[U-Boot] [PATCH] arm, davinci: make arch_cpu_init() in da850_lowlevel.c weak

[Christian Riesch]

Hello Heiko,

On Wed, Nov 9, 2011 at 1:27 PM, Heiko Schocher <hs at denx.de> wrote:
> Christian Riesch wrote:
>> On Wed, Nov 9, 2011 at 11:44 AM, Heiko Schocher <hs at denx.de> wrote:
>>>> Heiko, Christian: please negotiate who performs which part of the
>>>> cleanup. But I expect that with proper symbolic names instead of the
>>>> hardwired constants the need for a wek function will go away.
>>> I rework this.
>>
>> I'd like to help but since we have different custom boards it's a bit
>> hard to test. Therefore I'd like to get this lowlevel initialization
>
> Do you plan to post your patches for your custom board to the ML?

I'm not sure yet if it makes sense... Of course we will publish the
code to comply with GPL once the product development is finished. But
I would rather provide the features that I develop for my custom board
for the da850evm, so everyone with an evaluation board can test them
and adapt them for their own boards.

>
>> working on the da850evm board. Since it boots from SPI I guess I'll
>> need an SPL for SPI flash, right? Could we try to clean it up in a way
>
> Yes.
>
>> so it nicely integrates with your SPL code in
>> arch/arm/cpu/arm926ejs/davinci/spl.c?
>
> I soon post the cleanup for the da850_lowlevel.c file, so you can
> do this based on that patch. You only have to adapt board_init_f() in
> arch/arm/cpu/arm926ejs/davinci/spl.c. There for the da850 case,
> arch_cpu_init() or for the spl case better called da850_lowlevel_init()
> from arch/arm/cpu/arm926ejs/davinci/da850_lowlevel.c must be called.

Great! I'll try to do that.
Thanks for your comments!
Christian