[Bug 1906] CBS scheduler implementation

[bugzilla-daemon at rtems.org]

https://www.rtems.org/bugzilla/show_bug.cgi?id=1906

--- Comment #6 from Petr Benes <benesp16 at fel.cvut.cz> 2011-09-13 07:58:24 CDT ---
(In reply to comment #4)
> There are some warnings which I think I cleaned up OK.  Prototype for budget
> overrun callout should be:
> 
> void _Scheduler_CBS_Budget_overrun(Thread_Control *the_thread)
OK
> 
> We discussed not using a signal send from score.  I would prefer a callback to
> be registered.  The use of signals assumes you know that is that the end user
> wants to do and that the signal chosen doesn't conflict.
I blindly tried to change it to a callback. However, the difference is that a
service routine runs in a different context (I assume) because it is invoked
directly from the budget overrun callback. Thus, the user has to use printk()
instead of printf() in the callback. It is a little weird, what do you think? 
> 
> There are warnings in the tests for printf formats.  I fixed those for EDF
> tests before I committed them.
I'm fixing it.
> 
> If we can address the callback/signal issue, this should be OK to merge. 
> Double check Doxygen for warnings.
I do not know how to identify the Doxygen warnings, but I have found two spots
to fix.
> 
> I would like to get these algorithms merged and documented. :)

-- 
Configure bugmail: https://www.rtems.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.