Skip to contentSkip to conversation

KVM: x86: Replace BUG_ON() with WARN_ON_ONCE() on "bad" nested GPA translation

PATCH Applied Sean Christopherson 4 participants 5 messages Jun 18 – Jun 24

Subsystems (3)
lore
Share & tools
lore

Sean Christopherson

author maintainer
#0 lore raw

PATCH 1/1 KVM: x86: Replace BUG_ON() with WARN_ON_ONCE() on "bad" nested GPA translation

If KVM attempts to translate what it thinks is an L2 GPA with a non-nested MMU, simply WARN and return the GPA, i.e. trust the MMU more than the caller, as there is zero reason to potentially panic the host kernel just because KVM misused an API.

---
 arch/x86/kvm/svm/nested.c | 3 ++-
 arch/x86/kvm/vmx/nested.c | 3 ++-
 2 files changed, 4 insertions(+), 2 deletions(-)
arch/x86/kvm/svm/nested.c +2 -1
index 9aedb88c832d..3e6c671a8dc2 100644
@@ -2152,7 +2152,8 @@ static gpa_t svm_translate_nested_gpa(struct kvm_vcpu *vcpu, gpa_t gpa,
 	struct vcpu_svm *svm = to_svm(vcpu);
 	struct kvm_mmu *mmu = vcpu->arch.mmu;
 
-	BUG_ON(!mmu_is_nested(vcpu));
+	if (WARN_ON_ONCE(!mmu_is_nested(vcpu)))
+		return gpa;
 
 	/* Non-GMET walks are always user-walks */
 	if (!(svm->nested.ctl.misc_ctl & SVM_MISC_ENABLE_GMET))
arch/x86/kvm/vmx/nested.c +2 -1
index 3a293640d58c..6957bb6f5cf7 100644
@@ -7470,7 +7470,8 @@ static gpa_t vmx_translate_nested_gpa(struct kvm_vcpu *vcpu, gpa_t gpa,
 {
 	struct kvm_mmu *mmu = vcpu->arch.mmu;
 
-	BUG_ON(!mmu_is_nested(vcpu));
+	if (WARN_ON_ONCE(!mmu_is_nested(vcpu)))
+		return gpa;
 
 	/*
 	 * MBEC differentiates based on the effective U/S bit of

base-commit: 9d4853b044beefa21c4ee3e18c40653601a64ced

Signature · 1 line
2.55.0.rc0.738.g0c8ab3ebcc-goog
Reviewed-by on v1 1/1
#1 lore raw
Quoted 7 lines · Sean Christopherson

If KVM attempts to translate what it thinks is an L2 GPA with a non-nested MMU, simply WARN and return the GPA, i.e. trust the MMU more than the caller, as there is zero reason to potentially panic the host kernel just because KVM misused an API.

#2 lore raw
In reply to Sean Christopherson
Quoted 4 lines · On Thu, 18 Jun 2026 11:57:45 -0700 Sean Christopherson <[email protected]>…

If KVM attempts to translate what it thinks is an L2 GPA with a non-nested MMU, simply WARN and return the GPA, i.e. trust the MMU more than the caller, as there is zero reason to potentially panic the host kernel just because KVM misused an API.

Except that PANIC_ON_WARN stands a reasonable chance of being set. So it makes little difference.

David

Quoted 37 lines
---
 arch/x86/kvm/svm/nested.c | 3 ++-
 arch/x86/kvm/vmx/nested.c | 3 ++-
 2 files changed, 4 insertions(+), 2 deletions(-)
arch/x86/kvm/svm/nested.c +2 -1
index 9aedb88c832d..3e6c671a8dc2 100644
@@ -2152,7 +2152,8 @@ static gpa_t svm_translate_nested_gpa(struct kvm_vcpu *vcpu, gpa_t gpa,
 	struct vcpu_svm *svm = to_svm(vcpu);
 	struct kvm_mmu *mmu = vcpu->arch.mmu;
 
-	BUG_ON(!mmu_is_nested(vcpu));
+	if (WARN_ON_ONCE(!mmu_is_nested(vcpu)))
+		return gpa;
 
 	/* Non-GMET walks are always user-walks */
 	if (!(svm->nested.ctl.misc_ctl & SVM_MISC_ENABLE_GMET))
arch/x86/kvm/vmx/nested.c +2 -1
index 3a293640d58c..6957bb6f5cf7 100644
@@ -7470,7 +7470,8 @@ static gpa_t vmx_translate_nested_gpa(struct kvm_vcpu *vcpu, gpa_t gpa,
 {
 	struct kvm_mmu *mmu = vcpu->arch.mmu;
 
-	BUG_ON(!mmu_is_nested(vcpu));
+	if (WARN_ON_ONCE(!mmu_is_nested(vcpu)))
+		return gpa;
 
 	/*
 	 * MBEC differentiates based on the effective U/S bit of

base-commit: 9d4853b044beefa21c4ee3e18c40653601a64ced

author maintainer
#3 lore raw
Quoted 9 lines · David Laight
Quoted 4 lines · On Thu, 18 Jun 2026 11:57:45 -0700 Sean Christopherson <[email protected]>…

If KVM attempts to translate what it thinks is an L2 GPA with a non-nested MMU, simply WARN and return the GPA, i.e. trust the MMU more than the caller, as there is zero reason to potentially panic the host kernel just because KVM misused an API.

Except that PANIC_ON_WARN stands a reasonable chance of being set.

Not in cloud environments, or in any environment where the guest workload is untrusted.

Quoted 1 line

So it makes little difference.

I disagree, vehemently. There's a massive difference between opting in to minimizing risk of data corruption at the cost of availability, and forcing all KVM users to sacrifice availability, especially for no tangible beneft. Paolo and I are fully aligned on this:

https://lore.kernel.org/all/CABgObfZJV5hU_7WoPWLRH3-EvKts%[email protected]