kern/153610: nfe0 malfunction at boot time
rfg at tristatelogic.com
Sat Jan 1 23:10:11 UTC 2011
>Synopsis: nfe0 malfunction at boot time
>Arrival-Date: Sat Jan 01 23:10:10 UTC 2011
>Originator: Ronald F. Guilmette
>Release: FreeBSD 8.1-RELEASE amd64
Infinite Monkeys & Co.
FreeBSD 8.1-RELEASE amd64
Gigabyte GA-m55plus-S3G nmotherboard w/ Nvidia 430 Southbridge
Onboard Ethernet port hardwired up to a stock Linksys WRT54G wireless
router configured for static (i.e. NO DHCP)
Upon boot-up of the GA-m55plus-S3G based system, I get an (unending?)
stream of console/syslog messages exactly like this:
nfe0: discard frame w/o leading ethernet header (len 0 pkt len 0)
After googling around awhile, I learned that this is an old, and apparently
still unsolved problem:
The thread above suggested that I try manually down-ing and then up-ing
the nfe0 interface. I did that, and yes, that causes the problem to go
away and after that up/down the interface seems to be functioning normally.
However this server SHOULD be able to recover gracefully from, e.g.,
power failures, even when I am out of town, so what am I supposed to do?
Should I add "manual" down & up commands for this interface to my
Ideally, I shouldn't have to do that (and I'm not even 100% sure that even
doing that will produce reliable results).
See above. Get yourself a motherboard with an Nvidia 430 Southbridge
on it, and then wire up the ethernet port from that to a stock Linksys
WRT54G, set the Linksys box to NO DHCP (although that probably doesn't
matter), give the nfe0 device a static IP in your /etc/rc.conf file
(although that probably doesn't matter either), and then do a cold restart.
Beats me. I am forced to workaround this for now by installing a cheapie
Realtek ethernet card (which most certainly works).
That's a pity, because I wanted to try using the gigabit capability of the
onboard Nvidia ethernet port. Oh well. I'll live.
More information about the freebsd-bugs