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:
  dma: Add SoF and EoF debugging to ipu_idmac.c, minor cleanup
  dw_dmac: add cyclic API to DW DMA driver
  dmaengine: Add privatecnt to revert DMA_PRIVATE property
  dmatest: add dma interrupts and callbacks
  dmatest: add xor test
  dmaengine: allow dma support for async_tx to be toggled
  async_tx: provide __async_inline for HAS_DMA=n archs
  dmaengine: kill some unused headers
  dmaengine: initialize tx_list in dma_async_tx_descriptor_init
  dma: i.MX31 IPU DMA robustness improvements
  dma: improve section assignment in i.MX31 IPU DMA driver
  dma: ipu_idmac driver cosmetic clean-up
  dmaengine: fail device registration if channel registration fails

16 files changed, 925 insertions, 241 deletions

diff --git a/crypto/async_tx/async_tx.c b/crypto/async_tx/async_tx.c
index f21147f3626a..06eb6cc09fef 100644
--- a/crypto/async_tx/async_tx.c
+++ b/crypto/async_tx/async_tx.c
@@ -30,7 +30,7 @@
 #ifdef CONFIG_DMA_ENGINE
 static int __init async_tx_init(void)
 {
-	dmaengine_get();
+	async_dmaengine_get();
 
 	printk(KERN_INFO "async_tx: api initialized (async)\n");
 
@@ -39,7 +39,7 @@ static int __init async_tx_init(void)
 
 static void __exit async_tx_exit(void)
 {
-	dmaengine_put();
+	async_dmaengine_put();
 }
 
 /**
@@ -56,7 +56,7 @@ __async_tx_find_channel(struct dma_async_tx_descriptor *depend_tx,
 	if (depend_tx &&
 	    dma_has_cap(tx_type, depend_tx->chan->device->cap_mask))
 		return depend_tx->chan;
-	return dma_find_channel(tx_type);
+	return async_dma_find_channel(tx_type);
 }
 EXPORT_SYMBOL_GPL(__async_tx_find_channel);
 #else
diff --git a/crypto/async_tx/async_xor.c b/crypto/async_tx/async_xor.c
index 595b78672b36..95fe2c8d6c51 100644
--- a/crypto/async_tx/async_xor.c
+++ b/crypto/async_tx/async_xor.c
@@ -30,11 +30,8 @@
 #include <linux/raid/xor.h>
 #include <linux/async_tx.h>
 
-/* do_async_xor - dma map the pages and perform the xor with an engine.
- * 	This routine is marked __always_inline so it can be compiled away
- * 	when CONFIG_DMA_ENGINE=n
- */
-static __always_inline struct dma_async_tx_descriptor *
+/* do_async_xor - dma map the pages and perform the xor with an engine */
+static __async_inline struct dma_async_tx_descriptor *
 do_async_xor(struct dma_chan *chan, struct page *dest, struct page **src_list,
 	     unsigned int offset, int src_cnt, size_t len,
 	     enum async_tx_flags flags,
diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig
index 48ea59e79672..3b3c01b6f1ee 100644
--- a/drivers/dma/Kconfig
+++ b/drivers/dma/Kconfig
@@ -98,6 +98,17 @@ config NET_DMA
 	  Say Y here if you enabled INTEL_IOATDMA or FSL_DMA, otherwise
 	  say N.
 
+config ASYNC_TX_DMA
+	bool "Async_tx: Offload support for the async_tx api"
+	depends on DMA_ENGINE
+	help
+	  This allows the async_tx api to take advantage of offload engines for
+	  memcpy, memset, xor, and raid6 p+q operations.  If your platform has
+	  a dma engine that can perform raid operations and you have enabled
+	  MD_RAID456 say Y.
+
+	  If unsure, say N.
+
 config DMATEST
 	tristate "DMA Test client"
 	depends on DMA_ENGINE
diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c
index 280a9d263eb3..92438e9dacc3 100644
--- a/drivers/dma/dmaengine.c
+++ b/drivers/dma/dmaengine.c
@@ -507,6 +507,7 @@ struct dma_chan *__dma_request_channel(dma_cap_mask_t *mask, dma_filter_fn fn, v
 			 * published in the general-purpose allocator
 			 */
 			dma_cap_set(DMA_PRIVATE, device->cap_mask);
+			device->privatecnt++;
 			err = dma_chan_get(chan);
 
 			if (err == -ENODEV) {
@@ -518,6 +519,8 @@ struct dma_chan *__dma_request_channel(dma_cap_mask_t *mask, dma_filter_fn fn, v
 				       dma_chan_name(chan), err);
 			else
 				break;
+			if (--device->privatecnt == 0)
+				dma_cap_clear(DMA_PRIVATE, device->cap_mask);
 			chan->private = NULL;
 			chan = NULL;
 		}
@@ -537,6 +540,9 @@ void dma_release_channel(struct dma_chan *chan)
 	WARN_ONCE(chan->client_count != 1,
 		  "chan reference count %d != 1\n", chan->client_count);
 	dma_chan_put(chan);
+	/* drop PRIVATE cap enabled by __dma_request_channel() */
+	if (--chan->device->privatecnt == 0)
+		dma_cap_clear(DMA_PRIVATE, chan->device->cap_mask);
 	chan->private = NULL;
 	mutex_unlock(&dma_list_mutex);
 }
@@ -602,6 +608,24 @@ void dmaengine_put(void)
 }
 EXPORT_SYMBOL(dmaengine_put);
 
+static int get_dma_id(struct dma_device *device)
+{
+	int rc;
+
+ idr_retry:
+	if (!idr_pre_get(&dma_idr, GFP_KERNEL))
+		return -ENOMEM;
+	mutex_lock(&dma_list_mutex);
+	rc = idr_get_new(&dma_idr, NULL, &device->dev_id);
+	mutex_unlock(&dma_list_mutex);
+	if (rc == -EAGAIN)
+		goto idr_retry;
+	else if (rc != 0)
+		return rc;
+
+	return 0;
+}
+
 /**
  * dma_async_device_register - registers DMA devices found
  * @device: &dma_device
@@ -640,27 +664,25 @@ int dma_async_device_register(struct dma_device *device)
 	idr_ref = kmalloc(sizeof(*idr_ref), GFP_KERNEL);
 	if (!idr_ref)
 		return -ENOMEM;
-	atomic_set(idr_ref, 0);
- idr_retry:
-	if (!idr_pre_get(&dma_idr, GFP_KERNEL))
-		return -ENOMEM;
-	mutex_lock(&dma_list_mutex);
-	rc = idr_get_new(&dma_idr, NULL, &device->dev_id);
-	mutex_unlock(&dma_list_mutex);
-	if (rc == -EAGAIN)
-		goto idr_retry;
-	else if (rc != 0)
+	rc = get_dma_id(device);
+	if (rc != 0) {
+		kfree(idr_ref);
 		return rc;
+	}
+
+	atomic_set(idr_ref, 0);
 
 	/* represent channels in sysfs. Probably want devs too */
 	list_for_each_entry(chan, &device->channels, device_node) {
+		rc = -ENOMEM;
 		chan->local = alloc_percpu(typeof(*chan->local));
 		if (chan->local == NULL)
-			continue;
+			goto err_out;
 		chan->dev = kzalloc(sizeof(*chan->dev), GFP_KERNEL);
 		if (chan->dev == NULL) {
 			free_percpu(chan->local);
-			continue;
+			chan->local = NULL;
+			goto err_out;
 		}
 
 		chan->chan_id = chancnt++;
@@ -677,6 +699,8 @@ int dma_async_device_register(struct dma_device *device)
 		if (rc) {
 			free_percpu(chan->local);
 			chan->local = NULL;
+			kfree(chan->dev);
+			atomic_dec(idr_ref);
 			goto err_out;
 		}
 		chan->client_count = 0;
@@ -701,12 +725,23 @@ int dma_async_device_register(struct dma_device *device)
 			}
 		}
 	list_add_tail_rcu(&device->global_node, &dma_device_list);
+	if (dma_has_cap(DMA_PRIVATE, device->cap_mask))
+		device->privatecnt++;	/* Always private */
 	dma_channel_rebalance();
 	mutex_unlock(&dma_list_mutex);
 
 	return 0;
 
 err_out:
+	/* if we never registered a channel just release the idr */
+	if (atomic_read(idr_ref) == 0) {
+		mutex_lock(&dma_list_mutex);
+		idr_remove(&dma_idr, device->dev_id);
+		mutex_unlock(&dma_list_mutex);
+		kfree(idr_ref);
+		return rc;
+	}
+
 	list_for_each_entry(chan, &device->channels, device_node) {
 		if (chan->local == NULL)
 			continue;
@@ -893,6 +928,7 @@ void dma_async_tx_descriptor_init(struct dma_async_tx_descriptor *tx,
 {
 	tx->chan = chan;
 	spin_lock_init(&tx->lock);
+	INIT_LIST_HEAD(&tx->tx_list);
 }
 EXPORT_SYMBOL(dma_async_tx_descriptor_init);
 
diff --git a/drivers/dma/dmatest.c b/drivers/dma/dmatest.c
index e190d8b30700..a27c0fb1bc11 100644
--- a/drivers/dma/dmatest.c
+++ b/drivers/dma/dmatest.c
@@ -38,6 +38,11 @@ module_param(max_channels, uint, S_IRUGO);
 MODULE_PARM_DESC(max_channels,
 		"Maximum number of channels to use (default: all)");
 
+static unsigned int xor_sources = 3;
+module_param(xor_sources, uint, S_IRUGO);
+MODULE_PARM_DESC(xor_sources,
+		"Number of xor source buffers (default: 3)");
+
 /*
  * Initialization patterns. All bytes in the source buffer has bit 7
  * set, all bytes in the destination buffer has bit 7 cleared.
@@ -59,8 +64,9 @@ struct dmatest_thread {
 	struct list_head	node;
 	struct task_struct	*task;
 	struct dma_chan		*chan;
-	u8			*srcbuf;
-	u8			*dstbuf;
+	u8			**srcs;
+	u8			**dsts;
+	enum dma_transaction_type type;
 };
 
 struct dmatest_chan {
@@ -98,30 +104,37 @@ static unsigned long dmatest_random(void)
 	return buf;
 }
 
-static void dmatest_init_srcbuf(u8 *buf, unsigned int start, unsigned int len)
+static void dmatest_init_srcs(u8 **bufs, unsigned int start, unsigned int len)
 {
 	unsigned int i;
-
-	for (i = 0; i < start; i++)
-		buf[i] = PATTERN_SRC | (~i & PATTERN_COUNT_MASK);
-	for ( ; i < start + len; i++)
-		buf[i] = PATTERN_SRC | PATTERN_COPY
-			| (~i & PATTERN_COUNT_MASK);;
-	for ( ; i < test_buf_size; i++)
-		buf[i] = PATTERN_SRC | (~i & PATTERN_COUNT_MASK);
+	u8 *buf;
+
+	for (; (buf = *bufs); bufs++) {
+		for (i = 0; i < start; i++)
+			buf[i] = PATTERN_SRC | (~i & PATTERN_COUNT_MASK);
+		for ( ; i < start + len; i++)
+			buf[i] = PATTERN_SRC | PATTERN_COPY
+				| (~i & PATTERN_COUNT_MASK);;
+		for ( ; i < test_buf_size; i++)
+			buf[i] = PATTERN_SRC | (~i & PATTERN_COUNT_MASK);
+		buf++;
+	}
 }
 
-static void dmatest_init_dstbuf(u8 *buf, unsigned int start, unsigned int len)
+static void dmatest_init_dsts(u8 **bufs, unsigned int start, unsigned int len)
 {
 	unsigned int i;
-
-	for (i = 0; i < start; i++)
-		buf[i] = PATTERN_DST | (~i & PATTERN_COUNT_MASK);
-	for ( ; i < start + len; i++)
-		buf[i] = PATTERN_DST | PATTERN_OVERWRITE
-			| (~i & PATTERN_COUNT_MASK);
-	for ( ; i < test_buf_size; i++)
-		buf[i] = PATTERN_DST | (~i & PATTERN_COUNT_MASK);
+	u8 *buf;
+
+	for (; (buf = *bufs); bufs++) {
+		for (i = 0; i < start; i++)
+			buf[i] = PATTERN_DST | (~i & PATTERN_COUNT_MASK);
+		for ( ; i < start + len; i++)
+			buf[i] = PATTERN_DST | PATTERN_OVERWRITE
+				| (~i & PATTERN_COUNT_MASK);
+		for ( ; i < test_buf_size; i++)
+			buf[i] = PATTERN_DST | (~i & PATTERN_COUNT_MASK);
+	}
 }
 
 static void dmatest_mismatch(u8 actual, u8 pattern, unsigned int index,
@@ -150,23 +163,30 @@ static void dmatest_mismatch(u8 actual, u8 pattern, unsigned int index,
 				thread_name, index, expected, actual);
 }
 
-static unsigned int dmatest_verify(u8 *buf, unsigned int start,
+static unsigned int dmatest_verify(u8 **bufs, unsigned int start,
 		unsigned int end, unsigned int counter, u8 pattern,
 		bool is_srcbuf)
 {
 	unsigned int i;
 	unsigned int error_count = 0;
 	u8 actual;
-
-	for (i = start; i < end; i++) {
-		actual = buf[i];
-		if (actual != (pattern | (~counter & PATTERN_COUNT_MASK))) {
-			if (error_count < 32)
-				dmatest_mismatch(actual, pattern, i, counter,
-						is_srcbuf);
-			error_count++;
+	u8 expected;
+	u8 *buf;
+	unsigned int counter_orig = counter;
+
+	for (; (buf = *bufs); bufs++) {
+		counter = counter_orig;
+		for (i = start; i < end; i++) {
+			actual = buf[i];
+			expected = pattern | (~counter & PATTERN_COUNT_MASK);
+			if (actual != expected) {
+				if (error_count < 32)
+					dmatest_mismatch(actual, pattern, i,
+							 counter, is_srcbuf);
+				error_count++;
+			}
+			counter++;
 		}
-		counter++;
 	}
 
 	if (error_count > 32)
@@ -176,12 +196,17 @@ static unsigned int dmatest_verify(u8 *buf, unsigned int start,
 	return error_count;
 }
 
+static void dmatest_callback(void *completion)
+{
+	complete(completion);
+}
+
 /*
  * This function repeatedly tests DMA transfers of various lengths and
- * offsets until it is told to exit by kthread_stop(). There may be
- * multiple threads running this function in parallel for a single
- * channel, and there may be multiple channels being tested in
- * parallel.
+ * offsets for a given operation type until it is told to exit by
+ * kthread_stop(). There may be multiple threads running this function
+ * in parallel for a single channel, and there may be multiple channels
+ * being tested in parallel.
  *
  * Before each test, the source and destination buffer is initialized
  * with a known pattern. This pattern is different depending on
@@ -201,25 +226,57 @@ static int dmatest_func(void *data)
 	unsigned int		total_tests = 0;
 	dma_cookie_t		cookie;
 	enum dma_status		status;
+	enum dma_ctrl_flags 	flags;
 	int			ret;
+	int			src_cnt;
+	int			dst_cnt;
+	int			i;
 
 	thread_name = current->comm;
 
 	ret = -ENOMEM;
-	thread->srcbuf = kmalloc(test_buf_size, GFP_KERNEL);
-	if (!thread->srcbuf)
-		goto err_srcbuf;
-	thread->dstbuf = kmalloc(test_buf_size, GFP_KERNEL);
-	if (!thread->dstbuf)
-		goto err_dstbuf;
 
 	smp_rmb();
 	chan = thread->chan;
+	if (thread->type == DMA_MEMCPY)
+		src_cnt = dst_cnt = 1;
+	else if (thread->type == DMA_XOR) {
+		src_cnt = xor_sources | 1; /* force odd to ensure dst = src */
+		dst_cnt = 1;
+	} else
+		goto err_srcs;
+
+	thread->srcs = kcalloc(src_cnt+1, sizeof(u8 *), GFP_KERNEL);
+	if (!thread->srcs)
+		goto err_srcs;
+	for (i = 0; i < src_cnt; i++) {
+		thread->srcs[i] = kmalloc(test_buf_size, GFP_KERNEL);
+		if (!thread->srcs[i])
+			goto err_srcbuf;
+	}
+	thread->srcs[i] = NULL;
+
+	thread->dsts = kcalloc(dst_cnt+1, sizeof(u8 *), GFP_KERNEL);
+	if (!thread->dsts)
+		goto err_dsts;
+	for (i = 0; i < dst_cnt; i++) {
+		thread->dsts[i] = kmalloc(test_buf_size, GFP_KERNEL);
+		if (!thread->dsts[i])
+			goto err_dstbuf;
+	}
+	thread->dsts[i] = NULL;
+
+	set_user_nice(current, 10);
+
+	flags = DMA_CTRL_ACK | DMA_COMPL_SKIP_DEST_UNMAP | DMA_PREP_INTERRUPT;
 
 	while (!kthread_should_stop()) {
 		struct dma_device *dev = chan->device;
-		struct dma_async_tx_descriptor *tx;
-		dma_addr_t dma_src, dma_dest;
+		struct dma_async_tx_descriptor *tx = NULL;
+		dma_addr_t dma_srcs[src_cnt];
+		dma_addr_t dma_dsts[dst_cnt];
+		struct completion cmp;
+		unsigned long tmo = msecs_to_jiffies(3000);
 
 		total_tests++;
 
@@ -227,22 +284,41 @@ static int dmatest_func(void *data)
 		src_off = dmatest_random() % (test_buf_size - len + 1);
 		dst_off = dmatest_random() % (test_buf_size - len + 1);
 
-		dmatest_init_srcbuf(thread->srcbuf, src_off, len);
-		dmatest_init_dstbuf(thread->dstbuf, dst_off, len);
+		dmatest_init_srcs(thread->srcs, src_off, len);
+		dmatest_init_dsts(thread->dsts, dst_off, len);
 
-		dma_src = dma_map_single(dev->dev, thread->srcbuf + src_off,
-				len, DMA_TO_DEVICE);
+		for (i = 0; i < src_cnt; i++) {
+			u8 *buf = thread->srcs[i] + src_off;
+
+			dma_srcs[i] = dma_map_single(dev->dev, buf, len,
+						     DMA_TO_DEVICE);
+		}
 		/* map with DMA_BIDIRECTIONAL to force writeback/invalidate */
-		dma_dest = dma_map_single(dev->dev, thread->dstbuf,
-				test_buf_size, DMA_BIDIRECTIONAL);
+		for (i = 0; i < dst_cnt; i++) {
+			dma_dsts[i] = dma_map_single(dev->dev, thread->dsts[i],
+						     test_buf_size,
+						     DMA_BIDIRECTIONAL);
+		}
+
+		if (thread->type == DMA_MEMCPY)
+			tx = dev->device_prep_dma_memcpy(chan,
+							 dma_dsts[0] + dst_off,
+							 dma_srcs[0], len,
+							 flags);
+		else if (thread->type == DMA_XOR)
+			tx = dev->device_prep_dma_xor(chan,
+						      dma_dsts[0] + dst_off,
+						      dma_srcs, xor_sources,
+						      len, flags);
 
-		tx = dev->device_prep_dma_memcpy(chan, dma_dest + dst_off,
-				dma_src, len,
-				DMA_CTRL_ACK | DMA_COMPL_SKIP_DEST_UNMAP);
 		if (!tx) {
-			dma_unmap_single(dev->dev, dma_src, len, DMA_TO_DEVICE);
-			dma_unmap_single(dev->dev, dma_dest,
-					test_buf_size, DMA_BIDIRECTIONAL);
+			for (i = 0; i < src_cnt; i++)
+				dma_unmap_single(dev->dev, dma_srcs[i], len,
+						 DMA_TO_DEVICE);
+			for (i = 0; i < dst_cnt; i++)
+				dma_unmap_single(dev->dev, dma_dsts[i],
+						 test_buf_size,
+						 DMA_BIDIRECTIONAL);
 			pr_warning("%s: #%u: prep error with src_off=0x%x "
 					"dst_off=0x%x len=0x%x\n",
 					thread_name, total_tests - 1,
@@ -251,7 +327,10 @@ static int dmatest_func(void *data)
 			failed_tests++;
 			continue;
 		}
-		tx->callback = NULL;
+
+		init_completion(&cmp);
+		tx->callback = dmatest_callback;
+		tx->callback_param = &cmp;
 		cookie = tx->tx_submit(tx);
 
 		if (dma_submit_error(cookie)) {
@@ -263,44 +342,50 @@ static int dmatest_func(void *data)
 			failed_tests++;
 			continue;
 		}
-		dma_async_memcpy_issue_pending(chan);
+		dma_async_issue_pending(chan);
 
-		do {
-			msleep(1);
-			status = dma_async_memcpy_complete(
-					chan, cookie, NULL, NULL);
-		} while (status == DMA_IN_PROGRESS);
+		tmo = wait_for_completion_timeout(&cmp, tmo);
+		status = dma_async_is_tx_complete(chan, cookie, NULL, NULL);
 
-		if (status == DMA_ERROR) {
-			pr_warning("%s: #%u: error during copy\n",
-					thread_name, total_tests - 1);
+		if (tmo == 0) {
+			pr_warning("%s: #%u: test timed out\n",
+				   thread_name, total_tests - 1);
+			failed_tests++;
+			continue;
+		} else if (status != DMA_SUCCESS) {
+			pr_warning("%s: #%u: got completion callback,"
+				   " but status is \'%s\'\n",
+				   thread_name, total_tests - 1,
+				   status == DMA_ERROR ? "error" : "in progress");
 			failed_tests++;
 			continue;
 		}
+
 		/* Unmap by myself (see DMA_COMPL_SKIP_DEST_UNMAP above) */
-		dma_unmap_single(dev->dev, dma_dest,
-				test_buf_size, DMA_BIDIRECTIONAL);
+		for (i = 0; i < dst_cnt; i++)
+			dma_unmap_single(dev->dev, dma_dsts[i], test_buf_size,
+					 DMA_BIDIRECTIONAL);
 
 		error_count = 0;
 
 		pr_debug("%s: verifying source buffer...\n", thread_name);
-		error_count += dmatest_verify(thread->srcbuf, 0, src_off,
+		error_count += dmatest_verify(thread->srcs, 0, src_off,
 				0, PATTERN_SRC, true);
-		error_count += dmatest_verify(thread->srcbuf, src_off,
+		error_count += dmatest_verify(thread->srcs, src_off,
 				src_off + len, src_off,
 				PATTERN_SRC | PATTERN_COPY, true);
-		error_count += dmatest_verify(thread->srcbuf, src_off + len,
+		error_count += dmatest_verify(thread->srcs, src_off + len,
 				test_buf_size, src_off + len,
 				PATTERN_SRC, true);
 
 		pr_debug("%s: verifying dest buffer...\n",
 				thread->task->comm);
-		error_count += dmatest_verify(thread->dstbuf, 0, dst_off,
+		error_count += dmatest_verify(thread->dsts, 0, dst_off,
 				0, PATTERN_DST, false);
-		error_count += dmatest_verify(thread->dstbuf, dst_off,
+		error_count += dmatest_verify(thread->dsts, dst_off,
 				dst_off + len, src_off,
 				PATTERN_SRC | PATTERN_COPY, false);
-		error_count += dmatest_verify(thread->dstbuf, dst_off + len,
+		error_count += dmatest_verify(thread->dsts, dst_off + len,
 				test_buf_size, dst_off + len,
 				PATTERN_DST, false);
 
@@ -319,10 +404,16 @@ static int dmatest_func(void *data)
 	}
 
 	ret = 0;
-	kfree(thread->dstbuf);
+	for (i = 0; thread->dsts[i]; i++)
+		kfree(thread->dsts[i]);
 err_dstbuf:
-	kfree(thread->srcbuf);
+	kfree(thread->dsts);
+err_dsts:
+	for (i = 0; thread->srcs[i]; i++)
+		kfree(thread->srcs[i]);
 err_srcbuf:
+	kfree(thread->srcs);
+err_srcs:
 	pr_notice("%s: terminating after %u tests, %u failures (status %d)\n",
 			thread_name, total_tests, failed_tests, ret);
 	return ret;
@@ -344,35 +435,36 @@ static void dmatest_cleanup_channel(struct dmatest_chan *dtc)
 	kfree(dtc);
 }
 
-static int dmatest_add_channel(struct dma_chan *chan)
+static int dmatest_add_threads(struct dmatest_chan *dtc, enum dma_transaction_type type)
 {
-	struct dmatest_chan	*dtc;
-	struct dmatest_thread	*thread;
-	unsigned int		i;
-
-	dtc = kmalloc(sizeof(struct dmatest_chan), GFP_KERNEL);
-	if (!dtc) {
-		pr_warning("dmatest: No memory for %s\n", dma_chan_name(chan));
-		return -ENOMEM;
-	}
+	struct dmatest_thread *thread;
+	struct dma_chan *chan = dtc->chan;
+	char *op;
+	unsigned int i;
 
-	dtc->chan = chan;
-	INIT_LIST_HEAD(&dtc->threads);
+	if (type == DMA_MEMCPY)
+		op = "copy";
+	else if (type == DMA_XOR)
+		op = "xor";
+	else
+		return -EINVAL;
 
 	for (i = 0; i < threads_per_chan; i++) {
 		thread = kzalloc(sizeof(struct dmatest_thread), GFP_KERNEL);
 		if (!thread) {
-			pr_warning("dmatest: No memory for %s-test%u\n",
-				   dma_chan_name(chan), i);
+			pr_warning("dmatest: No memory for %s-%s%u\n",
+				   dma_chan_name(chan), op, i);
+
 			break;
 		}
 		thread->chan = dtc->chan;
+		thread->type = type;
 		smp_wmb();
-		thread->task = kthread_run(dmatest_func, thread, "%s-test%u",
-				dma_chan_name(chan), i);
+		thread->task = kthread_run(dmatest_func, thread, "%s-%s%u",
+				dma_chan_name(chan), op, i);
 		if (IS_ERR(thread->task)) {
-			pr_warning("dmatest: Failed to run thread %s-test%u\n",
-					dma_chan_name(chan), i);
+			pr_warning("dmatest: Failed to run thread %s-%s%u\n",
+					dma_chan_name(chan), op, i);
 			kfree(thread);
 			break;
 		}
@@ -382,7 +474,36 @@ static int dmatest_add_channel(struct dma_chan *chan)
 		list_add_tail(&thread->node, &dtc->threads);
 	}
 
-	pr_info("dmatest: Started %u threads using %s\n", i, dma_chan_name(chan));
+	return i;
+}
+
+static int dmatest_add_channel(struct dma_chan *chan)
+{
+	struct dmatest_chan	*dtc;
+	struct dma_device	*dma_dev = chan->device;
+	unsigned int		thread_count = 0;
+	unsigned int		cnt;
+
+	dtc = kmalloc(sizeof(struct dmatest_chan), GFP_KERNEL);
+	if (!dtc) {
+		pr_warning("dmatest: No memory for %s\n", dma_chan_name(chan));
+		return -ENOMEM;
+	}
+
+	dtc->chan = chan;
+	INIT_LIST_HEAD(&dtc->threads);
+
+	if (dma_has_cap(DMA_MEMCPY, dma_dev->cap_mask)) {
+		cnt = dmatest_add_threads(dtc, DMA_MEMCPY);
+		thread_count += cnt > 0 ?: 0;
+	}
+	if (dma_has_cap(DMA_XOR, dma_dev->cap_mask)) {
+		cnt = dmatest_add_threads(dtc, DMA_XOR);
+		thread_count += cnt > 0 ?: 0;
+	}
+
+	pr_info("dmatest: Started %u threads using %s\n",
+		thread_count, dma_chan_name(chan));
 
 	list_add_tail(&dtc->node, &dmatest_channels);
 	nr_channels++;
diff --git a/drivers/dma/dw_dmac.c b/drivers/dma/dw_dmac.c
index 20ad3d26bec2..98c9a847bf51 100644
--- a/drivers/dma/dw_dmac.c
+++ b/drivers/dma/dw_dmac.c
@@ -363,6 +363,82 @@ static void dwc_handle_error(struct dw_dma *dw, struct dw_dma_chan *dwc)
 	dwc_descriptor_complete(dwc, bad_desc);
 }
 
+/* --------------------- Cyclic DMA API extensions -------------------- */
+
+inline dma_addr_t dw_dma_get_src_addr(struct dma_chan *chan)
+{
+	struct dw_dma_chan *dwc = to_dw_dma_chan(chan);
+	return channel_readl(dwc, SAR);
+}
+EXPORT_SYMBOL(dw_dma_get_src_addr);
+
+inline dma_addr_t dw_dma_get_dst_addr(struct dma_chan *chan)
+{
+	struct dw_dma_chan *dwc = to_dw_dma_chan(chan);
+	return channel_readl(dwc, DAR);
+}
+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)
+{
+	if (status_block & 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;
+		if (callback) {
+			spin_unlock(&dwc->lock);
+			callback(callback_param);
+			spin_lock(&dwc->lock);
+		}
+	}
+
+	/*
+	 * Error and transfer complete are highly unlikely, and will most
+	 * likely be due to a configuration error by the user.
+	 */
+	if (unlikely(status_err & dwc->mask) ||
+			unlikely(status_xfer & dwc->mask)) {
+		int i;
+
+		dev_err(chan2dev(&dwc->chan), "cyclic DMA unexpected %s "
+				"interrupt, stopping DMA transfer\n",
+				status_xfer ? "xfer" : "error");
+		dev_err(chan2dev(&dwc->chan),
+			"  SAR: 0x%x DAR: 0x%x LLP: 0x%x CTL: 0x%x:%08x\n",
+			channel_readl(dwc, SAR),
+			channel_readl(dwc, DAR),
+			channel_readl(dwc, LLP),
+			channel_readl(dwc, CTL_HI),
+			channel_readl(dwc, CTL_LO));
+
+		channel_clear_bit(dw, CH_EN, dwc->mask);
+		while (dma_readl(dw, CH_EN) & dwc->mask)
+			cpu_relax();
+
+		/* make sure DMA does not restart by loading a new list */
+		channel_writel(dwc, LLP, 0);
+		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);
+
+		for (i = 0; i < dwc->cdesc->periods; i++)
+			dwc_dump_lli(dwc, &dwc->cdesc->desc[i]->lli);
+	}
+}
+
+/* ------------------------------------------------------------------------- */
+
 static void dw_dma_tasklet(unsigned long data)
 {
 	struct dw_dma *dw = (struct dw_dma *)data;
@@ -382,7 +458,10 @@ 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 (status_err & (1 << i))
+		if (test_bit(DW_DMA_IS_CYCLIC, &dwc->flags))
+			dwc_handle_cyclic(dw, dwc, status_block, status_err,
+					status_xfer);
+		else if (status_err & (1 << i))
 			dwc_handle_error(dw, dwc);
 		else if ((status_block | status_xfer) & (1 << i))
 			dwc_scan_descriptors(dw, dwc);
@@ -826,7 +905,6 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
 		dma_async_tx_descriptor_init(&desc->txd, chan);
 		desc->txd.tx_submit = dwc_tx_submit;
 		desc->txd.flags = DMA_CTRL_ACK;
-		INIT_LIST_HEAD(&desc->txd.tx_list);
 		desc->txd.phys = dma_map_single(chan2parent(chan), &desc->lli,
 				sizeof(desc->lli), DMA_TO_DEVICE);
 		dwc_desc_put(dwc, desc);
@@ -884,6 +962,257 @@ static void dwc_free_chan_resources(struct dma_chan *chan)
 	dev_vdbg(chan2dev(chan), "free_chan_resources done\n");
 }
 
+/* --------------------- Cyclic DMA API extensions -------------------- */
+
+/**
+ * dw_dma_cyclic_start - start the cyclic DMA transfer
+ * @chan: the DMA channel to start
+ *
+ * Must be called with soft interrupts disabled. Returns zero on success or
+ * -errno on failure.
+ */
+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);
+
+	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);
+
+	/* assert channel is idle */
+	if (dma_readl(dw, CH_EN) & dwc->mask) {
+		dev_err(chan2dev(&dwc->chan),
+			"BUG: Attempted to start non-idle channel\n");
+		dev_err(chan2dev(&dwc->chan),
+			"  SAR: 0x%x DAR: 0x%x LLP: 0x%x CTL: 0x%x:%08x\n",
+			channel_readl(dwc, SAR),
+			channel_readl(dwc, DAR),
+			channel_readl(dwc, LLP),
+			channel_readl(dwc, CTL_HI),
+			channel_readl(dwc, CTL_LO));
+		spin_unlock(&dwc->lock);
+		return -EBUSY;
+	}
+
+	dma_writel(dw, CLEAR.BLOCK, dwc->mask);
+	dma_writel(dw, CLEAR.ERROR, dwc->mask);
+	dma_writel(dw, CLEAR.XFER, dwc->mask);
+
+	/* setup DMAC channel registers */
+	channel_writel(dwc, LLP, dwc->cdesc->desc[0]->txd.phys);
+	channel_writel(dwc, CTL_LO, DWC_CTLL_LLP_D_EN | DWC_CTLL_LLP_S_EN);
+	channel_writel(dwc, CTL_HI, 0);
+
+	channel_set_bit(dw, CH_EN, dwc->mask);
+
+	spin_unlock(&dwc->lock);
+
+	return 0;
+}
+EXPORT_SYMBOL(dw_dma_cyclic_start);
+
+/**
+ * dw_dma_cyclic_stop - stop the cyclic DMA transfer
+ * @chan: the DMA channel to stop
+ *
+ * Must be called with soft interrupts disabled.
+ */
+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);
+
+	spin_lock(&dwc->lock);
+
+	channel_clear_bit(dw, CH_EN, dwc->mask);
+	while (dma_readl(dw, CH_EN) & dwc->mask)
+		cpu_relax();
+
+	spin_unlock(&dwc->lock);
+}
+EXPORT_SYMBOL(dw_dma_cyclic_stop);
+
+/**
+ * dw_dma_cyclic_prep - 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
+ *
+ * Must be called before trying to start the transfer. Returns a valid struct
+ * dw_cyclic_desc if successful or an ERR_PTR(-errno) if not successful.
+ */
+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)
+{
+	struct dw_dma_chan		*dwc = to_dw_dma_chan(chan);
+	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;
+	unsigned int			i;
+
+	spin_lock_bh(&dwc->lock);
+	if (!list_empty(&dwc->queue) || !list_empty(&dwc->active_list)) {
+		spin_unlock_bh(&dwc->lock);
+		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);
+	if (was_cyclic) {
+		dev_dbg(chan2dev(&dwc->chan),
+				"channel already prepared for cyclic DMA\n");
+		return ERR_PTR(-EBUSY);
+	}
+
+	retval = ERR_PTR(-EINVAL);
+	reg_width = dws->reg_width;
+	periods = buf_len / period_len;
+
+	/* Check for too big/unaligned periods and unaligned DMA buffer. */
+	if (period_len > (DWC_MAX_COUNT << reg_width))
+		goto out_err;
+	if (unlikely(period_len & ((1 << reg_width) - 1)))
+		goto out_err;
+	if (unlikely(buf_addr & ((1 << reg_width) - 1)))
+		goto out_err;
+	if (unlikely(!(direction & (DMA_TO_DEVICE | DMA_FROM_DEVICE))))
+		goto out_err;
+
+	retval = ERR_PTR(-ENOMEM);
+
+	if (periods > NR_DESCS_PER_CHANNEL)
+		goto out_err;
+
+	cdesc = kzalloc(sizeof(struct dw_cyclic_desc), GFP_KERNEL);
+	if (!cdesc)
+		goto out_err;
+
+	cdesc->desc = kzalloc(sizeof(struct dw_desc *) * periods, GFP_KERNEL);
+	if (!cdesc->desc)
+		goto out_err_alloc;
+
+	for (i = 0; i < periods; i++) {
+		desc = dwc_desc_get(dwc);
+		if (!desc)
+			goto out_err_desc_get;
+
+		switch (direction) {
+		case DMA_TO_DEVICE:
+			desc->lli.dar = dws->tx_reg;
+			desc->lli.sar = buf_addr + (period_len * i);
+			desc->lli.ctllo = (DWC_DEFAULT_CTLLO
+					| DWC_CTLL_DST_WIDTH(reg_width)
+					| DWC_CTLL_SRC_WIDTH(reg_width)
+					| DWC_CTLL_DST_FIX
+					| DWC_CTLL_SRC_INC
+					| DWC_CTLL_FC_M2P
+					| DWC_CTLL_INT_EN);
+			break;
+		case DMA_FROM_DEVICE:
+			desc->lli.dar = buf_addr + (period_len * i);
+			desc->lli.sar = dws->rx_reg;
+			desc->lli.ctllo = (DWC_DEFAULT_CTLLO
+					| DWC_CTLL_SRC_WIDTH(reg_width)
+					| DWC_CTLL_DST_WIDTH(reg_width)
+					| DWC_CTLL_DST_INC
+					| DWC_CTLL_SRC_FIX
+					| DWC_CTLL_FC_P2M
+					| DWC_CTLL_INT_EN);
+			break;
+		default:
+			break;
+		}
+
+		desc->lli.ctlhi = (period_len >> reg_width);
+		cdesc->desc[i] = desc;
+
+		if (last) {
+			last->lli.llp = desc->txd.phys;
+			dma_sync_single_for_device(chan2parent(chan),
+					last->txd.phys, sizeof(last->lli),
+					DMA_TO_DEVICE);
+		}
+
+		last = desc;
+	}
+
+	/* lets make a cyclic list */
+	last->lli.llp = cdesc->desc[0]->txd.phys;
+	dma_sync_single_for_device(c