linux-kernel - [PATCH v2 12/22] KVM: nSVM: Move SMI vmexit handling to svm_check_nested

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20200425070154.251290-3-pbonzini@redhat.com>
Date:   Sat, 25 Apr 2020 03:01:44 -0400
From:   Paolo Bonzini <pbonzini@...hat.com>
To:     linux-kernel@...r.kernel.org, kvm@...r.kernel.org
Cc:     wei.huang2@....com, cavery@...hat.com, vkuznets@...hat.com,
        Sean Christopherson <sean.j.christopherson@...el.com>,
        Oliver Upton <oupton@...gle.com>,
        Jim Mattson <jmattson@...gle.com>
Subject: [PATCH v2 12/22] KVM: nSVM: Move SMI vmexit handling to svm_check_nested_events()

Unlike VMX, SVM allows a hypervisor to take a SMI vmexit without having
any special SMM-monitor enablement sequence.  Therefore, it has to be
handled like interrupts and NMIs.  Check for an unblocked SMI in
svm_check_nested_events() so that pending SMIs are correctly prioritized
over IRQs and NMIs when the latter events will trigger VM-Exit.

Note that there is no need to test explicitly for SMI vmexits, because
guests always runs outside SMM and therefore can never get an SMI while
they are blocked.

Signed-off-by: Paolo Bonzini <pbonzini@...hat.com>
---
 arch/x86/kvm/svm/nested.c | 16 ++++++++++++++++
 arch/x86/kvm/svm/svm.c    |  8 --------
 arch/x86/kvm/svm/svm.h    |  5 +++++
 3 files changed, 21 insertions(+), 8 deletions(-)

diff --git a/arch/x86/kvm/svm/nested.c b/arch/x86/kvm/svm/nested.c
index 748b01220aac..226d5a0d677b 100644
--- a/arch/x86/kvm/svm/nested.c
+++ b/arch/x86/kvm/svm/nested.c
@@ -776,6 +776,15 @@ int nested_svm_check_exception(struct vcpu_svm *svm, unsigned nr,
 	return vmexit;
 }
 
+static void nested_svm_smi(struct vcpu_svm *svm)
+{
+	svm->vmcb->control.exit_code = SVM_EXIT_SMI;
+	svm->vmcb->control.exit_info_1 = 0;
+	svm->vmcb->control.exit_info_2 = 0;
+
+	nested_svm_vmexit(svm);
+}
+
 static void nested_svm_nmi(struct vcpu_svm *svm)
 {
 	svm->vmcb->control.exit_code = SVM_EXIT_NMI;
@@ -807,6 +816,13 @@ static int svm_check_nested_events(struct kvm_vcpu *vcpu)
 		kvm_event_needs_reinjection(vcpu) || svm->nested.exit_required ||
 		svm->nested.nested_run_pending;
 
+	if (vcpu->arch.smi_pending && nested_exit_on_smi(svm)) {
+		if (block_nested_events)
+			return -EBUSY;
+		nested_svm_smi(svm);
+		return 0;
+	}
+
 	if (vcpu->arch.nmi_pending && nested_exit_on_nmi(svm)) {
 		if (block_nested_events)
 			return -EBUSY;
diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
index 3f1f80737f9e..1f9577b281e6 100644
--- a/arch/x86/kvm/svm/svm.c
+++ b/arch/x86/kvm/svm/svm.c
@@ -3766,14 +3766,6 @@ static bool svm_smi_allowed(struct kvm_vcpu *vcpu)
 	if (!gif_set(svm))
 		return false;
 
-	if (is_guest_mode(&svm->vcpu) &&
-	    svm->nested.intercept & (1ULL << INTERCEPT_SMI)) {
-		/* TODO: Might need to set exit_info_1 and exit_info_2 here */
-		svm->vmcb->control.exit_code = SVM_EXIT_SMI;
-		svm->nested.exit_required = true;
-		return false;
-	}
-
 	return !is_smm(vcpu);
 }
 
diff --git a/arch/x86/kvm/svm/svm.h b/arch/x86/kvm/svm/svm.h
index d8ae654340d4..4dc6d2b4b721 100644
--- a/arch/x86/kvm/svm/svm.h
+++ b/arch/x86/kvm/svm/svm.h
@@ -378,6 +378,11 @@ static inline bool svm_nested_virtualize_tpr(struct kvm_vcpu *vcpu)
 	return is_guest_mode(vcpu) && (vcpu->arch.hflags & HF_VINTR_MASK);
 }
 
+static inline bool nested_exit_on_smi(struct vcpu_svm *svm)
+{
+	return (svm->nested.intercept & (1ULL << INTERCEPT_SMI));
+}
+
 static inline bool nested_exit_on_nmi(struct vcpu_svm *svm)
 {
 	return (svm->nested.intercept & (1ULL << INTERCEPT_NMI));
-- 
2.18.2