Re: [issue2025] CPU P-States not available to powerd(8)

[Sepherosa Ziehau <sepherosa@xxxxxxxxx>]

On Thu, Mar 17, 2011 at 1:33 AM, Max Herrgard (via DragonFly issue
tracker) <bugs@crater.dragonflybsd.org> wrote:
>
> Max Herrgard <herrgard@gmail.com> added the comment:
>
> Yes, now it shows this:
>
> cpu_cst0: <ACPI CPU C-State> [tentative] on cpu0
> cpu_cst0: switching to generic Cx mode
> cpu_cst0: <ACPI CPU C-State> [attached!] on cpu0
> cpu_pst0.cpu0.acpi0.nexus0.root0
> cpu_pst0: <ACPI CPU P-State> [tentative] on cpu0
> cpu_pst0: control reg 127 0
> cpu_pst0: status reg 127 0
> cpu_pst0: freq 1800, pwr 67000, xlat 100, blat 9, cv e020298a, sv 0000018a
> cpu_pst0: freq 1000, pwr 21000, xlat 100, blat 9, cv e0202c82, sv 00000482
> cpu_pst0: _PPC 0
> cpu_pst0: <ACPI CPU P-State> [attached!] on cpu0
> ...
> cpu0: only one P-State is supported
> ACPI: domain0 P-State configuration check failed
>
> So I guess my box doesn't support it. But why does it then print two available
> frequencies (1000 MHz and 1800 MHz)?

The check is conducted according to AMD #32559 page 300.

Well, try this patch:
http://leaf.dragonflybsd.org/~sephe/0Fh_fid.diff

It may be dangerous, if BIOS does generate wrong P-States.

Best Regards,
sephe

-- 
Tomorrow Will Never Die