On Wed, Jun 8, 2022 at 12:40 AM Fuad Tabba <tabba@xxxxxxxxxx> wrote:
>
> Hi Peter,
>
> On Tue, Jun 7, 2022 at 1:42 AM Peter Collingbourne <pcc@xxxxxxxxxx> wrote:
> >
> > On Thu, May 19, 2022 at 7:40 AM Will Deacon <will@xxxxxxxxxx> wrote:
> > >
> > > From: Fuad Tabba <tabba@xxxxxxxxxx>
> > >
> > > Return an error (-EINVAL) if trying to enable MTE on a protected
> > > vm.
> > >
> > > Signed-off-by: Fuad Tabba <tabba@xxxxxxxxxx>
> > > ---
> > > arch/arm64/kvm/arm.c | 4 +++-
> > > 1 file changed, 3 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
> > > index 10e036bf06e3..8a1b4ba1dfa7 100644
> > > --- a/arch/arm64/kvm/arm.c
> > > +++ b/arch/arm64/kvm/arm.c
> > > @@ -90,7 +90,9 @@ int kvm_vm_ioctl_enable_cap(struct kvm *kvm,
> > > break;
> > > case KVM_CAP_ARM_MTE:
> > > mutex_lock(&kvm->lock);
> > > - if (!system_supports_mte() || kvm->created_vcpus) {
> > > + if (!system_supports_mte() ||
> > > + kvm_vm_is_protected(kvm) ||
> >
> > Should this check be added to kvm_vm_ioctl_check_extension() as well?
>
> No need. kvm_vm_ioctl_check_extension() calls pkvm_check_extension()
> for protected vms, which functions as an allow list rather than a
> block list.
I see. I guess I got confused when reading the code because I saw this
in kvm_check_extension():
case KVM_CAP_ARM_NISV_TO_USER:
r = !kvm || !kvm_vm_is_protected(kvm);
break;
This can probably be simplified to "r = 1;".
Peter
