thad@public.BTR.COM (Thaddeus P. Floryan) (02/19/91)
jim@syteke.be (Jim Sanchez) in <1792@syteke.be> writes:
I have had the WIN/TCP package and ethernet card for my 3b1 for
about nine months now and it has had problems from day one. I use
it to transfer files from my DOS pc which runs FTP's TCP/IP package
(version 2.03). With fairly high probability when I try transfering
files to/from the 3b1 I get a kernel panic - something about a NMI.
This means a reboot and it pretty frustrating. I have played around
with the window values on the pc but going from 512 to 2048 doesn't
seem to change anything. Any ideas will be appreciated. I have a
combo card w/ram if that matters and use 3.51m version of unix.
MS-DOS vadanya. :-)
Seriously, have you considered the possibilty the PC could be the culprit?
Most of the stuff in comp.protocols.tcp-ip and related newsgroups suggests
that PC-based Ethernet products suck dead bunnies through a straw.
You and I bought our Ethernet cards at the same time from the same source, and
I personally checked your card so I'm convinced your hardware is OK. The ONLY
difference between your setup and mine is that I have 3B1s, Amiga, and some
other CT products on my net, and your net has a 3B1 and the PC.
HOWEVER: I did have a lot of problems at first (tossed the WIN/3B sendmail
crap due to excessive defunct processes) and ported some of the 4.3BSD
networking stuff over to replace the "stock" stuff, but there is one more
thing you should check: the /etc/lddrv/drivers file. From my experience,
the "ether" MUST BE THE LAST ENTRY. NO EXCEPTIONS. My file is:
lipc
cmb
[...]
starlan
voice
ether
and the output of "/etc/lddrv/lddrv -s" is:
DEVNAME ID BLK CHAR LINE SIZE ADDR FLAGS
wind 0 -1 7 -1 0x9000 0x53000 ALLOC BOUND
lipc 1 -1 -1 -1 0x7000 0x360000 ALLOC BOUND
cmb 2 -1 -1 -1 0x3000 0x5c000 ALLOC BOUND
[...]
starlan 5 -1 11 -1 0x14000 0x3de000 ALLOC BOUND
voice 6 -1 13 -1 0xa000 0x33e000 ALLOC BOUND
ether 7 -1 14 -1 0x13000 0x348000 ALLOC BOUND
Absolutely NO OTHER COMBINATION would work on my system. I transfer over
1MB a day between systems over the Ethernet, and lookee here:
thadlabs ksh 7432/7652> date
Tue Feb 19 03:12:52 PST 1991
thadlabs ksh 7432/7652> who -b
. system boot Nov 12 03:44
thadlabs ksh 7432/7652> uptime
up 98 days 23:36:52 booted Mon Nov 12 03:36:08 1990
thadlabs ksh 7432/7652> ruptime
thadlabs up ??:??, 2 users, load 0.06, 0.05, 0.00
tlabs3 up 7+20:07, 0 users, load 0.01, 0.00, 0.00
tlabs4 up 19+31:23, 0 users, load 0.00, 0.00, 0.00
tlabs9 up 42+22:31, 0 users, load 0.00, 0.00, 0.00
Sheesh, another bug in Wollongong software; I guess they never expected a
system could remain up so long. I thought I first saw that "??:??" shortly
after the 90-day period but forgot about it 'til just now. Sigh, guess I
gotta write my own ruptime (or find a 4.3BSD version).
Thad Floryan [ thad@btr.com (OR) {decwrl, mips, fernwood}!btr!thad ]