[RFC PATCH] clk: scmi: Provide channel information to every clock

[Michal Simek]

On 10/5/23 13:49, Michal Simek wrote:
> When assigned-clocks/assigned-clock-rates is defined U-Boot DM core finds
> out leaf which is registered via clk_register().
> But new clock doesn't have information about channel which is stored in
> private data. That's why copy parents private data (channel only now) to
> all childs.
> It will fix the issue when there is reference to clk child and operations
> with it.
> 
> Signed-off-by: Michal Simek <michal.simek at amd.com>
> ---
> 
> I am not really sure that this is the right way how to solve this problem.
> I had another patch which is pretty much written in a way if current device
> doesn't have private data it looks at parents private data.
> Anyway I am sending it as RFC to start to have discussion about it.
> ---
>   drivers/clk/clk_scmi.c | 2 ++
>   1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/clk/clk_scmi.c b/drivers/clk/clk_scmi.c
> index d172fed24c9d..98a779fdc81b 100644
> --- a/drivers/clk/clk_scmi.c
> +++ b/drivers/clk/clk_scmi.c
> @@ -12,6 +12,7 @@
>   #include <scmi_protocols.h>
>   #include <asm/types.h>
>   #include <linux/clk-provider.h>
> +#include <dm/device-internal.h>
>   
>   /**
>    * struct scmi_clk_priv - Private data for SCMI clocks
> @@ -186,6 +187,7 @@ static int scmi_clk_probe(struct udevice *dev)
>   				return ret;
>   			}
>   
> +			dev_set_priv(clk->dev, dev_get_priv(dev));
>   			clk_dm(i, clk);
>   		}
>   	}

Please ignore this patch.
Takahiro's series is fixing this problem
https://lore.kernel.org/all/20230926065750.734440-1-takahiro.akashi@linaro.org/

Thanks,
Michal