From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-4.0 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_PASS autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 79B87C43381 for ; Mon, 25 Mar 2019 10:16:30 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 4909420879 for ; Mon, 25 Mar 2019 10:16:30 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1730492AbfCYKQ3 (ORCPT ); Mon, 25 Mar 2019 06:16:29 -0400 Received: from mx2.suse.de ([195.135.220.15]:57948 "EHLO mx1.suse.de" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1729935AbfCYKQ3 (ORCPT ); Mon, 25 Mar 2019 06:16:29 -0400 X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.220.254]) by mx1.suse.de (Postfix) with ESMTP id 9024FAF50; Mon, 25 Mar 2019 10:16:28 +0000 (UTC) MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII; format=flowed Content-Transfer-Encoding: 7bit Date: Mon, 25 Mar 2019 11:16:28 +0100 From: Roman Penyaev To: Jens Axboe , linux-block@vger.kernel.org Subject: io_uring: -EAGAIN on write path in case of O_DIRECT Message-ID: X-Sender: rpenyaev@suse.de User-Agent: Roundcube Webmail Sender: linux-block-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-block@vger.kernel.org Hi Jens, I gave a try to use io_uring and stumbled upon -EAGAIN on write path in direct mode if page cache is already populated or has been populated in-between by some buffered read. I'am talking about generic_file_direct_write() call, which checks filemap_range_has_page() on IOCB_NOWAIT path. To proceed further with tests I simply did the same thing, like you did in io_read(), and in case of -EAGAIN async worker does the rest. So the following chunk works well: diff --git a/fs/io_uring.c b/fs/io_uring.c index 6aaa30580a2b..ccb656168ae4 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -1022,6 +1022,8 @@ static int io_write(struct io_kiocb *req, const struct sqe_submit *s, ret = rw_verify_area(WRITE, file, &kiocb->ki_pos, iov_count); if (!ret) { + ssize_t ret2; + /* * Open-code file_start_write here to grab freeze protection, * which will be released by another thread in @@ -1036,7 +1038,19 @@ static int io_write(struct io_kiocb *req, const struct sqe_submit *s, SB_FREEZE_WRITE); } kiocb->ki_flags |= IOCB_WRITE; - io_rw_done(kiocb, call_write_iter(file, kiocb, &iter)); + + ret2 = call_write_iter(file, kiocb, &iter); + if (!force_nonblock || ret2 != -EAGAIN) { + io_rw_done(kiocb, ret2); + } else { + /* + * If ->needs_lock is true, we're already in async + * context. + */ + if (!s->needs_lock) + io_async_list_note(WRITE, req, iov_count); + ret = -EAGAIN; + } Does it make sense? -- Roman