From: Jens Axboe Date: Tue, 11 Aug 2020 15:50:19 +0000 (-0600) Subject: io_uring: fail poll arm on queue proc failure X-Git-Tag: v5.7.16~37 X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=ae189bfb784786e94b3b8826d01253500f95b605;p=thirdparty%2Fkernel%2Fstable.git io_uring: fail poll arm on queue proc failure commit a36da65c46565d2527eec3efdb546251e38253fd upstream. Check the ipt.error value, it must have been either cleared to zero or set to another error than the default -EINVAL if we don't go through the waitqueue proc addition. Just give up on poll at that point and return failure, this will fallback to async work. io_poll_add() doesn't suffer from this failure case, as it returns the error value directly. Cc: stable@vger.kernel.org # v5.7+ Reported-by: syzbot+a730016dc0bdce4f6ff5@syzkaller.appspotmail.com Reviewed-by: Stefano Garzarella Signed-off-by: Jens Axboe Signed-off-by: Greg Kroah-Hartman --- diff --git a/fs/io_uring.c b/fs/io_uring.c index 13c915d254d44..31cad88ec4852 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -4544,7 +4544,7 @@ static bool io_arm_poll_handler(struct io_kiocb *req) ret = __io_arm_poll_handler(req, &apoll->poll, &ipt, mask, io_async_wake); - if (ret) { + if (ret || ipt.error) { io_poll_remove_double(req, apoll->double_poll); spin_unlock_irq(&ctx->completion_lock); memcpy(&req->work, &apoll->work, sizeof(req->work));