From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [NETFILTER]: bridge: fix missing link layer headers on outgoing routed packets Date: Thu, 13 Dec 2007 09:32:46 -0800 (PST) Message-ID: <20071213.093246.83297959.davem@davemloft.net> References: <47616BC6.9010208@trash.net> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: netfilter-devel@vger.kernel.org To: kaber@trash.net Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:36233 "EHLO sunset.davemloft.net" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1755010AbXLMRcq (ORCPT ); Thu, 13 Dec 2007 12:32:46 -0500 In-Reply-To: <47616BC6.9010208@trash.net> Sender: netfilter-devel-owner@vger.kernel.org List-ID: From: Patrick McHardy Date: Thu, 13 Dec 2007 18:28:38 +0100 > [NETFILTER]: bridge: fix missing link layer headers on outgoing routed packets > > As reported by Damien Thebault, the double POSTROUTING hook invocation > fix caused outgoing packets routed between two bridges to appear without > a link-layer header. The reason for this is that we're skipping the > br_nf_post_routing hook for routed packets now and don't save the > original link layer header, but nevertheless tries to restore it on > output, causing corruption. > > The root cause for this is that skb->nf_bridge has no clearly defined > lifetime and is used to indicate all kind of things, but that is > quite complicated to fix. For now simply don't touch these packets > and handle them like packets from any other device. > > Tested-by: Damien Thebault > Signed-off-by: Patrick McHardy Applied to net-2.6, thanks Patrick.