dm raid: use read_disk_sb() throughout
authorHeinz Mauelshagen <heinzm@redhat.com>
Sat, 14 Jan 2017 02:53:10 +0000 (03:53 +0100)
committerMike Snitzer <snitzer@redhat.com>
Wed, 25 Jan 2017 11:49:07 +0000 (12:49 +0100)
For consistency, call read_disk_sb() from
attempt_restore_of_faulty_devices() instead
of calling sync_page_io() directly.

Explicitly set device to faulty on superblock read error.

Signed-off-by: Heinz Mauelshagen <heinzm@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
drivers/md/dm-raid.c

index e52c493212d05a0a20a748d161b636bbccd2e933..c982649497a5fc8327bf618ae35195b3b628a94d 100644 (file)
@@ -1936,18 +1936,21 @@ static int rs_check_reshape(struct raid_set *rs)
        return -EPERM;
 }
 
-static int read_disk_sb(struct md_rdev *rdev, int size)
+static int read_disk_sb(struct md_rdev *rdev, int size, bool force_reload)
 {
        BUG_ON(!rdev->sb_page);
 
-       if (rdev->sb_loaded)
+       if (rdev->sb_loaded && !force_reload)
                return 0;
 
+       rdev->sb_loaded = 0;
+
        if (!sync_page_io(rdev, 0, size, rdev->sb_page, REQ_OP_READ, 0, true)) {
                DMERR("Failed to read superblock of device at position %d",
                      rdev->raid_disk);
                md_error(rdev->mddev, rdev);
-               return -EINVAL;
+               set_bit(Faulty, &rdev->flags);
+               return -EIO;
        }
 
        rdev->sb_loaded = 1;
@@ -2075,7 +2078,7 @@ static int super_load(struct md_rdev *rdev, struct md_rdev *refdev)
                return -EINVAL;
        }
 
-       r = read_disk_sb(rdev, rdev->sb_size);
+       r = read_disk_sb(rdev, rdev->sb_size, false);
        if (r)
                return r;
 
@@ -2453,7 +2456,6 @@ static int analyse_superblocks(struct dm_target *ti, struct raid_set *rs)
                         * The rdev has to stay on the same_set list to allow for
                         * the attempt to restore faulty devices on second resume.
                         */
-                       set_bit(Faulty, &rdev->flags);
                        rdev->raid_disk = rdev->saved_raid_disk = -1;
                        break;
                }
@@ -3548,9 +3550,8 @@ static void attempt_restore_of_faulty_devices(struct raid_set *rs)
                if (test_bit(Journal, &r->flags))
                        continue;
 
-               if (test_bit(Faulty, &r->flags) && r->sb_page &&
-                   sync_page_io(r, 0, r->sb_size, r->sb_page,
-                                REQ_OP_READ, 0, true)) {
+               if (test_bit(Faulty, &r->flags) &&
+                   r->meta_bdev && !read_disk_sb(r, r->sb_size, true)) {
                        DMINFO("Faulty %s device #%d has readable super block."
                               "  Attempting to revive it.",
                               rs->raid_type->name, i);