6 files changed, 38 insertions, 9 deletions

diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
index b6a74bcbb08f..2a7f0dd6abab 100644
--- a/drivers/md/bcache/request.c
+++ b/drivers/md/bcache/request.c
@@ -1000,7 +1000,7 @@ static void request_write(struct cached_dev *dc, struct search *s)
 
 		if (bio->bi_rw & REQ_FLUSH) {
 			/* Also need to send a flush to the backing device */
-			struct bio *flush = bio_alloc_bioset(0, GFP_NOIO,
+			struct bio *flush = bio_alloc_bioset(GFP_NOIO, 0,
 							     dc->disk.bio_split);
 
 			flush->bi_rw	= WRITE_FLUSH;
diff --git a/drivers/md/dm-snap-persistent.c b/drivers/md/dm-snap-persistent.c
index 4caa8e6d59d7..2d2b1b7588d7 100644
--- a/drivers/md/dm-snap-persistent.c
+++ b/drivers/md/dm-snap-persistent.c
@@ -269,6 +269,14 @@ static chunk_t area_location(struct pstore *ps, chunk_t area)
 	return NUM_SNAPSHOT_HDR_CHUNKS + ((ps->exceptions_per_area + 1) * area);
 }
 
+static void skip_metadata(struct pstore *ps)
+{
+	uint32_t stride = ps->exceptions_per_area + 1;
+	chunk_t next_free = ps->next_free;
+	if (sector_div(next_free, stride) == NUM_SNAPSHOT_HDR_CHUNKS)
+		ps->next_free++;
+}
+
 /*
  * Read or write a metadata area.  Remembering to skip the first
  * chunk which holds the header.
@@ -502,6 +510,8 @@ static int read_exceptions(struct pstore *ps,
 
 	ps->current_area--;
 
+	skip_metadata(ps);
+
 	return 0;
 }
 
@@ -616,8 +626,6 @@ static int persistent_prepare_exception(struct dm_exception_store *store,
 					struct dm_exception *e)
 {
 	struct pstore *ps = get_info(store);
-	uint32_t stride;
-	chunk_t next_free;
 	sector_t size = get_dev_size(dm_snap_cow(store->snap)->bdev);
 
 	/* Is there enough room ? */
@@ -630,10 +638,8 @@ static int persistent_prepare_exception(struct dm_exception_store *store,
 	 * Move onto the next free pending, making sure to take
 	 * into account the location of the metadata chunks.
 	 */
-	stride = (ps->exceptions_per_area + 1);
-	next_free = ++ps->next_free;
-	if (sector_div(next_free, stride) == 1)
-		ps->next_free++;
+	ps->next_free++;
+	skip_metadata(ps);
 
 	atomic_inc(&ps->pending_count);
 	return 0;
diff --git a/drivers/md/md.c b/drivers/md/md.c
index adf4d7e1d5e1..561a65f82e26 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -8111,6 +8111,7 @@ static int md_set_badblocks(struct badblocks *bb, sector_t s, int sectors,
 	u64 *p;
 	int lo, hi;
 	int rv = 1;
+	unsigned long flags;
 
 	if (bb->shift < 0)
 		/* badblocks are disabled */
@@ -8125,7 +8126,7 @@ static int md_set_badblocks(struct badblocks *bb, sector_t s, int sectors,
 		sectors = next - s;
 	}
 
-	write_seqlock_irq(&bb->lock);
+	write_seqlock_irqsave(&bb->lock, flags);
 
 	p = bb->page;
 	lo = 0;
@@ -8241,7 +8242,7 @@ static int md_set_badblocks(struct badblocks *bb, sector_t s, int sectors,
 	bb->changed = 1;
 	if (!acknowledged)
 		bb->unacked_exist = 1;
-	write_sequnlock_irq(&bb->lock);
+	write_sequnlock_irqrestore(&bb->lock, flags);
 
 	return rv;
 }
diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index d60412c7f995..aacf6bf352d8 100644
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -1479,6 +1479,7 @@ static int raid1_spare_active(struct mddev *mddev)
 			}
 		}
 		if (rdev
+		    && rdev->recovery_offset == MaxSector
 		    && !test_bit(Faulty, &rdev->flags)
 		    && !test_and_set_bit(In_sync, &rdev->flags)) {
 			count++;
diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index df7b0a06b0ea..73dc8a377522 100644
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -1782,6 +1782,7 @@ static int raid10_spare_active(struct mddev *mddev)
 			}
 			sysfs_notify_dirent_safe(tmp->replacement->sysfs_state);
 		} else if (tmp->rdev
+			   && tmp->rdev->recovery_offset == MaxSector
 			   && !test_bit(Faulty, &tmp->rdev->flags)
 			   && !test_and_set_bit(In_sync, &tmp->rdev->flags)) {
 			count++;
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 7ff4f252ca1a..f8b906843926 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -778,6 +778,12 @@ static void ops_run_io(struct stripe_head *sh, struct stripe_head_state *s)
 			bi->bi_io_vec[0].bv_len = STRIPE_SIZE;
 			bi->bi_io_vec[0].bv_offset = 0;
 			bi->bi_size = STRIPE_SIZE;
+			/*
+			 * If this is discard request, set bi_vcnt 0. We don't
+			 * want to confuse SCSI because SCSI will replace payload
+			 */
+			if (rw & REQ_DISCARD)
+				bi->bi_vcnt = 0;
 			if (rrdev)
 				set_bit(R5_DOUBLE_LOCKED, &sh->dev[i].flags);
 
@@ -816,6 +822,12 @@ static void ops_run_io(struct stripe_head *sh, struct stripe_head_state *s)
 			rbi->bi_io_vec[0].bv_len = STRIPE_SIZE;
 			rbi->bi_io_vec[0].bv_offset = 0;
 			rbi->bi_size = STRIPE_SIZE;
+			/*
+			 * If this is discard request, set bi_vcnt 0. We don't
+			 * want to confuse SCSI because SCSI will replace payload
+			 */
+			if (rw & REQ_DISCARD)
+				rbi->bi_vcnt = 0;
 			if (conf->mddev->gendisk)
 				trace_block_bio_remap(bdev_get_queue(rbi->bi_bdev),
 						      rbi, disk_devt(conf->mddev->gendisk),
@@ -2910,6 +2922,14 @@ static void handle_stripe_clean_event(struct r5conf *conf,
 		}
 		/* now that discard is done we can proceed with any sync */
 		clear_bit(STRIPE_DISCARD, &sh->state);
+		/*
+		 * SCSI discard will change some bio fields and the stripe has
+		 * no updated data, so remove it from hash list and the stripe
+		 * will be reinitialized
+		 */
+		spin_lock_irq(&conf->device_lock);
+		remove_hash(sh);
+		spin_unlock_irq(&conf->device_lock);
 		if (test_bit(STRIPE_SYNC_REQUESTED, &sh->state))
 			set_bit(STRIPE_HANDLE, &sh->state);