Re: [PATCH 2/2] Interrupt unhalts vcpu when it shouldn't

[Avi Kivity <avi@redhat.com>]

Gleb Natapov wrote:
> kvm_vcpu_block() unhalts vpu on an interrupt/timer without checking
> if interrupt window is actually opened.
>
>  
> +static int svm_interrupt_allowed(struct kvm_vcpu *vcpu)
> +{
> +	struct vcpu_svm *svm = to_svm(vcpu);
> +	struct vmcb *vmcb = svm->vmcb;
> +	return (vmcb->save.rflags & X86_EFLAGS_IF) && 
> +		!(vmcb->control.int_state & SVM_INTERRUPT_SHADOW_MASK) &&
> +		(svm->vcpu.arch.hflags & HF_GIF_MASK);
> +}
> +
>  
> +static int vmx_interrupt_allowed(struct kvm_vcpu *vcpu)
> +{
> +	vmx_update_window_states(vcpu);
> +	return vcpu->arch.interrupt_window_open;
> +}
> +
>  static void kvm_do_inject_irq(struct kvm_vcpu *vcpu)
>  }
> +
> +int kvm_arch_interrupt_allowed(struct kvm_vcpu *vcpu)
> +{
> +	return kvm_x86_ops->interrupt_allowed(vcpu);
> +}
>   

If the guest enables interrupts but sets tpr/cr8 to block interrupts, 
we'll spin (like we do now).

So I think this should be called kvm_arch_can_accept_interrupt() and 
take tpr into account.

-- 
error compiling committee.c: too many arguments to function