On 11/19/2012 03:36 PM, Richard Mortimer wrote:
[...]ok watch-net-all /pci@1d,700000/network@2,1 Timed out waiting for Autonegotiation to complete Check cable and try again Link Down /pci@1d,700000/network@2 100 Mbps full duplex Link up Looking for Ethernet Packets. '.' is a Good Packet. 'X' is a Bad Packet. Type any key to stop. .................................................................. /pci@1f,700000/network@2,1 Timed out waiting for Autonegotiation to complete Check cable and try again Link Down /pci@1f,700000/network@2 100 Mbps full duplex Link up Looking for Ethernet Packets. '.' is a Good Packet. 'X' is a Bad Packet. Type any key to stop.........................................................................................................So as I suspected eth0 and eth1 are the ports labelled net0 and net2Did you try connecting the cable to those ports and testing. From what I've see and what others have said I suspect that will just work.
I've connected cables to eth0 and eth1 or NICs 0 and 1. In addition I tried NIC 2 but because Debian doesn't show the card itself I wasn't able to get any connectivity.
... snip ...Additionally it looks like the 2nd network device on each PCI bus is not getting attached properly. Does dmesg | grep tg3[ 39.625716] tg3.c:v3.116 (December 3, 2010) [ 40.264438] tg3 0000:00:02.0: eth0: Tigon3 [partno(none) rev 2100] (PCI:66MHz:64-bit) MAC address 00:14:4f:5d:1e:7e [ 40.401775] tg3 0000:00:02.0: eth0: attached PHY is 5704 (10/100/1000Base-T Ethernet) (WireSpeed[1]) [ 40.521950] tg3 0000:00:02.0: eth0: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[0] TSOcap[1] [ 40.626026] tg3 0000:00:02.0: eth0: dma_rwctrl[763f0000] dma_mask[32-bit][ 41.439201] tg3 0000:00:02.1: BAR 2: can't reserve mem region [0x7f700000000-0x7f70000ffff] [ 41.549032] tg3 0000:00:02.1: Cannot obtain PCI resources, aborting [ 41.734027] tg3: probe of 0000:00:02.1 failed with error -16This explains why the other two ports are not showing up. A quick search shows up. http://www.spinics.net/lists/netdev/msg200255.html Not sure if that has been addressed in later kernel versions yet.
Hmm.... if this is the case at present as I predicted I am thinking of switching to FreeBSD for the time being to gain usage out of the other NICs. Both Solaris 10 and OpenBSD 5.1 work perfectly so as a temporary measure I could try that.
I know it's not something to be mentioning on a Debian mailing list however, if I have no other choice at present it is unfortunate :-(
[...] Regards, Kaya