Re: [PATCH v2 2/2] usb: typec: tcpci_maxim: configure charging & data paths

From: Badhri Jagan Sridharan
Date: Fri Mar 12 2021 - 02:06:24 EST


On Thu, Mar 11, 2021 at 10:39 PM Guenter Roeck <linux@xxxxxxxxxxxx> wrote:
>
> On 3/11/21 9:24 PM, Badhri Jagan Sridharan wrote:
> > The change exposes the data_role and the orientation as a extcon
> > interface for configuring the USB data controller.
> >
> > Signed-off-by: Badhri Jagan Sridharan <badhri@xxxxxxxxxx>
> > ---
> > Changes since V1:
> > - Dropped changes related to get_/set_current_limit and pd_capable
> > callback. Will send them in as separate patches.
> > ---
> > drivers/usb/typec/tcpm/tcpci_maxim.c | 56 ++++++++++++++++++++++++++++
> > 1 file changed, 56 insertions(+)
> >
> > diff --git a/drivers/usb/typec/tcpm/tcpci_maxim.c b/drivers/usb/typec/tcpm/tcpci_maxim.c
> > index 041a1c393594..1210445713ee 100644
> > --- a/drivers/usb/typec/tcpm/tcpci_maxim.c
> > +++ b/drivers/usb/typec/tcpm/tcpci_maxim.c
> > @@ -7,6 +7,8 @@
> >
> > #include <linux/interrupt.h>
> > #include <linux/i2c.h>
> > +#include <linux/extcon.h>
> > +#include <linux/extcon-provider.h>
> > #include <linux/kernel.h>
> > #include <linux/module.h>
> > #include <linux/regmap.h>
> > @@ -46,6 +48,8 @@ struct max_tcpci_chip {
> > struct device *dev;
> > struct i2c_client *client;
> > struct tcpm_port *port;
> > + bool attached;
> > + struct extcon_dev *extcon;
> > };
> >
> > static const struct regmap_range max_tcpci_tcpci_range[] = {
> > @@ -439,6 +443,39 @@ static int tcpci_init(struct tcpci *tcpci, struct tcpci_data *data)
> > return -1;
> > }
> >
> > +static void max_tcpci_set_roles(struct tcpci *tcpci, struct tcpci_data *data, bool attached,
> > + enum typec_role role, enum typec_data_role data_role)
> > +{
> > + struct max_tcpci_chip *chip = tdata_to_max_tcpci(data);
> > +
> > + chip->attached = attached;
> > +
> > + if (!attached) {
> > + extcon_set_state_sync(chip->extcon, EXTCON_USB_HOST, 0);
> > + extcon_set_state_sync(chip->extcon, EXTCON_USB, 0);
> > + return;
> > + }
> > +
> > + extcon_set_state_sync(chip->extcon, data_role == TYPEC_HOST ? EXTCON_USB_HOST : EXTCON_USB,
> > + 1);
> > +}
> > +
> > +static void max_tcpci_set_cc_polarity(struct tcpci *tcpci, struct tcpci_data *data,
> > + enum typec_cc_polarity polarity)
> > +{
> > + struct max_tcpci_chip *chip = tdata_to_max_tcpci(data);
> > +
> > + extcon_set_property(chip->extcon, EXTCON_USB, EXTCON_PROP_USB_TYPEC_POLARITY,
> > + (union extcon_property_value)(int)polarity);
> > + extcon_set_property(chip->extcon, EXTCON_USB_HOST, EXTCON_PROP_USB_TYPEC_POLARITY,
> > + (union extcon_property_value)(int)polarity);
> > +}
> > +
> > +static const unsigned int usbpd_extcon[] = {
> > + EXTCON_USB,
> > + EXTCON_USB_HOST,
> > +};
> > +
> > static int max_tcpci_probe(struct i2c_client *client, const struct i2c_device_id *i2c_id)
> > {
> > int ret;
> > @@ -472,6 +509,8 @@ static int max_tcpci_probe(struct i2c_client *client, const struct i2c_device_id
> > chip->data.auto_discharge_disconnect = true;
> > chip->data.vbus_vsafe0v = true;
> > chip->data.set_partner_usb_comm_capable = max_tcpci_set_partner_usb_comm_capable;
> > + chip->data.set_roles = max_tcpci_set_roles;
> > + chip->data.set_cc_polarity = max_tcpci_set_cc_polarity;
> >
> > max_tcpci_init_regs(chip);
> > chip->tcpci = tcpci_register_port(chip->dev, &chip->data);
> > @@ -484,6 +523,23 @@ static int max_tcpci_probe(struct i2c_client *client, const struct i2c_device_id
> > if (ret < 0)
> > goto unreg_port;
> >
> > + chip->extcon = devm_extcon_dev_allocate(&client->dev, usbpd_extcon);
> > + if (IS_ERR(chip->extcon)) {
> > + dev_err(&client->dev, "Error allocating extcon: %ld\n", PTR_ERR(chip->extcon));
> > + ret = PTR_ERR(chip->extcon);
> > + goto unreg_port;
> > + }
> > +
> > + ret = devm_extcon_dev_register(&client->dev, chip->extcon);
> > + if (ret < 0) {
> > + dev_err(&client->dev, "failed to register extcon device");
> > + goto unreg_port;
> > + }
>
> Effectively this mandates extcon support to be able to use this driver/chip.
> Does that make sense ? If this is indeed mandatory, how did it work so far ?
Hi Guenter,

We had this in our downstream branch but didnt get a chance to send it
to linux upstream.
I should wrap it in "if(IS_ENABLED(CONFIG_EXTCON))", the tcpc can work
without the
extcon.

>
> Also, what makes this code chip specific ?

Extcon here as is not chip code specific, but, the driver which
subscribes to the extcon interface is chip specific.
I hope it's ok to still send this.

Thanks,
Badhri


>
> Thanks,
> Guenter
>
> > +
> > + extcon_set_property_capability(chip->extcon, EXTCON_USB, EXTCON_PROP_USB_TYPEC_POLARITY);
> > + extcon_set_property_capability(chip->extcon, EXTCON_USB_HOST,
> > + EXTCON_PROP_USB_TYPEC_POLARITY);
> > +
> > device_init_wakeup(chip->dev, true);
> > return 0;
> >
> >
>