Current tulip driver from 2.4.5 is plain broken

Current tulip driver from 2.4.5 is plain broken

[Michal Jaegermann]

I mentioned that before but this should be stated clearly.  As far as I
am concerned "Linux Tulip driver version 0.9.15-pre2 (May 16, 2001)",
as used in 2.4.5 - and other kernels - is totally buggered.  It comes up,
and ethernet interfaces can be configured, but does not matter how I am
playing with options I cannot get a single packet through.

Replacing it in 2.4.5 with "Linux Tulip driver version 0.9.14d (April 3,
2001)", which I have handy, restores sanity immediately and a network
simply works without any heroic efforts.

My NIC is "Digital DS21143 Tulip rev 65 at 0x8800".  BTW - a version
"tulip-1.1.7" from sourceforge behaves exactly like 0.9.15-pre2.

This is an output of 'tulip-diag -af' from a working setup:

tulip-diag.c:v2.06 1/8/2001 Donald Becker (becker@scyld.com)
 http://www.scyld.com/diag/index.html
Index #1: Found a Digital DS21143 Tulip adapter at 0x8800.
Digital DS21143 Tulip chip registers at 0x8800:
 0x00: f8a0e000 ffffffff ffffffff 0d572000 0d572200 f0660000 b2422002 fbfffbff
 0x40: e0000000 fff583ff ffffffff 00000000 000052ca ffff0001 fffbffff 8ffd0008
 Port selection is 10mpbs-serial, half-duplex.
 Transmit started, Receive started, half-duplex.
  The Rx process state is 'Waiting for packets'.
  The Tx process state is 'Idle'.
  The transmit threshold is 72.
  The NWay status register is 000052ca.
  Internal autonegotiation state is 'Negotiation complete'.

And this what shows up when I am trying to use "the-new-and-improved":

tulip-diag.c:v2.06 1/8/2001 Donald Becker (becker@scyld.com)
 http://www.scyld.com/diag/index.html
Index #1: Found a Digital DS21143 Tulip adapter at 0x8800.
Digital DS21143 Tulip chip registers at 0x8800:
 0x00: f8a0d000 ffffffff ffffffff 0b9f4000 0b9f4200 f0000000 b2420200 fbfffbff
 0x40: e0000000 fff483ff ffffffff 00000000 000060ca ffff0001 fffbffff 8ffd0000
 Port selection is 10mpbs-serial, full-duplex.
 Transmit stopped, Receive stopped, full-duplex.
  The Rx process state is 'Stopped'.
  The Tx process state is 'Stopped'.
  The transmit threshold is 72.
  The NWay status register is 000060ca.
  Internal autonegotiation state is 'Link check'.

Comments?

    Michal
    michal@harddata.com

Re: Current tulip driver from 2.4.5 is plain broken

[J Brook]

Michal Jaegermann wrote:
> I mentioned that before but this should be stated clearly.  As far
> as I am concerned "Linux Tulip driver version 0.9.15-pre2 (May 16,
> 2001)", as used in 2.4.5 - and other kernels - is totally buggered.
> It comes up, and ethernet interfaces can be configured, but does
> not matter how I am playing with options I cannot get a single
> packet through.
> 
> Replacing it in 2.4.5 with "Linux Tulip driver version 0.9.14d 
> (April 3, 2001)", which I have handy, restores sanity immediately 
> and a network simply works without any heroic efforts.
> 
> My NIC is "Digital DS21143 Tulip rev 65 at 0x8800".  BTW - a
> version "tulip-1.1.7" from sourceforge behaves exactly like 
> 0.9.15-pre2.

 I see exactly the same (broken!) behaviour here. The last kernel
that
works for me in 2.4.4-ac6, which I'm running at the moment. All
subsequent -ac kernels and 2.4.5-pre4 and above are broken. I
reported
the bug last week. Quick system summary: RH7.1, Duron, KT133, Network
card chip "Digital 21041-AA". I get the same problem as above
(working
kernels set half-duplex, broken kernels set full-duplex). More
details
available on request, or at
http://boudicca.tux.org/hypermail/linux-kernel/2001week21/0278.html

  Thanks!

    John
----------------
jbk@postmark.net

Re: Current tulip driver from 2.4.5 is plain broken

[Terry Shull]

J Brook wrote:


>  I see exactly the same (broken!) behaviour here. The last kernel
> that
> works for me in 2.4.4-ac6, which I'm running at the moment. All
> subsequent -ac kernels and 2.4.5-pre4 and above are broken. I
> reported
> the bug last week. Quick system summary: RH7.1, Duron, KT133, Network
> card chip "Digital 21041-AA". I get the same problem as above
> (working
> kernels set half-duplex, broken kernels set full-duplex). More
> details
> available on request, or at
> http://boudicca.tux.org/hypermail/linux-kernel/2001week21/0278.html
> 
>   Thanks!
> 
>     John

RH 7.1 running kernel 2.4.5-ac3, Dual Xeon 450, Supermicro S2DGE,
Network card
Macronix, Inc. [MXIC] MX987x5.

tulip-diag.c:v2.08 5/15/2001 Donald Becker (becker@scyld.com)
 http://www.scyld.com/diag/index.html
Index #1: Found a Macronix 98715 PMAC adapter at 0xe800.
Macronix 98715 PMAC chip registers at 0xe800:
 0x00: fff88000 ffffffff ffffffff 1fefc000 1fefc200 e4660000 01a82202
e7ffebef
 0x40: fffe0000 0fffcf08 ffffffff fffe0000 40a1d0cc ffff0000 ffffffff
fff00000
 Extended registers:
 80: 0f340000 0f340000 0f340000 0f340000 0f340000 0f340000 fc4ffff0
f84ffff0
 a0: 2001527f 2001527f 2001527f 2001527f 2001527f 2001527f 2001527f
2001527f
 c0: 2001527f 2001527f 2001527f 2001527f 2001527f 2001527f 2001527f
2001527f
 e0: 2001527f 2001527f 2001527f 2001527f 2001527f 2001527f 2001527f
2001527f
 Port selection is 10mpbs-serial 100baseTx scrambler, full-duplex.
 Transmit started, Receive started, full-duplex.
  The Rx process state is 'Waiting for packets'.
  The Tx process state is 'Idle'.
  The transmit unit is set to store-and-forward.
EEPROM 64 words, 6 address bits.
 A simplifed EEPROM data table was found.
 The EEPROM does not contain transceiver control information.
   No MII transceivers found!

Works great, no problems at all.