xref: /minix/lib/libm/src/s_rintl.c (revision 00b67f09)
1 /*	$NetBSD: s_rintl.c,v 1.5 2013/08/21 13:04:44 martin Exp $	*/
2 
3 /*-
4  * Copyright (c) 2008 David Schultz <das@FreeBSD.ORG>
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  *
16  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26  * SUCH DAMAGE.
27  */
28 
29 #include <sys/cdefs.h>
30 #if 0
31 __FBSDID("$FreeBSD: src/lib/msun/src/s_rintl.c,v 1.5 2008/02/22 11:59:05 bde Exp $");
32 #else
33 __RCSID("$NetBSD: s_rintl.c,v 1.5 2013/08/21 13:04:44 martin Exp $");
34 #endif
35 
36 #include <float.h>
37 #include <machine/ieee.h>
38 
39 #include "math.h"
40 #include "math_private.h"
41 
42 #ifdef __HAVE_LONG_DOUBLE
43 static const float
44 shift[2] = {
45 #if EXT_FRACBITS == 64
46 	0x1.0p63, -0x1.0p63
47 #elif EXT_FRACBITS == 113
48 	0x1.0p112, -0x1.0p112
49 #elif EXT_FRACBITS == 112
50 	0x1.0p111, -0x1.0p111
51 #else
52 #error "Unsupported long double format"
53 #endif
54 };
55 static const float zero[2] = { 0.0, -0.0 };
56 
57 long double
58 rintl(long double x)
59 {
60 	union ieee_ext_u u;
61 	uint32_t expsign;
62 	int ex, sign;
63 
64 	u.extu_ld = x;
65 	u.extu_ext.ext_frach &= ~0x80000000;
66 	expsign = u.extu_ext.ext_sign;
67 	ex = expsign & 0x7fff;
68 
69 	if (ex >= EXT_EXP_BIAS + EXT_FRACBITS - 1) {
70 		if (ex == EXT_EXP_BIAS + EXT_FRACBITS)
71 			return (x + x);	/* Inf, NaN, or unsupported format */
72 		return (x);		/* finite and already an integer */
73 	}
74 	sign = expsign >> 15;
75 
76 	/*
77 	 * The following code assumes that intermediate results are
78 	 * evaluated in long double precision. If they are evaluated in
79 	 * greater precision, double rounding may occur, and if they are
80 	 * evaluated in less precision (as on i386), results will be
81 	 * wildly incorrect.
82 	 */
83 	x += shift[sign];
84 	x -= shift[sign];
85 
86 	/*
87 	 * If the result is +-0, then it must have the same sign as x, but
88 	 * the above calculation doesn't always give this.  Fix up the sign.
89 	 */
90 	if (ex < EXT_EXP_BIAS && x == 0.0L)
91 		return (zero[sign]);
92 
93 	return (x);
94 }
95 #endif
96