iio: adc: intel_mrfld_adc: Use be16_to_cpu() instead of get_unaligned_be16()

There is no need to call unaligned helpers on stack placed variables
because compiler will align them correctly, accordingly to architectural
ABI. Moreover, using bitwise type makes it explicit to see what we are
reading in bulk transfer. On top of that, use sizeof() instead of
magic value.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

drivers/iio/adc/intel_mrfld_adc.c

@@ -75,7 +75,7 @@ static int mrfld_adc_single_conv(struct iio_dev *indio_dev,
 	struct regmap *regmap = adc->regmap;
 	unsigned int req;
 	long timeout;
-	u8 buf[2];
+	__be16 value;
 	int ret;
 
 	reinit_completion(&adc->completion);
@@ -105,11 +105,11 @@ static int mrfld_adc_single_conv(struct iio_dev *indio_dev,
 		goto done;
 	}
 
-	ret = regmap_bulk_read(regmap, chan->address, buf, 2);
+	ret = regmap_bulk_read(regmap, chan->address, &value, sizeof(value));
 	if (ret)
 		goto done;
 
-	*result = get_unaligned_be16(buf);
+	*result = be16_to_cpu(value);
 	ret = IIO_VAL_INT;
 
 done: