| Date | Sun, 05 Feb 2012 23:11:06 +0100 | From | Willy Tarreau <> | Subject | [PATCH 77/91] block: add proper state guards to __elv_next_request |
| |
2.6.27-longterm review patch. If anyone has any objections, please let us know.
------------------
commit 0a58e077eb600d1efd7e54ad9926a75a39d7f8ae upstream.
blk_cleanup_queue() calls elevator_exit() and after this, we can't touch the elevator without oopsing. __elv_next_request() must check for this state because in the refcounted queue model, we can still call it after blk_cleanup_queue() has been called.
This was reported as causing an oops attributable to scsi.
[WT: in 2.6.27, __elv_next_request() is in elevator.c]
Signed-off-by: James Bottomley <James.Bottomley@suse.de> Signed-off-by: Jens Axboe <jaxboe@fusionio.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de> --- block/clk.h | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-)
Index: longterm-2.6.27/block/elevator.c =================================================================== --- longterm-2.6.27.orig/block/elevator.c 2012-02-05 22:34:32.624915301 +0100 +++ longterm-2.6.27/block/elevator.c 2012-02-05 22:34:45.849915230 +0100 @@ -729,7 +729,8 @@ return rq; } - if (!q->elevator->ops->elevator_dispatch_fn(q, 0)) + if (test_bit(QUEUE_FLAG_DEAD, &q->queue_flags) || + !q->elevator->ops->elevator_dispatch_fn(q, 0)) return NULL; } }
|