[tulip] FA310tx causes linux box to hang

[Matthew Horoschun]

Hi All,

We've got a bunch of NetGear FA310TX cards that are semirandomly 
causing a 2.2.5-22 Linux box to halt (and much less often to reboot).

Dec 12 21:29:52 cor1 kernel: tulip.c:v0.89K.1 3/16/99 Originally written by becker@cesdis.gsfc.nasa.gov 
Dec 12 21:29:52 cor1 kernel: Driver modified by Netgear for FA310TX 
Dec 12 21:29:52 cor1 kernel: Netgear technical support: support@netgear.matrixx.net 
Dec 12 21:29:52 cor1 kernel: eth0: NETGEAR NGMC169 MAC at 0xb000, 00 a0 cc d0 3b 37, IRQ 5. 
Dec 12 21:29:52 cor1 kernel: eth0: Checking for MII transceivers... 
Dec 12 21:29:52 cor1 kernel: eth0:  MII transceiver found at MDIO address 1, config 3000 status 7809. 
Dec 12 21:29:52 cor1 kernel: eth1: NETGEAR NGMC169 MAC at 0xa800, 00 a0 cc d0 3b 38, IRQ 12. 
Dec 12 21:29:52 cor1 kernel: eth1: Checking for MII transceivers... 
Dec 12 21:29:52 cor1 kernel: eth1:  MII transceiver found at MDIO address 1, config 3000 status 7829. 
Dec 12 21:29:52 cor1 kernel: eth2: NETGEAR NGMC169 MAC at 0xa400, 00 a0 cc d1 22 61, IRQ 10. 
Dec 12 21:29:52 cor1 kernel: eth2: Checking for MII transceivers... 
Dec 12 21:29:52 cor1 kernel: eth2:  MII transceiver found at MDIO address 1, config 3000 status 7829. 
Dec 12 21:29:52 cor1 kernel: eth3: NETGEAR NGMC169 MAC at 0xa000, 00 a0 cc d1 25 26, IRQ 11. 
Dec 12 21:29:52 cor1 kernel: eth3: Checking for MII transceivers... 
Dec 12 21:29:52 cor1 kernel: eth3:  MII transceiver found at MDIO address 1, config 3000 status 7829. 
Dec 12 21:29:52 cor1 kernel: eth4: NETGEAR NGMC169 MAC at 0x9800, 00 a0 cc d1 22 5f, IRQ 5. 
Dec 12 21:29:52 cor1 kernel: eth4: Checking for MII transceivers... 
Dec 12 21:29:52 cor1 kernel: eth4:  MII transceiver found at MDIO address 1, config 3000 status 7809. 
Dec 12 21:29:52 cor1 kernel: eth5: NETGEAR NGMC169 MAC at 0x9400, 00 a0 cc d1 22 62, IRQ 12. 
Dec 12 21:29:52 cor1 kernel: eth5: Checking for MII transceivers... 
Dec 12 21:29:52 cor1 kernel: eth5:  MII transceiver found at MDIO address 1, config 3000 status 7829.

It has been running perfectly for many months, and then suddenly we've 
had a couple of problems in the space of 48 hours.

We got the following messages from it a few hours before it hung:

Dec 11 15:09:16 cor1 kernel: eth4: The transmitter stopped!  CSR5 is 1ebef, CSR6 7da10.
Dec 11 15:09:16 cor1 kernel: eth4: Something Wicked happened! 0001ebef.

And then the last messages were:

Dec 11 21:23:41 cor1 kernel: eth4: The transmitter stopped!  CSR5 is 800f7aa, CSR6 814c2012.
Dec 11 21:23:41 cor1 kernel: eth4: The transmitter stopped!  CSR5 is 800f7aa, CSR6 814c2012.

We rebooted it once this morning and got:

Dec 12 21:15:05 cor1 kernel:   The PCI BIOS has not enabled the device at 0/104!  Updating PCI command 0000->0005. 
Dec 12 21:15:05 cor1 kernel: eth4: NETGEAR NGMC169 MAC at 0x9810, ff ff ff ff ff ff, IRQ 21. 
Dec 12 21:15:05 cor1 kernel: eth4: Checking for MII transceivers... 
Dec 12 21:15:05 cor1 kernel:   PCI latency timer (CFLT) is unreasonably low at 0.  Setting to 64 clocks.

(a cold boot fixed the problem and it booted normally).

Any hints/ideas on where the problem might be?

Thanks

Matthew

Matthew Horoschun
Network Administrator
ActWEB.NET Australia