Yes, something like that should do...
I'd prefer not to call ACPI functions from outside of drivers/acpi,
but this could be cleaned later.
On 5/9/07, Rafael J. Wysocki <rjw@xxxxxxx> wrote:
On Wednesday, 9 May 2007 15:34, Alexey Starikovskiy wrote:
> we needed to call pm_prepare() on boot kernel only to disable GPEs to
> resolve 7887. It will also blink the LEDs during the resume image
> load, but it's possible to live without it.
> If you drop call to pm_prepare(), we still need to disable GPEs, so
> they will not fire during image switch. So if we disable GPE in
> suspend() method of EC it will work as soon as there are no other so
> much active GPEs at boot time...
Okay, I just wanted to clarify.
In that case, I think, we should add a special callback to disable the GPEs
in the restore code path (before we restore the image).
I also think that it would be sufficient if this callback only executed the
following loop:
status = acpi_hw_disable_all_gpes();
if (ACPI_FAILURE(status)) {
return_ACPI_STATUS(status);
}
and it should be called right before disable_nonboot_cpus().
Is that correct?
Now, assumig that it's correct, we'll need a complementary callback to be
executed in the restore error code paths, right after enable_nonboot_cpus(),
to enable the GPEs if there's an error during the restore. I think it might
contain the following code:
status = acpi_hw_enable_all_runtime_gpes();
if (ACPI_FAILURE(status)) {
return_ACPI_STATUS(status);
}
Greetings,
Rafael
> On 5/9/07, Rafael J. Wysocki <rjw@xxxxxxx> wrote:
> > On Wednesday, 9 May 2007 13:59, Alexey Starikovskiy wrote:
> > > This patch will disable EC GPE, which seems to be major problem.
> >
> > Could you please explain to me why you think so?
> >
> > Greetings,
> > Rafael
> >
>
>
--
If you don't have the time to read,
you don't have the time or the tools to write.
- Stephen King
_______________________________________________
linux-pm mailing list
linux-pm@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linux-foundation.org/mailman/listinfo/linux-pm
