Hi Shimoda-san, On 06 July 2015 08:18, Shimoda-san wrote: > Hi Phil-san, > > Thank you very much for the patch! > > > Sent: Thursday, July 02, 2015 5:06 PM > < snip > > > +/* VBUS change IRQ handler */ > > +static irqreturn_t gpio_vbus_irq(int irq, void *data) > > +{ > > + struct rcar_gen2_channel *channel = data; > > + > > + /* Wait 20ms before doing anything as VBUS needs to settle */ > > + schedule_delayed_work(&channel->work, msecs_to_jiffies(20)); > > + > > + return IRQ_HANDLED; > > +} > > + > > +static int probe_otg(struct platform_device *pdev, > > + struct rcar_gen2_phy_driver *drv) > > +{ > > + struct device *dev = &pdev->dev; > > + struct rcar_gen2_channel *ch = drv->channels; > > + int irq; > > + int ret; > > + > > + /* GPIOs for Host/Fn switching */ > > + ch->gpio_id = of_get_named_gpio_flags(dev->of_node, > > + "renesas,id-gpio", 0, NULL); > > + ch->gpio_vbus = of_get_named_gpio_flags(dev->of_node, > > + "renesas,vbus-gpio", 0, NULL); > > + > > + /* Need valid ID and VBUS gpios for Host/Fn switching */ > > + if (gpio_is_valid(ch->gpio_id) && gpio_is_valid(ch->gpio_vbus)) { > > + ch->use_otg = 1; > > + > > + /* GPIO for ID input */ > > + ret = devm_gpio_request_one(dev, ch->gpio_id, GPIOF_IN, > "id"); > > + if (ret) > > + return ret; > > + > > + /* GPIO for VBUS input */ > > + ret = devm_gpio_request_one(dev, ch->gpio_vbus, GPIOF_IN, > "vbus"); > > According to the checkpatch.pl, "line over 80 characters". As I understand it, the 80 chars is not a hard rule. One reason for the 80 char guideline is to avoid overly complex code, but I don't think that can be said of the above line! > > + if (ret) > > + return ret; > > + > > + irq = gpio_to_irq(ch->gpio_vbus); > > + ret = devm_request_irq(dev, irq, gpio_vbus_irq, > VBUS_IRQ_FLAGS, > > + "vbus_detect", ch); > > + if (ret) > > + return ret; > > + > > + /* Optional GPIO for VBUS power */ > > + ch->gpio_vbus_pwr = of_get_named_gpio_flags(dev->of_node, > > + "renesas,vbus-pwr-gpio", 0, > NULL); > > Same above. Since this code was already split over two lines, I agree with you here. > > > + if (gpio_is_valid(ch->gpio_id)) { > > + ret = devm_gpio_request_one(dev, ch->gpio_vbus_pwr, > > + GPIOF_OUT_INIT_LOW, "vbus-pwr"); > > + if (ret) > > + return ret; > > + } > > + > > + } else if (gpio_is_valid(ch->gpio_id)) { > > + dev_err(dev, "Failed to get VBUS gpio\n"); > > + return ch->gpio_vbus; > > + } else if (gpio_is_valid(ch->gpio_vbus)) { > > + dev_err(dev, "Failed to get ID gpio\n"); > > + return ch->gpio_id; > > + } > > + > > + return 0; > > +} > > + > > +/* bind/unbind the peripheral controller */ > > +static int rcar_gen2_usb_set_peripheral(struct usb_otg *otg, > > + struct usb_gadget *gadget) > > +{ > > + otg->gadget = gadget; > > + if (!gadget) { > > + usb_gadget_vbus_disconnect(otg->gadget); > > Since the otg->gadget is NULL, this driver should not call this function here. Ah, yes you are correct! > > + otg->state = OTG_STATE_UNDEFINED; > > + } > > + > > + return 0; > > +} > > + > > static int rcar_gen2_phy_probe(struct platform_device *pdev) > > { > > struct device *dev = &pdev->dev; > < snip > > > @@ -323,6 +540,14 @@ static int rcar_gen2_phy_probe(struct platform_device > *pdev) > > > > dev_set_drvdata(dev, drv); > > > > + /* > > + * If we already have something plugged into USB0, we won't get an edge > > + * on VBUS, so we have to manually schedule work to look at the VBUS > > + * and ID signals. > > + */ > > + if (drv->channels->use_otg) > > + schedule_delayed_work(&drv->channels->work, > msecs_to_jiffies(100)); > > This line is also "line over 80 characters". I don't think it is necessary to force this into the 80 char recommendation. > Best regards, > Yoshihiro Shimoda > > > + > > return 0; > > } > > > > -- > > 1.9.1 Best regards Phil -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html