Hello,

On Sun, Jun 19, 2011 at 12:06 AM, Mark Brown
[off-list ref] wrote:

On Thu, Jun 16, 2011 at 05:30:02PM +0900, MyungJoo Ham wrote:

quoted

+ ? ? adc->vdd = regulator_get(dev, S3C_ADC_REGULATOR_NAME);

I'm not convinced that the #define for the name is terribly good style
here, especially given that you actually call it vdd in the code.

Then, would it be fine to use as [ regulator_get(dev, "vdd"); ] ?

quoted

+ ? ? if (IS_ERR_OR_NULL(adc->vdd)) {
+ ? ? ? ? ? ? dev_dbg(dev, "operating without regulator %s.\n", S3C_ADC_REGULATOR_NAME);
+ ? ? ? ? ? ? adc->vdd = NULL; /* Do not control regulator */
+ ? ? }
+

No, don't do this. ?Just unconditionally assume the regulator is present
if power is essential for use of the device. ?The regulator API will
stub out correctly if it's not in use to allow things to proceed and if
vdd is genuinely not hooked up then the driver can't function.

This ADC driver is for every ADC from S3C24xx series to Exynos4 (and
its successors as well).
The regulator (VDD for ADC) is essential for the recent chips
(S5PC110, S5PV210, and Exynos4).
I was just worried about the old boards using the same ADC driver
(mach-s3c2410/mach-*.c, mach-s3c6410/mach-*.c, and so on) without
ADC-VDD regulators defined.

However, no s3c compliance defconfigs have ever used CONFIG_REGULATOR.
Thus, it seems that it's safe to enforce using "vdd" with regulators
in plat-samsung's ADC driver.
I'll proceed as you have commented.

quoted

+ ? ? if (adc->vdd)
+ ? ? ? ? ? ? regulator_enable(adc->vdd);

You're not checking the return value here or anywhere else after the
inital get().

Ok. I'll let it handle errors from regulator_enable.


Thank you!

- MyungJoo.
-- 
MyungJoo Ham (???), Ph.D.
Mobile Software Platform Lab,
Digital Media and Communications (DMC) Business
Samsung Electronics
cell: 82-10-6714-2858