Re: [RFC 5/5] ACPI GPE based wakeup event detection

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Tuesday, 9 of September 2008, Li, Shaohua wrote:
> >> > +
> >> >  /**
> >> >   * acpi_bus_notify
> >> >   * ---------------
> >> > @@ -506,6 +519,8 @@ static void acpi_bus_notify(acpi_handle
> >> >     int result = 0;
> >> >     struct acpi_device *device = NULL;
> >> >
> >> > +   blocking_notifier_call_chain(&acpi_bus_notify_list,
> >> > +                   type, (void *)handle);
> >>
> >> Hm, perhaps I'm too tired and I'm missing something obvious, but can you
> >> tell me please why that has to be a notifier chain?  It looks like you
> >add only
> >> one notifier to it, so seemingly it could be replaced by a direct call to
> >a
> >> function like acpi_gpe_pme_handler() (with modified list of arguments).
> >When the notifier chain is used, it can work regardless of whether the
> >CONFIG_ACPI_GPE_WAKEUP is set.
> >If the CONFIG_ACPI_GPE_WAKEUP is set, what you said is also OK.
> I actually had another usage for this (I sent a patchset for docking station, which uses it)
> 
> >> >     if (acpi_bus_get_device(handle, &device))
> >> >             return;
> >> > Index: linux/drivers/acpi/sleep/wakeup.c
> >> > ===================================================================
> >> > --- linux.orig/drivers/acpi/sleep/wakeup.c  2008-09-08
> >14:28:55.000000000 +0800
> >> > +++ linux/drivers/acpi/sleep/wakeup.c       2008-09-08
> >15:04:23.000000000 +0800
> >> > @@ -142,6 +142,70 @@ void acpi_disable_wakeup_device(u8 sleep
> >> >     spin_unlock(&acpi_device_lock);
> >> >  }
> >> >
> >> > +#ifdef CONFIG_ACPI_GPE_WAKEUP
> >> > +static int acpi_gpe_pme_check(struct acpi_device *dev)
> >> > +{
> >> > +   struct device *ldev;
> >> > +
> >> > +   ldev = acpi_get_physical_device(dev->handle);
> >> > +   if (!ldev)
> >> > +           return -ENODEV;
> >> > +   /*
> >> > +    * AML code might already clear the event, so ignore the return
> >value.
> >> > +    * Actually we can't correctly detect which device invokes GPE if
> >the
> >> > +    * event is cleared.
> >> > +    */
> >> > +   if (ldev->bus->pm && ldev->bus->pm->base.wakeup_event)
> >> > +           ldev->bus->pm->base.wakeup_event(ldev);
> >> > +
> >> > +   /*
> >> > +    * We always send the event. AML code usually identifies the exact
> >> > +    * device of the GPE, let's trust it
> >> > +    */
> >> > +   device_receive_wakeup_event(ldev);
> >> > +
> >> > +   put_device(ldev);
> >> > +   return 0;
> >> > +}
> >> > +
> >> > +static int acpi_gpe_pme_handler(struct notifier_block *nb,
> >> > +   unsigned long type, void *data)
> >> > +{
> >> > +   int ret;
> >> > +   acpi_handle handle = data;
> >> > +   struct acpi_device *dev;
> >> > +
> >> > +   if (type != ACPI_NOTIFY_DEVICE_WAKE)
> >> > +           return NOTIFY_DONE;
> >> > +
> >> > +   if (acpi_bus_get_device(handle, &dev))
> >> > +           return NOTIFY_DONE;
> >> > +
> >> > +   ret = acpi_gpe_pme_check(dev);
> >> > +
> >> > +   acpi_disable_gpe(dev->wakeup.gpe_device, dev->wakeup.gpe_number,
> >> > +           ACPI_NOT_ISR);
> >> > +
> >>
> >> At which point are dev->wakeup.gpe_device and dev->wakeup.gpe_number
> >> determined, for example, for PCI devices, and how?
> >In the boot phase a PCI device will be bound with an ACPI device. In the
> >system based on ACPI(GPE) when a PCI device generates a wakeup event, an
> >ACPI interrupt will be triggered and the ACPI AML code will send the
> >device wakeup event to the corresponding ACPI device. In such case we
> >can get the gpe number for the PCI device.
> This is correct. For example:
> 
>             Method (_L07, 0, NotSerialized)
>             {
>                 Notify (\_SB.PCI0.SMBS, 0x02)
>             }
> 
>             Method (_L05, 0, NotSerialized)
>             {
>                 Notify (\_SB.PCI0.MODM, 0x02)
>             }
> 
> GPE 5,7 are wakeup GPEs

Well, it is not clear to me, though, which one of them is associated with the
PME# signal.

Also, the question was what piece of code in the kernel was responsible for the
identification of the wake-up GPEs and for setting dev->wakeup.gpe_device
and dev->wakeup.gpe_number as appropriate.  In particular, how do we figure
out which wake-up GPE will be used for signalling PME# for devices that are
not on-board?

Thanks,
Rafael