Michael Niewöhner has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/45000 )
Change subject: Revert "soc/intel/cannonlake: Enable ACPI timer emulation if PM timer is disabled" ......................................................................
Patch Set 6:
Patch Set 6:
Patch Set 6:
Patch Set 6: Code-Review+2
Patch Set 6: Code-Review+1
(1 comment)
Hey folks, does anyone have an opinion on this? I'd prefer to get it in, simply because it avoids complexity. I don't think, but can't prove, that it's really harmless to have both the emulation and the actual timer available.
Sorry about the delay in response. I digged through my notes as well as the specs I have access to, but I don't see any reason why we cannot have multiple timers available. I believe the confusion was around the comment in some of the specs that emulated timer needs to be enabled when the ACPI PM hardware timer is disabled which is required for power savings. But, it does not apply the other way around i.e. having emulated timer enabled does not require the ACPI PM hardware timer to be disabled.
Thanks for checking this. That's how I understood it, too
But keep in mind that the ACPI PM hardware timer would never be accessed because the emulation enablement will trap the IO access in the CPU. So it's sort of moot.
That's mostly right, but there's one exception: PCH itself can access the PM timer, for example TCO, which needs the hardware timer