---
id: "CVE-2026-98637"
kind: "vulnerability"
aliases: ["CVE-2026-98637"]
title: "Linux KVM on arm64: SVE init failure leaves NULL state for guests"
severity: "unknown"
cvss: null
epss: null
exploited: false
status: "open"
category: "runtime"
products: []
ecosystems: []
tags: ["arm64", "kvm", "linux-kernel", "pkvm", "runtime"]
not_affected_if: []
published_at: "2026-10-18T16:17:16+00:00"
first_seen_at: "2026-10-23T21:57:16+00:00"
updated_at: "2026-10-25T05:20:58+00:00"
sources: [{"source": "nvd", "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-98637", "first_seen_at": "2026-10-23T21:57:16+00:00"}]
---

# Linux KVM on arm64: SVE init failure leaves NULL state for guests

When SVE setup fails for one vCPU in a pKVM arm64 guest, later vCPUs on the same VM skip SVE setup and are registered with a NULL SVE state. The guest's first floating-point access then passes that NULL pointer into sve_load_state(), which can crash the host kernel.

_Summary, check and fix written by a model from the sources; labelled as such._

## How to check

- unknown

## Fix

- Update the Linux kernel to a release containing the upstream commit that returns the error from pkvm_vcpu_init_sve() without clearing KVM_ARM_VCPU_SVE (version unknown from sources)

## Timeline

- 2026-10-23T21:57:16+00:00 nvd https://nvd.nist.gov/vuln/detail/CVE-2026-98637

## From the sources

In the Linux kernel, the following vulnerability has been resolved:

KVM: arm64: Do not clear VM-wide SVE feature on vCPU init failure

pkvm_vcpu_init_sve() clears KVM_ARM_VCPU_SVE in kvm->arch.vcpu_features
when it fails, but vcpu_has_sve() tests KVM_ARCH_FLAG_GUEST_HAS_SVE,
which is left set. Later vCPUs on that VM then skip the SVE setup and
register with a NULL sve_state, which the guest's first FP access hands
to sve_load_state().

Return the error without touching vcpu_features.

---

_Free to read at a human pace. An API with keys and an MCP server are in preparation; early access: hello@infrastack.ai_