RE: [PATCH v3 5/9] phy: freescale: pcie: initialize the imx8 pcie standalone phy driver

[Richard Zhu <hongxing.zhu@nxp.com>]

> -----Original Message-----
> From: Lucas Stach <l.stach@pengutronix.de>
> Sent: Saturday, October 16, 2021 2:56 AM
> To: Richard Zhu <hongxing.zhu@nxp.com>; tharvey@gateworks.com;
> kishon@ti.com; vkoul@kernel.org; robh@kernel.org;
> galak@kernel.crashing.org; shawnguo@kernel.org
> Cc: linux-phy@lists.infradead.org; devicetree@vger.kernel.org;
> linux-arm-kernel@lists.infradead.org; linux-kernel@vger.kernel.org;
> kernel@pengutronix.de; dl-linux-imx <linux-imx@nxp.com>
> Subject: Re: [PATCH v3 5/9] phy: freescale: pcie: initialize the imx8 pcie
> standalone phy driver
> 
> Am Dienstag, dem 12.10.2021 um 16:41 +0800 schrieb Richard Zhu:
> > Add the standalone i.MX8 PCIe PHY driver.
> > Some reset bits should be manipulated between PHY configurations and
> > status check(internal PLL is locked or not).
> > So, do the PHY configuration in the phy_calibrate().
> > And check the PHY is ready or not in the phy_init().
> >
> > Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com>
> > ---
> >  drivers/phy/freescale/Kconfig              |   9 +
> >  drivers/phy/freescale/Makefile             |   1 +
> >  drivers/phy/freescale/phy-fsl-imx8m-pcie.c | 218
> > +++++++++++++++++++++
> >  3 files changed, 228 insertions(+)
> >  create mode 100644 drivers/phy/freescale/phy-fsl-imx8m-pcie.c
> >
> > diff --git a/drivers/phy/freescale/Kconfig
> > b/drivers/phy/freescale/Kconfig index 320630ffe3cd..fb08e5242602
> > 100644
> > --- a/drivers/phy/freescale/Kconfig
> > +++ b/drivers/phy/freescale/Kconfig
> > @@ -14,3 +14,12 @@ config PHY_MIXEL_MIPI_DPHY
> >  	help
> >  	  Enable this to add support for the Mixel DSI PHY as found
> >  	  on NXP's i.MX8 family of SOCs.
> > +
> > +config PHY_FSL_IMX8M_PCIE
> > +	tristate "Freescale i.MX8 PCIE PHY"
> > +	depends on OF && HAS_IOMEM
> > +	select GENERIC_PHY
> > +	default ARCH_MXC
> 
> Not sure if we need this default, but if you want to keep it add at least a "&&
> ARM64" to not enable it on old 32bit i.MX SoCs.
[Richard Zhu] Add the "&& ARM64" is preferred. Thanks.
> 
> > +	help
> > +	  Enable this to add support for the PCIE PHY as found on
> > +	  i.MX8M family of SOCs.
> > diff --git a/drivers/phy/freescale/Makefile
> > b/drivers/phy/freescale/Makefile index 1d02e3869b45..55d07c742ab0
> > 100644
> > --- a/drivers/phy/freescale/Makefile
> > +++ b/drivers/phy/freescale/Makefile
> > @@ -1,3 +1,4 @@
> >  # SPDX-License-Identifier: GPL-2.0-only
> >  obj-$(CONFIG_PHY_FSL_IMX8MQ_USB)	+= phy-fsl-imx8mq-usb.o
> >  obj-$(CONFIG_PHY_MIXEL_MIPI_DPHY)	+= phy-fsl-imx8-mipi-dphy.o
> > +obj-$(CONFIG_PHY_FSL_IMX8M_PCIE)	+= phy-fsl-imx8m-pcie.o
> > diff --git a/drivers/phy/freescale/phy-fsl-imx8m-pcie.c
> > b/drivers/phy/freescale/phy-fsl-imx8m-pcie.c
> > new file mode 100644
> > index 000000000000..317cf61bff37
> > --- /dev/null
> > +++ b/drivers/phy/freescale/phy-fsl-imx8m-pcie.c
> > @@ -0,0 +1,218 @@
> > +// SPDX-License-Identifier: GPL-2.0+
> > +/*
> > + * Copyright 2021 NXP
> > + */
> > +
> > +#include <linux/clk.h>
> > +#include <linux/io.h>
> > +#include <linux/iopoll.h>
> > +#include <linux/delay.h>
> > +#include <linux/mfd/syscon.h>
> > +#include <linux/mfd/syscon/imx7-iomuxc-gpr.h>
> > +#include <linux/module.h>
> > +#include <linux/phy/phy.h>
> > +#include <linux/platform_device.h>
> > +#include <linux/regmap.h>
> > +#include <linux/reset.h>
> > +#include <dt-bindings/phy/phy-imx8-pcie.h>
> > +
> > +#define IMX8MM_PCIE_PHY_CMN_REG061	0x184
> > +#define  ANA_PLL_CLK_OUT_TO_EXT_IO_EN	BIT(0)
> > +#define IMX8MM_PCIE_PHY_CMN_REG062	0x188
> > +#define  ANA_PLL_CLK_OUT_TO_EXT_IO_SEL	BIT(3)
> > +#define IMX8MM_PCIE_PHY_CMN_REG063	0x18C
> > +#define  AUX_PLL_REFCLK_SEL_SYS_PLL	GENMASK(7, 6)
> > +#define IMX8MM_PCIE_PHY_CMN_REG064	0x190
> > +#define  ANA_AUX_RX_TX_SEL_TX		BIT(7)
> > +#define  ANA_AUX_RX_TERM_GND_EN		BIT(3)
> > +#define  ANA_AUX_TX_TERM		BIT(2)
> > +#define IMX8MM_PCIE_PHY_CMN_REG065	0x194
> > +#define  ANA_AUX_RX_TERM		(BIT(7) | BIT(4))
> > +#define  ANA_AUX_TX_LVL			GENMASK(3, 0)
> > +#define IMX8MM_PCIE_PHY_CMN_REG75	0x1D4
> > +#define  PCIE_PHY_CMN_REG75_PLL_DONE	0x3
> > +#define PCIE_PHY_TRSV_REG5		0x414
> > +#define  PCIE_PHY_TRSV_REG5_GEN1_DEEMP	0x2D
> > +#define PCIE_PHY_TRSV_REG6		0x418
> > +#define  PCIE_PHY_TRSV_REG6_GEN2_DEEMP	0xF
> > +
> > +#define IMX8MM_GPR_PCIE_REF_CLK_SEL	GENMASK(25, 24)
> > +#define IMX8MM_GPR_PCIE_REF_CLK_PLL
> 	FIELD_PREP(IMX8MM_GPR_PCIE_REF_CLK_SEL, 0x3)
> > +#define IMX8MM_GPR_PCIE_REF_CLK_EXT
> 	FIELD_PREP(IMX8MM_GPR_PCIE_REF_CLK_SEL, 0x2)
> > +#define IMX8MM_GPR_PCIE_AUX_EN		BIT(19)
> > +#define IMX8MM_GPR_PCIE_CMN_RST		BIT(18)
> > +#define IMX8MM_GPR_PCIE_POWER_OFF	BIT(17)
> > +#define IMX8MM_GPR_PCIE_SSC_EN		BIT(16)
> > +#define IMX8MM_GPR_PCIE_REF_USE_PAD	BIT(9)
> > +
> > +struct imx8_pcie_phy {
> > +	u32			refclk_pad_mode;
> 
> Moving this to the end of the structure will get rid of a hole due to padding.
[Richard Zhu] Got that, thanks.

> 
> > +	void __iomem		*base;
> > +	struct clk		*clk;
> > +	struct phy		*phy;
> > +	struct regmap		*iomuxc_gpr;
> > +	struct reset_control	*reset;
> > +};
> > +
> > +static int imx8_pcie_phy_init(struct phy *phy) {
> > +	int ret;
> > +	u32 val, pad_mode;
> > +	struct imx8_pcie_phy *imx8_phy = phy_get_drvdata(phy);
> > +
> > +	reset_control_assert(imx8_phy->reset);
> > +
> > +	regmap_update_bits(imx8_phy->iomuxc_gpr, IOMUXC_GPR14,
> > +			   IMX8MM_GPR_PCIE_REF_USE_PAD,
> > +			   imx8_phy->refclk_pad_mode == 1 ?
> > +			   IMX8MM_GPR_PCIE_REF_USE_PAD : 0);
> > +	regmap_update_bits(imx8_phy->iomuxc_gpr, IOMUXC_GPR14,
> > +			   IMX8MM_GPR_PCIE_AUX_EN,
> > +			   IMX8MM_GPR_PCIE_AUX_EN);
> > +	regmap_update_bits(imx8_phy->iomuxc_gpr, IOMUXC_GPR14,
> > +			   IMX8MM_GPR_PCIE_POWER_OFF, 0);
> > +	regmap_update_bits(imx8_phy->iomuxc_gpr, IOMUXC_GPR14,
> > +			   IMX8MM_GPR_PCIE_SSC_EN, 0);
> > +
> > +	regmap_update_bits(imx8_phy->iomuxc_gpr, IOMUXC_GPR14,
> > +			   IMX8MM_GPR_PCIE_REF_CLK_SEL,
> > +			   imx8_phy->refclk_pad_mode == 1 ?
> > +			   IMX8MM_GPR_PCIE_REF_CLK_EXT :
> > +			   IMX8MM_GPR_PCIE_REF_CLK_PLL);
> > +	usleep_range(100, 200);
> > +
> > +	/* Do the PHY common block reset */
> > +	regmap_update_bits(imx8_phy->iomuxc_gpr, IOMUXC_GPR14,
> > +			   IMX8MM_GPR_PCIE_CMN_RST,
> > +			   IMX8MM_GPR_PCIE_CMN_RST);
> > +	usleep_range(200, 500);
> > +
> > +
> > +	pad_mode = imx8_phy->refclk_pad_mode;
> > +	if (pad_mode == IMX8_PCIE_REFCLK_PAD_INPUT) {
> > +		/* Configure the pad as input */
> > +		val = readl(imx8_phy->base + IMX8MM_PCIE_PHY_CMN_REG061);
> > +		writel(val & ~ANA_PLL_CLK_OUT_TO_EXT_IO_EN,
> > +		       imx8_phy->base + IMX8MM_PCIE_PHY_CMN_REG061);
> > +	} else if (pad_mode == IMX8_PCIE_REFCLK_PAD_OUTPUT) {
> > +		/* Configure the PHY to output the refclock via pad */
> > +		writel(ANA_PLL_CLK_OUT_TO_EXT_IO_EN,
> > +		       imx8_phy->base + IMX8MM_PCIE_PHY_CMN_REG061);
> > +		writel(ANA_PLL_CLK_OUT_TO_EXT_IO_SEL,
> > +		       imx8_phy->base + IMX8MM_PCIE_PHY_CMN_REG062);
> > +		writel(AUX_PLL_REFCLK_SEL_SYS_PLL,
> > +		       imx8_phy->base + IMX8MM_PCIE_PHY_CMN_REG063);
> > +		val = ANA_AUX_RX_TX_SEL_TX | ANA_AUX_TX_TERM;
> > +		writel(val | ANA_AUX_RX_TERM_GND_EN,
> > +		       imx8_phy->base + IMX8MM_PCIE_PHY_CMN_REG064);
> > +		writel(ANA_AUX_RX_TERM | ANA_AUX_TX_LVL,
> > +		       imx8_phy->base + IMX8MM_PCIE_PHY_CMN_REG065);
> > +	}
> > +
> > +	/* Tune PHY de-emphasis setting to pass PCIe compliance. */
> > +	writel(PCIE_PHY_TRSV_REG5_GEN1_DEEMP,
> > +	       imx8_phy->base + PCIE_PHY_TRSV_REG5);
> > +	writel(PCIE_PHY_TRSV_REG6_GEN2_DEEMP,
> > +	       imx8_phy->base + PCIE_PHY_TRSV_REG6);
> 
> I guess those are at least in part board specific. Maybe it would be a good
> idea to add DT properties to allow each board to specify the actual
> de-emphasis values?
[Richard Zhu] Agree. Would use two properties to specified them by HW board.
Thanks.

> 
> > +
> > +	reset_control_deassert(imx8_phy->reset);
> > +
> > +	/* Polling to check the phy is ready or not. */
> > +	ret = readl_poll_timeout(imx8_phy->base +
> IMX8MM_PCIE_PHY_CMN_REG75,
> > +				 val, val == PCIE_PHY_CMN_REG75_PLL_DONE,
> > +				 10, 20000);
> > +	return ret;
> > +}
> > +
> > +static int imx8_pcie_phy_power_on(struct phy *phy) {
> > +	struct imx8_pcie_phy *imx8_phy = phy_get_drvdata(phy);
> > +
> > +	return clk_prepare_enable(imx8_phy->clk);
> > +}
> > +
> > +static int imx8_pcie_phy_power_off(struct phy *phy) {
> > +	struct imx8_pcie_phy *imx8_phy = phy_get_drvdata(phy);
> > +
> > +	clk_disable_unprepare(imx8_phy->clk);
> > +
> > +	return 0;
> > +}
> > +
> > +static const struct phy_ops imx8_pcie_phy_ops = {
> > +	.init		= imx8_pcie_phy_init,
> > +	.power_on	= imx8_pcie_phy_power_on,
> > +	.power_off	= imx8_pcie_phy_power_off,
> > +	.owner		= THIS_MODULE,
> > +};
> > +
> > +static int imx8_pcie_phy_probe(struct platform_device *pdev) {
> > +	struct phy_provider *phy_provider;
> > +	struct device *dev = &pdev->dev;
> > +	struct device_node *np = dev->of_node;
> > +	struct imx8_pcie_phy *imx8_phy;
> > +	struct resource *res;
> > +
> > +	imx8_phy = devm_kzalloc(dev, sizeof(*imx8_phy), GFP_KERNEL);
> > +	if (!imx8_phy)
> > +		return -ENOMEM;
> > +
> > +	/* get PHY refclk pad mode */
> > +	of_property_read_u32(np, "fsl,refclk-pad-mode",
> > +			     &imx8_phy->refclk_pad_mode);
> > +
> > +	imx8_phy->clk = devm_clk_get(dev, "phy");
> > +	if (IS_ERR(imx8_phy->clk)) {
> > +		dev_err(dev, "failed to get imx pcie phy clock\n");
> > +		return PTR_ERR(imx8_phy->clk);
> > +	}
> > +
> > +	/* Grab GPR config register range */
> > +	imx8_phy->iomuxc_gpr =
> > +		 syscon_regmap_lookup_by_compatible("fsl,imx6q-iomuxc-gpr");
> 
> I would prefer if we could get the syscon via a phandle and make the GPR
> register to use another cell of the handle.
> 
> That's not really important for the 8MM, where there is only one PCIe PHY,
> but if we want to retrofit this model to the i.MX8MQ, we need to deal with
> different GPR offsets for both PHYs and I would like to have this explicit in the
> DT. Doing the same here would be good for consistency.
[Richard Zhu] It seems that this is a good idea.
But I'm not clear about how to do about that.
IOMUXC_GPR module had been defined with 64Kbytes size and "syscon" compatible.
On i.MX8MQ, GPR14/15 are used for PCIe0 PHY, and GPR16/17 are used for PCIe1 PHY.
How to explicit these different offset in the different phandles?

BR
Richard

> 
> Overall, looks much better than the last submission.
> 
> Regards,
> Lucas
> 
> > +	if (IS_ERR(imx8_phy->iomuxc_gpr)) {
> > +		dev_err(dev, "unable to find iomuxc registers\n");
> > +		return PTR_ERR(imx8_phy->iomuxc_gpr);
> > +	}
> > +
> > +	imx8_phy->reset = devm_reset_control_get_exclusive(dev, "pciephy");
> > +	if (IS_ERR(imx8_phy->reset)) {
> > +		dev_err(dev, "Failed to get PCIEPHY reset control\n");
> > +		return PTR_ERR(imx8_phy->reset);
> > +	}
> > +
> > +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> > +	imx8_phy->base = devm_ioremap_resource(dev, res);
> > +	if (IS_ERR(imx8_phy->base))
> > +		return PTR_ERR(imx8_phy->base);
> > +
> > +	imx8_phy->phy = devm_phy_create(dev, NULL, &imx8_pcie_phy_ops);
> > +	if (IS_ERR(imx8_phy->phy))
> > +		return PTR_ERR(imx8_phy->phy);
> > +
> > +	phy_set_drvdata(imx8_phy->phy, imx8_phy);
> > +
> > +	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 imx8_pcie_phy_of_match[] = {
> > +	{.compatible = "fsl,imx8mm-pcie-phy",},
> > +	{ },
> > +};
> > +MODULE_DEVICE_TABLE(of, imx8_pcie_phy_of_match);
> > +
> > +static struct platform_driver imx8_pcie_phy_driver = {
> > +	.probe	= imx8_pcie_phy_probe,
> > +	.driver = {
> > +		.name	= "imx8-pcie-phy",
> > +		.of_match_table	= imx8_pcie_phy_of_match,
> > +	}
> > +};
> > +module_platform_driver(imx8_pcie_phy_driver);
> > +
> > +MODULE_DESCRIPTION("FSL IMX8 PCIE PHY driver");
> > +MODULE_LICENSE("GPL");
>