Re: [RFC V2 1/2] ASoC: img: Add binding document for Pistachio audio card

[Damien Horsley]

On 27/01/16 16:00, Mark Brown wrote:
> On Wed, Jan 27, 2016 at 03:18:20PM +0000, Damien Horsley wrote:
>> On 27/01/16 14:57, Mark Brown wrote:
>>> On Tue, Jan 26, 2016 at 02:34:26PM +0000, Damien Horsley wrote:
> 
>>>> +  - clock-names : Includes the following entries:
>>>> +        "audio_pll"  The audio PLL
>>>> +        "i2s_mclk"   The i2s reference clock
>>>> +                     Also connected to i2s_out_0_mclk output
>>>> +        "dac_mclk"   Dac reference clock. Connected to i2s_dac_clk output
> 
>>> Why are these (especially the dac_mclk and i2s_mclk) properties of the
>>> card and not of the drivers for the respective devices?
> 
>> Due to the shared nature of these clocks. Individual components cannot
>> be responsible for controlling these as this could break configurations
>> for other components that are sharing the clocks. Only the card driver
>> has visibility of all of the components and their requirements.
> 
> You're talking about the code that decides what rates to set the clock
> at, not where the properties are placed in the DT.
> 
>> i2s_mclk and dac_mclk can be used by both the i2s in and i2s out paths
>> on some boards
> 
> Multiple devices can reference the same clock.
> 

audio_pll is referenced exclusively by the card device

i2s_mclk and dac_mclk can also be referenced by other devices. The i2s
out controller references i2s_mclk, and codec devices can reference
i2s_mclk/dac_mclk dependent on their system clock requirements

without a reference to i2s_mclk and dac_mclk in the card driver, it
would not be possible to control the divisors and gates for these clocks
in the following cases:

Simplistic codecs that do not have drivers

Codec drivers that do not call clk_set_rate and clk_enable/clk_disable