From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Kiszka Subject: Re: [PATCH v4 5/6] KVM: nVMX: Fix conditions for NMI injection Date: Sun, 14 Apr 2013 18:35:14 +0200 Message-ID: <516ADAC2.7090306@web.de> References: <523b4691a695525a22102c496a078fda6a13d392.1365934368.git.jan.kiszka@web.de> <20130414152359.GT17919@redhat.com> <516AD0E1.5000000@web.de> <20130414161806.GA6542@redhat.com> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="----enig2VBDWUGRWTWMXXVHCQQCG" Cc: Marcelo Tosatti , kvm , Paolo Bonzini , Nadav Har'El To: Gleb Natapov Return-path: Received: from mout.web.de ([212.227.15.3]:58599 "EHLO mout.web.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751580Ab3DNQfT (ORCPT ); Sun, 14 Apr 2013 12:35:19 -0400 In-Reply-To: <20130414161806.GA6542@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: This is an OpenPGP/MIME signed message (RFC 4880 and 3156) ------enig2VBDWUGRWTWMXXVHCQQCG Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable On 2013-04-14 18:18, Gleb Natapov wrote: > On Sun, Apr 14, 2013 at 05:53:05PM +0200, Jan Kiszka wrote: >> On 2013-04-14 17:23, Gleb Natapov wrote: >>> On Sun, Apr 14, 2013 at 12:12:49PM +0200, Jan Kiszka wrote: >>>> From: Jan Kiszka >>>> >>>> The logic for checking if interrupts can be injected has to be appli= ed >>>> also on NMIs. The difference is that if NMI interception is on these= >>>> events are consumed and blocked by the VM exit. >>>> >>>> Signed-off-by: Jan Kiszka >>>> --- >>>> arch/x86/kvm/vmx.c | 28 ++++++++++++++++++++++++++++ >>>> 1 files changed, 28 insertions(+), 0 deletions(-) >>>> >>>> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c >>>> index 56e7519..ad9b4bc 100644 >>>> --- a/arch/x86/kvm/vmx.c >>>> +++ b/arch/x86/kvm/vmx.c >>>> @@ -4190,6 +4190,12 @@ static bool nested_exit_on_intr(struct kvm_vc= pu *vcpu) >>>> PIN_BASED_EXT_INTR_MASK; >>>> } >>>> =20 >>>> +static bool nested_exit_on_nmi(struct kvm_vcpu *vcpu) >>>> +{ >>>> + return get_vmcs12(vcpu)->pin_based_vm_exec_control & >>>> + PIN_BASED_NMI_EXITING; >>>> +} >>>> + >>>> static void enable_irq_window(struct kvm_vcpu *vcpu) >>>> { >>>> u32 cpu_based_vm_exec_control; >>>> @@ -4315,6 +4321,28 @@ static void vmx_set_nmi_mask(struct kvm_vcpu = *vcpu, bool masked) >>>> =20 >>>> static int vmx_nmi_allowed(struct kvm_vcpu *vcpu) >>>> { >>>> + if (is_guest_mode(vcpu)) { >>>> + struct vmcs12 *vmcs12 =3D get_vmcs12(vcpu); >>>> + >>>> + if (to_vmx(vcpu)->nested.nested_run_pending || >>>> + vmcs_read32(GUEST_ACTIVITY_STATE) =3D=3D >>>> + GUEST_ACTIVITY_WAIT_SIPI) >>> The same is true for interrupt too, >> >> Yes, but aren't we already waiting with interrupts disabled in that st= ate? >> > Why? L1 can do vmcs_write(GUEST_ACTIVITY_STATE, > GUEST_ACTIVITY_WAIT_SIPI) at any point. Hmm, ok. >=20 >>> but I do not think that we should allow >>> nested guest directly enter GUEST_ACTIVITY_WAIT_SIPI state or any oth= er >>> state except ACTIVE. They should be emulated instead. >> >> Well, they aren't emulated yet but directly applied. So I think the >> patch is correct in the current context at least. >> > If my understanding is correct the facts that it is directly applied is= > a serious bug and should be fixed. Yeah, L1 could put the vCPU in wait-for-SIPI, and only that physical signal will wake it up again. But L0 will not send it... >=20 >> What negative effects do you expect from entering those states with L2= ? >> > As I wrote below (and you misunderstood me :)) it looks like L0 externa= l > interrupts do not generate vmexit while active VMCS is in Wait-For-SIPI= > state. In any case we should carefully examine what are the implication= s > of this state since KVM never uses it and does not know how to handle i= t. OK, but this should be conceptually unrelated to this patch. So, given that you applied patch 4, should I simply remove the activity state check from this one in order to proceed with it? Jan ------enig2VBDWUGRWTWMXXVHCQQCG Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.16 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iEYEARECAAYFAlFq2sIACgkQitSsb3rl5xR7SwCgi9Q/R5o4ZttwqieZAw7FbjSj 3nIAn2vlszxFFlb+fZ7Q9HDjGKN2p66r =QAat -----END PGP SIGNATURE----- ------enig2VBDWUGRWTWMXXVHCQQCG--