Re: [PATCH v7 018/120] x86/cpuid: Parse CPUID(0x16)

[Ahmed S. Darwish]

On Thu, 18 Jun 2026, Borislav Petkov wrote:
>
> On Thu, May 28, 2026 at 05:37:40PM +0200, Ahmed S. Darwish wrote:
> > diff --git a/arch/x86/kernel/cpu/cpuid_parser.h b/arch/x86/kernel/cpu/cpuid_parser.h
> > index 8b0d44b745c5..ee1958f3d369 100644
> > --- a/arch/x86/kernel/cpu/cpuid_parser.h
> > +++ b/arch/x86/kernel/cpu/cpuid_parser.h
> > @@ -144,6 +144,7 @@ struct cpuid_parse_entry {
> >   */
> >  #define CPUID_COMMON_ENTRIES									\
> >  	/*			Leaf		Subleaf		Reader function */		\
> > +	CPUID_PARSE_ENTRY   (	0x16,		0,		generic			),	\
> >  	CPUID_PARSE_ENTRY   (	0x80000000,	0,		0x80000000		),	\
> >  	CPUID_PARSE_ENTRY   (	0x80000002,	0,		generic			),	\
> >  	CPUID_PARSE_ENTRY   (	0x80000003,	0,		generic			),	\
> > @@ -180,5 +181,6 @@ struct cpuid_vendor_entry {
> >
> >  #define CPUID_VENDOR_ENTRIES								\
> >  	/*		   Leaf		Vendor list		    */			\
> > +	CPUID_VENDOR_ENTRY(0x16,	X86_VENDOR_INTEL),				\
> >
>
> If this leaf is Intel-specific, why is it being added to the common entries
> too?

CPUID leaf parsing entries are split into:

  CPUID_EARLY_ENTRIES
  CPUID_COMMON_ENTRIES

CPUID_VENDOR_ENTRIES is not a parser entries table.  It just attaches
vendor tags to a leaf if/when a leaf needs it.

The confusion stems from naming the vendor tags table with an _ENTRIES
suffix, just like EARLY and COMMON.

I'll find a better name for it.  I guess CPUID_VENDOR_TAGS should work.

>
>  * For vendor-specific leaves, a matching entry must be added to the CPUID leaf
>  * vendor table later defined.  Leaves which are here, but without a matching
>  * vendor entry, are treated by the CPUID parser as valid for all x86 vendors.
>
> Why?
>
>
> I guess it has to do with that vendor parsing in 016 but parsing things twice
> better have a very good justification.
>

If a CPUID leaf parser entry has no vendor tag, then it's parsed for all
x86 vendors.

There's no parsing things twice here.  Sorry, I guess it just that way
because of the CPUID_VENDOR_ENTRIES name misnormer.

Thanks!
Ahmed