From mboxrd@z Thu Jan 1 00:00:00 1970 From: Fabian Hugelshofer Subject: Re: Conntrack Events Performance - Multipart Messages? Date: Fri, 25 Jul 2008 09:44:55 +0100 Message-ID: <48899287.50402@gmx.ch> References: <487E24FC.60700@gmx.ch> <487F18DA.7030208@netfilter.org> <487FFBEE.90409@trash.net> <4884B068.4050306@gmx.ch> <4884B270.5010104@trash.net> <4884CC17.3020905@gmx.ch> <488740E7.3040005@gmx.ch> <48874272.1020503@trash.net> <48875887.8040209@gmx.ch> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Cc: Patrick McHardy , Pablo Neira Ayuso To: netfilter-devel@vger.kernel.org Return-path: Received: from mail.gmx.net ([213.165.64.20]:59768 "HELO mail.gmx.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1755470AbYGYIo6 (ORCPT ); Fri, 25 Jul 2008 04:44:58 -0400 In-Reply-To: <48875887.8040209@gmx.ch> Sender: netfilter-devel-owner@vger.kernel.org List-ID: Fabian Hugelshofer wrote: > Patrick McHardy wrote: >> The first thing to try would be to use sane allocation sizes >> for the event messages. This patch doesn't implement it properly >> (uses probing), but should be enough to test whether it helps. > > Thanks a lot. This patch already decreased the CPU usage for ctevtest > from 85% to 44%. Sweet... I just rerun the test. The 85% CPU usage was with 1700pps and 44% with 1500pps. The wireless channel does not provide the same performance all the time and I did not pay attention to the packet rate. Without the allocation patch and 1500pps the CPU usage is 56%. But 10% less is still quite nice.