io_uring/rw: ensure retry condition isn't lost

A previous commit removed the checking on whether or not it was possible
to retry a request, since it's now possible to retry any of them. This
would previously have caused the request to have been ended with an error,
but now the retry condition can simply get lost instead.

Cleanup the retry handling and always just punt it to task_work, which
will queue it with io-wq appropriately.
Reported-by: Changhui Zhong <czhong@redhat.com>
Tested-by: Ming Lei <ming.lei@redhat.com>
Fixes: cca65713 ("io_uring/rw: cleanup retry path")
Signed-off-by: Jens Axboe <axboe@kernel.dk>

io_uring/io_uring.c

@@ -527,6 +527,19 @@ static void io_queue_iowq(struct io_kiocb *req)
 		io_queue_linked_timeout(link);
 }
 
+static void io_tw_requeue_iowq(struct io_kiocb *req, struct io_tw_state *ts)
+{
+	req->flags &= ~REQ_F_REISSUE;
+	io_queue_iowq(req);
+}
+
+void io_tw_queue_iowq(struct io_kiocb *req)
+{
+	req->flags |= REQ_F_REISSUE | REQ_F_BL_NO_RECYCLE;
+	req->io_task_work.func = io_tw_requeue_iowq;
+	io_req_task_work_add(req);
+}
+
 static __cold void io_queue_deferred(struct io_ring_ctx *ctx)
 {
 	while (!list_empty(&ctx->defer_list)) {

io_uring/io_uring.h

@@ -75,6 +75,7 @@ struct file *io_file_get_fixed(struct io_kiocb *req, int fd,
 void __io_req_task_work_add(struct io_kiocb *req, unsigned flags);
 bool io_alloc_async_data(struct io_kiocb *req);
 void io_req_task_queue(struct io_kiocb *req);
+void io_tw_queue_iowq(struct io_kiocb *req);
 void io_req_task_complete(struct io_kiocb *req, struct io_tw_state *ts);
 void io_req_task_queue_fail(struct io_kiocb *req, int ret);
 void io_req_task_submit(struct io_kiocb *req, struct io_tw_state *ts);

io_uring/rw.c

@@ -455,7 +455,7 @@ static bool __io_complete_rw_common(struct io_kiocb *req, long res)
 			 * current cycle.
 			 */
 			io_req_io_end(req);
-			req->flags |= REQ_F_REISSUE | REQ_F_BL_NO_RECYCLE;
+			io_tw_queue_iowq(req);
 			return true;
 		}
 		req_set_fail(req);
@@ -521,7 +521,7 @@ static void io_complete_rw_iopoll(struct kiocb *kiocb, long res)
 		io_req_end_write(req);
 	if (unlikely(res != req->cqe.res)) {
 		if (res == -EAGAIN && io_rw_should_reissue(req)) {
-			req->flags |= REQ_F_REISSUE | REQ_F_BL_NO_RECYCLE;
+			io_tw_queue_iowq(req);
 			return;
 		}
 		req->cqe.res = res;
@@ -839,7 +839,8 @@ static int __io_read(struct io_kiocb *req, unsigned int issue_flags)
 	ret = io_iter_do_read(rw, &io->iter);
 
 	if (ret == -EAGAIN || (req->flags & REQ_F_REISSUE)) {
-		req->flags &= ~REQ_F_REISSUE;
+		if (req->flags & REQ_F_REISSUE)
+			return IOU_ISSUE_SKIP_COMPLETE;
 		/* If we can poll, just do that. */
 		if (io_file_can_poll(req))
 			return -EAGAIN;
@@ -1034,10 +1035,8 @@ int io_write(struct io_kiocb *req, unsigned int issue_flags)
 	else
 		ret2 = -EINVAL;
 
-	if (req->flags & REQ_F_REISSUE) {
-		req->flags &= ~REQ_F_REISSUE;
-		ret2 = -EAGAIN;
-	}
+	if (req->flags & REQ_F_REISSUE)
+		return IOU_ISSUE_SKIP_COMPLETE;
 
 	/*
 	 * Raw bdev writes will return -EOPNOTSUPP for IOCB_NOWAIT. Just