RE: [Xen-devel] [PATCH 04/14] Nested Virtualization: function hooks

["Dong, Eddie" <eddie.dong@xxxxxxxxx>]

> +int hvm_nestedhvm_vcpu_vmentry(struct vcpu *v)
> +{
> +    if (hvm_funcs.nestedhvm_vcpu_vmentry)
> +        return hvm_funcs.nestedhvm_vcpu_vmentry(v);
> +    return -EOPNOTSUPP;
> +}

I am not sure if we need a generic vmentry here. Will this require us to 
pre-read many vendor specific control state (i.e. vmcs & vmcs field) into a 
common data structure. From patch 10, it seems yes. Then it may be an 
overwhelmed access to VMCS registers.

> +
> +int hvm_nestedhvm_vcpu_vmexit(struct vcpu *v, struct cpu_user_regs
> *regs, +                             uint64_t exitcode)
> +{
> +    if (hvm_funcs.nestedhvm_vcpu_vmexit)
> +        return hvm_funcs.nestedhvm_vcpu_vmexit(v, regs, exitcode);
> +    return -EOPNOTSUPP;
> +}
> +

ditto

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel