Re: [PATCH] x86: KVM: Add feature flag for AMD's FsGsKernelGsBaseNonSerializing

[Jim Mattson]

On Tue, Oct 3, 2023 at 8:27 PM Dave Hansen <dave.hansen@xxxxxxxxx> wrote:
>
> On 10/3/23 19:44, Jim Mattson wrote:
> > I'm a little surprised at the pushback, TBH. Are you implying that
> > there is some advantage to *not* passing this bit through?
>
> I'm not really trying to push back.  I'm honestly just curious.  Linux
> obviously doesn't cat about the bit.  So is this for some future Linux
> or some other OS?

It's not for any particular guest OS. It's just for correctness of the
virtual CPU. Pedantically, hardware that enumerates this bit cannot
run a guest that doesn't. Pragmatically, it almost certainly doesn't
matter. Getting it right is trivial and has no impact on performance
or code size, so why not just do it?