论文部分内容阅读
嵌入式软件需要处理大量中断。中断的随机特性导致越来越复杂的嵌入式系统测试困难。该文基于遗传算法的中断过载动态检测方法,使用均匀变异增加中断类型多样性,非均匀变异算法的使用提高了局部搜索能力;构造产生堆栈深度的中断执行序列测试中断处理的堆栈溢出错误。采用自主研发基于SPARC V8架构的Virtualsparc仿真平台进行中断过载仿真测试,实验表明:该方法可以快速检测堆栈中待处理的中断丢失率,随着堆栈深度的动态修改,大大降低了中断处理在堆栈里的丢失率,同时对中断过载进行定位来提高嵌入式软件中断缺陷检测效率。
Embedded software needs to handle a large number of interrupts. The random nature of interrupts leads to more and more complex embedded system testing difficulties. In this paper, based on genetic algorithm, the algorithm of dynamic overload detection is used to increase the diversity of interrupt type by using uniform mutation. The use of non-uniform mutation algorithm improves the local search ability. The stack overflow error is constructed by interrupt execution sequence test. The experiment of self-developed Virtual Sparc simulation platform based on SPARC V8 architecture shows that this method can quickly detect the interrupt loss rate to be processed in the stack and greatly reduce the interrupt handling in the stack with the dynamic modification of the stack depth Loss rate, while positioning the interrupt overload to improve embedded software interrupt defect detection efficiency.