Re: [Xen-users] CPU lockup detected in dom0

[Reiner Dassing <dassing@xxxxxxxxxxxxxxxx>]

Hi Hugo,

I had a simliar problem and I played around to get an idea
of the reason "BUG: soft lockup detected on CPU#0".

Summary:
The error may be produced by enabling SMP during kernel build
and no SMP processor is there,just a single processor.

Well, I had the following:
We have here two DELL PowerEdge Servers.
I installed Xen-3.0.2 on the first, made it running and put it in the
production state(!).
Everything was and is fine.

Then I bought another one and copied the .config file to the new one.
During generation of the DEBIAN operating system I got the mentioned
bug. This can be reproduced!

After disabling the SMP capability of the kernel on the new machine,
which just have a single processor, the kernel build was successful.

Regards

Reiner


 Rodenburg wrote:

> Hi list,
>  
> I've been playing around with Xen for a few months now, and I'm very 
> excited about this product.
> A few days ago I started a fresh install on an Compaq Proliant DL580. 
> It's the first generation of this machine with 4x p3 700 Mhz (1mb cache btw)
> The dom0 is using Debian 3.1 (sarge) and works perfectly.
> Only after installing xen, and booting xen, I'm getting a _lot_ of the 
> following messages after a while:
>  
> Apr 17 19:31:38 pgcore02 kernel: BUG: soft lockup detected on CPU#3!
> Apr 17 19:31:38 pgcore02 kernel:
> Apr 17 19:31:38 pgcore02 kernel: Pid: 0, comm:              swapper
> Apr 17 19:31:38 pgcore02 kernel: EIP: 0061:[hypercall_page+935/4096] CPU: 3
> Apr 17 19:31:38 pgcore02 kernel: EIP is at 0xc01013a7
> Apr 17 19:31:38 pgcore02 kernel:  EFLAGS: 00000292    Not tainted  
> (2.6.16-xen0-pgcore-1-xen3testing #1)
> Apr 17 19:31:38 pgcore02 kernel: EAX: 00000000 EBX: 00000001 ECX: 
> 00000000 EDX: 00000018
> Apr 17 19:31:38 pgcore02 kernel: ESI: 00000000 EDI: 00000001 EBP: 
> c0646000 DS: 007b ES: 007b
> Apr 17 19:31:38 pgcore02 kernel: CR0: 8005003b CR2: b7f2f000 CR3: 
> 0f8ec000 CR4: 00000640
> Apr 17 19:31:38 pgcore02 kernel:  [xen_idle+83/176] xen_idle+0x53/0xb0
> Apr 17 19:31:38 pgcore02 kernel:  [cpu_idle+117/128] cpu_idle+0x75/0x80
> Apr 17 19:31:38 pgcore02 kernel: BUG: soft lockup detected on CPU#1!
> Apr 17 19:31:38 pgcore02 kernel:
> Apr 17 19:31:38 pgcore02 kernel: Pid: 0, comm:              swapper
> Apr 17 19:31:38 pgcore02 kernel: EIP: 0061:[hypercall_page+935/4096] CPU: 1
> Apr 17 19:31:38 pgcore02 kernel: EIP is at 0xc01013a7
> Apr 17 19:31:38 pgcore02 kernel:  EFLAGS: 00000292    Not tainted  
> (2.6.16-xen0-pgcore-1-xen3testing #1)
> Apr 17 19:31:38 pgcore02 kernel: EAX: 00000000 EBX: 00000001 ECX: 
> 00000000 EDX: 00000018
> Apr 17 19:31:38 pgcore02 kernel: ESI: 00000000 EDI: 00000001 EBP: 
> c0642000 DS: 007b ES: 007b
> Apr 17 19:31:38 pgcore02 kernel: CR0: 8005003b CR2: bfe85fec CR3: 
> 0f52f000 CR4: 00000640
> Apr 17 19:31:38 pgcore02 kernel:  [xen_idle+83/176] xen_idle+0x53/0xb0
> Apr 17 19:31:38 pgcore02 kernel:  [cpu_idle+117/128] cpu_idle+0x75/0x80
> Apr 17 19:31:38 pgcore02 kernel: BUG: soft lockup detected on CPU#2!
> Apr 17 19:31:38 pgcore02 kernel:
> Apr 17 19:31:38 pgcore02 kernel: Pid: 0, comm:              swapper
> Apr 17 19:31:38 pgcore02 kernel: EIP: 0061:[hypercall_page+935/4096] CPU: 2
> Apr 17 19:31:38 pgcore02 kernel: EIP is at 0xc01013a7
> Apr 17 19:31:38 pgcore02 kernel:  EFLAGS: 00000292    Not tainted  
> (2.6.16-xen0-pgcore-1-xen3testing #1)
> Apr 17 19:31:38 pgcore02 kernel: EAX: 00000000 EBX: 00000001 ECX: 
> 00000000 EDX: 00000018
> Apr 17 19:31:38 pgcore02 kernel: ESI: 00000000 EDI: 00000001 EBP: 
> c0644000 DS: 007b ES: 007b
> Apr 17 19:31:38 pgcore02 kernel: CR0: 8005003b CR2: 080e3b3c CR3: 
> 0d088000 CR4: 00000640
> Apr 17 19:31:38 pgcore02 kernel:  [xen_idle+83/176] xen_idle+0x53/0xb0
> Apr 17 19:31:38 pgcore02 kernel:  [cpu_idle+117/128] cpu_idle+0x75/0x80
> There aren't any domU's running yet. And after a while I've noticed that 
> this error occurs on every cpu, except CPU 0 ?
>  
> I also found this short conversation in the mailing list, but are 
> curious if there is any news on this subject.
> http://groups.google.nl/group/xen-devel-archive/browse_thread/thread/6391b9fc16f3b591/1b41dff3f57e9186?lnk=st
>  
> Keir also suggests to "turn off" the messages, but is this a solution or 
> just a workaround ?
>  
> I also tried running it on a different p3 xeon box (with 2 processors) 
> and got the same errors.
> Then I tried running the same config on a normal p4 desktop pc, and I 
> got no errors at all.
>  
> I hope someone has any ideas or suggestions of what this problem might 
> cause, or can tell me what I'm missing here.
>  
> Thanks in advance,
>  
> Hugo Rodenburg
>  
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Xen-users mailing list
> Xen-users@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-users

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