From 8c4f28cd81fe86033918eec69d5280b532c05842 Mon Sep 17 00:00:00 2001 From: Yosry Ahmed Date: Thu, 27 Feb 2025 01:27:11 +0000 Subject: [PATCH] KVM: nVMX: Always use IBPB to properly virtualize IBRS On synthesized nested VM-exits in VMX, an IBPB is performed if IBRS is advertised to the guest to properly provide separate prediction domains for L1 and L2. However, this is currently conditional on X86_FEATURE_USE_IBPB, which depends on the host spectre_v2_user mitigation. In short, if spectre_v2_user=no, IBRS is not virtualized correctly and L1 becomes susceptible to attacks from L2. Fix this by performing the IBPB regardless of X86_FEATURE_USE_IBPB. Fixes: 2e7eab81425a ("KVM: VMX: Execute IBPB on emulated VM-exit when guest has IBRS") Signed-off-by: Yosry Ahmed Signed-off-by: Ingo Molnar Reviewed-by: Jim Mattson Acked-by: Josh Poimboeuf Acked-by: Sean Christopherson Link: https://lore.kernel.org/r/20250227012712.3193063-6-yosry.ahmed@linux.dev --- arch/x86/kvm/vmx/nested.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c index 1df427a822dd..8a7af02d466e 100644 --- a/arch/x86/kvm/vmx/nested.c +++ b/arch/x86/kvm/vmx/nested.c @@ -5026,8 +5026,7 @@ void nested_vmx_vmexit(struct kvm_vcpu *vcpu, u32 vm_exit_reason, * doesn't isolate different VMCSs, i.e. in this case, doesn't provide * separate modes for L2 vs L1. */ - if (guest_cpu_cap_has(vcpu, X86_FEATURE_SPEC_CTRL) && - cpu_feature_enabled(X86_FEATURE_USE_IBPB)) + if (guest_cpu_cap_has(vcpu, X86_FEATURE_SPEC_CTRL)) indirect_branch_prediction_barrier(); /* Update any VMCS fields that might have changed while L2 ran */