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