On 10 April 2016 at 16:17, Liming Wang <liming.wang@canonical.com> wrote:

>> Sorry, I just found there is hikey_defconfig. Using this config, rootfs
>> can be mounted successfully now.
>> But the multi cpus still can't come onlie at my side.
>> So would you please send me your config, compiled kernel and dtb to test?
>I have shared my images into the folder:
>It includes boot images for l-loader.bin/fip.bin and kernel/dtb images.
>Also include kernel's config, I have not changed anything for kernel's
>config with using below command to build kernel:
Hi Leo,

I just found I used https://github.com/96boards/linux but you used
https://github.com/96boards-hikey/linux. It's different, so which tree should
I use to test?

After using https://github.com/96boards-hikey/linux, I could generate the same
kernel config with yours, but generate a different dtb with yours.
My tree's HEAD commit is below, please help me to confirm we used the same commit:

commit 41d0e8c16d18110718fbdc6cd8c1482100d077fb
Merge: 1025574 ddf5088
Author: Guodong Xu <guodong.xu@linaro.org>
Date:   Thu Mar 17 21:32:32 2016 +0800

    Merge remote-tracking branch 'linaro-android/test/linaro-android-4.4' into working-hikey-mainline-rebase-v4.4-c15

Then I tested your l-loader.bin/fip.bin and kernel/dtb images, and I found some issues:

1. Your dtb didn't fit my board, the kernel still panic. I have attached the log.
If I replace with my generated dtb, the kernel worked well and multi cpus also
could come online. So I wonder whether we use the same board. I used leMaker
version, 1GB RAM and 8GB Flash.

2. I found multi cpus issue is not related to kernel, but u-boot. The key problem is
that in u-boot README, it provides a old version mcuimage link:
We know there are two version mcuimages and above is old. With the old mcuimage, the kernel
could have multi cpus issue. We should provide a patch to fix this link.

I can certainly send a patch to update the readme.

Is there a link which I can use which will always contain the latest mcuimage.bin?

Also does anyone know if there is a published changelog for these mcuimage.bin
binaries? IIRC in the past an updated mcuimage.bin has effected the ability to halt
the processors via JTAG (using OpenOCD) So knowing what changes between
mcuimage.bin versions would be useful.

I must admit to being slightly overwhelmed by the various repos, branches etc
mentioned in this email thread. It reminded me of our mission to reduce
fragmentation ;)