> Since RAM usually has ECC or at least parity,

We wish. :-/  Unfortunately there's a lot of non-parity (or
logic-parity, which amounts to the same thing as far as detecting
storage errors goes) memory out there.

> a bad SIMM should cause a memory fault not a data fault.

And maybe it does - does NetBSD/sparc do anything with memory faults?
I'm fairly sure it doesn't on the 1+; I've got three known-bad SIMMs
that I use as memory fault detection test cases, and I don't think I've
ever seen NetBSD/sparc notice memory errors as memory errors.

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B