Sean Christopherson writes:
> On Mon, Mar 23, 2020 at 01:10:40PM +0100, Vitaly Kuznetsov wrote:
>> Sean Christopherson writes:
>>
>> > +
>> > + .runtime_ops = &svm_x86_ops,
>> > +};
>>
>> Unrelated to your patch but I think we can make the naming of some of
>> these functions more consistend
On 23/03/20 17:24, Vitaly Kuznetsov wrote:
> Sounds cool! (not sure that with only two implementations people won't
> call it 'over-engineered' but cool).
Yes, something like
#define KVM_X86_OP(name) .name = vmx_##name
(svm_##name for svm.c) and then
KVM_X86_OP(check_nested_events)
et
On Mon, Mar 23, 2020 at 05:24:56PM +0100, Vitaly Kuznetsov wrote:
> Sean Christopherson writes:
>
> > On Mon, Mar 23, 2020 at 01:10:40PM +0100, Vitaly Kuznetsov wrote:
> >> Sean Christopherson writes:
> >>
> >> > +
> >> > +.runtime_ops = &svm_x86_ops,
> >> > +};
> >>
> >> Unrelated to
On Mon, Mar 23, 2020 at 01:10:40PM +0100, Vitaly Kuznetsov wrote:
> Sean Christopherson writes:
>
> > +
> > + .runtime_ops = &svm_x86_ops,
> > +};
>
> Unrelated to your patch but I think we can make the naming of some of
> these functions more consistend on SVM/VMX, in particular I'd suggest
Sean Christopherson writes:
> Move the kvm_x86_ops functions that are used only within the scope of
> kvm_init() into a separate struct, kvm_x86_init_ops. In addition to
> identifying the init-only functions without restorting to code comments,
> this also sets the stage for waiting until after
Move the kvm_x86_ops functions that are used only within the scope of
kvm_init() into a separate struct, kvm_x86_init_ops. In addition to
identifying the init-only functions without restorting to code comments,
this also sets the stage for waiting until after ->hardware_setup() to
set kvm_x86_ops.