Added some primative support for using software calibration in a way

that can be integrated with boards that don't use software calibrations.

include/comedilib.h

@@ -205,7 +205,7 @@ typedef struct
 	unsigned int channel;
 	unsigned int value;
 } comedi_caldac_t;
-#define COMEDI_MAX_NUM_POLYNOMIAL_COEFFICIENTS (4)
+#define COMEDI_MAX_NUM_POLYNOMIAL_COEFFICIENTS 4
 typedef struct
 {
 	double coefficients[COMEDI_MAX_NUM_POLYNOMIAL_COEFFICIENTS];
@@ -248,7 +248,24 @@ void comedi_cleanup_calibration( comedi_calibration_t *calibration );
 int comedi_apply_calibration( comedi_t *dev, unsigned int subdev, unsigned int channel,
 	unsigned int range, unsigned int aref, const char *cal_file_path);
 
-
+/* New stuff to provide conversion between integers and physical values that
+* can support software calibrations. */
+enum comedi_conversion_direction
+{
+	COMEDI_TO_PHYSICAL,
+	COMEDI_FROM_PHYSICAL
+};
+int comedi_get_softcal_converter(
+	unsigned subdevice, unsigned channel, unsigned range,
+	enum comedi_conversion_direction direction,
+	const comedi_calibration_t *calibration, comedi_polynomial_t* polynomial);
+int comedi_get_hardcal_converter(
+	comedi_t *dev, unsigned subdevice, unsigned channel, unsigned range,
+	enum comedi_conversion_direction direction, comedi_polynomial_t* polynomial);
+double comedi_to_physical(lsampl_t data,
+	const comedi_polynomial_t *conversion_polynomial);
+lsampl_t comedi_from_physical(double data,
+	const comedi_polynomial_t *conversion_polynomial);
 
 #endif
 

lib/calib.c

@@ -22,6 +22,8 @@
 
 #define _GNU_SOURCE
 
+#include <assert.h>
+#include <math.h>
 #include <stdlib.h>
 #include <stdio.h>
 #include <string.h>
@@ -242,3 +244,99 @@ int _comedi_apply_calibration( comedi_t *dev, unsigned int subdev, unsigned int
 
 	return retval;
 }
+
+EXPORT_ALIAS_DEFAULT(_comedi_get_hardcal_converter, comedi_get_hardcal_converter, 0.7.23);
+int _comedi_get_hardcal_converter(
+	comedi_t *dev, unsigned subdevice, unsigned channel, unsigned range,
+	enum comedi_conversion_direction direction,
+	comedi_polynomial_t* polynomial)
+{
+	comedi_range *range_ptr = comedi_get_range(dev, subdevice, channel, range);
+	if(range_ptr == NULL)
+	{
+		return -1;
+	}
+	lsampl_t maxdata = comedi_get_maxdata(dev, subdevice, channel);
+	if(maxdata == 0)
+	{
+		return -1;
+	}
+	polynomial->order = 1;
+	switch(direction)
+	{
+	case COMEDI_TO_PHYSICAL:
+		polynomial->expansion_origin = 0.;
+		polynomial->coefficients[0] = range_ptr->min;
+		polynomial->coefficients[1] = (range_ptr->max - range_ptr->min) / maxdata;
+		break;
+	case COMEDI_FROM_PHYSICAL:
+		polynomial->expansion_origin = range_ptr->min;
+		polynomial->coefficients[0] = 0.;
+		polynomial->coefficients[1] =  maxdata / (range_ptr->max - range_ptr->min);
+		break;
+	}
+	return 0;
+}
+
+EXPORT_ALIAS_DEFAULT(_comedi_get_softcal_converter, comedi_get_softcal_converter, 0.7.23);
+int _comedi_get_softcal_converter(
+	unsigned subdevice, unsigned channel, unsigned range,
+	enum comedi_conversion_direction direction,
+	const comedi_calibration_t *calibration, comedi_polynomial_t* polynomial)
+{
+	unsigned i;
+
+	for(i = 0; i < calibration->num_settings; ++i)
+	{
+		if(calibration->settings[i].subdevice != subdevice) continue;
+		if(valid_channel(calibration, i, channel) == 0) continue;
+		if(valid_range(calibration, i, range) == 0) continue;
+		switch(direction)
+		{
+		case COMEDI_TO_PHYSICAL:
+			if(calibration->settings[i].soft_calibration.to_phys == NULL)
+			{
+				continue;
+			}
+			*polynomial = *calibration->settings[i].soft_calibration.to_phys;
+			break;
+		case COMEDI_FROM_PHYSICAL:
+			if(calibration->settings[i].soft_calibration.from_phys == NULL)
+			{
+				continue;
+			}
+			*polynomial = *calibration->settings[i].soft_calibration.from_phys;
+			break;
+		}
+		return 0;
+	}
+	return -1;
+}
+
+static double apply_polynomial(const comedi_polynomial_t *polynomial, double input)
+{
+	double value = 0.;
+	double term = 1.;
+	unsigned i;
+	assert(polynomial->order < COMEDI_MAX_NUM_POLYNOMIAL_COEFFICIENTS);
+	for(i = 0; i <= polynomial->order; ++i)
+	{
+		value += polynomial->coefficients[i] * term;
+		term *= input - polynomial->expansion_origin;
+	}
+	return value;
+}
+
+EXPORT_ALIAS_DEFAULT(_comedi_to_physical, comedi_to_physical, 0.7.23);
+double _comedi_to_physical(lsampl_t data,
+	const comedi_polynomial_t *conversion_polynomial)
+{
+	return apply_polynomial(conversion_polynomial, data);
+}
+
+EXPORT_ALIAS_DEFAULT(_comedi_from_physical, comedi_from_physical, 0.7.23);
+lsampl_t _comedi_from_physical(double data,
+	const comedi_polynomial_t *conversion_polynomial)
+{
+	return nearbyint(apply_polynomial(conversion_polynomial, data));
+}

lib/version_script

@@ -83,7 +83,11 @@ v0.7.20 {
 
 v0.7.23 {
 	global:
+		comedi_from_physical;
 		comedi_dio_get_config;
+		comedi_get_hardcal_converter;
+		comedi_get_softcal_converter;
 		comedi_get_subdevice_flags;
 		comedi_mark_buffer_written;
+		comedi_to_physical;
 } v0.7.20;