Re: [RFC 00/10] Process-local memory allocations for hiding KVM secrets

From: Andy Lutomirski
Date: Mon Jun 17 2019 - 11:59:48 EST

On Mon, Jun 17, 2019 at 8:50 AM Dave Hansen <dave.hansen@xxxxxxxxx> wrote:
> On 6/17/19 12:38 AM, Alexander Graf wrote:
> >> Yes I know, but as a benefit we could get rid of all the GSBASE
> >> horrors in
> >> the entry code as we could just put the percpu space into the local PGD.
> >
> > Would that mean that with Meltdown affected CPUs we open speculation
> > attacks against the mmlocal memory from KVM user space?
> Not necessarily. There would likely be a _set_ of local PGDs. We could
> still have pair of PTI PGDs just like we do know, they'd just be a local
> PGD pair.

Unfortunately, this would mean that we need to sync twice as many
top-level entries when we context switch.