Method for discovering memory destruction vulnerability based on natural language understanding function prototype

Abstract

The invention provides a method for discovering a memory destruction vulnerability based on a natural language understanding function prototype, and relates to the field of software security. The method comprises the following steps: 1, extracting a function prototype in a source code as a to-be-compared function prototype; 2, extracting corpora in the annotations to generate a self-adaptive corpus; 3, performing function prototype segmentation on the to-be-compared function prototype by using the adaptive corpus, and segmenting the to-be-compared function prototype into a group of words or phrases; 4, selecting a function prototype with known function semantics as a reference function prototype, and judging the type of the function prototype to be compared through the reference function prototype; and 5, marking the source code to be detected by using the matching information in the function matching list, and detecting the memory destructive vulnerability by taking the function vulnerability rule as the input of symbolic execution. In the method, the missing report generated in the static source code vulnerability detection is reduced by extracting the memory management function in the source code.

Description

technical field [0001] The invention relates to the field of software security, in particular to a method for discovering memory corruption loopholes based on natural language understanding function prototypes. Background technique [0002] A large part of the reason for memory corruption vulnerabilities is that software developers do not properly manage dynamic memory. Managing dynamic memory usually includes relatively complex memory management logic, so software developers usually use memory management functions for dynamic memory management. However, incorrect use of memory management functions can lead to memory destructive vulnerabilities (for example, after freeing memory, use this memory immediately and release the same memory repeatedly), in static source code vulnerability detection, detection Methods for memory corruption vulnerabilities include: pattern matching, symbolic execution, pointer analysis, etc. Effectively discovering possible memory corruption vulne...

AI Technical Summary

Problems solved by technology

However, in the actual software development process, software developers will not only use standard library memory management functions, but also third-party library memory management functions. In this case, the existing static source code vulnerability detection still only relies on standard The memory management functions of the library are not enough, and false negatives will occur in the actual static source code vulnerability detection

Images