Hi Rob, On 01/04/22 05:50, Rob Herring wrote: > On Thu, Mar 24, 2022 at 12:53:08PM +0100, Krzysztof Kozlowski wrote: >> On 24/03/2022 12:40, Aswath Govindraju wrote: >>> Hi Krzysztof, >>> >>> On 24/03/22 16:37, Krzysztof Kozlowski wrote: >>>> On 24/03/2022 08:34, Aswath Govindraju wrote: >>>>> Add bindings for the TI's AM62 wrapper module for the Synopsys USBSS-DRD >>>>> controller. >>>>> >>>>> Signed-off-by: Aswath Govindraju <a-govindraju@xxxxxx> >>>>> --- >>>>> >>>>> Changes since v1: >>>>> - made correction in grammer of clocks property description >>>>> and added maxItems in the interrupts property based on comments >>>>> received from Roger >>>>> - corrected the title, fixed the description of >>>>> ti,syscon-phy-pll-refclk, added pattern properties and child node >>>>> in the example based on the comments from Krzysztof. >>>>> >>>>> .../devicetree/bindings/usb/ti,am62-usb.yaml | 117 ++++++++++++++++++ >>>>> 1 file changed, 117 insertions(+) >>>>> create mode 100644 Documentation/devicetree/bindings/usb/ti,am62-usb.yaml >>>>> >>>>> diff --git a/Documentation/devicetree/bindings/usb/ti,am62-usb.yaml b/Documentation/devicetree/bindings/usb/ti,am62-usb.yaml >>>>> new file mode 100644 >>>>> index 000000000000..452bfdc6fb09 >>>>> --- /dev/null >>>>> +++ b/Documentation/devicetree/bindings/usb/ti,am62-usb.yaml >>>>> @@ -0,0 +1,117 @@ >>>>> +# SPDX-License-Identifier: GPL-2.0-only OR BSD-2-Clause >>>>> +%YAML 1.2 >>>>> +--- >>>>> +$id: http://devicetree.org/schemas/usb/ti,am62-usb.yaml# >>>>> +$schema: http://devicetree.org/meta-schemas/core.yaml# >>>>> + >>>>> +title: TI's AM62 wrapper module for the Synopsys USBSS-DRD controller >>>>> + >>>>> +maintainers: >>>>> + - Aswath Govindraju <a-govindraju@xxxxxx> >>>>> + >>>>> +properties: >>>>> + compatible: >>>>> + const: ti,am62-usb >>>>> + >>>>> + reg: >>>>> + maxItems: 1 >>>>> + >>>>> + ranges: true >>>>> + >>>>> + power-domains: >>>>> + description: >>>>> + PM domain provider node and an args specifier containing >>>>> + the USB ISO device id value. See, >>>>> + Documentation/devicetree/bindings/soc/ti/sci-pm-domain.yaml >>>>> + maxItems: 1 >>>>> + >>>>> + clocks: >>>>> + description: Clock phandle to usb2_refclk >>>>> + maxItems: 1 >>>>> + >>>>> + clock-names: >>>>> + items: >>>>> + - const: ref >>>>> + >>>>> + id-gpio: >>>>> + description: >>>>> + GPIO to be used as ID pin >>>>> + maxItems: 1 >>>> >>>> I have doubts about this. If you USB controller handles the ID pin, then >>>> probably this should be moved to usb-connector.yaml. I did not see >>>> id-gpio in any other USB controller blocks. >>>> >>> >>> Yes, the USB wrapper handles the ID pin operation only. It also reads >>> the status of VBUS by reading a register from its MMR and not using a >>> gpio. After evaluating the role the based on the states if id pin and >>> VBUS, this role is communicated to the dwc3 core driver using extcon. >>> There is no way for the dwc3 driver to detect the role on its own. >>> >>> >>> The usb-connector(drivers/usb/common/usb-conn-gpio.c) driver, seems to >>> be implemented for driving the VBUS, based on ID and VBUS pin status. >>> However, in case of the above implementation we need to communicate the >>> detected role to the dwc3 core driver. Also, the wrapper does not >>> control VBUS but it is the dwc3 core driver that drives the VBUS. >>> Therefore, I think the usb-connector implementation cannot be used here. >> >> I don't think about usb-conn-gpio.c but using the binding generic >> binding for usb-X-connector and define a connector with ID. >> >> Actually Rob could help here. >> >> Rob, >> Should the id-gpio be modeled as a property in this glue/wrapper driver >> or rather as part of usb-connector child node? > > That's a simple question. Where does the ID GPIO signal go to? The > connector, so it goes in the connector node. > Thank you for the clarification. Here ID-gpio is directly read by the wrapper and hence, I have modeled it as a property in the wrapper dt node. May I know if this wrong and should the modelling be looked at differently? Thanks, Aswath > If we have a driver for the usb-connector node, that's news to me. Not > that we couldn't, but that has nothing to do with designing the binding. > > Rob
