Hacking the Control Flow error detection mechanism

Many techniques have been proposed in literature to cope with transient, permanent and malicious faults in computing systems. Among these techniques for reliability improvement and fault tolerance, Control Flow Checking allows covering any fault affecting the part of the storing elements containing the executable program, as well as all the hardware components handling the program itself and its flow. In [1] the authors proposed a low-overhead solution implementing hardware based control flow monitoring technique. They suggested that control flow error detection could be also used as a solution for enhancing the security of a computing system, preventing the insertion of malicious code in an application. In this paper we present a technique to map a malicious program into another one without structure violation and thus bypassing the control flow detection method.