gold/stringpool.h

   1 // stringpool.h -- a string pool for gold    -*- C++ -*-
   2
   3 // Copyright 2006, 2007 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
  26 #ifndef GOLD_STRINGPOOL_H
  27 #define GOLD_STRINGPOOL_H
  28
  29 namespace gold
  30 {
  31
  32 class Output_file;
  33
  34 // A Stringpool is a pool of unique strings.  It provides the
  35 // following features:
  36
  37 // Every string in the pool is unique.  Thus, if you have two strings
  38 // in the Stringpool, you can compare them for equality by using
  39 // pointer comparison rather than string comparison.
  40
  41 // There is a key associated with every string in the pool.  If you
  42 // add strings to the Stringpool in the same order, then the key for
  43 // each string will always be the same for any run of the linker.
  44 // This is not true of the string pointers themselves, as they may
  45 // change due to address space randomization.  Some parts of the
  46 // linker (e.g., the symbol table) use the key value instead of the
  47 // string pointer so that repeated runs of the linker will generate
  48 // precisely the same output.
  49
  50 // When you add a string to a Stringpool, Stringpool will optionally
  51 // make a copy of it.  Thus there is no requirement to keep a copy
  52 // elsewhere.
  53
  54 // A Stringpool can be turned into a string table, a sequential series
  55 // of null terminated strings.  The first string may optionally be a
  56 // single zero byte, as required for SHT_STRTAB sections.  This
  57 // conversion is only permitted after all strings have been added to
  58 // the Stringpool.  After doing this conversion, you can ask for the
  59 // offset of any string in the stringpool in the string table, and you
  60 // can write the resulting string table to an output file.
  61
  62 // When a Stringpool is turned into a string table, then as an
  63 // optimization it will reuse string suffixes to avoid duplicating
  64 // strings.  That is, given the strings "abc" and "bc", only the
  65 // string "abc" will be stored, and "bc" will be represented by an
  66 // offset into the middle of the string "abc".
  67
  68 // Stringpools are implemented in terms of Stringpool_template, which
  69 // is generalized on the type of character used for the strings.  Most
  70 // uses will want the Stringpool type which uses char.  Other cases
  71 // are used for merging wide string constants.
  72
  73 template<typename Stringpool_char>
  74 class Stringpool_template
  75 {
  76  public:
  77   // The type of a key into the stringpool.  As described above, a key
  78   // value will always be the same during any run of the linker.  Zero
  79   // is never a valid key value.
  80   typedef size_t Key;
  81
  82   // Create a Stringpool.
  83   Stringpool_template();
  84
  85   ~Stringpool_template();
  86
  87   // Clear all the data from the stringpool.
  88   void
  89   clear();
  90
  91   // Hint to the stringpool class that you intend to insert n additional
  92   // elements.  The stringpool class can use this info however it likes;
  93   // in practice it will resize its internal hashtables to make room.
  94   void
  95   reserve(unsigned int n);
  96
  97   // Indicate that we should not reserve offset 0 to hold the empty
  98   // string when converting the stringpool to a string table.  This
  99   // should not be called for a proper ELF SHT_STRTAB section.
 100   void
 101   set_no_zero_null()
 102   { this->zero_null_ = false; }
 103
 104   // Add the string S to the pool.  This returns a canonical permanent
 105   // pointer to the string in the pool.  If COPY is true, the string
 106   // is copied into permanent storage.  If PKEY is not NULL, this sets
 107   // *PKEY to the key for the string.
 108   const Stringpool_char*
 109   add(const Stringpool_char* s, bool copy, Key* pkey);
 110
 111   // Add the prefix of length LEN of string S to the pool.
 112   const Stringpool_char*
 113   add_prefix(const Stringpool_char* s, size_t len, Key* pkey);
 114
 115   // If the string S is present in the pool, return the canonical
 116   // string pointer.  Otherwise, return NULL.  If PKEY is not NULL,
 117   // set *PKEY to the key.
 118   const Stringpool_char*
 119   find(const Stringpool_char* s, Key* pkey) const;
 120
 121   // Turn the stringpool into a string table: determine the offsets of
 122   // all the strings.  After this is called, no more strings may be
 123   // added to the stringpool.
 124   void
 125   set_string_offsets();
 126
 127   // Get the offset of the string S in the string table.  This returns
 128   // the offset in bytes, not in units of Stringpool_char.  This may
 129   // only be called after set_string_offsets has been called.
 130   off_t
 131   get_offset(const Stringpool_char* s) const;
 132
 133   // Get the offset of the string S in the string table.
 134   off_t
 135   get_offset(const std::basic_string<Stringpool_char>& s) const
 136   { return this->get_offset(s.c_str()); }
 137
 138   // Get the size of the string table.  This returns the number of
 139   // bytes, not in units of Stringpool_char.
 140   off_t
 141   get_strtab_size() const
 142   {
 143     gold_assert(this->strtab_size_ != 0);
 144     return this->strtab_size_;
 145   }
 146
 147   // Write the string table into the output file at the specified
 148   // offset.
 149   void
 150   write(Output_file*, off_t offset);
 151
 152   // Write the string table into the specified buffer, of the
 153   // specified size.  buffer_size should be at least
 154   // get_strtab_size().
 155   void
 156   write_to_buffer(unsigned char* buffer, size_t buffer_size);
 157
 158   // Dump statistical information to stderr.
 159   void
 160   print_stats(const char*) const;
 161
 162  private:
 163   Stringpool_template(const Stringpool_template&);
 164   Stringpool_template& operator=(const Stringpool_template&);
 165
 166   // Return the length of a string in units of Stringpool_char.
 167   static size_t
 168   string_length(const Stringpool_char*);
 169
 170   // Return whether two strings are equal.
 171   static bool
 172   string_equal(const Stringpool_char*, const Stringpool_char*);
 173
 174   // Compute a hash code for a string.  LENGTH is the length of the
 175   // string in characters.
 176   static size_t
 177   string_hash(const Stringpool_char*, size_t length);
 178
 179   // We store the actual data in a list of these buffers.
 180   struct Stringdata
 181   {
 182     // Length of data in buffer.
 183     size_t len;
 184     // Allocated size of buffer.
 185     size_t alc;
 186     // Buffer index.
 187     unsigned int index;
 188     // Buffer.
 189     char data[1];
 190   };
 191
 192   // Copy a string into the buffers, returning a canonical string.
 193   const Stringpool_char*
 194   add_string(const Stringpool_char*, size_t, Key*);
 195
 196   // Return whether s1 is a suffix of s2.
 197   static bool
 198   is_suffix(const Stringpool_char* s1, size_t len1,
 199             const Stringpool_char* s2, size_t len2);
 200
 201   // The hash table key includes the string, the length of the string,
 202   // and the hash code for the string.  We put the hash code
 203   // explicitly into the key so that we can do a find()/insert()
 204   // sequence without having to recompute the hash.  Computing the
 205   // hash code is a significant user of CPU time in the linker.
 206   struct Hashkey
 207   {
 208     const Stringpool_char* string;
 209     // Length is in characters, not bytes.
 210     size_t length;
 211     size_t hash_code;
 212
 213     // This goes in an STL container, so we need a default
 214     // constructor.
 215     Hashkey()
 216       : string(NULL), length(0), hash_code(0)
 217     { }
 218
 219     // Note that these constructors are relatively expensive, because
 220     // they compute the hash code.
 221     Hashkey(const Stringpool_char* s)
 222       : string(s), length(string_length(s)), hash_code(string_hash(s, length))
 223     { }
 224
 225     Hashkey(const Stringpool_char* s, size_t len)
 226       : string(s), length(len), hash_code(string_hash(s, len))
 227     { }
 228   };
 229
 230   // Hash function.  This is trivial, since we have already computed
 231   // the hash.
 232   struct Stringpool_hash
 233   {
 234     size_t
 235     operator()(const Hashkey& hk) const
 236     { return hk.hash_code; }
 237   };
 238
 239   // Equality comparison function for hash table.
 240   struct Stringpool_eq
 241   {
 242     bool
 243     operator()(const Hashkey&, const Hashkey&) const;
 244   };
 245
 246   // The hash table is a map from strings to a pair of Key and string
 247   // table offsets.  We only use the offsets if we turn this into an
 248   // string table section.
 249
 250   typedef std::pair<Key, off_t> Hashval;
 251
 252   typedef Unordered_map<Hashkey, Hashval, Stringpool_hash,
 253                         Stringpool_eq> String_set_type;
 254
 255   // Comparison routine used when sorting into a string table.
 256
 257   typedef typename String_set_type::iterator Stringpool_sort_info;
 258
 259   struct Stringpool_sort_comparison
 260   {
 261     bool
 262     operator()(const Stringpool_sort_info&, const Stringpool_sort_info&) const;
 263   };
 264
 265   // List of Stringdata structures.
 266   typedef std::list<Stringdata*> Stringdata_list;
 267
 268   // Mapping from const char* to namepool entry.
 269   String_set_type string_set_;
 270   // List of buffers.
 271   Stringdata_list strings_;
 272   // Size of string table.
 273   off_t strtab_size_;
 274   // Next Stringdata index.
 275   unsigned int next_index_;
 276   // Next key value for a string we don't copy.
 277   int next_uncopied_key_;
 278   // Whether to reserve offset 0 to hold the null string.
 279   bool zero_null_;
 280 };
 281
 282 // The most common type of Stringpool.
 283 typedef Stringpool_template<char> Stringpool;
 284
 285 } // End namespace gold.
 286
 287 #endif // !defined(GOLD_STRINGPOOL_H)