Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752358AbbGQALQ (ORCPT ); Thu, 16 Jul 2015 20:11:16 -0400 Received: from v094114.home.net.pl ([79.96.170.134]:61247 "HELO v094114.home.net.pl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1751743AbbGQALO (ORCPT ); Thu, 16 Jul 2015 20:11:14 -0400 From: "Rafael J. Wysocki" To: Tomeu Vizoso Cc: Alan Stern , "linux-pm@vger.kernel.org" , "linux-kernel@vger.kernel.org" , Len Brown , Greg Kroah-Hartman , Pavel Machek Subject: Re: [PATCH v4 2/3] PM / sleep: Go direct_complete if driver has no callbacks Date: Fri, 17 Jul 2015 02:37:53 +0200 Message-ID: <1960688.DPHZXQdxmJ@vostro.rjw.lan> User-Agent: KMail/4.11.5 (Linux/4.1.0-rc5+; KDE/4.11.5; x86_64; ; ) In-Reply-To: References: <6740752.elSTxeiQAg@vostro.rjw.lan> MIME-Version: 1.0 Content-Transfer-Encoding: 7Bit Content-Type: text/plain; charset="utf-8" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2983 Lines: 75 On Thursday, July 16, 2015 10:47:51 AM Tomeu Vizoso wrote: > On 16 July 2015 at 02:41, Rafael J. Wysocki wrote: > > On Wednesday, July 15, 2015 02:47:50 PM Alan Stern wrote: > >> On Wed, 15 Jul 2015, Tomeu Vizoso wrote: > >> > >> > If a suitable prepare callback cannot be found for a given device and > >> > its driver has no PM callbacks at all, assume that it can go direct to > >> > complete when the system goes to sleep. > >> > > >> > The reason for this is that there's lots of devices in a system that do > >> > no PM at all and there's no reason for them to prevent their ancestors > >> > to do direct_complete if they can support it. > >> > > >> > Signed-off-by: Tomeu Vizoso > >> > --- > >> > > >> > drivers/base/power/main.c | 17 +++++++++++++++++ > >> > 1 file changed, 17 insertions(+) > >> > > >> > diff --git a/drivers/base/power/main.c b/drivers/base/power/main.c > >> > index 1710c26ba097..edda3f233c7c 100644 > >> > --- a/drivers/base/power/main.c > >> > +++ b/drivers/base/power/main.c > >> > @@ -1540,6 +1540,21 @@ int dpm_suspend(pm_message_t state) > >> > return error; > >> > } > >> > > >> > +static bool driver_has_no_pm_callbacks(struct device_driver *drv) > >> > +{ > >> > + if (!drv->pm) > >> > + return true; > >> > + > >> > + return !drv->pm->prepare && > >> > + !drv->pm->suspend && > >> > + !drv->pm->suspend_late && > >> > + !drv->pm->suspend_noirq && > >> > + !drv->pm->resume_noirq && > >> > + !drv->pm->resume_early && > >> > + !drv->pm->resume && > >> > + !drv->pm->complete; > >> > +} > >> > >> This isn't exactly what I meant. We also need to check the dev_pm_ops > >> fields in dev->pm_domain, dev->type, dev->class, and dev->bus. Only if > >> _all_ of these callbacks are missing should we use direct_complete. > > > > Also checking that on every suspend is kind of wasteful, because those things > > do not change very often. > > Do you have any suggestion on when would be a good time to do that > check? device_pm_sleep_init() and device_pm_add() are unfortunately > too early. The time to check that is (a) when the device is registered (for the bus type/class/device type), (b) when it is added to (or removed from) a PM domain and (c) when a driver is bound to (unbound from) it. > Alternatively we could check once on the first suspend and cache it, That information may be stale by the time we suspend next time. > but I'm not sure that complexity would be worth it. I don't think that adding extra overhead to *every* suspend can be justified easily, however. -- I speak only for myself. Rafael J. Wysocki, Intel Open Source Technology Center. -- 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/