Merge branch 'linus' into oprofile-v2

Conflicts:
	arch/x86/kernel/apic_32.c
	arch/x86/oprofile/nmi_int.c
	include/linux/pci_ids.h

18 files changed, 485 insertions, 256 deletions

diff --git a/drivers/md/bitmap.c b/drivers/md/bitmap.c
index 621a272..ac89a5d 100644
--- a/drivers/md/bitmap.c
+++ b/drivers/md/bitmap.c
@@ -238,15 +238,47 @@ static struct page *read_sb_page(mddev_t *mddev, long offset, unsigned long inde
 
 }
 
+static mdk_rdev_t *next_active_rdev(mdk_rdev_t *rdev, mddev_t *mddev)
+{
+	/* Iterate the disks of an mddev, using rcu to protect access to the
+	 * linked list, and raising the refcount of devices we return to ensure
+	 * they don't disappear while in use.
+	 * As devices are only added or removed when raid_disk is < 0 and
+	 * nr_pending is 0 and In_sync is clear, the entries we return will
+	 * still be in the same position on the list when we re-enter
+	 * list_for_each_continue_rcu.
+	 */
+	struct list_head *pos;
+	rcu_read_lock();
+	if (rdev == NULL)
+		/* start at the beginning */
+		pos = &mddev->disks;
+	else {
+		/* release the previous rdev and start from there. */
+		rdev_dec_pending(rdev, mddev);
+		pos = &rdev->same_set;
+	}
+	list_for_each_continue_rcu(pos, &mddev->disks) {
+		rdev = list_entry(pos, mdk_rdev_t, same_set);
+		if (rdev->raid_disk >= 0 &&
+		    test_bit(In_sync, &rdev->flags) &&
+		    !test_bit(Faulty, &rdev->flags)) {
+			/* this is a usable devices */
+			atomic_inc(&rdev->nr_pending);
+			rcu_read_unlock();
+			return rdev;
+		}
+	}
+	rcu_read_unlock();
+	return NULL;
+}
+
 static int write_sb_page(struct bitmap *bitmap, struct page *page, int wait)
 {
-	mdk_rdev_t *rdev;
+	mdk_rdev_t *rdev = NULL;
 	mddev_t *mddev = bitmap->mddev;
 
-	rcu_read_lock();
-	rdev_for_each_rcu(rdev, mddev)
-		if (test_bit(In_sync, &rdev->flags)
-		    && !test_bit(Faulty, &rdev->flags)) {
+	while ((rdev = next_active_rdev(rdev, mddev)) != NULL) {
 			int size = PAGE_SIZE;
 			if (page->index == bitmap->file_pages-1)
 				size = roundup(bitmap->last_page_size,
@@ -281,8 +313,7 @@ static int write_sb_page(struct bitmap *bitmap, struct page *page, int wait)
 				       + page->index * (PAGE_SIZE/512),
 				       size,
 				       page);
-		}
-	rcu_read_unlock();
+	}
 
 	if (wait)
 		md_super_wait(mddev);
@@ -1234,7 +1265,7 @@ int bitmap_startwrite(struct bitmap *bitmap, sector_t offset, unsigned long sect
 		case 0:
 			bitmap_file_set_bit(bitmap, offset);
 			bitmap_count_page(bitmap,offset, 1);
-			blk_plug_device(bitmap->mddev->queue);
+			blk_plug_device_unlocked(bitmap->mddev->queue);
 			/* fall through */
 		case 1:
 			*bmc = 2;
diff --git a/drivers/md/dm-crypt.c b/drivers/md/dm-crypt.c
index 1395643..682ef9e 100644
--- a/drivers/md/dm-crypt.c
+++ b/drivers/md/dm-crypt.c
@@ -333,7 +333,6 @@ static void crypt_convert_init(struct crypt_config *cc,
 	ctx->idx_out = bio_out ? bio_out->bi_idx : 0;
 	ctx->sector = sector + cc->iv_offset;
 	init_completion(&ctx->restart);
-	atomic_set(&ctx->pending, 1);
 }
 
 static int crypt_convert_block(struct crypt_config *cc,
@@ -408,6 +407,8 @@ static int crypt_convert(struct crypt_config *cc,
 {
 	int r;
 
+	atomic_set(&ctx->pending, 1);
+
 	while(ctx->idx_in < ctx->bio_in->bi_vcnt &&
 	      ctx->idx_out < ctx->bio_out->bi_vcnt) {
 
@@ -456,9 +457,11 @@ static void dm_crypt_bio_destructor(struct bio *bio)
 /*
  * Generate a new unfragmented bio with the given size
  * This should never violate the device limitations
- * May return a smaller bio when running out of pages
+ * May return a smaller bio when running out of pages, indicated by
+ * *out_of_pages set to 1.
  */
-static struct bio *crypt_alloc_buffer(struct dm_crypt_io *io, unsigned size)
+static struct bio *crypt_alloc_buffer(struct dm_crypt_io *io, unsigned size,
+				      unsigned *out_of_pages)
 {
 	struct crypt_config *cc = io->target->private;
 	struct bio *clone;
@@ -472,11 +475,14 @@ static struct bio *crypt_alloc_buffer(struct dm_crypt_io *io, unsigned size)
 		return NULL;
 
 	clone_init(io, clone);
+	*out_of_pages = 0;
 
 	for (i = 0; i < nr_iovecs; i++) {
 		page = mempool_alloc(cc->page_pool, gfp_mask);
-		if (!page)
+		if (!page) {
+			*out_of_pages = 1;
 			break;
+		}
 
 		/*
 		 * if additional pages cannot be allocated without waiting,
@@ -517,6 +523,27 @@ static void crypt_free_buffer_pages(struct crypt_config *cc, struct bio *clone)
 	}
 }
 
+static struct dm_crypt_io *crypt_io_alloc(struct dm_target *ti,
+					  struct bio *bio, sector_t sector)
+{
+	struct crypt_config *cc = ti->private;
+	struct dm_crypt_io *io;
+
+	io = mempool_alloc(cc->io_pool, GFP_NOIO);
+	io->target = ti;
+	io->base_bio = bio;
+	io->sector = sector;
+	io->error = 0;
+	atomic_set(&io->pending, 0);
+
+	return io;
+}
+
+static void crypt_inc_pending(struct dm_crypt_io *io)
+{
+	atomic_inc(&io->pending);
+}
+
 /*
  * One of the bios was finished. Check for completion of
  * the whole request and correctly clean up the buffer.
@@ -591,7 +618,7 @@ static void kcryptd_io_read(struct dm_crypt_io *io)
 	struct bio *base_bio = io->base_bio;
 	struct bio *clone;
 
-	atomic_inc(&io->pending);
+	crypt_inc_pending(io);
 
 	/*
 	 * The block layer might modify the bvec array, so always
@@ -653,6 +680,7 @@ static void kcryptd_crypt_write_io_submit(struct dm_crypt_io *io,
 		crypt_free_buffer_pages(cc, clone);
 		bio_put(clone);
 		io->error = -EIO;
+		crypt_dec_pending(io);
 		return;
 	}
 
@@ -664,28 +692,34 @@ static void kcryptd_crypt_write_io_submit(struct dm_crypt_io *io,
 
 	if (async)
 		kcryptd_queue_io(io);
-	else {
-		atomic_inc(&io->pending);
+	else
 		generic_make_request(clone);
-	}
 }
 
-static void kcryptd_crypt_write_convert_loop(struct dm_crypt_io *io)
+static void kcryptd_crypt_write_convert(struct dm_crypt_io *io)
 {
 	struct crypt_config *cc = io->target->private;
 	struct bio *clone;
+	int crypt_finished;
+	unsigned out_of_pages = 0;
 	unsigned remaining = io->base_bio->bi_size;
 	int r;
 
 	/*
+	 * Prevent io from disappearing until this function completes.
+	 */
+	crypt_inc_pending(io);
+	crypt_convert_init(cc, &io->ctx, NULL, io->base_bio, io->sector);
+
+	/*
 	 * The allocated buffers can be smaller than the whole bio,
 	 * so repeat the whole process until all the data can be handled.
 	 */
 	while (remaining) {
-		clone = crypt_alloc_buffer(io, remaining);
+		clone = crypt_alloc_buffer(io, remaining, &out_of_pages);
 		if (unlikely(!clone)) {
 			io->error = -ENOMEM;
-			return;
+			break;
 		}
 
 		io->ctx.bio_out = clone;
@@ -693,37 +727,32 @@ static void kcryptd_crypt_write_convert_loop(struct dm_crypt_io *io)
 
 		remaining -= clone->bi_size;
 
+		crypt_inc_pending(io);
 		r = crypt_convert(cc, &io->ctx);
+		crypt_finished = atomic_dec_and_test(&io->ctx.pending);
 
-		if (atomic_dec_and_test(&io->ctx.pending)) {
-			/* processed, no running async crypto  */
+		/* Encryption was already finished, submit io now */
+		if (crypt_finished) {
 			kcryptd_crypt_write_io_submit(io, r, 0);
-			if (unlikely(r < 0))
-				return;
-		} else
-			atomic_inc(&io->pending);
 
-		/* out of memory -> run queues */
-		if (unlikely(remaining)) {
-			/* wait for async crypto then reinitialize pending */
-			wait_event(cc->writeq, !atomic_read(&io->ctx.pending));
-			atomic_set(&io->ctx.pending, 1);
-			congestion_wait(WRITE, HZ/100);
+			/*
+			 * If there was an error, do not try next fragments.
+			 * For async, error is processed in async handler.
+			 */
+			if (unlikely(r < 0))
+				break;
 		}
-	}
-}
 
-static void kcryptd_crypt_write_convert(struct dm_crypt_io *io)
-{
-	struct crypt_config *cc = io->target->private;
-
-	/*
-	 * Prevent io from disappearing until this function completes.
-	 */
-	atomic_inc(&io->pending);
+		/*
+		 * Out of memory -> run queues
+		 * But don't wait if split was due to the io size restriction
+		 */
+		if (unlikely(out_of_pages))
+			congestion_wait(WRITE, HZ/100);
 
-	crypt_convert_init(cc, &io->ctx, NULL, io->base_bio, io->sector);
-	kcryptd_crypt_write_convert_loop(io);
+		if (unlikely(remaining))
+			wait_event(cc->writeq, !atomic_read(&io->ctx.pending));
+	}
 
 	crypt_dec_pending(io);
 }
@@ -741,7 +770,7 @@ static void kcryptd_crypt_read_convert(struct dm_crypt_io *io)
 	struct crypt_config *cc = io->target->private;
 	int r = 0;
 
-	atomic_inc(&io->pending);
+	crypt_inc_pending(io);
 
 	crypt_convert_init(cc, &io->ctx, io->base_bio, io->base_bio,
 			   io->sector);
@@ -1108,15 +1137,9 @@ static void crypt_dtr(struct dm_target *ti)
 static int crypt_map(struct dm_target *ti, struct bio *bio,
 		     union map_info *map_context)
 {
-	struct crypt_config *cc = ti->private;
 	struct dm_crypt_io *io;
 
-	io = mempool_alloc(cc->io_pool, GFP_NOIO);
-	io->target = ti;
-	io->base_bio = bio;
-	io->sector = bio->bi_sector - ti->begin;
-	io->error = 0;
-	atomic_set(&io->pending, 0);
+	io = crypt_io_alloc(ti, bio, bio->bi_sector - ti->begin);
 
 	if (bio_data_dir(io->base_bio) == READ)
 		kcryptd_queue_io(io);
diff --git a/drivers/md/dm-exception-store.c b/drivers/md/dm-exception-store.c
index 41f4080..769ab67 100644
--- a/drivers/md/dm-exception-store.c
+++ b/drivers/md/dm-exception-store.c
@@ -108,12 +108,12 @@ struct pstore {
 	 * Used to keep track of which metadata area the data in
 	 * 'chunk' refers to.
 	 */
-	uint32_t current_area;
+	chunk_t current_area;
 
 	/*
 	 * The next free chunk for an exception.
 	 */
-	uint32_t next_free;
+	chunk_t next_free;
 
 	/*
 	 * The index of next free exception in the current
@@ -175,7 +175,7 @@ static void do_metadata(struct work_struct *work)
 /*
  * Read or write a chunk aligned and sized block of data from a device.
  */
-static int chunk_io(struct pstore *ps, uint32_t chunk, int rw, int metadata)
+static int chunk_io(struct pstore *ps, chunk_t chunk, int rw, int metadata)
 {
 	struct dm_io_region where = {
 		.bdev = ps->snap->cow->bdev,
@@ -209,16 +209,23 @@ static int chunk_io(struct pstore *ps, uint32_t chunk, int rw, int metadata)
 }
 
 /*
+ * Convert a metadata area index to a chunk index.
+ */
+static chunk_t area_location(struct pstore *ps, chunk_t area)
+{
+	return 1 + ((ps->exceptions_per_area + 1) * area);
+}
+
+/*
  * Read or write a metadata area.  Remembering to skip the first
  * chunk which holds the header.
  */
-static int area_io(struct pstore *ps, uint32_t area, int rw)
+static int area_io(struct pstore *ps, chunk_t area, int rw)
 {
 	int r;
-	uint32_t chunk;
+	chunk_t chunk;
 
-	/* convert a metadata area index to a chunk index */
-	chunk = 1 + ((ps->exceptions_per_area + 1) * area);
+	chunk = area_location(ps, area);
 
 	r = chunk_io(ps, chunk, rw, 0);
 	if (r)
@@ -228,7 +235,7 @@ static int area_io(struct pstore *ps, uint32_t area, int rw)
 	return 0;
 }
 
-static int zero_area(struct pstore *ps, uint32_t area)
+static int zero_area(struct pstore *ps, chunk_t area)
 {
 	memset(ps->area, 0, ps->snap->chunk_size << SECTOR_SHIFT);
 	return area_io(ps, area, WRITE);
@@ -404,7 +411,7 @@ static int insert_exceptions(struct pstore *ps, int *full)
 
 static int read_exceptions(struct pstore *ps)
 {
-	uint32_t area;
+	chunk_t area;
 	int r, full = 1;
 
 	/*
@@ -517,6 +524,7 @@ static int persistent_prepare(struct exception_store *store,
 {
 	struct pstore *ps = get_info(store);
 	uint32_t stride;
+	chunk_t next_free;
 	sector_t size = get_dev_size(store->snap->cow->bdev);
 
 	/* Is there enough room ? */
@@ -530,7 +538,8 @@ static int persistent_prepare(struct exception_store *store,
 	 * into account the location of the metadata chunks.
 	 */
 	stride = (ps->exceptions_per_area + 1);
-	if ((++ps->next_free % stride) == 1)
+	next_free = ++ps->next_free;
+	if (sector_div(next_free, stride) == 1)
 		ps->next_free++;
 
 	atomic_inc(&ps->pending_count);
diff --git a/drivers/md/dm-ioctl.c b/drivers/md/dm-ioctl.c
index b262c00..dca401d 100644
--- a/drivers/md/dm-ioctl.c
+++ b/drivers/md/dm-ioctl.c
@@ -426,7 +426,7 @@ static int list_devices(struct dm_ioctl *param, size_t param_size)
 				old_nl->next = (uint32_t) ((void *) nl -
 							   (void *) old_nl);
 			disk = dm_disk(hc->md);
-			nl->dev = huge_encode_dev(MKDEV(disk->major, disk->first_minor));
+			nl->dev = huge_encode_dev(disk_devt(disk));
 			nl->next = 0;
 			strcpy(nl->name, hc->name);
 
@@ -539,7 +539,7 @@ static int __dev_status(struct mapped_device *md, struct dm_ioctl *param)
 	if (dm_suspended(md))
 		param->flags |= DM_SUSPEND_FLAG;
 
-	param->dev = huge_encode_dev(MKDEV(disk->major, disk->first_minor));
+	param->dev = huge_encode_dev(disk_devt(disk));
 
 	/*
 	 * Yes, this will be out of date by the time it gets back
@@ -548,7 +548,7 @@ static int __dev_status(struct mapped_device *md, struct dm_ioctl *param)
 	 */
 	param->open_count = dm_open_count(md);
 
-	if (disk->policy)
+	if (get_disk_ro(disk))
 		param->flags |= DM_READONLY_FLAG;
 
 	param->event_nr = dm_get_event_nr(md);
@@ -1131,7 +1131,7 @@ static void retrieve_deps(struct dm_table *table,
 	unsigned int count = 0;
 	struct list_head *tmp;
 	size_t len, needed;
-	struct dm_dev *dd;
+	struct dm_dev_internal *dd;
 	struct dm_target_deps *deps;
 
 	deps = get_result_buffer(param, param_size, &len);
@@ -1157,7 +1157,7 @@ static void retrieve_deps(struct dm_table *table,
 	deps->count = count;
 	count = 0;
 	list_for_each_entry (dd, dm_table_get_devices(table), list)
-		deps->dev[count++] = huge_encode_dev(dd->bdev->bd_dev);
+		deps->dev[count++] = huge_encode_dev(dd->dm_dev.bdev->bd_dev);
 
 	param->data_size = param->data_start + needed;
 }
diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c
index fea966d..103304c 100644
--- a/drivers/md/dm-mpath.c
+++ b/drivers/md/dm-mpath.c
@@ -30,9 +30,11 @@ struct pgpath {
 	struct list_head list;
 
 	struct priority_group *pg;	/* Owning PG */
+	unsigned is_active;		/* Path status */
 	unsigned fail_count;		/* Cumulative failure count */
 
 	struct dm_path path;
+	struct work_struct deactivate_path;
 };
 
 #define path_to_pgpath(__pgp) container_of((__pgp), struct pgpath, path)
@@ -63,6 +65,7 @@ struct multipath {
 
 	const char *hw_handler_name;
 	struct work_struct activate_path;
+	struct pgpath *pgpath_to_activate;
 	unsigned nr_priority_groups;
 	struct list_head priority_groups;
 	unsigned pg_init_required;	/* pg_init needs calling? */
@@ -111,6 +114,7 @@ static struct workqueue_struct *kmultipathd, *kmpath_handlerd;
 static void process_queued_ios(struct work_struct *work);
 static void trigger_event(struct work_struct *work);
 static void activate_path(struct work_struct *work);
+static void deactivate_path(struct work_struct *work);
 
 
 /*-----------------------------------------------
@@ -121,8 +125,10 @@ static struct pgpath *alloc_pgpath(void)
 {
 	struct pgpath *pgpath = kzalloc(sizeof(*pgpath), GFP_KERNEL);
 
-	if (pgpath)
-		pgpath->path.is_active = 1;
+	if (pgpath) {
+		pgpath->is_active = 1;
+		INIT_WORK(&pgpath->deactivate_path, deactivate_path);
+	}
 
 	return pgpath;
 }
@@ -132,6 +138,14 @@ static void free_pgpath(struct pgpath *pgpath)
 	kfree(pgpath);
 }
 
+static void deactivate_path(struct work_struct *work)
+{
+	struct pgpath *pgpath =
+		container_of(work, struct pgpath, deactivate_path);
+
+	blk_abort_queue(pgpath->path.dev->bdev->bd_disk->queue);
+}
+
 static struct priority_group *alloc_priority_group(void)
 {
 	struct priority_group *pg;
@@ -146,11 +160,19 @@ static struct priority_group *alloc_priority_group(void)
 
 static void free_pgpaths(struct list_head *pgpaths, struct dm_target *ti)
 {
+	unsigned long flags;
 	struct pgpath *pgpath, *tmp;
+	struct multipath *m = ti->private;
 
 	list_for_each_entry_safe(pgpath, tmp, pgpaths, list) {
 		list_del(&pgpath->list);
+		if (m->hw_handler_name)
+			scsi_dh_detach(bdev_get_queue(pgpath->path.dev->bdev));
 		dm_put_device(ti, pgpath->path.dev);
+		spin_lock_irqsave(&m->lock, flags);
+		if (m->pgpath_to_activate == pgpath)
+			m->pgpath_to_activate = NULL;
+		spin_unlock_irqrestore(&m->lock, flags);
 		free_pgpath(pgpath);
 	}
 }
@@ -418,6 +440,7 @@ static void process_queued_ios(struct work_struct *work)
 		__choose_pgpath(m);
 
 	pgpath = m->current_pgpath;
+	m->pgpath_to_activate = m->current_pgpath;
 
 	if ((pgpath && !m->queue_io) ||
 	    (!pgpath && !m->queue_if_no_path))
@@ -548,16 +571,17 @@ static struct pgpath *parse_path(struct arg_set *as, struct path_selector *ps,
 {
 	int r;
 	struct pgpath *p;
+	struct multipath *m = ti->private;
 
 	/* we need at least a path arg */
 	if (as->argc < 1) {
 		ti->error = "no device given";
-		return NULL;
+		return ERR_PTR(-EINVAL);
 	}
 
 	p = alloc_pgpath();
 	if (!p)
-		return NULL;
+		return ERR_PTR(-ENOMEM);
 
 	r = dm_get_device(ti, shift(as), ti->begin, ti->len,
 			  dm_table_get_mode(ti->table), &p->path.dev);
@@ -566,6 +590,15 @@ static struct pgpath *parse_path(struct arg_set *as, struct path_selector *ps,
 		goto bad;
 	}
 
+	if (m->hw_handler_name) {
+		r = scsi_dh_attach(bdev_get_queue(p->path.dev->bdev),
+				   m->hw_handler_name);
+		if (r < 0) {
+			dm_put_device(ti, p->path.dev);
+			goto bad;
+		}
+	}
+
 	r = ps->type->add_path(ps, &p->path, as->argc, as->argv, &ti->error);
 	if (r) {
 		dm_put_device(ti, p->path.dev);
@@ -576,7 +609,7 @@ static struct pgpath *parse_path(struct arg_set *as, struct path_selector *ps,
 
  bad:
 	free_pgpath(p);
-	return NULL;
+	return ERR_PTR(r);
 }
 
 static struct priority_group *parse_priority_group(struct arg_set *as,
@@ -594,14 +627,14 @@ static struct priority_group *parse_priority_group(struct arg_set *as,
 
 	if (as->argc < 2) {
 		as->argc = 0;
-		ti->error = "not enough priority group aruments";
-		return NULL;
+		ti->error = "not enough priority group arguments";
+		return ERR_PTR(-EINVAL);
 	}
 
 	pg = alloc_priority_group();
 	if (!pg) {
 		ti->error = "couldn't allocate priority group";
-		return NULL;
+		return ERR_PTR(-ENOMEM);
 	}
 	pg->m = m;
 
@@ -634,8 +667,10 @@ static struct priority_group *parse_priority_group(struct arg_set *as,
 		path_args.argv = as->argv;
 
 		pgpath = parse_path(&path_args, &pg->ps, ti);
-		if (!pgpath)
+		if (IS_ERR(pgpath)) {
+			r = PTR_ERR(pgpath);
 			goto bad;
+		}
 
 		pgpath->pg = pg;
 		list_add_tail(&pgpath->list, &pg->pgpaths);
@@ -646,7 +681,7 @@ static struct priority_group *parse_priority_group(struct arg_set *as,
 
  bad:
 	free_priority_group(pg, ti);
-	return NULL;
+	return ERR_PTR(r);
 }
 
 static int parse_hw_handler(struct arg_set *as, struct multipath *m)
@@ -765,8 +800,8 @@ static int multipath_ctr(struct dm_target *ti, unsigned int argc,
 		struct priority_group *pg;
 
 		pg = parse_priority_group(&as, m);
-		if (!pg) {
-			r = -EINVAL;
+		if (IS_ERR(pg)) {
+			r = PTR_ERR(pg);
 			goto bad;
 		}
 
@@ -832,13 +867,13 @@ static int fail_path(struct pgpath *pgpath)
 
 	spin_lock_irqsave(&m->lock, flags);
 
-	if (!pgpath->path.is_active)
+	if (!pgpath->is_active)
 		goto out;
 
 	DMWARN("Failing path %s.", pgpath->path.dev->name);
 
 	pgpath->pg->ps.type->fail_path(&pgpath->pg->ps, &pgpath->path);
-	pgpath->path.is_active = 0;
+	pgpath->is_active = 0;
 	pgpath->fail_count++;
 
 	m->nr_valid_paths--;
@@ -850,6 +885,7 @@ static int fail_path(struct pgpath *pgpath)
 		      pgpath->path.dev->name, m->nr_valid_paths);
 
 	queue_work(kmultipathd, &m->trigger_event);
+	queue_work(kmultipathd, &pgpath->deactivate_path);
 
 out:
 	spin_unlock_irqrestore(&m->lock, flags);
@@ -868,7 +904,7 @@ static int reinstate_path(struct pgpath *pgpath)
 
 	spin_lock_irqsave(&m->lock, flags);
 
-	if (pgpath->path.is_active)
+	if (pgpath->is_active)
 		goto out;
 
 	if (!pgpath->pg->ps.type->reinstate_path) {
@@ -882,7 +918,7 @@ static int reinstate_path(struct pgpath *pgpath)
 	if (r)
 		goto out;
 
-	pgpath->path.is_active = 1;
+	pgpath->is_active = 1;
 
 	m->current_pgpath = NULL;
 	if (!m->nr_valid_paths++ && m->queue_size)
@@ -1080,8 +1116,15 @@ static void activate_path(struct work_struct *work)
 	int ret;
 	struct multipath *m =
 		container_of(work, struct multipath, activate_path);
-	struct dm_path *path = &m->current_pgpath->path;
+	struct dm_path *path;
+	unsigned long flags;
 
+	spin_lock_irqsave(&m->lock, flags);
+	path = &m->pgpath_to_activate->path;
+	m->pgpath_to_activate = NULL;
+	spin_unlock_irqrestore(&m->lock, flags);
+	if (!path)
+		return;
 	ret = scsi_dh_activate(bdev_get_queue(path->dev->bdev));
 	pg_init_done(path, ret);
 }
@@ -1263,7 +1306,7 @@ static int multipath_status(struct dm_target *ti, status_type_t type,
 
 			list_for_each_entry(p, &pg->pgpaths, list) {
 				DMEMIT("%s %s %u ", p->path.dev->name,
-				       p->path.is_active ? "A" : "F",
+				       p->is_active ? "A" : "F",
 				       p->fail_count);
 				if (pg->ps.type->status)
 					sz += pg->ps.type->status(&pg->ps,
diff --git a/drivers/md/dm-mpath.h b/drivers/md/dm-mpath.h
index c198b85..e230f71 100644
--- a/drivers/md/dm-mpath.h
+++ b/drivers/md/dm-mpath.h
@@ -13,8 +13,6 @@ struct dm_dev;
 
 struct dm_path {
 	struct dm_dev *dev;	/* Read-only */
-	unsigned is_active;	/* Read-only */
-
 	void *pscontext;	/* For path-selector use */
 };
 
diff --git a/drivers/md/dm-raid1.c b/drivers/md/dm-raid1.c
index ff05fe8..29913e4 100644
--- a/drivers/md/dm-raid1.c
+++ b/drivers/md/dm-raid1.c
@@ -842,7 +842,9 @@ static int recover(struct mirror_set *ms, struct region *reg)
 	}
 
 	/* hand to kcopyd */
-	set_bit(DM_KCOPYD_IGNORE_ERROR, &flags);
+	if (!errors_handled(ms))
+		set_bit(DM_KCOPYD_IGNORE_ERROR, &flags);
+
 	r = dm_kcopyd_copy(ms->kcopyd_client, &from, ms->nr_mirrors - 1, to,
 			   flags, recovery_complete, reg);
 
diff --git a/drivers/md/dm-stripe.c b/drivers/md/dm-stripe.c
index 4de90ab..b745d8a 100644
--- a/drivers/md/dm-stripe.c
+++ b/drivers/md/dm-stripe.c
@@ -284,8 +284,8 @@ static int stripe_end_io(struct dm_target *ti, struct bio *bio,
 
 	memset(major_minor, 0, sizeof(major_minor));
 	sprintf(major_minor, "%d:%d",
-		bio->bi_bdev->bd_disk->major,
-		bio->bi_bdev->bd_disk->first_minor);
+		MAJOR(disk_devt(bio->bi_bdev->bd_disk)),
+		MINOR(disk_devt(bio->bi_bdev->bd_disk)));
 
 	/*
 	 * Test to see which stripe drive triggered the event
diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
index 798e468..a740a69 100644
--- a/drivers/md/dm-table.c
+++ b/drivers/md/dm-table.c
@@ -250,7 +250,8 @@ static void free_devices(struct list_head *devices)
 	struct list_head *tmp, *next;
 
 	list_for_each_safe(tmp, next, devices) {
-		struct dm_dev *dd = list_entry(tmp, struct dm_dev, list);
+		struct dm_dev_internal *dd =
+		    list_entry(tmp, struct dm_dev_internal, list);
 		kfree(dd);
 	}
 }
@@ -316,40 +317,23 @@ static inline int check_space(struct dm_table *t)
  */
 static int lookup_device(const char *path, dev_t *dev)
 {
-	int r;
-	struct nameidata nd;
-	struct inode *inode;
-
-	if ((r = path_lookup(path, LOOKUP_FOLLOW, &nd)))
-		return r;
-
-	inode = nd.path.dentry->d_inode;
-	if (!inode) {
-		r = -ENOENT;
-		goto out;
-	}
-
-	if (!S_ISBLK(inode->i_mode)) {
-		r = -ENOTBLK;
-		goto out;
-	}
-
-	*dev = inode->i_rdev;
-
- out:
-	path_put(&nd.path);
-	return r;
+	struct block_device *bdev = lookup_bdev(path);
+	if (IS_ERR(bdev))
+		return PTR_ERR(bdev);
+	*dev = bdev->bd_dev;
+	bdput(bdev);
+	return 0;
 }
 
 /*
  * See if we've already got a device in the list.
  */
-static struct dm_dev *find_device(struct list_head *l, dev_t dev)
+static struct dm_dev_internal *find_device(struct list_head *l, dev_t dev)
 {
-	struct dm_dev *dd;
+	struct dm_dev_internal *dd;
 
 	list_for_each_entry (dd, l, list)
-		if (dd->bdev->bd_dev == dev)
+		if (dd->dm_dev.bdev->bd_dev == dev)
 			return dd;
 
 	return NULL;
@@ -358,45 +342,47 @@ static struct dm_dev *find_device(struct list_head *l, dev_t dev)
 /*
  * Open a device so we can use it as a map destination.
  */
-static int open_dev(struct dm_dev *d, dev_t dev, struct mapped_device *md)
+static int open_dev(struct dm_dev_internal *d, dev_t dev,
+		    struct mapped_device *md)
 {
 	static char *_claim_ptr = "I belong to device-mapper";
 	struct block_device *bdev;
 
 	int r;
 
-	BUG_ON(d->bdev);
+	BUG_ON(d->dm_dev.bdev);
 
-	bdev = open_by_devnum(dev, d->mode);
+	bdev = open_by_devnum(dev, d->dm_dev.mode);
 	if (IS_ERR(bdev))
 		return PTR_ERR(bdev);
 	r = bd_claim_by_disk(bdev, _claim_ptr, dm_disk(md));
 	if (r)
 		blkdev_put(bdev);
 	else
-		d->bdev = bdev;
+		d->dm_dev.bdev = bdev;
 	return r;
 }
 
 /*
  * Close a device that we've been using.
  */
-static void close_dev(struct dm_dev *d, struct mapped_device *md)
+static void close_dev(struct dm_dev_internal *d, struct mapped_device *md)
 {
-	if (!d->bdev)
+	if (!d->dm_dev.bdev)
 		return;
 
-	bd_release_from_disk(d->bdev, dm_disk(md));
-	blkdev_put(d->bdev);
-	d->bdev = NULL;
+	bd_release_from_disk(d->dm_dev.bdev, dm_disk(md));
+	blkdev_put(d->dm_dev.bdev);
+	d->dm_dev.bdev = NULL;
 }
 
 /*
  * If possible, this checks an area of a destination device is valid.
  */
-static int check_device_area(struct dm_dev *dd, sector_t start, sector_t len)
+static int check_device_area(struct dm_dev_internal *dd, sector_t start,
+			     sector_t len)
 {
-	sector_t dev_size = dd->bdev->bd_inode->i_size >> SECTOR_SHIFT;
+	sector_t dev_size = dd->dm_dev.bdev->bd_inode->i_size >> SECTOR_SHIFT;
 
 	if (!dev_size)
 		return 1;
@@ -409,16 +395,17 @@ static int check_device_area(struct dm_dev *dd, sector_t start, sector_t len)
  * careful to leave things as they were if we fail to reopen the
  * device.
  */
-static int upgrade_mode(struct dm_dev *dd, int new_mode, struct mapped_device *md)
+static int upgrade_mode(struct dm_dev_internal *dd, int new_mode,
+			struct mapped_device *md)
 {
 	int r;
-	struct dm_dev dd_copy;
-	dev_t dev = dd->bdev->bd_dev;
+	struct dm_dev_internal dd_copy;
+	dev_t dev = dd->dm_dev.bdev->bd_dev;
 
 	dd_copy = *dd;
 
-	dd->mode |= new_mode;
-	dd->bdev = NULL;
+	dd->dm_dev.mode |= new_mode;
+	dd->dm_dev.bdev = NULL;
 	r = open_dev(dd, dev, md);
 	if (!r)
 		close_dev(&dd_copy, md);
@@ -438,7 +425,7 @@ static int __table_get_device(struct dm_table *t, struct dm_target *ti,
 {
 	int r;
 	dev_t uninitialized_var(dev);
-	struct dm_dev *dd;
+	struct dm_dev_internal *dd;
 	unsigned int major, minor;
 
 	BUG_ON(!t);
@@ -460,20 +447,20 @@ static int __table_get_device(struct dm_table *t, struct dm_target *ti,
 		if (!dd)
 			return -ENOMEM;
 
-		dd->mode = mode;
-		dd->bdev = NULL;
+		dd->dm_dev.mode = mode;
+		dd->dm_dev.bdev = NULL;
 
 		if ((r = open_dev(dd, dev, t->md))) {
 			kfree(dd);
 			return r;
 		}
 
-		format_dev_t(dd->name, dev);
+		format_dev_t(dd->dm_dev.name, dev);
 
 		atomic_set(&dd->count, 0);
 		list_add(&dd->list, &t->devices);
 
-	} else if (dd->mode != (mode | dd->mode)) {
+	} else if (dd->dm_dev.mode != (mode | dd->dm_dev.mode)) {
 		r = upgrade_mode(dd, mode, t->md);
 		if (r)
 			return r;
@@ -482,11 +469,11 @@ static int __table_get_device(struct dm_table *t, struct dm_target *ti,
 
 	if (!check_device_area(dd, start, len)) {
 		DMWARN("device %s too small for target", path);
-		dm_put_device(ti, dd);
+		dm_put_device(ti, &dd->dm_dev);
 		return -EINVAL;
 	}
 
-	*result = dd;
+	*result = &dd->dm_dev;
 
 	return 0;
 }
@@ -495,6 +482,13 @@ void dm_set_device_limits(struct dm_target *ti, struct block_device *bdev)
 {
 	struct request_queue *q = bdev_get_queue(bdev);
 	struct io_restrictions *rs = &ti->limits;
+	char b[BDEVNAME_SIZE];
+
+	if (unlikely(!q)) {
+		DMWARN("%s: Cannot set limits for nonexistent device %s",
+		       dm_device_name(ti->table->md), bdevname(bdev, b));
+		return;
+	}
 
 	/*
 	 * Combine the device limits low.
@@ -557,8 +551,11 @@ int dm_get_device(struct dm_target *ti, const char *path, sector_t start,
 /*
  * Decrement a devices use count and remove it if necessary.
  */
-void dm_put_device(struct dm_target *ti, struct dm_dev *dd)
+void dm_put_device(struct dm_target *ti, struct dm_dev *d)
 {
+	struct dm_dev_internal *dd = container_of(d, struct dm_dev_internal,
+						  dm_dev);
+
 	if (atomic_dec_and_test(&dd->count)) {
 		close_dev(dd, ti->table->md);
 		list_del(&dd->list);
@@ -954,13 +951,20 @@ int dm_table_resume_targets(struct dm_table *t)
 
 int dm_table_any_congested(struct dm_table *t, int bdi_bits)
 {
-	struct dm_dev *dd;
+	struct dm_dev_internal *dd;
 	struct list_head *devices = dm_table_get_devices(t);
 	int r = 0;
 
 	list_for_each_entry(dd, devices, list) {
-		struct request_queue *q = bdev_get_queue(dd->bdev);
-		r |= bdi_congested(&q->backing_dev_info, bdi_bits);
+		struct request_queue *q = bdev_get_queue(dd->dm_dev.bdev);
+		char b[BDEVNAME_SIZE];
+
+		if (likely(q))
+			r |= bdi_congested(&q->backing_dev_info, bdi_bits);
+		else
+			DMWARN_LIMIT("%s: any_congested: nonexistent device %s",
+				     dm_device_name(t->md),
+				     bdevname(dd->dm_dev.bdev, b));
 	}
 
 	return r;
@@ -968,13 +972,19 @@ int dm_table_any_congested(struct dm_table *t, int bdi_bits)
 
 void dm_table_unplug_all(struct dm_table *t)
 {
-	struct dm_dev *dd;
+	struct dm_dev_internal *dd;
 	struct list_head *devices = dm_table_get_devices(t);
 
 	list_for_each_entry(dd, devices, list) {
-		struct request_queue *q = bdev_get_queue(dd->bdev);
-
-		blk_unplug(q);
+		struct request_queue *q = bdev_get_queue(dd->dm_dev.bdev);
+		char b[BDEVNAME_SIZE];
+
+		if (likely(q))
+			blk_unplug(q);
+		else
+			DMWARN_LIMIT("%s: Cannot unplug nonexistent device %s",
+				     dm_device_name(t->md),
+				     bdevname(dd->dm_dev.bdev, b));
 	}
 }
 
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index bca448e..327de03 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -377,13 +377,14 @@ static void free_tio(struct mapped_device *md, struct dm_target_io *tio)
 static void start_io_acct(struct dm_io *io)
 {
 	struct mapped_device *md = io->md;
+	int cpu;
 
 	io->start_time = jiffies;
 
-	preempt_disable();
-	disk_round_stats(dm_disk(md));
-	preempt_enable();
-	dm_disk(md)->in_flight = atomic_inc_return(&md->pending);
+	cpu = part_stat_lock();
+	part_round_stats(cpu, &dm_disk(md)->part0);
+	part_stat_unlock();
+	dm_disk(md)->part0.in_flight = atomic_inc_return(&md->pending);
 }
 
 static int end_io_acct(struct dm_io *io)
@@ -391,15 +392,16 @@ static int end_io_acct(struct dm_io *io)
 	struct mapped_device *md = io->md;
 	struct bio *bio = io->bio;
 	unsigned long duration = jiffies - io->start_time;
-	int pending;
+	int pending, cpu;
 	int rw = bio_data_dir(bio);
 
-	preempt_disable();
-	disk_round_stats(dm_disk(md));
-	preempt_enable();
-	dm_disk(md)->in_flight = pending = atomic_dec_return(&md->pending);
+	cpu = part_stat_lock();
+	part_round_stats(cpu, &dm_disk(md)->part0);
+	part_stat_add(cpu, &dm_disk(md)->part0, ticks[rw], duration);
+	part_stat_unlock();
 
-	disk_stat_add(dm_disk(md), ticks[rw], duration);
+	dm_disk(md)->part0.in_flight = pending =
+		atomic_dec_return(&md->pending);
 
 	return !pending;
 }
@@ -837,12 +839,14 @@ static int dm_merge_bvec(struct request_queue *q,
 	struct dm_table *map = dm_get_table(md);
 	struct dm_target *ti;
 	sector_t max_sectors;
-	int max_size;
+	int max_size = 0;
 
 	if (unlikely(!map))
-		return 0;
+		goto out;
 
 	ti = dm_table_find_target(map, bvm->bi_sector);
+	if (!dm_target_is_valid(ti))
+		goto out_table;
 
 	/*
 	 * Find maximum amount of I/O that won't need splitting
@@ -861,14 +865,16 @@ static int dm_merge_bvec(struct request_queue *q,
 	if (max_size && ti->type->merge)
 		max_size = ti->type->merge(ti, bvm, biovec, max_size);
 
+out_table:
+	dm_table_put(map);
+
+out:
 	/*
 	 * Always allow an entire first page
 	 */
 	if (max_size <= biovec->bv_len && !(bvm->bi_size >> SECTOR_SHIFT))
 		max_size = biovec->bv_len;
 
-	dm_table_put(map);
-
 	return max_size;
 }
 
@@ -881,6 +887,7 @@ static int dm_request(struct request_queue *q, struct bio *bio)
 	int r = -EIO;
 	int rw = bio_data_dir(bio);
 	struct mapped_device *md = q->queuedata;
+	int cpu;
 
 	/*
 	 * There is no use in forwarding any barrier request since we can't
@@ -893,8 +900,10 @@ static int dm_request(struct request_queue *q, struct bio *bio)
 
 	down_read(&md->io_lock);
 
-	disk_stat_inc(dm_disk(md), ios[rw]);
-	disk_stat_add(dm_disk(md), sectors[rw], bio_sectors(bio));
+	cpu = part_stat_lock();
+	part_stat_inc(cpu, &dm_disk(md)->part0, ios[rw]);
+	part_stat_add(cpu, &dm_disk(md)->part0, sectors[rw], bio_sectors(bio));
+	part_stat_unlock();
 
 	/*
 	 * If we're suspended we have to queue
@@ -1142,7 +1151,7 @@ static void unlock_fs(struct mapped_device *md);
 
 static void free_dev(struct mapped_device *md)
 {
-	int minor = md->disk->first_minor;
+	int minor = MINOR(disk_devt(md->disk));
 
 	if (md->suspended_bdev) {
 		unlock_fs(md);
@@ -1178,7 +1187,7 @@ static void event_callback(void *context)
 	list_splice_init(&md->uevent_list, &uevents);
 	spin_unlock_irqrestore(&md->uevent_lock, flags);
 
-	dm_send_uevents(&uevents, &md->disk->dev.kobj);
+	dm_send_uevents(&uevents, &disk_to_dev(md->disk)->kobj);
 
 	atomic_inc(&md->event_nr);
 	wake_up(&md->eventq);
@@ -1263,7 +1272,7 @@ static struct mapped_device *dm_find_md(dev_t dev)
 
 	md = idr_find(&_minor_idr, minor);
 	if (md && (md == MINOR_ALLOCED ||
-		   (dm_disk(md)->first_minor != minor) ||
+		   (MINOR(disk_devt(dm_disk(md))) != minor) ||
 		   test_bit(DMF_FREEING, &md->flags))) {
 		md = NULL;
 		goto out;
@@ -1314,7 +1323,8 @@ void dm_put(struct mapped_device *md)
 
 	if (atomic_dec_and_lock(&md->holders, &_minor_lock)) {
 		map = dm_get_table(md);
-		idr_replace(&_minor_idr, MINOR_ALLOCED, dm_disk(md)->first_minor);
+		idr_replace(&_minor_idr, MINOR_ALLOCED,
+			    MINOR(disk_devt(dm_disk(md))));
 		set_bit(DMF_FREEING, &md->flags);
 		spin_unlock(&_minor_lock);
 		if (!dm_suspended(md)) {
@@ -1634,7 +1644,7 @@ out:
  *---------------------------------------------------------------*/
 void dm_kobject_uevent(struct mapped_device *md)
 {
-	kobject_uevent(&md->disk->dev.kobj, KOBJ_CHANGE);
+	kobject_uevent(&disk_to_dev(md->disk)->kobj, KOBJ_CHANGE);
 }
 
 uint32_t dm_next_uevent_seq(struct mapped_device *md)
diff --git a/drivers/md/dm.h b/drivers/md/dm.h
index 1e59a0b..cd189da 100644
--- a/drivers/md/dm.h
+++ b/drivers/md/dm.h
@@ -25,13 +25,10 @@
 /*
  * List of devices that a metadevice uses and should open/close.
  */
-struct dm_dev {
+struct dm_dev_internal {
 	struct list_head list;
-
 	atomic_t count;
-	int mode;
-	struct block_device *bdev;
-	char name[16];
+	struct dm_dev dm_dev;
 };
 
 struct dm_table;
@@ -49,7 +46,6 @@ void dm_table_presuspend_targets(struct dm_table *t);
 void dm_table_postsuspend_targets(struct dm_table *t);
 int dm_table_resume_targets(struct dm_table *t);
 int dm_table_any_congested(struct dm_table *t, int bdi_bits);
-void dm_table_unplug_all(struct dm_table *t);
 
 /*
  * To check the return value from dm_table_find_target().
@@ -93,8 +89,6 @@ void dm_linear_exit(void);
 int dm_stripe_init(void);
 void dm_stripe_exit(void);
 
-void *dm_vcalloc(unsigned long nmemb, unsigned long elem_size);
-union map_info *dm_get_mapinfo(struct bio *bio);
 int dm_open_count(struct mapped_device *md);
 int dm_lock_for_deletion(struct mapped_device *md);
 
diff --git a/drivers/md/linear.c b/drivers/md/linear.c
index b1eebf8..b9cbee68 100644
--- a/drivers/md/linear.c
+++ b/drivers/md/linear.c
@@ -318,14 +318,18 @@ static int linear_make_request (struct request_queue *q, struct bio *bio)
 	mddev_t *mddev = q->queuedata;
 	dev_info_t *tmp_dev;
 	sector_t block;
+	int cpu;
 
 	if (unlikely(bio_barrier(bio))) {
 		bio_endio(bio, -EOPNOTSUPP);
 		return 0;
 	}
 
-	disk_stat_inc(mddev->gendisk, ios[rw]);
-	disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bio));
+	cpu = part_stat_lock();
+	part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]);
+	part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw],
+		      bio_sectors(bio));
+	part_stat_unlock();
 
 	tmp_dev = which_dev(mddev, bio->bi_sector);
 	block = bio->bi_sector >> 1;
@@ -349,7 +353,7 @@ static int linear_make_request (struct request_queue *q, struct bio *bio)
 		 * split it.
 		 */
 		struct bio_pair *bp;
-		bp = bio_split(bio, bio_split_pool,
+		bp = bio_split(bio,
 			       ((tmp_dev->offset + tmp_dev->size)<<1) - bio->bi_sector);
 		if (linear_make_request(q, &bp->bio1))
 			generic_make_request(&bp->bio1);
diff --git a/drivers/md/md.c b/drivers/md/md.c
index c2ff77c..0a3a4bd 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -1464,10 +1464,7 @@ static int bind_rdev_to_array(mdk_rdev_t * rdev, mddev_t * mddev)
 	if ((err = kobject_add(&rdev->kobj, &mddev->kobj, "dev-%s", b)))
 		goto fail;
 
-	if (rdev->bdev->bd_part)
-		ko = &rdev->bdev->bd_part->dev.kobj;
-	else
-		ko = &rdev->bdev->bd_disk->dev.kobj;
+	ko = &part_to_dev(rdev->bdev->bd_part)->kobj;
 	if ((err = sysfs_create_link(&rdev->kobj, ko, "block"))) {
 		kobject_del(&rdev->kobj);
 		goto fail;
@@ -2393,6 +2390,8 @@ static void analyze_sbs(mddev_t * mddev)
 
 }
 
+static void md_safemode_timeout(unsigned long data);
+
 static ssize_t
 safe_delay_show(mddev_t *mddev, char *page)
 {
@@ -2432,9 +2431,12 @@ safe_delay_store(mddev_t *mddev, const char *cbuf, size_t len)
 	if (msec == 0)
 		mddev->safemode_delay = 0;
 	else {
+		unsigned long old_delay = mddev->safemode_delay;
 		mddev->safemode_delay = (msec*HZ)/1000;
 		if (mddev->safemode_delay == 0)
 			mddev->safemode_delay = 1;
+		if (mddev->safemode_delay < old_delay)
+			md_safemode_timeout((unsigned long)mddev);
 	}
 	return len;
 }
@@ -3465,8 +3467,8 @@ static struct kobject *md_probe(dev_t dev, int *part, void *data)
 	disk->queue = mddev->queue;
 	add_disk(disk);
 	mddev->gendisk = disk;
-	error = kobject_init_and_add(&mddev->kobj, &md_ktype, &disk->dev.kobj,
-				     "%s", "md");
+	error = kobject_init_and_add(&mddev->kobj, &md_ktype,
+				     &disk_to_dev(disk)->kobj, "%s", "md");
 	mutex_unlock(&disks_mutex);
 	if (error)
 		printk(KERN_WARNING "md: cannot register %s/md - name in use\n",
@@ -3483,7 +3485,7 @@ static void md_safemode_timeout(unsigned long data)
 	if (!atomic_read(&mddev->writes_pending)) {
 		mddev->safemode = 1;
 		if (mddev->external)
-			sysfs_notify(&mddev->kobj, NULL, "array_state");
+			set_bit(MD_NOTIFY_ARRAY_STATE, &mddev->flags);
 	}
 	md_wakeup_thread(mddev->thread);
 }
@@ -3756,7 +3758,7 @@ static int do_md_run(mddev_t * mddev)
 	sysfs_notify(&mddev->kobj, NULL, "array_state");
 	sysfs_notify(&mddev->kobj, NULL, "sync_action");
 	sysfs_notify(&mddev->kobj, NULL, "degraded");
-	kobject_uevent(&mddev->gendisk->dev.kobj, KOBJ_CHANGE);
+	kobject_uevent(&disk_to_dev(mddev->gendisk)->kobj, KOBJ_CHANGE);
 	return 0;
 }
 
@@ -3836,8 +3838,6 @@ static int do_md_stop(mddev_t * mddev, int mode, int is_open)
 
 		del_timer_sync(&mddev->safemode_timer);
 
-		invalidate_partition(disk, 0);
-
 		switch(mode) {
 		case 1: /* readonly */
 			err  = -ENXIO;
@@ -4634,6 +4634,11 @@ static int update_size(mddev_t *mddev, sector_t num_sectors)
 	 */
 	if (mddev->sync_thread)
 		return -EBUSY;
+	if (mddev->bitmap)
+		/* Sorry, cannot grow a bitmap yet, just remove it,
+		 * grow, and re-add.
+		 */
+		return -EBUSY;
 	rdev_for_each(rdev, tmp, mddev) {
 		sector_t avail;
 		avail = rdev->size * 2;
@@ -5541,8 +5546,8 @@ static int is_mddev_idle(mddev_t *mddev)
 	rcu_read_lock();
 	rdev_for_each_rcu(rdev, mddev) {
 		struct gendisk *disk = rdev->bdev->bd_contains->bd_disk;
-		curr_events = disk_stat_read(disk, sectors[0]) + 
-				disk_stat_read(disk, sectors[1]) - 
+		curr_events = part_stat_read(&disk->part0, sectors[0]) +
+				part_stat_read(&disk->part0, sectors[1]) -
 				atomic_read(&disk->sync_io);
 		/* sync IO will cause sync_io to increase before the disk_stats
 		 * as sync_io is counted when a request starts, and
@@ -5753,7 +5758,11 @@ void md_do_sync(mddev_t *mddev)
 					 * time 'round when curr_resync == 2
 					 */
 					continue;
-				prepare_to_wait(&resync_wait, &wq, TASK_UNINTERRUPTIBLE);
+				/* We need to wait 'interruptible' so as not to
+				 * contribute to the load average, and not to
+				 * be caught by 'softlockup'
+				 */
+				prepare_to_wait(&resync_wait, &wq, TASK_INTERRUPTIBLE);
 				if (!kthread_should_stop() &&
 				    mddev2->curr_resync >= mddev->curr_resync) {
 					printk(KERN_INFO "md: delaying %s of %s"
@@ -5761,6 +5770,8 @@ void md_do_sync(mddev_t *mddev)
 					       " share one or more physical units)\n",
 					       desc, mdname(mddev), mdname(mddev2));
 					mddev_put(mddev2);
+					if (signal_pending(current))
+						flush_signals(current);
 					schedule();
 					finish_wait(&resync_wait, &wq);
 					goto try_again;
@@ -5993,10 +6004,11 @@ static int remove_and_add_spares(mddev_t *mddev)
 			}
 		}
 
-	if (mddev->degraded) {
+	if (mddev->degraded && ! mddev->ro) {
 		rdev_for_each(rdev, rtmp, mddev) {
 			if (rdev->raid_disk >= 0 &&
-			    !test_bit(In_sync, &rdev->flags))
+			    !test_bit(In_sync, &rdev->flags) &&
+			    !test_bit(Blocked, &rdev->flags))
 				spares++;
 			if (rdev->raid_disk < 0
 			    && !test_bit(Faulty, &rdev->flags)) {
@@ -6051,6 +6063,9 @@ void md_check_recovery(mddev_t *mddev)
 	if (mddev->bitmap)
 		bitmap_daemon_work(mddev->bitmap);
 
+	if (test_and_clear_bit(MD_NOTIFY_ARRAY_STATE, &mddev->flags))
+		sysfs_notify(&mddev->kobj, NULL, "array_state");
+
 	if (mddev->ro)
 		return;
 
@@ -6063,6 +6078,8 @@ void md_check_recovery(mddev_t *mddev)
 		flush_signals(current);
 	}
 
+	if (mddev->ro && !test_bit(MD_RECOVERY_NEEDED, &mddev->recovery))
+		return;
 	if ( ! (
 		(mddev->flags && !mddev->external) ||
 		test_bit(MD_RECOVERY_NEEDED, &mddev->recovery) ||
@@ -6076,6 +6093,15 @@ void md_check_recovery(mddev_t *mddev)
 	if (mddev_trylock(mddev)) {
 		int spares = 0;
 
+		if (mddev->ro) {
+			/* Only thing we do on a ro array is remove
+			 * failed devices.
+			 */
+			remove_and_add_spares(mddev);
+			clear_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
+			goto unlock;
+		}
+
 		if (!mddev->external) {
 			int did_change = 0;
 			spin_lock_irq(&mddev->write_lock);
@@ -6113,7 +6139,8 @@ void md_check_recovery(mddev_t *mddev)
 			/* resync has finished, collect result */
 			md_unregister_thread(mddev->sync_thread);
 			mddev->sync_thread = NULL;
-			if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery)) {
+			if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery) &&
+			    !test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery)) {
 				/* success...*/
 				/* activate any spares */
 				if (mddev->pers->spare_active(mddev))
@@ -6165,6 +6192,7 @@ void md_check_recovery(mddev_t *mddev)
 		} else if ((spares = remove_and_add_spares(mddev))) {
 			clear_bit(MD_RECOVERY_SYNC, &mddev->recovery);
 			clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
+			clear_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
 			set_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
 		} else if (mddev->recovery_cp < MaxSector) {
 			set_bit(MD_RECOVERY_SYNC, &mddev->recovery);
@@ -6228,7 +6256,11 @@ static int md_notify_reboot(struct notifier_block *this,
 
 		for_each_mddev(mddev, tmp)
 			if (mddev_trylock(mddev)) {
-				do_md_stop (mddev, 1, 0);
+				/* Force a switch to readonly even array
+				 * appears to still be in use.  Hence
+				 * the '100'.
+				 */
+				do_md_stop (mddev, 1, 100);
 				mddev_unlock(mddev);
 			}
 		/*
diff --git a/drivers/md/multipath.c b/drivers/md/multipath.c
index c4779cc..8bb8794 100644
--- a/drivers/md/multipath.c
+++ b/drivers/md/multipath.c
@@ -147,6 +147,7 @@ static int multipath_make_request (struct request_queue *q, struct bio * bio)
 	struct multipath_bh * mp_bh;
 	struct multipath_info *multipath;
 	const int rw = bio_data_dir(bio);
+	int cpu;
 
 	if (unlikely(bio_barrier(bio))) {
 		bio_endio(bio, -EOPNOTSUPP);
@@ -158,8 +159,11 @@ static int multipath_make_request (struct request_queue *q, struct bio * bio)
 	mp_bh->master_bio = bio;
 	mp_bh->mddev = mddev;
 
-	disk_stat_inc(mddev->gendisk, ios[rw]);
-	disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bio));
+	cpu = part_stat_lock();
+	part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]);
+	part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw],
+		      bio_sectors(bio));
+	part_stat_unlock();
 
 	mp_bh->path = multipath_map(conf);
 	if (mp_bh->path < 0) {
diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c
index 1836106..53508a8 100644
--- a/drivers/md/raid0.c
+++ b/drivers/md/raid0.c
@@ -399,14 +399,18 @@ static int raid0_make_request (struct request_queue *q, struct bio *bio)
 	sector_t chunk;
 	sector_t block, rsect;
 	const int rw = bio_data_dir(bio);
+	int cpu;
 
 	if (unlikely(bio_barrier(bio))) {
 		bio_endio(bio, -EOPNOTSUPP);
 		return 0;
 	}
 
-	disk_stat_inc(mddev->gendisk, ios[rw]);
-	disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bio));
+	cpu = part_stat_lock();
+	part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]);
+	part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw],
+		      bio_sectors(bio));
+	part_stat_unlock();
 
 	chunk_size = mddev->chunk_size >> 10;
 	chunk_sects = mddev->chunk_size >> 9;
@@ -423,7 +427,7 @@ static int raid0_make_request (struct request_queue *q, struct bio *bio)
 		/* This is a one page bio that upper layers
 		 * refuse to split for us, so we need to split it.
 		 */
-		bp = bio_split(bio, bio_split_pool, chunk_sects - (bio->bi_sector & (chunk_sects - 1)) );
+		bp = bio_split(bio, chunk_sects - (bio->bi_sector & (chunk_sects - 1)));
 		if (raid0_make_request(q, &bp->bio1))
 			generic_make_request(&bp->bio1);
 		if (raid0_make_request(q, &bp->bio2))
diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index 03a5ab7..b976442 100644
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -779,7 +779,7 @@ static int make_request(struct request_queue *q, struct bio * bio)
 	struct page **behind_pages = NULL;
 	const int rw = bio_data_dir(bio);
 	const int do_sync = bio_sync(bio);
-	int do_barriers;
+	int cpu, do_barriers;
 	mdk_rdev_t *blocked_rdev;
 
 	/*
@@ -804,8 +804,11 @@ static int make_request(struct request_queue *q, struct bio * bio)
 
 	bitmap = mddev->bitmap;
 
-	disk_stat_inc(mddev->gendisk, ios[rw]);
-	disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bio));
+	cpu = part_stat_lock();
+	part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]);
+	part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw],
+		      bio_sectors(bio));
+	part_stat_unlock();
 
 	/*
 	 * make_request() can abort the operation when READA is being
@@ -1302,9 +1305,6 @@ static void sync_request_write(mddev_t *mddev, r1bio_t *r1_bio)
 					sbio->bi_size = r1_bio->sectors << 9;
 					sbio->bi_idx = 0;
 					sbio->bi_phys_segments = 0;
-					sbio->bi_hw_segments = 0;
-					sbio->bi_hw_front_size = 0;
-					sbio->bi_hw_back_size = 0;
 					sbio->bi_flags &= ~(BIO_POOL_MASK - 1);
 					sbio->bi_flags |= 1 << BIO_UPTODATE;
 					sbio->bi_next = NULL;
@@ -1790,7 +1790,6 @@ static sector_t sync_request(mddev_t *mddev, sector_t sector_nr, int *skipped, i
 		bio->bi_vcnt = 0;
 		bio->bi_idx = 0;
 		bio->bi_phys_segments = 0;
-		bio->bi_hw_segments = 0;
 		bio->bi_size = 0;
 		bio->bi_end_io = NULL;
 		bio->bi_private = NULL;
diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index 159535d..8bdc9bf 100644
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -76,11 +76,13 @@ static void r10bio_pool_free(void *r10_bio, void *data)
 	kfree(r10_bio);
 }
 
+/* Maximum size of each resync request */
 #define RESYNC_BLOCK_SIZE (64*1024)
-//#define RESYNC_BLOCK_SIZE PAGE_SIZE
-#define RESYNC_SECTORS (RESYNC_BLOCK_SIZE >> 9)
 #define RESYNC_PAGES ((RESYNC_BLOCK_SIZE + PAGE_SIZE-1) / PAGE_SIZE)
-#define RESYNC_WINDOW (2048*1024)
+/* amount of memory to reserve for resync requests */
+#define RESYNC_WINDOW (1024*1024)
+/* maximum number of concurrent requests, memory permitting */
+#define RESYNC_DEPTH (32*1024*1024/RESYNC_BLOCK_SIZE)
 
 /*
  * When performing a resync, we need to read and compare, so
@@ -215,6 +217,9 @@ static void reschedule_retry(r10bio_t *r10_bio)
 	conf->nr_queued ++;
 	spin_unlock_irqrestore(&conf->device_lock, flags);
 
+	/* wake up frozen array... */
+	wake_up(&conf->wait_barrier);
+
 	md_wakeup_thread(mddev->thread);
 }
 
@@ -687,7 +692,6 @@ static int flush_pending_writes(conf_t *conf)
  *    there is no normal IO happeing.  It must arrange to call
  *    lower_barrier when the particular background IO completes.
  */
-#define RESYNC_DEPTH 32
 
 static void raise_barrier(conf_t *conf, int force)
 {
@@ -785,6 +789,7 @@ static int make_request(struct request_queue *q, struct bio * bio)
 	mirror_info_t *mirror;
 	r10bio_t *r10_bio;
 	struct bio *read_bio;
+	int cpu;
 	int i;
 	int chunk_sects = conf->chunk_mask + 1;
 	const int rw = bio_data_dir(bio);
@@ -812,7 +817,7 @@ static int make_request(struct request_queue *q, struct bio * bio)
 		/* This is a one page bio that upper layers
 		 * refuse to split for us, so we need to split it.
 		 */
-		bp = bio_split(bio, bio_split_pool,
+		bp = bio_split(bio,
 			       chunk_sects - (bio->bi_sector & (chunk_sects - 1)) );
 		if (make_request(q, &bp->bio1))
 			generic_make_request(&bp->bio1);
@@ -839,8 +844,11 @@ static int make_request(struct request_queue *q, struct bio * bio)
 	 */
 	wait_barrier(conf);
 
-	disk_stat_inc(mddev->gendisk, ios[rw]);
-	disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bio));
+	cpu = part_stat_lock();
+	part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]);
+	part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw],
+		      bio_sectors(bio));
+	part_stat_unlock();
 
 	r10_bio = mempool_alloc(conf->r10bio_pool, GFP_NOIO);
 
@@ -1341,9 +1349,6 @@ static void sync_request_write(mddev_t *mddev, r10bio_t *r10_bio)
 		tbio->bi_size = r10_bio->sectors << 9;
 		tbio->bi_idx = 0;
 		tbio->bi_phys_segments = 0;
-		tbio->bi_hw_segments = 0;
-		tbio->bi_hw_front_size = 0;
-		tbio->bi_hw_back_size = 0;
 		tbio->bi_flags &= ~(BIO_POOL_MASK - 1);
 		tbio->bi_flags |= 1 << BIO_UPTODATE;
 		tbio->bi_next = NULL;
@@ -1943,7 +1948,6 @@ static sector_t sync_request(mddev_t *mddev, sector_t sector_nr, int *skipped, i
 		bio->bi_vcnt = 0;
 		bio->bi_idx = 0;
 		bio->bi_phys_segments = 0;
-		bio->bi_hw_segments = 0;
 		bio->bi_size = 0;
 	}
 
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 55e7c56..ae16794 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -101,6 +101,40 @@
 const char raid6_empty_zero_page[PAGE_SIZE] __attribute__((aligned(256)));
 #endif
 
+/*
+ * We maintain a biased count of active stripes in the bottom 16 bits of
+ * bi_phys_segments, and a count of processed stripes in the upper 16 bits
+ */
+static inline int raid5_bi_phys_segments(struct bio *bio)
+{
+	return bio->bi_phys_segments & 0xffff;
+}
+
+static inline int raid5_bi_hw_segments(struct bio *bio)
+{
+	return (bio->bi_phys_segments >> 16) & 0xffff;
+}
+
+static inline int raid5_dec_bi_phys_segments(struct bio *bio)
+{
+	--bio->bi_phys_segments;
+	return raid5_bi_phys_segments(bio);
+}
+
+static inline int raid5_dec_bi_hw_segments(struct bio *bio)
+{
+	unsigned short val = raid5_bi_hw_segments(bio);
+
+	--val;
+	bio->bi_phys_segments = (val << 16) | raid5_bi_phys_segments(bio);
+	return val;
+}
+
+static inline void raid5_set_bi_hw_segments(struct bio *bio, unsigned int cnt)
+{
+	bio->bi_phys_segments = raid5_bi_phys_segments(bio) || (cnt << 16);
+}
+
 static inline int raid6_next_disk(int disk, int raid_disks)
 {
 	disk++;
@@ -507,7 +541,7 @@ static void ops_complete_biofill(void *stripe_head_ref)
 			while (rbi && rbi->bi_sector <
 				dev->sector + STRIPE_SECTORS) {
 				rbi2 = r5_next_bio(rbi, dev->sector);
-				if (--rbi->bi_phys_segments == 0) {
+				if (!raid5_dec_bi_phys_segments(rbi)) {
 					rbi->bi_next = return_bi;
 					return_bi = rbi;
 				}
@@ -1725,7 +1759,7 @@ static int add_stripe_bio(struct stripe_head *sh, struct bio *bi, int dd_idx, in
 	if (*bip)
 		bi->bi_next = *bip;
 	*bip = bi;
-	bi->bi_phys_segments ++;
+	bi->bi_phys_segments++;
 	spin_unlock_irq(&conf->device_lock);
 	spin_unlock(&sh->lock);
 
@@ -1819,7 +1853,7 @@ handle_failed_stripe(raid5_conf_t *conf, struct stripe_head *sh,
 			sh->dev[i].sector + STRIPE_SECTORS) {
 			struct bio *nextbi = r5_next_bio(bi, sh->dev[i].sector);
 			clear_bit(BIO_UPTODATE, &bi->bi_flags);
-			if (--bi->bi_phys_segments == 0) {
+			if (!raid5_dec_bi_phys_segments(bi)) {
 				md_write_end(conf->mddev);
 				bi->bi_next = *return_bi;
 				*return_bi = bi;
@@ -1834,7 +1868,7 @@ handle_failed_stripe(raid5_conf_t *conf, struct stripe_head *sh,
 		       sh->dev[i].sector + STRIPE_SECTORS) {
 			struct bio *bi2 = r5_next_bio(bi, sh->dev[i].sector);
 			clear_bit(BIO_UPTODATE, &bi->bi_flags);
-			if (--bi->bi_phys_segments == 0) {
+			if (!raid5_dec_bi_phys_segments(bi)) {
 				md_write_end(conf->mddev);
 				bi->bi_next = *return_bi;
 				*return_bi = bi;
@@ -1858,7 +1892,7 @@ handle_failed_stripe(raid5_conf_t *conf, struct stripe_head *sh,
 				struct bio *nextbi =
 					r5_next_bio(bi, sh->dev[i].sector);
 				clear_bit(BIO_UPTODATE, &bi->bi_flags);
-				if (--bi->bi_phys_segments == 0) {
+				if (!raid5_dec_bi_phys_segments(bi)) {
 					bi->bi_next = *return_bi;
 					*return_bi = bi;
 				}
@@ -2033,7 +2067,7 @@ static void handle_stripe_clean_event(raid5_conf_t *conf,
 				while (wbi && wbi->bi_sector <
 					dev->sector + STRIPE_SECTORS) {
 					wbi2 = r5_next_bio(wbi, dev->sector);
-					if (--wbi->bi_phys_segments == 0) {
+					if (!raid5_dec_bi_phys_segments(wbi)) {
 						md_write_end(conf->mddev);
 						wbi->bi_next = *return_bi;
 						*return_bi = wbi;
@@ -2507,7 +2541,7 @@ static void handle_stripe_expansion(raid5_conf_t *conf, struct stripe_head *sh,
  *
  */
 
-static void handle_stripe5(struct stripe_head *sh)
+static bool handle_stripe5(struct stripe_head *sh)
 {
 	raid5_conf_t *conf = sh->raid_conf;
 	int disks = sh->disks, i;
@@ -2568,10 +2602,10 @@ static void handle_stripe5(struct stripe_head *sh)
 		if (dev->written)
 			s.written++;
 		rdev = rcu_dereference(conf->disks[i].rdev);
-		if (rdev && unlikely(test_bit(Blocked, &rdev->flags))) {
+		if (blocked_rdev == NULL &&
+		    rdev && unlikely(test_bit(Blocked, &rdev->flags))) {
 			blocked_rdev = rdev;
 			atomic_inc(&rdev->nr_pending);
-			break;
 		}
 		if (!rdev || !test_bit(In_sync, &rdev->flags)) {
 			/* The ReadError flag will just be confusing now */
@@ -2588,8 +2622,14 @@ static void handle_stripe5(struct stripe_head *sh)
 	rcu_read_unlock();
 
 	if (unlikely(blocked_rdev)) {
-		set_bit(STRIPE_HANDLE, &sh->state);
-		goto unlock;
+		if (s.syncing || s.expanding || s.expanded ||
+		    s.to_write || s.written) {
+			set_bit(STRIPE_HANDLE, &sh->state);
+			goto unlock;
+		}
+		/* There is nothing for the blocked_rdev to block */
+		rdev_dec_pending(blocked_rdev, conf->mddev);
+		blocked_rdev = NULL;
 	}
 
 	if (s.to_fill && !test_bit(STRIPE_BIOFILL_RUN, &sh->state)) {
@@ -2717,10 +2757,11 @@ static void handle_stripe5(struct stripe_head *sh)
 	if (sh->reconstruct_state == reconstruct_state_result) {
 		sh->reconstruct_state = reconstruct_state_idle;
 		clear_bit(STRIPE_EXPANDING, &sh->state);
-		for (i = conf->raid_disks; i--; )
+		for (i = conf->raid_disks; i--; ) {
 			set_bit(R5_Wantwrite, &sh->dev[i].flags);
-			set_bit(R5_LOCKED, &dev->flags);
+			set_bit(R5_LOCKED, &sh->dev[i].flags);
 			s.locked++;
+		}
 	}
 
 	if (s.expanded && test_bit(STRIPE_EXPANDING, &sh->state) &&
@@ -2754,9 +2795,11 @@ static void handle_stripe5(struct stripe_head *sh)
 	ops_run_io(sh, &s);
 
 	return_io(return_bi);
+
+	return blocked_rdev == NULL;
 }
 
-static void handle_stripe6(struct stripe_head *sh, struct page *tmp_page)
+static bool handle_stripe6(struct stripe_head *sh, struct page *tmp_page)
 {
 	raid6_conf_t *conf = sh->raid_conf;
 	int disks = sh->disks;
@@ -2805,7 +2848,7 @@ static void handle_stripe6(struct stripe_head *sh, struct page *tmp_page)
 				copy_data(0, rbi, dev->page, dev->sector);
 				rbi2 = r5_next_bio(rbi, dev->sector);
 				spin_lock_irq(&conf->device_lock);
-				if (--rbi->bi_phys_segments == 0) {
+				if (!raid5_dec_bi_phys_segments(rbi)) {
 					rbi->bi_next = return_bi;
 					return_bi = rbi;
 				}
@@ -2829,10 +2872,10 @@ static void handle_stripe6(struct stripe_head *sh, struct page *tmp_page)
 		if (dev->written)
 			s.written++;
 		rdev = rcu_dereference(conf->disks[i].rdev);
-		if (rdev && unlikely(test_bit(Blocked, &rdev->flags))) {
+		if (blocked_rdev == NULL &&
+		    rdev && unlikely(test_bit(Blocked, &rdev->flags))) {
 			blocked_rdev = rdev;
 			atomic_inc(&rdev->nr_pending);
-			break;
 		}
 		if (!rdev || !test_bit(In_sync, &rdev->flags)) {
 			/* The ReadError flag will just be confusing now */
@@ -2850,9 +2893,16 @@ static void handle_stripe6(struct stripe_head *sh, struct page *tmp_page)
 	rcu_read_unlock();
 
 	if (unlikely(blocked_rdev)) {
-		set_bit(STRIPE_HANDLE, &sh->state);
-		goto unlock;
+		if (s.syncing || s.expanding || s.expanded ||
+		    s.to_write || s.written) {
+			set_bit(STRIPE_HANDLE, &sh->state);
+			goto unlock;
+		}
+		/* There is nothing for the blocked_rdev to block */
+		rdev_dec_pending(blocked_rdev, conf->mddev);
+		blocked_rdev = NULL;
 	}
+
 	pr_debug("locked=%d uptodate=%d to_read=%d"
 	       " to_write=%d failed=%d failed_num=%d,%d\n",
 	       s.locked, s.uptodate, s.to_read, s.to_write, s.failed,
@@ -2967,14 +3017,17 @@ static void handle_stripe6(struct stripe_head *sh, struct page *tmp_page)
 	ops_run_io(sh, &s);
 
 	return_io(return_bi);
+
+	return blocked_rdev == NULL;
 }
 
-static void handle_stripe(struct stripe_head *sh, struct page *tmp_page)
+/* returns true if the stripe was handled */
+static bool handle_stripe(struct stripe_head *sh, struct page *tmp_page)
 {
 	if (sh->raid_conf->level == 6)
-		handle_stripe6(sh, tmp_page);
+		return handle_stripe6(sh, tmp_page);
 	else
-		handle_stripe5(sh);
+		return handle_stripe5(sh);
 }
 
 
@@ -3136,8 +3189,11 @@ static struct bio *remove_bio_from_retry(raid5_conf_t *conf)
 	if(bi) {
 		conf->retry_read_aligned_list = bi->bi_next;
 		bi->bi_next = NULL;
+		/*
+		 * this sets the active strip count to 1 and the processed
+		 * strip count to zero (upper 8 bits)
+		 */
 		bi->bi_phys_segments = 1; /* biased count of active stripes */
-		bi->bi_hw_segments = 0; /* count of processed stripes */
 	}
 
 	return bi;
@@ -3187,8 +3243,7 @@ static int bio_fits_rdev(struct bio *bi)
 	if ((bi->bi_size>>9) > q->max_sectors)
 		return 0;
 	blk_recount_segments(q, bi);
-	if (bi->bi_phys_segments > q->max_phys_segments ||
-	    bi->bi_hw_segments > q->max_hw_segments)
+	if (bi->bi_phys_segments > q->max_phys_segments)
 		return 0;
 
 	if (q->merge_bvec_fn)
@@ -3332,7 +3387,7 @@ static int make_request(struct request_queue *q, struct bio * bi)
 	sector_t logical_sector, last_sector;
 	struct stripe_head *sh;
 	const int rw = bio_data_dir(bi);
-	int remaining;
+	int cpu, remaining;
 
 	if (unlikely(bio_barrier(bi))) {
 		bio_endio(bi, -EOPNOTSUPP);
@@ -3341,8 +3396,11 @@ static int make_request(struct request_queue *q, struct bio * bi)
 
 	md_write_start(mddev, bi);
 
-	disk_stat_inc(mddev->gendisk, ios[rw]);
-	disk_stat_add(mddev->gendisk, sectors[rw], bio_sectors(bi));
+	cpu = part_stat_lock();
+	part_stat_inc(cpu, &mddev->gendisk->part0, ios[rw]);
+	part_stat_add(cpu, &mddev->gendisk->part0, sectors[rw],
+		      bio_sectors(bi));
+	part_stat_unlock();
 
 	if (rw == READ &&
 	     mddev->reshape_position == MaxSector &&
@@ -3449,7 +3507,7 @@ static int make_request(struct request_queue *q, struct bio * bi)
 			
 	}
 	spin_lock_irq(&conf->device_lock);
-	remaining = --bi->bi_phys_segments;
+	remaining = raid5_dec_bi_phys_segments(bi);
 	spin_unlock_irq(&conf->device_lock);
 	if (remaining == 0) {
 
@@ -3692,7 +3750,9 @@ static inline sector_t sync_request(mddev_t *mddev, sector_t sector_nr, int *ski
 	clear_bit(STRIPE_INSYNC, &sh->state);
 	spin_unlock(&sh->lock);
 
-	handle_stripe(sh, NULL);
+	/* wait for any blocked device to be handled */
+	while(unlikely(!handle_stripe(sh, NULL)))
+		;
 	release_stripe(sh);
 
 	return STRIPE_SECTORS;
@@ -3731,7 +3791,7 @@ static int  retry_aligned_read(raid5_conf_t *conf, struct bio *raid_bio)
 		     sector += STRIPE_SECTORS,
 		     scnt++) {
 
-		if (scnt < raid_bio->bi_hw_segments)
+		if (scnt < raid5_bi_hw_segments(raid_bio))
 			/* already done this stripe */
 			continue;
 
@@ -3739,7 +3799,7 @@ static int  retry_aligned_read(raid5_conf_t *conf, struct bio *raid_bio)
 
 		if (!sh) {
 			/* failed to get a stripe - must wait */
-			raid_bio->bi_hw_segments = scnt;
+			raid5_set_bi_hw_segments(raid_bio, scnt);
 			conf->retry_read_aligned = raid_bio;
 			return handled;
 		}
@@ -3747,7 +3807,7 @@ static int  retry_aligned_read(raid5_conf_t *conf, struct bio *raid_bio)
 		set_bit(R5_ReadError, &sh->dev[dd_idx].flags);
 		if (!add_stripe_bio(sh, raid_bio, dd_idx, 0)) {
 			release_stripe(sh);
-			raid_bio->bi_hw_segments = scnt;
+			raid5_set_bi_hw_segments(raid_bio, scnt);
 			conf->retry_read_aligned = raid_bio;
 			return handled;
 		}
@@ -3757,7 +3817,7 @@ static int  retry_aligned_read(raid5_conf_t *conf, struct bio *raid_bio)
 		handled++;
 	}
 	spin_lock_irq(&conf->device_lock);
-	remaining = --raid_bio->bi_phys_segments;
+	remaining = raid5_dec_bi_phys_segments(raid_bio);
 	spin_unlock_irq(&conf->device_lock);
 	if (remaining == 0)
 		bio_endio(raid_bio, 0);
@@ -3811,10 +3871,8 @@ static void raid5d(mddev_t *mddev)
 
 		sh = __get_priority_stripe(conf);
 
-		if (!sh) {
-			async_tx_issue_pending_all();
+		if (!sh)
 			break;
-		}
 		spin_unlock_irq(&conf->device_lock);
 		
 		handled++;
@@ -3827,6 +3885,7 @@ static void raid5d(mddev_t *mddev)
 
 	spin_unlock_irq(&conf->device_lock);
 
+	async_tx_issue_pending_all();
 	unplug_slaves(mddev);
 
 	pr_debug("--- raid5d inactive\n");
@@ -4439,6 +4498,9 @@ static int raid5_check_reshape(mddev_t *mddev)
 		return -EINVAL; /* Cannot shrink array or change level yet */
 	if (mddev->delta_disks == 0)
 		return 0; /* nothing to do */
+	if (mddev->bitmap)
+		/* Cannot grow a bitmap yet */
+		return -EBUSY;
 
 	/* Can only proceed if there are plenty of stripe_heads.
 	 * We need a minimum of one full stripe,, and for sensible progress