[CVE-2026-89914][MODERATE REGULAR] KVM: arm64: Sign-extend VA for range-based TLBI invalidation

[CVE-2026-89914][MODERATE REGULAR] KVM: arm64: Sign-extend VA for range-based TLBI invalidation

[AL-KERNEL]

CVE: CVE-2026-89914
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: KVM: arm64: Sign-extend VA for range-based TLBI invalidation
Commit: 72bce82c4171bf330919ff1b64dc0a36c254ec7d
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=72bce82c4171bf330919ff1b64dc0a36c254ec7d
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89914
Analysis date: Wed, 16 Sep 2026 07:19:16 -0400
ActionableScore: 4
ActionableScore lower bound: 3
Actionable bucket: Borderline / manual review recommended
Manual review required: YES

Summary:
No one-sentence report phrase was found.

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

The original Linux kernel CVE announcement for CVE-2026-89914 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89914

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 REGULAR
Manual review required: YES

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

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

CVE-2026-89914	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-89914 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-682;CWE-681;CWE-693;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:L';BEST CVSS score: '5.8';DESCR 'KVM/arm64 range based TLBI decoding could use a non sign extended virtual address after the helper was reused for Stage 1 invalidations. This may cause KVM to invalidate the wrong address range and leave stale TLB translations active for a guest or nested guest. For the CVSS the PR:L is used because the realistic attacker needs local ability to run or control a guest or nested guest that can issue the relevant TLBI operation, but full host administrator privileges are not required. The issue is not directly network reachable and would normally be exercised through virtualization control or guest execution rather than packet traffic. Impact is at least a guest correctness or availability problem. In the paranoid view it may allow limited confidentiality or integrity impact through stale translations, so manual review is preferred over automatic closure.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Borderline / manual review recommended (with actual score 4)  SKIP CVE-2026-89914 UNKNOWN SKIP The Fixes patch not applied yet, so unlikely that actual: url YES  NO NO unknown 	MAYBE	VIRT SIMPLEFIX HARDWARE LINUS 	-	-	checked

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

ActionableScore=4
ActionableScoreLower=3
PrimitiveClass=CLASS_B
RealUAF=NO
HostAdminRequired=NO
NamespaceScopedPrivilege=NO
ImportantCandidate=NO
ManualReview=RECOMMENDED
AutoDowngradeAllowed=NO

1. ActionableScore

* Raw conservative score: 3
* Conservative score: 3. No floor or cap changed the score.
* Raw paranoid score: 4
* Paranoid score: 4. No floor or cap changed the score.
* Final recommended bucket: **Borderline / manual review recommended**::

2. Signal breakdown

Conservative signals:

* +2. CLASS_B active translation invalidation consistency primitive. The patch establishes that range-based TLBI decoding used a non sign-extended VA, so KVM could invalidate the wrong address range and leave live stale translations.
* +1. Active translation synchronization concern. The affected state is operational TLB or translation state, not passive metadata, and stale translations may continue to affect execution after the intended invalidation.
* +0. Reliable host kernel crash is not established by the supplied patch.
* +0. Remote reachability is not established.
* +0. Confidentiality and integrity impact are not awarded conservatively because no concrete unauthorized read, write, or boundary-crossing memory access primitive is demonstrated.

Paranoid signals:

* +1. Limited unresolved stale-translation impact. In a KVM/MMU context, wrong TLBI range invalidation may plausibly affect guest or nested guest correctness and potentially limited isolation semantics, but the evidence does not establish arbitrary host read/write or guest-to-host compromise.

Non-awarded material signals:

* Real UAF not awarded. The patch does not show freed host memory, stale heap object use, RCU lifetime break, or repeated release/refcount mutation.
* Strong exploitation leverage not awarded. There is no controlled reclaim, controlled object replacement, callback influence, directed host kernel write, or demonstrated guest-to-host memory corruption.
* KVM guest-to-host strong leverage not awarded. Guest-influenced KVM state plus stale translation behavior is not by itself a concrete guest-to-host exploitation primitive.
* Extra availability not awarded. The supplied evidence does not show a reliable host panic or persistent host-wide outage.
* CLASS_A not used. The primitive is an active translation consistency/invalidation bug, not a demonstrated arbitrary write, authorization bypass, protected resource theft, or LPE.

3. Reachability analysis

The evaluated attacker is a KVM guest or nested-virtualization actor that can cause the relevant TLBI operation to be handled by host KVM, or a local VMM process controlling guest execution. This is below host-admin in the relevant security model, so HostAdminRequired=NO.

This is not a network-triggered issue. It is exercised through virtualization execution or virtualization control paths, not packet traffic.

NamespaceScopedPrivilege=NO because the supplied evidence is about KVM/arm64 TLBI handling, not user or network namespace scoped capabilities.

Prior host setup is likely required in the practical sense that KVM and nested virtualization support must be present and a VM must exist, but the vulnerable operation itself is not shown to require host root once the guest/VMM context is available.

No special rare physical hardware requirement is established beyond arm64 KVM support and the relevant nested/TLBI feature path. No hard timing race is established.

4. Primitive and severity interpretation

The exact primitive is incorrect decoding of a range-based TLBI base address due to missing sign extension from the 48-bit VA/IPA domain. This can cause invalidation to target the wrong range and leave stale translations active.

This is not fixed stale storage and not a heap lifetime UAF. It is an active translation-state consistency issue in KVM/MMU code.

No concrete repeated release/refcount operation exists. No first lifetime operation freeing an object and no second mutation through the same dangling reference are established.

PrimitiveClass=CLASS_B because the bug affects active translation retirement/invalidation semantics. It is stronger than a passive arithmetic bug, but weaker than a demonstrated memory corruption or guest-to-host exploit primitive.

Demonstrated consequence is wrong or incomplete TLBI invalidation and possible stale translations. Speculative consequence is limited confidentiality or integrity impact through stale translations, but this remains unresolved and is not enough for CLASS_A.

ImportantCandidate=NO because no strong exploitation primitive, LPE, arbitrary R/W, or demonstrated guest-to-host boundary compromise is established. Manual review is still recommended because KVM/MMU stale translation bugs can be security-sensitive and are easy to underestimate.

AutoDowngradeAllowed=NO because automatic closure could miss unresolved security impact from stale operational translation state in a virtualization boundary. This does not mean Strong Important. It means manual review is preferred before downgrading.

Likely operational severity is Borderline to Actionable Moderate, depending on whether further review shows practical stale-translation effects beyond guest correctness or availability.

5. One-sentence report phrase

KVM arm64 range-based TLBI decoding missed VA sign extension, which can cause invalidation of the wrong range and leave stale translations active, requiring manual review because KVM/MMU stale translation effects may exceed a simple correctness bug.

6. Manual review recommendation

MANUAL CHECK RECOMMENDED

The patch affects KVM/MMU TLBI invalidation semantics and stale active translations, but supplied evidence does not establish UAF, arbitrary host memory access, privilege escalation, or a reliable host crash.

7. Call-site confidence

Call-site confidence: medium

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

Patch: KVM: arm64: Sign-extend VA for range-based TLBI invalidation
Commit: 72bce82c4171bf330919ff1b64dc0a36c254ec7d
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=72bce82c4171bf330919ff1b64dc0a36c254ec7d

Commit description:

When the decode_range_tlbi() helper was moved to be used for S1 TLBIs,
the required sign extension was omitted. Add it.

As a result, special care must be taken to not overflow PA bits when
this is used for S2 invalidation.

Fixes: 85bba00 ("KVM: arm64: nv: Move TLBI range decoding to a helper")
Reported-by: sashiko-bot@kernel.org
Link: https://lore.kernel.org/r/20260801130337.EB2BA1F00AC4@smtp.kernel.org
Signed-off-by: Marc Zyngier <maz@kernel.org>
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20260806091026.620700-7-maz@kernel.org
Signed-off-by: Oliver Upton <oupton@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Changed files:
  arch/arm64/include/asm/kvm_nested.h
  arch/arm64/kvm/sys_regs.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=72bce82c4171bf330919ff1b64dc0a36c254ec7d

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

The original Linux kernel CVE announcement for CVE-2026-89914 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89914

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:N/I:N/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: 3
  Paranoid ActionableScore: 4

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: NO
Published priority for this report (same as in Subject): MODERATE REGULAR

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).