A Memory-Related Vulnerability Detection Approach Based on Vulnerability Features

Developing secure software systems is a major challenge in the software industry due to errors or weaknesses that bring vulnerabilities to the software system. To address this challenge, researchers often use the source code features of vulnerabilities to improve vulnerability detection. Notwithstanding the success achieved by these techniques, the existing studies mainly focus on the conceptual description without an accurate definition of vulnerability features. In this study, we introduce a novel and efficient Memory-Related Vulnerability Detection Approach using Vulnerability Features (MRVDAVF). Our framework uses three distinct strategies to improve vulnerability detection. In the first stage, we introduce an improved Control Flow Graph (CFG) and Pointer-related Control Flow Graph (PCFG) to describe the features of some common vulnerabilities, including memory leak, double-free, and use-after-free. Afterward, two algorithms, namely Vulnerability Judging algorithm based on Vulnerability Feature (VJVF) and Feature Judging (FJ) algorithm, are employed to detect memory-related vulnerabilities. Finally, the proposed model is validated using three test cases obtained from Juliet Test Suite. The experimental results show that the proposed approach is feasible and effective.