[CVE-2026-68101][MODERATE REGULAR] drm/amdgpu: fix check in amdgpu_hmm_invalidate_gfx For a short moment during alloc/free the userptr BO is not part of his VM, so bo->vm_bo can be NULL. Keep a reference to the VM root PD as parent of the userptr BO so that we can always use that to wait for all submissions of the VM instead of only the one involving the userptr BO. Signed-off-by: Christian König <christian.koenig@amd.com> Fixes: 9125089 ("drm/amdgpu: fix waiting for all submissions for userptrs") Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/5399 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from

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

CVE: CVE-2026-68101
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: drm/amdgpu: fix check in amdgpu_hmm_invalidate_gfx For a short moment during alloc/free the userptr BO is not part of his VM, so bo->vm_bo can be NULL. Keep a reference to the VM root PD as parent of the userptr BO so that we can always use that to wait for all submissions of the VM instead of only the one involving the userptr BO. Signed-off-by: Christian König <christian.koenig@amd.com> Fixes: 9125089 ("drm/amdgpu: fix waiting for all submissions for userptrs") Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/5399 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from
Commit: 52f650963d8825e97a0ccdd2b616f8a01d9d3d38
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=52f650963d8825e97a0ccdd2b616f8a01d9d3d38
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-68101
Analysis date: Mon, 10 Aug 2026 15:02:11 -0400
ActionableScore: 3
ActionableScore lower bound: 2
Actionable bucket: Borderline Moderate. Manual review recommended
Manual review required: YES

Summary:
A local process with access to the AMDGPU DRM device can race userptr BO alloc/free with HMM invalidation so that amdgpu_hmm_invalidate_gfx dereferences a NULL bo vm_bo pointer, causing a kernel crash or GPU driver denial of service.

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

The original Linux kernel CVE announcement for CVE-2026-68101 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-68101

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-2026-68101	MODERATE	CHECK WITH IMPACT FROM ORIG NN LOW	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:H';**CWE-476;CWE-362;CWE-703;BEST CVSS score: '4.7';DESCR 'A NULL pointer dereference can occur in the AMDGPU HMM invalidate path because amdgpu_hmm_invalidate_gfx may dereference bo vm_bo while a userptr BO is briefly not attached to its VM during allocation or free. A local process with access to the AMDGPU DRM device can potentially trigger the race through userptr memory activity and MMU notifier invalidation, causing a kernel crash or GPU driver denial of service. For the CVSS the PR:L is used because reliable triggering requires local code execution with access to the DRM render or GPU device rather than only remote network reachability. The issue is not network reachable. Impact is denial of service only based on the visible primitive, because the patch supports a NULL dereference race but does not show UAF, OOB write, arbitrary write, or information disclosure.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Borderline Moderate. Manual review recommended (with actual score 3)  SKIP CVE-2026-68101 UNKNOWN SKIP No affected files built, so skip this CVE NO  - - unknown 	MAYBE	SIMPLEFIX 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: **Borderline Moderate. Manual review recommended**

## 2. Signal breakdown

Conservative signals:

* Local unprivileged trigger: +1. A local process with access to the AMDGPU DRM render or GPU device can create userptr BOs and cause MMU notifier activity.
* Reliable kernel crash / strong DoS: +1. The patch prevents dereferencing bo->vm_bo when it can be NULL, which is a plausible kernel NULL dereference crash.
* Broad/common affected deployment on matching hardware: +1. AMDGPU render nodes are commonly exposed to local users on systems with AMD GPUs.
* Hard or unreliable race / special timing required: -1. The vulnerable window exists only briefly during userptr BO alloc/free and must coincide with HMM/MMU invalidation.

Paranoid additional signal:

* GPU object lifetime / MMU notifier path needs manual review: +1. The path involves userptr BO lifetime, VM association state, and asynchronous invalidation callbacks, so it deserves review even though the visible primitive is only NULL dereference.

Not counted:

* No generic strong memory corruption. The patch shows NULL dereference, not UAF, OOB write, double free, arbitrary write, type confusion, or attacker-controlled reuse.
* No privilege escalation bonus. LPE is not supported by the visible primitive.
* No confidentiality or integrity impact. There is no shown read-back, info leak, write primitive, or policy bypass.
* No remote/network reachability.

Call-site confidence: high. The patch includes both the userptr ioctl setup path and the HMM invalidate callback where the bad dereference occurs.

## 3. Reachability analysis

The likely trigger is local. An attacker needs local code execution and access to the AMDGPU DRM render node or GPU device. This is often available to desktop users or users in render/video-related groups, but it is not a network-exposed path.

Namespaces and containers may affect reachability only if the GPU device node is passed through or exposed to the container. Without device access, the bug is not reachable. The affected path is hardware and driver dependent, requiring AMDGPU plus userptr/HMM behavior, so it is not universal across all Linux systems.

Realistic exploitation conditions require racing userptr BO allocation or free against MMU notifier invalidation. That makes the bug less reliable than a direct ioctl crash.

## 4. Severity interpretation

This behaves like an ordinary to borderline Moderate kernel DoS issue, not an Important-class vulnerability. The theoretical concern comes from a race in GPU memory management, but the demonstrated primitive is a NULL pointer dereference. There is no evidence of stale object reuse, UAF reclaim, writable stale object access, callback control, refcount takeover, or arbitrary memory corruption.

The paranoid score is raised only enough to prevent auto-close without human review, because GPU userptr and HMM invalidation paths are complex and locally reachable on affected systems.

## 5. One-sentence report phrase

A local process with access to the AMDGPU DRM device can race userptr BO alloc/free with HMM invalidation so that amdgpu_hmm_invalidate_gfx dereferences a NULL bo vm_bo pointer, causing a kernel crash or GPU driver denial of service.

## 6. Manual review recommendation

MANUAL CHECK RECOMMENDED

Manual review is recommended because this is a local race in a GPU memory-management lifetime path, but current evidence supports DoS only and does not justify Important-class handling.

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

Patch: drm/amdgpu: fix check in amdgpu_hmm_invalidate_gfx For a short moment during alloc/free the userptr BO is not part of his VM, so bo->vm_bo can be NULL. Keep a reference to the VM root PD as parent of the userptr BO so that we can always use that to wait for all submissions of the VM instead of only the one involving the userptr BO. Signed-off-by: Christian König <christian.koenig@amd.com> Fixes: 9125089 ("drm/amdgpu: fix waiting for all submissions for userptrs") Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/5399 Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from
Commit: 52f650963d8825e97a0ccdd2b616f8a01d9d3d38
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=52f650963d8825e97a0ccdd2b616f8a01d9d3d38

Changed files:
  drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
  drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.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=52f650963d8825e97a0ccdd2b616f8a01d9d3d38

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

The original Linux kernel CVE announcement for CVE-2026-68101 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-68101

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: Not available
  The Best / paranoid CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:H
  The Best / paranoid CVSS score: 4.7

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).