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=-5.2 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,MAILING_LIST_MULTI,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 24BD5C433B4 for ; Wed, 14 Apr 2021 11:24:05 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id EEC1A61164 for ; Wed, 14 Apr 2021 11:24:04 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S233014AbhDNLX4 (ORCPT ); Wed, 14 Apr 2021 07:23:56 -0400 Received: from out30-57.freemail.mail.aliyun.com ([115.124.30.57]:48560 "EHLO out30-57.freemail.mail.aliyun.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S234078AbhDNLXz (ORCPT ); Wed, 14 Apr 2021 07:23:55 -0400 X-Alimail-AntiSpam: AC=PASS;BC=-1|-1;BR=01201311R751e4;CH=green;DM=||false|;DS=||;FP=0|-1|-1|-1|0|-1|-1|-1;HT=e01e04400;MF=haoxu@linux.alibaba.com;NM=1;PH=DS;RN=3;SR=0;TI=SMTPD_---0UVXqQIm_1618399412; Received: from B-25KNML85-0107.local(mailfrom:haoxu@linux.alibaba.com fp:SMTPD_---0UVXqQIm_1618399412) by smtp.aliyun-inc.com(127.0.0.1); Wed, 14 Apr 2021 19:23:32 +0800 To: Jens Axboe From: Hao Xu Subject: 12aceb89b0bc eventfd: convert to f_op->read_iter() causes 6.7% performance regression Cc: io-uring , Alexander Viro Message-ID: <1157fb4c-9975-3515-70b0-fefefa3ddfae@linux.alibaba.com> Date: Wed, 14 Apr 2021 19:23:32 +0800 User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:78.0) Gecko/20100101 Thunderbird/78.9.1 MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 7bit Precedence: bulk List-ID: X-Mailing-List: io-uring@vger.kernel.org Hi Jens, We found that 12aceb89b0bc eventfd: convert to f_op->read_iter() introduced 6.7% peformance regression in test suite will-it-scale The test commands are: ./runtest.py eventfd1 10 thread 1 and ./runtest.py eventfd1 10 thread 128 the test is a simple while which is like: while(time not over) { write(eventfd); read(eventfd); } By perf analysis, I saw that the main cost is new_sync_read(). Though I think this regression only happens when there is no 'schedule/sleep'(otherwise sleep occupies most time), I feel it still necessary to report it.