mirror of
https://sourceware.org/git/binutils-gdb.git
synced 2024-11-30 21:44:19 +08:00
7eb5191ac3
aout64.h.
520 lines
19 KiB
C
520 lines
19 KiB
C
/* `a.out' object-file definitions, including extensions to 64-bit fields
|
||
|
||
Copyright 2001, 2003 Free Software Foundation, Inc.
|
||
|
||
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 2 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, write to the Free Software
|
||
Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
|
||
|
||
#ifndef __A_OUT_64_H__
|
||
#define __A_OUT_64_H__
|
||
|
||
#ifndef BYTES_IN_WORD
|
||
#define BYTES_IN_WORD 4
|
||
#endif
|
||
|
||
/* This is the layout on disk of the 32-bit or 64-bit exec header. */
|
||
|
||
#ifndef external_exec
|
||
struct external_exec
|
||
{
|
||
bfd_byte e_info[4]; /* Magic number and stuff. */
|
||
bfd_byte e_text[BYTES_IN_WORD]; /* Length of text section in bytes. */
|
||
bfd_byte e_data[BYTES_IN_WORD]; /* Length of data section in bytes. */
|
||
bfd_byte e_bss[BYTES_IN_WORD]; /* Length of bss area in bytes. */
|
||
bfd_byte e_syms[BYTES_IN_WORD]; /* Length of symbol table in bytes. */
|
||
bfd_byte e_entry[BYTES_IN_WORD]; /* Start address. */
|
||
bfd_byte e_trsize[BYTES_IN_WORD]; /* Length of text relocation info. */
|
||
bfd_byte e_drsize[BYTES_IN_WORD]; /* Length of data relocation info. */
|
||
};
|
||
|
||
#define EXEC_BYTES_SIZE (4 + BYTES_IN_WORD * 7)
|
||
|
||
/* Magic numbers for a.out files. */
|
||
|
||
#if ARCH_SIZE==64
|
||
#define OMAGIC 0x1001 /* Code indicating object file. */
|
||
#define ZMAGIC 0x1002 /* Code indicating demand-paged executable. */
|
||
#define NMAGIC 0x1003 /* Code indicating pure executable. */
|
||
|
||
/* There is no 64-bit QMAGIC as far as I know. */
|
||
|
||
#define N_BADMAG(x) (N_MAGIC(x) != OMAGIC \
|
||
&& N_MAGIC(x) != NMAGIC \
|
||
&& N_MAGIC(x) != ZMAGIC)
|
||
#else
|
||
#define OMAGIC 0407 /* Object file or impure executable. */
|
||
#define NMAGIC 0410 /* Code indicating pure executable. */
|
||
#define ZMAGIC 0413 /* Code indicating demand-paged executable. */
|
||
#define BMAGIC 0415 /* Used by a b.out object. */
|
||
|
||
/* This indicates a demand-paged executable with the header in the text.
|
||
It is used by 386BSD (and variants) and Linux, at least. */
|
||
#ifndef QMAGIC
|
||
#define QMAGIC 0314
|
||
#endif
|
||
# ifndef N_BADMAG
|
||
# define N_BADMAG(x) (N_MAGIC(x) != OMAGIC \
|
||
&& N_MAGIC(x) != NMAGIC \
|
||
&& N_MAGIC(x) != ZMAGIC \
|
||
&& N_MAGIC(x) != QMAGIC)
|
||
# endif /* N_BADMAG */
|
||
#endif
|
||
|
||
#endif
|
||
|
||
#ifdef QMAGIC
|
||
#define N_IS_QMAGIC(x) (N_MAGIC (x) == QMAGIC)
|
||
#else
|
||
#define N_IS_QMAGIC(x) (0)
|
||
#endif
|
||
|
||
/* The difference between TARGET_PAGE_SIZE and N_SEGSIZE is that TARGET_PAGE_SIZE is
|
||
the finest granularity at which you can page something, thus it
|
||
controls the padding (if any) before the text segment of a ZMAGIC
|
||
file. N_SEGSIZE is the resolution at which things can be marked as
|
||
read-only versus read/write, so it controls the padding between the
|
||
text segment and the data segment (in memory; on disk the padding
|
||
between them is TARGET_PAGE_SIZE). TARGET_PAGE_SIZE and N_SEGSIZE are the same
|
||
for most machines, but different for sun3. */
|
||
|
||
/* By default, segment size is constant. But some machines override this
|
||
to be a function of the a.out header (e.g. machine type). */
|
||
|
||
#ifndef N_SEGSIZE
|
||
#define N_SEGSIZE(x) SEGMENT_SIZE
|
||
#endif
|
||
|
||
/* Virtual memory address of the text section.
|
||
This is getting very complicated. A good reason to discard a.out format
|
||
for something that specifies these fields explicitly. But til then...
|
||
|
||
* OMAGIC and NMAGIC files:
|
||
(object files: text for "relocatable addr 0" right after the header)
|
||
start at 0, offset is EXEC_BYTES_SIZE, size as stated.
|
||
* The text address, offset, and size of ZMAGIC files depend
|
||
on the entry point of the file:
|
||
* entry point below TEXT_START_ADDR:
|
||
(hack for SunOS shared libraries)
|
||
start at 0, offset is 0, size as stated.
|
||
* If N_HEADER_IN_TEXT(x) is true (which defaults to being the
|
||
case when the entry point is EXEC_BYTES_SIZE or further into a page):
|
||
no padding is needed; text can start after exec header. Sun
|
||
considers the text segment of such files to include the exec header;
|
||
for BFD's purposes, we don't, which makes more work for us.
|
||
start at TEXT_START_ADDR + EXEC_BYTES_SIZE, offset is EXEC_BYTES_SIZE,
|
||
size as stated minus EXEC_BYTES_SIZE.
|
||
* If N_HEADER_IN_TEXT(x) is false (which defaults to being the case when
|
||
the entry point is less than EXEC_BYTES_SIZE into a page (e.g. page
|
||
aligned)): (padding is needed so that text can start at a page boundary)
|
||
start at TEXT_START_ADDR, offset TARGET_PAGE_SIZE, size as stated.
|
||
|
||
Specific configurations may want to hardwire N_HEADER_IN_TEXT,
|
||
for efficiency or to allow people to play games with the entry point.
|
||
In that case, you would #define N_HEADER_IN_TEXT(x) as 1 for sunos,
|
||
and as 0 for most other hosts (Sony News, Vax Ultrix, etc).
|
||
(Do this in the appropriate bfd target file.)
|
||
(The default is a heuristic that will break if people try changing
|
||
the entry point, perhaps with the ld -e flag.)
|
||
|
||
* QMAGIC is always like a ZMAGIC for which N_HEADER_IN_TEXT is true,
|
||
and for which the starting address is TARGET_PAGE_SIZE (or should this be
|
||
SEGMENT_SIZE?) (TEXT_START_ADDR only applies to ZMAGIC, not to QMAGIC). */
|
||
|
||
/* This macro is only relevant for ZMAGIC files; QMAGIC always has the header
|
||
in the text. */
|
||
#ifndef N_HEADER_IN_TEXT
|
||
#define N_HEADER_IN_TEXT(x) \
|
||
(((x).a_entry & (TARGET_PAGE_SIZE-1)) >= EXEC_BYTES_SIZE)
|
||
#endif
|
||
|
||
/* Sun shared libraries, not linux. This macro is only relevant for ZMAGIC
|
||
files. */
|
||
#ifndef N_SHARED_LIB
|
||
#if defined (TEXT_START_ADDR) && TEXT_START_ADDR == 0
|
||
#define N_SHARED_LIB(x) (0)
|
||
#else
|
||
#define N_SHARED_LIB(x) ((x).a_entry < TEXT_START_ADDR)
|
||
#endif
|
||
#endif
|
||
|
||
/* Returning 0 not TEXT_START_ADDR for OMAGIC and NMAGIC is based on
|
||
the assumption that we are dealing with a .o file, not an
|
||
executable. This is necessary for OMAGIC (but means we don't work
|
||
right on the output from ld -N); more questionable for NMAGIC. */
|
||
|
||
#ifndef N_TXTADDR
|
||
#define N_TXTADDR(x) \
|
||
(/* The address of a QMAGIC file is always one page in, \
|
||
with the header in the text. */ \
|
||
N_IS_QMAGIC (x) \
|
||
? (bfd_vma) TARGET_PAGE_SIZE + EXEC_BYTES_SIZE \
|
||
: (N_MAGIC (x) != ZMAGIC \
|
||
? (bfd_vma) 0 /* Object file or NMAGIC. */ \
|
||
: (N_SHARED_LIB (x) \
|
||
? (bfd_vma) 0 \
|
||
: (N_HEADER_IN_TEXT (x) \
|
||
? (bfd_vma) TEXT_START_ADDR + EXEC_BYTES_SIZE \
|
||
: (bfd_vma) TEXT_START_ADDR))))
|
||
#endif
|
||
|
||
/* If N_HEADER_IN_TEXT is not true for ZMAGIC, there is some padding
|
||
to make the text segment start at a certain boundary. For most
|
||
systems, this boundary is TARGET_PAGE_SIZE. But for Linux, in the
|
||
time-honored tradition of crazy ZMAGIC hacks, it is 1024 which is
|
||
not what TARGET_PAGE_SIZE needs to be for QMAGIC. */
|
||
|
||
#ifndef ZMAGIC_DISK_BLOCK_SIZE
|
||
#define ZMAGIC_DISK_BLOCK_SIZE TARGET_PAGE_SIZE
|
||
#endif
|
||
|
||
#define N_DISK_BLOCK_SIZE(x) \
|
||
(N_MAGIC(x) == ZMAGIC ? ZMAGIC_DISK_BLOCK_SIZE : TARGET_PAGE_SIZE)
|
||
|
||
/* Offset in an a.out of the start of the text section. */
|
||
#ifndef N_TXTOFF
|
||
#define N_TXTOFF(x) \
|
||
(/* For {O,N,Q}MAGIC, no padding. */ \
|
||
N_MAGIC (x) != ZMAGIC \
|
||
? EXEC_BYTES_SIZE \
|
||
: (N_SHARED_LIB (x) \
|
||
? 0 \
|
||
: (N_HEADER_IN_TEXT (x) \
|
||
? EXEC_BYTES_SIZE /* No padding. */ \
|
||
: ZMAGIC_DISK_BLOCK_SIZE /* A page of padding. */)))
|
||
#endif
|
||
/* Size of the text section. It's always as stated, except that we
|
||
offset it to `undo' the adjustment to N_TXTADDR and N_TXTOFF
|
||
for ZMAGIC files that nominally include the exec header
|
||
as part of the first page of text. (BFD doesn't consider the
|
||
exec header to be part of the text segment.) */
|
||
#ifndef N_TXTSIZE
|
||
#define N_TXTSIZE(x) \
|
||
(/* For QMAGIC, we don't consider the header part of the text section. */\
|
||
N_IS_QMAGIC (x) \
|
||
? (x).a_text - EXEC_BYTES_SIZE \
|
||
: ((N_MAGIC (x) != ZMAGIC || N_SHARED_LIB (x)) \
|
||
? (x).a_text \
|
||
: (N_HEADER_IN_TEXT (x) \
|
||
? (x).a_text - EXEC_BYTES_SIZE /* No padding. */ \
|
||
: (x).a_text /* A page of padding. */ )))
|
||
#endif
|
||
/* The address of the data segment in virtual memory.
|
||
It is the text segment address, plus text segment size, rounded
|
||
up to a N_SEGSIZE boundary for pure or pageable files. */
|
||
#ifndef N_DATADDR
|
||
#define N_DATADDR(x) \
|
||
(N_MAGIC (x) == OMAGIC \
|
||
? (N_TXTADDR (x) + N_TXTSIZE (x)) \
|
||
: (N_SEGSIZE (x) + ((N_TXTADDR (x) + N_TXTSIZE (x) - 1) \
|
||
& ~ (bfd_vma) (N_SEGSIZE (x) - 1))))
|
||
#endif
|
||
/* The address of the BSS segment -- immediately after the data segment. */
|
||
|
||
#define N_BSSADDR(x) (N_DATADDR (x) + (x).a_data)
|
||
|
||
/* Offsets of the various portions of the file after the text segment. */
|
||
|
||
/* For {Q,Z}MAGIC, there is padding to make the data segment start on
|
||
a page boundary. Most of the time the a_text field (and thus
|
||
N_TXTSIZE) already contains this padding. It is possible that for
|
||
BSDI and/or 386BSD it sometimes doesn't contain the padding, and
|
||
perhaps we should be adding it here. But this seems kind of
|
||
questionable and probably should be BSDI/386BSD-specific if we do
|
||
do it.
|
||
|
||
For NMAGIC (at least for hp300 BSD, probably others), there is
|
||
padding in memory only, not on disk, so we must *not* ever pad here
|
||
for NMAGIC. */
|
||
|
||
#ifndef N_DATOFF
|
||
#define N_DATOFF(x) (N_TXTOFF (x) + N_TXTSIZE (x))
|
||
#endif
|
||
#ifndef N_TRELOFF
|
||
#define N_TRELOFF(x) (N_DATOFF (x) + (x).a_data)
|
||
#endif
|
||
#ifndef N_DRELOFF
|
||
#define N_DRELOFF(x) (N_TRELOFF (x) + (x).a_trsize)
|
||
#endif
|
||
#ifndef N_SYMOFF
|
||
#define N_SYMOFF(x) (N_DRELOFF (x) + (x).a_drsize)
|
||
#endif
|
||
#ifndef N_STROFF
|
||
#define N_STROFF(x) (N_SYMOFF (x) + (x).a_syms)
|
||
#endif
|
||
|
||
/* Symbols */
|
||
#ifndef external_nlist
|
||
struct external_nlist
|
||
{
|
||
bfd_byte e_strx[BYTES_IN_WORD]; /* Index into string table of name. */
|
||
bfd_byte e_type[1]; /* Type of symbol. */
|
||
bfd_byte e_other[1]; /* Misc info (usually empty). */
|
||
bfd_byte e_desc[2]; /* Description field. */
|
||
bfd_byte e_value[BYTES_IN_WORD]; /* Value of symbol. */
|
||
};
|
||
#define EXTERNAL_NLIST_SIZE (BYTES_IN_WORD+4+BYTES_IN_WORD)
|
||
#endif
|
||
|
||
struct internal_nlist
|
||
{
|
||
unsigned long n_strx; /* Index into string table of name. */
|
||
unsigned char n_type; /* Type of symbol. */
|
||
unsigned char n_other; /* Misc info (usually empty). */
|
||
unsigned short n_desc; /* Description field. */
|
||
bfd_vma n_value; /* Value of symbol. */
|
||
};
|
||
|
||
/* The n_type field is the symbol type, containing: */
|
||
|
||
#define N_UNDF 0 /* Undefined symbol. */
|
||
#define N_ABS 2 /* Absolute symbol -- defined at particular addr. */
|
||
#define N_TEXT 4 /* Text sym -- defined at offset in text seg. */
|
||
#define N_DATA 6 /* Data sym -- defined at offset in data seg. */
|
||
#define N_BSS 8 /* BSS sym -- defined at offset in zero'd seg. */
|
||
#define N_COMM 0x12 /* Common symbol (visible after shared lib dynlink). */
|
||
#define N_FN 0x1f /* File name of .o file. */
|
||
#define N_FN_SEQ 0x0C /* N_FN from Sequent compilers (sigh). */
|
||
/* Note: N_EXT can only be usefully OR-ed with N_UNDF, N_ABS, N_TEXT,
|
||
N_DATA, or N_BSS. When the low-order bit of other types is set,
|
||
(e.g. N_WARNING versus N_FN), they are two different types. */
|
||
#define N_EXT 1 /* External symbol (as opposed to local-to-this-file). */
|
||
#define N_TYPE 0x1e
|
||
#define N_STAB 0xe0 /* If any of these bits are on, it's a debug symbol. */
|
||
|
||
#define N_INDR 0x0a
|
||
|
||
/* The following symbols refer to set elements.
|
||
All the N_SET[ATDB] symbols with the same name form one set.
|
||
Space is allocated for the set in the text section, and each set
|
||
elements value is stored into one word of the space.
|
||
The first word of the space is the length of the set (number of elements).
|
||
|
||
The address of the set is made into an N_SETV symbol
|
||
whose name is the same as the name of the set.
|
||
This symbol acts like a N_DATA global symbol
|
||
in that it can satisfy undefined external references. */
|
||
|
||
/* These appear as input to LD, in a .o file. */
|
||
#define N_SETA 0x14 /* Absolute set element symbol. */
|
||
#define N_SETT 0x16 /* Text set element symbol. */
|
||
#define N_SETD 0x18 /* Data set element symbol. */
|
||
#define N_SETB 0x1A /* Bss set element symbol. */
|
||
|
||
/* This is output from LD. */
|
||
#define N_SETV 0x1C /* Pointer to set vector in data area. */
|
||
|
||
/* Warning symbol. The text gives a warning message, the next symbol
|
||
in the table will be undefined. When the symbol is referenced, the
|
||
message is printed. */
|
||
|
||
#define N_WARNING 0x1e
|
||
|
||
/* Weak symbols. These are a GNU extension to the a.out format. The
|
||
semantics are those of ELF weak symbols. Weak symbols are always
|
||
externally visible. The N_WEAK? values are squeezed into the
|
||
available slots. The value of a N_WEAKU symbol is 0. The values
|
||
of the other types are the definitions. */
|
||
#define N_WEAKU 0x0d /* Weak undefined symbol. */
|
||
#define N_WEAKA 0x0e /* Weak absolute symbol. */
|
||
#define N_WEAKT 0x0f /* Weak text symbol. */
|
||
#define N_WEAKD 0x10 /* Weak data symbol. */
|
||
#define N_WEAKB 0x11 /* Weak bss symbol. */
|
||
|
||
/* Relocations
|
||
|
||
There are two types of relocation flavours for a.out systems,
|
||
standard and extended. The standard form is used on systems where the
|
||
instruction has room for all the bits of an offset to the operand, whilst
|
||
the extended form is used when an address operand has to be split over n
|
||
instructions. Eg, on the 68k, each move instruction can reference
|
||
the target with a displacement of 16 or 32 bits. On the sparc, move
|
||
instructions use an offset of 14 bits, so the offset is stored in
|
||
the reloc field, and the data in the section is ignored. */
|
||
|
||
/* This structure describes a single relocation to be performed.
|
||
The text-relocation section of the file is a vector of these structures,
|
||
all of which apply to the text section.
|
||
Likewise, the data-relocation section applies to the data section. */
|
||
|
||
struct reloc_std_external
|
||
{
|
||
bfd_byte r_address[BYTES_IN_WORD]; /* Offset of of data to relocate. */
|
||
bfd_byte r_index[3]; /* Symbol table index of symbol. */
|
||
bfd_byte r_type[1]; /* Relocation type. */
|
||
};
|
||
|
||
#define RELOC_STD_BITS_PCREL_BIG ((unsigned int) 0x80)
|
||
#define RELOC_STD_BITS_PCREL_LITTLE ((unsigned int) 0x01)
|
||
|
||
#define RELOC_STD_BITS_LENGTH_BIG ((unsigned int) 0x60)
|
||
#define RELOC_STD_BITS_LENGTH_SH_BIG 5
|
||
#define RELOC_STD_BITS_LENGTH_LITTLE ((unsigned int) 0x06)
|
||
#define RELOC_STD_BITS_LENGTH_SH_LITTLE 1
|
||
|
||
#define RELOC_STD_BITS_EXTERN_BIG ((unsigned int) 0x10)
|
||
#define RELOC_STD_BITS_EXTERN_LITTLE ((unsigned int) 0x08)
|
||
|
||
#define RELOC_STD_BITS_BASEREL_BIG ((unsigned int) 0x08)
|
||
#define RELOC_STD_BITS_BASEREL_LITTLE ((unsigned int) 0x10)
|
||
|
||
#define RELOC_STD_BITS_JMPTABLE_BIG ((unsigned int) 0x04)
|
||
#define RELOC_STD_BITS_JMPTABLE_LITTLE ((unsigned int) 0x20)
|
||
|
||
#define RELOC_STD_BITS_RELATIVE_BIG ((unsigned int) 0x02)
|
||
#define RELOC_STD_BITS_RELATIVE_LITTLE ((unsigned int) 0x40)
|
||
|
||
#define RELOC_STD_SIZE (BYTES_IN_WORD + 3 + 1) /* Bytes per relocation entry. */
|
||
|
||
struct reloc_std_internal
|
||
{
|
||
bfd_vma r_address; /* Address (within segment) to be relocated. */
|
||
/* The meaning of r_symbolnum depends on r_extern. */
|
||
unsigned int r_symbolnum:24;
|
||
/* Nonzero means value is a pc-relative offset
|
||
and it should be relocated for changes in its own address
|
||
as well as for changes in the symbol or section specified. */
|
||
unsigned int r_pcrel:1;
|
||
/* Length (as exponent of 2) of the field to be relocated.
|
||
Thus, a value of 2 indicates 1<<2 bytes. */
|
||
unsigned int r_length:2;
|
||
/* 1 => relocate with value of symbol.
|
||
r_symbolnum is the index of the symbol
|
||
in files the symbol table.
|
||
0 => relocate with the address of a segment.
|
||
r_symbolnum is N_TEXT, N_DATA, N_BSS or N_ABS
|
||
(the N_EXT bit may be set also, but signifies nothing). */
|
||
unsigned int r_extern:1;
|
||
/* The next three bits are for SunOS shared libraries, and seem to
|
||
be undocumented. */
|
||
unsigned int r_baserel:1; /* Linkage table relative. */
|
||
unsigned int r_jmptable:1; /* pc-relative to jump table. */
|
||
unsigned int r_relative:1; /* "relative relocation". */
|
||
/* unused */
|
||
unsigned int r_pad:1; /* Padding -- set to zero. */
|
||
};
|
||
|
||
|
||
/* EXTENDED RELOCS. */
|
||
|
||
struct reloc_ext_external
|
||
{
|
||
bfd_byte r_address[BYTES_IN_WORD]; /* Offset of of data to relocate. */
|
||
bfd_byte r_index[3]; /* Symbol table index of symbol. */
|
||
bfd_byte r_type[1]; /* Relocation type. */
|
||
bfd_byte r_addend[BYTES_IN_WORD]; /* Datum addend. */
|
||
};
|
||
|
||
#ifndef RELOC_EXT_BITS_EXTERN_BIG
|
||
#define RELOC_EXT_BITS_EXTERN_BIG ((unsigned int) 0x80)
|
||
#endif
|
||
|
||
#ifndef RELOC_EXT_BITS_EXTERN_LITTLE
|
||
#define RELOC_EXT_BITS_EXTERN_LITTLE ((unsigned int) 0x01)
|
||
#endif
|
||
|
||
#ifndef RELOC_EXT_BITS_TYPE_BIG
|
||
#define RELOC_EXT_BITS_TYPE_BIG ((unsigned int) 0x1F)
|
||
#endif
|
||
|
||
#ifndef RELOC_EXT_BITS_TYPE_SH_BIG
|
||
#define RELOC_EXT_BITS_TYPE_SH_BIG 0
|
||
#endif
|
||
|
||
#ifndef RELOC_EXT_BITS_TYPE_LITTLE
|
||
#define RELOC_EXT_BITS_TYPE_LITTLE ((unsigned int) 0xF8)
|
||
#endif
|
||
|
||
#ifndef RELOC_EXT_BITS_TYPE_SH_LITTLE
|
||
#define RELOC_EXT_BITS_TYPE_SH_LITTLE 3
|
||
#endif
|
||
|
||
/* Bytes per relocation entry. */
|
||
#define RELOC_EXT_SIZE (BYTES_IN_WORD + 3 + 1 + BYTES_IN_WORD)
|
||
|
||
enum reloc_type
|
||
{
|
||
/* Simple relocations. */
|
||
RELOC_8, /* data[0:7] = addend + sv */
|
||
RELOC_16, /* data[0:15] = addend + sv */
|
||
RELOC_32, /* data[0:31] = addend + sv */
|
||
/* PC-rel displacement. */
|
||
RELOC_DISP8, /* data[0:7] = addend - pc + sv */
|
||
RELOC_DISP16, /* data[0:15] = addend - pc + sv */
|
||
RELOC_DISP32, /* data[0:31] = addend - pc + sv */
|
||
/* Special. */
|
||
RELOC_WDISP30, /* data[0:29] = (addend + sv - pc)>>2 */
|
||
RELOC_WDISP22, /* data[0:21] = (addend + sv - pc)>>2 */
|
||
RELOC_HI22, /* data[0:21] = (addend + sv)>>10 */
|
||
RELOC_22, /* data[0:21] = (addend + sv) */
|
||
RELOC_13, /* data[0:12] = (addend + sv) */
|
||
RELOC_LO10, /* data[0:9] = (addend + sv) */
|
||
RELOC_SFA_BASE,
|
||
RELOC_SFA_OFF13,
|
||
/* P.I.C. (base-relative). */
|
||
RELOC_BASE10, /* Not sure - maybe we can do this the */
|
||
RELOC_BASE13, /* right way now */
|
||
RELOC_BASE22,
|
||
/* For some sort of pc-rel P.I.C. (?) */
|
||
RELOC_PC10,
|
||
RELOC_PC22,
|
||
/* P.I.C. jump table. */
|
||
RELOC_JMP_TBL,
|
||
/* Reputedly for shared libraries somehow. */
|
||
RELOC_SEGOFF16,
|
||
RELOC_GLOB_DAT,
|
||
RELOC_JMP_SLOT,
|
||
RELOC_RELATIVE,
|
||
|
||
RELOC_11,
|
||
RELOC_WDISP2_14,
|
||
RELOC_WDISP19,
|
||
RELOC_HHI22, /* data[0:21] = (addend + sv) >> 42 */
|
||
RELOC_HLO10, /* data[0:9] = (addend + sv) >> 32 */
|
||
|
||
/* 29K relocation types. */
|
||
RELOC_JUMPTARG,
|
||
RELOC_CONST,
|
||
RELOC_CONSTH,
|
||
|
||
/* All the new ones I can think of, for sparc v9. */
|
||
RELOC_64, /* data[0:63] = addend + sv */
|
||
RELOC_DISP64, /* data[0:63] = addend - pc + sv */
|
||
RELOC_WDISP21, /* data[0:20] = (addend + sv - pc)>>2 */
|
||
RELOC_DISP21, /* data[0:20] = addend - pc + sv */
|
||
RELOC_DISP14, /* data[0:13] = addend - pc + sv */
|
||
/* Q .
|
||
What are the other ones,
|
||
Since this is a clean slate, can we throw away the ones we dont
|
||
understand ? Should we sort the values ? What about using a
|
||
microcode format like the 68k ? */
|
||
NO_RELOC
|
||
};
|
||
|
||
|
||
struct reloc_internal
|
||
{
|
||
bfd_vma r_address; /* Offset of of data to relocate. */
|
||
long r_index; /* Symbol table index of symbol. */
|
||
enum reloc_type r_type; /* Relocation type. */
|
||
bfd_vma r_addend; /* Datum addend. */
|
||
};
|
||
|
||
/* Q.
|
||
Should the length of the string table be 4 bytes or 8 bytes ?
|
||
|
||
Q.
|
||
What about archive indexes ? */
|
||
|
||
#endif /* __A_OUT_64_H__ */
|