[CVE-2026-98367][MODERATE REGULAR] RDMA/siw: Clear association under lock if siw_qp_modify fails in siw_accept [ Upstream

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

CVE: CVE-2026-98367
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: RDMA/siw: Clear association under lock if siw_qp_modify fails in siw_accept [ Upstream
Commit: f11e09fe2fc3a11ccdf8f932b68181b0bb1d2078
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=f11e09fe2fc3a11ccdf8f932b68181b0bb1d2078
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98367
Analysis date: Thu, 08 Oct 2026 07:54:03 -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-2026-98367 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98367

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-2026-98367	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-98367 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:H/A:H';*CWE-416;CWE-362;CWE-667;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:H';BEST CVSS score: '7';DESCR 'A use after free can occur in the RDMA siw connection manager when siw_qp_modify fails in siw_accept and the QP state lock is released before error cleanup completes. A concurrent local ibv_modify_qp transition of the same QP to ERROR can drop the cep reference and free the cep object. The failing siw_accept path can then write to cep after it has been freed. For the CVSS the PR:L value is used because reliable triggering requires local access to an siw RDMA QP and the ability to issue QP modification operations. The race requires precise timing and is therefore AC:H. A remote peer may help cause an accept failure but cannot independently trigger the required local concurrent QP transition. Impact is at least kernel denial of service. The stale write to a freed kernel object also provides a plausible memory corruption path with potential confidentiality integrity and privilege escalation consequences and should be reviewed manually.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Borderline / manual review recommended (with actual score 4)	MAYBE	DANGER RACE ERRORPATH NETWORK LINUS  DECREASED_TO_MODERATE70_BASED_ON_ACTIONABLESCORELOWERTHAN5 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: 2
* Conservative score: 4
* Raw paranoid score: 2
* Paranoid score: 4
* Applied conservative floor: 2 -> 4. A real UAF of an active structured connection endpoint object is directly established, and host-admin-equivalent privilege is not positively established.
* Applied paranoid floor: 2 -> 4.
* Final recommended bucket: **Borderline / manual review recommended**::

2. Signal breakdown

Positive signals:

* +2 Real UAF / active lifetime violation. `siw_cep_put(qp->cep)` in the concurrent QP ERROR transition can free `cep`, while the failing `siw_accept()` path still retains its local `cep` pointer and subsequently writes `cep->qp = NULL`.
* +0 for reliable kernel crash. A stale post-free write is established, but the supplied evidence does not demonstrate a reliable Oops or panic. The practical consequence may be a crash or silent slab corruption depending on allocator state.

Floors:

* Real UAF FLOOR 4 applied. `cep` is an active structured SIW connection-endpoint object, not fixed DMA or fixed stale storage. The patch directly establishes a post-free write through a stale pointer. No proven host-admin-equivalent requirement permits applying a privilege-based reduction.

Important non-awarded signals:

* +0 Controlled reclaim or object replacement. No evidence shows that the attacker can replace freed `cep` storage with attacker-shaped data before the stale write.
* +0 Arbitrary or directed write. The established stale write is fixed: it stores NULL into the `qp` field at a fixed offset.
* +0 Strong or weak LPE leverage. A UAF write exists, but no callback, function pointer, attacker-controlled replacement, sensitive field target, refcount takeover, or privilege-escalation chain is established.
* +0 Remote reachability. A remote peer may contribute to a failed connection setup, but the required concurrent `ibv_modify_qp(ERROR)` operation is local and no remote-only trigger is established.
* +0 Local unprivileged trigger. The supplied material does not establish that an ordinary unprivileged user can create and operate the required SIW RDMA QP and execute the full trigger sequence.
* +0 Concrete repeated release/refcount operation. The first `siw_cep_put()` can end the `cep` lifetime, and the exact stale `cep` pointer survives in `siw_accept()`. However, the later operation is `cep->qp = NULL`, not a second `put`, `release`, `free`, or refcount mutation. The dedicated repeated-release/refcount signal does not apply.
* +0 Paranoid repeated-refcount leverage. Not applicable because there is no second lifetime-mutating operation.
* +0 Structured stale-object leverage. The object is structured, but the demonstrated post-free use is a fixed cleanup write. No meaningful post-reuse field consumption is established.
* +0 Confidentiality and integrity consequence. The stale write can theoretically affect reused slab memory, but no disclosure, protected-state modification, or attacker-controlled target is demonstrated.
* +0 additional Availability. No distinct availability consequence beyond possible kernel fault or corruption is established.

3. Reachability analysis

The evaluated attacker is a local RDMA user or process that can operate on the same SIW QP through RDMA verbs, including an `ibv_modify_qp()` transition to ERROR while `siw_accept()` is handling a failed QP modification.

The patch does not establish the complete privilege boundary for this sequence. Access to RDMA uverbs devices may be delegated through device-node permissions, groups, containers, or service accounts, but the supplied evidence does not prove that an ordinary unprivileged user can create the required SIW connection and QP without prior privileged setup. Conversely, it also does not prove that host-admin-equivalent privilege is mandatory. Therefore, `HostAdminRequired=UNKNOWN`.

`NamespaceScopedPrivilege=UNKNOWN`. No supplied evidence establishes that SIW device provisioning, RDMA device access, connection setup, QP creation, and the relevant QP modification operation can all be performed with attacker-obtainable namespace-scoped privileges. The mere possibility of network namespaces or container deployment is insufficient.

The race requires a failed `siw_qp_modify()` during accept and a concurrent QP ERROR transition. The synchronization window is explicit, but the patch does not establish that it is unusually narrow, one-shot, or externally uncontrollable. No hard-race negative point is applied.

No special physical hardware is required because SIW is software iWARP over TCP, but an enabled and accessible SIW/RDMA environment is a practical prerequisite.

4. Primitive and severity interpretation

The exact primitive is a post-free write: one execution path drops the `cep` reference and can free the SIW connection endpoint, while another path retains a stale local pointer and writes `cep->qp = NULL`.

This is an active structured-object UAF, not fixed asynchronous storage. However, the stale use is located in an error cleanup path and performs only a fixed NULL assignment to a fixed field. The supplied patch does not show controlled reclamation, attacker-shaped replacement contents, callback dispatch, arbitrary read/write, or a security-boundary bypass.

This is not a concrete repeated-release/refcount case. The first lifetime operation is `siw_cep_put()` in the concurrent ERROR transition and may free the object. The exact pointer survives in the failing accept path, but the second operation is a direct stale field write rather than another release or refcount mutation.

The primitive is therefore `CLASS_B`: a real kernel lifetime violation with a concrete stale write, but without evidence of a strong exploitation primitive. The demonstrated consequence is possible kernel memory corruption and likely denial of service. Privilege escalation, confidentiality impact, integrity impact against a protected target, and controlled kernel corruption remain speculative.

`ImportantCandidate=NO` because the currently established stale write has fixed payload and fixed field offset, and the patch provides no meaningful post-free reuse or attacker-control mechanism. `AutoDowngradeAllowed=YES` because this is currently a constrained cleanup-path Class B UAF without demonstrated controlled reclaim, repeated lifetime mutation, security-boundary effect, or concrete non-crash exploitation leverage. Manual review is still recommended because the patch explicitly documents a real UAF write rather than a simple NULL dereference.

Likely operational severity is Borderline Moderate. It should be patched in RDMA/SIW-enabled environments, but the supplied evidence does not support Important classification without additional reachability or exploitation evidence.

5. One-sentence report phrase

A race between failed SIW connection acceptance and a concurrent QP ERROR transition can free a connection endpoint before the failing accept path clears its association, resulting in a fixed post-free kernel write and possible denial of service.

6. Manual review recommendation

MANUAL CHECK RECOMMENDED

The patch explicitly establishes a real UAF write to a freed SIW connection endpoint, but available evidence does not establish below-host-admin reachability, controlled object reuse, attacker-controlled post-free contents, or a privilege-escalation primitive.

7. Call-site confidence

Call-site confidence: medium

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

Patch: RDMA/siw: Clear association under lock if siw_qp_modify fails in siw_accept [ Upstream
Commit: f11e09fe2fc3a11ccdf8f932b68181b0bb1d2078
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=f11e09fe2fc3a11ccdf8f932b68181b0bb1d2078

Changed files:
  drivers/infiniband/sw/siw/siw_cm.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=f11e09fe2fc3a11ccdf8f932b68181b0bb1d2078

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

The original Linux kernel CVE announcement for CVE-2026-98367 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98367

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: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: 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).