[PATCH v6 1/4] phy: cadence: Add Cadence D-PHY Rx driver
Pratyush Yadav
p.yadav at ti.com
Wed Dec 22 10:45:48 PST 2021
On 20/12/21 04:31PM, Paul Kocialkowski wrote:
> Hi Pratyush,
>
> On Tue 14 Dec 21, 23:37, Pratyush Yadav wrote:
> > The Cadence D-PHY can be configured in Tx (DSI) mode or Rx (CSI) mode.
> > Both modes have a different programming sequence and share little among
> > them. In addition, a PHY configured in Tx mode cannot be used in Rx mode
> > and vice versa. For this reason, create a separate driver for the Rx
> > mode to make it easier to read and maintain.
> >
> > Signed-off-by: Pratyush Yadav <p.yadav at ti.com>
> >
> > ---
> >
> > Changes in v6:
> > - Move to a separate driver.
> >
> > Changes in v5:
> > - Use the new cdns_dphy_info to specify PHY ops.
> > - Re-order include in alphabetical order.
> > - Make bands const.
> > - Drop num_bands.
> > - Make i, lanes unsigned.
> > - Drop the maximum check in cdns_dphy_rx_get_band_ctrl(). Let the loop
> > complete and return -EOPNOTSUPP when we reach the end.
> > - Drop the "rate < bands[i].min_rate" check since the bands are in
> > ascending order.
> > - Move data_lane_ctrl to start of function and make it static const.
> >
> > Changes in v4:
> > - Drop the submode parts. Use a different compatible for the Rx ops.
> > - Make bands and num_bands static.
> >
> > Changes in v3:
> > - Use a table to select the band.
> > - Use a table to poll the data lane ready bits.
> > - Multiply the DPHY HS clock rate by 2 to get the bit rate since the
> > clock is DDR.
> >
> > drivers/phy/cadence/Kconfig | 8 +
> > drivers/phy/cadence/Makefile | 1 +
> > drivers/phy/cadence/cdns-dphy-rx.c | 250 +++++++++++++++++++++++++++++
> > 3 files changed, 259 insertions(+)
> > create mode 100644 drivers/phy/cadence/cdns-dphy-rx.c
> >
> > diff --git a/drivers/phy/cadence/Kconfig b/drivers/phy/cadence/Kconfig
> > index a62910ff5591..1adde2d99ae7 100644
> > --- a/drivers/phy/cadence/Kconfig
> > +++ b/drivers/phy/cadence/Kconfig
> > @@ -22,6 +22,14 @@ config PHY_CADENCE_DPHY
> > system. If M is selected, the module will be called
> > cdns-dphy.
> >
> > +config PHY_CADENCE_DPHY_RX
> > + tristate "Cadence D-PHY Rx Support"
> > + depends on HAS_IOMEM && OF
> > + select GENERIC_PHY
> > + select GENERIC_PHY_MIPI_DPHY
> > + help
> > + Support for Cadence D-PHY in Rx configuration.
> > +
> > config PHY_CADENCE_SIERRA
> > tristate "Cadence Sierra PHY Driver"
> > depends on OF && HAS_IOMEM && RESET_CONTROLLER
> > diff --git a/drivers/phy/cadence/Makefile b/drivers/phy/cadence/Makefile
> > index 26e16bd34efe..e17f035ddece 100644
> > --- a/drivers/phy/cadence/Makefile
> > +++ b/drivers/phy/cadence/Makefile
> > @@ -1,5 +1,6 @@
> > # SPDX-License-Identifier: GPL-2.0-only
> > obj-$(CONFIG_PHY_CADENCE_TORRENT) += phy-cadence-torrent.o
> > obj-$(CONFIG_PHY_CADENCE_DPHY) += cdns-dphy.o
> > +obj-$(CONFIG_PHY_CADENCE_DPHY_RX) += cdns-dphy-rx.o
> > obj-$(CONFIG_PHY_CADENCE_SIERRA) += phy-cadence-sierra.o
> > obj-$(CONFIG_PHY_CADENCE_SALVO) += phy-cadence-salvo.o
> > diff --git a/drivers/phy/cadence/cdns-dphy-rx.c b/drivers/phy/cadence/cdns-dphy-rx.c
> > new file mode 100644
> > index 000000000000..fb75e645e662
> > --- /dev/null
> > +++ b/drivers/phy/cadence/cdns-dphy-rx.c
> > @@ -0,0 +1,250 @@
> > +// SPDX-License-Identifier: GPL-2.0+
> > +/*
> > + * Copyright (C) 2021 Texas Instruments Incorporated - https://www.ti.com/
> > + */
> > +
> > +#include <linux/bitfield.h>
> > +#include <linux/bitops.h>
> > +#include <linux/io.h>
> > +#include <linux/iopoll.h>
> > +#include <linux/module.h>
> > +#include <linux/phy/phy.h>
> > +#include <linux/phy/phy-mipi-dphy.h>
> > +#include <linux/platform_device.h>
> > +
> > +#define DPHY_PMA_CMN(reg) (reg)
> > +#define DPHY_PCS(reg) (0xb00 + (reg))
> > +#define DPHY_ISO(reg) (0xc00 + (reg))
> > +
> > +#define DPHY_CMN_SSM DPHY_PMA_CMN(0x20)
> > +#define DPHY_CMN_RX_MODE_EN BIT(10)
> > +#define DPHY_CMN_RX_BANDGAP_TIMER_MASK GENMASK(8, 1)
> > +#define DPHY_CMN_SSM_EN BIT(0)
> > +
> > +#define DPHY_CMN_RX_BANDGAP_TIMER 0x14
> > +
> > +#define DPHY_BAND_CFG DPHY_PCS(0x0)
> > +#define DPHY_BAND_CFG_RIGHT_BAND GENMASK(9, 5)
> > +#define DPHY_BAND_CFG_LEFT_BAND GENMASK(4, 0)
> > +
> > +#define DPHY_POWER_ISLAND_EN_DATA DPHY_PCS(0x8)
> > +#define DPHY_POWER_ISLAND_EN_DATA_VAL 0xaaaaaaaa
> > +
> > +#define DPHY_POWER_ISLAND_EN_CLK DPHY_PCS(0xc)
> > +#define DPHY_POWER_ISLAND_EN_CLK_VAL 0xaa
> > +
> > +#define DPHY_ISO_CL_CTRL_L DPHY_ISO(0x10)
> > +#define DPHY_ISO_DL_CTRL_L0 DPHY_ISO(0x14)
> > +#define DPHY_ISO_DL_CTRL_L1 DPHY_ISO(0x20)
> > +#define DPHY_ISO_DL_CTRL_L2 DPHY_ISO(0x30)
> > +#define DPHY_ISO_DL_CTRL_L3 DPHY_ISO(0x3c)
> > +
> > +#define DPHY_ISO_LANE_READY_BIT 0
> > +#define DPHY_ISO_LANE_READY_TIMEOUT_MS 100UL
> > +
> > +#define DPHY_LANES_MIN 1
> > +#define DPHY_LANES_MAX 4
> > +
> > +struct cdns_dphy_rx {
> > + void __iomem *regs;
> > + struct device *dev;
> > + struct phy *phy;
> > +};
> > +
> > +struct cdns_dphy_rx_band {
> > + /* Rates are in Mbps. */
> > + unsigned int min_rate;
> > + unsigned int max_rate;
> > +};
> > +
> > +/* Order of bands is important since the index is the band number. */
> > +static const struct cdns_dphy_rx_band bands[] = {
> > + {80, 100}, {100, 120}, {120, 160}, {160, 200}, {200, 240},
> > + {240, 280}, {280, 320}, {320, 360}, {360, 400}, {400, 480},
> > + {480, 560}, {560, 640}, {640, 720}, {720, 800}, {800, 880},
> > + {880, 1040}, {1040, 1200}, {1200, 1350}, {1350, 1500}, {1500, 1750},
> > + {1750, 2000}, {2000, 2250}, {2250, 2500}
>
> Cosmetic suggestion: add whitespaces after { and before }.
I think it would just add noise. I would like to keep it compact.
>
> > +};
> > +
> > +static int cdns_dphy_rx_power_on(struct phy *phy)
> > +{
> > + struct cdns_dphy_rx *dphy = phy_get_drvdata(phy);
> > +
> > + /* Start RX state machine. */
> > + writel(DPHY_CMN_SSM_EN | DPHY_CMN_RX_MODE_EN |
> > + FIELD_PREP(DPHY_CMN_RX_BANDGAP_TIMER_MASK,
> > + DPHY_CMN_RX_BANDGAP_TIMER),
> > + dphy->regs + DPHY_CMN_SSM);
> > +
> > + return 0;
> > +}
> > +
> > +static int cdns_dphy_rx_power_off(struct phy *phy)
> > +{
> > + struct cdns_dphy_rx *dphy = phy_get_drvdata(phy);
> > +
> > + writel(0, dphy->regs + DPHY_CMN_SSM);
> > +
> > + return 0;
> > +}
> > +
> > +static int cdns_dphy_rx_get_band_ctrl(unsigned long hs_clk_rate)
> > +{
> > + unsigned int rate, i;
> > +
> > + rate = hs_clk_rate / 1000000UL;
> > + /* Since CSI-2 clock is DDR, the bit rate is twice the clock rate. */
> > + rate *= 2;
> > +
> > + if (rate < bands[0].min_rate)
> > + return -EOPNOTSUPP;
> > +
> > + for (i = 0; i < ARRAY_SIZE(bands); i++) {
> > + if (rate < bands[i].max_rate)
> > + return i;
> > + }
>
> Comstic: you don't need the wrapping { and } here.
Yes, but I generally like to add it for cases like this to make the
scope clearer. But I have no strong preferences, I can drop them.
>
> > +
> > + return -EOPNOTSUPP;
> > +}
> > +
> > +static int cdns_dphy_rx_wait_for_bit(void __iomem *addr, unsigned int bit)
>
> You can probably inline this one.
The compiler will likely do it anyway, but sure I will add it.
>
> > +{
> > + u32 val;
> > +
> > + return readl_relaxed_poll_timeout(addr, val, val & BIT(bit), 10,
> > + DPHY_ISO_LANE_READY_TIMEOUT_MS * 1000);
> > +}
> > +
> > +static int cdns_dphy_rx_wait_lane_ready(struct cdns_dphy_rx *dphy,
> > + unsigned int lanes)
> > +{
> > + static const u32 data_lane_ctrl[] = {DPHY_ISO_DL_CTRL_L0,
> > + DPHY_ISO_DL_CTRL_L1,
> > + DPHY_ISO_DL_CTRL_L2,
> > + DPHY_ISO_DL_CTRL_L3};
> > + void __iomem *reg = dphy->regs;
> > + unsigned int i;
> > + int ret;
> > +
> > + /* Data lanes. Minimum one lane is mandatory. */
> > + if (lanes < DPHY_LANES_MIN || lanes > DPHY_LANES_MAX)
> > + return -EINVAL;
> > +
> > + /* Clock lane */
> > + ret = cdns_dphy_rx_wait_for_bit(reg + DPHY_ISO_CL_CTRL_L,
> > + DPHY_ISO_LANE_READY_BIT);
> > + if (ret)
> > + return ret;
> > +
> > + for (i = 0; i < lanes; i++) {
> > + ret = cdns_dphy_rx_wait_for_bit(reg + data_lane_ctrl[i],
> > + DPHY_ISO_LANE_READY_BIT);
> > + if (ret)
> > + return ret;
> > + }
> > +
> > + return 0;
> > +}
> > +
> > +static int cdns_dphy_rx_configure(struct phy *phy,
> > + union phy_configure_opts *opts)
> > +{
> > + struct cdns_dphy_rx *dphy = phy_get_drvdata(phy);
> > + unsigned int reg;
> > + int band_ctrl, ret;
> > +
> > + band_ctrl = cdns_dphy_rx_get_band_ctrl(opts->mipi_dphy.hs_clk_rate);
> > + if (band_ctrl < 0)
> > + return band_ctrl;
> > +
> > + reg = FIELD_PREP(DPHY_BAND_CFG_LEFT_BAND, band_ctrl) |
> > + FIELD_PREP(DPHY_BAND_CFG_RIGHT_BAND, band_ctrl);
> > + writel(reg, dphy->regs + DPHY_BAND_CFG);
> > +
> > + /*
> > + * Set the required power island phase 2 time. This is mandated by DPHY
> > + * specs.
> > + */
> > + reg = DPHY_POWER_ISLAND_EN_DATA_VAL;
> > + writel(reg, dphy->regs + DPHY_POWER_ISLAND_EN_DATA);
> > + reg = DPHY_POWER_ISLAND_EN_CLK_VAL;
> > + writel(reg, dphy->regs + DPHY_POWER_ISLAND_EN_CLK);
> > +
> > + ret = cdns_dphy_rx_wait_lane_ready(dphy, opts->mipi_dphy.lanes);
> > + if (ret) {
> > + dev_err(dphy->dev, "DPHY wait for lane ready timeout\n");
> > + return ret;
> > + }
> > +
> > + return 0;
> > +}
> > +
> > +static int cdns_dphy_rx_validate(struct phy *phy, enum phy_mode mode,
> > + int submode, union phy_configure_opts *opts)
> > +{
> > + int ret;
> > +
> > + if (mode != PHY_MODE_MIPI_DPHY)
> > + return -EINVAL;
> > +
> > + ret = cdns_dphy_rx_get_band_ctrl(opts->mipi_dphy.hs_clk_rate);
> > + if (ret < 0)
> > + return ret;
> > +
> > + return phy_mipi_dphy_config_validate(&opts->mipi_dphy);
> > +}
> > +
> > +static const struct phy_ops cdns_dphy_rx_ops = {
> > + .power_on = cdns_dphy_rx_power_on,
> > + .power_off = cdns_dphy_rx_power_off,
> > + .configure = cdns_dphy_rx_configure,
> > + .validate = cdns_dphy_rx_validate,
> > +};
> > +
> > +static int cdns_dphy_rx_probe(struct platform_device *pdev)
> > +{
> > + struct device *dev = &pdev->dev;
> > + struct phy_provider *provider;
> > + struct cdns_dphy_rx *dphy;
> > +
> > + dphy = devm_kzalloc(dev, sizeof(*dphy), GFP_KERNEL);
> > + if (!dphy)
> > + return -ENOMEM;
> > +
> > + dev_set_drvdata(dev, dphy);
>
> It looks like you're never getting the dphy pointer from dev so this seems a bit
> pointless.
Yes, we don't need it since there is no remove hook as of now. But I
would still like to keep it. It would avoid extra churn when someone
does want to add one.
>
> > + dphy->dev = dev;
> > +
> > + dphy->regs = devm_platform_ioremap_resource(pdev, 0);
> > + if (IS_ERR(dphy->regs))
> > + return PTR_ERR(dphy->regs);
> > +
> > + dphy->phy = devm_phy_create(dev, NULL, &cdns_dphy_rx_ops);
> > + if (IS_ERR(dphy->phy)) {
> > + dev_err(dev, "Failed to create PHY: %d\n", PTR_ERR(dphy->phy));
> > + return PTR_ERR(dphy->phy);
> > + }
> > +
> > + phy_set_drvdata(dphy->phy, dphy);
> > + provider = devm_of_phy_provider_register(dev, of_phy_simple_xlate);
>
> You might want to print an error message if it failed to register.
Ok, will do.
>
> > +
> > + return PTR_ERR_OR_ZERO(provider);
> > +}
> > +
> > +static const struct of_device_id cdns_dphy_rx_of_match[] = {
> > + { .compatible = "cdns,dphy-rx" },
> > + { /* sentinel */ },
> > +};
> > +MODULE_DEVICE_TABLE(of, cdns_dphy_rx_of_match);
> > +
> > +static struct platform_driver cdns_dphy_rx_platform_driver = {
> > + .probe = cdns_dphy_rx_probe,
> > + .driver = {
> > + .name = "cdns-mipi-dphy-rx",
> > + .of_match_table = cdns_dphy_rx_of_match,
> > + },
> > +};
> > +module_platform_driver(cdns_dphy_rx_platform_driver);
> > +
> > +MODULE_AUTHOR("Pratyush Yadav <p.yadav at ti.com>");
> > +MODULE_DESCRIPTION("Cadence D-PHY Rx Driver");
> > +MODULE_LICENSE("GPL v2");
>
> Otherwise this looks good to me!
Thanks for reviewing.
>
> Cheers,
>
> Paul
>
> --
> Paul Kocialkowski, Bootlin
> Embedded Linux and kernel engineering
> https://bootlin.com
--
Regards,
Pratyush Yadav
Texas Instruments Inc.
More information about the linux-phy
mailing list