gold/archive.cc

   1 // archive.cc -- archive support for gold
   2
   3 // Copyright 2006, 2007 Free Software Foundation, Inc.
   4 // Written by Ian Lance Taylor <iant@google.com>.
   5
   6 // This file is part of gold.
   7
   8 // This program is free software; you can redistribute it and/or modify
   9 // it under the terms of the GNU General Public License as published by
  10 // the Free Software Foundation; either version 3 of the License, or
  11 // (at your option) any later version.
  12
  13 // This program is distributed in the hope that it will be useful,
  14 // but WITHOUT ANY WARRANTY; without even the implied warranty of
  15 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16 // GNU General Public License for more details.
  17
  18 // You should have received a copy of the GNU General Public License
  19 // along with this program; if not, write to the Free Software
  20 // Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston,
  21 // MA 02110-1301, USA.
  22
  23 #include "gold.h"
  24
  25 #include <cerrno>
  26 #include <cstring>
  27 #include <climits>
  28 #include <vector>
  29
  30 #include "elfcpp.h"
  31 #include "options.h"
  32 #include "fileread.h"
  33 #include "readsyms.h"
  34 #include "symtab.h"
  35 #include "object.h"
  36 #include "archive.h"
  37
  38 namespace gold
  39 {
  40
  41 // The header of an entry in the archive.  This is all readable text,
  42 // padded with spaces where necesary.  If the contents of an archive
  43 // are all text file, the entire archive is readable.
  44
  45 struct Archive::Archive_header
  46 {
  47   // The entry name.
  48   char ar_name[16];
  49   // The file modification time.
  50   char ar_date[12];
  51   // The user's UID in decimal.
  52   char ar_uid[6];
  53   // The user's GID in decimal.
  54   char ar_gid[6];
  55   // The file mode in octal.
  56   char ar_mode[8];
  57   // The file size in decimal.
  58   char ar_size[10];
  59   // The final magic code.
  60   char ar_fmag[2];
  61 };
  62
  63 // Archive methods.
  64
  65 const char Archive::armag[sarmag] =
  66 {
  67   '!', '<', 'a', 'r', 'c', 'h', '>', '\n'
  68 };
  69
  70 const char Archive::arfmag[2] = { '`', '\n' };
  71
  72 // Set up the archive: read the symbol map and the extended name
  73 // table.
  74
  75 void
  76 Archive::setup()
  77 {
  78   // The first member of the archive should be the symbol table.
  79   std::string armap_name;
  80   off_t armap_size = this->read_header(sarmag, &armap_name);
  81   off_t off;
  82   if (armap_name.empty())
  83     {
  84       this->read_armap(sarmag + sizeof(Archive_header), armap_size);
  85       off = sarmag + sizeof(Archive_header) + armap_size;
  86     }
  87   else if (!this->input_file_->options().include_whole_archive())
  88     {
  89       fprintf(stderr, _("%s: %s: no archive symbol table (run ranlib)\n"),
  90               program_name, this->name().c_str());
  91       gold_exit(false);
  92     }
  93   else
  94     off = sarmag;
  95
  96   // See if there is an extended name table.
  97   if ((off & 1) != 0)
  98     ++off;
  99   std::string xname;
 100   off_t extended_size = this->read_header(off, &xname);
 101   if (xname == "/")
 102     {
 103       const unsigned char* p = this->get_view(off + sizeof(Archive_header),
 104                                               extended_size);
 105       const char* px = reinterpret_cast<const char*>(p);
 106       this->extended_names_.assign(px, extended_size);
 107     }
 108
 109   // Opening the file locked it.  Unlock it now.
 110   this->input_file_->file().unlock();
 111 }
 112
 113 // Read the archive symbol map.
 114
 115 void
 116 Archive::read_armap(off_t start, off_t size)
 117 {
 118   // Read in the entire armap.
 119   const unsigned char* p = this->get_view(start, size);
 120
 121   // Numbers in the armap are always big-endian.
 122   const elfcpp::Elf_Word* pword = reinterpret_cast<const elfcpp::Elf_Word*>(p);
 123   unsigned int nsyms = elfcpp::Swap<32, true>::readval(pword);
 124   ++pword;
 125
 126   // Note that the addition is in units of sizeof(elfcpp::Elf_Word).
 127   const char* pnames = reinterpret_cast<const char*>(pword + nsyms);
 128
 129   this->armap_.resize(nsyms);
 130
 131   for (unsigned int i = 0; i < nsyms; ++i)
 132     {
 133       this->armap_[i].name = pnames;
 134       this->armap_[i].offset = elfcpp::Swap<32, true>::readval(pword);
 135       pnames += strlen(pnames) + 1;
 136       ++pword;
 137     }
 138
 139   if (reinterpret_cast<const unsigned char*>(pnames) - p > size)
 140     {
 141       fprintf(stderr, _("%s: %s: bad archive symbol table names\n"),
 142               program_name, this->name().c_str());
 143       gold_exit(false);
 144     }
 145
 146   // This array keeps track of which symbols are for archive elements
 147   // which we have already included in the link.
 148   this->armap_checked_.resize(nsyms);
 149 }
 150
 151 // Read the header of an archive member at OFF.  Fail if something
 152 // goes wrong.  Return the size of the member.  Set *PNAME to the name
 153 // of the member.
 154
 155 off_t
 156 Archive::read_header(off_t off, std::string* pname)
 157 {
 158   const unsigned char* p = this->get_view(off, sizeof(Archive_header));
 159   const Archive_header* hdr = reinterpret_cast<const Archive_header*>(p);
 160   return this->interpret_header(hdr, off,  pname);
 161 }
 162
 163 // Interpret the header of HDR, the header of the archive member at
 164 // file offset OFF.  Fail if something goes wrong.  Return the size of
 165 // the member.  Set *PNAME to the name of the member.
 166
 167 off_t
 168 Archive::interpret_header(const Archive_header* hdr, off_t off,
 169                           std::string* pname)
 170 {
 171   if (memcmp(hdr->ar_fmag, arfmag, sizeof arfmag) != 0)
 172     {
 173       fprintf(stderr, _("%s; %s: malformed archive header at %ld\n"),
 174               program_name, this->name().c_str(),
 175               static_cast<long>(off));
 176       gold_exit(false);
 177     }
 178
 179   const int size_string_size = sizeof hdr->ar_size;
 180   char size_string[size_string_size + 1];
 181   memcpy(size_string, hdr->ar_size, size_string_size);
 182   char* ps = size_string + size_string_size;
 183   while (ps[-1] == ' ')
 184     --ps;
 185   *ps = '\0';
 186
 187   errno = 0;
 188   char* end;
 189   off_t member_size = strtol(size_string, &end, 10);
 190   if (*end != '\0'
 191       || member_size < 0
 192       || (member_size == LONG_MAX && errno == ERANGE))
 193     {
 194       fprintf(stderr, _("%s: %s: malformed archive header size at %ld\n"),
 195               program_name, this->name().c_str(),
 196               static_cast<long>(off));
 197       gold_exit(false);
 198     }
 199
 200   if (hdr->ar_name[0] != '/')
 201     {
 202       const char* name_end = strchr(hdr->ar_name, '/');
 203       if (name_end == NULL
 204           || name_end - hdr->ar_name >= static_cast<int>(sizeof hdr->ar_name))
 205         {
 206           fprintf(stderr, _("%s: %s: malformed archive header name at %ld\n"),
 207                   program_name, this->name().c_str(),
 208                   static_cast<long>(off));
 209           gold_exit(false);
 210         }
 211       pname->assign(hdr->ar_name, name_end - hdr->ar_name);
 212     }
 213   else if (hdr->ar_name[1] == ' ')
 214     {
 215       // This is the symbol table.
 216       pname->clear();
 217     }
 218   else if (hdr->ar_name[1] == '/')
 219     {
 220       // This is the extended name table.
 221       pname->assign(1, '/');
 222     }
 223   else
 224     {
 225       errno = 0;
 226       long x = strtol(hdr->ar_name + 1, &end, 10);
 227       if (*end != ' '
 228           || x < 0
 229           || (x == LONG_MAX && errno == ERANGE)
 230           || static_cast<size_t>(x) >= this->extended_names_.size())
 231         {
 232           fprintf(stderr, _("%s: %s: bad extended name index at %ld\n"),
 233                   program_name, this->name().c_str(),
 234                   static_cast<long>(off));
 235           gold_exit(false);
 236         }
 237
 238       const char* name = this->extended_names_.data() + x;
 239       const char* name_end = strchr(name, '/');
 240       if (static_cast<size_t>(name_end - name) > this->extended_names_.size()
 241           || name_end[1] != '\n')
 242         {
 243           fprintf(stderr, _("%s: %s: bad extended name entry at header %ld\n"),
 244                   program_name, this->name().c_str(),
 245                   static_cast<long>(off));
 246           gold_exit(false);
 247         }
 248       pname->assign(name, name_end - name);
 249     }
 250
 251   return member_size;
 252 }
 253
 254 // Select members from the archive and add them to the link.  We walk
 255 // through the elements in the archive map, and look each one up in
 256 // the symbol table.  If it exists as a strong undefined symbol, we
 257 // pull in the corresponding element.  We have to do this in a loop,
 258 // since pulling in one element may create new undefined symbols which
 259 // may be satisfied by other objects in the archive.
 260
 261 void
 262 Archive::add_symbols(Symbol_table* symtab, Layout* layout,
 263                      Input_objects* input_objects)
 264 {
 265   if (this->input_file_->options().include_whole_archive())
 266     return this->include_all_members(symtab, layout, input_objects);
 267
 268   const size_t armap_size = this->armap_.size();
 269
 270   // This is a quick optimization, since we usually see many symbols
 271   // in a row with the same offset.  last_seen_offset holds the last
 272   // offset we saw that was present in the seen_offsets_ set.
 273   off_t last_seen_offset = -1;
 274
 275   // Track which symbols in the symbol table we've already found to be
 276   // defined.
 277
 278   bool added_new_object;
 279   do
 280     {
 281       added_new_object = false;
 282       for (size_t i = 0; i < armap_size; ++i)
 283         {
 284           if (this->armap_checked_[i])
 285             continue;
 286           if (this->armap_[i].offset == last_seen_offset)
 287             {
 288               this->armap_checked_[i] = true;
 289               continue;
 290             }
 291           if (this->seen_offsets_.find(this->armap_[i].offset)
 292               != this->seen_offsets_.end())
 293             {
 294               this->armap_checked_[i] = true;
 295               last_seen_offset = this->armap_[i].offset;
 296               continue;
 297             }
 298
 299           Symbol* sym = symtab->lookup(this->armap_[i].name);
 300           if (sym == NULL)
 301             continue;
 302           else if (!sym->is_undefined())
 303             {
 304               this->armap_checked_[i] = true;
 305               continue;
 306             }
 307           else if (sym->binding() == elfcpp::STB_WEAK)
 308             continue;
 309
 310           // We want to include this object in the link.
 311           last_seen_offset = this->armap_[i].offset;
 312           this->seen_offsets_.insert(last_seen_offset);
 313           this->armap_checked_[i] = true;
 314           this->include_member(symtab, layout, input_objects,
 315                                last_seen_offset);
 316           added_new_object = true;
 317         }
 318     }
 319   while (added_new_object);
 320 }
 321
 322 // Include all the archive members in the link.  This is for --whole-archive.
 323
 324 void
 325 Archive::include_all_members(Symbol_table* symtab, Layout* layout,
 326                              Input_objects* input_objects)
 327 {
 328   off_t off = sarmag;
 329   while (true)
 330     {
 331       off_t bytes;
 332       const unsigned char* p = this->get_view(off, sizeof(Archive_header),
 333                                               &bytes);
 334       if (bytes < sizeof(Archive_header))
 335         {
 336           if (bytes != 0)
 337             {
 338               fprintf(stderr, _("%s: %s: short archive header at %ld\n"),
 339                       program_name, this->name().c_str(),
 340                       static_cast<long>(off));
 341               gold_exit(false);
 342             }
 343
 344           break;
 345         }
 346
 347       const Archive_header* hdr = reinterpret_cast<const Archive_header*>(p);
 348       std::string name;
 349       off_t size = this->interpret_header(hdr, off, &name);
 350       if (name.empty())
 351         {
 352           // Symbol table.
 353         }
 354       else if (name == "/")
 355         {
 356           // Extended name table.
 357         }
 358       else
 359         this->include_member(symtab, layout, input_objects, off);
 360
 361       off += sizeof(Archive_header) + size;
 362       if ((off & 1) != 0)
 363         ++off;
 364     }
 365 }
 366
 367 // Include an archive member in the link.  OFF is the file offset of
 368 // the member header.
 369
 370 void
 371 Archive::include_member(Symbol_table* symtab, Layout* layout,
 372                         Input_objects* input_objects, off_t off)
 373 {
 374   std::string n;
 375   this->read_header(off, &n);
 376
 377   size_t memoff = off + sizeof(Archive_header);
 378
 379   // Read enough of the file to pick up the entire ELF header.
 380   int ehdr_size = elfcpp::Elf_sizes<64>::ehdr_size;
 381   off_t bytes;
 382   const unsigned char* p = this->input_file_->file().get_view(memoff,
 383                                                               ehdr_size,
 384                                                               &bytes);
 385   if (bytes < 4)
 386     {
 387       fprintf(stderr, _("%s: %s: member at %ld is not an ELF object"),
 388               program_name, this->name().c_str(),
 389               static_cast<long>(off));
 390       gold_exit(false);
 391     }
 392
 393   static unsigned char elfmagic[4] =
 394     {
 395       elfcpp::ELFMAG0, elfcpp::ELFMAG1,
 396       elfcpp::ELFMAG2, elfcpp::ELFMAG3
 397     };
 398   if (memcmp(p, elfmagic, 4) != 0)
 399     {
 400       fprintf(stderr, _("%s: %s: member at %ld is not an ELF object"),
 401               program_name, this->name().c_str(),
 402               static_cast<long>(off));
 403       gold_exit(false);
 404     }
 405
 406   Object* obj = make_elf_object((std::string(this->input_file_->filename())
 407                                  + "(" + n + ")"),
 408                                 this->input_file_, memoff, p, bytes);
 409
 410   input_objects->add_object(obj);
 411
 412   Read_symbols_data sd;
 413   obj->read_symbols(&sd);
 414   obj->layout(symtab, layout, &sd);
 415   obj->add_symbols(symtab, &sd);
 416 }
 417
 418 // Add_archive_symbols methods.
 419
 420 Add_archive_symbols::~Add_archive_symbols()
 421 {
 422   if (this->this_blocker_ != NULL)
 423     delete this->this_blocker_;
 424   // next_blocker_ is deleted by the task associated with the next
 425   // input file.
 426 }
 427
 428 // Return whether we can add the archive symbols.  We are blocked by
 429 // this_blocker_.  We block next_blocker_.  We also lock the file.
 430
 431 Task::Is_runnable_type
 432 Add_archive_symbols::is_runnable(Workqueue*)
 433 {
 434   if (this->this_blocker_ != NULL && this->this_blocker_->is_blocked())
 435     return IS_BLOCKED;
 436   return IS_RUNNABLE;
 437 }
 438
 439 class Add_archive_symbols::Add_archive_symbols_locker : public Task_locker
 440 {
 441  public:
 442   Add_archive_symbols_locker(Task_token& token, Workqueue* workqueue,
 443                              File_read& file)
 444     : blocker_(token, workqueue), filelock_(file)
 445   { }
 446
 447  private:
 448   Task_locker_block blocker_;
 449   Task_locker_obj<File_read> filelock_;
 450 };
 451
 452 Task_locker*
 453 Add_archive_symbols::locks(Workqueue* workqueue)
 454 {
 455   return new Add_archive_symbols_locker(*this->next_blocker_,
 456                                         workqueue,
 457                                         this->archive_->file());
 458 }
 459
 460 void
 461 Add_archive_symbols::run(Workqueue*)
 462 {
 463   this->archive_->add_symbols(this->symtab_, this->layout_,
 464                               this->input_objects_);
 465
 466   if (this->input_group_ != NULL)
 467     this->input_group_->add_archive(this->archive_);
 468   else
 469     {
 470       // We no longer need to know about this archive.
 471       delete this->archive_;
 472     }
 473 }
 474
 475 } // End namespace gold.