Re: [PATCH v4] pwm: atmel: Enable clk when pwm already enabled in bootloader

[Guiting Shen]

On Wed,Jul 12,2023 at 21:55:53PM GMT+8, Uwe Kleine-König wrote:
> On Wed, Jul 12, 2023 at 09:43:47PM +0800, Guiting Shen wrote:
> > The driver would never call clk_enable() if the PWM channel was already
> > enabled in bootloader which lead to dump the warning message "the PWM
> > clock already disabled" when turning off the PWM channel.
> >
> > Add atmel_pwm_enable_clk_if_on() in probe function to enable clock if
> > the PWM channel was already enabled in bootloader.
> >
> > Signed-off-by: Guiting Shen <aarongt.shen@xxxxxxxxx>
> > ---
> >   drivers/pwm/pwm-atmel.c | 40 +++++++++++++++++++++++++++++++++++++++-
> >   1 file changed, 39 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/pwm/pwm-atmel.c b/drivers/pwm/pwm-atmel.c
> > index cdbc23649032..28ea0f7267ca 100644
> > --- a/drivers/pwm/pwm-atmel.c
> > +++ b/drivers/pwm/pwm-atmel.c
> > @@ -36,7 +36,7 @@
> >   #define PWM_SR			0x0C
> >   #define PWM_ISR			0x1C
> >   /* Bit field in SR */
> > -#define PWM_SR_ALL_CH_ON	0x0F
> > +#define PWM_SR_ALL_CH_MASK	0x0F
> >   
> >   /* The following register is PWM channel related registers */
> >   #define PWM_CH_REG_OFFSET	0x200
> > @@ -464,6 +464,37 @@ static const struct of_device_id atmel_pwm_dt_ids[] = {
> >   };
> >   MODULE_DEVICE_TABLE(of, atmel_pwm_dt_ids);
> >   
> > +static int atmel_pwm_enable_clk_if_on(struct atmel_pwm_chip *atmel_pwm)
> > +{
> > +	unsigned int i;
> > +	int err;
> > +	u32 sr;
> > +
> > +	sr = atmel_pwm_readl(atmel_pwm, PWM_SR);
> > +	if (!(sr & PWM_SR_ALL_CH_MASK))
> > +		return 0;
> > +
> > +	for (i = 0; i < atmel_pwm->chip.npwm; i++) {
> > +		if (!(sr & (1 << i)))
> > +			continue;
> > +
> > +		err = clk_enable(atmel_pwm->clk);
> > +		if (err) {
> > +			dev_err(atmel_pwm->chip.dev,
> > +				"failed to enable clock for pwm #%d: %pe\n",
> > +							i, ERR_PTR(err));
> > +
> > +			while (i--) {
> > +				if (sr & (1 << i))
> > +					clk_disable(atmel_pwm->clk);
> > +			}
> > +			return err;
> > +		}
> > +	}
> > +
> > +	return 0;
> > +}
> > +
> >   static int atmel_pwm_probe(struct platform_device *pdev)
> >   {
> >   	struct atmel_pwm_chip *atmel_pwm;
> > @@ -504,8 +535,15 @@ static int atmel_pwm_probe(struct platform_device *pdev)
> >   
> >   	platform_set_drvdata(pdev, atmel_pwm);
> >   
> > +	ret = atmel_pwm_enable_clk_if_on(atmel_pwm);
> > +	if (ret < 0)
> > +		goto remove_pwmchip;
> > +
> >   	return ret;
> >   
> > +remove_pwmchip:
> > +	pwmchip_remove(&atmel_pwm->chip);
> > +
>
> I'd consider it more natural to do the atmel_pwm_enable_clk_if_on() call
> before registering the pwmchip. But I guess it works like this, too.
> (Well unless it's possible that there are set bits in PWM_SR and the
> clock is off.)

It need to invoke something like atmel_pwm_disable_clk_if_on() after 
failing to register the pwmchip when call the 
atmel_pwm_enable_clk_if_on() first which I feel inconvenient.
And it's impossible to enable PWM channels before enabling the clock.
The bits set in PWM_SR means that the PWM channel and the clock turn on.


--
Best regards,
Guiting Shen