io_uring: don't try io-wq polling if not supported
authorPavel Begunkov <asml.silence@gmail.com>
Sat, 23 Oct 2021 11:13:59 +0000 (12:13 +0100)
committerJens Axboe <axboe@kernel.dk>
Mon, 25 Oct 2021 13:42:33 +0000 (07:42 -0600)
If an opcode doesn't support polling, just let it be executed
synchronously in iowq, otherwise it will do a nonblock attempt just to
fail in io_arm_poll_handler() and return back to blocking execution.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/6401256db01b88f448f15fcd241439cb76f5b940.1634987320.git.asml.silence@gmail.com
Reviewed-by: Hao Xu <haoxu@linux.alibaba.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
fs/io_uring.c

index bff911f..c6f32fc 100644 (file)
@@ -6741,9 +6741,13 @@ static void io_wq_submit_work(struct io_wq_work *work)
        }
 
        if (req->flags & REQ_F_FORCE_ASYNC) {
-               needs_poll = req->file && file_can_poll(req->file);
-               if (needs_poll)
+               const struct io_op_def *def = &io_op_defs[req->opcode];
+               bool opcode_poll = def->pollin || def->pollout;
+
+               if (opcode_poll && file_can_poll(req->file)) {
+                       needs_poll = true;
                        issue_flags |= IO_URING_F_NONBLOCK;
+               }
        }
 
        do {