[U-Boot] [PATCH 1/9 V4] add new CONFIG_AT91_LEGACY

[Tom]

Daniel Gorsulowski wrote:
> Tom wrote:
>> Daniel Gorsulowski wrote:
>>> Hello Tom,
>>>
>>> Tom wrote:
>>>> Jens Scharsig wrote:
>>>>> * add's the new temporary CONFIG_AT91_LEGACY to all board configs
>>>>>  This will need for backward compatiblity, while change the SoC access
>>>>>  to c structures. If CONFIG_AT91_LEGACY is defined, the deprecated
>>>>>  SoC is used.
>>>>>
>>>>>
>>>>> Signed-off-by: Jens Scharsig <js_at_ng at scharsoft.de>
>>>> This looks good.
>>>> The only problem is the new at91 target otc570 is breaking.
>>>>
>>>> Please take a look at errors in otc570
>>>>
>>>> otc570 :
>>>>
>>>> at91sam9263_devices.c: In function 'at91_serial0_hw_init':
>>>> at91sam9263_devices.c:40: warning: implicit declaration of function 
>>>> 'at91_set_a_periph'
>>>> at91sam9263_devices.c: In function 'at91_spi0_hw_init':
>>>> at91sam9263_devices.c:96: warning: implicit declaration of function 
>>>> 'at91_set_b_periph'
>>>> at91sam9263_devices.c:116: warning: implicit declaration of function 
>>>> 'at91_set_pio_output'
>>>> clock.c: In function 'at91_clock_init':
>>>> clock.c:160: warning: implicit declaration of function 'at91_sys_read'
>>>> clock.c:160: error: 'AT91_CKGR_MCFR' undeclared (first use in this function)
>>>> clock.c:160: error: (Each undeclared identifier is reported only once
>>>> clock.c:160: error: for each function it appears in.)
>>>>
>>>> Tom
>>> This is caused by missing defines in include\configs\otc570.h:
>>> #define CONFIG_AT91_LEGACY
>>> and
>>> #define CONFIG_AT91_GPIO	1
>>>
>>> Should I send a patch to fix this, or should I wait for Jens patches coming
>>> mainline?
>>>
>> Please send a patch.
>> It should be a 2-3 liner.
>> I will combine it with Jens' patchset and push them together.
>> Tom
>>
> You probably recognized that I did not sent a patch with CONFIG_AT91_LEGACY
> support but I updated the otc570 board to new SoC access.
> http://lists.denx.de/pipermail/u-boot/2010-February/067691.html
> I guess, this is the best solution.
> 
Yes.
This is a better solution.
Thanks
Tom


> Best regards,
> Daniel Gorsulowski