[CVE-2026-89808][LOW] drm/amdkfd: Fix the case that vm range is hole at svm_migrate_copy_to_vram

[CVE-2026-89808][LOW] drm/amdkfd: Fix the case that vm range is hole at svm_migrate_copy_to_vram

[AL-KERNEL]

CVE: CVE-2026-89808
Priority: LOW
AL-KERNEL base severity: LOW
KPANIC flag: NO
Patch: drm/amdkfd: Fix the case that vm range is hole at svm_migrate_copy_to_vram
Commit: ae806a95b28fcecb913430cfa45a252e91a945d6
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=ae806a95b28fcecb913430cfa45a252e91a945d6
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89808
Analysis date: Wed, 16 Sep 2026 14:05:40 -0400
ActionableScore: 1
ActionableScore lower bound: unknown
Actionable bucket: Ordinary Moderate
Manual review required: NO

Summary:
No one-sentence report phrase was found.

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

The original Linux kernel CVE announcement for CVE-2026-89808 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89808

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: LOW
Manual review required: NO

A detailed explanation of the methodology and priority rules is included
at the end of this message.

======================================================================
AL-KERNEL CLASSIFICATION RESULT
======================================================================

CVE-2026-89808	LOW	https://www.kernelcve.org/list/?q=CVE-2026-89808 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:L/A:L';*CWE-457;CWE-703;CWE-754;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:L';BEST CVSS score: '3.6';DESCR 'A logic error in AMD KFD SVM migration can process a CPU side hole with no valid pages and use an uninitialized status value after a zero length copy operation. This can incorrectly treat the migration as failed and release VRAM pages prepared for the current operation, causing migration failure or disruption of an affected GPU workload. For the CVSS the PR:L value is used because an attacker needs local access to AMD KFD and must run a workload that exercises SVM migration with the required hole condition. The issue is not network reachable. The available code does not demonstrate a use after free, out of bounds access, arbitrary write, information disclosure, or privilege escalation primitive. Impact is primarily local availability loss for the affected GPU workload. The paranoid assessment includes I:L only for possible loss or corruption of application level GPU computation results.';NO MANUAL CHECK; ,and ActionableScore result is Ordinary Moderate (with actual score 1)  SKIP CVE-2026-89808 UNKNOWN SKIP No affected files built, so skip this CVE NO  - - unknown 	MAYBE	SIMPLEFIX DMAorINTERRUPT HARDWARE MEMORY  DECREASED_TO_MODERATE_REGULAR_FROM_MODERATE7_BASED_ON_GUESSCVSS DECREASED_TO_LOW_FROM_MODERATE_BASED_ON_ACTIONABLESCORELESSTHAN3	-	-	checked

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

ActionableScore=1
PrimitiveClass=NONE
RealUAF=NO
HostAdminRequired=NO
NamespaceScopedPrivilege=NO
ImportantCandidate=NO
ManualReview=NO
AutoDowngradeAllowed=YES

1. ActionableScore

* Raw conservative score: 1
* Conservative score: 1
* Raw paranoid score: 1
* Paranoid score: 1
* Final recommended bucket: **Ordinary Moderate**::

The conservative and paranoid scores are identical because the patch establishes a migration failure and erroneous cleanup outcome, but does not establish a memory-corruption, ownership-violation, or security-boundary primitive.

2. Signal breakdown

* +1 Availability impact realistic. A local AMD KFD workload can encounter a CPU-side migration hole, causing the operation to be treated as failed and newly prepared VRAM pages to be released. This can disrupt the affected GPU compute workload.
* +0 Uninitialized status variable. The uninitialized value is used as an error-status result and selects the existing error-cleanup path. The supplied code does not show that it is used as an address, size, index, callback, DMA descriptor, or other corruption-relevant value.
* +0 No host-wide kernel crash is demonstrated. The stated result is release of VRAM pages prepared for the current migration operation.
* +0 No persistent resource leak is established. The relevant behavior is an erroneous cleanup, not a missing release.
* +0 No concrete repeated release or refcount operation is established. The patch does not show the same freed object being released again through a surviving dangling reference.
* +0 Controlled reclaim not demonstrated.
* +0 Arbitrary or directed kernel write not demonstrated.
* +0 Information disclosure not demonstrated.
* +0 Integrity impact is not awarded. Loss or interruption of a users GPU computation is not evidence of unauthorized modification of protected kernel state or another principals data.
* +0 Remote reachability not established. The issue is in a local GPU SVM migration path.
* +0 Strong device-management corruption is not established. Although the code is in AMD KFD and handles VRAM and DMA mappings, the supplied evidence does not show invalid DMA submission, invalid DMA unmap execution, page-table corruption, or stale ownership after the cleanup path.

3. Reachability analysis

The evaluated attacker is a local process using AMD KFD GPU-compute functionality and exercising SVM migration. The trigger requires a specific migration state where a CPU-side range is a hole, with MIGRATE_PFN_MIGRATE set and MIGRATE_PFN_VALID unset.

Host-admin-equivalent privilege is not required by the vulnerable operation itself. Access to AMD KFD is normally controlled through the relevant device-node permissions and is therefore a local low-privilege or delegated-device-access scenario rather than a host-administration operation. The supplied evidence supports HostAdminRequired=NO, but does not establish that every ordinary local account can access the affected KFD device.

The complete trigger is not namespace-scoped. It depends on a host GPU, the AMDGPU and AMDKFD driver stack, and an SVM migration path. A user namespace or network namespace does not independently provide the required GPU device access or establish the complete vulnerable operation.

No race is established. This is an ordered invalid-state path caused by a CPU-side migration hole and a zero-length final copy attempt, not a timing-dependent lifetime race. Common AMD GPU hardware may be involved, but no rare-hardware penalty is applied because the problem is not physical-access-only and AMD GPU compute systems are a normal deployment class.

4. Primitive and severity interpretation

The exact primitive is an uninitialized integer status value in svm_migrate_copy_memory_gart and svm_migrate_copy_to_vram when a migration range contains no valid source pages for the relevant copy segment. The uninitialized status can be interpreted as an error and cause the existing out_free_vram_pages cleanup path to release VRAM pages allocated for the current migration attempt.

This is neither active structured stale-object use nor fixed stale storage. No object is shown to be freed while retained references remain. No post-free dereference, second put, refcount mutation, callback dispatch, stale DMA descriptor use, or reused-object interpretation is established.

The cleanup path is relevant, but it is ordinary error cleanup of pages created for the migration operation. The supplied patch does not establish that those pages have already become active mappings, remain reachable through another owner, or are accessed after being released. The added src entry clearing on DMA-map failure is defensive cleanup correctness and does not independently demonstrate an invalid DMA-unmap or memory-corruption primitive.

PrimitiveClass=NONE is appropriate because the demonstrated impact is migration failure or workload disruption without a meaningful corruption, lifetime, authorization, or ownership primitive. The likely operational severity is low and limited to affected GPU SVM workloads. ImportantCandidate=NO and AutoDowngradeAllowed=YES are appropriate because no unresolved Important-grade leverage remains after excluding speculative outcomes.

5. One-sentence report phrase

An AMD KFD SVM migration path can use an uninitialized status value for CPU-side memory holes and incorrectly release VRAM pages prepared for the current migration, causing failure of an affected local GPU workload.

6. Manual review recommendation

NO MANUAL CHECK NEEDED

The supplied patch supports a local GPU-workload availability issue only and does not establish UAF, double free, invalid DMA execution, persistent leak, page ownership corruption, privilege escalation, or a security-boundary bypass.

7. Call-site confidence

Call-site confidence: medium

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

Patch: drm/amdkfd: Fix the case that vm range is hole at svm_migrate_copy_to_vram
Commit: ae806a95b28fcecb913430cfa45a252e91a945d6
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=ae806a95b28fcecb913430cfa45a252e91a945d6

Commit description:

When migration vm range is hole at cpu side(MIGRATE_PFN_MIGRATE set +
MIGRATE_PFN_VALID unset) driver still allocates device pages. There is no
dma map of src pages and migration. j is 0 and svm_migrate_copy_memory_gart()
will return an uninitialized r. That can trigger out_free_vram_pages to drop
all VRAM just set up.

Initialize r and only call the last svm_migrate_copy_memory_gart if j > 0.

Current code postponed the last page to the final copy. This patch flushes on
the last page when reach to the end of current drm_buddy_block; avoids another
svm_migrate_copy_memory_gart.

Cc: stable@vger.kernel.org
Signed-off-by: Xiaogang Chen <xiaogang.chen@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@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/amdkfd/kfd_migrate.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=ae806a95b28fcecb913430cfa45a252e91a945d6

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

The original Linux kernel CVE announcement for CVE-2026-89808 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89808

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:N/I:N/A:L
  The Best / paranoid CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L
  The Best / paranoid CVSS score: 3.6

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: 1

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: LOW
KPANIC detected for this report: NO
Published priority for this report (same as in Subject): LOW

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).