> 
> > >0x4c00048f =    stcmi   4, cr0, [r0], #-572
> > >
> > >Co-processor 4 is not a standard co-processor, so my guess is you are 
> > >trying to execute data -- ie your program jumped to never-never land.
> > 
> > Hmmm, I guess a better question would be, "Why do I get this irregularly
> > from make and cc when doing a make build?"  Is this a possible indicator
> > of bad memory, perhaps?
> 
> Bad memory? hmm... I have that ;-)
> 
> Excuse me if this is not relevant, but why doesn't the jump itself produce some 
> sort of segmentation violation & crap out _before_ executing the data?
> 

I don't think NetBSD restricts execute permission on data and stack areas: 
if you can read it you can (try to) run it.