On Thursday 29 January 2009 15:54:34 ext Mark Brown wrote:
On Thu, Jan 29, 2009 at 02:57:49PM +0200, Peter Ujfalusi wrote:
With the this implementation, there is no visible difference in power consumption compared to the previous, codec off way. So when the codec is in SND_SOC_BIAS_STANDBY mode and all of the bypasses are disabled, the additional power consumption is minimal, if any.
It isn't ideal since what should happen is that the bias is kept on while any paths are active (not just digital ones) but it shouldn't cause any problems since all that should happen is that the checks for bypass paths being active will just always return false. I'll the patch - when the core becomes more capable the patch can always be revisited.
You mean, that if one of the bypass is active, and the core wants to set the codec to SND_SOC_BIAS_STANDBY, that the twl4030_set_bias_level should return with false?
Also, if the codec was in SND_SOC_BIAS_STANDBY and one of the bypass gots enabled, than it should set the codec bias_state to SND_SOC_BIAS_ON?
So the STANDBY state is really STANDBY?