Re: [PATCH 2/4] KVM: VMX: prepare sync_pir_to_irr for running with APICv disabled

[David Matlack]

On Mon, Nov 22, 2021 at 07:43:09PM -0500, Paolo Bonzini wrote:
> If APICv is disabled for this vCPU, assigned devices may
> still attempt to post interrupts.  In that case, we need
> to cancel the vmentry and deliver the interrupt with
> KVM_REQ_EVENT.  Extend the existing code that handles
> injection of L1 interrupts into L2 to cover this case
> as well.
> 
> vmx_hwapic_irr_update is only called when APICv is active
> so it would be confusing to add a check for
> vcpu->arch.apicv_active in there.  Instead, just use
> vmx_set_rvi directly in vmx_sync_pir_to_irr.
> 
> Cc: stable@xxxxxxxxxxxxxxx
> Signed-off-by: Paolo Bonzini <pbonzini@xxxxxxxxxx>

Reviewed-by: David Matlack <dmatlack@xxxxxxxxxx>

(Although I agree with Sean's suggestions and 1 nit below.)

> ---
>  arch/x86/kvm/vmx/vmx.c | 35 +++++++++++++++++++++++------------
>  1 file changed, 23 insertions(+), 12 deletions(-)
> 
> diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
> index ba66c171d951..cccf1eab58ac 100644
> --- a/arch/x86/kvm/vmx/vmx.c
> +++ b/arch/x86/kvm/vmx/vmx.c
> @@ -6264,7 +6264,7 @@ static int vmx_sync_pir_to_irr(struct kvm_vcpu *vcpu)
>  	int max_irr;
>  	bool max_irr_updated;
>  
> -	if (KVM_BUG_ON(!vcpu->arch.apicv_active, vcpu->kvm))
> +	if (KVM_BUG_ON(!enable_apicv, vcpu->kvm))
>  		return -EIO;
>  
>  	if (pi_test_on(&vmx->pi_desc)) {
> @@ -6276,20 +6276,31 @@ static int vmx_sync_pir_to_irr(struct kvm_vcpu *vcpu)
>  		smp_mb__after_atomic();
>  		max_irr_updated =
>  			kvm_apic_update_irr(vcpu, vmx->pi_desc.pir, &max_irr);
> -
> -		/*
> -		 * If we are running L2 and L1 has a new pending interrupt
> -		 * which can be injected, this may cause a vmexit or it may
> -		 * be injected into L2.  Either way, this interrupt will be
> -		 * processed via KVM_REQ_EVENT, not RVI, because we do not use
> -		 * virtual interrupt delivery to inject L1 interrupts into L2.
> -		 */
> -		if (is_guest_mode(vcpu) && max_irr_updated)
> -			kvm_make_request(KVM_REQ_EVENT, vcpu);
>  	} else {
>  		max_irr = kvm_lapic_find_highest_irr(vcpu);
> +		max_irr_updated = false;
>  	}
> -	vmx_hwapic_irr_update(vcpu, max_irr);
> +
> +	/*
> +	 * If virtual interrupt delivery is not in use, the interrupt
> +	 * will be processed via KVM_REQ_EVENT, not RVI.  This can happen
> +	 * in two cases:
> +	 *
> +	 * 1) If we are running L2 and L1 has a new pending interrupt
> +	 * which can be injected, this may cause a vmexit or it may
> +	 * be injected into L2.  We do not use virtual interrupt
> +	 * delivery to inject L1 interrupts into L2.
> +	 *
> +	 * 2) If APICv is disabled for this vCPU, assigned devices may
> +	 * still attempt to post interrupts.  The posted interrupt
> +	 * vector will cause a vmexit and the subsequent entry will
> +	 * call sync_pir_to_irr.
> +	 */
> +	if (!is_guest_mode(vcpu) && vcpu->arch.apicv_active)
> +		vmx_set_rvi(max_irr);
> +	else if (max_irr_updated)
> +		kvm_make_request(KVM_REQ_EVENT, vcpu);

nit: In the version of this code that is currently in kvm/queue the
indentation of the previous 3 lines uses spaces instead of tabs. I see
tabs in this mail though.

> +
>  	return max_irr;
>  }
>  
> -- 
> 2.27.0
> 
>