Bug 2248

Summary:	(net 3c59x) Cyclone Cards just recognized on cold boot not on restart
Product:	Drivers	Reporter:	Haldir
Component:	Network	Assignee:	Jeff Garzik (jgarzik)
Status:	CLOSED PATCH_ALREADY_AVAILABLE
Severity:	blocking
Priority:	P2
Hardware:	i386
OS:	Linux
Kernel Version:	2.6.3	Subsystem:
Regression:	---	Bisected commit-id:

Description Haldir 2004-03-04 09:52:12 UTC

Distribution: Gentoo
Hardware Environment: Dual P3 800 on Via Chipset
Software Environment:
Problem Description: My 3Com Cyclone (905B) cards are just recognized properly
after booting from power-down, a normal restart fails
Logs after Fail:
Mar  1 19:04:40 [kernel] 3c59x: Donald Becker and others.
www.scyld.com/network/vortex.html
Mar  1 19:04:40 [kernel] 0000:00:0b.0: 3Com PCI 3c905B Cyclone 100baseTx at
0x1000. Vers LK1.1.19
Mar  1 19:04:40 [kernel] PCI: Setting latency timer of device 0000:00:0b.0 to 64
Mar  1 19:04:40 [kernel]   ***WARNING*** No MII transceivers found!

(this happens for each card)
Subsequent actions on the card fail like:

Mar  1 19:04:44 [kernel] eth0: command 0x5800 did not complete! Status=0xffff
Mar  1 19:04:45 [kernel] eth0: command 0x2804 did not complete! Status=0xffff

Additionally the HWID (MAC) of the eth0 devices are FF:FF:FF:FF

Logs for "working" boot:
Mar  4 18:28:15 [kernel] 3c59x: Donald Becker and others.
www.scyld.com/network/vortex.html
Mar  4 18:28:15 [kernel] 0000:00:0b.0: 3Com PCI 3c905B Cyclone 100baseTx at
0xdc00. Vers LK1.1.19
Mar  4 18:28:15 [kernel] 0000:00:0c.0: 3Com PCI 3c905B Cyclone 100baseTx at
0xe000. Vers LK1.1.19

No further error Messages, MAC ID is correct

This happens since I updated to 2.6.3 Kernel (didn't happen in 2.4.2x)
Steps to reproduce:
SMP Kernel 2.6.3 with 3Com drivers compiled in for 905B and MII support compiled in.
After reboot fails with the errors above, after complete shutdown with power off    
the cards work properly.

Comment 1 Chris Lent 2004-11-23 06:32:54 UTC

Seems like this might be related to:
(net 3c59x) problems with WOL changes
Bug#:   	 1394
http://bugzilla.kernel.org/show_bug.cgi?id=1394

Comment 2 Alexander Nyberg 2004-12-06 13:50:11 UTC

In mail exchange Haldir tells me this is no longer a problem in 2.6.8, closing.