2022-08-19 18:35:27

by Paolo Bonzini

Subject: [PATCH] KVM: SVM: remove unnecessary check on INIT intercept

Since svm_check_nested_events() is now handling INIT signals, there is
no need to latch it until the VMEXIT is injected. The only condition
under which INIT signals are latched is GIF=0.

Suggested-by: Maxim Levitsky <[email protected]>
Signed-off-by: Paolo Bonzini <[email protected]>
---
arch/x86/kvm/svm/svm.c | 10 +---------
1 file changed, 1 insertion(+), 9 deletions(-)

diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
index f3813dbacb9f..26a348389ece 100644
--- a/arch/x86/kvm/svm/svm.c
+++ b/arch/x86/kvm/svm/svm.c
@@ -4697,15 +4697,7 @@ static bool svm_apic_init_signal_blocked(struct kvm_vcpu *vcpu)
{
struct vcpu_svm *svm = to_svm(vcpu);

- /*
- * TODO: Last condition latch INIT signals on vCPU when
- * vCPU is in guest-mode and vmcb12 defines intercept on INIT.
- * To properly emulate the INIT intercept,
- * svm_check_nested_events() should call nested_svm_vmexit()
- * if an INIT signal is pending.
- */
- return !gif_set(svm) ||
- (vmcb_is_intercept(&svm->vmcb->control, INTERCEPT_INIT));
+ return !gif_set(svm);
}

static void svm_vcpu_deliver_sipi_vector(struct kvm_vcpu *vcpu, u8 vector)
--
2.31.1

2022-08-21 09:49:46

by Maxim Levitsky

Subject: Re: [PATCH] KVM: SVM: remove unnecessary check on INIT intercept

On Fri, 2022-08-19 at 12:56 -0400, Paolo Bonzini wrote:
> Since svm_check_nested_events() is now handling INIT signals, there is
> no need to latch it until the VMEXIT is injected.  The only condition
> under which INIT signals are latched is GIF=0.
>
> Suggested-by: Maxim Levitsky <[email protected]>
> Signed-off-by: Paolo Bonzini <[email protected]>
> ---
>  arch/x86/kvm/svm/svm.c | 10 +---------
>  1 file changed, 1 insertion(+), 9 deletions(-)
>
> diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
> index f3813dbacb9f..26a348389ece 100644
> --- a/arch/x86/kvm/svm/svm.c
> +++ b/arch/x86/kvm/svm/svm.c
> @@ -4697,15 +4697,7 @@ static bool svm_apic_init_signal_blocked(struct kvm_vcpu *vcpu)
>  {
>         struct vcpu_svm *svm = to_svm(vcpu);
>  
> -       /*
> -        * TODO: Last condition latch INIT signals on vCPU when
> -        * vCPU is in guest-mode and vmcb12 defines intercept on INIT.
> -        * To properly emulate the INIT intercept,
> -        * svm_check_nested_events() should call nested_svm_vmexit()
> -        * if an INIT signal is pending.
> -        */
> -       return !gif_set(svm) ||
> -                  (vmcb_is_intercept(&svm->vmcb->control, INTERCEPT_INIT));
> +       return !gif_set(svm);
>  }
>  
>  static void svm_vcpu_deliver_sipi_vector(struct kvm_vcpu *vcpu, u8 vector)

Reviewed-by: Maxim Levitsky <[email protected]>

Best regards,
Maxim Levitsky

2022-08-30 22:10:21

by Sean Christopherson

Subject: Re: [PATCH] KVM: SVM: remove unnecessary check on INIT intercept

On Fri, Aug 19, 2022, Paolo Bonzini wrote:
> Since svm_check_nested_events() is now handling INIT signals, there is
> no need to latch it until the VMEXIT is injected. The only condition
> under which INIT signals are latched is GIF=0.
>
> Suggested-by: Maxim Levitsky <[email protected]>
> Signed-off-by: Paolo Bonzini <[email protected]>
> ---

Pushed to branch `for_paolo/6.1` at:

https://github.com/sean-jc/linux.git

Unless you hear otherwise, it will make its way to kvm/queue "soon", thus completing
the circle.

Note, the commit IDs are not guaranteed to be stable.