1 .\" Copyright (c) 1993 Andrew Moore, Talke Studio
2 .\" All rights reserved.
4 .\" Redistribution and use in source and binary forms, with or without
5 .\" modification, are permitted provided that the following conditions
7 .\" 1. Redistributions of source code must retain the above copyright
8 .\" notice, this list of conditions and the following disclaimer.
9 .\" 2. Redistributions in binary form must reproduce the above copyright
10 .\" notice, this list of conditions and the following disclaimer in the
11 .\" documentation and/or other materials provided with the distribution.
12 .\" 3. All advertising materials mentioning features or use of this software
13 .\" must display the following acknowledgement:
14 .\" This product includes software developed by the University of
15 .\" California, Berkeley and its contributors.
16 .\" 4. Neither the name of the University nor the names of its contributors
17 .\" may be used to endorse or promote products derived from this software
18 .\" without specific prior written permission.
20 .\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 .\" @(#)fpgetround.3 1.0 (Berkeley) 9/23/93
47 .Nd IEEE floating point interface
52 FP_RN, /* round to nearest */
53 FP_RM, /* round down to minus infinity */
54 FP_RP, /* round up to plus infinity */
61 .Fn fpsetround "fp_rnd_t direction"
64 FP_PS, /* 24 bit (single-precision) */
65 FP_PRS, /* reserved */
66 FP_PD, /* 53 bit (double-precision) */
67 FP_PE /* 64 bit (extended-precision) */
73 .Fn fpsetprec "fp_prec_t precision"
75 #define fp_except_t int
76 #define FP_X_INV 0x01 /* invalid operation */
77 #define FP_X_DNML 0x02 /* denormal */
78 #define FP_X_DZ 0x04 /* zero divide */
79 #define FP_X_OFL 0x08 /* overflow */
80 #define FP_X_UFL 0x10 /* underflow */
81 #define FP_X_IMP 0x20 /* (im)precision */
82 #define FP_X_STK 0x40 /* stack fault */
87 .Fn fpsetmask "fp_except_t mask"
91 .Fn fpresetsticky "fp_except_t sticky"
93 The routines described herein are deprecated.
94 New code should use the functionality provided by
97 When a floating point exception is detected, the exception sticky flag is
98 set and the exception mask is tested.
99 If the mask is set, then a trap
101 These routines allow both setting the floating point exception
102 masks, and resetting the exception sticky flags after an exception is
104 In addition, they allow setting the floating point rounding mode
110 returns the current floating point rounding mode.
115 sets the floating point rounding mode and returns
121 returns the current floating point precision.
126 sets the floating point precision and returns
127 the previous precision.
132 returns the current floating point exception masks.
137 sets the floating point exception masks and returns the
143 returns the current floating point sticky flags.
148 clears the floating point sticky flags and returns
151 Sample code which prevents a trap on divide-by-zero:
152 .Bd -literal -offset indent
157 fpresetsticky(FP_X_DZ);
160 .Sh IMPLEMENTATION NOTES
165 functions provide functionality unavailable on many platforms.
166 At present, they are implemented only on the i386 and amd64 platforms.
167 Changing precision isn't a supported feature:
168 it may be ineffective when code is compiled to take advantage of SSE,
169 and many library functions and compiler optimizations depend upon the
170 default precision for correct behavior.
175 These routines are based on SysV/386 routines of the same name.
177 After a floating point exception and before a mask is set, the sticky
179 If another exception occurs before the sticky
180 flags are reset, then a wrong exception type may be signaled.