md/raid5: add missing call to schedule() after prepare_to_wait()
authorDan Williams <dan.j.williams@intel.com>
Tue, 16 Jun 2009 23:00:33 +0000 (16:00 -0700)
committerNeilBrown <neilb@suse.de>
Wed, 17 Jun 2009 22:50:18 +0000 (08:50 +1000)
In the unlikely event that reshape progresses past the current request
while it is waiting for a stripe we need to schedule() before retrying
for 2 reasons:
1/ Prevent list corruption from duplicated list_add() calls without
   intervening list_del().
2/ Give the reshape code a chance to make some progress to resolve the
   conflict.

Cc: <stable@kernel.org>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: NeilBrown <neilb@suse.de>
drivers/md/raid5.c

index 8f2e374..59e29c2 100644 (file)
@@ -3695,6 +3695,7 @@ static int make_request(struct request_queue *q, struct bio * bi)
                                spin_unlock_irq(&conf->device_lock);
                                if (must_retry) {
                                        release_stripe(sh);
+                                       schedule();
                                        goto retry;
                                }
                        }