Re: [PATCH v5 2/6] ASoC: sgtl5000: Improve VAG power and mute control

[Cezary Rojewski]

On 2019-07-18 20:42, Cezary Rojewski wrote:
> On 2019-07-18 11:02, Oleksandr Suvorov wrote:
> > +enum {
> > +ÂÂÂ HP_POWER_EVENT,
> > +ÂÂÂ DAC_POWER_EVENT,
> > +ÂÂÂ ADC_POWER_EVENT,
> > +ÂÂÂ LAST_POWER_EVENT
> > +};
> > +
> > +static u16 mute_mask[] = {
> > +ÂÂÂ SGTL5000_HP_MUTE,
> > +ÂÂÂ SGTL5000_OUTPUTS_MUTE,
> > +ÂÂÂ SGTL5000_OUTPUTS_MUTE
> > +};
>
> If mute_mask[] is only used within common handler, you may consider 
> declaring const array within said handler instead (did not check that 
> myself).
> Otherwise, simple comment for the second _OUTPUTS_MUTE should suffice - 
> its not self explanatory why you doubled that mask.
>
> > +
> > Â /* sgtl5000 private structure in codec */
> > Â struct sgtl5000_priv {
> > ÂÂÂÂÂ int sysclk;ÂÂÂ /* sysclk rate */
> > @@ -137,8 +157,109 @@ struct sgtl5000_priv {
> > ÂÂÂÂÂ u8 micbias_voltage;
> > ÂÂÂÂÂ u8 lrclk_strength;
> > ÂÂÂÂÂ u8 sclk_strength;
> > +ÂÂÂ u16 mute_state[LAST_POWER_EVENT];
> > Â };
>
> When I spoke of LAST enum constant, I did not really had this specific 
> usage in mind.
>
>  From design perspective, _LAST_ does not exist and should never be 
> referred to as "the next option" i.e.: new enum constant.
> That is way preferred usage is:
> u16 mute_state[ADC_POWER_EVENT+1;
> -or-
> u16 mute_state[LAST_POWER_EVENT+1];
>
> Maybe I'm just being radical here :)
>
> Czarek

Forgive me for double posting. Comment above is targeted towards:

>> +enum {
>> +    HP_POWER_EVENT,
>> +    DAC_POWER_EVENT,
>> +    ADC_POWER_EVENT,
>> +    LAST_POWER_EVENT
>> +};

as LAST_POWER_EVENT is not assigned explicitly to ADC_POWER_EVENT and 
thus generates implicit "new option" of value 3.

Czarek