Re...

>>What version kernel was this? I remember people with uVAX IIs complaing
>>that 1.2 didn't work, but 1.1a was stable (kernel-wise). The VAX 3xxx
>>series is classified as uVAXIII (to some people) so some similarities in
>>the hardware (clock stuff perhaps?) may be causing the problem? I don't
>>know that's really the case though, because people have been using
>>3[3,4,5]00's with Kernel version 1.2 with relatively few problems. The big
>>problem is that people "neglect" to read the documentation, and wonder why
>>they can't use DSSI disks.

Many thanks for the reply. 

At the weekend I tried to install 1.2 on the 3500 as per the 3550, with
exactly the same results. The 3500 has a kda50 and ra70 instead of the
KFQSA on the `3550, but other than that and less memory, a similar
configuration. Either there is some rune that i'm not giving it, or
there is a problem with the miniroot binary. Anyhow, downloaded 1.1a
stuff in desparation and lo and behold, miniroot installed and came up
with no problems on both the 3550 and 3500. After downloaded most of
1.1a binaries (Don't want to see the phone bill :-(, have networking and
a more or less complete system working with a current uptime of around
40 hours.

Interesting point about the DSSI disks. ISTR, Ragge mentioned in a
previous message that they are supported only on KFQSA. 1.1a works via
KFQSA here anyway. 

>> I'm vowing to write the drivers...

Having a dig through some of the sources, it seems that there is driver
source for the VSII/III 8 plane graphics, two / three board set, even
though it's not mentioned in the man pages. As the 3500 had a set of
these boards, plugged em in but no console whatsoever, so miniroot at
least doesn't offer support right now. Would be really good to get these
going so that I can run X. It seems that there are two required modes of
operation: The first during bootup where (correct me if i'm wrong) the
device csr's etc are mapped into physical memory space to allow probing
- the driver under these conditions being simple character terminal
mode. The second mode is where the system is up and the device registers
are mapped into virtual memory. From looking at the driver sources, it
seems that it only supports the latter case, so a driver project
(character mode) beckons. Time to download sources and get a build
environment running methinks.

Finally, (rambling on) how many versions of 1.2 are there ?. Is it
possible I have a early, non working copy ?. The same procedure was used
for both installations, so why 1.1a should work and 1.2 doesn't is a
mystery. 

Regards, Chris