• DocumentCode
    2177463
  • Title

    Accurate call graph extraction of programs with function pointers using type signatures

  • Author

    Atkinson, D.C.

  • Author_Institution
    Dept. of Comput. Eng., Santa Clara Univ., CA, USA
  • fYear
    2004
  • fDate
    30 Nov.-3 Dec. 2004
  • Firstpage
    326
  • Lastpage
    335
  • Abstract
    Software engineers need to understand programs in order to effectively maintain them. The call graph, which presents the calling relationships between functions, is a useful representation of a program that can aid understanding. For programs that do not use function pointers, the call graph can be extracted simply by parsing the program. However, for programs that use function pointers, call graph extraction is nontrivial. Many widely used C programs utilize function pointers for efficiency and ease of implementation. We present a technique called type signature filtering for improving call graph extraction in the presence of function pointers. Filtering can be accomplished in a single pass after pointer analysis is complete, making it reusable across different analyses. Our results show that for many programs our technique yields a call graph that is nearly identical to the true call graph, even if a naive pointer analysis is used.
  • Keywords
    data structures; functions; graph theory; reverse engineering; software maintenance; C program; function pointer; parsing; pointer analysis; program call graph extraction; software engineer; software maintenance; type signature filtering; Computer architecture; Costs; Design engineering; Documentation; Filtering; Maintenance engineering; NIST; Programming profession; Software maintenance; Software systems;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Software Engineering Conference, 2004. 11th Asia-Pacific
  • ISSN
    1530-1362
  • Print_ISBN
    0-7695-2245-9
  • Type

    conf

  • DOI
    10.1109/APSEC.2004.16
  • Filename
    1371935