Subject: Re: [PATCH 3/4] dt-bindings: clock: Add AST2700 clock bindings

On 19/08/2024 07:55, Ryan Chen wrote:

quoted

quoted

Subject: Re: [PATCH 3/4] dt-bindings: clock: Add AST2700 clock
bindings

On 13/08/2024 03:53, Ryan Chen wrote:

quoted

quoted

Drop the define for number of clocks from the header, because it is
not a

*NUMBER OF CLOCKS*

quoted

quoted

binding. You can put it in the driver or not, I don't care and do
not provide guidance on this because I don't know if it makes sense at all.
What I know is that number of clocks is not related to binding. It
is not needed

*NUMBER OF CLOCKS*

quoted

quoted

in the binding, either.

Sorry, I am confused.
if you think that number of clocks is not related to binding.

*NUMBER OF CLOCKS*

quoted

How dtsi claim for clk?
For example in dtsi.
include <dt-bindings/clock/aspeed,ast2700-clk.h>
usb3bhp: usb3bhp {
....
clocks = <&syscon0 SCU0_CLK_GATE_PORTAUSB>;

And where is *NUMBER OF CLOCKS* here? I don't see any problem. No
useless SCU0_CLK_GATE_NUM define here.

Understood now, I will remove those *NUMBER OF CLOCKS*.
And will replace to
#define SCU0_CLK_END  34

NAK, it's like you keep ignoring my comments entirely. Even if you call it
"SCU0_CLK_NOT_END" it does not change. Do you understand that it is not
about name? Read my first comment.

quoted

Refer:
https://github.com/torvalds/linux/blob/master/include/dt-bindings/cloc
k/imx8-clock.h#L87

So you found a bug and this allows you to create the same bug?

Sorry, I don't see this is a bug.
But I try to understand your point, you prefer following for clock nums, am I correct?
https://github.com/torvalds/linux/blob/master/drivers/clk/meson/g12a.c#L5558-L5559

dt-binding is index table like following.
https://github.com/torvalds/linux/blob/master/drivers/clk/meson/g12a.c#L4380

Best regards,
Krzysztof