Merge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx: (33 commits)
  x86: poll waiting for I/OAT DMA channel status
  maintainers: add dma engine tree details
  dmaengine: add TODO items for future work on dma drivers
  dmaengine: Add API documentation for slave dma usage
  dmaengine/dw_dmac: Update maintainer-ship
  dmaengine: move link order
  dmaengine/dw_dmac: implement pause and resume in dwc_control
  dmaengine/dw_dmac: Replace spin_lock* with irqsave variants and enable submission from callback
  dmaengine/dw_dmac: Divide one sg to many desc, if sg len is greater than DWC_MAX_COUNT
  dmaengine/dw_dmac: set residue as total len in dwc_tx_status if status is !DMA_SUCCESS
  dmaengine/dw_dmac: don't call callback routine in case dmaengine_terminate_all() is called
  dmaengine: at_hdmac: pause: no need to wait for FIFO empty
  pch_dma: modify pci device table definition
  pch_dma: Support new device ML7223 IOH
  pch_dma: Support I2S for ML7213 IOH
  pch_dma: Fix DMA setting issue
  pch_dma: modify for checkpatch
  pch_dma: fix dma direction issue for ML7213 IOH video-in
  dmaengine: at_hdmac: use descriptor chaining help function
  dmaengine: at_hdmac: implement pause and resume in atc_control
  ...

Fix up trivial conflict in drivers/dma/dw_dmac.c

18 files changed, 727 insertions, 237 deletions

diff --git a/Documentation/dmaengine.txt b/Documentation/dmaengine.txt
index 0c1c2f63c0a9..5a0cb1ef6164 100644
--- a/Documentation/dmaengine.txt
+++ b/Documentation/dmaengine.txt
@@ -1 +1,96 @@
-See Documentation/crypto/async-tx-api.txt
+			DMA Engine API Guide
+			====================
+
+		 Vinod Koul <vinod dot koul at intel.com>
+
+NOTE: For DMA Engine usage in async_tx please see:
+	Documentation/crypto/async-tx-api.txt
+
+
+Below is a guide to device driver writers on how to use the Slave-DMA API of the
+DMA Engine. This is applicable only for slave DMA usage only.
+
+The slave DMA usage consists of following steps
+1. Allocate a DMA slave channel
+2. Set slave and controller specific parameters
+3. Get a descriptor for transaction
+4. Submit the transaction and wait for callback notification
+
+1. Allocate a DMA slave channel
+Channel allocation is slightly different in the slave DMA context, client
+drivers typically need a channel from a particular DMA controller only and even
+in some cases a specific channel is desired. To request a channel
+dma_request_channel() API is used.
+
+Interface:
+struct dma_chan *dma_request_channel(dma_cap_mask_t mask,
+		dma_filter_fn filter_fn,
+		void *filter_param);
+where dma_filter_fn is defined as:
+typedef bool (*dma_filter_fn)(struct dma_chan *chan, void *filter_param);
+
+When the optional 'filter_fn' parameter is set to NULL dma_request_channel
+simply returns the first channel that satisfies the capability mask.  Otherwise,
+when the mask parameter is insufficient for specifying the necessary channel,
+the filter_fn routine can be used to disposition the available channels in the
+system. The filter_fn routine is called once for each free channel in the
+system.  Upon seeing a suitable channel filter_fn returns DMA_ACK which flags
+that channel to be the return value from dma_request_channel.  A channel
+allocated via this interface is exclusive to the caller, until
+dma_release_channel() is called.
+
+2. Set slave and controller specific parameters
+Next step is always to pass some specific information to the DMA driver. Most of
+the generic information which a slave DMA can use is in struct dma_slave_config.
+It allows the clients to specify DMA direction, DMA addresses, bus widths, DMA
+burst lengths etc. If some DMA controllers have more parameters to be sent then
+they should try to embed struct dma_slave_config in their controller specific
+structure. That gives flexibility to client to pass more parameters, if
+required.
+
+Interface:
+int dmaengine_slave_config(struct dma_chan *chan,
+					  struct dma_slave_config *config)
+
+3. Get a descriptor for transaction
+For slave usage the various modes of slave transfers supported by the
+DMA-engine are:
+slave_sg	- DMA a list of scatter gather buffers from/to a peripheral
+dma_cyclic	- Perform a cyclic DMA operation from/to a peripheral till the
+		  operation is explicitly stopped.
+The non NULL return of this transfer API represents a "descriptor" for the given
+transaction.
+
+Interface:
+struct dma_async_tx_descriptor *(*chan->device->device_prep_dma_sg)(
+		struct dma_chan *chan,
+		struct scatterlist *dst_sg, unsigned int dst_nents,
+		struct scatterlist *src_sg, unsigned int src_nents,
+		unsigned long flags);
+struct dma_async_tx_descriptor *(*chan->device->device_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);
+
+4. Submit the transaction and wait for callback notification
+To schedule the transaction to be scheduled by dma device, the "descriptor"
+returned in above (3) needs to be submitted.
+To tell the dma driver that a transaction is ready to be serviced, the
+descriptor->submit() callback needs to be invoked. This chains the descriptor to
+the pending queue.
+The transactions in the pending queue can be activated by calling the
+issue_pending API. If channel is idle then the first transaction in queue is
+started and subsequent ones queued up.
+On completion of the DMA operation the next in queue is submitted and a tasklet
+triggered. The tasklet would then call the client driver completion callback
+routine for notification, if set.
+Interface:
+void dma_async_issue_pending(struct dma_chan *chan);
+
+==============================================================================
+
+Additional usage notes for dma driver writers
+1/ Although DMA engine specifies that completion callback routines cannot submit
+any new operations, but typically for slave DMA subsequent transaction may not
+be available for submit prior to callback routine being called. This requirement
+is not a requirement for DMA-slave devices. But they should take care to drop
+the spin-lock they might be holding before calling the callback routine
diff --git a/MAINTAINERS b/MAINTAINERS
index b9f5aee36375..2e94220898d7 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2178,6 +2178,8 @@ M:	Dan Williams <dan.j.williams@intel.com>
 S:	Supported
 F:	drivers/dma/
 F:	include/linux/dma*
+T:	git git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx.git
+T:	git git://git.infradead.org/users/vkoul/slave-dma.git (slave-dma)
 
 DME1737 HARDWARE MONITOR DRIVER
 M:	Juerg Haefliger <juergh@gmail.com>
@@ -5451,6 +5453,13 @@ L:	linux-serial@vger.kernel.org
 S:	Maintained
 F:	drivers/tty/serial
 
+SYNOPSYS DESIGNWARE DMAC DRIVER
+M:	Viresh Kumar <viresh.kumar@st.com>
+S:	Maintained
+F:	include/linux/dw_dmac.h
+F:	drivers/dma/dw_dmac_regs.h
+F:	drivers/dma/dw_dmac.c
+
 TIMEKEEPING, NTP
 M:	John Stultz <johnstul@us.ibm.com>
 M:	Thomas Gleixner <tglx@linutronix.de>
diff --git a/drivers/Makefile b/drivers/Makefile
index 6b17f5864340..09f3232bcdcd 100644
--- a/drivers/Makefile
+++ b/drivers/Makefile
@@ -17,6 +17,9 @@ obj-$(CONFIG_SFI)		+= sfi/
 # was used and do nothing if so
 obj-$(CONFIG_PNP)		+= pnp/
 obj-$(CONFIG_ARM_AMBA)		+= amba/
+# Many drivers will want to use DMA so this has to be made available
+# really early.
+obj-$(CONFIG_DMA_ENGINE)	+= dma/
 
 obj-$(CONFIG_VIRTIO)		+= virtio/
 obj-$(CONFIG_XEN)		+= xen/
@@ -92,7 +95,6 @@ obj-$(CONFIG_EISA)		+= eisa/
 obj-y				+= lguest/
 obj-$(CONFIG_CPU_FREQ)		+= cpufreq/
 obj-$(CONFIG_CPU_IDLE)		+= cpuidle/
-obj-$(CONFIG_DMA_ENGINE)	+= dma/
 obj-$(CONFIG_MMC)		+= mmc/
 obj-$(CONFIG_MEMSTICK)		+= memstick/
 obj-y				+= leds/
diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig
index a572600e44eb..25cf327cd1cb 100644
--- a/drivers/dma/Kconfig
+++ b/drivers/dma/Kconfig
@@ -200,16 +200,18 @@ config PL330_DMA
 	  platform_data for a dma-pl330 device.
 
 config PCH_DMA
-	tristate "Intel EG20T PCH / OKI SEMICONDUCTOR ML7213 IOH DMA support"
+	tristate "Intel EG20T PCH / OKI Semi IOH(ML7213/ML7223) DMA support"
 	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 ML7213 IOH(Input/
-	  Output Hub) which is for IVI(In-Vehicle Infotainment) use.
-	  ML7213 is companion chip for Intel Atom E6xx series.
-	  ML7213 is completely compatible for Intel EG20T PCH.
+	  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.
 
 config IMX_SDMA
 	tristate "i.MX SDMA support"
diff --git a/drivers/dma/TODO b/drivers/dma/TODO
new file mode 100644
index 000000000000..a4af8589330c
--- /dev/null
+++ b/drivers/dma/TODO
@@ -0,0 +1,14 @@
+TODO for slave dma
+
+1. Move remaining drivers to use new slave interface
+2. Remove old slave pointer machansim
+3. Make issue_pending to start the transaction in below drivers
+	- mpc512x_dma
+	- imx-dma
+	- imx-sdma
+	- mxs-dma.c
+	- dw_dmac
+	- intel_mid_dma
+	- ste_dma40
+4. Check other subsystems for dma drivers and merge/move to dmaengine
+5. Remove dma_slave_config's dma direction.
diff --git a/drivers/dma/at_hdmac.c b/drivers/dma/at_hdmac.c
index 235f53bf494e..36144f88d718 100644
--- a/drivers/dma/at_hdmac.c
+++ b/drivers/dma/at_hdmac.c
@@ -37,8 +37,8 @@
 
 #define	ATC_DEFAULT_CFG		(ATC_FIFOCFG_HALFFIFO)
 #define	ATC_DEFAULT_CTRLA	(0)
-#define	ATC_DEFAULT_CTRLB	(ATC_SIF(0)	\
-				|ATC_DIF(1))
+#define	ATC_DEFAULT_CTRLB	(ATC_SIF(AT_DMA_MEM_IF) \
+				|ATC_DIF(AT_DMA_MEM_IF))
 
 /*
  * Initial number of descriptors to allocate for each channel. This could
@@ -165,6 +165,29 @@ static void atc_desc_put(struct at_dma_chan *atchan, struct at_desc *desc)
 }
 
 /**
+ * atc_desc_chain - build chain adding a descripor
+ * @first: address of first descripor of the chain
+ * @prev: address of previous descripor of the chain
+ * @desc: descriptor to queue
+ *
+ * Called from prep_* functions
+ */
+static void atc_desc_chain(struct at_desc **first, struct at_desc **prev,
+			   struct at_desc *desc)
+{
+	if (!(*first)) {
+		*first = desc;
+	} else {
+		/* inform the HW lli about chaining */
+		(*prev)->lli.dscr = desc->txd.phys;
+		/* insert the link descriptor to the LD ring */
+		list_add_tail(&desc->desc_node,
+				&(*first)->tx_list);
+	}
+	*prev = desc;
+}
+
+/**
  * atc_assign_cookie - compute and assign new cookie
  * @atchan: channel we work on
  * @desc: descriptor to assign cookie for
@@ -237,16 +260,12 @@ static void atc_dostart(struct at_dma_chan *atchan, struct at_desc *first)
 static void
 atc_chain_complete(struct at_dma_chan *atchan, struct at_desc *desc)
 {
-	dma_async_tx_callback		callback;
-	void				*param;
 	struct dma_async_tx_descriptor	*txd = &desc->txd;
 
 	dev_vdbg(chan2dev(&atchan->chan_common),
 		"descriptor %u complete\n", txd->cookie);
 
 	atchan->completed_cookie = txd->cookie;
-	callback = txd->callback;
-	param = txd->callback_param;
 
 	/* move children to free_list */
 	list_splice_init(&desc->tx_list, &atchan->free_list);
@@ -278,12 +297,19 @@ atc_chain_complete(struct at_dma_chan *atchan, struct at_desc *desc)
 		}
 	}
 
-	/*
-	 * The API requires that no submissions are done from a
-	 * callback, so we don't need to drop the lock here
-	 */
-	if (callback)
-		callback(param);
+	/* for cyclic transfers,
+	 * no need to replay callback function while stopping */
+	if (!test_bit(ATC_IS_CYCLIC, &atchan->status)) {
+		dma_async_tx_callback	callback = txd->callback;
+		void			*param = txd->callback_param;
+
+		/*
+		 * The API requires that no submissions are done from a
+		 * callback, so we don't need to drop the lock here
+		 */
+		if (callback)
+			callback(param);
+	}
 
 	dma_run_dependencies(txd);
 }
@@ -419,6 +445,26 @@ static void atc_handle_error(struct at_dma_chan *atchan)
 	atc_chain_complete(atchan, bad_desc);
 }
 
+/**
+ * atc_handle_cyclic - at the end of a period, run callback function
+ * @atchan: channel used for cyclic operations
+ *
+ * Called with atchan->lock held and bh disabled
+ */
+static void atc_handle_cyclic(struct at_dma_chan *atchan)
+{
+	struct at_desc			*first = atc_first_active(atchan);
+	struct dma_async_tx_descriptor	*txd = &first->txd;
+	dma_async_tx_callback		callback = txd->callback;
+	void				*param = txd->callback_param;
+
+	dev_vdbg(chan2dev(&atchan->chan_common),
+			"new cyclic period llp 0x%08x\n",
+			channel_readl(atchan, DSCR));
+
+	if (callback)
+		callback(param);
+}
 
 /*--  IRQ & Tasklet  ---------------------------------------------------*/
 
@@ -426,16 +472,11 @@ static void atc_tasklet(unsigned long data)
 {
 	struct at_dma_chan *atchan = (struct at_dma_chan *)data;
 
-	/* Channel cannot be enabled here */
-	if (atc_chan_is_enabled(atchan)) {
-		dev_err(chan2dev(&atchan->chan_common),
-			"BUG: channel enabled in tasklet\n");
-		return;
-	}
-
 	spin_lock(&atchan->lock);
-	if (test_and_clear_bit(0, &atchan->error_status))
+	if (test_and_clear_bit(ATC_IS_ERROR, &atchan->status))
 		atc_handle_error(atchan);
+	else if (test_bit(ATC_IS_CYCLIC, &atchan->status))
+		atc_handle_cyclic(atchan);
 	else
 		atc_advance_work(atchan);
 
@@ -464,12 +505,13 @@ static irqreturn_t at_dma_interrupt(int irq, void *dev_id)
 
 		for (i = 0; i < atdma->dma_common.chancnt; i++) {
 			atchan = &atdma->chan[i];
-			if (pending & (AT_DMA_CBTC(i) | AT_DMA_ERR(i))) {
+			if (pending & (AT_DMA_BTC(i) | AT_DMA_ERR(i))) {
 				if (pending & AT_DMA_ERR(i)) {
 					/* Disable channel on AHB error */
-					dma_writel(atdma, CHDR, atchan->mask);
+					dma_writel(atdma, CHDR,
+						AT_DMA_RES(i) | atchan->mask);
 					/* Give information to tasklet */
-					set_bit(0, &atchan->error_status);
+					set_bit(ATC_IS_ERROR, &atchan->status);
 				}
 				tasklet_schedule(&atchan->tasklet);
 				ret = IRQ_HANDLED;
@@ -549,7 +591,7 @@ atc_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
 	}
 
 	ctrla =   ATC_DEFAULT_CTRLA;
-	ctrlb =   ATC_DEFAULT_CTRLB
+	ctrlb =   ATC_DEFAULT_CTRLB | ATC_IEN
 		| ATC_SRC_ADDR_MODE_INCR
 		| ATC_DST_ADDR_MODE_INCR
 		| ATC_FC_MEM2MEM;
@@ -584,16 +626,7 @@ atc_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
 
 		desc->txd.cookie = 0;
 
-		if (!first) {
-			first = desc;
-		} else {
-			/* inform the HW lli about chaining */
-			prev->lli.dscr = desc->txd.phys;
-			/* insert the link descriptor to the LD ring */
-			list_add_tail(&desc->desc_node,
-					&first->tx_list);
-		}
-		prev = desc;
+		atc_desc_chain(&first, &prev, desc);
 	}
 
 	/* First descriptor of the chain embedds additional information */
@@ -639,7 +672,8 @@ atc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 	struct scatterlist	*sg;
 	size_t			total_len = 0;
 
-	dev_vdbg(chan2dev(chan), "prep_slave_sg: %s f0x%lx\n",
+	dev_vdbg(chan2dev(chan), "prep_slave_sg (%d): %s f0x%lx\n",
+			sg_len,
 			direction == DMA_TO_DEVICE ? "TO DEVICE" : "FROM DEVICE",
 			flags);
 
@@ -651,14 +685,15 @@ atc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 	reg_width = atslave->reg_width;
 
 	ctrla = ATC_DEFAULT_CTRLA | atslave->ctrla;
-	ctrlb = ATC_DEFAULT_CTRLB | ATC_IEN;
+	ctrlb = ATC_IEN;
 
 	switch (direction) {
 	case DMA_TO_DEVICE:
 		ctrla |=  ATC_DST_WIDTH(reg_width);
 		ctrlb |=  ATC_DST_ADDR_MODE_FIXED
 			| ATC_SRC_ADDR_MODE_INCR
-			| ATC_FC_MEM2PER;
+			| ATC_FC_MEM2PER
+			| ATC_SIF(AT_DMA_MEM_IF) | ATC_DIF(AT_DMA_PER_IF);
 		reg = atslave->tx_reg;
 		for_each_sg(sgl, sg, sg_len, i) {
 			struct at_desc	*desc;
@@ -682,16 +717,7 @@ atc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 					| len >> mem_width;
 			desc->lli.ctrlb = ctrlb;
 
-			if (!first) {
-				first = desc;
-			} else {
-				/* inform the HW lli about chaining */
-				prev->lli.dscr = desc->txd.phys;
-				/* insert the link descriptor to the LD ring */
-				list_add_tail(&desc->desc_node,
-						&first->tx_list);
-			}
-			prev = desc;
+			atc_desc_chain(&first, &prev, desc);
 			total_len += len;
 		}
 		break;
@@ -699,7 +725,8 @@ atc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 		ctrla |=  ATC_SRC_WIDTH(reg_width);
 		ctrlb |=  ATC_DST_ADDR_MODE_INCR
 			| ATC_SRC_ADDR_MODE_FIXED
-			| ATC_FC_PER2MEM;
+			| ATC_FC_PER2MEM
+			| ATC_SIF(AT_DMA_PER_IF) | ATC_DIF(AT_DMA_MEM_IF);
 
 		reg = atslave->rx_reg;
 		for_each_sg(sgl, sg, sg_len, i) {
@@ -724,16 +751,7 @@ atc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 					| len >> reg_width;
 			desc->lli.ctrlb = ctrlb;
 
-			if (!first) {
-				first = desc;
-			} else {
-				/* inform the HW lli about chaining */
-				prev->lli.dscr = desc->txd.phys;
-				/* insert the link descriptor to the LD ring */
-				list_add_tail(&desc->desc_node,
-						&first->tx_list);
-			}
-			prev = desc;
+			atc_desc_chain(&first, &prev, desc);
 			total_len += len;
 		}
 		break;
@@ -759,41 +777,211 @@ err_desc_get:
 	return NULL;
 }
 
+/**
+ * atc_dma_cyclic_check_values
+ * Check for too big/unaligned periods and unaligned DMA buffer
+ */
+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)
+{
+	if (period_len > (ATC_BTSIZE_MAX << reg_width))
+		goto err_out;
+	if (unlikely(period_len & ((1 << reg_width) - 1)))
+		goto err_out;
+	if (unlikely(buf_addr & ((1 << reg_width) - 1)))
+		goto err_out;
+	if (unlikely(!(direction & (DMA_TO_DEVICE | DMA_FROM_DEVICE))))
+		goto err_out;
+
+	return 0;
+
+err_out:
+	return -EINVAL;
+}
+
+/**
+ * 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,
+		unsigned int period_index, dma_addr_t buf_addr,
+		size_t period_len, enum dma_data_direction direction)
+{
+	u32		ctrla;
+	unsigned int	reg_width = atslave->reg_width;
+
+	/* prepare common CRTLA value */
+	ctrla =   ATC_DEFAULT_CTRLA | atslave->ctrla
+		| ATC_DST_WIDTH(reg_width)
+		| ATC_SRC_WIDTH(reg_width)
+		| period_len >> reg_width;
+
+	switch (direction) {
+	case DMA_TO_DEVICE:
+		desc->lli.saddr = buf_addr + (period_len * period_index);
+		desc->lli.daddr = atslave->tx_reg;
+		desc->lli.ctrla = ctrla;
+		desc->lli.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);
+		break;
+
+	case DMA_FROM_DEVICE:
+		desc->lli.saddr = atslave->rx_reg;
+		desc->lli.daddr = buf_addr + (period_len * period_index);
+		desc->lli.ctrla = ctrla;
+		desc->lli.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);
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+/**
+ * atc_prep_dma_cyclic - prepare the cyclic DMA transfer
+ * @chan: the DMA channel to prepare
+ * @buf_addr: physical DMA address where the buffer starts
+ * @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
+ */
+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)
+{
+	struct at_dma_chan	*atchan = to_at_dma_chan(chan);
+	struct at_dma_slave	*atslave = chan->private;
+	struct at_desc		*first = NULL;
+	struct at_desc		*prev = NULL;
+	unsigned long		was_cyclic;
+	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",
+			buf_addr,
+			periods, buf_len, period_len);
+
+	if (unlikely(!atslave || !buf_len || !period_len)) {
+		dev_dbg(chan2dev(chan), "prep_dma_cyclic: length is zero!\n");
+		return NULL;
+	}
+
+	was_cyclic = test_and_set_bit(ATC_IS_CYCLIC, &atchan->status);
+	if (was_cyclic) {
+		dev_dbg(chan2dev(chan), "prep_dma_cyclic: channel in use!\n");
+		return NULL;
+	}
+
+	/* Check for too big/unaligned periods and unaligned DMA buffer */
+	if (atc_dma_cyclic_check_values(atslave->reg_width, buf_addr,
+					period_len, direction))
+		goto err_out;
+
+	/* build cyclic linked list */
+	for (i = 0; i < periods; i++) {
+		struct at_desc	*desc;
+
+		desc = atc_desc_get(atchan);
+		if (!desc)
+			goto err_desc_get;
+
+		if (atc_dma_cyclic_fill_desc(atslave, desc, i, buf_addr,
+						period_len, direction))
+			goto err_desc_get;
+
+		atc_desc_chain(&first, &prev, desc);
+	}
+
+	/* lets make a cyclic list */
+	prev->lli.dscr = first->txd.phys;
+
+	/* First descriptor of the chain embedds additional information */
+	first->txd.cookie = -EBUSY;
+	first->len = buf_len;
+
+	return &first->txd;
+
+err_desc_get:
+	dev_err(chan2dev(chan), "not enough descriptors available\n");
+	atc_desc_put(atchan, first);
+err_out:
+	clear_bit(ATC_IS_CYCLIC, &atchan->status);
+	return NULL;
+}
+
+
 static int atc_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
 		       unsigned long arg)
 {
 	struct at_dma_chan	*atchan = to_at_dma_chan(chan);
 	struct at_dma		*atdma = to_at_dma(chan->device);
-	struct at_desc		*desc, *_desc;
+	int			chan_id = atchan->chan_common.chan_id;
+
 	LIST_HEAD(list);
 
-	/* Only supports DMA_TERMINATE_ALL */
-	if (cmd != DMA_TERMINATE_ALL)
-		return -ENXIO;
+	dev_vdbg(chan2dev(chan), "atc_control (%d)\n", cmd);
 
-	/*
-	 * This is only called when something went wrong elsewhere, so
-	 * we don't really care about the data. Just disable the
-	 * channel. We still have to poll the channel enable bit due
-	 * to AHB/HSB limitations.
-	 */
-	spin_lock_bh(&atchan->lock);
+	if (cmd == DMA_PAUSE) {
+		spin_lock_bh(&atchan->lock);
 
-	dma_writel(atdma, CHDR, atchan->mask);
+		dma_writel(atdma, CHER, AT_DMA_SUSP(chan_id));
+		set_bit(ATC_IS_PAUSED, &atchan->status);
 
-	/* confirm that this channel is disabled */
-	while (dma_readl(atdma, CHSR) & atchan->mask)
-		cpu_relax();
+		spin_unlock_bh(&atchan->lock);
+	} else if (cmd == DMA_RESUME) {
+		if (!test_bit(ATC_IS_PAUSED, &atchan->status))
+			return 0;
 
-	/* active_list entries will end up before queued entries */
-	list_splice_init(&atchan->queue, &list);
-	list_splice_init(&atchan->active_list, &list);
+		spin_lock_bh(&atchan->lock);
 
-	/* Flush all pending and queued descriptors */
-	list_for_each_entry_safe(desc, _desc, &list, desc_node)
-		atc_chain_complete(atchan, desc);
+		dma_writel(atdma, CHDR, AT_DMA_RES(chan_id));
+		clear_bit(ATC_IS_PAUSED, &atchan->status);
 
-	spin_unlock_bh(&atchan->lock);
+		spin_unlock_bh(&atchan->lock);
+	} else if (cmd == DMA_TERMINATE_ALL) {
+		struct at_desc	*desc, *_desc;
+		/*
+		 * This is only called when something went wrong elsewhere, so
+		 * we don't really care about the data. Just disable the
+		 * channel. We still have to poll the channel enable bit due
+		 * to AHB/HSB limitations.
+		 */
+		spin_lock_bh(&atchan->lock);
+
+		/* disabling channel: must also remove suspend state */
+		dma_writel(atdma, CHDR, AT_DMA_RES(chan_id) | atchan->mask);
+
+		/* confirm that this channel is disabled */
+		while (dma_readl(atdma, CHSR) & atchan->mask)
+			cpu_relax();
+
+		/* active_list entries will end up before queued entries */
+		list_splice_init(&atchan->queue, &list);
+		list_splice_init(&atchan->active_list, &list);
+
+		/* Flush all pending and queued descriptors */
+		list_for_each_entry_safe(desc, _desc, &list, desc_node)
+			atc_chain_complete(atchan, desc);
+
+		clear_bit(ATC_IS_PAUSED, &atchan->status);
+		/* if channel dedicated to cyclic operations, free it */
+		clear_bit(ATC_IS_CYCLIC, &atchan->status);
+
+		spin_unlock_bh(&atchan->lock);
+	} else {
+		return -ENXIO;
+	}
 
 	return 0;
 }
@@ -835,9 +1023,17 @@ atc_tx_status(struct dma_chan *chan,
 
 	spin_unlock_bh(&atchan->lock);
 
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
-	dev_vdbg(chan2dev(chan), "tx_status: %d (d%d, u%d)\n",
-		 cookie, last_complete ? last_complete : 0,
+	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);
+
+	if (test_bit(ATC_IS_PAUSED, &atchan->status))
+		ret = DMA_PAUSED;
+
+	dev_vdbg(chan2dev(chan), "tx_status %d: cookie = %d (d%d, u%d)\n",
+		 ret, cookie, last_complete ? last_complete : 0,
 		 last_used ? last_used : 0);
 
 	return ret;
@@ -853,6 +1049,10 @@ static void atc_issue_pending(struct dma_chan *chan)
 
 	dev_vdbg(chan2dev(chan), "issue_pending\n");
 
+	/* Not needed for cyclic transfers */
+	if (test_bit(ATC_IS_CYCLIC, &atchan->status))
+		return;
+
 	spin_lock_bh(&atchan->lock);
 	if (!atc_chan_is_enabled(atchan)) {
 		atc_advance_work(atchan);
@@ -959,6 +1159,7 @@ static void atc_free_chan_resources(struct dma_chan *chan)
 	}
 	list_splice_init(&atchan->free_list, &list);
 	atchan->descs_allocated = 0;
+	atchan->status = 0;
 
 	dev_vdbg(chan2dev(chan), "free_chan_resources: done\n");
 }
@@ -1092,10 +1293,15 @@ static int __init at_dma_probe(struct platform_device *pdev)
 	if (dma_has_cap(DMA_MEMCPY, atdma->dma_common.cap_mask))
 		atdma->dma_common.device_prep_dma_memcpy = atc_prep_dma_memcpy;
 
-	if (dma_has_cap(DMA_SLAVE, atdma->dma_common.cap_mask)) {
+	if (dma_has_cap(DMA_SLAVE, atdma->dma_common.cap_mask))
 		atdma->dma_common.device_prep_slave_sg = atc_prep_slave_sg;
+
+	if (dma_has_cap(DMA_CYCLIC, atdma->dma_common.cap_mask))
+		atdma->dma_common.device_prep_dma_cyclic = atc_prep_dma_cyclic;
+
+	if (dma_has_cap(DMA_SLAVE, atdma->dma_common.cap_mask) ||
+	    dma_has_cap(DMA_CYCLIC, atdma->dma_common.cap_mask))
 		atdma->dma_common.device_control = atc_control;
-	}
 
 	dma_writel(atdma, EN, AT_DMA_ENABLE);
 
diff --git a/drivers/dma/at_hdmac_regs.h b/drivers/dma/at_hdmac_regs.h
index 495457e3dc4b..087dbf1dd39c 100644
--- a/drivers/dma/at_hdmac_regs.h
+++ b/drivers/dma/at_hdmac_regs.h
@@ -103,6 +103,10 @@
 /* Bitfields in CTRLB */
 #define	ATC_SIF(i)		(0x3 & (i))	/* Src tx done via AHB-Lite Interface i */
 #define	ATC_DIF(i)		((0x3 & (i)) <<  4)	/* Dst tx done via AHB-Lite Interface i */
+				  /* Specify AHB interfaces */
+#define AT_DMA_MEM_IF		0 /* interface 0 as memory interface */
+#define AT_DMA_PER_IF		1 /* interface 1 as peripheral interface */
+
 #define	ATC_SRC_PIP		(0x1 <<  8)	/* Source Picture-in-Picture enabled */
 #define	ATC_DST_PIP		(0x1 << 12)	/* Destination Picture-in-Picture enabled */
 #define	ATC_SRC_DSCR_DIS	(0x1 << 16)	/* Src Descriptor fetch disable */
@@ -181,12 +185,23 @@ txd_to_at_desc(struct dma_async_tx_descriptor *txd)
 /*--  Channels  --------------------------------------------------------*/
 
 /**
+ * atc_status - information bits stored in channel status flag
+ *
+ * Manipulated with atomic operations.
+ */
+enum atc_status {
+	ATC_IS_ERROR = 0,
+	ATC_IS_PAUSED = 1,
+	ATC_IS_CYCLIC = 24,
+};
+
+/**
  * struct at_dma_chan - internal representation of an Atmel HDMAC channel
  * @chan_common: common dmaengine channel object members
  * @device: parent device
  * @ch_regs: memory mapped register base
  * @mask: channel index in a mask
- * @error_status: transmit error status information from irq handler
+ * @status: transmit status information from irq/prep* functions
  *                to tasklet (use atomic operations)
  * @tasklet: bottom half to finish transaction work
  * @lock: serializes enqueue/dequeue operations to descriptors lists
@@ -201,7 +216,7 @@ struct at_dma_chan {
 	struct at_dma		*device;
 	void __iomem		*ch_regs;
 	u8			mask;
-	unsigned long		error_status;
+	unsigned long		status;
 	struct tasklet_struct	tasklet;
 
 	spinlock_t		lock;
@@ -309,8 +324,8 @@ static void atc_setup_irq(struct at_dma_chan *atchan, int on)
 	struct at_dma	*atdma = to_at_dma(atchan->chan_common.device);
 	u32		ebci;
 
-	/* enable interrupts on buffer chain completion & error */
-	ebci =    AT_DMA_CBTC(atchan->chan_common.chan_id)
+	/* enable interrupts on buffer transfer completion & error */
+	ebci =    AT_DMA_BTC(atchan->chan_common.chan_id)
 		| AT_DMA_ERR(atchan->chan_common.chan_id);
 	if (on)
 		dma_writel(atdma, EBCIER, ebci);
@@ -347,7 +362,12 @@ static inline int atc_chan_is_enabled(struct at_dma_chan *atchan)
  */
 static void set_desc_eol(struct at_desc *desc)
 {
-	desc->lli.ctrlb |= ATC_SRC_DSCR_DIS | ATC_DST_DSCR_DIS;
+	u32 ctrlb = desc->lli.ctrlb;
+
+	ctrlb &= ~ATC_IEN;
+	ctrlb |= ATC_SRC_DSCR_DIS | ATC_DST_DSCR_DIS;
+
+	desc->lli.ctrlb = ctrlb;
 	desc->lli.dscr = 0;
 }
 
diff --git a/drivers/dma/coh901318.c b/drivers/dma/coh901318.c
index f48e54006518..af8c0b5ed70f 100644
--- a/drivers/dma/coh901318.c
+++ b/drivers/dma/coh901318.c
@@ -1610,7 +1610,7 @@ int __init coh901318_init(void)
 {
 	return platform_driver_probe(&coh901318_driver, coh901318_probe);
 }
-arch_initcall(coh901318_init);
+subsys_initcall(coh901318_init);
 
 void __exit coh901318_exit(void)
 {
diff --git a/drivers/dma/dw_dmac.c b/drivers/dma/dw_dmac.c
index 2a2e2fa00e91..4d180ca9a1d8 100644
--- a/drivers/dma/dw_dmac.c
+++ b/drivers/dma/dw_dmac.c
@@ -3,6 +3,7 @@
  * AVR32 systems.)
  *
  * Copyright (C) 2007-2008 Atmel Corporation
+ * Copyright (C) 2010-2011 ST Microelectronics
  *
  * 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
@@ -93,8 +94,9 @@ static struct dw_desc *dwc_desc_get(struct dw_dma_chan *dwc)
 	struct dw_desc *desc, *_desc;
 	struct dw_desc *ret = NULL;
 	unsigned int i = 0;
+	unsigned long flags;
 
-	spin_lock_bh(&dwc->lock);
+	spin_lock_irqsave(&dwc->lock, flags);
 	list_for_each_entry_safe(desc, _desc, &dwc->free_list, desc_node) {
 		if (async_tx_test_ack(&desc->txd)) {
 			list_del(&desc->desc_node);
@@ -104,7 +106,7 @@ static struct dw_desc *dwc_desc_get(struct dw_dma_chan *dwc)
 		dev_dbg(chan2dev(&dwc->chan), "desc %p not ACKed\n", desc);
 		i++;
 	}
-	spin_unlock_bh(&dwc->lock);
+	spin_unlock_irqrestore(&dwc->lock, flags);
 
 	dev_vdbg(chan2dev(&dwc->chan), "scanned %u descriptors on freelist\n", i);
 
@@ -130,12 +132,14 @@ static void dwc_sync_desc_for_cpu(struct dw_dma_chan *dwc, struct dw_desc *desc)
  */
 static void dwc_desc_put(struct dw_dma_chan *dwc, struct dw_desc *desc)
 {
+	unsigned long flags;
+
 	if (desc) {
 		struct dw_desc *child;
 
 		dwc_sync_desc_for_cpu(dwc, desc);
 
-		spin_lock_bh(&dwc->lock);
+		spin_lock_irqsave(&dwc->lock, flags);
 		list_for_each_entry(child, &desc->tx_list, desc_node)
 			dev_vdbg(chan2dev(&dwc->chan),
 					"moving child desc %p to freelist\n",
@@ -143,7 +147,7 @@ static void dwc_desc_put(struct dw_dma_chan *dwc, struct dw_desc *desc)
 		list_splice_init(&desc->tx_list, &dwc->free_list);
 		dev_vdbg(chan2dev(&dwc->chan), "moving desc %p to freelist\n", desc);
 		list_add(&desc->desc_node, &dwc->free_list);
-		spin_unlock_bh(&dwc->lock);
+		spin_unlock_irqrestore(&dwc->lock, flags);
 	}
 }
 
@@ -195,18 +199,23 @@ static void dwc_dostart(struct dw_dma_chan *dwc, struct dw_desc *first)
 /*----------------------------------------------------------------------*/
 
 static void
-dwc_descriptor_complete(struct dw_dma_chan *dwc, struct dw_desc *desc)
+dwc_descriptor_complete(struct dw_dma_chan *dwc, struct dw_desc *desc,
+		bool callback_required)
 {
-	dma_async_tx_callback		callback;
-	void				*param;
+	dma_async_tx_callback		callback = NULL;
+	void				*param = NULL;
 	struct dma_async_tx_descriptor	*txd = &desc->txd;
 	struct dw_desc			*child;
+	unsigned long			flags;
 
 	dev_vdbg(chan2dev(&dwc->chan), "descriptor %u complete\n", txd->cookie);
 
+	spin_lock_irqsave(&dwc->lock, flags);
 	dwc->completed = txd->cookie;
-	callback = txd->callback;
-	param = txd->callback_param;
+	if (callback_required) {
+		callback = txd->callback;
+		param = txd->callback_param;
+	}
 
 	dwc_sync_desc_for_cpu(dwc, desc);
 
@@ -238,11 +247,9 @@ dwc_descriptor_complete(struct dw_dma_chan *dwc, struct dw_desc *desc)
 		}
 	}
 
-	/*
-	 * The API requires that no submissions are done from a
-	 * callback, so we don't need to drop the lock here
-	 */
-	if (callback)
+	spin_unlock_irqrestore(&dwc->lock, flags);
+
+	if (callback_required && callback)
 		callback(param);
 }
 
@@ -250,7 +257,9 @@ static void dwc_complete_all(struct dw_dma *dw, struct dw_dma_chan *dwc)
 {
 	struct dw_desc *desc, *_desc;
 	LIST_HEAD(list);
+	unsigned long flags;
 
+	spin_lock_irqsave(&dwc->lock, flags);
 	if (dma_readl(dw, CH_EN) & dwc->mask) {
 		dev_err(chan2dev(&dwc->chan),
 			"BUG: XFER bit set, but channel not idle!\n");
@@ -271,8 +280,10 @@ static void dwc_complete_all(struct dw_dma *dw, struct dw_dma_chan *dwc)
 		dwc_dostart(dwc, dwc_first_active(dwc));
 	}
 
+	spin_unlock_irqrestore(&dwc->lock, flags);
+
 	list_for_each_entry_safe(desc, _desc, &list, desc_node)
-		dwc_descriptor_complete(dwc, desc);
+		dwc_descriptor_complete(dwc, desc, true);
 }
 
 static void dwc_scan_descriptors(struct dw_dma *dw, struct dw_dma_chan *dwc)
@@ -281,7 +292,9 @@ static void dwc_scan_descriptors(struct dw_dma *dw, struct dw_dma_chan *dwc)
 	struct dw_desc *desc, *_desc;
 	struct dw_desc *child;
 	u32 status_xfer;
+	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
@@ -294,30 +307,47 @@ static void dwc_scan_descriptors(struct dw_dma *dw, struct dw_dma_chan *dwc)
 	if (status_xfer & dwc->mask) {
 		/* Everything we've submitted is done */
 		dma_writel(dw, CLEAR.XFER, dwc->mask);
+		spin_unlock_irqrestore(&dwc->lock, flags);
+
 		dwc_complete_all(dw, dwc);
 		return;
 	}
 
-	if (list_empty(&dwc->active_list))
+	if (list_empty(&dwc->active_list)) {
+		spin_unlock_irqrestore(&dwc->lock, flags);
 		return;
+	}
 
 	dev_vdbg(chan2dev(&dwc->chan), "scan_descriptors: llp=0x%x\n", llp);
 
 	list_for_each_entry_safe(desc, _desc, &dwc->active_list, desc_node) {
-		if (desc->lli.llp == llp)
+		/* check first descriptors addr */
+		if (desc->txd.phys == llp) {
+			spin_unlock_irqrestore(&dwc->lock, flags);
+			return;
+		}
+
+		/* check first descriptors llp */
+		if (desc->lli.llp == llp) {
 			/* This one is currently in progress */
+			spin_unlock_irqrestore(&dwc->lock, flags);
 			return;
+		}
 
 		list_for_each_entry(child, &desc->tx_list, desc_node)
-			if (child->lli.llp == llp)
+			if (child->lli.llp == llp) {
 				/* Currently in progress */
+				spin_unlock_irqrestore(&dwc->lock, flags);
 				return;
+			}
 
 		/*
 		 * No descriptors so far seem to be in progress, i.e.
 		 * this one must be done.
 		 */
-		dwc_descriptor_complete(dwc, desc);
+		spin_unlock_irqrestore(&dwc->lock, flags);
+		dwc_descriptor_complete(dwc, desc, true);
+		spin_lock_irqsave(&dwc->lock, flags);
 	}
 
 	dev_err(chan2dev(&dwc->chan),
@@ -332,6 +362,7 @@ static void dwc_scan_descriptors(struct dw_dma *dw, struct dw_dma_chan *dwc)
 		list_move(dwc->queue.next, &dwc->active_list);
 		dwc_dostart(dwc, dwc_first_active(dwc));
 	}
+	spin_unlock_irqrestore(&dwc->lock, flags);
 }
 
 static void dwc_dump_lli(struct dw_dma_chan *dwc, struct dw_lli *lli)
@@ -346,9 +377,12 @@ static void dwc_handle_error(struct dw_dma *dw, struct dw_dma_chan *dwc)
 {
 	struct dw_desc *bad_desc;
 	struct dw_desc *child;
+	unsigned long flags;
 
 	dwc_scan_descriptors(dw, dwc);
 
+	spin_lock_irqsave(&dwc->lock, flags);
+
 	/*
 	 * The descriptor currently at the head of the active list is
 	 * borked. Since we don't have any way to report errors, we'll
@@ -378,8 +412,10 @@ static void dwc_handle_error(struct dw_dma *dw, struct dw_dma_chan *dwc)
 	list_for_each_entry(child, &bad_desc->tx_list, desc_node)
 		dwc_dump_lli(dwc, &child->lli);
 
+	spin_unlock_irqrestore(&dwc->lock, flags);
+
 	/* Pretend the descriptor completed successfully */
-	dwc_descriptor_complete(dwc, bad_desc);
+	dwc_descriptor_complete(dwc, bad_desc, true);
 }
 
 /* --------------------- Cyclic DMA API extensions -------------------- */
@@ -402,6 +438,8 @@ EXPORT_SYMBOL(dw_dma_get_dst_addr);
 static void dwc_handle_cyclic(struct dw_dma *dw, struct dw_dma_chan *dwc,
 		u32 status_block, u32 status_err, u32 status_xfer)
 {
+	unsigned long flags;
+
 	if (status_block & dwc->mask) {
 		void (*callback)(void *param);
 		void *callback_param;
@@ -412,11 +450,9 @@ static void dwc_handle_cyclic(struct dw_dma *dw, struct dw_dma_chan *dwc,
 
 		callback = dwc->cdesc->period_callback;
 		callback_param = dwc->cdesc->period_callback_param;
-		if (callback) {
-			spin_unlock(&dwc->lock);
+
+		if (callback)
 			callback(callback_param);
-			spin_lock(&dwc->lock);
-		}
 	}
 
 	/*
@@ -430,6 +466,9 @@ static void dwc_handle_cyclic(struct dw_dma *dw, struct dw_dma_chan *dwc,
 		dev_err(chan2dev(&dwc->chan), "cyclic DMA unexpected %s "
 				"interrupt, stopping DMA transfer\n",
 				status_xfer ? "xfer" : "error");
+
+		spin_lock_irqsave(&dwc->lock, flags);
+
 		dev_err(chan2dev(&dwc->chan),
 			"  SAR: 0x%x DAR: 0x%x LLP: 0x%x CTL: 0x%x:%08x\n",
 			channel_readl(dwc, SAR),
@@ -453,6 +492,8 @@ static void dwc_handle_cyclic(struct dw_dma *dw, struct dw_dma_chan *dwc,
 
 		for (i = 0; i < dwc->cdesc->periods; i++)
 			dwc_dump_lli(dwc, &dwc->cdesc->desc[i]->lli);
+
+		spin_unlock_irqrestore(&dwc->lock, flags);
 	}
 }
 
@@ -476,7 +517,6 @@ static void dw_dma_tasklet(unsigned long data)
 
 	for (i = 0; i < dw->dma.chancnt; i++) {
 		dwc = &dw->chan[i];
-		spin_lock(&dwc->lock);
 		if (test_bit(DW_DMA_IS_CYCLIC, &dwc->flags))
 			dwc_handle_cyclic(dw, dwc, status_block, status_err,
 					status_xfer);
@@ -484,7 +524,6 @@ static void dw_dma_tasklet(unsigned long data)
 			dwc_handle_error(dw, dwc);
 		else if ((status_block | status_xfer) & (1 << i))
 			dwc_scan_descriptors(dw, dwc);
-		spin_unlock(&dwc->lock);
 	}
 
 	/*
@@ -539,8 +578,9 @@ static dma_cookie_t dwc_tx_submit(struct dma_async_tx_descriptor *tx)
 	struct dw_desc		*desc = txd_to_dw_desc(tx);
 	struct dw_dma_chan	*dwc = to_dw_dma_chan(tx->chan);
 	dma_cookie_t		cookie;
+	unsigned long		flags;
 
-	spin_lock_bh(&dwc->lock);
+	spin_lock_irqsave(&dwc->lock, flags);
 	cookie = dwc_assign_cookie(dwc, desc);
 
 	/*
@@ -560,7 +600,7 @@ static dma_cookie_t dwc_tx_submit(struct dma_async_tx_descriptor *tx)
 		list_add_tail(&desc->desc_node, &dwc->queue);
 	}
 
-	spin_unlock_bh(&dwc->lock);
+	spin_unlock_irqrestore(&dwc->lock, flags);
 
 	return cookie;
 }
@@ -689,9 +729,15 @@ dwc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 		reg = dws->tx_reg;
 		for_each_sg(sgl, sg, sg_len, i) {
 			struct dw_desc	*desc;
-			u32		len;
-			u32		mem;
+			u32		len, dlen, mem;
+
+			mem = sg_phys(sg);
+			len = sg_dma_len(sg);
+			mem_width = 2;
+			if (unlikely(mem & 3 || len & 3))
+				mem_width = 0;
 
+slave_sg_todev_fill_desc:
 			desc = dwc_desc_get(dwc);
 			if (!desc) {
 				dev_err(chan2dev(chan),
@@ -699,16 +745,19 @@ dwc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 				goto err_desc_get;
 			}
 
-			mem = sg_phys(sg);
-			len = sg_dma_len(sg);
-			mem_width = 2;
-			if (unlikely(mem & 3 || len & 3))
-				mem_width = 0;
-
 			desc->lli.sar = mem;
 			desc->lli.dar = reg;
 			desc->lli.ctllo = ctllo | DWC_CTLL_SRC_WIDTH(mem_width);
-			desc->lli.ctlhi = len >> mem_width;
+			if ((len >> mem_width) > DWC_MAX_COUNT) {
+				dlen = DWC_MAX_COUNT << mem_width;
+				mem += dlen;
+				len -= dlen;
+			} else {
+				dlen = len;
+				len = 0;
+			}
+
+			desc->lli.ctlhi = dlen >> mem_width;
 
 			if (!first) {
 				first = desc;
@@ -722,7 +771,10 @@ dwc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 						&first->tx_list);
 			}
 			prev = desc;
-			total_len += len;
+			total_len += dlen;
+
+			if (len)
+				goto slave_sg_todev_fill_desc;
 		}
 		break;
 	case DMA_FROM_DEVICE:
@@ -735,15 +787,7 @@ dwc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 		reg = dws->rx_reg;
 		for_each_sg(sgl, sg, sg_len, i) {
 			struct dw_desc	*desc;
-			u32		len;
-			u32		mem;
-
-			desc = dwc_desc_get(dwc);
-			if (!desc) {
-				dev_err(chan2dev(chan),
-					"not enough descriptors available\n");
-				goto err_desc_get;
-			}
+			u32		len, dlen, mem;
 
 			mem = sg_phys(sg);
 			len = sg_dma_len(sg);
@@ -751,10 +795,26 @@ dwc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 			if (unlikely(mem & 3 || len & 3))
 				mem_width = 0;
 
+slave_sg_fromdev_fill_desc:
+			desc = dwc_desc_get(dwc);
+			if (!desc) {
+				dev_err(chan2dev(chan),
+						"not enough descriptors available\n");
+				goto err_desc_get;
+			}
+
 			desc->lli.sar = reg;
 			desc->lli.dar = mem;
 			desc->lli.ctllo = ctllo | DWC_CTLL_DST_WIDTH(mem_width);
-			desc->lli.ctlhi = len >> reg_width;
+			if ((len >> reg_width) > DWC_MAX_COUNT) {
+				dlen = DWC_MAX_COUNT << reg_width;
+				mem += dlen;
+				len -= dlen;
+			} else {
+				dlen = len;
+				len = 0;
+			}
+			desc->lli.ctlhi = dlen >> reg_width;
 
 			if (!first) {
 				first = desc;
@@ -768,7 +828,10 @@ dwc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
 						&first->tx_list);
 			}
 			prev = desc;
-			total_len += len;
+			total_len += dlen;
+
+			if (len)
+				goto slave_sg_fromdev_fill_desc;
 		}
 		break;
 	default:
@@ -799,34 +862,51 @@ static int dwc_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
 	struct dw_dma_chan	*dwc = to_dw_dma_chan(chan);
 	struct dw_dma		*dw = to_dw_dma(chan->device);
 	struct dw_desc		*desc, *_desc;
+	unsigned long		flags;
+	u32			cfglo;
 	LIST_HEAD(list);
 
-	/* Only supports DMA_TERMINATE_ALL */
-	if (cmd != DMA_TERMINATE_ALL)
-		return -ENXIO;
+	if (cmd == DMA_PAUSE) {
+		spin_lock_irqsave(&dwc->lock, flags);
 
-	/*
-	 * This is only called when something went wrong elsewhere, so
-	 * we don't really care about the data. Just disable the
-	 * channel. We still have to poll the channel enable bit due
-	 * to AHB/HSB limitations.
-	 */
-	spin_lock_bh(&dwc->lock);
+		cfglo = channel_readl(dwc, CFG_LO);
+		channel_writel(dwc, CFG_LO, cfglo | DWC_CFGL_CH_SUSP);
+		while (!(channel_readl(dwc, CFG_LO) & DWC_CFGL_FIFO_EMPTY))
+			cpu_relax();
 
-	channel_clear_bit(dw, CH_EN, dwc->mask);
+		dwc->paused = true;
+		spin_unlock_irqrestore(&dwc->lock, flags);
+	} else if (cmd == DMA_RESUME) {
+		if (!dwc->paused)
+			return 0;
 
-	while (dma_readl(dw, CH_EN) & dwc->mask)
-		cpu_relax();
+		spin_lock_irqsave(&dwc->lock, flags);
 
-	/* active_list entries will end up before queued entries */
-	list_splice_init(&dwc->queue, &list);
-	list_splice_init(&dwc->active_list, &list);
+		cfglo = channel_readl(dwc, CFG_LO);
+		channel_writel(dwc, CFG_LO, cfglo & ~DWC_CFGL_CH_SUSP);
+		dwc->paused = false;
 
-	spin_unlock_bh(&dwc->lock);
+		spin_unlock_irqrestore(&dwc->lock, flags);
+	} else if (cmd == DMA_TERMINATE_ALL) {
+		spin_lock_irqsave(&dwc->lock, flags);
 
-	/* Flush all pending and queued descriptors */
-	list_for_each_entry_safe(desc, _desc, &list, desc_node)
-		dwc_descriptor_complete(dwc, desc);
+		channel_clear_bit(dw, CH_EN, dwc->mask);
+		while (dma_readl(dw, CH_EN) & dwc->mask)
+			cpu_relax();
+
+		dwc->paused = false;
+
+		/* active_list entries will end up before queued entries */
+		list_splice_init(&dwc->queue, &list);
+		list_splice_init(&dwc->active_list, &list);
+
+		spin_unlock_irqrestore(&dwc->lock, flags);
+
+		/* Flush all pending and queued descriptors */
+		list_for_each_entry_safe(desc, _desc, &list, desc_node)
+			dwc_descriptor_complete(dwc, desc, false);
+	} else
+		return -ENXIO;
 
 	return 0;
 }
@@ -846,9 +926,7 @@ dwc_tx_status(struct dma_chan *chan,
 
 	ret = dma_async_is_complete(cookie, last_complete, last_used);
 	if (ret != DMA_SUCCESS) {
-		spin_lock_bh(&dwc->lock);
 		dwc_scan_descriptors(to_dw_dma(chan->device), dwc);
-		spin_unlock_bh(&dwc->lock);
 
 		last_complete = dwc->completed;
 		last_used = chan->cookie;
@@ -856,7 +934,14 @@ dwc_tx_status(struct dma_chan *chan,
 		ret = dma_async_is_complete(cookie, last_complete, last_used);
 	}
 
-	dma_set_tx_state(txstate, last_complete, last_used, 0);
+	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);
+
+	if (dwc->paused)
+		return DMA_PAUSED;
 
 	return ret;
 }
@@ -865,10 +950,8 @@ static void dwc_issue_pending(struct dma_chan *chan)
 {
 	struct dw_dma_chan	*dwc = to_dw_dma_chan(chan);
 
-	spin_lock_bh(&dwc->lock);
 	if (!list_empty(&dwc->queue))
 		dwc_scan_descriptors(to_dw_dma(chan->device), dwc);
-	spin_unlock_bh(&dwc->lock);
 }
 
 static int dwc_alloc_chan_resources(struct dma_chan *chan)
@@ -880,6 +963,7 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
 	int			i;
 	u32			cfghi;
 	u32			cfglo;
+	unsigned long		flags;
 
 	dev_vdbg(chan2dev(chan), "alloc_chan_resources\n");
 
@@ -917,16 +1001,16 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
 	 * doesn't mean what you think it means), and status writeback.
 	 */
 
-	spin_lock_bh(&dwc->lock);
+	spin_lock_irqsave(&dwc->lock, flags);
 	i = dwc->descs_allocated;
 	while (dwc->descs_allocated < NR_DESCS_PER_CHANNEL) {
-		spin_unlock_bh(&dwc->lock);
+		spin_unlock_irqrestore(&dwc->lock, flags);
 
 		desc = kzalloc(sizeof(struct dw_desc), GFP_KERNEL);
 		if (!desc) {
 			dev_info(chan2dev(chan),
 				"only allocated %d descriptors\n", i);
-			spin_lock_bh(&dwc->lock);
+			spin_lock_irqsave(&dwc->lock, flags);
 			break;
 		}
 
@@ -938,7 +1022,7 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
 				sizeof(desc->lli), DMA_TO_DEVICE);
 		dwc_desc_put(dwc, desc);
 
-		spin_lock_bh(&dwc->lock);
+		spin_lock_irqsave(&dwc->lock, flags);
 		i = ++dwc->descs_allocated;
 	}
 
@@ -947,7 +1031,7 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
 	channel_set_bit(dw, MASK.BLOCK, dwc->mask);
 	channel_set_bit(dw, MASK.ERROR, dwc->mask);
 
-	spin_unlock_bh(&dwc->lock);
+	spin_unlock_irqrestore(&dwc->lock, flags);
 
 	dev_dbg(chan2dev(chan),
 		"alloc_chan_resources allocated %d descriptors\n", i);
@@ -960,6 +1044,7 @@ static void dwc_free_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, *_desc;
+	unsigned long		flags;
 	LIST_HEAD(list);
 
 	dev_dbg(chan2dev(chan), "free_chan_resources (descs allocated=%u)\n",
@@ -970,7 +1055,7 @@ static void dwc_free_chan_resources(struct dma_chan *chan)
 	BUG_ON(!list_empty(&dwc->queue));
 	BUG_ON(dma_readl(to_dw_dma(chan->device), CH_EN) & dwc->mask);
 
-	spin_lock_bh(&dwc->lock);
+	spin_lock_irqsave(&dwc->lock, flags);
 	list_splice_init(&dwc->free_list, &list);
 	dwc->descs_allocated = 0;
 
@@ -979,7 +1064,7 @@ static void dwc_free_chan_resources(struct dma_chan *chan)
 	channel_clear_bit(dw, MASK.BLOCK, dwc->mask);
 	channel_clear_bit(dw, MASK.ERROR, dwc->mask);
 
-	spin_unlock_bh(&dwc->lock);
+	spin_unlock_irqrestore(&dwc->lock, flags);
 
 	list_for_each_entry_safe(desc, _desc, &list, desc_node) {
 		dev_vdbg(chan2dev(chan), "  freeing descriptor %p\n", desc);
@@ -1004,13 +1089,14 @@ int dw_dma_cyclic_start(struct dma_chan *chan)
 {
 	struct dw_dma_chan	*dwc = to_dw_dma_chan(chan);
 	struct dw_dma		*dw = to_dw_dma(dwc->chan.device);
+	unsigned long		flags;
 
 	if (!test_bit(DW_DMA_IS_CYCLIC, &dwc->flags)) {
 		dev_err(chan2dev(&dwc->chan), "missing prep for cyclic DMA\n");
 		return -ENODEV;
 	}
 
-	spin_lock(&dwc->lock);
+	spin_lock_irqsave(&dwc->lock, flags);
 
 	/* assert channel is idle */
 	if (dma_readl(dw, CH_EN) & dwc->mask) {
@@ -1023,7 +1109,7 @@ int dw_dma_cyclic_start(struct dma_chan *chan)
 			channel_readl(dwc, LLP),
 			channel_readl(dwc, CTL_HI),
 			channel_readl(dwc, CTL_LO));
-		spin_unlock(&dwc->lock);
+		spin_unlock_irqrestore(&dwc->lock, flags);
 		return -EBUSY;
 	}
 
@@ -1038,7 +1124,7 @@ int dw_dma_cyclic_start(struct dma_chan *chan)
 
 	channel_set_bit(dw, CH_EN, dwc->mask);
 
-	spin_unlock(&dwc->lock);
+	spin_unlock_irqrestore(&dwc->lock, flags);
 
 	return 0;
 }
@@ -1054,14 +1140,15 @@ void dw_dma_cyclic_stop(struct dma_chan *chan)
 {
 	struct dw_dma_chan	*dwc = to_dw_dma_chan(chan);
 	struct dw_dma		*dw = to_dw_dma(dwc->chan.device);
+	unsigned long		flags;
 
-	spin_lock(&dwc->lock);
+	spin_lock_irqsave(&dwc->lock, flags);
 
 	channel_clear_bit(dw, CH_EN, dwc->mask);
 	while (dma_readl(dw, CH_EN) & dwc->mask)
 		cpu_relax();
 
-	spin_unlock(&dwc->lock);
+	spin_unlock_irqrestore(&dwc->lock, flags);
 }
 EXPORT_SYMBOL(dw_dma_cyclic_stop);
 
@@ -1090,17 +1177,18 @@ struct dw_cyclic_desc *dw_dma_cyclic_prep(struct dma_chan *chan,
 	unsigned int			reg_width;
 	unsigned int			periods;
 	unsigned int			i;
+	unsigned long			flags;
 
-	spin_lock_bh(&dwc->lock);
+	spin_lock_irqsave(&dwc->lock, flags);
 	if (!list_empty(&dwc->queue) || !list_empty(&dwc->active_list)) {
-		spin_unlock_bh(&dwc->lock);
+		spin_unlock_irqrestore(&dwc->lock, flags);
 		dev_dbg(chan2dev(&dwc->chan),
 				"queue and/or active list are not empty\n");
 		return ERR_PTR(-EBUSY);
 	}
 
 	was_cyclic = test_and_set_bit(DW_DMA_IS_CYCLIC, &dwc->flags);
-	spin_unlock_bh(&dwc->lock);
+	spin_unlock_irqrestore(&dwc->lock, flags);
 	if (was_cyclic) {
 		dev_dbg(chan2dev(&dwc->chan),
 				"channel already prepared for cyclic DMA\n");
@@ -1214,13 +1302,14 @@ void dw_dma_cyclic_free(struct dma_chan *chan)
 	struct dw_dma		*dw = to_dw_dma(dwc->chan.device);
 	struct dw_cyclic_desc	*cdesc = dwc->cdesc;
 	int			i;
+	unsigned long		flags;
 
 	dev_dbg(chan2dev(&dwc->chan), "cyclic free\n");
 
 	if (!cdesc)
 		return;
 
-	spin_lock_bh(&dwc->lock);
+	spin_lock_irqsave(&dwc->lock, flags);
 
 	channel_clear_bit(dw, CH_EN, dwc->mask);
 	while (dma_readl(dw, CH_EN) & dwc->mask)
@@ -1230,7 +1319,7 @@ void dw_dma_cyclic_free(struct dma_chan *chan)
 	dma_writel(dw, CLEAR.ERROR, dwc->mask);
 	dma_writel(dw, CLEAR.XFER, dwc->mask);
 
-	spin_unlock_bh(&dwc->lock);
+	spin_unlock_irqrestore(&dwc->lock, flags);
 
 	for (i = 0; i < cdesc->periods; i++)
 		dwc_desc_put(dwc, cdesc->desc[i]);
@@ -1487,3 +1576,4 @@ module_exit(dw_exit);
 MODULE_LICENSE("GPL v2");
 MODULE_DESCRIPTION("Synopsys DesignWare DMA Controller driver");
 MODULE_AUTHOR("Haavard Skinnemoen (Atmel)");
+MODULE_AUTHOR("Viresh Kumar <viresh.kumar@st.com>");
diff --git a/drivers/dma/dw_dmac_regs.h b/drivers/dma/dw_dmac_regs.h
index 720f821527f8..c3419518d701 100644
--- a/drivers/dma/dw_dmac_regs.h
+++ b/drivers/dma/dw_dmac_regs.h
@@ -2,6 +2,7 @@
  * Driver for the Synopsys DesignWare AHB DMA Controller
  *
  * Copyright (C) 2005-2007 Atmel Corporation
+ * Copyright (C) 2010-2011 ST Microelectronics
  *
  * 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
@@ -138,6 +139,7 @@ struct dw_dma_chan {
 	void __iomem		*ch_regs;
 	u8			mask;
 	u8			priority;
+	bool			paused;
 
 	spinlock_t		lock;
 
diff --git a/drivers/dma/intel_mid_dma.c b/drivers/dma/intel_mid_dma.c
index 3d4ec38b9b62..f653517ef744 100644
--- a/drivers/dma/intel_mid_dma.c
+++ b/drivers/dma/intel_mid_dma.c
@@ -1292,8 +1292,7 @@ static int __devinit intel_mid_dma_probe(struct pci_dev *pdev,
 	if (err)
 		goto err_dma;
 
-	pm_runtime_set_active(&pdev->dev);
-	pm_runtime_enable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
 	pm_runtime_allow(&pdev->dev);
 	return 0;
 
@@ -1322,6 +1321,9 @@ err_enable_device:
 static void __devexit intel_mid_dma_remove(struct pci_dev *pdev)
 {
 	struct middma_device *device = pci_get_drvdata(pdev);
+
+	pm_runtime_get_noresume(&pdev->dev);
+	pm_runtime_forbid(&pdev->dev);
 	middma_shutdown(pdev);
 	pci_dev_put(pdev);
 	kfree(device);
@@ -1385,13 +1387,20 @@ int dma_resume(struct pci_dev *pci)
 static int dma_runtime_suspend(struct device *dev)
 {
 	struct pci_dev *pci_dev = to_pci_dev(dev);
-	return dma_suspend(pci_dev, PMSG_SUSPEND);
+	struct middma_device *device = pci_get_drvdata(pci_dev);
+
+	device->state = SUSPENDED;
+	return 0;
 }
 
 static int dma_runtime_resume(struct device *dev)
 {
 	struct pci_dev *pci_dev = to_pci_dev(dev);
-	return dma_resume(pci_dev);
+	struct middma_device *device = pci_get_drvdata(pci_dev);
+
+	device->state = RUNNING;
+	iowrite32(REG_BIT0, device->dma_base + DMA_CFG);
+	return 0;
 }
 
 static int dma_runtime_idle(struct device *dev)
diff --git a/drivers/dma/ioat/dma_v2.c b/drivers/dma/ioat/dma_v2.c
index f4a51d4d0349..5d65f8377971 100644
--- a/drivers/dma/ioat/dma_v2.c
+++ b/drivers/dma/ioat/dma_v2.c
@@ -508,6 +508,7 @@ int ioat2_alloc_chan_resources(struct dma_chan *c)
 	struct ioat_ring_ent **ring;
 	u64 status;
 	int order;
+	int i = 0;
 
 	/* have we already been set up? */
 	if (ioat->ring)
@@ -548,8 +549,11 @@ int ioat2_alloc_chan_resources(struct dma_chan *c)
 	ioat2_start_null_desc(ioat);
 
 	/* check that we got off the ground */
-	udelay(5);
-	status = ioat_chansts(chan);
+	do {
+		udelay(1);
+		status = ioat_chansts(chan);
+	} while (i++ < 20 && !is_ioat_active(status) && !is_ioat_idle(status));
+
 	if (is_ioat_active(status) || is_ioat_idle(status)) {
 		set_bit(IOAT_RUN, &chan->state);
 		return 1 << ioat->alloc_order;
diff --git a/drivers/dma/iop-adma.c b/drivers/dma/iop-adma.c
index c6b01f535b29..e03f811a83dd 100644
--- a/drivers/dma/iop-adma.c
+++ b/drivers/dma/iop-adma.c
@@ -619,7 +619,7 @@ iop_adma_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dma_dest,
 
 	if (unlikely(!len))
 		return NULL;
-	BUG_ON(unlikely(len > IOP_ADMA_MAX_BYTE_COUNT));
+	BUG_ON(len > IOP_ADMA_MAX_BYTE_COUNT);
 
 	dev_dbg(iop_chan->device->common.dev, "%s len: %u\n",
 		__func__, len);
@@ -652,7 +652,7 @@ iop_adma_prep_dma_memset(struct dma_chan *chan, dma_addr_t dma_dest,
 
 	if (unlikely(!len))
 		return NULL;
-	BUG_ON(unlikely(len > IOP_ADMA_MAX_BYTE_COUNT));
+	BUG_ON(len > IOP_ADMA_MAX_BYTE_COUNT);
 
 	dev_dbg(iop_chan->device->common.dev, "%s len: %u\n",
 		__func__, len);
@@ -686,7 +686,7 @@ iop_adma_prep_dma_xor(struct dma_chan *chan, dma_addr_t dma_dest,
 
 	if (unlikely(!len))
 		return NULL;
-	BUG_ON(unlikely(len > IOP_ADMA_XOR_MAX_BYTE_COUNT));
+	BUG_ON(len > IOP_ADMA_XOR_MAX_BYTE_COUNT);
 
 	dev_dbg(iop_chan->device->common.dev,
 		"%s src_cnt: %d len: %u flags: %lx\n",
diff --git a/drivers/dma/mv_xor.c b/drivers/dma/mv_xor.c
index a25f5f61e0e0..954e334e01bb 100644
--- a/drivers/dma/mv_xor.c
+++ b/drivers/dma/mv_xor.c
@@ -671,7 +671,7 @@ mv_xor_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
 	if (unlikely(len < MV_XOR_MIN_BYTE_COUNT))
 		return NULL;
 
-	BUG_ON(unlikely(len > MV_XOR_MAX_BYTE_COUNT));
+	BUG_ON(len > MV_XOR_MAX_BYTE_COUNT);
 
 	spin_lock_bh(&mv_chan->lock);
 	slot_cnt = mv_chan_memcpy_slot_count(len);
@@ -710,7 +710,7 @@ mv_xor_prep_dma_memset(struct dma_chan *chan, dma_addr_t dest, int value,
 	if (unlikely(len < MV_XOR_MIN_BYTE_COUNT))
 		return NULL;
 
-	BUG_ON(unlikely(len > MV_XOR_MAX_BYTE_COUNT));
+	BUG_ON(len > MV_XOR_MAX_BYTE_COUNT);
 
 	spin_lock_bh(&mv_chan->lock);
 	slot_cnt = mv_chan_memset_slot_count(len);
@@ -744,7 +744,7 @@ mv_xor_prep_dma_xor(struct dma_chan *chan, dma_addr_t dest, dma_addr_t *src,
 	if (unlikely(len < MV_XOR_MIN_BYTE_COUNT))
 		return NULL;
 
-	BUG_ON(unlikely(len > MV_XOR_MAX_BYTE_COUNT));
+	BUG_ON(len > MV_XOR_MAX_BYTE_COUNT);
 
 	dev_dbg(mv_chan->device->common.dev,
 		"%s src_cnt: %d len: dest %x %u flags: %ld\n",
diff --git a/drivers/dma/pch_dma.c b/drivers/dma/pch_dma.c
index 8d8fef1480a9..ff5b38f9d45b 100644
--- a/drivers/dma/pch_dma.c
+++ b/drivers/dma/pch_dma.c
@@ -77,10 +77,10 @@ struct pch_dma_regs {
 	u32	dma_ctl0;
 	u32	dma_ctl1;
 	u32	dma_ctl2;
-	u32	reserved1;
+	u32	dma_ctl3;
 	u32	dma_sts0;
 	u32	dma_sts1;
-	u32	reserved2;
+	u32	dma_sts2;
 	u32	reserved3;
 	struct pch_dma_desc_regs desc[MAX_CHAN_NR];
 };
@@ -130,6 +130,7 @@ struct pch_dma {
 #define PCH_DMA_CTL0	0x00
 #define PCH_DMA_CTL1	0x04
 #define PCH_DMA_CTL2	0x08
+#define PCH_DMA_CTL3	0x0C
 #define PCH_DMA_STS0	0x10
 #define PCH_DMA_STS1	0x14
 
@@ -138,7 +139,8 @@ struct pch_dma {
 #define dma_writel(pd, name, val) \
 	writel((val), (pd)->membase + PCH_DMA_##name)
 
-static inline struct pch_dma_desc *to_pd_desc(struct dma_async_tx_descriptor *txd)
+static inline
+struct pch_dma_desc *to_pd_desc(struct dma_async_tx_descriptor *txd)
 {
 	return container_of(txd, struct pch_dma_desc, txd);
 }
@@ -163,13 +165,15 @@ static inline struct device *chan2parent(struct dma_chan *chan)
 	return chan->dev->device.parent;
 }
 
-static inline struct pch_dma_desc *pdc_first_active(struct pch_dma_chan *pd_chan)
+static inline
+struct pch_dma_desc *pdc_first_active(struct pch_dma_chan *pd_chan)
 {
 	return list_first_entry(&pd_chan->active_list,
 				struct pch_dma_desc, desc_node);
 }
 
-static inline struct pch_dma_desc *pdc_first_queued(struct pch_dma_chan *pd_chan)
+static inline
+struct pch_dma_desc *pdc_first_queued(struct pch_dma_chan *pd_chan)
 {
 	return list_first_entry(&pd_chan->queue,
 				struct pch_dma_desc, desc_node);
@@ -199,16 +203,30 @@ static void pdc_set_dir(struct dma_chan *chan)
 	struct pch_dma *pd = to_pd(chan->device);
 	u32 val;
 
-	val = dma_readl(pd, CTL0);
+	if (chan->chan_id < 8) {
+		val = dma_readl(pd, CTL0);
 
-	if (pd_chan->dir == DMA_TO_DEVICE)
-		val |= 0x1 << (DMA_CTL0_BITS_PER_CH * chan->chan_id +
-			       DMA_CTL0_DIR_SHIFT_BITS);
-	else
-		val &= ~(0x1 << (DMA_CTL0_BITS_PER_CH * chan->chan_id +
-				 DMA_CTL0_DIR_SHIFT_BITS));
+		if (pd_chan->dir == DMA_TO_DEVICE)
+			val |= 0x1 << (DMA_CTL0_BITS_PER_CH * chan->chan_id +
+				       DMA_CTL0_DIR_SHIFT_BITS);
+		else
+			val &= ~(0x1 << (DMA_CTL0_BITS_PER_CH * chan->chan_id +
+					 DMA_CTL0_DIR_SHIFT_BITS));
+
+		dma_writel(pd, CTL0, val);
+	} else {
+		int ch = chan->chan_id - 8; /* ch8-->0 ch9-->1 ... ch11->3 */
+		val = dma_readl(pd, CTL3);
 
-	dma_writel(pd, CTL0, val);
+		if (pd_chan->dir == DMA_TO_DEVICE)
+			val |= 0x1 << (DMA_CTL0_BITS_PER_CH * ch +
+				       DMA_CTL0_DIR_SHIFT_BITS);
+		else
+			val &= ~(0x1 << (DMA_CTL0_BITS_PER_CH * ch +
+					 DMA_CTL0_DIR_SHIFT_BITS));
+
+		dma_writel(pd, CTL3, val);
+	}
 
 	dev_dbg(chan2dev(chan), "pdc_set_dir: chan %d -> %x\n",
 		chan->chan_id, val);
@@ -219,13 +237,26 @@ static void pdc_set_mode(struct dma_chan *chan, u32 mode)
 	struct pch_dma *pd = to_pd(chan->device);
 	u32 val;
 
-	val = dma_readl(pd, CTL0);
+	if (chan->chan_id < 8) {
+		val = dma_readl(pd, CTL0);
+
+		val &= ~(DMA_CTL0_MODE_MASK_BITS <<
+			(DMA_CTL0_BITS_PER_CH * chan->chan_id));
+		val |= mode << (DMA_CTL0_BITS_PER_CH * chan->chan_id);
 
-	val &= ~(DMA_CTL0_MODE_MASK_BITS <<
-		(DMA_CTL0_BITS_PER_CH * chan->chan_id));
-	val |= mode << (DMA_CTL0_BITS_PER_CH * chan->chan_id);
+		dma_writel(pd, CTL0, val);
+	} else {
+		int ch = chan->chan_id - 8; /* ch8-->0 ch9-->1 ... ch11->3 */
+
+		val = dma_readl(pd, CTL3);
+
+		val &= ~(DMA_CTL0_MODE_MASK_BITS <<
+			(DMA_CTL0_BITS_PER_CH * ch));
+		val |= mode << (DMA_CTL0_BITS_PER_CH * ch);
 
-	dma_writel(pd, CTL0, val);
+		dma_writel(pd, CTL3, val);
+
+	}
 
 	dev_dbg(chan2dev(chan), "pdc_set_mode: chan %d -> %x\n",
 		chan->chan_id, val);
@@ -251,9 +282,6 @@ static bool pdc_is_idle(struct pch_dma_chan *pd_chan)
 
 static void pdc_dostart(struct pch_dma_chan *pd_chan, struct pch_dma_desc* desc)
 {
-	struct pch_dma *pd = to_pd(pd_chan->chan.device);
-	u32 val;
-
 	if (!pdc_is_idle(pd_chan)) {
 		dev_err(chan2dev(&pd_chan->chan),
 			"BUG: Attempt to start non-idle channel\n");
@@ -279,10 +307,6 @@ static void pdc_dostart(struct pch_dma_chan *pd_chan, struct pch_dma_desc* desc)
 		channel_writel(pd_chan, NEXT, desc->txd.phys);
 		pdc_set_mode(&pd_chan->chan, DMA_CTL0_SG);
 	}
-
-	val = dma_readl(pd, CTL2);
-	val |= 1 << (DMA_CTL2_START_SHIFT_BITS + pd_chan->chan.chan_id);
-	dma_writel(pd, CTL2, val);
 }
 
 static void pdc_chain_complete(struct pch_dma_chan *pd_chan,
@@ -403,7 +427,7 @@ static struct pch_dma_desc *pdc_desc_get(struct pch_dma_chan *pd_chan)
 {
 	struct pch_dma_desc *desc, *_d;
 	struct pch_dma_desc *ret = NULL;
-	int i;
+	int i = 0;
 
 	spin_lock(&pd_chan->lock);
 	list_for_each_entry_safe(desc, _d, &pd_chan->free_list, desc_node) {
@@ -478,7 +502,6 @@ static int pd_alloc_chan_resources(struct dma_chan *chan)
 	spin_unlock_bh(&pd_chan->lock);
 
 	pdc_enable_irq(chan, 1);
-	pdc_set_dir(chan);
 
 	return pd_chan->descs_allocated;
 }
@@ -561,6 +584,9 @@ static struct dma_async_tx_descriptor *pd_prep_slave_sg(struct dma_chan *chan,
 	else
 		return NULL;
 
+	pd_chan->dir = direction;
+	pdc_set_dir(chan);
+
 	for_each_sg(sgl, sg, sg_len, i) {
 		desc = pdc_desc_get(pd_chan);
 
@@ -703,6 +729,7 @@ static void pch_dma_save_regs(struct pch_dma *pd)
 	pd->regs.dma_ctl0 = dma_readl(pd, CTL0);
 	pd->regs.dma_ctl1 = dma_readl(pd, CTL1);
 	pd->regs.dma_ctl2 = dma_readl(pd, CTL2);
+	pd->regs.dma_ctl3 = dma_readl(pd, CTL3);
 
 	list_for_each_entry_safe(chan, _c, &pd->dma.channels, device_node) {
 		pd_chan = to_pd_chan(chan);
@@ -725,6 +752,7 @@ static void pch_dma_restore_regs(struct pch_dma *pd)
 	dma_writel(pd, CTL0, pd->regs.dma_ctl0);
 	dma_writel(pd, CTL1, pd->regs.dma_ctl1);
 	dma_writel(pd, CTL2, pd->regs.dma_ctl2);
+	dma_writel(pd, CTL3, pd->regs.dma_ctl3);
 
 	list_for_each_entry_safe(chan, _c, &pd->dma.channels, device_node) {
 		pd_chan = to_pd_chan(chan);
@@ -850,8 +878,6 @@ static int __devinit pch_dma_probe(struct pci_dev *pdev,
 
 		pd_chan->membase = &regs->desc[i];
 
-		pd_chan->dir = (i % 2) ? DMA_FROM_DEVICE : DMA_TO_DEVICE;
-
 		spin_lock_init(&pd_chan->lock);
 
 		INIT_LIST_HEAD(&pd_chan->active_list);
@@ -929,13 +955,23 @@ static void __devexit pch_dma_remove(struct pci_dev *pdev)
 #define PCI_DEVICE_ID_ML7213_DMA1_8CH	0x8026
 #define PCI_DEVICE_ID_ML7213_DMA2_8CH	0x802B
 #define PCI_DEVICE_ID_ML7213_DMA3_4CH	0x8034
+#define PCI_DEVICE_ID_ML7213_DMA4_12CH	0x8032
+#define PCI_DEVICE_ID_ML7223_DMA1_4CH	0x800B
+#define PCI_DEVICE_ID_ML7223_DMA2_4CH	0x800E
+#define PCI_DEVICE_ID_ML7223_DMA3_4CH	0x8017
+#define PCI_DEVICE_ID_ML7223_DMA4_4CH	0x803B
 
-static const struct pci_device_id pch_dma_id_table[] = {
+DEFINE_PCI_DEVICE_TABLE(pch_dma_id_table) = {
 	{ PCI_VDEVICE(INTEL, PCI_DEVICE_ID_EG20T_PCH_DMA_8CH), 8 },
 	{ PCI_VDEVICE(INTEL, PCI_DEVICE_ID_EG20T_PCH_DMA_4CH), 4 },
 	{ PCI_VDEVICE(ROHM, PCI_DEVICE_ID_ML7213_DMA1_8CH), 8}, /* UART Video */
 	{ PCI_VDEVICE(ROHM, PCI_DEVICE_ID_ML7213_DMA2_8CH), 8}, /* PCMIF SPI */
 	{ PCI_VDEVICE(ROHM, PCI_DEVICE_ID_ML7213_DMA3_4CH), 4}, /* FPGA */
+	{ PCI_VDEVICE(ROHM, PCI_DEVICE_ID_ML7213_DMA4_12CH), 12}, /* I2S */
+	{ PCI_VDEVICE(ROHM, PCI_DEVICE_ID_ML7223_DMA1_4CH), 4}, /* UART */
+	{ 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 */
 	{ 0, },
 };
 
diff --git a/drivers/dma/ppc4xx/adma.c b/drivers/dma/ppc4xx/adma.c
index 3b0247e74cc4..fc457a7e8832 100644
--- a/drivers/dma/ppc4xx/adma.c
+++ b/drivers/dma/ppc4xx/adma.c
@@ -2313,7 +2313,7 @@ static struct dma_async_tx_descriptor *ppc440spe_adma_prep_dma_memcpy(
 	if (unlikely(!len))
 		return NULL;
 
-	BUG_ON(unlikely(len > PPC440SPE_ADMA_DMA_MAX_BYTE_COUNT));
+	BUG_ON(len > PPC440SPE_ADMA_DMA_MAX_BYTE_COUNT);
 
 	spin_lock_bh(&ppc440spe_chan->lock);
 
@@ -2354,7 +2354,7 @@ static struct dma_async_tx_descriptor *ppc440spe_adma_prep_dma_memset(
 	if (unlikely(!len))
 		return NULL;
 
-	BUG_ON(unlikely(len > PPC440SPE_ADMA_DMA_MAX_BYTE_COUNT));
+	BUG_ON(len > PPC440SPE_ADMA_DMA_MAX_BYTE_COUNT);
 
 	spin_lock_bh(&ppc440spe_chan->lock);
 
@@ -2397,7 +2397,7 @@ static struct dma_async_tx_descriptor *ppc440spe_adma_prep_dma_xor(
 				     dma_dest, dma_src, src_cnt));
 	if (unlikely(!len))
 		return NULL;
-	BUG_ON(unlikely(len > PPC440SPE_ADMA_XOR_MAX_BYTE_COUNT));
+	BUG_ON(len > PPC440SPE_ADMA_XOR_MAX_BYTE_COUNT);
 
 	dev_dbg(ppc440spe_chan->device->common.dev,
 		"ppc440spe adma%d: %s src_cnt: %d len: %u int_en: %d\n",
@@ -2887,7 +2887,7 @@ static struct dma_async_tx_descriptor *ppc440spe_adma_prep_dma_pq(
 	ADMA_LL_DBG(prep_dma_pq_dbg(ppc440spe_chan->device->id,
 				    dst, src, src_cnt));
 	BUG_ON(!len);
-	BUG_ON(unlikely(len > PPC440SPE_ADMA_XOR_MAX_BYTE_COUNT));
+	BUG_ON(len > PPC440SPE_ADMA_XOR_MAX_BYTE_COUNT);
 	BUG_ON(!src_cnt);
 
 	if (src_cnt == 1 && dst[1] == src[0]) {
diff --git a/drivers/dma/ste_dma40.c b/drivers/dma/ste_dma40.c
index 94ee15dd3aed..8f222d4db7de 100644
--- a/drivers/dma/ste_dma40.c
+++ b/drivers/dma/ste_dma40.c
@@ -1829,7 +1829,7 @@ d40_get_dev_addr(struct d40_chan *chan, enum dma_data_direction direction)
 {
 	struct stedma40_platform_data *plat = chan->base->plat_data;
 	struct stedma40_chan_cfg *cfg = &chan->dma_cfg;
-	dma_addr_t addr;
+	dma_addr_t addr = 0;
 
 	if (chan->runtime_addr)
 		return chan->runtime_addr;
@@ -2962,4 +2962,4 @@ static int __init stedma40_init(void)
 {
 	return platform_driver_probe(&d40_driver, d40_probe);
 }
-arch_initcall(stedma40_init);
+subsys_initcall(stedma40_init);
diff --git a/include/linux/dw_dmac.h b/include/linux/dw_dmac.h
index 6998d9376ef9..4bfe0a2f7d50 100644
--- a/include/linux/dw_dmac.h
+++ b/include/linux/dw_dmac.h
@@ -3,6 +3,7 @@
  * AVR32 systems.)
  *
  * Copyright (C) 2007 Atmel Corporation
+ * Copyright (C) 2010-2011 ST Microelectronics
  *
  * 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