From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: 2.6.39-rc7-git11, x86/32, failed on ppp2897'th interface, PERCPU: allocation failed Date: Thu, 19 May 2011 09:44:31 +0200 Message-ID: <1305791071.3028.1.camel@edumazet-laptop> References: <1305787158.3019.12.camel@edumazet-laptop> <1305788113.3019.19.camel@edumazet-laptop> <10f61af229a48d77d06d154b4647cdde@visp.net.lb> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: netdev@vger.kernel.org To: Denys Fedoryshchenko Return-path: Received: from mail-ww0-f44.google.com ([74.125.82.44]:39944 "EHLO mail-ww0-f44.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753299Ab1ESHof (ORCPT ); Thu, 19 May 2011 03:44:35 -0400 Received: by wwa36 with SMTP id 36so2570870wwa.1 for ; Thu, 19 May 2011 00:44:34 -0700 (PDT) In-Reply-To: <10f61af229a48d77d06d154b4647cdde@visp.net.lb> Sender: netdev-owner@vger.kernel.org List-ID: Le jeudi 19 mai 2011 =C3=A0 10:28 +0300, Denys Fedoryshchenko a =C3=A9c= rit : > On Thu, 19 May 2011 08:55:13 +0200, Eric Dumazet wrote: > > Le jeudi 19 mai 2011 =C3=A0 08:39 +0200, Eric Dumazet a =C3=A9crit = : > > > >> Its a known problem : When ipv6 is enabled, we allocate percpu=20 > >> memory to > >> hold per device snmp counters. > >> > >> make sure kernel idea of max possible cpus matches real number of=20 > >> cpus. > >> > >> And yes, switching to 64bit kernel helps a lot. > >> > >> > > > > Looking at snmp6_alloc_dev(), we allocate three mib per device : > > > > ipstats_mib (30 * sizeof(u64) * number_of_possible_cpus) > > icmpv6_mib (4 * sizeof(long) * number_of_possible_cpus) > > icmpv6msg_mib (26 * sizeof(long)) > 1920 + > 256 + > 208 =3D 2386 * 3000ppp's =3D 7152000, i think it is not that much at= any=20 > case, if i am not wrong. >=20 > But at any case i will try 64bit. If you really want to stay 32bit, you might try to enlarge vmalloc area= d (128 Mbytes default) to get room for pcpu data : grep pcpu /proc/vmallocinfo=20 boot param : vmalloc=3D256M