/* ELF executable support for BFD.
- Copyright 1993 Free Software Foundation, Inc.
+ Copyright 1993, 94, 95, 96, 97, 98, 1999 Free Software Foundation, Inc.
This file is part of BFD, the Binary File Descriptor library.
#include "bfdlink.h"
#include "libbfd.h"
#define ARCH_SIZE 0
-#include "libelf.h"
-
-static file_ptr map_program_segments PARAMS ((bfd *, file_ptr,
- Elf_Internal_Shdr *,
- Elf_Internal_Shdr **,
- bfd_size_type));
-static boolean assign_file_positions_except_relocs PARAMS ((bfd *, boolean));
+#include "elf-bfd.h"
+
+static INLINE struct elf_segment_map *make_mapping
+ PARAMS ((bfd *, asection **, unsigned int, unsigned int, boolean));
+static boolean map_sections_to_segments PARAMS ((bfd *));
+static int elf_sort_sections PARAMS ((const PTR, const PTR));
+static boolean assign_file_positions_for_segments PARAMS ((bfd *));
+static boolean assign_file_positions_except_relocs PARAMS ((bfd *));
static boolean prep_headers PARAMS ((bfd *));
-static boolean swap_out_syms PARAMS ((bfd *, struct bfd_strtab_hash **));
+static boolean swap_out_syms PARAMS ((bfd *, struct bfd_strtab_hash **, int));
+static boolean copy_private_bfd_data PARAMS ((bfd *, bfd *));
+static char *elf_read PARAMS ((bfd *, long, unsigned int));
+static void elf_fake_sections PARAMS ((bfd *, asection *, PTR));
+static boolean assign_section_numbers PARAMS ((bfd *));
+static INLINE int sym_is_global PARAMS ((bfd *, asymbol *));
+static boolean elf_map_symbols PARAMS ((bfd *));
+static bfd_size_type get_program_header_size PARAMS ((bfd *));
+
+/* Swap version information in and out. The version information is
+ currently size independent. If that ever changes, this code will
+ need to move into elfcode.h. */
+
+/* Swap in a Verdef structure. */
+
+void
+_bfd_elf_swap_verdef_in (abfd, src, dst)
+ bfd *abfd;
+ const Elf_External_Verdef *src;
+ Elf_Internal_Verdef *dst;
+{
+ dst->vd_version = bfd_h_get_16 (abfd, src->vd_version);
+ dst->vd_flags = bfd_h_get_16 (abfd, src->vd_flags);
+ dst->vd_ndx = bfd_h_get_16 (abfd, src->vd_ndx);
+ dst->vd_cnt = bfd_h_get_16 (abfd, src->vd_cnt);
+ dst->vd_hash = bfd_h_get_32 (abfd, src->vd_hash);
+ dst->vd_aux = bfd_h_get_32 (abfd, src->vd_aux);
+ dst->vd_next = bfd_h_get_32 (abfd, src->vd_next);
+}
+
+/* Swap out a Verdef structure. */
+
+void
+_bfd_elf_swap_verdef_out (abfd, src, dst)
+ bfd *abfd;
+ const Elf_Internal_Verdef *src;
+ Elf_External_Verdef *dst;
+{
+ bfd_h_put_16 (abfd, src->vd_version, dst->vd_version);
+ bfd_h_put_16 (abfd, src->vd_flags, dst->vd_flags);
+ bfd_h_put_16 (abfd, src->vd_ndx, dst->vd_ndx);
+ bfd_h_put_16 (abfd, src->vd_cnt, dst->vd_cnt);
+ bfd_h_put_32 (abfd, src->vd_hash, dst->vd_hash);
+ bfd_h_put_32 (abfd, src->vd_aux, dst->vd_aux);
+ bfd_h_put_32 (abfd, src->vd_next, dst->vd_next);
+}
+
+/* Swap in a Verdaux structure. */
+
+void
+_bfd_elf_swap_verdaux_in (abfd, src, dst)
+ bfd *abfd;
+ const Elf_External_Verdaux *src;
+ Elf_Internal_Verdaux *dst;
+{
+ dst->vda_name = bfd_h_get_32 (abfd, src->vda_name);
+ dst->vda_next = bfd_h_get_32 (abfd, src->vda_next);
+}
+
+/* Swap out a Verdaux structure. */
+
+void
+_bfd_elf_swap_verdaux_out (abfd, src, dst)
+ bfd *abfd;
+ const Elf_Internal_Verdaux *src;
+ Elf_External_Verdaux *dst;
+{
+ bfd_h_put_32 (abfd, src->vda_name, dst->vda_name);
+ bfd_h_put_32 (abfd, src->vda_next, dst->vda_next);
+}
+
+/* Swap in a Verneed structure. */
+
+void
+_bfd_elf_swap_verneed_in (abfd, src, dst)
+ bfd *abfd;
+ const Elf_External_Verneed *src;
+ Elf_Internal_Verneed *dst;
+{
+ dst->vn_version = bfd_h_get_16 (abfd, src->vn_version);
+ dst->vn_cnt = bfd_h_get_16 (abfd, src->vn_cnt);
+ dst->vn_file = bfd_h_get_32 (abfd, src->vn_file);
+ dst->vn_aux = bfd_h_get_32 (abfd, src->vn_aux);
+ dst->vn_next = bfd_h_get_32 (abfd, src->vn_next);
+}
+
+/* Swap out a Verneed structure. */
+
+void
+_bfd_elf_swap_verneed_out (abfd, src, dst)
+ bfd *abfd;
+ const Elf_Internal_Verneed *src;
+ Elf_External_Verneed *dst;
+{
+ bfd_h_put_16 (abfd, src->vn_version, dst->vn_version);
+ bfd_h_put_16 (abfd, src->vn_cnt, dst->vn_cnt);
+ bfd_h_put_32 (abfd, src->vn_file, dst->vn_file);
+ bfd_h_put_32 (abfd, src->vn_aux, dst->vn_aux);
+ bfd_h_put_32 (abfd, src->vn_next, dst->vn_next);
+}
+
+/* Swap in a Vernaux structure. */
+
+void
+_bfd_elf_swap_vernaux_in (abfd, src, dst)
+ bfd *abfd;
+ const Elf_External_Vernaux *src;
+ Elf_Internal_Vernaux *dst;
+{
+ dst->vna_hash = bfd_h_get_32 (abfd, src->vna_hash);
+ dst->vna_flags = bfd_h_get_16 (abfd, src->vna_flags);
+ dst->vna_other = bfd_h_get_16 (abfd, src->vna_other);
+ dst->vna_name = bfd_h_get_32 (abfd, src->vna_name);
+ dst->vna_next = bfd_h_get_32 (abfd, src->vna_next);
+}
+
+/* Swap out a Vernaux structure. */
+
+void
+_bfd_elf_swap_vernaux_out (abfd, src, dst)
+ bfd *abfd;
+ const Elf_Internal_Vernaux *src;
+ Elf_External_Vernaux *dst;
+{
+ bfd_h_put_32 (abfd, src->vna_hash, dst->vna_hash);
+ bfd_h_put_16 (abfd, src->vna_flags, dst->vna_flags);
+ bfd_h_put_16 (abfd, src->vna_other, dst->vna_other);
+ bfd_h_put_32 (abfd, src->vna_name, dst->vna_name);
+ bfd_h_put_32 (abfd, src->vna_next, dst->vna_next);
+}
+
+/* Swap in a Versym structure. */
+
+void
+_bfd_elf_swap_versym_in (abfd, src, dst)
+ bfd *abfd;
+ const Elf_External_Versym *src;
+ Elf_Internal_Versym *dst;
+{
+ dst->vs_vers = bfd_h_get_16 (abfd, src->vs_vers);
+}
+
+/* Swap out a Versym structure. */
+
+void
+_bfd_elf_swap_versym_out (abfd, src, dst)
+ bfd *abfd;
+ const Elf_Internal_Versym *src;
+ Elf_External_Versym *dst;
+{
+ bfd_h_put_16 (abfd, src->vs_vers, dst->vs_vers);
+}
/* Standard ELF hash function. Do not change this function; you will
- cause invalid hash tables to be generated. (Well, you would if this
- were being used yet.) */
+ cause invalid hash tables to be generated. */
+
unsigned long
-bfd_elf_hash (name)
- CONST unsigned char *name;
+bfd_elf_hash (namearg)
+ const char *namearg;
{
+ const unsigned char *name = (const unsigned char *) namearg;
unsigned long h = 0;
unsigned long g;
int ch;
if ((g = (h & 0xf0000000)) != 0)
{
h ^= g >> 24;
- h &= ~g;
+ /* The ELF ABI says `h &= ~g', but this is equivalent in
+ this case and on some machines one insn instead of two. */
+ h ^= g;
}
}
return h;
char *buf;
if ((buf = bfd_alloc (abfd, size)) == NULL)
- {
- bfd_set_error (bfd_error_no_memory);
- return NULL;
- }
+ return NULL;
if (bfd_seek (abfd, offset, SEEK_SET) == -1)
return NULL;
if (bfd_read ((PTR) buf, size, 1, abfd) != size)
}
boolean
-elf_mkobject (abfd)
+bfd_elf_mkobject (abfd)
bfd * abfd;
{
/* this just does initialization */
elf_tdata (abfd) = (struct elf_obj_tdata *)
bfd_zalloc (abfd, sizeof (struct elf_obj_tdata));
if (elf_tdata (abfd) == 0)
- {
- bfd_set_error (bfd_error_no_memory);
- return false;
- }
+ return false;
/* since everything is done at close time, do we need any
initialization? */
return true;
}
+boolean
+bfd_elf_mkcorefile (abfd)
+ bfd * abfd;
+{
+ /* I think this can be done just like an object file. */
+ return bfd_elf_mkobject (abfd);
+}
+
char *
bfd_elf_get_str_section (abfd, shindex)
bfd * abfd;
&& bfd_elf_get_str_section (abfd, shindex) == NULL)
return NULL;
+ if (strindex >= hdr->sh_size)
+ {
+ (*_bfd_error_handler)
+ (_("%s: invalid string offset %u >= %lu for section `%s'"),
+ bfd_get_filename (abfd), strindex, (unsigned long) hdr->sh_size,
+ ((shindex == elf_elfheader(abfd)->e_shstrndx
+ && strindex == hdr->sh_name)
+ ? ".shstrtab"
+ : elf_string_from_elf_strtab (abfd, hdr->sh_name)));
+ return "";
+ }
+
return ((char *) hdr->contents) + strindex;
}
|| strncmp (name, ".stab", sizeof ".stab" - 1) == 0)
flags |= SEC_DEBUGGING;
+ /* As a GNU extension, if the name begins with .gnu.linkonce, we
+ only link a single copy of the section. This is used to support
+ g++. g++ will emit each template expansion in its own section.
+ The symbols will be defined as weak, so that multiple definitions
+ are permitted. The GNU linker extension is to actually discard
+ all but one of the sections. */
+ if (strncmp (name, ".gnu.linkonce", sizeof ".gnu.linkonce" - 1) == 0)
+ flags |= SEC_LINK_ONCE | SEC_LINK_DUPLICATES_DISCARD;
+
if (! bfd_set_section_flags (abfd, newsect, flags))
return false;
+ if ((flags & SEC_ALLOC) != 0)
+ {
+ Elf_Internal_Phdr *phdr;
+ unsigned int i;
+
+ /* Look through the phdrs to see if we need to adjust the lma.
+ If all the p_paddr fields are zero, we ignore them, since
+ some ELF linkers produce such output. */
+ phdr = elf_tdata (abfd)->phdr;
+ for (i = 0; i < elf_elfheader (abfd)->e_phnum; i++, phdr++)
+ {
+ if (phdr->p_paddr != 0)
+ break;
+ }
+ if (i < elf_elfheader (abfd)->e_phnum)
+ {
+ phdr = elf_tdata (abfd)->phdr;
+ for (i = 0; i < elf_elfheader (abfd)->e_phnum; i++, phdr++)
+ {
+ if (phdr->p_type == PT_LOAD
+ && phdr->p_vaddr != phdr->p_paddr
+ && phdr->p_vaddr <= hdr->sh_addr
+ && (phdr->p_vaddr + phdr->p_memsz
+ >= hdr->sh_addr + hdr->sh_size)
+ && ((flags & SEC_LOAD) == 0
+ || (phdr->p_offset <= (bfd_vma) hdr->sh_offset
+ && (phdr->p_offset + phdr->p_filesz
+ >= hdr->sh_offset + hdr->sh_size))))
+ {
+ newsect->lma += phdr->p_paddr - phdr->p_vaddr;
+ break;
+ }
+ }
+ }
+ }
+
hdr->bfd_section = newsect;
elf_section_data (newsect)->this_hdr = *hdr;
i_shdrp = elf_elfsections (abfd);
if (i_shdrp != NULL)
{
- shstrtab = bfd_elf_get_str_section (abfd, elf_elfheader (abfd)->e_shstrndx);
+ shstrtab = bfd_elf_get_str_section
+ (abfd, elf_elfheader (abfd)->e_shstrndx);
if (shstrtab != NULL)
{
max = elf_elfheader (abfd)->e_shnum;
input_section,
output_bfd,
error_message)
- bfd *abfd;
+ bfd *abfd ATTRIBUTE_UNUSED;
arelent *reloc_entry;
asymbol *symbol;
- PTR data;
+ PTR data ATTRIBUTE_UNUSED;
asection *input_section;
bfd *output_bfd;
- char **error_message;
+ char **error_message ATTRIBUTE_UNUSED;
{
if (output_bfd != (bfd *) NULL
&& (symbol->flags & BSF_SECTION_SYM) == 0
return bfd_reloc_continue;
}
\f
+/* Print out the program headers. */
+
+boolean
+_bfd_elf_print_private_bfd_data (abfd, farg)
+ bfd *abfd;
+ PTR farg;
+{
+ FILE *f = (FILE *) farg;
+ Elf_Internal_Phdr *p;
+ asection *s;
+ bfd_byte *dynbuf = NULL;
+
+ p = elf_tdata (abfd)->phdr;
+ if (p != NULL)
+ {
+ unsigned int i, c;
+
+ fprintf (f, _("\nProgram Header:\n"));
+ c = elf_elfheader (abfd)->e_phnum;
+ for (i = 0; i < c; i++, p++)
+ {
+ const char *s;
+ char buf[20];
+
+ switch (p->p_type)
+ {
+ case PT_NULL: s = "NULL"; break;
+ case PT_LOAD: s = "LOAD"; break;
+ case PT_DYNAMIC: s = "DYNAMIC"; break;
+ case PT_INTERP: s = "INTERP"; break;
+ case PT_NOTE: s = "NOTE"; break;
+ case PT_SHLIB: s = "SHLIB"; break;
+ case PT_PHDR: s = "PHDR"; break;
+ default: sprintf (buf, "0x%lx", p->p_type); s = buf; break;
+ }
+ fprintf (f, "%8s off 0x", s);
+ fprintf_vma (f, p->p_offset);
+ fprintf (f, " vaddr 0x");
+ fprintf_vma (f, p->p_vaddr);
+ fprintf (f, " paddr 0x");
+ fprintf_vma (f, p->p_paddr);
+ fprintf (f, " align 2**%u\n", bfd_log2 (p->p_align));
+ fprintf (f, " filesz 0x");
+ fprintf_vma (f, p->p_filesz);
+ fprintf (f, " memsz 0x");
+ fprintf_vma (f, p->p_memsz);
+ fprintf (f, " flags %c%c%c",
+ (p->p_flags & PF_R) != 0 ? 'r' : '-',
+ (p->p_flags & PF_W) != 0 ? 'w' : '-',
+ (p->p_flags & PF_X) != 0 ? 'x' : '-');
+ if ((p->p_flags &~ (PF_R | PF_W | PF_X)) != 0)
+ fprintf (f, " %lx", p->p_flags &~ (PF_R | PF_W | PF_X));
+ fprintf (f, "\n");
+ }
+ }
+
+ s = bfd_get_section_by_name (abfd, ".dynamic");
+ if (s != NULL)
+ {
+ int elfsec;
+ unsigned long link;
+ bfd_byte *extdyn, *extdynend;
+ size_t extdynsize;
+ void (*swap_dyn_in) PARAMS ((bfd *, const PTR, Elf_Internal_Dyn *));
+
+ fprintf (f, _("\nDynamic Section:\n"));
+
+ dynbuf = (bfd_byte *) bfd_malloc (s->_raw_size);
+ if (dynbuf == NULL)
+ goto error_return;
+ if (! bfd_get_section_contents (abfd, s, (PTR) dynbuf, (file_ptr) 0,
+ s->_raw_size))
+ goto error_return;
+
+ elfsec = _bfd_elf_section_from_bfd_section (abfd, s);
+ if (elfsec == -1)
+ goto error_return;
+ link = elf_elfsections (abfd)[elfsec]->sh_link;
+
+ extdynsize = get_elf_backend_data (abfd)->s->sizeof_dyn;
+ swap_dyn_in = get_elf_backend_data (abfd)->s->swap_dyn_in;
+
+ extdyn = dynbuf;
+ extdynend = extdyn + s->_raw_size;
+ for (; extdyn < extdynend; extdyn += extdynsize)
+ {
+ Elf_Internal_Dyn dyn;
+ const char *name;
+ char ab[20];
+ boolean stringp;
+
+ (*swap_dyn_in) (abfd, (PTR) extdyn, &dyn);
+
+ if (dyn.d_tag == DT_NULL)
+ break;
+
+ stringp = false;
+ switch (dyn.d_tag)
+ {
+ default:
+ sprintf (ab, "0x%lx", (unsigned long) dyn.d_tag);
+ name = ab;
+ break;
+
+ case DT_NEEDED: name = "NEEDED"; stringp = true; break;
+ case DT_PLTRELSZ: name = "PLTRELSZ"; break;
+ case DT_PLTGOT: name = "PLTGOT"; break;
+ case DT_HASH: name = "HASH"; break;
+ case DT_STRTAB: name = "STRTAB"; break;
+ case DT_SYMTAB: name = "SYMTAB"; break;
+ case DT_RELA: name = "RELA"; break;
+ case DT_RELASZ: name = "RELASZ"; break;
+ case DT_RELAENT: name = "RELAENT"; break;
+ case DT_STRSZ: name = "STRSZ"; break;
+ case DT_SYMENT: name = "SYMENT"; break;
+ case DT_INIT: name = "INIT"; break;
+ case DT_FINI: name = "FINI"; break;
+ case DT_SONAME: name = "SONAME"; stringp = true; break;
+ case DT_RPATH: name = "RPATH"; stringp = true; break;
+ case DT_SYMBOLIC: name = "SYMBOLIC"; break;
+ case DT_REL: name = "REL"; break;
+ case DT_RELSZ: name = "RELSZ"; break;
+ case DT_RELENT: name = "RELENT"; break;
+ case DT_PLTREL: name = "PLTREL"; break;
+ case DT_DEBUG: name = "DEBUG"; break;
+ case DT_TEXTREL: name = "TEXTREL"; break;
+ case DT_JMPREL: name = "JMPREL"; break;
+ case DT_AUXILIARY: name = "AUXILIARY"; stringp = true; break;
+ case DT_FILTER: name = "FILTER"; stringp = true; break;
+ case DT_VERSYM: name = "VERSYM"; break;
+ case DT_VERDEF: name = "VERDEF"; break;
+ case DT_VERDEFNUM: name = "VERDEFNUM"; break;
+ case DT_VERNEED: name = "VERNEED"; break;
+ case DT_VERNEEDNUM: name = "VERNEEDNUM"; break;
+ }
+
+ fprintf (f, " %-11s ", name);
+ if (! stringp)
+ fprintf (f, "0x%lx", (unsigned long) dyn.d_un.d_val);
+ else
+ {
+ const char *string;
+
+ string = bfd_elf_string_from_elf_section (abfd, link,
+ dyn.d_un.d_val);
+ if (string == NULL)
+ goto error_return;
+ fprintf (f, "%s", string);
+ }
+ fprintf (f, "\n");
+ }
+
+ free (dynbuf);
+ dynbuf = NULL;
+ }
+
+ if ((elf_dynverdef (abfd) != 0 && elf_tdata (abfd)->verdef == NULL)
+ || (elf_dynverref (abfd) != 0 && elf_tdata (abfd)->verref == NULL))
+ {
+ if (! _bfd_elf_slurp_version_tables (abfd))
+ return false;
+ }
+
+ if (elf_dynverdef (abfd) != 0)
+ {
+ Elf_Internal_Verdef *t;
+
+ fprintf (f, _("\nVersion definitions:\n"));
+ for (t = elf_tdata (abfd)->verdef; t != NULL; t = t->vd_nextdef)
+ {
+ fprintf (f, "%d 0x%2.2x 0x%8.8lx %s\n", t->vd_ndx,
+ t->vd_flags, t->vd_hash, t->vd_nodename);
+ if (t->vd_auxptr->vda_nextptr != NULL)
+ {
+ Elf_Internal_Verdaux *a;
+
+ fprintf (f, "\t");
+ for (a = t->vd_auxptr->vda_nextptr;
+ a != NULL;
+ a = a->vda_nextptr)
+ fprintf (f, "%s ", a->vda_nodename);
+ fprintf (f, "\n");
+ }
+ }
+ }
+
+ if (elf_dynverref (abfd) != 0)
+ {
+ Elf_Internal_Verneed *t;
+
+ fprintf (f, _("\nVersion References:\n"));
+ for (t = elf_tdata (abfd)->verref; t != NULL; t = t->vn_nextref)
+ {
+ Elf_Internal_Vernaux *a;
+
+ fprintf (f, _(" required from %s:\n"), t->vn_filename);
+ for (a = t->vn_auxptr; a != NULL; a = a->vna_nextptr)
+ fprintf (f, " 0x%8.8lx 0x%2.2x %2.2d %s\n", a->vna_hash,
+ a->vna_flags, a->vna_other, a->vna_nodename);
+ }
+ }
+
+ return true;
+
+ error_return:
+ if (dynbuf != NULL)
+ free (dynbuf);
+ return false;
+}
+
/* Display ELF-specific fields of a symbol. */
+
void
-bfd_elf_print_symbol (ignore_abfd, filep, symbol, how)
- bfd *ignore_abfd;
+bfd_elf_print_symbol (abfd, filep, symbol, how)
+ bfd *abfd;
PTR filep;
asymbol *symbol;
bfd_print_symbol_type how;
(bfd_is_com_section (symbol->section)
? ((elf_symbol_type *) symbol)->internal_elf_sym.st_value
: ((elf_symbol_type *) symbol)->internal_elf_sym.st_size));
+
+ /* If we have version information, print it. */
+ if (elf_tdata (abfd)->dynversym_section != 0
+ && (elf_tdata (abfd)->dynverdef_section != 0
+ || elf_tdata (abfd)->dynverref_section != 0))
+ {
+ unsigned int vernum;
+ const char *version_string;
+
+ vernum = ((elf_symbol_type *) symbol)->version & VERSYM_VERSION;
+
+ if (vernum == 0)
+ version_string = "";
+ else if (vernum == 1)
+ version_string = "Base";
+ else if (vernum <= elf_tdata (abfd)->cverdefs)
+ version_string =
+ elf_tdata (abfd)->verdef[vernum - 1].vd_nodename;
+ else
+ {
+ Elf_Internal_Verneed *t;
+
+ version_string = "";
+ for (t = elf_tdata (abfd)->verref;
+ t != NULL;
+ t = t->vn_nextref)
+ {
+ Elf_Internal_Vernaux *a;
+
+ for (a = t->vn_auxptr; a != NULL; a = a->vna_nextptr)
+ {
+ if (a->vna_other == vernum)
+ {
+ version_string = a->vna_nodename;
+ break;
+ }
+ }
+ }
+ }
+
+ if ((((elf_symbol_type *) symbol)->version & VERSYM_HIDDEN) == 0)
+ fprintf (file, " %-11s", version_string);
+ else
+ {
+ int i;
+
+ fprintf (file, " (%s)", version_string);
+ for (i = 10 - strlen (version_string); i > 0; --i)
+ putc (' ', file);
+ }
+ }
+
+ /* If the st_other field is not zero, print it. */
+ if (((elf_symbol_type *) symbol)->internal_elf_sym.st_other != 0)
+ fprintf (file, " 0x%02x",
+ ((unsigned int)
+ ((elf_symbol_type *) symbol)->internal_elf_sym.st_other));
+
fprintf (file, " %s", symbol->name);
}
break;
ret = ((struct elf_link_hash_entry *)
bfd_hash_allocate (table, sizeof (struct elf_link_hash_entry)));
if (ret == (struct elf_link_hash_entry *) NULL)
- {
- bfd_set_error (bfd_error_no_memory);
- return (struct bfd_hash_entry *) ret;
- }
+ return (struct bfd_hash_entry *) ret;
/* Call the allocation method of the superclass. */
ret = ((struct elf_link_hash_entry *)
ret->dynindx = -1;
ret->dynstr_index = 0;
ret->weakdef = NULL;
- ret->got_offset = (bfd_vma) -1;
- ret->plt_offset = (bfd_vma) -1;
+ ret->got.offset = (bfd_vma) -1;
+ ret->plt.offset = (bfd_vma) -1;
+ ret->linker_section_pointer = (elf_linker_section_pointers_t *)0;
+ ret->verinfo.verdef = NULL;
+ ret->vtable_entries_used = NULL;
+ ret->vtable_entries_size = 0;
+ ret->vtable_parent = NULL;
ret->type = STT_NOTYPE;
- ret->elf_link_hash_flags = 0;
+ ret->other = 0;
+ /* Assume that we have been called by a non-ELF symbol reader.
+ This flag is then reset by the code which reads an ELF input
+ file. This ensures that a symbol created by a non-ELF symbol
+ reader will have the flag set correctly. */
+ ret->elf_link_hash_flags = ELF_LINK_NON_ELF;
}
return (struct bfd_hash_entry *) ret;
table->dynstr = NULL;
table->bucketcount = 0;
table->needed = NULL;
+ table->hgot = NULL;
+ table->stab_info = NULL;
return _bfd_link_hash_table_init (&table->root, abfd, newfunc);
}
ret = ((struct elf_link_hash_table *)
bfd_alloc (abfd, sizeof (struct elf_link_hash_table)));
if (ret == (struct elf_link_hash_table *) NULL)
- {
- bfd_set_error (bfd_error_no_memory);
- return NULL;
- }
+ return NULL;
if (! _bfd_elf_link_hash_table_init (ret, abfd, _bfd_elf_link_hash_newfunc))
{
bfd *abfd;
const char *name;
{
- elf_dt_needed_name (abfd) = name;
+ if (bfd_get_flavour (abfd) == bfd_target_elf_flavour
+ && bfd_get_format (abfd) == bfd_object)
+ elf_dt_name (abfd) = name;
}
-/* Get the list of DT_NEEDED entries for a link. */
+/* Get the list of DT_NEEDED entries for a link. This is a hook for
+ the linker ELF emulation code. */
-struct bfd_elf_link_needed_list *
+struct bfd_link_needed_list *
bfd_elf_get_needed_list (abfd, info)
- bfd *abfd;
+ bfd *abfd ATTRIBUTE_UNUSED;
struct bfd_link_info *info;
{
+ if (info->hash->creator->flavour != bfd_target_elf_flavour)
+ return NULL;
return elf_hash_table (info)->needed;
}
+
+/* Get the name actually used for a dynamic object for a link. This
+ is the SONAME entry if there is one. Otherwise, it is the string
+ passed to bfd_elf_set_dt_needed_name, or it is the filename. */
+
+const char *
+bfd_elf_get_dt_soname (abfd)
+ bfd *abfd;
+{
+ if (bfd_get_flavour (abfd) == bfd_target_elf_flavour
+ && bfd_get_format (abfd) == bfd_object)
+ return elf_dt_name (abfd);
+ return NULL;
+}
+
+/* Get the list of DT_NEEDED entries from a BFD. This is a hook for
+ the ELF linker emulation code. */
+
+boolean
+bfd_elf_get_bfd_needed_list (abfd, pneeded)
+ bfd *abfd;
+ struct bfd_link_needed_list **pneeded;
+{
+ asection *s;
+ bfd_byte *dynbuf = NULL;
+ int elfsec;
+ unsigned long link;
+ bfd_byte *extdyn, *extdynend;
+ size_t extdynsize;
+ void (*swap_dyn_in) PARAMS ((bfd *, const PTR, Elf_Internal_Dyn *));
+
+ *pneeded = NULL;
+
+ if (bfd_get_flavour (abfd) != bfd_target_elf_flavour
+ || bfd_get_format (abfd) != bfd_object)
+ return true;
+
+ s = bfd_get_section_by_name (abfd, ".dynamic");
+ if (s == NULL || s->_raw_size == 0)
+ return true;
+
+ dynbuf = (bfd_byte *) bfd_malloc (s->_raw_size);
+ if (dynbuf == NULL)
+ goto error_return;
+
+ if (! bfd_get_section_contents (abfd, s, (PTR) dynbuf, (file_ptr) 0,
+ s->_raw_size))
+ goto error_return;
+
+ elfsec = _bfd_elf_section_from_bfd_section (abfd, s);
+ if (elfsec == -1)
+ goto error_return;
+
+ link = elf_elfsections (abfd)[elfsec]->sh_link;
+
+ extdynsize = get_elf_backend_data (abfd)->s->sizeof_dyn;
+ swap_dyn_in = get_elf_backend_data (abfd)->s->swap_dyn_in;
+
+ extdyn = dynbuf;
+ extdynend = extdyn + s->_raw_size;
+ for (; extdyn < extdynend; extdyn += extdynsize)
+ {
+ Elf_Internal_Dyn dyn;
+
+ (*swap_dyn_in) (abfd, (PTR) extdyn, &dyn);
+
+ if (dyn.d_tag == DT_NULL)
+ break;
+
+ if (dyn.d_tag == DT_NEEDED)
+ {
+ const char *string;
+ struct bfd_link_needed_list *l;
+
+ string = bfd_elf_string_from_elf_section (abfd, link,
+ dyn.d_un.d_val);
+ if (string == NULL)
+ goto error_return;
+
+ l = (struct bfd_link_needed_list *) bfd_alloc (abfd, sizeof *l);
+ if (l == NULL)
+ goto error_return;
+
+ l->by = abfd;
+ l->name = string;
+ l->next = *pneeded;
+ *pneeded = l;
+ }
+ }
+
+ free (dynbuf);
+
+ return true;
+
+ error_return:
+ if (dynbuf != NULL)
+ free (dynbuf);
+ return false;
+}
\f
/* Allocate an ELF string table--force the first byte to be zero. */
case SHT_DYNAMIC: /* Dynamic linking information. */
case SHT_NOBITS: /* .bss section. */
case SHT_HASH: /* .hash section. */
+ case SHT_NOTE: /* .note section. */
return _bfd_elf_make_section_from_shdr (abfd, hdr, name);
case SHT_SYMTAB: /* A symbol table */
BFD_ASSERT (elf_onesymtab (abfd) == 0);
elf_onesymtab (abfd) = shindex;
elf_tdata (abfd)->symtab_hdr = *hdr;
- elf_elfsections (abfd)[shindex] = &elf_tdata (abfd)->symtab_hdr;
+ elf_elfsections (abfd)[shindex] = hdr = &elf_tdata (abfd)->symtab_hdr;
abfd->flags |= HAS_SYMS;
/* Sometimes a shared object will map in the symbol table. If
BFD_ASSERT (elf_dynsymtab (abfd) == 0);
elf_dynsymtab (abfd) = shindex;
elf_tdata (abfd)->dynsymtab_hdr = *hdr;
- elf_elfsections (abfd)[shindex] = &elf_tdata (abfd)->dynsymtab_hdr;
+ elf_elfsections (abfd)[shindex] = hdr = &elf_tdata (abfd)->dynsymtab_hdr;
abfd->flags |= HAS_SYMS;
/* Besides being a symbol table, we also treat this as a regular
if (elf_dynsymtab (abfd) == i)
{
elf_tdata (abfd)->dynstrtab_hdr = *hdr;
- elf_elfsections (abfd)[shindex] =
+ elf_elfsections (abfd)[shindex] = hdr =
&elf_tdata (abfd)->dynstrtab_hdr;
/* We also treat this as a regular section, so
that objcopy can handle it. */
{
asection *target_sect;
Elf_Internal_Shdr *hdr2;
- int use_rela_p = get_elf_backend_data (abfd)->use_rela_p;
+
+ /* Check for a bogus link to avoid crashing. */
+ if (hdr->sh_link >= ehdr->e_shnum)
+ {
+ ((*_bfd_error_handler)
+ (_("%s: invalid link %lu for reloc section %s (index %u)"),
+ bfd_get_filename (abfd), hdr->sh_link, name, shindex));
+ return _bfd_elf_make_section_from_shdr (abfd, hdr, name);
+ }
/* For some incomprehensible reason Oracle distributes
libraries for Solaris in which some of the objects have
if (hdr->sh_link != elf_onesymtab (abfd))
return _bfd_elf_make_section_from_shdr (abfd, hdr, name);
- /* Don't allow REL relocations on a machine that uses RELA and
- vice versa. */
- /* @@ Actually, the generic ABI does suggest that both might be
- used in one file. But the four ABI Processor Supplements I
- have access to right now all specify that only one is used on
- each of those architectures. It's conceivable that, e.g., a
- bunch of absolute 32-bit relocs might be more compact in REL
- form even on a RELA machine... */
- BFD_ASSERT (use_rela_p
- ? (hdr->sh_type == SHT_RELA
- && hdr->sh_entsize == bed->s->sizeof_rela)
- : (hdr->sh_type == SHT_REL
- && hdr->sh_entsize == bed->s->sizeof_rel));
-
if (! bfd_section_from_shdr (abfd, hdr->sh_info))
return false;
target_sect = bfd_section_from_elf_index (abfd, hdr->sh_info);
if (target_sect == NULL)
return false;
- hdr2 = &elf_section_data (target_sect)->rel_hdr;
+ if ((target_sect->flags & SEC_RELOC) == 0
+ || target_sect->reloc_count == 0)
+ hdr2 = &elf_section_data (target_sect)->rel_hdr;
+ else
+ {
+ BFD_ASSERT (elf_section_data (target_sect)->rel_hdr2 == NULL);
+ hdr2 = (Elf_Internal_Shdr *) bfd_alloc (abfd, sizeof (*hdr2));
+ elf_section_data (target_sect)->rel_hdr2 = hdr2;
+ }
*hdr2 = *hdr;
elf_elfsections (abfd)[shindex] = hdr2;
- target_sect->reloc_count = hdr->sh_size / hdr->sh_entsize;
+ target_sect->reloc_count += hdr->sh_size / hdr->sh_entsize;
target_sect->flags |= SEC_RELOC;
target_sect->relocation = NULL;
target_sect->rel_filepos = hdr->sh_offset;
+ /* In the section to which the relocations apply, mark whether
+ its relocations are of the REL or RELA variety. */
+ elf_section_data (target_sect)->use_rela_p
+ = (hdr->sh_type == SHT_RELA);
abfd->flags |= HAS_RELOC;
return true;
}
break;
- case SHT_NOTE:
+ case SHT_GNU_verdef:
+ elf_dynverdef (abfd) = shindex;
+ elf_tdata (abfd)->dynverdef_hdr = *hdr;
+ return _bfd_elf_make_section_from_shdr (abfd, hdr, name);
+ break;
+
+ case SHT_GNU_versym:
+ elf_dynversym (abfd) = shindex;
+ elf_tdata (abfd)->dynversym_hdr = *hdr;
+ return _bfd_elf_make_section_from_shdr (abfd, hdr, name);
+ break;
+
+ case SHT_GNU_verneed:
+ elf_dynverref (abfd) = shindex;
+ elf_tdata (abfd)->dynverref_hdr = *hdr;
+ return _bfd_elf_make_section_from_shdr (abfd, hdr, name);
break;
case SHT_SHLIB:
{
struct bfd_elf_section_data *sdata;
- sdata = (struct bfd_elf_section_data *) bfd_alloc (abfd, sizeof (*sdata));
+ sdata = (struct bfd_elf_section_data *) bfd_zalloc (abfd, sizeof (*sdata));
if (!sdata)
- {
- bfd_set_error (bfd_error_no_memory);
- return false;
- }
+ return false;
sec->used_by_bfd = (PTR) sdata;
- memset (sdata, 0, sizeof (*sdata));
+
+ /* Indicate whether or not this section should use RELA relocations. */
+ sdata->use_rela_p
+ = get_elf_backend_data (abfd)->default_use_rela_p;
+
return true;
}
char namebuf[64];
int split;
- split = ((hdr->p_memsz > 0) &&
- (hdr->p_filesz > 0) &&
- (hdr->p_memsz > hdr->p_filesz));
+ split = ((hdr->p_memsz > 0)
+ && (hdr->p_filesz > 0)
+ && (hdr->p_memsz > hdr->p_filesz));
sprintf (namebuf, split ? "segment%da" : "segment%d", index);
name = bfd_alloc (abfd, strlen (namebuf) + 1);
if (!name)
- {
- bfd_set_error (bfd_error_no_memory);
- return false;
- }
+ return false;
strcpy (name, namebuf);
newsect = bfd_make_section (abfd, name);
if (newsect == NULL)
sprintf (namebuf, "segment%db", index);
name = bfd_alloc (abfd, strlen (namebuf) + 1);
if (!name)
- {
- bfd_set_error (bfd_error_no_memory);
- return false;
- }
+ return false;
strcpy (name, namebuf);
newsect = bfd_make_section (abfd, name);
if (newsect == NULL)
return true;
}
+/* Initialize REL_HDR, the section-header for new section, containing
+ relocations against ASECT. If USE_RELA_P is true, we use RELA
+ relocations; otherwise, we use REL relocations. */
+
+boolean
+_bfd_elf_init_reloc_shdr (abfd, rel_hdr, asect, use_rela_p)
+ bfd *abfd;
+ Elf_Internal_Shdr *rel_hdr;
+ asection *asect;
+ boolean use_rela_p;
+{
+ char *name;
+ struct elf_backend_data *bed;
+
+ bed = get_elf_backend_data (abfd);
+ name = bfd_alloc (abfd, sizeof ".rela" + strlen (asect->name));
+ if (name == NULL)
+ return false;
+ sprintf (name, "%s%s", use_rela_p ? ".rela" : ".rel", asect->name);
+ rel_hdr->sh_name =
+ (unsigned int) _bfd_stringtab_add (elf_shstrtab (abfd), name,
+ true, false);
+ if (rel_hdr->sh_name == (unsigned int) -1)
+ return false;
+ rel_hdr->sh_type = use_rela_p ? SHT_RELA : SHT_REL;
+ rel_hdr->sh_entsize = (use_rela_p
+ ? bed->s->sizeof_rela
+ : bed->s->sizeof_rel);
+ rel_hdr->sh_addralign = bed->s->file_align;
+ rel_hdr->sh_flags = 0;
+ rel_hdr->sh_addr = 0;
+ rel_hdr->sh_size = 0;
+ rel_hdr->sh_offset = 0;
+
+ return true;
+}
+
/* Set up an ELF internal section header for a section. */
/*ARGSUSED*/
this_hdr->sh_flags = 0;
- /* FIXME: This should really use vma, rather than lma. However,
- that would mean that the lma information was lost, which would
- mean that the AT keyword in linker scripts would not work.
- Fortunately, native scripts do not use the AT keyword, so we can
- get away with using lma here. The right way to handle this is to
- 1) read the program headers as well as the section headers, and
- set the lma fields of the BFD sections based on the p_paddr
- fields of the program headers, and 2) set the p_paddr fields of
- the program headers based on the section lma fields when writing
- them out. */
- if ((asect->flags & SEC_ALLOC) != 0)
- this_hdr->sh_addr = asect->lma;
+ if ((asect->flags & SEC_ALLOC) != 0
+ || asect->user_set_vma)
+ this_hdr->sh_addr = asect->vma;
else
this_hdr->sh_addr = 0;
this_hdr->sh_offset = 0;
this_hdr->sh_size = asect->_raw_size;
this_hdr->sh_link = 0;
- this_hdr->sh_info = 0;
this_hdr->sh_addralign = 1 << asect->alignment_power;
- this_hdr->sh_entsize = 0;
+ /* The sh_entsize and sh_info fields may have been set already by
+ copy_private_section_data. */
this_hdr->bfd_section = asect;
this_hdr->contents = NULL;
else if (strcmp (asect->name, ".hash") == 0)
{
this_hdr->sh_type = SHT_HASH;
- this_hdr->sh_entsize = bed->s->arch_size / 8;
+ this_hdr->sh_entsize = bed->s->sizeof_hash_entry;
}
else if (strcmp (asect->name, ".dynsym") == 0)
{
this_hdr->sh_entsize = bed->s->sizeof_dyn;
}
else if (strncmp (asect->name, ".rela", 5) == 0
- && get_elf_backend_data (abfd)->use_rela_p)
+ && get_elf_backend_data (abfd)->may_use_rela_p)
{
this_hdr->sh_type = SHT_RELA;
this_hdr->sh_entsize = bed->s->sizeof_rela;
}
else if (strncmp (asect->name, ".rel", 4) == 0
- && ! get_elf_backend_data (abfd)->use_rela_p)
+ && get_elf_backend_data (abfd)->may_use_rel_p)
{
this_hdr->sh_type = SHT_REL;
this_hdr->sh_entsize = bed->s->sizeof_rel;
}
- else if (strcmp (asect->name, ".note") == 0)
+ else if (strncmp (asect->name, ".note", 5) == 0)
this_hdr->sh_type = SHT_NOTE;
else if (strncmp (asect->name, ".stab", 5) == 0
&& strcmp (asect->name + strlen (asect->name) - 3, "str") == 0)
this_hdr->sh_type = SHT_STRTAB;
+ else if (strcmp (asect->name, ".gnu.version") == 0)
+ {
+ this_hdr->sh_type = SHT_GNU_versym;
+ this_hdr->sh_entsize = sizeof (Elf_External_Versym);
+ }
+ else if (strcmp (asect->name, ".gnu.version_d") == 0)
+ {
+ this_hdr->sh_type = SHT_GNU_verdef;
+ this_hdr->sh_entsize = 0;
+ /* objcopy or strip will copy over sh_info, but may not set
+ cverdefs. The linker will set cverdefs, but sh_info will be
+ zero. */
+ if (this_hdr->sh_info == 0)
+ this_hdr->sh_info = elf_tdata (abfd)->cverdefs;
+ else
+ BFD_ASSERT (elf_tdata (abfd)->cverdefs == 0
+ || this_hdr->sh_info == elf_tdata (abfd)->cverdefs);
+ }
+ else if (strcmp (asect->name, ".gnu.version_r") == 0)
+ {
+ this_hdr->sh_type = SHT_GNU_verneed;
+ this_hdr->sh_entsize = 0;
+ /* objcopy or strip will copy over sh_info, but may not set
+ cverrefs. The linker will set cverrefs, but sh_info will be
+ zero. */
+ if (this_hdr->sh_info == 0)
+ this_hdr->sh_info = elf_tdata (abfd)->cverrefs;
+ else
+ BFD_ASSERT (elf_tdata (abfd)->cverrefs == 0
+ || this_hdr->sh_info == elf_tdata (abfd)->cverrefs);
+ }
else if ((asect->flags & SEC_ALLOC) != 0
&& (asect->flags & SEC_LOAD) != 0)
this_hdr->sh_type = SHT_PROGBITS;
else if ((asect->flags & SEC_ALLOC) != 0
&& ((asect->flags & SEC_LOAD) == 0))
- {
- BFD_ASSERT (strcmp (asect->name, ".bss") == 0
- || strcmp (asect->name, ".sbss") == 0
- || strcmp (asect->name, ".scommon") == 0
- || strcmp (asect->name, "COMMON") == 0);
- this_hdr->sh_type = SHT_NOBITS;
- }
+ this_hdr->sh_type = SHT_NOBITS;
else
{
/* Who knows? */
this_hdr->sh_flags |= SHF_EXECINSTR;
/* Check for processor-specific section types. */
- {
- struct elf_backend_data *bed = get_elf_backend_data (abfd);
-
- if (bed->elf_backend_fake_sections)
- (*bed->elf_backend_fake_sections) (abfd, this_hdr, asect);
- }
+ if (bed->elf_backend_fake_sections)
+ (*bed->elf_backend_fake_sections) (abfd, this_hdr, asect);
/* If the section has relocs, set up a section header for the
- SHT_REL[A] section. */
- if ((asect->flags & SEC_RELOC) != 0)
- {
- Elf_Internal_Shdr *rela_hdr;
- int use_rela_p = get_elf_backend_data (abfd)->use_rela_p;
- char *name;
-
- rela_hdr = &elf_section_data (asect)->rel_hdr;
- name = bfd_alloc (abfd, sizeof ".rela" + strlen (asect->name));
- if (name == NULL)
- {
- bfd_set_error (bfd_error_no_memory);
- *failedptr = true;
- return;
- }
- sprintf (name, "%s%s", use_rela_p ? ".rela" : ".rel", asect->name);
- rela_hdr->sh_name =
- (unsigned int) _bfd_stringtab_add (elf_shstrtab (abfd), name,
- true, false);
- if (rela_hdr->sh_name == (unsigned int) -1)
- {
- *failedptr = true;
- return;
- }
- rela_hdr->sh_type = use_rela_p ? SHT_RELA : SHT_REL;
- rela_hdr->sh_entsize = (use_rela_p
- ? bed->s->sizeof_rela
- : bed->s->sizeof_rel);
- rela_hdr->sh_addralign = bed->s->file_align;
- rela_hdr->sh_flags = 0;
- rela_hdr->sh_addr = 0;
- rela_hdr->sh_size = 0;
- rela_hdr->sh_offset = 0;
- }
+ SHT_REL[A] section. If two relocation sections are required for
+ this section, it is up to the processor-specific back-end to
+ create the other. */
+ if ((asect->flags & SEC_RELOC) != 0
+ && !_bfd_elf_init_reloc_shdr (abfd,
+ &elf_section_data (asect)->rel_hdr,
+ asect,
+ elf_section_data (asect)->use_rela_p))
+ *failedptr = true;
}
/* Assign all ELF section numbers. The dummy first section is handled here
d->rel_idx = 0;
else
d->rel_idx = section_number++;
+
+ if (d->rel_hdr2)
+ d->rel_idx2 = section_number++;
+ else
+ d->rel_idx2 = 0;
}
t->shstrtab_section = section_number++;
elf_elfheader (abfd)->e_shstrndx = t->shstrtab_section;
t->shstrtab_hdr.sh_size = _bfd_stringtab_size (elf_shstrtab (abfd));
- if (abfd->symcount > 0)
+ if (bfd_get_symcount (abfd) > 0)
{
t->symtab_section = section_number++;
t->strtab_section = section_number++;
i_shdrp = ((Elf_Internal_Shdr **)
bfd_alloc (abfd, section_number * sizeof (Elf_Internal_Shdr *)));
if (i_shdrp == NULL)
- {
- bfd_set_error (bfd_error_no_memory);
- return false;
- }
+ return false;
i_shdrp[0] = ((Elf_Internal_Shdr *)
bfd_alloc (abfd, sizeof (Elf_Internal_Shdr)));
if (i_shdrp[0] == NULL)
{
bfd_release (abfd, i_shdrp);
- bfd_set_error (bfd_error_no_memory);
return false;
}
memset (i_shdrp[0], 0, sizeof (Elf_Internal_Shdr));
elf_elfsections (abfd) = i_shdrp;
i_shdrp[t->shstrtab_section] = &t->shstrtab_hdr;
- if (abfd->symcount > 0)
+ if (bfd_get_symcount (abfd) > 0)
{
i_shdrp[t->symtab_section] = &t->symtab_hdr;
i_shdrp[t->strtab_section] = &t->strtab_hdr;
i_shdrp[d->this_idx] = &d->this_hdr;
if (d->rel_idx != 0)
i_shdrp[d->rel_idx] = &d->rel_hdr;
+ if (d->rel_idx2 != 0)
+ i_shdrp[d->rel_idx2] = d->rel_hdr2;
/* Fill in the sh_link and sh_info fields while we're at it. */
d->rel_hdr.sh_link = t->symtab_section;
d->rel_hdr.sh_info = d->this_idx;
}
+ if (d->rel_idx2 != 0)
+ {
+ d->rel_hdr2->sh_link = t->symtab_section;
+ d->rel_hdr2->sh_info = d->this_idx;
+ }
switch (d->this_hdr.sh_type)
{
char *alc;
len = strlen (sec->name);
- alc = (char *) malloc (len - 2);
+ alc = (char *) bfd_malloc (len - 2);
if (alc == NULL)
- {
- bfd_set_error (bfd_error_no_memory);
- return false;
- }
+ return false;
strncpy (alc, sec->name, len - 3);
alc[len - 3] = '\0';
s = bfd_get_section_by_name (abfd, alc);
case SHT_DYNAMIC:
case SHT_DYNSYM:
+ case SHT_GNU_verneed:
+ case SHT_GNU_verdef:
/* sh_link is the section header index of the string table
- used for the dynamic entries or symbol table. */
+ used for the dynamic entries, or the symbol table, or the
+ version strings. */
s = bfd_get_section_by_name (abfd, ".dynstr");
if (s != NULL)
d->this_hdr.sh_link = elf_section_data (s)->this_idx;
break;
case SHT_HASH:
+ case SHT_GNU_versym:
/* sh_link is the section header index of the symbol table
- this hash table is for. */
+ this hash table or version table is for. */
s = bfd_get_section_by_name (abfd, ".dynsym");
if (s != NULL)
d->this_hdr.sh_link = elf_section_data (s)->this_idx;
int idx;
asection *asect;
asymbol **new_syms;
+ asymbol *sym;
#ifdef DEBUG
fprintf (stderr, "elf_map_symbols\n");
max_index++;
sect_syms = (asymbol **) bfd_zalloc (abfd, max_index * sizeof (asymbol *));
if (sect_syms == NULL)
- {
- bfd_set_error (bfd_error_no_memory);
- return false;
- }
+ return false;
elf_section_syms (abfd) = sect_syms;
for (idx = 0; idx < symcount; idx++)
{
- if ((syms[idx]->flags & BSF_SECTION_SYM) != 0
- && syms[idx]->value == 0)
+ sym = syms[idx];
+
+ if ((sym->flags & BSF_SECTION_SYM) != 0
+ && sym->value == 0)
{
asection *sec;
- sec = syms[idx]->section;
+ sec = sym->section;
+
if (sec->owner != NULL)
{
if (sec->owner != abfd)
{
if (sec->output_offset != 0)
continue;
+
sec = sec->output_section;
+
+ /* Empty sections in the input files may have had a section
+ symbol created for them. (See the comment near the end of
+ _bfd_generic_link_output_symbols in linker.c). If the linker
+ script discards such sections then we will reach this point.
+ Since we know that we cannot avoid this case, we detect it
+ and skip the abort and the assignment to the sect_syms array.
+ To reproduce this particular case try running the linker
+ testsuite test ld-scripts/weak.exp for an ELF port that uses
+ the generic linker. */
+ if (sec->owner == NULL)
+ continue;
+
BFD_ASSERT (sec->owner == abfd);
}
sect_syms[sec->index] = syms[idx];
for (asect = abfd->sections; asect; asect = asect->next)
{
- asymbol *sym;
-
if (sect_syms[asect->index] != NULL)
continue;
num_sections++;
#ifdef DEBUG
fprintf (stderr,
- "creating section symbol, name = %s, value = 0x%.8lx, index = %d, section = 0x%.8lx\n",
+ _("creating section symbol, name = %s, value = 0x%.8lx, index = %d, section = 0x%.8lx\n"),
asect->name, (long) asect->vma, asect->index, (long) asect);
#endif
}
bfd_alloc (abfd,
(num_locals + num_globals) * sizeof (asymbol *)));
if (new_syms == NULL)
- {
- bfd_set_error (bfd_error_no_memory);
- return false;
- }
+ return false;
for (idx = 0; idx < symcount; idx++)
{
return true;
}
+/* Align to the maximum file alignment that could be required for any
+ ELF data structure. */
+
+static INLINE file_ptr align_file_position PARAMS ((file_ptr, int));
+static INLINE file_ptr
+align_file_position (off, align)
+ file_ptr off;
+ int align;
+{
+ return (off + align - 1) & ~(align - 1);
+}
+
+/* Assign a file position to a section, optionally aligning to the
+ required section alignment. */
+
+INLINE file_ptr
+_bfd_elf_assign_file_position_for_section (i_shdrp, offset, align)
+ Elf_Internal_Shdr *i_shdrp;
+ file_ptr offset;
+ boolean align;
+{
+ if (align)
+ {
+ unsigned int al;
+
+ al = i_shdrp->sh_addralign;
+ if (al > 1)
+ offset = BFD_ALIGN (offset, al);
+ }
+ i_shdrp->sh_offset = offset;
+ if (i_shdrp->bfd_section != NULL)
+ i_shdrp->bfd_section->filepos = offset;
+ if (i_shdrp->sh_type != SHT_NOBITS)
+ offset += i_shdrp->sh_size;
+ return offset;
+}
+
/* Compute the file positions we are going to put the sections at, and
otherwise prepare to begin writing out the ELF file. If LINK_INFO
is not NULL, this is being called by the ELF backend linker. */
if (! prep_headers (abfd))
return false;
+ /* Post process the headers if necessary. */
+ if (bed->elf_backend_post_process_headers)
+ (*bed->elf_backend_post_process_headers) (abfd, link_info);
+
failed = false;
bfd_map_over_sections (abfd, elf_fake_sections, &failed);
if (failed)
return false;
/* The backend linker builds symbol table information itself. */
- if (link_info == NULL)
+ if (link_info == NULL && bfd_get_symcount (abfd) > 0)
{
- if (! swap_out_syms (abfd, &strtab))
+ /* Non-zero if doing a relocatable link. */
+ int relocatable_p = ! (abfd->flags & (EXEC_P | DYNAMIC));
+
+ if (! swap_out_syms (abfd, &strtab, relocatable_p))
return false;
}
shstrtab_hdr->sh_entsize = 0;
shstrtab_hdr->sh_link = 0;
shstrtab_hdr->sh_info = 0;
- /* sh_offset is set in assign_file_positions_for_symtabs_and_strtabs. */
+ /* sh_offset is set in assign_file_positions_except_relocs. */
shstrtab_hdr->sh_addralign = 1;
- if (!assign_file_positions_except_relocs (abfd,
- link_info == NULL ? true : false))
+ if (!assign_file_positions_except_relocs (abfd))
return false;
- if (link_info == NULL)
+ if (link_info == NULL && bfd_get_symcount (abfd) > 0)
{
+ file_ptr off;
+ Elf_Internal_Shdr *hdr;
+
+ off = elf_tdata (abfd)->next_file_pos;
+
+ hdr = &elf_tdata (abfd)->symtab_hdr;
+ off = _bfd_elf_assign_file_position_for_section (hdr, off, true);
+
+ hdr = &elf_tdata (abfd)->strtab_hdr;
+ off = _bfd_elf_assign_file_position_for_section (hdr, off, true);
+
+ elf_tdata (abfd)->next_file_pos = off;
+
/* Now that we know where the .strtab section goes, write it
out. */
- if ((bfd_seek (abfd, elf_tdata (abfd)->strtab_hdr.sh_offset, SEEK_SET)
- != 0)
+ if (bfd_seek (abfd, hdr->sh_offset, SEEK_SET) != 0
|| ! _bfd_stringtab_emit (abfd, strtab))
return false;
_bfd_stringtab_free (strtab);
return true;
}
+/* Create a mapping from a set of sections to a program segment. */
-/* Align to the maximum file alignment that could be required for any
- ELF data structure. */
-
-static INLINE file_ptr align_file_position PARAMS ((file_ptr, int));
-static INLINE file_ptr
-align_file_position (off, align)
- file_ptr off;
- int align;
+static INLINE struct elf_segment_map *
+make_mapping (abfd, sections, from, to, phdr)
+ bfd *abfd;
+ asection **sections;
+ unsigned int from;
+ unsigned int to;
+ boolean phdr;
{
- return (off + align - 1) & ~(align - 1);
-}
-
-/* Assign a file position to a section, aligning to the required
- section alignment. */
+ struct elf_segment_map *m;
+ unsigned int i;
+ asection **hdrpp;
-INLINE file_ptr
-_bfd_elf_assign_file_position_for_section (i_shdrp, offset)
- Elf_Internal_Shdr *i_shdrp;
- file_ptr offset;
-{
- unsigned int al;
+ m = ((struct elf_segment_map *)
+ bfd_zalloc (abfd,
+ (sizeof (struct elf_segment_map)
+ + (to - from - 1) * sizeof (asection *))));
+ if (m == NULL)
+ return NULL;
+ m->next = NULL;
+ m->p_type = PT_LOAD;
+ for (i = from, hdrpp = sections + from; i < to; i++, hdrpp++)
+ m->sections[i - from] = *hdrpp;
+ m->count = to - from;
- /* Align the offst. */
- al = i_shdrp->sh_addralign;
- if (al > 1)
- offset = BFD_ALIGN (offset, al);
+ if (from == 0 && phdr)
+ {
+ /* Include the headers in the first PT_LOAD segment. */
+ m->includes_filehdr = 1;
+ m->includes_phdrs = 1;
+ }
- i_shdrp->sh_offset = offset;
- if (i_shdrp->bfd_section != NULL)
- i_shdrp->bfd_section->filepos = offset;
- if (i_shdrp->sh_type != SHT_NOBITS)
- offset += i_shdrp->sh_size;
- return offset;
+ return m;
}
-/* Get the size of the program header.
-
- SORTED_HDRS, if non-NULL, is an array of COUNT pointers to headers sorted
- by VMA. Non-allocated sections (!SHF_ALLOC) must appear last. All
- section VMAs and sizes are known so we can compute the correct value.
- (??? This may not be perfectly true. What cases do we miss?)
-
- If SORTED_HDRS is NULL we assume there are two segments: text and data
- (exclusive of .interp and .dynamic).
-
- If this is called by the linker before any of the section VMA's are set, it
- can't calculate the correct value for a strange memory layout. This only
- happens when SIZEOF_HEADERS is used in a linker script. In this case,
- SORTED_HDRS is NULL and we assume the normal scenario of one text and one
- data segment (exclusive of .interp and .dynamic).
-
- ??? User written scripts must either not use SIZEOF_HEADERS, or assume there
- will be two segments. */
+/* Set up a mapping from BFD sections to program segments. */
-static bfd_size_type
-get_program_header_size (abfd, sorted_hdrs, count, maxpagesize)
+static boolean
+map_sections_to_segments (abfd)
bfd *abfd;
- Elf_Internal_Shdr **sorted_hdrs;
- unsigned int count;
- bfd_vma maxpagesize;
{
- size_t segs;
+ asection **sections = NULL;
asection *s;
- struct elf_backend_data *bed = get_elf_backend_data (abfd);
-
- /* We can't return a different result each time we're called. */
- if (elf_tdata (abfd)->program_header_size != 0)
- return elf_tdata (abfd)->program_header_size;
+ unsigned int i;
+ unsigned int count;
+ struct elf_segment_map *mfirst;
+ struct elf_segment_map **pm;
+ struct elf_segment_map *m;
+ asection *last_hdr;
+ unsigned int phdr_index;
+ bfd_vma maxpagesize;
+ asection **hdrpp;
+ boolean phdr_in_segment = true;
+ boolean writable;
+ asection *dynsec;
+
+ if (elf_tdata (abfd)->segment_map != NULL)
+ return true;
- if (sorted_hdrs != NULL)
- {
- unsigned int i;
- unsigned int last_type;
- Elf_Internal_Shdr **hdrpp;
- /* What we think the current segment's offset is. */
- bfd_vma p_offset;
- /* What we think the current segment's address is. */
- bfd_vma p_vaddr;
- /* How big we think the current segment is. */
- bfd_vma p_memsz;
- /* What we think the current file offset is. */
- bfd_vma file_offset;
- bfd_vma next_offset;
-
- /* Scan the headers and compute the number of segments required. This
- code is intentionally similar to the code in map_program_segments.
-
- The `sh_offset' field isn't valid at this point, so we keep our own
- running total in `file_offset'.
-
- This works because section VMAs are already known. */
-
- segs = 1;
- /* Make sure the first section goes in the first segment. */
- file_offset = p_offset = sorted_hdrs[0]->sh_addr % maxpagesize;
- p_vaddr = sorted_hdrs[0]->sh_addr;
- p_memsz = 0;
- last_type = SHT_PROGBITS;
-
- for (i = 0, hdrpp = sorted_hdrs; i < count; i++, hdrpp++)
- {
- Elf_Internal_Shdr *hdr;
+ if (bfd_count_sections (abfd) == 0)
+ return true;
- hdr = *hdrpp;
+ /* Select the allocated sections, and sort them. */
- /* Ignore any section which will not be part of the process
- image. */
- if ((hdr->sh_flags & SHF_ALLOC) == 0)
- continue;
-
- /* Keep track of where this and the next sections go.
- The section VMA must equal the file position modulo
- the page size. */
- file_offset += (hdr->sh_addr - file_offset) % maxpagesize;
- next_offset = file_offset;
- if (hdr->sh_type != SHT_NOBITS)
- next_offset = file_offset + hdr->sh_size;
-
- /* If this section fits in the segment we are constructing, add
- it in. */
- if ((file_offset - (p_offset + p_memsz)
- == hdr->sh_addr - (p_vaddr + p_memsz))
- && (last_type != SHT_NOBITS || hdr->sh_type == SHT_NOBITS))
- {
- bfd_size_type adjust;
+ sections = (asection **) bfd_malloc (bfd_count_sections (abfd)
+ * sizeof (asection *));
+ if (sections == NULL)
+ goto error_return;
- adjust = hdr->sh_addr - (p_vaddr + p_memsz);
- p_memsz += hdr->sh_size + adjust;
- file_offset = next_offset;
- last_type = hdr->sh_type;
- continue;
- }
+ i = 0;
+ for (s = abfd->sections; s != NULL; s = s->next)
+ {
+ if ((s->flags & SEC_ALLOC) != 0)
+ {
+ sections[i] = s;
+ ++i;
+ }
+ }
+ BFD_ASSERT (i <= bfd_count_sections (abfd));
+ count = i;
- /* The section won't fit, start a new segment. */
- ++segs;
+ qsort (sections, (size_t) count, sizeof (asection *), elf_sort_sections);
- /* Initialize the segment. */
- p_vaddr = hdr->sh_addr;
- p_memsz = hdr->sh_size;
- p_offset = file_offset;
- file_offset = next_offset;
+ /* Build the mapping. */
- last_type = hdr->sh_type;
- }
- }
- else
- {
- /* Assume we will need exactly two PT_LOAD segments: one for text
- and one for data. */
- segs = 2;
- }
+ mfirst = NULL;
+ pm = &mfirst;
+ /* If we have a .interp section, then create a PT_PHDR segment for
+ the program headers and a PT_INTERP segment for the .interp
+ section. */
s = bfd_get_section_by_name (abfd, ".interp");
if (s != NULL && (s->flags & SEC_LOAD) != 0)
{
- /* If we have a loadable interpreter section, we need a
- PT_INTERP segment. In this case, assume we also need a
- PT_PHDR segment, although that may not be true for all
- targets. */
- segs += 2;
+ m = ((struct elf_segment_map *)
+ bfd_zalloc (abfd, sizeof (struct elf_segment_map)));
+ if (m == NULL)
+ goto error_return;
+ m->next = NULL;
+ m->p_type = PT_PHDR;
+ /* FIXME: UnixWare and Solaris set PF_X, Irix 5 does not. */
+ m->p_flags = PF_R | PF_X;
+ m->p_flags_valid = 1;
+ m->includes_phdrs = 1;
+
+ *pm = m;
+ pm = &m->next;
+
+ m = ((struct elf_segment_map *)
+ bfd_zalloc (abfd, sizeof (struct elf_segment_map)));
+ if (m == NULL)
+ goto error_return;
+ m->next = NULL;
+ m->p_type = PT_INTERP;
+ m->count = 1;
+ m->sections[0] = s;
+
+ *pm = m;
+ pm = &m->next;
}
- if (bfd_get_section_by_name (abfd, ".dynamic") != NULL)
+ /* Look through the sections. We put sections in the same program
+ segment when the start of the second section can be placed within
+ a few bytes of the end of the first section. */
+ last_hdr = NULL;
+ phdr_index = 0;
+ maxpagesize = get_elf_backend_data (abfd)->maxpagesize;
+ writable = false;
+ dynsec = bfd_get_section_by_name (abfd, ".dynamic");
+ if (dynsec != NULL
+ && (dynsec->flags & SEC_LOAD) == 0)
+ dynsec = NULL;
+
+ /* Deal with -Ttext or something similar such that the first section
+ is not adjacent to the program headers. This is an
+ approximation, since at this point we don't know exactly how many
+ program headers we will need. */
+ if (count > 0)
{
- /* We need a PT_DYNAMIC segment. */
- ++segs;
+ bfd_size_type phdr_size;
+
+ phdr_size = elf_tdata (abfd)->program_header_size;
+ if (phdr_size == 0)
+ phdr_size = get_elf_backend_data (abfd)->s->sizeof_phdr;
+ if ((abfd->flags & D_PAGED) == 0
+ || sections[0]->lma < phdr_size
+ || sections[0]->lma % maxpagesize < phdr_size % maxpagesize)
+ phdr_in_segment = false;
}
- /* Let the backend count up any program headers it might need. */
- if (bed->elf_backend_create_program_headers)
- segs = ((*bed->elf_backend_create_program_headers)
- (abfd, (Elf_Internal_Phdr *) NULL, segs));
+ for (i = 0, hdrpp = sections; i < count; i++, hdrpp++)
+ {
+ asection *hdr;
+ boolean new_segment;
- elf_tdata (abfd)->program_header_size = segs * bed->s->sizeof_phdr;
- return elf_tdata (abfd)->program_header_size;
-}
+ hdr = *hdrpp;
-/* Create the program header. OFF is the file offset where the
- program header should be written. FIRST is the first loadable ELF
- section. SORTED_HDRS is the ELF sections sorted by section
- address. PHDR_SIZE is the size of the program header as returned
- by get_program_header_size. */
+ /* See if this section and the last one will fit in the same
+ segment. */
-static file_ptr
-map_program_segments (abfd, off, first, sorted_hdrs, phdr_size)
- bfd *abfd;
- file_ptr off;
- Elf_Internal_Shdr *first;
- Elf_Internal_Shdr **sorted_hdrs;
- bfd_size_type phdr_size;
-{
- Elf_Internal_Phdr phdrs[10];
- unsigned int phdr_count;
- Elf_Internal_Phdr *phdr;
- int phdr_size_adjust;
- unsigned int i;
- Elf_Internal_Shdr **hdrpp;
- asection *sinterp, *sdyn;
- unsigned int last_type;
- Elf_Internal_Ehdr *i_ehdrp;
- struct elf_backend_data *bed = get_elf_backend_data (abfd);
+ if (last_hdr == NULL)
+ {
+ /* If we don't have a segment yet, then we don't need a new
+ one (we build the last one after this loop). */
+ new_segment = false;
+ }
+ else if (last_hdr->lma - last_hdr->vma != hdr->lma - hdr->vma)
+ {
+ /* If this section has a different relation between the
+ virtual address and the load address, then we need a new
+ segment. */
+ new_segment = true;
+ }
+ else if (BFD_ALIGN (last_hdr->lma + last_hdr->_raw_size, maxpagesize)
+ < BFD_ALIGN (hdr->lma, maxpagesize))
+ {
+ /* If putting this section in this segment would force us to
+ skip a page in the segment, then we need a new segment. */
+ new_segment = true;
+ }
+ else if ((last_hdr->flags & SEC_LOAD) == 0
+ && (hdr->flags & SEC_LOAD) != 0)
+ {
+ /* We don't want to put a loadable section after a
+ nonloadable section in the same segment. */
+ new_segment = true;
+ }
+ else if ((abfd->flags & D_PAGED) == 0)
+ {
+ /* If the file is not demand paged, which means that we
+ don't require the sections to be correctly aligned in the
+ file, then there is no other reason for a new segment. */
+ new_segment = false;
+ }
+ else if (! writable
+ && (hdr->flags & SEC_READONLY) == 0
+ && (BFD_ALIGN (last_hdr->lma + last_hdr->_raw_size, maxpagesize)
+ == hdr->lma))
+ {
+ /* We don't want to put a writable section in a read only
+ segment, unless they are on the same page in memory
+ anyhow. We already know that the last section does not
+ bring us past the current section on the page, so the
+ only case in which the new section is not on the same
+ page as the previous section is when the previous section
+ ends precisely on a page boundary. */
+ new_segment = true;
+ }
+ else
+ {
+ /* Otherwise, we can use the same segment. */
+ new_segment = false;
+ }
+
+ if (! new_segment)
+ {
+ if ((hdr->flags & SEC_READONLY) == 0)
+ writable = true;
+ last_hdr = hdr;
+ continue;
+ }
- BFD_ASSERT ((abfd->flags & (EXEC_P | DYNAMIC)) != 0);
- BFD_ASSERT (phdr_size / sizeof (Elf_Internal_Phdr)
- <= sizeof phdrs / sizeof (phdrs[0]));
+ /* We need a new program segment. We must create a new program
+ header holding all the sections from phdr_index until hdr. */
- phdr_count = 0;
- phdr = phdrs;
+ m = make_mapping (abfd, sections, phdr_index, i, phdr_in_segment);
+ if (m == NULL)
+ goto error_return;
- if (bed->want_hdr_in_seg)
- phdr_size_adjust = first->sh_offset - phdr_size;
- else
- phdr_size_adjust = 0;
+ *pm = m;
+ pm = &m->next;
+
+ if ((hdr->flags & SEC_READONLY) == 0)
+ writable = true;
+ else
+ writable = false;
+
+ last_hdr = hdr;
+ phdr_index = i;
+ phdr_in_segment = false;
+ }
- /* If we have a loadable .interp section, we must create a PT_INTERP
- segment which must precede all PT_LOAD segments. We assume that
- we must also create a PT_PHDR segment, although that may not be
- true for all targets. */
- sinterp = bfd_get_section_by_name (abfd, ".interp");
- if (sinterp != NULL && (sinterp->flags & SEC_LOAD) != 0)
+ /* Create a final PT_LOAD program segment. */
+ if (last_hdr != NULL)
{
- BFD_ASSERT (first != NULL);
+ m = make_mapping (abfd, sections, phdr_index, i, phdr_in_segment);
+ if (m == NULL)
+ goto error_return;
+
+ *pm = m;
+ pm = &m->next;
+ }
- phdr->p_type = PT_PHDR;
+ /* If there is a .dynamic section, throw in a PT_DYNAMIC segment. */
+ if (dynsec != NULL)
+ {
+ m = ((struct elf_segment_map *)
+ bfd_zalloc (abfd, sizeof (struct elf_segment_map)));
+ if (m == NULL)
+ goto error_return;
+ m->next = NULL;
+ m->p_type = PT_DYNAMIC;
+ m->count = 1;
+ m->sections[0] = dynsec;
+
+ *pm = m;
+ pm = &m->next;
+ }
- phdr->p_offset = off;
+ /* For each loadable .note section, add a PT_NOTE segment. We don't
+ use bfd_get_section_by_name, because if we link together
+ nonloadable .note sections and loadable .note sections, we will
+ generate two .note sections in the output file. FIXME: Using
+ names for section types is bogus anyhow. */
+ for (s = abfd->sections; s != NULL; s = s->next)
+ {
+ if ((s->flags & SEC_LOAD) != 0
+ && strncmp (s->name, ".note", 5) == 0)
+ {
+ m = ((struct elf_segment_map *)
+ bfd_zalloc (abfd, sizeof (struct elf_segment_map)));
+ if (m == NULL)
+ goto error_return;
+ m->next = NULL;
+ m->p_type = PT_NOTE;
+ m->count = 1;
+ m->sections[0] = s;
+
+ *pm = m;
+ pm = &m->next;
+ }
+ }
- /* Account for any adjustment made because of the alignment of
- the first loadable section. */
- phdr_size_adjust = (first->sh_offset - phdr_size) - off;
- BFD_ASSERT (phdr_size_adjust >= 0 && phdr_size_adjust < 128);
+ free (sections);
+ sections = NULL;
- /* The program header precedes all loadable sections. This lets
- us compute its loadable address. This depends on the linker
- script. */
- phdr->p_vaddr = first->sh_addr - (phdr_size + phdr_size_adjust);
+ elf_tdata (abfd)->segment_map = mfirst;
+ return true;
- phdr->p_paddr = 0;
- phdr->p_filesz = phdr_size;
- phdr->p_memsz = phdr_size;
+ error_return:
+ if (sections != NULL)
+ free (sections);
+ return false;
+}
- /* FIXME: UnixWare and Solaris set PF_X, Irix 5 does not. */
- phdr->p_flags = PF_R | PF_X;
+/* Sort sections by address. */
- phdr->p_align = bed->s->file_align;
- BFD_ASSERT ((phdr->p_vaddr - phdr->p_offset) % bed->s->file_align == 0);
+static int
+elf_sort_sections (arg1, arg2)
+ const PTR arg1;
+ const PTR arg2;
+{
+ const asection *sec1 = *(const asection **) arg1;
+ const asection *sec2 = *(const asection **) arg2;
- /* Include the ELF header in the first loadable segment. */
- phdr_size_adjust += off;
+ /* Sort by LMA first, since this is the address used to
+ place the section into a segment. */
+ if (sec1->lma < sec2->lma)
+ return -1;
+ else if (sec1->lma > sec2->lma)
+ return 1;
- ++phdr_count;
- ++phdr;
+ /* Then sort by VMA. Normally the LMA and the VMA will be
+ the same, and this will do nothing. */
+ if (sec1->vma < sec2->vma)
+ return -1;
+ else if (sec1->vma > sec2->vma)
+ return 1;
- phdr->p_type = PT_INTERP;
- phdr->p_offset = sinterp->filepos;
- phdr->p_vaddr = sinterp->vma;
- phdr->p_paddr = 0;
- phdr->p_filesz = sinterp->_raw_size;
- phdr->p_memsz = sinterp->_raw_size;
- phdr->p_flags = PF_R;
- phdr->p_align = 1 << bfd_get_section_alignment (abfd, sinterp);
+ /* Put !SEC_LOAD sections after SEC_LOAD ones. */
- ++phdr_count;
- ++phdr;
- }
+#define TOEND(x) (((x)->flags & SEC_LOAD) == 0)
- /* Look through the sections to see how they will be divided into
- program segments. The sections must be arranged in order by
- sh_addr for this to work correctly. */
- phdr->p_type = PT_NULL;
- last_type = SHT_PROGBITS;
- for (i = 1, hdrpp = sorted_hdrs;
- i < elf_elfheader (abfd)->e_shnum;
- i++, hdrpp++)
+ if (TOEND (sec1))
{
- Elf_Internal_Shdr *hdr;
+ if (TOEND (sec2))
+ return sec1->target_index - sec2->target_index;
+ else
+ return 1;
+ }
- hdr = *hdrpp;
+ if (TOEND (sec2))
+ return -1;
- /* Ignore any section which will not be part of the process
- image. */
- if ((hdr->sh_flags & SHF_ALLOC) == 0)
- continue;
+#undef TOEND
- /* If this section fits in the segment we are constructing, add
- it in. */
- if (phdr->p_type != PT_NULL
- && (hdr->sh_offset - (phdr->p_offset + phdr->p_memsz)
- == hdr->sh_addr - (phdr->p_vaddr + phdr->p_memsz))
- && (last_type != SHT_NOBITS || hdr->sh_type == SHT_NOBITS))
- {
- bfd_size_type adjust;
-
- adjust = hdr->sh_addr - (phdr->p_vaddr + phdr->p_memsz);
- phdr->p_memsz += hdr->sh_size + adjust;
- if (hdr->sh_type != SHT_NOBITS)
- phdr->p_filesz += hdr->sh_size + adjust;
- if ((hdr->sh_flags & SHF_WRITE) != 0)
- phdr->p_flags |= PF_W;
- if ((hdr->sh_flags & SHF_EXECINSTR) != 0)
- phdr->p_flags |= PF_X;
- last_type = hdr->sh_type;
- continue;
- }
+ /* Sort by size, to put zero sized sections before others at the
+ same address. */
- /* The section won't fit, start a new segment. If we're already in one,
- move to the next one. */
- if (phdr->p_type != PT_NULL)
- {
- ++phdr;
- ++phdr_count;
- }
+ if (sec1->_raw_size < sec2->_raw_size)
+ return -1;
+ if (sec1->_raw_size > sec2->_raw_size)
+ return 1;
- /* Initialize the segment. */
- phdr->p_type = PT_LOAD;
- phdr->p_offset = hdr->sh_offset;
- phdr->p_vaddr = hdr->sh_addr;
- phdr->p_paddr = 0;
- if (hdr->sh_type == SHT_NOBITS)
- phdr->p_filesz = 0;
- else
- phdr->p_filesz = hdr->sh_size;
- phdr->p_memsz = hdr->sh_size;
- phdr->p_flags = PF_R;
- if ((hdr->sh_flags & SHF_WRITE) != 0)
- phdr->p_flags |= PF_W;
- if ((hdr->sh_flags & SHF_EXECINSTR) != 0)
- phdr->p_flags |= PF_X;
- phdr->p_align = bed->maxpagesize;
-
- if (hdr == first
- && (bed->want_hdr_in_seg
- || (sinterp != NULL
- && (sinterp->flags & SEC_LOAD) != 0)))
- {
- phdr->p_offset -= phdr_size + phdr_size_adjust;
- phdr->p_vaddr -= phdr_size + phdr_size_adjust;
- phdr->p_filesz += phdr_size + phdr_size_adjust;
- phdr->p_memsz += phdr_size + phdr_size_adjust;
- }
+ return sec1->target_index - sec2->target_index;
+}
- last_type = hdr->sh_type;
- }
+/* Assign file positions to the sections based on the mapping from
+ sections to segments. This function also sets up some fields in
+ the file header, and writes out the program headers. */
- if (phdr->p_type != PT_NULL)
+static boolean
+assign_file_positions_for_segments (abfd)
+ bfd *abfd;
+{
+ const struct elf_backend_data *bed = get_elf_backend_data (abfd);
+ unsigned int count;
+ struct elf_segment_map *m;
+ unsigned int alloc;
+ Elf_Internal_Phdr *phdrs;
+ file_ptr off, voff;
+ bfd_vma filehdr_vaddr, filehdr_paddr;
+ bfd_vma phdrs_vaddr, phdrs_paddr;
+ Elf_Internal_Phdr *p;
+
+ if (elf_tdata (abfd)->segment_map == NULL)
{
- ++phdr;
- ++phdr_count;
+ if (! map_sections_to_segments (abfd))
+ return false;
}
- /* If we have a .dynamic section, create a PT_DYNAMIC segment. */
- sdyn = bfd_get_section_by_name (abfd, ".dynamic");
- if (sdyn != NULL && (sdyn->flags & SEC_LOAD) != 0)
+ if (bed->elf_backend_modify_segment_map)
{
- phdr->p_type = PT_DYNAMIC;
- phdr->p_offset = sdyn->filepos;
- phdr->p_vaddr = sdyn->vma;
- phdr->p_paddr = 0;
- phdr->p_filesz = sdyn->_raw_size;
- phdr->p_memsz = sdyn->_raw_size;
- phdr->p_flags = PF_R;
- if ((sdyn->flags & SEC_READONLY) == 0)
- phdr->p_flags |= PF_W;
- if ((sdyn->flags & SEC_CODE) != 0)
- phdr->p_flags |= PF_X;
- phdr->p_align = 1 << bfd_get_section_alignment (abfd, sdyn);
-
- ++phdr;
- ++phdr_count;
+ if (! (*bed->elf_backend_modify_segment_map) (abfd))
+ return false;
}
- /* Let the backend create additional program headers. */
- if (bed->elf_backend_create_program_headers)
- phdr_count = (*bed->elf_backend_create_program_headers) (abfd,
- phdrs,
- phdr_count);
+ count = 0;
+ for (m = elf_tdata (abfd)->segment_map; m != NULL; m = m->next)
+ ++count;
+
+ elf_elfheader (abfd)->e_phoff = bed->s->sizeof_ehdr;
+ elf_elfheader (abfd)->e_phentsize = bed->s->sizeof_phdr;
+ elf_elfheader (abfd)->e_phnum = count;
+
+ if (count == 0)
+ return true;
- /* Make sure the return value from get_program_header_size matches
- what we computed here. Actually, it's OK if we allocated too
- much space in the program header. */
- if (phdr_count > phdr_size / bed->s->sizeof_phdr)
+ /* If we already counted the number of program segments, make sure
+ that we allocated enough space. This happens when SIZEOF_HEADERS
+ is used in a linker script. */
+ alloc = elf_tdata (abfd)->program_header_size / bed->s->sizeof_phdr;
+ if (alloc != 0 && count > alloc)
{
((*_bfd_error_handler)
- ("%s: Not enough room for program headers (allocated %lu, need %u)",
- bfd_get_filename (abfd),
- (unsigned long) (phdr_size / bed->s->sizeof_phdr),
- phdr_count));
+ (_("%s: Not enough room for program headers (allocated %u, need %u)"),
+ bfd_get_filename (abfd), alloc, count));
bfd_set_error (bfd_error_bad_value);
- return (file_ptr) -1;
+ return false;
}
- /* Set up program header information. */
- i_ehdrp = elf_elfheader (abfd);
- i_ehdrp->e_phentsize = bed->s->sizeof_phdr;
- i_ehdrp->e_phoff = off;
- i_ehdrp->e_phnum = phdr_count;
+ if (alloc == 0)
+ alloc = count;
- /* Save the program headers away. I don't think anybody uses this
- information right now. */
- elf_tdata (abfd)->phdr = ((Elf_Internal_Phdr *)
- bfd_alloc (abfd,
- (phdr_count
- * sizeof (Elf_Internal_Phdr))));
- if (elf_tdata (abfd)->phdr == NULL && phdr_count != 0)
- {
- bfd_set_error (bfd_error_no_memory);
- return (file_ptr) -1;
- }
- memcpy (elf_tdata (abfd)->phdr, phdrs,
- phdr_count * sizeof (Elf_Internal_Phdr));
+ phdrs = ((Elf_Internal_Phdr *)
+ bfd_alloc (abfd, alloc * sizeof (Elf_Internal_Phdr)));
+ if (phdrs == NULL)
+ return false;
- /* Write out the program headers. */
- if (bfd_seek (abfd, off, SEEK_SET) != 0)
- return (file_ptr) -1;
+ off = bed->s->sizeof_ehdr;
+ off += alloc * bed->s->sizeof_phdr;
- if (bed->s->write_out_phdrs (abfd, phdrs, phdr_count) != 0)
- return (file_ptr) -1;
+ filehdr_vaddr = 0;
+ filehdr_paddr = 0;
+ phdrs_vaddr = 0;
+ phdrs_paddr = 0;
- return off + phdr_count * bed->s->sizeof_phdr;
-}
+ for (m = elf_tdata (abfd)->segment_map, p = phdrs;
+ m != NULL;
+ m = m->next, p++)
+ {
+ unsigned int i;
+ asection **secpp;
-/* Work out the file positions of all the sections. This is called by
- _bfd_elf_compute_section_file_positions. All the section sizes and
- VMAs must be known before this is called.
+ /* If elf_segment_map is not from map_sections_to_segments, the
+ sections may not be correctly ordered. */
+ if (m->count > 0)
+ qsort (m->sections, (size_t) m->count, sizeof (asection *),
+ elf_sort_sections);
+
+ p->p_type = m->p_type;
+
+ if (m->p_flags_valid)
+ p->p_flags |= m->p_flags;
+
+ if (p->p_type == PT_LOAD
+ && m->count > 0
+ && (m->sections[0]->flags & SEC_ALLOC) != 0)
+ {
+ if ((abfd->flags & D_PAGED) != 0)
+ off += (m->sections[0]->vma - off) % bed->maxpagesize;
+ else
+ {
+ bfd_size_type align;
+
+ align = 0;
+ for (i = 0, secpp = m->sections; i < m->count; i++, secpp++)
+ {
+ bfd_size_type secalign;
+
+ secalign = bfd_get_section_alignment (abfd, *secpp);
+ if (secalign > align)
+ align = secalign;
+ }
+
+ off += (m->sections[0]->vma - off) % (1 << align);
+ }
+ }
+
+ if (m->count == 0)
+ p->p_vaddr = 0;
+ else
+ p->p_vaddr = m->sections[0]->vma;
+
+ if (m->p_paddr_valid)
+ p->p_paddr = m->p_paddr;
+ else if (m->count == 0)
+ p->p_paddr = 0;
+ else
+ p->p_paddr = m->sections[0]->lma;
+
+ if (p->p_type == PT_LOAD
+ && (abfd->flags & D_PAGED) != 0)
+ p->p_align = bed->maxpagesize;
+ else if (m->count == 0)
+ p->p_align = bed->s->file_align;
+ else
+ p->p_align = 0;
+
+ p->p_offset = 0;
+ p->p_filesz = 0;
+ p->p_memsz = 0;
+
+ if (m->includes_filehdr)
+ {
+ if (! m->p_flags_valid)
+ p->p_flags |= PF_R;
+ p->p_offset = 0;
+ p->p_filesz = bed->s->sizeof_ehdr;
+ p->p_memsz = bed->s->sizeof_ehdr;
+ if (m->count > 0)
+ {
+ BFD_ASSERT (p->p_type == PT_LOAD);
+
+ if (p->p_vaddr < (bfd_vma) off)
+ {
+ _bfd_error_handler (_("%s: Not enough room for program headers, try linking with -N"),
+ bfd_get_filename (abfd));
+ bfd_set_error (bfd_error_bad_value);
+ return false;
+ }
+
+ p->p_vaddr -= off;
+ if (! m->p_paddr_valid)
+ p->p_paddr -= off;
+ }
+ if (p->p_type == PT_LOAD)
+ {
+ filehdr_vaddr = p->p_vaddr;
+ filehdr_paddr = p->p_paddr;
+ }
+ }
+
+ if (m->includes_phdrs)
+ {
+ if (! m->p_flags_valid)
+ p->p_flags |= PF_R;
+
+ if (m->includes_filehdr)
+ {
+ if (p->p_type == PT_LOAD)
+ {
+ phdrs_vaddr = p->p_vaddr + bed->s->sizeof_ehdr;
+ phdrs_paddr = p->p_paddr + bed->s->sizeof_ehdr;
+ }
+ }
+ else
+ {
+ p->p_offset = bed->s->sizeof_ehdr;
+
+ if (m->count > 0)
+ {
+ BFD_ASSERT (p->p_type == PT_LOAD);
+ p->p_vaddr -= off - p->p_offset;
+ if (! m->p_paddr_valid)
+ p->p_paddr -= off - p->p_offset;
+ }
+
+ if (p->p_type == PT_LOAD)
+ {
+ phdrs_vaddr = p->p_vaddr;
+ phdrs_paddr = p->p_paddr;
+ }
+ else
+ phdrs_vaddr = bed->maxpagesize + bed->s->sizeof_ehdr;
+ }
+
+ p->p_filesz += alloc * bed->s->sizeof_phdr;
+ p->p_memsz += alloc * bed->s->sizeof_phdr;
+ }
+
+ if (p->p_type == PT_LOAD
+ || (p->p_type == PT_NOTE && bfd_get_format (abfd) == bfd_core))
+ {
+ if (! m->includes_filehdr && ! m->includes_phdrs)
+ p->p_offset = off;
+ else
+ {
+ file_ptr adjust;
+
+ adjust = off - (p->p_offset + p->p_filesz);
+ p->p_filesz += adjust;
+ p->p_memsz += adjust;
+ }
+ }
+
+ voff = off;
+
+ for (i = 0, secpp = m->sections; i < m->count; i++, secpp++)
+ {
+ asection *sec;
+ flagword flags;
+ bfd_size_type align;
+
+ sec = *secpp;
+ flags = sec->flags;
+ align = 1 << bfd_get_section_alignment (abfd, sec);
+
+ /* The section may have artificial alignment forced by a
+ link script. Notice this case by the gap between the
+ cumulative phdr vma and the section's vma. */
+ if (p->p_vaddr + p->p_memsz < sec->vma)
+ {
+ bfd_vma adjust = sec->vma - (p->p_vaddr + p->p_memsz);
+
+ p->p_memsz += adjust;
+ off += adjust;
+ voff += adjust;
+ if ((flags & SEC_LOAD) != 0)
+ p->p_filesz += adjust;
+ }
+
+ if (p->p_type == PT_LOAD)
+ {
+ bfd_signed_vma adjust;
+
+ if ((flags & SEC_LOAD) != 0)
+ {
+ adjust = sec->lma - (p->p_paddr + p->p_memsz);
+ if (adjust < 0)
+ adjust = 0;
+ }
+ else if ((flags & SEC_ALLOC) != 0)
+ {
+ /* The section VMA must equal the file position
+ modulo the page size. FIXME: I'm not sure if
+ this adjustment is really necessary. We used to
+ not have the SEC_LOAD case just above, and then
+ this was necessary, but now I'm not sure. */
+ if ((abfd->flags & D_PAGED) != 0)
+ adjust = (sec->vma - voff) % bed->maxpagesize;
+ else
+ adjust = (sec->vma - voff) % align;
+ }
+ else
+ adjust = 0;
+
+ if (adjust != 0)
+ {
+ if (i == 0)
+ {
+ (* _bfd_error_handler)
+ (_("Error: First section in segment (%s) starts at 0x%x"),
+ bfd_section_name (abfd, sec), sec->lma);
+ (* _bfd_error_handler)
+ (_(" whereas segment starts at 0x%x"),
+ p->p_paddr);
+
+ return false;
+ }
+ p->p_memsz += adjust;
+ off += adjust;
+ voff += adjust;
+ if ((flags & SEC_LOAD) != 0)
+ p->p_filesz += adjust;
+ }
+
+ sec->filepos = off;
+
+ /* We check SEC_HAS_CONTENTS here because if NOLOAD is
+ used in a linker script we may have a section with
+ SEC_LOAD clear but which is supposed to have
+ contents. */
+ if ((flags & SEC_LOAD) != 0
+ || (flags & SEC_HAS_CONTENTS) != 0)
+ off += sec->_raw_size;
+
+ if ((flags & SEC_ALLOC) != 0)
+ voff += sec->_raw_size;
+ }
+
+ if (p->p_type == PT_NOTE && bfd_get_format (abfd) == bfd_core)
+ {
+ if (i == 0) /* the actual "note" segment */
+ { /* this one actually contains everything. */
+ sec->filepos = off;
+ p->p_filesz = sec->_raw_size;
+ off += sec->_raw_size;
+ voff = off;
+ }
+ else /* fake sections -- don't need to be written */
+ {
+ sec->filepos = 0;
+ sec->_raw_size = 0;
+ flags = sec->flags = 0; /* no contents */
+ }
+ p->p_memsz = 0;
+ p->p_align = 1;
+ }
+ else
+ {
+ p->p_memsz += sec->_raw_size;
+
+ if ((flags & SEC_LOAD) != 0)
+ p->p_filesz += sec->_raw_size;
+
+ if (align > p->p_align
+ && (p->p_type != PT_LOAD || (abfd->flags & D_PAGED) == 0))
+ p->p_align = align;
+ }
+
+ if (! m->p_flags_valid)
+ {
+ p->p_flags |= PF_R;
+ if ((flags & SEC_CODE) != 0)
+ p->p_flags |= PF_X;
+ if ((flags & SEC_READONLY) == 0)
+ p->p_flags |= PF_W;
+ }
+ }
+ }
+
+ /* Now that we have set the section file positions, we can set up
+ the file positions for the non PT_LOAD segments. */
+ for (m = elf_tdata (abfd)->segment_map, p = phdrs;
+ m != NULL;
+ m = m->next, p++)
+ {
+ if (p->p_type != PT_LOAD && m->count > 0)
+ {
+ BFD_ASSERT (! m->includes_filehdr && ! m->includes_phdrs);
+ p->p_offset = m->sections[0]->filepos;
+ }
+ if (m->count == 0)
+ {
+ if (m->includes_filehdr)
+ {
+ p->p_vaddr = filehdr_vaddr;
+ if (! m->p_paddr_valid)
+ p->p_paddr = filehdr_paddr;
+ }
+ else if (m->includes_phdrs)
+ {
+ p->p_vaddr = phdrs_vaddr;
+ if (! m->p_paddr_valid)
+ p->p_paddr = phdrs_paddr;
+ }
+ }
+ }
+
+ /* Clear out any program headers we allocated but did not use. */
+ for (; count < alloc; count++, p++)
+ {
+ memset (p, 0, sizeof *p);
+ p->p_type = PT_NULL;
+ }
+
+ elf_tdata (abfd)->phdr = phdrs;
+
+ elf_tdata (abfd)->next_file_pos = off;
+
+ /* Write out the program headers. */
+ if (bfd_seek (abfd, bed->s->sizeof_ehdr, SEEK_SET) != 0
+ || bed->s->write_out_phdrs (abfd, phdrs, alloc) != 0)
+ return false;
+
+ return true;
+}
+
+/* Get the size of the program header.
+
+ If this is called by the linker before any of the section VMA's are set, it
+ can't calculate the correct value for a strange memory layout. This only
+ happens when SIZEOF_HEADERS is used in a linker script. In this case,
+ SORTED_HDRS is NULL and we assume the normal scenario of one text and one
+ data segment (exclusive of .interp and .dynamic).
+
+ ??? User written scripts must either not use SIZEOF_HEADERS, or assume there
+ will be two segments. */
+
+static bfd_size_type
+get_program_header_size (abfd)
+ bfd *abfd;
+{
+ size_t segs;
+ asection *s;
+ struct elf_backend_data *bed = get_elf_backend_data (abfd);
+
+ /* We can't return a different result each time we're called. */
+ if (elf_tdata (abfd)->program_header_size != 0)
+ return elf_tdata (abfd)->program_header_size;
+
+ if (elf_tdata (abfd)->segment_map != NULL)
+ {
+ struct elf_segment_map *m;
+
+ segs = 0;
+ for (m = elf_tdata (abfd)->segment_map; m != NULL; m = m->next)
+ ++segs;
+ elf_tdata (abfd)->program_header_size = segs * bed->s->sizeof_phdr;
+ return elf_tdata (abfd)->program_header_size;
+ }
+
+ /* Assume we will need exactly two PT_LOAD segments: one for text
+ and one for data. */
+ segs = 2;
+
+ s = bfd_get_section_by_name (abfd, ".interp");
+ if (s != NULL && (s->flags & SEC_LOAD) != 0)
+ {
+ /* If we have a loadable interpreter section, we need a
+ PT_INTERP segment. In this case, assume we also need a
+ PT_PHDR segment, although that may not be true for all
+ targets. */
+ segs += 2;
+ }
+
+ if (bfd_get_section_by_name (abfd, ".dynamic") != NULL)
+ {
+ /* We need a PT_DYNAMIC segment. */
+ ++segs;
+ }
+
+ for (s = abfd->sections; s != NULL; s = s->next)
+ {
+ if ((s->flags & SEC_LOAD) != 0
+ && strncmp (s->name, ".note", 5) == 0)
+ {
+ /* We need a PT_NOTE segment. */
+ ++segs;
+ }
+ }
+
+ /* Let the backend count up any program headers it might need. */
+ if (bed->elf_backend_additional_program_headers)
+ {
+ int a;
+
+ a = (*bed->elf_backend_additional_program_headers) (abfd);
+ if (a == -1)
+ abort ();
+ segs += a;
+ }
+
+ elf_tdata (abfd)->program_header_size = segs * bed->s->sizeof_phdr;
+ return elf_tdata (abfd)->program_header_size;
+}
+
+/* Work out the file positions of all the sections. This is called by
+ _bfd_elf_compute_section_file_positions. All the section sizes and
+ VMAs must be known before this is called.
We do not consider reloc sections at this point, unless they form
part of the loadable image. Reloc sections are assigned file
positions in assign_file_positions_for_relocs, which is called by
write_object_contents and final_link.
- If DOSYMS is false, we do not assign file positions for the symbol
- table or the string table. */
-
-static int elf_sort_hdrs PARAMS ((const PTR, const PTR));
+ We also don't set the positions of the .symtab and .strtab here. */
static boolean
-assign_file_positions_except_relocs (abfd, dosyms)
+assign_file_positions_except_relocs (abfd)
bfd *abfd;
- boolean dosyms;
{
struct elf_obj_tdata * const tdata = elf_tdata (abfd);
Elf_Internal_Ehdr * const i_ehdrp = elf_elfheader (abfd);
file_ptr off;
struct elf_backend_data *bed = get_elf_backend_data (abfd);
- /* Start after the ELF header. */
- off = i_ehdrp->e_ehsize;
-
- if ((abfd->flags & (EXEC_P | DYNAMIC)) == 0)
+ if ((abfd->flags & (EXEC_P | DYNAMIC)) == 0
+ && bfd_get_format (abfd) != bfd_core)
{
Elf_Internal_Shdr **hdrpp;
unsigned int i;
+ /* Start after the ELF header. */
+ off = i_ehdrp->e_ehsize;
+
/* We are not creating an executable, which means that we are
not creating a program header, and that the actual order of
the sections in the file is unimportant. */
hdr->sh_offset = -1;
continue;
}
- if (! dosyms
- && (i == tdata->symtab_section
- || i == tdata->strtab_section))
+ if (i == tdata->symtab_section
+ || i == tdata->strtab_section)
{
hdr->sh_offset = -1;
continue;
}
-
- off = _bfd_elf_assign_file_position_for_section (hdr, off);
+
+ off = _bfd_elf_assign_file_position_for_section (hdr, off, true);
}
}
else
{
- file_ptr phdr_off;
- bfd_size_type phdr_size;
- bfd_vma maxpagesize;
- size_t hdrppsize;
- Elf_Internal_Shdr **sorted_hdrs;
- Elf_Internal_Shdr **hdrpp;
unsigned int i;
- Elf_Internal_Shdr *first;
- file_ptr phdr_map;
-
- /* We are creating an executable. */
-
- maxpagesize = get_elf_backend_data (abfd)->maxpagesize;
- if (maxpagesize == 0)
- maxpagesize = 1;
-
- /* We must sort the sections. The GNU linker will always create
- the sections in an appropriate order, but the Irix 5 linker
- will not. We don't include the dummy first section in the
- sort. We sort sections which are not SHF_ALLOC to the end. */
- hdrppsize = (i_ehdrp->e_shnum - 1) * sizeof (Elf_Internal_Shdr *);
- sorted_hdrs = (Elf_Internal_Shdr **) malloc (hdrppsize);
- if (sorted_hdrs == NULL)
- {
- bfd_set_error (bfd_error_no_memory);
- return false;
- }
+ Elf_Internal_Shdr **hdrpp;
- memcpy (sorted_hdrs, i_shdrpp + 1, hdrppsize);
- qsort (sorted_hdrs, (size_t) i_ehdrp->e_shnum - 1,
- sizeof (Elf_Internal_Shdr *), elf_sort_hdrs);
-
- /* We can't actually create the program header until we have set the
- file positions for the sections, and we can't do that until we know
- how big the header is going to be. */
- off = align_file_position (off, bed->s->file_align);
- phdr_size = get_program_header_size (abfd,
- sorted_hdrs, i_ehdrp->e_shnum - 1,
- maxpagesize);
- if (phdr_size == (bfd_size_type) -1)
+ /* Assign file positions for the loaded sections based on the
+ assignment of sections to segments. */
+ if (! assign_file_positions_for_segments (abfd))
return false;
- /* Compute the file offsets of each section. */
- phdr_off = off;
- off += phdr_size;
- first = NULL;
- for (i = 1, hdrpp = sorted_hdrs; i < i_ehdrp->e_shnum; i++, hdrpp++)
+ /* Assign file positions for the other sections. */
+
+ off = elf_tdata (abfd)->next_file_pos;
+ for (i = 1, hdrpp = i_shdrpp + 1; i < i_ehdrp->e_shnum; i++, hdrpp++)
{
Elf_Internal_Shdr *hdr;
hdr = *hdrpp;
- if ((hdr->sh_flags & SHF_ALLOC) == 0)
+ if (hdr->bfd_section != NULL
+ && hdr->bfd_section->filepos != 0)
+ hdr->sh_offset = hdr->bfd_section->filepos;
+ else if ((hdr->sh_flags & SHF_ALLOC) != 0)
{
- if (hdr->sh_type == SHT_REL || hdr->sh_type == SHT_RELA)
- {
- hdr->sh_offset = -1;
- continue;
- }
- if (! dosyms
- && (hdr == i_shdrpp[tdata->symtab_section]
- || hdr == i_shdrpp[tdata->strtab_section]))
- {
- hdr->sh_offset = -1;
- continue;
- }
- off = _bfd_elf_assign_file_position_for_section (hdr, off);
+ ((*_bfd_error_handler)
+ (_("%s: warning: allocated section `%s' not in segment"),
+ bfd_get_filename (abfd),
+ (hdr->bfd_section == NULL
+ ? "*unknown*"
+ : hdr->bfd_section->name)));
+ if ((abfd->flags & D_PAGED) != 0)
+ off += (hdr->sh_addr - off) % bed->maxpagesize;
+ else
+ off += (hdr->sh_addr - off) % hdr->sh_addralign;
+ off = _bfd_elf_assign_file_position_for_section (hdr, off,
+ false);
}
+ else if (hdr->sh_type == SHT_REL
+ || hdr->sh_type == SHT_RELA
+ || hdr == i_shdrpp[tdata->symtab_section]
+ || hdr == i_shdrpp[tdata->strtab_section])
+ hdr->sh_offset = -1;
else
- {
- if (first == NULL)
- first = hdr;
-
- /* The section VMA must equal the file position modulo
- the page size. This is required by the program
- header. */
- off += (hdr->sh_addr - off) % maxpagesize;
- off = _bfd_elf_assign_file_position_for_section (hdr, off);
- }
+ off = _bfd_elf_assign_file_position_for_section (hdr, off, true);
}
-
- /* Create the program header. */
- phdr_map = map_program_segments (abfd, phdr_off, first, sorted_hdrs,
- phdr_size);
- if (phdr_map == (file_ptr) -1)
- return false;
- BFD_ASSERT ((bfd_size_type) phdr_map
- <= (bfd_size_type) phdr_off + phdr_size);
-
- free (sorted_hdrs);
}
/* Place the section headers. */
return true;
}
-/* Sort the ELF headers by VMA. We sort headers which are not
- SHF_ALLOC to the end. */
-static int
-elf_sort_hdrs (arg1, arg2)
- const PTR arg1;
- const PTR arg2;
-{
- int ret;
- const Elf_Internal_Shdr *hdr1 = *(const Elf_Internal_Shdr **) arg1;
- const Elf_Internal_Shdr *hdr2 = *(const Elf_Internal_Shdr **) arg2;
-
-#define TOEND(x) (((x)->sh_flags & SHF_ALLOC)==0)
-
- if (TOEND (hdr1))
- if (TOEND (hdr2))
- return 0;
- else
- return 1;
-
- if (TOEND (hdr2))
- return -1;
-
- if (hdr1->sh_addr < hdr2->sh_addr)
- return -1;
- else if (hdr1->sh_addr > hdr2->sh_addr)
- return 1;
-
- /* Put !SHT_NOBITS sections before SHT_NOBITS ones.
- The main loop in map_program_segments requires this. */
-
- ret = (hdr1->sh_type == SHT_NOBITS) - (hdr2->sh_type == SHT_NOBITS);
-
- if (ret != 0)
- return ret;
- if (hdr1->sh_size < hdr2->sh_size)
- return -1;
- if (hdr1->sh_size > hdr2->sh_size)
- return 1;
- return 0;
-}
-
static boolean
prep_headers (abfd)
bfd *abfd;
{
Elf_Internal_Ehdr *i_ehdrp; /* Elf file header, internal form */
- Elf_Internal_Phdr *i_phdrp = 0; /* Program header table, internal form */
+ Elf_Internal_Phdr *i_phdrp = 0; /* Program header table, internal form */
Elf_Internal_Shdr **i_shdrp; /* Section header table, internal form */
int count;
struct bfd_strtab_hash *shstrtab;
i_ehdrp->e_ident[EI_CLASS] = bed->s->elfclass;
i_ehdrp->e_ident[EI_DATA] =
- abfd->xvec->byteorder_big_p ? ELFDATA2MSB : ELFDATA2LSB;
+ bfd_big_endian (abfd) ? ELFDATA2MSB : ELFDATA2LSB;
i_ehdrp->e_ident[EI_VERSION] = bed->s->ev_current;
+ i_ehdrp->e_ident[EI_OSABI] = ELFOSABI_SYSV;
+ i_ehdrp->e_ident[EI_ABIVERSION] = 0;
+
for (count = EI_PAD; count < EI_NIDENT; count++)
i_ehdrp->e_ident[count] = 0;
i_ehdrp->e_type = ET_DYN;
else if ((abfd->flags & EXEC_P) != 0)
i_ehdrp->e_type = ET_EXEC;
+ else if (bfd_get_format (abfd) == bfd_core)
+ i_ehdrp->e_type = ET_CORE;
else
i_ehdrp->e_type = ET_REL;
break;
case bfd_arch_sparc:
if (bed->s->arch_size == 64)
- i_ehdrp->e_machine = EM_SPARC64;
+ i_ehdrp->e_machine = EM_SPARCV9;
else
i_ehdrp->e_machine = EM_SPARC;
break;
case bfd_arch_i860:
i_ehdrp->e_machine = EM_860;
break;
+ case bfd_arch_i960:
+ i_ehdrp->e_machine = EM_960;
+ break;
case bfd_arch_mips: /* MIPS Rxxxx */
i_ehdrp->e_machine = EM_MIPS; /* only MIPS R3000 */
break;
case bfd_arch_powerpc:
i_ehdrp->e_machine = EM_PPC;
break;
-/* start-sanitize-arc */
- case bfd_arch_arc:
+ case bfd_arch_alpha:
+ i_ehdrp->e_machine = EM_ALPHA;
+ break;
+ case bfd_arch_sh:
+ i_ehdrp->e_machine = EM_SH;
+ break;
+ case bfd_arch_d10v:
+ i_ehdrp->e_machine = EM_CYGNUS_D10V;
+ break;
+ case bfd_arch_d30v:
+ i_ehdrp->e_machine = EM_CYGNUS_D30V;
+ break;
+ case bfd_arch_fr30:
+ i_ehdrp->e_machine = EM_CYGNUS_FR30;
+ break;
+ case bfd_arch_mcore:
+ i_ehdrp->e_machine = EM_MCORE;
+ break;
+ case bfd_arch_v850:
+ switch (bfd_get_mach (abfd))
+ {
+ default:
+ case 0: i_ehdrp->e_machine = EM_CYGNUS_V850; break;
+ }
+ break;
+ case bfd_arch_arc:
i_ehdrp->e_machine = EM_CYGNUS_ARC;
break;
-/* end-sanitize-arc */
+ case bfd_arch_arm:
+ i_ehdrp->e_machine = EM_ARM;
+ break;
+ case bfd_arch_m32r:
+ i_ehdrp->e_machine = EM_CYGNUS_M32R;
+ break;
+ case bfd_arch_mn10200:
+ i_ehdrp->e_machine = EM_CYGNUS_MN10200;
+ break;
+ case bfd_arch_mn10300:
+ i_ehdrp->e_machine = EM_CYGNUS_MN10300;
+ break;
/* also note that EM_M32, AT&T WE32100 is unknown to bfd */
default:
i_ehdrp->e_machine = EM_NONE;
shdrp = *shdrpp;
if ((shdrp->sh_type == SHT_REL || shdrp->sh_type == SHT_RELA)
&& shdrp->sh_offset == -1)
- off = _bfd_elf_assign_file_position_for_section (shdrp, off);
+ off = _bfd_elf_assign_file_position_for_section (shdrp, off, true);
}
elf_tdata (abfd)->next_file_pos = off;
unsigned int count;
if (! abfd->output_has_begun
- && ! _bfd_elf_compute_section_file_positions (abfd,
- (struct bfd_link_info *) NULL))
+ && ! _bfd_elf_compute_section_file_positions
+ (abfd, (struct bfd_link_info *) NULL))
return false;
i_shdrp = elf_elfsections (abfd);
bfd_map_over_sections (abfd, bed->s->write_relocs, &failed);
if (failed)
return false;
+
_bfd_elf_assign_file_positions_for_relocs (abfd);
/* After writing the headers, we need to write the sections too... */
return bed->s->write_shdrs_and_ehdr (abfd);
}
+boolean
+_bfd_elf_write_corefile_contents (abfd)
+ bfd *abfd;
+{
+ /* Hopefully this can be done just like an object file. */
+ return _bfd_elf_write_object_contents (abfd);
+}
/* given a section, search the header to find them... */
int
_bfd_elf_section_from_bfd_section (abfd, asect)
if (bfd_is_und_section (asect))
return SHN_UNDEF;
+ bfd_set_error (bfd_error_nonrepresentable_section);
+
return -1;
}
-/* given a symbol, return the bfd index for that symbol. */
- int
+/* Given a BFD symbol, return the index in the ELF symbol table, or -1
+ on error. */
+
+int
_bfd_elf_symbol_from_bfd_symbol (abfd, asym_ptr_ptr)
bfd *abfd;
- struct symbol_cache_entry **asym_ptr_ptr;
+ asymbol **asym_ptr_ptr;
{
- struct symbol_cache_entry *asym_ptr = *asym_ptr_ptr;
+ asymbol *asym_ptr = *asym_ptr_ptr;
int idx;
flagword flags = asym_ptr->flags;
}
idx = asym_ptr->udata.i;
- BFD_ASSERT (idx != 0);
+
+ if (idx == 0)
+ {
+ /* This case can occur when using --strip-symbol on a symbol
+ which is used in a relocation entry. */
+ (*_bfd_error_handler)
+ (_("%s: symbol `%s' required but not present"),
+ bfd_get_filename (abfd), bfd_asymbol_name (asym_ptr));
+ bfd_set_error (bfd_error_no_symbols);
+ return -1;
+ }
#if DEBUG & 4
{
fprintf (stderr,
- "elf_symbol_from_bfd_symbol 0x%.8lx, name = %s, sym num = %d, flags = 0x%.8lx%s\n",
- (long) asym_ptr, asym_ptr->name, idx, flags, elf_symbol_flags (flags));
+ _("elf_symbol_from_bfd_symbol 0x%.8lx, name = %s, sym num = %d, flags = 0x%.8lx%s\n"),
+ (long) asym_ptr, asym_ptr->name, idx, flags,
+ elf_symbol_flags (flags));
fflush (stderr);
}
#endif
return idx;
}
+/* Copy private BFD data. This copies any program header information. */
+
static boolean
-swap_out_syms (abfd, sttp)
- bfd *abfd;
- struct bfd_strtab_hash **sttp;
+copy_private_bfd_data (ibfd, obfd)
+ bfd *ibfd;
+ bfd *obfd;
{
- struct elf_backend_data *bed = get_elf_backend_data (abfd);
+ Elf_Internal_Ehdr *iehdr;
+ struct elf_segment_map *mfirst;
+ struct elf_segment_map **pm;
+ struct elf_segment_map *m;
+ Elf_Internal_Phdr *p;
+ unsigned int i;
+ unsigned int num_segments;
+ boolean phdr_included = false;
- if (!elf_map_symbols (abfd))
+ if (bfd_get_flavour (ibfd) != bfd_target_elf_flavour
+ || bfd_get_flavour (obfd) != bfd_target_elf_flavour)
+ return true;
+
+ if (elf_tdata (ibfd)->phdr == NULL)
+ return true;
+
+ iehdr = elf_elfheader (ibfd);
+
+ mfirst = NULL;
+ pm = &mfirst;
+
+ num_segments = elf_elfheader (ibfd)->e_phnum;
+
+#define IS_CONTAINED_BY(addr, len, bottom, phdr) \
+ ((addr) >= (bottom) \
+ && ( ((addr) + (len)) <= ((bottom) + (phdr)->p_memsz) \
+ || ((addr) + (len)) <= ((bottom) + (phdr)->p_filesz)))
+
+ /* Special case: corefile "NOTE" section containing regs, prpsinfo etc. */
+
+#define IS_COREFILE_NOTE(p, s) \
+ (p->p_type == PT_NOTE \
+ && bfd_get_format (ibfd) == bfd_core \
+ && s->vma == 0 && s->lma == 0 \
+ && (bfd_vma) s->filepos >= p->p_offset \
+ && (bfd_vma) s->filepos + s->_raw_size \
+ <= p->p_offset + p->p_filesz)
+
+ /* The complicated case when p_vaddr is 0 is to handle the Solaris
+ linker, which generates a PT_INTERP section with p_vaddr and
+ p_memsz set to 0. */
+
+#define IS_SOLARIS_PT_INTERP(p, s) \
+ (p->p_vaddr == 0 \
+ && p->p_filesz > 0 \
+ && (s->flags & SEC_HAS_CONTENTS) != 0 \
+ && s->_raw_size > 0 \
+ && (bfd_vma) s->filepos >= p->p_offset \
+ && ((bfd_vma) s->filepos + s->_raw_size \
+ <= p->p_offset + p->p_filesz))
+
+ /* Scan through the segments specified in the program header
+ of the input BFD. */
+ for (i = 0, p = elf_tdata (ibfd)->phdr; i < num_segments; i++, p++)
+ {
+ unsigned int csecs;
+ asection *s;
+ asection **sections;
+ asection *os;
+ unsigned int isec;
+ bfd_vma matching_lma;
+ bfd_vma suggested_lma;
+ unsigned int j;
+
+ /* For each section in the input BFD, decide if it should be
+ included in the current segment. A section will be included
+ if it is within the address space of the segment, and it is
+ an allocated segment, and there is an output section
+ associated with it. */
+ csecs = 0;
+ for (s = ibfd->sections; s != NULL; s = s->next)
+ if (s->output_section != NULL)
+ {
+ if ((IS_CONTAINED_BY (s->vma, s->_raw_size, p->p_vaddr, p)
+ || IS_SOLARIS_PT_INTERP (p, s))
+ && (s->flags & SEC_ALLOC) != 0)
+ ++csecs;
+ else if (IS_COREFILE_NOTE (p, s))
+ ++csecs;
+ }
+
+ /* Allocate a segment map big enough to contain all of the
+ sections we have selected. */
+ m = ((struct elf_segment_map *)
+ bfd_alloc (obfd,
+ (sizeof (struct elf_segment_map)
+ + ((size_t) csecs - 1) * sizeof (asection *))));
+ if (m == NULL)
+ return false;
+
+ /* Initialise the fields of the segment map. Default to
+ using the physical address of the segment in the input BFD. */
+ m->next = NULL;
+ m->p_type = p->p_type;
+ m->p_flags = p->p_flags;
+ m->p_flags_valid = 1;
+ m->p_paddr = p->p_paddr;
+ m->p_paddr_valid = 1;
+
+ /* Determine if this segment contains the ELF file header
+ and if it contains the program headers themselves. */
+ m->includes_filehdr = (p->p_offset == 0
+ && p->p_filesz >= iehdr->e_ehsize);
+
+ m->includes_phdrs = 0;
+
+ if (! phdr_included || p->p_type != PT_LOAD)
+ {
+ m->includes_phdrs =
+ (p->p_offset <= (bfd_vma) iehdr->e_phoff
+ && (p->p_offset + p->p_filesz
+ >= ((bfd_vma) iehdr->e_phoff
+ + iehdr->e_phnum * iehdr->e_phentsize)));
+ if (p->p_type == PT_LOAD && m->includes_phdrs)
+ phdr_included = true;
+ }
+
+ if (csecs == 0)
+ {
+ /* Special segments, such as the PT_PHDR segment, may contain
+ no sections, but ordinary, loadable segments should contain
+ something. */
+
+ if (p->p_type == PT_LOAD)
+ _bfd_error_handler
+ (_("%s: warning: Empty loadable segment detected\n"),
+ bfd_get_filename (ibfd));
+
+ m->count = 0;
+ *pm = m;
+ pm = &m->next;
+
+ continue;
+ }
+
+ /* Now scan the sections in the input BFD again and attempt
+ to add their corresponding output sections to the segment map.
+ The problem here is how to handle an output section which has
+ been moved (ie had its LMA changed). There are four possibilities:
+
+ 1. None of the sections have been moved.
+ In this case we can continue to use the segment LMA from the
+ input BFD.
+
+ 2. All of the sections have been moved by the same amount.
+ In this case we can change the segment's LMA to match the LMA
+ of the first section.
+
+ 3. Some of the sections have been moved, others have not.
+ In this case those sections which have not been moved can be
+ placed in the current segment which will have to have its size,
+ and possibly its LMA changed, and a new segment or segments will
+ have to be created to contain the other sections.
+
+ 4. The sections have been moved, but not be the same amount.
+ In this case we can change the segment's LMA to match the LMA
+ of the first section and we will have to create a new segment
+ or segments to contain the other sections.
+
+ In order to save time, we allocate an array to hold the section
+ pointers that we are interested in. As these sections get assigned
+ to a segment, they are removed from this array. */
+
+ sections = (asection **) bfd_malloc (sizeof (asection *) * csecs);
+ if (sections == NULL)
+ return false;
+
+ /* Step One: Scan for segment vs section LMA conflicts.
+ Also add the sections to the section array allocated above.
+ Also add the sections to the current segment. In the common
+ case, where the sections have not been moved, this means that
+ we have completely filled the segment, and there is nothing
+ more to do. */
+
+ isec = 0;
+ matching_lma = false;
+ suggested_lma = 0;
+
+ for (j = 0, s = ibfd->sections; s != NULL; s = s->next)
+ {
+ os = s->output_section;
+
+ if ((((IS_CONTAINED_BY (s->vma, s->_raw_size, p->p_vaddr, p)
+ || IS_SOLARIS_PT_INTERP (p, s))
+ && (s->flags & SEC_ALLOC) != 0)
+ || IS_COREFILE_NOTE (p, s))
+ && os != NULL)
+ {
+ sections[j++] = s;
+
+ /* The Solaris native linker always sets p_paddr to 0.
+ We try to catch that case here, and set it to the
+ correct value. */
+ if (p->p_paddr == 0
+ && p->p_vaddr != 0
+ && isec == 0
+ && os->lma != 0
+ && (os->vma == (p->p_vaddr
+ + (m->includes_filehdr
+ ? iehdr->e_ehsize
+ : 0)
+ + (m->includes_phdrs
+ ? iehdr->e_phnum * iehdr->e_phentsize
+ : 0))))
+ m->p_paddr = p->p_vaddr;
+
+ /* Match up the physical address of the segment with the
+ LMA address of the output section. */
+ if (IS_CONTAINED_BY (os->lma, os->_raw_size, m->p_paddr, p)
+ || IS_COREFILE_NOTE (p, s))
+ {
+ if (matching_lma == 0)
+ matching_lma = os->lma;
+
+ /* We assume that if the section fits within the segment
+ that it does not overlap any other section within that
+ segment. */
+ m->sections[isec++] = os;
+ }
+ else if (suggested_lma == 0)
+ suggested_lma = os->lma;
+ }
+ }
+
+ BFD_ASSERT (j == csecs);
+
+ /* Step Two: Adjust the physical address of the current segment,
+ if necessary. */
+ if (isec == csecs)
+ {
+ /* All of the sections fitted within the segment as currently
+ specified. This is the default case. Add the segment to
+ the list of built segments and carry on to process the next
+ program header in the input BFD. */
+ m->count = csecs;
+ *pm = m;
+ pm = &m->next;
+
+ free (sections);
+ continue;
+ }
+ else if (matching_lma != 0)
+ {
+ /* At least one section fits inside the current segment.
+ Keep it, but modify its physical address to match the
+ LMA of the first section that fitted. */
+
+ m->p_paddr = matching_lma;
+ }
+ else
+ {
+ /* None of the sections fitted inside the current segment.
+ Change the current segment's physical address to match
+ the LMA of the first section. */
+
+ m->p_paddr = suggested_lma;
+ }
+
+ /* Step Three: Loop over the sections again, this time assigning
+ those that fit to the current segment and remvoing them from the
+ sections array; but making sure not to leave large gaps. Once all
+ possible sections have been assigned to the current segment it is
+ added to the list of built segments and if sections still remain
+ to be assigned, a new segment is constructed before repeating
+ the loop. */
+ isec = 0;
+ do
+ {
+ m->count = 0;
+ suggested_lma = 0;
+
+ /* Fill the current segment with sections that fit. */
+ for (j = 0; j < csecs; j++)
+ {
+ s = sections[j];
+
+ if (s == NULL)
+ continue;
+
+ os = s->output_section;
+
+ if (IS_CONTAINED_BY (os->lma, os->_raw_size, m->p_paddr, p)
+ || IS_COREFILE_NOTE (p, s))
+ {
+ if (m->count == 0)
+ {
+ /* If the first section in a segment does not start at
+ the beginning of the segment, then something is wrong. */
+ if (os->lma != m->p_paddr)
+ abort ();
+ }
+ else
+ {
+ asection * prev_sec;
+ bfd_vma maxpagesize;
+
+ prev_sec = m->sections[m->count - 1];
+ maxpagesize = get_elf_backend_data (obfd)->maxpagesize;
+
+ /* If the gap between the end of the previous section
+ and the start of this section is more than maxpagesize
+ then we need to start a new segment. */
+ if (BFD_ALIGN (prev_sec->lma + prev_sec->_raw_size, maxpagesize)
+ < BFD_ALIGN (os->lma, maxpagesize))
+ {
+ if (suggested_lma == 0)
+ suggested_lma = os->lma;
+
+ continue;
+ }
+ }
+
+ m->sections[m->count++] = os;
+ ++isec;
+ sections[j] = NULL;
+ }
+ else if (suggested_lma == 0)
+ suggested_lma = os->lma;
+ }
+
+ BFD_ASSERT (m->count > 0);
+
+ /* Add the current segment to the list of built segments. */
+ *pm = m;
+ pm = &m->next;
+
+ if (isec < csecs)
+ {
+ /* We still have not allocated all of the sections to
+ segments. Create a new segment here, initialise it
+ and carry on looping. */
+
+ m = ((struct elf_segment_map *)
+ bfd_alloc (obfd,
+ (sizeof (struct elf_segment_map)
+ + ((size_t) csecs - 1) * sizeof (asection *))));
+ if (m == NULL)
+ return false;
+
+ /* Initialise the fields of the segment map. Set the physical
+ physical address to the LMA of the first section that has
+ not yet been assigned. */
+
+ m->next = NULL;
+ m->p_type = p->p_type;
+ m->p_flags = p->p_flags;
+ m->p_flags_valid = 1;
+ m->p_paddr = suggested_lma;
+ m->p_paddr_valid = 1;
+ m->includes_filehdr = 0;
+ m->includes_phdrs = 0;
+ }
+ }
+ while (isec < csecs);
+
+ free (sections);
+ }
+
+ /* The Solaris linker creates program headers in which all the
+ p_paddr fields are zero. When we try to objcopy or strip such a
+ file, we get confused. Check for this case, and if we find it
+ reset the p_paddr_valid fields. */
+ for (m = mfirst; m != NULL; m = m->next)
+ if (m->p_paddr != 0)
+ break;
+ if (m == NULL)
+ {
+ for (m = mfirst; m != NULL; m = m->next)
+ m->p_paddr_valid = 0;
+ }
+
+ elf_tdata (obfd)->segment_map = mfirst;
+
+#if 0
+ /* Final Step: Sort the segments into ascending order of physical address. */
+ if (mfirst != NULL)
+ {
+ struct elf_segment_map* prev;
+
+ prev = mfirst;
+ for (m = mfirst->next; m != NULL; prev = m, m = m->next)
+ {
+ /* Yes I know - its a bubble sort....*/
+ if (m->next != NULL && (m->next->p_paddr < m->p_paddr))
+ {
+ /* swap m and m->next */
+ prev->next = m->next;
+ m->next = m->next->next;
+ prev->next->next = m;
+
+ /* restart loop. */
+ m = mfirst;
+ }
+ }
+ }
+#endif
+
+#undef IS_CONTAINED_BY
+#undef IS_SOLARIS_PT_INTERP
+#undef IS_COREFILE_NOTE
+ return true;
+}
+
+/* Copy private section information. This copies over the entsize
+ field, and sometimes the info field. */
+
+boolean
+_bfd_elf_copy_private_section_data (ibfd, isec, obfd, osec)
+ bfd *ibfd;
+ asection *isec;
+ bfd *obfd;
+ asection *osec;
+{
+ Elf_Internal_Shdr *ihdr, *ohdr;
+
+ if (ibfd->xvec->flavour != bfd_target_elf_flavour
+ || obfd->xvec->flavour != bfd_target_elf_flavour)
+ return true;
+
+ /* Copy over private BFD data if it has not already been copied.
+ This must be done here, rather than in the copy_private_bfd_data
+ entry point, because the latter is called after the section
+ contents have been set, which means that the program headers have
+ already been worked out. */
+ if (elf_tdata (obfd)->segment_map == NULL
+ && elf_tdata (ibfd)->phdr != NULL)
+ {
+ asection *s;
+
+ /* Only set up the segments if there are no more SEC_ALLOC
+ sections. FIXME: This won't do the right thing if objcopy is
+ used to remove the last SEC_ALLOC section, since objcopy
+ won't call this routine in that case. */
+ for (s = isec->next; s != NULL; s = s->next)
+ if ((s->flags & SEC_ALLOC) != 0)
+ break;
+ if (s == NULL)
+ {
+ if (! copy_private_bfd_data (ibfd, obfd))
+ return false;
+ }
+ }
+
+ ihdr = &elf_section_data (isec)->this_hdr;
+ ohdr = &elf_section_data (osec)->this_hdr;
+
+ ohdr->sh_entsize = ihdr->sh_entsize;
+
+ if (ihdr->sh_type == SHT_SYMTAB
+ || ihdr->sh_type == SHT_DYNSYM
+ || ihdr->sh_type == SHT_GNU_verneed
+ || ihdr->sh_type == SHT_GNU_verdef)
+ ohdr->sh_info = ihdr->sh_info;
+
+ elf_section_data (osec)->use_rela_p
+ = elf_section_data (isec)->use_rela_p;
+
+ return true;
+}
+
+/* Copy private symbol information. If this symbol is in a section
+ which we did not map into a BFD section, try to map the section
+ index correctly. We use special macro definitions for the mapped
+ section indices; these definitions are interpreted by the
+ swap_out_syms function. */
+
+#define MAP_ONESYMTAB (SHN_LORESERVE - 1)
+#define MAP_DYNSYMTAB (SHN_LORESERVE - 2)
+#define MAP_STRTAB (SHN_LORESERVE - 3)
+#define MAP_SHSTRTAB (SHN_LORESERVE - 4)
+
+boolean
+_bfd_elf_copy_private_symbol_data (ibfd, isymarg, obfd, osymarg)
+ bfd *ibfd;
+ asymbol *isymarg;
+ bfd *obfd;
+ asymbol *osymarg;
+{
+ elf_symbol_type *isym, *osym;
+
+ if (bfd_get_flavour (ibfd) != bfd_target_elf_flavour
+ || bfd_get_flavour (obfd) != bfd_target_elf_flavour)
+ return true;
+
+ isym = elf_symbol_from (ibfd, isymarg);
+ osym = elf_symbol_from (obfd, osymarg);
+
+ if (isym != NULL
+ && osym != NULL
+ && bfd_is_abs_section (isym->symbol.section))
+ {
+ unsigned int shndx;
+
+ shndx = isym->internal_elf_sym.st_shndx;
+ if (shndx == elf_onesymtab (ibfd))
+ shndx = MAP_ONESYMTAB;
+ else if (shndx == elf_dynsymtab (ibfd))
+ shndx = MAP_DYNSYMTAB;
+ else if (shndx == elf_tdata (ibfd)->strtab_section)
+ shndx = MAP_STRTAB;
+ else if (shndx == elf_tdata (ibfd)->shstrtab_section)
+ shndx = MAP_SHSTRTAB;
+ osym->internal_elf_sym.st_shndx = shndx;
+ }
+
+ return true;
+}
+
+/* Swap out the symbols. */
+
+static boolean
+swap_out_syms (abfd, sttp, relocatable_p)
+ bfd *abfd;
+ struct bfd_strtab_hash **sttp;
+ int relocatable_p;
+{
+ struct elf_backend_data *bed = get_elf_backend_data (abfd);
+
+ if (!elf_map_symbols (abfd))
return false;
/* Dump out the symtabs. */
outbound_syms = bfd_alloc (abfd,
(1 + symcount) * bed->s->sizeof_sym);
if (outbound_syms == NULL)
- {
- bfd_set_error (bfd_error_no_memory);
- return false;
- }
+ return false;
symtab_hdr->contents = (PTR) outbound_syms;
/* now generate the data (for "contents") */
bfd_vma value = syms[idx]->value;
elf_symbol_type *type_ptr;
flagword flags = syms[idx]->flags;
+ int type;
if (flags & BSF_SECTION_SYM)
/* Section symbols have no names. */
type_ptr = elf_symbol_from (abfd, syms[idx]);
- if (bfd_is_com_section (syms[idx]->section))
+ if ((flags & BSF_SECTION_SYM) == 0
+ && bfd_is_com_section (syms[idx]->section))
{
/* ELF common symbols put the alignment into the `value' field,
and the size into the `size' field. This is backwards from
sym.st_value = value >= 16 ? 16 : (1 << bfd_log2 (value));
else
sym.st_value = type_ptr->internal_elf_sym.st_value;
- sym.st_shndx = _bfd_elf_section_from_bfd_section (abfd,
- syms[idx]->section);
+ sym.st_shndx = _bfd_elf_section_from_bfd_section
+ (abfd, syms[idx]->section);
}
else
{
value += sec->output_offset;
sec = sec->output_section;
}
- value += sec->vma;
+ /* Don't add in the section vma for relocatable output. */
+ if (! relocatable_p)
+ value += sec->vma;
sym.st_value = value;
sym.st_size = type_ptr ? type_ptr->internal_elf_sym.st_size : 0;
- sym.st_shndx = shndx = _bfd_elf_section_from_bfd_section (abfd, sec);
- if (shndx == -1)
+
+ if (bfd_is_abs_section (sec)
+ && type_ptr != NULL
+ && type_ptr->internal_elf_sym.st_shndx != 0)
+ {
+ /* This symbol is in a real ELF section which we did
+ not create as a BFD section. Undo the mapping done
+ by copy_private_symbol_data. */
+ shndx = type_ptr->internal_elf_sym.st_shndx;
+ switch (shndx)
+ {
+ case MAP_ONESYMTAB:
+ shndx = elf_onesymtab (abfd);
+ break;
+ case MAP_DYNSYMTAB:
+ shndx = elf_dynsymtab (abfd);
+ break;
+ case MAP_STRTAB:
+ shndx = elf_tdata (abfd)->strtab_section;
+ break;
+ case MAP_SHSTRTAB:
+ shndx = elf_tdata (abfd)->shstrtab_section;
+ break;
+ default:
+ break;
+ }
+ }
+ else
{
- asection *sec2;
- /* Writing this would be a hell of a lot easier if we had
- some decent documentation on bfd, and knew what to expect
- of the library, and what to demand of applications. For
- example, it appears that `objcopy' might not set the
- section of a symbol to be a section that is actually in
- the output file. */
- sec2 = bfd_get_section_by_name (abfd, sec->name);
- BFD_ASSERT (sec2 != 0);
- sym.st_shndx = shndx = _bfd_elf_section_from_bfd_section (abfd, sec2);
- BFD_ASSERT (shndx != -1);
+ shndx = _bfd_elf_section_from_bfd_section (abfd, sec);
+
+ if (shndx == -1)
+ {
+ asection *sec2;
+
+ /* Writing this would be a hell of a lot easier if
+ we had some decent documentation on bfd, and
+ knew what to expect of the library, and what to
+ demand of applications. For example, it
+ appears that `objcopy' might not set the
+ section of a symbol to be a section that is
+ actually in the output file. */
+ sec2 = bfd_get_section_by_name (abfd, sec->name);
+ BFD_ASSERT (sec2 != 0);
+ shndx = _bfd_elf_section_from_bfd_section (abfd, sec2);
+ BFD_ASSERT (shndx != -1);
+ }
}
+
+ sym.st_shndx = shndx;
}
- if (bfd_is_com_section (syms[idx]->section))
- sym.st_info = ELF_ST_INFO (STB_GLOBAL, STT_OBJECT);
+ if ((flags & BSF_FUNCTION) != 0)
+ type = STT_FUNC;
+ else if ((flags & BSF_OBJECT) != 0)
+ type = STT_OBJECT;
+ else
+ type = STT_NOTYPE;
+
+ /* Processor-specific types */
+ if (bed->elf_backend_get_symbol_type)
+ type = (*bed->elf_backend_get_symbol_type) (&type_ptr->internal_elf_sym, type);
+
+ if (flags & BSF_SECTION_SYM)
+ sym.st_info = ELF_ST_INFO (STB_LOCAL, STT_SECTION);
+ else if (bfd_is_com_section (syms[idx]->section))
+ sym.st_info = ELF_ST_INFO (STB_GLOBAL, type);
else if (bfd_is_und_section (syms[idx]->section))
sym.st_info = ELF_ST_INFO (((flags & BSF_WEAK)
? STB_WEAK
: STB_GLOBAL),
- ((flags & BSF_FUNCTION)
- ? STT_FUNC
- : STT_NOTYPE));
- else if (flags & BSF_SECTION_SYM)
- sym.st_info = ELF_ST_INFO (STB_LOCAL, STT_SECTION);
+ type);
else if (flags & BSF_FILE)
sym.st_info = ELF_ST_INFO (STB_LOCAL, STT_FILE);
else
{
int bind = STB_LOCAL;
- int type = STT_OBJECT;
if (flags & BSF_LOCAL)
bind = STB_LOCAL;
else if (flags & BSF_GLOBAL)
bind = STB_GLOBAL;
- if (flags & BSF_FUNCTION)
- type = STT_FUNC;
-
sym.st_info = ELF_ST_INFO (bind, type);
}
- sym.st_other = 0;
+ if (type_ptr != NULL)
+ sym.st_other = type_ptr->internal_elf_sym.st_other;
+ else
+ sym.st_other = 0;
+
bed->s->swap_symbol_out (abfd, &sym, (PTR) outbound_syms);
outbound_syms += bed->s->sizeof_sym;
}
long
_bfd_elf_get_reloc_upper_bound (abfd, asect)
- bfd *abfd;
+ bfd *abfd ATTRIBUTE_UNUSED;
sec_ptr asect;
{
return (asect->reloc_count + 1) * sizeof (arelent *);
arelent *tblptr;
unsigned int i;
- if (! get_elf_backend_data (abfd)->s->slurp_reloc_table (abfd, section, symbols))
+ if (! get_elf_backend_data (abfd)->s->slurp_reloc_table (abfd,
+ section,
+ symbols,
+ false))
return -1;
tblptr = section->relocation;
bfd *abfd;
asymbol **alocation;
{
- long symcount = get_elf_backend_data (abfd)->s->slurp_symbol_table (abfd, alocation, false);
+ long symcount = get_elf_backend_data (abfd)->s->slurp_symbol_table
+ (abfd, alocation, false);
if (symcount >= 0)
bfd_get_symcount (abfd) = symcount;
bfd *abfd;
asymbol **alocation;
{
- return get_elf_backend_data (abfd)->s->slurp_symbol_table (abfd, alocation, true);
+ return get_elf_backend_data (abfd)->s->slurp_symbol_table
+ (abfd, alocation, true);
}
-asymbol *
-_bfd_elf_make_empty_symbol (abfd)
+/* Return the size required for the dynamic reloc entries. Any
+ section that was actually installed in the BFD, and has type
+ SHT_REL or SHT_RELA, and uses the dynamic symbol table, is
+ considered to be a dynamic reloc section. */
+
+long
+_bfd_elf_get_dynamic_reloc_upper_bound (abfd)
bfd *abfd;
{
- elf_symbol_type *newsym;
+ long ret;
+ asection *s;
- newsym = (elf_symbol_type *) bfd_zalloc (abfd, sizeof (elf_symbol_type));
- if (!newsym)
- {
- bfd_set_error (bfd_error_no_memory);
- return NULL;
- }
- else
+ if (elf_dynsymtab (abfd) == 0)
{
- newsym->symbol.the_bfd = abfd;
- return &newsym->symbol;
+ bfd_set_error (bfd_error_invalid_operation);
+ return -1;
}
-}
-void
-_bfd_elf_get_symbol_info (ignore_abfd, symbol, ret)
- bfd *ignore_abfd;
- asymbol *symbol;
- symbol_info *ret;
-{
- bfd_symbol_info (symbol, ret);
-}
+ ret = sizeof (arelent *);
+ for (s = abfd->sections; s != NULL; s = s->next)
+ if (elf_section_data (s)->this_hdr.sh_link == elf_dynsymtab (abfd)
+ && (elf_section_data (s)->this_hdr.sh_type == SHT_REL
+ || elf_section_data (s)->this_hdr.sh_type == SHT_RELA))
+ ret += ((s->_raw_size / elf_section_data (s)->this_hdr.sh_entsize)
+ * sizeof (arelent *));
-alent *
-_bfd_elf_get_lineno (ignore_abfd, symbol)
- bfd *ignore_abfd;
- asymbol *symbol;
-{
- abort ();
- return NULL;
+ return ret;
}
-boolean
-_bfd_elf_set_arch_mach (abfd, arch, machine)
+/* Canonicalize the dynamic relocation entries. Note that we return
+ the dynamic relocations as a single block, although they are
+ actually associated with particular sections; the interface, which
+ was designed for SunOS style shared libraries, expects that there
+ is only one set of dynamic relocs. Any section that was actually
+ installed in the BFD, and has type SHT_REL or SHT_RELA, and uses
+ the dynamic symbol table, is considered to be a dynamic reloc
+ section. */
+
+long
+_bfd_elf_canonicalize_dynamic_reloc (abfd, storage, syms)
bfd *abfd;
- enum bfd_architecture arch;
- unsigned long machine;
+ arelent **storage;
+ asymbol **syms;
{
- /* If this isn't the right architecture for this backend, and this
- isn't the generic backend, fail. */
- if (arch != get_elf_backend_data (abfd)->arch
- && arch != bfd_arch_unknown
- && get_elf_backend_data (abfd)->arch != bfd_arch_unknown)
- return false;
+ boolean (*slurp_relocs) PARAMS ((bfd *, asection *, asymbol **, boolean));
+ asection *s;
+ long ret;
- return bfd_default_set_arch_mach (abfd, arch, machine);
-}
+ if (elf_dynsymtab (abfd) == 0)
+ {
+ bfd_set_error (bfd_error_invalid_operation);
+ return -1;
+ }
-/* Find the nearest line to a particular section and offset, for error
- reporting. */
+ slurp_relocs = get_elf_backend_data (abfd)->s->slurp_reloc_table;
+ ret = 0;
+ for (s = abfd->sections; s != NULL; s = s->next)
+ {
+ if (elf_section_data (s)->this_hdr.sh_link == elf_dynsymtab (abfd)
+ && (elf_section_data (s)->this_hdr.sh_type == SHT_REL
+ || elf_section_data (s)->this_hdr.sh_type == SHT_RELA))
+ {
+ arelent *p;
+ long count, i;
+
+ if (! (*slurp_relocs) (abfd, s, syms, true))
+ return -1;
+ count = s->_raw_size / elf_section_data (s)->this_hdr.sh_entsize;
+ p = s->relocation;
+ for (i = 0; i < count; i++)
+ *storage++ = p++;
+ ret += count;
+ }
+ }
+
+ *storage = NULL;
+
+ return ret;
+}
+\f
+/* Read in the version information. */
boolean
-_bfd_elf_find_nearest_line (abfd,
- section,
- symbols,
- offset,
- filename_ptr,
- functionname_ptr,
- line_ptr)
+_bfd_elf_slurp_version_tables (abfd)
bfd *abfd;
- asection *section;
- asymbol **symbols;
- bfd_vma offset;
- CONST char **filename_ptr;
- CONST char **functionname_ptr;
- unsigned int *line_ptr;
{
- const char *filename;
- asymbol *func;
- asymbol **p;
-
- if (symbols == NULL)
- return false;
+ bfd_byte *contents = NULL;
+
+ if (elf_dynverdef (abfd) != 0)
+ {
+ Elf_Internal_Shdr *hdr;
+ Elf_External_Verdef *everdef;
+ Elf_Internal_Verdef *iverdef;
+ unsigned int i;
+
+ hdr = &elf_tdata (abfd)->dynverdef_hdr;
+
+ elf_tdata (abfd)->verdef =
+ ((Elf_Internal_Verdef *)
+ bfd_zalloc (abfd, hdr->sh_info * sizeof (Elf_Internal_Verdef)));
+ if (elf_tdata (abfd)->verdef == NULL)
+ goto error_return;
+
+ elf_tdata (abfd)->cverdefs = hdr->sh_info;
+
+ contents = (bfd_byte *) bfd_malloc (hdr->sh_size);
+ if (contents == NULL)
+ goto error_return;
+ if (bfd_seek (abfd, hdr->sh_offset, SEEK_SET) != 0
+ || bfd_read ((PTR) contents, 1, hdr->sh_size, abfd) != hdr->sh_size)
+ goto error_return;
+
+ everdef = (Elf_External_Verdef *) contents;
+ iverdef = elf_tdata (abfd)->verdef;
+ for (i = 0; i < hdr->sh_info; i++, iverdef++)
+ {
+ Elf_External_Verdaux *everdaux;
+ Elf_Internal_Verdaux *iverdaux;
+ unsigned int j;
+
+ _bfd_elf_swap_verdef_in (abfd, everdef, iverdef);
+
+ iverdef->vd_bfd = abfd;
+
+ iverdef->vd_auxptr = ((Elf_Internal_Verdaux *)
+ bfd_alloc (abfd,
+ (iverdef->vd_cnt
+ * sizeof (Elf_Internal_Verdaux))));
+ if (iverdef->vd_auxptr == NULL)
+ goto error_return;
+
+ everdaux = ((Elf_External_Verdaux *)
+ ((bfd_byte *) everdef + iverdef->vd_aux));
+ iverdaux = iverdef->vd_auxptr;
+ for (j = 0; j < iverdef->vd_cnt; j++, iverdaux++)
+ {
+ _bfd_elf_swap_verdaux_in (abfd, everdaux, iverdaux);
+
+ iverdaux->vda_nodename =
+ bfd_elf_string_from_elf_section (abfd, hdr->sh_link,
+ iverdaux->vda_name);
+ if (iverdaux->vda_nodename == NULL)
+ goto error_return;
+
+ if (j + 1 < iverdef->vd_cnt)
+ iverdaux->vda_nextptr = iverdaux + 1;
+ else
+ iverdaux->vda_nextptr = NULL;
+
+ everdaux = ((Elf_External_Verdaux *)
+ ((bfd_byte *) everdaux + iverdaux->vda_next));
+ }
+
+ iverdef->vd_nodename = iverdef->vd_auxptr->vda_nodename;
+
+ if (i + 1 < hdr->sh_info)
+ iverdef->vd_nextdef = iverdef + 1;
+ else
+ iverdef->vd_nextdef = NULL;
+
+ everdef = ((Elf_External_Verdef *)
+ ((bfd_byte *) everdef + iverdef->vd_next));
+ }
+
+ free (contents);
+ contents = NULL;
+ }
+
+ if (elf_dynverref (abfd) != 0)
+ {
+ Elf_Internal_Shdr *hdr;
+ Elf_External_Verneed *everneed;
+ Elf_Internal_Verneed *iverneed;
+ unsigned int i;
+
+ hdr = &elf_tdata (abfd)->dynverref_hdr;
+
+ elf_tdata (abfd)->verref =
+ ((Elf_Internal_Verneed *)
+ bfd_zalloc (abfd, hdr->sh_info * sizeof (Elf_Internal_Verneed)));
+ if (elf_tdata (abfd)->verref == NULL)
+ goto error_return;
+
+ elf_tdata (abfd)->cverrefs = hdr->sh_info;
+
+ contents = (bfd_byte *) bfd_malloc (hdr->sh_size);
+ if (contents == NULL)
+ goto error_return;
+ if (bfd_seek (abfd, hdr->sh_offset, SEEK_SET) != 0
+ || bfd_read ((PTR) contents, 1, hdr->sh_size, abfd) != hdr->sh_size)
+ goto error_return;
+
+ everneed = (Elf_External_Verneed *) contents;
+ iverneed = elf_tdata (abfd)->verref;
+ for (i = 0; i < hdr->sh_info; i++, iverneed++)
+ {
+ Elf_External_Vernaux *evernaux;
+ Elf_Internal_Vernaux *ivernaux;
+ unsigned int j;
+
+ _bfd_elf_swap_verneed_in (abfd, everneed, iverneed);
+
+ iverneed->vn_bfd = abfd;
+
+ iverneed->vn_filename =
+ bfd_elf_string_from_elf_section (abfd, hdr->sh_link,
+ iverneed->vn_file);
+ if (iverneed->vn_filename == NULL)
+ goto error_return;
+
+ iverneed->vn_auxptr =
+ ((Elf_Internal_Vernaux *)
+ bfd_alloc (abfd,
+ iverneed->vn_cnt * sizeof (Elf_Internal_Vernaux)));
+
+ evernaux = ((Elf_External_Vernaux *)
+ ((bfd_byte *) everneed + iverneed->vn_aux));
+ ivernaux = iverneed->vn_auxptr;
+ for (j = 0; j < iverneed->vn_cnt; j++, ivernaux++)
+ {
+ _bfd_elf_swap_vernaux_in (abfd, evernaux, ivernaux);
+
+ ivernaux->vna_nodename =
+ bfd_elf_string_from_elf_section (abfd, hdr->sh_link,
+ ivernaux->vna_name);
+ if (ivernaux->vna_nodename == NULL)
+ goto error_return;
+
+ if (j + 1 < iverneed->vn_cnt)
+ ivernaux->vna_nextptr = ivernaux + 1;
+ else
+ ivernaux->vna_nextptr = NULL;
+
+ evernaux = ((Elf_External_Vernaux *)
+ ((bfd_byte *) evernaux + ivernaux->vna_next));
+ }
+
+ if (i + 1 < hdr->sh_info)
+ iverneed->vn_nextref = iverneed + 1;
+ else
+ iverneed->vn_nextref = NULL;
+
+ everneed = ((Elf_External_Verneed *)
+ ((bfd_byte *) everneed + iverneed->vn_next));
+ }
+
+ free (contents);
+ contents = NULL;
+ }
+
+ return true;
+
+ error_return:
+ if (contents == NULL)
+ free (contents);
+ return false;
+}
+\f
+asymbol *
+_bfd_elf_make_empty_symbol (abfd)
+ bfd *abfd;
+{
+ elf_symbol_type *newsym;
+
+ newsym = (elf_symbol_type *) bfd_zalloc (abfd, sizeof (elf_symbol_type));
+ if (!newsym)
+ return NULL;
+ else
+ {
+ newsym->symbol.the_bfd = abfd;
+ return &newsym->symbol;
+ }
+}
+
+void
+_bfd_elf_get_symbol_info (ignore_abfd, symbol, ret)
+ bfd *ignore_abfd ATTRIBUTE_UNUSED;
+ asymbol *symbol;
+ symbol_info *ret;
+{
+ bfd_symbol_info (symbol, ret);
+}
+
+/* Return whether a symbol name implies a local symbol. Most targets
+ use this function for the is_local_label_name entry point, but some
+ override it. */
+
+boolean
+_bfd_elf_is_local_label_name (abfd, name)
+ bfd *abfd ATTRIBUTE_UNUSED;
+ const char *name;
+{
+ /* Normal local symbols start with ``.L''. */
+ if (name[0] == '.' && name[1] == 'L')
+ return true;
+
+ /* At least some SVR4 compilers (e.g., UnixWare 2.1 cc) generate
+ DWARF debugging symbols starting with ``..''. */
+ if (name[0] == '.' && name[1] == '.')
+ return true;
+
+ /* gcc will sometimes generate symbols beginning with ``_.L_'' when
+ emitting DWARF debugging output. I suspect this is actually a
+ small bug in gcc (it calls ASM_OUTPUT_LABEL when it should call
+ ASM_GENERATE_INTERNAL_LABEL, and this causes the leading
+ underscore to be emitted on some ELF targets). For ease of use,
+ we treat such symbols as local. */
+ if (name[0] == '_' && name[1] == '.' && name[2] == 'L' && name[3] == '_')
+ return true;
+
+ return false;
+}
+
+alent *
+_bfd_elf_get_lineno (ignore_abfd, symbol)
+ bfd *ignore_abfd ATTRIBUTE_UNUSED;
+ asymbol *symbol ATTRIBUTE_UNUSED;
+{
+ abort ();
+ return NULL;
+}
+
+boolean
+_bfd_elf_set_arch_mach (abfd, arch, machine)
+ bfd *abfd;
+ enum bfd_architecture arch;
+ unsigned long machine;
+{
+ /* If this isn't the right architecture for this backend, and this
+ isn't the generic backend, fail. */
+ if (arch != get_elf_backend_data (abfd)->arch
+ && arch != bfd_arch_unknown
+ && get_elf_backend_data (abfd)->arch != bfd_arch_unknown)
+ return false;
+
+ return bfd_default_set_arch_mach (abfd, arch, machine);
+}
+
+/* Find the nearest line to a particular section and offset, for error
+ reporting. */
+
+boolean
+_bfd_elf_find_nearest_line (abfd,
+ section,
+ symbols,
+ offset,
+ filename_ptr,
+ functionname_ptr,
+ line_ptr)
+ bfd *abfd;
+ asection *section;
+ asymbol **symbols;
+ bfd_vma offset;
+ CONST char **filename_ptr;
+ CONST char **functionname_ptr;
+ unsigned int *line_ptr;
+{
+ boolean found;
+ const char *filename;
+ asymbol *func;
+ bfd_vma low_func;
+ asymbol **p;
+
+ if (_bfd_dwarf1_find_nearest_line (abfd, section, symbols, offset,
+ filename_ptr, functionname_ptr,
+ line_ptr))
+ return true;
+
+ if (_bfd_dwarf2_find_nearest_line (abfd, section, symbols, offset,
+ filename_ptr, functionname_ptr,
+ line_ptr, 0))
+ return true;
+
+ if (! _bfd_stab_section_find_nearest_line (abfd, symbols, section, offset,
+ &found, filename_ptr,
+ functionname_ptr, line_ptr,
+ &elf_tdata (abfd)->line_info))
+ return false;
+ if (found)
+ return true;
+
+ if (symbols == NULL)
+ return false;
filename = NULL;
func = NULL;
+ low_func = 0;
for (p = symbols; *p != NULL; p++)
{
case STT_FILE:
filename = bfd_asymbol_name (&q->symbol);
break;
+ case STT_NOTYPE:
case STT_FUNC:
- if (func == NULL
- || q->symbol.value <= offset)
- func = (asymbol *) q;
+ if (q->symbol.section == section
+ && q->symbol.value >= low_func
+ && q->symbol.value <= offset)
+ {
+ func = (asymbol *) q;
+ low_func = q->symbol.value;
+ }
break;
}
}
ret = get_elf_backend_data (abfd)->s->sizeof_ehdr;
if (! reloc)
- ret += get_program_header_size (abfd, (Elf_Internal_Shdr **) NULL, 0,
- (bfd_vma) 0);
+ ret += get_program_header_size (abfd);
return ret;
}
Elf_Internal_Shdr *hdr;
if (! abfd->output_has_begun
- && ! _bfd_elf_compute_section_file_positions (abfd,
- (struct bfd_link_info *) NULL))
+ && ! _bfd_elf_compute_section_file_positions
+ (abfd, (struct bfd_link_info *) NULL))
return false;
hdr = &elf_section_data (section)->this_hdr;
void
_bfd_elf_no_info_to_howto (abfd, cache_ptr, dst)
- bfd *abfd;
- arelent *cache_ptr;
- Elf_Internal_Rela *dst;
+ bfd *abfd ATTRIBUTE_UNUSED;
+ arelent *cache_ptr ATTRIBUTE_UNUSED;
+ Elf_Internal_Rela *dst ATTRIBUTE_UNUSED;
{
abort ();
}
abort ();
}
#endif
+
+/* Try to convert a non-ELF reloc into an ELF one. */
+
+boolean
+_bfd_elf_validate_reloc (abfd, areloc)
+ bfd *abfd;
+ arelent *areloc;
+{
+ /* Check whether we really have an ELF howto. */
+
+ if ((*areloc->sym_ptr_ptr)->the_bfd->xvec != abfd->xvec)
+ {
+ bfd_reloc_code_real_type code;
+ reloc_howto_type *howto;
+
+ /* Alien reloc: Try to determine its type to replace it with an
+ equivalent ELF reloc. */
+
+ if (areloc->howto->pc_relative)
+ {
+ switch (areloc->howto->bitsize)
+ {
+ case 8:
+ code = BFD_RELOC_8_PCREL;
+ break;
+ case 12:
+ code = BFD_RELOC_12_PCREL;
+ break;
+ case 16:
+ code = BFD_RELOC_16_PCREL;
+ break;
+ case 24:
+ code = BFD_RELOC_24_PCREL;
+ break;
+ case 32:
+ code = BFD_RELOC_32_PCREL;
+ break;
+ case 64:
+ code = BFD_RELOC_64_PCREL;
+ break;
+ default:
+ goto fail;
+ }
+
+ howto = bfd_reloc_type_lookup (abfd, code);
+
+ if (areloc->howto->pcrel_offset != howto->pcrel_offset)
+ {
+ if (howto->pcrel_offset)
+ areloc->addend += areloc->address;
+ else
+ areloc->addend -= areloc->address; /* addend is unsigned!! */
+ }
+ }
+ else
+ {
+ switch (areloc->howto->bitsize)
+ {
+ case 8:
+ code = BFD_RELOC_8;
+ break;
+ case 14:
+ code = BFD_RELOC_14;
+ break;
+ case 16:
+ code = BFD_RELOC_16;
+ break;
+ case 26:
+ code = BFD_RELOC_26;
+ break;
+ case 32:
+ code = BFD_RELOC_32;
+ break;
+ case 64:
+ code = BFD_RELOC_64;
+ break;
+ default:
+ goto fail;
+ }
+
+ howto = bfd_reloc_type_lookup (abfd, code);
+ }
+
+ if (howto)
+ areloc->howto = howto;
+ else
+ goto fail;
+ }
+
+ return true;
+
+ fail:
+ (*_bfd_error_handler)
+ (_("%s: unsupported relocation type %s"),
+ bfd_get_filename (abfd), areloc->howto->name);
+ bfd_set_error (bfd_error_bad_value);
+ return false;
+}
+
+boolean
+_bfd_elf_close_and_cleanup (abfd)
+ bfd *abfd;
+{
+ if (bfd_get_format (abfd) == bfd_object)
+ {
+ if (elf_shstrtab (abfd) != NULL)
+ _bfd_stringtab_free (elf_shstrtab (abfd));
+ }
+
+ return _bfd_generic_close_and_cleanup (abfd);
+}
+
+/* For Rel targets, we encode meaningful data for BFD_RELOC_VTABLE_ENTRY
+ in the relocation's offset. Thus we cannot allow any sort of sanity
+ range-checking to interfere. There is nothing else to do in processing
+ this reloc. */
+
+bfd_reloc_status_type
+_bfd_elf_rel_vtable_reloc_fn (abfd, re, symbol, data, is, obfd, errmsg)
+ bfd *abfd ATTRIBUTE_UNUSED;
+ arelent *re ATTRIBUTE_UNUSED;
+ struct symbol_cache_entry *symbol ATTRIBUTE_UNUSED;
+ PTR data ATTRIBUTE_UNUSED;
+ asection *is ATTRIBUTE_UNUSED;
+ bfd *obfd ATTRIBUTE_UNUSED;
+ char **errmsg ATTRIBUTE_UNUSED;
+{
+ return bfd_reloc_ok;
+}
+
+\f
+/* Elf core file support. Much of this only works on native
+ toolchains, since we rely on knowing the
+ machine-dependent procfs structure in order to pick
+ out details about the corefile. */
+
+#ifdef HAVE_SYS_PROCFS_H
+# include <sys/procfs.h>
+#endif
+
+
+/* Define offsetof for those systems which lack it. */
+
+#ifndef offsetof
+# define offsetof(TYPE, MEMBER) ((unsigned long) &((TYPE *)0)->MEMBER)
+#endif
+
+
+/* FIXME: this is kinda wrong, but it's what gdb wants. */
+
+static int
+elfcore_make_pid (abfd)
+ bfd* abfd;
+{
+ return ((elf_tdata (abfd)->core_lwpid << 16)
+ + (elf_tdata (abfd)->core_pid));
+}
+
+
+/* If there isn't a section called NAME, make one, using
+ data from SECT. Note, this function will generate a
+ reference to NAME, so you shouldn't deallocate or
+ overwrite it. */
+
+static boolean
+elfcore_maybe_make_sect (abfd, name, sect)
+ bfd* abfd;
+ char* name;
+ asection* sect;
+{
+ asection* sect2;
+
+ if (bfd_get_section_by_name (abfd, name) != NULL)
+ return true;
+
+ sect2 = bfd_make_section (abfd, name);
+ if (sect2 == NULL)
+ return false;
+
+ sect2->_raw_size = sect->_raw_size;
+ sect2->filepos = sect->filepos;
+ sect2->flags = sect->flags;
+ sect2->alignment_power = sect->alignment_power;
+ return true;
+}
+
+
+/* prstatus_t exists on:
+ solaris 2.[567]
+ linux 2.[01] + glibc
+ unixware 4.2
+*/
+
+#if defined (HAVE_PRSTATUS_T)
+static boolean
+elfcore_grok_prstatus (abfd, note)
+ bfd* abfd;
+ Elf_Internal_Note* note;
+{
+ prstatus_t prstat;
+ char buf[100];
+ char* name;
+ asection* sect;
+
+ if (note->descsz != sizeof (prstat))
+ return true;
+
+ memcpy (&prstat, note->descdata, sizeof (prstat));
+
+ elf_tdata (abfd)->core_signal = prstat.pr_cursig;
+ elf_tdata (abfd)->core_pid = prstat.pr_pid;
+
+ /* pr_who exists on:
+ solaris 2.[567]
+ unixware 4.2
+ pr_who doesn't exist on:
+ linux 2.[01]
+ */
+#if defined (HAVE_PRSTATUS_T_PR_WHO)
+ elf_tdata (abfd)->core_lwpid = prstat.pr_who;
+#endif
+
+ /* Make a ".reg/999" section. */
+
+ sprintf (buf, ".reg/%d", elfcore_make_pid (abfd));
+ name = bfd_alloc (abfd, strlen (buf) + 1);
+ if (name == NULL)
+ return false;
+ strcpy (name, buf);
+
+ sect = bfd_make_section (abfd, name);
+ if (sect == NULL)
+ return false;
+ sect->_raw_size = sizeof (prstat.pr_reg);
+ sect->filepos = note->descpos + offsetof (prstatus_t, pr_reg);
+ sect->flags = SEC_HAS_CONTENTS;
+ sect->alignment_power = 2;
+
+ if (! elfcore_maybe_make_sect (abfd, ".reg", sect))
+ return false;
+
+ return true;
+}
+#endif /* defined (HAVE_PRSTATUS_T) */
+
+
+/* There isn't a consistent prfpregset_t across platforms,
+ but it doesn't matter, because we don't have to pick this
+ data structure apart. */
+
+static boolean
+elfcore_grok_prfpreg (abfd, note)
+ bfd* abfd;
+ Elf_Internal_Note* note;
+{
+ char buf[100];
+ char* name;
+ asection* sect;
+
+ /* Make a ".reg2/999" section. */
+
+ sprintf (buf, ".reg2/%d", elfcore_make_pid (abfd));
+ name = bfd_alloc (abfd, strlen (buf) + 1);
+ if (name == NULL)
+ return false;
+ strcpy (name, buf);
+
+ sect = bfd_make_section (abfd, name);
+ if (sect == NULL)
+ return false;
+ sect->_raw_size = note->descsz;
+ sect->filepos = note->descpos;
+ sect->flags = SEC_HAS_CONTENTS;
+ sect->alignment_power = 2;
+
+ if (! elfcore_maybe_make_sect (abfd, ".reg2", sect))
+ return false;
+
+ return true;
+}
+
+#if defined (HAVE_PRPSINFO_T)
+# define elfcore_psinfo_t prpsinfo_t
+#endif
+
+#if defined (HAVE_PSINFO_T)
+# define elfcore_psinfo_t psinfo_t
+#endif
+
+
+#if defined (HAVE_PRPSINFO_T) || defined (HAVE_PSINFO_T)
+
+/* return a malloc'ed copy of a string at START which is at
+ most MAX bytes long, possibly without a terminating '\0'.
+ the copy will always have a terminating '\0'. */
+
+static char*
+elfcore_strndup (abfd, start, max)
+ bfd* abfd;
+ char* start;
+ int max;
+{
+ char* dup;
+ char* end = memchr (start, '\0', max);
+ int len;
+
+ if (end == NULL)
+ len = max;
+ else
+ len = end - start;
+
+ dup = bfd_alloc (abfd, len + 1);
+ if (dup == NULL)
+ return NULL;
+
+ memcpy (dup, start, len);
+ dup[len] = '\0';
+
+ return dup;
+}
+
+static boolean
+elfcore_grok_psinfo (abfd, note)
+ bfd* abfd;
+ Elf_Internal_Note* note;
+{
+ elfcore_psinfo_t psinfo;
+
+ if (note->descsz != sizeof (elfcore_psinfo_t))
+ return true;
+
+ memcpy (&psinfo, note->descdata, note->descsz);
+
+ elf_tdata (abfd)->core_program
+ = elfcore_strndup (abfd, psinfo.pr_fname, sizeof (psinfo.pr_fname));
+
+ elf_tdata (abfd)->core_command
+ = elfcore_strndup (abfd, psinfo.pr_psargs, sizeof (psinfo.pr_psargs));
+
+ /* Note that for some reason, a spurious space is tacked
+ onto the end of the args in some (at least one anyway)
+ implementations, so strip it off if it exists. */
+
+ {
+ char* command = elf_tdata (abfd)->core_command;
+ int n = strlen (command);
+
+ if (0 < n && command[n - 1] == ' ')
+ command[n - 1] = '\0';
+ }
+
+ return true;
+}
+#endif /* defined (HAVE_PRPSINFO_T) || defined (HAVE_PSINFO_T) */
+
+
+#if defined (HAVE_PSTATUS_T)
+static boolean
+elfcore_grok_pstatus (abfd, note)
+ bfd* abfd;
+ Elf_Internal_Note* note;
+{
+ pstatus_t pstat;
+
+ if (note->descsz != sizeof (pstat))
+ return true;
+
+ memcpy (&pstat, note->descdata, sizeof (pstat));
+
+ elf_tdata (abfd)->core_pid = pstat.pr_pid;
+
+ /* Could grab some more details from the "representative"
+ lwpstatus_t in pstat.pr_lwp, but we'll catch it all in an
+ NT_LWPSTATUS note, presumably. */
+
+ return true;
+}
+#endif /* defined (HAVE_PSTATUS_T) */
+
+
+#if defined (HAVE_LWPSTATUS_T)
+static boolean
+elfcore_grok_lwpstatus (abfd, note)
+ bfd* abfd;
+ Elf_Internal_Note* note;
+{
+ lwpstatus_t lwpstat;
+ char buf[100];
+ char* name;
+ asection* sect;
+
+ if (note->descsz != sizeof (lwpstat))
+ return true;
+
+ memcpy (&lwpstat, note->descdata, sizeof (lwpstat));
+
+ elf_tdata (abfd)->core_lwpid = lwpstat.pr_lwpid;
+ elf_tdata (abfd)->core_signal = lwpstat.pr_cursig;
+
+ /* Make a ".reg/999" section. */
+
+ sprintf (buf, ".reg/%d", elfcore_make_pid (abfd));
+ name = bfd_alloc (abfd, strlen (buf) + 1);
+ if (name == NULL)
+ return false;
+ strcpy (name, buf);
+
+ sect = bfd_make_section (abfd, name);
+ if (sect == NULL)
+ return false;
+
+#if defined (HAVE_LWPSTATUS_T_PR_CONTEXT)
+ sect->_raw_size = sizeof (lwpstat.pr_context.uc_mcontext.gregs);
+ sect->filepos = note->descpos
+ + offsetof (lwpstatus_t, pr_context.uc_mcontext.gregs);
+#endif
+
+#if defined (HAVE_LWPSTATUS_T_PR_REG)
+ sect->_raw_size = sizeof (lwpstat.pr_reg);
+ sect->filepos = note->descpos + offsetof (lwpstatus_t, pr_reg);
+#endif
+
+ sect->flags = SEC_HAS_CONTENTS;
+ sect->alignment_power = 2;
+
+ if (!elfcore_maybe_make_sect (abfd, ".reg", sect))
+ return false;
+
+ /* Make a ".reg2/999" section */
+
+ sprintf (buf, ".reg2/%d", elfcore_make_pid (abfd));
+ name = bfd_alloc (abfd, strlen (buf) + 1);
+ if (name == NULL)
+ return false;
+ strcpy (name, buf);
+
+ sect = bfd_make_section (abfd, name);
+ if (sect == NULL)
+ return false;
+
+#if defined (HAVE_LWPSTATUS_T_PR_CONTEXT)
+ sect->_raw_size = sizeof (lwpstat.pr_context.uc_mcontext.fpregs);
+ sect->filepos = note->descpos
+ + offsetof (lwpstatus_t, pr_context.uc_mcontext.fpregs);
+#endif
+
+#if defined (HAVE_LWPSTATUS_T_PR_FPREG)
+ sect->_raw_size = sizeof (lwpstat.pr_fpreg);
+ sect->filepos = note->descpos + offsetof (lwpstatus_t, pr_fpreg);
+#endif
+
+ sect->flags = SEC_HAS_CONTENTS;
+ sect->alignment_power = 2;
+
+ if (!elfcore_maybe_make_sect (abfd, ".reg2", sect))
+ return false;
+
+ return true;
+}
+#endif /* defined (HAVE_LWPSTATUS_T) */
+
+
+
+static boolean
+elfcore_grok_note (abfd, note)
+ bfd* abfd;
+ Elf_Internal_Note* note;
+{
+ switch (note->type)
+ {
+ default:
+ return true;
+
+#if defined (HAVE_PRSTATUS_T)
+ case NT_PRSTATUS:
+ return elfcore_grok_prstatus (abfd, note);
+#endif
+
+#if defined (HAVE_PSTATUS_T)
+ case NT_PSTATUS:
+ return elfcore_grok_pstatus (abfd, note);
+#endif
+
+#if defined (HAVE_LWPSTATUS_T)
+ case NT_LWPSTATUS:
+ return elfcore_grok_lwpstatus (abfd, note);
+#endif
+
+ case NT_FPREGSET: /* FIXME: rename to NT_PRFPREG */
+ return elfcore_grok_prfpreg (abfd, note);
+
+#if defined (HAVE_PRPSINFO_T) || defined (HAVE_PSINFO_T)
+ case NT_PRPSINFO:
+ case NT_PSINFO:
+ return elfcore_grok_psinfo (abfd, note);
+#endif
+ }
+}
+
+
+static boolean
+elfcore_read_notes (abfd, offset, size)
+ bfd* abfd;
+ bfd_vma offset;
+ bfd_vma size;
+{
+ char* buf;
+ char* p;
+
+ if (size <= 0)
+ return true;
+
+ if (bfd_seek (abfd, offset, SEEK_SET) == -1)
+ return false;
+
+ buf = bfd_malloc ((size_t) size);
+ if (buf == NULL)
+ return false;
+
+ if (bfd_read (buf, size, 1, abfd) != size)
+ {
+ error:
+ free (buf);
+ return false;
+ }
+
+ p = buf;
+ while (p < buf + size)
+ {
+ /* FIXME: bad alignment assumption. */
+ Elf_External_Note* xnp = (Elf_External_Note*) p;
+ Elf_Internal_Note in;
+
+ in.type = bfd_h_get_32 (abfd, (bfd_byte *) xnp->type);
+
+ in.namesz = bfd_h_get_32 (abfd, (bfd_byte *) xnp->namesz);
+ in.namedata = xnp->name;
+
+ in.descsz = bfd_h_get_32 (abfd, (bfd_byte *) xnp->descsz);
+ in.descdata = in.namedata + BFD_ALIGN (in.namesz, 4);
+ in.descpos = offset + (in.descdata - buf);
+
+ if (! elfcore_grok_note (abfd, &in))
+ goto error;
+
+ p = in.descdata + BFD_ALIGN (in.descsz, 4);
+ }
+
+ free (buf);
+ return true;
+}
+
+
+
+boolean
+_bfd_elfcore_section_from_phdr (abfd, phdr, sec_num)
+ bfd* abfd;
+ Elf_Internal_Phdr* phdr;
+ int sec_num;
+{
+ if (! bfd_section_from_phdr (abfd, phdr, sec_num))
+ return false;
+
+ if (phdr->p_type == PT_NOTE
+ && ! elfcore_read_notes (abfd, phdr->p_offset, phdr->p_filesz))
+ return false;
+
+ return true;
+}
+
projects / deliverable / binutils-gdb.git / blobdiff