RE: [PATCH] ACPI: Execute the _PTS method when system reboot

From: Ocean HY1 He
Date: Thu Jun 23 2016 - 23:07:16 EST




Regards,
Ocean He
SW Development Dept.
Beijing Design Center
Enterprise Product Group
Mobile: 18911778926
E-mail: hehy1@xxxxxxxxxx
No.6 Chuang Ye Road, Haidian District, Beijing, China 100085


> -----Original Message-----
> From: rjwysocki@xxxxxxxxx [mailto:rjwysocki@xxxxxxxxx] On Behalf Of
> Rafael J. Wysocki
> Sent: Thursday, June 23, 2016 9:13 PM
> To: Ocean HY1 He
> Cc: Rafael J. Wysocki; lenb@xxxxxxxxxx; linux-acpi@xxxxxxxxxxxxxxx;
> linux-kernel@xxxxxxxxxxxxxxx; David Tanaka; Nagananda Chumbalkar
> Subject: Re: [PATCH] ACPI: Execute the _PTS method when system reboot
>
> On Thu, Jun 23, 2016 at 2:55 PM, Ocean HY1 He <hehy1@xxxxxxxxxx>
> wrote:
> > Hi Rafael,
> > Please see my reply in below.
> >
> > Regards,
> > Ocean He
> > SW Development Dept.
> > Beijing Design Center
> > Enterprise Product Group
> > Mobile: 18911778926
> > E-mail: hehy1@xxxxxxxxxx
> > No.6 Chuang Ye Road, Haidian District, Beijing, China 100085
> >
> >> -----Original Message-----
> >> From: Rafael J. Wysocki [mailto:rjw@xxxxxxxxxxxxx]
> >> Sent: Wednesday, June 22, 2016 7:56 AM
> >> To: Ocean HY1 He
> >> Cc: lenb@xxxxxxxxxx; linux-acpi@xxxxxxxxxxxxxxx;
> >> linux-kernel@xxxxxxxxxxxxxxx; David Tanaka; Nagananda Chumbalkar
> >> Subject: Re: [PATCH] ACPI: Execute the _PTS method when system
> reboot
> >>
> >> On Monday, May 09, 2016 05:50:11 AM Ocean HY1 He wrote:
> >> > The _PTS control method is defined in the section 7.4.1 of acpi 6.0
> >> > spec. The _PTS control method is executed by the OS during the sleep
> >> > transition process for S1, S2, S3, S4, and for orderly S5 shutdown.
> >> > The sleeping state value (For example, 1, 2, 3, 4 or 5 for the S5
> >> > soft-off state) is passed to the _PTS control method. This method
> >> > is called after OSPM has notified native device drivers of the sleep
> >> > state transition and before the OSPM has had a chance to fully
> >> > prepare the system for a sleep state transition.
> >> >
> >> > The _PTS control method provides the BIOS a mechanism for
> performing
> >> > some housekeeping, such as writing the sleep type value to the
> >> embedded
> >> > controller, before entering the system sleeping state.
> >> >
> >> > According to section 7.5 of acpi 6.0 spec, _PTS should run after _TTS.
> >> >
> >> > Thus, a _PTS block notifier is added to the reboot notifier list so that
> >> > the _PTS object will also be evaluated when the system reboot.
> >>
> >> So I understand why it may be necessary to evaluate _PTS before
> entering
> >> S5,
> >> but I'm totally unsure about reboot.
> >>
> >> What does reboot have to do with S5?
> >>
> > In ACPI spec, there is no explicit words saying _PTS should be
> > executed when reboot. But reboot could be equal to the
> > process S0->S5->S0.
>
> Not in general.
>
> In particular, wakeup devices that would be set up for S5 need not be
> set up for that. Also the mechanism by which transitions to S5 are
> entered is different from the reboot one, at least from the OS
> perspective.
>
> > Thus _PTS should be executed when reboot.
>
> No, it doesn't follow.
>
> > I am thinking this is the same as _TTS. In ACPI spec, there is also
> > no explicit words saying _TTS should be executed when reboot.
> > But kernel executes _TTS when reboot indeed.
>
> Yes, it does. Maybe it shouldn't?
>
> It may not hurt to call _PTS before reboot too, but is it guaranteed
> to work across the board on all systems everywhere?
>
I try to clarify the key point of this case: does devices should go to S5(shutdown) when reboot?

I think the answer is yes.
And It has no hurt to let devices go to S5 before reboot is invoked, here is the reasons:
#1 The new _PTS codes block nothing thus reboot can be guaranteed to be invoked eventually.
#2. Devices are mandatory to support S5 state, this means go to S5 could be a safe trip.
#3 Reboot would cause devices re-initialization from the scratch.

What's your decision then? ;-)

Regards,
Ocean.

> >> > Signed-off-by: Ocean He <hehy1@xxxxxxxxxx>
> >> > Signed-off-by: Nagananda Chumbalkar <nchumbalkar@xxxxxxxxxx>
> >> > ---
> >> > drivers/acpi/sleep.c | 27 +++++++++++++++++++++++++++
> >> > 1 file changed, 27 insertions(+)
> >> >
> >> > diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c
> >> > index 2a8b596..8b290fb 100644
> >> > --- a/drivers/acpi/sleep.c
> >> > +++ b/drivers/acpi/sleep.c
> >> > @@ -55,6 +55,26 @@ static struct notifier_block tts_notifier = {
> >> > .priority = 0,
> >> > };
> >> >
> >> > +static int pts_notify_reboot(struct notifier_block *this,
> >> > + unsigned long code, void *x)
> >> > +{
> >> > + acpi_status status;
> >> > +
> >> > + status = acpi_execute_simple_method(NULL, "\\_PTS",
> >> ACPI_STATE_S5);
> >> > + if (ACPI_FAILURE(status) && status != AE_NOT_FOUND) {
> >> > + /* It won't break anything. */
> >> > + printk(KERN_NOTICE "Failure in evaluating _PTS
> object\n");
> >> > + }
> >> > +
> >> > + return NOTIFY_DONE;
> >> > +}
> >> > +
> >> > +static struct notifier_block pts_notifier = {
> >> > + .notifier_call = pts_notify_reboot,
> >> > + .next = NULL,
> >> > + .priority = 0,
> >> > +};
> >> > +
> >> > static int acpi_sleep_prepare(u32 acpi_state)
> >> > {
> >> > #ifdef CONFIG_ACPI_SLEEP
> >> > @@ -896,5 +916,12 @@ int __init acpi_sleep_init(void)
> >> > * object can also be evaluated when the system enters S5.
> >> > */
> >> > register_reboot_notifier(&tts_notifier);
> >> > +
> >> > + /*
> >> > + * According to section 7.5 of acpi 6.0 spec, _PTS should run
> after
> >> > + * _TTS when the system enters S5.
> >> > + */
> >> > + register_reboot_notifier(&pts_notifier);
> >>
> >> Why do you have to add a second notifier?
> >>
> >> Why can't _TTS and _PTS be evaluated from one notifier?
> >>
> > If execute _PTS method in tts_notify_reboot(), then it would break
> > definition of tts_notify_reboot().
>
> What exactly would it break?
>
> > My intention is to keep new codes
> > has limited impact on existed codes.
>
> Even if that makes a little sense?
>
> > Of course, it's possible to merge _TTS and _PTS into one unified notifier.
> > The advantage is more actions could be added into the unified notifier
> in future.
> > Which way you prefer?
>
> I would just use one notifier.
I follow you.
>
> Thanks,
> Rafael