A Coverage-Oriented Fuzzing Test Method for Embedded Firmware

When it comes to firmware vulnerability detection in embedded systems, traditional static analysis approaches struggle to capture deep-level program execution logic, resulting in unsatisfactory effectiveness. Based on common fuzzer, this paper proposes a coverage-oriented dynamic fuzzing test method called dynamic instrumentation marking coverage-based fuzzer to support simulation and test for ARM and MIPS architecture-based IoT embedded device firmware. Furthermore, two experiments were designed. One proved the fuzzer can effectively discover vulnerabilities in firmware; the other one demonstrated that compared to the AFL fuzzer, the method we proposed can generate more test samples, execute most of them and trigger more crash records.