• DocumentCode
    685502
  • Title

    Refining the Pointer Analysis by Exploiting Constraints on the CFL-Paths

  • Author

    Xiaoshan Sun ; Liang Cheng ; Yang Zhang ; Yi Deng ; Jingbiao Hou

  • Author_Institution
    Trusted Comput. & Inf. Assurance Lab., Inst. of Software, Beijing, China
  • Volume
    1
  • fYear
    2013
  • fDate
    2-5 Dec. 2013
  • Firstpage
    17
  • Lastpage
    24
  • Abstract
    The pointer analysis finds out what a pointer points to in a program. This analysis is useful in static program analysis, bug detection, source navigations and so on. We propose an approach to generate new constraints for one points-to and we use the constraints to refine it. The constraints can be used to refine other program analysis as well. This new constraints are based on the sequencings among the edges on a CFL-path. We propose algorithms to compute the sequencing relations for a CFL-path. Based on the news constraints, we propose a method to check the context-sensitivity of a points to by checking whether there is a trace of the C program satisfying the sequencing relations. We implement our approach in a proto-type tool called TCPA to evaluate the efficiency of our approach. Our approach can be useful in static program analysis, source code bug detection and source code understanding.
  • Keywords
    C language; program diagnostics; reachability analysis; C program; CFL-Path; TCPA; context-sensitivity; pointer analysis; source code bug detection; source code understanding; static program analysis; Accuracy; Algorithm design and analysis; Computational modeling; Context modeling; Grammar; Motorcycles; Sequential analysis; CFL-reachability problem; constraints solving; pointer analysis; program analysis;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Software Engineering Conference (APSEC), 2013 20th Asia-Pacific
  • Conference_Location
    Bangkok
  • ISSN
    1530-1362
  • Print_ISBN
    978-1-4799-2143-0
  • Type

    conf

  • DOI
    10.1109/APSEC.2013.14
  • Filename
    6805385