Abstracting Strings for Model Checking of C Programs

Applied Sciences

Martina Olliaro¹  Petr Roc̆kai¹  Henrich Lauko¹  Agostino Cortesi² 

[1] Faculty of Informatics, Masaryk University, Botanickà 68A, 60200 Brno, Czech Republic;Scientific Campus, Ca’ Foscari University of Venice, Via Torino 155, Mestre, 30172 Venice, Italy;

关键词: string analysis;    model checking;    abstract interpretation;    abstract domain;   

DOI  :  10.3390/app10217853

来源: DOAJ

【 摘 要 】

Data type abstraction plays a crucial role in software verification. In this paper, we introduce a domain for abstracting strings in the C programming language, where strings are managed as null-terminated arrays of characters. The new domain M-String is parametrized on an index (bound) domain and a character domain. By means of these different constituent domains, M-Strings captures shape information on the array structure as well as value information on the characters occurring in the string. By tuning these two parameters, M-String can be easily tailored for specific verification tasks, balancing precision against complexity. The concrete and the abstract semantics of basic operations on strings are carefully formalized, and soundness proofs are fully detailed. Moreover, for a selection of functions contained in the standard C library, we provide the semantics for character access and update, enabling an automatic lifting of arbitrary string-manipulating code into our new domain. An implementation of abstract operations is provided within a tool that automatically lifts existing programs into the M-String domain along with an explicit-state model checker. The accuracy of the proposed domain is experimentally evaluated on real-case test programs, showing that M-String can efficiently detect real-world bugs as well as to prove that program does not contain them after they are fixed.

【 授权许可】

Unknown