Technical impact
A bug in the Linux kernel’s KVM module mishandles the SS bit when interpreting nested page-fault error codes from Hyper-V guests. This incorrect bit handling causes an array-index-out-of-bounds check inside permission_fault(), which can lead to out-of-bounds memory writes and kernel memory corruption. The flaw could potentially allow a malicious guest to elevate privileges on the host or to crash the kernel, resulting in denial of service.
Affected Systems
Linux distributions that ship the Linux kernel with KVM support and the nSVM path (including kernels 6.18 and newer) are affected when running Hyper-V guests on Windows 11. The issue originates in the KVM’s mmu handling code and is limited to hosts that execute the faulty commit sequence; kernels that have applied the subsequent fix are not impacted.
Risk and Exploitability
The CVSS score of 4.7 reflects moderate severity and an EPSS below 1% indicates low likelihood of real-world exploitation. The vulnerability requires a Hyper-V guest to trigger a nested page fault with an unexpected SS bit; thus the attack surface is narrow and depends on a specific host configuration. Successful exploitation could lead to kernel corruption and privilege escalation, but the narrow trigger conditions reduce the probability of widespread impact.