2021-04-21 03:27:32

by Haiwei Li

[permalink] [raw]
Subject: [PATCH] KVM: x86: Take advantage of kvm_arch_dy_has_pending_interrupt()

From: Haiwei Li <[email protected]>

`kvm_arch_dy_runnable` checks the pending_interrupt as the code in
`kvm_arch_dy_has_pending_interrupt`. So take advantage of it.

Signed-off-by: Haiwei Li <[email protected]>
---
arch/x86/kvm/x86.c | 21 +++++++++------------
1 file changed, 9 insertions(+), 12 deletions(-)

diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index d696a9f..08bd616 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -11125,28 +11125,25 @@ int kvm_arch_vcpu_runnable(struct kvm_vcpu *vcpu)
return kvm_vcpu_running(vcpu) || kvm_vcpu_has_events(vcpu);
}

-bool kvm_arch_dy_runnable(struct kvm_vcpu *vcpu)
+bool kvm_arch_dy_has_pending_interrupt(struct kvm_vcpu *vcpu)
{
- if (READ_ONCE(vcpu->arch.pv.pv_unhalted))
- return true;
-
- if (kvm_test_request(KVM_REQ_NMI, vcpu) ||
- kvm_test_request(KVM_REQ_SMI, vcpu) ||
- kvm_test_request(KVM_REQ_EVENT, vcpu))
- return true;
-
if (vcpu->arch.apicv_active && static_call(kvm_x86_dy_apicv_has_pending_interrupt)(vcpu))
return true;

return false;
}

-bool kvm_arch_dy_has_pending_interrupt(struct kvm_vcpu *vcpu)
+bool kvm_arch_dy_runnable(struct kvm_vcpu *vcpu)
{
- if (vcpu->arch.apicv_active && static_call(kvm_x86_dy_apicv_has_pending_interrupt)(vcpu))
+ if (READ_ONCE(vcpu->arch.pv.pv_unhalted))
return true;

- return false;
+ if (kvm_test_request(KVM_REQ_NMI, vcpu) ||
+ kvm_test_request(KVM_REQ_SMI, vcpu) ||
+ kvm_test_request(KVM_REQ_EVENT, vcpu))
+ return true;
+
+ return kvm_arch_dy_has_pending_interrupt(vcpu);
}

bool kvm_arch_vcpu_in_kernel(struct kvm_vcpu *vcpu)
--
1.8.3.1


2021-04-22 09:23:34

by Wanpeng Li

[permalink] [raw]
Subject: Re: [PATCH] KVM: x86: Take advantage of kvm_arch_dy_has_pending_interrupt()

On Wed, 21 Apr 2021 at 11:26, <[email protected]> wrote:
>
> From: Haiwei Li <[email protected]>
>
> `kvm_arch_dy_runnable` checks the pending_interrupt as the code in
> `kvm_arch_dy_has_pending_interrupt`. So take advantage of it.
>
> Signed-off-by: Haiwei Li <[email protected]>

Reviewed-by: Wanpeng Li <[email protected]>

2021-04-22 10:38:20

by Paolo Bonzini

[permalink] [raw]
Subject: Re: [PATCH] KVM: x86: Take advantage of kvm_arch_dy_has_pending_interrupt()

On 21/04/21 05:25, [email protected] wrote:
> From: Haiwei Li <[email protected]>
>
> `kvm_arch_dy_runnable` checks the pending_interrupt as the code in
> `kvm_arch_dy_has_pending_interrupt`. So take advantage of it.
>
> Signed-off-by: Haiwei Li <[email protected]>
> ---
> arch/x86/kvm/x86.c | 21 +++++++++------------
> 1 file changed, 9 insertions(+), 12 deletions(-)
>
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index d696a9f..08bd616 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -11125,28 +11125,25 @@ int kvm_arch_vcpu_runnable(struct kvm_vcpu *vcpu)
> return kvm_vcpu_running(vcpu) || kvm_vcpu_has_events(vcpu);
> }
>
> -bool kvm_arch_dy_runnable(struct kvm_vcpu *vcpu)
> +bool kvm_arch_dy_has_pending_interrupt(struct kvm_vcpu *vcpu)
> {
> - if (READ_ONCE(vcpu->arch.pv.pv_unhalted))
> - return true;
> -
> - if (kvm_test_request(KVM_REQ_NMI, vcpu) ||
> - kvm_test_request(KVM_REQ_SMI, vcpu) ||
> - kvm_test_request(KVM_REQ_EVENT, vcpu))
> - return true;
> -
> if (vcpu->arch.apicv_active && static_call(kvm_x86_dy_apicv_has_pending_interrupt)(vcpu))
> return true;
>
> return false;
> }
>
> -bool kvm_arch_dy_has_pending_interrupt(struct kvm_vcpu *vcpu)
> +bool kvm_arch_dy_runnable(struct kvm_vcpu *vcpu)
> {
> - if (vcpu->arch.apicv_active && static_call(kvm_x86_dy_apicv_has_pending_interrupt)(vcpu))
> + if (READ_ONCE(vcpu->arch.pv.pv_unhalted))
> return true;
>
> - return false;
> + if (kvm_test_request(KVM_REQ_NMI, vcpu) ||
> + kvm_test_request(KVM_REQ_SMI, vcpu) ||
> + kvm_test_request(KVM_REQ_EVENT, vcpu))
> + return true;
> +
> + return kvm_arch_dy_has_pending_interrupt(vcpu);
> }
>
> bool kvm_arch_vcpu_in_kernel(struct kvm_vcpu *vcpu)
>

Looks good, but I'd like to take a look at the other patches for
directed yield first. Thanks!

Paolo