• DocumentCode
    235015
  • Title

    SymWalker: Symbolic Execution in Routines of Binary Code

  • Author

    Jinxin Ma ; Guowei Dong ; Puhan Zhang ; Tao Guo

  • Author_Institution
    China Inf. Technol. Security Evaluation Center, Beijing, China
  • fYear
    2014
  • fDate
    15-16 Nov. 2014
  • Firstpage
    694
  • Lastpage
    698
  • Abstract
    Detecting vulnerabilities in binary codes is one of the most difficult problems due to the lack of type information and symbols. We propose a novel tool to perform symbolic execution inside the routines of binary codes, providing easy static analysis for vulnerability detection. Compared with existing systems, our tool has four properties: first, it could work on binary codes without source codes, second, it employs the VEX language for program analysis, thus having no side effects, third, it could deliver high coverage by statically executing on control flow graphs of disassembly codes, fourth, two security property rules are summarized to detect the corresponding vulnerabilities, based on which a convenient interface is provided for developers to detecting vulnerabilities, such as buffer overflow, improper memory access, and etc. Experimental results on real software binary files show that our tool could efficiently detect different types of vulnerabilities.
  • Keywords
    binary codes; flow graphs; program diagnostics; programming languages; symbol manipulation; SymWalker; VEX language; binary code routines; binary code vulnerability detection; control flow graphs; disassembly codes; program analysis; security property rules; software binary files; source codes; static analysis; symbolic execution; Binary codes; Computer bugs; Computer languages; Computers; Registers; Security; Software; Symbolic execution; control flow analysis; security property; vulnerabilities;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computational Intelligence and Security (CIS), 2014 Tenth International Conference on
  • Conference_Location
    Kunming
  • Print_ISBN
    978-1-4799-7433-7
  • Type

    conf

  • DOI
    10.1109/CIS.2014.16
  • Filename
    7016986