[CVE-2026-89798][MODERATE REGULAR] rpcrdma: arm rn_done before publishing the notification [ Upstream

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

CVE: CVE-2026-89798
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: rpcrdma: arm rn_done before publishing the notification [ Upstream
Commit: 3c97b8e76ca2bba9e8770571413aed694068e78e
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=3c97b8e76ca2bba9e8770571413aed694068e78e
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89798
Analysis date: Wed, 16 Sep 2026 10:23:09 -0400
ActionableScore: 2
ActionableScore lower bound: 1
Actionable bucket: Ordinary Moderate
Manual review required: NO

Summary:
A race in `rpcrdma_rn_register()` can publish a notification object before `rn_done` is initialized, allowing concurrent RDMA device removal to call a NULL callback and crash the kernel.

======================================================================
ABOUT THIS REPORT
======================================================================

The original Linux kernel CVE announcement for CVE-2026-89798 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89798

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: NO

A detailed explanation of the methodology and priority rules is included
at the end of this message.

======================================================================
AL-KERNEL CLASSIFICATION RESULT
======================================================================

CVE-2026-89798	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-89798 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:N/I:N/A:H';CWE-362;**CWE-476;CWE-667;Other CVSS 'AV:L/AC:H/PR:H/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '4.7';DESCR 'rpcrdma_rn_register can publish a notification object in rd_xa before the rn_done callback pointer is armed. A concurrent RDMA device removal path can then find the object and call rn_done while it is still NULL, causing a NULL function pointer call and a kernel crash. For the CVSS the PR:L is used in the paranoid interpretation because a reduced privilege local actor with access to RDMA or NFS RDMA client operations may be able to race registration with device removal or recovery events. The issue is not directly network reachable and requires a local race with RDMA device lifecycle handling. Impact is denial of service via kernel crash. No concrete UAF, OOB write, information leak, or arbitrary write primitive is shown by the code context.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Ordinary Moderate (with actual score 2)	YES	REMOTE OOB INIT RACE NETWORK HARDWARE TEST  DECREASED_TO_MODERATE_REGULAR_FROM_MODERATE7_BASED_ON_GUESSCVSS_AND_CIANNH	YES	NO	checked

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

ActionableScore=2
ActionableScoreLower=1
PrimitiveClass=CLASS_C
RealUAF=NO
HostAdminRequired=UNKNOWN
NamespaceScopedPrivilege=NO
ImportantCandidate=NO
ManualReview=RECOMMENDED
AutoDowngradeAllowed=YES

## 1. ActionableScore

* Raw conservative score: 1
* Conservative score: 1
* Raw paranoid score: 2
* Paranoid score: 2
* Final recommended bucket: **Ordinary Moderate**::

No floor or cap was applied.

## 2. Signal breakdown

Conservative signals:

* +1 Weak indirect invalid-state primitive. `rpcrdma_rn_register()` publishes `rn` into `rd_xa` before `rn->rn_done` is initialized, allowing another path to observe a partially initialized notification object.
* +1 Reliable kernel crash / strong DoS. `rpcrdma_remove_one()` can call `rn->rn_done(rn)` while `rn_done` is still NULL, causing a NULL function pointer call and kernel crash.
* -1 Hard or one-shot race. The demonstrated trigger depends on RDMA device removal racing a fresh notification registration, which is a lifecycle/removal timing window rather than a normal packet-processing path.

Paranoid difference:

* The paranoid score withholds the -1 race penalty because repeated registration and recovery/removal attempts may be possible in some RDMA/NFS-RDMA operational environments, but this does not create a stronger exploitation primitive.

Important non-awarded signals:

* RealUAF not awarded. The established primitive is a NULL callback invocation from a partially initialized published object, not post-free use of an object.
* Concrete repeated release/refcount not awarded. The `rn_done` cleanup discussion in the commit explains why the new early store must be cleared on `xa_alloc()` failure, but the original vulnerable path does not establish a first release followed by a second lifetime mutation on the same dangling reference.
* Controlled reclaim not demonstrated.
* Arbitrary write not demonstrated.
* Callback hijack not demonstrated. The callback pointer is observed as NULL, not attacker-controlled.
* Confidentiality and integrity not awarded. No disclosure, overwrite, or meaningful security-state modification is established.
* Extra availability not awarded because the same kernel crash is already counted once.
* Remote reachability not established. The issue is in RDMA device lifecycle and RPC/RDMA registration handling, not direct processing of attacker network packets.

Repeated-release/refcount checklist:

* First release can end lifetime: not established.
* Exact same dangling reference survives: not established.
* Second operation mutates lifetime/refcount: not established for the original bug.
* Concrete repeated release/refcount +1 awarded: no.
* Paranoid repeated-refcount leverage +1 awarded: no.
* FLOOR 5 applied: no.

## 3. Reachability analysis

The trigger requires a fresh RPC/RDMA notification registration to race with RDMA device removal. The supplied evidence does not prove that an ordinary local user can perform the complete trigger chain. NFS-RDMA setup, RDMA device lifecycle operations, and device removal are commonly controlled by privileged services or host administration, but the prompt rules require positive evidence before setting `HostAdminRequired=YES`.

Therefore `HostAdminRequired=UNKNOWN`.

No evidence shows that the complete operation is namespace-scoped. RDMA device removal and RPC/RDMA client registration are not established as attacker-created namespace-local operations. Therefore `NamespaceScopedPrivilege=NO`.

No direct network trigger is established. A remote NFS/RDMA server or network peer might influence higher-level client behavior in some deployments, but the vulnerable operation described here is a local registration vs device-removal race. Treating it as network reachable would require additional evidence.

Prior setup likely includes an RDMA-capable environment and use of RPC-over-RDMA/NFS-RDMA, but this is not scored as a rare-hardware penalty because the prompt requires concrete evidence for such negative scoring. The race is tied to device lifecycle handling and is not shown to be a repeatedly attacker-controlled data-plane operation.

## 4. Primitive and severity interpretation

The exact primitive is publication of a partially initialized notification object in an xarray before its callback pointer is armed. A concurrent removal path can find the object and invoke the callback while it is NULL.

This is an active structured object state bug, but the stale/invalid use shown is limited to a NULL function pointer call. It is not a demonstrated UAF, not a double release, and not a controlled callback overwrite. The demonstrated consequence is kernel crash / DoS.

PrimitiveClass is `CLASS_C` because the observed behavior is primarily crash/invalid-state behavior without an independently useful corruption primitive.

ImportantCandidate is `NO` because there is no established UAF, repeated refcount mutation, controlled reclaim, attacker-controlled callback, arbitrary write, disclosure, authorization bypass, or privilege-escalation path.

AutoDowngradeAllowed is `YES` because the established impact is crash-only and the unresolved exploitation leverage is weak. Manual review is still recommended because this is a race around a kernel callback pointer in RDMA lifecycle code, but the evidence does not support preserving it as an Important-class candidate.

Likely operational severity is ordinary Moderate or Low-like DoS depending on deployment exposure and who can cause RDMA device removal.

## 5. One-sentence report phrase

A race in `rpcrdma_rn_register()` can publish a notification object before `rn_done` is initialized, allowing concurrent RDMA device removal to call a NULL callback and crash the kernel.

## 6. Manual review recommendation

MANUAL CHECK RECOMMENDED

Reason: the bug is a kernel race involving a callback pointer and RDMA device lifecycle handling, but the supplied evidence supports only NULL function pointer DoS and does not establish UAF, repeated refcount corruption, controlled callback hijack, or privilege escalation.

## 7. Call-site confidence

Call-site confidence: medium

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

Patch: rpcrdma: arm rn_done before publishing the notification [ Upstream
Commit: 3c97b8e76ca2bba9e8770571413aed694068e78e
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=3c97b8e76ca2bba9e8770571413aed694068e78e

Changed files:
  net/sunrpc/xprtrdma/ib_client.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=3c97b8e76ca2bba9e8770571413aed694068e78e

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

The original Linux kernel CVE announcement for CVE-2026-89798 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89798

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:H/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:N/I:N/A:H
  The Best / paranoid CVSS score: 4.7

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: 1
  Paranoid ActionableScore: 2

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).