For blk-mq, queue release handler is usually called into after blk_mq_freeze_queue_wait() returns. However, q_usage_counter->release() handler may not be started yet at that time, so cause user-after-free. Fix the issue by moving percpu_ref_exit() into blk_free_queue_rcu() since ->release() is called with rcu read lock held, since it is concluded that the race should be covered in caller per discussion from the two links. Reported-by: Zhang Wensheng <zhangwensheng@xxxxxxxxxxxxxxx> Reported-by: Zhong Jinghua <zhongjinghua@xxxxxxxxxx> Link: https://lore.kernel.org/linux-block/Y5prfOjyyjQKUrtH@T590/T/#u Link: https://lore.kernel.org/lkml/Y4%2FmzMd4evRg9yDi@fedora/ Cc: Hillf Danton <hdanton@xxxxxxxx> Cc: Yu Kuai <yukuai3@xxxxxxxxxx> Cc: Dennis Zhou <dennis@xxxxxxxxxx> Fixes: 2b0d3d3e4fcf ("percpu_ref: reduce memory footprint of percpu_ref in fast path") Signed-off-by: Ming Lei <ming.lei@xxxxxxxxxx> --- block/blk-core.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/block/blk-core.c b/block/blk-core.c index 3866b6c4cd88..9321767470dc 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -254,14 +254,15 @@ EXPORT_SYMBOL_GPL(blk_clear_pm_only); static void blk_free_queue_rcu(struct rcu_head *rcu_head) { - kmem_cache_free(blk_requestq_cachep, - container_of(rcu_head, struct request_queue, rcu_head)); + struct request_queue *q = container_of(rcu_head, + struct request_queue, rcu_head); + + percpu_ref_exit(&q->q_usage_counter); + kmem_cache_free(blk_requestq_cachep, q); } static void blk_free_queue(struct request_queue *q) { - percpu_ref_exit(&q->q_usage_counter); - if (q->poll_stat) blk_stat_remove_callback(q, q->poll_cb); blk_stat_free_callback(q->poll_cb); -- 2.31.1
