[eepro100] which driver to use?

[Jan Just Keijser]

Hi all,

I've been reading up on eepro100 drivers for a while and I've been
getting pretty confused as of late; which eepro100.c driver should I
use? For uni-processor boxes? For SMP boxes?

Here's the problem
- PC's with dual eepro100 adapters (lspci reports 82557 Rev 8)
- Linux 2.2.16-3 kernel from RedHat

We've been seeing wait_for_cmd timeout messages pretty frequently, esp.
out at a customer site, of course and I've been trying to find a working
driver; I've upgraded to the Linux 2.2.18 eepro100.c driver (with the VA
Linux patch), and for a while things were looking better, but now I am
getting:

Mar  8 19:08:32 localhost kernel: eepro100.c:v1.09j-t 9/29/99 Donald
Becker http://cesdis.gsfc.nasa.gov/linux/drivers/eepro100.html
Mar  8 19:08:32 localhost kernel: eepro100.c: $Revision: 1.20.2.10 $
2000/05/31 Modified by Andrey V. Savochkin <saw@saw.sw.com.sg> and
others
Mar  8 19:08:32 localhost kernel: eepro100.c: VA Linux custom, Dragan
Stancevic <visitor@valinux.com> 2000/11/15
Mar  8 19:08:32 localhost kernel: eth0: Intel PCI EtherExpress Pro100
82557, 00:04:9A:49:3E:16, IRQ 9.
Mar  8 19:08:32 localhost kernel:   Board assembly 668081-002, Physical
connectors present: RJ45
Mar  8 19:08:32 localhost kernel:   Primary interface chip i82555 PHY
#1.
Mar  8 19:08:32 localhost kernel:   General self-test: passed.
Mar  8 19:08:32 localhost kernel:   Serial sub-system self-test: passed.

Mar  8 19:08:32 localhost kernel:   Internal registers self-test:
passed.
Mar  8 19:08:32 localhost kernel:   ROM checksum self-test: passed
(0x04f4518b).
Mar  8 19:08:32 localhost kernel:   Receiver lock-up workaround
activated.
Mar  8 19:08:32 localhost kernel: eth1: Intel PCI EtherExpress Pro100
82557, 00:04:9A:49:3E:17, IRQ 10.
Mar  8 19:08:32 localhost kernel:   Board assembly 668081-002, Physical
connectors present: RJ45
Mar  8 19:08:32 localhost kernel:   Primary interface chip i82555 PHY
#1.
Mar  8 19:08:32 localhost kernel:   General self-test: passed.
Mar  8 19:08:32 localhost kernel:   Serial sub-system self-test: passed.

Mar  8 19:08:32 localhost kernel:   Internal registers self-test:
passed.
Mar  8 19:08:32 localhost kernel:   ROM checksum self-test: passed
(0x04f4518b).
Mar  8 19:08:32 localhost kernel:   Receiver lock-up workaround
activated.
Mar  8 19:08:32 localhost kernel: eepro100: cmd_wait for(0x70) timedout
with(0x70)!
Mar  8 19:08:32 localhost kernel: eepro100: cmd_wait for(0x10) timedout
with(0x10)!
Mar  8 19:08:32 localhost kernel: eepro100: cmd_wait for(0xffffff90)
timedout with(0xffffff90)!
Mar  8 19:08:33 localhost last message repeated 53 times
Mar  8 19:08:32 localhost crond: crond startup succeeded
Mar  8 19:08:33 localhost kernel: eepro100: cmd_wait for(0xffffff90)
timedout with(0xffffff90)!

and we're actually worse off than before. I've downloaded the v1.13
driver from www.scyld.com and compiled it and things are now looking OK
again (different box, same hardware):

Mar 13 09:48:11 dhcp-89-2-190 kernel: eepro100.c:v1.13 1/9/2001 Donald
Becker <becker@scyld.com>
Mar 13 09:48:11 dhcp-89-2-190 kernel:
http://www.scyld.com/network/eepro100.html
Mar 13 09:48:11 dhcp-89-2-190 kernel: eth0: Intel PCI EtherExpress
Pro100 at 0xd084a000, 00:04:9A:49:3D:00, IRQ 9.
Mar 13 09:48:11 dhcp-89-2-190 kernel:   Board assembly 668081-002,
Physical connectors present: RJ45
Mar 13 09:48:11 dhcp-89-2-190 kernel:   Primary interface chip i82555
PHY #1.
Mar 13 09:48:11 dhcp-89-2-190 kernel:   General self-test: passed.
Mar 13 09:48:11 dhcp-89-2-190 kernel:   Serial sub-system self-test:
passed.
Mar 13 09:48:11 dhcp-89-2-190 kernel:   Internal registers self-test:
passed.
Mar 13 09:48:11 dhcp-89-2-190 kernel:   ROM checksum self-test: passed
(0x04f4518b).
Mar 13 09:48:11 dhcp-89-2-190 kernel:   Receiver lock-up workaround
activated.
Mar 13 09:48:11 dhcp-89-2-190 kernel: eth1: Intel PCI EtherExpress
Pro100 at 0xd084c000, 00:04:9A:49:3D:01, IRQ 10.
Mar 13 09:48:11 dhcp-89-2-190 kernel:   Board assembly 668081-002,
Physical connectors present: RJ45
Mar 13 09:48:11 dhcp-89-2-190 kernel:   Primary interface chip i82555
PHY #1.
Mar 13 09:48:11 dhcp-89-2-190 kernel:   General self-test: passed.
Mar 13 09:48:11 dhcp-89-2-190 kernel:   Serial sub-system self-test:
passed.
Mar 13 09:48:11 dhcp-89-2-190 kernel:   Internal registers self-test:
passed.
Mar 13 09:48:11 dhcp-89-2-190 kernel:   ROM checksum self-test: passed
(0x04f4518b).
Mar 13 09:48:11 dhcp-89-2-190 kernel:   Receiver lock-up workaround
activated.

but...
- am I using the right driver now, finally?
- can I use this driver for SMP boxes as well?
- how can I know that the problem *really* is fixed now - this thing is
so intermittent!
- is there a test to see if my hardware is affected (because as I
understand we're looking at an Intel hardware boo-boo)?

thanks for any pointers, suggestions, help,

JJK / Jan Just Keijser
Cisco Systems Inc
Amsterdam Office