1FENV(3) Linux Programmer's Manual FENV(3)
2
6 feclearexcept, fegetexceptflag, feraiseexcept, fesetexceptflag, fetes‐
7 texcept, fegetenv, fegetround, feholdexcept, fesetround, fesetenv,
8 feupdateenv - C99 floating point rounding and exception handling
9
11 #include <fenv.h>
12 int feclearexcept(int excepts);
14 int fegetexceptflag(fexcept_t *flagp, int excepts);
15 int feraiseexcept(int excepts);
16 int fesetexceptflag(const fexcept_t *flagp, int excepts);
17 int fetestexcept(int excepts);
18 int fegetround(void);
20 int fesetround(int rounding_mode);
21 int fegetenv(fenv_t *envp);
23 int feholdexcept(fenv_t *envp);
24 int fesetenv(const fenv_t *envp);
25 int feupdateenv(const fenv_t *envp);
26
28 These eleven functions were defined in C99, and describe the handling
29 of floating point rounding and exceptions (overflow, zero-divide etc.).
30 Exceptions
32 The DivideByZero exception occurs when an operation on finite numbers
33 produces infinity as exact answer.
34 The Overflow exception occurs when a result has to be represented as a
36 floating point number, but has (much) larger absolute value than the
37 largest (finite) floating point number that is representable.
38 The Underflow exception occurs when a result has to be represented as a
40 floating point number, but has smaller absolute value than the smallest
41 positive normalized floating point number (and would lose much accuracy
42 when represented as a denormalized number).
43 The Inexact exception occurs when the rounded result of an operation is
45 not equal to the infinite precision result. It may occur whenever
46 Overflow or Underflow occurs.
47 The Invalid exception occurs when there is no well-defined result for
49 an operation, as for 0/0 or infinity - infinity or sqrt(-1).
50 Exception handling
52 Exceptions are represented in two ways: as a single bit (exception
53 present/absent), and these bits correspond in some implementation-
54 defined way with bit positions in an integer, and also as an opaque
55 structure that may contain more information about the exception (per‐
56 haps the code address where it occurred).
57 Each of the macros FE_DIVBYZERO, FE_INEXACT, FE_INVALID, FE_OVERFLOW,
59 FE_UNDERFLOW is defined when the implementation supports handling of
60 the corresponding exception, and if so then defines the corresponding
61 bit(s), so that one can call exception handling functions e.g. using
62 the integer argument FE_OVERFLOW|FE_UNDERFLOW. Other exceptions may be
63 supported. The macro FE_ALL_EXCEPT is the bitwise OR of all bits corre‐
64 sponding to supported exceptions.
65 The feclearexcept() function clears the supported exceptions repre‐
67 sented by the bits in its argument.
68 The fegetexceptflag() function stores a representation of the state of
70 the exception flags represented by the argument excepts in the opaque
71 object *flagp.
72 The feraiseexcept() function raises the supported exceptions repre‐
74 sented by the bits in excepts.
75 The fesetexceptflag() function sets the complete status for the excep‐
77 tions represented by excepts to the value *flagp. This value must have
78 been obtained by an earlier call of fegetexceptflag() with a last argu‐
79 ment that contained all bits in excepts.
80 The fetestexcept() function returns a word in which the bits are set
82 that were set in the argument excepts and for which the corresponding
83 exception is currently set.
84 Rounding
86 Each of the macros FE_DOWNWARD, FE_TONEAREST, FE_TOWARDZERO, FE_UPWARD
87 is defined when the implementation supports getting and setting the
88 corresponding rounding direction.
89 The fegetround() function returns the macro corresponding to the cur‐
91 rent rounding mode.
92 The fesetround() function sets the rounding mode as specified by its
94 argument and returns zero when it was successful.
95 Floating point environment
97 The entire floating point environment, including control modes and sta‐
98 tus flags, can be handled as one opaque object, of type fenv_t. The
99 default environment is denoted by FE_DFL_ENV (of type const fenv_t *).
100 This is the environment setup at program start and it is defined by ISO
101 C to have round to nearest, all exceptions cleared and a non-stop (con‐
102 tinue on exceptions) mode.
103 The fegetenv() function saves the current floating point environment in
105 the object *envp.
106 The feholdexcept() function does the same, then clears all exception
108 flags, and sets a non-stop (continue on exceptions) mode, if available.
109 It returns zero when successful.
110 The fesetenv() function restores the floating point environment from
112 the object *envp. This object must be known to be valid, e.g., the
113 result of a call to fegetenv() or feholdexcept() or equal to
114 FE_DFL_ENV. This call does not raise exceptions.
115 The feupdateenv() function installs the floating-point environment rep‐
117 resented by the object *envp, except that currently raised exceptions
118 are not cleared. After calling this function, the raised exceptions
119 will be a bitwise OR of those previously set with those in *envp. As
120 before, the object *envp must be known to be valid.
121
123 These functions return zero on success and non-zero if an error
124 occurred.
125
127 If possible, the GNU C Library defines a macro FE_NOMASK_ENV which rep‐
128 resents an environment where every exception raised causes a trap to
129 occur. You can test for this macro using #ifdef. It is only defined
130 if _GNU_SOURCE is defined. The C99 standard does not define a way to
131 set individual bits in the floating point mask, e.g. to trap on spe‐
132 cific flags. glibc 2.2 supports the functions feenableexcept() and
133 fedisableexcept() to set individual floating point traps, and fegetex‐
134 cept() to query the state.
135 #define _GNU_SOURCE
137 #include <fenv.h>
138 int feenableexcept (int excepts);
140 int fedisableexcept (int excepts);
141 int fegetexcept (void);
142 The feenableexcept() and fedisableexcept() functions enable (disable)
144 traps for each of the exceptions represented by excepts and return the
145 previous set of enabled exceptions when successful, and -1 otherwise.
146 The fegetexcept() function returns the set of all currently enabled
147 exceptions.
148
150 Link with -lm.
151
153 IEC 60559 (IEC 559:1989), ANSI/IEEE 854, C99.
154
156 feature_test_macros(7)
157Linux Manpage 2000-08-12 FENV(3)