From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753301AbaIYOf4 (ORCPT ); Thu, 25 Sep 2014 10:35:56 -0400 Received: from www.linutronix.de ([62.245.132.108]:33915 "EHLO Galois.linutronix.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752914AbaIYOfz (ORCPT ); Thu, 25 Sep 2014 10:35:55 -0400 Date: Thu, 25 Sep 2014 16:35:49 +0200 (CEST) From: Thomas Gleixner To: Dave Jones cc: linux-mm@kvack.org, Frederic Weisbecker , Peter Zijlstra , LKML Subject: Re: hrtimer deadlock caused by nohz_full In-Reply-To: <20140925141425.GA21702@redhat.com> Message-ID: References: <20140925141425.GA21702@redhat.com> User-Agent: Alpine 2.10 (DEB 1266 2009-07-14) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Linutronix-Spam-Score: -1.0 X-Linutronix-Spam-Level: - X-Linutronix-Spam-Status: No , -1.0 points, 5.0 required, ALL_TRUSTED=-1,SHORTCIRCUIT=-0.0001 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, 25 Sep 2014, Dave Jones wrote: > Got this on a box that had been fuzzing for 12 hours or so. > There's also some timer stuff going on htere, so cc'ing the usual suspects. And it's a hrtimer lockup > [] ? hrtimer_try_to_cancel+0x58/0x1f0 > [] ? lock_release+0x1d/0x300 > [] ? lock_release+0x1d/0x300 > [] ? hrtimer_try_to_cancel+0x58/0x1f0 > [] ? lock_release+0x1d/0x300 > <> [] _raw_spin_unlock_irqrestore+0x24/0x70 > [] hrtimer_try_to_cancel+0x58/0x1f0 > [] hrtimer_cancel+0x1a/0x30 > [] tick_nohz_restart+0x17/0x90 > [] __tick_nohz_full_check+0xc8/0xe0 > [] nohz_full_kick_work_func+0xe/0x10 > [] irq_work_run_list+0x4f/0x70 > [] irq_work_run+0x2a/0x60 > [] update_process_times+0x5b/0x70 > [] tick_sched_handle.isra.21+0x25/0x60 > [] tick_sched_timer+0x41/0x60 > [] __run_hrtimer+0x81/0x480 > [] ? tick_sched_do_timer+0x90/0x90 > [] hrtimer_interrupt+0x107/0x260 > [] local_apic_timer_interrupt+0x34/0x60 > [] smp_apic_timer_interrupt+0x3f/0x60 > [] apic_timer_interrupt+0x6f/0x80 hrtimer_interrupt tick_sched_timer tick_sched_handle update_process_times irq_work_run irq_work_run_list nohz_full_kick_work_func __tick_nohz_full_check tick_nohz_restart hrtimer_cancel And that hrtimer_cancel is: static void tick_nohz_restart(struct tick_sched *ts, ktime_t now) { hrtimer_cancel(&ts->sched_timer); Now, that's really bad because we are in the timer callback of ts->sched_timer. So hrtimer_cancel will loop forever waiting for the callback to complete. Frederic !?!? Thanks, tglx