Re: [PATCH] KVM: x86: reset RVI upon system reset

["Chen, Tiejun" <tiejun.chen@intel.com>]

On 2014/11/5 10:53, Wei Wang wrote:
> A bug was reported as follows: when running Windows 7 32-bit guests on qemu-kvm,
> sometimes the guests run into blue screen during reboot. The problem was that a
> guest's RVI was not cleared when it rebooted. This patch has fixed the problem.
>
> Signed-off-by: Wei Wang <wei.w.wang@intel.com>
> Signed-off-by: Yang Zhang <yang.z.zhang@intel.com>
> Tested-by: Rongrong Liu <rongrongx.liu@intel.com>, Da Chun <ngugc@qq.com>
> ---
>   arch/x86/kvm/lapic.c |    3 +++
>   arch/x86/kvm/vmx.c   |   12 ++++++------
>   2 files changed, 9 insertions(+), 6 deletions(-)
>
> diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
> index 66dd173..6942742 100644
> --- a/arch/x86/kvm/lapic.c
> +++ b/arch/x86/kvm/lapic.c
> @@ -1712,6 +1712,9 @@ void kvm_apic_post_state_restore(struct kvm_vcpu *vcpu,
>   	apic->isr_count = kvm_apic_vid_enabled(vcpu->kvm) ?
>   				1 : count_vectors(apic->regs + APIC_ISR);
>   	apic->highest_isr_cache = -1;
> +	if (kvm_x86_ops->hwapic_irr_update)
> +		kvm_x86_ops->hwapic_irr_update(vcpu,
> +				apic_find_highest_irr(apic));

Could we pass 0 directly? Because looks we just need to clear RVI.

kvm_x86_ops->hwapic_irr_update(vcpu, 0);

I think this already makes sense based on your description.

Thanks
Tiejun

>   	kvm_x86_ops->hwapic_isr_update(vcpu->kvm, apic_find_highest_isr(apic));
>   	kvm_make_request(KVM_REQ_EVENT, vcpu);
>   	kvm_rtc_eoi_tracking_restore_one(vcpu);
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index fe4d2f4..d632548 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -7292,19 +7292,19 @@ static void vmx_set_rvi(int vector)
>   static void vmx_hwapic_irr_update(struct kvm_vcpu *vcpu, int max_irr)
>   {
>   	if (max_irr == -1)
> +		max_irr = 0;
> +
> +	if (!is_guest_mode(vcpu)) {
> +		vmx_set_rvi(max_irr);
>   		return;
> +	}
>
>   	/*
>   	 * If a vmexit is needed, vmx_check_nested_events handles it.
>   	 */
> -	if (is_guest_mode(vcpu) && nested_exit_on_intr(vcpu))
> +	if ((is_guest_mode(vcpu) && nested_exit_on_intr(vcpu)) || max_irr == 0)
>   		return;
>
> -	if (!is_guest_mode(vcpu)) {
> -		vmx_set_rvi(max_irr);
> -		return;
> -	}
> -
>   	/*
>   	 * Fall back to pre-APICv interrupt injection since L2
>   	 * is run without virtual interrupt delivery.
>