> Getting tracebacks from the other processes would be useful.
> 
> t/t 0t<pid>
>     
> Processes 1 and 4 would likely be most interesting.
> 
> Looks like it's "stuck" in the middle of init forking something.
> (Jason's been making a bunch of changes to VM code.. I'd suspect that
> first..)

*sigh*

My bust.  I noticed on a cvs update (when there wasn't a slew of
changes coming in) that I didn't have an up to date kern_descrip.c.
When I diffed it, it was missing one of the two pool inits.  I'm
not sure how that happened or why it didn't cause a more spectacular
or verbose failure...  I was touching that bit of code when trying
to debug the mac68k problems a while back, so this is probably fallout
from that somehow--I've been rather distracted from NetBSD lately.  :-(

Sorry to trouble.

-allen