[meta-freescale] [meta-fsl-arm-extra][PATCH 0/3] Initial support for Nitrogen6X

[Otavio Salvador]

On Tue, Dec 11, 2012 at 12:09 PM, Eric Nelson <
eric.nelson at boundarydevices.com> wrote:

> On 12/11/2012 01:23 AM, Eric Bénard wrote:
>
>> Hi Eric,
>>
>> Le Mon, 10 Dec 2012 20:58:22 -0700,
>> Eric Nelson <eric.nelson at boundarydevices.**com<eric.nelson at boundarydevices.com>>
>> a écrit :
>>
>>> | 6x_bootscript-nitrogen6x.txt: No such file or directory
>>>
>>
>> Does that work if you do bibake u-boot-script-boundary before building
>> the image ?
>>
>>
> Thanks Eric,
>
> That allows things to work, so it seems we have a missing dependency.
>
> We still have an issue though. The generic boot script in our git
> repository has a couple of references that presume a single
> partition:
>
>         https://github.com/**boundarydevices/u-boot-imx6/**
> blob/production/board/**boundary/nitrogen6x/6x_**bootscript.txt#L56<https://github.com/boundarydevices/u-boot-imx6/blob/production/board/boundary/nitrogen6x/6x_bootscript.txt#L56>
>
> In particular, the 'root=/dev/mmcblk0p1' clause and the 1 in the
> command:
>         ${fs}load mmc ${disk}:1
>
> The SD card created by bitbake is configured for two:
>
>         ~/yocto/build$ fdisk -l tmp/deploy/images/fsl-image-**
> test-nitrogen6x-**20121211024022.rootfs.sdcard
>         ...
>         Device Boot      Start         End      Blocks   Id  System
>      ...sdcard1            8192       24575        8192   83  Linux
>      ...sdcard2           24576      753663      364544   83  Linux
>
> It trivially easy to create a boot script specifically for Yocto
> images, but this should probably be discussed.
>

We can apply a patch in the script for it. Can you please generate a patch
to adapt the script for our environment?


> Why should we use two partitions? It seems that unless we're loading
> a RAM disk, there's no benefit in having multiple partitions.
>
> I suspect that there's a way to create a RAM disk in Yocto, but
> I'm not sure where to start besides Googling the MLs and grepping
> the sources.
>

The real benefit here is to avoid too much complexity in sdcard class. For
accomplish it in a single partition we'd need to flexibilize the class even
more and I am not sure it is worth the effort.


> If we decide to use a single partition, it also seems that the
> .sdcard target is a bit of overkill, and a tar-ball of the RFS
> is sufficient.
>

The idea to provide a sdcard image is to allow for easier test and use of
the build images. It is a matter of dd the fine, directly in the sdcard and
be done with that.

-- 
Otavio Salvador                             O.S. Systems
E-mail: otavio at ossystems.com.br  http://www.ossystems.com.br
Mobile: +55 53 9981-7854              http://projetos.ossystems.com.br
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.yoctoproject.org/pipermail/meta-freescale/attachments/20121211/19ccf9eb/attachment.html>