Re: [Xen-devel] [PATCH v14 05/17] pvh: Introduce PVH guest type

[Tim Deegan <tim@xxxxxxx>]

At 11:21 +0000 on 07 Nov (1383819695), George Dunlap wrote:
> On 06/11/13 23:28, Tim Deegan wrote:
> > At 12:14 +0000 on 04 Nov (1383563694), George Dunlap wrote:
> >> --- a/xen/common/domain.c
> >> +++ b/xen/common/domain.c
> >> @@ -239,6 +239,17 @@ struct domain *domain_create(
> >>   
> >>       if ( domcr_flags & DOMCRF_hvm )
> >>           d->guest_type = guest_type_hvm;
> >> +    else if ( domcr_flags & DOMCRF_pvh )
> >> +    {
> >> +        if ( !(domcr_flags & DOMCRF_hap) )
> >> +        {
> >> +            err = -EOPNOTSUPP;
> >> +            printk(XENLOG_INFO "PVH guest must have HAP on\n");
> >> +            goto fail;
> >> +        }
> >> +        d->guest_type = guest_type_pvh;
> >> +        printk("Creating PVH guest d%d\n", d->domain_id);
> >> +    }
> > This check seems like it should be in arch-specific code.  If it
> > were in arch_domain_create(), it would also correctly handle the
> > case where the tools asked for PVH+HAP but HAP wasn't available.
> 
> Looking at the HVM case, this (and Jan's comments on the vmcs patch from 
> v13) should probably be handled the same way: i.e., at start of day, see 
> if we have the necessary hardware support to run in pvh mode; and set 
> "pvh_enabled" (analog to hvm_enabled) accordingly. Then we can check 
> this in hvm_domain_initialise() just as we do for hvm guests.

Yep, that sounds like a good idea.

Tim.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel