On Fri, Apr 06, 2018 at 10:40:23AM +0100, Andrew Cooper wrote:
> On 06/04/18 10:36, Wei Liu wrote:
> > Signed-off-by: Wei Liu <wei.l...@citrix.com>
> For what purpose?  This is very deliberate that, if faulting is
> available, levelling never gets touched.
> Levelling, as a mechanism, is strictly inferior to faulting.  On Intel,
> the sets of hardware with levelling and faulting are disjoint.

AFAICT levelling_caps is changed by probe_cpuid_faulting. That's what
I'm getting at.


Xen-devel mailing list

Reply via email to