[PATCH v2 06/11] KVM: arm64: Use vcpu_idx for invalidation tracking

[Marc Zyngier]

On Thu, 21 Sep 2023 10:16:42 +0100,
Zenghui Yu <yuzenghui at huawei.com> wrote:
> 
> On 2023/9/21 2:17, Marc Zyngier wrote:
> > Signed-off-by: Marc Zyngier <maz at kernel.org>
> > ---
> >  arch/arm64/kvm/arm.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c
> > index 872679a0cbd7..23c22dbd1969 100644
> > --- a/arch/arm64/kvm/arm.c
> > +++ b/arch/arm64/kvm/arm.c
> > @@ -438,9 +438,9 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
> >  	 * We might get preempted before the vCPU actually runs, but
> >  	 * over-invalidation doesn't affect correctness.
> >  	 */
> > -	if (*last_ran != vcpu->vcpu_id) {
> > +	if (*last_ran != vcpu->vcpu_idx) {
> >  		kvm_call_hyp(__kvm_flush_cpu_context, mmu);
> > -		*last_ran = vcpu->vcpu_id;
> > +		*last_ran = vcpu->vcpu_idx;
> >  	}
> >   	vcpu->cpu = cpu;
> 
> Isn't the original code (using vcpu_id) enough to detect a different
> previously run VCPU? What am I missing?

It is in theory enough. However, I couldn't convince myself of the
*unicity* of the vcpu_id field. It really feels like something as
crucial as the TLB invalidation shouldn't rely on something that is
controlled by userspace.

And I really should write a commit message to capture this.

Thanks,

	M.

-- 
Without deviation from the norm, progress is not possible.