Return-path: Received: from mga02.intel.com ([134.134.136.20]:42692 "EHLO mga02.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751088Ab2ATGvs convert rfc822-to-8bit (ORCPT ); Fri, 20 Jan 2012 01:51:48 -0500 From: "Grumbach, Emmanuel" To: Dave Jones CC: Emmanuel Grumbach , Dan Williams , "netdev@vger.kernel.org" , "Guy, Wey-Yi W" , Linux Wireless Subject: RE: iwlwifi worse than ever in current git. Date: Fri, 20 Jan 2012 06:50:36 +0000 Message-ID: <0BA3FCBA62E2DC44AF3030971E174FB3079B13@HASMSX103.ger.corp.intel.com> (sfid-20120120_075154_385801_4B06E32C) References: <20120110184917.GA26973@redhat.com> <1326221614.5067.6.camel@dcbw.foobar.com> <0BA3FCBA62E2DC44AF3030971E174FB3063085@HASMSX103.ger.corp.intel.com> <20120111014243.GA22066@redhat.com> <0BA3FCBA62E2DC44AF3030971E174FB306429E@HASMSX103.ger.corp.intel.com> <20120111195055.GA5140@redhat.com> <0BA3FCBA62E2DC44AF3030971E174FB3064AF5@HASMSX103.ger.corp.intel.com> <20120119182955.GA1630@redhat.com> <0BA3FCBA62E2DC44AF3030971E174FB3079A3F@HASMSX103.ger.corp.intel.com> <20120119220410.GA11980@redhat.com> In-Reply-To: <20120119220410.GA11980@redhat.com> Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Sender: linux-wireless-owner@vger.kernel.org List-ID: > > On Thu, Jan 19, 2012 at 08:38:35PM +0000, Grumbach, Emmanuel wrote: > > > > > > > > > > > On Thu, Jan 19, 2012 at 08:47:28AM +0200, Emmanuel Grumbach wrote: > > > > > Ok, I finally had a change to take a closer look at this. Please run > > > > > with debug=0xC0000000. > > > > > > > > > > This will be noisy, but I guess we have no choice... > > > > > I will be travelling next week, so... > > > > > > > > Ah. Got it. I had to do a large download (a kernel tarball was enough) > to > > > > trigger it. > > > > Log is at http://davej.fedorapeople.org/iwl-debug.txt > > > > > > > > The download began at around 13:17 in the log. > > > > It hung shortly afterwards. > > > > > > > > Dave > > > > > > Thanks for that one. Really helpful. I need to send you a patch to get a > > > little more data. I don't know when I will be able to do it though. > > > > Can you please retry with the patch attached ? > > Hung again. > > http://davej.fedorapeople.org/iwl-debug2.txt > > Dave Thanks. From what I see, a management frame is sent from an AC queue and when we get the Tx response for this frame, we get sta_id = AP, tid = 0 (and the bug is probably here). So we modify the counters for sta_id = AP, and tid = 0, and this messes up everything. I guess I need to see why I don't see this issue, and see how to give a hint to the Tx response path that counters shouldn't be modified. --------------------------------------------------------------------- Intel Israel (74) Limited This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies.