Okay, after seeing how -O2 broke on what was it, nfs_vfsops.c I think,
I took it completely out of sys/arch/sun3/conf/Makefile.sun3 and built
a kernel completely without optimization.  Of course, it's bigger.
That's not the problem.

The problem is that now, the machine seems to be mysteriously hanging.
It will get itself into a state where all running binaries continue to
run, apparently normally, but attempts to start further ones all hang.
As if all exec() attempts are locking up.

And I can't see any way to debug it because NetBSD/sun3 doesn't do
kernel coredumps (yet), and ddb doesn't have any obvious way to trace
the kernel stack of any process but curproc; it's not clear how to get
a valid <frame_addr> for the trace command from ps output.

What's hard about kernel coredumps?  Perhaps that would be the most
worthwhile place for me to invest effort....

					der Mouse

			    mouse@collatz.mcrcim.mcgill.edu