Re: [RFC PATCH v2 66/69] KVM: TDX: Add "basic" support for building and running Trust Domains

[Xiaoyao Li]

On 10/22/2021 5:44 AM, Sagi Shahar wrote:
>   On Fri, Jul 2, 2021 at 3:06 PM, Isaku Yamahata
> <isaku.yamahata@xxxxxxxxx> wrote:
> > Subject: [RFC PATCH v2 66/69] KVM: TDX: Add "basic" support for
> > building and running Trust Domains
> >
> >
> > +static int tdx_map_gpa(struct kvm_vcpu *vcpu)
> > +{
> > +       gpa_t gpa = tdvmcall_p1_read(vcpu);
> > +       gpa_t size = tdvmcall_p2_read(vcpu);
> > +
> > +       if (!IS_ALIGNED(gpa, 4096) || !IS_ALIGNED(size, 4096) ||
> > +           (gpa + size) < gpa ||
> > +           (gpa + size) > vcpu->kvm->arch.gfn_shared_mask << (PAGE_SHIFT + 1))
> > +               tdvmcall_set_return_code(vcpu, TDG_VP_VMCALL_INVALID_OPERAND);
> > +       else
> > +               tdvmcall_set_return_code(vcpu, TDG_VP_VMCALL_SUCCESS);
> > +
> > +       return 1;
> > +}
>
> This function looks like a no op in case of success. Is this
> intentional? Is this mapping handled somewhere else later on?

Yes, it's intentional.

The mapping will be exactly set up in EPT violation handler when the GPA 
is really accessed.