Predicate Abstraction Refinement in Thread-Modular Analysis

Veronika Pavlovna Rudenchik, Pavel Sergeevich Andrianov
{"title":"Predicate Abstraction Refinement in Thread-Modular Analysis","authors":"Veronika Pavlovna Rudenchik, Pavel Sergeevich Andrianov","doi":"10.15514/ispras-2023-35(3)-14","DOIUrl":null,"url":null,"abstract":"Thread-modular approach over predicate abstraction is an efficient technique for software verification of complicated real-world source code. One of the main problems in the technique is a predicate abstraction refinement in a multithreaded case. A default predicate refiner considers only a path related to one thread, and does not refine the thread-modular environment. For instance, if we have applied an effect from the second thread to the current one, then the path in the second thread to the applied effect is not refined. Our goal was to develop a more precise refinement procedure, reusing a default predicate refiner to refine both: a path in a current thread and a path to an effect in the environment. The idea is to construct a joined boolean formula from these two paths. Since some variables may be common, a key challenge is to correctly rename and equate variables in two parts of the formula to accurately represent the way threads interact. It is essential to get reliable predicates that can potentially prove spuriousness of the path. The proposed approach is implemented on top of CPAchecker framework. It is evaluated on standard SV-COMP benchmark set, and the results show some benefit. Evaluation on the real-world software does not demonstrate significant accuracy increase, as the described flaw of predicate refinement is not the only reason of false positive results. While the proposed approach can successfully prove some specific paths to be spurious, it is not enough to fully prove correctness of some programs. However, the approach has further potential for improvements.","PeriodicalId":33459,"journal":{"name":"Trudy Instituta sistemnogo programmirovaniia RAN","volume":"57 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2023-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Trudy Instituta sistemnogo programmirovaniia RAN","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.15514/ispras-2023-35(3)-14","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

Abstract

Thread-modular approach over predicate abstraction is an efficient technique for software verification of complicated real-world source code. One of the main problems in the technique is a predicate abstraction refinement in a multithreaded case. A default predicate refiner considers only a path related to one thread, and does not refine the thread-modular environment. For instance, if we have applied an effect from the second thread to the current one, then the path in the second thread to the applied effect is not refined. Our goal was to develop a more precise refinement procedure, reusing a default predicate refiner to refine both: a path in a current thread and a path to an effect in the environment. The idea is to construct a joined boolean formula from these two paths. Since some variables may be common, a key challenge is to correctly rename and equate variables in two parts of the formula to accurately represent the way threads interact. It is essential to get reliable predicates that can potentially prove spuriousness of the path. The proposed approach is implemented on top of CPAchecker framework. It is evaluated on standard SV-COMP benchmark set, and the results show some benefit. Evaluation on the real-world software does not demonstrate significant accuracy increase, as the described flaw of predicate refinement is not the only reason of false positive results. While the proposed approach can successfully prove some specific paths to be spurious, it is not enough to fully prove correctness of some programs. However, the approach has further potential for improvements.
线程模块化分析中的谓词抽象细化
线程模块化方法优于谓词抽象,是一种对复杂的实际源代码进行软件验证的有效技术。该技术的主要问题之一是在多线程情况下对谓词抽象进行细化。默认谓词细化器只考虑与一个线程相关的路径,而不细化线程模块化环境。例如,如果我们从第二个线程应用了一个效果到当前线程,那么第二个线程中到应用效果的路径不会被细化。我们的目标是开发一个更精确的细化过程,重用默认谓词细化器来细化:当前线程中的路径和环境中某个效果的路径。其思想是从这两条路径构造一个连接的布尔公式。由于一些变量可能是通用的,因此一个关键的挑战是正确地重命名和等同公式中两个部分的变量,以准确地表示线程交互的方式。获得可靠的谓词是至关重要的,这些谓词可以潜在地证明路径的虚假性。该方法是在CPAchecker框架之上实现的。在标准SV-COMP基准集上进行了评估,结果显示出一定的效益。对真实软件的评估没有显示出显著的准确性提高,因为所描述的谓词细化缺陷并不是假阳性结果的唯一原因。虽然所提出的方法可以成功地证明某些特定路径是假的,但还不足以完全证明某些程序的正确性。然而,这种方法还有进一步改进的潜力。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
自引率
0.00%
发文量
18
审稿时长
4 weeks
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信