From mboxrd@z Thu Jan 1 00:00:00 1970 From: ebiederm@xmission.com (Eric W. Biederman) Subject: Re: [PATCH] ipv6: Revert 'administrative down' address handling changes. Date: Tue, 25 Jan 2011 10:55:31 -0800 Message-ID: References: <20110123.234101.71098533.davem@davemloft.net> <20110124.233813.246546891.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: netdev@vger.kernel.org, shemminger@vyatta.com, brian.haley@hp.com, lorenzo@google.com, herbert@gondor.hengli.com.au To: David Miller Return-path: Received: from out01.mta.xmission.com ([166.70.13.231]:54020 "EHLO out01.mta.xmission.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752478Ab1AYSzm (ORCPT ); Tue, 25 Jan 2011 13:55:42 -0500 In-Reply-To: (Eric W. Biederman's message of "Tue, 25 Jan 2011 00:51:19 -0800") Sender: netdev-owner@vger.kernel.org List-ID: ebiederm@xmission.com (Eric W. Biederman) writes: > David Miller writes: > >> From: David Miller >> Date: Sun, 23 Jan 2011 23:41:01 -0800 (PST) >> >>> Eric B. and co., please do some testing to make sure all of your >>> disable_ipv6 cases are functioning properly with this applied. >> >> Ping? > > In progress. I had to make a small change to your patch to get it > to apply against 2.6.37. neigh_ifdown has not been removed from the > beginning of addrconf_ifdown there. The piece that was failing for > me is not failing now so, so far so good. > > It was reported that in 2.6.37 there was a new regression that > 1connecting to ::1 when ipv6 was disabled would not fail immediately but > would have to wait a while. With your patch applied I am not seeing > that behavior either. > > Tomorrow I should know if I see any weird side effects with your patch, > after my regression tests for everything else have finished running. I have to admit I am seeing weird side effects. A test that was mysteriously failing because it could not create a vlan on top of a tap device has started working again ;) So this change is looking really good for me. I have to track through some additional failures that I think are test bugs, but it looks like I might finally have 2.6.37 in a shape where I can use it. Eric