gold/stringpool.h

   1 // stringpool.h -- a string pool for gold    -*- C++ -*-
   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 <string>
  24 #include <list>
  25 #include <vector>
  26
  27 #ifndef GOLD_STRINGPOOL_H
  28 #define GOLD_STRINGPOOL_H
  29
  30 namespace gold
  31 {
  32
  33 class Output_file;
  34
  35 // A Stringpool is a pool of unique strings.  It provides the
  36 // following features:
  37
  38 // Every string in the pool is unique.  Thus, if you have two strings
  39 // in the Stringpool, you can compare them for equality by using
  40 // pointer comparison rather than string comparison.
  41
  42 // There is a key associated with every string in the pool.  If you
  43 // add strings to the Stringpool in the same order, then the key for
  44 // each string will always be the same for any run of the linker.
  45 // This is not true of the string pointers themselves, as they may
  46 // change due to address space randomization.  Some parts of the
  47 // linker (e.g., the symbol table) use the key value instead of the
  48 // string pointer so that repeated runs of the linker will generate
  49 // precisely the same output.
  50
  51 // When you add a string to a Stringpool, Stringpool will optionally
  52 // make a copy of it.  Thus there is no requirement to keep a copy
  53 // elsewhere.
  54
  55 // A Stringpool can be turned into a string table, a sequential series
  56 // of null terminated strings.  The first string may optionally be a
  57 // single zero byte, as required for SHT_STRTAB sections.  This
  58 // conversion is only permitted after all strings have been added to
  59 // the Stringpool.  After doing this conversion, you can ask for the
  60 // offset of any string (or any key) in the stringpool in the string
  61 // table, and you can write the resulting string table to an output
  62 // file.
  63
  64 // When a Stringpool is turned into a string table, then as an
  65 // optimization it will reuse string suffixes to avoid duplicating
  66 // strings.  That is, given the strings "abc" and "bc", only the
  67 // string "abc" will be stored, and "bc" will be represented by an
  68 // offset into the middle of the string "abc".
  69
  70
  71 // A simple chunked vector class--this is a subset of std::vector
  72 // which stores memory in chunks.  We don't provide iterators, because
  73 // we don't need them.
  74
  75 template<typename Element>
  76 class Chunked_vector
  77 {
  78  public:
  79   Chunked_vector()
  80     : chunks_()
  81   { }
  82
  83   // Clear the elements.
  84   void
  85   clear()
  86   { this->chunks_.clear(); }
  87
  88   // Reserve elements.
  89   void
  90   reserve(unsigned int n)
  91   {
  92     n += chunk_size - 1;
  93     while (n >= chunk_size)
  94       {
  95         this->chunks_.push_back(Element_vector());
  96         this->chunks_.back().reserve(chunk_size);
  97         n -= chunk_size;
  98       }
  99   }
 100
 101   // Get the number of elements.
 102   size_t
 103   size() const
 104   {
 105     if (this->chunks_.empty())
 106       return 0;
 107     else
 108       return ((this->chunks_.size() - 1) * chunk_size
 109               + this->chunks_.back().size());
 110   }
 111
 112   // Push a new element on the back of the vector.
 113   void
 114   push_back(const Element& element)
 115   {
 116     if (this->chunks_.empty() || this->chunks_.back().size() == chunk_size)
 117       {
 118         this->chunks_.push_back(Element_vector());
 119         this->chunks_.back().reserve(chunk_size);
 120       }
 121     this->chunks_.back().push_back(element);
 122   }
 123
 124   // Return a reference to an entry in the vector.
 125   Element&
 126   operator[](size_t i)
 127   { return this->chunks_[i / chunk_size][i % chunk_size]; }
 128
 129   const Element&
 130   operator[](size_t i) const
 131   { return this->chunks_[i / chunk_size][i % chunk_size]; }
 132
 133  private:
 134   static const unsigned int chunk_size = 8192;
 135
 136   typedef std::vector<Element> Element_vector;
 137   typedef std::vector<Element_vector> Chunk_vector;
 138
 139   Chunk_vector chunks_;
 140 };
 141
 142
 143 // Stringpools are implemented in terms of Stringpool_template, which
 144 // is generalized on the type of character used for the strings.  Most
 145 // uses will want the Stringpool type which uses char.  Other cases
 146 // are used for merging wide string constants.
 147
 148 template<typename Stringpool_char>
 149 class Stringpool_template
 150 {
 151  public:
 152   // The type of a key into the stringpool.  As described above, a key
 153   // value will always be the same during any run of the linker.  Zero
 154   // is never a valid key value.
 155   typedef size_t Key;
 156
 157   // Create a Stringpool.
 158   Stringpool_template();
 159
 160   ~Stringpool_template();
 161
 162   // Clear all the data from the stringpool.
 163   void
 164   clear();
 165
 166   // Hint to the stringpool class that you intend to insert n additional
 167   // elements.  The stringpool class can use this info however it likes;
 168   // in practice it will resize its internal hashtables to make room.
 169   void
 170   reserve(unsigned int n);
 171
 172   // Indicate that we should not reserve offset 0 to hold the empty
 173   // string when converting the stringpool to a string table.  This
 174   // should not be called for a proper ELF SHT_STRTAB section.
 175   void
 176   set_no_zero_null()
 177   { this->zero_null_ = false; }
 178
 179   // Add the string S to the pool.  This returns a canonical permanent
 180   // pointer to the string in the pool.  If COPY is true, the string
 181   // is copied into permanent storage.  If PKEY is not NULL, this sets
 182   // *PKEY to the key for the string.
 183   const Stringpool_char*
 184   add(const Stringpool_char* s, bool copy, Key* pkey);
 185
 186   // Add string S of length LEN characters to the pool.  If COPY is
 187   // true, S need not be null terminated.
 188   const Stringpool_char*
 189   add_with_length(const Stringpool_char* s, size_t len, bool copy, Key* pkey);
 190
 191   // If the string S is present in the pool, return the canonical
 192   // string pointer.  Otherwise, return NULL.  If PKEY is not NULL,
 193   // set *PKEY to the key.
 194   const Stringpool_char*
 195   find(const Stringpool_char* s, Key* pkey) const;
 196
 197   // Turn the stringpool into a string table: determine the offsets of
 198   // all the strings.  After this is called, no more strings may be
 199   // added to the stringpool.
 200   void
 201   set_string_offsets();
 202
 203   // Get the offset of the string S in the string table.  This returns
 204   // the offset in bytes, not in units of Stringpool_char.  This may
 205   // only be called after set_string_offsets has been called.
 206   section_offset_type
 207   get_offset(const Stringpool_char* s) const;
 208
 209   // Get the offset of the string S in the string table.
 210   section_offset_type
 211   get_offset(const std::basic_string<Stringpool_char>& s) const
 212   { return this->get_offset_with_length(s.c_str(), s.size()); }
 213
 214   // Get the offset of string S, with length LENGTH characters, in the
 215   // string table.
 216   section_offset_type
 217   get_offset_with_length(const Stringpool_char* s, size_t length) const;
 218
 219   // Get the offset of the string with key K.
 220   section_offset_type
 221   get_offset_from_key(Key k) const
 222   {
 223     gold_assert(k <= this->key_to_offset_.size());
 224     return this->key_to_offset_[k - 1];
 225   }
 226
 227   // Get the size of the string table.  This returns the number of
 228   // bytes, not in units of Stringpool_char.
 229   section_size_type
 230   get_strtab_size() const
 231   {
 232     gold_assert(this->strtab_size_ != 0);
 233     return this->strtab_size_;
 234   }
 235
 236   // Write the string table into the output file at the specified
 237   // offset.
 238   void
 239   write(Output_file*, off_t offset);
 240
 241   // Write the string table into the specified buffer, of the
 242   // specified size.  buffer_size should be at least
 243   // get_strtab_size().
 244   void
 245   write_to_buffer(unsigned char* buffer, section_size_type buffer_size);
 246
 247   // Dump statistical information to stderr.
 248   void
 249   print_stats(const char*) const;
 250
 251  private:
 252   Stringpool_template(const Stringpool_template&);
 253   Stringpool_template& operator=(const Stringpool_template&);
 254
 255   // Return the length of a string in units of Stringpool_char.
 256   static size_t
 257   string_length(const Stringpool_char*);
 258
 259   // Return whether two strings are equal.
 260   static bool
 261   string_equal(const Stringpool_char*, const Stringpool_char*);
 262
 263   // Compute a hash code for a string.  LENGTH is the length of the
 264   // string in characters.
 265   static size_t
 266   string_hash(const Stringpool_char*, size_t length);
 267
 268   // We store the actual data in a list of these buffers.
 269   struct Stringdata
 270   {
 271     // Length of data in buffer.
 272     size_t len;
 273     // Allocated size of buffer.
 274     size_t alc;
 275     // Buffer.
 276     char data[1];
 277   };
 278
 279   // Copy a string into the buffers, returning a canonical string.
 280   const Stringpool_char*
 281   add_string(const Stringpool_char*, size_t);
 282
 283   // Return whether s1 is a suffix of s2.
 284   static bool
 285   is_suffix(const Stringpool_char* s1, size_t len1,
 286             const Stringpool_char* s2, size_t len2);
 287
 288   // The hash table key includes the string, the length of the string,
 289   // and the hash code for the string.  We put the hash code
 290   // explicitly into the key so that we can do a find()/insert()
 291   // sequence without having to recompute the hash.  Computing the
 292   // hash code is a significant user of CPU time in the linker.
 293   struct Hashkey
 294   {
 295     const Stringpool_char* string;
 296     // Length is in characters, not bytes.
 297     size_t length;
 298     size_t hash_code;
 299
 300     // This goes in an STL container, so we need a default
 301     // constructor.
 302     Hashkey()
 303       : string(NULL), length(0), hash_code(0)
 304     { }
 305
 306     // Note that these constructors are relatively expensive, because
 307     // they compute the hash code.
 308     explicit Hashkey(const Stringpool_char* s)
 309       : string(s), length(string_length(s)), hash_code(string_hash(s, length))
 310     { }
 311
 312     Hashkey(const Stringpool_char* s, size_t len)
 313       : string(s), length(len), hash_code(string_hash(s, len))
 314     { }
 315   };
 316
 317   // Hash function.  This is trivial, since we have already computed
 318   // the hash.
 319   struct Stringpool_hash
 320   {
 321     size_t
 322     operator()(const Hashkey& hk) const
 323     { return hk.hash_code; }
 324   };
 325
 326   // Equality comparison function for hash table.
 327   struct Stringpool_eq
 328   {
 329     bool
 330     operator()(const Hashkey&, const Hashkey&) const;
 331   };
 332
 333   // The hash table is a map from strings to Keys.
 334
 335   typedef Key Hashval;
 336
 337   typedef Unordered_map<Hashkey, Hashval, Stringpool_hash,
 338                         Stringpool_eq> String_set_type;
 339
 340   // Comparison routine used when sorting into a string table.
 341
 342   typedef typename String_set_type::iterator Stringpool_sort_info;
 343
 344   struct Stringpool_sort_comparison
 345   {
 346     bool
 347     operator()(const Stringpool_sort_info&, const Stringpool_sort_info&) const;
 348   };
 349
 350   // Keys map to offsets via a Chunked_vector.  We only use the
 351   // offsets if we turn this into an string table section.
 352   typedef Chunked_vector<section_offset_type> Key_to_offset;
 353
 354   // List of Stringdata structures.
 355   typedef std::list<Stringdata*> Stringdata_list;
 356
 357   // Mapping from const char* to namepool entry.
 358   String_set_type string_set_;
 359   // Mapping from Key to string table offset.
 360   Key_to_offset key_to_offset_;
 361   // List of buffers.
 362   Stringdata_list strings_;
 363   // Size of string table.
 364   section_size_type strtab_size_;
 365   // Whether to reserve offset 0 to hold the null string.
 366   bool zero_null_;
 367 };
 368
 369 // The most common type of Stringpool.
 370 typedef Stringpool_template<char> Stringpool;
 371
 372 } // End namespace gold.
 373
 374 #endif // !defined(GOLD_STRINGPOOL_H)