Hi!
Yesterday i tried out RTAI. I'd like to say that Paolo (and friends)
did a rather interesting and good job (but the development team of
the original RT-Linux implementation from NMT did a well job, too,
of course :-) Btw: will there be efforts to (re-)join both projects?)
Experiences
===========
After installation, it was realy simple to port my existing RT-Linux-
applications to RTL based on RTAI, but i had a sort of a negative experience:
When runing a RT-Task, sometimes (maybe one time per minute) the srceen
got black for a few microseconds, so you can recognize a black beam
of maybe 3cm in width up to the whole screen. I didn't tried rtai-0.1,
yet (will do it today evening). What could be the problem?
(System: RH 5.2, Kernel 2.2.5, Dual-Celeron 300A, GA-686BXD, Banshee-GFX)
Questions
=========
The most interessting point about RTAI is the SMP-scheduler and the
possibility to lock task to CPUs. BTW: will that feature apear in
NMT-RT-Linux as well?
I had a look into the docs and source-codes of RTAI, but i still have
a lot of questions:
1. The 8254 is used for both, one-shot an periodic mode?
=> in peridoc-mode you only have a fixed time-raster of i.e. 50 us,
but you don't need to reprogram the timer-registers for task-scheduling?
=> there will be a timer-interrupt every i.e. 50 us, and the
scheduler tests if some tasks need to be started?
Or am i completly wrong?
2. The Time Stamp Counter is used for "fine-tuning", only?
That means, the timer interrupt will occur i.e. 10 us earlier as it
should and then the scheduler does a spin-locked wait up to the correct
time?
Is that done in both, periodic and one-shot mode?
Is it possible to disable that feature, if necessary?
Or am i completly wrong as well?
3. There will be used the local-APIC as timer for future schedulers,
to get rid of 8254 and TSC?
=> But fallback is possible in case of using an embedded 486 (i.e. AMD SC400)
Or am i - once more - complety wrong?
Thanks in advance for your answers
--
-------------------------------------------------------------------
| Bernhard Kuhn (kuhn[at]lpr.ei.tum.de) O|||OO||OO| |
| Laboratory for Process Control and Real-Time Systems O|||O|O|O|O |
| Technische Universität München Tel.+49-89-289-23732 O|||OO||OO| |
| 80290 München, Germany Room 3944 Fax -23555 OOO|O|||O|O |
--------------------------------------------------------------------
--- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
----
For more information on Real-Time Linux see:
http://www.rtlinux.org/~rtlinux/
