Bug 213991

Summary: xen networking doen't bring bridge up with xen 3.0.3
Product: [Fedora] Fedora Reporter: Henning Schmiedehausen <hps>
Component: xenAssignee: Xen Maintainance List <xen-maint>
Status: CLOSED WONTFIX QA Contact: Martin Jenner <mjenner>
Severity: high Docs Contact:
Priority: medium    
Version: 5CC: bstein, katzj, rene
Target Milestone: ---   
Target Release: ---   
Hardware: i386   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2008-02-26 18:38:46 EST Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Attachments:
Description Flags
dmesg of the xen host
none
dmesg of the xen client none

Description Henning Schmiedehausen 2006-11-04 07:37:52 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.0.7) Gecko/20060930 Fedora/1.5.0.7-1.1.fc5.im Firefox/1.5.0.7

Description of problem:
After upgrading FC5 to the current versions of kernel and xen, the virtual hosts finally start again. However, the networking does not work. 

brctl show
xenbr0          8000.feffffffffff       no              peth3
                                                        vif0.0
                                                        vif2.0
                                                        vif2.1

if you run "ifconfig vif2.0 up", network starts working.

In the dmesg file of the host I found:

device vif2.0 entered promiscuous mode
audit(1162642195.512:11): dev=vif2.0 prom=256 old_prom=0 auid=4294967295
ADDRCONF(NETDEV_UP): vif2.0: link is not ready
xenbr0: port 3(vif2.0) entering disabled state


That seems to be the problem. 



Version-Release number of selected component (if applicable):
xen-3.0.3-1.fc5 kernel-2.6.18-1.2200.fc5 (xen0 and xenU)

How reproducible:
Always


Steps to Reproduce:
1. update FC5 with all patches
2. start a xen client with bridged networking
3. look at dmesg

Actual Results:
networking does not work

Expected Results:
the bridge should pick up the topology changes (client interface coming up) and networking should work. It worked up to 2187

Additional info:
Set prio to high because this breaks existing xen installations.
Comment 1 Henning Schmiedehausen 2006-11-04 07:39:05 EST
Created attachment 140350 [details]
dmesg of the xen host
Comment 2 Henning Schmiedehausen 2006-11-04 07:39:38 EST
Created attachment 140351 [details]
dmesg of the xen client
Comment 3 Rene Cunningham 2007-01-08 00:54:59 EST
Do you have a network interface that uses a Broadcom chipset?

I too had the same problem and i had to upgrade the firmware for my broadcom NIC. 

See the following thread that describes the problem.

http://lists.xensource.com/archives/html/xen-users/2006-08/msg00553.html

The following page (mywiki.ncsa.uiuc.edu was down at the time of writing this
comment, hence google cache) described how to upgrade the firmware.

http://72.14.253.104/search?q=cache:ZJlR5NbKrHIJ:mywiki.ncsa.uiuc.edu/wiki/Dell_PE1950_NIC_Firmware_Workaround+Dell_PE1950_NIC_Firmware_Workaround&hl=en&ct=clnk&cd=3

IBM had a patch released for the broadcom chipset which fixed this error for me.
Comment 4 Henning Schmiedehausen 2007-01-08 02:44:07 EST
No, I'm using Intel EEPro 100 cards. Actually the problem was that the
networking config no longer detects reliably to which a card a bridge is
connected. Adding the mac addresses explicitly now works. In my
/etc/xen/auto/<host> I have

vif = [ "mac=00:16:3e:3e:05:af,bridge=xenbr0",
"mac=00:16:3e:3e:05:ae,bridge=xenbr1" ]

and that works well. Without the mac=<xxx> parameters it does not. 

Seems to be more of a documentation issue than a real bug.
Comment 5 Red Hat Bugzilla 2007-07-24 20:02:32 EDT
change QA contact
Comment 6 Chris Lalancette 2008-02-26 18:38:46 EST
This report targets FC6, which is now end-of-life.

Please re-test against Fedora 7 or later, and if the issue persists, open a new bug.

Thanks