I am fully aware of the fact that Linux is no RTOS.

This uClinux system I am working on now, I wouldn't categorize as
hard-real time. If it were, we wouldn't use uClinux. We use uClinux
because there is a lot of good usable software available; good network
stack, JFFS2 support, NFS support etc.
We have one event that preferably should be handled 'fast' and Greg
already provided a good hint as how to hand-tailor handling of an
interrupt. If 12.5us is what we can get, we'll make our system work
with that. If we run into some trouble, we can redesign the FPGA code
accordingly.

Now I was merely curious if anyone else has some latency numbers for
any CPU, and for m68k in particular. That is all I am asking about.
For future reference, it could be good to know 'typical' latencies for
different CPUs.

Thanks,
Harry

On Nov 22, 2007 5:30 AM, Michael Schnell <[EMAIL PROTECTED]> wrote:
> Linux is not a hard realtime OS !
>
> So no maximum latency can be specified. Even not for ISRs (there is no
> spec how low at max the interrupt might be disabled). For Linux there is
> just some _soft_ realtime spec, meaning that the probability for a delay
> in lower than a certain value.
>
> Kernel 2.4 is outdated and know for much worse soft realtime abilities.
> 2.6 is better and with 2.6.22 some more improvements have been implemented
>
> If you need hard realtime you need to add things like RTAI or PikeOS.
>
> -Michael
>
> _______________________________________________
> uClinux-dev mailing list
> [email protected]
> http://mailman.uclinux.org/mailman/listinfo/uclinux-dev
> This message was resent by [email protected]
> To unsubscribe see:
> http://mailman.uclinux.org/mailman/options/uclinux-dev
>
_______________________________________________
uClinux-dev mailing list
[email protected]
http://mailman.uclinux.org/mailman/listinfo/uclinux-dev
This message was resent by [email protected]
To unsubscribe see:
http://mailman.uclinux.org/mailman/options/uclinux-dev

Reply via email to