Re: [PATCH 0/3] Add CpuidUserDis support

[Alejandro Vallejo <alejandro.vallejo@xxxxxxxxx>]

On Mon, May 08, 2023 at 11:06:31AM +0200, Jan Beulich wrote:
> On 05.05.2023 19:57, Alejandro Vallejo wrote:
> > Nowadays AMD supports trapping the CPUID instruction from ring3 to ring0,
> 
> Since it's relevant for PV32: Their doc talks about CPL > 0, i.e. not just
> ring 3. Therefore I wonder whether ...

Very true. It's CPL>0, not just ring3. Noted and changed on v2.

> 
> > (CpuidUserDis)
> 
> ... we shouldn't deviate from the PM and avoid the misleading use of "user"
> in our internal naming.
> 
> Jan
> 

IMO it's going to be confusing enough as is. We'll eventually have
virtualized versions of both Intel and AMD that may or may not be
cross-hooked with each other (e.g: virtualized Intel interface on
an AMD host) due to backward compatibility. That means we'll probably
want:

 1. A name for the Intel mechanism, currently "CPUID faulting"
 2. A name for the AMD mechanism, currently "CpuidUserDis"
 3. A generic name for the cpuid-can-be-trapped behaviour, currently
    overloaded with the Intel name (but could do with a Xen-specific one).
    It doesn't matter a lot now, but it will once the AMD interface is
    virtualized.

Sure, we could give it an alternative name on AMD, but we still need yet
another one to disambiguate the trapping behaviour from the specific
mechanism that does it.

Using the AMD manual name does have the upside that it's easier to check
the manual without having to remember the AMD-specific feature that
corresponds to a Xen-specific name. That said, if you have a good suggestion
I'm happy to change it. So long as in the end result is (1), (2) and (3)
have non-ambiguous names.

Alejandro