From e38f2308273c8a51ec45f013d22c963590917cca Mon Sep 17 00:00:00 2001 From: Mark Murray Date: Fri, 1 Mar 2024 15:53:58 +0000 Subject: [PATCH] lib/msun: Fix tgammal(3) on IEEE 128-bit platforms Undo the 80-bit "stub" implementation of the 128-bit long double tgammal(3) function. The latest (as of Feb 2024) version of the src/contrib/arm-optimised-routines library includes a standalone, full 128-bit replacement. This needs a small bit of wrapping to fit it in, but is otherwise a drop-in replacement. Testing this is hard, as most maths packages blow up as soon as their 80-bit floating-point capability is exceeded. With 128-bit tgammal(), this is easy to do, and this is the range that needs to be checked the most carefully. Using my copy of Maple, I was able to check that the output was within a few ULP of the correct answer, right up to the point of 128-bit over- and underflow. Additionally, the results are no worse, and indeed better than the 80-bit version. Steve Kargl sent me his libm testing code, which I used to verify that the excpetions for certain key values were correct. Tested in this case were +-Inf, +-NaN, +-1 and +-0. Differential Revision: https://reviews.freebsd.org/D44168 Reviewed by: theraven, andrew, imp --- lib/msun/Makefile | 2 +- lib/msun/ld128/b_tgammal.c | 53 ++++---------------------------------- lib/msun/man/lgamma.3 | 19 +++++++------- 3 files changed, 15 insertions(+), 59 deletions(-) diff --git a/lib/msun/Makefile b/lib/msun/Makefile index 6c8c6c3e8009..24989749a502 100644 --- a/lib/msun/Makefile +++ b/lib/msun/Makefile @@ -26,7 +26,7 @@ CFLAGS+= -I${.CURDIR}/x86 CFLAGS+= -I${.CURDIR}/ld80 .elif ${LDBL_PREC} == 113 .PATH: ${.CURDIR}/ld128 -CFLAGS+= -I${.CURDIR}/ld128 +CFLAGS+= -I${.CURDIR}/ld128 -I${SRCTOP}/contrib/arm-optimized-routines/math .endif CFLAGS+= -I${.CURDIR}/${ARCH_SUBDIR} diff --git a/lib/msun/ld128/b_tgammal.c b/lib/msun/ld128/b_tgammal.c index 4bae4f3aded6..6df7264a4c9e 100644 --- a/lib/msun/ld128/b_tgammal.c +++ b/lib/msun/ld128/b_tgammal.c @@ -1,55 +1,12 @@ /*- * SPDX-License-Identifier: BSD-2-Clause * - * Copyright (c) 2013 David Chisnall - * All rights reserved. - * - * Redistribution and use in source and binary forms, with or without - * modification, are permitted provided that the following conditions - * are met: - * 1. Redistributions of source code must retain the above copyright - * notice, this list of conditions and the following disclaimer. - * 2. Redistributions in binary form must reproduce the above copyright - * notice, this list of conditions and the following disclaimer in the - * documentation and/or other materials provided with the distribution. - * - * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND - * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE - * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE - * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE - * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL - * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS - * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) - * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT - * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY - * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF - * SUCH DAMAGE. + * Copyright (c) 2024 The FreeBSD Foundation */ -#include -#include - /* - * If long double is not the same size as double, then these will lose - * precision and we should emit a warning whenever something links against - * them. + * This is a pure C function generously donated by ARM. + * See src/contrib/arm-optimized-routines/math/tgamma128.[ch]. */ -#if (LDBL_MANT_DIG > 53) -#define WARN_IMPRECISE(x) \ - __warn_references(x, # x " has lower than advertised precision"); -#else -#define WARN_IMPRECISE(x) -#endif -/* - * Declare the functions as weak variants so that other libraries providing - * real versions can override them. - */ -#define DECLARE_WEAK(x)\ - __weak_reference(imprecise_## x, x);\ - WARN_IMPRECISE(x) - -#define DECLARE_IMPRECISE(f) \ - long double imprecise_ ## f ## l(long double v) { return f(v); }\ - DECLARE_WEAK(f ## l) - -DECLARE_IMPRECISE(tgamma); +#define tgamma128 tgammal +#include "tgamma128.c" diff --git a/lib/msun/man/lgamma.3 b/lib/msun/man/lgamma.3 index 8c0298ec8299..a77d524373fb 100644 --- a/lib/msun/man/lgamma.3 +++ b/lib/msun/man/lgamma.3 @@ -25,7 +25,7 @@ .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF .\" SUCH DAMAGE. .\" -.Dd December 8, 2017 +.Dd March 1, 2024 .Dt LGAMMA 3 .Os .Sh NAME @@ -167,15 +167,6 @@ non-positive integers. For large non-integer negative values, .Fn tgamma will underflow. -.Sh BUGS -To conform with newer C/C++ standards, a stub implementation for -.Nm tgammal -was committed to the math library, where -.Nm tgammal -is mapped to -.Nm tgamma . -Thus, the numerical accuracy is at most that of the 53-bit double -precision implementation. .Sh SEE ALSO .Xr math 3 .Sh STANDARDS @@ -212,3 +203,11 @@ The .Fn tgamma function appeared in .Fx 5.0 . +The 128-bit +.Ft long double +version of +.Fn tgammal +replaced the 80-bit stub version in +version in +.Fx 16 , +thanks to an appropriate implementation from Arm.