From mboxrd@z Thu Jan 1 00:00:00 1970 From: Auke Kok Subject: Re: [BUG] e100: eth0 appers many times in /proc/interrupts after resume Date: Mon, 22 Jan 2007 16:15:29 -0800 Message-ID: <45B553A1.5010303@intel.com> References: <1167599458.2662.8.camel@nigel.suspend2.net> <1167605481.12328.0.camel@localhost> <1167607994.2662.39.camel@nigel.suspend2.net> <1167644970.7142.6.camel@localhost> <1168317278.6948.9.camel@nigel.suspend2.net> <1168448689.7430.1.camel@localhost> <1168463852.3205.1.camel@nigel.suspend2.net> <1169407062.1932.4.camel@localhost> <20070121212209.GB8958@slug> <45B3DEF7.8020505@intel.com> <20070121220457.GC8958@slug> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Andrei Popa , linux-kernel@vger.kernel.org, nigel@suspend2.net, NetDev Return-path: Received: from mga02.intel.com ([134.134.136.20]:35415 "EHLO mga02.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932509AbXAWAP5 (ORCPT ); Mon, 22 Jan 2007 19:15:57 -0500 To: Frederik Deweerdt In-Reply-To: <20070121220457.GC8958@slug> Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org Frederik Deweerdt wrote: > On Sun, Jan 21, 2007 at 01:45:27PM -0800, Auke Kok wrote: >> Frederik Deweerdt wrote: >>> On Sun, Jan 21, 2007 at 09:17:41PM +0200, Andrei Popa wrote: >>>> It's the 10th resume and in /proc/interrupts eth0 appers 10 times. >>> The e100_resume() function should be calling netif_device_detach and >>> free_irq. Could you try the following (compile tested) patch? >> I just fixed suspend/shutdown for e100 in 2.6.19, not sure why the problem still shows up. Since it's a driver/net issue, you >> should CC netdev on it tho, otherwise it might go unnoticed. > Thanks for adding the CC >> I'll open up the can-o-worms on this issue and see what's up with it. >> >> I'm not so sure that this patch is OK, and I wonder why it stopped working, because I spent quite some time fixing it only a >> few months ago. Did swsup change again? sigh... > > I may well be wrong (It appears that most of the time I am :)), but the > unbalanced netif_device_attach (in resume) looks suspicious. resume() > also calls request_irq, so calling free_irq on suspend seemed logical. I just tested the patch and looked it over again. The patch works good and indeed fixes the problem, and netconsole works great. It even shows much of the suspend/resume over the wire, something which I can't remember seeing before with netconsole. reboot -f also works OK. I'll push the patch upstream, thanks! Auke