Chris,
thanks for your reply. I had actually tried something like this. I have
had problems with this technique, I was tempted to ask during my first post,
but so many questions...
When I try simply running the RT task without making it periodic, it runs
perfectly but at its completion, the computer behaves very strangely,
responding to keyboard but not mouse, cron won't shutdown on reboot, etc.
This persists until I remove rtl_sched. It is almost as though linux
interrupts are disabled (computer has a sense of humour), although I'm not
sure that explains why I have keyboard access (is it poll driven?).
The other thing is, I presume running the RT task does not prevent the PICs
from raising the INTR line and forcing the CPU to perform some RTLinux calls
like queuing the IRQs, or is this negligible in CPU time? I just want the
jitter that is visibly introduced into the parallel port signal whenever I
bump the mouse to go away.
thanks again for your reply, I don't know whether you have any thoughts on
the above matters, but I thought you might be interested and any input would
be appreciated,
regards,
Craig
> Craig -- the following is a dangerous, but an effective solution
>-- do not set your thread to be periodic -- meaning, fork it off, but do
>not call pthread_make_periodic_np. The side effect of this appears to be
>that all interrupts are effectively blocked an the Linux kernel never gets
>any time slices until your RT task completes. The downside is that you can
>lock yourself out of your system and have to reboot it.
>
> There may be a more fine grained solution -- but I am not sure
>what it is...
>
> Good luck,
> Chris
>
>--------------------------------------------------------------------
>Christopher D. Carothers
>On Sun, 26 Nov 2000, Craig Ireland wrote:
>
> > Hello everybody,
> >
> > I have some questions that I'd be grateful for any answers to-
> >
> > Why does RTLinux provide the rtl_hard_disable_irq call? Does this
>perform
> > an analogous function to masking the interrupts by directly outb'ing to
>the
> > PICs? I am trying to simply disable the serial port interrupts for the
> > duration of a real time task (not just one period) so as to prevent
>mouse
> > movement introducing greater jitter into my parallel port output.
>Currently
> > I've tried calling rtl_hard_disable_irq as well as directly outbing the
>PICs
> > in both my init_module and in my fifo handler. This seems to be utterly
> > ineffectual - my serial mouse remains obstinately alive. Please, even
>if
> > there is a trivial solution to this, could someone please enlighten me.
> > Also, throughout the mailing list and documentation there are dark hints
>to
> > 'ensure interrupts are disabled for only a short time'. What will
>happen
> > otherwise?
> >
> > thanks in advance,
> > Craig Ireland
> >
_____________________________________________________________________________________
Get more from the Web. FREE MSN Explorer download : http://explorer.msn.com
-- [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/
