[PATCH v4 2/3] clocksource: timer-cadence-ttc: Support TTC device configured as PWM

[Uwe Kleine-König]

On Wed, Jan 15, 2025 at 05:05:55PM +0530, Mubin Sayyed wrote:
> TTC device can act either as clocksource/clockevent or PWM generator,
> it would be decided by pwm-cells property. If pwm-cells property is
> present in TTC node, it would be treated as PWM device, and clocksource
> driver just calls probe function for PWM functionality, so that TTC
> device would be registered with PWM framework.
> 
> Signed-off-by: Mubin Sayyed <mubin.sayyed at amd.com>
> ---
> Changes for v4:
>     - In case of pwm-cells property call probe function for PWM
>       feature instead of returning error.
> Changes for v3:
>     - None
> Changes for v2:
>     - Added comment regarding pwm-cells property
> ---
>  drivers/clocksource/timer-cadence-ttc.c | 34 +++++++++++++++++++++++++
>  1 file changed, 34 insertions(+)
> 
> diff --git a/drivers/clocksource/timer-cadence-ttc.c b/drivers/clocksource/timer-cadence-ttc.c
> index 2f33d4c40153..c5ecad9332c9 100644
> --- a/drivers/clocksource/timer-cadence-ttc.c
> +++ b/drivers/clocksource/timer-cadence-ttc.c
> @@ -35,6 +35,10 @@
>   * obtained from device tree. The pre-scaler of 32 is used.
>   */
>  
> +struct ttc_timer_config {
> +	bool is_pwm_mode;
> +};
> +
>  /**
>   * struct ttc_timer - This definition defines local timer structure
>   *
> @@ -453,6 +457,7 @@ static int __init ttc_setup_clockevent(struct clk *clk,
>  
>  static int __init ttc_timer_probe(struct platform_device *pdev)
>  {
> +	struct ttc_timer_config *ttc_config;
>  	unsigned int irq;
>  	void __iomem *timer_baseaddr;
>  	struct clk *clk_cs, *clk_ce;
> @@ -461,6 +466,24 @@ static int __init ttc_timer_probe(struct platform_device *pdev)
>  	u32 timer_width = 16;
>  	struct device_node *timer = pdev->dev.of_node;
>  
> +	ttc_config = devm_kzalloc(&pdev->dev, sizeof(*ttc_config), GFP_KERNEL);
> +	if (!ttc_config)
> +		return -ENOMEM;
> +
> +	/*
> +	 * If pwm-cells property is present in TTC node,
> +	 * it would be treated as PWM device.
> +	 */
> +	if (of_property_read_bool(timer, "#pwm-cells")) {
> +		#if defined(CONFIG_PWM_CADENCE)
> +		ttc_config->is_pwm_mode = true;
> +			return ttc_pwm_probe(pdev);

strange indention. Maybe use

	if (IS_REACHABLE(CONFIG_PWM_CADENCE))

This is an unusal way to bind the PWM driver. I'd prefer creation of separate
device in the PWM case. I wonder if it can happen that ttc_pwm_probe() is
called during boot before pwm_init() completed. Or use an auxbus device
to distinguish between timer and pwm?

> +		#else
> +			return -ENODEV;
> +		#endif
> +	}
> +	dev_set_drvdata(&pdev->dev, ttc_config);
> +
>  	if (initialized)
>  		return 0;
>  

Best regards
Uwe
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20250324/17d1804d/attachment.sig>