From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753401Ab3LNKjd (ORCPT ); Sat, 14 Dec 2013 05:39:33 -0500 Received: from order.stressinduktion.org ([87.106.68.36]:55030 "EHLO order.stressinduktion.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752169Ab3LNKjb (ORCPT ); Sat, 14 Dec 2013 05:39:31 -0500 Date: Sat, 14 Dec 2013 11:39:29 +0100 From: Hannes Frederic Sowa To: David Miller Cc: dvlasenk@redhat.com, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, kuznet@ms2.inr.ac.ru, jmorris@namei.org, yoshfuji@linux-ipv6.org, kaber@trash.net, jpirko@redhat.com Subject: Re: [patch net-next v3] ipv6: log autoconfiguration failures Message-ID: <20131214103929.GE23563@order.stressinduktion.org> Mail-Followup-To: David Miller , dvlasenk@redhat.com, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, kuznet@ms2.inr.ac.ru, jmorris@namei.org, yoshfuji@linux-ipv6.org, kaber@trash.net, jpirko@redhat.com References: <1386949547-23457-1-git-send-email-dvlasenk@redhat.com> <20131213.175709.414078764436451428.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <20131213.175709.414078764436451428.davem@davemloft.net> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Dec 13, 2013 at 05:57:09PM -0500, David Miller wrote: > From: Denys Vlasenko > Date: Fri, 13 Dec 2013 16:45:47 +0100 > > > If ipv6 auto-configuration does not work, currently it's hard > > to track what's going on. This change adds log messages > > (at debug level) on every code path where ipv6 autoconf fails. > > > > v3: changed pr_debug's to pr_warn's. > > > > Signed-off-by: Denys Vlasenko > > I already made it clear I want to see statistics added instead of log > messages for this. Hmm... I gave this a bit more thought. I would go with counters, too. What if we add something like per interface softerror states. So that we e.g. can see that one protocol had an error on an interface and it needs attention of an administrator? E.g. 2: eth0: Would catch the eyes more quickly. If counters are implemented we could do that completly in user space and add that flag as soon as one of the observed counters are above 0 (and of course, iproute needs colors). What do you think?