From: Pavel Begunkov Date: Thu, 7 Jul 2022 14:13:15 +0000 (+0100) Subject: io_uring: don't race double poll setting REQ_F_ASYNC_DATA X-Git-Url: http://git.maquefel.me/?a=commitdiff_plain;h=ceff501790a9789c748a9b851f30f4f7e2fe4d72;p=linux.git io_uring: don't race double poll setting REQ_F_ASYNC_DATA Just as with io_poll_double_prepare() setting REQ_F_DOUBLE_POLL, we can race with the first poll entry when setting REQ_F_ASYNC_DATA. Move it under io_poll_double_prepare(). Fixes: a18427bb2d9b ("io_uring: optimise submission side poll_refs") Signed-off-by: Pavel Begunkov Link: https://lore.kernel.org/r/df6920f509c11115aa2bce8b34dc5fdb0eb98920.1657203020.git.asml.silence@gmail.com Signed-off-by: Jens Axboe --- diff --git a/io_uring/poll.c b/io_uring/poll.c index 3710a0a46a877..c1359d45a396f 100644 --- a/io_uring/poll.c +++ b/io_uring/poll.c @@ -410,6 +410,8 @@ static void io_poll_double_prepare(struct io_kiocb *req) spin_lock_irq(&head->lock); req->flags |= REQ_F_DOUBLE_POLL; + if (req->opcode == IORING_OP_POLL_ADD) + req->flags |= REQ_F_ASYNC_DATA; if (head) spin_unlock_irq(&head->lock); @@ -448,13 +450,11 @@ static void __io_queue_proc(struct io_poll *poll, struct io_poll_table *pt, return; } - io_poll_double_prepare(req); /* mark as double wq entry */ wqe_private |= IO_WQE_F_DOUBLE; io_init_poll_iocb(poll, first->events, first->wait.func); + io_poll_double_prepare(req); *poll_ptr = poll; - if (req->opcode == IORING_OP_POLL_ADD) - req->flags |= REQ_F_ASYNC_DATA; } else { /* fine to modify, there is no poll queued to race with us */ req->flags |= REQ_F_SINGLE_POLL;