Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Markus Deuling <deuling@de.ibm.com>
To: Ulrich Weigand <uweigand@de.ibm.com>
Cc: GDB Patches <gdb-patches@sourceware.org>,
	Eli Zaretskii <eliz@gnu.org>,
	        Daniel Jacobowitz <drow@false.org>
Subject: Re: [patch] Remove BITS_BIG_ENDIAN from defs.h
Date: Tue, 15 Jan 2008 09:06:00 -0000	[thread overview]
Message-ID: <478C76F5.2000402@de.ibm.com> (raw)
In-Reply-To: <200801141811.m0EIBEFw007402@d12av02.megacenter.de.ibm.com>

[-- Attachment #1: Type: text/plain, Size: 2075 bytes --]

Ulrich Weigand schrieb:
> Markus Deuling wrote:
> 
>> this patch removes BITS_BIG_ENDIAN from defs.h by replacing it with
>> its expression. The way to recognize endianess of a target is
>> gdbarch_byte_order (current_gdbarch) == BFD_ENDIAN_BIG which is
>> widely used by most files.  So this macro is unnecessary.
> 
> As Dan and Eli pointed out, there *is* a difference between the two:
> - gdbarch_byte_order says whether in a multi-byte value, the least
>   or most significant *byte* comes first
> - BITS_BIG_ENDIAN is intended to say whether in a bitfield, the
>   least of most significant *bit* comes first
> 
> Looking at GCC, there are targets with big endian byte order but
> little endian bit order (or vice versa), even though those are 
> not supported in GDB today (or maybe they are but bitfields simply
> do not work correctly).
> 
> Thus I think we should introduce a new gdbarch property to allow
> to set the bit order.
> 


Hi,

thank you very much for all of your valuable feedback. I rewrote the patch
and introduce a new gdbarch property bits_big_endian now to replace BITS_BIG_ENDIAN.

I tested this patch on both Little and Big Endian machine (x86/ppc) without regression.
If this patch is ok I'll post another one for the documentation. Ok to commit?

ChangeLog:

	* gdbarch.sh (function_list): Add new property bits_big_endian to
	gdbarch structure.
	* gdbarch.{c,h}: Regenerate.

	* value.c (struct value): Replace BITS_BIG_ENDIAN by
	gdbarch_bits_big_endian (comment).
	(unpack_field_as_long, modify_field): Likewise.
	* value.h: Likewise (comment).
	* valops.c (value_slice): Likewise.
	* valarith.c (value_subscript, value_bit_index): Likewise.
	* gdbtypes.h (field): Likewise (comment).
	* eval.c (evaluate_subexp_standard): Likewise.
	* dwarf2read.c (dwarf2_add_field): Likewise.
	* ada-lang.c (decode_packed_array, ada_value_primitive_packed_val)
	(move_bits, ada_value_assign, value_assign_to_component): Likewise.
	
	* defs.h (BITS_BIG_ENDIAN): Remove.




-- 
  Markus Deuling
  GNU Toolchain for Linux on Cell BE
  deuling@de.ibm.com



[-- Attachment #2: diff-BITS_BIG_ENDIAN --]
[-- Type: text/plain, Size: 12337 bytes --]

diff -urpN src/gdb/ada-lang.c dev/gdb/ada-lang.c
--- src/gdb/ada-lang.c	2008-01-09 11:01:28.000000000 +0100
+++ dev/gdb/ada-lang.c	2008-01-15 08:02:37.000000000 +0100
@@ -1875,7 +1875,8 @@ decode_packed_array (struct value *arr)
       return NULL;
     }
 
-  if (BITS_BIG_ENDIAN && ada_is_modular_type (value_type (arr)))
+  if (gdbarch_bits_big_endian (current_gdbarch)
+      && ada_is_modular_type (value_type (arr)))
     {
        /* This is a (right-justified) modular type representing a packed
  	 array with no wrapper.  In order to interpret the value through
@@ -1998,7 +1999,7 @@ ada_value_primitive_packed_val (struct v
   int len = (bit_size + bit_offset + HOST_CHAR_BIT - 1) / 8;
   /* Transmit bytes from least to most significant; delta is the direction
      the indices move.  */
-  int delta = BITS_BIG_ENDIAN ? -1 : 1;
+  int delta = gdbarch_bits_big_endian (current_gdbarch) ? -1 : 1;
 
   type = ada_check_typedef (type);
 
@@ -2047,7 +2048,7 @@ ada_value_primitive_packed_val (struct v
       memset (unpacked, 0, TYPE_LENGTH (type));
       return v;
     }
-  else if (BITS_BIG_ENDIAN)
+  else if (gdbarch_bits_big_endian (current_gdbarch))
     {
       src = len - 1;
       if (has_negatives (type)
@@ -2141,7 +2142,7 @@ move_bits (gdb_byte *target, int targ_of
   targ_offset %= HOST_CHAR_BIT;
   source += src_offset / HOST_CHAR_BIT;
   src_offset %= HOST_CHAR_BIT;
-  if (BITS_BIG_ENDIAN)
+  if (gdbarch_bits_big_endian (current_gdbarch))
     {
       accum = (unsigned char) *source;
       source += 1;
@@ -2229,7 +2230,7 @@ ada_value_assign (struct value *toval, s
         fromval = value_cast (type, fromval);
 
       read_memory (to_addr, buffer, len);
-      if (BITS_BIG_ENDIAN)
+      if (gdbarch_bits_big_endian (current_gdbarch))
         move_bits (buffer, value_bitpos (toval),
                    value_contents (fromval),
                    TYPE_LENGTH (value_type (fromval)) * TARGET_CHAR_BIT -
@@ -2276,7 +2277,7 @@ value_assign_to_component (struct value 
   else
     bits = value_bitsize (component);
 
-  if (BITS_BIG_ENDIAN)
+  if (gdbarch_bits_big_endian (current_gdbarch))
     move_bits (value_contents_writeable (container) + offset_in_container, 
 	       value_bitpos (container) + bit_offset_in_container,
 	       value_contents (val),
diff -urpN src/gdb/defs.h dev/gdb/defs.h
--- src/gdb/defs.h	2008-01-11 13:28:02.000000000 +0100
+++ dev/gdb/defs.h	2008-01-15 08:04:32.000000000 +0100
@@ -1043,14 +1043,6 @@ enum { MAX_REGISTER_SIZE = 16 };
 #define HOST_CHAR_BIT TARGET_CHAR_BIT
 #endif
 
-/* The bit byte-order has to do just with numbering of bits in
-   debugging symbols and such.  Conceptually, it's quite separate
-   from byte/word byte order.  */
-
-#if !defined (BITS_BIG_ENDIAN)
-#define BITS_BIG_ENDIAN (gdbarch_byte_order (current_gdbarch) == BFD_ENDIAN_BIG)
-#endif
-
 /* In findvar.c.  */
 
 extern LONGEST extract_signed_integer (const gdb_byte *, int);
diff -urpN src/gdb/dwarf2read.c dev/gdb/dwarf2read.c
--- src/gdb/dwarf2read.c	2008-01-11 14:32:31.000000000 +0100
+++ dev/gdb/dwarf2read.c	2008-01-15 07:55:28.000000000 +0100
@@ -3502,7 +3502,7 @@ dwarf2_add_field (struct field_info *fip
       attr = dwarf2_attr (die, DW_AT_bit_offset, cu);
       if (attr)
 	{
-	  if (BITS_BIG_ENDIAN)
+	  if (gdbarch_bits_big_endian (current_gdbarch))
 	    {
 	      /* For big endian bits, the DW_AT_bit_offset gives the
 	         additional bit offset from the MSB of the containing
diff -urpN src/gdb/eval.c dev/gdb/eval.c
--- src/gdb/eval.c	2008-01-08 11:22:24.000000000 +0100
+++ dev/gdb/eval.c	2008-01-15 07:53:28.000000000 +0100
@@ -686,7 +686,7 @@ evaluate_subexp_standard (struct type *e
 	      for (; range_low <= range_high; range_low++)
 		{
 		  int bit_index = (unsigned) range_low % TARGET_CHAR_BIT;
-		  if (BITS_BIG_ENDIAN)
+		  if (gdbarch_bits_big_endian (current_gdbarch))
 		    bit_index = TARGET_CHAR_BIT - 1 - bit_index;
 		  valaddr[(unsigned) range_low / TARGET_CHAR_BIT]
 		    |= 1 << bit_index;
diff -urpN src/gdb/gdbarch.c dev/gdb/gdbarch.c
--- src/gdb/gdbarch.c	2008-01-11 14:20:52.000000000 +0100
+++ dev/gdb/gdbarch.c	2008-01-15 08:04:34.000000000 +0100
@@ -130,6 +130,7 @@ struct gdbarch
 
      */
 
+  int bits_big_endian;
   int short_bit;
   int int_bit;
   int long_bit;
@@ -251,6 +252,7 @@ struct gdbarch startup_gdbarch =
   /*per-architecture data-pointers and swap regions */
   0, NULL, NULL,
   /* Multi-arch values */
+  0,  /* bits_big_endian */
   8 * sizeof (short),  /* short_bit */
   8 * sizeof (int),  /* int_bit */
   8 * sizeof (long),  /* long_bit */
@@ -382,6 +384,7 @@ gdbarch_alloc (const struct gdbarch_info
   gdbarch->target_desc = info->target_desc;
 
   /* Force the explicit initialization of these. */
+  gdbarch->bits_big_endian = (gdbarch->byte_order == BFD_ENDIAN_BIG);
   gdbarch->short_bit = 2*TARGET_CHAR_BIT;
   gdbarch->int_bit = 4*TARGET_CHAR_BIT;
   gdbarch->long_bit = 4*TARGET_CHAR_BIT;
@@ -480,6 +483,7 @@ verify_gdbarch (struct gdbarch *gdbarch)
   if (gdbarch->bfd_arch_info == NULL)
     fprintf_unfiltered (log, "\n\tbfd_arch_info");
   /* Check those that need to be defined for the given multi-arch level. */
+  /* Skip verify of bits_big_endian, invalid_p == 0 */
   /* Skip verify of short_bit, invalid_p == 0 */
   /* Skip verify of int_bit, invalid_p == 0 */
   /* Skip verify of long_bit, invalid_p == 0 */
@@ -648,6 +652,9 @@ gdbarch_dump (struct gdbarch *gdbarch, s
                       "gdbarch_dump: bfd_arch_info = %s\n",
                       gdbarch_bfd_arch_info (gdbarch)->printable_name);
   fprintf_unfiltered (file,
+                      "gdbarch_dump: bits_big_endian = %s\n",
+                      paddr_d (gdbarch->bits_big_endian));
+  fprintf_unfiltered (file,
                       "gdbarch_dump: breakpoint_from_pc = <0x%lx>\n",
                       (long) gdbarch->breakpoint_from_pc);
   fprintf_unfiltered (file,
@@ -1060,6 +1067,23 @@ gdbarch_target_desc (struct gdbarch *gdb
 }
 
 int
+gdbarch_bits_big_endian (struct gdbarch *gdbarch)
+{
+  gdb_assert (gdbarch != NULL);
+  /* Skip verify of bits_big_endian, invalid_p == 0 */
+  if (gdbarch_debug >= 2)
+    fprintf_unfiltered (gdb_stdlog, "gdbarch_bits_big_endian called\n");
+  return gdbarch->bits_big_endian;
+}
+
+void
+set_gdbarch_bits_big_endian (struct gdbarch *gdbarch,
+                             int bits_big_endian)
+{
+  gdbarch->bits_big_endian = bits_big_endian;
+}
+
+int
 gdbarch_short_bit (struct gdbarch *gdbarch)
 {
   gdb_assert (gdbarch != NULL);
diff -urpN src/gdb/gdbarch.h dev/gdb/gdbarch.h
--- src/gdb/gdbarch.h	2008-01-11 14:20:52.000000000 +0100
+++ dev/gdb/gdbarch.h	2008-01-15 08:04:28.000000000 +0100
@@ -71,6 +71,12 @@ extern const struct target_desc * gdbarc
 
 /* The following are initialized by the target dependent code. */
 
+/* The bit byte-order has to do just with numbering of bits in debugging symbols
+   and such.  Conceptually, it's quite separate from byte/word byte order. */
+
+extern int gdbarch_bits_big_endian (struct gdbarch *gdbarch);
+extern void set_gdbarch_bits_big_endian (struct gdbarch *gdbarch, int bits_big_endian);
+
 /* Number of bits in a char or unsigned char for the target machine.
    Just like CHAR_BIT in <limits.h> but describes the target machine.
    v:TARGET_CHAR_BIT:int:char_bit::::8 * sizeof (char):8::0:
diff -urpN src/gdb/gdbarch.sh dev/gdb/gdbarch.sh
--- src/gdb/gdbarch.sh	2008-01-11 14:20:52.000000000 +0100
+++ dev/gdb/gdbarch.sh	2008-01-15 08:04:22.000000000 +0100
@@ -343,6 +343,11 @@ i:int:byte_order:::BFD_ENDIAN_BIG
 i:enum gdb_osabi:osabi:::GDB_OSABI_UNKNOWN
 #
 i:const struct target_desc *:target_desc:::::::paddr_d ((long) gdbarch->target_desc)
+
+# The bit byte-order has to do just with numbering of bits in debugging symbols
+# and such.  Conceptually, it's quite separate from byte/word byte order.
+v:int:bits_big_endian:::0:(gdbarch->byte_order == BFD_ENDIAN_BIG)::0
+
 # Number of bits in a char or unsigned char for the target machine.
 # Just like CHAR_BIT in <limits.h> but describes the target machine.
 # v:TARGET_CHAR_BIT:int:char_bit::::8 * sizeof (char):8::0:
diff -urpN src/gdb/gdbtypes.h dev/gdb/gdbtypes.h
--- src/gdb/gdbtypes.h	2008-01-01 23:53:10.000000000 +0100
+++ dev/gdb/gdbtypes.h	2008-01-15 07:52:38.000000000 +0100
@@ -403,8 +403,8 @@ struct main_type
     {
       /* Position of this field, counting in bits from start of
 	 containing structure.
-	 For BITS_BIG_ENDIAN=1 targets, it is the bit offset to the MSB.
-	 For BITS_BIG_ENDIAN=0 targets, it is the bit offset to the LSB.
+	 For gdbarch_bits_big_endian=1 targets, it is the bit offset to the MSB.
+	 For gdbarch_bits_big_endian=0 targets, it is the bit offset to the LSB.
 	 For a range bound or enum value, this is the value itself. */
 
       int bitpos;
diff -urpN src/gdb/valarith.c dev/gdb/valarith.c
--- src/gdb/valarith.c	2008-01-08 11:22:24.000000000 +0100
+++ dev/gdb/valarith.c	2008-01-15 07:51:31.000000000 +0100
@@ -230,7 +230,8 @@ value_subscript (struct value *array, st
       offset = index / TARGET_CHAR_BIT;
       byte = *((char *) value_contents (array) + offset);
       bit_index = index % TARGET_CHAR_BIT;
-      byte >>= (BITS_BIG_ENDIAN ? TARGET_CHAR_BIT - 1 - bit_index : bit_index);
+      byte >>= (gdbarch_bits_big_endian (current_gdbarch) ?
+		TARGET_CHAR_BIT - 1 - bit_index : bit_index);
       v = value_from_longest (LA_BOOL_TYPE, byte & 1);
       set_value_bitpos (v, bit_index);
       set_value_bitsize (v, 1);
@@ -1573,7 +1574,7 @@ value_bit_index (struct type *type, cons
   word = unpack_long (builtin_type_unsigned_char,
 		      valaddr + (rel_index / TARGET_CHAR_BIT));
   rel_index %= TARGET_CHAR_BIT;
-  if (BITS_BIG_ENDIAN)
+  if (gdbarch_bits_big_endian (current_gdbarch))
     rel_index = TARGET_CHAR_BIT - 1 - rel_index;
   return (word >> rel_index) & 1;
 }
diff -urpN src/gdb/valops.c dev/gdb/valops.c
--- src/gdb/valops.c	2008-01-08 11:22:24.000000000 +0100
+++ dev/gdb/valops.c	2008-01-15 07:28:43.000000000 +0100
@@ -2743,7 +2743,7 @@ value_slice (struct value *array, int lo
 	  else if (element > 0)
 	    {
 	      int j = i % TARGET_CHAR_BIT;
-	      if (BITS_BIG_ENDIAN)
+	      if (gdbarch_bits_big_endian (current_gdbarch))
 		j = TARGET_CHAR_BIT - 1 - j;
 	      value_contents_raw (slice)[i / TARGET_CHAR_BIT] |= (1 << j);
 	    }
diff -urpN src/gdb/value.c dev/gdb/value.c
--- src/gdb/value.c	2008-01-08 11:22:24.000000000 +0100
+++ dev/gdb/value.c	2008-01-15 07:26:01.000000000 +0100
@@ -72,8 +72,8 @@ struct value
   int bitsize;
 
   /* Only used for bitfields; position of start of field.  For
-     BITS_BIG_ENDIAN=0 targets, it is the position of the LSB.  For
-     BITS_BIG_ENDIAN=1 targets, it is the position of the MSB. */
+     gdbarch_bits_big_endian=0 targets, it is the position of the LSB.  For
+     gdbarch_bits_big_endian=1 targets, it is the position of the MSB. */
   int bitpos;
 
   /* Frame register value is relative to.  This will be described in
@@ -1481,7 +1481,7 @@ unpack_field_as_long (struct type *type,
 
   /* Extract bits.  See comment above. */
 
-  if (BITS_BIG_ENDIAN)
+  if (gdbarch_bits_big_endian (current_gdbarch))
     lsbcount = (sizeof val * 8 - bitpos % 8 - bitsize);
   else
     lsbcount = (bitpos % 8);
@@ -1537,7 +1537,7 @@ modify_field (gdb_byte *addr, LONGEST fi
   oword = extract_unsigned_integer (addr, sizeof oword);
 
   /* Shifting for bit field depends on endianness of the target machine.  */
-  if (BITS_BIG_ENDIAN)
+  if (gdbarch_bits_big_endian (current_gdbarch))
     bitpos = sizeof (oword) * 8 - bitpos - bitsize;
 
   oword &= ~(mask << bitpos);
diff -urpN src/gdb/value.h dev/gdb/value.h
--- src/gdb/value.h	2008-01-08 11:22:24.000000000 +0100
+++ dev/gdb/value.h	2008-01-15 07:27:58.000000000 +0100
@@ -62,8 +62,8 @@ extern int value_bitsize (struct value *
 extern void set_value_bitsize (struct value *, int bit);
 
 /* Only used for bitfields; position of start of field.  For
-   BITS_BIG_ENDIAN=0 targets, it is the position of the LSB.  For
-   BITS_BIG_ENDIAN=1 targets, it is the position of the MSB.  */
+   gdbarch_bits_big_endian=0 targets, it is the position of the LSB.  For
+   gdbarch_bits_big_endian=1 targets, it is the position of the MSB.  */
 
 extern int value_bitpos (struct value *);
 extern void set_value_bitpos (struct value *, int bit);


  reply	other threads:[~2008-01-15  9:06 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-01-11 18:20 Markus Deuling
2008-01-11 19:24 ` Daniel Jacobowitz
2008-01-11 20:59 ` Eli Zaretskii
2008-01-14 18:11 ` Ulrich Weigand
2008-01-15  9:06   ` Markus Deuling [this message]
2008-01-15 21:01     ` Ulrich Weigand
2008-01-16  5:59       ` Markus Deuling
2008-01-16 18:43         ` Eli Zaretskii
2008-01-17  7:09           ` Markus Deuling
2008-01-18 16:21             ` Eli Zaretskii
2008-01-18 17:08               ` Markus Deuling

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=478C76F5.2000402@de.ibm.com \
    --to=deuling@de.ibm.com \
    --cc=drow@false.org \
    --cc=eliz@gnu.org \
    --cc=gdb-patches@sourceware.org \
    --cc=uweigand@de.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox