PrimLock: Leveraging Linear Programming to Optimize Primitive Logic Locking
Faisal Rasheed, Xiao Tong Cui, Israr Ahmad · 计算机科学辑要 · 2025
This paper presents PrimLock, a logic locking framework that employs integer linear programming (ILP) to optimize key gate placement in integrated circuits (ICs), addressing vulnerabilities like IP theft and hardware Trojans in untrusted supply chains. Existing methods suffer from heuristic key placement, scalability limitations, and susceptibility to SAT attacks. PrimLock introduces a du-al-circuit ILP model that maximizes output corruption under incorrect keys while minimizing computational overhead. A standardized security metric quantifies resilience as the percentage of corrupted outputs, validated on the C17 benchmark. Results show 100% corruption in critical test cases, out-per-forming heuristic and SAT-resistant approaches. Constraint reduction tech-niques reduce computational complexity significantly, enabling scalability to industrial designs. This work establishes a mathematically rigorous founda-tion for logic locking, balancing provable security with practical efficiency.