>>If I try to boot (sd@3,0) now I've this drive going, it still stops at
>>the same spot. This blows the theory it's looking at wd0, as it at
>>least now exists. Worst case I can just use this as the boot drive..
>>though this shouldn't be necessary. :(
> 
> True. Did you try a -current kernel? I'd like to know if it still hangs
> there.

Nope, but that's an idea.

> Hmm, the card apparently has its own PCI bridge, I remember some people
> having trouble with these.
> 
> So - please try a -current kernel, I'll build one for you if you don't
> want to go through the trouble of rolling your own.  
> I /think/ the code that maps the OF device name to NetBSD's may choke on
> your boot path: /pci@1f,0/pci@1/pci@3/SUNW,isptwo@4/sd@3,0:a/netbsd
> three nested instances of pci aren't /that/ common after all and macppc
> had a similar problem a while ago ( it didn't hang though, just didn't
> use the right device ) 

I do recall reading something to that effect. The commas, I believe?

> If -current still has the problem it should definitely be fixed for 3.0
> You may want to file a bug report about it too, just to make sure we
> don't forget about it ;)

Indeed!

> Hmm, your CPU looks a bit odd too, I thought all UltraSPARC-IIi modules
> for U5/U10 boxes with 300MHz or more had 2MB cache.

AFAIR the 333MHz and 400MHz modules have 2M cache, the 270MHz, 360MHz 
and 440MHz (U10 only, unofficial in newer U5 boards?) have 256K.

> Does the HME on the card work properly?

Under Solaris, yes.

Actually one thing I noticed just after sending the last message and 
looking over the dmesg, was that hme0 and hme1 seem to have the same MAC 
address!

hme0 at pci1 dev 1 function 1: Sun Happy Meal Ethernet, rev. 1
hme0: interrupting at ivec 3021
hme0: Ethernet address 08:00:20:b6:14:e8

hme1 at pci3 dev 0 function 1: Sun Happy Meal Ethernet, rev. 1
hme1: interrupting at ivec 3019
hme1: Ethernet address 08:00:20:b6:14:e8

The first being the onboard interface. I'm not sure if this would affect 
the SCSI bit of the card, but this certainly isn't right, is it?

I'll work on getting -current built sometime this week.

Thanks