Hi

I have a Netgear FA311 revision A1 network card in my NetBSD/i386 1.5.2 (aka duron) machine, connected to a FreeBSD 4.4R (aka pentium) machine with an identical card via a Netgear hub.  I cannot ping 'pentium' from 'duron' or vice versa.  However, the network works fine when I boot the NetBSD machine into Win2k.  I haven't managed to find a solution to my problem in the list archives, although others have had similar problems.

The card is probed successfully by NetBSD during bootup:

> sip0 at pci0 dev 10 function 0: NatSemi DP83815 10/100 Ethernet
> sip0: interrupting at irq 11
> sip0: Ethernet address 00:02:e3:22:75:89
> ukphy0 at sip0 phy 0: Generic IEEE 802.3u media interface
> ukphy0: OUI 0x080017, model 0x0002, rev. 1
> ukphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto

/etc/ifconfig.sip0 contains:

inet 10.0.0.1 netmask 255.0.0.0 media auto
ifconfig -a reports:
> duron% ifconfig -a
> sip0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
>         address: 00:02:e3:22:75:89
>         media: Ethernet autoselect (100baseTX)
>         status: active
>         inet 10.0.0.1 netmask 0xff000000 broadcast 10.255.255.255

Pinging the FreeBSD machine, 10.0.0.2, from the NetBSD machine, 10.0.0.1, reports:

> duron% ping 10.0.0.2
> PING pentium (10.0.0.2): 56 data bytes
> ping: sendto: Host is down
> ping: sendto: Host is down
> ping: sendto: Host is down> ---8<---------------------
Similarly, pinging duron from pentium reports:

> pentium% ping 10.0.0.1
> ping: sendto: Host is down
> ping: sendto: Host is down
> ping: sendto: Host is down
> ---8<---------------------

The tcpdump output on the NetBSD machine during the ping is:

> duron# tcpdump
> tcpdump: listening on sip0
> 22:02:01.363112 arp who-has pentium tell duron
> 22:02:02.370135 arp who-has pentium tell duron
> 22:02:03.380141 arp who-has pentium tell duron
> 22:02:04.380144 arp who-has pentium tell duron
> 22:02:05.380135 arp who-has pentium tell duron

Ending tcpdump indicates that the correct number of packets were received and that 0 were dropped during the ping.  The activity light on the hub flashes a few times but stops when ping reports that the host is down.

tcpdump on the FreeBSD machine gives the following output when it is pinged by the NetBSD machine:

> pentium# tcpdump
> tcpdump: listening on sis0
> 22:04:22.487905 arp who-has pentium tell duron
> 22:04:22.487935 arp reply pentium is-at 0:2:e3:22:75:8d
> 22:04:23.487986 arp who-has pentium tell duron
> 22:04:23.488006 arp reply pentium is-at 0:2:e3:22:75:8d
> 22:04:24.488073 arp who-has pentium tell duron
> 22:04:24.488091 arp reply pentium is-at 0:2:e3:22:75:8d
> 22:04:25.488145 arp who-has pentium tell duron
> 22:04:25.488162 arp reply pentium is-at 0:2:e3:22:75:8d
> 22:04:26.488228 arp who-has pentium tell duron
> 22:04:26.488247 arp reply pentium is-at 0:2:e3:22:75:8d
Conversely, when pinging NetBSD from FreeBSD, the output of tcpdump on the FreeBSD machine is:

pentium# tcpdump
tcpdump: listening on sis0
22:23:16.520162 pentium > duron: icmp: echo request
22:23:17.528083 pentium > duron: icmp: echo request
22:23:18.538091 pentium > duron: icmp: echo request
22:23:19.548112 pentium > duron: icmp: echo request
22:23:20.558129 pentium > duron: icmp: echo request
22:23:21.568140 pentium > duron: icmp: echo request
22:23:22.578155 pentium > duron: icmp: echo request
22:23:23.588166 pentium > duron: icmp: echo request

The output of tcpdump on the NetBSD machine when it is being pinged is:

> duron# tcpdump
> tcpdump: listening on sip0
> ^C   
> 0 packets received by filter
> 0 packets dropped by kernel
It seems that the NetBSD machine can send packets, but not receive them.

I would most appreciate any suggestions on how to remedy this problem.
In lieu of that, suggestions for well supported 10/100Mbit PCI cards in the US$30 (AU$60) region would be nice :-)

Many thanks (& sorry about the long post)

Cale Pearce
foo@graffiti.net
-- 

_______________________________________________
Get your free email from http://www.graffiti.net

Powered by Outblaze