[CVE-2026-89814][MODERATE REGULAR] drm/amdgpu: clamp the isolation index for rings outside a partition

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

CVE: CVE-2026-89814
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: drm/amdgpu: clamp the isolation index for rings outside a partition
Commit: 964de255497ffd7cb8a86e405b8ac6d927e7e177
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=964de255497ffd7cb8a86e405b8ac6d927e7e177
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89814
Analysis date: Wed, 16 Sep 2026 08:29:54 -0400
ActionableScore: 3
ActionableScore lower bound: 2
Actionable bucket: Borderline / manual review recommended
Manual review required: YES

Summary:
No one-sentence report phrase was found.

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

The original Linux kernel CVE announcement for CVE-2026-89814 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89814

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-2026-89814	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-89814 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:H/I:H/A:H';CWE-129;CWE-125;*CWE-787;Other CVSS 'AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '7';DESCR 'An invalid isolation array index can occur when an AMDGPU ring is not assigned to an XCP partition and ring xcp_id retains AMDGPU_XCP_NO_PARTITION. This value is approximately zero inverted and was used directly to index adev isolation during SDMA submission and VM flush processing, producing an out of bounds pointer and triggering UBSAN. In the isolation enforcement path the pointer was formed before the ring type check. In the VM flush path the invalid pointer can reach later isolation state accesses, so the impact may extend beyond a diagnostic warning to a kernel crash or memory corruption. For the CVSS the PR:L value is selected because a local unprivileged process may submit GPU workloads when it has access to the AMDGPU render node. The issue is not directly network reachable. The base impact is denial of service. The paranoid impact uses C:H I:H A:H because an out of bounds kernel pointer used by later VM flush logic may provide a memory corruption path, although reliable privilege escalation is not demonstrated and requires manual source review.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Borderline / manual review recommended (with actual score 3)  SKIP CVE-2026-89814 UNKNOWN SKIP No affected files built, so skip this CVE NO  - - unknown 	MAYBE	KASAN OOB SIMPLEFIX HARDWARE MEMORY  DECREASED_TO_MODERATEREG_BASED_ON_ACTIONABLESCORELESSTHAN4	-	-	checked

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

ActionableScore=3
ActionableScoreLower=2
PrimitiveClass=CLASS_C
RealUAF=NO
HostAdminRequired=NO
NamespaceScopedPrivilege=NO
ImportantCandidate=NO
ManualReview=RECOMMENDED
AutoDowngradeAllowed=YES

1. ActionableScore

* Raw conservative score: 2
* Conservative score: 2
* Raw paranoid score: 3
* Paranoid score: 3
* No floor or cap was applied.
* Final recommended bucket: **Borderline / manual review recommended**::

2. Signal breakdown

Conservative signals:

* +1 Local unprivileged trigger. A process with access to the AMDGPU render node can normally submit SDMA workloads without host-admin-equivalent privilege.
* +1 Weak or indirect corruption candidate. `AMDGPU_XCP_NO_PARTITION`, represented by an all-ones sentinel, is used as an unchecked index and creates an out-of-bounds `struct amdgpu_isolation` pointer.
* +1 Broad and realistically exercised exposure. AMDGPU is broadly deployed, and SDMA submission is an ordinary GPU data-plane operation on affected hardware.
* -1 No attacker-controlled payload or target. The invalid index is a fixed sentinel associated with the selected ring. The supplied evidence does not show that the attacker can direct the pointer toward a chosen kernel object.
* Total: 1 + 1 + 1 - 1 = 2.

Paranoid-only signal:

* +1 Plausible strong DoS. The VM-flush path creates the invalid pointer without the early ring-type return present in the isolation-enforcement path. A later dereference can plausibly cause a kernel fault, although the supplied context directly demonstrates only an out-of-bounds UBSAN report.
* Total: 2 + 1 = 3.

Important signals not awarded:

* No strong OOB write was established.
* No attacker-selected destination or data was established.
* No disclosure path was established from an OOB read.
* No controlled replacement, reclaim, type confusion, callback influence, or arbitrary read/write was demonstrated.
* No UAF, repeated release, refcount mutation, or lifetime violation exists.
* No confidentiality or integrity points were awarded merely because the previous CVSS analysis used `C:H/I:H`.
* No extra availability point was awarded for the same potential kernel fault.
* Remote reachability was not established.
* Hardware rarity penalty was not applied because AMD GPUs are not inherently rare, although only affected AMDGPU devices are exposed.

3. Reachability analysis

An ordinary local process can potentially reach the path by opening an accessible AMDGPU render node and submitting work that uses an SDMA ring outside an XCP partition. DRM master, KMS control, compositor status, and host-administrator privilege are not normally required for render-node command submission.

The complete trigger remains dependent on an affected AMD GPU, the AMDGPU driver, a ring retaining `AMDGPU_XCP_NO_PARTITION`, and the relevant SDMA or VM-flush path. These are device and driver prerequisites, not evidence of physical-only or rare hardware.

No capability namespace governs ordinary render-node access. Access is based on the device node and its file permissions or an already supplied DRM file descriptor. Therefore namespace-scoped privilege is not part of the demonstrated trigger.

There is no race or narrow timing requirement. The invalid index follows directly from the persistent sentinel value. No direct network trigger is present.

4. Primitive and severity interpretation

The exact primitive is creation of an out-of-bounds pointer by indexing `adev->isolation[]` with the fixed `AMDGPU_XCP_NO_PARTITION` sentinel.

In `amdgpu_device_enforce_isolation()`, the old code formed the invalid pointer before checking whether the ring type made isolation relevant. For an SDMA ring, the function can return without consuming the pointer, so this instance can be limited to undefined pointer arithmetic and an UBSAN report.

In `amdgpu_vm_flush()`, the pointer is created on the active VM-flush path. Later use of `isolation` is plausible from normal function semantics, but the supplied context does not show the subsequent accesses or prove whether they are reads, writes, or only comparisons. Consequently, a concrete exploitable OOB write cannot be assigned.

This is not stale storage, UAF, or a lifetime-management defect. It is an invalid-index and invalid-pointer issue whose established symptom is UBSAN detection and whose plausible practical consequence is a kernel fault or GPU workload failure. It is therefore CLASS_C rather than CLASS_A or CLASS_B.

ImportantCandidate is NO because no independently useful corruption primitive beyond the possible crash has been established. AutoDowngradeAllowed is YES because there is no unresolved controlled corruption, ownership violation, lifetime primitive, or security-boundary bypass. Manual review remains recommended to inspect every later `isolation` dereference in `amdgpu_vm_flush()` and determine whether any write occurs before treating the issue as conclusively crash-only.

The likely operational severity is ordinary Moderate or Low-like, with a borderline score retained because the vulnerable path is locally reachable below host-admin privilege and executes in the kernel.

5. One-sentence report phrase

A local process with AMDGPU render-node access can submit SDMA work that uses the no-partition sentinel as an out-of-bounds isolation-array index, triggering UBSAN and potentially causing a kernel fault, but no controlled write or privilege-escalation primitive is established.

6. Manual review recommendation

MANUAL CHECK RECOMMENDED

Review the complete `amdgpu_vm_flush()` function and all accesses through `isolation` to determine whether the invalid pointer can be dereferenced for a write or only produces a warning or fault.

7. Call-site confidence

Call-site confidence: medium

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

Patch: drm/amdgpu: clamp the isolation index for rings outside a partition
Commit: 964de255497ffd7cb8a86e405b8ac6d927e7e177
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=964de255497ffd7cb8a86e405b8ac6d927e7e177

Commit description:

adev->isolation[] has one slot per partition, but a ring that is not
assigned to one keeps AMDGPU_XCP_NO_PARTITION, which is ~0, so indexing
the array with it is out of bounds. SDMA submissions hit this on both
the isolation enforcement and the VM flush path and trip UBSAN.

Fall back to the first slot the way the cleaner shader path already
does, and stop taking the address before the ring type check that makes
it relevant.

Cc: stable@vger.kernel.org
Signed-off-by: Xiang Liu <xiang.liu@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Changed files:
  drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
  drivers/gpu/drm/amd/amdgpu/amdgpu_vm.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=964de255497ffd7cb8a86e405b8ac6d927e7e177

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

The original Linux kernel CVE announcement for CVE-2026-89814 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89814

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: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: 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).