Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755915AbcCRQKF (ORCPT ); Fri, 18 Mar 2016 12:10:05 -0400 Received: from mail-wm0-f65.google.com ([74.125.82.65]:36823 "EHLO mail-wm0-f65.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756010AbcCRQJq (ORCPT ); Fri, 18 Mar 2016 12:09:46 -0400 From: Paolo Bonzini To: linux-kernel@vger.kernel.org, kvm@vger.kernel.org Cc: jmontleo@redhat.com, bdas@redhat.com, stable@vger.kernel.org Subject: [PATCH 3/3] KVM: VMX: fix nested vpid for old KVM guests Date: Fri, 18 Mar 2016 17:09:39 +0100 Message-Id: <1458317379-8363-4-git-send-email-pbonzini@redhat.com> X-Mailer: git-send-email 2.5.0 In-Reply-To: <1458317379-8363-1-git-send-email-pbonzini@redhat.com> References: <1458317379-8363-1-git-send-email-pbonzini@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1887 Lines: 55 Old KVM guests invoke single-context invvpid without actually checking whether it is supported. This was fixed by commit 518c8ae ("KVM: VMX: Make sure single type invvpid is supported before issuing invvpid instruction", 2010-08-01) and the patch after, but pre-2.6.36 kernels lack it including RHEL 6. Reported-by: jmontleo@redhat.com Cc: stable@vger.kernel.org Fixes: 99b83ac893b84ed1a62ad6d1f2b6cc32026b9e85 Signed-off-by: Paolo Bonzini --- arch/x86/kvm/vmx.c | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-) diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index 728f31bd350e..5cf66c14aac0 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -2712,8 +2712,15 @@ static void nested_vmx_setup_ctls_msrs(struct vcpu_vmx *vmx) } else vmx->nested.nested_vmx_ept_caps = 0; + /* + * Old versions of KVM use the single-context version without + * checking for support, so declare that it is supported even + * though it is treated as global context. The alternative is + * not failing the single-context invvpid, and it is worse. + */ if (enable_vpid) vmx->nested.nested_vmx_vpid_caps = VMX_VPID_INVVPID_BIT | + VMX_VPID_EXTENT_SINGLE_CONTEXT_BIT | VMX_VPID_EXTENT_GLOBAL_CONTEXT_BIT; else vmx->nested.nested_vmx_vpid_caps = 0; @@ -7475,12 +7482,17 @@ static int handle_invvpid(struct kvm_vcpu *vcpu) } switch (type) { + case VMX_VPID_EXTENT_SINGLE_CONTEXT: + /* + * Old versions of KVM use the single-context version so we + * have to support it; just treat it the same as all-context. + */ case VMX_VPID_EXTENT_ALL_CONTEXT: __vmx_flush_tlb(vcpu, to_vmx(vcpu)->nested.vpid02); nested_vmx_succeed(vcpu); break; default: - /* Trap single context invalidation invvpid calls */ + /* Trap individual address invalidation invvpid calls */ BUG_ON(1); break; } -- 1.8.3.1