[CVE-2026-89801][MODERATE 7.0] drm/nouveau/uvmm: fix premature region free on failed OP_UNMAP_SPARSE

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

CVE: CVE-2026-89801
Priority: MODERATE 7.0
AL-KERNEL base severity: MODERATE
KPANIC flag: YES
Patch: drm/nouveau/uvmm: fix premature region free on failed OP_UNMAP_SPARSE
Commit: ba42d8a1c2c7629c61914df11a7d63ac77449ffb
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=ba42d8a1c2c7629c61914df11a7d63ac77449ffb
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89801
Analysis date: Wed, 16 Sep 2026 15:10:53 -0400
ActionableScore: 6
ActionableScore lower bound: 5
Actionable bucket: Actionable Moderate / Strong Important candidate requiring manual review
Manual review required: YES

Summary:
A use-after-free in the nouveau UVMM VM_BIND error path can prematurely free a live region and later perform another cleanup operation on the same freed region, requiring local access to the nouveau DRM device and causing at least kernel crash DoS with unresolved memory-corruption leverage.

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

The original Linux kernel CVE announcement for CVE-2026-89801 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89801

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-2026-89801	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-89801 CHECK WITH IMPACT FROM ORIG NN MODERATE	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H';*CWE-416;*CWE-664;*CWE-672;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:H';BEST CVSS score: '7';DESCR 'A use after free can occur in the nouveau UVMM VM_BIND path because OP_UNMAP_SPARSE stores a region pointer returned by nouveau_uvma_region_find without taking a reference. On selected failure paths, such as a busy or dirty region check or a drm_gpuvm_sm_unmap_ops_create failure, op reg remains set even though the submit path is unwound. Cleanup can then remove and put a region that the failing job did not create or own, dropping the sole tree reference and freeing a live region. A later cleanup or another job using the same region can dereference the freed object, causing a kernel crash and possible memory corruption. For the CVSS the PR:L is used because exploitation requires local access to the nouveau DRM device or render node, but does not inherently require full administrative privileges. The issue is not network reachable. Impact is at least local denial of service via kernel crash and in the paranoid case may allow confidentiality or integrity impact due to a kernel use after free, so manual review is needed.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 6)	YES	UAF HARDWARE LINUS KPANIC  KPANIC INCREASED_TO_HIGH_BASED_ON_GUESSCVSS DECREASED_TO_MODERATE7_BASED_ON_FALSEPOSCHECKOFDH	YES	NO	checked

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

ActionableScore=6
ActionableScoreLower=5
PrimitiveClass=CLASS_B
RealUAF=YES
HostAdminRequired=NO
NamespaceScopedPrivilege=NO
ImportantCandidate=YES
ManualReview=REQUIRED
AutoDowngradeAllowed=NO

## 1. ActionableScore

* Raw conservative score: 5
* Conservative score: 5
* Raw paranoid score: 6
* Paranoid score: 6
* Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**::

No floor or cap changed the conservative score. No floor or cap changed the paranoid score.

## 2. Signal breakdown

Conservative positive signals:

* +2 Real UAF / active lifetime violation. `op->reg` can point to a live `nouveau_uvma_region` found without taking a reference. Cleanup can remove and put a region that the failing job did not create or own, dropping the sole tree reference.
* +1 Concrete repeated release/refcount operation. The first cleanup can remove and put the live region and end its lifetime. Another job can still hold the same region pointer and later cleanup can remove and put the already freed region.
* +1 Reliable kernel crash / strong DoS. The established consequence is use-after-free on a freed region, with realistic kernel crash potential.
* +1 Local below-host-admin reachability. The vulnerable path is a nouveau DRM VM_BIND uAPI path and supplied context indicates local access to the nouveau DRM device or render node is sufficient, not full host administrative privilege.

Conservative negative signals:

* No host-admin penalty. Host-admin-equivalent privilege is not established and the supplied context supports PR:L style local device access.
* No race/timing penalty. This is an ordered error/unwind lifetime bug, not a narrow timing race.
* No hardware rarity penalty. Nouveau is hardware-specific, but common GPU driver code is not automatically rare or physical-only.

Paranoid additional signal:

* +1 Paranoid repeated-refcount leverage. The second operation is not merely a stale read. It can perform another lifetime/refcount mutation through storage whose original object lifetime may already have ended, leaving post-lifetime corruption leverage unresolved.

Material non-awarded signals:

* Controlled reclaim not demonstrated.
* Arbitrary or directed kernel write not demonstrated.
* Callback or function-pointer control not demonstrated.
* Confidentiality and integrity impact are not awarded as separate ActionableScore points merely because UAF could theoretically be exploitable.
* Structured stale-object leverage is not separately awarded because the meaningful extra fact is already counted as concrete repeated release/refcount mutation.
* Weak LPE is not separately awarded for the same repeated-release fact.
* Extra Availability is not awarded because the same crash is already counted once.
* Remote reachability is not established.

Repeated-release/refcount checks:

* First release can end lifetime: yes. `nouveau_uvma_region_put()` can drop the sole reference after `nouveau_uvma_region_remove()`.
* Exact same object reference survives: yes, another job may own the same region pointer.
* Second operation is a real lifetime/refcount mutation: yes, later cleanup can again remove and put the freed region.
* +1 Concrete repeated release/refcount awarded: yes.
* Paranoid repeated-refcount leverage +1 awarded: yes, because the stale operation itself mutates lifetime/refcount state after possible lifetime end.
* FLOOR 5 applied: not needed because raw conservative score already equals 5.

## 3. Reachability analysis

The trigger is local through the nouveau DRM VM_BIND path, specifically the `OP_UNMAP_SPARSE` arm of `nouveau_uvmm_bind_job_submit()`. The attacker needs access to the nouveau DRM device or render node and must be able to submit VM_BIND jobs that reach this failure/unwind logic.

This is not network reachable. No namespace-scoped capability chain is established. `NamespaceScopedPrivilege=NO` because the issue is a local DRM/GPU device interface issue, not a namespace-created object path.

`HostAdminRequired=NO` is used because the supplied context indicates local device or render-node access is sufficient and full host administrative privilege is not inherently required. This should still be validated against product device-node permissions, udev policy, and whether the affected nouveau VM_BIND interface is exposed to ordinary logged-in users, compositor contexts, container workloads, or only trusted graphics services.

No prior host setup beyond having the nouveau driver and affected hardware/interface available is established. No special rare hardware penalty is applied. The failure path may require conditions such as `drm_gpuvm_sm_unmap_ops_create()` returning `-ENOMEM` or a busy/dirty region case, but this is not a hard race. It is an error-path lifetime bug.

## 4. Primitive and severity interpretation

The exact primitive is premature release of a live `nouveau_uvma_region` due to stale `op->reg` ownership state on submit failure. The object is an active structured kernel object participating in GPU virtual memory management and region lifetime tracking.

This is stronger than an ordinary stale cleanup dereference. The important distinction is:

A. ordinary stale cleanup dereference: stale pointer is passed to cleanup and immediately faults.

B. second lifetime/refcount mutation through a potentially dangling object: one job can remove and put a region it does not own, potentially freeing it, while another job can later remove and put the same freed region.

This case matches B. Controlled reclaim, arbitrary R/W, callback control, or demonstrated LPE are not established, so it remains `CLASS_B`, not `CLASS_A`.

Demonstrated consequence: local kernel crash / DoS via UAF.

Speculative consequence: memory corruption beyond crash and possible privilege escalation may be possible in a worst case, but no controlled exploitation primitive is established in the supplied evidence.

`ImportantCandidate=YES` because a concrete active-object lifetime bug plus concrete repeated lifetime/refcount mutation leaves meaningful Important-vs-Moderate exploitation leverage unresolved.

`AutoDowngradeAllowed=NO` because automatically closing this as ordinary Moderate would discard a real kernel lifetime corruption primitive with a concrete second lifetime mutation through a potentially freed object.

Likely operational severity: actionable Moderate at minimum, with Strong Important review warranted if product exposure confirms ordinary local-user render-node access and further analysis finds usable reclaim or structured reuse.

## 5. One-sentence report phrase

A use-after-free in the nouveau UVMM VM_BIND error path can prematurely free a live region and later perform another cleanup operation on the same freed region, requiring local access to the nouveau DRM device and causing at least kernel crash DoS with unresolved memory-corruption leverage.

## 6. Manual review recommendation

MANUAL CHECK REQUIRED

Reason: this is a real active-object UAF with a concrete repeated release/refcount mutation, not merely a crash-only NULL dereference or passive cleanup warning.

## 7. Call-site confidence

Call-site confidence: high

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

Patch: drm/nouveau/uvmm: fix premature region free on failed OP_UNMAP_SPARSE
Commit: ba42d8a1c2c7629c61914df11a7d63ac77449ffb
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=ba42d8a1c2c7629c61914df11a7d63ac77449ffb

Commit description:

In nouveau_uvmm_bind_job_submit()'s OP_UNMAP_SPARSE arm, op->reg is set
from nouveau_uvma_region_find(), which only looks the region up and takes
no reference; a region's sole reference is its membership in
uvmm->region_mt. Two failure paths leave op->reg set: the -ENOENT check
when the region is busy, and the drm_gpuvm_sm_unmap_ops_create() failure.
The sibling nouveau_uvmm_sm_unmap_prepare() failure just below clears
op->reg; these two do not.

unwind_continue steps back one op, so the failing op is skipped by the
unwind loop and its op->reg stays set. nouveau_uvmm_bind_job_cleanup()
then enters its if (op->reg) branch and calls nouveau_uvma_region_remove()
and nouveau_uvma_region_put() on it, dropping the tree's sole reference
and freeing a region this job never created. The comment above the
cleanup loop documents the broken invariant: op->reg must be NULL on
submit failure.

This frees a live region on an unrelated failure, reachable single-job
when drm_gpuvm_sm_unmap_ops_create() returns -ENOMEM; if another job owns
the same region, its cleanup then removes and puts the freed region, a
use-after-free. Clear op->reg on both failure paths.

Fixes: b88baab ("drm/nouveau: implement new VM_BIND uAPI")
Reported-by: Yuhao Jiang <danisjiang@gmail.com>
Assisted-by: Claude:claude-opus-5
Cc: stable@vger.kernel.org
Signed-off-by: Zhenhao Wan <whi4ed0g@gmail.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Link: https://patch.msgid.link/20260811-nouveau-uvmm-vmbind-fixes-v2-2-aaee4b395d04@gmail.com
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Changed files:
  drivers/gpu/drm/nouveau/nouveau_uvmm.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=ba42d8a1c2c7629c61914df11a7d63ac77449ffb

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

The original Linux kernel CVE announcement for CVE-2026-89801 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89801

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:L/I:L/A:H
  The Best / paranoid CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H
  The Best / paranoid CVSS score: 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: 5
  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).