Re: [PATCH v2 4/7] KVM: x86: allow kvm_x86_ops.set_efer to return a value

[Sean Christopherson]

On Thu, Aug 20, 2020 at 02:43:56PM -0700, Jim Mattson wrote:
> On Thu, Aug 20, 2020 at 6:34 AM Maxim Levitsky <mlevitsk@xxxxxxxxxx> wrote:
> >
> > This will be used later to return an error when setting this msr fails.
> >
> > For VMX, it already has an error condition when EFER is
> > not in the shared MSR list, so return an error in this case.
> >
> > Signed-off-by: Maxim Levitsky <mlevitsk@xxxxxxxxxx>
> > ---
> 
> > --- a/arch/x86/kvm/x86.c
> > +++ b/arch/x86/kvm/x86.c
> > @@ -1471,7 +1471,8 @@ static int set_efer(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
> >         efer &= ~EFER_LMA;
> >         efer |= vcpu->arch.efer & EFER_LMA;
> >
> > -       kvm_x86_ops.set_efer(vcpu, efer);
> > +       if (kvm_x86_ops.set_efer(vcpu, efer))
> > +               return 1;
> 
> This seems like a userspace ABI change to me. Previously, it looks
> like userspace could always use KVM_SET_MSRS to set MSR_EFER to 0 or
> EFER_SCE, and it would always succeed. Now, it looks like it will fail
> on CPUs that don't support EFER in hardware. (Perhaps it should fail,
> but it didn't before, AFAICT.)

KVM emulates SYSCALL, presumably that also works when EFER doesn't exist in
hardware.

The above also adds weirdness to nested VMX as vmx_set_efer() simply can't
fail.