Merge branch 'md-next' into md-linus

1 files changed, 364 insertions, 337 deletions

diff --git a/drivers/md/md.c b/drivers/md/md.c
index f975cd08923d..82821ee0d57f 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -30,6 +30,18 @@
    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.
+
+   Errors, Warnings, etc.
+   Please use:
+     pr_crit() for error conditions that risk data loss
+     pr_err() for error conditions that are unexpected, like an IO error
+         or internal inconsistency
+     pr_warn() for error conditions that could have been predicated, like
+         adding a device to an array when it has incompatible metadata
+     pr_info() for every interesting, very rare events, like an array starting
+         or stopping, or resync starting or stopping
+     pr_debug() for everything else.
+
 */
 
 #include <linux/kthread.h>
@@ -52,6 +64,7 @@
 #include <linux/raid/md_p.h>
 #include <linux/raid/md_u.h>
 #include <linux/slab.h>
+#include <trace/events/block.h>
 #include "md.h"
 #include "bitmap.h"
 #include "md-cluster.h"
@@ -684,11 +697,8 @@ static inline sector_t calc_dev_sboffset(struct md_rdev *rdev)
 static int alloc_disk_sb(struct md_rdev *rdev)
 {
 	rdev->sb_page = alloc_page(GFP_KERNEL);
-	if (!rdev->sb_page) {
-		printk(KERN_ALERT "md: out of memory.\n");
+	if (!rdev->sb_page)
 		return -ENOMEM;
-	}
-
 	return 0;
 }
 
@@ -715,9 +725,15 @@ static void super_written(struct bio *bio)
 	struct mddev *mddev = rdev->mddev;
 
 	if (bio->bi_error) {
-		printk("md: super_written gets error=%d\n", bio->bi_error);
+		pr_err("md: super_written gets error=%d\n", bio->bi_error);
 		md_error(mddev, rdev);
-	}
+		if (!test_bit(Faulty, &rdev->flags)
+		    && (bio->bi_opf & MD_FAILFAST)) {
+			set_bit(MD_SB_NEED_REWRITE, &mddev->sb_flags);
+			set_bit(LastDev, &rdev->flags);
+		}
+	} else
+		clear_bit(LastDev, &rdev->flags);
 
 	if (atomic_dec_and_test(&mddev->pending_writes))
 		wake_up(&mddev->sb_wait);
@@ -734,7 +750,13 @@ void md_super_write(struct mddev *mddev, struct md_rdev *rdev,
 	 * if zero is reached.
 	 * If an error occurred, call md_error
 	 */
-	struct bio *bio = bio_alloc_mddev(GFP_NOIO, 1, mddev);
+	struct bio *bio;
+	int ff = 0;
+
+	if (test_bit(Faulty, &rdev->flags))
+		return;
+
+	bio = bio_alloc_mddev(GFP_NOIO, 1, mddev);
 
 	atomic_inc(&rdev->nr_pending);
 
@@ -743,16 +765,24 @@ void md_super_write(struct mddev *mddev, struct md_rdev *rdev,
 	bio_add_page(bio, page, size, 0);
 	bio->bi_private = rdev;
 	bio->bi_end_io = super_written;
-	bio->bi_opf = REQ_OP_WRITE | REQ_PREFLUSH | REQ_FUA;
+
+	if (test_bit(MD_FAILFAST_SUPPORTED, &mddev->flags) &&
+	    test_bit(FailFast, &rdev->flags) &&
+	    !test_bit(LastDev, &rdev->flags))
+		ff = MD_FAILFAST;
+	bio->bi_opf = REQ_OP_WRITE | REQ_PREFLUSH | REQ_FUA | ff;
 
 	atomic_inc(&mddev->pending_writes);
 	submit_bio(bio);
 }
 
-void md_super_wait(struct mddev *mddev)
+int md_super_wait(struct mddev *mddev)
 {
 	/* wait for all superblock writes that were scheduled to complete */
 	wait_event(mddev->sb_wait, atomic_read(&mddev->pending_writes)==0);
+	if (test_and_clear_bit(MD_SB_NEED_REWRITE, &mddev->sb_flags))
+		return -EAGAIN;
+	return 0;
 }
 
 int sync_page_io(struct md_rdev *rdev, sector_t sector, int size,
@@ -795,8 +825,8 @@ static int read_disk_sb(struct md_rdev *rdev, int size)
 	return 0;
 
 fail:
-	printk(KERN_WARNING "md: disabled device %s, could not read superblock.\n",
-		bdevname(rdev->bdev,b));
+	pr_err("md: disabled device %s, could not read superblock.\n",
+	       bdevname(rdev->bdev,b));
 	return -EINVAL;
 }
 
@@ -818,7 +848,6 @@ static int sb_equal(mdp_super_t *sb1, mdp_super_t *sb2)
 
 	if (!tmp1 || !tmp2) {
 		ret = 0;
-		printk(KERN_INFO "md.c sb_equal(): failed to allocate memory!\n");
 		goto abort;
 	}
 
@@ -932,7 +961,7 @@ int md_check_no_bitmap(struct mddev *mddev)
 {
 	if (!mddev->bitmap_info.file && !mddev->bitmap_info.offset)
 		return 0;
-	printk(KERN_ERR "%s: bitmaps are not supported for %s\n",
+	pr_warn("%s: bitmaps are not supported for %s\n",
 		mdname(mddev), mddev->pers->name);
 	return 1;
 }
@@ -956,7 +985,8 @@ static int super_90_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor
 	rdev->sb_start = calc_dev_sboffset(rdev);
 
 	ret = read_disk_sb(rdev, MD_SB_BYTES);
-	if (ret) return ret;
+	if (ret)
+		return ret;
 
 	ret = -EINVAL;
 
@@ -964,17 +994,15 @@ static int super_90_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor
 	sb = page_address(rdev->sb_page);
 
 	if (sb->md_magic != MD_SB_MAGIC) {
-		printk(KERN_ERR "md: invalid raid superblock magic on %s\n",
-		       b);
+		pr_warn("md: invalid raid superblock magic on %s\n", b);
 		goto abort;
 	}
 
 	if (sb->major_version != 0 ||
 	    sb->minor_version < 90 ||
 	    sb->minor_version > 91) {
-		printk(KERN_WARNING "Bad version number %d.%d on %s\n",
-			sb->major_version, sb->minor_version,
-			b);
+		pr_warn("Bad version number %d.%d on %s\n",
+			sb->major_version, sb->minor_version, b);
 		goto abort;
 	}
 
@@ -982,8 +1010,7 @@ static int super_90_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor
 		goto abort;
 
 	if (md_csum_fold(calc_sb_csum(sb)) != md_csum_fold(sb->sb_csum)) {
-		printk(KERN_WARNING "md: invalid superblock checksum on %s\n",
-			b);
+		pr_warn("md: invalid superblock checksum on %s\n", b);
 		goto abort;
 	}
 
@@ -1004,14 +1031,13 @@ static int super_90_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor
 		__u64 ev1, ev2;
 		mdp_super_t *refsb = page_address(refdev->sb_page);
 		if (!uuid_equal(refsb, sb)) {
-			printk(KERN_WARNING "md: %s has different UUID to %s\n",
+			pr_warn("md: %s has different UUID to %s\n",
 				b, bdevname(refdev->bdev,b2));
 			goto abort;
 		}
 		if (!sb_equal(refsb, sb)) {
-			printk(KERN_WARNING "md: %s has same UUID"
-			       " but different superblock to %s\n",
-			       b, bdevname(refdev->bdev, b2));
+			pr_warn("md: %s has same UUID but different superblock to %s\n",
+				b, bdevname(refdev->bdev, b2));
 			goto abort;
 		}
 		ev1 = md_event(sb);
@@ -1158,6 +1184,8 @@ static int super_90_validate(struct mddev *mddev, struct md_rdev *rdev)
 		}
 		if (desc->state & (1<<MD_DISK_WRITEMOSTLY))
 			set_bit(WriteMostly, &rdev->flags);
+		if (desc->state & (1<<MD_DISK_FAILFAST))
+			set_bit(FailFast, &rdev->flags);
 	} else /* MULTIPATH are always insync */
 		set_bit(In_sync, &rdev->flags);
 	return 0;
@@ -1283,6 +1311,8 @@ static void super_90_sync(struct mddev *mddev, struct md_rdev *rdev)
 		}
 		if (test_bit(WriteMostly, &rdev2->flags))
 			d->state |= (1<<MD_DISK_WRITEMOSTLY);
+		if (test_bit(FailFast, &rdev2->flags))
+			d->state |= (1<<MD_DISK_FAILFAST);
 	}
 	/* now set the "removed" and "faulty" bits on any missing devices */
 	for (i=0 ; i < mddev->raid_disks ; i++) {
@@ -1324,9 +1354,10 @@ super_90_rdev_size_change(struct md_rdev *rdev, sector_t num_sectors)
 	if (IS_ENABLED(CONFIG_LBDAF) && (u64)num_sectors >= (2ULL << 32) &&
 	    rdev->mddev->level >= 1)
 		num_sectors = (sector_t)(2ULL << 32) - 2;
-	md_super_write(rdev->mddev, rdev, rdev->sb_start, rdev->sb_size,
+	do {
+		md_super_write(rdev->mddev, rdev, rdev->sb_start, rdev->sb_size,
 		       rdev->sb_page);
-	md_super_wait(rdev->mddev);
+	} while (md_super_wait(rdev->mddev) < 0);
 	return num_sectors;
 }
 
@@ -1413,13 +1444,13 @@ static int super_1_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor_
 		return -EINVAL;
 
 	if (calc_sb_1_csum(sb) != sb->sb_csum) {
-		printk("md: invalid superblock checksum on %s\n",
+		pr_warn("md: invalid superblock checksum on %s\n",
 			bdevname(rdev->bdev,b));
 		return -EINVAL;
 	}
 	if (le64_to_cpu(sb->data_size) < 10) {
-		printk("md: data_size too small on %s\n",
-		       bdevname(rdev->bdev,b));
+		pr_warn("md: data_size too small on %s\n",
+			bdevname(rdev->bdev,b));
 		return -EINVAL;
 	}
 	if (sb->pad0 ||
@@ -1503,8 +1534,7 @@ static int super_1_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor_
 		    sb->level != refsb->level ||
 		    sb->layout != refsb->layout ||
 		    sb->chunksize != refsb->chunksize) {
-			printk(KERN_WARNING "md: %s has strangely different"
-				" superblock to %s\n",
+			pr_warn("md: %s has strangely different superblock to %s\n",
 				bdevname(rdev->bdev,b),
 				bdevname(refdev->bdev,b2));
 			return -EINVAL;
@@ -1646,8 +1676,7 @@ static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev)
 		case MD_DISK_ROLE_JOURNAL: /* journal device */
 			if (!(le32_to_cpu(sb->feature_map) & MD_FEATURE_JOURNAL)) {
 				/* journal device without journal feature */
-				printk(KERN_WARNING
-				  "md: journal device provided without journal feature, ignoring the device\n");
+				pr_warn("md: journal device provided without journal feature, ignoring the device\n");
 				return -EINVAL;
 			}
 			set_bit(Journal, &rdev->flags);
@@ -1669,6 +1698,8 @@ static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev)
 		}
 		if (sb->devflags & WriteMostly1)
 			set_bit(WriteMostly, &rdev->flags);
+		if (sb->devflags & FailFast1)
+			set_bit(FailFast, &rdev->flags);
 		if (le32_to_cpu(sb->feature_map) & MD_FEATURE_REPLACEMENT)
 			set_bit(Replacement, &rdev->flags);
 	} else /* MULTIPATH are always insync */
@@ -1707,6 +1738,10 @@ static void super_1_sync(struct mddev *mddev, struct md_rdev *rdev)
 	sb->chunksize = cpu_to_le32(mddev->chunk_sectors);
 	sb->level = cpu_to_le32(mddev->level);
 	sb->layout = cpu_to_le32(mddev->layout);
+	if (test_bit(FailFast, &rdev->flags))
+		sb->devflags |= FailFast1;
+	else
+		sb->devflags &= ~FailFast1;
 
 	if (test_bit(WriteMostly, &rdev->flags))
 		sb->devflags |= WriteMostly1;
@@ -1863,9 +1898,10 @@ super_1_rdev_size_change(struct md_rdev *rdev, sector_t num_sectors)
 	sb->data_size = cpu_to_le64(num_sectors);
 	sb->super_offset = rdev->sb_start;
 	sb->sb_csum = calc_sb_1_csum(sb);
-	md_super_write(rdev->mddev, rdev, rdev->sb_start, rdev->sb_size,
-		       rdev->sb_page);
-	md_super_wait(rdev->mddev);
+	do {
+		md_super_write(rdev->mddev, rdev, rdev->sb_start, rdev->sb_size,
+			       rdev->sb_page);
+	} while (md_super_wait(rdev->mddev) < 0);
 	return num_sectors;
 
 }
@@ -2004,9 +2040,9 @@ int md_integrity_register(struct mddev *mddev)
 	blk_integrity_register(mddev->gendisk,
 			       bdev_get_integrity(reference->bdev));
 
-	printk(KERN_NOTICE "md: data integrity enabled on %s\n", mdname(mddev));
+	pr_debug("md: data integrity enabled on %s\n", mdname(mddev));
 	if (bioset_integrity_create(mddev->bio_set, BIO_POOL_SIZE)) {
-		printk(KERN_ERR "md: failed to create integrity pool for %s\n",
+		pr_err("md: failed to create integrity pool for %s\n",
 		       mdname(mddev));
 		return -EINVAL;
 	}
@@ -2034,8 +2070,8 @@ int md_integrity_add_rdev(struct md_rdev *rdev, struct mddev *mddev)
 		return 0;
 
 	if (blk_integrity_compare(mddev->gendisk, rdev->bdev->bd_disk) != 0) {
-		printk(KERN_NOTICE "%s: incompatible integrity profile for %s\n",
-				mdname(mddev), bdevname(rdev->bdev, name));
+		pr_err("%s: incompatible integrity profile for %s\n",
+		       mdname(mddev), bdevname(rdev->bdev, name));
 		return -ENXIO;
 	}
 
@@ -2089,15 +2125,15 @@ static int bind_rdev_to_array(struct md_rdev *rdev, struct mddev *mddev)
 	rcu_read_unlock();
 	if (!test_bit(Journal, &rdev->flags) &&
 	    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);
+		pr_warn("md: %s: array is limited to %d devices\n",
+			mdname(mddev), mddev->max_disks);
 		return -EBUSY;
 	}
 	bdevname(rdev->bdev,b);
 	strreplace(b, '/', '!');
 
 	rdev->mddev = mddev;
-	printk(KERN_INFO "md: bind<%s>\n", b);
+	pr_debug("md: bind<%s>\n", b);
 
 	if ((err = kobject_add(&rdev->kobj, &mddev->kobj, "dev-%s", b)))
 		goto fail;
@@ -2116,8 +2152,8 @@ static int bind_rdev_to_array(struct md_rdev *rdev, struct mddev *mddev)
 	return 0;
 
  fail:
-	printk(KERN_WARNING "md: failed to register dev-%s for %s\n",
-	       b, mdname(mddev));
+	pr_warn("md: failed to register dev-%s for %s\n",
+		b, mdname(mddev));
 	return err;
 }
 
@@ -2134,7 +2170,7 @@ 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);
-	printk(KERN_INFO "md: unbind<%s>\n", bdevname(rdev->bdev,b));
+	pr_debug("md: unbind<%s>\n", bdevname(rdev->bdev,b));
 	rdev->mddev = NULL;
 	sysfs_remove_link(&rdev->kobj, "block");
 	sysfs_put(rdev->sysfs_state);
@@ -2164,8 +2200,7 @@ static int lock_rdev(struct md_rdev *rdev, dev_t dev, int shared)
 	bdev = blkdev_get_by_dev(dev, FMODE_READ|FMODE_WRITE|FMODE_EXCL,
 				 shared ? (struct md_rdev *)lock_rdev : rdev);
 	if (IS_ERR(bdev)) {
-		printk(KERN_ERR "md: could not open %s.\n",
-			__bdevname(dev, b));
+		pr_warn("md: could not open %s.\n", __bdevname(dev, b));
 		return PTR_ERR(bdev);
 	}
 	rdev->bdev = bdev;
@@ -2185,8 +2220,7 @@ static void export_rdev(struct md_rdev *rdev)
 {
 	char b[BDEVNAME_SIZE];
 
-	printk(KERN_INFO "md: export_rdev(%s)\n",
-		bdevname(rdev->bdev,b));
+	pr_debug("md: export_rdev(%s)\n", bdevname(rdev->bdev,b));
 	md_rdev_clear(rdev);
 #ifndef MODULE
 	if (test_bit(AutoDetected, &rdev->flags))
@@ -2288,24 +2322,24 @@ void md_update_sb(struct mddev *mddev, int force_change)
 
 	if (mddev->ro) {
 		if (force_change)
-			set_bit(MD_CHANGE_DEVS, &mddev->flags);
+			set_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags);
 		return;
 	}
 
 repeat:
 	if (mddev_is_clustered(mddev)) {
-		if (test_and_clear_bit(MD_CHANGE_DEVS, &mddev->flags))
+		if (test_and_clear_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags))
 			force_change = 1;
-		if (test_and_clear_bit(MD_CHANGE_CLEAN, &mddev->flags))
+		if (test_and_clear_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags))
 			nospares = 1;
 		ret = md_cluster_ops->metadata_update_start(mddev);
 		/* Has someone else has updated the sb */
 		if (!does_sb_need_changing(mddev)) {
 			if (ret == 0)
 				md_cluster_ops->metadata_update_cancel(mddev);
-			bit_clear_unless(&mddev->flags, BIT(MD_CHANGE_PENDING),
-							 BIT(MD_CHANGE_DEVS) |
-							 BIT(MD_CHANGE_CLEAN));
+			bit_clear_unless(&mddev->sb_flags, BIT(MD_SB_CHANGE_PENDING),
+							 BIT(MD_SB_CHANGE_DEVS) |
+							 BIT(MD_SB_CHANGE_CLEAN));
 			return;
 		}
 	}
@@ -2321,10 +2355,10 @@ repeat:
 
 	}
 	if (!mddev->persistent) {
-		clear_bit(MD_CHANGE_CLEAN, &mddev->flags);
-		clear_bit(MD_CHANGE_DEVS, &mddev->flags);
+		clear_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags);
+		clear_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags);
 		if (!mddev->external) {
-			clear_bit(MD_CHANGE_PENDING, &mddev->flags);
+			clear_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags);
 			rdev_for_each(rdev, mddev) {
 				if (rdev->badblocks.changed) {
 					rdev->badblocks.changed = 0;
@@ -2344,9 +2378,9 @@ repeat:
 
 	mddev->utime = ktime_get_real_seconds();
 
-	if (test_and_clear_bit(MD_CHANGE_DEVS, &mddev->flags))
+	if (test_and_clear_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags))
 		force_change = 1;
-	if (test_and_clear_bit(MD_CHANGE_CLEAN, &mddev->flags))
+	if (test_and_clear_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags))
 		/* just a clean<-> dirty transition, possibly leave spares alone,
 		 * though if events isn't the right even/odd, we will have to do
 		 * spares after all
@@ -2402,6 +2436,9 @@ repeat:
 	pr_debug("md: updating %s RAID superblock on device (in sync %d)\n",
 		 mdname(mddev), mddev->in_sync);
 
+	if (mddev->queue)
+		blk_add_trace_msg(mddev->queue, "md md_update_sb");
+rewrite:
 	bitmap_update_sb(mddev->bitmap);
 	rdev_for_each(rdev, mddev) {
 		char b[BDEVNAME_SIZE];
@@ -2433,15 +2470,16 @@ repeat:
 			/* only need to write one superblock... */
 			break;
 	}
-	md_super_wait(mddev);
-	/* if there was a failure, MD_CHANGE_DEVS was set, and we re-write super */
+	if (md_super_wait(mddev) < 0)
+		goto rewrite;
+	/* if there was a failure, MD_SB_CHANGE_DEVS was set, and we re-write super */
 
 	if (mddev_is_clustered(mddev) && ret == 0)
 		md_cluster_ops->metadata_update_finish(mddev);
 
 	if (mddev->in_sync != sync_req ||
-	    !bit_clear_unless(&mddev->flags, BIT(MD_CHANGE_PENDING),
-			       BIT(MD_CHANGE_DEVS) | BIT(MD_CHANGE_CLEAN)))
+	    !bit_clear_unless(&mddev->sb_flags, BIT(MD_SB_CHANGE_PENDING),
+			       BIT(MD_SB_CHANGE_DEVS) | BIT(MD_SB_CHANGE_CLEAN)))
 		/* have to write it out again */
 		goto repeat;
 	wake_up(&mddev->sb_wait);
@@ -2485,7 +2523,7 @@ static int add_bound_rdev(struct md_rdev *rdev)
 	}
 	sysfs_notify_dirent_safe(rdev->sysfs_state);
 
-	set_bit(MD_CHANGE_DEVS, &mddev->flags);
+	set_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags);
 	if (mddev->degraded)
 		set_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
 	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
@@ -2523,51 +2561,41 @@ struct rdev_sysfs_entry {
 static ssize_t
 state_show(struct md_rdev *rdev, char *page)
 {
-	char *sep = "";
+	char *sep = ",";
 	size_t len = 0;
 	unsigned long flags = ACCESS_ONCE(rdev->flags);
 
 	if (test_bit(Faulty, &flags) ||
-	    rdev->badblocks.unacked_exist) {
-		len+= sprintf(page+len, "%sfaulty",sep);
-		sep = ",";
-	}
-	if (test_bit(In_sync, &flags)) {
-		len += sprintf(page+len, "%sin_sync",sep);
-		sep = ",";
-	}
-	if (test_bit(Journal, &flags)) {
-		len += sprintf(page+len, "%sjournal",sep);
-		sep = ",";
-	}
-	if (test_bit(WriteMostly, &flags)) {
-		len += sprintf(page+len, "%swrite_mostly",sep);
-		sep = ",";
-	}
+	    (!test_bit(ExternalBbl, &flags) &&
+	    rdev->badblocks.unacked_exist))
+		len += sprintf(page+len, "faulty%s", sep);
+	if (test_bit(In_sync, &flags))
+		len += sprintf(page+len, "in_sync%s", sep);
+	if (test_bit(Journal, &flags))
+		len += sprintf(page+len, "journal%s", sep);
+	if (test_bit(WriteMostly, &flags))
+		len += sprintf(page+len, "write_mostly%s", sep);
 	if (test_bit(Blocked, &flags) ||
 	    (rdev->badblocks.unacked_exist
-	     && !test_bit(Faulty, &flags))) {
-		len += sprintf(page+len, "%sblocked", sep);
-		sep = ",";
-	}
+	     && !test_bit(Faulty, &flags)))
+		len += sprintf(page+len, "blocked%s", sep);
 	if (!test_bit(Faulty, &flags) &&
 	    !test_bit(Journal, &flags) &&
-	    !test_bit(In_sync, &flags)) {
-		len += sprintf(page+len, "%sspare", sep);
-		sep = ",";
-	}
-	if (test_bit(WriteErrorSeen, &flags)) {
-		len += sprintf(page+len, "%swrite_error", sep);
-		sep = ",";
-	}
-	if (test_bit(WantReplacement, &flags)) {
-		len += sprintf(page+len, "%swant_replacement", sep);
-		sep = ",";
-	}
-	if (test_bit(Replacement, &flags)) {
-		len += sprintf(page+len, "%sreplacement", sep);
-		sep = ",";
-	}
+	    !test_bit(In_sync, &flags))
+		len += sprintf(page+len, "spare%s", sep);
+	if (test_bit(WriteErrorSeen, &flags))
+		len += sprintf(page+len, "write_error%s", sep);
+	if (test_bit(WantReplacement, &flags))
+		len += sprintf(page+len, "want_replacement%s", sep);
+	if (test_bit(Replacement, &flags))
+		len += sprintf(page+len, "replacement%s", sep);
+	if (test_bit(ExternalBbl, &flags))
+		len += sprintf(page+len, "external_bbl%s", sep);
+	if (test_bit(FailFast, &flags))
+		len += sprintf(page+len, "failfast%s", sep);
+
+	if (len)
+		len -= strlen(sep);
 
 	return len+sprintf(page+len, "\n");
 }
@@ -2587,6 +2615,7 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len)
 	 *            so that it gets rebuilt based on bitmap
 	 *  write_error - sets WriteErrorSeen
 	 *  -write_error - clears WriteErrorSeen
+	 *  {,-}failfast - set/clear FailFast
 	 */
 	int err = -EINVAL;
 	if (cmd_match(buf, "faulty") && rdev->mddev->pers) {
@@ -2610,8 +2639,10 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len)
 
 			if (err == 0) {
 				md_kick_rdev_from_array(rdev);
-				if (mddev->pers)
-					md_update_sb(mddev, 1);
+				if (mddev->pers) {
+					set_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags);
+					md_wakeup_thread(mddev->thread);
+				}
 				md_new_event(mddev);
 			}
 		}
@@ -2626,6 +2657,7 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len)
 		err = 0;
 	} else if (cmd_match(buf, "-blocked")) {
 		if (!test_bit(Faulty, &rdev->flags) &&
+		    !test_bit(ExternalBbl, &rdev->flags) &&
 		    rdev->badblocks.unacked_exist) {
 			/* metadata handler doesn't understand badblocks,
 			 * so we need to fail the device
@@ -2642,6 +2674,12 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len)
 	} else if (cmd_match(buf, "insync") && rdev->raid_disk == -1) {
 		set_bit(In_sync, &rdev->flags);
 		err = 0;
+	} else if (cmd_match(buf, "failfast")) {
+		set_bit(FailFast, &rdev->flags);
+		err = 0;
+	} else if (cmd_match(buf, "-failfast")) {
+		clear_bit(FailFast, &rdev->flags);
+		err = 0;
 	} else if (cmd_match(buf, "-insync") && rdev->raid_disk >= 0 &&
 		   !test_bit(Journal, &rdev->flags)) {
 		if (rdev->mddev->pers == NULL) {
@@ -2708,6 +2746,13 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len)
 			}
 		} else
 			err = -EBUSY;
+	} else if (cmd_match(buf, "external_bbl") && (rdev->mddev->external)) {
+		set_bit(ExternalBbl, &rdev->flags);
+		rdev->badblocks.shift = 0;
+		err = 0;
+	} else if (cmd_match(buf, "-external_bbl") && (rdev->mddev->external)) {
+		clear_bit(ExternalBbl, &rdev->flags);
+		err = 0;
 	}
 	if (!err)
 		sysfs_notify_dirent_safe(rdev->sysfs_state);
@@ -3211,10 +3256,8 @@ static struct md_rdev *md_import_device(dev_t newdev, int super_format, int supe
 	sector_t size;
 
 	rdev = kzalloc(sizeof(*rdev), GFP_KERNEL);
-	if (!rdev) {
-		printk(KERN_ERR "md: could not alloc mem for new device!\n");
+	if (!rdev)
 		return ERR_PTR(-ENOMEM);
-	}
 
 	err = md_rdev_init(rdev);
 	if (err)
@@ -3231,8 +3274,7 @@ 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
-			"md: %s has zero or unknown size, marking faulty!\n",
+		pr_warn("md: %s has zero or unknown size, marking faulty!\n",
 			bdevname(rdev->bdev,b));
 		err = -EINVAL;
 		goto abort_free;
@@ -3242,16 +3284,13 @@ static struct md_rdev *md_import_device(dev_t newdev, int super_format, int supe
 		err = super_types[super_format].
 			load_super(rdev, NULL, super_minor);
 		if (err == -EINVAL) {
-			printk(KERN_WARNING
-				"md: %s does not have a valid v%d.%d "
-			       "superblock, not importing!\n",
+			pr_warn("md: %s does not have a valid v%d.%d superblock, not importing!\n",
 				bdevname(rdev->bdev,b),
-			       super_format, super_minor);
+				super_format, super_minor);
 			goto abort_free;
 		}
 		if (err < 0) {
-			printk(KERN_WARNING
-				"md: could not read %s's sb, not importing!\n",
+			pr_warn("md: could not read %s's sb, not importing!\n",
 				bdevname(rdev->bdev,b));
 			goto abort_free;
 		}
@@ -3287,9 +3326,7 @@ static void analyze_sbs(struct mddev *mddev)
 		case 0:
 			break;
 		default:
-			printk( KERN_ERR \
-				"md: fatal superblock inconsistency in %s"
-				" -- removing from array\n",
+			pr_warn("md: fatal superblock inconsistency in %s -- removing from array\n",
 				bdevname(rdev->bdev,b));
 			md_kick_rdev_from_array(rdev);
 		}
@@ -3302,18 +3339,16 @@ static void analyze_sbs(struct mddev *mddev)
 		if (mddev->max_disks &&
 		    (rdev->desc_nr >= mddev->max_disks ||
 		     i > mddev->max_disks)) {
-			printk(KERN_WARNING
-			       "md: %s: %s: only %d devices permitted\n",
-			       mdname(mddev), bdevname(rdev->bdev, b),
-			       mddev->max_disks);
+			pr_warn("md: %s: %s: only %d devices permitted\n",
+				mdname(mddev), bdevname(rdev->bdev, b),
+				mddev->max_disks);
 			md_kick_rdev_from_array(rdev);
 			continue;
 		}
 		if (rdev != freshest) {
 			if (super_types[mddev->major_version].
 			    validate_super(mddev, rdev)) {
-				printk(KERN_WARNING "md: kicking non-fresh %s"
-					" from array!\n",
+				pr_warn("md: kicking non-fresh %s from array!\n",
 					bdevname(rdev->bdev,b));
 				md_kick_rdev_from_array(rdev);
 				continue;
@@ -3384,7 +3419,7 @@ safe_delay_store(struct mddev *mddev, const char *cbuf, size_t len)
 	unsigned long msec;
 
 	if (mddev_is_clustered(mddev)) {
-		pr_info("md: Safemode is disabled for clustered mode\n");
+		pr_warn("md: Safemode is disabled for clustered mode\n");
 		return -EINVAL;
 	}
 
@@ -3472,8 +3507,8 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 
 	rv = -EINVAL;
 	if (!mddev->pers->quiesce) {
-		printk(KERN_WARNING "md: %s: %s does not support online personality change\n",
-		       mdname(mddev), mddev->pers->name);
+		pr_warn("md: %s: %s does not support online personality change\n",
+			mdname(mddev), mddev->pers->name);
 		goto out_unlock;
 	}
 
@@ -3491,7 +3526,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 	pers = find_pers(level, clevel);
 	if (!pers || !try_module_get(pers->owner)) {
 		spin_unlock(&pers_lock);
-		printk(KERN_WARNING "md: personality %s not loaded\n", clevel);
+		pr_warn("md: personality %s not loaded\n", clevel);
 		rv = -EINVAL;
 		goto out_unlock;
 	}
@@ -3505,8 +3540,8 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 	}
 	if (!pers->takeover) {
 		module_put(pers->owner);
-		printk(KERN_WARNING "md: %s: %s does not support personality takeover\n",
-		       mdname(mddev), clevel);
+		pr_warn("md: %s: %s does not support personality takeover\n",
+			mdname(mddev), clevel);
 		rv = -EINVAL;
 		goto out_unlock;
 	}
@@ -3526,8 +3561,8 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 		mddev->delta_disks = 0;
 		mddev->reshape_backwards = 0;
 		module_put(pers->owner);
-		printk(KERN_WARNING "md: %s: %s would not accept array\n",
-		       mdname(mddev), clevel);
+		pr_warn("md: %s: %s would not accept array\n",
+			mdname(mddev), clevel);
 		rv = PTR_ERR(priv);
 		goto out_unlock;
 	}
@@ -3570,9 +3605,8 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 	    pers->sync_request != NULL) {
 		/* need to add the md_redundancy_group */
 		if (sysfs_create_group(&mddev->kobj, &md_redundancy_group))
-			printk(KERN_WARNING
-			       "md: cannot register extra attributes for %s\n",
-			       mdname(mddev));
+			pr_warn("md: cannot register extra attributes for %s\n",
+				mdname(mddev));
 		mddev->sysfs_action = sysfs_get_dirent(mddev->kobj.sd, "sync_action");
 	}
 	if (oldpers->sync_request != NULL &&
@@ -3603,9 +3637,8 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 			clear_bit(In_sync, &rdev->flags);
 		else {
 			if (sysfs_link_rdev(mddev, rdev))
-				printk(KERN_WARNING "md: cannot register rd%d"
-				       " for %s after level change\n",
-				       rdev->raid_disk, mdname(mddev));
+				pr_warn("md: cannot register rd%d for %s after level change\n",
+					rdev->raid_disk, mdname(mddev));
 		}
 	}
 
@@ -3618,7 +3651,7 @@ level_store(struct mddev *mddev, const char *buf, size_t len)
 	}
 	blk_set_stacking_limits(&mddev->queue->limits);
 	pers->run(mddev);
-	set_bit(MD_CHANGE_DEVS, &mddev->flags);
+	set_bit(MD_SB_CHANGE_DEVS, &mddev->sb_flags);
 	mddev_resume(mddev);
 	if (!mddev->thread)
 		md_update_sb(mddev, 1);
@@ -3813,7 +3846,7 @@ resync_start_store(struct mddev *mddev, const char *buf, size_t len)
 	if (!err) {
 		mddev->recovery_cp = n;
 		if (mddev->pers)
-			set_bit(MD_CHANGE_CLEAN, &mddev->flags);
+			set_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags);
 	}
 	mddev_unlock(mddev);
 	return err ?: len;
@@ -3887,7 +3920,7 @@ array_state_show(struct mddev *mddev, char *page)
 			st = read_auto;
 			break;
 		case 0:
-			if (test_bit(MD_CHANGE_PENDING, &mddev->flags))
+			if (test_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags))
 				st = write_pending;
 			else if (mddev->in_sync)
 				st = clean;
@@ -3925,7 +3958,8 @@ array_state_store(struct mddev *mddev, const char *buf, size_t len)
 		spin_lock(&mddev->lock);
 		if (st == active) {
 			restart_array(mddev);
-			clear_bit(MD_CHANGE_PENDING, &mddev->flags);
+			clear_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags);
+			md_wakeup_thread(mddev->thread);
 			wake_up(&mddev->sb_wait);
 			err = 0;
 		} else /* st == clean */ {
@@ -3935,7 +3969,7 @@ array_state_store(struct mddev *mddev, const char *buf, size_t len)
 					mddev->in_sync = 1;
 					if (mddev->safemode == 1)
 						mddev->safemode = 0;
-					set_bit(MD_CHANGE_CLEAN, &mddev->flags);
+					set_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags);
 				}
 				err = 0;
 			} else
@@ -4001,7 +4035,7 @@ array_state_store(struct mddev *mddev, const char *buf, size_t len)
 					mddev->in_sync = 1;
 					if (mddev->safemode == 1)
 						mddev->safemode = 0;
-					set_bit(MD_CHANGE_CLEAN, &mddev->flags);
+					set_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags);
 				}
 				err = 0;
 			} else
@@ -4015,7 +4049,7 @@ array_state_store(struct mddev *mddev, const char *buf, size_t len)
 			err = restart_array(mddev);
 			if (err)
 				break;
-			clear_bit(MD_CHANGE_PENDING, &mddev->flags);
+			clear_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags);
 			wake_up(&mddev->sb_wait);
 			err = 0;
 		} else {
@@ -5071,13 +5105,13 @@ static int md_alloc(dev_t dev, char *name)
 		/* This isn't possible, but as kobject_init_and_add is marked
 		 * __must_check, we must do something with the result
 		 */
-		printk(KERN_WARNING "md: cannot register %s/md - name in use\n",
-		       disk->disk_name);
+		pr_debug("md: cannot register %s/md - name in use\n",
+			 disk->disk_name);
 		error = 0;
 	}
 	if (mddev->kobj.sd &&
 	    sysfs_create_group(&mddev->kobj, &md_bitmap_group))
-		printk(KERN_DEBUG "pointless warning\n");
+		pr_debug("pointless warning\n");
 	mutex_unlock(&mddev->open_mutex);
  abort:
 	mutex_unlock(&disks_mutex);
@@ -5179,15 +5213,15 @@ int md_run(struct mddev *mddev)
 			if (mddev->dev_sectors &&
 			    rdev->data_offset + mddev->dev_sectors
 			    > rdev->sb_start) {
-				printk("md: %s: data overlaps metadata\n",
-				       mdname(mddev));
+				pr_warn("md: %s: data overlaps metadata\n",
+					mdname(mddev));
 				return -EINVAL;
 			}
 		} else {
 			if (rdev->sb_start + rdev->sb_size/512
 			    > rdev->data_offset) {
-				printk("md: %s: metadata overlaps data\n",
-				       mdname(mddev));
+				pr_warn("md: %s: metadata overlaps data\n",
+					mdname(mddev));
 				return -EINVAL;
 			}
 		}
@@ -5202,11 +5236,11 @@ int md_run(struct mddev *mddev)
 	if (!pers || !try_module_get(pers->owner)) {
 		spin_unlock(&pers_lock);
 		if (mddev->level != LEVEL_NONE)
-			printk(KERN_WARNING "md: personality for level %d is not loaded!\n",
-			       mddev->level);
+			pr_warn("md: personality for level %d is not loaded!\n",
+				mddev->level);
 		else
-			printk(KERN_WARNING "md: personality for level %s is not loaded!\n",
-			       mddev->clevel);
+			pr_warn("md: personality for level %s is not loaded!\n",
+				mddev->clevel);
 		return -EINVAL;
 	}
 	spin_unlock(&pers_lock);
@@ -5236,21 +5270,16 @@ int md_run(struct mddev *mddev)
 				if (rdev < rdev2 &&
 				    rdev->bdev->bd_contains ==
 				    rdev2->bdev->bd_contains) {
-					printk(KERN_WARNING
-					       "%s: WARNING: %s appears to be"
-					       " on the same physical disk as"
-					       " %s.\n",
-					       mdname(mddev),
-					       bdevname(rdev->bdev,b),
-					       bdevname(rdev2->bdev,b2));
+					pr_warn("%s: WARNING: %s appears to be on the same physical disk as %s.\n",
+						mdname(mddev),
+						bdevname(rdev->bdev,b),
+						bdevname(rdev2->bdev,b2));
 					warned = 1;
 				}
 			}
 
 		if (warned)
-			printk(KERN_WARNING
-			       "True protection against single-disk"
-			       " failure might be compromised.\n");
+			pr_warn("True protection against single-disk failure might be compromised.\n");
 	}
 
 	mddev->recovery = 0;
@@ -5264,14 +5293,14 @@ int md_run(struct mddev *mddev)
 
 	err = pers->run(mddev);
 	if (err)
-		printk(KERN_ERR "md: pers->run() failed ...\n");
+		pr_warn("md: pers->run() failed ...\n");
 	else if (pers->size(mddev, 0, 0) < mddev->array_sectors) {
-		WARN_ONCE(!mddev->external_size, "%s: default size too small,"
-			  " but 'external_size' not in effect?\n", __func__);
-		printk(KERN_ERR
-		       "md: invalid array_size %llu > default size %llu\n",
-		       (unsigned long long)mddev->array_sectors / 2,
-		       (unsigned long long)pers->size(mddev, 0, 0) / 2);
+		WARN_ONCE(!mddev->external_size,
+			  "%s: default size too small, but 'external_size' not in effect?\n",
+			  __func__);
+		pr_warn("md: invalid array_size %llu > default size %llu\n",
+			(unsigned long long)mddev->array_sectors / 2,
+			(unsigned long long)pers->size(mddev, 0, 0) / 2);
 		err = -EINVAL;
 	}
 	if (err == 0 && pers->sync_request &&
@@ -5281,8 +5310,8 @@ int md_run(struct mddev *mddev)
 		bitmap = bitmap_create(mddev, -1);
 		if (IS_ERR(bitmap)) {
 			err = PTR_ERR(bitmap);
-			printk(KERN_ERR "%s: failed to create bitmap (%d)\n",
-			       mdname(mddev), err);
+			pr_warn("%s: failed to create bitmap (%d)\n",
+				mdname(mddev), err);
 		} else
 			mddev->bitmap = bitmap;
 
@@ -5318,9 +5347,8 @@ int md_run(struct mddev *mddev)
 	if (pers->sync_request) {
 		if (mddev->kobj.sd &&
 		    sysfs_create_group(&mddev->kobj, &md_redundancy_group))
-			printk(KERN_WARNING
-			       "md: cannot register extra attributes for %s\n",
-			       mdname(mddev));
+			pr_warn("md: cannot register extra attributes for %s\n",
+				mdname(mddev));
 		mddev->sysfs_action = sysfs_get_dirent_safe(mddev->kobj.sd, "sync_action");
 	} else if (mddev->ro == 2) /* auto-readonly not meaningful */
 		mddev->ro = 0;
@@ -5350,7 +5378,7 @@ int md_run(struct mddev *mddev)
 		set_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
 	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
 
-	if (mddev->flags & MD_UPDATE_SB_FLAGS)
+	if (mddev->sb_flags)
 		md_update_sb(mddev, 0);
 
 	md_new_event(mddev);
@@ -5421,8 +5449,7 @@ static int restart_array(struct mddev *mddev)
 	mddev->safemode = 0;
 	mddev->ro = 0;
 	set_disk_ro(disk, 0);
-	printk(KERN_INFO "md: %s switched to read-write mode.\n",
-		mdname(mddev));
+	pr_debug("md: %s switched to read-write mode.\n", mdname(mddev));
 	/* Kick recovery or resync if necessary */
 	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
 	md_wakeup_thread(mddev->thread);
@@ -5446,6 +5473,7 @@ static void md_clean(struct mddev *mddev)
 	mddev->level = LEVEL_NONE;
 	mddev->clevel[0] = 0;
 	mddev->flags = 0;
+	mddev->sb_flags = 0;
 	mddev->ro = 0;
 	mddev->metadata_type[0] = 0;
 	mddev->chunk_sectors = 0;
@@ -5490,12 +5518,15 @@ static void __md_stop_writes(struct mddev *mddev)
 
 	del_timer_sync(&mddev->safemode_timer);
 
+	if (mddev->pers && mddev->pers->quiesce) {
+		mddev->pers->quiesce(mddev, 1);
+		mddev->pers->quiesce(mddev, 0);
+	}
 	bitmap_flush(mddev);
-	md_super_wait(mddev);
 
 	if (mddev->ro == 0 &&
 	    ((!mddev->in_sync && !mddev_is_clustered(mddev)) ||
-	     (mddev->flags & MD_UPDATE_SB_FLAGS))) {
+	     mddev->sb_flags)) {
 		/* mark array as shutdown cleanly */
 		if (!mddev_is_clustered(mddev))