According to the APM Volume #2, 15.5, Canonicalization and Consistency Checks (24593—Rev. 3.42—March 2024), the following condition (among others) results in a #VMEXIT with VMEXIT_INVALID (aka SVM_EXIT_ERR):
EFER.LME, CR0.PG, CR4.PAE, CS.L, and CS.D are all non-zero.
Add the missing consistency check. This is functionally a nop because the nested VMRUN results in SVM_EXIT_ERR in HW, which is forwarded to L1, but KVM makes all consistency checks before a VMRUN is actually attempted.
Fixes: 3d6368ef580a ("KVM: SVM: Add VMRUN handler") Cc: stable@vger.kernel.org Signed-off-by: Yosry Ahmed yosry.ahmed@linux.dev --- arch/x86/kvm/svm/nested.c | 7 +++++++ arch/x86/kvm/svm/svm.h | 1 + 2 files changed, 8 insertions(+)
diff --git a/arch/x86/kvm/svm/nested.c b/arch/x86/kvm/svm/nested.c index 5184e5ae8ccf..47d4316b126d 100644 --- a/arch/x86/kvm/svm/nested.c +++ b/arch/x86/kvm/svm/nested.c @@ -382,6 +382,11 @@ static bool nested_vmcb_check_save(struct kvm_vcpu *vcpu, CC(!(save->cr0 & X86_CR0_PE)) || CC(!kvm_vcpu_is_legal_cr3(vcpu, save->cr3))) return false; + + if (CC((save->cr4 & X86_CR4_PAE) && + (save->cs.attrib & SVM_SELECTOR_L_MASK) && + (save->cs.attrib & SVM_SELECTOR_DB_MASK))) + return false; }
/* Note, SVM doesn't have any additional restrictions on CR4. */ @@ -458,6 +463,8 @@ static void __nested_copy_vmcb_save_to_cache(struct vmcb_save_area_cached *to, * Copy only fields that are validated, as we need them * to avoid TOC/TOU races. */ + to->cs = from->cs; + to->efer = from->efer; to->cr0 = from->cr0; to->cr3 = from->cr3; diff --git a/arch/x86/kvm/svm/svm.h b/arch/x86/kvm/svm/svm.h index 9aa60924623f..d0de7d390889 100644 --- a/arch/x86/kvm/svm/svm.h +++ b/arch/x86/kvm/svm/svm.h @@ -139,6 +139,7 @@ struct kvm_vmcb_info { };
struct vmcb_save_area_cached { + struct vmcb_seg cs; u64 efer; u64 cr4; u64 cr3;