[U-Boot] [PATCH v3] Consolidate bootcount code into drivers/bootcount

[Andreas Bießmann]

Hi Stefan,

On 13.08.2012 16:51, Stefan Roese wrote:
> Hi Andreas,
> 
> On 08/13/2012 03:48 PM, Andreas Bießmann wrote:
>>>>> +COBJS-y				+= bootcount.o
>>>>> +COBJS-$(CONFIG_AT91SAM9XE)	+= bootcount_at91.o
>>>>
>>>> I tend to NAK this. Before it was available to all at91 processors (keep
>>>> in mind nearly all at91 have this gpbr register). Now it is only
>>>> available to AT91SAM9XE processor series which is the only user for
>>>> bootcount in mainline.
>>>
>>> Then we should choose a different CONFIG_ option here. One that selects
>>> all AT91 boards potentially supporting this feature. You are the expert
>>> here, please make a suggestion.
>>
>> Unfortunately there is no such config option yet. We could add all the
>> SoC explicitly like this:
>>
>> ---8<---
>> | +COBJS-$(CONFIG_AT91SAM9260)    += bootcount_at91.o
>> | +COBJS-$(CONFIG_AT91SAM9261)    += bootcount_at91.o
>> | +COBJS-$(CONFIG_AT91SAM9263)    += bootcount_at91.o
>> | +COBJS-$(CONFIG_AT91SAM9G10)    += bootcount_at91.o
>> | +COBJS-$(CONFIG_AT91SAM9G20)    += bootcount_at91.o
>> | +COBJS-$(CONFIG_AT91SAM9M10G45) += bootcount_at91.o
>> | +COBJS-$(CONFIG_AT91SAM9RL)     += bootcount_at91.o
>> |  COBJS-$(CONFIG_AT91SAM9XE)     += bootcount_at91.o
>> | +COBJS-$(CONFIG_AT91SAM9G20)    += bootcount_at91.o
>> --->8---
>>
>> Maybe there is some make foo to get this easier?
> 
> Maybe. But nothing I can think of quickly. The better solution would be
> to add a new common CONFIG_AT91 (or similar) define, which is defined
> for all at91 platforms. Might be helpful in other places as well.
> 
> Not sure, how to best add this global AT91 define though. Easy wold be
> to add it to config header files.

I think we put it in the asm/arch/hardware.h files, something like
CONFIG_AT91_GPBR.

>>>> I fear we may break some not mainline boards
>>>> here.
>>>
>>> Maybe. But we usually don't care about out-of-tree ports.
>>
>> That is true, we could just wait for patches adding this feature to
>> other at91 SoC.
>>
>>>> I would prefer something that includes all different at91 SoC by
>>>> default (except rm9200).
>>>>
>>>> I have no solution yet but send this to prevent a v4. Will send a
>>>> proposal for at91 later this day.
>>>
>>> Okay. But I would really like to see this patch go in soon. I still have
>>> a new board support patch waiting here for quite a long time depending
>>> on this bootcount stuff.
>>
>> I'm with you, do you have a suggestion how to do the make foo nice?
> 
> See above. Perhaps somebody else has other suggestions.
> 
> Nevertheless I think we can postpone this "AT91 bootcount tuning" to a
> follow-up patch.

This is ok for me. Just realized there is a v4 on patchwork (but did not
hit my MUA ...).

Just add a

Acked-by: Andreas Bießmann <andreas.devel at googlemail.com>

to v4.

Best regards

Andreas Bießmann