From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751681AbaG1T0N (ORCPT ); Mon, 28 Jul 2014 15:26:13 -0400 Received: from mail-pa0-f49.google.com ([209.85.220.49]:41224 "EHLO mail-pa0-f49.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750960AbaG1T0K (ORCPT ); Mon, 28 Jul 2014 15:26:10 -0400 Message-ID: <53D6A3CF.7010204@linaro.org> Date: Mon, 28 Jul 2014 12:26:07 -0700 From: John Stultz User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.0 MIME-Version: 1.0 To: "Rafael J. Wysocki" , Soren Brinkmann CC: Thomas Gleixner , Pavel Machek , Len Brown , linux-kernel@vger.kernel.org, linux-pm@vger.kernel.org, Daniel Lezcano Subject: Re: [PATCH RFC v2] PM / sleep: Fix racing timers References: <53D083C5.2000501@linaro.org> <1406322408-15408-1-git-send-email-soren.brinkmann@xilinx.com> <1461880.klCWmqfb3F@vostro.rjw.lan> In-Reply-To: <1461880.klCWmqfb3F@vostro.rjw.lan> X-Enigmail-Version: 1.6 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 07/27/2014 04:18 PM, Rafael J. Wysocki wrote: > On Friday, July 25, 2014 02:06:48 PM Soren Brinkmann wrote: >> On platforms that do not power off during suspend, successfully entering >> suspend races with timers. >> >> The race happening in a couple of location is: >> >> 1. disable IRQs (e.g. arch_suspend_disable_irqs()) >> ... >> 2. syscore_suspend() >> -> timekeeping_suspend() >> -> clockevents_notify(SUSPEND) >> -> tick_suspend() (timers are turned off here) >> ... >> 3. wfi (wait for wake-IRQ here) >> >> Between steps 1 and 2 the timers can still generate interrupts that are >> not handled and stay pending until step 3. That pending IRQ causes an >> immediate - spurious - wake. >> >> The solution is to move the clockevents suspend/resume notification >> out of the syscore_suspend step and explictly call them at the appropriate >> time in the suspend/hibernation paths. I.e. timers are suspend _before_ >> IRQs get disabled. And accordingly in the resume path. >> >> Signed-off-by: Soren Brinkmann >> --- >> Hi, >> >> This is my second shot at this. I followed John's suggestion to keep the >> timekeeping suspend where it is and just move the shutdown of the clockevent >> devices around. > John, what do you think? I've not had the chance to take a closer look and do any testing. I suspect we'll need tgxl's input here as well. The change makes sense, but ordering modifications in this area tend to be fragile, as there are lots of implicit dependencies. thanks -john