On Fri, Jan 27, 2012 at 04:10:23PM +0100, Wolfram Sang wrote:

> Mark: Still no DAPM support. We don't have PM on this board, and since I
> couldn't easily figure what is expected from me here, I'd like to leave this
> for people who have real interest in that.

I don't understand what was unclear here...

You should at least use set_bias_level() to manage the shutdown GPIO I
guess.

> +static const struct snd_kcontrol_new max9768_volume[] = {
> +	SOC_SINGLE_TLV("Playback Volume", MAX9768_VOL, 0, 63, 0, volume_tlv),
> +};

> +static const struct snd_kcontrol_new max9768_mute[] = {
> +	SOC_SINGLE_BOOL_EXT("Mute Switch", 0, max9768_get_gpio, max9768_set_gpio),
> +};

Should be "Playback Switch" to match the volume control - that way
userspace can present a single control to applications.

> +static bool max9768_always_false(struct device *dev, unsigned int reg)
> +{
> +	return false;
> +}

This should be the default in regmap for formats that use format_write
(and hence can't read back), please fix in regmap rather than add this.  

> +	if (pdata) {
> +		/* Mute on powerup to avoid clicks */
> +		err = gpio_request_one(pdata->mute_gpio, GPIOF_INIT_HIGH, "MAX9768 Mute");

This need to avoid clicks is the reason I was asking for the DAPM stuff.

> +		/* Activate chip by releasing shutdown, enables I2C */
> +		err = gpio_request_one(pdata->shdn_gpio, GPIOF_INIT_HIGH, "MAX9768 Shutdown");
> +		max9768->shdn_gpio = err ?: pdata->shdn_gpio;

Eeew.