Re: [PATCH] linux_2.6.38: update support for PcontrolG20 HW rev 2

["Eric Bénard" <eric@eukrea.com>]

Hi Peter,

On 16/05/2011 11:55, Peter Gsellmann wrote:
> Am Samstag 14 Mai 2011, 19:06:24 schrieb Eric Bénard:
>> On 29/04/2011 11:45, Peter Gsellmann wrote:
>>> All prototypes hardware rev 1 are drawn back
>>> so default should be rev 2
>>>
>>> Signed-off-by: Peter Gsellmann<pgsellmann@portner-elektronik.at>
>>> ---
>>>    recipes/linux/linux-2.6.38/pcontrolg20/defconfig   | 1761 ++++++++++++++++++++
>>>    .../pcontrolg20/pcontrolg20_2.6.38.patch           |  235 +++
>>>    recipes/linux/linux_2.6.38.bb                      |    2 +
>>>    3 files changed, 1998 insertions(+), 0 deletions(-)
>>>    create mode 100644 recipes/linux/linux-2.6.38/pcontrolg20/defconfig
>>>    create mode 100644 recipes/linux/linux-2.6.38/pcontrolg20/pcontrolg20_2.6.38.patch
>>>
>> I applied this patch and tried to build the kernel but it seems there is no
>> machine definition for pcontrolg20 : can you please resend a patch serie to
>> add the kernel support and the machine conf file.
> the machine configuration for pcontrolg20 is still not in a stable state:
> At least the kernel patch is stable, but for the at91bootstrap i have made changes for lowlevel initialization and memory detection which i have to bring in sync with mainline; right now my at91bootstrap patch looks ugly and would be rejected deservedly.
> Same is for U-boot. I patched a specific git-revision with a number of additional commands i need for factory testing.
> Even worse, the perpetual discussions about kernel relocation/decompression convinced me to evaluate alternatives like barebox.
>
good choice ;-)

> So my ToDo-list is:
> (*) commit kernel.patch
> * commit at91bootstrap.patch
> * commit U-boot or Barebox patches
> * commit conf/machine/pcontrolg20.conf
> * commit recipes/pcontrolg20-image.bb
>
> If its ok for you this would be the first commit.
>
> I could commit the pcontrolg20.conf also, but the entries for at91boosttrap and u-boot would be commented out.

yes a kernel recipe without the machine conf file can't be tested : please 
send the 2 patches (kernel + conf file), then you will be able to add entries 
for at91bootstrap & u-boot to the conf file later when you add the support for 
you board in their respective recipes.

thanks
Eric