Bug 148470 - XEN: eth0 doesn't work (MAC 00:00:00:00:00:00)
Summary: XEN: eth0 doesn't work (MAC 00:00:00:00:00:00)
Status: VERIFIED FIXED
Alias: None
Product: SUSE Linux 10.1
Classification: openSUSE
Component: Xen (show other bugs)
Version: Beta 2
Hardware: Other Other
: P5 - None : Normal (vote)
Target Milestone: ---
Assignee: Kirk Allan
QA Contact: E-mail List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2006-02-06 18:41 UTC by Harald Koenig
Modified: 2006-06-15 19:04 UTC (History)
0 users

See Also:
Found By: Other
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Harald Koenig 2006-02-06 18:41:39 UTC
when booting 10.1-beta2 with "XEN" boot label, eth0 in dom0 doesn't work
(no problem without XEN, eth0 is setup with static IP address).  reason is
that eth0 now claims to have MAC 00:00:00:00:00:00.
ifconfig output:
        eth0      Protokoll:Ethernet  Hardware Adresse 00:00:00:00:00:00
        eth1      Protokoll:Ethernet  Hardware Adresse 00:30:48:75:61:F5

it's not possible to (re)set the hwaddr using "ifconfig eth0 hw ether ....",
I always get "device busy" even after "ifconfig eth0 down".

"rmmod tg3 ; modprobe tg3" doesn't help either, but gives correct MAC addr
in kernel messages:

        tg3.c:v3.48 (Jan 16, 2006)
        GSI 19 sharing vector 0xC1 and IRQ 19
        ACPI: PCI Interrupt 0000:02:05.0[A] -> GSI 26 (level, low) -> IRQ 19
        eth0: Tigon3 [partno(BCM95704A6) rev 2100 PHY(5704)] (PCIX:100MHz:64-bit) 10/100/1000BaseT Ethernet 00:30:48:75:61:f4
        eth0: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[1] Split[0] WireSpeed[1] TSOcap[0]
        eth0: dma_rwctrl[769f4000]
        GSI 20 sharing vector 0xC9 and IRQ 20
        ACPI: PCI Interrupt 0000:02:05.1[B] -> GSI 27 (level, low) -> IRQ 20
        eth1: Tigon3 [partno(BCM95704A6) rev 2100 PHY(5704)] (PCIX:100MHz:64-bit) 10/100/1000BaseT Ethernet 00:30:48:75:61:f5
        eth1: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[0] Split[0] WireSpeed[1] TSOcap[1]
        eth1: dma_rwctrl[769f4000]

so I did not manage to get any network connection with XEN :-((
Comment 1 Harald Koenig 2006-02-08 12:35:21 UTC
update with eth0 setup in XEN:

beta-3 XEN x86_64 always crashed at startup of dom0, so I switched to KOTD kernel-xen-2.6.16_rc2_git2-20060206180157 (see bug #148719).

now with KOTD xen kernel, eth0 setup seems better, but still very flaky!
the following happened with various boots of xen dom0, without any changes in system setup, cabeling, whatever:

1st boot:  eth0 mac address ok, but network not working.  dhcpcd reports "cable is connected" :

              No configuration found for vif0.7
unusedWaiting for mandatory devices:  eth-id-00:30:48:75:61:f4 __NSC__
15 13 11 
    eth0      device: Broadcom Corporation NetXtreme BCM5704 Gigabit Ethernet (rev 10)
    eth0      configuration: eth-id-00:30:48:75:61:f4
    eth0      ifplugd is running
    eth0      cable is connected
    eth0      DHCP client NOT running
waiting
Setting up service network  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .done



2nd boot:  network not working.  network startup looks for (non-existing) config for "eth0" instead of "eth-id-..."

this time, the MAC address of eth0 is 00:00:00:00:00:00 again:-(  but now (unline beta-2) I can set the correct MAC using

    ifconfig eth0 hw ether   00:30:48:75:61:f4

but restarting dhcpcd doesn't help, I only get Tx-packets, but no packets are received.  ifconfig eth0 shows:

eth0      Protokoll:Ethernet  Hardware Adresse 00:30:48:75:61:F4  
          inet6 Adresse: fe80::230:48ff:fe75:61f4/64 Gültigkeitsbereich:Verbindung
          UP BROADCAST NOTRAILERS RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:13 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 Sendewarteschlangenlänge:0 
          RX bytes:0 (0.0 b)  TX bytes:4598 (4.4 Kb)



3rd boot:  now,  MAC address is ok, dhcpcd says "cable is connected" and network start works "out of the box"



any ideas what's going wrong here ?

which addition information do you need for debugging ?


thanks for any help!!
Comment 2 Kirk Allan 2006-02-08 23:53:06 UTC
I tried this on SULE Linux 10.1 Beta 2 for x86_64 with different LAN adapters and got the same results: eth0 had MAC of 00:00:00:00:00:00.

Try going into YaST, select “Network Device”, “Network Card”, and edit the card corresponding to eth0.  Select the “General” tab and make sure the “Device Activation” is set to “At Boot Time”.  Mine was set to “On Cable Connection” by default.  Exit out of YaST and reboot back into Xen.

This fixed the problem for me.  eth0 received its correct MAC address and networking worked correctly.  Let me know if this fixes the problem.
Comment 3 Harald Koenig 2006-02-14 08:56:56 UTC
(In reply to comment #2)

> Try going into YaST, select “Network Device”, “Network Card”, and edit
> the card corresponding to eth0.  Select the “General” tab and make sure the
> “Device Activation” is set to "At Boot Time”.  Mine was set to “On
> Cable Connection” by default.  Exit out of YaST and reboot back into Xen.
> 
> This fixed the problem for me.  eth0 received its correct MAC address and
> networking worked correctly.  Let me know if this fixes the problem.

sorry for the late reply.  using "At Boot Time" network setup seems to be a workaround, the system booted with correct network setup 2 or 3 times now.

but obviously for the future (even GM?) it would be nice if one woul dnot have to know about this special problems and "On Cable Connection" would work within dom0 as usual (e.g. for notebooks with optional xen boot...)


thanks for the pointer -- this is a great help for now!
Comment 4 Kirk Allan 2006-03-17 15:30:51 UTC
The default device activation has be changed to "At Boot Time" in SLES 10 Beta8.  Also the xen network scripts have been fixed to allow proper functionality if the NIC is configured to "On Cable Connection" in Beta8.