On Tue, 2012-02-07 at 17:52 +0530, Ashish Jangam wrote:

Can you consider adding yourself or David as maintainer ? Not a must, but no one else would really be able to support the driver.

Thank for this, you can add my name.

...

+Supported chips:

• • Dialog Semiconductors DA9052-BC and DA9053-AA/Bx PMICs
• Prefix: 'da9052'
• Datasheet: Kindly visit www.dialog-semiconductor.com and request for the

•           official datasheet.
  

Maybe, but that doesn't seem to be straightforward. The web site has no search field, and I am not sure who to contact in the contact list. And since there does not seem to be support in the US, I am not sure if one is expected to contact someone in Europe or Taiwan or some other place.

Maybe you can let me know how exactly to get the datasheets - I like to keep those around for chips we are supporting.

You need to request for DA9052/53 datasheet from www.dialog-semiconductor.com/DA9053.php

...

+static ssize_t da9052_read_tjunc(struct device *dev,

•                            struct device_attribute *devattr, char *buf)
  

+{

•   struct da9052_hwmon *hwmon = dev_get_drvdata(dev);
  

•   int tjunc, toffset, result;
  

•   tjunc = da9052_reg_read(hwmon->da9052, DA9052_TJUNC_RES_REG);
  

•   if (tjunc < 0)
  

•           return tjunc;
  

•   toffset = da9052_reg_read(hwmon->da9052, DA9052_T_OFFSET_REG);
  

•   if (toffset < 0)
  

•           return toffset;
  

•   if ((tjunc - toffset) == 0)
  

•           result = 0;
  

•   else
  

•           /* Degrees celsius = 1.708 * (TJUNC_RES - T_OFFSET) - 108.8 */
  

•           result = 1708 * (tjunc - toffset) - 108800;
  

Hmm. tjunc = 1, toffset = 2: return 1708 * (-1) - 108800 = -108800 - 1708 = -110508 tjunc = 2, toffset = 2: return 0 tjunc = 3, toffset = 2: return 1780 * 1 - 108800 = -108800 + 1708 = -107092

I am not saying this is wrong, not having access to the data sheet, but it is a bit odd. Why is the 0-difference case handled differently, and why does it return 0 and not, say, -108800 ?

If tjunc - toffset < 0 is invalid (no idea if it is, of course), it might make sense to detect that condition.

toffset is the trim value used for improving the calculated temperature accuracy so considering this tjunc - toffset <=0 seems to be not applicable cases. I will correct this along with your other comments and will re-submit the patch.