RE: [PATCH v2] x86/cpu: Fix x86_match_cpu() to match just X86_VENDOR_INTEL

[Thomas Gleixner]

On Fri, May 17 2024 at 18:17, Luck, Tony wrote:
>> I'm confused. Why not simply use say -1 for wildcard vendor match, -2
>> for no vendor ID (no CPUID or other known probing mechanism) and -3
>> for unrecognized vendor (vendor detectable but not known.)
>
> It was really convenient to have "0" be the wildcard for all of
> vendor, family, model, stepping, feature because users of
> x86_match_cpu() could just initialize the fields they cared about in
> the struct x86_cpu_id and have the compiler make the rest be 0
> automagically.
>
> But X86_VENDOR_INTEL being zero has always been a thorn in that scheme.

As all initializers are macro based, that's really a non-problem.

Thanks,

        tglx