Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit e691063a authored by NeilBrown's avatar NeilBrown Committed by Linus Torvalds
Browse files

md: support 'external' metadata for md arrays 



- Add a state flag 'external' to indicate that the metadata is managed
  externally (by user-space) so important changes need to be
  left of user-space to handle.
  Alternates are non-persistant ('none') where there is no stable metadata -
  after the  array is stopped there is no record of it's status - and
  internal which can be version 0.90 or version 1.x
  These are selected by writing to the 'metadata' attribute.

- move the updating of superblocks (sync_sbs) to after we have checked if
  there are any superblocks or not.

- New array state 'write_pending'.  This means that the metadata records
  the array as 'clean', but a write has been requested, so the metadata has
  to be updated to record a 'dirty' array before the write can continue.
  This change is reported to md by writing 'active' to the array_state
  attribute.

- tidy up marking of sb_dirty:
   - don't set sb_dirty when resync finishes as md_check_recovery
     calls md_update_sb when the sync thread finishes anyway.
   - Don't set sb_dirty in multipath_run as the array might not be dirty.
   - don't mark superblock dirty when switching to 'clean' if there
     is no internal superblock (if external, userspace can choose to
     update the superblock whenever it chooses to).

Signed-off-by: default avatarNeil Brown <neilb@suse.de>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
parent b47490c9

drivers/md/md.c

+58 −19
Original line number Original line Diff line number Diff line
@@ -778,7 +778,8 @@ static int super_90_validate(mddev_t *mddev, mdk_rdev_t *rdev) 
		mddev->major_version = 0;

		mddev->major_version = 0;

		mddev->minor_version = sb->minor_version;

		mddev->minor_version = sb->minor_version;

		mddev->patch_version = sb->patch_version;

		mddev->patch_version = sb->patch_version;

		mddev->persistent = ! sb->not_persistent;

		mddev->persistent = 1;

		mddev->external = 0;

		mddev->chunk_size = sb->chunk_size;

		mddev->chunk_size = sb->chunk_size;

		mddev->ctime = sb->ctime;

		mddev->ctime = sb->ctime;

		mddev->utime = sb->utime;

		mddev->utime = sb->utime;
@@ -904,7 +905,7 @@ static void super_90_sync(mddev_t *mddev, mdk_rdev_t *rdev) 
	sb->size  = mddev->size;

	sb->size  = mddev->size;

	sb->raid_disks = mddev->raid_disks;

	sb->raid_disks = mddev->raid_disks;

	sb->md_minor = mddev->md_minor;

	sb->md_minor = mddev->md_minor;

	sb->not_persistent = !mddev->persistent;

	sb->not_persistent = 0;

	sb->utime = mddev->utime;

	sb->utime = mddev->utime;

	sb->state = 0;

	sb->state = 0;

	sb->events_hi = (mddev->events>>32);

	sb->events_hi = (mddev->events>>32);
@@ -1158,6 +1159,7 @@ static int super_1_validate(mddev_t *mddev, mdk_rdev_t *rdev) 
		mddev->major_version = 1;

		mddev->major_version = 1;

		mddev->patch_version = 0;

		mddev->patch_version = 0;

		mddev->persistent = 1;

		mddev->persistent = 1;

		mddev->external = 0;

		mddev->chunk_size = le32_to_cpu(sb->chunksize) << 9;

		mddev->chunk_size = le32_to_cpu(sb->chunksize) << 9;

		mddev->ctime = le64_to_cpu(sb->ctime) & ((1ULL << 32)-1);

		mddev->ctime = le64_to_cpu(sb->ctime) & ((1ULL << 32)-1);

		mddev->utime = le64_to_cpu(sb->utime) & ((1ULL << 32)-1);

		mddev->utime = le64_to_cpu(sb->utime) & ((1ULL << 32)-1);
@@ -1696,18 +1698,20 @@ repeat: 
		MD_BUG();

		MD_BUG();

		mddev->events --;

		mddev->events --;

	}

	}

	sync_sbs(mddev, nospares);





	/*

	/*

	 * do not write anything to disk if using

	 * do not write anything to disk if using

	 * nonpersistent superblocks

	 * nonpersistent superblocks

	 */

	 */

	if (!mddev->persistent) {

	if (!mddev->persistent) {

		if (!mddev->external)

			clear_bit(MD_CHANGE_PENDING, &mddev->flags);

			clear_bit(MD_CHANGE_PENDING, &mddev->flags);



		spin_unlock_irq(&mddev->write_lock);

		spin_unlock_irq(&mddev->write_lock);

		wake_up(&mddev->sb_wait);

		wake_up(&mddev->sb_wait);

		return;

		return;

	}

	}

	sync_sbs(mddev, nospares);

	spin_unlock_irq(&mddev->write_lock);

	spin_unlock_irq(&mddev->write_lock);





	dprintk(KERN_INFO 

	dprintk(KERN_INFO
@@ -2425,6 +2429,8 @@ array_state_show(mddev_t *mddev, char *page) 
		case 0:

		case 0:

			if (mddev->in_sync)

			if (mddev->in_sync)

				st = clean;

				st = clean;

			else if (test_bit(MD_CHANGE_CLEAN, &mddev->flags))

				st = write_pending;

			else if (mddev->safemode)

			else if (mddev->safemode)

				st = active_idle;

				st = active_idle;

			else

			else
@@ -2455,11 +2461,9 @@ array_state_store(mddev_t *mddev, const char *buf, size_t len) 
		break;

		break;

	case clear:

	case clear:

		/* stopping an active array */

		/* stopping an active array */

		if (mddev->pers) {

		if (atomic_read(&mddev->active) > 1)

		if (atomic_read(&mddev->active) > 1)

			return -EBUSY;

			return -EBUSY;

		err = do_md_stop(mddev, 0);

		err = do_md_stop(mddev, 0);

		}

		break;

		break;

	case inactive:

	case inactive:

		/* stopping an active array */

		/* stopping an active array */
@@ -2467,7 +2471,8 @@ array_state_store(mddev_t *mddev, const char *buf, size_t len) 
			if (atomic_read(&mddev->active) > 1)

			if (atomic_read(&mddev->active) > 1)

				return -EBUSY;

				return -EBUSY;

			err = do_md_stop(mddev, 2);

			err = do_md_stop(mddev, 2);

		}

		} else

			err = 0; /* already inactive */

		break;

		break;

	case suspended:

	case suspended:

		break; /* not supported yet */

		break; /* not supported yet */
@@ -2495,9 +2500,15 @@ array_state_store(mddev_t *mddev, const char *buf, size_t len) 
			restart_array(mddev);

			restart_array(mddev);

			spin_lock_irq(&mddev->write_lock);

			spin_lock_irq(&mddev->write_lock);

			if (atomic_read(&mddev->writes_pending) == 0) {

			if (atomic_read(&mddev->writes_pending) == 0) {

				if (mddev->in_sync == 0) {

					mddev->in_sync = 1;

					mddev->in_sync = 1;

				set_bit(MD_CHANGE_CLEAN, &mddev->flags);

					if (mddev->persistent)

						set_bit(MD_CHANGE_CLEAN,

							&mddev->flags);

				}

				}

				err = 0;

			} else

				err = -EBUSY;

			spin_unlock_irq(&mddev->write_lock);

			spin_unlock_irq(&mddev->write_lock);

		} else {

		} else {

			mddev->ro = 0;

			mddev->ro = 0;
@@ -2508,6 +2519,7 @@ array_state_store(mddev_t *mddev, const char *buf, size_t len) 
	case active:

	case active:

		if (mddev->pers) {

		if (mddev->pers) {

			restart_array(mddev);

			restart_array(mddev);

			if (mddev->external)

				clear_bit(MD_CHANGE_CLEAN, &mddev->flags);

				clear_bit(MD_CHANGE_CLEAN, &mddev->flags);

			wake_up(&mddev->sb_wait);

			wake_up(&mddev->sb_wait);

			err = 0;

			err = 0;
@@ -2659,7 +2671,9 @@ __ATTR(component_size, S_IRUGO|S_IWUSR, size_show, size_store); 








/* Metdata version.

/* Metdata version.

 * This is either 'none' for arrays with externally managed metadata,

 * This is one of

 *   'none' for arrays with no metadata (good luck...)

 *   'external' for arrays with externally managed metadata,

 * or N.M for internally known formats

 * or N.M for internally known formats

 */

 */

static ssize_t

static ssize_t
@@ -2668,6 +2682,8 @@ metadata_show(mddev_t *mddev, char *page) 
	if (mddev->persistent)

	if (mddev->persistent)

		return sprintf(page, "%d.%d\n",

		return sprintf(page, "%d.%d\n",

			       mddev->major_version, mddev->minor_version);

			       mddev->major_version, mddev->minor_version);

	else if (mddev->external)

		return sprintf(page, "external:%s\n", mddev->metadata_type);

	else

	else

		return sprintf(page, "none\n");

		return sprintf(page, "none\n");

}

}
@@ -2682,6 +2698,21 @@ metadata_store(mddev_t *mddev, const char *buf, size_t len) 




	if (cmd_match(buf, "none")) {

	if (cmd_match(buf, "none")) {

		mddev->persistent = 0;

		mddev->persistent = 0;

		mddev->external = 0;

		mddev->major_version = 0;

		mddev->minor_version = 90;

		return len;

	}

	if (strncmp(buf, "external:", 9) == 0) {

		int namelen = len-9;

		if (namelen >= sizeof(mddev->metadata_type))

			namelen = sizeof(mddev->metadata_type)-1;

		strncpy(mddev->metadata_type, buf+9, namelen);

		mddev->metadata_type[namelen] = 0;

		if (namelen && mddev->metadata_type[namelen-1] == '\n')

			mddev->metadata_type[--namelen] = 0;

		mddev->persistent = 0;

		mddev->external = 1;

		mddev->major_version = 0;

		mddev->major_version = 0;

		mddev->minor_version = 90;

		mddev->minor_version = 90;

		return len;

		return len;
@@ -2698,6 +2729,7 @@ metadata_store(mddev_t *mddev, const char *buf, size_t len) 
	mddev->major_version = major;

	mddev->major_version = major;

	mddev->minor_version = minor;

	mddev->minor_version = minor;

	mddev->persistent = 1;

	mddev->persistent = 1;

	mddev->external = 0;

	return len;

	return len;

}

}
@@ -3524,6 +3556,7 @@ static int do_md_stop(mddev_t * mddev, int mode) 
		mddev->raid_disks = 0;

		mddev->raid_disks = 0;

		mddev->recovery_cp = 0;

		mddev->recovery_cp = 0;

		mddev->reshape_position = MaxSector;

		mddev->reshape_position = MaxSector;

		mddev->external = 0;





	} else if (mddev->pers)

	} else if (mddev->pers)

		printk(KERN_INFO "md: %s switched to read-only mode.\n",

		printk(KERN_INFO "md: %s switched to read-only mode.\n",
@@ -4165,12 +4198,14 @@ static int set_array_info(mddev_t * mddev, mdu_array_info_t *info) 
	else

	else

		mddev->recovery_cp = 0;

		mddev->recovery_cp = 0;

	mddev->persistent    = ! info->not_persistent;

	mddev->persistent    = ! info->not_persistent;

	mddev->external	     = 0;





	mddev->layout        = info->layout;

	mddev->layout        = info->layout;

	mddev->chunk_size    = info->chunk_size;

	mddev->chunk_size    = info->chunk_size;





	mddev->max_disks     = MD_SB_DISKS;

	mddev->max_disks     = MD_SB_DISKS;





	if (mddev->persistent)

		mddev->flags         = 0;

		mddev->flags         = 0;

	set_bit(MD_CHANGE_DEVS, &mddev->flags);

	set_bit(MD_CHANGE_DEVS, &mddev->flags);
@@ -4982,7 +5017,10 @@ static int md_seq_show(struct seq_file *seq, void *v) 
					   mddev->major_version,

					   mddev->major_version,

					   mddev->minor_version);

					   mddev->minor_version);

			}

			}

		} else

		} else if (mddev->external)

			seq_printf(seq, " super external:%s",

				   mddev->metadata_type);

		else

			seq_printf(seq, " super non-persistent");

			seq_printf(seq, " super non-persistent");





		if (mddev->pers) {

		if (mddev->pers) {
@@ -5589,7 +5627,7 @@ void md_check_recovery(mddev_t *mddev) 
	}

	}





	if ( ! (

	if ( ! (

		mddev->flags ||

		(mddev->flags && !mddev->external) ||

		test_bit(MD_RECOVERY_NEEDED, &mddev->recovery) ||

		test_bit(MD_RECOVERY_NEEDED, &mddev->recovery) ||

		test_bit(MD_RECOVERY_DONE, &mddev->recovery) ||

		test_bit(MD_RECOVERY_DONE, &mddev->recovery) ||

		(mddev->safemode == 1) ||

		(mddev->safemode == 1) ||
@@ -5605,6 +5643,7 @@ void md_check_recovery(mddev_t *mddev) 
		if (mddev->safemode && !atomic_read(&mddev->writes_pending) &&

		if (mddev->safemode && !atomic_read(&mddev->writes_pending) &&

		    !mddev->in_sync && mddev->recovery_cp == MaxSector) {

		    !mddev->in_sync && mddev->recovery_cp == MaxSector) {

			mddev->in_sync = 1;

			mddev->in_sync = 1;

			if (mddev->persistent)

				set_bit(MD_CHANGE_CLEAN, &mddev->flags);

				set_bit(MD_CHANGE_CLEAN, &mddev->flags);

		}

		}

		if (mddev->safemode == 1)

		if (mddev->safemode == 1)

include/linux/raid/md_k.h

+3 −0
Original line number Original line Diff line number Diff line
@@ -130,6 +130,9 @@ struct mddev_s 
					minor_version,

					minor_version,

					patch_version;

					patch_version;

	int				persistent;

	int				persistent;

	int 				external;	/* metadata is

							 * managed externally */

	char				metadata_type[17]; /* externally set*/

	int				chunk_size;

	int				chunk_size;

	time_t				ctime, utime;

	time_t				ctime, utime;

	int				level, layout;

	int				level, layout;