From 746e7d285dcb96caa1845fbbb62b14bf4010cdfb Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Wed, 7 May 2025 08:07:09 -0600 Subject: [PATCH] io_uring: ensure deferred completions are posted for multishot Commit 687b2bae0efff9b25e071737d6af5004e6e35af5 upstream. Multishot normally uses io_req_post_cqe() to post completions, but when stopping it, it may finish up with a deferred completion. This is fine, except if another multishot event triggers before the deferred completions get flushed. If this occurs, then CQEs may get reordered in the CQ ring, and cause confusion on the application side. When multishot posting via io_req_post_cqe(), flush any pending deferred completions first, if any. Cc: stable@vger.kernel.org # 6.1+ Reported-by: Norman Maurer Reported-by: Christian Mazakas Signed-off-by: Jens Axboe Signed-off-by: Greg Kroah-Hartman --- io_uring/io_uring.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index 4f4ac40fc605..db592fa549b7 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -919,6 +919,14 @@ static bool __io_post_aux_cqe(struct io_ring_ctx *ctx, u64 user_data, s32 res, u { bool filled; + /* + * If multishot has already posted deferred completions, ensure that + * those are flushed first before posting this one. If not, CQEs + * could get reordered. + */ + if (!wq_list_empty(&ctx->submit_state.compl_reqs)) + __io_submit_flush_completions(ctx); + io_cq_lock(ctx); filled = io_fill_cqe_aux(ctx, user_data, res, cflags); if (!filled && allow_overflow)