Re: FW: [PATCH 2/2] usb: typec: tcpci_rt1718s: Add Richtek RT1718S tcpci driver

[Gene Chen]

> -----Original Message-----
> From: Guenter Roeck [mailto:groeck7@xxxxxxxxx] On Behalf Of Guenter Roeck
> Sent: Tuesday, January 10, 2023 2:52 AM
> To: gene_chen(陳俊宇) <gene_chen@xxxxxxxxxxx>; robh+dt@xxxxxxxxxx; krzysztof.kozlowski+dt@xxxxxxxxxx; heikki.krogerus@xxxxxxxxxxxxxxx
> Cc: gregkh@xxxxxxxxxxxxxxxxxxx; linux-usb@xxxxxxxxxxxxxxx; devicetree@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx
> Subject: Re: [PATCH 2/2] usb: typec: tcpci_rt1718s: Add Richtek RT1718S tcpci driver
>
> On 1/9/23 01:31, gene_chen@xxxxxxxxxxx wrote:
> > From: Gene Chen <gene_chen@xxxxxxxxxxx>
> >
> > Richtek RT1718S is highly integrated TCPC and Power Delivery (PD)
> > controller with IEC-ESD Protection on SBU/CC/DP/DM, USB2.0 Switch,
> > Charging Port Controller and Power-Path Control.
> >
> > Signed-off-by: Gene Chen <gene_chen@xxxxxxxxxxx>
> > ---
> >   drivers/usb/typec/tcpm/Kconfig         |   9 +
> >   drivers/usb/typec/tcpm/Makefile        |   1 +
> >   drivers/usb/typec/tcpm/tcpci_rt1718s.c | 349 +++++++++++++++++++++++++++++++++
> >   3 files changed, 359 insertions(+)
> >   create mode 100644 drivers/usb/typec/tcpm/tcpci_rt1718s.c
> >
> > diff --git a/drivers/usb/typec/tcpm/Kconfig
> > b/drivers/usb/typec/tcpm/Kconfig index e6b88ca..f0efb34 100644
> > --- a/drivers/usb/typec/tcpm/Kconfig
> > +++ b/drivers/usb/typec/tcpm/Kconfig
> > @@ -27,6 +27,15 @@ config TYPEC_RT1711H
> >     Type-C Port Controller Manager to provide USB PD and USB
> >     Type-C functionalities.
> >
> > +config TYPEC_RT1718S
> > +tristate "Richtek RT1718S Type-C chip driver"
> > +depends on I2C
> > +help
> > +  Richtek RT1718S Type-C chip driver that works with
> > +  Type-C Port Controller Manager to provide USB PD and USB
> > +  Type-C functionalities.
> > +  Additionally, it supports BC1.2 and power-path control.
> > +
> >   config TYPEC_MT6360
> >   tristate "Mediatek MT6360 Type-C driver"
> >   depends on MFD_MT6360
> > diff --git a/drivers/usb/typec/tcpm/Makefile
> > b/drivers/usb/typec/tcpm/Makefile index 906d9dc..db33ffc 100644
> > --- a/drivers/usb/typec/tcpm/Makefile
> > +++ b/drivers/usb/typec/tcpm/Makefile
> > @@ -5,6 +5,7 @@ obj-$(CONFIG_TYPEC_WCOVE)+= typec_wcove.o
> >   typec_wcove-y:= wcove.o
> >   obj-$(CONFIG_TYPEC_TCPCI)+= tcpci.o
> >   obj-$(CONFIG_TYPEC_RT1711H)+= tcpci_rt1711h.o
> > +obj-$(CONFIG_TYPEC_RT1718S)+= tcpci_rt1718s.o
> >   obj-$(CONFIG_TYPEC_MT6360)+= tcpci_mt6360.o
> >   obj-$(CONFIG_TYPEC_TCPCI_MT6370)+= tcpci_mt6370.o
> >   obj-$(CONFIG_TYPEC_TCPCI_MAXIM)+= tcpci_maxim.o
> > diff --git a/drivers/usb/typec/tcpm/tcpci_rt1718s.c
> > b/drivers/usb/typec/tcpm/tcpci_rt1718s.c
> > new file mode 100644
> > index 00000000..305b39c
> > --- /dev/null
> > +++ b/drivers/usb/typec/tcpm/tcpci_rt1718s.c
> > +
> > +if (*(u8 *)reg == RT1718S_P1PREFIX) {
> > +xfer[0].len = 1,
> > +xfer[0].buf = (u8 *)(reg + 1);
>
> Pointer arithmetic on void * (see below).
>

yes, should be xfer[0].buf = (u8 *)reg + 1;

> > +}
>
> There is a lot of context here which needs explanation. The code extracts the upper 8 bit of the register address and drops it if the value is 0.
> So the address is either 8 bit or 16 bit depending on the register address.
> Also, this only works because reg_format_endian is set to big endian.
>
> This really needs to be documented.
>
> Assigning the values to len and buf twice is really bad style.
> Please use if/else.
>

Because of rt1718s address which is more than 1 page(0x00-0xFF), hw
design i2c pattern with register has 2 byte to access another
page(0xF200-0xF2FF).
The upper byte is set to 0xF2 indicated page 2.

> > +
> > +if (*(u8 *)val == RT1718S_P1PREFIX) {
> > +xfer.len = val_size - 1;
> > +xfer.buf = (u8 *)(val + 1);
> > +}
> > +
>
> Same comments as above. Also, typecast seems wrong. Shouldn't it be
> ((u8 *)val) + 1) ? My memory may defeat me, but I think that pointer arithmetic on void * is undefined or even illegal.
>
>

You are right, I will fix it, thanks.

> > +if (ret < 0)
> > +return ret;
> > +if (ret != 1)
> > +return -EIO;
> > +
> > +return 0;
> > +}
> > +
> > +static const struct regmap_bus rt1718s_regmap_bus = {
> > +.read= rt1718s_regmap_read,
> > +.write= rt1718s_regmap_write,
> > +};
> > +
>
> This needs documentation: Why not use standard regmap functions ?
> Yes, I know, it is because of the ubusual addressing format used by the chip, but it still needs to be explained. Not everyone should have to read the datasheet to understand the code.
>

    Should I add comment before declare rt1718s_regmap_bus?

/*
 * Because of rt1718s address which is more than 1 page(0x00-0xFF),
 * hw design i2c pattern with register has 2 byte to access another
page(0xF200-0xF2FF).
 * The upper byte is set to 0xF2 indicated page 2.
 */

> > +static int rt1718s_sw_reset(struct rt1718s_chip *chip) {
> > +int ret;
> > +
> > +ret = regmap_update_bits(chip->tdata.regmap, RT1718S_SYS_CTRL3,
> > + RT1718S_SWRESET_MASK, RT1718S_SWRESET_MASK);
> > +if (ret < 0)
> > +return ret;
> > +
> > +/* Wait for IC to reset done*/
> > +usleep_range(1000, 2000);
> > +
>
> "RT1718S will not response to the I2C commands for 2ms after writing SOFT_RESET"
>
> Timeout needs to be at least 2 ms.
>

ACK

> > +return 0;
> > +}
> > +
> > +static int rt1718s_check_chip_exist(struct i2c_client *i2c) {
> > +int ret;
> > +
> > +ret = i2c_smbus_read_word_data(i2c, TCPC_VENDOR_ID);
> > +if (ret < 0)
> > +return ret;
> > +if (ret != RT1718S_VID) {
> > +dev_err(&i2c->dev, "vid is not correct, 0x%04x\n", ret);
> > +return -ENODEV;
> > +}
> > +ret = i2c_smbus_read_word_data(i2c, TCPC_PRODUCT_ID);
> > +if (ret < 0)
> > +return ret;
> > +if (ret != RT1718S_PID) {
> > +dev_err(&i2c->dev, "pid is not correct, 0x%04x\n", ret);
> > +return -ENODEV;
> > +}
> > +return 0;
> > +}
> > +
> > +static int rt1718s_probe(struct i2c_client *i2c) {
> > +struct rt1718s_chip *chip;
> > +struct gpio_desc *gpiod;
> > +int ret;
> > +u16 clr_events = 0;
> > +
> > +ret = rt1718s_check_chip_exist(i2c);
> > +if (ret < 0) {
> > +dev_err(&i2c->dev, "check vid/pid fail(%d)\n", ret);
>
> Double error message.
>

ACK, I will remove it.

> > +return ret;
> > +}
> > +
> > +chip = devm_kzalloc(&i2c->dev, sizeof(*chip), GFP_KERNEL);
> > +if (!chip)
> > +return -ENOMEM;
> > +
> > +chip->dev = &i2c->dev;
> > +
> > +chip->tdata.regmap = devm_regmap_init(&i2c->dev,
> > +      &rt1718s_regmap_bus, &i2c->dev,
> > +      &rt1718s_regmap_config);
> > +if (IS_ERR(chip->tdata.regmap))
> > +return dev_err_probe(&i2c->dev, PTR_ERR(chip->tdata.regmap),
> > +     "Failed to init regmap\n");
> > +
> > +chip->vbus = devm_regulator_get(&i2c->dev, "vbus");
> > +if (IS_ERR(chip->vbus))
> > +return dev_err_probe(&i2c->dev, PTR_ERR(chip->vbus),
> > +     "Failed to get vbus regulator\n");
> > +
> > +ret = rt1718s_sw_reset(chip);
> > +if (ret < 0)
> > +return ret;
> > +
> > +ret = regmap_raw_write(chip->tdata.regmap, TCPC_ALERT_MASK, &clr_events,
> > +       sizeof(clr_events));
> > +
> > +chip->tdata.init = rt1718s_init;
> > +chip->tdata.set_vbus = rt1718s_set_vbus;
> > +chip->tcpci = tcpci_register_port(&i2c->dev, &chip->tdata);
> > +if (IS_ERR(chip->tcpci))
> > +return dev_err_probe(&i2c->dev, PTR_ERR(chip->tcpci),
> > +     "Failed to register tcpci port\n");
> > +
> > +/* for platform set gpio default inpull_high */
> > +gpiod = devm_gpiod_get(&i2c->dev, NULL, GPIOD_IN);
> > +if (IS_ERR(gpiod))
> > +return dev_err_probe(&i2c->dev, PTR_ERR(gpiod),
> > +     "Failed to get gpio\n");
> > +
> > +ret = devm_request_threaded_irq(&i2c->dev, i2c->irq, NULL,
> > +rt1718s_irq, IRQF_ONESHOT,
> > +dev_name(&i2c->dev), chip);
> > +if (ret) {
> > +dev_err(chip->dev, "Failed to register irq\n");
> > +tcpci_unregister_port(chip->tcpci);
> > +return ret;
> > +}
> > +
> > +device_init_wakeup(&i2c->dev, true);
> > +i2c_set_clientdata(i2c, chip);
> > +
> > +dev_info(&i2c->dev, "%s:successfully\n", __func__);
>
> Nore logging noise.
>

ACK

> > +return 0;
> > +}
> > +
> > +static void rt1718s_remove(struct i2c_client *i2c) {
> > +struct rt1718s_chip *chip = i2c_get_clientdata(i2c);
> > +
> > +tcpci_unregister_port(chip->tcpci);
> > +}
> > +
> > +static int __maybe_unused rt1718s_suspend(struct device *dev) {
> > +struct i2c_client *i2c = to_i2c_client(dev);
> > +
> > +if (device_may_wakeup(dev))
> > +enable_irq_wake(i2c->irq);
> > +disable_irq(i2c->irq);
> > +
> > +return 0;
> > +}
> > +
> > +static int __maybe_unused rt1718s_resume(struct device *dev) {
> > +struct i2c_client *i2c = to_i2c_client(dev);
> > +
> > +if (device_may_wakeup(dev))
> > +disable_irq_wake(i2c->irq);
> > +enable_irq(i2c->irq);
> > +
> > +return 0;
> > +}
> > +
> > +static SIMPLE_DEV_PM_OPS(rt1718s_pm_ops, rt1718s_suspend,
> > +rt1718s_resume);
> > +
> > +static const struct of_device_id __maybe_unused rt1718s_of_id[] = {
> > +{ .compatible = "richtek,rt1718s", },
> > +{},
> > +};
> > +MODULE_DEVICE_TABLE(of, rt1718s_of_id);
> > +
> > +static struct i2c_driver rt1718s_i2c_driver = {
> > +.driver = {
> > +.name = "rt1718s",
> > +.pm = &rt1718s_pm_ops,
> > +.of_match_table = rt1718s_of_id,
> > +},
> > +.probe_new = rt1718s_probe,
> > +.remove = rt1718s_remove,
> > +};
> > +module_i2c_driver(rt1718s_i2c_driver);
> > +
> > +MODULE_AUTHOR("Gene Chen <gene_chen@xxxxxxxxxxx>");
> > +MODULE_DESCRIPTION("RT1718S USB Type-C Port Controller Interface
> > +Driver"); MODULE_LICENSE("GPL");