[PATCH] KVM: x86: Add EOI exit bitmap handlers for Hyper-V SynIC vectors

[PATCH] KVM: x86: Add EOI exit bitmap handlers for Hyper-V SynIC vectors

[Wang Guangju]

From: wangguangju <wangguangju@baidu.com>

Hyper-V SynIC vectors were added into EOI exit bitmap in func
synic_set_sint().But when the Windows VM VMEXIT due to
EXIT_REASON_EOI_INDUCED, there are no EOI exit bitmap handlers
for Hyper-V SynIC vectors.

This patch fix it.

Change-Id: I2404ebf7bda60326be3f6786e0e34e63aa81bbd4
Signed-off-by: wangguangju <wangguangju@baidu.com>
---
 arch/x86/kvm/lapic.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
index 0e68b4c..59096f8 100644
--- a/arch/x86/kvm/lapic.c
+++ b/arch/x86/kvm/lapic.c
@@ -1303,6 +1303,10 @@ void kvm_apic_set_eoi_accelerated(struct kvm_vcpu *vcpu, int vector)
 
 	trace_kvm_eoi(apic, vector);
 
+	if (to_hv_vcpu(apic->vcpu) &&
+	    test_bit(vector, to_hv_synic(apic->vcpu)->vec_bitmap))
+		kvm_hv_synic_send_eoi(apic->vcpu, vector);
+
 	kvm_ioapic_send_eoi(apic, vector);
 	kvm_make_request(KVM_REQ_EVENT, apic->vcpu);
 }
-- 
2.9.4

Re: [PATCH] KVM: x86: Add EOI exit bitmap handlers for Hyper-V SynIC vectors

[Vitaly Kuznetsov]

Wang Guangju <wangguangju@baidu.com> writes:

> From: wangguangju <wangguangju@baidu.com>
>
> Hyper-V SynIC vectors were added into EOI exit bitmap in func
> synic_set_sint().But when the Windows VM VMEXIT due to
> EXIT_REASON_EOI_INDUCED, there are no EOI exit bitmap handlers
> for Hyper-V SynIC vectors.

My take:

"When EOI virtualization is performed on VMX,
kvm_apic_set_eoi_accelerated() is called upon EXIT_REASON_EOI_INDUCED
but unlike its non-accelerated apic_set_eoi() sibling, Hyper-V SINT
vectors are left unhandled.
"

>
> This patch fix it.
>
> Change-Id: I2404ebf7bda60326be3f6786e0e34e63aa81bbd4

In case this is not something publicly available it doesn't belong to
kernel changelog as it doesn't bring any value.

> Signed-off-by: wangguangju <wangguangju@baidu.com>
> ---
>  arch/x86/kvm/lapic.c | 4 ++++
>  1 file changed, 4 insertions(+)
>
> diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
> index 0e68b4c..59096f8 100644
> --- a/arch/x86/kvm/lapic.c
> +++ b/arch/x86/kvm/lapic.c
> @@ -1303,6 +1303,10 @@ void kvm_apic_set_eoi_accelerated(struct kvm_vcpu *vcpu, int vector)
>  
>  	trace_kvm_eoi(apic, vector);
>  
> +	if (to_hv_vcpu(apic->vcpu) &&
> +	    test_bit(vector, to_hv_synic(apic->vcpu)->vec_bitmap))
> +		kvm_hv_synic_send_eoi(apic->vcpu, vector);
> +
>  	kvm_ioapic_send_eoi(apic, vector);
>  	kvm_make_request(KVM_REQ_EVENT, apic->vcpu);


This whole part:

	if (to_hv_vcpu(apic->vcpu) &&
	    test_bit(vector, to_hv_synic(apic->vcpu)->vec_bitmap))
		kvm_hv_synic_send_eoi(apic->vcpu, vector);

	kvm_ioapic_send_eoi(apic, vector);
	kvm_make_request(KVM_REQ_EVENT, apic->vcpu);

could be split into an inline function, something like (completely
untested):

diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
index 6ff17d5a2ae3..9d19c7c738c0 100644
--- a/arch/x86/kvm/lapic.c
+++ b/arch/x86/kvm/lapic.c
@@ -1269,6 +1269,16 @@ static void kvm_ioapic_send_eoi(struct kvm_lapic *apic, int vector)
        kvm_ioapic_update_eoi(apic->vcpu, vector, trigger_mode);
 }
 
+static inline void apic_set_eoi_vector(struct kvm_lapic *apic, int vector)
+{
+       if (to_hv_vcpu(apic->vcpu) &&
+           test_bit(vector, to_hv_synic(apic->vcpu)->vec_bitmap))
+               kvm_hv_synic_send_eoi(apic->vcpu, vector);
+
+       kvm_ioapic_send_eoi(apic, vector);
+       kvm_make_request(KVM_REQ_EVENT, apic->vcpu);
+}
+
 static int apic_set_eoi(struct kvm_lapic *apic)
 {
        int vector = apic_find_highest_isr(apic);
@@ -1285,12 +1295,8 @@ static int apic_set_eoi(struct kvm_lapic *apic)
        apic_clear_isr(vector, apic);
        apic_update_ppr(apic);
 
-       if (to_hv_vcpu(apic->vcpu) &&
-           test_bit(vector, to_hv_synic(apic->vcpu)->vec_bitmap))
-               kvm_hv_synic_send_eoi(apic->vcpu, vector);
+       apic_set_eoi_vector(apic, vector);
 
-       kvm_ioapic_send_eoi(apic, vector);
-       kvm_make_request(KVM_REQ_EVENT, apic->vcpu);
        return vector;
 }
 
@@ -1304,8 +1310,7 @@ void kvm_apic_set_eoi_accelerated(struct kvm_vcpu *vcpu, int vector)
 
        trace_kvm_eoi(apic, vector);
 
-       kvm_ioapic_send_eoi(apic, vector);
-       kvm_make_request(KVM_REQ_EVENT, apic->vcpu);
+       apic_set_eoi_vector(apic, vector);
 }
 EXPORT_SYMBOL_GPL(kvm_apic_set_eoi_accelerated);


>  }

-- 
Vitaly

答复: [PATCH] KVM: x86: Add EOI exit bitmap handlers for Hyper-V SynIC vectors

[Wang,Guangju]

> This whole part:
>
> 	if (to_hv_vcpu(apic->vcpu) &&
>	    test_bit(vector, to_hv_synic(apic->vcpu)->vec_bitmap))
>		kvm_hv_synic_send_eoi(apic->vcpu, vector);
>
>	kvm_ioapic_send_eoi(apic, vector);
>	kvm_make_request(KVM_REQ_EVENT, apic->vcpu);

> could be split into an inline function, something like (completely
> untested):

Thank you, I will modify and test it and then send a new patch.

--
Wang