From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750885AbWJVXxe (ORCPT ); Sun, 22 Oct 2006 19:53:34 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750894AbWJVXxe (ORCPT ); Sun, 22 Oct 2006 19:53:34 -0400 Received: from terminus.zytor.com ([192.83.249.54]:42893 "EHLO terminus.zytor.com") by vger.kernel.org with ESMTP id S1750885AbWJVXxd (ORCPT ); Sun, 22 Oct 2006 19:53:33 -0400 Message-ID: <453C0470.6090503@zytor.com> Date: Sun, 22 Oct 2006 16:53:20 -0700 From: "H. Peter Anvin" User-Agent: Thunderbird 1.5.0.7 (X11/20060913) MIME-Version: 1.0 To: Bernardo Innocenti CC: Cristian Grigoriu , b.innocenti@develer.com, lkml Subject: Re: NAT failure with TCP, too References: <4538B314.2020309@provus.ro> <453C01AE.7060103@develer.com> In-Reply-To: <453C01AE.7060103@develer.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Bernardo Innocenti wrote: > > > Cristian Grigoriu wrote: >> Hi, >> >> I can confirm the same bug you reported here >> http://www.ussg.iu.edu/hypermail/linux/kernel/0509.2/0279.html >> This time it happens with TCP connections originating from the same >> TCP port (1234) from multiple machines. The SNAT simply doesn't take >> place and the normal routing occurs. >> >> Kernel is Debian stock 2.6.18-1. >> >> Please let me know if you have find a workaround. > > It turned out that the real thing that was triggering the bug > for me was unloading and reloading the ip_nat module without > also reloading ip_conntrack. > > The connection tracking tuple would remain in the kernel, visible > in /proc/net/ip_conntrack, but no longer linked to the SNAT rule. > I'd consider this a bug, but very few users will ever be affected. > > The workaround for me was to remove my hand-cracted iptables > rules from ppp's ip-up.local and move them to the distro-supplied > iptables firewall instead. The only downside is that I must now > hardcode the destination ip of the SNAT rule because it's too early > to read the interface address of ppp0. > You may want to use the MASQUERADE target instead of the SNAT target. -hpa