Subject: Re: CVS commit: src/sys/arch/i386/i386
To: Julio M. Merino Vidal <jmmv84@gmail.com>
From: Juan RP <juan@xtrarom.org>
List: source-changes
Date: 07/25/2007 16:51:52
On Wed, 25 Jul 2007 16:45:52 +0200
"Julio M. Merino Vidal" <jmmv84@gmail.com> wrote:
> 3? I guess you meant 6 there.
Yes, I meant 6 sorry.
> x86/include/cputypes.h defines much more CPUVENDOR_* values.
Yeah, but the check about CPUID2_EST will only be true with
CPUVENDOR_INTEL and CPUVENDOR_IDT.
> > What do you see wrong here? I don't understand you.
>
> I was just asking you if you were certain that the change was right,
> because it makes the code behave differently than it did even though
> the commit message did not say why.
>
> The problem is that p3_get_bus_clock may now be called with some
> values it is supposed not to take. I am not sure if this *can*
> happen or not, but it might. If that's the case, people with family
> 6, non-Intel, non-IDT CPUs will start to see an extra message during
> boot saying that the calculation of the bus clock failed.
I'm aware about all the things that you said, and it's not possible to
pass any other cpu_vendor than CPUVENDOR_INTEL/IDT, because they
are the only CPUs that implement Enhanced Speedstep.
Correct me if I'm wrong.
--
Juan Romero Pardines - The NetBSD Project
http://plog.xtrarom.org - NetBSD/pkgsrc news in Spanish