Re: [Xen-devel] [RFC] x86: PV SMAP for 64-bit guests

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 29.01.14 at 19:00, David Vrabel <david.vrabel@xxxxxxxxxx> wrote:
> Each copy_from_user() and copy_to_user() and get_user()/put_user() would
> thus require two hypercalls, at least one of which would do a TLB flush?

Right.

> This does sound rather expensive and thus not something we (XenServer)
> would be especially interested in using.

And I'd consider this only of interest for the very security conscious.
The plan would certainly be to have this default disabled in Linux.

> Do you have any figures for the performance impact on guests not using
> this feature?

Guests not using this feature would only suffer from the extra
instructions added to Xen's entry.S paths. I'm not sure that would
be directly measurable in other than micro benchmarks, but as said
I'm still somewhat concerned - I especially wouldn't want to add
such if people don't think this is at least of reasonable use in some
environments.

Jan


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