Re: [PATCH v2 2/6] KVM: nSVM: Simplify error handling of nested_svm_copy_vmcb12_to_cache()

[Yosry Ahmed]

> I strongly dislike the if-elif approach, because it makes unnecessarily hard to
> see that *all* ret !=0 cases are handled, i.e. that overwriting ret below is ok.
>
> The comment is also super confusing, because there's no #UD in sight, but there
> is a #GP.
>
> This is what I have locally and am planning on pushing to kvm-x86/next.
>
>         ret = nested_svm_copy_vmcb12_to_cache(vcpu, vmcb12_gpa);
>         if (ret) {
>                 if (ret == -EFAULT) {
>                         kvm_inject_gp(vcpu, 0);
>                         return 1;
>                 }
>
>                 /* Advance RIP past VMRUN as part of the nested #VMEXIT. */
>                 return kvm_skip_emulated_instruction(vcpu);
>         }
>
>         /* At this point, VMRUN is guaranteed to not fault; advance RIP. */
>         ret = kvm_skip_emulated_instruction(vcpu);

Looks good. I will rebase on top of this and drop the patch before
sending the next version.