On 2020/11/19 17:07, Markus Elfring wrote:
>> Refine hi3620_mmc_clk_init() to use of_clk_add_hw_provider()
>> instead of of_clk_add_provider(), …
>
> …
>> +++ b/drivers/clk/hisilicon/clk-hi3620.c
> …
>> @@ -439,17 +440,22 @@ static struct clk *hisi_register_clk_mmc(struct
>> hisi_mmc_clock *mmc_clk,
> …
>> + err = clk_hw_register(NULL, hw);
>> +
>> + if (err) {
>
> I suggest to omit a blank line here.
>
>
> …
>> +++ b/drivers/clk/hisilicon/clk.c
>> @@ -65,25 +65,26 @@ struct hisi_clock_data *hisi_clk_init(struct
>> device_node *np,
> …
>> of_clk_add_provider(np, of_clk_src_onecell_get, &clk_data->clk_data);
>> return clk_data;
>> -err_data:
>> +free_clk_data:
>
> How do you think about to adjust also such a function call for this function
> implementation?
>
> Will further collateral evolution become interesting?
> https://elixir.bootlin.com/linux/v5.10-rc4/C/ident/of_clk_add_provider
Thanks for the review.
How about we adjust such a function call in another series patches? I suggest
do it in another series.
For this patch, how about we firstly merge it after I fix the above
comments(omit a blank line)?
>
> Regards,
> Markus
> .
>
