[CVE-2026-99150][MODERATE REGULAR] drm/amdgpu: don't drop the runtime PM reference twice in amdgpu_gpr

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

CVE: CVE-2026-99150
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: drm/amdgpu: don't drop the runtime PM reference twice in amdgpu_gpr
Commit: 652b1f26125fe48aed5f4944698a85b457eef491
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=652b1f26125fe48aed5f4944698a85b457eef491
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-99150
Analysis date: Wed, 28 Oct 2026 22:50:34 -0400
ActionableScore: 3
ActionableScore lower bound: unknown
Actionable bucket: Actionable Moderate / Important-candidate manual review
Manual review required: YES

Summary:
A failed GPR debugfs userspace copy performs a second runtime-PM put, corrupting the GPU usage count and potentially suspending the device while another workload still holds a reference.

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

The original Linux kernel CVE announcement for CVE-2026-99150 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-99150

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-2026-99150	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-99150 CHECK WITH IMPACT FROM ORIG NN LOW | AEGIS IMPACT LOW AEGIS_AS=2 LEGACY=MODERATE RESULT=DIFF (final Legacy and Aegis differ; FINAL: MODERATE, Aegis: 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-911;CWE-675;Other CVSS 'AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L';BEST CVSS score: '5.5';DESCR 'amdgpu_debugfs_gpr_read released its runtime PM reference after reading the wave registers and released it again when put_user failed on an invalid userspace buffer. This can steal a runtime PM reference held by another GPU user or underflow the usage counter, potentially causing premature GPU suspend and disruption of graphics or compute workloads. For the CVSS the PR:L rating reflects that reliable triggering requires local permission to read the amdgpu GPR debugfs file, which is normally restricted but may be delegated to a diagnostic process or a reduced capability administrator. The demonstrated primitive is runtime PM reference count corruption rather than memory corruption. There is no supported UAF, OOB access, information leak, arbitrary write, or privilege escalation path. 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 / Important-candidate 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 CVSS score, vector or bug class threshold rule 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_HIGHCVSSSCORE and DECREASED_TO_MODERATEREG_BASED_ON_ACTIONABLESCORELESSTHAN4. The independent weaker Aegis shadow evaluated CVSS 3.3 and ActionableScore 2 with primitive class CLASS_B and manual review REQUIRED. Its independently reconciled impact was LOW. The models disagreed but the aggregation rules retained the production result. The final reported impact is MODERATE.';NO MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Important-candidate manual review (with actual score 3)	MAYBE	OOB SIMPLEFIX LEAK ERRORPATH HARDWARE DEBUGFS  INCREASED_FROM_LOW_BASED_ON_HIGHCVSSSCORE DECREASED_TO_MODERATEREG_BASED_ON_ACTIONABLESCORELESSTHAN4	-	-	checked

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

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

1. ActionableScore

* Raw base score: 3.
* Conservative score: 3; no floor or cap applied.
* Raw paranoid score: 3.
* Paranoid score: 3; no additional exploit effect is technically established.
* Final recommended bucket: **Actionable Moderate / Important-candidate manual review**::

2. Signal breakdown

* +1 — Concrete repeated resource/refcount mutation: after the normal `pm_runtime_put_autosuspend()`, a failed userspace write reaches `err` and performs the same PM put again. The first put may end the active power hold; the second changes the runtime PM count again.
* +1 — Integrity impact: the commit explicitly shows runtime usage-count underflow or theft of a reference held by another GPU user.
* +1 — Availability impact: the incorrect zero count can permit premature GPU autosuspend and disrupt graphics or compute workloads.
* No negative privilege, race, or rare-hardware penalty: exact debugfs permissions are not supplied, there is no timing race, and Navi31 is hardware-specific but not proven rare.
* No UAF, heap OOB, arbitrary write, controlled reclaim, kernel panic, disclosure, LPE, or authorization bypass is shown.
* The repeated-reference fact is counted once; no separate weak-corruption or crash point is stacked.

The first put can end the runtime-active resource state, and the exact device reference reaches a second real PM-count mutation. This does not free the `amdgpu_device`, so `RealUAF=NO`, the UAF floor does not apply, and no paranoid repeated-UAF point is awarded.

3. Reachability analysis

A process able to read the GPR debugfs file can supply an invalid userspace destination so that `put_user()` fails after the hardware read. Whether such access requires host administration, host UID 0, or delegated diagnostic permissions is not established; therefore both effective host-admin and UID requirements remain unknown. No namespace-scoped capability chain is shown. The trigger is ordered rather than raced. AMDGPU/Navi31 and enabled debugfs access are prerequisites, but no separate unrelated vulnerability is needed.

4. Primitive and severity reading

The primitive is active runtime-PM reference-state corruption, not memory corruption: a second PM put can underflow the count or consume another holder’s reference. The privileged runtime-PM machinery later consumes that corrupted active state and may autosuspend the GPU.

The credible but unresolved boundary is between a potentially delegated debugfs accessor and other GPU users whose PM hold and workloads can be affected. Under the active-state/trust-boundary rule, this specific write, later privileged consumer, and unresolved delegated-interface boundary require Important-candidate review. The shown effect is GPU state/workload disruption; host panic, memory corruption, protected-data modification, and privilege escalation remain unproven.

5. One-sentence report phrase

A failed GPR debugfs userspace copy performs a second runtime-PM put, corrupting the GPU usage count and potentially suspending the device while another workload still holds a reference.

6. Manual review advice

MANUAL CHECK REQUIRED

Review debugfs delegation and cross-user GPU impact because active PM state is corrupted and later trusted by kernel power-management code.

7. Call-site confidence

Call-site confidence: high

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

Patch: drm/amdgpu: don't drop the runtime PM reference twice in amdgpu_gpr
Commit: 652b1f26125fe48aed5f4944698a85b457eef491
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=652b1f26125fe48aed5f4944698a85b457eef491

Commit description:

amdgpu_debugfs_gpr_read() drops its runtime PM reference right after
reading the wave registers. When copying them to userspace fails, it
jumps to the err label, which drops the reference a second time. A read
into a bad buffer on Navi31 either steals a reference somebody else
holds (power/runtime_usage goes from 1 to 0) or underflows the count:

  amdgpu 0000:03:00.0: Runtime PM usage count underflow!

The registers are already in the kernel buffer at that point, so finish
the hardware access first and copy the buffer with a single
copy_to_user() instead of the put_user() loop with its own error path.

Fixes: 9eee152 ("drm/amdgpu/debugfs: fix ref count leak when pm_runtime_get_sync fails")
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Jesse Zhang <Jesse.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 31035bbd74f7c3e10c163db875b8e67f8b50ee6a)
Cc: stable@vger.kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Changed files:
  drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.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=652b1f26125fe48aed5f4944698a85b457eef491

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

The original Linux kernel CVE announcement for CVE-2026-99150 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-99150

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: unknown
  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).