From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jamal Hadi Salim <jhs@mojatatu.com>
Subject: Re: [PATCH 1/1] iproute2 bridge: bring to par with brctl show macs
Date: Tue, 27 May 2014 17:43:43 -0400
Message-ID: <5385070F.1030802@mojatatu.com>
References: <1401024151-8774-1-git-send-email-jhs@emojatatu.com> <27A51C19-553E-4BA0-A8B2-F4251FABA5F4@cumulusnetworks.com> <53850672.1030107@mojatatu.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: Stephen Hemminger <stephen@networkplumber.org>,
	netdev <netdev@vger.kernel.org>,
	Vlad Yasevich <vyasevic@redhat.com>,
	john.r.fastabend@intel.com,
	Roopa Prabhu <roopa@cumulusnetworks.com>
To: Scott Feldman <sfeldma@cumulusnetworks.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from mail-ig0-f181.google.com ([209.85.213.181]:57794 "EHLO
	mail-ig0-f181.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1752215AbaE0Vnq (ORCPT
	<rfc822;netdev@vger.kernel.org>); Tue, 27 May 2014 17:43:46 -0400
Received: by mail-ig0-f181.google.com with SMTP id h3so1687565igd.8
        for <netdev@vger.kernel.org>; Tue, 27 May 2014 14:43:45 -0700 (PDT)
In-Reply-To: <53850672.1030107@mojatatu.com>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

On 05/27/14 17:41, Jamal Hadi Salim wrote:

> filtering by brport is very useful.
> I wasnt as much concerned about that use case because i can already
> filter in user space for brport when i dump (iproute2 does as well).
> But i couldnt filter by bridge since that info is not even
> present in the dumped data
> I just saw the patch posted by Roopa - I think that now gives me the
> ability to also filter by bridge in user space.
>
> Note: All i am doing is stealing the ndm ifindex - which before was
> a zero and ignore by the kernel as the bridge ifindex.
> I think what you are suggesting is: if the user doesnt specify the
> bridge i should use the ndm ifindex to send down the brport as
> the filter. But that becomes an bridge "or" bridgeport filter
> choice. A more useful scenario is to have bridge "and"
> bridgeport filter. So almost seems like i need an attribute to send
> down for the bridge port.
> Sounds like a separate patch...
>

Hrm. I wonder - I could use the NDA_MASTER attribute that Roopa
posted to define the bridge filter. I could then use the ndm ifindex to
define the bridgeport filter.

Thoughts?

cheers,
jamal