Re: [PATCH 4/4] KVM: async_pf: Let host know whether the guest support delivery async_pf as #PF vmexit

["Radim Krčmář" <rkrcmar@redhat.com>]

2017-06-12 23:08-0700, Wanpeng Li:
> From: Wanpeng Li <wanpeng.li@hotmail.com>
> 
> Adds another flag bit (bit 2) to MSR_KVM_ASYNC_PF_EN. If bit 2 is 1, async 
> page faults are delivered to L1 as #PF vmexits; if bit 2 is 0, kvm_can_do_async_pf 
> returns 0 if in guest mode.
> 
> Cc: Paolo Bonzini <pbonzini@redhat.com>
> Cc: Radim Krčmář <rkrcmar@redhat.com>
> Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
> ---

I think KVM (L1) should also do something like

  diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
  index dd274db9bf77..c15a9f178e60 100644
  --- a/arch/x86/kvm/vmx.c
  +++ b/arch/x86/kvm/vmx.c
  @@ -7991,7 +7991,7 @@ static bool nested_vmx_exit_handled(struct kvm_vcpu *vcpu)
   		if (is_nmi(intr_info))
   			return false;
   		else if (is_page_fault(intr_info))
  -			return enable_ept;
  +			return !vmx->apf_reason && enable_ept;
   		else if (is_no_device(intr_info) &&
   			 !(vmcs12->guest_cr0 & X86_CR0_TS))
   			return false;

so it doesn't pass the APF directed towards it (L1) into L2 if there is
L3 at the moment.