Re: [PATCH 3/8] acpi-cpufreq: merge acpi functionality of cpufreq drivers

[Bruno Ducrot <ducrot@poupinou.org>]

On Thu, Aug 03, 2006 at 12:20:24PM -0400, Dave Jones wrote:
> On Thu, Aug 03, 2006 at 06:11:30PM +0200, Bruno Ducrot wrote:
>  > On Wed, Aug 02, 2006 at 10:44:34PM +0400, Alexey Starikovskiy wrote:
>  > > Bruno Ducrot wrote:
>  > > > I've not checked carrefully but it seems to me you are breaking
>  > > > powernow-k8.
>  > > We do not touch powernow-k8, and tried to make sure we do not change
>  > > acpi-cpufreq to exclude any unknown hardware. What we tried to do, is to merge
>  > > different ways of accessing cpufreq with ACPI under single driver. Could you be
>  > > more specific how we break powernow-k8?
>  > 
>  > I haven't seen the if (!check_speedstep_cpu(cpu)) {...} onto
>  > apci_cpufreq_cpu_init_acpi() indeed.
>  > 
>  > Therefore you are correct.
>  > 
>  > But it's funny to see that how acpi_cpufreq driver become
>  > so Intel specific.
> 
> I've not looked over these patches in detail yet, but one precondition
> is that acpi-cpufreq _must_ continue to work on non-Intel.

I think it's the case.

> Newer VIA CPUs rely on this driver rather than longhaul for example.

Ah I see.  In fact we could add them already to
speedstep-centrino.c easily.
It seems also (though I've not checked carrefully yet)
that it might be better to add newest AMD powernow stuff
to speedstep-centrino.c instead of powernow-k8.c.

Therefore, it might be better to add all that stuff
to acpi-cpufreq.c instead of speedstep-centrino.c after all.

-- 
Bruno Ducrot

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.