iio: magnetometer: mmc35240: Fix broken processed value

The current computation for fractional part of the magnetic
field is broken. This patch fixes it by taking a different
approach. We expose the raw reading in milli Gauss (to avoid
rounding errors) with a scale of 0.001.

Thus the final computation is done in userspace where floating
point operation are more relaxed.

Fixes: abeb6b1e ("iio: magnetometer: Add support for MEMSIC MMC35240")
Signed-off-by: Daniel Baluta <daniel.baluta@intel.com>
Signed-off-by: Jonathan Cameron <jic23@kernel.org>

drivers/iio/magnetometer/mmc35240.c

@@ -113,8 +113,9 @@ static IIO_CONST_ATTR_SAMP_FREQ_AVAIL("100 200 333 666");
 	.modified = 1, \
 	.channel2 = IIO_MOD_ ## _axis, \
 	.address = AXIS_ ## _axis, \
-	.info_mask_separate = BIT(IIO_CHAN_INFO_PROCESSED), \
-	.info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SAMP_FREQ), \
+	.info_mask_separate = BIT(IIO_CHAN_INFO_RAW), \
+	.info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SAMP_FREQ) | \
+			BIT(IIO_CHAN_INFO_SCALE), \
 }
 
 static const struct iio_chan_spec mmc35240_channels[] = {
@@ -241,9 +242,19 @@ static int mmc35240_read_measurement(struct mmc35240_data *data, __le16 buf[3])
 				3 * sizeof(__le16));
 }
 
-static int mmc35240_raw_to_gauss(struct mmc35240_data *data, int index,
-				 __le16 buf[],
-				 int *val, int *val2)
+/**
+ * mmc35240_raw_to_mgauss - convert raw readings to milli gauss. Also apply
+			    compensation for output value.
+ *
+ * @data: device private data
+ * @index: axis index for which we want the conversion
+ * @buf: raw data to be converted, 2 bytes in little endian format
+ * @val: compensated output reading (unit is milli gauss)
+ *
+ * Returns: 0 in case of success, -EINVAL when @index is not valid
+ */
+static int mmc35240_raw_to_mgauss(struct mmc35240_data *data, int index,
+				  __le16 buf[], int *val)
 {
 	int raw_x, raw_y, raw_z;
 	int sens_x, sens_y, sens_z;
@@ -261,18 +272,15 @@ static int mmc35240_raw_to_gauss(struct mmc35240_data *data, int index,
 
 	switch (index) {
 	case AXIS_X:
-		*val = (raw_x - nfo) / sens_x;
-		*val2 = ((raw_x - nfo) % sens_x) * 1000000;
+		*val = (raw_x - nfo) * 1000 / sens_x;
 		break;
 	case AXIS_Y:
-		*val = (raw_y - nfo) / sens_y - (raw_z - nfo) / sens_z;
-		*val2 = (((raw_y - nfo) % sens_y - (raw_z - nfo) % sens_z))
-			* 1000000;
+		*val = (raw_y - nfo) * 1000 / sens_y -
+			(raw_z - nfo)  * 1000 / sens_z;
 		break;
 	case AXIS_Z:
-		*val = (raw_y - nfo) / sens_y + (raw_z - nfo) / sens_z;
-		*val2 = (((raw_y - nfo) % sens_y + (raw_z - nfo) % sens_z))
-			* 1000000;
+		*val = (raw_y - nfo) * 1000 / sens_y +
+			(raw_z - nfo) * 1000 / sens_z;
 		break;
 	default:
 		return -EINVAL;
@@ -290,16 +298,19 @@ static int mmc35240_read_raw(struct iio_dev *indio_dev,
 	__le16 buf[3];
 
 	switch (mask) {
-	case IIO_CHAN_INFO_PROCESSED:
+	case IIO_CHAN_INFO_RAW:
 		mutex_lock(&data->mutex);
 		ret = mmc35240_read_measurement(data, buf);
 		mutex_unlock(&data->mutex);
 		if (ret < 0)
 			return ret;
-		ret = mmc35240_raw_to_gauss(data, chan->address,
-					    buf, val, val2);
+		ret = mmc35240_raw_to_mgauss(data, chan->address, buf, val);
 		if (ret < 0)
 			return ret;
+		return IIO_VAL_INT;
+	case IIO_CHAN_INFO_SCALE:
+		*val = 0;
+		*val2 = 1000;
 		return IIO_VAL_INT_PLUS_MICRO;
 	case IIO_CHAN_INFO_SAMP_FREQ:
 		mutex_lock(&data->mutex);