Quoting Peter Chen (2016-07-07 02:14:51) > diff --git a/drivers/usb/chipidea/host.c b/drivers/usb/chipidea/host.c > index 053bac9..55120ef 100644 > --- a/drivers/usb/chipidea/host.c > +++ b/drivers/usb/chipidea/host.c > @@ -109,15 +109,25 @@ static int host_start(struct ci_hdrc *ci) > struct ehci_hcd *ehci; > struct ehci_ci_priv *priv; > int ret; > + struct device *dev = ci->dev; > > - if (usb_disabled()) > + if (usb_disabled() || !dev) Does that ever happen? > return -ENODEV; > > - hcd = usb_create_hcd(&ci_ehci_hc_driver, ci->dev, dev_name(ci->dev)); > + /* > + * USB Core will try to get child node under roothub, > + * but chipidea core has no of_node, and the child node > + * for controller is located at glue layer's node which > + * is chipidea core's parent. > + */ > + if (dev->parent && dev->parent->of_node) > + dev->of_node = dev->parent->of_node; Why not do this during the ci device probe? How is host special for having an of_node for the child device. Doing that would also make my ULPI bus DT discovery patch simpler because we wouldn't have to search the parent of the parent for a DT node to find the ULPI node that's a child of the ci glue device. -- 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