On Tue, 2022-11-01 at 13:24 -0400, Sven van Ashbrook wrote: > On Mon, Oct 31, 2022 at 4:55 PM Limonciello, Mario > <mario.limonciello@xxxxxxx> wrote: > > I just spun together an RFC series for this idea and while doing it I > > had the same realization. So I left the warning messages in place for > > both drivers. > > > > You can take a look at the series here: > > > > https://lore.kernel.org/platform-driver-x86/20221031204320.22464-1-mario.limonciello@xxxxxxx/T/#m6c7db55c98b8a3ce8c48d451fc01c1d9b0df37fb > > > > We've had some internal discussions within ChromeOS intel big core, > and we believe this is a worthwhile effort, and we are supportive, as > long as our current S0ix fail detection will not break for the > foreseeable future, i.e. as long as the warning message and register > dump stays in place. Which is the case for your RFC. Yeah, I did not see this as a replacement for anything in the pmc drivers. Given the prevalence of S0ix, and that hardware based low power idle support is indicated in the FADT (so part of the standard) it makes sense to have it tracked by the suspend core, particularly when it's being used as a replacement for S3. We don't need to collect any implementation or debug details there. Only detect when it's available, being used for suspend, and being achieved. Maybe residency information as well if available but that's it. Other information is separate and should be contained to the individual drivers which have the detailed platform knowledge. David > > +swboyd@xxxxxxxxxxxx who expressed interest in doing something similar for > ARM.
