[CVE-2026-68276][MODERATE 7.0] drm/amdgpu/gfx: fix cleaner shader IB buffer overflow [ Upstream

[AL-KERNEL <alkernel@kernelcve.org>]

CVE: CVE-2026-68276
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: drm/amdgpu/gfx: fix cleaner shader IB buffer overflow [ Upstream
Commit: 201633f47b542a99bb7baafdfcda7781249fb3d9
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=201633f47b542a99bb7baafdfcda7781249fb3d9
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-68276
Analysis date: Mon, 10 Aug 2026 19:08:17 -0400
ActionableScore: 6
ActionableScore lower bound: 3
Actionable bucket: Actionable Moderate at minimum
Manual review required: YES

Summary:
An AMDGPU cleaner shader sysfs path allocated a 16 dword IB buffer but filled align_mask plus 1 dwords, causing an out-of-bounds kernel write and a likely local kernel crash, with manual review warranted because the bug is real driver memory corruption even though attacker control appears constrained.

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

The original Linux kernel CVE announcement for CVE-2026-68276 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-68276

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-2026-68276	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:H/I:H/A:H';*CWE-787;CWE-122;CWE-119;Other CVSS 'AV:L/AC:L/PR:H/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '7.8';DESCR 'An out of bounds write in the AMDGPU cleaner shader sysfs path occurs because a 16 dword IB buffer is allocated but the code fills align_mask plus 1 dwords. On GFX rings align_mask can be 0xff, so 256 dwords can be written into a 64 byte buffer and this can cause a kernel page fault or broader kernel memory corruption. For the CVSS the PR:L is used in the paranoid score because reliable triggering normally requires local access to the cleaner shader control path, but some deployments may expose GPU control to a service account, container root, or delegated admin user rather than full host root. The issue is not network reachable and requires local interaction with the AMDGPU sysfs control path. Impact is at least local denial of service via kernel crash and in worst case may allow confidentiality or integrity impact due to kernel memory corruption. YES REQUIRES MANUAL CHECK';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate at minimum (with actual score 5)  SKIP CVE-2026-68276 UNKNOWN SKIP No affected files built, so skip this CVE NO  - - unknown 	MAYBE	OOB SIMPLEFIX HARDWARE KPANIC MEMORY  KPANIC INCREASED_TO_HIGH_BASED_ON_GUESSCVSS KPANIC DECREASED_TO_MODERATE70_BASED_ON_ACTIONABLESCORELOWERTHAN7	-	-	checked

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

ActionableScore=6
ActionableScoreLower=3

## 1. ActionableScore

* Conservative score: 3
* Paranoid score: 6
* Final recommended bucket: **Actionable Moderate at minimum**

## 2. Signal breakdown

### Conservative score: 3

* Constrained kernel buffer overwrite: +1
  The bug writes 256 dwords into a 16 dword IB buffer, but the written value is the ring NOP opcode, not an attacker chosen payload.

* Reliable kernel crash / strong DoS: +1
  The commit explicitly states that the overflow caused a kernel page fault.

* Privileged kernel/device-management memory corruption path: +1
  The corruption occurs inside the AMDGPU job submission / GFX cleaner shader path, which is a trusted kernel driver and device-management context.

* Availability impact realistic: +1
  The failure mode is a kernel page fault, so system availability impact is realistic.

* Requires admin/root/CAP-like access in typical deployments: -2
  The trigger is a cleaner shader sysfs control path, which is normally not exposed to arbitrary unprivileged users.

* Broad/default/common subsystem exposure: +1
  AMDGPU is a broadly deployed GPU driver, although this specific cleaner shader sysfs path is narrower than normal GPU rendering interfaces.

### Paranoid score: 6

* Memory corruption, strong corruption primitive: +2
  This is an actual out-of-bounds write from kernel code into memory beyond the allocated IB buffer.

* Weak LPE concern: +1
  Privilege escalation is not demonstrated, but kernel heap/driver memory corruption makes LPE worth manual review.

* Reliable kernel crash / strong DoS: +1
  The observed result is a kernel page fault.

* Privileged kernel/device-management memory corruption path: +1
  The overflow occurs in a GPU driver job/IB construction path.

* Availability impact realistic: +1
  Crash impact is concrete and host-wide.

* Broad/default/common subsystem exposure: +1
  AMDGPU is common enough that this should not be dismissed as purely exotic.

* Corruption primitive highly constrained or no attacker-controlled payload/target: -1
  The overwrite value is a fixed NOP and the target is adjacent allocation state, not an attacker-selected address.

## 3. Reachability analysis

The vulnerable path is local and is reached through the AMDGPU cleaner shader sysfs interface. In typical deployments this is expected to require root or a privileged service managing the GPU, so the conservative interpretation treats it as privileged local access.

A more paranoid interpretation is justified when GPU control sysfs access is delegated to a service account, container root, reduced-capability root, or an operational automation process that is not equivalent to full host root. This does not make the bug remotely reachable, but it can reduce the practical privilege barrier below full administrative control.

The path is not network-triggerable. It requires an affected AMDGPU device and the cleaner shader interface introduced by the referenced commit. Exploitation evidence supports kernel crash. The OOB write primitive is real, but payload and target control appear limited from the supplied patch.

Call-site confidence: high. The affected allocation, fill loop, IB length assignment, and job submission call are all visible in the provided diff.

## 4. Severity interpretation

This is stronger than an ordinary Moderate issue because the patch fixes a real kernel out-of-bounds write, not only a validation warning or resource leak. The realistic demonstrated impact is local denial of service through a kernel page fault.

The Important-class argument is weaker because the overwrite payload is fixed to the ring NOP value and no controlled target, reclaim primitive, callback overwrite, or object replacement is shown. Therefore the conservative score stays low-to-borderline, while the paranoid score reaches Actionable Moderate due to actual kernel memory corruption.

## 5. One-sentence report phrase

An AMDGPU cleaner shader sysfs path allocated a 16 dword IB buffer but filled align_mask plus 1 dwords, causing an out-of-bounds kernel write and a likely local kernel crash, with manual review warranted because the bug is real driver memory corruption even though attacker control appears constrained.

## 6. Manual review recommendation

MANUAL CHECK REQUIRED

Reason: the bug is an OOB write in kernel driver memory. Even though the observed impact is a kernel page fault and the overwrite appears constrained to fixed NOP values, memory corruption in a GPU job/IB path should not be auto-closed without checking sysfs permissions, allocator adjacency, affected device generations, and whether delegated GPU management access can make the trigger PR:L in practice.

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

Patch: drm/amdgpu/gfx: fix cleaner shader IB buffer overflow [ Upstream
Commit: 201633f47b542a99bb7baafdfcda7781249fb3d9
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=201633f47b542a99bb7baafdfcda7781249fb3d9

Changed files:
  drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.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=201633f47b542a99bb7baafdfcda7781249fb3d9

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

The original Linux kernel CVE announcement for CVE-2026-68276 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-68276

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:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H
  The Best / paranoid CVSS score: 7.8

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: 3
  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).