On Wed, 31 Jul 2024, Marek Maślanka wrote:
> On Wed, Jul 31, 2024 at 6:33 PM Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote:
> >
> > Marek!
> >
> > On Wed, Jul 31 2024 at 16:44, Marek Maślanka wrote:
> > > On Tue, Jul 30, 2024 at 6:08 PM Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote:
> > >> On Tue, Jul 30 2024 at 12:05, Marek Maslanka wrote:
> > >> +static void acpi_pm_disable(struct clocksource *cs)
> > >> +{
> > >> + acpi_pm_enabled = false;
> > >> + if (enable_callback)
> > >> + enable_callback(false);
> > >> +}
> > >> +
> > >> static struct clocksource clocksource_acpi_pm = {
> > >> .name = "acpi_pm",
> > >> .rating = 200,
> > >> .read = acpi_pm_read,
> > >> .mask = (u64)ACPI_PM_MASK,
> > >> .flags = CLOCK_SOURCE_IS_CONTINUOUS,
> > >> + .enable = acpi_pm_enable,
> > >> + .disable = acpi_pm_disable,
> > >> };
> > >>
> > > Thanks. I'll try do this in that way. But I need to disable/enable
> > > ACPI PM timer only on suspend/resume, so I'll use suspend/resume
> > > callbacks.
> >
> > Why? What's the point of keeping it running when nothing uses it?
> >
> > Thanks,
> >
> > tglx
>
> In case of Intel CPUs the watchdog (iTCO/wdat_wdt) is driven by ACPI PM
> Timer. But it may also be used by others that I don't know about, so I don't
> want to disable it.
Hi Marek,
This kind of non-obvious information should be put into the changelog
because it helps if after ten years somebody is looking into this change
and asks similar why questions.
--
i.
