Merge Linus' tree to be be to apply submitted patches to newer code than

current trivial.git base

45 files changed, 1167 insertions, 978 deletions

diff --git a/drivers/md/bcache/alloc.c b/drivers/md/bcache/alloc.c
index 443d03fbac47..8eeab72b93e2 100644
--- a/drivers/md/bcache/alloc.c
+++ b/drivers/md/bcache/alloc.c
@@ -331,7 +331,7 @@ static int bch_allocator_thread(void *arg)
 				mutex_unlock(&ca->set->bucket_lock);
 				blkdev_issue_discard(ca->bdev,
 					bucket_to_sector(ca->set, bucket),
-					ca->sb.block_size, GFP_KERNEL, 0);
+					ca->sb.bucket_size, GFP_KERNEL, 0);
 				mutex_lock(&ca->set->bucket_lock);
 			}
 
diff --git a/drivers/md/bcache/bcache.h b/drivers/md/bcache/bcache.h
index d2ebcf323094..04f7bc28ef83 100644
--- a/drivers/md/bcache/bcache.h
+++ b/drivers/md/bcache/bcache.h
@@ -477,9 +477,13 @@ struct gc_stat {
  * CACHE_SET_STOPPING always gets set first when we're closing down a cache set;
  * we'll continue to run normally for awhile with CACHE_SET_STOPPING set (i.e.
  * flushing dirty data).
+ *
+ * CACHE_SET_RUNNING means all cache devices have been registered and journal
+ * replay is complete.
  */
 #define CACHE_SET_UNREGISTERING		0
 #define	CACHE_SET_STOPPING		1
+#define	CACHE_SET_RUNNING		2
 
 struct cache_set {
 	struct closure		cl;
diff --git a/drivers/md/bcache/bset.c b/drivers/md/bcache/bset.c
index 545416415305..646fe85261c1 100644
--- a/drivers/md/bcache/bset.c
+++ b/drivers/md/bcache/bset.c
@@ -1182,7 +1182,7 @@ static void __btree_sort(struct btree_keys *b, struct btree_iter *iter,
 {
 	uint64_t start_time;
 	bool used_mempool = false;
-	struct bset *out = (void *) __get_free_pages(__GFP_NOWARN|GFP_NOIO,
+	struct bset *out = (void *) __get_free_pages(__GFP_NOWARN|GFP_NOWAIT,
 						     order);
 	if (!out) {
 		struct page *outp;
diff --git a/drivers/md/bcache/bset.h b/drivers/md/bcache/bset.h
index 5f6728d5d4dd..ae964624efb2 100644
--- a/drivers/md/bcache/bset.h
+++ b/drivers/md/bcache/bset.h
@@ -453,7 +453,7 @@ static inline bool bch_bkey_equal_header(const struct bkey *l,
 {
 	return (KEY_DIRTY(l) == KEY_DIRTY(r) &&
 		KEY_PTRS(l) == KEY_PTRS(r) &&
-		KEY_CSUM(l) == KEY_CSUM(l));
+		KEY_CSUM(l) == KEY_CSUM(r));
 }
 
 /* Keylists */
diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index 7347b6100961..00cde40db572 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -117,9 +117,9 @@
 ({									\
 	int _r, l = (b)->level - 1;					\
 	bool _w = l <= (op)->lock;					\
-	struct btree *_child = bch_btree_node_get((b)->c, op, key, l, _w);\
+	struct btree *_child = bch_btree_node_get((b)->c, op, key, l,	\
+						  _w, b);		\
 	if (!IS_ERR(_child)) {						\
-		_child->parent = (b);					\
 		_r = bch_btree_ ## fn(_child, op, ##__VA_ARGS__);	\
 		rw_unlock(_w, _child);					\
 	} else								\
@@ -142,7 +142,6 @@
 		rw_lock(_w, _b, _b->level);				\
 		if (_b == (c)->root &&					\
 		    _w == insert_lock(op, _b)) {			\
-			_b->parent = NULL;				\
 			_r = bch_btree_ ## fn(_b, op, ##__VA_ARGS__);	\
 		}							\
 		rw_unlock(_w, _b);					\
@@ -202,7 +201,7 @@ void bch_btree_node_read_done(struct btree *b)
 	struct bset *i = btree_bset_first(b);
 	struct btree_iter *iter;
 
-	iter = mempool_alloc(b->c->fill_iter, GFP_NOWAIT);
+	iter = mempool_alloc(b->c->fill_iter, GFP_NOIO);
 	iter->size = b->c->sb.bucket_size / b->c->sb.block_size;
 	iter->used = 0;
 
@@ -421,7 +420,7 @@ static void do_btree_node_write(struct btree *b)
 	SET_PTR_OFFSET(&k.key, 0, PTR_OFFSET(&k.key, 0) +
 		       bset_sector_offset(&b->keys, i));
 
-	if (!bio_alloc_pages(b->bio, GFP_NOIO)) {
+	if (!bio_alloc_pages(b->bio, __GFP_NOWARN|GFP_NOWAIT)) {
 		int j;
 		struct bio_vec *bv;
 		void *base = (void *) ((unsigned long) i & ~(PAGE_SIZE - 1));
@@ -967,7 +966,8 @@ err:
  * level and op->lock.
  */
 struct btree *bch_btree_node_get(struct cache_set *c, struct btree_op *op,
-				 struct bkey *k, int level, bool write)
+				 struct bkey *k, int level, bool write,
+				 struct btree *parent)
 {
 	int i = 0;
 	struct btree *b;
@@ -1002,6 +1002,7 @@ retry:
 		BUG_ON(b->level != level);
 	}
 
+	b->parent = parent;
 	b->accessed = 1;
 
 	for (; i <= b->keys.nsets && b->keys.set[i].size; i++) {
@@ -1022,15 +1023,16 @@ retry:
 	return b;
 }
 
-static void btree_node_prefetch(struct cache_set *c, struct bkey *k, int level)
+static void btree_node_prefetch(struct btree *parent, struct bkey *k)
 {
 	struct btree *b;
 
-	mutex_lock(&c->bucket_lock);
-	b = mca_alloc(c, NULL, k, level);
-	mutex_unlock(&c->bucket_lock);
+	mutex_lock(&parent->c->bucket_lock);
+	b = mca_alloc(parent->c, NULL, k, parent->level - 1);
+	mutex_unlock(&parent->c->bucket_lock);
 
 	if (!IS_ERR_OR_NULL(b)) {
+		b->parent = parent;
 		bch_btree_node_read(b);
 		rw_unlock(true, b);
 	}
@@ -1060,15 +1062,16 @@ static void btree_node_free(struct btree *b)
 	mutex_unlock(&b->c->bucket_lock);
 }
 
-struct btree *bch_btree_node_alloc(struct cache_set *c, struct btree_op *op,
-				   int level)
+struct btree *__bch_btree_node_alloc(struct cache_set *c, struct btree_op *op,
+				     int level, bool wait,
+				     struct btree *parent)
 {
 	BKEY_PADDED(key) k;
 	struct btree *b = ERR_PTR(-EAGAIN);
 
 	mutex_lock(&c->bucket_lock);
 retry:
-	if (__bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, 1, op != NULL))
+	if (__bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, 1, wait))
 		goto err;
 
 	bkey_put(c, &k.key);
@@ -1085,6 +1088,7 @@ retry:
 	}
 
 	b->accessed = 1;
+	b->parent = parent;
 	bch_bset_init_next(&b->keys, b->keys.set->data, bset_magic(&b->c->sb));
 
 	mutex_unlock(&c->bucket_lock);
@@ -1096,14 +1100,21 @@ err_free:
 err:
 	mutex_unlock(&c->bucket_lock);
 
-	trace_bcache_btree_node_alloc_fail(b);
+	trace_bcache_btree_node_alloc_fail(c);
 	return b;
 }
 
+static struct btree *bch_btree_node_alloc(struct cache_set *c,
+					  struct btree_op *op, int level,
+					  struct btree *parent)
+{
+	return __bch_btree_node_alloc(c, op, level, op != NULL, parent);
+}
+
 static struct btree *btree_node_alloc_replacement(struct btree *b,
 						  struct btree_op *op)
 {
-	struct btree *n = bch_btree_node_alloc(b->c, op, b->level);
+	struct btree *n = bch_btree_node_alloc(b->c, op, b->level, b->parent);
 	if (!IS_ERR_OR_NULL(n)) {
 		mutex_lock(&n->write_lock);
 		bch_btree_sort_into(&b->keys, &n->keys, &b->c->sort);
@@ -1403,6 +1414,7 @@ static int btree_gc_coalesce(struct btree *b, struct btree_op *op,
 	BUG_ON(btree_bset_first(new_nodes[0])->keys);
 	btree_node_free(new_nodes[0]);
 	rw_unlock(true, new_nodes[0]);
+	new_nodes[0] = NULL;
 
 	for (i = 0; i < nodes; i++) {
 		if (__bch_keylist_realloc(&keylist, bkey_u64s(&r[i].b->key)))
@@ -1516,7 +1528,7 @@ static int btree_gc_recurse(struct btree *b, struct btree_op *op,
 		k = bch_btree_iter_next_filter(&iter, &b->keys, bch_ptr_bad);
 		if (k) {
 			r->b = bch_btree_node_get(b->c, op, k, b->level - 1,
-						  true);
+						  true, b);
 			if (IS_ERR(r->b)) {
 				ret = PTR_ERR(r->b);
 				break;
@@ -1811,7 +1823,7 @@ static int bch_btree_check_recurse(struct btree *b, struct btree_op *op)
 			k = bch_btree_iter_next_filter(&iter, &b->keys,
 						       bch_ptr_bad);
 			if (k)
-				btree_node_prefetch(b->c, k, b->level - 1);
+				btree_node_prefetch(b, k);
 
 			if (p)
 				ret = btree(check_recurse, p, b, op);
@@ -1976,12 +1988,12 @@ static int btree_split(struct btree *b, struct btree_op *op,
 
 		trace_bcache_btree_node_split(b, btree_bset_first(n1)->keys);
 
-		n2 = bch_btree_node_alloc(b->c, op, b->level);
+		n2 = bch_btree_node_alloc(b->c, op, b->level, b->parent);
 		if (IS_ERR(n2))
 			goto err_free1;
 
 		if (!b->parent) {
-			n3 = bch_btree_node_alloc(b->c, op, b->level + 1);
+			n3 = bch_btree_node_alloc(b->c, op, b->level + 1, NULL);
 			if (IS_ERR(n3))
 				goto err_free2;
 		}
diff --git a/drivers/md/bcache/btree.h b/drivers/md/bcache/btree.h
index 91dfa5e69685..5c391fa01bed 100644
--- a/drivers/md/bcache/btree.h
+++ b/drivers/md/bcache/btree.h
@@ -242,9 +242,10 @@ void __bch_btree_node_write(struct btree *, struct closure *);
 void bch_btree_node_write(struct btree *, struct closure *);
 
 void bch_btree_set_root(struct btree *);
-struct btree *bch_btree_node_alloc(struct cache_set *, struct btree_op *, int);
+struct btree *__bch_btree_node_alloc(struct cache_set *, struct btree_op *,
+				     int, bool, struct btree *);
 struct btree *bch_btree_node_get(struct cache_set *, struct btree_op *,
-				 struct bkey *, int, bool);
+				 struct bkey *, int, bool, struct btree *);
 
 int bch_btree_insert_check_key(struct btree *, struct btree_op *,
 			       struct bkey *);
diff --git a/drivers/md/bcache/extents.c b/drivers/md/bcache/extents.c
index 3a0de4cf9771..243de0bf15cd 100644
--- a/drivers/md/bcache/extents.c
+++ b/drivers/md/bcache/extents.c
@@ -474,9 +474,8 @@ out:
 	return false;
 }
 
-static bool bch_extent_invalid(struct btree_keys *bk, const struct bkey *k)
+bool __bch_extent_invalid(struct cache_set *c, const struct bkey *k)
 {
-	struct btree *b = container_of(bk, struct btree, keys);
 	char buf[80];
 
 	if (!KEY_SIZE(k))
@@ -485,16 +484,22 @@ static bool bch_extent_invalid(struct btree_keys *bk, const struct bkey *k)
 	if (KEY_SIZE(k) > KEY_OFFSET(k))
 		goto bad;
 
-	if (__ptr_invalid(b->c, k))
+	if (__ptr_invalid(c, k))
 		goto bad;
 
 	return false;
 bad:
 	bch_extent_to_text(buf, sizeof(buf), k);
-	cache_bug(b->c, "spotted extent %s: %s", buf, bch_ptr_status(b->c, k));
+	cache_bug(c, "spotted extent %s: %s", buf, bch_ptr_status(c, k));
 	return true;
 }
 
+static bool bch_extent_invalid(struct btree_keys *bk, const struct bkey *k)
+{
+	struct btree *b = container_of(bk, struct btree, keys);
+	return __bch_extent_invalid(b->c, k);
+}
+
 static bool bch_extent_bad_expensive(struct btree *b, const struct bkey *k,
 				     unsigned ptr)
 {
diff --git a/drivers/md/bcache/extents.h b/drivers/md/bcache/extents.h
index e4e23409782d..e2ed54054e7a 100644
--- a/drivers/md/bcache/extents.h
+++ b/drivers/md/bcache/extents.h
@@ -9,5 +9,6 @@ struct cache_set;
 
 void bch_extent_to_text(char *, size_t, const struct bkey *);
 bool __bch_btree_ptr_invalid(struct cache_set *, const struct bkey *);
+bool __bch_extent_invalid(struct cache_set *, const struct bkey *);
 
 #endif /* _BCACHE_EXTENTS_H */
diff --git a/drivers/md/bcache/journal.c b/drivers/md/bcache/journal.c
index 59e82021b5bb..fe080ad0e558 100644
--- a/drivers/md/bcache/journal.c
+++ b/drivers/md/bcache/journal.c
@@ -7,6 +7,7 @@
 #include "bcache.h"
 #include "btree.h"
 #include "debug.h"
+#include "extents.h"
 
 #include <trace/events/bcache.h>
 
@@ -189,11 +190,15 @@ int bch_journal_read(struct cache_set *c, struct list_head *list)
 			if (read_bucket(l))
 				goto bsearch;
 
-		if (list_empty(list))
+		/* no journal entries on this device? */
+		if (l == ca->sb.njournal_buckets)
 			continue;
 bsearch:
+		BUG_ON(list_empty(list));
+
 		/* Binary search */
-		m = r = find_next_bit(bitmap, ca->sb.njournal_buckets, l + 1);
+		m = l;
+		r = find_next_bit(bitmap, ca->sb.njournal_buckets, l + 1);
 		pr_debug("starting binary search, l %u r %u", l, r);
 
 		while (l + 1 < r) {
@@ -291,15 +296,16 @@ void bch_journal_mark(struct cache_set *c, struct list_head *list)
 
 		for (k = i->j.start;
 		     k < bset_bkey_last(&i->j);
-		     k = bkey_next(k)) {
-			unsigned j;
+		     k = bkey_next(k))
+			if (!__bch_extent_invalid(c, k)) {
+				unsigned j;
 
-			for (j = 0; j < KEY_PTRS(k); j++)
-				if (ptr_available(c, k, j))
-					atomic_inc(&PTR_BUCKET(c, k, j)->pin);
+				for (j = 0; j < KEY_PTRS(k); j++)
+					if (ptr_available(c, k, j))
+						atomic_inc(&PTR_BUCKET(c, k, j)->pin);
 
-			bch_initial_mark_key(c, 0, k);
-		}
+				bch_initial_mark_key(c, 0, k);
+			}
 	}
 }
 
diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
index 15fff4f68a7c..62e6e98186b5 100644
--- a/drivers/md/bcache/request.c
+++ b/drivers/md/bcache/request.c
@@ -311,7 +311,8 @@ void bch_data_insert(struct closure *cl)
 {
 	struct data_insert_op *op = container_of(cl, struct data_insert_op, cl);
 
-	trace_bcache_write(op->bio, op->writeback, op->bypass);
+	trace_bcache_write(op->c, op->inode, op->bio,
+			   op->writeback, op->bypass);
 
 	bch_keylist_init(&op->insert_keys);
 	bio_get(op->bio);
diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 926ded8ccbf5..4dd2bb7167f0 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -733,8 +733,6 @@ static void bcache_device_detach(struct bcache_device *d)
 static void bcache_device_attach(struct bcache_device *d, struct cache_set *c,
 				 unsigned id)
 {
-	BUG_ON(test_bit(CACHE_SET_STOPPING, &c->flags));
-
 	d->id = id;
 	d->c = c;
 	c->devices[id] = d;
@@ -844,6 +842,7 @@ static int bcache_device_init(struct bcache_device *d, unsigned block_size,
 	q->limits.logical_block_size	= block_size;
 	q->limits.physical_block_size	= block_size;
 	set_bit(QUEUE_FLAG_NONROT,	&d->disk->queue->queue_flags);
+	clear_bit(QUEUE_FLAG_ADD_RANDOM, &d->disk->queue->queue_flags);
 	set_bit(QUEUE_FLAG_DISCARD,	&d->disk->queue->queue_flags);
 
 	blk_queue_flush(q, REQ_FLUSH|REQ_FUA);
@@ -927,6 +926,7 @@ static void cached_dev_detach_finish(struct work_struct *w)
 	list_move(&dc->list, &uncached_devices);
 
 	clear_bit(BCACHE_DEV_DETACHING, &dc->disk.flags);
+	clear_bit(BCACHE_DEV_UNLINK_DONE, &dc->disk.flags);
 
 	mutex_unlock(&bch_register_lock);
 
@@ -1041,6 +1041,9 @@ int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c)
 	 */
 	atomic_set(&dc->count, 1);
 
+	if (bch_cached_dev_writeback_start(dc))
+		return -ENOMEM;
+
 	if (BDEV_STATE(&dc->sb) == BDEV_STATE_DIRTY) {
 		bch_sectors_dirty_init(dc);
 		atomic_set(&dc->has_dirty, 1);
@@ -1070,7 +1073,8 @@ static void cached_dev_free(struct closure *cl)
 	struct cached_dev *dc = container_of(cl, struct cached_dev, disk.cl);
 
 	cancel_delayed_work_sync(&dc->writeback_rate_update);
-	kthread_stop(dc->writeback_thread);
+	if (!IS_ERR_OR_NULL(dc->writeback_thread))
+		kthread_stop(dc->writeback_thread);
 
 	mutex_lock(&bch_register_lock);
 
@@ -1081,12 +1085,8 @@ static void cached_dev_free(struct closure *cl)
 
 	mutex_unlock(&bch_register_lock);
 
-	if (!IS_ERR_OR_NULL(dc->bdev)) {
-		if (dc->bdev->bd_disk)
-			blk_sync_queue(bdev_get_queue(dc->bdev));
-
+	if (!IS_ERR_OR_NULL(dc->bdev))
 		blkdev_put(dc->bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
-	}
 
 	wake_up(&unregister_wait);
 
@@ -1213,7 +1213,9 @@ void bch_flash_dev_release(struct kobject *kobj)
 static void flash_dev_free(struct closure *cl)
 {
 	struct bcache_device *d = container_of(cl, struct bcache_device, cl);
+	mutex_lock(&bch_register_lock);
 	bcache_device_free(d);
+	mutex_unlock(&bch_register_lock);
 	kobject_put(&d->kobj);
 }
 
@@ -1221,7 +1223,9 @@ static void flash_dev_flush(struct closure *cl)
 {
 	struct bcache_device *d = container_of(cl, struct bcache_device, cl);
 
+	mutex_lock(&bch_register_lock);
 	bcache_device_unlink(d);
+	mutex_unlock(&bch_register_lock);
 	kobject_del(&d->kobj);
 	continue_at(cl, flash_dev_free, system_wq);
 }
@@ -1277,6 +1281,9 @@ int bch_flash_dev_create(struct cache_set *c, uint64_t size)
 	if (test_bit(CACHE_SET_STOPPING, &c->flags))
 		return -EINTR;
 
+	if (!test_bit(CACHE_SET_RUNNING, &c->flags))
+		return -EPERM;
+
 	u = uuid_find_empty(c);
 	if (!u) {
 		pr_err("Can't create volume, no room for UUID");
@@ -1346,8 +1353,11 @@ static void cache_set_free(struct closure *cl)
 	bch_journal_free(c);
 
 	for_each_cache(ca, c, i)
-		if (ca)
+		if (ca) {
+			ca->set = NULL;
+			c->cache[ca->sb.nr_this_dev] = NULL;
 			kobject_put(&ca->kobj);
+		}
 
 	bch_bset_sort_state_free(&c->sort);
 	free_pages((unsigned long) c->uuids, ilog2(bucket_pages(c)));
@@ -1405,9 +1415,11 @@ static void cache_set_flush(struct closure *cl)
 		if (ca->alloc_thread)
 			kthread_stop(ca->alloc_thread);
 
-	cancel_delayed_work_sync(&c->journal.work);
-	/* flush last journal entry if needed */
-	c->journal.work.work.func(&c->journal.work.work);
+	if (c->journal.cur) {
+		cancel_delayed_work_sync(&c->journal.work);
+		/* flush last journal entry if needed */
+		c->journal.work.work.func(&c->journal.work.work);
+	}
 
 	closure_return(cl);
 }
@@ -1586,7 +1598,7 @@ static void run_cache_set(struct cache_set *c)
 			goto err;
 
 		err = "error reading btree root";
-		c->root = bch_btree_node_get(c, NULL, k, j->btree_level, true);
+		c->root = bch_btree_node_get(c, NULL, k, j->btree_level, true, NULL);
 		if (IS_ERR_OR_NULL(c->root))
 			goto err;
 
@@ -1661,7 +1673,7 @@ static void run_cache_set(struct cache_set *c)
 			goto err;
 
 		err = "cannot allocate new btree root";
-		c->root = bch_btree_node_alloc(c, NULL, 0);
+		c->root = __bch_btree_node_alloc(c, NULL, 0, true, NULL);
 		if (IS_ERR_OR_NULL(c->root))
 			goto err;
 
@@ -1697,6 +1709,7 @@ static void run_cache_set(struct cache_set *c)
 
 	flash_devs_run(c);
 
+	set_bit(CACHE_SET_RUNNING, &c->flags);
 	return;
 err:
 	closure_sync(&cl);
@@ -1760,6 +1773,7 @@ found:
 		pr_debug("set version = %llu", c->sb.version);
 	}
 
+	kobject_get(&ca->kobj);
 	ca->set = c;
 	ca->set->cache[ca->sb.nr_this_dev] = ca;
 	c->cache_by_alloc[c->caches_loaded++] = ca;
@@ -1780,8 +1794,10 @@ void bch_cache_release(struct kobject *kobj)
 	struct cache *ca = container_of(kobj, struct cache, kobj);
 	unsigned i;
 
-	if (ca->set)
+	if (ca->set) {
+		BUG_ON(ca->set->cache[ca->sb.nr_this_dev] != ca);
 		ca->set->cache[ca->sb.nr_this_dev] = NULL;
+	}
 
 	bio_split_pool_free(&ca->bio_split_hook);
 
@@ -1798,10 +1814,8 @@ void bch_cache_release(struct kobject *kobj)
 	if (ca->sb_bio.bi_inline_vecs[0].bv_page)
 		put_page(ca->sb_bio.bi_io_vec[0].bv_page);
 
-	if (!IS_ERR_OR_NULL(ca->bdev)) {
-		blk_sync_queue(bdev_get_queue(ca->bdev));
+	if (!IS_ERR_OR_NULL(ca->bdev))
 		blkdev_put(ca->bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
-	}
 
 	kfree(ca);
 	module_put(THIS_MODULE);
@@ -1844,7 +1858,7 @@ static int cache_alloc(struct cache_sb *sb, struct cache *ca)
 }
 
 static void register_cache(struct cache_sb *sb, struct page *sb_page,
-				  struct block_device *bdev, struct cache *ca)
+				struct block_device *bdev, struct cache *ca)
 {
 	char name[BDEVNAME_SIZE];
 	const char *err = "cannot allocate memory";
@@ -1877,10 +1891,12 @@ static void register_cache(struct cache_sb *sb, struct page *sb_page,
 		goto err;
 
 	pr_info("registered cache device %s", bdevname(bdev, name));
+out:
+	kobject_put(&ca->kobj);
 	return;
 err:
 	pr_notice("error opening %s: %s", bdevname(bdev, name), err);
-	kobject_put(&ca->kobj);
+	goto out;
 }
 
 /* Global interfaces/init */
@@ -1945,10 +1961,12 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 	if (IS_ERR(bdev)) {
 		if (bdev == ERR_PTR(-EBUSY)) {
 			bdev = lookup_bdev(strim(path));
+			mutex_lock(&bch_register_lock);
 			if (!IS_ERR(bdev) && bch_is_open(bdev))
 				err = "device already registered";
 			else
 				err = "device busy";
+			mutex_unlock(&bch_register_lock);
 		}
 		goto err;
 	}
diff --git a/drivers/md/bcache/util.h b/drivers/md/bcache/util.h
index ac7d0d1f70d7..98df7572b5f7 100644
--- a/drivers/md/bcache/util.h
+++ b/drivers/md/bcache/util.h
@@ -416,8 +416,8 @@ do {									\
 			  average_frequency,	frequency_units);	\
 	__print_time_stat(stats, name,					\
 			  average_duration,	duration_units);	\
-	__print_time_stat(stats, name,					\
-			  max_duration,		duration_units);	\
+	sysfs_print(name ## _ ##max_duration ## _ ## duration_units,	\
+			div_u64((stats)->max_duration, NSEC_PER_ ## duration_units));\
 									\
 	sysfs_print(name ## _last_ ## frequency_units, (stats)->last	\
 		    ? div_s64(local_clock() - (stats)->last,		\
diff --git a/drivers/md/bcache/writeback.c b/drivers/md/bcache/writeback.c
index f4300e4c0114..f1986bcd1bf0 100644
--- a/drivers/md/bcache/writeback.c
+++ b/drivers/md/bcache/writeback.c
@@ -239,7 +239,7 @@ static void read_dirty(struct cached_dev *dc)
 		if (KEY_START(&w->key) != dc->last_read ||
 		    jiffies_to_msecs(delay) > 50)
 			while (!kthread_should_stop() && delay)
-				delay = schedule_timeout_uninterruptible(delay);
+				delay = schedule_timeout_interruptible(delay);
 
 		dc->last_read	= KEY_OFFSET(&w->key);
 
@@ -436,7 +436,7 @@ static int bch_writeback_thread(void *arg)
 			while (delay &&
 			       !kthread_should_stop() &&
 			       !test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags))
-				delay = schedule_timeout_uninterruptible(delay);
+				delay = schedule_timeout_interruptible(delay);
 		}
 	}
 
@@ -478,7 +478,7 @@ void bch_sectors_dirty_init(struct cached_dev *dc)
 	dc->disk.sectors_dirty_last = bcache_dev_sectors_dirty(&dc->disk);
 }
 
-int bch_cached_dev_writeback_init(struct cached_dev *dc)
+void bch_cached_dev_writeback_init(struct cached_dev *dc)
 {
 	sema_init(&dc->in_flight, 64);
 	init_rwsem(&dc->writeback_lock);
@@ -494,14 +494,20 @@ int bch_cached_dev_writeback_init(struct cached_dev *dc)
 	dc->writeback_rate_d_term	= 30;
 	dc->writeback_rate_p_term_inverse = 6000;
 
+	INIT_DELAYED_WORK(&dc->writeback_rate_update, update_writeback_rate);
+}
+
+int bch_cached_dev_writeback_start(struct cached_dev *dc)
+{
 	dc->writeback_thread = kthread_create(bch_writeback_thread, dc,
 					      "bcache_writeback");
 	if (IS_ERR(dc->writeback_thread))
 		return PTR_ERR(dc->writeback_thread);
 
-	INIT_DELAYED_WORK(&dc->writeback_rate_update, update_writeback_rate);
 	schedule_delayed_work(&dc->writeback_rate_update,
 			      dc->writeback_rate_update_seconds * HZ);
 
+	bch_writeback_queue(dc);
+
 	return 0;
 }
diff --git a/drivers/md/bcache/writeback.h b/drivers/md/bcache/writeback.h
index e2f8598937ac..0a9dab187b79 100644
--- a/drivers/md/bcache/writeback.h
+++ b/drivers/md/bcache/writeback.h
@@ -85,6 +85,7 @@ static inline void bch_writeback_add(struct cached_dev *dc)
 void bcache_dev_sectors_dirty_add(struct cache_set *, unsigned, uint64_t, int);
 
 void bch_sectors_dirty_init(struct cached_dev *dc);
-int bch_cached_dev_writeback_init(struct cached_dev *);
+void bch_cached_dev_writeback_init(struct cached_dev *);
+int bch_cached_dev_writeback_start(struct cached_dev *);
 
 #endif
diff --git a/drivers/md/bitmap.c b/drivers/md/bitmap.c
index 67f8b31e2054..da3604e73e8a 100644
--- a/drivers/md/bitmap.c
+++ b/drivers/md/bitmap.c
@@ -879,7 +879,6 @@ void bitmap_unplug(struct bitmap *bitmap)
 {
 	unsigned long i;
 	int dirty, need_write;
-	int wait = 0;
 
 	if (!bitmap || !bitmap->storage.filemap ||
 	    test_bit(BITMAP_STALE, &bitmap->flags))
@@ -897,16 +896,13 @@ void bitmap_unplug(struct bitmap *bitmap)
 			clear_page_attr(bitmap, i, BITMAP_PAGE_PENDING);
 			write_page(bitmap, bitmap->storage.filemap[i], 0);
 		}
-		if (dirty)
-			wait = 1;
-	}
-	if (wait) { /* if any writes were performed, we need to wait on them */
-		if (bitmap->storage.file)
-			wait_event(bitmap->write_wait,
-				   atomic_read(&bitmap->pending_writes)==0);
-		else
-			md_super_wait(bitmap->mddev);
 	}
+	if (bitmap->storage.file)
+		wait_event(bitmap->write_wait,
+			   atomic_read(&bitmap->pending_writes)==0);
+	else
+		md_super_wait(bitmap->mddev);
+
 	if (test_bit(BITMAP_WRITE_ERROR, &bitmap->flags))
 		bitmap_file_kick(bitmap);
 }
diff --git a/drivers/md/dm-bufio.c b/drivers/md/dm-bufio.c
index ab472c557d18..afe79719ea32 100644
--- a/drivers/md/dm-bufio.c
+++ b/drivers/md/dm-bufio.c
@@ -465,6 +465,7 @@ static void __relink_lru(struct dm_buffer *b, int dirty)
 	c->n_buffers[dirty]++;
 	b->list_mode = dirty;
 	list_move(&b->lru_list, &c->lru[dirty]);
+	b->last_accessed = jiffies;
 }
 
 /*----------------------------------------------------------------
@@ -720,7 +721,6 @@ static void __wait_for_free_buffer(struct dm_bufio_client *c)
 
 	io_schedule();
 
-	set_task_state(current, TASK_RUNNING);
 	remove_wait_queue(&c->free_buffer_wait, &wait);
 
 	dm_bufio_lock(c);
@@ -1434,9 +1434,9 @@ static void drop_buffers(struct dm_bufio_client *c)
 
 /*
  * Test if the buffer is unused and too old, and commit it.
- * At if noio is set, we must not do any I/O because we hold
- * dm_bufio_clients_lock and we would risk deadlock if the I/O gets rerouted to
- * different bufio client.
+ * And if GFP_NOFS is used, we must not do any I/O because we hold
+ * dm_bufio_clients_lock and we would risk deadlock if the I/O gets
+ * rerouted to different bufio client.
  */
 static int __cleanup_old_buffer(struct dm_buffer *b, gfp_t gfp,
 				unsigned long max_jiffies)
@@ -1444,7 +1444,7 @@ static int __cleanup_old_buffer(struct dm_buffer *b, gfp_t gfp,
 	if (jiffies - b->last_accessed < max_jiffies)
 		return 0;
 
-	if (!(gfp & __GFP_IO)) {
+	if (!(gfp & __GFP_FS)) {
 		if (test_bit(B_READING, &b->state) ||
 		    test_bit(B_WRITING, &b->state) ||
 		    test_bit(B_DIRTY, &b->state))
@@ -1472,9 +1472,9 @@ static long __scan(struct dm_bufio_client *c, unsigned long nr_to_scan,
 		list_for_each_entry_safe_reverse(b, tmp, &c->lru[l], lru_list) {
 			freed += __cleanup_old_buffer(b, gfp_mask, 0);
 			if (!--nr_to_scan)
-				break;
+				return freed;
+			dm_bufio_cond_resched();
 		}
-		dm_bufio_cond_resched();
 	}
 	return freed;
 }
@@ -1486,7 +1486,7 @@ dm_bufio_shrink_scan(struct shrinker *shrink, struct shrink_control *sc)
 	unsigned long freed;
 
 	c = container_of(shrink, struct dm_bufio_client, shrinker);
-	if (sc->gfp_mask & __GFP_IO)
+	if (sc->gfp_mask & __GFP_FS)
 		dm_bufio_lock(c);
 	else if (!dm_bufio_trylock(c))
 		return SHRINK_STOP;
@@ -1503,7 +1503,7 @@ dm_bufio_shrink_count(struct shrinker *shrink, struct shrink_control *sc)
 	unsigned long count;
 
 	c = container_of(shrink, struct dm_bufio_client, shrinker);
-	if (sc->gfp_mask & __GFP_IO)
+	if (sc->gfp_mask & __GFP_FS)
 		dm_bufio_lock(c);
 	else if (!dm_bufio_trylock(c))
 		return 0;
diff --git a/drivers/md/dm-cache-metadata.c b/drivers/md/dm-cache-metadata.c
index d2899e7eb3aa..06709257adde 100644
--- a/drivers/md/dm-cache-metadata.c
+++ b/drivers/md/dm-cache-metadata.c
@@ -330,7 +330,7 @@ static int __write_initial_superblock(struct dm_cache_metadata *cmd)
 	disk_super->discard_root = cpu_to_le64(cmd->discard_root);
 	disk_super->discard_block_size = cpu_to_le64(cmd->discard_block_size);
 	disk_super->discard_nr_blocks = cpu_to_le64(from_oblock(cmd->discard_nr_blocks));
-	disk_super->metadata_block_size = cpu_to_le32(DM_CACHE_METADATA_BLOCK_SIZE >> SECTOR_SHIFT);
+	disk_super->metadata_block_size = cpu_to_le32(DM_CACHE_METADATA_BLOCK_SIZE);
 	disk_super->data_block_size = cpu_to_le32(cmd->data_block_size);
 	disk_super->cache_blocks = cpu_to_le32(0);
 
@@ -478,7 +478,7 @@ static int __create_persistent_data_objects(struct dm_cache_metadata *cmd,
 					    bool may_format_device)
 {
 	int r;
-	cmd->bm = dm_block_manager_create(cmd->bdev, DM_CACHE_METADATA_BLOCK_SIZE,
+	cmd->bm = dm_block_manager_create(cmd->bdev, DM_CACHE_METADATA_BLOCK_SIZE << SECTOR_SHIFT,
 					  CACHE_METADATA_CACHE_SIZE,
 					  CACHE_MAX_CONCURRENT_LOCKS);
 	if (IS_ERR(cmd->bm)) {
diff --git a/drivers/md/dm-cache-metadata.h b/drivers/md/dm-cache-metadata.h
index cd70a78623a3..7383c90ccdb8 100644
--- a/drivers/md/dm-cache-metadata.h
+++ b/drivers/md/dm-cache-metadata.h
@@ -9,19 +9,17 @@
 
 #include "dm-cache-block-types.h"
 #include "dm-cache-policy-internal.h"
+#include "persistent-data/dm-space-map-metadata.h"
 
 /*----------------------------------------------------------------*/
 
-#define DM_CACHE_METADATA_BLOCK_SIZE 4096
+#define DM_CACHE_METADATA_BLOCK_SIZE DM_SM_METADATA_BLOCK_SIZE
 
 /* FIXME: remove this restriction */
 /*
  * The metadata device is currently limited in size.
- *
- * We have one block of index, which can hold 255 index entries.  Each
- * index entry contains allocation info about 16k metadata blocks.
  */
-#define DM_CACHE_METADATA_MAX_SECTORS (255 * (1 << 14) * (DM_CACHE_METADATA_BLOCK_SIZE / (1 << SECTOR_SHIFT)))
+#define DM_CACHE_METADATA_MAX_SECTORS DM_SM_METADATA_MAX_SECTORS
 
 /*
  * A metadata device larger than 16GB triggers a warning.
diff --git a/drivers/md/dm-cache-target.c b/drivers/md/dm-cache-target.c
index 2c63326638b6..7130505c2425 100644
--- a/drivers/md/dm-cache-target.c
+++ b/drivers/md/dm-cache-target.c
@@ -718,6 +718,22 @@ static int bio_triggers_commit(struct cache *cache, struct bio *bio)
 	return bio->bi_rw & (REQ_FLUSH | REQ_FUA);
 }
 
+/*
+ * You must increment the deferred set whilst the prison cell is held.  To
+ * encourage this, we ask for 'cell' to be passed in.
+ */
+static void inc_ds(struct cache *cache, struct bio *bio,
+		   struct dm_bio_prison_cell *cell)
+{
+	size_t pb_data_size = get_per_bio_data_size(cache);
+	struct per_bio_data *pb = get_per_bio_data(bio, pb_data_size);
+
+	BUG_ON(!cell);
+	BUG_ON(pb->all_io_entry);
+
+	pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
+}
+
 static void issue(struct cache *cache, struct bio *bio)
 {
 	unsigned long flags;
@@ -737,6 +753,12 @@ static void issue(struct cache *cache, struct bio *bio)
 	spin_unlock_irqrestore(&cache->lock, flags);
 }
 
+static void inc_and_issue(struct cache *cache, struct bio *bio, struct dm_bio_prison_cell *cell)
+{
+	inc_ds(cache, bio, cell);
+	issue(cache, bio);
+}
+
 static void defer_writethrough_bio(struct cache *cache, struct bio *bio)
 {
 	unsigned long flags;
@@ -873,8 +895,8 @@ static void migration_success_pre_commit(struct dm_cache_migration *mg)
 	struct cache *cache = mg->cache;
 
 	if (mg->writeback) {
-		cell_defer(cache, mg->old_ocell, false);
 		clear_dirty(cache, mg->old_oblock, mg->cblock);
+		cell_defer(cache, mg->old_ocell, false);
 		cleanup_migration(mg);
 		return;
 
@@ -929,13 +951,13 @@ static void migration_success_post_commit(struct dm_cache_migration *mg)
 		}
 
 	} else {
+		clear_dirty(cache, mg->new_oblock, mg->cblock);
 		if (mg->requeue_holder)
 			cell_defer(cache, mg->new_ocell, true);
 		else {
 			bio_endio(mg->new_ocell->holder, 0);
 			cell_defer(cache, mg->new_ocell, false);
 		}
-		clear_dirty(cache, mg->new_oblock, mg->cblock);
 		cleanup_migration(mg);
 	}
 }
@@ -1015,6 +1037,11 @@ static void issue_overwrite(struct dm_cache_migration *mg, struct bio *bio)
 
 	dm_hook_bio(&pb->hook_info, bio, overwrite_endio, mg);
 	remap_to_cache_dirty(mg->cache, bio, mg->new_oblock, mg->cblock);
+
+	/*
+	 * No need to inc_ds() here, since the cell will be held for the
+	 * duration of the io.
+	 */
 	generic_make_request(bio);
 }
 
@@ -1115,8 +1142,7 @@ static void check_for_quiesced_migrations(struct cache *cache,
 		return;
 
 	INIT_LIST_HEAD(&work);
-	if (pb->all_io_entry)
-		dm_deferred_entry_dec(pb->all_io_entry, &work);
+	dm_deferred_entry_dec(pb->all_io_entry, &work);
 
 	if (!list_empty(&work))
 		queue_quiesced_migrations(cache, &work);
@@ -1252,6 +1278,11 @@ static void process_flush_bio(struct cache *cache, struct bio *bio)
 	else
 		remap_to_cache(cache, bio, 0);
 
+	/*
+	 * REQ_FLUSH is not directed at any particular block so we don't
+	 * need to inc_ds().  REQ_FUA's are split into a write + REQ_FLUSH
+	 * by dm-core.
+	 */
 	issue(cache, bio);
 }
 
@@ -1301,15 +1332,6 @@ static void inc_miss_counter(struct cache *cache, struct bio *bio)
 		   &cache->stats.read_miss : &cache->stats.write_miss);
 }
 
-static void issue_cache_bio(struct cache *cache, struct bio *bio,
-			    struct per_bio_data *pb,
-			    dm_oblock_t oblock, dm_cblock_t cblock)
-{
-	pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
-	remap_to_cache_dirty(cache, bio, oblock, cblock);
-	issue(cache, bio);
-}
-
 static void process_bio(struct cache *cache, struct prealloc *structs,
 			struct bio *bio)
 {
@@ -1318,8 +1340,6 @@ static void process_bio(struct cache *cache, struct prealloc *structs,
 	dm_oblock_t block = get_bio_block(cache, bio);
 	struct dm_bio_prison_cell *cell_prealloc, *old_ocell, *new_ocell;
 	struct policy_result lookup_result;
-	size_t pb_data_size = get_per_bio_data_size(cache);
-	struct per_bio_data *pb = get_per_bio_data(bio, pb_data_size);
 	bool discarded_block = is_discarded_oblock(cache, block);
 	bool passthrough = passthrough_mode(&cache->features);
 	bool can_migrate = !passthrough && (discarded_block || spare_migration_bandwidth(cache));
@@ -1359,9 +1379,8 @@ static void process_bio(struct cache *cache, struct prealloc *structs,
 
 			} else {
 				/* FIXME: factor out issue_origin() */
-				pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
 				remap_to_origin_clear_discard(cache, bio, block);
-				issue(cache, bio);
+				inc_and_issue(cache, bio, new_ocell);
 			}
 		} else {
 			inc_hit_counter(cache, bio);
@@ -1369,20 +1388,21 @@ static void process_bio(struct cache *cache, struct prealloc *structs,
 			if (bio_data_dir(bio) == WRITE &&
 			    writethrough_mode(&cache->features) &&
 			    !is_dirty(cache, lookup_result.cblock)) {
-				pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
 				remap_to_origin_then_cache(cache, bio, block, lookup_result.cblock);
-				issue(cache, bio);
-			} else
-				issue_cache_bio(cache, bio, pb, block, lookup_result.cblock);
+				inc_and_issue(cache, bio, new_ocell);
+
+			} else  {
+				remap_to_cache_dirty(cache, bio, block, lookup_result.cblock);
+				inc_and_issue(cache, bio, new_ocell);
+			}
 		}
 
 		break;
 
 	case POLICY_MISS:
 		inc_miss_counter(cache, bio);
-		pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
 		remap_to_origin_clear_discard(cache, bio, block);
-		issue(cache, bio);
+		inc_and_issue(cache, bio, new_ocell);
 		break;
 
 	case POLICY_NEW:
@@ -1501,6 +1521,9 @@ static void process_deferred_flush_bios(struct cache *cache, bool submit_bios)
 	bio_list_init(&cache->deferred_flush_bios);
 	spin_unlock_irqrestore(&cache->lock, flags);
 
+	/*
+	 * These bios have already been through inc_ds()
+	 */
 	while ((bio = bio_list_pop(&bios)))
 		submit_bios ? generic_make_request(bio) : bio_io_error(bio);
 }
@@ -1518,6 +1541,9 @@ static void process_deferred_writethrough_bios(struct cache *cache)
 	bio_list_init(&cache->deferred_writethrough_bios);
 	spin_unlock_irqrestore(&cache->lock, flags);
 
+	/*
+	 * These bios have already been through inc_ds()
+	 */
 	while ((bio = bio_list_pop(&bios)))
 		generic_make_request(bio);
 }
@@ -1694,6 +1720,7 @@ static void do_worker(struct work_struct *ws)
 
 		if (commit_if_needed(cache)) {
 			process_deferred_flush_bios(cache, false);
+			process_migrations(cache, &cache->need_commit_migrations, migration_failure);
 
 			/*
 			 * FIXME: rollback metadata or just go into a
@@ -2406,16 +2433,13 @@ out:
 	return r;
 }
 
-static int cache_map(struct dm_target *ti, struct bio *bio)
+static int __cache_map(struct cache *cache, struct bio *bio, struct dm_bio_prison_cell **cell)
 {
-	struct cache *cache = ti->private;
-
 	int r;
 	dm_oblock_t block = get_bio_block(cache, bio);
 	size_t pb_data_size = get_per_bio_data_size(cache);
 	bool can_migrate = false;
 	bool discarded_block;
-	struct dm_bio_prison_cell *cell;
 	struct policy_result lookup_result;
 	struct per_bio_data *pb = init_per_bio_data(bio, pb_data_size);
 
@@ -2437,15 +2461,15 @@ static int cache_map(struct dm_target *ti, struct bio *bio)
 	/*
 	 * Check to see if that block is currently migrating.
 	 */
-	cell = alloc_prison_cell(cache);
-	if (!cell) {
+	*cell = alloc_prison_cell(cache);
+	if (!*cell) {
 		defer_bio(cache, bio);
 		return DM_MAPIO_SUBMITTED;
 	}
 
-	r = bio_detain(cache, block, bio, cell,
+	r = bio_detain(cache, block, bio, *cell,
 		       (cell_free_fn) free_prison_cell,
-		       cache, &cell);
+		       cache, cell);
 	if (r) {
 		if (r < 0)
 			defer_bio(cache, bio);
@@ -2458,11 +2482,12 @@ static int cache_map(struct dm_target *ti, struct bio *bio)
 	r = policy_map(cache->policy, block, false, can_migrate, discarded_block,
 		       bio, &lookup_result);
 	if (r == -EWOULDBLOCK) {
-		cell_defer(cache, cell, true);
+		cell_defer(cache, *cell, true);
 		return DM_MAPIO_SUBMITTED;
 
 	} else if (r) {
 		DMERR_LIMIT("Unexpected return from cache replacement policy: %d", r);
+		cell_defer(cache, *cell, false);
 		bio_io_error(bio);
 		return DM_MAPIO_SUBMITTED;
 	}
@@ -2476,52 +2501,44 @@ static int cache_map(struct dm_target *ti, struct bio *bio)
 				 * We need to invalidate this block, so
 				 * defer for the worker thread.
 				 */
-				cell_defer(cache, cell, true);
+				cell_defer(cache, *cell, true);
 				r = DM_MAPIO_SUBMITTED;
 
 			} else {
-				pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
 				inc_miss_counter(cache, bio);
 				remap_to_origin_clear_discard(cache, bio, block);
-
-				cell_defer(cache, cell, false);
 			}
 
 		} else {
 			inc_hit_counter(cache, bio);
-			pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
-
 			if (bio_data_dir(bio) == WRITE && writethrough_mode(&cache->features) &&
 			    !is_dirty(cache, lookup_result.cblock))
 				remap_to_origin_then_cache(cache, bio, block, lookup_result.cblock);
 			else
 				remap_to_cache_dirty(cache, bio, block, lookup_result.cblock);
-
-			cell_defer(cache, cell, false);
 		}
 		break;
 
 	case POLICY_MISS:
 		inc_miss_counter(cache, bio);
-		pb->all_io_entry = dm_deferred_entry_inc(cache->all_io_ds);
-
 		if (pb->req_nr != 0) {
 			/*
 			 * This is a duplicate writethrough io that is no
 			 * longer needed because the block has been demoted.
 			 */
 			bio_endio(bio, 0);
-			cell_defer(cache, cell, false);
-			return DM_MAPIO_SUBMITTED;
-		} else {
+			cell_defer(cache, *cell, false);
+			r = DM_MAPIO_SUBMITTED;
+
+		} else
 			remap_to_origin_clear_discard(cache, bio, block);
-			cell_defer(cache, cell, false);
-		}
+
 		break;
 
 	default:
 		DMERR_LIMIT("%s: erroring bio: unknown policy op: %u", __func__,
 			    (unsigned) lookup_result.op);
+		cell_defer(cache, *cell, false);
 		bio_io_error(bio);
 		r = DM_MAPIO_SUBMITTED;
 	}
@@ -2529,6 +2546,21 @@ static int cache_map(struct dm_target *ti, struct bio *bio)
 	return r;
 }
 
+static int cache_map(struct dm_target *ti, struct bio *bio)
+{
+	int r;
+	struct dm_bio_prison_cell *cell;
+	struct cache *cache = ti->private;
+
+	r = __cache_map(cache, bio, &cell);
+	if (r == DM_MAPIO_REMAPPED) {
+		inc_ds(cache, bio, cell);
+		cell_defer(cache, cell, false);
+	}
+
+	return r;
+}
+
 static int cache_end_io(struct dm_target *ti, struct bio *bio, int error)
 {
 	struct cache *cache = ti->private;
@@ -2808,7 +2840,7 @@ static void cache_status(struct dm_target *ti, status_type_t type,
 		residency = policy_residency(cache->policy);
 
 		DMEMIT("%u %llu/%llu %u %llu/%llu %u %u %u %u %u %u %lu ",
-		       (unsigned)(DM_CACHE_METADATA_BLOCK_SIZE >> SECTOR_SHIFT),
+		       (unsigned)DM_CACHE_METADATA_BLOCK_SIZE,
 		       (unsigned long long)(nr_blocks_metadata - nr_free_blocks_metadata),
 		       (unsigned long long)nr_blocks_metadata,
 		       cache->sectors_per_block,
@@ -3062,7 +3094,7 @@ static void cache_io_hints(struct dm_target *ti, struct queue_limits *limits)
 	 */
 	if (io_opt_sectors < cache->sectors_per_block ||
 	    do_div(io_opt_sectors, cache->sectors_per_block)) {
-		blk_limits_io_min(limits, 0);
+		blk_limits_io_min(limits, cache->sectors_per_block << SECTOR_SHIFT);
 		blk_limits_io_opt(limits, cache->sectors_per_block << SECTOR_SHIFT);
 	}
 	set_discard_limits(cache, limits);
@@ -3072,7 +3104,7 @@ static void cache_io_hints(struct dm_target *ti, struct queue_limits *limits)
 
 static struct target_type cache_target = {
 	.name = "cache",
-	.version = {1, 4, 0},
+	.version = {1, 5, 0},
 	.module = THIS_MODULE,
 	.ctr = cache_ctr,
 	.dtr = cache_dtr,
diff --git a/drivers/md/dm-crypt.c b/drivers/md/dm-crypt.c
index 4cba2d808afb..fc93b9330af4 100644
--- a/drivers/md/dm-crypt.c
+++ b/drivers/md/dm-crypt.c
@@ -59,7 +59,7 @@ struct dm_crypt_io {
 	int error;
 	sector_t sector;
 	struct dm_crypt_io *base_io;
-};
+} CRYPTO_MINALIGN_ATTR;
 
 struct dm_crypt_request {
 	struct convert_context *ctx;
@@ -162,6 +162,8 @@ struct crypt_config {
 	 */
 	unsigned int dmreq_start;
 
+	unsigned int per_bio_data_size;
+
 	unsigned long flags;
 	unsigned int key_size;
 	unsigned int key_parts;      /* independent parts in key buffer */
@@ -524,29 +526,26 @@ static int crypt_iv_lmk_one(struct crypt_config *cc, u8 *iv,
 			    u8 *data)
 {
 	struct iv_lmk_private *lmk = &cc->iv_gen_private.lmk;
-	struct {
-		struct shash_desc desc;
-		char ctx[crypto_shash_descsize(lmk->hash_tfm)];
-	} sdesc;
+	SHASH_DESC_ON_STACK(desc, lmk->hash_tfm);
 	struct md5_state md5state;
 	__le32 buf[4];
 	int i, r;
 
-	sdesc.desc.tfm = lmk->hash_tfm;
-	sdesc.desc.flags = CRYPTO_TFM_REQ_MAY_SLEEP;
+	desc->tfm = lmk->hash_tfm;
+	desc->flags = CRYPTO_TFM_REQ_MAY_SLEEP;
 
-	r = crypto_shash_init(&sdesc.desc);
+	r = crypto_shash_init(desc);
 	if (r)
 		return r;
 
 	if (lmk->seed) {
-		r = crypto_shash_update(&sdesc.desc, lmk->seed, LMK_SEED_SIZE);
+		r = crypto_shash_update(desc, lmk->seed, LMK_SEED_SIZE);
 		if (r)
 			return r;
 	}
 
 	/* Sector is always 512B, block size 16, add data of blocks 1-31 */
-	r = crypto_shash_update(&sdesc.desc, data + 16, 16 * 31);
+	r = crypto_shash_update(desc, data + 16, 16 * 31);
 	if (r)
 		return r;
 
@@ -555,12 +554,12 @@ static int crypt_iv_lmk_one(struct crypt_config *cc, u8 *iv,
 	buf[1] = cpu_to_le32((((u64)dmreq->iv_sector >> 32) & 0x00FFFFFF) | 0x80000000);
 	buf[2] = cpu_to_le32(4024);
 	buf[3] = 0;
-	r = crypto_shash_update(&sdesc.desc, (u8 *)buf, sizeof(buf));
+	r = crypto_shash_update(desc, (u8 *)buf, sizeof(buf));
 	if (r)
 		return r;
 
 	/* No MD5 padding here */
-	r = crypto_shash_export(&sdesc.desc, &md5state);
+	r = crypto_shash_export(desc, &md5state);
 	if (r)
 		return r;
 
@@ -677,10 +676,7 @@ static int crypt_iv_tcw_whitening(struct crypt_config *cc,
 	struct iv_tcw_private *tcw = &cc->iv_gen_private.tcw;
 	u64 sector = cpu_to_le64((u64)dmreq->iv_sector);
 	u8 buf[TCW_WHITENING_SIZE];
-	struct {
-		struct shash_desc desc;
-		char ctx[crypto_shash_descsize(tcw->crc32_tfm)];
-	} sdesc;
+	SHASH_DESC_ON_STACK(desc, tcw->crc32_tfm);
 	int i, r;
 
 	/* xor whitening with sector number */
@@ -689,16 +685,16 @@ static int crypt_iv_tcw_whitening(struct crypt_config *cc,
 	crypto_xor(&buf[8], (u8 *)&sector, 8);
 
 	/* calculate crc32 for every 32bit part and xor it */
-	sdesc.desc.tfm = tcw->crc32_tfm;
-	sdesc.desc.flags = CRYPTO_TFM_REQ_MAY_SLEEP;
+	desc->tfm = tcw->crc32_tfm;
+	desc->flags = CRYPTO_TFM_REQ_MAY_SLEEP;
 	for (i = 0; i < 4; i++) {
-		r = crypto_shash_init(&sdesc.desc);
+		r = crypto_shash_init(desc);
 		if (r)
 			goto out;
-		r = crypto_shash_update(&sdesc.desc, &buf[i * 4], 4);
+		r = crypto_shash_update(desc, &buf[i * 4], 4);
 		if (r)
 			goto out;
-		r = crypto_shash_final(&sdesc.desc, &buf[i * 4]);
+		r = crypto_shash_final(desc, &buf[i * 4]);
 		if (r)
 			goto out;
 	}
@@ -895,6 +891,15 @@ static void crypt_alloc_req(struct crypt_config *cc,
 	    kcryptd_async_done, dmreq_of_req(cc, ctx->req));
 }
 
+static void crypt_free_req(struct crypt_config *cc,
+			   struct ablkcipher_request *req, struct bio *base_bio)
+{
+	struct dm_crypt_io *io = dm_per_bio_data(base_bio, cc->per_bio_data_size);
+
+	if ((struct ablkcipher_request *)(io + 1) != req)
+		mempool_free(req, cc->req_pool);
+}
+
 /*
  * Encrypt / decrypt data from one bio to another one (can be the same one)
  */
@@ -1008,12 +1013,9 @@ static void crypt_free_buffer_pages(struct crypt_config *cc, struct bio *clone)
 	}
 }
 
-static struct dm_crypt_io *crypt_io_alloc(struct crypt_config *cc,
-					  struct bio *bio, sector_t sector)
+static void crypt_io_init(struct dm_crypt_io *io, struct crypt_config *cc,
+			  struct bio *bio, sector_t sector)
 {
-	struct dm_crypt_io *io;
-
-	io = mempool_alloc(cc->io_pool, GFP_NOIO);
 	io->cc = cc;
 	io->base_bio = bio;
 	io->sector = sector;
@@ -1021,8 +1023,6 @@ static struct dm_crypt_io *crypt_io_alloc(struct crypt_config *cc,
 	io->base_io = NULL;
 	io->ctx.req = NULL;
 	atomic_set(&io->io_pending, 0);
-
-	return io;
 }
 
 static void crypt_inc_pending(struct dm_crypt_io *io)
@@ -1046,8 +1046,9 @@ static void crypt_dec_pending(struct dm_crypt_io *io)
 		return;
 
 	if (io->ctx.req)
-		mempool_free(io->ctx.req, cc->req_pool);
-	mempool_free(io, cc->io_pool);
+		crypt_free_req(cc, io->ctx.req, base_bio);
+	if (io != dm_per_bio_data(base_bio, cc->per_bio_data_size))
+		mempool_free(io, cc->io_pool);
 
 	if (likely(!base_io))
 		bio_endio(base_bio, error);
@@ -1255,8 +1256,8 @@ static void kcryptd_crypt_write_convert(struct dm_crypt_io *io)
 		 * between fragments, so switch to a new dm_crypt_io structure.
 		 */
 		if (unlikely(!crypt_finished && remaining)) {
-			new_io = crypt_io_alloc(io->cc, io->base_bio,
-						sector);
+			new_io = mempool_alloc(cc->io_pool, GFP_NOIO);
+			crypt_io_init(new_io, io->cc, io->base_bio, sector);
 			crypt_inc_pending(new_io);
 			crypt_convert_init(cc, &new_io->ctx, NULL,
 					   io->base_bio, sector);
@@ -1325,7 +1326,7 @@ static void kcryptd_async_done(struct crypto_async_request *async_req,
 	if (error < 0)
 		io->error = -EIO;
 
-	mempool_free(req_of_dmreq(cc, dmreq), cc->req_pool);
+	crypt_free_req(cc, req_of_dmreq(cc, dmreq), io->base_bio);
 
 	if (!atomic_dec_and_test(&ctx->cc_pending))
 		return;
@@ -1681,6 +1682,7 @@ static int crypt_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 	unsigned int key_size, opt_params;
 	unsigned long long tmpll;
 	int ret;
+	size_t iv_size_padding;
 	struct dm_arg_set as;
 	const char *opt_string;
 	char dummy;
@@ -1717,17 +1719,33 @@ static int crypt_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 
 	cc->dmreq_start = sizeof(struct ablkcipher_request);
 	cc->dmreq_start += crypto_ablkcipher_reqsize(any_tfm(cc));
-	cc->dmreq_start = ALIGN(cc->dmreq_start, crypto_tfm_ctx_alignment());
-	cc->dmreq_start += crypto_ablkcipher_alignmask(any_tfm(cc)) &
-			   ~(crypto_tfm_ctx_alignment() - 1);
+	cc->dmreq_start = ALIGN(cc->dmreq_start, __alignof__(struct dm_crypt_request));
+
+	if (crypto_ablkcipher_alignmask(any_tfm(cc)) < CRYPTO_MINALIGN) {
+		/* Allocate the padding exactly */
+		iv_size_padding = -(cc->dmreq_start + sizeof(struct dm_crypt_request))
+				& crypto_ablkcipher_alignmask(any_tfm(cc));
+	} else {
+		/*
+		 * If the cipher requires greater alignment than kmalloc
+		 * alignment, we don't know the exact position of the
+		 * initialization vector. We must assume worst case.
+		 */
+		iv_size_padding = crypto_ablkcipher_alignmask(any_tfm(cc));
+	}
 
 	cc->req_pool = mempool_create_kmalloc_pool(MIN_IOS, cc->dmreq_start +
-			sizeof(struct dm_crypt_request) + cc->iv_size);
+			sizeof(struct dm_crypt_request) + iv_size_padding + cc->iv_size);
 	if (!cc->req_pool) {
 		ti->error = "Cannot allocate crypt request mempool";
 		goto bad;
 	}
 
+	cc->per_bio_data_size = ti->per_bio_data_size =
+		ALIGN(sizeof(struct dm_crypt_io) + cc->dmreq_start +
+		      sizeof(struct dm_crypt_request) + iv_size_padding + cc->iv_size,
+		      ARCH_KMALLOC_MINALIGN);
+
 	cc->page_pool = mempool_create_page_pool(MIN_POOL_PAGES, 0);
 	if (!cc->page_pool) {
 		ti->error = "Cannot allocate page mempool";
@@ -1824,7 +1842,9 @@ static int crypt_map(struct dm_target *ti, struct bio *bio)
 		return DM_MAPIO_REMAPPED;
 	}
 
-	io = crypt_io_alloc(cc, bio, dm_target_offset(ti, bio->bi_iter.bi_sector));
+	io = dm_per_bio_data(bio, cc->per_bio_data_size);
+	crypt_io_init(io, cc, bio, dm_target_offset(ti, bio->bi_iter.bi_sector));
+	io->ctx.req = (struct ablkcipher_request *)(io + 1);
 
 	if (bio_data_dir(io->base_bio) == READ) {
 		if (kcryptd_io_read(io, GFP_NOWAIT))
diff --git a/drivers/md/dm-io.c b/drivers/md/dm-io.c
index db404a0f7e2c..c09359db3a90 100644
--- a/drivers/md/dm-io.c
+++ b/drivers/md/dm-io.c
@@ -33,7 +33,6 @@ struct dm_io_client {
 struct io {
 	unsigned long error_bits;
 	atomic_t count;
-	struct completion *wait;
 	struct dm_io_client *client;
 	io_notify_fn callback;
 	void *context;
@@ -112,28 +111,27 @@ static void retrieve_io_and_region_from_bio(struct bio *bio, struct io **io,
  * We need an io object to keep track of the number of bios that
  * have been dispatched for a particular io.
  *---------------------------------------------------------------*/
-static void dec_count(struct io *io, unsigned int region, int error)
+static void complete_io(struct io *io)
 {
-	if (error)
-		set_bit(region, &io->error_bits);
+	unsigned long error_bits = io->error_bits;
+	io_notify_fn fn = io->callback;
+	void *context = io->context;
 
-	if (atomic_dec_and_test(&io->count)) {
-		if (io->vma_invalidate_size)
-			invalidate_kernel_vmap_range(io->vma_invalidate_address,
-						     io->vma_invalidate_size);
+	if (io->vma_invalidate_size)
+		invalidate_kernel_vmap_range(io->vma_invalidate_address,
+					     io->vma_invalidate_size);
 
-		if (io->wait)
-			complete(io->wait);
+	mempool_free(io, io->client->pool);
+	fn(error_bits, context);
+}
 
-		else {
-			unsigned long r = io->error_bits;
-			io_notify_fn fn = io->callback;
-			void *context = io->context;
+static void dec_count(struct io *io, unsigned int region, int error)
+{
+	if (error)
+		set_bit(region, &io->error_bits);
 
-			mempool_free(io, io->client->pool);
-			fn(r, context);
-		}
-	}
+	if (atomic_dec_and_test(&io->count))
+		complete_io(io);
 }
 
 static void endio(struct bio *bio, int error)
@@ -376,41 +374,51 @@ static void dispatch_io(int rw, unsigned int num_regions,
 	dec_count(io, 0, 0);
 }
 
+struct sync_io {
+	unsigned long error_bits;
+	struct completion wait;
+};
+
+static void sync_io_complete(unsigned long error, void *context)
+{
+	struct sync_io *sio = context;
+
+	sio->error_bits = error;
+	complete(&sio->wait);
+}
+
 static int sync_io(struct dm_io_client *client, unsigned int num_regions,
 		   struct dm_io_region *where, int rw, struct dpages *dp,
 		   unsigned long *error_bits)
 {
-	/*
-	 * gcc <= 4.3 can't do the alignment for stack variables, so we must
-	 * align it on our own.
-	 * volatile prevents the optimizer from removing or reusing
-	 * "io_" field from the stack frame (allowed in ANSI C).
-	 */
-	volatile char io_[sizeof(struct io) + __alignof__(struct io) - 1];
-	struct io *io = (struct io *)PTR_ALIGN(&io_, __alignof__(struct io));
-	DECLARE_COMPLETION_ONSTACK(wait);
+	struct io *io;
+	struct sync_io sio;
 
 	if (num_regions > 1 && (rw & RW_MASK) != WRITE) {
 		WARN_ON(1);
 		return -EIO;
 	}
 
+	init_completion(&sio.wait);
+
+	io = mempool_alloc(client->pool, GFP_NOIO);
 	io->error_bits = 0;
 	atomic_set(&io->count, 1); /* see dispatch_io() */
-	io->wait = &wait;
 	io->client = client;
+	io->callback = sync_io_complete;
+	io->context = &sio;
 
 	io->vma_invalidate_address = dp->vma_invalidate_address;
 	io->vma_invalidate_size = dp->vma_invalidate_size;
 
 	dispatch_io(rw, num_regions, where, dp, io, 1);
 
-	wait_for_completion_io(&wait);
+	wait_for_completion_io(&sio.wait);
 
 	if (error_bits)
-		*error_bits = io->error_bits;
+		*error_bits = sio.error_bits;
 
-	return io->error_bits ? -EIO : 0;
+	return sio.error_bits ? -EIO : 0;
 }
 
 static int async_io(struct dm_io_client *client, unsigned int num_regions,
@@ -428,7 +436,6 @@ static int async_io(struct dm_io_client *client, unsigned int num_regions,
 	io = mempool_alloc(client->pool, GFP_NOIO);
 	io->error_bits = 0;
 	atomic_set(&io->count, 1); /* see dispatch_io() */
-	io->wait = NULL;
 	io->client = client;
 	io->callback = fn;
 	io->context = context;
@@ -481,9 +488,9 @@ static int dp_init(struct dm_io_request *io_req, struct dpages *dp,
  * New collapsed (a)synchronous interface.
  *
  * If the IO is asynchronous (i.e. it has notify.fn), you must either unplug
- * the queue with blk_unplug() some time later or set REQ_SYNC in
-io_req->bi_rw. If you fail to do one of these, the IO will be submitted to
- * the disk after q->unplug_delay, which defaults to 3ms in blk-settings.c.
+ * the queue with blk_unplug() some time later or set REQ_SYNC in io_req->bi_rw.
+ * If you fail to do one of these, the IO will be submitted to the disk after
+ * q->unplug_delay, which defaults to 3ms in blk-settings.c.
  */
 int dm_io(struct dm_io_request *io_req, unsigned num_regions,
 	  struct dm_io_region *where, unsigned long *sync_error_bits)
diff --git a/drivers/md/dm-ioctl.c b/drivers/md/dm-ioctl.c
index 51521429fb59..0be9381365d7 100644
--- a/drivers/md/dm-ioctl.c
+++ b/drivers/md/dm-ioctl.c
@@ -1418,7 +1418,7 @@ 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->dm_dev.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-log-userspace-transfer.c b/drivers/md/dm-log-userspace-transfer.c
index b428c0ae63d5..39ad9664d397 100644
--- a/drivers/md/dm-log-userspace-transfer.c
+++ b/drivers/md/dm-log-userspace-transfer.c
@@ -272,7 +272,7 @@ int dm_ulog_tfr_init(void)
 
 	r = cn_add_callback(&ulog_cn_id, "dmlogusr", cn_ulog_callback);
 	if (r) {
-		cn_del_callback(&ulog_cn_id);
+		kfree(prealloced_cn_msg);
 		return r;
 	}
 
diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c
index f4167b013d99..7b6b0f0f831a 100644
--- a/drivers/md/dm-mpath.c
+++ b/drivers/md/dm-mpath.c
@@ -317,8 +317,10 @@ static void __choose_pgpath(struct multipath *m, size_t nr_bytes)
 	struct priority_group *pg;
 	unsigned bypassed = 1;
 
-	if (!m->nr_valid_paths)
+	if (!m->nr_valid_paths) {
+		m->queue_io = 0;
 		goto failed;
+	}
 
 	/* Were we instructed to switch PG? */
 	if (m->next_pg) {
@@ -373,8 +375,6 @@ static int __must_push_back(struct multipath *m)
 		 dm_noflush_suspending(m->ti)));
 }
 
-#define pg_ready(m) (!(m)->queue_io && !(m)->pg_init_required)
-
 /*
  * Map cloned requests
  */
@@ -402,11 +402,11 @@ static int multipath_map(struct dm_target *ti, struct request *clone,
 		if (!__must_push_back(m))
 			r = -EIO;	/* Failed */
 		goto out_unlock;
-	}
-	if (!pg_ready(m)) {
+	} else if (m->queue_io || m->pg_init_required) {
 		__pg_init_all_paths(m);
 		goto out_unlock;
 	}
+
 	if (set_mapinfo(m, map_context) < 0)
 		/* ENOMEM, requeue */
 		goto out_unlock;
diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c
index 4880b69e2e9e..07c0fa0fa284 100644
--- a/drivers/md/dm-raid.c
+++ b/drivers/md/dm-raid.c
@@ -1,6 +1,6 @@
 /*
  * Copyright (C) 2010-2011 Neil Brown
- * Copyright (C) 2010-2011 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2010-2014 Red Hat, Inc. All rights reserved.
  *
  * This file is released under the GPL.
  */
@@ -18,6 +18,8 @@
 
 #define DM_MSG_PREFIX "raid"
 
+static bool devices_handle_discard_safely = false;
+
 /*
  * The following flags are used by dm-raid.c to set up the array state.
  * They must be cleared before md_run is called.
@@ -475,6 +477,8 @@ too_many:
  *                                      will form the "stripe"
  *    [[no]sync]			Force or prevent recovery of the
  *                                      entire array
+ *    [devices_handle_discard_safely]	Allow discards on RAID4/5/6; useful if RAID
+ *					member device(s) properly support TRIM/UNMAP
  *    [rebuild <idx>]			Rebuild the drive indicated by the index
  *    [daemon_sleep <ms>]		Time between bitmap daemon work to
  *                                      clear bits
@@ -785,8 +789,7 @@ struct dm_raid_superblock {
 	__le32 layout;
 	__le32 stripe_sectors;
 
-	__u8 pad[452];		/* Round struct to 512 bytes. */
-				/* Always set to 0 when writing. */
+	/* Remainder of a logical block is zero-filled when writing (see super_sync()). */
 } __packed;
 
 static int read_disk_sb(struct md_rdev *rdev, int size)
@@ -823,7 +826,7 @@ static void super_sync(struct mddev *mddev, struct md_rdev *rdev)
 		    test_bit(Faulty, &(rs->dev[i].rdev.flags)))
 			failed_devices |= (1ULL << i);
 
-	memset(sb, 0, sizeof(*sb));
+	memset(sb + 1, 0, rdev->sb_size - sizeof(*sb));
 
 	sb->magic = cpu_to_le32(DM_RAID_MAGIC);
 	sb->features = cpu_to_le32(0);	/* No features yet */
@@ -858,7 +861,11 @@ static int super_load(struct md_rdev *rdev, struct md_rdev *refdev)
 	uint64_t events_sb, events_refsb;
 
 	rdev->sb_start = 0;
-	rdev->sb_size = sizeof(*sb);
+	rdev->sb_size = bdev_logical_block_size(rdev->meta_bdev);
+	if (rdev->sb_size < sizeof(*sb) || rdev->sb_size > PAGE_SIZE) {
+		DMERR("superblock size of a logical block is no longer valid");
+		return -EINVAL;
+	}
 
 	ret = read_disk_sb(rdev, rdev->sb_size);
 	if (ret)
@@ -1150,6 +1157,53 @@ static int analyse_superblocks(struct dm_target *ti, struct raid_set *rs)
 }
 
 /*
+ * Enable/disable discard support on RAID set depending on
+ * RAID level and discard properties of underlying RAID members.
+ */
+static void configure_discard_support(struct dm_target *ti, struct raid_set *rs)
+{
+	int i;
+	bool raid456;
+
+	/* Assume discards not supported until after checks below. */
+	ti->discards_supported = false;
+
+	/* RAID level 4,5,6 require discard_zeroes_data for data integrity! */
+	raid456 = (rs->md.level == 4 || rs->md.level == 5 || rs->md.level == 6);
+
+	for (i = 0; i < rs->md.raid_disks; i++) {
+		struct request_queue *q;
+
+		if (!rs->dev[i].rdev.bdev)
+			continue;
+
+		q = bdev_get_queue(rs->dev[i].rdev.bdev);
+		if (!q || !blk_queue_discard(q))
+			return;
+
+		if (raid456) {
+			if (!q->limits.discard_zeroes_data)
+				return;
+			if (!devices_handle_discard_safely) {
+				DMERR("raid456 discard support disabled due to discard_zeroes_data uncertainty.");
+				DMERR("Set dm-raid.devices_handle_discard_safely=Y to override.");
+				return;
+			}
+		}
+	}
+
+	/* All RAID members properly support discards */
+	ti->discards_supported = true;
+
+	/*
+	 * RAID1 and RAID10 personalities require bio splitting,
+	 * RAID0/4/5/6 don't and process large discard bios properly.
+	 */
+	ti->split_discard_bios = !!(rs->md.level == 1 || rs->md.level == 10);
+	ti->num_discard_bios = 1;
+}
+
+/*
  * Construct a RAID4/5/6 mapping:
  * Args:
  *	<raid_type> <#raid_params> <raid_params>		\
@@ -1231,6 +1285,11 @@ static int raid_ctr(struct dm_target *ti, unsigned argc, char **argv)
 	ti->private = rs;
 	ti->num_flush_bios = 1;
 
+	/*
+	 * Disable/enable discard support on RAID set.
+	 */
+	configure_discard_support(ti, rs);
+
 	mutex_lock(&rs->md.reconfig_mutex);
 	ret = md_run(&rs->md);
 	rs->md.in_sync = 0; /* Assume already marked dirty */
@@ -1652,7 +1711,7 @@ static void raid_resume(struct dm_target *ti)
 
 static struct target_type raid_target = {
 	.name = "raid",
-	.version = {1, 5, 2},
+	.version = {1, 6, 0},
 	.module = THIS_MODULE,
 	.ctr = raid_ctr,
 	.dtr = raid_dtr,
@@ -1683,6 +1742,10 @@ static void __exit dm_raid_exit(void)
 module_init(dm_raid_init);
 module_exit(dm_raid_exit);
 
+module_param(devices_handle_discard_safely, bool, 0644);
+MODULE_PARM_DESC(devices_handle_discard_safely,
+		 "Set to Y if all devices in each array reliably return zeroes on reads from discarded regions");
+
 MODULE_DESCRIPTION(DM_NAME " raid4/5/6 target");
 MODULE_ALIAS("dm-raid1");
 MODULE_ALIAS("dm-raid10");
diff --git a/drivers/md/dm-stats.c b/drivers/md/dm-stats.c
index 28a90122a5a8..87f86c77b094 100644
--- a/drivers/md/dm-stats.c
+++ b/drivers/md/dm-stats.c
@@ -548,7 +548,7 @@ void dm_stats_account_io(struct dm_stats *stats, unsigned long bi_rw,
 		 * A race condition can at worst result in the merged flag being
 		 * misrepresented, so we don't have to disable preemption here.
 		 */
-		last = __this_cpu_ptr(stats->last);
+		last = raw_cpu_ptr(stats->last);
 		stats_aux->merged =
 			(bi_sector == (ACCESS_ONCE(last->last_sector) &&
 				       ((bi_rw & (REQ_WRITE | REQ_DISCARD)) ==
diff --git a/drivers/md/dm-stripe.c b/drivers/md/dm-stripe.c
index d1600d2aa2e2..f8b37d4c05d8 100644
--- a/drivers/md/dm-stripe.c
+++ b/drivers/md/dm-stripe.c
@@ -159,8 +159,10 @@ static int stripe_ctr(struct dm_target *ti, unsigned int argc, char **argv)
 		sc->stripes_shift = __ffs(stripes);
 
 	r = dm_set_target_max_io_len(ti, chunk_size);
-	if (r)
+	if (r) {
+		kfree(sc);
 		return r;
+	}
 
 	ti->num_flush_bios = stripes;
 	ti->num_discard_bios = stripes;
diff --git a/drivers/md/dm-switch.c b/drivers/md/dm-switch.c
index 09a688b3d48c..50fca469cafd 100644
--- a/drivers/md/dm-switch.c
+++ b/drivers/md/dm-switch.c
@@ -137,13 +137,23 @@ static void switch_get_position(struct switch_ctx *sctx, unsigned long region_nr
 	*bit *= sctx->region_table_entry_bits;
 }
 
+static unsigned switch_region_table_read(struct switch_ctx *sctx, unsigned long region_nr)
+{
+	unsigned long region_index;
+	unsigned bit;
+
+	switch_get_position(sctx, region_nr, &region_index, &bit);
+
+	return (ACCESS_ONCE(sctx->region_table[region_index]) >> bit) &
+		((1 << sctx->region_table_entry_bits) - 1);
+}
+
 /*
  * Find which path to use at given offset.
  */
 static unsigned switch_get_path_nr(struct switch_ctx *sctx, sector_t offset)
 {
-	unsigned long region_index;
-	unsigned bit, path_nr;
+	unsigned path_nr;
 	sector_t p;
 
 	p = offset;
@@ -152,9 +162,7 @@ static unsigned switch_get_path_nr(struct switch_ctx *sctx, sector_t offset)
 	else
 		sector_div(p, sctx->region_size);
 
-	switch_get_position(sctx, p, &region_index, &bit);
-	path_nr = (ACCESS_ONCE(sctx->region_table[region_index]) >> bit) &
-	       ((1 << sctx->region_table_entry_bits) - 1);
+	path_nr = switch_region_table_read(sctx, p);
 
 	/* This can only happen if the processor uses non-atomic stores. */
 	if (unlikely(path_nr >= sctx->nr_paths))
@@ -363,7 +371,7 @@ static __always_inline unsigned long parse_hex(const char **string)
 }
 
 static int process_set_region_mappings(struct switch_ctx *sctx,
-			     unsigned argc, char **argv)
+				       unsigned argc, char **argv)
 {
 	unsigned i;
 	unsigned long region_index = 0;
@@ -372,6 +380,51 @@ static int process_set_region_mappings(struct switch_ctx *sctx,
 		unsigned long path_nr;
 		const char *string = argv[i];
 
+		if ((*string & 0xdf) == 'R') {
+			unsigned long cycle_length, num_write;
+
+			string++;
+			if (unlikely(*string == ',')) {
+				DMWARN("invalid set_region_mappings argument: '%s'", argv[i]);
+				return -EINVAL;
+			}
+			cycle_length = parse_hex(&string);
+			if (unlikely(*string != ',')) {
+				DMWARN("invalid set_region_mappings argument: '%s'", argv[i]);
+				return -EINVAL;
+			}
+			string++;
+			if (unlikely(!*string)) {
+				DMWARN("invalid set_region_mappings argument: '%s'", argv[i]);
+				return -EINVAL;
+			}
+			num_write = parse_hex(&string);
+			if (unlikely(*string)) {
+				DMWARN("invalid set_region_mappings argument: '%s'", argv[i]);
+				return -EINVAL;
+			}
+
+			if (unlikely(!cycle_length) || unlikely(cycle_length - 1 > region_index)) {
+				DMWARN("invalid set_region_mappings cycle length: %lu > %lu",
+				       cycle_length - 1, region_index);
+				return -EINVAL;
+			}
+			if (unlikely(region_index + num_write < region_index) ||
+			    unlikely(region_index + num_write >= sctx->nr_regions)) {
+				DMWARN("invalid set_region_mappings region number: %lu + %lu >= %lu",
+				       region_index, num_write, sctx->nr_regions);
+				return -EINVAL;
+			}
+
+			while (num_write--) {
+				region_index++;
+				path_nr = switch_region_table_read(sctx, region_index - cycle_length);
+				switch_region_table_write(sctx, region_index, path_nr);
+			}
+
+			continue;
+		}
+
 		if (*string == ':')
 			region_index++;
 		else {
@@ -500,7 +553,7 @@ static int switch_iterate_devices(struct dm_target *ti,
 
 static struct target_type switch_target = {
 	.name = "switch",
-	.version = {1, 0, 0},
+	.version = {1, 1, 0},
 	.module = THIS_MODULE,
 	.ctr = switch_ctr,
 	.dtr = switch_dtr,
diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
index 5f59f1e3e5b1..b2bd1ebf4562 100644
--- a/drivers/md/dm-table.c
+++ b/drivers/md/dm-table.c
@@ -210,15 +210,16 @@ int dm_table_create(struct dm_table **result, fmode_t mode,
 	return 0;
 }
 
-static void free_devices(struct list_head *devices)
+static void free_devices(struct list_head *devices, struct mapped_device *md)
 {
 	struct list_head *tmp, *next;
 
 	list_for_each_safe(tmp, next, devices) {
 		struct dm_dev_internal *dd =
 		    list_entry(tmp, struct dm_dev_internal, list);
-		DMWARN("dm_table_destroy: dm_put_device call missing for %s",
-		       dd->dm_dev.name);
+		DMWARN("%s: dm_table_destroy: dm_put_device call missing for %s",
+		       dm_device_name(md), dd->dm_dev->name);
+		dm_put_table_device(md, dd->dm_dev);
 		kfree(dd);
 	}
 }
@@ -247,7 +248,7 @@ void dm_table_destroy(struct dm_table *t)
 	vfree(t->highs);
 
 	/* free the device list */
-	free_devices(&t->devices);
+	free_devices(&t->devices, t->md);
 
 	dm_free_md_mempools(t->mempools);
 
@@ -262,53 +263,13 @@ static struct dm_dev_internal *find_device(struct list_head *l, dev_t dev)
 	struct dm_dev_internal *dd;
 
 	list_for_each_entry (dd, l, list)
-		if (dd->dm_dev.bdev->bd_dev == dev)
+		if (dd->dm_dev->bdev->bd_dev == dev)
 			return dd;
 
 	return NULL;
 }
 
 /*
- * Open a device so we can use it as a map destination.
- */
-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->dm_dev.bdev);
-
-	bdev = blkdev_get_by_dev(dev, d->dm_dev.mode | FMODE_EXCL, _claim_ptr);
-	if (IS_ERR(bdev))
-		return PTR_ERR(bdev);
-
-	r = bd_link_disk_holder(bdev, dm_disk(md));
-	if (r) {
-		blkdev_put(bdev, d->dm_dev.mode | FMODE_EXCL);
-		return r;
-	}
-
-	d->dm_dev.bdev = bdev;
-	return 0;
-}
-
-/*
- * Close a device that we've been using.
- */
-static void close_dev(struct dm_dev_internal *d, struct mapped_device *md)
-{
-	if (!d->dm_dev.bdev)
-		return;
-
-	bd_unlink_disk_holder(d->dm_dev.bdev, dm_disk(md));
-	blkdev_put(d->dm_dev.bdev, d->dm_dev.mode | FMODE_EXCL);
-	d->dm_dev.bdev = NULL;
-}
-
-/*
  * If possible, this checks an area of a destination device is invalid.
  */
 static int device_area_is_invalid(struct dm_target *ti, struct dm_dev *dev,
@@ -386,19 +347,17 @@ static int upgrade_mode(struct dm_dev_internal *dd, fmode_t new_mode,
 			struct mapped_device *md)
 {
 	int r;
-	struct dm_dev_internal dd_new, dd_old;
-
-	dd_new = dd_old = *dd;
+	struct dm_dev *old_dev, *new_dev;
 
-	dd_new.dm_dev.mode |= new_mode;
-	dd_new.dm_dev.bdev = NULL;
+	old_dev = dd->dm_dev;
 
-	r = open_dev(&dd_new, dd->dm_dev.bdev->bd_dev, md);
+	r = dm_get_table_device(md, dd->dm_dev->bdev->bd_dev,
+				dd->dm_dev->mode | new_mode, &new_dev);
 	if (r)
 		return r;
 
-	dd->dm_dev.mode |= new_mode;
-	close_dev(&dd_old, md);
+	dd->dm_dev = new_dev;
+	dm_put_table_device(md, old_dev);
 
 	return 0;
 }
@@ -440,27 +399,22 @@ int dm_get_device(struct dm_target *ti, const char *path, fmode_t mode,
 		if (!dd)
 			return -ENOMEM;
 
-		dd->dm_dev.mode = mode;
-		dd->dm_dev.bdev = NULL;
-
-		if ((r = open_dev(dd, dev, t->md))) {
+		if ((r = dm_get_table_device(t->md, dev, mode, &dd->dm_dev))) {
 			kfree(dd);
 			return r;
 		}
 
-		format_dev_t(dd->dm_dev.name, dev);
-
 		atomic_set(&dd->count, 0);
 		list_add(&dd->list, &t->devices);
 
-	} else if (dd->dm_dev.mode != (mode | dd->dm_dev.mode)) {
+	} else if (dd->dm_dev->mode != (mode | dd->dm_dev->mode)) {
 		r = upgrade_mode(dd, mode, t->md);
 		if (r)
 			return r;
 	}
 	atomic_inc(&dd->count);
 
-	*result = &dd->dm_dev;
+	*result = dd->dm_dev;
 	return 0;
 }
 EXPORT_SYMBOL(dm_get_device);
@@ -505,11 +459,23 @@ static int dm_set_device_limits(struct dm_target *ti, struct dm_dev *dev,
  */
 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);
+	int found = 0;
+	struct list_head *devices = &ti->table->devices;
+	struct dm_dev_internal *dd;
 
+	list_for_each_entry(dd, devices, list) {
+		if (dd->dm_dev == d) {
+			found = 1;
+			break;
+		}
+	}
+	if (!found) {
+		DMWARN("%s: device %s not in table devices list",
+		       dm_device_name(ti->table->md), d->name);
+		return;
+	}
 	if (atomic_dec_and_test(&dd->count)) {
-		close_dev(dd, ti->table->md);
+		dm_put_table_device(ti->table->md, d);
 		list_del(&dd->list);
 		kfree(dd);
 	}
@@ -906,7 +872,7 @@ static int dm_table_set_type(struct dm_table *t)
 	/* Non-request-stackable devices can't be used for request-based dm */
 	devices = dm_table_get_devices(t);
 	list_for_each_entry(dd, devices, list) {
-		if (!blk_queue_stackable(bdev_get_queue(dd->dm_dev.bdev))) {
+		if (!blk_queue_stackable(bdev_get_queue(dd->dm_dev->bdev))) {
 			DMWARN("table load rejected: including"
 			       " non-request-stackable devices");
 			return -EINVAL;
@@ -1043,7 +1009,7 @@ static struct gendisk * dm_table_get_integrity_disk(struct dm_table *t,
 	struct gendisk *prev_disk = NULL, *template_disk = NULL;
 
 	list_for_each_entry(dd, devices, list) {
-		template_disk = dd->dm_dev.bdev->bd_disk;
+		template_disk = dd->dm_dev->bdev->bd_disk;
 		if (!blk_get_integrity(template_disk))
 			goto no_integrity;
 		if (!match_all && !blk_integrity_is_initialized(template_disk))
@@ -1386,6 +1352,14 @@ static int device_is_not_random(struct dm_target *ti, struct dm_dev *dev,
 	return q && !blk_queue_add_random(q);
 }
 
+static int queue_supports_sg_merge(struct dm_target *ti, struct dm_dev *dev,
+				   sector_t start, sector_t len, void *data)
+{
+	struct request_queue *q = bdev_get_queue(dev->bdev);
+
+	return q && !test_bit(QUEUE_FLAG_NO_SG_MERGE, &q->queue_flags);
+}
+
 static bool dm_table_all_devices_attribute(struct dm_table *t,
 					   iterate_devices_callout_fn func)
 {
@@ -1430,6 +1404,43 @@ static bool dm_table_supports_write_same(struct dm_table *t)
 	return true;
 }
 
+static int device_discard_capable(struct dm_target *ti, struct dm_dev *dev,
+				  sector_t start, sector_t len, void *data)
+{
+	struct request_queue *q = bdev_get_queue(dev->bdev);
+
+	return q && blk_queue_discard(q);
+}
+
+static bool dm_table_supports_discards(struct dm_table *t)
+{
+	struct dm_target *ti;
+	unsigned i = 0;
+
+	/*
+	 * 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 selectively must provide.
+	 */
+	while (i < dm_table_get_num_targets(t)) {
+		ti = dm_table_get_target(t, i++);
+
+		if (!ti->num_discard_bios)
+			continue;
+
+		if (ti->discards_supported)
+			return 1;
+
+		if (ti->type->iterate_devices &&
+		    ti->type->iterate_devices(ti, device_discard_capable, NULL))
+			return 1;
+	}
+
+	return 0;
+}
+
 void dm_table_set_restrictions(struct dm_table *t, struct request_queue *q,
 			       struct queue_limits *limits)
 {
@@ -1464,6 +1475,11 @@ void dm_table_set_restrictions(struct dm_table *t, struct request_queue *q,
 	if (!dm_table_supports_write_same(t))
 		q->limits.max_write_same_sectors = 0;
 
+	if (dm_table_all_devices_attribute(t, queue_supports_sg_merge))
+		queue_flag_clear_unlocked(QUEUE_FLAG_NO_SG_MERGE, q);
+	else
+		queue_flag_set_unlocked(QUEUE_FLAG_NO_SG_MERGE, q);
+
 	dm_table_set_integrity(t);
 
 	/*
@@ -1579,7 +1595,7 @@ int dm_table_any_congested(struct dm_table *t, int bdi_bits)
 	int r = 0;
 
 	list_for_each_entry(dd, devices, list) {
-		struct request_queue *q = bdev_get_queue(dd->dm_dev.bdev);
+		struct request_queue *q = bdev_get_queue(dd->dm_dev->bdev);
 		char b[BDEVNAME_SIZE];
 
 		if (likely(q))
@@ -1587,7 +1603,7 @@ int dm_table_any_congested(struct dm_table *t, int bdi_bits)
 		else
 			DMWARN_LIMIT("%s: any_congested: nonexistent device %s",
 				     dm_device_name(t->md),
-				     bdevname(dd->dm_dev.bdev, b));
+				     bdevname(dd->dm_dev->bdev, b));
 	}
 
 	list_for_each_entry(cb, &t->target_callbacks, list)
@@ -1636,39 +1652,3 @@ void dm_table_run_md_queue_async(struct dm_table *t)
 }
 EXPORT_SYMBOL(dm_table_run_md_queue_async);
 
-static int device_discard_capable(struct dm_target *ti, struct dm_dev *dev,
-				  sector_t start, sector_t len, void *data)
-{
-	struct request_queue *q = bdev_get_queue(dev->bdev);
-
-	return q && blk_queue_discard(q);
-}
-
-bool dm_table_supports_discards(struct dm_table *t)
-{
-	struct dm_target *ti;
-	unsigned i = 0;
-
-	/*
-	 * 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 selectively must provide.
-	 */
-	while (i < dm_table_get_num_targets(t)) {
-		ti = dm_table_get_target(t, i++);
-
-		if (!ti->num_discard_bios)
-			continue;
-
-		if (ti->discards_supported)
-			return 1;
-
-		if (ti->type->iterate_devices &&
-		    ti->type->iterate_devices(ti, device_discard_capable, NULL))
-			return 1;
-	}
-
-	return 0;
-}
diff --git a/drivers/md/dm-thin.c b/drivers/md/dm-thin.c
index fc9c848a60c9..0f86d802b533 100644
--- a/drivers/md/dm-thin.c
+++ b/drivers/md/dm-thin.c
@@ -227,6 +227,7 @@ struct thin_c {
 	struct list_head list;
 	struct dm_dev *pool_dev;
 	struct dm_dev *origin_dev;
+	sector_t origin_size;
 	dm_thin_id dev_id;
 
 	struct pool *pool;
@@ -554,11 +555,16 @@ static void remap_and_issue(struct thin_c *tc, struct bio *bio,
 struct dm_thin_new_mapping {
 	struct list_head list;
 
-	bool quiesced:1;
-	bool prepared:1;
 	bool pass_discard:1;
 	bool definitely_not_shared:1;
 
+	/*
+	 * Track quiescing, copying and zeroing preparation actions.  When this
+	 * counter hits zero the block is prepared and can be inserted into the
+	 * btree.
+	 */
+	atomic_t prepare_actions;
+
 	int err;
 	struct thin_c *tc;
 	dm_block_t virt_block;
@@ -575,43 +581,41 @@ struct dm_thin_new_mapping {
 	bio_end_io_t *saved_bi_end_io;
 };
 
-static void __maybe_add_mapping(struct dm_thin_new_mapping *m)
+static void __complete_mapping_preparation(struct dm_thin_new_mapping *m)
 {
 	struct pool *pool = m->tc->pool;
 
-	if (m->quiesced && m->prepared) {
+	if (atomic_dec_and_test(&m->prepare_actions)) {
 		list_add_tail(&m->list, &pool->prepared_mappings);
 		wake_worker(pool);
 	}
 }
 
-static void copy_complete(int read_err, unsigned long write_err, void *context)
+static void complete_mapping_preparation(struct dm_thin_new_mapping *m)
 {
 	unsigned long flags;
-	struct dm_thin_new_mapping *m = context;
 	struct pool *pool = m->tc->pool;
 
-	m->err = read_err || write_err ? -EIO : 0;
-
 	spin_lock_irqsave(&pool->lock, flags);
-	m->prepared = true;
-	__maybe_add_mapping(m);
+	__complete_mapping_preparation(m);
 	spin_unlock_irqrestore(&pool->lock, flags);
 }
 
+static void copy_complete(int read_err, unsigned long write_err, void *context)
+{
+	struct dm_thin_new_mapping *m = context;
+
+	m->err = read_err || write_err ? -EIO : 0;
+	complete_mapping_preparation(m);
+}
+
 static void overwrite_endio(struct bio *bio, int err)
 {
-	unsigned long flags;
 	struct dm_thin_endio_hook *h = dm_per_bio_data(bio, sizeof(struct dm_thin_endio_hook));
 	struct dm_thin_new_mapping *m = h->overwrite_mapping;
-	struct pool *pool = m->tc->pool;
 
 	m->err = err;
-
-	spin_lock_irqsave(&pool->lock, flags);
-	m->prepared = true;
-	__maybe_add_mapping(m);
-	spin_unlock_irqrestore(&pool->lock, flags);
+	complete_mapping_preparation(m);
 }
 
 /*----------------------------------------------------------------*/
@@ -821,10 +825,31 @@ static struct dm_thin_new_mapping *get_next_mapping(struct pool *pool)
 	return m;
 }
 
+static void ll_zero(struct thin_c *tc, struct dm_thin_new_mapping *m,
+		    sector_t begin, sector_t end)
+{
+	int r;
+	struct dm_io_region to;
+
+	to.bdev = tc->pool_dev->bdev;
+	to.sector = begin;
+	to.count = end - begin;
+
+	r = dm_kcopyd_zero(tc->pool->copier, 1, &to, 0, copy_complete, m);
+	if (r < 0) {
+		DMERR_LIMIT("dm_kcopyd_zero() failed");
+		copy_complete(1, 1, m);
+	}
+}
+
+/*
+ * A partial copy also needs to zero the uncopied region.
+ */
 static void schedule_copy(struct thin_c *tc, dm_block_t virt_block,
 			  struct dm_dev *origin, dm_block_t data_origin,
 			  dm_block_t data_dest,
-			  struct dm_bio_prison_cell *cell, struct bio *bio)
+			  struct dm_bio_prison_cell *cell, struct bio *bio,
+			  sector_t len)
 {
 	int r;
 	struct pool *pool = tc->pool;
@@ -835,8 +860,15 @@ static void schedule_copy(struct thin_c *tc, dm_block_t virt_block,
 	m->data_block = data_dest;
 	m->cell = cell;
 
+	/*
+	 * quiesce action + copy action + an extra reference held for the
+	 * duration of this function (we may need to inc later for a
+	 * partial zero).
+	 */
+	atomic_set(&m->prepare_actions, 3);
+
 	if (!dm_deferred_set_add_work(pool->shared_read_ds, &m->list))
-		m->quiesced = true;
+		complete_mapping_preparation(m); /* already quiesced */
 
 	/*
 	 * IO to pool_dev remaps to the pool target's data_dev.
@@ -857,20 +889,38 @@ static void schedule_copy(struct thin_c *tc, dm_block_t virt_block,
 
 		from.bdev = origin->bdev;
 		from.sector = data_origin * pool->sectors_per_block;
-		from.count = pool->sectors_per_block;
+		from.count = len;
 
 		to.bdev = tc->pool_dev->bdev;
 		to.sector = data_dest * pool->sectors_per_block;
-		to.count = pool->sectors_per_block;
+		to.count = len;
 
 		r = dm_kcopyd_copy(pool->copier, &from, 1, &to,
 				   0, copy_complete, m);
 		if (r < 0) {
-			mempool_free(m, pool->mapping_pool);
 			DMERR_LIMIT("dm_kcopyd_copy() failed");
-			cell_error(pool, cell);
+			copy_complete(1, 1, m);
+
+			/*
+			 * We allow the zero to be issued, to simplify the
+			 * error path.  Otherwise we'd need to start
+			 * worrying about decrementing the prepare_actions
+			 * counter.
+			 */
+		}
+
+		/*
+		 * Do we need to zero a tail region?
+		 */
+		if (len < pool->sectors_per_block && pool->pf.zero_new_blocks) {
+			atomic_inc(&m->prepare_actions);
+			ll_zero(tc, m,
+				data_dest * pool->sectors_per_block + len,
+				(data_dest + 1) * pool->sectors_per_block);
 		}
 	}
+
+	complete_mapping_preparation(m); /* drop our ref */
 }
 
 static void schedule_internal_copy(struct thin_c *tc, dm_block_t virt_block,
@@ -878,15 +928,8 @@ static void schedule_internal_copy(struct thin_c *tc, dm_block_t virt_block,
 				   struct dm_bio_prison_cell *cell, struct bio *bio)
 {
 	schedule_copy(tc, virt_block, tc->pool_dev,
-		      data_origin, data_dest, cell, bio);
-}
-
-static void schedule_external_copy(struct thin_c *tc, dm_block_t virt_block,
-				   dm_block_t data_dest,
-				   struct dm_bio_prison_cell *cell, struct bio *bio)
-{
-	schedule_copy(tc, virt_block, tc->origin_dev,
-		      virt_block, data_dest, cell, bio);
+		      data_origin, data_dest, cell, bio,
+		      tc->pool->sectors_per_block);
 }
 
 static void schedule_zero(struct thin_c *tc, dm_block_t virt_block,
@@ -896,8 +939,7 @@ static void schedule_zero(struct thin_c *tc, dm_block_t virt_block,
 	struct pool *pool = tc->pool;
 	struct dm_thin_new_mapping *m = get_next_mapping(pool);
 
-	m->quiesced = true;
-	m->prepared = false;
+	atomic_set(&m->prepare_actions, 1); /* no need to quiesce */
 	m->tc = tc;
 	m->virt_block = virt_block;
 	m->data_block = data_block;
@@ -919,21 +961,33 @@ static void schedule_zero(struct thin_c *tc, dm_block_t virt_block,
 		save_and_set_endio(bio, &m->saved_bi_end_io, overwrite_endio);
 		inc_all_io_entry(pool, bio);
 		remap_and_issue(tc, bio, data_block);
-	} else {
-		int r;
-		struct dm_io_region to;
 
-		to.bdev = tc->pool_dev->bdev;
-		to.sector = data_block * pool->sectors_per_block;
-		to.count = pool->sectors_per_block;
+	} else
+		ll_zero(tc, m,
+			data_block * pool->sectors_per_block,
+			(data_block + 1) * pool->sectors_per_block);
+}
 
-		r = dm_kcopyd_zero(pool->copier, 1, &to, 0, copy_complete, m);
-		if (r < 0) {
-			mempool_free(m, pool->mapping_pool);
-			DMERR_LIMIT("dm_kcopyd_zero() failed");
-			cell_error(pool, cell);
-		}
-	}
+static void schedule_external_copy(struct thin_c *tc, dm_block_t virt_block,
+				   dm_block_t data_dest,
+				   struct dm_bio_prison_cell *cell, struct bio *bio)
+{
+	struct pool *pool = tc->pool;
+	sector_t virt_block_begin = virt_block * pool->sectors_per_block;
+	sector_t virt_block_end = (virt_block + 1) * pool->sectors_per_block;
+
+	if (virt_block_end <= tc->origin_size)
+		schedule_copy(tc, virt_block, tc->origin_dev,
+			      virt_block, data_dest, cell, bio,
+			      pool->sectors_per_block);
+
+	else if (virt_block_begin < tc->origin_size)
+		schedule_copy(tc, virt_block, tc->origin_dev,
+			      virt_block, data_dest, cell, bio,
+			      tc->origin_size - virt_block_begin);
+
+	else
+		schedule_zero(tc, virt_block, data_dest, cell, bio);
 }
 
 /*
@@ -1315,7 +1369,18 @@ static void process_bio(struct thin_c *tc, struct bio *bio)
 			inc_all_io_entry(pool, bio);
 			cell_defer_no_holder(tc, cell);
 
-			remap_to_origin_and_issue(tc, bio);
+			if (bio_end_sector(bio) <= tc->origin_size)
+				remap_to_origin_and_issue(tc, bio);
+
+			else if (bio->bi_iter.bi_sector < tc->origin_size) {
+				zero_fill_bio(bio);
+				bio->bi_iter.bi_size = (tc->origin_size - bio->bi_iter.bi_sector) << SECTOR_SHIFT;
+				remap_to_origin_and_issue(tc, bio);
+
+			} else {
+				zero_fill_bio(bio);
+				bio_endio(bio, 0);
+			}
 		} else
 			provision_block(tc, bio, block, cell);
 		break;
@@ -1871,6 +1936,14 @@ static int thin_bio_map(struct dm_target *ti, struct bio *bio)
 		return DM_MAPIO_SUBMITTED;
 	}
 
+	/*
+	 * We must hold the virtual cell before doing the lookup, otherwise
+	 * there's a race with discard.
+	 */
+	build_virtual_key(tc->td, block, &key);
+	if (dm_bio_detain(tc->pool->prison, &key, bio, &cell1, &cell_result))
+		return DM_MAPIO_SUBMITTED;
+
 	r = dm_thin_find_block(td, block, 0, &result);
 
 	/*
@@ -1894,13 +1967,10 @@ static int thin_bio_map(struct dm_target *ti, struct bio *bio)
 			 * shared flag will be set in their case.
 			 */
 			thin_defer_bio(tc, bio);
+			cell_defer_no_holder_no_free(tc, &cell1);
 			return DM_MAPIO_SUBMITTED;
 		}
 
-		build_virtual_key(tc->td, block, &key);
-		if (dm_bio_detain(tc->pool->prison, &key, bio, &cell1, &cell_result))
-			return DM_MAPIO_SUBMITTED;
-
 		build_data_key(tc->td, result.block, &key);
 		if (dm_bio_detain(tc->pool->prison, &key, bio, &cell2, &cell_result)) {
 			cell_defer_no_holder_no_free(tc, &cell1);
@@ -1921,6 +1991,7 @@ static int thin_bio_map(struct dm_target *ti, struct bio *bio)
 			 * of doing so.
 			 */
 			handle_unserviceable_bio(tc->pool, bio);
+			cell_defer_no_holder_no_free(tc, &cell1);
 			return DM_MAPIO_SUBMITTED;
 		}
 		/* fall through */
@@ -1931,6 +2002,7 @@ static int thin_bio_map(struct dm_target *ti, struct bio *bio)
 		 * provide the hint to load the metadata into cache.
 		 */
 		thin_defer_bio(tc, bio);
+		cell_defer_no_holder_no_free(tc, &cell1);
 		return DM_MAPIO_SUBMITTED;
 
 	default:
@@ -1940,6 +2012,7 @@ static int thin_bio_map(struct dm_target *ti, struct bio *bio)
 		 * pool is switched to fail-io mode.
 		 */
 		bio_io_error(bio);
+		cell_defer_no_holder_no_free(tc, &cell1);
 		return DM_MAPIO_SUBMITTED;
 	}
 }
@@ -3112,7 +3185,7 @@ static void pool_io_hints(struct dm_target *ti, struct queue_limits *limits)
 	 */
 	if (io_opt_sectors < pool->sectors_per_block ||
 	    do_div(io_opt_sectors, pool->sectors_per_block)) {
-		blk_limits_io_min(limits, 0);
+		blk_limits_io_min(limits, pool->sectors_per_block << SECTOR_SHIFT);
 		blk_limits_io_opt(limits, pool->sectors_per_block << SECTOR_SHIFT);
 	}
 
@@ -3141,7 +3214,7 @@ static struct target_type pool_target = {
 	.name = "thin-pool",
 	.features = DM_TARGET_SINGLETON | DM_TARGET_ALWAYS_WRITEABLE |
 		    DM_TARGET_IMMUTABLE,
-	.version = {1, 12, 0},
+	.version = {1, 13, 0},
 	.module = THIS_MODULE,
 	.ctr = pool_ctr,
 	.dtr = pool_dtr,
@@ -3361,8 +3434,7 @@ static int thin_endio(struct dm_target *ti, struct bio *bio, int err)
 		spin_lock_irqsave(&pool->lock, flags);
 		list_for_each_entry_safe(m, tmp, &work, list) {
 			list_del(&m->list);
-			m->quiesced = true;
-			__maybe_add_mapping(m);
+			__complete_mapping_preparation(m);
 		}
 		spin_unlock_irqrestore(&pool->lock, flags);
 	}
@@ -3401,6 +3473,16 @@ static void thin_postsuspend(struct dm_target *ti)
 	noflush_work(tc, do_noflush_stop);
 }
 
+static int thin_preresume(struct dm_target *ti)
+{
+	struct thin_c *tc = ti->private;
+
+	if (tc->origin_dev)
+		tc->origin_size = get_dev_size(tc->origin_dev->bdev);
+
+	return 0;
+}
+
 /*
  * <nr mapped sectors> <highest mapped sector>
  */
@@ -3483,12 +3565,13 @@ static int thin_iterate_devices(struct dm_target *ti,
 
 static struct target_type thin_target = {
 	.name = "thin",
-	.version = {1, 12, 0},
+	.version = {1, 13, 0},
 	.module	= THIS_MODULE,
 	.ctr = thin_ctr,
 	.dtr = thin_dtr,
 	.map = thin_map,
 	.end_io = thin_endio,
+	.preresume = thin_preresume,
 	.presuspend = thin_presuspend,
 	.postsuspend = thin_postsuspend,
 	.status = thin_status,
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 32b958dbc499..58f3927fd7cc 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -142,6 +142,9 @@ struct mapped_device {
 	 */
 	struct dm_table *map;
 
+	struct list_head table_devices;
+	struct mutex table_devices_lock;
+
 	unsigned long flags;
 
 	struct request_queue *queue;
@@ -212,6 +215,12 @@ struct dm_md_mempools {
 	struct bio_set *bs;
 };
 
+struct table_device {
+	struct list_head list;
+	atomic_t count;
+	struct dm_dev dm_dev;
+};
+
 #define RESERVED_BIO_BASED_IOS		16
 #define RESERVED_REQUEST_BASED_IOS	256
 #define RESERVED_MAX_IOS		1024
@@ -670,6 +679,120 @@ static void dm_put_live_table_fast(struct mapped_device *md) __releases(RCU)
 }
 
 /*
+ * Open a table device so we can use it as a map destination.
+ */
+static int open_table_device(struct table_device *td, dev_t dev,
+			     struct mapped_device *md)
+{
+	static char *_claim_ptr = "I belong to device-mapper";
+	struct block_device *bdev;
+
+	int r;
+
+	BUG_ON(td->dm_dev.bdev);
+
+	bdev = blkdev_get_by_dev(dev, td->dm_dev.mode | FMODE_EXCL, _claim_ptr);
+	if (IS_ERR(bdev))
+		return PTR_ERR(bdev);
+
+	r = bd_link_disk_holder(bdev, dm_disk(md));
+	if (r) {
+		blkdev_put(bdev, td->dm_dev.mode | FMODE_EXCL);
+		return r;
+	}
+
+	td->dm_dev.bdev = bdev;
+	return 0;
+}
+
+/*
+ * Close a table device that we've been using.
+ */
+static void close_table_device(struct table_device *td, struct mapped_device *md)
+{
+	if (!td->dm_dev.bdev)
+		return;
+
+	bd_unlink_disk_holder(td->dm_dev.bdev, dm_disk(md));
+	blkdev_put(td->dm_dev.bdev, td->dm_dev.mode | FMODE_EXCL);
+	td->dm_dev.bdev = NULL;
+}
+
+static struct table_device *find_table_device(struct list_head *l, dev_t dev,
+					      fmode_t mode) {
+	struct table_device *td;
+
+	list_for_each_entry(td, l, list)
+		if (td->dm_dev.bdev->bd_dev == dev && td->dm_dev.mode == mode)
+			return td;
+
+	return NULL;
+}
+
+int dm_get_table_device(struct mapped_device *md, dev_t dev, fmode_t mode,
+			struct dm_dev **result) {
+	int r;
+	struct table_device *td;
+
+	mutex_lock(&md->table_devices_lock);
+	td = find_table_device(&md->table_devices, dev, mode);
+	if (!td) {
+		td = kmalloc(sizeof(*td), GFP_KERNEL);
+		if (!td) {
+			mutex_unlock(&md->table_devices_lock);
+			return -ENOMEM;
+		}
+
+		td->dm_dev.mode = mode;
+		td->dm_dev.bdev = NULL;
+
+		if ((r = open_table_device(td, dev, md))) {
+			mutex_unlock(&md->table_devices_lock);
+			kfree(td);
+			return r;
+		}
+
+		format_dev_t(td->dm_dev.name, dev);
+
+		atomic_set(&td->count, 0);
+		list_add(&td->list, &md->table_devices);
+	}
+	atomic_inc(&td->count);
+	mutex_unlock(&md->table_devices_lock);
+
+	*result = &td->dm_dev;
+	return 0;
+}
+EXPORT_SYMBOL_GPL(dm_get_table_device);
+
+void dm_put_table_device(struct mapped_device *md, struct dm_dev *d)
+{
+	struct table_device *td = container_of(d, struct table_device, dm_dev);
+
+	mutex_lock(&md->table_devices_lock);
+	if (atomic_dec_and_test(&td->count)) {
+		close_table_device(td, md);
+		list_del(&td->list);
+		kfree(td);
+	}
+	mutex_unlock(&md->table_devices_lock);
+}
+EXPORT_SYMBOL(dm_put_table_device);
+
+static void free_table_devices(struct list_head *devices)
+{
+	struct list_head *tmp, *next;
+
+	list_for_each_safe(tmp, next, devices) {
+		struct table_device *td = list_entry(tmp, struct table_device, list);
+
+		DMWARN("dm_destroy: %s still exists with %d references",
+		       td->dm_dev.name, atomic_read(&td->count));
+		kfree(td);
+	}
+}
+
+/*
  * Get the geometry associated with a dm device
  */
 int dm_get_geometry(struct mapped_device *md, struct hd_geometry *geo)
@@ -1249,13 +1372,13 @@ static void clone_bio(struct dm_target_io *tio, struct bio *bio,
 }
 
 static struct dm_target_io *alloc_tio(struct clone_info *ci,
-				      struct dm_target *ti, int nr_iovecs,
+				      struct dm_target *ti,
 				      unsigned target_bio_nr)
 {
 	struct dm_target_io *tio;
 	struct bio *clone;
 
-	clone = bio_alloc_bioset(GFP_NOIO, nr_iovecs, ci->md->bs);
+	clone = bio_alloc_bioset(GFP_NOIO, 0, ci->md->bs);
 	tio = container_of(clone, struct dm_target_io, clone);
 
 	tio->io = ci->io;
@@ -1269,17 +1392,12 @@ static void __clone_and_map_simple_bio(struct clone_info *ci,
 				       struct dm_target *ti,
 				       unsigned target_bio_nr, unsigned *len)
 {
-	struct dm_target_io *tio = alloc_tio(ci, ti, ci->bio->bi_max_vecs, target_bio_nr);
+	struct dm_target_io *tio = alloc_tio(ci, ti, target_bio_nr);
 	struct bio *clone = &tio->clone;
 
 	tio->len_ptr = len;
 
-	/*
-	 * Discard requests require the bio's inline iovecs be initialized.
-	 * ci->bio->bi_max_vecs is BIO_INLINE_VECS anyway, for both flush
-	 * and discard, so no need for concern about wasted bvec allocations.
-	 */
-	 __bio_clone_fast(clone, ci->bio);
+	__bio_clone_fast(clone, ci->bio);
 	if (len)
 		bio_setup_sector(clone, ci->sector, *len);
 
@@ -1322,7 +1440,7 @@ static void __clone_and_map_data_bio(struct clone_info *ci, struct dm_target *ti
 		num_target_bios = ti->num_write_bios(ti, bio);
 
 	for (target_bio_nr = 0; target_bio_nr < num_target_bios; target_bio_nr++) {
-		tio = alloc_tio(ci, ti, 0, target_bio_nr);
+		tio = alloc_tio(ci, ti, target_bio_nr);
 		tio->len_ptr = len;
 		clone_bio(tio, bio, sector, *len);
 		__map_bio(tio);
@@ -1949,12 +2067,14 @@ static struct mapped_device *alloc_dev(int minor)
 	md->type = DM_TYPE_NONE;
 	mutex_init(&md->suspend_lock);
 	mutex_init(&md->type_lock);
+	mutex_init(&md->table_devices_lock);
 	spin_lock_init(&md->deferred_lock);
 	atomic_set(&md->holders, 1);
 	atomic_set(&md->open_count, 0);
 	atomic_set(&md->event_nr, 0);
 	atomic_set(&md->uevent_seq, 0);
 	INIT_LIST_HEAD(&md->uevent_list);
+	INIT_LIST_HEAD(&md->table_devices);
 	spin_lock_init(&md->uevent_lock);
 
 	md->queue = blk_alloc_queue(GFP_KERNEL);
@@ -2040,6 +2160,7 @@ static void free_dev(struct mapped_device *md)
 	blk_integrity_unregister(md->disk);
 	del_gendisk(md->disk);
 	cleanup_srcu_struct(&md->io_barrier);
+	free_table_devices(&md->table_devices);
 	free_minor(minor);
 
 	spin_lock(&_minor_lock);
@@ -2900,7 +3021,7 @@ struct dm_md_mempools *dm_alloc_md_mempools(unsigned type, unsigned integrity, u
 	if (!pools->io_pool)
 		goto out;
 
-	pools->bs = bioset_create(pool_size, front_pad);
+	pools->bs = bioset_create_nobvec(pool_size, front_pad);
 	if (!pools->bs)
 		goto out;
 
diff --git a/drivers/md/dm.h b/drivers/md/dm.h
index ed76126aac54..988c7fb7b145 100644
--- a/drivers/md/dm.h
+++ b/drivers/md/dm.h
@@ -44,7 +44,7 @@
 struct dm_dev_internal {
 	struct list_head list;
 	atomic_t count;
-	struct dm_dev dm_dev;
+	struct dm_dev *dm_dev;
 };
 
 struct dm_table;
@@ -72,7 +72,6 @@ int dm_table_any_busy_target(struct dm_table *t);
 unsigned dm_table_get_type(struct dm_table *t);
 struct target_type *dm_table_get_immutable_target_type(struct dm_table *t);
 bool dm_table_request_based(struct dm_table *t);
-bool dm_table_supports_discards(struct dm_table *t);
 void dm_table_free_md_mempools(struct dm_table *t);
 struct dm_md_mempools *dm_table_get_md_mempools(struct dm_table *t);
 
@@ -189,6 +188,9 @@ int dm_cancel_deferred_remove(struct mapped_device *md);
 int dm_request_based(struct mapped_device *md);
 sector_t dm_get_size(struct mapped_device *md);
 struct request_queue *dm_get_md_queue(struct mapped_device *md);
+int dm_get_table_device(struct mapped_device *md, dev_t dev, fmode_t mode,
+			struct dm_dev **result);
+void dm_put_table_device(struct mapped_device *md, struct dm_dev *d);
 struct dm_stats *dm_get_stats(struct mapped_device *md);
 
 int dm_kobject_uevent(struct mapped_device *md, enum kobject_action action,
diff --git a/drivers/md/linear.c b/drivers/md/linear.c
index 56f534b4a2d2..64713b77df1c 100644
--- a/drivers/md/linear.c
+++ b/drivers/md/linear.c
@@ -10,10 +10,10 @@
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation; either version 2, or (at your option)
    any later version.
-   
+
    You should have received a copy of the GNU General Public License
    (for example /usr/src/linux/COPYING); if not, write to the Free
-   Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.  
+   Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 */
 
 #include <linux/blkdev.h>
@@ -25,7 +25,7 @@
 #include "linear.h"
 
 /*
- * find which device holds a particular offset 
+ * find which device holds a particular offset
  */
 static inline struct dev_info *which_dev(struct mddev *mddev, sector_t sector)
 {
@@ -355,7 +355,6 @@ static void linear_status (struct seq_file *seq, struct mddev *mddev)
 	seq_printf(seq, " %dk rounding", mddev->chunk_sectors / 2);
 }
 
-
 static struct md_personality linear_personality =
 {
 	.name		= "linear",
@@ -379,7 +378,6 @@ static void linear_exit (void)
 	unregister_md_personality (&linear_personality);
 }
 
-
 module_init(linear_init);
 module_exit(linear_exit);
 MODULE_LICENSE("GPL");
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 32fc19c540d4..9233c71138f1 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -1,6 +1,6 @@
 /*
    md.c : Multiple Devices driver for Linux
-	  Copyright (C) 1998, 1999, 2000 Ingo Molnar
+     Copyright (C) 1998, 1999, 2000 Ingo Molnar
 
      completely rewritten, based on the MD driver code from Marc Zyngier
 
@@ -66,8 +66,6 @@ static void autostart_arrays(int part);
 static LIST_HEAD(pers_list);
 static DEFINE_SPINLOCK(pers_lock);
 
-static void md_print_devices(void);
-
 static DECLARE_WAIT_QUEUE_HEAD(resync_wait);
 static struct workqueue_struct *md_wq;
 static struct workqueue_struct *md_misc_wq;
@@ -75,8 +73,6 @@ static struct workqueue_struct *md_misc_wq;
 static int remove_and_add_spares(struct mddev *mddev,
 				 struct md_rdev *this);
 
-#define MD_BUG(x...) { printk("md: bug in file %s, line %d\n", __FILE__, __LINE__); md_print_devices(); }
-
 /*
  * Default number of read corrections we'll attempt on an rdev
  * before ejecting it from the array. We divide the read error
@@ -218,7 +214,6 @@ static void md_new_event_inintr(struct mddev *mddev)
 static LIST_HEAD(all_mddevs);
 static DEFINE_SPINLOCK(all_mddevs_lock);
 
-
 /*
  * iterates through all used mddevs in the system.
  * We take care to grab the all_mddevs_lock whenever navigating
@@ -228,7 +223,7 @@ static DEFINE_SPINLOCK(all_mddevs_lock);
  */
 #define for_each_mddev(_mddev,_tmp)					\
 									\
-	for (({ spin_lock(&all_mddevs_lock); 				\
+	for (({ spin_lock(&all_mddevs_lock);				\
 		_tmp = all_mddevs.next;					\
 		_mddev = NULL;});					\
 	     ({ if (_tmp != &all_mddevs)				\
@@ -241,7 +236,6 @@ static DEFINE_SPINLOCK(all_mddevs_lock);
 		_tmp = _tmp->next;})					\
 		)
 
-
 /* Rather than calling directly into the personality make_request function,
  * IO requests come here first so that we can check if the device is
  * being suspended pending a reconfiguration.
@@ -488,7 +482,7 @@ void mddev_init(struct mddev *mddev)
 }
 EXPORT_SYMBOL_GPL(mddev_init);
 
-static struct mddev * mddev_find(dev_t unit)
+static struct mddev *mddev_find(dev_t unit)
 {
 	struct mddev *mddev, *new = NULL;
 
@@ -530,7 +524,7 @@ static struct mddev * mddev_find(dev_t unit)
 				kfree(new);
 				return NULL;
 			}
-				
+
 			is_free = 1;
 			list_for_each_entry(mddev, &all_mddevs, all_mddevs)
 				if (mddev->unit == dev) {
@@ -562,7 +556,7 @@ static struct mddev * mddev_find(dev_t unit)
 	goto retry;
 }
 
-static inline int __must_check mddev_lock(struct mddev * mddev)
+static inline int __must_check mddev_lock(struct mddev *mddev)
 {
 	return mutex_lock_interruptible(&mddev->reconfig_mutex);
 }
@@ -570,7 +564,7 @@ static inline int __must_check mddev_lock(struct mddev * mddev)
 /* Sometimes we need to take the lock in a situation where
  * failure due to interrupts is not acceptable.
  */
-static inline void mddev_lock_nointr(struct mddev * mddev)
+static inline void mddev_lock_nointr(struct mddev *mddev)
 {
 	mutex_lock(&mddev->reconfig_mutex);
 }
@@ -580,14 +574,14 @@ static inline int mddev_is_locked(struct mddev *mddev)
 	return mutex_is_locked(&mddev->reconfig_mutex);
 }
 
-static inline int mddev_trylock(struct mddev * mddev)
+static inline int mddev_trylock(struct mddev *mddev)
 {
 	return mutex_trylock(&mddev->reconfig_mutex);
 }
 
 static struct attribute_group md_redundancy_group;
 
-static void mddev_unlock(struct mddev * mddev)
+static void mddev_unlock(struct mddev *mddev)
 {
 	if (mddev->to_remove) {
 		/* These cannot be removed under reconfig_mutex as
@@ -630,17 +624,6 @@ static void mddev_unlock(struct mddev * mddev)
 	spin_unlock(&pers_lock);
 }
 
-static struct md_rdev * find_rdev_nr(struct mddev *mddev, int nr)
-{
-	struct md_rdev *rdev;
-
-	rdev_for_each(rdev, mddev)
-		if (rdev->desc_nr == nr)
-			return rdev;
-
-	return NULL;
-}
-
 static struct md_rdev *find_rdev_nr_rcu(struct mddev *mddev, int nr)
 {
 	struct md_rdev *rdev;
@@ -693,11 +676,8 @@ static inline sector_t calc_dev_sboffset(struct md_rdev *rdev)
 	return MD_NEW_SIZE_SECTORS(num_sectors);
 }
 
-static int alloc_disk_sb(struct md_rdev * rdev)
+static int alloc_disk_sb(struct md_rdev *rdev)
 {
-	if (rdev->sb_page)
-		MD_BUG();
-
 	rdev->sb_page = alloc_page(GFP_KERNEL);
 	if (!rdev->sb_page) {
 		printk(KERN_ALERT "md: out of memory.\n");
@@ -766,14 +746,7 @@ void md_super_write(struct mddev *mddev, struct md_rdev *rdev,
 void md_super_wait(struct mddev *mddev)
 {
 	/* wait for all superblock writes that were scheduled to complete */
-	DEFINE_WAIT(wq);
-	for(;;) {
-		prepare_to_wait(&mddev->sb_wait, &wq, TASK_UNINTERRUPTIBLE);
-		if (atomic_read(&mddev->pending_writes)==0)
-			break;
-		schedule();
-	}
-	finish_wait(&mddev->sb_wait, &wq);
+	wait_event(mddev->sb_wait, atomic_read(&mddev->pending_writes)==0);
 }
 
 int sync_page_io(struct md_rdev *rdev, sector_t sector, int size,
@@ -801,17 +774,13 @@ int sync_page_io(struct md_rdev *rdev, sector_t sector, int size,
 }
 EXPORT_SYMBOL_GPL(sync_page_io);
 
-static int read_disk_sb(struct md_rdev * rdev, int size)
+static int read_disk_sb(struct md_rdev *rdev, int size)
 {
 	char b[BDEVNAME_SIZE];
-	if (!rdev->sb_page) {
-		MD_BUG();
-		return -EINVAL;
-	}
+
 	if (rdev->sb_loaded)
 		return 0;
 
-
 	if (!sync_page_io(rdev, 0, size, rdev->sb_page, READ, true))
 		goto fail;
 	rdev->sb_loaded = 1;
@@ -825,7 +794,7 @@ fail:
 
 static int uuid_equal(mdp_super_t *sb1, mdp_super_t *sb2)
 {
-	return 	sb1->set_uuid0 == sb2->set_uuid0 &&
+	return	sb1->set_uuid0 == sb2->set_uuid0 &&
 		sb1->set_uuid1 == sb2->set_uuid1 &&
 		sb1->set_uuid2 == sb2->set_uuid2 &&
 		sb1->set_uuid3 == sb2->set_uuid3;
@@ -861,14 +830,13 @@ abort:
 	return ret;
 }
 
-
 static u32 md_csum_fold(u32 csum)
 {
 	csum = (csum & 0xffff) + (csum >> 16);
 	return (csum & 0xffff) + (csum >> 16);
 }
 
-static unsigned int calc_sb_csum(mdp_super_t * sb)
+static unsigned int calc_sb_csum(mdp_super_t *sb)
 {
 	u64 newcsum = 0;
 	u32 *sb32 = (u32*)sb;
@@ -882,7 +850,6 @@ static unsigned int calc_sb_csum(mdp_super_t * sb)
 		newcsum += sb32[i];
 	csum = (newcsum & 0xffffffff) + (newcsum>>32);
 
-
 #ifdef CONFIG_ALPHA
 	/* This used to use csum_partial, which was wrong for several
 	 * reasons including that different results are returned on
@@ -899,7 +866,6 @@ static unsigned int calc_sb_csum(mdp_super_t * sb)
 	return csum;
 }
 
-
 /*
  * Handle superblock details.
  * We want to be able to handle multiple superblock formats
@@ -965,7 +931,7 @@ int md_check_no_bitmap(struct mddev *mddev)
 EXPORT_SYMBOL(md_check_no_bitmap);
 
 /*
- * load_super for 0.90.0 
+ * load_super for 0.90.0
  */
 static int super_90_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor_version)
 {
@@ -1044,7 +1010,7 @@ static int super_90_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor
 		ev2 = md_event(refsb);
 		if (ev1 > ev2)
 			ret = 1;
-		else 
+		else
 			ret = 0;
 	}
 	rdev->sectors = rdev->sb_start;
@@ -1118,7 +1084,7 @@ static int super_90_validate(struct mddev *mddev, struct md_rdev *rdev)
 		if (sb->state & (1<<MD_SB_CLEAN))
 			mddev->recovery_cp = MaxSector;
 		else {
-			if (sb->events_hi == sb->cp_events_hi && 
+			if (sb->events_hi == sb->cp_events_hi &&
 				sb->events_lo == sb->cp_events_lo) {
 				mddev->recovery_cp = sb->recovery_cp;
 			} else
@@ -1146,7 +1112,7 @@ static int super_90_validate(struct mddev *mddev, struct md_rdev *rdev)
 		++ev1;
 		if (sb->disks[rdev->desc_nr].state & (
 			    (1<<MD_DISK_SYNC) | (1 << MD_DISK_ACTIVE)))
-			if (ev1 < mddev->events) 
+			if (ev1 < mddev->events)
 				return -EINVAL;
 	} else if (mddev->bitmap) {
 		/* if adding to array with a bitmap, then we can accept an
@@ -1197,7 +1163,6 @@ static void super_90_sync(struct mddev *mddev, struct md_rdev *rdev)
 	struct md_rdev *rdev2;
 	int next_spare = mddev->raid_disks;
 
-
 	/* make rdev->sb match mddev data..
 	 *
 	 * 1/ zero out disks
@@ -1366,7 +1331,7 @@ super_90_allow_new_offset(struct md_rdev *rdev, unsigned long long new_offset)
  * version 1 superblock
  */
 
-static __le32 calc_sb_1_csum(struct mdp_superblock_1 * sb)
+static __le32 calc_sb_1_csum(struct mdp_superblock_1 *sb)
 {
 	__le32 disk_csum;
 	u32 csum;
@@ -1430,7 +1395,6 @@ static int super_1_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor_
 	ret = read_disk_sb(rdev, 4096);
 	if (ret) return ret;
 
-
 	sb = page_address(rdev->sb_page);
 
 	if (sb->magic != cpu_to_le32(MD_SB_MAGIC) ||
@@ -1817,7 +1781,7 @@ retry:
 
 	for (i=0; i<max_dev;i++)
 		sb->dev_roles[i] = cpu_to_le16(0xfffe);
-	
+
 	rdev_for_each(rdev2, mddev) {
 		i = rdev2->desc_nr;
 		if (test_bit(Faulty, &rdev2->flags))
@@ -2033,18 +1997,13 @@ void md_integrity_add_rdev(struct md_rdev *rdev, struct mddev *mddev)
 }
 EXPORT_SYMBOL(md_integrity_add_rdev);
 
-static int bind_rdev_to_array(struct md_rdev * rdev, struct mddev * mddev)
+static int bind_rdev_to_array(struct md_rdev *rdev, struct mddev *mddev)
 {
 	char b[BDEVNAME_SIZE];
 	struct kobject *ko;
 	char *s;
 	int err;
 
-	if (rdev->mddev) {
-		MD_BUG();
-		return -EINVAL;
-	}
-
 	/* prevent duplicates */
 	if (find_rdev(mddev, rdev->bdev->bd_dev))
 		return -EEXIST;
@@ -2067,16 +2026,21 @@ static int bind_rdev_to_array(struct md_rdev * rdev, struct mddev * mddev)
 	 * If it is -1, assign a free number, else
 	 * check number is not in use
 	 */
+	rcu_read_lock();
 	if (rdev->desc_nr < 0) {
 		int choice = 0;
-		if (mddev->pers) choice = mddev->raid_disks;
-		while (find_rdev_nr(mddev, choice))
+		if (mddev->pers)
+			choice = mddev->raid_disks;
+		while (find_rdev_nr_rcu(mddev, choice))
 			choice++;
 		rdev->desc_nr = choice;
 	} else {
-		if (find_rdev_nr(mddev, rdev->desc_nr))
+		if (find_rdev_nr_rcu(mddev, rdev->desc_nr)) {
+			rcu_read_unlock();
 			return -EBUSY;
+		}
 	}
+	rcu_read_unlock();
 	if (mddev->max_disks && rdev->desc_nr >= mddev->max_disks) {
 		printk(KERN_WARNING "md: %s: array is limited to %d devices\n",
 		       mdname(mddev), mddev->max_disks);
@@ -2118,13 +2082,10 @@ static void md_delayed_delete(struct work_struct *ws)
 	kobject_put(&rdev->kobj);
 }
 
-static void unbind_rdev_from_array(struct md_rdev * rdev)
+static void unbind_rdev_from_array(struct md_rdev *rdev)
 {
 	char b[BDEVNAME_SIZE];
-	if (!rdev->mddev) {
-		MD_BUG();
-		return;
-	}
+
 	bd_unlink_disk_holder(rdev->bdev, rdev->mddev->gendisk);
 	list_del_rcu(&rdev->same_set);
 	printk(KERN_INFO "md: unbind<%s>\n", bdevname(rdev->bdev,b));
@@ -2169,20 +2130,17 @@ static void unlock_rdev(struct md_rdev *rdev)
 {
 	struct block_device *bdev = rdev->bdev;
 	rdev->bdev = NULL;
-	if (!bdev)
-		MD_BUG();
 	blkdev_put(bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
 }
 
 void md_autodetect_dev(dev_t dev);
 
-static void export_rdev(struct md_rdev * rdev)
+static void export_rdev(struct md_rdev *rdev)
 {
 	char b[BDEVNAME_SIZE];
+
 	printk(KERN_INFO "md: export_rdev(%s)\n",
 		bdevname(rdev->bdev,b));
-	if (rdev->mddev)
-		MD_BUG();
 	md_rdev_clear(rdev);
 #ifndef MODULE
 	if (test_bit(AutoDetected, &rdev->flags))
@@ -2192,7 +2150,7 @@ static void export_rdev(struct md_rdev * rdev)
 	kobject_put(&rdev->kobj);
 }
 
-static void kick_rdev_from_array(struct md_rdev * rdev)
+static void kick_rdev_from_array(struct md_rdev *rdev)
 {
 	unbind_rdev_from_array(rdev);
 	export_rdev(rdev);
@@ -2200,153 +2158,18 @@ static void kick_rdev_from_array(struct md_rdev * rdev)
 
 static void export_array(struct mddev *mddev)
 {
-	struct md_rdev *rdev, *tmp;
+	struct md_rdev *rdev;
 
-	rdev_for_each_safe(rdev, tmp, mddev) {
-		if (!rdev->mddev) {
-			MD_BUG();
-			continue;
-		}
+	while (!list_empty(&mddev->disks)) {
+		rdev = list_first_entry(&mddev->disks, struct md_rdev,
+					same_set);
 		kick_rdev_from_array(rdev);
 	}
-	if (!list_empty(&mddev->disks))
-		MD_BUG();
 	mddev->raid_disks = 0;
 	mddev->major_version = 0;
 }
 
-static void print_desc(mdp_disk_t *desc)
-{
-	printk(" DISK<N:%d,(%d,%d),R:%d,S:%d>\n", desc->number,
-		desc->major,desc->minor,desc->raid_disk,desc->state);
-}
-
-static void print_sb_90(mdp_super_t *sb)
-{
-	int i;
-
-	printk(KERN_INFO 
-		"md:  SB: (V:%d.%d.%d) ID:<%08x.%08x.%08x.%08x> CT:%08x\n",
-		sb->major_version, sb->minor_version, sb->patch_version,
-		sb->set_uuid0, sb->set_uuid1, sb->set_uuid2, sb->set_uuid3,
-		sb->ctime);
-	printk(KERN_INFO "md:     L%d S%08d ND:%d RD:%d md%d LO:%d CS:%d\n",
-		sb->level, sb->size, sb->nr_disks, sb->raid_disks,
-		sb->md_minor, sb->layout, sb->chunk_size);
-	printk(KERN_INFO "md:     UT:%08x ST:%d AD:%d WD:%d"
-		" FD:%d SD:%d CSUM:%08x E:%08lx\n",
-		sb->utime, sb->state, sb->active_disks, sb->working_disks,
-		sb->failed_disks, sb->spare_disks,
-		sb->sb_csum, (unsigned long)sb->events_lo);
-
-	printk(KERN_INFO);
-	for (i = 0; i < MD_SB_DISKS; i++) {
-		mdp_disk_t *desc;
-
-		desc = sb->disks + i;
-		if (desc->number || desc->major || desc->minor ||
-		    desc->raid_disk || (desc->state && (desc->state != 4))) {
-			printk("     D %2d: ", i);
-			print_desc(desc);
-		}
-	}
-	printk(KERN_INFO "md:     THIS: ");
-	print_desc(&sb->this_disk);
-}
-
-static void print_sb_1(struct mdp_superblock_1 *sb)
-{
-	__u8 *uuid;
-
-	uuid = sb->set_uuid;
-	printk(KERN_INFO
-	       "md:  SB: (V:%u) (F:0x%08x) Array-ID:<%pU>\n"
-	       "md:    Name: \"%s\" CT:%llu\n",
-		le32_to_cpu(sb->major_version),
-		le32_to_cpu(sb->feature_map),
-		uuid,
-		sb->set_name,
-		(unsigned long long)le64_to_cpu(sb->ctime)
-		       & MD_SUPERBLOCK_1_TIME_SEC_MASK);
-
-	uuid = sb->device_uuid;
-	printk(KERN_INFO
-	       "md:       L%u SZ%llu RD:%u LO:%u CS:%u DO:%llu DS:%llu SO:%llu"
-			" RO:%llu\n"
-	       "md:     Dev:%08x UUID: %pU\n"
-	       "md:       (F:0x%08x) UT:%llu Events:%llu ResyncOffset:%llu CSUM:0x%08x\n"
-	       "md:         (MaxDev:%u) \n",
-		le32_to_cpu(sb->level),
-		(unsigned long long)le64_to_cpu(sb->size),
-		le32_to_cpu(sb->raid_disks),
-		le32_to_cpu(sb->layout),
-		le32_to_cpu(sb->chunksize),
-		(unsigned long long)le64_to_cpu(sb->data_offset),
-		(unsigned long long)le64_to_cpu(sb->data_size),
-		(unsigned long long)le64_to_cpu(sb->super_offset),
-		(unsigned long long)le64_to_cpu(sb->recovery_offset),
-		le32_to_cpu(sb->dev_number),
-		uuid,
-		sb->devflags,
-		(unsigned long long)le64_to_cpu(sb->utime) & MD_SUPERBLOCK_1_TIME_SEC_MASK,
-		(unsigned long long)le64_to_cpu(sb->events),
-		(unsigned long long)le64_to_cpu(sb->resync_offset),
-		le32_to_cpu(sb->sb_csum),
-		le32_to_cpu(sb->max_dev)
-		);
-}
-
-static void print_rdev(struct md_rdev *rdev, int major_version)
-{
-	char b[BDEVNAME_SIZE];
-	printk(KERN_INFO "md: rdev %s, Sect:%08llu F:%d S:%d DN:%u\n",
-		bdevname(rdev->bdev, b), (unsigned long long)rdev->sectors,
-	        test_bit(Faulty, &rdev->flags), test_bit(In_sync, &rdev->flags),
-	        rdev->desc_nr);
-	if (rdev->sb_loaded) {
-		printk(KERN_INFO "md: rdev superblock (MJ:%d):\n", major_version);
-		switch (major_version) {
-		case 0:
-			print_sb_90(page_address(rdev->sb_page));
-			break;
-		case 1:
-			print_sb_1(page_address(rdev->sb_page));
-			break;
-		}
-	} else
-		printk(KERN_INFO "md: no rdev superblock!\n");
-}
-
-static void md_print_devices(void)
-{
-	struct list_head *tmp;
-	struct md_rdev *rdev;
-	struct mddev *mddev;
-	char b[BDEVNAME_SIZE];
-
-	printk("\n");
-	printk("md:	**********************************\n");
-	printk("md:	* <COMPLETE RAID STATE PRINTOUT> *\n");
-	printk("md:	**********************************\n");
-	for_each_mddev(mddev, tmp) {
-
-		if (mddev->bitmap)
-			bitmap_print_sb(mddev->bitmap);
-		else
-			printk("%s: ", mdname(mddev));
-		rdev_for_each(rdev, mddev)
-			printk("<%s>", bdevname(rdev->bdev,b));
-		printk("\n");
-
-		rdev_for_each(rdev, mddev)
-			print_rdev(rdev, mddev->major_version);
-	}
-	printk("md:	**********************************\n");
-	printk("\n");
-}
-
-
-static void sync_sbs(struct mddev * mddev, int nospares)
+static void sync_sbs(struct mddev *mddev, int nospares)
 {
 	/* Update each superblock (in-memory image), but
 	 * if we are allowed to, skip spares which already
@@ -2369,7 +2192,7 @@ static void sync_sbs(struct mddev * mddev, int nospares)
 	}
 }
 
-static void md_update_sb(struct mddev * mddev, int force_change)
+static void md_update_sb(struct mddev *mddev, int force_change)
 {
 	struct md_rdev *rdev;
 	int sync_req;
@@ -2390,7 +2213,7 @@ repeat:
 		    mddev->curr_resync_completed > rdev->recovery_offset)
 				rdev->recovery_offset = mddev->curr_resync_completed;
 
-	}	
+	}
 	if (!mddev->persistent) {
 		clear_bit(MD_CHANGE_CLEAN, &mddev->flags);
 		clear_bit(MD_CHANGE_DEVS, &mddev->flags);
@@ -2453,15 +2276,12 @@ repeat:
 		mddev->can_decrease_events = nospares;
 	}
 
-	if (!mddev->events) {
-		/*
-		 * oops, this 64-bit counter should never wrap.
-		 * Either we are in around ~1 trillion A.C., assuming
-		 * 1 reboot per second, or we have a bug:
-		 */
-		MD_BUG();
-		mddev->events --;
-	}
+	/*
+	 * This 64-bit counter should never wrap.
+	 * Either we are in around ~1 trillion A.C., assuming
+	 * 1 reboot per second, or we have a bug...
+	 */
+	WARN_ON(mddev->events == 0);
 
 	rdev_for_each(rdev, mddev) {
 		if (rdev->badblocks.changed)
@@ -2668,10 +2488,12 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len)
 		set_bit(In_sync, &rdev->flags);
 		err = 0;
 	} else if (cmd_match(buf, "-insync") && rdev->raid_disk >= 0) {
-		clear_bit(In_sync, &rdev->flags);
-		rdev->saved_raid_disk = rdev->raid_disk;
-		rdev->raid_disk = -1;
-		err = 0;
+		if (rdev->mddev->pers == NULL) {
+			clear_bit(In_sync, &rdev->flags);
+			rdev->saved_raid_disk = rdev->raid_disk;
+			rdev->raid_disk = -1;
+			err = 0;
+		}
 	} else if (cmd_match(buf, "write_error")) {
 		set_bit(WriteErrorSeen, &rdev->flags);
 		err = 0;
@@ -2829,7 +2651,6 @@ slot_store(struct md_rdev *rdev, const char *buf, size_t len)
 	return len;
 }
 
-
 static struct rdev_sysfs_entry rdev_slot =
 __ATTR(slot, S_IRUGO|S_IWUSR, slot_show, slot_store);
 
@@ -2980,20 +2801,20 @@ rdev_size_store(struct md_rdev *rdev, const char *buf, size_t len)
 
 	rdev->sectors = sectors;
 	if (sectors > oldsectors && my_mddev->external) {
-		/* need to check that all other rdevs with the same ->bdev
-		 * do not overlap.  We need to unlock the mddev to avoid
-		 * a deadlock.  We have already changed rdev->sectors, and if
-		 * we have to change it back, we will have the lock again.
+		/* Need to check that all other rdevs with the same
+		 * ->bdev do not overlap.  'rcu' is sufficient to walk
+		 * the rdev lists safely.
+		 * This check does not provide a hard guarantee, it
+		 * just helps avoid dangerous mistakes.
 		 */
 		struct mddev *mddev;
 		int overlap = 0;
 		struct list_head *tmp;
 
-		mddev_unlock(my_mddev);
+		rcu_read_lock();
 		for_each_mddev(mddev, tmp) {
 			struct md_rdev *rdev2;
 
-			mddev_lock_nointr(mddev);
 			rdev_for_each(rdev2, mddev)
 				if (rdev->bdev == rdev2->bdev &&
 				    rdev != rdev2 &&
@@ -3003,13 +2824,12 @@ rdev_size_store(struct md_rdev *rdev, const char *buf, size_t len)
 					overlap = 1;
 					break;
 				}
-			mddev_unlock(mddev);
 			if (overlap) {
 				mddev_put(mddev);
 				break;
 			}
 		}
-		mddev_lock_nointr(my_mddev);
+		rcu_read_unlock();
 		if (overlap) {
 			/* Someone else could have slipped in a size
 			 * change here, but doing so is just silly.
@@ -3027,7 +2847,6 @@ rdev_size_store(struct md_rdev *rdev, const char *buf, size_t len)
 static struct rdev_sysfs_entry rdev_size =
 __ATTR(size, S_IRUGO|S_IWUSR, rdev_size_show, rdev_size_store);
 
-
 static ssize_t recovery_start_show(struct md_rdev *rdev, char *page)
 {
 	unsigned long long recovery_start = rdev->recovery_offset;
@@ -3063,7 +2882,6 @@ static ssize_t recovery_start_store(struct md_rdev *rdev, const char *buf, size_
 static struct rdev_sysfs_entry rdev_recovery_start =
 __ATTR(recovery_start, S_IRUGO|S_IWUSR, recovery_start_show, recovery_start_store);
 
-
 static ssize_t
 badblocks_show(struct badblocks *bb, char *page, int unack);
 static ssize_t
@@ -3084,7 +2902,6 @@ static ssize_t bb_store(struct md_rdev *rdev, const char *page, size_t len)
 static struct rdev_sysfs_entry rdev_bad_blocks =
 __ATTR(bad_blocks, S_IRUGO|S_IWUSR, bb_show, bb_store);
 
-
 static ssize_t ubb_show(struct md_rdev *rdev, char *page)
 {
 	return badblocks_show(&rdev->badblocks, page, 1);
@@ -3241,7 +3058,7 @@ static struct md_rdev *md_import_device(dev_t newdev, int super_format, int supe
 
 	size = i_size_read(rdev->bdev->bd_inode) >> BLOCK_SIZE_BITS;
 	if (!size) {
-		printk(KERN_WARNING 
+		printk(KERN_WARNING
 			"md: %s has zero or unknown size, marking faulty!\n",
 			bdevname(rdev->bdev,b));
 		err = -EINVAL;
@@ -3260,7 +3077,7 @@ static struct md_rdev *md_import_device(dev_t newdev, int super_format, int supe
 			goto abort_free;
 		}
 		if (err < 0) {
-			printk(KERN_WARNING 
+			printk(KERN_WARNING
 				"md: could not read %s's sb, not importing!\n",
 				bdevname(rdev->bdev,b));
 			goto abort_free;
@@ -3281,8 +3098,7 @@ abort_free:
  * Check a full RAID array for plausibility
  */
 
-
-static void analyze_sbs(struct mddev * mddev)
+static void analyze_sbs(struct mddev *mddev)
 {
 	int i;
 	struct md_rdev *rdev, *freshest, *tmp;
@@ -3300,12 +3116,11 @@ static void analyze_sbs(struct mddev * mddev)
 		default:
 			printk( KERN_ERR \
 				"md: fatal superblock inconsistency in %s"
-				" -- removing from array\n", 
+				" -- removing from array\n",
 				bdevname(rdev->bdev,b));
 			kick_rdev_from_array(rdev);
 		}
 
-
 	super_types[mddev->major_version].
 		validate_super(mddev, freshest);
 
@@ -3344,7 +3159,7 @@ static void analyze_sbs(struct mddev * mddev)
 /* Read a fixed-point number.
  * Numbers in sysfs attributes should be in "standard" units where
  * possible, so time should be in seconds.
- * However we internally use a a much smaller unit such as 
+ * However we internally use a a much smaller unit such as
  * milliseconds or jiffies.
  * This function takes a decimal number with a possible fractional
  * component, and produces an integer which is the result of
@@ -3381,7 +3196,6 @@ int strict_strtoul_scaled(const char *cp, unsigned long *res, int scale)
 	return 0;
 }
 
-
 static void md_safemode_timeout(unsigned long data);
 
 static ssize_t
@@ -3524,7 +3338,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 	/* Looks like we have a winner */
 	mddev_suspend(mddev);
 	mddev->pers->stop(mddev);
-	
+
 	if (mddev->pers->sync_request == NULL &&
 	    pers->sync_request != NULL) {
 		/* need to add the md_redundancy_group */
@@ -3533,7 +3347,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 			       "md: cannot register extra attributes for %s\n",
 			       mdname(mddev));
 		mddev->sysfs_action = sysfs_get_dirent(mddev->kobj.sd, "sync_action");
-	}		
+	}
 	if (mddev->pers->sync_request != NULL &&
 	    pers->sync_request == NULL) {
 		/* need to remove the md_redundancy_group */
@@ -3611,7 +3425,6 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 static struct md_sysfs_entry md_level =
 __ATTR(level, S_IRUGO|S_IWUSR, level_show, level_store);
 
-
 static ssize_t
 layout_show(struct mddev *mddev, char *page)
 {
@@ -3654,7 +3467,6 @@ layout_store(struct mddev *mddev, const char *buf, size_t len)
 static struct md_sysfs_entry md_layout =
 __ATTR(layout, S_IRUGO|S_IWUSR, layout_show, layout_store);
 
-
 static ssize_t
 raid_disks_show(struct mddev *mddev, char *page)
 {
@@ -3859,9 +3671,9 @@ array_state_show(struct mddev *mddev, char *page)
 	return sprintf(page, "%s\n", array_states[st]);
 }
 
-static int do_md_stop(struct mddev * mddev, int ro, struct block_device *bdev);
-static int md_set_readonly(struct mddev * mddev, struct block_device *bdev);
-static int do_md_run(struct mddev * mddev);
+static int do_md_stop(struct mddev *mddev, int ro, struct block_device *bdev);
+static int md_set_readonly(struct mddev *mddev, struct block_device *bdev);
+static int do_md_run(struct mddev *mddev);
 static int restart_array(struct mddev *mddev);
 
 static ssize_t
@@ -4012,7 +3824,6 @@ new_dev_store(struct mddev *mddev, const char *buf, size_t len)
 	    minor != MINOR(dev))
 		return -EOVERFLOW;
 
-
 	if (mddev->persistent) {
 		rdev = md_import_device(dev, mddev->major_version,
 					mddev->minor_version);
@@ -4108,7 +3919,6 @@ size_store(struct mddev *mddev, const char *buf, size_t len)
 static struct md_sysfs_entry md_size =
 __ATTR(component_size, S_IRUGO|S_IWUSR, size_show, size_store);
 
-
 /* Metadata version.
  * This is one of
  *   'none' for arrays with no metadata (good luck...)
@@ -4490,7 +4300,7 @@ suspend_lo_store(struct mddev *mddev, const char *buf, size_t len)
 	unsigned long long new = simple_strtoull(buf, &e, 10);
 	unsigned long long old = mddev->suspend_lo;
 
-	if (mddev->pers == NULL || 
+	if (mddev->pers == NULL ||
 	    mddev->pers->quiesce == NULL)
 		return -EINVAL;
 	if (buf == e || (*e && *e != '\n'))
@@ -4510,7 +4320,6 @@ suspend_lo_store(struct mddev *mddev, const char *buf, size_t len)
 static struct md_sysfs_entry md_suspend_lo =
 __ATTR(suspend_lo, S_IRUGO|S_IWUSR, suspend_lo_show, suspend_lo_store);
 
-
 static ssize_t
 suspend_hi_show(struct mddev *mddev, char *page)
 {
@@ -4698,7 +4507,6 @@ static struct attribute_group md_redundancy_group = {
 	.attrs = md_redundancy_attrs,
 };
 
-
 static ssize_t
 md_attr_show(struct kobject *kobj, struct attribute *attr, char *page)
 {
@@ -5111,7 +4919,7 @@ int md_run(struct mddev *mddev)
 	} else if (mddev->ro == 2) /* auto-readonly not meaningful */
 		mddev->ro = 0;
 
- 	atomic_set(&mddev->writes_pending,0);
+	atomic_set(&mddev->writes_pending,0);
 	atomic_set(&mddev->max_corr_read_errors,
 		   MD_DEFAULT_MAX_CORRECTED_READ_ERRORS);
 	mddev->safemode = 0;
@@ -5125,9 +4933,9 @@ int md_run(struct mddev *mddev)
 		if (rdev->raid_disk >= 0)
 			if (sysfs_link_rdev(mddev, rdev))
 				/* failure here is OK */;
-	
+
 	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
-	
+
 	if (mddev->flags & MD_UPDATE_SB_FLAGS)
 		md_update_sb(mddev, 0);
 
@@ -5307,12 +5115,13 @@ static int md_set_readonly(struct mddev *mddev, struct block_device *bdev)
 	mddev_lock_nointr(mddev);
 
 	mutex_lock(&mddev->open_mutex);
-	if (atomic_read(&mddev->openers) > !!bdev ||
+	if ((mddev->pers && atomic_read(&mddev->openers) > !!bdev) ||
 	    mddev->sync_thread ||
 	    (bdev && !test_bit(MD_STILL_CLOSED, &mddev->flags))) {
 		printk("md: %s still in use.\n",mdname(mddev));
 		if (did_freeze) {
 			clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
+			set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
 			md_wakeup_thread(mddev->thread);
 		}
 		err = -EBUSY;
@@ -5327,6 +5136,8 @@ static int md_set_readonly(struct mddev *mddev, struct block_device *bdev)
 		mddev->ro = 1;
 		set_disk_ro(mddev->gendisk, 1);
 		clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
+		set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
+		md_wakeup_thread(mddev->thread);
 		sysfs_notify_dirent_safe(mddev->sysfs_state);
 		err = 0;
 	}
@@ -5339,7 +5150,7 @@ out:
  *   0 - completely stop and dis-assemble array
  *   2 - stop but do not disassemble array
  */
-static int do_md_stop(struct mddev * mddev, int mode,
+static int do_md_stop(struct mddev *mddev, int mode,
 		      struct block_device *bdev)
 {
 	struct gendisk *disk = mddev->gendisk;
@@ -5362,7 +5173,7 @@ static int do_md_stop(struct mddev * mddev, int mode,
 	mddev_lock_nointr(mddev);
 
 	mutex_lock(&mddev->open_mutex);
-	if (atomic_read(&mddev->openers) > !!bdev ||
+	if ((mddev->pers && atomic_read(&mddev->openers) > !!bdev) ||
 	    mddev->sysfs_active ||
 	    mddev->sync_thread ||
 	    (bdev && !test_bit(MD_STILL_CLOSED, &mddev->flags))) {
@@ -5370,6 +5181,7 @@ static int do_md_stop(struct mddev * mddev, int mode,
 		mutex_unlock(&mddev->open_mutex);
 		if (did_freeze) {
 			clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
+			set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
 			md_wakeup_thread(mddev->thread);
 		}
 		return -EBUSY;
@@ -5512,12 +5324,12 @@ static void autorun_devices(int part)
 				"md: cannot allocate memory for md drive.\n");
 			break;
 		}
-		if (mddev_lock(mddev)) 
+		if (mddev_lock(mddev))
 			printk(KERN_WARNING "md: %s locked, cannot run\n",
 			       mdname(mddev));
 		else if (mddev->raid_disks || mddev->major_version
 			 || !list_empty(&mddev->disks)) {
-			printk(KERN_WARNING 
+			printk(KERN_WARNING
 				"md: %s already running, cannot run %s\n",
 				mdname(mddev), bdevname(rdev0->bdev,b));
 			mddev_unlock(mddev);
@@ -5545,7 +5357,7 @@ static void autorun_devices(int part)
 }
 #endif /* !MODULE */
 
-static int get_version(void __user * arg)
+static int get_version(void __user *arg)
 {
 	mdu_version_t ver;
 
@@ -5559,7 +5371,7 @@ static int get_version(void __user * arg)
 	return 0;
 }
 
-static int get_array_info(struct mddev * mddev, void __user * arg)
+static int get_array_info(struct mddev *mddev, void __user *arg)
 {
 	mdu_array_info_t info;
 	int nr,working,insync,failed,spare;
@@ -5574,7 +5386,7 @@ static int get_array_info(struct mddev * mddev, void __user * arg)
 		else {
 			working++;
 			if (test_bit(In_sync, &rdev->flags))
-				insync++;	
+				insync++;
 			else
 				spare++;
 		}
@@ -5614,7 +5426,7 @@ static int get_array_info(struct mddev * mddev, void __user * arg)
 	return 0;
 }
 
-static int get_bitmap_file(struct mddev * mddev, void __user * arg)
+static int get_bitmap_file(struct mddev *mddev, void __user * arg)
 {
 	mdu_bitmap_file_t *file = NULL; /* too big for stack allocation */
 	char *ptr, *buf = NULL;
@@ -5652,7 +5464,7 @@ out:
 	return err;
 }
 
-static int get_disk_info(struct mddev * mddev, void __user * arg)
+static int get_disk_info(struct mddev *mddev, void __user * arg)
 {
 	mdu_disk_info_t info;
 	struct md_rdev *rdev;
@@ -5688,7 +5500,7 @@ static int get_disk_info(struct mddev * mddev, void __user * arg)
 	return 0;
 }
 
-static int add_new_disk(struct mddev * mddev, mdu_disk_info_t *info)
+static int add_new_disk(struct mddev *mddev, mdu_disk_info_t *info)
 {
 	char b[BDEVNAME_SIZE], b2[BDEVNAME_SIZE];
 	struct md_rdev *rdev;
@@ -5702,7 +5514,7 @@ static int add_new_disk(struct mddev * mddev, mdu_disk_info_t *info)
 		/* expecting a device which has a superblock */
 		rdev = md_import_device(dev, mddev->major_version, mddev->minor_version);
 		if (IS_ERR(rdev)) {
-			printk(KERN_WARNING 
+			printk(KERN_WARNING
 				"md: md_import_device returned %ld\n",
 				PTR_ERR(rdev));
 			return PTR_ERR(rdev);
@@ -5714,9 +5526,9 @@ static int add_new_disk(struct mddev * mddev, mdu_disk_info_t *info)
 			err = super_types[mddev->major_version]
 				.load_super(rdev, rdev0, mddev->minor_version);
 			if (err < 0) {
-				printk(KERN_WARNING 
+				printk(KERN_WARNING
 					"md: %s has different UUID to %s\n",
-					bdevname(rdev->bdev,b), 
+					bdevname(rdev->bdev,b),
 					bdevname(rdev0->bdev,b2));
 				export_rdev(rdev);
 				return -EINVAL;
@@ -5736,7 +5548,7 @@ static int add_new_disk(struct mddev * mddev, mdu_disk_info_t *info)
 	if (mddev->pers) {
 		int err;
 		if (!mddev->pers->hot_add_disk) {
-			printk(KERN_WARNING 
+			printk(KERN_WARNING
 				"%s: personality does not support diskops!\n",
 			       mdname(mddev));
 			return -EINVAL;
@@ -5747,7 +5559,7 @@ static int add_new_disk(struct mddev * mddev, mdu_disk_info_t *info)
 		else
 			rdev = md_import_device(dev, -1, -1);
 		if (IS_ERR(rdev)) {
-			printk(KERN_WARNING 
+			printk(KERN_WARNING
 				"md: md_import_device returned %ld\n",
 				PTR_ERR(rdev));
 			return PTR_ERR(rdev);
@@ -5821,7 +5633,7 @@ static int add_new_disk(struct mddev * mddev, mdu_disk_info_t *info)
 		int err;
 		rdev = md_import_device(dev, -1, 0);
 		if (IS_ERR(rdev)) {
-			printk(KERN_WARNING 
+			printk(KERN_WARNING
 				"md: error, md_import_device() returned %ld\n",
 				PTR_ERR(rdev));
 			return PTR_ERR(rdev);
@@ -5856,7 +5668,7 @@ static int add_new_disk(struct mddev * mddev, mdu_disk_info_t *info)
 	return 0;
 }
 
-static int hot_remove_disk(struct mddev * mddev, dev_t dev)
+static int hot_remove_disk(struct mddev *mddev, dev_t dev)
 {
 	char b[BDEVNAME_SIZE];
 	struct md_rdev *rdev;
@@ -5882,7 +5694,7 @@ busy:
 	return -EBUSY;
 }
 
-static int hot_add_disk(struct mddev * mddev, dev_t dev)
+static int hot_add_disk(struct mddev *mddev, dev_t dev)
 {
 	char b[BDEVNAME_SIZE];
 	int err;
@@ -5898,7 +5710,7 @@ static int hot_add_disk(struct mddev * mddev, dev_t dev)
 		return -EINVAL;
 	}
 	if (!mddev->pers->hot_add_disk) {
-		printk(KERN_WARNING 
+		printk(KERN_WARNING
 			"%s: personality does not support diskops!\n",
 			mdname(mddev));
 		return -EINVAL;
@@ -5906,7 +5718,7 @@ static int hot_add_disk(struct mddev * mddev, dev_t dev)
 
 	rdev = md_import_device(dev, -1, 0);
 	if (IS_ERR(rdev)) {
-		printk(KERN_WARNING 
+		printk(KERN_WARNING
 			"md: error, md_import_device() returned %ld\n",
 			PTR_ERR(rdev));
 		return -EINVAL;
@@ -5920,7 +5732,7 @@ static int hot_add_disk(struct mddev * mddev, dev_t dev)
 	rdev->sectors = rdev->sb_start;
 
 	if (test_bit(Faulty, &rdev->flags)) {
-		printk(KERN_WARNING 
+		printk(KERN_WARNING
 			"md: can not hot-add faulty %s disk to %s!\n",
 			bdevname(rdev->bdev,b), mdname(mddev));
 		err = -EINVAL;
@@ -5961,14 +5773,13 @@ static int set_bitmap_file(struct mddev *mddev, int fd)
 	int err = 0;
 
 	if (mddev->pers) {
-		if (!mddev->pers->quiesce)
+		if (!mddev->pers->quiesce || !mddev->thread)
 			return -EBUSY;
 		if (mddev->recovery || mddev->sync_thread)
 			return -EBUSY;
 		/* we should be able to change the bitmap.. */
 	}
 
-
 	if (fd >= 0) {
 		struct inode *inode;
 		if (mddev->bitmap)
@@ -6039,7 +5850,7 @@ static int set_bitmap_file(struct mddev *mddev, int fd)
  *  The minor and patch _version numbers are also kept incase the
  *  super_block handler wishes to interpret them.
  */
-static int set_array_info(struct mddev * mddev, mdu_array_info_t *info)
+static int set_array_info(struct mddev *mddev, mdu_array_info_t *info)
 {
 
 	if (info->raid_disks == 0) {
@@ -6048,7 +5859,7 @@ static int set_array_info(struct mddev * mddev, mdu_array_info_t *info)
 		    info->major_version >= ARRAY_SIZE(super_types) ||
 		    super_types[info->major_version].name == NULL) {
 			/* maybe try to auto-load a module? */
-			printk(KERN_INFO 
+			printk(KERN_INFO
 				"md: superblock version %d not known\n",
 				info->major_version);
 			return -EINVAL;
@@ -6196,7 +6007,6 @@ static int update_raid_disks(struct mddev *mddev, int raid_disks)
 	return rv;
 }
 
-
 /*
  * update_array_info is used to change the configuration of an
  * on-line array.
@@ -6263,7 +6073,7 @@ static int update_array_info(struct mddev *mddev, mdu_array_info_t *info)
 		rv = update_raid_disks(mddev, info->raid_disks);
 
 	if ((state ^ info->state) & (1<<MD_SB_BITMAP_PRESENT)) {
-		if (mddev->pers->quiesce == NULL)
+		if (mddev->pers->quiesce == NULL || mddev->thread == NULL)
 			return -EINVAL;
 		if (mddev->recovery || mddev->sync_thread)
 			return -EBUSY;
@@ -6347,7 +6157,6 @@ static inline bool md_ioctl_valid(unsigned int cmd)
 	case GET_DISK_INFO:
 	case HOT_ADD_DISK:
 	case HOT_REMOVE_DISK:
-	case PRINT_RAID_DEBUG:
 	case RAID_AUTORUN:
 	case RAID_VERSION:
 	case RESTART_ARRAY_RW:
@@ -6391,18 +6200,13 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 	switch (cmd) {
 	case RAID_VERSION:
 		err = get_version(argp);
-		goto done;
-
-	case PRINT_RAID_DEBUG:
-		err = 0;
-		md_print_devices();
-		goto done;
+		goto out;
 
 #ifndef MODULE
 	case RAID_AUTORUN:
 		err = 0;
 		autostart_arrays(arg);
-		goto done;
+		goto out;
 #endif
 	default:;
 	}
@@ -6415,7 +6219,7 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 
 	if (!mddev) {
 		BUG();
-		goto abort;
+		goto out;
 	}
 
 	/* Some actions do not requires the mutex */
@@ -6425,18 +6229,18 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 			err = -ENODEV;
 		else
 			err = get_array_info(mddev, argp);
-		goto abort;
+		goto out;
 
 	case GET_DISK_INFO:
 		if (!mddev->raid_disks && !mddev->external)
 			err = -ENODEV;
 		else
 			err = get_disk_info(mddev, argp);
-		goto abort;
+		goto out;
 
 	case SET_DISK_FAULTY:
 		err = set_disk_faulty(mddev, new_decode_dev(arg));
-		goto abort;
+		goto out;
 	}
 
 	if (cmd == ADD_NEW_DISK)
@@ -6454,10 +6258,10 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 		 * and writes
 		 */
 		mutex_lock(&mddev->open_mutex);
-		if (atomic_read(&mddev->openers) > 1) {
+		if (mddev->pers && atomic_read(&mddev->openers) > 1) {
 			mutex_unlock(&mddev->open_mutex);
 			err = -EBUSY;
-			goto abort;
+			goto out;
 		}
 		set_bit(MD_STILL_CLOSED, &mddev->flags);
 		mutex_unlock(&mddev->open_mutex);
@@ -6465,10 +6269,10 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 	}
 	err = mddev_lock(mddev);
 	if (err) {
-		printk(KERN_INFO 
+		printk(KERN_INFO
 			"md: ioctl lock interrupted, reason %d, cmd %d\n",
 			err, cmd);
-		goto abort;
+		goto out;
 	}
 
 	if (cmd == SET_ARRAY_INFO) {
@@ -6477,38 +6281,38 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 			memset(&info, 0, sizeof(info));
 		else if (copy_from_user(&info, argp, sizeof(info))) {
 			err = -EFAULT;
-			goto abort_unlock;
+			goto unlock;
 		}
 		if (mddev->pers) {
 			err = update_array_info(mddev, &info);
 			if (err) {
 				printk(KERN_WARNING "md: couldn't update"
 				       " array info. %d\n", err);
-				goto abort_unlock;
+				goto unlock;
 			}
-			goto done_unlock;
+			goto unlock;
 		}
 		if (!list_empty(&mddev->disks)) {
 			printk(KERN_WARNING
 			       "md: array %s already has disks!\n",
 			       mdname(mddev));
 			err = -EBUSY;
-			goto abort_unlock;
+			goto unlock;
 		}
 		if (mddev->raid_disks) {
 			printk(KERN_WARNING
 			       "md: array %s already initialised!\n",
 			       mdname(mddev));
 			err = -EBUSY;
-			goto abort_unlock;
+			goto unlock;
 		}
 		err = set_array_info(mddev, &info);
 		if (err) {
 			printk(KERN_WARNING "md: couldn't set"
 			       " array info. %d\n", err);
-			goto abort_unlock;
+			goto unlock;
 		}
-		goto done_unlock;
+		goto unlock;
 	}
 
 	/*
@@ -6521,7 +6325,7 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 	    && cmd != RUN_ARRAY && cmd != SET_BITMAP_FILE
 	    && cmd != GET_BITMAP_FILE) {
 		err = -ENODEV;
-		goto abort_unlock;
+		goto unlock;
 	}
 
 	/*
@@ -6530,23 +6334,23 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 	switch (cmd) {
 	case GET_BITMAP_FILE:
 		err = get_bitmap_file(mddev, argp);
-		goto done_unlock;
+		goto unlock;
 
 	case RESTART_ARRAY_RW:
 		err = restart_array(mddev);
-		goto done_unlock;
+		goto unlock;
 
 	case STOP_ARRAY:
 		err = do_md_stop(mddev, 0, bdev);
-		goto done_unlock;
+		goto unlock;
 
 	case STOP_ARRAY_RO:
 		err = md_set_readonly(mddev, bdev);
-		goto done_unlock;
+		goto unlock;
 
 	case HOT_REMOVE_DISK:
 		err = hot_remove_disk(mddev, new_decode_dev(arg));
-		goto done_unlock;
+		goto unlock;
 
 	case ADD_NEW_DISK:
 		/* We can support ADD_NEW_DISK on read-only arrays
@@ -6562,14 +6366,14 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 				break;
 			else
 				err = add_new_disk(mddev, &info);
-			goto done_unlock;
+			goto unlock;
 		}
 		break;
 
 	case BLKROSET:
 		if (get_user(ro, (int __user *)(arg))) {
 			err = -EFAULT;
-			goto done_unlock;
+			goto unlock;
 		}
 		err = -EINVAL;
 
@@ -6577,11 +6381,11 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 		 * does not matter, no writes are coming
 		 */
 		if (ro)
-			goto done_unlock;
+			goto unlock;
 
 		/* are we are already prepared for writes? */
 		if (mddev->ro != 1)
-			goto done_unlock;
+			goto unlock;
 
 		/* transitioning to readauto need only happen for
 		 * arrays that call md_write_start
@@ -6593,17 +6397,14 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 				set_disk_ro(mddev->gendisk, 0);
 			}
 		}
-		goto done_unlock;
+		goto unlock;
 	}
 
 	/*
 	 * The remaining ioctls are changing the state of the
 	 * superblock, so we do not allow them on read-only arrays.
-	 * However non-MD ioctls (e.g. get-size) will still come through
-	 * here and hit the 'default' below, so only disallow
-	 * 'md' ioctls, and switch to rw mode if started auto-readonly.
 	 */
-	if (_IOC_TYPE(cmd) == MD_MAJOR && mddev->ro && mddev->pers) {
+	if (mddev->ro && mddev->pers) {
 		if (mddev->ro == 2) {
 			mddev->ro = 0;
 			sysfs_notify_dirent_safe(mddev->sysfs_state);
@@ -6621,7 +6422,7 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 			}
 		} else {
 			err = -EROFS;
-			goto abort_unlock;
+			goto unlock;
 		}
 	}
 
@@ -6633,38 +6434,32 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 			err = -EFAULT;
 		else
 			err = add_new_disk(mddev, &info);
-		goto done_unlock;
+		goto unlock;
 	}
 
 	case HOT_ADD_DISK:
 		err = hot_add_disk(mddev, new_decode_dev(arg));
-		goto done_unlock;
+		goto unlock;
 
 	case RUN_ARRAY:
 		err = do_md_run(mddev);
-		goto done_unlock;
+		goto unlock;
 
 	case SET_BITMAP_FILE:
 		err = set_bitmap_file(mddev, (int)arg);
-		goto done_unlock;
+		goto unlock;
 
 	default:
 		err = -EINVAL;
-		goto abort_unlock;
+		goto unlock;
 	}
 
-done_unlock:
-abort_unlock:
+unlock:
 	if (mddev->hold_active == UNTIL_IOCTL &&
 	    err != -EINVAL)
 		mddev->hold_active = 0;
 	mddev_unlock(mddev);
-
-	return err;
-done:
-	if (err)
-		MD_BUG();
-abort:
+out:
 	return err;
 }
 #ifdef CONFIG_COMPAT
@@ -6726,7 +6521,7 @@ static int md_open(struct block_device *bdev, fmode_t mode)
 
 static void md_release(struct gendisk *disk, fmode_t mode)
 {
- 	struct mddev *mddev = disk->private_data;
+	struct mddev *mddev = disk->private_data;
 
 	BUG_ON(!mddev);
 	atomic_dec(&mddev->openers);
@@ -6761,7 +6556,7 @@ static const struct block_device_operations md_fops =
 	.revalidate_disk= md_revalidate,
 };
 
-static int md_thread(void * arg)
+static int md_thread(void *arg)
 {
 	struct md_thread *thread = arg;
 
@@ -6810,6 +6605,7 @@ void md_wakeup_thread(struct md_thread *thread)
 		wake_up(&thread->wqueue);
 	}
 }
+EXPORT_SYMBOL(md_wakeup_thread);
 
 struct md_thread *md_register_thread(void (*run) (struct md_thread *),
 		struct mddev *mddev, const char *name)
@@ -6835,6 +6631,7 @@ struct md_thread *md_register_thread(void (*run) (struct md_thread *),
 	}
 	return thread;
 }
+EXPORT_SYMBOL(md_register_thread);
 
 void md_unregister_thread(struct md_thread **threadp)
 {
@@ -6852,14 +6649,10 @@ void md_unregister_thread(struct md_thread **threadp)
 	kthread_stop(thread->tsk);
 	kfree(thread);
 }
+EXPORT_SYMBOL(md_unregister_thread);
 
 void md_error(struct mddev *mddev, struct md_rdev *rdev)
 {
-	if (!mddev) {
-		MD_BUG();
-		return;
-	}
-
 	if (!rdev || test_bit(Faulty, &rdev->flags))
 		return;
 
@@ -6876,6 +6669,7 @@ void md_error(struct mddev *mddev, struct md_rdev *rdev)
 		queue_work(md_misc_wq, &mddev->event_work);
 	md_new_event_inintr(mddev);
 }
+EXPORT_SYMBOL(md_error);
 
 /* seq_file implementation /proc/mdstat */
 
@@ -6898,8 +6692,7 @@ static void status_unused(struct seq_file *seq)
 	seq_printf(seq, "\n");
 }
 
-
-static void status_resync(struct seq_file *seq, struct mddev * mddev)
+static void status_resync(struct seq_file *seq, struct mddev *mddev)
 {
 	sector_t max_sectors, resync, res;
 	unsigned long dt, db;
@@ -6919,13 +6712,7 @@ static void status_resync(struct seq_file *seq, struct mddev * mddev)
 	else
 		max_sectors = mddev->dev_sectors;
 
-	/*
-	 * Should not happen.
-	 */
-	if (!max_sectors) {
-		MD_BUG();
-		return;
-	}
+	WARN_ON(max_sectors == 0);
 	/* Pick 'scale' such that (resync>>scale)*1000 will fit
 	 * in a sector_t, and (max_sectors>>scale) will fit in a
 	 * u32, as those are the requirements for sector_div.
@@ -7021,7 +6808,7 @@ static void *md_seq_next(struct seq_file *seq, void *v, loff_t *pos)
 {
 	struct list_head *tmp;
 	struct mddev *next_mddev, *mddev = v;
-	
+
 	++*pos;
 	if (v == (void*)2)
 		return NULL;
@@ -7036,7 +6823,7 @@ static void *md_seq_next(struct seq_file *seq, void *v, loff_t *pos)
 	else {
 		next_mddev = (void*)2;
 		*pos = 0x10000;
-	}		
+	}
 	spin_unlock(&all_mddevs_lock);
 
 	if (v != (void*)1)
@@ -7132,7 +6919,7 @@ static int md_seq_show(struct seq_file *seq, void *v)
 
 		if (mddev->pers) {
 			mddev->pers->status(seq, mddev);
-	 		seq_printf(seq, "\n      ");
+			seq_printf(seq, "\n      ");
 			if (mddev->pers->sync_request) {
 				if (mddev->curr_resync > 2) {
 					status_resync(seq, mddev);
@@ -7150,7 +6937,7 @@ static int md_seq_show(struct seq_file *seq, void *v)
 		seq_printf(seq, "\n");
 	}
 	mddev_unlock(mddev);
-	
+
 	return 0;
 }
 
@@ -7204,12 +6991,14 @@ static const struct file_operations md_seq_fops = {
 
 int register_md_personality(struct md_personality *p)
 {
+	printk(KERN_INFO "md: %s personality registered for level %d\n",
+						p->name, p->level);
 	spin_lock(&pers_lock);
 	list_add_tail(&p->list, &pers_list);
-	printk(KERN_INFO "md: %s personality registered for level %d\n", p->name, p->level);
 	spin_unlock(&pers_lock);
 	return 0;
 }
+EXPORT_SYMBOL(register_md_personality);
 
 int unregister_md_personality(struct md_personality *p)
 {
@@ -7219,10 +7008,11 @@ int unregister_md_personality(struct md_personality *p)
 	spin_unlock(&pers_lock);
 	return 0;
 }
+EXPORT_SYMBOL(unregister_md_personality);
 
 static int is_mddev_idle(struct mddev *mddev, int init)
 {
-	struct md_rdev * rdev;
+	struct md_rdev *rdev;
 	int idle;
 	int curr_events;
 
@@ -7276,7 +7066,7 @@ void md_done_sync(struct mddev *mddev, int blocks, int ok)
 		// stop recovery, signal do_sync ....
 	}
 }
-
+EXPORT_SYMBOL(md_done_sync);
 
 /* md_write_start(mddev, bi)
  * If we need to update some array metadata (e.g. 'active' flag
@@ -7317,6 +7107,7 @@ void md_write_start(struct mddev *mddev, struct bio *bi)
 	wait_event(mddev->sb_wait,
 		   !test_bit(MD_CHANGE_PENDING, &mddev->flags));
 }
+EXPORT_SYMBOL(md_write_start);
 
 void md_write_end(struct mddev *mddev)
 {
@@ -7327,6 +7118,7 @@ void md_write_end(struct mddev *mddev)
 			mod_timer(&mddev->safemode_timer, jiffies + mddev->safemode_delay);
 	}
 }
+EXPORT_SYMBOL(md_write_end);
 
 /* md_allow_write(mddev)
  * Calling this ensures that the array is marked 'active' so that writes
@@ -7376,7 +7168,7 @@ void md_do_sync(struct md_thread *thread)
 	struct mddev *mddev2;
 	unsigned int currspeed = 0,
 		 window;
-	sector_t max_sectors,j, io_sectors;
+	sector_t max_sectors,j, io_sectors, recovery_done;
 	unsigned long mark[SYNC_MARKS];
 	unsigned long update_time;
 	sector_t mark_cnt[SYNC_MARKS];
@@ -7652,7 +7444,8 @@ void md_do_sync(struct md_thread *thread)
 		 */
 		cond_resched();
 
-		currspeed = ((unsigned long)(io_sectors-mddev->resync_mark_cnt))/2
+		recovery_done = io_sectors - atomic_read(&mddev->recovery_active);
+		currspeed = ((unsigned long)(recovery_done - mddev->resync_mark_cnt))/2
 			/((jiffies-mddev->resync_mark)/HZ +1) +1;
 
 		if (currspeed > speed_min(mddev)) {
@@ -7783,6 +7576,33 @@ no_add:
 	return spares;
 }
 
+static void md_start_sync(struct work_struct *ws)
+{
+	struct mddev *mddev = container_of(ws, struct mddev, del_work);
+
+	mddev->sync_thread = md_register_thread(md_do_sync,
+						mddev,
+						"resync");
+	if (!mddev->sync_thread) {
+		printk(KERN_ERR "%s: could not start resync"
+		       " thread...\n",
+		       mdname(mddev));
+		/* leave the spares where they are, it shouldn't hurt */
+		clear_bit(MD_RECOVERY_SYNC, &mddev->recovery);
+		clear_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
+		clear_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
+		clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
+		clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
+		if (test_and_clear_bit(MD_RECOVERY_RECOVER,
+				       &mddev->recovery))
+			if (mddev->sysfs_action)
+				sysfs_notify_dirent_safe(mddev->sysfs_action);
+	} else
+		md_wakeup_thread(mddev->sync_thread);
+	sysfs_notify_dirent_safe(mddev->sysfs_action);
+	md_new_event(mddev);
+}
+
 /*
  * This routine is regularly called by all per-raid-array threads to
  * deal with generic issues like resync and super-block update.
@@ -7899,7 +7719,7 @@ void md_check_recovery(struct mddev *mddev)
 
 		if (!test_and_clear_bit(MD_RECOVERY_NEEDED, &mddev->recovery) ||
 		    test_bit(MD_RECOVERY_FROZEN, &mddev->recovery))
-			goto unlock;
+			goto not_running;
 		/* no recovery is running.
 		 * remove any failed drives, then
 		 * add spares if possible.
@@ -7911,7 +7731,7 @@ void md_check_recovery(struct mddev *mddev)
 			if (mddev->pers->check_reshape == NULL ||
 			    mddev->pers->check_reshape(mddev) != 0)
 				/* Cannot proceed */
-				goto unlock;
+				goto not_running;
 			set_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
 			clear_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
 		} else if ((spares = remove_and_add_spares(mddev, NULL))) {
@@ -7924,7 +7744,7 @@ void md_check_recovery(struct mddev *mddev)
 			clear_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
 		} else if (!test_bit(MD_RECOVERY_SYNC, &mddev->recovery))
 			/* nothing to be done ... */
-			goto unlock;
+			goto not_running;
 
 		if (mddev->pers->sync_request) {
 			if (spares) {
@@ -7934,27 +7754,11 @@ void md_check_recovery(struct mddev *mddev)
 				 */
 				bitmap_write_all(mddev->bitmap);
 			}
-			mddev->sync_thread = md_register_thread(md_do_sync,
-								mddev,
-								"resync");
-			if (!mddev->sync_thread) {
-				printk(KERN_ERR "%s: could not start resync"
-					" thread...\n", 
-					mdname(mddev));
-				/* leave the spares where they are, it shouldn't hurt */
-				clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
-				clear_bit(MD_RECOVERY_SYNC, &mddev->recovery);
-				clear_bit(MD_RECOVERY_RESHAPE, &mddev->recovery);
-				clear_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
-				clear_bit(MD_RECOVERY_CHECK, &mddev->recovery);
-			} else
-				md_wakeup_thread(mddev->sync_thread);
-			sysfs_notify_dirent_safe(mddev->sysfs_action);
-			md_new_event(mddev);
+			INIT_WORK(&mddev->del_work, md_start_sync);
+			queue_work(md_misc_wq, &mddev->del_work);
+			goto unlock;
 		}
-	unlock:
-		wake_up(&mddev->sb_wait);
-
+	not_running:
 		if (!mddev->sync_thread) {
 			clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
 			if (test_and_clear_bit(MD_RECOVERY_RECOVER,
@@ -7962,9 +7766,12 @@ void md_check_recovery(struct mddev *mddev)
 				if (mddev->sysfs_action)
 					sysfs_notify_dirent_safe(mddev->sysfs_action);
 		}
+	unlock:
+		wake_up(&mddev->sb_wait);
 		mddev_unlock(mddev);
 	}
 }
+EXPORT_SYMBOL(md_check_recovery);
 
 void md_reap_sync_thread(struct mddev *mddev)
 {
@@ -8007,6 +7814,7 @@ void md_reap_sync_thread(struct mddev *mddev)
 	if (mddev->event_work.func)
 		queue_work(md_misc_wq, &mddev->event_work);
 }
+EXPORT_SYMBOL(md_reap_sync_thread);
 
 void md_wait_for_blocked_rdev(struct md_rdev *rdev, struct mddev *mddev)
 {
@@ -8592,7 +8400,7 @@ static int __init md_init(void)
 		goto err_mdp;
 	mdp_major = ret;
 
-	blk_register_region(MKDEV(MD_MAJOR, 0), 1UL<<MINORBITS, THIS_MODULE,
+	blk_register_region(MKDEV(MD_MAJOR, 0), 512, THIS_MODULE,
 			    md_probe, NULL, NULL);
 	blk_register_region(MKDEV(mdp_major, 0), 1UL<<MINORBITS, THIS_MODULE,
 			    md_probe, NULL, NULL);
@@ -8640,7 +8448,6 @@ void md_autodetect_dev(dev_t dev)
 	}
 }
 
-
 static void autostart_arrays(int part)
 {
 	struct md_rdev *rdev;
@@ -8664,10 +8471,9 @@ static void autostart_arrays(int part)
 		if (IS_ERR(rdev))
 			continue;
 
-		if (test_bit(Faulty, &rdev->flags)) {
-			MD_BUG();
+		if (test_bit(Faulty, &rdev->flags))
 			continue;
-		}
+
 		set_bit(AutoDetected, &rdev->flags);
 		list_add(&rdev->same_set, &pending_raid_disks);
 		i_passed++;
@@ -8687,7 +8493,7 @@ static __exit void md_exit(void)
 	struct list_head *tmp;
 	int delay = 1;
 
-	blk_unregister_region(MKDEV(MD_MAJOR,0), 1U << MINORBITS);
+	blk_unregister_region(MKDEV(MD_MAJOR,0), 512);
 	blk_unregister_region(MKDEV(mdp_major,0), 1U << MINORBITS);
 
 	unregister_blkdev(MD_MAJOR,"md");
@@ -8735,20 +8541,8 @@ static int set_ro(const char *val, struct kernel_param *kp)
 
 module_param_call(start_ro, set_ro, get_ro, NULL, S_IRUSR|S_IWUSR);
 module_param(start_dirty_degraded, int, S_IRUGO|S_IWUSR);
-
 module_param_call(new_array, add_named_array, NULL, NULL, S_IWUSR);
 
-EXPORT_SYMBOL(register_md_personality);
-EXPORT_SYMBOL(unregister_md_personality);
-EXPORT_SYMBOL(md_error);
-EXPORT_SYMBOL(md_done_sync);
-EXPORT_SYMBOL(md_write_start);
-EXPORT_SYMBOL(md_write_end);
-EXPORT_SYMBOL(md_register_thread);
-EXPORT_SYMBOL(md_unregister_thread);
-EXPORT_SYMBOL(md_wakeup_thread);
-EXPORT_SYMBOL(md_check_recovery);
-EXPORT_SYMBOL(md_reap_sync_thread);
 MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("MD RAID framework");
 MODULE_ALIAS("md");
diff --git a/drivers/md/md.h b/drivers/md/md.h
index a49d991f3fe1..03cec5bdcaae 100644
--- a/drivers/md/md.h
+++ b/drivers/md/md.h
@@ -1,15 +1,15 @@
 /*
    md.h : kernel internal structure of the Linux MD driver
           Copyright (C) 1996-98 Ingo Molnar, Gadi Oxman
-	  
+
    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation; either version 2, or (at your option)
    any later version.
-   
+
    You should have received a copy of the GNU General Public License
    (for example /usr/src/linux/COPYING); if not, write to the Free
-   Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.  
+   Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 */
 
 #ifndef _MD_MD_H
@@ -56,7 +56,7 @@ struct md_rdev {
 	__u64		sb_events;
 	sector_t	data_offset;	/* start of data in array */
 	sector_t	new_data_offset;/* only relevant while reshaping */
-	sector_t 	sb_start;	/* offset of the super block (in 512byte sectors) */
+	sector_t	sb_start;	/* offset of the super block (in 512byte sectors) */
 	int		sb_size;	/* bytes in the superblock */
 	int		preferred_minor;	/* autorun support */
 
@@ -239,7 +239,7 @@ struct mddev {
 					minor_version,
 					patch_version;
 	int				persistent;
-	int 				external;	/* metadata is
+	int				external;	/* metadata is
 							 * managed externally */
 	char				metadata_type[17]; /* externally set*/
 	int				chunk_sectors;
@@ -248,7 +248,7 @@ struct mddev {
 	char				clevel[16];
 	int				raid_disks;
 	int				max_disks;
-	sector_t			dev_sectors; 	/* used size of
+	sector_t			dev_sectors;	/* used size of
 							 * component devices */
 	sector_t			array_sectors; /* exported array size */
 	int				external_size; /* size managed
@@ -312,7 +312,7 @@ struct mddev {
 	int				parallel_resync;
 
 	int				ok_start_degraded;
-	/* recovery/resync flags 
+	/* recovery/resync flags
 	 * NEEDED:   we might need to start a resync/recover
 	 * RUNNING:  a thread is running, or about to be started
 	 * SYNC:     actually doing a resync, not a recovery
@@ -392,20 +392,20 @@ struct mddev {
 
 	unsigned int			safemode;	/* if set, update "clean" superblock
 							 * when no writes pending.
-							 */ 
+							 */
 	unsigned int			safemode_delay;
 	struct timer_list		safemode_timer;
-	atomic_t			writes_pending; 
+	atomic_t			writes_pending;
 	struct request_queue		*queue;	/* for plugging ... */
 
-	struct bitmap                   *bitmap; /* the bitmap for the device */
+	struct bitmap			*bitmap; /* the bitmap for the device */
 	struct {
 		struct file		*file; /* the bitmap file */
 		loff_t			offset; /* offset from superblock of
 						 * start of bitmap. May be
 						 * negative, but not '0'
 						 * For external metadata, offset
-						 * from start of device. 
+						 * from start of device.
 						 */
 		unsigned long		space; /* space available at this offset */
 		loff_t			default_offset; /* this is the offset to use when
@@ -421,7 +421,7 @@ struct mddev {
 		int			external;
 	} bitmap_info;
 
-	atomic_t 			max_corr_read_errors; /* max read retries */
+	atomic_t			max_corr_read_errors; /* max read retries */
 	struct list_head		all_mddevs;
 
 	struct attribute_group		*to_remove;
@@ -439,7 +439,6 @@ struct mddev {
 	void (*sync_super)(struct mddev *mddev, struct md_rdev *rdev);
 };
 
-
 static inline void rdev_dec_pending(struct md_rdev *rdev, struct mddev *mddev)
 {
 	int faulty = test_bit(Faulty, &rdev->flags);
@@ -449,7 +448,7 @@ static inline void rdev_dec_pending(struct md_rdev *rdev, struct mddev *mddev)
 
 static inline void md_sync_acct(struct block_device *bdev, unsigned long nr_sectors)
 {
-        atomic_add(nr_sectors, &bdev->bd_contains->bd_disk->sync_io);
+	atomic_add(nr_sectors, &bdev->bd_contains->bd_disk->sync_io);
 }
 
 struct md_personality
@@ -463,7 +462,7 @@ struct md_personality
 	int (*stop)(struct mddev *mddev);
 	void (*status)(struct seq_file *seq, struct mddev *mddev);
 	/* error_handler must set ->faulty and clear ->in_sync
-	 * if appropriate, and should abort recovery if needed 
+	 * if appropriate, and should abort recovery if needed
 	 */
 	void (*error_handler)(struct mddev *mddev, struct md_rdev *rdev);
 	int (*hot_add_disk) (struct mddev *mddev, struct md_rdev *rdev);
@@ -493,7 +492,6 @@ struct md_personality
 	void *(*takeover) (struct mddev *mddev);
 };
 
-
 struct md_sysfs_entry {
 	struct attribute attr;
 	ssize_t (*show)(struct mddev *, char *);
@@ -560,7 +558,7 @@ struct md_thread {
 	void			(*run) (struct md_thread *thread);
 	struct mddev		*mddev;
 	wait_queue_head_t	wqueue;
-	unsigned long           flags;
+	unsigned long		flags;
 	struct task_struct	*tsk;
 	unsigned long		timeout;
 	void			*private;
@@ -594,7 +592,7 @@ extern void md_flush_request(struct mddev *mddev, struct bio *bio);
 extern void md_super_write(struct mddev *mddev, struct md_rdev *rdev,
 			   sector_t sector, int size, struct page *page);
 extern void md_super_wait(struct mddev *mddev);
-extern int sync_page_io(struct md_rdev *rdev, sector_t sector, int size, 
+extern int sync_page_io(struct md_rdev *rdev, sector_t sector, int size,
 			struct page *page, int rw, bool metadata_op);
 extern void md_do_sync(struct md_thread *thread);
 extern void md_new_event(struct mddev *mddev);
diff --git a/drivers/md/multipath.c b/drivers/md/multipath.c
index 849ad39f547b..399272f9c042 100644
--- a/drivers/md/multipath.c
+++ b/drivers/md/multipath.c
@@ -31,13 +31,12 @@
 
 #define	NR_RESERVED_BUFS	32
 
-
 static int multipath_map (struct mpconf *conf)
 {
 	int i, disks = conf->raid_disks;
 
 	/*
-	 * Later we do read balancing on the read side 
+	 * Later we do read balancing on the read side
 	 * now we use the first available disk.
 	 */
 
@@ -68,7 +67,6 @@ static void multipath_reschedule_retry (struct multipath_bh *mp_bh)
 	md_wakeup_thread(mddev->thread);
 }
 
-
 /*
  * multipath_end_bh_io() is called when we have finished servicing a multipathed
  * operation and are ready to return a success/failure code to the buffer
@@ -98,8 +96,8 @@ static void multipath_end_request(struct bio *bio, int error)
 		 */
 		char b[BDEVNAME_SIZE];
 		md_error (mp_bh->mddev, rdev);
-		printk(KERN_ERR "multipath: %s: rescheduling sector %llu\n", 
-		       bdevname(rdev->bdev,b), 
+		printk(KERN_ERR "multipath: %s: rescheduling sector %llu\n",
+		       bdevname(rdev->bdev,b),
 		       (unsigned long long)bio->bi_iter.bi_sector);
 		multipath_reschedule_retry(mp_bh);
 	} else
@@ -145,12 +143,12 @@ static void multipath_status (struct seq_file *seq, struct mddev *mddev)
 {
 	struct mpconf *conf = mddev->private;
 	int i;
-	
+
 	seq_printf (seq, " [%d/%d] [", conf->raid_disks,
 		    conf->raid_disks - mddev->degraded);
 	for (i = 0; i < conf->raid_disks; i++)
 		seq_printf (seq, "%s",
-			       conf->multipaths[i].rdev && 
+			       conf->multipaths[i].rdev &&
 			       test_bit(In_sync, &conf->multipaths[i].rdev->flags) ? "U" : "_");
 	seq_printf (seq, "]");
 }
@@ -195,7 +193,7 @@ static void multipath_error (struct mddev *mddev, struct md_rdev *rdev)
 		 * first check if this is a queued request for a device
 		 * which has just failed.
 		 */
-		printk(KERN_ALERT 
+		printk(KERN_ALERT
 		       "multipath: only one IO path left and IO error.\n");
 		/* leave it active... it's all we have */
 		return;
@@ -242,7 +240,6 @@ static void print_multipath_conf (struct mpconf *conf)
 	}
 }
 
-
 static int multipath_add_disk(struct mddev *mddev, struct md_rdev *rdev)
 {
 	struct mpconf *conf = mddev->private;
@@ -325,8 +322,6 @@ abort:
 	return err;
 }
 
-
-
 /*
  * This is a kernel thread which:
  *
@@ -356,7 +351,7 @@ static void multipathd(struct md_thread *thread)
 
 		bio = &mp_bh->bio;
 		bio->bi_iter.bi_sector = mp_bh->master_bio->bi_iter.bi_sector;
-		
+
 		if ((mp_bh->path = multipath_map (conf))<0) {
 			printk(KERN_ALERT "multipath: %s: unrecoverable IO read"
 				" error for block %llu\n",
@@ -414,7 +409,7 @@ static int multipath_run (struct mddev *mddev)
 	conf = kzalloc(sizeof(struct mpconf), GFP_KERNEL);
 	mddev->private = conf;
 	if (!conf) {
-		printk(KERN_ERR 
+		printk(KERN_ERR
 			"multipath: couldn't allocate memory for %s\n",
 			mdname(mddev));
 		goto out;
@@ -423,7 +418,7 @@ static int multipath_run (struct mddev *mddev)
 	conf->multipaths = kzalloc(sizeof(struct multipath_info)*mddev->raid_disks,
 				   GFP_KERNEL);
 	if (!conf->multipaths) {
-		printk(KERN_ERR 
+		printk(KERN_ERR
 			"multipath: couldn't allocate memory for %s\n",
 			mdname(mddev));
 		goto out_free_conf;
@@ -469,7 +464,7 @@ static int multipath_run (struct mddev *mddev)
 	conf->pool = mempool_create_kmalloc_pool(NR_RESERVED_BUFS,
 						 sizeof(struct multipath_bh));
 	if (conf->pool == NULL) {
-		printk(KERN_ERR 
+		printk(KERN_ERR
 			"multipath: couldn't allocate memory for %s\n",
 			mdname(mddev));
 		goto out_free_conf;
@@ -485,7 +480,7 @@ static int multipath_run (struct mddev *mddev)
 		}
 	}
 
-	printk(KERN_INFO 
+	printk(KERN_INFO
 		"multipath: array %s active with %d out of %d IO paths\n",
 		mdname(mddev), conf->raid_disks - mddev->degraded,
 	       mddev->raid_disks);
@@ -512,7 +507,6 @@ out:
 	return -EIO;
 }
 
-
 static int multipath_stop (struct mddev *mddev)
 {
 	struct mpconf *conf = mddev->private;
diff --git a/drivers/md/persistent-data/dm-btree-internal.h b/drivers/md/persistent-data/dm-btree-internal.h
index 37d367bb9aa8..bf2b80d5c470 100644
--- a/drivers/md/persistent-data/dm-btree-internal.h
+++ b/drivers/md/persistent-data/dm-btree-internal.h
@@ -42,6 +42,12 @@ struct btree_node {
 } __packed;
 
 
+/*
+ * Locks a block using the btree node validator.
+ */
+int bn_read_lock(struct dm_btree_info *info, dm_block_t b,
+		 struct dm_block **result);
+
 void inc_children(struct dm_transaction_manager *tm, struct btree_node *n,
 		  struct dm_btree_value_type *vt);
 
diff --git a/drivers/md/persistent-data/dm-btree-spine.c b/drivers/md/persistent-data/dm-btree-spine.c
index cf9fd676ae44..1b5e13ec7f96 100644
--- a/drivers/md/persistent-data/dm-btree-spine.c
+++ b/drivers/md/persistent-data/dm-btree-spine.c
@@ -92,7 +92,7 @@ struct dm_block_validator btree_node_validator = {
 
 /*----------------------------------------------------------------*/
 
-static int bn_read_lock(struct dm_btree_info *info, dm_block_t b,
+int bn_read_lock(struct dm_btree_info *info, dm_block_t b,
 		 struct dm_block **result)
 {
 	return dm_tm_read_lock(info->tm, b, &btree_node_validator, result);
diff --git a/drivers/md/persistent-data/dm-btree.c b/drivers/md/persistent-data/dm-btree.c
index 416060c25709..200ac12a1d40 100644
--- a/drivers/md/persistent-data/dm-btree.c
+++ b/drivers/md/persistent-data/dm-btree.c
@@ -847,22 +847,26 @@ EXPORT_SYMBOL_GPL(dm_btree_find_lowest_key);
  * FIXME: We shouldn't use a recursive algorithm when we have limited stack
  * space.  Also this only works for single level trees.
  */
-static int walk_node(struct ro_spine *s, dm_block_t block,
+static int walk_node(struct dm_btree_info *info, dm_block_t block,
 		     int (*fn)(void *context, uint64_t *keys, void *leaf),
 		     void *context)
 {
 	int r;
 	unsigned i, nr;
+	struct dm_block *node;
 	struct btree_node *n;
 	uint64_t keys;
 
-	r = ro_step(s, block);
-	n = ro_node(s);
+	r = bn_read_lock(info, block, &node);
+	if (r)
+		return r;
+
+	n = dm_block_data(node);
 
 	nr = le32_to_cpu(n->header.nr_entries);
 	for (i = 0; i < nr; i++) {
 		if (le32_to_cpu(n->header.flags) & INTERNAL_NODE) {
-			r = walk_node(s, value64(n, i), fn, context);
+			r = walk_node(info, value64(n, i), fn, context);
 			if (r)
 				goto out;
 		} else {
@@ -874,7 +878,7 @@ static int walk_node(struct ro_spine *s, dm_block_t block,
 	}
 
 out:
-	ro_pop(s);
+	dm_tm_unlock(info->tm, node);
 	return r;
 }
 
@@ -882,15 +886,7 @@ int dm_btree_walk(struct dm_btree_info *info, dm_block_t root,
 		  int (*fn)(void *context, uint64_t *keys, void *leaf),
 		  void *context)
 {
-	int r;
-	struct ro_spine spine;
-
 	BUG_ON(info->levels > 1);
-
-	init_ro_spine(&spine, info);
-	r = walk_node(&spine, root, fn, context);
-	exit_ro_spine(&spine);
-
-	return r;
+	return walk_node(info, root, fn, context);
 }
 EXPORT_SYMBOL_GPL(dm_btree_walk);
diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c
index 407a99e46f69..ba6b85de96d2 100644
--- a/drivers/md/raid0.c
+++ b/drivers/md/raid0.c
@@ -1,10 +1,9 @@
 /*
    raid0.c : Multiple Devices driver for Linux
-             Copyright (C) 1994-96 Marc ZYNGIER
+	     Copyright (C) 1994-96 Marc ZYNGIER
 	     <zyngier@ufr-info-p7.ibp.fr> or
 	     <maz@gloups.fdn.fr>
-             Copyright (C) 1999, 2000 Ingo Molnar, Red Hat
-
+	     Copyright (C) 1999, 2000 Ingo Molnar, Red Hat
 
    RAID-0 management functions.
 
@@ -12,10 +11,10 @@
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation; either version 2, or (at your option)
    any later version.
-   
+
    You should have received a copy of the GNU General Public License
    (for example /usr/src/linux/COPYING); if not, write to the Free
-   Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.  
+   Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 */
 
 #include <linux/blkdev.h>
@@ -685,6 +684,12 @@ static void *raid0_takeover(struct mddev *mddev)
 	 *  raid10 - assuming we have all necessary active disks
 	 *  raid1 - with (N -1) mirror drives faulty
 	 */
+
+	if (mddev->bitmap) {
+		printk(KERN_ERR "md/raid0: %s: cannot takeover array with bitmap\n",
+		       mdname(mddev));
+		return ERR_PTR(-EBUSY);
+	}
 	if (mddev->level == 4)
 		return raid0_takeover_raid45(mddev);
 
diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index 56e24c072b62..40b35be34f8d 100644
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -494,7 +494,6 @@ static void raid1_end_write_request(struct bio *bio, int error)
 		bio_put(to_put);
 }
 
-
 /*
  * This routine returns the disk from which the requested read should
  * be done. There is a per-array 'next expected sequential IO' sector
@@ -540,11 +539,7 @@ static int read_balance(struct r1conf *conf, struct r1bio *r1_bio, int *max_sect
 	has_nonrot_disk = 0;
 	choose_next_idle = 0;
 
-	if (conf->mddev->recovery_cp < MaxSector &&
-	    (this_sector + sectors >= conf->next_resync))
-		choose_first = 1;
-	else
-		choose_first = 0;
+	choose_first = (conf->mddev->recovery_cp < this_sector + sectors);
 
 	for (disk = 0 ; disk < conf->raid_disks * 2 ; disk++) {
 		sector_t dist;
@@ -831,7 +826,7 @@ static void flush_pending_writes(struct r1conf *conf)
  *    there is no normal IO happeing.  It must arrange to call
  *    lower_barrier when the particular background IO completes.
  */
-static void raise_barrier(struct r1conf *conf)
+static void raise_barrier(struct r1conf *conf, sector_t sector_nr)
 {
 	spin_lock_irq(&conf->resync_lock);
 
@@ -841,6 +836,7 @@ static void raise_barrier(struct r1conf *conf)
 
 	/* block any new IO from starting */
 	conf->barrier++;
+	conf->next_resync = sector_nr;
 
 	/* For these conditions we must wait:
 	 * A: while the array is in frozen state
@@ -849,14 +845,17 @@ static void raise_barrier(struct r1conf *conf)
 	 * C: next_resync + RESYNC_SECTORS > start_next_window, meaning
 	 *    next resync will reach to the window which normal bios are
 	 *    handling.
+	 * D: while there are any active requests in the current window.
 	 */
 	wait_event_lock_irq(conf->wait_barrier,
 			    !conf->array_frozen &&
 			    conf->barrier < RESYNC_DEPTH &&
+			    conf->current_window_requests == 0 &&
 			    (conf->start_next_window >=
 			     conf->next_resync + RESYNC_SECTORS),
 			    conf->resync_lock);
 
+	conf->nr_pending++;
 	spin_unlock_irq(&conf->resync_lock);
 }
 
@@ -866,6 +865,7 @@ static void lower_barrier(struct r1conf *conf)
 	BUG_ON(conf->barrier <= 0);
 	spin_lock_irqsave(&conf->resync_lock, flags);
 	conf->barrier--;
+	conf->nr_pending--;
 	spin_unlock_irqrestore(&conf->resync_lock, flags);
 	wake_up(&conf->wait_barrier);
 }
@@ -877,12 +877,10 @@ static bool need_to_wait_for_sync(struct r1conf *conf, struct bio *bio)
 	if (conf->array_frozen || !bio)
 		wait = true;
 	else if (conf->barrier && bio_data_dir(bio) == WRITE) {
-		if (conf->next_resync < RESYNC_WINDOW_SECTORS)
-			wait = true;
-		else if ((conf->next_resync - RESYNC_WINDOW_SECTORS
-				>= bio_end_sector(bio)) ||
-			 (conf->next_resync + NEXT_NORMALIO_DISTANCE
-				<= bio->bi_iter.bi_sector))
+		if ((conf->mddev->curr_resync_completed
+		     >= bio_end_sector(bio)) ||
+		    (conf->next_resync + NEXT_NORMALIO_DISTANCE
+		     <= bio->bi_iter.bi_sector))
 			wait = false;
 		else
 			wait = true;
@@ -902,25 +900,25 @@ static sector_t wait_barrier(struct r1conf *conf, struct bio *bio)
 		 * However if there are already pending
 		 * requests (preventing the barrier from
 		 * rising completely), and the
-		 * pre-process bio queue isn't empty,
+		 * per-process bio queue isn't empty,
 		 * then don't wait, as we need to empty
-		 * that queue to get the nr_pending
-		 * count down.
+		 * that queue to allow conf->start_next_window
+		 * to increase.
 		 */
 		wait_event_lock_irq(conf->wait_barrier,
 				    !conf->array_frozen &&
 				    (!conf->barrier ||
-				    ((conf->start_next_window <
-				      conf->next_resync + RESYNC_SECTORS) &&
-				     current->bio_list &&
-				     !bio_list_empty(current->bio_list))),
+				     ((conf->start_next_window <
+				       conf->next_resync + RESYNC_SECTORS) &&
+				      current->bio_list &&
+				      !bio_list_empty(current->bio_list))),
 				    conf->resync_lock);
 		conf->nr_waiting--;
 	}
 
 	if (bio && bio_data_dir(bio) == WRITE) {
-		if (conf->next_resync + NEXT_NORMALIO_DISTANCE
-		    <= bio->bi_iter.bi_sector) {
+		if (bio->bi_iter.bi_sector >=
+		    conf->mddev->curr_resync_completed) {
 			if (conf->start_next_window == MaxSector)
 				conf->start_next_window =
 					conf->next_resync +
@@ -1002,8 +1000,7 @@ static void unfreeze_array(struct r1conf *conf)
 	spin_unlock_irq(&conf->resync_lock);
 }
 
-
-/* duplicate the data pages for behind I/O 
+/* duplicate the data pages for behind I/O
  */
 static void alloc_behind_pages(struct bio *bio, struct r1bio *r1_bio)
 {
@@ -1186,6 +1183,7 @@ read_again:
 				   atomic_read(&bitmap->behind_writes) == 0);
 		}
 		r1_bio->read_disk = rdisk;
+		r1_bio->start_next_window = 0;
 
 		read_bio = bio_clone_mddev(bio, GFP_NOIO, mddev);
 		bio_trim(read_bio, r1_bio->sector - bio->bi_iter.bi_sector,
@@ -1471,7 +1469,6 @@ static void status(struct seq_file *seq, struct mddev *mddev)
 	seq_printf(seq, "]");
 }
 
-
 static void error(struct mddev *mddev, struct md_rdev *rdev)
 {
 	char b[BDEVNAME_SIZE];
@@ -1501,12 +1498,12 @@ static void error(struct mddev *mddev, struct md_rdev *rdev)
 		mddev->degraded++;
 		set_bit(Faulty, &rdev->flags);
 		spin_unlock_irqrestore(&conf->device_lock, flags);
-		/*
-		 * if recovery is running, make sure it aborts.
-		 */
-		set_bit(MD_RECOVERY_INTR, &mddev->recovery);
 	} else
 		set_bit(Faulty, &rdev->flags);
+	/*
+	 * if recovery is running, make sure it aborts.
+	 */
+	set_bit(MD_RECOVERY_INTR, &mddev->recovery);
 	set_bit(MD_CHANGE_DEVS, &mddev->flags);
 	printk(KERN_ALERT
 	       "md/raid1:%s: Disk failure on %s, disabling device.\n"
@@ -1548,8 +1545,13 @@ static void close_sync(struct r1conf *conf)
 	mempool_destroy(conf->r1buf_pool);
 	conf->r1buf_pool = NULL;
 
+	spin_lock_irq(&conf->resync_lock);
 	conf->next_resync = 0;
 	conf->start_next_window = MaxSector;
+	conf->current_window_requests +=
+		conf->next_window_requests;
+	conf->next_window_requests = 0;
+	spin_unlock_irq(&conf->resync_lock);
 }
 
 static int raid1_spare_active(struct mddev *mddev)
@@ -1560,7 +1562,7 @@ static int raid1_spare_active(struct mddev *mddev)
 	unsigned long flags;
 
 	/*
-	 * Find all failed disks within the RAID1 configuration 
+	 * Find all failed disks within the RAID1 configuration
 	 * and mark them readable.
 	 * Called under mddev lock, so rcu protection not needed.
 	 */
@@ -1601,7 +1603,6 @@ static int raid1_spare_active(struct mddev *mddev)
 	return count;
 }
 
-
 static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev)
 {
 	struct r1conf *conf = mddev->private;
@@ -1730,7 +1731,6 @@ abort:
 	return err;
 }
 
-
 static void end_sync_read(struct bio *bio, int error)
 {
 	struct r1bio *r1_bio = bio->bi_private;
@@ -1942,7 +1942,7 @@ static int fix_sync_read_error(struct r1bio *r1_bio)
 	return 1;
 }
 
-static int process_checks(struct r1bio *r1_bio)
+static void process_checks(struct r1bio *r1_bio)
 {
 	/* We have read all readable devices.  If we haven't
 	 * got the block, then there is no hope left.
@@ -2034,7 +2034,6 @@ static int process_checks(struct r1bio *r1_bio)
 
 		bio_copy_data(sbio, pbio);
 	}
-	return 0;
 }
 
 static void sync_request_write(struct mddev *mddev, struct r1bio *r1_bio)
@@ -2052,8 +2051,8 @@ static void sync_request_write(struct mddev *mddev, struct r1bio *r1_bio)
 			return;
 
 	if (test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery))
-		if (process_checks(r1_bio) < 0)
-			return;
+		process_checks(r1_bio);
+
 	/*
 	 * schedule writes
 	 */
@@ -2150,7 +2149,7 @@ static void fix_read_error(struct r1conf *conf, int read_disk,
 			d--;
 			rdev = conf->mirrors[d].rdev;
 			if (rdev &&
-			    test_bit(In_sync, &rdev->flags))
+			    !test_bit(Faulty, &rdev->flags))
 				r1_sync_page_io(rdev, sect, s,
 						conf->tmppage, WRITE);
 		}
@@ -2162,7 +2161,7 @@ static void fix_read_error(struct r1conf *conf, int read_disk,
 			d--;
 			rdev = conf->mirrors[d].rdev;
 			if (rdev &&
-			    test_bit(In_sync, &rdev->flags)) {
+			    !test_bit(Faulty, &rdev->flags)) {
 				if (r1_sync_page_io(rdev, sect, s,
 						    conf->tmppage, READ)) {
 					atomic_add(s, &rdev->corrected_errors);
@@ -2453,7 +2452,6 @@ static void raid1d(struct md_thread *thread)
 	blk_finish_plug(&plug);
 }
 
-
 static int init_resync(struct r1conf *conf)
 {
 	int buffs;
@@ -2541,9 +2539,8 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, int *skipp
 
 	bitmap_cond_end_sync(mddev->bitmap, sector_nr);
 	r1_bio = mempool_alloc(conf->r1buf_pool, GFP_NOIO);
-	raise_barrier(conf);
 
-	conf->next_resync = sector_nr;
+	raise_barrier(conf, sector_nr);
 
 	rcu_read_lock();
 	/*
@@ -2718,7 +2715,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, int *skipp
 						/* remove last page from this bio */
 						bio->bi_vcnt--;
 						bio->bi_iter.bi_size -= len;
-						bio->bi_flags &= ~(1<< BIO_SEG_VALID);
+						__clear_bit(BIO_SEG_VALID, &bio->bi_flags);
 					}
 					goto bio_full;
 				}
@@ -2943,9 +2940,9 @@ static int run(struct mddev *mddev)
 		printk(KERN_NOTICE "md/raid1:%s: not clean"
 		       " -- starting background reconstruction\n",
 		       mdname(mddev));
-	printk(KERN_INFO 
+	printk(KERN_INFO
 		"md/raid1:%s: active with %d out of %d mirrors\n",
-		mdname(mddev), mddev->raid_disks - mddev->degraded, 
+		mdname(mddev), mddev->raid_disks - mddev->degraded,
 		mddev->raid_disks);
 
 	/*
diff --git a/drivers/md/raid1.h b/drivers/md/raid1.h
index 9bebca7bff2f..33bda55ef9f7 100644
--- a/drivers/md/raid1.h
+++ b/drivers/md/raid1.h
@@ -90,7 +90,6 @@ struct r1conf {
 	 */
 	int			recovery_disabled;
 
-
 	/* poolinfo contains information about the content of the
 	 * mempools - it changes when the array grows or shrinks
 	 */
@@ -103,7 +102,6 @@ struct r1conf {
 	 */
 	struct page		*tmppage;
 
-
 	/* When taking over an array from a different personality, we store
 	 * the new thread here until we fully activate the array.
 	 */
diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index cb882aae9e20..32e282f4c83c 100644
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -366,7 +366,6 @@ static void raid10_end_read_request(struct bio *bio, int error)
 	struct md_rdev *rdev;
 	struct r10conf *conf = r10_bio->mddev->private;
 
-
 	slot = r10_bio->read_slot;
 	dev = r10_bio->devs[slot].devnum;
 	rdev = r10_bio->devs[slot].rdev;
@@ -1559,7 +1558,6 @@ static void make_request(struct mddev *mddev, struct bio *bio)
 
 	md_write_start(mddev, bio);
 
-
 	do {
 
 		/*
@@ -1684,13 +1682,12 @@ static void error(struct mddev *mddev, struct md_rdev *rdev)
 		spin_unlock_irqrestore(&conf->device_lock, flags);
 		return;
 	}
-	if (test_and_clear_bit(In_sync, &rdev->flags)) {
+	if (test_and_clear_bit(In_sync, &rdev->flags))
 		mddev->degraded++;
-			/*
-		 * if recovery is running, make sure it aborts.
-		 */
-		set_bit(MD_RECOVERY_INTR, &mddev->recovery);
-	}
+	/*
+	 * If recovery is running, make sure it aborts.
+	 */
+	set_bit(MD_RECOVERY_INTR, &mddev->recovery);
 	set_bit(Blocked, &rdev->flags);
 	set_bit(Faulty, &rdev->flags);
 	set_bit(MD_CHANGE_DEVS, &mddev->flags);
@@ -1783,7 +1780,6 @@ static int raid10_spare_active(struct mddev *mddev)
 	return count;
 }
 
-
 static int raid10_add_disk(struct mddev *mddev, struct md_rdev *rdev)
 {
 	struct r10conf *conf = mddev->private;
@@ -1930,7 +1926,6 @@ abort:
 	return err;
 }
 
-
 static void end_sync_read(struct bio *bio, int error)
 {
 	struct r10bio *r10_bio = bio->bi_private;
@@ -2296,7 +2291,6 @@ static void recovery_request_write(struct mddev *mddev, struct r10bio *r10_bio)
 	}
 }
 
-
 /*
  * Used by fix_read_error() to decay the per rdev read_errors.
  * We halve the read error count for every hour that has elapsed
@@ -2853,7 +2847,6 @@ static void raid10d(struct md_thread *thread)
 	blk_finish_plug(&plug);
 }
 
-
 static int init_resync(struct r10conf *conf)
 {
 	int buffs;
@@ -2954,6 +2947,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr,
 		 */
 		if (test_bit(MD_RECOVERY_RESHAPE, &mddev->recovery)) {
 			end_reshape(conf);
+			close_sync(conf);
 			return 0;
 		}
 
@@ -3082,6 +3076,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr,
 			}
 
 			r10_bio = mempool_alloc(conf->r10buf_pool, GFP_NOIO);
+			r10_bio->state = 0;
 			raise_barrier(conf, rb2 != NULL);
 			atomic_set(&r10_bio->remaining, 0);
 
@@ -3270,6 +3265,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr,
 		if (sync_blocks < max_sync)
 			max_sync = sync_blocks;
 		r10_bio = mempool_alloc(conf->r10buf_pool, GFP_NOIO);
+		r10_bio->state = 0;
 
 		r10_bio->mddev = mddev;
 		atomic_set(&r10_bio->remaining, 0);
@@ -3386,7 +3382,7 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr,
 				/* remove last page from this bio */
 				bio2->bi_vcnt--;
 				bio2->bi_iter.bi_size -= len;
-				bio2->bi_flags &= ~(1<< BIO_SEG_VALID);
+				__clear_bit(BIO_SEG_VALID, &bio2->bi_flags);
 			}
 			goto bio_full;
 		}
@@ -3774,7 +3770,6 @@ static int run(struct mddev *mddev)
 		blk_queue_merge_bvec(mddev->queue, raid10_mergeable_bvec);
 	}
 
-
 	if (md_integrity_register(mddev))
 		goto out_free_conf;
 
@@ -3832,6 +3827,8 @@ static int stop(struct mddev *mddev)
 		mempool_destroy(conf->r10bio_pool);
 	safe_put_page(conf->tmppage);
 	kfree(conf->mirrors);
+	kfree(conf->mirrors_old);
+	kfree(conf->mirrors_new);
 	kfree(conf);
 	mddev->private = NULL;
 	return 0;
@@ -4119,7 +4116,7 @@ static int raid10_start_reshape(struct mddev *mddev)
 		memcpy(conf->mirrors_new, conf->mirrors,
 		       sizeof(struct raid10_info)*conf->prev.raid_disks);
 		smp_mb();
-		kfree(conf->mirrors_old); /* FIXME and elsewhere */
+		kfree(conf->mirrors_old);
 		conf->mirrors_old = conf->mirrors;
 		conf->mirrors = conf->mirrors_new;
 		conf->mirrors_new = NULL;
@@ -4385,6 +4382,7 @@ static sector_t reshape_request(struct mddev *mddev, sector_t sector_nr,
 read_more:
 	/* Now schedule reads for blocks from sector_nr to last */
 	r10_bio = mempool_alloc(conf->r10buf_pool, GFP_NOIO);
+	r10_bio->state = 0;
 	raise_barrier(conf, sectors_done != 0);
 	atomic_set(&r10_bio->remaining, 0);
 	r10_bio->mddev = mddev;
@@ -4399,6 +4397,7 @@ read_more:
 		 * on all the target devices.
 		 */
 		// FIXME
+		mempool_free(r10_bio, conf->r10buf_pool);
 		set_bit(MD_RECOVERY_INTR, &mddev->recovery);
 		return sectors_done;
 	}
@@ -4411,8 +4410,8 @@ read_more:
 	read_bio->bi_private = r10_bio;
 	read_bio->bi_end_io = end_sync_read;
 	read_bio->bi_rw = READ;
-	read_bio->bi_flags &= ~(BIO_POOL_MASK - 1);
-	read_bio->bi_flags |= 1 << BIO_UPTODATE;
+	read_bio->bi_flags &= (~0UL << BIO_RESET_BITS);
+	__set_bit(BIO_UPTODATE, &read_bio->bi_flags);
 	read_bio->bi_vcnt = 0;
 	read_bio->bi_iter.bi_size = 0;
 	r10_bio->master_bio = read_bio;
@@ -4469,7 +4468,7 @@ read_more:
 				/* Remove last page from this bio */
 				bio2->bi_vcnt--;
 				bio2->bi_iter.bi_size -= len;
-				bio2->bi_flags &= ~(1<<BIO_SEG_VALID);
+				__clear_bit(BIO_SEG_VALID, &bio2->bi_flags);
 			}
 			goto bio_full;
 		}
@@ -4571,7 +4570,6 @@ static void end_reshape(struct r10conf *conf)
 	conf->fullsync = 0;
 }
 
-
 static int handle_reshape_read_error(struct mddev *mddev,
 				     struct r10bio *r10_bio)
 {
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 6234b2e84587..9c66e5997fc8 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -64,6 +64,10 @@
 #define cpu_to_group(cpu) cpu_to_node(cpu)
 #define ANY_GROUP NUMA_NO_NODE
 
+static bool devices_handle_discard_safely = false;
+module_param(devices_handle_discard_safely, bool, 0644);
+MODULE_PARM_DESC(devices_handle_discard_safely,
+		 "Set to Y if all devices in each array reliably return zeroes on reads from discarded regions");
 static struct workqueue_struct *raid5_wq;
 /*
  * Stripe cache
@@ -459,7 +463,6 @@ static inline void insert_hash(struct r5conf *conf, struct stripe_head *sh)
 	hlist_add_head(&sh->hash, hp);
 }
 
-
 /* find an idle stripe, make sure it is unhashed, and return it. */
 static struct stripe_head *get_free_stripe(struct r5conf *conf, int hash)
 {
@@ -527,9 +530,7 @@ static void init_stripe(struct stripe_head *sh, sector_t sector, int previous)
 	BUG_ON(stripe_operations_active(sh));
 
 	pr_debug("init_stripe called, stripe %llu\n",
-		(unsigned long long)sh->sector);
-
-	remove_hash(sh);
+		(unsigned long long)sector);
 retry:
 	seq = read_seqcount_begin(&conf->gen_lock);
 	sh->generation = conf->generation - previous;
@@ -538,7 +539,6 @@ retry:
 	stripe_set_idx(sector, conf, previous, sh);
 	sh->state = 0;
 
-
 	for (i = sh->disks; i--; ) {
 		struct r5dev *dev = &sh->dev[i];
 
@@ -1346,7 +1346,6 @@ ops_run_compute6_2(struct stripe_head *sh, struct raid5_percpu *percpu)
 	}
 }
 
-
 static void ops_complete_prexor(void *stripe_head_ref)
 {
 	struct stripe_head *sh = stripe_head_ref;
@@ -2415,7 +2414,6 @@ static sector_t raid5_compute_sector(struct r5conf *conf, sector_t r_sector,
 	return new_sector;
 }
 
-
 static sector_t compute_blocknr(struct stripe_head *sh, int i, int previous)
 {
 	struct r5conf *conf = sh->raid_conf;
@@ -2433,7 +2431,6 @@ static sector_t compute_blocknr(struct stripe_head *sh, int i, int previous)
 	sector_t r_sector;
 	struct stripe_head sh2;
 
-
 	chunk_offset = sector_div(new_sector, sectors_per_chunk);
 	stripe = new_sector;
 
@@ -2537,7 +2534,6 @@ static sector_t compute_blocknr(struct stripe_head *sh, int i, int previous)
 	return r_sector;
 }
 
-
 static void
 schedule_reconstruction(struct stripe_head *sh, struct stripe_head_state *s,
 			 int rcw, int expand)
@@ -2922,7 +2918,7 @@ static int fetch_block(struct stripe_head *sh, struct stripe_head_state *s,
 	      (!test_bit(R5_Insync, &dev->flags) || test_bit(STRIPE_PREREAD_ACTIVE, &sh->state)) &&
 	      !test_bit(R5_OVERWRITE, &fdev[0]->flags)) ||
 	     (sh->raid_conf->level == 6 && s->failed && s->to_write &&
-	      s->to_write < sh->raid_conf->raid_disks - 2 &&
+	      s->to_write - s->non_overwrite < sh->raid_conf->raid_disks - 2 &&
 	      (!test_bit(R5_Insync, &dev->flags) || test_bit(STRIPE_PREREAD_ACTIVE, &sh->state))))) {
 		/* we would like to get this block, possibly by computing it,
 		 * otherwise read it if the backing disk is insync
@@ -3009,7 +3005,6 @@ static void handle_stripe_fill(struct stripe_head *sh,
 	set_bit(STRIPE_HANDLE, &sh->state);
 }
 
-
 /* handle_stripe_clean_event
  * any written block on an uptodate or failed drive can be returned.
  * Note that if we 'wrote' to a failed drive, it will be UPTODATE, but
@@ -3300,7 +3295,6 @@ static void handle_parity_checks5(struct r5conf *conf, struct stripe_head *sh,
 	}
 }
 
-
 static void handle_parity_checks6(struct r5conf *conf, struct stripe_head *sh,
 				  struct stripe_head_state *s,
 				  int disks)
@@ -3817,6 +3811,8 @@ static void handle_stripe(struct stripe_head *sh)
 				set_bit(R5_Wantwrite, &dev->flags);
 				if (prexor)
 					continue;
+				if (s.failed > 1)
+					continue;
 				if (!test_bit(R5_Insync, &dev->flags) ||
 				    ((i == sh->pd_idx || i == sh->qd_idx)  &&
 				     s.failed == 0))
@@ -3933,7 +3929,6 @@ static void handle_stripe(struct stripe_head *sh)
 			}
 		}
 
-
 	/* Finish reconstruct operations initiated by the expansion process */
 	if (sh->reconstruct_state == reconstruct_state_result) {
 		struct stripe_head *sh_src
@@ -4131,7 +4126,6 @@ static int raid5_mergeable_bvec(struct request_queue *q,
 		return max;
 }
 
-
 static int in_chunk_boundary(struct mddev *mddev, struct bio *bio)
 {
 	sector_t sector = bio->bi_iter.bi_sector + get_start_sect(bio->bi_bdev);
@@ -4161,7 +4155,6 @@ static void add_bio_to_retry(struct bio *bi,struct r5conf *conf)
 	md_wakeup_thread(conf->mddev->thread);
 }
 
-
 static struct bio *remove_bio_from_retry(struct r5conf *conf)
 {
 	struct bio *bi;
@@ -4185,7 +4178,6 @@ static struct bio *remove_bio_from_retry(struct r5conf *conf)
 	return bi;
 }
 
-
 /*
  *  The "raid5_align_endio" should check if the read succeeded and if it
  *  did, call bio_endio on the original bio (having bio_put the new bio
@@ -4218,7 +4210,6 @@ static void raid5_align_endio(struct bio *bi, int error)
 		return;
 	}
 
-
 	pr_debug("raid5_align_endio : io error...handing IO for a retry\n");
 
 	add_bio_to_retry(raid_bi, conf);
@@ -4243,7 +4234,6 @@ static int bio_fits_rdev(struct bio *bi)
 	return 1;
 }
 
-
 static int chunk_aligned_read(struct mddev *mddev, struct bio * raid_bio)
 {
 	struct r5conf *conf = mddev->private;
@@ -4295,7 +4285,7 @@ static int chunk_aligned_read(struct mddev *mddev, struct bio * raid_bio)
 		rcu_read_unlock();
 		raid_bio->bi_next = (void*)rdev;
 		align_bi->bi_bdev =  rdev->bdev;
-		align_bi->bi_flags &= ~(1 << BIO_SEG_VALID);
+		__clear_bit(BIO_SEG_VALID, &align_bi->bi_flags);
 
 		if (!bio_fits_rdev(align_bi) ||
 		    is_badblock(rdev, align_bi->bi_iter.bi_sector,
@@ -5440,7 +5430,6 @@ raid5_skip_copy = __ATTR(skip_copy, S_IRUGO | S_IWUSR,
 					raid5_show_skip_copy,
 					raid5_store_skip_copy);
 
-
 static ssize_t
 stripe_cache_active_show(struct mddev *mddev, char *page)
 {
@@ -5892,7 +5881,6 @@ static struct r5conf *setup_conf(struct mddev *mddev)
 		return ERR_PTR(-ENOMEM);
 }
 
-
 static int only_parity(int raid_disk, int algo, int raid_disks, int max_degraded)
 {
 	switch (algo) {
@@ -5905,7 +5893,7 @@ static int only_parity(int raid_disk, int algo, int raid_disks, int max_degraded
 			return 1;
 		break;
 	case ALGORITHM_PARITY_0_6:
-		if (raid_disk == 0 || 
+		if (raid_disk == 0 ||
 		    raid_disk == raid_disks - 1)
 			return 1;
 		break;
@@ -6159,7 +6147,6 @@ static int run(struct mddev *mddev)
 							"reshape");
 	}
 
-
 	/* Ok, everything is just fine now */
 	if (mddev->to_remove == &raid5_attrs_group)
 		mddev->to_remove = NULL;
@@ -6206,7 +6193,7 @@ static int run(struct mddev *mddev)
 		mddev->queue->limits.discard_granularity = stripe;
 		/*
 		 * unaligned part of discard request will be ignored, so can't
-		 * guarantee discard_zerors_data
+		 * guarantee discard_zeroes_data
 		 */
 		mddev->queue->limits.discard_zeroes_data = 0;
 
@@ -6231,6 +6218,18 @@ static int run(struct mddev *mddev)
 			    !bdev_get_queue(rdev->bdev)->
 						limits.discard_zeroes_data)
 				discard_supported = false;
+			/* Unfortunately, discard_zeroes_data is not currently
+			 * a guarantee - just a hint.  So we only allow DISCARD
+			 * if the sysadmin has confirmed that only safe devices
+			 * are in use by setting a module parameter.
+			 */
+			if (!devices_handle_discard_safely) {
+				if (discard_supported) {
+					pr_info("md/raid456: discard support disabled due to uncertainty.\n");
+					pr_info("Set raid456.devices_handle_discard_safely=Y to override.\n");
+				}
+				discard_supported = false;
+			}
 		}
 
 		if (discard_supported &&
@@ -6796,7 +6795,6 @@ static void raid5_quiesce(struct mddev *mddev, int state)
 	}
 }
 
-
 static void *raid45_takeover_raid0(struct mddev *mddev, int level)
 {
 	struct r0conf *raid0_conf = mddev->private;
@@ -6823,7 +6821,6 @@ static void *raid45_takeover_raid0(struct mddev *mddev, int level)
 	return setup_conf(mddev);
 }
 
-
 static void *raid5_takeover_raid1(struct mddev *mddev)
 {
 	int chunksect;
@@ -6884,7 +6881,6 @@ static void *raid5_takeover_raid6(struct mddev *mddev)
 	return setup_conf(mddev);
 }
 
-
 static int raid5_check_reshape(struct mddev *mddev)
 {
 	/* For a 2-drive array, the layout and chunk size can be changed
@@ -7033,7 +7029,6 @@ static void *raid6_takeover(struct mddev *mddev)
 	return setup_conf(mddev);
 }
 
-
 static struct md_personality raid6_personality =
 {
 	.name		= "raid6",
diff --git a/drivers/md/raid5.h b/drivers/md/raid5.h
index bc72cd4be5f8..d59f5ca743cd 100644
--- a/drivers/md/raid5.h
+++ b/drivers/md/raid5.h
@@ -155,7 +155,7 @@
  */
 
 /*
- * Operations state - intermediate states that are visible outside of 
+ * Operations state - intermediate states that are visible outside of
  *   STRIPE_ACTIVE.
  * In general _idle indicates nothing is running, _run indicates a data
  * processing operation is active, and _result means the data processing result
@@ -364,7 +364,6 @@ enum {
  * HANDLE gets cleared if stripe_handle leaves nothing locked.
  */
 
-
 struct disk_info {
 	struct md_rdev	*rdev, *replacement;
 };
@@ -528,7 +527,6 @@ struct r5conf {
 #define ALGORITHM_ROTATING_N_RESTART	9 /* DDF PRL=6 RLQ=2 */
 #define ALGORITHM_ROTATING_N_CONTINUE	10 /*DDF PRL=6 RLQ=3 */
 
-
 /* For every RAID5 algorithm we define a RAID6 algorithm
  * with exactly the same layout for data and parity, and
  * with the Q block always on the last device (N-1).