On 6/28/20 3:52 AM, Pavel Begunkov wrote: > Don't forget to wake up a process to which io_rw_reissue() added > task_work. > > Signed-off-by: Pavel Begunkov <asml.silence@xxxxxxxxx> > --- > fs/io_uring.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/fs/io_uring.c b/fs/io_uring.c > index 5e0196393c4f..79d5680219d1 100644 > --- a/fs/io_uring.c > +++ b/fs/io_uring.c > @@ -2172,6 +2172,7 @@ static bool io_rw_reissue(struct io_kiocb *req, long res) > ret = task_work_add(tsk, &req->task_work, true); > if (!ret) > return true; > + wake_up_process(tsk); > #endif > return false; > } > Actually, I spoke too soon, you'll want this wake_up_process() to be for the success case. I've dropped this patch, please resend it. -- Jens Axboe
