[CVE-2026-74691][MODERATE REGULAR] net: thunderbolt: Tear down DMA paths before stopping the rings [ Upstream

[CVE-2026-74691][MODERATE REGULAR] net: thunderbolt: Tear down DMA paths before stopping the rings [ Upstream

[AL-KERNEL]

CVE: CVE-2026-74691
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: net: thunderbolt: Tear down DMA paths before stopping the rings [ Upstream
Commit: 7cce39109206bc5497e0953806563644b88bfc44
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=7cce39109206bc5497e0953806563644b88bfc44
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-74691
Analysis date: Sat, 22 Aug 2026 14:23:00 -0400
ActionableScore: 2
ActionableScore lower bound: 0
Actionable bucket: Ordinary Moderate / hardware-specific DoS, manual review recommended
Manual review required: YES

Summary:
Thunderbolt networking teardown stopped rings and freed frame buffers before disabling DMA paths, which can leave pending DMA hops stuck on affected host routers and eventually take down the XDomain control channel until controller power cycle.

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

The original Linux kernel CVE announcement for CVE-2026-74691 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-74691

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-2026-74691	MODERATE	CHECK WITH IMPACT FROM ORIG NN LOW	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:A/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H';CWE-404;*CWE-664;**CWE-400;Other CVSS 'AV:L/AC:L/PR:H/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '5.3';DESCR 'Thunderbolt networking teardown can leave DMA paths active after Rx and Tx rings are stopped and frame buffers are unmapped and freed. On affected host routers this can make the hop pending bit never clear, causing repeated 500 ms timeout failures and eventually taking the XDomain control channel down until a controller power cycle. For the CVSS the PR:N value is used for the paranoid case because a connected Thunderbolt peer or physical link event may be able to trigger teardown without a local account, while the normal administrative teardown path is closer to PR:H. The issue is not Internet reachable and requires Thunderbolt link proximity or local device administration. Impact is denial of service through loss of the Thunderbolt network peer or controller function. No direct confidentiality or integrity impact is claimed because the patch context shows a stuck DMA teardown and controller hang, not a demonstrated arbitrary DMA write or information leak primitive.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Ordinary Moderate / hardware-specific DoS, manual review recommended (with actual score 1)	YES	INIT NETWORK DMAorINTERRUPT HARDWARE PACKET  DECREASED_TO_MODERATE_REGULAR_FROM_MODERATE7_BASED_ON_GUESSCVSS_AND_CIANNH	NO	NO	checked

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

ActionableScore=2
ActionableScoreLower=0

## 1. ActionableScore

* Conservative score: 0
* Paranoid score: 2
* Final recommended bucket: **Ordinary Moderate / hardware-specific DoS, manual review recommended**::

## 2. Signal breakdown

Conservative signals:

* Reliable kernel crash / strong DoS: +1
  The issue can repeatedly break Thunderbolt networking teardown and eventually take down the XDomain control channel.
* Availability impact realistic: +1
  The impact can persist until controller power cycle, so it is more than a transient warning.
* Requires admin/root/CAP_NET_ADMIN in typical deployments: -2
  The normal trigger is local device or netdev administration, such as bringing the interface down, bonding changes, unload, reset, or teardown.
* Physical-only or rare hardware-only condition: -2
  The demonstrated failure depends on Thunderbolt hardware and was specifically observed on ASMedia ASM4242 host router behavior.
* Conservative floor applied: score treated as 0.

Paranoid additional interpretation:

* Firmware-mediated external influence: +1
  A connected Thunderbolt peer or physical link event may influence teardown behavior without a local user account.
* Privileged kernel/device-management DMA-adjacent state corruption path: +1
  The bug involves DMA paths, ring descriptor bases, unmapped frame buffers, and teardown ordering in a trusted device-management path. This is not enough to claim arbitrary DMA memory corruption, but it justifies manual review.
* Physical-only or rare hardware-only condition: -2
  Still applies because realistic triggering requires Thunderbolt proximity and affected router behavior.

Not awarded:

* No generic memory corruption +2.
  The patch does not demonstrate UAF reclaim, arbitrary write, double free, OOB write, type confusion, or controlled object replacement.
* No privilege escalation plausible +1 or +2.
  The evidence supports device/controller DoS, not LPE.
* No confidentiality or integrity impact.
  The patch context shows stuck DMA teardown and controller state loss, not a demonstrated information leak or unauthorized write primitive.
* No remote network-triggerable +2.
  Thunderbolt networking is not Internet-reachable IP packet parsing in this path.

## 3. Reachability analysis

The typical trigger requires local administrative control over the Thunderbolt netdev or related device lifecycle operations. CAP_NET_ADMIN or equivalent root-level device management is normally needed to reliably bring down the interface, alter bonding state, unload or reset the path, or otherwise force repeated teardown.

A more paranoid interpretation allows influence from a connected Thunderbolt peer or physical link event, but this is adjacent or firmware-mediated rather than remotely network-triggerable. Containers and namespaces usually do not make this broadly reachable unless Thunderbolt device management or the relevant netdev is explicitly delegated into the container. The path is not a default Internet-facing attack surface and depends on Thunderbolt hardware plus affected router behavior.

## 4. Severity interpretation

This behaves more like an ordinary Moderate hardware/device DoS than an Important-class kernel vulnerability. The practical impact is persistent loss of the Thunderbolt networking peer or controller function until power cycle. The theoretical concern is DMA-adjacent lifetime ordering because buffers are unmapped and freed while paths may still be active, but the provided patch and commit message do not establish a concrete memory corruption primitive, attacker-controlled DMA target, information disclosure, or privilege escalation path.

The issue should not be treated as a strong Important candidate based only on the words DMA and freed buffers. It does deserve manual review because teardown ordering around DMA paths can hide more serious consequences on some hardware.

## 5. One-sentence report phrase

Thunderbolt networking teardown stopped rings and freed frame buffers before disabling DMA paths, which can leave pending DMA hops stuck on affected host routers and eventually take down the XDomain control channel until controller power cycle.

## 6. Manual review recommendation

MANUAL CHECK RECOMMENDED

Reason: the supported impact is DoS-only and hardware-specific, but the bug is DMA-adjacent and involves freeing or unmapping frame buffers before path deactivation, so a short manual review is justified to confirm that no platform exposes a stronger DMA memory corruption or cross-boundary impact.

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

Patch: net: thunderbolt: Tear down DMA paths before stopping the rings [ Upstream
Commit: 7cce39109206bc5497e0953806563644b88bfc44
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=7cce39109206bc5497e0953806563644b88bfc44

Changed files:
  drivers/net/thunderbolt.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=7cce39109206bc5497e0953806563644b88bfc44

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

The original Linux kernel CVE announcement for CVE-2026-74691 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-74691

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:A/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H
  The Best / paranoid CVSS score: 5.3

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: 0
  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).