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)
  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   // This is the hash function used by the dynamic linker for
 155   // DT_GNU_HASH entries.  I compared this to a Fowler/Noll/Vo hash
 156   // for a C++ program with 385,775 global symbols.  This hash
 157   // function was very slightly worse.  However, it is much faster to
 158   // compute.  Overall wall clock time was a win.
 159   const unsigned char* p = reinterpret_cast<const unsigned char*>(s);
 160   size_t h = 5381;
 161   for (size_t i = 0; i < length * sizeof(Stringpool_char); ++i)
 162     h = h * 33 + *p++;
 163   return h;
 164 }
 165
 166 // Add the string S to the list of canonical strings.  Return a
 167 // pointer to the canonical string.  If PKEY is not NULL, set *PKEY to
 168 // the key.  LENGTH is the length of S in characters.  Note that S may
 169 // not be NUL terminated.
 170
 171 template<typename Stringpool_char>
 172 const Stringpool_char*
 173 Stringpool_template<Stringpool_char>::add_string(const Stringpool_char* s,
 174                                                  size_t len)
 175 {
 176   // We are in trouble if we've already computed the string offsets.
 177   gold_assert(this->strtab_size_ == 0);
 178
 179   // The size we allocate for a new Stringdata.
 180   const size_t buffer_size = 1000;
 181   // The amount we multiply the Stringdata index when calculating the
 182   // key.
 183   const size_t key_mult = 1024;
 184   gold_assert(key_mult >= buffer_size);
 185
 186   // Convert len to the number of bytes we need to allocate, including
 187   // the null character.
 188   len = (len + 1) * sizeof(Stringpool_char);
 189
 190   size_t alc;
 191   bool front = true;
 192   if (len > buffer_size)
 193     {
 194       alc = sizeof(Stringdata) + len;
 195       front = false;
 196     }
 197   else if (this->strings_.empty())
 198     alc = sizeof(Stringdata) + buffer_size;
 199   else
 200     {
 201       Stringdata *psd = this->strings_.front();
 202       if (len > psd->alc - psd->len)
 203         alc = sizeof(Stringdata) + buffer_size;
 204       else
 205         {
 206           char* ret = psd->data + psd->len;
 207           memcpy(ret, s, len - sizeof(Stringpool_char));
 208           memset(ret + len - sizeof(Stringpool_char), 0,
 209                  sizeof(Stringpool_char));
 210
 211           psd->len += len;
 212
 213           return reinterpret_cast<const Stringpool_char*>(ret);
 214         }
 215     }
 216
 217   Stringdata *psd = reinterpret_cast<Stringdata*>(new char[alc]);
 218   psd->alc = alc - sizeof(Stringdata);
 219   memcpy(psd->data, s, len - sizeof(Stringpool_char));
 220   memset(psd->data + len - sizeof(Stringpool_char), 0,
 221          sizeof(Stringpool_char));
 222   psd->len = len;
 223
 224   if (front)
 225     this->strings_.push_front(psd);
 226   else
 227     this->strings_.push_back(psd);
 228
 229   return reinterpret_cast<const Stringpool_char*>(psd->data);
 230 }
 231
 232 // Add a string to a string pool.
 233
 234 template<typename Stringpool_char>
 235 const Stringpool_char*
 236 Stringpool_template<Stringpool_char>::add(const Stringpool_char* s, bool copy,
 237                                           Key* pkey)
 238 {
 239   return this->add_with_length(s, string_length(s), copy, pkey);
 240 }
 241
 242 template<typename Stringpool_char>
 243 const Stringpool_char*
 244 Stringpool_template<Stringpool_char>::add_with_length(const Stringpool_char* s,
 245                                                       size_t length,
 246                                                       bool copy,
 247                                                       Key* pkey)
 248 {
 249   typedef std::pair<typename String_set_type::iterator, bool> Insert_type;
 250
 251   // We add 1 so that 0 is always invalid.
 252   const Key k = this->key_to_offset_.size() + 1;
 253
 254   if (!copy)
 255     {
 256       // When we don't need to copy the string, we can call insert
 257       // directly.
 258
 259       std::pair<Hashkey, Hashval> element(Hashkey(s, length), k);
 260
 261       Insert_type ins = this->string_set_.insert(element);
 262
 263       typename String_set_type::const_iterator p = ins.first;
 264
 265       if (ins.second)
 266         {
 267           // We just added the string.  The key value has now been
 268           // used.
 269           this->key_to_offset_.push_back(0);
 270         }
 271       else
 272         {
 273           gold_assert(k != p->second);
 274         }
 275
 276       if (pkey != NULL)
 277         *pkey = p->second;
 278       return p->first.string;
 279     }
 280
 281   // When we have to copy the string, we look it up twice in the hash
 282   // table.  The problem is that we can't insert S before we
 283   // canonicalize it by copying it into the canonical list. The hash
 284   // code will only be computed once.
 285
 286   Hashkey hk(s, length);
 287   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 288   if (p != this->string_set_.end())
 289     {
 290       if (pkey != NULL)
 291         *pkey = p->second;
 292       return p->first.string;
 293     }
 294
 295   this->key_to_offset_.push_back(0);
 296
 297   hk.string = this->add_string(s, length);
 298   // The contents of the string stay the same, so we don't need to
 299   // adjust hk.hash_code or hk.length.
 300
 301   std::pair<Hashkey, Hashval> element(hk, k);
 302
 303   Insert_type ins = this->string_set_.insert(element);
 304   gold_assert(ins.second);
 305
 306   if (pkey != NULL)
 307     *pkey = k;
 308   return hk.string;
 309 }
 310
 311 template<typename Stringpool_char>
 312 const Stringpool_char*
 313 Stringpool_template<Stringpool_char>::find(const Stringpool_char* s,
 314                                            Key* pkey) const
 315 {
 316   Hashkey hk(s);
 317   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 318   if (p == this->string_set_.end())
 319     return NULL;
 320
 321   if (pkey != NULL)
 322     *pkey = p->second;
 323
 324   return p->first.string;
 325 }
 326
 327 // Comparison routine used when sorting into an ELF strtab.  We want
 328 // to sort this so that when one string is a suffix of another, we
 329 // always see the shorter string immediately after the longer string.
 330 // For example, we want to see these strings in this order:
 331 //   abcd
 332 //   cd
 333 //   d
 334 // When strings are not suffixes, we don't care what order they are
 335 // in, but we need to ensure that suffixes wind up next to each other.
 336 // So we do a reversed lexicographic sort on the reversed string.
 337
 338 template<typename Stringpool_char>
 339 bool
 340 Stringpool_template<Stringpool_char>::Stringpool_sort_comparison::operator()(
 341   const Stringpool_sort_info& sort_info1,
 342   const Stringpool_sort_info& sort_info2) const
 343 {
 344   const Hashkey& h1(sort_info1->first);
 345   const Hashkey& h2(sort_info2->first);
 346   const Stringpool_char* s1 = h1.string;
 347   const Stringpool_char* s2 = h2.string;
 348   const size_t len1 = h1.length;
 349   const size_t len2 = h2.length;
 350   const size_t minlen = len1 < len2 ? len1 : len2;
 351   const Stringpool_char* p1 = s1 + len1 - 1;
 352   const Stringpool_char* p2 = s2 + len2 - 1;
 353   for (size_t i = minlen; i > 0; --i, --p1, --p2)
 354     {
 355       if (*p1 != *p2)
 356         return *p1 > *p2;
 357     }
 358   return len1 > len2;
 359 }
 360
 361 // Return whether s1 is a suffix of s2.
 362
 363 template<typename Stringpool_char>
 364 bool
 365 Stringpool_template<Stringpool_char>::is_suffix(const Stringpool_char* s1,
 366                                                 size_t len1,
 367                                                 const Stringpool_char* s2,
 368                                                 size_t len2)
 369 {
 370   if (len1 > len2)
 371     return false;
 372   return memcmp(s1, s2 + len2 - len1, len1 * sizeof(Stringpool_char)) == 0;
 373 }
 374
 375 // Turn the stringpool into an ELF strtab: determine the offsets of
 376 // each string in the table.
 377
 378 template<typename Stringpool_char>
 379 void
 380 Stringpool_template<Stringpool_char>::set_string_offsets()
 381 {
 382   if (this->strtab_size_ != 0)
 383     {
 384       // We've already computed the offsets.
 385       return;
 386     }
 387
 388   const size_t charsize = sizeof(Stringpool_char);
 389
 390   // Offset 0 may be reserved for the empty string.
 391   section_offset_type offset = this->zero_null_ ? charsize : 0;
 392
 393   // Sorting to find suffixes can take over 25% of the total CPU time
 394   // used by the linker.  Since it's merely an optimization to reduce
 395   // the strtab size, and gives a relatively small benefit (it's
 396   // typically rare for a symbol to be a suffix of another), we only
 397   // take the time to sort when the user asks for heavy optimization.
 398   if (parameters->options().optimize() < 2)
 399     {
 400       for (typename String_set_type::iterator curr = this->string_set_.begin();
 401            curr != this->string_set_.end();
 402            curr++)
 403         {
 404           section_offset_type* poff = &this->key_to_offset_[curr->second - 1];
 405           if (this->zero_null_ && curr->first.string[0] == 0)
 406             *poff = 0;
 407           else
 408             {
 409               *poff = offset;
 410               offset += (curr->first.length + 1) * charsize;
 411             }
 412         }
 413     }
 414   else
 415     {
 416       size_t count = this->string_set_.size();
 417
 418       std::vector<Stringpool_sort_info> v;
 419       v.reserve(count);
 420
 421       for (typename String_set_type::iterator p = this->string_set_.begin();
 422            p != this->string_set_.end();
 423            ++p)
 424         v.push_back(Stringpool_sort_info(p));
 425
 426       std::sort(v.begin(), v.end(), Stringpool_sort_comparison());
 427
 428       section_offset_type last_offset = -1;
 429       for (typename std::vector<Stringpool_sort_info>::iterator last = v.end(),
 430              curr = v.begin();
 431            curr != v.end();
 432            last = curr++)
 433         {
 434           section_offset_type this_offset;
 435           if (this->zero_null_ && (*curr)->first.string[0] == 0)
 436             this_offset = 0;
 437           else if (last != v.end()
 438                    && is_suffix((*curr)->first.string,
 439                                 (*curr)->first.length,
 440                                 (*last)->first.string,
 441                                 (*last)->first.length))
 442             this_offset = (last_offset
 443                            + (((*last)->first.length - (*curr)->first.length)
 444                               * charsize));
 445           else
 446             {
 447               this_offset = offset;
 448               offset += ((*curr)->first.length + 1) * charsize;
 449             }
 450           this->key_to_offset_[(*curr)->second - 1] = this_offset;
 451           last_offset = this_offset;
 452         }
 453     }
 454
 455   this->strtab_size_ = offset;
 456 }
 457
 458 // Get the offset of a string in the ELF strtab.  The string must
 459 // exist.
 460
 461 template<typename Stringpool_char>
 462 section_offset_type
 463 Stringpool_template<Stringpool_char>::get_offset(const Stringpool_char* s)
 464   const
 465 {
 466   return this->get_offset_with_length(s, string_length(s));
 467 }
 468
 469 template<typename Stringpool_char>
 470 section_offset_type
 471 Stringpool_template<Stringpool_char>::get_offset_with_length(
 472     const Stringpool_char* s,
 473     size_t length) const
 474 {
 475   gold_assert(this->strtab_size_ != 0);
 476   Hashkey hk(s, length);
 477   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 478   if (p != this->string_set_.end())
 479     return this->key_to_offset_[p->second - 1];
 480   gold_unreachable();
 481 }
 482
 483 // Write the ELF strtab into the buffer.
 484
 485 template<typename Stringpool_char>
 486 void
 487 Stringpool_template<Stringpool_char>::write_to_buffer(
 488     unsigned char* buffer,
 489     section_size_type bufsize)
 490 {
 491   gold_assert(this->strtab_size_ != 0);
 492   gold_assert(bufsize >= this->strtab_size_);
 493   if (this->zero_null_)
 494     buffer[0] = '\0';
 495   for (typename String_set_type::const_iterator p = this->string_set_.begin();
 496        p != this->string_set_.end();
 497        ++p)
 498     {
 499       const int len = (p->first.length + 1) * sizeof(Stringpool_char);
 500       const section_offset_type offset = this->key_to_offset_[p->second - 1];
 501       gold_assert(static_cast<section_size_type>(offset) + len
 502                   <= this->strtab_size_);
 503       memcpy(buffer + offset, p->first.string, len);
 504     }
 505 }
 506
 507 // Write the ELF strtab into the output file at the specified offset.
 508
 509 template<typename Stringpool_char>
 510 void
 511 Stringpool_template<Stringpool_char>::write(Output_file* of, off_t offset)
 512 {
 513   gold_assert(this->strtab_size_ != 0);
 514   unsigned char* view = of->get_output_view(offset, this->strtab_size_);
 515   this->write_to_buffer(view, this->strtab_size_);
 516   of->write_output_view(offset, this->strtab_size_, view);
 517 }
 518
 519 // Print statistical information to stderr.  This is used for --stats.
 520
 521 template<typename Stringpool_char>
 522 void
 523 Stringpool_template<Stringpool_char>::print_stats(const char* name) const
 524 {
 525 #if defined(HAVE_TR1_UNORDERED_MAP) || defined(HAVE_EXT_HASH_MAP)
 526   fprintf(stderr, _("%s: %s entries: %zu; buckets: %zu\n"),
 527           program_name, name, this->string_set_.size(),
 528           this->string_set_.bucket_count());
 529 #else
 530   fprintf(stderr, _("%s: %s entries: %zu\n"),
 531           program_name, name, this->table_.size());
 532 #endif
 533   fprintf(stderr, _("%s: %s Stringdata structures: %zu\n"),
 534           program_name, name, this->strings_.size());
 535 }
 536
 537 // Instantiate the templates we need.
 538
 539 template
 540 class Stringpool_template<char>;
 541
 542 template
 543 class Stringpool_template<uint16_t>;
 544
 545 template
 546 class Stringpool_template<uint32_t>;
 547
 548 } // End namespace gold.