxs_chan->status;
 }
@@ -580,7 +569,7 @@ static int __init mxs_dma_init(struct mxs_dma_engine *mxs_dma)
 
 	ret = clk_prepare_enable(mxs_dma->clk);
 	if (ret)
-		goto err_out;
+		return ret;
 
 	ret = mxs_reset_block(mxs_dma->base);
 	if (ret)
@@ -604,11 +593,8 @@ static int __init mxs_dma_init(struct mxs_dma_engine *mxs_dma)
 	writel(MXS_DMA_CHANNELS_MASK << MXS_DMA_CHANNELS,
 		mxs_dma->base + HW_APBHX_CTRL1 + MXS_SET_ADDR);
 
-	clk_disable_unprepare(mxs_dma->clk);
-
-	return 0;
-
 err_out:
+	clk_disable_unprepare(mxs_dma->clk);
 	return ret;
 }
 
@@ -657,6 +643,7 @@ static int __init mxs_dma_probe(struct platform_device *pdev)
 
 		mxs_chan->mxs_dma = mxs_dma;
 		mxs_chan->chan.device = &mxs_dma->dma_device;
+		dma_cookie_init(&mxs_chan->chan);
 
 		tasklet_init(&mxs_chan->tasklet, mxs_dma_tasklet,
 			     (unsigned long) mxs_chan);
diff --git a/drivers/dma/pch_dma.c b/drivers/dma/pch_dma.c
index a6d0e3dbed07..65c0495a6d40 100644
--- a/drivers/dma/pch_dma.c
+++ b/drivers/dma/pch_dma.c
@@ -1,7 +1,7 @@
 /*
  * Topcliff PCH DMA controller driver
  * Copyright (c) 2010 Intel Corporation
- * Copyright (C) 2011 OKI SEMICONDUCTOR CO., LTD.
+ * Copyright (C) 2011 LAPIS Semiconductor Co., Ltd.
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 2 as
@@ -25,6 +25,8 @@
 #include <linux/module.h>
 #include <linux/pch_dma.h>
 
+#include "dmaengine.h"
+
 #define DRV_NAME "pch-dma"
 
 #define DMA_CTL0_DISABLE		0x0
@@ -99,13 +101,12 @@ struct pch_dma_desc {
 struct pch_dma_chan {
 	struct dma_chan		chan;
 	void __iomem *membase;
-	enum dma_data_direction	dir;
+	enum dma_transfer_direction dir;
 	struct tasklet_struct	tasklet;
 	unsigned long		err_status;
 
 	spinlock_t		lock;
 
-	dma_cookie_t		completed_cookie;
 	struct list_head	active_list;
 	struct list_head	queue;
 	struct list_head	free_list;
@@ -224,7 +225,7 @@ static void pdc_set_dir(struct dma_chan *chan)
 		mask_ctl = DMA_MASK_CTL0_MODE & ~(DMA_CTL0_MODE_MASK_BITS <<
 				       (DMA_CTL0_BITS_PER_CH * chan->chan_id));
 		val &= mask_mode;
-		if (pd_chan->dir == DMA_TO_DEVICE)
+		if (pd_chan->dir == DMA_MEM_TO_DEV)
 			val |= 0x1 << (DMA_CTL0_BITS_PER_CH * chan->chan_id +
 				       DMA_CTL0_DIR_SHIFT_BITS);
 		else
@@ -242,7 +243,7 @@ static void pdc_set_dir(struct dma_chan *chan)
 		mask_ctl = DMA_MASK_CTL2_MODE & ~(DMA_CTL0_MODE_MASK_BITS <<
 						 (DMA_CTL0_BITS_PER_CH * ch));
 		val &= mask_mode;
-		if (pd_chan->dir == DMA_TO_DEVICE)
+		if (pd_chan->dir == DMA_MEM_TO_DEV)
 			val |= 0x1 << (DMA_CTL0_BITS_PER_CH * ch +
 				       DMA_CTL0_DIR_SHIFT_BITS);
 		else
@@ -416,20 +417,6 @@ static void pdc_advance_work(struct pch_dma_chan *pd_chan)
 	}
 }
 
-static dma_cookie_t pdc_assign_cookie(struct pch_dma_chan *pd_chan,
-				      struct pch_dma_desc *desc)
-{
-	dma_cookie_t cookie = pd_chan->chan.cookie;
-
-	if (++cookie < 0)
-		cookie = 1;
-
-	pd_chan->chan.cookie = cookie;
-	desc->txd.cookie = cookie;
-
-	return cookie;
-}
-
 static dma_cookie_t pd_tx_submit(struct dma_async_tx_descriptor *txd)
 {
 	struct pch_dma_desc *desc = to_pd_desc(txd);
@@ -437,7 +424,7 @@ static dma_cookie_t pd_tx_submit(struct dma_async_tx_descriptor *txd)
 	dma_cookie_t cookie;
 
 	spin_lock(&pd_chan->lock);
-	cookie = pdc_assign_cookie(pd_chan, desc);
+	cookie = dma_cookie_assign(txd);
 
 	if (list_empty(&pd_chan->active_list)) {
 		list_add_tail(&desc->desc_node, &pd_chan->active_list);
@@ -544,7 +531,7 @@ static int pd_alloc_chan_resources(struct dma_chan *chan)
 	spin_lock_irq(&pd_chan->lock);
 	list_splice(&tmp_list, &pd_chan->free_list);
 	pd_chan->descs_allocated = i;
-	pd_chan->completed_cookie = chan->cookie = 1;
+	dma_cookie_init(chan);
 	spin_unlock_irq(&pd_chan->lock);
 
 	pdc_enable_irq(chan, 1);
@@ -578,19 +565,12 @@ static enum dma_status pd_tx_status(struct dma_chan *chan, dma_cookie_t cookie,
 				    struct dma_tx_state *txstate)
 {
 	struct pch_dma_chan *pd_chan = to_pd_chan(chan);
-	dma_cookie_t last_used;
-	dma_cookie_t last_completed;
-	int ret;
+	enum dma_status ret;
 
 	spin_lock_irq(&pd_chan->lock);
-	last_completed = pd_chan->completed_cookie;
-	last_used = chan->cookie;
+	ret = dma_cookie_status(chan, cookie, txstate);
 	spin_unlock_irq(&pd_chan->lock);
 
-	ret = dma_async_is_complete(cookie, last_completed, last_used);
-
-	dma_set_tx_state(txstate, last_completed, last_used, 0);
-
 	return ret;
 }
 
@@ -607,7 +587,8 @@ static void pd_issue_pending(struct dma_chan *chan)
 
 static struct dma_async_tx_descriptor *pd_prep_slave_sg(struct dma_chan *chan,
 			struct scatterlist *sgl, unsigned int sg_len,
-			enum dma_data_direction direction, unsigned long flags)
+			enum dma_transfer_direction direction, unsigned long flags,
+			void *context)
 {
 	struct pch_dma_chan *pd_chan = to_pd_chan(chan);
 	struct pch_dma_slave *pd_slave = chan->private;
@@ -623,9 +604,9 @@ static struct dma_async_tx_descriptor *pd_prep_slave_sg(struct dma_chan *chan,
 		return NULL;
 	}
 
-	if (direction == DMA_FROM_DEVICE)
+	if (direction == DMA_DEV_TO_MEM)
 		reg = pd_slave->rx_reg;
-	else if (direction == DMA_TO_DEVICE)
+	else if (direction == DMA_MEM_TO_DEV)
 		reg = pd_slave->tx_reg;
 	else
 		return NULL;
@@ -932,7 +913,7 @@ static int __devinit pch_dma_probe(struct pci_dev *pdev,
 		struct pch_dma_chan *pd_chan = &pd->channels[i];
 
 		pd_chan->chan.device = &pd->dma;
-		pd_chan->chan.cookie = 1;
+		dma_cookie_init(&pd_chan->chan);
 
 		pd_chan->membase = &regs->desc[i];
 
@@ -1018,6 +999,8 @@ static void __devexit pch_dma_remove(struct pci_dev *pdev)
 #define PCI_DEVICE_ID_ML7223_DMA2_4CH	0x800E
 #define PCI_DEVICE_ID_ML7223_DMA3_4CH	0x8017
 #define PCI_DEVICE_ID_ML7223_DMA4_4CH	0x803B
+#define PCI_DEVICE_ID_ML7831_DMA1_8CH	0x8810
+#define PCI_DEVICE_ID_ML7831_DMA2_4CH	0x8815
 
 DEFINE_PCI_DEVICE_TABLE(pch_dma_id_table) = {
 	{ PCI_VDEVICE(INTEL, PCI_DEVICE_ID_EG20T_PCH_DMA_8CH), 8 },
@@ -1030,6 +1013,8 @@ DEFINE_PCI_DEVICE_TABLE(pch_dma_id_table) = {
 	{ PCI_VDEVICE(ROHM, PCI_DEVICE_ID_ML7223_DMA2_4CH), 4}, /* Video SPI */
 	{ PCI_VDEVICE(ROHM, PCI_DEVICE_ID_ML7223_DMA3_4CH), 4}, /* Security */
 	{ PCI_VDEVICE(ROHM, PCI_DEVICE_ID_ML7223_DMA4_4CH), 4}, /* FPGA */
+	{ PCI_VDEVICE(ROHM, PCI_DEVICE_ID_ML7831_DMA1_8CH), 8}, /* UART */
+	{ PCI_VDEVICE(ROHM, PCI_DEVICE_ID_ML7831_DMA2_4CH), 4}, /* SPI */
 	{ 0, },
 };
 
@@ -1057,7 +1042,7 @@ static void __exit pch_dma_exit(void)
 module_init(pch_dma_init);
 module_exit(pch_dma_exit);
 
-MODULE_DESCRIPTION("Intel EG20T PCH / OKI SEMICONDUCTOR ML7213 IOH "
+MODULE_DESCRIPTION("Intel EG20T PCH / LAPIS Semicon ML7213/ML7223/ML7831 IOH "
 		   "DMA controller driver");
 MODULE_AUTHOR("Yong Wang <yong.y.wang@intel.com>");
 MODULE_LICENSE("GPL v2");
diff --git a/drivers/dma/pl330.c b/drivers/dma/pl330.c
index 09adcfcd953e..282caf118be8 100644
--- a/drivers/dma/pl330.c
+++ b/drivers/dma/pl330.c
@@ -1,4 +1,6 @@
-/* linux/drivers/dma/pl330.c
+/*
+ * Copyright (c) 2012 Samsung Electronics Co., Ltd.
+ *		http://www.samsung.com
  *
  * Copyright (C) 2010 Samsung Electronics Co. Ltd.
  *	Jaswinder Singh <jassi.brar@samsung.com>
@@ -9,10 +11,15 @@
  * (at your option) any later version.
  */
 
+#include <linux/kernel.h>
 #include <linux/io.h>
 #include <linux/init.h>
 #include <linux/slab.h>
 #include <linux/module.h>
+#include <linux/string.h>
+#include <linux/delay.h>
+#include <linux/interrupt.h>
+#include <linux/dma-mapping.h>
 #include <linux/dmaengine.h>
 #include <linux/interrupt.h>
 #include <linux/amba/bus.h>
@@ -21,8 +28,497 @@
 #include <linux/scatterlist.h>
 #include <linux/of.h>
 
+#include "dmaengine.h"
+#define PL330_MAX_CHAN		8
+#define PL330_MAX_IRQS		32
+#define PL330_MAX_PERI		32
+
+enum pl330_srccachectrl {
+	SCCTRL0,	/* Noncacheable and nonbufferable */
+	SCCTRL1,	/* Bufferable only */
+	SCCTRL2,	/* Cacheable, but do not allocate */
+	SCCTRL3,	/* Cacheable and bufferable, but do not allocate */
+	SINVALID1,
+	SINVALID2,
+	SCCTRL6,	/* Cacheable write-through, allocate on reads only */
+	SCCTRL7,	/* Cacheable write-back, allocate on reads only */
+};
+
+enum pl330_dstcachectrl {
+	DCCTRL0,	/* Noncacheable and nonbufferable */
+	DCCTRL1,	/* Bufferable only */
+	DCCTRL2,	/* Cacheable, but do not allocate */
+	DCCTRL3,	/* Cacheable and bufferable, but do not allocate */
+	DINVALID1,	/* AWCACHE = 0x1000 */
+	DINVALID2,
+	DCCTRL6,	/* Cacheable write-through, allocate on writes only */
+	DCCTRL7,	/* Cacheable write-back, allocate on writes only */
+};
+
+enum pl330_byteswap {
+	SWAP_NO,
+	SWAP_2,
+	SWAP_4,
+	SWAP_8,
+	SWAP_16,
+};
+
+enum pl330_reqtype {
+	MEMTOMEM,
+	MEMTODEV,
+	DEVTOMEM,
+	DEVTODEV,
+};
+
+/* Register and Bit field Definitions */
+#define DS			0x0
+#define DS_ST_STOP		0x0
+#define DS_ST_EXEC		0x1
+#define DS_ST_CMISS		0x2
+#define DS_ST_UPDTPC		0x3
+#define DS_ST_WFE		0x4
+#define DS_ST_ATBRR		0x5
+#define DS_ST_QBUSY		0x6
+#define DS_ST_WFP		0x7
+#define DS_ST_KILL		0x8
+#define DS_ST_CMPLT		0x9
+#define DS_ST_FLTCMP		0xe
+#define DS_ST_FAULT		0xf
+
+#define DPC			0x4
+#define INTEN			0x20
+#define ES			0x24
+#define INTSTATUS		0x28
+#define INTCLR			0x2c
+#define FSM			0x30
+#define FSC			0x34
+#define FTM			0x38
+
+#define _FTC			0x40
+#define FTC(n)			(_FTC + (n)*0x4)
+
+#define _CS			0x100
+#define CS(n)			(_CS + (n)*0x8)
+#define CS_CNS			(1 << 21)
+
+#define _CPC			0x104
+#define CPC(n)			(_CPC + (n)*0x8)
+
+#define _SA			0x400
+#define SA(n)			(_SA + (n)*0x20)
+
+#define _DA			0x404
+#define DA(n)			(_DA + (n)*0x20)
+
+#define _CC			0x408
+#define CC(n)			(_CC + (n)*0x20)
+
+#define CC_SRCINC		(1 << 0)
+#define CC_DSTINC		(1 << 14)
+#define CC_SRCPRI		(1 << 8)
+#define CC_DSTPRI		(1 << 22)
+#define CC_SRCNS		(1 << 9)
+#define CC_DSTNS		(1 << 23)
+#define CC_SRCIA		(1 << 10)
+#define CC_DSTIA		(1 << 24)
+#define CC_SRCBRSTLEN_SHFT	4
+#define CC_DSTBRSTLEN_SHFT	18
+#define CC_SRCBRSTSIZE_SHFT	1
+#define CC_DSTBRSTSIZE_SHFT	15
+#define CC_SRCCCTRL_SHFT	11
+#define CC_SRCCCTRL_MASK	0x7
+#define CC_DSTCCTRL_SHFT	25
+#define CC_DRCCCTRL_MASK	0x7
+#define CC_SWAP_SHFT		28
+
+#define _LC0			0x40c
+#define LC0(n)			(_LC0 + (n)*0x20)
+
+#define _LC1			0x410
+#define LC1(n)			(_LC1 + (n)*0x20)
+
+#define DBGSTATUS		0xd00
+#define DBG_BUSY		(1 << 0)
+
+#define DBGCMD			0xd04
+#define DBGINST0		0xd08
+#define DBGINST1		0xd0c
+
+#define CR0			0xe00
+#define CR1			0xe04
+#define CR2			0xe08
+#define CR3			0xe0c
+#define CR4			0xe10
+#define CRD			0xe14
+
+#define PERIPH_ID		0xfe0
+#define PERIPH_REV_SHIFT	20
+#define PERIPH_REV_MASK		0xf
+#define PERIPH_REV_R0P0		0
+#define PERIPH_REV_R1P0		1
+#define PERIPH_REV_R1P1		2
+#define PCELL_ID		0xff0
+
+#define CR0_PERIPH_REQ_SET	(1 << 0)
+#define CR0_BOOT_EN_SET		(1 << 1)
+#define CR0_BOOT_MAN_NS		(1 << 2)
+#define CR0_NUM_CHANS_SHIFT	4
+#define CR0_NUM_CHANS_MASK	0x7
+#define CR0_NUM_PERIPH_SHIFT	12
+#define CR0_NUM_PERIPH_MASK	0x1f
+#define CR0_NUM_EVENTS_SHIFT	17
+#define CR0_NUM_EVENTS_MASK	0x1f
+
+#define CR1_ICACHE_LEN_SHIFT	0
+#define CR1_ICACHE_LEN_MASK	0x7
+#define CR1_NUM_ICACHELINES_SHIFT	4
+#define CR1_NUM_ICACHELINES_MASK	0xf
+
+#define CRD_DATA_WIDTH_SHIFT	0
+#define CRD_DATA_WIDTH_MASK	0x7
+#define CRD_WR_CAP_SHIFT	4
+#define CRD_WR_CAP_MASK		0x7
+#define CRD_WR_Q_DEP_SHIFT	8
+#define CRD_WR_Q_DEP_MASK	0xf
+#define CRD_RD_CAP_SHIFT	12
+#define CRD_RD_CAP_MASK		0x7
+#define CRD_RD_Q_DEP_SHIFT	16
+#define CRD_RD_Q_DEP_MASK	0xf
+#define CRD_DATA_BUFF_SHIFT	20
+#define CRD_DATA_BUFF_MASK	0x3ff
+
+#define PART			0x330
+#define DESIGNER		0x41
+#define REVISION		0x0
+#define INTEG_CFG		0x0
+#define PERIPH_ID_VAL		((PART << 0) | (DESIGNER << 12))
+
+#define PCELL_ID_VAL		0xb105f00d
+
+#define PL330_STATE_STOPPED		(1 << 0)
+#define PL330_STATE_EXECUTING		(1 << 1)
+#define PL330_STATE_WFE			(1 << 2)
+#define PL330_STATE_FAULTING		(1 << 3)
+#define PL330_STATE_COMPLETING		(1 << 4)
+#define PL330_STATE_WFP			(1 << 5)
+#define PL330_STATE_KILLING		(1 << 6)
+#define PL330_STATE_FAULT_COMPLETING	(1 << 7)
+#define PL330_STATE_CACHEMISS		(1 << 8)
+#define PL330_STATE_UPDTPC		(1 << 9)
+#define PL330_STATE_ATBARRIER		(1 << 10)
+#define PL330_STATE_QUEUEBUSY		(1 << 11)
+#define PL330_STATE_INVALID		(1 << 15)
+
+#define PL330_STABLE_STATES (PL330_STATE_STOPPED | PL330_STATE_EXECUTING \
+				| PL330_STATE_WFE | PL330_STATE_FAULTING)
+
+#define CMD_DMAADDH		0x54
+#define CMD_DMAEND		0x00
+#define CMD_DMAFLUSHP		0x35
+#define CMD_DMAGO		0xa0
+#define CMD_DMALD		0x04
+#define CMD_DMALDP		0x25
+#define CMD_DMALP		0x20
+#define CMD_DMALPEND		0x28
+#define CMD_DMAKILL		0x01
+#define CMD_DMAMOV		0xbc
+#define CMD_DMANOP		0x18
+#define CMD_DMARMB		0x12
+#define CMD_DMASEV		0x34
+#define CMD_DMAST		0x08
+#define CMD_DMASTP		0x29
+#define CMD_DMASTZ		0x0c
+#define CMD_DMAWFE		0x36
+#define CMD_DMAWFP		0x30
+#define CMD_DMAWMB		0x13
+
+#define SZ_DMAADDH		3
+#define SZ_DMAEND		1
+#define SZ_DMAFLUSHP		2
+#define SZ_DMALD		1
+#define SZ_DMALDP		2
+#define SZ_DMALP		2
+#define SZ_DMALPEND		2
+#define SZ_DMAKILL		1
+#define SZ_DMAMOV		6
+#define SZ_DMANOP		1
+#define SZ_DMARMB		1
+#define SZ_DMASEV		2
+#define SZ_DMAST		1
+#define SZ_DMASTP		2
+#define SZ_DMASTZ		1
+#define SZ_DMAWFE		2
+#define SZ_DMAWFP		2
+#define SZ_DMAWMB		1
+#define SZ_DMAGO		6
+
+#define BRST_LEN(ccr)		((((ccr) >> CC_SRCBRSTLEN_SHFT) & 0xf) + 1)
+#define BRST_SIZE(ccr)		(1 << (((ccr) >> CC_SRCBRSTSIZE_SHFT) & 0x7))
+
+#define BYTE_TO_BURST(b, ccr)	((b) / BRST_SIZE(ccr) / BRST_LEN(ccr))
+#define BURST_TO_BYTE(c, ccr)	((c) * BRST_SIZE(ccr) * BRST_LEN(ccr))
+
+/*
+ * With 256 bytes, we can do more than 2.5MB and 5MB xfers per req
+ * at 1byte/burst for P<->M and M<->M respectively.
+ * For typical scenario, at 1word/burst, 10MB and 20MB xfers per req
+ * should be enough for P<->M and M<->M respectively.
+ */
+#define MCODE_BUFF_PER_REQ	256
+
+/* If the _pl330_req is available to the client */
+#define IS_FREE(req)	(*((u8 *)((req)->mc_cpu)) == CMD_DMAEND)
+
+/* Use this _only_ to wait on transient states */
+#define UNTIL(t, s)	while (!(_state(t) & (s))) cpu_relax();
+
+#ifdef PL330_DEBUG_MCGEN
+static unsigned cmd_line;
+#define PL330_DBGCMD_DUMP(off, x...)	do { \
+						printk("%x:", cmd_line); \
+						printk(x); \
+						cmd_line += off; \
+					} while (0)
+#define PL330_DBGMC_START(addr)		(cmd_line = addr)
+#else
+#define PL330_DBGCMD_DUMP(off, x...)	do {} while (0)
+#define PL330_DBGMC_START(addr)		do {} while (0)
+#endif
+
+/* The number of default descriptors */
+
 #define NR_DEFAULT_DESC	16
 
+/* Populated by the PL330 core driver for DMA API driver's info */
+struct pl330_config {
+	u32	periph_id;
+	u32	pcell_id;
+#define DMAC_MODE_NS	(1 << 0)
+	unsigned int	mode;
+	unsigned int	data_bus_width:10; /* In number of bits */
+	unsigned int	data_buf_dep:10;
+	unsigned int	num_chan:4;
+	unsigned int	num_peri:6;
+	u32		peri_ns;
+	unsigned int	num_events:6;
+	u32		irq_ns;
+};
+
+/* Handle to the DMAC provided to the PL330 core */
+struct pl330_info {
+	/* Owning device */
+	struct device *dev;
+	/* Size of MicroCode buffers for each channel. */
+	unsigned mcbufsz;
+	/* ioremap'ed address of PL330 registers. */
+	void __iomem	*base;
+	/* Client can freely use it. */
+	void	*client_data;
+	/* PL330 core data, Client must not touch it. */
+	void	*pl330_data;
+	/* Populated by the PL330 core driver during pl330_add */
+	struct pl330_config	pcfg;
+	/*
+	 * If the DMAC has some reset mechanism, then the
+	 * client may want to provide pointer to the method.
+	 */
+	void (*dmac_reset)(struct pl330_info *pi);
+};
+
+/**
+ * Request Configuration.
+ * The PL330 core does not modify this and uses the last
+ * working configuration if the request doesn't provide any.
+ *
+ * The Client may want to provide this info only for the
+ * first request and a request with new settings.
+ */
+struct pl330_reqcfg {
+	/* Address Incrementing */
+	unsigned dst_inc:1;
+	unsigned src_inc:1;
+
+	/*
+	 * For now, the SRC & DST protection levels
+	 * and burst size/length are assumed same.
+	 */
+	bool nonsecure;
+	bool privileged;
+	bool insnaccess;
+	unsigned brst_len:5;
+	unsigned brst_size:3; /* in power of 2 */
+
+	enum pl330_dstcachectrl dcctl;
+	enum pl330_srccachectrl scctl;
+	enum pl330_byteswap swap;
+	struct pl330_config *pcfg;
+};
+
+/*
+ * One cycle of DMAC operation.
+ * There may be more than one xfer in a request.
+ */
+struct pl330_xfer {
+	u32 src_addr;
+	u32 dst_addr;
+	/* Size to xfer */
+	u32 bytes;
+	/*
+	 * Pointer to next xfer in the list.
+	 * The last xfer in the req must point to NULL.
+	 */
+	struct pl330_xfer *next;
+};
+
+/* The xfer callbacks are made with one of these arguments. */
+enum pl330_op_err {
+	/* The all xfers in the request were success. */
+	PL330_ERR_NONE,
+	/* If req aborted due to global error. */
+	PL330_ERR_ABORT,
+	/* If req failed due to problem with Channel. */
+	PL330_ERR_FAIL,
+};
+
+/* A request defining Scatter-Gather List ending with NULL xfer. */
+struct pl330_req {
+	enum pl330_reqtype rqtype;
+	/* Index of peripheral for the xfer. */
+	unsigned peri:5;
+	/* Unique token for this xfer, set by the client. */
+	void *token;
+	/* Callback to be called after xfer. */
+	void (*xfer_cb)(void *token, enum pl330_op_err err);
+	/* If NULL, req will be done at last set parameters. */
+	struct pl330_reqcfg *cfg;
+	/* Pointer to first xfer in the request. */
+	struct pl330_xfer *x;
+};
+
+/*
+ * To know the status of the channel and DMAC, the client
+ * provides a pointer to this structure. The PL330 core
+ * fills it with current information.
+ */
+struct pl330_chanstatus {
+	/*
+	 * If the DMAC engine halted due to some error,
+	 * the client should remove-add DMAC.
+	 */
+	bool dmac_halted;
+	/*
+	 * If channel is halted due to some error,
+	 * the client should ABORT/FLUSH and START the channel.
+	 */
+	bool faulting;
+	/* Location of last load */
+	u32 src_addr;
+	/* Location of last store */
+	u32 dst_addr;
+	/*
+	 * Pointer to the currently active req, NULL if channel is
+	 * inactive, even though the requests may be present.
+	 */
+	struct pl330_req *top_req;
+	/* Pointer to req waiting second in the queue if any. */
+	struct pl330_req *wait_req;
+};
+
+enum pl330_chan_op {
+	/* Start the channel */
+	PL330_OP_START,
+	/* Abort the active xfer */
+	PL330_OP_ABORT,
+	/* Stop xfer and flush queue */
+	PL330_OP_FLUSH,
+};
+
+struct _xfer_spec {
+	u32 ccr;
+	struct pl330_req *r;
+	struct pl330_xfer *x;
+};
+
+enum dmamov_dst {
+	SAR = 0,
+	CCR,
+	DAR,
+};
+
+enum pl330_dst {
+	SRC = 0,
+	DST,
+};
+
+enum pl330_cond {
+	SINGLE,
+	BURST,
+	ALWAYS,
+};
+
+struct _pl330_req {
+	u32 mc_bus;
+	void *mc_cpu;
+	/* Number of bytes taken to setup MC for the req */
+	u32 mc_len;
+	struct pl330_req *r;
+	/* Hook to attach to DMAC's list of reqs with due callback */
+	struct list_head rqd;
+};
+
+/* ToBeDone for tasklet */
+struct _pl330_tbd {
+	bool reset_dmac;
+	bool reset_mngr;
+	u8 reset_chan;
+};
+
+/* A DMAC Thread */
+struct pl330_thread {
+	u8 id;
+	int ev;
+	/* If the channel is not yet acquired by any client */
+	bool free;
+	/* Parent DMAC */
+	struct pl330_dmac *dmac;
+	/* Only two at a time */
+	struct _pl330_req req[2];
+	/* Index of the last enqueued request */
+	unsigned lstenq;
+	/* Index of the last submitted request or -1 if the DMA is stopped */
+	int req_running;
+};
+
+enum pl330_dmac_state {
+	UNINIT,
+	INIT,
+	DYING,
+};
+
+/* A DMAC */
+struct pl330_dmac {
+	spinlock_t		lock;
+	/* Holds list of reqs with due callbacks */
+	struct list_head	req_done;
+	/* Pointer to platform specific stuff */
+	struct pl330_info	*pinfo;
+	/* Maximum possible events/irqs */
+	int			events[32];
+	/* BUS address of MicroCode buffer */
+	u32			mcode_bus;
+	/* CPU address of MicroCode buffer */
+	void			*mcode_cpu;
+	/* List of all Channel threads */
+	struct pl330_thread	*channels;
+	/* Pointer to the MANAGER thread */
+	struct pl330_thread	*manager;
+	/* To handle bad news in interrupt */
+	struct tasklet_struct	tasks;
+	struct _pl330_tbd	dmac_tbd;
+	/* State of DMAC operation */
+	enum pl330_dmac_state	state;
+};
+
 enum desc_status {
 	/* In the DMAC pool */
 	FREE,
@@ -51,9 +547,6 @@ struct dma_pl330_chan {
 	/* DMA-Engine Channel */
 	struct dma_chan chan;
 
-	/* Last completed cookie */
-	dma_cookie_t completed;
-
 	/* List of to be xfered descriptors */
 	struct list_head work_list;
 
@@ -117,6 +610,1599 @@ struct dma_pl330_desc {
 	struct dma_pl330_chan *pchan;
 };
 
+static inline void _callback(struct pl330_req *r, enum pl330_op_err err)
+{
+	if (r && r->xfer_cb)
+		r->xfer_cb(r->token, err);
+}
+
+static inline bool _queue_empty(struct pl330_thread *thrd)
+{
+	return (IS_FREE(&thrd->req[0]) && IS_FREE(&thrd->req[1]))
+		? true : false;
+}
+
+static inline bool _queue_full(struct pl330_thread *thrd)
+{
+	return (IS_FREE(&thrd->req[0]) || IS_FREE(&thrd->req[1]))
+		? false : true;
+}
+
+static inline bool is_manager(struct pl330_thread *thrd)
+{
+	struct pl330_dmac *pl330 = thrd->dmac;
+
+	/* MANAGER is indexed at the end */
+	if (thrd->id == pl330->pinfo->pcfg.num_chan)
+		return true;
+	else
+		return false;
+}
+
+/* If manager of the thread is in Non-Secure mode */
+static inline bool _manager_ns(struct pl330_thread *thrd)
+{
+	struct pl330_dmac *pl330 = thrd->dmac;
+
+	return (pl330->pinfo->pcfg.mode & DMAC_MODE_NS) ? true : false;
+}
+
+static inline u32 get_id(struct pl330_info *pi, u32 off)
+{
+	void __iomem *regs = pi->base;
+	u32 id = 0;
+
+	id |= (readb(regs + off + 0x0) << 0);
+	id |= (readb(regs + off + 0x4) << 8);
+	id |= (readb(regs + off + 0x8) << 16);
+	id |= (readb(regs + off + 0xc) << 24);
+
+	return id;
+}
+
+static inline u32 get_revision(u32 periph_id)
+{
+	return (periph_id >> PERIPH_REV_SHIFT) & PERIPH_REV_MASK;
+}
+
+static inline u32 _emit_ADDH(unsigned dry_run, u8 buf[],
+		enum pl330_dst da, u16 val)
+{
+	if (dry_run)
+		return SZ_DMAADDH;
+
+	buf[0] = CMD_DMAADDH;
+	buf[0] |= (da << 1);
+	*((u16 *)&buf[1]) = val;
+
+	PL330_DBGCMD_DUMP(SZ_DMAADDH, "\tDMAADDH %s %u\n",
+		da == 1 ? "DA" : "SA", val);
+
+	return SZ_DMAADDH;
+}
+
+static inline u32 _emit_END(unsigned dry_run, u8 buf[])
+{
+	if (dry_run)
+		return SZ_DMAEND;
+
+	buf[0] = CMD_DMAEND;
+
+	PL330_DBGCMD_DUMP(SZ_DMAEND, "\tDMAEND\n");
+
+	return SZ_DMAEND;
+}
+
+static inline u32 _emit_FLUSHP(unsigned dry_run, u8 buf[], u8 peri)
+{
+	if (dry_run)
+		return SZ_DMAFLUSHP;
+
+	buf[0] = CMD_DMAFLUSHP;
+
+	peri &= 0x1f;
+	peri <<= 3;
+	buf[1] = peri;
+
+	PL330_DBGCMD_DUMP(SZ_DMAFLUSHP, "\tDMAFLUSHP %u\n", peri >> 3);
+
+	return SZ_DMAFLUSHP;
+}
+
+static inline u32 _emit_LD(unsigned dry_run, u8 buf[],	enum pl330_cond cond)
+{
+	if (dry_run)
+		return SZ_DMALD;
+
+	buf[0] = CMD_DMALD;
+
+	if (cond == SINGLE)
+		buf[0] |= (0 << 1) | (1 << 0);
+	else if (cond == BURST)
+		buf[0] |= (1 << 1) | (1 << 0);
+
+	PL330_DBGCMD_DUMP(SZ_DMALD, "\tDMALD%c\n",
+		cond == SINGLE ? 'S' : (cond == BURST ? 'B' : 'A'));
+
+	return SZ_DMALD;
+}
+
+static inline u32 _emit_LDP(unsigned dry_run, u8 buf[],
+		enum pl330_cond cond, u8 peri)
+{
+	if (dry_run)
+		return SZ_DMALDP;
+
+	buf[0] = CMD_DMALDP;
+
+	if (cond == BURST)
+		buf[0] |= (1 << 1);
+
+	peri &= 0x1f;
+	peri <<= 3;
+	buf[1] = peri;
+
+	PL330_DBGCMD_DUMP(SZ_DMALDP, "\tDMALDP%c %u\n",
+		cond == SINGLE ? 'S' : 'B', peri >> 3);
+
+	return SZ_DMALDP;
+}
+
+static inline u32 _emit_LP(unsigned dry_run, u8 buf[],
+		unsigned loop, u8 cnt)
+{
+	if (dry_run)
+		return SZ_DMALP;
+
+	buf[0] = CMD_DMALP;
+
+	if (loop)
+		buf[0] |= (1 << 1);
+
+	cnt--; /* DMAC increments by 1 internally */
+	buf[1] = cnt;
+
+	PL330_DBGCMD_DUMP(SZ_DMALP, "\tDMALP_%c %u\n", loop ? '1' : '0', cnt);
+
+	return SZ_DMALP;
+}
+
+struct _arg_LPEND {
+	enum pl330_cond cond;
+	bool forever;
+	unsigned loop;
+	u8 bjump;
+};
+
+static inline u32 _emit_LPEND(unsigned dry_run, u8 buf[],
+		const struct _arg_LPEND *arg)
+{
+	enum pl330_cond cond = arg->cond;
+	bool forever = arg->forever;
+	unsigned loop = arg->loop;
+	u8 bjump = arg->bjump;
+
+	if (dry_run)
+		return SZ_DMALPEND;
+
+	buf[0] = CMD_DMALPEND;
+
+	if (loop)
+		buf[0] |= (1 << 2);
+
+	if (!forever)
+		buf[0] |= (1 << 4);
+
+	if (cond == SINGLE)
+		buf[0] |= (0 << 1) | (1 << 0);
+	else if (cond == BURST)
+		buf[0] |= (1 << 1) | (1 << 0);
+
+	buf[1] = bjump;
+
+	PL330_DBGCMD_DUMP(SZ_DMALPEND, "\tDMALP%s%c_%c bjmpto_%x\n",
+			forever ? "FE" : "END",
+			cond == SINGLE ? 'S' : (cond == BURST ? 'B' : 'A'),
+			loop ? '1' : '0',
+			bjump);
+
+	return SZ_DMALPEND;
+}
+
+static inline u32 _emit_KILL(unsigned dry_run, u8 buf[])
+{
+	if (dry_run)
+		return SZ_DMAKILL;
+
+	buf[0] = CMD_DMAKILL;
+
+	return SZ_DMAKILL;
+}
+
+static inline u32 _emit_MOV(unsigned dry_run, u8 buf[],
+		enum dmamov_dst dst, u32 val)
+{
+	if (dry_run)
+		return SZ_DMAMOV;
+
+	buf[0] = CMD_DMAMOV;
+	buf[1] = dst;
+	*((u32 *)&buf[2]) = val;
+
+	PL330_DBGCMD_DUMP(SZ_DMAMOV, "\tDMAMOV %s 0x%x\n",
+		dst == SAR ? "SAR" : (dst == DAR ? "DAR" : "CCR"), val);
+
+	return SZ_DMAMOV;
+}
+
+static inline u32 _emit_NOP(unsigned dry_run, u8 buf[])
+{
+	if (dry_run)
+		return SZ_DMANOP;
+
+	buf[0] = CMD_DMANOP;
+
+	PL330_DBGCMD_DUMP(SZ_DMANOP, "\tDMANOP\n");
+
+	return SZ_DMANOP;
+}
+
+static inline u32 _emit_RMB(unsigned dry_run, u8 buf[])
+{
+	if (dry_run)
+		return SZ_DMARMB;
+
+	buf[0] = CMD_DMARMB;
+
+	PL330_DBGCMD_DUMP(SZ_DMARMB, "\tDMARMB\n");
+
+	return SZ_DMARMB;
+}
+
+static inline u32 _emit_SEV(unsigned dry_run, u8 buf[], u8 ev)
+{
+	if (dry_run)
+		return SZ_DMASEV;
+
+	buf[0] = CMD_DMASEV;
+
+	ev &= 0x1f;
+	ev <<= 3;
+	buf[1] = ev;
+
+	PL330_DBGCMD_DUMP(SZ_DMASEV, "\tDMASEV %u\n", ev >> 3);
+
+	return SZ_DMASEV;
+}
+
+static inline u32 _emit_ST(unsigned dry_run, u8 buf[], enum pl330_cond cond)
+{
+	if (dry_run)
+		return SZ_DMAST;
+
+	buf[0] = CMD_DMAST;
+
+	if (cond == SINGLE)
+		buf[0] |= (0 << 1) | (1 << 0);
+	else if (cond == BURST)
+		buf[0] |= (1 << 1) | (1 << 0);
+
+	PL330_DBGCMD_DUMP(SZ_DMAST, "\tDMAST%c\n",
+		cond == SINGLE ? 'S' : (cond == BURST ? 'B' : 'A'));
+
+	return SZ_DMAST;
+}
+
+static inline u32 _emit_STP(unsigned dry_run, u8 buf[],
+		enum pl330_cond cond, u8 peri)
+{
+	if (dry_run)
+		return SZ_DMASTP;
+
+	buf[0] = CMD_DMASTP;
+
+	if (cond == BURST)
+		buf[0] |= (1 << 1);
+
+	peri &= 0x1f;
+	peri <<= 3;
+	buf[1] = peri;
+
+	PL330_DBGCMD_DUMP(SZ_DMASTP, "\tDMASTP%c %u\n",
+		cond == SINGLE ? 'S' : 'B', peri >> 3);
+
+	return SZ_DMASTP;
+}
+
+static inline u32 _emit_STZ(unsigned dry_run, u8 buf[])
+{
+	if (dry_run)
+		return SZ_DMASTZ;
+
+	buf[0] = CMD_DMASTZ;
+
+	PL330_DBGCMD_DUMP(SZ_DMASTZ, "\tDMASTZ\n");
+
+	return SZ_DMASTZ;
+}
+
+static inline u32 _emit_WFE(unsigned dry_run, u8 buf[], u8 ev,
+		unsigned invalidate)
+{
+	if (dry_run)
+		return SZ_DMAWFE;
+
+	buf[0] = CMD_DMAWFE;
+
+	ev &= 0x1f;
+	ev <<= 3;
+	buf[1] = ev;
+
+	if (invalidate)
+		buf[1] |= (1 << 1);
+
+	PL330_DBGCMD_DUMP(SZ_DMAWFE, "\tDMAWFE %u%s\n",
+		ev >> 3, invalidate ? ", I" : "");
+
+	return SZ_DMAWFE;
+}
+
+static inline u32 _emit_WFP(unsigned dry_run, u8 buf[],
+		enum pl330_cond cond, u8 peri)
+{
+	if (dry_run)
+		return SZ_DMAWFP;
+
+	buf[0] = CMD_DMAWFP;
+
+	if (cond == SINGLE)
+		buf[0] |= (0 << 1) | (0 << 0);
+	else if (cond == BURST)
+		buf[0] |= (1 << 1) | (0 << 0);
+	else
+		buf[0] |= (0 << 1) | (1 << 0);
+
+	peri &= 0x1f;
+	peri <<= 3;
+	buf[1] = peri;
+
+	PL330_DBGCMD_DUMP(SZ_DMAWFP, "\tDMAWFP%c %u\n",
+		cond == SINGLE ? 'S' : (cond == BURST ? 'B' : 'P'), peri >> 3);
+
+	return SZ_DMAWFP;
+}
+
+static inline u32 _emit_WMB(unsigned dry_run, u8 buf[])
+{
+	if (dry_run)
+		return SZ_DMAWMB;
+
+	buf[0] = CMD_DMAWMB;
+
+	PL330_DBGCMD_DUMP(SZ_DMAWMB, "\tDMAWMB\n");
+
+	return SZ_DMAWMB;
+}
+
+struct _arg_GO {
+	u8 chan;
+	u32 addr;
+	unsigned ns;
+};
+
+static inline u32 _emit_GO(unsigned dry_run, u8 buf[],
+		const struct _arg_GO *arg)
+{
+	u8 chan = arg->chan;
+	u32 addr = arg->addr;
+	unsigned ns = arg->ns;
+
+	if (dry_run)
+		return SZ_DMAGO;
+
+	buf[0] = CMD_DMAGO;
+	buf[0] |= (ns << 1);
+
+	buf[1] = chan & 0x7;
+
+	*((u32 *)&buf[2]) = addr;
+
+	return SZ_DMAGO;
+}
+
+#define msecs_to_loops(t) (loops_per_jiffy / 1000 * HZ * t)
+
+/* Returns Time-Out */
+static bool _until_dmac_idle(struct pl330_thread *thrd)
+{
+	void __iomem *regs = thrd->dmac->pinfo->base;
+	unsigned long loops = msecs_to_loops(5);
+
+	do {
+		/* Until Manager is Idle */
+		if (!(readl(regs + DBGSTATUS) & DBG_BUSY))
+			break;
+
+		cpu_relax();
+	} while (--loops);
+
+	if (!loops)
+		return true;
+
+	return false;
+}
+
+static inline void _execute_DBGINSN(struct pl330_thread *thrd,
+		u8 insn[], bool as_manager)
+{
+	void __iomem *regs = thrd->dmac->pinfo->base;
+	u32 val;
+
+	val = (insn[0] << 16) | (insn[1] << 24);
+	if (!as_manager) {
+		val |= (1 << 0);
+		val |= (thrd->id << 8); /* Channel Number */
+	}
+	writel(val, regs + DBGINST0);
+
+	val = *((u32 *)&insn[2]);
+	writel(val, regs + DBGINST1);
+
+	/* If timed out due to halted state-machine */
+	if (_until_dmac_idle(thrd)) {
+		dev_err(thrd->dmac->pinfo->dev, "DMAC halted!\n");
+		return;
+	}
+
+	/* Get going */
+	writel(0, regs + DBGCMD);
+}
+
+/*
+ * Mark a _pl330_req as free.
+ * We do it by writing DMAEND as the first instruction
+ * because no valid request is going to have DMAEND as
+ * its first instruction to execute.
+ */
+static void mark_free(struct pl330_thread *thrd, int idx)
+{
+	struct _pl330_req *req = &thrd->req[idx];
+
+	_emit_END(0, req->mc_cpu);
+	req->mc_len = 0;
+
+	thrd->req_running = -1;
+}
+
+static inline u32 _state(struct pl330_thread *thrd)
+{
+	void __iomem *regs = thrd->dmac->pinfo->base;
+	u32 val;
+
+	if (is_manager(thrd))
+		val = readl(regs + DS) & 0xf;
+	else
+		val = readl(regs + CS(thrd->id)) & 0xf;
+
+	switch (val) {
+	case DS_ST_STOP:
+		return PL330_STATE_STOPPED;
+	case DS_ST_EXEC:
+		return PL330_STATE_EXECUTING;
+	case DS_ST_CMISS:
+		return PL330_STATE_CACHEMISS;
+	case DS_ST_UPDTPC:
+		return PL330_STATE_UPDTPC;
+	case DS_ST_WFE:
+		return PL330_STATE_WFE;
+	case DS_ST_FAULT:
+		return PL330_STATE_FAULTING;
+	case DS_ST_ATBRR:
+		if (is_manager(thrd))
+			return PL330_STATE_INVALID;
+		else
+			return PL330_STATE_ATBARRIER;
+	case DS_ST_QBUSY:
+		if (is_manager(thrd))
+			return PL330_STATE_INVALID;
+		else
+			return PL330_STATE_QUEUEBUSY;
+	case DS_ST_WFP:
+		if (is_manager(thrd))
+			return PL330_STATE_INVALID;
+		else
+			return PL330_STATE_WFP;
+	case DS_ST_KILL:
+		if (is_manager(thrd))
+			return PL330_STATE_INVALID;
+		else
+			return PL330_STATE_KILLING;
+	case DS_ST_CMPLT:
+		if (is_manager(thrd))
+			return PL330_STATE_INVALID;
+		else
+			return PL330_STATE_COMPLETING;
+	case DS_ST_FLTCMP:
+		if (is_manager(thrd))
+			return PL330_STATE_INVALID;
+		else
+			return PL330_STATE_FAULT_COMPLETING;
+	default:
+		return PL330_STATE_INVALID;
+	}
+}
+
+static void _stop(struct pl330_thread *thrd)
+{
+	void __iomem *regs = thrd->dmac->pinfo->base;
+	u8 insn[6] = {0, 0, 0, 0, 0, 0};
+
+	if (_state(thrd) == PL330_STATE_FAULT_COMPLETING)
+		UNTIL(thrd, PL330_STATE_FAULTING | PL330_STATE_KILLING);
+
+	/* Return if nothing needs to be done */
+	if (_state(thrd) == PL330_STATE_COMPLETING
+		  || _state(thrd) == PL330_STATE_KILLING
+		  || _state(thrd) == PL330_STATE_STOPPED)
+		return;
+
+	_emit_KILL(0, insn);
+
+	/* Stop generating interrupts for SEV */
+	writel(readl(regs + INTEN) & ~(1 << thrd->ev), regs + INTEN);
+
+	_execute_DBGINSN(thrd, insn, is_manager(thrd));
+}
+
+/* Start doing req 'idx' of thread 'thrd' */
+static bool _trigger(struct pl330_thread *thrd)
+{
+	void __iomem *regs = thrd->dmac->pinfo->base;
+	struct _pl330_req *req;
+	struct pl330_req *r;
+	struct _arg_GO go;
+	unsigned ns;
+	u8 insn[6] = {0, 0, 0, 0, 0, 0};
+	int idx;
+
+	/* Return if already ACTIVE */
+	if (_state(thrd) != PL330_STATE_STOPPED)
+		return true;
+
+	idx = 1 - thrd->lstenq;
+	if (!IS_FREE(&thrd->req[idx]))
+		req = &thrd->req[idx];
+	else {
+		idx = thrd->lstenq;
+		if (!IS_FREE(&thrd->req[idx]))
+			req = &thrd->req[idx];
+		else
+			req = NULL;
+	}
+
+	/* Return if no request */
+	if (!req || !req->r)
+		return true;
+
+	r = req->r;
+
+	if (r->cfg)
+		ns = r->cfg->nonsecure ? 1 : 0;
+	else if (readl(regs + CS(thrd->id)) & CS_CNS)
+		ns = 1;
+	else
+		ns = 0;
+
+	/* See 'Abort Sources' point-4 at Page 2-25 */
+	if (_manager_ns(thrd) && !ns)
+		dev_info(thrd->dmac->pinfo->dev, "%s:%d Recipe for ABORT!\n",
+			__func__, __LINE__);
+
+	go.chan = thrd->id;
+	go.addr = req->mc_bus;
+	go.ns = ns;
+	_emit_GO(0, insn, &go);
+
+	/* Set to generate interrupts for SEV */
+	writel(readl(regs + INTEN) | (1 << thrd->ev), regs + INTEN);
+
+	/* Only manager can execute GO */
+	_execute_DBGINSN(thrd, insn, true);
+
+	thrd->req_running = idx;
+
+	return true;
+}
+
+static bool _start(struct pl330_thread *thrd)
+{
+	switch (_state(thrd)) {
+	case PL330_STATE_FAULT_COMPLETING:
+		UNTIL(thrd, PL330_STATE_FAULTING | PL330_STATE_KILLING);
+
+		if (_state(thrd) == PL330_STATE_KILLING)
+			UNTIL(thrd, PL330_STATE_STOPPED)
+
+	case PL330_STATE_FAULTING:
+		_stop(thrd);
+
+	case PL330_STATE_KILLING:
+	case PL330_STATE_COMPLETING:
+		UNTIL(thrd, PL330_STATE_STOPPED)
+
+	case PL330_STATE_STOPPED:
+		return _trigger(thrd);
+
+	case PL330_STATE_WFP:
+	case PL330_STATE_QUEUEBUSY:
+	case PL330_STATE_ATBARRIER:
+	case PL330_STATE_UPDTPC:
+	case PL330_STATE_CACHEMISS:
+	case PL330_STATE_EXECUTING:
+		return true;
+
+	case PL330_STATE_WFE: /* For RESUME, nothing yet */
+	default:
+		return false;
+	}
+}
+
+static inline int _ldst_memtomem(unsigned dry_run, u8 buf[],
+		const struct _xfer_spec *pxs, int cyc)
+{
+	int off = 0;
+	struct pl330_config *pcfg = pxs->r->cfg->pcfg;
+
+	/* check lock-up free version */
+	if (get_revision(pcfg->periph_id) >= PERIPH_REV_R1P0) {
+		while (cyc--) {
+			off += _emit_LD(dry_run, &buf[off], ALWAYS);
+			off += _emit_ST(dry_run, &buf[off], ALWAYS);
+		}
+	} else {
+		while (cyc--) {
+			off += _emit_LD(dry_run, &buf[off], ALWAYS);
+			off += _emit_RMB(dry_run, &buf[off]);
+			off += _emit_ST(dry_run, &buf[off], ALWAYS);
+			off += _emit_WMB(dry_run, &buf[off]);
+		}
+	}
+
+	return off;
+}
+
+static inline int _ldst_devtomem(unsigned dry_run, u8 buf[],
+		const struct _xfer_spec *pxs, int cyc)
+{
+	int off = 0;
+
+	while (cyc--) {
+		off += _emit_WFP(dry_run, &buf[off], SINGLE, pxs->r->peri);
+		off += _emit_LDP(dry_run, &buf[off], SINGLE, pxs->r->peri);
+		off += _emit_ST(dry_run, &buf[off], ALWAYS);
+		off += _emit_FLUSHP(dry_run, &buf[off], pxs->r->peri);
+	}
+
+	return off;
+}
+
+static inline int _ldst_memtodev(unsigned dry_run, u8 buf[],
+		const struct _xfer_spec *pxs, int cyc)
+{
+	int off = 0;
+
+	while (cyc--) {
+		off += _emit_WFP(dry_run, &buf[off], SINGLE, pxs->r->peri);
+		off += _emit_LD(dry_run, &buf[off], ALWAYS);
+		off += _emit_STP(dry_run, &buf[off], SINGLE, pxs->r->peri);
+		off += _emit_FLUSHP(dry_run, &buf[off], pxs->r->peri);
+	}
+
+	return off;
+}
+
+static int _bursts(unsigned dry_run, u8 buf[],
+		const struct _xfer_spec *pxs, int cyc)
+{
+	int off = 0;
+
+	switch (pxs->r->rqtype) {
+	case MEMTODEV:
+		off += _ldst_memtodev(dry_run, &buf[off], pxs, cyc);
+		break;
+	case DEVTOMEM:
+		off += _ldst_devtomem(dry_run, &buf[off], pxs, cyc);
+		break;
+	case MEMTOMEM:
+		off += _ldst_memtomem(dry_run, &buf[off], pxs, cyc);
+		break;
+	default:
+		off += 0x40000000; /* Scare off the Client */
+		break;
+	}
+
+	return off;
+}
+
+/* Returns bytes consumed and updates bursts */
+static inline int _loop(unsigned dry_run, u8 buf[],
+		unsigned long *bursts, const struct _xfer_spec *pxs)
+{
+	int cyc, cycmax, szlp, szlpend, szbrst, off;
+	unsigned lcnt0, lcnt1, ljmp0, ljmp1;
+	struct _arg_LPEND lpend;
+
+	/* Max iterations possible in DMALP is 256 */
+	if (*bursts >= 256*256) {
+		lcnt1 = 256;
+		lcnt0 = 256;
+		cyc = *bursts / lcnt1 / lcnt0;
+	} else if (*bursts > 256) {
+		lcnt1 = 256;
+		lcnt0 = *bursts / lcnt1;
+		cyc = 1;
+	} else {
+		lcnt1 = *bursts;
+		lcnt0 = 0;
+		cyc = 1;
+	}
+
+	szlp = _emit_LP(1, buf, 0, 0);
+	szbrst = _bursts(1, buf, pxs, 1);
+
+	lpend.cond = ALWAYS;
+	lpend.forever = false;
+	lpend.loop = 0;
+	lpend.bjump = 0;
+	szlpend = _emit_LPEND(1, buf, &lpend);
+
+	if (lcnt0) {
+		szlp *= 2;
+		szlpend *= 2;
+	}
+
+	/*
+	 * Max bursts that we can unroll due to limit on the
+	 * size of backward jump that can be encoded in DMALPEND
+	 * which is 8-bits and hence 255
+	 */
+	cycmax = (255 - (szlp + szlpend)) / szbrst;
+
+	cyc = (cycmax < cyc) ? cycmax : cyc;
+
+	off = 0;
+
+	if (lcnt0) {
+		off += _emit_LP(dry_run, &buf[off], 0, lcnt0);
+		ljmp0 = off;
+	}
+
+	off += _emit_LP(dry_run, &buf[off], 1, lcnt1);
+	ljmp1 = off;
+
+	off += _bursts(dry_run, &buf[off], pxs, cyc);
+
+	lpend.cond = ALWAYS;
+	lpend.forever = false;
+	lpend.loop = 1;
+	lpend.bjump = off - ljmp1;
+	off += _emit_LPEND(dry_run, &buf[off], &lpend);
+
+	if (lcnt0) {
+		lpend.cond = ALWAYS;
+		lpend.forever = false;
+		lpend.loop = 0;
+		lpend.bjump = off - ljmp0;
+		off += _emit_LPEND(dry_run, &buf[off], &lpend);
+	}
+
+	*bursts = lcnt1 * cyc;
+	if (lcnt0)
+		*bursts *= lcnt0;
+
+	return off;
+}
+
+static inline int _setup_loops(unsigned dry_run, u8 buf[],
+		const struct _xfer_spec *pxs)
+{
+	struct pl330_xfer *x = pxs->x;
+	u32 ccr = pxs->ccr;
+	unsigned long c, bursts = BYTE_TO_BURST(x->bytes, ccr);
+	int off = 0;
+
+	while (bursts) {
+		c = bursts;
+		off += _loop(dry_run, &buf[off], &c, pxs);
+		bursts -= c;
+	}
+
+	return off;
+}
+
+static inline int _setup_xfer(unsigned dry_run, u8 buf[],
+		const struct _xfer_spec *pxs)
+{
+	struct pl330_xfer *x = pxs->x;
+	int off = 0;
+
+	/* DMAMOV SAR, x->src_addr */
+	off += _emit_MOV(dry_run, &buf[off], SAR, x->src_addr);
+	/* DMAMOV DAR, x->dst_addr */
+	off += _emit_MOV(dry_run, &buf[off], DAR, x->dst_addr);
+
+	/* Setup Loop(s) */
+	off += _setup_loops(dry_run, &buf[off], pxs);
+
+	return off;
+}
+
+/*
+ * A req is a sequence of one or more xfer units.
+ * Returns the number of bytes taken to setup the MC for the req.
+ */
+static int _setup_req(unsigned dry_run, struct pl330_thread *thrd,
+		unsigned index, struct _xfer_spec *pxs)
+{
+	struct _pl330_req *req = &thrd->req[index];
+	struct pl330_xfer *x;
+	u8 *buf = req->mc_cpu;
+	int off = 0;
+
+	PL330_DBGMC_START(req->mc_bus);
+
+	/* DMAMOV CCR, ccr */
+	off += _emit_MOV(dry_run, &buf[off], CCR, pxs->ccr);
+
+	x = pxs->r->x;
+	do {
+		/* Error if xfer length is not aligned at burst size */
+		if (x->bytes % (BRST_SIZE(pxs->ccr) * BRST_LEN(pxs->ccr)))
+			return -EINVAL;
+
+		pxs->x = x;
+		off += _setup_xfer(dry_run, &buf[off], pxs);
+
+		x = x->next;
+	} while (x);
+
+	/* DMASEV peripheral/event */
+	off += _emit_SEV(dry_run, &buf[off], thrd->ev);
+	/* DMAEND */
+	off += _emit_END(dry_run, &buf[off]);
+
+	return off;
+}
+
+static inline u32 _prepare_ccr(const struct pl330_reqcfg *rqc)
+{
+	u32 ccr = 0;
+
+	if (rqc->src_inc)
+		ccr |= CC_SRCINC;
+
+	if (rqc->dst_inc)
+		ccr |= CC_DSTINC;
+
+	/* We set same protection levels for Src and DST for now */
+	if (rqc->privileged)
+		ccr |= CC_SRCPRI | CC_DSTPRI;
+	if (rqc->nonsecure)
+		ccr |= CC_SRCNS | CC_DSTNS;
+	if (rqc->insnaccess)
+		ccr |= CC_SRCIA | CC_DSTIA;
+
+	ccr |= (((rqc->brst_len - 1) & 0xf) << CC_SRCBRSTLEN_SHFT);
+	ccr |= (((rqc->brst_len - 1) & 0xf) << CC_DSTBRSTLEN_SHFT);
+
+	ccr |= (rqc->brst_size << CC_SRCBRSTSIZE_SHFT);
+	ccr |= (rqc->brst_size << CC_DSTBRSTSIZE_SHFT);
+
+	ccr |= (rqc->scctl << CC_SRCCCTRL_SHFT);
+	ccr |= (rqc->dcctl << CC_DSTCCTRL_SHFT);
+
+	ccr |= (rqc->swap << CC_SWAP_SHFT);
+
+	return ccr;
+}
+
+static inline bool _is_valid(u32 ccr)
+{
+	enum pl330_dstcachectrl dcctl;
+	enum pl330_srccachectrl scctl;
+
+	dcctl = (ccr >> CC_DSTCCTRL_SHFT) & CC_DRCCCTRL_MASK;
+	scctl = (ccr >> CC_SRCCCTRL_SHFT) & CC_SRCCCTRL_MASK;
+
+	if (dcctl == DINVALID1 || dcctl == DINVALID2
+			|| scctl == SINVALID1 || scctl == SINVALID2)
+		return false;
+	else
+		return true;
+}
+
+/*
+ * Submit a list of xfers after which the client wants notification.
+ * Client is not notified after each xfer unit, just once after all
+ * xfer units are done or some error occurs.
+ */
+static int pl330_submit_req(void *ch_id, struct pl330_req *r)
+{
+	struct pl330_thread *thrd = ch_id;
+	struct pl330_dmac *pl330;
+	struct pl330_info *pi;
+	struct _xfer_spec xs;
+	unsigned long flags;
+	void __iomem *regs;
+	unsigned idx;
+	u32 ccr;
+	int ret = 0;
+
+	/* No Req or Unacquired Channel or DMAC */
+	if (!r || !thrd || thrd->free)
+		return -EINVAL;
+
+	pl330 = thrd->dmac;
+	pi = pl330->pinfo;
+	regs = pi->base;
+
+	if (pl330->state == DYING
+		|| pl330->dmac_tbd.reset_chan & (1 << thrd->id)) {
+		dev_info(thrd->dmac->pinfo->dev, "%s:%d\n",
+			__func__, __LINE__);
+		return -EAGAIN;
+	}
+
+	/* If request for non-existing peripheral */
+	if (r->rqtype != MEMTOMEM && r->peri >= pi->pcfg.num_peri) {
+		dev_info(thrd->dmac->pinfo->dev,
+				"%s:%d Invalid peripheral(%u)!\n",
+				__func__, __LINE__, r->peri);
+		return -EINVAL;
+	}
+
+	spin_lock_irqsave(&pl330->lock, flags);
+
+	if (_queue_full(thrd)) {
+		ret = -EAGAIN;
+		goto xfer_exit;
+	}
+
+	/* Prefer Secure Channel */
+	if (!_manager_ns(thrd))
+		r->cfg->nonsecure = 0;
+	else
+		r->cfg->nonsecure = 1;
+
+	/* Use last settings, if not provided */
+	if (r->cfg)
+		ccr = _prepare_ccr(r->cfg);
+	else
+		ccr = readl(regs + CC(thrd->id));
+
+	/* If this req doesn't have valid xfer settings */
+	if (!_is_valid(ccr)) {
+		ret = -EINVAL;
+		dev_info(thrd->dmac->pinfo->dev, "%s:%d Invalid CCR(%x)!\n",
+			__func__, __LINE__, ccr);
+		goto xfer_exit;
+	}
+
+	idx = IS_FREE(&thrd->req[0]) ? 0 : 1;
+
+	xs.ccr = ccr;
+	xs.r = r;
+
+	/* First dry run to check if req is acceptable */
+	ret = _setup_req(1, thrd, idx, &xs);
+	if (ret < 0)
+		goto xfer_exit;
+
+	if (ret > pi->mcbufsz / 2) {
+		dev_info(thrd->dmac->pinfo->dev,
+			"%s:%d Trying increasing mcbufsz\n",
+				__func__, __LINE__);
+		ret = -ENOMEM;
+		goto xfer_exit;
+	}
+
+	/* Hook the request */
+	thrd->lstenq = idx;
+	thrd->req[idx].mc_len = _setup_req(0, thrd, idx, &xs);
+	thrd->req[idx].r = r;
+
+	ret = 0;
+
+xfer_exit:
+	spin_unlock_irqrestore(&pl330->lock, flags);
+
+	return ret;
+}
+
+static void pl330_dotask(unsigned long data)
+{
+	struct pl330_dmac *pl330 = (struct pl330_dmac *) data;
+	struct pl330_info *pi = pl330->pinfo;
+	unsigned long flags;
+	int i;
+
+	spin_lock_irqsave(&pl330->lock, flags);
+
+	/* The DMAC itself gone nuts */
+	if (pl330->dmac_tbd.reset_dmac) {
+		pl330->state = DYING;
+		/* Reset the manager too */
+		pl330->dmac_tbd.reset_mngr = true;
+		/* Clear the reset flag */
+		pl330->dmac_tbd.reset_dmac = false;
+	}
+
+	if (pl330->dmac_tbd.reset_mngr) {
+		_stop(pl330->manager);
+		/* Reset all channels */
+		pl330->dmac_tbd.reset_chan = (1 << pi->pcfg.num_chan) - 1;
+		/* Clear the reset flag */
+		pl330->dmac_tbd.reset_mngr = false;
+	}
+
+	for (i = 0; i < pi->pcfg.num_chan; i++) {
+
+		if (pl330->dmac_tbd.reset_chan & (1 << i)) {
+			struct pl330_thread *thrd = &pl330->channels[i];
+			void __iomem *regs = pi->base;
+			enum pl330_op_err err;
+
+			_stop(thrd);
+
+			if (readl(regs + FSC) & (1 << thrd->id))
+				err = PL330_ERR_FAIL;
+			else
+				err = PL330_ERR_ABORT;
+
+			spin_unlock_irqrestore(&pl330->lock, flags);
+
+			_callback(thrd->req[1 - thrd->lstenq].r, err);
+			_callback(thrd->req[thrd->lstenq].r, err);
+
+			spin_lock_irqsave(&pl330->lock, flags);
+
+			thrd->req[0].r = NULL;
+			thrd->req[1].r = NULL;
+			mark_free(thrd, 0);
+			mark_free(thrd, 1);
+
+			/* Clear the reset flag */
+			pl330->dmac_tbd.reset_chan &= ~(1 << i);
+		}
+	}
+
+	spin_unlock_irqrestore(&pl330->lock, flags);
+
+	return;
+}
+
+/* Returns 1 if state was updated, 0 otherwise */
+static int pl330_update(const struct pl330_info *pi)
+{
+	struct _pl330_req *rqdone;
+	struct pl330_dmac *pl330;
+	unsigned long flags;
+	void __iomem *regs;
+	u32 val;
+	int id, ev, ret = 0;
+
+	if (!pi || !pi->pl330_data)
+		return 0;
+
+	regs = pi->base;
+	pl330 = pi->pl330_data;
+
+	spin_lock_irqsave(&pl330->lock, flags);
+
+	val = readl(regs + FSM) & 0x1;
+	if (val)
+		pl330->dmac_tbd.reset_mngr = true;
+	else
+		pl330->dmac_tbd.reset_mngr = false;
+
+	val = readl(regs + FSC) & ((1 << pi->pcfg.num_chan) - 1);
+	pl330->dmac_tbd.reset_chan |= val;
+	if (val) {
+		int i = 0;
+		while (i < pi->pcfg.num_chan) {
+			if (val & (1 << i)) {
+				dev_info(pi->dev,
+					"Reset Channel-%d\t CS-%x FTC-%x\n",
+						i, readl(regs + CS(i)),
+						readl(regs + FTC(i)));
+				_stop(&pl330->channels[i]);
+			}
+			i++;
+		}
+	}
+
+	/* Check which event happened i.e, thread notified */
+	val = readl(regs + ES);
+	if (pi->pcfg.num_events < 32
+			&& val & ~((1 << pi->pcfg.num_events) - 1)) {
+		pl330->dmac_tbd.reset_dmac = true;
+		dev_err(pi->dev, "%s:%d Unexpected!\n", __func__, __LINE__);
+		ret = 1;
+		goto updt_exit;
+	}
+
+	for (ev = 0; ev < pi->pcfg.num_events; ev++) {
+		if (val & (1 << ev)) { /* Event occurred */
+			struct pl330_thread *thrd;
+			u32 inten = readl(regs + INTEN);
+			int active;
+
+			/* Clear the event */
+			if (inten & (1 << ev))
+				writel(1 << ev, regs + INTCLR);
+
+			ret = 1;
+
+			id = pl330->events[ev];
+
+			thrd = &pl330->channels[id];
+
+			active = thrd->req_running;
+			if (active == -1) /* Aborted */
+				continue;
+
+			rqdone = &thrd->req[active];
+			mark_free(thrd, active);
+
+			/* Get going again ASAP */
+			_start(thrd);
+
+			/* For now, just make a list of callbacks to be done */
+			list_add_tail(&rqdone->rqd, &pl330->req_done);
+		}
+	}
+
+	/* Now that we are in no hurry, do the callbacks */
+	while (!list_empty(&pl330->req_done)) {
+		struct pl330_req *r;
+
+		rqdone = container_of(pl330->req_done.next,
+					struct _pl330_req, rqd);
+
+		list_del_init(&rqdone->rqd);
+
+		/* Detach the req */
+		r = rqdone->r;
+		rqdone->r = NULL;
+
+		spin_unlock_irqrestore(&pl330->lock, flags);
+		_callback(r, PL330_ERR_NONE);
+		spin_lock_irqsave(&pl330->lock, flags);
+	}
+
+updt_exit:
+	spin_unlock_irqrestore(&pl330->lock, flags);
+
+	if (pl330->dmac_tbd.reset_dmac
+			|| pl330->dmac_tbd.reset_mngr
+			|| pl330->dmac_tbd.reset_chan) {
+		ret = 1;
+		tasklet_schedule(&pl330->tasks);
+	}
+
+	return ret;
+}
+
+static int pl330_chan_ctrl(void *ch_id, enum pl330_chan_op op)
+{
+	struct pl330_thread *thrd = ch_id;
+	struct pl330_dmac *pl330;
+	unsigned long flags;
+	int ret = 0, active;
+
+	if (!thrd || thrd->free || thrd->dmac->state == DYING)
+		return -EINVAL;
+
+	pl330 = thrd->dmac;
+	active = thrd->req_running;
+
+	spin_lock_irqsave(&pl330->lock, flags);
+
+	switch (op) {
+	case PL330_OP_FLUSH:
+		/* Make sure the channel is stopped */
+		_stop(thrd);
+
+		thrd->req[0].r = NULL;
+		thrd->req[1].r = NULL;
+		mark_free(thrd, 0);
+		mark_free(thrd, 1);
+		break;
+
+	case PL330_OP_ABORT:
+		/* Make sure the channel is stopped */
+		_stop(thrd);
+
+		/* ABORT is only for the active req */
+		if (active == -1)
+			break;
+
+		thrd->req[active].r = NULL;
+		mark_free(thrd, active);
+
+		/* Start the next */
+	case PL330_OP_START:
+		if ((active == -1) && !_start(thrd))
+			ret = -EIO;
+		break;
+
+	default:
+		ret = -EINVAL;
+	}
+
+	spin_unlock_irqrestore(&pl330->lock, flags);
+	return ret;
+}
+
+/* Reserve an event */
+static inline int _alloc_event(struct pl330_thread *thrd)
+{
+	struct pl330_dmac *pl330 = thrd->dmac;
+	struct pl330_info *pi = pl330->pinfo;
+	int ev;
+
+	for (ev = 0; ev < pi->pcfg.num_events; ev++)
+		if (pl330->events[ev] == -1) {
+			pl330->events[ev] = thrd->id;
+			return ev;
+		}
+
+	return -1;
+}
+
+static bool _chan_ns(const struct pl330_info *pi, int i)
+{
+	return pi->pcfg.irq_ns & (1 << i);
+}
+
+/* Upon success, returns IdentityToken for the
+ * allocated channel, NULL otherwise.
+ */
+static void *pl330_request_channel(const struct pl330_info *pi)
+{
+	struct pl330_thread *thrd = NULL;
+	struct pl330_dmac *pl330;
+	unsigned long flags;
+	int chans, i;
+
+	if (!pi || !pi->pl330_data)
+		return NULL;
+
+	pl330 = pi->pl330_data;
+
+	if (pl330->state == DYING)
+		return NULL;
+
+	chans = pi->pcfg.num_chan;
+
+	spin_lock_irqsave(&pl330->lock, flags);
+
+	for (i = 0; i < chans; i++) {
+		thrd = &pl330->channels[i];
+		if ((thrd->free) && (!_manager_ns(thrd) ||
+					_chan_ns(pi, i))) {
+			thrd->ev = _alloc_event(thrd);
+			if (thrd->ev >= 0) {
+				thrd->free = false;
+				thrd->lstenq = 1;
+				thrd->req[0].r = NULL;
+				mark_free(thrd, 0);
+				thrd->req[1].r = NULL;
+				mark_free(thrd, 1);
+				break;
+			}
+		}
+		thrd = NULL;
+	}
+
+	spin_unlock_irqrestore(&pl330->lock, flags);
+
+	return thrd;
+}
+
+/* Release an event */
+static inline void _free_event(struct pl330_thread *thrd, int ev)
+{
+	struct pl330_dmac *pl330 = thrd->dmac;
+	struct pl330_info *pi = pl330->pinfo;
+
+	/* If the event is valid and was held by the thread */
+	if (ev >= 0 && ev < pi->pcfg.num_events
+			&& pl330->events[ev] == thrd->id)
+		pl330->events[ev] = -1;
+}
+
+static void pl330_release_channel(void *ch_id)
+{
+	struct pl330_thread *thrd = ch_id;
+	struct pl330_dmac *pl330;
+	unsigned long flags;
+
+	if (!thrd || thrd->free)
+		return;
+
+	_stop(thrd);
+
+	_callback(thrd->req[1 - thrd->lstenq].r, PL330_ERR_ABORT);
+	_callback(thrd->req[thrd->lstenq].r, PL330_ERR_ABORT);
+
+	pl330 = thrd->dmac;
+
+	spin_lock_irqsave(&pl330->lock, flags);
+	_free_event(thrd, thrd->ev);
+	thrd->free = true;
+	spin_unlock_irqrestore(&pl330->lock, flags);
+}
+
+/* Initialize the structure for PL330 configuration, that can be used
+ * by the client driver the make best use of the DMAC
+ */
+static void read_dmac_config(struct pl330_info *pi)
+{
+	void __iomem *regs = pi->base;
+	u32 val;
+
+	val = readl(regs + CRD) >> CRD_DATA_WIDTH_SHIFT;
+	val &= CRD_DATA_WIDTH_MASK;
+	pi->pcfg.data_bus_width = 8 * (1 << val);
+
+	val = readl(regs + CRD) >> CRD_DATA_BUFF_SHIFT;
+	val &= CRD_DATA_BUFF_MASK;
+	pi->pcfg.data_buf_dep = val + 1;
+
+	val = readl(regs + CR0) >> CR0_NUM_CHANS_SHIFT;
+	val &= CR0_NUM_CHANS_MASK;
+	val += 1;
+	pi->pcfg.num_chan = val;
+
+	val = readl(regs + CR0);
+	if (val & CR0_PERIPH_REQ_SET) {
+		val = (val >> CR0_NUM_PERIPH_SHIFT) & CR0_NUM_PERIPH_MASK;
+		val += 1;
+		pi->pcfg.num_peri = val;
+		pi->pcfg.peri_ns = readl(regs + CR4);
+	} else {
+		pi->pcfg.num_peri = 0;
+	}
+
+	val = readl(regs + CR0);
+	if (val & CR0_BOOT_MAN_NS)
+		pi->pcfg.mode |= DMAC_MODE_NS;
+	else
+		pi->pcfg.mode &= ~DMAC_MODE_NS;
+
+	val = readl(regs + CR0) >> CR0_NUM_EVENTS_SHIFT;
+	val &= CR0_NUM_EVENTS_MASK;
+	val += 1;
+	pi->pcfg.num_events = val;
+
+	pi->pcfg.irq_ns = readl(regs + CR3);
+
+	pi->pcfg.periph_id = get_id(pi, PERIPH_ID);
+	pi->pcfg.pcell_id = get_id(pi, PCELL_ID);
+}
+
+static inline void _reset_thread(struct pl330_thread *thrd)
+{
+	struct pl330_dmac *pl330 = thrd->dmac;
+	struct pl330_info *pi = pl330->pinfo;
+
+	thrd->req[0].mc_cpu = pl330->mcode_cpu
+				+ (thrd->id * pi->mcbufsz);
+	thrd->req[0].mc_bus = pl330->mcode_bus
+				+ (thrd->id * pi->mcbufsz);
+	thrd->req[0].r = NULL;
+	mark_free(thrd, 0);
+
+	thrd->req[1].mc_cpu = thrd->req[0].mc_cpu
+				+ pi->mcbufsz / 2;
+	thrd->req[1].mc_bus = thrd->req[0].mc_bus
+				+ pi->mcbufsz / 2;
+	thrd->req[1].r = NULL;
+	mark_free(thrd, 1);
+}
+
+static int dmac_alloc_threads(struct pl330_dmac *pl330)
+{
+	struct pl330_info *pi = pl330->pinfo;
+	int chans = pi->pcfg.num_chan;
+	struct pl330_thread *thrd;
+	int i;
+
+	/* Allocate 1 Manager and 'chans' Channel threads */
+	pl330->channels = kzalloc((1 + chans) * sizeof(*thrd),
+					GFP_KERNEL);
+	if (!pl330->channels)
+		return -ENOMEM;
+
+	/* Init Channel threads */
+	for (i = 0; i < chans; i++) {
+		thrd = &pl330->channels[i];
+		thrd->id = i;
+		thrd->dmac = pl330;
+		_reset_thread(thrd);
+		thrd->free = true;
+	}
+
+	/* MANAGER is indexed at the end */
+	thrd = &pl330->channels[chans];
+	thrd->id = chans;
+	thrd->dmac = pl330;
+	thrd->free = false;
+	pl330->manager = thrd;
+
+	return 0;
+}
+
+static int dmac_alloc_resources(struct pl330_dmac *pl330)
+{
+	struct pl330_info *pi = pl330->pinfo;
+	int chans = pi->pcfg.num_chan;
+	int ret;
+
+	/*
+	 * Alloc MicroCode buffer for 'chans' Channel threads.
+	 * A channel's buffer offset is (Channel_Id * MCODE_BUFF_PERCHAN)
+	 */
+	pl330->mcode_cpu = dma_alloc_coherent(pi->dev,
+				chans * pi->mcbufsz,
+				&pl330->mcode_bus, GFP_KERNEL);
+	if (!pl330->mcode_cpu) {
+		dev_err(pi->dev, "%s:%d Can't allocate memory!\n",
+			__func__, __LINE__);
+		return -ENOMEM;
+	}
+
+	ret = dmac_alloc_threads(pl330);
+	if (ret) {
+		dev_err(pi->dev, "%s:%d Can't to create channels for DMAC!\n",
+			__func__, __LINE__);
+		dma_free_coherent(pi->dev,
+				chans * pi->mcbufsz,
+				pl330->mcode_cpu, pl330->mcode_bus);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int pl330_add(struct pl330_info *pi)
+{
+	struct pl330_dmac *pl330;
+	void __iomem *regs;
+	int i, ret;
+
+	if (!pi || !pi->dev)
+		return -EINVAL;
+
+	/* If already added */
+	if (pi->pl330_data)
+		return -EINVAL;
+
+	/*
+	 * If the SoC can perform reset on the DMAC, then do it
+	 * before reading its configuration.
+	 */
+	if (pi->dmac_reset)
+		pi->dmac_reset(pi);
+
+	regs = pi->base;
+
+	/* Check if we can handle this DMAC */
+	if ((get_id(pi, PERIPH_ID) & 0xfffff) != PERIPH_ID_VAL
+	   || get_id(pi, PCELL_ID) != PCELL_ID_VAL) {
+		dev_err(pi->dev, "PERIPH_ID 0x%x, PCELL_ID 0x%x !\n",
+			get_id(pi, PERIPH_ID), get_id(pi, PCELL_ID));
+		return -EINVAL;
+	}
+
+	/* Read the configuration of the DMAC */
+	read_dmac_config(pi);
+
+	if (pi->pcfg.num_events == 0) {
+		dev_err(pi->dev, "%s:%d Can't work without events!\n",
+			__func__, __LINE__);
+		return -EINVAL;
+	}
+
+	pl330 = kzalloc(sizeof(*pl330), GFP_KERNEL);
+	if (!pl330) {
+		dev_err(pi->dev, "%s:%d Can't allocate memory!\n",
+			__func__, __LINE__);
+		return -ENOMEM;
+	}
+
+	/* Assign the info structure and private data */
+	pl330->pinfo = pi;
+	pi->pl330_data = pl330;
+
+	spin_lock_init(&pl330->lock);
+
+	INIT_LIST_HEAD(&pl330->req_done);
+
+	/* Use default MC buffer size if not provided */
+	if (!pi->mcbufsz)
+		pi->mcbufsz = MCODE_BUFF_PER_REQ * 2;
+
+	/* Mark all events as free */
+	for (i = 0; i < pi->pcfg.num_events; i++)
+		pl330->events[i] = -1;
+
+	/* Allocate resources needed by the DMAC */
+	ret = dmac_alloc_resources(pl330);
+	if (ret) {
+		dev_err(pi->dev, "Unable to create channels for DMAC\n");
+		kfree(pl330);
+		return ret;
+	}
+
+	tasklet_init(&pl330->tasks, pl330_dotask, (unsigned long) pl330);
+
+	pl330->state = INIT;
+
+	return 0;
+}
+
+static int dmac_free_threads(struct pl330_dmac *pl330)
+{
+	struct pl330_info *pi = pl330->pinfo;
+	int chans = pi->pcfg.num_chan;
+	struct pl330_thread *thrd;
+	int i;
+
+	/* Release Channel threads */
+	for (i = 0; i < chans; i++) {
+		thrd = &pl330->channels[i];
+		pl330_release_channel((void *)thrd);
+	}
+
+	/* Free memory */
+	kfree(pl330->channels);
+
+	return 0;
+}
+
+static void dmac_free_resources(struct pl330_dmac *pl330)
+{
+	struct pl330_info *pi = pl330->pinfo;
+	int chans = pi->pcfg.num_chan;
+
+	dmac_free_threads(pl330);
+
+	dma_free_coherent(pi->dev, chans * pi->mcbufsz,
+				pl330->mcode_cpu, pl330->mcode_bus);
+}
+
+static void pl330_del(struct pl330_info *pi)
+{
+	struct pl330_dmac *pl330;
+
+	if (!pi || !pi->pl330_data)
+		return;
+
+	pl330 = pi->pl330_data;
+
+	pl330->state = UNINIT;
+
+	tasklet_kill(&pl330->tasks);
+
+	/* Free DMAC resources */
+	dmac_free_resources(pl330);
+
+	kfree(pl330);
+	pi->pl330_data = NULL;
+}
+
 /* forward declaration */
 static struct amba_driver pl330_driver;
 
@@ -234,7 +2320,7 @@ static void pl330_tasklet(unsigned long data)
 	/* Pick up ripe tomatoes */
 	list_for_each_entry_safe(desc, _dt, &pch->work_list, node)
 		if (desc->status == DONE) {
-			pch->completed = desc->txd.cookie;
+			dma_cookie_complete(&desc->txd);
 			list_move_tail(&desc->node, &list);
 		}
 
@@ -305,7 +2391,7 @@ static int pl330_alloc_chan_resources(struct dma_chan *chan)
 
 	spin_lock_irqsave(&pch->lock, flags);
 
-	pch->completed = chan->cookie = 1;
+	dma_cookie_init(chan);
 	pch->cyclic = false;
 
 	pch->pl330_chid = pl330_request_channel(&pdmac->pif);
@@ -340,7 +2426,6 @@ static int pl330_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd, unsigned
 		/* Mark all desc done */
 		list_for_each_entry_safe(desc, _dt, &pch->work_list , node) {
 			desc->status = DONE;
-			pch->completed = desc->txd.cookie;
 			list_move_tail(&desc->node, &list);
 		}
 
@@ -350,14 +2435,14 @@ static int pl330_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd, unsigned
 	case DMA_SLAVE_CONFIG:
 		slave_config = (struct dma_slave_config *)arg;
 
-		if (slave_config->direction == DMA_TO_DEVICE) {
+		if (slave_config->direction == DMA_MEM_TO_DEV) {
 			if (slave_config->dst_addr)
 				pch->fifo_addr = slave_config->dst_addr;
 			if (slave_config->dst_addr_width)
 				pch->burst_sz = __ffs(slave_config->dst_addr_width);
 			if (slave_config->dst_maxburst)
 				pch->burst_len = slave_config->dst_maxburst;
-		} else if (slave_config->direction == DMA_FROM_DEVICE) {
+		} else if (slave_config->direction == DMA_DEV_TO_MEM) {
 			if (slave_config->src_addr)
 				pch->fifo_addr = slave_config->src_addr;
 			if (slave_config->src_addr_width)
@@ -396,18 +2481,7 @@ static enum dma_status
 pl330_tx_status(struct dma_chan *chan, dma_cookie_t cookie,
 		 struct dma_tx_state *txstate)
 {
-	struct dma_pl330_chan *pch = to_pchan(chan);
-	dma_cookie_t last_done, last_used;
-	int ret;
-
-	last_done = pch->completed;
-	last_used = chan->cookie;
-
-	ret = dma_async_is_complete(cookie, last_done, last_used);
-
-	dma_set_tx_state(txstate, last_done, last_used, 0);
-
-	return ret;
+	return dma_cookie_status(chan, cookie, txstate);
 }
 
 static void pl330_issue_pending(struct dma_chan *chan)
@@ -430,26 +2504,16 @@ static dma_cookie_t pl330_tx_submit(struct dma_async_tx_descriptor *tx)
 	spin_lock_irqsave(&pch->lock, flags);
 
 	/* Assign cookies to all nodes */
-	cookie = tx->chan->cookie;
-
 	while (!list_empty(&last->node)) {
 		desc = list_entry(last->node.next, struct dma_pl330_desc, node);
 
-		if (++cookie < 0)
-			cookie = 1;
-		desc->txd.cookie = cookie;
+		dma_cookie_assign(&desc->txd);
 
 		list_move_tail(&desc->node, &pch->work_list);
 	}
 
-	if (++cookie < 0)
-		cookie = 1;
-	last->txd.cookie = cookie;
-
+	cookie = dma_cookie_assign(&last->txd);
 	list_add_tail(&last->node, &pch->work_list);
-
-	tx->chan->cookie = cookie;
-
 	spin_unlock_irqrestore(&pch->lock, flags);
 
 	return cookie;
@@ -553,6 +2617,7 @@ static struct dma_pl330_desc *pl330_get_desc(struct dma_pl330_chan *pch)
 	async_tx_ack(&desc->txd);
 
 	desc->req.peri = peri_id ? pch->chan.chan_id : 0;
+	desc->rqcfg.pcfg = &pch->dmac->pif.pcfg;
 
 	dma_async_tx_descriptor_init(&desc->txd, &pch->chan);
 
@@ -621,7 +2686,8 @@ static inline int get_burst_len(struct dma_pl330_desc *desc, size_t len)
 
 static struct dma_async_tx_descriptor *pl330_prep_dma_cyclic(
 		struct dma_chan *chan, dma_addr_t dma_addr, size_t len,
-		size_t period_len, enum dma_data_direction direction)
+		size_t period_len, enum dma_transfer_direction direction,
+		void *context)
 {
 	struct dma_pl330_desc *desc;
 	struct dma_pl330_chan *pch = to_pchan(chan);
@@ -636,14 +2702,14 @@ static struct dma_async_tx_descriptor *pl330_prep_dma_cyclic(
 	}
 
 	switch (direction) {
-	case DMA_TO_DEVICE:
+	case DMA_MEM_TO_DEV:
 		desc->rqcfg.src_inc = 1;
 		desc->rqcfg.dst_inc = 0;
 		desc->req.rqtype = MEMTODEV;
 		src = dma_addr;
 		dst = pch->fifo_addr;
 		break;
-	case DMA_FROM_DEVICE:
+	case DMA_DEV_TO_MEM:
 		desc->rqcfg.src_inc = 0;
 		desc->rqcfg.dst_inc = 1;
 		desc->req.rqtype = DEVTOMEM;
@@ -710,8 +2776,8 @@ pl330_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dst,
 
 static struct dma_async_tx_descriptor *
 pl330_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
-		unsigned int sg_len, enum dma_data_direction direction,
-		unsigned long flg)
+		unsigned int sg_len, enum dma_transfer_direction direction,
+		unsigned long flg, void *context)
 {
 	struct dma_pl330_desc *first, *desc = NULL;
 	struct dma_pl330_chan *pch = to_pchan(chan);
@@ -759,7 +2825,7 @@ pl330_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 		else
 			list_add_tail(&desc->node, &first->node);
 
-		if (direction == DMA_TO_DEVICE) {
+		if (direction == DMA_MEM_TO_DEV) {
 			desc->rqcfg.src_inc = 1;
 			desc->rqcfg.dst_inc = 0;
 			desc->req.rqtype = MEMTODEV;
@@ -829,22 +2895,12 @@ pl330_probe(struct amba_device *adev, const struct amba_id *id)
 	if (IS_ERR(pdmac->clk)) {
 		dev_err(&adev->dev, "Cannot get operation clock.\n");
 		ret = -EINVAL;
-		goto probe_err1;
+		goto probe_err2;
 	}
 
 	amba_set_drvdata(adev, pdmac);
 
-#ifdef CONFIG_PM_RUNTIME
-	/* to use the runtime PM helper functions */
-	pm_runtime_enable(&adev->dev);
-
-	/* enable the power domain */
-	if (pm_runtime_get_sync(&adev->dev)) {
-		dev_err(&adev->dev, "failed to get runtime pm\n");
-		ret = -ENODEV;
-		goto probe_err1;
-	}
-#else
+#ifndef CONFIG_PM_RUNTIME
 	/* enable dma clk */
 	clk_enable(pdmac->clk);
 #endif
@@ -853,11 +2909,11 @@ pl330_probe(struct amba_device *adev, const struct amba_id *id)
 	ret = request_irq(irq, pl330_irq_handler, 0,
 			dev_name(&adev->dev), pi);
 	if (ret)
-		goto probe_err2;
+		goto probe_err3;
 
 	ret = pl330_add(pi);
 	if (ret)
-		goto probe_err3;
+		goto probe_err4;
 
 	INIT_LIST_HEAD(&pdmac->desc_pool);
 	spin_lock_init(&pdmac->pool_lock);
@@ -914,7 +2970,7 @@ pl330_probe(struct amba_device *adev, const struct amba_id *id)
 	ret = dma_async_device_register(pd);
 	if (ret) {
 		dev_err(&adev->dev, "unable to register DMAC\n");
-		goto probe_err4;
+		goto probe_err5;
 	}
 
 	dev_info(&adev->dev,
@@ -927,10 +2983,15 @@ pl330_probe(struct amba_device *adev, const struct amba_id *id)
 
 	return 0;
 
-probe_err4:
+probe_err5:
 	pl330_del(pi);
-probe_err3:
+probe_err4:
 	free_irq(irq, pi);
+probe_err3:
+#ifndef CONFIG_PM_RUNTIME
+	clk_disable(pdmac->clk);
+#endif
+	clk_put(pdmac->clk);
 probe_err2:
 	iounmap(pi->base);
 probe_err1:
@@ -977,10 +3038,7 @@ static int __devexit pl330_remove(struct amba_device *adev)
 	res = &adev->res;
 	release_mem_region(res->start, resource_size(res));
 
-#ifdef CONFIG_PM_RUNTIME
-	pm_runtime_put(&adev->dev);
-	pm_runtime_disable(&adev->dev);
-#else
+#ifndef CONFIG_PM_RUNTIME
 	clk_disable(pdmac->clk);
 #endif
 
@@ -1048,18 +3106,7 @@ static struct amba_driver pl330_driver = {
 	.remove = pl330_remove,
 };
 
-static int __init pl330_init(void)
-{
-	return amba_driver_register(&pl330_driver);
-}
-module_init(pl330_init);
-
-static void __exit pl330_exit(void)
-{
-	amba_driver_unregister(&pl330_driver);
-	return;
-}
-module_exit(pl330_exit);
+module_amba_driver(pl330_driver);
 
 MODULE_AUTHOR("Jaswinder Singh <jassi.brar@samsung.com>");
 MODULE_DESCRIPTION("API Driver for PL330 DMAC");
diff --git a/drivers/dma/ppc4xx/adma.c b/drivers/dma/ppc4xx/adma.c
index fc457a7e8832..ced98826684a 100644
--- a/drivers/dma/ppc4xx/adma.c
+++ b/drivers/dma/ppc4xx/adma.c
@@ -46,6 +46,7 @@
 #include <asm/dcr.h>
 #include <asm/dcr-regs.h>
 #include "adma.h"
+#include "../dmaengine.h"
 
 enum ppc_adma_init_code {
 	PPC_ADMA_INIT_OK = 0,
@@ -1930,7 +1931,7 @@ static void __ppc440spe_adma_slot_cleanup(struct ppc440spe_adma_chan *chan)
 				if (end_of_chain && slot_cnt) {
 					/* Should wait for ZeroSum completion */
 					if (cookie > 0)
-						chan->completed_cookie = cookie;
+						chan->common.completed_cookie = cookie;
 					return;
 				}
 
@@ -1960,7 +1961,7 @@ static void __ppc440spe_adma_slot_cleanup(struct ppc440spe_adma_chan *chan)
 	BUG_ON(!seen_current);
 
 	if (cookie > 0) {
-		chan->completed_cookie = cookie;
+		chan->common.completed_cookie = cookie;
 		pr_debug("\tcompleted cookie %d\n", cookie);
 	}
 
@@ -2150,22 +2151,6 @@ static int ppc440spe_adma_alloc_chan_resources(struct dma_chan *chan)
 }
 
 /**
- * ppc440spe_desc_assign_cookie - assign a cookie
- */
-static dma_cookie_t ppc440spe_desc_assign_cookie(
-		struct ppc440spe_adma_chan *chan,
-		struct ppc440spe_adma_desc_slot *desc)
-{
-	dma_cookie_t cookie = chan->common.cookie;
-
-	cookie++;
-	if (cookie < 0)
-		cookie = 1;
-	chan->common.cookie = desc->async_tx.cookie = cookie;
-	return cookie;
-}
-
-/**
  * ppc440spe_rxor_set_region_data -
  */
 static void ppc440spe_rxor_set_region(struct ppc440spe_adma_desc_slot *desc,
@@ -2235,8 +2220,7 @@ static dma_cookie_t ppc440spe_adma_tx_submit(struct dma_async_tx_descriptor *tx)
 	slots_per_op = group_start->slots_per_op;
 
 	spin_lock_bh(&chan->lock);
-
-	cookie = ppc440spe_desc_assign_cookie(chan, sw_desc);
+	cookie = dma_cookie_assign(tx);
 
 	if (unlikely(list_empty(&chan->chain))) {
 		/* first peer */
@@ -3944,28 +3928,16 @@ static enum dma_status ppc440spe_adma_tx_status(struct dma_chan *chan,
 			dma_cookie_t cookie, struct dma_tx_state *txstate)
 {
 	struct ppc440spe_adma_chan *ppc440spe_chan;
-	dma_cookie_t last_used;
-	dma_cookie_t last_complete;
 	enum dma_status ret;
 
 	ppc440spe_chan = to_ppc440spe_adma_chan(chan);
-	last_used = chan->cookie;
-	last_complete = ppc440spe_chan->completed_cookie;
-
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
-
-	ret = dma_async_is_complete(cookie, last_complete, last_used);
+	ret = dma_cookie_status(chan, cookie, txstate);
 	if (ret == DMA_SUCCESS)
 		return ret;
 
 	ppc440spe_adma_slot_cleanup(ppc440spe_chan);
 
-	last_used = chan->cookie;
-	last_complete = ppc440spe_chan->completed_cookie;
-
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
-
-	return dma_async_is_complete(cookie, last_complete, last_used);
+	return dma_cookie_status(chan, cookie, txstate);
 }
 
 /**
@@ -4050,16 +4022,12 @@ static void ppc440spe_chan_start_null_xor(struct ppc440spe_adma_chan *chan)
 		async_tx_ack(&sw_desc->async_tx);
 		ppc440spe_desc_init_null_xor(group_start);
 
-		cookie = chan->common.cookie;
-		cookie++;
-		if (cookie <= 1)
-			cookie = 2;
+		cookie = dma_cookie_assign(&sw_desc->async_tx);
 
 		/* initialize the completed cookie to be less than
 		 * the most recently used cookie
 		 */
-		chan->completed_cookie = cookie - 1;
-		chan->common.cookie = sw_desc->async_tx.cookie = cookie;
+		chan->common.completed_cookie = cookie - 1;
 
 		/* channel should not be busy */
 		BUG_ON(ppc440spe_chan_is_busy(chan));
@@ -4529,6 +4497,7 @@ static int __devinit ppc440spe_adma_probe(struct platform_device *ofdev)
 	INIT_LIST_HEAD(&chan->all_slots);
 	chan->device = adev;
 	chan->common.device = &adev->common;
+	dma_cookie_init(&chan->common);
 	list_add_tail(&chan->common.device_node, &adev->common.channels);
 	tasklet_init(&chan->irq_tasklet, ppc440spe_adma_tasklet,
 		     (unsigned long)chan);
diff --git a/drivers/dma/ppc4xx/adma.h b/drivers/dma/ppc4xx/adma.h
index 8ada5a812e3b..26b7a5ed9ac7 100644
--- a/drivers/dma/ppc4xx/adma.h
+++ b/drivers/dma/ppc4xx/adma.h
@@ -81,7 +81,6 @@ struct ppc440spe_adma_device {
  * @common: common dmaengine channel object members
  * @all_slots: complete domain of slots usable by the channel
  * @pending: allows batching of hardware operations
- * @completed_cookie: identifier for the most recently completed operation
  * @slots_allocated: records the actual size of the descriptor slot pool
  * @hw_chain_inited: h/w descriptor chain initialization flag
  * @irq_tasklet: bottom half where ppc440spe_adma_slot_cleanup runs
@@ -99,7 +98,6 @@ struct ppc440spe_adma_chan {
 	struct list_head all_slots;
 	struct ppc440spe_adma_desc_slot *last_used;
 	int pending;
-	dma_cookie_t completed_cookie;
 	int slots_allocated;
 	int hw_chain_inited;
 	struct tasklet_struct irq_tasklet;
diff --git a/drivers/dma/sa11x0-dma.c b/drivers/dma/sa11x0-dma.c
new file mode 100644
index 000000000000..ec78ccef9132
--- /dev/null
+++ b/drivers/dma/sa11x0-dma.c
@@ -0,0 +1,1109 @@
+/*
+ * SA11x0 DMAengine support
+ *
+ * Copyright (C) 2012 Russell King
+ *   Derived in part from arch/arm/mach-sa1100/dma.c,
+ *   Copyright (C) 2000, 2001 by Nicolas Pitre
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/sched.h>
+#include <linux/device.h>
+#include <linux/dmaengine.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/sa11x0-dma.h>
+#include <linux/slab.h>
+#include <linux/spinlock.h>
+
+#define NR_PHY_CHAN	6
+#define DMA_ALIGN	3
+#define DMA_MAX_SIZE	0x1fff
+#define DMA_CHUNK_SIZE	0x1000
+
+#define DMA_DDAR	0x00
+#define DMA_DCSR_S	0x04
+#define DMA_DCSR_C	0x08
+#define DMA_DCSR_R	0x0c
+#define DMA_DBSA	0x10
+#define DMA_DBTA	0x14
+#define DMA_DBSB	0x18
+#define DMA_DBTB	0x1c
+#define DMA_SIZE	0x20
+
+#define DCSR_RUN	(1 << 0)
+#define DCSR_IE		(1 << 1)
+#define DCSR_ERROR	(1 << 2)
+#define DCSR_DONEA	(1 << 3)
+#define DCSR_STRTA	(1 << 4)
+#define DCSR_DONEB	(1 << 5)
+#define DCSR_STRTB	(1 << 6)
+#define DCSR_BIU	(1 << 7)
+
+#define DDAR_RW		(1 << 0)	/* 0 = W, 1 = R */
+#define DDAR_E		(1 << 1)	/* 0 = LE, 1 = BE */
+#define DDAR_BS		(1 << 2)	/* 0 = BS4, 1 = BS8 */
+#define DDAR_DW		(1 << 3)	/* 0 = 8b, 1 = 16b */
+#define DDAR_Ser0UDCTr	(0x0 << 4)
+#define DDAR_Ser0UDCRc	(0x1 << 4)
+#define DDAR_Ser1SDLCTr	(0x2 << 4)
+#define DDAR_Ser1SDLCRc	(0x3 << 4)
+#define DDAR_Ser1UARTTr	(0x4 << 4)
+#define DDAR_Ser1UARTRc	(0x5 << 4)
+#define DDAR_Ser2ICPTr	(0x6 << 4)
+#define DDAR_Ser2ICPRc	(0x7 << 4)
+#define DDAR_Ser3UARTTr	(0x8 << 4)
+#define DDAR_Ser3UARTRc	(0x9 << 4)
+#define DDAR_Ser4MCP0Tr	(0xa << 4)
+#define DDAR_Ser4MCP0Rc	(0xb << 4)
+#define DDAR_Ser4MCP1Tr	(0xc << 4)
+#define DDAR_Ser4MCP1Rc	(0xd << 4)
+#define DDAR_Ser4SSPTr	(0xe << 4)
+#define DDAR_Ser4SSPRc	(0xf << 4)
+
+struct sa11x0_dma_sg {
+	u32			addr;
+	u32			len;
+};
+
+struct sa11x0_dma_desc {
+	struct dma_async_tx_descriptor tx;
+	u32			ddar;
+	size_t			size;
+
+	/* maybe protected by c->lock */
+	struct list_head	node;
+	unsigned		sglen;
+	struct sa11x0_dma_sg	sg[0];
+};
+
+struct sa11x0_dma_phy;
+
+struct sa11x0_dma_chan {
+	struct dma_chan		chan;
+	spinlock_t		lock;
+	dma_cookie_t		lc;
+
+	/* protected by c->lock */
+	struct sa11x0_dma_phy	*phy;
+	enum dma_status		status;
+	struct list_head	desc_submitted;
+	struct list_head	desc_issued;
+
+	/* protected by d->lock */
+	struct list_head	node;
+
+	u32			ddar;
+	const char		*name;
+};
+
+struct sa11x0_dma_phy {
+	void __iomem		*base;
+	struct sa11x0_dma_dev	*dev;
+	unsigned		num;
+
+	struct sa11x0_dma_chan	*vchan;
+
+	/* Protected by c->lock */
+	unsigned		sg_load;
+	struct sa11x0_dma_desc	*txd_load;
+	unsigned		sg_done;
+	struct sa11x0_dma_desc	*txd_done;
+#ifdef CONFIG_PM_SLEEP
+	u32			dbs[2];
+	u32			dbt[2];
+	u32			dcsr;
+#endif
+};
+
+struct sa11x0_dma_dev {
+	struct dma_device	slave;
+	void __iomem		*base;
+	spinlock_t		lock;
+	struct tasklet_struct	task;
+	struct list_head	chan_pending;
+	struct list_head	desc_complete;
+	struct sa11x0_dma_phy	phy[NR_PHY_CHAN];
+};
+
+static struct sa11x0_dma_chan *to_sa11x0_dma_chan(struct dma_chan *chan)
+{
+	return container_of(chan, struct sa11x0_dma_chan, chan);
+}
+
+static struct sa11x0_dma_dev *to_sa11x0_dma(struct dma_device *dmadev)
+{
+	return container_of(dmadev, struct sa11x0_dma_dev, slave);
+}
+
+static struct sa11x0_dma_desc *to_sa11x0_dma_tx(struct dma_async_tx_descriptor *tx)
+{
+	return container_of(tx, struct sa11x0_dma_desc, tx);
+}
+
+static struct sa11x0_dma_desc *sa11x0_dma_next_desc(struct sa11x0_dma_chan *c)
+{
+	if (list_empty(&c->desc_issued))
+		return NULL;
+
+	return list_first_entry(&c->desc_issued, struct sa11x0_dma_desc, node);
+}
+
+static void sa11x0_dma_start_desc(struct sa11x0_dma_phy *p, struct sa11x0_dma_desc *txd)
+{
+	list_del(&txd->node);
+	p->txd_load = txd;
+	p->sg_load = 0;
+
+	dev_vdbg(p->dev->slave.dev, "pchan %u: txd %p[%x]: starting: DDAR:%x\n",
+		p->num, txd, txd->tx.cookie, txd->ddar);
+}
+
+static void noinline sa11x0_dma_start_sg(struct sa11x0_dma_phy *p,
+	struct sa11x0_dma_chan *c)
+{
+	struct sa11x0_dma_desc *txd = p->txd_load;
+	struct sa11x0_dma_sg *sg;
+	void __iomem *base = p->base;
+	unsigned dbsx, dbtx;
+	u32 dcsr;
+
+	if (!txd)
+		return;
+
+	dcsr = readl_relaxed(base + DMA_DCSR_R);
+
+	/* Don't try to load the next transfer if both buffers are started */
+	if ((dcsr & (DCSR_STRTA | DCSR_STRTB)) == (DCSR_STRTA | DCSR_STRTB))
+		return;
+
+	if (p->sg_load == txd->sglen) {
+		struct sa11x0_dma_desc *txn = sa11x0_dma_next_desc(c);
+
+		/*
+		 * We have reached the end of the current descriptor.
+		 * Peek at the next descriptor, and if compatible with
+		 * the current, start processing it.
+		 */
+		if (txn && txn->ddar == txd->ddar) {
+			txd = txn;
+			sa11x0_dma_start_desc(p, txn);
+		} else {
+			p->txd_load = NULL;
+			return;
+		}
+	}
+
+	sg = &txd->sg[p->sg_load++];
+
+	/* Select buffer to load according to channel status */
+	if (((dcsr & (DCSR_BIU | DCSR_STRTB)) == (DCSR_BIU | DCSR_STRTB)) ||
+	    ((dcsr & (DCSR_BIU | DCSR_STRTA)) == 0)) {
+		dbsx = DMA_DBSA;
+		dbtx = DMA_DBTA;
+		dcsr = DCSR_STRTA | DCSR_IE | DCSR_RUN;
+	} else {
+		dbsx = DMA_DBSB;
+		dbtx = DMA_DBTB;
+		dcsr = DCSR_STRTB | DCSR_IE | DCSR_RUN;
+	}
+
+	writel_relaxed(sg->addr, base + dbsx);
+	writel_relaxed(sg->len, base + dbtx);
+	writel(dcsr, base + DMA_DCSR_S);
+
+	dev_dbg(p->dev->slave.dev, "pchan %u: load: DCSR:%02x DBS%c:%08x DBT%c:%08x\n",
+		p->num, dcsr,
+		'A' + (dbsx == DMA_DBSB), sg->addr,
+		'A' + (dbtx == DMA_DBTB), sg->len);
+}
+
+static void noinline sa11x0_dma_complete(struct sa11x0_dma_phy *p,
+	struct sa11x0_dma_chan *c)
+{
+	struct sa11x0_dma_desc *txd = p->txd_done;
+
+	if (++p->sg_done == txd->sglen) {
+		struct sa11x0_dma_dev *d = p->dev;
+
+		dev_vdbg(d->slave.dev, "pchan %u: txd %p[%x]: completed\n",
+			p->num, p->txd_done, p->txd_done->tx.cookie);
+
+		c->lc = txd->tx.cookie;
+
+		spin_lock(&d->lock);
+		list_add_tail(&txd->node, &d->desc_complete);
+		spin_unlock(&d->lock);
+
+		p->sg_done = 0;
+		p->txd_done = p->txd_load;
+
+		tasklet_schedule(&d->task);
+	}
+
+	sa11x0_dma_start_sg(p, c);
+}
+
+static irqreturn_t sa11x0_dma_irq(int irq, void *dev_id)
+{
+	struct sa11x0_dma_phy *p = dev_id;
+	struct sa11x0_dma_dev *d = p->dev;
+	struct sa11x0_dma_chan *c;
+	u32 dcsr;
+
+	dcsr = readl_relaxed(p->base + DMA_DCSR_R);
+	if (!(dcsr & (DCSR_ERROR | DCSR_DONEA | DCSR_DONEB)))
+		return IRQ_NONE;
+
+	/* Clear reported status bits */
+	writel_relaxed(dcsr & (DCSR_ERROR | DCSR_DONEA | DCSR_DONEB),
+		p->base + DMA_DCSR_C);
+
+	dev_dbg(d->slave.dev, "pchan %u: irq: DCSR:%02x\n", p->num, dcsr);
+
+	if (dcsr & DCSR_ERROR) {
+		dev_err(d->slave.dev, "pchan %u: error. DCSR:%02x DDAR:%08x DBSA:%08x DBTA:%08x DBSB:%08x DBTB:%08x\n",
+			p->num, dcsr,
+			readl_relaxed(p->base + DMA_DDAR),
+			readl_relaxed(p->base + DMA_DBSA),
+			readl_relaxed(p->base + DMA_DBTA),
+			readl_relaxed(p->base + DMA_DBSB),
+			readl_relaxed(p->base + DMA_DBTB));
+	}
+
+	c = p->vchan;
+	if (c) {
+		unsigned long flags;
+
+		spin_lock_irqsave(&c->lock, flags);
+		/*
+		 * Now that we're holding the lock, check that the vchan
+		 * really is associated with this pchan before touching the
+		 * hardware.  This should always succeed, because we won't
+		 * change p->vchan or c->phy while the channel is actively
+		 * transferring.
+		 */
+		if (c->phy == p) {
+			if (dcsr & DCSR_DONEA)
+				sa11x0_dma_complete(p, c);
+			if (dcsr & DCSR_DONEB)
+				sa11x0_dma_complete(p, c);
+		}
+		spin_unlock_irqrestore(&c->lock, flags);
+	}
+
+	return IRQ_HANDLED;
+}
+
+static void sa11x0_dma_start_txd(struct sa11x0_dma_chan *c)
+{
+	struct sa11x0_dma_desc *txd = sa11x0_dma_next_desc(c);
+
+	/* If the issued list is empty, we have no further txds to process */
+	if (txd) {
+		struct sa11x0_dma_phy *p = c->phy;
+
+		sa11x0_dma_start_desc(p, txd);
+		p->txd_done = txd;
+		p->sg_done = 0;
+
+		/* The channel should not have any transfers started */
+		WARN_ON(readl_relaxed(p->base + DMA_DCSR_R) &
+				      (DCSR_STRTA | DCSR_STRTB));
+
+		/* Clear the run and start bits before changing DDAR */
+		writel_relaxed(DCSR_RUN | DCSR_STRTA | DCSR_STRTB,
+			       p->base + DMA_DCSR_C);
+		writel_relaxed(txd->ddar, p->base + DMA_DDAR);
+
+		/* Try to start both buffers */
+		sa11x0_dma_start_sg(p, c);
+		sa11x0_dma_start_sg(p, c);
+	}
+}
+
+static void sa11x0_dma_tasklet(unsigned long arg)
+{
+	struct sa11x0_dma_dev *d = (struct sa11x0_dma_dev *)arg;
+	struct sa11x0_dma_phy *p;
+	struct sa11x0_dma_chan *c;
+	struct sa11x0_dma_desc *txd, *txn;
+	LIST_HEAD(head);
+	unsigned pch, pch_alloc = 0;
+
+	dev_dbg(d->slave.dev, "tasklet enter\n");
+
+	/* Get the completed tx descriptors */
+	spin_lock_irq(&d->lock);
+	list_splice_init(&d->desc_complete, &head);
+	spin_unlock_irq(&d->lock);
+
+	list_for_each_entry(txd, &head, node) {
+		c = to_sa11x0_dma_chan(txd->tx.chan);
+
+		dev_dbg(d->slave.dev, "vchan %p: txd %p[%x] completed\n",
+			c, txd, txd->tx.cookie);
+
+		spin_lock_irq(&c->lock);
+		p = c->phy;
+		if (p) {
+			if (!p->txd_done)
+				sa11x0_dma_start_txd(c);
+			if (!p->txd_done) {
+				/* No current txd associated with this channel */
+				dev_dbg(d->slave.dev, "pchan %u: free\n", p->num);
+
+				/* Mark this channel free */
+				c->phy = NULL;
+				p->vchan = NULL;
+			}
+		}
+		spin_unlock_irq(&c->lock);
+	}
+
+	spin_lock_irq(&d->lock);
+	for (pch = 0; pch < NR_PHY_CHAN; pch++) {
+		p = &d->phy[pch];
+
+		if (p->vchan == NULL && !list_empty(&d->chan_pending)) {
+			c = list_first_entry(&d->chan_pending,
+				struct sa11x0_dma_chan, node);
+			list_del_init(&c->node);
+
+			pch_alloc |= 1 << pch;
+
+			/* Mark this channel allocated */
+			p->vchan = c;
+
+			dev_dbg(d->slave.dev, "pchan %u: alloc vchan %p\n", pch, c);
+		}
+	}
+	spin_unlock_irq(&d->lock);
+
+	for (pch = 0; pch < NR_PHY_CHAN; pch++) {
+		if (pch_alloc & (1 << pch)) {
+			p = &d->phy[pch];
+			c = p->vchan;
+
+			spin_lock_irq(&c->lock);
+			c->phy = p;
+
+			sa11x0_dma_start_txd(c);
+			spin_unlock_irq(&c->lock);
+		}
+	}
+
+	/* Now free the completed tx descriptor, and call their callbacks */
+	list_for_each_entry_safe(txd, txn, &head, node) {
+		dma_async_tx_callback callback = txd->tx.callback;
+		void *callback_param = txd->tx.callback_param;
+
+		dev_dbg(d->slave.dev, "txd %p[%x]: callback and free\n",
+			txd, txd->tx.cookie);
+
+		kfree(txd);
+
+		if (callback)
+			callback(callback_param);
+	}
+
+	dev_dbg(d->slave.dev, "tasklet exit\n");
+}
+
+
+static void sa11x0_dma_desc_free(struct sa11x0_dma_dev *d, struct list_head *head)
+{
+	struct sa11x0_dma_desc *txd, *txn;
+
+	list_for_each_entry_safe(txd, txn, head, node) {
+		dev_dbg(d->slave.dev, "txd %p: freeing\n", txd);
+		kfree(txd);
+	}
+}
+
+static int sa11x0_dma_alloc_chan_resources(struct dma_chan *chan)
+{
+	return 0;
+}
+
+static void sa11x0_dma_free_chan_resources(struct dma_chan *chan)
+{
+	struct sa11x0_dma_chan *c = to_sa11x0_dma_chan(chan);
+	struct sa11x0_dma_dev *d = to_sa11x0_dma(chan->device);
+	unsigned long flags;
+	LIST_HEAD(head);
+
+	spin_lock_irqsave(&c->lock, flags);
+	spin_lock(&d->lock);
+	list_del_init(&c->node);
+	spin_unlock(&d->lock);
+
+	list_splice_tail_init(&c->desc_submitted, &head);
+	list_splice_tail_init(&c->desc_issued, &head);
+	spin_unlock_irqrestore(&c->lock, flags);
+
+	sa11x0_dma_desc_free(d, &head);
+}
+
+static dma_addr_t sa11x0_dma_pos(struct sa11x0_dma_phy *p)
+{
+	unsigned reg;
+	u32 dcsr;
+
+	dcsr = readl_relaxed(p->base + DMA_DCSR_R);
+
+	if ((dcsr & (DCSR_BIU | DCSR_STRTA)) == DCSR_STRTA ||
+	    (dcsr & (DCSR_BIU | DCSR_STRTB)) == DCSR_BIU)
+		reg = DMA_DBSA;
+	else
+		reg = DMA_DBSB;
+
+	return readl_relaxed(p->base + reg);
+}
+
+static enum dma_status sa11x0_dma_tx_status(struct dma_chan *chan,
+	dma_cookie_t cookie, struct dma_tx_state *state)
+{
+	struct sa11x0_dma_chan *c = to_sa11x0_dma_chan(chan);
+	struct sa11x0_dma_dev *d = to_sa11x0_dma(chan->device);
+	struct sa11x0_dma_phy *p;
+	struct sa11x0_dma_desc *txd;
+	dma_cookie_t last_used, last_complete;
+	unsigned long flags;
+	enum dma_status ret;
+	size_t bytes = 0;
+
+	last_used = c->chan.cookie;
+	last_complete = c->lc;
+
+	ret = dma_async_is_complete(cookie, last_complete, last_used);
+	if (ret == DMA_SUCCESS) {
+		dma_set_tx_state(state, last_complete, last_used, 0);
+		return ret;
+	}
+
+	spin_lock_irqsave(&c->lock, flags);
+	p = c->phy;
+	ret = c->status;
+	if (p) {
+		dma_addr_t addr = sa11x0_dma_pos(p);
+
+		dev_vdbg(d->slave.dev, "tx_status: addr:%x\n", addr);
+
+		txd = p->txd_done;
+		if (txd) {
+			unsigned i;
+
+			for (i = 0; i < txd->sglen; i++) {
+				dev_vdbg(d->slave.dev, "tx_status: [%u] %x+%x\n",
+					i, txd->sg[i].addr, txd->sg[i].len);
+				if (addr >= txd->sg[i].addr &&
+				    addr < txd->sg[i].addr + txd->sg[i].len) {
+					unsigned len;
+
+					len = txd->sg[i].len -
+						(addr - txd->sg[i].addr);
+					dev_vdbg(d->slave.dev, "tx_status: [%u] +%x\n",
+						i, len);
+					bytes += len;
+					i++;
+					break;
+				}
+			}
+			for (; i < txd->sglen; i++) {
+				dev_vdbg(d->slave.dev, "tx_status: [%u] %x+%x ++\n",
+					i, txd->sg[i].addr, txd->sg[i].len);
+				bytes += txd->sg[i].len;
+			}
+		}
+		if (txd != p->txd_load && p->txd_load)
+			bytes += p->txd_load->size;
+	}
+	list_for_each_entry(txd, &c->desc_issued, node) {
+		bytes += txd->size;
+	}
+	spin_unlock_irqrestore(&c->lock, flags);
+
+	dma_set_tx_state(state, last_complete, last_used, bytes);
+
+	dev_vdbg(d->slave.dev, "tx_status: bytes 0x%zx\n", bytes);
+
+	return ret;
+}
+
+/*
+ * Move pending txds to the issued list, and re-init pending list.
+ * If not already pending, add this channel to the list of pending
+ * channels and trigger the tasklet to run.
+ */
+static void sa11x0_dma_issue_pending(struct dma_chan *chan)
+{
+	struct sa11x0_dma_chan *c = to_sa11x0_dma_chan(chan);
+	struct sa11x0_dma_dev *d = to_sa11x0_dma(chan->device);
+	unsigned long flags;
+
+	spin_lock_irqsave(&c->lock, flags);
+	list_splice_tail_init(&c->desc_submitted, &c->desc_issued);
+	if (!list_empty(&c->desc_issued)) {
+		spin_lock(&d->lock);
+		if (!c->phy && list_empty(&c->node)) {
+			list_add_tail(&c->node, &d->chan_pending);
+			tasklet_schedule(&d->task);
+			dev_dbg(d->slave.dev, "vchan %p: issued\n", c);
+		}
+		spin_unlock(&d->lock);
+	} else
+		dev_dbg(d->slave.dev, "vchan %p: nothing to issue\n", c);
+	spin_unlock_irqrestore(&c->lock, flags);
+}
+
+static dma_cookie_t sa11x0_dma_tx_submit(struct dma_async_tx_descriptor *tx)
+{
+	struct sa11x0_dma_chan *c = to_sa11x0_dma_chan(tx->chan);
+	struct sa11x0_dma_desc *txd = to_sa11x0_dma_tx(tx);
+	unsigned long flags;
+
+	spin_lock_irqsave(&c->lock, flags);
+	c->chan.cookie += 1;
+	if (c->chan.cookie < 0)
+		c->chan.cookie = 1;
+	txd->tx.cookie = c->chan.cookie;
+
+	list_add_tail(&txd->node, &c->desc_submitted);
+	spin_unlock_irqrestore(&c->lock, flags);
+
+	dev_dbg(tx->chan->device->dev, "vchan %p: txd %p[%x]: submitted\n",
+		c, txd, txd->tx.cookie);
+
+	return txd->tx.cookie;
+}
+
+static struct dma_async_tx_descriptor *sa11x0_dma_prep_slave_sg(
+	struct dma_chan *chan, struct scatterlist *sg, unsigned int sglen,
+	enum dma_transfer_direction dir, unsigned long flags, void *context)
+{
+	struct sa11x0_dma_chan *c = to_sa11x0_dma_chan(chan);
+	struct sa11x0_dma_desc *txd;
+	struct scatterlist *sgent;
+	unsigned i, j = sglen;
+	size_t size = 0;
+
+	/* SA11x0 channels can only operate in their native direction */
+	if (dir != (c->ddar & DDAR_RW ? DMA_DEV_TO_MEM : DMA_MEM_TO_DEV)) {
+		dev_err(chan->device->dev, "vchan %p: bad DMA direction: DDAR:%08x dir:%u\n",
+			c, c->ddar, dir);
+		return NULL;
+	}
+
+	/* Do not allow zero-sized txds */
+	if (sglen == 0)
+		return NULL;
+
+	for_each_sg(sg, sgent, sglen, i) {
+		dma_addr_t addr = sg_dma_address(sgent);
+		unsigned int len = sg_dma_len(sgent);
+
+		if (len > DMA_MAX_SIZE)
+			j += DIV_ROUND_UP(len, DMA_MAX_SIZE & ~DMA_ALIGN) - 1;
+		if (addr & DMA_ALIGN) {
+			dev_dbg(chan->device->dev, "vchan %p: bad buffer alignment: %08x\n",
+				c, addr);
+			return NULL;
+		}
+	}
+
+	txd = kzalloc(sizeof(*txd) + j * sizeof(txd->sg[0]), GFP_ATOMIC);
+	if (!txd) {
+		dev_dbg(chan->device->dev, "vchan %p: kzalloc failed\n", c);
+		return NULL;
+	}
+
+	j = 0;
+	for_each_sg(sg, sgent, sglen, i) {
+		dma_addr_t addr = sg_dma_address(sgent);
+		unsigned len = sg_dma_len(sgent);
+
+		size += len;
+
+		do {
+			unsigned tlen = len;
+
+			/*
+			 * Check whether the transfer will fit.  If not, try
+			 * to split the transfer up such that we end up with
+			 * equal chunks - but make sure that we preserve the
+			 * alignment.  This avoids small segments.
+			 */
+			if (tlen > DMA_MAX_SIZE) {
+				unsigned mult = DIV_ROUND_UP(tlen,
+					DMA_MAX_SIZE & ~DMA_ALIGN);
+
+				tlen = (tlen / mult) & ~DMA_ALIGN;
+			}
+
+			txd->sg[j].addr = addr;
+			txd->sg[j].len = tlen;
+
+			addr += tlen;
+			len -= tlen;
+			j++;
+		} while (len);
+	}
+
+	dma_async_tx_descriptor_init(&txd->tx, &c->chan);
+	txd->tx.flags = flags;
+	txd->tx.tx_submit = sa11x0_dma_tx_submit;
+	txd->ddar = c->ddar;
+	txd->size = size;
+	txd->sglen = j;
+
+	dev_dbg(chan->device->dev, "vchan %p: txd %p: size %u nr %u\n",
+		c, txd, txd->size, txd->sglen);
+
+	return &txd->tx;
+}
+
+static int sa11x0_dma_slave_config(struct sa11x0_dma_chan *c, struct dma_slave_config *cfg)
+{
+	u32 ddar = c->ddar & ((0xf << 4) | DDAR_RW);
+	dma_addr_t addr;
+	enum dma_slave_buswidth width;
+	u32 maxburst;
+
+	if (ddar & DDAR_RW) {
+		addr = cfg->src_addr;
+		width = cfg->src_addr_width;
+		maxburst = cfg->src_maxburst;
+	} else {
+		addr = cfg->dst_addr;
+		width = cfg->dst_addr_width;
+		maxburst = cfg->dst_maxburst;
+	}
+
+	if ((width != DMA_SLAVE_BUSWIDTH_1_BYTE &&
+	     width != DMA_SLAVE_BUSWIDTH_2_BYTES) ||
+	    (maxburst != 4 && maxburst != 8))
+		return -EINVAL;
+
+	if (width == DMA_SLAVE_BUSWIDTH_2_BYTES)
+		ddar |= DDAR_DW;
+	if (maxburst == 8)
+		ddar |= DDAR_BS;
+
+	dev_dbg(c->chan.device->dev, "vchan %p: dma_slave_config addr %x width %u burst %u\n",
+		c, addr, width, maxburst);
+
+	c->ddar = ddar | (addr & 0xf0000000) | (addr & 0x003ffffc) << 6;
+
+	return 0;
+}
+
+static int sa11x0_dma_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
+	unsigned long arg)
+{
+	struct sa11x0_dma_chan *c = to_sa11x0_dma_chan(chan);
+	struct sa11x0_dma_dev *d = to_sa11x0_dma(chan->device);
+	struct sa11x0_dma_phy *p;
+	LIST_HEAD(head);
+	unsigned long flags;
+	int ret;
+
+	switch (cmd) {
+	case DMA_SLAVE_CONFIG:
+		return sa11x0_dma_slave_config(c, (struct dma_slave_config *)arg);
+
+	case DMA_TERMINATE_ALL:
+		dev_dbg(d->slave.dev, "vchan %p: terminate all\n", c);
+		/* Clear the tx descriptor lists */
+		spin_lock_irqsave(&c->lock, flags);
+		list_splice_tail_init(&c->desc_submitted, &head);
+		list_splice_tail_init(&c->desc_issued, &head);
+
+		p = c->phy;
+		if (p) {
+			struct sa11x0_dma_desc *txd, *txn;
+
+			dev_dbg(d->slave.dev, "pchan %u: terminating\n", p->num);
+			/* vchan is assigned to a pchan - stop the channel */
+			writel(DCSR_RUN | DCSR_IE |
+				DCSR_STRTA | DCSR_DONEA |
+				DCSR_STRTB | DCSR_DONEB,
+				p->base + DMA_DCSR_C);
+
+			list_for_each_entry_safe(txd, txn, &d->desc_complete, node)
+				if (txd->tx.chan == &c->chan)
+					list_move(&txd->node, &head);
+
+			if (p->txd_load) {
+				if (p->txd_load != p->txd_done)
+					list_add_tail(&p->txd_load->node, &head);
+				p->txd_load = NULL;
+			}
+			if (p->txd_done) {
+				list_add_tail(&p->txd_done->node, &head);
+				p->txd_done = NULL;
+			}
+			c->phy = NULL;
+			spin_lock(&d->lock);
+			p->vchan = NULL;
+			spin_unlock(&d->lock);
+			tasklet_schedule(&d->task);
+		}
+		spin_unlock_irqrestore(&c->lock, flags);
+		sa11x0_dma_desc_free(d, &head);
+		ret = 0;
+		break;
+
+	case DMA_PAUSE:
+		dev_dbg(d->slave.dev, "vchan %p: pause\n", c);
+		spin_lock_irqsave(&c->lock, flags);
+		if (c->status == DMA_IN_PROGRESS) {
+			c->status = DMA_PAUSED;
+
+			p = c->phy;
+			if (p) {
+				writel(DCSR_RUN | DCSR_IE, p->base + DMA_DCSR_C);
+			} else {
+				spin_lock(&d->lock);
+				list_del_init(&c->node);
+				spin_unlock(&d->lock);
+			}
+		}
+		spin_unlock_irqrestore(&c->lock, flags);
+		ret = 0;
+		break;
+
+	case DMA_RESUME:
+		dev_dbg(d->slave.dev, "vchan %p: resume\n", c);
+		spin_lock_irqsave(&c->lock, flags);
+		if (c->status == DMA_PAUSED) {
+			c->status = DMA_IN_PROGRESS;
+
+			p = c->phy;
+			if (p) {
+				writel(DCSR_RUN | DCSR_IE, p->base + DMA_DCSR_S);
+			} else if (!list_empty(&c->desc_issued)) {
+				spin_lock(&d->lock);
+				list_add_tail(&c->node, &d->chan_pending);
+				spin_unlock(&d->lock);
+			}
+		}
+		spin_unlock_irqrestore(&c->lock, flags);
+		ret = 0;
+		break;
+
+	default:
+		ret = -ENXIO;
+		break;
+	}
+
+	return ret;
+}
+
+struct sa11x0_dma_channel_desc {
+	u32 ddar;
+	const char *name;
+};
+
+#define CD(d1, d2) { .ddar = DDAR_##d1 | d2, .name = #d1 }
+static const struct sa11x0_dma_channel_desc chan_desc[] = {
+	CD(Ser0UDCTr, 0),
+	CD(Ser0UDCRc, DDAR_RW),
+	CD(Ser1SDLCTr, 0),
+	CD(Ser1SDLCRc, DDAR_RW),
+	CD(Ser1UARTTr, 0),
+	CD(Ser1UARTRc, DDAR_RW),
+	CD(Ser2ICPTr, 0),
+	CD(Ser2ICPRc, DDAR_RW),
+	CD(Ser3UARTTr, 0),
+	CD(Ser3UARTRc, DDAR_RW),
+	CD(Ser4MCP0Tr, 0),
+	CD(Ser4MCP0Rc, DDAR_RW),
+	CD(Ser4MCP1Tr, 0),
+	CD(Ser4MCP1Rc, DDAR_RW),
+	CD(Ser4SSPTr, 0),
+	CD(Ser4SSPRc, DDAR_RW),
+};
+
+static int __devinit sa11x0_dma_init_dmadev(struct dma_device *dmadev,
+	struct device *dev)
+{
+	unsigned i;
+
+	dmadev->chancnt = ARRAY_SIZE(chan_desc);
+	INIT_LIST_HEAD(&dmadev->channels);
+	dmadev->dev = dev;
+	dmadev->device_alloc_chan_resources = sa11x0_dma_alloc_chan_resources;
+	dmadev->device_free_chan_resources = sa11x0_dma_free_chan_resources;
+	dmadev->device_control = sa11x0_dma_control;
+	dmadev->device_tx_status = sa11x0_dma_tx_status;
+	dmadev->device_issue_pending = sa11x0_dma_issue_pending;
+
+	for (i = 0; i < dmadev->chancnt; i++) {
+		struct sa11x0_dma_chan *c;
+
+		c = kzalloc(sizeof(*c), GFP_KERNEL);
+		if (!c) {
+			dev_err(dev, "no memory for channel %u\n", i);
+			return -ENOMEM;
+		}
+
+		c->chan.device = dmadev;
+		c->status = DMA_IN_PROGRESS;
+		c->ddar = chan_desc[i].ddar;
+		c->name = chan_desc[i].name;
+		spin_lock_init(&c->lock);
+		INIT_LIST_HEAD(&c->desc_submitted);
+		INIT_LIST_HEAD(&c->desc_issued);
+		INIT_LIST_HEAD(&c->node);
+		list_add_tail(&c->chan.device_node, &dmadev->channels);
+	}
+
+	return dma_async_device_register(dmadev);
+}
+
+static int sa11x0_dma_request_irq(struct platform_device *pdev, int nr,
+	void *data)
+{
+	int irq = platform_get_irq(pdev, nr);
+
+	if (irq <= 0)
+		return -ENXIO;
+
+	return request_irq(irq, sa11x0_dma_irq, 0, dev_name(&pdev->dev), data);
+}
+
+static void sa11x0_dma_free_irq(struct platform_device *pdev, int nr,
+	void *data)
+{
+	int irq = platform_get_irq(pdev, nr);
+	if (irq > 0)
+		free_irq(irq, data);
+}
+
+static void sa11x0_dma_free_channels(struct dma_device *dmadev)
+{
+	struct sa11x0_dma_chan *c, *cn;
+
+	list_for_each_entry_safe(c, cn, &dmadev->channels, chan.device_node) {
+		list_del(&c->chan.device_node);
+		kfree(c);
+	}
+}
+
+static int __devinit sa11x0_dma_probe(struct platform_device *pdev)
+{
+	struct sa11x0_dma_dev *d;
+	struct resource *res;
+	unsigned i;
+	int ret;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res)
+		return -ENXIO;
+
+	d = kzalloc(sizeof(*d), GFP_KERNEL);
+	if (!d) {
+		ret = -ENOMEM;
+		goto err_alloc;
+	}
+
+	spin_lock_init(&d->lock);
+	INIT_LIST_HEAD(&d->chan_pending);
+	INIT_LIST_HEAD(&d->desc_complete);
+
+	d->base = ioremap(res->start, resource_size(res));
+	if (!d->base) {
+		ret = -ENOMEM;
+		goto err_ioremap;
+	}
+
+	tasklet_init(&d->task, sa11x0_dma_tasklet, (unsigned long)d);
+
+	for (i = 0; i < NR_PHY_CHAN; i++) {
+		struct sa11x0_dma_phy *p = &d->phy[i];
+
+		p->dev = d;
+		p->num = i;
+		p->base = d->base + i * DMA_SIZE;
+		writel_relaxed(DCSR_RUN | DCSR_IE | DCSR_ERROR |
+			DCSR_DONEA | DCSR_STRTA | DCSR_DONEB | DCSR_STRTB,
+			p->base + DMA_DCSR_C);
+		writel_relaxed(0, p->base + DMA_DDAR);
+
+		ret = sa11x0_dma_request_irq(pdev, i, p);
+		if (ret) {
+			while (i) {
+				i--;
+				sa11x0_dma_free_irq(pdev, i, &d->phy[i]);
+			}
+			goto err_irq;
+		}
+	}
+
+	dma_cap_set(DMA_SLAVE, d->slave.cap_mask);
+	d->slave.device_prep_slave_sg = sa11x0_dma_prep_slave_sg;
+	ret = sa11x0_dma_init_dmadev(&d->slave, &pdev->dev);
+	if (ret) {
+		dev_warn(d->slave.dev, "failed to register slave async device: %d\n",
+			ret);
+		goto err_slave_reg;
+	}
+
+	platform_set_drvdata(pdev, d);
+	return 0;
+
+ err_slave_reg:
+	sa11x0_dma_free_channels(&d->slave);
+	for (i = 0; i < NR_PHY_CHAN; i++)
+		sa11x0_dma_free_irq(pdev, i, &d->phy[i]);
+ err_irq:
+	tasklet_kill(&d->task);
+	iounmap(d->base);
+ err_ioremap:
+	kfree(d);
+ err_alloc:
+	return ret;
+}
+
+static int __devexit sa11x0_dma_remove(struct platform_device *pdev)
+{
+	struct sa11x0_dma_dev *d = platform_get_drvdata(pdev);
+	unsigned pch;
+
+	dma_async_device_unregister(&d->slave);
+
+	sa11x0_dma_free_channels(&d->slave);
+	for (pch = 0; pch < NR_PHY_CHAN; pch++)
+		sa11x0_dma_free_irq(pdev, pch, &d->phy[pch]);
+	tasklet_kill(&d->task);
+	iounmap(d->base);
+	kfree(d);
+
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int sa11x0_dma_suspend(struct device *dev)
+{
+	struct sa11x0_dma_dev *d = dev_get_drvdata(dev);
+	unsigned pch;
+
+	for (pch = 0; pch < NR_PHY_CHAN; pch++) {
+		struct sa11x0_dma_phy *p = &d->phy[pch];
+		u32 dcsr, saved_dcsr;
+
+		dcsr = saved_dcsr = readl_relaxed(p->base + DMA_DCSR_R);
+		if (dcsr & DCSR_RUN) {
+			writel(DCSR_RUN | DCSR_IE, p->base + DMA_DCSR_C);
+			dcsr = readl_relaxed(p->base + DMA_DCSR_R);
+		}
+
+		saved_dcsr &= DCSR_RUN | DCSR_IE;
+		if (dcsr & DCSR_BIU) {
+			p->dbs[0] = readl_relaxed(p->base + DMA_DBSB);
+			p->dbt[0] = readl_relaxed(p->base + DMA_DBTB);
+			p->dbs[1] = readl_relaxed(p->base + DMA_DBSA);
+			p->dbt[1] = readl_relaxed(p->base + DMA_DBTA);
+			saved_dcsr |= (dcsr & DCSR_STRTA ? DCSR_STRTB : 0) |
+				      (dcsr & DCSR_STRTB ? DCSR_STRTA : 0);
+		} else {
+			p->dbs[0] = readl_relaxed(p->base + DMA_DBSA);
+			p->dbt[0] = readl_relaxed(p->base + DMA_DBTA);
+			p->dbs[1] = readl_relaxed(p->base + DMA_DBSB);
+			p->dbt[1] = readl_relaxed(p->base + DMA_DBTB);
+			saved_dcsr |= dcsr & (DCSR_STRTA | DCSR_STRTB);
+		}
+		p->dcsr = saved_dcsr;
+
+		writel(DCSR_STRTA | DCSR_STRTB, p->base + DMA_DCSR_C);
+	}
+
+	return 0;
+}
+
+static int sa11x0_dma_resume(struct device *dev)
+{
+	struct sa11x0_dma_dev *d = dev_get_drvdata(dev);
+	unsigned pch;
+
+	for (pch = 0; pch < NR_PHY_CHAN; pch++) {
+		struct sa11x0_dma_phy *p = &d->phy[pch];
+		struct sa11x0_dma_desc *txd = NULL;
+		u32 dcsr = readl_relaxed(p->base + DMA_DCSR_R);
+
+		WARN_ON(dcsr & (DCSR_BIU | DCSR_STRTA | DCSR_STRTB | DCSR_RUN));
+
+		if (p->txd_done)
+			txd = p->txd_done;
+		else if (p->txd_load)
+			txd = p->txd_load;
+
+		if (!txd)
+			continue;
+
+		writel_relaxed(txd->ddar, p->base + DMA_DDAR);
+
+		writel_relaxed(p->dbs[0], p->base + DMA_DBSA);
+		writel_relaxed(p->dbt[0], p->base + DMA_DBTA);
+		writel_relaxed(p->dbs[1], p->base + DMA_DBSB);
+		writel_relaxed(p->dbt[1], p->base + DMA_DBTB);
+		writel_relaxed(p->dcsr, p->base + DMA_DCSR_S);
+	}
+
+	return 0;
+}
+#endif
+
+static const struct dev_pm_ops sa11x0_dma_pm_ops = {
+	.suspend_noirq = sa11x0_dma_suspend,
+	.resume_noirq = sa11x0_dma_resume,
+	.freeze_noirq = sa11x0_dma_suspend,
+	.thaw_noirq = sa11x0_dma_resume,
+	.poweroff_noirq = sa11x0_dma_suspend,
+	.restore_noirq = sa11x0_dma_resume,
+};
+
+static struct platform_driver sa11x0_dma_driver = {
+	.driver = {
+		.name	= "sa11x0-dma",
+		.owner	= THIS_MODULE,
+		.pm	= &sa11x0_dma_pm_ops,
+	},
+	.probe		= sa11x0_dma_probe,
+	.remove		= __devexit_p(sa11x0_dma_remove),
+};
+
+bool sa11x0_dma_filter_fn(struct dma_chan *chan, void *param)
+{
+	if (chan->device->dev->driver == &sa11x0_dma_driver.driver) {
+		struct sa11x0_dma_chan *c = to_sa11x0_dma_chan(chan);
+		const char *p = param;
+
+		return !strcmp(c->name, p);
+	}
+	return false;
+}
+EXPORT_SYMBOL(sa11x0_dma_filter_fn);
+
+static int __init sa11x0_dma_init(void)
+{
+	return platform_driver_register(&sa11x0_dma_driver);
+}
+subsys_initcall(sa11x0_dma_init);
+
+static void __exit sa11x0_dma_exit(void)
+{
+	platform_driver_unregister(&sa11x0_dma_driver);
+}
+module_exit(sa11x0_dma_exit);
+
+MODULE_AUTHOR("Russell King");
+MODULE_DESCRIPTION("SA-11x0 DMA driver");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS("platform:sa11x0-dma");
diff --git a/drivers/dma/shdma.c b/drivers/dma/shdma.c
index 81809c2b46ab..19d7a8d3975d 100644
--- a/drivers/dma/shdma.c
+++ b/drivers/dma/shdma.c
@@ -23,7 +23,6 @@
 #include <linux/interrupt.h>
 #include <linux/dmaengine.h>
 #include <linux/delay.h>
-#include <linux/dma-mapping.h>
 #include <linux/platform_device.h>
 #include <linux/pm_runtime.h>
 #include <linux/sh_dma.h>
@@ -31,6 +30,8 @@
 #include <linux/kdebug.h>
 #include <linux/spinlock.h>
 #include <linux/rculist.h>
+
+#include "dmaengine.h"
 #include "shdma.h"
 
 /* DMA descriptor control */
@@ -57,6 +58,15 @@ static LIST_HEAD(sh_dmae_devices);
 static unsigned long sh_dmae_slave_used[BITS_TO_LONGS(SH_DMA_SLAVE_NUMBER)];
 
 static void sh_dmae_chan_ld_cleanup(struct sh_dmae_chan *sh_chan, bool all);
+static void sh_chan_xfer_ld_queue(struct sh_dmae_chan *sh_chan);
+
+static void chclr_write(struct sh_dmae_chan *sh_dc, u32 data)
+{
+	struct sh_dmae_device *shdev = to_sh_dev(sh_dc);
+
+	__raw_writel(data, shdev->chan_reg +
+		     shdev->pdata->channel[sh_dc->id].chclr_offset);
+}
 
 static void sh_dmae_writel(struct sh_dmae_chan *sh_dc, u32 data, u32 reg)
 {
@@ -129,6 +139,15 @@ static int sh_dmae_rst(struct sh_dmae_device *shdev)
 
 	dmaor = dmaor_read(shdev) & ~(DMAOR_NMIF | DMAOR_AE | DMAOR_DME);
 
+	if (shdev->pdata->chclr_present) {
+		int i;
+		for (i = 0; i < shdev->pdata->channel_num; i++) {
+			struct sh_dmae_chan *sh_chan = shdev->chan[i];
+			if (sh_chan)
+				chclr_write(sh_chan, 0);
+		}
+	}
+
 	dmaor_write(shdev, dmaor | shdev->pdata->dmaor_init);
 
 	dmaor = dmaor_read(shdev);
@@ -139,6 +158,10 @@ static int sh_dmae_rst(struct sh_dmae_device *shdev)
 		dev_warn(shdev->common.dev, "Can't initialize DMAOR.\n");
 		return -EIO;
 	}
+	if (shdev->pdata->dmaor_init & ~dmaor)
+		dev_warn(shdev->common.dev,
+			 "DMAOR=0x%x hasn't latched the initial value 0x%x.\n",
+			 dmaor, shdev->pdata->dmaor_init);
 	return 0;
 }
 
@@ -259,8 +282,6 @@ static int dmae_set_dmars(struct sh_dmae_chan *sh_chan, u16 val)
 	return 0;
 }
 
-static void sh_chan_xfer_ld_queue(struct sh_dmae_chan *sh_chan);
-
 static dma_cookie_t sh_dmae_tx_submit(struct dma_async_tx_descriptor *tx)
 {
 	struct sh_desc *desc = tx_to_sh_desc(tx), *chunk, *last = desc, *c;
@@ -277,13 +298,7 @@ static dma_cookie_t sh_dmae_tx_submit(struct dma_async_tx_descriptor *tx)
 	else
 		power_up = false;
 
-	cookie = sh_chan->common.cookie;
-	cookie++;
-	if (cookie < 0)
-		cookie = 1;
-
-	sh_chan->common.cookie = cookie;
-	tx->cookie = cookie;
+	cookie = dma_cookie_assign(tx);
 
 	/* Mark all chunks of this descriptor as submitted, move to the queue */
 	list_for_each_entry_safe(chunk, c, desc->node.prev, node) {
@@ -340,6 +355,8 @@ static dma_cookie_t sh_dmae_tx_submit(struct dma_async_tx_descriptor *tx)
 				sh_chan_xfer_ld_queue(sh_chan);
 			sh_chan->pm_state = DMAE_PM_ESTABLISHED;
 		}
+	} else {
+		sh_chan->pm_state = DMAE_PM_PENDING;
 	}
 
 	spin_unlock_irq(&sh_chan->desc_lock);
@@ -479,19 +496,19 @@ static void sh_dmae_free_chan_resources(struct dma_chan *chan)
  * @sh_chan:	DMA channel
  * @flags:	DMA transfer flags
  * @dest:	destination DMA address, incremented when direction equals
- *		DMA_FROM_DEVICE or DMA_BIDIRECTIONAL
+ *		DMA_DEV_TO_MEM
  * @src:	source DMA address, incremented when direction equals
- *		DMA_TO_DEVICE or DMA_BIDIRECTIONAL
+ *		DMA_MEM_TO_DEV
  * @len:	DMA transfer length
  * @first:	if NULL, set to the current descriptor and cookie set to -EBUSY
  * @direction:	needed for slave DMA to decide which address to keep constant,
- *		equals DMA_BIDIRECTIONAL for MEMCPY
+ *		equals DMA_MEM_TO_MEM for MEMCPY
  * Returns 0 or an error
  * Locks: called with desc_lock held
  */
 static struct sh_desc *sh_dmae_add_desc(struct sh_dmae_chan *sh_chan,
 	unsigned long flags, dma_addr_t *dest, dma_addr_t *src, size_t *len,
-	struct sh_desc **first, enum dma_data_direction direction)
+	struct sh_desc **first, enum dma_transfer_direction direction)
 {
 	struct sh_desc *new;
 	size_t copy_size;
@@ -531,9 +548,9 @@ static struct sh_desc *sh_dmae_add_desc(struct sh_dmae_chan *sh_chan,
 	new->direction = direction;
 
 	*len -= copy_size;
-	if (direction == DMA_BIDIRECTIONAL || direction == DMA_TO_DEVICE)
+	if (direction == DMA_MEM_TO_MEM || direction == DMA_MEM_TO_DEV)
 		*src += copy_size;
-	if (direction == DMA_BIDIRECTIONAL || direction == DMA_FROM_DEVICE)
+	if (direction == DMA_MEM_TO_MEM || direction == DMA_DEV_TO_MEM)
 		*dest += copy_size;
 
 	return new;
@@ -546,12 +563,12 @@ static struct sh_desc *sh_dmae_add_desc(struct sh_dmae_chan *sh_chan,
  * converted to scatter-gather to guarantee consistent locking and a correct
  * list manipulation. For slave DMA direction carries the usual meaning, and,
  * logically, the SG list is RAM and the addr variable contains slave address,
- * e.g., the FIFO I/O register. For MEMCPY direction equals DMA_BIDIRECTIONAL
+ * e.g., the FIFO I/O register. For MEMCPY direction equals DMA_MEM_TO_MEM
  * and the SG list contains only one element and points at the source buffer.
  */
 static struct dma_async_tx_descriptor *sh_dmae_prep_sg(struct sh_dmae_chan *sh_chan,
 	struct scatterlist *sgl, unsigned int sg_len, dma_addr_t *addr,
-	enum dma_data_direction direction, unsigned long flags)
+	enum dma_transfer_direction direction, unsigned long flags)
 {
 	struct scatterlist *sg;
 	struct sh_desc *first = NULL, *new = NULL /* compiler... */;
@@ -592,7 +609,7 @@ static struct dma_async_tx_descriptor *sh_dmae_prep_sg(struct sh_dmae_chan *sh_c
 			dev_dbg(sh_chan->dev, "Add SG #%d@%p[%d], dma %llx\n",
 				i, sg, len, (unsigned long long)sg_addr);
 
-			if (direction == DMA_FROM_DEVICE)
+			if (direction == DMA_DEV_TO_MEM)
 				new = sh_dmae_add_desc(sh_chan, flags,
 						&sg_addr, addr, &len, &first,
 						direction);
@@ -646,13 +663,14 @@ static struct dma_async_tx_descriptor *sh_dmae_prep_memcpy(
 	sg_dma_address(&sg) = dma_src;
 	sg_dma_len(&sg) = len;
 
-	return sh_dmae_prep_sg(sh_chan, &sg, 1, &dma_dest, DMA_BIDIRECTIONAL,
+	return sh_dmae_prep_sg(sh_chan, &sg, 1, &dma_dest, DMA_MEM_TO_MEM,
 			       flags);
 }
 
 static struct dma_async_tx_descriptor *sh_dmae_prep_slave_sg(
 	struct dma_chan *chan, struct scatterlist *sgl, unsigned int sg_len,
-	enum dma_data_direction direction, unsigned long flags)
+	enum dma_transfer_direction direction, unsigned long flags,
+	void *context)
 {
 	struct sh_dmae_slave *param;
 	struct sh_dmae_chan *sh_chan;
@@ -743,12 +761,12 @@ static dma_async_tx_callback __ld_cleanup(struct sh_dmae_chan *sh_chan, bool all
 			cookie = tx->cookie;
 
 		if (desc->mark == DESC_COMPLETED && desc->chunks == 1) {
-			if (sh_chan->completed_cookie != desc->cookie - 1)
+			if (sh_chan->common.completed_cookie != desc->cookie - 1)
 				dev_dbg(sh_chan->dev,
 					"Completing cookie %d, expected %d\n",
 					desc->cookie,
-					sh_chan->completed_cookie + 1);
-			sh_chan->completed_cookie = desc->cookie;
+					sh_chan->common.completed_cookie + 1);
+			sh_chan->common.completed_cookie = desc->cookie;
 		}
 
 		/* Call callback on the last chunk */
@@ -802,7 +820,7 @@ static dma_async_tx_callback __ld_cleanup(struct sh_dmae_chan *sh_chan, bool all
 		 * Terminating and the loop completed normally: forgive
 		 * uncompleted cookies
 		 */
-		sh_chan->completed_cookie = sh_chan->common.cookie;
+		sh_chan->common.completed_cookie = sh_chan->common.cookie;
 
 	spin_unlock_irqrestore(&sh_chan->desc_lock, flags);
 
@@ -862,23 +880,14 @@ static enum dma_status sh_dmae_tx_status(struct dma_chan *chan,
 					struct dma_tx_state *txstate)
 {
 	struct sh_dmae_chan *sh_chan = to_sh_chan(chan);
-	dma_cookie_t last_used;
-	dma_cookie_t last_complete;
 	enum dma_status status;
 	unsigned long flags;
 
 	sh_dmae_chan_ld_cleanup(sh_chan, false);
 
-	/* First read completed cookie to avoid a skew */
-	last_complete = sh_chan->completed_cookie;
-	rmb();
-	last_used = chan->cookie;
-	BUG_ON(last_complete < 0);
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
-
 	spin_lock_irqsave(&sh_chan->desc_lock, flags);
 
-	status = dma_async_is_complete(cookie, last_complete, last_used);
+	status = dma_cookie_status(chan, cookie, txstate);
 
 	/*
 	 * If we don't find cookie on the queue, it has been aborted and we have
@@ -996,7 +1005,7 @@ static void dmae_do_tasklet(unsigned long data)
 	spin_lock_irq(&sh_chan->desc_lock);
 	list_for_each_entry(desc, &sh_chan->ld_queue, node) {
 		if (desc->mark == DESC_SUBMITTED &&
-		    ((desc->direction == DMA_FROM_DEVICE &&
+		    ((desc->direction == DMA_DEV_TO_MEM &&
 		      (desc->hw.dar + desc->hw.tcr) == dar_buf) ||
 		     (desc->hw.sar + desc->hw.tcr) == sar_buf)) {
 			dev_dbg(sh_chan->dev, "done #%d@%p dst %u\n",
@@ -1081,6 +1090,7 @@ static int __devinit sh_dmae_chan_probe(struct sh_dmae_device *shdev, int id,
 
 	/* reference struct dma_device */
 	new_sh_chan->common.device = &shdev->common;
+	dma_cookie_init(&new_sh_chan->common);
 
 	new_sh_chan->dev = shdev->common.dev;
 	new_sh_chan->id = id;
@@ -1225,6 +1235,8 @@ static int __init sh_dmae_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, shdev);
 
+	shdev->common.dev = &pdev->dev;
+
 	pm_runtime_enable(&pdev->dev);
 	pm_runtime_get_sync(&pdev->dev);
 
@@ -1239,7 +1251,8 @@ static int __init sh_dmae_probe(struct platform_device *pdev)
 
 	INIT_LIST_HEAD(&shdev->common.channels);
 
-	dma_cap_set(DMA_MEMCPY, shdev->common.cap_mask);
+	if (!pdata->slave_only)
+		dma_cap_set(DMA_MEMCPY, shdev->common.cap_mask);
 	if (pdata->slave && pdata->slave_num)
 		dma_cap_set(DMA_SLAVE, shdev->common.cap_mask);
 
@@ -1254,7 +1267,6 @@ static int __init sh_dmae_probe(struct platform_device *pdev)
 	shdev->common.device_prep_slave_sg = sh_dmae_prep_slave_sg;
 	shdev->common.device_control = sh_dmae_control;
 
-	shdev->common.dev = &pdev->dev;
 	/* Default transfer size of 32 bytes requires 32-byte alignment */
 	shdev->common.copy_align = LOG2_DEFAULT_XFER_SIZE;
 
@@ -1435,22 +1447,17 @@ static int sh_dmae_runtime_resume(struct device *dev)
 #ifdef CONFIG_PM
 static int sh_dmae_suspend(struct device *dev)
 {
-	struct sh_dmae_device *shdev = dev_get_drvdata(dev);
-	int i;
-
-	for (i = 0; i < shdev->pdata->channel_num; i++) {
-		struct sh_dmae_chan *sh_chan = shdev->chan[i];
-		if (sh_chan->descs_allocated)
-			sh_chan->pm_error = pm_runtime_put_sync(dev);
-	}
-
 	return 0;
 }
 
 static int sh_dmae_resume(struct device *dev)
 {
 	struct sh_dmae_device *shdev = dev_get_drvdata(dev);
-	int i;
+	int i, ret;
+
+	ret = sh_dmae_rst(shdev);
+	if (ret < 0)
+		dev_err(dev, "Failed to reset!\n");
 
 	for (i = 0; i < shdev->pdata->channel_num; i++) {
 		struct sh_dmae_chan *sh_chan = shdev->chan[i];
@@ -1459,9 +1466,6 @@ static int sh_dmae_resume(struct device *dev)
 		if (!sh_chan->descs_allocated)
 			continue;
 
-		if (!sh_chan->pm_error)
-			pm_runtime_get_sync(dev);
-
 		if (param) {
 			const struct sh_dmae_slave_config *cfg = param->config;
 			dmae_set_dmars(sh_chan, cfg->mid_rid);
diff --git a/drivers/dma/shdma.h b/drivers/dma/shdma.h
index 2b55a276dc5b..0b1d2c105f02 100644
--- a/drivers/dma/shdma.h
+++ b/drivers/dma/shdma.h
@@ -30,7 +30,6 @@ enum dmae_pm_state {
 };
 
 struct sh_dmae_chan {
-	dma_cookie_t completed_cookie;	/* The maximum cookie completed */
 	spinlock_t desc_lock;		/* Descriptor operation lock */
 	struct list_head ld_queue;	/* Link descriptors queue */
 	struct list_head ld_free;	/* Link descriptors free */
diff --git a/drivers/dma/sirf-dma.c b/drivers/dma/sirf-dma.c
new file mode 100644
index 000000000000..434ad31174f2
--- /dev/null
+++ b/drivers/dma/sirf-dma.c
@@ -0,0 +1,698 @@
+/*
+ * DMA controller driver for CSR SiRFprimaII
+ *
+ * Copyright (c) 2011 Cambridge Silicon Radio Limited, a CSR plc group company.
+ *
+ * Licensed under GPLv2 or later.
+ */
+
+#include <linux/module.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/slab.h>
+#include <linux/of_irq.h>
+#include <linux/of_address.h>
+#include <linux/of_device.h>
+#include <linux/of_platform.h>
+#include <linux/sirfsoc_dma.h>
+
+#include "dmaengine.h"
+
+#define SIRFSOC_DMA_DESCRIPTORS                 16
+#define SIRFSOC_DMA_CHANNELS                    16
+
+#define SIRFSOC_DMA_CH_ADDR                     0x00
+#define SIRFSOC_DMA_CH_XLEN                     0x04
+#define SIRFSOC_DMA_CH_YLEN                     0x08
+#define SIRFSOC_DMA_CH_CTRL                     0x0C
+
+#define SIRFSOC_DMA_WIDTH_0                     0x100
+#define SIRFSOC_DMA_CH_VALID                    0x140
+#define SIRFSOC_DMA_CH_INT                      0x144
+#define SIRFSOC_DMA_INT_EN                      0x148
+#define SIRFSOC_DMA_CH_LOOP_CTRL                0x150
+
+#define SIRFSOC_DMA_MODE_CTRL_BIT               4
+#define SIRFSOC_DMA_DIR_CTRL_BIT                5
+
+/* xlen and dma_width register is in 4 bytes boundary */
+#define SIRFSOC_DMA_WORD_LEN			4
+
+struct sirfsoc_dma_desc {
+	struct dma_async_tx_descriptor	desc;
+	struct list_head		node;
+
+	/* SiRFprimaII 2D-DMA parameters */
+
+	int             xlen;           /* DMA xlen */
+	int             ylen;           /* DMA ylen */
+	int             width;          /* DMA width */
+	int             dir;
+	bool            cyclic;         /* is loop DMA? */
+	u32             addr;		/* DMA buffer address */
+};
+
+struct sirfsoc_dma_chan {
+	struct dma_chan			chan;
+	struct list_head		free;
+	struct list_head		prepared;
+	struct list_head		queued;
+	struct list_head		active;
+	struct list_head		completed;
+	unsigned long			happened_cyclic;
+	unsigned long			completed_cyclic;
+
+	/* Lock for this structure */
+	spinlock_t			lock;
+
+	int				mode;
+};
+
+struct sirfsoc_dma {
+	struct dma_device		dma;
+	struct tasklet_struct		tasklet;
+	struct sirfsoc_dma_chan		channels[SIRFSOC_DMA_CHANNELS];
+	void __iomem			*base;
+	int				irq;
+};
+
+#define DRV_NAME	"sirfsoc_dma"
+
+/* Convert struct dma_chan to struct sirfsoc_dma_chan */
+static inline
+struct sirfsoc_dma_chan *dma_chan_to_sirfsoc_dma_chan(struct dma_chan *c)
+{
+	return container_of(c, struct sirfsoc_dma_chan, chan);
+}
+
+/* Convert struct dma_chan to struct sirfsoc_dma */
+static inline struct sirfsoc_dma *dma_chan_to_sirfsoc_dma(struct dma_chan *c)
+{
+	struct sirfsoc_dma_chan *schan = dma_chan_to_sirfsoc_dma_chan(c);
+	return container_of(schan, struct sirfsoc_dma, channels[c->chan_id]);
+}
+
+/* Execute all queued DMA descriptors */
+static void sirfsoc_dma_execute(struct sirfsoc_dma_chan *schan)
+{
+	struct sirfsoc_dma *sdma = dma_chan_to_sirfsoc_dma(&schan->chan);
+	int cid = schan->chan.chan_id;
+	struct sirfsoc_dma_desc *sdesc = NULL;
+
+	/*
+	 * lock has been held by functions calling this, so we don't hold
+	 * lock again
+	 */
+
+	sdesc = list_first_entry(&schan->queued, struct sirfsoc_dma_desc,
+		node);
+	/* Move the first queued descriptor to active list */
+	list_move_tail(&schan->queued, &schan->active);
+
+	/* Start the DMA transfer */
+	writel_relaxed(sdesc->width, sdma->base + SIRFSOC_DMA_WIDTH_0 +
+		cid * 4);
+	writel_relaxed(cid | (schan->mode << SIRFSOC_DMA_MODE_CTRL_BIT) |
+		(sdesc->dir << SIRFSOC_DMA_DIR_CTRL_BIT),
+		sdma->base + cid * 0x10 + SIRFSOC_DMA_CH_CTRL);
+	writel_relaxed(sdesc->xlen, sdma->base + cid * 0x10 +
+		SIRFSOC_DMA_CH_XLEN);
+	writel_relaxed(sdesc->ylen, sdma->base + cid * 0x10 +
+		SIRFSOC_DMA_CH_YLEN);
+	writel_relaxed(readl_relaxed(sdma->base + SIRFSOC_DMA_INT_EN) |
+		(1 << cid), sdma->base + SIRFSOC_DMA_INT_EN);
+
+	/*
+	 * writel has an implict memory write barrier to make sure data is
+	 * flushed into memory before starting DMA
+	 */
+	writel(sdesc->addr >> 2, sdma->base + cid * 0x10 + SIRFSOC_DMA_CH_ADDR);
+
+	if (sdesc->cyclic) {
+		writel((1 << cid) | 1 << (cid + 16) |
+			readl_relaxed(sdma->base + SIRFSOC_DMA_CH_LOOP_CTRL),
+			sdma->base + SIRFSOC_DMA_CH_LOOP_CTRL);
+		schan->happened_cyclic = schan->completed_cyclic = 0;
+	}
+}
+
+/* Interrupt handler */
+static irqreturn_t sirfsoc_dma_irq(int irq, void *data)
+{
+	struct sirfsoc_dma *sdma = data;
+	struct sirfsoc_dma_chan *schan;
+	struct sirfsoc_dma_desc *sdesc = NULL;
+	u32 is;
+	int ch;
+
+	is = readl(sdma->base + SIRFSOC_DMA_CH_INT);
+	while ((ch = fls(is) - 1) >= 0) {
+		is &= ~(1 << ch);
+		writel_relaxed(1 << ch, sdma->base + SIRFSOC_DMA_CH_INT);
+		schan = &sdma->channels[ch];
+
+		spin_lock(&schan->lock);
+
+		sdesc = list_first_entry(&schan->active, struct sirfsoc_dma_desc,
+			node);
+		if (!sdesc->cyclic) {
+			/* Execute queued descriptors */
+			list_splice_tail_init(&schan->active, &schan->completed);
+			if (!list_empty(&schan->queued))
+				sirfsoc_dma_execute(schan);
+		} else
+			schan->happened_cyclic++;
+
+		spin_unlock(&schan->lock);
+	}
+
+	/* Schedule tasklet */
+	tasklet_schedule(&sdma->tasklet);
+
+	return IRQ_HANDLED;
+}
+
+/* process completed descriptors */
+static void sirfsoc_dma_process_completed(struct sirfsoc_dma *sdma)
+{
+	dma_cookie_t last_cookie = 0;
+	struct sirfsoc_dma_chan *schan;
+	struct sirfsoc_dma_desc *sdesc;
+	struct dma_async_tx_descriptor *desc;
+	unsigned long flags;
+	unsigned long happened_cyclic;
+	LIST_HEAD(list);
+	int i;
+
+	for (i = 0; i < sdma->dma.chancnt; i++) {
+		schan = &sdma->channels[i];
+
+		/* Get all completed descriptors */
+		spin_lock_irqsave(&schan->lock, flags);
+		if (!list_empty(&schan->completed)) {
+			list_splice_tail_init(&schan->completed, &list);
+			spin_unlock_irqrestore(&schan->lock, flags);
+
+			/* Execute callbacks and run dependencies */
+			list_for_each_entry(sdesc, &list, node) {
+				desc = &sdesc->desc;
+
+				if (desc->callback)
+					desc->callback(desc->callback_param);
+
+				last_cookie = desc->cookie;
+				dma_run_dependencies(desc);
+			}
+
+			/* Free descriptors */
+			spin_lock_irqsave(&schan->lock, flags);
+			list_splice_tail_init(&list, &schan->free);
+			schan->chan.completed_cookie = last_cookie;
+			spin_unlock_irqrestore(&schan->lock, flags);
+		} else {
+			/* for cyclic channel, desc is always in active list */
+			sdesc = list_first_entry(&schan->active, struct sirfsoc_dma_desc,
+				node);
+
+			if (!sdesc || (sdesc && !sdesc->cyclic)) {
+				/* without active cyclic DMA */
+				spin_unlock_irqrestore(&schan->lock, flags);
+				continue;
+			}
+
+			/* cyclic DMA */
+			happened_cyclic = schan->happened_cyclic;
+			spin_unlock_irqrestore(&schan->lock, flags);
+
+			desc = &sdesc->desc;
+			while (happened_cyclic != schan->completed_cyclic) {
+				if (desc->callback)
+					desc->callback(desc->callback_param);
+				schan->completed_cyclic++;
+			}
+		}
+	}
+}
+
+/* DMA Tasklet */
+static void sirfsoc_dma_tasklet(unsigned long data)
+{
+	struct sirfsoc_dma *sdma = (void *)data;
+
+	sirfsoc_dma_process_completed(sdma);
+}
+
+/* Submit descriptor to hardware */
+static dma_cookie_t sirfsoc_dma_tx_submit(struct dma_async_tx_descriptor *txd)
+{
+	struct sirfsoc_dma_chan *schan = dma_chan_to_sirfsoc_dma_chan(txd->chan);
+	struct sirfsoc_dma_desc *sdesc;
+	unsigned long flags;
+	dma_cookie_t cookie;
+
+	sdesc = container_of(txd, struct sirfsoc_dma_desc, desc);
+
+	spin_lock_irqsave(&schan->lock, flags);
+
+	/* Move descriptor to queue */
+	list_move_tail(&sdesc->node, &schan->queued);
+
+	cookie = dma_cookie_assign(txd);
+
+	spin_unlock_irqrestore(&schan->lock, flags);
+
+	return cookie;
+}
+
+static int sirfsoc_dma_slave_config(struct sirfsoc_dma_chan *schan,
+	struct dma_slave_config *config)
+{
+	unsigned long flags;
+
+	if ((config->src_addr_width != DMA_SLAVE_BUSWIDTH_4_BYTES) ||
+		(config->dst_addr_width != DMA_SLAVE_BUSWIDTH_4_BYTES))
+		return -EINVAL;
+
+	spin_lock_irqsave(&schan->lock, flags);
+	schan->mode = (config->src_maxburst == 4 ? 1 : 0);
+	spin_unlock_irqrestore(&schan->lock, flags);
+
+	return 0;
+}
+
+static int sirfsoc_dma_terminate_all(struct sirfsoc_dma_chan *schan)
+{
+	struct sirfsoc_dma *sdma = dma_chan_to_sirfsoc_dma(&schan->chan);
+	int cid = schan->chan.chan_id;
+	unsigned long flags;
+
+	writel_relaxed(readl_relaxed(sdma->base + SIRFSOC_DMA_INT_EN) &
+		~(1 << cid), sdma->base + SIRFSOC_DMA_INT_EN);
+	writel_relaxed(1 << cid, sdma->base + SIRFSOC_DMA_CH_VALID);
+
+	writel_relaxed(readl_relaxed(sdma->base + SIRFSOC_DMA_CH_LOOP_CTRL)
+		& ~((1 << cid) | 1 << (cid + 16)),
+			sdma->base + SIRFSOC_DMA_CH_LOOP_CTRL);
+
+	spin_lock_irqsave(&schan->lock, flags);
+	list_splice_tail_init(&schan->active, &schan->free);
+	list_splice_tail_init(&schan->queued, &schan->free);
+	spin_unlock_irqrestore(&schan->lock, flags);
+
+	return 0;
+}
+
+static int sirfsoc_dma_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
+	unsigned long arg)
+{
+	struct dma_slave_config *config;
+	struct sirfsoc_dma_chan *schan = dma_chan_to_sirfsoc_dma_chan(chan);
+
+	switch (cmd) {
+	case DMA_TERMINATE_ALL:
+		return sirfsoc_dma_terminate_all(schan);
+	case DMA_SLAVE_CONFIG:
+		config = (struct dma_slave_config *)arg;
+		return sirfsoc_dma_slave_config(schan, config);
+
+	default:
+		break;
+	}
+
+	return -ENOSYS;
+}
+
+/* Alloc channel resources */
+static int sirfsoc_dma_alloc_chan_resources(struct dma_chan *chan)
+{
+	struct sirfsoc_dma *sdma = dma_chan_to_sirfsoc_dma(chan);
+	struct sirfsoc_dma_chan *schan = dma_chan_to_sirfsoc_dma_chan(chan);
+	struct sirfsoc_dma_desc *sdesc;
+	unsigned long flags;
+	LIST_HEAD(descs);
+	int i;
+
+	/* Alloc descriptors for this channel */
+	for (i = 0; i < SIRFSOC_DMA_DESCRIPTORS; i++) {
+		sdesc = kzalloc(sizeof(*sdesc), GFP_KERNEL);
+		if (!sdesc) {
+			dev_notice(sdma->dma.dev, "Memory allocation error. "
+				"Allocated only %u descriptors\n", i);
+			break;
+		}
+
+		dma_async_tx_descriptor_init(&sdesc->desc, chan);
+		sdesc->desc.flags = DMA_CTRL_ACK;
+		sdesc->desc.tx_submit = sirfsoc_dma_tx_submit;
+
+		list_add_tail(&sdesc->node, &descs);
+	}
+
+	/* Return error only if no descriptors were allocated */
+	if (i == 0)
+		return -ENOMEM;
+
+	spin_lock_irqsave(&schan->lock, flags);
+
+	list_splice_tail_init(&descs, &schan->free);
+	spin_unlock_irqrestore(&schan->lock, flags);
+
+	return i;
+}
+
+/* Free channel resources */
+static void sirfsoc_dma_free_chan_resources(struct dma_chan *chan)
+{
+	struct sirfsoc_dma_chan *schan = dma_chan_to_sirfsoc_dma_chan(chan);
+	struct sirfsoc_dma_desc *sdesc, *tmp;
+	unsigned long flags;
+	LIST_HEAD(descs);
+
+	spin_lock_irqsave(&schan->lock, flags);
+
+	/* Channel must be idle */
+	BUG_ON(!list_empty(&schan->prepared));
+	BUG_ON(!list_empty(&schan->queued));
+	BUG_ON(!list_empty(&schan->active));
+	BUG_ON(!list_empty(&schan->completed));
+
+	/* Move data */
+	list_splice_tail_init(&schan->free, &descs);
+
+	spin_unlock_irqrestore(&schan->lock, flags);
+
+	/* Free descriptors */
+	list_for_each_entry_safe(sdesc, tmp, &descs, node)
+		kfree(sdesc);
+}
+
+/* Send pending descriptor to hardware */
+static void sirfsoc_dma_issue_pending(struct dma_chan *chan)
+{
+	struct sirfsoc_dma_chan *schan = dma_chan_to_sirfsoc_dma_chan(chan);
+	unsigned long flags;
+
+	spin_lock_irqsave(&schan->lock, flags);
+
+	if (list_empty(&schan->active) && !list_empty(&schan->queued))
+		sirfsoc_dma_execute(schan);
+
+	spin_unlock_irqrestore(&schan->lock, flags);
+}
+
+/* Check request completion status */
+static enum dma_status
+sirfsoc_dma_tx_status(struct dma_chan *chan, dma_cookie_t cookie,
+	struct dma_tx_state *txstate)
+{
+	struct sirfsoc_dma_chan *schan = dma_chan_to_sirfsoc_dma_chan(chan);
+	unsigned long flags;
+	enum dma_status ret;
+
+	spin_lock_irqsave(&schan->lock, flags);
+	ret = dma_cookie_status(chan, cookie, txstate);
+	spin_unlock_irqrestore(&schan->lock, flags);
+
+	return ret;
+}
+
+static struct dma_async_tx_descriptor *sirfsoc_dma_prep_interleaved(
+	struct dma_chan *chan, struct dma_interleaved_template *xt,
+	unsigned long flags)
+{
+	struct sirfsoc_dma *sdma = dma_chan_to_sirfsoc_dma(chan);
+	struct sirfsoc_dma_chan *schan = dma_chan_to_sirfsoc_dma_chan(chan);
+	struct sirfsoc_dma_desc *sdesc = NULL;
+	unsigned long iflags;
+	int ret;
+
+	if ((xt->dir != DMA_MEM_TO_DEV) || (xt->dir != DMA_DEV_TO_MEM)) {
+		ret = -EINVAL;
+		goto err_dir;
+	}
+
+	/* Get free descriptor */
+	spin_lock_irqsave(&schan->lock, iflags);
+	if (!list_empty(&schan->free)) {
+		sdesc = list_first_entry(&schan->free, struct sirfsoc_dma_desc,
+			node);
+		list_del(&sdesc->node);
+	}
+	spin_unlock_irqrestore(&schan->lock, iflags);
+
+	if (!sdesc) {
+		/* try to free completed descriptors */
+		sirfsoc_dma_process_completed(sdma);
+		ret = 0;
+		goto no_desc;
+	}
+
+	/* Place descriptor in prepared list */
+	spin_lock_irqsave(&schan->lock, iflags);
+
+	/*
+	 * Number of chunks in a frame can only be 1 for prima2
+	 * and ylen (number of frame - 1) must be at least 0
+	 */
+	if ((xt->frame_size == 1) && (xt->numf > 0)) {
+		sdesc->cyclic = 0;
+		sdesc->xlen = xt->sgl[0].size / SIRFSOC_DMA_WORD_LEN;
+		sdesc->width = (xt->sgl[0].size + xt->sgl[0].icg) /
+				SIRFSOC_DMA_WORD_LEN;
+		sdesc->ylen = xt->numf - 1;
+		if (xt->dir == DMA_MEM_TO_DEV) {
+			sdesc->addr = xt->src_start;
+			sdesc->dir = 1;
+		} else {
+			sdesc->addr = xt->dst_start;
+			sdesc->dir = 0;
+		}
+
+		list_add_tail(&sdesc->node, &schan->prepared);
+	} else {
+		pr_err("sirfsoc DMA Invalid xfer\n");
+		ret = -EINVAL;
+		goto err_xfer;
+	}
+	spin_unlock_irqrestore(&schan->lock, iflags);
+
+	return &sdesc->desc;
+err_xfer:
+	spin_unlock_irqrestore(&schan->lock, iflags);
+no_desc:
+err_dir:
+	return ERR_PTR(ret);
+}
+
+static struct dma_async_tx_descriptor *
+sirfsoc_dma_prep_cyclic(struct dma_chan *chan, dma_addr_t addr,
+	size_t buf_len, size_t period_len,
+	enum dma_transfer_direction direction, void *context)
+{
+	struct sirfsoc_dma_chan *schan = dma_chan_to_sirfsoc_dma_chan(chan);
+	struct sirfsoc_dma_desc *sdesc = NULL;
+	unsigned long iflags;
+
+	/*
+	 * we only support cycle transfer with 2 period
+	 * If the X-length is set to 0, it would be the loop mode.
+	 * The DMA address keeps increasing until reaching the end of a loop
+	 * area whose size is defined by (DMA_WIDTH x (Y_LENGTH + 1)). Then
+	 * the DMA address goes back to the beginning of this area.
+	 * In loop mode, the DMA data region is divided into two parts, BUFA
+	 * and BUFB. DMA controller generates interrupts twice in each loop:
+	 * when the DMA address reaches the end of BUFA or the end of the
+	 * BUFB
+	 */
+	if (buf_len !=  2 * period_len)
+		return ERR_PTR(-EINVAL);
+
+	/* Get free descriptor */
+	spin_lock_irqsave(&schan->lock, iflags);
+	if (!list_empty(&schan->free)) {
+		sdesc = list_first_entry(&schan->free, struct sirfsoc_dma_desc,
+			node);
+		list_del(&sdesc->node);
+	}
+	spin_unlock_irqrestore(&schan->lock, iflags);
+
+	if (!sdesc)
+		return 0;
+
+	/* Place descriptor in prepared list */
+	spin_lock_irqsave(&schan->lock, iflags);
+	sdesc->addr = addr;
+	sdesc->cyclic = 1;
+	sdesc->xlen = 0;
+	sdesc->ylen = buf_len / SIRFSOC_DMA_WORD_LEN - 1;
+	sdesc->width = 1;
+	list_add_tail(&sdesc->node, &schan->prepared);
+	spin_unlock_irqrestore(&schan->lock, iflags);
+
+	return &sdesc->desc;
+}
+
+/*
+ * The DMA controller consists of 16 independent DMA channels.
+ * Each channel is allocated to a different function
+ */
+bool sirfsoc_dma_filter_id(struct dma_chan *chan, void *chan_id)
+{
+	unsigned int ch_nr = (unsigned int) chan_id;
+
+	if (ch_nr == chan->chan_id +
+		chan->device->dev_id * SIRFSOC_DMA_CHANNELS)
+		return true;
+
+	return false;
+}
+EXPORT_SYMBOL(sirfsoc_dma_filter_id);
+
+static int __devinit sirfsoc_dma_probe(struct platform_device *op)
+{
+	struct device_node *dn = op->dev.of_node;
+	struct device *dev = &op->dev;
+	struct dma_device *dma;
+	struct sirfsoc_dma *sdma;
+	struct sirfsoc_dma_chan *schan;
+	struct resource res;
+	ulong regs_start, regs_size;
+	u32 id;
+	int ret, i;
+
+	sdma = devm_kzalloc(dev, sizeof(*sdma), GFP_KERNEL);
+	if (!sdma) {
+		dev_err(dev, "Memory exhausted!\n");
+		return -ENOMEM;
+	}
+
+	if (of_property_read_u32(dn, "cell-index", &id)) {
+		dev_err(dev, "Fail to get DMAC index\n");
+		ret = -ENODEV;
+		goto free_mem;
+	}
+
+	sdma->irq = irq_of_parse_and_map(dn, 0);
+	if (sdma->irq == NO_IRQ) {
+		dev_err(dev, "Error mapping IRQ!\n");
+		ret = -EINVAL;
+		goto free_mem;
+	}
+
+	ret = of_address_to_resource(dn, 0, &res);
+	if (ret) {
+		dev_err(dev, "Error parsing memory region!\n");
+		goto free_mem;
+	}
+
+	regs_start = res.start;
+	regs_size = resource_size(&res);
+
+	sdma->base = devm_ioremap(dev, regs_start, regs_size);
+	if (!sdma->base) {
+		dev_err(dev, "Error mapping memory region!\n");
+		ret = -ENOMEM;
+		goto irq_dispose;
+	}
+
+	ret = devm_request_irq(dev, sdma->irq, &sirfsoc_dma_irq, 0, DRV_NAME,
+		sdma);
+	if (ret) {
+		dev_err(dev, "Error requesting IRQ!\n");
+		ret = -EINVAL;
+		goto unmap_mem;
+	}
+
+	dma = &sdma->dma;
+	dma->dev = dev;
+	dma->chancnt = SIRFSOC_DMA_CHANNELS;
+
+	dma->device_alloc_chan_resources = sirfsoc_dma_alloc_chan_resources;
+	dma->device_free_chan_resources = sirfsoc_dma_free_chan_resources;
+	dma->device_issue_pending = sirfsoc_dma_issue_pending;
+	dma->device_control = sirfsoc_dma_control;
+	dma->device_tx_status = sirfsoc_dma_tx_status;
+	dma->device_prep_interleaved_dma = sirfsoc_dma_prep_interleaved;
+	dma->device_prep_dma_cyclic = sirfsoc_dma_prep_cyclic;
+
+	INIT_LIST_HEAD(&dma->channels);
+	dma_cap_set(DMA_SLAVE, dma->cap_mask);
+	dma_cap_set(DMA_CYCLIC, dma->cap_mask);
+	dma_cap_set(DMA_INTERLEAVE, dma->cap_mask);
+	dma_cap_set(DMA_PRIVATE, dma->cap_mask);
+
+	for (i = 0; i < dma->chancnt; i++) {
+		schan = &sdma->channels[i];
+
+		schan->chan.device = dma;
+		dma_cookie_init(&schan->chan);
+
+		INIT_LIST_HEAD(&schan->free);
+		INIT_LIST_HEAD(&schan->prepared);
+		INIT_LIST_HEAD(&schan->queued);
+		INIT_LIST_HEAD(&schan->active);
+		INIT_LIST_HEAD(&schan->completed);
+
+		spin_lock_init(&schan->lock);
+		list_add_tail(&schan->chan.device_node, &dma->channels);
+	}
+
+	tasklet_init(&sdma->tasklet, sirfsoc_dma_tasklet, (unsigned long)sdma);
+
+	/* Register DMA engine */
+	dev_set_drvdata(dev, sdma);
+	ret = dma_async_device_register(dma);
+	if (ret)
+		goto free_irq;
+
+	dev_info(dev, "initialized SIRFSOC DMAC driver\n");
+
+	return 0;
+
+free_irq:
+	devm_free_irq(dev, sdma->irq, sdma);
+irq_dispose:
+	irq_dispose_mapping(sdma->irq);
+unmap_mem:
+	iounmap(sdma->base);
+free_mem:
+	devm_kfree(dev, sdma);
+	return ret;
+}
+
+static int __devexit sirfsoc_dma_remove(struct platform_device *op)
+{
+	struct device *dev = &op->dev;
+	struct sirfsoc_dma *sdma = dev_get_drvdata(dev);
+
+	dma_async_device_unregister(&sdma->dma);
+	devm_free_irq(dev, sdma->irq, sdma);
+	irq_dispose_mapping(sdma->irq);
+	iounmap(sdma->base);
+	devm_kfree(dev, sdma);
+	return 0;
+}
+
+static struct of_device_id sirfsoc_dma_match[] = {
+	{ .compatible = "sirf,prima2-dmac", },
+	{},
+};
+
+static struct platform_driver sirfsoc_dma_driver = {
+	.probe		= sirfsoc_dma_probe,
+	.remove		= __devexit_p(sirfsoc_dma_remove),
+	.driver = {
+		.name = DRV_NAME,
+		.owner = THIS_MODULE,
+		.of_match_table	= sirfsoc_dma_match,
+	},
+};
+
+module_platform_driver(sirfsoc_dma_driver);
+
+MODULE_AUTHOR("Rongjun Ying <rongjun.ying@csr.com>, "
+	"Barry Song <baohua.song@csr.com>");
+MODULE_DESCRIPTION("SIRFSOC DMA control driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/dma/ste_dma40.c b/drivers/dma/ste_dma40.c
index 13259cad0ceb..bdd41d4bfa8d 100644
--- a/drivers/dma/ste_dma40.c
+++ b/drivers/dma/ste_dma40.c
@@ -14,11 +14,14 @@
 #include <linux/platform_device.h>
 #include <linux/clk.h>
 #include <linux/delay.h>
+#include <linux/pm.h>
+#include <linux/pm_runtime.h>
 #include <linux/err.h>
 #include <linux/amba/bus.h>
 
 #include <plat/ste_dma40.h>
 
+#include "dmaengine.h"
 #include "ste_dma40_ll.h"
 
 #define D40_NAME "dma40"
@@ -32,6 +35,9 @@
 /* Maximum iterations taken before giving up suspending a channel */
 #define D40_SUSPEND_MAX_IT 500
 
+/* Milliseconds */
+#define DMA40_AUTOSUSPEND_DELAY	100
+
 /* Hardware requirement on LCLA alignment */
 #define LCLA_ALIGNMENT 0x40000
 
@@ -62,6 +68,55 @@ enum d40_command {
 	D40_DMA_SUSPENDED	= 3
 };
 
+/*
+ * These are the registers that has to be saved and later restored
+ * when the DMA hw is powered off.
+ * TODO: Add save/restore of D40_DREG_GCC on dma40 v3 or later, if that works.
+ */
+static u32 d40_backup_regs[] = {
+	D40_DREG_LCPA,
+	D40_DREG_LCLA,
+	D40_DREG_PRMSE,
+	D40_DREG_PRMSO,
+	D40_DREG_PRMOE,
+	D40_DREG_PRMOO,
+};
+
+#define BACKUP_REGS_SZ ARRAY_SIZE(d40_backup_regs)
+
+/* TODO: Check if all these registers have to be saved/restored on dma40 v3 */
+static u32 d40_backup_regs_v3[] = {
+	D40_DREG_PSEG1,
+	D40_DREG_PSEG2,
+	D40_DREG_PSEG3,
+	D40_DREG_PSEG4,
+	D40_DREG_PCEG1,
+	D40_DREG_PCEG2,
+	D40_DREG_PCEG3,
+	D40_DREG_PCEG4,
+	D40_DREG_RSEG1,
+	D40_DREG_RSEG2,
+	D40_DREG_RSEG3,
+	D40_DREG_RSEG4,
+	D40_DREG_RCEG1,
+	D40_DREG_RCEG2,
+	D40_DREG_RCEG3,
+	D40_DREG_RCEG4,
+};
+
+#define BACKUP_REGS_SZ_V3 ARRAY_SIZE(d40_backup_regs_v3)
+
+static u32 d40_backup_regs_chan[] = {
+	D40_CHAN_REG_SSCFG,
+	D40_CHAN_REG_SSELT,
+	D40_CHAN_REG_SSPTR,
+	D40_CHAN_REG_SSLNK,
+	D40_CHAN_REG_SDCFG,
+	D40_CHAN_REG_SDELT,
+	D40_CHAN_REG_SDPTR,
+	D40_CHAN_REG_SDLNK,
+};
+
 /**
  * struct d40_lli_pool - Structure for keeping LLIs in memory
  *
@@ -96,7 +151,7 @@ struct d40_lli_pool {
  * during a transfer.
  * @node: List entry.
  * @is_in_client_list: true if the client owns this descriptor.
- * the previous one.
+ * @cyclic: true if this is a cyclic job
  *
  * This descriptor is used for both logical and physical transfers.
  */
@@ -143,6 +198,7 @@ struct d40_lcla_pool {
  * channels.
  *
  * @lock: A lock protection this entity.
+ * @reserved: True if used by secure world or otherwise.
  * @num: The physical channel number of this entity.
  * @allocated_src: Bit mapped to show which src event line's are mapped to
  * this physical channel. Can also be free or physically allocated.
@@ -152,6 +208,7 @@ struct d40_lcla_pool {
  */
 struct d40_phy_res {
 	spinlock_t lock;
+	bool	   reserved;
 	int	   num;
 	u32	   allocated_src;
 	u32	   allocated_dst;
@@ -164,8 +221,6 @@ struct d40_base;
  *
  * @lock: A spinlock to protect this struct.
  * @log_num: The logical number, if any of this channel.
- * @completed: Starts with 1, after first interrupt it is set to dma engine's
- * current cookie.
  * @pending_tx: The number of pending transfers. Used between interrupt handler
  * and tasklet.
  * @busy: Set to true when transfer is ongoing on this channel.
@@ -185,7 +240,6 @@ struct d40_base;
  * @src_def_cfg: Default cfg register setting for src.
  * @dst_def_cfg: Default cfg register setting for dst.
  * @log_def: Default logical channel settings.
- * @lcla: Space for one dst src pair for logical channel transfers.
  * @lcpa: Pointer to dst and src lcpa settings.
  * @runtime_addr: runtime configured address.
  * @runtime_direction: runtime configured direction.
@@ -195,8 +249,6 @@ struct d40_base;
 struct d40_chan {
 	spinlock_t			 lock;
 	int				 log_num;
-	/* ID of the most recent completed transfer */
-	int				 completed;
 	int				 pending_tx;
 	bool				 busy;
 	struct d40_phy_res		*phy_chan;
@@ -217,7 +269,7 @@ struct d40_chan {
 	struct d40_log_lli_full		*lcpa;
 	/* Runtime reconfiguration */
 	dma_addr_t			runtime_addr;
-	enum dma_data_direction		runtime_direction;
+	enum dma_transfer_direction	runtime_direction;
 };
 
 /**
@@ -241,6 +293,7 @@ struct d40_chan {
  * @dma_both: dma_device channels that can do both memcpy and slave transfers.
  * @dma_slave: dma_device channels that can do only do slave transfers.
  * @dma_memcpy: dma_device channels that can do only do memcpy transfers.
+ * @phy_chans: Room for all possible physical channels in system.
  * @log_chans: Room for all possible logical channels in system.
  * @lookup_log_chans: Used to map interrupt number to logical channel. Points
  * to log_chans entries.
@@ -248,12 +301,20 @@ struct d40_chan {
  * to phy_chans entries.
  * @plat_data: Pointer to provided platform_data which is the driver
  * configuration.
+ * @lcpa_regulator: Pointer to hold the regulator for the esram bank for lcla.
  * @phy_res: Vector containing all physical channels.
  * @lcla_pool: lcla pool settings and data.
  * @lcpa_base: The virtual mapped address of LCPA.
  * @phy_lcpa: The physical address of the LCPA.
  * @lcpa_size: The size of the LCPA area.
  * @desc_slab: cache for descriptors.
+ * @reg_val_backup: Here the values of some hardware registers are stored
+ * before the DMA is powered off. They are restored when the power is back on.
+ * @reg_val_backup_v3: Backup of registers that only exits on dma40 v3 and
+ * later.
+ * @reg_val_backup_chan: Backup data for standard channel parameter registers.
+ * @gcc_pwr_off_mask: Mask to maintain the channels that can be turned off.
+ * @initialized: true if the dma has been initialized
  */
 struct d40_base {
 	spinlock_t			 interrupt_lock;
@@ -275,6 +336,7 @@ struct d40_base {
 	struct d40_chan			**lookup_log_chans;
 	struct d40_chan			**lookup_phy_chans;
 	struct stedma40_platform_data	 *plat_data;
+	struct regulator		 *lcpa_regulator;
 	/* Physical half channels */
 	struct d40_phy_res		 *phy_res;
 	struct d40_lcla_pool		  lcla_pool;
@@ -282,6 +344,11 @@ struct d40_base {
 	dma_addr_t			  phy_lcpa;
 	resource_size_t			  lcpa_size;
 	struct kmem_cache		 *desc_slab;
+	u32				  reg_val_backup[BACKUP_REGS_SZ];
+	u32				  reg_val_backup_v3[BACKUP_REGS_SZ_V3];
+	u32				 *reg_val_backup_chan;
+	u16				  gcc_pwr_off_mask;
+	bool				  initialized;
 };
 
 /**
@@ -479,13 +546,14 @@ static struct d40_desc *d40_desc_get(struct d40_chan *d40c)
 		struct d40_desc *d;
 		struct d40_desc *_d;
 
-		list_for_each_entry_safe(d, _d, &d40c->client, node)
+		list_for_each_entry_safe(d, _d, &d40c->client, node) {
 			if (async_tx_test_ack(&d->txd)) {
 				d40_desc_remove(d);
 				desc = d;
 				memset(desc, 0, sizeof(*desc));
 				break;
 			}
+		}
 	}
 
 	if (!desc)
@@ -536,6 +604,7 @@ static void d40_log_lli_to_lcxa(struct d40_chan *chan, struct d40_desc *desc)
 	bool cyclic = desc->cyclic;
 	int curr_lcla = -EINVAL;
 	int first_lcla = 0;
+	bool use_esram_lcla = chan->base->plat_data->use_esram_lcla;
 	bool linkback;
 
 	/*
@@ -608,11 +677,16 @@ static void d40_log_lli_to_lcxa(struct d40_chan *chan, struct d40_desc *desc)
 				       &lli->src[lli_current],
 				       next_lcla, flags);
 
-		dma_sync_single_range_for_device(chan->base->dev,
-					pool->dma_addr, lcla_offset,
-					2 * sizeof(struct d40_log_lli),
-					DMA_TO_DEVICE);
-
+		/*
+		 * Cache maintenance is not needed if lcla is
+		 * mapped in esram
+		 */
+		if (!use_esram_lcla) {
+			dma_sync_single_range_for_device(chan->base->dev,
+						pool->dma_addr, lcla_offset,
+						2 * sizeof(struct d40_log_lli),
+						DMA_TO_DEVICE);
+		}
 		curr_lcla = next_lcla;
 
 		if (curr_lcla == -EINVAL || curr_lcla == first_lcla) {
@@ -740,7 +814,61 @@ static int d40_sg_2_dmalen(struct scatterlist *sgl, int sg_len,
 	return len;
 }
 
-/* Support functions for logical channels */
+
+#ifdef CONFIG_PM
+static void dma40_backup(void __iomem *baseaddr, u32 *backup,
+			 u32 *regaddr, int num, bool save)
+{
+	int i;
+
+	for (i = 0; i < num; i++) {
+		void __iomem *addr = baseaddr + regaddr[i];
+
+		if (save)
+			backup[i] = readl_relaxed(addr);
+		else
+			writel_relaxed(backup[i], addr);
+	}
+}
+
+static void d40_save_restore_registers(struct d40_base *base, bool save)
+{
+	int i;
+
+	/* Save/Restore channel specific registers */
+	for (i = 0; i < base->num_phy_chans; i++) {
+		void __iomem *addr;
+		int idx;
+
+		if (base->phy_res[i].reserved)
+			continue;
+
+		addr = base->virtbase + D40_DREG_PCBASE + i * D40_DREG_PCDELTA;
+		idx = i * ARRAY_SIZE(d40_backup_regs_chan);
+
+		dma40_backup(addr, &base->reg_val_backup_chan[idx],
+			     d40_backup_regs_chan,
+			     ARRAY_SIZE(d40_backup_regs_chan),
+			     save);
+	}
+
+	/* Save/Restore global registers */
+	dma40_backup(base->virtbase, base->reg_val_backup,
+		     d40_backup_regs, ARRAY_SIZE(d40_backup_regs),
+		     save);
+
+	/* Save/Restore registers only existing on dma40 v3 and later */
+	if (base->rev >= 3)
+		dma40_backup(base->virtbase, base->reg_val_backup_v3,
+			     d40_backup_regs_v3,
+			     ARRAY_SIZE(d40_backup_regs_v3),
+			     save);
+}
+#else
+static void d40_save_restore_registers(struct d40_base *base, bool save)
+{
+}
+#endif
 
 static int d40_channel_execute_command(struct d40_chan *d40c,
 				       enum d40_command command)
@@ -973,6 +1101,10 @@ static void d40_config_write(struct d40_chan *d40c)
 		/* Set LIDX for lcla */
 		writel(lidx, chanbase + D40_CHAN_REG_SSELT);
 		writel(lidx, chanbase + D40_CHAN_REG_SDELT);
+
+		/* Clear LNK which will be used by d40_chan_has_events() */
+		writel(0, chanbase + D40_CHAN_REG_SSLNK);
+		writel(0, chanbase + D40_CHAN_REG_SDLNK);
 	}
 }
 
@@ -1013,6 +1145,7 @@ static int d40_pause(struct d40_chan *d40c)
 	if (!d40c->busy)
 		return 0;
 
+	pm_runtime_get_sync(d40c->base->dev);
 	spin_lock_irqsave(&d40c->lock, flags);
 
 	res = d40_channel_execute_command(d40c, D40_DMA_SUSPEND_REQ);
@@ -1025,7 +1158,8 @@ static int d40_pause(struct d40_chan *d40c)
 								  D40_DMA_RUN);
 		}
 	}
-
+	pm_runtime_mark_last_busy(d40c->base->dev);
+	pm_runtime_put_autosuspend(d40c->base->dev);
 	spin_unlock_irqrestore(&d40c->lock, flags);
 	return res;
 }
@@ -1039,7 +1173,7 @@ static int d40_resume(struct d40_chan *d40c)
 		return 0;
 
 	spin_lock_irqsave(&d40c->lock, flags);
-
+	pm_runtime_get_sync(d40c->base->dev);
 	if (d40c->base->rev == 0)
 		if (chan_is_logical(d40c)) {
 			res = d40_channel_execute_command(d40c,
@@ -1057,6 +1191,8 @@ static int d40_resume(struct d40_chan *d40c)
 	}
 
 no_suspend:
+	pm_runtime_mark_last_busy(d40c->base->dev);
+	pm_runtime_put_autosuspend(d40c->base->dev);
 	spin_unlock_irqrestore(&d40c->lock, flags);
 	return res;
 }
@@ -1084,21 +1220,14 @@ static dma_cookie_t d40_tx_submit(struct dma_async_tx_descriptor *tx)
 					     chan);
 	struct d40_desc *d40d = container_of(tx, struct d40_desc, txd);
 	unsigned long flags;
+	dma_cookie_t cookie;
 
 	spin_lock_irqsave(&d40c->lock, flags);
-
-	d40c->chan.cookie++;
-
-	if (d40c->chan.cookie < 0)
-		d40c->chan.cookie = 1;
-
-	d40d->txd.cookie = d40c->chan.cookie;
-
+	cookie = dma_cookie_assign(tx);
 	d40_desc_queue(d40c, d40d);
-
 	spin_unlock_irqrestore(&d40c->lock, flags);
 
-	return tx->cookie;
+	return cookie;
 }
 
 static int d40_start(struct d40_chan *d40c)
@@ -1129,7 +1258,10 @@ static struct d40_desc *d40_queue_start(struct d40_chan *d40c)
 	d40d = d40_first_queued(d40c);
 
 	if (d40d != NULL) {
-		d40c->busy = true;
+		if (!d40c->busy)
+			d40c->busy = true;
+
+		pm_runtime_get_sync(d40c->base->dev);
 
 		/* Remove from queue */
 		d40_desc_remove(d40d);
@@ -1190,6 +1322,8 @@ static void dma_tc_handle(struct d40_chan *d40c)
 
 		if (d40_queue_start(d40c) == NULL)
 			d40c->busy = false;
+		pm_runtime_mark_last_busy(d40c->base->dev);
+		pm_runtime_put_autosuspend(d40c->base->dev);
 	}
 
 	d40c->pending_tx++;
@@ -1213,7 +1347,7 @@ static void dma_tasklet(unsigned long data)
 		goto err;
 
 	if (!d40d->cyclic)
-		d40c->completed = d40d->txd.cookie;
+		dma_cookie_complete(&d40d->txd);
 
 	/*
 	 * If terminating a channel pending_tx is set to zero.
@@ -1405,11 +1539,16 @@ static int d40_validate_conf(struct d40_chan *d40c,
 	return res;
 }
 
-static bool d40_alloc_mask_set(struct d40_phy_res *phy, bool is_src,
-			       int log_event_line, bool is_log)
+static bool d40_alloc_mask_set(struct d40_phy_res *phy,
+			       bool is_src, int log_event_line, bool is_log,
+			       bool *first_user)
 {
 	unsigned long flags;
 	spin_lock_irqsave(&phy->lock, flags);
+
+	*first_user = ((phy->allocated_src | phy->allocated_dst)
+			== D40_ALLOC_FREE);
+
 	if (!is_log) {
 		/* Physical interrupts are masked per physical full channel */
 		if (phy->allocated_src == D40_ALLOC_FREE &&
@@ -1490,7 +1629,7 @@ out:
 	return is_free;
 }
 
-static int d40_allocate_channel(struct d40_chan *d40c)
+static int d40_allocate_channel(struct d40_chan *d40c, bool *first_phy_user)
 {
 	int dev_type;
 	int event_group;
@@ -1526,7 +1665,8 @@ static int d40_allocate_channel(struct d40_chan *d40c)
 			for (i = 0; i < d40c->base->num_phy_chans; i++) {
 
 				if (d40_alloc_mask_set(&phys[i], is_src,
-						       0, is_log))
+						       0, is_log,
+						       first_phy_user))
 					goto found_phy;
 			}
 		} else
@@ -1536,7 +1676,8 @@ static int d40_allocate_channel(struct d40_chan *d40c)
 					if (d40_alloc_mask_set(&phys[i],
 							       is_src,
 							       0,
-							       is_log))
+							       is_log,
+							       first_phy_user))
 						goto found_phy;
 				}
 			}
@@ -1552,6 +1693,25 @@ found_phy:
 	/* Find logical channel */
 	for (j = 0; j < d40c->base->num_phy_chans; j += 8) {
 		int phy_num = j + event_group * 2;
+
+		if (d40c->dma_cfg.use_fixed_channel) {
+			i = d40c->dma_cfg.phy_channel;
+
+			if ((i != phy_num) && (i != phy_num + 1)) {
+				dev_err(chan2dev(d40c),
+					"invalid fixed phy channel %d\n", i);
+				return -EINVAL;
+			}
+
+			if (d40_alloc_mask_set(&phys[i], is_src, event_line,
+					       is_log, first_phy_user))
+				goto found_log;
+
+			dev_err(chan2dev(d40c),
+				"could not allocate fixed phy channel %d\n", i);
+			return -EINVAL;
+		}
+
 		/*
 		 * Spread logical channels across all available physical rather
 		 * than pack every logical channel at the first available phy
@@ -1560,13 +1720,15 @@ found_phy:
 		if (is_src) {
 			for (i = phy_num; i < phy_num + 2; i++) {
 				if (d40_alloc_mask_set(&phys[i], is_src,
-						       event_line, is_log))
+						       event_line, is_log,
+						       first_phy_user))
 					goto found_log;
 			}
 		} else {
 			for (i = phy_num + 1; i >= phy_num; i--) {
 				if (d40_alloc_mask_set(&phys[i], is_src,
-						       event_line, is_log))
+						       event_line, is_log,
+						       first_phy_user))
 					goto found_log;
 			}
 		}
@@ -1643,10 +1805,11 @@ static int d40_free_dma(struct d40_chan *d40c)
 		return -EINVAL;
 	}
 
+	pm_runtime_get_sync(d40c->base->dev);
 	res = d40_channel_execute_command(d40c, D40_DMA_SUSPEND_REQ);
 	if (res) {
 		chan_err(d40c, "suspend failed\n");
-		return res;
+		goto out;
 	}
 
 	if (chan_is_logical(d40c)) {
@@ -1664,13 +1827,11 @@ static int d40_free_dma(struct d40_chan *d40c)
 			if (d40_chan_has_events(d40c)) {
 				res = d40_channel_execute_command(d40c,
 								  D40_DMA_RUN);
-				if (res) {
+				if (res)
 					chan_err(d40c,
 						"Executing RUN command\n");
-					return res;
-				}
 			}
-			return 0;
+			goto out;
 		}
 	} else {
 		(void) d40_alloc_mask_free(phy, is_src, 0);
@@ -1680,13 +1841,23 @@ static int d40_free_dma(struct d40_chan *d40c)
 	res = d40_channel_execute_command(d40c, D40_DMA_STOP);
 	if (res) {
 		chan_err(d40c, "Failed to stop channel\n");
-		return res;
+		goto out;
+	}
+
+	if (d40c->busy) {
+		pm_runtime_mark_last_busy(d40c->base->dev);
+		pm_runtime_put_autosuspend(d40c->base->dev);
 	}
+
+	d40c->busy = false;
 	d40c->phy_chan = NULL;
 	d40c->configured = false;
 	d40c->base->lookup_phy_chans[phy->num] = NULL;
+out:
 
-	return 0;
+	pm_runtime_mark_last_busy(d40c->base->dev);
+	pm_runtime_put_autosuspend(d40c->base->dev);
+	return res;
 }
 
 static bool d40_is_paused(struct d40_chan *d40c)
@@ -1855,7 +2026,7 @@ err:
 }
 
 static dma_addr_t
-d40_get_dev_addr(struct d40_chan *chan, enum dma_data_direction direction)
+d40_get_dev_addr(struct d40_chan *chan, enum dma_transfer_direction direction)
 {
 	struct stedma40_platform_data *plat = chan->base->plat_data;
 	struct stedma40_chan_cfg *cfg = &chan->dma_cfg;
@@ -1864,9 +2035,9 @@ d40_get_dev_addr(struct d40_chan *chan, enum dma_data_direction direction)
 	if (chan->runtime_addr)
 		return chan->runtime_addr;
 
-	if (direction == DMA_FROM_DEVICE)
+	if (direction == DMA_DEV_TO_MEM)
 		addr = plat->dev_rx[cfg->src_dev_type];
-	else if (direction == DMA_TO_DEVICE)
+	else if (direction == DMA_MEM_TO_DEV)
 		addr = plat->dev_tx[cfg->dst_dev_type];
 
 	return addr;
@@ -1875,7 +2046,7 @@ d40_get_dev_addr(struct d40_chan *chan, enum dma_data_direction direction)
 static struct dma_async_tx_descriptor *
 d40_prep_sg(struct dma_chan *dchan, struct scatterlist *sg_src,
 	    struct scatterlist *sg_dst, unsigned int sg_len,
-	    enum dma_data_direction direction, unsigned long dma_flags)
+	    enum dma_transfer_direction direction, unsigned long dma_flags)
 {
 	struct d40_chan *chan = container_of(dchan, struct d40_chan, chan);
 	dma_addr_t src_dev_addr = 0;
@@ -1902,9 +2073,9 @@ d40_prep_sg(struct dma_chan *dchan, struct scatterlist *sg_src,
 	if (direction != DMA_NONE) {
 		dma_addr_t dev_addr = d40_get_dev_addr(chan, direction);
 
-		if (direction == DMA_FROM_DEVICE)
+		if (direction == DMA_DEV_TO_MEM)
 			src_dev_addr = dev_addr;
-		else if (direction == DMA_TO_DEVICE)
+		else if (direction == DMA_MEM_TO_DEV)
 			dst_dev_addr = dev_addr;
 	}
 
@@ -2001,7 +2172,7 @@ static int d40_alloc_chan_resources(struct dma_chan *chan)
 	bool is_free_phy;
 	spin_lock_irqsave(&d40c->lock, flags);
 
-	d40c->completed = chan->cookie = 1;
+	dma_cookie_init(chan);
 
 	/* If no dma configuration is set use default configuration (memcpy) */
 	if (!d40c->configured) {
@@ -2011,14 +2182,15 @@ static int d40_alloc_chan_resources(struct dma_chan *chan)
 			goto fail;
 		}
 	}
-	is_free_phy = (d40c->phy_chan == NULL);
 
-	err = d40_allocate_channel(d40c);
+	err = d40_allocate_channel(d40c, &is_free_phy);
 	if (err) {
 		chan_err(d40c, "Failed to allocate channel\n");
+		d40c->configured = false;
 		goto fail;
 	}
 
+	pm_runtime_get_sync(d40c->base->dev);
 	/* Fill in basic CFG register values */
 	d40_phy_cfg(&d40c->dma_cfg, &d40c->src_def_cfg,
 		    &d40c->dst_def_cfg, chan_is_logical(d40c));
@@ -2038,6 +2210,12 @@ static int d40_alloc_chan_resources(struct dma_chan *chan)
 			  D40_LCPA_CHAN_SIZE + D40_LCPA_CHAN_DST_DELTA;
 	}
 
+	dev_dbg(chan2dev(d40c), "allocated %s channel (phy %d%s)\n",
+		 chan_is_logical(d40c) ? "logical" : "physical",
+		 d40c->phy_chan->num,
+		 d40c->dma_cfg.use_fixed_channel ? ", fixed" : "");
+
+
 	/*
 	 * Only write channel configuration to the DMA if the physical
 	 * resource is free. In case of multiple logical channels
@@ -2046,6 +2224,8 @@ static int d40_alloc_chan_resources(struct dma_chan *chan)
 	if (is_free_phy)
 		d40_config_write(d40c);
 fail:
+	pm_runtime_mark_last_busy(d40c->base->dev);
+	pm_runtime_put_autosuspend(d40c->base->dev);
 	spin_unlock_irqrestore(&d40c->lock, flags);
 	return err;
 }
@@ -2108,10 +2288,11 @@ d40_prep_memcpy_sg(struct dma_chan *chan,
 static struct dma_async_tx_descriptor *d40_prep_slave_sg(struct dma_chan *chan,
 							 struct scatterlist *sgl,
 							 unsigned int sg_len,
-							 enum dma_data_direction direction,
-							 unsigned long dma_flags)
+							 enum dma_transfer_direction direction,
+							 unsigned long dma_flags,
+							 void *context)
 {
-	if (direction != DMA_FROM_DEVICE && direction != DMA_TO_DEVICE)
+	if (direction != DMA_DEV_TO_MEM && direction != DMA_MEM_TO_DEV)
 		return NULL;
 
 	return d40_prep_sg(chan, sgl, sgl, sg_len, direction, dma_flags);
@@ -2120,7 +2301,7 @@ static struct dma_async_tx_descriptor *d40_prep_slave_sg(struct dma_chan *chan,
 static struct dma_async_tx_descriptor *
 dma40_prep_dma_cyclic(struct dma_chan *chan, dma_addr_t dma_addr,
 		     size_t buf_len, size_t period_len,
-		     enum dma_data_direction direction)
+		     enum dma_transfer_direction direction, void *context)
 {
 	unsigned int periods = buf_len / period_len;
 	struct dma_async_tx_descriptor *txd;
@@ -2152,25 +2333,19 @@ static enum dma_status d40_tx_status(struct dma_chan *chan,
 				     struct dma_tx_state *txstate)
 {
 	struct d40_chan *d40c = container_of(chan, struct d40_chan, chan);
-	dma_cookie_t last_used;
-	dma_cookie_t last_complete;
-	int ret;
+	enum dma_status ret;
 
 	if (d40c->phy_chan == NULL) {
 		chan_err(d40c, "Cannot read status of unallocated channel\n");
 		return -EINVAL;
 	}
 
-	last_complete = d40c->completed;
-	last_used = chan->cookie;
+	ret = dma_cookie_status(chan, cookie, txstate);
+	if (ret != DMA_SUCCESS)
+		dma_set_residue(txstate, stedma40_residue(chan));
 
 	if (d40_is_paused(d40c))
 		ret = DMA_PAUSED;
-	else
-		ret = dma_async_is_complete(cookie, last_complete, last_used);
-
-	dma_set_tx_state(txstate, last_complete, last_used,
-			 stedma40_residue(chan));
 
 	return ret;
 }
@@ -2269,7 +2444,7 @@ static int d40_set_runtime_config(struct dma_chan *chan,
 	dst_addr_width = config->dst_addr_width;
 	dst_maxburst = config->dst_maxburst;
 
-	if (config->direction == DMA_FROM_DEVICE) {
+	if (config->direction == DMA_DEV_TO_MEM) {
 		dma_addr_t dev_addr_rx =
 			d40c->base->plat_data->dev_rx[cfg->src_dev_type];
 
@@ -2292,7 +2467,7 @@ static int d40_set_runtime_config(struct dma_chan *chan,
 		if (dst_maxburst == 0)
 			dst_maxburst = src_maxburst;
 
-	} else if (config->direction == DMA_TO_DEVICE) {
+	} else if (config->direction == DMA_MEM_TO_DEV) {
 		dma_addr_t dev_addr_tx =
 			d40c->base->plat_data->dev_tx[cfg->dst_dev_type];
 
@@ -2357,7 +2532,7 @@ static int d40_set_runtime_config(struct dma_chan *chan,
 		"configured channel %s for %s, data width %d/%d, "
 		"maxburst %d/%d elements, LE, no flow control\n",
 		dma_chan_name(chan),
-		(config->direction == DMA_FROM_DEVICE) ? "RX" : "TX",
+		(config->direction == DMA_DEV_TO_MEM) ? "RX" : "TX",
 		src_addr_width, dst_addr_width,
 		src_maxburst, dst_maxburst);
 
@@ -2519,6 +2694,72 @@ failure1:
 	return err;
 }
 
+/* Suspend resume functionality */
+#ifdef CONFIG_PM
+static int dma40_pm_suspend(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct d40_base *base = platform_get_drvdata(pdev);
+	int ret = 0;
+	if (!pm_runtime_suspended(dev))
+		return -EBUSY;
+
+	if (base->lcpa_regulator)
+		ret = regulator_disable(base->lcpa_regulator);
+	return ret;
+}
+
+static int dma40_runtime_suspend(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct d40_base *base = platform_get_drvdata(pdev);
+
+	d40_save_restore_registers(base, true);
+
+	/* Don't disable/enable clocks for v1 due to HW bugs */
+	if (base->rev != 1)
+		writel_relaxed(base->gcc_pwr_off_mask,
+			       base->virtbase + D40_DREG_GCC);
+
+	return 0;
+}
+
+static int dma40_runtime_resume(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct d40_base *base = platform_get_drvdata(pdev);
+
+	if (base->initialized)
+		d40_save_restore_registers(base, false);
+
+	writel_relaxed(D40_DREG_GCC_ENABLE_ALL,
+		       base->virtbase + D40_DREG_GCC);
+	return 0;
+}
+
+static int dma40_resume(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct d40_base *base = platform_get_drvdata(pdev);
+	int ret = 0;
+
+	if (base->lcpa_regulator)
+		ret = regulator_enable(base->lcpa_regulator);
+
+	return ret;
+}
+
+static const struct dev_pm_ops dma40_pm_ops = {
+	.suspend		= dma40_pm_suspend,
+	.runtime_suspend	= dma40_runtime_suspend,
+	.runtime_resume		= dma40_runtime_resume,
+	.resume			= dma40_resume,
+};
+#define DMA40_PM_OPS	(&dma40_pm_ops)
+#else
+#define DMA40_PM_OPS	NULL
+#endif
+
 /* Initialization functions. */
 
 static int __init d40_phy_res_init(struct d40_base *base)
@@ -2527,6 +2768,7 @@ static int __init d40_phy_res_init(struct d40_base *base)
 	int num_phy_chans_avail = 0;
 	u32 val[2];
 	int odd_even_bit = -2;
+	int gcc = D40_DREG_GCC_ENA;
 
 	val[0] = readl(base->virtbase + D40_DREG_PRSME);
 	val[1] = readl(base->virtbase + D40_DREG_PRSMO);
@@ -2538,9 +2780,17 @@ static int __init d40_phy_res_init(struct d40_base *base)
 			/* Mark security only channels as occupied */
 			base->phy_res[i].allocated_src = D40_ALLOC_PHY;
 			base->phy_res[i].allocated_dst = D40_ALLOC_PHY;
+			base->phy_res[i].reserved = true;
+			gcc |= D40_DREG_GCC_EVTGRP_ENA(D40_PHYS_TO_GROUP(i),
+						       D40_DREG_GCC_SRC);
+			gcc |= D40_DREG_GCC_EVTGRP_ENA(D40_PHYS_TO_GROUP(i),
+						       D40_DREG_GCC_DST);
+
+
 		} else {
 			base->phy_res[i].allocated_src = D40_ALLOC_FREE;
 			base->phy_res[i].allocated_dst = D40_ALLOC_FREE;
+			base->phy_res[i].reserved = false;
 			num_phy_chans_avail++;
 		}
 		spin_lock_init(&base->phy_res[i].lock);
@@ -2552,6 +2802,11 @@ static int __init d40_phy_res_init(struct d40_base *base)
 
 		base->phy_res[chan].allocated_src = D40_ALLOC_PHY;
 		base->phy_res[chan].allocated_dst = D40_ALLOC_PHY;
+		base->phy_res[chan].reserved = true;
+		gcc |= D40_DREG_GCC_EVTGRP_ENA(D40_PHYS_TO_GROUP(chan),
+					       D40_DREG_GCC_SRC);
+		gcc |= D40_DREG_GCC_EVTGRP_ENA(D40_PHYS_TO_GROUP(chan),
+					       D40_DREG_GCC_DST);
 		num_phy_chans_avail--;
 	}
 
@@ -2572,6 +2827,15 @@ static int __init d40_phy_res_init(struct d40_base *base)
 		val[0] = val[0] >> 2;
 	}
 
+	/*
+	 * To keep things simple, Enable all clocks initially.
+	 * The clocks will get managed later post channel allocation.
+	 * The clocks for the event lines on which reserved channels exists
+	 * are not managed here.
+	 */
+	writel(D40_DREG_GCC_ENABLE_ALL, base->virtbase + D40_DREG_GCC);
+	base->gcc_pwr_off_mask = gcc;
+
 	return num_phy_chans_avail;
 }
 
@@ -2699,10 +2963,15 @@ static struct d40_base * __init d40_hw_detect_init(struct platform_device *pdev)
 			goto failure;
 	}
 
-	base->lcla_pool.alloc_map = kzalloc(num_phy_chans *
-					    sizeof(struct d40_desc *) *
-					    D40_LCLA_LINK_PER_EVENT_GRP,
+	base->reg_val_backup_chan = kmalloc(base->num_phy_chans *
+					    sizeof(d40_backup_regs_chan),
 					    GFP_KERNEL);
+	if (!base->reg_val_backup_chan)
+		goto failure;
+
+	base->lcla_pool.alloc_map =
+		kzalloc(num_phy_chans * sizeof(struct d40_desc *)
+			* D40_LCLA_LINK_PER_EVENT_GRP, GFP_KERNEL);
 	if (!base->lcla_pool.alloc_map)
 		goto failure;
 
@@ -2741,9 +3010,9 @@ failure:
 static void __init d40_hw_init(struct d40_base *base)
 {
 
-	static const struct d40_reg_val dma_init_reg[] = {
+	static struct d40_reg_val dma_init_reg[] = {
 		/* Clock every part of the DMA block from start */
-		{ .reg = D40_DREG_GCC,    .val = 0x0000ff01},
+		{ .reg = D40_DREG_GCC,    .val = D40_DREG_GCC_ENABLE_ALL},
 
 		/* Interrupts on all logical channels */
 		{ .reg = D40_DREG_LCMIS0, .val = 0xFFFFFFFF},
@@ -2943,11 +3212,31 @@ static int __init d40_probe(struct platform_device *pdev)
 		d40_err(&pdev->dev, "Failed to ioremap LCPA region\n");
 		goto failure;
 	}
+	/* If lcla has to be located in ESRAM we don't need to allocate */
+	if (base->plat_data->use_esram_lcla) {
+		res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
+							"lcla_esram");
+		if (!res) {
+			ret = -ENOENT;
+			d40_err(&pdev->dev,
+				"No \"lcla_esram\" memory resource\n");
+			goto failure;
+		}
+		base->lcla_pool.base = ioremap(res->start,
+						resource_size(res));
+		if (!base->lcla_pool.base) {
+			ret = -ENOMEM;
+			d40_err(&pdev->dev, "Failed to ioremap LCLA region\n");
+			goto failure;
+		}
+		writel(res->start, base->virtbase + D40_DREG_LCLA);
 
-	ret = d40_lcla_allocate(base);
-	if (ret) {
-		d40_err(&pdev->dev, "Failed to allocate LCLA area\n");
-		goto failure;
+	} else {
+		ret = d40_lcla_allocate(base);
+		if (ret) {
+			d40_err(&pdev->dev, "Failed to allocate LCLA area\n");
+			goto failure;
+		}
 	}
 
 	spin_lock_init(&base->lcla_pool.lock);
@@ -2960,6 +3249,32 @@ static int __init d40_probe(struct platform_device *pdev)
 		goto failure;
 	}
 
+	pm_runtime_irq_safe(base->dev);
+	pm_runtime_set_autosuspend_delay(base->dev, DMA40_AUTOSUSPEND_DELAY);
+	pm_runtime_use_autosuspend(base->dev);
+	pm_runtime_enable(base->dev);
+	pm_runtime_resume(base->dev);
+
+	if (base->plat_data->use_esram_lcla) {
+
+		base->lcpa_regulator = regulator_get(base->dev, "lcla_esram");
+		if (IS_ERR(base->lcpa_regulator)) {
+			d40_err(&pdev->dev, "Failed to get lcpa_regulator\n");
+			base->lcpa_regulator = NULL;
+			goto failure;
+		}
+
+		ret = regulator_enable(base->lcpa_regulator);
+		if (ret) {
+			d40_err(&pdev->dev,
+				"Failed to enable lcpa_regulator\n");
+			regulator_put(base->lcpa_regulator);
+			base->lcpa_regulator = NULL;
+			goto failure;
+		}
+	}
+
+	base->initialized = true;
 	err = d40_dmaengine_init(base, num_reserved_chans);
 	if (err)
 		goto failure;
@@ -2976,6 +3291,11 @@ failure:
 		if (base->virtbase)
 			iounmap(base->virtbase);
 
+		if (base->lcla_pool.base && base->plat_data->use_esram_lcla) {
+			iounmap(base->lcla_pool.base);
+			base->lcla_pool.base = NULL;
+		}
+
 		if (base->lcla_pool.dma_addr)
 			dma_unmap_single(base->dev, base->lcla_pool.dma_addr,
 					 SZ_1K * base->num_phy_chans,
@@ -2998,6 +3318,11 @@ failure:
 			clk_put(base->clk);
 		}
 
+		if (base->lcpa_regulator) {
+			regulator_disable(base->lcpa_regulator);
+			regulator_put(base->lcpa_regulator);
+		}
+
 		kfree(base->lcla_pool.alloc_map);
 		kfree(base->lookup_log_chans);
 		kfree(base->lookup_phy_chans);
@@ -3013,6 +3338,7 @@ static struct platform_driver d40_driver = {
 	.driver = {
 		.owner = THIS_MODULE,
 		.name  = D40_NAME,
+		.pm = DMA40_PM_OPS,
 	},
 };
 
diff --git a/drivers/dma/ste_dma40_ll.h b/drivers/dma/ste_dma40_ll.h
index b44c455158de..8d3d490968a3 100644
--- a/drivers/dma/ste_dma40_ll.h
+++ b/drivers/dma/ste_dma40_ll.h
@@ -16,6 +16,8 @@
 
 #define D40_TYPE_TO_GROUP(type) (type / 16)
 #define D40_TYPE_TO_EVENT(type) (type % 16)
+#define D40_GROUP_SIZE 8
+#define D40_PHYS_TO_GROUP(phys) ((phys & (D40_GROUP_SIZE - 1)) / 2)
 
 /* Most bits of the CFG register are the same in log as in phy mode */
 #define D40_SREG_CFG_MST_POS		15
@@ -123,6 +125,15 @@
 
 /* DMA Register Offsets */
 #define D40_DREG_GCC		0x000
+#define D40_DREG_GCC_ENA	0x1
+/* This assumes that there are only 4 event groups */
+#define D40_DREG_GCC_ENABLE_ALL	0xff01
+#define D40_DREG_GCC_EVTGRP_POS 8
+#define D40_DREG_GCC_SRC 0
+#define D40_DREG_GCC_DST 1
+#define D40_DREG_GCC_EVTGRP_ENA(x, y) \
+	(1 << (D40_DREG_GCC_EVTGRP_POS + 2 * x + y))
+
 #define D40_DREG_PRTYP		0x004
 #define D40_DREG_PRSME		0x008
 #define D40_DREG_PRSMO		0x00C
diff --git a/drivers/dma/timb_dma.c b/drivers/dma/timb_dma.c
index a4a398f2ef61..4e0dff59901d 100644
--- a/drivers/dma/timb_dma.c
+++ b/drivers/dma/timb_dma.c
@@ -31,6 +31,8 @@
 
 #include <linux/timb_dma.h>
 
+#include "dmaengine.h"
+
 #define DRIVER_NAME "timb-dma"
 
 /* Global DMA registers */
@@ -84,13 +86,12 @@ struct timb_dma_chan {
 					especially the lists and descriptors,
 					from races between the tasklet and calls
 					from above */
-	dma_cookie_t		last_completed_cookie;
 	bool			ongoing;
 	struct list_head	active_list;
 	struct list_head	queue;
 	struct list_head	free_list;
 	unsigned int		bytes_per_line;
-	enum dma_data_direction	direction;
+	enum dma_transfer_direction	direction;
 	unsigned int		descs; /* Descriptors to allocate */
 	unsigned int		desc_elems; /* number of elems per descriptor */
 };
@@ -166,10 +167,10 @@ static void __td_unmap_desc(struct timb_dma_chan *td_chan, const u8 *dma_desc,
 
 	if (single)
 		dma_unmap_single(chan2dev(&td_chan->chan), addr, len,
-			td_chan->direction);
+			DMA_TO_DEVICE);
 	else
 		dma_unmap_page(chan2dev(&td_chan->chan), addr, len,
-			td_chan->direction);
+			DMA_TO_DEVICE);
 }
 
 static void __td_unmap_descs(struct timb_dma_desc *td_desc, bool single)
@@ -235,7 +236,7 @@ static void __td_start_dma(struct timb_dma_chan *td_chan)
 		"td_chan: %p, chan: %d, membase: %p\n",
 		td_chan, td_chan->chan.chan_id, td_chan->membase);
 
-	if (td_chan->direction == DMA_FROM_DEVICE) {
+	if (td_chan->direction == DMA_DEV_TO_MEM) {
 
 		/* descriptor address */
 		iowrite32(0, td_chan->membase + TIMBDMA_OFFS_RX_DHAR);
@@ -278,13 +279,13 @@ static void __td_finish(struct timb_dma_chan *td_chan)
 		txd->cookie);
 
 	/* make sure to stop the transfer */
-	if (td_chan->direction == DMA_FROM_DEVICE)
+	if (td_chan->direction == DMA_DEV_TO_MEM)
 		iowrite32(0, td_chan->membase + TIMBDMA_OFFS_RX_ER);
 /* Currently no support for stopping DMA transfers
 	else
 		iowrite32(0, td_chan->membase + TIMBDMA_OFFS_TX_DLAR);
 */
-	td_chan->last_completed_cookie = txd->cookie;
+	dma_cookie_complete(txd);
 	td_chan->ongoing = false;
 
 	callback = txd->callback;
@@ -349,12 +350,7 @@ static dma_cookie_t td_tx_submit(struct dma_async_tx_descriptor *txd)
 	dma_cookie_t cookie;
 
 	spin_lock_bh(&td_chan->lock);
-
-	cookie = txd->chan->cookie;
-	if (++cookie < 0)
-		cookie = 1;
-	txd->chan->cookie = cookie;
-	txd->cookie = cookie;
+	cookie = dma_cookie_assign(txd);
 
 	if (list_empty(&td_chan->active_list)) {
 		dev_dbg(chan2dev(txd->chan), "%s: started %u\n", __func__,
@@ -481,8 +477,7 @@ static int td_alloc_chan_resources(struct dma_chan *chan)
 	}
 
 	spin_lock_bh(&td_chan->lock);
-	td_chan->last_completed_cookie = 1;
-	chan->cookie = 1;
+	dma_cookie_init(chan);
 	spin_unlock_bh(&td_chan->lock);
 
 	return 0;
@@ -515,24 +510,13 @@ static void td_free_chan_resources(struct dma_chan *chan)
 static enum dma_status td_tx_status(struct dma_chan *chan, dma_cookie_t cookie,
 				    struct dma_tx_state *txstate)
 {
-	struct timb_dma_chan *td_chan =
-		container_of(chan, struct timb_dma_chan, chan);
-	dma_cookie_t		last_used;
-	dma_cookie_t		last_complete;
-	int			ret;
+	enum dma_status ret;
 
 	dev_dbg(chan2dev(chan), "%s: Entry\n", __func__);
 
-	last_complete = td_chan->last_completed_cookie;
-	last_used = chan->cookie;
-
-	ret = dma_async_is_complete(cookie, last_complete, last_used);
-
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
+	ret = dma_cookie_status(chan, cookie, txstate);
 
-	dev_dbg(chan2dev(chan),
-		"%s: exit, ret: %d, last_complete: %d, last_used: %d\n",
-		__func__, ret, last_complete, last_used);
+	dev_dbg(chan2dev(chan), "%s: exit, ret: %d\n", 	__func__, ret);
 
 	return ret;
 }
@@ -558,7 +542,8 @@ static void td_issue_pending(struct dma_chan *chan)
 
 static struct dma_async_tx_descriptor *td_prep_slave_sg(struct dma_chan *chan,
 	struct scatterlist *sgl, unsigned int sg_len,
-	enum dma_data_direction direction, unsigned long flags)
+	enum dma_transfer_direction direction, unsigned long flags,
+	void *context)
 {
 	struct timb_dma_chan *td_chan =
 		container_of(chan, struct timb_dma_chan, chan);
@@ -606,7 +591,7 @@ static struct dma_async_tx_descriptor *td_prep_slave_sg(struct dma_chan *chan,
 	}
 
 	dma_sync_single_for_device(chan2dmadev(chan), td_desc->txd.phys,
-		td_desc->desc_list_len, DMA_TO_DEVICE);
+		td_desc->desc_list_len, DMA_MEM_TO_DEV);
 
 	return &td_desc->txd;
 }
@@ -766,7 +751,7 @@ static int __devinit td_probe(struct platform_device *pdev)
 		}
 
 		td_chan->chan.device = &td->dma;
-		td_chan->chan.cookie = 1;
+		dma_cookie_init(&td_chan->chan);
 		spin_lock_init(&td_chan->lock);
 		INIT_LIST_HEAD(&td_chan->active_list);
 		INIT_LIST_HEAD(&td_chan->queue);
@@ -775,8 +760,8 @@ static int __devinit td_probe(struct platform_device *pdev)
 		td_chan->descs = pchan->descriptors;
 		td_chan->desc_elems = pchan->descriptor_elements;
 		td_chan->bytes_per_line = pchan->bytes_per_line;
-		td_chan->direction = pchan->rx ? DMA_FROM_DEVICE :
-			DMA_TO_DEVICE;
+		td_chan->direction = pchan->rx ? DMA_DEV_TO_MEM :
+			DMA_MEM_TO_DEV;
 
 		td_chan->membase = td->membase +
 			(i / 2) * TIMBDMA_INSTANCE_OFFSET +
@@ -841,17 +826,7 @@ static struct platform_driver td_driver = {
 	.remove	= __exit_p(td_remove),
 };
 
-static int __init td_init(void)
-{
-	return platform_driver_register(&td_driver);
-}
-module_init(td_init);
-
-static void __exit td_exit(void)
-{
-	platform_driver_unregister(&td_driver);
-}
-module_exit(td_exit);
+module_platform_driver(td_driver);
 
 MODULE_LICENSE("GPL v2");
 MODULE_DESCRIPTION("Timberdale DMA controller driver");
diff --git a/drivers/dma/txx9dmac.c b/drivers/dma/txx9dmac.c
index cbd83e362b5e..913f55c76c99 100644
--- a/drivers/dma/txx9dmac.c
+++ b/drivers/dma/txx9dmac.c
@@ -15,6 +15,8 @@
 #include <linux/platform_device.h>
 #include <linux/slab.h>
 #include <linux/scatterlist.h>
+
+#include "dmaengine.h"
 #include "txx9dmac.h"
 
 static struct txx9dmac_chan *to_txx9dmac_chan(struct dma_chan *chan)
@@ -279,21 +281,6 @@ static void txx9dmac_desc_put(struct txx9dmac_chan *dc,
 	}
 }
 
-/* Called with dc->lock held and bh disabled */
-static dma_cookie_t
-txx9dmac_assign_cookie(struct txx9dmac_chan *dc, struct txx9dmac_desc *desc)
-{
-	dma_cookie_t cookie = dc->chan.cookie;
-
-	if (++cookie < 0)
-		cookie = 1;
-
-	dc->chan.cookie = cookie;
-	desc->txd.cookie = cookie;
-
-	return cookie;
-}
-
 /*----------------------------------------------------------------------*/
 
 static void txx9dmac_dump_regs(struct txx9dmac_chan *dc)
@@ -424,7 +411,7 @@ txx9dmac_descriptor_complete(struct txx9dmac_chan *dc,
 	dev_vdbg(chan2dev(&dc->chan), "descriptor %u %p complete\n",
 		 txd->cookie, desc);
 
-	dc->completed = txd->cookie;
+	dma_cookie_complete(txd);
 	callback = txd->callback;
 	param = txd->callback_param;
 
@@ -738,7 +725,7 @@ static dma_cookie_t txx9dmac_tx_submit(struct dma_async_tx_descriptor *tx)
 	dma_cookie_t cookie;
 
 	spin_lock_bh(&dc->lock);
-	cookie = txx9dmac_assign_cookie(dc, desc);
+	cookie = dma_cookie_assign(tx);
 
 	dev_vdbg(chan2dev(tx->chan), "tx_submit: queued %u %p\n",
 		 desc->txd.cookie, desc);
@@ -845,8 +832,8 @@ txx9dmac_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
 
 static struct dma_async_tx_descriptor *
 txx9dmac_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
-		unsigned int sg_len, enum dma_data_direction direction,
-		unsigned long flags)
+		unsigned int sg_len, enum dma_transfer_direction direction,
+		unsigned long flags, void *context)
 {
 	struct txx9dmac_chan *dc = to_txx9dmac_chan(chan);
 	struct txx9dmac_dev *ddev = dc->ddev;
@@ -860,9 +847,9 @@ txx9dmac_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 
 	BUG_ON(!ds || !ds->reg_width);
 	if (ds->tx_reg)
-		BUG_ON(direction != DMA_TO_DEVICE);
+		BUG_ON(direction != DMA_MEM_TO_DEV);
 	else
-		BUG_ON(direction != DMA_FROM_DEVICE);
+		BUG_ON(direction != DMA_DEV_TO_MEM);
 	if (unlikely(!sg_len))
 		return NULL;
 
@@ -882,7 +869,7 @@ txx9dmac_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 		mem = sg_dma_address(sg);
 
 		if (__is_dmac64(ddev)) {
-			if (direction == DMA_TO_DEVICE) {
+			if (direction == DMA_MEM_TO_DEV) {
 				desc->hwdesc.SAR = mem;
 				desc->hwdesc.DAR = ds->tx_reg;
 			} else {
@@ -891,7 +878,7 @@ txx9dmac_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 			}
 			desc->hwdesc.CNTR = sg_dma_len(sg);
 		} else {
-			if (direction == DMA_TO_DEVICE) {
+			if (direction == DMA_MEM_TO_DEV) {
 				desc->hwdesc32.SAR = mem;
 				desc->hwdesc32.DAR = ds->tx_reg;
 			} else {
@@ -900,7 +887,7 @@ txx9dmac_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 			}
 			desc->hwdesc32.CNTR = sg_dma_len(sg);
 		}
-		if (direction == DMA_TO_DEVICE) {
+		if (direction == DMA_MEM_TO_DEV) {
 			sai = ds->reg_width;
 			dai = 0;
 		} else {
@@ -972,27 +959,17 @@ txx9dmac_tx_status(struct dma_chan *chan, dma_cookie_t cookie,
 		   struct dma_tx_state *txstate)
 {
 	struct txx9dmac_chan *dc = to_txx9dmac_chan(chan);
-	dma_cookie_t last_used;
-	dma_cookie_t last_complete;
-	int ret;
+	enum dma_status ret;
 
-	last_complete = dc->completed;
-	last_used = chan->cookie;
-
-	ret = dma_async_is_complete(cookie, last_complete, last_used);
+	ret = dma_cookie_status(chan, cookie, txstate);
 	if (ret != DMA_SUCCESS) {
 		spin_lock_bh(&dc->lock);
 		txx9dmac_scan_descriptors(dc);
 		spin_unlock_bh(&dc->lock);
 
-		last_complete = dc->completed;
-		last_used = chan->cookie;
-
-		ret = dma_async_is_complete(cookie, last_complete, last_used);
+		ret = dma_cookie_status(chan, cookie, txstate);
 	}
 
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
-
 	return ret;
 }
 
@@ -1057,7 +1034,7 @@ static int txx9dmac_alloc_chan_resources(struct dma_chan *chan)
 		return -EIO;
 	}
 
-	dc->completed = chan->cookie = 1;
+	dma_cookie_init(chan);
 
 	dc->ccr = TXX9_DMA_CCR_IMMCHN | TXX9_DMA_CCR_INTENE | CCR_LE;
 	txx9dmac_chan_set_SMPCHN(dc);
@@ -1186,7 +1163,7 @@ static int __init txx9dmac_chan_probe(struct platform_device *pdev)
 	dc->ddev->chan[ch] = dc;
 	dc->chan.device = &dc->dma;
 	list_add_tail(&dc->chan.device_node, &dc->chan.device->channels);
-	dc->chan.cookie = dc->completed = 1;
+	dma_cookie_init(&dc->chan);
 
 	if (is_dmac64(dc))
 		dc->ch_regs = &__txx9dmac_regs(dc->ddev)->CHAN[ch];
diff --git a/drivers/dma/txx9dmac.h b/drivers/dma/txx9dmac.h
index 365d42366b9f..f5a760598882 100644
--- a/drivers/dma/txx9dmac.h
+++ b/drivers/dma/txx9dmac.h
@@ -172,7 +172,6 @@ struct txx9dmac_chan {
 	spinlock_t		lock;
 
 	/* these other elements are all protected by lock */
-	dma_cookie_t		completed;
 	struct list_head	active_list;
 	struct list_head	queue;
 	struct list_head	free_list;