Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6:
  firewire: sbp2: parallelize login, reconnect, logout
  firewire: sbp2: octlet AT payloads can be stack-allocated
  firewire: sbp2: omit Scsi_Host lock from queuecommand
  firewire: core: use non-reentrant workqueue with rescuer
  firewire: optimize iso queueing by setting wake only after the last packet
  firewire: octlet AT payloads can be stack-allocated
  firewire: ohci: optimize find_branch_descriptor()
  firewire: ohci: avoid separate DMA mapping for small AT payloads
  firewire: ohci: do not start DMA contexts before link is enabled

16 files changed, 140 insertions, 117 deletions

diff --git a/drivers/firewire/core-card.c b/drivers/firewire/core-card.c
index 3c44fbc..29d2423 100644
--- a/drivers/firewire/core-card.c
+++ b/drivers/firewire/core-card.c
@@ -228,8 +228,8 @@ void fw_schedule_bus_reset(struct fw_card *card, bool delayed, bool short_reset)
 
 	/* Use an arbitrary short delay to combine multiple reset requests. */
 	fw_card_get(card);
-	if (!schedule_delayed_work(&card->br_work,
-				   delayed ? DIV_ROUND_UP(HZ, 100) : 0))
+	if (!queue_delayed_work(fw_workqueue, &card->br_work,
+				delayed ? DIV_ROUND_UP(HZ, 100) : 0))
 		fw_card_put(card);
 }
 EXPORT_SYMBOL(fw_schedule_bus_reset);
@@ -241,7 +241,7 @@ static void br_work(struct work_struct *work)
 	/* Delay for 2s after last reset per IEEE 1394 clause 8.2.1. */
 	if (card->reset_jiffies != 0 &&
 	    time_before64(get_jiffies_64(), card->reset_jiffies + 2 * HZ)) {
-		if (!schedule_delayed_work(&card->br_work, 2 * HZ))
+		if (!queue_delayed_work(fw_workqueue, &card->br_work, 2 * HZ))
 			fw_card_put(card);
 		return;
 	}
@@ -258,8 +258,7 @@ static void allocate_broadcast_channel(struct fw_card *card, int generation)
 
 	if (!card->broadcast_channel_allocated) {
 		fw_iso_resource_manage(card, generation, 1ULL << 31,
-				       &channel, &bandwidth, true,
-				       card->bm_transaction_data);
+				       &channel, &bandwidth, true);
 		if (channel != 31) {
 			fw_notify("failed to allocate broadcast channel\n");
 			return;
@@ -294,6 +293,7 @@ static void bm_work(struct work_struct *work)
 	bool root_device_is_cmc;
 	bool irm_is_1394_1995_only;
 	bool keep_this_irm;
+	__be32 transaction_data[2];
 
 	spin_lock_irq(&card->lock);
 
@@ -355,21 +355,21 @@ static void bm_work(struct work_struct *work)
 			goto pick_me;
 		}
 
-		card->bm_transaction_data[0] = cpu_to_be32(0x3f);
-		card->bm_transaction_data[1] = cpu_to_be32(local_id);
+		transaction_data[0] = cpu_to_be32(0x3f);
+		transaction_data[1] = cpu_to_be32(local_id);
 
 		spin_unlock_irq(&card->lock);
 
 		rcode = fw_run_transaction(card, TCODE_LOCK_COMPARE_SWAP,
 				irm_id, generation, SCODE_100,
 				CSR_REGISTER_BASE + CSR_BUS_MANAGER_ID,
-				card->bm_transaction_data, 8);
+				transaction_data, 8);
 
 		if (rcode == RCODE_GENERATION)
 			/* Another bus reset, BM work has been rescheduled. */
 			goto out;
 
-		bm_id = be32_to_cpu(card->bm_transaction_data[0]);
+		bm_id = be32_to_cpu(transaction_data[0]);
 
 		spin_lock_irq(&card->lock);
 		if (rcode == RCODE_COMPLETE && generation == card->generation)
@@ -490,11 +490,11 @@ static void bm_work(struct work_struct *work)
 		/*
 		 * Make sure that the cycle master sends cycle start packets.
 		 */
-		card->bm_transaction_data[0] = cpu_to_be32(CSR_STATE_BIT_CMSTR);
+		transaction_data[0] = cpu_to_be32(CSR_STATE_BIT_CMSTR);
 		rcode = fw_run_transaction(card, TCODE_WRITE_QUADLET_REQUEST,
 				root_id, generation, SCODE_100,
 				CSR_REGISTER_BASE + CSR_STATE_SET,
-				card->bm_transaction_data, 4);
+				transaction_data, 4);
 		if (rcode == RCODE_GENERATION)
 			goto out;
 	}
@@ -630,6 +630,10 @@ static int dummy_queue_iso(struct fw_iso_context *ctx, struct fw_iso_packet *p,
 	return -ENODEV;
 }
 
+static void dummy_flush_queue_iso(struct fw_iso_context *ctx)
+{
+}
+
 static const struct fw_card_driver dummy_driver_template = {
 	.read_phy_reg		= dummy_read_phy_reg,
 	.update_phy_reg		= dummy_update_phy_reg,
@@ -641,6 +645,7 @@ static const struct fw_card_driver dummy_driver_template = {
 	.start_iso		= dummy_start_iso,
 	.set_iso_channels	= dummy_set_iso_channels,
 	.queue_iso		= dummy_queue_iso,
+	.flush_queue_iso	= dummy_flush_queue_iso,
 };
 
 void fw_card_release(struct kref *kref)
diff --git a/drivers/firewire/core-cdev.c b/drivers/firewire/core-cdev.c
index 62ac111..b1c1177 100644
--- a/drivers/firewire/core-cdev.c
+++ b/drivers/firewire/core-cdev.c
@@ -141,7 +141,6 @@ struct iso_resource {
 	int generation;
 	u64 channels;
 	s32 bandwidth;
-	__be32 transaction_data[2];
 	struct iso_resource_event *e_alloc, *e_dealloc;
 };
 
@@ -150,7 +149,7 @@ static void release_iso_resource(struct client *, struct client_resource *);
 static void schedule_iso_resource(struct iso_resource *r, unsigned long delay)
 {
 	client_get(r->client);
-	if (!schedule_delayed_work(&r->work, delay))
+	if (!queue_delayed_work(fw_workqueue, &r->work, delay))
 		client_put(r->client);
 }
 
@@ -1108,6 +1107,7 @@ static int ioctl_queue_iso(struct client *client, union ioctl_arg *arg)
 		payload += u.packet.payload_length;
 		count++;
 	}
+	fw_iso_context_queue_flush(ctx);
 
 	a->size    -= uptr_to_u64(p) - a->packets;
 	a->packets  = uptr_to_u64(p);
@@ -1229,8 +1229,7 @@ static void iso_resource_work(struct work_struct *work)
 			r->channels, &channel, &bandwidth,
 			todo == ISO_RES_ALLOC ||
 			todo == ISO_RES_REALLOC ||
-			todo == ISO_RES_ALLOC_ONCE,
-			r->transaction_data);
+			todo == ISO_RES_ALLOC_ONCE);
 	/*
 	 * Is this generation outdated already?  As long as this resource sticks
 	 * in the idr, it will be scheduled again for a newer generation or at
diff --git a/drivers/firewire/core-device.c b/drivers/firewire/core-device.c
index 9a26243..95a4714 100644
--- a/drivers/firewire/core-device.c
+++ b/drivers/firewire/core-device.c
@@ -725,6 +725,15 @@ struct fw_device *fw_device_get_by_devt(dev_t devt)
 	return device;
 }
 
+struct workqueue_struct *fw_workqueue;
+EXPORT_SYMBOL(fw_workqueue);
+
+static void fw_schedule_device_work(struct fw_device *device,
+				    unsigned long delay)
+{
+	queue_delayed_work(fw_workqueue, &device->work, delay);
+}
+
 /*
  * These defines control the retry behavior for reading the config
  * rom.  It shouldn't be necessary to tweak these; if the device
@@ -750,7 +759,7 @@ static void fw_device_shutdown(struct work_struct *work)
 	if (time_before64(get_jiffies_64(),
 			  device->card->reset_jiffies + SHUTDOWN_DELAY)
 	    && !list_empty(&device->card->link)) {
-		schedule_delayed_work(&device->work, SHUTDOWN_DELAY);
+		fw_schedule_device_work(device, SHUTDOWN_DELAY);
 		return;
 	}
 
@@ -862,7 +871,7 @@ static int lookup_existing_device(struct device *dev, void *data)
 		fw_notify("rediscovered device %s\n", dev_name(dev));
 
 		PREPARE_DELAYED_WORK(&old->work, fw_device_update);
-		schedule_delayed_work(&old->work, 0);
+		fw_schedule_device_work(old, 0);
 
 		if (current_node == card->root_node)
 			fw_schedule_bm_work(card, 0);
@@ -953,7 +962,7 @@ static void fw_device_init(struct work_struct *work)
 		if (device->config_rom_retries < MAX_RETRIES &&
 		    atomic_read(&device->state) == FW_DEVICE_INITIALIZING) {
 			device->config_rom_retries++;
-			schedule_delayed_work(&device->work, RETRY_DELAY);
+			fw_schedule_device_work(device, RETRY_DELAY);
 		} else {
 			if (device->node->link_on)
 				fw_notify("giving up on config rom for node id %x\n",
@@ -1019,7 +1028,7 @@ static void fw_device_init(struct work_struct *work)
 			   FW_DEVICE_INITIALIZING,
 			   FW_DEVICE_RUNNING) == FW_DEVICE_GONE) {
 		PREPARE_DELAYED_WORK(&device->work, fw_device_shutdown);
-		schedule_delayed_work(&device->work, SHUTDOWN_DELAY);
+		fw_schedule_device_work(device, SHUTDOWN_DELAY);
 	} else {
 		if (device->config_rom_retries)
 			fw_notify("created device %s: GUID %08x%08x, S%d00, "
@@ -1098,7 +1107,7 @@ static void fw_device_refresh(struct work_struct *work)
 		if (device->config_rom_retries < MAX_RETRIES / 2 &&
 		    atomic_read(&device->state) == FW_DEVICE_INITIALIZING) {
 			device->config_rom_retries++;
-			schedule_delayed_work(&device->work, RETRY_DELAY / 2);
+			fw_schedule_device_work(device, RETRY_DELAY / 2);
 
 			return;
 		}
@@ -1131,7 +1140,7 @@ static void fw_device_refresh(struct work_struct *work)
 		if (device->config_rom_retries < MAX_RETRIES &&
 		    atomic_read(&device->state) == FW_DEVICE_INITIALIZING) {
 			device->config_rom_retries++;
-			schedule_delayed_work(&device->work, RETRY_DELAY);
+			fw_schedule_device_work(device, RETRY_DELAY);
 
 			return;
 		}
@@ -1158,7 +1167,7 @@ static void fw_device_refresh(struct work_struct *work)
  gone:
 	atomic_set(&device->state, FW_DEVICE_GONE);
 	PREPARE_DELAYED_WORK(&device->work, fw_device_shutdown);
-	schedule_delayed_work(&device->work, SHUTDOWN_DELAY);
+	fw_schedule_device_work(device, SHUTDOWN_DELAY);
  out:
 	if (node_id == card->root_node->node_id)
 		fw_schedule_bm_work(card, 0);
@@ -1214,7 +1223,7 @@ void fw_node_event(struct fw_card *card, struct fw_node *node, int event)
 		 * first config rom scan half a second after bus reset.
 		 */
 		INIT_DELAYED_WORK(&device->work, fw_device_init);
-		schedule_delayed_work(&device->work, INITIAL_DELAY);
+		fw_schedule_device_work(device, INITIAL_DELAY);
 		break;
 
 	case FW_NODE_INITIATED_RESET:
@@ -1230,7 +1239,7 @@ void fw_node_event(struct fw_card *card, struct fw_node *node, int event)
 			    FW_DEVICE_RUNNING,
 			    FW_DEVICE_INITIALIZING) == FW_DEVICE_RUNNING) {
 			PREPARE_DELAYED_WORK(&device->work, fw_device_refresh);
-			schedule_delayed_work(&device->work,
+			fw_schedule_device_work(device,
 				device->is_local ? 0 : INITIAL_DELAY);
 		}
 		break;
@@ -1245,7 +1254,7 @@ void fw_node_event(struct fw_card *card, struct fw_node *node, int event)
 		device->generation = card->generation;
 		if (atomic_read(&device->state) == FW_DEVICE_RUNNING) {
 			PREPARE_DELAYED_WORK(&device->work, fw_device_update);
-			schedule_delayed_work(&device->work, 0);
+			fw_schedule_device_work(device, 0);
 		}
 		break;
 
@@ -1270,7 +1279,7 @@ void fw_node_event(struct fw_card *card, struct fw_node *node, int event)
 		if (atomic_xchg(&device->state,
 				FW_DEVICE_GONE) == FW_DEVICE_RUNNING) {
 			PREPARE_DELAYED_WORK(&device->work, fw_device_shutdown);
-			schedule_delayed_work(&device->work,
+			fw_schedule_device_work(device,
 				list_empty(&card->link) ? 0 : SHUTDOWN_DELAY);
 		}
 		break;
diff --git a/drivers/firewire/core-iso.c b/drivers/firewire/core-iso.c
index 481056d..57c3973 100644
--- a/drivers/firewire/core-iso.c
+++ b/drivers/firewire/core-iso.c
@@ -185,6 +185,12 @@ int fw_iso_context_queue(struct fw_iso_context *ctx,
 }
 EXPORT_SYMBOL(fw_iso_context_queue);
 
+void fw_iso_context_queue_flush(struct fw_iso_context *ctx)
+{
+	ctx->card->driver->flush_queue_iso(ctx);
+}
+EXPORT_SYMBOL(fw_iso_context_queue_flush);
+
 int fw_iso_context_stop(struct fw_iso_context *ctx)
 {
 	return ctx->card->driver->stop_iso(ctx);
@@ -196,9 +202,10 @@ EXPORT_SYMBOL(fw_iso_context_stop);
  */
 
 static int manage_bandwidth(struct fw_card *card, int irm_id, int generation,
-			    int bandwidth, bool allocate, __be32 data[2])
+			    int bandwidth, bool allocate)
 {
 	int try, new, old = allocate ? BANDWIDTH_AVAILABLE_INITIAL : 0;
+	__be32 data[2];
 
 	/*
 	 * On a 1394a IRM with low contention, try < 1 is enough.
@@ -233,9 +240,10 @@ static int manage_bandwidth(struct fw_card *card, int irm_id, int generation,
 }
 
 static int manage_channel(struct fw_card *card, int irm_id, int generation,
-		u32 channels_mask, u64 offset, bool allocate, __be32 data[2])
+		u32 channels_mask, u64 offset, bool allocate)
 {
 	__be32 bit, all, old;
+	__be32 data[2];
 	int channel, ret = -EIO, retry = 5;
 
 	old = all = allocate ? cpu_to_be32(~0) : 0;
@@ -284,7 +292,7 @@ static int manage_channel(struct fw_card *card, int irm_id, int generation,
 }
 
 static void deallocate_channel(struct fw_card *card, int irm_id,
-			       int generation, int channel, __be32 buffer[2])
+			       int generation, int channel)
 {
 	u32 mask;
 	u64 offset;
@@ -293,7 +301,7 @@ static void deallocate_channel(struct fw_card *card, int irm_id,
 	offset = channel < 32 ? CSR_REGISTER_BASE + CSR_CHANNELS_AVAILABLE_HI :
 				CSR_REGISTER_BASE + CSR_CHANNELS_AVAILABLE_LO;
 
-	manage_channel(card, irm_id, generation, mask, offset, false, buffer);
+	manage_channel(card, irm_id, generation, mask, offset, false);
 }
 
 /**
@@ -322,7 +330,7 @@ static void deallocate_channel(struct fw_card *card, int irm_id,
  */
 void fw_iso_resource_manage(struct fw_card *card, int generation,
 			    u64 channels_mask, int *channel, int *bandwidth,
-			    bool allocate, __be32 buffer[2])
+			    bool allocate)
 {
 	u32 channels_hi = channels_mask;	/* channels 31...0 */
 	u32 channels_lo = channels_mask >> 32;	/* channels 63...32 */
@@ -335,11 +343,11 @@ void fw_iso_resource_manage(struct fw_card *card, int generation,
 	if (channels_hi)
 		c = manage_channel(card, irm_id, generation, channels_hi,
 				CSR_REGISTER_BASE + CSR_CHANNELS_AVAILABLE_HI,
-				allocate, buffer);
+				allocate);
 	if (channels_lo && c < 0) {
 		c = manage_channel(card, irm_id, generation, channels_lo,
 				CSR_REGISTER_BASE + CSR_CHANNELS_AVAILABLE_LO,
-				allocate, buffer);
+				allocate);
 		if (c >= 0)
 			c += 32;
 	}
@@ -351,14 +359,13 @@ void fw_iso_resource_manage(struct fw_card *card, int generation,
 	if (*bandwidth == 0)
 		return;
 
-	ret = manage_bandwidth(card, irm_id, generation, *bandwidth,
-			       allocate, buffer);
+	ret = manage_bandwidth(card, irm_id, generation, *bandwidth, allocate);
 	if (ret < 0)
 		*bandwidth = 0;
 
 	if (allocate && ret < 0) {
 		if (c >= 0)
-			deallocate_channel(card, irm_id, generation, c, buffer);
+			deallocate_channel(card, irm_id, generation, c);
 		*channel = ret;
 	}
 }
diff --git a/drivers/firewire/core-transaction.c b/drivers/firewire/core-transaction.c
index d00f8ce..334b82a 100644
--- a/drivers/firewire/core-transaction.c
+++ b/drivers/firewire/core-transaction.c
@@ -36,6 +36,7 @@
 #include <linux/string.h>
 #include <linux/timer.h>
 #include <linux/types.h>
+#include <linux/workqueue.h>
 
 #include <asm/byteorder.h>
 
@@ -326,8 +327,8 @@ static int allocate_tlabel(struct fw_card *card)
  * It will contain tag, channel, and sy data instead of a node ID then.
  *
  * The payload buffer at @data is going to be DMA-mapped except in case of
- * quadlet-sized payload or of local (loopback) requests.  Hence make sure that
- * the buffer complies with the restrictions for DMA-mapped memory.  The
+ * @length <= 8 or of local (loopback) requests.  Hence make sure that the
+ * buffer complies with the restrictions of the streaming DMA mapping API.
  * @payload must not be freed before the @callback is called.
  *
  * In case of request types without payload, @data is NULL and @length is 0.
@@ -411,7 +412,8 @@ static void transaction_callback(struct fw_card *card, int rcode,
  *
  * Returns the RCODE.  See fw_send_request() for parameter documentation.
  * Unlike fw_send_request(), @data points to the payload of the request or/and
- * to the payload of the response.
+ * to the payload of the response.  DMA mapping restrictions apply to outbound
+ * request payloads of >= 8 bytes but not to inbound response payloads.
  */
 int fw_run_transaction(struct fw_card *card, int tcode, int destination_id,
 		       int generation, int speed, unsigned long long offset,
@@ -1212,13 +1214,21 @@ static int __init fw_core_init(void)
 {
 	int ret;
 
+	fw_workqueue = alloc_workqueue("firewire",
+				       WQ_NON_REENTRANT | WQ_MEM_RECLAIM, 0);
+	if (!fw_workqueue)
+		return -ENOMEM;
+
 	ret = bus_register(&fw_bus_type);
-	if (ret < 0)
+	if (ret < 0) {
+		destroy_workqueue(fw_workqueue);
 		return ret;
+	}
 
 	fw_cdev_major = register_chrdev(0, "firewire", &fw_device_ops);
 	if (fw_cdev_major < 0) {
 		bus_unregister(&fw_bus_type);
+		destroy_workqueue(fw_workqueue);
 		return fw_cdev_major;
 	}
 
@@ -1234,6 +1244,7 @@ static void __exit fw_core_cleanup(void)
 {
 	unregister_chrdev(fw_cdev_major, "firewire");
 	bus_unregister(&fw_bus_type);
+	destroy_workqueue(fw_workqueue);
 	idr_destroy(&fw_device_idr);
 }
 
diff --git a/drivers/firewire/core.h b/drivers/firewire/core.h
index 25e729c..0fe4e4e 100644
--- a/drivers/firewire/core.h
+++ b/drivers/firewire/core.h
@@ -97,6 +97,8 @@ struct fw_card_driver {
 			 struct fw_iso_buffer *buffer,
 			 unsigned long payload);
 
+	void (*flush_queue_iso)(struct fw_iso_context *ctx);
+
 	int (*stop_iso)(struct fw_iso_context *ctx);
 };
 
diff --git a/drivers/firewire/net.c b/drivers/firewire/net.c
index 3f04dd3..b9762d0 100644
--- a/drivers/firewire/net.c
+++ b/drivers/firewire/net.c
@@ -881,7 +881,9 @@ static void fwnet_receive_broadcast(struct fw_iso_context *context,
 
 	spin_unlock_irqrestore(&dev->lock, flags);
 
-	if (retval < 0)
+	if (retval >= 0)
+		fw_iso_context_queue_flush(dev->broadcast_rcv_context);
+	else
 		fw_error("requeue failed\n");
 }
 
diff --git a/drivers/firewire/ohci.c b/drivers/firewire/ohci.c
index 23d1468..438e6c8 100644
--- a/drivers/firewire/ohci.c
+++ b/drivers/firewire/ohci.c
@@ -1006,13 +1006,12 @@ static void ar_context_run(struct ar_context *ctx)
 
 static struct descriptor *find_branch_descriptor(struct descriptor *d, int z)
 {
-	int b, key;
+	__le16 branch;
 
-	b   = (le16_to_cpu(d->control) & DESCRIPTOR_BRANCH_ALWAYS) >> 2;
-	key = (le16_to_cpu(d->control) & DESCRIPTOR_KEY_IMMEDIATE) >> 8;
+	branch = d->control & cpu_to_le16(DESCRIPTOR_BRANCH_ALWAYS);
 
 	/* figure out which descriptor the branch address goes in */
-	if (z == 2 && (b == 3 || key == 2))
+	if (z == 2 && branch == cpu_to_le16(DESCRIPTOR_BRANCH_ALWAYS))
 		return d;
 	else
 		return d + z - 1;
@@ -1193,9 +1192,6 @@ static void context_append(struct context *ctx,
 	wmb(); /* finish init of new descriptors before branch_address update */
 	ctx->prev->branch_address = cpu_to_le32(d_bus | z);
 	ctx->prev = find_branch_descriptor(d, z);
-
-	reg_write(ctx->ohci, CONTROL_SET(ctx->regs), CONTEXT_WAKE);
-	flush_writes(ctx->ohci);
 }
 
 static void context_stop(struct context *ctx)
@@ -1218,6 +1214,7 @@ static void context_stop(struct context *ctx)
 }
 
 struct driver_data {
+	u8 inline_data[8];
 	struct fw_packet *packet;
 };
 
@@ -1301,20 +1298,28 @@ static int at_context_queue_packet(struct context *ctx,
 		return -1;
 	}
 
+	BUILD_BUG_ON(sizeof(struct driver_data) > sizeof(struct descriptor));
 	driver_data = (struct driver_data *) &d[3];
 	driver_data->packet = packet;
 	packet->driver_data = driver_data;
 
 	if (packet->payload_length > 0) {
-		payload_bus =
-			dma_map_single(ohci->card.device, packet->payload,
-				       packet->payload_length, DMA_TO_DEVICE);
-		if (dma_mapping_error(ohci->card.device, payload_bus)) {
-			packet->ack = RCODE_SEND_ERROR;
-			return -1;
+		if (packet->payload_length > sizeof(driver_data->inline_data)) {
+			payload_bus = dma_map_single(ohci->card.device,
+						     packet->payload,
+						     packet->payload_length,
+						     DMA_TO_DEVICE);
+			if (dma_mapping_error(ohci->card.device, payload_bus)) {
+				packet->ack = RCODE_SEND_ERROR;
+				return -1;
+			}
+			packet->payload_bus	= payload_bus;
+			packet->payload_mapped	= true;
+		} else {
+			memcpy(driver_data->inline_data, packet->payload,
+			       packet->payload_length);
+			payload_bus = d_bus + 3 * sizeof(*d);
 		}
-		packet->payload_bus	= payload_bus;
-		packet->payload_mapped	= true;
 
 		d[2].req_count    = cpu_to_le16(packet->payload_length);
 		d[2].data_address = cpu_to_le32(payload_bus);
@@ -1340,8 +1345,12 @@ static int at_context_queue_packet(struct context *ctx,
 
 	context_append(ctx, d, z, 4 - z);
 
-	if (!ctx->running)
+	if (ctx->running) {
+		reg_write(ohci, CONTROL_SET(ctx->regs), CONTEXT_WAKE);
+		flush_writes(ohci);
+	} else {
 		context_run(ctx, 0);
+	}
 
 	return 0;
 }
@@ -2066,8 +2075,6 @@ static int ohci_enable(struct fw_card *card,
 
 	reg_write(ohci, OHCI1394_SelfIDBuffer, ohci->self_id_bus);
 	reg_write(ohci, OHCI1394_LinkControlSet,
-		  OHCI1394_LinkControl_rcvSelfID |
-		  OHCI1394_LinkControl_rcvPhyPkt |
 		  OHCI1394_LinkControl_cycleTimerEnable |
 		  OHCI1394_LinkControl_cycleMaster);
 
@@ -2094,9 +2101,6 @@ static int ohci_enable(struct fw_card *card,
 	reg_write(ohci, OHCI1394_FairnessControl, 0);
 	card->priority_budget_implemented = ohci->pri_req_max != 0;
 
-	ar_context_run(&ohci->ar_request_ctx);
-	ar_context_run(&ohci->ar_response_ctx);
-
 	reg_write(ohci, OHCI1394_PhyUpperBound, 0x00010000);
 	reg_write(ohci, OHCI1394_IntEventClear, ~0);
 	reg_write(ohci, OHCI1394_IntMaskClear, ~0);
@@ -2186,7 +2190,13 @@ static int ohci_enable(struct fw_card *card,
 	reg_write(ohci, OHCI1394_HCControlSet,
 		  OHCI1394_HCControl_linkEnable |
 		  OHCI1394_HCControl_BIBimageValid);
-	flush_writes(ohci);
+
+	reg_write(ohci, OHCI1394_LinkControlSet,
+		  OHCI1394_LinkControl_rcvSelfID |
+		  OHCI1394_LinkControl_rcvPhyPkt);
+
+	ar_context_run(&ohci->ar_request_ctx);
+	ar_context_run(&ohci->ar_response_ctx); /* also flushes writes */
 
 	/* We are ready to go, reset bus to finish initialization. */
 	fw_schedule_bus_reset(&ohci->card, false, true);
@@ -3112,6 +3122,15 @@ static int ohci_queue_iso(struct fw_iso_context *base,
 	return ret;
 }
 
+static void ohci_flush_queue_iso(struct fw_iso_context *base)
+{
+	struct context *ctx =
+			&container_of(base, struct iso_context, base)->context;
+
+	reg_write(ctx->ohci, CONTROL_SET(ctx->regs), CONTEXT_WAKE);
+	flush_writes(ctx->ohci);
+}
+
 static const struct fw_card_driver ohci_driver = {
 	.enable			= ohci_enable,
 	.read_phy_reg		= ohci_read_phy_reg,
@@ -3128,6 +3147,7 @@ static const struct fw_card_driver ohci_driver = {
 	.free_iso_context	= ohci_free_iso_context,
 	.set_iso_channels	= ohci_set_iso_channels,
 	.queue_iso		= ohci_queue_iso,
+	.flush_queue_iso	= ohci_flush_queue_iso,
 	.start_iso		= ohci_start_iso,
 	.stop_iso		= ohci_stop_iso,
 };
diff --git a/drivers/firewire/sbp2.c b/drivers/firewire/sbp2.c
index 77ed589..41841a3 100644
--- a/drivers/firewire/sbp2.c
+++ b/drivers/firewire/sbp2.c
@@ -125,9 +125,6 @@ MODULE_PARM_DESC(workarounds, "Work around device bugs (default = 0"
 	", override internal blacklist = " __stringify(SBP2_WORKAROUND_OVERRIDE)
 	", or a combination)");
 
-/* I don't know why the SCSI stack doesn't define something like this... */
-typedef void (*scsi_done_fn_t)(struct scsi_cmnd *);
-
 static const char sbp2_driver_name[] = "sbp2";
 
 /*
@@ -261,7 +258,6 @@ struct sbp2_orb {
 	struct kref kref;
 	dma_addr_t request_bus;
 	int rcode;
-	struct sbp2_pointer pointer;
 	void (*callback)(struct sbp2_orb * orb, struct sbp2_status * status);
 	struct list_head link;
 };
@@ -314,7 +310,6 @@ struct sbp2_command_orb {
 		u8 command_block[SBP2_MAX_CDB_SIZE];
 	} request;
 	struct scsi_cmnd *cmd;
-	scsi_done_fn_t done;
 	struct sbp2_logical_unit *lu;
 
 	struct sbp2_pointer page_table[SG_ALL] __attribute__((aligned(8)));
@@ -494,10 +489,11 @@ static void sbp2_send_orb(struct sbp2_orb *orb, struct sbp2_logical_unit *lu,
 			  int node_id, int generation, u64 offset)
 {
 	struct fw_device *device = target_device(lu->tgt);
+	struct sbp2_pointer orb_pointer;
 	unsigned long flags;
 
-	orb->pointer.high = 0;
-	orb->pointer.low = cpu_to_be32(orb->request_bus);
+	orb_pointer.high = 0;
+	orb_pointer.low = cpu_to_be32(orb->request_bus);
 
 	spin_lock_irqsave(&device->card->lock, flags);
 	list_add_tail(&orb->link, &lu->orb_list);
@@ -508,7 +504,7 @@ static void sbp2_send_orb(struct sbp2_orb *orb, struct sbp2_logical_unit *lu,
 
 	fw_send_request(device->card, &orb->t, TCODE_WRITE_BLOCK_REQUEST,
 			node_id, generation, device->max_speed, offset,
-			&orb->pointer, 8, complete_transaction, orb);
+			&orb_pointer, 8, complete_transaction, orb);
 }
 
 static int sbp2_cancel_orbs(struct sbp2_logical_unit *lu)
@@ -830,8 +826,6 @@ static void sbp2_target_put(struct sbp2_target *tgt)
 	kref_put(&tgt->kref, sbp2_release_target);
 }
 
-static struct workqueue_struct *sbp2_wq;
-
 /*
  * Always get the target's kref when scheduling work on one its units.
  * Each workqueue job is responsible to call sbp2_target_put() upon return.
@@ -839,7 +833,7 @@ static struct workqueue_struct *sbp2_wq;
 static void sbp2_queue_work(struct sbp2_logical_unit *lu, unsigned long delay)
 {
 	sbp2_target_get(lu->tgt);
-	if (!queue_delayed_work(sbp2_wq, &lu->work, delay))
+	if (!queue_delayed_work(fw_workqueue, &lu->work, delay))
 		sbp2_target_put(lu->tgt);
 }
 
@@ -1398,7 +1392,7 @@ static void complete_command_orb(struct sbp2_orb *base_orb,
 	sbp2_unmap_scatterlist(device->card->device, orb);
 
 	orb->cmd->result = result;
-	orb->done(orb->cmd);
+	orb->cmd->scsi_done(orb->cmd);
 }
 
 static int sbp2_map_scatterlist(struct sbp2_command_orb *orb,
@@ -1463,7 +1457,8 @@ static int sbp2_map_scatterlist(struct sbp2_command_orb *orb,
 
 /* SCSI stack integration */
 
-static int sbp2_scsi_queuecommand_lck(struct scsi_cmnd *cmd, scsi_done_fn_t done)
+static int sbp2_scsi_queuecommand(struct Scsi_Host *shost,
+				  struct scsi_cmnd *cmd)
 {
 	struct sbp2_logical_unit *lu = cmd->device->hostdata;
 	struct fw_device *device = target_device(lu->tgt);
@@ -1477,7 +1472,7 @@ static int sbp2_scsi_queuecommand_lck(struct scsi_cmnd *cmd, scsi_done_fn_t done
 	if (cmd->sc_data_direction == DMA_BIDIRECTIONAL) {
 		fw_error("Can't handle DMA_BIDIRECTIONAL, rejecting command\n");
 		cmd->result = DID_ERROR << 16;
-		done(cmd);
+		cmd->scsi_done(cmd);
 		return 0;
 	}
 
@@ -1490,11 +1485,8 @@ static int sbp2_scsi_queuecommand_lck(struct scsi_cmnd *cmd, scsi_done_fn_t done
 	/* Initialize rcode to something not RCODE_COMPLETE. */
 	orb->base.rcode = -1;
 	kref_init(&orb->base.kref);
-
-	orb->lu   = lu;
-	orb->done = done;
-	orb->cmd  = cmd;
-
+	orb->lu = lu;
+	orb->cmd = cmd;
 	orb->request.next.high = cpu_to_be32(SBP2_ORB_NULL);
 	orb->request.misc = cpu_to_be32(
 		COMMAND_ORB_MAX_PAYLOAD(lu->tgt->max_payload) |
@@ -1529,8 +1521,6 @@ static int sbp2_scsi_queuecommand_lck(struct scsi_cmnd *cmd, scsi_done_fn_t done
 	return retval;
 }
 
-static DEF_SCSI_QCMD(sbp2_scsi_queuecommand)
-
 static int sbp2_scsi_slave_alloc(struct scsi_device *sdev)
 {
 	struct sbp2_logical_unit *lu = sdev->hostdata;
@@ -1653,17 +1643,12 @@ MODULE_ALIAS("sbp2");
 
 static int __init sbp2_init(void)
 {
-	sbp2_wq = create_singlethread_workqueue(KBUILD_MODNAME);
-	if (!sbp2_wq)
-		return -ENOMEM;
-
 	return driver_register(&sbp2_driver.driver);
 }
 
 static void __exit sbp2_cleanup(void)
 {
 	driver_unregister(&sbp2_driver.driver);
-	destroy_workqueue(sbp2_wq);
 }
 
 module_init(sbp2_init);
diff --git a/drivers/media/dvb/firewire/firedtv-avc.c b/drivers/media/dvb/firewire/firedtv-avc.c
index fc5ccd8..21c52e3 100644
--- a/drivers/media/dvb/firewire/firedtv-avc.c
+++ b/drivers/media/dvb/firewire/firedtv-avc.c
@@ -1320,14 +1320,10 @@ static int cmp_read(struct firedtv *fdtv, u64 addr, __be32 *data)
 {
 	int ret;
 
-	mutex_lock(&fdtv->avc_mutex);
-
 	ret = fdtv_read(fdtv, addr, data);
 	if (ret < 0)
 		dev_err(fdtv->device, "CMP: read I/O error\n");
 
-	mutex_unlock(&fdtv->avc_mutex);
-
 	return ret;
 }
 
@@ -1335,18 +1331,9 @@ static int cmp_lock(struct firedtv *fdtv, u64 addr, __be32 data[])
 {
 	int ret;
 
-	mutex_lock(&fdtv->avc_mutex);
-
-	/* data[] is stack-allocated and should not be DMA-mapped. */
-	memcpy(fdtv->avc_data, data, 8);
-
-	ret = fdtv_lock(fdtv, addr, fdtv->avc_data);
+	ret = fdtv_lock(fdtv, addr, data);
 	if (ret < 0)
 		dev_err(fdtv->device, "CMP: lock I/O error\n");
-	else
-		memcpy(data, fdtv->avc_data, 8);
-
-	mutex_unlock(&fdtv->avc_mutex);
 
 	return ret;
 }
diff --git a/drivers/media/dvb/firewire/firedtv-fw.c b/drivers/media/dvb/firewire/firedtv-fw.c
index 8022b74..864b627 100644
--- a/drivers/media/dvb/firewire/firedtv-fw.c
+++ b/drivers/media/dvb/firewire/firedtv-fw.c
@@ -125,6 +125,7 @@ static void handle_iso(struct fw_iso_context *context, u32 cycle,
 
 		i = (i + 1) & (N_PACKETS - 1);
 	}
+	fw_iso_context_queue_flush(ctx->context);
 	ctx->current_packet = i;
 }
 
diff --git a/include/linux/firewire.h b/include/linux/firewire.h
index c64f368..5e6f427 100644
--- a/include/linux/firewire.h
+++ b/include/linux/firewire.h
@@ -125,7 +125,6 @@ struct fw_card {
 	struct delayed_work bm_work; /* bus manager job */
 	int bm_retries;
 	int bm_generation;
-	__be32 bm_transaction_data[2];
 	int bm_node_id;
 	bool bm_abdicate;
 
@@ -441,12 +440,15 @@ int fw_iso_context_queue(struct fw_iso_context *ctx,
 			 struct fw_iso_packet *packet,
 			 struct fw_iso_buffer *buffer,
 			 unsigned long payload);
+void fw_iso_context_queue_flush(struct fw_iso_context *ctx);
 int fw_iso_context_start(struct fw_iso_context *ctx,
 			 int cycle, int sync, int tags);
 int fw_iso_context_stop(struct fw_iso_context *ctx);
 void fw_iso_context_destroy(struct fw_iso_context *ctx);
 void fw_iso_resource_manage(struct fw_card *card, int generation,
 			    u64 channels_mask, int *channel, int *bandwidth,
-			    bool allocate, __be32 buffer[2]);
+			    bool allocate);
+
+extern struct workqueue_struct *fw_workqueue;
 
 #endif /* _LINUX_FIREWIRE_H */
diff --git a/sound/firewire/amdtp.c b/sound/firewire/amdtp.c
index b18140f..87657dd 100644
--- a/sound/firewire/amdtp.c
+++ b/sound/firewire/amdtp.c
@@ -396,6 +396,7 @@ static void out_packet_callback(struct fw_iso_context *context, u32 cycle,
 
 	for (i = 0; i < packets; ++i)
 		queue_out_packet(s, ++cycle);
+	fw_iso_context_queue_flush(s->context);
 }
 
 static int queue_initial_skip_packets(struct amdtp_out_stream *s)
diff --git a/sound/firewire/cmp.c b/sound/firewire/cmp.c
index 4a37f3a..14cacbc 100644
--- a/sound/firewire/cmp.c
+++ b/sound/firewire/cmp.c
@@ -49,10 +49,9 @@ static int pcr_modify(struct cmp_connection *c,
 		      enum bus_reset_handling bus_reset_handling)
 {
 	struct fw_device *device = fw_parent_device(c->resources.unit);
-	__be32 *buffer = c->resources.buffer;
 	int generation = c->resources.generation;
 	int rcode, errors = 0;
-	__be32 old_arg;
+	__be32 old_arg, buffer[2];
 	int err;
 
 	buffer[0] = c->last_pcr_value;
diff --git a/sound/firewire/iso-resources.c b/sound/firewire/iso-resources.c
index 775dbd5..bb9c0c1 100644
--- a/sound/firewire/iso-resources.c
+++ b/sound/firewire/iso-resources.c
@@ -11,7 +11,6 @@
 #include <linux/jiffies.h>
 #include <linux/mutex.h>
 #include <linux/sched.h>
-#include <linux/slab.h>
 #include <linux/spinlock.h>
 #include "iso-resources.h"
 
@@ -25,10 +24,6 @@
  */
 int fw_iso_resources_init(struct fw_iso_resources *r, struct fw_unit *unit)
 {
-	r->buffer = kmalloc(2 * 4, GFP_KERNEL);
-	if (!r->buffer)
-		return -ENOMEM;
-
 	r->channels_mask = ~0uLL;
 	r->unit = fw_unit_get(unit);
 	mutex_init(&r->mutex);
@@ -44,7 +39,6 @@ int fw_iso_resources_init(struct fw_iso_resources *r, struct fw_unit *unit)
 void fw_iso_resources_destroy(struct fw_iso_resources *r)
 {
 	WARN_ON(r->allocated);
-	kfree(r->buffer);
 	mutex_destroy(&r->mutex);
 	fw_unit_put(r->unit);
 }
@@ -131,7 +125,7 @@ retry_after_bus_reset:
 
 	bandwidth = r->bandwidth + r->bandwidth_overhead;
 	fw_iso_resource_manage(card, r->generation, r->channels_mask,
-			       &channel, &bandwidth, true, r->buffer);
+			       &channel, &bandwidth, true);
 	if (channel == -EAGAIN) {
 		mutex_unlock(&r->mutex);
 		goto retry_after_bus_reset;
@@ -184,7 +178,7 @@ int fw_iso_resources_update(struct fw_iso_resources *r)
 	bandwidth = r->bandwidth + r->bandwidth_overhead;
 
 	fw_iso_resource_manage(card, r->generation, 1uLL << r->channel,
-			       &channel, &bandwidth, true, r->buffer);
+			       &channel, &bandwidth, true);
 	/*
 	 * When another bus reset happens, pretend that the allocation
 	 * succeeded; we will try again for the new generation later.
@@ -220,7 +214,7 @@ void fw_iso_resources_free(struct fw_iso_resources *r)
 	if (r->allocated) {
 		bandwidth = r->bandwidth + r->bandwidth_overhead;
 		fw_iso_resource_manage(card, r->generation, 1uLL << r->channel,
-				       &channel, &bandwidth, false, r->buffer);
+				       &channel, &bandwidth, false);
 		if (channel < 0)
 			dev_err(&r->unit->device,
 				"isochronous resource deallocation failed\n");
diff --git a/sound/firewire/iso-resources.h b/sound/firewire/iso-resources.h
index 3f0730e4..5a9af7c 100644
--- a/sound/firewire/iso-resources.h
+++ b/sound/firewire/iso-resources.h
@@ -24,7 +24,6 @@ struct fw_iso_resources {
 	unsigned int bandwidth_overhead;
 	int generation; /* in which allocation is valid */
 	bool allocated;
-	__be32 *buffer;
 };
 
 int fw_iso_resources_init(struct fw_iso_resources *r,