[CVE-2026-63896][MODERATE 7.0] usb: gadget: composite: fix integer underflow in WebUSB GET_URL handling

[CVE-2026-63896][MODERATE 7.0] usb: gadget: composite: fix integer underflow in WebUSB GET_URL handling

[AL-KERNEL]

CVE: CVE-2026-63896
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: usb: gadget: composite: fix integer underflow in WebUSB GET_URL handling
Commit: 046870ff6b6f7b743c953c061043a9b30700d491
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=046870ff6b6f7b743c953c061043a9b30700d491
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-63896
Analysis date: Sun, 19 Jul 2026 23:16:12 -0400
ActionableScore: 6
ActionableScore lower bound: 4
Actionable bucket: Actionable Moderate at minimum / Strong Important candidate for manual review
Manual review required: YES

Summary:
A malicious USB host can trigger an integer underflow in WebUSB GET_URL handling on configured USB gadgets, causing a huge memcpy into the gadget request buffer and resulting in kernel memory corruption and likely device crash.

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

The original Linux kernel CVE announcement for CVE-2026-63896 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-63896

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-2026-63896	MODERATE	CHECK WITH IMPACT FROM ORIG NN LOW	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:P/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H';CWE-191;*CWE-787;CWE-122;Other CVSS 'AV:P/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '6.8';DESCR 'WebUSB GET_URL handling in usb gadget composite has an integer underflow when a USB host sends wLength smaller than the URL descriptor header. The underflow makes landing_page_length wrap and can drive a very large memcpy from cdev landing_page into cdev req buf, producing a slab out of bounds write and a kernel crash. For the CVSS the PR:N is used because the attacking USB host does not need privileges on the gadget system beyond access to the USB control endpoint. The attack is not Internet reachable and requires a malicious or compromised USB host connected to a gadget with webusb use enabled and landingPage configured. Impact is at least denial of service, and because this is an out of bounds write in kernel memory, the paranoid interpretation should treat confidentiality and integrity impact as possible and require manual review.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate at minimum / Strong Important candidate for manual review (with actual score 5)	YES	WRITE KASAN OOB USB LINUS PACKET  KPANIC INCREASED_TO_MODERATE_FROM_LOW_BASED_ON_GUESSCVSS	YES	NO	checked

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

ActionableScore=6
ActionableScoreLower=4

## 1. ActionableScore

* Conservative score: 4
* Paranoid score: 6
* Final recommended bucket: **Actionable Moderate at minimum / Strong Important candidate for manual review**::

## 2. Signal breakdown

Conservative signals:

* Generic memory corruption: +2. The patch fixes an unsigned integer underflow that can drive a huge `memcpy()` into `cdev->req->buf`, producing a slab out-of-bounds write.
* Reliable kernel crash / strong DoS: +1. KASAN reports slab-out-of-bounds and FORTIFY_SOURCE traps the oversized memcpy.
* Privileged kernel/device-management memory corruption path: +1. The corruption occurs in the USB gadget control request handling path, which is a trusted kernel device-management context influenced by an external USB host.
* Physical-only condition: -2. The attacker must act as a USB host connected to the gadget device.
* Configuration-dependent exposure: -1. The gadget must have `webusb/use=1` and a configured `landingPage`.
* Availability impact realistic: +1. A single malformed SETUP request can trigger a host-wide kernel crash on affected gadget configurations.

Paranoid additional signals:

* Confidentiality impact plausible: +1. The oversized copy reads far beyond the intended landing page buffer before writing, so unintended kernel memory movement is plausible even if no direct disclosure path is shown.
* Integrity impact plausible: +1. The destination is kernel memory beyond the intended USB request buffer, so kernel memory corruption is plausible.
* Weak LPE concern: +1. This is not a demonstrated LPE primitive, but an attacker-triggered OOB write in kernel memory is enough to require manual review.

## 3. Reachability analysis

The bug is reachable by a malicious or compromised USB host sending a malformed WebUSB GET_URL SETUP packet with `wLength` smaller than the URL descriptor header. No privileges are required inside the attacked gadget system, so from the gadget OS perspective the triggering side is PR:N. This is not network reachable over IP and should not be scored as remote network exposure. The realistic deployment requirement is a Linux USB gadget configuration with WebUSB enabled and a landing page configured, which is not universal but is a valid production mode for embedded and device-side USB systems. Namespaces and containers are not the main factor here because the attacker is external to the gadget OS and interacts through the physical USB control endpoint.

Call-site confidence: high. The vulnerable arithmetic and the subsequent memcpy are both shown in the provided patch context.

## 4. Severity interpretation

This is more serious than an ordinary Moderate crash bug because the failure mode is an integer underflow leading to a very large kernel memcpy and a slab out-of-bounds write. Realistically, the most demonstrated impact is a reliable local-physical DoS of the gadget device. Theoretical confidentiality or integrity impact is plausible because this is kernel memory corruption, but the patch does not show attacker-controlled target selection, controlled payload placement, object reclaim, callback control, or a demonstrated privilege-escalation chain. Therefore the conservative score remains Borderline/manual-review, while the paranoid score reaches Actionable Moderate.

## 5. One-sentence report phrase

A malicious USB host can trigger an integer underflow in WebUSB GET_URL handling on configured USB gadgets, causing a huge memcpy into the gadget request buffer and resulting in kernel memory corruption and likely device crash.

## 6. Manual review recommendation

MANUAL CHECK REQUIRED

This should not be auto-closed because it is an externally triggerable kernel OOB write, even though the exposure is physical USB rather than network and practical exploitation beyond DoS is not demonstrated.

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

Patch: usb: gadget: composite: fix integer underflow in WebUSB GET_URL handling
Commit: 046870ff6b6f7b743c953c061043a9b30700d491
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=046870ff6b6f7b743c953c061043a9b30700d491

Commit description:

The WebUSB GET_URL handler in composite_setup() narrows
landing_page_length to fit the host-supplied wLength using

	landing_page_length = w_length
		- WEBUSB_URL_DESCRIPTOR_HEADER_LENGTH + landing_page_offset;

If wLength is smaller than WEBUSB_URL_DESCRIPTOR_HEADER_LENGTH the
unsigned subtraction wraps, and the subsequent

	memcpy(url_descriptor->URL,
	       cdev->landing_page + landing_page_offset,
	       landing_page_length - landing_page_offset);

ends up copying close to UINT_MAX bytes from cdev->landing_page into
cdev->req->buf.  KASAN reports a slab-out-of-bounds in composite_setup
on the kmalloc-2k gadget_info allocation, and FORTIFY_SOURCE traps the
memcpy as a 4294967293-byte field-spanning write into
url_descriptor->URL (size 252).

A USB host can reach this from a single SETUP packet against any
gadget that has webusb/use=1 and a landingPage configured.

Handle the small-wLength case before the math: when the host requested
fewer bytes than the URL descriptor header, only the header is
meaningful and no URL bytes need to be copied.  Setting
landing_page_length to landing_page_offset makes the existing memcpy a
no-op and leaves the descriptor returned to the host unchanged for all
larger wLength values.

Fixes: 93c4739 ("usb: gadget: add WebUSB landing page support")
Cc: stable <stable@kernel.org>
Signed-off-by: Jeremy Erazo <mendozayt13@gmail.com>
Link: https://patch.msgid.link/20260512160530.352318-1-mendozayt13@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Changed files:
  drivers/usb/gadget/composite.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=046870ff6b6f7b743c953c061043a9b30700d491

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

The original Linux kernel CVE announcement for CVE-2026-63896 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-63896

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:L/PR:N/UI:N/S:U/C:N/I:N/A:H
  The Best / paranoid CVSS vector: AV:P/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
  The Best / paranoid CVSS score: 6.8

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: 6

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).