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(uint64_t addralign)
  38   : string_set_(), key_to_offset_(), strings_(), strtab_size_(0),
  39     zero_null_(true), optimize_(false), offset_(sizeof(Stringpool_char)),
  40     addralign_(addralign)
  41 {
  42   if (parameters->options_valid() && parameters->options().optimize() >= 2)
  43     this->optimize_ = true;
  44 }
  45
  46 template<typename Stringpool_char>
  47 void
  48 Stringpool_template<Stringpool_char>::clear()
  49 {
  50   for (typename std::list<Stringdata*>::iterator p = this->strings_.begin();
  51        p != this->strings_.end();
  52        ++p)
  53     delete[] reinterpret_cast<char*>(*p);
  54   this->strings_.clear();
  55   this->key_to_offset_.clear();
  56   this->string_set_.clear();
  57 }
  58
  59 template<typename Stringpool_char>
  60 Stringpool_template<Stringpool_char>::~Stringpool_template()
  61 {
  62   this->clear();
  63 }
  64
  65 // Resize the internal hashtable with the expectation we'll get n new
  66 // elements.  Note that the hashtable constructor takes a "number of
  67 // buckets you'd like," rather than "number of elements you'd like,"
  68 // but that's the best we can do.
  69
  70 template<typename Stringpool_char>
  71 void
  72 Stringpool_template<Stringpool_char>::reserve(unsigned int n)
  73 {
  74   this->key_to_offset_.reserve(n);
  75
  76 #if defined(HAVE_UNORDERED_MAP)
  77   this->string_set_.rehash(this->string_set_.size() + n);
  78   return;
  79 #elif defined(HAVE_TR1_UNORDERED_MAP)
  80   // rehash() implementation is broken in gcc 4.0.3's stl
  81   //this->string_set_.rehash(this->string_set_.size() + n);
  82   //return;
  83 #elif defined(HAVE_EXT_HASH_MAP)
  84   this->string_set_.resize(this->string_set_.size() + n);
  85   return;
  86 #endif
  87
  88   // This is the generic "reserve" code, if no #ifdef above triggers.
  89   String_set_type new_string_set(this->string_set_.size() + n);
  90   new_string_set.insert(this->string_set_.begin(), this->string_set_.end());
  91   this->string_set_.swap(new_string_set);
  92 }
  93
  94 // Compare two strings of arbitrary character type for equality.
  95
  96 template<typename Stringpool_char>
  97 bool
  98 Stringpool_template<Stringpool_char>::string_equal(const Stringpool_char* s1,
  99                                                    const Stringpool_char* s2)
 100 {
 101   while (*s1 != 0)
 102     if (*s1++ != *s2++)
 103       return false;
 104   return *s2 == 0;
 105 }
 106
 107 // Specialize string_equal for char.
 108
 109 template<>
 110 inline bool
 111 Stringpool_template<char>::string_equal(const char* s1, const char* s2)
 112 {
 113   return strcmp(s1, s2) == 0;
 114 }
 115
 116 // Equality comparison function for the hash table.
 117
 118 template<typename Stringpool_char>
 119 inline bool
 120 Stringpool_template<Stringpool_char>::Stringpool_eq::operator()(
 121     const Hashkey& h1,
 122     const Hashkey& h2) const
 123 {
 124   return (h1.hash_code == h2.hash_code
 125           && h1.length == h2.length
 126           && (h1.string == h2.string
 127               || memcmp(h1.string, h2.string,
 128                         h1.length * sizeof(Stringpool_char)) == 0));
 129 }
 130
 131 // Hash function.  The length is in characters, not bytes.
 132
 133 template<typename Stringpool_char>
 134 size_t
 135 Stringpool_template<Stringpool_char>::string_hash(const Stringpool_char* s,
 136                                                   size_t length)
 137 {
 138   return gold::string_hash<Stringpool_char>(s, length);
 139 }
 140
 141 // Add the string S to the list of canonical strings.  Return a
 142 // pointer to the canonical string.  If PKEY is not NULL, set *PKEY to
 143 // the key.  LENGTH is the length of S in characters.  Note that S may
 144 // not be NUL terminated.
 145
 146 template<typename Stringpool_char>
 147 const Stringpool_char*
 148 Stringpool_template<Stringpool_char>::add_string(const Stringpool_char* s,
 149                                                  size_t len)
 150 {
 151   // We are in trouble if we've already computed the string offsets.
 152   gold_assert(this->strtab_size_ == 0);
 153
 154   // The size we allocate for a new Stringdata.
 155   const size_t buffer_size = 1000;
 156   // The amount we multiply the Stringdata index when calculating the
 157   // key.
 158   const size_t key_mult = 1024;
 159   gold_assert(key_mult >= buffer_size);
 160
 161   // Convert len to the number of bytes we need to allocate, including
 162   // the null character.
 163   len = (len + 1) * sizeof(Stringpool_char);
 164
 165   size_t alc;
 166   bool front = true;
 167   if (len > buffer_size)
 168     {
 169       alc = sizeof(Stringdata) + len;
 170       front = false;
 171     }
 172   else if (this->strings_.empty())
 173     alc = sizeof(Stringdata) + buffer_size;
 174   else
 175     {
 176       Stringdata* psd = this->strings_.front();
 177       if (len > psd->alc - psd->len)
 178         alc = sizeof(Stringdata) + buffer_size;
 179       else
 180         {
 181           char* ret = psd->data + psd->len;
 182           memcpy(ret, s, len - sizeof(Stringpool_char));
 183           memset(ret + len - sizeof(Stringpool_char), 0,
 184                  sizeof(Stringpool_char));
 185
 186           psd->len += len;
 187
 188           return reinterpret_cast<const Stringpool_char*>(ret);
 189         }
 190     }
 191
 192   Stringdata* psd = reinterpret_cast<Stringdata*>(new char[alc]);
 193   psd->alc = alc - sizeof(Stringdata);
 194   memcpy(psd->data, s, len - sizeof(Stringpool_char));
 195   memset(psd->data + len - sizeof(Stringpool_char), 0,
 196          sizeof(Stringpool_char));
 197   psd->len = len;
 198
 199   if (front)
 200     this->strings_.push_front(psd);
 201   else
 202     this->strings_.push_back(psd);
 203
 204   return reinterpret_cast<const Stringpool_char*>(psd->data);
 205 }
 206
 207 // Add a string to a string pool.
 208
 209 template<typename Stringpool_char>
 210 const Stringpool_char*
 211 Stringpool_template<Stringpool_char>::add(const Stringpool_char* s, bool copy,
 212                                           Key* pkey)
 213 {
 214   return this->add_with_length(s, string_length(s), copy, pkey);
 215 }
 216
 217 // Add a new key offset entry.
 218
 219 template<typename Stringpool_char>
 220 void
 221 Stringpool_template<Stringpool_char>::new_key_offset(size_t length)
 222 {
 223   section_offset_type offset;
 224   if (this->zero_null_ && length == 0)
 225     offset = 0;
 226   else
 227     {
 228       offset = this->offset_;
 229       // Align non-zero length strings.
 230       if (length != 0)
 231         offset = align_address(offset, this->addralign_);
 232       this->offset_ = offset + (length + 1) * sizeof(Stringpool_char);
 233     }
 234   this->key_to_offset_.push_back(offset);
 235 }
 236
 237 template<typename Stringpool_char>
 238 const Stringpool_char*
 239 Stringpool_template<Stringpool_char>::add_with_length(const Stringpool_char* s,
 240                                                       size_t length,
 241                                                       bool copy,
 242                                                       Key* pkey)
 243 {
 244   typedef std::pair<typename String_set_type::iterator, bool> Insert_type;
 245
 246   // We add 1 so that 0 is always invalid.
 247   const Key k = this->key_to_offset_.size() + 1;
 248
 249   if (!copy)
 250     {
 251       // When we don't need to copy the string, we can call insert
 252       // directly.
 253
 254       std::pair<Hashkey, Hashval> element(Hashkey(s, length), k);
 255
 256       Insert_type ins = this->string_set_.insert(element);
 257
 258       typename String_set_type::const_iterator p = ins.first;
 259
 260       if (ins.second)
 261         {
 262           // We just added the string.  The key value has now been
 263           // used.
 264           this->new_key_offset(length);
 265         }
 266       else
 267         {
 268           gold_assert(k != p->second);
 269         }
 270
 271       if (pkey != NULL)
 272         *pkey = p->second;
 273       return p->first.string;
 274     }
 275
 276   // When we have to copy the string, we look it up twice in the hash
 277   // table.  The problem is that we can't insert S before we
 278   // canonicalize it by copying it into the canonical list. The hash
 279   // code will only be computed once.
 280
 281   Hashkey hk(s, length);
 282   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 283   if (p != this->string_set_.end())
 284     {
 285       if (pkey != NULL)
 286         *pkey = p->second;
 287       return p->first.string;
 288     }
 289
 290   this->new_key_offset(length);
 291
 292   hk.string = this->add_string(s, length);
 293   // The contents of the string stay the same, so we don't need to
 294   // adjust hk.hash_code or hk.length.
 295
 296   std::pair<Hashkey, Hashval> element(hk, k);
 297
 298   Insert_type ins = this->string_set_.insert(element);
 299   gold_assert(ins.second);
 300
 301   if (pkey != NULL)
 302     *pkey = k;
 303   return hk.string;
 304 }
 305
 306 template<typename Stringpool_char>
 307 const Stringpool_char*
 308 Stringpool_template<Stringpool_char>::find(const Stringpool_char* s,
 309                                            Key* pkey) const
 310 {
 311   Hashkey hk(s);
 312   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 313   if (p == this->string_set_.end())
 314     return NULL;
 315
 316   if (pkey != NULL)
 317     *pkey = p->second;
 318
 319   return p->first.string;
 320 }
 321
 322 // Comparison routine used when sorting into an ELF strtab.  We want
 323 // to sort this so that when one string is a suffix of another, we
 324 // always see the shorter string immediately after the longer string.
 325 // For example, we want to see these strings in this order:
 326 //   abcd
 327 //   cd
 328 //   d
 329 // When strings are not suffixes, we don't care what order they are
 330 // in, but we need to ensure that suffixes wind up next to each other.
 331 // So we do a reversed lexicographic sort on the reversed string.
 332
 333 template<typename Stringpool_char>
 334 bool
 335 Stringpool_template<Stringpool_char>::Stringpool_sort_comparison::operator()(
 336   const Stringpool_sort_info& sort_info1,
 337   const Stringpool_sort_info& sort_info2) const
 338 {
 339   const Hashkey& h1(sort_info1->first);
 340   const Hashkey& h2(sort_info2->first);
 341   const Stringpool_char* s1 = h1.string;
 342   const Stringpool_char* s2 = h2.string;
 343   const size_t len1 = h1.length;
 344   const size_t len2 = h2.length;
 345   const size_t minlen = len1 < len2 ? len1 : len2;
 346   const Stringpool_char* p1 = s1 + len1 - 1;
 347   const Stringpool_char* p2 = s2 + len2 - 1;
 348   for (size_t i = minlen; i > 0; --i, --p1, --p2)
 349     {
 350       if (*p1 != *p2)
 351         return *p1 > *p2;
 352     }
 353   return len1 > len2;
 354 }
 355
 356 // Return whether s1 is a suffix of s2.
 357
 358 template<typename Stringpool_char>
 359 bool
 360 Stringpool_template<Stringpool_char>::is_suffix(const Stringpool_char* s1,
 361                                                 size_t len1,
 362                                                 const Stringpool_char* s2,
 363                                                 size_t len2)
 364 {
 365   if (len1 > len2)
 366     return false;
 367   return memcmp(s1, s2 + len2 - len1, len1 * sizeof(Stringpool_char)) == 0;
 368 }
 369
 370 // Turn the stringpool into an ELF strtab: determine the offsets of
 371 // each string in the table.
 372
 373 template<typename Stringpool_char>
 374 void
 375 Stringpool_template<Stringpool_char>::set_string_offsets()
 376 {
 377   if (this->strtab_size_ != 0)
 378     {
 379       // We've already computed the offsets.
 380       return;
 381     }
 382
 383   const size_t charsize = sizeof(Stringpool_char);
 384
 385   // Offset 0 may be reserved for the empty string.
 386   section_offset_type offset = this->zero_null_ ? charsize : 0;
 387
 388   // Sorting to find suffixes can take over 25% of the total CPU time
 389   // used by the linker.  Since it's merely an optimization to reduce
 390   // the strtab size, and gives a relatively small benefit (it's
 391   // typically rare for a symbol to be a suffix of another), we only
 392   // take the time to sort when the user asks for heavy optimization.
 393   if (!this->optimize_)
 394     {
 395       // If we are not optimizing, the offsets are already assigned.
 396       offset = this->offset_;
 397     }
 398   else
 399     {
 400       size_t count = this->string_set_.size();
 401
 402       std::vector<Stringpool_sort_info> v;
 403       v.reserve(count);
 404
 405       for (typename String_set_type::iterator p = this->string_set_.begin();
 406            p != this->string_set_.end();
 407            ++p)
 408         v.push_back(Stringpool_sort_info(p));
 409
 410       std::sort(v.begin(), v.end(), Stringpool_sort_comparison());
 411
 412       section_offset_type last_offset = -1;
 413       for (typename std::vector<Stringpool_sort_info>::iterator last = v.end(),
 414              curr = v.begin();
 415            curr != v.end();
 416            last = curr++)
 417         {
 418           section_offset_type this_offset;
 419           if (this->zero_null_ && (*curr)->first.string[0] == 0)
 420             this_offset = 0;
 421           else if (last != v.end()
 422                    && is_suffix((*curr)->first.string,
 423                                 (*curr)->first.length,
 424                                 (*last)->first.string,
 425                                 (*last)->first.length))
 426             this_offset = (last_offset
 427                            + (((*last)->first.length - (*curr)->first.length)
 428                               * charsize));
 429           else
 430             {
 431               this_offset = align_address(offset, this->addralign_);
 432               offset = this_offset + ((*curr)->first.length + 1) * charsize;
 433             }
 434           this->key_to_offset_[(*curr)->second - 1] = this_offset;
 435           last_offset = this_offset;
 436         }
 437     }
 438
 439   this->strtab_size_ = offset;
 440 }
 441
 442 // Get the offset of a string in the ELF strtab.  The string must
 443 // exist.
 444
 445 template<typename Stringpool_char>
 446 section_offset_type
 447 Stringpool_template<Stringpool_char>::get_offset(const Stringpool_char* s)
 448   const
 449 {
 450   return this->get_offset_with_length(s, string_length(s));
 451 }
 452
 453 template<typename Stringpool_char>
 454 section_offset_type
 455 Stringpool_template<Stringpool_char>::get_offset_with_length(
 456     const Stringpool_char* s,
 457     size_t length) const
 458 {
 459   gold_assert(this->strtab_size_ != 0);
 460   Hashkey hk(s, length);
 461   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 462   if (p != this->string_set_.end())
 463     return this->key_to_offset_[p->second - 1];
 464   gold_unreachable();
 465 }
 466
 467 // Write the ELF strtab into the buffer.
 468
 469 template<typename Stringpool_char>
 470 void
 471 Stringpool_template<Stringpool_char>::write_to_buffer(
 472     unsigned char* buffer,
 473     section_size_type bufsize)
 474 {
 475   gold_assert(this->strtab_size_ != 0);
 476   gold_assert(bufsize >= this->strtab_size_);
 477   if (this->zero_null_)
 478     buffer[0] = '\0';
 479   for (typename String_set_type::const_iterator p = this->string_set_.begin();
 480        p != this->string_set_.end();
 481        ++p)
 482     {
 483       const int len = (p->first.length + 1) * sizeof(Stringpool_char);
 484       const section_offset_type offset = this->key_to_offset_[p->second - 1];
 485       gold_assert(static_cast<section_size_type>(offset) + len
 486                   <= this->strtab_size_);
 487       memcpy(buffer + offset, p->first.string, len);
 488     }
 489 }
 490
 491 // Write the ELF strtab into the output file at the specified offset.
 492
 493 template<typename Stringpool_char>
 494 void
 495 Stringpool_template<Stringpool_char>::write(Output_file* of, off_t offset)
 496 {
 497   gold_assert(this->strtab_size_ != 0);
 498   unsigned char* view = of->get_output_view(offset, this->strtab_size_);
 499   this->write_to_buffer(view, this->strtab_size_);
 500   of->write_output_view(offset, this->strtab_size_, view);
 501 }
 502
 503 // Print statistical information to stderr.  This is used for --stats.
 504
 505 template<typename Stringpool_char>
 506 void
 507 Stringpool_template<Stringpool_char>::print_stats(const char* name) const
 508 {
 509 #if defined(HAVE_UNORDERED_MAP) || defined(HAVE_TR1_UNORDERED_MAP) || defined(HAVE_EXT_HASH_MAP)
 510   fprintf(stderr, _("%s: %s entries: %zu; buckets: %zu\n"),
 511           program_name, name, this->string_set_.size(),
 512           this->string_set_.bucket_count());
 513 #else
 514   fprintf(stderr, _("%s: %s entries: %zu\n"),
 515           program_name, name, this->table_.size());
 516 #endif
 517   fprintf(stderr, _("%s: %s Stringdata structures: %zu\n"),
 518           program_name, name, this->strings_.size());
 519 }
 520
 521 // Instantiate the templates we need.
 522
 523 template
 524 class Stringpool_template<char>;
 525
 526 template
 527 class Stringpool_template<uint16_t>;
 528
 529 template
 530 class Stringpool_template<uint32_t>;
 531
 532 } // End namespace gold.