A binary-oriented fuzz testing method and system

Abstract

The invention relates to the field of binary-oriented fuzz testing in the field of software system security detection and vulnerability mining methods, and aims to provide a binary-oriented fuzz testing method and system. This method adopts a lightweight static analysis method and binary source code instrumentation technology. The specific content includes: first, through static analysis, extract the comparison instruction information that hinders the fuzz test from penetrating into the program, and then insert the binary file based on the obtained information. The stub is used to obtain the specific value of the operand in the process of fuzzing, and then mutate the input file byte by byte, and then speculate on the dependency relationship between the input file and the comparison instruction, and finally according to the dependency relationship and the comparison progress information. Guided mutation of files. The system does not depend on program source code, reduces the blindness of mutation in fuzz testing, saves computing resources, and has higher versatility.

Description

technical field [0001] The invention relates to the field of software system safety detection, in particular to the field of binary-oriented fuzzy testing in vulnerability mining technology. Background technique [0002] With the widespread use of computers and the rapid development of computer networks, people's lives and even various industries such as financial securities, national defense technology, medical and health care, etc. are inseparable from various computer software, and the number of software is increasing day by day. The hidden dangers of security vulnerabilities have also become more prominent. Hidden dangers of software security vulnerabilities provide opportunities for attackers, and attacks by attackers using software vulnerabilities have emerged in an endless stream in recent years. Therefore, discovering and patching vulnerabilities as early as possible has become a hot issue in the field of software security research. [0003] In the field of vulnera...

AI Technical Summary

Problems solved by technology

The current popular fuzz testing tools use genetic algorithms and use code coverage as feedback to screen offspring. Compared with traditional fuzz testing, this type of fuzz testing tool improves the efficiency of fuzz testing, but this type of tool still has certain limitations. If you don’t know which bytes in the input file are worth mutating and how to mutate the bytes in the input file, it will often take a lot of time to mutate some insignificant bytes, even if sometimes the mutation occurs in a key position. However, since the tool does not know how to mutate, it still cannot effectively bypass the program check

These problems make it difficult for the input file to trigger the complex logic code part of the program, so it is difficult to find potential loopholes deep in the program

Images