From mboxrd@z Thu Jan 1 00:00:00 1970 From: AL-KERNEL To: kernel-cve@kernelcve.org Subject: [CVE-2026-68316][MODERATE 7.0] accel: ethosu: Fix element size accounting for cmd stream validation [ Upstream Date: Mon, 10 Aug 2026 08:51:19 -0400 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-AL-KERNEL-CVE: CVE-2026-68316 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: 7 X-AL-KERNEL-ActionableScore-Lower: 5 X-AL-KERNEL-Commit: b4ae748f8e6cb65bb86e5a281bbb5b5e5f106527 List-Id: CVE: CVE-2026-68316 Priority: MODERATE 7.0 AL-KERNEL base severity: MODERATE KPANIC flag: YES Patch: accel: ethosu: Fix element size accounting for cmd stream validation [ Upstream Commit: b4ae748f8e6cb65bb86e5a281bbb5b5e5f106527 Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=b4ae748f8e6cb65bb86e5a281bbb5b5e5f106527 Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-68316 Analysis date: Mon, 10 Aug 2026 08:51:19 -0400 ActionableScore: 7 ActionableScore lower bound: 5 Actionable bucket: Strong Important candidate / Actionable Moderate at minimum Manual review required: YES Summary: The Ethos U command stream validator can underestimate DMA buffer sizes for 16/32/64 bit elements, allowing a local user with accelerator access to submit commands that may cause out-of-bounds NPU DMA reads or writes, leading to DoS and possible confidentiality or integrity impact. ====================================================================== ABOUT THIS REPORT ====================================================================== The original Linux kernel CVE announcement for CVE-2026-68316 is available here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-68316 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-2026-68316 MODERATE CHECK WITH IMPACT FROM ORIG NN MODERATE Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H';CWE-131;*CWE-787;CWE-125;*CWE-20;Other CVSS 'AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:H';BEST CVSS score: '7.8';DESCR 'The Ethos U command stream validator can underestimate the required feature matrix DMA range when 16/32/64 bit elements are used. The NHWC path misses the element size in the channel offset calculation and the IFM/IFM2 and OFM precision bitfields are decoded differently. A local user with access to the Ethos U accelerator device may craft a command stream that passes validation with buffers that are smaller than the actual NPU memory access. For the CVSS the PR:L is used because triggering requires local access to the accelerator device or an equivalent service account with permission to submit command streams. The issue is not network reachable by itself. Impact is at least local denial of service and possible limited confidentiality or integrity impact via out of bounds DMA reads or writes. For the paranoid score, choose the highest still defensible interpretation supported by the bug class and patch context, with preference for manual review sensitivity over autoclosed false negatives.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Strong Important candidate / Actionable Moderate at minimum (with actual score 7) SKIP CVE-2026-68316 UNKNOWN SKIP No affected files built, so skip this CVE NO - - unknown MAYBE SIMPLEFIX LINUS KPANIC MEMORY KPANIC INCREASED_TO_HIGH_BASED_ON_GUESSCVSS DECREASED_TO_MODERATE7_BASED_ON_FALSEPOSCHECKOFKP - - checked ====================================================================== ACTIONABLESCORE ANALYSIS ====================================================================== ActionableScore=7 ActionableScoreLower=5 ## 1. ActionableScore * Conservative score: 5 * Paranoid score: 7 * Final recommended bucket: **Strong Important candidate / Actionable Moderate at minimum**:: ## 2. Signal breakdown Conservative score signals: * Local unprivileged trigger: +1. A local user or service with access to the Ethos U accelerator device can submit command streams. * Constrained kernel or device buffer overwrite: +1. The validator can approve buffers smaller than the actual NPU DMA access size. * Privileged kernel/device-management memory corruption path: +1. The issue is in accelerator command stream validation and affects DMA-adjacent memory access decisions. * Confidentiality impact plausible: +1. IFM/IFM2 paths can plausibly cause out-of-bounds DMA reads from adjacent mapped memory. * Integrity impact plausible: +1. OFM output paths can plausibly cause out-of-bounds DMA writes. * Availability impact realistic: +1. Invalid DMA access can plausibly crash the driver, fault the device, or destabilize the system. * Rare hardware/subsystem exposure: -1. Ethos U requires specific accelerator hardware and driver exposure. Conservative total: 5. Paranoid additional interpretation: * Weak LPE concern: +1. Out-of-bounds DMA writes may become security-relevant if attacker-controlled command streams can influence adjacent DMA mappings. * Stronger memory corruption interpretation: +1 over the conservative constrained overwrite model. In a weak IOMMU or permissive DMA mapping setup, OFM DMA writes may cross from validated buffers into other sensitive memory. Paranoid total: 7. ## 3. Reachability analysis The bug is locally reachable through the Arm Ethos U accelerator command stream submission path. It is not network reachable by itself. The practical attacker needs access to the accelerator device node or an equivalent local service account allowed to submit workloads. Namespaces and containers matter only if the accelerator device is passed through or exposed into the container. If the device node is not exposed to untrusted users, practical reachability is reduced. If it is exposed as a render or accelerator device to ordinary users, PR:L is appropriate. The path is not broadly default-enabled across Linux systems because it depends on Ethos U hardware and the new accelerator driver. However, on affected embedded or edge AI systems where the device is exposed to applications, crafted command streams are a realistic input source. Call-site confidence: high. The patch shows the validator calculating feature matrix lengths for IFM, IFM2, and OFM before accepting command stream memory ranges. ## 4. Severity interpretation This is more than an ordinary Moderate validation bug because the validation error can make DMA range checks too small. The realistic impact is at least local DoS and possible limited confidentiality or integrity impact through out-of-bounds DMA reads or writes. The theoretical worst case is stronger: if DMA isolation is weak or adjacent mappings are attacker-influenced, OFM writes could become a memory corruption primitive with LPE relevance. That path is not proven by the patch alone, so the conservative score stays Actionable Moderate, while the paranoid score is a Strong Important candidate. ## 5. One-sentence report phrase The Ethos U command stream validator can underestimate DMA buffer sizes for 16/32/64 bit elements, allowing a local user with accelerator access to submit commands that may cause out-of-bounds NPU DMA reads or writes, leading to DoS and possible confidentiality or integrity impact. ## 6. Manual review recommendation MANUAL CHECK REQUIRED. Reason: the patch fixes under-sized DMA validation for attacker-controlled command streams, including OFM output writes, which is a plausible memory corruption class even though practical exploitability depends on device-node permissions and DMA isolation. ====================================================================== UPSTREAM PATCH SUMMARY ====================================================================== Patch: accel: ethosu: Fix element size accounting for cmd stream validation [ Upstream Commit: b4ae748f8e6cb65bb86e5a281bbb5b5e5f106527 Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=b4ae748f8e6cb65bb86e5a281bbb5b5e5f106527 Changed files: drivers/accel/ethosu/ethosu_gem.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=b4ae748f8e6cb65bb86e5a281bbb5b5e5f106527 ====================================================================== DETAILED REPORT METHODOLOGY ====================================================================== The original Linux kernel CVE announcement for CVE-2026-68316 can be found here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-68316 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:L/PR:L/UI:N/S:U/C:L/I:L/A:H The Best / paranoid CVSS vector: AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H The Best / paranoid CVSS score: 7.8 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: 5 Paranoid ActionableScore: 7 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).