Conditional weighted pushdown systems and applications

  • Authors:

  • Xin Li;Mizuhito Ogawa

  • Affiliations:

  • Japan Advanced Institute of Science and Technology, Nomi, Japan;Japan Advanced Institute of Science and Technology, Nomi, Japan

  • Venue:
  • Proceedings of the 2010 ACM SIGPLAN workshop on Partial evaluation and program manipulation
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Pushdown systems are well understood as abstract models of programs with (recursive) procedures. Reps et al. recently extended pushdown systems into weighted pushdown systems, which serve as a generalized framework for solving certain kinds of meet-over-all-path problems in program analysis. In this paper, we extend weighted pushdown systems to conditional weighted pushdown systems, by further specifying conditions under which a pushdown transition rule can be applied, and show that model checking problems on conditional weighted pushdown systems can be reduced to those on weighted pushdown systems. There are wider applications of conditional weighted pushdown systems when analyzing programs with objected-oriented features, for which weighted pushdown systems is not precise enough under a direct application. As an example, we lift a stacking-based points-to analysis for Java designed in the framework of weighted pushdown systems to a more precise counterpart in the framework of conditional weighted pushdown systems. In addition to the fundamental context-sensitivity in terms of valid paths, the lifted points-to analysis algorithm further enjoys context-sensitivity with respect to objected-oriented features, including call graph construction, heap abstraction, and heap access. These context-sensitive properties are shown to be crucial to the analysis precision in practice.