Profile-guided code identification and hardening using return oriented programming

Abstract

Protecting code from a security breach needs a systematic defense. The most promising solution in this task is to protect the control flow of the code. A majority of the existing approaches for protecting the code using control flow information suffer from high performance overhead. In this paper, we propose a method for assisted discovery of profile-guided techniques for the cold-code identification and low-overhead software security hardening. Our method proceeds with identifying cold-code and hot-code by performing static and dynamic analysis on the input code. Furthermore, develops heuristics to identify security-sensitive code. The outcome of this research is the identification of a minimum number of security sensitive points with minimal overhead. This research develops a common framework for profile-guided code self-verification based on predetermined security and performance requirements. We use Return Oriented Programming (ROP) which is used mostly by adversaries for the defensive purpose to build self-verifying ROP chain. The ROP chain inserted into the code diversifies and obfuscates execution of the program hence, making it harder to breach.