Re: [PATCH v3 4/4] KVM: TDX: Enable Bus Lock VM exit

[Sean Christopherson <seanjc@google.com>]

On Thu, Aug 13, 2026, Xiaoyao Li wrote:
> On 8/13/2026 6:58 AM, Edgecombe, Rick P wrote:
> > On Wed, 2026-08-12 at 16:02 +0800, Xiaoyao Li wrote:
> > > -	if (unlikely(READ_ONCE(to_kvm_tdx(vcpu->kvm)->wait_for_sept_zap)))
> > > +	if (unlikely(READ_ONCE(to_kvm_tdx(vcpu->kvm)->wait_for_sept_zap))) {
> > > +		vt->exit_reason.bus_lock_detected = 0;
> > >   		return EXIT_FASTPATH_EXIT_HANDLED;
> > > +	}
> > 
> > This still feels hacky to me. It at least deserves a comment I think. If you
> > spin another version.
> 
> I spent more time on this today and I find there is an existing issue. When
> the previous Exit Reason is EXIT_REASON_EXTERNAL_INTERRUPT, the early return
> here can go to the path due to the stale exit_reason.
> 
> kvm_x86_call(handle_exit_irqoff)(vcpu)
>   vmx_handle_exit_irqoff()
>     handle_external_interrupt_irqoff()
> 
> This makes the host process the external interrupt twice, and maybe more
> times if the wait_for_sept_zap remains longer.
> 
> I think we need an separate fix to set
> 
> 	vt->exit_reasons.full = TDX_INVALID_EXIT_REASON;
> 
> so that if some patch is going to consume the stale Exit Reason, it can be
> caught by TDX_INVALID_EXIT_REASON.

The more I look at this, the more I'm against shoving garbage into vt->exit_reasons.
With tdx_is_exit_reason_valid(), this is trivially easy to handle, *and* explicitly
captures the logic instead of subtly rerouting KVM away from meaningful handling.

diff --git a/arch/x86/kvm/vmx/main.c b/arch/x86/kvm/vmx/main.c
index 95d89d809c19..a1f5b5dc1fa3 100644
--- a/arch/x86/kvm/vmx/main.c
+++ b/arch/x86/kvm/vmx/main.c
@@ -157,6 +157,14 @@ static fastpath_t vt_vcpu_run(struct kvm_vcpu *vcpu, u64 run_flags)
        return vmx_vcpu_run(vcpu, run_flags);
 }
 
+static void vt_handle_exit_irqoff(struct kvm_vcpu *vcpu)
+{
+       if (is_td_vcpu(vcpu) && !tdx_is_exit_reason_valid(vcpu))
+               return;
+
+       vmx_handle_exit_irqoff();
+}
+
 static int vt_handle_exit(struct kvm_vcpu *vcpu,
                          enum exit_fastpath_completion fastpath)
 {
@@ -1071,7 +1079,7 @@ struct kvm_x86_ops vt_x86_ops __initdata = {
        .load_mmu_pgd = vt_op(load_mmu_pgd),
 
        .check_intercept = vmx_check_intercept,
-       .handle_exit_irqoff = vmx_handle_exit_irqoff,
+       .handle_exit_irqoff = vt_op(handle_exit_irqoff),
 
        .update_cpu_dirty_logging = vt_op(update_cpu_dirty_logging),