Re: [RFC 02/10] x86/hvm: separate AMD-V and Intel VT-x hvm_function_table initializers

[Xenia Ragiadakou <burzalodowa@xxxxxxxxx>]

On 2/13/23 18:47, Jan Beulich wrote:
> On 13.02.2023 15:57, Xenia Ragiadakou wrote:
> > --- a/xen/arch/x86/include/asm/hvm/hvm.h
> > +++ b/xen/arch/x86/include/asm/hvm/hvm.h
> > @@ -261,8 +261,16 @@ extern struct hvm_function_table hvm_funcs;
> >   extern bool_t hvm_enabled;
> >   extern s8 hvm_port80_allowed;
> >   
> > +#ifdef CONFIG_AMD_SVM
> >   extern const struct hvm_function_table *start_svm(void);
> > +#else
> > +static inline const struct hvm_function_table *start_svm(void) { return NULL; }
> > +#endif
> > +#ifdef CONFIG_INTEL_VMX
> >   extern const struct hvm_function_table *start_vmx(void);
> > +#else
> > +static inline const struct hvm_function_table *start_vmx(void) { return NULL; }
> > +#endif
> >   
> >   int hvm_domain_initialise(struct domain *d,
> >                             const struct xen_domctl_createdomain *config);
> Instead of this (which I consider harder to read), may I suggest
>
>      if ( IS_ENABLED(CONFIG_VMX) && cpu_has_vmx )
>          fns = start_vmx();
>      else if ( IS_ENABLED(CONFIG_SVM) && cpu_has_svm )
>          fns = start_svm();
>
> in hvm_enable() instead (with DCE taking care of removing the dead
> calls)?
Sure, it looks much better this way.

> Jan
--
Xenia