From mboxrd@z Thu Jan 1 00:00:00 1970 From: AL-KERNEL To: kernel-cve@kernelcve.org Subject: [CVE-2026-89916][MODERATE 7.0] KVM: arm64: Make VNCR invalidation participate in MMU invalidation retry Date: Wed, 16 Sep 2026 07:17:31 -0400 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-AL-KERNEL-CVE: CVE-2026-89916 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: 4 X-AL-KERNEL-Commit: d8580e7aa189b0c6c5bb4d94b8352728a7c35bbd List-Id: CVE: CVE-2026-89916 Priority: MODERATE 7.0 AL-KERNEL base severity: MODERATE KPANIC flag: YES Patch: KVM: arm64: Make VNCR invalidation participate in MMU invalidation retry Commit: d8580e7aa189b0c6c5bb4d94b8352728a7c35bbd Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=d8580e7aa189b0c6c5bb4d94b8352728a7c35bbd Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89916 Analysis date: Wed, 16 Sep 2026 07:17:31 -0400 ActionableScore: 5 ActionableScore lower bound: 4 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-2026-89916 is available here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89916 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-2026-89916 MODERATE https://www.kernelcve.org/list/?q=CVE-2026-89916 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:L/A:H';CWE-362;CWE-667;*CWE-664;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L';BEST CVSS score: '5.8';DESCR 'KVM arm64 nested virtualization can install a stale VNCR TLB entry because a VNCR VA invalidation did not participate in the mmu_invalidate_seq retry protocol while another vCPU was translating the same VNCR page. A malicious or buggy guest with control over nested S1 page tables and TLBI operations could win this race and make KVM cache a translation for a page that was invalidated during a BBM sequence. For the CVSS the PR:L is used because reliable triggering requires code execution or control inside a guest vCPU context rather than unauthenticated network access, but it does not require full host administrator privileges. The issue is not network reachable from outside the host and is triggered through the KVM guest execution path. Impact is at least integrity and availability risk from stale translation state, and in the paranoid interpretation may include limited confidentiality or integrity impact, so manual review is recommended.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 5) SKIP CVE-2026-89916 UNKNOWN SKIP The Fixes patch not applied yet, so unlikely that actual: url YES NO NO unknown MAYBE VIRT HARDWARE LINUS KPANIC INCREASED_TO_MODERATE70_BASED_ON_ACTIONABLESCOREHIGHEREQTHAN5 - - checked ====================================================================== ACTIONABLESCORE ANALYSIS ====================================================================== ActionableScore=5 ActionableScoreLower=4 PrimitiveClass=CLASS_B RealUAF=NO HostAdminRequired=NO NamespaceScopedPrivilege=NO ImportantCandidate=YES ManualReview=REQUIRED AutoDowngradeAllowed=NO 1. ActionableScore * Raw conservative score: 4 * Conservative score: 4. No floor or cap applied. * Raw paranoid score: 5 * Paranoid score: 5. No floor or cap applied. * Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**:: 2. Signal breakdown Conservative signals: * +2. CLASS_B active KVM translation state lifetime and consistency violation. A VNCR translation can be walked before an S1 invalidation and then inserted into the VNCR TLB after the guest has invalidated that address. * +1. Below-host-admin guest execution path. The described scenario is driven by vCPU activity inside an already created VM or nested virtualization setup, not by host administrator code directly. * +1. Security-relevant stale active translation state. The stale VNCR TLB entry remains operational after the underlying guest S1 page table state has changed, which is stronger than passive stale metadata. Paranoid additional signal: * +1. Active semantic translation-domain transition. The old translation can survive across a transition from valid translated state to invalid guest S1 state without participating in the MMU invalidation retry protocol. Non-awarded signals: * RealUAF not awarded. The patch establishes stale translation state, not a heap object use-after-free. * Controlled reclaim not demonstrated. * Arbitrary or directed host kernel read or write not demonstrated. * Strong guest-to-host exploitation leverage not awarded. The evidence does not establish attacker-shaped replacement contents, callback control, host arbitrary R/W, or a demonstrated escape. * Reliable host crash point not separately awarded because the supplied text establishes stale TLB insertion but does not show a concrete panic trace. * Race penalty not applied. The bug is a race, but the supplied evidence does not prove a one-shot or externally uncontrollable timing window. * Remote reachability not awarded. This is KVM guest execution / local virtualization state, not a network packet path. * Extra Availability not awarded. No distinct persistent availability consequence beyond stale invalid translation risk is established. * C/I direct impact not awarded conservatively. Limited C/I concern is reflected only in the paranoid translation-state interpretation, not as a proven disclosure or write primitive. Repeated release/refcount checklist: * First release can end lifetime: not applicable. * Exact same reference survives: not applicable. * Second operation mutates lifetime/refcount: not applicable. * Concrete repeated release/refcount +1 awarded: no. * Paranoid repeated-refcount leverage +1 awarded: no. * FLOOR 5 applied: no. 3. Reachability analysis The evaluated attacker is primarily malicious guest or nested guest code running under an already created KVM VM. The commit scenario explicitly involves one vCPU walking page tables for VNCR translation while another vCPU updates guest S1 page tables and issues TLBI S1E2 for the same VA. This is not an unauthenticated network attack. It is also not a host-admin-only operation once the VM and relevant nested virtualization features are exposed. Host administrator or VMM setup may be needed to create the VM and expose arm64 nested virtualization features, but the actual invalidation and racing translation are guest/vCPU operations. The complete trigger appears guest-driven within the virtualization execution path, not namespace-scoped. User/network namespaces are not relevant here. Special prerequisites exist: arm64 KVM nested virtualization and VNCR/NV2-related support must be in use. I did not subtract a rare-configuration penalty because the patch itself does not establish that the feature is so rare or unreachable that it should reduce triage priority. Race repeatability is unresolved. It requires overlap between VNCR translation and VNCR/S1 invalidation, but the two operations are plausible with multiple vCPUs and are not shown to be a one-time teardown race. KVM attacker-role separation: * Malicious guest: controls guest S1 page table activity and TLBI operations described in the commit. * Local VMM / /dev/kvm process: controls VM creation and configuration, and may be responsible for exposing nested virtualization features. * Host administrator: may perform initial host virtualization setup, but the vulnerability mechanics do not require host-admin code to execute the racing VNCR operations. * Trigger model: mostly guest execution path under prior VM setup, not a pure host-admin management operation. 4. Primitive and severity interpretation The exact primitive is stale active KVM translation state: a VNCR TLB entry can be inserted after the guest has invalidated the corresponding S1 translation because VNCR invalidation did not bump `mmu_invalidate_seq` and the walker sampled the sequence too late. This is not a RealUAF and not a concrete repeated release/refcount mutation. It is an active MMU/TLB synchronization bug in host KVM state. The stale state is active structured translation state, not fixed passive storage. Demonstrated consequence: stale VNCR TLB insertion for a page that is now invalid. Speculative consequence: limited guest/host memory access confusion or stronger guest-to-host impact might be possible, but the supplied patch does not establish controlled host memory corruption, arbitrary R/W, controlled reclaim, or privilege escalation. PrimitiveClass is CLASS_B because this is an active translation-state lifetime/consistency violation in KVM/MMU code requiring exploitation-leverage analysis, but it is not a CLASS_A strong primitive. ImportantCandidate=YES because stale active translation in KVM nested virtualization is security-boundary relevant and should not be automatically downgraded without manual review. AutoDowngradeAllowed=NO because a concrete active translation invalidation/retry failure remains unresolved from a security perspective. This does not mean proven Important severity, but automatic closure as ordinary Moderate would risk missing a meaningful KVM boundary issue. Likely operational severity: Actionable Moderate at minimum, with Strong Important review warranted if further analysis shows guest-controlled access to host-owned memory or another concrete boundary-crossing primitive. 5. One-sentence report phrase KVM arm64 nested virtualization can cache a stale VNCR TLB translation when VNCR invalidation races with a parallel S1 page-table walk, allowing guest-driven stale translation state in host KVM and requiring manual review for possible security-boundary impact. 6. Manual review recommendation MANUAL CHECK REQUIRED The issue affects active KVM MMU/TLB state and a guest-controlled invalidation/translation race, so the absence of a demonstrated arbitrary read/write primitive is not enough to safely auto-close it. 7. Call-site confidence Call-site confidence: medium ====================================================================== UPSTREAM PATCH SUMMARY ====================================================================== Patch: KVM: arm64: Make VNCR invalidation participate in MMU invalidation retry Commit: d8580e7aa189b0c6c5bb4d94b8352728a7c35bbd Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=d8580e7aa189b0c6c5bb4d94b8352728a7c35bbd Commit description: A VNCR TLB invalidation can occur on one vcpu while another vcpu is faulting in this same page. Without correctly handling this, we can end up with the following scenario: - vcpu A walks the PTs to translate VNCR - before vcpu A is able to grab the MMU lock to insert the TLB, vcpu B updates the S1 PTs with an invalid entry, and issues a TLBI S1E2 for this VA - vcpu A inserts the TLB for something that is now invalid This isn't a new problem, and we manage S2 by having the MMU notifier to bump up mmu_invalidate_seq on invalidation so that the fault can be replayed. We can perform something similar here, and extend invalidate_vncr_va() to update the same counter, clearly indicating that the context has changed under our feet. This is safe as the invalidation always happen while holding the MMU lock for write, and that we sample the sequence number before walking S1. Fixes: 4ffa72a ("KVM: arm64: nv: Add S1 TLB invalidation primitive for VNCR_EL2") Reported-by: sashiko-bot@kernel.org Link: https://lore.kernel.org/r/20260801130454.5D9F11F00AC4@smtp.kernel.org Signed-off-by: Marc Zyngier Cc: stable@vger.kernel.org Link: https://patch.msgid.link/20260806091026.620700-8-maz@kernel.org Signed-off-by: Oliver Upton Signed-off-by: Greg Kroah-Hartman Changed files: arch/arm64/kvm/nested.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=d8580e7aa189b0c6c5bb4d94b8352728a7c35bbd ====================================================================== DETAILED REPORT METHODOLOGY ====================================================================== The original Linux kernel CVE announcement for CVE-2026-89916 can be found here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89916 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:L/A:L The Best / paranoid CVSS vector: AV:L/AC:H/PR:L/UI:N/S:U/C:L/I:L/A:H The Best / paranoid CVSS score: 5.8 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: 4 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).