Ahem. I disagree with Gernot. There is no new *kernel mechanism*
necessary, the simple priority scheduler will do for RM. But there's
still a little bit to do.
The priority scheduler picks between runnable tasks by their priority.
You still need to ensure that tasks become runnable/not-runnable at the
right times, either by messaging and blocking-on-IPC, or having a
supervisor suspend/resume them. That will probably require a
high-priority user-level task controlling a timer device of its own.
Good luck,
Thomas.
On 07/03/17 11:34, [email protected] wrote:
>> On 7 Mar 2017, at 0:47 , wang <[email protected]> wrote:
>>
>> Hello,
>>
>> I'm now studying seL4/seL4test and try to implement the RM scheduling on
>> seL4test.
>> I understand how the RM scheduling works, but just don't know how to make it
>> happen on seL4test.
>> So where can I start?
> There’s nothing to be done.
>
> RM is a static priority assignment, all you need to do is correctly set the
> priority of the threads you create.
>
>> In addition, I put the sel4test driver image on my board and run, it shows
>> "206/206 tests passed. All is well in the universe" at the last line.
>> Does it means that the seL4test project runs correctly?
> Yes
>
> Gernot
> _______________________________________________
> Devel mailing list
> [email protected]
> https://sel4.systems/lists/listinfo/devel
_______________________________________________
Devel mailing list
[email protected]
https://sel4.systems/lists/listinfo/devel
