The vulnerability exists due to improper state isolation inside the multi-threaded flow tracking worker queues. High-velocity concurrent network streams can cause worker threads to race when updating global flow indices, resulting in unexpected pointer dereferences and kernel-level thread crashes.