Re: [PATCH v5 2/6] dt-bindings: usb: add binding for Type-B GPIO connector driver

[Chunfeng Yun]

On Tue, 2019-05-14 at 13:12 -0500, Rob Herring wrote:
> On Tue, May 14, 2019 at 04:47:19PM +0800, Chunfeng Yun wrote:
> > It's used to support dual role switch via GPIO when use Type-B
> > receptacle, typically the USB ID pin is connected to an input
> > GPIO pin
> > 
> > Signed-off-by: Chunfeng Yun <chunfeng.yun@xxxxxxxxxxxx>
> > ---
> > v5 changes:
> >  1. treat type-B connector as child device of USB controller's, but not
> >     as a separate virtual device, suggested by Rob
> >  2. put connector's port node under connector node, suggested by Rob
> > 
> > v4 no changes
> > 
> > v3 changes:
> >  1. treat type-B connector as a virtual device, but not child device of
> >     USB controller's
> > 
> > v2 changes:
> >   1. new patch to make binding clear suggested by Hans
> > ---
> >  .../bindings/usb/typeb-conn-gpio.txt          | 42 +++++++++++++++++++
> >  1 file changed, 42 insertions(+)
> >  create mode 100644 Documentation/devicetree/bindings/usb/typeb-conn-gpio.txt
> > 
> > diff --git a/Documentation/devicetree/bindings/usb/typeb-conn-gpio.txt b/Documentation/devicetree/bindings/usb/typeb-conn-gpio.txt
> > new file mode 100644
> > index 000000000000..20dd3499a348
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/usb/typeb-conn-gpio.txt
> > @@ -0,0 +1,42 @@
> > +USB Type-B GPIO Connector
> > +
> > +This is used to switch dual role mode from the USB ID pin connected to
> > +an input GPIO pin.
> > +
> > +Required properties:
> > +- compatible : should include "linux,typeb-conn-gpio" and "usb-b-connector".
> 
> I don't think we need "linux,typeb-conn-gpio". 
Not all usb-b-connector child node need bind this driver, by adding the
new compatible can avoid unnecessary binding.

> A driver can decide to 
> handle GPIO lines if they present
Yes, the driver, e.g. USB controller driver can do it, but here I want
to provide a common driver to handle this special case, like
extcon-usb-gpio driver does, and try to keep transparency from USB
controller driver. 

>  or we assume the parent device handles 
> ID and/or Vbus if they are not present.
Yes, it will
> 
> > +- id-gpios, vbus-gpios : either one of them must be present, and both
> > +	can be present as well.
> 
> Please clarify that vbus-gpios is an input to sense Vbus presence as an 
> output it should be modelled as a regulator only.
Ok, will add more description.
> 
> These should be added to usb-connector.txt.
Already add them in [1/6].
> 
> The result of all this is you don't need this file. Just additions to 
> usb-connector.txt.
Here add more constrains for id-gpios and vbus-gpios, at least one
should be present, although they are both optional, this is not true for
some cases, so not suitable to add into usb-connector.txt.
> 
> > +- vbus-supply : can be present if needed when supports dual role mode or
> > +	host mode.
> > +	see connector/usb-connector.txt
> > +
> > +Sub-nodes:
> > +- port : should be present.
> > +	see graph.txt
> > +
> > +Example:
> > +
> > +&mtu3 {
> > +	status = "okay";
> 
> Don't show status in examples.
Ok, will drop it.
> 
> > +
> > +	connector {
> > +		compatible = "linux,typeb-conn-gpio", "usb-b-connector";
> > +		label = "micro-USB";
> > +		type = "micro";
> > +		id-gpios = <&pio 12 GPIO_ACTIVE_HIGH>;
> > +		vbus-supply = <&usb_p0_vbus>;
> > +
> > +		port {
> > +			bconn_ep: endpoint@0 {
> > +				remote-endpoint = <&usb_role_sw>;
> > +			};
> > +		};
> > +	};
> > +
> > +	port {
> > +		usb_role_sw: endpoint@0 {
> > +			remote-endpoint = <&bconn_ep>;
> > +		};
> > +	};
> 
> When the host controller is the parent of the connector, you don't need 
> the graph unless you're describing the alternate modes in Type-C.
Ok, got it.

Thanks a lot.

> 
> > +};
> > -- 
> > 2.21.0
> >