Re: [RFC][PATCH 3/3] x86: make MP a required-feature on 64-bit

[H. Peter Anvin]

On 06/20/2014 11:15 AM, Borislav Petkov wrote:
> On Fri, Jun 20, 2014 at 10:50:25AM -0700, H. Peter Anvin wrote:
>> Looking at the AMD init code, there is a whole bunch of other 32/64-bit
>> differences that are clearly bogus.  I also see that amd_k7_smp_check()
>> doesn't even *exist* on 64 bits, and that init_amd_k7() which calls
>> amd_k7_smp_check() only is ever called for family == 6, despite having
>> tests for family 7 and above in it.
> 
> No, K7 is family 6. The tests are for c->x86_model, or am I looking at
> the wrong place?
> 
> OTOH, init_amd() could probably use a cleanup of moving the per-family
> code into init_amd_<fam>() functions and extending the switch-case.
> 
> I'll take a look.
> 

Ah, yes, you're right.

This code is clearly not applicable to any 64-bit CPU, so cpu_has_mp is
simply a noop on 64 bits... so no need for Dave H. to worry about it at
all; we should get rid of it and replace it with cpu_has() in the AMD code.

I actually have Linus' old dual-processor K7 sitting in my garage, but
$DEITY knows if it actually runs.

	-hpa

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/