Add narrowing subtract functions.

This patch adds the narrowing subtract functions from TS 18661-1 to
glibc's libm: fsub, fsubl, dsubl, f32subf64, f32subf32x, f32xsubf64
for all configurations; f32subf64x, f32subf128, f64subf64x,
f64subf128, f32xsubf64x, f32xsubf128, f64xsubf128 for configurations
with _Float64x and _Float128; __nldbl_dsubl for ldbl-opt.

The changes are essentially the same as for the narrowing add
functions, so the description of those generally applies to this patch
as well.

Tested for x86_64, x86, mips64 (all three ABIs, both hard and soft
float) and powerpc, and with build-many-glibcs.py.

	* math/Makefile (libm-narrow-fns): Add sub.
	(libm-test-funcs-narrow): Likewise.
	* math/Versions (GLIBC_2.28): Add narrowing subtract functions.
	* math/bits/mathcalls-narrow.h (sub): Use __MATHCALL_NARROW.
	* math/gen-auto-libm-tests.c (test_functions): Add sub.
	* math/math-narrow.h (CHECK_NARROW_SUB): New macro.
	(NARROW_SUB_ROUND_TO_ODD): Likewise.
	(NARROW_SUB_TRIVIAL): Likewise.
	* sysdeps/ieee754/float128/float128_private.h (__fsubl): New
	macro.
	(__dsubl): Likewise.
	* sysdeps/ieee754/ldbl-opt/Makefile (libnldbl-calls): Add fsub and
	dsub.
	(CFLAGS-nldbl-dsub.c): New variable.
	(CFLAGS-nldbl-fsub.c): Likewise.
	* sysdeps/ieee754/ldbl-opt/Versions (GLIBC_2.28): Add
	__nldbl_dsubl.
	* sysdeps/ieee754/ldbl-opt/nldbl-compat.h (__nldbl_dsubl): New
	prototype.
	* manual/arith.texi (Misc FP Arithmetic): Document fsub, fsubl,
	dsubl, fMsubfN, fMsubfNx, fMxsubfN and fMxsubfNx.
	* math/auto-libm-test-in: Add tests of sub.
	* math/auto-libm-test-out-narrow-sub: New generated file.
	* math/libm-test-narrow-sub.inc: New file.
	* sysdeps/i386/fpu/s_f32xsubf64.c: Likewise.
	* sysdeps/ieee754/dbl-64/s_f32xsubf64.c: Likewise.
	* sysdeps/ieee754/dbl-64/s_fsub.c: Likewise.
	* sysdeps/ieee754/float128/s_f32subf128.c: Likewise.
	* sysdeps/ieee754/float128/s_f64subf128.c: Likewise.
	* sysdeps/ieee754/float128/s_f64xsubf128.c: Likewise.
	* sysdeps/ieee754/ldbl-128/s_dsubl.c: Likewise.
	* sysdeps/ieee754/ldbl-128/s_f64xsubf128.c: Likewise.
	* sysdeps/ieee754/ldbl-128/s_fsubl.c: Likewise.
	* sysdeps/ieee754/ldbl-128ibm/s_dsubl.c: Likewise.
	* sysdeps/ieee754/ldbl-128ibm/s_fsubl.c: Likewise.
	* sysdeps/ieee754/ldbl-96/s_dsubl.c: Likewise.
	* sysdeps/ieee754/ldbl-96/s_fsubl.c: Likewise.
	* sysdeps/ieee754/ldbl-opt/nldbl-dsub.c: Likewise.
	* sysdeps/ieee754/ldbl-opt/nldbl-fsub.c: Likewise.
	* sysdeps/ieee754/soft-fp/s_dsubl.c: Likewise.
	* sysdeps/ieee754/soft-fp/s_fsub.c: Likewise.
	* sysdeps/ieee754/soft-fp/s_fsubl.c: Likewise.
	* sysdeps/powerpc/fpu/libm-test-ulps: Update.
	* sysdeps/mach/hurd/i386/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/aarch64/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/alpha/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/arm/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/hppa/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/i386/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/ia64/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/m68k/coldfire/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/m68k/m680x0/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/microblaze/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/mips/mips32/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/mips/mips64/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/nios2/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/powerpc/powerpc32/fpu/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/powerpc/powerpc64/libm-le.abilist: Likewise.
	* sysdeps/unix/sysv/linux/powerpc/powerpc64/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/riscv/rv64/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/s390/s390-32/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/s390/s390-64/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/sh/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/sparc/sparc32/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/sparc/sparc64/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/tile/tilegx32/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/tile/tilegx64/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/x86_64/64/libm.abilist: Likewise.
	* sysdeps/unix/sysv/linux/x86_64/x32/libm.abilist: Likewise.

21 files changed, 542 insertions, 2 deletions

diff --git a/sysdeps/ieee754/dbl-64/s_f32xsubf64.c b/sysdeps/ieee754/dbl-64/s_f32xsubf64.c
new file mode 100644
index 0000000000..568f43a880
--- /dev/null
+++ b/sysdeps/ieee754/dbl-64/s_f32xsubf64.c
@@ -0,0 +1,30 @@
+/* Subtract _Float64 values, converting the result to _Float32x.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#define dsubl __hide_dsubl
+#include <math.h>
+#undef dsubl
+
+#include <math-narrow.h>
+
+_Float32x
+__f32xsubf64 (_Float64 x, _Float64 y)
+{
+  NARROW_SUB_TRIVIAL (x, y, _Float32x);
+}
+libm_alias_float32x_float64 (sub)
diff --git a/sysdeps/ieee754/dbl-64/s_fsub.c b/sysdeps/ieee754/dbl-64/s_fsub.c
new file mode 100644
index 0000000000..c3b911e801
--- /dev/null
+++ b/sysdeps/ieee754/dbl-64/s_fsub.c
@@ -0,0 +1,34 @@
+/* Subtract double values, narrowing the result to float.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#define f32subf64 __hide_f32subf64
+#define f32subf32x __hide_f32subf32x
+#define fsubl __hide_fsubl
+#include <math.h>
+#undef f32subf64
+#undef f32subf32x
+#undef fsubl
+
+#include <math-narrow.h>
+
+float
+__fsub (double x, double y)
+{
+  NARROW_SUB_ROUND_TO_ODD (x, y, float, union ieee754_double, , mantissa1);
+}
+libm_alias_float_double (sub)
diff --git a/sysdeps/ieee754/float128/float128_private.h b/sysdeps/ieee754/float128/float128_private.h
index eebf19af93..7606026593 100644
--- a/sysdeps/ieee754/float128/float128_private.h
+++ b/sysdeps/ieee754/float128/float128_private.h
@@ -258,6 +258,8 @@
 
 #define __faddl __f32addf128
 #define __daddl __f64addf128
+#define __fsubl __f32subf128
+#define __dsubl __f64subf128
 
 /* __nexttowardf128 is not _Float128 API. */
 #define __nexttowardl __nexttowardf128_do_not_use
diff --git a/sysdeps/ieee754/float128/s_f32subf128.c b/sysdeps/ieee754/float128/s_f32subf128.c
new file mode 100644
index 0000000000..f1373f68ac
--- /dev/null
+++ b/sysdeps/ieee754/float128/s_f32subf128.c
@@ -0,0 +1,6 @@
+#define f32subf64x __hide_f32subf64x
+#define f32subf128 __hide_f32subf128
+#include <float128_private.h>
+#undef f32subf64x
+#undef f32subf128
+#include "../ldbl-128/s_fsubl.c"
diff --git a/sysdeps/ieee754/float128/s_f64subf128.c b/sysdeps/ieee754/float128/s_f64subf128.c
new file mode 100644
index 0000000000..89ef5a7cc6
--- /dev/null
+++ b/sysdeps/ieee754/float128/s_f64subf128.c
@@ -0,0 +1,10 @@
+#define f32xsubf64x __hide_f32xsubf64x
+#define f32xsubf128 __hide_f32xsubf128
+#define f64subf64x __hide_f64subf64x
+#define f64subf128 __hide_f64subf128
+#include <float128_private.h>
+#undef f32xsubf64x
+#undef f32xsubf128
+#undef f64subf64x
+#undef f64subf128
+#include "../ldbl-128/s_dsubl.c"
diff --git a/sysdeps/ieee754/float128/s_f64xsubf128.c b/sysdeps/ieee754/float128/s_f64xsubf128.c
new file mode 100644
index 0000000000..64a0d6d82c
--- /dev/null
+++ b/sysdeps/ieee754/float128/s_f64xsubf128.c
@@ -0,0 +1,2 @@
+#include <float128_private.h>
+#include "../ldbl-128/s_f64xsubf128.c"
diff --git a/sysdeps/ieee754/ldbl-128/s_dsubl.c b/sysdeps/ieee754/ldbl-128/s_dsubl.c
new file mode 100644
index 0000000000..46d0f95b4c
--- /dev/null
+++ b/sysdeps/ieee754/ldbl-128/s_dsubl.c
@@ -0,0 +1,37 @@
+/* Subtract long double (ldbl-128) values, narrowing the result to double.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#define f32xsubf64x __hide_f32xsubf64x
+#define f32xsubf128 __hide_f32xsubf128
+#define f64subf64x __hide_f64subf64x
+#define f64subf128 __hide_f64subf128
+#include <math.h>
+#undef f32xsubf64x
+#undef f32xsubf128
+#undef f64subf64x
+#undef f64subf128
+
+#include <math-narrow.h>
+
+double
+__dsubl (_Float128 x, _Float128 y)
+{
+  NARROW_SUB_ROUND_TO_ODD (x, y, double, union ieee854_long_double, l,
+			   mantissa3);
+}
+libm_alias_double_ldouble (sub)
diff --git a/sysdeps/ieee754/ldbl-128/s_f64xsubf128.c b/sysdeps/ieee754/ldbl-128/s_f64xsubf128.c
new file mode 100644
index 0000000000..c9a63d9e84
--- /dev/null
+++ b/sysdeps/ieee754/ldbl-128/s_f64xsubf128.c
@@ -0,0 +1,38 @@
+/* Subtract _Float128 values, converting the result to _Float64x.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#include <math.h>
+#include <math-narrow.h>
+
+/* math_ldbl.h defines _Float128 to long double for this directory,
+   but when they are different, this function must be defined with
+   _Float128 arguments to avoid defining an alias with an incompatible
+   type.  */
+#undef _Float128
+
+_Float64x
+__f64xsubf128 (_Float128 x, _Float128 y)
+{
+#if __HAVE_FLOAT64X_LONG_DOUBLE && __HAVE_DISTINCT_FLOAT128
+  NARROW_SUB_ROUND_TO_ODD (x, y, _Float64x, union ieee854_long_double, l,
+			   mantissa3);
+#else
+  NARROW_SUB_TRIVIAL (x, y, _Float64x);
+#endif
+}
+libm_alias_float64x_float128 (sub)
diff --git a/sysdeps/ieee754/ldbl-128/s_fsubl.c b/sysdeps/ieee754/ldbl-128/s_fsubl.c
new file mode 100644
index 0000000000..6a5e6b3173
--- /dev/null
+++ b/sysdeps/ieee754/ldbl-128/s_fsubl.c
@@ -0,0 +1,33 @@
+/* Subtract long double (ldbl-128) values, narrowing the result to float.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#define f32subf64x __hide_f32subf64x
+#define f32subf128 __hide_f32subf128
+#include <math.h>
+#undef f32subf64x
+#undef f32subf128
+
+#include <math-narrow.h>
+
+float
+__fsubl (_Float128 x, _Float128 y)
+{
+  NARROW_SUB_ROUND_TO_ODD (x, y, float, union ieee854_long_double, l,
+			   mantissa3);
+}
+libm_alias_float_ldouble (sub)
diff --git a/sysdeps/ieee754/ldbl-128ibm/s_dsubl.c b/sysdeps/ieee754/ldbl-128ibm/s_dsubl.c
new file mode 100644
index 0000000000..e0e80f26a4
--- /dev/null
+++ b/sysdeps/ieee754/ldbl-128ibm/s_dsubl.c
@@ -0,0 +1,27 @@
+/* Subtract long double (ldbl-128ibm) values, narrowing the result to double.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#include <math.h>
+#include <math-narrow.h>
+
+double
+__dsubl (long double x, long double y)
+{
+  NARROW_SUB_TRIVIAL (x, y, double);
+}
+libm_alias_double_ldouble (sub)
diff --git a/sysdeps/ieee754/ldbl-128ibm/s_fsubl.c b/sysdeps/ieee754/ldbl-128ibm/s_fsubl.c
new file mode 100644
index 0000000000..aa6662de05
--- /dev/null
+++ b/sysdeps/ieee754/ldbl-128ibm/s_fsubl.c
@@ -0,0 +1,27 @@
+/* Subtract long double (ldbl-128ibm) values, narrowing the result to float.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#include <math.h>
+#include <math-narrow.h>
+
+float
+__fsubl (long double x, long double y)
+{
+  NARROW_SUB_TRIVIAL (x, y, float);
+}
+libm_alias_float_ldouble (sub)
diff --git a/sysdeps/ieee754/ldbl-96/s_dsubl.c b/sysdeps/ieee754/ldbl-96/s_dsubl.c
new file mode 100644
index 0000000000..5a855790f6
--- /dev/null
+++ b/sysdeps/ieee754/ldbl-96/s_dsubl.c
@@ -0,0 +1,33 @@
+/* Subtract long double (ldbl-96) values, narrowing the result to double.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#define f32xsubf64x __hide_f32xsubf64x
+#define f64subf64x __hide_f64subf64x
+#include <math.h>
+#undef f32xsubf64x
+#undef f64subf64x
+
+#include <math-narrow.h>
+
+double
+__dsubl (long double x, long double y)
+{
+  NARROW_SUB_ROUND_TO_ODD (x, y, double, union ieee854_long_double, l,
+			   mantissa1);
+}
+libm_alias_double_ldouble (sub)
diff --git a/sysdeps/ieee754/ldbl-96/s_fsubl.c b/sysdeps/ieee754/ldbl-96/s_fsubl.c
new file mode 100644
index 0000000000..54aaf68b74
--- /dev/null
+++ b/sysdeps/ieee754/ldbl-96/s_fsubl.c
@@ -0,0 +1,31 @@
+/* Subtract long double (ldbl-96) values, narrowing the result to float.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#define f32subf64x __hide_f32subf64x
+#include <math.h>
+#undef f32subf64x
+
+#include <math-narrow.h>
+
+float
+__fsubl (long double x, long double y)
+{
+  NARROW_SUB_ROUND_TO_ODD (x, y, float, union ieee854_long_double, l,
+			   mantissa1);
+}
+libm_alias_float_ldouble (sub)
diff --git a/sysdeps/ieee754/ldbl-opt/Makefile b/sysdeps/ieee754/ldbl-opt/Makefile
index 2340e4ce29..7cdeaad2f2 100644
--- a/sysdeps/ieee754/ldbl-opt/Makefile
+++ b/sysdeps/ieee754/ldbl-opt/Makefile
@@ -44,7 +44,8 @@ libnldbl-calls = asprintf dprintf fprintf fscanf fwprintf fwscanf iovfscanf \
 		 isoc99_vwscanf isoc99_vfwscanf isoc99_vswscanf \
 		 nextup nextdown totalorder totalordermag getpayload \
 		 canonicalize setpayload setpayloadsig llogb fmaxmag fminmag \
-		 roundeven fromfp ufromfp fromfpx ufromfpx fadd dadd
+		 roundeven fromfp ufromfp fromfpx ufromfpx fadd dadd \
+		 fsub dsub
 libnldbl-routines = $(libnldbl-calls:%=nldbl-%)
 libnldbl-inhibit-o = $(object-suffixes)
 libnldbl-static-only-routines = $(libnldbl-routines)
@@ -87,6 +88,7 @@ CFLAGS-nldbl-csqrt.c = -fno-builtin-csqrtl
 CFLAGS-nldbl-ctan.c = -fno-builtin-ctanl
 CFLAGS-nldbl-ctanh.c = -fno-builtin-ctanhl
 CFLAGS-nldbl-dadd.c = -fno-builtin-daddl
+CFLAGS-nldbl-dsub.c = -fno-builtin-dsubl
 CFLAGS-nldbl-erf.c = -fno-builtin-erfl
 CFLAGS-nldbl-erfc.c = -fno-builtin-erfcl
 CFLAGS-nldbl-exp.c = -fno-builtin-expl
@@ -107,6 +109,7 @@ CFLAGS-nldbl-fmod.c = -fno-builtin-fmodl
 CFLAGS-nldbl-frexp.c = -fno-builtin-frexpl
 CFLAGS-nldbl-fromfp.c = -fno-builtin-fromfpl
 CFLAGS-nldbl-fromfpx.c = -fno-builtin-fromfpxl
+CFLAGS-nldbl-fsub.c = -fno-builtin-fsubl
 CFLAGS-nldbl-gamma.c = -fno-builtin-gammal
 CFLAGS-nldbl-getpayload.c = -fno-builtin-getpayloadl
 CFLAGS-nldbl-hypot.c = -fno-builtin-hypotl
diff --git a/sysdeps/ieee754/ldbl-opt/Versions b/sysdeps/ieee754/ldbl-opt/Versions
index e50efa3520..c33dc42d81 100644
--- a/sysdeps/ieee754/ldbl-opt/Versions
+++ b/sysdeps/ieee754/ldbl-opt/Versions
@@ -101,6 +101,6 @@ libm {
     # Functions taking long double = double argument and rounding
     # result to double (same as f32x*f64 functions, but those names
     # are not reserved in TS 18661-1).
-    __nldbl_daddl;
+    __nldbl_daddl; __nldbl_dsubl;
   }
 }
diff --git a/sysdeps/ieee754/ldbl-opt/nldbl-compat.h b/sysdeps/ieee754/ldbl-opt/nldbl-compat.h
index 74f0e459fa..4f5ae2dd26 100644
--- a/sysdeps/ieee754/ldbl-opt/nldbl-compat.h
+++ b/sysdeps/ieee754/ldbl-opt/nldbl-compat.h
@@ -104,5 +104,6 @@ extern void __nldbl___vsyslog_chk (int, int, const char *, va_list);
 /* The original declarations of these were hidden by the including
    file.  */
 extern double __nldbl_daddl (double, double) __THROW;
+extern double __nldbl_dsubl (double, double) __THROW;
 
 #endif /* __NLDBL_COMPAT_H */
diff --git a/sysdeps/ieee754/ldbl-opt/nldbl-dsub.c b/sysdeps/ieee754/ldbl-opt/nldbl-dsub.c
new file mode 100644
index 0000000000..1d31a75f13
--- /dev/null
+++ b/sysdeps/ieee754/ldbl-opt/nldbl-dsub.c
@@ -0,0 +1,28 @@
+/* Compatibility routine for IEEE double as long double for dsub.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#define dsubl __hide_dsubl
+#include "nldbl-compat.h"
+#undef dsubl
+
+double
+attribute_hidden
+dsubl (double x, double y)
+{
+  return __nldbl_dsubl (x, y);
+}
diff --git a/sysdeps/ieee754/ldbl-opt/nldbl-fsub.c b/sysdeps/ieee754/ldbl-opt/nldbl-fsub.c
new file mode 100644
index 0000000000..36dfe11a75
--- /dev/null
+++ b/sysdeps/ieee754/ldbl-opt/nldbl-fsub.c
@@ -0,0 +1,28 @@
+/* Compatibility routine for IEEE double as long double for fsub.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#define fsubl __hide_fsubl
+#include "nldbl-compat.h"
+#undef fsubl
+
+float
+attribute_hidden
+fsubl (double x, double y)
+{
+  return fsub (x, y);
+}
diff --git a/sysdeps/ieee754/soft-fp/s_dsubl.c b/sysdeps/ieee754/soft-fp/s_dsubl.c
new file mode 100644
index 0000000000..52f489c6ed
--- /dev/null
+++ b/sysdeps/ieee754/soft-fp/s_dsubl.c
@@ -0,0 +1,59 @@
+/* Subtract long double (ldbl-128) values, narrowing the result to
+   double, using soft-fp.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#define f32xsubf64x __hide_f32xsubf64x
+#define f32xsubf128 __hide_f32xsubf128
+#define f64subf64x __hide_f64subf64x
+#define f64subf128 __hide_f64subf128
+#include <math.h>
+#undef f32xsubf64x
+#undef f32xsubf128
+#undef f64subf64x
+#undef f64subf128
+
+#include <math-narrow.h>
+#include <soft-fp.h>
+#include <double.h>
+#include <quad.h>
+
+double
+__dsubl (_Float128 x, _Float128 y)
+{
+  FP_DECL_EX;
+  FP_DECL_Q (X);
+  FP_DECL_Q (Y);
+  FP_DECL_Q (R);
+  FP_DECL_D (RN);
+  double ret;
+
+  FP_INIT_ROUNDMODE;
+  FP_UNPACK_SEMIRAW_Q (X, x);
+  FP_UNPACK_SEMIRAW_Q (Y, y);
+  FP_SUB_Q (R, X, Y);
+#if (2 * _FP_W_TYPE_SIZE) < _FP_FRACBITS_Q
+  FP_TRUNC (D, Q, 2, 4, RN, R);
+#else
+  FP_TRUNC (D, Q, 1, 2, RN, R);
+#endif
+  FP_PACK_SEMIRAW_D (ret, RN);
+  FP_HANDLE_EXCEPTIONS;
+  CHECK_NARROW_SUB (ret, x, y);
+  return ret;
+}
+libm_alias_double_ldouble (sub)
diff --git a/sysdeps/ieee754/soft-fp/s_fsub.c b/sysdeps/ieee754/soft-fp/s_fsub.c
new file mode 100644
index 0000000000..824a3abbc7
--- /dev/null
+++ b/sysdeps/ieee754/soft-fp/s_fsub.c
@@ -0,0 +1,56 @@
+/* Subtract double values, narrowing the result to float, using soft-fp.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#define f32subf64 __hide_f32subf64
+#define f32subf32x __hide_f32subf32x
+#define fsubl __hide_fsubl
+#include <math.h>
+#undef f32subf64
+#undef f32subf32x
+#undef fsubl
+
+#include <math-narrow.h>
+#include <soft-fp.h>
+#include <single.h>
+#include <double.h>
+
+float
+__fsub (double x, double y)
+{
+  FP_DECL_EX;
+  FP_DECL_D (X);
+  FP_DECL_D (Y);
+  FP_DECL_D (R);
+  FP_DECL_S (RN);
+  float ret;
+
+  FP_INIT_ROUNDMODE;
+  FP_UNPACK_SEMIRAW_D (X, x);
+  FP_UNPACK_SEMIRAW_D (Y, y);
+  FP_SUB_D (R, X, Y);
+#if _FP_W_TYPE_SIZE < _FP_FRACBITS_D
+  FP_TRUNC (S, D, 1, 2, RN, R);
+#else
+  FP_TRUNC (S, D, 1, 1, RN, R);
+#endif
+  FP_PACK_SEMIRAW_S (ret, RN);
+  FP_HANDLE_EXCEPTIONS;
+  CHECK_NARROW_SUB (ret, x, y);
+  return ret;
+}
+libm_alias_float_double (sub)
diff --git a/sysdeps/ieee754/soft-fp/s_fsubl.c b/sysdeps/ieee754/soft-fp/s_fsubl.c
new file mode 100644
index 0000000000..84b1013797
--- /dev/null
+++ b/sysdeps/ieee754/soft-fp/s_fsubl.c
@@ -0,0 +1,55 @@
+/* Subtract long double (ldbl-128) values, narrowing the result to
+   float, using soft-fp.
+   Copyright (C) 2018 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
+   <http://www.gnu.org/licenses/>.  */
+
+#define f32subf64x __hide_f32subf64x
+#define f32subf128 __hide_f32subf128
+#include <math.h>
+#undef f32subf64x
+#undef f32subf128
+
+#include <math-narrow.h>
+#include <soft-fp.h>
+#include <single.h>
+#include <quad.h>
+
+float
+__fsubl (_Float128 x, _Float128 y)
+{
+  FP_DECL_EX;
+  FP_DECL_Q (X);
+  FP_DECL_Q (Y);
+  FP_DECL_Q (R);
+  FP_DECL_S (RN);
+  float ret;
+
+  FP_INIT_ROUNDMODE;
+  FP_UNPACK_SEMIRAW_Q (X, x);
+  FP_UNPACK_SEMIRAW_Q (Y, y);
+  FP_SUB_Q (R, X, Y);
+#if (2 * _FP_W_TYPE_SIZE) < _FP_FRACBITS_Q
+  FP_TRUNC (S, Q, 1, 4, RN, R);
+#else
+  FP_TRUNC (S, Q, 1, 2, RN, R);
+#endif
+  FP_PACK_SEMIRAW_S (ret, RN);
+  FP_HANDLE_EXCEPTIONS;
+  CHECK_NARROW_SUB (ret, x, y);
+  return ret;
+}
+libm_alias_float_ldouble (sub)