From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail.candelatech.com ([208.74.158.172]:52840 "EHLO ns3.lanforge.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755032Ab0KKFvc (ORCPT ); Thu, 11 Nov 2010 00:51:32 -0500 Message-ID: <4CDB845E.5040906@candelatech.com> Date: Wed, 10 Nov 2010 21:51:26 -0800 From: Ben Greear MIME-Version: 1.0 To: Johannes Berg CC: "linux-wireless@vger.kernel.org" Subject: Re: ath5k/mac80211: Reproducible deadlock with 64-stations. References: <4CDB2488.4040802@candelatech.com> <4CDB3F7E.1000107@candelatech.com> <1289437415.3748.26.camel@jlt3.sipsolutions.net> In-Reply-To: <1289437415.3748.26.camel@jlt3.sipsolutions.net> Content-Type: text/plain; charset=UTF-8; format=flowed Sender: linux-wireless-owner@vger.kernel.org List-ID: On 11/10/2010 05:03 PM, Johannes Berg wrote: > On Wed, 2010-11-10 at 16:57 -0800, Ben Greear wrote: > >> I think the attached trace might be more useful. It shows >> the blocked tasks spammage that we saw in an earlier >> run. > > Yeah but we already knew that these tasks were blocking, and where ... > nothing really new here. I don't see it, yet anyway. I think at least one of the 'ip' processes has RTNL held, according to it's stack trace, and then it went and did something in mac80211. I haven't tried to track everything down in the code yet, however. The system starts being very sluggish, and then locks hard. It's possible that those particular processes might eventually recover or get a bit farther, perhaps the hard-lock is elsewhere. We have lockdep enabled, and it never spits anything out, for what that's worth. I'll get some more traces tomorrow to see if I can find any similarities. Thanks, Ben -- Ben Greear Candela Technologies Inc http://www.candelatech.com