md: restart recovery cleanly after device failure.

[linux-2.6-omap-h63xx.git] / drivers / md / raid10.c

diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index 8536ede1e7129c37fc30303e15577c383e68569c..1de17da34a956e8383efb1b42474c3c4d3d15dbe 100644 (file)
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -1020,7 +1020,7 @@ static void error(mddev_t *mddev, mdk_rdev_t *rdev)
                /*
                 * if recovery is running, make sure it aborts.
                 */
-               set_bit(MD_RECOVERY_ERR, &mddev->recovery);
+               set_bit(MD_RECOVERY_INTR, &mddev->recovery);
        }
        set_bit(Faulty, &rdev->flags);
        set_bit(MD_CHANGE_DEVS, &mddev->flags);
@@ -1171,6 +1171,14 @@ static int raid10_remove_disk(mddev_t *mddev, int number)
                        err = -EBUSY;
                        goto abort;
                }
+               /* Only remove faulty devices in recovery
+                * is not possible.
+                */
+               if (!test_bit(Faulty, &rdev->flags) &&
+                   enough(conf)) {
+                       err = -EBUSY;
+                       goto abort;
+               }
                p->rdev = NULL;
                synchronize_rcu();
                if (atomic_read(&rdev->nr_pending)) {
@@ -1237,6 +1245,7 @@ static void end_sync_write(struct bio *bio, int error)
 
        if (!uptodate)
                md_error(mddev, conf->mirrors[d].rdev);
+
        update_head_pos(i, r10_bio);
 
        while (atomic_dec_and_test(&r10_bio->remaining)) {
@@ -1844,7 +1853,8 @@ static sector_t sync_request(mddev_t *mddev, sector_t sector_nr, int *skipped, i
                                        if (rb2)
                                                atomic_dec(&rb2->remaining);
                                        r10_bio = rb2;
-                                       if (!test_and_set_bit(MD_RECOVERY_ERR, &mddev->recovery))
+                                       if (!test_and_set_bit(MD_RECOVERY_INTR,
+                                                             &mddev->recovery))
                                                printk(KERN_INFO "raid10: %s: insufficient working devices for recovery.\n",
                                                       mdname(mddev));
                                        break;