Detected object program feature extraction method for vulnerability detection

Abstract

The invention relates to a detected object program feature extraction method for vulnerability detection, which comprises the following steps of: 1) loading vulnerability, namely establishing a data structure of a vulnerability model consisting of a set of states of a vulnerability state machine and a set of operation of the vulnerability state machine, wherein the states represent the state of the state machine and each state has state number, type and feature attributes; and 2) loading a program and analyzing and extracting features, namely carrying out abstraction on a software code which is subjected to gcc compilation, constructing a program feature sequence for detection of the vulnerability model based on a detected object, extracting program feature information related to the vulnerability state from a program control flow graph, organizing the program feature information by taking a function as a unit and organizing each function into a plurality of operation sequence link list indexed by the detected object. The method solves the problems of a great amount of repeated work, influence on detection speed and inaccurate parameter analysis of the conventional software vulnerability detection method so as to judge whether the vulnerability model exists in the code or not.

Description

technical field [0001] The invention relates to a method for detecting computer software, in particular to a method for detecting software loopholes. Background technique [0002] Software suffers losses due to the existence of hidden security loopholes, which are easy to be attacked, and its security threats tend to increase continuously. Discovering and eliminating security loopholes in software is one of the important means to reduce security threats and security risks by performing vulnerability detection on software codes. Vulnerability model detection is a representative code vulnerability detection method. In order to conduct vulnerability detection based on the vulnerability state machine model, we propose a program feature extraction method for vulnerability model detection, which is used to abstract the program code to be tested and establish a program feature model for vulnerability detection. [0003] The current vulnerability model software vulnerability detec...

AI Technical Summary

Problems solved by technology

[0004] The object of the present invention is: facing the target of software vulnerability model detection based on the vulnerability model, aiming at the problems of a large amount of repetitive work, affecting the detection speed, and inaccurate parameter analysis in the current software vulnerability detection method, based on the state machine model of the software vulnerability, the utilization control Flow graph and data flow analysis technology, centering on the key detection objects, extracts the program features related to the state machine model of software vulnerabilities, and abstracts the software code to model, which can be used for static detection of vulnerabilities to judge whether the vulnerability model is presence in code

Images