mirror of
https://sourceware.org/git/glibc.git
synced 2024-11-10 15:20:10 +00:00
681900d296
The behavior of isnan/__builtin_isnan on bit patterns that do not correspond to something that the CPU would produce from valid inputs is currently under-defined in the toolchain. (The GCC built-in and glibc disagree.) The isnan check in PRINTF_FP_FETCH in stdio-common/printf_fp.c assumes the GCC behavior that returns true for non-normal numbers which are not specified as NaN. (The glibc implementation returns false for such numbers.) At present, passing non-normal numbers to __mpn_extract_long_double causes this function to produce irregularly shaped multi-precision integers, triggering undefined behavior in __printf_fp_l. With GCC 10 and glibc 2.32, this behavior is not visible because __builtin_isnan is used, which avoids calling __mpn_extract_long_double in this case. This commit updates the implementation of __mpn_extract_long_double so that regularly shaped multi-precision integers are produced in this case, avoiding undefined behavior in __printf_fp_l.
129 lines
4.1 KiB
C
129 lines
4.1 KiB
C
/* Copyright (C) 1995-2020 Free Software Foundation, Inc.
|
|
This file is part of the GNU C Library.
|
|
|
|
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
|
|
<https://www.gnu.org/licenses/>. */
|
|
|
|
#include "gmp.h"
|
|
#include "gmp-impl.h"
|
|
#include "longlong.h"
|
|
#include <ieee754.h>
|
|
#include <float.h>
|
|
#include <stdlib.h>
|
|
|
|
/* Convert a `long double' in IEEE854 standard double-precision format to a
|
|
multi-precision integer representing the significand scaled up by its
|
|
number of bits (64 for long double) and an integral power of two
|
|
(MPN frexpl). */
|
|
|
|
mp_size_t
|
|
__mpn_extract_long_double (mp_ptr res_ptr, mp_size_t size,
|
|
int *expt, int *is_neg,
|
|
long double value)
|
|
{
|
|
union ieee854_long_double u;
|
|
u.d = value;
|
|
|
|
*is_neg = u.ieee.negative;
|
|
*expt = (int) u.ieee.exponent - IEEE854_LONG_DOUBLE_BIAS;
|
|
|
|
#if BITS_PER_MP_LIMB == 32
|
|
res_ptr[0] = u.ieee.mantissa1; /* Low-order 32 bits of fraction. */
|
|
res_ptr[1] = u.ieee.mantissa0; /* High-order 32 bits. */
|
|
#define N 2
|
|
#elif BITS_PER_MP_LIMB == 64
|
|
/* Hopefully the compiler will combine the two bitfield extracts
|
|
and this composition into just the original quadword extract. */
|
|
res_ptr[0] = ((mp_limb_t) u.ieee.mantissa0 << 32) | u.ieee.mantissa1;
|
|
#define N 1
|
|
#else
|
|
#error "mp_limb size " BITS_PER_MP_LIMB "not accounted for"
|
|
#endif
|
|
|
|
if (u.ieee.exponent == 0)
|
|
{
|
|
/* A biased exponent of zero is a special case.
|
|
Either it is a zero or it is a denormal number. */
|
|
if (res_ptr[0] == 0 && res_ptr[N - 1] == 0) /* Assumes N<=2. */
|
|
/* It's zero. */
|
|
*expt = 0;
|
|
else
|
|
{
|
|
/* It is a denormal number, meaning it has no implicit leading
|
|
one bit, and its exponent is in fact the format minimum. */
|
|
int cnt;
|
|
|
|
/* One problem with Intel's 80-bit format is that the explicit
|
|
leading one in the normalized representation has to be zero
|
|
for denormalized number. If it is one, the number is according
|
|
to Intel's specification an invalid number. We make the
|
|
representation unique by explicitly clearing this bit. */
|
|
res_ptr[N - 1] &= ~((mp_limb_t) 1 << ((LDBL_MANT_DIG - 1) % BITS_PER_MP_LIMB));
|
|
|
|
if (res_ptr[N - 1] != 0)
|
|
{
|
|
count_leading_zeros (cnt, res_ptr[N - 1]);
|
|
if (cnt != 0)
|
|
{
|
|
#if N == 2
|
|
res_ptr[N - 1] = res_ptr[N - 1] << cnt
|
|
| (res_ptr[0] >> (BITS_PER_MP_LIMB - cnt));
|
|
res_ptr[0] <<= cnt;
|
|
#else
|
|
res_ptr[N - 1] <<= cnt;
|
|
#endif
|
|
}
|
|
*expt = LDBL_MIN_EXP - 1 - cnt;
|
|
}
|
|
else if (res_ptr[0] != 0)
|
|
{
|
|
count_leading_zeros (cnt, res_ptr[0]);
|
|
res_ptr[N - 1] = res_ptr[0] << cnt;
|
|
res_ptr[0] = 0;
|
|
*expt = LDBL_MIN_EXP - 1 - BITS_PER_MP_LIMB - cnt;
|
|
}
|
|
else
|
|
{
|
|
/* This is the special case of the pseudo denormal number
|
|
with only the implicit leading bit set. The value is
|
|
in fact a normal number and so we have to treat this
|
|
case differently. */
|
|
#if N == 2
|
|
res_ptr[N - 1] = 0x80000000ul;
|
|
#else
|
|
res_ptr[0] = 0x8000000000000000ul;
|
|
#endif
|
|
*expt = LDBL_MIN_EXP - 1;
|
|
}
|
|
}
|
|
}
|
|
else if (u.ieee.exponent < 0x7fff
|
|
#if N == 2
|
|
&& res_ptr[0] == 0
|
|
#endif
|
|
&& res_ptr[N - 1] == 0)
|
|
/* Pseudo zero. */
|
|
*expt = 0;
|
|
else
|
|
/* Unlike other floating point formats, the most significant bit
|
|
is explicit and expected to be set for normal numbers. Set it
|
|
in case it is cleared in the input. Otherwise, callers will
|
|
not be able to produce the expected multi-precision integer
|
|
layout by shifting. */
|
|
res_ptr[N - 1] |= (mp_limb_t) 1 << (LDBL_MANT_DIG - 1
|
|
- ((N - 1) * BITS_PER_MP_LIMB));
|
|
|
|
return N;
|
|
}
|