gold/archive.cc

   1 // archive.cc -- archive support for gold
   2
   3 // Copyright 2006, 2007, 2008 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 #include "libiberty.h"
  30 #include "filenames.h"
  31
  32 #include "elfcpp.h"
  33 #include "options.h"
  34 #include "fileread.h"
  35 #include "readsyms.h"
  36 #include "symtab.h"
  37 #include "object.h"
  38 #include "archive.h"
  39
  40 namespace gold
  41 {
  42
  43 // The header of an entry in the archive.  This is all readable text,
  44 // padded with spaces where necesary.  If the contents of an archive
  45 // are all text file, the entire archive is readable.
  46
  47 struct Archive::Archive_header
  48 {
  49   // The entry name.
  50   char ar_name[16];
  51   // The file modification time.
  52   char ar_date[12];
  53   // The user's UID in decimal.
  54   char ar_uid[6];
  55   // The user's GID in decimal.
  56   char ar_gid[6];
  57   // The file mode in octal.
  58   char ar_mode[8];
  59   // The file size in decimal.
  60   char ar_size[10];
  61   // The final magic code.
  62   char ar_fmag[2];
  63 };
  64
  65 // Archive methods.
  66
  67 const char Archive::armag[sarmag] =
  68 {
  69   '!', '<', 'a', 'r', 'c', 'h', '>', '\n'
  70 };
  71
  72 const char Archive::armagt[sarmag] =
  73 {
  74   '!', '<', 't', 'h', 'i', 'n', '>', '\n'
  75 };
  76
  77 const char Archive::arfmag[2] = { '`', '\n' };
  78
  79 // Set up the archive: read the symbol map and the extended name
  80 // table.
  81
  82 void
  83 Archive::setup()
  84 {
  85   // We need to ignore empty archives.
  86   if (this->input_file_->file().filesize() == sarmag)
  87     return;
  88
  89   // The first member of the archive should be the symbol table.
  90   std::string armap_name;
  91   section_size_type armap_size =
  92     convert_to_section_size_type(this->read_header(sarmag, false,
  93                                                    &armap_name, NULL));
  94   off_t off = sarmag;
  95   if (armap_name.empty())
  96     {
  97       this->read_armap(sarmag + sizeof(Archive_header), armap_size);
  98       off = sarmag + sizeof(Archive_header) + armap_size;
  99     }
 100   else if (!this->input_file_->options().whole_archive())
 101     gold_error(_("%s: no archive symbol table (run ranlib)"),
 102                this->name().c_str());
 103
 104   // See if there is an extended name table.  We cache these views
 105   // because it is likely that we will want to read the following
 106   // header in the add_symbols routine.
 107   if ((off & 1) != 0)
 108     ++off;
 109   std::string xname;
 110   section_size_type extended_size =
 111     convert_to_section_size_type(this->read_header(off, true, &xname, NULL));
 112   if (xname == "/")
 113     {
 114       const unsigned char* p = this->get_view(off + sizeof(Archive_header),
 115                                               extended_size, true);
 116       const char* px = reinterpret_cast<const char*>(p);
 117       this->extended_names_.assign(px, extended_size);
 118     }
 119 }
 120
 121 // Unlock any nested archives.
 122
 123 void
 124 Archive::unlock_nested_archives()
 125 {
 126   for (Nested_archive_table::iterator p = this->nested_archives_.begin();
 127        p != this->nested_archives_.end();
 128        ++p)
 129     {
 130       p->second->unlock(this->task_);
 131     }
 132 }
 133
 134 // Read the archive symbol map.
 135
 136 void
 137 Archive::read_armap(off_t start, section_size_type size)
 138 {
 139   // Read in the entire armap.
 140   const unsigned char* p = this->get_view(start, size, false);
 141
 142   // Numbers in the armap are always big-endian.
 143   const elfcpp::Elf_Word* pword = reinterpret_cast<const elfcpp::Elf_Word*>(p);
 144   unsigned int nsyms = elfcpp::Swap<32, true>::readval(pword);
 145   ++pword;
 146
 147   // Note that the addition is in units of sizeof(elfcpp::Elf_Word).
 148   const char* pnames = reinterpret_cast<const char*>(pword + nsyms);
 149   section_size_type names_size =
 150     reinterpret_cast<const char*>(p) + size - pnames;
 151   this->armap_names_.assign(pnames, names_size);
 152
 153   this->armap_.resize(nsyms);
 154
 155   section_offset_type name_offset = 0;
 156   for (unsigned int i = 0; i < nsyms; ++i)
 157     {
 158       this->armap_[i].name_offset = name_offset;
 159       this->armap_[i].file_offset = elfcpp::Swap<32, true>::readval(pword);
 160       name_offset += strlen(pnames + name_offset) + 1;
 161       ++pword;
 162     }
 163
 164   if (static_cast<section_size_type>(name_offset) > names_size)
 165     gold_error(_("%s: bad archive symbol table names"),
 166                this->name().c_str());
 167
 168   // This array keeps track of which symbols are for archive elements
 169   // which we have already included in the link.
 170   this->armap_checked_.resize(nsyms);
 171 }
 172
 173 // Read the header of an archive member at OFF.  Fail if something
 174 // goes wrong.  Return the size of the member.  Set *PNAME to the name
 175 // of the member.
 176
 177 off_t
 178 Archive::read_header(off_t off, bool cache, std::string* pname,
 179                      off_t* nested_off)
 180 {
 181   const unsigned char* p = this->get_view(off, sizeof(Archive_header), cache);
 182   const Archive_header* hdr = reinterpret_cast<const Archive_header*>(p);
 183   return this->interpret_header(hdr, off,  pname, nested_off);
 184 }
 185
 186 // Interpret the header of HDR, the header of the archive member at
 187 // file offset OFF.  Fail if something goes wrong.  Return the size of
 188 // the member.  Set *PNAME to the name of the member.
 189
 190 off_t
 191 Archive::interpret_header(const Archive_header* hdr, off_t off,
 192                           std::string* pname, off_t* nested_off)
 193 {
 194   if (memcmp(hdr->ar_fmag, arfmag, sizeof arfmag) != 0)
 195     {
 196       gold_error(_("%s: malformed archive header at %zu"),
 197                  this->name().c_str(), static_cast<size_t>(off));
 198       return this->input_file_->file().filesize() - off;
 199     }
 200
 201   const int size_string_size = sizeof hdr->ar_size;
 202   char size_string[size_string_size + 1];
 203   memcpy(size_string, hdr->ar_size, size_string_size);
 204   char* ps = size_string + size_string_size;
 205   while (ps[-1] == ' ')
 206     --ps;
 207   *ps = '\0';
 208
 209   errno = 0;
 210   char* end;
 211   off_t member_size = strtol(size_string, &end, 10);
 212   if (*end != '\0'
 213       || member_size < 0
 214       || (member_size == LONG_MAX && errno == ERANGE))
 215     {
 216       gold_error(_("%s: malformed archive header size at %zu"),
 217                  this->name().c_str(), static_cast<size_t>(off));
 218       return this->input_file_->file().filesize() - off;
 219     }
 220
 221   if (hdr->ar_name[0] != '/')
 222     {
 223       const char* name_end = strchr(hdr->ar_name, '/');
 224       if (name_end == NULL
 225           || name_end - hdr->ar_name >= static_cast<int>(sizeof hdr->ar_name))
 226         {
 227           gold_error(_("%s: malformed archive header name at %zu"),
 228                      this->name().c_str(), static_cast<size_t>(off));
 229           return this->input_file_->file().filesize() - off;
 230         }
 231       pname->assign(hdr->ar_name, name_end - hdr->ar_name);
 232       if (nested_off != NULL)
 233         *nested_off = 0;
 234     }
 235   else if (hdr->ar_name[1] == ' ')
 236     {
 237       // This is the symbol table.
 238       pname->clear();
 239     }
 240   else if (hdr->ar_name[1] == '/')
 241     {
 242       // This is the extended name table.
 243       pname->assign(1, '/');
 244     }
 245   else
 246     {
 247       errno = 0;
 248       long x = strtol(hdr->ar_name + 1, &end, 10);
 249       long y = 0;
 250       if (*end == ':')
 251         y = strtol(end + 1, &end, 10);
 252       if (*end != ' '
 253           || x < 0
 254           || (x == LONG_MAX && errno == ERANGE)
 255           || static_cast<size_t>(x) >= this->extended_names_.size())
 256         {
 257           gold_error(_("%s: bad extended name index at %zu"),
 258                      this->name().c_str(), static_cast<size_t>(off));
 259           return this->input_file_->file().filesize() - off;
 260         }
 261
 262       const char* name = this->extended_names_.data() + x;
 263       const char* name_end = strchr(name, '\n');
 264       if (static_cast<size_t>(name_end - name) > this->extended_names_.size()
 265           || name_end[-1] != '/')
 266         {
 267           gold_error(_("%s: bad extended name entry at header %zu"),
 268                      this->name().c_str(), static_cast<size_t>(off));
 269           return this->input_file_->file().filesize() - off;
 270         }
 271       pname->assign(name, name_end - 1 - name);
 272       if (nested_off != NULL)
 273         *nested_off = y;
 274     }
 275
 276   return member_size;
 277 }
 278
 279 // Select members from the archive and add them to the link.  We walk
 280 // through the elements in the archive map, and look each one up in
 281 // the symbol table.  If it exists as a strong undefined symbol, we
 282 // pull in the corresponding element.  We have to do this in a loop,
 283 // since pulling in one element may create new undefined symbols which
 284 // may be satisfied by other objects in the archive.
 285
 286 void
 287 Archive::add_symbols(Symbol_table* symtab, Layout* layout,
 288                      Input_objects* input_objects)
 289 {
 290   if (this->input_file_->options().whole_archive())
 291     return this->include_all_members(symtab, layout, input_objects);
 292
 293   const size_t armap_size = this->armap_.size();
 294
 295   // This is a quick optimization, since we usually see many symbols
 296   // in a row with the same offset.  last_seen_offset holds the last
 297   // offset we saw that was present in the seen_offsets_ set.
 298   off_t last_seen_offset = -1;
 299
 300   // Track which symbols in the symbol table we've already found to be
 301   // defined.
 302
 303   bool added_new_object;
 304   do
 305     {
 306       added_new_object = false;
 307       for (size_t i = 0; i < armap_size; ++i)
 308         {
 309           if (this->armap_checked_[i])
 310             continue;
 311           if (this->armap_[i].file_offset == last_seen_offset)
 312             {
 313               this->armap_checked_[i] = true;
 314               continue;
 315             }
 316           if (this->seen_offsets_.find(this->armap_[i].file_offset)
 317               != this->seen_offsets_.end())
 318             {
 319               this->armap_checked_[i] = true;
 320               last_seen_offset = this->armap_[i].file_offset;
 321               continue;
 322             }
 323
 324           const char* sym_name = (this->armap_names_.data()
 325                                   + this->armap_[i].name_offset);
 326           Symbol* sym = symtab->lookup(sym_name);
 327           if (sym == NULL)
 328             continue;
 329           else if (!sym->is_undefined())
 330             {
 331               this->armap_checked_[i] = true;
 332               continue;
 333             }
 334           else if (sym->binding() == elfcpp::STB_WEAK)
 335             continue;
 336
 337           // We want to include this object in the link.
 338           last_seen_offset = this->armap_[i].file_offset;
 339           this->seen_offsets_.insert(last_seen_offset);
 340           this->armap_checked_[i] = true;
 341           this->include_member(symtab, layout, input_objects,
 342                                last_seen_offset);
 343           added_new_object = true;
 344         }
 345     }
 346   while (added_new_object);
 347 }
 348
 349 // Include all the archive members in the link.  This is for --whole-archive.
 350
 351 void
 352 Archive::include_all_members(Symbol_table* symtab, Layout* layout,
 353                              Input_objects* input_objects)
 354 {
 355   off_t off = sarmag;
 356   off_t filesize = this->input_file_->file().filesize();
 357   while (true)
 358     {
 359       if (filesize - off < static_cast<off_t>(sizeof(Archive_header)))
 360         {
 361           if (filesize != off)
 362             gold_error(_("%s: short archive header at %zu"),
 363                        this->name().c_str(), static_cast<size_t>(off));
 364           break;
 365         }
 366
 367       unsigned char hdr_buf[sizeof(Archive_header)];
 368       this->input_file_->file().read(off, sizeof(Archive_header), hdr_buf);
 369
 370       const Archive_header* hdr =
 371         reinterpret_cast<const Archive_header*>(hdr_buf);
 372       std::string name;
 373       off_t size = this->interpret_header(hdr, off, &name, NULL);
 374       if (name.empty())
 375         {
 376           // Symbol table.
 377         }
 378       else if (name == "/")
 379         {
 380           // Extended name table.
 381         }
 382       else
 383         this->include_member(symtab, layout, input_objects, off);
 384
 385       off += sizeof(Archive_header);
 386       if (!this->is_thin_archive_)
 387         off += size;
 388       if ((off & 1) != 0)
 389         ++off;
 390     }
 391 }
 392
 393 // Include an archive member in the link.  OFF is the file offset of
 394 // the member header.
 395
 396 void
 397 Archive::include_member(Symbol_table* symtab, Layout* layout,
 398                         Input_objects* input_objects, off_t off)
 399 {
 400   std::string n;
 401   off_t nested_off;
 402   this->read_header(off, false, &n, &nested_off);
 403
 404   Input_file* input_file;
 405   off_t memoff;
 406
 407   if (!this->is_thin_archive_)
 408     {
 409       input_file = this->input_file_;
 410       memoff = off + static_cast<off_t>(sizeof(Archive_header));
 411     }
 412   else
 413     {
 414       // Adjust a relative pathname so that it is relative
 415       // to the directory containing the archive.
 416       if (!IS_ABSOLUTE_PATH(n.c_str()))
 417         {
 418           const char *arch_path = this->name().c_str();
 419           const char *basename = lbasename(arch_path);
 420           if (basename > arch_path)
 421             n.replace(0, 0, this->name().substr(0, basename - arch_path));
 422         }
 423       if (nested_off > 0)
 424         {
 425           // This is a member of a nested archive.  Open the containing
 426           // archive if we don't already have it open, then do a recursive
 427           // call to include the member from that archive.
 428           Archive* arch;
 429           Nested_archive_table::const_iterator p =
 430             this->nested_archives_.find(n);
 431           if (p != this->nested_archives_.end())
 432             arch = p->second;
 433           else
 434             {
 435               Input_file_argument* input_file_arg =
 436                 new Input_file_argument(n.c_str(), false, "", false,
 437                                         parameters->options());
 438               input_file = new Input_file(input_file_arg);
 439               if (!input_file->open(parameters->options(), *this->dirpath_,
 440                                     this->task_))
 441                 return;
 442               arch = new Archive(n, input_file, false, this->dirpath_,
 443                                  this->task_);
 444               arch->setup();
 445               std::pair<Nested_archive_table::iterator, bool> ins =
 446                 this->nested_archives_.insert(std::make_pair(n, arch));
 447               gold_assert(ins.second);
 448             }
 449           arch->include_member(symtab, layout, input_objects, nested_off);
 450           return;
 451         }
 452       // This is an external member of a thin archive.  Open the
 453       // file as a regular relocatable object file.
 454       Input_file_argument* input_file_arg =
 455           new Input_file_argument(n.c_str(), false, "", false,
 456                                   this->input_file_->options());
 457       input_file = new Input_file(input_file_arg);
 458       if (!input_file->open(parameters->options(), *this->dirpath_,
 459                             this->task_))
 460         {
 461           return;
 462         }
 463       memoff = 0;
 464     }
 465
 466   // Read enough of the file to pick up the entire ELF header.
 467   unsigned char ehdr_buf[elfcpp::Elf_sizes<64>::ehdr_size];
 468
 469   off_t filesize = input_file->file().filesize();
 470   int read_size = elfcpp::Elf_sizes<64>::ehdr_size;
 471   if (filesize - memoff < read_size)
 472     read_size = filesize - memoff;
 473
 474   if (read_size < 4)
 475     {
 476       gold_error(_("%s: member at %zu is not an ELF object"),
 477                  this->name().c_str(), static_cast<size_t>(off));
 478       return;
 479     }
 480
 481   input_file->file().read(memoff, read_size, ehdr_buf);
 482
 483   static unsigned char elfmagic[4] =
 484     {
 485       elfcpp::ELFMAG0, elfcpp::ELFMAG1,
 486       elfcpp::ELFMAG2, elfcpp::ELFMAG3
 487     };
 488   if (memcmp(ehdr_buf, elfmagic, 4) != 0)
 489     {
 490       gold_error(_("%s: member at %zu is not an ELF object"),
 491                  this->name().c_str(), static_cast<size_t>(off));
 492       return;
 493     }
 494
 495   Object* obj = make_elf_object((std::string(this->input_file_->filename())
 496                                  + "(" + n + ")"),
 497                                 input_file, memoff, ehdr_buf,
 498                                 read_size);
 499
 500   if (input_objects->add_object(obj))
 501     {
 502       Read_symbols_data sd;
 503       obj->read_symbols(&sd);
 504       obj->layout(symtab, layout, &sd);
 505       obj->add_symbols(symtab, &sd);
 506     }
 507   else
 508     {
 509       // FIXME: We need to close the descriptor here.
 510       delete obj;
 511     }
 512
 513   if (this->is_thin_archive_)
 514     {
 515       // Opening the file locked it.  Unlock it now.
 516       input_file->file().unlock(this->task_);
 517     }
 518 }
 519
 520 // Add_archive_symbols methods.
 521
 522 Add_archive_symbols::~Add_archive_symbols()
 523 {
 524   if (this->this_blocker_ != NULL)
 525     delete this->this_blocker_;
 526   // next_blocker_ is deleted by the task associated with the next
 527   // input file.
 528 }
 529
 530 // Return whether we can add the archive symbols.  We are blocked by
 531 // this_blocker_.  We block next_blocker_.  We also lock the file.
 532
 533 Task_token*
 534 Add_archive_symbols::is_runnable()
 535 {
 536   if (this->this_blocker_ != NULL && this->this_blocker_->is_blocked())
 537     return this->this_blocker_;
 538   return NULL;
 539 }
 540
 541 void
 542 Add_archive_symbols::locks(Task_locker* tl)
 543 {
 544   tl->add(this, this->next_blocker_);
 545   tl->add(this, this->archive_->token());
 546 }
 547
 548 void
 549 Add_archive_symbols::run(Workqueue*)
 550 {
 551   this->archive_->add_symbols(this->symtab_, this->layout_,
 552                               this->input_objects_);
 553
 554   this->archive_->unlock_nested_archives();
 555
 556   this->archive_->release();
 557
 558   if (this->input_group_ != NULL)
 559     this->input_group_->add_archive(this->archive_);
 560   else
 561     {
 562       // We no longer need to know about this archive.
 563       delete this->archive_;
 564       this->archive_ = NULL;
 565     }
 566 }
 567
 568 } // End namespace gold.