gold/stringpool.cc

   1 // stringpool.cc -- a string pool for gold
   2
   3 // Copyright (C) 2006-2023 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 <cstring>
  26 #include <algorithm>
  27 #include <vector>
  28
  29 #include "output.h"
  30 #include "parameters.h"
  31 #include "stringpool.h"
  32
  33 namespace gold
  34 {
  35
  36 template<typename Stringpool_char>
  37 Stringpool_template<Stringpool_char>::Stringpool_template(uint64_t addralign)
  38   : string_set_(), key_to_offset_(), strings_(), strtab_size_(0),
  39     zero_null_(true), optimize_(false), offset_(sizeof(Stringpool_char)),
  40     addralign_(addralign)
  41 {
  42   if (parameters->options_valid()
  43       && parameters->options().optimize() >= 2
  44       && addralign <= sizeof(Stringpool_char))
  45     this->optimize_ = true;
  46 }
  47
  48 template<typename Stringpool_char>
  49 void
  50 Stringpool_template<Stringpool_char>::clear()
  51 {
  52   for (typename std::list<Stringdata*>::iterator p = this->strings_.begin();
  53        p != this->strings_.end();
  54        ++p)
  55     delete[] reinterpret_cast<char*>(*p);
  56   this->strings_.clear();
  57   this->key_to_offset_.clear();
  58   this->string_set_.clear();
  59 }
  60
  61 template<typename Stringpool_char>
  62 Stringpool_template<Stringpool_char>::~Stringpool_template()
  63 {
  64   this->clear();
  65 }
  66
  67 // Resize the internal hashtable with the expectation we'll get n new
  68 // elements.  Note that the hashtable constructor takes a "number of
  69 // buckets you'd like," rather than "number of elements you'd like,"
  70 // but that's the best we can do.
  71
  72 template<typename Stringpool_char>
  73 void
  74 Stringpool_template<Stringpool_char>::reserve(unsigned int n)
  75 {
  76   this->key_to_offset_.reserve(n);
  77
  78 #if defined(HAVE_UNORDERED_MAP)
  79   this->string_set_.rehash(this->string_set_.size() + n);
  80   return;
  81 #elif defined(HAVE_TR1_UNORDERED_MAP)
  82   // rehash() implementation is broken in gcc 4.0.3's stl
  83   //this->string_set_.rehash(this->string_set_.size() + n);
  84   //return;
  85 #elif defined(HAVE_EXT_HASH_MAP)
  86   this->string_set_.resize(this->string_set_.size() + n);
  87   return;
  88 #endif
  89
  90   // This is the generic "reserve" code, if no #ifdef above triggers.
  91   String_set_type new_string_set(this->string_set_.size() + n);
  92   new_string_set.insert(this->string_set_.begin(), this->string_set_.end());
  93   this->string_set_.swap(new_string_set);
  94 }
  95
  96 // Compare two strings of arbitrary character type for equality.
  97
  98 template<typename Stringpool_char>
  99 bool
 100 Stringpool_template<Stringpool_char>::string_equal(const Stringpool_char* s1,
 101                                                    const Stringpool_char* s2)
 102 {
 103   while (*s1 != 0)
 104     if (*s1++ != *s2++)
 105       return false;
 106   return *s2 == 0;
 107 }
 108
 109 // Specialize string_equal for char.
 110
 111 template<>
 112 inline bool
 113 Stringpool_template<char>::string_equal(const char* s1, const char* s2)
 114 {
 115   return strcmp(s1, s2) == 0;
 116 }
 117
 118 // Equality comparison function for the hash table.
 119
 120 template<typename Stringpool_char>
 121 inline bool
 122 Stringpool_template<Stringpool_char>::Stringpool_eq::operator()(
 123     const Hashkey& h1,
 124     const Hashkey& h2) const
 125 {
 126   return (h1.hash_code == h2.hash_code
 127           && h1.length == h2.length
 128           && (h1.string == h2.string
 129               || memcmp(h1.string, h2.string,
 130                         h1.length * sizeof(Stringpool_char)) == 0));
 131 }
 132
 133 // Hash function.  The length is in characters, not bytes.
 134
 135 template<typename Stringpool_char>
 136 size_t
 137 Stringpool_template<Stringpool_char>::string_hash(const Stringpool_char* s,
 138                                                   size_t length)
 139 {
 140   return gold::string_hash<Stringpool_char>(s, length);
 141 }
 142
 143 // Add the string S to the list of canonical strings.  Return a
 144 // pointer to the canonical string.  If PKEY is not NULL, set *PKEY to
 145 // the key.  LENGTH is the length of S in characters.  Note that S may
 146 // not be NUL terminated.
 147
 148 template<typename Stringpool_char>
 149 const Stringpool_char*
 150 Stringpool_template<Stringpool_char>::add_string(const Stringpool_char* s,
 151                                                  size_t len)
 152 {
 153   // We are in trouble if we've already computed the string offsets.
 154   gold_assert(this->strtab_size_ == 0);
 155
 156   // The size we allocate for a new Stringdata.
 157   const size_t buffer_size = 1000;
 158   // The amount we multiply the Stringdata index when calculating the
 159   // key.
 160   const size_t key_mult = 1024;
 161   gold_assert(key_mult >= buffer_size);
 162
 163   // Convert len to the number of bytes we need to allocate, including
 164   // the null character.
 165   len = (len + 1) * sizeof(Stringpool_char);
 166
 167   size_t alc;
 168   bool front = true;
 169   if (len > buffer_size)
 170     {
 171       alc = sizeof(Stringdata) + len;
 172       front = false;
 173     }
 174   else if (this->strings_.empty())
 175     alc = sizeof(Stringdata) + buffer_size;
 176   else
 177     {
 178       Stringdata* psd = this->strings_.front();
 179       if (len > psd->alc - psd->len)
 180         alc = sizeof(Stringdata) + buffer_size;
 181       else
 182         {
 183           char* ret = psd->data + psd->len;
 184           memcpy(ret, s, len - sizeof(Stringpool_char));
 185           memset(ret + len - sizeof(Stringpool_char), 0,
 186                  sizeof(Stringpool_char));
 187
 188           psd->len += len;
 189
 190           return reinterpret_cast<const Stringpool_char*>(ret);
 191         }
 192     }
 193
 194   Stringdata* psd = reinterpret_cast<Stringdata*>(new char[alc]);
 195   psd->alc = alc - sizeof(Stringdata);
 196   memcpy(psd->data, s, len - sizeof(Stringpool_char));
 197   memset(psd->data + len - sizeof(Stringpool_char), 0,
 198          sizeof(Stringpool_char));
 199   psd->len = len;
 200
 201   if (front)
 202     this->strings_.push_front(psd);
 203   else
 204     this->strings_.push_back(psd);
 205
 206   return reinterpret_cast<const Stringpool_char*>(psd->data);
 207 }
 208
 209 // Add a string to a string pool.
 210
 211 template<typename Stringpool_char>
 212 const Stringpool_char*
 213 Stringpool_template<Stringpool_char>::add(const Stringpool_char* s, bool copy,
 214                                           Key* pkey)
 215 {
 216   return this->add_with_length(s, string_length(s), copy, pkey);
 217 }
 218
 219 // Add a new key offset entry.
 220
 221 template<typename Stringpool_char>
 222 void
 223 Stringpool_template<Stringpool_char>::new_key_offset(size_t length)
 224 {
 225   section_offset_type offset;
 226   if (this->zero_null_ && length == 0)
 227     offset = 0;
 228   else
 229     {
 230       offset = this->offset_;
 231       // Align strings.
 232       offset = align_address(offset, this->addralign_);
 233       this->offset_ = offset + (length + 1) * sizeof(Stringpool_char);
 234     }
 235   this->key_to_offset_.push_back(offset);
 236 }
 237
 238 template<typename Stringpool_char>
 239 const Stringpool_char*
 240 Stringpool_template<Stringpool_char>::add_with_length(const Stringpool_char* s,
 241                                                       size_t length,
 242                                                       bool copy,
 243                                                       Key* pkey)
 244 {
 245   typedef std::pair<typename String_set_type::iterator, bool> Insert_type;
 246
 247   // We add 1 so that 0 is always invalid.
 248   const Key k = this->key_to_offset_.size() + 1;
 249
 250   if (!copy)
 251     {
 252       // When we don't need to copy the string, we can call insert
 253       // directly.
 254
 255       std::pair<Hashkey, Hashval> element(Hashkey(s, length), k);
 256
 257       Insert_type ins = this->string_set_.insert(element);
 258
 259       typename String_set_type::const_iterator p = ins.first;
 260
 261       if (ins.second)
 262         {
 263           // We just added the string.  The key value has now been
 264           // used.
 265           this->new_key_offset(length);
 266         }
 267       else
 268         {
 269           gold_assert(k != p->second);
 270         }
 271
 272       if (pkey != NULL)
 273         *pkey = p->second;
 274       return p->first.string;
 275     }
 276
 277   // When we have to copy the string, we look it up twice in the hash
 278   // table.  The problem is that we can't insert S before we
 279   // canonicalize it by copying it into the canonical list. The hash
 280   // code will only be computed once.
 281
 282   Hashkey hk(s, length);
 283   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 284   if (p != this->string_set_.end())
 285     {
 286       if (pkey != NULL)
 287         *pkey = p->second;
 288       return p->first.string;
 289     }
 290
 291   this->new_key_offset(length);
 292
 293   hk.string = this->add_string(s, length);
 294   // The contents of the string stay the same, so we don't need to
 295   // adjust hk.hash_code or hk.length.
 296
 297   std::pair<Hashkey, Hashval> element(hk, k);
 298
 299   Insert_type ins = this->string_set_.insert(element);
 300   gold_assert(ins.second);
 301
 302   if (pkey != NULL)
 303     *pkey = k;
 304   return hk.string;
 305 }
 306
 307 template<typename Stringpool_char>
 308 const Stringpool_char*
 309 Stringpool_template<Stringpool_char>::find(const Stringpool_char* s,
 310                                            Key* pkey) const
 311 {
 312   Hashkey hk(s);
 313   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 314   if (p == this->string_set_.end())
 315     return NULL;
 316
 317   if (pkey != NULL)
 318     *pkey = p->second;
 319
 320   return p->first.string;
 321 }
 322
 323 // Comparison routine used when sorting into an ELF strtab.  We want
 324 // to sort this so that when one string is a suffix of another, we
 325 // always see the shorter string immediately after the longer string.
 326 // For example, we want to see these strings in this order:
 327 //   abcd
 328 //   cd
 329 //   d
 330 // When strings are not suffixes, we don't care what order they are
 331 // in, but we need to ensure that suffixes wind up next to each other.
 332 // So we do a reversed lexicographic sort on the reversed string.
 333
 334 template<typename Stringpool_char>
 335 bool
 336 Stringpool_template<Stringpool_char>::Stringpool_sort_comparison::operator()(
 337   const Stringpool_sort_info& sort_info1,
 338   const Stringpool_sort_info& sort_info2) const
 339 {
 340   const Hashkey& h1(sort_info1->first);
 341   const Hashkey& h2(sort_info2->first);
 342   const Stringpool_char* s1 = h1.string;
 343   const Stringpool_char* s2 = h2.string;
 344   const size_t len1 = h1.length;
 345   const size_t len2 = h2.length;
 346   const size_t minlen = len1 < len2 ? len1 : len2;
 347   const Stringpool_char* p1 = s1 + len1 - 1;
 348   const Stringpool_char* p2 = s2 + len2 - 1;
 349   for (size_t i = minlen; i > 0; --i, --p1, --p2)
 350     {
 351       if (*p1 != *p2)
 352         return *p1 > *p2;
 353     }
 354   return len1 > len2;
 355 }
 356
 357 // Return whether s1 is a suffix of s2.
 358
 359 template<typename Stringpool_char>
 360 bool
 361 Stringpool_template<Stringpool_char>::is_suffix(const Stringpool_char* s1,
 362                                                 size_t len1,
 363                                                 const Stringpool_char* s2,
 364                                                 size_t len2)
 365 {
 366   if (len1 > len2)
 367     return false;
 368   return memcmp(s1, s2 + len2 - len1, len1 * sizeof(Stringpool_char)) == 0;
 369 }
 370
 371 // Turn the stringpool into an ELF strtab: determine the offsets of
 372 // each string in the table.
 373
 374 template<typename Stringpool_char>
 375 void
 376 Stringpool_template<Stringpool_char>::set_string_offsets()
 377 {
 378   if (this->strtab_size_ != 0)
 379     {
 380       // We've already computed the offsets.
 381       return;
 382     }
 383
 384   const size_t charsize = sizeof(Stringpool_char);
 385
 386   // Offset 0 may be reserved for the empty string.
 387   section_offset_type offset = this->zero_null_ ? charsize : 0;
 388
 389   // Sorting to find suffixes can take over 25% of the total CPU time
 390   // used by the linker.  Since it's merely an optimization to reduce
 391   // the strtab size, and gives a relatively small benefit (it's
 392   // typically rare for a symbol to be a suffix of another), we only
 393   // take the time to sort when the user asks for heavy optimization.
 394   if (!this->optimize_)
 395     {
 396       // If we are not optimizing, the offsets are already assigned.
 397       offset = this->offset_;
 398     }
 399   else
 400     {
 401       size_t count = this->string_set_.size();
 402
 403       std::vector<Stringpool_sort_info> v;
 404       v.reserve(count);
 405
 406       for (typename String_set_type::iterator p = this->string_set_.begin();
 407            p != this->string_set_.end();
 408            ++p)
 409         v.push_back(Stringpool_sort_info(p));
 410
 411       std::sort(v.begin(), v.end(), Stringpool_sort_comparison());
 412
 413       section_offset_type last_offset = -1;
 414       for (typename std::vector<Stringpool_sort_info>::iterator last = v.end(),
 415              curr = v.begin();
 416            curr != v.end();
 417            last = curr++)
 418         {
 419           section_offset_type this_offset;
 420           if (this->zero_null_ && (*curr)->first.string[0] == 0)
 421             this_offset = 0;
 422           else if (last != v.end()
 423                    && ((((*curr)->first.length - (*last)->first.length)
 424                         % this->addralign_) == 0)
 425                    && is_suffix((*curr)->first.string,
 426                                 (*curr)->first.length,
 427                                 (*last)->first.string,
 428                                 (*last)->first.length))
 429             this_offset = (last_offset
 430                            + (((*last)->first.length - (*curr)->first.length)
 431                               * charsize));
 432           else
 433             {
 434               this_offset = align_address(offset, this->addralign_);
 435               offset = this_offset + ((*curr)->first.length + 1) * charsize;
 436             }
 437           this->key_to_offset_[(*curr)->second - 1] = this_offset;
 438           last_offset = this_offset;
 439         }
 440     }
 441
 442   this->strtab_size_ = offset;
 443 }
 444
 445 // Get the offset of a string in the ELF strtab.  The string must
 446 // exist.
 447
 448 template<typename Stringpool_char>
 449 section_offset_type
 450 Stringpool_template<Stringpool_char>::get_offset(const Stringpool_char* s)
 451   const
 452 {
 453   return this->get_offset_with_length(s, string_length(s));
 454 }
 455
 456 template<typename Stringpool_char>
 457 section_offset_type
 458 Stringpool_template<Stringpool_char>::get_offset_with_length(
 459     const Stringpool_char* s,
 460     size_t length) const
 461 {
 462   gold_assert(this->strtab_size_ != 0);
 463   Hashkey hk(s, length);
 464   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 465   if (p != this->string_set_.end())
 466     return this->key_to_offset_[p->second - 1];
 467   gold_unreachable();
 468 }
 469
 470 // Write the ELF strtab into the buffer.
 471
 472 template<typename Stringpool_char>
 473 void
 474 Stringpool_template<Stringpool_char>::write_to_buffer(
 475     unsigned char* buffer,
 476     section_size_type bufsize)
 477 {
 478   gold_assert(this->strtab_size_ != 0);
 479   gold_assert(bufsize >= this->strtab_size_);
 480   if (this->zero_null_)
 481     buffer[0] = '\0';
 482   for (typename String_set_type::const_iterator p = this->string_set_.begin();
 483        p != this->string_set_.end();
 484        ++p)
 485     {
 486       const int len = (p->first.length + 1) * sizeof(Stringpool_char);
 487       const section_offset_type offset = this->key_to_offset_[p->second - 1];
 488       gold_assert(static_cast<section_size_type>(offset) + len
 489                   <= this->strtab_size_);
 490       memcpy(buffer + offset, p->first.string, len);
 491     }
 492 }
 493
 494 // Write the ELF strtab into the output file at the specified offset.
 495
 496 template<typename Stringpool_char>
 497 void
 498 Stringpool_template<Stringpool_char>::write(Output_file* of, off_t offset)
 499 {
 500   gold_assert(this->strtab_size_ != 0);
 501   unsigned char* view = of->get_output_view(offset, this->strtab_size_);
 502   this->write_to_buffer(view, this->strtab_size_);
 503   of->write_output_view(offset, this->strtab_size_, view);
 504 }
 505
 506 // Print statistical information to stderr.  This is used for --stats.
 507
 508 template<typename Stringpool_char>
 509 void
 510 Stringpool_template<Stringpool_char>::print_stats(const char* name) const
 511 {
 512 #if defined(HAVE_UNORDERED_MAP) || defined(HAVE_TR1_UNORDERED_MAP) || defined(HAVE_EXT_HASH_MAP)
 513   fprintf(stderr, _("%s: %s entries: %zu; buckets: %zu\n"),
 514           program_name, name, this->string_set_.size(),
 515           this->string_set_.bucket_count());
 516 #else
 517   fprintf(stderr, _("%s: %s entries: %zu\n"),
 518           program_name, name, this->table_.size());
 519 #endif
 520   fprintf(stderr, _("%s: %s Stringdata structures: %zu\n"),
 521           program_name, name, this->strings_.size());
 522 }
 523
 524 // Instantiate the templates we need.
 525
 526 template
 527 class Stringpool_template<char>;
 528
 529 template
 530 class Stringpool_template<uint16_t>;
 531
 532 template
 533 class Stringpool_template<uint32_t>;
 534
 535 } // End namespace gold.