A Static Detection Method for State-Dependent Defects in Large-Scale Systems

Abstract

The invention relates to a static detection method for state-related defects in a large-scale system. Firstly, the state-related defects in the large-scale system are abstracted into a form represented by a finite state machine. Secondly, the Interprocedural Control Flow Execution Tree (ICFET) is generated based on the symbolic execution technique, and a new interval-based path encoding / decoding algorithm is proposed to represent path constraints. Alias ​​analysis and data flow analysis are then performed based on ICFET in a context-sensitive and path-sensitive manner, tracking the flow of each object of each specified type to identify the sequence of events that may occur on the object. Finally, according to the program graph generated by ICFET and alias analysis and data flow analysis, the graph reachability calculation based on constraints is performed. If the reachable edge can reach the undefined or error state in the finite state machine, there is an object or a set of The sequence of events in which an object drives to an undefined or erroneous state on a specification, generating an error report.

Description

technical field [0001] The invention relates to a static detection method for state-related defects in a large-scale system, belonging to the application fields of software engineering, computer systems, and high-performance computing intersecting technologies. Background technique [0002] Large-scale software systems, including operating systems, web browsers, databases, database processing engines, etc., form the backbone of modern computing. Since these systems are widely used in various fields, ensuring their reliability is of paramount importance. Despite continuous efforts by industry and academia to make these systems more reliable, errors still occur frequently in large-scale software systems. [0003] A common type of error in large-scale software systems is state-dependent: after a finite sequence of events has occurred to an object of interest, an error occurs that drives the object into a wrong state. Examples include acquiring locks that are not released when...

AI Technical Summary

Problems solved by technology

[0006] Technical problem: The present invention provides a static detection method for state-related defects in large-scale systems. Accurate and scalable static checking of state properties, efficiently and accurately detecting a variety of errors in large-scale systems, including source-sink problems (such as resource leaks), exception handling errors (such as missing error handling code), or type status related errors (such as inappropriate use of file handlers), etc.

Images