考虑中断和上下文切换开销的响应时间分析
作者:
作者单位:

作者简介:

于广良(1986-),男,山东莱州人,博士,工程师,主要研究领域为实时系统,嵌入式系统,可信软件;杨孟飞(1962-),男,博士,研究员,博士生导师,CCF高级会员,主要研究领域为空间飞行器设计,控制计算机,可信软件.

通讯作者:

于广良,E-mail:ygl_222@126.com

基金项目:

国家自然科学基金(91118007,61502031,61632005)


Response Time Analysis of Embedded Systems by Taking Interrupt and Context Switch Overheads into Account
Author:
Affiliation:

Fund Project:

National Natural Science Foundation of China (91118007, 61502031, 61632005)

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
    摘要:

    实时嵌入式系统多采用中断和上下文切换实现多任务间调度,在对此类系统进行可调度性分析时,在任务的最差响应时间计算中必须包含中断和上下文切换开销.现有包含这些开销的方法是将中断作为高优先级任务,同时将上下文切换开销加入到任务最差执行时间中进行分析,然而这些方法过于粗略,缺乏对实际系统细节的考虑,计算得到的最差响应时间并不精确.首先,对中断和上下文切换的机制和时间流程进行详细的阐述,进而分析中断和上下文切换对任务关键性时刻的影响;接着,给出包含上述开销的更加精确的响应时间计算方法;最后进行仿真验证.扩展了包含系统调度开销的响应时间计算方法,可用于资源受限的硬实时系统中需要精确计算响应时间的场合.

    Abstract:

    Interrupt and context switch are basic mechanisms for multi-task scheduling in real-time and embedded systems. During schedulability analysis, the overheads of the interrupt and context switch should be considered in the calculation of tasks' worst-case response time. The current calculation methods of response time add interrupt as task with high priority, and simply add the overheads of context switch in the meantime. However, these methods neglect the details of practical systems and roughly give an inaccurate worst-case response time. In this paper, the mechanisms and time flow of interrupt and context switch are thoroughly described. In addition, their influence on tasks' critical instant is discussed. More importantly, a much more accurate calculation method for response time is presented. At last, simulations is conducted to validate the improvement in accuracy of this new method. The response time analysis accounting for scheduling overheads is extended in this paper for the resource-constrained hard real-time systems which need to accurately calculate the response time.

    参考文献
    相似文献
    引证文献
引用本文

于广良,杨孟飞.考虑中断和上下文切换开销的响应时间分析.软件学报,2018,29(6):1681-1698

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
历史
  • 收稿日期:2017-07-01
  • 最后修改日期:2017-09-01
  • 录用日期:2017-11-06
  • 在线发布日期: 2017-12-28
您是第位访问者
版权所有:中国科学院软件研究所 京ICP备05046678号-3
地址:北京市海淀区中关村南四街4号,邮政编码:100190
电话:010-62562563 传真:010-62562533 Email:jos@iscas.ac.cn
技术支持:北京勤云科技发展有限公司

京公网安备 11040202500063号