Hi, Wesley Cheng <[email protected]> writes: > @@ -190,6 +195,73 @@ static int dwc3_qcom_register_extcon(struct dwc3_qcom > *qcom) > return 0; > } > > +static int dwc3_qcom_usb_role_switch_set(struct usb_role_switch *sw, > + enum usb_role role) > +{ > + struct dwc3_qcom *qcom = usb_role_switch_get_drvdata(sw); > + struct fwnode_handle *child; > + bool enable = false; > + > + if (!qcom->dwc3_drd_sw) { > + child = device_get_next_child_node(qcom->dev, NULL); > + if (child) { > + qcom->dwc3_drd_sw = > usb_role_switch_find_by_fwnode(child); > + fwnode_handle_put(child); > + if (IS_ERR(qcom->dwc3_drd_sw)) { > + qcom->dwc3_drd_sw = NULL; > + return 0; > + } > + } > + } > + > + usb_role_switch_set_role(qcom->dwc3_drd_sw, role);
why is this done at the glue layer instead of core.c? > + if (role == USB_ROLE_DEVICE) > + enable = true; > + else > + enable = false; > + > + qcom->mode = (role == USB_ROLE_HOST) ? USB_DR_MODE_HOST : > + USB_DR_MODE_PERIPHERAL; > + dwc3_qcom_vbus_overrride_enable(qcom, enable); could you add a patch fixing this typo? -- balbi
signature.asc
Description: PGP signature
