The Simulation Environment Of Embedded Binary Software Defect Detection System

The quickly development of embedded software industry has been made itbecome one of the most popular and promising IT applications. This article solvessome problems encountered in the process of developing embedded software. Mainlyusing virtual simulation and defect detection technology to accelerate embeddedsoftware development,establish a complete system and deepen the developmentmodel which uses the virtual simulationenvironment to develop embedded software.From the hardware simulation to software, the detectionfor software defects such asmemory leak and uninitialized variableetc is all included in the system, andthroughout the whole processof developing embedded software.An embedded software defect detection system based on the binary code is adetection system which is aimed at detecting embedded binary code. The detectionsystem consists of a virtual simulation environment; a defect detection engine anddefect report generation. Firstly, make instruction analysis on a binary code programwhich needs to be measured. Secondly, do instrumentation and insert thecorresponding function codes into the program. Thirdly, collect execution informationwhen the program operates in a virtual simulation environment. Finally, form a defectdetectionreport about the embedded software.The main work and achievements of this thesis can be summarized as followingpoints.1. An in-depth study of virtualization technology and application scenarios andexplore of feasibility of embedded software defect detection which based on asimulation environment.2. After having analysed the source code and the working principle of anexisting open source software QEMU, we implemented a multi-architecture simulatorwhich named Emu. Although currently it only supports embedded ARM developmentplatform series, but thanks to the open and extensible development approach, thelatter continue to improve supported hardware platforms and architectures becomeeasy.3. After having researched on correlative techniques about defect detection, weput forward three solutions which offer simulation support for the defect detection engine. These solutions have their pros and cons,and have been verified about theirfeasibility and versatility.They are emulation support programs respectively based onthe resident program, the network communication and the depth integration.4. Having designed and implemented a complete embedded hardwaresimulation environment by using C/C++language. With simulation support, defectdetection and analysis tools can detect memory leaks defects, uninitialized memorydefects, stack overflow and integer overflow etc, which all belong to embeddedsoftware common pitfalls.