Hi Yi, Thanks for the report! On Mon, Dec 06, 2021 at 09:38:35AM +0800, Yi Zhang wrote: > This is the first commit that observed this issue. > Kernel repo: > https://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux-block.git > Commit: c7d61010b991 - Merge branch 'for-5.17/block' into for-next The issue should be fixed by the following patch: >From 8d30d9a46d4ede16c61ad48c2a1ceaf2ec29d044 Mon Sep 17 00:00:00 2001 From: Ming Lei <ming.lei@xxxxxxxxxx> Date: Mon, 6 Dec 2021 11:33:50 +0800 Subject: [PATCH] blk-mq: don't use plug->mq_list->q directly in blk_mq_run_dispatch_ops() blk_mq_run_dispatch_ops() is defined as one macro, and plug->mq_list will be changed when running 'dispatch_ops', so add one local variable for holding request queue. Reported-by: Yi Zhang <yi.zhang@xxxxxxxxxx> Fixes: 4cafe86c9267 ("blk-mq: run dispatch lock once in case of issuing from list") Signed-off-by: Ming Lei <ming.lei@xxxxxxxxxx> --- block/blk-mq.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/block/blk-mq.c b/block/blk-mq.c index 22ec21aa0c22..537295f6e0e9 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -2521,7 +2521,9 @@ void blk_mq_flush_plug_list(struct blk_plug *plug, bool from_schedule) plug->rq_count = 0; if (!plug->multiple_queues && !plug->has_elevator && !from_schedule) { - blk_mq_run_dispatch_ops(plug->mq_list->q, + struct request_queue *q = rq_list_peek(&plug->mq_list)->q; + + blk_mq_run_dispatch_ops(q, blk_mq_plug_issue_direct(plug, false)); if (rq_list_empty(plug->mq_list)) return; -- 2.31.1 Thanks, Ming
