mirror of
https://sourceware.org/git/binutils-gdb.git
synced 2024-12-15 04:31:49 +08:00
1ff6de0312
This is quite complicated because the CTF section's contents depend on the final contents of the symtab and strtab, because it has two sections whose contents are shuffled to be in 1:1 correspondence with the symtab, and an internal strtab that gets deduplicated against the ELF strtab (with offsets adjusted to point into the ELF strtab instead). It is also compressed if large enough, so its size depends on its contents! So we cannot construct it as early as most sections: we cannot even *begin* construction until after the symtab and strtab are finalized. Thankfully there is already one section treated similarly: compressed debugging sections: the only differences are that compressed debugging sections have extra handling to deal with their changing name if compressed (CTF sections are always called ".ctf" for now, though we have reserved ".ctf.*" against future use), and that compressed debugging sections have previously-uncompressed content which has to be stashed away for later compression, while CTF sections have no content at all until we generate it (very late). BFD also cannot do the link itself: libctf knows how to do it, and BFD cannot call libctf directly because libctf already depends on bfd for file I/O. So we have to use a pair of callbacks, one, examine_strtab, which allows a caller to examine the symtab and strtab after finalization (called from elf_link_swap_symbols_out(), right before the symtabs are written, and after the strtab has been finalized), and one which actually does the emission (called emit_ctf simply because it is grouped with a bunch of section-specific late-emission function calls at the bottom of bfd_elf_final_link, and a section-specific name seems best for that). emit_ctf is actually called *twice*: once from lang_process if the emulation suggests that this bfd target does not examine the symtab or strtab, and once via a bfd callback if it does. (This means that non-ELF targets still get CTF emitted, even though the late CTF emission stage is never called for them). v2: merged with non-ELF support patch: slight commit message adjustments. v3: do not spend time merging CTF, or crash, if the CTF section is explicitly discarded. Do not try to merge or compress CTF unless linking. v4: add CTF_COMPRESSION_THRESHOLD. Annul the freed input ctf_file_t's after writeout: set SEC_IN_MEMORY on the output contents so a future bfd enhancement knows it could free it. Add SEC_LINKER_CREATED | SEC_KEEP to avoid having to add .ctf to the linker script. Drop now-unnecessary ldlang.h-level elf-bfd.h include and hackery around it. Adapt to elf32.em->elf.em and elf-generic.em->ldelf*.c changes. v5: fix tabdamage. Drop #inclusions in .h files: include in .c files, .em files, and use struct forwards instead. Use bfd_section_is_ctf inline function rather than SECTION_IS_CTF macro. Move a few comments. * Makefile.def (dependencies): all-ld depends on all-libctf. * Makefile.in: Regenerated. include/ * bfdlink.h (elf_strtab_hash): New forward. (elf_sym_strtab): Likewise. (struct bfd_link_callbacks <examine_strtab>): New. (struct bfd_link_callbacks <emit_ctf>): Likewise. bfd/ * elf-bfd.h (bfd_section_is_ctf): New inline function. * elf.c (special_sections_c): Add ".ctf". (assign_file_positions_for_non_load_sections): Note that compressed debugging sections etc are not assigned here. Treat CTF sections like SEC_ELF_COMPRESS sections when is_linker_output: sh_offset -1. (assign_file_positions_except_relocs): Likewise. (find_section_in_list): Note that debugging and CTF sections, as well as reloc sections, are assigned later. (_bfd_elf_assign_file_positions_for_non_load): CTF sections get their size and contents updated. (_bfd_elf_set_section_contents): Skip CTF sections: unlike compressed sections, they have no uncompressed content to copy at this stage. * elflink.c (elf_link_swap_symbols_out): Call the examine_strtab callback right before the strtab is written out. (bfd_elf_final_link): Don't cache the section contents of CTF sections: they are not populated yet. Call the emit_ctf callback right at the end, after all the symbols and strings are flushed out. ld/ * ldlang.h: (struct lang_input_statement_struct): Add the_ctf. (struct elf_sym_strtab): Add forward. (struct elf_strtab_hash): Likewise. (ldlang_ctf_apply_strsym): Declare. (ldlang_write_ctf_late): Likewise. * ldemul.h (ldemul_emit_ctf_early): New. (ldemul_examine_strtab_for_ctf): Likewise. (ld_emulation_xfer_type) <emit_ctf_early>: Likewise. (ld_emulation_xfer_type) <examine_strtab_for_ctf>: Likewise. * ldemul.c (ldemul_emit_ctf_early): New. (ldemul_examine_strtab_for_ctf): Likewise. * ldlang.c: Include ctf-api.h. (CTF_COMPRESSION_THRESHOLD): New. (ctf_output): New. Initialized in... (ldlang_open_ctf): ... this new function. Open all the CTF sections in the input files: mark them non-loaded and empty so as not to copy their contents to the output, but linker-created so the section gets created in the target. (ldlang_merge_ctf): New, merge types via ctf_link_add_ctf and ctf_link. (ldlang_ctf_apply_strsym): New, an examine_strtab callback: wrap ldemul_examine_strtab_for_ctf. (lang_write_ctf): New, write out the CTF section. (ldlang_write_ctf_late): New, late call via bfd's emit_ctf hook. (lang_process): Call ldlang_open_ctf, ldlang_merge_ctf, and lang_write_ctf. * ldmain.c (link_callbacks): Add ldlang_ctf_apply_strsym, ldlang_write_ctf_late. * emultempl/aix.em: Add ctf-api.h. * emultempl/armcoff.em: Likewise. * emultempl/beos.em: Likewise. * emultempl/elf.em: Likewise. * emultempl/generic.em: Likewise. * emultempl/linux.em: Likewise. * emultempl/msp430.em: Likewise. * emultempl/pe.em: Likewise. * emultempl/pep.em: Likewise. * emultempl/ticoff.em: Likewise. * emultempl/vanilla.em: Likewise. * ldcref.c: Likewise. * ldctor.c: Likewise. * ldelf.c: Likewise. * ldelfgen.c: Likewise. * ldemul.c: Likewise. * ldexp.c: Likewise. * ldfile.c: Likewise. * ldgram.c: Likewise. * ldlex.l: Likewise. * ldmain.c: Likewise. * ldmisc.c: Likewise. * ldver.c: Likewise. * ldwrite.c: Likewise. * lexsup.c: Likewise. * mri.c: Likewise. * pe-dll.c: Likewise. * plugin.c: Likewise. * ldelfgen.c (ldelf_emit_ctf_early): New. (ldelf_examine_strtab_for_ctf): tell libctf about the symtab and strtab. (struct ctf_strsym_iter_cb_arg): New, state to do so. (ldelf_ctf_strtab_iter_cb): New: tell libctf about each string in the strtab in turn. (ldelf_ctf_symbols_iter_cb): New, tell libctf about each symbol in the symtab in turn. * ldelfgen.h (struct elf_sym_strtab): Add forward. (struct elf_strtab_hash): Likewise. (struct ctf_file): Likewise. (ldelf_emit_ctf_early): Declare. (ldelf_examine_strtab_for_ctf): Likewise. * emultempl/elf-generic.em (LDEMUL_EMIT_CTF_EARLY): Set it. (LDEMUL_EXAMINE_STRTAB_FOR_CTF): Likewise. * emultempl/aix.em (ld_${EMULATION_NAME}_emulation): Add emit_ctf_early and examine_strtab_for_ctf, NULL by default. * emultempl/armcoff.em (ld_${EMULATION_NAME}_emulation): Likewise. * emultempl/beos.em (ld_${EMULATION_NAME}_emulation): Likewise. * emultempl/elf.em (ld_${EMULATION_NAME}_emulation): Likewise. * emultempl/generic.em (ld_${EMULATION_NAME}_emulation): Likewise. * emultempl/linux.em (ld_${EMULATION_NAME}_emulation): Likewise. * emultempl/msp430.em (ld_${EMULATION_NAME}_emulation): Likewise. * emultempl/pe.em (ld_${EMULATION_NAME}_emulation): Likewise. * emultempl/pep.em (ld_${EMULATION_NAME}_emulation): Likewise. * emultempl/ticoff.em (ld_${EMULATION_NAME}_emulation): Likewise. * emultempl/vanilla.em (ld_vanilla_emulation): Likewise. * Makefile.am: Pull in libctf (and zlib, a transitive requirement for compressed CTF section emission). Pass it on to DejaGNU. * configure.ac: Add AM_ZLIB. * aclocal.m4: Added zlib.m4. * Makefile.in: Regenerated. * testsuite/ld-bootstrap/bootstrap.exp: Use it when relinking ld.
2137 lines
59 KiB
C
2137 lines
59 KiB
C
/* ELF emulation code for targets using elf.em.
|
|
Copyright (C) 1991-2019 Free Software Foundation, Inc.
|
|
|
|
This file is part of the GNU Binutils.
|
|
|
|
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. */
|
|
|
|
#include "sysdep.h"
|
|
#include "bfd.h"
|
|
#include "libiberty.h"
|
|
#include "filenames.h"
|
|
#include "safe-ctype.h"
|
|
#include "bfdlink.h"
|
|
#include "ctf-api.h"
|
|
#include "ld.h"
|
|
#include "ldmain.h"
|
|
#include "ldmisc.h"
|
|
#include "ldexp.h"
|
|
#include "ldlang.h"
|
|
#include "ldfile.h"
|
|
#include "ldemul.h"
|
|
#include "ldbuildid.h"
|
|
#include <ldgram.h>
|
|
#include "elf-bfd.h"
|
|
#ifdef HAVE_GLOB
|
|
#include <glob.h>
|
|
#endif
|
|
#include "ldelf.h"
|
|
|
|
struct dt_needed
|
|
{
|
|
bfd *by;
|
|
const char *name;
|
|
};
|
|
|
|
/* Style of .note.gnu.build-id section. */
|
|
const char *ldelf_emit_note_gnu_build_id;
|
|
|
|
/* These variables are required to pass information back and forth
|
|
between after_open and check_needed and stat_needed and vercheck. */
|
|
|
|
static struct bfd_link_needed_list *global_needed;
|
|
static lang_input_statement_type *global_found;
|
|
static struct stat global_stat;
|
|
static struct bfd_link_needed_list *global_vercheck_needed;
|
|
static bfd_boolean global_vercheck_failed;
|
|
|
|
void
|
|
ldelf_after_parse (void)
|
|
{
|
|
if (bfd_link_pie (&link_info))
|
|
link_info.flags_1 |= (bfd_vma) DF_1_PIE;
|
|
|
|
if (bfd_link_executable (&link_info)
|
|
&& link_info.nointerp)
|
|
{
|
|
if (link_info.dynamic_undefined_weak > 0)
|
|
einfo (_("%P: warning: -z dynamic-undefined-weak ignored\n"));
|
|
link_info.dynamic_undefined_weak = 0;
|
|
}
|
|
after_parse_default ();
|
|
}
|
|
|
|
/* Handle the generation of DT_NEEDED tags. */
|
|
|
|
bfd_boolean
|
|
ldelf_load_symbols (lang_input_statement_type *entry)
|
|
{
|
|
int link_class = 0;
|
|
|
|
/* Tell the ELF linker that we don't want the output file to have a
|
|
DT_NEEDED entry for this file, unless it is used to resolve
|
|
references in a regular object. */
|
|
if (entry->flags.add_DT_NEEDED_for_regular)
|
|
link_class = DYN_AS_NEEDED;
|
|
|
|
/* Tell the ELF linker that we don't want the output file to have a
|
|
DT_NEEDED entry for any dynamic library in DT_NEEDED tags from
|
|
this file at all. */
|
|
if (!entry->flags.add_DT_NEEDED_for_dynamic)
|
|
link_class |= DYN_NO_ADD_NEEDED;
|
|
|
|
if (entry->flags.just_syms
|
|
&& (bfd_get_file_flags (entry->the_bfd) & DYNAMIC) != 0)
|
|
einfo (_("%F%P: %pB: --just-symbols may not be used on DSO\n"),
|
|
entry->the_bfd);
|
|
|
|
if (link_class == 0
|
|
|| (bfd_get_file_flags (entry->the_bfd) & DYNAMIC) == 0)
|
|
return FALSE;
|
|
|
|
bfd_elf_set_dyn_lib_class (entry->the_bfd,
|
|
(enum dynamic_lib_link_class) link_class);
|
|
|
|
/* Continue on with normal load_symbols processing. */
|
|
return FALSE;
|
|
}
|
|
|
|
/* On Linux, it's possible to have different versions of the same
|
|
shared library linked against different versions of libc. The
|
|
dynamic linker somehow tags which libc version to use in
|
|
/etc/ld.so.cache, and, based on the libc that it sees in the
|
|
executable, chooses which version of the shared library to use.
|
|
|
|
We try to do a similar check here by checking whether this shared
|
|
library needs any other shared libraries which may conflict with
|
|
libraries we have already included in the link. If it does, we
|
|
skip it, and try to find another shared library farther on down the
|
|
link path.
|
|
|
|
This is called via lang_for_each_input_file.
|
|
GLOBAL_VERCHECK_NEEDED is the list of objects needed by the object
|
|
which we are checking. This sets GLOBAL_VERCHECK_FAILED if we find
|
|
a conflicting version. */
|
|
|
|
static void
|
|
ldelf_vercheck (lang_input_statement_type *s)
|
|
{
|
|
const char *soname;
|
|
struct bfd_link_needed_list *l;
|
|
|
|
if (global_vercheck_failed)
|
|
return;
|
|
if (s->the_bfd == NULL
|
|
|| (bfd_get_file_flags (s->the_bfd) & DYNAMIC) == 0)
|
|
return;
|
|
|
|
soname = bfd_elf_get_dt_soname (s->the_bfd);
|
|
if (soname == NULL)
|
|
soname = lbasename (bfd_get_filename (s->the_bfd));
|
|
|
|
for (l = global_vercheck_needed; l != NULL; l = l->next)
|
|
{
|
|
const char *suffix;
|
|
|
|
if (filename_cmp (soname, l->name) == 0)
|
|
{
|
|
/* Probably can't happen, but it's an easy check. */
|
|
continue;
|
|
}
|
|
|
|
if (strchr (l->name, '/') != NULL)
|
|
continue;
|
|
|
|
suffix = strstr (l->name, ".so.");
|
|
if (suffix == NULL)
|
|
continue;
|
|
|
|
suffix += sizeof ".so." - 1;
|
|
|
|
if (filename_ncmp (soname, l->name, suffix - l->name) == 0)
|
|
{
|
|
/* Here we know that S is a dynamic object FOO.SO.VER1, and
|
|
the object we are considering needs a dynamic object
|
|
FOO.SO.VER2, and VER1 and VER2 are different. This
|
|
appears to be a version mismatch, so we tell the caller
|
|
to try a different version of this library. */
|
|
global_vercheck_failed = TRUE;
|
|
return;
|
|
}
|
|
}
|
|
}
|
|
|
|
|
|
/* See if an input file matches a DT_NEEDED entry by running stat on
|
|
the file. */
|
|
|
|
static void
|
|
ldelf_stat_needed (lang_input_statement_type *s)
|
|
{
|
|
struct stat st;
|
|
const char *suffix;
|
|
const char *soname;
|
|
|
|
if (global_found != NULL)
|
|
return;
|
|
if (s->the_bfd == NULL)
|
|
return;
|
|
|
|
/* If this input file was an as-needed entry, and wasn't found to be
|
|
needed at the stage it was linked, then don't say we have loaded it. */
|
|
if ((bfd_elf_get_dyn_lib_class (s->the_bfd) & DYN_AS_NEEDED) != 0)
|
|
return;
|
|
|
|
if (bfd_stat (s->the_bfd, &st) != 0)
|
|
{
|
|
einfo (_("%P: %pB: bfd_stat failed: %E\n"), s->the_bfd);
|
|
return;
|
|
}
|
|
|
|
/* Some operating systems, e.g. Windows, do not provide a meaningful
|
|
st_ino; they always set it to zero. (Windows does provide a
|
|
meaningful st_dev.) Do not indicate a duplicate library in that
|
|
case. While there is no guarantee that a system that provides
|
|
meaningful inode numbers will never set st_ino to zero, this is
|
|
merely an optimization, so we do not need to worry about false
|
|
negatives. */
|
|
if (st.st_dev == global_stat.st_dev
|
|
&& st.st_ino == global_stat.st_ino
|
|
&& st.st_ino != 0)
|
|
{
|
|
global_found = s;
|
|
return;
|
|
}
|
|
|
|
/* We issue a warning if it looks like we are including two
|
|
different versions of the same shared library. For example,
|
|
there may be a problem if -lc picks up libc.so.6 but some other
|
|
shared library has a DT_NEEDED entry of libc.so.5. This is a
|
|
heuristic test, and it will only work if the name looks like
|
|
NAME.so.VERSION. FIXME: Depending on file names is error-prone.
|
|
If we really want to issue warnings about mixing version numbers
|
|
of shared libraries, we need to find a better way. */
|
|
|
|
if (strchr (global_needed->name, '/') != NULL)
|
|
return;
|
|
suffix = strstr (global_needed->name, ".so.");
|
|
if (suffix == NULL)
|
|
return;
|
|
suffix += sizeof ".so." - 1;
|
|
|
|
soname = bfd_elf_get_dt_soname (s->the_bfd);
|
|
if (soname == NULL)
|
|
soname = lbasename (s->filename);
|
|
|
|
if (filename_ncmp (soname, global_needed->name,
|
|
suffix - global_needed->name) == 0)
|
|
einfo (_("%P: warning: %s, needed by %pB, may conflict with %s\n"),
|
|
global_needed->name, global_needed->by, soname);
|
|
}
|
|
|
|
/* This function is called for each possible name for a dynamic object
|
|
named by a DT_NEEDED entry. The FORCE parameter indicates whether
|
|
to skip the check for a conflicting version. */
|
|
|
|
static bfd_boolean
|
|
ldelf_try_needed (struct dt_needed *needed, int force, int is_linux)
|
|
{
|
|
bfd *abfd;
|
|
const char *name = needed->name;
|
|
const char *soname;
|
|
int link_class;
|
|
|
|
abfd = bfd_openr (name, bfd_get_target (link_info.output_bfd));
|
|
if (abfd == NULL)
|
|
{
|
|
if (verbose)
|
|
info_msg (_("attempt to open %s failed\n"), name);
|
|
return FALSE;
|
|
}
|
|
|
|
/* Linker needs to decompress sections. */
|
|
abfd->flags |= BFD_DECOMPRESS;
|
|
|
|
if (! bfd_check_format (abfd, bfd_object))
|
|
{
|
|
bfd_close (abfd);
|
|
return FALSE;
|
|
}
|
|
if ((bfd_get_file_flags (abfd) & DYNAMIC) == 0)
|
|
{
|
|
bfd_close (abfd);
|
|
return FALSE;
|
|
}
|
|
|
|
/* For DT_NEEDED, they have to match. */
|
|
if (abfd->xvec != link_info.output_bfd->xvec)
|
|
{
|
|
bfd_close (abfd);
|
|
return FALSE;
|
|
}
|
|
|
|
/* Check whether this object would include any conflicting library
|
|
versions. If FORCE is set, then we skip this check; we use this
|
|
the second time around, if we couldn't find any compatible
|
|
instance of the shared library. */
|
|
|
|
if (!force)
|
|
{
|
|
struct bfd_link_needed_list *needs;
|
|
|
|
if (! bfd_elf_get_bfd_needed_list (abfd, &needs))
|
|
einfo (_("%F%P: %pB: bfd_elf_get_bfd_needed_list failed: %E\n"), abfd);
|
|
|
|
if (needs != NULL)
|
|
{
|
|
global_vercheck_needed = needs;
|
|
global_vercheck_failed = FALSE;
|
|
lang_for_each_input_file (ldelf_vercheck);
|
|
if (global_vercheck_failed)
|
|
{
|
|
bfd_close (abfd);
|
|
/* Return FALSE to force the caller to move on to try
|
|
another file on the search path. */
|
|
return FALSE;
|
|
}
|
|
|
|
/* But wait! It gets much worse. On Linux, if a shared
|
|
library does not use libc at all, we are supposed to skip
|
|
it the first time around in case we encounter a shared
|
|
library later on with the same name which does use the
|
|
version of libc that we want. This is much too horrible
|
|
to use on any system other than Linux. */
|
|
if (is_linux)
|
|
{
|
|
struct bfd_link_needed_list *l;
|
|
|
|
for (l = needs; l != NULL; l = l->next)
|
|
if (CONST_STRNEQ (l->name, "libc.so"))
|
|
break;
|
|
if (l == NULL)
|
|
{
|
|
bfd_close (abfd);
|
|
return FALSE;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
/* We've found a dynamic object matching the DT_NEEDED entry. */
|
|
|
|
/* We have already checked that there is no other input file of the
|
|
same name. We must now check again that we are not including the
|
|
same file twice. We need to do this because on many systems
|
|
libc.so is a symlink to, e.g., libc.so.1. The SONAME entry will
|
|
reference libc.so.1. If we have already included libc.so, we
|
|
don't want to include libc.so.1 if they are the same file, and we
|
|
can only check that using stat. */
|
|
|
|
if (bfd_stat (abfd, &global_stat) != 0)
|
|
einfo (_("%F%P: %pB: bfd_stat failed: %E\n"), abfd);
|
|
|
|
/* First strip off everything before the last '/'. */
|
|
soname = lbasename (bfd_get_filename (abfd));
|
|
|
|
if (verbose)
|
|
info_msg (_("found %s at %s\n"), soname, name);
|
|
|
|
global_found = NULL;
|
|
lang_for_each_input_file (ldelf_stat_needed);
|
|
if (global_found != NULL)
|
|
{
|
|
/* Return TRUE to indicate that we found the file, even though
|
|
we aren't going to do anything with it. */
|
|
return TRUE;
|
|
}
|
|
|
|
/* Specify the soname to use. */
|
|
bfd_elf_set_dt_needed_name (abfd, soname);
|
|
|
|
/* Tell the ELF linker that we don't want the output file to have a
|
|
DT_NEEDED entry for this file, unless it is used to resolve
|
|
references in a regular object. */
|
|
link_class = DYN_DT_NEEDED;
|
|
|
|
/* Tell the ELF linker that we don't want the output file to have a
|
|
DT_NEEDED entry for this file at all if the entry is from a file
|
|
with DYN_NO_ADD_NEEDED. */
|
|
if (needed->by != NULL
|
|
&& (bfd_elf_get_dyn_lib_class (needed->by) & DYN_NO_ADD_NEEDED) != 0)
|
|
link_class |= DYN_NO_NEEDED | DYN_NO_ADD_NEEDED;
|
|
|
|
bfd_elf_set_dyn_lib_class (abfd, (enum dynamic_lib_link_class) link_class);
|
|
|
|
/* Add this file into the symbol table. */
|
|
if (! bfd_link_add_symbols (abfd, &link_info))
|
|
einfo (_("%F%P: %pB: error adding symbols: %E\n"), abfd);
|
|
|
|
return TRUE;
|
|
}
|
|
|
|
/* Search for a needed file in a path. */
|
|
|
|
static bfd_boolean
|
|
ldelf_search_needed (const char *path, struct dt_needed *n, int force,
|
|
int is_linux, int elfsize)
|
|
{
|
|
const char *s;
|
|
const char *name = n->name;
|
|
size_t len;
|
|
struct dt_needed needed;
|
|
|
|
if (name[0] == '/')
|
|
return ldelf_try_needed (n, force, is_linux);
|
|
|
|
if (path == NULL || *path == '\0')
|
|
return FALSE;
|
|
|
|
needed.by = n->by;
|
|
needed.name = n->name;
|
|
|
|
len = strlen (name);
|
|
while (1)
|
|
{
|
|
unsigned offset = 0;
|
|
char * var;
|
|
char *filename, *sset;
|
|
|
|
s = strchr (path, config.rpath_separator);
|
|
if (s == NULL)
|
|
s = path + strlen (path);
|
|
|
|
#if HAVE_DOS_BASED_FILE_SYSTEM
|
|
/* Assume a match on the second char is part of drive specifier. */
|
|
else if (config.rpath_separator == ':'
|
|
&& s == path + 1
|
|
&& ISALPHA (*path))
|
|
{
|
|
s = strchr (s + 1, config.rpath_separator);
|
|
if (s == NULL)
|
|
s = path + strlen (path);
|
|
}
|
|
#endif
|
|
filename = (char *) xmalloc (s - path + len + 2);
|
|
if (s == path)
|
|
sset = filename;
|
|
else
|
|
{
|
|
memcpy (filename, path, s - path);
|
|
filename[s - path] = '/';
|
|
sset = filename + (s - path) + 1;
|
|
}
|
|
strcpy (sset, name);
|
|
|
|
/* PR 20535: Support the same pseudo-environment variables that
|
|
are supported by ld.so. Namely, $ORIGIN, $LIB and $PLATFORM.
|
|
Since there can be more than one occurrence of these tokens in
|
|
the path we loop until no more are found. Since we might not
|
|
be able to substitute some of the tokens we maintain an offset
|
|
into the filename for where we should begin our scan. */
|
|
while ((var = strchr (filename + offset, '$')) != NULL)
|
|
{
|
|
/* The ld.so manual page does not say, but I am going to assume that
|
|
these tokens are terminated by a directory separator character
|
|
(/) or the end of the string. There is also an implication that
|
|
$ORIGIN should only be used at the start of a path, but that is
|
|
not enforced here.
|
|
|
|
The ld.so manual page also states that it allows ${ORIGIN},
|
|
${LIB} and ${PLATFORM}, so these are supported as well.
|
|
|
|
FIXME: The code could be a lot cleverer about allocating space
|
|
for the processed string. */
|
|
char * end = strchr (var, '/');
|
|
const char *replacement = NULL;
|
|
char * v = var + 1;
|
|
char * freeme = NULL;
|
|
unsigned flen = strlen (filename);
|
|
|
|
if (end != NULL)
|
|
/* Temporarily terminate the filename at the end of the token. */
|
|
* end = 0;
|
|
|
|
if (*v == '{')
|
|
++ v;
|
|
switch (*v++)
|
|
{
|
|
case 'O':
|
|
if (strcmp (v, "RIGIN") == 0 || strcmp (v, "RIGIN}") == 0)
|
|
{
|
|
/* ORIGIN - replace with the full path to the directory
|
|
containing the program or shared object. */
|
|
if (needed.by == NULL)
|
|
{
|
|
if (link_info.output_bfd == NULL)
|
|
{
|
|
break;
|
|
}
|
|
else
|
|
replacement = bfd_get_filename (link_info.output_bfd);
|
|
}
|
|
else
|
|
replacement = bfd_get_filename (needed.by);
|
|
|
|
if (replacement)
|
|
{
|
|
char * slash;
|
|
|
|
if (replacement[0] == '/')
|
|
freeme = xstrdup (replacement);
|
|
else
|
|
{
|
|
char * current_dir = getpwd ();
|
|
|
|
freeme = xmalloc (strlen (replacement)
|
|
+ strlen (current_dir) + 2);
|
|
sprintf (freeme, "%s/%s", current_dir, replacement);
|
|
}
|
|
|
|
replacement = freeme;
|
|
if ((slash = strrchr (replacement, '/')) != NULL)
|
|
* slash = 0;
|
|
}
|
|
}
|
|
break;
|
|
|
|
case 'L':
|
|
if (strcmp (v, "IB") == 0 || strcmp (v, "IB}") == 0)
|
|
{
|
|
/* LIB - replace with "lib" in 32-bit environments
|
|
and "lib64" in 64-bit environments. */
|
|
|
|
switch (elfsize)
|
|
{
|
|
case 32: replacement = "lib"; break;
|
|
case 64: replacement = "lib64"; break;
|
|
default:
|
|
abort ();
|
|
}
|
|
}
|
|
break;
|
|
|
|
case 'P':
|
|
/* Supporting $PLATFORM in a cross-hosted environment is not
|
|
possible. Supporting it in a native environment involves
|
|
loading the <sys/auxv.h> header file which loads the
|
|
system <elf.h> header file, which conflicts with the
|
|
"include/elf/mips.h" header file. */
|
|
/* Fall through. */
|
|
default:
|
|
break;
|
|
}
|
|
|
|
if (replacement)
|
|
{
|
|
char * filename2 = xmalloc (flen + strlen (replacement));
|
|
|
|
if (end)
|
|
{
|
|
sprintf (filename2, "%.*s%s/%s",
|
|
(int)(var - filename), filename,
|
|
replacement, end + 1);
|
|
offset = (var - filename) + 1 + strlen (replacement);
|
|
}
|
|
else
|
|
{
|
|
sprintf (filename2, "%.*s%s",
|
|
(int)(var - filename), filename,
|
|
replacement);
|
|
offset = var - filename + strlen (replacement);
|
|
}
|
|
|
|
free (filename);
|
|
filename = filename2;
|
|
/* There is no need to restore the path separator (when
|
|
end != NULL) as we have replaced the entire string. */
|
|
}
|
|
else
|
|
{
|
|
if (verbose)
|
|
/* We only issue an "unrecognised" message in verbose mode
|
|
as the $<foo> token might be a legitimate component of
|
|
a path name in the target's file system. */
|
|
info_msg (_("unrecognised or unsupported token "
|
|
"'%s' in search path\n"), var);
|
|
if (end)
|
|
/* Restore the path separator. */
|
|
* end = '/';
|
|
|
|
/* PR 20784: Make sure that we resume the scan *after*
|
|
the token that we could not replace. */
|
|
offset = (var + 1) - filename;
|
|
}
|
|
|
|
free (freeme);
|
|
}
|
|
|
|
needed.name = filename;
|
|
|
|
if (ldelf_try_needed (&needed, force, is_linux))
|
|
return TRUE;
|
|
|
|
free (filename);
|
|
|
|
if (*s == '\0')
|
|
break;
|
|
path = s + 1;
|
|
}
|
|
|
|
return FALSE;
|
|
}
|
|
|
|
/* Prefix the sysroot to absolute paths in PATH, a string containing
|
|
paths separated by config.rpath_separator. If running on a DOS
|
|
file system, paths containing a drive spec won't have the sysroot
|
|
prefix added, unless the sysroot also specifies the same drive. */
|
|
|
|
static const char *
|
|
ldelf_add_sysroot (const char *path)
|
|
{
|
|
size_t len, extra;
|
|
const char *p;
|
|
char *ret, *q;
|
|
int dos_drive_sysroot = HAS_DRIVE_SPEC (ld_sysroot);
|
|
|
|
len = strlen (ld_sysroot);
|
|
for (extra = 0, p = path; ; )
|
|
{
|
|
int dos_drive = HAS_DRIVE_SPEC (p);
|
|
|
|
if (dos_drive)
|
|
p += 2;
|
|
if (IS_DIR_SEPARATOR (*p)
|
|
&& (!dos_drive
|
|
|| (dos_drive_sysroot
|
|
&& ld_sysroot[0] == p[-2])))
|
|
{
|
|
if (dos_drive && dos_drive_sysroot)
|
|
extra += len - 2;
|
|
else
|
|
extra += len;
|
|
}
|
|
p = strchr (p, config.rpath_separator);
|
|
if (!p)
|
|
break;
|
|
++p;
|
|
}
|
|
|
|
ret = xmalloc (strlen (path) + extra + 1);
|
|
|
|
for (q = ret, p = path; ; )
|
|
{
|
|
const char *end;
|
|
int dos_drive = HAS_DRIVE_SPEC (p);
|
|
|
|
if (dos_drive)
|
|
{
|
|
*q++ = *p++;
|
|
*q++ = *p++;
|
|
}
|
|
if (IS_DIR_SEPARATOR (*p)
|
|
&& (!dos_drive
|
|
|| (dos_drive_sysroot
|
|
&& ld_sysroot[0] == p[-2])))
|
|
{
|
|
if (dos_drive && dos_drive_sysroot)
|
|
{
|
|
strcpy (q, ld_sysroot + 2);
|
|
q += len - 2;
|
|
}
|
|
else
|
|
{
|
|
strcpy (q, ld_sysroot);
|
|
q += len;
|
|
}
|
|
}
|
|
end = strchr (p, config.rpath_separator);
|
|
if (end)
|
|
{
|
|
size_t n = end - p + 1;
|
|
strncpy (q, p, n);
|
|
q += n;
|
|
p += n;
|
|
}
|
|
else
|
|
{
|
|
strcpy (q, p);
|
|
break;
|
|
}
|
|
}
|
|
|
|
return ret;
|
|
}
|
|
|
|
/* Read the system search path the FreeBSD way rather than the Linux way. */
|
|
#ifdef HAVE_ELF_HINTS_H
|
|
#include <elf-hints.h>
|
|
#else
|
|
#include "elf-hints-local.h"
|
|
#endif
|
|
|
|
static bfd_boolean
|
|
ldelf_check_ld_elf_hints (const struct bfd_link_needed_list *l, int force,
|
|
int elfsize)
|
|
{
|
|
static bfd_boolean initialized;
|
|
static const char *ld_elf_hints;
|
|
struct dt_needed needed;
|
|
|
|
if (!initialized)
|
|
{
|
|
FILE *f;
|
|
char *tmppath;
|
|
|
|
tmppath = concat (ld_sysroot, _PATH_ELF_HINTS, (const char *) NULL);
|
|
f = fopen (tmppath, FOPEN_RB);
|
|
free (tmppath);
|
|
if (f != NULL)
|
|
{
|
|
struct elfhints_hdr hdr;
|
|
|
|
if (fread (&hdr, 1, sizeof (hdr), f) == sizeof (hdr)
|
|
&& hdr.magic == ELFHINTS_MAGIC
|
|
&& hdr.version == 1)
|
|
{
|
|
if (fseek (f, hdr.strtab + hdr.dirlist, SEEK_SET) != -1)
|
|
{
|
|
char *b;
|
|
|
|
b = xmalloc (hdr.dirlistlen + 1);
|
|
if (fread (b, 1, hdr.dirlistlen + 1, f) ==
|
|
hdr.dirlistlen + 1)
|
|
ld_elf_hints = ldelf_add_sysroot (b);
|
|
|
|
free (b);
|
|
}
|
|
}
|
|
fclose (f);
|
|
}
|
|
|
|
initialized = TRUE;
|
|
}
|
|
|
|
if (ld_elf_hints == NULL)
|
|
return FALSE;
|
|
|
|
needed.by = l->by;
|
|
needed.name = l->name;
|
|
return ldelf_search_needed (ld_elf_hints, &needed, force, FALSE, elfsize);
|
|
}
|
|
|
|
/* For a native linker, check the file /etc/ld.so.conf for directories
|
|
in which we may find shared libraries. /etc/ld.so.conf is really
|
|
only meaningful on Linux. */
|
|
|
|
struct ldelf_ld_so_conf
|
|
{
|
|
char *path;
|
|
size_t len, alloc;
|
|
};
|
|
|
|
static bfd_boolean
|
|
ldelf_parse_ld_so_conf (struct ldelf_ld_so_conf *, const char *);
|
|
|
|
static void
|
|
ldelf_parse_ld_so_conf_include (struct ldelf_ld_so_conf *info,
|
|
const char *filename,
|
|
const char *pattern)
|
|
{
|
|
char *newp = NULL;
|
|
#ifdef HAVE_GLOB
|
|
glob_t gl;
|
|
#endif
|
|
|
|
if (pattern[0] != '/')
|
|
{
|
|
char *p = strrchr (filename, '/');
|
|
size_t patlen = strlen (pattern) + 1;
|
|
|
|
newp = xmalloc (p - filename + 1 + patlen);
|
|
memcpy (newp, filename, p - filename + 1);
|
|
memcpy (newp + (p - filename + 1), pattern, patlen);
|
|
pattern = newp;
|
|
}
|
|
|
|
#ifdef HAVE_GLOB
|
|
if (glob (pattern, 0, NULL, &gl) == 0)
|
|
{
|
|
size_t i;
|
|
|
|
for (i = 0; i < gl.gl_pathc; ++i)
|
|
ldelf_parse_ld_so_conf (info, gl.gl_pathv[i]);
|
|
globfree (&gl);
|
|
}
|
|
#else
|
|
/* If we do not have glob, treat the pattern as a literal filename. */
|
|
ldelf_parse_ld_so_conf (info, pattern);
|
|
#endif
|
|
|
|
if (newp)
|
|
free (newp);
|
|
}
|
|
|
|
static bfd_boolean
|
|
ldelf_parse_ld_so_conf (struct ldelf_ld_so_conf *info, const char *filename)
|
|
{
|
|
FILE *f = fopen (filename, FOPEN_RT);
|
|
char *line;
|
|
size_t linelen;
|
|
|
|
if (f == NULL)
|
|
return FALSE;
|
|
|
|
linelen = 256;
|
|
line = xmalloc (linelen);
|
|
do
|
|
{
|
|
char *p = line, *q;
|
|
|
|
/* Normally this would use getline(3), but we need to be portable. */
|
|
while ((q = fgets (p, linelen - (p - line), f)) != NULL
|
|
&& strlen (q) == linelen - (p - line) - 1
|
|
&& line[linelen - 2] != '\n')
|
|
{
|
|
line = xrealloc (line, 2 * linelen);
|
|
p = line + linelen - 1;
|
|
linelen += linelen;
|
|
}
|
|
|
|
if (q == NULL && p == line)
|
|
break;
|
|
|
|
p = strchr (line, '\n');
|
|
if (p)
|
|
*p = '\0';
|
|
|
|
/* Because the file format does not know any form of quoting we
|
|
can search forward for the next '#' character and if found
|
|
make it terminating the line. */
|
|
p = strchr (line, '#');
|
|
if (p)
|
|
*p = '\0';
|
|
|
|
/* Remove leading whitespace. NUL is no whitespace character. */
|
|
p = line;
|
|
while (*p == ' ' || *p == '\f' || *p == '\r' || *p == '\t' || *p == '\v')
|
|
++p;
|
|
|
|
/* If the line is blank it is ignored. */
|
|
if (p[0] == '\0')
|
|
continue;
|
|
|
|
if (CONST_STRNEQ (p, "include") && (p[7] == ' ' || p[7] == '\t'))
|
|
{
|
|
char *dir, c;
|
|
p += 8;
|
|
do
|
|
{
|
|
while (*p == ' ' || *p == '\t')
|
|
++p;
|
|
|
|
if (*p == '\0')
|
|
break;
|
|
|
|
dir = p;
|
|
|
|
while (*p != ' ' && *p != '\t' && *p)
|
|
++p;
|
|
|
|
c = *p;
|
|
*p++ = '\0';
|
|
if (dir[0] != '\0')
|
|
ldelf_parse_ld_so_conf_include (info, filename, dir);
|
|
}
|
|
while (c != '\0');
|
|
}
|
|
else
|
|
{
|
|
char *dir = p;
|
|
while (*p && *p != '=' && *p != ' ' && *p != '\t' && *p != '\f'
|
|
&& *p != '\r' && *p != '\v')
|
|
++p;
|
|
|
|
while (p != dir && p[-1] == '/')
|
|
--p;
|
|
if (info->path == NULL)
|
|
{
|
|
info->alloc = p - dir + 1 + 256;
|
|
info->path = xmalloc (info->alloc);
|
|
info->len = 0;
|
|
}
|
|
else
|
|
{
|
|
if (info->len + 1 + (p - dir) >= info->alloc)
|
|
{
|
|
info->alloc += p - dir + 256;
|
|
info->path = xrealloc (info->path, info->alloc);
|
|
}
|
|
info->path[info->len++] = config.rpath_separator;
|
|
}
|
|
memcpy (info->path + info->len, dir, p - dir);
|
|
info->len += p - dir;
|
|
info->path[info->len] = '\0';
|
|
}
|
|
}
|
|
while (! feof (f));
|
|
free (line);
|
|
fclose (f);
|
|
return TRUE;
|
|
}
|
|
|
|
static bfd_boolean
|
|
ldelf_check_ld_so_conf (const struct bfd_link_needed_list *l, int force,
|
|
int elfsize)
|
|
{
|
|
static bfd_boolean initialized;
|
|
static const char *ld_so_conf;
|
|
struct dt_needed needed;
|
|
|
|
if (! initialized)
|
|
{
|
|
char *tmppath;
|
|
struct ldelf_ld_so_conf info;
|
|
|
|
info.path = NULL;
|
|
info.len = info.alloc = 0;
|
|
tmppath = concat (ld_sysroot, "${prefix}/etc/ld.so.conf",
|
|
(const char *) NULL);
|
|
if (!ldelf_parse_ld_so_conf (&info, tmppath))
|
|
{
|
|
free (tmppath);
|
|
tmppath = concat (ld_sysroot, "/etc/ld.so.conf",
|
|
(const char *) NULL);
|
|
ldelf_parse_ld_so_conf (&info, tmppath);
|
|
}
|
|
free (tmppath);
|
|
|
|
if (info.path)
|
|
{
|
|
ld_so_conf = ldelf_add_sysroot (info.path);
|
|
free (info.path);
|
|
}
|
|
initialized = TRUE;
|
|
}
|
|
|
|
if (ld_so_conf == NULL)
|
|
return FALSE;
|
|
|
|
|
|
needed.by = l->by;
|
|
needed.name = l->name;
|
|
return ldelf_search_needed (ld_so_conf, &needed, force, TRUE, elfsize);
|
|
}
|
|
|
|
/* See if an input file matches a DT_NEEDED entry by name. */
|
|
|
|
static void
|
|
ldelf_check_needed (lang_input_statement_type *s)
|
|
{
|
|
const char *soname;
|
|
|
|
/* Stop looking if we've found a loaded lib. */
|
|
if (global_found != NULL
|
|
&& (bfd_elf_get_dyn_lib_class (global_found->the_bfd)
|
|
& DYN_AS_NEEDED) == 0)
|
|
return;
|
|
|
|
if (s->filename == NULL || s->the_bfd == NULL)
|
|
return;
|
|
|
|
/* Don't look for a second non-loaded as-needed lib. */
|
|
if (global_found != NULL
|
|
&& (bfd_elf_get_dyn_lib_class (s->the_bfd) & DYN_AS_NEEDED) != 0)
|
|
return;
|
|
|
|
if (filename_cmp (s->filename, global_needed->name) == 0)
|
|
{
|
|
global_found = s;
|
|
return;
|
|
}
|
|
|
|
if (s->flags.search_dirs)
|
|
{
|
|
const char *f = strrchr (s->filename, '/');
|
|
if (f != NULL
|
|
&& filename_cmp (f + 1, global_needed->name) == 0)
|
|
{
|
|
global_found = s;
|
|
return;
|
|
}
|
|
}
|
|
|
|
soname = bfd_elf_get_dt_soname (s->the_bfd);
|
|
if (soname != NULL
|
|
&& filename_cmp (soname, global_needed->name) == 0)
|
|
{
|
|
global_found = s;
|
|
return;
|
|
}
|
|
}
|
|
|
|
/* This is called after all the input files have been opened. */
|
|
|
|
void
|
|
ldelf_after_open (int use_libpath, int native, int is_linux, int is_freebsd,
|
|
int elfsize)
|
|
{
|
|
struct bfd_link_needed_list *needed, *l;
|
|
struct elf_link_hash_table *htab;
|
|
asection *s;
|
|
bfd *abfd;
|
|
|
|
after_open_default ();
|
|
|
|
htab = elf_hash_table (&link_info);
|
|
if (!is_elf_hash_table (htab))
|
|
return;
|
|
|
|
if (command_line.out_implib_filename)
|
|
{
|
|
unlink_if_ordinary (command_line.out_implib_filename);
|
|
link_info.out_implib_bfd
|
|
= bfd_openw (command_line.out_implib_filename,
|
|
bfd_get_target (link_info.output_bfd));
|
|
|
|
if (link_info.out_implib_bfd == NULL)
|
|
{
|
|
einfo (_("%F%P: %s: can't open for writing: %E\n"),
|
|
command_line.out_implib_filename);
|
|
}
|
|
}
|
|
|
|
if (ldelf_emit_note_gnu_build_id != NULL)
|
|
{
|
|
/* Find an ELF input. */
|
|
for (abfd = link_info.input_bfds;
|
|
abfd != (bfd *) NULL; abfd = abfd->link.next)
|
|
if (bfd_get_flavour (abfd) == bfd_target_elf_flavour
|
|
&& bfd_count_sections (abfd) != 0
|
|
&& !bfd_input_just_syms (abfd))
|
|
break;
|
|
|
|
/* PR 10555: If there are no ELF input files do not try to
|
|
create a .note.gnu-build-id section. */
|
|
if (abfd == NULL
|
|
|| !ldelf_setup_build_id (abfd))
|
|
{
|
|
free ((char *) ldelf_emit_note_gnu_build_id);
|
|
ldelf_emit_note_gnu_build_id = NULL;
|
|
}
|
|
}
|
|
|
|
get_elf_backend_data (link_info.output_bfd)->setup_gnu_properties (&link_info);
|
|
|
|
if (bfd_link_relocatable (&link_info))
|
|
{
|
|
if (link_info.execstack == !link_info.noexecstack)
|
|
{
|
|
/* PR ld/16744: If "-z [no]execstack" has been specified on the
|
|
command line and we are perfoming a relocatable link then no
|
|
PT_GNU_STACK segment will be created and so the
|
|
linkinfo.[no]execstack values set in _handle_option() will have no
|
|
effect. Instead we create a .note.GNU-stack section in much the
|
|
same way as the assembler does with its --[no]execstack option. */
|
|
flagword flags = SEC_READONLY | (link_info.execstack ? SEC_CODE : 0);
|
|
(void) bfd_make_section_with_flags (link_info.input_bfds,
|
|
".note.GNU-stack", flags);
|
|
}
|
|
return;
|
|
}
|
|
|
|
if (!link_info.traditional_format)
|
|
{
|
|
bfd *elfbfd = NULL;
|
|
bfd_boolean warn_eh_frame = FALSE;
|
|
int seen_type = 0;
|
|
|
|
for (abfd = link_info.input_bfds; abfd; abfd = abfd->link.next)
|
|
{
|
|
int type = 0;
|
|
|
|
if (bfd_input_just_syms (abfd))
|
|
continue;
|
|
|
|
for (s = abfd->sections; s && type < COMPACT_EH_HDR; s = s->next)
|
|
{
|
|
const char *name = bfd_section_name (s);
|
|
|
|
if (bfd_is_abs_section (s->output_section))
|
|
continue;
|
|
if (CONST_STRNEQ (name, ".eh_frame_entry"))
|
|
type = COMPACT_EH_HDR;
|
|
else if (strcmp (name, ".eh_frame") == 0 && s->size > 8)
|
|
type = DWARF2_EH_HDR;
|
|
}
|
|
|
|
if (type != 0)
|
|
{
|
|
if (seen_type == 0)
|
|
{
|
|
seen_type = type;
|
|
}
|
|
else if (seen_type != type)
|
|
{
|
|
einfo (_("%F%P: compact frame descriptions incompatible with"
|
|
" DWARF2 .eh_frame from %pB\n"),
|
|
type == DWARF2_EH_HDR ? abfd : elfbfd);
|
|
break;
|
|
}
|
|
|
|
if (!elfbfd
|
|
&& (type == COMPACT_EH_HDR
|
|
|| link_info.eh_frame_hdr_type != 0))
|
|
{
|
|
if (bfd_get_flavour (abfd) == bfd_target_elf_flavour)
|
|
elfbfd = abfd;
|
|
|
|
warn_eh_frame = TRUE;
|
|
}
|
|
}
|
|
|
|
if (seen_type == COMPACT_EH_HDR)
|
|
link_info.eh_frame_hdr_type = COMPACT_EH_HDR;
|
|
}
|
|
if (elfbfd)
|
|
{
|
|
const struct elf_backend_data *bed;
|
|
|
|
bed = get_elf_backend_data (elfbfd);
|
|
s = bfd_make_section_with_flags (elfbfd, ".eh_frame_hdr",
|
|
bed->dynamic_sec_flags
|
|
| SEC_READONLY);
|
|
if (s != NULL
|
|
&& bfd_set_section_alignment (s, 2))
|
|
{
|
|
htab->eh_info.hdr_sec = s;
|
|
warn_eh_frame = FALSE;
|
|
}
|
|
}
|
|
if (warn_eh_frame)
|
|
einfo (_("%P: warning: cannot create .eh_frame_hdr section,"
|
|
" --eh-frame-hdr ignored\n"));
|
|
}
|
|
|
|
/* Get the list of files which appear in DT_NEEDED entries in
|
|
dynamic objects included in the link (often there will be none).
|
|
For each such file, we want to track down the corresponding
|
|
library, and include the symbol table in the link. This is what
|
|
the runtime dynamic linker will do. Tracking the files down here
|
|
permits one dynamic object to include another without requiring
|
|
special action by the person doing the link. Note that the
|
|
needed list can actually grow while we are stepping through this
|
|
loop. */
|
|
needed = bfd_elf_get_needed_list (link_info.output_bfd, &link_info);
|
|
for (l = needed; l != NULL; l = l->next)
|
|
{
|
|
struct bfd_link_needed_list *ll;
|
|
struct dt_needed n, nn;
|
|
int force;
|
|
|
|
/* If the lib that needs this one was --as-needed and wasn't
|
|
found to be needed, then this lib isn't needed either. */
|
|
if (l->by != NULL
|
|
&& (bfd_elf_get_dyn_lib_class (l->by) & DYN_AS_NEEDED) != 0)
|
|
continue;
|
|
|
|
/* Skip the lib if --no-copy-dt-needed-entries and
|
|
--allow-shlib-undefined is in effect. */
|
|
if (l->by != NULL
|
|
&& link_info.unresolved_syms_in_shared_libs == RM_IGNORE
|
|
&& (bfd_elf_get_dyn_lib_class (l->by) & DYN_NO_ADD_NEEDED) != 0)
|
|
continue;
|
|
|
|
/* If we've already seen this file, skip it. */
|
|
for (ll = needed; ll != l; ll = ll->next)
|
|
if ((ll->by == NULL
|
|
|| (bfd_elf_get_dyn_lib_class (ll->by) & DYN_AS_NEEDED) == 0)
|
|
&& strcmp (ll->name, l->name) == 0)
|
|
break;
|
|
if (ll != l)
|
|
continue;
|
|
|
|
/* See if this file was included in the link explicitly. */
|
|
global_needed = l;
|
|
global_found = NULL;
|
|
lang_for_each_input_file (ldelf_check_needed);
|
|
if (global_found != NULL
|
|
&& (bfd_elf_get_dyn_lib_class (global_found->the_bfd)
|
|
& DYN_AS_NEEDED) == 0)
|
|
continue;
|
|
|
|
n.by = l->by;
|
|
n.name = l->name;
|
|
nn.by = l->by;
|
|
if (verbose)
|
|
info_msg (_("%s needed by %pB\n"), l->name, l->by);
|
|
|
|
/* As-needed libs specified on the command line (or linker script)
|
|
take priority over libs found in search dirs. */
|
|
if (global_found != NULL)
|
|
{
|
|
nn.name = global_found->filename;
|
|
if (ldelf_try_needed (&nn, TRUE, is_linux))
|
|
continue;
|
|
}
|
|
|
|
/* We need to find this file and include the symbol table. We
|
|
want to search for the file in the same way that the dynamic
|
|
linker will search. That means that we want to use
|
|
rpath_link, rpath, then the environment variable
|
|
LD_LIBRARY_PATH (native only), then the DT_RPATH/DT_RUNPATH
|
|
entries (native only), then the linker script LIB_SEARCH_DIRS.
|
|
We do not search using the -L arguments.
|
|
|
|
We search twice. The first time, we skip objects which may
|
|
introduce version mismatches. The second time, we force
|
|
their use. See ldelf_vercheck comment. */
|
|
for (force = 0; force < 2; force++)
|
|
{
|
|
size_t len;
|
|
search_dirs_type *search;
|
|
const char *path;
|
|
struct bfd_link_needed_list *rp;
|
|
int found;
|
|
|
|
if (ldelf_search_needed (command_line.rpath_link, &n, force,
|
|
is_linux, elfsize))
|
|
break;
|
|
|
|
if (use_libpath)
|
|
{
|
|
path = command_line.rpath;
|
|
if (path)
|
|
{
|
|
path = ldelf_add_sysroot (path);
|
|
found = ldelf_search_needed (path, &n, force,
|
|
is_linux, elfsize);
|
|
free ((char *) path);
|
|
if (found)
|
|
break;
|
|
}
|
|
}
|
|
if (native)
|
|
{
|
|
if (command_line.rpath_link == NULL
|
|
&& command_line.rpath == NULL)
|
|
{
|
|
path = (const char *) getenv ("LD_RUN_PATH");
|
|
if (path
|
|
&& ldelf_search_needed (path, &n, force,
|
|
is_linux, elfsize))
|
|
break;
|
|
}
|
|
path = (const char *) getenv ("LD_LIBRARY_PATH");
|
|
if (path
|
|
&& ldelf_search_needed (path, &n, force,
|
|
is_linux, elfsize))
|
|
break;
|
|
}
|
|
if (use_libpath)
|
|
{
|
|
found = 0;
|
|
rp = bfd_elf_get_runpath_list (link_info.output_bfd, &link_info);
|
|
for (; !found && rp != NULL; rp = rp->next)
|
|
{
|
|
path = ldelf_add_sysroot (rp->name);
|
|
found = (rp->by == l->by
|
|
&& ldelf_search_needed (path, &n, force,
|
|
is_linux, elfsize));
|
|
free ((char *) path);
|
|
}
|
|
if (found)
|
|
break;
|
|
|
|
if (is_freebsd
|
|
&& ldelf_check_ld_elf_hints (l, force, elfsize))
|
|
break;
|
|
|
|
if (is_linux
|
|
&& ldelf_check_ld_so_conf (l, force, elfsize))
|
|
break;
|
|
}
|
|
|
|
len = strlen (l->name);
|
|
for (search = search_head; search != NULL; search = search->next)
|
|
{
|
|
char *filename;
|
|
|
|
if (search->cmdline)
|
|
continue;
|
|
filename = (char *) xmalloc (strlen (search->name) + len + 2);
|
|
sprintf (filename, "%s/%s", search->name, l->name);
|
|
nn.name = filename;
|
|
if (ldelf_try_needed (&nn, force, is_linux))
|
|
break;
|
|
free (filename);
|
|
}
|
|
if (search != NULL)
|
|
break;
|
|
}
|
|
|
|
if (force < 2)
|
|
continue;
|
|
|
|
einfo (_("%P: warning: %s, needed by %pB, not found "
|
|
"(try using -rpath or -rpath-link)\n"),
|
|
l->name, l->by);
|
|
}
|
|
|
|
if (link_info.eh_frame_hdr_type == COMPACT_EH_HDR)
|
|
if (!bfd_elf_parse_eh_frame_entries (NULL, &link_info))
|
|
einfo (_("%F%P: failed to parse EH frame entries\n"));
|
|
}
|
|
|
|
static bfd_size_type
|
|
id_note_section_size (bfd *abfd ATTRIBUTE_UNUSED)
|
|
{
|
|
const char *style = ldelf_emit_note_gnu_build_id;
|
|
bfd_size_type size;
|
|
bfd_size_type build_id_size;
|
|
|
|
size = offsetof (Elf_External_Note, name[sizeof "GNU"]);
|
|
size = (size + 3) & -(bfd_size_type) 4;
|
|
|
|
build_id_size = compute_build_id_size (style);
|
|
if (build_id_size)
|
|
size += build_id_size;
|
|
else
|
|
size = 0;
|
|
|
|
return size;
|
|
}
|
|
|
|
static bfd_boolean
|
|
write_build_id (bfd *abfd)
|
|
{
|
|
const struct elf_backend_data *bed = get_elf_backend_data (abfd);
|
|
struct elf_obj_tdata *t = elf_tdata (abfd);
|
|
const char *style;
|
|
asection *asec;
|
|
Elf_Internal_Shdr *i_shdr;
|
|
unsigned char *contents, *id_bits;
|
|
bfd_size_type size;
|
|
file_ptr position;
|
|
Elf_External_Note *e_note;
|
|
|
|
style = t->o->build_id.style;
|
|
asec = t->o->build_id.sec;
|
|
if (bfd_is_abs_section (asec->output_section))
|
|
{
|
|
einfo (_("%P: warning: .note.gnu.build-id section discarded,"
|
|
" --build-id ignored\n"));
|
|
return TRUE;
|
|
}
|
|
i_shdr = &elf_section_data (asec->output_section)->this_hdr;
|
|
|
|
if (i_shdr->contents == NULL)
|
|
{
|
|
if (asec->contents == NULL)
|
|
asec->contents = (unsigned char *) xmalloc (asec->size);
|
|
contents = asec->contents;
|
|
}
|
|
else
|
|
contents = i_shdr->contents + asec->output_offset;
|
|
|
|
e_note = (Elf_External_Note *) contents;
|
|
size = offsetof (Elf_External_Note, name[sizeof "GNU"]);
|
|
size = (size + 3) & -(bfd_size_type) 4;
|
|
id_bits = contents + size;
|
|
size = asec->size - size;
|
|
|
|
bfd_h_put_32 (abfd, sizeof "GNU", &e_note->namesz);
|
|
bfd_h_put_32 (abfd, size, &e_note->descsz);
|
|
bfd_h_put_32 (abfd, NT_GNU_BUILD_ID, &e_note->type);
|
|
memcpy (e_note->name, "GNU", sizeof "GNU");
|
|
|
|
generate_build_id (abfd, style, bed->s->checksum_contents, id_bits, size);
|
|
|
|
position = i_shdr->sh_offset + asec->output_offset;
|
|
size = asec->size;
|
|
return (bfd_seek (abfd, position, SEEK_SET) == 0
|
|
&& bfd_bwrite (contents, size, abfd) == size);
|
|
}
|
|
|
|
/* Make .note.gnu.build-id section, and set up elf_tdata->build_id. */
|
|
|
|
bfd_boolean
|
|
ldelf_setup_build_id (bfd *ibfd)
|
|
{
|
|
asection *s;
|
|
bfd_size_type size;
|
|
flagword flags;
|
|
|
|
size = id_note_section_size (ibfd);
|
|
if (size == 0)
|
|
{
|
|
einfo (_("%P: warning: unrecognized --build-id style ignored\n"));
|
|
return FALSE;
|
|
}
|
|
|
|
flags = (SEC_ALLOC | SEC_LOAD | SEC_IN_MEMORY
|
|
| SEC_LINKER_CREATED | SEC_READONLY | SEC_DATA);
|
|
s = bfd_make_section_with_flags (ibfd, ".note.gnu.build-id", flags);
|
|
if (s != NULL && bfd_set_section_alignment (s, 2))
|
|
{
|
|
struct elf_obj_tdata *t = elf_tdata (link_info.output_bfd);
|
|
t->o->build_id.after_write_object_contents = &write_build_id;
|
|
t->o->build_id.style = ldelf_emit_note_gnu_build_id;
|
|
t->o->build_id.sec = s;
|
|
elf_section_type (s) = SHT_NOTE;
|
|
s->size = size;
|
|
return TRUE;
|
|
}
|
|
|
|
einfo (_("%P: warning: cannot create .note.gnu.build-id section,"
|
|
" --build-id ignored\n"));
|
|
return FALSE;
|
|
}
|
|
|
|
/* Look through an expression for an assignment statement. */
|
|
|
|
static void
|
|
ldelf_find_exp_assignment (etree_type *exp)
|
|
{
|
|
bfd_boolean provide = FALSE;
|
|
|
|
switch (exp->type.node_class)
|
|
{
|
|
case etree_provide:
|
|
case etree_provided:
|
|
provide = TRUE;
|
|
/* Fallthru */
|
|
case etree_assign:
|
|
/* We call record_link_assignment even if the symbol is defined.
|
|
This is because if it is defined by a dynamic object, we
|
|
actually want to use the value defined by the linker script,
|
|
not the value from the dynamic object (because we are setting
|
|
symbols like etext). If the symbol is defined by a regular
|
|
object, then, as it happens, calling record_link_assignment
|
|
will do no harm. */
|
|
if (strcmp (exp->assign.dst, ".") != 0)
|
|
{
|
|
if (!bfd_elf_record_link_assignment (link_info.output_bfd,
|
|
&link_info,
|
|
exp->assign.dst, provide,
|
|
exp->assign.hidden))
|
|
einfo (_("%F%P: failed to record assignment to %s: %E\n"),
|
|
exp->assign.dst);
|
|
}
|
|
ldelf_find_exp_assignment (exp->assign.src);
|
|
break;
|
|
|
|
case etree_binary:
|
|
ldelf_find_exp_assignment (exp->binary.lhs);
|
|
ldelf_find_exp_assignment (exp->binary.rhs);
|
|
break;
|
|
|
|
case etree_trinary:
|
|
ldelf_find_exp_assignment (exp->trinary.cond);
|
|
ldelf_find_exp_assignment (exp->trinary.lhs);
|
|
ldelf_find_exp_assignment (exp->trinary.rhs);
|
|
break;
|
|
|
|
case etree_unary:
|
|
ldelf_find_exp_assignment (exp->unary.child);
|
|
break;
|
|
|
|
default:
|
|
break;
|
|
}
|
|
}
|
|
|
|
/* This is called by the before_allocation routine via
|
|
lang_for_each_statement. It locates any assignment statements, and
|
|
tells the ELF backend about them, in case they are assignments to
|
|
symbols which are referred to by dynamic objects. */
|
|
|
|
static void
|
|
ldelf_find_statement_assignment (lang_statement_union_type *s)
|
|
{
|
|
if (s->header.type == lang_assignment_statement_enum)
|
|
ldelf_find_exp_assignment (s->assignment_statement.exp);
|
|
}
|
|
|
|
/* Used by before_allocation and handle_option. */
|
|
|
|
void
|
|
ldelf_append_to_separated_string (char **to, char *op_arg)
|
|
{
|
|
if (*to == NULL)
|
|
*to = xstrdup (op_arg);
|
|
else
|
|
{
|
|
size_t to_len = strlen (*to);
|
|
size_t op_arg_len = strlen (op_arg);
|
|
char *buf;
|
|
char *cp = *to;
|
|
|
|
/* First see whether OPTARG is already in the path. */
|
|
do
|
|
{
|
|
if (strncmp (op_arg, cp, op_arg_len) == 0
|
|
&& (cp[op_arg_len] == 0
|
|
|| cp[op_arg_len] == config.rpath_separator))
|
|
/* We found it. */
|
|
break;
|
|
|
|
/* Not yet found. */
|
|
cp = strchr (cp, config.rpath_separator);
|
|
if (cp != NULL)
|
|
++cp;
|
|
}
|
|
while (cp != NULL);
|
|
|
|
if (cp == NULL)
|
|
{
|
|
buf = xmalloc (to_len + op_arg_len + 2);
|
|
sprintf (buf, "%s%c%s", *to,
|
|
config.rpath_separator, op_arg);
|
|
free (*to);
|
|
*to = buf;
|
|
}
|
|
}
|
|
}
|
|
|
|
/* This is called after the sections have been attached to output
|
|
sections, but before any sizes or addresses have been set. */
|
|
|
|
void
|
|
ldelf_before_allocation (char *audit, char *depaudit,
|
|
const char *default_interpreter_name)
|
|
{
|
|
const char *rpath;
|
|
asection *sinterp;
|
|
bfd *abfd;
|
|
struct bfd_link_hash_entry *ehdr_start = NULL;
|
|
unsigned char ehdr_start_save_type = 0;
|
|
char ehdr_start_save_u[sizeof ehdr_start->u
|
|
- sizeof ehdr_start->u.def.next] = "";
|
|
|
|
if (is_elf_hash_table (link_info.hash))
|
|
{
|
|
_bfd_elf_tls_setup (link_info.output_bfd, &link_info);
|
|
|
|
/* Make __ehdr_start hidden if it has been referenced, to
|
|
prevent the symbol from being dynamic. */
|
|
if (!bfd_link_relocatable (&link_info))
|
|
{
|
|
struct elf_link_hash_table *htab = elf_hash_table (&link_info);
|
|
struct elf_link_hash_entry *h
|
|
= elf_link_hash_lookup (htab, "__ehdr_start", FALSE, FALSE, TRUE);
|
|
|
|
/* Only adjust the export class if the symbol was referenced
|
|
and not defined, otherwise leave it alone. */
|
|
if (h != NULL
|
|
&& (h->root.type == bfd_link_hash_new
|
|
|| h->root.type == bfd_link_hash_undefined
|
|
|| h->root.type == bfd_link_hash_undefweak
|
|
|| h->root.type == bfd_link_hash_common))
|
|
{
|
|
const struct elf_backend_data *bed;
|
|
bed = get_elf_backend_data (link_info.output_bfd);
|
|
(*bed->elf_backend_hide_symbol) (&link_info, h, TRUE);
|
|
if (ELF_ST_VISIBILITY (h->other) != STV_INTERNAL)
|
|
h->other = (h->other & ~ELF_ST_VISIBILITY (-1)) | STV_HIDDEN;
|
|
/* Don't leave the symbol undefined. Undefined hidden
|
|
symbols typically won't have dynamic relocations, but
|
|
we most likely will need dynamic relocations for
|
|
__ehdr_start if we are building a PIE or shared
|
|
library. */
|
|
ehdr_start = &h->root;
|
|
ehdr_start_save_type = ehdr_start->type;
|
|
memcpy (ehdr_start_save_u,
|
|
(char *) &ehdr_start->u + sizeof ehdr_start->u.def.next,
|
|
sizeof ehdr_start_save_u);
|
|
ehdr_start->type = bfd_link_hash_defined;
|
|
ehdr_start->u.def.section = bfd_abs_section_ptr;
|
|
ehdr_start->u.def.value = 0;
|
|
}
|
|
}
|
|
|
|
/* If we are going to make any variable assignments, we need to
|
|
let the ELF backend know about them in case the variables are
|
|
referred to by dynamic objects. */
|
|
lang_for_each_statement (ldelf_find_statement_assignment);
|
|
}
|
|
|
|
/* Let the ELF backend work out the sizes of any sections required
|
|
by dynamic linking. */
|
|
rpath = command_line.rpath;
|
|
if (rpath == NULL)
|
|
rpath = (const char *) getenv ("LD_RUN_PATH");
|
|
|
|
for (abfd = link_info.input_bfds; abfd; abfd = abfd->link.next)
|
|
if (bfd_get_flavour (abfd) == bfd_target_elf_flavour)
|
|
{
|
|
const char *audit_libs = elf_dt_audit (abfd);
|
|
|
|
/* If the input bfd contains an audit entry, we need to add it as
|
|
a dep audit entry. */
|
|
if (audit_libs && *audit_libs != '\0')
|
|
{
|
|
char *cp = xstrdup (audit_libs);
|
|
do
|
|
{
|
|
int more = 0;
|
|
char *cp2 = strchr (cp, config.rpath_separator);
|
|
|
|
if (cp2)
|
|
{
|
|
*cp2 = '\0';
|
|
more = 1;
|
|
}
|
|
|
|
if (cp != NULL && *cp != '\0')
|
|
ldelf_append_to_separated_string (&depaudit, cp);
|
|
|
|
cp = more ? ++cp2 : NULL;
|
|
}
|
|
while (cp != NULL);
|
|
}
|
|
}
|
|
|
|
if (! (bfd_elf_size_dynamic_sections
|
|
(link_info.output_bfd, command_line.soname, rpath,
|
|
command_line.filter_shlib, audit, depaudit,
|
|
(const char * const *) command_line.auxiliary_filters,
|
|
&link_info, &sinterp)))
|
|
einfo (_("%F%P: failed to set dynamic section sizes: %E\n"));
|
|
|
|
if (sinterp != NULL)
|
|
{
|
|
/* Let the user override the dynamic linker we are using. */
|
|
if (command_line.interpreter != NULL)
|
|
default_interpreter_name = command_line.interpreter;
|
|
if (default_interpreter_name != NULL)
|
|
{
|
|
sinterp->contents = (bfd_byte *) default_interpreter_name;
|
|
sinterp->size = strlen ((char *) sinterp->contents) + 1;
|
|
}
|
|
}
|
|
|
|
/* Look for any sections named .gnu.warning. As a GNU extensions,
|
|
we treat such sections as containing warning messages. We print
|
|
out the warning message, and then zero out the section size so
|
|
that it does not get copied into the output file. */
|
|
|
|
{
|
|
LANG_FOR_EACH_INPUT_STATEMENT (is)
|
|
{
|
|
asection *s;
|
|
bfd_size_type sz;
|
|
char *msg;
|
|
|
|
if (is->flags.just_syms)
|
|
continue;
|
|
|
|
s = bfd_get_section_by_name (is->the_bfd, ".gnu.warning");
|
|
if (s == NULL)
|
|
continue;
|
|
|
|
sz = s->size;
|
|
msg = (char *) xmalloc ((size_t) (sz + 1));
|
|
if (! bfd_get_section_contents (is->the_bfd, s, msg,
|
|
(file_ptr) 0, sz))
|
|
einfo (_("%F%P: %pB: can't read contents of section .gnu.warning: %E\n"),
|
|
is->the_bfd);
|
|
msg[sz] = '\0';
|
|
(*link_info.callbacks->warning) (&link_info, msg,
|
|
(const char *) NULL, is->the_bfd,
|
|
(asection *) NULL, (bfd_vma) 0);
|
|
free (msg);
|
|
|
|
/* Clobber the section size, so that we don't waste space
|
|
copying the warning into the output file. If we've already
|
|
sized the output section, adjust its size. The adjustment
|
|
is on rawsize because targets that size sections early will
|
|
have called lang_reset_memory_regions after sizing. */
|
|
if (s->output_section != NULL
|
|
&& s->output_section->rawsize >= s->size)
|
|
s->output_section->rawsize -= s->size;
|
|
|
|
s->size = 0;
|
|
|
|
/* Also set SEC_EXCLUDE, so that local symbols defined in the
|
|
warning section don't get copied to the output. */
|
|
s->flags |= SEC_EXCLUDE | SEC_KEEP;
|
|
}
|
|
}
|
|
|
|
before_allocation_default ();
|
|
|
|
if (!bfd_elf_size_dynsym_hash_dynstr (link_info.output_bfd, &link_info))
|
|
einfo (_("%F%P: failed to set dynamic section sizes: %E\n"));
|
|
|
|
if (ehdr_start != NULL)
|
|
{
|
|
/* If we twiddled __ehdr_start to defined earlier, put it back
|
|
as it was. */
|
|
ehdr_start->type = ehdr_start_save_type;
|
|
memcpy ((char *) &ehdr_start->u + sizeof ehdr_start->u.def.next,
|
|
ehdr_start_save_u,
|
|
sizeof ehdr_start_save_u);
|
|
}
|
|
}
|
|
/* Try to open a dynamic archive. This is where we know that ELF
|
|
dynamic libraries have an extension of .so (or .sl on oddball systems
|
|
like hpux). */
|
|
|
|
bfd_boolean
|
|
ldelf_open_dynamic_archive (const char *arch, search_dirs_type *search,
|
|
lang_input_statement_type *entry)
|
|
{
|
|
const char *filename;
|
|
char *string;
|
|
size_t len;
|
|
bfd_boolean opened = FALSE;
|
|
|
|
if (! entry->flags.maybe_archive)
|
|
return FALSE;
|
|
|
|
filename = entry->filename;
|
|
len = strlen (search->name) + strlen (filename);
|
|
if (entry->flags.full_name_provided)
|
|
{
|
|
len += sizeof "/";
|
|
string = (char *) xmalloc (len);
|
|
sprintf (string, "%s/%s", search->name, filename);
|
|
}
|
|
else
|
|
{
|
|
size_t xlen = 0;
|
|
|
|
len += strlen (arch) + sizeof "/lib.so";
|
|
#ifdef EXTRA_SHLIB_EXTENSION
|
|
xlen = (strlen (EXTRA_SHLIB_EXTENSION) > 3
|
|
? strlen (EXTRA_SHLIB_EXTENSION) - 3
|
|
: 0);
|
|
#endif
|
|
string = (char *) xmalloc (len + xlen);
|
|
sprintf (string, "%s/lib%s%s.so", search->name, filename, arch);
|
|
#ifdef EXTRA_SHLIB_EXTENSION
|
|
/* Try the .so extension first. If that fails build a new filename
|
|
using EXTRA_SHLIB_EXTENSION. */
|
|
opened = ldfile_try_open_bfd (string, entry);
|
|
if (!opened)
|
|
strcpy (string + len - 4, EXTRA_SHLIB_EXTENSION);
|
|
#endif
|
|
}
|
|
|
|
if (!opened && !ldfile_try_open_bfd (string, entry))
|
|
{
|
|
free (string);
|
|
return FALSE;
|
|
}
|
|
|
|
entry->filename = string;
|
|
|
|
/* We have found a dynamic object to include in the link. The ELF
|
|
backend linker will create a DT_NEEDED entry in the .dynamic
|
|
section naming this file. If this file includes a DT_SONAME
|
|
entry, it will be used. Otherwise, the ELF linker will just use
|
|
the name of the file. For an archive found by searching, like
|
|
this one, the DT_NEEDED entry should consist of just the name of
|
|
the file, without the path information used to find it. Note
|
|
that we only need to do this if we have a dynamic object; an
|
|
archive will never be referenced by a DT_NEEDED entry.
|
|
|
|
FIXME: This approach--using bfd_elf_set_dt_needed_name--is not
|
|
very pretty. I haven't been able to think of anything that is
|
|
pretty, though. */
|
|
if (bfd_check_format (entry->the_bfd, bfd_object)
|
|
&& (entry->the_bfd->flags & DYNAMIC) != 0)
|
|
{
|
|
ASSERT (entry->flags.maybe_archive && entry->flags.search_dirs);
|
|
|
|
/* Rather than duplicating the logic above. Just use the
|
|
filename we recorded earlier. */
|
|
|
|
if (!entry->flags.full_name_provided)
|
|
filename = lbasename (entry->filename);
|
|
bfd_elf_set_dt_needed_name (entry->the_bfd, filename);
|
|
}
|
|
|
|
return TRUE;
|
|
}
|
|
|
|
/* A variant of lang_output_section_find used by place_orphan. */
|
|
|
|
static lang_output_section_statement_type *
|
|
output_rel_find (int isdyn, int rela)
|
|
{
|
|
lang_output_section_statement_type *lookup;
|
|
lang_output_section_statement_type *last = NULL;
|
|
lang_output_section_statement_type *last_alloc = NULL;
|
|
lang_output_section_statement_type *last_ro_alloc = NULL;
|
|
lang_output_section_statement_type *last_rel = NULL;
|
|
lang_output_section_statement_type *last_rel_alloc = NULL;
|
|
|
|
for (lookup = &lang_os_list.head->output_section_statement;
|
|
lookup != NULL;
|
|
lookup = lookup->next)
|
|
{
|
|
if (lookup->constraint >= 0
|
|
&& CONST_STRNEQ (lookup->name, ".rel"))
|
|
{
|
|
int lookrela = lookup->name[4] == 'a';
|
|
|
|
/* .rel.dyn must come before all other reloc sections, to suit
|
|
GNU ld.so. */
|
|
if (isdyn)
|
|
break;
|
|
|
|
/* Don't place after .rel.plt as doing so results in wrong
|
|
dynamic tags. */
|
|
if (strcmp (".plt", lookup->name + 4 + lookrela) == 0)
|
|
break;
|
|
|
|
if (rela == lookrela || last_rel == NULL)
|
|
last_rel = lookup;
|
|
if ((rela == lookrela || last_rel_alloc == NULL)
|
|
&& lookup->bfd_section != NULL
|
|
&& (lookup->bfd_section->flags & SEC_ALLOC) != 0)
|
|
last_rel_alloc = lookup;
|
|
}
|
|
|
|
last = lookup;
|
|
if (lookup->bfd_section != NULL
|
|
&& (lookup->bfd_section->flags & SEC_ALLOC) != 0)
|
|
{
|
|
last_alloc = lookup;
|
|
if ((lookup->bfd_section->flags & SEC_READONLY) != 0)
|
|
last_ro_alloc = lookup;
|
|
}
|
|
}
|
|
|
|
if (last_rel_alloc)
|
|
return last_rel_alloc;
|
|
|
|
if (last_rel)
|
|
return last_rel;
|
|
|
|
if (last_ro_alloc)
|
|
return last_ro_alloc;
|
|
|
|
if (last_alloc)
|
|
return last_alloc;
|
|
|
|
return last;
|
|
}
|
|
|
|
/* Return whether IN is suitable to be part of OUT. */
|
|
|
|
static bfd_boolean
|
|
elf_orphan_compatible (asection *in, asection *out)
|
|
{
|
|
/* Non-zero sh_info implies a section with SHF_INFO_LINK with
|
|
unknown semantics for the generic linker, or a SHT_REL/SHT_RELA
|
|
section where sh_info specifies a symbol table. (We won't see
|
|
SHT_GROUP, SHT_SYMTAB or SHT_DYNSYM sections here.) We clearly
|
|
can't merge SHT_REL/SHT_RELA using differing symbol tables, and
|
|
shouldn't merge sections with differing unknown semantics. */
|
|
if (elf_section_data (out)->this_hdr.sh_info
|
|
!= elf_section_data (in)->this_hdr.sh_info)
|
|
return FALSE;
|
|
/* We can't merge with member of output section group nor merge two
|
|
sections with differing SHF_EXCLUDE when doing a relocatable link.
|
|
*/
|
|
if (bfd_link_relocatable (&link_info)
|
|
&& (elf_next_in_group (out) != NULL
|
|
|| ((elf_section_flags (out) ^ elf_section_flags (in))
|
|
& SHF_EXCLUDE) != 0))
|
|
return FALSE;
|
|
return _bfd_elf_match_sections_by_type (link_info.output_bfd, out,
|
|
in->owner, in);
|
|
}
|
|
|
|
/* Place an orphan section. We use this to put random SHF_ALLOC
|
|
sections in the right segment. */
|
|
|
|
lang_output_section_statement_type *
|
|
ldelf_place_orphan (asection *s, const char *secname, int constraint)
|
|
{
|
|
static struct orphan_save hold[] =
|
|
{
|
|
{ ".text",
|
|
SEC_HAS_CONTENTS | SEC_ALLOC | SEC_LOAD | SEC_READONLY | SEC_CODE,
|
|
0, 0, 0, 0 },
|
|
{ ".rodata",
|
|
SEC_HAS_CONTENTS | SEC_ALLOC | SEC_LOAD | SEC_READONLY | SEC_DATA,
|
|
0, 0, 0, 0 },
|
|
{ ".tdata",
|
|
SEC_HAS_CONTENTS | SEC_ALLOC | SEC_LOAD | SEC_DATA | SEC_THREAD_LOCAL,
|
|
0, 0, 0, 0 },
|
|
{ ".data",
|
|
SEC_HAS_CONTENTS | SEC_ALLOC | SEC_LOAD | SEC_DATA,
|
|
0, 0, 0, 0 },
|
|
{ ".bss",
|
|
SEC_ALLOC,
|
|
0, 0, 0, 0 },
|
|
{ 0,
|
|
SEC_HAS_CONTENTS | SEC_ALLOC | SEC_LOAD | SEC_READONLY | SEC_DATA,
|
|
0, 0, 0, 0 },
|
|
{ ".interp",
|
|
SEC_HAS_CONTENTS | SEC_ALLOC | SEC_LOAD | SEC_READONLY | SEC_DATA,
|
|
0, 0, 0, 0 },
|
|
{ ".sdata",
|
|
SEC_HAS_CONTENTS | SEC_ALLOC | SEC_LOAD | SEC_DATA | SEC_SMALL_DATA,
|
|
0, 0, 0, 0 },
|
|
{ ".comment",
|
|
SEC_HAS_CONTENTS,
|
|
0, 0, 0, 0 },
|
|
};
|
|
enum orphan_save_index
|
|
{
|
|
orphan_text = 0,
|
|
orphan_rodata,
|
|
orphan_tdata,
|
|
orphan_data,
|
|
orphan_bss,
|
|
orphan_rel,
|
|
orphan_interp,
|
|
orphan_sdata,
|
|
orphan_nonalloc
|
|
};
|
|
static int orphan_init_done = 0;
|
|
struct orphan_save *place;
|
|
lang_output_section_statement_type *after;
|
|
lang_output_section_statement_type *os;
|
|
lang_output_section_statement_type *match_by_name = NULL;
|
|
int isdyn = 0;
|
|
int elfinput = s->owner->xvec->flavour == bfd_target_elf_flavour;
|
|
int elfoutput = link_info.output_bfd->xvec->flavour == bfd_target_elf_flavour;
|
|
unsigned int sh_type = elfinput ? elf_section_type (s) : SHT_NULL;
|
|
flagword flags;
|
|
asection *nexts;
|
|
|
|
if (!bfd_link_relocatable (&link_info)
|
|
&& link_info.combreloc
|
|
&& (s->flags & SEC_ALLOC))
|
|
{
|
|
if (elfinput)
|
|
switch (sh_type)
|
|
{
|
|
case SHT_RELA:
|
|
secname = ".rela.dyn";
|
|
isdyn = 1;
|
|
break;
|
|
case SHT_REL:
|
|
secname = ".rel.dyn";
|
|
isdyn = 1;
|
|
break;
|
|
default:
|
|
break;
|
|
}
|
|
else if (CONST_STRNEQ (secname, ".rel"))
|
|
{
|
|
secname = secname[4] == 'a' ? ".rela.dyn" : ".rel.dyn";
|
|
isdyn = 1;
|
|
}
|
|
}
|
|
|
|
if (!bfd_link_relocatable (&link_info)
|
|
&& elfinput
|
|
&& elfoutput
|
|
&& (s->flags & SEC_ALLOC) != 0
|
|
&& (elf_tdata (s->owner)->has_gnu_osabi & elf_gnu_osabi_mbind) != 0
|
|
&& (elf_section_flags (s) & SHF_GNU_MBIND) != 0)
|
|
{
|
|
/* Find the output mbind section with the same type, attributes
|
|
and sh_info field. */
|
|
for (os = &lang_os_list.head->output_section_statement;
|
|
os != NULL;
|
|
os = os->next)
|
|
if (os->bfd_section != NULL
|
|
&& !bfd_is_abs_section (os->bfd_section)
|
|
&& (elf_section_flags (os->bfd_section) & SHF_GNU_MBIND) != 0
|
|
&& ((s->flags & (SEC_ALLOC
|
|
| SEC_LOAD
|
|
| SEC_HAS_CONTENTS
|
|
| SEC_READONLY
|
|
| SEC_CODE))
|
|
== (os->bfd_section->flags & (SEC_ALLOC
|
|
| SEC_LOAD
|
|
| SEC_HAS_CONTENTS
|
|
| SEC_READONLY
|
|
| SEC_CODE)))
|
|
&& (elf_section_data (os->bfd_section)->this_hdr.sh_info
|
|
== elf_section_data (s)->this_hdr.sh_info))
|
|
{
|
|
lang_add_section (&os->children, s, NULL, os);
|
|
return os;
|
|
}
|
|
|
|
/* Create the output mbind section with the ".mbind." prefix
|
|
in section name. */
|
|
if ((s->flags & (SEC_LOAD | SEC_HAS_CONTENTS)) == 0)
|
|
secname = ".mbind.bss";
|
|
else if ((s->flags & SEC_READONLY) == 0)
|
|
secname = ".mbind.data";
|
|
else if ((s->flags & SEC_CODE) == 0)
|
|
secname = ".mbind.rodata";
|
|
else
|
|
secname = ".mbind.text";
|
|
elf_tdata (link_info.output_bfd)->has_gnu_osabi |= elf_gnu_osabi_mbind;
|
|
}
|
|
|
|
/* Look through the script to see where to place this section. The
|
|
script includes entries added by previous lang_insert_orphan
|
|
calls, so this loop puts multiple compatible orphans of the same
|
|
name into a single output section. */
|
|
if (constraint == 0)
|
|
for (os = lang_output_section_find (secname);
|
|
os != NULL;
|
|
os = next_matching_output_section_statement (os, 0))
|
|
{
|
|
/* If we don't match an existing output section, tell
|
|
lang_insert_orphan to create a new output section. */
|
|
constraint = SPECIAL;
|
|
|
|
/* Check to see if we already have an output section statement
|
|
with this name, and its bfd section has compatible flags.
|
|
If the section already exists but does not have any flags
|
|
set, then it has been created by the linker, possibly as a
|
|
result of a --section-start command line switch. */
|
|
if (os->bfd_section != NULL
|
|
&& (os->bfd_section->flags == 0
|
|
|| (((s->flags ^ os->bfd_section->flags)
|
|
& (SEC_LOAD | SEC_ALLOC)) == 0
|
|
&& (!elfinput
|
|
|| !elfoutput
|
|
|| elf_orphan_compatible (s, os->bfd_section)))))
|
|
{
|
|
lang_add_section (&os->children, s, NULL, os);
|
|
return os;
|
|
}
|
|
|
|
/* Save unused output sections in case we can match them
|
|
against orphans later. */
|
|
if (os->bfd_section == NULL)
|
|
match_by_name = os;
|
|
}
|
|
|
|
/* If we didn't match an active output section, see if we matched an
|
|
unused one and use that. */
|
|
if (match_by_name)
|
|
{
|
|
lang_add_section (&match_by_name->children, s, NULL, match_by_name);
|
|
return match_by_name;
|
|
}
|
|
|
|
if (!orphan_init_done)
|
|
{
|
|
struct orphan_save *ho;
|
|
|
|
for (ho = hold; ho < hold + sizeof (hold) / sizeof (hold[0]); ++ho)
|
|
if (ho->name != NULL)
|
|
{
|
|
ho->os = lang_output_section_find (ho->name);
|
|
if (ho->os != NULL && ho->os->flags == 0)
|
|
ho->os->flags = ho->flags;
|
|
}
|
|
orphan_init_done = 1;
|
|
}
|
|
|
|
/* If this is a final link, then always put .gnu.warning.SYMBOL
|
|
sections into the .text section to get them out of the way. */
|
|
if (bfd_link_executable (&link_info)
|
|
&& CONST_STRNEQ (s->name, ".gnu.warning.")
|
|
&& hold[orphan_text].os != NULL)
|
|
{
|
|
os = hold[orphan_text].os;
|
|
lang_add_section (&os->children, s, NULL, os);
|
|
return os;
|
|
}
|
|
|
|
flags = s->flags;
|
|
if (!bfd_link_relocatable (&link_info))
|
|
{
|
|
nexts = s;
|
|
while ((nexts = bfd_get_next_section_by_name (nexts->owner, nexts))
|
|
!= NULL)
|
|
if (nexts->output_section == NULL
|
|
&& (nexts->flags & SEC_EXCLUDE) == 0
|
|
&& ((nexts->flags ^ flags) & (SEC_LOAD | SEC_ALLOC)) == 0
|
|
&& (nexts->owner->flags & DYNAMIC) == 0
|
|
&& !bfd_input_just_syms (nexts->owner)
|
|
&& _bfd_elf_match_sections_by_type (nexts->owner, nexts,
|
|
s->owner, s))
|
|
flags = (((flags ^ SEC_READONLY)
|
|
| (nexts->flags ^ SEC_READONLY))
|
|
^ SEC_READONLY);
|
|
}
|
|
|
|
/* Decide which segment the section should go in based on the
|
|
section name and section flags. We put loadable .note sections
|
|
right after the .interp section, so that the PT_NOTE segment is
|
|
stored right after the program headers where the OS can read it
|
|
in the first page. */
|
|
|
|
place = NULL;
|
|
if ((flags & (SEC_ALLOC | SEC_DEBUGGING)) == 0)
|
|
place = &hold[orphan_nonalloc];
|
|
else if ((flags & SEC_ALLOC) == 0)
|
|
;
|
|
else if ((flags & SEC_LOAD) != 0
|
|
&& (elfinput
|
|
? sh_type == SHT_NOTE
|
|
: CONST_STRNEQ (secname, ".note")))
|
|
place = &hold[orphan_interp];
|
|
else if ((flags & (SEC_LOAD | SEC_HAS_CONTENTS | SEC_THREAD_LOCAL)) == 0)
|
|
place = &hold[orphan_bss];
|
|
else if ((flags & SEC_SMALL_DATA) != 0)
|
|
place = &hold[orphan_sdata];
|
|
else if ((flags & SEC_THREAD_LOCAL) != 0)
|
|
place = &hold[orphan_tdata];
|
|
else if ((flags & SEC_READONLY) == 0)
|
|
place = &hold[orphan_data];
|
|
else if ((flags & SEC_LOAD) != 0
|
|
&& (elfinput
|
|
? sh_type == SHT_RELA || sh_type == SHT_REL
|
|
: CONST_STRNEQ (secname, ".rel")))
|
|
place = &hold[orphan_rel];
|
|
else if ((flags & SEC_CODE) == 0)
|
|
place = &hold[orphan_rodata];
|
|
else
|
|
place = &hold[orphan_text];
|
|
|
|
after = NULL;
|
|
if (place != NULL)
|
|
{
|
|
if (place->os == NULL)
|
|
{
|
|
if (place->name != NULL)
|
|
place->os = lang_output_section_find (place->name);
|
|
else
|
|
{
|
|
int rela = elfinput ? sh_type == SHT_RELA : secname[4] == 'a';
|
|
place->os = output_rel_find (isdyn, rela);
|
|
}
|
|
}
|
|
after = place->os;
|
|
if (after == NULL)
|
|
after
|
|
= lang_output_section_find_by_flags (s, flags, &place->os,
|
|
_bfd_elf_match_sections_by_type);
|
|
if (after == NULL)
|
|
/* *ABS* is always the first output section statement. */
|
|
after = &lang_os_list.head->output_section_statement;
|
|
}
|
|
|
|
return lang_insert_orphan (s, secname, constraint, after, place, NULL, NULL);
|
|
}
|