Merge commit 'async_tx/next'

12 files changed, 217 insertions, 141 deletions

diff --git a/crypto/async_tx/async_memcpy.c b/crypto/async_tx/async_memcpy.c
index 0f6282207b32..84caa4efc0d4 100644
--- a/crypto/async_tx/async_memcpy.c
+++ b/crypto/async_tx/async_memcpy.c
@@ -66,11 +66,11 @@ async_memcpy(struct page *dest, struct page *src, unsigned int dest_offset,
 	}
 
 	if (tx) {
-		pr_debug("%s: (async) len: %zu\n", __FUNCTION__, len);
+		pr_debug("%s: (async) len: %zu\n", __func__, len);
 		async_tx_submit(chan, tx, flags, depend_tx, cb_fn, cb_param);
 	} else {
 		void *dest_buf, *src_buf;
-		pr_debug("%s: (sync) len: %zu\n", __FUNCTION__, len);
+		pr_debug("%s: (sync) len: %zu\n", __func__, len);
 
 		/* wait for any prerequisite operations */
 		if (depend_tx) {
@@ -80,7 +80,7 @@ async_memcpy(struct page *dest, struct page *src, unsigned int dest_offset,
 			BUG_ON(depend_tx->ack);
 			if (dma_wait_for_async_tx(depend_tx) == DMA_ERROR)
 				panic("%s: DMA_ERROR waiting for depend_tx\n",
-					__FUNCTION__);
+					__func__);
 		}
 
 		dest_buf = kmap_atomic(dest, KM_USER0) + dest_offset;
diff --git a/crypto/async_tx/async_memset.c b/crypto/async_tx/async_memset.c
index 09c0e83664bc..f5ff3906b035 100644
--- a/crypto/async_tx/async_memset.c
+++ b/crypto/async_tx/async_memset.c
@@ -63,11 +63,11 @@ async_memset(struct page *dest, int val, unsigned int offset,
 	}
 
 	if (tx) {
-		pr_debug("%s: (async) len: %zu\n", __FUNCTION__, len);
+		pr_debug("%s: (async) len: %zu\n", __func__, len);
 		async_tx_submit(chan, tx, flags, depend_tx, cb_fn, cb_param);
 	} else { /* run the memset synchronously */
 		void *dest_buf;
-		pr_debug("%s: (sync) len: %zu\n", __FUNCTION__, len);
+		pr_debug("%s: (sync) len: %zu\n", __func__, len);
 
 		dest_buf = (void *) (((char *) page_address(dest)) + offset);
 
@@ -79,7 +79,7 @@ async_memset(struct page *dest, int val, unsigned int offset,
 			BUG_ON(depend_tx->ack);
 			if (dma_wait_for_async_tx(depend_tx) == DMA_ERROR)
 				panic("%s: DMA_ERROR waiting for depend_tx\n",
-					__FUNCTION__);
+					__func__);
 		}
 
 		memset(dest_buf, val, len);
diff --git a/crypto/async_tx/async_tx.c b/crypto/async_tx/async_tx.c
index 562882189de5..69756164b61d 100644
--- a/crypto/async_tx/async_tx.c
+++ b/crypto/async_tx/async_tx.c
@@ -89,13 +89,19 @@ dma_wait_for_async_tx(struct dma_async_tx_descriptor *tx)
 		iter = tx;
 
 		/* find the root of the unsubmitted dependency chain */
-		while (iter->cookie == -EBUSY) {
+		do {
 			parent = iter->parent;
-			if (parent && parent->cookie == -EBUSY)
-				iter = iter->parent;
-			else
+			if (!parent)
 				break;
-		}
+			else
+				iter = parent;
+		} while (parent);
+
+		/* there is a small window for ->parent == NULL and
+		 * ->cookie == -EBUSY
+		 */
+		while (iter->cookie == -EBUSY)
+			cpu_relax();
 
 		status = dma_sync_wait(iter->chan, iter->cookie);
 	} while (status == DMA_IN_PROGRESS || (iter != tx));
@@ -111,24 +117,33 @@ EXPORT_SYMBOL_GPL(dma_wait_for_async_tx);
 void
 async_tx_run_dependencies(struct dma_async_tx_descriptor *tx)
 {
-	struct dma_async_tx_descriptor *dep_tx, *_dep_tx;
-	struct dma_device *dev;
+	struct dma_async_tx_descriptor *next = tx->next;
 	struct dma_chan *chan;
 
-	list_for_each_entry_safe(dep_tx, _dep_tx, &tx->depend_list,
-		depend_node) {
-		chan = dep_tx->chan;
-		dev = chan->device;
-		/* we can't depend on ourselves */
-		BUG_ON(chan == tx->chan);
-		list_del(&dep_tx->depend_node);
-		tx->tx_submit(dep_tx);
-
-		/* we need to poke the engine as client code does not
-		 * know about dependency submission events
-		 */
-		dev->device_issue_pending(chan);
+	if (!next)
+		return;
+
+	tx->next = NULL;
+	chan = next->chan;
+
+	/* keep submitting up until a channel switch is detected
+	 * in that case we will be called again as a result of
+	 * processing the interrupt from async_tx_channel_switch
+	 */
+	while (next && next->chan == chan) {
+		struct dma_async_tx_descriptor *_next;
+
+		spin_lock_bh(&next->lock);
+		next->parent = NULL;
+		_next = next->next;
+		next->next = NULL;
+		spin_unlock_bh(&next->lock);
+
+		next->tx_submit(next);
+		next = _next;
 	}
+
+	chan->device->device_issue_pending(chan);
 }
 EXPORT_SYMBOL_GPL(async_tx_run_dependencies);
 
@@ -397,6 +412,92 @@ static void __exit async_tx_exit(void)
 }
 #endif
 
+
+/**
+ * async_tx_channel_switch - queue an interrupt descriptor with a dependency
+ * 	pre-attached.
+ * @depend_tx: the operation that must finish before the new operation runs
+ * @tx: the new operation
+ */
+static void
+async_tx_channel_switch(struct dma_async_tx_descriptor *depend_tx,
+			struct dma_async_tx_descriptor *tx)
+{
+	struct dma_chan *chan;
+	struct dma_device *device;
+	struct dma_async_tx_descriptor *intr_tx = (void *) ~0;
+
+	/* first check to see if we can still append to depend_tx */
+	spin_lock_bh(&depend_tx->lock);
+	if (depend_tx->parent && depend_tx->chan == tx->chan) {
+		tx->parent = depend_tx;
+		depend_tx->next = tx;
+		intr_tx = NULL;
+	}
+	spin_unlock_bh(&depend_tx->lock);
+
+	if (!intr_tx)
+		return;
+
+	chan = depend_tx->chan;
+	device = chan->device;
+
+	/* see if we can schedule an interrupt
+	 * otherwise poll for completion
+	 */
+	if (dma_has_cap(DMA_INTERRUPT, device->cap_mask))
+		intr_tx = device->device_prep_dma_interrupt(chan);
+	else
+		intr_tx = NULL;
+
+	if (intr_tx) {
+		intr_tx->callback = NULL;
+		intr_tx->callback_param = NULL;
+		tx->parent = intr_tx;
+		/* safe to set ->next outside the lock since we know we are
+		 * not submitted yet
+		 */
+		intr_tx->next = tx;
+
+		/* check if we need to append */
+		spin_lock_bh(&depend_tx->lock);
+		if (depend_tx->parent) {
+			intr_tx->parent = depend_tx;
+			depend_tx->next = intr_tx;
+			async_tx_ack(intr_tx);
+			intr_tx = NULL;
+		}
+		spin_unlock_bh(&depend_tx->lock);
+
+		if (intr_tx) {
+			intr_tx->parent = NULL;
+			intr_tx->tx_submit(intr_tx);
+			async_tx_ack(intr_tx);
+		}
+	} else {
+		if (dma_wait_for_async_tx(depend_tx) == DMA_ERROR)
+			panic("%s: DMA_ERROR waiting for depend_tx\n",
+			      __func__);
+		tx->tx_submit(tx);
+	}
+}
+
+
+/**
+ * submit_disposition - while holding depend_tx->lock we must avoid submitting
+ * 	new operations to prevent a circular locking dependency with
+ * 	drivers that already hold a channel lock when calling
+ * 	async_tx_run_dependencies.
+ * @ASYNC_TX_SUBMITTED: we were able to append the new operation under the lock
+ * @ASYNC_TX_CHANNEL_SWITCH: when the lock is dropped schedule a channel switch
+ * @ASYNC_TX_DIRECT_SUBMIT: when the lock is dropped submit directly
+ */
+enum submit_disposition {
+	ASYNC_TX_SUBMITTED,
+	ASYNC_TX_CHANNEL_SWITCH,
+	ASYNC_TX_DIRECT_SUBMIT,
+};
+
 void
 async_tx_submit(struct dma_chan *chan, struct dma_async_tx_descriptor *tx,
 	enum async_tx_flags flags, struct dma_async_tx_descriptor *depend_tx,
@@ -405,28 +506,54 @@ async_tx_submit(struct dma_chan *chan, struct dma_async_tx_descriptor *tx,
 	tx->callback = cb_fn;
 	tx->callback_param = cb_param;
 
-	/* set this new tx to run after depend_tx if:
-	 * 1/ a dependency exists (depend_tx is !NULL)
-	 * 2/ the tx can not be submitted to the current channel
-	 */
-	if (depend_tx && depend_tx->chan != chan) {
-		/* if ack is already set then we cannot be sure
+	if (depend_tx) {
+		enum submit_disposition s;
+
+		/* sanity check the dependency chain:
+		 * 1/ if ack is already set then we cannot be sure
 		 * we are referring to the correct operation
+		 * 2/ dependencies are 1:1 i.e. two transactions can
+		 * not depend on the same parent
 		 */
-		BUG_ON(depend_tx->ack);
+		BUG_ON(depend_tx->ack || depend_tx->next || tx->parent);
 
-		tx->parent = depend_tx;
+		/* the lock prevents async_tx_run_dependencies from missing
+		 * the setting of ->next when ->parent != NULL
+		 */
 		spin_lock_bh(&depend_tx->lock);
-		list_add_tail(&tx->depend_node, &depend_tx->depend_list);
-		if (depend_tx->cookie == 0) {
-			struct dma_chan *dep_chan = depend_tx->chan;
-			struct dma_device *dep_dev = dep_chan->device;
-			dep_dev->device_dependency_added(dep_chan);
+		if (depend_tx->parent) {
+			/* we have a parent so we can not submit directly
+			 * if we are staying on the same channel: append
+			 * else: channel switch
+			 */
+			if (depend_tx->chan == chan) {
+				tx->parent = depend_tx;
+				depend_tx->next = tx;
+				s = ASYNC_TX_SUBMITTED;
+			} else
+				s = ASYNC_TX_CHANNEL_SWITCH;
+		} else {
+			/* we do not have a parent so we may be able to submit
+			 * directly if we are staying on the same channel
+			 */
+			if (depend_tx->chan == chan)
+				s = ASYNC_TX_DIRECT_SUBMIT;
+			else
+				s = ASYNC_TX_CHANNEL_SWITCH;
 		}
 		spin_unlock_bh(&depend_tx->lock);
 
-		/* schedule an interrupt to trigger the channel switch */
-		async_trigger_callback(ASYNC_TX_ACK, depend_tx, NULL, NULL);
+		switch (s) {
+		case ASYNC_TX_SUBMITTED:
+			break;
+		case ASYNC_TX_CHANNEL_SWITCH:
+			async_tx_channel_switch(depend_tx, tx);
+			break;
+		case ASYNC_TX_DIRECT_SUBMIT:
+			tx->parent = NULL;
+			tx->tx_submit(tx);
+			break;
+		}
 	} else {
 		tx->parent = NULL;
 		tx->tx_submit(tx);
@@ -472,11 +599,11 @@ async_trigger_callback(enum async_tx_flags flags,
 		tx = NULL;
 
 	if (tx) {
-		pr_debug("%s: (async)\n", __FUNCTION__);
+		pr_debug("%s: (async)\n", __func__);
 
 		async_tx_submit(chan, tx, flags, depend_tx, cb_fn, cb_param);
 	} else {
-		pr_debug("%s: (sync)\n", __FUNCTION__);
+		pr_debug("%s: (sync)\n", __func__);
 
 		/* wait for any prerequisite operations */
 		if (depend_tx) {
@@ -486,7 +613,7 @@ async_trigger_callback(enum async_tx_flags flags,
 			BUG_ON(depend_tx->ack);
 			if (dma_wait_for_async_tx(depend_tx) == DMA_ERROR)
 				panic("%s: DMA_ERROR waiting for depend_tx\n",
-					__FUNCTION__);
+					__func__);
 		}
 
 		async_tx_sync_epilog(flags, depend_tx, cb_fn, cb_param);
diff --git a/crypto/async_tx/async_xor.c b/crypto/async_tx/async_xor.c
index 2259a4ff15cb..7a9db353f198 100644
--- a/crypto/async_tx/async_xor.c
+++ b/crypto/async_tx/async_xor.c
@@ -47,7 +47,7 @@ do_async_xor(struct dma_device *device,
 	int i;
 	unsigned long dma_prep_flags = cb_fn ? DMA_PREP_INTERRUPT : 0;
 
-	pr_debug("%s: len: %zu\n", __FUNCTION__, len);
+	pr_debug("%s: len: %zu\n", __func__, len);
 
 	dma_dest = dma_map_page(device->dev, dest, offset, len,
 				DMA_FROM_DEVICE);
@@ -86,7 +86,7 @@ do_sync_xor(struct page *dest, struct page **src_list, unsigned int offset,
 	void *_dest;
 	int i;
 
-	pr_debug("%s: len: %zu\n", __FUNCTION__, len);
+	pr_debug("%s: len: %zu\n", __func__, len);
 
 	/* reuse the 'src_list' array to convert to buffer pointers */
 	for (i = 0; i < src_cnt; i++)
@@ -196,7 +196,7 @@ async_xor(struct page *dest, struct page **src_list, unsigned int offset,
 					DMA_ERROR)
 					panic("%s: DMA_ERROR waiting for "
 						"depend_tx\n",
-						__FUNCTION__);
+						__func__);
 			}
 
 			do_sync_xor(dest, &src_list[src_off], offset,
@@ -276,7 +276,7 @@ async_xor_zero_sum(struct page *dest, struct page **src_list,
 		unsigned long dma_prep_flags = cb_fn ? DMA_PREP_INTERRUPT : 0;
 		int i;
 
-		pr_debug("%s: (async) len: %zu\n", __FUNCTION__, len);
+		pr_debug("%s: (async) len: %zu\n", __func__, len);
 
 		for (i = 0; i < src_cnt; i++)
 			dma_src[i] = dma_map_page(device->dev, src_list[i],
@@ -299,7 +299,7 @@ async_xor_zero_sum(struct page *dest, struct page **src_list,
 	} else {
 		unsigned long xor_flags = flags;
 
-		pr_debug("%s: (sync) len: %zu\n", __FUNCTION__, len);
+		pr_debug("%s: (sync) len: %zu\n", __func__, len);
 
 		xor_flags |= ASYNC_TX_XOR_DROP_DST;
 		xor_flags &= ~ASYNC_TX_ACK;
@@ -310,7 +310,7 @@ async_xor_zero_sum(struct page *dest, struct page **src_list,
 		if (tx) {
 			if (dma_wait_for_async_tx(tx) == DMA_ERROR)
 				panic("%s: DMA_ERROR waiting for tx\n",
-					__FUNCTION__);
+					__func__);
 			async_tx_ack(tx);
 		}
 
diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c
index 29965231b912..5ca0d947462d 100644
--- a/drivers/dma/dmaengine.c
+++ b/drivers/dma/dmaengine.c
@@ -362,7 +362,6 @@ int dma_async_device_register(struct dma_device *device)
 
 	BUG_ON(!device->device_alloc_chan_resources);
 	BUG_ON(!device->device_free_chan_resources);
-	BUG_ON(!device->device_dependency_added);
 	BUG_ON(!device->device_is_tx_complete);
 	BUG_ON(!device->device_issue_pending);
 	BUG_ON(!device->dev);
@@ -600,8 +599,6 @@ void dma_async_tx_descriptor_init(struct dma_async_tx_descriptor *tx,
 {
 	tx->chan = chan;
 	spin_lock_init(&tx->lock);
-	INIT_LIST_HEAD(&tx->depend_node);
-	INIT_LIST_HEAD(&tx->depend_list);
 }
 EXPORT_SYMBOL(dma_async_tx_descriptor_init);
 
diff --git a/drivers/dma/fsldma.c b/drivers/dma/fsldma.c
index cc9a68158d99..9ee8f5c352b8 100644
--- a/drivers/dma/fsldma.c
+++ b/drivers/dma/fsldma.c
@@ -57,12 +57,12 @@ static void dma_init(struct fsl_dma_chan *fsl_chan)
 
 }
 
-static void set_sr(struct fsl_dma_chan *fsl_chan, dma_addr_t val)
+static void set_sr(struct fsl_dma_chan *fsl_chan, u32 val)
 {
 	DMA_OUT(fsl_chan, &fsl_chan->reg_base->sr, val, 32);
 }
 
-static dma_addr_t get_sr(struct fsl_dma_chan *fsl_chan)
+static u32 get_sr(struct fsl_dma_chan *fsl_chan)
 {
 	return DMA_IN(fsl_chan, &fsl_chan->reg_base->sr, 32);
 }
@@ -436,7 +436,7 @@ static struct dma_async_tx_descriptor *fsl_dma_prep_memcpy(
 		dev_dbg(fsl_chan->dev, "new link desc alloc %p\n", new);
 #endif
 
-		copy = min(len, FSL_DMA_BCR_MAX_CNT);
+		copy = min(len, (size_t)FSL_DMA_BCR_MAX_CNT);
 
 		set_desc_cnt(fsl_chan, &new->hw, copy);
 		set_desc_src(fsl_chan, &new->hw, dma_src);
@@ -581,8 +581,8 @@ static void fsl_chan_xfer_ld_queue(struct fsl_dma_chan *fsl_chan)
 	if (ld_node != &fsl_chan->ld_queue) {
 		/* Get the ld start address from ld_queue */
 		next_dest_addr = to_fsl_desc(ld_node)->async_tx.phys;
-		dev_dbg(fsl_chan->dev, "xfer LDs staring from 0x%016llx\n",
-				(u64)next_dest_addr);
+		dev_dbg(fsl_chan->dev, "xfer LDs staring from %p\n",
+				(void *)next_dest_addr);
 		set_cdar(fsl_chan, next_dest_addr);
 		dma_start(fsl_chan);
 	} else {
@@ -625,13 +625,6 @@ static void fsl_dma_memcpy_issue_pending(struct dma_chan *chan)
 	fsl_chan_xfer_ld_queue(fsl_chan);
 }
 
-static void fsl_dma_dependency_added(struct dma_chan *chan)
-{
-	struct fsl_dma_chan *fsl_chan = to_fsl_chan(chan);
-
-	fsl_chan_ld_cleanup(fsl_chan);
-}
-
 /**
  * fsl_dma_is_complete - Determine the DMA status
  * @fsl_chan : Freescale DMA channel
@@ -662,7 +655,7 @@ static enum dma_status fsl_dma_is_complete(struct dma_chan *chan,
 static irqreturn_t fsl_dma_chan_do_interrupt(int irq, void *data)
 {
 	struct fsl_dma_chan *fsl_chan = (struct fsl_dma_chan *)data;
-	dma_addr_t stat;
+	u32 stat;
 
 	stat = get_sr(fsl_chan);
 	dev_dbg(fsl_chan->dev, "event: channel %d, stat = 0x%x\n",
@@ -681,9 +674,8 @@ static irqreturn_t fsl_dma_chan_do_interrupt(int irq, void *data)
 	 */
 	if (stat & FSL_DMA_SR_EOSI) {
 		dev_dbg(fsl_chan->dev, "event: End-of-segments INT\n");
-		dev_dbg(fsl_chan->dev, "event: clndar 0x%016llx, "
-				"nlndar 0x%016llx\n", (u64)get_cdar(fsl_chan),
-				(u64)get_ndar(fsl_chan));
+		dev_dbg(fsl_chan->dev, "event: clndar %p, nlndar %p\n",
+			(void *)get_cdar(fsl_chan), (void *)get_ndar(fsl_chan));
 		stat &= ~FSL_DMA_SR_EOSI;
 	}
 
@@ -726,12 +718,15 @@ static void dma_do_tasklet(unsigned long data)
 	fsl_chan_ld_cleanup(fsl_chan);
 }
 
+#ifdef FSL_DMA_CALLBACKTEST
 static void fsl_dma_callback_test(struct fsl_dma_chan *fsl_chan)
 {
 	if (fsl_chan)
 		dev_info(fsl_chan->dev, "selftest: callback is ok!\n");
 }
+#endif
 
+#ifdef CONFIG_FSL_DMA_SELFTEST
 static int fsl_dma_self_test(struct fsl_dma_chan *fsl_chan)
 {
 	struct dma_chan *chan;
@@ -837,9 +832,9 @@ static int fsl_dma_self_test(struct fsl_dma_chan *fsl_chan)
 	if (err) {
 		for (i = 0; (*(src + i) == *(dest + i)) && (i < test_size);
 				i++);
-		dev_err(fsl_chan->dev, "selftest: Test failed, data %d/%d is "
+		dev_err(fsl_chan->dev, "selftest: Test failed, data %d/%ld is "
 				"error! src 0x%x, dest 0x%x\n",
-				i, test_size, *(src + i), *(dest + i));
+				i, (long)test_size, *(src + i), *(dest + i));
 	}
 
 free_resources:
@@ -848,6 +843,7 @@ out:
 	kfree(src);
 	return err;
 }
+#endif
 
 static int __devinit of_fsl_dma_chan_probe(struct of_device *dev,
 			const struct of_device_id *match)
@@ -1008,8 +1004,8 @@ static int __devinit of_fsl_dma_probe(struct of_device *dev,
 	}
 
 	dev_info(&dev->dev, "Probe the Freescale DMA driver for %s "
-			"controller at 0x%08x...\n",
-			match->compatible, fdev->reg.start);
+			"controller at %p...\n",
+			match->compatible, (void *)fdev->reg.start);
 	fdev->reg_base = ioremap(fdev->reg.start, fdev->reg.end
 						- fdev->reg.start + 1);
 
@@ -1020,7 +1016,6 @@ static int __devinit of_fsl_dma_probe(struct of_device *dev,
 	fdev->common.device_prep_dma_memcpy = fsl_dma_prep_memcpy;
 	fdev->common.device_is_tx_complete = fsl_dma_is_complete;
 	fdev->common.device_issue_pending = fsl_dma_memcpy_issue_pending;
-	fdev->common.device_dependency_added = fsl_dma_dependency_added;
 	fdev->common.dev = &dev->dev;
 
 	irq = irq_of_parse_and_map(dev->node, 0);
diff --git a/drivers/dma/ioat_dma.c b/drivers/dma/ioat_dma.c
index 4017d9e7acd2..1517fe4e2d14 100644
--- a/drivers/dma/ioat_dma.c
+++ b/drivers/dma/ioat_dma.c
@@ -924,17 +924,6 @@ static void ioat_dma_memcpy_cleanup(struct ioat_dma_chan *ioat_chan)
 	spin_unlock_bh(&ioat_chan->cleanup_lock);
 }
 
-static void ioat_dma_dependency_added(struct dma_chan *chan)
-{
-	struct ioat_dma_chan *ioat_chan = to_ioat_chan(chan);
-	spin_lock_bh(&ioat_chan->desc_lock);
-	if (ioat_chan->pending == 0) {
-		spin_unlock_bh(&ioat_chan->desc_lock);
-		ioat_dma_memcpy_cleanup(ioat_chan);
-	} else
-		spin_unlock_bh(&ioat_chan->desc_lock);
-}
-
 /**
  * ioat_dma_is_complete - poll the status of a IOAT DMA transaction
  * @chan: IOAT DMA channel handle
@@ -1316,7 +1305,6 @@ struct ioatdma_device *ioat_dma_probe(struct pci_dev *pdev,
 
 	dma_cap_set(DMA_MEMCPY, device->common.cap_mask);
 	device->common.device_is_tx_complete = ioat_dma_is_complete;
-	device->common.device_dependency_added = ioat_dma_dependency_added;
 	switch (device->version) {
 	case IOAT_VER_1_2:
 		device->common.device_prep_dma_memcpy = ioat1_dma_prep_memcpy;
diff --git a/drivers/dma/iop-adma.c b/drivers/dma/iop-adma.c
index 3986d54492bd..93252294f32b 100644
--- a/drivers/dma/iop-adma.c
+++ b/drivers/dma/iop-adma.c
@@ -63,7 +63,6 @@ iop_adma_run_tx_complete_actions(struct iop_adma_desc_slot *desc,
 	struct iop_adma_chan *iop_chan, dma_cookie_t cookie)
 {
 	BUG_ON(desc->async_tx.cookie < 0);
-	spin_lock_bh(&desc->async_tx.lock);
 	if (desc->async_tx.cookie > 0) {
 		cookie = desc->async_tx.cookie;
 		desc->async_tx.cookie = 0;
@@ -101,7 +100,6 @@ iop_adma_run_tx_complete_actions(struct iop_adma_desc_slot *desc,
 
 	/* run dependent operations */
 	async_tx_run_dependencies(&desc->async_tx);
-	spin_unlock_bh(&desc->async_tx.lock);
 
 	return cookie;
 }
@@ -140,7 +138,7 @@ static void __iop_adma_slot_cleanup(struct iop_adma_chan *iop_chan)
 	int busy = iop_chan_is_busy(iop_chan);
 	int seen_current = 0, slot_cnt = 0, slots_per_op = 0;
 
-	dev_dbg(iop_chan->device->common.dev, "%s\n", __FUNCTION__);
+	dev_dbg(iop_chan->device->common.dev, "%s\n", __func__);
 	/* free completed slots from the chain starting with
 	 * the oldest descriptor
 	 */
@@ -257,8 +255,6 @@ static void __iop_adma_slot_cleanup(struct iop_adma_chan *iop_chan)
 
 	BUG_ON(!seen_current);
 
-	iop_chan_idle(busy, iop_chan);
-
 	if (cookie > 0) {
 		iop_chan->completed_cookie = cookie;
 		pr_debug("\tcompleted cookie %d\n", cookie);
@@ -275,8 +271,11 @@ iop_adma_slot_cleanup(struct iop_adma_chan *iop_chan)
 
 static void iop_adma_tasklet(unsigned long data)
 {
-	struct iop_adma_chan *chan = (struct iop_adma_chan *) data;
-	__iop_adma_slot_cleanup(chan);
+	struct iop_adma_chan *iop_chan = (struct iop_adma_chan *) data;
+
+	spin_lock(&iop_chan->lock);
+	__iop_adma_slot_cleanup(iop_chan);
+	spin_unlock(&iop_chan->lock);
 }
 
 static struct iop_adma_desc_slot *
@@ -438,7 +437,7 @@ iop_adma_tx_submit(struct dma_async_tx_descriptor *tx)
 	spin_unlock_bh(&iop_chan->lock);
 
 	dev_dbg(iop_chan->device->common.dev, "%s cookie: %d slot: %d\n",
-		__FUNCTION__, sw_desc->async_tx.cookie, sw_desc->idx);
+		__func__, sw_desc->async_tx.cookie, sw_desc->idx);
 
 	return cookie;
 }
@@ -520,7 +519,7 @@ iop_adma_prep_dma_interrupt(struct dma_chan *chan)
 	struct iop_adma_desc_slot *sw_desc, *grp_start;
 	int slot_cnt, slots_per_op;
 
-	dev_dbg(iop_chan->device->common.dev, "%s\n", __FUNCTION__);
+	dev_dbg(iop_chan->device->common.dev, "%s\n", __func__);
 
 	spin_lock_bh(&iop_chan->lock);
 	slot_cnt = iop_chan_interrupt_slot_count(&slots_per_op, iop_chan);
@@ -548,7 +547,7 @@ iop_adma_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dma_dest,
 	BUG_ON(unlikely(len > IOP_ADMA_MAX_BYTE_COUNT));
 
 	dev_dbg(iop_chan->device->common.dev, "%s len: %u\n",
-		__FUNCTION__, len);
+		__func__, len);
 
 	spin_lock_bh(&iop_chan->lock);
 	slot_cnt = iop_chan_memcpy_slot_count(len, &slots_per_op);
@@ -580,7 +579,7 @@ iop_adma_prep_dma_memset(struct dma_chan *chan, dma_addr_t dma_dest,
 	BUG_ON(unlikely(len > IOP_ADMA_MAX_BYTE_COUNT));
 
 	dev_dbg(iop_chan->device->common.dev, "%s len: %u\n",
-		__FUNCTION__, len);
+		__func__, len);
 
 	spin_lock_bh(&iop_chan->lock);
 	slot_cnt = iop_chan_memset_slot_count(len, &slots_per_op);
@@ -614,7 +613,7 @@ iop_adma_prep_dma_xor(struct dma_chan *chan, dma_addr_t dma_dest,
 
 	dev_dbg(iop_chan->device->common.dev,
 		"%s src_cnt: %d len: %u flags: %lx\n",
-		__FUNCTION__, src_cnt, len, flags);
+		__func__, src_cnt, len, flags);
 
 	spin_lock_bh(&iop_chan->lock);
 	slot_cnt = iop_chan_xor_slot_count(len, src_cnt, &slots_per_op);
@@ -648,7 +647,7 @@ iop_adma_prep_dma_zero_sum(struct dma_chan *chan, dma_addr_t *dma_src,
 		return NULL;
 
 	dev_dbg(iop_chan->device->common.dev, "%s src_cnt: %d len: %u\n",
-		__FUNCTION__, src_cnt, len);
+		__func__, src_cnt, len);
 
 	spin_lock_bh(&iop_chan->lock);
 	slot_cnt = iop_chan_zero_sum_slot_count(len, src_cnt, &slots_per_op);
@@ -659,7 +658,7 @@ iop_adma_prep_dma_zero_sum(struct dma_chan *chan, dma_addr_t *dma_src,
 		iop_desc_set_zero_sum_byte_count(grp_start, len);
 		grp_start->xor_check_result = result;
 		pr_debug("\t%s: grp_start->xor_check_result: %p\n",
-			__FUNCTION__, grp_start->xor_check_result);
+			__func__, grp_start->xor_check_result);
 		sw_desc->unmap_src_cnt = src_cnt;
 		sw_desc->unmap_len = len;
 		while (src_cnt--)
@@ -671,12 +670,6 @@ iop_adma_prep_dma_zero_sum(struct dma_chan *chan, dma_addr_t *dma_src,
 	return sw_desc ? &sw_desc->async_tx : NULL;
 }
 
-static void iop_adma_dependency_added(struct dma_chan *chan)
-{
-	struct iop_adma_chan *iop_chan = to_iop_adma_chan(chan);
-	tasklet_schedule(&iop_chan->irq_tasklet);
-}
-
 static void iop_adma_free_chan_resources(struct dma_chan *chan)
 {
 	struct iop_adma_chan *iop_chan = to_iop_adma_chan(chan);
@@ -700,7 +693,7 @@ static void iop_adma_free_chan_resources(struct dma_chan *chan)
 	iop_chan->last_used = NULL;
 
 	dev_dbg(iop_chan->device->common.dev, "%s slots_allocated %d\n",
-		__FUNCTION__, iop_chan->slots_allocated);
+		__func__, iop_chan->slots_allocated);
 	spin_unlock_bh(&iop_chan->lock);
 
 	/* one is ok since we left it on there on purpose */
@@ -753,7 +746,7 @@ static irqreturn_t iop_adma_eot_handler(int irq, void *data)
 {
 	struct iop_adma_chan *chan = data;
 
-	dev_dbg(chan->device->common.dev, "%s\n", __FUNCTION__);
+	dev_dbg(chan->device->common.dev, "%s\n", __func__);
 
 	tasklet_schedule(&chan->irq_tasklet);
 
@@ -766,7 +759,7 @@ static irqreturn_t iop_adma_eoc_handler(int irq, void *data)
 {
 	struct iop_adma_chan *chan = data;
 
-	dev_dbg(chan->device->common.dev, "%s\n", __FUNCTION__);
+	dev_dbg(chan->device->common.dev, "%s\n", __func__);
 
 	tasklet_schedule(&chan->irq_tasklet);
 
@@ -823,7 +816,7 @@ static int __devinit iop_adma_memcpy_self_test(struct iop_adma_device *device)
 	int err = 0;
 	struct iop_adma_chan *iop_chan;
 
-	dev_dbg(device->common.dev, "%s\n", __FUNCTION__);
+	dev_dbg(device->common.dev, "%s\n", __func__);
 
 	src = kzalloc(sizeof(u8) * IOP_ADMA_TEST_SIZE, GFP_KERNEL);
 	if (!src)
@@ -906,7 +899,7 @@ iop_adma_xor_zero_sum_self_test(struct iop_adma_device *device)
 	int err = 0;
 	struct iop_adma_chan *iop_chan;
 
-	dev_dbg(device->common.dev, "%s\n", __FUNCTION__);
+	dev_dbg(device->common.dev, "%s\n", __func__);
 
 	for (src_idx = 0; src_idx < IOP_ADMA_NUM_SRC_TEST; src_idx++) {
 		xor_srcs[src_idx] = alloc_page(GFP_KERNEL);
@@ -1159,7 +1152,7 @@ static int __devinit iop_adma_probe(struct platform_device *pdev)
 	}
 
 	dev_dbg(&pdev->dev, "%s: allocted descriptor pool virt %p phys %p\n",
-		__FUNCTION__, adev->dma_desc_pool_virt,
+		__func__, adev->dma_desc_pool_virt,
 		(void *) adev->dma_desc_pool);
 
 	adev->id = plat_data->hw_id;
@@ -1177,7 +1170,6 @@ static int __devinit iop_adma_probe(struct platform_device *pdev)
 	dma_dev->device_free_chan_resources = iop_adma_free_chan_resources;
 	dma_dev->device_is_tx_complete = iop_adma_is_complete;
 	dma_dev->device_issue_pending = iop_adma_issue_pending;
-	dma_dev->device_dependency_added = iop_adma_dependency_added;
 	dma_dev->dev = &pdev->dev;
 
 	/* set prep routines based on capability */
@@ -1232,9 +1224,6 @@ static int __devinit iop_adma_probe(struct platform_device *pdev)
 	}
 
 	spin_lock_init(&iop_chan->lock);
-	init_timer(&iop_chan->cleanup_watchdog);
-	iop_chan->cleanup_watchdog.data = (unsigned long) iop_chan;
-	iop_chan->cleanup_watchdog.function = iop_adma_tasklet;
 	INIT_LIST_HEAD(&iop_chan->chain);
 	INIT_LIST_HEAD(&iop_chan->all_slots);
 	INIT_RCU_HEAD(&iop_chan->common.rcu);
@@ -1289,7 +1278,7 @@ static void iop_chan_start_null_memcpy(struct iop_adma_chan *iop_chan)
 	dma_cookie_t cookie;
 	int slot_cnt, slots_per_op;
 
-	dev_dbg(iop_chan->device->common.dev, "%s\n", __FUNCTION__);
+	dev_dbg(iop_chan->device->common.dev, "%s\n", __func__);
 
 	spin_lock_bh(&iop_chan->lock);
 	slot_cnt = iop_chan_memcpy_slot_count(0, &slots_per_op);
@@ -1346,7 +1335,7 @@ static void iop_chan_start_null_xor(struct iop_adma_chan *iop_chan)
 	dma_cookie_t cookie;
 	int slot_cnt, slots_per_op;
 
-	dev_dbg(iop_chan->device->common.dev, "%s\n", __FUNCTION__);
+	dev_dbg(iop_chan->device->common.dev, "%s\n", __func__);
 
 	spin_lock_bh(&iop_chan->lock);
 	slot_cnt = iop_chan_xor_slot_count(0, 2, &slots_per_op);
diff --git a/include/asm-arm/arch-iop13xx/adma.h b/include/asm-arm/arch-iop13xx/adma.h
index efd9a5eb1008..90d14ee564f5 100644
--- a/include/asm-arm/arch-iop13xx/adma.h
+++ b/include/asm-arm/arch-iop13xx/adma.h
@@ -454,11 +454,6 @@ static inline void iop_chan_append(struct iop_adma_chan *chan)
 	__raw_writel(adma_accr, ADMA_ACCR(chan));
 }
 
-static inline void iop_chan_idle(int busy, struct iop_adma_chan *chan)
-{
-	do { } while (0);
-}
-
 static inline u32 iop_chan_get_status(struct iop_adma_chan *chan)
 {
 	return __raw_readl(ADMA_ACSR(chan));
diff --git a/include/asm-arm/hardware/iop3xx-adma.h b/include/asm-arm/hardware/iop3xx-adma.h
index 5c529e6a5e3b..84d635b0a71a 100644
--- a/include/asm-arm/hardware/iop3xx-adma.h
+++ b/include/asm-arm/hardware/iop3xx-adma.h
@@ -767,20 +767,12 @@ static inline int iop_desc_get_zero_result(struct iop_adma_desc_slot *desc)
 static inline void iop_chan_append(struct iop_adma_chan *chan)
 {
 	u32 dma_chan_ctrl;
-	/* workaround dropped interrupts on 3xx */
-	mod_timer(&chan->cleanup_watchdog, jiffies + msecs_to_jiffies(3));
 
 	dma_chan_ctrl = __raw_readl(DMA_CCR(chan));
 	dma_chan_ctrl |= 0x2;
 	__raw_writel(dma_chan_ctrl, DMA_CCR(chan));
 }
 
-static inline void iop_chan_idle(int busy, struct iop_adma_chan *chan)
-{
-	if (!busy)
-		del_timer(&chan->cleanup_watchdog);
-}
-
 static inline u32 iop_chan_get_status(struct iop_adma_chan *chan)
 {
 	return __raw_readl(DMA_CSR(chan));
diff --git a/include/asm-arm/hardware/iop_adma.h b/include/asm-arm/hardware/iop_adma.h
index ca8e71f44346..cb7e3611bcba 100644
--- a/include/asm-arm/hardware/iop_adma.h
+++ b/include/asm-arm/hardware/iop_adma.h
@@ -51,7 +51,6 @@ struct iop_adma_device {
  * @common: common dmaengine channel object members
  * @last_used: place holder for allocation to continue from where it left off
  * @all_slots: complete domain of slots usable by the channel
- * @cleanup_watchdog: workaround missed interrupts on iop3xx
  * @slots_allocated: records the actual size of the descriptor slot pool
  * @irq_tasklet: bottom half where iop_adma_slot_cleanup runs
  */
@@ -65,7 +64,6 @@ struct iop_adma_chan {
 	struct dma_chan common;
 	struct iop_adma_desc_slot *last_used;
 	struct list_head all_slots;
-	struct timer_list cleanup_watchdog;
 	int slots_allocated;
 	struct tasklet_struct irq_tasklet;
 };
diff --git a/include/linux/dmaengine.h b/include/linux/dmaengine.h
index 261e43a4c873..4e413c76d440 100644
--- a/include/linux/dmaengine.h
+++ b/include/linux/dmaengine.h
@@ -221,11 +221,9 @@ typedef void (*dma_async_tx_callback)(void *dma_async_param);
  * @callback: routine to call after this operation is complete
  * @callback_param: general parameter to pass to the callback routine
  * ---async_tx api specific fields---
- * @depend_list: at completion this list of transactions are submitted
- * @depend_node: allow this transaction to be executed after another
- *	transaction has completed, possibly on another channel
+ * @next: at completion submit this descriptor
  * @parent: pointer to the next level up in the dependency chain
- * @lock: protect the dependency list
+ * @lock: protect the parent and next pointers
  */
 struct dma_async_tx_descriptor {
 	dma_cookie_t cookie;
@@ -236,8 +234,7 @@ struct dma_async_tx_descriptor {
 	dma_cookie_t (*tx_submit)(struct dma_async_tx_descriptor *tx);
 	dma_async_tx_callback callback;
 	void *callback_param;
-	struct list_head depend_list;
-	struct list_head depend_node;
+	struct dma_async_tx_descriptor *next;
 	struct dma_async_tx_descriptor *parent;
 	spinlock_t lock;
 };
@@ -261,7 +258,6 @@ struct dma_async_tx_descriptor {
  * @device_prep_dma_zero_sum: prepares a zero_sum operation
  * @device_prep_dma_memset: prepares a memset operation
  * @device_prep_dma_interrupt: prepares an end of chain interrupt operation
- * @device_dependency_added: async_tx notifies the channel about new deps
  * @device_issue_pending: push pending transactions to hardware
  */
 struct dma_device {
@@ -296,7 +292,6 @@ struct dma_device {
 	struct dma_async_tx_descriptor *(*device_prep_dma_interrupt)(
 		struct dma_chan *chan);
 
-	void (*device_dependency_added)(struct dma_chan *chan);
 	enum dma_status (*device_is_tx_complete)(struct dma_chan *chan,
 			dma_cookie_t cookie, dma_cookie_t *last,
 			dma_cookie_t *used);