This is a note to let you know that I've just added the patch titled KVM: VMX: Execute IBPB on emulated VM-exit when guest has IBRS to the 5.10-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary The filename of the patch is: kvm-vmx-execute-ibpb-on-emulated-vm-exit-when-guest-.patch and it can be found in the queue-5.10 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable@xxxxxxxxxxxxxxx> know about it. commit c1dbbed2206ddef9a3eecf988a76d5089f574aa1 Author: Jim Mattson <jmattson@xxxxxxxxxx> Date: Wed Oct 19 14:36:20 2022 -0700 KVM: VMX: Execute IBPB on emulated VM-exit when guest has IBRS [ Upstream commit 2e7eab81425ad6c875f2ed47c0ce01e78afc38a5 ] According to Intel's document on Indirect Branch Restricted Speculation, "Enabling IBRS does not prevent software from controlling the predicted targets of indirect branches of unrelated software executed later at the same predictor mode (for example, between two different user applications, or two different virtual machines). Such isolation can be ensured through use of the Indirect Branch Predictor Barrier (IBPB) command." This applies to both basic and enhanced IBRS. Since L1 and L2 VMs share hardware predictor modes (guest-user and guest-kernel), hardware IBRS is not sufficient to virtualize IBRS. (The way that basic IBRS is implemented on pre-eIBRS parts, hardware IBRS is actually sufficient in practice, even though it isn't sufficient architecturally.) For virtual CPUs that support IBRS, add an indirect branch prediction barrier on emulated VM-exit, to ensure that the predicted targets of indirect branches executed in L1 cannot be controlled by software that was executed in L2. Since we typically don't intercept guest writes to IA32_SPEC_CTRL, perform the IBPB at emulated VM-exit regardless of the current IA32_SPEC_CTRL.IBRS value, even though the IBPB could technically be deferred until L1 sets IA32_SPEC_CTRL.IBRS, if IA32_SPEC_CTRL.IBRS is clear at emulated VM-exit. This is CVE-2022-2196. Fixes: 5c911beff20a ("KVM: nVMX: Skip IBPB when switching between vmcs01 and vmcs02") Cc: Sean Christopherson <seanjc@xxxxxxxxxx> Signed-off-by: Jim Mattson <jmattson@xxxxxxxxxx> Reviewed-by: Sean Christopherson <seanjc@xxxxxxxxxx> Link: https://lore.kernel.org/r/20221019213620.1953281-3-jmattson@xxxxxxxxxx Signed-off-by: Sean Christopherson <seanjc@xxxxxxxxxx> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx> diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c index f15ddf58a5bcd..91371b01eae0c 100644 --- a/arch/x86/kvm/vmx/nested.c +++ b/arch/x86/kvm/vmx/nested.c @@ -4556,6 +4556,17 @@ void nested_vmx_vmexit(struct kvm_vcpu *vcpu, u32 vm_exit_reason, vmx_switch_vmcs(vcpu, &vmx->vmcs01); + /* + * If IBRS is advertised to the vCPU, KVM must flush the indirect + * branch predictors when transitioning from L2 to L1, as L1 expects + * hardware (KVM in this case) to provide separate predictor modes. + * Bare metal isolates VMX root (host) from VMX non-root (guest), but + * doesn't isolate different VMCSs, i.e. in this case, doesn't provide + * separate modes for L2 vs L1. + */ + if (guest_cpuid_has(vcpu, X86_FEATURE_SPEC_CTRL)) + indirect_branch_prediction_barrier(); + /* Update any VMCS fields that might have changed while L2 ran */ vmcs_write32(VM_EXIT_MSR_LOAD_COUNT, vmx->msr_autoload.host.nr); vmcs_write32(VM_ENTRY_MSR_LOAD_COUNT, vmx->msr_autoload.guest.nr); diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c index 8f7152e158e28..c37cbd3fdd852 100644 --- a/arch/x86/kvm/vmx/vmx.c +++ b/arch/x86/kvm/vmx/vmx.c @@ -1431,8 +1431,10 @@ void vmx_vcpu_load_vmcs(struct kvm_vcpu *vcpu, int cpu, /* * No indirect branch prediction barrier needed when switching - * the active VMCS within a guest, e.g. on nested VM-Enter. - * The L1 VMM can protect itself with retpolines, IBPB or IBRS. + * the active VMCS within a vCPU, unless IBRS is advertised to + * the vCPU. To minimize the number of IBPBs executed, KVM + * performs IBPB on nested VM-Exit (a single nested transition + * may switch the active VMCS multiple times). */ if (!buddy || WARN_ON_ONCE(buddy->vmcs != prev)) indirect_branch_prediction_barrier();