Quoting Ian Clark <ic@smoke.com.au>:

> Well I can report that there is at least 1 (one) DIO SCSI card in
> existance! I pulled it out of a pile of scrap  300s. And yes it is 
> the ONLY one I have ever seen. Believe me, I checked every other 
> machine in the pile...

NetBSD 1.4.3 (NEWMUULI) #1: Sat Mar  3 00:59:20 PST 2001
    jarkko@newmuuli:/usr/src/sys/arch/hp300/compile/NEWMUULI
HP 9000/380 (25MHz MC68040 CPU+MMU+FPU, 4k on-chip physical I/D caches)
cpu: delay divisor 30, mmuid 2
real mem  = 29347840
avail mem = 23035904
.
.
oscsi0 at dio0 scode 14 ipl 3: 32 bit dma, async, no parity, scsi id 7 
sd0 at oscsi0 targ 0 lun 0: <HP, D2076A, 8870> (SCSI-2)
sd0: 1872 cylinders, 15 heads, 2059140 blocks, 512 bytes/block
sd1 at oscsi0 targ 1 lun 0: <HP, C2247-300, 0BA4> (SCSI-2)
sd1: 2051 cylinders, 13 heads, 2054864 blocks, 512 bytes/block
sd2 at oscsi0 targ 2 lun 0: <MICROP, 4421-07   0502SJ, 0502> (SCSI-2)
sd2: 4049 cylinders, 7 heads, 4193360 blocks, 512 bytes/block
dcm0 at dio0 scode 17 ipl 5
internal parallel at dio0 scode 24 not configured
le0 at dio0 scode 25 ipl 4: address 08:00:09:07:9b:98
le0: 8 receive buffers, 2 transmit buffers
fhpib0 at dio0 scode 29 ipl 3: 98625A/98625B HP-IB
hpibbus1 at fhpib0
oscsi1 at dio0 scode 30 ipl 3:async, scsi id 7
interrupt levels: bio = 3, net = 4, tty = 5
 
I'm not sure if they're rare but I can say that they're not really common.
Wish I had more :-)
-- 
jht