20 Jun
2010
20 Jun
'10
3:11 p.m.
On Sat, Jun 19, 2010 at 04:49:53PM +0200, Lars-Peter Clausen wrote:
This looks good, just one thing:
+#ifdef CONFIG_PM_SLEEP
+static int jz4740_codec_suspend(struct device *dev) +{
- struct jz4740_codec *jz4740_codec = dev_get_drvdata(dev);
- return jz4740_codec_set_bias_level(&jz4740_codec->codec,
SND_SOC_BIAS_OFF);+}
You've got these set up on the CODEC platform device rather than the ASoC CODEC. This means that the suspend and resume will happen out of sequence with the rest of the ASoC suspend and resume which could result in poor performance or bugs if the device is suspended while the core still thinks it's active. For example, ASoC will use DAPM to shut down the CODEC and it's possible that the CODEC could be suspended (and generate an audible noise) while an external amplifier is still powered, worsening the problem.