diff options
Diffstat (limited to 'usr/src/lib/libm/common/Q/rintl.c')
-rw-r--r-- | usr/src/lib/libm/common/Q/rintl.c | 75 |
1 files changed, 75 insertions, 0 deletions
diff --git a/usr/src/lib/libm/common/Q/rintl.c b/usr/src/lib/libm/common/Q/rintl.c new file mode 100644 index 0000000000..a721e4231b --- /dev/null +++ b/usr/src/lib/libm/common/Q/rintl.c @@ -0,0 +1,75 @@ +/* + * CDDL HEADER START + * + * The contents of this file are subject to the terms of the + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. + * + * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE + * or http://www.opensolaris.org/os/licensing. + * See the License for the specific language governing permissions + * and limitations under the License. + * + * When distributing Covered Code, include this CDDL HEADER in each + * file and include the License file at usr/src/OPENSOLARIS.LICENSE. + * If applicable, add the following below this CDDL HEADER, with the + * fields enclosed by brackets "[]" replaced with your own identifying + * information: Portions Copyright [yyyy] [name of copyright owner] + * + * CDDL HEADER END + */ + +/* + * Copyright 2011 Nexenta Systems, Inc. All rights reserved. + */ +/* + * Copyright 2006 Sun Microsystems, Inc. All rights reserved. + * Use is subject to license terms. + */ + +#pragma weak rintl = __rintl + +/* + * rintl(long double x) return x rounded to integral according to + * the prevailing rounding direction + * + * NOTE: aintl(x), anintl(x), ceill(x), floorl(x), and rintl(x) return result + * with the same sign as x's, including 0.0L. + */ + +#include "libm.h" +#include "longdouble.h" + +extern enum fp_precision_type __swapRP(enum fp_precision_type); + +static const double one = 1.0; +static const long double qzero = 0.0L; + +long double +rintl(long double x) { + enum fp_precision_type rp; + long double t, w, two112; + int *pt = (int *) &two112; + + if (!finitel(x)) + return (x + x); + + if (*(int *) &one != 0) { /* set two112 = 2^112 */ + pt[0] = 0x406f0000; + pt[1] = pt[2] = pt[3] = 0; + } else { + pt[3] = 0x406f0000; + pt[0] = pt[1] = pt[2] = 0; + } + + if (fabsl(x) >= two112) + return (x); /* already an integer */ + t = copysignl(two112, x); + rp = __swapRP(fp_extended); /* make sure precision is long double */ + w = x + t; /* x+sign(x)*2^112 rounded to integer */ + (void) __swapRP(rp); /* restore precision mode */ + if (w == t) + return (copysignl(qzero, x)); /* x rounded to zero */ + else + return (w - t); +} |