From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.6) Gecko/20040510 Description of problem: After an upgrade from FC1 (AMD64), my 3Com Corporation 3c940 10/100/1000Base-T [Marvell] (rev 12) no longer works. It's the onboard ethernet controller on the Albatron K8X800 mainboard. From /var/log/messages: May 22 00:18:46 lisa kernel: sk98lin: Network Device Driver v6.23 May 22 00:18:46 lisa kernel: (C)Copyright 1999-2004 Marvell(R). May 22 00:18:46 lisa kernel: eth0: 3Com Gigabit LOM (3C940) May 22 00:18:46 lisa kernel: PrefPort:A RlmtMode:Check Link State May 22 00:18:46 lisa kernel: Module sk98lin cannot be unloaded due to unsafe usage in drivers/net/sk98lin/skge.c:792 Here is what it used to look like (FC1): May 21 20:47:01 lisa kernel: sk98lin: Network Device Driver v6.15 May 21 20:47:01 lisa kernel: (C)Copyright 1999-2003 Marvell(R). May 21 20:47:01 lisa kernel: eth1: 3Com Gigabit LOM (3C940) May 21 20:47:01 lisa kernel: PrefPort:A RlmtMode:Check Link State [...] May 21 20:47:02 lisa kernel: sk98lin: Network Device Driver v6.15 May 21 20:47:02 lisa kernel: (C)Copyright 1999-2003 Marvell(R). May 21 20:47:02 lisa kernel: eth1: 3Com Gigabit LOM (3C940) May 21 20:47:02 lisa kernel: PrefPort:A RlmtMode:Check Link State May 21 20:47:02 lisa kernel: eth1: network connection up using port A May 21 20:47:02 lisa kernel: speed: 1000 May 21 20:47:02 lisa kernel: autonegotiation: yes May 21 20:47:02 lisa kernel: duplex mode: full May 21 20:47:02 lisa kernel: flowctrl: symmetric May 21 20:47:02 lisa kernel: role: slave May 21 20:47:02 lisa kernel: scatter-gather: enabled Version-Release number of selected component (if applicable): 2.6.5-1.358 How reproducible: Always Steps to Reproduce: 1. Boot 2. 3. Actual Results: No network! Expected Results: Network! Additional info:
May 22 00:18:46 lisa kernel: eth0: 3Com Gigabit LOM (3C940) May 22 00:18:46 lisa kernel: PrefPort:A RlmtMode:Check Link State well it found eth0... while it was eth1 in fc1... are you sure the networking config for eth0 is right ?
Yes, sorry, but I'm sure. I had to change my network configuration in order to file the bug. :-) (I have an RTL8029 PCI card in the system as well, connecting just one old Amiga 4000 via coax. This card still works, so I swapped configurations and the cable, and now use this card as my main interface towards the rest of the LAN and the internet.)
Just tried to use sk98lin on a Asus A7N8X-E motherboard, which has an onboard Gigabit interface in addition to the one provided by the NForce chipset. With kernel 2.6.5-358 it works ok. However, with kernel 2.6.6-1-435.2.3 it does not. The module loads, in dmesg the following message appears: sk98lin: Network Device Driver v6.23 (C)Copyright 1999-2004 Marvell(R). divert: allocating divert_blk for eth0 eth0: Yukon Gigabit Ethernet 10/100/1000Base-T Adapter PrefPort:A RlmtMode:Check Link State Module sk98lin cannot be unloaded due to unsafe usage in drivers/net/sk98lin/skge.c:792 The link with the switch is established, but there is absolulety no traffic through the interface, not even ARP. lspci identifies the card as: 01:04.0 Ethernet controller: Marvell Yukon Gigabit Ethernet 10/100/1000Base-T Adapter (rev 13) As i sidenote, I have the feeling that the new kernel swaps eth0 and eth1. I have tried all possible configurations, including switching cables, but the Gigabit intetface seems to not allow traffic through.
I have noticed similar symptoms on a (infamous) Asus P4P800 motherboard with a 3Com Gigabit LOM (3C940) onboard ethernet interface. It goes like this: When booting with kernels up to and including 2.6.6-1.435.2.3 there are no problems, or any logged errors or warnings. With kernels 2.6.7-1.494.2.2 and 2.6.8-1.521 I still do not see any errors or obvious problems, however the interface's performance drops significantly: tcp throughput is ridiculously slow, down to a crawl, always on the same LAN. This behavior occurs both with SMP and UP kernels. The only other thing I notice is that lsmod shows that sk98lin is used by [unsafe] and can not be unloaded. I hope this helps, even if it's not so specific...
Fedora Core 2 has now reached end of life, and no further updates will be provided by Red Hat. The Fedora legacy project will be producing further kernel updates for security problems only. If this bug has not been fixed in the latest Fedora Core 2 update kernel, please try to reproduce it under Fedora Core 3, and reopen if necessary, changing the product version accordingly. Thank you.
Yup, FC3 did not have this problem (at least not on my system).