1999-05-03 09:29:11 +02:00
|
|
|
/* Generic BFD support for file formats.
|
2019-01-01 11:31:27 +01:00
|
|
|
Copyright (C) 1990-2019 Free Software Foundation, Inc.
|
1999-05-03 09:29:11 +02:00
|
|
|
Written by Cygnus Support.
|
|
|
|
|
2003-02-14 12:16:09 +01:00
|
|
|
This file is part of BFD, the Binary File Descriptor library.
|
1999-05-03 09:29:11 +02:00
|
|
|
|
2003-02-14 12:16:09 +01:00
|
|
|
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
|
2007-07-03 16:26:43 +02:00
|
|
|
the Free Software Foundation; either version 3 of the License, or
|
2003-02-14 12:16:09 +01:00
|
|
|
(at your option) any later version.
|
1999-05-03 09:29:11 +02:00
|
|
|
|
2003-02-14 12:16:09 +01:00
|
|
|
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.
|
1999-05-03 09:29:11 +02:00
|
|
|
|
2003-02-14 12:16:09 +01:00
|
|
|
You should have received a copy of the GNU General Public License
|
|
|
|
along with this program; if not, write to the Free Software
|
2007-07-03 16:26:43 +02:00
|
|
|
Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston,
|
|
|
|
MA 02110-1301, USA. */
|
|
|
|
|
1999-05-03 09:29:11 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
SECTION
|
|
|
|
File formats
|
|
|
|
|
|
|
|
A format is a BFD concept of high level file contents type. The
|
2001-01-23 21:27:54 +01:00
|
|
|
formats supported by BFD are:
|
1999-05-03 09:29:11 +02:00
|
|
|
|
|
|
|
o <<bfd_object>>
|
|
|
|
|
|
|
|
The BFD may contain data, symbols, relocations and debug info.
|
|
|
|
|
|
|
|
o <<bfd_archive>>
|
|
|
|
|
|
|
|
The BFD contains other BFDs and an optional index.
|
|
|
|
|
|
|
|
o <<bfd_core>>
|
|
|
|
|
|
|
|
The BFD contains the result of an executable core dump.
|
|
|
|
|
2005-10-05 23:24:23 +02:00
|
|
|
SUBSECTION
|
|
|
|
File format functions
|
1999-05-03 09:29:11 +02:00
|
|
|
*/
|
|
|
|
|
|
|
|
#include "sysdep.h"
|
2007-04-26 16:47:00 +02:00
|
|
|
#include "bfd.h"
|
1999-05-03 09:29:11 +02:00
|
|
|
#include "libbfd.h"
|
|
|
|
|
|
|
|
/* IMPORT from targets.c. */
|
|
|
|
extern const size_t _bfd_target_vector_entries;
|
|
|
|
|
|
|
|
/*
|
|
|
|
FUNCTION
|
|
|
|
bfd_check_format
|
|
|
|
|
|
|
|
SYNOPSIS
|
2003-02-14 12:16:09 +01:00
|
|
|
bfd_boolean bfd_check_format (bfd *abfd, bfd_format format);
|
1999-05-03 09:29:11 +02:00
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
Verify if the file attached to the BFD @var{abfd} is compatible
|
|
|
|
with the format @var{format} (i.e., one of <<bfd_object>>,
|
|
|
|
<<bfd_archive>> or <<bfd_core>>).
|
|
|
|
|
|
|
|
If the BFD has been set to a specific target before the
|
|
|
|
call, only the named target and format combination is
|
|
|
|
checked. If the target has not been set, or has been set to
|
|
|
|
<<default>>, then all the known target backends is
|
|
|
|
interrogated to determine a match. If the default target
|
|
|
|
matches, it is used. If not, exactly one target must recognize
|
|
|
|
the file, or an error results.
|
|
|
|
|
2002-11-30 09:39:46 +01:00
|
|
|
The function returns <<TRUE>> on success, otherwise <<FALSE>>
|
2001-01-23 21:27:54 +01:00
|
|
|
with one of the following error codes:
|
1999-05-03 09:29:11 +02:00
|
|
|
|
|
|
|
o <<bfd_error_invalid_operation>> -
|
|
|
|
if <<format>> is not one of <<bfd_object>>, <<bfd_archive>> or
|
|
|
|
<<bfd_core>>.
|
|
|
|
|
|
|
|
o <<bfd_error_system_call>> -
|
|
|
|
if an error occured during a read - even some file mismatches
|
|
|
|
can cause bfd_error_system_calls.
|
|
|
|
|
|
|
|
o <<file_not_recognised>> -
|
|
|
|
none of the backends recognised the file format.
|
|
|
|
|
|
|
|
o <<bfd_error_file_ambiguously_recognized>> -
|
|
|
|
more than one backend recognised the file format.
|
|
|
|
*/
|
|
|
|
|
2002-11-30 09:39:46 +01:00
|
|
|
bfd_boolean
|
2003-06-29 12:06:40 +02:00
|
|
|
bfd_check_format (bfd *abfd, bfd_format format)
|
1999-05-03 09:29:11 +02:00
|
|
|
{
|
|
|
|
return bfd_check_format_matches (abfd, format, NULL);
|
|
|
|
}
|
|
|
|
|
2013-01-26 03:08:01 +01:00
|
|
|
struct bfd_preserve
|
|
|
|
{
|
|
|
|
void *marker;
|
|
|
|
void *tdata;
|
|
|
|
flagword flags;
|
|
|
|
const struct bfd_arch_info *arch_info;
|
|
|
|
struct bfd_section *sections;
|
|
|
|
struct bfd_section *section_last;
|
|
|
|
unsigned int section_count;
|
PR22458, failure to choose a matching ELF target
https://sourceware.org/ml/binutils/2013-05/msg00271.html was supposed
to banish "file format is ambiguous" errors for ELF. It didn't,
because the code supposedly detecting formats that implement
match_priority didn't work. That was due to not placing all matching
targets into the vector of matching targets. ELF objects should all
match the generic ELF target (priority 2), plus one or more machine
specific targets (priority 1), and perhaps a single machine specific
target with OS/ABI set (priority 0, best match). So the armel object
in the testcase actually matches elf32-littlearm,
elf32-littlearm-symbian, and elf32-littlearm-vxworks (all priority 1),
and elf32-little (priority 2). As the PR reported, elf32-little
wasn't seen as matching. Fixing that part of the problem wasn't too
difficult but matching the generic ELF target as well as the ARM ELF
targets resulted in ARM testsuite failures.
These proved to be the annoying reordering of stubs that occurs from
time to time due to the stub names containing the section id.
Matching another target causes more sections to be created in
elf_object_p. If section ids change, stub names change, which results
in different hashing and can therefore result in different hash table
traversal and stub creation order. That particular problem is fixed
by resetting section_id to the initial state before attempting each
target match, and taking a snapshot of its value after a successful
match.
PR 22458
* format.c (struct bfd_preserve): Add section_id.
(bfd_preserve_save, bfd_preserve_restore): Save and restore
_bfd_section_id.
(bfd_reinit): Set _bfd_section_id.
(bfd_check_format_matches): Put all matches of any priority into
matching_vector. Save initial section id and start each attempted
match at that section id.
* libbfd-in.h (_bfd_section_id): Declare.
* section.c (_bfd_section_id): Rename from section_id and make
global. Adjust uses.
(bfd_get_next_section_id): Delete.
* elf64-ppc.c (ppc64_elf_setup_section_lists): Replace use of
bfd_get_section_id with _bfd_section_id.
* libbfd.h: Regenerate.
* bfd-in2.h: Regenerate.
2018-05-16 04:03:48 +02:00
|
|
|
unsigned int section_id;
|
2013-01-26 03:08:01 +01:00
|
|
|
struct bfd_hash_table section_htab;
|
2016-09-14 14:49:16 +02:00
|
|
|
const struct bfd_build_id *build_id;
|
2013-01-26 03:08:01 +01:00
|
|
|
};
|
|
|
|
|
|
|
|
/* When testing an object for compatibility with a particular target
|
|
|
|
back-end, the back-end object_p function needs to set up certain
|
|
|
|
fields in the bfd on successfully recognizing the object. This
|
|
|
|
typically happens in a piecemeal fashion, with failures possible at
|
|
|
|
many points. On failure, the bfd is supposed to be restored to its
|
|
|
|
initial state, which is virtually impossible. However, restoring a
|
|
|
|
subset of the bfd state works in practice. This function stores
|
|
|
|
the subset. */
|
|
|
|
|
|
|
|
static bfd_boolean
|
|
|
|
bfd_preserve_save (bfd *abfd, struct bfd_preserve *preserve)
|
|
|
|
{
|
|
|
|
preserve->tdata = abfd->tdata.any;
|
|
|
|
preserve->arch_info = abfd->arch_info;
|
|
|
|
preserve->flags = abfd->flags;
|
|
|
|
preserve->sections = abfd->sections;
|
|
|
|
preserve->section_last = abfd->section_last;
|
|
|
|
preserve->section_count = abfd->section_count;
|
PR22458, failure to choose a matching ELF target
https://sourceware.org/ml/binutils/2013-05/msg00271.html was supposed
to banish "file format is ambiguous" errors for ELF. It didn't,
because the code supposedly detecting formats that implement
match_priority didn't work. That was due to not placing all matching
targets into the vector of matching targets. ELF objects should all
match the generic ELF target (priority 2), plus one or more machine
specific targets (priority 1), and perhaps a single machine specific
target with OS/ABI set (priority 0, best match). So the armel object
in the testcase actually matches elf32-littlearm,
elf32-littlearm-symbian, and elf32-littlearm-vxworks (all priority 1),
and elf32-little (priority 2). As the PR reported, elf32-little
wasn't seen as matching. Fixing that part of the problem wasn't too
difficult but matching the generic ELF target as well as the ARM ELF
targets resulted in ARM testsuite failures.
These proved to be the annoying reordering of stubs that occurs from
time to time due to the stub names containing the section id.
Matching another target causes more sections to be created in
elf_object_p. If section ids change, stub names change, which results
in different hashing and can therefore result in different hash table
traversal and stub creation order. That particular problem is fixed
by resetting section_id to the initial state before attempting each
target match, and taking a snapshot of its value after a successful
match.
PR 22458
* format.c (struct bfd_preserve): Add section_id.
(bfd_preserve_save, bfd_preserve_restore): Save and restore
_bfd_section_id.
(bfd_reinit): Set _bfd_section_id.
(bfd_check_format_matches): Put all matches of any priority into
matching_vector. Save initial section id and start each attempted
match at that section id.
* libbfd-in.h (_bfd_section_id): Declare.
* section.c (_bfd_section_id): Rename from section_id and make
global. Adjust uses.
(bfd_get_next_section_id): Delete.
* elf64-ppc.c (ppc64_elf_setup_section_lists): Replace use of
bfd_get_section_id with _bfd_section_id.
* libbfd.h: Regenerate.
* bfd-in2.h: Regenerate.
2018-05-16 04:03:48 +02:00
|
|
|
preserve->section_id = _bfd_section_id;
|
2013-01-26 03:08:01 +01:00
|
|
|
preserve->section_htab = abfd->section_htab;
|
|
|
|
preserve->marker = bfd_alloc (abfd, 1);
|
2016-09-14 14:49:16 +02:00
|
|
|
preserve->build_id = abfd->build_id;
|
2013-01-26 03:08:01 +01:00
|
|
|
if (preserve->marker == NULL)
|
|
|
|
return FALSE;
|
|
|
|
|
|
|
|
return bfd_hash_table_init (&abfd->section_htab, bfd_section_hash_newfunc,
|
|
|
|
sizeof (struct section_hash_entry));
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Clear out a subset of BFD state. */
|
|
|
|
|
|
|
|
static void
|
PR22458, failure to choose a matching ELF target
https://sourceware.org/ml/binutils/2013-05/msg00271.html was supposed
to banish "file format is ambiguous" errors for ELF. It didn't,
because the code supposedly detecting formats that implement
match_priority didn't work. That was due to not placing all matching
targets into the vector of matching targets. ELF objects should all
match the generic ELF target (priority 2), plus one or more machine
specific targets (priority 1), and perhaps a single machine specific
target with OS/ABI set (priority 0, best match). So the armel object
in the testcase actually matches elf32-littlearm,
elf32-littlearm-symbian, and elf32-littlearm-vxworks (all priority 1),
and elf32-little (priority 2). As the PR reported, elf32-little
wasn't seen as matching. Fixing that part of the problem wasn't too
difficult but matching the generic ELF target as well as the ARM ELF
targets resulted in ARM testsuite failures.
These proved to be the annoying reordering of stubs that occurs from
time to time due to the stub names containing the section id.
Matching another target causes more sections to be created in
elf_object_p. If section ids change, stub names change, which results
in different hashing and can therefore result in different hash table
traversal and stub creation order. That particular problem is fixed
by resetting section_id to the initial state before attempting each
target match, and taking a snapshot of its value after a successful
match.
PR 22458
* format.c (struct bfd_preserve): Add section_id.
(bfd_preserve_save, bfd_preserve_restore): Save and restore
_bfd_section_id.
(bfd_reinit): Set _bfd_section_id.
(bfd_check_format_matches): Put all matches of any priority into
matching_vector. Save initial section id and start each attempted
match at that section id.
* libbfd-in.h (_bfd_section_id): Declare.
* section.c (_bfd_section_id): Rename from section_id and make
global. Adjust uses.
(bfd_get_next_section_id): Delete.
* elf64-ppc.c (ppc64_elf_setup_section_lists): Replace use of
bfd_get_section_id with _bfd_section_id.
* libbfd.h: Regenerate.
* bfd-in2.h: Regenerate.
2018-05-16 04:03:48 +02:00
|
|
|
bfd_reinit (bfd *abfd, unsigned int section_id)
|
2013-01-26 03:08:01 +01:00
|
|
|
{
|
|
|
|
abfd->tdata.any = NULL;
|
|
|
|
abfd->arch_info = &bfd_default_arch_struct;
|
|
|
|
abfd->flags &= BFD_FLAGS_SAVED;
|
|
|
|
bfd_section_list_clear (abfd);
|
PR22458, failure to choose a matching ELF target
https://sourceware.org/ml/binutils/2013-05/msg00271.html was supposed
to banish "file format is ambiguous" errors for ELF. It didn't,
because the code supposedly detecting formats that implement
match_priority didn't work. That was due to not placing all matching
targets into the vector of matching targets. ELF objects should all
match the generic ELF target (priority 2), plus one or more machine
specific targets (priority 1), and perhaps a single machine specific
target with OS/ABI set (priority 0, best match). So the armel object
in the testcase actually matches elf32-littlearm,
elf32-littlearm-symbian, and elf32-littlearm-vxworks (all priority 1),
and elf32-little (priority 2). As the PR reported, elf32-little
wasn't seen as matching. Fixing that part of the problem wasn't too
difficult but matching the generic ELF target as well as the ARM ELF
targets resulted in ARM testsuite failures.
These proved to be the annoying reordering of stubs that occurs from
time to time due to the stub names containing the section id.
Matching another target causes more sections to be created in
elf_object_p. If section ids change, stub names change, which results
in different hashing and can therefore result in different hash table
traversal and stub creation order. That particular problem is fixed
by resetting section_id to the initial state before attempting each
target match, and taking a snapshot of its value after a successful
match.
PR 22458
* format.c (struct bfd_preserve): Add section_id.
(bfd_preserve_save, bfd_preserve_restore): Save and restore
_bfd_section_id.
(bfd_reinit): Set _bfd_section_id.
(bfd_check_format_matches): Put all matches of any priority into
matching_vector. Save initial section id and start each attempted
match at that section id.
* libbfd-in.h (_bfd_section_id): Declare.
* section.c (_bfd_section_id): Rename from section_id and make
global. Adjust uses.
(bfd_get_next_section_id): Delete.
* elf64-ppc.c (ppc64_elf_setup_section_lists): Replace use of
bfd_get_section_id with _bfd_section_id.
* libbfd.h: Regenerate.
* bfd-in2.h: Regenerate.
2018-05-16 04:03:48 +02:00
|
|
|
_bfd_section_id = section_id;
|
2013-01-26 03:08:01 +01:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Restores bfd state saved by bfd_preserve_save. */
|
|
|
|
|
|
|
|
static void
|
|
|
|
bfd_preserve_restore (bfd *abfd, struct bfd_preserve *preserve)
|
|
|
|
{
|
|
|
|
bfd_hash_table_free (&abfd->section_htab);
|
|
|
|
|
|
|
|
abfd->tdata.any = preserve->tdata;
|
|
|
|
abfd->arch_info = preserve->arch_info;
|
|
|
|
abfd->flags = preserve->flags;
|
|
|
|
abfd->section_htab = preserve->section_htab;
|
|
|
|
abfd->sections = preserve->sections;
|
|
|
|
abfd->section_last = preserve->section_last;
|
|
|
|
abfd->section_count = preserve->section_count;
|
PR22458, failure to choose a matching ELF target
https://sourceware.org/ml/binutils/2013-05/msg00271.html was supposed
to banish "file format is ambiguous" errors for ELF. It didn't,
because the code supposedly detecting formats that implement
match_priority didn't work. That was due to not placing all matching
targets into the vector of matching targets. ELF objects should all
match the generic ELF target (priority 2), plus one or more machine
specific targets (priority 1), and perhaps a single machine specific
target with OS/ABI set (priority 0, best match). So the armel object
in the testcase actually matches elf32-littlearm,
elf32-littlearm-symbian, and elf32-littlearm-vxworks (all priority 1),
and elf32-little (priority 2). As the PR reported, elf32-little
wasn't seen as matching. Fixing that part of the problem wasn't too
difficult but matching the generic ELF target as well as the ARM ELF
targets resulted in ARM testsuite failures.
These proved to be the annoying reordering of stubs that occurs from
time to time due to the stub names containing the section id.
Matching another target causes more sections to be created in
elf_object_p. If section ids change, stub names change, which results
in different hashing and can therefore result in different hash table
traversal and stub creation order. That particular problem is fixed
by resetting section_id to the initial state before attempting each
target match, and taking a snapshot of its value after a successful
match.
PR 22458
* format.c (struct bfd_preserve): Add section_id.
(bfd_preserve_save, bfd_preserve_restore): Save and restore
_bfd_section_id.
(bfd_reinit): Set _bfd_section_id.
(bfd_check_format_matches): Put all matches of any priority into
matching_vector. Save initial section id and start each attempted
match at that section id.
* libbfd-in.h (_bfd_section_id): Declare.
* section.c (_bfd_section_id): Rename from section_id and make
global. Adjust uses.
(bfd_get_next_section_id): Delete.
* elf64-ppc.c (ppc64_elf_setup_section_lists): Replace use of
bfd_get_section_id with _bfd_section_id.
* libbfd.h: Regenerate.
* bfd-in2.h: Regenerate.
2018-05-16 04:03:48 +02:00
|
|
|
_bfd_section_id = preserve->section_id;
|
2016-09-14 14:49:16 +02:00
|
|
|
abfd->build_id = preserve->build_id;
|
2013-01-26 03:08:01 +01:00
|
|
|
|
|
|
|
/* bfd_release frees all memory more recently bfd_alloc'd than
|
|
|
|
its arg, as well as its arg. */
|
|
|
|
bfd_release (abfd, preserve->marker);
|
|
|
|
preserve->marker = NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Called when the bfd state saved by bfd_preserve_save is no longer
|
|
|
|
needed. */
|
|
|
|
|
|
|
|
static void
|
|
|
|
bfd_preserve_finish (bfd *abfd ATTRIBUTE_UNUSED, struct bfd_preserve *preserve)
|
|
|
|
{
|
|
|
|
/* It would be nice to be able to free more memory here, eg. old
|
|
|
|
tdata, but that's not possible since these blocks are sitting
|
|
|
|
inside bfd_alloc'd memory. The section hash is on a separate
|
|
|
|
objalloc. */
|
|
|
|
bfd_hash_table_free (&preserve->section_htab);
|
|
|
|
preserve->marker = NULL;
|
|
|
|
}
|
|
|
|
|
1999-05-03 09:29:11 +02:00
|
|
|
/*
|
|
|
|
FUNCTION
|
|
|
|
bfd_check_format_matches
|
|
|
|
|
|
|
|
SYNOPSIS
|
2003-06-29 12:06:40 +02:00
|
|
|
bfd_boolean bfd_check_format_matches
|
|
|
|
(bfd *abfd, bfd_format format, char ***matching);
|
1999-05-03 09:29:11 +02:00
|
|
|
|
|
|
|
DESCRIPTION
|
2002-11-30 09:39:46 +01:00
|
|
|
Like <<bfd_check_format>>, except when it returns FALSE with
|
1999-05-03 09:29:11 +02:00
|
|
|
<<bfd_errno>> set to <<bfd_error_file_ambiguously_recognized>>. In that
|
|
|
|
case, if @var{matching} is not NULL, it will be filled in with
|
|
|
|
a NULL-terminated list of the names of the formats that matched,
|
|
|
|
allocated with <<malloc>>.
|
|
|
|
Then the user may choose a format and try again.
|
|
|
|
|
|
|
|
When done with the list that @var{matching} points to, the caller
|
2001-01-23 21:27:54 +01:00
|
|
|
should free it.
|
1999-05-03 09:29:11 +02:00
|
|
|
*/
|
|
|
|
|
2002-11-30 09:39:46 +01:00
|
|
|
bfd_boolean
|
2003-06-29 12:06:40 +02:00
|
|
|
bfd_check_format_matches (bfd *abfd, bfd_format format, char ***matching)
|
1999-05-03 09:29:11 +02:00
|
|
|
{
|
|
|
|
extern const bfd_target binary_vec;
|
Fix BFD format matching for x86_64-w64-mingw32 -m32 LTO.
In <https://sourceware.org/ml/binutils/2015-12/msg00190.html> (commit
4a07dc81356ed8728e204e9aabeb256703c59aef), Kwok fixed a problem with
the template used for a dummy BFD for an IR file for LTO on MinGW,
where the input and output formats are not the same.
A problem, however, remains in the case of linking for
x86_64-w64-mingw32 -m32, where LTO linking reports an ambiguity
between the pe-i386 and pei-i386 formats. An object (pe-i386) with
plugin data is being tested by the linker to see what formats match.
The default format initially set by the linker when
bfd_check_format_matches is called is pei-i386 (as that's the output
format from the linker script), which does not match, so the function
goes on to the loop over possible BFD vectors. The pe-i386 vector
matches, as it should. One other vector matches: the plugin vector.
bfd_check_format_matches tests a vector for matching by temporarily
modifying the BFD object to use that vector then using
_bfd_check_format on it. So the BFD object is temporarily using
plugin_vec. _bfd_check_format ends up using bfd_plugin_object_p which
uses plugin_object_p from ld which uses plugin_get_ir_dummy_bfd which
succeeds, having created a BFD based on link_info.output_bfd (because
srctemplate is the BFD temporarily using plugin_vec, even after Kwok's
patch link_info.output_bfd is all that's available to base the dummy
BFD on). So we end up with a match from the plugin vector which uses
the pei-i386 vector even though the pei-i386 vector itself does not
match the input object. (In the i686-mingw32 case, as opposed to this
multilib case, pe-i386 is the default BFD target, which would
short-circuit that logic.)
There are two cases of the linker handling inputs with a plugin: they
may be inputs that are also accepted by some non-plugin BFD format, as
here, or they may be a format that would not be recognized at all, as
with some tests in the ld testsuite. In the former case, there is no
need for BFD to accept the objects using the plugin vector, as the
linker has its own logic to allow plugins to claim objects accepted by
some other BFD vector. Thus, this patch arranges for the plugin
vector to have the lowest match priority, and for the priority from
that vector to be used in the relevant case (the attempted match to
the plugin vector results in TEMP pointing to the pei-i386 vector).
Tested for GCC and Binutils testsuites for x86_64-pc-linux-gnu, as
well as verifying that it fixes the observed LTO issue for
x86_64-w64-mingw32.
* plugin.c (plugin_vec): Set match priority to 255.
* format.c (bfd_check_format_matches) [BFD_SUPPORTS_PLUGINS]: When
matching against the plugin vector, take priority from there not
from TEMP.
2016-02-16 01:27:11 +01:00
|
|
|
#if BFD_SUPPORTS_PLUGINS
|
|
|
|
extern const bfd_target plugin_vec;
|
|
|
|
#endif
|
2002-11-05 02:18:56 +01:00
|
|
|
const bfd_target * const *target;
|
|
|
|
const bfd_target **matching_vector = NULL;
|
* targets.c (bfd_target): Make ar_max_namelen an unsigned char.
Add match_priority.
* configure.in: Bump bfd version.
* elfcode.h (elf_object_p): Delete hacks preventing match of
EM_NONE and ELFOSABI_NONE targets when a better match exists.
* elfxx-target.h (elf_match_priority): Define and use.
* format.c (bfd_check_format_matches): Use target match_priority
to choose best of multiple matching targets. In cases with multiple
matches rerun _bfd_check_format if we don't choose the last match.
* aout-adobe.c, * aout-arm.c, * aout-target.h, * aout-tic30.c,
* binary.c, * bout.c, * coff-alpha.c, * coff-i386.c, * coff-i860.c,
* coff-i960.c, * coff-ia64.c, * coff-mips.c, * coff-or32.c,
* coff-ppc.c, * coff-rs6000.c, * coff-sh.c, * coff-tic30.c,
* coff-tic54x.c, * coff-x86_64.c, * coff64-rs6000.c, * coffcode.h,
* i386msdos.c, * i386os9k.c, * ieee.c, * ihex.c, * mach-o-target.c,
* mipsbsd.c, * mmo.c, * nlm-target.h, * oasys.c, * pdp11.c,
* pe-mips.c, * pef.c, * plugin.c, * ppcboot.c, * som.c, * srec.c,
* tekhex.c, * trad-core.c, * verilog.c, * versados.c, * vms-alpha.c,
* vms-lib.c, * xsym.c: Init match_priority field.
* configure: Regenerate.
* bfd-in2.h: Regenerate.
2011-06-06 03:26:05 +02:00
|
|
|
const bfd_target *save_targ, *right_targ, *ar_right_targ, *match_targ;
|
|
|
|
int match_count, best_count, best_match;
|
2001-08-17 05:19:01 +02:00
|
|
|
int ar_match_index;
|
PR22458, failure to choose a matching ELF target
https://sourceware.org/ml/binutils/2013-05/msg00271.html was supposed
to banish "file format is ambiguous" errors for ELF. It didn't,
because the code supposedly detecting formats that implement
match_priority didn't work. That was due to not placing all matching
targets into the vector of matching targets. ELF objects should all
match the generic ELF target (priority 2), plus one or more machine
specific targets (priority 1), and perhaps a single machine specific
target with OS/ABI set (priority 0, best match). So the armel object
in the testcase actually matches elf32-littlearm,
elf32-littlearm-symbian, and elf32-littlearm-vxworks (all priority 1),
and elf32-little (priority 2). As the PR reported, elf32-little
wasn't seen as matching. Fixing that part of the problem wasn't too
difficult but matching the generic ELF target as well as the ARM ELF
targets resulted in ARM testsuite failures.
These proved to be the annoying reordering of stubs that occurs from
time to time due to the stub names containing the section id.
Matching another target causes more sections to be created in
elf_object_p. If section ids change, stub names change, which results
in different hashing and can therefore result in different hash table
traversal and stub creation order. That particular problem is fixed
by resetting section_id to the initial state before attempting each
target match, and taking a snapshot of its value after a successful
match.
PR 22458
* format.c (struct bfd_preserve): Add section_id.
(bfd_preserve_save, bfd_preserve_restore): Save and restore
_bfd_section_id.
(bfd_reinit): Set _bfd_section_id.
(bfd_check_format_matches): Put all matches of any priority into
matching_vector. Save initial section id and start each attempted
match at that section id.
* libbfd-in.h (_bfd_section_id): Declare.
* section.c (_bfd_section_id): Rename from section_id and make
global. Adjust uses.
(bfd_get_next_section_id): Delete.
* elf64-ppc.c (ppc64_elf_setup_section_lists): Replace use of
bfd_get_section_id with _bfd_section_id.
* libbfd.h: Regenerate.
* bfd-in2.h: Regenerate.
2018-05-16 04:03:48 +02:00
|
|
|
unsigned int initial_section_id = _bfd_section_id;
|
2013-01-26 03:08:01 +01:00
|
|
|
struct bfd_preserve preserve;
|
1999-05-03 09:29:11 +02:00
|
|
|
|
2007-09-14 07:19:27 +02:00
|
|
|
if (matching != NULL)
|
|
|
|
*matching = NULL;
|
|
|
|
|
2001-08-17 05:19:01 +02:00
|
|
|
if (!bfd_read_p (abfd)
|
2001-08-17 10:57:42 +02:00
|
|
|
|| (unsigned int) abfd->format >= (unsigned int) bfd_type_end)
|
2000-07-20 02:31:39 +02:00
|
|
|
{
|
|
|
|
bfd_set_error (bfd_error_invalid_operation);
|
2002-11-30 09:39:46 +01:00
|
|
|
return FALSE;
|
2000-07-20 02:31:39 +02:00
|
|
|
}
|
1999-05-03 09:29:11 +02:00
|
|
|
|
|
|
|
if (abfd->format != bfd_unknown)
|
2002-11-30 09:39:46 +01:00
|
|
|
return abfd->format == format;
|
1999-05-03 09:29:11 +02:00
|
|
|
|
2007-09-14 07:19:27 +02:00
|
|
|
if (matching != NULL || *bfd_associated_vector != NULL)
|
1999-05-03 09:29:11 +02:00
|
|
|
{
|
Touches most files in bfd/, so likely will be blamed for everything..
o bfd_read and bfd_write lose an unnecessary param and become
bfd_bread and bfd_bwrite.
o bfd_*alloc now all take a bfd_size_type arg, and will error if
size_t is too small. eg. 32 bit host, 64 bit bfd, verrry big files
or bugs in linker scripts etc.
o file_ptr becomes a bfd_signed_vma. Besides matching sizes with
various other types involved in handling sections, this should make
it easier for bfd to support a 64 bit off_t on 32 bit hosts that
provide it.
o I've made the H_GET_* and H_PUT_* macros (which invoke bfd_h_{get,put}_*)
generally available. They now cast their args to bfd_vma and
bfd_byte * as appropriate, which removes a swag of casts from the
source.
o Bug fixes to bfd_get8, aix386_core_vec, elf32_h8_relax_section, and
aout-encap.c.
o Zillions of formatting and -Wconversion fixes.
2001-09-18 11:57:26 +02:00
|
|
|
bfd_size_type amt;
|
|
|
|
|
2002-11-05 02:18:56 +01:00
|
|
|
amt = sizeof (*matching_vector) * 2 * _bfd_target_vector_entries;
|
2009-09-09 23:38:59 +02:00
|
|
|
matching_vector = (const bfd_target **) bfd_malloc (amt);
|
1999-05-03 09:29:11 +02:00
|
|
|
if (!matching_vector)
|
2002-11-30 09:39:46 +01:00
|
|
|
return FALSE;
|
1999-05-03 09:29:11 +02:00
|
|
|
}
|
2001-01-23 21:27:54 +01:00
|
|
|
|
2000-07-20 02:31:39 +02:00
|
|
|
/* Presume the answer is yes. */
|
1999-05-03 09:29:11 +02:00
|
|
|
abfd->format = format;
|
2013-01-26 03:08:01 +01:00
|
|
|
save_targ = abfd->xvec;
|
|
|
|
preserve.marker = NULL;
|
1999-05-03 09:29:11 +02:00
|
|
|
|
|
|
|
/* If the target type was explicitly specified, just check that target. */
|
2000-07-20 02:31:39 +02:00
|
|
|
if (!abfd->target_defaulted)
|
|
|
|
{
|
Touches most files in bfd/, so likely will be blamed for everything..
o bfd_read and bfd_write lose an unnecessary param and become
bfd_bread and bfd_bwrite.
o bfd_*alloc now all take a bfd_size_type arg, and will error if
size_t is too small. eg. 32 bit host, 64 bit bfd, verrry big files
or bugs in linker scripts etc.
o file_ptr becomes a bfd_signed_vma. Besides matching sizes with
various other types involved in handling sections, this should make
it easier for bfd to support a 64 bit off_t on 32 bit hosts that
provide it.
o I've made the H_GET_* and H_PUT_* macros (which invoke bfd_h_{get,put}_*)
generally available. They now cast their args to bfd_vma and
bfd_byte * as appropriate, which removes a swag of casts from the
source.
o Bug fixes to bfd_get8, aix386_core_vec, elf32_h8_relax_section, and
aout-encap.c.
o Zillions of formatting and -Wconversion fixes.
2001-09-18 11:57:26 +02:00
|
|
|
if (bfd_seek (abfd, (file_ptr) 0, SEEK_SET) != 0) /* rewind! */
|
2007-09-14 07:19:27 +02:00
|
|
|
goto err_ret;
|
2001-01-23 21:27:54 +01:00
|
|
|
|
2000-07-20 02:31:39 +02:00
|
|
|
right_targ = BFD_SEND_FMT (abfd, _bfd_check_format, (abfd));
|
2001-01-23 21:27:54 +01:00
|
|
|
|
2000-07-20 02:31:39 +02:00
|
|
|
if (right_targ)
|
2007-09-14 07:19:27 +02:00
|
|
|
goto ok_ret;
|
2000-07-20 02:31:39 +02:00
|
|
|
|
|
|
|
/* For a long time the code has dropped through to check all
|
|
|
|
targets if the specified target was wrong. I don't know why,
|
|
|
|
and I'm reluctant to change it. However, in the case of an
|
|
|
|
archive, it can cause problems. If the specified target does
|
|
|
|
not permit archives (e.g., the binary target), then we should
|
|
|
|
not allow some other target to recognize it as an archive, but
|
|
|
|
should instead allow the specified target to recognize it as an
|
|
|
|
object. When I first made this change, it broke the PE target,
|
|
|
|
because the specified pei-i386 target did not recognize the
|
|
|
|
actual pe-i386 archive. Since there may be other problems of
|
|
|
|
this sort, I changed this test to check only for the binary
|
|
|
|
target. */
|
|
|
|
if (format == bfd_archive && save_targ == &binary_vec)
|
2007-09-14 07:19:27 +02:00
|
|
|
goto err_unrecog;
|
2000-07-20 02:31:39 +02:00
|
|
|
}
|
|
|
|
|
2013-01-26 03:08:01 +01:00
|
|
|
/* Since the target type was defaulted, check them all in the hope
|
|
|
|
that one will be uniquely recognized. */
|
|
|
|
right_targ = NULL;
|
|
|
|
ar_right_targ = NULL;
|
|
|
|
match_targ = NULL;
|
|
|
|
best_match = 256;
|
|
|
|
best_count = 0;
|
|
|
|
match_count = 0;
|
|
|
|
ar_match_index = _bfd_target_vector_entries;
|
|
|
|
|
2000-07-20 02:31:39 +02:00
|
|
|
for (target = bfd_target_vector; *target != NULL; target++)
|
|
|
|
{
|
|
|
|
const bfd_target *temp;
|
2001-01-23 21:27:54 +01:00
|
|
|
|
2005-12-13 16:34:28 +01:00
|
|
|
/* Don't check the default target twice. */
|
|
|
|
if (*target == &binary_vec
|
PR22458, failure to choose a matching ELF target
https://sourceware.org/ml/binutils/2013-05/msg00271.html was supposed
to banish "file format is ambiguous" errors for ELF. It didn't,
because the code supposedly detecting formats that implement
match_priority didn't work. That was due to not placing all matching
targets into the vector of matching targets. ELF objects should all
match the generic ELF target (priority 2), plus one or more machine
specific targets (priority 1), and perhaps a single machine specific
target with OS/ABI set (priority 0, best match). So the armel object
in the testcase actually matches elf32-littlearm,
elf32-littlearm-symbian, and elf32-littlearm-vxworks (all priority 1),
and elf32-little (priority 2). As the PR reported, elf32-little
wasn't seen as matching. Fixing that part of the problem wasn't too
difficult but matching the generic ELF target as well as the ARM ELF
targets resulted in ARM testsuite failures.
These proved to be the annoying reordering of stubs that occurs from
time to time due to the stub names containing the section id.
Matching another target causes more sections to be created in
elf_object_p. If section ids change, stub names change, which results
in different hashing and can therefore result in different hash table
traversal and stub creation order. That particular problem is fixed
by resetting section_id to the initial state before attempting each
target match, and taking a snapshot of its value after a successful
match.
PR 22458
* format.c (struct bfd_preserve): Add section_id.
(bfd_preserve_save, bfd_preserve_restore): Save and restore
_bfd_section_id.
(bfd_reinit): Set _bfd_section_id.
(bfd_check_format_matches): Put all matches of any priority into
matching_vector. Save initial section id and start each attempted
match at that section id.
* libbfd-in.h (_bfd_section_id): Declare.
* section.c (_bfd_section_id): Rename from section_id and make
global. Adjust uses.
(bfd_get_next_section_id): Delete.
* elf64-ppc.c (ppc64_elf_setup_section_lists): Replace use of
bfd_get_section_id with _bfd_section_id.
* libbfd.h: Regenerate.
* bfd-in2.h: Regenerate.
2018-05-16 04:03:48 +02:00
|
|
|
|| (!abfd->target_defaulted && *target == save_targ))
|
2000-07-20 02:31:39 +02:00
|
|
|
continue;
|
2001-01-23 21:27:54 +01:00
|
|
|
|
2013-01-26 03:08:01 +01:00
|
|
|
/* If we already tried a match, the bfd is modified and may
|
|
|
|
have sections attached, which will confuse the next
|
|
|
|
_bfd_check_format call. */
|
PR22458, failure to choose a matching ELF target
https://sourceware.org/ml/binutils/2013-05/msg00271.html was supposed
to banish "file format is ambiguous" errors for ELF. It didn't,
because the code supposedly detecting formats that implement
match_priority didn't work. That was due to not placing all matching
targets into the vector of matching targets. ELF objects should all
match the generic ELF target (priority 2), plus one or more machine
specific targets (priority 1), and perhaps a single machine specific
target with OS/ABI set (priority 0, best match). So the armel object
in the testcase actually matches elf32-littlearm,
elf32-littlearm-symbian, and elf32-littlearm-vxworks (all priority 1),
and elf32-little (priority 2). As the PR reported, elf32-little
wasn't seen as matching. Fixing that part of the problem wasn't too
difficult but matching the generic ELF target as well as the ARM ELF
targets resulted in ARM testsuite failures.
These proved to be the annoying reordering of stubs that occurs from
time to time due to the stub names containing the section id.
Matching another target causes more sections to be created in
elf_object_p. If section ids change, stub names change, which results
in different hashing and can therefore result in different hash table
traversal and stub creation order. That particular problem is fixed
by resetting section_id to the initial state before attempting each
target match, and taking a snapshot of its value after a successful
match.
PR 22458
* format.c (struct bfd_preserve): Add section_id.
(bfd_preserve_save, bfd_preserve_restore): Save and restore
_bfd_section_id.
(bfd_reinit): Set _bfd_section_id.
(bfd_check_format_matches): Put all matches of any priority into
matching_vector. Save initial section id and start each attempted
match at that section id.
* libbfd-in.h (_bfd_section_id): Declare.
* section.c (_bfd_section_id): Rename from section_id and make
global. Adjust uses.
(bfd_get_next_section_id): Delete.
* elf64-ppc.c (ppc64_elf_setup_section_lists): Replace use of
bfd_get_section_id with _bfd_section_id.
* libbfd.h: Regenerate.
* bfd-in2.h: Regenerate.
2018-05-16 04:03:48 +02:00
|
|
|
bfd_reinit (abfd, initial_section_id);
|
2013-01-26 03:08:01 +01:00
|
|
|
|
|
|
|
/* Change BFD's target temporarily. */
|
|
|
|
abfd->xvec = *target;
|
2001-01-23 21:27:54 +01:00
|
|
|
|
Touches most files in bfd/, so likely will be blamed for everything..
o bfd_read and bfd_write lose an unnecessary param and become
bfd_bread and bfd_bwrite.
o bfd_*alloc now all take a bfd_size_type arg, and will error if
size_t is too small. eg. 32 bit host, 64 bit bfd, verrry big files
or bugs in linker scripts etc.
o file_ptr becomes a bfd_signed_vma. Besides matching sizes with
various other types involved in handling sections, this should make
it easier for bfd to support a 64 bit off_t on 32 bit hosts that
provide it.
o I've made the H_GET_* and H_PUT_* macros (which invoke bfd_h_{get,put}_*)
generally available. They now cast their args to bfd_vma and
bfd_byte * as appropriate, which removes a swag of casts from the
source.
o Bug fixes to bfd_get8, aix386_core_vec, elf32_h8_relax_section, and
aout-encap.c.
o Zillions of formatting and -Wconversion fixes.
2001-09-18 11:57:26 +02:00
|
|
|
if (bfd_seek (abfd, (file_ptr) 0, SEEK_SET) != 0)
|
2007-09-14 07:19:27 +02:00
|
|
|
goto err_ret;
|
2001-01-23 21:27:54 +01:00
|
|
|
|
2000-07-20 02:31:39 +02:00
|
|
|
/* If _bfd_check_format neglects to set bfd_error, assume
|
|
|
|
bfd_error_wrong_format. We didn't used to even pay any
|
|
|
|
attention to bfd_error, so I suspect that some
|
|
|
|
_bfd_check_format might have this problem. */
|
|
|
|
bfd_set_error (bfd_error_wrong_format);
|
|
|
|
|
|
|
|
temp = BFD_SEND_FMT (abfd, _bfd_check_format, (abfd));
|
* targets.c (bfd_target): Make ar_max_namelen an unsigned char.
Add match_priority.
* configure.in: Bump bfd version.
* elfcode.h (elf_object_p): Delete hacks preventing match of
EM_NONE and ELFOSABI_NONE targets when a better match exists.
* elfxx-target.h (elf_match_priority): Define and use.
* format.c (bfd_check_format_matches): Use target match_priority
to choose best of multiple matching targets. In cases with multiple
matches rerun _bfd_check_format if we don't choose the last match.
* aout-adobe.c, * aout-arm.c, * aout-target.h, * aout-tic30.c,
* binary.c, * bout.c, * coff-alpha.c, * coff-i386.c, * coff-i860.c,
* coff-i960.c, * coff-ia64.c, * coff-mips.c, * coff-or32.c,
* coff-ppc.c, * coff-rs6000.c, * coff-sh.c, * coff-tic30.c,
* coff-tic54x.c, * coff-x86_64.c, * coff64-rs6000.c, * coffcode.h,
* i386msdos.c, * i386os9k.c, * ieee.c, * ihex.c, * mach-o-target.c,
* mipsbsd.c, * mmo.c, * nlm-target.h, * oasys.c, * pdp11.c,
* pe-mips.c, * pef.c, * plugin.c, * ppcboot.c, * som.c, * srec.c,
* tekhex.c, * trad-core.c, * verilog.c, * versados.c, * vms-alpha.c,
* vms-lib.c, * xsym.c: Init match_priority field.
* configure: Regenerate.
* bfd-in2.h: Regenerate.
2011-06-06 03:26:05 +02:00
|
|
|
if (temp)
|
2013-01-26 03:08:01 +01:00
|
|
|
{
|
Fix BFD format matching for x86_64-w64-mingw32 -m32 LTO.
In <https://sourceware.org/ml/binutils/2015-12/msg00190.html> (commit
4a07dc81356ed8728e204e9aabeb256703c59aef), Kwok fixed a problem with
the template used for a dummy BFD for an IR file for LTO on MinGW,
where the input and output formats are not the same.
A problem, however, remains in the case of linking for
x86_64-w64-mingw32 -m32, where LTO linking reports an ambiguity
between the pe-i386 and pei-i386 formats. An object (pe-i386) with
plugin data is being tested by the linker to see what formats match.
The default format initially set by the linker when
bfd_check_format_matches is called is pei-i386 (as that's the output
format from the linker script), which does not match, so the function
goes on to the loop over possible BFD vectors. The pe-i386 vector
matches, as it should. One other vector matches: the plugin vector.
bfd_check_format_matches tests a vector for matching by temporarily
modifying the BFD object to use that vector then using
_bfd_check_format on it. So the BFD object is temporarily using
plugin_vec. _bfd_check_format ends up using bfd_plugin_object_p which
uses plugin_object_p from ld which uses plugin_get_ir_dummy_bfd which
succeeds, having created a BFD based on link_info.output_bfd (because
srctemplate is the BFD temporarily using plugin_vec, even after Kwok's
patch link_info.output_bfd is all that's available to base the dummy
BFD on). So we end up with a match from the plugin vector which uses
the pei-i386 vector even though the pei-i386 vector itself does not
match the input object. (In the i686-mingw32 case, as opposed to this
multilib case, pe-i386 is the default BFD target, which would
short-circuit that logic.)
There are two cases of the linker handling inputs with a plugin: they
may be inputs that are also accepted by some non-plugin BFD format, as
here, or they may be a format that would not be recognized at all, as
with some tests in the ld testsuite. In the former case, there is no
need for BFD to accept the objects using the plugin vector, as the
linker has its own logic to allow plugins to claim objects accepted by
some other BFD vector. Thus, this patch arranges for the plugin
vector to have the lowest match priority, and for the priority from
that vector to be used in the relevant case (the attempted match to
the plugin vector results in TEMP pointing to the pei-i386 vector).
Tested for GCC and Binutils testsuites for x86_64-pc-linux-gnu, as
well as verifying that it fixes the observed LTO issue for
x86_64-w64-mingw32.
* plugin.c (plugin_vec): Set match priority to 255.
* format.c (bfd_check_format_matches) [BFD_SUPPORTS_PLUGINS]: When
matching against the plugin vector, take priority from there not
from TEMP.
2016-02-16 01:27:11 +01:00
|
|
|
int match_priority = temp->match_priority;
|
|
|
|
#if BFD_SUPPORTS_PLUGINS
|
|
|
|
/* If this object can be handled by a plugin, give that the
|
|
|
|
lowest priority; objects both handled by a plugin and
|
|
|
|
with an underlying object format will be claimed
|
|
|
|
separately by the plugin. */
|
|
|
|
if (*target == &plugin_vec)
|
|
|
|
match_priority = (*target)->match_priority;
|
|
|
|
#endif
|
|
|
|
|
2013-01-26 03:08:01 +01:00
|
|
|
match_targ = temp;
|
|
|
|
if (preserve.marker != NULL)
|
|
|
|
bfd_preserve_finish (abfd, &preserve);
|
2001-01-23 21:27:54 +01:00
|
|
|
|
2013-01-28 08:58:16 +01:00
|
|
|
if (abfd->format != bfd_archive
|
|
|
|
|| (bfd_has_map (abfd)
|
|
|
|
&& bfd_get_error () != bfd_error_wrong_object_format))
|
* targets.c (bfd_target): Make ar_max_namelen an unsigned char.
Add match_priority.
* configure.in: Bump bfd version.
* elfcode.h (elf_object_p): Delete hacks preventing match of
EM_NONE and ELFOSABI_NONE targets when a better match exists.
* elfxx-target.h (elf_match_priority): Define and use.
* format.c (bfd_check_format_matches): Use target match_priority
to choose best of multiple matching targets. In cases with multiple
matches rerun _bfd_check_format if we don't choose the last match.
* aout-adobe.c, * aout-arm.c, * aout-target.h, * aout-tic30.c,
* binary.c, * bout.c, * coff-alpha.c, * coff-i386.c, * coff-i860.c,
* coff-i960.c, * coff-ia64.c, * coff-mips.c, * coff-or32.c,
* coff-ppc.c, * coff-rs6000.c, * coff-sh.c, * coff-tic30.c,
* coff-tic54x.c, * coff-x86_64.c, * coff64-rs6000.c, * coffcode.h,
* i386msdos.c, * i386os9k.c, * ieee.c, * ihex.c, * mach-o-target.c,
* mipsbsd.c, * mmo.c, * nlm-target.h, * oasys.c, * pdp11.c,
* pe-mips.c, * pef.c, * plugin.c, * ppcboot.c, * som.c, * srec.c,
* tekhex.c, * trad-core.c, * verilog.c, * versados.c, * vms-alpha.c,
* vms-lib.c, * xsym.c: Init match_priority field.
* configure: Regenerate.
* bfd-in2.h: Regenerate.
2011-06-06 03:26:05 +02:00
|
|
|
{
|
2013-01-28 08:58:16 +01:00
|
|
|
/* If this is the default target, accept it, even if
|
|
|
|
other targets might match. People who want those
|
|
|
|
other targets have to set the GNUTARGET variable. */
|
|
|
|
if (temp == bfd_default_vector[0])
|
|
|
|
goto ok_ret;
|
|
|
|
|
|
|
|
if (matching_vector)
|
|
|
|
matching_vector[match_count] = temp;
|
|
|
|
match_count++;
|
|
|
|
|
Fix BFD format matching for x86_64-w64-mingw32 -m32 LTO.
In <https://sourceware.org/ml/binutils/2015-12/msg00190.html> (commit
4a07dc81356ed8728e204e9aabeb256703c59aef), Kwok fixed a problem with
the template used for a dummy BFD for an IR file for LTO on MinGW,
where the input and output formats are not the same.
A problem, however, remains in the case of linking for
x86_64-w64-mingw32 -m32, where LTO linking reports an ambiguity
between the pe-i386 and pei-i386 formats. An object (pe-i386) with
plugin data is being tested by the linker to see what formats match.
The default format initially set by the linker when
bfd_check_format_matches is called is pei-i386 (as that's the output
format from the linker script), which does not match, so the function
goes on to the loop over possible BFD vectors. The pe-i386 vector
matches, as it should. One other vector matches: the plugin vector.
bfd_check_format_matches tests a vector for matching by temporarily
modifying the BFD object to use that vector then using
_bfd_check_format on it. So the BFD object is temporarily using
plugin_vec. _bfd_check_format ends up using bfd_plugin_object_p which
uses plugin_object_p from ld which uses plugin_get_ir_dummy_bfd which
succeeds, having created a BFD based on link_info.output_bfd (because
srctemplate is the BFD temporarily using plugin_vec, even after Kwok's
patch link_info.output_bfd is all that's available to base the dummy
BFD on). So we end up with a match from the plugin vector which uses
the pei-i386 vector even though the pei-i386 vector itself does not
match the input object. (In the i686-mingw32 case, as opposed to this
multilib case, pe-i386 is the default BFD target, which would
short-circuit that logic.)
There are two cases of the linker handling inputs with a plugin: they
may be inputs that are also accepted by some non-plugin BFD format, as
here, or they may be a format that would not be recognized at all, as
with some tests in the ld testsuite. In the former case, there is no
need for BFD to accept the objects using the plugin vector, as the
linker has its own logic to allow plugins to claim objects accepted by
some other BFD vector. Thus, this patch arranges for the plugin
vector to have the lowest match priority, and for the priority from
that vector to be used in the relevant case (the attempted match to
the plugin vector results in TEMP pointing to the pei-i386 vector).
Tested for GCC and Binutils testsuites for x86_64-pc-linux-gnu, as
well as verifying that it fixes the observed LTO issue for
x86_64-w64-mingw32.
* plugin.c (plugin_vec): Set match priority to 255.
* format.c (bfd_check_format_matches) [BFD_SUPPORTS_PLUGINS]: When
matching against the plugin vector, take priority from there not
from TEMP.
2016-02-16 01:27:11 +01:00
|
|
|
if (match_priority < best_match)
|
2013-01-28 08:58:16 +01:00
|
|
|
{
|
Fix BFD format matching for x86_64-w64-mingw32 -m32 LTO.
In <https://sourceware.org/ml/binutils/2015-12/msg00190.html> (commit
4a07dc81356ed8728e204e9aabeb256703c59aef), Kwok fixed a problem with
the template used for a dummy BFD for an IR file for LTO on MinGW,
where the input and output formats are not the same.
A problem, however, remains in the case of linking for
x86_64-w64-mingw32 -m32, where LTO linking reports an ambiguity
between the pe-i386 and pei-i386 formats. An object (pe-i386) with
plugin data is being tested by the linker to see what formats match.
The default format initially set by the linker when
bfd_check_format_matches is called is pei-i386 (as that's the output
format from the linker script), which does not match, so the function
goes on to the loop over possible BFD vectors. The pe-i386 vector
matches, as it should. One other vector matches: the plugin vector.
bfd_check_format_matches tests a vector for matching by temporarily
modifying the BFD object to use that vector then using
_bfd_check_format on it. So the BFD object is temporarily using
plugin_vec. _bfd_check_format ends up using bfd_plugin_object_p which
uses plugin_object_p from ld which uses plugin_get_ir_dummy_bfd which
succeeds, having created a BFD based on link_info.output_bfd (because
srctemplate is the BFD temporarily using plugin_vec, even after Kwok's
patch link_info.output_bfd is all that's available to base the dummy
BFD on). So we end up with a match from the plugin vector which uses
the pei-i386 vector even though the pei-i386 vector itself does not
match the input object. (In the i686-mingw32 case, as opposed to this
multilib case, pe-i386 is the default BFD target, which would
short-circuit that logic.)
There are two cases of the linker handling inputs with a plugin: they
may be inputs that are also accepted by some non-plugin BFD format, as
here, or they may be a format that would not be recognized at all, as
with some tests in the ld testsuite. In the former case, there is no
need for BFD to accept the objects using the plugin vector, as the
linker has its own logic to allow plugins to claim objects accepted by
some other BFD vector. Thus, this patch arranges for the plugin
vector to have the lowest match priority, and for the priority from
that vector to be used in the relevant case (the attempted match to
the plugin vector results in TEMP pointing to the pei-i386 vector).
Tested for GCC and Binutils testsuites for x86_64-pc-linux-gnu, as
well as verifying that it fixes the observed LTO issue for
x86_64-w64-mingw32.
* plugin.c (plugin_vec): Set match priority to 255.
* format.c (bfd_check_format_matches) [BFD_SUPPORTS_PLUGINS]: When
matching against the plugin vector, take priority from there not
from TEMP.
2016-02-16 01:27:11 +01:00
|
|
|
best_match = match_priority;
|
2013-01-28 08:58:16 +01:00
|
|
|
best_count = 0;
|
|
|
|
}
|
PR22458, failure to choose a matching ELF target
https://sourceware.org/ml/binutils/2013-05/msg00271.html was supposed
to banish "file format is ambiguous" errors for ELF. It didn't,
because the code supposedly detecting formats that implement
match_priority didn't work. That was due to not placing all matching
targets into the vector of matching targets. ELF objects should all
match the generic ELF target (priority 2), plus one or more machine
specific targets (priority 1), and perhaps a single machine specific
target with OS/ABI set (priority 0, best match). So the armel object
in the testcase actually matches elf32-littlearm,
elf32-littlearm-symbian, and elf32-littlearm-vxworks (all priority 1),
and elf32-little (priority 2). As the PR reported, elf32-little
wasn't seen as matching. Fixing that part of the problem wasn't too
difficult but matching the generic ELF target as well as the ARM ELF
targets resulted in ARM testsuite failures.
These proved to be the annoying reordering of stubs that occurs from
time to time due to the stub names containing the section id.
Matching another target causes more sections to be created in
elf_object_p. If section ids change, stub names change, which results
in different hashing and can therefore result in different hash table
traversal and stub creation order. That particular problem is fixed
by resetting section_id to the initial state before attempting each
target match, and taking a snapshot of its value after a successful
match.
PR 22458
* format.c (struct bfd_preserve): Add section_id.
(bfd_preserve_save, bfd_preserve_restore): Save and restore
_bfd_section_id.
(bfd_reinit): Set _bfd_section_id.
(bfd_check_format_matches): Put all matches of any priority into
matching_vector. Save initial section id and start each attempted
match at that section id.
* libbfd-in.h (_bfd_section_id): Declare.
* section.c (_bfd_section_id): Rename from section_id and make
global. Adjust uses.
(bfd_get_next_section_id): Delete.
* elf64-ppc.c (ppc64_elf_setup_section_lists): Replace use of
bfd_get_section_id with _bfd_section_id.
* libbfd.h: Regenerate.
* bfd-in2.h: Regenerate.
2018-05-16 04:03:48 +02:00
|
|
|
if (match_priority <= best_match)
|
|
|
|
{
|
|
|
|
/* This format checks out as ok! */
|
|
|
|
right_targ = temp;
|
|
|
|
best_count++;
|
|
|
|
}
|
2013-01-28 08:58:16 +01:00
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
/* An archive with no armap or objects of the wrong
|
|
|
|
type. We want this target to match if we get no
|
|
|
|
better matches. */
|
|
|
|
if (ar_right_targ != bfd_default_vector[0])
|
|
|
|
ar_right_targ = *target;
|
|
|
|
if (matching_vector)
|
|
|
|
matching_vector[ar_match_index] = *target;
|
|
|
|
ar_match_index++;
|
* targets.c (bfd_target): Make ar_max_namelen an unsigned char.
Add match_priority.
* configure.in: Bump bfd version.
* elfcode.h (elf_object_p): Delete hacks preventing match of
EM_NONE and ELFOSABI_NONE targets when a better match exists.
* elfxx-target.h (elf_match_priority): Define and use.
* format.c (bfd_check_format_matches): Use target match_priority
to choose best of multiple matching targets. In cases with multiple
matches rerun _bfd_check_format if we don't choose the last match.
* aout-adobe.c, * aout-arm.c, * aout-target.h, * aout-tic30.c,
* binary.c, * bout.c, * coff-alpha.c, * coff-i386.c, * coff-i860.c,
* coff-i960.c, * coff-ia64.c, * coff-mips.c, * coff-or32.c,
* coff-ppc.c, * coff-rs6000.c, * coff-sh.c, * coff-tic30.c,
* coff-tic54x.c, * coff-x86_64.c, * coff64-rs6000.c, * coffcode.h,
* i386msdos.c, * i386os9k.c, * ieee.c, * ihex.c, * mach-o-target.c,
* mipsbsd.c, * mmo.c, * nlm-target.h, * oasys.c, * pdp11.c,
* pe-mips.c, * pef.c, * plugin.c, * ppcboot.c, * som.c, * srec.c,
* tekhex.c, * trad-core.c, * verilog.c, * versados.c, * vms-alpha.c,
* vms-lib.c, * xsym.c: Init match_priority field.
* configure: Regenerate.
* bfd-in2.h: Regenerate.
2011-06-06 03:26:05 +02:00
|
|
|
}
|
2013-01-26 03:08:01 +01:00
|
|
|
|
2013-01-28 08:58:16 +01:00
|
|
|
if (!bfd_preserve_save (abfd, &preserve))
|
|
|
|
goto err_ret;
|
|
|
|
}
|
|
|
|
else if (bfd_get_error () != bfd_error_wrong_format)
|
2013-01-26 03:08:01 +01:00
|
|
|
goto err_ret;
|
2000-07-20 02:31:39 +02:00
|
|
|
}
|
2001-01-23 21:27:54 +01:00
|
|
|
|
* targets.c (bfd_target): Make ar_max_namelen an unsigned char.
Add match_priority.
* configure.in: Bump bfd version.
* elfcode.h (elf_object_p): Delete hacks preventing match of
EM_NONE and ELFOSABI_NONE targets when a better match exists.
* elfxx-target.h (elf_match_priority): Define and use.
* format.c (bfd_check_format_matches): Use target match_priority
to choose best of multiple matching targets. In cases with multiple
matches rerun _bfd_check_format if we don't choose the last match.
* aout-adobe.c, * aout-arm.c, * aout-target.h, * aout-tic30.c,
* binary.c, * bout.c, * coff-alpha.c, * coff-i386.c, * coff-i860.c,
* coff-i960.c, * coff-ia64.c, * coff-mips.c, * coff-or32.c,
* coff-ppc.c, * coff-rs6000.c, * coff-sh.c, * coff-tic30.c,
* coff-tic54x.c, * coff-x86_64.c, * coff64-rs6000.c, * coffcode.h,
* i386msdos.c, * i386os9k.c, * ieee.c, * ihex.c, * mach-o-target.c,
* mipsbsd.c, * mmo.c, * nlm-target.h, * oasys.c, * pdp11.c,
* pe-mips.c, * pef.c, * plugin.c, * ppcboot.c, * som.c, * srec.c,
* tekhex.c, * trad-core.c, * verilog.c, * versados.c, * vms-alpha.c,
* vms-lib.c, * xsym.c: Init match_priority field.
* configure: Regenerate.
* bfd-in2.h: Regenerate.
2011-06-06 03:26:05 +02:00
|
|
|
if (best_count == 1)
|
|
|
|
match_count = 1;
|
|
|
|
|
2001-08-17 05:19:01 +02:00
|
|
|
if (match_count == 0)
|
|
|
|
{
|
|
|
|
/* Try partial matches. */
|
|
|
|
right_targ = ar_right_targ;
|
2003-02-14 12:16:09 +01:00
|
|
|
|
2001-08-17 05:19:01 +02:00
|
|
|
if (right_targ == bfd_default_vector[0])
|
|
|
|
{
|
|
|
|
match_count = 1;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
match_count = ar_match_index - _bfd_target_vector_entries;
|
2003-02-14 12:16:09 +01:00
|
|
|
|
2007-09-14 07:19:27 +02:00
|
|
|
if (matching_vector && match_count > 1)
|
2003-02-14 12:16:09 +01:00
|
|
|
memcpy (matching_vector,
|
|
|
|
matching_vector + _bfd_target_vector_entries,
|
|
|
|
sizeof (*matching_vector) * match_count);
|
2002-11-05 02:18:56 +01:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-05-23 05:35:59 +02:00
|
|
|
/* We have more than one equally good match. If any of the best
|
|
|
|
matches is a target in config.bfd targ_defvec or targ_selvecs,
|
|
|
|
choose it. */
|
2007-09-14 07:19:27 +02:00
|
|
|
if (match_count > 1)
|
2002-11-05 02:18:56 +01:00
|
|
|
{
|
|
|
|
const bfd_target * const *assoc = bfd_associated_vector;
|
|
|
|
|
|
|
|
while ((right_targ = *assoc++) != NULL)
|
|
|
|
{
|
|
|
|
int i = match_count;
|
|
|
|
|
|
|
|
while (--i >= 0)
|
2013-05-23 05:35:59 +02:00
|
|
|
if (matching_vector[i] == right_targ
|
|
|
|
&& right_targ->match_priority <= best_match)
|
2002-11-05 02:18:56 +01:00
|
|
|
break;
|
|
|
|
|
|
|
|
if (i >= 0)
|
|
|
|
{
|
|
|
|
match_count = 1;
|
|
|
|
break;
|
2001-08-17 05:19:01 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-05-23 05:35:59 +02:00
|
|
|
/* We still have more than one equally good match, and at least some
|
|
|
|
of the targets support match priority. Choose the first of the
|
|
|
|
best matches. */
|
2014-12-16 15:17:15 +01:00
|
|
|
if (matching_vector && match_count > 1 && best_count != match_count)
|
2013-05-23 05:35:59 +02:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < match_count; i++)
|
|
|
|
{
|
|
|
|
right_targ = matching_vector[i];
|
|
|
|
if (right_targ->match_priority <= best_match)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
match_count = 1;
|
|
|
|
}
|
|
|
|
|
2013-01-26 03:08:01 +01:00
|
|
|
/* There is way too much undoing of half-known state here. We
|
|
|
|
really shouldn't iterate on live bfd's. Note that saving the
|
|
|
|
whole bfd and restoring it would be even worse; the first thing
|
|
|
|
you notice is that the cached bfd file position gets out of sync. */
|
|
|
|
if (preserve.marker != NULL)
|
|
|
|
bfd_preserve_restore (abfd, &preserve);
|
|
|
|
|
2000-07-20 02:31:39 +02:00
|
|
|
if (match_count == 1)
|
|
|
|
{
|
* targets.c (bfd_target): Make ar_max_namelen an unsigned char.
Add match_priority.
* configure.in: Bump bfd version.
* elfcode.h (elf_object_p): Delete hacks preventing match of
EM_NONE and ELFOSABI_NONE targets when a better match exists.
* elfxx-target.h (elf_match_priority): Define and use.
* format.c (bfd_check_format_matches): Use target match_priority
to choose best of multiple matching targets. In cases with multiple
matches rerun _bfd_check_format if we don't choose the last match.
* aout-adobe.c, * aout-arm.c, * aout-target.h, * aout-tic30.c,
* binary.c, * bout.c, * coff-alpha.c, * coff-i386.c, * coff-i860.c,
* coff-i960.c, * coff-ia64.c, * coff-mips.c, * coff-or32.c,
* coff-ppc.c, * coff-rs6000.c, * coff-sh.c, * coff-tic30.c,
* coff-tic54x.c, * coff-x86_64.c, * coff64-rs6000.c, * coffcode.h,
* i386msdos.c, * i386os9k.c, * ieee.c, * ihex.c, * mach-o-target.c,
* mipsbsd.c, * mmo.c, * nlm-target.h, * oasys.c, * pdp11.c,
* pe-mips.c, * pef.c, * plugin.c, * ppcboot.c, * som.c, * srec.c,
* tekhex.c, * trad-core.c, * verilog.c, * versados.c, * vms-alpha.c,
* vms-lib.c, * xsym.c: Init match_priority field.
* configure: Regenerate.
* bfd-in2.h: Regenerate.
2011-06-06 03:26:05 +02:00
|
|
|
abfd->xvec = right_targ;
|
|
|
|
/* If we come out of the loop knowing that the last target that
|
|
|
|
matched is the one we want, then ABFD should still be in a usable
|
|
|
|
state (except possibly for XVEC). */
|
|
|
|
if (match_targ != right_targ)
|
|
|
|
{
|
PR22458, failure to choose a matching ELF target
https://sourceware.org/ml/binutils/2013-05/msg00271.html was supposed
to banish "file format is ambiguous" errors for ELF. It didn't,
because the code supposedly detecting formats that implement
match_priority didn't work. That was due to not placing all matching
targets into the vector of matching targets. ELF objects should all
match the generic ELF target (priority 2), plus one or more machine
specific targets (priority 1), and perhaps a single machine specific
target with OS/ABI set (priority 0, best match). So the armel object
in the testcase actually matches elf32-littlearm,
elf32-littlearm-symbian, and elf32-littlearm-vxworks (all priority 1),
and elf32-little (priority 2). As the PR reported, elf32-little
wasn't seen as matching. Fixing that part of the problem wasn't too
difficult but matching the generic ELF target as well as the ARM ELF
targets resulted in ARM testsuite failures.
These proved to be the annoying reordering of stubs that occurs from
time to time due to the stub names containing the section id.
Matching another target causes more sections to be created in
elf_object_p. If section ids change, stub names change, which results
in different hashing and can therefore result in different hash table
traversal and stub creation order. That particular problem is fixed
by resetting section_id to the initial state before attempting each
target match, and taking a snapshot of its value after a successful
match.
PR 22458
* format.c (struct bfd_preserve): Add section_id.
(bfd_preserve_save, bfd_preserve_restore): Save and restore
_bfd_section_id.
(bfd_reinit): Set _bfd_section_id.
(bfd_check_format_matches): Put all matches of any priority into
matching_vector. Save initial section id and start each attempted
match at that section id.
* libbfd-in.h (_bfd_section_id): Declare.
* section.c (_bfd_section_id): Rename from section_id and make
global. Adjust uses.
(bfd_get_next_section_id): Delete.
* elf64-ppc.c (ppc64_elf_setup_section_lists): Replace use of
bfd_get_section_id with _bfd_section_id.
* libbfd.h: Regenerate.
* bfd-in2.h: Regenerate.
2018-05-16 04:03:48 +02:00
|
|
|
bfd_reinit (abfd, initial_section_id);
|
* targets.c (bfd_target): Make ar_max_namelen an unsigned char.
Add match_priority.
* configure.in: Bump bfd version.
* elfcode.h (elf_object_p): Delete hacks preventing match of
EM_NONE and ELFOSABI_NONE targets when a better match exists.
* elfxx-target.h (elf_match_priority): Define and use.
* format.c (bfd_check_format_matches): Use target match_priority
to choose best of multiple matching targets. In cases with multiple
matches rerun _bfd_check_format if we don't choose the last match.
* aout-adobe.c, * aout-arm.c, * aout-target.h, * aout-tic30.c,
* binary.c, * bout.c, * coff-alpha.c, * coff-i386.c, * coff-i860.c,
* coff-i960.c, * coff-ia64.c, * coff-mips.c, * coff-or32.c,
* coff-ppc.c, * coff-rs6000.c, * coff-sh.c, * coff-tic30.c,
* coff-tic54x.c, * coff-x86_64.c, * coff64-rs6000.c, * coffcode.h,
* i386msdos.c, * i386os9k.c, * ieee.c, * ihex.c, * mach-o-target.c,
* mipsbsd.c, * mmo.c, * nlm-target.h, * oasys.c, * pdp11.c,
* pe-mips.c, * pef.c, * plugin.c, * ppcboot.c, * som.c, * srec.c,
* tekhex.c, * trad-core.c, * verilog.c, * versados.c, * vms-alpha.c,
* vms-lib.c, * xsym.c: Init match_priority field.
* configure: Regenerate.
* bfd-in2.h: Regenerate.
2011-06-06 03:26:05 +02:00
|
|
|
if (bfd_seek (abfd, (file_ptr) 0, SEEK_SET) != 0)
|
|
|
|
goto err_ret;
|
|
|
|
match_targ = BFD_SEND_FMT (abfd, _bfd_check_format, (abfd));
|
2013-01-26 03:08:01 +01:00
|
|
|
BFD_ASSERT (match_targ != NULL);
|
* targets.c (bfd_target): Make ar_max_namelen an unsigned char.
Add match_priority.
* configure.in: Bump bfd version.
* elfcode.h (elf_object_p): Delete hacks preventing match of
EM_NONE and ELFOSABI_NONE targets when a better match exists.
* elfxx-target.h (elf_match_priority): Define and use.
* format.c (bfd_check_format_matches): Use target match_priority
to choose best of multiple matching targets. In cases with multiple
matches rerun _bfd_check_format if we don't choose the last match.
* aout-adobe.c, * aout-arm.c, * aout-target.h, * aout-tic30.c,
* binary.c, * bout.c, * coff-alpha.c, * coff-i386.c, * coff-i860.c,
* coff-i960.c, * coff-ia64.c, * coff-mips.c, * coff-or32.c,
* coff-ppc.c, * coff-rs6000.c, * coff-sh.c, * coff-tic30.c,
* coff-tic54x.c, * coff-x86_64.c, * coff64-rs6000.c, * coffcode.h,
* i386msdos.c, * i386os9k.c, * ieee.c, * ihex.c, * mach-o-target.c,
* mipsbsd.c, * mmo.c, * nlm-target.h, * oasys.c, * pdp11.c,
* pe-mips.c, * pef.c, * plugin.c, * ppcboot.c, * som.c, * srec.c,
* tekhex.c, * trad-core.c, * verilog.c, * versados.c, * vms-alpha.c,
* vms-lib.c, * xsym.c: Init match_priority field.
* configure: Regenerate.
* bfd-in2.h: Regenerate.
2011-06-06 03:26:05 +02:00
|
|
|
}
|
2001-01-23 21:27:54 +01:00
|
|
|
|
* targets.c (bfd_target): Make ar_max_namelen an unsigned char.
Add match_priority.
* configure.in: Bump bfd version.
* elfcode.h (elf_object_p): Delete hacks preventing match of
EM_NONE and ELFOSABI_NONE targets when a better match exists.
* elfxx-target.h (elf_match_priority): Define and use.
* format.c (bfd_check_format_matches): Use target match_priority
to choose best of multiple matching targets. In cases with multiple
matches rerun _bfd_check_format if we don't choose the last match.
* aout-adobe.c, * aout-arm.c, * aout-target.h, * aout-tic30.c,
* binary.c, * bout.c, * coff-alpha.c, * coff-i386.c, * coff-i860.c,
* coff-i960.c, * coff-ia64.c, * coff-mips.c, * coff-or32.c,
* coff-ppc.c, * coff-rs6000.c, * coff-sh.c, * coff-tic30.c,
* coff-tic54x.c, * coff-x86_64.c, * coff64-rs6000.c, * coffcode.h,
* i386msdos.c, * i386os9k.c, * ieee.c, * ihex.c, * mach-o-target.c,
* mipsbsd.c, * mmo.c, * nlm-target.h, * oasys.c, * pdp11.c,
* pe-mips.c, * pef.c, * plugin.c, * ppcboot.c, * som.c, * srec.c,
* tekhex.c, * trad-core.c, * verilog.c, * versados.c, * vms-alpha.c,
* vms-lib.c, * xsym.c: Init match_priority field.
* configure: Regenerate.
* bfd-in2.h: Regenerate.
2011-06-06 03:26:05 +02:00
|
|
|
ok_ret:
|
2005-05-17 21:44:55 +02:00
|
|
|
/* If the file was opened for update, then `output_has_begun'
|
|
|
|
some time ago when the file was created. Do not recompute
|
|
|
|
sections sizes or alignments in _bfd_set_section_contents.
|
|
|
|
We can not set this flag until after checking the format,
|
|
|
|
because it will interfere with creation of BFD sections. */
|
|
|
|
if (abfd->direction == both_direction)
|
|
|
|
abfd->output_has_begun = TRUE;
|
|
|
|
|
2007-09-14 07:19:27 +02:00
|
|
|
if (matching_vector)
|
|
|
|
free (matching_vector);
|
2013-01-26 03:08:01 +01:00
|
|
|
|
|
|
|
/* File position has moved, BTW. */
|
|
|
|
return TRUE;
|
1999-05-03 09:29:11 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
if (match_count == 0)
|
|
|
|
{
|
2007-09-14 07:19:27 +02:00
|
|
|
err_unrecog:
|
1999-05-03 09:29:11 +02:00
|
|
|
bfd_set_error (bfd_error_file_not_recognized);
|
2007-09-14 07:19:27 +02:00
|
|
|
err_ret:
|
|
|
|
abfd->xvec = save_targ;
|
|
|
|
abfd->format = bfd_unknown;
|
|
|
|
if (matching_vector)
|
2003-06-29 12:06:40 +02:00
|
|
|
free (matching_vector);
|
2013-01-26 03:08:01 +01:00
|
|
|
if (preserve.marker != NULL)
|
|
|
|
bfd_preserve_restore (abfd, &preserve);
|
2007-09-14 07:19:27 +02:00
|
|
|
return FALSE;
|
1999-05-03 09:29:11 +02:00
|
|
|
}
|
2001-08-17 05:19:01 +02:00
|
|
|
|
2013-01-26 03:08:01 +01:00
|
|
|
/* Restore original target type and format. */
|
|
|
|
abfd->xvec = save_targ;
|
|
|
|
abfd->format = bfd_unknown;
|
2007-09-14 07:19:27 +02:00
|
|
|
bfd_set_error (bfd_error_file_ambiguously_recognized);
|
|
|
|
|
|
|
|
if (matching)
|
|
|
|
{
|
|
|
|
*matching = (char **) matching_vector;
|
|
|
|
matching_vector[match_count] = NULL;
|
|
|
|
/* Return target names. This is a little nasty. Maybe we
|
|
|
|
should do another bfd_malloc? */
|
|
|
|
while (--match_count >= 0)
|
2001-08-17 05:19:01 +02:00
|
|
|
{
|
2007-09-14 07:19:27 +02:00
|
|
|
const char *name = matching_vector[match_count]->name;
|
|
|
|
*(const char **) &matching_vector[match_count] = name;
|
2001-08-17 05:19:01 +02:00
|
|
|
}
|
|
|
|
}
|
2002-11-30 09:39:46 +01:00
|
|
|
return FALSE;
|
1999-05-03 09:29:11 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
FUNCTION
|
|
|
|
bfd_set_format
|
|
|
|
|
|
|
|
SYNOPSIS
|
2003-02-14 12:16:09 +01:00
|
|
|
bfd_boolean bfd_set_format (bfd *abfd, bfd_format format);
|
1999-05-03 09:29:11 +02:00
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
This function sets the file format of the BFD @var{abfd} to the
|
|
|
|
format @var{format}. If the target set in the BFD does not
|
|
|
|
support the format requested, the format is invalid, or the BFD
|
|
|
|
is not open for writing, then an error occurs.
|
|
|
|
*/
|
|
|
|
|
2002-11-30 09:39:46 +01:00
|
|
|
bfd_boolean
|
2003-06-29 12:06:40 +02:00
|
|
|
bfd_set_format (bfd *abfd, bfd_format format)
|
1999-05-03 09:29:11 +02:00
|
|
|
{
|
2001-08-17 05:19:01 +02:00
|
|
|
if (bfd_read_p (abfd)
|
2001-08-17 10:57:42 +02:00
|
|
|
|| (unsigned int) abfd->format >= (unsigned int) bfd_type_end)
|
2000-07-20 02:31:39 +02:00
|
|
|
{
|
|
|
|
bfd_set_error (bfd_error_invalid_operation);
|
2002-11-30 09:39:46 +01:00
|
|
|
return FALSE;
|
2000-07-20 02:31:39 +02:00
|
|
|
}
|
1999-05-03 09:29:11 +02:00
|
|
|
|
|
|
|
if (abfd->format != bfd_unknown)
|
2002-11-30 09:39:46 +01:00
|
|
|
return abfd->format == format;
|
1999-05-03 09:29:11 +02:00
|
|
|
|
2000-07-20 02:31:39 +02:00
|
|
|
/* Presume the answer is yes. */
|
1999-05-03 09:29:11 +02:00
|
|
|
abfd->format = format;
|
|
|
|
|
2000-07-20 02:31:39 +02:00
|
|
|
if (!BFD_SEND_FMT (abfd, _bfd_set_format, (abfd)))
|
|
|
|
{
|
|
|
|
abfd->format = bfd_unknown;
|
2002-11-30 09:39:46 +01:00
|
|
|
return FALSE;
|
2000-07-20 02:31:39 +02:00
|
|
|
}
|
1999-05-03 09:29:11 +02:00
|
|
|
|
2002-11-30 09:39:46 +01:00
|
|
|
return TRUE;
|
1999-05-03 09:29:11 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
FUNCTION
|
|
|
|
bfd_format_string
|
|
|
|
|
|
|
|
SYNOPSIS
|
2003-02-14 12:16:09 +01:00
|
|
|
const char *bfd_format_string (bfd_format format);
|
1999-05-03 09:29:11 +02:00
|
|
|
|
|
|
|
DESCRIPTION
|
|
|
|
Return a pointer to a const string
|
|
|
|
<<invalid>>, <<object>>, <<archive>>, <<core>>, or <<unknown>>,
|
|
|
|
depending upon the value of @var{format}.
|
|
|
|
*/
|
|
|
|
|
2001-08-17 05:19:01 +02:00
|
|
|
const char *
|
2003-06-29 12:06:40 +02:00
|
|
|
bfd_format_string (bfd_format format)
|
1999-05-03 09:29:11 +02:00
|
|
|
{
|
2003-06-29 12:06:40 +02:00
|
|
|
if (((int) format < (int) bfd_unknown)
|
|
|
|
|| ((int) format >= (int) bfd_type_end))
|
1999-05-03 09:29:11 +02:00
|
|
|
return "invalid";
|
2001-01-23 21:27:54 +01:00
|
|
|
|
2000-07-20 02:31:39 +02:00
|
|
|
switch (format)
|
|
|
|
{
|
|
|
|
case bfd_object:
|
2003-11-30 19:40:41 +01:00
|
|
|
return "object"; /* Linker/assembler/compiler output. */
|
2001-01-23 21:27:54 +01:00
|
|
|
case bfd_archive:
|
2000-07-20 02:31:39 +02:00
|
|
|
return "archive"; /* Object archive file. */
|
2001-01-23 21:27:54 +01:00
|
|
|
case bfd_core:
|
2000-07-20 02:31:39 +02:00
|
|
|
return "core"; /* Core dump. */
|
2001-01-23 21:27:54 +01:00
|
|
|
default:
|
2000-07-20 02:31:39 +02:00
|
|
|
return "unknown";
|
|
|
|
}
|
1999-05-03 09:29:11 +02:00
|
|
|
}
|