From 433f0b31ebec09bb89c2414eb76571ced670a202 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Wed, 11 Aug 2021 21:56:15 -0700 Subject: [PATCH] KVM: nVMX: Use vmx_need_pf_intercept() when deciding if L0 wants a #PF commit 18712c13709d2de9516c5d3414f707c4f0a9c190 upstream. Use vmx_need_pf_intercept() when determining if L0 wants to handle a #PF in L2 or if the VM-Exit should be forwarded to L1. The current logic fails to account for the case where #PF is intercepted to handle guest.MAXPHYADDR < host.MAXPHYADDR and ends up reflecting all #PFs into L1. At best, L1 will complain and inject the #PF back into L2. At worst, L1 will eat the unexpected fault and cause L2 to hang on infinite page faults. Note, while the bug was technically introduced by the commit that added support for the MAXPHYADDR madness, the shame is all on commit a0c134347baf ("KVM: VMX: introduce vmx_need_pf_intercept"). Fixes: 1dbf5d68af6f ("KVM: VMX: Add guest physical address check in EPT violation and misconfig") Cc: stable@vger.kernel.org Cc: Peter Shier Cc: Oliver Upton Cc: Jim Mattson Signed-off-by: Sean Christopherson Message-Id: <20210812045615.3167686-1-seanjc@google.com> Signed-off-by: Paolo Bonzini Signed-off-by: Greg Kroah-Hartman --- arch/x86/kvm/vmx/nested.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c index 67554bc7adb26..e0c7910207c0f 100644 --- a/arch/x86/kvm/vmx/nested.c +++ b/arch/x86/kvm/vmx/nested.c @@ -5779,7 +5779,8 @@ static bool nested_vmx_l0_wants_exit(struct kvm_vcpu *vcpu, if (is_nmi(intr_info)) return true; else if (is_page_fault(intr_info)) - return vcpu->arch.apf.host_apf_flags || !enable_ept; + return vcpu->arch.apf.host_apf_flags || + vmx_need_pf_intercept(vcpu); else if (is_debug(intr_info) && vcpu->guest_debug & (KVM_GUESTDBG_SINGLESTEP | KVM_GUESTDBG_USE_HW_BP))