glibc/sysdeps/ieee754/flt-32/s_remquof.c
Joseph Myers a820f9b3c0 Fix remquo spurious overflows (bug 17978).
Various remquo implementations, when computing the last three bits of
the quotient, have spurious overflows when 4 times the second argument
to remquo overflows.  These overflows can in turn cause bad results in
rounding modes where that overflow results in a finite value.  This
patch adds tests to avoid the problem multiplications in cases where
they would overflow, similar to those that control an earlier
multiplication by 8.

Tested for x86_64, x86, mips64 and powerpc.

	[BZ #17978]
	* sysdeps/ieee754/dbl-64/s_remquo.c (__remquo): Do not form
	products 4 * y and 2 * y where those would overflow.
	* sysdeps/ieee754/dbl-64/wordsize-64/s_remquo.c (__remquo):
	Likewise.
	* sysdeps/ieee754/flt-32/s_remquof.c (__remquof): Likewise.
	* sysdeps/ieee754/ldbl-128/s_remquol.c (__remquol): Likewise.
	* sysdeps/ieee754/ldbl-128ibm/s_remquol.c (__remquol): Likewise.
	* sysdeps/ieee754/ldbl-96/s_remquol.c (__remquol): Likewise.
	* math/libm-test.inc (remquo_test_data): Add more tests.
2015-02-16 22:38:28 +00:00

108 lines
2.2 KiB
C

/* Compute remainder and a congruent to the quotient.
Copyright (C) 1997-2015 Free Software Foundation, Inc.
This file is part of the GNU C Library.
Contributed by Ulrich Drepper <drepper@cygnus.com>, 1997.
The GNU C Library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.
The GNU C Library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with the GNU C Library; if not, see
<http://www.gnu.org/licenses/>. */
#include <math.h>
#include <math_private.h>
static const float zero = 0.0;
float
__remquof (float x, float y, int *quo)
{
int32_t hx,hy;
u_int32_t sx;
int cquo, qs;
GET_FLOAT_WORD (hx, x);
GET_FLOAT_WORD (hy, y);
sx = hx & 0x80000000;
qs = sx ^ (hy & 0x80000000);
hy &= 0x7fffffff;
hx &= 0x7fffffff;
/* Purge off exception values. */
if (hy == 0)
return (x * y) / (x * y); /* y = 0 */
if ((hx >= 0x7f800000) /* x not finite */
|| (hy > 0x7f800000)) /* y is NaN */
return (x * y) / (x * y);
if (hy <= 0x7dffffff)
x = __ieee754_fmodf (x, 8 * y); /* now x < 8y */
if ((hx - hy) == 0)
{
*quo = qs ? -1 : 1;
return zero * x;
}
x = fabsf (x);
y = fabsf (y);
cquo = 0;
if (hy <= 0x7e7fffff && x >= 4 * y)
{
x -= 4 * y;
cquo += 4;
}
if (hy <= 0x7effffff && x >= 2 * y)
{
x -= 2 * y;
cquo += 2;
}
if (hy < 0x01000000)
{
if (x + x > y)
{
x -= y;
++cquo;
if (x + x >= y)
{
x -= y;
++cquo;
}
}
}
else
{
float y_half = 0.5 * y;
if (x > y_half)
{
x -= y;
++cquo;
if (x >= y_half)
{
x -= y;
++cquo;
}
}
}
*quo = qs ? -cquo : cquo;
if (sx)
x = -x;
return x;
}
weak_alias (__remquof, remquof)