floatingpoint(3) UNIX System V(BSD Compatibility Package) floatingpoint(3)
NAME
floatingpoint - IEEE floating point definitions
SYNOPSIS
cc [ flag. . . ] file . . . -lucb
#include <sys/ieeefp.h>
#include <fp.h>
DESCRIPTION
This file defines constants, types, variables, and functions used to
implement standard floating point according to ANSI/IEEE Std 754-1985.
The variables and functions are implemented in libucb.a. The included
file <sys/ieeefp.h> defines certain types of interest to the kernel.
IEEE Rounding Modes:
fpdirectiontype
The type of the IEEE rounding direction mode. Note: the
order of enumeration varies according to hardware.
fpdirection The IEEE rounding direction mode currently in force.
This is a global variable that is intended to reflect
the hardware state, so it should only be written
indirectly through a function that also sets the
hardware state.
fpprecisiontype
The type of the IEEE rounding precision mode, which only
applies on systems that support extended precision.
fpprecision The IEEE rounding precision mode currently in force.
This is a global variable that is intended to reflect
the hardware state on systems with extended precision,
so it should only be written indirectly.
SIGFPE Handling:
sigfpecodetype The type of a SIGFPE code.
sigfpehandlertype
The type of a user-definable SIGFPE exception handler
called to handle a particular SIGFPE code.
SIGFPEDEFAULT A macro indicating the default SIGFPE exception handling,
namely to perform the exception handling specified by
calls to ieeehandler(3M), if any, and otherwise to dump
core using abort(3).
SIGFPEIGNORE A macro indicating an alternate SIGFPE exception
handling, namely to ignore and continue execution.
10/89 Page 1
floatingpoint(3) UNIX System V(BSD Compatibility Package) floatingpoint(3)
SIGFPEABORT A macro indicating an alternate SIGFPE exception
handling, namely to abort with a core dump.
IEEE Exception Handling:
NIEEEEXCEPTION The number of distinct IEEE floating-point exceptions.
fpexceptiontype
The type of the NIEEEEXCEPTION exceptions. Each
exception is given a bit number.
fpexceptionfieldtype
The type intended to hold at least NIEEEEXCEPTION bits
corresponding to the IEEE exceptions numbered by
fpexceptiontype. Thus fpinexact corresponds to the
least significant bit and fpinvalid to the fifth least
significant bit. Note: some operations may set more
than one exception.
fpaccruedexceptions
The IEEE exceptions between the time this global
variable was last cleared, and the last time a function
was called to update the variable by obtaining the
hardware state.
ieeehandlers An array of user-specifiable signal handlers for use by
the standard SIGFPE handler for IEEE arithmetic-related
SIGFPE codes. Since IEEE trapping modes correspond to
hardware modes, elements of this array should only be
modified with a function like ieeehandler(3M) that
performs the appropriate hardware mode update. If no
sigfpehandler has been declared for a particular IEEE-
related SIGFPE code, then the related ieeehandlers will
be invoked.
IEEE Formats and Classification:
single;extended Definitions of IEEE formats.
fpclasstype An enumeration of the various classes of IEEE values and
symbols.
IEEE Base Conversion:
The functions described under floatingtodecimal(3) and
decimaltofloating(3) not only satisfy the IEEE Standard, but also the
stricter requirements of correct rounding for all arguments.
DECIMALSTRINGLENGTH
The length of a decimalstring.
Page 2 10/89
floatingpoint(3) UNIX System V(BSD Compatibility Package) floatingpoint(3)
decimalstring The digit buffer in a decimalrecord.
decimalrecord The canonical form for representing an unpacked decimal
floating-point number.
decimalform The type used to specify fixed or floating binary to
decimal conversion.
decimalmode A struct that contains specifications for conversion
between binary and decimal.
decimalstringform
An enumeration of possible valid character strings
representing floating-point numbers, infinities, or
NaNs.
FILES
/usr/include/sys/ieeefp.h
/usr/include/fp.h
/usr/ucblib/libucb.a
SEE ALSO
decimaltofloating(3), econvert(3), floatingtodecimal(3),
ieeehandler(3M), sigfpe(3)
abort(3), strtod(3) in the Programmer's Reference Manual
10/89 Page 3