just a short quick little note about the status of the
sparc port.

first, it runs. it's not the most solid or best thing around
of course. the sunos compatibility code currently has a few
problems, but most significant binaries do run.

i picked up a second sparcstation so i could test-boot the kernel as i
developed it, and have been plagued by one of it's disks (the disk
that netbsd uses as it's root partition) going bad. at that time, the
linker was still a little bit unstable. I'll get at compiling a full
netbsd userland in a bit.

so.. i have not yet managed to get a build of libc to do the right
thing. as well, there is a tty-association bug in the console driver
code, or more likely, sunos /sbin/init does not do the right thing for
the netbsd kernel.

yes, i am booting (boot netbsd -bs) with a my / and /usr
partitions containing only sunos binaries. a number of things
do the wrong thing, here's a short list:
	ps... well, of course
	df    gets the numbers wrong
	vi    finds some or other ioctl missing

emacs does work though, and i'm able to fsck/mount ufs
filesystems. nfs does not work yet either.

anyways, as a number of people know, not much will be happening
on the sparc code this week, it's holiday time (i'm at berkeley
for a few days.. had to come in and physically "touch" sun-lamp 8-)

if others want to get the kernel build process up and running
on their own machines, it is not difficult. i can't help much
though, as a lot has changed since i setup my cross-compilation
environment. basically, on a sparcstation running sunos4.1.x...

	1. compile up berkeley make
	2. use gcc2.4.5, sun's linker, sun's assembler, and the
	   cpp included with gcc2.4.5 for everything.
	3. compile up a libkern.a. This can take a bit of work,
	   i actually used "gas" to compile pieces of this. the
	   m4 script to generate {u,s}{div,rem}.s is a bit of
	   a pain... enjoy :-)
	4. i actually never compiled up config.new under sunos;
	   rather i run it over nfs from a real netbsd box i
	   have.
	5. config.new a kernel, and bmake it.

	There are likely to be a few problems compiling up
	genassym.c -- nothing truly difficult.

it'll take N>15 hours or so to get this all setup and working
I guess.. if you want something that is known to work you'll
have to wait a bit longer, but if anyone wants to help me finish
this thing off i'd appreciate help.
 <tdr.

ps. cgd has seen it, it does exist. we booted it on a sparcstation
here tonight... and he typed fsck, ls, du, and a few other things
and it worked..

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