[CVE-2026-98637][MODERATE REGULAR] KVM: arm64: Do not clear VM-wide SVE feature on vCPU init failure [ Upstream

[CVE-2026-98637][MODERATE REGULAR] KVM: arm64: Do not clear VM-wide SVE feature on vCPU init failure [ Upstream

[AL-KERNEL]

CVE: CVE-2026-98637
Priority: MODERATE REGULAR
AL-KERNEL base severity: MODERATE
KPANIC flag: NO
Patch: KVM: arm64: Do not clear VM-wide SVE feature on vCPU init failure [ Upstream
Commit: bd46b41f412fb81c50bac1a0929ec31320f3a807
Upstream patch: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=bd46b41f412fb81c50bac1a0929ec31320f3a807
Original CVE announcement: https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98637
Analysis date: Mon, 19 Oct 2026 04:12:44 -0400
ActionableScore: 1
ActionableScore lower bound: unknown
Actionable bucket: Ordinary Moderate
Manual review required: NO

Summary:
A failed pKVM SVE vCPU initialization can corrupt VM-wide feature bookkeeping so that a later vCPU reaches `sve_load_state()` with a NULL SVE state on its first guest FP access, potentially crashing the hypervisor or host.

======================================================================
ABOUT THIS REPORT
======================================================================

The original Linux kernel CVE announcement for CVE-2026-98637 is available here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98637

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: NO

A detailed explanation of the methodology and priority rules is included
at the end of this message.

======================================================================
AL-KERNEL CLASSIFICATION RESULT
======================================================================

CVE-2026-98637	MODERATE	https://www.kernelcve.org/list/?q=CVE-2026-98637 CHECK WITH IMPACT FROM ORIG NN MODERATE | AEGIS IMPACT MODERATE_KPANIC AEGIS_AS=0 LEGACY=MODERATE RESULT=DIFF (final Legacy and Aegis differ; FINAL: MODERATE, Aegis: MODERATE_KPANIC)	Maybe valid. Check manually. Hints by AL-KERNEL: The best (paranoid) CVSS is 'AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H';**CWE-476;CWE-665;CWE-703;Other CVSS 'AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:N/A:H';BEST CVSS score: '5.5';DESCR 'A VM wide SVE state inconsistency can arise when pkvm_vcpu_init_sve() fails and clears KVM_ARM_VCPU_SVE while KVM_ARCH_FLAG_GUEST_HAS_SVE remains set. Later vCPUs can then skip the required SVE setup and register with a NULL sve_state. The first guest FP or SVE access can pass this NULL pointer to sve_load_state() and may crash the hypervisor or host. For the CVSS the PR:L reflects that a local process normally needs access to KVM or control of the VMM and must create the affected VM and vCPU state. The issue is not directly network reachable. The confirmed impact is denial of service. No UAF, out of bounds access, information disclosure, or arbitrary write primitive is demonstrated. The affected configuration is limited to arm64 KVM environments using pKVM and guest SVE, and triggering also depends on an earlier SVE vCPU initialization failure. Automated impact decision path. The original SimpleNet classifier selected MODERATE. The primary LLM estimated paranoid CVSS 5.5 and the production ActionableScore returned 1 with recommendation Ordinary Moderate. The production rules then evaluated CVSS thresholds and ActionableScore thresholds. No severity changing production threshold or specialist override was recorded. The legacy KP specialist has no review artifact for this run. The legacy FD or DH downgrade specialist has no review artifact for this run. Exact triggered rule markers were V41_INHERITED_KPANIC_PROMOTION_BLOCKED and MANUAL_REVIEW_REQUIRED. The independent weaker Aegis shadow evaluated CVSS 4.8 and ActionableScore 0 with primitive class CLASS_C and manual review REQUIRED. Its independently reconciled impact was MODERATE_KPANIC. The models disagreed but the aggregation rules retained the production result. The final reported impact is MODERATE.';YES REQUIRES MANUAL CHECK; ,and ActionableScore result is Ordinary Moderate (with actual score 1)	MAYBE	VIRT INIT SIMPLEFIX HARDWARE LINUS TEST  V41_INHERITED_KPANIC_PROMOTION_BLOCKED MANUAL_REVIEW_REQUIRED	-	-	checked

======================================================================
ACTIONABLESCORE ANALYSIS
======================================================================

ActionableScore=1
PrimitiveClass=CLASS_C
RealUAF=NO
HostAdminRequired=UNKNOWN
NamespaceScopedPrivilege=UNKNOWN
ImportantCandidate=NO
ManualReview=RECOMMENDED
AutoDowngradeAllowed=YES
RequiresHostUID0=UNKNOWN
AuthorizationDecisionBypass=NO
ProtectedResourceImpact=NONE
B4PendingReview=NO
B4PendingConfirmed=NO

1. ActionableScore

* Raw conservative score: 1
* Conservative score: 1
* Raw paranoid score: 1
* Paranoid score: 1
* No floor or cap was applied.
* Final recommended bucket: **Ordinary Moderate**::

2. Signal breakdown

Positive signals:

* +1 Reliable kernel or hypervisor crash concern. A failed SVE initialization can leave later vCPUs with a NULL `sve_state`, which is passed to `sve_load_state()` on the first guest FP or SVE access.
* +0 Invalid-state consumption. The inconsistent VM-wide feature state is concrete, but its demonstrated sink is a NULL-pointer dereference rather than an independently useful corruption primitive.

Negative signals:

* No negative privilege penalty. The complete privilege requirement for deliberately causing the initial SVE initialization failure is unresolved.
* No niche or hardware penalty. Arm64 pKVM with guest SVE is configuration-specific, but the evidence does not establish that it is both rare and inaccessible to the relevant actor.

Important non-awarded signals:

* No UAF, OOB write, repeated release, refcount corruption, controlled reclaim, arbitrary write, disclosure, or authorization bypass is established.
* No strong guest-to-host corruption point was awarded. Guest FP access triggers the final NULL use, but the guest is not shown to control the preceding host-side SVE initialization failure.
* No local-unprivileged bonus was awarded because access to `/dev/kvm` and control over the complete failure sequence are not established for an ordinary user.
* No extra Availability point was awarded because it would duplicate the same possible host or hypervisor crash.
* NEW52 B5-R does not apply because there is no heap OOB write or retry-accumulated offset.
* NEW53 B3-PENDING does not apply because no shared-backing write exists.
* V33 and V32 do not apply because no attacker-influenced active networking metadata write exists.
* NEW54 B4-PENDING does not apply because no DENY-to-ALLOW authorization decision changes.

3. Reachability analysis

The trigger has two distinct stages and potentially different principals:

* A local VMM or process with `/dev/kvm` access creates and initializes the VM and its vCPUs. It controls host-side KVM configuration but is not automatically equivalent to host root.
* An SVE initialization attempt must first fail because of an invalid vector length, a NULL host SVE state, or failure of `hyp_pin_shared_mem()`. The supplied evidence does not establish that a guest alone can cause this failure reliably.
* A later vCPU is registered after the VM-wide SVE feature bitmap was incorrectly cleared. Guest code then performs its first FP or SVE access and reaches `sve_load_state()` with the NULL pointer.
* The evaluated path is therefore a compound VMM plus guest sequence unless the same local VMM controls the guest workload.
* Host-administrator privilege is not positively required, but complete below-host-admin reachability is also not proven. `HostAdminRequired=UNKNOWN` is therefore appropriate.
* No user-namespace or namespace-scoped capability path is shown.
* Required prerequisites are arm64, KVM pKVM, guest SVE, multiple vCPU initialization attempts, and an initial SVE setup failure.
* This is an ordered error-state sequence, not a timing race.

4. Primitive and severity interpretation

The exact primitive is inconsistent VM-wide feature state followed by consumption of a NULL SVE state pointer. The vulnerable code clears `KVM_ARM_VCPU_SVE` after a per-vCPU initialization failure while `KVM_ARCH_FLAG_GUEST_HAS_SVE` remains set. Later initialization decisions therefore disagree about whether SVE setup is required.

The affected state is active structured KVM state, but the demonstrated operation is a direct NULL-pointer use. There is no post-lifetime object use, replacement object, stale callback, secondary index corruption, or write through attacker-shaped state. This is therefore `CLASS_C`, not Class A or Class B.

The patch-supported consequence is a hypervisor or host availability failure. A guest-to-host privilege escalation, host memory corruption, confidentiality loss, and integrity modification are speculative and unsupported.

`ImportantCandidate=NO` because no Important-preserving corruption or boundary-bypass primitive is established. `AutoDowngradeAllowed=YES` because the demonstrated defect remains an invalid-state crash path without unresolved exploitation leverage beyond denial of service.

5. One-sentence report phrase

A failed pKVM SVE vCPU initialization can corrupt VM-wide feature bookkeeping so that a later vCPU reaches `sve_load_state()` with a NULL SVE state on its first guest FP access, potentially crashing the hypervisor or host.

6. Manual review recommendation

MANUAL CHECK RECOMMENDED

Confirm whether an untrusted VMM or guest can reliably induce the initial SVE initialization failure and whether the resulting EL2 NULL dereference crashes the complete host or is contained to the affected VM.

7. Call-site confidence

Call-site confidence: medium

======================================================================
UPSTREAM PATCH SUMMARY
======================================================================

Patch: KVM: arm64: Do not clear VM-wide SVE feature on vCPU init failure [ Upstream
Commit: bd46b41f412fb81c50bac1a0929ec31320f3a807
Upstream URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=bd46b41f412fb81c50bac1a0929ec31320f3a807

Changed files:
  arch/arm64/kvm/hyp/nvhe/pkvm.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=bd46b41f412fb81c50bac1a0929ec31320f3a807

======================================================================
DETAILED REPORT METHODOLOGY
======================================================================

The original Linux kernel CVE announcement for CVE-2026-98637 can be found here:
https://lore.kernel.org/linux-cve-announce/?q=CVE-2026-98637

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:N/A:H
  The Best / paranoid CVSS vector: AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
  The Best / paranoid CVSS score: 5.5

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: 1

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).