Re: [PATCH v3 02/10] bindings: ipmi: Add binding for IPMB device intf

[Ninad Palsule <ninad@linux.ibm.com>]

Hello Cprey,

On 1/10/25 11:53, Corey Minyard wrote:
> On Fri, Jan 10, 2025 at 10:07:13AM -0600, Rob Herring wrote:
>> On Wed, Jan 08, 2025 at 10:36:30AM -0600, Ninad Palsule wrote:
>>> Add device tree binding document for the IPMB device interface.
>>> This device is already in use in both driver and .dts files.
>>>
>>> Signed-off-by: Ninad Palsule <ninad@linux.ibm.com>
>>> ---
>>>   .../devicetree/bindings/ipmi/ipmb-dev.yaml    | 44 +++++++++++++++++++
>>>   1 file changed, 44 insertions(+)
>>>   create mode 100644 Documentation/devicetree/bindings/ipmi/ipmb-dev.yaml
>>>
>>> diff --git a/Documentation/devicetree/bindings/ipmi/ipmb-dev.yaml b/Documentation/devicetree/bindings/ipmi/ipmb-dev.yaml
>>> new file mode 100644
>>> index 000000000000..a8f46f1b883e
>>> --- /dev/null
>>> +++ b/Documentation/devicetree/bindings/ipmi/ipmb-dev.yaml
>>> @@ -0,0 +1,44 @@
>>> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
>>> +%YAML 1.2
>>> +---
>>> +$id: http://devicetree.org/schemas/ipmi/ipmb-dev.yaml#
>>> +$schema: http://devicetree.org/meta-schemas/core.yaml#
>>> +
>>> +title: IPMB Device
>>> +
>>> +description: IPMB Device interface to receive request and send response
> First, thank you, this does need to be documented.
>
>> IPMB is not defined anywhere.
> Indeed.  At least reference the spec, but better do that and provide a
> basic description.
>
>> Which side of the interface does this apply to? How do I know if I have
>> an ipmb-dev?
>>
>> This document needs to stand on its own. Bindings exist in a standalone
>> tree without kernel drivers or docs.
> At least to someone who knows what IPMB is, it's pretty clear that you
> are saying "The i2c device this node is in is on an IPMB bus." However,
> to someone who is not, this is all a foreign language.  This definitely
> needs better documentation.
>
> Why do you have a "reg" property?  I don't see it referenced in the
> driver.  I assume that's the I2C address, but that's going to be the
> same as what's in the containing I2C node.  I don't think it's
> necessary.

Sorry forgot to answer this question. We are setting extra values for 
reg for I2C slave.

I am not sure how it is used in the driver.

Regards,

Ninad

>
> -corey
>
>>> +
>>> +maintainers:
>>> +  - Ninad Palsule <ninad@linux.ibm.com>
>>> +
>>> +properties:
>>> +  compatible:
>>> +    enum:
>>> +      - ipmb-dev
>>> +
>>> +  reg:
>>> +    maxItems: 1
>>> +
>>> +  i2c-protocol:
>>> +    description:
>>> +      Use I2C block transfer instead of SMBUS block transfer.
>>> +    type: boolean
>>> +
>>> +required:
>>> +  - compatible
>>> +  - reg
>>> +
>>> +additionalProperties: false
>>> +
>>> +examples:
>>> +  - |
>>> +    i2c {
>>> +        #address-cells = <1>;
>>> +        #size-cells = <0>;
>>> +
>>> +        ipmb-dev@10 {
>>> +            compatible = "ipmb-dev";
>>> +            reg = <0x10>;
>>> +            i2c-protocol;
>>> +        };
>>> +    };
>>> -- 
>>> 2.43.0
>>>