From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?utf-8?B?U8O2cmVu?= Brinkmann Subject: Re: [PATCH RESEND] PM / sleep: Fix racing timers Date: Tue, 28 Oct 2014 13:40:11 -0700 Message-ID: <83c78fd2985d4530a48a5947caef21be@BL2FFO11FD052.protection.gbl> References: <1411405623-7869-1-git-send-email-soren.brinkmann@xilinx.com> <1860362.OyoLA3GxVJ@vostro.rjw.lan> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Received: from mail-bl2on0076.outbound.protection.outlook.com ([65.55.169.76]:29654 "EHLO na01-bl2-obe.outbound.protection.outlook.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1751190AbaJ1UkQ (ORCPT ); Tue, 28 Oct 2014 16:40:16 -0400 Content-Disposition: inline In-Reply-To: Sender: linux-pm-owner@vger.kernel.org List-Id: linux-pm@vger.kernel.org To: "Rafael J. Wysocki" Cc: Thomas Gleixner , John Stultz , Pavel Machek , Len Brown , linux-kernel@vger.kernel.org, linux-pm@vger.kernel.org Hi Rafael, any opinion on this? Thanks, S=C3=B6ren On Thu, 2014-10-02 at 09:01AM -0700, S=C3=B6ren Brinkmann wrote: > Hi Rafael, >=20 > On Tue, 2014-09-23 at 01:01AM +0200, Rafael J. Wysocki wrote: > > On Monday, September 22, 2014 10:07:03 AM Soren Brinkmann wrote: > > > On platforms that do not power off during suspend, successfully e= ntering > > > suspend races with timers. > > >=20 > > > The race happening in a couple of location is: > > >=20 > > > 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) > > >=20 > > > Between steps 1 and 2 the timers can still generate interrupts th= at are > > > not handled and stay pending until step 3. That pending IRQ cause= s an > > > immediate - spurious - wake. > > >=20 > > > The solution is to move the clockevents suspend/resume notificati= on > > > out of the syscore_suspend step and explictly call them at the ap= propriate > > > time in the suspend/hibernation paths. I.e. timers are suspend _b= efore_ > > > IRQs get disabled. And accordingly in the resume path. > > >=20 > > > Signed-off-by: Soren Brinkmann > > > --- > > > Hi, > > >=20 > > > there was not a lot of discussion on the last submission. Just on= e comment from > > > Rafael (https://lkml.org/lkml/2014/8/26/780), which - as I outlin= ed in my > > > response, does not apply, IMHO, since the platform does not re-en= able > > > interrupts. > >=20 > > Well, you just don't agree with it. > >=20 > > The problem with your approach is that timer interrupts aren't actu= ally as > > special as you think and any other IRQF_NO_SUSPEND interrupts would= have caused > > similar issues to appear under specific conditions. > >=20 > > The solution I would suggest and that actually covers all IRQF_NO_S= USPEND > > interrupts would be to use a wait_event() loop like the one in free= ze_enter() > > (on top of the current linux-next or the pm-genirq branch of linux-= pm.git), > > but wait for pm_abort_suspend to become true, to implement system s= uspend. >=20 > sorry, it took me a while since I needed to get some dependencies por= ted > to the pm-genirq base. Once I had that, it reproduced my original iss= ue. > So far so good. I then looked into finding a solution following your > guidance. I'm not sure I really found what you had in mind, but below= is > what I came up with, which seems to do it. > Please let me know how far off I am. >=20 > Thanks, > S=C3=B6ren >=20 > -------8<------------------8<----------------8<----------------8<----= ----------- > diff --git a/drivers/base/power/wakeup.c b/drivers/base/power/wakeup.= c > index c2744b30d5d9..a4f9914571f1 100644 > --- a/drivers/base/power/wakeup.c > +++ b/drivers/base/power/wakeup.c > @@ -25,7 +25,7 @@ > bool events_check_enabled __read_mostly; > =20 > /* If set and the system is suspending, terminate the suspend. */ > -static bool pm_abort_suspend __read_mostly; > +bool pm_abort_suspend __read_mostly; > =20 > /* > * Combined counters of registered wakeup events and wakeup events i= n progress. > diff --git a/kernel/power/suspend.c b/kernel/power/suspend.c > index 6dadb25cb0d8..e6a6de8f76d0 100644 > --- a/kernel/power/suspend.c > +++ b/kernel/power/suspend.c > @@ -33,6 +33,7 @@ > =20 > static const char *pm_labels[] =3D { "mem", "standby", "freeze", }; > const char *pm_states[PM_SUSPEND_MAX]; > +extern bool pm_abort_suspend; > =20 > static const struct platform_suspend_ops *suspend_ops; > static const struct platform_freeze_ops *freeze_ops; > @@ -294,25 +295,27 @@ static int suspend_enter(suspend_state_t state,= bool *wakeup) > if (error || suspend_test(TEST_CPUS)) > goto Enable_cpus; > =20 > - arch_suspend_disable_irqs(); > - BUG_ON(!irqs_disabled()); > - > - error =3D syscore_suspend(); > - if (!error) { > - *wakeup =3D pm_wakeup_pending(); > - if (!(suspend_test(TEST_CORE) || *wakeup)) { > - trace_suspend_resume(TPS("machine_suspend"), > - state, true); > - error =3D suspend_ops->enter(state); > - trace_suspend_resume(TPS("machine_suspend"), > - state, false); > - events_check_enabled =3D false; > + while (!pm_abort_suspend) { > + arch_suspend_disable_irqs(); > + BUG_ON(!irqs_disabled()); > + > + error =3D syscore_suspend(); > + if (!error) { > + *wakeup =3D pm_wakeup_pending(); > + if (!(suspend_test(TEST_CORE) || *wakeup)) { > + trace_suspend_resume(TPS("machine_suspend"), > + state, true); > + error =3D suspend_ops->enter(state); > + trace_suspend_resume(TPS("machine_suspend"), > + state, false); > + events_check_enabled =3D false; > + } > + syscore_resume(); > } > - syscore_resume(); > - } > =20 > - arch_suspend_enable_irqs(); > - BUG_ON(irqs_disabled()); > + arch_suspend_enable_irqs(); > + BUG_ON(irqs_disabled()); > + } > =20 > Enable_cpus: > enable_nonboot_cpus(); > -- > To unsubscribe from this list: send the line "unsubscribe linux-kerne= l" 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/