Hello,

I have recently acquired a VAXstation 4000/90, on which I want to run
NetBSD. I have some questions, both regarding NetBSD and the hardware
itself; I hope this is the right place for them. (If not, just tell me
to go play somewhere else.)

I should add that I have no experience whatsoever with NetBSD/vax; but I
do know Linux, NetBSD/sparc, FreeBSD and SCO OpenServer (in decreasing
order of familiarity.)

Ok, here it goes ...



When booting with a serial console, the VAX displays the following
message:

KA49-A V1.0-006-V4.0
08-00-2B-2C-F3-97
32MB
                                       
?? 001   2     LCSPX  0512
?? 000   8       SYS  0512
.
>>>

Can you tell me what those error messages are, and how I can convince it
to boot unattended?



I boot the VAX off a Linux (2.2.16) server. When the VAX starts, the
mopd on the Linux box displays the following messages:

MOP RC 8:0:2b:2c:f3:97   > ab:0:0:2:0:0      len   37 code 07 SID 
MOP RC 8:0:2b:2c:f3:97   > ab:0:0:2:0:0      len   37 code 07 SID 
MOP RC 8:0:2b:2c:f3:97   > ab:0:0:2:0:0      len   37 code 07 SID 
MOP RC 8:0:2b:2c:f3:97   > ab:0:0:2:0:0      len   37 code 07 SID 
MOP RC 8:0:2b:2c:f3:97   > ab:0:0:2:0:0      len   37 code 07 SID 
MOP RC 8:0:2b:2c:f3:97   > ab:0:0:2:0:0      len   37 code 07 SID 
MOP RC 8:0:2b:2c:f3:97   > ab:0:0:2:0:0      len   37 code 07 SID 
MOP RC 8:0:2b:2c:f3:97   > ab:0:0:2:0:0      len   37 code 07 SID 
MOP RC 8:0:2b:2c:f3:97   > ab:0:0:2:0:0      len   37 code 07 SID 
MOP RC 8:0:2b:2c:f3:97   > ab:0:0:2:0:0      len   37 code 07 SID 
MOP DL 8:0:2b:2c:f3:97   > ab:0:0:1:0:0      len   39 code 08 RPR 
MOP DL 0:50:56:8a:0:0    > 8:0:2b:2c:f3:97   len    1 code 03 ASV 
MOP DL 8:0:2b:2c:f3:97   > ab:0:0:1:0:0      len   39 code 08 RPR 
MOP DL 0:80:ad:73:5c:14  > 8:0:2b:2c:f3:97   len    1 code 03 ASV 

The first messages ("MOP RC") take a rather long time, and as far as I
can make out, the VAX doesn't get a reply to these messages. Is this
normal, or can I tell the VAX not to bother with these messages, so that
it will boot faster?



If I do not add the VAX's ethernet address to Linux's ARP cache by hand,
downloading the kernel does not work; the error messages look like this:

>>> b eza0

 .
-EZA0
>> NetBSD/vax boot [Sep 11 2000 14:15:24] <<
>> Press any key to abort autoboot 0
Trying BOOTP
Using IP address: 192.168.232.8
myip:  (192.168.232.8)
root addr=192.168.232.1 path=/export/leonide
open netbsd.vax: No such file or directory
> boot netbsd
open netbsd: No such file or directory
netbsd: boot failed: No such file or directory
> boot netbsd.gz
1152336read text: Unknown error: code 60
netbsd.gz: boot failed: Unknown error: code 60
> boot netbsd.old
open netbsd.old: No such file or directory
netbsd.old: boot failed: No such file or directory
> boot gennetbsd
open gennetbsd: No such file or directory
gennetbsd: boot failed: No such file or directory
> 

The corresponding parts of the tcpdump look like this:

19:57:50.303630 leonide.priv.272 > moebius.priv.nfs: 104 read [|nfs]
19:57:50.303891 moebius.priv.nfs > leonide.priv.272: reply ok 1124 read [|nfs]
19:57:50.354994 arp who-has leonide.priv tell moebius.priv
19:57:50.359010 leonide.priv.273 > moebius.priv.nfs: 104 read [|nfs]
19:57:50.359450 moebius.priv.nfs > leonide.priv.273: reply ok 1124 read [|nfs]
19:57:50.360873 arp reply leonide.priv is-at 8:0:2b:2c:f3:97
19:57:52.225079 leonide.priv.273 > moebius.priv.nfs: 104 read [|nfs]
19:57:52.225543 moebius.priv.nfs > leonide.priv.273: reply ok 1124 read [|nfs]
[...]
19:58:00.363480 leonide.priv.282 > moebius.priv.nfs: 104 read [|nfs]
19:58:00.363985 moebius.priv.nfs > leonide.priv.282: reply ok 1124 read [|nfs]
19:58:08.356823 leonide.priv.282 > moebius.priv.nfs: 104 read [|nfs]
19:58:08.357335 moebius.priv.nfs > leonide.priv.282: reply ok 1124 read [|nfs]
19:58:25.344153 leonide.priv.283 > moebius.priv.nfs: 108 lookup [|nfs]
19:58:25.344670 moebius.priv.nfs > leonide.priv.283: reply ok 28 lookup [|nfs]
19:58:25.515806 leonide.priv.284 > moebius.priv.nfs: 108 lookup [|nfs]
19:58:25.516144 moebius.priv.nfs > leonide.priv.284: reply ok 28 lookup [|nfs]
19:58:30.334992 arp who-has leonide.priv tell moebius.priv
19:58:31.334997 arp who-has leonide.priv tell moebius.priv
19:58:32.334992 arp who-has leonide.priv tell moebius.priv
19:58:33.334992 arp who-has leonide.priv tell moebius.priv
19:58:34.334989 arp who-has leonide.priv tell moebius.priv
19:58:35.334990 arp who-has leonide.priv tell moebius.priv

Apparently, Linux decides to clear its ARP cache during the transfer,
and the VAX does not properly reply to ARP requests during that
time. Any ideas how I can solve this problem? (I tried to increase the
validity of an ARP entry to at least 60 seconds by saying "echo 120 >
/proc/sys/net/ipv4/neigh/eth0/base_reachable_time", but this didn't
change anything.)



I understand that there is neither a SCSI driver nor a console driver
for this model at the moment. Is there some work being done in this
field, ie. can I expect to see a driver eventually, sometime or other,
without holding my breath? (See, I have this SZ12 with two RZ56, which
I'd rather like to use ...) (I'm sorry, I'd like to help, but both my
programming and my VAX experience are low to non-existent. In a few
years, perhaps ...)



NetBSD boots OK on the machine, by the way (once I get it to boot at
all.) I'm using the GENERIC kernel from NetBSD-1.5_ALPHA2. I haven't
tried building a custom kernel yet (mostly, because I couldn't find the
system sources for 1.5_ALPHA ...) Took some time, though, until I was
able to log into it via telnet ... ;-)



Oh, and ... before I forget it ... It's a great work you're doing with
NetBSD/vax, as well as with the other ports! It runs simply marvellously
on my SS IPX, and I hope to be able to properly use it on the VS4k90
too, since I don't want this very nifty machine go to waste.

-- 
Yours, Sebastian Kirsch <skirsch@moebius.inka.de>