From mboxrd@z Thu Jan 1 00:00:00 1970 From: Patrick McHardy Subject: Re: [PATCH 20/31] netns ct: NOTRACK in netns Date: Mon, 21 Jul 2008 19:08:43 +0200 Message-ID: <4884C29B.7060107@trash.net> References: <20080721165028.GF13390@martell.zuzino.mipt.ru> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-15; format=flowed Content-Transfer-Encoding: 7bit Cc: Alexey Dobriyan , netdev@vger.kernel.org, netfilter-devel@vger.kernel.org, dev@openvz.org, xemul@openvz.org, ebiederm@xmission.com, dlezcano@fr.ibm.com, benjamin.thery@bull.net To: Jan Engelhardt Return-path: Received: from stinky.trash.net ([213.144.137.162]:42802 "EHLO stinky.trash.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751842AbYGURIp (ORCPT ); Mon, 21 Jul 2008 13:08:45 -0400 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: Jan Engelhardt wrote: > On Monday 2008-07-21 18:50, Alexey Dobriyan wrote: > >> +++ b/net/netfilter/xt_state.c >> @@ -30,7 +31,7 @@ state_mt(const struct sk_buff *skb, const struct net_device *in, >> enum ip_conntrack_info ctinfo; >> unsigned int statebit; >> >> - if (nf_ct_is_untracked(skb)) >> + if (nf_ct_is_untracked(dev_net(in ? in : out), skb)) > > Does this even do the right thing? Given the forward path, where > both in!=NULL and out!=NULL, why should it specifically be 'in' > that is selected? It doesn't matter, the namespace during forwaring is always equal for the input and output devices.