[PATCH 2/2] phy: nuvoton: add new driver for the Nuvoton MA35 SoC USB 2.0 PHY

[Hui-Ping Chen]

Dear Krzysztof,

Thank you for your reply.



On 2024/7/29 下午 03:32, Krzysztof Kozlowski wrote:
> On 29/07/2024 08:15, hpchen0 wrote:
>> Nuvoton MA35 SoCs support DWC2 USB controller.
>> Add the driver to drive the USB 2.0 PHY transceivers.
>>
>> Signed-off-by: hpchen0 <hpchen0nvt at gmail.com>
>> +
>> +	ret = clk_prepare_enable(p_phy->clk);
>> +	if (ret < 0) {
>> +		dev_err(p_phy->dev, "Failed to enable PHY clock: %d\n", ret);
>> +		return ret;
>> +	}
>> +
>> +	regmap_read(p_phy->sysreg, MA35_SYS_REG_USBPMISCR, &val);
>> +	if (val & PHY0SUSPEND) {
>> +		/*
>> +		 * USB PHY0 is in operation mode already
>> +		 * make sure USB PHY 60 MHz UTMI Interface Clock ready
>> +		 */
>> +		timeout = jiffies + msecs_to_jiffies(200);
>> +		while (time_before(jiffies, timeout)) {
>> +			regmap_read(p_phy->sysreg, MA35_SYS_REG_USBPMISCR, &val);
>> +			if (val & PHY0DEVCKSTB)
>> +				return 0;
>> +			usleep_range(1000, 1500);
>> +		}
> You want some readl_poll_timeout version here.

Okay. The readl_poll_timeout function will be used instead.



>> +	}
>> +
>> +	/*
>> +	 * reset USB PHY0.
>> +	 * wait until USB PHY0 60 MHz UTMI Interface Clock ready
>> +	 */
>> +	regmap_update_bits(p_phy->sysreg, MA35_SYS_REG_USBPMISCR, 0x7, (PHY0POR | PHY0SUSPEND));
>> +	timeout = jiffies + msecs_to_jiffies(200);
>> +	while (time_before(jiffies, timeout)) {
>> +		regmap_read(p_phy->sysreg, MA35_SYS_REG_USBPMISCR, &val);
>> +		if (val & PHY0DEVCKSTB)
>> +			break;
>> +		usleep_range(1000, 1500);
>> +	}
>> +
>> +	/* make USB PHY0 enter operation mode */
>> +	regmap_update_bits(p_phy->sysreg, MA35_SYS_REG_USBPMISCR, 0x7, PHY0SUSPEND);
>> +
>> +	/* make sure USB PHY 60 MHz UTMI Interface Clock ready */
>> +	timeout = jiffies + msecs_to_jiffies(200);
>> +	while (time_before(jiffies, timeout)) {
>> +		regmap_read(p_phy->sysreg, MA35_SYS_REG_USBPMISCR, &val);
>> +		if (val & PHY0DEVCKSTB)
>> +			return 0;
>> +		usleep_range(1000, 1500);
>> +	}
>> +
>> +	dev_err(p_phy->dev, "Timed out waiting for PHY to power on\n");
>> +	ret = -ETIMEDOUT;
>> +
>> +	clk_disable_unprepare(p_phy->clk);
>> +	return ret;
>> +}
>> +
>> +static int ma35_usb_phy_power_off(struct phy *phy)
>> +{
>> +	struct ma35_usb_phy *p_phy = phy_get_drvdata(phy);
>> +
>> +	clk_disable_unprepare(p_phy->clk);
>> +	return 0;
>> +}
>> +
>> +static const struct phy_ops ma35_usb_phy_ops = {
>> +	.power_on = ma35_usb_phy_power_on,
>> +	.power_off = ma35_usb_phy_power_off,
>> +	.owner = THIS_MODULE,
>> +};
>> +
>> +static int ma35_usb_phy_probe(struct platform_device *pdev)
>> +{
>> +	struct phy_provider *provider;
>> +	struct ma35_usb_phy *p_phy;
>> +	const char *clkgate;
>> +	struct phy *phy;
>> +
>> +	p_phy = devm_kzalloc(&pdev->dev, sizeof(*p_phy), GFP_KERNEL);
>> +	if (!p_phy)
>> +		return -ENOMEM;
>> +
>> +	p_phy->dev = &pdev->dev;
>> +	platform_set_drvdata(pdev, p_phy);
>> +
>> +	p_phy->sysreg = syscon_regmap_lookup_by_phandle(p_phy->dev->of_node, "nuvoton,sys");
>> +	if (IS_ERR(p_phy->sysreg))
>> +		return dev_err_probe(&pdev->dev, PTR_ERR(p_phy->sysreg),
>> +				     "Failed to get SYS registers\n");
>> +
>> +	/* enable clock */
>> +	of_property_read_string(p_phy->dev->of_node, "clock-enable", &clkgate);
> There is no such property.

I'm sorry, I forgot to remove this part. I will remove it and correct it.



>> +	p_phy->clk = devm_clk_get(p_phy->dev, clkgate);
> Don't mix styles of variables: you were using pdev->dev but now entirely
> different. Stick to pdev->dev.

Okay. I will consistently  use pdev->dev|.|

Thank you for the reminder.



>> +	if (IS_ERR(p_phy->clk))
>> +		return dev_err_probe(&pdev->dev, PTR_ERR(p_phy->clk),
> And here again pdev->dev... Bring some consistency, not random coding style.

Okay. I will consistently  use pdev->dev|.|

Thank you for the reminder.



>> +				     "Failed to get usb_phy clock\n");
>> +
>
> Best regards,
> Krzysztof


Best regards,

Hui-Ping Chen