EtWExplorer: Multi-Priority Scheduling Path Exploration Technology Based on Abstract Syntax Tree Analysis

Author:

He XingluORCID,Wang Pengfei,Lu Kai,Zhou XuORCID

Abstract

Symbolic execution is well known as a dynamic vulnerability discovery technique. Its greatest advantage is the capability to analyze the execution information of the program and to explore the path in the program deterministically. This is a more accurate way to determine if there are vulnerabilities in a program than randomized testing by fuzzing. In addition, symbolic execution does not suffer from the problem of decreasing the capability to discover new paths as more paths are discovered, similar to that caused by random-based fuzzing. However, the reason why symbolic execution is not widely used in vulnerability discovery is mainly due to the state space explosion in the program. The state space explosion severely affects the applicability of symbolic execution. To further improve the applicability of symbolic execution, this paper proposes a path exploration technology based on abstract syntax tree analysis. With the distance between the expression generated by the symbolic execution of the repeat location and the “unsatisfiable” condition of the “unsat” state, we can perform multi-priority scheduling for the repeat location state, thus mitigating the impact of the state space explosion on path exploration. We proposed and implemented EtWExplorer, a multi-priority scheduling technique based on abstract syntax tree analysis. With this technique, we can significantly improve the capability of symbolic execution to discover unknown paths even in state space exploration. Experiments show that EtWExplorer introduces a performance overhead of 72% in the worst case and can improve performance by 294% in the best case. EtWExplorer has a 95% improvement in state space explosion mitigation capability and a 199% to 983% improvement in the path exploration capability of block coverage and a 181% to 1047% improvement in the path exploration capability of edge coverage when facing programs that cause a state space explosion.

Funder

National High-level Personnel for Defense Technology Program

National Natural Science Foundation of China

Publisher

MDPI AG

Subject

Fluid Flow and Transfer Processes,Computer Science Applications,Process Chemistry and Technology,General Engineering,Instrumentation,General Materials Science

Reference17 articles.

1. Driller: Augmenting Fuzzing through Selective Symbolic Execution;Stephens;Proceedings of the NDSS,2016

2. {QSYM}: A practical concolic execution engine tailored for hybrid fuzzing;Yun;Proceedings of the 27th USENIX Security Symposium (USENIX Security 18),2018

3. Savior: Towards bug-driven hybrid testing;Chen;Proceedings of the 2020 IEEE Symposium on Security and Privacy (SP),2020

4. HFL: Hybrid Fuzzing on the Linux Kernel;Kim;Proceedings of the NDSS,2020

5. Pangolin: Incremental hybrid fuzzing with polyhedral path abstraction;Huang;Proceedings of the 2020 IEEE Symposium on Security and Privacy (SP),2020

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3