From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Korsgaard Subject: Re: [PATCH] libxt_TCPMSS.c: unbreak build without ipv6 after ea2a02f7 Date: Fri, 22 Jul 2011 12:49:35 +0200 Message-ID: <87zkk6y4hs.fsf@macbook.be.48ers.dk> References: <1311312260-16102-1-git-send-email-jacmet@sunsite.dk> <874o2ezk67.fsf@macbook.be.48ers.dk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: netfilter-devel@vger.kernel.org To: Jan Engelhardt Return-path: Received: from mail-wy0-f174.google.com ([74.125.82.174]:39934 "EHLO mail-wy0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753848Ab1GVKtw (ORCPT ); Fri, 22 Jul 2011 06:49:52 -0400 Received: by wyg8 with SMTP id 8so1409064wyg.19 for ; Fri, 22 Jul 2011 03:49:51 -0700 (PDT) In-Reply-To: (Jan Engelhardt's message of "Fri, 22 Jul 2011 12:31:04 +0200 (CEST)") Sender: netfilter-devel-owner@vger.kernel.org List-ID: >>>>> "Jan" == Jan Engelhardt writes: Hi, >> But those structs define the network format of ipv4 and ipv6 headers, so >> they are not likely to change. Jan> I was pondering on a hypothetical implementation with CHAR_BIT=16. Using iptables? Even so, the ipv6 header would need to be the same to be able to communicate with other ipv6 hosts. Jan> How about Jan> http://dev.medozas.de/gitweb.cgi?p=iptables;a=commitdiff;h=4d8656ad9d0afd04820f125a85a7b673c7e74fe6 That works as well (if we assume that all systems have netinet/ip.h even when built with --disable-ipv4): checking size of struct ip6_hdr... 0 So you end up with a wrong TCPMSS6_opts.max value (UINT16_MAX), but that presumably doesn't matter if you don't use ipv6. -- Bye, Peter Korsgaard