linux-kernel - [PATCH 06/28] KVM: SVM: always update CR3 in VMCB

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20200526172308.111575-7-pbonzini@redhat.com>
Date:   Tue, 26 May 2020 13:22:46 -0400
From:   Paolo Bonzini <pbonzini@...hat.com>
To:     linux-kernel@...r.kernel.org, kvm@...r.kernel.org
Cc:     vkuznets@...hat.com, mlevitsk@...hat.com,
        Sean Christopherson <sean.j.christopherson@...el.com>,
        Jim Mattson <jmattson@...gle.com>
Subject: [PATCH 06/28] KVM: SVM: always update CR3 in VMCB

svm_load_mmu_pgd is delaying the write of GUEST_CR3 to prepare_vmcs02 as
an optimization, but this is only correct before the nested vmentry.
If userspace is modifying CR3 with KVM_SET_SREGS after the VM has
already been put in guest mode, the value of CR3 will not be updated.
Remove the optimization, which almost never triggers anyway.
This was was added in commit 689f3bf21628 ("KVM: x86: unify callbacks
to load paging root", 2020-03-16) just to keep the two vendor-specific
modules closer, but we'll fix VMX too.

Fixes: 689f3bf21628 ("KVM: x86: unify callbacks to load paging root")
Signed-off-by: Paolo Bonzini <pbonzini@...hat.com>
---
 arch/x86/kvm/svm/nested.c |  6 +-----
 arch/x86/kvm/svm/svm.c    | 16 +++++-----------
 2 files changed, 6 insertions(+), 16 deletions(-)

diff --git a/arch/x86/kvm/svm/nested.c b/arch/x86/kvm/svm/nested.c
index 166b88fc9509..81e0fbd5e267 100644
--- a/arch/x86/kvm/svm/nested.c
+++ b/arch/x86/kvm/svm/nested.c
@@ -256,11 +256,7 @@ void enter_svm_guest_mode(struct vcpu_svm *svm, u64 vmcb_gpa,
 	svm_set_efer(&svm->vcpu, nested_vmcb->save.efer);
 	svm_set_cr0(&svm->vcpu, nested_vmcb->save.cr0);
 	svm_set_cr4(&svm->vcpu, nested_vmcb->save.cr4);
-	if (npt_enabled) {
-		svm->vmcb->save.cr3 = nested_vmcb->save.cr3;
-		svm->vcpu.arch.cr3 = nested_vmcb->save.cr3;
-	} else
-		(void)kvm_set_cr3(&svm->vcpu, nested_vmcb->save.cr3);
+	(void)kvm_set_cr3(&svm->vcpu, nested_vmcb->save.cr3);
 
 	/* Guest paging mode is active - reset mmu */
 	kvm_mmu_reset_context(&svm->vcpu);
diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
index 270061fa6cfa..abe277a3216b 100644
--- a/arch/x86/kvm/svm/svm.c
+++ b/arch/x86/kvm/svm/svm.c
@@ -3447,7 +3447,6 @@ static fastpath_t svm_vcpu_run(struct kvm_vcpu *vcpu)
 static void svm_load_mmu_pgd(struct kvm_vcpu *vcpu, unsigned long root)
 {
 	struct vcpu_svm *svm = to_svm(vcpu);
-	bool update_guest_cr3 = true;
 	unsigned long cr3;
 
 	cr3 = __sme_set(root);
@@ -3456,18 +3455,13 @@ static void svm_load_mmu_pgd(struct kvm_vcpu *vcpu, unsigned long root)
 		mark_dirty(svm->vmcb, VMCB_NPT);
 
 		/* Loading L2's CR3 is handled by enter_svm_guest_mode.  */
-		if (is_guest_mode(vcpu))
-			update_guest_cr3 = false;
-		else if (test_bit(VCPU_EXREG_CR3, (ulong *)&vcpu->arch.regs_avail))
-			cr3 = vcpu->arch.cr3;
-		else /* CR3 is already up-to-date.  */
-			update_guest_cr3 = false;
+		if (!test_bit(VCPU_EXREG_CR3, (ulong *)&vcpu->arch.regs_avail))
+			return;
+		cr3 = vcpu->arch.cr3;
 	}
 
-	if (update_guest_cr3) {
-		svm->vmcb->save.cr3 = cr3;
-		mark_dirty(svm->vmcb, VMCB_CR);
-	}
+	svm->vmcb->save.cr3 = cr3;
+	mark_dirty(svm->vmcb, VMCB_CR);
 }
 
 static int is_disabled(void)
-- 
2.26.2