The Linux kernel contains a logic flaw in the uncore device initialization routine, where the function uncore_device_to_die() returns -1 when all CPUs linked to a UBOX device are offline or when NUMA is disabled. This incorrect return value leads to the scanner breaking out of its loop and consequently skipping other UBOX devices. Additionally, the condition was incorrectly treated as an error, causing PMON units on Intel SPR and EMR platforms to be omitted from the runtime data structures. The consequence is that performance monitoring data is incomplete or misleading, but the flaw does not allow code execution, confidentiality compromise, or direct availability impact.

Any Linux kernel containing the perf/x86/intel/uncore submodule before the fix is affected. The CPE list shows that the vulnerability exists in the 7.0 release candidates (rc1 through rc7) and in the generic linux_kernel CPE, implying that earlier stable releases that include the same code are also affected. Kernel releases that incorporate the referenced commit series resolve the issue and are not impacted.

The CVSS score of 5.5 indicates moderate severity, while the EPSS score of <1% reflects a very low likelihood of exploitation. The vulnerability is not listed in CISA’s KEV catalog, and the attack vector is local, requiring privileged kernel access during boot or when a privileged process re‑initializes uncore topology. Because the flaw only affects performance monitoring logic and not core system functionality, the overall security risk is limited but can result in inaccurate monitoring data for administrators.