From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail.candelatech.com ([208.74.158.172]:39350 "EHLO ns3.lanforge.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754241Ab1FIVZ0 (ORCPT ); Thu, 9 Jun 2011 17:25:26 -0400 Message-ID: <4DF13A43.6020500@candelatech.com> Date: Thu, 09 Jun 2011 14:25:23 -0700 From: Ben Greear To: Steve Dickson CC: linux-nfs@vger.kernel.org Subject: Re: [PATCH v2] nfs-utils: Support binding to source address. References: <1307574083-27608-1-git-send-email-greearb@candelatech.com> <4DF10053.9050901@RedHat.com> <4DF102E0.5000201@candelatech.com> <4DF13842.1020307@RedHat.com> In-Reply-To: <4DF13842.1020307@RedHat.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Sender: linux-nfs-owner@vger.kernel.org List-ID: MIME-Version: 1.0 On 06/09/2011 02:16 PM, Steve Dickson wrote: > > > On 06/09/2011 01:29 PM, Ben Greear wrote: >> On 06/09/2011 10:18 AM, Steve Dickson wrote: >>> Hello, >>> >>> >>> On 06/08/2011 07:01 PM, greearb@candelatech.com wrote: >>>> From: Ben Greear >>>> >>>> This lets one specify the source IP address for >>>> sockets, allowing users to leverage routing rules >>>> on multi-homed systems. >>> Right out of the box I do think some support like >>> this would be good... but I not convinced on how needed >>> something like this is which means there is a lot of >>> change here for basically not... But lets continue >>> to talk about to see where we end up.... >>> >>> One thing you must do is break up this patch in to >>> a more readable patch series. There is no way I going >>> to commit a patch that is close to 1500 lines >>> that so invasive. >> >> The vast majority of the patch is adding an extra >> argument to various methods to get things to compile. > Yeah I know... I saw that... But there has to be > some logical way to break this up.. > >> >> Do you want me to break those changes into individual >> chunks, or should I just separate the parsing& documentation >> chunks out? > I guess that would be start.... but just breaking this up > unto just a couple patches probably will not do anything for > readability either... > >> >> I can dummy stuff out with passing NULL around to make >> lots of smaller patches, but I'm not sure it helps >> any readability. > I don't know either... but it make me very nervous > to apply a 1400+ line patch especially when its basically > rewriting mount... I've broken mount before... Its not > a pretty sight... ;-) How about an initial patch that just passes a container down to the bind logic with the 'sap' and 'salen' into a container struct. Should have zero functionality change, but fairly large patch (basically the same places as where I added the 'local_ip' argument). Then, I can post a patch to enable packaging a src-addr in that container as well. That should be a much smaller patch and so may seem less scary. Thanks, Ben -- Ben Greear Candela Technologies Inc http://www.candelatech.com