Re: [PATCH] ASoC: core - Fix dai usage in pcm_new() platform handlers

[Mark Brown <broonie@opensource.wolfsonmicro.com>]

On Wed, Jun 22, 2011 at 08:05:29PM +0200, Daniel Mack wrote:
> Commit 552d1ef ("ASoC: core - Optimise and refactor pcm_new() to pass
> only rtd") refactored the ASoC internal core to only pass a pointer
> to a runtime as argument to individual pcm_new() functions. However, the
> new code that was added to the handlers to get access to the dai now
> uses rtd->cpu_dai, while before this patch, rtd->codec_dai was used.

Which is a problem because...

> diff --git a/sound/soc/atmel/atmel-pcm.c b/sound/soc/atmel/atmel-pcm.c
> index f81d4c3..d1140c1 100644
> --- a/sound/soc/atmel/atmel-pcm.c
> +++ b/sound/soc/atmel/atmel-pcm.c
> @@ -367,7 +367,7 @@ static u64 atmel_pcm_dmamask = 0xffffffff;
>  static int atmel_pcm_new(struct snd_soc_pcm_runtime *rtd)
>  {
>  	struct snd_card *card = rtd->card->snd_card;
> -	struct snd_soc_dai *dai = rtd->cpu_dai;
> +	struct snd_soc_dai *dai = rtd->codec_dai;
>  	struct snd_pcm *pcm = rtd->pcm;
>  	int ret = 0;
>  

The original code doesn't look obviously wrong while if the CPU driver
is peering at the CODEC driver that smells very bad.  What's the actual
problem here?