Well all in all I've had a lot of success with NetBSD on cobalt hardware so
far, I've got a Qube 2 using the siop driver with a DC-390F scsi card from
Tekram (it fits like a glove, and its the fastest I could find), and this
machine works more or less flawlessly (I haven't found anything wonky yet).
I got all saucy with this box as its going to be my new firewall/nat box
and even decided to remove the green LEDs and replace them with blue ones.
I chose some with a typical mcd of 350 (the dimmest DigiKey has) and its
still pretty darn bright but its got a nice high swank-factor now.  As soon
as I coerce a buddy of mine with a digital camera to come over I'll try to
get some pictures on the web for grins.

I've got a Raq 1 also up and running, it works beautifully with one
annoying wart, and thats the clock is way too fast.  It gains about 82
seconds every 10 minutes.  For some damn reason that I can't figure out I
can't make ntpd actually sync either.  ntpdate works OK so I've got a cron
hack going, but if anyone has a clear idea of what I'm doing wrong I'd love
to hear it.  I'm pretty new to NetBSD so it may very well just be a config
step I missed in birthing this machine, though why it only affects this raq
and not my qube2 I couldn't say.  Does anyone else with a Raq 1 have time
issues like this or is it just me?

The Qube 2700 I've been battling with is still just sitting here not
getting anywhere.  I got the schematic for the serial interface for it
(thanks Chris) which I'll scan here pretty soon and toss a nice jpeg up
(the only other thing I saw online was an AutoCad file), maybe we should
add it to the general site documentation.  If there's interest I'd be happy
to re-draw it even in xfig so we could get some printer-friendly formats.
Ofcourse all this may be moot as I can't find a DS14C239 for the life of
me.  If anyone knows where I could score one let me know, the other chip
(the 85C30) seems to be standard fare and is evidently made by several
different folks.  I had another thought - as cobalt products tend to
sometimes favor a specific kernel name on boot (vmlinux.gz vmlinux_RAQ.gz
etc.) might the Qube 2700 have a special name of its own?  I didn't even
think to look before I blew that linux disk away...  I kinda doubt this is
the issue but I'm all for an easy out. :)

And finally I'm trying to bring up another raq 1 right now but I'm having
an annoying problem with panics and "TLB out of universe"-style-joy.
Everytime I run fdisk it panics.

# fdisk
wd0: no disk label
panic: TLB out of universe: ksp was 0xc4daff98
Stopped at      0x80123cb8:     lw      v0,392(s2)
db> t
trap+3c8 (10400002,2462ffe8,d,0) ra 28421000 sz 80
PC 0x28421000: not in kernel space
0+28421000 (10400002,2462ffe8,d,0) ra 0 sz 0
User-level: curproc NULL

I know squat about kernel debugging so sorry if this information is
worthless.  The disk is a Quantum Fireball EL and I've had a royal pain of
a time trying to get it partitioned correctly.  My PC's BIOS wants to call
it one thing but NetBSD seems to recognize it as:
bytes/sector: 512
sectors/track: 63
tracks/cylinder: 16
sectors/cylinder: 1008
cylinders: 16383
total sectors: 20044080

The drive however has different geometry mentioned on it, so I'm gonna give
that a shot and see where it leads me.


-- 
Jamie Heilman                   http://audible.transient.net/~jamie/
"Most people wouldn't know music if it came up and bit them on the ass."
                                                        -Frank Zappa