1997-09-24 19:39:15 +04:00
|
|
|
|
/* `a.out' object-file definitions, including extensions to 64-bit fields */
|
|
|
|
|
|
|
|
|
|
#ifndef __A_OUT_64_H__
|
|
|
|
|
#define __A_OUT_64_H__
|
|
|
|
|
|
|
|
|
|
/* 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
|
|
|
|
|
#define N_SHARED_LIB(x) ((x).a_entry < TEXT_START_ADDR)
|
|
|
|
|
#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) ? TARGET_PAGE_SIZE + EXEC_BYTES_SIZE : \
|
|
|
|
|
N_MAGIC(x) != ZMAGIC ? 0 : /* object file or NMAGIC */\
|
|
|
|
|
N_SHARED_LIB(x) ? 0 : \
|
|
|
|
|
N_HEADER_IN_TEXT(x) ? \
|
|
|
|
|
TEXT_START_ADDR + EXEC_BYTES_SIZE : /* no padding */\
|
|
|
|
|
TEXT_START_ADDR /* a page of padding */\
|
|
|
|
|
)
|
|
|
|
|
#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) & ~(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) */
|
1998-09-13 13:24:33 +04:00
|
|
|
|
#define N_SIZE 0x0C /* A size sybol -- used by NetBSD */
|
1997-09-24 19:39:15 +04:00
|
|
|
|
/* Note: N_EXT can only be usefully OR-ed with N_UNDF, N_ABS, N_TEXT,
|
1998-09-13 13:24:33 +04:00
|
|
|
|
N_DATA, N_BSS or N_SIZE. When the low-order bit of other types is set,
|
1997-09-24 19:39:15 +04:00
|
|
|
|
(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 */
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
#define RELOC_EXT_BITS_EXTERN_BIG ((unsigned int) 0x80)
|
|
|
|
|
#define RELOC_EXT_BITS_EXTERN_LITTLE ((unsigned int) 0x01)
|
|
|
|
|
|
|
|
|
|
#define RELOC_EXT_BITS_TYPE_BIG ((unsigned int) 0x1F)
|
|
|
|
|
#define RELOC_EXT_BITS_TYPE_SH_BIG 0
|
|
|
|
|
#define RELOC_EXT_BITS_TYPE_LITTLE ((unsigned int) 0xF8)
|
|
|
|
|
#define RELOC_EXT_BITS_TYPE_SH_LITTLE 3
|
|
|
|
|
|
|
|
|
|
/* 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__ */
|