* [CVE-2026-97498][MODERATE REGULAR] drm/amdgpu/userq: pin mqd and fw object bo to avoid eviction [ Upstream
@ 2026-09-24 22:35 21% AL-KERNEL
0 siblings, 0 replies; 1+ results
From: AL-KERNEL @ 2026-09-24 22:35 UTC (permalink / raw)
To: kernel-cve
CVE: CVE-2026-97498
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: drm/amdgpu/userq: pin mqd and fw object bo to avoid eviction [ Upstream
Commit: a16d9986b7b27b08a696524a4dbf03dbf5433a16
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=a16d9986b7b27b08a696524a4dbf03dbf5433a16
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-97498
Analysis date: Thu, 24 Sep 2026 18:35:50 -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-2026-97498 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-97498
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-2026-97498 MODERATE https://www.kernelcve.org/list/?q=CVE-2026-97498 CHECK WITH IMPACT FROM ORIG NN 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-664;*CWE-672;**CWE-400;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '5.5';DESCR 'AMDGPU user queue MQD and firmware buffer objects can remain referenced by GPU hardware without being pinned in GTT. Memory pressure can evict or move these objects while the hardware continues to use their previous GPU addresses, resulting in GPU page faults and a system hang. For the CVSS the PR:L because a local process with access to an AMDGPU render node or user queue interface may create queues without full administrative privileges. The issue is not directly network reachable. The demonstrated impact is local denial of service with high availability impact. No concrete CPU use after free, information disclosure, out of bounds write, or arbitrary write primitive is established by the supplied code. The paranoid score lowers attack complexity because controlled TTM memory pressure may make eviction repeatable, but it does not change the issue into a memory corruption vulnerability.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Borderline / manual review recommended (with actual score 3) SKIP CVE-2026-97498 UNKNOWN SKIP No affected files built, so skip this CVE NO - - unknown MAYBE HARDWARE INCREASED_FROM_LOW_BASED_ON_REQUIREMANUALCHECK 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 score 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 and user queue interface can create the relevant queue without host-admin-equivalent privilege.
* +1 Reliable strong DoS. The commit directly reports GPU page faults followed by a system hang when the active queue objects are evicted.
* Total: 2.
Additional paranoid signal:
* +1 Local unprivileged high-control kernel data-plane API. The attacker can create and keep user queues active while generating GPU and TTM memory pressure, making eviction of the unpinned queue objects plausibly repeatable.
* Total: 3.
No negative signals were applied:
* No hardware rarity penalty. AMDGPU hardware is device-specific but is not inherently rare.
* No race penalty. The defect is a residency failure under eviction, not a demonstrated narrow or externally uncontrollable race.
* No host-admin penalty. The supplied context supports access through a low-privilege render-node user.
* No uncommon-subsystem penalty. User queues may depend on supported hardware and driver configuration, but the evidence does not establish a genuinely rare or inaccessible configuration.
Material non-awarded signals:
* No UAF signal. The BO remains allocated, but its required GTT residency is lost while hardware retains its GPU address.
* No strong or constrained overwrite signal. The evidence establishes invalid-address access, page faults, and a hang, not a successful DMA write into another object.
* No controlled reclaim, arbitrary write, information disclosure, callback influence, or privilege escalation is demonstrated.
* No additional Availability point was awarded because the page faults and system hang describe the same DoS consequence.
* No C or I impact was inferred from the theoretical possibility of device access through an invalid GPU address.
* No repeated release or refcount operation exists. The added unpin calls balance the newly added pin and do not fix a double release.
* No active-state retirement or translation-domain bonus was awarded. The fix prevents eviction by pinning rather than restoring a bypassed zap, invalidation, rmap removal, or synchronization protocol.
3. Reachability analysis
The evaluated attacker is a local process with permission to open the AMDGPU render node and use the user queue interface. This is below host administrator in normal render-node deployments, so HostAdminRequired is NO.
The operation is based on DRM device-file access rather than a capability evaluated in an attacker-created user or network namespace. NamespaceScopedPrivilege is therefore NO.
The attacker can instantiate user queues and submit GPU work. The paranoid interpretation also assumes that the same process can create sufficient GPU or TTM pressure to make eviction practical while the queue remains active. The conservative score does not assume fully deterministic eviction control.
Prior system setup is limited to supported AMDGPU hardware, the affected driver, enabled user queue functionality, and render-node access. No network trigger is established. No special timing race is required by the patch. The issue instead depends on eviction occurring while hardware still consumes the queue objects.
4. Primitive and severity interpretation
The exact primitive is loss of required GTT residency for MQD and firmware BOs that remain actively referenced by GPU hardware. Hardware subsequently uses stale or invalid GPU addresses, causing page faults and a reported system hang.
This is fixed stale device-address state, not a CPU use-after-free. The objects are not shown to have been freed, reclaimed, or replaced with attacker-controlled storage. There is no concrete repeated release or refcount mutation.
PrimitiveClass is CLASS_C because the demonstrated behavior is invalid-state device access resulting in page faults and DoS, without an independently useful corruption primitive.
The demonstrated consequence is a local system hang. A DMA access to unintended memory is a theoretical concern worth checking against the GPU VM and IOMMU behavior, but it is not established by this patch and is not scored.
ImportantCandidate is NO because neither meaningful kernel corruption leverage nor a security-boundary bypass is demonstrated. AutoDowngradeAllowed is YES because the established issue is a constrained local DoS and no unresolved Class A or strong Class B primitive remains.
The likely operational severity is Moderate for systems exposing affected AMDGPU user queues to untrusted local users.
5. One-sentence report phrase
An unprivileged AMDGPU render-node user may keep a user queue active while its unpinned MQD or firmware buffer is evicted, causing GPU page faults and a system hang without a demonstrated memory-corruption or privilege-escalation primitive.
6. Manual review recommendation
MANUAL CHECK RECOMMENDED
Confirm whether the affected GPU VM and IOMMU configurations always fault stale GPU addresses or whether any supported configuration could redirect hardware access to another mapped object.
7. Call-site confidence
Call-site confidence: medium
======================================================================
UPSTREAM PATCH SUMMARY
======================================================================
Patch: drm/amdgpu/userq: pin mqd and fw object bo to avoid eviction [ Upstream
Commit: a16d9986b7b27b08a696524a4dbf03dbf5433a16
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=a16d9986b7b27b08a696524a4dbf03dbf5433a16
Changed files:
drivers/gpu/drm/amd/amdgpu/amdgpu_userq.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=a16d9986b7b27b08a696524a4dbf03dbf5433a16
======================================================================
DETAILED REPORT METHODOLOGY
======================================================================
The original Linux kernel CVE announcement for CVE-2026-97498 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-97498
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:H
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: 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 [email protected] (and both
send reply to CVE record itself too and see "reply" button below for howto reply).
^ permalink raw reply [relevance 21%]
Results 1-1 of 1 | reverse | sort options + mbox downloads above
-- links below jump to the message on this page --
2026-09-24 22:35 21% [CVE-2026-97498][MODERATE REGULAR] drm/amdgpu/userq: pin mqd and fw object bo to avoid eviction [ Upstream AL-KERNEL
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox