gold/stringpool.cc

   1 // stringpool.cc -- a string pool for gold
   2
   3 // Copyright 2006, 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 <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()
  38   : string_set_(), strings_(), strtab_size_(0), next_index_(1),
  39     next_uncopied_key_(-1), zero_null_(true)
  40 {
  41 }
  42
  43 template<typename Stringpool_char>
  44 void
  45 Stringpool_template<Stringpool_char>::clear()
  46 {
  47   for (typename std::list<Stringdata*>::iterator p = this->strings_.begin();
  48        p != this->strings_.end();
  49        ++p)
  50     delete[] reinterpret_cast<char*>(*p);
  51   this->strings_.clear();
  52   this->string_set_.clear();
  53 }
  54
  55 template<typename Stringpool_char>
  56 Stringpool_template<Stringpool_char>::~Stringpool_template()
  57 {
  58   this->clear();
  59 }
  60
  61 // Resize the internal hashtable with the expectation we'll get n new
  62 // elements.  Note that the hashtable constructor takes a "number of
  63 // buckets you'd like," rather than "number of elements you'd like,"
  64 // but that's the best we can do.
  65
  66 template<typename Stringpool_char>
  67 void
  68 Stringpool_template<Stringpool_char>::reserve(unsigned int n)
  69 {
  70 #if defined(HAVE_TR1_UNORDERED_MAP)
  71   // rehash() implementation is broken in gcc 4.0.3's stl
  72   //this->string_set_.rehash(this->string_set_.size() + n);
  73   //return;
  74 #elif defined(HAVE_EXT_HASH_MAP)
  75   this->string_set_.resize(this->string_set_.size() + n);
  76   return;
  77 #endif
  78
  79   // This is the generic "reserve" code, if no #ifdef above triggers.
  80   String_set_type new_string_set(this->string_set_.size() + n);
  81   new_string_set.insert(this->string_set_.begin(), this->string_set_.end());
  82   this->string_set_.swap(new_string_set);
  83 }
  84
  85 // Return the length of a string of arbitrary character type.
  86
  87 template<typename Stringpool_char>
  88 size_t
  89 Stringpool_template<Stringpool_char>::string_length(const Stringpool_char* p)
  90 {
  91   size_t len = 0;
  92   for (; *p != 0; ++p)
  93     ++len;
  94   return len;
  95 }
  96
  97 // Specialize string_length for char.  Maybe we could just use
  98 // std::char_traits<>::length?
  99
 100 template<>
 101 inline size_t
 102 Stringpool_template<char>::string_length(const char* p)
 103 {
 104   return strlen(p);
 105 }
 106
 107 // Compare two strings of arbitrary character type for equality.
 108
 109 template<typename Stringpool_char>
 110 bool
 111 Stringpool_template<Stringpool_char>::string_equal(const Stringpool_char* s1,
 112                                                    const Stringpool_char* s2)
 113 {
 114   while (*s1 != 0)
 115     if (*s1++ != *s2++)
 116       return false;
 117   return *s2 == 0;
 118 }
 119
 120 // Specialize string_equal for char.
 121
 122 template<>
 123 inline bool
 124 Stringpool_template<char>::string_equal(const char* s1, const char* s2)
 125 {
 126   return strcmp(s1, s2) == 0;
 127 }
 128
 129 // Equality comparison function for the hash table.
 130
 131 template<typename Stringpool_char>
 132 inline bool
 133 Stringpool_template<Stringpool_char>::Stringpool_eq::operator()(
 134     const Hashkey& h1,
 135     const Hashkey& h2) const
 136 {
 137   return (h1.hash_code == h2.hash_code
 138           && h1.length == h2.length
 139           && (h1.string == h2.string
 140               || memcmp(h1.string, h2.string,
 141                         h1.length * sizeof(Stringpool_char)) == 0));
 142 }
 143
 144 // Hash function.  The length is in characters, not bytes.
 145
 146 template<typename Stringpool_char>
 147 size_t
 148 Stringpool_template<Stringpool_char>::string_hash(const Stringpool_char* s,
 149                                                   size_t length)
 150 {
 151   // Fowler/Noll/Vo (FNV) hash (type FNV-1a).
 152   const unsigned char* p = reinterpret_cast<const unsigned char*>(s);
 153   if (sizeof(size_t) > 4)
 154     {
 155       size_t result = static_cast<size_t>(14695981039346656037ULL);
 156       for (size_t i = 0; i < length * sizeof(Stringpool_char); ++i)
 157         {
 158           result ^= static_cast<size_t>(*p++);
 159           result *= 1099511628211ULL;
 160         }
 161       return result;
 162     }
 163   else
 164     {
 165       size_t result = 2166136261UL;
 166       for (size_t i = 0; i < length * sizeof(Stringpool_char); ++i)
 167         {
 168           result ^= static_cast<size_t>(*p++);
 169           result *= 16777619UL;
 170         }
 171       return result;
 172     }
 173 }
 174
 175 // Add the string S to the list of canonical strings.  Return a
 176 // pointer to the canonical string.  If PKEY is not NULL, set *PKEY to
 177 // the key.  LENGTH is the length of S in characters.  Note that S may
 178 // not be NUL terminated.
 179
 180 template<typename Stringpool_char>
 181 const Stringpool_char*
 182 Stringpool_template<Stringpool_char>::add_string(const Stringpool_char* s,
 183                                                  size_t len,
 184                                                  Key* pkey)
 185 {
 186   // We are in trouble if we've already computed the string offsets.
 187   gold_assert(this->strtab_size_ == 0);
 188
 189   // The size we allocate for a new Stringdata.
 190   const size_t buffer_size = 1000;
 191   // The amount we multiply the Stringdata index when calculating the
 192   // key.
 193   const size_t key_mult = 1024;
 194   gold_assert(key_mult >= buffer_size);
 195
 196   // Convert len to the number of bytes we need to allocate, including
 197   // the null character.
 198   len = (len + 1) * sizeof(Stringpool_char);
 199
 200   size_t alc;
 201   bool front = true;
 202   if (len > buffer_size)
 203     {
 204       alc = sizeof(Stringdata) + len;
 205       front = false;
 206     }
 207   else if (this->strings_.empty())
 208     alc = sizeof(Stringdata) + buffer_size;
 209   else
 210     {
 211       Stringdata *psd = this->strings_.front();
 212       if (len > psd->alc - psd->len)
 213         alc = sizeof(Stringdata) + buffer_size;
 214       else
 215         {
 216           char* ret = psd->data + psd->len;
 217           memcpy(ret, s, len - sizeof(Stringpool_char));
 218           memset(ret + len - sizeof(Stringpool_char), 0,
 219                  sizeof(Stringpool_char));
 220
 221           if (pkey != NULL)
 222             *pkey = psd->index * key_mult + psd->len;
 223
 224           psd->len += len;
 225
 226           return reinterpret_cast<const Stringpool_char*>(ret);
 227         }
 228     }
 229
 230   Stringdata *psd = reinterpret_cast<Stringdata*>(new char[alc]);
 231   psd->alc = alc - sizeof(Stringdata);
 232   memcpy(psd->data, s, len - sizeof(Stringpool_char));
 233   memset(psd->data + len - sizeof(Stringpool_char), 0,
 234          sizeof(Stringpool_char));
 235   psd->len = len;
 236   psd->index = this->next_index_;
 237   ++this->next_index_;
 238
 239   if (pkey != NULL)
 240     {
 241       *pkey = psd->index * key_mult;
 242       // Ensure there was no overflow.
 243       gold_assert(*pkey / key_mult == psd->index);
 244     }
 245
 246   if (front)
 247     this->strings_.push_front(psd);
 248   else
 249     this->strings_.push_back(psd);
 250
 251   return reinterpret_cast<const Stringpool_char*>(psd->data);
 252 }
 253
 254 // Add a string to a string pool.
 255
 256 template<typename Stringpool_char>
 257 const Stringpool_char*
 258 Stringpool_template<Stringpool_char>::add(const Stringpool_char* s, bool copy,
 259                                           Key* pkey)
 260 {
 261   return this->add_with_length(s, string_length(s), copy, pkey);
 262 }
 263
 264 template<typename Stringpool_char>
 265 const Stringpool_char*
 266 Stringpool_template<Stringpool_char>::add_with_length(const Stringpool_char* s,
 267                                                       size_t length,
 268                                                       bool copy,
 269                                                       Key* pkey)
 270 {
 271   typedef std::pair<typename String_set_type::iterator, bool> Insert_type;
 272
 273   if (!copy)
 274     {
 275       // When we don't need to copy the string, we can call insert
 276       // directly.
 277
 278       const Key k = this->next_uncopied_key_;
 279       const section_offset_type ozero = 0;
 280       std::pair<Hashkey, Hashval> element(Hashkey(s, length),
 281                                           std::make_pair(k, ozero));
 282
 283       Insert_type ins = this->string_set_.insert(element);
 284
 285       typename String_set_type::const_iterator p = ins.first;
 286
 287       if (ins.second)
 288         {
 289           // We just added the string.  The key value has now been
 290           // used.
 291           --this->next_uncopied_key_;
 292         }
 293       else
 294         {
 295           gold_assert(k != p->second.first);
 296         }
 297
 298       if (pkey != NULL)
 299         *pkey = p->second.first;
 300       return p->first.string;
 301     }
 302
 303   // When we have to copy the string, we look it up twice in the hash
 304   // table.  The problem is that we can't insert S before we
 305   // canonicalize it by copying it into the canonical list. The hash
 306   // code will only be computed once.
 307
 308   Hashkey hk(s, length);
 309   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 310   if (p != this->string_set_.end())
 311     {
 312       if (pkey != NULL)
 313         *pkey = p->second.first;
 314       return p->first.string;
 315     }
 316
 317   Key k;
 318   hk.string = this->add_string(s, length, &k);
 319   // The contents of the string stay the same, so we don't need to
 320   // adjust hk.hash_code or hk.length.
 321
 322   const section_offset_type ozero = 0;
 323   std::pair<Hashkey, Hashval> element(hk, std::make_pair(k, ozero));
 324
 325   Insert_type ins = this->string_set_.insert(element);
 326   gold_assert(ins.second);
 327
 328   if (pkey != NULL)
 329     *pkey = k;
 330   return hk.string;
 331 }
 332
 333 template<typename Stringpool_char>
 334 const Stringpool_char*
 335 Stringpool_template<Stringpool_char>::find(const Stringpool_char* s,
 336                                            Key* pkey) const
 337 {
 338   Hashkey hk(s);
 339   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 340   if (p == this->string_set_.end())
 341     return NULL;
 342
 343   if (pkey != NULL)
 344     *pkey = p->second.first;
 345
 346   return p->first.string;
 347 }
 348
 349 // Comparison routine used when sorting into an ELF strtab.  We want
 350 // to sort this so that when one string is a suffix of another, we
 351 // always see the shorter string immediately after the longer string.
 352 // For example, we want to see these strings in this order:
 353 //   abcd
 354 //   cd
 355 //   d
 356 // When strings are not suffixes, we don't care what order they are
 357 // in, but we need to ensure that suffixes wind up next to each other.
 358 // So we do a reversed lexicographic sort on the reversed string.
 359
 360 template<typename Stringpool_char>
 361 bool
 362 Stringpool_template<Stringpool_char>::Stringpool_sort_comparison::operator()(
 363   const Stringpool_sort_info& sort_info1,
 364   const Stringpool_sort_info& sort_info2) const
 365 {
 366   const Hashkey& h1(sort_info1->first);
 367   const Hashkey& h2(sort_info2->first);
 368   const Stringpool_char* s1 = h1.string;
 369   const Stringpool_char* s2 = h2.string;
 370   const size_t len1 = h1.length;
 371   const size_t len2 = h2.length;
 372   const size_t minlen = len1 < len2 ? len1 : len2;
 373   const Stringpool_char* p1 = s1 + len1 - 1;
 374   const Stringpool_char* p2 = s2 + len2 - 1;
 375   for (size_t i = minlen; i > 0; --i, --p1, --p2)
 376     {
 377       if (*p1 != *p2)
 378         return *p1 > *p2;
 379     }
 380   return len1 > len2;
 381 }
 382
 383 // Return whether s1 is a suffix of s2.
 384
 385 template<typename Stringpool_char>
 386 bool
 387 Stringpool_template<Stringpool_char>::is_suffix(const Stringpool_char* s1,
 388                                                 size_t len1,
 389                                                 const Stringpool_char* s2,
 390                                                 size_t len2)
 391 {
 392   if (len1 > len2)
 393     return false;
 394   return memcmp(s1, s2 + len2 - len1, len1 * sizeof(Stringpool_char)) == 0;
 395 }
 396
 397 // Turn the stringpool into an ELF strtab: determine the offsets of
 398 // each string in the table.
 399
 400 template<typename Stringpool_char>
 401 void
 402 Stringpool_template<Stringpool_char>::set_string_offsets()
 403 {
 404   if (this->strtab_size_ != 0)
 405     {
 406       // We've already computed the offsets.
 407       return;
 408     }
 409
 410   const size_t charsize = sizeof(Stringpool_char);
 411
 412   // Offset 0 may be reserved for the empty string.
 413   section_offset_type offset = this->zero_null_ ? charsize : 0;
 414
 415   // Sorting to find suffixes can take over 25% of the total CPU time
 416   // used by the linker.  Since it's merely an optimization to reduce
 417   // the strtab size, and gives a relatively small benefit (it's
 418   // typically rare for a symbol to be a suffix of another), we only
 419   // take the time to sort when the user asks for heavy optimization.
 420   if (parameters->optimization_level() < 2)
 421     {
 422       for (typename String_set_type::iterator curr = this->string_set_.begin();
 423            curr != this->string_set_.end();
 424            curr++)
 425         {
 426           if (this->zero_null_ && curr->first.string[0] == 0)
 427             curr->second.second = 0;
 428           else
 429             {
 430               curr->second.second = offset;
 431               offset += (curr->first.length + 1) * charsize;
 432             }
 433         }
 434     }
 435   else
 436     {
 437       size_t count = this->string_set_.size();
 438
 439       std::vector<Stringpool_sort_info> v;
 440       v.reserve(count);
 441
 442       for (typename String_set_type::iterator p = this->string_set_.begin();
 443            p != this->string_set_.end();
 444            ++p)
 445         v.push_back(Stringpool_sort_info(p));
 446
 447       std::sort(v.begin(), v.end(), Stringpool_sort_comparison());
 448
 449       for (typename std::vector<Stringpool_sort_info>::iterator last = v.end(),
 450              curr = v.begin();
 451            curr != v.end();
 452            last = curr++)
 453         {
 454           if (this->zero_null_ && (*curr)->first.string[0] == 0)
 455             (*curr)->second.second = 0;
 456           else if (last != v.end()
 457                    && is_suffix((*curr)->first.string,
 458                                 (*curr)->first.length,
 459                                 (*last)->first.string,
 460                                 (*last)->first.length))
 461             (*curr)->second.second = ((*last)->second.second
 462                                       + (((*last)->first.length
 463                                           - (*curr)->first.length)
 464                                          * charsize));
 465           else
 466             {
 467               (*curr)->second.second = offset;
 468               offset += ((*curr)->first.length + 1) * charsize;
 469             }
 470         }
 471     }
 472
 473   this->strtab_size_ = offset;
 474 }
 475
 476 // Get the offset of a string in the ELF strtab.  The string must
 477 // exist.
 478
 479 template<typename Stringpool_char>
 480 section_offset_type
 481 Stringpool_template<Stringpool_char>::get_offset(const Stringpool_char* s)
 482   const
 483 {
 484   return this->get_offset_with_length(s, string_length(s));
 485 }
 486
 487 template<typename Stringpool_char>
 488 section_offset_type
 489 Stringpool_template<Stringpool_char>::get_offset_with_length(
 490     const Stringpool_char* s,
 491     size_t length) const
 492 {
 493   gold_assert(this->strtab_size_ != 0);
 494   Hashkey hk(s, length);
 495   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 496   if (p != this->string_set_.end())
 497     return p->second.second;
 498   gold_unreachable();
 499 }
 500
 501 // Write the ELF strtab into the buffer.
 502
 503 template<typename Stringpool_char>
 504 void
 505 Stringpool_template<Stringpool_char>::write_to_buffer(
 506     unsigned char* buffer,
 507     section_size_type bufsize)
 508 {
 509   gold_assert(this->strtab_size_ != 0);
 510   gold_assert(bufsize >= this->strtab_size_);
 511   if (this->zero_null_)
 512     buffer[0] = '\0';
 513   for (typename String_set_type::const_iterator p = this->string_set_.begin();
 514        p != this->string_set_.end();
 515        ++p)
 516     {
 517       const int len = (p->first.length + 1) * sizeof(Stringpool_char);
 518       gold_assert(static_cast<section_size_type>(p->second.second) + len
 519                   <= this->strtab_size_);
 520       memcpy(buffer + p->second.second, p->first.string, len);
 521     }
 522 }
 523
 524 // Write the ELF strtab into the output file at the specified offset.
 525
 526 template<typename Stringpool_char>
 527 void
 528 Stringpool_template<Stringpool_char>::write(Output_file* of, off_t offset)
 529 {
 530   gold_assert(this->strtab_size_ != 0);
 531   unsigned char* view = of->get_output_view(offset, this->strtab_size_);
 532   this->write_to_buffer(view, this->strtab_size_);
 533   of->write_output_view(offset, this->strtab_size_, view);
 534 }
 535
 536 // Print statistical information to stderr.  This is used for --stats.
 537
 538 template<typename Stringpool_char>
 539 void
 540 Stringpool_template<Stringpool_char>::print_stats(const char* name) const
 541 {
 542 #if defined(HAVE_TR1_UNORDERED_MAP) || defined(HAVE_EXT_HASH_MAP)
 543   fprintf(stderr, _("%s: %s entries: %zu; buckets: %zu\n"),
 544           program_name, name, this->string_set_.size(),
 545           this->string_set_.bucket_count());
 546 #else
 547   fprintf(stderr, _("%s: %s entries: %zu\n"),
 548           program_name, name, this->table_.size());
 549 #endif
 550   fprintf(stderr, _("%s: %s Stringdata structures: %zu\n"),
 551           program_name, name, this->strings_.size());
 552 }
 553
 554 // Instantiate the templates we need.
 555
 556 template
 557 class Stringpool_template<char>;
 558
 559 template
 560 class Stringpool_template<uint16_t>;
 561
 562 template
 563 class Stringpool_template<uint32_t>;
 564
 565 } // End namespace gold.