DSLR–: A low-overhead data structure layout randomization for defending data-oriented programming.

This article focuses on mitigating Data-Oriented Programming (DOP) attacks, which manipulate non-control data to bypass traditional control-flow defenses. Through static analysis of 20 real-world applications, the authors identify that only about 3% of program data are targeted by DOP gadgets and dispatchers. They propose DSLR–, a low-overhead Data Structure Layout Randomization method that selectively randomizes the memory layout of these target data, disrupting attackers' ability to locate and manipulate them. Experimental results show that DSLR– effectively mitigates DOP attacks while reducing performance overhead by 71.2% and memory overhead by 82.5% compared to existing DSLR techniques. The study also introduces a semi-automated tool to assist in identifying dispatchers and chaining operations on target data, advancing the understanding and defense against DOP attacks.