Re: [Xen-devel] [XenARM] XEN tools for ARM with Virtualization Extensions

[Ian Campbell <ian.campbell@xxxxxxxxxx>]

On Thu, 2013-07-11 at 16:46 +0000, Eric Trudeau wrote:
> > > >> > What functions should I call to implement  XEN_DOMCTL_bind_pt_irq on
> > > >> ARM?
> > > >>
> > > >> There's a function like route_irq_to_guest which we use to route IRQs 
> > > >> to
> > > >> dom0 during boot. In principal that could also be used to reroute an 
> > > >> IRQ
> > > >> to a guest, but I'm not sure how it will interact with the 
> > > >> reassignment,
> > > >> since in your case the IRQ starts off bound to dom0. Hopefully it's 
> > > >> just
> > > >> a small change to make it work for this case.
> > > >>
> 
> I have IRQs for devices being passed to ARM DomU guests now.  Thanks for your
> help.

Excellent! Do you have any useful patches?

> The IRQ mapping needs to be unmapped when the domU is destroyed and it is not
> happening presently.  I don't see a call to PHYSDEVOP_unmap_pirq or anything.
> Should the mapped IRQs for a guest be unmapped when Xen destroys the domain
> based on the IRQ rangeset?  Where do you suggest doing this?
> Or should the tools call an xc_physdev_unmap_pirq() when destroying the 
> domain?
> I would think it should be implicit in Hypervisor code to relinquish all 
> resources of a
> domain upon destruction...

Having the hypervisor cleanup on destroy sounds sensible to me, x86
seems to call free_domain_pirqs from arch_domain_destroy.

I think the unmap_pirq path is for hotunplug


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