From mboxrd@z Thu Jan 1 00:00:00 1970 From: Patrick McHardy Subject: Re: nf_conntrack_proto_generic, assigned to 0 Date: Mon, 26 Jan 2009 14:19:33 +0100 Message-ID: <497DB865.8050003@trash.net> References: <200901241622.36634.christoph.paasch@student.uclouvain.be> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-15; format=flowed Content-Transfer-Encoding: 7bit Cc: netfilter-devel@vger.kernel.org To: Christoph Paasch Return-path: Received: from stinky.trash.net ([213.144.137.162]:43623 "EHLO stinky.trash.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750738AbZAZNTf (ORCPT ); Mon, 26 Jan 2009 08:19:35 -0500 In-Reply-To: <200901241622.36634.christoph.paasch@student.uclouvain.be> Sender: netfilter-devel-owner@vger.kernel.org List-ID: Christoph Paasch wrote: > Hi, > > while reading through the source-code I remarked, that for the generic l4 > protocol handler (nf_conntrack_proto_generic.c) the protocol number 0 is > used. While linux/in.h defines IPPROTO_IP = 0 (and says "a dummy protocol for > tcp", whatever that may mean), in net/ipv6.h, the protocol 0 is used for the > hop-by-hop header. And as IANA states, 0 is assigned to the IPv6 Hop-by-hop > option. > > Even, if netfilter doesn't tracks the hop-by-hop protocol, it shouldn't use > this number for the generic protocol handler, because this might be confusing. > In my opinion, 255 should be used, which is assigned as "Reserved" by IANA. > > What do you think? It doesn't make a difference currently since in both cases we'd use generic for HBH. But I think you're suggestion makes sense, could you send a patch for this?