Re: [Xen-users] Hypervisor trap when dumping register

[Chenxiao Zhao <chenxiao.zhao@xxxxxxxxx>]

On Tue, May 24, 2016 at 12:20 PM Julien Grall <julien.grall@xxxxxxx> wrote:

On 24/05/2016 05:00, Chenxiao Zhao wrote:
> Hi

Hello,

> I'm playing xen on Hikey board which has octa-core with Cortex-A53. When
> I am trying to dump core registers on xen console (press 'd'), I got
> Hypervisor trap error.
>
> Anyone can explain why it happen and how to fix it?
>
> Thanks.
>
> (XEN) 'd' pressed -> dumping registers
> (XEN)
> (XEN) *** Dumping CPU0 host state: ***
> (XEN) ----[ Xen-4.7.0-rc  arm64  debug=y  Not tainted ]----
> (XEN) CPU:    0
> (XEN) PC:     0000000000245ed8Hypervisor Trap. HSR=0x96000007 EC=0x25

EC=0x25 (see D7.2.27 in ARM DDI 0487A.i) is a data abort from the
hypervisor. Can you please paste all the logs, give the rc you are using...

Hi, Julien

This is all the logs and the system get halt on the last line. I am using 4.7.0-rc which is checked out from git repository master branch.

I found it is caused by a printk in xen/arch/arm/trap.c where is trying to print out symblo for PC with %pS in show_registers_64(). It will finally call a vsnprintf to format the string. 

When I comment out all the lines that has printk with %pS or changed the %pS to %p, every thing works fine. So I guess this is maybe a bug on arm64 with libc?

I have tried several versions of linaro's latest gcc and still got no luck.

Any advise?

Thanks

 

Regards,

--
Julien Grall

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxx
http://lists.xen.org/xen-users