On Wed, Oct 19, 2022 at 04:00:06PM +0000, Quentin Perret wrote:
> On Monday 17 Oct 2022 at 12:51:57 (+0100), Will Deacon wrote:
> > +struct kvm_protected_vm {
> > + pkvm_handle_t handle;
> > + struct mutex vm_lock;
>
> Why is this lock needed btw? Isn't kvm->lock good enough?
Good question. I can't see why kvm->lock wouldn't work for this series,
so I'll drop this for now. I have vague recollections that we ran into
lock ordering issues in an early implementation wrt vcpu locking, but
for now let's drop it.
Cheers,
Will
