2020-10-15 11:52:01

by Borislav Petkov

[permalink] [raw]
Subject: DPTF_POWER and DPTF_PCH_FIVR

Dudes,

what are those new symbols in Kconfig and why do they wanna get enabled
on my box when doing oldconfig? Depends on X86? Really, that widespread?

The help text doesn't say why do I need this...

Thx.

--
Regards/Gruss,
Boris.

https://people.kernel.org/tglx/notes-about-netiquette


2020-10-15 12:35:29

by Borislav Petkov

[permalink] [raw]
Subject: Re: DPTF_POWER and DPTF_PCH_FIVR

On Thu, Oct 15, 2020 at 01:31:55PM +0200, Rafael J. Wysocki wrote:
> Well, they depend on ACPI too and it doesn't look like there is anything
> more fine grained that could be used here.
>
> Also I'm not sure if adding ACPI_DPTF (bool, disabled by default and
> depending on x86) would help a lot.

So you're saying this is going to be present in current and future
hardware so might as well enable it?

> Obviously an oversight on my part, sorry.
>
> I think that changing it to "dptf_power" (and analogously for the FIVR
> participant driver) should work.  Srinivas?

Makes sense.

Thx.

--
Regards/Gruss,
Boris.

https://people.kernel.org/tglx/notes-about-netiquette

2020-10-15 12:40:40

by Rafael J. Wysocki

[permalink] [raw]
Subject: Re: DPTF_POWER and DPTF_PCH_FIVR

On Thu, Oct 15, 2020 at 2:06 PM Borislav Petkov <[email protected]> wrote:
>
> On Thu, Oct 15, 2020 at 01:31:55PM +0200, Rafael J. Wysocki wrote:
> > Well, they depend on ACPI too and it doesn't look like there is anything
> > more fine grained that could be used here.
> >
> > Also I'm not sure if adding ACPI_DPTF (bool, disabled by default and
> > depending on x86) would help a lot.
>
> So you're saying this is going to be present in current and future
> hardware so might as well enable it?

Yes. [Only on clients AFAICS, but still.]

In the meantime I have concluded that having a submenu for it would
make sense, though. At least from the clarity POV.

2020-10-15 18:26:47

by Borislav Petkov

[permalink] [raw]
Subject: Re: DPTF_POWER and DPTF_PCH_FIVR

On Thu, Oct 15, 2020 at 11:53:11AM +0200, Borislav Petkov wrote:
> Dudes,
>
> what are those new symbols in Kconfig and why do they wanna get enabled
> on my box when doing oldconfig? Depends on X86? Really, that widespread?
>
> The help text doesn't say why do I need this...

And that thing in sysfs:

"/sys/bus/platform/drivers/DPTF Platform Power"

has spaces in the name. What's going on?!

--
Regards/Gruss,
Boris.

https://people.kernel.org/tglx/notes-about-netiquette

2020-10-15 18:27:17

by Wysocki, Rafael J

[permalink] [raw]
Subject: Re: DPTF_POWER and DPTF_PCH_FIVR

On 10/15/2020 11:58 AM, Borislav Petkov wrote:
> On Thu, Oct 15, 2020 at 11:53:11AM +0200, Borislav Petkov wrote:
>> Dudes,
>>
>> what are those new symbols in Kconfig and why do they wanna get enabled
>> on my box when doing oldconfig? Depends on X86? Really, that widespread?

Well, they depend on ACPI too and it doesn't look like there is anything
more fine grained that could be used here.

Also I'm not sure if adding ACPI_DPTF (bool, disabled by default and
depending on x86) would help a lot.


>> The help text doesn't say why do I need this...
> And that thing in sysfs:
>
> "/sys/bus/platform/drivers/DPTF Platform Power"
>
> has spaces in the name. What's going on?!
>
Obviously an oversight on my part, sorry.

I think that changing it to "dptf_power" (and analogously for the FIVR
participant driver) should work.  Srinivas?