BDM locks up occasionally on overnight tests

[Eric Norum]

> "Benshoof, William {AVL~Roswell}" wrote:
> 
> We are using RTEMS with a P&E ICD_BDM, GDB 4.18/DDD 3.2.1/Pavel's
> patches, and a Motorola 68332 CPU.  Overall, our setup is very stable,
> and we aren't having significant troubles.  However, when we set up a
> test to run overnight or for several hours, sometimes we get an
> unexpected (and inexplicable) interrupt.
> 
> We're going to be running this sort of test out of ROM as well as with
> the BDM.  We're also working on implementing leak-detection and usage
> statistics, and we have the stack checker turned on.
> 
> Any ideas on what we might find?  Is there a "this happens sometimes
> with the BDM" category?
> 

I've had RTEMS/BDM/gdb systems running large (>1Mbyte of code, several
Mbytes of data/bss) applications on MC68360 systems with the `5-chip'
BDM adapter for days at a time without problems.  What kind of interrupt
are you seeing?  A `spurious interrupt' (vector 0x60, interrupt 24) is
often the result of clearing a device interrupt enable bit while
processor interrupts are enabled......

-- 
Eric Norum                                 eric.norum at usask.ca
Department of Electrical Engineering       Phone: (306) 966-5394
University of Saskatchewan                 FAX:   (306) 966-5407
Saskatoon, Canada.