1 files changed, 121 insertions, 43 deletions

diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index 36a8fc0..9f7f8be 100644
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -46,6 +46,20 @@
  */
 #define	NR_RAID1_BIOS 256
 
+/* when we get a read error on a read-only array, we redirect to another
+ * device without failing the first device, or trying to over-write to
+ * correct the read error.  To keep track of bad blocks on a per-bio
+ * level, we store IO_BLOCKED in the appropriate 'bios' pointer
+ */
+#define IO_BLOCKED ((struct bio *)1)
+/* When we successfully write to a known bad-block, we need to remove the
+ * bad-block marking which must be done from process context.  So we record
+ * the success by setting devs[n].bio to IO_MADE_GOOD
+ */
+#define IO_MADE_GOOD ((struct bio *)2)
+
+#define BIO_SPECIAL(bio) ((unsigned long)bio <= 2)
+
 /* When there are this many requests queue to be written by
  * the raid1 thread, we become 'congested' to provide back-pressure
  * for writeback.
@@ -483,12 +497,14 @@ static int read_balance(struct r1conf *conf, struct r1bio *r1_bio, int *max_sect
 	const sector_t this_sector = r1_bio->sector;
 	int sectors;
 	int best_good_sectors;
-	int start_disk;
-	int best_disk;
-	int i;
+	int best_disk, best_dist_disk, best_pending_disk;
+	int has_nonrot_disk;
+	int disk;
 	sector_t best_dist;
+	unsigned int min_pending;
 	struct md_rdev *rdev;
 	int choose_first;
+	int choose_next_idle;
 
 	rcu_read_lock();
 	/*
@@ -499,26 +515,26 @@ static int read_balance(struct r1conf *conf, struct r1bio *r1_bio, int *max_sect
  retry:
 	sectors = r1_bio->sectors;
 	best_disk = -1;
+	best_dist_disk = -1;
 	best_dist = MaxSector;
+	best_pending_disk = -1;
+	min_pending = UINT_MAX;
 	best_good_sectors = 0;
+	has_nonrot_disk = 0;
+	choose_next_idle = 0;
 
 	if (conf->mddev->recovery_cp < MaxSector &&
-	    (this_sector + sectors >= conf->next_resync)) {
+	    (this_sector + sectors >= conf->next_resync))
 		choose_first = 1;
-		start_disk = 0;
-	} else {
+	else
 		choose_first = 0;
-		start_disk = conf->last_used;
-	}
 
-	for (i = 0 ; i < conf->raid_disks * 2 ; i++) {
+	for (disk = 0 ; disk < conf->raid_disks * 2 ; disk++) {
 		sector_t dist;
 		sector_t first_bad;
 		int bad_sectors;
-
-		int disk = start_disk + i;
-		if (disk >= conf->raid_disks * 2)
-			disk -= conf->raid_disks * 2;
+		unsigned int pending;
+		bool nonrot;
 
 		rdev = rcu_dereference(conf->mirrors[disk].rdev);
 		if (r1_bio->bios[disk] == IO_BLOCKED
@@ -577,22 +593,77 @@ static int read_balance(struct r1conf *conf, struct r1bio *r1_bio, int *max_sect
 		} else
 			best_good_sectors = sectors;
 
+		nonrot = blk_queue_nonrot(bdev_get_queue(rdev->bdev));
+		has_nonrot_disk |= nonrot;
+		pending = atomic_read(&rdev->nr_pending);
 		dist = abs(this_sector - conf->mirrors[disk].head_position);
-		if (choose_first
-		    /* Don't change to another disk for sequential reads */
-		    || conf->next_seq_sect == this_sector
-		    || dist == 0
-		    /* If device is idle, use it */
-		    || atomic_read(&rdev->nr_pending) == 0) {
+		if (choose_first) {
+			best_disk = disk;
+			break;
+		}
+		/* Don't change to another disk for sequential reads */
+		if (conf->mirrors[disk].next_seq_sect == this_sector
+		    || dist == 0) {
+			int opt_iosize = bdev_io_opt(rdev->bdev) >> 9;
+			struct raid1_info *mirror = &conf->mirrors[disk];
+
+			best_disk = disk;
+			/*
+			 * If buffered sequential IO size exceeds optimal
+			 * iosize, check if there is idle disk. If yes, choose
+			 * the idle disk. read_balance could already choose an
+			 * idle disk before noticing it's a sequential IO in
+			 * this disk. This doesn't matter because this disk
+			 * will idle, next time it will be utilized after the
+			 * first disk has IO size exceeds optimal iosize. In
+			 * this way, iosize of the first disk will be optimal
+			 * iosize at least. iosize of the second disk might be
+			 * small, but not a big deal since when the second disk
+			 * starts IO, the first disk is likely still busy.
+			 */
+			if (nonrot && opt_iosize > 0 &&
+			    mirror->seq_start != MaxSector &&
+			    mirror->next_seq_sect > opt_iosize &&
+			    mirror->next_seq_sect - opt_iosize >=
+			    mirror->seq_start) {
+				choose_next_idle = 1;
+				continue;
+			}
+			break;
+		}
+		/* If device is idle, use it */
+		if (pending == 0) {
 			best_disk = disk;
 			break;
 		}
+
+		if (choose_next_idle)
+			continue;
+
+		if (min_pending > pending) {
+			min_pending = pending;
+			best_pending_disk = disk;
+		}
+
 		if (dist < best_dist) {
 			best_dist = dist;
-			best_disk = disk;
+			best_dist_disk = disk;
 		}
 	}
 
+	/*
+	 * If all disks are rotational, choose the closest disk. If any disk is
+	 * non-rotational, choose the disk with less pending request even the
+	 * disk is rotational, which might/might not be optimal for raids with
+	 * mixed ratation/non-rotational disks depending on workload.
+	 */
+	if (best_disk == -1) {
+		if (has_nonrot_disk)
+			best_disk = best_pending_disk;
+		else
+			best_disk = best_dist_disk;
+	}
+
 	if (best_disk >= 0) {
 		rdev = rcu_dereference(conf->mirrors[best_disk].rdev);
 		if (!rdev)
@@ -606,8 +677,11 @@ static int read_balance(struct r1conf *conf, struct r1bio *r1_bio, int *max_sect
 			goto retry;
 		}
 		sectors = best_good_sectors;
-		conf->next_seq_sect = this_sector + sectors;
-		conf->last_used = best_disk;
+
+		if (conf->mirrors[best_disk].next_seq_sect != this_sector)
+			conf->mirrors[best_disk].seq_start = this_sector;
+
+		conf->mirrors[best_disk].next_seq_sect = this_sector + sectors;
 	}
 	rcu_read_unlock();
 	*max_sectors = sectors;
@@ -873,7 +947,7 @@ do_sync_io:
 static void make_request(struct mddev *mddev, struct bio * bio)
 {
 	struct r1conf *conf = mddev->private;
-	struct mirror_info *mirror;
+	struct raid1_info *mirror;
 	struct r1bio *r1_bio;
 	struct bio *read_bio;
 	int i, disks;
@@ -1364,7 +1438,7 @@ static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev)
 	struct r1conf *conf = mddev->private;
 	int err = -EEXIST;
 	int mirror = 0;
-	struct mirror_info *p;
+	struct raid1_info *p;
 	int first = 0;
 	int last = conf->raid_disks - 1;
 	struct request_queue *q = bdev_get_queue(rdev->bdev);
@@ -1433,7 +1507,7 @@ static int raid1_remove_disk(struct mddev *mddev, struct md_rdev *rdev)
 	struct r1conf *conf = mddev->private;
 	int err = 0;
 	int number = rdev->raid_disk;
-	struct mirror_info *p = conf->mirrors+ number;
+	struct raid1_info *p = conf->mirrors + number;
 
 	if (rdev != p->rdev)
 		p = conf->mirrors + conf->raid_disks + number;
@@ -2370,6 +2444,18 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, int *skipp
 				bio->bi_rw = READ;
 				bio->bi_end_io = end_sync_read;
 				read_targets++;
+			} else if (!test_bit(WriteErrorSeen, &rdev->flags) &&
+				test_bit(MD_RECOVERY_SYNC, &mddev->recovery) &&
+				!test_bit(MD_RECOVERY_CHECK, &mddev->recovery)) {
+				/*
+				 * The device is suitable for reading (InSync),
+				 * but has bad block(s) here. Let's try to correct them,
+				 * if we are doing resync or repair. Otherwise, leave
+				 * this device alone for this sync request.
+				 */
+				bio->bi_rw = WRITE;
+				bio->bi_end_io = end_sync_write;
+				write_targets++;
 			}
 		}
 		if (bio->bi_end_io) {
@@ -2427,7 +2513,10 @@ static sector_t sync_request(struct mddev *mddev, sector_t sector_nr, int *skipp
 		/* There is nowhere to write, so all non-sync
 		 * drives must be failed - so we are finished
 		 */
-		sector_t rv = max_sector - sector_nr;
+		sector_t rv;
+		if (min_bad > 0)
+			max_sector = sector_nr + min_bad;
+		rv = max_sector - sector_nr;
 		*skipped = 1;
 		put_buf(r1_bio);
 		return rv;
@@ -2520,7 +2609,7 @@ static struct r1conf *setup_conf(struct mddev *mddev)
 {
 	struct r1conf *conf;
 	int i;
-	struct mirror_info *disk;
+	struct raid1_info *disk;
 	struct md_rdev *rdev;
 	int err = -ENOMEM;
 
@@ -2528,7 +2617,7 @@ static struct r1conf *setup_conf(struct mddev *mddev)
 	if (!conf)
 		goto abort;
 
-	conf->mirrors = kzalloc(sizeof(struct mirror_info)
+	conf->mirrors = kzalloc(sizeof(struct raid1_info)
 				* mddev->raid_disks * 2,
 				 GFP_KERNEL);
 	if (!conf->mirrors)
@@ -2571,6 +2660,7 @@ static struct r1conf *setup_conf(struct mddev *mddev)
 			mddev->merge_check_needed = 1;
 
 		disk->head_position = 0;
+		disk->seq_start = MaxSector;
 	}
 	conf->raid_disks = mddev->raid_disks;
 	conf->mddev = mddev;
@@ -2584,7 +2674,6 @@ static struct r1conf *setup_conf(struct mddev *mddev)
 	conf->recovery_disabled = mddev->recovery_disabled - 1;
 
 	err = -EIO;
-	conf->last_used = -1;
 	for (i = 0; i < conf->raid_disks * 2; i++) {
 
 		disk = conf->mirrors + i;
@@ -2610,19 +2699,9 @@ static struct r1conf *setup_conf(struct mddev *mddev)
 			if (disk->rdev &&
 			    (disk->rdev->saved_raid_disk < 0))
 				conf->fullsync = 1;
-		} else if (conf->last_used < 0)
-			/*
-			 * The first working device is used as a
-			 * starting point to read balancing.
-			 */
-			conf->last_used = i;
+		}
 	}
 
-	if (conf->last_used < 0) {
-		printk(KERN_ERR "md/raid1:%s: no operational mirrors\n",
-		       mdname(mddev));
-		goto abort;
-	}
 	err = -ENOMEM;
 	conf->thread = md_register_thread(raid1d, mddev, "raid1");
 	if (!conf->thread) {
@@ -2797,7 +2876,7 @@ static int raid1_reshape(struct mddev *mddev)
 	 */
 	mempool_t *newpool, *oldpool;
 	struct pool_info *newpoolinfo;
-	struct mirror_info *newmirrors;
+	struct raid1_info *newmirrors;
 	struct r1conf *conf = mddev->private;
 	int cnt, raid_disks;
 	unsigned long flags;
@@ -2840,7 +2919,7 @@ static int raid1_reshape(struct mddev *mddev)
 		kfree(newpoolinfo);
 		return -ENOMEM;
 	}
-	newmirrors = kzalloc(sizeof(struct mirror_info) * raid_disks * 2,
+	newmirrors = kzalloc(sizeof(struct raid1_info) * raid_disks * 2,
 			     GFP_KERNEL);
 	if (!newmirrors) {
 		kfree(newpoolinfo);
@@ -2879,7 +2958,6 @@ static int raid1_reshape(struct mddev *mddev)
 	conf->raid_disks = mddev->raid_disks = raid_disks;
 	mddev->delta_disks = 0;
 
-	conf->last_used = 0; /* just make sure it is in-range */
 	lower_barrier(conf);
 
 	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);