[LITMUS^RT] request for review: patches for arbitrary deadlines
Jeremy Erickson
jerickso at cs.unc.edu
Wed Jul 18 00:51:28 CEST 2012
A few comments:
1.) In base_task.c and base_mt_task.c, you replaced tabs with spaces while
defining PERIOD and EXEC_COST
2.) Why are you requiring that deadline not be more than period? This
constraint is only necessary for HRT systems. For SRT, you just ensure a
tardiness bound of at least $C_i - D_i$. There are a couple of reasons you
could want to do this:
a.) A system designer may have a deadline that is larger than the
average-case execution time, but smaller than the WCET, if doing something
stochastic.
b.) If the deadlines aren't really deadlines but are just priority points.
For example, without your restriction, G-FL can be implemented by
requesting $D_i - (m-1)/m C_i$ as the "deadline", but with your
restriction, a task can get rejected unnecessarily. Technically this case
could actually apply to HRT systems as well.
Otherwise, looks good.
-Jeremy Erickson
On Tue, Jul 17, 2012 at 5:37 PM, Glenn Elliott <gelliott at cs.unc.edu> wrote:
> Thanks for the feedback, everyone.
>
> Please see the revised patches attached.
>
> -Glenn
>
> _______________________________________________
> litmus-dev mailing list
> litmus-dev at lists.litmus-rt.org
> https://lists.litmus-rt.org/listinfo/litmus-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.litmus-rt.org/pipermail/litmus-dev/attachments/20120717/4fbbed82/attachment.html>
More information about the litmus-dev
mailing list