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(Task* task)
  77 {
  78   // We need to ignore empty archives.
  79   if (this->input_file_->file().filesize() == sarmag)
  80     {
  81       this->input_file_->file().unlock(task);
  82       return;
  83     }
  84
  85   // The first member of the archive should be the symbol table.
  86   std::string armap_name;
  87   off_t armap_size = this->read_header(sarmag, &armap_name);
  88   off_t off = sarmag;
  89   if (armap_name.empty())
  90     {
  91       this->read_armap(sarmag + sizeof(Archive_header), armap_size);
  92       off = sarmag + sizeof(Archive_header) + armap_size;
  93     }
  94   else if (!this->input_file_->options().include_whole_archive())
  95     gold_error(_("%s: no archive symbol table (run ranlib)"),
  96                this->name().c_str());
  97
  98   // See if there is an extended name table.
  99   if ((off & 1) != 0)
 100     ++off;
 101   std::string xname;
 102   off_t extended_size = this->read_header(off, &xname);
 103   if (xname == "/")
 104     {
 105       const unsigned char* p = this->get_view(off + sizeof(Archive_header),
 106                                               extended_size, false);
 107       const char* px = reinterpret_cast<const char*>(p);
 108       this->extended_names_.assign(px, extended_size);
 109     }
 110
 111   // Opening the file locked it.  Unlock it now.
 112   this->input_file_->file().unlock(task);
 113 }
 114
 115 // Read the archive symbol map.
 116
 117 void
 118 Archive::read_armap(off_t start, off_t size)
 119 {
 120   // Read in the entire armap.
 121   const unsigned char* p = this->get_view(start, size, false);
 122
 123   // Numbers in the armap are always big-endian.
 124   const elfcpp::Elf_Word* pword = reinterpret_cast<const elfcpp::Elf_Word*>(p);
 125   unsigned int nsyms = elfcpp::Swap<32, true>::readval(pword);
 126   ++pword;
 127
 128   // Note that the addition is in units of sizeof(elfcpp::Elf_Word).
 129   const char* pnames = reinterpret_cast<const char*>(pword + nsyms);
 130   off_t names_size = reinterpret_cast<const char*>(p) + size - pnames;
 131   this->armap_names_.assign(pnames, names_size);
 132
 133   this->armap_.resize(nsyms);
 134
 135   off_t name_offset = 0;
 136   for (unsigned int i = 0; i < nsyms; ++i)
 137     {
 138       this->armap_[i].name_offset = name_offset;
 139       this->armap_[i].file_offset = elfcpp::Swap<32, true>::readval(pword);
 140       name_offset += strlen(pnames + name_offset) + 1;
 141       ++pword;
 142     }
 143
 144   if (reinterpret_cast<const unsigned char*>(pnames) - p > size)
 145     gold_error(_("%s: bad archive symbol table names"),
 146                this->name().c_str());
 147
 148   // This array keeps track of which symbols are for archive elements
 149   // which we have already included in the link.
 150   this->armap_checked_.resize(nsyms);
 151 }
 152
 153 // Read the header of an archive member at OFF.  Fail if something
 154 // goes wrong.  Return the size of the member.  Set *PNAME to the name
 155 // of the member.
 156
 157 off_t
 158 Archive::read_header(off_t off, std::string* pname)
 159 {
 160   const unsigned char* p = this->get_view(off, sizeof(Archive_header), false);
 161   const Archive_header* hdr = reinterpret_cast<const Archive_header*>(p);
 162   return this->interpret_header(hdr, off,  pname);
 163 }
 164
 165 // Interpret the header of HDR, the header of the archive member at
 166 // file offset OFF.  Fail if something goes wrong.  Return the size of
 167 // the member.  Set *PNAME to the name of the member.
 168
 169 off_t
 170 Archive::interpret_header(const Archive_header* hdr, off_t off,
 171                           std::string* pname)
 172 {
 173   if (memcmp(hdr->ar_fmag, arfmag, sizeof arfmag) != 0)
 174     {
 175       gold_error(_("%s: malformed archive header at %zu"),
 176                  this->name().c_str(), static_cast<size_t>(off));
 177       return this->input_file_->file().filesize() - off;
 178     }
 179
 180   const int size_string_size = sizeof hdr->ar_size;
 181   char size_string[size_string_size + 1];
 182   memcpy(size_string, hdr->ar_size, size_string_size);
 183   char* ps = size_string + size_string_size;
 184   while (ps[-1] == ' ')
 185     --ps;
 186   *ps = '\0';
 187
 188   errno = 0;
 189   char* end;
 190   off_t member_size = strtol(size_string, &end, 10);
 191   if (*end != '\0'
 192       || member_size < 0
 193       || (member_size == LONG_MAX && errno == ERANGE))
 194     {
 195       gold_error(_("%s: malformed archive header size at %zu"),
 196                  this->name().c_str(), static_cast<size_t>(off));
 197       return this->input_file_->file().filesize() - off;
 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           gold_error(_("%s: malformed archive header name at %zu"),
 207                      this->name().c_str(), static_cast<size_t>(off));
 208           return this->input_file_->file().filesize() - off;
 209         }
 210       pname->assign(hdr->ar_name, name_end - hdr->ar_name);
 211     }
 212   else if (hdr->ar_name[1] == ' ')
 213     {
 214       // This is the symbol table.
 215       pname->clear();
 216     }
 217   else if (hdr->ar_name[1] == '/')
 218     {
 219       // This is the extended name table.
 220       pname->assign(1, '/');
 221     }
 222   else
 223     {
 224       errno = 0;
 225       long x = strtol(hdr->ar_name + 1, &end, 10);
 226       if (*end != ' '
 227           || x < 0
 228           || (x == LONG_MAX && errno == ERANGE)
 229           || static_cast<size_t>(x) >= this->extended_names_.size())
 230         {
 231           gold_error(_("%s: bad extended name index at %zu"),
 232                      this->name().c_str(), static_cast<size_t>(off));
 233           return this->input_file_->file().filesize() - off;
 234         }
 235
 236       const char* name = this->extended_names_.data() + x;
 237       const char* name_end = strchr(name, '/');
 238       if (static_cast<size_t>(name_end - name) > this->extended_names_.size()
 239           || name_end[1] != '\n')
 240         {
 241           gold_error(_("%s: bad extended name entry at header %zu"),
 242                      this->name().c_str(), static_cast<size_t>(off));
 243           return this->input_file_->file().filesize() - off;
 244         }
 245       pname->assign(name, name_end - name);
 246     }
 247
 248   return member_size;
 249 }
 250
 251 // Select members from the archive and add them to the link.  We walk
 252 // through the elements in the archive map, and look each one up in
 253 // the symbol table.  If it exists as a strong undefined symbol, we
 254 // pull in the corresponding element.  We have to do this in a loop,
 255 // since pulling in one element may create new undefined symbols which
 256 // may be satisfied by other objects in the archive.
 257
 258 void
 259 Archive::add_symbols(Symbol_table* symtab, Layout* layout,
 260                      Input_objects* input_objects)
 261 {
 262   if (this->input_file_->options().include_whole_archive())
 263     return this->include_all_members(symtab, layout, input_objects);
 264
 265   const size_t armap_size = this->armap_.size();
 266
 267   // This is a quick optimization, since we usually see many symbols
 268   // in a row with the same offset.  last_seen_offset holds the last
 269   // offset we saw that was present in the seen_offsets_ set.
 270   off_t last_seen_offset = -1;
 271
 272   // Track which symbols in the symbol table we've already found to be
 273   // defined.
 274
 275   bool added_new_object;
 276   do
 277     {
 278       added_new_object = false;
 279       for (size_t i = 0; i < armap_size; ++i)
 280         {
 281           if (this->armap_checked_[i])
 282             continue;
 283           if (this->armap_[i].file_offset == last_seen_offset)
 284             {
 285               this->armap_checked_[i] = true;
 286               continue;
 287             }
 288           if (this->seen_offsets_.find(this->armap_[i].file_offset)
 289               != this->seen_offsets_.end())
 290             {
 291               this->armap_checked_[i] = true;
 292               last_seen_offset = this->armap_[i].file_offset;
 293               continue;
 294             }
 295
 296           const char* sym_name = (this->armap_names_.data()
 297                                   + this->armap_[i].name_offset);
 298           Symbol* sym = symtab->lookup(sym_name);
 299           if (sym == NULL)
 300             continue;
 301           else if (!sym->is_undefined())
 302             {
 303               this->armap_checked_[i] = true;
 304               continue;
 305             }
 306           else if (sym->binding() == elfcpp::STB_WEAK)
 307             continue;
 308
 309           // We want to include this object in the link.
 310           last_seen_offset = this->armap_[i].file_offset;
 311           this->seen_offsets_.insert(last_seen_offset);
 312           this->armap_checked_[i] = true;
 313           this->include_member(symtab, layout, input_objects,
 314                                last_seen_offset);
 315           added_new_object = true;
 316         }
 317     }
 318   while (added_new_object);
 319 }
 320
 321 // Include all the archive members in the link.  This is for --whole-archive.
 322
 323 void
 324 Archive::include_all_members(Symbol_table* symtab, Layout* layout,
 325                              Input_objects* input_objects)
 326 {
 327   off_t off = sarmag;
 328   off_t filesize = this->input_file_->file().filesize();
 329   while (true)
 330     {
 331       if (filesize - off < static_cast<off_t>(sizeof(Archive_header)))
 332         {
 333           if (filesize != off)
 334             gold_error(_("%s: short archive header at %zu"),
 335                        this->name().c_str(), static_cast<size_t>(off));
 336           break;
 337         }
 338
 339       unsigned char hdr_buf[sizeof(Archive_header)];
 340       this->input_file_->file().read(off, sizeof(Archive_header), hdr_buf);
 341
 342       const Archive_header* hdr =
 343         reinterpret_cast<const Archive_header*>(hdr_buf);
 344       std::string name;
 345       off_t size = this->interpret_header(hdr, off, &name);
 346       if (name.empty())
 347         {
 348           // Symbol table.
 349         }
 350       else if (name == "/")
 351         {
 352           // Extended name table.
 353         }
 354       else
 355         this->include_member(symtab, layout, input_objects, off);
 356
 357       off += sizeof(Archive_header) + size;
 358       if ((off & 1) != 0)
 359         ++off;
 360     }
 361 }
 362
 363 // Include an archive member in the link.  OFF is the file offset of
 364 // the member header.
 365
 366 void
 367 Archive::include_member(Symbol_table* symtab, Layout* layout,
 368                         Input_objects* input_objects, off_t off)
 369 {
 370   std::string n;
 371   this->read_header(off, &n);
 372
 373   const off_t memoff = off + static_cast<off_t>(sizeof(Archive_header));
 374
 375   // Read enough of the file to pick up the entire ELF header.
 376   unsigned char ehdr_buf[elfcpp::Elf_sizes<64>::ehdr_size];
 377
 378   off_t filesize = this->input_file_->file().filesize();
 379   int read_size = elfcpp::Elf_sizes<64>::ehdr_size;
 380   if (filesize - memoff < read_size)
 381     read_size = filesize - memoff;
 382
 383   if (read_size < 4)
 384     {
 385       gold_error(_("%s: member at %zu is not an ELF object"),
 386                  this->name().c_str(), static_cast<size_t>(off));
 387       return;
 388     }
 389
 390   this->input_file_->file().read(memoff, read_size, ehdr_buf);
 391
 392   static unsigned char elfmagic[4] =
 393     {
 394       elfcpp::ELFMAG0, elfcpp::ELFMAG1,
 395       elfcpp::ELFMAG2, elfcpp::ELFMAG3
 396     };
 397   if (memcmp(ehdr_buf, elfmagic, 4) != 0)
 398     {
 399       gold_error(_("%s: member at %zu is not an ELF object"),
 400                  this->name().c_str(), static_cast<size_t>(off));
 401       return;
 402     }
 403
 404   Object* obj = make_elf_object((std::string(this->input_file_->filename())
 405                                  + "(" + n + ")"),
 406                                 this->input_file_, memoff, ehdr_buf,
 407                                 read_size);
 408
 409   if (input_objects->add_object(obj))
 410     {
 411       Read_symbols_data sd;
 412       obj->read_symbols(&sd);
 413       obj->layout(symtab, layout, &sd);
 414       obj->add_symbols(symtab, &sd);
 415     }
 416   else
 417     {
 418       // FIXME: We need to close the descriptor here.
 419       delete obj;
 420     }
 421
 422 }
 423
 424 // Add_archive_symbols methods.
 425
 426 Add_archive_symbols::~Add_archive_symbols()
 427 {
 428   if (this->this_blocker_ != NULL)
 429     delete this->this_blocker_;
 430   // next_blocker_ is deleted by the task associated with the next
 431   // input file.
 432 }
 433
 434 // Return whether we can add the archive symbols.  We are blocked by
 435 // this_blocker_.  We block next_blocker_.  We also lock the file.
 436
 437 Task_token*
 438 Add_archive_symbols::is_runnable()
 439 {
 440   if (this->this_blocker_ != NULL && this->this_blocker_->is_blocked())
 441     return this->this_blocker_;
 442   return NULL;
 443 }
 444
 445 void
 446 Add_archive_symbols::locks(Task_locker* tl)
 447 {
 448   tl->add(this, this->next_blocker_);
 449   tl->add(this, this->archive_->token());
 450 }
 451
 452 void
 453 Add_archive_symbols::run(Workqueue*)
 454 {
 455   this->archive_->add_symbols(this->symtab_, this->layout_,
 456                               this->input_objects_);
 457
 458   this->archive_->release();
 459
 460   if (this->input_group_ != NULL)
 461     this->input_group_->add_archive(this->archive_);
 462   else
 463     {
 464       // We no longer need to know about this archive.
 465       delete this->archive_;
 466       this->archive_ = NULL;
 467     }
 468 }
 469
 470 } // End namespace gold.