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=-6.7 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,MAILING_LIST_MULTI,NICE_REPLY_A,SPF_HELO_NONE, SPF_PASS,UNPARSEABLE_RELAY,USER_AGENT_SANE_1 autolearn=no 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 9A0B0C433EF for ; Fri, 17 Sep 2021 18:49:23 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 714FE61241 for ; Fri, 17 Sep 2021 18:49:23 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1343615AbhIQSuo (ORCPT ); Fri, 17 Sep 2021 14:50:44 -0400 Received: from out30-45.freemail.mail.aliyun.com ([115.124.30.45]:40931 "EHLO out30-45.freemail.mail.aliyun.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S236020AbhIQSua (ORCPT ); Fri, 17 Sep 2021 14:50:30 -0400 X-Alimail-AntiSpam: AC=PASS;BC=-1|-1;BR=01201311R171e4;CH=green;DM=||false|;DS=||;FP=0|-1|-1|-1|0|-1|-1|-1;HT=alimailimapcm10staff010182156082;MF=haoxu@linux.alibaba.com;NM=1;PH=DS;RN=4;SR=0;TI=SMTPD_---0UoiV53d_1631904545; Received: from B-25KNML85-0107.local(mailfrom:haoxu@linux.alibaba.com fp:SMTPD_---0UoiV53d_1631904545) by smtp.aliyun-inc.com(127.0.0.1); Sat, 18 Sep 2021 02:49:06 +0800 Subject: Re: [PATCH 1/6] io_uring: enhance flush completion logic From: Hao Xu To: Pavel Begunkov , Jens Axboe Cc: io-uring@vger.kernel.org, Joseph Qi References: <20210903110049.132958-1-haoxu@linux.alibaba.com> <20210903110049.132958-2-haoxu@linux.alibaba.com> <302430f6-f83e-4096-448e-9d35f8f4303e@linux.alibaba.com> <5dd28d14-24b5-e2dc-aa55-a68cb5d9f4e8@gmail.com> <53d4f2b4-b0eb-fa4a-9659-794c2c363a7f@linux.alibaba.com> Message-ID: <00d6d99c-946a-39b4-9f12-d46d4fee8e61@linux.alibaba.com> Date: Sat, 18 Sep 2021 02:49:05 +0800 User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:78.0) Gecko/20100101 Thunderbird/78.13.0 MIME-Version: 1.0 In-Reply-To: <53d4f2b4-b0eb-fa4a-9659-794c2c363a7f@linux.alibaba.com> Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 8bit Precedence: bulk List-ID: X-Mailing-List: io-uring@vger.kernel.org 在 2021/9/3 下午9:38, Hao Xu 写道: > 在 2021/9/3 下午8:27, Pavel Begunkov 写道: >> On 9/3/21 1:08 PM, Hao Xu wrote: >>> 在 2021/9/3 下午7:42, Pavel Begunkov 写道: >>>> On 9/3/21 12:00 PM, Hao Xu wrote: >>>>> Though currently refcount of a req is always one when we flush inline >>>> >>>> It can be refcounted and != 1. E.g. poll requests, or consider >>> It seems poll requests don't leverage comp cache, do I miss anything? >> >> Hmm, looks so. Not great that it doesn't, but probably it's because >> of trying to submit next reqs right in io_poll_task_func(). >> >> I'll be pushing for some changes around tw, with it should be easy >> to hook poll completion batching with no drawbacks. Would be great >> if you will be willing to take a shot on it. > Sure, I'll take a look. >> Tried to integrate poll logic to completion cache, I did test it with an echo-server program, both multishot and singleshot mode, not difference on req/s.