io_uring/rw: add write support for IOCB_DIO_CALLER_COMP

If the filesystem dio handler understands IOCB_DIO_CALLER_COMP, we'll
get a kiocb->ki_complete() callback with kiocb->dio_complete set. In
that case, rather than complete the IO directly through task_work, queue
up an intermediate task_work handler that first processes this callback
and then immediately completes the request.

For XFS, this avoids a punt through a workqueue, which is a lot less
efficient and adds latency to lower queue depth (or sync) O_DIRECT
writes.

Only do this for non-polled IO, as polled IO doesn't need this kind
of deferral as it always completes within the task itself. This then
avoids a check for deferral in the polled IO completion handler.
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

io_uring/rw.c

@@ -105,6 +105,7 @@ int io_prep_rw(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 	} else {
 		rw->kiocb.ki_ioprio = get_current_ioprio();
 	}
+	rw->kiocb.dio_complete = NULL;
 
 	rw->addr = READ_ONCE(sqe->addr);
 	rw->len = READ_ONCE(sqe->len);
@@ -285,6 +286,15 @@ static inline int io_fixup_rw_res(struct io_kiocb *req, long res)
 
 void io_req_rw_complete(struct io_kiocb *req, struct io_tw_state *ts)
 {
+	struct io_rw *rw = io_kiocb_to_cmd(req, struct io_rw);
+	struct kiocb *kiocb = &rw->kiocb;
+
+	if ((kiocb->ki_flags & IOCB_DIO_CALLER_COMP) && kiocb->dio_complete) {
+		long res = kiocb->dio_complete(rw->kiocb.private);
+
+		io_req_set_res(req, io_fixup_rw_res(req, res), 0);
+	}
+
 	io_req_io_end(req);
 
 	if (req->flags & (REQ_F_BUFFER_SELECTED|REQ_F_BUFFER_RING)) {
@@ -300,9 +310,11 @@ static void io_complete_rw(struct kiocb *kiocb, long res)
 	struct io_rw *rw = container_of(kiocb, struct io_rw, kiocb);
 	struct io_kiocb *req = cmd_to_io_kiocb(rw);
 
-	if (__io_complete_rw_common(req, res))
-		return;
-	io_req_set_res(req, io_fixup_rw_res(req, res), 0);
+	if (!kiocb->dio_complete || !(kiocb->ki_flags & IOCB_DIO_CALLER_COMP)) {
+		if (__io_complete_rw_common(req, res))
+			return;
+		io_req_set_res(req, io_fixup_rw_res(req, res), 0);
+	}
 	req->io_task_work.func = io_req_rw_complete;
 	__io_req_task_work_add(req, IOU_F_TWQ_LAZY_WAKE);
 }
@@ -916,6 +928,15 @@ int io_write(struct io_kiocb *req, unsigned int issue_flags)
 	}
 	kiocb->ki_flags |= IOCB_WRITE;
 
+	/*
+	 * For non-polled IO, set IOCB_DIO_CALLER_COMP, stating that our handler
+	 * groks deferring the completion to task context. This isn't
+	 * necessary and useful for polled IO as that can always complete
+	 * directly.
+	 */
+	if (!(kiocb->ki_flags & IOCB_HIPRI))
+		kiocb->ki_flags |= IOCB_DIO_CALLER_COMP;
+
 	if (likely(req->file->f_op->write_iter))
 		ret2 = call_write_iter(req->file, kiocb, &s->iter);
 	else if (req->file->f_op->write)