Re: [PATCH 2/3] x86/split_lock: Bits in IA32_CORE_CAPABILITIES are not architectural

[Sean Christopherson]

On Sat, Apr 18, 2020 at 12:15:57PM +0800, Xiaoyao Li wrote:
> So now, it's tightly associated with CPU model, which makes it harder to
> expose this feature to guest. For guest, the CPU model can be configured to
> anything.
> 
> As suggested by Sean internally, we'd better use a KVM CPUID to expose it to
> guest, which makes it independent of CPU model.

Making this a paravirt feature from a KVM perspective would also let us do
the whole STICKY bit thing straight away.  I don't like paravirtualizing
something that could be emulated as-is, I but dislike it less than exposing
features based on CPU model.