[PATCH net-next v4 2/9] dpll: spec: Add Netlink spec in YAML

[Kubalewski, Arkadiusz]

>From: Jakub Kicinski <kuba at kernel.org>
>Sent: Tuesday, August 15, 2023 4:44 AM
>
>On Fri, 11 Aug 2023 21:03:33 +0100 Vadim Fedorenko wrote:
>> +attribute-sets:
>> +  -
>> +    name: dpll
>> +    enum-name: dpll_a
>> +    attributes:
>> +      -
>> +        name: id
>> +        type: u32
>> +        value: 1
>
>value: 1 is the default
>
>> +      -
>> +        name: module-name
>> +        type: string
>> +      -
>> +        name: clock-id
>> +        type: u64
>
>I don't see a pad, you have a u64 without a pad?
>
>> +      -
>> +        name: mode
>> +        type: u8
>
>It's an enum, should always be u32 or bigger at protocol level.
>Netlink pads to 4B, you're not saving any space.
>
>> +        enum: mode
>> +      -
>> +        name: mode-supported
>> +        type: u8
>
>Same.
>
>> +        enum: mode
>> +        multi-attr: true
>> +      -
>> +        name: lock-status
>> +        type: u8
>
>And here.
>
>> +        enum: lock-status
>> +      -
>> +        name: temp
>> +        type: s32
>> +      -
>> +        name: type
>> +        type: u8
>
>And here, etc.
>
>Why are all attributes in a single attr space? :(
>More than half of them are prefixed with a pin- does it really
>not scream to you that they belong to a different space?
>

I agree, but there is an issue with this, currently:

name: pin-parent-device
subset-of: dpll
attributes:
  -
    name: id
    type: u32
  -
    name: pin-direction
    type: u32
  -
    name: pin-prio
    type: u32
  -
    name: pin-state
    type: u32

Where "id" is a part of device space, rest attrs would be a pin space..
Shall we have another argument for device id in a pin space?

Thank you!
Arkadiusz


>> +operations:
>> +  enum-name: dpll_cmd
>> +  list:
>> +    -
>> +      name: device-id-get
>> +      doc: |
>> +        Get id of dpll device that matches given attributes
>> +      value: 1
>
>is the default
>--
>pw-bot: cr