Return-path: Received: from mail-yw0-f46.google.com ([209.85.213.46]:34602 "EHLO mail-yw0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750963Ab2BJR3H convert rfc822-to-8bit (ORCPT ); Fri, 10 Feb 2012 12:29:07 -0500 Received: by yhoo21 with SMTP id o21so1600380yho.19 for ; Fri, 10 Feb 2012 09:29:05 -0800 (PST) MIME-Version: 1.0 In-Reply-To: <1328563980.3466.10.camel@jlt3.sipsolutions.net> References: <1328455870.7290.43.camel@wwguy-huron> <1DC40B07CD6EC041A66726C271A73AE611F67E@IRSMSX102.ger.corp.intel.com> <1328563980.3466.10.camel@jlt3.sipsolutions.net> From: Andrew Lutomirski Date: Fri, 10 Feb 2012 09:28:45 -0800 Message-ID: (sfid-20120210_182911_636308_C41840AE) Subject: Re: [Ilw] Re: 6300agn: queue stuck and driver doesn't recover To: Johannes Berg Cc: Alexander Schnaidt , "Guy, Wey-Yi W" , "ilw@linux.intel.com" , "linux-wireless@vger.kernel.org" Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Mon, Feb 6, 2012 at 1:33 PM, Johannes Berg wrote: > On Mon, 2012-02-06 at 22:27 +0100, Alexander Schnaidt wrote: > >> > I think you probably had something like "options iwlagn >> 11n_disable=1" or so in /etc/modprobe.d/ somewhere. This caused the >> problem to happen only with the driver called "iwlwifi", since >> "iwlagn" would get 11n disabled. > >> You're almost spot on!, somewhere along the bisect I put swcrypto=1 into >> modprobe.conf... so much for scientific reproducibility. Sorry for that. >> >> I redid the whole bisect with no options in modprobe.conf which defaults >> to swcrypto=0 on >=3.0 if I'm not mistaken. > > ... > >> If my interpretation is right the culprit would be: >> >> commit 5a3d9882b84edf5fa8e8ca33a5d6df25e2e727a5 >> Author: Johannes Berg >> Date: ? Fri Jul 15 13:03:12 2011 -0700 >> >> ? ? iwlagn: rewrite HW crypto > >> To me it makes sense considering that swcrypto=1 fixes >> my problems all the way up till kernel 3.3.0-rc2. >> >> ps.: 11n_disable=1 as suggested by Wey-Yi does not >> change the faulty behavior. >> >> Again I apologize for my blunder and thanks for the effort. > > Ok this makes sense. A few other people also recently reported this. I'm > currently (still) away from home so I can't reproduce this issue, and I > didn't find anything by just looking at the code. Traces for mac80211 > and iwlwifi subsystems would be useful to look at in this case, I think, > if you want to record the failure case see: > http://wireless.kernel.org/en/developers/Documentation/mac80211/tracing > (-e mac80211 -e iwlwifi). If you do record it, please don't send the > traces to the list, bzip2 it and send it to me directly. I just reproduced the problem (or at least the similar problem affecting me) with swcrypto=1. I'll try to get a trace. --Andy