[PATCH] pwm: mediatek: Add error handling for zero rate in PWM state

[Uwe Kleine-König]

Hello,

the From: line of your patch is strange.

On Mon, Dec 15, 2025 at 06:40:32PM +0800, 20190311120055 created wrote:
> From: Payne Lin <payne.lin at mediatek.com>
> 
> Added a check to handle cases where the rate is zero in the
> mtk_disp_pwm_get_state function. This prevents division by zero errors
> when calculating the period.
> - Added error message for zero rate scenario
> 
> Signed-off-by: Payne Lin <payne.lin at mediatek.com>
> ---
>  drivers/pwm/pwm-mtk-disp.c | 31 +++++++++++++++++++------------
>  1 file changed, 19 insertions(+), 12 deletions(-)
> 
> diff --git a/drivers/pwm/pwm-mtk-disp.c b/drivers/pwm/pwm-mtk-disp.c
> index bafd6b6195f6..dd0ae041af70 100644
> --- a/drivers/pwm/pwm-mtk-disp.c
> +++ b/drivers/pwm/pwm-mtk-disp.c
> @@ -176,19 +176,18 @@ static int mtk_disp_pwm_get_state(struct pwm_chip *chip,
>  	struct mtk_disp_pwm *mdp = to_mtk_disp_pwm(chip);
>  	u64 rate, period, high_width;
>  	u32 clk_div, pwm_en, con0, con1;
> -	int err;
> +	int ret;

If you keep the variable name, the patch gets quite a bit more obivious.

> -	err = clk_prepare_enable(mdp->clk_main);
> -	if (err < 0) {
> -		dev_err(pwmchip_parent(chip), "Can't enable mdp->clk_main: %pe\n", ERR_PTR(err));
> -		return err;
> +	ret = clk_prepare_enable(mdp->clk_main);
> +	if (ret < 0) {
> +		dev_err(pwmchip_parent(chip), "Can't enable mdp->clk_main: %pe\n", ERR_PTR(ret));
> +		goto err_handle;
>  	}

Fixing error handling should go into a separate patch please.

> -	err = clk_prepare_enable(mdp->clk_mm);
> -	if (err < 0) {
> -		dev_err(pwmchip_parent(chip), "Can't enable mdp->clk_mm: %pe\n", ERR_PTR(err));
> -		clk_disable_unprepare(mdp->clk_main);
> -		return err;
> +	ret = clk_prepare_enable(mdp->clk_mm);
> +	if (ret < 0) {
> +		dev_err(pwmchip_parent(chip), "Can't enable mdp->clk_mm: %pe\n", ERR_PTR(ret));
> +		goto err_disable_clk_main;
>  	}
>  
>  	/*
> @@ -212,15 +211,23 @@ static int mtk_disp_pwm_get_state(struct pwm_chip *chip,
>  	 * period has 12 bits, clk_div 11 and NSEC_PER_SEC has 30,
>  	 * so period * (clk_div + 1) * NSEC_PER_SEC doesn't overflow.
>  	 */
That comment belongs to the assignment to state->period. Don't separate
them with your addition.

> +	if (rate == 0) {
> +		dev_err(pwmchip_parent(chip), "rate is zero, cannot calculate period\n");

All the other messages (at least those in the context above) start with
a capital letter.

> +		ret = -EINVAL;
> +		goto err_disable_clk_mm;
> +	}
>  	state->period = DIV64_U64_ROUND_UP(period * (clk_div + 1) * NSEC_PER_SEC, rate);
>  	high_width = FIELD_GET(PWM_HIGH_WIDTH_MASK, con1);
>  	state->duty_cycle = DIV64_U64_ROUND_UP(high_width * (clk_div + 1) * NSEC_PER_SEC,
>  					       rate);
>  	state->polarity = PWM_POLARITY_NORMAL;
> +
> +err_disable_clk_mm:
>  	clk_disable_unprepare(mdp->clk_mm);
> +err_disable_clk_main:
>  	clk_disable_unprepare(mdp->clk_main);
> -
> -	return 0;
> +err_handle:
> +	return ret;
>  }

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-mediatek/attachments/20251215/d95dfd66/attachment.sig>