From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dawid Ciezarkiewicz Subject: Re: [RFC] wrr (weighted round-robin) bonding Date: Fri, 20 Oct 2006 22:52:00 +0200 Message-ID: <200610202252.01390.dpc@asn.pl> References: <200610162021.12884.dpc@asn.pl> <200610202141.15140.dpc@asn.pl> <200610201953.k9KJr2Zt000305@death.nxdomain.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit Cc: Andy Gospodarek , netdev@vger.kernel.org Return-path: Received: from apollo.asn.pl ([85.14.104.1]:4997 "HELO apollo.asn.pl") by vger.kernel.org with SMTP id S1030259AbWJTUwU (ORCPT ); Fri, 20 Oct 2006 16:52:20 -0400 To: Jay Vosburgh In-Reply-To: <200610201953.k9KJr2Zt000305@death.nxdomain.ibm.com> Content-Disposition: inline Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Friday, 20 October 2006 21:53, Jay Vosburgh wrote: > Dawid Ciezarkiewicz wrote: > > >On Thursday, 19 October 2006 21:04, Andy Gospodarek wrote: > >> It would seem to me that extending an existing mode would be more > >> desirable than adding yet another mode to worry about. I don't even > >> like the fact that there are as many as there are, but I understand why > >> they are there. > > > >Ack. I will probably update wrr bonding patch to replace rr mode. > > Also, if acceptance into the mainline is your ultimate goal, the > ioctl control interface will be a very difficult sell. You'd want to > look into some other control mechanism, most likely an additional sysfs > entry. Oh. I'm quite puzzled here. What is current policy? I'd like sysfs interfaces better than ioctl - they are much cleaner etc. - but I thought ioctl will be better here because current bonding control uses ioctl and extending it is much simpler. More generally: should I use sysfs always when adding anything to kernel and use it even when extending older functionality? If wrr bonding have chances for inclusion into mainline I'll do any necessary changes to let it meet requirements. I'd like to know if functionality from Andy's patch is desired to be a part of such round-robin bonding extension. I think I could easily integrate it with wrr.