From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jiri Pirko Subject: Re: [PATCH v8 net-next 2/2] bonding: Simplify the xmit function for modes that use xmit_hash Date: Mon, 19 Oct 2015 20:07:18 +0200 Message-ID: <20151019180718.GA10106@nanopsycho.orion> References: <1412469901-27451-1-git-send-email-maheshb@google.com> <20151019163556.GK2288@nanopsycho.orion> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Jay Vosburgh , Andy Gospodarek , Veaceslav Falico , Nikolay Aleksandrov , David Miller , netdev , Eric Dumazet , Maciej Zenczykowski To: Mahesh Bandewar Return-path: Received: from mail-lf0-f52.google.com ([209.85.215.52]:32908 "EHLO mail-lf0-f52.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751705AbbJSSHV (ORCPT ); Mon, 19 Oct 2015 14:07:21 -0400 Received: by lffv3 with SMTP id v3so117045250lff.0 for ; Mon, 19 Oct 2015 11:07:19 -0700 (PDT) Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: Mon, Oct 19, 2015 at 08:01:24PM CEST, maheshb@google.com wrote: >On Mon, Oct 19, 2015 at 9:35 AM, Jiri Pirko wrote: >> Sun, Oct 05, 2014 at 02:45:01AM CEST, maheshb@google.com wrote: >>>Earlier change to use usable slave array for TLB mode had an additional >>>performance advantage. So extending the same logic to all other modes >>>that use xmit-hash for slave selection (viz 802.3AD, and XOR modes). >>>Also consolidating this with the earlier TLB change. >>> >>>The main idea is to build the usable slaves array in the control path >>>and use that array for slave selection during xmit operation. >>> >>>Measured performance in a setup with a bond of 4x1G NICs with 200 >>>instances of netperf for the modes involved (3ad, xor, tlb) >>>cmd: netperf -t TCP_RR -H -l 60 -s 5 >>> >>>Mode TPS-Before TPS-After >>> >>>802.3ad : 468,694 493,101 >>>TLB (lb=0): 392,583 392,965 >>>XOR : 475,696 484,517 >> >> Looking over bonding code now and came across this commit. It gave me a >> little headache. Why don't you guys just use team instead of doing >> bonding megapatches like this? This is making bonding code even more >> complicated (screwed-up) :/ I don't understand. > >Why are you looking at bonding to have headache in first place go use team. ;) > >The commit message was very clear about the intention and the purpose >is to avoid list-traversal cost in fast path for modes that use >xmit-hash. If you feel something is screwed-up please send a patch to >fix it. No, it's about the concept. You introduced something in bonding which team already has and made the code even more complicated and scary. I just don't understand why you didn't use team instead of this :/ Just my 50 cents.