[En-Nut-Discussion] Ethernut Lockups

[Harald Kipp]

Steve,

I'd suggest to use the Watchdog for an immediate workaround.

At 10:22 06.03.2006 +1000, you wrote:
>Hey All
>
>Now I thought the problem could be related to heat as it seemed to
>happen when the air-con was turned off in our office at night.
>But that may just be a coincidence at this stage.

Ethernut 2 becomes really warm, but problems with heat are
extremely rare. Though, electromagnetic fields may cause the
system to reset or hang up. See
http://www.ethernut.de/en/hardware/enut2/index.html
(Known problems)
Switching on/off of aircons or fridges may introduce
such problems.


>Two things I have done of which im not sure if it has any effect on my
>problem is that 1. I have changed the Crystal to 16Mhz. 2. I use Uart1
>for my RS485 comms and just use the standard imagecraft #pragma's for
>the interrupt routines.

We ran Ethernuts at 18 MHz+ without problem.




>Does it matter that I don't use the Nut Interrupt routine registrations
>etc?

No. Running native interrupts without registration is the
fastest way to handle them and fully OK. But you need to
follow certain restrictions like not using Nut/OS API
calls except NutEventPostFromIrq() and only a few more.
http://www.ethernut.de/en/documents/at91-timer-irq.html

>The longest time I seem to have had the ethernut working before a lockup
>Is 24 hours. It then does not respond to any TCP traffic or 485 comms.
>The stacks seem fine I have tried extending them quite a bit without any
>luck.

In the first place I'd blame the software. There's no other
way than trying to isolate the problem by reducing your
application to a bare minimum.

You may also consider using syslog.

Harald