[PATCH] md: yet another attempt to get bitmap-based resync to do the right thing...
authorNeilBrown <neilb@cse.unsw.edu.au>
Thu, 4 Aug 2005 19:53:34 +0000 (12:53 -0700)
committerLinus Torvalds <torvalds@g5.osdl.org>
Thu, 4 Aug 2005 20:00:54 +0000 (13:00 -0700)
Firstly, R1BIO_Degraded was being set in a number of places in the resync
code, but is never used there, so get rid of those settings.

Then: When doing a resync, we want to clear the bit in the bitmap iff the
array will be non-degraded when the sync has completed.  However the current
code would clear the bitmap if the array was non-degraded when the resync
*started*, which obviously isn't right (it is for 'resync' but not for
'recovery' - i.e.  rebuilding a failed drive).

This patch calculated 'still_degraded' and uses the to tell bitmap_start_sync
whether this sync should clear the corresponding bit.

Signed-off-by: Neil Brown <neilb@cse.unsw.edu.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
drivers/md/raid1.c

index d3a64a04a6d857558e16b54e3794fb7830e5f2ec..51d9645ed09c5e8ff79aa9e7a6604e94dc274a54 100644 (file)
@@ -893,7 +893,6 @@ static int end_sync_read(struct bio *bio, unsigned int bytes_done, int error)
        if (!uptodate) {
                md_error(r1_bio->mddev,
                         conf->mirrors[r1_bio->read_disk].rdev);
-               set_bit(R1BIO_Degraded, &r1_bio->state);
        } else
                set_bit(R1BIO_Uptodate, &r1_bio->state);
        rdev_dec_pending(conf->mirrors[r1_bio->read_disk].rdev, conf->mddev);
@@ -918,10 +917,9 @@ static int end_sync_write(struct bio *bio, unsigned int bytes_done, int error)
                        mirror = i;
                        break;
                }
-       if (!uptodate) {
+       if (!uptodate)
                md_error(mddev, conf->mirrors[mirror].rdev);
-               set_bit(R1BIO_Degraded, &r1_bio->state);
-       }
+
        update_head_pos(mirror, r1_bio);
 
        if (atomic_dec_and_test(&r1_bio->remaining)) {
@@ -1109,6 +1107,7 @@ static sector_t sync_request(mddev_t *mddev, sector_t sector_nr, int *skipped, i
        int i;
        int write_targets = 0;
        int sync_blocks;
+       int still_degraded = 0;
 
        if (!conf->r1buf_pool)
        {
@@ -1137,7 +1136,10 @@ static sector_t sync_request(mddev_t *mddev, sector_t sector_nr, int *skipped, i
                return 0;
        }
 
-       if (!bitmap_start_sync(mddev->bitmap, sector_nr, &sync_blocks, mddev->degraded) &&
+       /* before building a request, check if we can skip these blocks..
+        * This call the bitmap_start_sync doesn't actually record anything
+        */
+       if (!bitmap_start_sync(mddev->bitmap, sector_nr, &sync_blocks, 1) &&
            !conf->fullsync) {
                /* We can skip this block, and probably several more */
                *skipped = 1;
@@ -1203,24 +1205,23 @@ static sector_t sync_request(mddev_t *mddev, sector_t sector_nr, int *skipped, i
                if (i == disk) {
                        bio->bi_rw = READ;
                        bio->bi_end_io = end_sync_read;
-               } else if (conf->mirrors[i].rdev &&
-                          !conf->mirrors[i].rdev->faulty &&
-                          (!conf->mirrors[i].rdev->in_sync ||
-                           sector_nr + RESYNC_SECTORS > mddev->recovery_cp)) {
+               } else if (conf->mirrors[i].rdev == NULL ||
+                          conf->mirrors[i].rdev->faulty) {
+                       still_degraded = 1;
+                       continue;
+               } else if (!conf->mirrors[i].rdev->in_sync ||
+                          sector_nr + RESYNC_SECTORS > mddev->recovery_cp) {
                        bio->bi_rw = WRITE;
                        bio->bi_end_io = end_sync_write;
                        write_targets ++;
                } else
+                       /* no need to read or write here */
                        continue;
                bio->bi_sector = sector_nr + conf->mirrors[i].rdev->data_offset;
                bio->bi_bdev = conf->mirrors[i].rdev->bdev;
                bio->bi_private = r1_bio;
        }
 
-       if (write_targets + 1 < conf->raid_disks)
-               /* array degraded, can't clear bitmap */
-               set_bit(R1BIO_Degraded, &r1_bio->state);
-
        if (write_targets == 0) {
                /* There is nowhere to write, so all non-sync
                 * drives must be failed - so we are finished
@@ -1243,7 +1244,7 @@ static sector_t sync_request(mddev_t *mddev, sector_t sector_nr, int *skipped, i
                        break;
                if (sync_blocks == 0) {
                        if (!bitmap_start_sync(mddev->bitmap, sector_nr,
-                                       &sync_blocks, mddev->degraded) &&
+                                       &sync_blocks, still_degraded) &&
                                        !conf->fullsync)
                                break;
                        if (sync_blocks < (PAGE_SIZE>>9))