A vulnerability analysis method based on implicit taint propagation

Abstract

The invention provides a vulnerability analysis method based on implicit stain propagation. The method comprises the steps of extracting a control flow graph of a to-be-analyzed program; calculating a control dependence relationship of front and back inevitable nodes of each instruction in the control flow graph; starting the program and monitoring each instruction performed by the program; adding a return value of a monitored read data operation function called by the program to a stain record; tracking the instructions performed by the program and calculating a forward stain propagation process until an abnormity occurs in the program or an analysis process is interrupted; and if only an explicit stain propagation path obtained in the forward stain propagation calculation process exist, starting reverse backtracking by taking a control flow hijack point existing in the program as a starting point, and if the backtracking is interrupted, employing implicit stain propagation, thereby enabling a backtracking process to be carried out continuously.

Description

technical field [0001] The invention relates to the technical field of network security, in particular to a vulnerability analysis method based on implicit taint propagation. Background technique [0002] The current software vulnerability analysis mainly adopts the taint propagation analysis method, but the traditional taint propagation is usually analyzed for data dependence. It is better for file vulnerability analysis, but it cannot track browser vulnerabilities caused by Java Script and other scripts. The reason is that Script-like programs are usually executed by parsing, and in program logic, they appear as an implicit taint propagation process in which input data generates control dependencies and affects program execution. Therefore, it is necessary to improve the ability of implicit taint propagation analysis for control dependencies. [0003] The current analysis techniques based on explicit taint propagation usually use the following methods: [0004] 1. Interc...

AI Technical Summary

Problems solved by technology

Currently, browsers use Java Script technology extensively. Vulnerability attack codes use Java Script to control memory layout and overwrite control flow transfer addresses. These operations are implemented by parsing the input script string through the script analysis engine, and the input script content is not consistent with the program Direct data dependencies are generated. Existing vulnerability analysis methods are implemented based on explicit taint propagation, and only analyze data dependencies between instructions and input data. When analyzing script vulnerabilities, the analysis process will be suspended and cannot Achieve analysis effect

Images