glibc/math/math-narrow.h
Joseph Myers b3f27d8150 Add narrowing fma functions
This patch adds the narrowing fused multiply-add functions from TS
18661-1 / TS 18661-3 / C2X to glibc's libm: ffma, ffmal, dfmal,
f32fmaf64, f32fmaf32x, f32xfmaf64 for all configurations; f32fmaf64x,
f32fmaf128, f64fmaf64x, f64fmaf128, f32xfmaf64x, f32xfmaf128,
f64xfmaf128 for configurations with _Float64x and _Float128;
__f32fmaieee128 and __f64fmaieee128 aliases in the powerpc64le case
(for calls to ffmal and dfmal when long double is IEEE binary128).
Corresponding tgmath.h macro support is also added.

The changes are mostly similar to those for the other narrowing
functions previously added, especially that for sqrt, so the
description of those generally applies to this patch as well.  As with
sqrt, I reused the same test inputs in auto-libm-test-in as for
non-narrowing fma rather than adding extra or separate inputs for
narrowing fma.  The tests in libm-test-narrow-fma.inc also follow
those for non-narrowing fma.

The non-narrowing fma has a known bug (bug 6801) that it does not set
errno on errors (overflow, underflow, Inf * 0, Inf - Inf).  Rather
than fixing this or having narrowing fma check for errors when
non-narrowing does not (complicating the cases when narrowing fma can
otherwise be an alias for a non-narrowing function), this patch does
not attempt to check for errors from narrowing fma and set errno; the
CHECK_NARROW_FMA macro is still present, but as a placeholder that
does nothing, and this missing errno setting is considered to be
covered by the existing bug rather than needing a separate open bug.
missing-errno annotations are duly added to many of the
auto-libm-test-in test inputs for fma.

This completes adding all the new functions from TS 18661-1 to glibc,
so will be followed by corresponding stdc-predef.h changes to define
__STDC_IEC_60559_BFP__ and __STDC_IEC_60559_COMPLEX__, as the support
for TS 18661-1 will be at a similar level to that for C standard
floating-point facilities up to C11 (pragmas not implemented, but
library functions done).  (There are still further changes to be done
to implement changes to the types of fromfp functions from N2548.)

Tested as followed: natively with the full glibc testsuite for x86_64
(GCC 11, 7, 6) and x86 (GCC 11); with build-many-glibcs.py with GCC
11, 7 and 6; cross testing of math/ tests for powerpc64le, powerpc32
hard float, mips64 (all three ABIs, both hard and soft float).  The
different GCC versions are to cover the different cases in tgmath.h
and tgmath.h tests properly (GCC 6 has _Float* only as typedefs in
glibc headers, GCC 7 has proper _Float* support, GCC 8 adds
__builtin_tgmath).
2021-09-22 21:25:31 +00:00

399 lines
13 KiB
C

/* Helper macros for functions returning a narrower type.
Copyright (C) 2018-2021 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/>. */
#ifndef _MATH_NARROW_H
#define _MATH_NARROW_H 1
#include <bits/floatn.h>
#include <bits/long-double.h>
#include <errno.h>
#include <fenv.h>
#include <ieee754.h>
#include <math-barriers.h>
#include <math_private.h>
#include <fenv_private.h>
#include <math-narrow-alias.h>
#include <stdbool.h>
/* Carry out a computation using round-to-odd. The computation is
EXPR; the union type in which to store the result is UNION and the
subfield of the "ieee" field of that union with the low part of the
mantissa is MANTISSA; SUFFIX is the suffix for both underlying libm
functions for the argument type (for computations where a libm
function rather than a C operator is used when argument and result
types are the same) and the libc_fe* macros to ensure that the
correct rounding mode is used, for platforms with multiple rounding
modes where those macros set only the relevant mode.
CLEAR_UNDERFLOW indicates whether underflow exceptions must be
cleared (in the case where a round-toward-zero underflow might not
indicate an underflow after narrowing, when that narrowing only
reduces precision not exponent range and the architecture uses
before-rounding tininess detection). This macro does not work
correctly if the sign of an exact zero result depends on the
rounding mode, so that case must be checked for separately. */
#define ROUND_TO_ODD(EXPR, UNION, SUFFIX, MANTISSA, CLEAR_UNDERFLOW) \
({ \
fenv_t env; \
UNION u; \
\
libc_feholdexcept_setround ## SUFFIX (&env, FE_TOWARDZERO); \
u.d = (EXPR); \
math_force_eval (u.d); \
if (CLEAR_UNDERFLOW) \
feclearexcept (FE_UNDERFLOW); \
u.ieee.MANTISSA \
|= libc_feupdateenv_test ## SUFFIX (&env, FE_INEXACT) != 0; \
\
u.d; \
})
/* Check for error conditions from a narrowing add function returning
RET with arguments X and Y and set errno as needed. Overflow and
underflow can occur for finite arguments and a domain error for
infinite ones. */
#define CHECK_NARROW_ADD(RET, X, Y) \
do \
{ \
if (!isfinite (RET)) \
{ \
if (isnan (RET)) \
{ \
if (!isnan (X) && !isnan (Y)) \
__set_errno (EDOM); \
} \
else if (isfinite (X) && isfinite (Y)) \
__set_errno (ERANGE); \
} \
else if ((RET) == 0 && (X) != -(Y)) \
__set_errno (ERANGE); \
} \
while (0)
/* Implement narrowing add using round-to-odd. The arguments are X
and Y, the return type is TYPE and UNION, MANTISSA and SUFFIX are
as for ROUND_TO_ODD. */
#define NARROW_ADD_ROUND_TO_ODD(X, Y, TYPE, UNION, SUFFIX, MANTISSA) \
do \
{ \
TYPE ret; \
\
/* Ensure a zero result is computed in the original rounding \
mode. */ \
if ((X) == -(Y)) \
ret = (TYPE) ((X) + (Y)); \
else \
ret = (TYPE) ROUND_TO_ODD (math_opt_barrier (X) + (Y), \
UNION, SUFFIX, MANTISSA, false); \
\
CHECK_NARROW_ADD (ret, (X), (Y)); \
return ret; \
} \
while (0)
/* Implement a narrowing add function that is not actually narrowing
or where no attempt is made to be correctly rounding (the latter
only applies to IBM long double). The arguments are X and Y and
the return type is TYPE. */
#define NARROW_ADD_TRIVIAL(X, Y, TYPE) \
do \
{ \
TYPE ret; \
\
ret = (TYPE) ((X) + (Y)); \
CHECK_NARROW_ADD (ret, (X), (Y)); \
return ret; \
} \
while (0)
/* Check for error conditions from a narrowing subtract function
returning RET with arguments X and Y and set errno as needed.
Overflow and underflow can occur for finite arguments and a domain
error for infinite ones. */
#define CHECK_NARROW_SUB(RET, X, Y) \
do \
{ \
if (!isfinite (RET)) \
{ \
if (isnan (RET)) \
{ \
if (!isnan (X) && !isnan (Y)) \
__set_errno (EDOM); \
} \
else if (isfinite (X) && isfinite (Y)) \
__set_errno (ERANGE); \
} \
else if ((RET) == 0 && (X) != (Y)) \
__set_errno (ERANGE); \
} \
while (0)
/* Implement narrowing subtract using round-to-odd. The arguments are
X and Y, the return type is TYPE and UNION, MANTISSA and SUFFIX are
as for ROUND_TO_ODD. */
#define NARROW_SUB_ROUND_TO_ODD(X, Y, TYPE, UNION, SUFFIX, MANTISSA) \
do \
{ \
TYPE ret; \
\
/* Ensure a zero result is computed in the original rounding \
mode. */ \
if ((X) == (Y)) \
ret = (TYPE) ((X) - (Y)); \
else \
ret = (TYPE) ROUND_TO_ODD (math_opt_barrier (X) - (Y), \
UNION, SUFFIX, MANTISSA, false); \
\
CHECK_NARROW_SUB (ret, (X), (Y)); \
return ret; \
} \
while (0)
/* Implement a narrowing subtract function that is not actually
narrowing or where no attempt is made to be correctly rounding (the
latter only applies to IBM long double). The arguments are X and Y
and the return type is TYPE. */
#define NARROW_SUB_TRIVIAL(X, Y, TYPE) \
do \
{ \
TYPE ret; \
\
ret = (TYPE) ((X) - (Y)); \
CHECK_NARROW_SUB (ret, (X), (Y)); \
return ret; \
} \
while (0)
/* Check for error conditions from a narrowing multiply function
returning RET with arguments X and Y and set errno as needed.
Overflow and underflow can occur for finite arguments and a domain
error for Inf * 0. */
#define CHECK_NARROW_MUL(RET, X, Y) \
do \
{ \
if (!isfinite (RET)) \
{ \
if (isnan (RET)) \
{ \
if (!isnan (X) && !isnan (Y)) \
__set_errno (EDOM); \
} \
else if (isfinite (X) && isfinite (Y)) \
__set_errno (ERANGE); \
} \
else if ((RET) == 0 && (X) != 0 && (Y) != 0) \
__set_errno (ERANGE); \
} \
while (0)
/* Implement narrowing multiply using round-to-odd. The arguments are
X and Y, the return type is TYPE and UNION, MANTISSA, SUFFIX and
CLEAR_UNDERFLOW are as for ROUND_TO_ODD. */
#define NARROW_MUL_ROUND_TO_ODD(X, Y, TYPE, UNION, SUFFIX, MANTISSA, \
CLEAR_UNDERFLOW) \
do \
{ \
TYPE ret; \
\
ret = (TYPE) ROUND_TO_ODD (math_opt_barrier (X) * (Y), \
UNION, SUFFIX, MANTISSA, \
CLEAR_UNDERFLOW); \
\
CHECK_NARROW_MUL (ret, (X), (Y)); \
return ret; \
} \
while (0)
/* Implement a narrowing multiply function that is not actually
narrowing or where no attempt is made to be correctly rounding (the
latter only applies to IBM long double). The arguments are X and Y
and the return type is TYPE. */
#define NARROW_MUL_TRIVIAL(X, Y, TYPE) \
do \
{ \
TYPE ret; \
\
ret = (TYPE) ((X) * (Y)); \
CHECK_NARROW_MUL (ret, (X), (Y)); \
return ret; \
} \
while (0)
/* Check for error conditions from a narrowing divide function
returning RET with arguments X and Y and set errno as needed.
Overflow, underflow and divide-by-zero can occur for finite
arguments and a domain error for Inf / Inf and 0 / 0. */
#define CHECK_NARROW_DIV(RET, X, Y) \
do \
{ \
if (!isfinite (RET)) \
{ \
if (isnan (RET)) \
{ \
if (!isnan (X) && !isnan (Y)) \
__set_errno (EDOM); \
} \
else if (isfinite (X)) \
__set_errno (ERANGE); \
} \
else if ((RET) == 0 && (X) != 0 && !isinf (Y)) \
__set_errno (ERANGE); \
} \
while (0)
/* Implement narrowing divide using round-to-odd. The arguments are X
and Y, the return type is TYPE and UNION, MANTISSA, SUFFIX and
CLEAR_UNDERFLOW are as for ROUND_TO_ODD. */
#define NARROW_DIV_ROUND_TO_ODD(X, Y, TYPE, UNION, SUFFIX, MANTISSA, \
CLEAR_UNDERFLOW) \
do \
{ \
TYPE ret; \
\
ret = (TYPE) ROUND_TO_ODD (math_opt_barrier (X) / (Y), \
UNION, SUFFIX, MANTISSA, \
CLEAR_UNDERFLOW); \
\
CHECK_NARROW_DIV (ret, (X), (Y)); \
return ret; \
} \
while (0)
/* Implement a narrowing divide function that is not actually
narrowing or where no attempt is made to be correctly rounding (the
latter only applies to IBM long double). The arguments are X and Y
and the return type is TYPE. */
#define NARROW_DIV_TRIVIAL(X, Y, TYPE) \
do \
{ \
TYPE ret; \
\
ret = (TYPE) ((X) / (Y)); \
CHECK_NARROW_DIV (ret, (X), (Y)); \
return ret; \
} \
while (0)
/* Check for error conditions from a narrowing square root function
returning RET with argument X and set errno as needed. Overflow
and underflow can occur for finite positive arguments and a domain
error for negative arguments. */
#define CHECK_NARROW_SQRT(RET, X) \
do \
{ \
if (!isfinite (RET)) \
{ \
if (isnan (RET)) \
{ \
if (!isnan (X)) \
__set_errno (EDOM); \
} \
else if (isfinite (X)) \
__set_errno (ERANGE); \
} \
else if ((RET) == 0 && (X) != 0) \
__set_errno (ERANGE); \
} \
while (0)
/* Implement narrowing square root using round-to-odd. The argument
is X, the return type is TYPE and UNION, MANTISSA and SUFFIX are as
for ROUND_TO_ODD. */
#define NARROW_SQRT_ROUND_TO_ODD(X, TYPE, UNION, SUFFIX, MANTISSA) \
do \
{ \
TYPE ret; \
\
ret = (TYPE) ROUND_TO_ODD (sqrt ## SUFFIX (math_opt_barrier (X)), \
UNION, SUFFIX, MANTISSA, false); \
\
CHECK_NARROW_SQRT (ret, (X)); \
return ret; \
} \
while (0)
/* Implement a narrowing square root function where no attempt is made
to be correctly rounding (this only applies to IBM long double; the
case where the function is not actually narrowing is handled by
aliasing other sqrt functions in libm, not using this macro). The
argument is X and the return type is TYPE. */
#define NARROW_SQRT_TRIVIAL(X, TYPE, SUFFIX) \
do \
{ \
TYPE ret; \
\
ret = (TYPE) (sqrt ## SUFFIX (X)); \
CHECK_NARROW_SQRT (ret, (X)); \
return ret; \
} \
while (0)
/* Check for error conditions from a narrowing fused multiply-add
function returning RET with arguments X, Y and Z and set errno as
needed. Checking for error conditions for fma (either narrowing or
not) and setting errno is not currently implemented. See bug
6801. */
#define CHECK_NARROW_FMA(RET, X, Y, Z) \
do \
{ \
} \
while (0)
/* Implement narrowing fused multiply-add using round-to-odd. The
arguments are X, Y and Z, the return type is TYPE and UNION,
MANTISSA, SUFFIX and CLEAR_UNDERFLOW are as for ROUND_TO_ODD. */
#define NARROW_FMA_ROUND_TO_ODD(X, Y, Z, TYPE, UNION, SUFFIX, MANTISSA, \
CLEAR_UNDERFLOW) \
do \
{ \
typeof (X) tmp; \
TYPE ret; \
\
tmp = ROUND_TO_ODD (fma ## SUFFIX (math_opt_barrier (X), (Y), \
(Z)), \
UNION, SUFFIX, MANTISSA, CLEAR_UNDERFLOW); \
/* If the round-to-odd result is zero, the result is an exact \
zero and must be recomputed in the original rounding mode. */ \
if (tmp == 0) \
ret = (TYPE) (math_opt_barrier (X) * (Y) + (Z)); \
else \
ret = (TYPE) tmp; \
\
CHECK_NARROW_FMA (ret, (X), (Y), (Z)); \
return ret; \
} \
while (0)
/* Implement a narrowing fused multiply-add function where no attempt
is made to be correctly rounding (this only applies to IBM long
double; the case where the function is not actually narrowing is
handled by aliasing other fma functions in libm, not using this
macro). The arguments are X, Y and Z and the return type is
TYPE. */
#define NARROW_FMA_TRIVIAL(X, Y, Z, TYPE, SUFFIX) \
do \
{ \
TYPE ret; \
\
ret = (TYPE) (fma ## SUFFIX ((X), (Y), (Z))); \
CHECK_NARROW_FMA (ret, (X), (Y), (Z)); \
return ret; \
} \
while (0)
#endif /* math-narrow.h. */