Re: [PATCH v4 2/2] iio: frequency: adf4350: add clk provider

[Nuno Sá]

On Wed, 2024-06-12 at 16:02 +0300, Antoniu Miclaus wrote:
> Add clk provider feature for the adf4350.
> 
> Even though the driver was sent as an IIO driver in most cases the
> device is actually seen as a clock provider.
> 
> This patch aims to cover actual usecases requested by users in order to
> completely control the output frequencies from userspace.
> 
> Signed-off-by: Antoniu Miclaus <antoniu.miclaus@xxxxxxxxxx>
> ---
> changes in v4:
>  - rename macro to `to_adf4350_state`
>  - do not expose ADF4350_FREQ and ADF4350_FREQ_REFIN if driver is used as clk
>    provider.
>  - initialize flags with CLK_SET_RATE_PARENT
>  drivers/iio/frequency/adf4350.c | 134 +++++++++++++++++++++++++++++++-
>  1 file changed, 133 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/iio/frequency/adf4350.c b/drivers/iio/frequency/adf4350.c
> index 4abf80f75ef5..8309ddfca9af 100644
> --- a/drivers/iio/frequency/adf4350.c
> +++ b/drivers/iio/frequency/adf4350.c
> @@ -19,6 +19,7 @@
>  #include <linux/gpio/consumer.h>
>  #include <asm/div64.h>
>  #include <linux/clk.h>
> +#include <linux/clk-provider.h>
>  
>  #include <linux/iio/iio.h>
>  #include <linux/iio/sysfs.h>
> @@ -36,6 +37,9 @@ struct adf4350_state {
>  	struct gpio_desc		*lock_detect_gpiod;
>  	struct adf4350_platform_data	*pdata;
>  	struct clk			*clk;
> +	struct clk			*clkout;
> +	const char			*clk_out_name;
> +	struct clk_hw			hw;
>  	unsigned long			clkin;
>  	unsigned long			chspc; /* Channel Spacing */
>  	unsigned long			fpfd; /* Phase Frequency Detector */
> @@ -61,6 +65,8 @@ struct adf4350_state {
>  	__be32				val __aligned(IIO_DMA_MINALIGN);
>  };
>  
> +#define to_adf4350_state(_hw) container_of(_hw, struct adf4350_state, hw)
> +
>  static struct adf4350_platform_data default_pdata = {
>  	.channel_spacing = 10000,
>  	.r2_user_settings = ADF4350_REG2_PD_POLARITY_POS |
> @@ -370,6 +376,12 @@ static const struct iio_chan_spec_ext_info adf4350_ext_info[]
> = {
>  	{ },
>  };
>  
> +static const struct iio_chan_spec_ext_info adf4350_clk_ext_info[] = {
> +	_ADF4350_EXT_INFO("frequency_resolution", ADF4350_FREQ_RESOLUTION),
> +	_ADF4350_EXT_INFO("powerdown", ADF4350_PWRDOWN),
> +	{ },

Do we really need powerdown? Dunno :). I would expect one of unprepare/disable to
take care of that. Moreover, imagine userspace powers down the device while an in
kernel consumer was using it? Not cool right ehhe?

Even the frequency_resolution is arguable as that is also a DT property but to keep
consistency why not? That one I can live with...

Also note that you're still not including the clock maintainers in the loop.

- Nuno Sá