Re: [PATCH v4 00/34] x86: Introduce a centralized CPUID data model

[David Woodhouse <dwmw2@infradead.org>]

[-- Attachment #1: Type: text/plain, Size: 1157 bytes --]

On Fri, 2025-08-15 at 09:01 +0200, Ahmed S. Darwish wrote:
> 
> This series introduces a centralized CPUID model for the x86
> subsystem.
> 
> Rationale for this work can be found at:
> 
>     https://lore.kernel.org/lkml/874ixernra.ffs@tglx
> 
>     https://gitlab.com/x86-cpuid.org/x86-cpuid-db
> 
> The first 5 patches can be independently applied.

Nice work. Looks like you haven't attempted to address hypervisor CPUID
yet. I've attempted to document that in a section at the end of 
http://david.woodhou.se/ExtDestId.pdf — I wonder if we should find
somewhere to publish it as canonical?

I suspect our loop in cpuid_base_hypervisor() should be 'fixed' to
comply with the new rule I just made up, that it should only scan each
block at 0x4000_0.00 until it finds an empty block, rather than going
all the way up to 0x4001_0000?

Are there any hypervisors which provide more than one block, that
*aren't* just putting the Hyper-V leaves at 0x4000_0000 and their own
native leaves at 0x4000_0100 ?

Even when we support Xen under KVM we expose *only* the Xen leaves and
don't put the KVM leaves at 0x4000_0100 too.

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 5069 bytes --]