On Thu, Aug 13, 2009 at 5:08 PM, Mark Brownbroonie@opensource.wolfsonmicro.com wrote:
On Thu, Aug 13, 2009 at 11:41:13AM +0800, Barry Song wrote:
On Wed, Aug 12, 2009 at 9:46 PM, Mark Brownbroonie@opensource.wolfsonmicro.com wrote:
Just "ad1836" - the fact that it's controlled via SPI is clear from the fact that this is a struct spi_driver.
I agree the spi structure implies the meaning of spi. But files in arch/blackfin/mach... are using names like "xxx-spi" for almost all spi devices. How about keeping the coherence?
Oh dear. Looks like you've essentially got a choice between being consistent with the rest of the world and being consistent with what the blackfin stuff is doing.
Looks like the rest world really doesn't define a spi device with modalias "xxx-spi" but just "xxx" in arch. So I agree we can change blackfin stuff as Mike.
The power control will be handled by DAPM but your driver still needs to restore things like the volume settings - when the driver is powered off the register settings will be forgotten.
My test shows registers setting doesn't lose after dac/adc shutdown :-)
Probably your board is maintaining power to the chip even during suspend.
If so, all codec drivers need to save setting then restore them when resume. The physical shutdown depends on boards great. If we get an individual programmable PMU/LDO to control the voltage input to this chip, I can shutdown power-input to codec physically and save the registers when suspend, then restore them when resume. If no this unit, should every codec driver handle this case?