I've just `solved' a weird problem in a way that makes very little
sense to me. Quite frankly, I wouldn't beleive this story if someone
else told it to me.

I built a new 1.0A kernel for my 1.0 sparc 1+ several days ago. I
tried to boot it, and right after the rc printed `starting network'
something got a bus error, dumped core, and then the kernel dropped
into ddb with an alignment fault.  I played around with it a little,
tried some different kernel options, without luck, but soon ran out of
time so I left it alone.

Today I built a new kernel after another sup, booted it - same
problem.  Ok, I tried with a reset from the prom first. No help. So I
booted -s to trace through and find what was getting the bus error.

I walked through /etc/rc and /etc/netstart, running the commands
manually. Everything worked properly. After I'd passed /etc/netstart
and gotten to the point where it should have printed something else, I
hit ^D to quit the single-user shell, and let it come up.

I was now somewhat confused. I looked around, and found a (useless)
core-file from hostname in /etc, with a date/time corresponding to
when i was trying the other day. Shrugging to myself, I decided to
reboot again, let it fall over, and see what ddb could tell me.

The machine came up happily all by itself, and has done so quite a
number of times ever since. THIS MAKES NO SENSE!! AAAARGH!

So, while it seems I don't really have a problem anymore, I am totally
at a loss as to what was going on. Does anyone have any plausible
theories? I'm not really happy with just thinking that it needed a
little help once and after that could manage for itself.

--
Dan.