Fedora xen - Jan 2006 - Hang when starting xenU using riel''s packages

Hi all,

I am using kernel-xenU, kernel-xen0, and xen-3 from 
http://people.redhat.com/~riel/xen_for_fc4/. xen0 boots up fine, but when I 
start xenU, it hangs at the below screen. I have already moved /lib/tls out 
of the way.
---------
  ***************************************************************
  ***************************************************************
  ** WARNING: Currently emulating unsupported memory accesses  **
  **          in /lib/tls libraries. The emulation is very     **
  **          slow. To ensure full performance you should      **
  **          execute the following as root:                   **
  **          mv /lib/tls /lib/tls.disabled                    **
  ** Offending process: init (pid=1)                           **
  ***************************************************************
  ***************************************************************

Continuing...
----------

My xen config file is as follows:

kernel ="/boot/vmlinuz-2.6.12-1.1454_FC4xenU"
memory = 128
name = "waipeng"
nics = 1
disk = [''phy:/servers/waipeng,hda1,w'']
root = "/dev/hda1"
extra = "ro selinux=0 3

The log shows this:

[2006-01-19 23:43:17 xend] DEBUG (image:156) initDomain: cpu=-1 mem_kb=131072 
ssidref=-1 dom=3
[2006-01-19 23:43:17 xend] DEBUG (XendDomainInfo:843) init_domain> Created 
domain=3 name=waipeng memory=128
[2006-01-19 23:43:17 xend] INFO (XendRoot:141) EVENT> xend.domain.create 
[''waipeng'', 3]
[2006-01-19 23:43:17 xend] INFO (image:217) buildDomain os=linux dom=3 vcpus=1
[2006-01-19 23:43:17 xend] DEBUG (image:255) dom            = 3
[2006-01-19 23:43:17 xend] DEBUG (image:256) image          
= /boot/vmlinuz-2.6.12-1.1454_FC4xenU
[2006-01-19 23:43:17 xend] DEBUG (image:257) store_evtchn   = 1
[2006-01-19 23:43:17 xend] DEBUG (image:258) console_evtchn = 2
[2006-01-19 23:43:17 xend] DEBUG (image:259) cmdline        =  root=/dev/hda1 
ro selinux=0 3
[2006-01-19 23:43:17 xend] DEBUG (image:260) ramdisk        [2006-01-19 23:43:17
xend] DEBUG (image:261) flags          = 0
[2006-01-19 23:43:17 xend] DEBUG (image:262) vcpus          = 1
[2006-01-19 23:43:17 xend] INFO (xsobj:326) DBMap>introduceDomain> 3 59700
<EventChannel dom1:0:17 
dom2:3:1> /domain/499d6865-9920-4f87-bc34-6fd4b1933bb3
[2006-01-19 23:43:17 xend] DEBUG (XendDomainInfo:698) writing key 
availability=online to path /cpu/0 in store
[2006-01-19 23:43:17 xend] WARNING (XendDomainInfo:1100) Unknown config field 
boot
[2006-01-19 23:43:17 xend] ERROR (XendDomainInfo:443) [''vif'',
[''mac'',
''aa:00:00:7d:9a:01'']]
[2006-01-19 23:43:17 xend] INFO (XendRoot:141) EVENT> xend.domain.unpause 
[''waipeng'', 3]

Questions:
1) Anybody have encountered this before? How did you managed to solve this?
2) Does the mac error matters? Why is it there in the first place?

I would appreciate any help given. Thanks!

Wai Peng

Hi,

On Thu, Jan 19, 2006 at 11:52:30PM +0800, Yip Wai Peng wrote:
 
> I am using kernel-xenU, kernel-xen0, and xen-3 from 
> http://people.redhat.com/~riel/xen_for_fc4/. xen0 boots up fine, but when I
> start xenU, it hangs at the below screen. I have already moved /lib/tls out
> of the way.
> ---------
>   ***************************************************************
>   ***************************************************************
>   ** WARNING: Currently emulating unsupported memory accesses  **
>   **          in /lib/tls libraries. The emulation is very     **
>   **          slow. To ensure full performance you should      **
>   **          execute the following as root:                   **
>   **          mv /lib/tls /lib/tls.disabled                    **
>   ** Offending process: init (pid=1)                           **
>   ***************************************************************
>   ***************************************************************
That message is OK --- this just turns up when the Xen kernel comes
across a statically linked binary.  The distro now has magic to
disable tls dynamically at run time for Xen kernels, but that depends
on running shared libraries, so static binaries still trigger the tls
warning.  We have disabled the tls warning in the fc5t2 xen kernels
for this reason.

The hang is more worrying ---  things are supposed to continue after a
5-second timeout after that message, so a hang implies that timers
aren''t working properly.  Can you reproduce this on fc5t2, which has a
MUCH newer xen HV/tools/kernel combination? 

Thanks,
 Stephen

Hi folks!

I also installed xen and it works fine for me.

Have you ever tried to wait some secoonds until boot continues?
My xen config file looks like:
----
kernel="/boot/vmlinuz-2.6-xen"
memory=640
name="fedora1"
nics=1
disk=[''file:/opt/xen/fedora1.img,sda1,w'' , 
file:/opt/xen/fedora1.swap.img,sda2,w'']
root="/dev/sda1"
extra="ro selinux=0 3"
hostname="fedora"
vcpus=1
----

Works just fine ....

Hope this helps
Wolfi