diff options
| author | Pavel Begunkov | 2023-05-19 15:21:16 +0100 |
|---|---|---|
| committer | Jens Axboe | 2023-05-19 19:56:56 -0600 |
| commit | 5498bf28d8f2bd63a46ad40f4427518615fb793f (patch) | |
| tree | 309046326582714d3cab7d2ff7cb5f09edbf95a3 | |
| parent | 3af0356c162c299a8216576b644eb72715e97cb2 (diff) | |
io_uring: annotate offset timeout races
It's racy to read ->cached_cq_tail without taking proper measures
(usually grabbing ->completion_lock) as timeout requests with CQE
offsets do, however they have never had a good semantics for from
when they start counting. Annotate racy reads with data_race().
Reported-by: syzbot+cb265db2f3f3468ef436@syzkaller.appspotmail.com
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/4de3685e185832a92a572df2be2c735d2e21a83d.1684506056.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
| -rw-r--r-- | io_uring/timeout.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/io_uring/timeout.c b/io_uring/timeout.c index fc950177e2e1..350eb830b485 100644 --- a/io_uring/timeout.c +++ b/io_uring/timeout.c @@ -594,7 +594,7 @@ int io_timeout(struct io_kiocb *req, unsigned int issue_flags) goto add; } - tail = ctx->cached_cq_tail - atomic_read(&ctx->cq_timeouts); + tail = data_race(ctx->cached_cq_tail) - atomic_read(&ctx->cq_timeouts); timeout->target_seq = tail + off; /* Update the last seq here in case io_flush_timeouts() hasn't. |