Task management method for enhancing real-time performance

Abstract

A task management method for enhancing real-time performance of a RTEMS (real-time executive for multiprocessor system) includes steps: A, scheduling a certain task by sending semaphore; B, setting up a co-task prior task ready queue, arranging a task to the co-task prior task ready queue if the task is supposed to be put into the ready array according to the state value of the semaphore of the task, and otherwise, forcedly occupying processor resources and performing the task; C, occupying the threshold of an appointed priority level of the task and judging whether the task is allowed to be occupied or not according to the threshold; D, when the task is added to the co-task prior task ready queue, judging the priority level of the task according to the semaphore of the task and the occupied threshold, and positioning an insertion position in the co-task prior task ready queue to form an ordered task block array.

Description

technical field [0001] The invention relates to the field of real-time system operating system kernel design, in particular to a task management method for enhancing real-time performance in an RTEMS real-time operating system. Background technique [0002] With the increasing demand and scale of real-time systems, the traditional small-scale single-chip system can no longer meet the needs of large-scale real-time control systems. While the real-time system is getting larger in scale, there are also many design factors that conflict with real-time performance and time determinism. More optimizations must be made on the basis of the existing real-time operating system in order to make the entire system more stable and have Greater timing determinism allows application developers greater flexibility. [0003] As an open source real-time operating system, RTEMS has many advantages, but if it is applied to a high-precision control real-time system, there are still many shortcom...

AI Technical Summary

Problems solved by technology

2. When the task scale is relatively large, when multiple tasks compete for a resource, the time delay fluctuates greatly, which cannot meet the time deterministic requirements of hard real-time systems

It cannot control the running order of tasks of the same priority

4. The design of the scheduling algorithm is relatively simple. If the operation of a certain task is triggered according to the semaphore, a series of system calls need to be executed, which is time-consuming, and the specified task cannot be enforced.

Images