Am Die, 13 Jun 2000 schrieben Sie:
> > Hallo,
> >
> > I changed the frequency of my rt-mod from 1kHz to 4kHz with the result, the
> > system is frozen (higher frequencies as well). SysRQ doesn't work. Who has an
> > idea why? I'm using rh6.2/rtl-2.2. The module itself is using mbuff, comedi and
> > alocates about 500k memory on the heap by vmalloc. It's designed for use with
> > the posix interface and uses rtf's for commands. The thread code itself uses
> > rtl_hard_savef_and_cli(flags)/rtl_hard_restore_flags(flags).
> >
> how long is the loop on the realtime side ?? you can "feez" a box with
> the measurement example if you set the loop long enough and advance high
> enough , all that happens in this case is what is supposed to happen, the
> realtime job get ALL the CPU time and Linux doesn't get any .
>
> to find out if this is happening there are a fiew ways I guess , my
> pragmatic aproach would be to simply set increase the frequency in a fiew
> steps and check if the Linux system starts getting slow. you could also
> try measuring the time of your internal loop , or simply let it print
> something to the console once in a while and I guess you would see that
> your realtime app goes on happily printing to the console and Linux
> "looks frozen" ...
I found the problem - it isn't related to the realtime task which can run up to
20 kHz (I haven't tryed more because is enough for me). It's related to comedi
and the DAQ.
The system looks frozen at a point, no "slowing/frozen" like jam or so near
this frequency. My design does analog output (zero) at idle
time, so I didn't distinguish it at first time. (After this: I have a bad design
in my module as a conclusion ...).
I guess the code for measuring time I can find in the example dir from the rtl
dist (jitter measurement) or not ?
Thanks Olaf
-- [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/
