[CVE-2026-98382][MODERATE 7.0] bpf: Reject dev-bound-only programs on other devices [ Upstream

[CVE-2026-98382][MODERATE 7.0] bpf: Reject dev-bound-only programs on other devices [ Upstream

[AL-KERNEL]

CVE: CVE-2026-98382
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: bpf: Reject dev-bound-only programs on other devices [ Upstream
Commit: e57f04194361574493e3e6cf0b9e9c69e4ae9791
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=e57f04194361574493e3e6cf0b9e9c69e4ae9791
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98382
Analysis date: Fri, 09 Oct 2026 22:08:52 -0400
ActionableScore: 5
ActionableScore lower bound: unknown
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-2026-98382 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98382

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-2026-98382	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-98382 CHECK WITH IMPACT FROM ORIG NN LOW | AEGIS IMPACT MODERATE_KPANIC AEGIS_AS=2 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:L/PR:L/UI:N/S:U/C:L/I:N/A:H';CWE-843;CWE-125;**CWE-476;CWE-863;Other CVSS 'AV:L/AC:L/PR:H/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '6.1';DESCR 'A device matching logic error in __bpf_offload_dev_match can allow a device bound XDP program to be attached to a different network device. Driver specific metadata kfuncs may then interpret the target driver xdp_buff as a larger and incompatible private structure. The demonstrated veth to tun combination reads beyond the actual tun xdp_buff and causes a general protection fault followed by a kernel panic. For the CVSS the PR:L in the paranoid score reflects that CAP_BPF and CAP_NET_ADMIN may be delegated to a constrained service or container root without granting complete host administration. The base assessment uses PR:H because both capabilities are normally restricted and provide substantial control over BPF and networking. The issue is locally triggered through BPF_LINK_CREATE and is not directly network reachable. Impact is at least denial of service through a kernel panic. A limited confidentiality impact is plausible because a driver metadata kfunc can read fields beyond an incompatible object and may return the resulting value to the BPF program. Integrity impact is not assigned because no UAF OOB write or arbitrary write primitive is established. Privilege escalation is not demonstrated but the cross driver type confusion and out of bounds access require manual review.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 5)	MAYBE	KASAN OOB NULLPTR BPF HARDWARE LINUS  KPANIC INCREASED_FROM_LOW_BASED_ON_REQUIREMANUALCHECK	-	-	checked

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

ActionableScore=5
PrimitiveClass=CLASS_A
RealUAF=NO
HostAdminRequired=UNKNOWN
NamespaceScopedPrivilege=UNKNOWN
ImportantCandidate=YES
ManualReview=REQUIRED
AutoDowngradeAllowed=NO
RequiresHostUID0=UNKNOWN
AuthorizationDecisionBypass=NO
ProtectedResourceImpact=NONE
B4PendingReview=NO
B4PendingConfirmed=NO

1. ActionableScore

* Raw conservative score: 3
* Conservative score: 5
* Raw paranoid score: 4
* Paranoid score: 5
* Applied conservative manual-review floor: 3 -> 5 because the patch establishes attacker-selected cross-driver semantic reinterpretation and a concrete kernel OOB read.
* Applied paranoid manual-review floor: 4 -> 5 for the same established primitive. The BPF cap was checked but did not reduce this score.
* Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**::

2. Signal breakdown

Positive conservative signals:

* +2 Strong attacker-influenced semantic reinterpretation. A program specialized for one netdevice driver can be attached to another driver, causing a tun `xdp_buff` to be interpreted as a larger `veth_xdp_buff`.
* +1 Reliable kernel crash. The supplied trace establishes an Oops followed by a fatal kernel panic.

Additional paranoid signal:

* +1 Confidentiality plausible. A metadata kfunc reads driver-private fields outside the actual target object and normally returns metadata to the BPF program. This is a technically plausible limited disclosure path, although the demonstrated combination faults before useful disclosure is shown.

Negative signals:

* No negative privilege score. CAP_BPF and CAP_NET_ADMIN are established, but the evidence does not resolve whether the complete trigger requires host-admin-equivalent capabilities or can run under delegated capabilities.
* No race penalty. The mismatch is deterministic after the incompatible attachment.
* No rare-hardware penalty. The demonstrated path uses software devices, veth and tun.

Important non-awarded signals:

* No UAF, controlled reclaim, OOB write, arbitrary write, callback hijack, or demonstrated LPE.
* No extra availability point because it would double-count the same kernel panic.
* No local-unprivileged bonus because ordinary-user access to the complete BPF load, bind, attach, and trigger sequence is not established.
* No remote-reachability bonus.
* NEW52 B5-R does not apply because there is no retry-accumulated OOB write.
* NEW53 B3-PENDING does not apply because no write to shared backing is established.
* NEW54 B4-PENDING does not apply because this is a device-compatibility validation failure, not an authorization DENY-to-ALLOW transition.

3. Reachability analysis

The demonstrated actor needs CAP_BPF and CAP_NET_ADMIN, creates or possesses a device-bound XDP program, invokes `BPF_LINK_CREATE` with a different target ifindex, and then exercises the target data path. The supplied evidence does not prove that host UID 0 is required because capabilities can exist independently of numeric UID.

The complete namespace boundary is unresolved. Network-device administration may be namespace-scoped, but the evidence does not establish that program loading, device-bound setup, attachment, tun access, and triggering all work within an attacker-created namespace. Therefore both host-admin equivalence and complete namespace-scoped reachability remain unknown.

No prior firmware or physical hardware is required by the demonstrated veth-to-tun case. The issue is local rather than remotely network-triggerable. It is not a timing race and does not require repeated probabilistic attempts.

4. Primitive and severity interpretation

The exact primitive is a controlled cross-driver type confusion or semantic reinterpretation followed by a heap-or-stack-adjacent OOB read from an incompatible `xdp_buff` representation. The target buffer is an active packet-processing object, not stale storage, but there is no lifetime violation and therefore no real UAF.

The behavior is broader than an ordinary invalid pointer supplied accidentally. The caller selects the bound program and incompatible target device, after which a driver-specific metadata kfunc consumes the target object using the wrong private layout. This establishes Class A semantic reinterpretation under the prompt rules, while the demonstrated consequence remains kernel panic.

A limited disclosure is defensible as a paranoid possibility because metadata kfunc output can be consumed by the BPF program. Useful disclosure, arbitrary read, write corruption, and privilege escalation are not demonstrated. ImportantCandidate is YES because the cross-driver interpretation primitive and resulting OOB access require review across other metadata kfunc and driver combinations. AutoDowngradeAllowed is NO because automatically reducing the issue to crash-only behavior would discard the established attacker-selected semantic mismatch.

Likely operational severity is Actionable Moderate. Important classification remains a review candidate rather than a demonstrated LPE or high-impact confidentiality/integrity vulnerability.

5. One-sentence report phrase

A CAP_BPF and CAP_NET_ADMIN caller can attach a device-bound XDP program to an incompatible netdevice, causing driver-specific metadata kfuncs to reinterpret and read beyond the target xdp_buff and potentially panic the kernel.

6. Manual review recommendation

MANUAL CHECK REQUIRED

Review all device-bound metadata kfunc and cross-driver combinations for non-crashing disclosure or stronger memory-access effects beyond the demonstrated veth-to-tun panic.

7. Call-site confidence

Call-site confidence: high

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

Patch: bpf: Reject dev-bound-only programs on other devices [ Upstream
Commit: e57f04194361574493e3e6cf0b9e9c69e4ae9791
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=e57f04194361574493e3e6cf0b9e9c69e4ae9791

Changed files:
  drivers/net/veth.c
  drivers/net/tun.c
  fs/read_write.c
  arch/x86/entry/syscall_64.c
  arch/x86/entry/entry_64.S
  kernel/bpf/offload.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=e57f04194361574493e3e6cf0b9e9c69e4ae9791

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

The original Linux kernel CVE announcement for CVE-2026-98382 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98382

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:N/A:H
  The Best / paranoid CVSS vector: AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:H
  The Best / paranoid CVSS score: 6.1

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: 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).