Re: [PATCH 03/13] x86/HV: Add new hvcall guest address host visibility support

[Tianyu Lan]

On 7/29/2021 10:09 PM, Dave Hansen wrote:
> On 7/29/21 6:01 AM, Tianyu Lan wrote:
> > On 7/29/2021 1:06 AM, Dave Hansen wrote:
> > > On 7/28/21 7:52 AM, Tianyu Lan wrote:
> > > > @@ -1986,7 +1988,9 @@ static int __set_memory_enc_dec(unsigned long
> > > > addr, int numpages, bool enc)
> > > >        int ret;
> > > >          /* Nothing to do if memory encryption is not active */
> > > > -    if (!mem_encrypt_active())
> > > > +    if (hv_is_isolation_supported())
> > > > +        return hv_set_mem_enc(addr, numpages, enc);
> > > > +    else if (!mem_encrypt_active())
> > > >            return 0;
> > >
> > > One more thing.  If you're going to be patching generic code, please
> > > start using feature checks that can get optimized away at runtime.
> > > hv_is_isolation_supported() doesn't look like the world's cheapest
> > > check.  It can't be inlined and costs at least a function call.
> >
> > Yes, you are right. How about adding a static branch key for the check
> > of isolation VM? This may reduce the check cost.
>
> I don't think you need a static key.
>
> There are basically three choices:
> 1. Use an existing X86_FEATURE bit.  I think there's already one for
>     when you are running under a hypervisor.  It's not super precise,
>     but it's better than what you have.
> 2. Define a new X86_FEATURE bit for when you are running under
>     Hyper-V.
> 3. Define a new X86_FEATURE bit specifically for Hyper-V isolation VM
>     support.  This particular feature might be a little uncommon to
>     deserve its own bit.
>
> I'd probably just do #2.

There is x86_hyper_type to identify hypervisor type and we may check 
this variable after checking X86_FEATURE_HYPERVISOR.

static inline bool hv_is_isolation_supported(void)
{
	if (!cpu_feature_enabled(X86_FEATURE_HYPERVISOR))
		return 0;

        if (x86_hyper_type != X86_HYPER_MS_HYPERV)
                return 0;

	// out of line function call:
	return __hv_is_isolation_supported();
}