A Control Flow Graph Reconstruction Method from Binaries Based on XML

The first step of decompilation and other reverse analysis for binary codes is constructing control flow graph of program. Due to the loss of structure in compiled code, extracting a control flow graph from an executable is not simply a matter. The usual way which scans the branch instructions in the disassembly file and creates the CFG can not suitable for multiply architecture and file format. This paper analyzes the structure of assembly program and designs a data information description language based on XML-reverse meta language, which is to describe structured disassembling information and then brings forward the control flow graph reconstruction algorithm. This method abstracts disassembly representations and supports multiple architectures. The testing result shows that this method can reconstruct control flow graph of binaries effectively.