[CVE-2026-89929][MODERATE 7.0] KVM: nVM: Ensure INVVPID is emulated on the correct physical CPU

[AL-KERNEL <alkernel@kernelcve.org>]

CVE: CVE-2026-89929
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: KVM: nVM: Ensure INVVPID is emulated on the correct physical CPU
Commit: d03e721801ac54b9fdf05668aa0c069f07913b68
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=d03e721801ac54b9fdf05668aa0c069f07913b68
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89929
Analysis date: Wed, 16 Sep 2026 07:44:13 -0400
ActionableScore: 5
ActionableScore lower bound: 4
Actionable bucket: Actionable Moderate
Manual review required: NO

Summary:
KVM nested VMX may emulate INVVPID on the wrong physical CPU, leaving stale VPID TLB entries active for L2 and causing a nested virtualization translation correctness or isolation failure without evidence of direct host memory corruption.

======================================================================
ABOUT THIS REPORT
======================================================================

The original Linux kernel CVE announcement for CVE-2026-89929 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89929

The original announcement does not normally provide a security severity
estimate, CVSS assessment, or enough information to determine whether the
reported kernel bug represents a practically relevant security issue.

This report was generated by AL-KERNEL, an AI-assisted Linux kernel
vulnerability analysis system developed by Alexander Larkin. It combines
an autonomous classifier with LLM-assisted technical analysis and a
separate ActionableScore mechanism.

The purpose of this report is to prioritize Linux kernel CVEs before
manual review, identify cases that require prompt investigation, and
support automatic closure of issues that are unlikely to have meaningful
security impact.

Published priority for this report: MODERATE 7.0
Manual review required: NO

A detailed explanation of the methodology and priority rules is included
at the end of this message.

======================================================================
AL-KERNEL CLASSIFICATION RESULT
======================================================================

CVE-2026-89929	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-89929 CHECK WITH IMPACT FROM ORIG NN MODERATE	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:H';CWE-362;CWE-693;*CWE-664;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:L';BEST CVSS score: '5.8';DESCR 'KVM nested VMX can emulate INVVPID on a different physical CPU than the one where L2 last ran. This can leave stale VPID tagged TLB entries on the original CPU after L1 believes the invalidation completed, causing L2 to run with stale address translations. For the CVSS the PR:L because an attacker needs local ability to run or control a KVM guest with nested VMX rather than only sending network traffic. The issue is not network reachable. Impact is mainly guest isolation or correctness failure through stale translations, with possible limited confidentiality or integrity impact inside the affected VM and possible availability impact if the guest is destabilized. It does not by itself indicate host memory corruption or a host escape primitive.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate (with actual score 5)	YES	VIRT INIT HARDWARE  KPANIC INCREASED_TO_MODERATE70_BASED_ON_ACTIONABLESCOREHIGHEREQTHAN5	NO	NO	checked

======================================================================
ACTIONABLESCORE ANALYSIS
======================================================================

ActionableScore=5
ActionableScoreLower=4
PrimitiveClass=CLASS_B
RealUAF=NO
HostAdminRequired=NO
NamespaceScopedPrivilege=NO
ImportantCandidate=NO
ManualReview=RECOMMENDED
AutoDowngradeAllowed=NO

## 1. ActionableScore

* Raw conservative score: 4
* Conservative score: 4
* Raw paranoid score: 5
* Paranoid score: 5
* Final recommended bucket: **Actionable Moderate**::

No floor or cap changed the conservative score.
No floor or cap changed the paranoid score.

## 2. Signal breakdown

Conservative signals:

* +1 Active-state invalidation bypass. KVM may execute INVVPID on a pCPU different from the pCPU where L2 last ran, leaving live VPID-tagged TLB entries uninvalidated.
* +1 Active translation synchronization signal. The fix relies on detecting the pCPU mismatch and forcing a full VPID flush on the next nested VM-Enter, which shows stale hardware translation state can remain operational.
* +1 Local below-host-admin KVM reachability. The relevant actor is a local VMM or guest-control context with `/dev/kvm` and nested VMX access, not necessarily host root.
* +1 Limited virtualization correctness or isolation impact. Stale TLB translations can cause L2 to continue using translations after L1 believes invalidation completed.

Paranoid additional signal:

* +1 Unresolved nested virtualization isolation impact. The stale translation state is in a security-sensitive KVM MMU or TLB path, and the exact consequence beyond correctness failure may require manual review.

Non-awarded signals:

* RealUAF not awarded. The patch does not establish use-after-free or stale heap object use.
* Controlled reclaim not demonstrated.
* Arbitrary host kernel read or write not demonstrated.
* Guest-to-host exploitation leverage not awarded. The evidence shows stale TLB state, not a concrete host memory corruption or host escape primitive.
* Remote reachability not established. This is not packet-triggered.
* Race penalty not applied. The scenario involves CPU migration and ordering, but the prompt rules say not to penalize merely because timing or migration is involved.
* Extra availability point not awarded. No host panic, reliable crash, or persistent host-wide DoS is shown.
* Class A not used because there is no demonstrated controlled corruption, authorization bypass, privileged resource theft, or LPE.

## 3. Reachability analysis

The complete practical trigger requires a KVM nested-VMX setup. The evaluated attacker is best modeled as a local VMM or local user/process with access to `/dev/kvm` that can run or control an L1 guest and nested L2 execution.

A malicious L1 can execute INVVPID while the vCPU migrates between host pCPUs. The bug is not established as guest-only from L2, because L2 does not itself perform the host-side INVVPID emulation path in the same way. It is also not a direct remote network vulnerability.

Host administrator privilege is not positively required by the supplied evidence, because KVM access can be delegated to non-root local users through common `/dev/kvm` permissions. However, some deployments restrict `/dev/kvm` to trusted users, so the operational exposure depends on local KVM access policy.

Namespace-scoped privilege does not apply. This is not a netns or container capability issue.

Prior setup exists in the sense that nested VMX and KVM must be available and the attacker must be able to run a nested virtualization workload. No special rare hardware beyond Intel VMX/VPID-capable nested virtualization is established as a strong reducing factor.

## 4. Primitive and severity interpretation

The primitive is an active translation-state synchronization failure. KVM invalidates VPID-tagged translations on the wrong physical CPU, so stale TLB entries on the CPU where L2 previously ran can remain active.

This is not a UAF, double free, refcount bug, or heap corruption primitive. It is also not an ordinary passive bookkeeping issue, because the stale state is hardware translation state that can still affect later L2 execution.

PrimitiveClass is CLASS_B because the bug affects live KVM translation state and a required invalidation protocol, but no controlled host-memory corruption or host escape primitive is demonstrated.

Demonstrated consequence: stale L2 translations after L1 believes INVVPID completed.
Speculative consequence: limited confidentiality, integrity, or availability impact inside the nested virtualization context if stale translations allow execution or access under outdated mappings.
Not demonstrated: arbitrary host read/write, host privilege escalation, guest-to-host escape, or controlled replacement of host objects.

ImportantCandidate is NO because the evidence does not establish a strong Important-grade primitive. ManualReview is RECOMMENDED because KVM translation invalidation bugs can affect isolation semantics and should not be automatically dismissed from patch wording alone.

AutoDowngradeAllowed is NO because the issue is an actionable KVM active-translation synchronization failure, not a simple warning-only or cosmetic bug.

Likely operational severity: Actionable Moderate. It should be reviewed by KVM/security maintainers or virtualization security owners, especially where untrusted users can run nested virtualization.

## 5. One-sentence report phrase

KVM nested VMX may emulate INVVPID on the wrong physical CPU, leaving stale VPID TLB entries active for L2 and causing a nested virtualization translation correctness or isolation failure without evidence of direct host memory corruption.

## 6. Manual review recommendation

MANUAL CHECK RECOMMENDED

Reason: the bug is not a UAF or proven host escape, but it affects live KVM TLB invalidation semantics in nested virtualization and stale translations may have security-relevant consequences depending on the L1/L2 threat model.

## 7. Call-site confidence

Call-site confidence: medium

======================================================================
UPSTREAM PATCH SUMMARY
======================================================================

Patch: KVM: nVM: Ensure INVVPID is emulated on the correct physical CPU
Commit: d03e721801ac54b9fdf05668aa0c069f07913b68
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=d03e721801ac54b9fdf05668aa0c069f07913b68

Commit description:

When emulating INVVPID, KVM executes INVVPID on the physical CPU using
vpid02 (instead of the L1 assigned VPID), after doing some validations
on the operands. However, it is possible that the physical CPU KVM
executes INVVPID on is different from the CPU L2 is running on.

For example, in the following scenario:
- L2 runs on CPU #1 and exits to L1 (vmx->nested.vmcs02.cpu=1)
- L1 migrates to CPU #2 and executes INVVPID
- KVM executes INVVPID on CPU #2
- L1 migrates back to CPU #1 and runs L2 (vmx->nested.vmcs02.cpu=1)

The TLB entries on CPU #1 are never invalidated, because INVVPID was
executed on CPU #2, and vmcs02 never ran on a different pCPU (i.e.
vmx_vcpu_load_vmcs() will *not* request KVM_REQ_TLB_FLUSH).

Ensure that INVVPID is being executed on the same pCPU that L2 last ran
on, and if not, fallback to clearing last_vpid=0 to trigger a full VPID
flush on the next nested VM-Enter (as KVM will detect L1 using a
different VPID for L2). If L2 ends up running on a different pCPU, KVM
will flush the TLB anyway through vmx_vcpu_load_vmcs().

Cc: stable@vger.kernel.org
Signed-off-by: Yosry Ahmed <yosry@kernel.org>
Reviewed-by: Kai Huang <kai.huang@intel.com>
Link: https://patch.msgid.link/20260616214652.2157032-4-yosry@kernel.org
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Changed files:
  arch/x86/kvm/vmx/nested.c

Diff excerpt:

Not included in this email. See the upstream URL for the full patch.

Full patch:
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=d03e721801ac54b9fdf05668aa0c069f07913b68

======================================================================
DETAILED REPORT METHODOLOGY
======================================================================

The original Linux kernel CVE announcement for CVE-2026-89929 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89929

The original CVE announcement normally does not include a security-level
estimate. In particular, it may not contain a CVSS assessment, an impact
level, or enough information to determine whether the reported bug is a
practically relevant security issue. One purpose of this parallel CVE list
is to provide that missing technical and prioritization information.

The original goal of the AL-KERNEL project was to prioritize Linux kernel
CVE analysis automatically before manual review. The system can also help
identify non-security issues that may be suitable for automatic closure.

This report was generated by AL-KERNEL, an AI-assisted Linux kernel
vulnerability analysis system developed by Alexander Larkin.

The first analysis stage combines an autonomous classifier with additional
LLM-based analysis. The autonomous classifier runs locally on a CPU and is
based on a backpropagation neural network. Together, these mechanisms
produce a technical vulnerability description, identify likely weakness
types, estimate CVSS severity, and provide input for ActionableScore.

Two CVSS estimates are retained because incomplete kernel vulnerability
information often permits more than one defensible interpretation:

  Conservative CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:L
  The Best / paranoid CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:H
  The Best / paranoid CVSS score: 5.8

The conservative vector represents a lower-impact interpretation.
The Best/paranoid vector intentionally represents a plausible upper-bound
interpretation and should not automatically be treated as demonstrated
real-world impact.

CVSS may also need to be adjusted for a particular Linux deployment,
because actual reachability, privileges, enabled kernel configuration,
hardware, namespaces, exposed device nodes, and other environmental
conditions can differ significantly between systems.

A separate ActionableScore mechanism evaluates practical remediation
urgency. Its analysis may include reachability, attack prerequisites,
subsystem exposure, memory-corruption characteristics, denial-of-service
reliability, and possible confidentiality, integrity, or
privilege-escalation impact.

  Conservative ActionableScore: 4
  Paranoid ActionableScore: 5

The final base severity is taken directly from the second tab-separated
field of the AL-KERNEL classification result. ActionableScore does not
replace or independently override that final AL-KERNEL decision, and
if ActionableScore adjusted impact level of ALKERNEL, then you would see
self-readable flags above like INCREASED_TO_HIGH_BASED_ON_ACTIONABLESCOREHIGHEREQTHAN7.

For an AL-KERNEL result of MODERATE, this report uses the following
additional presentation split:

  ActionableScore below 5   -> MODERATE REGULAR
  ActionableScore 5 or more -> MODERATE 7.0

The distinction between MODERATE REGULAR and MODERATE 7.0 makes it
possible to identify Moderate issues that should receive manual analysis
and fixes before lower-priority MODERATE REGULAR issues. In many cases,
MODERATE REGULAR fixes may wait for a later rebase or routine update.

There is one override in which MODERATE REGULAR becomes MODERATE 7.0
even when the ActionableScore is below 5. When the AL-KERNEL result
contains the KPANIC flag, a MODERATE result is always presented as
MODERATE 7.0. The KPANIC flag selected with few regexps without
usage of AI at all, so it helps to detect cases when Kernel Crash happens
and similar (to filter False-Negative results from the LLM usage).

KPANIC indicates that a reliable kernel crash, kernel panic, or similarly
serious kernel availability impact was identified by the classification
workflow.

AL-KERNEL base severity for this report: MODERATE
KPANIC detected for this report: YES
Published priority for this report (same as in Subject): MODERATE 7.0

These results are intended to support engineering triage. They are
machine-generated estimates, and cases marked for manual review should
be validated by a human security engineer before final disposition.
For more info read docs linked from here: https://kernelcve.org/
(and you can submit you own patch there to generate such a report
for non-existant CVE-id yet).

Note that in many cases this AI tool selects higher severity, than
real is (means you can expect Importants instead of Moderate 7.0 or
Moderates 7.0 instead of regular Moderates). If you see such cases,
please use reply email interface to add additional manual analyses
info to this particular CVE.
And please, please, let me know when you see Lows instead of Importants
or Important instead of Low (because particular for such cases I
need to tune this AI tool to make it better for this one and next similar).
My contact email for such notifications is alexanjelausa@gmail.com (and both
send reply to CVE record itself too and see "reply" button below for howto reply).