Re: [PATCH v8 4/5] xen/riscv: probe the A/D update scheme if the DT advertises neither

[Oleksii Kurochko <oleksii.kurochko@xxxxxxxxx>]

On 10/26/26 5:59 PM, Jan Beulich wrote:
> On 26.10.2026 16:35, Oleksii Kurochko wrote:
> > On 10/26/26 4:14 PM, Baptiste Le Duc wrote:
> > > +/*
> > > + * Probe whether hardware updates the PTE A/D bits by writing to a local
> > > + * variable through an alias mapping whose A and D bits are clear.
> > > + *
> > > + * Check the D bit rather than the A bit, as the A bit may be set
> > > + * speculatively while the D bit can't.
> Following Andrew's clarification, I don't think this is relevant anymore
> (but I also don't mind it to stay the way it is): If we observed the A
> bit set (and no fault), we can still deduce Svadu, no matter whether the
> update was due to an architectural or a speculative access. Svade would
> implicitly disallow speculative updates of A, as exceptions may only be
> raised for architectural accesses (unless there are provisions in the
> spec that I'm unaware of).
>
> Both of you: Please clarify whether you want this change to go in as-is,
> or whether this wants updating again.
I'd prefer to update it: the statement itself is true, but it isn't a 
reason to check D rather than A, since, as you say, observing A set 
without a fault would equally imply Svadu.
If you're fine adjusting it while committing, the second paragraph could 
become:
  Check the D bit, as it is set only once the store through the alias
  has been performed.

(or just be dropped), with the matching sentence in the description 
("The D bit is checked rather than the A bit, as the A bit may be set 
speculatively.") dropped as well. Otherwise Baptiste, please send an 
updated version.
~ Oleksii

> > Generally, looks good to me:
> >
> > Reviewed-by: Oleksii Kurochko <oleksii.kurochko@xxxxxxxxx>
> Acked-by: Jan Beulich <jbeulich@xxxxxxxx>