]> git.ipfire.org Git - thirdparty/linux.git/commit
io_uring: don't touch sqd->thread off tw add
authorPavel Begunkov <asml.silence@gmail.com>
Fri, 10 Jan 2025 20:36:45 +0000 (20:36 +0000)
committerJens Axboe <axboe@kernel.dk>
Fri, 10 Jan 2025 21:00:25 +0000 (14:00 -0700)
commitbd2703b42decebdcddf76e277ba76b4c4a142d73
treefb40c213c149d6bc89a6ca538b87ac07ce5491c6
parent4b7cfa8b6c28a9fa22b86894166a1a34f6d630ba
io_uring: don't touch sqd->thread off tw add

With IORING_SETUP_SQPOLL all requests are created by the SQPOLL task,
which means that req->task should always match sqd->thread. Since
accesses to sqd->thread should be separately protected, use req->task
in io_req_normal_work_add() instead.

Note, in the eyes of io_req_normal_work_add(), the SQPOLL task struct
is always pinned and alive, and sqd->thread can either be the task or
NULL. It's only problematic if the compiler decides to reload the value
after the null check, which is not so likely.

Cc: stable@vger.kernel.org
Cc: Bui Quang Minh <minhquangbui99@gmail.com>
Reported-by: lizetao <lizetao1@huawei.com>
Fixes: 78f9b61bd8e54 ("io_uring: wake SQPOLL task when task_work is added to an empty queue")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/1cbbe72cf32c45a8fee96026463024cd8564a7d7.1736541357.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
io_uring/io_uring.c