acpi_cpu: _PDC vs _OSC

[John Baldwin]

On Thursday 04 February 2010 5:13:29 pm Andriy Gapon wrote:
> on 04/02/2010 23:57 John Baldwin said the following:
> > On Thursday 04 February 2010 2:42:47 pm Andriy Gapon wrote:
> >> -	 * TODO: evaluate failure of _OSC.
> >> +	 * On some systems evaluation of _OSC/_PDC dynamically
> >> +	 * loads the _PSS and other methods.
> >>  	 */
> > 
> > I would only say _OSC here.  I don't think we've seen any systems that 
load 
> > something when _PDC is invoked, only when _OSC is invoked.
> 
> Actually, I think that the way it's written should be OK.
> I've seen a few DSDTs where both are present and both do the same thing.
> E.g.:
> 
>     Scope (\_PR.CPU0)
>     {
>         Name (HI0, Zero)
>         Name (HC0, Zero)
>         Method (_PDC, 1, NotSerialized)
>         {
>             Store (CPDC (Arg0), Local0)
>             GCAP (Local0)
>             Return (Local0)
>         }
> 
>         Method (_OSC, 4, NotSerialized)
>         {
>             Store (COSC (Arg0, Arg1, Arg2, Arg3), Local0)
>             GCAP (Local0)
>             Return (Local0)
>         }
> ...
> 
> Looks like CPDC is "Convert _PDC" and COSC is "Convert _OSC" and GCAP is 
"G...
> capabilities", whatever "G..." could mean.

But is GCAP loading an additional SSDT?  That is what the "loading something" 
refers to and I think we've only observed that occurring with _OSC.  I'd 
rather we only document unexpected quirks that someone has actually reported 
and not assume that just because an _OSC method on some box did it, there's 
bound to be a _PDC method on some other box that does it.  In truth, the 
comment is probably not needed now anyway since this will always do _OSC 
first.

-- 
John Baldwin