On Thu, Nov 26, 2009 at 06:59:00PM +0100, Daniel Mack wrote:
+static struct regulator_consumer_supply audio_va_consumer_supply =
- REGULATOR_SUPPLY("va", "soc-audio.0");
I hadn't spotted this in your cs4270 patches but the device used should be the I2C device, not the soc-audio device. I'd also expect to see this regulator setup done in arch/arm - the stuff I'd expect to see in here is the resource claiming stuff. Having it here is also fine, it's just a little surprising.
+static int raumfeld_ak4104_init(struct snd_soc_codec *codec) +{
- return 0;
+}
Remove empty functions like this (I think this is the only one that's left).
+static int __init raumfeld_audio_init(void) +{
- int ret;
- if (!machine_is_raumfeld_speaker() &&
!machine_is_raumfeld_connector())return 0;- max9486_client = i2c_new_device(i2c_get_adapter(0),
&max9486_hwmon_info);- if (!max9486_client)
return -ENOMEM;
I'd still really prefer to see this as a regular I2C device driver with the code in arch/arm (even if the driver itself ends up embedded in this file, which is what GTA01 did for a similar situation) but I guess I'm OK. max9486_hwmon_info could be renamed, though.
platform_add_devices(regulator_devices, ARRAY_SIZE(regulator_devices));
Nothing ever reverses this (one of the reasons to shove this bit into arch/arm :) ).