The Linux kernel vulnerability CVE-2026-64352 concerns the BPF subsystem's handling of LPM trie map access from sleepable BPF programs. The trie_lookup_elem() function used rcu_dereference_check() with an annotation that did not recognize the RCU locking context of sleepable BPF programs, causing lockdep warnings on debug kernels. Similarly, trie_update_elem() and trie_delete_elem() used rcu_dereference() with assumptions about RCU locks that did not hold for sleepable BPF programs. While no use-after-free occurs due to serialization by Tasks Trace RCU, the incorrect annotations cause lockdep warnings and console spam in debug builds. The fix changes the RCU annotations to bpf_rcu_lock_held() for lookups and rcu_dereference_protected() for updates/deletes, aligning with actual locking and eliminating the warnings.

This vulnerability does not cause memory corruption or use-after-free conditions but triggers lockdep warnings and debug console spam on kernels compiled with debugging enabled when sleepable BPF programs access LPM tries. There are no known exploits in the wild. The issue affects kernel debugging and may hinder developers or administrators monitoring debug kernel logs.

A fix for this issue has been implemented in the Linux kernel by correcting RCU locking annotations in the BPF LPM trie code. Users running debug kernels should update to a kernel version that includes this fix to eliminate lockdep warnings and console spam. For non-debug kernels, this issue does not cause functional problems. Patch status is not explicitly stated; check the vendor advisory or Linux kernel mailing lists for the specific fixed kernel versions.