实时系统-2.ppt

RTA with Non-Preemptive Scheduling A special case: all tasks are non-preemptive RTA for non-preemptive Fixed-Priority Scheduling? RTA with Non-Preemptive Sections RTA with Non-Preemptive Sections RTA with Non-Preemptive Sections We need to look into multiple instances Some Well-known Mistakes RTA for Non-preemptive scheduling CAN bus analysis Sporadic Server in POSIX RTA for cyclic graphs … … The Jitter Problem So far …… We assumed that tasks are released exactly at the period boundaries This is realistic and can be implemented, but sometimes, the release time may change a little, but the jitter is bounded by some constant value J Jitters may cause tasks to miss deadline! * Jitter – An Example An example T1 = (20, 100), T2 = (40, 150), T3 = (20, after T2) T3’s release time depends on T2’s response time * Jitter – Definitions Jb : maximal delay from period start Js : minimal delay from period start Jitter = Jb – Js Jitter The maximal length of the interval in which a task may be released non-deterministically Jb = Js : no jitter * Jitter – An Example An example T1 = (20, 100), T2 = (40, 150), T3 = (20, after T2) Prio(T1) Prio(T2) Prio(T3) * Jb = T2’s worst-case response time = 60; Js = T2’s best-case response time = 40 Jitter = Jb – Js = 60 – 40 = 20 Jitter – An Example An example T1 = (20, 100), T2 = (40, 150), T3 = (20, anytime during T2) Prio(T1) Prio(T2) Prio(T3) * Jb = T2’s worst-case exe time = 60; Js = 0 Jitter = Jb – Js = 60 – 0 = 60 The Number of Preemptions Due to Jitter * The Number of Preemptions Due to Jitter * The Number of Preemptions Due to Jitter * The Number of Preemptions Due to Jitter * The Number of Preemptions Due to Jitter The number of preemptions experienced by an LP task due to the jitter of an HP task is * Handling Jitter in Schedulability Analysis * * Worst case response time calculation * Worst case response time calculation Let Ri stand for the response time for task i. Then Ci is the computing time I(i,j) is the