1 /*
   2  * Copyright (c) 1997, 2017, Oracle and/or its affiliates. All rights reserved.
   3  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
   4  *
   5  * This code is free software; you can redistribute it and/or modify it
   6  * under the terms of the GNU General Public License version 2 only, as
   7  * published by the Free Software Foundation.
   8  *
   9  * This code is distributed in the hope that it will be useful, but WITHOUT
  10  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
  11  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
  12  * version 2 for more details (a copy is included in the LICENSE file that
  13  * accompanied this code).
  14  *
  15  * You should have received a copy of the GNU General Public License version
  16  * 2 along with this work; if not, write to the Free Software Foundation,
  17  * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
  18  *
  19  * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
  20  * or visit www.oracle.com if you need additional information or have any
  21  * questions.
  22  *
  23  */
  24 
  25 #ifndef SHARE_VM_CLASSFILE_SYMBOLTABLE_HPP
  26 #define SHARE_VM_CLASSFILE_SYMBOLTABLE_HPP
  27 
  28 #include "memory/allocation.hpp"
  29 #include "oops/symbol.hpp"
  30 #include "utilities/hashtable.hpp"
  31 
  32 // The symbol table holds all Symbol*s and corresponding interned strings.
  33 // Symbol*s and literal strings should be canonicalized.
  34 //
  35 // The interned strings are created lazily.
  36 //
  37 // It is implemented as an open hash table with a fixed number of buckets.
  38 //
  39 // %note:
  40 //  - symbolTableEntrys are allocated in blocks to reduce the space overhead.
  41 
  42 class BoolObjectClosure;
  43 class outputStream;
  44 class SerializeClosure;
  45 
  46 // TempNewSymbol acts as a handle class in a handle/body idiom and is
  47 // responsible for proper resource management of the body (which is a Symbol*).
  48 // The body is resource managed by a reference counting scheme.
  49 // TempNewSymbol can therefore be used to properly hold a newly created or referenced
  50 // Symbol* temporarily in scope.
  51 //
  52 // Routines in SymbolTable will initialize the reference count of a Symbol* before
  53 // it becomes "managed" by TempNewSymbol instances. As a handle class, TempNewSymbol
  54 // needs to maintain proper reference counting in context of copy semantics.
  55 //
  56 // In SymbolTable, new_symbol() and lookup() will create a Symbol* if not already in the
  57 // symbol table and add to the symbol's reference count.
  58 // probe() and lookup_only() will increment the refcount if symbol is found.
  59 class TempNewSymbol : public StackObj {
  60   Symbol* _temp;
  61 
  62  public:
  63   TempNewSymbol() : _temp(NULL) {}
  64 
  65   // Conversion from a Symbol* to a TempNewSymbol.
  66   // Does not increment the current reference count.
  67   TempNewSymbol(Symbol *s) : _temp(s) {}
  68 
  69   // Copy constructor increments reference count.
  70   TempNewSymbol(const TempNewSymbol& rhs) : _temp(rhs._temp) {
  71     if (_temp != NULL) {
  72       _temp->increment_refcount();
  73     }
  74   }
  75 
  76   // Assignment operator uses a c++ trick called copy and swap idiom.
  77   // rhs is passed by value so within the scope of this method it is a copy.
  78   // At method exit it contains the former value of _temp, triggering the correct refcount
  79   // decrement upon destruction.
  80   void operator=(TempNewSymbol rhs) {
  81     Symbol* tmp = rhs._temp;
  82     rhs._temp = _temp;
  83     _temp = tmp;
  84   }
  85 
  86   // Decrement reference counter so it can go away if it's unused
  87   ~TempNewSymbol() {
  88     if (_temp != NULL) {
  89       _temp->decrement_refcount();
  90     }
  91   }
  92 
  93   // Symbol* conversion operators
  94   Symbol* operator -> () const                   { return _temp; }
  95   bool    operator == (Symbol* o) const          { return _temp == o; }
  96   operator Symbol*()                             { return _temp; }
  97 };
  98 
  99 template <class T, class N> class CompactHashtable;
 100 
 101 class SymbolTable : public RehashableHashtable<Symbol*, mtSymbol> {
 102   friend class VMStructs;
 103   friend class ClassFileParser;
 104 
 105 private:
 106   // The symbol table
 107   static SymbolTable* _the_table;
 108 
 109   // Set if one bucket is out of balance due to hash algorithm deficiency
 110   static bool _needs_rehashing;
 111   static bool _lookup_shared_first;
 112 
 113   // For statistics
 114   static int _symbols_removed;
 115   static int _symbols_counted;
 116 
 117   // shared symbol table.
 118   static CompactHashtable<Symbol*, char> _shared_table;
 119 
 120   Symbol* allocate_symbol(const u1* name, int len, bool c_heap, TRAPS); // Assumes no characters larger than 0x7F
 121 
 122   // Adding elements
 123   Symbol* basic_add(int index, u1* name, int len, unsigned int hashValue,
 124                     bool c_heap, TRAPS);
 125   bool basic_add(ClassLoaderData* loader_data,
 126                  const constantPoolHandle& cp, int names_count,
 127                  const char** names, int* lengths, int* cp_indices,
 128                  unsigned int* hashValues, TRAPS);
 129 
 130   static void new_symbols(ClassLoaderData* loader_data,
 131                           const constantPoolHandle& cp, int names_count,
 132                           const char** name, int* lengths,
 133                           int* cp_indices, unsigned int* hashValues,
 134                           TRAPS) {
 135     add(loader_data, cp, names_count, name, lengths, cp_indices, hashValues, THREAD);
 136   }
 137 
 138   static Symbol* lookup_shared(const char* name, int len, unsigned int hash);
 139   Symbol* lookup_dynamic(int index, const char* name, int len, unsigned int hash);
 140   Symbol* lookup(int index, const char* name, int len, unsigned int hash);
 141 
 142   SymbolTable()
 143     : RehashableHashtable<Symbol*, mtSymbol>(SymbolTableSize, sizeof (HashtableEntry<Symbol*, mtSymbol>)) {}
 144 
 145   SymbolTable(HashtableBucket<mtSymbol>* t, int number_of_entries)
 146     : RehashableHashtable<Symbol*, mtSymbol>(SymbolTableSize, sizeof (HashtableEntry<Symbol*, mtSymbol>), t,
 147                 number_of_entries) {}
 148 
 149   // Arena for permanent symbols (null class loader) that are never unloaded
 150   static Arena*  _arena;
 151   static Arena* arena() { return _arena; }  // called for statistics
 152 
 153   static void initialize_symbols(int arena_alloc_size = 0);
 154 
 155   static volatile int _parallel_claimed_idx;
 156 
 157   typedef SymbolTable::BucketUnlinkContext BucketUnlinkContext;
 158   // Release any dead symbols. Unlinked bucket entries are collected in the given
 159   // context to be freed later.
 160   // This allows multiple threads to work on the table at once.
 161   static void buckets_unlink(int start_idx, int end_idx, BucketUnlinkContext* context);
 162 public:
 163   enum {
 164     symbol_alloc_batch_size = 8,
 165     // Pick initial size based on java -version size measurements
 166     symbol_alloc_arena_size = 360*K
 167   };
 168 
 169   // The symbol table
 170   static SymbolTable* the_table() { return _the_table; }
 171 
 172   // Size of one bucket in the string table.  Used when checking for rollover.
 173   static uint bucket_size() { return sizeof(HashtableBucket<mtSymbol>); }
 174 
 175   static void create_table() {
 176     assert(_the_table == NULL, "One symbol table allowed.");
 177     _the_table = new SymbolTable();
 178     initialize_symbols(symbol_alloc_arena_size);
 179   }
 180 
 181   static unsigned int hash_symbol(const char* s, int len);
 182   static unsigned int hash_shared_symbol(const char* s, int len);
 183 
 184   static Symbol* lookup(const char* name, int len, TRAPS);
 185   // lookup only, won't add. Also calculate hash.
 186   static Symbol* lookup_only(const char* name, int len, unsigned int& hash);
 187   // Only copy to C string to be added if lookup failed.
 188   static Symbol* lookup(const Symbol* sym, int begin, int end, TRAPS);
 189 
 190   static void release(Symbol* sym);
 191 
 192   // Look up the address of the literal in the SymbolTable for this Symbol*
 193   static Symbol** lookup_symbol_addr(Symbol* sym);
 194 
 195   // jchar (UTF16) version of lookups
 196   static Symbol* lookup_unicode(const jchar* name, int len, TRAPS);
 197   static Symbol* lookup_only_unicode(const jchar* name, int len, unsigned int& hash);
 198 
 199   static void add(ClassLoaderData* loader_data,
 200                   const constantPoolHandle& cp, int names_count,
 201                   const char** names, int* lengths, int* cp_indices,
 202                   unsigned int* hashValues, TRAPS);
 203 
 204   // Release any dead symbols
 205   static void unlink() {
 206     int processed = 0;
 207     int removed = 0;
 208     unlink(&processed, &removed);
 209   }
 210   static void unlink(int* processed, int* removed);
 211   // Release any dead symbols, possibly parallel version
 212   static void possibly_parallel_unlink(int* processed, int* removed);
 213 
 214   // iterate over symbols
 215   static void symbols_do(SymbolClosure *cl);
 216   static void metaspace_pointers_do(MetaspaceClosure* it);
 217 
 218   // Symbol creation
 219   static Symbol* new_symbol(const char* utf8_buffer, int length, TRAPS) {
 220     assert(utf8_buffer != NULL, "just checking");
 221     return lookup(utf8_buffer, length, THREAD);
 222   }
 223   static Symbol*       new_symbol(const char* name, TRAPS) {
 224     return new_symbol(name, (int)strlen(name), THREAD);
 225   }
 226   static Symbol*       new_symbol(const Symbol* sym, int begin, int end, TRAPS) {
 227     assert(begin <= end && end <= sym->utf8_length(), "just checking");
 228     return lookup(sym, begin, end, THREAD);
 229   }
 230 
 231   // Create a symbol in the arena for symbols that are not deleted
 232   static Symbol* new_permanent_symbol(const char* name, TRAPS);
 233 
 234   // Symbol lookup
 235   static Symbol* lookup(int index, const char* name, int len, TRAPS);
 236 
 237   // Needed for preloading classes in signatures when compiling.
 238   // Returns the symbol is already present in symbol table, otherwise
 239   // NULL.  NO ALLOCATION IS GUARANTEED!
 240   static Symbol* probe(const char* name, int len) {
 241     unsigned int ignore_hash;
 242     return lookup_only(name, len, ignore_hash);
 243   }
 244   static Symbol* probe_unicode(const jchar* name, int len) {
 245     unsigned int ignore_hash;
 246     return lookup_only_unicode(name, len, ignore_hash);
 247   }
 248 
 249   // Histogram
 250   static void print_histogram()     PRODUCT_RETURN;
 251   static void print()     PRODUCT_RETURN;
 252 
 253   // Debugging
 254   static void verify();
 255   static void dump(outputStream* st, bool verbose=false);
 256   static void read(const char* filename, TRAPS);
 257 
 258   // Sharing
 259   static void write_to_archive();
 260   static void serialize(SerializeClosure* soc);
 261   static u4 encode_shared(Symbol* sym);
 262   static Symbol* decode_shared(u4 offset);
 263 
 264   // Rehash the symbol table if it gets out of balance
 265   static void rehash_table();
 266   static bool needs_rehashing()         { return _needs_rehashing; }
 267   // Parallel chunked scanning
 268   static void clear_parallel_claimed_index() { _parallel_claimed_idx = 0; }
 269   static int parallel_claimed_index()        { return _parallel_claimed_idx; }
 270 };
 271 
 272 #endif // SHARE_VM_CLASSFILE_SYMBOLTABLE_HPP