Re: [PATCH] doc: devicetree: Add bindings documentation for omap-des driver

[Joel Fernandes <joelf@ti.com>]

On 11/11/2013 05:01 AM, Mark Rutland wrote:
> Hi Joel,
> 
> I realise I'm a little late in replying to this, but there are a few
> things that would be nice to fix up.
> 
> On Fri, Nov 08, 2013 at 12:37:09AM +0000, Joel Fernandes wrote:
>> Add documentation for the generic OMAP DES crypto module describing the device
>> tree bindings.
>>
>> Signed-off-by: Joel Fernandes <joelf@ti.com>
>> ---
>>  .../devicetree/bindings/crypto/omap-des.txt        | 28 ++++++++++++++++++++++
>>  1 file changed, 28 insertions(+)
>>  create mode 100644 Documentation/devicetree/bindings/crypto/omap-des.txt
>>
>> diff --git a/Documentation/devicetree/bindings/crypto/omap-des.txt b/Documentation/devicetree/bindings/crypto/omap-des.txt
>> new file mode 100644
>> index 0000000..0637647
>> --- /dev/null
>> +++ b/Documentation/devicetree/bindings/crypto/omap-des.txt
>> @@ -0,0 +1,28 @@
>> +OMAP SoC DES crypto Module
>> +
>> +Required properties:
>> +
>> +- compatible : Should be "ti,omap4-des"
> 
> Nit: s/Should be/Should contain/

ok

> 
>> +- ti,hwmods: Name of the hwmod associated with the DES module
>> +- reg : Offset and length of the register set for the module
>> +- interrupts : the interrupt-specifier for the DES module
> 
> Just to check: there's only one interrupt?

Yes.

>> +- clocks : phandle to the functional clock node of the DES module
>> +- clock-names : Name of the functional clock
> 
> What name is expected here? The intent of clock-names is to describe
> which clock input lines the clocks are wired to, so this should be
> well-defined.
> 
> From the looks of the example, the clock input is called "fck". Is that
> correct?

Yes. The clock name is used by PM code to get the clock for the particular
device. Yes the code assumes it to be "fck", I can change doc to be so.

> When you have clock-names, the nicest thing to do is to define clocks
> in terms of clock-names. Something like:
> 
> - clocks: A phandle + clock-specifier pair for each entry in
>   clock-names.
> 
> - clock-names: Should contain:
>   * "fck" for the functional clock
> 
> Which implies that the correct way to find clocks is to look in
> clock-names first to find the clock's index, rather than grabbing the
> clock by index. This makes it far easier to add/remove/change clocks in
> future.

Ok. I'll reword documentation of clocks as above.

> 
>> +
>> +Optional properties:
>> +- dmas: DMA specifiers for tx and rx dma. See the DMA client binding,
>> +	Documentation/devicetree/bindings/dma/dma.txt
> 
> Similarly here it would be nice to have dmas refer to dma-names for
> the description of which dmas exist.

Ok, will make the above 3 changes and resubmit, if its ok will add your Reviewed-by.

thanks,

-Joel