Defining New Non-preemptive Dispatching and Locking Policies for Ada
In many high-integrity applications, non-preemptive execution is preferred. This paper motivates and defines such a dispatching policy for Ada. It also describes an associated locking policy for protected objects. To deliver non-preemptive execution a new processing resource is introduced – the execution token. In addition to defining these new policies the paper also reviews how non-preemptive tasking systems can be analysed for their schedulability.
KeywordsPriority Task Preemptive Schedule High Priority Task Protected Object Ready Queue
Unable to display preview. Download preview PDF.
- 1.A. Burns. Non-preemptive dispatching and locking policies. In M.G. Harbour, editor, Proceedings of the 10th International Real-Time Ada Workshop. ACM Ada Letters, 2001.Google Scholar
- 2.L. Asplund, B. Johnson, and K. Lundqvist. Session summary: The Ravenscar pro_le and implementation issues. In A. Burns, editor, Proceedings of the 9th International Real-Time Ada Workshop, volume XIX(2), pages 12–14. ACM Ada Letters, June 1999.Google Scholar
- 4.T. Baker and T. Vardanega. Session summary: Tasking profiles. In A.J. Wellings, editor, Proceedings of the 8th International Real-Time Ada Workshop, pages 5–7. ACM Ada Letters, 1997.Google Scholar
- 5.A. Burns and A.J. Welling. Restricted tasking models. In A.J. Wellings, editor, Proceedings of the 8th International Real-Time Ada Workshop, pages 27–32. ACM Ada Letters, 1997.Google Scholar
- 6.A. Burns and A. J. Wellings. Real-Time Systems and Programming Languages. Addison Wesley Longman, 3rd edition, 2001.Google Scholar