Message-ID: <dfcb1556-fa09-4d2e-b128-341e9725d3de@linaro.org>
Date: Thu, 15 Feb 2024 09:34:53 +0100
From: Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>
To: Stephen Boyd <swboyd@...omium.org>, chrome-platform@...ts.linux.dev
Cc: linux-kernel@...r.kernel.org, patches@...ts.linux.dev,
 linux-arm-kernel@...ts.infradead.org, devicetree@...r.kernel.org,
 linux-arm-msm@...r.kernel.org, Douglas Anderson <dianders@...omium.org>,
 Pin-yen Lin <treapking@...omium.org>, Rob Herring <robh+dt@...nel.org>,
 Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
 Conor Dooley <conor+dt@...nel.org>, Lee Jones <lee@...nel.org>,
 Benson Leung <bleung@...omium.org>, Guenter Roeck <groeck@...omium.org>,
 Prashant Malani <pmalani@...omium.org>, Tzung-Bi Shih <tzungbi@...nel.org>
Subject: Re: [PATCH 13/22] dt-bindings: chrome: Add
 google,cros-ec-typec-switch binding

On 15/02/2024 02:52, Stephen Boyd wrote:
> Quoting Krzysztof Kozlowski (2024-02-11 05:34:25)
>> On 10/02/2024 08:09, Stephen Boyd wrote:
>>> Add a binding for the USB type-c switch controls found on some ChromeOS
>>> Embedded Controllers (ECs). When this device is a mode switch, it takes
>>> one DisplayPort (DP) port as input and some number (possibly zero) of
>>> USB SuperSpeed ports (bundles of USB SS lanes) as input, and muxes those
>>> lanes into USB type-c SuperSpeed lanes suitable for the SSTRX1/2 pins on
>>> a usb-c-connector. When this device is an orientation switch, it
>>> redirects the DP lanes to the proper USB type-c SSTRX lanes.
>>>
>>> Cc: Rob Herring <robh+dt@...nel.org>
>>> Cc: Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>
>>> Cc: Conor Dooley <conor+dt@...nel.org>
>>> Cc: Lee Jones <lee@...nel.org>
>>> Cc: Benson Leung <bleung@...omium.org>
>>> Cc: Guenter Roeck <groeck@...omium.org>
>>> Cc: Prashant Malani <pmalani@...omium.org>
>>> Cc: Tzung-Bi Shih <tzungbi@...nel.org>
>>> Cc: <devicetree@...r.kernel.org>
>>> Cc: <chrome-platform@...ts.linux.dev>
>>> Cc: Pin-yen Lin <treapking@...omium.org>
>>> Signed-off-by: Stephen Boyd <swboyd@...omium.org>
>>> ---
>>>  .../chrome/google,cros-ec-typec-switch.yaml   | 365 ++++++++++++++++++
>>>  .../bindings/mfd/google,cros-ec.yaml          |   5 +
>>>  2 files changed, 370 insertions(+)
>>>  create mode 100644 Documentation/devicetree/bindings/chrome/google,cros-ec-typec-switch.yaml
>>>
>>> diff --git a/Documentation/devicetree/bindings/chrome/google,cros-ec-typec-switch.yaml b/Documentation/devicetree/bindings/chrome/google,cros-ec-typec-switch.yaml
>>> new file mode 100644
>>> index 000000000000..17a0ba928f5d
>>> --- /dev/null
>>> +++ b/Documentation/devicetree/bindings/chrome/google,cros-ec-typec-switch.yaml
>>> @@ -0,0 +1,365 @@
>>> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
>>> +%YAML 1.2
>>> +---
>>> +$id: http://devicetree.org/schemas/chrome/google,cros-ec-typec-switch.yaml#
>>> +$schema: http://devicetree.org/meta-schemas/core.yaml#
>>> +
>>> +title: Google Chrome OS EC(Embedded Controller) USB Type C Switch
>>> +
>>> +maintainers:
>>> +  - Benson Leung <bleung@...omium.org>
>>> +  - Prashant Malani <pmalani@...omium.org>
>>> +  - Stephen Boyd <swboyd@...omium.org>
>>> +
>>> +description:
>>> +  Chrome OS devices have an Embedded Controller(EC) which has access to USB
>>> +  Type C switching. This node is intended to allow the OS to control Type C
>>> +  signal muxing for USB-C orientation and alternate modes. The node for this
>>> +  device should be under a cros-ec node like google,cros-ec-spi.
>>> +
>>
>> If this is USB Type C switch, then you miss reference to
>> usb-switch.yaml, but then ports look a bit different.
> 
> Thanks for the pointer. I suspect that's in linux-next only? I'm going

Yes, Greg's tree.

> to move this to the cros-ec-typec.yaml file but I'll still need some
> sort of property like 'mode-switch' or 'orientation-switch' to describe
> what needs to be done in the kernel by changing lane assignments in the
> drm_bridge.
> 
> One problem with those properties is that they're boolean for the whole
> device. If I have a google,cros-ec-typec node that has two input DP
> ports and two output USB-C ports then it may be that one port needs
> orientation switching and the other only needs to do mode switching.
> This needs to be expressed somehow and a top-level boolean property
> doesn't do that. It could be part of the DP endpoint node itself.

Maybe it can be added to the connector binding as well.

> 
> Also I don't know how to indicate that mode switching can't be changed
> here directly. For example, on Trogdor the mode is switched by the EC,
> and the kernel can't change the mode. Changing DP lane assignments isn't
> going to change the situation either. The mode is still going to be
> something like DP+USB, or just DP, etc. Maybe there needs to be a
> different property, like 'dp-mode-lane-assignment = <array of ports>',
> that indicates which DP ports need to do lane assignment or
> 'dp-orientation-lane-assignment = <array of ports>' for orientation
> control. Either way, I'm saying that 'mode-switch' and
> 'orientation-switch' properties don't make any sense here. I was using
> them to wedge the code into the typec switch callbacks, but if I move
> this into the cros-ec-typec driver then I don't need them.

OK

> 
>>
>>> +
>>> +  no-hpd:
>>> +    description: Indicates this device doesn't signal HPD for DisplayPort
>>> +    type: boolean
>>> +
>>> +  ports:
> [...]
>>> +
>>> +          endpoint@7:
>>> +            $ref: /schemas/graph.yaml#/$defs/endpoint-base
>>> +            description: USB-C data for EC's 7th type-c port
>>> +            unevaluatedProperties: false
>>> +            properties:
>>> +              data-lanes:
>>> +                $ref: /schemas/types.yaml#/definitions/uint32-array
>>> +                description: |
>>> +                  An array of physical USB-C data lane indexes.
>>> +                  - 0 is SSRX1 lane
>>> +                  - 1 is SSTX1 lane
>>> +                  - 2 is SSTX2 lane
>>> +                  - 3 is SSRX2 lane
>>> +                minItems: 4
>>> +                maxItems: 4
>>> +                items:
>>> +                  maximum: 3
>>> +
>>> +        anyOf:
>>> +          - required:
>>> +              - endpoint@0
>>
>> I don't get what you want to say here. This anyOf should have no effect.
> 
> I'm trying to say that there should be at least one usb-c data lane
> output endpoint if there's a port@2 (usb-c output).

Wait, my bad, this is good. At least one endpoint will be required.

> 
>>
>>> +          - required:
>>> +              - endpoint@1
>>> +          - required:
>>> +              - endpoint@2
> [...]
>>
>>> +
>>> +required:
>>> +  - compatible
>>> +  - ports
>>> +
>>> +allOf:
>>> +  - if:
>>> +      properties:
>>> +        no-hpd: true
>>
>> I don't understand this either. What is it for? Where did you see such
>> syntax?
> 
> This is saying that if the no-hpd property is present then the port@0
> port (DP input port) is required. Otherwise no-hpd doesn't really make
> any sense because this device isn't muxing DP onto USB type-c connectors.

Then you only need if: required: like here:

https://elixir.bootlin.com/linux/v6.4-rc7/source/Documentation/devicetree/bindings/net/qcom,ipa.yaml#L174

> 
> I found this syntax by searching the schema website and reading this
> page[1]. The last yellow note about "country" not being a required
> property led me to this syntax.
> 
> [1] https://json-schema.org/understanding-json-schema/reference/conditionals#ifthenelse

Best regards,
Krzysztof

Powered by blists - more mailing lists