[tulip] DC21143 rev 65 - hopeless case?

[John Sutton]

Hi

Is this Intel chip with tulip driver utterly beyond hope?  I've got quite a few
in service so it's a pain to have to replace them all, especially since I know
that they are perfectly usable under Windoze ;-(

I've just returned a failed Kingston KNE100TX to my supplier, hoping that the
replacement would NOT be a rev 65...  but it is.  So either my supplier is
holding some very old stock OR Kingston are still shipping these chips?

Am I the only one with this problem?  (See posting below from 21st Dec for
details).

TIA
John Sutton

****************************

Hi there

I'm trying to get to the bottom of a persistent problem I have been having with
Kingston KNE100TX cards.  I've tried upgrading to driver v0.92 but (see
previous post - "[tulip] Multiple detection?") have temporarily abandoned that
and reverted to v0.91g.

I've been using these cards for over a year now in a server/diskless-client
configuration and everything works fine *most of the time*.  But occasionally,
and usually when I'm doing a big file transer, it all locks up.  In particular,
what usually happens is that the server card locks up and although restarting
the network layer on the server usually helps, it's never quite right again
until I do a full reboot.

Anyways, in an effort to get a handle on this, I've downloaded and built the
diagnostics.  And here's the output of tulip-diag after a lockup has occured:

tulip-diag.c:v2.04 9/26/2000 Donald Becker (becker@scyld.com)
http://www.scyld.com/diag/index.html
Index #1: Found a Digital DS21143 Tulip adapter at 0xf400.
Port selection is MII, full-duplex.
Transmit started, Receive started, full-duplex.
The Rx process state is 'Suspended -- no Rx buffers'.
The Tx process state is 'Idle'.
The transmit threshold is 1024.
Interrupt sources are pending!  CSR5 is f06988c7.
Tx done indication.
Tx complete indication.
Tx out of buffers indication.
Rx Done indication.
Receiver out of buffers indication.
Timer expired indication.
The NWay status register is 000000c6.
Internal autonegotiation state is 'Autonegotiation disabled'.
Use '-a' or '-aa' to show device registers,
'-e' to show EEPROM contents, -ee for parsed contents,
or '-m' or '-mm' to show MII management registers.

Any ideas?

I've never been sure whether the root cause of this problem is actually NFS, or
the card driver.  But I'm beginning to suspect the card driver because I also
get a lockup when using a Win95 machine as client and doing transfers over SMB.
And Win95 to Win95 I've never had a lockup of any sort (amazing but true ;-).

Also, to try to get a handle on the problem, I've have been trying to change the
speed/duplex/etc of the cards by passing kernel parameters e.g.
ether=0,0,10,eth0 at the LILO prompt (for the server) and in /etc/bootptab (for
the clients).  But nothing I do makes the slightest bit of difference ;-(  What
am I doing wrong?

TIA

***************************************************
John Sutton
SCL Computer Services
URL http://www.scl.co.uk/
Tel. +44 (0) 1239 621021
***************************************************