Re: Scheduling behaviour issue

[Niklas Holsti <niklas.holsti@tidorum.invalid>]

On 2020-04-22 21:05, Anh Vo wrote:
> On Wednesday, April 22, 2020 at 10:20:27 AM UTC-7, Simon Wright wrote:
>> fabien.chouteau@gmail.com writes:
>>
>>> On Wednesday, April 22, 2020 at 1:34:53 PM UTC+2, Simon Wright wrote:
>>>> but FreeRTOS adds the preempted task at the *tail* of its ready queue
>>>> ([4], section Prioritized Pre-emptive Scheduling (without Time Slicing),
>>>> on page 95 or thereabouts).
>>>
>>> I couldn't find the paragraph that says that.
>>
>> It's the dicussion after Figure 29.
>>
>> There are 2 "continuous" tasks (no interaction with the scheduler) of
>> the same priority, say A & B (this is so unlike any real-time system
>> I've encountered! You might have just one, doing some sort of
>> housekeeping or monitoring).
>>
>> Both A and B are ready. One of the tasks, A, runs until a
>> higher-priority task C preempts it; when the higher-priority task
>> finishes, *the other task* B runs!
>>
>>> That seems very strange to not resume the task that was just
>>> preempted.
>>
>> I so agree. I think it's because the mechanism used is derived from
>> FreeRTOS's round-robin scheduler.
> 
> If the preempted task was resumed again, then task B will have no
> chance to run.

Sure it will run, after other tasks at this priority terminate or block.

> Is it a undesired situation.

Only if you desire round-robin execution of tasks with the same priority.

-- 
Niklas Holsti
Tidorum Ltd
niklas holsti tidorum fi
       .      @       .