[PATCH v2] math: Remove ldbl-128/s_fma.c

[Wilco Dijkstra]

v2: Remove more redundant files

Remove ldbl-128/s_fma.c - it makes no sense to use emulated float128
operations to emulate FMA.  Benchmarking shows dbl-64/s_fma.c is about
twice as fast.  Remove redundant dbl-64/s_fma.c includes in targets
that were trying to work around this issue.

Passes buildmanyglibc on alpha/sparc/mips. OK for commit?

---

diff --git a/sysdeps/alpha/fpu/s_fma.c b/sysdeps/alpha/fpu/s_fma.c
deleted file mode 100644
index 8f62605870439f2a821b9ee95bc868a8794aa363..0000000000000000000000000000000000000000
--- a/sysdeps/alpha/fpu/s_fma.c
+++ /dev/null
@@ -1,2 +0,0 @@
-/* Always use dbl-64 version because long double is emulated in software.  */
-#include <sysdeps/ieee754/dbl-64/s_fma.c>
diff --git a/sysdeps/ieee754/ldbl-128/s_fma.c b/sysdeps/ieee754/ldbl-128/s_fma.c
deleted file mode 100644
index 01ca556260884553e2406581e94214730ce1f6eb..0000000000000000000000000000000000000000
--- a/sysdeps/ieee754/ldbl-128/s_fma.c
+++ /dev/null
@@ -1,67 +0,0 @@
-/* Compute x * y + z as ternary operation.
-   Copyright (C) 2010-2025 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/>.  */
-
-#define NO_MATH_REDIRECT
-#define dfmal __hide_dfmal
-#define f32xfmaf64 __hide_f32xfmaf64
-#include <math.h>
-#undef dfmal
-#undef f32xfmaf64
-#include <fenv.h>
-#include <ieee754.h>
-#include <libm-alias-double.h>
-#include <math-narrow-alias.h>
-#include <math-use-builtins.h>
-
-/* This implementation relies on long double being more than twice as
-   precise as double and uses rounding to odd in order to avoid problems
-   with double rounding.
-   See a paper by Boldo and Melquiond:
-   http://www.lri.fr/~melquion/doc/08-tc.pdf  */
-
-double
-__fma (double x, double y, double z)
-{
-#if USE_FMA_BUILTIN
-  return __builtin_fma (x, y, z);
-#else
-  fenv_t env;
-  /* Multiplication is always exact.  */
-  long double temp = (long double) x * (long double) y;
-
-  /* Ensure correct sign of an exact zero result by performing the
-     addition in the original rounding mode in that case.  */
-  if (temp == -z)
-    return (double) temp + z;
-
-  union ieee854_long_double u;
-  __feholdexcept (&env);
-  __fesetround (FE_TOWARDZERO);
-  /* Perform addition with round to odd.  */
-  u.d = temp + (long double) z;
-  if ((u.ieee.mantissa3 & 1) == 0 && u.ieee.exponent != 0x7fff)
-    u.ieee.mantissa3 |= __fetestexcept (FE_INEXACT) != 0;
-  __feupdateenv (&env);
-  /* And finally truncation with round to nearest.  */
-  return (double) u.d;
-#endif /* ! USE_FMA_BUILTIN  */
-}
-#ifndef __fma
-libm_alias_double (__fma, fma)
-libm_alias_double_narrow (__fma, fma)
-#endif
diff --git a/sysdeps/mips/mips64/n32/fpu/s_fma.c b/sysdeps/mips/mips64/n32/fpu/s_fma.c
deleted file mode 100644
index b61fa643de9ccc00d9d4da48cd90a446a1ef78d9..0000000000000000000000000000000000000000
--- a/sysdeps/mips/mips64/n32/fpu/s_fma.c
+++ /dev/null
@@ -1 +0,0 @@
-#include <sysdeps/ieee754/dbl-64/s_fma.c>
diff --git a/sysdeps/mips/mips64/n64/fpu/s_fma.c b/sysdeps/mips/mips64/n64/fpu/s_fma.c
deleted file mode 100644
index b61fa643de9ccc00d9d4da48cd90a446a1ef78d9..0000000000000000000000000000000000000000
--- a/sysdeps/mips/mips64/n64/fpu/s_fma.c
+++ /dev/null
@@ -1 +0,0 @@
-#include <sysdeps/ieee754/dbl-64/s_fma.c>
diff --git a/sysdeps/sparc/sparc32/fpu/s_fma.c b/sysdeps/sparc/sparc32/fpu/s_fma.c
deleted file mode 100644
index 8f62605870439f2a821b9ee95bc868a8794aa363..0000000000000000000000000000000000000000
--- a/sysdeps/sparc/sparc32/fpu/s_fma.c
+++ /dev/null
@@ -1,2 +0,0 @@
-/* Always use dbl-64 version because long double is emulated in software.  */
-#include <sysdeps/ieee754/dbl-64/s_fma.c>
diff --git a/sysdeps/sparc/sparc64/fpu/s_fma.c b/sysdeps/sparc/sparc64/fpu/s_fma.c
deleted file mode 100644
index 8f62605870439f2a821b9ee95bc868a8794aa363..0000000000000000000000000000000000000000
--- a/sysdeps/sparc/sparc64/fpu/s_fma.c
+++ /dev/null
@@ -1,2 +0,0 @@
-/* Always use dbl-64 version because long double is emulated in software.  */
-#include <sysdeps/ieee754/dbl-64/s_fma.c>