Add float128 support for ia64.

This patch enables float128 support for ia64, so that all the
configurations where GCC supports _Float128 / __float128 as an
ABI-distinct type now have glibc support as well.  bits/floatn.h
declares the support to be available for GCC 4.4 and later, which is
when the libgcc support was added.  The removal of
sysdeps/ia64/fpu/k_rem_pio2.c is because the generic k_rem_pio2.c
defines a function required by the float128 code.

Tested (compilation only) with build-many-glibcs.py for ia64 (GCC 6
and GCC 7).

Given how long it is since libm-test-ulps has been updated for ia64, I
think truncating the file and regenerating it from scratch would be a
good idea when doing a regeneration to add float128 ulps.  I expect
various ia64 libm issues (at least some already filed in Bugzilla) to
result in test failures even after ulps regeneration, but hopefully
the float128 code will pass tests as it's the same as used on other
architectures.

	* sysdeps/ia64/Implies: Add ieee754/float128.
	* sysdeps/ia64/bits/floatn.h: New file.
	* sysdeps/ia64/float128-abi.h: Likewise.
	* manual/math.texi (Mathematics): Document support for _Float128
	on ia64.
	* sysdeps/ia64/Makefile [$(subdir) = math] (CPPFLAGS): Append to
	Makefile variable.
	* sysdeps/ia64/fpu/e_sqrtf128.c: New file.
	* sysdeps/ia64/fpu/k_rem_pio2.c: Remove file.
	* sysdeps/ia64/fpu/sfp-machine.h: New file.  Based on libgcc.
	* sysdeps/ia64/math-tests.h: New file.
	* math/libm-test-support.h (XFAIL_FLOAT128_PAYLOAD): Also define
	based on TEST_COND_binary128 for [__ia64__].
	* sysdeps/unix/sysv/linux/ia64/libc.abilist: Update.
	* sysdeps/unix/sysv/linux/ia64/libm.abilist: Likewise,

10 files changed, 373 insertions, 1 deletions

diff --git a/sysdeps/ia64/Implies b/sysdeps/ia64/Implies
index 8524257020..22e4836e74 100644
--- a/sysdeps/ia64/Implies
+++ b/sysdeps/ia64/Implies
@@ -1,5 +1,6 @@
 wordsize-64
 # ia64 uses IEEE 754 floating point.
+ieee754/float128
 ieee754/ldbl-96
 ieee754/dbl-64
 ieee754/flt-32
diff --git a/sysdeps/ia64/Makefile b/sysdeps/ia64/Makefile
index 5ebca89afa..a4305d524d 100644
--- a/sysdeps/ia64/Makefile
+++ b/sysdeps/ia64/Makefile
@@ -1,6 +1,11 @@
 # The ia64 `long double' is a distinct type we support.
 long-double-fcts = yes
 
+ifeq ($(subdir),math)
+# sqrtf128 requires soft-fp.
+CPPFLAGS += -I../soft-fp
+endif
+
 ifeq ($(subdir),gmon)
 sysdep_routines += _mcount
 endif
diff --git a/sysdeps/ia64/bits/floatn.h b/sysdeps/ia64/bits/floatn.h
new file mode 100644
index 0000000000..c7b1df6d2c
--- /dev/null
+++ b/sysdeps/ia64/bits/floatn.h
@@ -0,0 +1,94 @@
+/* Macros to control TS 18661-3 glibc features on ia64.
+   Copyright (C) 2017 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/>.  */
+
+#ifndef _BITS_FLOATN_H
+#define _BITS_FLOATN_H
+
+#include <features.h>
+
+/* Defined to 1 if the current compiler invocation provides a
+   floating-point type with the IEEE 754 binary128 format, and this
+   glibc includes corresponding *f128 interfaces for it.  The required
+   libgcc support was added some time after the basic compiler
+   support.  */
+#if __GNUC_PREREQ (4, 4)
+# define __HAVE_FLOAT128 1
+#else
+# define __HAVE_FLOAT128 0
+#endif
+
+/* Defined to 1 if __HAVE_FLOAT128 is 1 and the type is ABI-distinct
+   from the default float, double and long double types in this glibc.  */
+#if __HAVE_FLOAT128
+# define __HAVE_DISTINCT_FLOAT128 1
+#else
+# define __HAVE_DISTINCT_FLOAT128 0
+#endif
+
+/* Defined to concatenate the literal suffix to be used with _Float128
+   types, if __HAVE_FLOAT128 is 1. */
+#if __HAVE_FLOAT128
+# if !__GNUC_PREREQ (7, 0) || defined __cplusplus
+/* The literal suffix f128 exists only since GCC 7.0.  */
+#  define __f128(x) x##q
+# else
+#  define __f128(x) x##f128
+# endif
+#endif
+
+/* Defined to a complex binary128 type if __HAVE_FLOAT128 is 1.  */
+#if __HAVE_FLOAT128
+# if !__GNUC_PREREQ (7, 0) || defined __cplusplus
+/* Add a typedef for older GCC compilers which don't natively support
+   _Complex _Float128.  */
+typedef _Complex float __cfloat128 __attribute__ ((__mode__ (__TC__)));
+#  define __CFLOAT128 __cfloat128
+# else
+#  define __CFLOAT128 _Complex _Float128
+# endif
+#endif
+
+/* The remaining of this file provides support for older compilers.  */
+#if __HAVE_FLOAT128
+
+/* The type _Float128 exists only since GCC 7.0.  */
+# if !__GNUC_PREREQ (7, 0) || defined __cplusplus
+typedef __float128 _Float128;
+# endif
+
+/* __builtin_huge_valf128 doesn't exist before GCC 7.0.  */
+# if !__GNUC_PREREQ (7, 0)
+#  define __builtin_huge_valf128() ((_Float128) __builtin_huge_val ())
+# endif
+
+/* Older GCC has only a subset of built-in functions for _Float128 on
+   ia64, and __builtin_infq is not usable in static initializers.
+   Converting a narrower sNaN to _Float128 produces a quiet NaN, so
+   attempts to use _Float128 sNaNs will not work properly with older
+   compilers.  */
+# if !__GNUC_PREREQ (7, 0)
+#  define __builtin_copysignf128 __builtin_copysignq
+#  define __builtin_fabsf128 __builtin_fabsq
+#  define __builtin_inff128() ((_Float128) __builtin_inf ())
+#  define __builtin_nanf128(x) ((_Float128) __builtin_nan (x))
+#  define __builtin_nansf128(x) ((_Float128) __builtin_nans (x))
+# endif
+
+#endif
+
+#endif /* _BITS_FLOATN_H */
diff --git a/sysdeps/ia64/float128-abi.h b/sysdeps/ia64/float128-abi.h
new file mode 100644
index 0000000000..6b954cc673
--- /dev/null
+++ b/sysdeps/ia64/float128-abi.h
@@ -0,0 +1,2 @@
+/* ABI version for _Float128 ABI introduction.  */
+#define FLOAT128_VERSION GLIBC_2.26
diff --git a/sysdeps/ia64/fpu/e_sqrtf128.c b/sysdeps/ia64/fpu/e_sqrtf128.c
new file mode 100644
index 0000000000..d5bcb79aaa
--- /dev/null
+++ b/sysdeps/ia64/fpu/e_sqrtf128.c
@@ -0,0 +1 @@
+#include <sysdeps/x86/fpu/e_sqrtf128.c>
diff --git a/sysdeps/ia64/fpu/k_rem_pio2.c b/sysdeps/ia64/fpu/k_rem_pio2.c
deleted file mode 100644
index 41254ae60a..0000000000
--- a/sysdeps/ia64/fpu/k_rem_pio2.c
+++ /dev/null
@@ -1 +0,0 @@
-/* Not needed. */
diff --git a/sysdeps/ia64/fpu/sfp-machine.h b/sysdeps/ia64/fpu/sfp-machine.h
new file mode 100644
index 0000000000..45e844d3da
--- /dev/null
+++ b/sysdeps/ia64/fpu/sfp-machine.h
@@ -0,0 +1,100 @@
+#define _FP_W_TYPE_SIZE		64
+#define _FP_W_TYPE		unsigned long
+#define _FP_WS_TYPE		signed long
+#define _FP_I_TYPE		long
+
+typedef int TItype __attribute__ ((mode (TI)));
+typedef unsigned int UTItype __attribute__ ((mode (TI)));
+
+#define TI_BITS (__CHAR_BIT__ * (int)sizeof(TItype))
+
+/* The type of the result of a floating point comparison.  This must
+   match `__libgcc_cmp_return__' in GCC for the target.  */
+typedef int __gcc_CMPtype __attribute__ ((mode (__libgcc_cmp_return__)));
+#define CMPtype __gcc_CMPtype
+
+#define _FP_MUL_MEAT_Q(R,X,Y)				\
+  _FP_MUL_MEAT_2_wide(_FP_WFRACBITS_Q,R,X,Y,umul_ppmm)
+
+#define _FP_DIV_MEAT_Q(R,X,Y)   _FP_DIV_MEAT_2_udiv(Q,R,X,Y)
+
+#define _FP_NANFRAC_S		_FP_QNANBIT_S
+#define _FP_NANFRAC_D		_FP_QNANBIT_D
+#define _FP_NANFRAC_E		_FP_QNANBIT_E, 0
+#define _FP_NANFRAC_Q		_FP_QNANBIT_Q, 0
+
+#define _FP_KEEPNANFRACP	1
+#define _FP_QNANNEGATEDP 0
+
+#define _FP_NANSIGN_S		1
+#define _FP_NANSIGN_D		1
+#define _FP_NANSIGN_E		1
+#define _FP_NANSIGN_Q		1
+
+/* Here is something Intel misdesigned: the specs don't define
+   the case where we have two NaNs with same mantissas, but
+   different sign. Different operations pick up different NaNs.  */
+#define _FP_CHOOSENAN(fs, wc, R, X, Y, OP)			\
+  do {								\
+    if (_FP_FRAC_GT_##wc(X, Y)					\
+	|| (_FP_FRAC_EQ_##wc(X,Y) && (OP == '+' || OP == '*')))	\
+      {								\
+	R##_s = X##_s;						\
+	_FP_FRAC_COPY_##wc(R,X);				\
+      }								\
+    else							\
+      {								\
+	R##_s = Y##_s;						\
+	_FP_FRAC_COPY_##wc(R,Y);				\
+      }								\
+    R##_c = FP_CLS_NAN;						\
+  } while (0)
+
+#define FP_EX_INVALID		0x01
+#define FP_EX_DENORM		0x02
+#define FP_EX_DIVZERO		0x04
+#define FP_EX_OVERFLOW		0x08
+#define FP_EX_UNDERFLOW		0x10
+#define FP_EX_INEXACT		0x20
+#define FP_EX_ALL \
+	(FP_EX_INVALID | FP_EX_DENORM | FP_EX_DIVZERO | FP_EX_OVERFLOW \
+	 | FP_EX_UNDERFLOW | FP_EX_INEXACT)
+
+#define _FP_TININESS_AFTER_ROUNDING 1
+
+void __sfp_handle_exceptions (int);
+
+#define FP_HANDLE_EXCEPTIONS			\
+  do {						\
+    if (__builtin_expect (_fex, 0))		\
+      __sfp_handle_exceptions (_fex);		\
+  } while (0);
+
+#define FP_TRAPPING_EXCEPTIONS	(~_fcw & FP_EX_ALL)
+
+#define FP_RND_NEAREST		0
+#define FP_RND_ZERO		0xc00L
+#define FP_RND_PINF		0x800L
+#define FP_RND_MINF		0x400L
+
+#define FP_RND_MASK		0xc00L
+
+#define _FP_DECL_EX \
+  unsigned long int _fcw __attribute__ ((unused)) = FP_RND_NEAREST
+
+#define FP_INIT_ROUNDMODE					\
+  do {								\
+    __asm__ __volatile__ ("mov.m %0 = ar.fpsr" : "=r" (_fcw));	\
+  } while (0)
+
+#define FP_ROUNDMODE		(_fcw & FP_RND_MASK)
+
+#define	__LITTLE_ENDIAN	1234
+#define	__BIG_ENDIAN	4321
+
+#define __BYTE_ORDER __LITTLE_ENDIAN
+
+/* Define ALIASNAME as a strong alias for NAME.  */
+#define strong_alias(name, aliasname) _strong_alias(name, aliasname)
+#define _strong_alias(name, aliasname) \
+  extern __typeof (name) aliasname __attribute__ ((alias (#name)));
diff --git a/sysdeps/ia64/math-tests.h b/sysdeps/ia64/math-tests.h
new file mode 100644
index 0000000000..439ac75e5d
--- /dev/null
+++ b/sysdeps/ia64/math-tests.h
@@ -0,0 +1,25 @@
+/* Configuration for math tests.  ia64 version.
+   Copyright (C) 2017 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/>.  */
+
+/* Before GCC 7, there is no built-in function to provide a __float128
+   sNaN, so most sNaN tests for this type cannot work.  */
+#if !__GNUC_PREREQ (7, 0)
+# define SNAN_TESTS_float128	0
+#endif
+
+#include_next <math-tests.h>
diff --git a/sysdeps/unix/sysv/linux/ia64/libc.abilist b/sysdeps/unix/sysv/linux/ia64/libc.abilist
index 39ccf2b5ee..62b67b859d 100644
--- a/sysdeps/unix/sysv/linux/ia64/libc.abilist
+++ b/sysdeps/unix/sysv/linux/ia64/libc.abilist
@@ -1885,11 +1885,18 @@ GLIBC_2.25 strfromd F
 GLIBC_2.25 strfromf F
 GLIBC_2.25 strfroml F
 GLIBC_2.26 GLIBC_2.26 A
+GLIBC_2.26 __strtof128_internal F
+GLIBC_2.26 __wcstof128_internal F
 GLIBC_2.26 preadv2 F
 GLIBC_2.26 preadv64v2 F
 GLIBC_2.26 pwritev2 F
 GLIBC_2.26 pwritev64v2 F
 GLIBC_2.26 reallocarray F
+GLIBC_2.26 strfromf128 F
+GLIBC_2.26 strtof128 F
+GLIBC_2.26 strtof128_l F
+GLIBC_2.26 wcstof128 F
+GLIBC_2.26 wcstof128_l F
 GLIBC_2.3 GLIBC_2.3 A
 GLIBC_2.3 __ctype_b_loc F
 GLIBC_2.3 __ctype_tolower_loc F
diff --git a/sysdeps/unix/sysv/linux/ia64/libm.abilist b/sysdeps/unix/sysv/linux/ia64/libm.abilist
index d7d80065fb..f165eb3957 100644
--- a/sysdeps/unix/sysv/linux/ia64/libm.abilist
+++ b/sysdeps/unix/sysv/linux/ia64/libm.abilist
@@ -402,4 +402,142 @@ GLIBC_2.25 ufromfpl F
 GLIBC_2.25 ufromfpx F
 GLIBC_2.25 ufromfpxf F
 GLIBC_2.25 ufromfpxl F
+GLIBC_2.26 GLIBC_2.26 A
+GLIBC_2.26 __acosf128_finite F
+GLIBC_2.26 __acoshf128_finite F
+GLIBC_2.26 __asinf128_finite F
+GLIBC_2.26 __atan2f128_finite F
+GLIBC_2.26 __atanhf128_finite F
+GLIBC_2.26 __coshf128_finite F
+GLIBC_2.26 __exp10f128_finite F
+GLIBC_2.26 __exp2f128_finite F
+GLIBC_2.26 __expf128_finite F
+GLIBC_2.26 __finitef128 F
+GLIBC_2.26 __fmodf128_finite F
+GLIBC_2.26 __fpclassifyf128 F
+GLIBC_2.26 __gammaf128_r_finite F
+GLIBC_2.26 __hypotf128_finite F
+GLIBC_2.26 __iseqsigf128 F
+GLIBC_2.26 __isinff128 F
+GLIBC_2.26 __isnanf128 F
+GLIBC_2.26 __issignalingf128 F
+GLIBC_2.26 __j0f128_finite F
+GLIBC_2.26 __j1f128_finite F
+GLIBC_2.26 __jnf128_finite F
+GLIBC_2.26 __lgammaf128_r_finite F
+GLIBC_2.26 __log10f128_finite F
+GLIBC_2.26 __log2f128_finite F
+GLIBC_2.26 __logf128_finite F
+GLIBC_2.26 __powf128_finite F
+GLIBC_2.26 __remainderf128_finite F
+GLIBC_2.26 __signbitf128 F
+GLIBC_2.26 __sinhf128_finite F
+GLIBC_2.26 __sqrtf128_finite F
+GLIBC_2.26 __y0f128_finite F
+GLIBC_2.26 __y1f128_finite F
+GLIBC_2.26 __ynf128_finite F
+GLIBC_2.26 acosf128 F
+GLIBC_2.26 acoshf128 F
+GLIBC_2.26 asinf128 F
+GLIBC_2.26 asinhf128 F
+GLIBC_2.26 atan2f128 F
+GLIBC_2.26 atanf128 F
+GLIBC_2.26 atanhf128 F
+GLIBC_2.26 cabsf128 F
+GLIBC_2.26 cacosf128 F
+GLIBC_2.26 cacoshf128 F
+GLIBC_2.26 canonicalizef128 F
+GLIBC_2.26 cargf128 F
+GLIBC_2.26 casinf128 F
+GLIBC_2.26 casinhf128 F
+GLIBC_2.26 catanf128 F
+GLIBC_2.26 catanhf128 F
+GLIBC_2.26 cbrtf128 F
+GLIBC_2.26 ccosf128 F
+GLIBC_2.26 ccoshf128 F
+GLIBC_2.26 ceilf128 F
+GLIBC_2.26 cexpf128 F
+GLIBC_2.26 cimagf128 F
+GLIBC_2.26 clog10f128 F
+GLIBC_2.26 clogf128 F
+GLIBC_2.26 conjf128 F
+GLIBC_2.26 copysignf128 F
+GLIBC_2.26 cosf128 F
+GLIBC_2.26 coshf128 F
+GLIBC_2.26 cpowf128 F
+GLIBC_2.26 cprojf128 F
+GLIBC_2.26 crealf128 F
+GLIBC_2.26 csinf128 F
+GLIBC_2.26 csinhf128 F
+GLIBC_2.26 csqrtf128 F
+GLIBC_2.26 ctanf128 F
+GLIBC_2.26 ctanhf128 F
+GLIBC_2.26 erfcf128 F
+GLIBC_2.26 erff128 F
+GLIBC_2.26 exp10f128 F
+GLIBC_2.26 exp2f128 F
+GLIBC_2.26 expf128 F
+GLIBC_2.26 expm1f128 F
+GLIBC_2.26 fabsf128 F
+GLIBC_2.26 fdimf128 F
+GLIBC_2.26 floorf128 F
+GLIBC_2.26 fmaf128 F
+GLIBC_2.26 fmaxf128 F
+GLIBC_2.26 fmaxmagf128 F
+GLIBC_2.26 fminf128 F
+GLIBC_2.26 fminmagf128 F
+GLIBC_2.26 fmodf128 F
+GLIBC_2.26 frexpf128 F
+GLIBC_2.26 fromfpf128 F
+GLIBC_2.26 fromfpxf128 F
+GLIBC_2.26 getpayloadf128 F
+GLIBC_2.26 hypotf128 F
+GLIBC_2.26 ilogbf128 F
+GLIBC_2.26 j0f128 F
+GLIBC_2.26 j1f128 F
+GLIBC_2.26 jnf128 F
+GLIBC_2.26 ldexpf128 F
+GLIBC_2.26 lgammaf128 F
+GLIBC_2.26 lgammaf128_r F
+GLIBC_2.26 llogbf128 F
+GLIBC_2.26 llrintf128 F
+GLIBC_2.26 llroundf128 F
+GLIBC_2.26 log10f128 F
+GLIBC_2.26 log1pf128 F
+GLIBC_2.26 log2f128 F
+GLIBC_2.26 logbf128 F
+GLIBC_2.26 logf128 F
+GLIBC_2.26 lrintf128 F
+GLIBC_2.26 lroundf128 F
+GLIBC_2.26 modff128 F
+GLIBC_2.26 nanf128 F
+GLIBC_2.26 nearbyintf128 F
+GLIBC_2.26 nextafterf128 F
+GLIBC_2.26 nextdownf128 F
+GLIBC_2.26 nextupf128 F
+GLIBC_2.26 powf128 F
+GLIBC_2.26 remainderf128 F
+GLIBC_2.26 remquof128 F
+GLIBC_2.26 rintf128 F
+GLIBC_2.26 roundevenf128 F
+GLIBC_2.26 roundf128 F
+GLIBC_2.26 scalblnf128 F
+GLIBC_2.26 scalbnf128 F
+GLIBC_2.26 setpayloadf128 F
+GLIBC_2.26 setpayloadsigf128 F
+GLIBC_2.26 sincosf128 F
+GLIBC_2.26 sinf128 F
+GLIBC_2.26 sinhf128 F
+GLIBC_2.26 sqrtf128 F
+GLIBC_2.26 tanf128 F
+GLIBC_2.26 tanhf128 F
+GLIBC_2.26 tgammaf128 F
+GLIBC_2.26 totalorderf128 F
+GLIBC_2.26 totalordermagf128 F
+GLIBC_2.26 truncf128 F
+GLIBC_2.26 ufromfpf128 F
+GLIBC_2.26 ufromfpxf128 F
+GLIBC_2.26 y0f128 F
+GLIBC_2.26 y1f128 F
+GLIBC_2.26 ynf128 F
 GLIBC_2.4 GLIBC_2.4 A