From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andi Kleen Subject: Re: [PATCH net-next-2.6] inetpeer: lower false sharing effect Date: Sat, 11 Jun 2011 09:09:53 +0200 Message-ID: <20110611070952.GA24330@one.firstfloor.org> References: <1307600810.3980.85.camel@edumazet-laptop> <1307664235.17300.44.camel@schen9-DESK> <20110609.204330.2090335955971650557.davem@davemloft.net> <1307745190.17300.85.camel@schen9-DESK> <1307768052.2872.50.camel@edumazet-laptop> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Eric Dumazet , Tim Chen , David Miller , netdev@vger.kernel.org, andi@firstfloor.org To: Changli Gao Return-path: Received: from one.firstfloor.org ([213.235.205.2]:32928 "EHLO one.firstfloor.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751546Ab1FKHJ4 (ORCPT ); Sat, 11 Jun 2011 03:09:56 -0400 Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Sat, Jun 11, 2011 at 02:17:24PM +0800, Changli Gao wrote: > On Sat, Jun 11, 2011 at 12:54 PM, Eric Dumazet wrote: > > Le samedi 11 juin 2011 =E0 08:54 +0800, Changli Gao a =E9crit : > > > > > > Whole point of the exercice is to prepare ground for routing cache > > removal :) > > > > If you want a server being hit by millions of clients around the wo= rld, > > routing cache is a real pain because of memory needs. > > >=20 > Yes. I know the routing cache removal is our goal. But for his > scenario, there aren't so many routing cache entries, so routing cach= e > may be a better option currently. However, if he just wants to The routing cache has terrible cache line bouncing in the reference count too. -andi