Pavel Begunkov <asml.silence@xxxxxxxxx> writes:
> task_work fallback is executed from a workqueue, so current and
> req->task are not necessarily the same. It's still safe to poke into it
> as the request holds a task_struct reference.
Makes sense to me. Feel free to add
Reviewed-by: Gabriel Krisman Bertazi <krisman@xxxxxxx>
>
> Signed-off-by: Pavel Begunkov <asml.silence@xxxxxxxxx>
> ---
> io_uring/io_uring.c | 2 +-
> io_uring/poll.c | 1 -
> 2 files changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
> index 94329c1ce91d..5a8a43fb6750 100644
> --- a/io_uring/io_uring.c
> +++ b/io_uring/io_uring.c
> @@ -1249,7 +1249,7 @@ static void io_req_task_cancel(struct io_kiocb *req, bool *locked)
> void io_req_task_submit(struct io_kiocb *req, bool *locked)
> {
> io_tw_lock(req->ctx, locked);
> - /* req->task == current here, checking PF_EXITING is safe */
> +
> if (likely(!(req->task->flags & PF_EXITING)))
> io_queue_sqe(req);
> else
> diff --git a/io_uring/poll.c b/io_uring/poll.c
> index 2830b7daf952..5d4a0a4a379a 100644
> --- a/io_uring/poll.c
> +++ b/io_uring/poll.c
> @@ -214,7 +214,6 @@ static int io_poll_check_events(struct io_kiocb *req, bool *locked)
> struct io_ring_ctx *ctx = req->ctx;
> int v, ret;
>
> - /* req->task == current here, checking PF_EXITING is safe */
> if (unlikely(req->task->flags & PF_EXITING))
> return -ECANCELED;
--
Gabriel Krisman Bertazi
