gold/archive.h

   1 // archive.h -- archive support for gold      -*- C++ -*-
   2
   3 // Copyright (C) 2006-2024 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 #ifndef GOLD_ARCHIVE_H
  24 #define GOLD_ARCHIVE_H
  25
  26 #include <string>
  27 #include <vector>
  28
  29 #include "fileread.h"
  30 #include "workqueue.h"
  31
  32 namespace gold
  33 {
  34
  35 class Task;
  36 class Input_argument;
  37 class Input_file;
  38 class Input_objects;
  39 class Input_group;
  40 class Layout;
  41 class Symbol_table;
  42 class Object;
  43 struct Read_symbols_data;
  44 class Input_file_lib;
  45 class Incremental_archive_entry;
  46
  47 // An entry in the archive map of offsets to members.
  48 struct Archive_member
  49 {
  50   Archive_member()
  51       : obj_(NULL), sd_(NULL), arg_serial_(0)
  52   { }
  53   Archive_member(Object* obj, Read_symbols_data* sd)
  54       : obj_(obj), sd_(sd), arg_serial_(0)
  55   { }
  56   // The object file.
  57   Object* obj_;
  58   // The data to pass from read_symbols() to add_symbols().
  59   Read_symbols_data* sd_;
  60   // The serial number of the file in the argument list.
  61   unsigned int arg_serial_;
  62 };
  63
  64 // This class serves as a base class for Archive and Lib_group objects.
  65
  66 class Library_base
  67 {
  68  public:
  69   Library_base(Task* task)
  70     : task_(task), incremental_info_(NULL)
  71   { }
  72
  73   virtual
  74   ~Library_base()
  75   { }
  76
  77   // The file name.
  78   const std::string&
  79   filename() const
  80   { return this->do_filename(); }
  81
  82   // The modification time of the archive file.
  83   Timespec
  84   get_mtime()
  85   { return this->do_get_mtime(); }
  86
  87   // When we see a symbol in an archive we might decide to include the member,
  88   // not include the member or be undecided. This enum represents these
  89   // possibilities.
  90
  91   enum Should_include
  92   {
  93     SHOULD_INCLUDE_NO,
  94     SHOULD_INCLUDE_YES,
  95     SHOULD_INCLUDE_UNKNOWN
  96   };
  97
  98   static Should_include
  99   should_include_member(Symbol_table* symtab, Layout*, const char* sym_name,
 100                         Symbol** symp, std::string* why, char** tmpbufp,
 101                         size_t* tmpbuflen);
 102
 103   // Store a pointer to the incremental link info for the library.
 104   void
 105   set_incremental_info(Incremental_archive_entry* info)
 106   { this->incremental_info_ = info; }
 107
 108   // Return the pointer to the incremental link info for the library.
 109   Incremental_archive_entry*
 110   incremental_info() const
 111   { return this->incremental_info_; }
 112
 113   // Abstract base class for processing unused symbols.
 114   class Symbol_visitor_base
 115   {
 116    public:
 117     Symbol_visitor_base()
 118     { }
 119
 120     virtual
 121     ~Symbol_visitor_base()
 122     { }
 123
 124     // This function will be called for each unused global
 125     // symbol in a library, with a pointer to the symbol name.
 126     virtual void
 127     visit(const char* /* name */) = 0;
 128   };
 129
 130   // Iterator for unused global symbols in the library.
 131   // Calls v->visit() for each global symbol defined
 132   // in each unused library member, passing a pointer to
 133   // the symbol name.
 134   void
 135   for_all_unused_symbols(Symbol_visitor_base* v) const
 136   { this->do_for_all_unused_symbols(v); }
 137
 138  protected:
 139   // The task reading this archive.
 140   Task *task_;
 141
 142  private:
 143   // The file name.
 144   virtual const std::string&
 145   do_filename() const = 0;
 146
 147   // Return the modification time of the archive file.
 148   virtual Timespec
 149   do_get_mtime() = 0;
 150
 151   // Iterator for unused global symbols in the library.
 152   virtual void
 153   do_for_all_unused_symbols(Symbol_visitor_base* v) const = 0;
 154
 155   // The incremental link information for this archive.
 156   Incremental_archive_entry* incremental_info_;
 157 };
 158
 159 // This class represents an archive--generally a libNAME.a file.
 160 // Archives have a symbol table and a list of objects.
 161
 162 class Archive : public Library_base
 163 {
 164  public:
 165   Archive(const std::string& name, Input_file* input_file,
 166           bool is_thin_archive, Dirsearch* dirpath, Task* task);
 167
 168   // The length of the magic string at the start of an archive.
 169   static const int sarmag = 8;
 170
 171   // The magic string at the start of an archive.
 172   static const char armag[sarmag];
 173   static const char armagt[sarmag];
 174
 175   // The string expected at the end of an archive member header.
 176   static const char arfmag[2];
 177
 178   // Name of 64-bit symbol table member.
 179   static const char sym64name[7];
 180
 181   // The name of the object.  This is the name used on the command
 182   // line; e.g., if "-lgcc" is on the command line, this will be
 183   // "gcc".
 184   const std::string&
 185   name() const
 186   { return this->name_; }
 187
 188   // The input file.
 189   const Input_file*
 190   input_file() const
 191   { return this->input_file_; }
 192
 193   // Set up the archive: read the symbol map.
 194   void
 195   setup();
 196
 197   // Get a reference to the underlying file.
 198   File_read&
 199   file()
 200   { return this->input_file_->file(); }
 201
 202   const File_read&
 203   file() const
 204   { return this->input_file_->file(); }
 205
 206   // Lock the underlying file.
 207   void
 208   lock(const Task* t)
 209   { this->input_file_->file().lock(t); }
 210
 211   // Unlock the underlying file.
 212   void
 213   unlock(const Task* t)
 214   { this->input_file_->file().unlock(t); }
 215
 216   // Return whether the underlying file is locked.
 217   bool
 218   is_locked() const
 219   { return this->input_file_->file().is_locked(); }
 220
 221   // Return the token, so that the task can be queued.
 222   Task_token*
 223   token()
 224   { return this->input_file_->file().token(); }
 225
 226   // Release the underlying file.
 227   void
 228   release()
 229   { this->input_file_->file().release(); }
 230
 231   // Clear uncached views in the underlying file.
 232   void
 233   clear_uncached_views()
 234   { this->input_file_->file().clear_uncached_views(); }
 235
 236   // Whether this is a thin archive.
 237   bool
 238   is_thin_archive() const
 239   { return this->is_thin_archive_; }
 240
 241   // Unlock any nested archives.
 242   void
 243   unlock_nested_archives();
 244
 245   // Select members from the archive as needed and add them to the
 246   // link.
 247   bool
 248   add_symbols(Symbol_table*, Layout*, Input_objects*, Mapfile*);
 249
 250   // Return whether the archive defines the symbol.
 251   bool
 252   defines_symbol(Symbol*) const;
 253
 254   // Dump statistical information to stderr.
 255   static void
 256   print_stats();
 257
 258   // Return the number of members in the archive.
 259   size_t
 260   count_members();
 261
 262   // Return the no-export flag.
 263   bool
 264   no_export()
 265   { return this->no_export_; }
 266
 267  private:
 268   Archive(const Archive&);
 269   Archive& operator=(const Archive&);
 270
 271   // The file name.
 272   const std::string&
 273   do_filename() const
 274   { return this->input_file_->filename(); }
 275
 276   // The modification time of the archive file.
 277   Timespec
 278   do_get_mtime()
 279   { return this->file().get_mtime(); }
 280
 281   struct Archive_header;
 282
 283   // Total number of archives seen.
 284   static unsigned int total_archives;
 285   // Total number of archive members seen.
 286   static unsigned int total_members;
 287   // Number of archive members loaded.
 288   static unsigned int total_members_loaded;
 289
 290   // Get a view into the underlying file.
 291   const unsigned char*
 292   get_view(off_t start, section_size_type size, bool aligned, bool cache)
 293   { return this->input_file_->file().get_view(0, start, size, aligned, cache); }
 294
 295   // Read the archive symbol map.
 296   template<int mapsize>
 297   void
 298   read_armap(off_t start, section_size_type size);
 299
 300   // Read an archive member header at OFF.  CACHE is whether to cache
 301   // the file view.  Return the size of the member, and set *PNAME to
 302   // the name.
 303   off_t
 304   read_header(off_t off, bool cache, std::string* pname, off_t* nested_off);
 305
 306   // Interpret an archive header HDR at OFF.  Return the size of the
 307   // member, and set *PNAME to the name.
 308   off_t
 309   interpret_header(const Archive_header* hdr, off_t off, std::string* pname,
 310                    off_t* nested_off) const;
 311
 312   // Get the file and offset for an archive member, which may be an
 313   // external member of a thin archive.  Set *INPUT_FILE to the
 314   // file containing the actual member, *MEMOFF to the offset
 315   // within that file (0 if not a nested archive), and *MEMBER_NAME
 316   // to the name of the archive member.  Return TRUE on success.
 317   bool
 318   get_file_and_offset(off_t off, Input_file** input_file, off_t* memoff,
 319                       off_t* memsize, std::string* member_name);
 320
 321   // Return an ELF object for the member at offset OFF.
 322   Object*
 323   get_elf_object_for_member(off_t off, bool*);
 324
 325   // Read the symbols from all the archive members in the link.
 326   void
 327   read_all_symbols();
 328
 329   // Read the symbols from an archive member in the link.  OFF is the file
 330   // offset of the member header.
 331   void
 332   read_symbols(off_t off);
 333
 334   // Include all the archive members in the link.
 335   bool
 336   include_all_members(Symbol_table*, Layout*, Input_objects*, Mapfile*);
 337
 338   // Include an archive member in the link.
 339   bool
 340   include_member(Symbol_table*, Layout*, Input_objects*, off_t off,
 341                  Mapfile*, Symbol*, const char* why);
 342
 343   // Return whether we found this archive by searching a directory.
 344   bool
 345   searched_for() const
 346   { return this->input_file_->will_search_for(); }
 347
 348   // Iterate over archive members.
 349   class const_iterator;
 350
 351   const_iterator
 352   begin();
 353
 354   const_iterator
 355   end();
 356
 357   friend class const_iterator;
 358
 359   // Iterator for unused global symbols in the library.
 360   void
 361   do_for_all_unused_symbols(Symbol_visitor_base* v) const;
 362
 363   // An entry in the archive map of symbols to object files.
 364   struct Armap_entry
 365   {
 366     // The offset to the symbol name in armap_names_.
 367     off_t name_offset;
 368     // The file offset to the object in the archive.
 369     off_t file_offset;
 370   };
 371
 372   // A simple hash code for off_t values.
 373   class Seen_hash
 374   {
 375    public:
 376     size_t operator()(off_t val) const
 377     { return static_cast<size_t>(val); }
 378   };
 379
 380   // For keeping track of open nested archives in a thin archive file.
 381   typedef Unordered_map<std::string, Archive*> Nested_archive_table;
 382
 383   // Name of object as printed to user.
 384   std::string name_;
 385   // For reading the file.
 386   Input_file* input_file_;
 387   // The archive map.
 388   std::vector<Armap_entry> armap_;
 389   // The names in the archive map.
 390   std::string armap_names_;
 391   // The extended name table.
 392   std::string extended_names_;
 393   // Track which symbols in the archive map are for elements which are
 394   // defined or which have already been included in the link.
 395   std::vector<bool> armap_checked_;
 396   // Track which elements have been included by offset.
 397   Unordered_set<off_t, Seen_hash> seen_offsets_;
 398   // Table of objects whose symbols have been pre-read.
 399   std::map<off_t, Archive_member> members_;
 400   // True if this is a thin archive.
 401   const bool is_thin_archive_;
 402   // True if we have included at least one object from this archive.
 403   bool included_member_;
 404   // Table of nested archives, indexed by filename.
 405   Nested_archive_table nested_archives_;
 406   // The directory search path.
 407   Dirsearch* dirpath_;
 408   // Number of members in this archive;
 409   unsigned int num_members_;
 410   // True if we exclude this library archive from automatic export.
 411   bool no_export_;
 412   // True if this library has been included as a --whole-archive.
 413   bool included_all_members_;
 414 };
 415
 416 // This class is used to read an archive and pick out the desired
 417 // elements and add them to the link.
 418
 419 class Add_archive_symbols : public Task
 420 {
 421  public:
 422   Add_archive_symbols(Symbol_table* symtab, Layout* layout,
 423                       Input_objects* input_objects, Dirsearch* dirpath,
 424                       int dirindex, Mapfile* mapfile,
 425                       const Input_argument* input_argument,
 426                       Archive* archive, Input_group* input_group,
 427                       Task_token* this_blocker,
 428                       Task_token* next_blocker)
 429     : symtab_(symtab), layout_(layout), input_objects_(input_objects),
 430       dirpath_(dirpath), dirindex_(dirindex), mapfile_(mapfile),
 431       input_argument_(input_argument), archive_(archive),
 432       input_group_(input_group), this_blocker_(this_blocker),
 433       next_blocker_(next_blocker)
 434   { }
 435
 436   ~Add_archive_symbols();
 437
 438   // The standard Task methods.
 439
 440   Task_token*
 441   is_runnable();
 442
 443   void
 444   locks(Task_locker*);
 445
 446   void
 447   run(Workqueue*);
 448
 449   std::string
 450   get_name() const
 451   {
 452     if (this->archive_ == NULL)
 453       return "Add_archive_symbols";
 454     return "Add_archive_symbols " + this->archive_->file().filename();
 455   }
 456
 457  private:
 458   Symbol_table* symtab_;
 459   Layout* layout_;
 460   Input_objects* input_objects_;
 461   Dirsearch* dirpath_;
 462   int dirindex_;
 463   Mapfile* mapfile_;
 464   const Input_argument* input_argument_;
 465   Archive* archive_;
 466   Input_group* input_group_;
 467   Task_token* this_blocker_;
 468   Task_token* next_blocker_;
 469 };
 470
 471 // This class represents the files surrounded by a --start-lib ... --end-lib.
 472
 473 class Lib_group : public Library_base
 474 {
 475  public:
 476   Lib_group(const Input_file_lib* lib, Task* task);
 477
 478   // Select members from the lib group as needed and add them to the link.
 479   void
 480   add_symbols(Symbol_table*, Layout*, Input_objects*);
 481
 482   // Include a member of the lib group in the link.
 483   void
 484   include_member(Symbol_table*, Layout*, Input_objects*, const Archive_member&);
 485
 486   Archive_member*
 487   get_member(int i)
 488   {
 489     return &this->members_[i];
 490   }
 491
 492   // Total number of archives seen.
 493   static unsigned int total_lib_groups;
 494   // Total number of archive members seen.
 495   static unsigned int total_members;
 496   // Number of archive members loaded.
 497   static unsigned int total_members_loaded;
 498
 499   // Dump statistical information to stderr.
 500   static void
 501   print_stats();
 502
 503  private:
 504   // The file name.
 505   const std::string&
 506   do_filename() const;
 507
 508   // A Lib_group does not have a modification time, since there is no
 509   // real library file.
 510   Timespec
 511   do_get_mtime()
 512   { return Timespec(0, 0); }
 513
 514   // Iterator for unused global symbols in the library.
 515   void
 516   do_for_all_unused_symbols(Symbol_visitor_base*) const;
 517
 518   // Table of the objects in the group.
 519   std::vector<Archive_member> members_;
 520 };
 521
 522 // This class is used to pick out the desired elements and add them to the link.
 523
 524 class Add_lib_group_symbols : public Task
 525 {
 526  public:
 527   Add_lib_group_symbols(Symbol_table* symtab, Layout* layout,
 528                         Input_objects* input_objects,
 529                         Lib_group* lib, Task_token* next_blocker)
 530       : symtab_(symtab), layout_(layout), input_objects_(input_objects),
 531         lib_(lib), readsyms_blocker_(NULL), this_blocker_(NULL),
 532         next_blocker_(next_blocker)
 533   { }
 534
 535   ~Add_lib_group_symbols();
 536
 537   // The standard Task methods.
 538
 539   Task_token*
 540   is_runnable();
 541
 542   void
 543   locks(Task_locker*);
 544
 545   void
 546   run(Workqueue*);
 547
 548   // Set the blocker to use for this task.
 549   void
 550   set_blocker(Task_token* readsyms_blocker, Task_token* this_blocker)
 551   {
 552     gold_assert(this->readsyms_blocker_ == NULL && this->this_blocker_ == NULL);
 553     this->readsyms_blocker_ = readsyms_blocker;
 554     this->this_blocker_ = this_blocker;
 555   }
 556
 557   std::string
 558   get_name() const
 559   {
 560     return "Add_lib_group_symbols";
 561   }
 562
 563  private:
 564   Symbol_table* symtab_;
 565   Layout* layout_;
 566   Input_objects* input_objects_;
 567   Lib_group* lib_;
 568   Task_token* readsyms_blocker_;
 569   Task_token* this_blocker_;
 570   Task_token* next_blocker_;
 571 };
 572
 573 } // End namespace gold.
 574
 575 #endif // !defined(GOLD_ARCHIVE_H)