[CVE-2026-92511][MODERATE 7.0] RDMA/core: Fix potential use after free in ib_destroy_cq_user() [ Upstream

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

CVE: CVE-2026-92511
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: RDMA/core: Fix potential use after free in ib_destroy_cq_user() [ Upstream
Commit: 42f7a0c7a94003059540cce448acabfa62ad119d
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=42f7a0c7a94003059540cce448acabfa62ad119d
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-92511
Analysis date: Fri, 18 Sep 2026 06:12:03 -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-2026-92511 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-92511

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-2026-92511	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-92511 CHECK WITH IMPACT FROM ORIG NN MODERATE	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-664;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '7';DESCR 'A race in RDMA CQ destruction can leave a CQ reachable through restrack while vendor specific CQ resources are already being destroyed. A local actor that can use RDMA objects and query the RDMA netlink restrack path may hit a use after free window when ib_destroy_cq_user runs concurrently with restrack access. For the CVSS the PR:L is used in the paranoid score because reliable triggering may only require a local user or service account with access to RDMA device interfaces rather than full host administrator privileges. The issue is not directly network reachable over IP and requires local interaction with RDMA control interfaces. Impact is at least local denial of service via kernel crash. In worst case the UAF nature may allow confidentiality or integrity impact depending on the vendor driver state that is freed and later observed or reused, so this should be reviewed manually.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 6)	YES	INIT UAF NETWORK LINUS KPANIC  KPANIC INCREASED_TO_HIGH_BASED_ON_GUESSCVSS DECREASED_TO_MODERATE7_BASED_ON_FALSEPOSCHECKOFKP	NO	NO	checked

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

ActionableScore=6
ActionableScoreLower=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
* Applied conservative floor/cap: Real UAF FLOOR 5 applied because the CQ remains reachable through restrack while internal vendor resources may already be destroyed, and the stale access is through an active netlink/restrack path rather than a purely local cleanup fault.
* Raw paranoid score: 6
* Paranoid score: 6
* Applied paranoid floor/cap: none
* Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**::

2. Signal breakdown

Positive conservative signals:

* +2 Real UAF / active lifetime violation. The commit directly states that the CQ can remain accessible through restrack after vendor-specific CQ resources may already be freed.
* +1 Reliable kernel crash / strong DoS. A stale CQ or stale CQ subresource accessed through the netlink restrack path can realistically fault the kernel.
* +1 Secondary index / registration-table staleness. The restrack table remains able to expose the CQ after destruction has progressed, and the fix moves removal from restrack before vendor resource destruction.

Positive paranoid-only signals:

* +1 Attacker-influenced active-object lifetime / repeatable lifetime manipulation. If a local actor can create and destroy RDMA CQs and concurrently query restrack, the actor can repeatedly drive the vulnerable destruction and lookup pattern.
* +1 Structured stale-object exploitation leverage. The stale object is an active CQ exposed through a kernel resource-tracking interface, and stale restrack access may consume structured CQ or vendor-resource state after destruction. This is not proof of controlled reclaim or arbitrary write, but it is enough for unresolved Class B review.

Non-awarded signals:

* Controlled reclaim not demonstrated.
* Arbitrary kernel read or write not demonstrated.
* Concrete repeated release/refcount mutation not established. The patch changes restrack begin/abort/commit ordering, but does not show the same dangling reference receiving a second lifetime/refcount mutation.
* Weak LPE not separately awarded because no distinct privilege-escalation primitive is established.
* Extra Availability not awarded because the same possible crash is already counted once.
* C/I not awarded directly. Confidentiality and integrity remain unresolved paranoid concerns through UAF analysis, not demonstrated consequences.
* Remote reachability not awarded. This is RDMA control/netlink/uverbs local interaction, not a direct IP packet trigger.
* Race penalty not applied. The window is described as short, but the create/destroy/query pattern appears repeatable enough that a hard one-shot timing penalty is not clearly supported.

Conservative arithmetic:

* +2 UAF
* +1 crash/DoS
* +1 restrack secondary-index staleness
* Raw conservative score: 4
* FLOOR 5 applied: 4 -> 5

Paranoid arithmetic:

* +2 UAF
* +1 crash/DoS
* +1 restrack secondary-index staleness
* +1 repeatable lifetime manipulation
* +1 structured stale-object leverage
* Raw paranoid score: 6

3. Reachability analysis

The vulnerable destroy path is `ib_destroy_cq_user()`, which is associated with user-created RDMA completion queues. The concurrent stale access is described as the RDMA netlink restrack path using `rdma_restrack_get()` as its synchronization mechanism.

A practical attacker likely needs local ability to use RDMA userspace objects and to query RDMA restrack through netlink. The supplied evidence does not prove that full host-admin-equivalent privilege is required, but it also does not fully prove ordinary unprivileged reachability for the complete chain on all systems. Therefore `HostAdminRequired=UNKNOWN`.

The complete namespace story is also unresolved. RDMA devices and RDMA netlink/resource tracking are not automatically equivalent to attacker-created network namespaces, and the supplied patch does not prove that all required stages are namespace-scoped. Therefore `NamespaceScopedPrivilege=UNKNOWN`.

No direct network reachability over IP is established. The issue is local/control-plane reachable, not packet-processing reachable. Prior setup may include RDMA hardware/device availability and permissions to access RDMA device nodes or RDMA management interfaces. Special RDMA-capable hardware or virtual RDMA setup may be needed, but the patch does not justify treating the hardware as rare or physical-only for scoring.

The bug involves a race between CQ destruction and restrack access. The commit says the window is short, but the relevant operations can plausibly be attempted repeatedly by an actor with RDMA object and netlink access, so no hard race penalty is applied.

4. Primitive and severity interpretation

The exact primitive is a lifecycle/order violation: `rdma_restrack_del()` used to occur after `device->ops.destroy_cq()`, so the CQ could remain published in restrack while vendor-specific resources associated with it were already released. A concurrent netlink restrack lookup could obtain or use a CQ that is in the middle of destruction, creating a directly implied UAF or stale subresource access.

This is an active structured object case, not fixed stale storage. The CQ is a nontrivial RDMA kernel object, and restrack is an active resource-tracking interface rather than a passive diagnostic field. The stale access is not merely a destructor-local cleanup dereference, because another path can still discover the CQ through restrack.

No concrete repeated release/refcount mutation is shown. This is not the repeated-release exception. The stronger concern is that the CQ remains externally discoverable through a secondary registration structure after internal destruction has started.

Primitive class is `CLASS_B`: a real kernel lifetime/UAF primitive exists, but controlled reclaim, arbitrary read/write, callback control, or demonstrated privilege escalation is not established.

Demonstrated consequence is at least local DoS through kernel crash. Speculative consequence is stronger memory-corruption leverage depending on what vendor-specific CQ resources are freed and what the restrack reader later consumes. That unresolved active-object UAF surface is why `ImportantCandidate=YES`, `ManualReview=REQUIRED`, and `AutoDowngradeAllowed=NO`.

Likely operational severity is actionable Moderate at minimum. It should not be automatically closed as an ordinary crash-only issue without checking RDMA netlink permissions, reachable drivers, vendor CQ layouts, and whether stale restrack access can consume attacker-influenced or reusable state.

5. One-sentence report phrase

A race in RDMA CQ destruction can leave a completion queue reachable through restrack after vendor-specific CQ resources have begun teardown, allowing local RDMA/netlink access to hit a potential use-after-free and kernel crash with unresolved memory-corruption leverage.

6. Manual review recommendation

MANUAL CHECK REQUIRED

Reason: this is a real active-object lifetime/UAF issue crossing the RDMA destroy path and restrack netlink visibility. Controlled exploitation is not demonstrated, but the object remains externally reachable after internal resources may be freed, so automatic downgrade would risk losing a meaningful unresolved Class B primitive.

7. Call-site confidence

Call-site confidence: medium

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

Patch: RDMA/core: Fix potential use after free in ib_destroy_cq_user() [ Upstream
Commit: 42f7a0c7a94003059540cce448acabfa62ad119d
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=42f7a0c7a94003059540cce448acabfa62ad119d

Changed files:
  drivers/infiniband/core/verbs.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=42f7a0c7a94003059540cce448acabfa62ad119d

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

The original Linux kernel CVE announcement for CVE-2026-92511 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-92511

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:N/I:N/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: 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).