[CVE-2026-99261][MODERATE REGULAR] drm/amdgpu: correct reservation fence slots for userq per-vm BOs eviction [ Upstream

[CVE-2026-99261][MODERATE REGULAR] drm/amdgpu: correct reservation fence slots for userq per-vm BOs eviction [ Upstream

[AL-KERNEL]

CVE: CVE-2026-99261
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: drm/amdgpu: correct reservation fence slots for userq per-vm BOs eviction [ Upstream
Commit: 76acfb84973bd316c9fffbdfc631f3cfb1cfa4cc
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=76acfb84973bd316c9fffbdfc631f3cfb1cfa4cc
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-99261
Analysis date: Tue, 03 Nov 2026 22:49:09 -0400
ActionableScore: 3
ActionableScore lower bound: 2
Actionable bucket: Actionable Moderate / Strong Important candidate requiring manual review
Manual review required: YES

Summary:
A local AMD GPU user-queue workload can under-reserve BO fence slots, omit an eviction fence, and expose later BO movement to inconsistent synchronization state, potentially causing a GPU hang or reset.

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

The original Linux kernel CVE announcement for CVE-2026-99261 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-99261

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-2026-99261	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-99261 CHECK WITH IMPACT FROM ORIG NN LOW | AEGIS IMPACT MODERATE AEGIS_AS=1 LEGACY=MODERATE RESULT=SAME (final reconciliation agrees; DECISION: both MODERATE)	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-667;CWE-682;*CWE-664;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:L';BEST CVSS score: '5.5';DESCR 'The reservation capacity for an amdgpu user queue per VM BO can be exhausted when TTM move fences consume the available slots and eviction requires one additional fence. Failure to publish the eviction fence can break GPU and BO movement ordering and may cause a failed eviction, GPU hang, or reset. For the CVSS the PR:L value reflects that a local process normally needs access to an AMD GPU render or device node and a user queue context. The base AC:H reflects the dependency on per VM BO eviction and fence consumption, while controlled VRAM pressure may make the paranoid AC:L interpretation defensible. The shown primitive is a reservation accounting and synchronization failure rather than a demonstrated UAF or OOB write, so confidentiality and integrity remain N. Automated impact decision path. The original SimpleNet classifier selected LOW. The primary LLM estimated paranoid CVSS 5.5 and the production ActionableScore returned 3 with recommendation Actionable Moderate / Strong Important candidate requiring manual review. The production rules then evaluated CVSS thresholds and ActionableScore thresholds. Recorded rule transitions were CVSS and initial triage stage LOW to MODERATE_KPANIC because the manual review requirement was applied. 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 INCREASED_FROM_LOW_BASED_ON_REQUIREMANUALCHECK and DECREASED_TO_MODERATEREG_BASED_ON_ACTIONABLESCORELESSTHAN4. The independent weaker Aegis shadow evaluated CVSS 0.0 and ActionableScore 1 with primitive class CLASS_C and manual review REQUIRED. Its independently reconciled impact was MODERATE. The shadow agreed with the final production impact and aggregation did not change that impact. The final reported impact is MODERATE.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 3)	MAYBE	LEAK HARDWARE  INCREASED_FROM_LOW_BASED_ON_REQUIREMANUALCHECK DECREASED_TO_MODERATEREG_BASED_ON_ACTIONABLESCORELESSTHAN4	-	-	checked

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

ActionableScore=3
ActionableScoreLower=2
PrimitiveClass=CLASS_B
RealUAF=NO
HostAdminRequired=NO
NamespaceScopedPrivilege=UNKNOWN
ImportantCandidate=YES
ManualReview=REQUIRED
AutoDowngradeAllowed=NO
RequiresHostUID0=NO
AuthorizationDecisionBypass=NO
ProtectedResourceImpact=NONE
B4PendingReview=NO
B4PendingConfirmed=NO

1. ActionableScore

* Raw base score: 2.
* Conservative score: 2; no floor or cap applied.
* Raw paranoid score: 3.
* Paranoid score: 3; no floor or cap applied.
* Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**::

2. Signal breakdown

* +1 weak/indirect active-state corruption: reserving only two fence slots can prevent publication of the additional eviction fence, leaving BO movement and eviction synchronization state inconsistent.
* +1 local below-host-admin trigger: supplied evidence grounds the path in a local AMD GPU user-queue process with render/device-node access.
* Paranoid +1 device availability: failed eviction and broken fence ordering plausibly produce a GPU hang or reset, but no host panic is shown.
* No OOB-write score: the patch does not confirm a heap overwrite, accumulated retry index, write target, or corrupted descriptor consumption. Therefore NEW52/B5-R is not established.
* No UAF, controlled reclaim, arbitrary write, confidentiality impact, protected-state integrity impact, or LPE is shown.
* No extra availability point: the possible hang/reset is one device-local consequence, not a distinct host-wide failure.

3. Reachability analysis

The supported actor is a below-host-admin local process with access to an AMD GPU node and a user-queue context. Host UID 0 is not required by the supplied evidence. The exact device-node authorization and whether the complete operation works inside an attacker-created namespace are not shown, so namespace-scoped privilege remains unknown. Eviction and fence-slot exhaustion require workload and memory-pressure conditions, but no narrow race is demonstrated and no timing penalty is applied.

4. Primitive and severity reading

The primitive is under-reservation of active reservation-fence state: TTM move fences can consume the available slots, leaving no slot for the eviction fence. Later BO movement/eviction machinery consumes that synchronization state. This is CLASS_B active structured-state inconsistency, not a demonstrated memory OOB write or UAF.

A credible delegated render-node-to-kernel/GPU trust boundary exists, while the consequences of consuming the missing-fence state remain incompletely characterized. Under the active-state trust-boundary rule, this warrants Important-candidate retention and blocks automatic downgrade, without proving host panic, arbitrary corruption, or Important severity.

5. One-sentence report phrase

A local AMD GPU user-queue workload can under-reserve BO fence slots, omit an eviction fence, and expose later BO movement to inconsistent synchronization state, potentially causing a GPU hang or reset.

6. Manual review advice

MANUAL CHECK REQUIRED

Confirm whether missing eviction ordering can expose stale/moving BO contents or unsafe GPU access beyond the supported device-local availability effect.

7. Call-site confidence

Call-site confidence: medium

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

Patch: drm/amdgpu: correct reservation fence slots for userq per-vm BOs eviction [ Upstream
Commit: 76acfb84973bd316c9fffbdfc631f3cfb1cfa4cc
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=76acfb84973bd316c9fffbdfc631f3cfb1cfa4cc

Changed files:
  drivers/gpu/drm/amd/amdgpu/amdgpu_userq.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=76acfb84973bd316c9fffbdfc631f3cfb1cfa4cc

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

The original Linux kernel CVE announcement for CVE-2026-99261 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-99261

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:H/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).