[linux-yocto] Qemuarm image fails to boot successfully ; stuck at scsi

[Bruce Ashfield]

On 2014-04-05, 12:12 AM, New B wrote:
> Hi,
>
> Build Env: yocto 1.5.1 final, running on ubuntu 12.10
>
> Build system is intact.  I followed the quickstart quide and was able to build and run a core-image-minimal for MACHINE ?= "qemuarm"
>
> I, then, followed the instructions to create and add my new bsp layer using "yocto-bsp create sampleBSP qemu", choosing 'arm' as the qemu architecture; default yocto-kernel (3.10); no new machine branch (for now); and default choices for the rest with the exception of having a touchscreen.
>
> I added my layer to the default set of bblayers in build/conf/ after the meta-yocto-bsp, and built the minimal image.
>
> I executed runqemu qemuarm <path to zImage.bin> <path to rootfs.ext3>
>
> ====> kernel does not boot successfully and is stuck in a loop resetting the scsi bus.  The last thing that shows prior to scsi rest loop is:
>
> PCI: enabling device 0000:00:0d.0 (0100 -> 0103)
> sym0: <895a> rev 0x0 at pci 0000:00:0d.0 irq 93
> sym0: No NVRAM, ID 7, Fast-40, LVD, parity checking
> sym0: SCSI BUS has been reset.
> scsi0: sym-2.2.3
>
> BTW: i rebuilt and successfully booted the standard qemuarm after these failures and setting MACHINE ?= "qemuarm" instead of MACHINE ?= "sampleBSP".  The meta-layer for sampleBSP is at the same dir level as meta-yocto-bsp
>
> I would really appreciate any help.  I have been stuck for a couple of days, googling and reading doc' and literally backing out each line (including comments !)  i added to the point where i have a meta-sampleBSP/conf/machine/sampleBSP.conf that matches exactly yocto's qemuarm.conf, and doing clean builds: no dice!  The closest i got was that ppl saying it was

I'd suspect the kernel config has gone wrong. Have you diff'd the .config
from the working qemuarm, and your 'sampleBSP' ?

Bruce

>
> Thanks!
>
>