Re: [RESEND PATCH v3 1/2] dt-bindings: i2c: Add Mediatek MT2701 i2c binding

[Matthias Brugger <matthias.bgg@gmail.com>]

On 23/05/17 05:34, Jun Gao wrote:
> On Wed, 2017-05-10 at 12:35 +0200, Matthias Brugger wrote:
>>
>> On 28/03/17 21:22, Wolfram Sang wrote:
>>> On Tue, Mar 28, 2017 at 05:50:12PM +0800, Jun Gao wrote:
>>>> On Wed, 2017-03-22 at 10:05 +0100, Wolfram Sang wrote:
>>>>> On Thu, Mar 09, 2017 at 11:13:04AM +0800, Jun Gao wrote:
>>>>>> From: Jun Gao <jun.gao@mediatek.com>
>>>>>>
>>>>>> Add MT2701 i2c binding to i2c-mt6577.txt and there is no need to
>>>>>> modify i2c driver.
>>>>>>
>>>>>> Signed-off-by: Jun Gao <jun.gao@mediatek.com>
>>>>>
>>>>> There is still the issue which Matthias Brugger pointed out on January,
>>>>> 24th: you need to describe the fallback compatibles.
>>>>>
>>>>
>>>> IMHO the value under compatible are just inputs which can be accepted.
>>>> It's little strange to add fallback information in binding. Some other
>>>> bindings do not describe so detailedly.
>>>> Is it OK to make binding as minimum standard?
>>>> If we describe it very detailedly, we will have to modify binding if
>>>> there are some changes for mt2701.
>>>
>>> My reading of the below is that I could simply use "mediatek,mt2701-i2c"
>>> as compatible and things will work. But it won't, we don't have that in
>>> the driver IIRC. So, we need a fallback for that to work.
>>>
>>>> Rob, could you give some suggestions?  Thanks!
>>>
>>> Would be welcome, yes. I lost track what the preferred solution is.
>>>
>>
>> We will need to define the fallback binding for each SoC.
>> As example take the rockchip mmc:
>> http://elixir.free-electrons.com/linux/latest/source/Documentation/devicetree/bindings/mmc/rockchip-dw-mshc.txt
>>
>> Regards,
>> Matthias
>>
> 
> Hi Rob,
> 
> There is no such requirement for other components.

The requirement is that the binding description reflects a compatible 
the driver is able to consume. Right now if you pass 
"mediatek,mt2701-i2c" the driver won't recognize this, therefor you need 
the fallback compatible. Which point isn't clear to you?

Regards,
Matthias

> Could you give us some suggestions?
> Thanks!
> 
> 
>>>>
>>>>>> ---
>>>>>>    .../devicetree/bindings/i2c/i2c-mt6577.txt         |   11 ++++++-----
>>>>>>    1 file changed, 6 insertions(+), 5 deletions(-)
>>>>>>
>>>>>> diff --git a/Documentation/devicetree/bindings/i2c/i2c-mt6577.txt b/Documentation/devicetree/bindings/i2c/i2c-mt6577.txt
>>>>>> index 0ce6fa3..27dbbf9 100644
>>>>>> --- a/Documentation/devicetree/bindings/i2c/i2c-mt6577.txt
>>>>>> +++ b/Documentation/devicetree/bindings/i2c/i2c-mt6577.txt
>>>>>> @@ -4,11 +4,12 @@ The Mediatek's I2C controller is used to interface with I2C devices.
>>>>>>    
>>>>>>    Required properties:
>>>>>>      - compatible: value should be either of the following.
>>>>>> -      (a) "mediatek,mt6577-i2c", for i2c compatible with mt6577 i2c.
>>>>>> -      (b) "mediatek,mt6589-i2c", for i2c compatible with mt6589 i2c.
>>>>>> -      (c) "mediatek,mt8127-i2c", for i2c compatible with mt8127 i2c.
>>>>>> -      (d) "mediatek,mt8135-i2c", for i2c compatible with mt8135 i2c.
>>>>>> -      (e) "mediatek,mt8173-i2c", for i2c compatible with mt8173 i2c.
>>>>>> +	"mediatek,mt2701-i2c"
>>>>>> +	"mediatek,mt6577-i2c"
>>>>>> +	"mediatek,mt6589-i2c"
>>>>>> +	"mediatek,mt8127-i2c"
>>>>>> +	"mediatek,mt8135-i2c"
>>>>>> +	"mediatek,mt8173-i2c"
>>>>>>      - reg: physical base address of the controller and dma base, length of memory
>>>>>>        mapped region.
>>>>>>      - interrupts: interrupt number to the cpu.
>>>>>> -- 
>>>>>> 1.7.9.5
>>>>>>
>>>>
>>>>
> 
>