In the Linux kernel, the following vulnerability has been resolved:bpf: Check rcu_read_lock_trace_held() in bpf_map_lookup_percpu_elem()bpf_map_lookup_percpu_elem() helper is also available for sleepable bpfprogram. When BPF JIT is disabled or under 32-bit host,bpf_map_lookup_percpu_elem() will not be inlined. Using it in asleepable bpf program will trigger the warning inbpf_map_lookup_percpu_elem(), because the bpf program only holdsrcu_read_lock_trace lock. Therefore, add the missed check.
No PoCs from references.
- https://github.com/w4zu/Debian_security