A methodology for machine language decompilation

Abstract

Machine language decompilation is the translation of machine (assembly) language instruction sequences into statements in a high-level-algebraic language such as PL/1. This process can be viewed as the inverse of compilation. Decompilation can be used as an aid for program conversion and program documentation. A general methodology for decompilation that is independent of a particular source and target language is presented. The basic approach is to map the source machine language to a high-level representation, which is relatively machine and language independent, and then translate to the chosen target language. An experimental decompiler was implemented to translate Knuth's MIXAL assembly language into PL/1.