On Wed, 2015-11-25 at 08:51 -0600, Pierre-Louis Bossart wrote:
@@ -689,7 +689,13 @@ int snd_soc_new_compress(struct snd_soc_pcm_runtime *rtd, int num) compr->ops->copy = soc_compr_copy;
mutex_init(&compr->lock);
- ret = snd_compress_new(rtd->card->snd_card, num, direction, compr);
- snprintf(new_name, sizeof(new_name), "%s %s-%d",
rtd->dai_link->stream_name,
rtd->codec_dai->name, num);
Adding an ID with a human-readable name sounds good. Wondering though if the codec_dai name is relevant for compressed streams?
I think it is, that was the original reason I wrote this patch. I'm writing some automated tests where I have to know what each stream is really connected to and the DAI name is important for that (in my case anyway). Since the PCM core already provides that in the info file it seemed reasonable to do the same for compressed.
In all the configurations we have with DPCM the name is snd-soc-dummy-dai, there is no connection to the codec proper. stream_name and id should be good enough, no?
There's going to be systems where the DAI name is something real and I can't see a disadvantage in providing that extra piece of information, just in case it's useful. I've kept the format the same as the equivalent for PCM