This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [patch 1/4] libdecnumber support


On Thu, 2007-10-11 at 11:54 -0400, Daniel Jacobowitz wrote: 
> I'll review all these now.

Thank you very much.

> Before you check them in, though, give
> me a chance to get libdecnumber merged into the src repository.  It
> shouldn't take long.

Thanks for taking care of this.

> On Thu, Sep 20, 2007 at 06:54:11PM -0300, Thiago Jung Bauermann wrote:
> > +#include <ctype.h>
> > +#include <endian.h>
> > +#include "defs.h"
> > +#include "dfp.h"
> 
> What do you need ctype.h for?  That will tell me whether you really
> want ctype.h or libiberty's locale-independent safe-ctype.h.
> 
> defs.h should always be first.

It turned out that ctype.h was not needed at all. I just removed it and
it made no difference. Probably a left-over from an older version of the
code.

> endian.h is not portable.  You should probably use the autoconf macro
> AC_C_BIGENDIAN.

Ok, I'm using it now. When I run autoreconf inside the src/gdb
directory, I get the following:

$ autoreconf
autoreconf: cannot create `cd $srcdir;pwd`/../../../..: No such file or
directory
autoreconf: cannot create `cd $srcdir;pwd`/../../..: No such file or
directory
autoreconf: cannot create `cd $srcdir;pwd`/..: No such file or directory
$ echo $?
0

Is this a problem? Things still seem to work fine. I used autoconf
version 2.61.

> > +/* Convert decimal type to its string representation.  LEN is the length
> > +   of the decimal type, 4 bytes for decimal32, 8 bytes for decimal64 and
> > +   16 bytes for decimal128.  */
> > +void
> > +decimal_to_string (const uint8_t *decbytes, int len, char *s)
> > +{
> > +  uint8_t *dec = (uint8_t *)malloc (len);
> 
> You didn't include anything that would give you uint8_t; use
> gdb_byte.

I included dfp.h, which included stdint.h. I see now that GDB doesn't
use C99 yet, so I changed to gdb_byte.

> Also we never use malloc, only xmalloc.  Or you could avoid
> the allocation by using gdb_byte dec[16].

Good idea. Just changed to use temporary array as suggested.

> > +        break;
> > +      default:
> > +	free (dec);
> 
> Spaces and tabs mixed up?  xfree, like xmalloc.

Fixed. No '\t' in the file now. By the way, what's the policy in GDB
w.r.t. tabs and spaces? Never use tabs?

> > +	error(_("Unknown decimal floating point type.\n"));
> 
> Space before parentheses please.

Fixed.

> > +/* There is a project intended to add DFP support into GCC, described in
> > +   http://gcc.gnu.org/wiki/Decimal%20Floating-Point.  This file is intended
> > +   to add DFP support into GDB.  */
> 
> This comment doesn't belong here.  The file is part of GDB once the
> patch is applied.

Ok, removed comment.

> > +#ifndef DFP_H
> > +#define DFP_H
> > +#include <string.h>
> > +#include <stdio.h>
> > +#include <stdint.h>
> 
> Don't include standard headers here; stick to what defs.h provides.
> Just as well, since neither string.h nor stdint.h is portable.

Ok, removed.
-- 
[]'s
Thiago Jung Bauermann
Software Engineer
IBM Linux Technology Center
Subject: libdecnumber support

This patch provides functions to interact with libdecnumber and changes
the Makefile to compile it.

2007-10-15  Wu Zhou  <woodzltc@cn.ibm.com> 
	    Thiago Jung Bauermann  <bauerman@br.ibm.com>

	* Makefile.in (LIBDECNUMBER_DIR, LIBDECNUMBER, LIBDECNUMBER_SRC
	LIBDECNUMBER_CFLAGS): New macros for libdecnumber.
	(INTERNAL_CFLAGS_BASE): Add LIBDECNUMBER_CFLAGS in.
	(INSTALLED_LIBS): Add -ldecnumber in.
	(CLIBS): Add LIBDECNUMBER in.
	(decimal128_h, decimal64_h, decimal32_h): New macros for decimal
	headers.
	(dfp_h): New macros for decimal floating point.
	(dfp.o): New target.
	(COMMON_OBS): Add dfp.o in.
	(c-exp.o): Add dfp_h as dependency.
	(valprint.o): Add dfp_h as dependency.
	(value.o): Add dfp_h as dependency.
	* dfp.h: New header file for decimal floating point support in
	GDB.
	* dfp.c: New source file for decimal floating point support in
	GDB.  Implement decimal_from_string and decimal_to_string based
	on libdecnumber API.
	* configure.ac: Add AC_C_BIGENDIAN test.
 
Index: gdb/Makefile.in
===================================================================
--- gdb/Makefile.in.orig	2007-10-15 14:25:14.000000000 -0200
+++ gdb/Makefile.in	2007-10-15 14:35:06.000000000 -0200
@@ -125,6 +125,12 @@ BFD = $(BFD_DIR)/libbfd.a
 BFD_SRC = $(srcdir)/$(BFD_DIR)
 BFD_CFLAGS = -I$(BFD_DIR) -I$(BFD_SRC)
 
+# Where is the decnumber library?  Typically in ../libdecnumber.
+LIBDECNUMBER_DIR = ../libdecnumber
+LIBDECNUMBER = $(LIBDECNUMBER_DIR)/libdecnumber.a
+LIBDECNUMBER_SRC = $(srcdir)/$(LIBDECNUMBER_DIR)
+LIBDECNUMBER_CFLAGS = -I$(LIBDECNUMBER_DIR) -I$(LIBDECNUMBER_SRC)
+
 # Where is the READLINE library?  Typically in ../readline.
 READLINE_DIR = ../readline
 READLINE_SRC = $(srcdir)/$(READLINE_DIR)
@@ -358,7 +364,7 @@ CXXFLAGS = -g -O
 INTERNAL_CFLAGS_BASE = \
 	$(CFLAGS) $(GLOBAL_CFLAGS) $(PROFILE_CFLAGS) \
 	$(GDB_CFLAGS) $(OPCODES_CFLAGS) $(READLINE_CFLAGS) \
-	$(BFD_CFLAGS) $(INCLUDE_CFLAGS) \
+	$(BFD_CFLAGS) $(INCLUDE_CFLAGS) $(LIBDECNUMBER_CFLAGS) \
 	$(INTL_CFLAGS) $(ENABLE_CFLAGS) $(INTERNAL_CPPFLAGS)
 INTERNAL_WARN_CFLAGS = $(INTERNAL_CFLAGS_BASE) $(GDB_WARN_CFLAGS)
 INTERNAL_CFLAGS = $(INTERNAL_WARN_CFLAGS) $(GDB_WERROR_CFLAGS)
@@ -381,10 +387,10 @@ INTERNAL_LDFLAGS = $(CFLAGS) $(GLOBAL_CF
 # LIBIBERTY appears twice on purpose.
 # If you have the Cygnus libraries installed,
 # you can use 'CLIBS=$(INSTALLED_LIBS)' 'CDEPS='
-INSTALLED_LIBS=-lbfd -lreadline -lopcodes -liberty \
+INSTALLED_LIBS=-lbfd -lreadline -lopcodes -liberty -ldecnumber \
 	$(XM_CLIBS) $(TM_CLIBS) $(NAT_CLIBS) $(GDBTKLIBS) @LIBS@ \
 	-lintl -liberty
-CLIBS = $(SIM) $(READLINE) $(OPCODES) $(BFD) $(INTL) $(LIBIBERTY) \
+CLIBS = $(SIM) $(READLINE) $(OPCODES) $(BFD) $(INTL) $(LIBIBERTY) $(LIBDECNUMBER) \
 	$(XM_CLIBS) $(TM_CLIBS) $(NAT_CLIBS) $(GDBTKLIBS) @LIBS@ \
 	$(LIBICONV) $(LIBEXPAT) \
 	$(LIBIBERTY) $(WIN32LIBS)
@@ -624,6 +630,10 @@ safe_ctype_h =  $(INCLUDE_DIR)/safe-ctyp
 hashtab_h =	$(INCLUDE_DIR)/hashtab.h
 filenames_h =	$(INCLUDE_DIR)/filenames.h
 
+decimal128_h = $(LIBDECNUMBER_DIR)/dpd/decimal128.h
+decimal64_h = $(LIBDECNUMBER_DIR)/dpd/decimal64.h
+decimal32_h = $(LIBDECNUMBER_DIR)/dpd/decimal32.h
+
 #
 # $BUILD/ headers
 #
@@ -690,6 +700,7 @@ dictionary_h = dictionary.h
 disasm_h = disasm.h
 doublest_h = doublest.h $(floatformat_h)
 dummy_frame_h = dummy-frame.h
+dfp_h = dfp.h
 dwarf2expr_h = dwarf2expr.h
 dwarf2_frame_h = dwarf2-frame.h
 dwarf2loc_h = dwarf2loc.h
@@ -938,7 +949,7 @@ COMMON_OBS = $(DEPFILES) $(CONFIG_OBS) $
 	auxv.o \
 	bfd-target.o \
 	blockframe.o breakpoint.o findvar.o regcache.o \
-	charset.o disasm.o dummy-frame.o \
+	charset.o disasm.o dummy-frame.o dfp.o \
 	source.o value.o eval.o valops.o valarith.o valprint.o printcmd.o \
 	block.o symtab.o symfile.o symmisc.o linespec.o dictionary.o \
 	infcall.o \
@@ -1871,7 +1882,7 @@ buildsym.o: buildsym.c $(defs_h) $(bfd_h
 	$(cp_support_h) $(dictionary_h) $(buildsym_h) $(stabsread_h)
 c-exp.o: c-exp.c $(defs_h) $(gdb_string_h) $(expression_h) $(value_h) \
 	$(parser_defs_h) $(language_h) $(c_lang_h) $(bfd_h) $(symfile_h) \
-	$(objfiles_h) $(charset_h) $(block_h) $(cp_support_h)
+	$(objfiles_h) $(charset_h) $(block_h) $(cp_support_h) $(dfp_h)
 charset.o: charset.c $(defs_h) $(charset_h) $(gdbcmd_h) $(gdb_assert_h) \
 	$(gdb_string_h)
 c-lang.o: c-lang.c $(defs_h) $(symtab_h) $(gdbtypes_h) $(expression_h) \
@@ -1954,6 +1965,7 @@ dsrec.o: dsrec.c $(defs_h) $(serial_h) $
 dummy-frame.o: dummy-frame.c $(defs_h) $(dummy_frame_h) $(regcache_h) \
 	$(frame_h) $(inferior_h) $(gdb_assert_h) $(frame_unwind_h) \
 	$(command_h) $(gdbcmd_h) $(gdb_string_h)
+dfp.o: dfp.c $(defs_h) $(dfp_h) $(decimal128_h) $(decimal64_h) $(decimal32_h)
 dwarf2expr.o: dwarf2expr.c $(defs_h) $(symtab_h) $(gdbtypes_h) $(value_h) \
 	$(gdbcore_h) $(elf_dwarf2_h) $(dwarf2expr_h)
 dwarf2-frame.o: dwarf2-frame.c $(defs_h) $(dwarf2expr_h) $(elf_dwarf2_h) \
@@ -2815,11 +2827,11 @@ valops.o: valops.c $(defs_h) $(symtab_h)
 valprint.o: valprint.c $(defs_h) $(gdb_string_h) $(symtab_h) $(gdbtypes_h) \
 	$(value_h) $(gdbcore_h) $(gdbcmd_h) $(target_h) $(language_h) \
 	$(annotate_h) $(valprint_h) $(floatformat_h) $(doublest_h) \
-	$(exceptions_h)
+	$(exceptions_h) $(dfp_h)
 value.o: value.c $(defs_h) $(gdb_string_h) $(symtab_h) $(gdbtypes_h) \
 	$(value_h) $(gdbcore_h) $(command_h) $(gdbcmd_h) $(target_h) \
 	$(language_h) $(demangle_h) $(doublest_h) \
-	$(gdb_assert_h) $(regcache_h) $(block_h)
+	$(gdb_assert_h) $(regcache_h) $(block_h) $(dfp_h)
 varobj.o: varobj.c $(defs_h) $(exceptions_h) $(value_h) $(expression_h) \
 	$(frame_h) $(language_h) $(wrapper_h) $(gdbcmd_h) $(block_h) \
 	$(gdb_assert_h) $(gdb_string_h) $(varobj_h) $(vec_h)
Index: gdb/dfp.c
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ gdb/dfp.c	2007-10-15 14:35:06.000000000 -0200
@@ -0,0 +1,113 @@
+/* Decimal floating point support for GDB.
+
+   Copyright 2007 Free Software Foundation, Inc.
+
+   This file is part of GDB.
+
+   This program 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 3 of the License, or
+   (at your option) any later version.
+
+   This program 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 this program.  If not, see <http://www.gnu.org/licenses/>.  */
+
+#include "defs.h"
+
+/* The order of the following headers is important for making sure
+   decNumber structure is large enough to hold decimal128 digits.  */
+
+#include "dpd/decimal128.h"
+#include "dpd/decimal64.h"
+#include "dpd/decimal32.h"
+
+/* In GDB, we are using an array of gdb_byte to represent decimal values.
+   They are stored in host byte order.  This routine does the conversion if
+   the target byte order is different.  */
+static void
+match_endianness (const gdb_byte *from, int len, gdb_byte *to)
+{
+  int i;
+
+#if WORDS_BIGENDIAN
+#define OPPOSITE_BYTE_ORDER BFD_ENDIAN_LITTLE
+#else
+#define OPPOSITE_BYTE_ORDER BFD_ENDIAN_BIG
+#endif
+
+  if (gdbarch_byte_order (current_gdbarch) == OPPOSITE_BYTE_ORDER)
+    for (i = 0; i < len; i++)
+      to[i] = from[len - i - 1];
+  else
+    for (i = 0; i < len; i++)
+      to[i] = from[i];
+
+  return;
+}
+
+/* Convert decimal type to its string representation.  LEN is the length
+   of the decimal type, 4 bytes for decimal32, 8 bytes for decimal64 and
+   16 bytes for decimal128.  */
+void
+decimal_to_string (const gdb_byte *decbytes, int len, char *s)
+{
+  gdb_byte dec[16];
+
+  match_endianness (decbytes, len, dec);
+  switch (len)
+    {
+      case 4:
+        decimal32ToString ((decimal32 *) dec, s);
+        break;
+      case 8:
+        decimal64ToString ((decimal64 *) dec, s);
+        break;
+      case 16:
+        decimal128ToString ((decimal128 *) dec, s);
+        break;
+      default:
+        error (_("Unknown decimal floating point type.\n"));
+        break;
+    }
+}
+
+/* Convert the string form of a decimal value to its decimal representation.
+   LEN is the length of the decimal type, 4 bytes for decimal32, 8 bytes for
+   decimal64 and 16 bytes for decimal128.  */
+int
+decimal_from_string (gdb_byte *decbytes, int len, const char *string)
+{
+  decContext set;
+  gdb_byte dec[16];
+
+  switch (len)
+    {
+      case 4:
+        decContextDefault (&set, DEC_INIT_DECIMAL32);
+        set.traps = 0;
+        decimal32FromString ((decimal32 *) dec, string, &set);
+        break;
+      case 8:
+        decContextDefault (&set, DEC_INIT_DECIMAL64);
+        set.traps = 0;
+        decimal64FromString ((decimal64 *) dec, string, &set);
+        break;
+      case 16:
+        decContextDefault (&set, DEC_INIT_DECIMAL128);
+        set.traps = 0;
+        decimal128FromString ((decimal128 *) dec, string, &set);
+        break;
+      default:
+        error (_("Unknown decimal floating point type.\n"));
+        break;
+    }
+
+  match_endianness (dec, len, decbytes);
+
+  return 1;
+}
Index: gdb/dfp.h
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ gdb/dfp.h	2007-10-15 14:35:06.000000000 -0200
@@ -0,0 +1,35 @@
+/* Decimal floating point support for GDB.
+
+   Copyright 2007 Free Software Foundation, Inc.
+
+   This file is part of GDB.
+
+   This program 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 3 of the License, or
+   (at your option) any later version.
+
+   This program 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 this program.  If not, see <http://www.gnu.org/licenses/>.  */
+
+/* Decimal floating point is one of the extension to IEEE 754, which is
+   described in http://grouper.ieee.org/groups/754/revision.html and
+   http://www2.hursley.ibm.com/decimal/.  It completes binary floating
+   point by representing floating point more exactly.  */
+
+#ifndef DFP_H
+#define DFP_H
+
+/* When using decimal128, this is the maximum string length + 1
+ * (value comes from libdecnumber's DECIMAL128_String constant).  */
+#define MAX_DECIMAL_STRING  43
+
+extern void decimal_to_string (const gdb_byte *, int, char *);
+extern int decimal_from_string (gdb_byte *, int, const char *);
+
+#endif
Index: gdb/configure.ac
===================================================================
--- gdb/configure.ac.orig	2007-10-01 14:20:02.000000000 -0300
+++ gdb/configure.ac	2007-10-15 14:35:06.000000000 -0200
@@ -496,6 +496,7 @@ AC_CHECK_TYPES(uintptr_t, [], [], [#incl
 
 AC_C_CONST
 AC_C_INLINE
+AC_C_BIGENDIAN
 
 # ------------------------------ #
 # Checks for library functions.  #

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]