This is the mail archive of the
gdb-patches@sources.redhat.com
mailing list for the GDB project.
Re: [patch/rfc] Remove last IEEE_FLOAT use
- From: Andrew Cagney <ac131313 at cygnus dot com>
- To: gdb-patches at sources dot redhat dot com
- Date: Sat, 19 Jan 2002 16:09:06 -0500
- Subject: Re: [patch/rfc] Remove last IEEE_FLOAT use
- References: <3C451C6C.5020508@cygnus.com>
> Hello,
>
> The attached removes the last remaning use of IEEE_FLOAT. I think it is ok. The relevant block is:
>
> - /* FIXME: kettenis/2001-01-20: The check for IEEE_FLOAT is probably
> - still necessary since GDB by default assumes that the target uses
> - the IEEE 754 representation for its floats and doubles. Of
> - course this is all crock and should be cleaned up. */
> -
> - if (len == TARGET_FLOAT_BIT / TARGET_CHAR_BIT && IEEE_FLOAT)
> - fmt = TARGET_FLOAT_FORMAT;
> - else if (len == TARGET_DOUBLE_BIT / TARGET_CHAR_BIT && IEEE_FLOAT)
> - fmt = TARGET_DOUBLE_FORMAT;
> - else if (len == TARGET_LONG_DOUBLE_BIT / TARGET_CHAR_BIT)
> - fmt = TARGET_LONG_DOUBLE_FORMAT;
> -
>
> Thinking hard about it. I can't see how IEEE_FLOAT can make a real difference. In all likelyhood both the host and target have IEEE float. If one of them doesn't, the floatformat stuff should be kicking in and correctly handle any needed conversion.
>
> Anyway, the above code block is replaced by a call to a new function floatformat_from_type().
>
> Any comments?
>
> Andrew
Oops, I forgot to attach the actual patch. This should help.
Andrew
Index: ChangeLog
2002-01-19 Andrew Cagney <ac131313@redhat.com>
* doublest.h (floatformat_from_type): Declare.
* doublest.c (floatformat_from_type): New function.
(convert_typed_floating): Use.
* valprint.c (print_floating): Replace checks for IEEE_FLOAT with
call to function floatformat_from_type.
* gdbarch.sh (IEEE_FLOAT): Delete.
* gdbarch.h, gdbarch.c: Re-generate.
* config/i960/tm-i960.h (IEEE_FLOAT): Delete macro.
* config/i386/tm-i386.h (IEEE_FLOAT): Ditto.
* config/z8k/tm-z8k.h (IEEE_FLOAT): Ditto.
* config/sparc/tm-sparc.h (IEEE_FLOAT): Ditto.
* config/pa/tm-hppa.h (IEEE_FLOAT): Ditto.
* config/m88k/tm-m88k.h (IEEE_FLOAT): Ditto.
* config/m68k/tm-m68k.h (IEEE_FLOAT): Ditto.
* config/h8500/tm-h8500.h (IEEE_FLOAT): Ditto.
* config/h8300/tm-h8300.h (IEEE_FLOAT): Ditto.
* config/fr30/tm-fr30.h (IEEE_FLOAT): Ditto.
* config/arm/tm-arm.h (IEEE_FLOAT): Ditto.
* config/arc/tm-arc.h (IEEE_FLOAT): Ditto.
* config/alpha/tm-alpha.h (IEEE_FLOAT): Ditto.
* s390-tdep.c (s390_gdbarch_init): Do not set ieee_float.
* x86-64-tdep.c (i386_gdbarch_init): Ditto.
* sparc-tdep.c (sparc_gdbarch_init): Ditto.
* sh-tdep.c (sh_gdbarch_init): Ditto.
* mips-tdep.c (mips_gdbarch_init): Ditto.
* m68hc11-tdep.c (m68hc11_gdbarch_init): Ditto.
* cris-tdep.c (cris_gdbarch_init): Ditto.
Index: doc/ChangeLog
2002-01-19 Andrew Cagney <ac131313@redhat.com>
* gdbint.texinfo (Target Architecture Definition): Remove
definition of IEEE_FLOAT.
Index: cris-tdep.c
===================================================================
RCS file: /cvs/src/src/gdb/cris-tdep.c,v
retrieving revision 1.10
diff -u -r1.10 cris-tdep.c
--- cris-tdep.c 2002/01/05 04:30:16 1.10
+++ cris-tdep.c 2002/01/19 20:58:16
@@ -4154,9 +4154,6 @@
which means we have to set this explicitly. */
set_gdbarch_long_double_bit (gdbarch, 64);
- /* Floating point is IEEE compatible. */
- set_gdbarch_ieee_float (gdbarch, 1);
-
/* There are 32 registers (some of which may not be implemented). */
set_gdbarch_num_regs (gdbarch, 32);
set_gdbarch_sp_regnum (gdbarch, 14);
Index: doublest.c
===================================================================
RCS file: /cvs/src/src/gdb/doublest.c,v
retrieving revision 1.6
diff -u -r1.6 doublest.c
--- doublest.c 2001/10/29 23:35:37 1.6
+++ doublest.c 2002/01/19 20:58:18
@@ -619,7 +619,7 @@
target-dependent code, the format of floating-point types is known,
but not passed on by GDB. This should be fixed. */
-static const struct floatformat *
+const struct floatformat *
floatformat_from_length (int len)
{
if (len * TARGET_CHAR_BIT == TARGET_FLOAT_BIT)
@@ -632,6 +632,16 @@
return NULL;
}
+const struct floatformat *
+floatformat_from_type (const struct type *type)
+{
+ gdb_assert (TYPE_CODE (type) == TYPE_CODE_FLT);
+ if (TYPE_FLOATFORMAT (type) != NULL)
+ return TYPE_FLOATFORMAT (type);
+ else
+ return floatformat_from_length (TYPE_LENGTH (type));
+}
+
/* If the host doesn't define NAN, use zero instead. */
#ifndef NAN
#define NAN 0.0
@@ -732,18 +742,11 @@
convert_typed_floating (const void *from, const struct type *from_type,
void *to, const struct type *to_type)
{
- const struct floatformat *from_fmt = TYPE_FLOATFORMAT (from_type);
- const struct floatformat *to_fmt = TYPE_FLOATFORMAT (to_type);
+ const struct floatformat *from_fmt = floatformat_from_type (from_type);
+ const struct floatformat *to_fmt = floatformat_from_type (to_type);
gdb_assert (TYPE_CODE (from_type) == TYPE_CODE_FLT);
gdb_assert (TYPE_CODE (to_type) == TYPE_CODE_FLT);
-
- /* If the floating-point format of FROM_TYPE or TO_TYPE isn't known,
- try to guess it from the type's length. */
- if (from_fmt == NULL)
- from_fmt = floatformat_from_length (TYPE_LENGTH (from_type));
- if (to_fmt == NULL)
- to_fmt = floatformat_from_length (TYPE_LENGTH (to_type));
if (from_fmt == NULL || to_fmt == NULL)
{
Index: doublest.h
===================================================================
RCS file: /cvs/src/src/gdb/doublest.h,v
retrieving revision 1.7
diff -u -r1.7 doublest.h
--- doublest.h 2001/10/29 23:35:37 1.7
+++ doublest.h 2002/01/19 20:58:19
@@ -68,6 +68,12 @@
extern DOUBLEST extract_floating (const void *addr, int len);
extern void store_floating (void *addr, int len, DOUBLEST val);
+/* Given TYPE, return its floatformat. TYPE_FLOATFORMAT() may return
+ NULL. type_floatformat() detects that and returns a floatformat
+ based on the type size when FLOATFORMAT is NULL. */
+
+const struct floatformat *floatformat_from_type (const struct type *type);
+
extern DOUBLEST extract_typed_floating (const void *addr,
const struct type *type);
extern void store_typed_floating (void *addr, const struct type *type,
Index: gdbarch.sh
===================================================================
RCS file: /cvs/src/src/gdb/gdbarch.sh,v
retrieving revision 1.103
diff -u -r1.103 gdbarch.sh
--- gdbarch.sh 2002/01/05 04:30:17 1.103
+++ gdbarch.sh 2002/01/19 20:58:40
@@ -403,8 +403,6 @@
# One if \`char' acts like \`signed char', zero if \`unsigned char'.
v::TARGET_CHAR_SIGNED:int:char_signed::::1:-1:1::::
#
-v::IEEE_FLOAT:int:ieee_float::::0:0::0:::
-#
f::TARGET_READ_PC:CORE_ADDR:read_pc:ptid_t ptid:ptid::0:generic_target_read_pc::0
f::TARGET_WRITE_PC:void:write_pc:CORE_ADDR val, ptid_t ptid:val, ptid::0:generic_target_write_pc::0
f::TARGET_READ_FP:CORE_ADDR:read_fp:void:::0:generic_target_read_fp::0
Index: m68hc11-tdep.c
===================================================================
RCS file: /cvs/src/src/gdb/m68hc11-tdep.c,v
retrieving revision 1.15
diff -u -r1.15 m68hc11-tdep.c
--- m68hc11-tdep.c 2001/07/22 12:37:02 1.15
+++ m68hc11-tdep.c 2002/01/19 20:58:45
@@ -1157,7 +1157,6 @@
set_gdbarch_stack_align (gdbarch, m68hc11_stack_align);
set_gdbarch_believe_pcc_promotion (gdbarch, 1);
- set_gdbarch_ieee_float (gdbarch, 1);
return gdbarch;
}
Index: mips-tdep.c
===================================================================
RCS file: /cvs/src/src/gdb/mips-tdep.c,v
retrieving revision 1.63
diff -u -r1.63 mips-tdep.c
--- mips-tdep.c 2002/01/05 04:30:17 1.63
+++ mips-tdep.c 2002/01/19 20:59:16
@@ -4283,7 +4283,6 @@
set_gdbarch_inner_than (gdbarch, core_addr_lessthan);
set_gdbarch_breakpoint_from_pc (gdbarch, mips_breakpoint_from_pc);
set_gdbarch_decr_pc_after_break (gdbarch, 0);
- set_gdbarch_ieee_float (gdbarch, 1);
set_gdbarch_skip_prologue (gdbarch, mips_skip_prologue);
set_gdbarch_saved_pc_after_call (gdbarch, mips_saved_pc_after_call);
Index: s390-tdep.c
===================================================================
RCS file: /cvs/src/src/gdb/s390-tdep.c,v
retrieving revision 1.39
diff -u -r1.39 s390-tdep.c
--- s390-tdep.c 2001/12/20 08:55:31 1.39
+++ s390-tdep.c 2002/01/19 20:59:21
@@ -1785,7 +1785,6 @@
but not always. */
set_gdbarch_decr_pc_after_break (gdbarch, 2);
set_gdbarch_pop_frame (gdbarch, s390_pop_frame);
- set_gdbarch_ieee_float (gdbarch, 1);
/* Stack grows downward. */
set_gdbarch_inner_than (gdbarch, core_addr_lessthan);
/* Offset from address of function to start of its code.
Index: sh-tdep.c
===================================================================
RCS file: /cvs/src/src/gdb/sh-tdep.c,v
retrieving revision 1.44
diff -u -r1.44 sh-tdep.c
--- sh-tdep.c 2002/01/12 16:38:47 1.44
+++ sh-tdep.c 2002/01/19 20:59:25
@@ -2123,7 +2123,6 @@
set_gdbarch_saved_pc_after_call (gdbarch, sh_saved_pc_after_call);
set_gdbarch_frame_num_args (gdbarch, frame_num_args_unknown);
set_gdbarch_believe_pcc_promotion (gdbarch, 1);
- set_gdbarch_ieee_float (gdbarch, 1);
return gdbarch;
}
Index: sparc-tdep.c
===================================================================
RCS file: /cvs/src/src/gdb/sparc-tdep.c,v
retrieving revision 1.20
diff -u -r1.20 sparc-tdep.c
--- sparc-tdep.c 2001/12/30 17:00:43 1.20
+++ sparc-tdep.c 2002/01/19 20:59:30
@@ -2963,7 +2963,6 @@
set_gdbarch_frameless_function_invocation (gdbarch,
frameless_look_for_prologue);
set_gdbarch_get_saved_register (gdbarch, sparc_get_saved_register);
- set_gdbarch_ieee_float (gdbarch, 1);
set_gdbarch_init_extra_frame_info (gdbarch, sparc_init_extra_frame_info);
set_gdbarch_inner_than (gdbarch, core_addr_lessthan);
set_gdbarch_int_bit (gdbarch, 4 * TARGET_CHAR_BIT);
Index: valprint.c
===================================================================
RCS file: /cvs/src/src/gdb/valprint.c,v
retrieving revision 1.20
diff -u -r1.20 valprint.c
--- valprint.c 2002/01/05 04:30:20 1.20
+++ valprint.c 2002/01/19 20:59:35
@@ -548,21 +548,9 @@
{
DOUBLEST doub;
int inv;
- const struct floatformat *fmt = &floatformat_unknown;
+ const struct floatformat *fmt = floatformat_from_type (type);
unsigned len = TYPE_LENGTH (type);
- /* FIXME: kettenis/2001-01-20: The check for IEEE_FLOAT is probably
- still necessary since GDB by default assumes that the target uses
- the IEEE 754 representation for its floats and doubles. Of
- course this is all crock and should be cleaned up. */
-
- if (len == TARGET_FLOAT_BIT / TARGET_CHAR_BIT && IEEE_FLOAT)
- fmt = TARGET_FLOAT_FORMAT;
- else if (len == TARGET_DOUBLE_BIT / TARGET_CHAR_BIT && IEEE_FLOAT)
- fmt = TARGET_DOUBLE_FORMAT;
- else if (len == TARGET_LONG_DOUBLE_BIT / TARGET_CHAR_BIT)
- fmt = TARGET_LONG_DOUBLE_FORMAT;
-
if (floatformat_is_nan (fmt, valaddr))
{
if (floatformat_is_negative (fmt, valaddr))
@@ -575,6 +563,13 @@
return;
}
+ /* FIXME: cagney/2002-01-15: The simpler extract_typed_floating()
+ routine could be used here only that routine has no way of
+ indicating that the floating point it extracted was invalid (As
+ indicated by INVALID_FLOAT). Instead, this code here could call
+ something like floating_invalid() to check for an invalid
+ floating point. */
+
doub = unpack_double (type, valaddr, &inv);
if (inv)
{
@@ -584,6 +579,10 @@
/* FIXME: kettenis/2001-01-20: The following code makes too much
assumptions about the host and target floating point format. */
+
+ /* FIXME: cagney/2002-01-15: The floatformat pointed to by FMT
+ should contain all the information needed to print the
+ floating-point value without host dependencies. */
if (len < sizeof (double))
fprintf_filtered (stream, "%.9g", (double) doub);
Index: x86-64-tdep.c
===================================================================
RCS file: /cvs/src/src/gdb/x86-64-tdep.c,v
retrieving revision 1.6
diff -u -r1.6 x86-64-tdep.c
--- x86-64-tdep.c 2001/12/27 15:22:27 1.6
+++ x86-64-tdep.c 2002/01/19 20:59:36
@@ -834,8 +834,6 @@
set_gdbarch_ptr_bit (gdbarch, 64);
set_gdbarch_long_double_format (gdbarch, &floatformat_i387_ext);
- set_gdbarch_ieee_float (gdbarch, 1);
-
set_gdbarch_num_regs (gdbarch, X86_64_NUM_REGS);
set_gdbarch_register_name (gdbarch, x86_64_register_name);
Index: config/alpha/tm-alpha.h
===================================================================
RCS file: /cvs/src/src/gdb/config/alpha/tm-alpha.h,v
retrieving revision 1.12
diff -u -r1.12 tm-alpha.h
--- tm-alpha.h 2002/01/19 20:22:28 1.12
+++ tm-alpha.h 2002/01/19 20:59:43
@@ -43,9 +43,6 @@
#define TARGET_LONG_LONG_BIT 64
#define TARGET_PTR_BIT 64
-/* Floating point is IEEE compliant */
-#define IEEE_FLOAT (1)
-
/* Number of traps that happen between exec'ing the shell
* to run an inferior, and when we finally get to
* the inferior code. This is 2 on most implementations.
Index: config/arc/tm-arc.h
===================================================================
RCS file: /cvs/src/src/gdb/config/arc/tm-arc.h,v
retrieving revision 1.7
diff -u -r1.7 tm-arc.h
--- tm-arc.h 2001/03/21 18:31:48 1.7
+++ tm-arc.h 2002/01/19 20:59:43
@@ -27,9 +27,6 @@
/* Byte order is selectable. */
#define TARGET_BYTE_ORDER_SELECTABLE
-/* We have IEEE floating point, if we have any float at all. */
-#define IEEE_FLOAT (1)
-
/* Offset from address of function to start of its code.
Zero on most machines. */
#define FUNCTION_START_OFFSET 0
Index: config/arm/tm-arm.h
===================================================================
RCS file: /cvs/src/src/gdb/config/arm/tm-arm.h,v
retrieving revision 1.13
diff -u -r1.13 tm-arm.h
--- tm-arm.h 2002/01/09 18:39:37 1.13
+++ tm-arm.h 2002/01/19 20:59:47
@@ -35,7 +35,6 @@
#define TARGET_BYTE_ORDER_DEFAULT BFD_ENDIAN_LITTLE
/* IEEE format floating point. */
-#define IEEE_FLOAT (1)
#define TARGET_DOUBLE_FORMAT (target_byte_order == BFD_ENDIAN_BIG \
? &floatformat_ieee_double_big \
: &floatformat_ieee_double_littlebyte_bigword)
Index: config/fr30/tm-fr30.h
===================================================================
RCS file: /cvs/src/src/gdb/config/fr30/tm-fr30.h,v
retrieving revision 1.7
diff -u -r1.7 tm-fr30.h
--- tm-fr30.h 2002/01/05 04:30:28 1.7
+++ tm-fr30.h 2002/01/19 20:59:47
@@ -187,9 +187,6 @@
/* Define this for Wingdb */
#define TARGET_FR30
-/* IEEE format floating point */
-#define IEEE_FLOAT (1)
-
/* Define other aspects of the stack frame. */
/* An expression that tells us whether the function invocation represented
Index: config/h8300/tm-h8300.h
===================================================================
RCS file: /cvs/src/src/gdb/config/h8300/tm-h8300.h,v
retrieving revision 1.8
diff -u -r1.8 tm-h8300.h
--- tm-h8300.h 2002/01/05 04:30:29 1.8
+++ tm-h8300.h 2002/01/19 20:59:48
@@ -58,7 +58,6 @@
extern void h8300_init_extra_frame_info ();
-#define IEEE_FLOAT (1)
/* Define the bit, byte, and word ordering of the machine. */
#define TARGET_BYTE_ORDER BFD_ENDIAN_BIG
#undef TARGET_INT_BIT
Index: config/h8500/tm-h8500.h
===================================================================
RCS file: /cvs/src/src/gdb/config/h8500/tm-h8500.h,v
retrieving revision 1.8
diff -u -r1.8 tm-h8500.h
--- tm-h8500.h 2002/01/05 04:30:30 1.8
+++ tm-h8500.h 2002/01/19 20:59:48
@@ -25,8 +25,6 @@
#define GDB_TARGET_IS_H8500
-#define IEEE_FLOAT (1)
-
/* Define the bit, byte, and word ordering of the machine. */
#define TARGET_BYTE_ORDER BFD_ENDIAN_BIG
Index: config/i386/tm-i386.h
===================================================================
RCS file: /cvs/src/src/gdb/config/i386/tm-i386.h,v
retrieving revision 1.27
diff -u -r1.27 tm-i386.h
--- tm-i386.h 2001/12/27 15:22:27 1.27
+++ tm-i386.h 2002/01/19 20:59:48
@@ -47,11 +47,6 @@
#define TARGET_LONG_DOUBLE_BIT 96
-/* Used for example in valprint.c:print_floating() to enable checking
- for NaN's */
-
-#define IEEE_FLOAT (1)
-
/* Number of traps that happen between exec'ing the shell to run an
inferior, and when we finally get to the inferior code. This is 2
on most implementations. */
Index: config/i960/tm-i960.h
===================================================================
RCS file: /cvs/src/src/gdb/config/i960/tm-i960.h,v
retrieving revision 1.6
diff -u -r1.6 tm-i960.h
--- tm-i960.h 2001/12/15 16:53:23 1.6
+++ tm-i960.h 2002/01/19 20:59:48
@@ -41,10 +41,6 @@
/* Byte order is configurable, but this machine runs little-endian. */
#define TARGET_BYTE_ORDER BFD_ENDIAN_LITTLE
-/* We have IEEE floating point, if we have any float at all. */
-
-#define IEEE_FLOAT (1)
-
/* Offset from address of function to start of its code.
Zero on most machines. */
Index: config/m68k/tm-m68k.h
===================================================================
RCS file: /cvs/src/src/gdb/config/m68k/tm-m68k.h,v
retrieving revision 1.8
diff -u -r1.8 tm-m68k.h
--- tm-m68k.h 2002/01/05 04:30:34 1.8
+++ tm-m68k.h 2002/01/19 20:59:49
@@ -23,8 +23,6 @@
/* Generic 68000 stuff, to be included by other tm-*.h files. */
-#define IEEE_FLOAT (1)
-
/* Define the bit, byte, and word ordering of the machine. */
#define TARGET_BYTE_ORDER BFD_ENDIAN_BIG
Index: config/m88k/tm-m88k.h
===================================================================
RCS file: /cvs/src/src/gdb/config/m88k/tm-m88k.h,v
retrieving revision 1.10
diff -u -r1.10 tm-m88k.h
--- tm-m88k.h 2002/01/05 04:30:36 1.10
+++ tm-m88k.h 2002/01/19 20:59:49
@@ -47,8 +47,6 @@
init_extra_frame_info (fromleaf, fi)
extern void init_extra_frame_info ();
-#define IEEE_FLOAT (1)
-
/* Offset from address of function to start of its code.
Zero on most machines. */
Index: config/pa/tm-hppa.h
===================================================================
RCS file: /cvs/src/src/gdb/config/pa/tm-hppa.h,v
retrieving revision 1.12
diff -u -r1.12 tm-hppa.h
--- tm-hppa.h 2002/01/15 02:06:47 1.12
+++ tm-hppa.h 2002/01/19 20:59:50
@@ -54,10 +54,6 @@
((X) >> (31 - (TO)) & ((1 << ((TO) - (FROM) + 1)) - 1))
#endif
-/* Watch out for NaNs */
-
-#define IEEE_FLOAT (1)
-
/* On the PA, any pass-by-value structure > 8 bytes is actually
passed via a pointer regardless of its type or the compiler
used. */
Index: config/sparc/tm-sparc.h
===================================================================
RCS file: /cvs/src/src/gdb/config/sparc/tm-sparc.h,v
retrieving revision 1.10
diff -u -r1.10 tm-sparc.h
--- tm-sparc.h 2002/01/05 04:30:39 1.10
+++ tm-sparc.h 2002/01/19 20:59:53
@@ -298,9 +298,6 @@
#define STACK_ALIGN(ADDR) sparc32_stack_align (ADDR)
extern CORE_ADDR sparc32_stack_align (CORE_ADDR addr);
-/* Floating point is IEEE compatible. */
-#define IEEE_FLOAT (1)
-
/* The Sparc returns long doubles on the stack. */
#define RETURN_VALUE_ON_STACK(TYPE) \
Index: config/z8k/tm-z8k.h
===================================================================
RCS file: /cvs/src/src/gdb/config/z8k/tm-z8k.h,v
retrieving revision 1.6
diff -u -r1.6 tm-z8k.h
--- tm-z8k.h 2002/01/05 04:30:40 1.6
+++ tm-z8k.h 2002/01/19 20:59:54
@@ -19,8 +19,6 @@
Foundation, Inc., 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA. */
-#define IEEE_FLOAT (1)
-
#undef TARGET_INT_BIT
#undef TARGET_LONG_BIT
#undef TARGET_SHORT_BIT
Index: doc/gdbint.texinfo
===================================================================
RCS file: /cvs/src/src/gdb/doc/gdbint.texinfo,v
retrieving revision 1.56
diff -u -r1.56 gdbint.texinfo
--- gdbint.texinfo 2002/01/18 04:51:10 1.56
+++ gdbint.texinfo 2002/01/19 21:00:13
@@ -3075,10 +3075,6 @@
@samp{$} or @samp{$$}. For example, @code{$$dyncall} is a millicode
routine that handles inter-space procedure calls on PA-RISC.
-@item IEEE_FLOAT
-@findex IEEE_FLOAT
-Define this if the target system uses IEEE-format floating point numbers.
-
@item INIT_EXTRA_FRAME_INFO (@var{fromleaf}, @var{frame})
@findex INIT_EXTRA_FRAME_INFO
If additional information about the frame is required this should be