Re: [PATCH v5] x86: make Viridian support optional

[Marek Marczykowski-Górecki <marmarek@xxxxxxxxxxxxxxxxxxxxxx>]

On Mon, Oct 20, 2025 at 09:44:29AM +0100, Roger Pau Monné wrote:
> On Sat, Oct 18, 2025 at 09:21:37PM -0400, Demi Marie Obenour wrote:
> > On 10/17/25 11:52, Alejandro Vallejo wrote:
> > > On Wed Oct 15, 2025 at 7:58 AM CEST, Demi Marie Obenour wrote:
> > >> On 10/13/25 06:01, Alejandro Vallejo wrote:
> > >>> On Wed Oct 8, 2025 at 6:04 PM CEST, Jan Beulich wrote:
> > >>>> On 30.09.2025 14:52, Grygorii Strashko wrote:
> > >>>>> --- a/xen/arch/x86/hvm/Kconfig
> > >>>>> +++ b/xen/arch/x86/hvm/Kconfig
> > >>>>> @@ -62,6 +62,16 @@ config ALTP2M
> > >>>>>  
> > >>>>>         If unsure, stay with defaults.
> > >>>>>  
> > >>>>> +config VIRIDIAN
> > >>>>> +     bool "Hyper-V enlightenments for guests" if EXPERT
> > >>>>> +     default y
> > >>>>> +     help
> > >>>>> +       Support optimizations for Hyper-V guests such as faster 
> > >>>>> hypercalls,
> > >>>>> +       efficient timer and interrupt handling, and enhanced 
> > >>>>> paravirtualized
> > >>>>> +       I/O. This is to improve performance and compatibility of 
> > >>>>> Windows VMs.
> > >>>>
> > >>>> What is "paravirtualized I/O" about in this context?
> > >>>
> > >>> Hypervisor-assisted IPIs, TLB flushes, etc. Or so I understood back 
> > >>> when I said
> > >>> that looked ok. I see there could be confusion with Xen PV device 
> > >>> protocols,
> > >>> but as far as the user of the help message is concerned it makes no 
> > >>> difference.
> > >>>
> > >>> One could even remove the examples and leave it as "... for Hyper-V 
> > >>> guests. This
> > >>> is to...". They are truly inconsequential.
> > >>>
> > >>> All that matters is that (modern) Windows won't run without it, and 
> > >>> that it
> > >>> provides some indeterminate hypervisor-provided assists to try to 
> > >>> reduce some
> > >>> virtualization overheads.
> > >>
> > >> Qubes OS doesn't expose Viridian at all, which is why it wasn't
> > >> vulnerable to XSA-472.  It still runs Windows guests just fine.
> > > 
> > > Can you run Windows 11?
> > 
> > I haven't tried it, but it is documented as working.
> > 
> > > I don't remember which, but I do know some versions of Windows refuse to 
> > > boot
> > > if they determine they are virtualised and don't see the mandatory parts 
> > > of
> > > the TLFS.
> > > 
> > > If 11 works, maybe Windows Server?
> > 
> > Windows Server is more likely.
> 
> FTR, for unrelated reasons I've tested Windows Server 2025 without
> Viridian, and it does work, albeit painfully slow.

Windows 11 works too, but also "painfully slow". Interestingly, Windows
10 works much better in the same environment (but still far from native
performance).

-- 
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab

Attachment: signature.asc
Description: PGP signature