From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [tbench regression fixes]: digging out smelly deadmen. Date: Fri, 31 Oct 2008 12:10:10 +0100 Message-ID: <490AE792.60200@cosmosbay.com> References: <20081031.005219.141937694.davem@davemloft.net> <20081031.025159.51432990.davem@davemloft.net> <490AE1CD.9040207@cosmosbay.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: David Miller , shemminger@vyatta.com, zbr@ioremap.net, rjw@sisk.pl, mingo@elte.hu, s0mbre@tservice.net.ru, a.p.zijlstra@chello.nl, LKML , Netdev , efault@gmx.de, Andrew Morton To: =?ISO-8859-1?Q?Ilpo_J=E4rvinen?= Return-path: Received: from gw1.cosmosbay.com ([86.65.150.130]:49028 "EHLO gw1.cosmosbay.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750790AbYJaLLI convert rfc822-to-8bit (ORCPT ); Fri, 31 Oct 2008 07:11:08 -0400 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: Ilpo J=E4rvinen a =E9crit : > On Fri, 31 Oct 2008, Eric Dumazet wrote: >=20 >> David Miller a =E9crit : >> > From: "Ilpo J=E4rvinen" >> > Date: Fri, 31 Oct 2008 11:40:16 +0200 (EET) >> > > > Let me remind that it is just a single process, so no ping-pon= g=20 >> & other >> > > lock related cache effects should play any significant role here= ,=20 >> no? (I'm >> > > no expert though :-)). >> > > Not locks or ping-pongs perhaps, I guess. So it just sends and >> > receives over a socket, implementing both ends of the communicatio= n >> > in the same process? >> > > If hash chain conflicts do happen for those 2 sockets, just=20 >> traversing >> > the chain 2 entries deep could show up. >> >> tbench is very sensible to cache line ping-pongs (on SMP machines of= =20 >> course) >=20 > ...Sorry to disappoint you but we were discussion there on my AIM9=20 > tcp_test results :-). >=20 Well, before you added AIM9 on this topic, we were focusing on tbench := ) Sorry to disappoint you :)