[Xen-devel] Re: [PATCH 2/2] xen/gnt{dev, alloc}: reserve event channels for notify

[Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>]

> That same SELinux category-based isolation mechanism is also a good solution 
> for xen
> qemu-dm processes, although moving qemu to a stubdom provides better 
> isolation since
> SELinux currently cannot talk to XSM to determine what domains a particular 
> qemu-dm 
> process should be able to manipulate.

<nods>
> 
> Only allowing event channels allocated by userspace to be used in gnt* notify 
> is
> a good idea, since there's no reason for userspace to need to manipulate an 
> event
> channel set up by the kernel.
> 
.. snip..
> > 
> > How would that work when the IRQ subsystem (so everything is setup in the 
> > kernel)
> > gets an event? Would the refcount be for that -1.. oh. You would only set
> > the refcnt when the _get/_put calls are made and not when in-kernel calls 
> > to setup
> > IRQs are done?
> > 
> 
> Right. The reference count would be a dual-purpose field indicating if the 
> event
> channel is kernel-internal (value -1) or userspace-visible (reference count > 
> 0).
> New event channels would start out at -1, and evtchn.c would change them to 1.

The tricky bit is going to be with the xen_free_irq which might have to deal 
with kernel
events and grantdev events... oh wait, the event_put is going to decrement it 
and then
call xen_free_irq, so that will come out to be the right number.

Looking forward to the patches! Thanks for doing this work.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel