On 2011-09-27 18:55, Linus Torvalds wrote:
> On Tue, Sep 27, 2011 at 5:47 PM, Jens Axboe <axboe@xxxxxxxxx> wrote:
>>
>> Do we have acks from the people who saw this oops, that this really does
>> fix it for them? I ask since this fix is vastly cleaner than the other
>> patch floating around, so I'd much rather pull this one in.
>
> I don't think I can't test it, since it seems to be some hard-to-hit
> race condition. It certainly doesn't seem to happen all the time -
> today was the first time I've seen it on one of my computers.
>
> But if you forward the actual patch to me (the one I see on lkml seems
> to be broken and doesn't compile cleanly because it's assiging a
> structure to a pointer), I'll try it out anyway.
Thanks, that would be great. It's inlined below.
diff --git a/block/blk-core.c b/block/blk-core.c
index b2ed78a..d34433a 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -348,9 +348,10 @@ void blk_put_queue(struct request_queue *q)
EXPORT_SYMBOL(blk_put_queue);
/*
- * Note: If a driver supplied the queue lock, it should not zap that lock
- * unexpectedly as some queue cleanup components like elevator_exit() and
- * blk_throtl_exit() need queue lock.
+ * Note: If a driver supplied the queue lock, it is disconnected
+ * by this function. The actual state of the lock doesn't matter
+ * here as the request_queue isn't accessible after this point
+ * (QUEUE_FLAG_DEAD is set) and no other requests will be queued.
*/
void blk_cleanup_queue(struct request_queue *q)
{
@@ -367,10 +368,8 @@ void blk_cleanup_queue(struct request_queue *q)
queue_flag_set_unlocked(QUEUE_FLAG_DEAD, q);
mutex_unlock(&q->sysfs_lock);
- if (q->elevator)
- elevator_exit(q->elevator);
-
- blk_throtl_exit(q);
+ if (q->queue_lock != &q->__queue_lock)
+ q->queue_lock = &q->__queue_lock;
blk_put_queue(q);
}
diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c
index e681805..60fda88 100644
--- a/block/blk-sysfs.c
+++ b/block/blk-sysfs.c
@@ -479,6 +479,11 @@ static void blk_release_queue(struct kobject *kobj)
blk_sync_queue(q);
+ if (q->elevator)
+ elevator_exit(q->elevator);
+
+ blk_throtl_exit(q);
+
if (rl->rq_pool)
mempool_destroy(rl->rq_pool);
--
Jens Axboe
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
