Re: [PATCH] x86/topo: Map vendor CPU types to generic Linux such types

On Wed, Jul 22, 2026 at 12:51:05PM -0700, Borislav Petkov wrote:
> On Wed, Jul 22, 2026 at 12:22:11PM -0700, Pawan Gupta wrote:
> > Unless I am misreading, TOPO_CPU_TYPE_ANY is introduced in this patch.
> > X86_*_ANY in x86_match_cpu() are much needed wildcards, I dont really
> > understand why x86_topology_cpu_type needs to carry a wildcard. It should
> > reflect the true state, PERFORMANCE, EFFICIENCY or UNKNOWN.
> > 
> > Between ANY and UNKNOWN, I would prefer UNKNOWN when CPU did not report a
> > type.
> 
> Because X86_CPU_TYPE_ANY is already there and it is 0 and TOPO_CPU_TYPE_ANY is
> supposed to be the same and mirror it.

This is exactly what I don't understand. For a wildcard X86_CPU_TYPE_ANY
that matches TOPO_CPU_TYPE_*, do we really need TOPO_CPU_TYPE_ANY for
matching to work?

> Did you even read my mail where I explain that we don't do multiple things in
> one patch?

Yes, but my main concern is about addition of TOPO_CPU_TYPE_ANY. If we
don't add TOPO_CPU_TYPE_ANY, second patch will not be needed. Sorry if I am
mis-interpreting things.

> > Ya. If need be, a mask of each possible cpu_type can be built for
> > dispatching work selectively.
>
> Instead of doing the completely sensible thing and keeping a per-CPU type?

My earlier comment was unnecessary, and putting us off-track here. Please
ignore this part.

(Just to clarify what I meant: For dispatching work to all P-cores in your
example scenario with something like on_each_cpu_mask() we would need a
mask of all P-cores. Such a mask can be built/cached using the per-CPU type
information).