From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: Multicast packet loss Date: Wed, 08 Apr 2009 10:12:33 +0200 Message-ID: <49DC5C71.2040505@cosmosbay.com> References: <49B4B909.7050002@cosmosbay.com> <20090313.145152.121603300.davem@davemloft.net> <49BADE87.40407@cosmosbay.com> <20090313.153851.11725991.davem@davemloft.net> <49BED109.3020504@cosmosbay.com> <49D66379.7070106@athenacr.com> <49D8B6DA.7050902@cosmosbay.com> <49DBB2BA.1010100@athenacr.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: David Miller , kchang@athenacr.com, netdev@vger.kernel.org, cl@linux-foundation.org To: Brian Bloniarz Return-path: Received: from gw1.cosmosbay.com ([212.99.114.194]:43157 "EHLO gw1.cosmosbay.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1761851AbZDHIN5 convert rfc822-to-8bit (ORCPT ); Wed, 8 Apr 2009 04:13:57 -0400 In-Reply-To: <49DBB2BA.1010100@athenacr.com> Sender: netdev-owner@vger.kernel.org List-ID: Brian Bloniarz a =E9crit : > Eric Dumazet wrote: >> Brian Bloniarz a =E9crit : >>> We've been experimenting with this softirq-delay patch in productio= n, > and >>> have seen some hard-to-reproduce crashes. We finally managed to > capture a >>> kexec crashdump this morning. >> >> Pointer being null might tell us that we managed to call > inet_def_readable() >> without socket lock hold... >=20 > False alarm -- I think I did the backport to 2.6.24 incorrectly. 2.6.= 24 was > before the UDP receive path started taking the socket lock, so > inet_def_readable's assumption doesn't hold. >=20 > Sorry to waste everyone's time. >=20 Thanks for doing this discovery work and analysis.=20 I am currently off-computers and could not do this until next week. So, if you want to use 2.6.24, we need to back port other patches as we= ll ?