[LITMUS^RT] Schedulability checking

[Björn Brandenburg]

On Feb 25, 2012, at 6:33 PM, Felipe Cerqueira wrote:

> Björn, thanks for the explanation.
> 
> I've been studying the overhead accounting approach in Devi's thesis. I have some questions.
> 


Hi Felipe,

sorry for the late reply, but I just wanted to mention that Uma Devi's discussion of overheads assume a quantum-driven implementation of G-EDF, whereas (the current) LITMUS^RT implements an event-driven G-EDF scheduler. The relevant overhead accounting can be found in Chapter 3 of my dissertation (http://www.cs.unc.edu/~bbb/diss/).


> 1) There are four types of overhead mentioned there. The three that doesn't depend on the application are scheduling, context-switching and tick overheads.
> 
> I've been comparing them with ft_tools overheads. "Plugin" overheads are completely included in their counterparts, right? So we don't use them.
> 
> Scheduling = SCHED + SCHED2
> Context-Switching = CXS
> Tick =  TICK

Correct.

> 
> Though the text doesn't deal with release overheads, is it ok to add them directly to the WCET? Releases only happen once per job.
> 
> T.e = T.e(b) + oh_release(p-EDF) + 2 · oh_sch(p-EDF) + ...

No, this is optimistic since it does not reflect the release costs of *other* jobs. Accounting for release interrupts under G-EDF is actually pretty tricky. I recommend using the preemption-centric method described in Chapter 3 of my dissertation.

Best regards,
Björn