mirror of
https://sourceware.org/git/binutils-gdb.git
synced 2024-11-27 03:51:15 +08:00
c8a1f2548b
* elf32-ppc.c (ppc_elf_relax_section): Work with a partial link. * bout.c (b_out_bfd_relax_section): Reject relocatable links. * elf32-m10300.c (mn10300_elf_relax_section): Likewise. * elf32-avr.c (elf32_avr_relax_section): Likewise. * elf32-frv.c (elf32_avr_relax_section): Likewise. * elf32-xtensa.c (elf_xtensa_relax_section): Likewise. * elf64-mmix.c (mmix_elf_relax_section): Likewise. * elfxx-ia64.c (elfNN_ia64_relax_section): Likewise. * elfxx-sparc.c (_bfd_sparc_elf_relax_section): Likewise. * reloc.c (bfd_generic_relax_section): Likewise. * reloc16.c (bfd_coff_reloc16_relax_section): Likewise. * vms.c (vms_bfd_relax_section): Likewise. ld/ * ldmain.c (main): Don't reject --relax -r. * ld.texinfo (PowerPC ELF32): Document behaviour of relaxing partial links. ld/testsuite/ * ld-powerpc/vxworks-relax-2.s: New. * ld-powerpc/vxworks-relax-2.rd: New. * ld-powerpc/powerpc.exp: Add it.
342 lines
9.9 KiB
C
342 lines
9.9 KiB
C
/* 8 and 16 bit COFF relocation functions, for BFD.
|
|
Copyright 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1998, 2000, 2001,
|
|
2002, 2003, 2004, 2007 Free Software Foundation, Inc.
|
|
Written by Cygnus Support.
|
|
|
|
This file is part of BFD, the Binary File Descriptor library.
|
|
|
|
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, write to the Free Software
|
|
Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston,
|
|
MA 02110-1301, USA. */
|
|
|
|
|
|
/* Most of this hacked by Steve Chamberlain <sac@cygnus.com>. */
|
|
|
|
/* These routines are used by coff-h8300 and coff-z8k to do
|
|
relocation.
|
|
|
|
FIXME: This code should be rewritten to support the new COFF
|
|
linker. Basically, they need to deal with COFF relocs rather than
|
|
BFD generic relocs. They should store the relocs in some location
|
|
where coff_link_input_bfd can find them (and coff_link_input_bfd
|
|
should be changed to use this location rather than rereading the
|
|
file) (unless info->keep_memory is FALSE, in which case they should
|
|
free up the relocs after dealing with them). */
|
|
|
|
#include "sysdep.h"
|
|
#include "bfd.h"
|
|
#include "libbfd.h"
|
|
#include "bfdlink.h"
|
|
#include "genlink.h"
|
|
#include "coff/internal.h"
|
|
#include "libcoff.h"
|
|
|
|
bfd_vma
|
|
bfd_coff_reloc16_get_value (reloc, link_info, input_section)
|
|
arelent *reloc;
|
|
struct bfd_link_info *link_info;
|
|
asection *input_section;
|
|
{
|
|
bfd_vma value;
|
|
asymbol *symbol = *(reloc->sym_ptr_ptr);
|
|
/* A symbol holds a pointer to a section, and an offset from the
|
|
base of the section. To relocate, we find where the section will
|
|
live in the output and add that in. */
|
|
|
|
if (bfd_is_und_section (symbol->section)
|
|
|| bfd_is_com_section (symbol->section))
|
|
{
|
|
struct bfd_link_hash_entry *h;
|
|
|
|
/* The symbol is undefined in this BFD. Look it up in the
|
|
global linker hash table. FIXME: This should be changed when
|
|
we convert this stuff to use a specific final_link function
|
|
and change the interface to bfd_relax_section to not require
|
|
the generic symbols. */
|
|
h = bfd_wrapped_link_hash_lookup (input_section->owner, link_info,
|
|
bfd_asymbol_name (symbol),
|
|
FALSE, FALSE, TRUE);
|
|
if (h != (struct bfd_link_hash_entry *) NULL
|
|
&& (h->type == bfd_link_hash_defined
|
|
|| h->type == bfd_link_hash_defweak))
|
|
value = (h->u.def.value
|
|
+ h->u.def.section->output_section->vma
|
|
+ h->u.def.section->output_offset);
|
|
else if (h != (struct bfd_link_hash_entry *) NULL
|
|
&& h->type == bfd_link_hash_common)
|
|
value = h->u.c.size;
|
|
else if (h != (struct bfd_link_hash_entry *) NULL
|
|
&& h->type == bfd_link_hash_undefweak)
|
|
/* This is a GNU extension. */
|
|
value = 0;
|
|
else
|
|
{
|
|
if (!((*link_info->callbacks->undefined_symbol)
|
|
(link_info, bfd_asymbol_name (symbol),
|
|
input_section->owner, input_section, reloc->address,
|
|
TRUE)))
|
|
abort ();
|
|
value = 0;
|
|
}
|
|
}
|
|
else
|
|
{
|
|
value = symbol->value
|
|
+ symbol->section->output_offset
|
|
+ symbol->section->output_section->vma;
|
|
}
|
|
|
|
/* Add the value contained in the relocation. */
|
|
value += reloc->addend;
|
|
|
|
return value;
|
|
}
|
|
|
|
void
|
|
bfd_perform_slip (abfd, slip, input_section, value)
|
|
bfd *abfd;
|
|
unsigned int slip;
|
|
asection *input_section;
|
|
bfd_vma value;
|
|
{
|
|
asymbol **s;
|
|
|
|
s = _bfd_generic_link_get_symbols (abfd);
|
|
BFD_ASSERT (s != (asymbol **) NULL);
|
|
|
|
/* Find all symbols past this point, and make them know
|
|
what's happened. */
|
|
while (*s)
|
|
{
|
|
asymbol *p = *s;
|
|
if (p->section == input_section)
|
|
{
|
|
/* This was pointing into this section, so mangle it. */
|
|
if (p->value > value)
|
|
{
|
|
p->value -= slip;
|
|
if (p->udata.p != NULL)
|
|
{
|
|
struct generic_link_hash_entry *h;
|
|
|
|
h = (struct generic_link_hash_entry *) p->udata.p;
|
|
BFD_ASSERT (h->root.type == bfd_link_hash_defined
|
|
|| h->root.type == bfd_link_hash_defweak);
|
|
h->root.u.def.value -= slip;
|
|
BFD_ASSERT (h->root.u.def.value == p->value);
|
|
}
|
|
}
|
|
}
|
|
s++;
|
|
}
|
|
}
|
|
|
|
bfd_boolean
|
|
bfd_coff_reloc16_relax_section (abfd, input_section, link_info, again)
|
|
bfd *abfd;
|
|
asection *input_section;
|
|
struct bfd_link_info *link_info;
|
|
bfd_boolean *again;
|
|
{
|
|
/* Get enough memory to hold the stuff. */
|
|
bfd *input_bfd = input_section->owner;
|
|
unsigned *shrinks;
|
|
unsigned shrink = 0;
|
|
long reloc_size = bfd_get_reloc_upper_bound (input_bfd, input_section);
|
|
arelent **reloc_vector = NULL;
|
|
long reloc_count;
|
|
|
|
if (link_info->relocatable)
|
|
(*link_info->callbacks->einfo)
|
|
(_("%P%F: --relax and -r may not be used together\n"));
|
|
|
|
/* We only do global relaxation once. It is not safe to do it multiple
|
|
times (see discussion of the "shrinks" array below). */
|
|
*again = FALSE;
|
|
|
|
if (reloc_size < 0)
|
|
return FALSE;
|
|
|
|
reloc_vector = (arelent **) bfd_malloc ((bfd_size_type) reloc_size);
|
|
if (!reloc_vector && reloc_size > 0)
|
|
return FALSE;
|
|
|
|
/* Get the relocs and think about them. */
|
|
reloc_count =
|
|
bfd_canonicalize_reloc (input_bfd, input_section, reloc_vector,
|
|
_bfd_generic_link_get_symbols (input_bfd));
|
|
if (reloc_count < 0)
|
|
{
|
|
free (reloc_vector);
|
|
return FALSE;
|
|
}
|
|
|
|
/* The reloc16.c and related relaxing code is very simple, the price
|
|
for that simplicity is we can only call this function once for
|
|
each section.
|
|
|
|
So, to get the best results within that limitation, we do multiple
|
|
relaxing passes over each section here. That involves keeping track
|
|
of the "shrink" at each reloc in the section. This allows us to
|
|
accurately determine the relative location of two relocs within
|
|
this section.
|
|
|
|
In theory, if we kept the "shrinks" array for each section for the
|
|
entire link, we could use the generic relaxing code in the linker
|
|
and get better results, particularly for jsr->bsr and 24->16 bit
|
|
memory reference relaxations. */
|
|
|
|
if (reloc_count > 0)
|
|
{
|
|
int another_pass = 0;
|
|
bfd_size_type amt;
|
|
|
|
/* Allocate and initialize the shrinks array for this section.
|
|
The last element is used as an accumulator of shrinks. */
|
|
amt = reloc_count + 1;
|
|
amt *= sizeof (unsigned);
|
|
shrinks = (unsigned *) bfd_zmalloc (amt);
|
|
|
|
/* Loop until nothing changes in this section. */
|
|
do
|
|
{
|
|
arelent **parent;
|
|
unsigned int i;
|
|
long j;
|
|
|
|
another_pass = 0;
|
|
|
|
for (i = 0, parent = reloc_vector; *parent; parent++, i++)
|
|
{
|
|
/* Let the target/machine dependent code examine each reloc
|
|
in this section and attempt to shrink it. */
|
|
shrink = bfd_coff_reloc16_estimate (abfd, input_section, *parent,
|
|
shrinks[i], link_info);
|
|
|
|
/* If it shrunk, note it in the shrinks array and set up for
|
|
another pass. */
|
|
if (shrink != shrinks[i])
|
|
{
|
|
another_pass = 1;
|
|
for (j = i + 1; j <= reloc_count; j++)
|
|
shrinks[j] += shrink - shrinks[i];
|
|
}
|
|
}
|
|
}
|
|
while (another_pass);
|
|
|
|
shrink = shrinks[reloc_count];
|
|
free ((char *) shrinks);
|
|
}
|
|
|
|
input_section->rawsize = input_section->size;
|
|
input_section->size -= shrink;
|
|
free ((char *) reloc_vector);
|
|
return TRUE;
|
|
}
|
|
|
|
bfd_byte *
|
|
bfd_coff_reloc16_get_relocated_section_contents (in_abfd,
|
|
link_info,
|
|
link_order,
|
|
data,
|
|
relocatable,
|
|
symbols)
|
|
bfd *in_abfd;
|
|
struct bfd_link_info *link_info;
|
|
struct bfd_link_order *link_order;
|
|
bfd_byte *data;
|
|
bfd_boolean relocatable;
|
|
asymbol **symbols;
|
|
{
|
|
/* Get enough memory to hold the stuff. */
|
|
bfd *input_bfd = link_order->u.indirect.section->owner;
|
|
asection *input_section = link_order->u.indirect.section;
|
|
long reloc_size = bfd_get_reloc_upper_bound (input_bfd, input_section);
|
|
arelent **reloc_vector;
|
|
long reloc_count;
|
|
bfd_size_type sz;
|
|
|
|
if (reloc_size < 0)
|
|
return NULL;
|
|
|
|
/* If producing relocatable output, don't bother to relax. */
|
|
if (relocatable)
|
|
return bfd_generic_get_relocated_section_contents (in_abfd, link_info,
|
|
link_order,
|
|
data, relocatable,
|
|
symbols);
|
|
|
|
/* Read in the section. */
|
|
sz = input_section->rawsize ? input_section->rawsize : input_section->size;
|
|
if (!bfd_get_section_contents (input_bfd, input_section, data, 0, sz))
|
|
return NULL;
|
|
|
|
reloc_vector = (arelent **) bfd_malloc ((bfd_size_type) reloc_size);
|
|
if (!reloc_vector && reloc_size != 0)
|
|
return NULL;
|
|
|
|
reloc_count = bfd_canonicalize_reloc (input_bfd,
|
|
input_section,
|
|
reloc_vector,
|
|
symbols);
|
|
if (reloc_count < 0)
|
|
{
|
|
free (reloc_vector);
|
|
return NULL;
|
|
}
|
|
|
|
if (reloc_count > 0)
|
|
{
|
|
arelent **parent = reloc_vector;
|
|
arelent *reloc;
|
|
unsigned int dst_address = 0;
|
|
unsigned int src_address = 0;
|
|
unsigned int run;
|
|
unsigned int idx;
|
|
|
|
/* Find how long a run we can do. */
|
|
while (dst_address < link_order->size)
|
|
{
|
|
reloc = *parent;
|
|
if (reloc)
|
|
{
|
|
/* Note that the relaxing didn't tie up the addresses in the
|
|
relocation, so we use the original address to work out the
|
|
run of non-relocated data. */
|
|
run = reloc->address - src_address;
|
|
parent++;
|
|
}
|
|
else
|
|
{
|
|
run = link_order->size - dst_address;
|
|
}
|
|
|
|
/* Copy the bytes. */
|
|
for (idx = 0; idx < run; idx++)
|
|
data[dst_address++] = data[src_address++];
|
|
|
|
/* Now do the relocation. */
|
|
if (reloc)
|
|
{
|
|
bfd_coff_reloc16_extra_cases (input_bfd, link_info, link_order,
|
|
reloc, data, &src_address,
|
|
&dst_address);
|
|
}
|
|
}
|
|
}
|
|
free ((char *) reloc_vector);
|
|
return data;
|
|
}
|