Re: [PATCH v2 3/3] ASoC: tlv320adcx140: Add proper support for master mode

[Camel Guo <camelg@axis.com>]

Please forget about this patch since Dan will upload a similar one.

@Dan, see my comment below.

On 9/15/20 8:50 PM, Dan Murphy wrote:
> Camel
> 
> On 9/11/20 3:07 AM, Camel Guo wrote:
>> From: Camel Guo <camelg@axis.com>
>>
>> Add setup of bclk-to-ws ratio and sample rate when in master mode,
>> as well as MCLK input pin setup.
>>
>> Signed-off-by: Camel Guo <camelg@axis.com>
>> ---
>>   v2:
>>    - Move GPIO setting into devicetree
>>    - Move master config register setting into a new function
>>
>>   sound/soc/codecs/tlv320adcx140.c | 139 ++++++++++++++++++++++++++++++-
>>   sound/soc/codecs/tlv320adcx140.h |  27 ++++++
>>   2 files changed, 162 insertions(+), 4 deletions(-)
>>
>> diff --git a/sound/soc/codecs/tlv320adcx140.c b/sound/soc/codecs/tlv320adcx140.c
>> index 97f16fbba441..685f5fd8b537 100644
>> --- a/sound/soc/codecs/tlv320adcx140.c
>> +++ b/sound/soc/codecs/tlv320adcx140.c
>> @@ -35,6 +35,7 @@ struct adcx140_priv {
>>        unsigned int dai_fmt;
>>        unsigned int tdm_delay;
>>        unsigned int slot_width;
>> +     bool master;
>>   };
>>   
>>   static const char * const gpo_config_names[] = {
>> @@ -651,11 +652,136 @@ static int adcx140_reset(struct adcx140_priv *adcx140)
>>        return ret;
>>   }
>>   
>> +static int adcx140_fs_bclk_ratio(unsigned int bclk_ratio)
>> +{
>> +     switch (bclk_ratio) {
>> +     case 16:
>> +             return ADCX140_RATIO_16;
>> +     case 24:
>> +             return ADCX140_RATIO_24;
>> +     case 32:
>> +             return ADCX140_RATIO_32;
>> +     case 48:
>> +             return ADCX140_RATIO_48;
>> +     case 64:
>> +             return ADCX140_RATIO_64;
>> +     case 96:
>> +             return ADCX140_RATIO_96;
>> +     case 128:
>> +             return ADCX140_RATIO_128;
>> +     case 192:
>> +             return ADCX140_RATIO_192;
>> +     case 256:
>> +             return ADCX140_RATIO_256;
>> +     case 384:
>> +             return ADCX140_RATIO_384;
>> +     case 512:
>> +             return ADCX140_RATIO_512;
>> +     case 1024:
>> +             return ADCX140_RATIO_1024;
>> +     case 2048:
>> +             return ADCX140_RATIO_2048;
>> +     default:
>> +             break;
>> +     }
>> +     return -EINVAL;
>> +}
>> +
>> +static int adcx140_fs_rate(unsigned int rate)
>> +{
>> +     switch (rate) {
>> +     case 7350:
>> +     case 8000:
>> +             return ADCX140_8_OR_7_35KHZ;
>> +     case 14700:
>> +     case 16000:
>> +             return ADCX140_16_OR_14_7KHZ;
>> +     case 22050:
>> +     case 24000:
>> +             return ADCX140_24_OR_22_05KHZ;
>> +     case 29400:
>> +     case 32000:
>> +             return ADCX140_32_OR_29_4KHZ;
>> +     case 44100:
>> +     case 48000:
>> +             return ADCX140_48_OR_44_1KHZ;
>> +     case 88200:
>> +     case 96000:
>> +             return ADCX140_96_OR_88_2KHZ;
>> +     case 176400:
>> +     case 192000:
>> +             return ADCX140_192_OR_176_4KHZ;
>> +     case 352800:
>> +     case 384000:
>> +             return ADCX140_384_OR_352_8KHZ;
>> +     case 705600:
>> +     case 768000:
>> +             return ADCX140_768_OR_705_6KHZ;
>> +     default:
>> +             break;
>> +     }
>> +     return -EINVAL;
>> +}
>> +
>> +static int adcx140_setup_master_config(struct snd_soc_component *component,
>> +                                    struct snd_pcm_hw_params *params)
>> +{
>> +     int ret = 0;
>> +     struct adcx140_priv *adcx140 = snd_soc_component_get_drvdata(component);
>> +
>> +     if (adcx140->master) {
> 
> Move this out to hw_params.  No reason to jump here just to jump back.
> 
> Check for master and if master then configure
> 
> This will eliminate the mixed code and variable declaration below which
> is what I wanted to avoid in v1.
> 
> This will also allow you to remove some of the variable initialization.
> 
>> +             u8 mst_cfg1 = 0;
>> +             u8 mst_cfg0 = 0;
> This can be init to mst_cfg0 = ADCX140_BCLK_FSYNC_MASTER no reason to
> set it here and then change it immediately.
>> +             unsigned int bclk_ratio;
>> +
>> +             mst_cfg0 = ADCX140_BCLK_FSYNC_MASTER;
>> +             if (params_rate(params) % 1000)
>> +                     mst_cfg0 |= ADCX140_FSYNCINV_BIT; /* 44.1 kHz et al */
>> +
>> +             ret = adcx140_fs_rate(params_rate(params));
>> +             if (ret < 0) {
>> +                     dev_err(adcx140->dev, "%s: Unsupported rate %d\n",
>> +                                     __func__, params_rate(params));
>> +                     return ret;
>> +             }
>> +             mst_cfg1 |= ret;
> Why the | here?  This is initialized to 0 so mst_cfg1 = ret. And why
> even use ret just return into mst_cfg1 and check that variable
>> +
>> +             /* In slave mode when using automatic clock configuration,
>> +              * the codec figures out the BCLK to FSYNC ratio itself. But
>> +              * here in master mode, we need to tell it.
>> +              */
>> +
>> +             bclk_ratio = snd_soc_params_to_frame_size(params);
>> +             ret = adcx140_fs_bclk_ratio(bclk_ratio);
>> +             if (ret < 0) {
>> +                     dev_err(adcx140->dev, "%s: Unsupported bclk_ratio %d\n",
>> +                                     __func__, bclk_ratio);
>> +                     return ret;
>> +             }
>> +             mst_cfg1 |= ret;
>> +
>> +             snd_soc_component_update_bits(component, ADCX140_MST_CFG1,
>> +                             ADCX140_FS_RATE_MSK |
>> +                             ADCX140_RATIO_MSK,
>> +                             mst_cfg1);
> 
> I don't understand the update_bits since you have calcualted both the
> Ratio and rate you can just write the register with the
> snd_soc_component_write.
> 
>> +
>> +             snd_soc_component_update_bits(component, ADCX140_MST_CFG0,
>> +                             ADCX140_FSYNCINV_BIT |
>> +                             ADCX140_BCLK_FSYNC_MASTER,
>> +                             mst_cfg0);
>> +
> 
> But this is ok.  I actually have other changes I am posting which move
> this to the set_dai_format.  So I am not sure if this will be needed
> after that patch is applied.

Maybe I can just wait for your patch. If it can make master mode work, 
there is no need for me to work on this one.

> 
> I will CC you on those patches.
> 
> Dan