[PATCH] ARM: config: Add BCM53xx defconfig

[Florian Fainelli]

On 11/7/22 14:25, Linus Walleij wrote:
> On Mon, Nov 7, 2022 at 6:32 PM Florian Fainelli <f.fainelli at gmail.com> wrote:
>> On 11/7/22 05:49, Linus Walleij wrote:
>>> This adds a BCM530x ARM-based Broadcom default config, typically
>>> for use with NorthStar systems. This is needed because it took
>>> me quite a lot of time to figure out all the different drivers and
>>> intricacies that needed to be enabled to bring a typical specimen
>>> from these systems up.
>>>
>>> By simply using this it is easy to configure and boot a mainline
>>> Linux kernel on these systems.
>>>
>>> Signed-off-by: Linus Walleij <linus.walleij at linaro.org>
>>> ---
>>> I suppose Florian can merge this, I can also send it directly to
>>> the SOC tree if preferred.
>>
>> ARM SoC maintainers would not let us have a cygnus_defconfig many years
>> ago, so I doubt they would accept this one. defconfigs are easy to
>> maintain out of tree, for better or for worse.

This is the thread I had in mind:

https://lore.kernel.org/linux-arm-kernel/1996601.jhvNxNYxnl@wuerfel/

I do remember pretty vividly that Arnd did not want to see more 
SoC-specific defconfig files but looking at the output of:

shows that we had defconfigs introduced as late as 2022, so either 
something changed and I missed it, or there was something else.

> 
> Why not, I had to spend hours to come up with this one, it has
> general use and is good for my compile tests. Also policies
> change, consistency isn't our best game.
> 
> Arnd, Olof: can I have this defconfig?
> 
> Itr's not like I can't deal with rejection but I want some rationale.
> 
> I can send it directly to SoC so you don't need to deal
> with it as long as you approve of the usefulness.

If it does land, I want to be able to track changes to that defconfig, 
so no it is not the route you want to seek, and I am not opposed to 
merging yet, just telling you that last I was aware it happened, the 
defconfig eventually did not land.

It makes it sound as if because I do not agree with your change, you are 
looking for a higher power to take on your changes which is effectively 
a form of bypass.

Maybe it would be more effective to introduce a generic 
Northstar/Northstar Plus defconfig since the same kernel image can run 
on both types of devices.
-- 
Florian