2018-02-11 02:05:23

by Chao Gao

Subject: [PATCH v3] KVM: VMX: Don't halt vcpu when L1 is injecting events to L2

Although L2 is in halt state, it will be in the active state after
VM entry if the VM entry is vectoring according to SDM 26.6.2 Activity
State. Halting the vcpu here means the event won't be injected to L2
and this decision isn't reported to L1. Thus L0 drops an event that
should be injected to L2.

Cc: Liran Alon <[email protected]>
Signed-off-by: Chao Gao <[email protected]>
---
changes in v3:
- s/VECTORING_INFO_VALID_MASK/INTR_INFO_VALID_MASK
- Changes of subject, comments.

Changes in v2:
- Remove VID stuff. Only handle event injection in this patch.

Signed-off-by: Chao Gao <[email protected]>
---
arch/x86/kvm/vmx.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index bb5b488..cf7c2fc 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -10985,7 +10985,12 @@ static int nested_vmx_run(struct kvm_vcpu *vcpu, bool launch)
if (ret)
return ret;

- if (vmcs12->guest_activity_state == GUEST_ACTIVITY_HLT)
+ /*
+ * If we're entering a halted L2 vcpu and the L2 vcpu won't be woken
+ * by event injection, halt vcpu.
+ */
+ if ((vmcs12->guest_activity_state == GUEST_ACTIVITY_HLT) &&
+ !(vmcs12->vm_entry_intr_info_field & INTR_INFO_VALID_MASK))
return kvm_vcpu_halt(vcpu);

vmx->nested.nested_run_pending = 1;
--
1.9.1

2018-02-11 08:52:28

by Liran Alon

Subject: Re: [PATCH v3] KVM: VMX: Don't halt vcpu when L1 is injecting events to L2

----- [email protected] wrote:

> Although L2 is in halt state, it will be in the active state after
> VM entry if the VM entry is vectoring according to SDM 26.6.2
> Activity
> State. Halting the vcpu here means the event won't be injected to L2
> and this decision isn't reported to L1. Thus L0 drops an event that
> should be injected to L2.
>
> Cc: Liran Alon <[email protected]>
> Signed-off-by: Chao Gao <[email protected]>
> ---
> changes in v3:
> - s/VECTORING_INFO_VALID_MASK/INTR_INFO_VALID_MASK
> - Changes of subject, comments.
>
> Changes in v2:
> - Remove VID stuff. Only handle event injection in this patch.
>
> Signed-off-by: Chao Gao <[email protected]>
> ---
> arch/x86/kvm/vmx.c | 7 ++++++-
> 1 file changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index bb5b488..cf7c2fc 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -10985,7 +10985,12 @@ static int nested_vmx_run(struct kvm_vcpu
> *vcpu, bool launch)
> if (ret)
> return ret;
>
> - if (vmcs12->guest_activity_state == GUEST_ACTIVITY_HLT)
> + /*
> + * If we're entering a halted L2 vcpu and the L2 vcpu won't be
> woken
> + * by event injection, halt vcpu.
> + */
> + if ((vmcs12->guest_activity_state == GUEST_ACTIVITY_HLT) &&
> + !(vmcs12->vm_entry_intr_info_field & INTR_INFO_VALID_MASK))
> return kvm_vcpu_halt(vcpu);
>
> vmx->nested.nested_run_pending = 1;
> --
> 1.9.1

Reviewed-by: Liran Alon <[email protected]>

(Paolo: I would recommend changing commit title from "KVM: VMX: ..." to "KVM: nVMX: ..."
when applying for convention. Easier to grep nVMX changes.)

2018-02-13 14:41:16

by Paolo Bonzini

Subject: Re: [PATCH v3] KVM: VMX: Don't halt vcpu when L1 is injecting events to L2

On 11/02/2018 03:06, Chao Gao wrote:
> Although L2 is in halt state, it will be in the active state after
> VM entry if the VM entry is vectoring according to SDM 26.6.2 Activity
> State. Halting the vcpu here means the event won't be injected to L2
> and this decision isn't reported to L1. Thus L0 drops an event that
> should be injected to L2.
>
> Cc: Liran Alon <[email protected]>
> Signed-off-by: Chao Gao <[email protected]>
> ---
> changes in v3:
> - s/VECTORING_INFO_VALID_MASK/INTR_INFO_VALID_MASK
> - Changes of subject, comments.
>
> Changes in v2:
> - Remove VID stuff. Only handle event injection in this patch.
>
> Signed-off-by: Chao Gao <[email protected]>
> ---
> arch/x86/kvm/vmx.c | 7 ++++++-
> 1 file changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index bb5b488..cf7c2fc 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -10985,7 +10985,12 @@ static int nested_vmx_run(struct kvm_vcpu *vcpu, bool launch)
> if (ret)
> return ret;
>
> - if (vmcs12->guest_activity_state == GUEST_ACTIVITY_HLT)
> + /*
> + * If we're entering a halted L2 vcpu and the L2 vcpu won't be woken
> + * by event injection, halt vcpu.
> + */
> + if ((vmcs12->guest_activity_state == GUEST_ACTIVITY_HLT) &&
> + !(vmcs12->vm_entry_intr_info_field & INTR_INFO_VALID_MASK))
> return kvm_vcpu_halt(vcpu);
>
> vmx->nested.nested_run_pending = 1;
>

Queued, thanks (haven't yet tried the new test).

Paolo