From mboxrd@z Thu Jan 1 00:00:00 1970 From: Nikolay Aleksandrov Subject: Re: [PATCH 1/5] bonding: mc addresses don't get deleted on enslave failure Date: Thu, 18 Apr 2013 19:15:07 +0200 Message-ID: <51702A1B.5080402@redhat.com> References: <1366295697-31037-1-git-send-email-nikolay@redhat.com> <1366295697-31037-2-git-send-email-nikolay@redhat.com> <14509.1366304280@death.nxdomain> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org, andy@greyhouse.net, davem@davemloft.net To: Jay Vosburgh Return-path: Received: from mx1.redhat.com ([209.132.183.28]:62547 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S967395Ab3DRRQk (ORCPT ); Thu, 18 Apr 2013 13:16:40 -0400 In-Reply-To: <14509.1366304280@death.nxdomain> Sender: netdev-owner@vger.kernel.org List-ID: On 18/04/13 18:58, Jay Vosburgh wrote: > Nikolay Aleksandrov wrote: > >> Add dev_mc_del after err_detach as that's the first error path >> after they're added. The main issue is the mc addresses' refcount >> which only gets bumped up. >> >> Signed-off-by: Nikolay Aleksandrov > > All 5 of these patches look good to me. The only minor nits are > that the above description says "dev_mc_del," but the actual function > call added is bond_mc_list_flush (which in turn does call dev_mc_dev), > and that this patch (#1) modifies the lacpdu_multicast variable > initialization, which isn't really necessary for the bug fix. > > -J > > Yes, that all is because initially I wrote it directly with dev_mc_del where I needed the variable and then I changed it to use bond_mc_list_flush. I'll re-post v2 with updated log message and without that initialization, sorry about this. Nik