[CVE-2026-98604][MODERATE 7.0] ovpn: replace bind when learning local endpoint [ Upstream

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

CVE: CVE-2026-98604
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: ovpn: replace bind when learning local endpoint [ Upstream
Commit: 21dbc9f4d0bdd3b22f91beabc7840064bf11594f
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=21dbc9f4d0bdd3b22f91beabc7840064bf11594f
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98604
Analysis date: Mon, 19 Oct 2026 00:01:47 -0400
ActionableScore: 5
ActionableScore lower bound: 4
Actionable bucket: Actionable Moderate / Strong Important candidate requiring manual review
Manual review required: YES

Summary:
A configured ovpn peer can race local IPv6 endpoint learning with UDP transmission, causing the host networking path to consume a torn source address and potentially disrupt the VPN tunnel, while broader routing or isolation consequences require manual review.

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

The original Linux kernel CVE announcement for CVE-2026-98604 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98604

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: YES

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

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

CVE-2026-98604	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-98604 CHECK WITH IMPACT FROM ORIG NN IMPORTANT | AEGIS IMPACT MODERATE_KPANIC AEGIS_AS=1 LEGACY=MODERATE_KPANIC RESULT=SAME (final reconciliation agrees; DECISION: both MODERATE_KPANIC)	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:H';CWE-362;CWE-662;Other CVSS 'AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:L';BEST CVSS score: '5.3';DESCR 'A race in the ovpn endpoint learning path can modify bind local in place while the UDP TX path reads the same RCU published object without peer lock. A concurrent IPv6 update can produce a torn address containing parts of the old and new local endpoints. This can select an invalid source address for outgoing UDP traffic and cause packet loss, routing failures, or disruption of the affected VPN tunnel. For the CVSS the PR:L value is used because reliable triggering is expected to require access as a configured VPN peer that can send traffic accepted for endpoint learning. The issue is network reachable through OpenVPN traffic, but AC:H applies because exploitation depends on winning a race between local endpoint learning and concurrent UDP transmission. Impact is primarily denial of service. The examined code does not provide a UAF, OOB access, arbitrary write, or another supported path to privilege escalation. Automated impact decision path. The original SimpleNet classifier selected IMPORTANT. The primary LLM estimated paranoid CVSS 5.3 and the production ActionableScore returned 5 with recommendation Actionable Moderate / Strong Important candidate requiring manual review. The production rules then evaluated CVSS thresholds and ActionableScore thresholds. Recorded rule transitions were CVSS and initial triage stage IMPORTANT to MODERATE_KPANIC because the CVSS score, vector or bug class threshold rule was applied. The legacy KP specialist produced a review artifact and its decision was processed by the applicable KP rules. The legacy FD or DH downgrade specialist has no review artifact for this run. Exact triggered rule markers were LOWERED_FROM_HIGH_BASED_ON_GUESSCVSS. The independent weaker Aegis shadow evaluated CVSS 2.7 and ActionableScore 1 with primitive class CLASS_C and manual review RECOMMENDED. Its independently reconciled impact was MODERATE_KPANIC. The shadow agreed with the final production impact and aggregation did not change that impact. The final reported impact is MODERATE_KPANIC.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 5)	MAYBE	OOB NETWORK SKB LINUS IPV6 KPANIC  KPANIC LOWERED_FROM_HIGH_BASED_ON_GUESSCVSS 	-	-	checked

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

ActionableScore=5
ActionableScoreLower=4
PrimitiveClass=CLASS_B
RealUAF=NO
HostAdminRequired=NO
NamespaceScopedPrivilege=NO
ImportantCandidate=YES
ManualReview=REQUIRED
AutoDowngradeAllowed=NO
RequiresHostUID0=NO
AuthorizationDecisionBypass=NO
ProtectedResourceImpact=NONE
B4PendingReview=NO
B4PendingConfirmed=NO

1. ActionableScore

* Raw conservative score: 4
* Conservative score: 4. No floor or cap applied.
* Raw paranoid score: 5
* Paranoid score: 5. No floor or cap applied.
* Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**::

2. Signal breakdown

Conservative signals:

* +1 active structured-state corruption. The writer replaces the 128-bit IPv6 value in `bind->local.ipv6` in place while an RCU reader can observe a mixture of old and new words.
* +1 attacker-influenced data-plane state. An accepted peer packet supplies the destination address used for local endpoint learning, so a configured peer can materially influence the value involved in the race.
* +1 privileged downstream consumption. Kernel UDP TX consumes the incoherent local endpoint as a source-address selection input after the unlocked read.
* +1 realistic availability impact. A torn source address can cause packet loss, routing failure, or disruption of the affected VPN tunnel.

Paranoid-only signal:

* +1 unresolved V33 trust-boundary consequence. Attacker-influenced active endpoint state is consumed by the privileged host networking path, while the full consequences for routing, filtering, and namespace or delegated-interface boundaries are not visible in the supplied call sites.

Material signals not awarded:

* No UAF, OOB access, heap overwrite, arbitrary write, controlled reclaim, or callback influence is established.
* No kernel panic or host-wide DoS point is awarded. The demonstrated consequence is tunnel or peer disruption.
* No confidentiality or protected-state integrity point is awarded. The evidence establishes an incoherent endpoint address, not disclosure or unauthorized modification of another principal data.
* No privilege-escalation point is awarded.
* No race penalty is applied merely because timing is involved. Repeated peer traffic and concurrent TX may provide repeated attempts, and the evidence does not establish a one-shot uncontrollable window.
* Direct unauthenticated remote reachability is not established. The evaluated attacker is an accepted or configured VPN peer.
* NEW52 B5-R does not apply because there is no heap OOB write or retry-accumulated index.
* NEW53 B3-PENDING does not apply because no shared or file-backed storage write is present.
* NEW54 B4-PENDING does not apply because no authorization guard changes from DENY to ALLOW.

3. Reachability analysis

The credible attacker is a configured remote OpenVPN peer able to send packets accepted by the ovpn data path. Such packets can influence the learned local endpoint through the packet destination address while concurrent UDP TX reads the published bind object.

The attacker does not require host UID 0 or host-administrator powers for the runtime trigger once the peer and tunnel have been configured. Prior privileged setup is required to create and configure the ovpn interface and peer, but that setup can be performed by a separate administrator and does not make the configured peer a host administrator.

No namespace capability is required by the evaluated remote-peer trigger. The supplied evidence does not establish a separate attacker-created namespace path. No special hardware, boot option, or unusual device is required.

The race requires overlap between local endpoint learning and UDP transmission. It is timing-dependent, but repeated network traffic can plausibly provide multiple attempts. The evidence does not establish reliable host-wide failure.

4. Primitive and severity interpretation

The exact primitive is an attacker-influenced torn read of active structured endpoint state. The old implementation modified a multiword IPv6 address inside an RCU-published `ovpn_bind` object while UDP TX could read it without `peer->lock`. The write itself remains within the valid object. The invalid state arises because the reader can observe a non-atomic mixture of two valid addresses.

This is broader than passive diagnostic metadata because the resulting address is consumed by the host UDP transmission path. It is not a lifetime violation, stale post-free access, or memory-safety corruption. There is no repeated release or refcount mutation.

PrimitiveClass is CLASS_B because a concrete attacker-influenced active-state consistency primitive reaches a privileged networking consumer across the peer-to-host data-plane boundary. The demonstrated consequence is malformed source-address selection and VPN availability disruption. Effects on broader routing, filtering, or isolation boundaries remain unresolved rather than demonstrated.

ImportantCandidate is YES under V33 because the active structured state, downstream privileged consumer, and peer-to-host trust boundary are independently identifiable. This is a review candidate and not proof of Important severity, host panic, or privilege escalation.

AutoDowngradeAllowed is NO because automatically treating the issue as ordinary packet loss would discard the unresolved active-state and trust-boundary analysis. The likely operational severity remains Actionable Moderate unless further call-site analysis establishes a broader routing or isolation consequence.

5. One-sentence report phrase

A configured ovpn peer can race local IPv6 endpoint learning with UDP transmission, causing the host networking path to consume a torn source address and potentially disrupt the VPN tunnel, while broader routing or isolation consequences require manual review.

6. Manual review recommendation

MANUAL CHECK REQUIRED

The patch establishes attacker-influenced incoherent active endpoint state consumed by privileged UDP TX, but the supplied context does not resolve all downstream routing, filtering, and namespace-boundary effects.

7. Call-site confidence

Call-site confidence: medium

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

Patch: ovpn: replace bind when learning local endpoint [ Upstream
Commit: 21dbc9f4d0bdd3b22f91beabc7840064bf11594f
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=21dbc9f4d0bdd3b22f91beabc7840064bf11594f

Changed files:
  drivers/net/ovpn/peer.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=21dbc9f4d0bdd3b22f91beabc7840064bf11594f

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

The original Linux kernel CVE announcement for CVE-2026-98604 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98604

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:N/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:L
  The Best / paranoid CVSS vector: AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:H
  The Best / paranoid CVSS score: 5.3

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).