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