From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Ahern Subject: Re: [net-next] net: l3mdev: address selection should only consider devices in L3 domain Date: Tue, 16 Feb 2016 16:49:03 -0700 Message-ID: <56C3B56F.4020106@cumulusnetworks.com> References: <1455663591-27615-1-git-send-email-dsa@cumulusnetworks.com> <20160216234115.GI659150@eidolon> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org, david@opensourcerouting.org To: David Lamparter Return-path: Received: from mail-io0-f169.google.com ([209.85.223.169]:34090 "EHLO mail-io0-f169.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933313AbcBPXtG (ORCPT ); Tue, 16 Feb 2016 18:49:06 -0500 Received: by mail-io0-f169.google.com with SMTP id 9so7204124iom.1 for ; Tue, 16 Feb 2016 15:49:06 -0800 (PST) In-Reply-To: <20160216234115.GI659150@eidolon> Sender: netdev-owner@vger.kernel.org List-ID: On 2/16/16 4:41 PM, David Lamparter wrote: > ... this won't do enough. If you look at the 4.3 patch I sent you, I > was adding a comment: > > + /* For VRFs, the VRF device takes the place of the loopback device, > + with addresses on it being preferred. Note in such cases the > + loopback device will be among the devices that fail the vrf_ifi > + equality check in the loop below. > + */ > Sure, I was focused on fixing the 'ignore devices not in the VRF' but certainly for unnumbered use cases you want the VRF loopback address to be preferred. Missed that earlier. Rather than spinning a patch on top of mine, fold this change into your and make it 1 patch to achieve the goal.