* [PATCH v8 1/6] Documentation: ABI: IIO: add new convdelay documentation
2025-06-03 14:36 [PATCH v8 0/6] iio: adc: add ad7606 calibration support Angelo Dureghello
@ 2025-06-03 14:36 ` Angelo Dureghello
2025-06-03 15:54 ` David Lechner
2025-06-03 14:36 ` [PATCH v8 2/6] iio: core: add ADC delay calibration definition Angelo Dureghello
` (4 subsequent siblings)
5 siblings, 1 reply; 9+ messages in thread
From: Angelo Dureghello @ 2025-06-03 14:36 UTC (permalink / raw)
To: Jonathan Cameron, David Lechner, Nuno Sá, Andy Shevchenko,
Lars-Peter Clausen, Michael Hennerich, Rob Herring,
Krzysztof Kozlowski, Conor Dooley
Cc: linux-iio, linux-kernel, Michael Hennerich, devicetree,
Angelo Dureghello
From: Angelo Dureghello <adureghello@baylibre.com>
Add new IIO "convdelay" documentation.
The ad7606 implements a phase calibation feature, in nanoseconds.
Being this a time delay, using the convdelay suffix.
Signed-off-by: Angelo Dureghello <adureghello@baylibre.com>
---
Documentation/ABI/testing/sysfs-bus-iio | 24 ++++++++++++++++++++++++
1 file changed, 24 insertions(+)
diff --git a/Documentation/ABI/testing/sysfs-bus-iio b/Documentation/ABI/testing/sysfs-bus-iio
index ef52c427a015cf47bb9847782e13afbee01e9f31..7e59cbd5acb85fd0909c1d56f9d76a84933d418a 100644
--- a/Documentation/ABI/testing/sysfs-bus-iio
+++ b/Documentation/ABI/testing/sysfs-bus-iio
@@ -559,6 +559,30 @@ Description:
- a small discrete set of values like "0 2 4 6 8"
- a range specified as "[min step max]"
+What: /sys/bus/iio/devices/iio:deviceX/in_voltageY_convdelay
+KernelVersion: 6.16
+Contact: linux-iio@vger.kernel.org
+Description:
+ Delay of start of conversion from common reference point shared
+ by all channels. Can be writable when used to compensate for
+ delay variation introduced by external filters feeding a
+ simultaneous sampling ADC.
+
+ E.g., for the ad7606 ADC series, this value is intended as a
+ configurable time delay in seconds, to correct delay introduced
+ by an optional external filtering circuit.
+
+What: /sys/bus/iio/devices/iio:deviceX/in_voltageY_convdelay_available
+KernelVersion: 6.16
+Contact: linux-iio@vger.kernel.org
+Description:
+ Available values of convdelay. Maybe expressed as:
+
+ - a range specified as "[min step max]"
+
+ If shared across all channels, <type>_convdelay_available
+ is used.
+
What: /sys/bus/iio/devices/iio:deviceX/in_accel_x_calibscale
What: /sys/bus/iio/devices/iio:deviceX/in_accel_y_calibscale
What: /sys/bus/iio/devices/iio:deviceX/in_accel_z_calibscale
--
2.49.0
^ permalink raw reply related [flat|nested] 9+ messages in thread* Re: [PATCH v8 1/6] Documentation: ABI: IIO: add new convdelay documentation
2025-06-03 14:36 ` [PATCH v8 1/6] Documentation: ABI: IIO: add new convdelay documentation Angelo Dureghello
@ 2025-06-03 15:54 ` David Lechner
0 siblings, 0 replies; 9+ messages in thread
From: David Lechner @ 2025-06-03 15:54 UTC (permalink / raw)
To: Angelo Dureghello, Jonathan Cameron, Nuno Sá,
Andy Shevchenko, Lars-Peter Clausen, Michael Hennerich,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: linux-iio, linux-kernel, devicetree
On 6/3/25 9:36 AM, Angelo Dureghello wrote:
> From: Angelo Dureghello <adureghello@baylibre.com>
>
> Add new IIO "convdelay" documentation.
>
> The ad7606 implements a phase calibation feature, in nanoseconds.
> Being this a time delay, using the convdelay suffix.
>
> Signed-off-by: Angelo Dureghello <adureghello@baylibre.com>
> ---
> Documentation/ABI/testing/sysfs-bus-iio | 24 ++++++++++++++++++++++++
> 1 file changed, 24 insertions(+)
>
> diff --git a/Documentation/ABI/testing/sysfs-bus-iio b/Documentation/ABI/testing/sysfs-bus-iio
> index ef52c427a015cf47bb9847782e13afbee01e9f31..7e59cbd5acb85fd0909c1d56f9d76a84933d418a 100644
> --- a/Documentation/ABI/testing/sysfs-bus-iio
> +++ b/Documentation/ABI/testing/sysfs-bus-iio
> @@ -559,6 +559,30 @@ Description:
> - a small discrete set of values like "0 2 4 6 8"
> - a range specified as "[min step max]"
>
> +What: /sys/bus/iio/devices/iio:deviceX/in_voltageY_convdelay
> +KernelVersion: 6.16
> +Contact: linux-iio@vger.kernel.org
> +Description:
> + Delay of start of conversion from common reference point shared
> + by all channels. Can be writable when used to compensate for
> + delay variation introduced by external filters feeding a
> + simultaneous sampling ADC.
> +
> + E.g., for the ad7606 ADC series, this value is intended as a
> + configurable time delay in seconds, to correct delay introduced
> + by an optional external filtering circuit.
> +
> +What: /sys/bus/iio/devices/iio:deviceX/in_voltageY_convdelay_available
> +KernelVersion: 6.16
Forgot to change these to 6.17?
> +Contact: linux-iio@vger.kernel.org
> +Description:
> + Available values of convdelay. Maybe expressed as:
> +
> + - a range specified as "[min step max]"
> +
> + If shared across all channels, <type>_convdelay_available
> + is used.
> +
> What: /sys/bus/iio/devices/iio:deviceX/in_accel_x_calibscale
> What: /sys/bus/iio/devices/iio:deviceX/in_accel_y_calibscale
> What: /sys/bus/iio/devices/iio:deviceX/in_accel_z_calibscale
>
^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCH v8 2/6] iio: core: add ADC delay calibration definition
2025-06-03 14:36 [PATCH v8 0/6] iio: adc: add ad7606 calibration support Angelo Dureghello
2025-06-03 14:36 ` [PATCH v8 1/6] Documentation: ABI: IIO: add new convdelay documentation Angelo Dureghello
@ 2025-06-03 14:36 ` Angelo Dureghello
2025-06-03 14:36 ` [PATCH v8 3/6] iio: adc: ad7606: add offset and phase calibration support Angelo Dureghello
` (3 subsequent siblings)
5 siblings, 0 replies; 9+ messages in thread
From: Angelo Dureghello @ 2025-06-03 14:36 UTC (permalink / raw)
To: Jonathan Cameron, David Lechner, Nuno Sá, Andy Shevchenko,
Lars-Peter Clausen, Michael Hennerich, Rob Herring,
Krzysztof Kozlowski, Conor Dooley
Cc: linux-iio, linux-kernel, Michael Hennerich, devicetree,
Angelo Dureghello
From: Angelo Dureghello <adureghello@baylibre.com>
ADCs as ad7606 implement a phase calibration as a delay. Add such
definition, needed for ad7606.
Signed-off-by: Angelo Dureghello <adureghello@baylibre.com>
---
drivers/iio/industrialio-core.c | 1 +
include/linux/iio/types.h | 1 +
2 files changed, 2 insertions(+)
diff --git a/drivers/iio/industrialio-core.c b/drivers/iio/industrialio-core.c
index 178e99b111debc59a247fcc3a6037e429db3bebf..f13c3aa470d774bfe655d6a9fb00c263789db637 100644
--- a/drivers/iio/industrialio-core.c
+++ b/drivers/iio/industrialio-core.c
@@ -188,6 +188,7 @@ static const char * const iio_chan_info_postfix[] = {
[IIO_CHAN_INFO_CALIBAMBIENT] = "calibambient",
[IIO_CHAN_INFO_ZEROPOINT] = "zeropoint",
[IIO_CHAN_INFO_TROUGH] = "trough_raw",
+ [IIO_CHAN_INFO_CONVDELAY] = "convdelay",
};
/**
* iio_device_id() - query the unique ID for the device
diff --git a/include/linux/iio/types.h b/include/linux/iio/types.h
index d89982c98368cf72c0fc30fa66ab001e48af4e8b..ad2761efcc8315e1f9907d2a7159447fb463333e 100644
--- a/include/linux/iio/types.h
+++ b/include/linux/iio/types.h
@@ -69,6 +69,7 @@ enum iio_chan_info_enum {
IIO_CHAN_INFO_CALIBAMBIENT,
IIO_CHAN_INFO_ZEROPOINT,
IIO_CHAN_INFO_TROUGH,
+ IIO_CHAN_INFO_CONVDELAY,
};
#endif /* _IIO_TYPES_H_ */
--
2.49.0
^ permalink raw reply related [flat|nested] 9+ messages in thread* [PATCH v8 3/6] iio: adc: ad7606: add offset and phase calibration support
2025-06-03 14:36 [PATCH v8 0/6] iio: adc: add ad7606 calibration support Angelo Dureghello
2025-06-03 14:36 ` [PATCH v8 1/6] Documentation: ABI: IIO: add new convdelay documentation Angelo Dureghello
2025-06-03 14:36 ` [PATCH v8 2/6] iio: core: add ADC delay calibration definition Angelo Dureghello
@ 2025-06-03 14:36 ` Angelo Dureghello
2025-06-03 14:36 ` [PATCH v8 4/6] dt-bindings: iio: adc: adi,ad7606: add gain " Angelo Dureghello
` (2 subsequent siblings)
5 siblings, 0 replies; 9+ messages in thread
From: Angelo Dureghello @ 2025-06-03 14:36 UTC (permalink / raw)
To: Jonathan Cameron, David Lechner, Nuno Sá, Andy Shevchenko,
Lars-Peter Clausen, Michael Hennerich, Rob Herring,
Krzysztof Kozlowski, Conor Dooley
Cc: linux-iio, linux-kernel, Michael Hennerich, devicetree,
Angelo Dureghello
From: Angelo Dureghello <adureghello@baylibre.com>
Add support for offset and phase calibration, only for
devices that support software mode, that are:
ad7606b
ad7606c-16
ad7606c-18
Tested-by: David Lechner <dlechner@baylibre.com>
Reviewed-by: Nuno Sá <nuno.sa@analog.com>
Signed-off-by: Angelo Dureghello <adureghello@baylibre.com>
---
drivers/iio/adc/ad7606.c | 160 +++++++++++++++++++++++++++++++++++++++++++++++
drivers/iio/adc/ad7606.h | 9 +++
2 files changed, 169 insertions(+)
diff --git a/drivers/iio/adc/ad7606.c b/drivers/iio/adc/ad7606.c
index 185243dee86ed2e9ebc43b578003d0c010e97a9f..9a9bef0cfbb37138f71ba5b1babeaa423eaf4d5a 100644
--- a/drivers/iio/adc/ad7606.c
+++ b/drivers/iio/adc/ad7606.c
@@ -95,6 +95,22 @@ static const unsigned int ad7616_oversampling_avail[8] = {
1, 2, 4, 8, 16, 32, 64, 128,
};
+static const int ad7606_calib_offset_avail[3] = {
+ -128, 1, 127,
+};
+
+static const int ad7606c_18bit_calib_offset_avail[3] = {
+ -512, 4, 508,
+};
+
+static const int ad7606b_calib_phase_avail[][2] = {
+ { 0, 0 }, { 0, 1250 }, { 0, 318750 },
+};
+
+static const int ad7606c_calib_phase_avail[][2] = {
+ { 0, 0 }, { 0, 1000 }, { 0, 255000 },
+};
+
static int ad7606c_18bit_chan_scale_setup(struct iio_dev *indio_dev,
struct iio_chan_spec *chan);
static int ad7606c_16bit_chan_scale_setup(struct iio_dev *indio_dev,
@@ -164,6 +180,8 @@ const struct ad7606_chip_info ad7606b_info = {
.scale_setup_cb = ad7606_16bit_chan_scale_setup,
.sw_setup_cb = ad7606b_sw_mode_setup,
.offload_storagebits = 32,
+ .calib_offset_avail = ad7606_calib_offset_avail,
+ .calib_phase_avail = ad7606b_calib_phase_avail,
};
EXPORT_SYMBOL_NS_GPL(ad7606b_info, "IIO_AD7606");
@@ -177,6 +195,8 @@ const struct ad7606_chip_info ad7606c_16_info = {
.scale_setup_cb = ad7606c_16bit_chan_scale_setup,
.sw_setup_cb = ad7606b_sw_mode_setup,
.offload_storagebits = 32,
+ .calib_offset_avail = ad7606_calib_offset_avail,
+ .calib_phase_avail = ad7606c_calib_phase_avail,
};
EXPORT_SYMBOL_NS_GPL(ad7606c_16_info, "IIO_AD7606");
@@ -226,6 +246,8 @@ const struct ad7606_chip_info ad7606c_18_info = {
.scale_setup_cb = ad7606c_18bit_chan_scale_setup,
.sw_setup_cb = ad7606b_sw_mode_setup,
.offload_storagebits = 32,
+ .calib_offset_avail = ad7606c_18bit_calib_offset_avail,
+ .calib_phase_avail = ad7606c_calib_phase_avail,
};
EXPORT_SYMBOL_NS_GPL(ad7606c_18_info, "IIO_AD7606");
@@ -681,6 +703,40 @@ static int ad7606_scan_direct(struct iio_dev *indio_dev, unsigned int ch,
return ret;
}
+static int ad7606_get_calib_offset(struct ad7606_state *st, int ch, int *val)
+{
+ int ret;
+
+ ret = st->bops->reg_read(st, AD7606_CALIB_OFFSET(ch));
+ if (ret < 0)
+ return ret;
+
+ *val = st->chip_info->calib_offset_avail[0] +
+ ret * st->chip_info->calib_offset_avail[1];
+
+ return 0;
+}
+
+static int ad7606_get_calib_phase(struct ad7606_state *st, int ch, int *val,
+ int *val2)
+{
+ int ret;
+
+ ret = st->bops->reg_read(st, AD7606_CALIB_PHASE(ch));
+ if (ret < 0)
+ return ret;
+
+ *val = 0;
+
+ /*
+ * ad7606b: phase delay from 0 to 318.75 μs in steps of 1.25 μs.
+ * ad7606c-16/18: phase delay from 0 µs to 255 µs in steps of 1 µs.
+ */
+ *val2 = ret * st->chip_info->calib_phase_avail[1][1];
+
+ return 0;
+}
+
static int ad7606_read_raw(struct iio_dev *indio_dev,
struct iio_chan_spec const *chan,
int *val,
@@ -715,6 +771,22 @@ static int ad7606_read_raw(struct iio_dev *indio_dev,
pwm_get_state(st->cnvst_pwm, &cnvst_pwm_state);
*val = DIV_ROUND_CLOSEST_ULL(NSEC_PER_SEC, cnvst_pwm_state.period);
return IIO_VAL_INT;
+ case IIO_CHAN_INFO_CALIBBIAS:
+ if (!iio_device_claim_direct(indio_dev))
+ return -EBUSY;
+ ret = ad7606_get_calib_offset(st, chan->scan_index, val);
+ iio_device_release_direct(indio_dev);
+ if (ret)
+ return ret;
+ return IIO_VAL_INT;
+ case IIO_CHAN_INFO_CONVDELAY:
+ if (!iio_device_claim_direct(indio_dev))
+ return -EBUSY;
+ ret = ad7606_get_calib_phase(st, chan->scan_index, val, val2);
+ iio_device_release_direct(indio_dev);
+ if (ret)
+ return ret;
+ return IIO_VAL_INT_PLUS_NANO;
}
return -EINVAL;
}
@@ -765,6 +837,64 @@ static int ad7606_write_os_hw(struct iio_dev *indio_dev, int val)
return 0;
}
+static int ad7606_set_calib_offset(struct ad7606_state *st, int ch, int val)
+{
+ int start_val, step_val, stop_val;
+ int offset;
+
+ start_val = st->chip_info->calib_offset_avail[0];
+ step_val = st->chip_info->calib_offset_avail[1];
+ stop_val = st->chip_info->calib_offset_avail[2];
+
+ if (val < start_val || val > stop_val)
+ return -EINVAL;
+
+ offset = (val - start_val) / step_val;
+
+ return st->bops->reg_write(st, AD7606_CALIB_OFFSET(ch), offset);
+}
+
+static int ad7606_set_calib_phase(struct ad7606_state *st, int ch, int val,
+ int val2)
+{
+ int wreg, start_ns, step_ns, stop_ns;
+
+ if (val != 0)
+ return -EINVAL;
+
+ start_ns = st->chip_info->calib_phase_avail[0][1];
+ step_ns = st->chip_info->calib_phase_avail[1][1];
+ stop_ns = st->chip_info->calib_phase_avail[2][1];
+
+ /*
+ * ad7606b: phase delay from 0 to 318.75 μs in steps of 1.25 μs.
+ * ad7606c-16/18: phase delay from 0 µs to 255 µs in steps of 1 µs.
+ */
+ if (val2 < start_ns || val2 > stop_ns)
+ return -EINVAL;
+
+ wreg = val2 / step_ns;
+
+ return st->bops->reg_write(st, AD7606_CALIB_PHASE(ch), wreg);
+}
+
+static int ad7606_write_raw_get_fmt(struct iio_dev *indio_dev,
+ struct iio_chan_spec const *chan, long info)
+{
+ switch (info) {
+ case IIO_CHAN_INFO_SCALE:
+ return IIO_VAL_INT_PLUS_MICRO;
+ case IIO_CHAN_INFO_SAMP_FREQ:
+ case IIO_CHAN_INFO_OVERSAMPLING_RATIO:
+ case IIO_CHAN_INFO_CALIBBIAS:
+ return IIO_VAL_INT;
+ case IIO_CHAN_INFO_CONVDELAY:
+ return IIO_VAL_INT_PLUS_NANO;
+ default:
+ return -EINVAL;
+ }
+}
+
static int ad7606_write_raw(struct iio_dev *indio_dev,
struct iio_chan_spec const *chan,
int val,
@@ -818,6 +948,18 @@ static int ad7606_write_raw(struct iio_dev *indio_dev,
if (val < 0 && val2 != 0)
return -EINVAL;
return ad7606_set_sampling_freq(st, val);
+ case IIO_CHAN_INFO_CALIBBIAS:
+ if (!iio_device_claim_direct(indio_dev))
+ return -EBUSY;
+ ret = ad7606_set_calib_offset(st, chan->scan_index, val);
+ iio_device_release_direct(indio_dev);
+ return ret;
+ case IIO_CHAN_INFO_CONVDELAY:
+ if (!iio_device_claim_direct(indio_dev))
+ return -EBUSY;
+ ret = ad7606_set_calib_phase(st, chan->scan_index, val, val2);
+ iio_device_release_direct(indio_dev);
+ return ret;
default:
return -EINVAL;
}
@@ -996,6 +1138,14 @@ static int ad7606_read_avail(struct iio_dev *indio_dev,
*type = IIO_VAL_INT_PLUS_MICRO;
return IIO_AVAIL_LIST;
+ case IIO_CHAN_INFO_CALIBBIAS:
+ *vals = st->chip_info->calib_offset_avail;
+ *type = IIO_VAL_INT;
+ return IIO_AVAIL_RANGE;
+ case IIO_CHAN_INFO_CONVDELAY:
+ *vals = (const int *)st->chip_info->calib_phase_avail;
+ *type = IIO_VAL_INT_PLUS_NANO;
+ return IIO_AVAIL_RANGE;
}
return -EINVAL;
}
@@ -1058,6 +1208,7 @@ static const struct iio_info ad7606_info_sw_mode = {
.read_raw = &ad7606_read_raw,
.write_raw = &ad7606_write_raw,
.read_avail = &ad7606_read_avail,
+ .write_raw_get_fmt = ad7606_write_raw_get_fmt,
.debugfs_reg_access = &ad7606_reg_access,
.validate_trigger = &ad7606_validate_trigger,
.update_scan_mode = &ad7606_update_scan_mode,
@@ -1250,6 +1401,15 @@ static int ad7606_probe_channels(struct iio_dev *indio_dev)
chan->info_mask_separate_available |=
BIT(IIO_CHAN_INFO_SCALE);
+ if (st->chip_info->calib_offset_avail) {
+ chan->info_mask_separate |=
+ BIT(IIO_CHAN_INFO_CALIBBIAS) |
+ BIT(IIO_CHAN_INFO_CONVDELAY);
+ chan->info_mask_separate_available |=
+ BIT(IIO_CHAN_INFO_CALIBBIAS) |
+ BIT(IIO_CHAN_INFO_CONVDELAY);
+ }
+
/*
* All chips with software mode support oversampling,
* so we skip the oversampling_available check. And the
diff --git a/drivers/iio/adc/ad7606.h b/drivers/iio/adc/ad7606.h
index 441e62c521bcbea69b4f70bb2d55f65334d22276..f613583a7fa4095115b0b28e3f8e51cd32b93524 100644
--- a/drivers/iio/adc/ad7606.h
+++ b/drivers/iio/adc/ad7606.h
@@ -40,6 +40,11 @@
#define AD7606_RANGE_CH_ADDR(ch) (0x03 + ((ch) >> 1))
#define AD7606_OS_MODE 0x08
+#define AD7606_CALIB_GAIN(ch) (0x09 + (ch))
+#define AD7606_CALIB_GAIN_MASK GENMASK(5, 0)
+#define AD7606_CALIB_OFFSET(ch) (0x11 + (ch))
+#define AD7606_CALIB_PHASE(ch) (0x19 + (ch))
+
struct ad7606_state;
typedef int (*ad7606_scale_setup_cb_t)(struct iio_dev *indio_dev,
@@ -61,6 +66,8 @@ typedef int (*ad7606_sw_setup_cb_t)(struct iio_dev *indio_dev);
* @init_delay_ms: required delay in milliseconds for initialization
* after a restart
* @offload_storagebits: storage bits used by the offload hw implementation
+ * @calib_offset_avail: pointer to offset calibration range/limits array
+ * @calib_phase_avail: pointer to phase calibration range/limits array
*/
struct ad7606_chip_info {
unsigned int max_samplerate;
@@ -74,6 +81,8 @@ struct ad7606_chip_info {
bool os_req_reset;
unsigned long init_delay_ms;
u8 offload_storagebits;
+ const int *calib_offset_avail;
+ const int (*calib_phase_avail)[2];
};
/**
--
2.49.0
^ permalink raw reply related [flat|nested] 9+ messages in thread* [PATCH v8 4/6] dt-bindings: iio: adc: adi,ad7606: add gain calibration support
2025-06-03 14:36 [PATCH v8 0/6] iio: adc: add ad7606 calibration support Angelo Dureghello
` (2 preceding siblings ...)
2025-06-03 14:36 ` [PATCH v8 3/6] iio: adc: ad7606: add offset and phase calibration support Angelo Dureghello
@ 2025-06-03 14:36 ` Angelo Dureghello
2025-06-03 14:36 ` [PATCH v8 5/6] iio: adc: ad7606: exit for invalid fdt dt_schema properties Angelo Dureghello
2025-06-03 14:36 ` [PATCH v8 6/6] iio: adc: ad7606: add gain calibration support Angelo Dureghello
5 siblings, 0 replies; 9+ messages in thread
From: Angelo Dureghello @ 2025-06-03 14:36 UTC (permalink / raw)
To: Jonathan Cameron, David Lechner, Nuno Sá, Andy Shevchenko,
Lars-Peter Clausen, Michael Hennerich, Rob Herring,
Krzysztof Kozlowski, Conor Dooley
Cc: linux-iio, linux-kernel, Michael Hennerich, devicetree,
Angelo Dureghello, Conor Dooley
From: Angelo Dureghello <adureghello@baylibre.com>
Add gain calibration support by a per-channel resistor value.
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Angelo Dureghello <adureghello@baylibre.com>
---
.../devicetree/bindings/iio/adc/adi,ad7606.yaml | 29 ++++++++++++++++++++++
1 file changed, 29 insertions(+)
diff --git a/Documentation/devicetree/bindings/iio/adc/adi,ad7606.yaml b/Documentation/devicetree/bindings/iio/adc/adi,ad7606.yaml
index 29f12d650442b8ff2eb455306ce59a0e87867ddd..6926f5f090ad6bbbe7bfd9327dc5ae17dafcd1fd 100644
--- a/Documentation/devicetree/bindings/iio/adc/adi,ad7606.yaml
+++ b/Documentation/devicetree/bindings/iio/adc/adi,ad7606.yaml
@@ -204,6 +204,15 @@ patternProperties:
considered a bipolar differential channel. Otherwise it is bipolar
single-ended.
+ adi,rfilter-ohms:
+ description:
+ For ADCs that supports gain calibration, this property must be set to
+ the value of the external RFilter resistor. Proper gain error
+ correction is applied based on this value.
+ default: 0
+ minimum: 0
+ maximum: 64512
+
required:
- reg
- bipolar
@@ -256,6 +265,25 @@ allOf:
properties:
adi,oversampling-ratio-gpios: false
+ - if:
+ properties:
+ compatible:
+ contains:
+ enum:
+ - adi,ad7605-4
+ - adi,ad7606-4
+ - adi,ad7606-6
+ - adi,ad7606-8
+ - adi,ad7607
+ - adi,ad7608
+ - adi,ad7609
+ - adi,ad7616
+ then:
+ patternProperties:
+ "^channel@[0-9a-f]+$":
+ properties:
+ adi,rfilter-ohms: false
+
- if:
properties:
compatible:
@@ -398,6 +426,7 @@ examples:
reg = <8>;
diff-channels = <8 8>;
bipolar;
+ adi,rfilter-ohms = <2048>;
};
};
--
2.49.0
^ permalink raw reply related [flat|nested] 9+ messages in thread* [PATCH v8 5/6] iio: adc: ad7606: exit for invalid fdt dt_schema properties
2025-06-03 14:36 [PATCH v8 0/6] iio: adc: add ad7606 calibration support Angelo Dureghello
` (3 preceding siblings ...)
2025-06-03 14:36 ` [PATCH v8 4/6] dt-bindings: iio: adc: adi,ad7606: add gain " Angelo Dureghello
@ 2025-06-03 14:36 ` Angelo Dureghello
2025-06-03 14:36 ` [PATCH v8 6/6] iio: adc: ad7606: add gain calibration support Angelo Dureghello
5 siblings, 0 replies; 9+ messages in thread
From: Angelo Dureghello @ 2025-06-03 14:36 UTC (permalink / raw)
To: Jonathan Cameron, David Lechner, Nuno Sá, Andy Shevchenko,
Lars-Peter Clausen, Michael Hennerich, Rob Herring,
Krzysztof Kozlowski, Conor Dooley
Cc: linux-iio, linux-kernel, Michael Hennerich, devicetree,
Angelo Dureghello
From: Angelo Dureghello <adureghello@baylibre.com>
Fix ad7606_get_chan_config() fdt parsing function to exit for error in
case of invalid dt_schema values.
Idea is to not proceed when there are values that are not allowed under
the dt_schema.
Signed-off-by: Angelo Dureghello <adureghello@baylibre.com>
---
drivers/iio/adc/ad7606.c | 10 ++++------
1 file changed, 4 insertions(+), 6 deletions(-)
diff --git a/drivers/iio/adc/ad7606.c b/drivers/iio/adc/ad7606.c
index 9a9bef0cfbb37138f71ba5b1babeaa423eaf4d5a..e0a666cc0e14255754e74daa9e1e88bc4ad1665c 100644
--- a/drivers/iio/adc/ad7606.c
+++ b/drivers/iio/adc/ad7606.c
@@ -319,15 +319,13 @@ static int ad7606_get_chan_config(struct iio_dev *indio_dev, int ch,
ret = fwnode_property_read_u32(child, "reg", ®);
if (ret)
- continue;
+ return ret;
/* channel number (here) is from 1 to num_channels */
- if (reg < 1 || reg > num_channels) {
- dev_warn(dev,
- "Invalid channel number (ignoring): %d\n", reg);
- continue;
- }
+ if (reg < 1 || reg > num_channels)
+ return -EINVAL;
+ /* Loop until we are in the right channel. */
if (reg != (ch + 1))
continue;
--
2.49.0
^ permalink raw reply related [flat|nested] 9+ messages in thread* [PATCH v8 6/6] iio: adc: ad7606: add gain calibration support
2025-06-03 14:36 [PATCH v8 0/6] iio: adc: add ad7606 calibration support Angelo Dureghello
` (4 preceding siblings ...)
2025-06-03 14:36 ` [PATCH v8 5/6] iio: adc: ad7606: exit for invalid fdt dt_schema properties Angelo Dureghello
@ 2025-06-03 14:36 ` Angelo Dureghello
2025-06-03 15:52 ` David Lechner
5 siblings, 1 reply; 9+ messages in thread
From: Angelo Dureghello @ 2025-06-03 14:36 UTC (permalink / raw)
To: Jonathan Cameron, David Lechner, Nuno Sá, Andy Shevchenko,
Lars-Peter Clausen, Michael Hennerich, Rob Herring,
Krzysztof Kozlowski, Conor Dooley
Cc: linux-iio, linux-kernel, Michael Hennerich, devicetree,
Angelo Dureghello
From: Angelo Dureghello <adureghello@baylibre.com>
Add gain calibration support, using resistor values set on devicetree,
values to be set accordingly with ADC external RFilter, as explained in
the ad7606c-16 datasheet, rev0, page 37.
Usage example in the fdt yaml documentation.
Tested-by: David Lechner <dlechner@baylibre.com>
Reviewed-by: Nuno Sá <nuno.sa@analog.com>
Signed-off-by: Angelo Dureghello <adureghello@baylibre.com>
---
drivers/iio/adc/ad7606.c | 66 ++++++++++++++++++++++++++++++++++++++++++++++++
drivers/iio/adc/ad7606.h | 6 +++++
2 files changed, 72 insertions(+)
diff --git a/drivers/iio/adc/ad7606.c b/drivers/iio/adc/ad7606.c
index e0a666cc0e14255754e74daa9e1e88bc4ad1665c..22dcb52ced57d4305db6401605c064fc438d5be4 100644
--- a/drivers/iio/adc/ad7606.c
+++ b/drivers/iio/adc/ad7606.c
@@ -33,6 +33,10 @@
#include "ad7606.h"
+#define AD7606_CALIB_GAIN_MIN 0
+#define AD7606_CALIB_GAIN_STEP 1024
+#define AD7606_CALIB_GAIN_MAX (63 * AD7606_CALIB_GAIN_STEP)
+
/*
* Scales are computed as 5000/32768 and 10000/32768 respectively,
* so that when applied to the raw values they provide mV values.
@@ -125,6 +129,7 @@ static int ad7609_chan_scale_setup(struct iio_dev *indio_dev,
struct iio_chan_spec *chan);
static int ad7616_sw_mode_setup(struct iio_dev *indio_dev);
static int ad7606b_sw_mode_setup(struct iio_dev *indio_dev);
+static int ad7606_chan_calib_gain_setup(struct iio_dev *indio_dev);
const struct ad7606_chip_info ad7605_4_info = {
.max_samplerate = 300 * KILO,
@@ -180,6 +185,7 @@ const struct ad7606_chip_info ad7606b_info = {
.scale_setup_cb = ad7606_16bit_chan_scale_setup,
.sw_setup_cb = ad7606b_sw_mode_setup,
.offload_storagebits = 32,
+ .calib_gain_setup_cb = ad7606_chan_calib_gain_setup,
.calib_offset_avail = ad7606_calib_offset_avail,
.calib_phase_avail = ad7606b_calib_phase_avail,
};
@@ -195,6 +201,7 @@ const struct ad7606_chip_info ad7606c_16_info = {
.scale_setup_cb = ad7606c_16bit_chan_scale_setup,
.sw_setup_cb = ad7606b_sw_mode_setup,
.offload_storagebits = 32,
+ .calib_gain_setup_cb = ad7606_chan_calib_gain_setup,
.calib_offset_avail = ad7606_calib_offset_avail,
.calib_phase_avail = ad7606c_calib_phase_avail,
};
@@ -246,6 +253,7 @@ const struct ad7606_chip_info ad7606c_18_info = {
.scale_setup_cb = ad7606c_18bit_chan_scale_setup,
.sw_setup_cb = ad7606b_sw_mode_setup,
.offload_storagebits = 32,
+ .calib_gain_setup_cb = ad7606_chan_calib_gain_setup,
.calib_offset_avail = ad7606c_18bit_calib_offset_avail,
.calib_phase_avail = ad7606c_calib_phase_avail,
};
@@ -355,6 +363,36 @@ static int ad7606_get_chan_config(struct iio_dev *indio_dev, int ch,
return 0;
}
+static int ad7606_chan_calib_gain_setup(struct iio_dev *indio_dev)
+{
+ struct ad7606_state *st = iio_priv(indio_dev);
+ unsigned int num_channels = st->chip_info->num_adc_channels;
+ struct device *dev = st->dev;
+ int ret;
+
+ device_for_each_child_node_scoped(dev, child) {
+ u32 reg, r_gain;
+
+ ret = fwnode_property_read_u32(child, "reg", ®);
+ if (ret)
+ return ret;
+
+ /* Chan reg is a 1-based index. */
+ if (reg < 1 || reg > num_channels)
+ return -EINVAL;
+
+ r_gain = 0;
+ ret = fwnode_property_read_u32(child, "adi,rfilter-ohms",
+ &r_gain);
+ if (r_gain > AD7606_CALIB_GAIN_MAX)
+ return -EINVAL;
+
+ st->r_gain[reg - 1] = r_gain;
+ }
+
+ return 0;
+}
+
static int ad7606c_18bit_chan_scale_setup(struct iio_dev *indio_dev,
struct iio_chan_spec *chan)
{
@@ -1352,6 +1390,21 @@ static int ad7606b_sw_mode_setup(struct iio_dev *indio_dev)
return st->bops->sw_mode_config(indio_dev);
}
+static int ad7606_set_gain_calib(struct ad7606_state *st)
+{
+ int i, ret;
+
+ for (i = 0; i < st->chip_info->num_adc_channels; i++) {
+ ret = st->bops->reg_write(st, AD7606_CALIB_GAIN(i),
+ DIV_ROUND_CLOSEST(st->r_gain[i],
+ AD7606_CALIB_GAIN_STEP));
+ if (ret)
+ return ret;
+ }
+
+ return 0;
+}
+
static int ad7606_probe_channels(struct iio_dev *indio_dev)
{
struct ad7606_state *st = iio_priv(indio_dev);
@@ -1444,6 +1497,13 @@ static int ad7606_probe_channels(struct iio_dev *indio_dev)
if (slow_bus)
channels[i] = (struct iio_chan_spec)IIO_CHAN_SOFT_TIMESTAMP(i);
+ /* Getting gain calibration values for all channels. */
+ if (st->sw_mode_en && st->chip_info->calib_gain_setup_cb) {
+ ret = st->chip_info->calib_gain_setup_cb(indio_dev);
+ if (ret)
+ return ret;
+ }
+
indio_dev->channels = channels;
return 0;
@@ -1620,6 +1680,12 @@ int ad7606_probe(struct device *dev, int irq, void __iomem *base_address,
st->chip_info->sw_setup_cb(indio_dev);
}
+ if (st->sw_mode_en && st->chip_info->calib_gain_setup_cb) {
+ ret = ad7606_set_gain_calib(st);
+ if (ret)
+ return ret;
+ }
+
return devm_iio_device_register(dev, indio_dev);
}
EXPORT_SYMBOL_NS_GPL(ad7606_probe, "IIO_AD7606");
diff --git a/drivers/iio/adc/ad7606.h b/drivers/iio/adc/ad7606.h
index f613583a7fa4095115b0b28e3f8e51cd32b93524..a5b0d318e2f4d73d3708288536e807957c5de68c 100644
--- a/drivers/iio/adc/ad7606.h
+++ b/drivers/iio/adc/ad7606.h
@@ -50,6 +50,7 @@ struct ad7606_state;
typedef int (*ad7606_scale_setup_cb_t)(struct iio_dev *indio_dev,
struct iio_chan_spec *chan);
typedef int (*ad7606_sw_setup_cb_t)(struct iio_dev *indio_dev);
+typedef int (*ad7606_calib_gain_setup_cb_t)(struct iio_dev *indio_dev);
/**
* struct ad7606_chip_info - chip specific information
@@ -66,6 +67,7 @@ typedef int (*ad7606_sw_setup_cb_t)(struct iio_dev *indio_dev);
* @init_delay_ms: required delay in milliseconds for initialization
* after a restart
* @offload_storagebits: storage bits used by the offload hw implementation
+ * @calib_gain_setup_cb: callback to setup of gain calibration
* @calib_offset_avail: pointer to offset calibration range/limits array
* @calib_phase_avail: pointer to phase calibration range/limits array
*/
@@ -81,6 +83,7 @@ struct ad7606_chip_info {
bool os_req_reset;
unsigned long init_delay_ms;
u8 offload_storagebits;
+ ad7606_calib_gain_setup_cb_t calib_gain_setup_cb;
const int *calib_offset_avail;
const int (*calib_phase_avail)[2];
};
@@ -131,6 +134,7 @@ struct ad7606_chan_scale {
* @data: buffer for reading data from the device
* @offload_en: SPI offload enabled
* @bus_data: bus-specific variables
+ * @r_gain: array to store gain calibration resistor value in ohm
* @d16: be16 buffer for reading data from the device
*/
struct ad7606_state {
@@ -161,6 +165,8 @@ struct ad7606_state {
bool offload_en;
void *bus_data;
+ int r_gain[AD760X_MAX_CHANNELS];
+
/*
* DMA (thus cache coherency maintenance) may require the
* transfer buffers to live in their own cache lines.
--
2.49.0
^ permalink raw reply related [flat|nested] 9+ messages in thread* Re: [PATCH v8 6/6] iio: adc: ad7606: add gain calibration support
2025-06-03 14:36 ` [PATCH v8 6/6] iio: adc: ad7606: add gain calibration support Angelo Dureghello
@ 2025-06-03 15:52 ` David Lechner
0 siblings, 0 replies; 9+ messages in thread
From: David Lechner @ 2025-06-03 15:52 UTC (permalink / raw)
To: Angelo Dureghello, Jonathan Cameron, Nuno Sá,
Andy Shevchenko, Lars-Peter Clausen, Michael Hennerich,
Rob Herring, Krzysztof Kozlowski, Conor Dooley
Cc: linux-iio, linux-kernel, devicetree
On 6/3/25 9:36 AM, Angelo Dureghello wrote:
> From: Angelo Dureghello <adureghello@baylibre.com>
>
> Add gain calibration support, using resistor values set on devicetree,
> values to be set accordingly with ADC external RFilter, as explained in
> the ad7606c-16 datasheet, rev0, page 37.
>
> Usage example in the fdt yaml documentation.
>
> Tested-by: David Lechner <dlechner@baylibre.com>
> Reviewed-by: Nuno Sá <nuno.sa@analog.com>
The patch has changed significantly since these tags were given, so
these tags should have been dropped.
> Signed-off-by: Angelo Dureghello <adureghello@baylibre.com>
> ---
> drivers/iio/adc/ad7606.c | 66 ++++++++++++++++++++++++++++++++++++++++++++++++
> drivers/iio/adc/ad7606.h | 6 +++++
> 2 files changed, 72 insertions(+)
>
> diff --git a/drivers/iio/adc/ad7606.c b/drivers/iio/adc/ad7606.c
> index e0a666cc0e14255754e74daa9e1e88bc4ad1665c..22dcb52ced57d4305db6401605c064fc438d5be4 100644
> --- a/drivers/iio/adc/ad7606.c
> +++ b/drivers/iio/adc/ad7606.c
> @@ -33,6 +33,10 @@
>
> #include "ad7606.h"
>
> +#define AD7606_CALIB_GAIN_MIN 0
> +#define AD7606_CALIB_GAIN_STEP 1024
> +#define AD7606_CALIB_GAIN_MAX (63 * AD7606_CALIB_GAIN_STEP)
> +
> /*
> * Scales are computed as 5000/32768 and 10000/32768 respectively,
> * so that when applied to the raw values they provide mV values.
> @@ -125,6 +129,7 @@ static int ad7609_chan_scale_setup(struct iio_dev *indio_dev,
> struct iio_chan_spec *chan);
> static int ad7616_sw_mode_setup(struct iio_dev *indio_dev);
> static int ad7606b_sw_mode_setup(struct iio_dev *indio_dev);
> +static int ad7606_chan_calib_gain_setup(struct iio_dev *indio_dev);
>
> const struct ad7606_chip_info ad7605_4_info = {
> .max_samplerate = 300 * KILO,
> @@ -180,6 +185,7 @@ const struct ad7606_chip_info ad7606b_info = {
> .scale_setup_cb = ad7606_16bit_chan_scale_setup,
> .sw_setup_cb = ad7606b_sw_mode_setup,
> .offload_storagebits = 32,
> + .calib_gain_setup_cb = ad7606_chan_calib_gain_setup,
> .calib_offset_avail = ad7606_calib_offset_avail,
> .calib_phase_avail = ad7606b_calib_phase_avail,
> };
> @@ -195,6 +201,7 @@ const struct ad7606_chip_info ad7606c_16_info = {
> .scale_setup_cb = ad7606c_16bit_chan_scale_setup,
> .sw_setup_cb = ad7606b_sw_mode_setup,
> .offload_storagebits = 32,
> + .calib_gain_setup_cb = ad7606_chan_calib_gain_setup,
> .calib_offset_avail = ad7606_calib_offset_avail,
> .calib_phase_avail = ad7606c_calib_phase_avail,
> };
> @@ -246,6 +253,7 @@ const struct ad7606_chip_info ad7606c_18_info = {
> .scale_setup_cb = ad7606c_18bit_chan_scale_setup,
> .sw_setup_cb = ad7606b_sw_mode_setup,
> .offload_storagebits = 32,
> + .calib_gain_setup_cb = ad7606_chan_calib_gain_setup,
> .calib_offset_avail = ad7606c_18bit_calib_offset_avail,
> .calib_phase_avail = ad7606c_calib_phase_avail,
> };
> @@ -355,6 +363,36 @@ static int ad7606_get_chan_config(struct iio_dev *indio_dev, int ch,
> return 0;
> }
>
> +static int ad7606_chan_calib_gain_setup(struct iio_dev *indio_dev)
> +{
> + struct ad7606_state *st = iio_priv(indio_dev);
> + unsigned int num_channels = st->chip_info->num_adc_channels;
> + struct device *dev = st->dev;
> + int ret;
> +
> + device_for_each_child_node_scoped(dev, child) {
Now that I had a deep dive into this in v7, I'm wondering what is the
benefit of having a separate function and iterating over child nodes
a second time compared to just having a bool flag in chip_info and
doing this in the existing iterator in ad7606_probe_channels().
It seems like we could do the same thing in much fewer lines of code
if we avoid adding a callback.
> + u32 reg, r_gain;
> +
> + ret = fwnode_property_read_u32(child, "reg", ®);
> + if (ret)
> + return ret;
> +
> + /* Chan reg is a 1-based index. */
> + if (reg < 1 || reg > num_channels)
> + return -EINVAL;
> +
> + r_gain = 0;
> + ret = fwnode_property_read_u32(child, "adi,rfilter-ohms",
> + &r_gain);
ret is set but never read here.
> + if (r_gain > AD7606_CALIB_GAIN_MAX)
> + return -EINVAL;
> +
> + st->r_gain[reg - 1] = r_gain;
> + }
> +
> + return 0;
> +}
> +
> static int ad7606c_18bit_chan_scale_setup(struct iio_dev *indio_dev,
> struct iio_chan_spec *chan)
> {
> @@ -1352,6 +1390,21 @@ static int ad7606b_sw_mode_setup(struct iio_dev *indio_dev)
> return st->bops->sw_mode_config(indio_dev);
> }
>
> +static int ad7606_set_gain_calib(struct ad7606_state *st)
> +{
> + int i, ret;
> +
> + for (i = 0; i < st->chip_info->num_adc_channels; i++) {
> + ret = st->bops->reg_write(st, AD7606_CALIB_GAIN(i),
> + DIV_ROUND_CLOSEST(st->r_gain[i],
> + AD7606_CALIB_GAIN_STEP));
I think typical kernel style would be to have at least one more tab
on this line (assuming aligning to "(" makes the line too long).
> + if (ret)
> + return ret;
> + }
> +
> + return 0;
> +}
> +
> static int ad7606_probe_channels(struct iio_dev *indio_dev)
> {
> struct ad7606_state *st = iio_priv(indio_dev);
> @@ -1444,6 +1497,13 @@ static int ad7606_probe_channels(struct iio_dev *indio_dev)
> if (slow_bus)
> channels[i] = (struct iio_chan_spec)IIO_CHAN_SOFT_TIMESTAMP(i);
>
> + /* Getting gain calibration values for all channels. */
> + if (st->sw_mode_en && st->chip_info->calib_gain_setup_cb) {
> + ret = st->chip_info->calib_gain_setup_cb(indio_dev);
> + if (ret)
> + return ret;
> + }
> +
> indio_dev->channels = channels;
>
> return 0;
> @@ -1620,6 +1680,12 @@ int ad7606_probe(struct device *dev, int irq, void __iomem *base_address,
> st->chip_info->sw_setup_cb(indio_dev);
> }
>
> + if (st->sw_mode_en && st->chip_info->calib_gain_setup_cb) {
> + ret = ad7606_set_gain_calib(st);
> + if (ret)
> + return ret;
> + }
> +
> return devm_iio_device_register(dev, indio_dev);
> }
> EXPORT_SYMBOL_NS_GPL(ad7606_probe, "IIO_AD7606");
> diff --git a/drivers/iio/adc/ad7606.h b/drivers/iio/adc/ad7606.h
> index f613583a7fa4095115b0b28e3f8e51cd32b93524..a5b0d318e2f4d73d3708288536e807957c5de68c 100644
> --- a/drivers/iio/adc/ad7606.h
> +++ b/drivers/iio/adc/ad7606.h
> @@ -50,6 +50,7 @@ struct ad7606_state;
> typedef int (*ad7606_scale_setup_cb_t)(struct iio_dev *indio_dev,
> struct iio_chan_spec *chan);
> typedef int (*ad7606_sw_setup_cb_t)(struct iio_dev *indio_dev);
> +typedef int (*ad7606_calib_gain_setup_cb_t)(struct iio_dev *indio_dev);
>
> /**
> * struct ad7606_chip_info - chip specific information
> @@ -66,6 +67,7 @@ typedef int (*ad7606_sw_setup_cb_t)(struct iio_dev *indio_dev);
> * @init_delay_ms: required delay in milliseconds for initialization
> * after a restart
> * @offload_storagebits: storage bits used by the offload hw implementation
> + * @calib_gain_setup_cb: callback to setup of gain calibration
> * @calib_offset_avail: pointer to offset calibration range/limits array
> * @calib_phase_avail: pointer to phase calibration range/limits array
> */
> @@ -81,6 +83,7 @@ struct ad7606_chip_info {
> bool os_req_reset;
> unsigned long init_delay_ms;
> u8 offload_storagebits;
> + ad7606_calib_gain_setup_cb_t calib_gain_setup_cb;
> const int *calib_offset_avail;
> const int (*calib_phase_avail)[2];
> };
> @@ -131,6 +134,7 @@ struct ad7606_chan_scale {
> * @data: buffer for reading data from the device
> * @offload_en: SPI offload enabled
> * @bus_data: bus-specific variables
> + * @r_gain: array to store gain calibration resistor value in ohm
> * @d16: be16 buffer for reading data from the device
> */
> struct ad7606_state {
> @@ -161,6 +165,8 @@ struct ad7606_state {
> bool offload_en;
> void *bus_data;
>
> + int r_gain[AD760X_MAX_CHANNELS];
This isn't used outside of probe, so we could possibly get away with
putting it somewhere else.
> +
> /*
> * DMA (thus cache coherency maintenance) may require the
> * transfer buffers to live in their own cache lines.
>
^ permalink raw reply [flat|nested] 9+ messages in thread