Tasks and Task Management任务和任务管理[精品].ppt

* Deadline Monotonic Priority Assignment Sporadic processes: period now provides a minimum (or average) bound on their arrival T = 20 ms is guaranteed not to arrive more than once in any 20 ms interval Example: a error handling routine, with timing derived from a fault model Assuming Deadline same as Period is no longer reasonable. DMPO: Fixed priority of a process is inversely proportional to its deadline /~iesag/ * Analysis Response time analysis Works perfectly for values of D less than T as long as the stopping criterion is changed to w_i D_i (instead of equality) Determine response time for a given priority ordering Deadline Monotonic Priority Ordering or DMPO is optimal That is, if for any process set that is schedulable by priority scheme, it is also schedulable by DMPO /~iesag/ * Process Interactions and Blocking /~iesag/ * Priority Inversion and Inheritance Priority inversion A higher priority process L4 waits for a lower priority process (because of resource locks) Result of fixed priority scheme Priority inheritance If a process p is suspended waiting for process q then the priority of q becomes that of p L1 will have priority of L4 And therefore run in preference to L3 and L2 /~iesag/ * /~iesag/ * Response Time Calculations with Blocking Very similar to the non-blocking case With priority inheritance Where usage is a 0/1 function: if resource k is used by at least one process with priority less than i and at least one process with priority greater than or equal to i. /~iesag/ * Dynamic Planning-Based Approaches Flexibility of dynamic approaches + predictability of approaches that check for feasibility On task arrival, before execution begins attempt made to create schedule that contains previously admitted tasks and the new arrival if attempt fails, alternative actions /~iesag/ * Dynamic Best Effort Approaches Task could be preempted any time during execution Don’t know whether timing constraint is met until the deadline arrives, or