gold/stringpool.cc

   1 // stringpool.cc -- a string pool for gold
   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 #include "gold.h"
  24
  25 #include <cstring>
  26 #include <algorithm>
  27 #include <vector>
  28 #include <uchar.h>
  29
  30 #include "output.h"
  31 #include "parameters.h"
  32 #include "stringpool.h"
  33
  34 namespace gold
  35 {
  36
  37 template<typename Stringpool_char>
  38 Stringpool_template<Stringpool_char>::Stringpool_template(uint64_t addralign)
  39   : string_set_(), key_to_offset_(), strings_(), strtab_size_(0),
  40     zero_null_(true), optimize_(false), offset_(sizeof(Stringpool_char)),
  41     addralign_(addralign)
  42 {
  43   if (parameters->options_valid()
  44       && parameters->options().optimize() >= 2
  45       && addralign <= sizeof(Stringpool_char))
  46     this->optimize_ = true;
  47 }
  48
  49 template<typename Stringpool_char>
  50 void
  51 Stringpool_template<Stringpool_char>::clear()
  52 {
  53   for (typename std::list<Stringdata*>::iterator p = this->strings_.begin();
  54        p != this->strings_.end();
  55        ++p)
  56     delete[] reinterpret_cast<char*>(*p);
  57   this->strings_.clear();
  58   this->key_to_offset_.clear();
  59   this->string_set_.clear();
  60 }
  61
  62 template<typename Stringpool_char>
  63 Stringpool_template<Stringpool_char>::~Stringpool_template()
  64 {
  65   this->clear();
  66 }
  67
  68 // Resize the internal hashtable with the expectation we'll get n new
  69 // elements.  Note that the hashtable constructor takes a "number of
  70 // buckets you'd like," rather than "number of elements you'd like,"
  71 // but that's the best we can do.
  72
  73 template<typename Stringpool_char>
  74 void
  75 Stringpool_template<Stringpool_char>::reserve(unsigned int n)
  76 {
  77   this->key_to_offset_.reserve(n);
  78
  79 #if defined(HAVE_UNORDERED_MAP)
  80   this->string_set_.rehash(this->string_set_.size() + n);
  81   return;
  82 #elif defined(HAVE_TR1_UNORDERED_MAP)
  83   // rehash() implementation is broken in gcc 4.0.3's stl
  84   //this->string_set_.rehash(this->string_set_.size() + n);
  85   //return;
  86 #elif defined(HAVE_EXT_HASH_MAP)
  87   this->string_set_.resize(this->string_set_.size() + n);
  88   return;
  89 #endif
  90
  91   // This is the generic "reserve" code, if no #ifdef above triggers.
  92   String_set_type new_string_set(this->string_set_.size() + n);
  93   new_string_set.insert(this->string_set_.begin(), this->string_set_.end());
  94   this->string_set_.swap(new_string_set);
  95 }
  96
  97 // Compare two strings of arbitrary character type for equality.
  98
  99 template<typename Stringpool_char>
 100 bool
 101 Stringpool_template<Stringpool_char>::string_equal(const Stringpool_char* s1,
 102                                                    const Stringpool_char* s2)
 103 {
 104   while (*s1 != 0)
 105     if (*s1++ != *s2++)
 106       return false;
 107   return *s2 == 0;
 108 }
 109
 110 // Specialize string_equal for char.
 111
 112 template<>
 113 inline bool
 114 Stringpool_template<char>::string_equal(const char* s1, const char* s2)
 115 {
 116   return strcmp(s1, s2) == 0;
 117 }
 118
 119 // Equality comparison function for the hash table.
 120
 121 template<typename Stringpool_char>
 122 inline bool
 123 Stringpool_template<Stringpool_char>::Stringpool_eq::operator()(
 124     const Hashkey& h1,
 125     const Hashkey& h2) const
 126 {
 127   return (h1.hash_code == h2.hash_code
 128           && h1.length == h2.length
 129           && (h1.string == h2.string
 130               || memcmp(h1.string, h2.string,
 131                         h1.length * sizeof(Stringpool_char)) == 0));
 132 }
 133
 134 // Hash function.  The length is in characters, not bytes.
 135
 136 template<typename Stringpool_char>
 137 size_t
 138 Stringpool_template<Stringpool_char>::string_hash(const Stringpool_char* s,
 139                                                   size_t length)
 140 {
 141   return gold::string_hash<Stringpool_char>(s, length);
 142 }
 143
 144 // Add the string S to the list of canonical strings.  Return a
 145 // pointer to the canonical string.  If PKEY is not NULL, set *PKEY to
 146 // the key.  LENGTH is the length of S in characters.  Note that S may
 147 // not be NUL terminated.
 148
 149 template<typename Stringpool_char>
 150 const Stringpool_char*
 151 Stringpool_template<Stringpool_char>::add_string(const Stringpool_char* s,
 152                                                  size_t len)
 153 {
 154   // We are in trouble if we've already computed the string offsets.
 155   gold_assert(this->strtab_size_ == 0);
 156
 157   // The size we allocate for a new Stringdata.
 158   const size_t buffer_size = 1000;
 159   // The amount we multiply the Stringdata index when calculating the
 160   // key.
 161   const size_t key_mult = 1024;
 162   gold_assert(key_mult >= buffer_size);
 163
 164   // Convert len to the number of bytes we need to allocate, including
 165   // the null character.
 166   len = (len + 1) * sizeof(Stringpool_char);
 167
 168   size_t alc;
 169   bool front = true;
 170   if (len > buffer_size)
 171     {
 172       alc = sizeof(Stringdata) + len;
 173       front = false;
 174     }
 175   else if (this->strings_.empty())
 176     alc = sizeof(Stringdata) + buffer_size;
 177   else
 178     {
 179       Stringdata* psd = this->strings_.front();
 180       if (len > psd->alc - psd->len)
 181         alc = sizeof(Stringdata) + buffer_size;
 182       else
 183         {
 184           char* ret = psd->data + psd->len;
 185           memcpy(ret, s, len - sizeof(Stringpool_char));
 186           memset(ret + len - sizeof(Stringpool_char), 0,
 187                  sizeof(Stringpool_char));
 188
 189           psd->len += len;
 190
 191           return reinterpret_cast<const Stringpool_char*>(ret);
 192         }
 193     }
 194
 195   Stringdata* psd = reinterpret_cast<Stringdata*>(new char[alc]);
 196   psd->alc = alc - sizeof(Stringdata);
 197   memcpy(psd->data, s, len - sizeof(Stringpool_char));
 198   memset(psd->data + len - sizeof(Stringpool_char), 0,
 199          sizeof(Stringpool_char));
 200   psd->len = len;
 201
 202   if (front)
 203     this->strings_.push_front(psd);
 204   else
 205     this->strings_.push_back(psd);
 206
 207   return reinterpret_cast<const Stringpool_char*>(psd->data);
 208 }
 209
 210 // Add a string to a string pool.
 211
 212 template<typename Stringpool_char>
 213 const Stringpool_char*
 214 Stringpool_template<Stringpool_char>::add(const Stringpool_char* s, bool copy,
 215                                           Key* pkey)
 216 {
 217   return this->add_with_length(s, string_length(s), copy, pkey);
 218 }
 219
 220 // Add a new key offset entry.
 221
 222 template<typename Stringpool_char>
 223 void
 224 Stringpool_template<Stringpool_char>::new_key_offset(size_t length)
 225 {
 226   section_offset_type offset;
 227   if (this->zero_null_ && length == 0)
 228     offset = 0;
 229   else
 230     {
 231       offset = this->offset_;
 232       // Align strings.
 233       offset = align_address(offset, this->addralign_);
 234       this->offset_ = offset + (length + 1) * sizeof(Stringpool_char);
 235     }
 236   this->key_to_offset_.push_back(offset);
 237 }
 238
 239 template<typename Stringpool_char>
 240 const Stringpool_char*
 241 Stringpool_template<Stringpool_char>::add_with_length(const Stringpool_char* s,
 242                                                       size_t length,
 243                                                       bool copy,
 244                                                       Key* pkey)
 245 {
 246   typedef std::pair<typename String_set_type::iterator, bool> Insert_type;
 247
 248   // We add 1 so that 0 is always invalid.
 249   const Key k = this->key_to_offset_.size() + 1;
 250
 251   if (!copy)
 252     {
 253       // When we don't need to copy the string, we can call insert
 254       // directly.
 255
 256       std::pair<Hashkey, Hashval> element(Hashkey(s, length), k);
 257
 258       Insert_type ins = this->string_set_.insert(element);
 259
 260       typename String_set_type::const_iterator p = ins.first;
 261
 262       if (ins.second)
 263         {
 264           // We just added the string.  The key value has now been
 265           // used.
 266           this->new_key_offset(length);
 267         }
 268       else
 269         {
 270           gold_assert(k != p->second);
 271         }
 272
 273       if (pkey != NULL)
 274         *pkey = p->second;
 275       return p->first.string;
 276     }
 277
 278   // When we have to copy the string, we look it up twice in the hash
 279   // table.  The problem is that we can't insert S before we
 280   // canonicalize it by copying it into the canonical list. The hash
 281   // code will only be computed once.
 282
 283   Hashkey hk(s, length);
 284   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 285   if (p != this->string_set_.end())
 286     {
 287       if (pkey != NULL)
 288         *pkey = p->second;
 289       return p->first.string;
 290     }
 291
 292   this->new_key_offset(length);
 293
 294   hk.string = this->add_string(s, length);
 295   // The contents of the string stay the same, so we don't need to
 296   // adjust hk.hash_code or hk.length.
 297
 298   std::pair<Hashkey, Hashval> element(hk, k);
 299
 300   Insert_type ins = this->string_set_.insert(element);
 301   gold_assert(ins.second);
 302
 303   if (pkey != NULL)
 304     *pkey = k;
 305   return hk.string;
 306 }
 307
 308 template<typename Stringpool_char>
 309 const Stringpool_char*
 310 Stringpool_template<Stringpool_char>::find(const Stringpool_char* s,
 311                                            Key* pkey) const
 312 {
 313   Hashkey hk(s);
 314   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 315   if (p == this->string_set_.end())
 316     return NULL;
 317
 318   if (pkey != NULL)
 319     *pkey = p->second;
 320
 321   return p->first.string;
 322 }
 323
 324 // Comparison routine used when sorting into an ELF strtab.  We want
 325 // to sort this so that when one string is a suffix of another, we
 326 // always see the shorter string immediately after the longer string.
 327 // For example, we want to see these strings in this order:
 328 //   abcd
 329 //   cd
 330 //   d
 331 // When strings are not suffixes, we don't care what order they are
 332 // in, but we need to ensure that suffixes wind up next to each other.
 333 // So we do a reversed lexicographic sort on the reversed string.
 334
 335 template<typename Stringpool_char>
 336 bool
 337 Stringpool_template<Stringpool_char>::Stringpool_sort_comparison::operator()(
 338   const Stringpool_sort_info& sort_info1,
 339   const Stringpool_sort_info& sort_info2) const
 340 {
 341   const Hashkey& h1(sort_info1->first);
 342   const Hashkey& h2(sort_info2->first);
 343   const Stringpool_char* s1 = h1.string;
 344   const Stringpool_char* s2 = h2.string;
 345   const size_t len1 = h1.length;
 346   const size_t len2 = h2.length;
 347   const size_t minlen = len1 < len2 ? len1 : len2;
 348   const Stringpool_char* p1 = s1 + len1 - 1;
 349   const Stringpool_char* p2 = s2 + len2 - 1;
 350   for (size_t i = minlen; i > 0; --i, --p1, --p2)
 351     {
 352       if (*p1 != *p2)
 353         return *p1 > *p2;
 354     }
 355   return len1 > len2;
 356 }
 357
 358 // Return whether s1 is a suffix of s2.
 359
 360 template<typename Stringpool_char>
 361 bool
 362 Stringpool_template<Stringpool_char>::is_suffix(const Stringpool_char* s1,
 363                                                 size_t len1,
 364                                                 const Stringpool_char* s2,
 365                                                 size_t len2)
 366 {
 367   if (len1 > len2)
 368     return false;
 369   return memcmp(s1, s2 + len2 - len1, len1 * sizeof(Stringpool_char)) == 0;
 370 }
 371
 372 // Turn the stringpool into an ELF strtab: determine the offsets of
 373 // each string in the table.
 374
 375 template<typename Stringpool_char>
 376 void
 377 Stringpool_template<Stringpool_char>::set_string_offsets()
 378 {
 379   if (this->strtab_size_ != 0)
 380     {
 381       // We've already computed the offsets.
 382       return;
 383     }
 384
 385   const size_t charsize = sizeof(Stringpool_char);
 386
 387   // Offset 0 may be reserved for the empty string.
 388   section_offset_type offset = this->zero_null_ ? charsize : 0;
 389
 390   // Sorting to find suffixes can take over 25% of the total CPU time
 391   // used by the linker.  Since it's merely an optimization to reduce
 392   // the strtab size, and gives a relatively small benefit (it's
 393   // typically rare for a symbol to be a suffix of another), we only
 394   // take the time to sort when the user asks for heavy optimization.
 395   if (!this->optimize_)
 396     {
 397       // If we are not optimizing, the offsets are already assigned.
 398       offset = this->offset_;
 399     }
 400   else
 401     {
 402       size_t count = this->string_set_.size();
 403
 404       std::vector<Stringpool_sort_info> v;
 405       v.reserve(count);
 406
 407       for (typename String_set_type::iterator p = this->string_set_.begin();
 408            p != this->string_set_.end();
 409            ++p)
 410         v.push_back(Stringpool_sort_info(p));
 411
 412       std::sort(v.begin(), v.end(), Stringpool_sort_comparison());
 413
 414       section_offset_type last_offset = -1;
 415       for (typename std::vector<Stringpool_sort_info>::iterator last = v.end(),
 416              curr = v.begin();
 417            curr != v.end();
 418            last = curr++)
 419         {
 420           section_offset_type this_offset;
 421           if (this->zero_null_ && (*curr)->first.string[0] == 0)
 422             this_offset = 0;
 423           else if (last != v.end()
 424                    && ((((*curr)->first.length - (*last)->first.length)
 425                         % this->addralign_) == 0)
 426                    && is_suffix((*curr)->first.string,
 427                                 (*curr)->first.length,
 428                                 (*last)->first.string,
 429                                 (*last)->first.length))
 430             this_offset = (last_offset
 431                            + (((*last)->first.length - (*curr)->first.length)
 432                               * charsize));
 433           else
 434             {
 435               this_offset = align_address(offset, this->addralign_);
 436               offset = this_offset + ((*curr)->first.length + 1) * charsize;
 437             }
 438           this->key_to_offset_[(*curr)->second - 1] = this_offset;
 439           last_offset = this_offset;
 440         }
 441     }
 442
 443   this->strtab_size_ = offset;
 444 }
 445
 446 // Get the offset of a string in the ELF strtab.  The string must
 447 // exist.
 448
 449 template<typename Stringpool_char>
 450 section_offset_type
 451 Stringpool_template<Stringpool_char>::get_offset(const Stringpool_char* s)
 452   const
 453 {
 454   return this->get_offset_with_length(s, string_length(s));
 455 }
 456
 457 template<typename Stringpool_char>
 458 section_offset_type
 459 Stringpool_template<Stringpool_char>::get_offset_with_length(
 460     const Stringpool_char* s,
 461     size_t length) const
 462 {
 463   gold_assert(this->strtab_size_ != 0);
 464   Hashkey hk(s, length);
 465   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 466   if (p != this->string_set_.end())
 467     return this->key_to_offset_[p->second - 1];
 468   gold_unreachable();
 469 }
 470
 471 // Write the ELF strtab into the buffer.
 472
 473 template<typename Stringpool_char>
 474 void
 475 Stringpool_template<Stringpool_char>::write_to_buffer(
 476     unsigned char* buffer,
 477     section_size_type bufsize)
 478 {
 479   gold_assert(this->strtab_size_ != 0);
 480   gold_assert(bufsize >= this->strtab_size_);
 481   if (this->zero_null_)
 482     buffer[0] = '\0';
 483   for (typename String_set_type::const_iterator p = this->string_set_.begin();
 484        p != this->string_set_.end();
 485        ++p)
 486     {
 487       const int len = (p->first.length + 1) * sizeof(Stringpool_char);
 488       const section_offset_type offset = this->key_to_offset_[p->second - 1];
 489       gold_assert(static_cast<section_size_type>(offset) + len
 490                   <= this->strtab_size_);
 491       memcpy(buffer + offset, p->first.string, len);
 492     }
 493 }
 494
 495 // Write the ELF strtab into the output file at the specified offset.
 496
 497 template<typename Stringpool_char>
 498 void
 499 Stringpool_template<Stringpool_char>::write(Output_file* of, off_t offset)
 500 {
 501   gold_assert(this->strtab_size_ != 0);
 502   unsigned char* view = of->get_output_view(offset, this->strtab_size_);
 503   this->write_to_buffer(view, this->strtab_size_);
 504   of->write_output_view(offset, this->strtab_size_, view);
 505 }
 506
 507 // Print statistical information to stderr.  This is used for --stats.
 508
 509 template<typename Stringpool_char>
 510 void
 511 Stringpool_template<Stringpool_char>::print_stats(const char* name) const
 512 {
 513 #if defined(HAVE_UNORDERED_MAP) || defined(HAVE_TR1_UNORDERED_MAP) || defined(HAVE_EXT_HASH_MAP)
 514   fprintf(stderr, _("%s: %s entries: %zu; buckets: %zu\n"),
 515           program_name, name, this->string_set_.size(),
 516           this->string_set_.bucket_count());
 517 #else
 518   fprintf(stderr, _("%s: %s entries: %zu\n"),
 519           program_name, name, this->table_.size());
 520 #endif
 521   fprintf(stderr, _("%s: %s Stringdata structures: %zu\n"),
 522           program_name, name, this->strings_.size());
 523 }
 524
 525 // Instantiate the templates we need.
 526
 527 template
 528 class Stringpool_template<char>;
 529
 530 template
 531 class Stringpool_template<char16_t>;
 532
 533 template
 534 class Stringpool_template<char32_t>;
 535
 536 } // End namespace gold.