mcx_ieee754.h

Go to the documentation of this file.

/***************************************************************************/
/***************************************************************************/
#ifndef MCX_IEEE754_H
#define MCX_IEEE754_H

#define MCX_LITTLE_ENDIAN 0x41424344UL
#define MCX_BIG_ENDIAN    0x44434241UL
#define MCX_PDP_ENDIAN    0x42414443UL
#define MCX_BYTE_ORDER    ('ABCD')

union ieee754_double {
    double d;

    /* This is the IEEE 754 double-precision format.  */
    struct {
#if MCX_BYTE_ORDER == MCX_BIG_ENDIAN
        unsigned int negative: 1;
        unsigned int exponent: 11;
        /* Together these comprise the mantissa.  */
        unsigned int mantissa0: 20;
        unsigned int mantissa1: 32;
#else
        /* Together these comprise the mantissa.  */
        unsigned int mantissa1: 32;
        unsigned int mantissa0: 20;
        unsigned int exponent: 11;
        unsigned int negative: 1;
#endif              /* Little endian.  */
    } ieee;

    /* This format makes it easier to see if a NaN is a signalling NaN.  */
    struct {
#if MCX_BYTE_ORDER == MCX_BIG_ENDIAN
        unsigned int negative: 1;
        unsigned int exponent: 11;
        unsigned int quiet_nan: 1;
        /* Together these comprise the mantissa.  */
        unsigned int mantissa0: 19;
        unsigned int mantissa1: 32;
#else
        /* Together these comprise the mantissa.  */
        unsigned int mantissa1: 32;
        unsigned int mantissa0: 19;
        unsigned int quiet_nan: 1;
        unsigned int exponent: 11;
        unsigned int negative: 1;
#endif
    } ieee_nan;
};

#define IEEE754_DOUBLE_BIAS 0x3ff /* Added to exponent.  */

#endif