From mboxrd@z Thu Jan 1 00:00:00 1970 From: AL-KERNEL To: kernel-cve@kernelcve.org Subject: [CVE-2026-97916][MODERATE REGULAR] accel/ivpu: Validate firmware log buffer metadata Date: Fri, 25 Sep 2026 10:46:43 -0400 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-AL-KERNEL-CVE: CVE-2026-97916 X-AL-KERNEL-Priority: MODERATE REGULAR X-AL-KERNEL-Severity: MODERATE REGULAR X-AL-KERNEL-Base-Severity: MODERATE X-AL-KERNEL-KPANIC: NO X-AL-KERNEL-ActionableScore: 4 X-AL-KERNEL-ActionableScore-Lower: unknown X-AL-KERNEL-Commit: 8ec3d6972f1440fae6c8dadd1859892775208759 List-Id: CVE: CVE-2026-97916 Priority: MODERATE REGULAR AL-KERNEL base severity: MODERATE KPANIC flag: NO Patch: accel/ivpu: Validate firmware log buffer metadata Commit: 8ec3d6972f1440fae6c8dadd1859892775208759 Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=8ec3d6972f1440fae6c8dadd1859892775208759 Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-97916 Analysis date: Fri, 25 Sep 2026 10:46:43 -0400 ActionableScore: 4 ActionableScore lower bound: unknown Actionable bucket: Borderline / manual review recommended Manual review required: YES Summary: A TOCTOU in IVPU firmware log parsing allows firmware-controlled DMA metadata to cause a host-kernel out-of-bounds read, potentially exposing adjacent data through diagnostic output or crashing the host. ====================================================================== ABOUT THIS REPORT ====================================================================== The original Linux kernel CVE announcement for CVE-2026-97916 is available here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-97916 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-2026-97916 MODERATE https://www.kernelcve.org/list/?q=CVE-2026-97916 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:L/I:N/A:H';CWE-367;CWE-125;*CWE-20;CWE-362;Other CVSS 'AV:L/AC:H/PR:H/UI:N/S:U/C:L/I:N/A:H';BEST CVSS score: '5.3';DESCR 'A TOCTOU race exists in the ivpu firmware log path because header_size and size are validated in DMA shared memory and then read again when fw_log_print_buffer() calculates the log data address and length. NPU firmware can modify this metadata between validation and use, causing the host kernel to read beyond the validated firmware log buffer while printing its contents. This provides an out of bounds read primitive that may expose adjacent mapped data through diagnostic output or trigger a host kernel fault and denial of service. There is no concrete out of bounds write or UAF primitive in the shown code, so integrity impact and privilege escalation are not established. For the CVSS the PR:L value in the paranoid score represents deployments where an untrusted local process can access the NPU and influence firmware workload, logging activity, and race timing. Direct firmware replacement or full firmware control would normally require higher privileges. The issue is local and is not network reachable. AC:H applies because successful triggering depends on changing shared metadata within the check to use timing window.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Borderline / manual review recommended (with actual score 4) SKIP CVE-2026-97916 UNKNOWN SKIP The Fixes patch not applied yet, so unlikely that actual: url YES NO NO unknown MAYBE OOB LOG DECREASED_TO_MODERATE_REGULAR_FROM_MODERATE7_BASED_ON_GUESSCVSS_AND_CIANNH - - checked ====================================================================== ACTIONABLESCORE ANALYSIS ====================================================================== ActionableScore=4 PrimitiveClass=CLASS_C RealUAF=NO HostAdminRequired=UNKNOWN NamespaceScopedPrivilege=UNKNOWN ImportantCandidate=NO ManualReview=RECOMMENDED AutoDowngradeAllowed=YES ## 1. ActionableScore * Raw conservative score: 4 * Conservative score: 4 * Raw paranoid score: 4 * Paranoid score: 4 * No floor or cap applied. * Final recommended bucket: **Borderline / manual review recommended**:: ## 2. Signal breakdown Positive signals: * +1 Weak or indirect corruption candidate. Mutable DMA metadata passes validation and is then reread, producing a concrete host-kernel OOB read. * +1 Firmware-mediated influence. NPU firmware can modify the shared `header_size` and `size` fields involved in the vulnerable calculation. * +1 Reliable kernel crash or strong DoS. An attacker-influenced address and length can make the host dereference memory beyond the mapped log buffer. * +1 Confidentiality plausible. The OOB data is consumed by a log-printing path using a DRM printer, providing a plausible disclosure path rather than a purely internal invalid read. No negative signals applied: * No race penalty. The patch establishes a TOCTOU, but does not establish that the timing is one-shot, externally uncontrollable, or difficult to repeat. * No host-admin penalty. Direct firmware replacement may require high privilege, but the complete trigger and privilege boundary are unresolved. * No rare-hardware penalty. An IVPU-capable NPU is required, but the evidence does not establish physical-only or genuinely rare hardware. * No metadata-only penalty because an actual OOB host read is established. Important non-awarded signals: * No UAF, stale lifetime reference, repeated release, or refcount mutation exists. * No OOB write, arbitrary write, controlled reclaim, callback influence, or integrity primitive is demonstrated. * No direct network reachability is established. * Strong LPE leverage is not awarded merely because firmware controls the source metadata. * Extra Availability is not awarded because it would double-count the same potential host crash. * No additional memory-layout point is awarded because it would duplicate the established OOB-read primitive. ## 3. Reachability analysis The NPU firmware directly controls the DMA-shared log metadata, while the host invokes the vulnerable path when parsing or printing firmware logs. A malicious or compromised firmware image can therefore provide the required metadata changes, but the supplied evidence does not establish whether an ordinary local NPU client can independently control those fields. A local user may be able to influence NPU workload and logging activity, but that alone does not prove control over the complete TOCTOU sequence. Direct firmware installation or replacement may require host administration, but this prerequisite is not positively established for every trigger scenario. Therefore `HostAdminRequired=UNKNOWN`. No namespace-scoped capability path is shown. The issue is not directly network-triggerable. An IVPU-capable NPU, loaded firmware, allocated firmware log BOs, and invocation of the log-printing path are prerequisites. Repeated firmware writes could make the race repeatable, so no hard-race penalty is justified. ## 4. Primitive and severity interpretation The exact primitive is a TOCTOU-driven OOB read from a DMA-shared firmware log buffer. The validated `header_size` and `size` values could be replaced before `fw_log_print_buffer()` reread them, causing `log_data` and `data_size` to describe memory outside the validated BO range. This is not a lifetime violation and does not involve stale storage after free. It is an active shared-metadata validation failure. The demonstrated consequences are an invalid host memory read and a plausible kernel fault. Because the data is processed by a printing path, disclosure of adjacent mapped bytes is plausible, although sensitive-data exposure is not demonstrated. There is no supported write primitive, UAF, controlled object replacement, or privilege-escalation chain. The issue is therefore `CLASS_C`, not Class A or Class B. `ImportantCandidate=NO` because no Important-preserving corruption or boundary-bypass primitive is established. `AutoDowngradeAllowed=YES` because the unresolved concern is limited to OOB read, possible disclosure, and DoS rather than active kernel-state corruption. Operationally, this is a borderline Moderate issue whose practical severity depends heavily on whether below-host-admin software can cause firmware to mutate the metadata deliberately. ## 5. One-sentence report phrase A TOCTOU in IVPU firmware log parsing allows firmware-controlled DMA metadata to cause a host-kernel out-of-bounds read, potentially exposing adjacent data through diagnostic output or crashing the host. ## 6. Manual review recommendation MANUAL CHECK RECOMMENDED Confirm who can invoke the log-printing interface, who can influence the NPU firmware metadata, what memory can be reached beyond the BO, and whether the DRM printer output is visible below host-admin privilege. ## 7. Call-site confidence Call-site confidence: medium ====================================================================== UPSTREAM PATCH SUMMARY ====================================================================== Patch: accel/ivpu: Validate firmware log buffer metadata Commit: 8ec3d6972f1440fae6c8dadd1859892775208759 Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=8ec3d6972f1440fae6c8dadd1859892775208759 Commit description: The tracing log headers parsed by fw_log_print_buffer() reside in DMA-shared BOs that the NPU firmware can write to. fw_log_from_bo() validated log->header_size and log->size, but fw_log_print_buffer() re-read those same fields from shared memory afterwards, allowing a TOCTOU where firmware changes them between the check and the use, and making the host dereference out-of-bounds addresses while printing logs. Snapshot the validated values once with READ_ONCE() and pass them down explicitly in a new struct ivpu_fw_log_desc instead of re-reading them from the shared struct. Cc: stable@vger.kernel.org Fixes: d4e4257 ("accel/ivpu: Add firmware tracing support") Signed-off-by: Magdalena Schulfer Signed-off-by: Dawid Osuchowski Reviewed-by: Karol Wachowski Signed-off-by: Karol Wachowski Link: https://patch.msgid.link/20260901125749.404338-3-dawid.osuchowski@linux.intel.com Signed-off-by: Greg Kroah-Hartman Changed files: drivers/accel/ivpu/ivpu_fw_log.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=8ec3d6972f1440fae6c8dadd1859892775208759 ====================================================================== DETAILED REPORT METHODOLOGY ====================================================================== The original Linux kernel CVE announcement for CVE-2026-97916 can be found here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-97916 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:H/UI:N/S:U/C:L/I:N/A:H The Best / paranoid CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:N/A:H The Best / paranoid CVSS score: 5.3 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: 4 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).