RESAA: A Removal and Structural Analysis Attack Against Compound Logic Locking

TL;DR

RESAA is a versatile framework that analyzes and attacks compound logic locking in integrated circuits, revealing vulnerabilities and aiding in security evaluation.

Contribution

It introduces a novel, LL technique-agnostic framework, RESAA, for classifying, analyzing, and attacking CLL-based IC security mechanisms.

Findings

RESAA achieves up to 92.6% accuracy on complex benchmarks.

All studied CLL variants are vulnerable to RESAA attacks.

RESAA effectively identifies critical gates and distinguishes LL segments.

Abstract

The semiconductor industry's paradigm shift towards fabless integrated circuit (IC) manufacturing has introduced security threats, including piracy, counterfeiting, hardware Trojans, and overproduction. In response to these challenges, various countermeasures, including Logic locking (LL), have been proposed to protect designs and mitigate security risks. LL is likely the most researched form of intellectual property (IP) protection for ICs. A significant advance has been made with the introduction of compound logic locking (CLL), where two LL techniques are concurrently utilized for improved resiliency against attacks. However, the vulnerabilities of LL techniques, particularly CLL, need to be explored further. This paper presents a novel framework, RESAA, designed to classify CLL-locked designs, identify critical gates, and execute various attacks to uncover secret keys. RESAA is agnostic to specific LL techniques, offering comprehensive insights into CLL's security scenarios. Experimental results demonstrate RESAA's efficacy in identifying critical gates, distinguishing segments corresponding to different LL techniques, and determining associated keys based on different threat models. In particular, for the oracle-less threat model, RESAA can achieve up to 92.6% accuracy on a relatively complex ITC'99 benchmark circuit. The results reported in this paper emphasize the significance of evaluation and thoughtful selection of LL techniques, as all studied CLL variants demonstrated vulnerability to our framework. RESAA is also open-sourced for the community at large.