[U-Boot] [PATCH v2 2/6] apalis/colibri_t20/t30: deactivate displaying board info

Stefan Agner stefan at agner.ch
Wed Oct 5 18:12:10 CEST 2016


On 2016-10-05 08:53, Stephen Warren wrote:
> On 10/03/2016 02:27 PM, Stefan Agner wrote:
>> On 03.10.2016 10:28, Stephen Warren wrote:
>>> On 09/30/2016 04:00 AM, Marcel Ziswiler wrote:
>>>> On Wed, 2016-09-28 at 12:00 -0600, Stephen Warren wrote:
>>>>> On 09/28/2016 03:35 AM, Marcel Ziswiler wrote:
>>>>>>
>>>>>> Avoid a checkboard() name clash with our upcoming custom
>>>>>> implementation
>>>>>> thereof.
>>>>> If you want to avoid naming conflicts, please simply name your new
>>>>> function something that doesn't conflict. That way it will avoid
>>>>> confusion is someone actually wants to enable the
>>>>> CONFIG_DISPLAY_BOARDINFO option themselves, plus it avoids taking
>>>>> the
>>>>> current feature set away.
>>>>
>>>> No, it is not just any function. We do want our custom checkboard() to
>>>> be called upon boot and not the Tegra generic one just printing a hard
>>>> coded string.
>>>>
>>>> I guess alternatively we could gate the checkboard() call
>>>> in arch/arm/mach-tegra/board2.c with a
>>>>
>>>> #if !defined(CONFIG_CUSTOM_BOARDINFO)
>>>>
>>>> just as introduced a while ago in common/board_info.c
>>>>
>>>> http://git.denx.de/?p=u-boot.git;a=blob;f=common/board_info.c;h=bd5dcfa
>>>> 066358c2cc44ce5d19fcc3e77d555cd09;hb=HEAD#l20
>>>>
>>>> in order to not print the hard coded name from the device tree.
>>>
>>> I'd prefer to keep the behaviour standard across all Tegra boards. If
>>> you want to do additional actions in the checkboard() function, I
>>> suggest making it call an optional additional function:
>>>
>>> __weak int tegra_checkboard(void)
>>> {
>>>         return 0;
>>> }
>>>
>>> int checkboard(void)
>>> {
>>>     ...
>>>     return tegra_checkboard();
>>> }
>>
>> Well that would print a message "Board: " ... twice, which is rather
>> strange.
> 
> Surely you simply make tegra_checkboard() not contain duplicate code?
> 
>> What do you think of my idea?
> 
> I'd rather not introduce any more ifdefs, but instead have a single
> path through the code-base.

Sorry, I was a bit unclear, with my other idea I meant the answer I sent
to patch 3/6 of this patchset:
http://lists.denx.de/pipermail/u-boot/2016-October/268669.html

It does remove a ifdef...

--
Stefan


More information about the U-Boot mailing list