Pierre Cloutier wrote:
>
> At 09:11 AM 05-06-99 -0600, you wrote:
> >RTLinux support FP . If you don't set use_fp, then you are relying
> >on good luck.
> >
>
> I did indeed set use_fp. The RT task uses the FPU and therefore rtl_sched
> _must_ save and restore the FPU register stack. That's understood.
>
> I also tried a floating point divide by 0.0 and the task still survives. A
> integer divide by zero results in the expected nasty crash.
> I notice that sqrt(-1.0) seems to return an extremely small number, whereas
> the divide by 0.0 seems to return a very large number.
>
> Is it possible Linux is configuring the FPU to return the nearest floating
> point number that makes sense? Does it mean we never get floating point
> exceptions? I think these are important questions.
>
> Regards
Simple in RTAI, purposely, and in RTL, I don not know if purposely as
well, the fpu is setup to do standard recovery and no trapping of fp
errors. Is up to you to control them.
Since in RTAI everything can be trapped without minding of the kernel
you can do it differently, if it is needed.
Ciao, Paolo.
--- [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/
