A race condition in the KVM ARM64 nested‑virtualization code can cause the kernel to dereference a null pointer when a nested guest performs write‑protect or unmap operations concurrently with tearing down the stage‑2 address space. The resulting kernel panic halts the host, providing a denial of service but no code execution. The CVE description does not state an attacker able to exploit this from outside; the attack requires control of a nested guest or compromise of a guest capable of triggering the race. Based on the description, it is inferred that the attacker would need privileged guest access and timely manipulation of page‑table operations.

Linux kernel builds that support KVM on ARM64 and enable nested virtualization are affected, provided they have not yet integrated the two commits that address the race (7270cc9157f4 and ec14c272408a). The vendor identifier appears as Linux, and no specific kernel version range is listed, so all pre‑patched kernel releases with these KVM paths, regardless of distribution, are at risk.

The CVSS score of 4.7 represents moderate severity, while the EPSS score of less than 1 % indicates a low probability of real‑world exploitation. It is not listed in the CISA KEV catalog, suggesting no documented exploitation. An attacker would need to control a nested guest and perform precise page‑table manipulation to trigger the race, making the overall risk moderate yet contingent on a sophisticated attacker who can reach the nested environment.