[CVE-2026-98430][MODERATE 7.0] RDMA/efa: Keep admin queues alive while IRQ is registered [ Upstream

[CVE-2026-98430][MODERATE 7.0] RDMA/efa: Keep admin queues alive while IRQ is registered [ Upstream

[AL-KERNEL]

CVE: CVE-2026-98430
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: RDMA/efa: Keep admin queues alive while IRQ is registered [ Upstream
Commit: 0a8a105978e5ace0d885964317c5aed8ab4c0a0c
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=0a8a105978e5ace0d885964317c5aed8ab4c0a0c
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98430
Analysis date: Thu, 15 Oct 2026 10:24:59 -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-2026-98430 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98430

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-2026-98430	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-98430 CHECK WITH IMPACT FROM ORIG NN MODERATE | AEGIS IMPACT MODERATE_KPANIC AEGIS_AS=0 LEGACY=MODERATE_KPANIC RESULT=SAME (legacy and Aegis reconciliation agree; DECISION: both 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-664;*CWE-672;Other CVSS 'AV:L/AC:H/PR:H/UI:N/S:U/C:L/I:L/A:H';BEST CVSS score: '7';DESCR 'The EFA management IRQ handler can access the admin completion queue and async event queue outside their valid lifetime. During device initialization an interrupt can arrive before these queues are fully constructed, while during removal an interrupt can race with queue destruction and dereference freed queue state. The device can also retain queue DMA addresses on an IRQ registration failure, creating a potential DMA access to released memory if the queues are destroyed before the device is reset. This is a race condition with a concrete UAF candidate and a plausible memory corruption path rather than a warning only issue. For the CVSS the PR:L value is used in the paranoid score because a local RDMA user or reduced capability service may generate device activity while reset or removal occurs independently, although reliable direct control of driver teardown normally requires PR:H. AC:H reflects the requirement to hit the IRQ lifetime window. The issue is not demonstrated to be directly network reachable because remote RDMA traffic alone does not provide control over queue destruction or PCI driver removal. Impact is at least local denial of service through a kernel crash and in the worst case may include confidentiality and integrity impact if stale CPU or DMA accesses reach reused memory. Automated impact decision path. The original SimpleNet classifier selected MODERATE_KPANIC. The primary LLM estimated paranoid CVSS 7 and the production ActionableScore returned 4 with recommendation Borderline / manual review recommended. The production rules then evaluated CVSS thresholds and ActionableScore thresholds. No severity changing production threshold or specialist override was recorded. The legacy KP specialist produced a review artifact and its decision was processed by the applicable KP rules. The legacy FD or DH downgrade specialist has no review artifact for this run. The independent weaker Aegis shadow evaluated CVSS 2.0 and ActionableScore 0 with primitive class CLASS_B and manual review RECOMMENDED. Its independently reconciled impact was MODERATE_KPANIC. The shadow agreed with the final production impact and aggregation did not change that impact. The final reported impact is MODERATE_KPANIC.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Borderline / manual review recommended (with actual score 4)	MAYBE	DANGER INIT UAF NETWORK DMAorINTERRUPT HARDWARE LINUS KPANIC 	-	-	checked

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

ActionableScore=4
PrimitiveClass=CLASS_B
RealUAF=YES
HostAdminRequired=UNKNOWN
NamespaceScopedPrivilege=UNKNOWN
ImportantCandidate=NO
ManualReview=RECOMMENDED
AutoDowngradeAllowed=YES
RequiresHostUID0=UNKNOWN
AuthorizationDecisionBypass=NO
ProtectedResourceImpact=NONE
B4PendingReview=NO
B4PendingConfirmed=NO

1. ActionableScore

* Raw conservative score: 4
* Conservative score: 4
* Raw paranoid score: 4
* Paranoid score: 4
* No floor or cap changed either score.
* Final recommended bucket: **Borderline / manual review recommended**::

2. Signal breakdown

* +2 Real UAF and lifetime violation. During removal, `efa_com_admin_destroy()` could release the admin completion and async event queues while the registered management IRQ handler could still access them.
* +1 Reliable kernel crash or strong DoS. An IRQ handler dereferencing released queue state can realistically produce an invalid access, Oops, or panic.
* +1 Privileged device-management memory corruption path. The device retains DMA addresses for the queues, and the added reset before destruction shows that DMA ownership must be revoked before releasing the backing resources.
* No timing penalty. The bug involves interrupt timing, but the patch does not establish an unusually narrow or difficult-to-repeat race.
* No privilege penalty. Host-admin control is plausible for probe, removal, and reset, but the supplied evidence does not positively establish the privilege required for the complete trigger.
* No hardware rarity penalty. EFA hardware is required, but the evidence does not establish physical attacker access or sufficient rarity to justify a numeric reduction.
* Controlled reclaim, attacker-selected target, attacker-controlled DMA payload, arbitrary write, callback control, disclosure, and LPE are not demonstrated.
* Structured stale-object leverage is not awarded because no attacker influence over reclaimed queue state or meaningful replacement is established.
* No extra Availability point is awarded because it would duplicate the possible kernel crash.
* NEW52 B5-R does not apply. No retry-accumulated heap OOB write exists.
* NEW53 B3-PENDING does not apply. No shared-backing or COW write is established.
* NEW54 B4-PENDING does not apply. There is no authorization decision or DENY-to-ALLOW transition.

3. Reachability analysis

The vulnerable windows occur during EFA probe, IRQ-registration failure, device reset, or removal. Device initialization and teardown are normally controlled by the PCI and driver-management paths, but the supplied patch does not establish whether a below-host-admin RDMA user can independently force the complete transition.

A local RDMA workload may influence device activity or interrupt generation, but that does not prove control over queue destruction or driver removal. Conversely, the absence of a demonstrated unprivileged trigger does not prove that host-admin privilege is mandatory. Therefore `HostAdminRequired=UNKNOWN`.

No evidence shows that the complete operation is available through attacker-created user or network namespaces. EFA hardware and an active driver are prerequisites. Direct remote reachability is not established because network traffic alone is not shown to initiate the required queue-lifetime transition. Race repeatability and attacker control over the interrupt window remain unresolved.

4. Primitive and severity interpretation

The concrete primitive is an asynchronous post-lifetime access. The management IRQ remains registered while the admin completion queue and async event queue can be destroyed, allowing the handler to consume stale queue state. The failed-IRQ path also risks leaving the device aware of DMA addresses until reset.

This is `CLASS_B`, not `CLASS_A`. The removal path establishes a real UAF, but no controlled reclaim, directed DMA write, attacker-shaped replacement object, or independently useful post-free primitive is demonstrated. The affected queues are structured objects, but the demonstrated stale use is tied to probe, failure cleanup, reset, and removal rather than general attacker-controlled active-object replacement.

There is no concrete repeated release or refcount mutation. The patch changes queue and IRQ destruction order, but it does not show the same dangling reference receiving a second `put`, `free`, or refcount decrement.

The demonstrated practical consequence is kernel crash or device/host instability. A device DMA access to released memory is technically plausible and motivates review, but its contents, timing, destination reuse, and attacker influence are unresolved. Therefore this remains a constrained Class B lifetime bug rather than a strong Important candidate.

`ImportantCandidate=NO` because meaningful exploitation leverage beyond crash has not been established. `AutoDowngradeAllowed=YES` because this is a teardown and initialization lifetime bug without controlled reclaim, repeated lifetime mutation, authorization bypass, or demonstrated non-crash corruption primitive.

5. One-sentence report phrase

The EFA management IRQ can outlive the admin queues it accesses, creating a teardown-time UAF and possible stale DMA exposure whose demonstrated practical impact is kernel crash or device instability.

6. Manual review recommendation

MANUAL CHECK RECOMMENDED

Review should confirm whether ordinary RDMA users can induce management interrupts during reset or removal and whether the device can write to released queue memory before reset, since either fact could strengthen reachability or corruption leverage.

7. Call-site confidence

Call-site confidence: medium

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

Patch: RDMA/efa: Keep admin queues alive while IRQ is registered [ Upstream
Commit: 0a8a105978e5ace0d885964317c5aed8ab4c0a0c
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=0a8a105978e5ace0d885964317c5aed8ab4c0a0c

Changed files:
  drivers/infiniband/hw/efa/efa_com.c
  drivers/infiniband/hw/efa/efa_main.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=0a8a105978e5ace0d885964317c5aed8ab4c0a0c

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

The original Linux kernel CVE announcement for CVE-2026-98430 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98430

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: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: 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).