[Xen-users] compatibility problem with xen 3.1 and xen 3.3

["Jia Rao" <rickenrao@xxxxxxxxx>]

Hi all,

I have several VMs created under xen 3.1. The VMs were created by copying file system in dom0 into domU. Everything was OK.

I just downloaded xen 3.3 and want to have a try. The xen 3.3 was compiled from source and dom0 was up with no problem.
When I copy the disk images to the new machine with xen 3.3. There is a problem with xm control command.

When I tried to change the number of VCPU in domU by using "xm vcpu-set domainID 2" there was a problem.
The domU was set to 4 VCPUs at start up. after I set VCPU to 2, the VCPU appears to be 3 in "xm list" and meanwhile, domU received following error:

 Unable to handle kernel NULL pointer dereference at 0000000000000018 RIP:
 [<ffffffff802185c2>] cacheinfo_cpu_callback+0x389/0x3f0
PGD 157e2067 PUD 0
Oops: 0002 [1] SMP
CPU 0
Modules linked in:
Pid: 2011, comm: xenwatch_cb Not tainted 2.6.18.8-xen #2
RIP: e030:[<ffffffff802185c2>]  [<ffffffff802185c2>] cacheinfo_cpu_callback+0x389/0x3f0
RSP: e02b:ffff88001e2dbdb0  EFLAGS: 00010297
RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
RDX: 0000000000000003 RSI: ffff88001f4b9b78 RDI: 0000000000000002
RBP: 0000000000000002 R08: 0000000000000007 R09: ffff88001f548578
R10: 0000000000000000 R11: ffffffff802b491c R12: ffff88001f4b9b60
R13: 0000000000000000 R14: 0000000000000000 R15: ffffffff80242c9d
FS:  00002aab51da7b80(0000) GS:ffffffff804d7000(0000) knlGS:0000000000000000
CS:  e033 DS: 0000 ES: 0000
Process xenwatch_cb (pid: 2011, threadinfo ffff88001e2da000, task ffff88001f44d100)
Stack:  ffff8800057550c0 0000000000000002 ffffffff80544308 ffffffff80242c57
 ffff88000567e9e0 ffffffff80470940 0000000000000002 0000000000000007
 0000000000000007 ffff880005767d80 ffffffff80242c9d ffffffff8023c9df
Call Trace:
 [<ffffffff80242c57>] kthread_stop+0x75/0x79
 [<ffffffff80242c9d>] keventd_create_kthread+0x0/0x61
 [<ffffffff8023c9df>] notifier_call_chain+0x20/0x32
 [<ffffffff8023cec4>] blocking_notifier_call_chain+0x22/0x36
 [<ffffffff80248da4>] cpu_down+0x165/0x255
 [<ffffffff8034ef8a>] vcpu_hotplug+0xb3/0xd1
 [<ffffffff80226d55>] __wake_up_common+0x3e/0x68
 [<ffffffff802eb70b>] strstr+0x4f/0x69
 [<ffffffff803522ec>] xenwatch_handle_callback+0x0/0x48
 [<ffffffff8034f08d>] handle_vcpu_hotplug_event+0x3a/0x3f
 [<ffffffff80352301>] xenwatch_handle_callback+0x15/0x48
 [<ffffffff80242f13>] kthread+0xd4/0x109
 [<ffffffff8020afe0>] child_rip+0xa/0x12
 [<ffffffff80242c9d>] keventd_create_kthread+0x0/0x61
 [<ffffffff80242e3f>] kthread+0x0/0x109
 [<ffffffff8020afd6>] child_rip+0x0/0x12


Code: f0 0f b3 68 18 49 8d 74 24 18 89 d7 e8 28 02 0d 00 83 f8 1f
RIP  [<ffffffff802185c2>] cacheinfo_cpu_callback+0x389/0x3f0
 RSP <ffff88001e2dbdb0>
CR2: 0000000000000018

There is nothing wrong in xm dmesg and xend.log
I thought there might be some compatiblity issue in xen 3.1 and xen 3.3. Then I recompile xen 3.3 within the VM, allowing the VM to have the same modules and kernels with the dom0. But it didnt work.

FYI:
OS: centOS 5.1
Kernel: 2.6.18-8
Xen: 3.3
CPU: two quad-core Intel cpu.
Box: Dell PowerEdge 1950

Thanks,
Jia

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users