gold/dwarf_reader.cc

   1 // dwarf_reader.cc -- parse dwarf2/3 debug information
   2
   3 // Copyright 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 "elfcpp_swap.h"
  26 #include "dwarf.h"
  27 #include "object.h"
  28 #include "reloc.h"
  29 #include "dwarf_reader.h"
  30
  31 namespace {
  32
  33 // Read an unsigned LEB128 number.  Each byte contains 7 bits of
  34 // information, plus one bit saying whether the number continues or
  35 // not.
  36
  37 uint64_t
  38 read_unsigned_LEB_128(const unsigned char* buffer, size_t* len)
  39 {
  40   uint64_t result = 0;
  41   size_t num_read = 0;
  42   unsigned int shift = 0;
  43   unsigned char byte;
  44
  45   do
  46     {
  47       byte = *buffer++;
  48       num_read++;
  49       result |= (static_cast<uint64_t>(byte & 0x7f)) << shift;
  50       shift += 7;
  51     }
  52   while (byte & 0x80);
  53
  54   *len = num_read;
  55
  56   return result;
  57 }
  58
  59 // Read a signed LEB128 number.  These are like regular LEB128
  60 // numbers, except the last byte may have a sign bit set.
  61
  62 int64_t
  63 read_signed_LEB_128(const unsigned char* buffer, size_t* len)
  64 {
  65   int64_t result = 0;
  66   int shift = 0;
  67   size_t num_read = 0;
  68   unsigned char byte;
  69
  70   do
  71     {
  72       byte = *buffer++;
  73       num_read++;
  74       result |= (static_cast<uint64_t>(byte & 0x7f) << shift);
  75       shift += 7;
  76     }
  77   while (byte & 0x80);
  78
  79   if ((shift < 8 * static_cast<int>(sizeof(result))) && (byte & 0x40))
  80     result |= -((static_cast<int64_t>(1)) << shift);
  81   *len = num_read;
  82   return result;
  83 }
  84
  85 } // End anonymous namespace.
  86
  87
  88 namespace gold {
  89
  90 // This is the format of a DWARF2/3 line state machine that we process
  91 // opcodes using.  There is no need for anything outside the lineinfo
  92 // processor to know how this works.
  93
  94 struct LineStateMachine
  95 {
  96   int file_num;
  97   uint64_t address;
  98   int line_num;
  99   int column_num;
 100   unsigned int shndx;    // the section address refers to
 101   bool is_stmt;          // stmt means statement.
 102   bool basic_block;
 103   bool end_sequence;
 104 };
 105
 106 static void
 107 ResetLineStateMachine(struct LineStateMachine* lsm, bool default_is_stmt)
 108 {
 109   lsm->file_num = 1;
 110   lsm->address = 0;
 111   lsm->line_num = 1;
 112   lsm->column_num = 0;
 113   lsm->shndx = -1U;
 114   lsm->is_stmt = default_is_stmt;
 115   lsm->basic_block = false;
 116   lsm->end_sequence = false;
 117 }
 118
 119 template<int size, bool big_endian>
 120 Dwarf_line_info<size, big_endian>::Dwarf_line_info(Object* object)
 121   : data_valid_(false), buffer_(NULL), symtab_buffer_(NULL),
 122     directories_(1), files_(1)
 123 {
 124   unsigned int debug_shndx;
 125   for (debug_shndx = 0; debug_shndx < object->shnum(); ++debug_shndx)
 126     if (object->section_name(debug_shndx) == ".debug_line")
 127       {
 128         off_t buffer_size;
 129         this->buffer_ = object->section_contents(
 130             debug_shndx, &buffer_size, false);
 131         this->buffer_end_ = this->buffer_ + buffer_size;
 132         break;
 133       }
 134   if (this->buffer_ == NULL)
 135     return;
 136
 137   // Find the relocation section for ".debug_line".
 138   bool got_relocs = false;
 139   for (unsigned int reloc_shndx = 0;
 140        reloc_shndx < object->shnum();
 141        ++reloc_shndx)
 142     {
 143       unsigned int reloc_sh_type = object->section_type(reloc_shndx);
 144       if ((reloc_sh_type == elfcpp::SHT_REL
 145            || reloc_sh_type == elfcpp::SHT_RELA)
 146           && object->section_info(reloc_shndx) == debug_shndx)
 147         {
 148           got_relocs = this->track_relocs_.initialize(object, reloc_shndx,
 149                                                       reloc_sh_type);
 150           break;
 151         }
 152     }
 153   if (!got_relocs)
 154     return;
 155
 156   // Finally, we need the symtab section to interpret the relocs.
 157   unsigned int symtab_shndx;
 158   for (symtab_shndx = 0; symtab_shndx < object->shnum(); ++symtab_shndx)
 159     if (object->section_type(symtab_shndx) == elfcpp::SHT_SYMTAB)
 160       {
 161         off_t symtab_size;
 162         this->symtab_buffer_ = object->section_contents(
 163             symtab_shndx, &symtab_size, false);
 164         this->symtab_buffer_end_ = this->symtab_buffer_ + symtab_size;
 165         break;
 166       }
 167   if (this->symtab_buffer_ == NULL)
 168     return;
 169
 170   // Now that we have successfully read all the data, parse the debug
 171   // info.
 172   this->data_valid_ = true;
 173   this->read_line_mappings();
 174 }
 175
 176 // Read the DWARF header.
 177
 178 template<int size, bool big_endian>
 179 const unsigned char*
 180 Dwarf_line_info<size, big_endian>::read_header_prolog(
 181     const unsigned char* lineptr)
 182 {
 183   uint32_t initial_length = elfcpp::Swap<32, big_endian>::readval(lineptr);
 184   lineptr += 4;
 185
 186   // In DWARF2/3, if the initial length is all 1 bits, then the offset
 187   // size is 8 and we need to read the next 8 bytes for the real length.
 188   if (initial_length == 0xffffffff)
 189     {
 190       header_.offset_size = 8;
 191       initial_length = elfcpp::Swap<64, big_endian>::readval(lineptr);
 192       lineptr += 8;
 193     }
 194   else
 195     header_.offset_size = 4;
 196
 197   header_.total_length = initial_length;
 198
 199   gold_assert(lineptr + header_.total_length <= buffer_end_);
 200
 201   header_.version = elfcpp::Swap<16, big_endian>::readval(lineptr);
 202   lineptr += 2;
 203
 204   if (header_.offset_size == 4)
 205     header_.prologue_length = elfcpp::Swap<32, big_endian>::readval(lineptr);
 206   else
 207     header_.prologue_length = elfcpp::Swap<64, big_endian>::readval(lineptr);
 208   lineptr += header_.offset_size;
 209
 210   header_.min_insn_length = *lineptr;
 211   lineptr += 1;
 212
 213   header_.default_is_stmt = *lineptr;
 214   lineptr += 1;
 215
 216   header_.line_base = *reinterpret_cast<const signed char*>(lineptr);
 217   lineptr += 1;
 218
 219   header_.line_range = *lineptr;
 220   lineptr += 1;
 221
 222   header_.opcode_base = *lineptr;
 223   lineptr += 1;
 224
 225   header_.std_opcode_lengths.reserve(header_.opcode_base + 1);
 226   header_.std_opcode_lengths[0] = 0;
 227   for (int i = 1; i < header_.opcode_base; i++)
 228     {
 229       header_.std_opcode_lengths[i] = *lineptr;
 230       lineptr += 1;
 231     }
 232
 233   return lineptr;
 234 }
 235
 236 // The header for a debug_line section is mildly complicated, because
 237 // the line info is very tightly encoded.
 238
 239 template<int size, bool big_endian>
 240 const unsigned char*
 241 Dwarf_line_info<size, big_endian>::read_header_tables(
 242     const unsigned char* lineptr)
 243 {
 244   // It is legal for the directory entry table to be empty.
 245   if (*lineptr)
 246     {
 247       int dirindex = 1;
 248       while (*lineptr)
 249         {
 250           const unsigned char* dirname = lineptr;
 251           gold_assert(dirindex == static_cast<int>(directories_.size()));
 252           directories_.push_back(reinterpret_cast<const char*>(dirname));
 253           lineptr += directories_.back().size() + 1;
 254           dirindex++;
 255         }
 256     }
 257   lineptr++;
 258
 259   // It is also legal for the file entry table to be empty.
 260   if (*lineptr)
 261     {
 262       int fileindex = 1;
 263       size_t len;
 264       while (*lineptr)
 265         {
 266           const char* filename = reinterpret_cast<const char*>(lineptr);
 267           lineptr += strlen(filename) + 1;
 268
 269           uint64_t dirindex = read_unsigned_LEB_128(lineptr, &len);
 270           if (dirindex >= directories_.size())
 271             dirindex = 0;
 272           lineptr += len;
 273
 274           read_unsigned_LEB_128(lineptr, &len);   // mod_time
 275           lineptr += len;
 276
 277           read_unsigned_LEB_128(lineptr, &len);   // filelength
 278           lineptr += len;
 279
 280           gold_assert(fileindex == static_cast<int>(files_.size()));
 281           files_.push_back(std::pair<int, std::string>(dirindex, filename));
 282           fileindex++;
 283         }
 284     }
 285   lineptr++;
 286
 287   return lineptr;
 288 }
 289
 290 // Process a single opcode in the .debug.line structure.
 291
 292 // Templating on size and big_endian would yield more efficient (and
 293 // simpler) code, but would bloat the binary.  Speed isn't important
 294 // here.
 295
 296 template<int size, bool big_endian>
 297 bool
 298 Dwarf_line_info<size, big_endian>::process_one_opcode(
 299     const unsigned char* start, struct LineStateMachine* lsm, size_t* len)
 300 {
 301   size_t oplen = 0;
 302   size_t templen;
 303   unsigned char opcode = *start;
 304   oplen++;
 305   start++;
 306
 307   // If the opcode is great than the opcode_base, it is a special
 308   // opcode. Most line programs consist mainly of special opcodes.
 309   if (opcode >= header_.opcode_base)
 310     {
 311       opcode -= header_.opcode_base;
 312       const int advance_address = ((opcode / header_.line_range)
 313                                    * header_.min_insn_length);
 314       lsm->address += advance_address;
 315
 316       const int advance_line = ((opcode % header_.line_range)
 317                                 + header_.line_base);
 318       lsm->line_num += advance_line;
 319       lsm->basic_block = true;
 320       *len = oplen;
 321       return true;
 322     }
 323
 324   // Otherwise, we have the regular opcodes
 325   switch (opcode)
 326     {
 327     case elfcpp::DW_LNS_copy:
 328       lsm->basic_block = false;
 329       *len = oplen;
 330       return true;
 331
 332     case elfcpp::DW_LNS_advance_pc:
 333       {
 334         const uint64_t advance_address
 335             = read_unsigned_LEB_128(start, &templen);
 336         oplen += templen;
 337         lsm->address += header_.min_insn_length * advance_address;
 338       }
 339       break;
 340
 341     case elfcpp::DW_LNS_advance_line:
 342       {
 343         const uint64_t advance_line = read_signed_LEB_128(start, &templen);
 344         oplen += templen;
 345         lsm->line_num += advance_line;
 346       }
 347       break;
 348
 349     case elfcpp::DW_LNS_set_file:
 350       {
 351         const uint64_t fileno = read_unsigned_LEB_128(start, &templen);
 352         oplen += templen;
 353         lsm->file_num = fileno;
 354       }
 355       break;
 356
 357     case elfcpp::DW_LNS_set_column:
 358       {
 359         const uint64_t colno = read_unsigned_LEB_128(start, &templen);
 360         oplen += templen;
 361         lsm->column_num = colno;
 362       }
 363       break;
 364
 365     case elfcpp::DW_LNS_negate_stmt:
 366       lsm->is_stmt = !lsm->is_stmt;
 367       break;
 368
 369     case elfcpp::DW_LNS_set_basic_block:
 370       lsm->basic_block = true;
 371       break;
 372
 373     case elfcpp::DW_LNS_fixed_advance_pc:
 374       {
 375         int advance_address;
 376         advance_address = elfcpp::Swap<16, big_endian>::readval(start);
 377         oplen += 2;
 378         lsm->address += advance_address;
 379       }
 380       break;
 381
 382     case elfcpp::DW_LNS_const_add_pc:
 383       {
 384         const int advance_address = (header_.min_insn_length
 385                                      * ((255 - header_.opcode_base)
 386                                         / header_.line_range));
 387         lsm->address += advance_address;
 388       }
 389       break;
 390
 391     case elfcpp::DW_LNS_extended_op:
 392       {
 393         const uint64_t extended_op_len
 394             = read_unsigned_LEB_128(start, &templen);
 395         start += templen;
 396         oplen += templen + extended_op_len;
 397
 398         const unsigned char extended_op = *start;
 399         start++;
 400
 401         switch (extended_op)
 402           {
 403           case elfcpp::DW_LNE_end_sequence:
 404             lsm->end_sequence = true;
 405             *len = oplen;
 406             return true;
 407
 408           case elfcpp::DW_LNE_set_address:
 409             {
 410               typename Reloc_map::const_iterator it
 411                   = reloc_map_.find(start - this->buffer_);
 412               if (it != reloc_map_.end())
 413                 {
 414                   // value + addend.
 415                   lsm->address =
 416                     (elfcpp::Swap<size, big_endian>::readval(start)
 417                      + it->second.second);
 418                   lsm->shndx = it->second.first;
 419                 }
 420               else
 421                 {
 422                   // Every set_address should have an associated
 423                   // relocation.
 424                   this->data_valid_ = false;
 425                 }
 426               break;
 427             }
 428           case elfcpp::DW_LNE_define_file:
 429             {
 430               const char* filename  = reinterpret_cast<const char*>(start);
 431               templen = strlen(filename) + 1;
 432               start += templen;
 433
 434               uint64_t dirindex = read_unsigned_LEB_128(start, &templen);
 435               if (dirindex >= directories_.size())
 436                 dirindex = 0;
 437               oplen += templen;
 438
 439               read_unsigned_LEB_128(start, &templen);   // mod_time
 440               oplen += templen;
 441
 442               read_unsigned_LEB_128(start, &templen);   // filelength
 443               oplen += templen;
 444
 445               files_.push_back(std::pair<int, std::string>(dirindex,
 446                                                            filename));
 447             }
 448             break;
 449           }
 450       }
 451       break;
 452
 453     default:
 454       {
 455         // Ignore unknown opcode  silently
 456         for (int i = 0; i < header_.std_opcode_lengths[opcode]; i++)
 457           {
 458             size_t templen;
 459             read_unsigned_LEB_128(start, &templen);
 460             start += templen;
 461             oplen += templen;
 462           }
 463       }
 464       break;
 465   }
 466   *len = oplen;
 467   return false;
 468 }
 469
 470 // Read the debug information at LINEPTR and store it in the line
 471 // number map.
 472
 473 template<int size, bool big_endian>
 474 unsigned const char*
 475 Dwarf_line_info<size, big_endian>::read_lines(unsigned const char* lineptr)
 476 {
 477   struct LineStateMachine lsm;
 478
 479   // LENGTHSTART is the place the length field is based on.  It is the
 480   // point in the header after the initial length field.
 481   const unsigned char* lengthstart = buffer_;
 482
 483   // In 64 bit dwarf, the initial length is 12 bytes, because of the
 484   // 0xffffffff at the start.
 485   if (header_.offset_size == 8)
 486     lengthstart += 12;
 487   else
 488     lengthstart += 4;
 489
 490   while (lineptr < lengthstart + header_.total_length)
 491     {
 492       ResetLineStateMachine(&lsm, header_.default_is_stmt);
 493       while (!lsm.end_sequence)
 494         {
 495           size_t oplength;
 496           bool add_line = this->process_one_opcode(lineptr, &lsm, &oplength);
 497           if (add_line)
 498             {
 499               Offset_to_lineno_entry entry
 500                   = { lsm.address, lsm.file_num, lsm.line_num };
 501               line_number_map_[lsm.shndx].push_back(entry);
 502             }
 503           lineptr += oplength;
 504         }
 505     }
 506
 507   return lengthstart + header_.total_length;
 508 }
 509
 510 // Looks in the symtab to see what section a symbol is in.
 511
 512 template<int size, bool big_endian>
 513 unsigned int
 514 Dwarf_line_info<size, big_endian>::symbol_section(
 515     unsigned int sym,
 516     typename elfcpp::Elf_types<size>::Elf_Addr* value)
 517 {
 518   const int symsize = elfcpp::Elf_sizes<size>::sym_size;
 519   gold_assert(this->symtab_buffer_ + sym * symsize < this->symtab_buffer_end_);
 520   elfcpp::Sym<size, big_endian> elfsym(this->symtab_buffer_ + sym * symsize);
 521   *value = elfsym.get_st_value();
 522   return elfsym.get_st_shndx();
 523 }
 524
 525 // Read the relocations into a Reloc_map.
 526
 527 template<int size, bool big_endian>
 528 void
 529 Dwarf_line_info<size, big_endian>::read_relocs()
 530 {
 531   if (this->symtab_buffer_ == NULL)
 532     return;
 533
 534   typename elfcpp::Elf_types<size>::Elf_Addr value;
 535   off_t reloc_offset;
 536   while ((reloc_offset = this->track_relocs_.next_offset()) != -1)
 537     {
 538       const unsigned int sym = this->track_relocs_.next_symndx();
 539       const unsigned int shndx = this->symbol_section(sym, &value);
 540       this->reloc_map_[reloc_offset] = std::make_pair(shndx, value);
 541       this->track_relocs_.advance(reloc_offset + 1);
 542     }
 543 }
 544
 545 // Read the line number info.
 546
 547 template<int size, bool big_endian>
 548 void
 549 Dwarf_line_info<size, big_endian>::read_line_mappings()
 550 {
 551   gold_assert(this->data_valid_ == true);
 552
 553   read_relocs();
 554   while (this->buffer_ < this->buffer_end_)
 555     {
 556       const unsigned char* lineptr = this->buffer_;
 557       lineptr = this->read_header_prolog(lineptr);
 558       lineptr = this->read_header_tables(lineptr);
 559       lineptr = this->read_lines(lineptr);
 560       this->buffer_ = lineptr;
 561     }
 562
 563   // Sort the lines numbers, so addr2line can use binary search.
 564   for (typename Lineno_map::iterator it = line_number_map_.begin();
 565        it != line_number_map_.end();
 566        ++it)
 567     // Each vector needs to be sorted by offset.
 568     std::sort(it->second.begin(), it->second.end());
 569 }
 570
 571 // Return a string for a file name and line number.
 572
 573 template<int size, bool big_endian>
 574 std::string
 575 Dwarf_line_info<size, big_endian>::addr2line(unsigned int shndx, off_t offset)
 576 {
 577   if (this->data_valid_ == false)
 578     return "";
 579
 580   const Offset_to_lineno_entry lookup_key = { offset, 0, 0 };
 581   std::vector<Offset_to_lineno_entry>& offsets = this->line_number_map_[shndx];
 582   if (offsets.empty())
 583     return "";
 584
 585   typename std::vector<Offset_to_lineno_entry>::const_iterator it
 586       = std::lower_bound(offsets.begin(), offsets.end(), lookup_key);
 587
 588   // If we found an exact match, great, otherwise find the last entry
 589   // before the passed-in offset.
 590   if (it->offset > offset)
 591     {
 592       if (it == offsets.begin())
 593         return "";
 594       --it;
 595       gold_assert(it->offset < offset);
 596     }
 597
 598   // Convert the file_num + line_num into a string.
 599   std::string ret;
 600   gold_assert(it->file_num < static_cast<int>(files_.size()));
 601   const std::pair<int, std::string>& filename_pair = files_[it->file_num];
 602   gold_assert(filename_pair.first < static_cast<int>(directories_.size()));
 603   const std::string& dirname = directories_[filename_pair.first];
 604   const std::string& filename = filename_pair.second;
 605   if (!dirname.empty())
 606     {
 607       ret += dirname;
 608       ret += "/";
 609     }
 610   ret += filename;
 611   if (ret.empty())
 612     ret = "(unknown)";
 613
 614   char buffer[64];   // enough to hold a line number
 615   snprintf(buffer, sizeof(buffer), "%d", it->line_num);
 616   ret += ":";
 617   ret += buffer;
 618
 619   return ret;
 620 }
 621
 622 #ifdef HAVE_TARGET_32_LITTLE
 623 template
 624 class Dwarf_line_info<32, false>;
 625 #endif
 626
 627 #ifdef HAVE_TARGET_32_BIG
 628 template
 629 class Dwarf_line_info<32, true>;
 630 #endif
 631
 632 #ifdef HAVE_TARGET_64_LITTLE
 633 template
 634 class Dwarf_line_info<64, false>;
 635 #endif
 636
 637 #ifdef HAVE_TARGET_64_BIG
 638 template
 639 class Dwarf_line_info<64, true>;
 640 #endif
 641
 642 } // End namespace gold.