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:
  dmaengine: ack to flags: make use of the unused bits in the 'ack' field
  iop-adma: remove the workaround for missed interrupts on iop3xx
  async_tx: kill ->device_dependency_added
  async_tx: fix multiple dependency submission
  fsldma: Split the MPC83xx event from MPC85xx and refine irq codes.
  fsldma: Remove CONFIG_FSL_DMA_SELFTEST, keep fsl_dma_self_test() running always.

crypto/async_tx/async_memcpy.c

@@ -77,7 +77,7 @@ async_memcpy(struct page *dest, struct page *src, unsigned int dest_offset,
 			/* if ack is already set then we cannot be sure
 			 * we are referring to the correct operation
 			 */
-			BUG_ON(depend_tx->ack);
+			BUG_ON(async_tx_test_ack(depend_tx));
 			if (dma_wait_for_async_tx(depend_tx) == DMA_ERROR)
 				panic("%s: DMA_ERROR waiting for depend_tx\n",
 					__func__);

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, 0);
+	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 +506,55 @@ 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(async_tx_test_ack(depend_tx) || 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);
@@ -467,7 +595,7 @@ async_trigger_callback(enum async_tx_flags flags,
 		if (device && !dma_has_cap(DMA_INTERRUPT, device->cap_mask))
 			device = NULL;
 
-		tx = device ? device->device_prep_dma_interrupt(chan) : NULL;
+		tx = device ? device->device_prep_dma_interrupt(chan, 0) : NULL;
 	} else
 		tx = NULL;
 
@@ -483,7 +611,7 @@ async_trigger_callback(enum async_tx_flags flags,
 			/* if ack is already set then we cannot be sure
 			 * we are referring to the correct operation
 			 */
-			BUG_ON(depend_tx->ack);
+			BUG_ON(async_tx_test_ack(depend_tx));
 			if (dma_wait_for_async_tx(depend_tx) == DMA_ERROR)
 				panic("%s: DMA_ERROR waiting for depend_tx\n",
 					__func__);

crypto/async_tx/async_xor.c

@@ -191,7 +191,7 @@ async_xor(struct page *dest, struct page **src_list, unsigned int offset,
 				/* if ack is already set then we cannot be sure
 				 * we are referring to the correct operation
 				 */
-				BUG_ON(depend_tx->ack);
+				BUG_ON(async_tx_test_ack(depend_tx));
 				if (dma_wait_for_async_tx(depend_tx) ==
 					DMA_ERROR)
 					panic("%s: DMA_ERROR waiting for "

drivers/dma/Kconfig

@@ -46,14 +46,6 @@ config FSL_DMA
 	  MPC8560/40, MPC8555, MPC8548 and MPC8641 processors.
 	  The MPC8349, MPC8360 is also supported.
 
-config FSL_DMA_SELFTEST
-	bool "Enable the self test for each DMA channel"
-	depends on FSL_DMA
-	default y
-	---help---
-	  Enable the self test for each DMA channel. A self test will be
-	  performed after the channel probed to ensure the DMA works well.
-
 config DMA_ENGINE
 	bool
 

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);
@@ -479,15 +478,15 @@ dma_async_memcpy_buf_to_buf(struct dma_chan *chan, void *dest,
 
 	dma_src = dma_map_single(dev->dev, src, len, DMA_TO_DEVICE);
 	dma_dest = dma_map_single(dev->dev, dest, len, DMA_FROM_DEVICE);
-	tx = dev->device_prep_dma_memcpy(chan, dma_dest, dma_src, len, 0);
+	tx = dev->device_prep_dma_memcpy(chan, dma_dest, dma_src, len,
+					 DMA_CTRL_ACK);
 
 	if (!tx) {
 		dma_unmap_single(dev->dev, dma_src, len, DMA_TO_DEVICE);
 		dma_unmap_single(dev->dev, dma_dest, len, DMA_FROM_DEVICE);
 		return -ENOMEM;
 	}
 
-	tx->ack = 1;
 	tx->callback = NULL;
 	cookie = tx->tx_submit(tx);
 
@@ -525,15 +524,15 @@ dma_async_memcpy_buf_to_pg(struct dma_chan *chan, struct page *page,
 
 	dma_src = dma_map_single(dev->dev, kdata, len, DMA_TO_DEVICE);
 	dma_dest = dma_map_page(dev->dev, page, offset, len, DMA_FROM_DEVICE);
-	tx = dev->device_prep_dma_memcpy(chan, dma_dest, dma_src, len, 0);
+	tx = dev->device_prep_dma_memcpy(chan, dma_dest, dma_src, len,
+					 DMA_CTRL_ACK);
 
 	if (!tx) {
 		dma_unmap_single(dev->dev, dma_src, len, DMA_TO_DEVICE);
 		dma_unmap_page(dev->dev, dma_dest, len, DMA_FROM_DEVICE);
 		return -ENOMEM;
 	}
 
-	tx->ack = 1;
 	tx->callback = NULL;
 	cookie = tx->tx_submit(tx);
 
@@ -574,15 +573,15 @@ dma_async_memcpy_pg_to_pg(struct dma_chan *chan, struct page *dest_pg,
 	dma_src = dma_map_page(dev->dev, src_pg, src_off, len, DMA_TO_DEVICE);
 	dma_dest = dma_map_page(dev->dev, dest_pg, dest_off, len,
 				DMA_FROM_DEVICE);
-	tx = dev->device_prep_dma_memcpy(chan, dma_dest, dma_src, len, 0);
+	tx = dev->device_prep_dma_memcpy(chan, dma_dest, dma_src, len,
+					 DMA_CTRL_ACK);
 
 	if (!tx) {
 		dma_unmap_page(dev->dev, dma_src, len, DMA_TO_DEVICE);
 		dma_unmap_page(dev->dev, dma_dest, len, DMA_FROM_DEVICE);
 		return -ENOMEM;
 	}
 
-	tx->ack = 1;
 	tx->callback = NULL;
 	cookie = tx->tx_submit(tx);
 
@@ -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);