probable typo in acpi_cpu.c

[John Baldwin]

On Monday 11 February 2008 06:59:19 pm Rui Paulo wrote:
> 
> On Feb 11, 2008, at 9:22 PM, Andriy Gapon wrote:
> 
> >
> > There is probably a typo in acpi_cpu.c, function acpi_cpu_startup:
> > 	    if (cpu_quirks && CPU_QUIRK_NO_C3) { <----- here
> > 		sc->cpu_cx_count = sc->cpu_non_c3 + 1;
> > 	    }
> >
> > Not sure how that affects the logic though.
> 
> You are right. This means that the C3 state will still be available on  
> systems like those based on PIIX4.
> 
> Nate, could you please commit a fix for it?

I've committed it, but I don't think it has that affect.  Instead, 
CPU_QUIRK_NO_C3 is always true, so it basically boils down to

	if (cpu_quirks) {
		/* Disable C3 */
	}

So I think the result is that C3 is disabled for other quirks (i.e. the 
NO_BM_CTRL quirk).  This includes SMP machines that don't have _CST.

-- 
John Baldwin