Merge tag 'md-3.3' of git://neil.brown.name/md

md update for 3.3

Big change is new hot-replacement.
A slot in an array can hold 2 devices - one that
wants-replacement and one that is the replacement.
Once the replacement is built - either from the
original or (in the case of errors) from elsewhere,
the wants-replacement device will be removed.

* tag 'md-3.3' of git://neil.brown.name/md: (36 commits)
  md/raid1: Mark device want_replacement when we see a write error.
  md/raid1: If there is a spare and a want_replacement device, start replacement.
  md/raid1: recognise replacements when assembling arrays.
  md/raid1: handle activation of replacement device when recovery completes.
  md/raid1: Allow a failed replacement device to be removed.
  md/raid1: Allocate spare to store replacement devices and their bios.
  md/raid1:  Replace use of mddev->raid_disks with conf->raid_disks.
  md/raid10: If there is a spare and a want_replacement device, start replacement.
  md/raid10: recognise replacements when assembling array.
  md/raid10: Allow replacement device to be replace old drive.
  md/raid10: handle recovery of replacement devices.
  md/raid10:  Handle replacement devices during resync.
  md/raid10: writes should get directed to replacement as well as original.
  md/raid10: allow removal of failed replacement devices.
  md/raid10: preferentially read from replacement device if possible.
  md/raid10:  change read_balance to return an rdev
  md/raid10: prepare data structures for handling replacement.
  md/raid5: Mark device want_replacement when we see a write error.
  md/raid5: If there is a spare and a want_replacement device, start replacement.
  md/raid5: recognise replacements when assembling array.
  ...

Documentation/md.txt

@@ -357,14 +357,14 @@ Each directory contains:
         written to, that device.
 
       state
-        A file recording the current state of the device in the array
+	A file recording the current state of the device in the array
 	which can be a comma separated list of
 	      faulty   - device has been kicked from active use due to
-                         a detected fault or it has unacknowledged bad
-                         blocks
+			 a detected fault, or it has unacknowledged bad
+			 blocks
 	      in_sync  - device is a fully in-sync member of the array
 	      writemostly - device will only be subject to read
-		         requests if there are no other options.
+			 requests if there are no other options.
 			 This applies only to raid1 arrays.
 	      blocked  - device has failed, and the failure hasn't been
 			 acknowledged yet by the metadata handler.
@@ -374,6 +374,13 @@ Each directory contains:
 			 This includes spares that are in the process
 			 of being recovered to
 	      write_error - device has ever seen a write error.
+	      want_replacement - device is (mostly) working but probably
+			 should be replaced, either due to errors or
+			 due to user request.
+	      replacement - device is a replacement for another active
+			 device with same raid_disk.
+
+
 	This list may grow in future.
 	This can be written to.
 	Writing "faulty"  simulates a failure on the device.
@@ -386,6 +393,13 @@ Each directory contains:
 	Writing "in_sync" sets the in_sync flag.
 	Writing "write_error" sets writeerrorseen flag.
 	Writing "-write_error" clears writeerrorseen flag.
+	Writing "want_replacement" is allowed at any time except to a
+		replacement device or a spare.  It sets the flag.
+	Writing "-want_replacement" is allowed at any time.  It clears
+		the flag.
+	Writing "replacement" or "-replacement" is only allowed before
+		starting the array.  It sets or clears the flag.
+
 
 	This file responds to select/poll. Any change to 'faulty'
 	or 'blocked' causes an event.

drivers/md/bitmap.c

@@ -1149,12 +1149,12 @@ void bitmap_daemon_work(struct mddev *mddev)
 		return;
 	}
 	if (time_before(jiffies, bitmap->daemon_lastrun
-			+ bitmap->mddev->bitmap_info.daemon_sleep))
+			+ mddev->bitmap_info.daemon_sleep))
 		goto done;
 
 	bitmap->daemon_lastrun = jiffies;
 	if (bitmap->allclean) {
-		bitmap->mddev->thread->timeout = MAX_SCHEDULE_TIMEOUT;
+		mddev->thread->timeout = MAX_SCHEDULE_TIMEOUT;
 		goto done;
 	}
 	bitmap->allclean = 1;
@@ -1206,7 +1206,7 @@ void bitmap_daemon_work(struct mddev *mddev)
 			 * sure that events_cleared is up-to-date.
 			 */
 			if (bitmap->need_sync &&
-			    bitmap->mddev->bitmap_info.external == 0) {
+			    mddev->bitmap_info.external == 0) {
 				bitmap_super_t *sb;
 				bitmap->need_sync = 0;
 				sb = kmap_atomic(bitmap->sb_page, KM_USER0);
@@ -1270,8 +1270,8 @@ void bitmap_daemon_work(struct mddev *mddev)
 
  done:
 	if (bitmap->allclean == 0)
-		bitmap->mddev->thread->timeout =
-			bitmap->mddev->bitmap_info.daemon_sleep;
+		mddev->thread->timeout =
+			mddev->bitmap_info.daemon_sleep;
 	mutex_unlock(&mddev->bitmap_info.mutex);
 }
 
@@ -1587,7 +1587,7 @@ static void bitmap_set_memory_bits(struct bitmap *bitmap, sector_t offset, int n
 	}
 	if (!*bmc) {
 		struct page *page;
-		*bmc = 1 | (needed ? NEEDED_MASK : 0);
+		*bmc = 2 | (needed ? NEEDED_MASK : 0);
 		bitmap_count_page(bitmap, offset, 1);
 		page = filemap_get_page(bitmap, offset >> CHUNK_BLOCK_SHIFT(bitmap));
 		set_page_attr(bitmap, page, BITMAP_PAGE_PENDING);

drivers/md/md.c

@@ -1713,6 +1713,8 @@ static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev)
 		}
 		if (sb->devflags & WriteMostly1)
 			set_bit(WriteMostly, &rdev->flags);
+		if (le32_to_cpu(sb->feature_map) & MD_FEATURE_REPLACEMENT)
+			set_bit(Replacement, &rdev->flags);
 	} else /* MULTIPATH are always insync */
 		set_bit(In_sync, &rdev->flags);
 
@@ -1766,6 +1768,9 @@ static void super_1_sync(struct mddev *mddev, struct md_rdev *rdev)
 		sb->recovery_offset =
 			cpu_to_le64(rdev->recovery_offset);
 	}
+	if (test_bit(Replacement, &rdev->flags))
+		sb->feature_map |=
+			cpu_to_le32(MD_FEATURE_REPLACEMENT);
 
 	if (mddev->reshape_position != MaxSector) {
 		sb->feature_map |= cpu_to_le32(MD_FEATURE_RESHAPE_ACTIVE);
@@ -2559,6 +2564,15 @@ state_show(struct md_rdev *rdev, char *page)
 		len += sprintf(page+len, "%swrite_error", sep);
 		sep = ",";
 	}
+	if (test_bit(WantReplacement, &rdev->flags)) {
+		len += sprintf(page+len, "%swant_replacement", sep);
+		sep = ",";
+	}
+	if (test_bit(Replacement, &rdev->flags)) {
+		len += sprintf(page+len, "%sreplacement", sep);
+		sep = ",";
+	}
+
 	return len+sprintf(page+len, "\n");
 }
 
@@ -2627,6 +2641,42 @@ state_store(struct md_rdev *rdev, const char *buf, size_t len)
 	} else if (cmd_match(buf, "-write_error")) {
 		clear_bit(WriteErrorSeen, &rdev->flags);
 		err = 0;
+	} else if (cmd_match(buf, "want_replacement")) {
+		/* Any non-spare device that is not a replacement can
+		 * become want_replacement at any time, but we then need to
+		 * check if recovery is needed.
+		 */
+		if (rdev->raid_disk >= 0 &&
+		    !test_bit(Replacement, &rdev->flags))
+			set_bit(WantReplacement, &rdev->flags);
+		set_bit(MD_RECOVERY_NEEDED, &rdev->mddev->recovery);
+		md_wakeup_thread(rdev->mddev->thread);
+		err = 0;
+	} else if (cmd_match(buf, "-want_replacement")) {
+		/* Clearing 'want_replacement' is always allowed.
+		 * Once replacements starts it is too late though.
+		 */
+		err = 0;
+		clear_bit(WantReplacement, &rdev->flags);
+	} else if (cmd_match(buf, "replacement")) {
+		/* Can only set a device as a replacement when array has not
+		 * yet been started.  Once running, replacement is automatic
+		 * from spares, or by assigning 'slot'.
+		 */
+		if (rdev->mddev->pers)
+			err = -EBUSY;
+		else {
+			set_bit(Replacement, &rdev->flags);
+			err = 0;
+		}
+	} else if (cmd_match(buf, "-replacement")) {
+		/* Similarly, can only clear Replacement before start */
+		if (rdev->mddev->pers)
+			err = -EBUSY;
+		else {
+			clear_bit(Replacement, &rdev->flags);
+			err = 0;
+		}
 	}
 	if (!err)
 		sysfs_notify_dirent_safe(rdev->sysfs_state);
@@ -2688,7 +2738,7 @@ slot_store(struct md_rdev *rdev, const char *buf, size_t len)
 		if (rdev->mddev->pers->hot_remove_disk == NULL)
 			return -EINVAL;
 		err = rdev->mddev->pers->
-			hot_remove_disk(rdev->mddev, rdev->raid_disk);
+			hot_remove_disk(rdev->mddev, rdev);
 		if (err)
 			return err;
 		sysfs_unlink_rdev(rdev->mddev, rdev);
@@ -2696,7 +2746,6 @@ slot_store(struct md_rdev *rdev, const char *buf, size_t len)
 		set_bit(MD_RECOVERY_NEEDED, &rdev->mddev->recovery);
 		md_wakeup_thread(rdev->mddev->thread);
 	} else if (rdev->mddev->pers) {
-		struct md_rdev *rdev2;
 		/* Activating a spare .. or possibly reactivating
 		 * if we ever get bitmaps working here.
 		 */
@@ -2710,10 +2759,6 @@ slot_store(struct md_rdev *rdev, const char *buf, size_t len)
 		if (rdev->mddev->pers->hot_add_disk == NULL)
 			return -EINVAL;
 
-		list_for_each_entry(rdev2, &rdev->mddev->disks, same_set)
-			if (rdev2->raid_disk == slot)
-				return -EEXIST;
-
 		if (slot >= rdev->mddev->raid_disks &&
 		    slot >= rdev->mddev->raid_disks + rdev->mddev->delta_disks)
 			return -ENOSPC;
@@ -6053,8 +6098,15 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 	struct mddev *mddev = NULL;
 	int ro;
 
-	if (!capable(CAP_SYS_ADMIN))
-		return -EACCES;
+	switch (cmd) {
+	case RAID_VERSION:
+	case GET_ARRAY_INFO:
+	case GET_DISK_INFO:
+		break;
+	default:
+		if (!capable(CAP_SYS_ADMIN))
+			return -EACCES;
+	}
 
 	/*
 	 * Commands dealing with the RAID driver but not any
@@ -6714,8 +6766,11 @@ static int md_seq_show(struct seq_file *seq, void *v)
 			if (test_bit(Faulty, &rdev->flags)) {
 				seq_printf(seq, "(F)");
 				continue;
-			} else if (rdev->raid_disk < 0)
+			}
+			if (rdev->raid_disk < 0)
 				seq_printf(seq, "(S)"); /* spare */
+			if (test_bit(Replacement, &rdev->flags))
+				seq_printf(seq, "(R)");
 			sectors += rdev->sectors;
 		}
 
@@ -7337,29 +7392,27 @@ static int remove_and_add_spares(struct mddev *mddev)
 		     ! test_bit(In_sync, &rdev->flags)) &&
 		    atomic_read(&rdev->nr_pending)==0) {
 			if (mddev->pers->hot_remove_disk(
-				    mddev, rdev->raid_disk)==0) {
+				    mddev, rdev) == 0) {
 				sysfs_unlink_rdev(mddev, rdev);
 				rdev->raid_disk = -1;
 			}
 		}
 
-	if (mddev->degraded) {
-		list_for_each_entry(rdev, &mddev->disks, same_set) {
-			if (rdev->raid_disk >= 0 &&
-			    !test_bit(In_sync, &rdev->flags) &&
-			    !test_bit(Faulty, &rdev->flags))
+	list_for_each_entry(rdev, &mddev->disks, same_set) {
+		if (rdev->raid_disk >= 0 &&
+		    !test_bit(In_sync, &rdev->flags) &&
+		    !test_bit(Faulty, &rdev->flags))
+			spares++;
+		if (rdev->raid_disk < 0
+		    && !test_bit(Faulty, &rdev->flags)) {
+			rdev->recovery_offset = 0;
+			if (mddev->pers->
+			    hot_add_disk(mddev, rdev) == 0) {
+				if (sysfs_link_rdev(mddev, rdev))
+					/* failure here is OK */;
 				spares++;
-			if (rdev->raid_disk < 0
-			    && !test_bit(Faulty, &rdev->flags)) {
-				rdev->recovery_offset = 0;
-				if (mddev->pers->
-				    hot_add_disk(mddev, rdev) == 0) {
-					if (sysfs_link_rdev(mddev, rdev))
-						/* failure here is OK */;
-					spares++;
-					md_new_event(mddev);
-					set_bit(MD_CHANGE_DEVS, &mddev->flags);
-				}
+				md_new_event(mddev);
+				set_bit(MD_CHANGE_DEVS, &mddev->flags);
 			}
 		}
 	}
@@ -7474,7 +7527,7 @@ void md_check_recovery(struct mddev *mddev)
 				    test_bit(Faulty, &rdev->flags) &&
 				    atomic_read(&rdev->nr_pending)==0) {
 					if (mddev->pers->hot_remove_disk(
-						    mddev, rdev->raid_disk)==0) {
+						    mddev, rdev) == 0) {
 						sysfs_unlink_rdev(mddev, rdev);
 						rdev->raid_disk = -1;
 					}

drivers/md/md.h

@@ -72,34 +72,7 @@ struct md_rdev {
 	 * This reduces the burden of testing multiple flags in many cases
 	 */
 
-	unsigned long	flags;
-#define	Faulty		1		/* device is known to have a fault */
-#define	In_sync		2		/* device is in_sync with rest of array */
-#define	WriteMostly	4		/* Avoid reading if at all possible */
-#define	AutoDetected	7		/* added by auto-detect */
-#define Blocked		8		/* An error occurred but has not yet
-					 * been acknowledged by the metadata
-					 * handler, so don't allow writes
-					 * until it is cleared */
-#define WriteErrorSeen	9		/* A write error has been seen on this
-					 * device
-					 */
-#define FaultRecorded	10		/* Intermediate state for clearing
-					 * Blocked.  The Fault is/will-be
-					 * recorded in the metadata, but that
-					 * metadata hasn't been stored safely
-					 * on disk yet.
-					 */
-#define BlockedBadBlocks 11		/* A writer is blocked because they
-					 * found an unacknowledged bad-block.
-					 * This can safely be cleared at any
-					 * time, and the writer will re-check.
-					 * It may be set at any time, and at
-					 * worst the writer will timeout and
-					 * re-check.  So setting it as
-					 * accurately as possible is good, but
-					 * not absolutely critical.
-					 */
+	unsigned long	flags;	/* bit set of 'enum flag_bits' bits. */
 	wait_queue_head_t blocked_wait;
 
 	int desc_nr;			/* descriptor index in the superblock */
@@ -152,6 +125,44 @@ struct md_rdev {
 		sector_t size;		/* in sectors */
 	} badblocks;
 };
+enum flag_bits {
+	Faulty,			/* device is known to have a fault */
+	In_sync,		/* device is in_sync with rest of array */
+	WriteMostly,		/* Avoid reading if at all possible */
+	AutoDetected,		/* added by auto-detect */
+	Blocked,		/* An error occurred but has not yet
+				 * been acknowledged by the metadata
+				 * handler, so don't allow writes
+				 * until it is cleared */
+	WriteErrorSeen,		/* A write error has been seen on this
+				 * device
+				 */
+	FaultRecorded,		/* Intermediate state for clearing
+				 * Blocked.  The Fault is/will-be
+				 * recorded in the metadata, but that
+				 * metadata hasn't been stored safely
+				 * on disk yet.
+				 */
+	BlockedBadBlocks,	/* A writer is blocked because they
+				 * found an unacknowledged bad-block.
+				 * This can safely be cleared at any
+				 * time, and the writer will re-check.
+				 * It may be set at any time, and at
+				 * worst the writer will timeout and
+				 * re-check.  So setting it as
+				 * accurately as possible is good, but
+				 * not absolutely critical.
+				 */
+	WantReplacement,	/* This device is a candidate to be
+				 * hot-replaced, either because it has
+				 * reported some faults, or because
+				 * of explicit request.
+				 */
+	Replacement,		/* This device is a replacement for
+				 * a want_replacement device with same
+				 * raid_disk number.
+				 */
+};
 
 #define BB_LEN_MASK	(0x00000000000001FFULL)
 #define BB_OFFSET_MASK	(0x7FFFFFFFFFFFFE00ULL)
@@ -428,7 +439,7 @@ struct md_personality
 	 */
 	void (*error_handler)(struct mddev *mddev, struct md_rdev *rdev);
 	int (*hot_add_disk) (struct mddev *mddev, struct md_rdev *rdev);
-	int (*hot_remove_disk) (struct mddev *mddev, int number);
+	int (*hot_remove_disk) (struct mddev *mddev, struct md_rdev *rdev);
 	int (*spare_active) (struct mddev *mddev);
 	sector_t (*sync_request)(struct mddev *mddev, sector_t sector_nr, int *skipped, int go_faster);
 	int (*resize) (struct mddev *mddev, sector_t sectors);
@@ -482,15 +493,20 @@ static inline char * mdname (struct mddev * mddev)
 static inline int sysfs_link_rdev(struct mddev *mddev, struct md_rdev *rdev)
 {
 	char nm[20];
-	sprintf(nm, "rd%d", rdev->raid_disk);
-	return sysfs_create_link(&mddev->kobj, &rdev->kobj, nm);
+	if (!test_bit(Replacement, &rdev->flags)) {
+		sprintf(nm, "rd%d", rdev->raid_disk);
+		return sysfs_create_link(&mddev->kobj, &rdev->kobj, nm);
+	} else
+		return 0;
 }
 
 static inline void sysfs_unlink_rdev(struct mddev *mddev, struct md_rdev *rdev)
 {
 	char nm[20];
-	sprintf(nm, "rd%d", rdev->raid_disk);
-	sysfs_remove_link(&mddev->kobj, nm);
+	if (!test_bit(Replacement, &rdev->flags)) {
+		sprintf(nm, "rd%d", rdev->raid_disk);
+		sysfs_remove_link(&mddev->kobj, nm);
+	}
 }
 
 /*

drivers/md/multipath.c

@@ -292,17 +292,16 @@ static int multipath_add_disk(struct mddev *mddev, struct md_rdev *rdev)
 	return err;
 }
 
-static int multipath_remove_disk(struct mddev *mddev, int number)
+static int multipath_remove_disk(struct mddev *mddev, struct md_rdev *rdev)
 {
 	struct mpconf *conf = mddev->private;
 	int err = 0;
-	struct md_rdev *rdev;
+	int number = rdev->raid_disk;
 	struct multipath_info *p = conf->multipaths + number;
 
 	print_multipath_conf(conf);
 
-	rdev = p->rdev;
-	if (rdev) {
+	if (rdev == p->rdev) {
 		if (test_bit(In_sync, &rdev->flags) ||
 		    atomic_read(&rdev->nr_pending)) {
 			printk(KERN_ERR "hot-remove-disk, slot %d is identified"

drivers/md/raid1.h

@@ -12,6 +12,9 @@ struct mirror_info {
  * pool was allocated for, so they know how much to allocate and free.
  * mddev->raid_disks cannot be used, as it can change while a pool is active
  * These two datums are stored in a kmalloced struct.
+ * The 'raid_disks' here is twice the raid_disks in r1conf.
+ * This allows space for each 'real' device can have a replacement in the
+ * second half of the array.
  */
 
 struct pool_info {
@@ -21,7 +24,9 @@ struct pool_info {
 
 struct r1conf {
 	struct mddev		*mddev;
-	struct mirror_info		*mirrors;
+	struct mirror_info	*mirrors;	/* twice 'raid_disks' to
+						 * allow for replacements.
+						 */
 	int			raid_disks;
 
 	/* When choose the best device for a read (read_balance())

drivers/md/raid10.h

@@ -2,7 +2,7 @@
 #define _RAID10_H
 
 struct mirror_info {
-	struct md_rdev	*rdev;
+	struct md_rdev	*rdev, *replacement;
 	sector_t	head_position;
 	int		recovery_disabled;	/* matches
 						 * mddev->recovery_disabled
@@ -18,12 +18,13 @@ struct r10conf {
 	spinlock_t		device_lock;
 
 	/* geometry */
-	int			near_copies;  /* number of copies laid out raid0 style */
+	int			near_copies;  /* number of copies laid out
+					       * raid0 style */
 	int 			far_copies;   /* number of copies laid out
 					       * at large strides across drives
 					       */
-	int			far_offset;   /* far_copies are offset by 1 stripe
-					       * instead of many
+	int			far_offset;   /* far_copies are offset by 1
+					       * stripe instead of many
 					       */
 	int			copies;	      /* near_copies * far_copies.
 					       * must be <= raid_disks
@@ -34,10 +35,11 @@ struct r10conf {
 					       * 1 stripe.
 					       */
 
-	sector_t		dev_sectors;  /* temp copy of mddev->dev_sectors */
+	sector_t		dev_sectors;  /* temp copy of
+					       * mddev->dev_sectors */
 
-	int chunk_shift; /* shift from chunks to sectors */
-	sector_t chunk_mask;
+	int			chunk_shift; /* shift from chunks to sectors */
+	sector_t		chunk_mask;
 
 	struct list_head	retry_list;
 	/* queue pending writes and submit them on unplug */
@@ -45,20 +47,22 @@ struct r10conf {
 	int			pending_count;
 
 	spinlock_t		resync_lock;
-	int nr_pending;
-	int nr_waiting;
-	int nr_queued;
-	int barrier;
+	int			nr_pending;
+	int			nr_waiting;
+	int			nr_queued;
+	int			barrier;
 	sector_t		next_resync;
 	int			fullsync;  /* set to 1 if a full sync is needed,
 					    * (fresh device added).
 					    * Cleared when a sync completes.
 					    */
-
+	int			have_replacement; /* There is at least one
+						   * replacement device.
+						   */
 	wait_queue_head_t	wait_barrier;
 
-	mempool_t *r10bio_pool;
-	mempool_t *r10buf_pool;
+	mempool_t		*r10bio_pool;
+	mempool_t		*r10buf_pool;
 	struct page		*tmppage;
 
 	/* When taking over an array from a different personality, we store
@@ -98,11 +102,18 @@ struct r10bio {
 	 * When resyncing we also use one for each copy.
 	 * When reconstructing, we use 2 bios, one for read, one for write.
 	 * We choose the number when they are allocated.
+	 * We sometimes need an extra bio to write to the replacement.
 	 */
 	struct {
-		struct bio		*bio;
-		sector_t addr;
-		int devnum;
+		struct bio	*bio;
+		union {
+			struct bio	*repl_bio; /* used for resync and
+						    * writes */
+			struct md_rdev	*rdev;	   /* used for reads
+						    * (read_slot >= 0) */
+		};
+		sector_t	addr;
+		int		devnum;
 	} devs[0];
 };
 
@@ -121,17 +132,19 @@ struct r10bio {
 #define BIO_SPECIAL(bio) ((unsigned long)bio <= 2)
 
 /* bits for r10bio.state */
-#define	R10BIO_Uptodate	0
-#define	R10BIO_IsSync	1
-#define	R10BIO_IsRecover 2
-#define	R10BIO_Degraded 3
+enum r10bio_state {
+	R10BIO_Uptodate,
+	R10BIO_IsSync,
+	R10BIO_IsRecover,
+	R10BIO_Degraded,
 /* Set ReadError on bios that experience a read error
  * so that raid10d knows what to do with them.
  */
-#define	R10BIO_ReadError 4
+	R10BIO_ReadError,
 /* If a write for this request means we can clear some
  * known-bad-block records, we set this flag.
  */
-#define	R10BIO_MadeGood 5
-#define	R10BIO_WriteError 6
+	R10BIO_MadeGood,
+	R10BIO_WriteError,
+};
 #endif

drivers/md/raid5.h

@@ -27,7 +27,7 @@
  * The possible state transitions are:
  *
  *  Empty -> Want   - on read or write to get old data for  parity calc
- *  Empty -> Dirty  - on compute_parity to satisfy write/sync request.(RECONSTRUCT_WRITE)
+ *  Empty -> Dirty  - on compute_parity to satisfy write/sync request.
  *  Empty -> Clean  - on compute_block when computing a block for failed drive
  *  Want  -> Empty  - on failed read
  *  Want  -> Clean  - on successful completion of read request
@@ -226,8 +226,11 @@ struct stripe_head {
 		#endif
 	} ops;
 	struct r5dev {
-		struct bio	req;
-		struct bio_vec	vec;
+		/* rreq and rvec are used for the replacement device when
+		 * writing data to both devices.
+		 */
+		struct bio	req, rreq;
+		struct bio_vec	vec, rvec;
 		struct page	*page;
 		struct bio	*toread, *read, *towrite, *written;
 		sector_t	sector;			/* sector of this page */
@@ -239,7 +242,13 @@ struct stripe_head {
  *     for handle_stripe.
  */
 struct stripe_head_state {
-	int syncing, expanding, expanded;
+	/* 'syncing' means that we need to read all devices, either
+	 * to check/correct parity, or to reconstruct a missing device.
+	 * 'replacing' means we are replacing one or more drives and
+	 * the source is valid at this point so we don't need to
+	 * read all devices, just the replacement targets.
+	 */
+	int syncing, expanding, expanded, replacing;
 	int locked, uptodate, to_read, to_write, failed, written;
 	int to_fill, compute, req_compute, non_overwrite;
 	int failed_num[2];
@@ -252,38 +261,41 @@ struct stripe_head_state {
 	int handle_bad_blocks;
 };
 
-/* Flags */
-#define	R5_UPTODATE	0	/* page contains current data */
-#define	R5_LOCKED	1	/* IO has been submitted on "req" */
-#define	R5_OVERWRITE	2	/* towrite covers whole page */
+/* Flags for struct r5dev.flags */
+enum r5dev_flags {
+	R5_UPTODATE,	/* page contains current data */
+	R5_LOCKED,	/* IO has been submitted on "req" */
+	R5_DOUBLE_LOCKED,/* Cannot clear R5_LOCKED until 2 writes complete */
+	R5_OVERWRITE,	/* towrite covers whole page */
 /* and some that are internal to handle_stripe */
-#define	R5_Insync	3	/* rdev && rdev->in_sync at start */
-#define	R5_Wantread	4	/* want to schedule a read */
-#define	R5_Wantwrite	5
-#define	R5_Overlap	7	/* There is a pending overlapping request on this block */
-#define	R5_ReadError	8	/* seen a read error here recently */
-#define	R5_ReWrite	9	/* have tried to over-write the readerror */
+	R5_Insync,	/* rdev && rdev->in_sync at start */
+	R5_Wantread,	/* want to schedule a read */
+	R5_Wantwrite,
+	R5_Overlap,	/* There is a pending overlapping request
+			 * on this block */
+	R5_ReadError,	/* seen a read error here recently */
+	R5_ReWrite,	/* have tried to over-write the readerror */
 
-#define	R5_Expanded	10	/* This block now has post-expand data */
-#define	R5_Wantcompute	11	/* compute_block in progress treat as
-				 * uptodate
-				 */
-#define	R5_Wantfill	12	/* dev->toread contains a bio that needs
-				 * filling
-				 */
-#define	R5_Wantdrain	13	/* dev->towrite needs to be drained */
-#define	R5_WantFUA	14	/* Write should be FUA */
-#define	R5_WriteError	15	/* got a write error - need to record it */
-#define	R5_MadeGood	16	/* A bad block has been fixed by writing to it*/
-/*
- * Write method
- */
-#define RECONSTRUCT_WRITE	1
-#define READ_MODIFY_WRITE	2
-/* not a write method, but a compute_parity mode */
-#define	CHECK_PARITY		3
-/* Additional compute_parity mode -- updates the parity w/o LOCKING */
-#define UPDATE_PARITY		4
+	R5_Expanded,	/* This block now has post-expand data */
+	R5_Wantcompute,	/* compute_block in progress treat as
+			 * uptodate
+			 */
+	R5_Wantfill,	/* dev->toread contains a bio that needs
+			 * filling
+			 */
+	R5_Wantdrain,	/* dev->towrite needs to be drained */
+	R5_WantFUA,	/* Write should be FUA */
+	R5_WriteError,	/* got a write error - need to record it */
+	R5_MadeGood,	/* A bad block has been fixed by writing to it */
+	R5_ReadRepl,	/* Will/did read from replacement rather than orig */
+	R5_MadeGoodRepl,/* A bad block on the replacement device has been
+			 * fixed by writing to it */
+	R5_NeedReplace,	/* This device has a replacement which is not
+			 * up-to-date at this stripe. */
+	R5_WantReplace, /* We need to update the replacement, we have read
+			 * data in, and now is a good time to write it out.
+			 */
+};
 
 /*
  * Stripe state
@@ -311,13 +323,14 @@ enum {
 /*
  * Operation request flags
  */
-#define STRIPE_OP_BIOFILL	0
-#define STRIPE_OP_COMPUTE_BLK	1
-#define STRIPE_OP_PREXOR	2
-#define STRIPE_OP_BIODRAIN	3
-#define STRIPE_OP_RECONSTRUCT	4
-#define STRIPE_OP_CHECK	5
-
+enum {
+	STRIPE_OP_BIOFILL,
+	STRIPE_OP_COMPUTE_BLK,
+	STRIPE_OP_PREXOR,
+	STRIPE_OP_BIODRAIN,
+	STRIPE_OP_RECONSTRUCT,
+	STRIPE_OP_CHECK,
+};
 /*
  * Plugging:
  *
@@ -344,13 +357,12 @@ enum {
 
 
 struct disk_info {
-	struct md_rdev	*rdev;
+	struct md_rdev	*rdev, *replacement;
 };
 
 struct r5conf {
 	struct hlist_head	*stripe_hashtbl;
 	struct mddev		*mddev;
-	struct disk_info	*spare;
 	int			chunk_sectors;
 	int			level, algorithm;
 	int			max_degraded;

include/linux/raid/md_p.h

@@ -277,7 +277,10 @@ struct mdp_superblock_1 {
 					   */
 #define	MD_FEATURE_RESHAPE_ACTIVE	4
 #define	MD_FEATURE_BAD_BLOCKS		8 /* badblock list is not empty */
-
-#define	MD_FEATURE_ALL			(1|2|4|8)
+#define	MD_FEATURE_REPLACEMENT		16 /* This device is replacing an
+					    * active device with same 'role'.
+					    * 'recovery_offset' is also set.
+					    */
+#define	MD_FEATURE_ALL			(1|2|4|8|16)
 
 #endif 

include/linux/raid/pq.h

@@ -132,7 +132,7 @@ void raid6_dual_recov(int disks, size_t bytes, int faila, int failb,
 						     PROT_READ|PROT_WRITE,   \
 						     MAP_PRIVATE|MAP_ANONYMOUS,\
 						     0, 0))
-# define free_pages(x, y)	munmap((void *)(x), (y)*PAGE_SIZE)
+# define free_pages(x, y)	munmap((void *)(x), PAGE_SIZE << (y))
 
 static inline void cpu_relax(void)
 {