The vulnerability occurs in the arm64 kernel helper function huge_ptep_get, which can be invoked with a virtual address that is not aligned to the base of a contpmd or contpte hugetlb folio. When the address is misaligned, the function may mistakenly treat a continuous PTE block as a PMD block, collecting incorrect bits and iterating past the end of the page table entries. This can result in dereferencing a kernel‑space pointer beyond the linear map, accessing memory that the kernel does not own and ultimately causing a kernel panic. The impact is kernel memory corruption followed by a denial‑of‑service event.

This issue affects all arm64 Linux kernels that include the vulnerable version of huge_ptep_get and have not yet incorporated the commit that aligns the PMDP pointer before comparison. It is applicable to any system running a kernel that supports hugetlb page mappings and where functions such as pagemap_hugetlb_range or page_mapped_in_vma may be invoked.

No public exploits have been reported and the exploitation likelihood is presently unknown; EPSS data is not available. The vulnerability is not listed in the CISA KEV catalog, indicating no widely known exploitation. Based on the description, it is inferred that the likely attack vector requires an attacker to invoke a local system call that triggers huge_ptep_get with a crafted unaligned address; such a call would need to be performed by code running with sufficient privileges to access the relevant kernel functions. Consequently, the risk is primarily a local denial‑of‑service threat rather than a remote code execution vector, but it remains important to mitigate.