Re: [PATCH v4] common: dom0less-bindings: introduce XSM labels

["Alejandro Vallejo" <alejandro.garciavallejo@amd.com>]

On Tue Oct 13, 2026 at 11:17 AM CEST, Jan Beulich wrote:
> On 13.10.2026 11:09, Alejandro Vallejo wrote:
> > On Mon Oct 12, 2026 at 12:15 PM CEST, Sergiy Kibrik wrote:
> >> --- a/xen/include/xsm/dummy.h
> >> +++ b/xen/include/xsm/dummy.h
> >> @@ -823,4 +823,11 @@ static XSM_INLINE int xsm_domain_resource_map(
> >>      return xsm_default_action(action, current->domain, d);
> >>  }
> >>  
> >> +static XSM_INLINE int xsm_lookup_sid(
> >> +    XSM_DEFAULT_ARG const char *ctx, uint32_t len, uint32_t *sid)
> >> +{
> >> +    /* dummy policy has no security contexts */
> >> +    return -EPERM;
> > 
> > nit: Perhaps -ENOSYS? or something like that. It's not that any
> > permission was violated, it's that the real version of this function is
> > absent.
>
> Yet then still e.g. -EOPNOTSUPP, but certainly not -ENOSYS.

That'd also be ok.

It would be helpful if there was a guide under docs/ in which we state
what each of our errnos are meant to be in Xen terms. ENOSYS/EOPNOTSUPP
are great examples of such mixed usage that it's disheartening trying to
reverse engineer the policy from their use.

Cheers,
Alejandro