From: Lucas Meneghel Rodrigues Subject: Re: [Autotest] [PATCH] KVM test: Add support for ipv6 addresses Date: Thu, 25 Feb 2010 17:09:55 -0300 Message-ID: <1267128595.15621.33.camel@localhost.localdomain> References: <1266984656-4036-1-git-send-email-lmr@redhat.com> <6ac58f4f1002232016n78d86867ke6b2f10cc7deda47@mail.gmail.com> <1267105301.4742.27.camel@yogi-laptop> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: autotest@test.kernel.org, kvm@vger.kernel.org, Yogananth Subramanian To: yogi Return-path: Received: from mx1.redhat.com ([209.132.183.28]:29676 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933525Ab0BYUKO (ORCPT ); Thu, 25 Feb 2010 15:10:14 -0500 In-Reply-To: <1267105301.4742.27.camel@yogi-laptop> Sender: kvm-owner@vger.kernel.org List-ID: On Thu, 2010-02-25 at 19:11 +0530, yogi wrote: > On Wed, 2010-02-24 at 01:16 -0300, Lucas Meneghel Rodrigues wrote: > > H i Yogi, I have read and rebased your patch against current trunk. I > > have one doubt: I noticed that the tcpdump command you use to monitor > > traffic is basically all traffic that goes to network routers > > > > /usr/sbin/tcpdump -npv ip6 -i any 'dst net ff02::2' > > > > However, the IPv4 implementation filters the traffic to the dhcp/bootp > > port, 68. Therefore the ipv6 mapper captures more (non necessarily > > useful) mappings than the ipv4 version. I made some brief research and > > couldn't figure out how to do it. Do you think it's possible to refine > > the tcpdump command to make it look only on the dhp request port (68)? > > > > Hello Lucas, thanks for reviewing my patches for ipv6. > Following tcpdump command will capture only the icmp6 packets and not > all the ipv6 packets > tcpdump -npv -i any icmp6 and 'dst net ff02::2' > > Since the patch uses ip6 link local address, we are looking for router > solicitation messages. Ok, so you think it'd be better to change the tcpdump command used in the patch to be "tcpdump -npv -i any icmp6 and 'dst net ff02::2'"? That said, I don't fully understand all gory ipv6 details, so please bear with me on this question: how would we be benefit from having an ipv6 address cache? I mean, it's not like we're going to use the mappings for any of the communications... This patch will stay in the queue so Michael can take a look at it and give opinions. Cheers, Lucas