[PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg

Samin Guo posted 2 patches 4 months ago
There is a newer version of this series
[PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
Posted by Samin Guo 4 months ago
The motorcomm phy (YT8531) supports the ability to adjust the drive
strength of the rx_clk/rx_data, the value range of pad driver
strength is 0 to 7.

Signed-off-by: Samin Guo <samin.guo@starfivetech.com>
---
 .../devicetree/bindings/net/motorcomm,yt8xxx.yaml    | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml b/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml
index 157e3bbcaf6f..29a1997a1577 100644
--- a/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml
+++ b/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml
@@ -52,6 +52,18 @@ properties:
       for a timer.
     type: boolean
 
+  motorcomm,rx-clk-driver-strength:
+    description: drive strength of rx_clk pad.
+    $ref: /schemas/types.yaml#/definitions/uint32
+    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
+    default: 3
+
+  motorcomm,rx-data-driver-strength:
+    description: drive strength of rx_data/rx_ctl rgmii pad.
+    $ref: /schemas/types.yaml#/definitions/uint32
+    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
+    default: 3
+
   motorcomm,tx-clk-adj-enabled:
     description: |
       This configuration is mainly to adapt to VF2 with JH7110 SoC.
-- 
2.17.1
Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
Posted by Conor Dooley 4 months ago
On Fri, May 26, 2023 at 05:05:01PM +0800, Samin Guo wrote:
> The motorcomm phy (YT8531) supports the ability to adjust the drive
> strength of the rx_clk/rx_data, the value range of pad driver
> strength is 0 to 7.
> 
> Signed-off-by: Samin Guo <samin.guo@starfivetech.com>
> ---
>  .../devicetree/bindings/net/motorcomm,yt8xxx.yaml    | 12 ++++++++++++
>  1 file changed, 12 insertions(+)
> 
> diff --git a/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml b/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml
> index 157e3bbcaf6f..29a1997a1577 100644
> --- a/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml
> +++ b/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml
> @@ -52,6 +52,18 @@ properties:
>        for a timer.
>      type: boolean
>  
> +  motorcomm,rx-clk-driver-strength:
> +    description: drive strength of rx_clk pad.
> +    $ref: /schemas/types.yaml#/definitions/uint32
> +    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]

I think you should use minimum & maximum instead of these listed out
enums. You have also had this comment since v1 & were reminded of it on
v2 by Krzysztof: "What do the numbers mean? What are the units? mA?"

This information should go into the binding, not sit in a thread on a
mailing list that noone will look at when trying to write a DT :(

Thanks,
Conor.

> +    default: 3
> +
> +  motorcomm,rx-data-driver-strength:
> +    description: drive strength of rx_data/rx_ctl rgmii pad.
> +    $ref: /schemas/types.yaml#/definitions/uint32
> +    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
> +    default: 3
> +
>    motorcomm,tx-clk-adj-enabled:
>      description: |
>        This configuration is mainly to adapt to VF2 with JH7110 SoC.
> -- 
> 2.17.1
> 
Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
Posted by Guo Samin 4 months ago
Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
From: Conor Dooley <conor@kernel.org>
to: Samin Guo <samin.guo@starfivetech.com>
data: 2023/5/27

> On Fri, May 26, 2023 at 05:05:01PM +0800, Samin Guo wrote:
>> The motorcomm phy (YT8531) supports the ability to adjust the drive
>> strength of the rx_clk/rx_data, the value range of pad driver
>> strength is 0 to 7.
>>
>> Signed-off-by: Samin Guo <samin.guo@starfivetech.com>
>> ---
>>  .../devicetree/bindings/net/motorcomm,yt8xxx.yaml    | 12 ++++++++++++
>>  1 file changed, 12 insertions(+)
>>
>> diff --git a/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml b/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml
>> index 157e3bbcaf6f..29a1997a1577 100644
>> --- a/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml
>> +++ b/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml
>> @@ -52,6 +52,18 @@ properties:
>>        for a timer.
>>      type: boolean
>>  
>> +  motorcomm,rx-clk-driver-strength:
>> +    description: drive strength of rx_clk pad.
>> +    $ref: /schemas/types.yaml#/definitions/uint32
>> +    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
> 
> I think you should use minimum & maximum instead of these listed out
> enums.

Thanks Conor, This can be improved in the next version.

 You have also had this comment since v1 & were reminded of it on
> v2 by Krzysztof: "What do the numbers mean? What are the units? mA?"
> 


The good news is that we just got some data about units from Motorcomm. 
Maybe I can post the data show of the unit later after I get the complete data.



> This information should go into the binding, not sit in a thread on a
> mailing list that noone will look at when trying to write a DT :(
> 
> Thanks,
> Conor.
>

Yes,when we have the complete 'unit' data, it will be placed in DT.

Best regards,
Samin
 
>> +    default: 3
>> +
>> +  motorcomm,rx-data-driver-strength:
>> +    description: drive strength of rx_data/rx_ctl rgmii pad.
>> +    $ref: /schemas/types.yaml#/definitions/uint32
>> +    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
>> +    default: 3
>> +
>>    motorcomm,tx-clk-adj-enabled:
>>      description: |
>>        This configuration is mainly to adapt to VF2 with JH7110 SoC.
>> -- 
>> 2.17.1
>>
Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
Posted by Guo Samin 3 months, 1 week ago
Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
From: Guo Samin <samin.guo@starfivetech.com>
to: Conor Dooley <conor@kernel.org>; Andrew Lunn <andrew@lunn.ch>
data: 2023/5/29

> Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
> From: Conor Dooley <conor@kernel.org>
> to: Samin Guo <samin.guo@starfivetech.com>
> data: 2023/5/27
> 
>> On Fri, May 26, 2023 at 05:05:01PM +0800, Samin Guo wrote:
>>> The motorcomm phy (YT8531) supports the ability to adjust the drive
>>> strength of the rx_clk/rx_data, the value range of pad driver
>>> strength is 0 to 7.
>>>
>>> Signed-off-by: Samin Guo <samin.guo@starfivetech.com>
>>> ---
>>>  .../devicetree/bindings/net/motorcomm,yt8xxx.yaml    | 12 ++++++++++++
>>>  1 file changed, 12 insertions(+)
>>>
>>> diff --git a/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml b/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml
>>> index 157e3bbcaf6f..29a1997a1577 100644
>>> --- a/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml
>>> +++ b/Documentation/devicetree/bindings/net/motorcomm,yt8xxx.yaml
>>> @@ -52,6 +52,18 @@ properties:
>>>        for a timer.
>>>      type: boolean
>>>  
>>> +  motorcomm,rx-clk-driver-strength:
>>> +    description: drive strength of rx_clk pad.
>>> +    $ref: /schemas/types.yaml#/definitions/uint32
>>> +    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
>>
>> I think you should use minimum & maximum instead of these listed out
>> enums.
> 
> Thanks Conor, This can be improved in the next version.
> 
>  You have also had this comment since v1 & were reminded of it on
>> v2 by Krzysztof: "What do the numbers mean? What are the units? mA?"
>>
> 
> 
> The good news is that we just got some data about units from Motorcomm. 
> Maybe I can post the data show of the unit later after I get the complete data.
>

Hi Andrew & Conor,

Sorry, haven't updated in a while.
I just got the detailed data of Driver Strength(DS) from Motorcomm , which applies to both rx_clk and rx_data.

|----------------------|
|     ds map table     |
|----------------------|
| DS(3b) | Current (mA)|
|--------|-------------|
|   000  |     1.20    |
|   001  |     2.10    |
|   010  |     2.70    |
|   011  |     2.91    |
|   100  |     3.11    |
|   101  |     3.60    |
|   110  |     3.97    |
|   111  |     4.35    |
|--------|-------------|

Since these currents are not integer values and have no regularity, it is not very good to use in the drive/dts in my opinion.

Therefore, I tend to continue to use DS(0-7) in dts/driver, and adding a description of the current value corresponding to DS in dt-bindings. 

Like This:

+  motorcomm,rx-clk-driver-strength:
+    description: drive strength of rx_clk pad.
+      |----------------------|
+      | rx_clk ds map table  |
+      |----------------------|
+      | DS(3b) | Current (mA)|
+      |   000  |     1.20    |
+      |   001  |     2.10    |
+      |   010  |     2.70    |
+      |   011  |     2.91    |
+      |   100  |     3.11    |
+      |   101  |     3.60    |
+      |   110  |     3.97    |
+      |   111  |     4.35    |
+      |--------|-------------|
+    $ref: /schemas/types.yaml#/definitions/uint32
+    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
+    default: 3
+
+  motorcomm,rx-data-driver-strength:
+    description: drive strength of rx_data/rx_ctl rgmii pad.
+      |----------------------|
+      | rx_data ds map table |
+      |----------------------|
+      | DS(3b) | Current (mA)|
+      |   000  |     1.20    |
+      |   001  |     2.10    |
+      |   010  |     2.70    |
+      |   011  |     2.91    |
+      |   100  |     3.11    |
+      |   101  |     3.60    |
+      |   110  |     3.97    |
+      |   111  |     4.35    |
+      |--------|-------------|
+    $ref: /schemas/types.yaml#/definitions/uint32
+    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
+    default: 3
+


Or use minimum & maximum instead of these listed out enums(Suggested by Conor)

+  motorcomm,rx-clk-driver-strength:
+    $ref: /schemas/types.yaml#/definitions/uint32
+    default: 3
+    minimum: 0
+    maximum: 7
+    description: drive strength of rx_clk pad.
+      |----------------------|
+      | rx_clk ds map table  |
+      |----------------------|
+      | DS(3b) | Current (mA)|
+      |   000  |     1.20    |
+      |   001  |     2.10    |
+      |   010  |     2.70    |
+      |   011  |     2.91    |
+      |   100  |     3.11    |
+      |   101  |     3.60    |
+      |   110  |     3.97    |
+      |   111  |     4.35    |
+      |--------|-------------|
+
+  motorcomm,rx-data-driver-strength:
+    $ref: /schemas/types.yaml#/definitions/uint32
+    default: 3
+    minimum: 0
+    maximum: 7
+    description: drive strength of rx_data/rx_ctl rgmii pad.
+      |----------------------|
+      | rx_data ds map table |
+      |----------------------|
+      | DS(3b) | Current (mA)|
+      |   000  |     1.20    |
+      |   001  |     2.10    |
+      |   010  |     2.70    |
+      |   011  |     2.91    |
+      |   100  |     3.11    |
+      |   101  |     3.60    |
+      |   110  |     3.97    |
+      |   111  |     4.35    |
+      |--------|-------------|
+


Looking forward to your suggestions.


Best regards,
Samin

> 
> 
>> This information should go into the binding, not sit in a thread on a
>> mailing list that noone will look at when trying to write a DT :(
>>
>> Thanks,
>> Conor.
>>
> 
> Yes,when we have the complete 'unit' data, it will be placed in DT.
> 
> Best regards,
> Samin
>  
>>> +    default: 3
>>> +
>>> +  motorcomm,rx-data-driver-strength:
>>> +    description: drive strength of rx_data/rx_ctl rgmii pad.
>>> +    $ref: /schemas/types.yaml#/definitions/uint32
>>> +    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
>>> +    default: 3
>>> +
>>>    motorcomm,tx-clk-adj-enabled:
>>>      description: |
>>>        This configuration is mainly to adapt to VF2 with JH7110 SoC.
>>> -- 
>>> 2.17.1
>>>
Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
Posted by Andrew Lunn 3 months, 1 week ago
> I just got the detailed data of Driver Strength(DS) from Motorcomm ,
> which applies to both rx_clk and rx_data.
> 
> |----------------------|
> |     ds map table     |
> |----------------------|
> | DS(3b) | Current (mA)|
> |--------|-------------|
> |   000  |     1.20    |
> |   001  |     2.10    |
> |   010  |     2.70    |
> |   011  |     2.91    |
> |   100  |     3.11    |
> |   101  |     3.60    |
> |   110  |     3.97    |
> |   111  |     4.35    |
> |--------|-------------|
>
> Since these currents are not integer values

Integers is not a problem. Simply use uA.

> and have no regularity, it is not very good to use in the drive/dts
> in my opinion.

I think they are fine to use. Add a lookup table, microamps to
register value. Return -EINVAL if the requested value is not in the
table. List the valid values in the schema, so the checker tool might
point out problems.

      Andrew
Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
Posted by Guo Samin 3 months, 1 week ago
Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
From: Andrew Lunn <andrew@lunn.ch>
to: Guo Samin <samin.guo@starfivetech.com>
data: 2023/6/20

>> I just got the detailed data of Driver Strength(DS) from Motorcomm ,
>> which applies to both rx_clk and rx_data.
>>
>> |----------------------|
>> |     ds map table     |
>> |----------------------|
>> | DS(3b) | Current (mA)|
>> |--------|-------------|
>> |   000  |     1.20    |
>> |   001  |     2.10    |
>> |   010  |     2.70    |
>> |   011  |     2.91    |
>> |   100  |     3.11    |
>> |   101  |     3.60    |
>> |   110  |     3.97    |
>> |   111  |     4.35    |
>> |--------|-------------|
>>
>> Since these currents are not integer values
> 
> Integers is not a problem. Simply use uA.
> 
>> and have no regularity, it is not very good to use in the drive/dts
>> in my opinion.
> 
> I think they are fine to use. Add a lookup table, microamps to
> register value. Return -EINVAL if the requested value is not in the
> table. List the valid values in the schema, so the checker tool might
> point out problems.
> 
>       Andrew

Thanks Andrew,
I'll use a lookup table to try.
Another thing we need to deal with DS under different IO voltages(1.8V/2.5V/3.3V).

The IO voltage can be configured via a hardware pull-up resistor (visionfive 2 is configured to 1.8V by default), 
and then the IO voltage can be obtained or set through the register(0xA001)

Chip_Config (EXT_0xA001)
|Bit  |Symbol  |Access  |Default  |Description |
|5:4  |Cfg_ldo |RW      |0x0      |Rgmii ldo voltage and RGMII/MDC/MDIO PAD's level shifter control. Depends on strapping.|
                                  |2'b11: 1.8v   2'b10: 1.8v    2'b01: 2.5v    2'b00: 3.3v                                |

      |----------------------|            
      | ds map table(1.8V)   |
      |----------------------|
      | DS(3b) | Current (mA)|
      |   000  |     1.20    |
      |   001  |     2.10    |
      |   010  |     2.70    |
      |   011  |     2.91    |
      |   100  |     3.11    |
      |   101  |     3.60    |
      |   110  |     3.97    |
      |   111  |     4.35    |
      |--------|-------------|


      |----------------------|
      | ds map table(3.3V)   |
      |----------------------|
      | DS(3b) | Current (mA)|
      |   000  |     3.07    |
      |   001  |     4.08    |
      |   010  |     4.37    |
      |   011  |     4.68    |
      |   100  |     5.02    |
      |   101  |     5.45    |
      |   110  |     5.74    |
      |   111  |     6.14    |
      |--------|-------------|
 
(The current value of 2.5V is not available to us now)

When we need to deal with current values at different voltages, using register values in drives may be simpler, comparing the current value.
Of course, I will also follow your suggestion and use the current value + lookup table to implement a solution. Thanks!

Best regards,
Samin
Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
Posted by Andrew Lunn 3 months, 1 week ago
> When we need to deal with current values at different voltages,
> using register values in drives may be simpler, comparing the
> current value.

Register values in DT are only allowed in the worst case if its all
undocumented vendor magic. You have the needed documentation, and its
easy code to write, lookup a value in a table.

     Andrew
Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
Posted by Conor Dooley 3 months, 1 week ago
Hey,

On Tue, Jun 20, 2023 at 11:09:52AM +0800, Guo Samin wrote:
> From: Guo Samin <samin.guo@starfivetech.com>
> > From: Conor Dooley <conor@kernel.org>
> >> On Fri, May 26, 2023 at 05:05:01PM +0800, Samin Guo wrote:
> >>> The motorcomm phy (YT8531) supports the ability to adjust the drive
> >>> strength of the rx_clk/rx_data, the value range of pad driver
> >>> strength is 0 to 7.

> >>> +  motorcomm,rx-clk-driver-strength:
> >>> +    description: drive strength of rx_clk pad.
> >>> +    $ref: /schemas/types.yaml#/definitions/uint32
> >>> +    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
> >>
> >> I think you should use minimum & maximum instead of these listed out
> >> enums.

> >  You have also had this comment since v1 & were reminded of it on
> >> v2 by Krzysztof: "What do the numbers mean? What are the units? mA?"

> > The good news is that we just got some data about units from Motorcomm. 
> > Maybe I can post the data show of the unit later after I get the complete data.

> Sorry, haven't updated in a while.

NW chief.

> I just got the detailed data of Driver Strength(DS) from Motorcomm , which applies to both rx_clk and rx_data.
> 
> |----------------------|
> |     ds map table     |
> |----------------------|
> | DS(3b) | Current (mA)|
> |--------|-------------|
> |   000  |     1.20    |
> |   001  |     2.10    |
> |   010  |     2.70    |
> |   011  |     2.91    |
> |   100  |     3.11    |
> |   101  |     3.60    |
> |   110  |     3.97    |
> |   111  |     4.35    |
> |--------|-------------|
> 
> Since these currents are not integer values and have no regularity,
> it is not very good to use in the drive/dts in my opinion.

Who says you have to use mA? What about uA?

> Therefore, I tend to continue to use DS(0-7) in dts/driver, and adding
> a description of the current value corresponding to DS in dt-bindings. 

I think this goes against not putting register values into the dts &
that the accurate description of the hardware are the currents.

> Like This:
> 
> +  motorcomm,rx-clk-driver-strength:
> +    description: drive strength of rx_clk pad.

You need "description: |" to preserve the formatting if you add tables,
but I don't think that this is a good idea. Put the values in here that
describe the hardware (IOW the currents) and then you don't need to have
this table.

> +      |----------------------|
> +      | rx_clk ds map table  |
> +      |----------------------|
> +      | DS(3b) | Current (mA)|
> +      |   000  |     1.20    |
> +      |   001  |     2.10    |
> +      |   010  |     2.70    |
> +      |   011  |     2.91    |
> +      |   100  |     3.11    |
> +      |   101  |     3.60    |
> +      |   110  |     3.97    |
> +      |   111  |     4.35    |
> +      |--------|-------------|
> +    $ref: /schemas/types.yaml#/definitions/uint32
> +    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
> +    default: 3
> +

> Or use minimum & maximum instead of these listed out enums

With the actual current values, enum rather than min + max.

Cheers,
Conor.
Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
Posted by Krzysztof Kozlowski 3 months, 1 week ago
On 20/06/2023 10:26, Conor Dooley wrote:
> Hey,
> 
> On Tue, Jun 20, 2023 at 11:09:52AM +0800, Guo Samin wrote:
>> From: Guo Samin <samin.guo@starfivetech.com>
>>> From: Conor Dooley <conor@kernel.org>
>>>> On Fri, May 26, 2023 at 05:05:01PM +0800, Samin Guo wrote:
>>>>> The motorcomm phy (YT8531) supports the ability to adjust the drive
>>>>> strength of the rx_clk/rx_data, the value range of pad driver
>>>>> strength is 0 to 7.
> 
>>>>> +  motorcomm,rx-clk-driver-strength:
>>>>> +    description: drive strength of rx_clk pad.
>>>>> +    $ref: /schemas/types.yaml#/definitions/uint32
>>>>> +    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
>>>>
>>>> I think you should use minimum & maximum instead of these listed out
>>>> enums.
> 
>>>  You have also had this comment since v1 & were reminded of it on
>>>> v2 by Krzysztof: "What do the numbers mean? What are the units? mA?"
> 
>>> The good news is that we just got some data about units from Motorcomm. 
>>> Maybe I can post the data show of the unit later after I get the complete data.
> 
>> Sorry, haven't updated in a while.
> 
> NW chief.
> 
>> I just got the detailed data of Driver Strength(DS) from Motorcomm , which applies to both rx_clk and rx_data.
>>
>> |----------------------|
>> |     ds map table     |
>> |----------------------|
>> | DS(3b) | Current (mA)|
>> |--------|-------------|
>> |   000  |     1.20    |
>> |   001  |     2.10    |
>> |   010  |     2.70    |
>> |   011  |     2.91    |
>> |   100  |     3.11    |
>> |   101  |     3.60    |
>> |   110  |     3.97    |
>> |   111  |     4.35    |
>> |--------|-------------|
>>
>> Since these currents are not integer values and have no regularity,

There is no mA unit in DT schema, so I don't see what by "not integer
values". 1200 uA is an integer.

>> it is not very good to use in the drive/dts in my opinion.
> 
> Who says you have to use mA? What about uA?

Yep

> 
>> Therefore, I tend to continue to use DS(0-7) in dts/driver, and adding
>> a description of the current value corresponding to DS in dt-bindings. 
> 
> I think this goes against not putting register values into the dts &
> that the accurate description of the hardware are the currents.

For vendor properties register values are often accepted, but logical
unit is much more readable in the DTS. Also allows further customization
or extending when new variant appears. You cannot do extend a property
easily when it holds a register value, without changing the meaning per
variant.

> 
>> Like This:
>>
>> +  motorcomm,rx-clk-driver-strength:
>> +    description: drive strength of rx_clk pad.
> 
> You need "description: |" to preserve the formatting if you add tables,
> but I don't think that this is a good idea. Put the values in here that
> describe the hardware (IOW the currents) and then you don't need to have
> this table.
> 
>> +      |----------------------|
>> +      | rx_clk ds map table  |
>> +      |----------------------|
>> +      | DS(3b) | Current (mA)|
>> +      |   000  |     1.20    |
>> +      |   001  |     2.10    |
>> +      |   010  |     2.70    |
>> +      |   011  |     2.91    |
>> +      |   100  |     3.11    |
>> +      |   101  |     3.60    |
>> +      |   110  |     3.97    |
>> +      |   111  |     4.35    |
>> +      |--------|-------------|
>> +    $ref: /schemas/types.yaml#/definitions/uint32
>> +    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
>> +    default: 3
>> +
> 
>> Or use minimum & maximum instead of these listed out enums
> 
> With the actual current values, enum rather than min + max.



Best regards,
Krzysztof
Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
Posted by Guo Samin 3 months, 1 week ago

-------- 原始信息 --------
主题: Re: [PATCH v3 1/2] dt-bindings: net: motorcomm: Add pad driver strength cfg
From: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
收件人: Conor Dooley <conor.dooley@microchip.com>, Guo Samin <samin.guo@starfivetech.com>
日期: 2023/6/20

> On 20/06/2023 10:26, Conor Dooley wrote:
>> Hey,
>>
>> On Tue, Jun 20, 2023 at 11:09:52AM +0800, Guo Samin wrote:
>>> From: Guo Samin <samin.guo@starfivetech.com>
>>>> From: Conor Dooley <conor@kernel.org>
>>>>> On Fri, May 26, 2023 at 05:05:01PM +0800, Samin Guo wrote:
>>>>>> The motorcomm phy (YT8531) supports the ability to adjust the drive
>>>>>> strength of the rx_clk/rx_data, the value range of pad driver
>>>>>> strength is 0 to 7.
>>
>>>>>> +  motorcomm,rx-clk-driver-strength:
>>>>>> +    description: drive strength of rx_clk pad.
>>>>>> +    $ref: /schemas/types.yaml#/definitions/uint32
>>>>>> +    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
>>>>>
>>>>> I think you should use minimum & maximum instead of these listed out
>>>>> enums.
>>
>>>>  You have also had this comment since v1 & were reminded of it on
>>>>> v2 by Krzysztof: "What do the numbers mean? What are the units? mA?"
>>
>>>> The good news is that we just got some data about units from Motorcomm. 
>>>> Maybe I can post the data show of the unit later after I get the complete data.
>>
>>> Sorry, haven't updated in a while.
>>
>> NW chief.
>>
>>> I just got the detailed data of Driver Strength(DS) from Motorcomm , which applies to both rx_clk and rx_data.
>>>
>>> |----------------------|
>>> |     ds map table     |
>>> |----------------------|
>>> | DS(3b) | Current (mA)|
>>> |--------|-------------|
>>> |   000  |     1.20    |
>>> |   001  |     2.10    |
>>> |   010  |     2.70    |
>>> |   011  |     2.91    |
>>> |   100  |     3.11    |
>>> |   101  |     3.60    |
>>> |   110  |     3.97    |
>>> |   111  |     4.35    |
>>> |--------|-------------|
>>>
>>> Since these currents are not integer values and have no regularity,
> 
> There is no mA unit in DT schema, so I don't see what by "not integer
> values". 1200 uA is an integer.
> 
>>> it is not very good to use in the drive/dts in my opinion.
>>
>> Who says you have to use mA? What about uA?
> 
> Yep

> 
>>
>>> Therefore, I tend to continue to use DS(0-7) in dts/driver, and adding
>>> a description of the current value corresponding to DS in dt-bindings. 
>>
>> I think this goes against not putting register values into the dts &
>> that the accurate description of the hardware are the currents.
> 
> For vendor properties register values are often accepted, but logical
> unit is much more readable in the DTS. Also allows further customization
> or extending when new variant appears. You cannot do extend a property
> easily when it holds a register value, without changing the meaning per
> variant.
>

Hi Conor & Krzysztof,
Thanks for taking the time to review.
Yes, uA can be used to avoid the problem of decimals.

In addition to this, we need to deal with the DS under different IOPAD.
these values were existed at IO 1.8V. When the IO is set to 3.3V, the values will change.

Chip_Config (EXT_0xA001)
|Bit  |Symbol  |Access  |Default  |Description |
|5:4  |Cfg_ldo |RW      |0x0      |Rgmii ldo voltage and RGMII/MDC/MDIO PAD's level shifter control. Depends on strapping.|
                                  |2'b11: 1.8v   2'b10: 1.8v    2'b01: 2.5v    2'b00: 3.3v                                |


      |----------------------|            
      | ds map table(1.8V)   |
      |----------------------|
      | DS(3b) | Current (mA)|
      |   000  |     1.20    |
      |   001  |     2.10    |
      |   010  |     2.70    |
      |   011  |     2.91    |
      |   100  |     3.11    |
      |   101  |     3.60    |
      |   110  |     3.97    |
      |   111  |     4.35    |
      |--------|-------------|


      |----------------------|
      | ds map table(3.3V)   |
      |----------------------|
      | DS(3b) | Current (mA)|
      |   000  |     3.07    |
      |   001  |     4.08    |
      |   010  |     4.37    |
      |   011  |     4.68    |
      |   100  |     5.02    |
      |   101  |     5.45    |
      |   110  |     5.74    |
      |   111  |     6.14    |
      |--------|-------------|


Best regards,
Samin
>>
>>> Like This:
>>>
>>> +  motorcomm,rx-clk-driver-strength:
>>> +    description: drive strength of rx_clk pad.
>>
>> You need "description: |" to preserve the formatting if you add tables,
>> but I don't think that this is a good idea. Put the values in here that
>> describe the hardware (IOW the currents) and then you don't need to have
>> this table.
>>
>>> +      |----------------------|
>>> +      | rx_clk ds map table  |
>>> +      |----------------------|
>>> +      | DS(3b) | Current (mA)|
>>> +      |   000  |     1.20    |
>>> +      |   001  |     2.10    |
>>> +      |   010  |     2.70    |
>>> +      |   011  |     2.91    |
>>> +      |   100  |     3.11    |
>>> +      |   101  |     3.60    |
>>> +      |   110  |     3.97    |
>>> +      |   111  |     4.35    |
>>> +      |--------|-------------|
>>> +    $ref: /schemas/types.yaml#/definitions/uint32
>>> +    enum: [ 0, 1, 2, 3, 4, 5, 6, 7 ]
>>> +    default: 3
>>> +
>>
>>> Or use minimum & maximum instead of these listed out enums
>>
>> With the actual current values, enum rather than min + max.
> 
> 
> 
> Best regards,
> Krzysztof
>