I have killed a few bugs in pmap.c and I can now run enough processes
(more than eight) that the pmap code has to fight over contexts and
page-map-entry-groups (PMEGs).  It will now run for a while under
these conditions (15 min. to an hour), but after a while, it still
shows signs of getting VM mappings trashed (SIGSEGV, SIGILL, etc.).

So, it's getting there I guess, but progress is slow and painful...

Currently one can boot this kernel on a Sun3/60 with the console
as either the keyboard/display, ttya or ttyb.  It (reportedly) will
boot on a Sun3/50 but only with ttya or ttyb (not keyboard/display).

Gordon W. Ross			  Internet: <gwr@mc.com>
Mercury Computer Systems	Voice mail: 508-256-0052x295
199 Riverneck Road		Front desk: 508-256-1300
Chelmsford, MA 01824-2820	 Facsimile: 508-256-3599

------------------------------------------------------------------------------