target-def.h (TARGET_DECIMAL_FLOAT_SUPPORTED_P): New.

2005-12-16  Jon Grimm  <jgrimm2@us.ibm.com>
	    Janis Johnson  <janis187@us.ibm.com>
	    Ben Elliston  <bje@au.ibm.com>

	* target-def.h (TARGET_DECIMAL_FLOAT_SUPPORTED_P): New.
	(TARGET_INITIALIZER): Add TARGET_DECIMAL_FLOAT_P.
	* target.h (struct gcc_target): Add decimal_float_supported_p.
	* targhooks.c (default_scalar_mode_supported_p): Handle
	MODE_DECIMAL_FLOAT.
	* builtins.def: Add new builtins for 32, 64 and 128 bit variants
	of inf, nan, finite, isinf and isnan.
	* builtin-types.def (BT_DFLOAT32, BT_DFLOAT64, BT_DFLOAT128,
	BT_DFLOAT32_PTR, BT_DFLOAT64_PTR, BT_DFLOAT128_PTR,
	BT_FN_DFLOAT32, BT_FN_DFLOAT64, BT_FN_DFLOAT128,
	BT_FN_INT_DFLOAT32, BT_FN_INT_DFLOAT64, BT_FN_INT_DFLOAT128,
	BT_FN_DFLOAT32_CONST_STRING, BT_FN_DFLOAT64_CONST_STRING,
	BT_FN_DFLOAT32_CONST_STRING, BT_FN_DFLOAT32_DFLOAT32,
	BT_FN_DFLOAT64_DFLOAT64, BT_FN_DFLOAT128_DFLOAT128): New.
	* c-decl.c (declspecs_add_type): Verify combos on type qualifiers.
	Pedwarn if decimal floating point types are used.  Error if
	decimal floating point is not supported by the target.
	(finish_declspecs): Return type from DFP typespec_word.
	* c-typeck.c (c_common_type): Choose the decimal floating point
	type with the greater precision when determining a common type.
	(convert_arguments): Warn if there is a mismatch between argument
	and prototype for decimal float types.  Warn of conversions with
	binary float types and of precision narrowing due to prototype.
	* c-parser.c (reswords): Add _Decimal32, _Decimal64, _Decimal128.
	(c_token_starts_typename): Handle RID_DFLOAT32/64/128.
	(c_token_starts_declspecs): Likewise.
	(c_parser_attributes): Likewise.
	* c-common.h (enum rid): Add new enumeration values RID_DFLOAT32,
	RID_DFLOAT64, RID_DFLOAT128.
	(T_D32, TEX_D32, T_D64, TEX_D64, T_D128, TEX_D128): New macros.
	* c-common.c (c_common_type_for_mode): Handle decimal float modes.
	(shorten_compare): Convert DFP/BFP operands to a common type.
	(c_common_modes_and_builtins): Register built-in decimal float
	types if the target supports them.
	(handle_mode_addtribute): Handle MODE_DECIMAL_FLOAT.
	* builtins.c (fold_builtin_1): Handle 32, 64 and 128 bit cases of
	inf, nan, finite, isinf and isnan builtins.
	* c-cppbuiltin.c (builtin_define_decimal_float_constants): New.
	(builtin_define_float_constants): Assert non-decimal radix.
	(c_cpp_builtins): Register built-in __DEC_EVAL_METHOD__ define.
	Call builtin_define_decimal_float_constants for each type.
	* c-lex.c (interpret_float): Decode decimal float types from CPP_N
	flags.  Use real_from_string3, which can handle binary or decimal
	floats.
	* c-tree.h (enum c_typespec_keyword): Add cts_dfloat32,
	cts_dfloat64, cts_dfloat128.
	* tree.c (build_common_tree_nodes_2): Add decimal float types.
	* tree.h (enum tree_index): Add new enumeration values
	TI_DFLOAT32_TYPE, TI_DFLOAT64_TYPE, TI_DFLOAT128_TYPE,
	TI_DFLOAT32_PTR_TYPE, TI_DFLOAT64_PTR_TYPE, TI_DFLOAT128_PTR_TYPE.
	(dfloat32_type_node): New macro.
	(dfloat64_type_node, dfloat128_type_node): Likewise.
	(dfloat32_ptr_type_node, dfloat64_ptr_type_node): Likewise.
	(dfloat128_ptr_type_node): Likewise.
	* c-pretty-print.c (pp_c_floating_constant): Append 32, 64 and 128
	bit decimal floating point types with "df", "dd" and "dl".
	* c-format.h (enum format_lengths): Add new enumeration values
	FMT_LEN_H, FMT_LEN_D and FMT_LEN_DD.
	* c-format.c (printf_length_specs, scanf_length_specs): Add
	entries for H, D, DD.
	(print_char_table, scan_char_table): Use new entries.
	(asm_fprintf_char_table, gcc_diag_char_table,
	gcc_cdiag_char_table, gcc_cxxdiag_char_table): Adjust for longer
	length arrays.
	* defaults.h (DECIMAL32_TYPE_SIZE): Define.
	(DECIMAL64_TYPE_SIZE): Likewise.
	(DECIMAL128_TYPE_SIZE): Likewise.
	(TARGET_DEC_EVAL_METHOD): Likewise.
	* doc/extend.texi (Decimal Float): New node.
	(Constructing Calls): Document decimal float built-ins.
	* doc/tm.texi: Document TARGET_DECIMAL_FLOAT_SUPPORTED_P hook.
	* Makefile.in (USER_H): Add $(srcdir)/ginclude/decfloat.h.
	* ginclude/decfloat.h: New file.

Co-Authored-By: Ben Elliston <bje@au.ibm.com>
Co-Authored-By: Janis Johnson <janis187@us.ibm.com>

From-SVN: r108629

1 files changed, 108 insertions, 0 deletions

diff --git a/gcc/ginclude/decfloat.h b/gcc/ginclude/decfloat.h
new file mode 100644
index 00000000000..fbf13463316
--- /dev/null
+++ b/gcc/ginclude/decfloat.h
@@ -0,0 +1,108 @@
+/* Copyright (C) 2005 Free Software Foundation, Inc.
+
+This file is part of GCC.
+
+GCC is free software; you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation; either version 2, or (at your option)
+any later version.
+
+GCC 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 General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with GCC; see the file COPYING.  If not, write to
+the Free Software Foundation, 51 Franklin Street, Fifth Floor,
+Boston, MA 02110-1301, USA.  */
+
+/* As a special exception, if you include this header file into source
+   files compiled by GCC, this header file does not by itself cause
+   the resulting executable to be covered by the GNU General Public
+   License.  This exception does not however invalidate any other
+   reasons why the executable file might be covered by the GNU General
+   Public License.  */
+
+/*
+ * Draft C Extension to support decimal floating-pointing arithmetic:  
+ * Characteristics of decimal floating types <decfloat.h>
+ */
+
+#ifndef _DECFLOAT_H___
+#define _DECFLOAT_H___
+
+/* Number of base-FLT_RADIX digits in the significand, p.  */
+#undef DEC32_MANT_DIG
+#undef DEC64_MANT_DIG
+#undef DEC128_MANT_DIG
+#define DEC32_MANT_DIG	__DEC32_MANT_DIG__
+#define DEC64_MANT_DIG	__DEC64_MANT_DIG__
+#define DEC128_MANT_DIG	__DEC128_MANT_DIG__
+
+/* Minimum exponent. */
+#undef DEC32_MIN_EXP
+#undef DEC64_MIN_EXP
+#undef DEC128_MIN_EXP
+#define DEC32_MIN_EXP	__DEC32_MIN_EXP__
+#define DEC64_MIN_EXP	__DEC64_MIN_EXP__
+#define DEC128_MIN_EXP	__DEC128_MIN_EXP__
+
+/* Maximum exponent. */
+#undef DEC32_MAX_EXP
+#undef DEC64_MAX_EXP
+#undef DEC128_MAX_EXP
+#define DEC32_MAX_EXP	__DEC32_MAX_EXP__
+#define DEC64_MAX_EXP	__DEC64_MAX_EXP__
+#define DEC128_MAX_EXP	__DEC128_MAX_EXP__
+
+/* Maximum representable finite decimal floating-point number
+   (there are 6, 15, and 33 9s after the decimal points respectively). */
+#undef DEC32_MAX
+#undef DEC64_MAX
+#undef DEC128_MAX
+#define DEC32_MAX   __DEC32_MAX__
+#define DEC64_MAX   __DEC64_MAX__
+#define DEC128_MAX  __DEC128_MAX__
+
+/* The difference between 1 and the least value greater than 1 that is
+   representable in the given floating point type. */
+#undef DEC32_EPSILON
+#undef DEC64_EPSILON
+#undef DEC128_EPSILON
+#define DEC32_EPSILON	__DEC32_EPSILON__
+#define DEC64_EPSILON	__DEC64_EPSILON__
+#define DEC128_EPSILON	__DEC128_EPSILON__
+
+/* Minimum normalized positive floating-point number. */
+#undef DEC32_MIN
+#undef DEC64_MIN
+#undef DEC128_MIN
+#define DEC32_MIN	__DEC32_MIN__
+#define DEC64_MIN	__DEC64_MIN__
+#define DEC128_MIN	__DEC128_MIN__
+
+/* Minimum denormalized positive floating-point number. */
+#undef DEC32_DEN
+#undef DEC64_DEN
+#undef DEC128_DEN
+#define DEC32_DEN       __DEC32_MIN__
+#define DEC64_DEN       __DEC64_MIN__
+#define DEC128_DEN      __DEC128_MIN__
+
+/* The floating-point expression evaluation method.
+         -1  indeterminate
+         0  evaluate all operations and constants just to the range and
+            precision of the type
+         1  evaluate operations and constants of type _Decimal32 
+	    and _Decimal64 to the range and precision of the _Decimal64 
+            type, evaluate _Decimal128 operations and constants to the 
+	    range and precision of the _Decimal128 type;
+	 2  evaluate all operations and constants to the range and
+	    precision of the _Decimal128 type.
+*/
+
+#undef DECFLT_EVAL_METHOD
+#define DECFLT_EVAL_METHOD	__DECFLT_EVAL_METHOD__
+
+#endif /* _DECFLOAT_H___ */