ANA-6944A/TX (!) + AMIBIOS + linux-2.1.122; 3/4 success!
John Kennedy
jk@csuchico.edu
Tue Sep 29 22:10:04 1998
09/29/98 @ 07:08:34 PM (Tuesday)
To repeatedly follow up to myself....
Ok, I got home. Compiled tulip.c with REVERSE_PROBE_ORDER defined
and the thing started behaving in a more more predictable fashion.
Now I'm having problems with eth3 (see below), but eth2 seems fine and
pretty soon I'm going to check out eth1. Now eth0 is the one that is
easily working (since the order is now reversed).
eth3: Something Wicked happened! 8001ffff.
eth3: The transmitter stopped! CSR5 is 800f7ba, CSR6 ffffffff.
If you down the interface, you can cause something like this:
eth3: 21140 transmit timed out, status 8001ffff,
SIA 7fffffff 00000000 7fffccff fcc00400, resetting...
Rx ring c0014810: 80000000 80000000 80000000 80000000
80000000 80000000 80000000 80000000 80000000 80000000
80000000 80000000 80000000 80000000 80000000 80000000
80000000 80000000 80000000 80000000 80000000 80000000
80000000 80000000 80000000 80000000 80000000 80000000
80000000 80000000 80000000 80000000
Tx ring c0014a10: 80000000 80000000 80000000 80000000
80000000 80000000 80000000 80000000 80000000 80000000
80000000 80000000 80000000 80000000 80000000 00000000
For comparison, the startup syslog now looks like this:
tulip.c:v0.89K 8/8/98 becker@cesdis.gsfc.nasa.gov
eth0: Digital DS21140 Tulip at 0xf800, 00 00 d1 1b 6d 41, IRQ 11.
eth0: EEPROM default media type Autosense.
eth0: Index #0 -
Media MII (#11) described by a 21140 MII PHY (1) block.
eth0: Checking for MII transceivers...
eth0: ***WARNING***: No MII transceiver found!
eth1: Digital DS21140 Tulip at 0xf880, EEPROM not present,
00 00 d1 1b 6d 42, IRQ 11.
eth1: Controller 1 of multiport board.
eth1: EEPROM default media type Autosense.
eth1: Index #0 -
Media MII (#11) described by a 21140 MII PHY (1) block.
eth1: Checking for MII transceivers...
eth1: ***WARNING***: No MII transceiver found!
eth2: Digital DS21140 Tulip at 0xfc00, EEPROM not present,
00 00 d1 1b 6d 43, IRQ 11.
eth2: Controller 2 of multiport board.
eth2: EEPROM default media type Autosense.
eth2: Index #0 -
Media MII (#11) described by a 21140 MII PHY (1) block.
eth2: Checking for MII transceivers...
eth2: ***WARNING***: No MII transceiver found!
eth3: Digital DS21140 Tulip at 0xfc80, EEPROM not present,
00 00 d1 1b 6d 44, IRQ 11.
eth3: Old style EEPROM -- no media selection information.
eth3: Checking for MII transceivers...
So, it looks like there are probably some driver issues with eth3 on
a 6944A. What can I do to help?
--- john
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
09/29/98 @ 03:06:48 PM (Tuesday)
*#*$#*!@!*# ):
s/9644A/6944A/g
--- john
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Date: Tue Sep 29 22:10:04 1998
From: John Kennedy <jk@csuchico.edu>
Message-Id: <199809291752.KAA07743@hircine.jk.ml.org>
Subject: ANA-9644A/TX + AMIBIOS + linux-2.1.122
09/29/98 @ 10:51:32 AM (Tuesday)
I just got this card and got it to work (partially), so I guess I'm
looking for success or known-bug stories.
I just got a ANA-9644A/TX card and I'm trying to stick it in my gateway
(which has AMIBIOS, which causes problems) with linux-2.1.122. In short,
I can only get eth3 to work reliably and the other ports report funny
problems, in win98 as well as linux (probably due to AMIBIOS).
I had lots of problems with 0.83 (stock 2.1.122) and I'm currently using
v0.89H, but I see that 0.89K is probably what I should be using. I'll do
that tonight when I get home. 0.89H let me get one port up and running,
which is at least as good as win98 will do.
I see this in the syslog:
tulip.c:v0.89H 5/23/98 becker@cesdis.gsfc.nasa.gov
eth0: Digital DS21140 Tulip at 0xfc80, EEPROM not present,
00 4c 69 6e 75 79, IRQ 0.
eth0: Old style EEPROM -- no media selection information.
eth1: Digital DS21140 Tulip at 0xfc00, EEPROM not present,
00 4c 69 6e 75 7a, IRQ 0.
eth1: Missing EEPROM, this interface may not work correctly!
eth2: Digital DS21140 Tulip at 0xf880, EEPROM not present,
00 4c 69 6e 75 7b, IRQ 0.
eth2: Missing EEPROM, this interface may not work correctly!
eth3: Digital DS21140 Tulip at 0xf800, 00 00 d1 1b 6d 41, IRQ 11.
eth3: EEPROM default media type Autosense.
eth3: Index #0 -
Media MII (#11) described by a 21140 MII PHY (1) block.
eth3: ***WARNING***: No MII transceiver found!
Right now I'm assuming that AMIBIOS is reversing the order and messing
everything up. Doing to probe in reverse is something else I'll be
trying to test tonight.
Under win98, trying to run the diagnostics got me messages saying that
it was having problems testing the bus-mastering ability of the 9644A.
Rearranging the cards got me sore fingers and little satisfaction.
I have a #9 motion 771 and AHA-2940UW in there as well. The #9 will only
be happy in the upper two PCI slots or so, the 2940UW will lose bits of
the drives if it is isn't in the top two as well, so I'm not positive
it is in a slot that supports bus-mastering. I can't find any useful
documentation from gateway that talks about the motherboard in a G6-300.
Again, color me unimpressed with gateway today.
Win98 shows the "top" 3 21140 ethernet cards with `!' marks, the last
one without and the last one (which is eth3 in linux) seems to be the
one that is working fine. I had the most recent drivers that I could
find (microsoft and the driver floppy that came with the 9644A agreed),
so this is probably AMIBIOS hosing win98 as well, perhaps unrecoverably
(for it; no -DREVERSE_PROBE for win98). (:{
Is there any way to test the bus-mastering ability of that slot? Both
work, at least for one port. Right now 9644A seems a little bit slower
than the 3c905 I had in there before, but I don't have hard timing data.
So: Any success stories with my configuration, am I barking up the
right tree and what can I do to test bus-mastering?
--- john