From: Maciej Trela <maciej.trela@xxxxxxxxx>
While raid10->raid0 takeover incorrect slot numbers are set for raid0. This was caused by invalidating dev slot numbers before calling pers->run(). We are invalidating all devs that has (rdev->raid_disk >= mddev->raid_disks) but rdev->raid_disk numbers are scaled in pers->run().
Another option to fix the problem:
If invalidating devs should rather be performed before run() we could use devlist[index] for invalidating - devlist is already updated after calling takeover...
---
drivers/md/md.c | 14 ++++++++------
1 files changed, 8 insertions(+), 6 deletions(-)
diff --git a/drivers/md/md.c b/drivers/md/md.c index 46b3a04..effe601 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -3052,12 +3052,6 @@ level_store(mddev_t *mddev, const char *buf, size_t len)
}
module_put(mddev->pers->owner);
- /* Invalidate devices that are now superfluous */
- list_for_each_entry(rdev, &mddev->disks, same_set)
- if (rdev->raid_disk >= mddev->raid_disks) {
- rdev->raid_disk = -1;
- clear_bit(In_sync, &rdev->flags);
- }
mddev->pers = pers;
mddev->private = priv;
strlcpy(mddev->clevel, pers->name, sizeof(mddev->clevel)); @@ -3072,7 +3066,15 @@ level_store(mddev_t *mddev, const char *buf, size_t len)
mddev->in_sync = 1;
del_timer_sync(&mddev->safemode_timer);
}
+
pers->run(mddev);
+
+ /* Invalidate devices that are now superfluous */
+ list_for_each_entry(rdev, &mddev->disks, same_set)
+ if (rdev->raid_disk >= mddev->raid_disks) {
+ rdev->raid_disk = -1;
+ clear_bit(In_sync, &rdev->flags);
+ }
mddev_resume(mddev);
set_bit(MD_CHANGE_DEVS, &mddev->flags);
set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
