Scholarlink

<p>This article presents a novel method for detecting copied code fragments called clones, which is then utilized to identify known common vulnerabilities and exposures copies. The proposed method is versatile and applicable to both source and binary code. It overcomes the limitations of existing tools that typically focus on detecting entire function clones and specializing in either source or binary code, but not both. The method outputs all the detected clones by comparing the provided code fragment against the target project. It employs program dependence graphs—a data structure that unifies data and control dependencies for the function to achieve high accuracy. Experimental evaluations of real-world projects and benchmarks demonstrate the high precision of the proposed method. Furthermore, we successfully applied this method to detect clones of known common vulnerabilities and exposures in source code and uncovered vulnerabilities in actual software. The detected vulnerabilities were confirmed by the community, validating the effectiveness of the proposed method.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Accurate code fragment clone detection and its application in identifying known CVE clones

  • 期刊论文

  • 发表日期：

    2025年1月5日

查看全文
  • Mariam Arutunian,
  • Sevak Sargsyan,
  • Hripsime Hovhannisyan,
  • Garnik Khroyan,
  • Albert Mkrtchyan,
  • Hovhannes Movsisyan,
  • Arutyun Avetisyan,
  • Hayk Aslanyan

摘要

This article presents a novel method for detecting copied code fragments called clones, which is then utilized to identify known common vulnerabilities and exposures copies. The proposed method is versatile and applicable to both source and binary code. It overcomes the limitations of existing tools that typically focus on detecting entire function clones and specializing in either source or binary code, but not both. The method outputs all the detected clones by comparing the provided code fragment against the target project. It employs program dependence graphs—a data structure that unifies data and control dependencies for the function to achieve high accuracy. Experimental evaluations of real-world projects and benchmarks demonstrate the high precision of the proposed method. Furthermore, we successfully applied this method to detect clones of known common vulnerabilities and exposures in source code and uncovered vulnerabilities in actual software. The detected vulnerabilities were confirmed by the community, validating the effectiveness of the proposed method.