[CVE-2026-98015][MODERATE 7.0] net/mlx5: E-Switch: fix use-after-free in mlx5_eswitch_termtbl_put [ Upstream

[CVE-2026-98015][MODERATE 7.0] net/mlx5: E-Switch: fix use-after-free in mlx5_eswitch_termtbl_put [ Upstream

[AL-KERNEL]

CVE: CVE-2026-98015
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: net/mlx5: E-Switch: fix use-after-free in mlx5_eswitch_termtbl_put [ Upstream
Commit: 304ea40ef4561671503c36e47da44a0a31b71e36
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=304ea40ef4561671503c36e47da44a0a31b71e36
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98015
Analysis date: Fri, 25 Sep 2026 09:15:15 -0400
ActionableScore: 5
ActionableScore lower bound: unknown
Actionable bucket: Actionable Moderate / Strong Important candidate requiring manual review
Manual review required: YES

Summary:
No one-sentence report phrase was found.

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

The original Linux kernel CVE announcement for CVE-2026-98015 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98015

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-2026-98015	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-98015 CHECK WITH IMPACT FROM ORIG NN IMPORTANT	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H';*CWE-416;CWE-362;CWE-415;Other CVSS 'AV:L/AC:H/PR:H/UI:N/S:U/C:L/I:L/A:H';BEST CVSS score: '7';DESCR 'A race in mlx5_eswitch_termtbl_put can access a termination table handle after another caller has reduced its reference count to zero and freed it. Two concurrent callers can also both observe the zero reference count and attempt to delete the same flow rule, destroy the same flow table, and free the same object. This creates a use-after-free and a plausible double cleanup or double-free condition that may crash the kernel and could potentially support privilege escalation through kernel memory corruption. For the CVSS the PR:L value is used for the paranoid score because mlx5 E-Switch control may be delegated to a container root or a service account with network administration capabilities, even though normal host configuration usually requires higher administrative privileges. AC:H applies because reliable triggering depends on winning a concurrency window between two cleanup operations. The issue is not directly network reachable through ordinary packet traffic and normally requires local access to the mlx5 E-Switch control plane. Impact is at least a local denial of service via kernel crash and in the worst case may include confidentiality and integrity compromise due to the plausible double-free condition.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 5)  SKIP CVE-2026-98015 UNKNOWN SKIP No affected files built, so skip this CVE NO  - - unknown 	MAYBE	REMOTE RACE UAF NETWORK KPANIC  KPANIC INCREASED_TO_HIGH_BASED_ON_GUESSCVSS DECREASED_TO_MODERATE7_BASED_ON_FALSEPOSCHECKOFDH	-	-	checked

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

ActionableScore=5
PrimitiveClass=CLASS_B
RealUAF=YES
HostAdminRequired=UNKNOWN
NamespaceScopedPrivilege=UNKNOWN
ImportantCandidate=YES
ManualReview=REQUIRED
AutoDowngradeAllowed=NO

1. ActionableScore

* Raw conservative score: 4
* Conservative score: 5
* Raw paranoid score: 5
* Paranoid score: 5
* Applied conservative floor: 4 -> 5 under the REAL UAF FLOOR 5 rule because the same handle can undergo concrete duplicate destruction and free operations.
* Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**::

2. Signal breakdown

* +2 Real UAF and lifetime violation. One caller can free `tt` while another caller subsequently reads `tt->ref_count`.
* +1 Concrete repeated release operation. Two callers can both observe the zero reference count and execute destruction followed by `kfree(tt)` on the same handle.
* +1 Reliable kernel crash or strong DoS. Duplicate flow-rule deletion, flow-table destruction, or object free can realistically fault or corrupt allocator state.
* Paranoid +1 Unresolved post-lifetime mutation. A second destructor and `kfree()` can operate on the exact object after its lifetime may already have ended.
* No race penalty. The patch establishes concurrency but does not establish that the window is one-shot, exceptionally narrow, or difficult to repeat.
* No privilege penalty. mlx5 E-Switch configuration may require substantial privileges, but the supplied evidence does not positively establish the complete trigger as host-admin-only.
* Controlled reclaim and attacker-shaped replacement contents are not demonstrated.
* Arbitrary or directed kernel read/write is not demonstrated.
* Structured stale-object leverage is not separately awarded. The stale `rule` and `termtbl` fields are consumed by cleanup helpers, but attacker influence over reclaimed contents is not established.
* Weak LPE is not separately awarded because the unresolved concern is already represented by the concrete repeated destruction and free.
* No additional Availability point is awarded because it would duplicate the same crash consequence.
* Confidentiality and integrity points are not awarded merely because duplicate destruction could theoretically become exploitable.
* Remote reachability through ordinary network traffic is not established.

Repeated-release checks:

* First release can end lifetime: yes. The first completed cleanup calls `kfree(tt)`.
* Exact reference survives: yes. The competing caller retains the same `tt` pointer after releasing `termtbl_mutex`.
* Second lifetime operation is concrete: yes. Both callers can enter the zero-reference cleanup and invoke destructors and `kfree()` for the same object.
* Concrete repeated release/refcount +1 awarded: yes.
* Paranoid repeated-lifetime leverage +1 awarded: yes, because the second free is a real lifetime mutation against storage whose original lifetime may already have ended.
* FLOOR 5 applied: yes, to the conservative raw score.

3. Reachability analysis

The trigger requires two concurrent `mlx5_eswitch_termtbl_put()` calls for the same termination-table handle. This is an mlx5 E-Switch offload control-path issue rather than a directly packet-triggered networking flaw.

The supplied context does not identify the complete userspace call chain or capability checks. It therefore does not prove either ordinary-user reachability or a host-admin-equivalent requirement. `HostAdminRequired` and `NamespaceScopedPrivilege` remain `UNKNOWN`. Delegation to a network-management service or container is possible but is not established for the complete operation.

An mlx5 device and an active E-Switch offload configuration are practical prerequisites. The hardware dependency is not scored as rare or physical-only because mlx5 hardware is commonly deployed in server environments. No prior privileged setup penalty is applied because the required provisioning boundary is unresolved.

The bug is a race, but no evidence establishes materially uncontrollable timing. Concurrent final puts may occur naturally during overlapping rule teardown, so no hard-race penalty is justified.

4. Primitive and severity interpretation

The exact primitive is an active structured-object UAF combined with possible duplicate destruction and double free. It is not fixed stale storage.

This is stronger than an ordinary cleanup stale dereference. In the ordinary case, the losing caller would only read freed memory and fault. Here, both callers can decide to perform final cleanup, causing duplicate `mlx5_del_flow_rules()`, duplicate `mlx5_destroy_flow_table()`, and duplicate `kfree()` operations on the same object.

The first cleanup can destroy and free the handle. The competing caller retains the exact same pointer and can perform a second lifetime-ending operation. This is concrete repeated lifetime corruption, but it remains `CLASS_B` because controlled reclaim, refcount takeover, directed corruption, callback control, and arbitrary kernel read/write are not demonstrated.

The demonstrated or directly supported consequence is kernel crash and allocator or object-lifetime corruption. Privilege escalation is a technically possible but unproven consequence. The unresolved duplicate-free behavior makes Important-vs-Moderate review necessary, so `ImportantCandidate=YES` and `AutoDowngradeAllowed=NO`.

Likely operational severity is Actionable Moderate at minimum, with manual review needed to determine whether concurrent teardown is reachable below host administration and whether reclaim between the two cleanup paths provides useful exploitation leverage.

5. One-sentence report phrase

A race in `mlx5_eswitch_termtbl_put()` allows concurrent callers to access and potentially destroy the same freed termination-table handle twice, creating a kernel UAF and double-free condition with at least denial-of-service impact.

6. Manual review recommendation

MANUAL CHECK REQUIRED

The concrete duplicate destruction and free leave unresolved post-lifetime corruption leverage, while attacker privilege, triggerability, and reclaim control require call-site analysis.

7. Call-site confidence

Call-site confidence: medium

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

Patch: net/mlx5: E-Switch: fix use-after-free in mlx5_eswitch_termtbl_put [ Upstream
Commit: 304ea40ef4561671503c36e47da44a0a31b71e36
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=304ea40ef4561671503c36e47da44a0a31b71e36

Changed files:
  drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads_termtbl.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=304ea40ef4561671503c36e47da44a0a31b71e36

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

The original Linux kernel CVE announcement for CVE-2026-98015 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98015

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:H/UI:N/S:U/C:L/I:L/A:H
  The Best / paranoid CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H
  The Best / paranoid CVSS score: 7

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: unknown
  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).