[{"id":1799735,"web_url":"http://patchwork.ozlabs.org/comment/1799735/","msgid":"<20171106143155.GA8738@salvia>","list_archive_url":null,"date":"2017-11-06T14:31:55","subject":"Re: [PATCH] netfilter: conntrack: use power efficient workqueue","submitter":{"id":1315,"url":"http://patchwork.ozlabs.org/api/people/1315/","name":"Pablo Neira Ayuso","email":"pablo@netfilter.org"},"content":"On Thu, 2 Nov 2017 15:16:07 +0100 Vincent Guittot wrote:\n> conntrack uses the bounded system_long_wq workqueue for its works that\n> don't have to run on the cpu they have been queued. Using bounded\n> workqueue prevents the scheduler to make smart decision about the best\n> place to schedule the work.\n>\n> This patch replaces system_long_wq with system_power_efficient_wq.\n> the work stays bounded to a cpu by default unless the\n> CONFIG_WQ_POWER_EFFICIENT is enable. In the latter case, the work can\n> be scheduled on the best cpu from a power or a performance point of\n> view.\n\nApplied, thanks.\n--\nTo unsubscribe from this list: send the line \"unsubscribe netfilter-devel\" in\nthe body of a message to majordomo@vger.kernel.org\nMore majordomo info at http://vger.kernel.org/majordomo-info.html","headers":{"Return-Path":"","X-Original-To":"incoming@patchwork.ozlabs.org","Delivered-To":"patchwork-incoming@bilbo.ozlabs.org","Authentication-Results":"ozlabs.org;\n\tspf=none (mailfrom) smtp.mailfrom=vger.kernel.org\n\t(client-ip=209.132.180.67; helo=vger.kernel.org;\n\tenvelope-from=netfilter-devel-owner@vger.kernel.org;\n\treceiver=)","Received":["from vger.kernel.org (vger.kernel.org [209.132.180.67])\n\tby ozlabs.org (Postfix) with ESMTP id 3yVw4W49HRz9ryT\n\tfor ;\n\tTue, 7 Nov 2017 01:32:11 +1100 (AEDT)","(majordomo@vger.kernel.org) by vger.kernel.org via listexpand\n\tid S1752756AbdKFOcK (ORCPT );\n\tMon, 6 Nov 2017 09:32:10 -0500","from mail.us.es ([193.147.175.20]:52542 \"EHLO mail.us.es\"\n\trhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP\n\tid S1752425AbdKFOcK (ORCPT );\n\tMon, 6 Nov 2017 09:32:10 -0500","from antivirus1-rhel7.int (unknown [192.168.2.11])\n\tby mail.us.es (Postfix) with ESMTP id EC5261B8419\n\tfor ;\n\tMon, 6 Nov 2017 15:32:08 +0100 (CET)","from antivirus1-rhel7.int (localhost [127.0.0.1])\n\tby antivirus1-rhel7.int (Postfix) with ESMTP id C6322DA793\n\tfor ;\n\tMon, 6 Nov 2017 15:32:08 +0100 (CET)","by antivirus1-rhel7.int (Postfix, from userid 99)\n\tid 7BB29DA3AA; Mon, 6 Nov 2017 15:32:08 +0100 (CET)","from antivirus1-rhel7.int (localhost [127.0.0.1])\n\tby antivirus1-rhel7.int (Postfix) with ESMTP id 12AFDDA392;\n\tMon, 6 Nov 2017 15:32:05 +0100 (CET)","from 192.168.1.97 (192.168.1.97) by antivirus1-rhel7.int\n\t(F-Secure/fsigk_smtp/550/antivirus1-rhel7.int); \n\tMon, 06 Nov 2017 15:32:05 +0100 (CET)","from us.es (unknown [27.1.3.195])\n\t(using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256\n\tbits)) (No client certificate requested)\n\t(Authenticated sender: 1984lsi)\n\tby entrada.int (Postfix) with ESMTPSA id C3243403DFA0;\n\tMon, 6 Nov 2017 15:32:03 +0100 (CET)"],"X-Spam-Checker-Version":"SpamAssassin 3.4.1 (2015-04-28) on\n\tantivirus1-rhel7.int","X-Spam-Level":"","X-Spam-Status":"No, score=-106.7 required=7.5 tests=ALL_TRUSTED,BAYES_50,\n\tFAKE_REPLY_C, HEADER_FROM_DIFFERENT_DOMAINS, SMTPAUTH_US2,\n\tUSER_IN_WHITELIST autolearn=disabled version=3.4.1","X-Virus-Status":"clean(F-Secure/fsigk_smtp/550/antivirus1-rhel7.int)","Date":"Mon, 6 Nov 2017 15:31:55 +0100","X-SMTPAUTHUS":"auth mail.us.es","From":"Pablo Neira Ayuso ","To":"Vincent Guittot ","Cc":"netfilter-devel@vger.kernel.org","Subject":"Re: [PATCH] netfilter: conntrack: use power efficient workqueue","Message-ID":"<20171106143155.GA8738@salvia>","MIME-Version":"1.0","Content-Type":"text/plain; charset=us-ascii","Content-Disposition":"inline","In-Reply-To":"<1509635767-12253-1-git-send-email-vincent.guittot@linaro.org>","User-Agent":"Mutt/1.5.23 (2014-03-12)","X-Virus-Scanned":"ClamAV using ClamSMTP","Sender":"netfilter-devel-owner@vger.kernel.org","Precedence":"bulk","List-ID":"","X-Mailing-List":"netfilter-devel@vger.kernel.org"}},{"id":1799760,"web_url":"http://patchwork.ozlabs.org/comment/1799760/","msgid":"<20171106145632.GA20091@salvia>","list_archive_url":null,"date":"2017-11-06T14:56:32","subject":"Re: [PATCH] netfilter: conntrack: use power efficient workqueue","submitter":{"id":1315,"url":"http://patchwork.ozlabs.org/api/people/1315/","name":"Pablo Neira Ayuso","email":"pablo@netfilter.org"},"content":"On Mon, Nov 06, 2017 at 03:31:55PM +0100, Pablo Neira Ayuso wrote:\n> On Thu, 2 Nov 2017 15:16:07 +0100 Vincent Guittot wrote:\n> > conntrack uses the bounded system_long_wq workqueue for its works that\n> > don't have to run on the cpu they have been queued. Using bounded\n> > workqueue prevents the scheduler to make smart decision about the best\n> > place to schedule the work.\n> >\n> > This patch replaces system_long_wq with system_power_efficient_wq.\n> > the work stays bounded to a cpu by default unless the\n> > CONFIG_WQ_POWER_EFFICIENT is enable. In the latter case, the work can\n> > be scheduled on the best cpu from a power or a performance point of\n> > view.\n> \n> Applied, thanks.\n\nI'm stepping back. According to what I'm reading\nsystem_power_efficient_wq becomes system_wq when disabled, which is\nnot semantically the same as system_long_wq that we have now.\n\nMy concern is that the conntrack garbage collector may run for quite a\nbit of time. Did you test this with a large conntrack table full of\nentries?\n\nThanks.\n--\nTo unsubscribe from this list: send the line \"unsubscribe netfilter-devel\" in\nthe body of a message to majordomo@vger.kernel.org\nMore majordomo info at http://vger.kernel.org/majordomo-info.html","headers":{"Return-Path":"","X-Original-To":"incoming@patchwork.ozlabs.org","Delivered-To":"patchwork-incoming@bilbo.ozlabs.org","Authentication-Results":"ozlabs.org;\n\tspf=none (mailfrom) smtp.mailfrom=vger.kernel.org\n\t(client-ip=209.132.180.67; helo=vger.kernel.org;\n\tenvelope-from=netfilter-devel-owner@vger.kernel.org;\n\treceiver=)","Received":["from vger.kernel.org (vger.kernel.org [209.132.180.67])\n\tby ozlabs.org (Postfix) with ESMTP id 3yVwcs6dBvz9s7G\n\tfor ;\n\tTue, 7 Nov 2017 01:56:45 +1100 (AEDT)","(majordomo@vger.kernel.org) by vger.kernel.org via listexpand\n\tid S932482AbdKFO4p (ORCPT );\n\tMon, 6 Nov 2017 09:56:45 -0500","from mail.us.es ([193.147.175.20]:32908 \"EHLO mail.us.es\"\n\trhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP\n\tid S932276AbdKFO4o (ORCPT ); \n\tMon, 6 Nov 2017 09:56:44 -0500","from antivirus1-rhel7.int (unknown [192.168.2.11])\n\tby mail.us.es (Postfix) with ESMTP id 0AA741B8401\n\tfor ;\n\tMon, 6 Nov 2017 15:56:43 +0100 (CET)","from antivirus1-rhel7.int (localhost [127.0.0.1])\n\tby antivirus1-rhel7.int (Postfix) with ESMTP id F0256DA385\n\tfor ;\n\tMon, 6 Nov 2017 15:56:42 +0100 (CET)","by antivirus1-rhel7.int (Postfix, from userid 99)\n\tid EF8A1DA38C; Mon, 6 Nov 2017 15:56:42 +0100 (CET)","from antivirus1-rhel7.int (localhost [127.0.0.1])\n\tby antivirus1-rhel7.int (Postfix) with ESMTP id CEBD2DA385;\n\tMon, 6 Nov 2017 15:56:39 +0100 (CET)","from 192.168.1.97 (192.168.1.97) by antivirus1-rhel7.int\n\t(F-Secure/fsigk_smtp/550/antivirus1-rhel7.int); \n\tMon, 06 Nov 2017 15:56:39 +0100 (CET)","from us.es (unknown [27.1.3.195])\n\t(using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256\n\tbits)) (No client certificate requested)\n\t(Authenticated sender: 1984lsi)\n\tby entrada.int (Postfix) with ESMTPSA id D66E7403DF98;\n\tMon, 6 Nov 2017 15:56:38 +0100 (CET)"],"X-Spam-Checker-Version":"SpamAssassin 3.4.1 (2015-04-28) on\n\tantivirus1-rhel7.int","X-Spam-Level":"","X-Spam-Status":"No, score=-108.2 required=7.5 tests=ALL_TRUSTED,BAYES_50,\n\tHEADER_FROM_DIFFERENT_DOMAINS,SMTPAUTH_US2,USER_IN_WHITELIST\n\tautolearn=disabled version=3.4.1","X-Virus-Status":"clean(F-Secure/fsigk_smtp/550/antivirus1-rhel7.int)","Date":"Mon, 6 Nov 2017 15:56:32 +0100","X-SMTPAUTHUS":"auth mail.us.es","From":"Pablo Neira Ayuso ","To":"Vincent Guittot ","Cc":"netfilter-devel@vger.kernel.org","Subject":"Re: [PATCH] netfilter: conntrack: use power efficient workqueue","Message-ID":"<20171106145632.GA20091@salvia>","References":"<1509635767-12253-1-git-send-email-vincent.guittot@linaro.org>\n\t<20171106143155.GA8738@salvia>","MIME-Version":"1.0","Content-Type":"text/plain; charset=us-ascii","Content-Disposition":"inline","In-Reply-To":"<20171106143155.GA8738@salvia>","User-Agent":"Mutt/1.5.23 (2014-03-12)","X-Virus-Scanned":"ClamAV using ClamSMTP","Sender":"netfilter-devel-owner@vger.kernel.org","Precedence":"bulk","List-ID":"","X-Mailing-List":"netfilter-devel@vger.kernel.org"}},{"id":1799774,"web_url":"http://patchwork.ozlabs.org/comment/1799774/","msgid":"","list_archive_url":null,"date":"2017-11-06T15:15:42","subject":"Re: [PATCH] netfilter: conntrack: use power efficient workqueue","submitter":{"id":22608,"url":"http://patchwork.ozlabs.org/api/people/22608/","name":"Vincent Guittot","email":"vincent.guittot@linaro.org"},"content":"Hi Pablo,\n\nOn 6 November 2017 at 15:56, Pablo Neira Ayuso wrote:\n> On Mon, Nov 06, 2017 at 03:31:55PM +0100, Pablo Neira Ayuso wrote:\n>> On Thu, 2 Nov 2017 15:16:07 +0100 Vincent Guittot wrote:\n>> > conntrack uses the bounded system_long_wq workqueue for its works that\n>> > don't have to run on the cpu they have been queued. Using bounded\n>> > workqueue prevents the scheduler to make smart decision about the best\n>> > place to schedule the work.\n>> >\n>> > This patch replaces system_long_wq with system_power_efficient_wq.\n>> > the work stays bounded to a cpu by default unless the\n>> > CONFIG_WQ_POWER_EFFICIENT is enable. In the latter case, the work can\n>> > be scheduled on the best cpu from a power or a performance point of\n>> > view.\n>>\n>> Applied, thanks.\n>\n> I'm stepping back. According to what I'm reading\n> system_power_efficient_wq becomes system_wq when disabled, which is\n> not semantically the same as system_long_wq that we have now.\n\nWhen disabled, system_power_efficient_wq behaves like system_wq (and\nsystem_long_wq) as the worqueue are bounded to a cpu but It stays a\ndifferent workqueue.\n\n>\n> My concern is that the conntrack garbage collector may run for quite a\n> bit of time. Did you test this with a large conntrack table full of\n\nNo, I haven't done specific tests with a large conntrack table full of entries.\n\nThere is no system_power_efficient_long_wq. I was not convinced that\nwe should create one that's why I have used system_power_efficient_wq\n\nVincent\n\n> entries?\n>\n> Thanks.\n--\nTo unsubscribe from this list: send the line \"unsubscribe netfilter-devel\" in\nthe body of a message to majordomo@vger.kernel.org\nMore majordomo info at http://vger.kernel.org/majordomo-info.html","headers":{"Return-Path":"","X-Original-To":"incoming@patchwork.ozlabs.org","Delivered-To":"patchwork-incoming@bilbo.ozlabs.org","Authentication-Results":["ozlabs.org;\n\tspf=none (mailfrom) smtp.mailfrom=vger.kernel.org\n\t(client-ip=209.132.180.67; helo=vger.kernel.org;\n\tenvelope-from=netfilter-devel-owner@vger.kernel.org;\n\treceiver=)","ozlabs.org;\n\tdkim=fail reason=\"signature verification failed\" (1024-bit key;\n\tunprotected) header.d=linaro.org header.i=@linaro.org\n\theader.b=\"dXgxMz9F\"; dkim-atps=neutral"],"Received":["from vger.kernel.org (vger.kernel.org [209.132.180.67])\n\tby ozlabs.org (Postfix) with ESMTP id 3yVx390hp2z9s7h\n\tfor ;\n\tTue, 7 Nov 2017 02:16:05 +1100 (AEDT)","(majordomo@vger.kernel.org) by vger.kernel.org via listexpand\n\tid S1752101AbdKFPQE (ORCPT );\n\tMon, 6 Nov 2017 10:16:04 -0500","from mail-it0-f67.google.com ([209.85.214.67]:45355 \"EHLO\n\tmail-it0-f67.google.com\" rhost-flags-OK-OK-OK-OK) by vger.kernel.org\n\twith ESMTP id S1751976AbdKFPQD (ORCPT\n\t);\n\tMon, 6 Nov 2017 10:16:03 -0500","by mail-it0-f67.google.com with SMTP id n195so5691663itg.0\n\tfor ;\n\tMon, 06 Nov 2017 07:16:03 -0800 (PST)","by 10.107.146.7 with HTTP; Mon, 6 Nov 2017 07:15:42 -0800 (PST)"],"DKIM-Signature":"v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google;\n\th=mime-version:in-reply-to:references:from:date:message-id:subject:to\n\t:cc; bh=GlyL+/uBIe2IpRs4KZ5Uk7mDB2LgwU94FRTvEC3m/LM=;\n\tb=dXgxMz9F1wmFA5LgSIl3t16arnhQSI0uz41DyITU1L+wZV11+ZDnn6pzArOpQz45Fu\n\tlOyrN7gFdKsanrNtcCvoMtKHpd5+eCKpv9+n00So9qoNc0IsnsswGck9VyucbkZ+KC7A\n\tiFaTcuvwlfIeSWjk6U2NuSJ5+G3oCIad9qQQk=","X-Google-DKIM-Signature":"v=1; a=rsa-sha256; c=relaxed/relaxed;\n\td=1e100.net; s=20161025;\n\th=x-gm-message-state:mime-version:in-reply-to:references:from:date\n\t:message-id:subject:to:cc;\n\tbh=GlyL+/uBIe2IpRs4KZ5Uk7mDB2LgwU94FRTvEC3m/LM=;\n\tb=sQTyCgfvEAyJi99Y60H2TzJOWZUEF8G483bHV4/UHF84fzI+wx+2oHh7MPeCZO0MPR\n\tJyD/29A7LtHEeTQ5RroVruJBQkeW+lqb6eyjgW0w/xDNZCoiH7m37N5dzz3JOUoFLxH7\n\tkGKw/ZFu00b3tSEUlCjuW2ITbIJ1cYa18Hb8yibGE9ZMP1E00WfC6QPBYx9Q04UJK95G\n\tOFmRSBfzzy9NcPnAWAFqoEapTY3pYZa7/TaolmlptnQsnOA+3qlCb42sHDr5YEgHmOA5\n\tGO85yiHxVGq8+T6pD5siwlkdDeYU6zBxv580fF0niE3bIHmoj9qskmtjVOvdAGhqg8GH\n\tJ6CQ==","X-Gm-Message-State":"AJaThX64QASbue9oYeFaUhI6SCwmbyLeIu3i2dQOkSb57QIq+figC5t7\n\thkNhR+xiin7w4yA86ZkMA2HrI+LaI01vINcUQT51Vd0JbnM=","X-Google-Smtp-Source":"ABhQp+Tqwrj8+CtmasVZnWv8PJ7fyO3CTM1IQN4SFck10zzE2Q4BHN8eMeBLZYJBeJYwaMbnhwuQpa3PjtNgCNWqFt0=","X-Received":"by 10.36.117.6 with SMTP id y6mr10208547itc.74.1509981362660;\n\tMon, 06 Nov 2017 07:16:02 -0800 (PST)","MIME-Version":"1.0","In-Reply-To":"<20171106145632.GA20091@salvia>","References":"<1509635767-12253-1-git-send-email-vincent.guittot@linaro.org>\n\t<20171106143155.GA8738@salvia> <20171106145632.GA20091@salvia>","From":"Vincent Guittot ","Date":"Mon, 6 Nov 2017 16:15:42 +0100","Message-ID":"","Subject":"Re: [PATCH] netfilter: conntrack: use power efficient workqueue","To":"Pablo Neira Ayuso ","Cc":"netfilter-devel@vger.kernel.org","Content-Type":"text/plain; charset=\"UTF-8\"","Sender":"netfilter-devel-owner@vger.kernel.org","Precedence":"bulk","List-ID":"","X-Mailing-List":"netfilter-devel@vger.kernel.org"}},{"id":1799783,"web_url":"http://patchwork.ozlabs.org/comment/1799783/","msgid":"<20171106152533.GA27072@salvia>","list_archive_url":null,"date":"2017-11-06T15:25:33","subject":"Re: [PATCH] netfilter: conntrack: use power efficient workqueue","submitter":{"id":1315,"url":"http://patchwork.ozlabs.org/api/people/1315/","name":"Pablo Neira Ayuso","email":"pablo@netfilter.org"},"content":"On Mon, Nov 06, 2017 at 04:15:42PM +0100, Vincent Guittot wrote:\n> Hi Pablo,\n> \n> On 6 November 2017 at 15:56, Pablo Neira Ayuso wrote:\n> > On Mon, Nov 06, 2017 at 03:31:55PM +0100, Pablo Neira Ayuso wrote:\n> >> On Thu, 2 Nov 2017 15:16:07 +0100 Vincent Guittot wrote:\n> >> > conntrack uses the bounded system_long_wq workqueue for its works that\n> >> > don't have to run on the cpu they have been queued. Using bounded\n> >> > workqueue prevents the scheduler to make smart decision about the best\n> >> > place to schedule the work.\n> >> >\n> >> > This patch replaces system_long_wq with system_power_efficient_wq.\n> >> > the work stays bounded to a cpu by default unless the\n> >> > CONFIG_WQ_POWER_EFFICIENT is enable. In the latter case, the work can\n> >> > be scheduled on the best cpu from a power or a performance point of\n> >> > view.\n> >>\n> >> Applied, thanks.\n> >\n> > I'm stepping back. According to what I'm reading\n> > system_power_efficient_wq becomes system_wq when disabled, which is\n> > not semantically the same as system_long_wq that we have now.\n> \n> When disabled, system_power_efficient_wq behaves like system_wq (and\n> system_long_wq) as the worqueue are bounded to a cpu but It stays a\n> different workqueue.\n>\n> > My concern is that the conntrack garbage collector may run for quite a\n> > bit of time. Did you test this with a large conntrack table full of\n> \n> No, I haven't done specific tests with a large conntrack table full of entries.\n> \n> There is no system_power_efficient_long_wq. I was not convinced that\n> we should create one that's why I have used system_power_efficient_wq\n\nMy concern is that this garbage collector may run intensively on busy\nconntrack tables to get rid of expired entries, so my question is if\nswitching from system_long_wq to system_wq is a real issue.\n\nThanks.\n--\nTo unsubscribe from this list: send the line \"unsubscribe netfilter-devel\" in\nthe body of a message to majordomo@vger.kernel.org\nMore majordomo info at http://vger.kernel.org/majordomo-info.html","headers":{"Return-Path":"","X-Original-To":"incoming@patchwork.ozlabs.org","Delivered-To":"patchwork-incoming@bilbo.ozlabs.org","Authentication-Results":"ozlabs.org;\n\tspf=none (mailfrom) smtp.mailfrom=vger.kernel.org\n\t(client-ip=209.132.180.67; helo=vger.kernel.org;\n\tenvelope-from=netfilter-devel-owner@vger.kernel.org;\n\treceiver=)","Received":["from vger.kernel.org (vger.kernel.org [209.132.180.67])\n\tby ozlabs.org (Postfix) with ESMTP id 3yVxGS1QvCz9s7p\n\tfor ;\n\tTue, 7 Nov 2017 02:25:51 +1100 (AEDT)","(majordomo@vger.kernel.org) by vger.kernel.org via listexpand\n\tid S1753940AbdKFPZt (ORCPT );\n\tMon, 6 Nov 2017 10:25:49 -0500","from mail.us.es ([193.147.175.20]:45912 \"EHLO mail.us.es\"\n\trhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP\n\tid S1753921AbdKFPZq (ORCPT );\n\tMon, 6 Nov 2017 10:25:46 -0500","from antivirus1-rhel7.int (unknown [192.168.2.11])\n\tby mail.us.es (Postfix) with ESMTP id 306051BCFC2\n\tfor ;\n\tMon, 6 Nov 2017 16:25:45 +0100 (CET)","from antivirus1-rhel7.int (localhost [127.0.0.1])\n\tby antivirus1-rhel7.int (Postfix) with ESMTP id 20796DA871\n\tfor ;\n\tMon, 6 Nov 2017 16:25:45 +0100 (CET)","by antivirus1-rhel7.int (Postfix, from userid 99)\n\tid 15BA1DA848; Mon, 6 Nov 2017 16:25:45 +0100 (CET)","from antivirus1-rhel7.int (localhost [127.0.0.1])\n\tby antivirus1-rhel7.int (Postfix) with ESMTP id 0D5C7DA848;\n\tMon, 6 Nov 2017 16:25:42 +0100 (CET)","from 192.168.1.97 (192.168.1.97) by antivirus1-rhel7.int\n\t(F-Secure/fsigk_smtp/550/antivirus1-rhel7.int); \n\tMon, 06 Nov 2017 16:25:42 +0100 (CET)","from us.es (unknown [27.1.3.195])\n\t(using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256\n\tbits)) (No client certificate requested)\n\t(Authenticated sender: 1984lsi)\n\tby entrada.int (Postfix) with ESMTPSA id B0E73403DFA0;\n\tMon, 6 Nov 2017 16:25:40 +0100 (CET)"],"X-Spam-Checker-Version":"SpamAssassin 3.4.1 (2015-04-28) on\n\tantivirus1-rhel7.int","X-Spam-Level":"","X-Spam-Status":"No, score=-108.2 required=7.5 tests=ALL_TRUSTED,BAYES_50,\n\tHEADER_FROM_DIFFERENT_DOMAINS,SMTPAUTH_US2,USER_IN_WHITELIST\n\tautolearn=disabled version=3.4.1","X-Virus-Status":"clean(F-Secure/fsigk_smtp/550/antivirus1-rhel7.int)","Date":"Mon, 6 Nov 2017 16:25:33 +0100","X-SMTPAUTHUS":"auth mail.us.es","From":"Pablo Neira Ayuso ","To":"Vincent Guittot ","Cc":"netfilter-devel@vger.kernel.org","Subject":"Re: [PATCH] netfilter: conntrack: use power efficient workqueue","Message-ID":"<20171106152533.GA27072@salvia>","References":"<1509635767-12253-1-git-send-email-vincent.guittot@linaro.org>\n\t<20171106143155.GA8738@salvia> <20171106145632.GA20091@salvia>\n\t","MIME-Version":"1.0","Content-Type":"text/plain; charset=us-ascii","Content-Disposition":"inline","In-Reply-To":"","User-Agent":"Mutt/1.5.23 (2014-03-12)","X-Virus-Scanned":"ClamAV using ClamSMTP","Sender":"netfilter-devel-owner@vger.kernel.org","Precedence":"bulk","List-ID":"","X-Mailing-List":"netfilter-devel@vger.kernel.org"}},{"id":1799822,"web_url":"http://patchwork.ozlabs.org/comment/1799822/","msgid":"","list_archive_url":null,"date":"2017-11-06T16:33:43","subject":"Re: [PATCH] netfilter: conntrack: use power efficient workqueue","submitter":{"id":22608,"url":"http://patchwork.ozlabs.org/api/people/22608/","name":"Vincent Guittot","email":"vincent.guittot@linaro.org"},"content":"On 6 November 2017 at 16:25, Pablo Neira Ayuso wrote:\n> On Mon, Nov 06, 2017 at 04:15:42PM +0100, Vincent Guittot wrote:\n>> Hi Pablo,\n>>\n>> On 6 November 2017 at 15:56, Pablo Neira Ayuso wrote:\n>> > On Mon, Nov 06, 2017 at 03:31:55PM +0100, Pablo Neira Ayuso wrote:\n>> >> On Thu, 2 Nov 2017 15:16:07 +0100 Vincent Guittot wrote:\n>> >> > conntrack uses the bounded system_long_wq workqueue for its works that\n>> >> > don't have to run on the cpu they have been queued. Using bounded\n>> >> > workqueue prevents the scheduler to make smart decision about the best\n>> >> > place to schedule the work.\n>> >> >\n>> >> > This patch replaces system_long_wq with system_power_efficient_wq.\n>> >> > the work stays bounded to a cpu by default unless the\n>> >> > CONFIG_WQ_POWER_EFFICIENT is enable. In the latter case, the work can\n>> >> > be scheduled on the best cpu from a power or a performance point of\n>> >> > view.\n>> >>\n>> >> Applied, thanks.\n>> >\n>> > I'm stepping back. According to what I'm reading\n>> > system_power_efficient_wq becomes system_wq when disabled, which is\n>> > not semantically the same as system_long_wq that we have now.\n>>\n>> When disabled, system_power_efficient_wq behaves like system_wq (and\n>> system_long_wq) as the worqueue are bounded to a cpu but It stays a\n>> different workqueue.\n>>\n>> > My concern is that the conntrack garbage collector may run for quite a\n>> > bit of time. Did you test this with a large conntrack table full of\n>>\n>> No, I haven't done specific tests with a large conntrack table full of entries.\n>>\n>> There is no system_power_efficient_long_wq. I was not convinced that\n>> we should create one that's why I have used system_power_efficient_wq\n>\n> My concern is that this garbage collector may run intensively on busy\n> conntrack tables to get rid of expired entries, so my question is if\n> switching from system_long_wq to system_wq is a real issue.\n\nsystem_long_wq and system_wq have the same configuration. They are\njust named differently so people will use one or the other but they\nbehave exactly the same.\nThen system_power_efficient_wq is a 3rd workqueue that behaves like\nsystem_wq and system_long_wq when CONFIG_WQ_POWER_EFFICIENT is disable\nbut uses ubounded workqueue when enabled\n\nI don't think that switching from system_long_wq to\nsystem_power_efficient_wq is an issue.\nI haven't seen any problem so far but i'm haven't done specific tests\nthat stress the system with a large conntrack table full of entries.\nFurthermore, these workqueue use default max active context (256)\nwhich means that other work should be able to run.\n\nRegards,\nVincent\n\n\n>\n> Thanks.\n--\nTo unsubscribe from this list: send the line \"unsubscribe netfilter-devel\" in\nthe body of a message to majordomo@vger.kernel.org\nMore majordomo info at http://vger.kernel.org/majordomo-info.html","headers":{"Return-Path":"","X-Original-To":"incoming@patchwork.ozlabs.org","Delivered-To":"patchwork-incoming@bilbo.ozlabs.org","Authentication-Results":["ozlabs.org;\n\tspf=none (mailfrom) smtp.mailfrom=vger.kernel.org\n\t(client-ip=209.132.180.67; helo=vger.kernel.org;\n\tenvelope-from=netfilter-devel-owner@vger.kernel.org;\n\treceiver=)","ozlabs.org;\n\tdkim=fail reason=\"signature verification failed\" (1024-bit key;\n\tunprotected) header.d=linaro.org header.i=@linaro.org\n\theader.b=\"WPSuCc0J\"; dkim-atps=neutral"],"Received":["from vger.kernel.org (vger.kernel.org [209.132.180.67])\n\tby ozlabs.org (Postfix) with ESMTP id 3yVynB1m24z9rxl\n\tfor ;\n\tTue, 7 Nov 2017 03:34:06 +1100 (AEDT)","(majordomo@vger.kernel.org) by vger.kernel.org via listexpand\n\tid S1753979AbdKFQeF (ORCPT );\n\tMon, 6 Nov 2017 11:34:05 -0500","from mail-io0-f194.google.com ([209.85.223.194]:48029 \"EHLO\n\tmail-io0-f194.google.com\" rhost-flags-OK-OK-OK-OK) by vger.kernel.org\n\twith ESMTP id S1753827AbdKFQeE (ORCPT\n\t);\n\tMon, 6 Nov 2017 11:34:04 -0500","by mail-io0-f194.google.com with SMTP id h70so16304814ioi.4\n\tfor ;\n\tMon, 06 Nov 2017 08:34:04 -0800 (PST)","by 10.107.146.7 with HTTP; Mon, 6 Nov 2017 08:33:43 -0800 (PST)"],"DKIM-Signature":"v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google;\n\th=mime-version:in-reply-to:references:from:date:message-id:subject:to\n\t:cc; bh=7mA82yFAzgvkfriqpS0Oc2MV90Fz/REDPqrLAQZphvo=;\n\tb=WPSuCc0J9bBjV0dJM3VmMZzl/Khzc0MSw6posDGsoeZQnqg+LXBuhDog+R9DXzK1Yr\n\tN3DuUg9foCkNl8tWnp3QxF/4BfV7e1CmhZiC82T/xJaOe24VwefnKo7e2XdYqyxvPIWG\n\tdKTfThcS8wC0HMpWw4ib35JFtncFs8oNn//T4=","X-Google-DKIM-Signature":"v=1; a=rsa-sha256; c=relaxed/relaxed;\n\td=1e100.net; s=20161025;\n\th=x-gm-message-state:mime-version:in-reply-to:references:from:date\n\t:message-id:subject:to:cc;\n\tbh=7mA82yFAzgvkfriqpS0Oc2MV90Fz/REDPqrLAQZphvo=;\n\tb=l3vkvb0U+QTbDiy5/1E5DPdf05s0QnAhN2IfXzG8OD98ov1nFHruwgcdpGEkgkXFy6\n\tqflelNyT1/1vXCLkDqKMtgzJ+6dP6ptbJPkQNTI5CbfRP+k8Y+7o4n4exQiNALNAuSgE\n\tiOJH7N201zQ1jG2gA73fApkSFKLUozuEspWgaJkz9OXk161IAkv7VHaSccU9O9NCAw0h\n\tCF0eAMcsufSWdW3HeXl4bCTOzdc8mz+gOP5eKZGDxsHMRBaaTHv8nlZRIHFJbvhSFY18\n\tFiy6zPwdlB/XNvC27uFIBvnZbgelEbyJhAehSkhjSGVLeZ8BsApS7SIdTfwvFJN3iEBl\n\tNoMg==","X-Gm-Message-State":"AMCzsaUKSq7MaX7H2v7eZsMnXJoZ+dtEq4FZEz/L+b1XpR7v+v9uEHW/\n\tmaqOyrP09h9Zy22MuUW943cxuA60E+lJ1ZAVLH+wQMS03zk=","X-Google-Smtp-Source":"ABhQp+SMwiiua5rZwqUQz3D8TPuhHZnzoXeL1GN3bS5wAqSAS+SgbEr1yrs72PrGfVkaAZ48ld7koBYsug3wjOOZstA=","X-Received":"by 10.107.69.25 with SMTP id s25mr19941512ioa.189.1509986043880; \n\tMon, 06 Nov 2017 08:34:03 -0800 (PST)","MIME-Version":"1.0","In-Reply-To":"<20171106152533.GA27072@salvia>","References":"<1509635767-12253-1-git-send-email-vincent.guittot@linaro.org>\n\t<20171106143155.GA8738@salvia> <20171106145632.GA20091@salvia>\n\t\n\t<20171106152533.GA27072@salvia>","From":"Vincent Guittot ","Date":"Mon, 6 Nov 2017 17:33:43 +0100","Message-ID":"","Subject":"Re: [PATCH] netfilter: conntrack: use power efficient workqueue","To":"Pablo Neira Ayuso ","Cc":"netfilter-devel@vger.kernel.org","Content-Type":"text/plain; charset=\"UTF-8\"","Sender":"netfilter-devel-owner@vger.kernel.org","Precedence":"bulk","List-ID":"","X-Mailing-List":"netfilter-devel@vger.kernel.org"}},{"id":1800133,"web_url":"http://patchwork.ozlabs.org/comment/1800133/","msgid":"<20171107000226.GA1364@salvia>","list_archive_url":null,"date":"2017-11-07T00:02:26","subject":"Re: [PATCH] netfilter: conntrack: use power efficient workqueue","submitter":{"id":1315,"url":"http://patchwork.ozlabs.org/api/people/1315/","name":"Pablo Neira Ayuso","email":"pablo@netfilter.org"},"content":"On Mon, Nov 06, 2017 at 05:33:43PM +0100, Vincent Guittot wrote:\n> On 6 November 2017 at 16:25, Pablo Neira Ayuso wrote:\n> > On Mon, Nov 06, 2017 at 04:15:42PM +0100, Vincent Guittot wrote:\n> >> Hi Pablo,\n> >>\n> >> On 6 November 2017 at 15:56, Pablo Neira Ayuso wrote:\n> >> > On Mon, Nov 06, 2017 at 03:31:55PM +0100, Pablo Neira Ayuso wrote:\n> >> >> On Thu, 2 Nov 2017 15:16:07 +0100 Vincent Guittot wrote:\n> >> >> > conntrack uses the bounded system_long_wq workqueue for its works that\n> >> >> > don't have to run on the cpu they have been queued. Using bounded\n> >> >> > workqueue prevents the scheduler to make smart decision about the best\n> >> >> > place to schedule the work.\n> >> >> >\n> >> >> > This patch replaces system_long_wq with system_power_efficient_wq.\n> >> >> > the work stays bounded to a cpu by default unless the\n> >> >> > CONFIG_WQ_POWER_EFFICIENT is enable. In the latter case, the work can\n> >> >> > be scheduled on the best cpu from a power or a performance point of\n> >> >> > view.\n> >> >>\n> >> >> Applied, thanks.\n> >> >\n> >> > I'm stepping back. According to what I'm reading\n> >> > system_power_efficient_wq becomes system_wq when disabled, which is\n> >> > not semantically the same as system_long_wq that we have now.\n> >>\n> >> When disabled, system_power_efficient_wq behaves like system_wq (and\n> >> system_long_wq) as the worqueue are bounded to a cpu but It stays a\n> >> different workqueue.\n> >>\n> >> > My concern is that the conntrack garbage collector may run for quite a\n> >> > bit of time. Did you test this with a large conntrack table full of\n> >>\n> >> No, I haven't done specific tests with a large conntrack table full of entries.\n> >>\n> >> There is no system_power_efficient_long_wq. I was not convinced that\n> >> we should create one that's why I have used system_power_efficient_wq\n> >\n> > My concern is that this garbage collector may run intensively on busy\n> > conntrack tables to get rid of expired entries, so my question is if\n> > switching from system_long_wq to system_wq is a real issue.\n> \n> system_long_wq and system_wq have the same configuration. They are\n> just named differently so people will use one or the other but they\n> behave exactly the same.\n\nI see, this confirms what I've been reading in the code, so it's just\nanother queue with a different name.\n\n> Then system_power_efficient_wq is a 3rd workqueue that behaves like\n> system_wq and system_long_wq when CONFIG_WQ_POWER_EFFICIENT is disable\n> but uses ubounded workqueue when enabled\n> \n> I don't think that switching from system_long_wq to\n> system_power_efficient_wq is an issue.\n> I haven't seen any problem so far but i'm haven't done specific tests\n> that stress the system with a large conntrack table full of entries.\n> Furthermore, these workqueue use default max active context (256)\n> which means that other work should be able to run.\n\nThanks for explaining, I think then system_long_wq is missing a bit\nits purpose, but this is out of my scope a bit, so I have pushed out\nthis to nf-next.git. Thanks!\n--\nTo unsubscribe from this list: send the line \"unsubscribe netfilter-devel\" in\nthe body of a message to majordomo@vger.kernel.org\nMore majordomo info at http://vger.kernel.org/majordomo-info.html","headers":{"Return-Path":"","X-Original-To":"incoming@patchwork.ozlabs.org","Delivered-To":"patchwork-incoming@bilbo.ozlabs.org","Authentication-Results":"ozlabs.org;\n\tspf=none (mailfrom) smtp.mailfrom=vger.kernel.org\n\t(client-ip=209.132.180.67; helo=vger.kernel.org;\n\tenvelope-from=netfilter-devel-owner@vger.kernel.org;\n\treceiver=)","Received":["from vger.kernel.org (vger.kernel.org [209.132.180.67])\n\tby ozlabs.org (Postfix) with ESMTP id 3yW8kp1MdVz9ryv\n\tfor ;\n\tTue, 7 Nov 2017 11:02:42 +1100 (AEDT)","(majordomo@vger.kernel.org) by vger.kernel.org via listexpand\n\tid S933092AbdKGACk (ORCPT );\n\tMon, 6 Nov 2017 19:02:40 -0500","from mail.us.es ([193.147.175.20]:52156 \"EHLO mail.us.es\"\n\trhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP\n\tid S932831AbdKGACi (ORCPT ); \n\tMon, 6 Nov 2017 19:02:38 -0500","from antivirus1-rhel7.int (unknown [192.168.2.11])\n\tby mail.us.es (Postfix) with ESMTP id B6CF41F4B65\n\tfor ;\n\tTue, 7 Nov 2017 01:02:36 +0100 (CET)","from antivirus1-rhel7.int (localhost [127.0.0.1])\n\tby antivirus1-rhel7.int (Postfix) with ESMTP id A4CF8DA7F1\n\tfor ;\n\tTue, 7 Nov 2017 01:02:36 +0100 (CET)","by antivirus1-rhel7.int (Postfix, from userid 99)\n\tid 9A707DA807; Tue, 7 Nov 2017 01:02:36 +0100 (CET)","from antivirus1-rhel7.int (localhost [127.0.0.1])\n\tby antivirus1-rhel7.int (Postfix) with ESMTP id 7581EDA7F1;\n\tTue, 7 Nov 2017 01:02:32 +0100 (CET)","from 192.168.1.97 (192.168.1.97) by antivirus1-rhel7.int\n\t(F-Secure/fsigk_smtp/550/antivirus1-rhel7.int); \n\tTue, 07 Nov 2017 01:02:32 +0100 (CET)","from us.es (unknown [61.40.109.130])\n\t(using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256\n\tbits)) (No client certificate requested)\n\t(Authenticated sender: 1984lsi)\n\tby entrada.int (Postfix) with ESMTPSA id 78570404FD19;\n\tTue, 7 Nov 2017 01:02:31 +0100 (CET)"],"X-Spam-Checker-Version":"SpamAssassin 3.4.1 (2015-04-28) on\n\tantivirus1-rhel7.int","X-Spam-Level":"","X-Spam-Status":"No, score=-108.2 required=7.5 tests=ALL_TRUSTED,BAYES_50,\n\tHEADER_FROM_DIFFERENT_DOMAINS,SMTPAUTH_US2,USER_IN_WHITELIST\n\tautolearn=disabled version=3.4.1","X-Virus-Status":"clean(F-Secure/fsigk_smtp/550/antivirus1-rhel7.int)","Date":"Tue, 7 Nov 2017 01:02:26 +0100","X-SMTPAUTHUS":"auth mail.us.es","From":"Pablo Neira Ayuso ","To":"Vincent Guittot ","Cc":"netfilter-devel@vger.kernel.org","Subject":"Re: [PATCH] netfilter: conntrack: use power efficient workqueue","Message-ID":"<20171107000226.GA1364@salvia>","References":"<1509635767-12253-1-git-send-email-vincent.guittot@linaro.org>\n\t<20171106143155.GA8738@salvia> <20171106145632.GA20091@salvia>\n\t\n\t<20171106152533.GA27072@salvia>\n\t","MIME-Version":"1.0","Content-Type":"text/plain; charset=us-ascii","Content-Disposition":"inline","In-Reply-To":"","User-Agent":"Mutt/1.5.23 (2014-03-12)","X-Virus-Scanned":"ClamAV using ClamSMTP","Sender":"netfilter-devel-owner@vger.kernel.org","Precedence":"bulk","List-ID":"","X-Mailing-List":"netfilter-devel@vger.kernel.org"}}]