Re: [Xen-devel] [PATCH] xen: update machine_to_phys_order on resume

[Ian Campbell <Ian.Campbell@eu.citrix.com>]

On Mon, 2011-07-18 at 09:47 +0100, Jan Beulich wrote:
> > @@ -960,6 +962,54 @@
> >      }
> >      break;
> >  
> > +    case XEN_DOMCTL_setfeatures:
> > +    {
> > +        struct domain *d;
> > +        ret = -ESRCH;
> > +        if ( (d = rcu_lock_domain_by_id(op->domain)) != NULL )
> > +        {
> > +            printk("dom%d set features[%d] = %#x\n", d->domain_id, op->u.setfeatures.submap_idx, op->u.setfeatures.submap);
> > +
> > +            switch (op->u.setfeatures.submap_idx) {
> > +            case 0:
> > +                if ( !paging_mode_translate(d) )
> 
> ... this condition looks inverted to me.

Quite possibly. I only ever actually tested this with a dodgy PV in HVM
container implementation.

> > +                {
> > +                    op->u.setfeatures.submap &= ~(1U<<XENFEAT_writable_page_tables);
> > +                    op->u.setfeatures.submap &= ~(1U<<XENFEAT_auto_translated_physmap);
> > +                }
> > +                if ( !is_pvhvm_domain(d) )
> > +                {
> > +                    op->u.setfeatures.submap &= ~(1U<<XENFEAT_supervisor_mode_kernel);
> > +                }
> > +
> > +                op->u.setfeatures.submap &= ~(1U<<XENFEAT_writable_descriptor_tables);
> 
> Why do you turn this off unconditionally?

Unless you build the hypervisor with supervisor_mode_kernel=1 (i.e.
never) then it does not support XENFEAT_writable_descriptor_tables. In
fact XENFEAT_supervisor_mode_kernel should also be unconditionally
cleared (the check is a remnant of the PV in HVM container stuff).

Note that XENFEAT_writable_descriptor_tables means that the guest kernel
should not make pagetable pages RO at all, which is different from the
hypervisor's support for writing to RO pagetables (i.e. emulating
pagetable updates).

> 
> > +
> > +                /* XXX other features */
> 
> That's perhaps also the place holder where the passed in information
> would actually get stored?

Yep ;-)

Ian.