The Linux kernel suffered a race condition in the page table dump (ptdump) workflow. During a ptdump operation, the kernel can free page tables while another MMA‑mm is reading them if the init_mm lock is not held. This could allow a malicious process to read or manipulate freed memory, leading to kernel corruption, denial of service or privilege escalation. The underlying weakness is a synchronization failure between concurrent kernel operations.

Vendors: Linux. The affected component is the Linux kernel, including architectures x86 and arm64 for which ptdump over non‑init_mm memory tables is possible. The advisory does not specify a particular major version, but the fix is present in the latest stable kernel releases containing the highlighted commits.

The advisory lists no EPSS score; exploitation likelihood cannot be quantified. The CVSS score is not provided, so the exact severity is unknown, but a race that can corrupt kernel memory is typically considered high risk. The vulnerability is not listed in CISA’s KEV catalog. Attackers would need local kernel access to trigger the race, typically by running privileged code to initiate a ptdump while another process frees page tables. No public exploitation has been reported and the fix applies a proper lock to prevent the race, eliminating the flaw.