From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755422Ab1LVRlk (ORCPT ); Thu, 22 Dec 2011 12:41:40 -0500 Received: from mail-ww0-f44.google.com ([74.125.82.44]:59727 "EHLO mail-ww0-f44.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752183Ab1LVRli (ORCPT ); Thu, 22 Dec 2011 12:41:38 -0500 Message-ID: <1324575692.2586.6.camel@edumazet-laptop> Subject: Re: BUG: unable to handle kernel NULL pointer dereference in ipv6_select_ident From: Eric Dumazet To: Chris Boot Cc: David Miller , lkml , netdev , Steffen Klassert Date: Thu, 22 Dec 2011 18:41:32 +0100 In-Reply-To: <4EF352A2.2090604@bootc.net> References: <4EF200BB.7000209@bootc.net> <1324484956.2301.24.camel@edumazet-HP-Compaq-6005-Pro-SFF-PC> <4EF2117F.6000803@bootc.net> <1324488984.2301.45.camel@edumazet-HP-Compaq-6005-Pro-SFF-PC> <1324490401.2301.46.camel@edumazet-HP-Compaq-6005-Pro-SFF-PC> <4EF23BF2.4000601@bootc.net> <1324499332.2621.7.camel@edumazet-laptop> <1324500775.2621.9.camel@edumazet-laptop> <4EF2568C.6040006@bootc.net> <1324528656.2621.19.camel@edumazet-laptop> <4EF300B2.3050903@bootc.net> <1324563353.2153.27.camel@edumazet-HP-Compaq-6005-Pro-SFF-PC> <4EF352A2.2090604@bootc.net> Content-Type: text/plain; charset="UTF-8" X-Mailer: Evolution 3.2.1- Content-Transfer-Encoding: 8bit Mime-Version: 1.0 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Le jeudi 22 décembre 2011 à 15:54 +0000, Chris Boot a écrit : > Eric, > > I'm seeing a new problem now with IPv6 on my bridge, I don't know if > it's related to any of the patches you gave me. > > Basically, I have eth0 and eth1 in a balance-rr bond (bond0), and this > is added as a port in a bridge. When the bond and bridge are freshly > brought up, IPv4 works fine but it appears to be deaf to IPv6 traffic. > It stays this way until I run 'tcpdump -i eth0' and 'tcpdump -i eth1'. > If I run tcpdump with the -p flag to not enable promiscuous mode, the > interfaces remain deaf. Only if they enter promiscuous mode do they > start to listen to each other. It also doesn't help to run tcpdump > against bond0 or br0 at all. > > I'm not sure if once I've killed tcpdump they become deaf again, but I > have enough IPv6 traffic to keep the neighbour entry alive. Thats probably a separate issue. It would be better to open a new thread on netdev, instead of continuing this one. IPv6 needs multicast capabilities, so maybe balance-rr bonding mode is not propagating multicat filters properly on both nics. It might be a problem on one of your NIC. (tcpdump automatically installs promiscous mode, so all multicast frames are received, regardless of previous multicast filters in place)