[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Weird NIC behaviour


I'm having a weird problem with a NIC in an OpenBSD 2.8 router. I have a /30
subnet on one port on a 4way NIC (D-Link 570TX). I have a Linux box
connected at the other end via a crossover cable. This has worked without a
hitch until today when the Linux box lost power. After that I am unable to
connect the to machines again. Config is:

(OpenBSD) de0 
IP:	 	xx.yy.158.9
BROADCAST:	xx.yy.158.11
NETWORK:	xx.yy.158.8

(Linux) eth1
IP:		xx.yy.158.10
BROADCAST:      xx.yy.158.11
NETWORK:        xx.yy.158.8

>From the OpenBSD box:

ifconfig -a shows:
        media: Ethernet autoselect (100baseTX)
        status: active
        inet xx.yy.158.9 netmask 0xfffffffc broadcast

netstat -rn shows:
xx.yy.158.8/30     link#3             UC          0        0   1500  de0
xx.yy.158.9        0:80:c8:ca:b8:31   UHL         0    25339   1500  lo0
xx.yy.158.10       0:2:b3:1c:2a:23    UHL         0       10   1500  de0

ping shows:
stone# ping xx.yy.158.10
PING xx.yy.158.10 (xx.yy.158.10): 56 data bytes
ping: sendto: No route to host
ping: wrote xx.yy.158.10 64 chars, ret=-1

I have tried to delete the MAC address in insert it manually. I have taken
the NIC down and up again. I have changed cables. Last a ran a tcpdump on
the OpenBSD box. It shows that pings from the Linux box gets through
(arp lookups). Tcpdump on the Linux box shows no traffic at all. I have
tried to change the mediaopt to full-duplex and it didn't help.

I am at loss here. The way I see everything should work and it used to work
before the poweroutage. If anyone has any ideas on how to fix this or better
an explanation to what is wrong...