[CVE-2026-93264][MODERATE REGULAR] RDMA/efa: Fix PBL chunk length computation [ Upstream

[CVE-2026-93264][MODERATE REGULAR] RDMA/efa: Fix PBL chunk length computation [ Upstream

[AL-KERNEL]

CVE: CVE-2026-93264
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: RDMA/efa: Fix PBL chunk length computation [ Upstream
Commit: 665cd418b8561a099c3854443f8ad8ae751b72a0
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=665cd418b8561a099c3854443f8ad8ae751b72a0
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93264
Analysis date: Thu, 24 Sep 2026 13:39:08 -0400
ActionableScore: 1
ActionableScore lower bound: unknown
Actionable bucket: Ordinary Moderate
Manual review required: NO

Summary:
No one-sentence report phrase was found.

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

The original Linux kernel CVE announcement for CVE-2026-93264 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93264

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-2026-93264	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-93264 CHECK WITH IMPACT FROM ORIG NN LOW	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L';CWE-125;CWE-193;CWE-682;BEST CVSS score: '3.3';DESCR 'An incorrect indirect PBL chunk length calculation in the EFA RDMA driver occurs when the number of PBL pages is an exact multiple of EFA_PTRS_PER_CHUNK. The last fully populated chunk can be described as containing only its control area, causing the EFA device to reject memory registration. The same condition also reaches one element past the chunks array while selecting the next chunk buffer. The observed out of bounds value is not subsequently used because page iteration is complete, and the patch does not demonstrate an out of bounds write, use after free, DMA corruption, information leak, or privilege escalation primitive. For the CVSS the PR:L reflects that a local process needs access to the EFA RDMA device and permission to register memory. The issue is not directly network reachable. Impact is limited to failed RDMA memory registration or local availability degradation.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Ordinary Moderate (with actual score 1)	YES	OOB SIMPLEFIX NETWORK DMAorINTERRUPT HARDWARE LINUS MEMORY  DECREASED_TO_MODERATE_REGULAR_FROM_MODERATE7_BASED_ON_GUESSCVSS	NO	NO	checked

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

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

1. ActionableScore

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

No floor or cap was applied.

2. Signal breakdown

* +1 Weak corruption concern. The original loop increments `chunk_idx` after filling the final full PBL chunk and then reads `chunk_list->chunks[chunk_idx].buf` when `chunk_idx == chunk_list_size`. This is a one-past-end array access.
* +0 Availability. The established functional result is rejected EFA memory registration because a fully populated final chunk is incorrectly described as having only the control-buffer length. This affects the registration operation, but the patch does not establish a host crash, device reset, persistent resource exhaustion, or system-wide DoS.

Important non-awarded signals:

* Controlled reclaim is not demonstrated.
* Arbitrary or directed kernel write is not demonstrated.
* The patch does not establish an OOB write. The affected operation reads the `buf` field of a one-past-end array element.
* The OOB-read value is not subsequently used on the shown exact-multiple path because processing has reached the final page and the iterator exits.
* Information disclosure is not demonstrated. No read-back, userspace copy, network output, or device-visible disclosure path is shown.
* Integrity impact is not demonstrated.
* Remote reachability is not established. Registering an MR is a local RDMA verbs operation.
* No UAF, stale reference, RCU lifetime mismatch, refcount mutation, double release, or ownership violation is present.
* No reliable kernel crash is established. Therefore no crash or additional availability point is awarded.
* No privilege escalation or security-boundary bypass is established.
* No negative privilege, niche-subsystem, or hardware penalty is applied because the supplied evidence does not establish that the complete trigger requires host-admin-equivalent privilege or that an attacker cannot access an already exposed EFA RDMA device.

3. Reachability analysis

The affected path is reached while registering a memory region through the EFA RDMA verbs driver. An attacker would need access to an EFA RDMA device and permission to create the relevant RDMA resources and register memory.

The supplied patch does not establish whether access to the EFA verbs device is available to an ordinary local user, a delegated service account, a container workload, or only a host administrator. Therefore `HostAdminRequired=UNKNOWN`.

The patch also does not show the relevant device-node permissions, RDMA cgroup policy, device assignment rules, user-namespace behavior, or whether the complete MR registration operation works in an attacker-controlled namespace. Therefore `NamespaceScopedPrivilege=UNKNOWN`.

An EFA device must already be provisioned and exposed by the host or cloud environment. However, this is a deployment prerequisite, not evidence that the attacker necessarily requires host-admin privileges after device access has been granted. The issue is not directly network-triggerable. Remote traffic does not directly invoke the vulnerable PBL construction path.

No race is involved. The fault condition is deterministic when the indirect PBL page count is an exact multiple of `EFA_PTRS_PER_CHUNK`.

4. Primitive and severity interpretation

The exact defect has two related effects.

First, when the PBL page count is exactly divisible by `EFA_PTRS_PER_CHUNK`, the final full chunk length is overwritten with the control-buffer-only size instead of retaining `EFA_CHUNK_USED_SIZE`. The DMA mapping and device-visible PBL metadata are therefore inconsistent with the actual populated chunk. The documented consequence is that the EFA device rejects memory registration.

Second, after writing the final page pointer into a full final chunk, the original loop increments `chunk_idx` and accesses `chunks[chunk_idx].buf` even though `chunk_idx` equals `chunk_list_size`. This is a limited one-past-end array read.

This is not an active structured-object lifetime violation. There is no release, free, put, refcount decrement, stale pointer, object replacement, callback dispatch, DMA write to an invalid target, or reuse of freed storage. The stale/OOB value is selected after the final payload entry and is not consumed by a subsequent iteration in the shown control flow.

The appropriate classification is `CLASS_C`, representing invalid-state behavior and a constrained invalid array access without an independently useful corruption primitive. The demonstrated operational impact is failed memory registration for a local EFA RDMA consumer.

`ImportantCandidate=NO` because the patch does not establish meaningful exploitation leverage beyond the failed MR operation and the unused one-past-end read. `AutoDowngradeAllowed=YES` because there is no unresolved UAF, write primitive, active-state corruption, authorization bypass, disclosure path, or repeated lifetime mutation that would be lost through automated lower-priority handling.

5. One-sentence report phrase

An EFA RDMA indirect PBL length calculation error can reject memory registration when the PBL page count exactly fills the final chunk, and the same condition performs a bounded one-past-end chunks array read whose result is not subsequently used.

6. Manual review recommendation

MANUAL CHECK RECOMMENDED

Confirm the allocation layout of `chunk_list->chunks`, compiler-generated behavior for the one-past-end field access, and whether any unshown caller or error path can consume the selected `cur_chunk_buf` value after the final iterator entry. Based on the supplied patch alone, this is not an Important candidate.

7. Call-site confidence

Call-site confidence: medium

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

Patch: RDMA/efa: Fix PBL chunk length computation [ Upstream
Commit: 665cd418b8561a099c3854443f8ad8ae751b72a0
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=665cd418b8561a099c3854443f8ad8ae751b72a0

Changed files:
  drivers/infiniband/hw/efa/efa_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=665cd418b8561a099c3854443f8ad8ae751b72a0

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

The original Linux kernel CVE announcement for CVE-2026-93264 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-93264

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: Not available
  The Best / paranoid CVSS vector: AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L
  The Best / paranoid CVSS score: 3.3

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: 1

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).