A high-severity race condition in the Linux kernel KVM arm64 implementation creates a use-after-free vulnerability that could allow a local attacker to compromise system integrity and availability.

This is a race condition vulnerability within the vgic_v3_save_pending_tables function. An authenticated local user can exploit the lack of proper reference counting on LPI structures to trigger a use-after-free condition during concurrent guest memory access or cache injection operations.

The vulnerability carries a CVSS score of 8.8, reflecting its potential for complete system compromise. Because the flaw exists within the KVM hypervisor layer, a successful exploit could allow an attacker to escape a guest environment or crash the host kernel, leading to unauthorized access, data corruption, or significant service disruption.

Immediate Action: Update the Linux kernel to version 6.18.51, 7.2.5, or later to incorporate the necessary reference counting fixes in the vgic-v3 subsystem.

Proactive Monitoring: Monitor system logs for kernel panic events or unusual KVM-related errors that may indicate exploitation attempts.

Compensating Controls: Restrict access to virtual machine management interfaces and ensure that only trusted users have the capability to initiate or interact with KVM guest instances.

Public Exploit Available: No

Given the severity of this kernel-level vulnerability and the potential for privilege escalation, organizations running KVM on arm64 architectures should prioritize patching as part of their next scheduled maintenance cycle. Ensure that all production kernels are updated to the identified safe versions to eliminate the underlying race condition.