Hello Andre,
On 2024-12-18 02:38, Andre Przywara wrote:
On Tue, 17 Dec 2024 23:02:43 +0100
Dragan Simic <dsimic@xxxxxxxxxxx> wrote:
On 2024-12-17 22:15, Andre Przywara wrote:
> On Tue, 17 Dec 2024 10:00:45 -0800
> Vasily Khoruzhick <anarsoul@xxxxxxxxx> wrote:
>> On Mon, Dec 16, 2024 at 11:33 PM Krzysztof Kozlowski <krzk@xxxxxxxxxx>
>> wrote:
>> >
>> > On Sat, Dec 14, 2024 at 09:34:57PM -0800, Vasily Khoruzhick wrote:
>> > > These will be used to explicitly select TCON0 clock parent in dts
>> > >
>> > > Fixes: ca1170b69968 ("clk: sunxi-ng: a64: force select PLL_MIPI in TCON0 mux")
>> > > Signed-off-by: Vasily Khoruzhick <anarsoul@xxxxxxxxx>
>> > > ---
>> > > drivers/clk/sunxi-ng/ccu-sun50i-a64.h | 2 --
>> > > include/dt-bindings/clock/sun50i-a64-ccu.h | 2 +
>>
>> > You cannot combine these changes.
>>
>> The patch basically moves defines out from ccu-sun50i-a64.h to
>> sun50i-a64-ccu.h. How do I split the change without introducing
>> compilation failure?
>
> You can just have the binding part first, adding the (same) definition
> to the binding headers. As long as the #define's are not conflicting,
> this is fine.
> Then remove the now redundant definitions in the kernel headers, with a
> subsequent patch.
Yes, that would be a way to make it formally correct, but also much
less readable and understandable later, as part of the source code
repository. FWIW, I find this to be an example of the form being
more important than the actual function.
Not sure I understand your last sentence, exactly,
Ah, sorry, I also saw the need to expand that sentence a bit, but only
after I had already sent my message. :/
but what Krzysztof pointed out is that one part (the header change in
include/dt-bindings) is a DT binding patch, so part of a spec, if you
like, the other is Linux *code*. There is the DT rebasing repo, which
cherry-picks DT patches, so they form a separate history there, and
Linux code has no place in there. U-Boot for instance pull this
repo now on a regular basis. So keeping those things strictly separate
is really important here.
Thanks a lot for your detailed explanation and for reminding me about
the DT rebasing repository. [*] I forgot about it for a moment, despite
being active in U-Boot development when the repository was introduced,
so yes, you're right that we're now in need to keep the bindings and the
kernel code separate in patches.
However, I can't stop myself from noticing that the way Krzysztof
replied
left a lot to be desired when it comes to friendliness in general. Such
an approach is probably what contributed to me forgetting about the DT
rebasing repository for a moment.
[*]
https://git.kernel.org/pub/scm/linux/kernel/git/devicetree/devicetree-rebasing.git/
>> > Please run scripts/checkpatch.pl and fix reported warnings. Then please
>> > run 'scripts/checkpatch.pl --strict' and (probably) fix more warnings.
>> > Some warnings can be ignored, especially from --strict run, but the code
>> > here looks like it needs a fix. Feel free to get in touch if the warning
>> > is not clear.
>>
>> Yeah, it is not clear what do you want me to do, assuming the previous
>> similar change to sun50i-a64-ccu.h did essentially the same, see
>> 71b597ef5d46a326fb0d5cbfc1c6ff1d73cdc7f9
