Re: [PATCH 08/10] usb: typec: Add support for usb role switch in rt1711h driver

[Heikki Krogerus]

On Sat, Oct 27, 2018 at 05:58:18PM +0800, Yu Chen wrote:
> This patch add struct device_connection to rt1711h driver to support
> find "usb-role-switch" in tcpm_register_port.
> 
> Cc: Heikki Krogerus <heikki.krogerus@xxxxxxxxxxxxxxx>
> Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> Cc: Lucas Tsai <lucas_tsai@xxxxxxxxxxx>
> Cc: shufan_lee <shufan_lee@xxxxxxxxxxx>
> Cc: John Stultz <john.stultz@xxxxxxxxxx>
> Cc: Binghui Wang <wangbinghui@xxxxxxxxxxxxx>
> Signed-off-by: Yu Chen <chenyu56@xxxxxxxxxx>
> ---
>  drivers/usb/typec/tcpm/tcpci_rt1711h.c | 15 +++++++++++++++
>  1 file changed, 15 insertions(+)
> 
> diff --git a/drivers/usb/typec/tcpm/tcpci_rt1711h.c b/drivers/usb/typec/tcpm/tcpci_rt1711h.c
> index 017389021b96..5f8b952612fe 100644
> --- a/drivers/usb/typec/tcpm/tcpci_rt1711h.c
> +++ b/drivers/usb/typec/tcpm/tcpci_rt1711h.c
> @@ -39,6 +39,7 @@ struct rt1711h_chip {
>  	struct tcpci_data data;
>  	struct tcpci *tcpci;
>  	struct device *dev;
> +	struct device_connection dev_conn;
>  };
>  
>  static int rt1711h_read16(struct rt1711h_chip *chip, unsigned int reg, u16 *val)
> @@ -237,6 +238,7 @@ static int rt1711h_probe(struct i2c_client *client,
>  {
>  	int ret;
>  	struct rt1711h_chip *chip;
> +	const char *dev_conn_end;
>  
>  	ret = rt1711h_check_revision(client);
>  	if (ret < 0) {
> @@ -254,6 +256,7 @@ static int rt1711h_probe(struct i2c_client *client,
>  		return PTR_ERR(chip->data.regmap);
>  
>  	chip->dev = &client->dev;
> +	dev_set_name(chip->dev, "rt1711h");
>  	i2c_set_clientdata(client, chip);
>  
>  	ret = rt1711h_sw_reset(chip);
> @@ -264,6 +267,15 @@ static int rt1711h_probe(struct i2c_client *client,
>  	if (ret < 0)
>  		return ret;
>  
> +	ret = device_property_read_string(chip->dev, "dev-conn-end",
> +			&dev_conn_end);
> +	if (!ret) {
> +		chip->dev_conn.endpoint[0] = "rt1711h";
> +		chip->dev_conn.endpoint[1] = dev_conn_end;
> +		chip->dev_conn.id = "usb-role-switch";
> +		device_connection_add(&chip->dev_conn);
> +	}

That is wrong. You should not read name of a device from a property.

Regardless of that, I don't think this patch is needed at all. You
need to describe this kind connection with OF device graph
(Documentation/devicetree/bindings/graph.txt).

I've prepared a proposal how we should be able handle graphs in the
device connection API:
https://lkml.org/lkml/2018/10/24/613


thanks,

-- 
heikki