From mboxrd@z Thu Jan 1 00:00:00 1970 From: AL-KERNEL To: kernel-cve@kernelcve.org Subject: [CVE-2026-97422][MODERATE 7.0] drm/amdkfd: fix SMI event cross-process information leak [ Upstream Date: Thu, 24 Sep 2026 19:23:17 -0400 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-AL-KERNEL-CVE: CVE-2026-97422 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: 5bb2dfa65d8da83e87b2a31c270bc567df69a5ca List-Id: CVE: CVE-2026-97422 Priority: MODERATE 7.0 AL-KERNEL base severity: MODERATE KPANIC flag: YES Patch: drm/amdkfd: fix SMI event cross-process information leak [ Upstream Commit: 5bb2dfa65d8da83e87b2a31c270bc567df69a5ca Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=5bb2dfa65d8da83e87b2a31c270bc567df69a5ca Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-97422 Analysis date: Thu, 24 Sep 2026 19:23:17 -0400 ActionableScore: 5 ActionableScore lower bound: unknown Actionable bucket: Actionable Moderate Manual review required: YES Summary: A local AMD KFD client can bypass per-process SMI event filtering and monitor another users GPU process lifecycle, command name, PID, and VM-fault activity without CAP_SYS_ADMIN. ====================================================================== ABOUT THIS REPORT ====================================================================== The original Linux kernel CVE announcement for CVE-2026-97422 is available here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-97422 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-2026-97422 MODERATE https://www.kernelcve.org/list/?q=CVE-2026-97422 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:L/I:N/A:N';*CWE-200;CWE-862;CWE-284;BEST CVSS score: '3.3';DESCR 'An access control error in the AMD KFD SMI event path allows events with pid 0 to bypass the existing per user privilege check. PROCESS_START, PROCESS_END, and VMFAULT events can therefore expose another process PID, command name, and GPU fault activity to unrelated local KFD clients. For the CVSS the PR:L value is used because an attacker needs local user access and permission to open /dev/kfd, commonly through membership in the render group. The issue is not network reachable and does not require interaction from another user. Impact is limited to cross process disclosure of GPU workload metadata. The available code does not provide an arbitrary memory read, memory corruption, data modification, denial of service, or privilege escalation primitive.';NO MANUAL CHECK; ,and ActionableScore result is Actionable Moderate (with actual score 5) SKIP CVE-2026-97422 UNKNOWN SKIP No affected files built, so skip this CVE NO - - unknown MAYBE LEAK HARDWARE KPANIC INCREASED_TO_MODERATE70_BASED_ON_ACTIONABLESCOREHIGHEREQTHAN5 - - checked ====================================================================== ACTIONABLESCORE ANALYSIS ====================================================================== ActionableScore=5 PrimitiveClass=CLASS_A RealUAF=NO 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 was applied. * Final recommended bucket: **Actionable Moderate**:: ## 2. Signal breakdown Positive signals: * +2 Concrete authorization decision bypass. Passing `pid=0` causes `kfd_smi_ev_enabled()` to skip the intended suser privilege check, changing event delivery from restricted to allowed. * +1 Local unprivileged trigger. Any local user already authorized to open `/dev/kfd`, commonly through the render group, can receive the affected events without host-admin privileges. * +1 Confidentiality impact. The delivered events disclose another process PID, command name, lifecycle, and GPU VM-fault activity. * +1 Central authorization helper fan-out. The defective per-client predicate affects multiple visible event classes: PROCESS_START, PROCESS_END, and VMFAULT. No negative signals were applied. Render-group access is below host-admin privilege and is the ordinary access model for KFD clients, not evidence of host-administrator control. Material non-awarded signals: * No UAF, OOB access, arbitrary read, arbitrary write, or other memory-corruption primitive is established. * No integrity or availability impact is established. * No remote or network trigger exists. * No protected FD, credential, memory contents, or privileged file is disclosed. * No privilege-escalation chain is demonstrated. * No extra practical-exploit-chain point was added because direct unauthorized event delivery is already represented by the authorization and confidentiality signals. * No paranoid escalation was applied because the disclosed information remains bounded GPU workload metadata. ## 3. Reachability analysis A local user with permission to open `/dev/kfd` can subscribe to SMI events and receive events belonging to unrelated GPU processes. Membership in the render group is sufficient according to the supplied commit description, so host root and host-admin-equivalent capabilities are not required. The attacker does not need `CAP_SYS_ADMIN`. That capability identifies subscribers that are intentionally permitted to monitor other processes, while the bug incorrectly gives comparable event visibility to ordinary KFD clients. No user-namespace or network-namespace capability is involved in the trigger. Prior provisioning of AMD KFD device access and suitable AMD GPU hardware is required, but this is normal device exposure rather than host-admin participation in each attack. There is no race or timing requirement. ## 4. Primitive and severity interpretation The exact primitive is a cross-process authorization bypass resulting in unauthorized disclosure of structured GPU workload metadata. It is not a memory-lifetime defect, stale-object use, or corruption primitive. The incorrect zero PID suppresses the intended ownership or privilege filtering. PROCESS_START, PROCESS_END, and VMFAULT records can then reveal another process PID, command name, and GPU fault activity to unrelated KFD clients. This is CLASS_A under the authorization calibration because a mandatory access-control decision is concretely bypassed and the fix restores process-specific filtering. The demonstrated consequence is limited information disclosure. Kernel memory disclosure, modification of another process, denial of service, and privilege escalation remain unsupported. ImportantCandidate is YES because the patch establishes a real security-boundary bypass, although the bounded metadata impact keeps the numeric score in Actionable Moderate rather than Strong Important. AutoDowngradeAllowed is NO because automatically treating it as an ordinary low-priority information leak would discard the concrete cross-process authorization failure. ## 5. One-sentence report phrase A local AMD KFD client can bypass per-process SMI event filtering and monitor another users GPU process lifecycle, command name, PID, and VM-fault activity without CAP_SYS_ADMIN. ## 6. Manual review recommendation MANUAL CHECK REQUIRED Confirm whether product-specific `/dev/kfd` policy exposes the interface to mutually untrusted render-group users and whether process names or GPU activity reveal sensitive tenant workload information. ## 7. Call-site confidence Call-site confidence: high ====================================================================== UPSTREAM PATCH SUMMARY ====================================================================== Patch: drm/amdkfd: fix SMI event cross-process information leak [ Upstream Commit: 5bb2dfa65d8da83e87b2a31c270bc567df69a5ca Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=5bb2dfa65d8da83e87b2a31c270bc567df69a5ca Changed files: drivers/gpu/drm/amd/amdkfd/kfd_smi_events.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=5bb2dfa65d8da83e87b2a31c270bc567df69a5ca ====================================================================== DETAILED REPORT METHODOLOGY ====================================================================== The original Linux kernel CVE announcement for CVE-2026-97422 can be found here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-97422 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: Not available The Best / paranoid CVSS vector: AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:N/A:N The Best / paranoid CVSS score: 3.3 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).