gold/stringpool.cc

   1 // stringpool.cc -- a string pool for gold
   2
   3 // Copyright 2006, 2007, 2008 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), optimize_(false)
  40 {
  41   if (parameters->options_valid() && parameters->options().optimize() >= 2)
  42     this->optimize_ = true;
  43 }
  44
  45 template<typename Stringpool_char>
  46 void
  47 Stringpool_template<Stringpool_char>::clear()
  48 {
  49   for (typename std::list<Stringdata*>::iterator p = this->strings_.begin();
  50        p != this->strings_.end();
  51        ++p)
  52     delete[] reinterpret_cast<char*>(*p);
  53   this->strings_.clear();
  54   this->key_to_offset_.clear();
  55   this->string_set_.clear();
  56 }
  57
  58 template<typename Stringpool_char>
  59 Stringpool_template<Stringpool_char>::~Stringpool_template()
  60 {
  61   this->clear();
  62 }
  63
  64 // Resize the internal hashtable with the expectation we'll get n new
  65 // elements.  Note that the hashtable constructor takes a "number of
  66 // buckets you'd like," rather than "number of elements you'd like,"
  67 // but that's the best we can do.
  68
  69 template<typename Stringpool_char>
  70 void
  71 Stringpool_template<Stringpool_char>::reserve(unsigned int n)
  72 {
  73   this->key_to_offset_.reserve(n);
  74
  75 #if defined(HAVE_TR1_UNORDERED_MAP)
  76   // rehash() implementation is broken in gcc 4.0.3's stl
  77   //this->string_set_.rehash(this->string_set_.size() + n);
  78   //return;
  79 #elif defined(HAVE_EXT_HASH_MAP)
  80   this->string_set_.resize(this->string_set_.size() + n);
  81   return;
  82 #endif
  83
  84   // This is the generic "reserve" code, if no #ifdef above triggers.
  85   String_set_type new_string_set(this->string_set_.size() + n);
  86   new_string_set.insert(this->string_set_.begin(), this->string_set_.end());
  87   this->string_set_.swap(new_string_set);
  88 }
  89
  90 // Return the length of a string of arbitrary character type.
  91
  92 template<typename Stringpool_char>
  93 size_t
  94 Stringpool_template<Stringpool_char>::string_length(const Stringpool_char* p)
  95 {
  96   size_t len = 0;
  97   for (; *p != 0; ++p)
  98     ++len;
  99   return len;
 100 }
 101
 102 // Specialize string_length for char.  Maybe we could just use
 103 // std::char_traits<>::length?
 104
 105 template<>
 106 inline size_t
 107 Stringpool_template<char>::string_length(const char* p)
 108 {
 109   return strlen(p);
 110 }
 111
 112 // Compare two strings of arbitrary character type for equality.
 113
 114 template<typename Stringpool_char>
 115 bool
 116 Stringpool_template<Stringpool_char>::string_equal(const Stringpool_char* s1,
 117                                                    const Stringpool_char* s2)
 118 {
 119   while (*s1 != 0)
 120     if (*s1++ != *s2++)
 121       return false;
 122   return *s2 == 0;
 123 }
 124
 125 // Specialize string_equal for char.
 126
 127 template<>
 128 inline bool
 129 Stringpool_template<char>::string_equal(const char* s1, const char* s2)
 130 {
 131   return strcmp(s1, s2) == 0;
 132 }
 133
 134 // Equality comparison function for the hash table.
 135
 136 template<typename Stringpool_char>
 137 inline bool
 138 Stringpool_template<Stringpool_char>::Stringpool_eq::operator()(
 139     const Hashkey& h1,
 140     const Hashkey& h2) const
 141 {
 142   return (h1.hash_code == h2.hash_code
 143           && h1.length == h2.length
 144           && (h1.string == h2.string
 145               || memcmp(h1.string, h2.string,
 146                         h1.length * sizeof(Stringpool_char)) == 0));
 147 }
 148
 149 // Hash function.  The length is in characters, not bytes.
 150
 151 template<typename Stringpool_char>
 152 size_t
 153 Stringpool_template<Stringpool_char>::string_hash(const Stringpool_char* s,
 154                                                   size_t length)
 155 {
 156   return gold::string_hash<Stringpool_char>(s, length);
 157 }
 158
 159 // Add the string S to the list of canonical strings.  Return a
 160 // pointer to the canonical string.  If PKEY is not NULL, set *PKEY to
 161 // the key.  LENGTH is the length of S in characters.  Note that S may
 162 // not be NUL terminated.
 163
 164 template<typename Stringpool_char>
 165 const Stringpool_char*
 166 Stringpool_template<Stringpool_char>::add_string(const Stringpool_char* s,
 167                                                  size_t len)
 168 {
 169   // We are in trouble if we've already computed the string offsets.
 170   gold_assert(this->strtab_size_ == 0);
 171
 172   // The size we allocate for a new Stringdata.
 173   const size_t buffer_size = 1000;
 174   // The amount we multiply the Stringdata index when calculating the
 175   // key.
 176   const size_t key_mult = 1024;
 177   gold_assert(key_mult >= buffer_size);
 178
 179   // Convert len to the number of bytes we need to allocate, including
 180   // the null character.
 181   len = (len + 1) * sizeof(Stringpool_char);
 182
 183   size_t alc;
 184   bool front = true;
 185   if (len > buffer_size)
 186     {
 187       alc = sizeof(Stringdata) + len;
 188       front = false;
 189     }
 190   else if (this->strings_.empty())
 191     alc = sizeof(Stringdata) + buffer_size;
 192   else
 193     {
 194       Stringdata *psd = this->strings_.front();
 195       if (len > psd->alc - psd->len)
 196         alc = sizeof(Stringdata) + buffer_size;
 197       else
 198         {
 199           char* ret = psd->data + psd->len;
 200           memcpy(ret, s, len - sizeof(Stringpool_char));
 201           memset(ret + len - sizeof(Stringpool_char), 0,
 202                  sizeof(Stringpool_char));
 203
 204           psd->len += len;
 205
 206           return reinterpret_cast<const Stringpool_char*>(ret);
 207         }
 208     }
 209
 210   Stringdata *psd = reinterpret_cast<Stringdata*>(new char[alc]);
 211   psd->alc = alc - sizeof(Stringdata);
 212   memcpy(psd->data, s, len - sizeof(Stringpool_char));
 213   memset(psd->data + len - sizeof(Stringpool_char), 0,
 214          sizeof(Stringpool_char));
 215   psd->len = len;
 216
 217   if (front)
 218     this->strings_.push_front(psd);
 219   else
 220     this->strings_.push_back(psd);
 221
 222   return reinterpret_cast<const Stringpool_char*>(psd->data);
 223 }
 224
 225 // Add a string to a string pool.
 226
 227 template<typename Stringpool_char>
 228 const Stringpool_char*
 229 Stringpool_template<Stringpool_char>::add(const Stringpool_char* s, bool copy,
 230                                           Key* pkey)
 231 {
 232   return this->add_with_length(s, string_length(s), copy, pkey);
 233 }
 234
 235 template<typename Stringpool_char>
 236 const Stringpool_char*
 237 Stringpool_template<Stringpool_char>::add_with_length(const Stringpool_char* s,
 238                                                       size_t length,
 239                                                       bool copy,
 240                                                       Key* pkey)
 241 {
 242   typedef std::pair<typename String_set_type::iterator, bool> Insert_type;
 243
 244   // We add 1 so that 0 is always invalid.
 245   const Key k = this->key_to_offset_.size() + 1;
 246
 247   if (!copy)
 248     {
 249       // When we don't need to copy the string, we can call insert
 250       // directly.
 251
 252       std::pair<Hashkey, Hashval> element(Hashkey(s, length), k);
 253
 254       Insert_type ins = this->string_set_.insert(element);
 255
 256       typename String_set_type::const_iterator p = ins.first;
 257
 258       if (ins.second)
 259         {
 260           // We just added the string.  The key value has now been
 261           // used.
 262           this->key_to_offset_.push_back(0);
 263         }
 264       else
 265         {
 266           gold_assert(k != p->second);
 267         }
 268
 269       if (pkey != NULL)
 270         *pkey = p->second;
 271       return p->first.string;
 272     }
 273
 274   // When we have to copy the string, we look it up twice in the hash
 275   // table.  The problem is that we can't insert S before we
 276   // canonicalize it by copying it into the canonical list. The hash
 277   // code will only be computed once.
 278
 279   Hashkey hk(s, length);
 280   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 281   if (p != this->string_set_.end())
 282     {
 283       if (pkey != NULL)
 284         *pkey = p->second;
 285       return p->first.string;
 286     }
 287
 288   this->key_to_offset_.push_back(0);
 289
 290   hk.string = this->add_string(s, length);
 291   // The contents of the string stay the same, so we don't need to
 292   // adjust hk.hash_code or hk.length.
 293
 294   std::pair<Hashkey, Hashval> element(hk, k);
 295
 296   Insert_type ins = this->string_set_.insert(element);
 297   gold_assert(ins.second);
 298
 299   if (pkey != NULL)
 300     *pkey = k;
 301   return hk.string;
 302 }
 303
 304 template<typename Stringpool_char>
 305 const Stringpool_char*
 306 Stringpool_template<Stringpool_char>::find(const Stringpool_char* s,
 307                                            Key* pkey) const
 308 {
 309   Hashkey hk(s);
 310   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 311   if (p == this->string_set_.end())
 312     return NULL;
 313
 314   if (pkey != NULL)
 315     *pkey = p->second;
 316
 317   return p->first.string;
 318 }
 319
 320 // Comparison routine used when sorting into an ELF strtab.  We want
 321 // to sort this so that when one string is a suffix of another, we
 322 // always see the shorter string immediately after the longer string.
 323 // For example, we want to see these strings in this order:
 324 //   abcd
 325 //   cd
 326 //   d
 327 // When strings are not suffixes, we don't care what order they are
 328 // in, but we need to ensure that suffixes wind up next to each other.
 329 // So we do a reversed lexicographic sort on the reversed string.
 330
 331 template<typename Stringpool_char>
 332 bool
 333 Stringpool_template<Stringpool_char>::Stringpool_sort_comparison::operator()(
 334   const Stringpool_sort_info& sort_info1,
 335   const Stringpool_sort_info& sort_info2) const
 336 {
 337   const Hashkey& h1(sort_info1->first);
 338   const Hashkey& h2(sort_info2->first);
 339   const Stringpool_char* s1 = h1.string;
 340   const Stringpool_char* s2 = h2.string;
 341   const size_t len1 = h1.length;
 342   const size_t len2 = h2.length;
 343   const size_t minlen = len1 < len2 ? len1 : len2;
 344   const Stringpool_char* p1 = s1 + len1 - 1;
 345   const Stringpool_char* p2 = s2 + len2 - 1;
 346   for (size_t i = minlen; i > 0; --i, --p1, --p2)
 347     {
 348       if (*p1 != *p2)
 349         return *p1 > *p2;
 350     }
 351   return len1 > len2;
 352 }
 353
 354 // Return whether s1 is a suffix of s2.
 355
 356 template<typename Stringpool_char>
 357 bool
 358 Stringpool_template<Stringpool_char>::is_suffix(const Stringpool_char* s1,
 359                                                 size_t len1,
 360                                                 const Stringpool_char* s2,
 361                                                 size_t len2)
 362 {
 363   if (len1 > len2)
 364     return false;
 365   return memcmp(s1, s2 + len2 - len1, len1 * sizeof(Stringpool_char)) == 0;
 366 }
 367
 368 // Turn the stringpool into an ELF strtab: determine the offsets of
 369 // each string in the table.
 370
 371 template<typename Stringpool_char>
 372 void
 373 Stringpool_template<Stringpool_char>::set_string_offsets()
 374 {
 375   if (this->strtab_size_ != 0)
 376     {
 377       // We've already computed the offsets.
 378       return;
 379     }
 380
 381   const size_t charsize = sizeof(Stringpool_char);
 382
 383   // Offset 0 may be reserved for the empty string.
 384   section_offset_type offset = this->zero_null_ ? charsize : 0;
 385
 386   // Sorting to find suffixes can take over 25% of the total CPU time
 387   // used by the linker.  Since it's merely an optimization to reduce
 388   // the strtab size, and gives a relatively small benefit (it's
 389   // typically rare for a symbol to be a suffix of another), we only
 390   // take the time to sort when the user asks for heavy optimization.
 391   if (!this->optimize_)
 392     {
 393       for (typename String_set_type::iterator curr = this->string_set_.begin();
 394            curr != this->string_set_.end();
 395            curr++)
 396         {
 397           section_offset_type* poff = &this->key_to_offset_[curr->second - 1];
 398           if (this->zero_null_ && curr->first.string[0] == 0)
 399             *poff = 0;
 400           else
 401             {
 402               *poff = offset;
 403               offset += (curr->first.length + 1) * charsize;
 404             }
 405         }
 406     }
 407   else
 408     {
 409       size_t count = this->string_set_.size();
 410
 411       std::vector<Stringpool_sort_info> v;
 412       v.reserve(count);
 413
 414       for (typename String_set_type::iterator p = this->string_set_.begin();
 415            p != this->string_set_.end();
 416            ++p)
 417         v.push_back(Stringpool_sort_info(p));
 418
 419       std::sort(v.begin(), v.end(), Stringpool_sort_comparison());
 420
 421       section_offset_type last_offset = -1;
 422       for (typename std::vector<Stringpool_sort_info>::iterator last = v.end(),
 423              curr = v.begin();
 424            curr != v.end();
 425            last = curr++)
 426         {
 427           section_offset_type this_offset;
 428           if (this->zero_null_ && (*curr)->first.string[0] == 0)
 429             this_offset = 0;
 430           else if (last != v.end()
 431                    && is_suffix((*curr)->first.string,
 432                                 (*curr)->first.length,
 433                                 (*last)->first.string,
 434                                 (*last)->first.length))
 435             this_offset = (last_offset
 436                            + (((*last)->first.length - (*curr)->first.length)
 437                               * charsize));
 438           else
 439             {
 440               this_offset = offset;
 441               offset += ((*curr)->first.length + 1) * charsize;
 442             }
 443           this->key_to_offset_[(*curr)->second - 1] = this_offset;
 444           last_offset = this_offset;
 445         }
 446     }
 447
 448   this->strtab_size_ = offset;
 449 }
 450
 451 // Get the offset of a string in the ELF strtab.  The string must
 452 // exist.
 453
 454 template<typename Stringpool_char>
 455 section_offset_type
 456 Stringpool_template<Stringpool_char>::get_offset(const Stringpool_char* s)
 457   const
 458 {
 459   return this->get_offset_with_length(s, string_length(s));
 460 }
 461
 462 template<typename Stringpool_char>
 463 section_offset_type
 464 Stringpool_template<Stringpool_char>::get_offset_with_length(
 465     const Stringpool_char* s,
 466     size_t length) const
 467 {
 468   gold_assert(this->strtab_size_ != 0);
 469   Hashkey hk(s, length);
 470   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 471   if (p != this->string_set_.end())
 472     return this->key_to_offset_[p->second - 1];
 473   gold_unreachable();
 474 }
 475
 476 // Write the ELF strtab into the buffer.
 477
 478 template<typename Stringpool_char>
 479 void
 480 Stringpool_template<Stringpool_char>::write_to_buffer(
 481     unsigned char* buffer,
 482     section_size_type bufsize)
 483 {
 484   gold_assert(this->strtab_size_ != 0);
 485   gold_assert(bufsize >= this->strtab_size_);
 486   if (this->zero_null_)
 487     buffer[0] = '\0';
 488   for (typename String_set_type::const_iterator p = this->string_set_.begin();
 489        p != this->string_set_.end();
 490        ++p)
 491     {
 492       const int len = (p->first.length + 1) * sizeof(Stringpool_char);
 493       const section_offset_type offset = this->key_to_offset_[p->second - 1];
 494       gold_assert(static_cast<section_size_type>(offset) + len
 495                   <= this->strtab_size_);
 496       memcpy(buffer + offset, p->first.string, len);
 497     }
 498 }
 499
 500 // Write the ELF strtab into the output file at the specified offset.
 501
 502 template<typename Stringpool_char>
 503 void
 504 Stringpool_template<Stringpool_char>::write(Output_file* of, off_t offset)
 505 {
 506   gold_assert(this->strtab_size_ != 0);
 507   unsigned char* view = of->get_output_view(offset, this->strtab_size_);
 508   this->write_to_buffer(view, this->strtab_size_);
 509   of->write_output_view(offset, this->strtab_size_, view);
 510 }
 511
 512 // Print statistical information to stderr.  This is used for --stats.
 513
 514 template<typename Stringpool_char>
 515 void
 516 Stringpool_template<Stringpool_char>::print_stats(const char* name) const
 517 {
 518 #if defined(HAVE_TR1_UNORDERED_MAP) || defined(HAVE_EXT_HASH_MAP)
 519   fprintf(stderr, _("%s: %s entries: %zu; buckets: %zu\n"),
 520           program_name, name, this->string_set_.size(),
 521           this->string_set_.bucket_count());
 522 #else
 523   fprintf(stderr, _("%s: %s entries: %zu\n"),
 524           program_name, name, this->table_.size());
 525 #endif
 526   fprintf(stderr, _("%s: %s Stringdata structures: %zu\n"),
 527           program_name, name, this->strings_.size());
 528 }
 529
 530 // Instantiate the templates we need.
 531
 532 template
 533 class Stringpool_template<char>;
 534
 535 template
 536 class Stringpool_template<uint16_t>;
 537
 538 template
 539 class Stringpool_template<uint32_t>;
 540
 541 } // End namespace gold.