From mboxrd@z Thu Jan 1 00:00:00 1970 From: Changli Gao Subject: Re: rps perfomance WAS(Re: rps: question Date: Thu, 15 Apr 2010 10:50:24 +0800 Message-ID: References: <1265568122.3688.36.camel@bigi> <1271271222.4567.51.camel@bigi> <20100414124426.6aee95c3@nehalam> <1271276568.4567.59.camel@bigi> <1271276855.16881.1756.camel@edumazet-laptop> <1271278661.16881.1761.camel@edumazet-laptop> <20100414160220.43a62999@nehalam> <1271299206.16881.1764.camel@edumazet-laptop> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Stephen Hemminger , Tom Herbert , hadi@cyberus.ca, netdev@vger.kernel.org, robert@herjulf.net, David Miller , Andi Kleen To: Eric Dumazet Return-path: Received: from mail-pv0-f174.google.com ([74.125.83.174]:43781 "EHLO mail-pv0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756519Ab0DOCup convert rfc822-to-8bit (ORCPT ); Wed, 14 Apr 2010 22:50:45 -0400 Received: by pva18 with SMTP id 18so540778pva.19 for ; Wed, 14 Apr 2010 19:50:44 -0700 (PDT) In-Reply-To: <1271299206.16881.1764.camel@edumazet-laptop> Sender: netdev-owner@vger.kernel.org List-ID: On Thu, Apr 15, 2010 at 10:40 AM, Eric Dumazet = wrote: > > Agree 100%, and irqbalance is the existing daemon. It should be used = and > changed if necessary. > > Changli, my stronges argument about your patches is that our schedule= r > and memory affinity api (numactl driven) is bitmask oriented, giving = the > same weight to individual cpu or individual memory node. > It works with the assumption: the workloads handled in non-schedulable context are less than the others. If most of work is done in non-schedulable(softirq) context, scheduler can't keep load balance. --=20 Regards=EF=BC=8C Changli Gao(xiaosuo@gmail.com)