ASoC: Intel: sof_sdw: move controls and widgets addition to sof_sdw_rtd_init

[bardliao]

Currently, we add card controls and dapm widgets one by one in the codec_info->dais->rtd_init callback. Duplicated controls and dapm widgets will be added if there are more than one types of amps in the dai link. Moving it to sof_sdw_rtd_init() and only add the controls/widgets of the first codec dai can avoid the duplications. And add "spk:" tag to components string for each amp dai type.

[plbossart]

Argh, conflict! @bardliao can you rebase?

[bardliao]

@plbossart @ujfalusi On the second thought, I would like to address @ujfalusi's comment in this PR.

[bardliao]

@ujfalusi @plbossart rebased and updated.

[bardliao]

@bardliao, I think there is a potential to simplify things further with something like this?

diff --git a/sound/soc/intel/boards/sof_sdw_rt711.c b/sound/soc/intel/boards/sof_sdw_rt711.c
index 7e54fc5cbe09..763be6d42bd6 100644
--- a/sound/soc/intel/boards/sof_sdw_rt711.c
+++ b/sound/soc/intel/boards/sof_sdw_rt711.c
@@ -69,24 +69,15 @@ static struct snd_soc_jack_pin rt711_jack_pins[] = {
  },
 };

-static const char * const jack_codecs[] = {
- "rt711"
-};
-
-int rt711_rtd_init(struct snd_soc_pcm_runtime *rtd)
+int rt711_rtd_init(struct snd_soc_pcm_runtime *rtd, struct snd_soc_dai *dai)
 {
  struct snd_soc_card *card = rtd->card;
  struct mc_private *ctx = snd_soc_card_get_drvdata(card);
- struct snd_soc_dai *codec_dai;
  struct snd_soc_component *component;
  struct snd_soc_jack *jack;
  int ret;

- codec_dai = get_codec_dai_by_name(rtd, jack_codecs, ARRAY_SIZE(jack_codecs));
- if (!codec_dai)
-     return -EINVAL;
-
- component = codec_dai->component;
+ component = dai->component;
  card->components = devm_kasprintf(card->dev, GFP_KERNEL,
                    "%s hs:rt711",
                    card->components);

or perhaps this to the mix:

  card->components = devm_kasprintf(card->dev, GFP_KERNEL,
                    "%s hs:%s",
                    card->components, dai->name);

but I'm not sure if that is correct...

Yes, we don't need get_codec_dai_by_name() anymore. I plan to remove get_codec_dai_by_name() in another PR.