Re: [PATCH v3 2/3] KVM: x86: Drop superfluous caching of KVM_ASYNC_PF_DELIVERY_AS_PF_VMEXIT

[Binbin Wu <binbin.wu@linux.intel.com>]

On 4/7/2026 6:53 AM, Sean Christopherson wrote:
> Drop kvm_vcpu_arch.delivery_as_pf_vmexit and instead use msr_en_val as the

Nit:
kvm_vcpu_arch.apf.elivery_as_pf_vmexit


> source of truth to reduce the probability of operating on stale data.  This
> fixes flaws where KVM fails to update delivery_as_pf_vmexit when APF is
> explicitly disabled by the guest or implicitly disabled by KVM on INIT.
> Absent other bugs, the flaws are benign as KVM *shouldn't* consume
> delivery_as_pf_vmexit when PV APF support is disabled.
> 
> Simply delete the field, as there's zero benefit to maintaining a separate
> "cache" of the state.
> 
> Fixes: 52a5c155cf79 ("KVM: async_pf: Let guest support delivery of async_pf from guest mode")
> Signed-off-by: Sean Christopherson <seanjc@google.com>


Reviewed-by: Binbin Wu <binbin.wu@linux.intel.com>

> ---
>  arch/x86/include/asm/kvm_host.h | 1 -
>  arch/x86/kvm/x86.c              | 3 +--
>  2 files changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
> index c470e40a00aa..fae1f4aeca5a 100644
> --- a/arch/x86/include/asm/kvm_host.h
> +++ b/arch/x86/include/asm/kvm_host.h
> @@ -1039,7 +1039,6 @@ struct kvm_vcpu_arch {
>  		u32 id;
>  		u32 host_apf_flags;
>  		bool send_always;
> -		bool delivery_as_pf_vmexit;
>  		bool pageready_pending;
>  	} apf;
>  
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index c35d359b56dd..4632222a5d1c 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -3666,7 +3666,6 @@ static int kvm_pv_enable_async_pf(struct kvm_vcpu *vcpu, u64 data)
>  	}
>  
>  	vcpu->arch.apf.send_always = (data & KVM_ASYNC_PF_SEND_ALWAYS);
> -	vcpu->arch.apf.delivery_as_pf_vmexit = data & KVM_ASYNC_PF_DELIVERY_AS_PF_VMEXIT;
>  
>  	kvm_async_pf_wakeup_all(vcpu);
>  
> @@ -14035,7 +14034,7 @@ static bool kvm_can_deliver_async_pf(struct kvm_vcpu *vcpu)
>  		 * L1 needs to opt into the special #PF vmexits that are
>  		 * used to deliver async page faults.
>  		 */
> -		return vcpu->arch.apf.delivery_as_pf_vmexit;
> +		return vcpu->arch.apf.msr_en_val & KVM_ASYNC_PF_DELIVERY_AS_PF_VMEXIT;
>  	} else {
>  		/*
>  		 * Play it safe in case the guest temporarily disables paging.