Design optimization for AUTOSAR models with preemption thresholds and mixed-criticality scheduling

详细信息    查看全文

• 作者：Qingling Zhao^a ; ^{zhaoqingling09@gmail.com}Author Vitae ; Zonghua Gu ; ^a ; ^{zonghua@gmail.com} ; ^{zgu@zju.edu.cn}Author Vitae ; Haibo Zeng^b ; ^{hbzeng@vt.edu}Author Vitae
• 关键词：Real-time embedded systems ; Real-time scheduling ; Mixed-criticality
• 刊名：Journal of Systems Architecture
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：72
• 期：Complete
• 页码：61-68
• 全文大小：962 K
• 卷排序：72

文摘

Preemption Threshold Scheduling (PTS) is an effective technique for reducing stack memory usage by selectively disabling preemption between pairs of tasks. We consider the AUTOSAR standard in automotive embedded software development, where each task consists of multiple runnables that are scheduled with static priority and preemption threshold. We address the problems of design synthesis from an AUTOSAR model to minimize stack usage for mixed-criticality systems with preemption threshold scheduling, and present algorithms for schedulability analysis and stack usage minimization. Experimental results demonstrate that our approach can significantly reduce the system stack usage.