A Novel Program Analysis Method Based on Execution Path Correlation

Abstract

Path-sensitive analysis is an important feature of static program analysis that can assist to performance the software testing and ensure the software quality. Currently most analysis models are based on symbolic execution analysis. Our approach faces on some challenges like how to deal with the "path explosion" problem due to the number of paths growing exponentially with the number of program branches and the inline process of inter-procedural callings. In this paper, we present a method called Path Correlation Analysis and an analytical model, Solving Time Symbolic Execution, which, working together, can alleviate the redundant analysis of the common nodes between different paths in conventional path-oriented analysis. We also give the description of the corresponding algorithms in the paper and prove that compared with the conventional methods, this method can efficiently reduce the time complexity, easing the effect of the "path explosion" problem.