From: Xie Yongji xieyongji@bytedance.com
commit 4b3749865374899e115aa8c48681709b086fe6d3 upstream.
We should defer eventfd_signal() to the workqueue when eventfd_signal_allowed() return false rather than return true.
Fixes: b542e383d8c0 ("eventfd: Make signal recursion protection a task bit") Signed-off-by: Xie Yongji xieyongji@bytedance.com Link: https://lore.kernel.org/r/20210913111928.98-1-xieyongji@bytedance.com Reviewed-by: Eric Biggers ebiggers@google.com Signed-off-by: Eric Biggers ebiggers@google.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- fs/aio.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
--- a/fs/aio.c +++ b/fs/aio.c @@ -1761,7 +1761,7 @@ static int aio_poll_wake(struct wait_que list_del_init(&req->wait.entry); list_del(&iocb->ki_list); iocb->ki_res.res = mangle_poll(mask); - if (iocb->ki_eventfd && eventfd_signal_allowed()) { + if (iocb->ki_eventfd && !eventfd_signal_allowed()) { iocb = NULL; INIT_WORK(&req->work, aio_poll_put_work); schedule_work(&req->work);