Method and device for detecting abnormal memory access

Abstract

The present application discloses a method and device for detecting memory access exceptions, which perform troubleshooting, lexical analysis, grammatical analysis, and semantic analysis on source codes to generate control flow graphs, data dependency graphs, first list files, and second list files; according to The control flow graph builds the global function call graph of the source code, and checks the matching of the source code for dynamic memory allocation according to the global function call graph, the second list file and the data dependency graph; builds a virtual execution platform, and uses the first list file, The control flow graph and data dependency graph extract the execution path, and realize the detection operation of the memory leak caused by the dynamic memory allocation and the memory access out of bounds when the program is running. By analyzing the first list file, the second list file, control flow graph, data dependency graph, and global function call graph, and constructing a virtual execution platform to extract the execution path, it is possible to fully mine the memory access exceptions in the source code and efficiently implement the Detecting operations for memory access exceptions.

Description

technical field [0001] The present application relates to the field of compiling technology, in particular to a method and device for detecting memory access exceptions. Background technique [0002] Memory access exceptions usually manifest as memory leaks caused by dynamic memory allocation, and memory access out-of-bounds when the program is running. Specifically, a memory leak means that the program uses functions such as malloc to dynamically apply for memory space from the operating system at runtime. Because the program temporarily exits or the programmer does not call the free function in the source code to release it, the process occupies memory resources for a long time. Out-of-bounds memory access refers to out-of-bounds access due to abnormal program calculations, irregular code writing, or illegal tampering of memory. These problems may lead to difficult-to-predict results during the running of the program. Therefore, how to detect and judge possible loopholes ...

AI Technical Summary

Problems solved by technology

However, when the program is running, it needs to search the first and last glyphs every time the memory is accessed, and judge whether it is out of bounds, which will seriously affect the execution performance of the program. In addition, the existing technology cannot handle the stack overflow scenario

[0005] It can be seen that the existing technology cannot accurately and efficiently detect the problem of abnormal memory access

Images