| Message ID | d880bc27fb8c3209b54641be4ff6ac02b0e5789a.1743679736.git.asml.silence@gmail.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | [1/1] io_uring: always do atomic put from iowq | expand |
On Thu, 03 Apr 2025 12:29:30 +0100, Pavel Begunkov wrote: > io_uring always switches requests to atomic refcounting for iowq > execution before there is any parallilism by setting REQ_F_REFCOUNT, > and the flag is not cleared until the request completes. That should be > fine as long as the compiler doesn't make up a non existing value for > the flags, however KCSAN still complains when the request owner changes > oter flag bits: > > [...] Applied, thanks! [1/1] io_uring: always do atomic put from iowq commit: 390513642ee6763c7ada07f0a1470474986e6c1c Best regards,
diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index 13e0b48d1aac..aad1fd9794b9 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -1800,7 +1800,7 @@ struct io_wq_work *io_wq_free_work(struct io_wq_work *work) struct io_kiocb *req = container_of(work, struct io_kiocb, work); struct io_kiocb *nxt = NULL; - if (req_ref_put_and_test(req)) { + if (req_ref_put_and_test_atomic(req)) { if (req->flags & IO_REQ_LINK_FLAGS) nxt = io_req_find_next(req); io_free_req(req); diff --git a/io_uring/refs.h b/io_uring/refs.h index 63982ead9f7d..0d928d87c4ed 100644 --- a/io_uring/refs.h +++ b/io_uring/refs.h @@ -17,6 +17,13 @@ static inline bool req_ref_inc_not_zero(struct io_kiocb *req) return atomic_inc_not_zero(&req->refs); } +static inline bool req_ref_put_and_test_atomic(struct io_kiocb *req) +{ + WARN_ON_ONCE(!(data_race(req->flags) & REQ_F_REFCOUNT)); + WARN_ON_ONCE(req_ref_zero_or_close_to_overflow(req)); + return atomic_dec_and_test(&req->refs); +} + static inline bool req_ref_put_and_test(struct io_kiocb *req) { if (likely(!(req->flags & REQ_F_REFCOUNT)))
io_uring always switches requests to atomic refcounting for iowq execution before there is any parallilism by setting REQ_F_REFCOUNT, and the flag is not cleared until the request completes. That should be fine as long as the compiler doesn't make up a non existing value for the flags, however KCSAN still complains when the request owner changes oter flag bits: BUG: KCSAN: data-race in io_req_task_cancel / io_wq_free_work ... read to 0xffff888117207448 of 8 bytes by task 3871 on cpu 0: req_ref_put_and_test io_uring/refs.h:22 [inline] Skip REQ_F_REFCOUNT checks for iowq, we know it's set. Reported-by: syzbot+903a2ad71fb3f1e47cf5@syzkaller.appspotmail.com Signed-off-by: Pavel Begunkov <asml.silence@gmail.com> --- io_uring/io_uring.c | 2 +- io_uring/refs.h | 7 +++++++ 2 files changed, 8 insertions(+), 1 deletion(-)