This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [patch 1/4] libdecnumber support
- From: Thiago Jung Bauermann <bauerman at br dot ibm dot com>
- To: Daniel Jacobowitz <drow at false dot org>
- Cc: gdb-patches at sourceware dot org
- Date: Mon, 15 Oct 2007 16:10:57 -0200
- Subject: Re: [patch 1/4] libdecnumber support
- References: <20070920215410.062714003@br.ibm.com> <20070920215539.570827491@br.ibm.com> <20071011155418.GA22982@caradoc.them.org>
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. #