The fixed point math API library module has the file name fxp.c. This module performs basic math operations and comparisons on floating point numbers using fixed point math. Floating point numbers are input in string form and stored as long integers. The position of the decimal point, the number as a string and the current value as a long are stored in a structure called fxp. The current position of the decimal point from the right side of the string is referred to in this document as the scale.
The calculation range of an fxp structure is the same as a signed long integer. All fxp structures are dynamically allocated and must be de-allocated using fxp_free only. This API has been coded to make use of the log manager API if it has been enabled.
Please see my floating point polemic paper for reasons why I use fixed point arithmetic and never use floating point.
Here is a list of functions in the fixed-point math API library module:
The following modules are required along with this module:
This module requires the header file:
Prototype : struct fxp *fxp_iconv(char *cval)
Parameters :
Name : cval
Description: floating-point number in string form
Returns : a pointer to a newly allocated fxp structure
This function will perform an input conversion on a floating point number. This function will dynamically create an fxp structure. It is the responsibility of the caller to de-allocate the structure after use.
Prototype : int fxp_oconv(struct fxp *tfxp, int tscale, char *outstr)
Parameters :
Name : tfxp
Description: fixed point input
Name : tscale
Description: output scale
Name : outstr
Description: output result
Returns : TRUE upon success, FALSE otherwise
This function will perform an output conversion on a fixed point number. Upon success, the string output result will be loaded with the result. An attempt is made to round the result to the precision specified by the output scale.
Prototype : struct fxp *fxp_add(struct fxp *ad1, struct fxp *ad2)
Parameters :
Name : ad1
Description: first operand
Name : ad2
Description: second operand
Returns : a pointer to a newly allocated fxp structure
This function will add two fixed point numbers. A new fxp structure will be dynamically allocated for the result. This must be de-allocated by the caller.
Prototype : struct fxp *fxp_sub(struct fxp *ad1, struct fxp *ad2)
Parameters :
Name : ad1
Description: first operand
Name : ad2
Description: second operand
Returns : a pointer to a newly allocated fxp structure
This function will subtract two fixed point numbers. A new fxp structure will be dynamically allocated for the result. This must be de-allocated by the caller. Note that the second operand is subtracted from the first operand.
Prototype : struct fxp *fxp_mul(struct fxp *ad1, struct fxp *ad2)
Parameters :
Name : ad1
Description: first operand
Name : ad2
Description: second operand
Returns : a pointer to a newly allocated fxp structure
This function will multiply two fixed point numbers. A new fxp structure will be dynamically allocated for the result. This must be de-allocated by the caller. Note that the first operand is multiplied by the second operand.
Prototype : struct fxp *fxp_div(struct fxp *ad1, struct fxp *ad2)
Parameters :
Name : ad1
Description: first operand
Name : ad2
Description: second operand
Returns : a pointer to a newly allocated fxp structure
This function will divide two fixed point numbers. A new fxp structure will be dynamically allocated for the result. This must be de-allocated by the caller. Note that the first operand is divided by the second operand.
Prototype : int fxp_eq(struct fxp *fxp1, struct fxp *fxp2)
Parameters :
Name : fxp1
Description: first operand
Name : fxp2
Description: second operand
Returns : TRUE if equal, FALSE otherwise
This function will perform an equality test of two fixed point numbers.
Prototype : int fxp_lt(struct fxp *fxp1, struct fxp *fxp2)
Parameters :
Name : fxp1
Description: first operand
Name : fxp2
Description: second operand
Returns : TRUE if first operand is less than
second operand, FALSE otherwise
This function will determine whether one fixed point number is less than the other.
Prototype : int fxp_gt(struct fxp *fxp1, struct fxp *fxp2)
Parameters :
Name : fxp1
Description: first operand
Name : fxp2
Description: second operand
Returns : TRUE if first operand is greater than
second operand, FALSE otherwise
This function will determine whether one fixed point number is greater than the other.
Prototype : int fxp_le(struct fxp *fxp1, struct fxp *fxp2)
Parameters :
Name : fxp1
Description: first operand
Name : fxp2
Description: second operand
Returns : TRUE if first operand is less than or equal to the
second operand, FALSE otherwise
This function will determine whether one fixed point number is less than or equal to the other.
Prototype : int fxp_ge(struct fxp *fxp1, struct fxp *fxp2)
Parameters :
Name : fxp1
Description: first operand
Name : fxp2
Description: second operand
Returns : TRUE if first operand is greater than or equal to
the second operand, FALSE otherwise
This function will determine whether one fixed point number is greater than or equal to the other.
Prototype : static struct fxp *fxp_math(struct fxp *ad1, struct fxp *ad2,
int opcode)
Parameters :
Name : ad1
Description: first operand
Name : ad2
Description: second operand
Name : opcode
Description: math operation code
Returns : a pointer to a newly allocated fxp structure
This private function will perform math on two fixed point numbers and return the result. A new fxp structure will be dynamically allocated for the result. This must be de-allocated by the caller.
Prototype : static struct fxp *fxp_comp(struct fxp *ad1, struct fxp *ad2,
int opcode)
Parameters :
Name : ad1
Description: first operand
Name : ad2
Description: second operand
Name : opcode
Description: comparison operation code
Returns : TRUE upon successful comparison, FALSE otherwise
This private function will perform a comparison of two fixed point numbers and return the result.
Prototype : int fxp_scale(struct fxp *tfxp, int tscale)
Parameters :
Name : tfxp
Description: input fixed-point number
Name : tscale
Description: desired scale/precision
Returns : TRUE upon success, FALSE otherwise
This function will change the scale of the input fixed point number. Digits will be added or taken away from the right side of the number.
Prototype : void fxp_free(struct fxp *tfxp)
Parameters :
Name : tfxp
Description: input fixed-point number
This function will de-allocate all dynamic memory used by a single fxp structure.
Prototype : struct fxp *fxp_dup(struct fxp *tfxp)
Parameters :
Name : tfxp
Description: input fixed-point number
Returns : a pointer to a newly allocated fxp structure
This function will duplicate an fxp structure.
Prototype : static struct fxp *fxp_new(void) Returns : a pointer to a newly allocated fxp structure
This private function will create a new fxp structure. Structure will be initialized.
Prototype : void fxp_debug(struct fxp *f1)
Parameters :
Name : f1
Description: input fixed-point number
This function will output the contents of an fxp structure to the logging manager. No data will be output unless the logging manager has been started by calling the function logman_start.
Prototype : void fxp_remove_decimal(char *cfin, char *cfout)
Parameters :
Name : cfin
Description: input floating-point number in string form
Name : cfout
Description: output number in string form
This function will copy the input floating-point number in string form removing the decimal point. Function returns the input floating-point number in string form (minus the decimal point) in the output number in string form upon success.