So I got a _really_ dirty Emulex UC07 controller today out of a VAX chassis 
that had been partially crushed :-(. I washed it off, replaced the firmware 
with the -R version, dried it out and set it up as a second disk controller 
on my test vax.

I finally figured out the CSR to use to get it to talk to me as a disk 
device (I have a CXA16 in this VAX) and booted the Disk FRD.

Connected to this box are two SCSI devices, a DEC RRD43 and a Seagate 2GB 
SCSI drive.

Going through the configuration the first time got me to the point where 
UQSSP disk controller 1 was showing DUA1 and DUA4 and the UC07 was showing 
up as UQSSP disk controller 2 and showing DUA1 and DUA3. I tried several 
different options to get the UC07 to call its disks DUB1 and DUB3 but it 
insisted so I added an offset of 8 to make them DUA8 and DUA9.

Then I configured and formatted the Seagate (very slow, took an hour!) Then 
I tried booting from the CD-ROM. No joy there, it just hung.

Then I tried booting NetBSD, it came up, sees the second controller and 
notes that there are disks attached to it. (the UC07 is calling the CDROM 
an RRD40) I can mount the CDROM and see files (I've got the NetBSD iso 
image in there and I can cat the sets to the stdout)

Then I try to write a disklabel to the seagate with disklabel -i. The UC07 
reports _only_ the number of sectors so the fictious label is weird (only 
an a and c partition). I try to change one and disklabel bombs out with a 
floating point exception. I create a label manually and use disklabel to 
write that. That works and now I can newfs the c partition. Great but the 
thing is slower than a pig in winter. To test it I try dd if=/dev/zero 
of=zero.file bs=512 count=5242880 (create a 10MB file of zeros) I finally 
kill it after it has written about half and dd reports 94651 bytes/second 
(yes boys and girls that is 94K bytes/second)

For comparison on the CQD-440/TM that is the primary MSCP controller I get 
from DD
107K bytes/second (same disk type)

This is on a MicroVAX 3400. It seems way to slow to me. (by further 
comparison I get over 1MB/second on my 4000/90) So is Q-bus 10x slower than 
a local SCSI disk?
Michael can you try this on your 4000/300 with SCSI disks and tell us what 
speed you get with 1.5?

--Chuck