From mboxrd@z Thu Jan 1 00:00:00 1970 From: Rick Jones Subject: Re: Performance regression on kernels 3.10 and newer Date: Mon, 18 Aug 2014 08:29:41 -0700 Message-ID: <53F21BE5.1030605@hp.com> References: <53ECFDAB.5010701@intel.com> <1408041962.6804.31.camel@edumazet-glaptop2.roam.corp.google.com> <53ED4354.9090904@intel.com> <20140814.162024.2218312002979492106.davem@davemloft.net> <53EE4023.6080902@intel.com> <53EE5B25.3040206@intel.com> <53EE967F.9090101@intel.com> <063D6719AE5E284EB5DD2968C1650D6D1747969E@AcuExch.aculab.com> <53F21A38.50905@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Cc: David Miller , Eric Dumazet , Linux Netdev List To: Alexander Duyck , David Laight , Tom Herbert Return-path: Received: from g2t2354.austin.hp.com ([15.217.128.53]:24907 "EHLO g2t2354.austin.hp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750834AbaHRP3o (ORCPT ); Mon, 18 Aug 2014 11:29:44 -0400 In-Reply-To: <53F21A38.50905@intel.com> Sender: netdev-owner@vger.kernel.org List-ID: Alex - Why don't you go ahead and shoot me a patch to netperf to conditionally put a poll() or select() in front of receive. I guess the "omni" code would be the place to do it. The netperf-talk or netperf-dev lists would be the place to send it I suspect. happy benchmarking, rick jones