From mboxrd@z Thu Jan 1 00:00:00 1970 From: AL-KERNEL To: kernel-cve@kernelcve.org Subject: [CVE-2026-97429][MODERATE 7.0] drm/amdkfd: fix UAF race in destroy_queue_cpsch [ Upstream Date: Thu, 24 Sep 2026 20:09:44 -0400 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-AL-KERNEL-CVE: CVE-2026-97429 X-AL-KERNEL-Priority: MODERATE 7.0 X-AL-KERNEL-Severity: MODERATE 7.0 X-AL-KERNEL-Base-Severity: MODERATE X-AL-KERNEL-KPANIC: YES X-AL-KERNEL-ActionableScore: 5 X-AL-KERNEL-ActionableScore-Lower: unknown X-AL-KERNEL-Commit: 41144829f4645c2bd4cb501d34d253858f750c0a List-Id: CVE: CVE-2026-97429 Priority: MODERATE 7.0 AL-KERNEL base severity: MODERATE KPANIC flag: YES Patch: drm/amdkfd: fix UAF race in destroy_queue_cpsch [ Upstream Commit: 41144829f4645c2bd4cb501d34d253858f750c0a Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=41144829f4645c2bd4cb501d34d253858f750c0a Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-97429 Analysis date: Thu, 24 Sep 2026 20:09:44 -0400 ActionableScore: 5 ActionableScore lower bound: unknown Actionable bucket: Actionable Moderate / Strong Important candidate requiring manual review Manual review required: YES Summary: No one-sentence report phrase was found. ====================================================================== ABOUT THIS REPORT ====================================================================== The original Linux kernel CVE announcement for CVE-2026-97429 is available here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-97429 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-2026-97429 MODERATE https://www.kernelcve.org/list/?q=CVE-2026-97429 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-362;CWE-667;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:H';BEST CVSS score: '7';DESCR 'A race in drm amdkfd queue destruction can cause a use-after-free because wait_on_destroy_queue drops the DQM locks while waiting for a suspended debug queue to resume. During this window another thread can destroy and free the same queue, after which the original destruction path may continue with a stale q pointer. A local user able to access /dev/kfd and create or manage KFD queues may attempt to trigger concurrent queue destruction. For the CVSS the PR:L reflects the requirement for a local user context and KFD device access without requiring full administrative control. AC:H reflects the timing dependent race window. The issue is not network reachable. Impact is at least a local denial of service through a kernel crash. In the worst defensible case the kernel use-after-free may affect confidentiality and integrity or permit privilege escalation, although a reliable write primitive is not demonstrated and manual review is required.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 5) SKIP CVE-2026-97429 UNKNOWN SKIP No affected files built, so skip this CVE NO - - unknown MAYBE RACE UAF HARDWARE KPANIC KPANIC INCREASED_TO_HIGH_BASED_ON_GUESSCVSS DECREASED_TO_MODERATE7_BASED_ON_FALSEPOSCHECKOFKP - - checked ====================================================================== ACTIONABLESCORE ANALYSIS ====================================================================== ActionableScore=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: 5 * Paranoid score: 5 * No floor or cap changed either score. * Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**:: 2. Signal breakdown Positive signals: * +2 Real UAF / active lifetime violation. While `wait_on_destroy_queue()` has dropped the DQM locks, a concurrent destroy can free the same queue that the waiting path still retains as `q`. * +1 Reliable kernel crash / strong DoS. Resuming the original path with a dangling queue pointer can realistically cause an invalid kernel dereference or allocator corruption. * +1 Local unprivileged trigger. An authorized non-administrative process with `/dev/kfd` access can create and manage its KFD queues without host-admin-equivalent privilege. * +1 Attacker-influenced active-object lifetime. The same process can create the queue and issue concurrent destruction operations while the suspended debug queue is waiting for resume. Negative signals: * No negative points. AMD GPU dependence is not treated as rare hardware. * No race penalty. The lock is deliberately dropped while waiting, providing a structured concurrency window, and concurrent queue operations can be driven repeatedly. * No host-admin penalty. KFD queue management is available to an authorized local GPU process and does not inherently require host root. Important non-awarded signals: * Controlled reclaim or attacker-shaped replacement contents are not demonstrated. * Arbitrary or directed kernel read/write is not demonstrated. * Structured stale-object leverage is not separately awarded because the visible stale activity remains in the queue-destruction path and the later field consumption is not shown. * Weak LPE is not awarded merely from the UAF class. * Confidentiality and integrity points are not awarded solely because stale queue storage might theoretically be reclaimed. * Remote reachability is not established. * Extra Availability is not awarded because it would duplicate the same potential kernel crash. Repeated-release assessment: * A concurrent destroy can end the queue lifetime. * The waiting path retains the exact `q` pointer across the unlocked interval. * A subsequent stale use is directly implied by the commit message. * The supplied code does not establish a specific second `put`, refcount decrement, `free`, or equivalent lifetime mutation by the waiting path. * Concrete repeated release/refcount +1 was therefore not awarded. * Paranoid repeated-refcount leverage +1 was withheld. * FLOOR 5 was not applied. The raw score independently reaches 5. 3. Reachability analysis The evaluated attacker is a local process authorized to use `/dev/kfd`, such as a GPU compute application or debugging process. Such access is normally below host-administrator privilege, although device-node permissions or a graphics group may gate access. The attacker can create and manage its own KFD queues and can issue concurrent queue-destruction operations. Triggering this specific path additionally requires debug-trap handling and a suspended queue so that `wait_on_destroy_queue()` drops the DQM locks while waiting for resume. No user-namespace or network-namespace capability is involved. The complete operation is device-FD based rather than namespace-capability based, so `NamespaceScopedPrivilege=NO`. Prior system setup consists of supported AMD GPU hardware, the AMDGPU/KFD driver, and user access to `/dev/kfd`. AMD GPU hardware is sufficiently common that no rare-hardware penalty is justified. The condition is not remotely reachable. The race is not shown to be a narrow, one-shot, externally uncontrollable interval. Waiting with locks intentionally released can give another thread a meaningful opportunity to destroy the queue, so no hard-race deduction is applied. 4. Primitive and severity interpretation The exact primitive is a use-after-free of an active structured KFD queue object. One destroy path waits with the manager locks released, another path can destroy and free the queue, and the original path still retains the stale queue pointer. This is broader than fixed asynchronous stale storage because the stale object is a live queue structure involved in queue-management processing. However, the supplied context primarily shows destruction-path use and does not establish controlled replacement, callback consumption, attacker-controlled post-free contents, or a directed memory-access primitive. No concrete second lifetime or refcount mutation is visible. This therefore differs from a proven double-release case in which an exact dangling reference receives a second `put`, refcount decrement, or `free`. The demonstrated primitive is CLASS_B. A kernel crash or corruption during stale queue processing is the practical supported consequence. Privilege escalation remains speculative because no controlled reclaim, arbitrary write, callback hijack, or credential impact is established. `ImportantCandidate=YES` because this is an active-object UAF reachable by a below-host-admin GPU process that directly influences the vulnerable lifetime transition. The unresolved stale-object behavior requires Important-versus-Moderate review. `AutoDowngradeAllowed=NO` because automatically reducing the issue to an ordinary crash-only bug would discard a concrete active-object lifetime violation with attacker-driven concurrent destruction. This does not independently prove Important severity, but it does require manual exploitation-leverage analysis. Likely operational severity is Actionable Moderate, with a possible Important classification if further source review establishes useful post-free field consumption, controllable reclaim, repeated lifetime mutation, or privilege escalation. 5. One-sentence report phrase A local process with access to `/dev/kfd` can race destruction of a suspended debug queue while `wait_on_destroy_queue()` has released the DQM locks, causing an active queue use-after-free and potential kernel crash, with stronger exploitation requiring manual review. 6. Manual review recommendation MANUAL CHECK REQUIRED The patch establishes an attacker-influenced active-object UAF below host-admin privilege, but the supplied context does not resolve which stale queue fields are subsequently consumed or whether reclaim can provide privilege-escalation leverage. 7. Call-site confidence Call-site confidence: medium ====================================================================== UPSTREAM PATCH SUMMARY ====================================================================== Patch: drm/amdkfd: fix UAF race in destroy_queue_cpsch [ Upstream Commit: 41144829f4645c2bd4cb501d34d253858f750c0a Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=41144829f4645c2bd4cb501d34d253858f750c0a Changed files: drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.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=41144829f4645c2bd4cb501d34d253858f750c0a ====================================================================== DETAILED REPORT METHODOLOGY ====================================================================== The original Linux kernel CVE announcement for CVE-2026-97429 can be found here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-97429 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: unknown Paranoid ActionableScore: 5 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).