block: fix a possible oops on elv_abort_queue()
authorKiyoshi Ueda <k-ueda@ct.jp.nec.com>
Tue, 2 Jun 2009 06:44:01 +0000 (08:44 +0200)
committerJens Axboe <jens.axboe@oracle.com>
Tue, 2 Jun 2009 06:44:01 +0000 (08:44 +0200)
I found one more mis-conversion to the 'request is always dequeued
when completing' model in elv_abort_queue() during code inspection.
Although I haven't hit any problem caused by this mis-conversion yet
and just done compile/boot test, please apply if you have no problem.

Request must be dequeued when it completes.
However, elv_abort_queue() completes requests without dequeueing.
This will cause oops in the __blk_end_request_all().
This patch fixes the oops.

Signed-off-by: Kiyoshi Ueda <k-ueda@ct.jp.nec.com>
Signed-off-by: Jun'ichi Nomura <j-nomura@ce.jp.nec.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
block/elevator.c

index bd78f69..a029cfe 100644 (file)
@@ -813,6 +813,11 @@ void elv_abort_queue(struct request_queue *q)
                rq = list_entry_rq(q->queue_head.next);
                rq->cmd_flags |= REQ_QUIET;
                trace_block_rq_abort(q, rq);
+               /*
+                * Mark this request as started so we don't trigger
+                * any debug logic in the end I/O path.
+                */
+               blk_start_request(rq);
                __blk_end_request_all(rq, -EIO);
        }
 }