[PATCH v4 5/8] phy: freescale: pcie: Initialize the imx8 pcie standalone phy driver
Richard Zhu
hongxing.zhu at nxp.com
Fri Oct 29 01:45:03 PDT 2021
> -----Original Message-----
> From: Marcel Ziswiler <marcel.ziswiler at toradex.com>
> Sent: Friday, October 29, 2021 4:13 PM
> To: kishon at ti.com; vkoul at kernel.org; robh at kernel.org;
> l.stach at pengutronix.de; shawnguo at kernel.org;
> tharvey at gateworks.com; galak at kernel.crashing.org; Richard Zhu
> <hongxing.zhu at nxp.com>
> Cc: linux-phy at lists.infradead.org; linux-arm-kernel at lists.infradead.org;
> kernel at pengutronix.de; devicetree at vger.kernel.org;
> linux-kernel at vger.kernel.org; dl-linux-imx <linux-imx at nxp.com>
> Subject: Re: [PATCH v4 5/8] phy: freescale: pcie: Initialize the imx8 pcie
> standalone phy driver
>
> On Thu, 2021-10-28 at 15:27 +0800, Richard Zhu wrote:
> > Add the standalone i.MX8 PCIe PHY driver.
> >
> > Signed-off-by: Richard Zhu <hongxing.zhu at nxp.com>
> > Tested-by: Marcel Ziswiler <marcel.ziswiler at toradex.com>
>
> Unfortunately, this version no longer works for our
> IMX8_PCIE_REFCLK_PAD_OUTPUT use-case. Further comments in- lined
> below.
>
[Richard Zhu] Sorry to hear about that. Then, it seems that this bit should be
set anyway.
It's hard to understand this bit refer to the RM document. Sigh ☹.
Would set the AUX_EN bit later.
> > ---
> > drivers/phy/freescale/Kconfig | 9 +
> > drivers/phy/freescale/Makefile | 1 +
> > drivers/phy/freescale/phy-fsl-imx8m-pcie.c | 234
> > +++++++++++++++++++++
> > 3 files changed, 244 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..de9ee7020f76
> > 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"
>
> Above description is missing the M as in i.MX 8M.
[Richard Zhu] Okay, would be added later.
Thanks.
>
> > + depends on OF && HAS_IOMEM
> > + select GENERIC_PHY
> > + default ARCH_MXC && ARM64
> > + 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..4b4402eaddcc
> > --- /dev/null
> > +++ b/drivers/phy/freescale/phy-fsl-imx8m-pcie.c
> > @@ -0,0 +1,234 @@
> > +// 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_AUX_EN_OVERRIDE BIT(9)
> > +
> > +struct imx8_pcie_phy {
> > + void __iomem *base;
> > + struct clk *clk;
> > + struct phy *phy;
> > + struct regmap *iomuxc_gpr;
> > + struct reset_control *reset;
> > + u32 refclk_pad_mode;
> > + u32 tx_deemph_gen1;
> > + u32 tx_deemph_gen2;
> > + bool clkreq_unused; };
> > +
> > +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);
> > +
> > + pad_mode = imx8_phy->refclk_pad_mode;
> > + /* Set AUX_EN_OVERRIDE 1'b0, when the CLKREQ# isn't
> hooked */
> > + regmap_update_bits(imx8_phy->iomuxc_gpr,
> IOMUXC_GPR14,
> >
> + IMX8MM_GPR_PCIE_AUX_EN_
> OVERRIDE,
> > + imx8_phy->clkreq_unused ?
> > + 0 :
> IMX8MM_GPR_PCIE_AUX_EN_OVERRIDE);
> > + regmap_update_bits(imx8_phy->iomuxc_gpr,
> IOMUXC_GPR14,
> > + IMX8MM_GPR_PCIE_AUX_EN,
> > + pad_mode ==
> IMX8_PCIE_REFCLK_PAD_INPUT ?
> > + IMX8MM_GPR_PCIE_AUX_EN :
> 0);
>
> V3 had IMX8MM_GPR_PCIE_AUX_EN always enabled. Turns out V4
> stopped working for our output use-case as it only enables it for the input
> use-case. If I enable this one always it starts working again.
[Richard Zhu] See my comment above.
Sorry to bring the regression on your board.
>
> > + regmap_update_bits(imx8_phy->iomuxc_gpr,
> IOMUXC_GPR14,
> >
> + IMX8MM_GPR_PCIE_POWER_O
> FF, 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_S
> EL,
> > + pad_mode ==
> IMX8_PCIE_REFCLK_PAD_INPUT ?
> >
> + 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);
> > +
> > +
> > + 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(imx8_phy->tx_deemph_gen1, imx8_phy->base +
> > +PCIE_PHY_TRSV_REG5);
> > + writel(imx8_phy->tx_deemph_gen2, imx8_phy->base +
> > +PCIE_PHY_TRSV_REG6);
> > +
> > + 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_mod
> e);
> > +
> > + if (of_property_read_u32(np, "fsl,tx-deemph-gen1",
> >
> + &imx8_phy->tx_deemp
> h_gen1))
> > + imx8_phy->tx_deemph_gen1 = 0;
>
> Is zero here really a sane default or should we rather use the previous
> hard-coded value?
>
> > +
> > + if (of_property_read_u32(np, "fsl,tx-deemph-gen2",
> >
> + &imx8_phy->tx_deemp
> h_gen2))
> > + imx8_phy->tx_deemph_gen2 = 0;
>
> Ditto.
[Richard Zhu] Refer to the RM document, the default values is zero.
How about to not touch the registers if the values of the tx_deemph_gen#
are zero?
In the end, thanks a lot for your tests on this series patches.
BR
Richard
>
> > +
> > + if (of_property_read_bool(np, "fsl,clkreq-unsupported"))
> > + imx8_phy->clkreq_unused = true;
> > + else
> > + imx8_phy->clkreq_unused = false;
> > +
> > + imx8_phy->clk = devm_clk_get(dev, "ref");
> > + 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");
> > + 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");
More information about the linux-arm-kernel
mailing list