more NetGear mising ARPs

[William J. Earl]

jferg@2boot.com writes:
 > Neale Banks wrote:
...
 > > tulip.c:v0.89H 5/23/98 becker@cesdis.gsfc.nasa.gov
 > > eth0: Lite-On 82c168 PNIC at 0xe800, 00 a0 cc 40 0a a7, IRQ 10.
 > > eth0:  MII transceiver found at MDIO address 1, config 1000 status 782d.
 > > eth0:  Advertising 01e1 on PHY 1, previously advertising 01e1.
 > > eth0: Changing PNIC configuration to half-duplex, CSR6 816e0000.
...
 > Check half/full duplex status of everything connected to hub (when you say hub, I assume
 > you mean a PHY-level repeater; they don't handle full-duplex very well)
...

     My configuration has a PHY-level repeater, and the other machine on
the repeater is running half-duplex as well.  I have the same situation
as above:

eth0: Lite-On 82c168 PNIC rev 33 at 0xea00, 00:A0:CC:3E:0B:32, IRQ 11. 
eth0:  MII transceiver #1 config 1000 status 782d advertising 01e1. 
eth0: The transmitter stopped.  CSR5 is 2678016, CSR6 814e2002, new CSR6 814e0000. 
eth0: Changing PNIC configuration to half-duplex, CSR6 814e0000. 

That is, the driver fails to negotiate full-duplex, so it falls back to half-duplex.

     Note that most traffic works just fine, and I can receive ARP
replies from many hosts on the subnet.  Sometimes I can receive ARP
replies from the other machine on the local repeater just fine, and
other times not at all.  That is, which machine is the "problem"
varies from time to time, although a "problem" machine tends to stay a
problem for minutes to hours.

     It seems like something about the size or timing of ARP replies
is somehow confusing the card.