[eepro100] EEPRO100 issues.
Heflin, Roger A.
Roger.A.Heflin@usa.conoco.com
Tue, 6 Mar 2001 09:18:49 -0600
Hello,
We are testing out a new motherboard (Tyan Thunder LE 2510 motherboard).
This motherboard is a serverworks LE motherboard with dual 1Ghz cpus.
We are running it under 2.2.19pre11, and the same release/build/setup is
working
correctly on 2 other machines without the built in networking (external pci
card),
both are supposed to be intel eepro implementations.
Under load after 5-60 minutes the network just stops. There are not
messages
indicating anything happened, or what happened. We get the expected NFS
timeout
and RPC timeout messages, but no messages from the driver saying it had a
problem.
Doing an ifdown/ifup fixes the issue for a while, and then it happens again
in 5-60
minutes.
The original (when working) diag program output is below, and the diff on
the working
and nonworking is also included. ETH0 is not configured, eth1 is
configured, on
the mii-diag eth1 changes a register from 0003 (working) to 0001(not
working), and 0001
is the same value as unconfigred eth0 has.
The machine has dual ethernet, we have duplicated it with either interface,
we have
tried noapic and it does not appear to make any difference.
I have tried debug=6 on the module load, but the messages put into
/var/log/messages
are very garbled compared to what dmesg produces, and dmesg does not keep
messages
long enough that I can get the failure messages.
It has:
Mar 5 13:07:47 poeplt02 kernel:
http://www.scyld.com/network/eepro100.html
Mar 5 13:07:47 poeplt02 kernel: eth0: OEM i82557/i82558 10/100 Ethernet at
0xc0919000, 00:E0:81:01:59:F9, IRQ 20.
Mar 5 13:07:47 poeplt02 kernel: Board assembly 123456-120, Physical
connectors present: RJ45
Mar 5 13:07:47 poeplt02 kernel: Primary interface chip i82555 PHY #1.
Mar 5 13:07:47 poeplt02 kernel: General self-test: passed.
Mar 5 13:07:47 poeplt02 kernel: Serial sub-system self-test: passed.
Mar 5 13:07:47 poeplt02 kernel: Internal registers self-test: passed.
Mar 5 13:07:47 poeplt02 kernel: ROM checksum self-test: passed
(0x04f4518b).
Mar 5 13:07:47 poeplt02 kernel: Receiver lock-up workaround activated.
Mar 5 13:07:47 poeplt02 kernel: eth1: OEM i82557/i82558 10/100 Ethernet at
0xc0925000, 00:E0:81:01:59:FA, IRQ 21.
Mar 5 13:07:47 poeplt02 kernel: Board assembly 123456-120, Physical
connectors present: RJ45
Mar 5 13:07:47 poeplt02 kernel: Primary interface chip i82555 PHY #1.
Mar 5 13:07:47 poeplt02 kernel: General self-test: passed.
Mar 5 13:07:47 poeplt02 kernel: Serial sub-system self-test: passed.
Mar 5 13:07:47 poeplt02 kernel: Internal registers self-test: passed.
Mar 5 13:07:47 poeplt02 kernel: ROM checksum self-test: passed
(0x04f4518b).
Mar 5 13:07:47 poeplt02 kernel: Receiver lock-up workaround activated.
eepro100-diag.c:v2.02 7/19/2000 Donald Becker (becker@scyld.com)
http://www.scyld.com/diag/index.html
Index #1: Found a Intel i82557 (or i82558) EtherExpressPro100B adapter at
0xd400.
i82557 chip registers at 0xd400:
00000000 00000000 00000000 00080002 18203000 00000000
No interrupt sources are pending.
The transmit unit state is 'Idle'.
The receive unit state is 'Idle'.
This status is unusual for an activated interface.
Index #2: Found a Intel i82557 (or i82558) EtherExpressPro100B adapter at
0xd000.
i82557 chip registers at 0xd000:
00000050 3cd720e4 00000000 00080002 182541e1 00000600
No interrupt sources are pending.
The transmit unit state is 'Suspended'.
The receive unit state is 'Ready'.
This status is normal for an activated but idle interface.
Diffs between good run and the bad run:
5c5
< 00000000 00000000 00000000 00080002 1821782d 00000000
---
> 00000000 00000000 00000000 00080002 18203000 00000000
mii-diag output:
Basic registers of MII PHY #1: 3000 782d 02a8 0154 05e1 41e1 0001 0000.
Basic mode control register 0x3000: Auto-negotiation enabled.
You have link beat, and everything is working OK.
Your link partner can do 41e1: 100baseTx-FD 100baseTx 10baseT-FD 10baseT.
Basic registers of MII PHY #1: 3000 782d 02a8 0154 05e1 41e1 0003 0000.
Basic mode control register 0x3000: Auto-negotiation enabled.
You have link beat, and everything is working OK.
Your link partner can do 41e1: 100baseTx-FD 100baseTx 10baseT-FD 10baseT.
Diffs between good and bad, the 0001 is what the eth0 interface reports, and
this
interface is plugged in, but not configured/up, originally eth1 reports 0003
but after
the problem occurs, it reports 0001.
5c5
< Basic registers of MII PHY #1: 3000 782d 02a8 0154 05e1 41e1 0003 0000.
---
> Basic registers of MII PHY #1: 3000 782d 02a8 0154 05e1 41e1 0001 0000.
pci-config.c:v2.02 1/8/2001 Donald Becker (becker@scyld.com)
http://www.scyld.com/diag/index.html
Device #1 at bus 0 device/function 0/0, 00091166.
Device #2 at bus 0 device/function 0/1, 00091166.
Device #3 at bus 0 device/function 1/0, 47521002.
Device #4 at bus 0 device/function 4/0, 12298086.
Device #5 at bus 0 device/function 5/0, 12298086.
Device #6 at bus 0 device/function 15/0, 02001166.
Device #7 at bus 0 device/function 15/1, 02111166.
Device #8 at bus 0 device/function 15/2, 02201166.