Re: [PATCH 02/12] ASoC: dapm: Implement stereo mixer control support

[Mark Brown <broonie-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>]

[-- Attachment #1: Type: text/plain, Size: 1292 bytes --]

On Mon, Oct 03, 2016 at 07:07:54PM +0800, Chen-Yu Tsai wrote:

>  	/* find dapm widget path assoc with kcontrol */
>  	dapm_kcontrol_for_each_path(path, kcontrol) {
> +		/*
> +		 * If status for the second channel was given ( >= 0 ),
> +		 * consider the second and later paths as the second
> +		 * channel.
> +		 */
> +		if (found && rconnect >= 0)
> +			soc_dapm_connect_path(path, rconnect, "mixer update");
> +		else
> +			soc_dapm_connect_path(path, connect, "mixer update");
>  		found = 1;
> -		soc_dapm_connect_path(path, connect, "mixer update");

This really only works for two channels with the current inteface - the
comment makes it sound like it'll work for more but we can only pass in
two (and there's only support for specifying two everywhere).

> -	change = dapm_kcontrol_set_value(kcontrol, val);
> +	/* This assumes field width < (bits in unsigned int / 2) */
> +	change = dapm_kcontrol_set_value(kcontrol, val | (rval << width));

That seems like a bit of an assumption in cases where we've got a single
control for both power and volume?  They're very rare though, I'm not
even sure any exist.  It'd be good to have a check in the code just in
case it does come up, it'll likely be a nightmare to debug if someone
does run into it.

Otherwise I think this makes sense.