|
@@ -80,16 +80,12 @@ struct bmp280_data {
|
|
|
s32 t_fine;
|
|
|
};
|
|
|
|
|
|
-/* Compensation parameters. */
|
|
|
-struct bmp280_comp_temp {
|
|
|
- u16 dig_t1;
|
|
|
- s16 dig_t2, dig_t3;
|
|
|
-};
|
|
|
-
|
|
|
-struct bmp280_comp_press {
|
|
|
- u16 dig_p1;
|
|
|
- s16 dig_p2, dig_p3, dig_p4, dig_p5, dig_p6, dig_p7, dig_p8, dig_p9;
|
|
|
-};
|
|
|
+/*
|
|
|
+ * These enums are used for indexing into the array of compensation
|
|
|
+ * parameters.
|
|
|
+ */
|
|
|
+enum { T1, T2, T3 };
|
|
|
+enum { P1, P2, P3, P4, P5, P6, P7, P8, P9 };
|
|
|
|
|
|
static const struct iio_chan_spec bmp280_channels[] = {
|
|
|
{
|
|
@@ -141,54 +137,6 @@ static const struct regmap_config bmp280_regmap_config = {
|
|
|
.volatile_reg = bmp280_is_volatile_reg,
|
|
|
};
|
|
|
|
|
|
-static int bmp280_read_compensation_temp(struct bmp280_data *data,
|
|
|
- struct bmp280_comp_temp *comp)
|
|
|
-{
|
|
|
- int ret;
|
|
|
- __le16 buf[BMP280_COMP_TEMP_REG_COUNT / 2];
|
|
|
-
|
|
|
- ret = regmap_bulk_read(data->regmap, BMP280_REG_COMP_TEMP_START,
|
|
|
- buf, BMP280_COMP_TEMP_REG_COUNT);
|
|
|
- if (ret < 0) {
|
|
|
- dev_err(&data->client->dev,
|
|
|
- "failed to read temperature calibration parameters\n");
|
|
|
- return ret;
|
|
|
- }
|
|
|
-
|
|
|
- comp->dig_t1 = (u16) le16_to_cpu(buf[0]);
|
|
|
- comp->dig_t2 = (s16) le16_to_cpu(buf[1]);
|
|
|
- comp->dig_t3 = (s16) le16_to_cpu(buf[2]);
|
|
|
-
|
|
|
- return 0;
|
|
|
-}
|
|
|
-
|
|
|
-static int bmp280_read_compensation_press(struct bmp280_data *data,
|
|
|
- struct bmp280_comp_press *comp)
|
|
|
-{
|
|
|
- int ret;
|
|
|
- __le16 buf[BMP280_COMP_PRESS_REG_COUNT / 2];
|
|
|
-
|
|
|
- ret = regmap_bulk_read(data->regmap, BMP280_REG_COMP_PRESS_START,
|
|
|
- buf, BMP280_COMP_PRESS_REG_COUNT);
|
|
|
- if (ret < 0) {
|
|
|
- dev_err(&data->client->dev,
|
|
|
- "failed to read pressure calibration parameters\n");
|
|
|
- return ret;
|
|
|
- }
|
|
|
-
|
|
|
- comp->dig_p1 = (u16) le16_to_cpu(buf[0]);
|
|
|
- comp->dig_p2 = (s16) le16_to_cpu(buf[1]);
|
|
|
- comp->dig_p3 = (s16) le16_to_cpu(buf[2]);
|
|
|
- comp->dig_p4 = (s16) le16_to_cpu(buf[3]);
|
|
|
- comp->dig_p5 = (s16) le16_to_cpu(buf[4]);
|
|
|
- comp->dig_p6 = (s16) le16_to_cpu(buf[5]);
|
|
|
- comp->dig_p7 = (s16) le16_to_cpu(buf[6]);
|
|
|
- comp->dig_p8 = (s16) le16_to_cpu(buf[7]);
|
|
|
- comp->dig_p9 = (s16) le16_to_cpu(buf[8]);
|
|
|
-
|
|
|
- return 0;
|
|
|
-}
|
|
|
-
|
|
|
/*
|
|
|
* Returns temperature in DegC, resolution is 0.01 DegC. Output value of
|
|
|
* "5123" equals 51.23 DegC. t_fine carries fine temperature as global
|
|
@@ -197,16 +145,33 @@ static int bmp280_read_compensation_press(struct bmp280_data *data,
|
|
|
* Taken from datasheet, Section 3.11.3, "Compensation formula".
|
|
|
*/
|
|
|
static s32 bmp280_compensate_temp(struct bmp280_data *data,
|
|
|
- struct bmp280_comp_temp *comp,
|
|
|
s32 adc_temp)
|
|
|
{
|
|
|
+ int ret;
|
|
|
s32 var1, var2, t;
|
|
|
+ __le16 buf[BMP280_COMP_TEMP_REG_COUNT / 2];
|
|
|
+
|
|
|
+ ret = regmap_bulk_read(data->regmap, BMP280_REG_COMP_TEMP_START,
|
|
|
+ buf, BMP280_COMP_TEMP_REG_COUNT);
|
|
|
+ if (ret < 0) {
|
|
|
+ dev_err(&data->client->dev,
|
|
|
+ "failed to read temperature calibration parameters\n");
|
|
|
+ return ret;
|
|
|
+ }
|
|
|
|
|
|
- var1 = (((adc_temp >> 3) - ((s32) comp->dig_t1 << 1)) *
|
|
|
- ((s32) comp->dig_t2)) >> 11;
|
|
|
- var2 = (((((adc_temp >> 4) - ((s32) comp->dig_t1)) *
|
|
|
- ((adc_temp >> 4) - ((s32) comp->dig_t1))) >> 12) *
|
|
|
- ((s32) comp->dig_t3)) >> 14;
|
|
|
+ /*
|
|
|
+ * The double casts are necessary because le16_to_cpu returns an
|
|
|
+ * unsigned 16-bit value. Casting that value directly to a
|
|
|
+ * signed 32-bit will not do proper sign extension.
|
|
|
+ *
|
|
|
+ * Conversely, T1 and P1 are unsigned values, so they can be
|
|
|
+ * cast straight to the larger type.
|
|
|
+ */
|
|
|
+ var1 = (((adc_temp >> 3) - ((s32)le16_to_cpu(buf[T1]) << 1)) *
|
|
|
+ ((s32)(s16)le16_to_cpu(buf[T2]))) >> 11;
|
|
|
+ var2 = (((((adc_temp >> 4) - ((s32)le16_to_cpu(buf[T1]))) *
|
|
|
+ ((adc_temp >> 4) - ((s32)le16_to_cpu(buf[T1])))) >> 12) *
|
|
|
+ ((s32)(s16)le16_to_cpu(buf[T3]))) >> 14;
|
|
|
|
|
|
data->t_fine = var1 + var2;
|
|
|
t = (data->t_fine * 5 + 128) >> 8;
|
|
@@ -222,27 +187,36 @@ static s32 bmp280_compensate_temp(struct bmp280_data *data,
|
|
|
* Taken from datasheet, Section 3.11.3, "Compensation formula".
|
|
|
*/
|
|
|
static u32 bmp280_compensate_press(struct bmp280_data *data,
|
|
|
- struct bmp280_comp_press *comp,
|
|
|
s32 adc_press)
|
|
|
{
|
|
|
+ int ret;
|
|
|
s64 var1, var2, p;
|
|
|
+ __le16 buf[BMP280_COMP_PRESS_REG_COUNT / 2];
|
|
|
+
|
|
|
+ ret = regmap_bulk_read(data->regmap, BMP280_REG_COMP_PRESS_START,
|
|
|
+ buf, BMP280_COMP_PRESS_REG_COUNT);
|
|
|
+ if (ret < 0) {
|
|
|
+ dev_err(&data->client->dev,
|
|
|
+ "failed to read pressure calibration parameters\n");
|
|
|
+ return ret;
|
|
|
+ }
|
|
|
|
|
|
- var1 = ((s64) data->t_fine) - 128000;
|
|
|
- var2 = var1 * var1 * (s64) comp->dig_p6;
|
|
|
- var2 = var2 + ((var1 * (s64) comp->dig_p5) << 17);
|
|
|
- var2 = var2 + (((s64) comp->dig_p4) << 35);
|
|
|
- var1 = ((var1 * var1 * (s64) comp->dig_p3) >> 8) +
|
|
|
- ((var1 * (s64) comp->dig_p2) << 12);
|
|
|
- var1 = (((((s64) 1) << 47) + var1)) * ((s64) comp->dig_p1) >> 33;
|
|
|
+ var1 = ((s64)data->t_fine) - 128000;
|
|
|
+ var2 = var1 * var1 * (s64)(s16)le16_to_cpu(buf[P6]);
|
|
|
+ var2 = var2 + ((var1 * (s64)(s16)le16_to_cpu(buf[P5])) << 17);
|
|
|
+ var2 = var2 + (((s64)(s16)le16_to_cpu(buf[P4])) << 35);
|
|
|
+ var1 = ((var1 * var1 * (s64)(s16)le16_to_cpu(buf[P3])) >> 8) +
|
|
|
+ ((var1 * (s64)(s16)le16_to_cpu(buf[P2])) << 12);
|
|
|
+ var1 = ((((s64)1) << 47) + var1) * ((s64)le16_to_cpu(buf[P1])) >> 33;
|
|
|
|
|
|
if (var1 == 0)
|
|
|
return 0;
|
|
|
|
|
|
- p = ((((s64) 1048576 - adc_press) << 31) - var2) * 3125;
|
|
|
+ p = ((((s64)1048576 - adc_press) << 31) - var2) * 3125;
|
|
|
p = div64_s64(p, var1);
|
|
|
- var1 = (((s64) comp->dig_p9) * (p >> 13) * (p >> 13)) >> 25;
|
|
|
- var2 = (((s64) comp->dig_p8) * p) >> 19;
|
|
|
- p = ((p + var1 + var2) >> 8) + (((s64) comp->dig_p7) << 4);
|
|
|
+ var1 = (((s64)(s16)le16_to_cpu(buf[P9])) * (p >> 13) * (p >> 13)) >> 25;
|
|
|
+ var2 = (((s64)(s16)le16_to_cpu(buf[P8])) * p) >> 19;
|
|
|
+ p = ((p + var1 + var2) >> 8) + (((s64)(s16)le16_to_cpu(buf[P7])) << 4);
|
|
|
|
|
|
return (u32) p;
|
|
|
}
|
|
@@ -253,11 +227,6 @@ static int bmp280_read_temp(struct bmp280_data *data,
|
|
|
int ret;
|
|
|
__be32 tmp = 0;
|
|
|
s32 adc_temp, comp_temp;
|
|
|
- struct bmp280_comp_temp comp;
|
|
|
-
|
|
|
- ret = bmp280_read_compensation_temp(data, &comp);
|
|
|
- if (ret < 0)
|
|
|
- return ret;
|
|
|
|
|
|
ret = regmap_bulk_read(data->regmap, BMP280_REG_TEMP_MSB,
|
|
|
(u8 *) &tmp, 3);
|
|
@@ -267,7 +236,7 @@ static int bmp280_read_temp(struct bmp280_data *data,
|
|
|
}
|
|
|
|
|
|
adc_temp = be32_to_cpu(tmp) >> 12;
|
|
|
- comp_temp = bmp280_compensate_temp(data, &comp, adc_temp);
|
|
|
+ comp_temp = bmp280_compensate_temp(data, adc_temp);
|
|
|
|
|
|
/*
|
|
|
* val might be NULL if we're called by the read_press routine,
|
|
@@ -288,11 +257,6 @@ static int bmp280_read_press(struct bmp280_data *data,
|
|
|
__be32 tmp = 0;
|
|
|
s32 adc_press;
|
|
|
u32 comp_press;
|
|
|
- struct bmp280_comp_press comp;
|
|
|
-
|
|
|
- ret = bmp280_read_compensation_press(data, &comp);
|
|
|
- if (ret < 0)
|
|
|
- return ret;
|
|
|
|
|
|
/* Read and compensate temperature so we get a reading of t_fine. */
|
|
|
ret = bmp280_read_temp(data, NULL);
|
|
@@ -307,7 +271,7 @@ static int bmp280_read_press(struct bmp280_data *data,
|
|
|
}
|
|
|
|
|
|
adc_press = be32_to_cpu(tmp) >> 12;
|
|
|
- comp_press = bmp280_compensate_press(data, &comp, adc_press);
|
|
|
+ comp_press = bmp280_compensate_press(data, adc_press);
|
|
|
|
|
|
*val = comp_press;
|
|
|
*val2 = 256000;
|