[CVE-2026-90359][MODERATE 7.0] bpf: Reject >8 byte return values on return-reading trampoline paths [ Upstream

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

CVE: CVE-2026-90359
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: bpf: Reject >8 byte return values on return-reading trampoline paths [ Upstream
Commit: d36ac53fc83baf843c51b7afd6d2471f36b713d8
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=d36ac53fc83baf843c51b7afd6d2471f36b713d8
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90359
Analysis date: Thu, 17 Sep 2026 14:04:54 -0400
ActionableScore: 5
ActionableScore lower bound: 4
Actionable bucket: Actionable Moderate
Manual review required: YES

Summary:
BPF return-reading trampoline paths allowed attachment to functions with return values larger than 8 bytes, causing the trampoline to preserve only half of the ABI return value and potentially corrupt the value observed by the real kernel caller.

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

The original Linux kernel CVE announcement for CVE-2026-90359 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90359

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-2026-90359	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-90359 CHECK WITH IMPACT FROM ORIG NN LOW	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:H';CWE-682;*CWE-20;CWE-754;Other CVSS 'AV:L/AC:L/PR:H/UI:N/S:U/C:N/I:L/A:L';BEST CVSS score: '5.8';DESCR 'BPF return reading trampoline paths accepted target functions with return values larger than 8 bytes. On common ABIs a 16 byte return value is passed in a register pair, but the trampoline preserves only the first 8 bytes, so fexit, fmod_ret, fsession, and struct_ops programs can corrupt the return value observed by the real caller and can observe only half of it. For the CVSS the PR:L is used for the paranoid score because a local user or service with delegated BPF tracing or struct_ops loading capability may be enough in some deployments, although typical systems restrict this to administrator level capabilities. The issue is not network reachable and requires local BPF program loading and attachment. Impact is mainly integrity corruption of kernel function results and possible denial of service, with no concrete information leak or arbitrary write primitive shown.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate (with actual score 5)	YES	OOB INIT BPF IMPROVEONLY LINUS  KPANIC INCREASED_FROM_LOW_BASED_ON_REQUIREMANUALCHECK	NO	NO	checked

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

ActionableScore=5
ActionableScoreLower=4
PrimitiveClass=CLASS_B
RealUAF=NO
HostAdminRequired=UNKNOWN
NamespaceScopedPrivilege=UNKNOWN
ImportantCandidate=NO
ManualReview=REQUIRED
AutoDowngradeAllowed=NO

## 1. ActionableScore

* Raw conservative score: 4
* Conservative score: 4. No floor or cap changed the score.
* Raw paranoid score: 5
* Paranoid score: 5. No floor or cap changed the score.
* Final recommended bucket: **Actionable Moderate**::

## 2. Signal breakdown

Conservative signals:

* +2: Concrete active semantic corruption primitive. BPF return-reading trampoline paths preserve only 8 bytes of a return value while allowing targets with larger return values, causing corruption of the value observed by the real caller.
* +1: Integrity impact is plausible. The bug can alter caller-visible kernel function return semantics, not merely produce passive bookkeeping disagreement.
* +1: Attacker-influenced target and attach-type selection exists once BPF attach privileges are available. The issue affects fexit, fmod_ret, fsession, related multi variants, and struct_ops.
* -0: No host-admin-equivalent penalty applied because the supplied evidence establishes privileged BPF gating but does not fully resolve whether the complete path always requires host-admin-equivalent privilege in the evaluated product.

Paranoid additional signal:

* +1: Target-dependent availability or stronger integrity consequence remains unresolved. Corrupting the high half of a kernel return value may be more serious for some eligible targets, but no arbitrary write, controlled reclaim, or demonstrated LPE is established.

Important non-awarded signals:

* RealUAF not awarded. No lifetime violation, stale pointer, double release, or post-free object use is shown.
* Controlled reclaim not demonstrated.
* Arbitrary or directed kernel write not demonstrated.
* Confidentiality not awarded because no disclosure or read-back path is established.
* Reliable kernel crash not awarded conservatively because the patch establishes return-value corruption, not a concrete panic path.
* Remote reachability not established. This is a local BPF control-plane issue.
* Strong LPE leverage not awarded. The patch does not show credential manipulation, callback hijack, arbitrary R/W, or a demonstrated privilege-escalation chain.
* BPF cleanup UAF floor logic does not apply because this is not a UAF.
* Concrete repeated release/refcount signal does not apply.

## 3. Reachability analysis

The trigger requires loading and attaching a BPF program through return-reading trampoline paths or using BPF struct_ops with a function returning more than 8 bytes. Relevant attach types include BPF_TRACE_FEXIT, BPF_MODIFY_RETURN, BPF_TRACE_FSESSION, their listed multi variants, and struct_ops.

This is not network reachable. Packet traffic alone does not reach the vulnerable condition.

The practical privilege boundary is BPF-specific. In common hardened deployments, BPF program loading and tracing attachment are gated by CAP_BPF, CAP_PERFMON, CAP_SYS_ADMIN, sysctl policy, or equivalent privileged setup. The supplied evidence does not prove that an ordinary unprivileged local user can complete load plus attach plus trigger. It also does not prove that host-admin-equivalent privilege is always required. Therefore HostAdminRequired is UNKNOWN.

NamespaceScopedPrivilege is UNKNOWN. The input does not establish that all required BPF load, attach, struct_ops, and trigger operations are available with only attacker-created namespace-scoped capabilities.

No special hardware dependency is established. No race or narrow timing condition is involved.

## 4. Primitive and severity interpretation

The exact primitive is caller-visible return-value corruption in BPF trampoline handling. A target function returning more than 8 bytes uses a register pair on common ABIs, but the trampoline saves and restores only the first 8 bytes. For return-reading attach types, the BPF program observes only part of the return value and the real caller may receive a corrupted return value.

This is an active semantic corruption primitive rather than a memory-safety primitive. It is not a UAF, not an OOB access, and not a double-free. It is also not merely a warning or validation-only cosmetic fix because the commit explicitly states that the real caller-visible return value is corrupted.

PrimitiveClass is CLASS_B because a concrete kernel semantic corruption primitive exists and target-dependent exploitation leverage needs review, but the patch does not establish CLASS_A evidence such as arbitrary write, controlled object replacement, callback control, root LPE, or a concrete authorization-boundary bypass.

Demonstrated consequence is integrity corruption of kernel function return semantics. Speculative consequences include DoS or stronger integrity effects depending on which eligible function is targeted. ImportantCandidate is NO because the currently supplied evidence does not establish an Important-grade exploitation primitive. AutoDowngradeAllowed is NO because automatically closing this as an ordinary low-risk issue would discard an unresolved active kernel return-value corruption primitive.

Likely operational severity is Actionable Moderate. It should receive manual security review, especially to identify affected 16-byte-return kernel functions reachable by these BPF attach mechanisms.

## 5. One-sentence report phrase

BPF return-reading trampoline paths allowed attachment to functions with return values larger than 8 bytes, causing the trampoline to preserve only half of the ABI return value and potentially corrupt the value observed by the real kernel caller.

## 6. Manual review recommendation

MANUAL CHECK REQUIRED

Manual review is required because the issue establishes active corruption of kernel function return semantics, and the real security impact depends on which eligible BPF targets with larger than 8 byte return values can be attached in the evaluated product.

## 7. Call-site confidence

Call-site confidence: high

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

Patch: bpf: Reject >8 byte return values on return-reading trampoline paths [ Upstream
Commit: d36ac53fc83baf843c51b7afd6d2471f36b713d8
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=d36ac53fc83baf843c51b7afd6d2471f36b713d8

Changed files:
  kernel/bpf/bpf_struct_ops.c
  kernel/bpf/verifier.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=d36ac53fc83baf843c51b7afd6d2471f36b713d8

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

The original Linux kernel CVE announcement for CVE-2026-90359 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-90359

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:H/UI:N/S:U/C:N/I:L/A:L
  The Best / paranoid CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:H
  The Best / paranoid CVSS score: 5.8

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: 4
  Paranoid ActionableScore: 5

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).