基于分段执行和摘要的动态程序分析

Dynamic program analysis based on partitioning and summarization

传统的动态程序分析需探索程序运行的所有可能路径。然而,实际的程序通常有大量甚至无限多的路径。由于不能完全探索所有路径,对程序的准确分析只限于一些简单的情形。具体符号测试方法结合符号分析和通过运行程序获得变量的具体数值。它提高了传统方法的适用性,但同样受到路径数目的限制。为解决这一问题,提出基于分段执行程序和对程序段落自动生成摘要的手段。该手段扩展了具体符号测试,提高其效率,并使之可应用于一些具有无限路径的程序。

Traditional dynamic program analysis relies on exploring all possible paths of program execution. In reality, however , programs have a large number of, or even infinitely many, paths. Since it is not feasible to explore all paths, precise analysis is limited to simple cases. Concolic testing method combines symbolic analysis and concrete values of variables obtained by executing programs. Though it improves applicability of traditional approaches, concolic testing is still limited by the number of explorable paths. To address this problem, this paper proposed a method here based on executing partitions of a program and automatically generating summaries for those partitions. This method extends concolic testing, improving its perfor-mance and making it applicable to certain programs with unbounded paths.