[CVE-2026-98370][MODERATE REGULAR] xfrm: fix compat ALLOCSPI request use-after-free [ Upstream

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

CVE: CVE-2026-98370
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: xfrm: fix compat ALLOCSPI request use-after-free [ Upstream
Commit: 494f2bee9d8d0ebcfa249ac41bed7fed26d119b4
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=494f2bee9d8d0ebcfa249ac41bed7fed26d119b4
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98370
Analysis date: Thu, 08 Oct 2026 09:47:21 -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-2026-98370 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98370

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-2026-98370	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-98370 CHECK WITH IMPACT FROM ORIG NN IMPORTANT | AEGIS IMPACT MODERATE_KPANIC AEGIS_AS=4 LEGACY=MODERATE RESULT=DIFF (legacy and Aegis reconciliation differ; DECISION: legacy MODERATE vs Aegis MODERATE_KPANIC)	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:N/A:H';*CWE-416;CWE-125;CWE-362;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:N/A:H';BEST CVSS score: '6.3';DESCR 'A use after free can occur in the XFRM compat ALLOCSPI netlink path because a redundant compat translation attaches a translated child skb to the inbound request skb frag_list. A multicast clone can retain access to the shared child while the request completion path frees it, racing a compat receiver that is still copying its contents. The redundant translation also interprets a 228 byte compat xfrm_userspi_info payload as a 232 byte native structure and reads four bytes past the declared payload. For the CVSS the PR:L is selected because sending an XFRM ALLOCSPI request normally requires CAP_NET_ADMIN, including delegated network administration capabilities available to a local container or service account. The issue is not directly network reachable. Impact is at least local denial of service through a kernel memory safety failure. In the paranoid case reused skb contents may be exposed through the UAF read race, but the patch does not demonstrate an arbitrary write primitive or reliable privilege escalation.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Borderline / manual review recommended (with actual score 4)	MAYBE	DANGER UAF SKB LINUS  LOWERED_FROM_HIGH_BASED_ON_GUESSCVSS  DECREASED_TO_MODERATEREG_BASED_ON_FALSEPOSCHECKOFKP	-	-	checked

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

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

1. ActionableScore

* Raw conservative score: 3
* Conservative score: 4
* Raw paranoid score: 3
* Paranoid score: 4
* Conservative floor applied: 3 -> 4 because a real UAF is established, the stale child skb is actively copied by a receiver, and host-admin-equivalent privilege is not positively established.
* Paranoid floor applied: 3 -> 4 for the same reason.
* Final recommended bucket: **Borderline / manual review recommended**::

2. Signal breakdown

Positive signals:

* +2 Real UAF / active lifetime violation. A translated child skb is attached to the inbound request skb frag_list. A multicast clone can retain access to that child while xfrm_user_rcv_msg frees it after the request handler returns. A compat receiver can still be copying the released child.
* +1 Confidentiality plausible. The stale skb child is consumed by a receiver copy path, which provides a plausible user-visible read-back path if the freed allocation is reused before copying completes.

Not awarded:

* +0 Strong corruption primitive. No controlled reclaim, attacker-controlled replacement object, arbitrary write, directed write, callback control, refcount takeover, or type confusion is demonstrated.
* +0 Structured stale-object leverage. The supplied evidence establishes copying from a freed skb child, but not meaningful attacker control over post-free replacement contents or subsequent structured field consumption.
* +0 Attacker-influenced active-object lifetime manipulation. Repeated requests may be possible, but the supplied context does not establish attacker control over the critical clone/free/copy ordering or useful object reuse.
* +0 Reliable kernel crash / strong DoS. The patch establishes a UAF race, but does not directly demonstrate a reproducible kernel panic, Oops, or host-wide crash.
* +0 Remote reachability. This is a local NETLINK_XFRM control-plane operation, not a packet-data path reachable by an unauthenticated remote peer.
* +0 Integrity impact. The stale operation is a receiver copy from freed storage. No post-free write or integrity-relevant mutation is established.
* +0 Strong LPE plausibility. Privilege escalation is not supported by a controlled reuse, write, callback, credential, or authorization primitive.
* +0 Concrete repeated release/refcount operation. The patch removes an unnecessary conversion and prevents publication of a child skb through frag_list. It does not show a second put, release, free, or refcount mutation through the same dangling reference.

No negative privilege, niche-subsystem, or hard-race penalty was applied. CAP_NET_ADMIN is relevant to XFRM state-management operations, but the supplied patch does not establish whether the complete trigger requires host-admin capability or can be completed within an attacker-controlled user and network namespace. The race exists, but the evidence does not establish a uniquely narrow or one-shot timing condition that warrants a timing penalty.

3. Reachability analysis

The triggering actor is a local process that can send a compat XFRM ALLOCSPI NETLINK_XFRM request and arrange for a compat receiver to observe the multicast-cloned request path.

XFRM state-management netlink operations normally require CAP_NET_ADMIN. The exact capability check and namespace ownership for all required stages are not included in the supplied context. XFRM objects are generally network-namespace scoped, and Linux commonly evaluates netlink network-administration checks relative to the target network namespace. However, the supplied material does not prove that an unprivileged user can create the required namespaces, obtain the relevant capability, issue ALLOCSPI, subscribe to the relevant multicast delivery, and complete the compat trigger sequence in the evaluated product.

Therefore HostAdminRequired is UNKNOWN rather than YES or NO. NamespaceScopedPrivilege is also UNKNOWN. No prior hardware, firmware, boot option, or unusual device is indicated. The compat aspect requires a 32-bit userspace process on a compatible 64-bit kernel configuration, but this is a software execution-mode prerequisite rather than rare physical hardware.

The race is between request completion freeing the attached child skb and the multicast receiver copying it. The bug may be retried through repeated netlink requests, but practical timing control and post-free object reuse are not established by the patch.

4. Primitive and severity interpretation

The exact primitive is a real use-after-free read race involving a translated child skb attached to the inbound request skb frag_list. The child becomes visible through shared skb_shared_info to a multicast clone. The request completion path frees the child while a compat receiver may still copy its data.

This is an active lifetime violation, not merely an invalid state or harmless bounds validation defect. The redundant compat conversion also reads four bytes past the declared compat xfrm_userspi_info payload because a 228-byte compat layout is interpreted as a 232-byte native layout. However, the supplied evidence does not independently establish that this limited OOB read discloses sensitive data or controls execution.

The stale object is an skb child used for data copying. It is not shown to be replaced by an attacker-controlled active object whose fields are later interpreted as pointers, callbacks, lengths, or refcounts. The demonstrated stale operation is a read/copy operation, not a second lifetime mutation. This is therefore CLASS_B rather than CLASS_A.

The demonstrated practical consequence is kernel memory-safety exposure and a plausible stale-data disclosure path to the receiver. A crash may be possible for some allocator reuse outcomes, but it is not directly demonstrated in the supplied patch context. Privilege escalation remains speculative because there is no demonstrated arbitrary write, controlled reclaim, structured object substitution, or security-boundary bypass.

ImportantCandidate is NO because the UAF is concrete but currently constrained to a copy-from-freed-skb path without established reuse leverage beyond potential stale-data exposure. AutoDowngradeAllowed is YES because the currently supported evidence does not establish controlled corruption, repeated lifetime mutation, active-object replacement, or a boundary-bypass primitive. Manual review remains recommended due to the real UAF and receiver-visible copy path.

Likely operational severity is borderline Moderate. It should be reviewed for namespace reachability and for whether skb allocator reuse can turn the copy path into a practical disclosure primitive in affected product configurations.

5. One-sentence report phrase

A local compat XFRM ALLOCSPI netlink request can race request cleanup with multicast receiver copying because a redundant compat translation attaches a child skb to the inbound request frag_list, causing a use-after-free read with potential stale-data disclosure.

6. Manual review recommendation

MANUAL CHECK RECOMMENDED

A real UAF and plausible receiver-visible stale-data copy are established, but host versus namespace privilege scope, practical allocator reuse, disclosure reliability, and crash behavior require validation before escalating to Important.

7. Call-site confidence

Call-site confidence: medium

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

Patch: xfrm: fix compat ALLOCSPI request use-after-free [ Upstream
Commit: 494f2bee9d8d0ebcfa249ac41bed7fed26d119b4
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=494f2bee9d8d0ebcfa249ac41bed7fed26d119b4

Changed files:
  net/xfrm/xfrm_user.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=494f2bee9d8d0ebcfa249ac41bed7fed26d119b4

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

The original Linux kernel CVE announcement for CVE-2026-98370 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98370

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:N/A:H
  The Best / paranoid CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:N/A:H
  The Best / paranoid CVSS score: 6.3

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: 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: 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).