Commit 2943c833 authored by Linus Torvalds's avatar Linus Torvalds
Browse files

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.
  ...
parents 98793265 19d67169

Documentation/md.txt

+18 −4
Original line number Diff line number Diff line
@@ -360,7 +360,7 @@ Each directory contains: 
	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

			 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
@@ -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

+6 −6
Original line number Diff line number Diff line
@@ -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

+80 −27
Original line number Diff line number Diff line
@@ -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;



	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;

		}
@@ -7337,13 +7392,12 @@ 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) &&
@@ -7362,7 +7416,6 @@ static int remove_and_add_spares(struct mddev *mddev) 
			}

		}

	}

	}

	return spares;

}
@@ -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

+49 −33
Original line number Diff line number Diff line
@@ -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);
@@ -482,16 +493,21 @@ static inline char * mdname (struct mddev * mddev) 
static inline int sysfs_link_rdev(struct mddev *mddev, struct md_rdev *rdev)

{

	char nm[20];

	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];

	if (!test_bit(Replacement, &rdev->flags)) {

		sprintf(nm, "rd%d", rdev->raid_disk);

		sysfs_remove_link(&mddev->kobj, nm);

	}

}



/*

 * iterates through some rdev ringlist. It's safe to remove the

drivers/md/multipath.c

+3 −4
Original line number Diff line number Diff line
@@ -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"

CRA Git | Maintained and supported by SUSTech CRA and CCSE