Can the token ring driver be compiled as a module?
Yes, it can be — and it works rather well.
There is an extra parameter that you can use
when it is compiled as a module.
If you ever need to ``spoof'' software install programs,
such as the redhat boot disks, into configuring your token ring card
as an ethernet device (for NFS/FTP installs, etc),
you can use the ``device'' parameter to force a device name
other then tr0.
You will probably have problems if you try this with multiple adapters.
It is mostly there to get around a few incompatabiliites. Ex.:
/sbin/insmod ibmtr device=eth0
I keep getting an error code ``0011''. Whats up?
Make sure that your connection to the network is good or that you have a loopback connector on your token ring card. This message just means that it could not open the ring. 99.99% of the times, it is just not plugged into one.
If you have a sound card in your machine, and it sits at IO 0x220, you may end up with a conflict with your token ring adapter at 0xa20. If you notice that a supported adapter does not seem to be working and you have a sound card, please try to either set your token ring card to 0xa24 or move/remove your sound card.
Also...
Here are some email messages that I have received about Token Ring and Linux. In some of the messages, I have removed parts that were not important to save space.
From: "Mr. Chuck Rickard" <chuck@gl.umbc.edu>
Subject: Re: Token Ring Kernel patch
I d/l'd the patch, applied it, and re-compiled. When booting it said,
"tr0: Can't assign device to adapter" and again for tr1. Any ideas?
Thanks!
Chuck Rickard
(chuck@umbc8.umbc.edu)
From: David Morris <dwm@shell.portal.com>
segment = inb(PIOaddr) in ibmtr.c)
was so out of range that there is no TR card at that IO address.
From: Mike Glover <glover@credit.erin.utoronto.ca>
Subject: Token ring problems.
Thanks for responding. I was starting to thing that I posted incorrectly,
and I was about to post again. Anyway, the following clip is part of my
/var/adm/messages file from when I boot up.
May 2 10:03:14 linux kernel: tr0: Unable to assign adapter to device.
May 2 10:03:14 linux kernel: tr1: Unable to assign adapter to device.
The section, tr0: Unable to assign adapter device. is what kinda confuses
me. The documentation is slim at best so I didn't know what to do with
the message. I know I didn't assign any token ring information,
and I didn't know where to do it.
The hardware:
Its a PC clone (Dell OMNIPLEX 560 to be exact)
The token ring card is a Olicom 16/4 Adapter.
I have DOS token ring drivers and it snaps into the network, so there is
nothing wrong with the card. I think, I am just missing something really
small in the config on the Linux side.
The Olicom 16/4 Adapter does not use the Tropic Chipset. Try using one of the cards that are listed at the top of the HOWTO.
From: Mike Glover <glover@credit.erin.utoronto.ca>
Subject: Found an IBM card...
I found myself an IBM token ring card and I got a little further,
but still not luck. (I did change /etc/rc.d/rc.inet1 ifconfig entry
from eth0 to tr0)
Anyway, here is part of the /var/adm/messages file:
May 2 16:23:07 linux kernel: IPX Portions Copyright (c) 1995 Caldera, Inc.
May 2 16:23:07 linux kernel: tr0: PIOaddr: a20 seg/intr: b8 mmio base:
000dc000 intr: 0
May 2 16:23:07 linux kernel: tr0: Channel ID string not found for PIOaddr: a20
May 2 16:23:07 linux kernel: tr0: Expected for ISA: 5049434f3631313039393020
May 2 16:23:07 linux kernel: tr0: found: 000902003021111000182000
May 2 16:23:07 linux kernel: tr0: Expected for MCA: 4d4152533633583435313820
May 2 16:23:07 linux kernel: tr0: Unable to assign adapter to device.
May 2 16:23:07 linux kernel: tr1: Unable to assign adapter to device.
May 2 16:23:07 linux kernel: PPP: version 0.2.7 (4 channels) NEW_TTY_DRIVERS
OPTIMIZE_FLAGS
Does this tell you anything. It tells me little.
From: David Morris <dwm@shell.portal.com>
From: Mike Glover <glover@credit.erin.utoronto.ca>
Subject: Almost there...
The following sample is what I'm getting on my messages file:
May 3 14:50:24 linux kernel: tr0: now opening the board...
May 3 14:50:24 linux kernel: tr0: board opened...
May 3 15:11:47 linux kernel: tr0: Arrg. Transmitter busy for more than 50 msec.
Donald resets adapter, but resetting
May 3 15:11:47 linux kernel: the IBM tokenring adapter takes a long time.
It might not even help when the
May 3 15:11:47 linux kernel: ring is very busy, so we just wait a little longer
and hope for the best.
May 3 15:11:47 linux kernel: tr0: Arrg. Transmitter busy for more than 50 msec.
Donald resets adapter, but resetting
May 3 15:11:47 linux kernel: the IBM tokenring adapter takes a long time.
It might not even help when the
May 3 15:11:47 linux kernel: ring is very busy, so we just wait a little longer
and hope for the best.
May 3 15:11:47 linux kernel: tr0: Arrg. Transmitter busy for more than 50 msec.
Donald resets adapter, but resetting
The Arrg entry only comes when I telnet, rlogin or ftp to a remote
machine. Once that happens, the connection hangs big time. I took out all
the other hardware (which was 1 SCSI card) and tried running the card all
by itself. Same thing, so it eliminates hardware IRQ's from getting into
a yelling match. The following are the DIP switches on the card:
1 Up |
2 Down |
3 Down | According to the manual this sets the
4 Up | base address to CC000 Which is fine for
5 Up | my machine.
6 Down |
7 Down + This sets the IRQ to 2. Which is also fine
8 Down +
9 Up X Primary Lan adapter. Which it is.
10 Up = 16 KB shared RAM size. This OK?
11 Down =
12 Up # 16 Mbps Data rate.
I was wondering if it is in fact the dip switches, or the way I have
configured my route table and other network info.
ttfn,
Mike
The Arrg problem is pretty much taken care of in the 2.0 kernels. All of the patches have not been ported back to 1.2, and I doubt that they will be. If you get excessive Arrg messages, do yourself a favour and move to Linux 2.0.