On Tue, 08 Feb 2000, Kevin Morgan wrote:
> MontaVista certainly won't recreate the work of Ingo! That is nice work.
> However, preemption points have their limitations, and are difficult to
> maintain over time. Also, preemption points don't address excessive
> interrupt off paths (which are significant and already reported as problems
> from more than one of our customers).
Agreed.
We are doing general characterization
> work now (with an immediate focus on our customer reported responsiveness
> problems), and will also do these characterizations on systems with Ingo's
> patches; we'll report the results here and on our web pages.
very nice !
Do you have an estimated date when first results will be available ?
Preliminary
> data indicates there is plenty of opportunity for dramatic response
> characteristic improvements in Linux, and we believe those opportunities
> extend well beyond what is achieveable with carefully placed preemption
> points. Our initial data on this will be forthcoming, and will flow out
> over time as we work through a sequence of improvement efforts.
Yes, the only thing to keep in mind is that we don't want do sacrifice too much
performance while in favor of latency.
I would accept a small overall performance decrease in exchange of good
responsiveness, but full preemtion would be overkill and would cause a big
performance hit.
> ps: yes all MontaVista product work in kernel space will of course be
> released under GPL, just as our compact PCI
> backplane networking and powerPC cross development packages have been
> released under GPL.
>
> pps: We have no illusions about getting native Linux down the the RTLinux
> range of responsiveness, 15-20 microsecond "guaranteed" kinds of levels.
> RTLinux is a fine solution for that level of real-time requirement.
I think most of us will be happy with 500-1ms max scheduling latency
when running SCHED_FIFO tasks.
Especially multimedia folks because this will enable linux to solve a wide range
of real-time multimedia problems like audio and video.
(RT audio being more critical than video since you can hear every little
deadline miss and using more than 5-20ms buffers will kill the responsiveness
completely.
Benno.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue Feb 15 2000 - 21:00:13 EST