[yocto] qemuarm: Unable to mount rootfs

Sergey Lychagin serg.lychagin at gmail.com
Fri Nov 18 01:31:43 PST 2011


Hello Jessica,

The answer for all you questions is yes. I made installation according
to the manual.
I've spent some time to investigate the issue. Unfortunately it is
still unresolved. But I found something interesting.
What I have checked:
0. Output from kernel has been redirected to file and analyzed: kernel
supports NFS (but cannot mount rootfs).
1. Firewall was disabled.
2. /etc/export has been updated with path to rootfs on host machine.
nfs daemon has bene restarted after that.
3. 'ifconfig' shows that tap0 interface has been created.
4. 'rpcinfo' shows that both programs (from rpc.mountd & rpc.nfsd) are running.
5. 'rpcinfo -m' points that there were some problems with mountd
(PMAP_GETPORT for 21111 has some failures)
6. '/opt/poky/1.1/sysroots/x86_64-pokysdk-linux/usr/bin/pseudolog
-F"%a(%m) %o %p %r %T"' indicates several failure with mountd:

-(755) stat /opt/poky/1.1/sysroots/x86_64-pokysdk-linux/usr/sbin/rpc.mountd fail

7. somewhere (i don't remember) i saw warnings that pseudo cannot write its log.

what i did then:

I introduced a pause into 'runqemu-internal' script (just before
starting 'distccc' and after start of 'rpc.mountd' & 'rpc.nfsd'). When
the pause happens I started 'qemu-system-arm' in different console.
The command-line for qemu start didn't contain
'mountprog=21111,nfsprog=11111,udp' parameters for kernel (the other
parameters were kept).
After that kernel started to loading. I saw a splash-screen and after
that it got stucked. Screenshot: http://imgur.com/7bL1M

So, I guess there is something wrong with rpc binding. Probably I need
to update some config to have default script working.
And new problem: yocto distro cannot be loaded finally (freezy on
splash-screen - progress bar is almost finished)
Do you have any ideas how to overcome it?

Thanks in advance
Sergey

On Wed, Nov 16, 2011 at 11:02 PM, Zhang, Jessica
<jessica.zhang at intel.com> wrote:
>
> Hi Sergey,
>
>
>
> Couple questions:
>
>
>
> 1.       So did you extract your qemuarm rootfs using runquem-extract-sdk to a directory?
>
> 2.       Point to that directory as sysroot in your Eclipse preference setup page (all of this should be covered in the ADT manual)?
>
> 3.       Can you open an xterm and look for qemu process and kill all of them and retry launch?
>
>
>
> Thanks,
>
> Jessica
>
>
>
> From: yocto-bounces at yoctoproject.org [mailto:yocto-bounces at yoctoproject.org] On Behalf Of Sergey Lychagin
> Sent: Wednesday, November 16, 2011 4:07 AM
> To: yocto at yoctoproject.org
> Subject: [yocto] qemuarm: Unable to mount rootfs
>
>
>
> Hello,
>
> I've faced the following problem: "VFS: unable to mount root fs via NFS" during staring pre-compiled kernel/rootfs images in quemu.
> The screen-shot: http://imgur.com/XtvRc
> I'm trying to start it via Eclipse as it is shown in http://vimeo.com/30557368
>
> I also followed the instructions described ADT user guide: http://www.yoctoproject.org/docs/current/adt-manual/adt-manual.html
> Target is qemuarm. kernel & rootfs (core-image-sato-sdkqemuarm & zImage-qemuarm.bin) are precompiled and downloaded from here: http://downloads.yoctoproject.org/releases/yocto/yocto-1.1/machines/qemu/qemuarm/
>
> As far as I know this kernel image should support NFS.
> Could you please help me?
>
> Thanks in advance
> Sergey



More information about the yocto mailing list