On Apr 18, 2014, at 4:23 AM, Thomas Petazzoni thomas.petazzoni@free-electrons.com wrote:

Dear Brian Austin,

On Thu, 17 Apr 2014 10:08:23 -0500, Brian Austin wrote:

...

can you try this quick check please?

static const struct snd_kcontrol_new cs42l51_snd_controls[] = { SOC_DOUBLE_R_SX_TLV("PCM Playback Volume", CS42L51_PCMA_VOL, CS42L51_PCMB_VOL,

•                   0, 0x19, 0x7F, adc_pcm_tlv),
  

•                   0, 0x19, 0x80, adc_pcm_tlv),
    SOC_DOUBLE_R("PCM Playback Switch",
                    CS42L51_PCMA_VOL, CS42L51_PCMB_VOL, 7, 1, 1),
    SOC_DOUBLE_R_SX_TLV("Analog Playback Volume",
                    CS42L51_AOUTA_VOL, CS42L51_AOUTB_VOL,
  

•                   0, 0x34, 0xE4, aout_tlv),
  

•                   0, 0x34, 0xE5, aout_tlv),
    SOC_DOUBLE_R_SX_TLV("ADC Mixer Volume",
                    CS42L51_ADCA_VOL, CS42L51_ADCB_VOL,
  

•                   0, 0x19, 0x7F, adc_pcm_tlv),
  

•                   0, 0x19, 0x80, adc_pcm_tlv),
    SOC_DOUBLE_R("ADC Mixer Switch",
                    CS42L51_ADCA_VOL, CS42L51_ADCB_VOL, 7, 1, 1),
    SOC_SINGLE("Playback Deemphasis Switch", CS42L51_DAC_CTL, 3, 1, 
  

0),

Unfortunately, it still doesn't work: at boot time the PCM and ADC mixer values are 100 which looks OK, but the Analog value is 170, which doesn't look ok. And inspecting the values set at the register level for the PCM control seems to reveal that things are still not working correctly (read on below).

But this time around, I've collected a lot of debugging informations, which hopefully will help understand the problem. My understanding is that:

SOC_DOUBLE_R_SX_TLV("Analog Playback Volume", CS42L51_AOUTA_VOL, CS42L51_AOUTB_VOL, 0, 0x34, 0xE5, aout_tlv),

cannot work, because it assumes that the range -102.0 dB to 12.0 dB described in the datasheet is represented by using register values from 0x34 to 0xE5. Which misses the fact that when going from -102.0 dB to 0 dB, the register value goes from 0x34 to 0, and then when going from 0 to 12.0 dB, the register value should go from 0 to 0xE5.

Ok, Thanks for the output. Something is for sure broken, but I could have sworn I checked this. I’ll hook up my L51 board today and see what the deal is.

Thanks, Brian