[CVE-2026-98763][MODERATE REGULAR] usb: core: clear both ep_in and ep_out for non-ep0 control endpoints

[CVE-2026-98763][MODERATE REGULAR] usb: core: clear both ep_in and ep_out for non-ep0 control endpoints

[AL-KERNEL]

CVE: CVE-2026-98763
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: usb: core: clear both ep_in and ep_out for non-ep0 control endpoints
Commit: 7a9f61476b0e8002092a52ee8295edeeb0a28d1e
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=7a9f61476b0e8002092a52ee8295edeeb0a28d1e
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98763
Analysis date: Mon, 19 Oct 2026 03:47:24 -0400
ActionableScore: 2
ActionableScore lower bound: 1
Actionable bucket: Borderline / manual review recommended
Manual review required: YES

Summary:
A malicious USB peripheral can leave a dangling non-ep0 control-endpoint pointer in the opposite endpoint array, potentially causing a host kernel crash if USB core later consumes that stale entry, while controlled reclaim or stronger compromise is not demonstrated.

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

The original Linux kernel CVE announcement for CVE-2026-98763 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98763

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: YES

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

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

CVE-2026-98763	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-98763 CHECK WITH IMPACT FROM ORIG NN MODERATE | AEGIS IMPACT MODERATE_KPANIC AEGIS_AS=1 LEGACY=MODERATE RESULT=DIFF (final Legacy and Aegis differ; FINAL: MODERATE, Aegis: MODERATE_KPANIC)	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:P/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H';*CWE-416;*CWE-672;*CWE-664;Other CVSS 'AV:P/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:H';BEST CVSS score: '6.4';DESCR 'A non-ep0 USB control endpoint is stored in both dev ep_in and dev ep_out, but the old disable path cleared only the slot selected by the endpoint direction. The opposite slot could retain a pointer to a disabled and later freed usb_host_endpoint, creating a potential use-after-free when that slot is accessed again. A malicious USB device can expose the unusual bidirectional control endpoint and rely on interface disable, reset, or reconfiguration followed by another endpoint operation. For the CVSS the PR:N value is used because a physical USB device attacker does not need an account or privileges on the target host. The attack has AV:P and AC:H because physical or equivalent USB attachment is required and successful triggering depends on endpoint lifetime transitions and later reuse of the stale slot. Impact is at least denial of service through a kernel crash and may include confidentiality or integrity impact if the stale endpoint object is reused as controlled memory. Direct network reachability is not established, although USB/IP deployments may require separate review. Automated impact decision path. The original SimpleNet classifier selected MODERATE_KPANIC. The primary LLM estimated paranoid CVSS 6.4 and the production ActionableScore returned 2 with recommendation Borderline / manual review recommended. The production rules then evaluated CVSS thresholds and ActionableScore thresholds. Recorded rule transitions were ActionableScore thresholds stage MODERATE_KPANIC to MODERATE because the ActionableScore threshold rule was applied. The legacy KP specialist has no review artifact for this run. The legacy FD or DH downgrade specialist has no review artifact for this run. Exact triggered rule markers were DECREASED_TO_MODERATEREG_BASED_ON_ACTIONABLESCORELESSTHAN4 and V41_INHERITED_KPANIC_PROMOTION_BLOCKED and MANUAL_REVIEW_REQUIRED. The independent weaker Aegis shadow evaluated CVSS 0.0 and ActionableScore 1 with primitive class CLASS_C and manual review RECOMMENDED. Its independently reconciled impact was MODERATE_KPANIC. The models disagreed but the aggregation rules retained the production result. The final reported impact is MODERATE.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Borderline / manual review recommended (with actual score 2)	MAYBE	USB UAF HARDWARE LINUS  DECREASED_TO_MODERATEREG_BASED_ON_ACTIONABLESCORELESSTHAN4 V41_INHERITED_KPANIC_PROMOTION_BLOCKED MANUAL_REVIEW_REQUIRED	-	-	checked

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

ActionableScore=2
ActionableScoreLower=1
PrimitiveClass=CLASS_B
RealUAF=YES
HostAdminRequired=UNKNOWN
NamespaceScopedPrivilege=NO
ImportantCandidate=NO
ManualReview=RECOMMENDED
AutoDowngradeAllowed=YES
RequiresHostUID0=NO
AuthorizationDecisionBypass=NO
ProtectedResourceImpact=NONE
B4PendingReview=NO
B4PendingConfirmed=NO

1. ActionableScore

* Raw conservative score: 1
* Conservative score: 1
* Raw paranoid score: 2
* Paranoid score: 2
* No floor or cap was applied. The physical-access prerequisite and constrained stale-use evidence prevent application of the general UAF floor.
* Final recommended bucket: **Borderline / manual review recommended**::

2. Signal breakdown

Conservative signals:

* +2 Real UAF or lifetime violation. A non-ep0 control endpoint is referenced by both endpoint arrays, but only one slot was cleared. The other slot can retain the same `usb_host_endpoint` after its lifetime ends.
* +1 Reliable kernel crash or strong DoS. A later endpoint lookup can dereference the dangling structured endpoint while submitting or validating a URB, plausibly causing an Oops or panic.
* -2 Physical-only access. The supported external attacker model requires attachment of a malicious USB peripheral exposing the unusual non-ep0 control endpoint. Network reachability is not established.

Paranoid-only signal:

* +1 Weak active-structured-object concern. The stale entry is not passive metadata. USB core and HCD paths may consume endpoint state, including its enabled state, descriptor, and controller-private endpoint data. The supplied evidence does not establish controlled reclaim, so this remains weak unresolved leverage.

Not awarded:

* Controlled reclaim, attacker-shaped replacement contents, arbitrary write, callback control, disclosure, and privilege escalation are not demonstrated.
* Structured stale-object leverage was not fully awarded because no concrete post-reclaim field consumption or attacker-controlled replacement is shown.
* Attacker-influenced repeatable lifetime manipulation was not awarded because the peripheral controls descriptors, but complete control over disable, free, and later stale traversal is unresolved.
* No additional Availability point was awarded because it would duplicate the same possible kernel crash.
* No confidentiality or integrity points were awarded solely from theoretical reuse.
* No repeated release or refcount signal exists. The patch shows a retained pointer, not a second `put`, `free`, or refcount mutation.
* NEW52 B5-R does not apply. There is no retry-accumulated heap OOB write.
* NEW53 B3-PENDING does not apply. There is no shared-backing write.
* V32, V33, and V34 are not established. There is no checksum metadata write or independently demonstrated cross-boundary active-state corruption consumer.
* Authorization rules NEW54 and NEW55 do not apply because no authorization decision changes from DENY to ALLOW.

Repeated-release checks:

* First release can end endpoint lifetime: potentially yes.
* Exact reference survives in the opposite endpoint-array slot: yes.
* Second lifetime or refcount mutation: no evidence.
* Concrete repeated release/refcount +1: not awarded.
* Paranoid repeated-refcount +1: not awarded.
* FLOOR 5: not applied.

3. Reachability analysis

The primary attacker is a malicious physical USB device that supplies a non-zero control endpoint descriptor. Such a peripheral does not need a host account or host UID 0.

The tested `dummy_hcd` and raw-gadget path normally represents a privileged local test configuration and does not establish an ordinary local-user trigger. Conversely, testing through raw-gadget does not prove that a real malicious peripheral requires host-administrator privileges.

The peripheral controls endpoint descriptors and enumeration input, but the supplied context does not prove that it independently controls the complete disable, endpoint destruction, and subsequent stale lookup sequence. Therefore `HostAdminRequired` remains `UNKNOWN`.

No user-namespace or network-namespace capability forms part of the evidenced physical-device path. USB/IP could provide a non-physical transport in some deployments, but no reachable USB/IP call chain is supplied, so remote reachability is not scored.

No timing race is required. The defect is a deterministic alias-cleanup error once the relevant endpoint and lifecycle sequence occur.

4. Primitive and severity interpretation

The exact primitive is a dangling pointer to the same `struct usb_host_endpoint` in the uncleared endpoint-direction array. This is an active structured object rather than a fixed transfer buffer, and later USB-core lookup can directly imply post-lifetime use.

The patch does not establish an OOB write, repeated free, refcount mutation, controlled object replacement, or attacker-controlled post-free contents. It therefore supports `CLASS_B`, not `CLASS_A`.

The demonstrated root cause is stronger than passive bookkeeping because a freed endpoint remains discoverable through a live lookup array. The practical supported consequence is a possible kernel crash. More useful post-reclaim behavior remains theoretical.

`ImportantCandidate=NO` because meaningful exploitation leverage beyond a stale endpoint dereference is not established. `AutoDowngradeAllowed=YES` because this is presently a constrained physical-device Class B case without controlled reclaim, repeated lifetime mutation, or a demonstrated security-boundary consequence.

Likely operational severity is Low to Moderate, with manual validation warranted because the precise endpoint-free and subsequent-lookup call chain is absent from the supplied diff.

5. One-sentence report phrase

A malicious USB peripheral can leave a dangling non-ep0 control-endpoint pointer in the opposite endpoint array, potentially causing a host kernel crash if USB core later consumes that stale entry, while controlled reclaim or stronger compromise is not demonstrated.

6. Manual review recommendation

MANUAL CHECK RECOMMENDED

Confirm the exact endpoint deallocation site, subsequent `ep_in` or `ep_out` lookup, real-hardware trigger sequence, and whether any HCD path consumes security-relevant fields before an immediate fault.

7. Call-site confidence

Call-site confidence: medium

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

Patch: usb: core: clear both ep_in and ep_out for non-ep0 control endpoints
Commit: 7a9f61476b0e8002092a52ee8295edeeb0a28d1e
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=7a9f61476b0e8002092a52ee8295edeeb0a28d1e

Commit description:

In usb_enable_endpoint(), non-zero control endpoints (where
usb_endpoint_xfer_control(&ep->desc) is true) populate both
dev->ep_in[epnum] and dev->ep_out[epnum] with the same struct
usb_host_endpoint pointer.

However, usb_disable_endpoint() only clears either dev->ep_out[epnum] or
dev->ep_in[epnum] depending on the direction bit of epaddr when
reset_hardware is set, leaving the opposite direction's array slot
pointing to the disabled/freed endpoint.

Clear both dev->ep_out[epnum] and dev->ep_in[epnum] when disabling a
non-ep0 control endpoint with reset_hardware set, and update the
function kerneldoc accordingly.

Tested in QEMU against Linux 7.3.0-rc3 using dummy_hcd and raw-gadget to
enumerate a USB device with a non-zero control endpoint
(bEndpointAddress 0x01, bmAttributes USB_ENDPOINT_XFER_CONTROL),
verifying that both dev->ep_in[1] and dev->ep_out[1] are cleared when
the interface is disabled.

Fixes: 1da177e ("Linux-2.6.12-rc2")
Cc: stable <stable@kernel.org>
Assisted-by: LLM
Signed-off-by: Hui Peng <benquike@gmail.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Link: https://patch.msgid.link/20260920231702.110963-1-benquike@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Changed files:
  drivers/usb/core/message.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=7a9f61476b0e8002092a52ee8295edeeb0a28d1e

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

The original Linux kernel CVE announcement for CVE-2026-98763 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98763

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:P/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:H
  The Best / paranoid CVSS vector: AV:P/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H
  The Best / paranoid CVSS score: 6.4

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).