Machine-Code Verification for Multiple Architectures - An Application of Decompilation into Logic

Abstract

Abstract-Realistic formal specifications of machine languages for commercial processors consist of thousands of lines of definitions. Current methods support trustworthy proofs of the correctness of programs for one such specification. However, these methods provide little or no support for reusing proofs of the same algorithm implemented in different machine languages. We describe an approach, based on proof-producing decompilation, which both makes machine-code verification tractable and supports proof reuse between different languages. We briefly present examples based on detailed models of machine code for ARM, PowerPC and x86. The theories and tools have been implemented in the HOL4 system.

Verified LISP Implementations on ARM, x86 and PowerPC

“…• The correctness of each code snippets is expressed as a machine-code Hoare triple [15]: { pre * pc p } p : code { post * pc (p + exit) }.…”

“…• For cons and equal we used previously developed proof automation [15], which allows for proof reuse in between different machine languages.…”

“…The correctness of the above implementations of car and plus is expressed formally by the two ARM Hoare triples [15] below. Here…”

“…Since then a decompiler [15] and compiler [16] have been developed to aid this project, which produced in total some 4,580 lines of proof automation and 16,130 lines of interactive proofs and definitions, excluding the definitions of the instruction set models [5,9,18]. Running through all of the proofs takes approximately 2.5 hours in HOL4 using PolyML.…”

“…This proof makes use of the following five proof rules derived from the definition of our machine-code Hoare triple, developed in previous work [15]. Formal definitions and detailed explanations are given in the first author's PhD thesis [14].…”