[U-Boot] [PATCH 12/14] microblaze: Move architecture to use generic board init

[Simon Glass]

Hi Michal,


On 29 January 2015 at 07:20, Michal Simek <michal.simek at xilinx.com> wrote:
> Hi Simon,
>
> On 01/29/2015 03:15 AM, Simon Glass wrote:
>> Hi Michal,
>>
>> On 27 January 2015 at 08:20, Michal Simek <michal.simek at xilinx.com> wrote:
>>> Compile code with -fPIC to get GOT. Do not build SPL
>>> with fPIC because it increasing SPL size for nothing.
>>>
>>> Signed-off-by: Michal Simek <michal.simek at xilinx.com>
>>> ---
>>>
>
> ...
>
>>> -#if defined(CONFIG_MIPS) || defined(CONFIG_PPC)
>>> +#if defined(CONFIG_MIPS) || defined(CONFIG_PPC) || defined(CONFIG_MICROBLAZE)
>>>  static int init_func_ram(void)
>>>  {
>>
>> Can you use dram_init() instead?
>
> No problem will do it in v2.
>
> Where dram_init(void) should be declared?
> I see that others have it in arch header. Maybe better to have
> it in any shared location.

Yes a shared location makes sense.

>
>>> --- a/common/board_r.c
>>> +++ b/common/board_r.c
>>> @@ -829,7 +829,7 @@ init_fnc_t init_sequence_r[] = {
>>>  #if defined(CONFIG_ARM)
>>>         initr_enable_interrupts,
>>>  #endif
>>> -#ifdef CONFIG_X86
>>> +#if defined(CONFIG_X86) || defined(CONFIG_MICROBLAZE)
>>>         timer_init,             /* initialize timer */
>>>  #endif
>>
>> Could you use interrupt_init() immediately before?
>
> What do you mean by that?
> Do you want to do timer_init in interrupt_init()?

Yes, then if we remove it from x86 also we can drop both #ifdefs. Do
you see any drawback with that?

Regards,
Simon