[CVE-2026-93250][MODERATE 7.0] vxlan: mdb: Fix use-after-free in vxlan_mdb_flush() [ Upstream

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

CVE: CVE-2026-93250
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: vxlan: mdb: Fix use-after-free in vxlan_mdb_flush() [ Upstream
Commit: f8a9b988e7a7bc674e74e8c901794d792c35689e
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=f8a9b988e7a7bc674e74e8c901794d792c35689e
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93250
Analysis date: Thu, 24 Sep 2026 15:00:07 -0400
ActionableScore: 4
ActionableScore lower bound: unknown
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-2026-93250 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93250

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-2026-93250	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-93250 CHECK WITH IMPACT FROM ORIG NN MODERATE	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H';*CWE-416;*CWE-664;*CWE-672;Other CVSS 'AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '7.8';DESCR 'A use after free in vxlan_mdb_flush occurs when flushing remotes of a star,G MDB entry removes a related S,G entry that was saved as the next element by hlist_for_each_entry_safe. The subsequent iteration operates on the stale entry and can write list poison values through freed list pointers, resulting in kernel memory corruption or a crash. The issue is reachable through RTM_DELMDB with NLM_F_BULK and through VXLAN device deletion. For the CVSS the PR:L is used because an attacker needs local control of VXLAN MDB configuration through rtnetlink and typically CAP_NET_ADMIN. This can include a reduced capability container process with delegated network administration privileges. The issue is not directly reachable through ordinary VXLAN packet traffic. Impact is at least local denial of service via kernel crash and in the worst case may allow confidentiality or integrity impact due to the confirmed use after free write and should be reviewed manually.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Borderline / manual review recommended (with actual score 4)	YES	DANGER INIT UAF NETWORK HARDWARE KPANIC  KPANIC INCREASED_TO_MODERATE7_BASED_ON_GUESSCVSS	YES	NO	checked

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

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

1. ActionableScore

* Raw conservative score: 3
* Conservative score: 4
* Raw paranoid score: 3
* Paranoid score: 4
* Applied conservative floor: Real UAF floor 4 applied because a freed MDB entry is subsequently processed through stale list pointers and host-admin-equivalent privilege is not required.
* Applied paranoid floor: Real UAF floor 4 applied for the same reason.
* Final recommended bucket: **Borderline / manual review recommended**::

2. Signal breakdown

Positive signals:

* +2 Real UAF / lifetime violation. A related S,G MDB entry can be freed while its hlist node remains saved as the next iterator element. The next loop iteration operates on this freed entry.
* +1 Reliable kernel crash / strong DoS. The supplied KASAN trace shows a wild write through poisoned list pointers during `vxlan_mdb_entry_put`.
* +1 Attacker-influenced active-object lifetime manipulation. The attacker can configure ordered MDB entries and trigger bulk MDB deletion or device deletion so that remotes flushing removes the saved next entry.

Negative signals:

* -1 Reduced or delegated privilege requirement. Triggering requires `CAP_NET_ADMIN` for VXLAN MDB and rtnetlink operations. This capability can be scoped to the relevant network namespace and is below host-admin-equivalent privilege.

Important non-awarded signals:

* Controlled reclaim or attacker-controlled replacement of the freed MDB entry is not demonstrated.
* Arbitrary or directed kernel write is not demonstrated. The observed write uses stale and poisoned hlist pointers.
* Remote reachability is not established. Ordinary VXLAN packet traffic does not reach this control-plane path.
* Structured stale-object exploitation leverage is not awarded. The stale use is a list cleanup operation and the evidence shows an immediate invalid list operation rather than meaningful consumption of attacker-shaped replacement fields.
* Weak LPE concern is not awarded. A UAF alone does not establish privilege escalation.
* Extra Availability is not awarded because the same stale-list operation already accounts for the demonstrated crash effect.
* Confidentiality and integrity impact are not awarded. The patch establishes a stale post-free list operation, but does not establish disclosure, controlled memory reuse, protected-state modification, or an arbitrary write.
* Concrete repeated release/refcount mutation is not established. `vxlan_mdb_entry_put` is invoked on a stale entry, but the supplied context does not establish a first and second lifetime mutation through the exact same surviving reference in the sense of a repeated `put` or refcount takeover primitive.
* Paranoid additional leverage is withheld. The supplied evidence supports a real UAF and list corruption, but not a post-free attacker-controlled mutation beyond the immediate poisoned-list failure.

3. Reachability analysis

The evaluated attacker is a local process controlling VXLAN multicast database configuration through rtnetlink. The supplied commit identifies `RTM_DELMDB` with `NLM_F_BULK` as a trigger, in addition to VXLAN device deletion.

VXLAN devices, MDB entries, and rtnetlink configuration are network-namespace-aware. A process with `CAP_NET_ADMIN` in the target network namespace can create and configure a namespace-local VXLAN device and exercise the relevant MDB control path. Therefore host-admin-equivalent privilege is not inherently required, and the complete relevant operation is namespace-scoped.

An ordinary unprivileged host user does not automatically have the necessary capability. Whether an attacker can create a user and network namespace and obtain the required namespaced capability depends on local user-namespace policy. No local-unprivileged-trigger bonus is awarded without that product-specific evidence.

No special hardware, firmware, boot parameter, or external network exposure is required. This is not a timing race. The required MDB ordering and subsequent bulk deletion are deterministic control-plane operations.

4. Primitive and severity interpretation

The exact primitive is a real use-after-free involving an active VXLAN MDB list entry. During iteration, `hlist_for_each_entry_safe` saves a following S,G entry. Flushing remotes from a preceding star,G entry can remove and free that saved S,G entry. The loop then processes the stale entry and performs an invalid hlist removal path, producing a write through poisoned list metadata.

This is an active structured object, but the demonstrated stale operation is limited to list and entry cleanup processing. The patch does not show controlled object replacement, attacker-controlled post-free contents, callback dispatch, arbitrary read or write, credential manipulation, or a cross-security-boundary effect.

This is CLASS_B because a genuine post-free kernel object operation and invalid memory write are established. It is not CLASS_A because the write destination and payload are not attacker-directed and useful post-free reuse is not established.

The demonstrated consequence is kernel memory-safety failure and plausible kernel crash or DoS. The speculative consequence is broader memory corruption if allocator reuse or list-pointer control were independently demonstrated, but that mechanism is not supported by the supplied patch. ImportantCandidate is NO because the observed stale operation is constrained list cleanup behavior with an immediate poisoned-pointer failure rather than an unresolved strong exploitation primitive. AutoDowngradeAllowed is YES, but manual review remains recommended to validate product-specific namespace exposure and allocator behavior.

Likely operational severity is Borderline Moderate. It should be prioritized above ordinary bookkeeping issues because it is a concrete UAF, but it is not currently supported as an Important-class exploitation primitive.

5. One-sentence report phrase

A local attacker with CAP_NET_ADMIN in the target network namespace can trigger a VXLAN MDB use-after-free through bulk MDB deletion, causing an invalid stale-list operation and kernel crash or denial of service.

6. Manual review recommendation

MANUAL CHECK RECOMMENDED

The patch demonstrates a genuine UAF with a post-free list write, but available evidence supports crash-oriented behavior rather than controlled corruption or privilege escalation. Review namespace exposure and whether the affected deployment permits untrusted tenants to manage VXLAN MDB state.

7. Call-site confidence

Call-site confidence: high

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

Patch: vxlan: mdb: Fix use-after-free in vxlan_mdb_flush() [ Upstream
Commit: f8a9b988e7a7bc674e74e8c901794d792c35689e
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=f8a9b988e7a7bc674e74e8c901794d792c35689e

Changed files:
  drivers/net/vxlan/vxlan_mdb.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=f8a9b988e7a7bc674e74e8c901794d792c35689e

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

The original Linux kernel CVE announcement for CVE-2026-93250 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93250

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:L/PR:L/UI:N/S:U/C:N/I:N/A:H
  The Best / paranoid CVSS vector: AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
  The Best / paranoid CVSS score: 7.8

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: unknown
  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: 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).