[PATCH 4/5] clk: sunxi-ng: Do not enable by default during compile testing

[Krzysztof Kozlowski]

On 04/04/2025 16:13, Andre Przywara wrote:
> On Fri, 04 Apr 2025 13:57:00 +0200
> Krzysztof Kozlowski <krzysztof.kozlowski at linaro.org> wrote:
> 
> Hi 
>> Enabling the compile test should not cause automatic enabling of all
>> drivers.  Restrict the default to ARCH also for individual drivers, even
>> though their choice is not visible without selecting parent Kconfig
>> symbol, because otherwise selecting parent would select the child during
>> compile testing.
> 
> so I remember we changed this to "default y", because there were some
> tricky problems with regards to RISC-V and ARM. See commits:
> 
> commit 0ff347db4c97cc16b4e428dc1db550ba3628f1e2
> Author: Samuel Holland <samuel at sholland.org>
> Date:   Sat Dec 31 17:14:25 2022 -0600
>     clk: sunxi-ng: Move SoC driver conditions to dependencies
> 
> and 
> 
> commit a26dc096f683ca27ac5e68703bfd3098b4212abd
> Author: Samuel Holland <samuel at sholland.org>
> Date:   Sat Dec 31 17:14:24 2022 -0600
>     clk: sunxi-ng: Remove duplicate ARCH_SUNXI dependencies
> 
> Don't remember what broke, exactly, but just wanted to give a heads up.
> 
Hm, I missed that. Quite unexpected commits.

I would expect they to change default to match exact case but apparently
for simplicity they became default for everyone? No clue...

Commit says:
" Do not duplicate the same expression on the `default` line, so the two
lines do not need to be kept in sync."
what it did, but not WHY.

Why is the most important thing...

The rest of the commit msg:
"Drivers stay disabled under COMPILE_TEST because of the `default
ARCH_SUNXI` applied to SUNXI_CCU."

is true, until you enable SUNXI_CCU. Then all of them became enabled by
default, which is not necessary and in general not welcomed. You should
compile test only things which you want, not everything.

Best regards,
Krzysztof