[PATCH v4] extcon: Palmas Extcon Driver

[Laxman Dewangan]

On Monday 06 May 2013 06:47 PM, Kishon Vijay Abraham I wrote:
> +
> +static irqreturn_t palmas_vbus_wakeup_irq(int irq, void *_palmas_usb)

Can we name the function to palams_vbus_irq_handler() for better 
understanding? Reserve the wakeup word for the suspend-wakeups.


> +
> +
> +static irqreturn_t palmas_id_wakeup_irq(int irq, void *_palmas_usb)

Same here for better name.


> +
> +static void palmas_set_vbus_work(struct work_struct *data)
> +{
> +       int ret;
> +       struct palmas_usb *palmas_usb = container_of(data, struct palmas_usb,
> +                                                               set_vbus_work);
> +
> +       if (IS_ERR_OR_NULL(palmas_usb->vbus_reg)) {
> +               dev_err(palmas_usb->dev, "invalid regulator\n");
> +               return;
> +       }

This error will keep coming if the vbus is not require as workqueue get 
scheduled always. I think we should remove it.


> +

> +static void palmas_dt_to_pdata(struct device_node *node,
> +               struct palmas_usb_platform_data *pdata)
> +{
> +       pdata->no_control_vbus = of_property_read_bool(node,
> +                                       "ti,no_control_vbus");


Can we change the variable names to enable_control_bus and logic 
accordingly as it looks more appropriate and easy to understand?


> +
> +       palmas_usb->irq1 = regmap_irq_get_virq(palmas->irq_data,
> +                                               PALMAS_ID_OTG_IRQ);
> +       palmas_usb->irq2 = regmap_irq_get_virq(palmas->irq_data,
> +                                               PALMAS_ID_IRQ);
> +       palmas_usb->irq3 = regmap_irq_get_virq(palmas->irq_data,
> +                                               PALMAS_VBUS_OTG_IRQ);
> +       palmas_usb->irq4 = regmap_irq_get_virq(palmas->irq_data,
> +                                               PALMAS_VBUS_IRQ);
> +

Better to name irq1, irq2 in more logical names for easy understanding.


> +
> +       if (device_create_file(&pdev->dev, &dev_attr_vbus))
> +               dev_warn(&pdev->dev, "could not create sysfs file\n");
> +
> +       palmas_usb->edev.name = "palmas_usb";
> +       palmas_usb->edev.supported_cable = palmas_extcon_cable;
> +       palmas_usb->edev.mutually_exclusive = mutually_exclusive;
> +
> +       ret = extcon_dev_register(&palmas_usb->edev, palmas_usb->dev);
> +       if (ret) {
> +               dev_err(&pdev->dev, "failed to register extcon device\n");
> +               return ret;

It need to destroy sysfs also.

> +       }
> +
> +       /* init spinlock for workqueue */
> +       spin_lock_init(&palmas_usb->lock);

It is already done above.

> +
> +       INIT_WORK(&palmas_usb->set_vbus_work, palmas_set_vbus_work);

Better to create the workqueu when control_vbus is require.


> +
> +
> diff --git a/include/linux/extcon/extcon_palmas.h b/include/linux/extcon/extcon_palmas.h
> new file mode 100644
> index 0000000..a5119c9
> --- /dev/null
> +++ b/include/linux/extcon/extcon_palmas.h

I think it can be use palama.h only. No need to have one more header for 
this.

> @@ -0,0 +1,26 @@
>
>
> -       u8 linkstat;
> +       int mailboxstat;
Do we really require mailboxstat?