From mboxrd@z Thu Jan 1 00:00:00 1970 From: AL-KERNEL To: kernel-cve@kernelcve.org Subject: [CVE-2026-89913][MODERATE 7.0] KVM: arm64: vgic-v3: take an LPI reference in vgic_v3_save_pending_tables Date: Wed, 16 Sep 2026 08:38:09 -0400 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-AL-KERNEL-CVE: CVE-2026-89913 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: 7631f95297560157d3a9283cb999e3be00103348 List-Id: CVE: CVE-2026-89913 Priority: MODERATE 7.0 AL-KERNEL base severity: MODERATE KPANIC flag: YES Patch: KVM: arm64: vgic-v3: take an LPI reference in vgic_v3_save_pending_tables Commit: 7631f95297560157d3a9283cb999e3be00103348 Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=7631f95297560157d3a9283cb999e3be00103348 Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89913 Analysis date: Wed, 16 Sep 2026 08:38:09 -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-2026-89913 is available here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89913 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-2026-89913 MODERATE https://www.kernelcve.org/list/?q=CVE-2026-89913 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-664;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 KVM arm64 VGICv3 LPI handling can let vgic_v3_save_pending_tables use a stale struct vgic_irq pointer after the LPI was removed from lpi_xa and freed through RCU. The vulnerable path reads LPI fields and performs guest memory operations after only temporary xarray RCU coverage, while a concurrent irqfd cached injection path and guest ITS DISCARD can drop the last LPI reference. For the CVSS the PR:L is used because triggering requires a local user or guest owner with access to a KVM VM or VMM interface rather than full host administrator privileges. The issue is not directly network reachable and is a local virtualization attack surface. Impact is at least denial of service via host kernel crash and in the worst case may allow confidentiality or integrity impact because the bug is a kernel use after free in host context.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Actionable Moderate / Strong Important candidate requiring manual review (with actual score 5) YES DANGER VIRT RACE UAF HARDWARE LINUS KPANIC NO NO checked ====================================================================== ACTIONABLESCORE ANALYSIS ====================================================================== ActionableScore=5 ActionableScoreLower=4 PrimitiveClass=CLASS_B RealUAF=YES HostAdminRequired=NO NamespaceScopedPrivilege=NO ImportantCandidate=YES ManualReview=REQUIRED AutoDowngradeAllowed=NO 1. ActionableScore * Raw conservative score: 3 * Conservative score: 4 * Applied conservative floor/cap: Real UAF FLOOR 4 applied because a real active structured host kernel object can be used after lifetime end and host-admin-equivalent privilege is not established. * Raw paranoid score: 5 * Paranoid score: 5 * Applied paranoid floor/cap: none * Final recommended bucket: **Actionable Moderate / Strong Important candidate requiring manual review**:: 2. Signal breakdown Positive conservative signals: * +2 Real UAF / lifetime violation. `vgic_v3_save_pending_tables()` can retain a `struct vgic_irq` pointer from `lpi_xa` without a stable reference while another path removes the LPI and frees it through `kfree_rcu()`. * +1 Reliable kernel crash / strong DoS. The stale pointer is dereferenced in host KVM context after the object may have been freed. Positive paranoid signals: * +2 Real UAF / lifetime violation. Same as conservative. * +1 Reliable kernel crash / strong DoS. Same as conservative. * +1 Attacker-influenced active-object lifetime manipulation. A local VM owner or VMM-controlled environment can plausibly coordinate guest ITS DISCARD, irqfd cached injection activity, and pending-table save operations, although a pure guest-only trigger is not established. * +1 Structured stale-object exploitation leverage. The stale `struct vgic_irq` is an active structured KVM object and the stale path consumes fields such as `target_vcpu` and pending state before performing guest memory operations. This is meaningfully beyond a shallow cleanup-only dereference, but it is not enough for Class A. Non-awarded or withheld signals: * +2 Strong LPE / strong exploitation leverage not awarded. Controlled reclaim, attacker-shaped replacement contents, callback control, arbitrary host read/write, refcount takeover, or demonstrated guest-to-host escape are not established. * Concrete repeated release/refcount signal not awarded. The bug is missing a stable reference before dereference, not a demonstrated first put followed by a second lifetime mutation through the exact same dangling reference. * Paranoid repeated-refcount leverage not awarded. No concrete second stale `put` or refcount mutation is shown. * Extra Availability not awarded. The host crash/DoS is already counted once. * Confidentiality and Integrity bonuses not awarded. C/I impact is possible only as an unresolved UAF consequence, not as an established disclosure or write primitive. * Direct remote reachability not awarded. This is a local virtualization attack surface, not a network packet-processing vulnerability. * KVM strong guest-to-host exploitation leverage not awarded. The evidence establishes host-side KVM UAF and possible host crash, but not a concrete guest-controlled host memory corruption primitive beyond the UAF itself. * Race penalty not applied. Although this is a race, the supplied evidence does not prove it is a narrow one-shot or externally uncontrollable timing window sufficient for a scoring penalty. * Host-admin penalty not applied. Full host-admin-equivalent privilege is not positively established for a VMM process with `/dev/kvm` access. 3. Reachability analysis The most defensible attacker model is a local VMM / VM-owner context with access to the KVM VM and the relevant arm64 VGICv3 ITS/LPI functionality. A malicious guest can influence LPI lifecycle through ITS DISCARD, but the complete trigger also appears to involve host-side KVM activity such as `vgic_v3_save_pending_tables()` and irqfd cached injection behavior. Therefore this should not be treated as a pure guest-only trigger unless additional evidence shows the guest can induce all required host-side stages. Host administrator privilege is not positively required because KVM VMs are commonly operated by non-root VMM processes with delegated `/dev/kvm` access. However, ordinary unprivileged local reachability is not established without that KVM access. The issue is not namespace-scoped. Network reachability is not established. The relevant prerequisites are arm64 KVM with VGICv3 ITS/LPI and irqfd/cached translation paths, which are virtualization-feature prerequisites rather than rare physical-only hardware requirements. Race repeatability is unresolved. The race involves concurrent pending-table save and LPI teardown after reference drops and RCU freeing. The operations appear repeatable in a VM lifecycle context, but no controlled reclaim or deterministic exploitation sequence is supplied. 4. Primitive and severity interpretation The exact primitive is a host-kernel use-after-free of `struct vgic_irq`. The object is active structured KVM state, not fixed stale storage. The vulnerable path is not merely destructor cleanup. It reads LPI fields, obtains `target_vcpu`, evaluates pending state, and performs guest memory read/write operations while lacking a stable LPI reference. This is PrimitiveClass=CLASS_B. It is a real lifetime/corruption primitive requiring exploitation-leverage analysis, but it is not Class A because no controlled reclaim, arbitrary write, callback hijack, refcount takeover, or demonstrated guest-to-host compromise is established. The demonstrated consequence is at least host kernel crash or DoS. The speculative consequence is stronger host impact if stale `vgic_irq` reuse can be shaped so that consumed fields drive useful host behavior, but that exploitation primitive is not demonstrated. ImportantCandidate=YES because this is a non-cleanup active-object UAF in KVM host context with unresolved exploitation leverage. AutoDowngradeAllowed=NO because automatically closing it as ordinary Moderate would discard a concrete host-kernel lifetime primitive on a virtualization boundary. Likely operational severity is actionable Moderate, with Strong Important review warranted if later analysis shows guest-only triggering, controlled reclaim, attacker-shaped replacement fields, or host memory corruption beyond crash. 5. One-sentence report phrase A race in KVM arm64 VGICv3 LPI handling can let `vgic_v3_save_pending_tables()` dereference a stale `struct vgic_irq` after removal from `lpi_xa` and RCU freeing, causing at least host kernel DoS and requiring manual review for unresolved guest/VMM-to-host UAF exploitation leverage. 6. Manual review recommendation MANUAL CHECK REQUIRED Reason: this is a real active-object UAF in host KVM state, with stale structured fields consumed outside a pure cleanup path, and the boundary/reachability implications are not sufficiently resolved for automatic downgrade. 7. Call-site confidence Call-site confidence: medium ====================================================================== UPSTREAM PATCH SUMMARY ====================================================================== Patch: KVM: arm64: vgic-v3: take an LPI reference in vgic_v3_save_pending_tables Commit: 7631f95297560157d3a9283cb999e3be00103348 Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=7631f95297560157d3a9283cb999e3be00103348 Commit description: vgic_v3_save_pending_tables() iterates dist->lpi_xa using xa_for_each() and dereferences the returned struct vgic_irq in the loop body without holding a reference on the LPI. The xarray iterator only provides temporary RCU coverage while looking up the current entry. That is not sufficient for this loop body, which reads fields from struct vgic_irq and performs guest memory accesses before the iteration completes. A concurrent path can trigger this race: the irqfd cached injection path (vgic_its_inject_cached_translation) obtains a transient LPI reference via vgic_its_check_cache() without holding kvm->lock, vcpu->mutex, config_lock, or its_lock. If guest ITS DISCARD then drops the cache and ITE references under its_lock, the transient inject reference may become the final one. When vgic_put_irq() drops it, the LPI is erased from lpi_xa and freed via kfree_rcu(). Meanwhile, vgic_v3_save_pending_tables() may still hold a stale pointer obtained from the xarray iterator and dereference it after the RCU grace period completes. Fix this by re-fetching each iterated LPI via vgic_get_irq(), which takes a stable reference, and dropping it with vgic_put_irq() on all paths. This matches the pattern already used by other lpi_xa iterators in the vgic ITS code. Cc: stable@vger.kernel.org Signed-off-by: Qihang Reviewed-by: Marc Zyngier Link: https://patch.msgid.link/20260807025534.34125-1-q.h.hack.winter@gmail.com Signed-off-by: Oliver Upton Signed-off-by: Greg Kroah-Hartman Changed files: arch/arm64/kvm/vgic/vgic-v3.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=7631f95297560157d3a9283cb999e3be00103348 ====================================================================== DETAILED REPORT METHODOLOGY ====================================================================== The original Linux kernel CVE announcement for CVE-2026-89913 can be found here: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-89913 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: 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).