Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S262448AbUKWKqE (ORCPT ); Tue, 23 Nov 2004 05:46:04 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S262464AbUKWKqE (ORCPT ); Tue, 23 Nov 2004 05:46:04 -0500 Received: from mx1.elte.hu ([157.181.1.137]:1171 "EHLO mx1.elte.hu") by vger.kernel.org with ESMTP id S262448AbUKWKon (ORCPT ); Tue, 23 Nov 2004 05:44:43 -0500 Date: Tue, 23 Nov 2004 12:46:39 +0100 From: Ingo Molnar To: Mark_H_Johnson@raytheon.com Cc: Amit Shah , Karsten Wiese , Bill Huey , Adam Heath , emann@mrv.com, Gunther Persoons , "K.R. Foley" , linux-kernel@vger.kernel.org, Florian Schmidt , Fernando Pablo Lopez-Lezcano , Lee Revell , Rui Nuno Capela , Shane Shrybman , Esben Nielsen , Thomas Gleixner , Michal Schmidt Subject: Re: [patch] Real-Time Preemption, -RT-2.6.10-rc2-mm2-V0.7.30-2 Message-ID: <20041123114639.GA23730@elte.hu> References: <20041122221224.GA13799@elte.hu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20041122221224.GA13799@elte.hu> User-Agent: Mutt/1.4.1i X-ELTE-SpamVersion: MailScanner 4.31.6-itk1 (ELTE 1.2) SpamAssassin 2.63 ClamAV 0.73 X-ELTE-VirusStatus: clean X-ELTE-SpamCheck: no X-ELTE-SpamCheck-Details: score=-2.201, required 5.9, BAYES_00 -4.90, SORTED_RECIPS 2.70 X-ELTE-SpamLevel: X-ELTE-SpamScore: -2 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1347 Lines: 34 * Ingo Molnar wrote: > * Mark_H_Johnson@raytheon.com wrote: > > > - echo 0 > /proc/sys/kernel/preempt_wakeup_timing [entered, but > > display was frozen at this point and did not see newline nor any > > further output] > > managed to reproduce this on an SMP box but not on an UP box, so i > think this is SMP related. It definitely happens almost immediately > after preempt_wakeup_timing is reset - or after preempt_max_timing is > reset. (Perhaps a dump_stack() from the wrong place, or something > like that.) The lockup was caused by the mutex wakeup being done under the PI lock, and if a new critical-section latency is reported within try_to_wake_up then the trylock done there deadlocked. The NMI watchdog triggered but the printks done there deadlocked as well. I fixed both the deadlock scenario, and made the NMI printout path more robust to get the messages out to the console in even such a case. The fixes are in the latest (-30-4) patch which can be found at the usual place: http://redhat.com/~mingo/realtime-preempt/ Ingo - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/