On 22 May 2014 16:00, Mark Brown <broonie@xxxxxxxxxx> wrote: > On Thu, May 22, 2014 at 02:47:07PM +0530, Tushar Behera wrote: > >> + max98090->mclk = devm_clk_get(codec->dev, "mclk"); >> + if (!IS_ERR(max98090->mclk)) >> + clk_prepare_enable(max98090->mclk); >> + > > Ths doesn't handle deferred probe, we need to at least return an error > if we get -EPROBE_DEFER. > Ok. > It'd also be better to move the enabling to set_bias_level() if possible > (I don't know if the clock is needed for register access) though less > essential. I tested with moving clk_enable/clk_disable calls to set_bias_level(): SND_SOC_BIAS_PREPARE. That works well for me. Does it look okay? @@ -1801,6 +1801,25 @@ static int max98090_set_bias_level(struct snd_soc_codec *codec, break; case SND_SOC_BIAS_PREPARE: + /* + * SND_SOC_BIAS_PREPARE is called while preparing for a + * transition to ON or away from ON. If current bias_level + * is SND_SOC_BIAS_ON, then it is preparing for a transition + * away from ON. Disable the clock in that case, otherwise + * enable it. + */ + if (!IS_ERR(max98090->mclk)) { + if (codec->dapm.bias_level == SND_SOC_BIAS_ON) + clk_disable(max98090->mclk); + else + clk_enable(max98090->mclk); + } break; -- Tushar Behera -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html