Buffer Overflow Vulnerability Detection Based on Static Analysis-assisted Symbolic Execution

Buffer overflow is a common exploitable high-risk vulnerability that can lead to arbitrary code execution, thus effective detection of buffer overflow vulnerability is of utmost importance for developers, enabling timely remendiation. Symbolic execution is a useful approach to detect it. In this paper, we propose VulDetection to improve the system resource utilization of symbolic execution through static analysis. Firstly, VulDetection uses static analysis techniques to extract the hotspot code segments in the program. Secondly, it uses the hotspot code segments to guide symbolic execution to explore only the vulnerability-related code parts. The experimental evaluation encompasses four vulnerable programs from the EDB vulnerability repository. The obtained experimental results demonstrate that VulDetection can improve the efficiency of vulnerability detection.