From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [2.6.33-rc5 regression] NULL pointer dereference in vlan_skb_recv - probably introduced by commit 9793241fe92f7d9303fb221e43fc598eb065f267 Date: Sun, 24 Jan 2010 14:42:54 +0100 Message-ID: <4B5C4E5E.2010507@gmail.com> References: <20100123165657.187c11e4@neptune.home> <20100123223132.0e62d8cb@neptune.home> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: "David S. Miller" , netdev@vger.kernel.org, linux-kernel@vger.kernel.org To: =?ISO-8859-1?Q?Bruno_Pr=E9mont?= Return-path: Received: from gw1.cosmosbay.com ([212.99.114.194]:56593 "EHLO gw1.cosmosbay.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932102Ab0AXNnI (ORCPT ); Sun, 24 Jan 2010 08:43:08 -0500 In-Reply-To: <20100123223132.0e62d8cb@neptune.home> Sender: netdev-owner@vger.kernel.org List-ID: Le 23/01/2010 22:31, Bruno Pr=E9mont a =E9crit : >> Above part of code did change between 2.6.32 and 2.6.33-rc5 with >> commit 9793241fe92f7d9303fb221e43fc598eb065f267 (vlan: Precise RX >> stats accounting) >> http://git.kernel.org/?p=3Dlinux/kernel/git/torvalds/linux-2.6.git;a= =3Dcommitdiff;h=3D9793241fe92f7d9303fb221e43fc598eb065f267 >=20 > Reverting just that commit gets the system running correctly. >=20 > Bruno I have no idea how this patch can break vlan networking. Your disassembly and .config seems to show your machine is not SMP Maybe something is broken on UP and alloc_percpu() ? Could you add a debug in vlan_dev_init() vlan_dev_info(dev)->vlan_rx_stats =3D alloc_percpu(struct vlan_rx_sta= ts); if (!vlan_dev_info(dev)->vlan_rx_stats) return -ENOMEM; + pr_err("vlan_dev_init() rx_stats=3D%p\n", vlan_dev_info(dev)->vlan_rx= _stats); This make sure vlan_dev_init() is called and percpu allocation is prope= rly done. Thanks