Re: [Xen-devel] FastModels Xen Crash - Guest data abort: Translation fault at level 2

[Sander Bogaert <sander.bogaert@xxxxxxxxxxxxx>]

On 18 March 2013 15:04, Ian Campbell <Ian.Campbell@xxxxxxxxxx> wrote:

On Mon, 2013-03-18 at 12:26 +0000, Sander Bogaert wrote:
> Hi,
>
>
> Using the latest Xen staging ( git://xenbits.xen.org/xen.git - staging
> ) and the lastest kernel for dom0
> ( git://github.com/torvalds/linux.git - master ) Xen crashes starting
> dom0. I tried 3 approaches:
>
>
> 1. mmc filesystem
>           xen,dom0-bootargs = "earlyprintk=xenboot console=ttyAMA0
> mem=2048M root=/dev/mmcblk0 rw ip=dhcp"
>           crashlog attached "xen_crash_1_mmc"
> 2. nfs filesystem
>           xen,dom0-bootargs = "earlyprintk=xenboot console=ttyAMA1 rw
> root=/dev/nfs nfsroot=157.193.205.141:/srv/nfsrootmin ip=dhcp";
>           crashlog attached "xen_crash_2_nfs"
> 3. ramdisk - this works!

Interesting. I'll concentrate on the first one since I know it works.
What version of the models are you running? What is your model command
line?

I'm running FastModels 8.0 that I downloaded last week. This is how I execute:

 ./model_shell ../resources/Cortex-A15x1/cadi_system_Linux-Release-GCC-4.4.so ../resources/xen_mmc -C motherboard.flashloader0.fname=../resources/kernel -C mothe rboard.mmc.p_mmc_file="../resources/disk.img" -C motherboard.smsc_91c111.enabled=1 -C motherboard.hostbridge.userNetworking=1

Where does your DTB come from? What hypervisor command line did it
include (it seems your logs only start after the early hypervisor bits,
please include everything if possible).

My dtb is compiled from this repository:

git://xenbits.xen.org/people/sstabellini/device-trees.git

I attached the whole log this time.

        (XEN) Guest data abort: Translation fault at level 2
        (XEN)     gva=ef7ff000
        (XEN)     gpa=00000000af7ff000
        (XEN)     instruction syndrome invalid
        (XEN)     eat=0 cm=0 s1ptw=0 dfsc=6
        (XEN) dom0 IPA 0x00000000af7ff000

0xaf7ff000 is a DRAM address. I noticed that the guest just logged
"Truncating memory at 0x80000000 to fit in 32-bit physical address
space". Might relate to how much memory the host and/or dom0 have been
given?

Where would I configure this?

 

        (XEN) P2M @ 02ffbfc0 mfn:0xffdfe
        (XEN) 1ST[0x2] = 0x00000000ffdfb6ff
        (XEN) 2ND[0x17b] = 0x0000000000000000
        (XEN) ----[ Xen-4.3-unstable  arm32  debug=y  Not tainted ]----
        (XEN) CPU:    0
        (XEN) PC:     c014c024


Can you translate 0xc014c024 into a line of the kernel? (e.g. with addr2line).

addr2line -f -e vmlinux 0xc014c024

__memzero

/home/dzan/thesis/source/kernels/linus-linux/arch/arm/lib/memzero.S:54

I'm not really sure this is correct since I did it on the vmlinux binary and I pass the zImage to FastModels.

 

        (XEN) CPSR:   200001d3 MODE:32-bit Guest SVC

Ian.

Attachment: crash_log.txt
Description: Text document

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