A static memory leak detection method for binary programs
Zhao Zhen · 2009
In order to conduct memory leak analysis upon executable programs,a binary-level memory leak analysis framework MLAB is designed and implemented.First,MLAB recovers control flow and data flow information from binary programs.Then a control flow automata is constructed based on the recovered control flow graph.After that,a model checking based algorithm is applied upon the automata to perform memory leak analysis.Several sample programs are further analyzed to show in detail how the method works.Experimental results on the benchmark suite MiBench shows the effectiveness of the method.