Merge tag 'for-5.3/block-20190708' of git://git.kernel.dk/linux-block

Pull block updates from Jens Axboe:
 "This is the main block updates for 5.3. Nothing earth shattering or
  major in here, just fixes, additions, and improvements all over the
  map. This contains:

   - Series of documentation fixes (Bart)

   - Optimization of the blk-mq ctx get/put (Bart)

   - null_blk removal race condition fix (Bob)

   - req/bio_op() cleanups (Chaitanya)

   - Series cleaning up the segment accounting, and request/bio mapping
     (Christoph)

   - Series cleaning up the page getting/putting for bios (Christoph)

   - block cgroup cleanups and moving it to where it is used (Christoph)

   - block cgroup fixes (Tejun)

   - Series of fixes and improvements to bcache, most notably a write
     deadlock fix (Coly)

   - blk-iolatency STS_AGAIN and accounting fixes (Dennis)

   - Series of improvements and fixes to BFQ (Douglas, Paolo)

   - debugfs_create() return value check removal for drbd (Greg)

   - Use struct_size(), where appropriate (Gustavo)

   - Two lighnvm fixes (Heiner, Geert)

   - MD fixes, including a read balance and corruption fix (Guoqing,
     Marcos, Xiao, Yufen)

   - block opal shadow mbr additions (Jonas, Revanth)

   - sbitmap compare-and-exhange improvemnts (Pavel)

   - Fix for potential bio->bi_size overflow (Ming)

   - NVMe pull requests:
       - improved PCIe suspent support (Keith Busch)
       - error injection support for the admin queue (Akinobu Mita)
       - Fibre Channel discovery improvements (James Smart)
       - tracing improvements including nvmetc tracing support (Minwoo Im)
       - misc fixes and cleanups (Anton Eidelman, Minwoo Im, Chaitanya
         Kulkarni)"

   - Various little fixes and improvements to drivers and core"

* tag 'for-5.3/block-20190708' of git://git.kernel.dk/linux-block: (153 commits)
  blk-iolatency: fix STS_AGAIN handling
  block: nr_phys_segments needs to be zero for REQ_OP_WRITE_ZEROES
  blk-mq: simplify blk_mq_make_request()
  blk-mq: remove blk_mq_put_ctx()
  sbitmap: Replace cmpxchg with xchg
  block: fix .bi_size overflow
  block: sed-opal: check size of shadow mbr
  block: sed-opal: ioctl for writing to shadow mbr
  block: sed-opal: add ioctl for done-mark of shadow mbr
  block: never take page references for ITER_BVEC
  direct-io: use bio_release_pages in dio_bio_complete
  block_dev: use bio_release_pages in bio_unmap_user
  block_dev: use bio_release_pages in blkdev_bio_end_io
  iomap: use bio_release_pages in iomap_dio_bio_end_io
  block: use bio_release_pages in bio_map_user_iov
  block: use bio_release_pages in bio_unmap_user
  block: optionally mark pages dirty in bio_release_pages
  block: move the BIO_NO_PAGE_REF check into bio_release_pages
  block: skd_main.c: Remove call to memset after dma_alloc_coherent
  block: mtip32xx: Remove call to memset after dma_alloc_coherent
  ...

19 files changed, 741 insertions, 270 deletions

diff --git a/drivers/md/bcache/alloc.c b/drivers/md/bcache/alloc.c
index f8986effcb50..6f776823b9ba 100644
--- a/drivers/md/bcache/alloc.c
+++ b/drivers/md/bcache/alloc.c
@@ -393,6 +393,11 @@ long bch_bucket_alloc(struct cache *ca, unsigned int reserve, bool wait)
 	struct bucket *b;
 	long r;
 
+
+	/* No allocation if CACHE_SET_IO_DISABLE bit is set */
+	if (unlikely(test_bit(CACHE_SET_IO_DISABLE, &ca->set->flags)))
+		return -1;
+
 	/* fastpath */
 	if (fifo_pop(&ca->free[RESERVE_NONE], r) ||
 	    fifo_pop(&ca->free[reserve], r))
@@ -484,6 +489,10 @@ int __bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve,
 {
 	int i;
 
+	/* No allocation if CACHE_SET_IO_DISABLE bit is set */
+	if (unlikely(test_bit(CACHE_SET_IO_DISABLE, &c->flags)))
+		return -1;
+
 	lockdep_assert_held(&c->bucket_lock);
 	BUG_ON(!n || n > c->caches_loaded || n > MAX_CACHES_PER_SET);
 
diff --git a/drivers/md/bcache/bcache.h b/drivers/md/bcache/bcache.h
index fdf75352e16a..013e35a9e317 100644
--- a/drivers/md/bcache/bcache.h
+++ b/drivers/md/bcache/bcache.h
@@ -705,8 +705,8 @@ struct cache_set {
 	atomic_long_t		writeback_keys_failed;
 
 	atomic_long_t		reclaim;
+	atomic_long_t		reclaimed_journal_buckets;
 	atomic_long_t		flush_write;
-	atomic_long_t		retry_flush_write;
 
 	enum			{
 		ON_ERROR_UNREGISTER,
@@ -726,8 +726,6 @@ struct cache_set {
 
 #define BUCKET_HASH_BITS	12
 	struct hlist_head	bucket_hash[1 << BUCKET_HASH_BITS];
-
-	DECLARE_HEAP(struct btree *, flush_btree);
 };
 
 struct bbio {
@@ -1006,7 +1004,7 @@ int bch_flash_dev_create(struct cache_set *c, uint64_t size);
 int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c,
 			  uint8_t *set_uuid);
 void bch_cached_dev_detach(struct cached_dev *dc);
-void bch_cached_dev_run(struct cached_dev *dc);
+int bch_cached_dev_run(struct cached_dev *dc);
 void bcache_device_stop(struct bcache_device *d);
 
 void bch_cache_set_unregister(struct cache_set *c);
diff --git a/drivers/md/bcache/bset.c b/drivers/md/bcache/bset.c
index 268f1b685084..08768796b543 100644
--- a/drivers/md/bcache/bset.c
+++ b/drivers/md/bcache/bset.c
@@ -347,22 +347,19 @@ EXPORT_SYMBOL(bch_btree_keys_alloc);
 void bch_btree_keys_init(struct btree_keys *b, const struct btree_keys_ops *ops,
 			 bool *expensive_debug_checks)
 {
-	unsigned int i;
-
 	b->ops = ops;
 	b->expensive_debug_checks = expensive_debug_checks;
 	b->nsets = 0;
 	b->last_set_unwritten = 0;
 
-	/* XXX: shouldn't be needed */
-	for (i = 0; i < MAX_BSETS; i++)
-		b->set[i].size = 0;
 	/*
-	 * Second loop starts at 1 because b->keys[0]->data is the memory we
-	 * allocated
+	 * struct btree_keys in embedded in struct btree, and struct
+	 * bset_tree is embedded into struct btree_keys. They are all
+	 * initialized as 0 by kzalloc() in mca_bucket_alloc(), and
+	 * b->set[0].data is allocated in bch_btree_keys_alloc(), so we
+	 * don't have to initiate b->set[].size and b->set[].data here
+	 * any more.
 	 */
-	for (i = 1; i < MAX_BSETS; i++)
-		b->set[i].data = NULL;
 }
 EXPORT_SYMBOL(bch_btree_keys_init);
 
@@ -970,45 +967,25 @@ static struct bset_search_iter bset_search_tree(struct bset_tree *t,
 	unsigned int inorder, j, n = 1;
 
 	do {
-		/*
-		 * A bit trick here.
-		 * If p < t->size, (int)(p - t->size) is a minus value and
-		 * the most significant bit is set, right shifting 31 bits
-		 * gets 1. If p >= t->size, the most significant bit is
-		 * not set, right shifting 31 bits gets 0.
-		 * So the following 2 lines equals to
-		 *	if (p >= t->size)
-		 *		p = 0;
-		 * but a branch instruction is avoided.
-		 */
 		unsigned int p = n << 4;
 
-		p &= ((int) (p - t->size)) >> 31;
-
-		prefetch(&t->tree[p]);
+		if (p < t->size)
+			prefetch(&t->tree[p]);
 
 		j = n;
 		f = &t->tree[j];
 
-		/*
-		 * Similar bit trick, use subtract operation to avoid a branch
-		 * instruction.
-		 *
-		 * n = (f->mantissa > bfloat_mantissa())
-		 *	? j * 2
-		 *	: j * 2 + 1;
-		 *
-		 * We need to subtract 1 from f->mantissa for the sign bit trick
-		 * to work  - that's done in make_bfloat()
-		 */
-		if (likely(f->exponent != 127))
-			n = j * 2 + (((unsigned int)
-				      (f->mantissa -
-				       bfloat_mantissa(search, f))) >> 31);
-		else
-			n = (bkey_cmp(tree_to_bkey(t, j), search) > 0)
-				? j * 2
-				: j * 2 + 1;
+		if (likely(f->exponent != 127)) {
+			if (f->mantissa >= bfloat_mantissa(search, f))
+				n = j * 2;
+			else
+				n = j * 2 + 1;
+		} else {
+			if (bkey_cmp(tree_to_bkey(t, j), search) > 0)
+				n = j * 2;
+			else
+				n = j * 2 + 1;
+		}
 	} while (n < t->size);
 
 	inorder = to_inorder(j, t);
diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index 773f5fdad25f..ba434d9ac720 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -35,7 +35,7 @@
 #include <linux/rcupdate.h>
 #include <linux/sched/clock.h>
 #include <linux/rculist.h>
-
+#include <linux/delay.h>
 #include <trace/events/bcache.h>
 
 /*
@@ -613,6 +613,10 @@ static void mca_data_alloc(struct btree *b, struct bkey *k, gfp_t gfp)
 static struct btree *mca_bucket_alloc(struct cache_set *c,
 				      struct bkey *k, gfp_t gfp)
 {
+	/*
+	 * kzalloc() is necessary here for initialization,
+	 * see code comments in bch_btree_keys_init().
+	 */
 	struct btree *b = kzalloc(sizeof(struct btree), gfp);
 
 	if (!b)
@@ -655,7 +659,25 @@ static int mca_reap(struct btree *b, unsigned int min_order, bool flush)
 		up(&b->io_mutex);
 	}
 
+retry:
+	/*
+	 * BTREE_NODE_dirty might be cleared in btree_flush_btree() by
+	 * __bch_btree_node_write(). To avoid an extra flush, acquire
+	 * b->write_lock before checking BTREE_NODE_dirty bit.
+	 */
 	mutex_lock(&b->write_lock);
+	/*
+	 * If this btree node is selected in btree_flush_write() by journal
+	 * code, delay and retry until the node is flushed by journal code
+	 * and BTREE_NODE_journal_flush bit cleared by btree_flush_write().
+	 */
+	if (btree_node_journal_flush(b)) {
+		pr_debug("bnode %p is flushing by journal, retry", b);
+		mutex_unlock(&b->write_lock);
+		udelay(1);
+		goto retry;
+	}
+
 	if (btree_node_dirty(b))
 		__bch_btree_node_write(b, &cl);
 	mutex_unlock(&b->write_lock);
@@ -778,10 +800,15 @@ void bch_btree_cache_free(struct cache_set *c)
 	while (!list_empty(&c->btree_cache)) {
 		b = list_first_entry(&c->btree_cache, struct btree, list);
 
-		if (btree_node_dirty(b))
+		/*
+		 * This function is called by cache_set_free(), no I/O
+		 * request on cache now, it is unnecessary to acquire
+		 * b->write_lock before clearing BTREE_NODE_dirty anymore.
+		 */
+		if (btree_node_dirty(b)) {
 			btree_complete_write(b, btree_current_write(b));
-		clear_bit(BTREE_NODE_dirty, &b->flags);
-
+			clear_bit(BTREE_NODE_dirty, &b->flags);
+		}
 		mca_data_free(b);
 	}
 
@@ -1067,11 +1094,25 @@ static void btree_node_free(struct btree *b)
 
 	BUG_ON(b == b->c->root);
 
+retry:
 	mutex_lock(&b->write_lock);
+	/*
+	 * If the btree node is selected and flushing in btree_flush_write(),
+	 * delay and retry until the BTREE_NODE_journal_flush bit cleared,
+	 * then it is safe to free the btree node here. Otherwise this btree
+	 * node will be in race condition.
+	 */
+	if (btree_node_journal_flush(b)) {
+		mutex_unlock(&b->write_lock);
+		pr_debug("bnode %p journal_flush set, retry", b);
+		udelay(1);
+		goto retry;
+	}
 
-	if (btree_node_dirty(b))
+	if (btree_node_dirty(b)) {
 		btree_complete_write(b, btree_current_write(b));
-	clear_bit(BTREE_NODE_dirty, &b->flags);
+		clear_bit(BTREE_NODE_dirty, &b->flags);
+	}
 
 	mutex_unlock(&b->write_lock);
 
diff --git a/drivers/md/bcache/btree.h b/drivers/md/bcache/btree.h
index d1c72ef64edf..76cfd121a486 100644
--- a/drivers/md/bcache/btree.h
+++ b/drivers/md/bcache/btree.h
@@ -158,11 +158,13 @@ enum btree_flags {
 	BTREE_NODE_io_error,
 	BTREE_NODE_dirty,
 	BTREE_NODE_write_idx,
+	BTREE_NODE_journal_flush,
 };
 
 BTREE_FLAG(io_error);
 BTREE_FLAG(dirty);
 BTREE_FLAG(write_idx);
+BTREE_FLAG(journal_flush);
 
 static inline struct btree_write *btree_current_write(struct btree *b)
 {
diff --git a/drivers/md/bcache/io.c b/drivers/md/bcache/io.c
index c25097968319..4d93f07f63e5 100644
--- a/drivers/md/bcache/io.c
+++ b/drivers/md/bcache/io.c
@@ -58,6 +58,18 @@ void bch_count_backing_io_errors(struct cached_dev *dc, struct bio *bio)
 
 	WARN_ONCE(!dc, "NULL pointer of struct cached_dev");
 
+	/*
+	 * Read-ahead requests on a degrading and recovering md raid
+	 * (e.g. raid6) device might be failured immediately by md
+	 * raid code, which is not a real hardware media failure. So
+	 * we shouldn't count failed REQ_RAHEAD bio to dc->io_errors.
+	 */
+	if (bio->bi_opf & REQ_RAHEAD) {
+		pr_warn_ratelimited("%s: Read-ahead I/O failed on backing device, ignore",
+				    dc->backing_dev_name);
+		return;
+	}
+
 	errors = atomic_add_return(1, &dc->io_errors);
 	if (errors < dc->error_limit)
 		pr_err("%s: IO error on backing device, unrecoverable",
diff --git a/drivers/md/bcache/journal.c b/drivers/md/bcache/journal.c
index 12dae9348147..be2a2a201603 100644
--- a/drivers/md/bcache/journal.c
+++ b/drivers/md/bcache/journal.c
@@ -100,6 +100,20 @@ reread:		left = ca->sb.bucket_size - offset;
 
 			blocks = set_blocks(j, block_bytes(ca->set));
 
+			/*
+			 * Nodes in 'list' are in linear increasing order of
+			 * i->j.seq, the node on head has the smallest (oldest)
+			 * journal seq, the node on tail has the biggest
+			 * (latest) journal seq.
+			 */
+
+			/*
+			 * Check from the oldest jset for last_seq. If
+			 * i->j.seq < j->last_seq, it means the oldest jset
+			 * in list is expired and useless, remove it from
+			 * this list. Otherwise, j is a condidate jset for
+			 * further following checks.
+			 */
 			while (!list_empty(list)) {
 				i = list_first_entry(list,
 					struct journal_replay, list);
@@ -109,13 +123,22 @@ reread:		left = ca->sb.bucket_size - offset;
 				kfree(i);
 			}
 
+			/* iterate list in reverse order (from latest jset) */
 			list_for_each_entry_reverse(i, list, list) {
 				if (j->seq == i->j.seq)
 					goto next_set;
 
+				/*
+				 * if j->seq is less than any i->j.last_seq
+				 * in list, j is an expired and useless jset.
+				 */
 				if (j->seq < i->j.last_seq)
 					goto next_set;
 
+				/*
+				 * 'where' points to first jset in list which
+				 * is elder then j.
+				 */
 				if (j->seq > i->j.seq) {
 					where = &i->list;
 					goto add;
@@ -129,10 +152,12 @@ add:
 			if (!i)
 				return -ENOMEM;
 			memcpy(&i->j, j, bytes);
+			/* Add to the location after 'where' points to */
 			list_add(&i->list, where);
 			ret = 1;
 
-			ja->seq[bucket_index] = j->seq;
+			if (j->seq > ja->seq[bucket_index])
+				ja->seq[bucket_index] = j->seq;
 next_set:
 			offset	+= blocks * ca->sb.block_size;
 			len	-= blocks * ca->sb.block_size;
@@ -268,7 +293,7 @@ bsearch:
 					    struct journal_replay,
 					    list)->j.seq;
 
-	return ret;
+	return 0;
 #undef read_bucket
 }
 
@@ -391,60 +416,90 @@ err:
 }
 
 /* Journalling */
-#define journal_max_cmp(l, r) \
-	(fifo_idx(&c->journal.pin, btree_current_write(l)->journal) < \
-	 fifo_idx(&(c)->journal.pin, btree_current_write(r)->journal))
-#define journal_min_cmp(l, r) \
-	(fifo_idx(&c->journal.pin, btree_current_write(l)->journal) > \
-	 fifo_idx(&(c)->journal.pin, btree_current_write(r)->journal))
 
 static void btree_flush_write(struct cache_set *c)
 {
-	/*
-	 * Try to find the btree node with that references the oldest journal
-	 * entry, best is our current candidate and is locked if non NULL:
-	 */
-	struct btree *b;
-	int i;
+	struct btree *b, *t, *btree_nodes[BTREE_FLUSH_NR];
+	unsigned int i, n;
+
+	if (c->journal.btree_flushing)
+		return;
+
+	spin_lock(&c->journal.flush_write_lock);
+	if (c->journal.btree_flushing) {
+		spin_unlock(&c->journal.flush_write_lock);
+		return;
+	}
+	c->journal.btree_flushing = true;
+	spin_unlock(&c->journal.flush_write_lock);
 
 	atomic_long_inc(&c->flush_write);
+	memset(btree_nodes, 0, sizeof(btree_nodes));
+	n = 0;
 
-retry:
-	spin_lock(&c->journal.lock);
-	if (heap_empty(&c->flush_btree)) {
-		for_each_cached_btree(b, c, i)
-			if (btree_current_write(b)->journal) {
-				if (!heap_full(&c->flush_btree))
-					heap_add(&c->flush_btree, b,
-						 journal_max_cmp);
-				else if (journal_max_cmp(b,
-					 heap_peek(&c->flush_btree))) {
-					c->flush_btree.data[0] = b;
-					heap_sift(&c->flush_btree, 0,
-						  journal_max_cmp);
-				}
-			}
+	mutex_lock(&c->bucket_lock);
+	list_for_each_entry_safe_reverse(b, t, &c->btree_cache, list) {
+		if (btree_node_journal_flush(b))
+			pr_err("BUG: flush_write bit should not be set here!");
+
+		mutex_lock(&b->write_lock);
 
-		for (i = c->flush_btree.used / 2 - 1; i >= 0; --i)
-			heap_sift(&c->flush_btree, i, journal_min_cmp);
+		if (!btree_node_dirty(b)) {
+			mutex_unlock(&b->write_lock);
+			continue;
+		}
+
+		if (!btree_current_write(b)->journal) {
+			mutex_unlock(&b->write_lock);
+			continue;
+		}
+
+		set_btree_node_journal_flush(b);
+
+		mutex_unlock(&b->write_lock);
+
+		btree_nodes[n++] = b;
+		if (n == BTREE_FLUSH_NR)
+			break;
 	}
+	mutex_unlock(&c->bucket_lock);
 
-	b = NULL;
-	heap_pop(&c->flush_btree, b, journal_min_cmp);
-	spin_unlock(&c->journal.lock);
+	for (i = 0; i < n; i++) {
+		b = btree_nodes[i];
+		if (!b) {
+			pr_err("BUG: btree_nodes[%d] is NULL", i);
+			continue;
+		}
+
+		/* safe to check without holding b->write_lock */
+		if (!btree_node_journal_flush(b)) {
+			pr_err("BUG: bnode %p: journal_flush bit cleaned", b);
+			continue;
+		}
 
-	if (b) {
 		mutex_lock(&b->write_lock);
 		if (!btree_current_write(b)->journal) {
+			clear_bit(BTREE_NODE_journal_flush, &b->flags);
+			mutex_unlock(&b->write_lock);
+			pr_debug("bnode %p: written by others", b);
+			continue;
+		}
+
+		if (!btree_node_dirty(b)) {
+			clear_bit(BTREE_NODE_journal_flush, &b->flags);
 			mutex_unlock(&b->write_lock);
-			/* We raced */
-			atomic_long_inc(&c->retry_flush_write);
-			goto retry;
+			pr_debug("bnode %p: dirty bit cleaned by others", b);
+			continue;
 		}
 
 		__bch_btree_node_write(b, NULL);
+		clear_bit(BTREE_NODE_journal_flush, &b->flags);
 		mutex_unlock(&b->write_lock);
 	}
+
+	spin_lock(&c->journal.flush_write_lock);
+	c->journal.btree_flushing = false;
+	spin_unlock(&c->journal.flush_write_lock);
 }
 
 #define last_seq(j)	((j)->seq - fifo_used(&(j)->pin) + 1)
@@ -559,6 +614,7 @@ static void journal_reclaim(struct cache_set *c)
 		k->ptr[n++] = MAKE_PTR(0,
 				  bucket_to_sector(c, ca->sb.d[ja->cur_idx]),
 				  ca->sb.nr_this_dev);
+		atomic_long_inc(&c->reclaimed_journal_buckets);
 	}
 
 	if (n) {
@@ -811,6 +867,10 @@ atomic_t *bch_journal(struct cache_set *c,
 	struct journal_write *w;
 	atomic_t *ret;
 
+	/* No journaling if CACHE_SET_IO_DISABLE set already */
+	if (unlikely(test_bit(CACHE_SET_IO_DISABLE, &c->flags)))
+		return NULL;
+
 	if (!CACHE_SYNC(&c->sb))
 		return NULL;
 
@@ -855,7 +915,6 @@ void bch_journal_free(struct cache_set *c)
 	free_pages((unsigned long) c->journal.w[1].data, JSET_BITS);
 	free_pages((unsigned long) c->journal.w[0].data, JSET_BITS);
 	free_fifo(&c->journal.pin);
-	free_heap(&c->flush_btree);
 }
 
 int bch_journal_alloc(struct cache_set *c)
@@ -863,6 +922,7 @@ int bch_journal_alloc(struct cache_set *c)
 	struct journal *j = &c->journal;
 
 	spin_lock_init(&j->lock);
+	spin_lock_init(&j->flush_write_lock);
 	INIT_DELAYED_WORK(&j->work, journal_write_work);
 
 	c->journal_delay_ms = 100;
@@ -870,8 +930,7 @@ int bch_journal_alloc(struct cache_set *c)
 	j->w[0].c = c;
 	j->w[1].c = c;
 
-	if (!(init_heap(&c->flush_btree, 128, GFP_KERNEL)) ||
-	    !(init_fifo(&j->pin, JOURNAL_PIN, GFP_KERNEL)) ||
+	if (!(init_fifo(&j->pin, JOURNAL_PIN, GFP_KERNEL)) ||
 	    !(j->w[0].data = (void *) __get_free_pages(GFP_KERNEL, JSET_BITS)) ||
 	    !(j->w[1].data = (void *) __get_free_pages(GFP_KERNEL, JSET_BITS)))
 		return -ENOMEM;
diff --git a/drivers/md/bcache/journal.h b/drivers/md/bcache/journal.h
index 66f0facff84b..f2ea34d5f431 100644
--- a/drivers/md/bcache/journal.h
+++ b/drivers/md/bcache/journal.h
@@ -103,6 +103,8 @@ struct journal_write {
 /* Embedded in struct cache_set */
 struct journal {
 	spinlock_t		lock;
+	spinlock_t		flush_write_lock;
+	bool			btree_flushing;
 	/* used when waiting because the journal was full */
 	struct closure_waitlist	wait;
 	struct closure		io;
@@ -154,6 +156,8 @@ struct journal_device {
 	struct bio_vec		bv[8];
 };
 
+#define BTREE_FLUSH_NR	8
+
 #define journal_pin_cmp(c, l, r)				\
 	(fifo_idx(&(c)->journal.pin, (l)) > fifo_idx(&(c)->journal.pin, (r)))
 
diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 1b63ac876169..26e374fbf57c 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -40,6 +40,7 @@ static const char invalid_uuid[] = {
 
 static struct kobject *bcache_kobj;
 struct mutex bch_register_lock;
+bool bcache_is_reboot;
 LIST_HEAD(bch_cache_sets);
 static LIST_HEAD(uncached_devices);
 
@@ -49,6 +50,7 @@ static wait_queue_head_t unregister_wait;
 struct workqueue_struct *bcache_wq;
 struct workqueue_struct *bch_journal_wq;
 
+
 #define BTREE_MAX_PAGES		(256 * 1024 / PAGE_SIZE)
 /* limitation of partitions number on single bcache device */
 #define BCACHE_MINORS		128
@@ -197,7 +199,9 @@ err:
 static void write_bdev_super_endio(struct bio *bio)
 {
 	struct cached_dev *dc = bio->bi_private;
-	/* XXX: error checking */
+
+	if (bio->bi_status)
+		bch_count_backing_io_errors(dc, bio);
 
 	closure_put(&dc->sb_write);
 }
@@ -691,6 +695,7 @@ static void bcache_device_link(struct bcache_device *d, struct cache_set *c,
 {
 	unsigned int i;
 	struct cache *ca;
+	int ret;
 
 	for_each_cache(ca, d->c, i)
 		bd_link_disk_holder(ca->bdev, d->disk);
@@ -698,9 +703,13 @@ static void bcache_device_link(struct bcache_device *d, struct cache_set *c,
 	snprintf(d->name, BCACHEDEVNAME_SIZE,
 		 "%s%u", name, d->id);
 
-	WARN(sysfs_create_link(&d->kobj, &c->kobj, "cache") ||
-	     sysfs_create_link(&c->kobj, &d->kobj, d->name),
-	     "Couldn't create device <-> cache set symlinks");
+	ret = sysfs_create_link(&d->kobj, &c->kobj, "cache");
+	if (ret < 0)
+		pr_err("Couldn't create device -> cache set symlink");
+
+	ret = sysfs_create_link(&c->kobj, &d->kobj, d->name);
+	if (ret < 0)
+		pr_err("Couldn't create cache set -> device symlink");
 
 	clear_bit(BCACHE_DEV_UNLINK_DONE, &d->flags);
 }
@@ -908,7 +917,7 @@ static int cached_dev_status_update(void *arg)
 }
 
 
-void bch_cached_dev_run(struct cached_dev *dc)
+int bch_cached_dev_run(struct cached_dev *dc)
 {
 	struct bcache_device *d = &dc->disk;
 	char *buf = kmemdup_nul(dc->sb.label, SB_LABEL_SIZE, GFP_KERNEL);
@@ -919,11 +928,19 @@ void bch_cached_dev_run(struct cached_dev *dc)
 		NULL,
 	};
 
+	if (dc->io_disable) {
+		pr_err("I/O disabled on cached dev %s",
+		       dc->backing_dev_name);
+		return -EIO;
+	}
+
 	if (atomic_xchg(&dc->running, 1)) {
 		kfree(env[1]);
 		kfree(env[2]);
 		kfree(buf);
-		return;
+		pr_info("cached dev %s is running already",
+		       dc->backing_dev_name);
+		return -EBUSY;
 	}
 
 	if (!d->c &&
@@ -949,8 +966,11 @@ void bch_cached_dev_run(struct cached_dev *dc)
 	kfree(buf);
 
 	if (sysfs_create_link(&d->kobj, &disk_to_dev(d->disk)->kobj, "dev") ||
-	    sysfs_create_link(&disk_to_dev(d->disk)->kobj, &d->kobj, "bcache"))
-		pr_debug("error creating sysfs link");
+	    sysfs_create_link(&disk_to_dev(d->disk)->kobj,
+			      &d->kobj, "bcache")) {
+		pr_err("Couldn't create bcache dev <-> disk sysfs symlinks");
+		return -ENOMEM;
+	}
 
 	dc->status_update_thread = kthread_run(cached_dev_status_update,
 					       dc, "bcache_status_update");
@@ -959,6 +979,8 @@ void bch_cached_dev_run(struct cached_dev *dc)
 			"continue to run without monitoring backing "
 			"device status");
 	}
+
+	return 0;
 }
 
 /*
@@ -996,7 +1018,6 @@ static void cached_dev_detach_finish(struct work_struct *w)
 	BUG_ON(!test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags));
 	BUG_ON(refcount_read(&dc->count));
 
-	mutex_lock(&bch_register_lock);
 
 	if (test_and_clear_bit(BCACHE_DEV_WB_RUNNING, &dc->disk.flags))
 		cancel_writeback_rate_update_dwork(dc);
@@ -1012,6 +1033,8 @@ static void cached_dev_detach_finish(struct work_struct *w)
 	bch_write_bdev_super(dc, &cl);
 	closure_sync(&cl);
 
+	mutex_lock(&bch_register_lock);
+
 	calc_cached_dev_sectors(dc->disk.c);
 	bcache_device_detach(&dc->disk);
 	list_move(&dc->list, &uncached_devices);
@@ -1054,6 +1077,7 @@ int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c,
 	uint32_t rtime = cpu_to_le32((u32)ktime_get_real_seconds());
 	struct uuid_entry *u;
 	struct cached_dev *exist_dc, *t;
+	int ret = 0;
 
 	if ((set_uuid && memcmp(set_uuid, c->sb.set_uuid, 16)) ||
 	    (!set_uuid && memcmp(dc->sb.set_uuid, c->sb.set_uuid, 16)))
@@ -1153,6 +1177,8 @@ int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c,
 	down_write(&dc->writeback_lock);
 	if (bch_cached_dev_writeback_start(dc)) {
 		up_write(&dc->writeback_lock);
+		pr_err("Couldn't start writeback facilities for %s",
+		       dc->disk.disk->disk_name);
 		return -ENOMEM;
 	}
 
@@ -1163,7 +1189,22 @@ int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c,
 
 	bch_sectors_dirty_init(&dc->disk);
 
-	bch_cached_dev_run(dc);
+	ret = bch_cached_dev_run(dc);
+	if (ret && (ret != -EBUSY)) {
+		up_write(&dc->writeback_lock);
+		/*
+		 * bch_register_lock is held, bcache_device_stop() is not
+		 * able to be directly called. The kthread and kworker
+		 * created previously in bch_cached_dev_writeback_start()
+		 * have to be stopped manually here.
+		 */
+		kthread_stop(dc->writeback_thread);
+		cancel_writeback_rate_update_dwork(dc);
+		pr_err("Couldn't run cached device %s",
+		       dc->backing_dev_name);
+		return ret;
+	}
+
 	bcache_device_link(&dc->disk, c, "bdev");
 	atomic_inc(&c->attached_dev_nr);
 
@@ -1190,18 +1231,16 @@ static void cached_dev_free(struct closure *cl)
 {
 	struct cached_dev *dc = container_of(cl, struct cached_dev, disk.cl);
 
-	mutex_lock(&bch_register_lock);
-
 	if (test_and_clear_bit(BCACHE_DEV_WB_RUNNING, &dc->disk.flags))
 		cancel_writeback_rate_update_dwork(dc);
 
 	if (!IS_ERR_OR_NULL(dc->writeback_thread))
 		kthread_stop(dc->writeback_thread);
-	if (dc->writeback_write_wq)
-		destroy_workqueue(dc->writeback_write_wq);
 	if (!IS_ERR_OR_NULL(dc->status_update_thread))
 		kthread_stop(dc->status_update_thread);
 
+	mutex_lock(&bch_register_lock);
+
 	if (atomic_read(&dc->running))
 		bd_unlink_disk_holder(dc->bdev, dc->disk.disk);
 	bcache_device_free(&dc->disk);
@@ -1290,6 +1329,7 @@ static int register_bdev(struct cache_sb *sb, struct page *sb_page,
 {
 	const char *err = "cannot allocate memory";
 	struct cache_set *c;
+	int ret = -ENOMEM;
 
 	bdevname(bdev, dc->backing_dev_name);
 	memcpy(&dc->sb, sb, sizeof(struct cache_sb));
@@ -1319,14 +1359,18 @@ static int register_bdev(struct cache_sb *sb, struct page *sb_page,
 		bch_cached_dev_attach(dc, c, NULL);
 
 	if (BDEV_STATE(&dc->sb) == BDEV_STATE_NONE ||
-	    BDEV_STATE(&dc->sb) == BDEV_STATE_STALE)
-		bch_cached_dev_run(dc);
+	    BDEV_STATE(&dc->sb) == BDEV_STATE_STALE) {
+		err = "failed to run cached device";
+		ret = bch_cached_dev_run(dc);
+		if (ret)
+			goto err;
+	}
 
 	return 0;
 err:
 	pr_notice("error %s: %s", dc->backing_dev_name, err);
 	bcache_device_stop(&dc->disk);
-	return -EIO;
+	return ret;
 }
 
 /* Flash only volumes */
@@ -1437,8 +1481,6 @@ int bch_flash_dev_create(struct cache_set *c, uint64_t size)
 
 bool bch_cached_dev_error(struct cached_dev *dc)
 {
-	struct cache_set *c;
-
 	if (!dc || test_bit(BCACHE_DEV_CLOSING, &dc->disk.flags))
 		return false;
 
@@ -1449,21 +1491,6 @@ bool bch_cached_dev_error(struct cached_dev *dc)
 	pr_err("stop %s: too many IO errors on backing device %s\n",
 		dc->disk.disk->disk_name, dc->backing_dev_name);
 
-	/*
-	 * If the cached device is still attached to a cache set,
-	 * even dc->io_disable is true and no more I/O requests
-	 * accepted, cache device internal I/O (writeback scan or
-	 * garbage collection) may still prevent bcache device from
-	 * being stopped. So here CACHE_SET_IO_DISABLE should be
-	 * set to c->flags too, to make the internal I/O to cache
-	 * device rejected and stopped immediately.
-	 * If c is NULL, that means the bcache device is not attached
-	 * to any cache set, then no CACHE_SET_IO_DISABLE bit to set.
-	 */
-	c = dc->disk.c;
-	if (c && test_and_set_bit(CACHE_SET_IO_DISABLE, &c->flags))
-		pr_info("CACHE_SET_IO_DISABLE already set");
-
 	bcache_device_stop(&dc->disk);
 	return true;
 }
@@ -1564,19 +1591,23 @@ static void cache_set_flush(struct closure *cl)
 	kobject_put(&c->internal);
 	kobject_del(&c->kobj);
 
-	if (c->gc_thread)
+	if (!IS_ERR_OR_NULL(c->gc_thread))
 		kthread_stop(c->gc_thread);
 
 	if (!IS_ERR_OR_NULL(c->root))
 		list_add(&c->root->list, &c->btree_cache);
 
-	/* Should skip this if we're unregistering because of an error */
-	list_for_each_entry(b, &c->btree_cache, list) {
-		mutex_lock(&b->write_lock);
-		if (btree_node_dirty(b))
-			__bch_btree_node_write(b, NULL);
-		mutex_unlock(&b->write_lock);
-	}
+	/*
+	 * Avoid flushing cached nodes if cache set is retiring
+	 * due to too many I/O errors detected.
+	 */
+	if (!test_bit(CACHE_SET_IO_DISABLE, &c->flags))
+		list_for_each_entry(b, &c->btree_cache, list) {
+			mutex_lock(&b->write_lock);
+			if (btree_node_dirty(b))
+				__bch_btree_node_write(b, NULL);
+			mutex_unlock(&b->write_lock);
+		}
 
 	for_each_cache(ca, c, i)
 		if (ca->alloc_thread)
@@ -1849,6 +1880,23 @@ static int run_cache_set(struct cache_set *c)
 		if (bch_btree_check(c))
 			goto err;
 
+		/*
+		 * bch_btree_check() may occupy too much system memory which
+		 * has negative effects to user space application (e.g. data
+		 * base) performance. Shrink the mca cache memory proactively
+		 * here to avoid competing memory with user space workloads..
+		 */
+		if (!c->shrinker_disabled) {
+			struct shrink_control sc;
+
+			sc.gfp_mask = GFP_KERNEL;
+			sc.nr_to_scan = c->btree_cache_used * c->btree_pages;
+			/* first run to clear b->accessed tag */
+			c->shrink.scan_objects(&c->shrink, &sc);
+			/* second run to reap non-accessed nodes */
+			c->shrink.scan_objects(&c->shrink, &sc);
+		}
+
 		bch_journal_mark(c, &journal);
 		bch_initial_gc_finish(c);
 		pr_debug("btree_check() done");
@@ -1957,7 +2005,7 @@ err:
 	}
 
 	closure_sync(&cl);
-	/* XXX: test this, it's broken */
+
 	bch_cache_set_error(c, "%s", err);
 
 	return -EIO;
@@ -2251,9 +2299,13 @@ err:
 
 static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 			       const char *buffer, size_t size);
+static ssize_t bch_pending_bdevs_cleanup(struct kobject *k,
+					 struct kobj_attribute *attr,
+					 const char *buffer, size_t size);
 
 kobj_attribute_write(register,		register_bcache);
 kobj_attribute_write(register_quiet,	register_bcache);
+kobj_attribute_write(pendings_cleanup,	bch_pending_bdevs_cleanup);
 
 static bool bch_is_open_backing(struct block_device *bdev)
 {
@@ -2301,6 +2353,11 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 	if (!try_module_get(THIS_MODULE))
 		return -EBUSY;
 
+	/* For latest state of bcache_is_reboot */
+	smp_mb();
+	if (bcache_is_reboot)
+		return -EBUSY;
+
 	path = kstrndup(buffer, size, GFP_KERNEL);
 	if (!path)
 		goto err;
@@ -2378,8 +2435,61 @@ err:
 	goto out;
 }
 
+
+struct pdev {
+	struct list_head list;
+	struct cached_dev *dc;
+};
+
+static ssize_t bch_pending_bdevs_cleanup(struct kobject *k,
+					 struct kobj_attribute *attr,
+					 const char *buffer,
+					 size_t size)
+{
+	LIST_HEAD(pending_devs);
+	ssize_t ret = size;
+	struct cached_dev *dc, *tdc;
+	struct pdev *pdev, *tpdev;
+	struct cache_set *c, *tc;
+
+	mutex_lock(&bch_register_lock);
+	list_for_each_entry_safe(dc, tdc, &uncached_devices, list) {
+		pdev = kmalloc(sizeof(struct pdev), GFP_KERNEL);
+		if (!pdev)
+			break;
+		pdev->dc = dc;
+		list_add(&pdev->list, &pending_devs);
+	}
+
+	list_for_each_entry_safe(pdev, tpdev, &pending_devs, list) {
+		list_for_each_entry_safe(c, tc, &bch_cache_sets, list) {
+			char *pdev_set_uuid = pdev->dc->sb.set_uuid;
+			char *set_uuid = c->sb.uuid;
+
+			if (!memcmp(pdev_set_uuid, set_uuid, 16)) {
+				list_del(&pdev->list);
+				kfree(pdev);
+				break;
+			}
+		}
+	}
+	mutex_unlock(&bch_register_lock);
+
+	list_for_each_entry_safe(pdev, tpdev, &pending_devs, list) {
+		pr_info("delete pdev %p", pdev);
+		list_del(&pdev->list);
+		bcache_device_stop(&pdev->dc->disk);
+		kfree(pdev);
+	}
+
+	return ret;
+}
+
 static int bcache_reboot(struct notifier_block *n, unsigned long code, void *x)
 {
+	if (bcache_is_reboot)
+		return NOTIFY_DONE;
+
 	if (code == SYS_DOWN ||
 	    code == SYS_HALT ||
 	    code == SYS_POWER_OFF) {
@@ -2392,19 +2502,45 @@ static int bcache_reboot(struct notifier_block *n, unsigned long code, void *x)
 
 		mutex_lock(&bch_register_lock);
 
+		if (bcache_is_reboot)
+			goto out;
+
+		/* New registration is rejected since now */
+		bcache_is_reboot = true;
+		/*
+		 * Make registering caller (if there is) on other CPU
+		 * core know bcache_is_reboot set to true earlier
+		 */
+		smp_mb();
+
 		if (list_empty(&bch_cache_sets) &&
 		    list_empty(&uncached_devices))
 			goto out;
 
+		mutex_unlock(&bch_register_lock);
+
 		pr_info("Stopping all devices:");
 
+		/*
+		 * The reason bch_register_lock is not held to call
+		 * bch_cache_set_stop() and bcache_device_stop() is to
+		 * avoid potential deadlock during reboot, because cache
+		 * set or bcache device stopping process will acqurie
+		 * bch_register_lock too.
+		 *
+		 * We are safe here because bcache_is_reboot sets to
+		 * true already, register_bcache() will reject new
+		 * registration now. bcache_is_reboot also makes sure
+		 * bcache_reboot() won't be re-entered on by other thread,
+		 * so there is no race in following list iteration by
+		 * list_for_each_entry_safe().
+		 */
 		list_for_each_entry_safe(c, tc, &bch_cache_sets, list)
 			bch_cache_set_stop(c);
 
 		list_for_each_entry_safe(dc, tdc, &uncached_devices, list)
 			bcache_device_stop(&dc->disk);
 
-		mutex_unlock(&bch_register_lock);
 
 		/*
 		 * Give an early chance for other kthreads and
@@ -2496,6 +2632,7 @@ static int __init bcache_init(void)
 	static const struct attribute *files[] = {
 		&ksysfs_register.attr,
 		&ksysfs_register_quiet.attr,
+		&ksysfs_pendings_cleanup.attr,
 		NULL
 	};
 
@@ -2531,6 +2668,8 @@ static int __init bcache_init(void)
 	bch_debug_init();
 	closure_debug_init();
 
+	bcache_is_reboot = false;
+
 	return 0;
 err:
 	bcache_exit();
diff --git a/drivers/md/bcache/sysfs.c b/drivers/md/bcache/sysfs.c
index bfb437ffb13c..9f0826712845 100644
--- a/drivers/md/bcache/sysfs.c
+++ b/drivers/md/bcache/sysfs.c
@@ -16,33 +16,31 @@
 #include <linux/sort.h>
 #include <linux/sched/clock.h>
 
+extern bool bcache_is_reboot;
+
 /* Default is 0 ("writethrough") */
 static const char * const bch_cache_modes[] = {
 	"writethrough",
 	"writeback",
 	"writearound",
-	"none",
-	NULL
+	"none"
 };
 
 /* Default is 0 ("auto") */
 static const char * const bch_stop_on_failure_modes[] = {
 	"auto",
-	"always",
-	NULL
+	"always"
 };
 
 static const char * const cache_replacement_policies[] = {
 	"lru",
 	"fifo",
-	"random",
-	NULL
+	"random"
 };
 
 static const char * const error_actions[] = {
 	"unregister",
-	"panic",
-	NULL
+	"panic"
 };
 
 write_attribute(attach);
@@ -84,8 +82,8 @@ read_attribute(bset_tree_stats);
 read_attribute(state);
 read_attribute(cache_read_races);
 read_attribute(reclaim);
+read_attribute(reclaimed_journal_buckets);
 read_attribute(flush_write);
-read_attribute(retry_flush_write);
 read_attribute(writeback_keys_done);
 read_attribute(writeback_keys_failed);
 read_attribute(io_errors);
@@ -180,7 +178,7 @@ SHOW(__bch_cached_dev)
 	var_print(writeback_percent);
 	sysfs_hprint(writeback_rate,
 		     wb ? atomic_long_read(&dc->writeback_rate.rate) << 9 : 0);
-	sysfs_hprint(io_errors,		atomic_read(&dc->io_errors));
+	sysfs_printf(io_errors,		"%i", atomic_read(&dc->io_errors));
 	sysfs_printf(io_error_limit,	"%i", dc->error_limit);
 	sysfs_printf(io_disable,	"%i", dc->io_disable);
 	var_print(writeback_rate_update_seconds);
@@ -271,6 +269,10 @@ STORE(__cached_dev)
 	struct cache_set *c;
 	struct kobj_uevent_env *env;
 
+	/* no user space access if system is rebooting */
+	if (bcache_is_reboot)
+		return -EBUSY;
+
 #define d_strtoul(var)		sysfs_strtoul(var, dc->var)
 #define d_strtoul_nonzero(var)	sysfs_strtoul_clamp(var, dc->var, 1, INT_MAX)
 #define d_strtoi_h(var)		sysfs_hatoi(var, dc->var)
@@ -329,11 +331,14 @@ STORE(__cached_dev)
 		bch_cache_accounting_clear(&dc->accounting);
 
 	if (attr == &sysfs_running &&
-	    strtoul_or_return(buf))
-		bch_cached_dev_run(dc);
+	    strtoul_or_return(buf)) {
+		v = bch_cached_dev_run(dc);
+		if (v)
+			return v;
+	}
 
 	if (attr == &sysfs_cache_mode) {
-		v = __sysfs_match_string(bch_cache_modes, -1, buf);
+		v = sysfs_match_string(bch_cache_modes, buf);
 		if (v < 0)
 			return v;
 
@@ -344,7 +349,7 @@ STORE(__cached_dev)
 	}
 
 	if (attr == &sysfs_stop_when_cache_set_failed) {
-		v = __sysfs_match_string(bch_stop_on_failure_modes, -1, buf);
+		v = sysfs_match_string(bch_stop_on_failure_modes, buf);
 		if (v < 0)
 			return v;
 
@@ -408,6 +413,10 @@ STORE(bch_cached_dev)
 	struct cached_dev *dc = container_of(kobj, struct cached_dev,
 					     disk.kobj);
 
+	/* no user space access if system is rebooting */
+	if (bcache_is_reboot)
+		return -EBUSY;
+
 	mutex_lock(&bch_register_lock);
 	size = __cached_dev_store(kobj, attr, buf, size);
 
@@ -464,7 +473,7 @@ static struct attribute *bch_cached_dev_files[] = {
 	&sysfs_writeback_rate_p_term_inverse,
 	&sysfs_writeback_rate_minimum,
 	&sysfs_writeback_rate_debug,
-	&sysfs_errors,
+	&sysfs_io_errors,
 	&sysfs_io_error_limit,
 	&sysfs_io_disable,
 	&sysfs_dirty_data,
@@ -511,6 +520,10 @@ STORE(__bch_flash_dev)
 					       kobj);
 	struct uuid_entry *u = &d->c->uuids[d->id];
 
+	/* no user space access if system is rebooting */
+	if (bcache_is_reboot)
+		return -EBUSY;
+
 	sysfs_strtoul(data_csum,	d->data_csum);
 
 	if (attr == &sysfs_size) {
@@ -693,12 +706,12 @@ SHOW(__bch_cache_set)
 	sysfs_print(reclaim,
 		    atomic_long_read(&c->reclaim));
 
+	sysfs_print(reclaimed_journal_buckets,
+		    atomic_long_read(&c->reclaimed_journal_buckets));
+
 	sysfs_print(flush_write,
 		    atomic_long_read(&c->flush_write));
 
-	sysfs_print(retry_flush_write,
-		    atomic_long_read(&c->retry_flush_write));
-
 	sysfs_print(writeback_keys_done,
 		    atomic_long_read(&c->writeback_keys_done));
 	sysfs_print(writeback_keys_failed,
@@ -746,6 +759,10 @@ STORE(__bch_cache_set)
 	struct cache_set *c = container_of(kobj, struct cache_set, kobj);
 	ssize_t v;
 
+	/* no user space access if system is rebooting */
+	if (bcache_is_reboot)
+		return -EBUSY;
+
 	if (attr == &sysfs_unregister)
 		bch_cache_set_unregister(c);
 
@@ -799,7 +816,7 @@ STORE(__bch_cache_set)
 			    0, UINT_MAX);
 
 	if (attr == &sysfs_errors) {
-		v = __sysfs_match_string(error_actions, -1, buf);
+		v = sysfs_match_string(error_actions, buf);
 		if (v < 0)
 			return v;
 
@@ -865,6 +882,10 @@ STORE(bch_cache_set_internal)
 {
 	struct cache_set *c = container_of(kobj, struct cache_set, internal);
 
+	/* no user space access if system is rebooting */
+	if (bcache_is_reboot)
+		return -EBUSY;
+
 	return bch_cache_set_store(&c->kobj, attr, buf, size);
 }
 
@@ -914,8 +935,8 @@ static struct attribute *bch_cache_set_internal_files[] = {
 	&sysfs_bset_tree_stats,
 	&sysfs_cache_read_races,
 	&sysfs_reclaim,
+	&sysfs_reclaimed_journal_buckets,
 	&sysfs_flush_write,
-	&sysfs_retry_flush_write,
 	&sysfs_writeback_keys_done,
 	&sysfs_writeback_keys_failed,
 
@@ -1050,6 +1071,10 @@ STORE(__bch_cache)
 	struct cache *ca = container_of(kobj, struct cache, kobj);
 	ssize_t v;
 
+	/* no user space access if system is rebooting */
+	if (bcache_is_reboot)
+		return -EBUSY;
+
 	if (attr == &sysfs_discard) {
 		bool v = strtoul_or_return(buf);
 
@@ -1063,7 +1088,7 @@ STORE(__bch_cache)
 	}
 
 	if (attr == &sysfs_cache_replacement_policy) {
-		v = __sysfs_match_string(cache_replacement_policies, -1, buf);
+		v = sysfs_match_string(cache_replacement_policies, buf);
 		if (v < 0)
 			return v;
 
diff --git a/drivers/md/bcache/util.h b/drivers/md/bcache/util.h
index 1fbced94e4cc..c029f7443190 100644
--- a/drivers/md/bcache/util.h
+++ b/drivers/md/bcache/util.h
@@ -113,8 +113,6 @@ do {									\
 
 #define heap_full(h)	((h)->used == (h)->size)
 
-#define heap_empty(h)	((h)->used == 0)
-
 #define DECLARE_FIFO(type, name)					\
 	struct {							\
 		size_t front, back, size, mask;				\
diff --git a/drivers/md/bcache/writeback.c b/drivers/md/bcache/writeback.c
index 73f0efac2b9f..d60268fe49e1 100644
--- a/drivers/md/bcache/writeback.c
+++ b/drivers/md/bcache/writeback.c
@@ -122,6 +122,9 @@ static void __update_writeback_rate(struct cached_dev *dc)
 static bool set_at_max_writeback_rate(struct cache_set *c,
 				       struct cached_dev *dc)
 {
+	/* Don't set max writeback rate if gc is running */
+	if (!c->gc_mark_valid)
+		return false;
 	/*
 	 * Idle_counter is increased everytime when update_writeback_rate() is
 	 * called. If all backing devices attached to the same cache set have
@@ -735,6 +738,10 @@ static int bch_writeback_thread(void *arg)
 		}
 	}
 
+	if (dc->writeback_write_wq) {
+		flush_workqueue(dc->writeback_write_wq);
+		destroy_workqueue(dc->writeback_write_wq);
+	}
 	cached_dev_put(dc);
 	wait_for_kthread_stop();
 
@@ -830,6 +837,7 @@ int bch_cached_dev_writeback_start(struct cached_dev *dc)
 					      "bcache_writeback");
 	if (IS_ERR(dc->writeback_thread)) {
 		cached_dev_put(dc);
+		destroy_workqueue(dc->writeback_write_wq);
 		return PTR_ERR(dc->writeback_thread);
 	}
 	dc->writeback_running = true;
diff --git a/drivers/md/md-bitmap.c b/drivers/md/md-bitmap.c
index c01d41198f5e..b092c7b5282f 100644
--- a/drivers/md/md-bitmap.c
+++ b/drivers/md/md-bitmap.c
@@ -1790,6 +1790,8 @@ void md_bitmap_destroy(struct mddev *mddev)
 		return;
 
 	md_bitmap_wait_behind_writes(mddev);
+	mempool_destroy(mddev->wb_info_pool);
+	mddev->wb_info_pool = NULL;
 
 	mutex_lock(&mddev->bitmap_info.mutex);
 	spin_lock(&mddev->lock);
@@ -1900,10 +1902,14 @@ int md_bitmap_load(struct mddev *mddev)
 	sector_t start = 0;
 	sector_t sector = 0;
 	struct bitmap *bitmap = mddev->bitmap;
+	struct md_rdev *rdev;
 
 	if (!bitmap)
 		goto out;
 
+	rdev_for_each(rdev, mddev)
+		mddev_create_wb_pool(mddev, rdev, true);
+
 	if (mddev_is_clustered(mddev))
 		md_cluster_ops->load_bitmaps(mddev, mddev->bitmap_info.nodes);
 
@@ -2462,12 +2468,26 @@ static ssize_t
 backlog_store(struct mddev *mddev, const char *buf, size_t len)
 {
 	unsigned long backlog;
+	unsigned long old_mwb = mddev->bitmap_info.max_write_behind;
 	int rv = kstrtoul(buf, 10, &backlog);
 	if (rv)
 		return rv;
 	if (backlog > COUNTER_MAX)
 		return -EINVAL;
 	mddev->bitmap_info.max_write_behind = backlog;
+	if (!backlog && mddev->wb_info_pool) {
+		/* wb_info_pool is not needed if backlog is zero */
+		mempool_destroy(mddev->wb_info_pool);
+		mddev->wb_info_pool = NULL;
+	} else if (backlog && !mddev->wb_info_pool) {
+		/* wb_info_pool is needed since backlog is not zero */
+		struct md_rdev *rdev;
+
+		rdev_for_each(rdev, mddev)
+			mddev_create_wb_pool(mddev, rdev, false);
+	}
+	if (old_mwb != backlog)
+		md_bitmap_update_sb(mddev->bitmap);
 	return len;
 }
 
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 9801d540fea1..a114b05e3db4 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -37,6 +37,7 @@
 
 */
 
+#include <linux/sched/mm.h>
 #include <linux/sched/signal.h>
 #include <linux/kthread.h>
 #include <linux/blkdev.h>
@@ -124,6 +125,77 @@ static inline int speed_max(struct mddev *mddev)
 		mddev->sync_speed_max : sysctl_speed_limit_max;
 }
 
+static int rdev_init_wb(struct md_rdev *rdev)
+{
+	if (rdev->bdev->bd_queue->nr_hw_queues == 1)
+		return 0;
+
+	spin_lock_init(&rdev->wb_list_lock);
+	INIT_LIST_HEAD(&rdev->wb_list);
+	init_waitqueue_head(&rdev->wb_io_wait);
+	set_bit(WBCollisionCheck, &rdev->flags);
+
+	return 1;
+}
+
+/*
+ * Create wb_info_pool if rdev is the first multi-queue device flaged
+ * with writemostly, also write-behind mode is enabled.
+ */
+void mddev_create_wb_pool(struct mddev *mddev, struct md_rdev *rdev,
+			  bool is_suspend)
+{
+	if (mddev->bitmap_info.max_write_behind == 0)
+		return;
+
+	if (!test_bit(WriteMostly, &rdev->flags) || !rdev_init_wb(rdev))
+		return;
+
+	if (mddev->wb_info_pool == NULL) {
+		unsigned int noio_flag;
+
+		if (!is_suspend)
+			mddev_suspend(mddev);
+		noio_flag = memalloc_noio_save();
+		mddev->wb_info_pool = mempool_create_kmalloc_pool(NR_WB_INFOS,
+							sizeof(struct wb_info));
+		memalloc_noio_restore(noio_flag);
+		if (!mddev->wb_info_pool)
+			pr_err("can't alloc memory pool for writemostly\n");
+		if (!is_suspend)
+			mddev_resume(mddev);
+	}
+}
+EXPORT_SYMBOL_GPL(mddev_create_wb_pool);
+
+/*
+ * destroy wb_info_pool if rdev is the last device flaged with WBCollisionCheck.
+ */
+static void mddev_destroy_wb_pool(struct mddev *mddev, struct md_rdev *rdev)
+{
+	if (!test_and_clear_bit(WBCollisionCheck, &rdev->flags))
+		return;
+
+	if (mddev->wb_info_pool) {
+		struct md_rdev *temp;
+		int num = 0;
+
+		/*
+		 * Check if other rdevs need wb_info_pool.
+		 */
+		rdev_for_each(temp, mddev)
+			if (temp != rdev &&
+			    test_bit(WBCollisionCheck, &temp->flags))
+				num++;
+		if (!num) {
+			mddev_suspend(rdev->mddev);
+			mempool_destroy(mddev->wb_info_pool);
+			mddev->wb_info_pool = NULL;
+			mddev_resume(rdev->mddev);
+		}
+	}
+}
+
 static struct ctl_table_header *raid_table_header;
 
 static struct ctl_table raid_table[] = {
@@ -2210,6 +2282,9 @@ static int bind_rdev_to_array(struct md_rdev *rdev, struct mddev *mddev)
 	rdev->mddev = mddev;
 	pr_debug("md: bind<%s>\n", b);
 
+	if (mddev->raid_disks)
+		mddev_create_wb_pool(mddev, rdev, false);
+
 	if ((err = kobject_add(&rdev->kobj, &mddev->kobj, "dev-%s", b)))
 		goto fail;
 
@@ -2246,6 +2321,7 @@ static void unbind_rdev_from_array(struct md_rdev *rdev)
 	bd_unlink_disk_holder(rdev->bdev, rdev->mddev->gendisk);
 	list_del_rcu(&rdev->same_set);
 	pr_debug("md: unbind<%s>\n", bdevname(rdev->bdev,b));
+	mddev_destroy_wb_pool(rdev->mddev, rdev);
 	rdev->mddev = NULL;
 	sysfs_remove_link(&rdev->kobj, "block");
 	sysfs_put(rdev->sysfs_state);
@@ -2758,8 +2834,10 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len)
 		}
 	} else if (cmd_match(buf, "writemostly")) {
 		set_bit(WriteMostly, &rdev->flags);
+		mddev_create_wb_pool(rdev->mddev, rdev, false);
 		err = 0;
 	} else if (cmd_match(buf, "-writemostly")) {
+		mddev_destroy_wb_pool(rdev->mddev, rdev);
 		clear_bit(WriteMostly, &rdev->flags);
 		err = 0;
 	} else if (cmd_match(buf, "blocked")) {
@@ -3356,7 +3434,7 @@ rdev_attr_show(struct kobject *kobj, struct attribute *attr, char *page)
 	if (!entry->show)
 		return -EIO;
 	if (!rdev->mddev)
-		return -EBUSY;
+		return -ENODEV;
 	return entry->show(rdev, page);
 }
 
@@ -5588,15 +5666,28 @@ int md_run(struct mddev *mddev)
 			mddev->bitmap = bitmap;
 
 	}
-	if (err) {
-		mddev_detach(mddev);
-		if (mddev->private)
-			pers->free(mddev, mddev->private);
-		mddev->private = NULL;
-		module_put(pers->owner);
-		md_bitmap_destroy(mddev);
-		goto abort;
+	if (err)
+		goto bitmap_abort;
+
+	if (mddev->bitmap_info.max_write_behind > 0) {
+		bool creat_pool = false;
+
+		rdev_for_each(rdev, mddev) {
+			if (test_bit(WriteMostly, &rdev->flags) &&
+			    rdev_init_wb(rdev))
+				creat_pool = true;
+		}
+		if (creat_pool && mddev->wb_info_pool == NULL) {
+			mddev->wb_info_pool =
+				mempool_create_kmalloc_pool(NR_WB_INFOS,
+						    sizeof(struct wb_info));
+			if (!mddev->wb_info_pool) {
+				err = -ENOMEM;
+				goto bitmap_abort;
+			}
+		}
 	}
+
 	if (mddev->queue) {
 		bool nonrot = true;
 
@@ -5639,8 +5730,7 @@ int md_run(struct mddev *mddev)
 	spin_unlock(&mddev->lock);
 	rdev_for_each(rdev, mddev)
 		if (rdev->raid_disk >= 0)
-			if (sysfs_link_rdev(mddev, rdev))
-				/* failure here is OK */;
+			sysfs_link_rdev(mddev, rdev); /* failure here is OK */
 
 	if (mddev->degraded && !mddev->ro)
 		/* This ensures that recovering status is reported immediately
@@ -5658,6 +5748,13 @@ int md_run(struct mddev *mddev)
 	sysfs_notify(&mddev->kobj, NULL, "degraded");
 	return 0;
 
+bitmap_abort:
+	mddev_detach(mddev);
+	if (mddev->private)
+		pers->free(mddev, mddev->private);
+	mddev->private = NULL;
+	module_put(pers->owner);
+	md_bitmap_destroy(mddev);
 abort:
 	bioset_exit(&mddev->bio_set);
 	bioset_exit(&mddev->sync_set);
@@ -5826,6 +5923,8 @@ static void __md_stop_writes(struct mddev *mddev)
 			mddev->in_sync = 1;
 		md_update_sb(mddev, 1);
 	}
+	mempool_destroy(mddev->wb_info_pool);
+	mddev->wb_info_pool = NULL;
 }
 
 void md_stop_writes(struct mddev *mddev)
@@ -8198,8 +8297,7 @@ void md_do_sync(struct md_thread *thread)
 {
 	struct mddev *mddev = thread->mddev;
 	struct mddev *mddev2;
-	unsigned int currspeed = 0,
-		 window;
+	unsigned int currspeed = 0, window;
 	sector_t max_sectors,j, io_sectors, recovery_done;
 	unsigned long mark[SYNC_MARKS];
 	unsigned long update_time;
@@ -8256,7 +8354,7 @@ void md_do_sync(struct md_thread *thread)
 	 * 0 == not engaged in resync at all
 	 * 2 == checking that there is no conflict with another sync
 	 * 1 == like 2, but have yielded to allow conflicting resync to
-	 *		commense
+	 *		commence
 	 * other == active in resync - this many blocks
 	 *
 	 * Before starting a resync we must have set curr_resync to
@@ -8387,7 +8485,7 @@ void md_do_sync(struct md_thread *thread)
 	/*
 	 * Tune reconstruction:
 	 */
-	window = 32*(PAGE_SIZE/512);
+	window = 32 * (PAGE_SIZE / 512);
 	pr_debug("md: using %dk window, over a total of %lluk.\n",
 		 window/2, (unsigned long long)max_sectors/2);
 
@@ -9200,7 +9298,6 @@ static void check_sb_changes(struct mddev *mddev, struct md_rdev *rdev)
 				 * perform resync with the new activated disk */
 				set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
 				md_wakeup_thread(mddev->thread);
-
 			}
 			/* device faulty
 			 * We just want to do the minimum to mark the disk
diff --git a/drivers/md/md.h b/drivers/md/md.h
index 7c930c091193..10f98200e2f8 100644
--- a/drivers/md/md.h
+++ b/drivers/md/md.h
@@ -109,6 +109,14 @@ struct md_rdev {
 					   * for reporting to userspace and storing
 					   * in superblock.
 					   */
+
+	/*
+	 * The members for check collision of write behind IOs.
+	 */
+	struct list_head wb_list;
+	spinlock_t wb_list_lock;
+	wait_queue_head_t wb_io_wait;
+
 	struct work_struct del_work;	/* used for delayed sysfs removal */
 
 	struct kernfs_node *sysfs_state; /* handle for 'state'
@@ -193,6 +201,10 @@ enum flag_bits {
 				 * it didn't fail, so don't use FailFast
 				 * any more for metadata
 				 */
+	WBCollisionCheck,	/*
+				 * multiqueue device should check if there
+				 * is collision between write behind bios.
+				 */
 };
 
 static inline int is_badblock(struct md_rdev *rdev, sector_t s, int sectors,
@@ -245,6 +257,14 @@ enum mddev_sb_flags {
 	MD_SB_NEED_REWRITE,	/* metadata write needs to be repeated */
 };
 
+#define NR_WB_INFOS	8
+/* record current range of write behind IOs */
+struct wb_info {
+	sector_t lo;
+	sector_t hi;
+	struct list_head list;
+};
+
 struct mddev {
 	void				*private;
 	struct md_personality		*pers;
@@ -461,6 +481,7 @@ struct mddev {
 					  */
 	struct work_struct flush_work;
 	struct work_struct event_work;	/* used by dm to report failure event */
+	mempool_t *wb_info_pool;
 	void (*sync_super)(struct mddev *mddev, struct md_rdev *rdev);
 	struct md_cluster_info		*cluster_info;
 	unsigned int			good_device_nr;	/* good device num within cluster raid */
@@ -709,6 +730,8 @@ extern struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
 extern void md_reload_sb(struct mddev *mddev, int raid_disk);
 extern void md_update_sb(struct mddev *mddev, int force);
 extern void md_kick_rdev_from_array(struct md_rdev * rdev);
+extern void mddev_create_wb_pool(struct mddev *mddev, struct md_rdev *rdev,
+				 bool is_suspend);
 struct md_rdev *md_find_rdev_nr_rcu(struct mddev *mddev, int nr);
 struct md_rdev *md_find_rdev_rcu(struct mddev *mddev, dev_t dev);
 
diff --git a/drivers/md/raid1-10.c b/drivers/md/raid1-10.c
index 400001b815db..54db34163968 100644
--- a/drivers/md/raid1-10.c
+++ b/drivers/md/raid1-10.c
@@ -3,12 +3,42 @@
 #define RESYNC_BLOCK_SIZE (64*1024)
 #define RESYNC_PAGES ((RESYNC_BLOCK_SIZE + PAGE_SIZE-1) / PAGE_SIZE)
 
+/*
+ * Number of guaranteed raid bios in case of extreme VM load:
+ */
+#define	NR_RAID_BIOS 256
+
+/* when we get a read error on a read-only array, we redirect to another
+ * device without failing the first device, or trying to over-write to
+ * correct the read error.  To keep track of bad blocks on a per-bio
+ * level, we store IO_BLOCKED in the appropriate 'bios' pointer
+ */
+#define IO_BLOCKED ((struct bio *)1)
+/* When we successfully write to a known bad-block, we need to remove the
+ * bad-block marking which must be done from process context.  So we record
+ * the success by setting devs[n].bio to IO_MADE_GOOD
+ */
+#define IO_MADE_GOOD ((struct bio *)2)
+
+#define BIO_SPECIAL(bio) ((unsigned long)bio <= 2)
+
+/* When there are this many requests queue to be written by
+ * the raid thread, we become 'congested' to provide back-pressure
+ * for writeback.
+ */
+static int max_queued_requests = 1024;
+
 /* for managing resync I/O pages */
 struct resync_pages {
 	void		*raid_bio;
 	struct page	*pages[RESYNC_PAGES];
 };
 
+static void rbio_pool_free(void *rbio, void *data)
+{
+	kfree(rbio);
+}
+
 static inline int resync_alloc_pages(struct resync_pages *rp,
 				     gfp_t gfp_flags)
 {
diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index 2aa36e570e04..34e26834ad28 100644
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -42,31 +42,6 @@
 	 (1L << MD_HAS_PPL) |		\
 	 (1L << MD_HAS_MULTIPLE_PPLS))
 
-/*
- * Number of guaranteed r1bios in case of extreme VM load:
- */
-#define	NR_RAID1_BIOS 256
-
-/* when we get a read error on a read-only array, we redirect to another
- * device without failing the first device, or trying to over-write to
- * correct the read error.  To keep track of bad blocks on a per-bio
- * level, we store IO_BLOCKED in the appropriate 'bios' pointer
- */
-#define IO_BLOCKED ((struct bio *)1)
-/* When we successfully write to a known bad-block, we need to remove the
- * bad-block marking which must be done from process context.  So we record
- * the success by setting devs[n].bio to IO_MADE_GOOD
- */
-#define IO_MADE_GOOD ((struct bio *)2)
-
-#define BIO_SPECIAL(bio) ((unsigned long)bio <= 2)
-
-/* When there are this many requests queue to be written by
- * the raid1 thread, we become 'congested' to provide back-pressure
- * for writeback.
- */
-static int max_queued_requests = 1024;
-
 static void allow_barrier(struct r1conf *conf, sector_t sector_nr);
 static void lower_barrier(struct r1conf *conf, sector_t sector_nr);
 
@@ -75,6 +50,57 @@ static void lower_barrier(struct r1conf *conf, sector_t sector_nr);
 
 #include "raid1-10.c"
 
+static int check_and_add_wb(struct md_rdev *rdev, sector_t lo, sector_t hi)
+{
+	struct wb_info *wi, *temp_wi;
+	unsigned long flags;
+	int ret = 0;
+	struct mddev *mddev = rdev->mddev;
+
+	wi = mempool_alloc(mddev->wb_info_pool, GFP_NOIO);
+
+	spin_lock_irqsave(&rdev->wb_list_lock, flags);
+	list_for_each_entry(temp_wi, &rdev->wb_list, list) {
+		/* collision happened */
+		if (hi > temp_wi->lo && lo < temp_wi->hi) {
+			ret = -EBUSY;
+			break;
+		}
+	}
+
+	if (!ret) {
+		wi->lo = lo;
+		wi->hi = hi;
+		list_add(&wi->list, &rdev->wb_list);
+	} else
+		mempool_free(wi, mddev->wb_info_pool);
+	spin_unlock_irqrestore(&rdev->wb_list_lock, flags);
+
+	return ret;
+}
+
+static void remove_wb(struct md_rdev *rdev, sector_t lo, sector_t hi)
+{
+	struct wb_info *wi;
+	unsigned long flags;
+	int found = 0;
+	struct mddev *mddev = rdev->mddev;
+
+	spin_lock_irqsave(&rdev->wb_list_lock, flags);
+	list_for_each_entry(wi, &rdev->wb_list, list)
+		if (hi == wi->hi && lo == wi->lo) {
+			list_del(&wi->list);
+			mempool_free(wi, mddev->wb_info_pool);
+			found = 1;
+			break;
+		}
+
+	if (!found)
+		WARN(1, "The write behind IO is not recorded\n");
+	spin_unlock_irqrestore(&rdev->wb_list_lock, flags);
+	wake_up(&rdev->wb_io_wait);
+}
+
 /*
  * for resync bio, r1bio pointer can be retrieved from the per-bio
  * 'struct resync_pages'.
@@ -93,11 +119,6 @@ static void * r1bio_pool_alloc(gfp_t gfp_flags, void *data)
 	return kzalloc(size, gfp_flags);
 }
 
-static void r1bio_pool_free(void *r1_bio, void *data)
-{
-	kfree(r1_bio);
-}
-
 #define RESYNC_DEPTH 32
 #define RESYNC_SECTORS (RESYNC_BLOCK_SIZE >> 9)
 #define RESYNC_WINDOW (RESYNC_BLOCK_SIZE * RESYNC_DEPTH)
@@ -173,7 +194,7 @@ out_free_bio:
 	kfree(rps);
 
 out_free_r1bio:
-	r1bio_pool_free(r1_bio, data);
+	rbio_pool_free(r1_bio, data);
 	return NULL;
 }
 
@@ -193,7 +214,7 @@ static void r1buf_pool_free(void *__r1_bio, void *data)
 	/* resync pages array stored in the 1st bio's .bi_private */
 	kfree(rp);
 
-	r1bio_pool_free(r1bio, data);
+	rbio_pool_free(r1bio, data);
 }
 
 static void put_all_bios(struct r1conf *conf, struct r1bio *r1_bio)
@@ -476,6 +497,12 @@ static void raid1_end_write_request(struct bio *bio)
 	}
 
 	if (behind) {
+		if (test_bit(WBCollisionCheck, &rdev->flags)) {
+			sector_t lo = r1_bio->sector;
+			sector_t hi = r1_bio->sector + r1_bio->sectors;
+
+			remove_wb(rdev, lo, hi);
+		}
 		if (test_bit(WriteMostly, &rdev->flags))
 			atomic_dec(&r1_bio->behind_remaining);
 
@@ -1449,7 +1476,6 @@ static void raid1_write_request(struct mddev *mddev, struct bio *bio,
 		if (!r1_bio->bios[i])
 			continue;
 
-
 		if (first_clone) {
 			/* do behind I/O ?
 			 * Not if there are too many, or cannot
@@ -1474,7 +1500,16 @@ static void raid1_write_request(struct mddev *mddev, struct bio *bio,
 			mbio = bio_clone_fast(bio, GFP_NOIO, &mddev->bio_set);
 
 		if (r1_bio->behind_master_bio) {
-			if (test_bit(WriteMostly, &conf->mirrors[i].rdev->flags))
+			struct md_rdev *rdev = conf->mirrors[i].rdev;
+
+			if (test_bit(WBCollisionCheck, &rdev->flags)) {
+				sector_t lo = r1_bio->sector;
+				sector_t hi = r1_bio->sector + r1_bio->sectors;
+
+				wait_event(rdev->wb_io_wait,
+					   check_and_add_wb(rdev, lo, hi) == 0);
+			}
+			if (test_bit(WriteMostly, &rdev->flags))
 				atomic_inc(&r1_bio->behind_remaining);
 		}
 
@@ -1729,9 +1764,8 @@ static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev)
 		first = last = rdev->saved_raid_disk;
 
 	for (mirror = first; mirror <= last; mirror++) {
-		p = conf->mirrors+mirror;
+		p = conf->mirrors + mirror;
 		if (!p->rdev) {
-
 			if (mddev->gendisk)
 				disk_stack_limits(mddev->gendisk, rdev->bdev,
 						  rdev->data_offset << 9);
@@ -2888,7 +2922,6 @@ static sector_t raid1_sync_request(struct mddev *mddev, sector_t sector_nr,
 		if (read_targets == 1)
 			bio->bi_opf &= ~MD_FAILFAST;
 		generic_make_request(bio);
-
 	}
 	return nr_sectors;
 }
@@ -2947,8 +2980,8 @@ static struct r1conf *setup_conf(struct mddev *mddev)
 	if (!conf->poolinfo)
 		goto abort;
 	conf->poolinfo->raid_disks = mddev->raid_disks * 2;
-	err = mempool_init(&conf->r1bio_pool, NR_RAID1_BIOS, r1bio_pool_alloc,
-			   r1bio_pool_free, conf->poolinfo);
+	err = mempool_init(&conf->r1bio_pool, NR_RAID_BIOS, r1bio_pool_alloc,
+			   rbio_pool_free, conf->poolinfo);
 	if (err)
 		goto abort;
 
@@ -3089,7 +3122,7 @@ static int raid1_run(struct mddev *mddev)
 	}
 
 	mddev->degraded = 0;
-	for (i=0; i < conf->raid_disks; i++)
+	for (i = 0; i < conf->raid_disks; i++)
 		if (conf->mirrors[i].rdev == NULL ||
 		    !test_bit(In_sync, &conf->mirrors[i].rdev->flags) ||
 		    test_bit(Faulty, &conf->mirrors[i].rdev->flags))
@@ -3124,7 +3157,7 @@ static int raid1_run(struct mddev *mddev)
 						  mddev->queue);
 	}
 
-	ret =  md_integrity_register(mddev);
+	ret = md_integrity_register(mddev);
 	if (ret) {
 		md_unregister_thread(&mddev->thread);
 		raid1_free(mddev, conf);
@@ -3232,8 +3265,8 @@ static int raid1_reshape(struct mddev *mddev)
 	newpoolinfo->mddev = mddev;
 	newpoolinfo->raid_disks = raid_disks * 2;
 
-	ret = mempool_init(&newpool, NR_RAID1_BIOS, r1bio_pool_alloc,
-			   r1bio_pool_free, newpoolinfo);
+	ret = mempool_init(&newpool, NR_RAID_BIOS, r1bio_pool_alloc,
+			   rbio_pool_free, newpoolinfo);
 	if (ret) {
 		kfree(newpoolinfo);
 		return ret;
diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index aea11476fee6..8a1354a08a1a 100644
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -64,31 +64,6 @@
  *    [B A] [D C]    [B A] [E C D]
  */
 
-/*
- * Number of guaranteed r10bios in case of extreme VM load:
- */
-#define	NR_RAID10_BIOS 256
-
-/* when we get a read error on a read-only array, we redirect to another
- * device without failing the first device, or trying to over-write to
- * correct the read error.  To keep track of bad blocks on a per-bio
- * level, we store IO_BLOCKED in the appropriate 'bios' pointer
- */
-#define IO_BLOCKED ((struct bio *)1)
-/* When we successfully write to a known bad-block, we need to remove the
- * bad-block marking which must be done from process context.  So we record
- * the success by setting devs[n].bio to IO_MADE_GOOD
- */
-#define IO_MADE_GOOD ((struct bio *)2)
-
-#define BIO_SPECIAL(bio) ((unsigned long)bio <= 2)
-
-/* When there are this many requests queued to be written by
- * the raid10 thread, we become 'congested' to provide back-pressure
- * for writeback.
- */
-static int max_queued_requests = 1024;
-
 static void allow_barrier(struct r10conf *conf);
 static void lower_barrier(struct r10conf *conf);
 static int _enough(struct r10conf *conf, int previous, int ignore);
@@ -123,11 +98,6 @@ static void * r10bio_pool_alloc(gfp_t gfp_flags, void *data)
 	return kzalloc(size, gfp_flags);
 }
 
-static void r10bio_pool_free(void *r10_bio, void *data)
-{
-	kfree(r10_bio);
-}
-
 #define RESYNC_SECTORS (RESYNC_BLOCK_SIZE >> 9)
 /* amount of memory to reserve for resync requests */
 #define RESYNC_WINDOW (1024*1024)
@@ -233,7 +203,7 @@ out_free_bio:
 	}
 	kfree(rps);
 out_free_r10bio:
-	r10bio_pool_free(r10_bio, conf);
+	rbio_pool_free(r10_bio, conf);
 	return NULL;
 }
 
@@ -261,7 +231,7 @@ static void r10buf_pool_free(void *__r10_bio, void *data)
 	/* resync pages array stored in the 1st bio's .bi_private */
 	kfree(rp);
 
-	r10bio_pool_free(r10bio, conf);
+	rbio_pool_free(r10bio, conf);
 }
 
 static void put_all_bios(struct r10conf *conf, struct r10bio *r10_bio)
@@ -737,15 +707,19 @@ static struct md_rdev *read_balance(struct r10conf *conf,
 	int sectors = r10_bio->sectors;
 	int best_good_sectors;
 	sector_t new_distance, best_dist;
-	struct md_rdev *best_rdev, *rdev = NULL;
+	struct md_rdev *best_dist_rdev, *best_pending_rdev, *rdev = NULL;
 	int do_balance;
-	int best_slot;
+	int best_dist_slot, best_pending_slot;
+	bool has_nonrot_disk = false;
+	unsigned int min_pending;
 	struct geom *geo = &conf->geo;
 
 	raid10_find_phys(conf, r10_bio);
 	rcu_read_lock();
-	best_slot = -1;
-	best_rdev = NULL;
+	best_dist_slot = -1;
+	min_pending = UINT_MAX;
+	best_dist_rdev = NULL;
+	best_pending_rdev = NULL;
 	best_dist = MaxSector;
 	best_good_sectors = 0;
 	do_balance = 1;
@@ -767,6 +741,8 @@ static struct md_rdev *read_balance(struct r10conf *conf,
 		sector_t first_bad;
 		int bad_sectors;
 		sector_t dev_sector;
+		unsigned int pending;
+		bool nonrot;
 
 		if (r10_bio->devs[slot].bio == IO_BLOCKED)
 			continue;
@@ -803,8 +779,8 @@ static struct md_rdev *read_balance(struct r10conf *conf,
 					first_bad - dev_sector;
 				if (good_sectors > best_good_sectors) {
 					best_good_sectors = good_sectors;
-					best_slot = slot;
-					best_rdev = rdev;
+					best_dist_slot = slot;
+					best_dist_rdev = rdev;
 				}
 				if (!do_balance)
 					/* Must read from here */
@@ -817,14 +793,23 @@ static struct md_rdev *read_balance(struct r10conf *conf,
 		if (!do_balance)
 			break;
 
-		if (best_slot >= 0)
+		nonrot = blk_queue_nonrot(bdev_get_queue(rdev->bdev));
+		has_nonrot_disk |= nonrot;
+		pending = atomic_read(&rdev->nr_pending);
+		if (min_pending > pending && nonrot) {
+			min_pending = pending;
+			best_pending_slot = slot;
+			best_pending_rdev = rdev;
+		}
+
+		if (best_dist_slot >= 0)
 			/* At least 2 disks to choose from so failfast is OK */
 			set_bit(R10BIO_FailFast, &r10_bio->state);
 		/* This optimisation is debatable, and completely destroys
 		 * sequential read speed for 'far copies' arrays.  So only
 		 * keep it for 'near' arrays, and review those later.
 		 */
-		if (geo->near_copies > 1 && !atomic_read(&rdev->nr_pending))
+		if (geo->near_copies > 1 && !pending)
 			new_distance = 0;
 
 		/* for far > 1 always use the lowest address */
@@ -833,15 +818,21 @@ static struct md_rdev *read_balance(struct r10conf *conf,
 		else
 			new_distance = abs(r10_bio->devs[slot].addr -
 					   conf->mirrors[disk].head_position);
+
 		if (new_distance < best_dist) {
 			best_dist = new_distance;
-			best_slot = slot;
-			best_rdev = rdev;
+			best_dist_slot = slot;
+			best_dist_rdev = rdev;
 		}
 	}
 	if (slot >= conf->copies) {
-		slot = best_slot;
-		rdev = best_rdev;
+		if (has_nonrot_disk) {
+			slot = best_pending_slot;
+			rdev = best_pending_rdev;
+		} else {
+			slot = best_dist_slot;
+			rdev = best_dist_rdev;
+		}
 	}
 
 	if (slot >= 0) {
@@ -3675,8 +3666,8 @@ static struct r10conf *setup_conf(struct mddev *mddev)
 
 	conf->geo = geo;
 	conf->copies = copies;
-	err = mempool_init(&conf->r10bio_pool, NR_RAID10_BIOS, r10bio_pool_alloc,
-			   r10bio_pool_free, conf);
+	err = mempool_init(&conf->r10bio_pool, NR_RAID_BIOS, r10bio_pool_alloc,
+			   rbio_pool_free, conf);
 	if (err)
 		goto out;
 
@@ -4780,8 +4771,7 @@ static int handle_reshape_read_error(struct mddev *mddev,
 	int idx = 0;
 	struct page **pages;
 
-	r10b = kmalloc(sizeof(*r10b) +
-	       sizeof(struct r10dev) * conf->copies, GFP_NOIO);
+	r10b = kmalloc(struct_size(r10b, devs, conf->copies), GFP_NOIO);
 	if (!r10b) {
 		set_bit(MD_RECOVERY_INTR, &mddev->recovery);
 		return -ENOMEM;
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index b83bce2beb66..3de4e13bde98 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -5251,7 +5251,6 @@ static int raid5_read_one_chunk(struct mddev *mddev, struct bio *raid_bio)
 		rcu_read_unlock();
 		raid_bio->bi_next = (void*)rdev;
 		bio_set_dev(align_bi, rdev->bdev);
-		bio_clear_flag(align_bi, BIO_SEG_VALID);
 
 		if (is_badblock(rdev, align_bi->bi_iter.bi_sector,
 				bio_sectors(align_bi),
@@ -7672,7 +7671,7 @@ abort:
 static int raid5_add_disk(struct mddev *mddev, struct md_rdev *rdev)
 {
 	struct r5conf *conf = mddev->private;
-	int err = -EEXIST;
+	int ret, err = -EEXIST;
 	int disk;
 	struct disk_info *p;
 	int first = 0;
@@ -7687,7 +7686,14 @@ static int raid5_add_disk(struct mddev *mddev, struct md_rdev *rdev)
 		 * The array is in readonly mode if journal is missing, so no
 		 * write requests running. We should be safe
 		 */
-		log_init(conf, rdev, false);
+		ret = log_init(conf, rdev, false);
+		if (ret)
+			return ret;
+
+		ret = r5l_start(conf->log);
+		if (ret)
+			return ret;
+
 		return 0;
 	}
 	if (mddev->recovery_disabled == conf->recovery_disabled)