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: (26 commits)
  firewire: fw-sbp2: Use sbp2 device-provided mgt orb timeout for logins
  firewire: fw-sbp2: increase login orb reply timeout, fix "failed to login"
  firewire: replace subtraction with bitwise and
  firewire: fw-core: react on bus resets while the config ROM is being fetched
  firewire: enforce access order between generation and node ID, fix "giving up on config rom"
  firewire: fw-cdev: use device generation, not card generation
  firewire: fw-sbp2: use device generation, not card generation
  firewire: fw-sbp2: try to increase reconnect_hold (speed up reconnection)
  firewire: fw-sbp2: skip unnecessary logout
  firewire vs. ieee1394: clarify MAINTAINERS
  firewire: fw-ohci: Dynamically allocate buffers for DMA descriptors
  firewire: fw-ohci: CycleTooLong interrupt management
  firewire: Fix extraction of source node id
  firewire: fw-ohci: Bug fixes for packet-per-buffer support
  firewire: fw-ohci: Fix for dualbuffer three-or-more buffers
  firewire: fw-sbp2: remove unused misleading macro
  firewire: fw-sbp2: prepare for s/g chaining
  firewire: fw-sbp2: refactor workq and kref handling
  ieee1394: ohci1394: don't schedule IT tasklets on IR events
  ieee1394: sbp2: raise default transfer size limit
  ...

MAINTAINERS

@@ -1595,7 +1595,7 @@ P:	Alexander Viro
 M:	viro@zeniv.linux.org.uk
 S:	Maintained
 
-FIREWIRE SUBSYSTEM
+FIREWIRE SUBSYSTEM (drivers/firewire, <linux/firewire*.h>)
 P:	Kristian Hoegsberg, Stefan Richter
 M:	krh@redhat.com, stefanr@s5r6.in-berlin.de
 L:	linux1394-devel@lists.sourceforge.net
@@ -1917,7 +1917,7 @@ L:	linux-ide@vger.kernel.org
 L:	linux-scsi@vger.kernel.org
 S:	Orphan
 
-IEEE 1394 SUBSYSTEM
+IEEE 1394 SUBSYSTEM (drivers/ieee1394)
 P:	Ben Collins
 M:	ben.collins@ubuntu.com
 P:	Stefan Richter

drivers/firewire/fw-cdev.c

@@ -206,12 +206,13 @@ fill_bus_reset_event(struct fw_cdev_event_bus_reset *event,
 
 	event->closure	     = client->bus_reset_closure;
 	event->type          = FW_CDEV_EVENT_BUS_RESET;
+	event->generation    = client->device->generation;
+	smp_rmb();           /* node_id must not be older than generation */
 	event->node_id       = client->device->node_id;
 	event->local_node_id = card->local_node->node_id;
 	event->bm_node_id    = 0; /* FIXME: We don't track the BM. */
 	event->irm_node_id   = card->irm_node->node_id;
 	event->root_node_id  = card->root_node->node_id;
-	event->generation    = card->generation;
 }
 
 static void

drivers/firewire/fw-device.c

@@ -27,6 +27,7 @@
 #include <linux/idr.h>
 #include <linux/rwsem.h>
 #include <asm/semaphore.h>
+#include <asm/system.h>
 #include <linux/ctype.h>
 #include "fw-transaction.h"
 #include "fw-topology.h"
@@ -182,9 +183,14 @@ static void fw_device_release(struct device *dev)
 
 int fw_device_enable_phys_dma(struct fw_device *device)
 {
+	int generation = device->generation;
+
+	/* device->node_id, accessed below, must not be older than generation */
+	smp_rmb();
+
 	return device->card->driver->enable_phys_dma(device->card,
 						     device->node_id,
-						     device->generation);
+						     generation);
 }
 EXPORT_SYMBOL(fw_device_enable_phys_dma);
 
@@ -384,17 +390,21 @@ complete_transaction(struct fw_card *card, int rcode,
 	complete(&callback_data->done);
 }
 
-static int read_rom(struct fw_device *device, int index, u32 * data)
+static int
+read_rom(struct fw_device *device, int generation, int index, u32 *data)
 {
 	struct read_quadlet_callback_data callback_data;
 	struct fw_transaction t;
 	u64 offset;
 
+	/* device->node_id, accessed below, must not be older than generation */
+	smp_rmb();
+
 	init_completion(&callback_data.done);
 
 	offset = 0xfffff0000400ULL + index * 4;
 	fw_send_request(device->card, &t, TCODE_READ_QUADLET_REQUEST,
-			device->node_id, device->generation, device->max_speed,
+			device->node_id, generation, device->max_speed,
 			offset, NULL, 4, complete_transaction, &callback_data);
 
 	wait_for_completion(&callback_data.done);
@@ -404,7 +414,14 @@ static int read_rom(struct fw_device *device, int index, u32 * data)
 	return callback_data.rcode;
 }
 
-static int read_bus_info_block(struct fw_device *device)
+/*
+ * Read the bus info block, perform a speed probe, and read all of the rest of
+ * the config ROM.  We do all this with a cached bus generation.  If the bus
+ * generation changes under us, read_bus_info_block will fail and get retried.
+ * It's better to start all over in this case because the node from which we
+ * are reading the ROM may have changed the ROM during the reset.
+ */
+static int read_bus_info_block(struct fw_device *device, int generation)
 {
 	static u32 rom[256];
 	u32 stack[16], sp, key;
@@ -414,7 +431,7 @@ static int read_bus_info_block(struct fw_device *device)
 
 	/* First read the bus info block. */
 	for (i = 0; i < 5; i++) {
-		if (read_rom(device, i, &rom[i]) != RCODE_COMPLETE)
+		if (read_rom(device, generation, i, &rom[i]) != RCODE_COMPLETE)
 			return -1;
 		/*
 		 * As per IEEE1212 7.2, during power-up, devices can
@@ -449,7 +466,8 @@ static int read_bus_info_block(struct fw_device *device)
 			device->max_speed = device->card->link_speed;
 
 		while (device->max_speed > SCODE_100) {
-			if (read_rom(device, 0, &dummy) == RCODE_COMPLETE)
+			if (read_rom(device, generation, 0, &dummy) ==
+			    RCODE_COMPLETE)
 				break;
 			device->max_speed--;
 		}
@@ -482,7 +500,7 @@ static int read_bus_info_block(struct fw_device *device)
 			return -1;
 
 		/* Read header quadlet for the block to get the length. */
-		if (read_rom(device, i, &rom[i]) != RCODE_COMPLETE)
+		if (read_rom(device, generation, i, &rom[i]) != RCODE_COMPLETE)
 			return -1;
 		end = i + (rom[i] >> 16) + 1;
 		i++;
@@ -501,7 +519,8 @@ static int read_bus_info_block(struct fw_device *device)
 		 * it references another block, and push it in that case.
 		 */
 		while (i < end) {
-			if (read_rom(device, i, &rom[i]) != RCODE_COMPLETE)
+			if (read_rom(device, generation, i, &rom[i]) !=
+			    RCODE_COMPLETE)
 				return -1;
 			if ((key >> 30) == 3 && (rom[i] >> 30) > 1 &&
 			    sp < ARRAY_SIZE(stack))
@@ -648,7 +667,7 @@ static void fw_device_init(struct work_struct *work)
 	 * device.
 	 */
 
-	if (read_bus_info_block(device) < 0) {
+	if (read_bus_info_block(device, device->generation) < 0) {
 		if (device->config_rom_retries < MAX_RETRIES) {
 			device->config_rom_retries++;
 			schedule_delayed_work(&device->work, RETRY_DELAY);
@@ -801,6 +820,7 @@ void fw_node_event(struct fw_card *card, struct fw_node *node, int event)
 
 		device = node->data;
 		device->node_id = node->node_id;
+		smp_wmb();  /* update node_id before generation */
 		device->generation = card->generation;
 		if (atomic_read(&device->state) == FW_DEVICE_RUNNING) {
 			PREPARE_DELAYED_WORK(&device->work, fw_device_update);

drivers/firewire/fw-device.h

@@ -35,6 +35,18 @@ struct fw_attribute_group {
 	struct attribute *attrs[11];
 };
 
+/*
+ * Note, fw_device.generation always has to be read before fw_device.node_id.
+ * Use SMP memory barriers to ensure this.  Otherwise requests will be sent
+ * to an outdated node_id if the generation was updated in the meantime due
+ * to a bus reset.
+ *
+ * Likewise, fw-core will take care to update .node_id before .generation so
+ * that whenever fw_device.generation is current WRT the actual bus generation,
+ * fw_device.node_id is guaranteed to be current too.
+ *
+ * The same applies to fw_device.card->node_id vs. fw_device.generation.
+ */
 struct fw_device {
 	atomic_t state;
 	struct fw_node *node;

drivers/firewire/fw-ohci.c

@@ -98,17 +98,48 @@ struct context;
 typedef int (*descriptor_callback_t)(struct context *ctx,
 				     struct descriptor *d,
 				     struct descriptor *last);
+
+/*
+ * A buffer that contains a block of DMA-able coherent memory used for
+ * storing a portion of a DMA descriptor program.
+ */
+struct descriptor_buffer {
+	struct list_head list;
+	dma_addr_t buffer_bus;
+	size_t buffer_size;
+	size_t used;
+	struct descriptor buffer[0];
+};
+
 struct context {
 	struct fw_ohci *ohci;
 	u32 regs;
+	int total_allocation;
 
-	struct descriptor *buffer;
+	/*
-	dma_addr_t buffer_bus;
+	 * List of page-sized buffers for storing DMA descriptors.
-	size_t buffer_size;
+	 * Head of list contains buffers in use and tail of list contains
-	struct descriptor *head_descriptor;
+	 * free buffers.
-	struct descriptor *tail_descriptor;
+	 */
-	struct descriptor *tail_descriptor_last;
+	struct list_head buffer_list;
-	struct descriptor *prev_descriptor;
+
+	/*
+	 * Pointer to a buffer inside buffer_list that contains the tail
+	 * end of the current DMA program.
+	 */
+	struct descriptor_buffer *buffer_tail;
+
+	/*
+	 * The descriptor containing the branch address of the first
+	 * descriptor that has not yet been filled by the device.
+	 */
+	struct descriptor *last;
+
+	/*
+	 * The last descriptor in the DMA program.  It contains the branch
+	 * address that must be updated upon appending a new descriptor.
+	 */
+	struct descriptor *prev;
 
 	descriptor_callback_t callback;
 
@@ -125,6 +156,7 @@ struct context {
 struct iso_context {
 	struct fw_iso_context base;
 	struct context context;
+	int excess_bytes;
 	void *header;
 	size_t header_length;
 };
@@ -197,8 +229,6 @@ static inline struct fw_ohci *fw_ohci(struct fw_card *card)
 #define SELF_ID_BUF_SIZE		0x800
 #define OHCI_TCODE_PHY_PACKET		0x0e
 #define OHCI_VERSION_1_1		0x010010
-#define ISO_BUFFER_SIZE			(64 * 1024)
-#define AT_BUFFER_SIZE			4096
 
 static char ohci_driver_name[] = KBUILD_MODNAME;
 
@@ -455,71 +485,108 @@ find_branch_descriptor(struct descriptor *d, int z)
 static void context_tasklet(unsigned long data)
 {
 	struct context *ctx = (struct context *) data;
-	struct fw_ohci *ohci = ctx->ohci;
 	struct descriptor *d, *last;
 	u32 address;
 	int z;
+	struct descriptor_buffer *desc;
 
-	dma_sync_single_for_cpu(ohci->card.device, ctx->buffer_bus,
+	desc = list_entry(ctx->buffer_list.next,
-				ctx->buffer_size, DMA_TO_DEVICE);
+			struct descriptor_buffer, list);
-
+	last = ctx->last;
-	d    = ctx->tail_descriptor;
-	last = ctx->tail_descriptor_last;
-
 	while (last->branch_address != 0) {
+		struct descriptor_buffer *old_desc = desc;
 		address = le32_to_cpu(last->branch_address);
 		z = address & 0xf;
-		d = ctx->buffer + (address - ctx->buffer_bus) / sizeof(*d);
+		address &= ~0xf;
+
+		/* If the branch address points to a buffer outside of the
+		 * current buffer, advance to the next buffer. */
+		if (address < desc->buffer_bus ||
+				address >= desc->buffer_bus + desc->used)
+			desc = list_entry(desc->list.next,
+					struct descriptor_buffer, list);
+		d = desc->buffer + (address - desc->buffer_bus) / sizeof(*d);
 		last = find_branch_descriptor(d, z);
 
 		if (!ctx->callback(ctx, d, last))
 			break;
 
-		ctx->tail_descriptor      = d;
+		if (old_desc != desc) {
-		ctx->tail_descriptor_last = last;
+			/* If we've advanced to the next buffer, move the
+			 * previous buffer to the free list. */
+			unsigned long flags;
+			old_desc->used = 0;
+			spin_lock_irqsave(&ctx->ohci->lock, flags);
+			list_move_tail(&old_desc->list, &ctx->buffer_list);
+			spin_unlock_irqrestore(&ctx->ohci->lock, flags);
+		}
+		ctx->last = last;
 	}
 }
 
+/*
+ * Allocate a new buffer and add it to the list of free buffers for this
+ * context.  Must be called with ohci->lock held.
+ */
+static int
+context_add_buffer(struct context *ctx)
+{
+	struct descriptor_buffer *desc;
+	dma_addr_t bus_addr;
+	int offset;
+
+	/*
+	 * 16MB of descriptors should be far more than enough for any DMA
+	 * program.  This will catch run-away userspace or DoS attacks.
+	 */
+	if (ctx->total_allocation >= 16*1024*1024)
+		return -ENOMEM;
+
+	desc = dma_alloc_coherent(ctx->ohci->card.device, PAGE_SIZE,
+			&bus_addr, GFP_ATOMIC);
+	if (!desc)
+		return -ENOMEM;
+
+	offset = (void *)&desc->buffer - (void *)desc;
+	desc->buffer_size = PAGE_SIZE - offset;
+	desc->buffer_bus = bus_addr + offset;
+	desc->used = 0;
+
+	list_add_tail(&desc->list, &ctx->buffer_list);
+	ctx->total_allocation += PAGE_SIZE;
+
+	return 0;
+}
+
 static int
 context_init(struct context *ctx, struct fw_ohci *ohci,
-	     size_t buffer_size, u32 regs,
+	     u32 regs, descriptor_callback_t callback)
-	     descriptor_callback_t callback)
 {
 	ctx->ohci = ohci;
 	ctx->regs = regs;
-	ctx->buffer_size = buffer_size;
+	ctx->total_allocation = 0;
-	ctx->buffer = kmalloc(buffer_size, GFP_KERNEL);
+
-	if (ctx->buffer == NULL)
+	INIT_LIST_HEAD(&ctx->buffer_list);
+	if (context_add_buffer(ctx) < 0)
 		return -ENOMEM;
 
+	ctx->buffer_tail = list_entry(ctx->buffer_list.next,
+			struct descriptor_buffer, list);
+
 	tasklet_init(&ctx->tasklet, context_tasklet, (unsigned long)ctx);
 	ctx->callback = callback;
 
-	ctx->buffer_bus =
-		dma_map_single(ohci->card.device, ctx->buffer,
-			       buffer_size, DMA_TO_DEVICE);
-	if (dma_mapping_error(ctx->buffer_bus)) {
-		kfree(ctx->buffer);
-		return -ENOMEM;
-	}
-
-	ctx->head_descriptor      = ctx->buffer;
-	ctx->prev_descriptor      = ctx->buffer;
-	ctx->tail_descriptor      = ctx->buffer;
-	ctx->tail_descriptor_last = ctx->buffer;
-
 	/*
 	 * We put a dummy descriptor in the buffer that has a NULL
 	 * branch address and looks like it's been sent.  That way we
-	 * have a descriptor to append DMA programs to.  Also, the
+	 * have a descriptor to append DMA programs to.
-	 * ring buffer invariant is that it always has at least one
-	 * element so that head == tail means buffer full.
 	 */
-
+	memset(ctx->buffer_tail->buffer, 0, sizeof(*ctx->buffer_tail->buffer));
-	memset(ctx->head_descriptor, 0, sizeof(*ctx->head_descriptor));
+	ctx->buffer_tail->buffer->control = cpu_to_le16(DESCRIPTOR_OUTPUT_LAST);
-	ctx->head_descriptor->control = cpu_to_le16(DESCRIPTOR_OUTPUT_LAST);
+	ctx->buffer_tail->buffer->transfer_status = cpu_to_le16(0x8011);
-	ctx->head_descriptor->transfer_status = cpu_to_le16(0x8011);
+	ctx->buffer_tail->used += sizeof(*ctx->buffer_tail->buffer);
-	ctx->head_descriptor++;
+	ctx->last = ctx->buffer_tail->buffer;
+	ctx->prev = ctx->buffer_tail->buffer;
 
 	return 0;
 }
@@ -528,35 +595,42 @@ static void
 context_release(struct context *ctx)
 {
 	struct fw_card *card = &ctx->ohci->card;
+	struct descriptor_buffer *desc, *tmp;
 
-	dma_unmap_single(card->device, ctx->buffer_bus,
+	list_for_each_entry_safe(desc, tmp, &ctx->buffer_list, list)
-			 ctx->buffer_size, DMA_TO_DEVICE);
+		dma_free_coherent(card->device, PAGE_SIZE, desc,
-	kfree(ctx->buffer);
+			desc->buffer_bus -
+			((void *)&desc->buffer - (void *)desc));
 }
 
+/* Must be called with ohci->lock held */
 static struct descriptor *
 context_get_descriptors(struct context *ctx, int z, dma_addr_t *d_bus)
 {
-	struct descriptor *d, *tail, *end;
+	struct descriptor *d = NULL;
+	struct descriptor_buffer *desc = ctx->buffer_tail;
 
-	d = ctx->head_descriptor;
+	if (z * sizeof(*d) > desc->buffer_size)
-	tail = ctx->tail_descriptor;
+		return NULL;
-	end = ctx->buffer + ctx->buffer_size / sizeof(*d);
 
-	if (d + z <= tail) {
+	if (z * sizeof(*d) > desc->buffer_size - desc->used) {
-		goto has_space;
+		/* No room for the descriptor in this buffer, so advance to the
-	} else if (d > tail && d + z <= end) {
+		 * next one. */
-		goto has_space;
+
-	} else if (d > tail && ctx->buffer + z <= tail) {
+		if (desc->list.next == &ctx->buffer_list) {
-		d = ctx->buffer;
+			/* If there is no free buffer next in the list,
-		goto has_space;
+			 * allocate one. */
+			if (context_add_buffer(ctx) < 0)
+				return NULL;
+		}
+		desc = list_entry(desc->list.next,
+				struct descriptor_buffer, list);
+		ctx->buffer_tail = desc;
 	}
 
-	return NULL;
+	d = desc->buffer + desc->used / sizeof(*d);
-
- has_space:
 	memset(d, 0, z * sizeof(*d));
-	*d_bus = ctx->buffer_bus + (d - ctx->buffer) * sizeof(*d);
+	*d_bus = desc->buffer_bus + desc->used;
 
 	return d;
 }
@@ -566,7 +640,7 @@ static void context_run(struct context *ctx, u32 extra)
 	struct fw_ohci *ohci = ctx->ohci;
 
 	reg_write(ohci, COMMAND_PTR(ctx->regs),
-		  le32_to_cpu(ctx->tail_descriptor_last->branch_address));
+		  le32_to_cpu(ctx->last->branch_address));
 	reg_write(ohci, CONTROL_CLEAR(ctx->regs), ~0);
 	reg_write(ohci, CONTROL_SET(ctx->regs), CONTEXT_RUN | extra);
 	flush_writes(ohci);
@@ -576,15 +650,13 @@ static void context_append(struct context *ctx,
 			   struct descriptor *d, int z, int extra)
 {
 	dma_addr_t d_bus;
+	struct descriptor_buffer *desc = ctx->buffer_tail;
 
-	d_bus = ctx->buffer_bus + (d - ctx->buffer) * sizeof(*d);
+	d_bus = desc->buffer_bus + (d - desc->buffer) * sizeof(*d);
 
-	ctx->head_descriptor = d + z + extra;
+	desc->used += (z + extra) * sizeof(*d);
-	ctx->prev_descriptor->branch_address = cpu_to_le32(d_bus | z);
+	ctx->prev->branch_address = cpu_to_le32(d_bus | z);
-	ctx->prev_descriptor = find_branch_descriptor(d, z);
+	ctx->prev = find_branch_descriptor(d, z);
-
-	dma_sync_single_for_device(ctx->ohci->card.device, ctx->buffer_bus,
-				   ctx->buffer_size, DMA_TO_DEVICE);
 
 	reg_write(ctx->ohci, CONTROL_SET(ctx->regs), CONTEXT_WAKE);
 	flush_writes(ctx->ohci);
@@ -1078,6 +1150,13 @@ static irqreturn_t irq_handler(int irq, void *data)
 	if (unlikely(event & OHCI1394_postedWriteErr))
 		fw_error("PCI posted write error\n");
 
+	if (unlikely(event & OHCI1394_cycleTooLong)) {
+		if (printk_ratelimit())
+			fw_notify("isochronous cycle too long\n");
+		reg_write(ohci, OHCI1394_LinkControlSet,
+			  OHCI1394_LinkControl_cycleMaster);
+	}
+
 	if (event & OHCI1394_cycle64Seconds) {
 		cycle_time = reg_read(ohci, OHCI1394_IsochronousCycleTimer);
 		if ((cycle_time & 0x80000000) == 0)
@@ -1151,8 +1230,8 @@ static int ohci_enable(struct fw_card *card, u32 *config_rom, size_t length)
 		  OHCI1394_RQPkt | OHCI1394_RSPkt |
 		  OHCI1394_reqTxComplete | OHCI1394_respTxComplete |
 		  OHCI1394_isochRx | OHCI1394_isochTx |
-		  OHCI1394_postedWriteErr | OHCI1394_cycle64Seconds |
+		  OHCI1394_postedWriteErr | OHCI1394_cycleTooLong |
-		  OHCI1394_masterIntEnable);
+		  OHCI1394_cycle64Seconds | OHCI1394_masterIntEnable);
 
 	/* Activate link_on bit and contender bit in our self ID packets.*/
 	if (ohci_update_phy_reg(card, 4, 0,
@@ -1408,9 +1487,13 @@ static int handle_ir_dualbuffer_packet(struct context *context,
 	void *p, *end;
 	int i;
 
-	if (db->first_res_count > 0 && db->second_res_count > 0)
+	if (db->first_res_count > 0 && db->second_res_count > 0) {
-		/* This descriptor isn't done yet, stop iteration. */
+		if (ctx->excess_bytes <= le16_to_cpu(db->second_req_count)) {
-		return 0;
+			/* This descriptor isn't done yet, stop iteration. */
+			return 0;
+		}
+		ctx->excess_bytes -= le16_to_cpu(db->second_req_count);
+	}
 
 	header_length = le16_to_cpu(db->first_req_count) -
 		le16_to_cpu(db->first_res_count);
@@ -1429,11 +1512,15 @@ static int handle_ir_dualbuffer_packet(struct context *context,
 		*(u32 *) (ctx->header + i) = __swab32(*(u32 *) (p + 4));
 		memcpy(ctx->header + i + 4, p + 8, ctx->base.header_size - 4);
 		i += ctx->base.header_size;
+		ctx->excess_bytes +=
+			(le32_to_cpu(*(u32 *)(p + 4)) >> 16) & 0xffff;
 		p += ctx->base.header_size + 4;
 	}
-
 	ctx->header_length = i;
 
+	ctx->excess_bytes -= le16_to_cpu(db->second_req_count) -
+		le16_to_cpu(db->second_res_count);
+
 	if (le16_to_cpu(db->control) & DESCRIPTOR_IRQ_ALWAYS) {
 		ir_header = (__le32 *) (db + 1);
 		ctx->base.callback(&ctx->base,
@@ -1452,24 +1539,24 @@ static int handle_ir_packet_per_buffer(struct context *context,
 {
 	struct iso_context *ctx =
 		container_of(context, struct iso_context, context);
-	struct descriptor *pd = d + 1;
+	struct descriptor *pd;
 	__le32 *ir_header;
-	size_t header_length;
+	void *p;
-	void *p, *end;
+	int i;
-	int i, z;
 
-	if (pd->res_count == pd->req_count)
+	for (pd = d; pd <= last; pd++) {
+		if (pd->transfer_status)
+			break;
+	}
+	if (pd > last)
 		/* Descriptor(s) not done yet, stop iteration */
 		return 0;
 
-	header_length = le16_to_cpu(d->req_count);
-
 	i   = ctx->header_length;
-	z   = le32_to_cpu(pd->branch_address) & 0xf;
+	p   = last + 1;
-	p   = d + z;
-	end = p + header_length;
 
-	while (p < end && i + ctx->base.header_size <= PAGE_SIZE) {
+	if (ctx->base.header_size > 0 &&
+			i + ctx->base.header_size <= PAGE_SIZE) {
 		/*
 		 * The iso header is byteswapped to little endian by
 		 * the controller, but the remaining header quadlets
@@ -1478,14 +1565,11 @@ static int handle_ir_packet_per_buffer(struct context *context,
 		 */
 		*(u32 *) (ctx->header + i) = __swab32(*(u32 *) (p + 4));
 		memcpy(ctx->header + i + 4, p + 8, ctx->base.header_size - 4);
-		i += ctx->base.header_size;
+		ctx->header_length += ctx->base.header_size;
-		p += ctx->base.header_size + 4;
 	}
 
-	ctx->header_length = i;
+	if (le16_to_cpu(last->control) & DESCRIPTOR_IRQ_ALWAYS) {
-
+		ir_header = (__le32 *) p;
-	if (le16_to_cpu(pd->control) & DESCRIPTOR_IRQ_ALWAYS) {
-		ir_header = (__le32 *) (d + z);
 		ctx->base.callback(&ctx->base,
 				   le32_to_cpu(ir_header[0]) & 0xffff,
 				   ctx->header_length, ctx->header,
@@ -1493,7 +1577,6 @@ static int handle_ir_packet_per_buffer(struct context *context,
 		ctx->header_length = 0;
 	}
 
-
 	return 1;
 }
 
@@ -1559,8 +1642,7 @@ ohci_allocate_iso_context(struct fw_card *card, int type, size_t header_size)
 	if (ctx->header == NULL)
 		goto out;
 
-	retval = context_init(&ctx->context, ohci, ISO_BUFFER_SIZE,
+	retval = context_init(&ctx->context, ohci, regs, callback);
-			      regs, callback);
 	if (retval < 0)
 		goto out_with_header;
 
@@ -1775,19 +1857,6 @@ ohci_queue_iso_receive_dualbuffer(struct fw_iso_context *base,
 	 * packet, retransmit or terminate..
 	 */
 
-	if (packet->skip) {
-		d = context_get_descriptors(&ctx->context, 2, &d_bus);
-		if (d == NULL)
-			return -ENOMEM;
-
-		db = (struct db_descriptor *) d;
-		db->control = cpu_to_le16(DESCRIPTOR_STATUS |
-					  DESCRIPTOR_BRANCH_ALWAYS |
-					  DESCRIPTOR_WAIT);
-		db->first_size = cpu_to_le16(ctx->base.header_size + 4);
-		context_append(&ctx->context, d, 2, 0);
-	}
-
 	p = packet;
 	z = 2;
 
@@ -1815,11 +1884,18 @@ ohci_queue_iso_receive_dualbuffer(struct fw_iso_context *base,
 		db->control = cpu_to_le16(DESCRIPTOR_STATUS |
 					  DESCRIPTOR_BRANCH_ALWAYS);
 		db->first_size = cpu_to_le16(ctx->base.header_size + 4);
-		db->first_req_count = cpu_to_le16(header_size);
+		if (p->skip && rest == p->payload_length) {
+			db->control |= cpu_to_le16(DESCRIPTOR_WAIT);
+			db->first_req_count = db->first_size;
+		} else {
+			db->first_req_count = cpu_to_le16(header_size);
+		}
 		db->first_res_count = db->first_req_count;
 		db->first_buffer = cpu_to_le32(d_bus + sizeof(*db));
 
-		if (offset + rest < PAGE_SIZE)
+		if (p->skip && rest == p->payload_length)
+			length = 4;
+		else if (offset + rest < PAGE_SIZE)
 			length = rest;
 		else
 			length = PAGE_SIZE - offset;
@@ -1835,7 +1911,8 @@ ohci_queue_iso_receive_dualbuffer(struct fw_iso_context *base,
 		context_append(&ctx->context, d, z, header_z);
 		offset = (offset + length) & ~PAGE_MASK;
 		rest -= length;
-		page++;
+		if (offset == 0)
+			page++;
 	}
 
 	return 0;
@@ -1849,67 +1926,70 @@ ohci_queue_iso_receive_packet_per_buffer(struct fw_iso_context *base,
 {
 	struct iso_context *ctx = container_of(base, struct iso_context, base);
 	struct descriptor *d = NULL, *pd = NULL;
-	struct fw_iso_packet *p;
+	struct fw_iso_packet *p = packet;
 	dma_addr_t d_bus, page_bus;
 	u32 z, header_z, rest;
-	int i, page, offset, packet_count, header_size;
+	int i, j, length;
-
+	int page, offset, packet_count, header_size, payload_per_buffer;
-	if (packet->skip) {
-		d = context_get_descriptors(&ctx->context, 1, &d_bus);
-		if (d == NULL)
-			return -ENOMEM;
-
-		d->control = cpu_to_le16(DESCRIPTOR_STATUS |
-					 DESCRIPTOR_INPUT_LAST |
-					 DESCRIPTOR_BRANCH_ALWAYS |
-					 DESCRIPTOR_WAIT);
-		context_append(&ctx->context, d, 1, 0);
-	}
-
-	/* one descriptor for header, one for payload */
-	/* FIXME: handle cases where we need multiple desc. for payload */
-	z = 2;
-	p = packet;
 
 	/*
 	 * The OHCI controller puts the status word in the
 	 * buffer too, so we need 4 extra bytes per packet.
 	 */
 	packet_count = p->header_length / ctx->base.header_size;
-	header_size  = packet_count * (ctx->base.header_size + 4);
+	header_size  = ctx->base.header_size + 4;
 
 	/* Get header size in number of descriptors. */
 	header_z = DIV_ROUND_UP(header_size, sizeof(*d));
 	page     = payload >> PAGE_SHIFT;
 	offset   = payload & ~PAGE_MASK;
-	rest     = p->payload_length;
+	payload_per_buffer = p->payload_length / packet_count;
 
 	for (i = 0; i < packet_count; i++) {
 		/* d points to the header descriptor */
+		z = DIV_ROUND_UP(payload_per_buffer + offset, PAGE_SIZE) + 1;
 		d = context_get_descriptors(&ctx->context,
-					    z + header_z, &d_bus);
+				z + header_z, &d_bus);
 		if (d == NULL)
 			return -ENOMEM;
 
-		d->control      = cpu_to_le16(DESCRIPTOR_INPUT_MORE);
+		d->control      = cpu_to_le16(DESCRIPTOR_STATUS |
+					      DESCRIPTOR_INPUT_MORE);
+		if (p->skip && i == 0)
+			d->control |= cpu_to_le16(DESCRIPTOR_WAIT);
 		d->req_count    = cpu_to_le16(header_size);
 		d->res_count    = d->req_count;
+		d->transfer_status = 0;
 		d->data_address = cpu_to_le32(d_bus + (z * sizeof(*d)));
 
-		/* pd points to the payload descriptor */
+		rest = payload_per_buffer;
-		pd = d + 1;
+		for (j = 1; j < z; j++) {
+			pd = d + j;
+			pd->control = cpu_to_le16(DESCRIPTOR_STATUS |
+						  DESCRIPTOR_INPUT_MORE);
+
+			if (offset + rest < PAGE_SIZE)
+				length = rest;
+			else
+				length = PAGE_SIZE - offset;
+			pd->req_count = cpu_to_le16(length);
+			pd->res_count = pd->req_count;
+			pd->transfer_status = 0;
+
+			page_bus = page_private(buffer->pages[page]);
+			pd->data_address = cpu_to_le32(page_bus + offset);
+
+			offset = (offset + length) & ~PAGE_MASK;
+			rest -= length;
+			if (offset == 0)
+				page++;
+		}
 		pd->control = cpu_to_le16(DESCRIPTOR_STATUS |
 					  DESCRIPTOR_INPUT_LAST |
 					  DESCRIPTOR_BRANCH_ALWAYS);
-		if (p->interrupt)
+		if (p->interrupt && i == packet_count - 1)
 			pd->control |= cpu_to_le16(DESCRIPTOR_IRQ_ALWAYS);
 
-		pd->req_count = cpu_to_le16(rest);
-		pd->res_count = pd->req_count;
-
-		page_bus = page_private(buffer->pages[page]);
-		pd->data_address = cpu_to_le32(page_bus + offset);
-
 		context_append(&ctx->context, d, z, header_z);
 	}
 
@@ -1923,16 +2003,22 @@ ohci_queue_iso(struct fw_iso_context *base,
 	       unsigned long payload)
 {
 	struct iso_context *ctx = container_of(base, struct iso_context, base);
+	unsigned long flags;
+	int retval;
 
+	spin_lock_irqsave(&ctx->context.ohci->lock, flags);
 	if (base->type == FW_ISO_CONTEXT_TRANSMIT)
-		return ohci_queue_iso_transmit(base, packet, buffer, payload);
+		retval = ohci_queue_iso_transmit(base, packet, buffer, payload);
 	else if (ctx->context.ohci->version >= OHCI_VERSION_1_1)
-		return ohci_queue_iso_receive_dualbuffer(base, packet,
+		retval = ohci_queue_iso_receive_dualbuffer(base, packet,
 							 buffer, payload);
 	else
-		return ohci_queue_iso_receive_packet_per_buffer(base, packet,
+		retval = ohci_queue_iso_receive_packet_per_buffer(base, packet,
 								buffer,
 								payload);
+	spin_unlock_irqrestore(&ctx->context.ohci->lock, flags);
+
+	return retval;
 }
 
 static const struct fw_card_driver ohci_driver = {
@@ -2004,10 +2090,10 @@ pci_probe(struct pci_dev *dev, const struct pci_device_id *ent)
 	ar_context_init(&ohci->ar_response_ctx, ohci,
 			OHCI1394_AsRspRcvContextControlSet);
 
-	context_init(&ohci->at_request_ctx, ohci, AT_BUFFER_SIZE,
+	context_init(&ohci->at_request_ctx, ohci,
 		     OHCI1394_AsReqTrContextControlSet, handle_at_packet);
 
-	context_init(&ohci->at_response_ctx, ohci, AT_BUFFER_SIZE,
+	context_init(&ohci->at_response_ctx, ohci,
 		     OHCI1394_AsRspTrContextControlSet, handle_at_packet);
 
 	reg_write(ohci, OHCI1394_IsoRecvIntMaskSet, ~0);

drivers/firewire/fw-sbp2.c

@@ -40,6 +40,7 @@
 #include <linux/stringify.h>
 #include <linux/timer.h>
 #include <linux/workqueue.h>
+#include <asm/system.h>
 
 #include <scsi/scsi.h>
 #include <scsi/scsi_cmnd.h>
@@ -148,18 +149,26 @@ struct sbp2_target {
 
 	unsigned workarounds;
 	struct list_head lu_list;
+
+	unsigned int mgt_orb_timeout;
 };
 
-#define SBP2_MAX_SG_ELEMENT_LENGTH	0xf000
+/*
-#define SBP2_MAX_SECTORS		255	/* Max sectors supported */
+ * Per section 7.4.8 of the SBP-2 spec, a mgt_ORB_timeout value can be
+ * provided in the config rom. Most devices do provide a value, which
+ * we'll use for login management orbs, but with some sane limits.
+ */
+#define SBP2_MIN_LOGIN_ORB_TIMEOUT	5000U	/* Timeout in ms */
+#define SBP2_MAX_LOGIN_ORB_TIMEOUT	40000U	/* Timeout in ms */
 #define SBP2_ORB_TIMEOUT		2000	/* Timeout in ms */
-
 #define SBP2_ORB_NULL			0x80000000
+#define SBP2_MAX_SG_ELEMENT_LENGTH	0xf000
 
 #define SBP2_DIRECTION_TO_MEDIA		0x0
 #define SBP2_DIRECTION_FROM_MEDIA	0x1
 
 /* Unit directory keys */
+#define SBP2_CSR_UNIT_CHARACTERISTICS	0x3a
 #define SBP2_CSR_FIRMWARE_REVISION	0x3c
 #define SBP2_CSR_LOGICAL_UNIT_NUMBER	0x14
 #define SBP2_CSR_LOGICAL_UNIT_DIRECTORY	0xd4
@@ -489,6 +498,7 @@ sbp2_send_management_orb(struct sbp2_logical_unit *lu, int node_id,
 {
 	struct fw_device *device = fw_device(lu->tgt->unit->device.parent);
 	struct sbp2_management_orb *orb;
+	unsigned int timeout;
 	int retval = -ENOMEM;
 
 	orb = kzalloc(sizeof(*orb), GFP_ATOMIC);
@@ -516,9 +526,13 @@ sbp2_send_management_orb(struct sbp2_logical_unit *lu, int node_id,
 	orb->request.status_fifo.low  = lu->address_handler.offset;
 
 	if (function == SBP2_LOGIN_REQUEST) {
+		/* Ask for 2^2 == 4 seconds reconnect grace period */
 		orb->request.misc |=
-			MANAGEMENT_ORB_EXCLUSIVE(sbp2_param_exclusive_login) |
+			MANAGEMENT_ORB_RECONNECT(2) |
-			MANAGEMENT_ORB_RECONNECT(0);
+			MANAGEMENT_ORB_EXCLUSIVE(sbp2_param_exclusive_login);
+		timeout = lu->tgt->mgt_orb_timeout;
+	} else {
+		timeout = SBP2_ORB_TIMEOUT;
 	}
 
 	fw_memcpy_to_be32(&orb->request, &orb->request, sizeof(orb->request));
@@ -535,8 +549,7 @@ sbp2_send_management_orb(struct sbp2_logical_unit *lu, int node_id,
 	sbp2_send_orb(&orb->base, lu, node_id, generation,
 		      lu->tgt->management_agent_address);
 
-	wait_for_completion_timeout(&orb->done,
+	wait_for_completion_timeout(&orb->done, msecs_to_jiffies(timeout));
-				    msecs_to_jiffies(SBP2_ORB_TIMEOUT));
 
 	retval = -EIO;
 	if (sbp2_cancel_orbs(lu) == 0) {
@@ -608,13 +621,17 @@ static void sbp2_release_target(struct kref *kref)
 	struct sbp2_logical_unit *lu, *next;
 	struct Scsi_Host *shost =
 		container_of((void *)tgt, struct Scsi_Host, hostdata[0]);
+	struct fw_device *device = fw_device(tgt->unit->device.parent);
 
 	list_for_each_entry_safe(lu, next, &tgt->lu_list, link) {
 		if (lu->sdev)
 			scsi_remove_device(lu->sdev);
 
-		sbp2_send_management_orb(lu, tgt->node_id, lu->generation,
+		if (!fw_device_is_shutdown(device))
-				SBP2_LOGOUT_REQUEST, lu->login_id, NULL);
+			sbp2_send_management_orb(lu, tgt->node_id,
+					lu->generation, SBP2_LOGOUT_REQUEST,
+					lu->login_id, NULL);
+
 		fw_core_remove_address_handler(&lu->address_handler);
 		list_del(&lu->link);
 		kfree(lu);
@@ -628,6 +645,21 @@ static void sbp2_release_target(struct kref *kref)
 
 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.
+ */
+static void sbp2_queue_work(struct sbp2_logical_unit *lu, unsigned long delay)
+{
+	if (queue_delayed_work(sbp2_wq, &lu->work, delay))
+		kref_get(&lu->tgt->kref);
+}
+
+static void sbp2_target_put(struct sbp2_target *tgt)
+{
+	kref_put(&tgt->kref, sbp2_release_target);
+}
+
 static void sbp2_reconnect(struct work_struct *work);
 
 static void sbp2_login(struct work_struct *work)
@@ -643,22 +675,19 @@ static void sbp2_login(struct work_struct *work)
 	struct sbp2_login_response response;
 	int generation, node_id, local_node_id;
 
-	generation    = device->card->generation;
+	generation    = device->generation;
-	node_id       = device->node->node_id;
+	smp_rmb();    /* node_id must not be older than generation */
-	local_node_id = device->card->local_node->node_id;
+	node_id       = device->node_id;
+	local_node_id = device->card->node_id;
 
 	if (sbp2_send_management_orb(lu, node_id, generation,
 				SBP2_LOGIN_REQUEST, lu->lun, &response) < 0) {
-		if (lu->retries++ < 5) {
+		if (lu->retries++ < 5)
-			if (queue_delayed_work(sbp2_wq, &lu->work,
+			sbp2_queue_work(lu, DIV_ROUND_UP(HZ, 5));
-					       DIV_ROUND_UP(HZ, 5)))
+		else
-				kref_get(&lu->tgt->kref);
-		} else {
 			fw_error("failed to login to %s LUN %04x\n",
 				 unit->device.bus_id, lu->lun);
-		}
+		goto out;
-		kref_put(&lu->tgt->kref, sbp2_release_target);
-		return;
 	}
 
 	lu->generation        = generation;
@@ -700,7 +729,8 @@ static void sbp2_login(struct work_struct *work)
 		lu->sdev = sdev;
 		scsi_device_put(sdev);
 	}
-	kref_put(&lu->tgt->kref, sbp2_release_target);
+ out:
+	sbp2_target_put(lu->tgt);
 }
 
 static int sbp2_add_logical_unit(struct sbp2_target *tgt, int lun_entry)
@@ -750,6 +780,7 @@ static int sbp2_scan_unit_dir(struct sbp2_target *tgt, u32 *directory,
 {
 	struct fw_csr_iterator ci;
 	int key, value;
+	unsigned int timeout;
 
 	fw_csr_iterator_init(&ci, directory);
 	while (fw_csr_iterator_next(&ci, &key, &value)) {
@@ -772,6 +803,21 @@ static int sbp2_scan_unit_dir(struct sbp2_target *tgt, u32 *directory,
 			*firmware_revision = value;
 			break;
 
+		case SBP2_CSR_UNIT_CHARACTERISTICS:
+			/* the timeout value is stored in 500ms units */
+			timeout = ((unsigned int) value >> 8 & 0xff) * 500;
+			timeout = max(timeout, SBP2_MIN_LOGIN_ORB_TIMEOUT);
+			tgt->mgt_orb_timeout =
+				  min(timeout, SBP2_MAX_LOGIN_ORB_TIMEOUT);
+
+			if (timeout > tgt->mgt_orb_timeout)
+				fw_notify("%s: config rom contains %ds "
+					  "management ORB timeout, limiting "
+					  "to %ds\n", tgt->unit->device.bus_id,
+					  timeout / 1000,
+					  tgt->mgt_orb_timeout / 1000);
+			break;
+
 		case SBP2_CSR_LOGICAL_UNIT_NUMBER:
 			if (sbp2_add_logical_unit(tgt, value) < 0)
 				return -ENOMEM;
@@ -865,18 +911,13 @@ static int sbp2_probe(struct device *dev)
 
 	get_device(&unit->device);
 
-	/*
+	/* Do the login in a workqueue so we can easily reschedule retries. */
-	 * We schedule work to do the login so we can easily
-	 * reschedule retries. Always get the ref before scheduling
-	 * work.
-	 */
 	list_for_each_entry(lu, &tgt->lu_list, link)
-		if (queue_delayed_work(sbp2_wq, &lu->work, 0))
+		sbp2_queue_work(lu, 0);
-			kref_get(&tgt->kref);
 	return 0;
 
  fail_tgt_put:
-	kref_put(&tgt->kref, sbp2_release_target);
+	sbp2_target_put(tgt);
 	return -ENOMEM;
 
  fail_shost_put:
@@ -889,7 +930,7 @@ static int sbp2_remove(struct device *dev)
 	struct fw_unit *unit = fw_unit(dev);
 	struct sbp2_target *tgt = unit->device.driver_data;
 
-	kref_put(&tgt->kref, sbp2_release_target);
+	sbp2_target_put(tgt);
 	return 0;
 }
 
@@ -901,9 +942,10 @@ static void sbp2_reconnect(struct work_struct *work)
 	struct fw_device *device = fw_device(unit->device.parent);
 	int generation, node_id, local_node_id;
 
-	generation    = device->card->generation;
+	generation    = device->generation;
-	node_id       = device->node->node_id;
+	smp_rmb();    /* node_id must not be older than generation */
-	local_node_id = device->card->local_node->node_id;
+	node_id       = device->node_id;
+	local_node_id = device->card->node_id;
 
 	if (sbp2_send_management_orb(lu, node_id, generation,
 				     SBP2_RECONNECT_REQUEST,
@@ -915,10 +957,8 @@ static void sbp2_reconnect(struct work_struct *work)
 			lu->retries = 0;
 			PREPARE_DELAYED_WORK(&lu->work, sbp2_login);
 		}
-		if (queue_delayed_work(sbp2_wq, &lu->work, DIV_ROUND_UP(HZ, 5)))
+		sbp2_queue_work(lu, DIV_ROUND_UP(HZ, 5));
-			kref_get(&lu->tgt->kref);
+		goto out;
-		kref_put(&lu->tgt->kref, sbp2_release_target);
-		return;
 	}
 
 	lu->generation        = generation;
@@ -930,8 +970,8 @@ static void sbp2_reconnect(struct work_struct *work)
 
 	sbp2_agent_reset(lu);
 	sbp2_cancel_orbs(lu);
-
+ out:
-	kref_put(&lu->tgt->kref, sbp2_release_target);
+	sbp2_target_put(lu->tgt);
 }
 
 static void sbp2_update(struct fw_unit *unit)
@@ -947,8 +987,7 @@ static void sbp2_update(struct fw_unit *unit)
 	 */
 	list_for_each_entry(lu, &tgt->lu_list, link) {
 		lu->retries = 0;
-		if (queue_delayed_work(sbp2_wq, &lu->work, 0))
+		sbp2_queue_work(lu, 0);
-			kref_get(&tgt->kref);
 	}
 }
 
@@ -1103,9 +1142,9 @@ sbp2_map_scatterlist(struct sbp2_command_orb *orb, struct fw_device *device,
 	 * elements larger than 65535 bytes, some IOMMUs may merge sg elements
 	 * during DMA mapping, and Linux currently doesn't prevent this.
 	 */
-	for (i = 0, j = 0; i < count; i++) {
+	for (i = 0, j = 0; i < count; i++, sg = sg_next(sg)) {
-		sg_len = sg_dma_len(sg + i);
+		sg_len = sg_dma_len(sg);
-		sg_addr = sg_dma_address(sg + i);
+		sg_addr = sg_dma_address(sg);
 		while (sg_len) {
 			/* FIXME: This won't get us out of the pinch. */
 			if (unlikely(j >= ARRAY_SIZE(orb->page_table))) {

drivers/firewire/fw-topology.c

@@ -21,6 +21,7 @@
 #include <linux/module.h>
 #include <linux/wait.h>
 #include <linux/errno.h>
+#include <asm/system.h>
 #include "fw-transaction.h"
 #include "fw-topology.h"
 
@@ -518,6 +519,11 @@ fw_core_handle_bus_reset(struct fw_card *card,
 		card->bm_retries = 0;
 
 	card->node_id = node_id;
+	/*
+	 * Update node_id before generation to prevent anybody from using
+	 * a stale node_id together with a current generation.
+	 */
+	smp_wmb();
 	card->generation = generation;
 	card->reset_jiffies = jiffies;
 	schedule_delayed_work(&card->work, 0);

drivers/firewire/fw-transaction.c

@@ -153,7 +153,7 @@ fw_fill_request(struct fw_packet *packet, int tcode, int tlabel,
 	int ext_tcode;
 
 	if (tcode > 0x10) {
-		ext_tcode = tcode - 0x10;
+		ext_tcode = tcode & ~0x10;
 		tcode = TCODE_LOCK_REQUEST;
 	} else
 		ext_tcode = 0;
@@ -650,7 +650,7 @@ fw_core_handle_request(struct fw_card *card, struct fw_packet *p)
 		 HEADER_GET_OFFSET_HIGH(p->header[1]) << 32) | p->header[2];
 	tcode       = HEADER_GET_TCODE(p->header[0]);
 	destination = HEADER_GET_DESTINATION(p->header[0]);
-	source      = HEADER_GET_SOURCE(p->header[0]);
+	source      = HEADER_GET_SOURCE(p->header[1]);
 
 	spin_lock_irqsave(&address_handler_lock, flags);
 	handler = lookup_enclosing_address_handler(&address_handler_list,

drivers/ieee1394/dma.c

@@ -231,37 +231,24 @@ void dma_region_sync_for_device(struct dma_region *dma, unsigned long offset,
 
 #ifdef CONFIG_MMU
 
-/* nopage() handler for mmap access */
+static int dma_region_pagefault(struct vm_area_struct *vma,
-
+				struct vm_fault *vmf)
-static struct page *dma_region_pagefault(struct vm_area_struct *area,
-					 unsigned long address, int *type)
 {
-	unsigned long offset;
+	struct dma_region *dma = (struct dma_region *)vma->vm_private_data;
-	unsigned long kernel_virt_addr;
-	struct page *ret = NOPAGE_SIGBUS;
-
-	struct dma_region *dma = (struct dma_region *)area->vm_private_data;
 
 	if (!dma->kvirt)
-		goto out;
+		return VM_FAULT_SIGBUS;
 
-	if ((address < (unsigned long)area->vm_start) ||
+	if (vmf->pgoff >= dma->n_pages)
-	    (address >
+		return VM_FAULT_SIGBUS;
-	     (unsigned long)area->vm_start + (dma->n_pages << PAGE_SHIFT)))
-		goto out;
 
-	if (type)
+	vmf->page = vmalloc_to_page(dma->kvirt + (vmf->pgoff << PAGE_SHIFT));
-		*type = VM_FAULT_MINOR;
+	get_page(vmf->page);
-	offset = address - area->vm_start;
+	return 0;
-	kernel_virt_addr = (unsigned long)dma->kvirt + offset;
-	ret = vmalloc_to_page((void *)kernel_virt_addr);
-	get_page(ret);
-      out:
-	return ret;
 }
 
 static struct vm_operations_struct dma_region_vm_ops = {
-	.nopage = dma_region_pagefault,
+	.fault = dma_region_pagefault,
 };
 
 /**
@@ -275,7 +262,7 @@ int dma_region_mmap(struct dma_region *dma, struct file *file,
 	if (!dma->kvirt)
 		return -EINVAL;
 
-	/* must be page-aligned */
+	/* must be page-aligned (XXX: comment is wrong, we could allow pgoff) */
 	if (vma->vm_pgoff != 0)
 		return -EINVAL;
 

drivers/ieee1394/ieee1394_transactions.c

@@ -570,71 +570,3 @@ int hpsb_write(struct hpsb_host *host, nodeid_t node, unsigned int generation,
 
 	return retval;
 }
-
-#if 0
-
-int hpsb_lock(struct hpsb_host *host, nodeid_t node, unsigned int generation,
-	      u64 addr, int extcode, quadlet_t * data, quadlet_t arg)
-{
-	struct hpsb_packet *packet;
-	int retval = 0;
-
-	BUG_ON(in_interrupt());	// We can't be called in an interrupt, yet
-
-	packet = hpsb_make_lockpacket(host, node, addr, extcode, data, arg);
-	if (!packet)
-		return -ENOMEM;
-
-	packet->generation = generation;
-	retval = hpsb_send_packet_and_wait(packet);
-	if (retval < 0)
-		goto hpsb_lock_fail;
-
-	retval = hpsb_packet_success(packet);
-
-	if (retval == 0) {
-		*data = packet->data[0];
-	}
-
-      hpsb_lock_fail:
-	hpsb_free_tlabel(packet);
-	hpsb_free_packet(packet);
-
-	return retval;
-}
-
-int hpsb_send_gasp(struct hpsb_host *host, int channel, unsigned int generation,
-		   quadlet_t * buffer, size_t length, u32 specifier_id,
-		   unsigned int version)
-{
-	struct hpsb_packet *packet;
-	int retval = 0;
-	u16 specifier_id_hi = (specifier_id & 0x00ffff00) >> 8;
-	u8 specifier_id_lo = specifier_id & 0xff;
-
-	HPSB_VERBOSE("Send GASP: channel = %d, length = %Zd", channel, length);
-
-	length += 8;
-
-	packet = hpsb_make_streampacket(host, NULL, length, channel, 3, 0);
-	if (!packet)
-		return -ENOMEM;
-
-	packet->data[0] = cpu_to_be32((host->node_id << 16) | specifier_id_hi);
-	packet->data[1] =
-	    cpu_to_be32((specifier_id_lo << 24) | (version & 0x00ffffff));
-
-	memcpy(&(packet->data[2]), buffer, length - 8);
-
-	packet->generation = generation;
-
-	packet->no_waiter = 1;
-
-	retval = hpsb_send_packet(packet);
-	if (retval < 0)
-		hpsb_free_packet(packet);
-
-	return retval;
-}
-
-#endif				/*  0  */

drivers/ieee1394/ohci1394.c

@@ -2126,10 +2126,14 @@ static void ohci_schedule_iso_tasklets(struct ti_ohci *ohci,
 	list_for_each_entry(t, &ohci->iso_tasklet_list, link) {
 		mask = 1 << t->context;
 
-		if (t->type == OHCI_ISO_TRANSMIT && tx_event & mask)
+		if (t->type == OHCI_ISO_TRANSMIT) {
-			tasklet_schedule(&t->tasklet);
+			if (tx_event & mask)
-		else if (rx_event & mask)
+				tasklet_schedule(&t->tasklet);
-			tasklet_schedule(&t->tasklet);
+		} else {
+			/* OHCI_ISO_RECEIVE or OHCI_ISO_MULTICHANNEL_RECEIVE */
+			if (rx_event & mask)
+				tasklet_schedule(&t->tasklet);
+		}
 	}
 
 	spin_unlock_irqrestore(&ohci->iso_tasklet_list_lock, flags);

drivers/ieee1394/raw1394.c

@@ -858,7 +858,7 @@ static int arm_read(struct hpsb_host *host, int nodeid, quadlet_t * buffer,
 	int found = 0, size = 0, rcode = -1;
 	struct arm_request_response *arm_req_resp = NULL;
 
-	DBGMSG("arm_read  called by node: %X"
+	DBGMSG("arm_read  called by node: %X "
 	       "addr: %4.4x %8.8x length: %Zu", nodeid,
 	       (u16) ((addr >> 32) & 0xFFFF), (u32) (addr & 0xFFFFFFFF),
 	       length);
@@ -1012,7 +1012,7 @@ static int arm_write(struct hpsb_host *host, int nodeid, int destid,
 	int found = 0, size = 0, rcode = -1, length_conflict = 0;
 	struct arm_request_response *arm_req_resp = NULL;
 
-	DBGMSG("arm_write called by node: %X"
+	DBGMSG("arm_write called by node: %X "
 	       "addr: %4.4x %8.8x length: %Zu", nodeid,
 	       (u16) ((addr >> 32) & 0xFFFF), (u32) (addr & 0xFFFFFFFF),
 	       length);

drivers/ieee1394/sbp2.c

@@ -51,6 +51,7 @@
  * Grep for inline FIXME comments below.
  */
 
+#include <linux/blkdev.h>
 #include <linux/compiler.h>
 #include <linux/delay.h>
 #include <linux/device.h>
@@ -127,17 +128,21 @@ MODULE_PARM_DESC(serialize_io, "Serialize requests coming from SCSI drivers "
 		 "(default = Y, faster but buggy = N)");
 
 /*
- * Bump up max_sectors if you'd like to support very large sized
+ * Adjust max_sectors if you'd like to influence how many sectors each SCSI
- * transfers. Please note that some older sbp2 bridge chips are broken for
+ * command can transfer at most. Please note that some older SBP-2 bridge
- * transfers greater or equal to 128KB.  Default is a value of 255
+ * chips are broken for transfers greater or equal to 128KB, therefore
- * sectors, or just under 128KB (at 512 byte sector size). I can note that
+ * max_sectors used to be a safe 255 sectors for many years. We now have a
- * the Oxsemi sbp2 chipsets have no problems supporting very large
+ * default of 0 here which means that we let the SCSI stack choose a limit.
- * transfer sizes.
+ *
+ * The SBP2_WORKAROUND_128K_MAX_TRANS flag, if set either in the workarounds
+ * module parameter or in the sbp2_workarounds_table[], will override the
+ * value of max_sectors. We should use sbp2_workarounds_table[] to cover any
+ * bridge chip which becomes known to need the 255 sectors limit.
  */
-static int sbp2_max_sectors = SBP2_MAX_SECTORS;
+static int sbp2_max_sectors;
 module_param_named(max_sectors, sbp2_max_sectors, int, 0444);
 MODULE_PARM_DESC(max_sectors, "Change max sectors per I/O supported "
-		 "(default = " __stringify(SBP2_MAX_SECTORS) ")");
+		 "(default = 0 = use SCSI stack's default)");
 
 /*
  * Exclusive login to sbp2 device? In most cases, the sbp2 driver should
@@ -1451,7 +1456,7 @@ static void sbp2_prep_command_orb_sg(struct sbp2_command_orb *orb,
 				     struct sbp2_fwhost_info *hi,
 				     struct sbp2_command_info *cmd,
 				     unsigned int scsi_use_sg,
-				     struct scatterlist *sgpnt,
+				     struct scatterlist *sg,
 				     u32 orb_direction,
 				     enum dma_data_direction dma_dir)
 {
@@ -1461,12 +1466,12 @@ static void sbp2_prep_command_orb_sg(struct sbp2_command_orb *orb,
 
 	/* special case if only one element (and less than 64KB in size) */
 	if ((scsi_use_sg == 1) &&
-	    (sgpnt[0].length <= SBP2_MAX_SG_ELEMENT_LENGTH)) {
+	    (sg_dma_len(sg) <= SBP2_MAX_SG_ELEMENT_LENGTH)) {
 
-		cmd->dma_size = sgpnt[0].length;
+		cmd->dma_size = sg_dma_len(sg);
 		cmd->dma_type = CMD_DMA_PAGE;
 		cmd->cmd_dma = dma_map_page(hi->host->device.parent,
-					    sg_page(&sgpnt[0]), sgpnt[0].offset,
+					    sg_page(sg), sg->offset,
 					    cmd->dma_size, cmd->dma_dir);
 
 		orb->data_descriptor_lo = cmd->cmd_dma;
@@ -1477,11 +1482,11 @@ static void sbp2_prep_command_orb_sg(struct sbp2_command_orb *orb,
 						&cmd->scatter_gather_element[0];
 		u32 sg_count, sg_len;
 		dma_addr_t sg_addr;
-		int i, count = dma_map_sg(hi->host->device.parent, sgpnt,
+		int i, count = dma_map_sg(hi->host->device.parent, sg,
 					  scsi_use_sg, dma_dir);
 
 		cmd->dma_size = scsi_use_sg;
-		cmd->sge_buffer = sgpnt;
+		cmd->sge_buffer = sg;
 
 		/* use page tables (s/g) */
 		orb->misc |= ORB_SET_PAGE_TABLE_PRESENT(0x1);
@@ -1489,9 +1494,9 @@ static void sbp2_prep_command_orb_sg(struct sbp2_command_orb *orb,
 
 		/* loop through and fill out our SBP-2 page tables
 		 * (and split up anything too large) */
-		for (i = 0, sg_count = 0 ; i < count; i++, sgpnt++) {
+		for (i = 0, sg_count = 0; i < count; i++, sg = sg_next(sg)) {
-			sg_len = sg_dma_len(sgpnt);
+			sg_len = sg_dma_len(sg);
-			sg_addr = sg_dma_address(sgpnt);
+			sg_addr = sg_dma_address(sg);
 			while (sg_len) {
 				sg_element[sg_count].segment_base_lo = sg_addr;
 				if (sg_len > SBP2_MAX_SG_ELEMENT_LENGTH) {
@@ -1521,11 +1526,10 @@ static void sbp2_create_command_orb(struct sbp2_lu *lu,
 				    unchar *scsi_cmd,
 				    unsigned int scsi_use_sg,
 				    unsigned int scsi_request_bufflen,
-				    void *scsi_request_buffer,
+				    struct scatterlist *sg,
 				    enum dma_data_direction dma_dir)
 {
 	struct sbp2_fwhost_info *hi = lu->hi;
-	struct scatterlist *sgpnt = (struct scatterlist *)scsi_request_buffer;
 	struct sbp2_command_orb *orb = &cmd->command_orb;
 	u32 orb_direction;
 
@@ -1560,7 +1564,7 @@ static void sbp2_create_command_orb(struct sbp2_lu *lu,
 		orb->data_descriptor_lo = 0x0;
 		orb->misc |= ORB_SET_DIRECTION(1);
 	} else
-		sbp2_prep_command_orb_sg(orb, hi, cmd, scsi_use_sg, sgpnt,
+		sbp2_prep_command_orb_sg(orb, hi, cmd, scsi_use_sg, sg,
 					 orb_direction, dma_dir);
 
 	sbp2util_cpu_to_be32_buffer(orb, sizeof(*orb));
@@ -1650,7 +1654,6 @@ static int sbp2_send_command(struct sbp2_lu *lu, struct scsi_cmnd *SCpnt,
 			     void (*done)(struct scsi_cmnd *))
 {
 	unchar *scsi_cmd = (unchar *)SCpnt->cmnd;
-	unsigned int request_bufflen = scsi_bufflen(SCpnt);
 	struct sbp2_command_info *cmd;
 
 	cmd = sbp2util_allocate_command_orb(lu, SCpnt, done);
@@ -1658,7 +1661,7 @@ static int sbp2_send_command(struct sbp2_lu *lu, struct scsi_cmnd *SCpnt,
 		return -EIO;
 
 	sbp2_create_command_orb(lu, cmd, scsi_cmd, scsi_sg_count(SCpnt),
-				request_bufflen, scsi_sglist(SCpnt),
+				scsi_bufflen(SCpnt), scsi_sglist(SCpnt),
 				SCpnt->sc_data_direction);
 	sbp2_link_orb_command(lu, cmd);
 
@@ -1987,6 +1990,8 @@ static int sbp2scsi_slave_configure(struct scsi_device *sdev)
 		sdev->skip_ms_page_8 = 1;
 	if (lu->workarounds & SBP2_WORKAROUND_FIX_CAPACITY)
 		sdev->fix_capacity = 1;
+	if (lu->workarounds & SBP2_WORKAROUND_128K_MAX_TRANS)
+		blk_queue_max_sectors(sdev->request_queue, 128 * 1024 / 512);
 	return 0;
 }
 
@@ -2093,9 +2098,6 @@ static int sbp2_module_init(void)
 		sbp2_shost_template.cmd_per_lun = 1;
 	}
 
-	if (sbp2_default_workarounds & SBP2_WORKAROUND_128K_MAX_TRANS &&
-	    (sbp2_max_sectors * 512) > (128 * 1024))
-		sbp2_max_sectors = 128 * 1024 / 512;
 	sbp2_shost_template.max_sectors = sbp2_max_sectors;
 
 	hpsb_register_highlevel(&sbp2_highlevel);

drivers/ieee1394/sbp2.h

@@ -222,7 +222,6 @@ struct sbp2_status_block {
  */
 
 #define SBP2_MAX_SG_ELEMENT_LENGTH		0xf000
-#define SBP2_MAX_SECTORS			255
 /* There is no real limitation of the queue depth (i.e. length of the linked
  * list of command ORBs) at the target. The chosen depth is merely an
  * implementation detail of the sbp2 driver. */