Re: [PATCH v7 3/3] phy: intel: Add Keem Bay eMMC PHY support

[Andy Shevchenko]

On Fri, Aug 21, 2020 at 07:37:47PM +0800, Wan Ahmad Zainie wrote:
> Add support for eMMC PHY on Intel Keem Bay SoC.

Thanks for an update. I have couple of nit-picks below, otherwise, FWIW,
Reviewed-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxx>

> Signed-off-by: Wan Ahmad Zainie <wan.ahmad.zainie.wan.mohamad@xxxxxxxxx>
> ---
>  drivers/phy/intel/Kconfig                  |  12 +
>  drivers/phy/intel/Makefile                 |   1 +
>  drivers/phy/intel/phy-intel-keembay-emmc.c | 310 +++++++++++++++++++++
>  3 files changed, 323 insertions(+)
>  create mode 100644 drivers/phy/intel/phy-intel-keembay-emmc.c
> 
> diff --git a/drivers/phy/intel/Kconfig b/drivers/phy/intel/Kconfig
> index db8586c3eed8..dad72fc64871 100644
> --- a/drivers/phy/intel/Kconfig
> +++ b/drivers/phy/intel/Kconfig
> @@ -2,6 +2,18 @@
>  #
>  # Phy drivers for Intel platforms
>  #
> +config PHY_INTEL_KEEMBAY_EMMC
> +	tristate "Intel Keem Bay EMMC PHY driver"

> +	depends on ARM64 || COMPILE_TEST
> +	depends on OF && HAS_IOMEM

It might be better from testing coverage perspective to have it like

	depends on HAS_IOMEM
	depends on (OF && ARM64) || COMPILE_TEST

> +	select GENERIC_PHY
> +	select REGMAP_MMIO
> +	help
> +	  Choose this option if you have an Intel Keem Bay SoC.
> +
> +	  To compile this driver as a module, choose M here: the module
> +	  will be called phy-keembay-emmc.ko.
> +
>  config PHY_INTEL_LGM_COMBO
>  	bool "Intel Lightning Mountain ComboPHY driver"
>  	depends on X86 || COMPILE_TEST
> diff --git a/drivers/phy/intel/Makefile b/drivers/phy/intel/Makefile
> index 662385d0a366..a5e0af5ccd75 100644
> --- a/drivers/phy/intel/Makefile
> +++ b/drivers/phy/intel/Makefile
> @@ -1,3 +1,4 @@
>  # SPDX-License-Identifier: GPL-2.0
> +obj-$(CONFIG_PHY_INTEL_KEEMBAY_EMMC)	+= phy-intel-keembay-emmc.o
>  obj-$(CONFIG_PHY_INTEL_LGM_COMBO)	+= phy-intel-lgm-combo.o
>  obj-$(CONFIG_PHY_INTEL_LGM_EMMC)	+= phy-intel-lgm-emmc.o
> diff --git a/drivers/phy/intel/phy-intel-keembay-emmc.c b/drivers/phy/intel/phy-intel-keembay-emmc.c
> new file mode 100644
> index 000000000000..c1a095c17165
> --- /dev/null
> +++ b/drivers/phy/intel/phy-intel-keembay-emmc.c
> @@ -0,0 +1,310 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * Intel Keem Bay eMMC PHY driver
> + * Copyright (C) 2020 Intel Corporation
> + */

Missed bitops.h

> +#include <linux/clk.h>
> +#include <linux/delay.h>
> +#include <linux/mfd/syscon.h>
> +#include <linux/module.h>
> +#include <linux/of.h>
> +#include <linux/of_address.h>
> +#include <linux/phy/phy.h>
> +#include <linux/platform_device.h>
> +#include <linux/regmap.h>
> +
> +/* eMMC/SD/SDIO core/phy configuration registers */
> +#define PHY_CFG_0		0x24
> +#define  SEL_DLY_TXCLK_MASK	BIT(29)
> +#define  OTAP_DLY_ENA_MASK	BIT(27)
> +#define  OTAP_DLY_SEL_MASK	GENMASK(26, 23)
> +#define  DLL_EN_MASK		BIT(10)
> +#define  PWR_DOWN_MASK		BIT(0)
> +
> +#define PHY_CFG_2		0x2c
> +#define  SEL_FREQ_MASK		GENMASK(12, 10)
> +
> +#define PHY_STAT		0x40
> +#define  CAL_DONE_MASK		BIT(6)
> +#define  IS_CALDONE(x)		((x) & CAL_DONE_MASK)
> +#define  DLL_RDY_MASK		BIT(5)
> +#define  IS_DLLRDY(x)		((x) & DLL_RDY_MASK)
> +
> +/* From ACS_eMMC51_16nFFC_RO1100_Userguide_v1p0.pdf p17 */
> +#define FREQSEL_200M_170M	0x0
> +#define FREQSEL_170M_140M	0x1
> +#define FREQSEL_140M_110M	0x2
> +#define FREQSEL_110M_80M	0x3
> +#define FREQSEL_80M_50M		0x4
> +
> +#define maskval(mask, val)	(((val) << (ffs(mask) - 1)) & mask)
> +
> +struct keembay_emmc_phy {
> +	struct regmap *syscfg;
> +	struct clk *emmcclk;
> +};
> +
> +static const struct regmap_config keembay_regmap_config = {
> +	.reg_bits = 32,
> +	.val_bits = 32,
> +	.reg_stride = 4,
> +};
> +
> +static int keembay_emmc_phy_power(struct phy *phy, bool on_off)
> +{
> +	struct keembay_emmc_phy *priv = phy_get_drvdata(phy);
> +	unsigned int caldone;
> +	unsigned int dllrdy;
> +	unsigned int freqsel;
> +	unsigned int mhz;
> +	int ret;
> +
> +	/*
> +	 * Keep phyctrl_pdb and phyctrl_endll low to allow
> +	 * initialization of CALIO state M/C DFFs
> +	 */
> +	ret = regmap_update_bits(priv->syscfg, PHY_CFG_0, PWR_DOWN_MASK,
> +				 maskval(PWR_DOWN_MASK, 0));
> +	if (ret) {
> +		dev_err(&phy->dev, "CALIO power down bar failed: %d\n", ret);
> +		return ret;
> +	}
> +
> +	ret = regmap_update_bits(priv->syscfg, PHY_CFG_0, DLL_EN_MASK,
> +				 maskval(DLL_EN_MASK, 0));
> +	if (ret) {
> +		dev_err(&phy->dev, "turn off the dll failed: %d\n", ret);
> +		return ret;
> +	}
> +
> +	/* Already finish power off above */
> +	if (!on_off)
> +		return 0;
> +
> +	mhz = DIV_ROUND_CLOSEST(clk_get_rate(priv->emmcclk), 1000000);
> +	if (mhz <= 200 && mhz >= 170)
> +		freqsel = FREQSEL_200M_170M;
> +	else if (mhz <= 170 && mhz >= 140)
> +		freqsel = FREQSEL_170M_140M;
> +	else if (mhz <= 140 && mhz >= 110)
> +		freqsel = FREQSEL_140M_110M;
> +	else if (mhz <= 110 && mhz >= 80)
> +		freqsel = FREQSEL_110M_80M;
> +	else if (mhz <= 80 && mhz >= 50)
> +		freqsel = FREQSEL_80M_50M;
> +	else
> +		freqsel = 0x0;
> +
> +	if (mhz < 50 || mhz > 200)
> +		dev_warn(&phy->dev, "Unsupported rate: %d MHz\n", mhz);
> +
> +	/*
> +	 * According to the user manual, calpad calibration
> +	 * cycle takes more than 2us without the minimal recommended
> +	 * value, so we may need a little margin here
> +	 */
> +	udelay(5);
> +
> +	ret = regmap_update_bits(priv->syscfg, PHY_CFG_0, PWR_DOWN_MASK,
> +				 maskval(PWR_DOWN_MASK, 1));
> +	if (ret) {
> +		dev_err(&phy->dev, "CALIO power down bar failed: %d\n", ret);
> +		return ret;
> +	}
> +
> +	/*
> +	 * According to the user manual, it asks driver to wait 5us for
> +	 * calpad busy trimming. However it is documented that this value is
> +	 * PVT(A.K.A. process, voltage and temperature) relevant, so some
> +	 * failure cases are found which indicates we should be more tolerant
> +	 * to calpad busy trimming.
> +	 */
> +	ret = regmap_read_poll_timeout(priv->syscfg, PHY_STAT,
> +				       caldone, IS_CALDONE(caldone),
> +				       0, 50);
> +	if (ret) {
> +		dev_err(&phy->dev, "caldone failed, ret=%d\n", ret);
> +		return ret;
> +	}
> +
> +	/* Set the frequency of the DLL operation */
> +	ret = regmap_update_bits(priv->syscfg, PHY_CFG_2, SEL_FREQ_MASK,
> +				 maskval(SEL_FREQ_MASK, freqsel));
> +	if (ret) {
> +		dev_err(&phy->dev, "set the frequency of dll failed:%d\n", ret);
> +		return ret;
> +	}
> +
> +	/* Turn on the DLL */
> +	ret = regmap_update_bits(priv->syscfg, PHY_CFG_0, DLL_EN_MASK,
> +				 maskval(DLL_EN_MASK, 1));
> +	if (ret) {
> +		dev_err(&phy->dev, "turn on the dll failed: %d\n", ret);
> +		return ret;
> +	}
> +
> +	/*
> +	 * We turned on the DLL even though the rate was 0 because we the
> +	 * clock might be turned on later.  ...but we can't wait for the DLL
> +	 * to lock when the rate is 0 because it will never lock with no
> +	 * input clock.
> +	 *
> +	 * Technically we should be checking the lock later when the clock
> +	 * is turned on, but for now we won't.
> +	 */
> +	if (mhz == 0)
> +		return 0;
> +
> +	/*
> +	 * After enabling analog DLL circuits docs say that we need 10.2 us if
> +	 * our source clock is at 50 MHz and that lock time scales linearly
> +	 * with clock speed. If we are powering on the PHY and the card clock
> +	 * is super slow (like 100kHz) this could take as long as 5.1 ms as
> +	 * per the math: 10.2 us * (50000000 Hz / 100000 Hz) => 5.1 ms
> +	 * hopefully we won't be running at 100 kHz, but we should still make
> +	 * sure we wait long enough.
> +	 *
> +	 * NOTE: There appear to be corner cases where the DLL seems to take
> +	 * extra long to lock for reasons that aren't understood. In some
> +	 * extreme cases we've seen it take up to over 10ms (!). We'll be
> +	 * generous and give it 50ms.
> +	 */
> +	ret = regmap_read_poll_timeout(priv->syscfg, PHY_STAT,
> +				       dllrdy, IS_DLLRDY(dllrdy),
> +				       0, 50 * USEC_PER_MSEC);
> +	if (ret)
> +		dev_err(&phy->dev, "dllrdy failed, ret=%d\n", ret);
> +
> +	return ret;
> +}
> +
> +static int keembay_emmc_phy_init(struct phy *phy)
> +{
> +	struct keembay_emmc_phy *priv = phy_get_drvdata(phy);
> +
> +	/*
> +	 * We purposely get the clock here and not in probe to avoid the
> +	 * circular dependency problem. We expect:
> +	 * - PHY driver to probe
> +	 * - SDHCI driver to start probe
> +	 * - SDHCI driver to register it's clock
> +	 * - SDHCI driver to get the PHY
> +	 * - SDHCI driver to init the PHY
> +	 *
> +	 * The clock is optional, so upon any error just return it like
> +	 * any other error to user.
> +	 */
> +	priv->emmcclk = clk_get_optional(&phy->dev, "emmcclk");
> +
> +	return PTR_ERR_OR_ZERO(priv->emmcclk);
> +}
> +
> +static int keembay_emmc_phy_exit(struct phy *phy)
> +{
> +	struct keembay_emmc_phy *priv = phy_get_drvdata(phy);
> +
> +	clk_put(priv->emmcclk);
> +
> +	return 0;
> +};
> +
> +static int keembay_emmc_phy_power_on(struct phy *phy)
> +{
> +	struct keembay_emmc_phy *priv = phy_get_drvdata(phy);
> +	int ret;
> +
> +	/* Delay chain based txclk: enable */
> +	ret = regmap_update_bits(priv->syscfg, PHY_CFG_0, SEL_DLY_TXCLK_MASK,
> +				 maskval(SEL_DLY_TXCLK_MASK, 1));
> +	if (ret) {
> +		dev_err(&phy->dev, "ERROR: delay chain txclk set: %d\n", ret);
> +		return ret;
> +	}
> +
> +	/* Output tap delay: enable */
> +	ret = regmap_update_bits(priv->syscfg, PHY_CFG_0, OTAP_DLY_ENA_MASK,
> +				 maskval(OTAP_DLY_ENA_MASK, 1));
> +	if (ret) {
> +		dev_err(&phy->dev, "ERROR: output tap delay set: %d\n", ret);
> +		return ret;
> +	}
> +
> +	/* Output tap delay */
> +	ret = regmap_update_bits(priv->syscfg, PHY_CFG_0, OTAP_DLY_SEL_MASK,
> +				 maskval(OTAP_DLY_SEL_MASK, 2));
> +	if (ret) {
> +		dev_err(&phy->dev, "ERROR: output tap delay select: %d\n", ret);
> +		return ret;
> +	}
> +
> +	/* Power up eMMC phy analog blocks */
> +	return keembay_emmc_phy_power(phy, true);
> +}
> +
> +static int keembay_emmc_phy_power_off(struct phy *phy)
> +{
> +	/* Power down eMMC phy analog blocks */
> +	return keembay_emmc_phy_power(phy, false);
> +}
> +
> +static const struct phy_ops ops = {
> +	.init		= keembay_emmc_phy_init,
> +	.exit		= keembay_emmc_phy_exit,
> +	.power_on	= keembay_emmc_phy_power_on,
> +	.power_off	= keembay_emmc_phy_power_off,
> +	.owner		= THIS_MODULE,
> +};
> +
> +static int keembay_emmc_phy_probe(struct platform_device *pdev)
> +{
> +	struct device *dev = &pdev->dev;
> +	struct device_node *np = dev->of_node;
> +	struct keembay_emmc_phy *priv;
> +	struct phy *generic_phy;
> +	struct phy_provider *phy_provider;
> +	void __iomem *base;
> +
> +	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
> +	if (!priv)
> +		return -ENOMEM;
> +
> +	base = devm_platform_ioremap_resource(pdev, 0);
> +	if (IS_ERR(base))
> +		return PTR_ERR(base);
> +
> +	priv->syscfg = devm_regmap_init_mmio(dev, base, &keembay_regmap_config);
> +	if (IS_ERR(priv->syscfg))
> +		return PTR_ERR(priv->syscfg);
> +
> +	generic_phy = devm_phy_create(dev, np, &ops);
> +	if (IS_ERR(generic_phy)) {

> +		dev_err(dev, "failed to create PHY\n");
> +		return PTR_ERR(generic_phy);

I guess now you may use

	return dev_err_probe(...);

> +	}
> +
> +	phy_set_drvdata(generic_phy, priv);
> +	phy_provider = devm_of_phy_provider_register(dev, of_phy_simple_xlate);
> +
> +	return PTR_ERR_OR_ZERO(phy_provider);
> +}
> +
> +static const struct of_device_id keembay_emmc_phy_dt_ids[] = {
> +	{ .compatible = "intel,keembay-emmc-phy" },
> +	{}
> +};
> +
> +MODULE_DEVICE_TABLE(of, keembay_emmc_phy_dt_ids);
> +
> +static struct platform_driver keembay_emmc_phy_driver = {
> +	.probe		= keembay_emmc_phy_probe,
> +	.driver		= {
> +		.name	= "keembay-emmc-phy",
> +		.of_match_table = keembay_emmc_phy_dt_ids,
> +	},
> +};
> +module_platform_driver(keembay_emmc_phy_driver);
> +
> +MODULE_AUTHOR("Wan Ahmad Zainie <wan.ahmad.zainie.wan.mohamad@xxxxxxxxx>");
> +MODULE_DESCRIPTION("Intel Keem Bay eMMC PHY driver");
> +MODULE_LICENSE("GPL v2");
> -- 
> 2.17.1
> 

-- 
With Best Regards,
Andy Shevchenko