Re: [PATCH for-4.9 v3 2/2] x86/io: move the list of guest to machine IO ports out of domain_iommu

[Roger Pau Monne <roger.pau@citrix.com>]

On Wed, Apr 05, 2017 at 03:17:17AM -0600, Jan Beulich wrote:
> >>> On 05.04.17 at 11:00, <roger.pau@citrix.com> wrote:
> > --- a/xen/arch/x86/domctl.c
> > +++ b/xen/arch/x86/domctl.c
> > @@ -721,14 +721,20 @@ long arch_do_domctl(
> >  
> >      case XEN_DOMCTL_ioport_mapping:
> >      {
> > -        struct domain_iommu *hd;
> >          unsigned int fgp = domctl->u.ioport_mapping.first_gport;
> >          unsigned int fmp = domctl->u.ioport_mapping.first_mport;
> >          unsigned int np = domctl->u.ioport_mapping.nr_ports;
> >          unsigned int add = domctl->u.ioport_mapping.add_mapping;
> > +        struct hvm_domain *hvm_domain;
> >          struct g2m_ioport *g2m_ioport;
> >          int found = 0;
> >  
> > +        ret = -EOPNOTSUPP;
> > +        if ( !is_hvm_domain(d) )
> > +        {
> > +            printk(XENLOG_G_ERR "ioport_map against non-HVM domain\n");
> > +            break;
> > +        }
> >          ret = -EINVAL;
> 
> There should be a blank line above this one, which can of course
> be added while committing. With that
> Reviewed-by: Jan Beulich <jbeulich@suse.com>
> but I think we should still get clarification on the supposed earlier
> regression before committing this.

Thanks.

The only callers of this domctl (xc_domain_ioport_mapping) are in QEMU, so it's
very unlikely that this domctl was ever used with PV guests (also taking into
account what Jan said about the iommu struct).

Roger.

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