Add narrowing subtract functions [committed]

Message ID alpine.DEB.2.20.1803200035320.19523@digraph.polyomino.org.uk
State New
Headers show
Series
  • Add narrowing subtract functions [committed]
Related show

Commit Message

Joseph Myers March 20, 2018, 12:36 a.m.
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.  Committed.

(auto-libm-test-out-narrow-sub contents omitted below.)

2018-03-20  Joseph Myers  <joseph@codesourcery.com>

	* 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.


-- 
Joseph S. Myers
joseph@codesourcery.com

Patch

diff --git a/NEWS b/NEWS
index 5c0a3c6..9fb4c7a 100644
--- a/NEWS
+++ b/NEWS
@@ -15,6 +15,9 @@  Major new features:
   - fadd, faddl, daddl and corresponding fMaddfN, fMaddfNx, fMxaddfN and
     fMxaddfNx functions.
 
+  - fsub, fsubl, dsubl and corresponding fMsubfN, fMsubfNx, fMxsubfN and
+    fMxsubfNx functions.
+
 * Nominative and genitive month names are now supported for the Catalan and
   Czech languages.  The Catalan and Greek languages now support abbreviated
   alternative month names.
diff --git a/manual/arith.texi b/manual/arith.texi
index 39f2a5d..116788b 100644
--- a/manual/arith.texi
+++ b/manual/arith.texi
@@ -2210,6 +2210,25 @@  function without any intermediate rounding to the type of the
 arguments.
 @end deftypefun
 
+@deftypefun float fsub (double @var{x}, double @var{y})
+@deftypefunx float fsubl (long double @var{x}, long double @var{y})
+@deftypefunx double dsubl (long double @var{x}, long double @var{y})
+@deftypefunx _FloatM fMsubfN (_Float@var{N} @var{x}, _Float@var{N} @var{y})
+@deftypefunx _FloatM fMsubfNx (_Float@var{N}x @var{x}, _Float@var{N}x @var{y})
+@deftypefunx _FloatMx fMxsubfN (_Float@var{N} @var{x}, _Float@var{N} @var{y})
+@deftypefunx _FloatMx fMxsubfNx (_Float@var{N}x @var{x}, _Float@var{N}x @var{y})
+@standards{TS 18661-1:2014, math.h}
+@standardsx{fMsubfN, TS 18661-3:2015, math.h}
+@standardsx{fMsubfNx, TS 18661-3:2015, math.h}
+@standardsx{fMxsubfN, TS 18661-3:2015, math.h}
+@standardsx{fMxsubfNx, TS 18661-3:2015, math.h}
+@safety{@prelim{}@mtsafe{}@assafe{}@acsafe{}}
+These functions, from TS 18661-1:2014 and TS 18661-3:2015, return
+@math{@var{x} - @var{y}}, rounded once to the return type of the
+function without any intermediate rounding to the type of the
+arguments.
+@end deftypefun
+
 @node Complex Numbers
 @section Complex Numbers
 @pindex complex.h
diff --git a/math/Makefile b/math/Makefile
index 12dcedd..7cc0d16 100644
--- a/math/Makefile
+++ b/math/Makefile
@@ -89,7 +89,7 @@  libm-compat-calls =							\
 	w_lgammaF_r_compat w_lgammaF_compat2 w_expF_compat		\
 	w_lgamma_compatF k_standardF
 
-libm-narrow-fns = add
+libm-narrow-fns = add sub
 libm-narrow-types-basic = s_fF s_f32xFf64
 libm-narrow-types-ldouble-yes = s_fFl s_dFl
 libm-narrow-types-float128-yes = s_f32Ff128 s_f64Ff128 s_f64xFf128
@@ -284,7 +284,7 @@  libm-test-funcs-noauto = canonicalize ceil cimag conj copysign cproj creal \
 			 scalbln scalbn setpayload setpayloadsig signbit \
 			 significand totalorder totalordermag trunc ufromfp \
 			 ufromfpx
-libm-test-funcs-narrow = add
+libm-test-funcs-narrow = add sub
 libm-test-funcs-all = $(libm-test-funcs-auto) $(libm-test-funcs-noauto)
 libm-test-c-auto = $(foreach f,$(libm-test-funcs-auto),libm-test-$(f).c)
 libm-test-c-noauto = $(foreach f,$(libm-test-funcs-noauto),libm-test-$(f).c)
diff --git a/math/Versions b/math/Versions
index 9191c0f..87c7ff6 100644
--- a/math/Versions
+++ b/math/Versions
@@ -555,9 +555,14 @@  libm {
     # Functions not involving _Float64x or _Float128, for all configurations.
     fadd; faddl; daddl;
     f32addf32x; f32addf64; f32xaddf64;
+    fsub; fsubl; dsubl;
+    f32subf32x; f32subf64; f32xsubf64;
     # Functions involving _Float64x or _Float128, for some configurations.
     f32addf64x; f32addf128;
     f32xaddf64x; f32xaddf128; f64addf64x; f64addf128;
     f64xaddf128;
+    f32subf64x; f32subf128;
+    f32xsubf64x; f32xsubf128; f64subf64x; f64subf128;
+    f64xsubf128;
   }
 }
diff --git a/math/auto-libm-test-in b/math/auto-libm-test-in
index fc407e9..bb3c325 100644
--- a/math/auto-libm-test-in
+++ b/math/auto-libm-test-in
@@ -7195,6 +7195,101 @@  sinh 0x2.c5d37700c6bb03a8p+12 no-test-inline
 sinh 0x2.c5d37700c6bb03a6c24b6c9b494cp+12 no-test-inline
 sinh 0x2.c5d37700c6bb03a6c24b6c9b494ep+12 no-test-inline
 
+sub 0 0
+sub 0 -0
+sub -0 0
+sub -0 -0
+sub max max
+sub max -max
+sub -max max
+sub -max -max
+sub min min missing-underflow:arg-ibm128
+sub min -min missing-underflow:arg-ibm128
+sub -min min missing-underflow:arg-ibm128
+sub -min -min missing-underflow:arg-ibm128
+sub min_subnorm min_subnorm missing-underflow:arg-ibm128
+sub min_subnorm -min_subnorm missing-underflow:arg-ibm128
+sub -min_subnorm min_subnorm missing-underflow:arg-ibm128
+sub -min_subnorm -min_subnorm missing-underflow:arg-ibm128
+sub 1 2
+sub 1 -2
+sub -1 2
+sub -1 -2
+sub 100.5 0.75
+sub 100.5 -0.75
+sub -100.5 0.75
+sub -100.5 -0.75
+sub 1 0x1p-23
+sub 1 0x1.7fp-23
+sub 1 0x1.8p-23
+sub 1 0x1.81p-23
+sub 1 0x1p-24
+sub 1 0x1.1p-24
+sub 1 0x0.fp-24
+sub 1 min
+sub 1 -min
+sub 1 min_subnorm
+sub 1 -min_subnorm
+sub -1 min
+sub -1 -min
+sub -1 min_subnorm
+sub -1 -min_subnorm
+# Cases where larger argument is half way between two values of a
+# floating-point format, so that double rounding would sometimes yield
+# the wrong result.
+sub 0x1.000001p0 min
+sub 0x1.000001p0 -min
+sub 0x1.000001p0 min_subnorm
+sub 0x1.000001p0 -min_subnorm
+sub -0x1.000001p0 min
+sub -0x1.000001p0 -min
+sub -0x1.000001p0 min_subnorm
+sub -0x1.000001p0 -min_subnorm
+sub 0x1.00000000000008p0 min
+sub 0x1.00000000000008p0 -min
+sub 0x1.00000000000008p0 min_subnorm
+sub 0x1.00000000000008p0 -min_subnorm
+sub -0x1.00000000000008p0 min
+sub -0x1.00000000000008p0 -min
+sub -0x1.00000000000008p0 min_subnorm
+sub -0x1.00000000000008p0 -min_subnorm
+sub 0x1.0000000000000001p0 min
+sub 0x1.0000000000000001p0 -min
+sub 0x1.0000000000000001p0 min_subnorm
+sub 0x1.0000000000000001p0 -min_subnorm
+sub -0x1.0000000000000001p0 min
+sub -0x1.0000000000000001p0 -min
+sub -0x1.0000000000000001p0 min_subnorm
+sub -0x1.0000000000000001p0 -min_subnorm
+# Two values representable in the same format with difference very
+# close to half way between two representable values.  In particular,
+# verify that there is no intermediate rounding to 64 bits when
+# subtracting two double values.
+sub 1 0x1.000002p-24
+sub 1 -0x1.000002p-24
+sub 1 0x0.ffffffp-24
+sub 1 -0x0.ffffffp-24
+sub 0x1.000002p0 0x1.000002p-24
+sub 0x1.000002p0 -0x1.000002p-24
+sub 0x1.000002p0 0x0.ffffffp-24
+sub 0x1.000002p0 -0x0.ffffffp-24
+sub 1 0x1.0000000000001p-53
+sub 1 -0x1.0000000000001p-53
+sub 1 0x0.fffffffffffff8p-53
+sub 1 -0x0.fffffffffffff8p-53
+sub 0x1.0000000000001p0 0x1.0000000000001p-53
+sub 0x1.0000000000001p0 -0x1.0000000000001p-53
+sub 0x1.0000000000001p0 0x0.fffffffffffff8p-53
+sub 0x1.0000000000001p0 -0x0.fffffffffffff8p-53
+sub 1 0x1.0000000000000002p-64
+sub 1 -0x1.0000000000000002p-64
+sub 1 0x0.ffffffffffffffffp-64
+sub 1 -0x0.ffffffffffffffffp-64
+sub 0x1.0000000000000002p0 0x1.0000000000000002p-64
+sub 0x1.0000000000000002p0 -0x1.0000000000000002p-64
+sub 0x1.0000000000000002p0 0x0.ffffffffffffffffp-64
+sub 0x1.0000000000000002p0 -0x0.ffffffffffffffffp-64
+
 sqrt 0
 sqrt -0
 sqrt 2209
diff --git a/math/bits/mathcalls-narrow.h b/math/bits/mathcalls-narrow.h
index 642d094..1082843 100644
--- a/math/bits/mathcalls-narrow.h
+++ b/math/bits/mathcalls-narrow.h
@@ -22,3 +22,6 @@ 
 
 /* Add.  */
 __MATHCALL_NARROW (__MATHCALL_NAME (add), __MATHCALL_REDIR_NAME (add), 2);
+
+/* Subtract.  */
+__MATHCALL_NARROW (__MATHCALL_NAME (sub), __MATHCALL_REDIR_NAME (sub), 2);
diff --git a/math/gen-auto-libm-tests.c b/math/gen-auto-libm-tests.c
index 9c244de..cf157f5 100644
--- a/math/gen-auto-libm-tests.c
+++ b/math/gen-auto-libm-tests.c
@@ -584,6 +584,7 @@  static test_function test_functions[] =
     FUNC ("sincos", ARGS1 (type_fp), RET2 (type_fp, type_fp), false, false,
 	  false, CALC (mpfr_f_11, mpfr_sin_cos)),
     FUNC_mpfr_f_f ("sinh", mpfr_sinh, false),
+    FUNC_mpfr_ff_f ("sub", mpfr_sub, true),
     FUNC_mpfr_f_f ("sqrt", mpfr_sqrt, true),
     FUNC_mpfr_f_f ("tan", mpfr_tan, false),
     FUNC_mpfr_f_f ("tanh", mpfr_tanh, false),
diff --git a/math/libm-test-narrow-sub.inc b/math/libm-test-narrow-sub.inc
new file mode 100644
index 0000000..3ced7a6
--- /dev/null
+++ b/math/libm-test-narrow-sub.inc
@@ -0,0 +1,179 @@ 
+/* Test narrowing subtract.
+   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 "libm-test-driver.c"
+
+static const struct test_aa_f_data sub_test_data[] =
+  {
+    TEST_aa_f (sub, arg_qnan_value, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_qnan_value, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_qnan_value, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_qnan_value, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_qnan_value, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_qnan_value, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_qnan_value, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_qnan_value, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_snan_value, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_snan_value, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_snan_value, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_snan_value, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+
+    TEST_aa_f (sub, arg_qnan_value, arg_plus_zero, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_qnan_value, arg_minus_zero, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_qnan_value, arg_plus_infty, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_qnan_value, arg_minus_infty, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_qnan_value, arg_min_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_qnan_value, -arg_min_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_qnan_value, arg_min_subnorm_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_qnan_value, -arg_min_subnorm_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_qnan_value, arg_max_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_qnan_value, -arg_max_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_plus_zero, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_zero, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_plus_infty, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_infty, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_min_value, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_min_value, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_min_subnorm_value, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_min_subnorm_value, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_max_value, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_max_value, arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_qnan_value, arg_plus_zero, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_qnan_value, arg_minus_zero, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_qnan_value, arg_plus_infty, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_qnan_value, arg_minus_infty, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_qnan_value, arg_min_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_qnan_value, -arg_min_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_qnan_value, arg_min_subnorm_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_qnan_value, -arg_min_subnorm_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_qnan_value, arg_max_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_qnan_value, -arg_max_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_plus_zero, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_zero, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_plus_infty, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_infty, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_min_value, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_min_value, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_min_subnorm_value, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_min_subnorm_value, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_max_value, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_max_value, -arg_qnan_value, qnan_value, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+
+    TEST_aa_f (sub, arg_snan_value, arg_plus_zero, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_snan_value, arg_minus_zero, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_snan_value, arg_plus_infty, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_snan_value, arg_minus_infty, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_snan_value, arg_min_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_snan_value, -arg_min_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_snan_value, arg_min_subnorm_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_snan_value, -arg_min_subnorm_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_snan_value, arg_max_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_snan_value, -arg_max_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_plus_zero, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_minus_zero, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_plus_infty, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_minus_infty, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_min_value, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_min_value, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_min_subnorm_value, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_min_subnorm_value, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_max_value, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_max_value, arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, arg_plus_zero, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, arg_minus_zero, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, arg_plus_infty, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, arg_minus_infty, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, arg_min_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, -arg_min_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, arg_min_subnorm_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, -arg_min_subnorm_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, arg_max_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_snan_value, -arg_max_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_plus_zero, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_minus_zero, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_plus_infty, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_minus_infty, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_min_value, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_min_value, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_min_subnorm_value, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_min_subnorm_value, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, arg_max_value, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+    TEST_aa_f (sub, -arg_max_value, -arg_snan_value, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION),
+
+    TEST_aa_f (sub, arg_plus_infty, arg_plus_infty, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION|ERRNO_EDOM),
+    TEST_aa_f (sub, arg_plus_infty, arg_minus_infty, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_infty, arg_plus_infty, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_infty, arg_minus_infty, qnan_value, NO_INEXACT_EXCEPTION|INVALID_EXCEPTION|ERRNO_EDOM),
+
+    TEST_aa_f (sub, arg_plus_infty, arg_plus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_plus_infty, arg_minus_zero, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_plus_infty, arg_min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_plus_infty, -arg_min_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_plus_infty, arg_min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_plus_infty, -arg_min_subnorm_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_plus_infty, arg_max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+    TEST_aa_f (sub, arg_plus_infty, -arg_max_value, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+    TEST_aa_f (sub, arg_plus_zero, arg_plus_infty, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_zero, arg_plus_infty, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_min_value, arg_plus_infty, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_min_value, arg_plus_infty, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_min_subnorm_value, arg_plus_infty, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_min_subnorm_value, arg_plus_infty, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_max_value, arg_plus_infty, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_max_value, arg_plus_infty, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_infty, arg_plus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_infty, arg_minus_zero, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_infty, arg_min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_infty, -arg_min_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_infty, arg_min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_infty, -arg_min_subnorm_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_infty, arg_max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+    TEST_aa_f (sub, arg_minus_infty, -arg_max_value, minus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED|XFAIL_ROUNDING_IBM128_LIBGCC),
+    TEST_aa_f (sub, arg_plus_zero, arg_minus_infty, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_minus_zero, arg_minus_infty, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_min_value, arg_minus_infty, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_min_value, arg_minus_infty, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_min_subnorm_value, arg_minus_infty, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_min_subnorm_value, arg_minus_infty, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, arg_max_value, arg_minus_infty, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+    TEST_aa_f (sub, -arg_max_value, arg_minus_infty, plus_infty, NO_INEXACT_EXCEPTION|ERRNO_UNCHANGED),
+
+    AUTO_TESTS_aa_f (sub),
+  };
+
+static void
+sub_test (void)
+{
+  ALL_RM_TEST (sub, 1, sub_test_data, RUN_TEST_LOOP_aa_f, END);
+}
+
+static void
+do_test (void)
+{
+  sub_test ();
+}
+
+/*
+ * Local Variables:
+ * mode:c
+ * End:
+ */
diff --git a/math/math-narrow.h b/math/math-narrow.h
index 03397a6..a062b8d 100644
--- a/math/math-narrow.h
+++ b/math/math-narrow.h
@@ -107,6 +107,64 @@ 
     }						\
   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);		\
+									\
+      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)
+
 /* The following macros declare aliases for a narrowing function.  The
    sole argument is the base name of a family of functions, such as
    "add".  If any platform changes long double format after the
diff --git a/sysdeps/i386/fpu/s_f32xsubf64.c b/sysdeps/i386/fpu/s_f32xsubf64.c
new file mode 100644
index 0000000..3a7dbca
--- /dev/null
+++ b/sysdeps/i386/fpu/s_f32xsubf64.c
@@ -0,0 +1,41 @@ 
+/* Subtract _Float64 values, converting the result to _Float32x.  i386 version.
+   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 <fpu_control.h>
+#include <math_private.h>
+#include <math-narrow.h>
+
+_Float32x
+__f32xsubf64 (_Float64 x, _Float64 y)
+{
+  /* To avoid double rounding, set double precision for the subtraction.
+     math_narrow_eval is still needed to eliminate excess range in the
+     case of overflow.  If the result of the subtraction is in the
+     subnormal range for double, it is exact, so no issues of double
+     rounding for subnormals arise.  */
+  fpu_control_t cw, cw_double;
+  _FPU_GETCW (cw);
+  cw_double = (cw & ~_FPU_EXTENDED) | _FPU_DOUBLE;
+  _FPU_SETCW (cw_double);
+  _Float32x ret = math_narrow_eval (x - y);
+  _FPU_SETCW (cw);
+  CHECK_NARROW_SUB (ret, x, y);
+  return ret;
+}
+libm_alias_float32x_float64 (sub)
diff --git a/sysdeps/ieee754/dbl-64/s_f32xsubf64.c b/sysdeps/ieee754/dbl-64/s_f32xsubf64.c
new file mode 100644
index 0000000..568f43a
--- /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 0000000..c3b911e
--- /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 eebf19a..7606026 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 0000000..f1373f6
--- /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 0000000..89ef5a7
--- /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 0000000..64a0d6d
--- /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 0000000..46d0f95
--- /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 0000000..c9a63d9
--- /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 0000000..6a5e6b3
--- /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 0000000..e0e80f2
--- /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 0000000..aa6662d
--- /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 0000000..5a85579
--- /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 0000000..54aaf68
--- /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 2340e4c..7cdeaad 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 e50efa3..c33dc42 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 74f0e45..4f5ae2d 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 0000000..1d31a75
--- /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 0000000..36dfe11
--- /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 0000000..52f489c
--- /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 0000000..824a3ab
--- /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 0000000..84b1013
--- /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)
diff --git a/sysdeps/mach/hurd/i386/libm.abilist b/sysdeps/mach/hurd/i386/libm.abilist
index 71fc544..83cd23d 100644
--- a/sysdeps/mach/hurd/i386/libm.abilist
+++ b/sysdeps/mach/hurd/i386/libm.abilist
@@ -1026,16 +1026,29 @@  GLIBC_2.27 ynf64 F
 GLIBC_2.27 ynf64x F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
diff --git a/sysdeps/powerpc/fpu/libm-test-ulps b/sysdeps/powerpc/fpu/libm-test-ulps
index 6eec66f..784633c 100644
--- a/sysdeps/powerpc/fpu/libm-test-ulps
+++ b/sysdeps/powerpc/fpu/libm-test-ulps
@@ -2635,6 +2635,10 @@  Function: "sqrt_upward":
 ildouble: 1
 ldouble: 1
 
+Function: "sub_ldouble":
+double: 1
+float: 1
+
 Function: "tan":
 float: 3
 float128: 1
diff --git a/sysdeps/unix/sysv/linux/aarch64/libm.abilist b/sysdeps/unix/sysv/linux/aarch64/libm.abilist
index dbb5326..c88a3ee 100644
--- a/sysdeps/unix/sysv/linux/aarch64/libm.abilist
+++ b/sysdeps/unix/sysv/linux/aarch64/libm.abilist
@@ -988,15 +988,28 @@  GLIBC_2.27 ynf64 F
 GLIBC_2.27 ynf64x F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
diff --git a/sysdeps/unix/sysv/linux/alpha/libm.abilist b/sysdeps/unix/sysv/linux/alpha/libm.abilist
index 00a7b8d..9b32581 100644
--- a/sysdeps/unix/sysv/linux/alpha/libm.abilist
+++ b/sysdeps/unix/sysv/linux/alpha/libm.abilist
@@ -998,19 +998,33 @@  GLIBC_2.27 ynf64 F
 GLIBC_2.27 ynf64x F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 __nldbl_daddl F
+GLIBC_2.28 __nldbl_dsubl F
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.3.4 GLIBC_2.3.4 A
 GLIBC_2.3.4 __c1_cabsf F
 GLIBC_2.3.4 __c1_cacosf F
diff --git a/sysdeps/unix/sysv/linux/arm/libm.abilist b/sysdeps/unix/sysv/linux/arm/libm.abilist
index 67ed656..543aaa7 100644
--- a/sysdeps/unix/sysv/linux/arm/libm.abilist
+++ b/sysdeps/unix/sysv/linux/arm/libm.abilist
@@ -437,11 +437,17 @@  GLIBC_2.27 ynf32x F
 GLIBC_2.27 ynf64 F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
 GLIBC_2.28 f32xaddf64 F
+GLIBC_2.28 f32xsubf64 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
 GLIBC_2.4 _LIB_VERSION D 0x4
 GLIBC_2.4 __clog10 F
diff --git a/sysdeps/unix/sysv/linux/hppa/libm.abilist b/sysdeps/unix/sysv/linux/hppa/libm.abilist
index 4cd41e3..399d1b1 100644
--- a/sysdeps/unix/sysv/linux/hppa/libm.abilist
+++ b/sysdeps/unix/sysv/linux/hppa/libm.abilist
@@ -749,10 +749,16 @@  GLIBC_2.27 ynf32x F
 GLIBC_2.27 ynf64 F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
 GLIBC_2.28 f32xaddf64 F
+GLIBC_2.28 f32xsubf64 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
 GLIBC_2.4 exp2l F
diff --git a/sysdeps/unix/sysv/linux/i386/libm.abilist b/sysdeps/unix/sysv/linux/i386/libm.abilist
index b9aa407..1eecf41 100644
--- a/sysdeps/unix/sysv/linux/i386/libm.abilist
+++ b/sysdeps/unix/sysv/linux/i386/libm.abilist
@@ -1035,16 +1035,29 @@  GLIBC_2.27 ynf64 F
 GLIBC_2.27 ynf64x F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
diff --git a/sysdeps/unix/sysv/linux/ia64/libm.abilist b/sysdeps/unix/sysv/linux/ia64/libm.abilist
index a7981b8..debb9d8 100644
--- a/sysdeps/unix/sysv/linux/ia64/libm.abilist
+++ b/sysdeps/unix/sysv/linux/ia64/libm.abilist
@@ -964,16 +964,29 @@  GLIBC_2.27 ynf64 F
 GLIBC_2.27 ynf64x F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
diff --git a/sysdeps/unix/sysv/linux/m68k/coldfire/libm.abilist b/sysdeps/unix/sysv/linux/m68k/coldfire/libm.abilist
index 67ed656..543aaa7 100644
--- a/sysdeps/unix/sysv/linux/m68k/coldfire/libm.abilist
+++ b/sysdeps/unix/sysv/linux/m68k/coldfire/libm.abilist
@@ -437,11 +437,17 @@  GLIBC_2.27 ynf32x F
 GLIBC_2.27 ynf64 F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
 GLIBC_2.28 f32xaddf64 F
+GLIBC_2.28 f32xsubf64 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
 GLIBC_2.4 _LIB_VERSION D 0x4
 GLIBC_2.4 __clog10 F
diff --git a/sysdeps/unix/sysv/linux/m68k/m680x0/libm.abilist b/sysdeps/unix/sysv/linux/m68k/m680x0/libm.abilist
index e833cdb..310085d 100644
--- a/sysdeps/unix/sysv/linux/m68k/m680x0/libm.abilist
+++ b/sysdeps/unix/sysv/linux/m68k/m680x0/libm.abilist
@@ -791,9 +791,15 @@  GLIBC_2.27 ynf32x F
 GLIBC_2.27 ynf64 F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
 GLIBC_2.28 f32xaddf64 F
+GLIBC_2.28 f32xsubf64 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
diff --git a/sysdeps/unix/sysv/linux/microblaze/libm.abilist b/sysdeps/unix/sysv/linux/microblaze/libm.abilist
index 3d779cb..c45bc41 100644
--- a/sysdeps/unix/sysv/linux/microblaze/libm.abilist
+++ b/sysdeps/unix/sysv/linux/microblaze/libm.abilist
@@ -748,8 +748,14 @@  GLIBC_2.27 ynf32x F
 GLIBC_2.27 ynf64 F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
 GLIBC_2.28 f32xaddf64 F
+GLIBC_2.28 f32xsubf64 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
diff --git a/sysdeps/unix/sysv/linux/mips/mips32/libm.abilist b/sysdeps/unix/sysv/linux/mips/mips32/libm.abilist
index 6b86a2d..0038994 100644
--- a/sysdeps/unix/sysv/linux/mips/mips32/libm.abilist
+++ b/sysdeps/unix/sysv/linux/mips/mips32/libm.abilist
@@ -750,11 +750,17 @@  GLIBC_2.27 ynf32x F
 GLIBC_2.27 ynf64 F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
 GLIBC_2.28 f32xaddf64 F
+GLIBC_2.28 f32xsubf64 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
 GLIBC_2.4 exp2l F
 _gp_disp _gp_disp A
diff --git a/sysdeps/unix/sysv/linux/mips/mips64/libm.abilist b/sysdeps/unix/sysv/linux/mips/mips64/libm.abilist
index 137e5f1..47044e6 100644
--- a/sysdeps/unix/sysv/linux/mips/mips64/libm.abilist
+++ b/sysdeps/unix/sysv/linux/mips/mips64/libm.abilist
@@ -990,16 +990,29 @@  GLIBC_2.27 ynf64 F
 GLIBC_2.27 ynf64x F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
diff --git a/sysdeps/unix/sysv/linux/nios2/libm.abilist b/sysdeps/unix/sysv/linux/nios2/libm.abilist
index 736c70f..64b40b6 100644
--- a/sysdeps/unix/sysv/linux/nios2/libm.abilist
+++ b/sysdeps/unix/sysv/linux/nios2/libm.abilist
@@ -748,8 +748,14 @@  GLIBC_2.27 ynf32x F
 GLIBC_2.27 ynf64 F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
 GLIBC_2.28 f32xaddf64 F
+GLIBC_2.28 f32xsubf64 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
diff --git a/sysdeps/unix/sysv/linux/powerpc/powerpc32/fpu/libm.abilist b/sysdeps/unix/sysv/linux/powerpc/powerpc32/fpu/libm.abilist
index ea84f1e..d39a1b3 100644
--- a/sysdeps/unix/sysv/linux/powerpc/powerpc32/fpu/libm.abilist
+++ b/sysdeps/unix/sysv/linux/powerpc/powerpc32/fpu/libm.abilist
@@ -793,12 +793,19 @@  GLIBC_2.27 ynf32x F
 GLIBC_2.27 ynf64 F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 __nldbl_daddl F
+GLIBC_2.28 __nldbl_dsubl F
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
 GLIBC_2.28 f32xaddf64 F
+GLIBC_2.28 f32xsubf64 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
 GLIBC_2.4 __clog10l F
 GLIBC_2.4 __finitel F
diff --git a/sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu/libm.abilist b/sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu/libm.abilist
index ee2593c..1bbb215 100644
--- a/sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu/libm.abilist
+++ b/sysdeps/unix/sysv/linux/powerpc/powerpc32/nofpu/libm.abilist
@@ -792,12 +792,19 @@  GLIBC_2.27 ynf32x F
 GLIBC_2.27 ynf64 F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 __nldbl_daddl F
+GLIBC_2.28 __nldbl_dsubl F
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
 GLIBC_2.28 f32xaddf64 F
+GLIBC_2.28 f32xsubf64 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
 GLIBC_2.4 __clog10l F
 GLIBC_2.4 __finitel F
diff --git a/sysdeps/unix/sysv/linux/powerpc/powerpc64/libm-le.abilist b/sysdeps/unix/sysv/linux/powerpc/powerpc64/libm-le.abilist
index 075f026..977021c 100644
--- a/sysdeps/unix/sysv/linux/powerpc/powerpc64/libm-le.abilist
+++ b/sysdeps/unix/sysv/linux/powerpc/powerpc64/libm-le.abilist
@@ -1029,16 +1029,30 @@  GLIBC_2.27 ynf64 F
 GLIBC_2.27 ynf64x F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 __nldbl_daddl F
+GLIBC_2.28 __nldbl_dsubl F
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
diff --git a/sysdeps/unix/sysv/linux/powerpc/powerpc64/libm.abilist b/sysdeps/unix/sysv/linux/powerpc/powerpc64/libm.abilist
index 9e6c7d3..ae8d4e8 100644
--- a/sysdeps/unix/sysv/linux/powerpc/powerpc64/libm.abilist
+++ b/sysdeps/unix/sysv/linux/powerpc/powerpc64/libm.abilist
@@ -468,12 +468,19 @@  GLIBC_2.27 ynf32x F
 GLIBC_2.27 ynf64 F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 __nldbl_daddl F
+GLIBC_2.28 __nldbl_dsubl F
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
 GLIBC_2.28 f32xaddf64 F
+GLIBC_2.28 f32xsubf64 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.3 GLIBC_2.3 A
 GLIBC_2.3 _LIB_VERSION D 0x4
 GLIBC_2.3 __clog10 F
diff --git a/sysdeps/unix/sysv/linux/riscv/rv64/libm.abilist b/sysdeps/unix/sysv/linux/riscv/rv64/libm.abilist
index 9a6bcb1..d6a2bad 100644
--- a/sysdeps/unix/sysv/linux/riscv/rv64/libm.abilist
+++ b/sysdeps/unix/sysv/linux/riscv/rv64/libm.abilist
@@ -970,15 +970,28 @@  GLIBC_2.27 ynf64x F
 GLIBC_2.27 ynl F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
diff --git a/sysdeps/unix/sysv/linux/s390/s390-32/libm.abilist b/sysdeps/unix/sysv/linux/s390/s390-32/libm.abilist
index 7b1854d..0f94d18 100644
--- a/sysdeps/unix/sysv/linux/s390/s390-32/libm.abilist
+++ b/sysdeps/unix/sysv/linux/s390/s390-32/libm.abilist
@@ -988,19 +988,33 @@  GLIBC_2.27 ynf64 F
 GLIBC_2.27 ynf64x F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 __nldbl_daddl F
+GLIBC_2.28 __nldbl_dsubl F
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
 GLIBC_2.4 __clog10l F
 GLIBC_2.4 __finitel F
diff --git a/sysdeps/unix/sysv/linux/s390/s390-64/libm.abilist b/sysdeps/unix/sysv/linux/s390/s390-64/libm.abilist
index 24b4c21..87168e2 100644
--- a/sysdeps/unix/sysv/linux/s390/s390-64/libm.abilist
+++ b/sysdeps/unix/sysv/linux/s390/s390-64/libm.abilist
@@ -986,19 +986,33 @@  GLIBC_2.27 ynf64 F
 GLIBC_2.27 ynf64x F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 __nldbl_daddl F
+GLIBC_2.28 __nldbl_dsubl F
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
 GLIBC_2.4 __clog10l F
 GLIBC_2.4 __finitel F
diff --git a/sysdeps/unix/sysv/linux/sh/libm.abilist b/sysdeps/unix/sysv/linux/sh/libm.abilist
index 011c2df..d1de781 100644
--- a/sysdeps/unix/sysv/linux/sh/libm.abilist
+++ b/sysdeps/unix/sysv/linux/sh/libm.abilist
@@ -749,10 +749,16 @@  GLIBC_2.27 ynf32x F
 GLIBC_2.27 ynf64 F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
 GLIBC_2.28 f32xaddf64 F
+GLIBC_2.28 f32xsubf64 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
 GLIBC_2.4 exp2l F
diff --git a/sysdeps/unix/sysv/linux/sparc/sparc32/libm.abilist b/sysdeps/unix/sysv/linux/sparc/sparc32/libm.abilist
index fdc125a..6c526a3 100644
--- a/sysdeps/unix/sysv/linux/sparc/sparc32/libm.abilist
+++ b/sysdeps/unix/sysv/linux/sparc/sparc32/libm.abilist
@@ -995,19 +995,33 @@  GLIBC_2.27 ynf64 F
 GLIBC_2.27 ynf64x F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 __nldbl_daddl F
+GLIBC_2.28 __nldbl_dsubl F
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
 GLIBC_2.4 __clog10l F
 GLIBC_2.4 __finitel F
diff --git a/sysdeps/unix/sysv/linux/sparc/sparc64/libm.abilist b/sysdeps/unix/sysv/linux/sparc/sparc64/libm.abilist
index 1fc848a..18f6551 100644
--- a/sysdeps/unix/sysv/linux/sparc/sparc64/libm.abilist
+++ b/sysdeps/unix/sysv/linux/sparc/sparc64/libm.abilist
@@ -989,16 +989,29 @@  GLIBC_2.27 ynf64 F
 GLIBC_2.27 ynf64x F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
diff --git a/sysdeps/unix/sysv/linux/tile/tilegx32/libm.abilist b/sysdeps/unix/sysv/linux/tile/tilegx32/libm.abilist
index a7c37e2..73ceb5a 100644
--- a/sysdeps/unix/sysv/linux/tile/tilegx32/libm.abilist
+++ b/sysdeps/unix/sysv/linux/tile/tilegx32/libm.abilist
@@ -749,8 +749,14 @@  GLIBC_2.27 ynf32x F
 GLIBC_2.27 ynf64 F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
 GLIBC_2.28 f32xaddf64 F
+GLIBC_2.28 f32xsubf64 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
diff --git a/sysdeps/unix/sysv/linux/tile/tilegx64/libm.abilist b/sysdeps/unix/sysv/linux/tile/tilegx64/libm.abilist
index a7c37e2..73ceb5a 100644
--- a/sysdeps/unix/sysv/linux/tile/tilegx64/libm.abilist
+++ b/sysdeps/unix/sysv/linux/tile/tilegx64/libm.abilist
@@ -749,8 +749,14 @@  GLIBC_2.27 ynf32x F
 GLIBC_2.27 ynf64 F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
 GLIBC_2.28 f32xaddf64 F
+GLIBC_2.28 f32xsubf64 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
diff --git a/sysdeps/unix/sysv/linux/x86_64/64/libm.abilist b/sysdeps/unix/sysv/linux/x86_64/64/libm.abilist
index 0d8c448..e4d0fa8 100644
--- a/sysdeps/unix/sysv/linux/x86_64/64/libm.abilist
+++ b/sysdeps/unix/sysv/linux/x86_64/64/libm.abilist
@@ -1024,16 +1024,29 @@  GLIBC_2.27 ynf64 F
 GLIBC_2.27 ynf64x F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F
 GLIBC_2.4 GLIBC_2.4 A
diff --git a/sysdeps/unix/sysv/linux/x86_64/x32/libm.abilist b/sysdeps/unix/sysv/linux/x86_64/x32/libm.abilist
index e7f368e..543b5f8 100644
--- a/sysdeps/unix/sysv/linux/x86_64/x32/libm.abilist
+++ b/sysdeps/unix/sysv/linux/x86_64/x32/libm.abilist
@@ -1023,15 +1023,28 @@  GLIBC_2.27 ynf64 F
 GLIBC_2.27 ynf64x F
 GLIBC_2.28 GLIBC_2.28 A
 GLIBC_2.28 daddl F
+GLIBC_2.28 dsubl F
 GLIBC_2.28 f32addf128 F
 GLIBC_2.28 f32addf32x F
 GLIBC_2.28 f32addf64 F
 GLIBC_2.28 f32addf64x F
+GLIBC_2.28 f32subf128 F
+GLIBC_2.28 f32subf32x F
+GLIBC_2.28 f32subf64 F
+GLIBC_2.28 f32subf64x F
 GLIBC_2.28 f32xaddf128 F
 GLIBC_2.28 f32xaddf64 F
 GLIBC_2.28 f32xaddf64x F
+GLIBC_2.28 f32xsubf128 F
+GLIBC_2.28 f32xsubf64 F
+GLIBC_2.28 f32xsubf64x F
 GLIBC_2.28 f64addf128 F
 GLIBC_2.28 f64addf64x F
+GLIBC_2.28 f64subf128 F
+GLIBC_2.28 f64subf64x F
 GLIBC_2.28 f64xaddf128 F
+GLIBC_2.28 f64xsubf128 F
 GLIBC_2.28 fadd F
 GLIBC_2.28 faddl F
+GLIBC_2.28 fsub F
+GLIBC_2.28 fsubl F