UnStrip: Unlocking Structurally Robust Stripped Functionality Logic Locking With ROBDD
Abstract
Logic locking (LL) protects integrated circuits (ICs) from intellectual property (IP) piracy and counterfeiting in a global supply chain model. Provably secure LL (PSLL) techniques involving stripped functionality LL (SFLL) are considered secure against the prominent Boolean SAT attack. Unfortunately, structural attacks have demonstrated vulnerabilities in PSLL techniques, rendering them impractical. Recently, structurally robust (SR)-SFLL has been proposed, proving improved security guarantees against known structural attacks on PSLL. In this work, we leverage reduced-order binary decision diagram (ROBDD) to mount a successful attack called UnStrip on 42 locked SR-SFLL circuits. UnStrip leverages the canonical representation of the locked circuits to identify the secret key. Initial evaluation of UnStrip on ISCAS-85 benchmarks locked with key sizes up to 120 bits yields promising results.