[coreboot-gerrit] Change in coreboot[master]: soc/intel/{common,skl}: set ACPI_FADT_PLATFORM_CLOCK based on Kconfig

[Michael Niewöhner (Code Review)]

Attention is currently required from:  Felix Singer, Furquan Shaikh, Mariusz Szafrański, Paul Menzel, Subrata Banik, Angel Pons, Patrick Rudolph, Lance Zhao, Nico Huber, Tim Wawrzynczak, Suresh Bellampalli, Vanessa Eusebio, Michal Motyl.
Michael Niewöhner has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/57934 )

Change subject: soc/intel/{common,skl}: set ACPI_FADT_PLATFORM_CLOCK based on Kconfig
......................................................................

Patch Set 22:

(1 comment)

Commit Message:

https://review.coreboot.org/c/coreboot/+/57934/comment/92dec6e9_917783f0 
PS21, Line 14:
> Please mention the platforms that change behavior, e.g. APL/GLK
> because for those USE_PM_ACPI_TIMER will always be false. Maybe
> even the boards that default to !USE_PM_ACPI_TIMER.

Yes, that's actually what this change is about: fix the problem, that we're not spec-compliant. I can mention this more explicitly.
> DNV does not change afaics, as its soc_fill_fadt() overwrites
> what acpi_fill_fadt() does, I guess.

Correct. Oh, and I just realize that while DNV uses common block acpi, it sets flags redundantly....
> Do we know the situation on Xeon-SP? If it doesn't select
> PM_ACPI_TIMER_OPTIONAL, it will change behavior too.

Xeon-SP does not set that flag and the PM timer can't be disabled at the time of this patch. So it's wrong, because of the guarding `if PM_ACPI_TIMER_OPTIONAL`.

The fix should be simple: I'll add PM_ACPI_TIMER_OPTIONAL to the check for now, so only platforms change, that have choosable USE_PM_ACPI_TIMER.

After CB:58017 this problem will vanish because that guard is dropped.