21143 problem

[Douglas Eadline]

Hi,

We recently replaced our "netgear 21140 tulip" with KTI KF-221TX/2
card that use the 21143.  I ran all test described here on both NICs
on the same systems (we just replaced the NICS)

kernel: 2.0.35 SMP (and UP)
tulip driver .90
system PII-266(and 450)

With the old cards (21140) we got great throughput.
Netperf results were very good (full wire speed for both UPD, TCP)

Now it seem the the 211143 has some problems with our switch
(Foundry Networks Workgroup) or something.

netperf calls:
./netperf -t UDP_STREAM -p 12865 -n 2 -l 60 -H coyote3 -- -s 65535 -m 1472
./netperf -t TCP_STREAM -p 12865 -n 2 -l 60 -H coyote3

The netperf results are:

UDP UNIDIRECTIONAL SEND TEST to coyote3
Socket  Message  Elapsed      Messages
Size    Size     Time         Okay Errors   Throughput
bytes   bytes    secs            #      #   10^6bits/sec

 65535    1472   60.00      487486      0      95.68
 65535           60.00      487458             95.68

TCP STREAM TEST to coyote3
Recv   Send    Send
Socket Socket  Message  Elapsed
Size   Size    Size     Time     Throughput
bytes  bytes   bytes    secs.    10^6bits/sec

 65535  65535  65535    60.04       3.65

>>>>The switch also reports a large amount of Aligment errors, FCS Erros,
and shorts for the receiver. 

The nics sometimes come up in half duplex.  Also, If I disconnect the
a cable from the nic, it drops into half duplex when reconnected.

Now here is the interesting thing.   If I use a crossover cable,
I get the following:

CP STREAM TEST to coyote3
Recv   Send    Send                          
Socket Socket  Message  Elapsed              
Size   Size    Size     Time     Throughput  
bytes  bytes   bytes    secs.    10^6bits/sec  

 65535  65535  65535    60.00      93.43 

So there is something going on between the switch and NIC that I do not 
quite understand.

Here is the diagnostic report for  the NIC:
===========================================

tulip-diag.c:v1.03 5/20/98 Donald Becker (becker@cesdis.gsfc.nasa.gov)
Chip Index #1: Found a DC21142 Tulip card at PCI bus 0, device 20 I/O
0xec80.
Digital DS21142 Tulip chip registers at 0xec80:
  ffa08000 ffffffff ffffffff 0000e028 0000e228 f0660000 b3862202 fbfffbff
  e0000000 ffffcbf8 ffffffff fffe0000 41e1d2cd ffff0001 fffbffff 8ff00008
  ffffffff ffffffff ffffffff ffffffff ffffffff ffffffff ffffffff ffffffff
  ffffffff ffffffff ffffffff ffffffff ffffffff ffffffff ffffffff ffffffff
 The Rx process state is 'Waiting for packets'.
 The Tx process state is 'Idle'.
Transmit started, Receive started, full-duplex.
 The transmit threshold is 128.
 Port selection is 100mbps-SYM/PCS 100baseTx scrambler, full-duplex.
EEPROM transceiver/media description for the DC21142 chip.

Leaf node at offset 30, default media type 0800 (Autosense).
 4 transceiver description blocks:
  Media 10baseT,  block type 2.
   Serial transceiver for 10baseT (media type 0).
    GP pin direction 08af  GP pin data 00a5.
  Media 10baseT-Full Duplex,  block type 2.
   Serial transceiver for 10baseT-Full Duplex (media type 4).
    GP pin direction 08af  GP pin data 00a5.
  Media 100baseTx,  block type 4.
   SYM transceiver for 100baseTx (media type 3).
    GP pin direction 08af  GP pin data 00a5.
    No media detection indication (command 80 61).
  Media 100baseTx Full Duplex,  block type 4.
   SYM transceiver for 100baseTx Full Duplex (media type 5).
    GP pin direction 08af  GP pin data 00a5.
    No media detection indication (command 80 61).

Does anyone have any idea?

Doug

-------------------------------------------------------------------
Paralogic, Inc.           |     PEAK     |      Voice:+610.861.6960
115 Research Drive        |   PARALLEL   |        Fax:+610.861.8247
Bethlehem, PA 18017 USA   |  PERFORMANCE |    http://www.plogic.com
-------------------------------------------------------------------