All,

I am having an issue getting a multi-image to boot on my PPC MPC8308-based board. I am using

U-Boot 2011.09

Kernel 3.0.30

Buildroot (rootfs) 2011.11

Using the mkimage tool I have composed a multi-image containing kernel, ramdisk(rootfs), and DTB using the following parameters:

mkimage -A ppc -O linux -T multi -a 0x8000 -e 0x8000 -C gzip -n "MX Image" -d zImage.gz:rootfs.ext2.lzma:my.dtb mxImage

The mxImage is programmed to my (NOR)flash at address 0xfc000000, and attempted started using the bootm-command

in U-Boot as follows:

...

bootm 0xfc000000

After some initialization/loading the board seem to “hang” somewhere along the point where

U-Boot hands-over control to the kernel. The following is a snapshot of the boot-result:

MPC8308=> bootm 0xfc000000

## Booting kernel from Legacy Image at fc000000 ...

Image Name:   MX Image

Image Type:   PowerPC Linux Multi-File Image (gzip compressed)

Data Size:    3663880 Bytes = 3.5 MiB

Load Address: 00008000

Entry Point:  00008000

Contents:

Image 0: 1607257 Bytes = 1.5 MiB

Image 1: 2050371 Bytes = 2 MiB

Image 2: 6232 Bytes = 6.1 KiB

Verifying Checksum ... OK

## Loading init Ramdisk from multi component Legacy Image at fc000000 ...

## Flattened Device Tree from multi component Image at FC000000

Booting using the fdt at 0xfc37cff0

Uncompressing Multi-File Image ... OK

Loading Ramdisk to 0fd43000, end 0ff37943 ... OK

Loading Device Tree to 0fd3e000, end 0fd42857 ... OK

………….board hangs, and must be manually reboot’ed.

Worth mentioning, perhaps, is that if I program the kernel, rootfs, and DTB separately and specify the bootm-command

with the three unique addresses; i.e.

...

bootm <kernel addr> <ramdisk addr> <DTB addr>

then the board boot-up is successful.

I would like to start using the multi-image principle to reduce number of MTD-partitions and hence get a more optimized use of the flash.

I have previous experience with the multi-image principle from a Cortex A9 ARM design (although with more recent

U-Boot, kernel, and Buildroot versions) and have tried to do the MPC8308 along the same lines – without success so far.

Appreciate any ideas.

Regards, Eldor