DAMAS: Control-Data Isolation at Runtime through Dynamic Binary Modification

Memory corruption attacks have been a major issue in software security for over two decades and are still one of the most dangerous and widespread types of attacks nowadays. Among these attacks, control-flow hijack attacks are the most popular and powerful, enabling the attacker to execute arbitrary code inside the target process. Many approaches have been developed to mitigate such attacks and to prevent them from happening. One of these approaches is the Control-Data Isolation (CDI) that tries to prevent such attacks by removing their trigger from the code, namely indirect branches. This approach has been implemented as a compiler pass that replaces every indirect branches in the program with a table that leads the control-flow to direct hard-written branches. The drawback of this approach is that it needs the recompilation of the program. In this paper we present an approach and its implementation, DAMAS, a framework capable of deploying protections on a running software and use runtime information to optimize them during the process execution. We implemented a coarse-grain CDI protection using our framework and evaluated its impact on performance.