[CVE-2026-90292][MODERATE 7.0] RDMA/siw: Fix use-after-free in siw_accept() [ Upstream

[CVE-2026-90292][MODERATE 7.0] RDMA/siw: Fix use-after-free in siw_accept() [ Upstream

[AL-KERNEL]

CVE: CVE-2026-90292
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: RDMA/siw: Fix use-after-free in siw_accept() [ Upstream
Commit: d57db36e74cca4763e1c2c0c6eb9f84d19813879
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=d57db36e74cca4763e1c2c0c6eb9f84d19813879
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90292
Analysis date: Thu, 17 Sep 2026 16:31:18 -0400
ActionableScore: 6
ActionableScore lower bound: 5
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-2026-90292 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90292

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-2026-90292	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-90292 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-672;CWE-703;Other CVSS 'AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:H';BEST CVSS score: '7.8';DESCR 'A use after free exists in RDMA siw siw_accept because cleanup checks only whether qp cep is non NULL and assumes the current accept path installed that association. If userspace supplies a QP that is already in RTS, qp cep may refer to an existing connection while the incoming CEP is the object whose reference is dropped. The incoming CEP can then be freed before later use of cep, and the existing QP association may also be cleared incorrectly. For the CVSS the PR:L is selected because reliable triggering requires a local process able to use the RDMA siw accept path and control the supplied QP state. The issue is not directly network reachable by packet traffic alone, although an incoming RDMA connection can be part of the scenario. Impact is at least local denial of service via kernel crash and in the paranoid case may include confidentiality or integrity impact due to kernel use after free.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 6)	YES	UAF NETWORK LINUS TEST KPANIC  KPANIC INCREASED_TO_HIGH_BASED_ON_GUESSCVSS DECREASED_TO_MODERATE7_BASED_ON_FALSEPOSCHECKOFDH	NO	NO	checked

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

ActionableScore=6
ActionableScoreLower=5
PrimitiveClass=CLASS_B
RealUAF=YES
HostAdminRequired=NO
NamespaceScopedPrivilege=UNKNOWN
ImportantCandidate=YES
ManualReview=REQUIRED
AutoDowngradeAllowed=NO

1. ActionableScore

* Raw conservative score: 5
* Conservative score: 5
* Raw paranoid score: 6
* Paranoid score: 6
* Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**::

No floor or cap changed the score.

2. Signal breakdown

Conservative signals:

* +2 Real UAF. The commit directly states that the incoming CEP can lose its remaining references and be freed before a later `cep->qp` store.
* +1 Reliable kernel crash / strong DoS. A post-lifetime access to a freed CEP in kernel RDMA connection-management code can realistically crash the kernel.
* +1 Weak exploitation / corruption concern. The stale operation is not only a read-only fault. It includes a post-free structured field store and incorrect clearing of an existing QP association.
* +1 Local below-host-admin reachability. The vulnerable path is driven by userspace supplying a QP to `siw_accept()`. This is a local RDMA/siw userspace attack surface rather than a host-admin-only kernel-internal path.

Paranoid additional signal:

* +1 Attacker-influenced active-object state / repeatable lifetime manipulation. A local process with RDMA/siw access can influence the supplied QP state, including the already-RTS case, and can plausibly repeat accept/error-path attempts with incoming CEP objects.

Non-awarded signals:

* Controlled reclaim is not demonstrated.
* Arbitrary or directed kernel write is not demonstrated.
* CLASS_A is not assigned because there is no evidence of controlled object replacement, callback control, arbitrary R/W, or demonstrated LPE.
* Remote reachability is not awarded. An incoming RDMA connection may participate, but packet traffic alone is not shown to select the QP and complete the vulnerable accept sequence.
* Extra Availability is not awarded because the same crash is already counted.
* Confidentiality and Integrity points are not separately awarded as CVSS-like impacts. The UAF justifies manual review, but no concrete disclosure or unauthorized data modification primitive is proven.

Repeated release / refcount check:

* First release can end lifetime: yes. `siw_cep_put(cep)` can drop the incoming CEP reference and may free it.
* Exact same reference survives: yes. The local `cep` reference remains used after the put.
* Second operation is a real lifetime/refcount mutation: no. The later established stale operation is a `cep->qp` store/use, not a second put/release/refcount mutation on the exact same reference.
* Concrete repeated release/refcount +1 awarded: no.
* Paranoid repeated-refcount leverage +1 awarded: no.
* FLOOR 5 applied: no explicit floor was needed because raw conservative score already reached 5.

3. Reachability analysis

The evaluated attacker is a local userspace process with access to the RDMA/siw accept path. The commit says the QP is supplied by userspace, and the vulnerable condition occurs when that QP is already in RTS before the incoming CEP is associated with it.

This is not directly network-triggerable by packets alone. A remote peer can be part of the connection scenario, but the evidence does not show that a remote-only attacker can choose the already-RTS QP or drive the local accept/error cleanup path without local userspace cooperation.

Host-admin-equivalent privilege is not positively established as required for the vulnerable operation. However, exact deployment permissions for RDMA device access, uverbs access, and siw setup may vary. Prior host setup may be needed to have siw/RDMA available, but the patch does not establish host root as required for triggering once the interface exists.

Namespace-scoped privilege is unresolved. The supplied evidence does not establish that the complete RDMA/siw operation is available inside attacker-created namespaces.

No special physical RDMA hardware is necessarily required because siw is software iWARP, but RDMA/siw availability is a deployment prerequisite.

This is not a timing race. It is an ordered error-cleanup/lifetime bug triggered by a specific QP state.

4. Primitive and severity interpretation

The exact primitive is a kernel use-after-free in `siw_accept()`. The incoming CEP can be freed after an erroneous `siw_cep_put(cep)`, and the same `cep` pointer can then be used again for a `cep->qp` store. The affected object is an active structured RDMA connection endpoint, not fixed stale DMA or passive storage.

This is stronger than an ordinary stale cleanup dereference because the stale path can write to a field of the freed CEP and also corrupt logical QP-to-CEP association state. It is still not CLASS_A because controlled reclaim, controlled replacement contents, arbitrary write, function-pointer influence, credential manipulation, or demonstrated LPE are not established.

This is not a concrete repeated-release/refcount case under the dedicated rule. The first put may free the CEP, but the later operation is a stale field store/use rather than a second lifetime/refcount mutation.

PrimitiveClass is CLASS_B because there is a real kernel lifetime violation with unresolved exploitation leverage. The demonstrated consequence is at least local kernel crash / DoS. The speculative consequence is possible stronger memory-corruption impact if post-free storage can be meaningfully reused or shaped, but that is not proven by the patch.

ImportantCandidate is YES because this is a real active-object UAF with post-free structured access reachable from userspace-controlled RDMA state. ManualReview is REQUIRED because automatically downgrading it as crash-only would discard unresolved lifetime-corruption leverage. AutoDowngradeAllowed is NO for the same reason.

Likely operational severity is Actionable Moderate at minimum, with Strong Important candidate review warranted if RDMA/siw access is available to non-host-admin local users.

5. One-sentence report phrase

A userspace-controlled RDMA/siw accept path can prematurely drop the incoming CEP reference when the supplied QP is already in RTS, causing a use-after-free before a later `cep->qp` store and requiring manual review for possible impact beyond local kernel DoS.

6. Manual review recommendation

MANUAL CHECK REQUIRED

Reason: real kernel UAF in an active RDMA connection-management object, with post-free structured field access and below-host-admin userspace reachability plausibly established.

7. Call-site confidence

Call-site confidence: medium

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

Patch: RDMA/siw: Fix use-after-free in siw_accept() [ Upstream
Commit: d57db36e74cca4763e1c2c0c6eb9f84d19813879
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=d57db36e74cca4763e1c2c0c6eb9f84d19813879

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=d57db36e74cca4763e1c2c0c6eb9f84d19813879

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

The original Linux kernel CVE announcement for CVE-2026-90292 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90292

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:L/I:L/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: 5
  Paranoid ActionableScore: 6

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).