In an attempt to better narrow down the problem I'm having with recent
-current kernels, I built a kernel from sources supped on October 8th which
works fine for me.

(Just a refresher: on recent -current kernels, every process (starting with
init) dies with a SIGBUS in apparantly random spots.  This is an old Leading
Edge 386/16, 8MB RAM, IDE drive).

So, I've been looking at the changes between October 8th and now.  The few
that seem to be in any way related to the problems I'm having now are the
recent changes to locore.s & friends.

What I'm wondering is:

- Does anyone else see this problem?  Newer -current's work fine on our 486,
  which leads me to think that it might be a 386-specific problem.  Have
  people ran newer -currents ("newer" being anything after October 20)
  successfully on 386's?

- Does anyone have any clue/suggestions/hints/ANYTHING on how to debug
  this problem?  I don't mind groping around with DDB, putting debugging
  statements in the kernel, etc etc ... as long as I have some idea as
  to what to do.  I'm not really sure to begin with this one :-(

Thanks for any help!

--Ken