MemSpate: Memory Usage Protocol Guided Fuzzing


연구 분야: Strategies



학회: International Conference on Formal Engineering Methods


초록

Memory safety vulnerabilities are high-risk and common vulnerabilities in software testing, often leading to a series of system errors. Fuzz testing is widely recognized as one of the most effective methods for detecting vulnerabilities, including memory safety ones. However, current fuzzing solutions typically only partially address memory usage, limiting their ability to detect memory safety vulnerabilities. In this paper, we introduce MemSpate, a dedicated fuzzer designed to detect memory safety vulnerabilities. Utilizing a more comprehensive memory usage protocol, MemSpate identifies the memory operation sequences that may violate the protocol and estimates the overall memory consumption to exceed an acceptable limit. It then monitors the coverage of these operation sequences and tracks the maximum memory consumption, both of which are used as a new feedback mechanism to guide the fuzzing process. We evaluated MemSpate on 12 real-world open-source programs and compared its performance with 5 state-of-the-art fuzzers. The results demonstrate that MemSpate surpasses all other fuzzers in terms of discovering memory safety vulnerabilities. Furthermore, our experiments have led to the discovery of 4 previously unknown vulnerabilities.


Author Profile
Jiacheng Jiang

College of Computer Science and Software Engineering Shenzhen University Shenzhen China

Andorra
Author Profile
Zhiyuan Fu

College of Computer Science and Software Engineering Shenzhen University Shenzhen China

Andorra
Author Profile
Cheng Wen

Guangzhou Institute of Technology Xidian University Xi’An China

China

📄 논문 정보

발행 연도 2024년
인용수 0
출판 국가 Andorra, China
사이트 Springer
좋아요 수 0

연관 논문 목록 (234건)