public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
From: Yosry Ahmed <yosry@kernel.org>
To: Sean Christopherson <seanjc@google.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>,
	kvm@vger.kernel.org, linux-kernel@vger.kernel.org,
	Yosry Ahmed <yosry@kernel.org>
Subject: [PATCH v6 26/31] KVM: nSVM: Restrict mapping VMCB12 on nested VMRUN
Date: Tue, 24 Feb 2026 22:34:00 +0000	[thread overview]
Message-ID: <20260224223405.3270433-27-yosry@kernel.org> (raw)
In-Reply-To: <20260224223405.3270433-1-yosry@kernel.org>

All accesses to the VMCB12 in the guest memory on nested VMRUN are
limited to nested_svm_vmrun() and nested_svm_vmrun_error_vmexit().
However, the VMCB12 remains mapped throughout nested_svm_vmrun().
Mapping and unmapping around usages is possible, but it becomes easy-ish
to introduce bugs where 'vmcb12' is used after being unmapped.

Move reading the VMCB12 and copying to cache from nested_svm_vmrun()
into a new helper, nested_svm_copy_vmcb12_to_cache(),  that maps the
VMCB12, caches the needed fields, and unmaps it. Use
kvm_vcpu_map_readonly() as only reading the VMCB12 is needed.

Similarly, move mapping the VMCB12 on VMRUN failure into
nested_svm_vmrun_error_vmexit(). Inject a triple fault if the mapping
fails, similar to nested_svm_vmexit().

Signed-off-by: Yosry Ahmed <yosry@kernel.org>
---
 arch/x86/kvm/svm/nested.c | 49 +++++++++++++++++++++++++++------------
 1 file changed, 34 insertions(+), 15 deletions(-)

diff --git a/arch/x86/kvm/svm/nested.c b/arch/x86/kvm/svm/nested.c
index 2159f5fbfc314..5c8449bc6fa0c 100644
--- a/arch/x86/kvm/svm/nested.c
+++ b/arch/x86/kvm/svm/nested.c
@@ -1102,28 +1102,54 @@ static void __nested_svm_vmexit(struct vcpu_svm *svm)
 		kvm_queue_exception(vcpu, DB_VECTOR);
 }
 
-static void nested_svm_vmrun_error_vmexit(struct kvm_vcpu *vcpu, struct vmcb *vmcb12)
+static void nested_svm_vmrun_error_vmexit(struct kvm_vcpu *vcpu, u64 vmcb12_gpa)
 {
 	struct vcpu_svm *svm = to_svm(vcpu);
+	struct kvm_host_map map;
+	struct vmcb *vmcb12;
 
 	WARN_ON_ONCE(svm->vmcb == svm->nested.vmcb02.ptr);
 
 	leave_guest_mode(vcpu);
 
+	if (kvm_vcpu_map(vcpu, gpa_to_gfn(vmcb12_gpa), &map)) {
+		kvm_make_request(KVM_REQ_TRIPLE_FAULT, vcpu);
+		goto out;
+	}
+
+	vmcb12 = map.hva;
 	vmcb12->control.exit_code = SVM_EXIT_ERR;
 	vmcb12->control.exit_info_1 = 0;
 	vmcb12->control.exit_info_2 = 0;
 	vmcb12->control.event_inj = 0;
 	vmcb12->control.event_inj_err = 0;
+	kvm_vcpu_unmap(vcpu, &map);
+out:
 	__nested_svm_vmexit(svm);
 }
 
+static int nested_svm_copy_vmcb12_to_cache(struct kvm_vcpu *vcpu, u64 vmcb12_gpa)
+{
+	struct vcpu_svm *svm = to_svm(vcpu);
+	struct kvm_host_map map;
+	struct vmcb *vmcb12;
+	int r;
+
+	r = kvm_vcpu_map_readonly(vcpu, gpa_to_gfn(vmcb12_gpa), &map);
+	if (r)
+		return r;
+
+	vmcb12 = map.hva;
+	nested_copy_vmcb_control_to_cache(svm, &vmcb12->control);
+	nested_copy_vmcb_save_to_cache(svm, &vmcb12->save);
+	kvm_vcpu_unmap(vcpu, &map);
+	return 0;
+}
+
 int nested_svm_vmrun(struct kvm_vcpu *vcpu)
 {
 	struct vcpu_svm *svm = to_svm(vcpu);
 	int ret;
-	struct vmcb *vmcb12;
-	struct kvm_host_map map;
 	u64 vmcb12_gpa;
 	struct vmcb *vmcb01 = svm->vmcb01.ptr;
 
@@ -1144,22 +1170,17 @@ int nested_svm_vmrun(struct kvm_vcpu *vcpu)
 		return ret;
 	}
 
+	if (WARN_ON_ONCE(!svm->nested.initialized))
+		return -EINVAL;
+
 	vmcb12_gpa = svm->vmcb->save.rax;
-	if (kvm_vcpu_map(vcpu, gpa_to_gfn(vmcb12_gpa), &map)) {
+	if (nested_svm_copy_vmcb12_to_cache(vcpu, vmcb12_gpa)) {
 		kvm_inject_gp(vcpu, 0);
 		return 1;
 	}
 
 	ret = kvm_skip_emulated_instruction(vcpu);
 
-	vmcb12 = map.hva;
-
-	if (WARN_ON_ONCE(!svm->nested.initialized))
-		return -EINVAL;
-
-	nested_copy_vmcb_control_to_cache(svm, &vmcb12->control);
-	nested_copy_vmcb_save_to_cache(svm, &vmcb12->save);
-
 	/*
 	 * Since vmcb01 is not in use, we can use it to store some of the L1
 	 * state.
@@ -1180,11 +1201,9 @@ int nested_svm_vmrun(struct kvm_vcpu *vcpu)
 		svm->nmi_l1_to_l2 = false;
 		svm->soft_int_injected = false;
 
-		nested_svm_vmrun_error_vmexit(vcpu, vmcb12);
+		nested_svm_vmrun_error_vmexit(vcpu, vmcb12_gpa);
 	}
 
-	kvm_vcpu_unmap(vcpu, &map);
-
 	return ret;
 }
 
-- 
2.53.0.414.gf7e9f6c205-goog


  parent reply	other threads:[~2026-02-24 22:34 UTC|newest]

Thread overview: 44+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-02-24 22:33 [PATCH 00/31] Nested SVM fixes, cleanups, and hardening Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 01/31] KVM: nSVM: Avoid clearing VMCB_LBR in vmcb12 Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 02/31] KVM: SVM: Switch svm_copy_lbrs() to a macro Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 03/31] KVM: SVM: Add missing save/restore handling of LBR MSRs Yosry Ahmed
2026-03-03  0:02   ` Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 04/31] KVM: selftests: Add a test for LBR save/restore (ft. nested) Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 05/31] KVM: nSVM: Always inject a #GP if mapping VMCB12 fails on nested VMRUN Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 06/31] KVM: nSVM: Refactor checking LBRV enablement in vmcb12 into a helper Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 07/31] KVM: nSVM: Refactor writing vmcb12 on nested #VMEXIT as " Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 08/31] KVM: nSVM: Triple fault if mapping VMCB12 fails on nested #VMEXIT Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 09/31] KVM: nSVM: Triple fault if restore host CR3 " Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 10/31] KVM: nSVM: Drop nested_vmcb_check_{save/control}() wrappers Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 11/31] KVM: nSVM: Call enter_guest_mode() before switching to VMCB02 Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 12/31] KVM: nSVM: Make nested_svm_merge_msrpm() return an errno Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 13/31] KVM: nSVM: Call nested_svm_merge_msrpm() from enter_svm_guest_mode() Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 14/31] KVM: nSVM: Call nested_svm_init_mmu_context() before switching to VMCB02 Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 15/31] KVM: nSVM: Refactor minimal #VMEXIT handling out of nested_svm_vmexit() Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 16/31] KVM: nSVM: Unify handling of VMRUN failures with proper cleanup Yosry Ahmed
2026-02-25 22:44   ` Yosry Ahmed
2026-02-28  0:07   ` Sean Christopherson
2026-02-28  0:46     ` Yosry Ahmed
2026-02-28  1:05       ` Sean Christopherson
2026-03-02 16:02         ` Yosry Ahmed
2026-03-02 16:11           ` Sean Christopherson
2026-02-24 22:33 ` [PATCH v6 17/31] KVM: nSVM: Clear EVENTINJ field in VMCB12 on nested #VMEXIT Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 18/31] KVM: nSVM: Drop the non-architectural consistency check for NP_ENABLE Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 19/31] KVM: nSVM: Add missing consistency check for nCR3 validity Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 20/31] KVM: nSVM: Add missing consistency check for hCR0.PG and NP_ENABLE Yosry Ahmed
2026-03-03  0:00   ` Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 21/31] KVM: nSVM: Add missing consistency check for EFER, CR0, CR4, and CS Yosry Ahmed
2026-03-03  0:03   ` Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 22/31] KVM: nSVM: Add missing consistency check for EVENTINJ Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 23/31] KVM: SVM: Rename vmcb->nested_ctl to vmcb->misc_ctl Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 24/31] KVM: SVM: Rename vmcb->virt_ext to vmcb->misc_ctl2 Yosry Ahmed
2026-02-24 22:33 ` [PATCH v6 25/31] KVM: nSVM: Cache all used fields from VMCB12 Yosry Ahmed
2026-02-24 23:58   ` Yosry Ahmed
2026-02-24 22:34 ` Yosry Ahmed [this message]
2026-02-24 22:34 ` [PATCH v6 27/31] KVM: nSVM: Use PAGE_MASK to drop lower bits of bitmap GPAs from vmcb12 Yosry Ahmed
2026-02-24 22:34 ` [PATCH v6 28/31] KVM: nSVM: Sanitize TLB_CONTROL field when copying " Yosry Ahmed
2026-02-24 22:34 ` [PATCH v6 29/31] KVM: nSVM: Sanitize INT/EVENTINJ fields " Yosry Ahmed
2026-02-24 22:34 ` [PATCH v6 30/31] KVM: nSVM: Only copy SVM_MISC_ENABLE_NP from VMCB01's misc_ctl Yosry Ahmed
2026-02-24 22:34 ` [PATCH v6 31/31] KVM: selftest: Add a selftest for VMRUN/#VMEXIT with unmappable vmcb12 Yosry Ahmed
2026-03-03  0:13   ` Yosry Ahmed
2026-02-24 22:37 ` [PATCH 00/31] Nested SVM fixes, cleanups, and hardening Yosry Ahmed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260224223405.3270433-27-yosry@kernel.org \
    --to=yosry@kernel.org \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=pbonzini@redhat.com \
    --cc=seanjc@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox