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_TR1_UNORDERED_MAP)
  77   // rehash() implementation is broken in gcc 4.0.3's stl
  78   //this->string_set_.rehash(this->string_set_.size() + n);
  79   //return;
  80 #elif defined(HAVE_EXT_HASH_MAP)
  81   this->string_set_.resize(this->string_set_.size() + n);
  82   return;
  83 #endif
  84
  85   // This is the generic "reserve" code, if no #ifdef above triggers.
  86   String_set_type new_string_set(this->string_set_.size() + n);
  87   new_string_set.insert(this->string_set_.begin(), this->string_set_.end());
  88   this->string_set_.swap(new_string_set);
  89 }
  90
  91 // Compare two strings of arbitrary character type for equality.
  92
  93 template<typename Stringpool_char>
  94 bool
  95 Stringpool_template<Stringpool_char>::string_equal(const Stringpool_char* s1,
  96                                                    const Stringpool_char* s2)
  97 {
  98   while (*s1 != 0)
  99     if (*s1++ != *s2++)
 100       return false;
 101   return *s2 == 0;
 102 }
 103
 104 // Specialize string_equal for char.
 105
 106 template<>
 107 inline bool
 108 Stringpool_template<char>::string_equal(const char* s1, const char* s2)
 109 {
 110   return strcmp(s1, s2) == 0;
 111 }
 112
 113 // Equality comparison function for the hash table.
 114
 115 template<typename Stringpool_char>
 116 inline bool
 117 Stringpool_template<Stringpool_char>::Stringpool_eq::operator()(
 118     const Hashkey& h1,
 119     const Hashkey& h2) const
 120 {
 121   return (h1.hash_code == h2.hash_code
 122           && h1.length == h2.length
 123           && (h1.string == h2.string
 124               || memcmp(h1.string, h2.string,
 125                         h1.length * sizeof(Stringpool_char)) == 0));
 126 }
 127
 128 // Hash function.  The length is in characters, not bytes.
 129
 130 template<typename Stringpool_char>
 131 size_t
 132 Stringpool_template<Stringpool_char>::string_hash(const Stringpool_char* s,
 133                                                   size_t length)
 134 {
 135   return gold::string_hash<Stringpool_char>(s, length);
 136 }
 137
 138 // Add the string S to the list of canonical strings.  Return a
 139 // pointer to the canonical string.  If PKEY is not NULL, set *PKEY to
 140 // the key.  LENGTH is the length of S in characters.  Note that S may
 141 // not be NUL terminated.
 142
 143 template<typename Stringpool_char>
 144 const Stringpool_char*
 145 Stringpool_template<Stringpool_char>::add_string(const Stringpool_char* s,
 146                                                  size_t len)
 147 {
 148   // We are in trouble if we've already computed the string offsets.
 149   gold_assert(this->strtab_size_ == 0);
 150
 151   // The size we allocate for a new Stringdata.
 152   const size_t buffer_size = 1000;
 153   // The amount we multiply the Stringdata index when calculating the
 154   // key.
 155   const size_t key_mult = 1024;
 156   gold_assert(key_mult >= buffer_size);
 157
 158   // Convert len to the number of bytes we need to allocate, including
 159   // the null character.
 160   len = (len + 1) * sizeof(Stringpool_char);
 161
 162   size_t alc;
 163   bool front = true;
 164   if (len > buffer_size)
 165     {
 166       alc = sizeof(Stringdata) + len;
 167       front = false;
 168     }
 169   else if (this->strings_.empty())
 170     alc = sizeof(Stringdata) + buffer_size;
 171   else
 172     {
 173       Stringdata* psd = this->strings_.front();
 174       if (len > psd->alc - psd->len)
 175         alc = sizeof(Stringdata) + buffer_size;
 176       else
 177         {
 178           char* ret = psd->data + psd->len;
 179           memcpy(ret, s, len - sizeof(Stringpool_char));
 180           memset(ret + len - sizeof(Stringpool_char), 0,
 181                  sizeof(Stringpool_char));
 182
 183           psd->len += len;
 184
 185           return reinterpret_cast<const Stringpool_char*>(ret);
 186         }
 187     }
 188
 189   Stringdata* psd = reinterpret_cast<Stringdata*>(new char[alc]);
 190   psd->alc = alc - sizeof(Stringdata);
 191   memcpy(psd->data, s, len - sizeof(Stringpool_char));
 192   memset(psd->data + len - sizeof(Stringpool_char), 0,
 193          sizeof(Stringpool_char));
 194   psd->len = len;
 195
 196   if (front)
 197     this->strings_.push_front(psd);
 198   else
 199     this->strings_.push_back(psd);
 200
 201   return reinterpret_cast<const Stringpool_char*>(psd->data);
 202 }
 203
 204 // Add a string to a string pool.
 205
 206 template<typename Stringpool_char>
 207 const Stringpool_char*
 208 Stringpool_template<Stringpool_char>::add(const Stringpool_char* s, bool copy,
 209                                           Key* pkey)
 210 {
 211   return this->add_with_length(s, string_length(s), copy, pkey);
 212 }
 213
 214 // Add a new key offset entry.
 215
 216 template<typename Stringpool_char>
 217 void
 218 Stringpool_template<Stringpool_char>::new_key_offset(size_t length)
 219 {
 220   section_offset_type offset;
 221   if (this->zero_null_ && length == 0)
 222     offset = 0;
 223   else
 224     {
 225       // Align non-zero length strings.
 226       if (length != 0)
 227         offset = align_address(this->offset_, this->addralign_);
 228       this->offset_ = offset + (length + 1) * sizeof(Stringpool_char);
 229     }
 230   this->key_to_offset_.push_back(offset);
 231 }
 232
 233 template<typename Stringpool_char>
 234 const Stringpool_char*
 235 Stringpool_template<Stringpool_char>::add_with_length(const Stringpool_char* s,
 236                                                       size_t length,
 237                                                       bool copy,
 238                                                       Key* pkey)
 239 {
 240   typedef std::pair<typename String_set_type::iterator, bool> Insert_type;
 241
 242   // We add 1 so that 0 is always invalid.
 243   const Key k = this->key_to_offset_.size() + 1;
 244
 245   if (!copy)
 246     {
 247       // When we don't need to copy the string, we can call insert
 248       // directly.
 249
 250       std::pair<Hashkey, Hashval> element(Hashkey(s, length), k);
 251
 252       Insert_type ins = this->string_set_.insert(element);
 253
 254       typename String_set_type::const_iterator p = ins.first;
 255
 256       if (ins.second)
 257         {
 258           // We just added the string.  The key value has now been
 259           // used.
 260           this->new_key_offset(length);
 261         }
 262       else
 263         {
 264           gold_assert(k != p->second);
 265         }
 266
 267       if (pkey != NULL)
 268         *pkey = p->second;
 269       return p->first.string;
 270     }
 271
 272   // When we have to copy the string, we look it up twice in the hash
 273   // table.  The problem is that we can't insert S before we
 274   // canonicalize it by copying it into the canonical list. The hash
 275   // code will only be computed once.
 276
 277   Hashkey hk(s, length);
 278   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 279   if (p != this->string_set_.end())
 280     {
 281       if (pkey != NULL)
 282         *pkey = p->second;
 283       return p->first.string;
 284     }
 285
 286   this->new_key_offset(length);
 287
 288   hk.string = this->add_string(s, length);
 289   // The contents of the string stay the same, so we don't need to
 290   // adjust hk.hash_code or hk.length.
 291
 292   std::pair<Hashkey, Hashval> element(hk, k);
 293
 294   Insert_type ins = this->string_set_.insert(element);
 295   gold_assert(ins.second);
 296
 297   if (pkey != NULL)
 298     *pkey = k;
 299   return hk.string;
 300 }
 301
 302 template<typename Stringpool_char>
 303 const Stringpool_char*
 304 Stringpool_template<Stringpool_char>::find(const Stringpool_char* s,
 305                                            Key* pkey) const
 306 {
 307   Hashkey hk(s);
 308   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 309   if (p == this->string_set_.end())
 310     return NULL;
 311
 312   if (pkey != NULL)
 313     *pkey = p->second;
 314
 315   return p->first.string;
 316 }
 317
 318 // Comparison routine used when sorting into an ELF strtab.  We want
 319 // to sort this so that when one string is a suffix of another, we
 320 // always see the shorter string immediately after the longer string.
 321 // For example, we want to see these strings in this order:
 322 //   abcd
 323 //   cd
 324 //   d
 325 // When strings are not suffixes, we don't care what order they are
 326 // in, but we need to ensure that suffixes wind up next to each other.
 327 // So we do a reversed lexicographic sort on the reversed string.
 328
 329 template<typename Stringpool_char>
 330 bool
 331 Stringpool_template<Stringpool_char>::Stringpool_sort_comparison::operator()(
 332   const Stringpool_sort_info& sort_info1,
 333   const Stringpool_sort_info& sort_info2) const
 334 {
 335   const Hashkey& h1(sort_info1->first);
 336   const Hashkey& h2(sort_info2->first);
 337   const Stringpool_char* s1 = h1.string;
 338   const Stringpool_char* s2 = h2.string;
 339   const size_t len1 = h1.length;
 340   const size_t len2 = h2.length;
 341   const size_t minlen = len1 < len2 ? len1 : len2;
 342   const Stringpool_char* p1 = s1 + len1 - 1;
 343   const Stringpool_char* p2 = s2 + len2 - 1;
 344   for (size_t i = minlen; i > 0; --i, --p1, --p2)
 345     {
 346       if (*p1 != *p2)
 347         return *p1 > *p2;
 348     }
 349   return len1 > len2;
 350 }
 351
 352 // Return whether s1 is a suffix of s2.
 353
 354 template<typename Stringpool_char>
 355 bool
 356 Stringpool_template<Stringpool_char>::is_suffix(const Stringpool_char* s1,
 357                                                 size_t len1,
 358                                                 const Stringpool_char* s2,
 359                                                 size_t len2)
 360 {
 361   if (len1 > len2)
 362     return false;
 363   return memcmp(s1, s2 + len2 - len1, len1 * sizeof(Stringpool_char)) == 0;
 364 }
 365
 366 // Turn the stringpool into an ELF strtab: determine the offsets of
 367 // each string in the table.
 368
 369 template<typename Stringpool_char>
 370 void
 371 Stringpool_template<Stringpool_char>::set_string_offsets()
 372 {
 373   if (this->strtab_size_ != 0)
 374     {
 375       // We've already computed the offsets.
 376       return;
 377     }
 378
 379   const size_t charsize = sizeof(Stringpool_char);
 380
 381   // Offset 0 may be reserved for the empty string.
 382   section_offset_type offset = this->zero_null_ ? charsize : 0;
 383
 384   // Sorting to find suffixes can take over 25% of the total CPU time
 385   // used by the linker.  Since it's merely an optimization to reduce
 386   // the strtab size, and gives a relatively small benefit (it's
 387   // typically rare for a symbol to be a suffix of another), we only
 388   // take the time to sort when the user asks for heavy optimization.
 389   if (!this->optimize_)
 390     {
 391       // If we are not optimizing, the offsets are already assigned.
 392       offset = this->offset_;
 393     }
 394   else
 395     {
 396       size_t count = this->string_set_.size();
 397
 398       std::vector<Stringpool_sort_info> v;
 399       v.reserve(count);
 400
 401       for (typename String_set_type::iterator p = this->string_set_.begin();
 402            p != this->string_set_.end();
 403            ++p)
 404         v.push_back(Stringpool_sort_info(p));
 405
 406       std::sort(v.begin(), v.end(), Stringpool_sort_comparison());
 407
 408       section_offset_type last_offset = -1;
 409       for (typename std::vector<Stringpool_sort_info>::iterator last = v.end(),
 410              curr = v.begin();
 411            curr != v.end();
 412            last = curr++)
 413         {
 414           section_offset_type this_offset;
 415           if (this->zero_null_ && (*curr)->first.string[0] == 0)
 416             this_offset = 0;
 417           else if (last != v.end()
 418                    && is_suffix((*curr)->first.string,
 419                                 (*curr)->first.length,
 420                                 (*last)->first.string,
 421                                 (*last)->first.length))
 422             this_offset = (last_offset
 423                            + (((*last)->first.length - (*curr)->first.length)
 424                               * charsize));
 425           else
 426             {
 427               this_offset = align_address(offset, this->addralign_);
 428               offset = this_offset + ((*curr)->first.length + 1) * charsize;
 429             }
 430           this->key_to_offset_[(*curr)->second - 1] = this_offset;
 431           last_offset = this_offset;
 432         }
 433     }
 434
 435   this->strtab_size_ = offset;
 436 }
 437
 438 // Get the offset of a string in the ELF strtab.  The string must
 439 // exist.
 440
 441 template<typename Stringpool_char>
 442 section_offset_type
 443 Stringpool_template<Stringpool_char>::get_offset(const Stringpool_char* s)
 444   const
 445 {
 446   return this->get_offset_with_length(s, string_length(s));
 447 }
 448
 449 template<typename Stringpool_char>
 450 section_offset_type
 451 Stringpool_template<Stringpool_char>::get_offset_with_length(
 452     const Stringpool_char* s,
 453     size_t length) const
 454 {
 455   gold_assert(this->strtab_size_ != 0);
 456   Hashkey hk(s, length);
 457   typename String_set_type::const_iterator p = this->string_set_.find(hk);
 458   if (p != this->string_set_.end())
 459     return this->key_to_offset_[p->second - 1];
 460   gold_unreachable();
 461 }
 462
 463 // Write the ELF strtab into the buffer.
 464
 465 template<typename Stringpool_char>
 466 void
 467 Stringpool_template<Stringpool_char>::write_to_buffer(
 468     unsigned char* buffer,
 469     section_size_type bufsize)
 470 {
 471   gold_assert(this->strtab_size_ != 0);
 472   gold_assert(bufsize >= this->strtab_size_);
 473   if (this->zero_null_)
 474     buffer[0] = '\0';
 475   for (typename String_set_type::const_iterator p = this->string_set_.begin();
 476        p != this->string_set_.end();
 477        ++p)
 478     {
 479       const int len = (p->first.length + 1) * sizeof(Stringpool_char);
 480       const section_offset_type offset = this->key_to_offset_[p->second - 1];
 481       gold_assert(static_cast<section_size_type>(offset) + len
 482                   <= this->strtab_size_);
 483       memcpy(buffer + offset, p->first.string, len);
 484     }
 485 }
 486
 487 // Write the ELF strtab into the output file at the specified offset.
 488
 489 template<typename Stringpool_char>
 490 void
 491 Stringpool_template<Stringpool_char>::write(Output_file* of, off_t offset)
 492 {
 493   gold_assert(this->strtab_size_ != 0);
 494   unsigned char* view = of->get_output_view(offset, this->strtab_size_);
 495   this->write_to_buffer(view, this->strtab_size_);
 496   of->write_output_view(offset, this->strtab_size_, view);
 497 }
 498
 499 // Print statistical information to stderr.  This is used for --stats.
 500
 501 template<typename Stringpool_char>
 502 void
 503 Stringpool_template<Stringpool_char>::print_stats(const char* name) const
 504 {
 505 #if defined(HAVE_TR1_UNORDERED_MAP) || defined(HAVE_EXT_HASH_MAP)
 506   fprintf(stderr, _("%s: %s entries: %zu; buckets: %zu\n"),
 507           program_name, name, this->string_set_.size(),
 508           this->string_set_.bucket_count());
 509 #else
 510   fprintf(stderr, _("%s: %s entries: %zu\n"),
 511           program_name, name, this->table_.size());
 512 #endif
 513   fprintf(stderr, _("%s: %s Stringdata structures: %zu\n"),
 514           program_name, name, this->strings_.size());
 515 }
 516
 517 // Instantiate the templates we need.
 518
 519 template
 520 class Stringpool_template<char>;
 521
 522 template
 523 class Stringpool_template<uint16_t>;
 524
 525 template
 526 class Stringpool_template<uint32_t>;
 527
 528 } // End namespace gold.