Re: PCI passthrough for HVM with stubdomain broken by "tools/libxl: handle the iomem parameter with the memory_mapping hcall"

["Marek Marczykowski-Górecki" <marmarek@invisiblethingslab.com>]

[-- Attachment #1.1: Type: text/plain, Size: 1531 bytes --]

On Thu, Jun 23, 2016 at 09:37:09AM -0600, Jan Beulich wrote:
> >>> On 23.06.16 at 17:22, <marmarek@invisiblethingslab.com> wrote:
> > xen: allow XEN_DOMCTL_getdomaininfo for device model domains
> > 
> > Allow device model domain to get info about its target domain.
> > It is used during PCI passthrough setup (xc_domain_memory_mapping
> > checks for guest being auto-translated). While it happens in stubdomain,
> > it failed, breaking PCI passthrough in such setup.
> 
> If that's the route to go (which I'm not convinced of, as I'm not sure
> we won't need other xenstore domain special casing later on) I'd
> really like to ask you to mention the other broken case too, as
> described in my original patch (unless you found I was wrong with
> that).

So, maybe something like this:
      case XEN_DOMCTL_getdomaininfo:
          if ( current-domain->is_xenstore )
              return xsm_default_action(XSM_XS_PRIV, current->domain, d);;
          return xsm_default_action(XSM_DM_PRIV, current->domain, d);


In your patch (changing XSM_XS_PRIV semantic), you implicitly considered
all domctls allowed for xenstore domain to be always a subset of those
allowed for device model domain. For now this is true, but if this set
is going to be extended in the future, your approach most likely will
lead to an error.

-- 
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?

[-- Attachment #1.2: signature.asc --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

[-- Attachment #2: Type: text/plain, Size: 126 bytes --]

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