MBA-Blast Code and Dataset for USENIX Security'21

August 6, 2021 ยท View on GitHub

MBA-Blast is a tool for simplification of MBA expressions.

Prerequisites:

Python 3.6

  1. Z3 solver: pip3 install z3-solver
  2. sympy: pip3 install sympy

Python 2.7

  1. Z3 solver: pip install z3-solver
  2. sympy: pip install sympy
  3. ast: pip install ast
  4. astunparse: pip install astparse

Arybo

Download it and install it by the link.

SSPAM

Download it and install it by the link.

Syntia

The source code is from the link, but we have download and refine it in the related folder "syntia".

Structure

MBA-Blast's code is structured in following parts: dataset, simplifying MBA expression in the dataset, peer simplification tools.

dataset

The files storing MBA expression.

  1. Dataset 1 is dataset1.txt
  2. Dataset 2 is dataset2_xbit.txt

mba-simplifier

Simplifying the mba expressions by the method describing in the paper. Functions to analyze and manipulate MBA expressions are in "tools" folder.

How to use

  1. Run MBA-Blast on Dataset1: make mba-simplify-1
  2. Run MBA-Blast on 8-bit dataset: make mba-simplify-8
  3. Run MBA-Blast on 16-bit dataset: make mba-simplify-16
  4. Run MBA-Blast on 32-bit dataset: make mba-simplify-32
  5. Run MBA-Blast on 64-bit dataset: make mba-simplify-64

Peer tools to do simplification:

Arybo

Run Arybo on Dataset1: make arybo-mba

SSPAM

Run SSPAM on Dataset1: make sspam-mba

Syntia

Run Syntia on Dataset1: program-synthesis-64, json-transformation-64, syntia-evaluation-64. Note that the main module mcts_synthesis is from the open source code of paper -- Syntia: Synthesizing the Semantics of Obfuscated Code.