On Mon, 2023-12-04 at 09:46 +0100, Jan Beulich wrote:
> On 01.12.2023 21:48, Oleksii Kurochko wrote:
> > --- a/xen/common/Kconfig
> > +++ b/xen/common/Kconfig
> > @@ -87,7 +87,7 @@ config MEM_ACCESS_ALWAYS_ON
> >  config MEM_ACCESS
> >     def_bool MEM_ACCESS_ALWAYS_ON
> >     prompt "Memory Access and VM events" if
> > !MEM_ACCESS_ALWAYS_ON
> > -   depends on HVM
> > +   depends on HVM && (ARM || X86)
> 
> While - unlike for GRANT_TABLE - I view going this route as an
> option, I
> still think it wants doing without explicitly naming architectures
> here.
> There wants to be a HAS_MEM_ACCESS, selected by Arm's Kconfig and by
> MEM_ACCESS_ALWAYS_ON (which in turn x86 selects). This could then
> also
> replace the HVM dependency here - x86 ought to select
> MEM_ACCESS_ALWAYS_ON
> only when HVM is enabled.
Thanks.

It will be definetely better to introduce HAS_MEM_ACCESS.


~ Oleksii

Reply via email to