History

KVM: nVMX: Assume TLB entries of L1 and L2 are tagged differently if L0 use EPT

Since commit 1313cc2bd8f6 ("kvm: mmu: Add guest_mode to kvm_mmu_page_role"), guest_mode was added to mmu-role and therefore if L0 use EPT, it will always run L1 and L2 with different EPTP. i.e. EPTP01!=EPTP02. Because TLB entries are tagged with EP4TA, KVM can assume TLB entries populated while running L2 are tagged differently than TLB entries populated while running L1. Therefore, update nested_has_guest_tlb_tag() to consider if L0 use EPT instead of if L1 use EPT.

Liran Alon committed 6 years ago

Commit 992edea1 parent 5637f601 file · +4 −2Browse files

Unified · Split

arch/x86/kvm/vmx/nested.c+4−2
@@ -1126,7 +1126,9 @@ static int nested_vmx_load_cr3(struct kvm_vcpu *vcpu, unsigned long cr3, bool ne · 1125 unmodified lines
11261126 * populated by L2 differently than TLB entries populated
11271127 * by L1.
11281128 *
1129− * If L1 uses EPT, then TLB entries are tagged with different EPTP.
1129+ * If L0 uses EPT, L1 and L2 run with different EPTP because
1130+ * guest_mode is part of kvm_mmu_page_role. Thus, TLB entries
1131+ * are tagged with different EPTP.
11301132 *
11311133 * If L1 uses VPID and we allocated a vpid02, TLB entries are tagged
11321134 * with different VPID (L1 entries are tagged with vmx->vpid
@@ -1136,7 +1138,7 @@ static bool nested_has_guest_tlb_tag(struct kvm_vcpu *vcpu) · 3 unmodified lines
11361138 {
11371139 struct vmcs12 *vmcs12 = get_vmcs12(vcpu);
11381140
1139− return nested_cpu_has_ept(vmcs12) ||
1141+ return enable_ept ||
11401142 (nested_cpu_has_vpid(vmcs12) && to_vmx(vcpu)->nested.vpid02);
11411143 }
11421144