Re: [PATCH v3 1/2] drivers: rtc: add max313xx series rtc driver

[Christophe JAILLET]

Le 08/11/2022 à 13:22, Ibrahim Tilki a écrit :
> Adding support for Analog Devices MAX313XX series RTCs.
>
> Signed-off-by: Ibrahim Tilki <Ibrahim.Tilki-OyLXuOCK7orQT0dZR+AlfA@xxxxxxxxxxxxxxxx>
> Signed-off-by: Zeynep Arslanbenzer <Zeynep.Arslanbenzer-OyLXuOCK7orQT0dZR+AlfA@xxxxxxxxxxxxxxxx>
> ---

[...]

> +static int max313xx_clkout_register(struct device *dev)
> +{
> +	struct max313xx *rtc = dev_get_drvdata(dev);
> +	int ret;
> +
> +	if (!device_property_present(dev, "#clock-cells"))
> +		return 0;
> +
> +	max313xx_clk_init.name = rtc->chip->clkout_name;
> +	device_property_read_string(dev, "clock-output-names",
> +				    &max313xx_clk_init.name);
> +	rtc->clkout.init = &max313xx_clk_init;
> +
> +	ret = devm_clk_hw_register(dev, &rtc->clkout);
> +	if (ret)
> +		return dev_err_probe(dev, ret, "cannot register clock\n");
> +
> +	return of_clk_add_provider(dev->of_node, of_clk_src_simple_get,
> +				   rtc->clkout.clk);

Hi,

No devm like functionality here?

devm_of_clk_add_hw_provider()? (not sure of the impact or not of the 
"_hw_" in the function name)

> +}

[...]

> +static int max313xx_irq_init(struct device *dev, const char *devname)
> +{
> +	struct max313xx *rtc = dev_get_drvdata(dev);
> +	bool wakeup;
> +	int ret;
> +
> +	rtc->irq = rtc->irqs[0];
> +
> +	switch (rtc->id) {
> +	case ID_MAX31328:
> +		/* max31328 sqw ant int pin is shared */
> +		if (rtc->id == ID_MAX31328 && rtc->irq > 0 && rtc->clkout.clk)
> +			return dev_err_probe(dev, -EOPNOTSUPP,
> +					     "cannot have both sqw clock output and irq enabled");
> +
> +		break;
> +	case ID_MAX31331:
> +	case ID_MAX31334:
> +		if (rtc->clkout.clk) {
> +			/* clockout needs to be enabled for using INTA pin */
> +			ret = clk_prepare_enable(rtc->clkout.clk);
> +			if (ret)
> +				return dev_err_probe(dev, ret,
> +						     "cannot enable clkout\n");
> +		} else {
> +			rtc->irq = rtc->irqs[1];
> +		}
> +		break;
> +	default:
> +		if (rtc->clkin) {
> +			rtc->irq = rtc->irqs[1];
> +
> +			/* wrong interrupt specified */
> +			if (rtc->irqs[0] > 0 && rtc->irqs[1] <= 0)
> +				dev_warn(dev, "INTA is specified but INTB required for irq when clkin is enabled\n");
> +
> +			if (rtc->clkout.clk && rtc->irq > 0)
> +				return dev_err_probe(dev, -EOPNOTSUPP,
> +						"irq not possible when both clkin and clkout are configured\n");
> +
> +			if (rtc->irq <= 0)
> +				break;
> +
> +			/* clkout needs to be disabled for using INTB pin */
> +			if (rtc->chip->clkout->en_invert)
> +				ret = regmap_set_bits(rtc->regmap,
> +						      rtc->chip->clkout->reg,
> +						      rtc->chip->clkout->en_bit);
> +			else
> +				ret = regmap_clear_bits(rtc->regmap,
> +							rtc->chip->clkout->reg,
> +							rtc->chip->clkout->en_bit);
> +
> +			if (ret)
> +				return ret;
> +		}
> +		break;
> +	}
> +
> +	if (rtc->irq > 0) {
> +		ret = devm_request_threaded_irq(dev, rtc->irq, NULL,
> +						&max313xx_irq, IRQF_ONESHOT,
> +						devname, rtc);
> +		if (ret)
> +			return ret;
> +
> +		wakeup = device_property_read_bool(dev, "wakeup-source");
> +		return device_init_wakeup(dev, wakeup);
> +	}
> +
> +	__clear_bit(RTC_FEATURE_ALARM, rtc->rtc->features);

Is it safe? Does it worth it to use __clear_bit() instead of clear_bit() 
here?

> +
> +	return 0;
> +}

[...]