linux-kernel - [PATCH v2 06/21] KVM: x86/pmu: WARN and bug the VM if PMU is refreshed after vCPU has run

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 for Android: free password hash cracker in your pocket

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

Message-ID: <20230210003148.2646712-7-seanjc@google.com>
Date:   Fri, 10 Feb 2023 00:31:33 +0000
From:   Sean Christopherson <seanjc@...gle.com>
To:     Sean Christopherson <seanjc@...gle.com>,
        Paolo Bonzini <pbonzini@...hat.com>
Cc:     kvm@...r.kernel.org, linux-kernel@...r.kernel.org,
        Xiaoyao Li <xiaoyao.li@...el.com>,
        Like Xu <like.xu.linux@...il.com>
Subject: [PATCH v2 06/21] KVM: x86/pmu: WARN and bug the VM if PMU is
 refreshed after vCPU has run

Now that KVM disallows changing feature MSRs, i.e. PERF_CAPABILITIES,
after running a vCPU, WARN and bug the VM if the PMU is refreshed after
the vCPU has run.

Note, KVM has disallowed CPUID updates after running a vCPU since commit
feb627e8d6f6 ("KVM: x86: Forbid KVM_SET_CPUID{,2} after KVM_RUN"), i.e.
PERF_CAPABILITIES was the only remaining way to trigger a PMU refresh
after KVM_RUN.

Cc: Like Xu <like.xu.linux@...il.com>
Signed-off-by: Sean Christopherson <seanjc@...gle.com>
---
 arch/x86/kvm/pmu.c |  3 +++
 arch/x86/kvm/x86.c | 10 +++++++++-
 2 files changed, 12 insertions(+), 1 deletion(-)

diff --git a/arch/x86/kvm/pmu.c b/arch/x86/kvm/pmu.c
index 612e6c70ce2e..7e974c4e61b0 100644
--- a/arch/x86/kvm/pmu.c
+++ b/arch/x86/kvm/pmu.c
@@ -589,6 +589,9 @@ int kvm_pmu_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
  */
 void kvm_pmu_refresh(struct kvm_vcpu *vcpu)
 {
+	if (KVM_BUG_ON(kvm_vcpu_has_run(vcpu), vcpu->kvm))
+		return;
+
 	static_call(kvm_x86_pmu_refresh)(vcpu);
 }
 
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 186cb6a81643..1b14632a94a3 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -3626,9 +3626,17 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 		if (data & ~kvm_caps.supported_perf_cap)
 			return 1;
 
+		/*
+		 * Note, this is not just a performance optimization!  KVM
+		 * disallows changing feature MSRs after the vCPU has run; PMU
+		 * refresh will bug the VM if called after the vCPU has run.
+		 */
+		if (vcpu->arch.perf_capabilities == data)
+			break;
+
 		vcpu->arch.perf_capabilities = data;
 		kvm_pmu_refresh(vcpu);
-		return 0;
+		break;
 	case MSR_EFER:
 		return set_efer(vcpu, msr_info);
 	case MSR_K7_HWCR:
-- 
2.39.1.581.gbfd45094c4-goog