15 files changed, 287 insertions, 168 deletions

diff --git a/drivers/md/bitmap.c b/drivers/md/bitmap.c
index 70bd738b8b99..574b09afedd3 100644
--- a/drivers/md/bitmap.c
+++ b/drivers/md/bitmap.c
@@ -534,6 +534,82 @@ void bitmap_print_sb(struct bitmap *bitmap)
 	kunmap_atomic(sb, KM_USER0);
 }
 
+/*
+ * bitmap_new_disk_sb
+ * @bitmap
+ *
+ * This function is somewhat the reverse of bitmap_read_sb.  bitmap_read_sb
+ * reads and verifies the on-disk bitmap superblock and populates bitmap_info.
+ * This function verifies 'bitmap_info' and populates the on-disk bitmap
+ * structure, which is to be written to disk.
+ *
+ * Returns: 0 on success, -Exxx on error
+ */
+static int bitmap_new_disk_sb(struct bitmap *bitmap)
+{
+	bitmap_super_t *sb;
+	unsigned long chunksize, daemon_sleep, write_behind;
+	int err = -EINVAL;
+
+	bitmap->sb_page = alloc_page(GFP_KERNEL);
+	if (IS_ERR(bitmap->sb_page)) {
+		err = PTR_ERR(bitmap->sb_page);
+		bitmap->sb_page = NULL;
+		return err;
+	}
+	bitmap->sb_page->index = 0;
+
+	sb = kmap_atomic(bitmap->sb_page, KM_USER0);
+
+	sb->magic = cpu_to_le32(BITMAP_MAGIC);
+	sb->version = cpu_to_le32(BITMAP_MAJOR_HI);
+
+	chunksize = bitmap->mddev->bitmap_info.chunksize;
+	BUG_ON(!chunksize);
+	if (!is_power_of_2(chunksize)) {
+		kunmap_atomic(sb, KM_USER0);
+		printk(KERN_ERR "bitmap chunksize not a power of 2\n");
+		return -EINVAL;
+	}
+	sb->chunksize = cpu_to_le32(chunksize);
+
+	daemon_sleep = bitmap->mddev->bitmap_info.daemon_sleep;
+	if (!daemon_sleep ||
+	    (daemon_sleep < 1) || (daemon_sleep > MAX_SCHEDULE_TIMEOUT)) {
+		printk(KERN_INFO "Choosing daemon_sleep default (5 sec)\n");
+		daemon_sleep = 5 * HZ;
+	}
+	sb->daemon_sleep = cpu_to_le32(daemon_sleep);
+	bitmap->mddev->bitmap_info.daemon_sleep = daemon_sleep;
+
+	/*
+	 * FIXME: write_behind for RAID1.  If not specified, what
+	 * is a good choice?  We choose COUNTER_MAX / 2 arbitrarily.
+	 */
+	write_behind = bitmap->mddev->bitmap_info.max_write_behind;
+	if (write_behind > COUNTER_MAX)
+		write_behind = COUNTER_MAX / 2;
+	sb->write_behind = cpu_to_le32(write_behind);
+	bitmap->mddev->bitmap_info.max_write_behind = write_behind;
+
+	/* keep the array size field of the bitmap superblock up to date */
+	sb->sync_size = cpu_to_le64(bitmap->mddev->resync_max_sectors);
+
+	memcpy(sb->uuid, bitmap->mddev->uuid, 16);
+
+	bitmap->flags |= BITMAP_STALE;
+	sb->state |= cpu_to_le32(BITMAP_STALE);
+	bitmap->events_cleared = bitmap->mddev->events;
+	sb->events_cleared = cpu_to_le64(bitmap->mddev->events);
+
+	bitmap->flags |= BITMAP_HOSTENDIAN;
+	sb->version = cpu_to_le32(BITMAP_MAJOR_HOSTENDIAN);
+
+	kunmap_atomic(sb, KM_USER0);
+
+	return 0;
+}
+
 /* read the superblock from the bitmap file and initialize some bitmap fields */
 static int bitmap_read_sb(struct bitmap *bitmap)
 {
@@ -575,7 +651,7 @@ static int bitmap_read_sb(struct bitmap *bitmap)
 		reason = "unrecognized superblock version";
 	else if (chunksize < 512)
 		reason = "bitmap chunksize too small";
-	else if ((1 << ffz(~chunksize)) != chunksize)
+	else if (!is_power_of_2(chunksize))
 		reason = "bitmap chunksize not a power of 2";
 	else if (daemon_sleep < 1 || daemon_sleep > MAX_SCHEDULE_TIMEOUT)
 		reason = "daemon sleep period out of range";
@@ -1076,8 +1152,8 @@ static int bitmap_init_from_disk(struct bitmap *bitmap, sector_t start)
 	}
 
 	printk(KERN_INFO "%s: bitmap initialized from disk: "
-		"read %lu/%lu pages, set %lu bits\n",
-		bmname(bitmap), bitmap->file_pages, num_pages, bit_cnt);
+	       "read %lu/%lu pages, set %lu of %lu bits\n",
+	       bmname(bitmap), bitmap->file_pages, num_pages, bit_cnt, chunks);
 
 	return 0;
 
@@ -1332,7 +1408,7 @@ int bitmap_startwrite(struct bitmap *bitmap, sector_t offset, unsigned long sect
 			return 0;
 		}
 
-		if (unlikely((*bmc & COUNTER_MAX) == COUNTER_MAX)) {
+		if (unlikely(COUNTER(*bmc) == COUNTER_MAX)) {
 			DEFINE_WAIT(__wait);
 			/* note that it is safe to do the prepare_to_wait
 			 * after the test as long as we do it before dropping
@@ -1404,10 +1480,10 @@ void bitmap_endwrite(struct bitmap *bitmap, sector_t offset, unsigned long secto
 			sysfs_notify_dirent_safe(bitmap->sysfs_can_clear);
 		}
 
-		if (!success && ! (*bmc & NEEDED_MASK))
+		if (!success && !NEEDED(*bmc))
 			*bmc |= NEEDED_MASK;
 
-		if ((*bmc & COUNTER_MAX) == COUNTER_MAX)
+		if (COUNTER(*bmc) == COUNTER_MAX)
 			wake_up(&bitmap->overflow_wait);
 
 		(*bmc)--;
@@ -1728,9 +1804,16 @@ int bitmap_create(mddev_t *mddev)
 		vfs_fsync(file, 1);
 	}
 	/* read superblock from bitmap file (this sets mddev->bitmap_info.chunksize) */
-	if (!mddev->bitmap_info.external)
-		err = bitmap_read_sb(bitmap);
-	else {
+	if (!mddev->bitmap_info.external) {
+		/*
+		 * If 'MD_ARRAY_FIRST_USE' is set, then device-mapper is
+		 * instructing us to create a new on-disk bitmap instance.
+		 */
+		if (test_and_clear_bit(MD_ARRAY_FIRST_USE, &mddev->flags))
+			err = bitmap_new_disk_sb(bitmap);
+		else
+			err = bitmap_read_sb(bitmap);
+	} else {
 		err = 0;
 		if (mddev->bitmap_info.chunksize == 0 ||
 		    mddev->bitmap_info.daemon_sleep == 0)
@@ -1754,9 +1837,6 @@ int bitmap_create(mddev_t *mddev)
 	bitmap->chunks = chunks;
 	bitmap->pages = pages;
 	bitmap->missing_pages = pages;
-	bitmap->counter_bits = COUNTER_BITS;
-
-	bitmap->syncchunk = ~0UL;
 
 #ifdef INJECT_FATAL_FAULT_1
 	bitmap->bp = NULL;
diff --git a/drivers/md/bitmap.h b/drivers/md/bitmap.h
index d0aeaf46d932..b2a127e891ac 100644
--- a/drivers/md/bitmap.h
+++ b/drivers/md/bitmap.h
@@ -85,7 +85,6 @@
 typedef __u16 bitmap_counter_t;
 #define COUNTER_BITS 16
 #define COUNTER_BIT_SHIFT 4
-#define COUNTER_BYTE_RATIO (COUNTER_BITS / 8)
 #define COUNTER_BYTE_SHIFT (COUNTER_BIT_SHIFT - 3)
 
 #define NEEDED_MASK ((bitmap_counter_t) (1 << (COUNTER_BITS - 1)))
@@ -196,19 +195,10 @@ struct bitmap {
 
 	mddev_t *mddev; /* the md device that the bitmap is for */
 
-	int counter_bits; /* how many bits per block counter */
-
 	/* bitmap chunksize -- how much data does each bit represent? */
 	unsigned long chunkshift; /* chunksize = 2^chunkshift (for bitops) */
 	unsigned long chunks; /* total number of data chunks for the array */
 
-	/* We hold a count on the chunk currently being synced, and drop
-	 * it when the last block is started.  If the resync is aborted
-	 * midway, we need to be able to drop that count, so we remember
-	 * the counted chunk..
-	 */
-	unsigned long syncchunk;
-
 	__u64	events_cleared;
 	int need_sync;
 
diff --git a/drivers/md/dm-io.c b/drivers/md/dm-io.c
index 76a5af00a26b..2067288f61f9 100644
--- a/drivers/md/dm-io.c
+++ b/drivers/md/dm-io.c
@@ -19,6 +19,8 @@
 #define DM_MSG_PREFIX "io"
 
 #define DM_IO_MAX_REGIONS	BITS_PER_LONG
+#define MIN_IOS		16
+#define MIN_BIOS	16
 
 struct dm_io_client {
 	mempool_t *pool;
@@ -41,33 +43,21 @@ struct io {
 static struct kmem_cache *_dm_io_cache;
 
 /*
- * io contexts are only dynamically allocated for asynchronous
- * io.  Since async io is likely to be the majority of io we'll
- * have the same number of io contexts as bios! (FIXME: must reduce this).
- */
-
-static unsigned int pages_to_ios(unsigned int pages)
-{
-	return 4 * pages;	/* too many ? */
-}
-
-/*
  * Create a client with mempool and bioset.
  */
-struct dm_io_client *dm_io_client_create(unsigned num_pages)
+struct dm_io_client *dm_io_client_create(void)
 {
-	unsigned ios = pages_to_ios(num_pages);
 	struct dm_io_client *client;
 
 	client = kmalloc(sizeof(*client), GFP_KERNEL);
 	if (!client)
 		return ERR_PTR(-ENOMEM);
 
-	client->pool = mempool_create_slab_pool(ios, _dm_io_cache);
+	client->pool = mempool_create_slab_pool(MIN_IOS, _dm_io_cache);
 	if (!client->pool)
 		goto bad;
 
-	client->bios = bioset_create(16, 0);
+	client->bios = bioset_create(MIN_BIOS, 0);
 	if (!client->bios)
 		goto bad;
 
@@ -81,13 +71,6 @@ struct dm_io_client *dm_io_client_create(unsigned num_pages)
 }
 EXPORT_SYMBOL(dm_io_client_create);
 
-int dm_io_client_resize(unsigned num_pages, struct dm_io_client *client)
-{
-	return mempool_resize(client->pool, pages_to_ios(num_pages),
-			      GFP_KERNEL);
-}
-EXPORT_SYMBOL(dm_io_client_resize);
-
 void dm_io_client_destroy(struct dm_io_client *client)
 {
 	mempool_destroy(client->pool);
diff --git a/drivers/md/dm-kcopyd.c b/drivers/md/dm-kcopyd.c
index 1bb73a13ca40..819e37eaaeba 100644
--- a/drivers/md/dm-kcopyd.c
+++ b/drivers/md/dm-kcopyd.c
@@ -27,15 +27,19 @@
 
 #include "dm.h"
 
+#define SUB_JOB_SIZE	128
+#define SPLIT_COUNT	8
+#define MIN_JOBS	8
+#define RESERVE_PAGES	(DIV_ROUND_UP(SUB_JOB_SIZE << SECTOR_SHIFT, PAGE_SIZE))
+
 /*-----------------------------------------------------------------
  * Each kcopyd client has its own little pool of preallocated
  * pages for kcopyd io.
  *---------------------------------------------------------------*/
 struct dm_kcopyd_client {
-	spinlock_t lock;
 	struct page_list *pages;
-	unsigned int nr_pages;
-	unsigned int nr_free_pages;
+	unsigned nr_reserved_pages;
+	unsigned nr_free_pages;
 
 	struct dm_io_client *io_client;
 
@@ -67,15 +71,18 @@ static void wake(struct dm_kcopyd_client *kc)
 	queue_work(kc->kcopyd_wq, &kc->kcopyd_work);
 }
 
-static struct page_list *alloc_pl(void)
+/*
+ * Obtain one page for the use of kcopyd.
+ */
+static struct page_list *alloc_pl(gfp_t gfp)
 {
 	struct page_list *pl;
 
-	pl = kmalloc(sizeof(*pl), GFP_KERNEL);
+	pl = kmalloc(sizeof(*pl), gfp);
 	if (!pl)
 		return NULL;
 
-	pl->page = alloc_page(GFP_KERNEL);
+	pl->page = alloc_page(gfp);
 	if (!pl->page) {
 		kfree(pl);
 		return NULL;
@@ -90,41 +97,56 @@ static void free_pl(struct page_list *pl)
 	kfree(pl);
 }
 
-static int kcopyd_get_pages(struct dm_kcopyd_client *kc,
-			    unsigned int nr, struct page_list **pages)
+/*
+ * Add the provided pages to a client's free page list, releasing
+ * back to the system any beyond the reserved_pages limit.
+ */
+static void kcopyd_put_pages(struct dm_kcopyd_client *kc, struct page_list *pl)
 {
-	struct page_list *pl;
-
-	spin_lock(&kc->lock);
-	if (kc->nr_free_pages < nr) {
-		spin_unlock(&kc->lock);
-		return -ENOMEM;
-	}
-
-	kc->nr_free_pages -= nr;
-	for (*pages = pl = kc->pages; --nr; pl = pl->next)
-		;
+	struct page_list *next;
 
-	kc->pages = pl->next;
-	pl->next = NULL;
+	do {
+		next = pl->next;
 
-	spin_unlock(&kc->lock);
+		if (kc->nr_free_pages >= kc->nr_reserved_pages)
+			free_pl(pl);
+		else {
+			pl->next = kc->pages;
+			kc->pages = pl;
+			kc->nr_free_pages++;
+		}
 
-	return 0;
+		pl = next;
+	} while (pl);
 }
 
-static void kcopyd_put_pages(struct dm_kcopyd_client *kc, struct page_list *pl)
+static int kcopyd_get_pages(struct dm_kcopyd_client *kc,
+			    unsigned int nr, struct page_list **pages)
 {
-	struct page_list *cursor;
+	struct page_list *pl;
+
+	*pages = NULL;
+
+	do {
+		pl = alloc_pl(__GFP_NOWARN | __GFP_NORETRY);
+		if (unlikely(!pl)) {
+			/* Use reserved pages */
+			pl = kc->pages;
+			if (unlikely(!pl))
+				goto out_of_memory;
+			kc->pages = pl->next;
+			kc->nr_free_pages--;
+		}
+		pl->next = *pages;
+		*pages = pl;
+	} while (--nr);
 
-	spin_lock(&kc->lock);
-	for (cursor = pl; cursor->next; cursor = cursor->next)
-		kc->nr_free_pages++;
+	return 0;
 
-	kc->nr_free_pages++;
-	cursor->next = kc->pages;
-	kc->pages = pl;
-	spin_unlock(&kc->lock);
+out_of_memory:
+	if (*pages)
+		kcopyd_put_pages(kc, *pages);
+	return -ENOMEM;
 }
 
 /*
@@ -141,13 +163,16 @@ static void drop_pages(struct page_list *pl)
 	}
 }
 
-static int client_alloc_pages(struct dm_kcopyd_client *kc, unsigned int nr)
+/*
+ * Allocate and reserve nr_pages for the use of a specific client.
+ */
+static int client_reserve_pages(struct dm_kcopyd_client *kc, unsigned nr_pages)
 {
-	unsigned int i;
+	unsigned i;
 	struct page_list *pl = NULL, *next;
 
-	for (i = 0; i < nr; i++) {
-		next = alloc_pl();
+	for (i = 0; i < nr_pages; i++) {
+		next = alloc_pl(GFP_KERNEL);
 		if (!next) {
 			if (pl)
 				drop_pages(pl);
@@ -157,17 +182,18 @@ static int client_alloc_pages(struct dm_kcopyd_client *kc, unsigned int nr)
 		pl = next;
 	}
 
+	kc->nr_reserved_pages += nr_pages;
 	kcopyd_put_pages(kc, pl);
-	kc->nr_pages += nr;
+
 	return 0;
 }
 
 static void client_free_pages(struct dm_kcopyd_client *kc)
 {
-	BUG_ON(kc->nr_free_pages != kc->nr_pages);
+	BUG_ON(kc->nr_free_pages != kc->nr_reserved_pages);
 	drop_pages(kc->pages);
 	kc->pages = NULL;
-	kc->nr_free_pages = kc->nr_pages = 0;
+	kc->nr_free_pages = kc->nr_reserved_pages = 0;
 }
 
 /*-----------------------------------------------------------------
@@ -216,16 +242,17 @@ struct kcopyd_job {
 	struct mutex lock;
 	atomic_t sub_jobs;
 	sector_t progress;
-};
 
-/* FIXME: this should scale with the number of pages */
-#define MIN_JOBS 512
+	struct kcopyd_job *master_job;
+};
 
 static struct kmem_cache *_job_cache;
 
 int __init dm_kcopyd_init(void)
 {
-	_job_cache = KMEM_CACHE(kcopyd_job, 0);
+	_job_cache = kmem_cache_create("kcopyd_job",
+				sizeof(struct kcopyd_job) * (SPLIT_COUNT + 1),
+				__alignof__(struct kcopyd_job), 0, NULL);
 	if (!_job_cache)
 		return -ENOMEM;
 
@@ -299,7 +326,12 @@ static int run_complete_job(struct kcopyd_job *job)
 
 	if (job->pages)
 		kcopyd_put_pages(kc, job->pages);
-	mempool_free(job, kc->job_pool);
+	/*
+	 * If this is the master job, the sub jobs have already
+	 * completed so we can free everything.
+	 */
+	if (job->master_job == job)
+		mempool_free(job, kc->job_pool);
 	fn(read_err, write_err, context);
 
 	if (atomic_dec_and_test(&kc->nr_jobs))
@@ -460,14 +492,14 @@ static void dispatch_job(struct kcopyd_job *job)
 	wake(kc);
 }
 
-#define SUB_JOB_SIZE 128
 static void segment_complete(int read_err, unsigned long write_err,
 			     void *context)
 {
 	/* FIXME: tidy this function */
 	sector_t progress = 0;
 	sector_t count = 0;
-	struct kcopyd_job *job = (struct kcopyd_job *) context;
+	struct kcopyd_job *sub_job = (struct kcopyd_job *) context;
+	struct kcopyd_job *job = sub_job->master_job;
 	struct dm_kcopyd_client *kc = job->kc;
 
 	mutex_lock(&job->lock);
@@ -498,8 +530,6 @@ static void segment_complete(int read_err, unsigned long write_err,
 
 	if (count) {
 		int i;
-		struct kcopyd_job *sub_job = mempool_alloc(kc->job_pool,
-							   GFP_NOIO);
 
 		*sub_job = *job;
 		sub_job->source.sector += progress;
@@ -511,7 +541,7 @@ static void segment_complete(int read_err, unsigned long write_err,
 		}
 
 		sub_job->fn = segment_complete;
-		sub_job->context = job;
+		sub_job->context = sub_job;
 		dispatch_job(sub_job);
 
 	} else if (atomic_dec_and_test(&job->sub_jobs)) {
@@ -531,19 +561,19 @@ static void segment_complete(int read_err, unsigned long write_err,
 }
 
 /*
- * Create some little jobs that will do the move between
- * them.
+ * Create some sub jobs to share the work between them.
  */
-#define SPLIT_COUNT 8
-static void split_job(struct kcopyd_job *job)
+static void split_job(struct kcopyd_job *master_job)
 {
 	int i;
 
-	atomic_inc(&job->kc->nr_jobs);
+	atomic_inc(&master_job->kc->nr_jobs);
 
-	atomic_set(&job->sub_jobs, SPLIT_COUNT);
-	for (i = 0; i < SPLIT_COUNT; i++)
-		segment_complete(0, 0u, job);
+	atomic_set(&master_job->sub_jobs, SPLIT_COUNT);
+	for (i = 0; i < SPLIT_COUNT; i++) {
+		master_job[i + 1].master_job = master_job;
+		segment_complete(0, 0u, &master_job[i + 1]);
+	}
 }
 
 int dm_kcopyd_copy(struct dm_kcopyd_client *kc, struct dm_io_region *from,
@@ -553,7 +583,8 @@ int dm_kcopyd_copy(struct dm_kcopyd_client *kc, struct dm_io_region *from,
 	struct kcopyd_job *job;
 
 	/*
-	 * Allocate a new job.
+	 * Allocate an array of jobs consisting of one master job
+	 * followed by SPLIT_COUNT sub jobs.
 	 */
 	job = mempool_alloc(kc->job_pool, GFP_NOIO);
 
@@ -577,10 +608,10 @@ int dm_kcopyd_copy(struct dm_kcopyd_client *kc, struct dm_io_region *from,
 
 	job->fn = fn;
 	job->context = context;
+	job->master_job = job;
 
-	if (job->source.count < SUB_JOB_SIZE)
+	if (job->source.count <= SUB_JOB_SIZE)
 		dispatch_job(job);
-
 	else {
 		mutex_init(&job->lock);
 		job->progress = 0;
@@ -606,17 +637,15 @@ int kcopyd_cancel(struct kcopyd_job *job, int block)
 /*-----------------------------------------------------------------
  * Client setup
  *---------------------------------------------------------------*/
-int dm_kcopyd_client_create(unsigned int nr_pages,
-			    struct dm_kcopyd_client **result)
+struct dm_kcopyd_client *dm_kcopyd_client_create(void)
 {
 	int r = -ENOMEM;
 	struct dm_kcopyd_client *kc;
 
 	kc = kmalloc(sizeof(*kc), GFP_KERNEL);
 	if (!kc)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
 
-	spin_lock_init(&kc->lock);
 	spin_lock_init(&kc->job_lock);
 	INIT_LIST_HEAD(&kc->complete_jobs);
 	INIT_LIST_HEAD(&kc->io_jobs);
@@ -633,12 +662,12 @@ int dm_kcopyd_client_create(unsigned int nr_pages,
 		goto bad_workqueue;
 
 	kc->pages = NULL;
-	kc->nr_pages = kc->nr_free_pages = 0;
-	r = client_alloc_pages(kc, nr_pages);
+	kc->nr_reserved_pages = kc->nr_free_pages = 0;
+	r = client_reserve_pages(kc, RESERVE_PAGES);
 	if (r)
 		goto bad_client_pages;
 
-	kc->io_client = dm_io_client_create(nr_pages);
+	kc->io_client = dm_io_client_create();
 	if (IS_ERR(kc->io_client)) {
 		r = PTR_ERR(kc->io_client);
 		goto bad_io_client;
@@ -647,8 +676,7 @@ int dm_kcopyd_client_create(unsigned int nr_pages,
 	init_waitqueue_head(&kc->destroyq);
 	atomic_set(&kc->nr_jobs, 0);
 
-	*result = kc;
-	return 0;
+	return kc;
 
 bad_io_client:
 	client_free_pages(kc);
@@ -659,7 +687,7 @@ bad_workqueue:
 bad_slab:
 	kfree(kc);
 
-	return r;
+	return ERR_PTR(r);
 }
 EXPORT_SYMBOL(dm_kcopyd_client_create);
 
diff --git a/drivers/md/dm-log.c b/drivers/md/dm-log.c
index a1f321889676..948e3f4925bf 100644
--- a/drivers/md/dm-log.c
+++ b/drivers/md/dm-log.c
@@ -449,8 +449,7 @@ static int create_log_context(struct dm_dirty_log *log, struct dm_target *ti,
 
 		lc->io_req.mem.type = DM_IO_VMA;
 		lc->io_req.notify.fn = NULL;
-		lc->io_req.client = dm_io_client_create(dm_div_up(buf_size,
-								   PAGE_SIZE));
+		lc->io_req.client = dm_io_client_create();
 		if (IS_ERR(lc->io_req.client)) {
 			r = PTR_ERR(lc->io_req.client);
 			DMWARN("couldn't allocate disk io client");
diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c
index a550a057d991..aa4e570c2cb5 100644
--- a/drivers/md/dm-mpath.c
+++ b/drivers/md/dm-mpath.c
@@ -1290,7 +1290,7 @@ static int do_end_io(struct multipath *m, struct request *clone,
 	if (!error && !clone->errors)
 		return 0;	/* I/O complete */
 
-	if (error == -EOPNOTSUPP || error == -EREMOTEIO)
+	if (error == -EOPNOTSUPP || error == -EREMOTEIO || error == -EILSEQ)
 		return error;
 
 	if (mpio->pgpath)
diff --git a/drivers/md/dm-raid1.c b/drivers/md/dm-raid1.c
index 976ad4688afc..9bfd057be686 100644
--- a/drivers/md/dm-raid1.c
+++ b/drivers/md/dm-raid1.c
@@ -22,8 +22,6 @@
 #define DM_MSG_PREFIX "raid1"
 
 #define MAX_RECOVERY 1	/* Maximum number of regions recovered in parallel. */
-#define DM_IO_PAGES 64
-#define DM_KCOPYD_PAGES 64
 
 #define DM_RAID1_HANDLE_ERRORS 0x01
 #define errors_handled(p)	((p)->features & DM_RAID1_HANDLE_ERRORS)
@@ -887,7 +885,7 @@ static struct mirror_set *alloc_context(unsigned int nr_mirrors,
 		return NULL;
 	}
 
-	ms->io_client = dm_io_client_create(DM_IO_PAGES);
+	ms->io_client = dm_io_client_create();
 	if (IS_ERR(ms->io_client)) {
 		ti->error = "Error creating dm_io client";
 		mempool_destroy(ms->read_record_pool);
@@ -1117,9 +1115,11 @@ static int mirror_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 		goto err_destroy_wq;
 	}
 
-	r = dm_kcopyd_client_create(DM_KCOPYD_PAGES, &ms->kcopyd_client);
-	if (r)
+	ms->kcopyd_client = dm_kcopyd_client_create();
+	if (IS_ERR(ms->kcopyd_client)) {
+		r = PTR_ERR(ms->kcopyd_client);
 		goto err_destroy_wq;
+	}
 
 	wakeup_mirrord(ms);
 	return 0;
diff --git a/drivers/md/dm-snap-persistent.c b/drivers/md/dm-snap-persistent.c
index 95891dfcbca0..135c2f1fdbfc 100644
--- a/drivers/md/dm-snap-persistent.c
+++ b/drivers/md/dm-snap-persistent.c
@@ -154,11 +154,6 @@ struct pstore {
 	struct workqueue_struct *metadata_wq;
 };
 
-static unsigned sectors_to_pages(unsigned sectors)
-{
-	return DIV_ROUND_UP(sectors, PAGE_SIZE >> 9);
-}
-
 static int alloc_area(struct pstore *ps)
 {
 	int r = -ENOMEM;
@@ -318,8 +313,7 @@ static int read_header(struct pstore *ps, int *new_snapshot)
 		chunk_size_supplied = 0;
 	}
 
-	ps->io_client = dm_io_client_create(sectors_to_pages(ps->store->
-							     chunk_size));
+	ps->io_client = dm_io_client_create();
 	if (IS_ERR(ps->io_client))
 		return PTR_ERR(ps->io_client);
 
@@ -368,11 +362,6 @@ static int read_header(struct pstore *ps, int *new_snapshot)
 		return r;
 	}
 
-	r = dm_io_client_resize(sectors_to_pages(ps->store->chunk_size),
-				ps->io_client);
-	if (r)
-		return r;
-
 	r = alloc_area(ps);
 	return r;
 
diff --git a/drivers/md/dm-snap.c b/drivers/md/dm-snap.c
index a2d330942cb2..9ecff5f3023a 100644
--- a/drivers/md/dm-snap.c
+++ b/drivers/md/dm-snap.c
@@ -40,11 +40,6 @@ static const char dm_snapshot_merge_target_name[] = "snapshot-merge";
 #define SNAPSHOT_COPY_PRIORITY 2
 
 /*
- * Reserve 1MB for each snapshot initially (with minimum of 1 page).
- */
-#define SNAPSHOT_PAGES (((1UL << 20) >> PAGE_SHIFT) ? : 1)
-
-/*
  * The size of the mempool used to track chunks in use.
  */
 #define MIN_IOS 256
@@ -1116,8 +1111,9 @@ static int snapshot_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 		goto bad_hash_tables;
 	}
 
-	r = dm_kcopyd_client_create(SNAPSHOT_PAGES, &s->kcopyd_client);
-	if (r) {
+	s->kcopyd_client = dm_kcopyd_client_create();
+	if (IS_ERR(s->kcopyd_client)) {
+		r = PTR_ERR(s->kcopyd_client);
 		ti->error = "Could not create kcopyd client";
 		goto bad_kcopyd;
 	}
diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
index cb8380c9767f..451c3bb176d2 100644
--- a/drivers/md/dm-table.c
+++ b/drivers/md/dm-table.c
@@ -362,6 +362,7 @@ static void close_dev(struct dm_dev_internal *d, struct mapped_device *md)
 static int device_area_is_invalid(struct dm_target *ti, struct dm_dev *dev,
 				  sector_t start, sector_t len, void *data)
 {
+	struct request_queue *q;
 	struct queue_limits *limits = data;
 	struct block_device *bdev = dev->bdev;
 	sector_t dev_size =
@@ -370,6 +371,22 @@ static int device_area_is_invalid(struct dm_target *ti, struct dm_dev *dev,
 		limits->logical_block_size >> SECTOR_SHIFT;
 	char b[BDEVNAME_SIZE];
 
+	/*
+	 * Some devices exist without request functions,
+	 * such as loop devices not yet bound to backing files.
+	 * Forbid the use of such devices.
+	 */
+	q = bdev_get_queue(bdev);
+	if (!q || !q->make_request_fn) {
+		DMWARN("%s: %s is not yet initialised: "
+		       "start=%llu, len=%llu, dev_size=%llu",
+		       dm_device_name(ti->table->md), bdevname(bdev, b),
+		       (unsigned long long)start,
+		       (unsigned long long)len,
+		       (unsigned long long)dev_size);
+		return 1;
+	}
+
 	if (!dev_size)
 		return 0;
 
@@ -1346,7 +1363,8 @@ bool dm_table_supports_discards(struct dm_table *t)
 		return 0;
 
 	/*
-	 * Ensure that at least one underlying device supports discards.
+	 * Unless any target used by the table set discards_supported,
+	 * require at least one underlying device to support discards.
 	 * t->devices includes internal dm devices such as mirror logs
 	 * so we need to use iterate_devices here, which targets
 	 * supporting discard must provide.
@@ -1354,6 +1372,9 @@ bool dm_table_supports_discards(struct dm_table *t)
 	while (i < dm_table_get_num_targets(t)) {
 		ti = dm_table_get_target(t, i++);
 
+		if (ti->discards_supported)
+			return 1;
+
 		if (ti->type->iterate_devices &&
 		    ti->type->iterate_devices(ti, device_discard_capable, NULL))
 			return 1;
diff --git a/drivers/md/md.c b/drivers/md/md.c
index aa640a85bb21..4332fc2f25d4 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -351,6 +351,9 @@ void mddev_resume(mddev_t *mddev)
 	mddev->suspended = 0;
 	wake_up(&mddev->sb_wait);
 	mddev->pers->quiesce(mddev, 0);
+
+	md_wakeup_thread(mddev->thread);
+	md_wakeup_thread(mddev->sync_thread); /* possibly kick off a reshape */
 }
 EXPORT_SYMBOL_GPL(mddev_resume);
 
@@ -1750,6 +1753,18 @@ static struct super_type super_types[] = {
 	},
 };
 
+static void sync_super(mddev_t *mddev, mdk_rdev_t *rdev)
+{
+	if (mddev->sync_super) {
+		mddev->sync_super(mddev, rdev);
+		return;
+	}
+
+	BUG_ON(mddev->major_version >= ARRAY_SIZE(super_types));
+
+	super_types[mddev->major_version].sync_super(mddev, rdev);
+}
+
 static int match_mddev_units(mddev_t *mddev1, mddev_t *mddev2)
 {
 	mdk_rdev_t *rdev, *rdev2;
@@ -1781,8 +1796,8 @@ int md_integrity_register(mddev_t *mddev)
 
 	if (list_empty(&mddev->disks))
 		return 0; /* nothing to do */
-	if (blk_get_integrity(mddev->gendisk))
-		return 0; /* already registered */
+	if (!mddev->gendisk || blk_get_integrity(mddev->gendisk))
+		return 0; /* shouldn't register, or already is */
 	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		/* skip spares and non-functional disks */
 		if (test_bit(Faulty, &rdev->flags))
@@ -2168,8 +2183,7 @@ static void sync_sbs(mddev_t * mddev, int nospares)
 			/* Don't update this superblock */
 			rdev->sb_loaded = 2;
 		} else {
-			super_types[mddev->major_version].
-				sync_super(mddev, rdev);
+			sync_super(mddev, rdev);
 			rdev->sb_loaded = 1;
 		}
 	}
@@ -2462,7 +2476,7 @@ slot_store(mdk_rdev_t *rdev, const char *buf, size_t len)
 		if (rdev->raid_disk == -1)
 			return -EEXIST;
 		/* personality does all needed checks */
-		if (rdev->mddev->pers->hot_add_disk == NULL)
+		if (rdev->mddev->pers->hot_remove_disk == NULL)
 			return -EINVAL;
 		err = rdev->mddev->pers->
 			hot_remove_disk(rdev->mddev, rdev->raid_disk);
@@ -4619,9 +4633,6 @@ int md_run(mddev_t *mddev)
 	if (mddev->flags)
 		md_update_sb(mddev, 0);
 
-	md_wakeup_thread(mddev->thread);
-	md_wakeup_thread(mddev->sync_thread); /* possibly kick off a reshape */
-
 	md_new_event(mddev);
 	sysfs_notify_dirent_safe(mddev->sysfs_state);
 	sysfs_notify_dirent_safe(mddev->sysfs_action);
@@ -4642,6 +4653,10 @@ static int do_md_run(mddev_t *mddev)
 		bitmap_destroy(mddev);
 		goto out;
 	}
+
+	md_wakeup_thread(mddev->thread);
+	md_wakeup_thread(mddev->sync_thread); /* possibly kick off a reshape */
+
 	set_capacity(mddev->gendisk, mddev->array_sectors);
 	revalidate_disk(mddev->gendisk);
 	mddev->changed = 1;
@@ -5259,6 +5274,8 @@ static int add_new_disk(mddev_t * mddev, mdu_disk_info_t *info)
 		if (mddev->degraded)
 			set_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
 		set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
+		if (!err)
+			md_new_event(mddev);
 		md_wakeup_thread(mddev->thread);
 		return err;
 	}
@@ -6866,8 +6883,8 @@ void md_do_sync(mddev_t *mddev)
 	 * Tune reconstruction:
 	 */
 	window = 32*(PAGE_SIZE/512);
-	printk(KERN_INFO "md: using %dk window, over a total of %llu blocks.\n",
-		window/2,(unsigned long long) max_sectors/2);
+	printk(KERN_INFO "md: using %dk window, over a total of %lluk.\n",
+		window/2, (unsigned long long)max_sectors/2);
 
 	atomic_set(&mddev->recovery_active, 0);
 	last_check = 0;
@@ -7045,7 +7062,6 @@ void md_do_sync(mddev_t *mddev)
 }
 EXPORT_SYMBOL_GPL(md_do_sync);
 
-
 static int remove_and_add_spares(mddev_t *mddev)
 {
 	mdk_rdev_t *rdev;
@@ -7157,6 +7173,9 @@ static void reap_sync_thread(mddev_t *mddev)
  */
 void md_check_recovery(mddev_t *mddev)
 {
+	if (mddev->suspended)
+		return;
+
 	if (mddev->bitmap)
 		bitmap_daemon_work(mddev);
 
diff --git a/drivers/md/md.h b/drivers/md/md.h
index 0b1fd3f1d85b..1c26c7a08ae6 100644
--- a/drivers/md/md.h
+++ b/drivers/md/md.h
@@ -124,6 +124,7 @@ struct mddev_s
 #define MD_CHANGE_DEVS	0	/* Some device status has changed */
 #define MD_CHANGE_CLEAN 1	/* transition to or from 'clean' */
 #define MD_CHANGE_PENDING 2	/* switch from 'clean' to 'active' in progress */
+#define MD_ARRAY_FIRST_USE 3    /* First use of array, needs initialization */
 
 	int				suspended;
 	atomic_t			active_io;
@@ -330,6 +331,7 @@ struct mddev_s
 	atomic_t flush_pending;
 	struct work_struct flush_work;
 	struct work_struct event_work;	/* used by dm to report failure event */
+	void (*sync_super)(mddev_t *mddev, mdk_rdev_t *rdev);
 };
 
 
diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index 5d096096f958..f7431b6d8447 100644
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -497,21 +497,19 @@ static int read_balance(conf_t *conf, r1bio_t *r1_bio)
 	return best_disk;
 }
 
-static int raid1_congested(void *data, int bits)
+int md_raid1_congested(mddev_t *mddev, int bits)
 {
-	mddev_t *mddev = data;
 	conf_t *conf = mddev->private;
 	int i, ret = 0;
 
-	if (mddev_congested(mddev, bits))
-		return 1;
-
 	rcu_read_lock();
 	for (i = 0; i < mddev->raid_disks; i++) {
 		mdk_rdev_t *rdev = rcu_dereference(conf->mirrors[i].rdev);
 		if (rdev && !test_bit(Faulty, &rdev->flags)) {
 			struct request_queue *q = bdev_get_queue(rdev->bdev);
 
+			BUG_ON(!q);
+
 			/* Note the '|| 1' - when read_balance prefers
 			 * non-congested targets, it can be removed
 			 */
@@ -524,7 +522,15 @@ static int raid1_congested(void *data, int bits)
 	rcu_read_unlock();
 	return ret;
 }
+EXPORT_SYMBOL_GPL(md_raid1_congested);
 
+static int raid1_congested(void *data, int bits)
+{
+	mddev_t *mddev = data;
+
+	return mddev_congested(mddev, bits) ||
+		md_raid1_congested(mddev, bits);
+}
 
 static void flush_pending_writes(conf_t *conf)
 {
@@ -1972,6 +1978,8 @@ static int run(mddev_t *mddev)
 		return PTR_ERR(conf);
 
 	list_for_each_entry(rdev, &mddev->disks, same_set) {
+		if (!mddev->gendisk)
+			continue;
 		disk_stack_limits(mddev->gendisk, rdev->bdev,
 				  rdev->data_offset << 9);
 		/* as we don't honour merge_bvec_fn, we must never risk
@@ -2013,8 +2021,10 @@ static int run(mddev_t *mddev)
 
 	md_set_array_sectors(mddev, raid1_size(mddev, 0, 0));
 
-	mddev->queue->backing_dev_info.congested_fn = raid1_congested;
-	mddev->queue->backing_dev_info.congested_data = mddev;
+	if (mddev->queue) {
+		mddev->queue->backing_dev_info.congested_fn = raid1_congested;
+		mddev->queue->backing_dev_info.congested_data = mddev;
+	}
 	return md_integrity_register(mddev);
 }
 
diff --git a/drivers/md/raid1.h b/drivers/md/raid1.h
index 5fc4ca1af863..e743a64fac4f 100644
--- a/drivers/md/raid1.h
+++ b/drivers/md/raid1.h
@@ -126,4 +126,6 @@ struct r1bio_s {
  */
 #define	R1BIO_Returned 6
 
+extern int md_raid1_congested(mddev_t *mddev, int bits);
+
 #endif
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 346e69bfdab3..b72edf35ec54 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -129,7 +129,7 @@ static inline int raid5_dec_bi_hw_segments(struct bio *bio)
 
 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);
+	bio->bi_phys_segments = raid5_bi_phys_segments(bio) | (cnt << 16);
 }
 
 /* Find first data disk in a raid6 stripe */
@@ -514,7 +514,7 @@ static void ops_run_io(struct stripe_head *sh, struct stripe_head_state *s)
 		bi = &sh->dev[i].req;
 
 		bi->bi_rw = rw;
-		if (rw == WRITE)
+		if (rw & WRITE)
 			bi->bi_end_io = raid5_end_write_request;
 		else
 			bi->bi_end_io = raid5_end_read_request;
@@ -548,13 +548,13 @@ static void ops_run_io(struct stripe_head *sh, struct stripe_head_state *s)
 			bi->bi_io_vec[0].bv_offset = 0;
 			bi->bi_size = STRIPE_SIZE;
 			bi->bi_next = NULL;
-			if (rw == WRITE &&
+			if ((rw & WRITE) &&
 			    test_bit(R5_ReWrite, &sh->dev[i].flags))
 				atomic_add(STRIPE_SECTORS,
 					&rdev->corrected_errors);
 			generic_make_request(bi);
 		} else {
-			if (rw == WRITE)
+			if (rw & WRITE)
 				set_bit(STRIPE_DEGRADED, &sh->state);
 			pr_debug("skip op %ld on disc %d for sector %llu\n",
 				bi->bi_rw, i, (unsigned long long)sh->sector);
@@ -585,7 +585,7 @@ async_copy_data(int frombio, struct bio *bio, struct page *page,
 	init_async_submit(&submit, flags, tx, NULL, NULL, NULL);
 
 	bio_for_each_segment(bvl, bio, i) {
-		int len = bio_iovec_idx(bio, i)->bv_len;
+		int len = bvl->bv_len;
 		int clen;
 		int b_offset = 0;
 
@@ -601,8 +601,8 @@ async_copy_data(int frombio, struct bio *bio, struct page *page,
 			clen = len;
 
 		if (clen > 0) {
-			b_offset += bio_iovec_idx(bio, i)->bv_offset;
-			bio_page = bio_iovec_idx(bio, i)->bv_page;
+			b_offset += bvl->bv_offset;
+			bio_page = bvl->bv_page;
 			if (frombio)
 				tx = async_memcpy(page, bio_page, page_offset,
 						  b_offset, clen, &submit);
@@ -4858,7 +4858,7 @@ static raid5_conf_t *setup_conf(mddev_t *mddev)
 			printk(KERN_INFO "md/raid:%s: device %s operational as raid"
 			       " disk %d\n",
 			       mdname(mddev), bdevname(rdev->bdev, b), raid_disk);
-		} else
+		} else if (rdev->saved_raid_disk != raid_disk)
 			/* Cannot rely on bitmap to complete recovery */
 			conf->fullsync = 1;
 	}