gold/reduced_debug_output.cc

   1 // reduced_debug_output.cc -- output reduced debugging information to save space
   2
   3 // Copyright 2008 Free Software Foundation, Inc.
   4 // Written by Caleb Howe <cshowe@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 "parameters.h"
  26 #include "options.h"
  27 #include "dwarf.h"
  28 #include "dwarf_reader.h"
  29 #include "reduced_debug_output.h"
  30
  31 #include <vector>
  32
  33 namespace gold
  34 {
  35
  36 void
  37 write_unsigned_LEB_128(std::vector<unsigned char>* buffer, uint64_t value)
  38 {
  39   do
  40     {
  41       unsigned char current_byte = value & 0x7f;
  42       value >>= 7;
  43       if (value != 0)
  44         {
  45           current_byte |= 0x80;
  46         }
  47       buffer->push_back(current_byte);
  48     }
  49   while (value != 0);
  50 }
  51
  52 size_t
  53 get_length_as_unsigned_LEB_128(uint64_t value)
  54 {
  55   size_t length = 0;
  56   do
  57     {
  58       unsigned char current_byte = value & 0x7f;
  59       value >>= 7;
  60       if (value != 0)
  61         {
  62           current_byte |= 0x80;
  63         }
  64       length++;
  65     }
  66   while (value != 0);
  67   return length;
  68 }
  69
  70 template <int valsize>
  71 void insert_into_vector(std::vector<unsigned char>* destination,
  72                         typename elfcpp::Valtype_base<valsize>::Valtype value)
  73 {
  74   unsigned char buffer[valsize / 8];
  75   if (parameters->target().is_big_endian())
  76     elfcpp::Swap_unaligned<valsize, true>::writeval(buffer, value);
  77   else
  78     elfcpp::Swap_unaligned<valsize, false>::writeval(buffer, value);
  79   destination->insert(destination->end(), buffer, buffer + valsize / 8);
  80 }
  81
  82 template <int valsize>
  83 typename elfcpp::Valtype_base<valsize>::Valtype
  84 read_from_pointer(unsigned char** source)
  85 {
  86   typename elfcpp::Valtype_base<valsize>::Valtype return_value;
  87   if (parameters->target().is_big_endian())
  88     return_value = elfcpp::Swap_unaligned<valsize, true>::readval(*source);
  89   else
  90     return_value = elfcpp::Swap_unaligned<valsize, false>::readval(*source);
  91   *source += valsize / 8;
  92   return return_value;
  93 }
  94
  95 // Given a pointer to the beginning of a die and the beginning of the associated
  96 // abbreviation fills in die_end with the end of the information entry.  If
  97 // successful returns true.  Get_die_end also takes a pointer to the end of the
  98 // buffer containing the die.  If die_end would be beyond the end of the
  99 // buffer, or if an unsupported dwarf form is encountered returns false.
 100 bool
 101 Output_reduced_debug_info_section::get_die_end(
 102     unsigned char* die, unsigned char* abbrev, unsigned char** die_end,
 103     unsigned char* buffer_end, int address_size, bool is64)
 104 {
 105   size_t LEB_size;
 106   uint64_t LEB_decoded;
 107   for(;;)
 108     {
 109       uint64_t attribute = read_unsigned_LEB_128(abbrev, &LEB_size);
 110       abbrev += LEB_size;
 111       elfcpp::DW_FORM form =
 112           static_cast<elfcpp::DW_FORM>(read_unsigned_LEB_128(abbrev,
 113                                                              &LEB_size));
 114       abbrev += LEB_size;
 115       if (!(attribute || form))
 116         break;
 117       if (die >= buffer_end)
 118         return false;
 119       switch(form)
 120         {
 121           case elfcpp::DW_FORM_null:
 122             break;
 123           case elfcpp::DW_FORM_strp:
 124             die += is64 ? 8 : 4;
 125             break;
 126           case elfcpp::DW_FORM_addr:
 127           case elfcpp::DW_FORM_ref_addr:
 128             die += address_size;
 129             break;
 130           case elfcpp::DW_FORM_block1:
 131             die += *die;
 132             die += 1;
 133             break;
 134           case elfcpp::DW_FORM_block2:
 135             {
 136               uint16_t block_size;
 137               block_size = read_from_pointer<16>(&die);
 138               die += block_size;
 139               break;
 140             }
 141           case elfcpp::DW_FORM_block4:
 142             {
 143               uint32_t block_size;
 144               block_size = read_from_pointer<32>(&die);
 145               die += block_size;
 146               break;
 147             }
 148           case elfcpp::DW_FORM_block:
 149             LEB_decoded = read_unsigned_LEB_128(die, &LEB_size);
 150             die += (LEB_decoded + LEB_size);
 151             break;
 152           case elfcpp::DW_FORM_data1:
 153           case elfcpp::DW_FORM_ref1:
 154           case elfcpp::DW_FORM_flag:
 155             die += 1;
 156             break;
 157           case elfcpp::DW_FORM_data2:
 158           case elfcpp::DW_FORM_ref2:
 159             die += 2;
 160             break;
 161           case elfcpp::DW_FORM_data4:
 162           case elfcpp::DW_FORM_ref4:
 163             die += 4;
 164             break;
 165           case elfcpp::DW_FORM_data8:
 166           case elfcpp::DW_FORM_ref8:
 167             die += 8;
 168             break;
 169           case elfcpp::DW_FORM_ref_udata:
 170           case elfcpp::DW_FORM_udata:
 171             read_unsigned_LEB_128(die, &LEB_size);
 172             die += LEB_size;
 173             break;
 174           case elfcpp::DW_FORM_string:
 175             {
 176               size_t length = strlen(reinterpret_cast<char*>(die));
 177               die += length + 1;
 178               break;
 179             }
 180           case elfcpp::DW_FORM_sdata:
 181           case elfcpp::DW_FORM_indirect:
 182             return false;
 183       }
 184     }
 185   *die_end = die;
 186   return true;
 187 }
 188
 189 void
 190 Output_reduced_debug_abbrev_section::set_final_data_size()
 191 {
 192   if (this->sized_ || this->failed_)
 193     return;
 194
 195   uint64_t abbrev_number;
 196   size_t LEB_size;
 197   unsigned char* abbrev_data = this->postprocessing_buffer();
 198   unsigned char* abbrev_end = this->postprocessing_buffer() +
 199                               this->postprocessing_buffer_size();
 200   this->write_to_postprocessing_buffer();
 201   while(abbrev_data < abbrev_end)
 202     {
 203       uint64_t abbrev_offset = abbrev_data - this->postprocessing_buffer();
 204       while((abbrev_number = read_unsigned_LEB_128(abbrev_data, &LEB_size)))
 205         {
 206           if (abbrev_data >= abbrev_end)
 207             {
 208               failed("Debug abbreviations extend beyond .debug_abbrev "
 209                      "section; failed to reduce debug abbreviations");
 210               return;
 211             }
 212           abbrev_data += LEB_size;
 213
 214           // Together with the abbreviation number these fields make up
 215           // the header for each abbreviation
 216           uint64_t abbrev_type = read_unsigned_LEB_128(abbrev_data, &LEB_size);
 217           abbrev_data += LEB_size;
 218
 219           // This would ordinarily be the has_children field of the
 220           // abbreviation.  But it's going to be false after reducting the
 221           // information, so there's no point in storing it
 222           abbrev_data++;
 223
 224           // Read to the end of the current abbreviation
 225           // This is indicated by two zero unsigned LEBs in a row.  We don't
 226           // need to parse the data yet, so we just scan through the data
 227           // looking for two consecutive 0 bytes indicating the end of the
 228           // abbreviation
 229           unsigned char* current_abbrev;
 230           for (current_abbrev = abbrev_data;
 231                current_abbrev[0] || current_abbrev[1];
 232                current_abbrev++)
 233             {
 234               if (current_abbrev >= abbrev_end)
 235                 {
 236                   this->failed(_("Debug abbreviations extend beyond "
 237                                  ".debug_abbrev section; failed to reduce "
 238                                  "debug abbreviations"));
 239                   return;
 240                 }
 241             }
 242           // Account for the two nulls and advance to the start of the
 243           // next abbreviation.
 244           current_abbrev += 2;
 245
 246           // We're eliminating every entry except for compile units, so we
 247           // only need to store abbreviations that describe them
 248           if (abbrev_type == elfcpp::DW_TAG_compile_unit)
 249             {
 250               write_unsigned_LEB_128(&this->data_, ++this->abbrev_count_);
 251               write_unsigned_LEB_128(&this->data_, abbrev_type);
 252               // has_children is false for all entries
 253               this->data_.push_back(0);
 254               this->abbrev_mapping_[std::make_pair(abbrev_offset,
 255                                                    abbrev_number)] =
 256                   std::make_pair(abbrev_count_, this->data_.size());
 257               this->data_.insert(this->data_.end(), abbrev_data,
 258                                  current_abbrev);
 259             }
 260           abbrev_data = current_abbrev;
 261         }
 262       gold_assert(LEB_size == 1);
 263       abbrev_data += LEB_size;
 264     }
 265   // Null terminate the list of abbreviations
 266   this->data_.push_back(0);
 267   this->set_data_size(data_.size());
 268   this->sized_ = true;
 269 }
 270
 271 void
 272 Output_reduced_debug_abbrev_section::do_write(Output_file* of)
 273 {
 274   off_t offset = this->offset();
 275   off_t data_size = this->data_size();
 276   unsigned char* view = of->get_output_view(offset, data_size);
 277   if (this->failed_)
 278     memcpy(view, this->postprocessing_buffer(),
 279            this->postprocessing_buffer_size());
 280   else
 281     memcpy(view, &this->data_.front(), data_size);
 282   of->write_output_view(offset, data_size, view);
 283 }
 284
 285 // Locates the abbreviation with abbreviation_number abbrev_number in the
 286 // abbreviation table at offset abbrev_offset.  abbrev_number is updated with
 287 // its new abbreviation number and a pointer to the beginning of the
 288 // abbreviation is returned.
 289 unsigned char*
 290 Output_reduced_debug_abbrev_section::get_new_abbrev(
 291   uint64_t* abbrev_number, uint64_t abbrev_offset)
 292 {
 293   set_final_data_size();
 294   std::pair<uint64_t, uint64_t> abbrev_info =
 295       this->abbrev_mapping_[std::make_pair(abbrev_offset, *abbrev_number)];
 296   *abbrev_number = abbrev_info.first;
 297   return &this->data_[abbrev_info.second];
 298 }
 299
 300 void Output_reduced_debug_info_section::set_final_data_size()
 301 {
 302   if (this->failed_)
 303     return;
 304   unsigned char* debug_info = this->postprocessing_buffer();
 305   unsigned char* debug_info_end = (this->postprocessing_buffer()
 306                                    + this->postprocessing_buffer_size());
 307   unsigned char* next_compile_unit;
 308   this->write_to_postprocessing_buffer();
 309
 310   while (debug_info < debug_info_end)
 311     {
 312       uint32_t compile_unit_start = read_from_pointer<32>(&debug_info);
 313       // The first 4 bytes of each compile unit determine whether or
 314       // not we're using dwarf32 or dwarf64.  This is not necessarily
 315       // related to whether the binary is 32 or 64 bits.
 316       if (compile_unit_start == 0xFFFFFFFF)
 317         {
 318           // Technically the size can be up to 96 bits.  Rather than handle
 319           // 96/128 bit integers we just truncate the size at 64 bits.
 320           if (0 != read_from_pointer<32>(&debug_info))
 321             {
 322               this->failed(_("Extremely large compile unit in debug info; "
 323                              "failed to reduce debug info"));
 324               return;
 325             }
 326           const int dwarf64_header_size = sizeof(uint64_t) + sizeof(uint16_t) +
 327                                           sizeof(uint64_t) + sizeof(uint8_t);
 328           if (debug_info + dwarf64_header_size >= debug_info_end)
 329             {
 330               this->failed(_("Debug info extends beyond .debug_info section;"
 331                              "failed to reduce debug info"));
 332               return;
 333             }
 334
 335           uint64_t compile_unit_size = read_from_pointer<64>(&debug_info);
 336           next_compile_unit = debug_info + compile_unit_size;
 337           uint16_t version = read_from_pointer<16>(&debug_info);
 338           uint64_t abbrev_offset = read_from_pointer<64>(&debug_info);
 339           uint8_t address_size = read_from_pointer<8>(&debug_info);
 340           size_t LEB_size;
 341           uint64_t abbreviation_number = read_unsigned_LEB_128(debug_info,
 342                                                                &LEB_size);
 343           debug_info += LEB_size;
 344           unsigned char* die_abbrev = this->associated_abbrev_->get_new_abbrev(
 345               &abbreviation_number, abbrev_offset);
 346           unsigned char* die_end;
 347           if (!this->get_die_end(debug_info, die_abbrev, &die_end,
 348                                  debug_info_end, address_size, true))
 349             {
 350               this->failed(_("Invalid DIE in debug info; "
 351                              "failed to reduce debug info"));
 352               return;
 353             }
 354
 355           insert_into_vector<32>(&this->data_, 0xFFFFFFFF);
 356           insert_into_vector<32>(&this->data_, 0);
 357           insert_into_vector<64>(
 358               &this->data_,
 359               (11 + get_length_as_unsigned_LEB_128(abbreviation_number)
 360                + die_end - debug_info));
 361           insert_into_vector<16>(&this->data_, version);
 362           insert_into_vector<64>(&this->data_, 0);
 363           insert_into_vector<8>(&this->data_, address_size);
 364           write_unsigned_LEB_128(&this->data_, abbreviation_number);
 365           this->data_.insert(this->data_.end(), debug_info, die_end);
 366         }
 367       else
 368         {
 369           const int dwarf32_header_size =
 370               sizeof(uint16_t) + sizeof(uint32_t) + sizeof(uint8_t);
 371           if (debug_info + dwarf32_header_size >= debug_info_end)
 372             {
 373               this->failed(_("Debug info extends beyond .debug_info section; "
 374                              "failed to reduce debug info"));
 375               return;
 376             }
 377           uint32_t compile_unit_size = compile_unit_start;
 378           next_compile_unit = debug_info + compile_unit_size;
 379           uint16_t version = read_from_pointer<16>(&debug_info);
 380           uint32_t abbrev_offset = read_from_pointer<32>(&debug_info);
 381           uint8_t address_size = read_from_pointer<8>(&debug_info);
 382           size_t LEB_size;
 383           uint64_t abbreviation_number = read_unsigned_LEB_128(debug_info,
 384                                                                &LEB_size);
 385           debug_info += LEB_size;
 386           unsigned char* die_abbrev = this->associated_abbrev_->get_new_abbrev(
 387               &abbreviation_number, abbrev_offset);
 388           unsigned char* die_end;
 389           if (!this->get_die_end(debug_info, die_abbrev, &die_end,
 390                                  debug_info_end, address_size, false))
 391             {
 392               this->failed(_("Invalid DIE in debug info; "
 393                              "failed to reduce debug info"));
 394               return;
 395             }
 396
 397           insert_into_vector<32>(
 398               &this->data_,
 399               (7 + get_length_as_unsigned_LEB_128(abbreviation_number)
 400                + die_end - debug_info));
 401           insert_into_vector<16>(&this->data_, version);
 402           insert_into_vector<32>(&this->data_, 0);
 403           insert_into_vector<8>(&this->data_, address_size);
 404           write_unsigned_LEB_128(&this->data_, abbreviation_number);
 405           this->data_.insert(this->data_.end(), debug_info, die_end);
 406         }
 407       debug_info = next_compile_unit;
 408     }
 409   this->set_data_size(data_.size());
 410 }
 411
 412 void Output_reduced_debug_info_section::do_write(Output_file* of)
 413 {
 414   off_t offset = this->offset();
 415   off_t data_size = this->data_size();
 416   unsigned char* view = of->get_output_view(offset, data_size);
 417   if (this->failed_)
 418     memcpy(view, this->postprocessing_buffer(),
 419            this->postprocessing_buffer_size());
 420   else
 421     memcpy(view, &this->data_.front(), data_size);
 422   of->write_output_view(offset, data_size, view);
 423 }
 424
 425 } // End namespace gold.