[CVE-2026-23034][MODERATE REGULAR] drm/amdgpu/userq: Fix fence reference leak on queue teardown v2 [ Upstream

[CVE-2026-23034][MODERATE REGULAR] drm/amdgpu/userq: Fix fence reference leak on queue teardown v2 [ Upstream

[AL-KERNEL]

CVE: CVE-2026-23034
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: drm/amdgpu/userq: Fix fence reference leak on queue teardown v2 [ Upstream
Commit: e1a30e1ab33fc522785d04bbf7e1b13a5c5c9175
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=e1a30e1ab33fc522785d04bbf7e1b13a5c5c9175
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-23034
Analysis date: Sat, 01 Aug 2026 14:01:02 -0400
ActionableScore: 3
ActionableScore lower bound: 2
Actionable bucket: Ordinary Moderate / Low-like, borderline only under sustained-abuse assumptions
Manual review required: NO

Summary:
A missing dma_fence_put() in AMDGPU user queue teardown can leak the last fence reference, allowing a local GPU user to potentially accumulate fence objects and cause resource exhaustion, but no memory corruption or privilege escalation primitive is shown.

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

The original Linux kernel CVE announcement for CVE-2026-23034 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-23034

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: NO

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

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

CVE-2026-23034	MODERATE	CHECK WITH IMPACT FROM ORIG NN LOW	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H';*CWE-772;*CWE-401;CWE-404;Other CVSS 'AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L';BEST CVSS score: '5.5';DESCR 'A reference leak in the AMDGPU user queue teardown path can keep the last dma_fence object alive because userq last_fence is not released when the queue is destroyed. A local user with access to the AMDGPU DRM user queue interface may be able to repeat queue create and destroy operations and accumulate leaked fence objects. For the CVSS the PR:L is used because reliable triggering requires local access to the GPU device interface but not necessarily full administrator privileges in systems exposing render nodes to users. The issue is not network reachable and does not provide a direct memory corruption primitive. Impact is denial of service through resource exhaustion, with no supported confidentiality or integrity impact.';NO MANUAL CHECK; ,and ActionableScore result is Ordinary Moderate / Low-like, borderline only under sustained-abuse assumptions (with actual score 3) 	MAYBE	INIT LEAK HARDWARE MEMORY  DECREASED_TO_MODERATE_REGULAR_FROM_MODERATE7_BASED_ON_GUESSCVSS_AND_CIANNH	-	-	checked

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

ActionableScore=3
ActionableScoreLower=2

## 1. ActionableScore

* Conservative score: 2
* Paranoid score: 3
* Final recommended bucket: **Ordinary Moderate / Low-like, borderline only under sustained-abuse assumptions**::

## 2. Signal breakdown

Triggered signals:

* Local unprivileged trigger: +1
  A local user with access to AMDGPU DRM user queue functionality may be able to create and tear down user queues.

* Real lifetime / refcount bug: +1
  The bug is a missing `dma_fence_put()` for `userq->last_fence`, causing a retained reference and leaked fence object.

Paranoid-only signal:

* Availability impact realistic: +1
  If repeated queue teardown can accumulate leaked `amdgpu_userq_fence` objects, this may cause resource exhaustion or make module unload fail noisily.

Not applied:

* Generic memory corruption: +0
  The patch fixes a reference leak, not UAF, double free, OOB write, type confusion, or arbitrary write.

* Privilege escalation plausible: +0
  No reclaim, stale object reuse, callback control, overwrite primitive, or refcount takeover is shown.

* Remote reachable: +0
  This is a local GPU driver interface issue, not a network-triggerable path.

* Reliable kernel crash / strong DoS: +0 conservative
  The demonstrated failure is during driver unload with remaining slab objects. That path usually requires administrative module unload, and the patch does not show an easy immediate host-wide crash from a single unprivileged action.

## 3. Reachability analysis

The likely trigger requires local access to the AMDGPU user queue interface. On systems exposing DRM render nodes to normal users, PR:L is a reasonable practical interpretation. If the affected user queue functionality is gated by device permissions, driver configuration, or unavailable hardware, exposure is reduced.

Namespaces and containers may matter only if GPU device nodes are passed through to the container. Without access to the DRM device, an unprivileged container user cannot normally trigger this path. The issue is not network reachable.

Call-site confidence: medium. The patch provides the teardown helper and the missing `dma_fence_put()` site, but not the full user queue create/destroy call path.

## 4. Severity interpretation

This behaves like an ordinary Moderate or Low-like resource lifetime bug. The realistic impact is resource leak or unload-time failure, not demonstrated memory corruption or LPE.

Theoretical impact is limited to sustained local resource exhaustion if queue creation and destruction can be repeated by an unprivileged GPU user. There is no evidence of UAF, attacker-controlled reclaim, object replacement, callback dispatch, arbitrary write, or sensitive data exposure.

## 5. One-sentence report phrase

A missing dma_fence_put() in AMDGPU user queue teardown can leak the last fence reference, allowing a local GPU user to potentially accumulate fence objects and cause resource exhaustion, but no memory corruption or privilege escalation primitive is shown.

## 6. Manual review recommendation

NO MANUAL CHECK NEEDED

This is a reference leak / resource exhaustion issue without a concrete memory corruption primitive, remote reachability, privilege escalation path, or sensitive-object access. It can be handled as non-urgent unless the target product exposes AMDGPU user queues broadly and has evidence of practical repeated resource exhaustion.

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

Patch: drm/amdgpu/userq: Fix fence reference leak on queue teardown v2 [ Upstream
Commit: e1a30e1ab33fc522785d04bbf7e1b13a5c5c9175
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=e1a30e1ab33fc522785d04bbf7e1b13a5c5c9175

Changed files:
  drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.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=e1a30e1ab33fc522785d04bbf7e1b13a5c5c9175

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

The original Linux kernel CVE announcement for CVE-2026-23034 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-23034

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:L/UI:N/S:U/C:N/I:N/A:L
  The Best / paranoid CVSS vector: AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
  The Best / paranoid CVSS score: 5.5

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: 2
  Paranoid ActionableScore: 3

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).