duke@435: /*
coleenp@4718:  * Copyright (c) 1997, 2013, Oracle and/or its affiliates. All rights reserved.
duke@435:  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
duke@435:  *
duke@435:  * This code is free software; you can redistribute it and/or modify it
duke@435:  * under the terms of the GNU General Public License version 2 only, as
duke@435:  * published by the Free Software Foundation.
duke@435:  *
duke@435:  * This code is distributed in the hope that it will be useful, but WITHOUT
duke@435:  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
duke@435:  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
duke@435:  * version 2 for more details (a copy is included in the LICENSE file that
duke@435:  * accompanied this code).
duke@435:  *
duke@435:  * You should have received a copy of the GNU General Public License version
duke@435:  * 2 along with this work; if not, write to the Free Software Foundation,
duke@435:  * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
duke@435:  *
trims@1907:  * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
trims@1907:  * or visit www.oracle.com if you need additional information or have any
trims@1907:  * questions.
duke@435:  *
duke@435:  */
duke@435: 
stefank@2314: #include "precompiled.hpp"
coleenp@3865: #include "classfile/altHashing.hpp"
stefank@2314: #include "classfile/javaClasses.hpp"
stefank@2314: #include "classfile/symbolTable.hpp"
stefank@2314: #include "classfile/systemDictionary.hpp"
stefank@2314: #include "gc_interface/collectedHeap.inline.hpp"
coleenp@3682: #include "memory/allocation.inline.hpp"
stefank@2314: #include "memory/filemap.hpp"
stefank@2314: #include "memory/gcLocker.inline.hpp"
stefank@2314: #include "oops/oop.inline.hpp"
stefank@2314: #include "oops/oop.inline2.hpp"
stefank@2314: #include "runtime/mutexLocker.hpp"
stefank@2314: #include "utilities/hashtable.inline.hpp"
duke@435: 
duke@435: // --------------------------------------------------------------------------
duke@435: 
duke@435: SymbolTable* SymbolTable::_the_table = NULL;
coleenp@3682: // Static arena for symbols that are not deallocated
coleenp@3682: Arena* SymbolTable::_arena = NULL;
coleenp@3865: bool SymbolTable::_needs_rehashing = false;
duke@435: 
coleenp@3682: Symbol* SymbolTable::allocate_symbol(const u1* name, int len, bool c_heap, TRAPS) {
coleenp@3875:   assert (len <= Symbol::max_length(), "should be checked by caller");
coleenp@3875: 
coleenp@3682:   Symbol* sym;
coleenp@4037: 
coleenp@4718:   if (DumpSharedSpaces) {
coleenp@4718:     // Allocate all symbols to CLD shared metaspace
coleenp@4718:     sym = new (len, ClassLoaderData::the_null_class_loader_data(), THREAD) Symbol(name, len, -1);
coleenp@4718:   } else if (c_heap) {
coleenp@3682:     // refcount starts as 1
coleenp@3682:     sym = new (len, THREAD) Symbol(name, len, 1);
coleenp@4037:     assert(sym != NULL, "new should call vm_exit_out_of_memory if C_HEAP is exhausted");
coleenp@4037:   } else {
coleenp@4718:     // Allocate to global arena
coleenp@3682:     sym = new (len, arena(), THREAD) Symbol(name, len, -1);
coleenp@3682:   }
coleenp@2497:   return sym;
coleenp@2497: }
coleenp@2497: 
coleenp@3682: void SymbolTable::initialize_symbols(int arena_alloc_size) {
coleenp@3682:   // Initialize the arena for global symbols, size passed in depends on CDS.
coleenp@3682:   if (arena_alloc_size == 0) {
zgu@3900:     _arena = new (mtSymbol) Arena();
coleenp@3682:   } else {
zgu@3900:     _arena = new (mtSymbol) Arena(arena_alloc_size);
coleenp@2497:   }
coleenp@2497: }
coleenp@2497: 
coleenp@2497: // Call function for all symbols in the symbol table.
coleenp@2497: void SymbolTable::symbols_do(SymbolClosure *cl) {
coleenp@2497:   const int n = the_table()->table_size();
coleenp@2497:   for (int i = 0; i < n; i++) {
zgu@3900:     for (HashtableEntry<Symbol*, mtSymbol>* p = the_table()->bucket(i);
coleenp@2497:          p != NULL;
coleenp@2497:          p = p->next()) {
coleenp@2497:       cl->do_symbol(p->literal_addr());
coleenp@2497:     }
coleenp@2497:   }
coleenp@2497: }
coleenp@2497: 
coleenp@2497: int SymbolTable::symbols_removed = 0;
coleenp@2497: int SymbolTable::symbols_counted = 0;
coleenp@2497: 
coleenp@2497: // Remove unreferenced symbols from the symbol table
coleenp@3682: // This is done late during GC.
coleenp@2497: void SymbolTable::unlink() {
coleenp@2497:   int removed = 0;
coleenp@2497:   int total = 0;
coleenp@2618:   size_t memory_total = 0;
coleenp@2497:   for (int i = 0; i < the_table()->table_size(); ++i) {
zgu@3900:     HashtableEntry<Symbol*, mtSymbol>** p = the_table()->bucket_addr(i);
zgu@3900:     HashtableEntry<Symbol*, mtSymbol>* entry = the_table()->bucket(i);
coleenp@3875:     while (entry != NULL) {
coleenp@3875:       // Shared entries are normally at the end of the bucket and if we run into
coleenp@3875:       // a shared entry, then there is nothing more to remove. However, if we
coleenp@3875:       // have rehashed the table, then the shared entries are no longer at the
coleenp@3875:       // end of the bucket.
coleenp@3875:       if (entry->is_shared() && !use_alternate_hashcode()) {
coleenp@2497:         break;
coleenp@2497:       }
coleenp@2497:       Symbol* s = entry->literal();
coleenp@4037:       memory_total += s->size();
coleenp@2497:       total++;
coleenp@2497:       assert(s != NULL, "just checking");
coleenp@2497:       // If reference count is zero, remove.
coleenp@2497:       if (s->refcount() == 0) {
coleenp@3875:         assert(!entry->is_shared(), "shared entries should be kept live");
coleenp@2497:         delete s;
coleenp@2497:         removed++;
coleenp@2497:         *p = entry->next();
coleenp@2497:         the_table()->free_entry(entry);
coleenp@2497:       } else {
coleenp@2497:         p = entry->next_addr();
coleenp@2497:       }
coleenp@3875:       // get next entry
zgu@3900:       entry = (HashtableEntry<Symbol*, mtSymbol>*)HashtableEntry<Symbol*, mtSymbol>::make_ptr(*p);
coleenp@2497:     }
coleenp@2497:   }
coleenp@2497:   symbols_removed += removed;
coleenp@2497:   symbols_counted += total;
coleenp@2618:   // Exclude printing for normal PrintGCDetails because people parse
coleenp@2618:   // this output.
coleenp@2618:   if (PrintGCDetails && Verbose && WizardMode) {
coleenp@2618:     gclog_or_tty->print(" [Symbols=%d size=" SIZE_FORMAT "K] ", total,
coleenp@2497:                         (memory_total*HeapWordSize)/1024);
coleenp@2497:   }
coleenp@2497: }
coleenp@2497: 
coleenp@3865: // Create a new table and using alternate hash code, populate the new table
coleenp@3865: // with the existing strings.   Set flag to use the alternate hash code afterwards.
coleenp@3865: void SymbolTable::rehash_table() {
coleenp@3865:   assert(SafepointSynchronize::is_at_safepoint(), "must be at safepoint");
coleenp@3875:   // This should never happen with -Xshare:dump but it might in testing mode.
coleenp@3875:   if (DumpSharedSpaces) return;
coleenp@3865:   // Create a new symbol table
coleenp@3865:   SymbolTable* new_table = new SymbolTable();
coleenp@3865: 
coleenp@3865:   the_table()->move_to(new_table);
coleenp@3865: 
coleenp@3865:   // Delete the table and buckets (entries are reused in new table).
coleenp@3865:   delete _the_table;
coleenp@3865:   // Don't check if we need rehashing until the table gets unbalanced again.
coleenp@3865:   // Then rehash with a new global seed.
coleenp@3865:   _needs_rehashing = false;
coleenp@3865:   _the_table = new_table;
coleenp@3865: }
coleenp@2497: 
duke@435: // Lookup a symbol in a bucket.
duke@435: 
coleenp@2497: Symbol* SymbolTable::lookup(int index, const char* name,
duke@435:                               int len, unsigned int hash) {
coleenp@3865:   int count = 0;
zgu@3900:   for (HashtableEntry<Symbol*, mtSymbol>* e = bucket(index); e != NULL; e = e->next()) {
coleenp@3865:     count++;  // count all entries in this bucket, not just ones with same hash
duke@435:     if (e->hash() == hash) {
coleenp@2497:       Symbol* sym = e->literal();
duke@435:       if (sym->equals(name, len)) {
coleenp@2497:         // something is referencing this symbol now.
coleenp@2497:         sym->increment_refcount();
duke@435:         return sym;
duke@435:       }
duke@435:     }
duke@435:   }
coleenp@3865:   // If the bucket size is too deep check if this hash code is insufficient.
zgu@3900:   if (count >= BasicHashtable<mtSymbol>::rehash_count && !needs_rehashing()) {
coleenp@3865:     _needs_rehashing = check_rehash_table(count);
coleenp@3865:   }
duke@435:   return NULL;
duke@435: }
duke@435: 
coleenp@3875: // Pick hashing algorithm.
coleenp@3875: unsigned int SymbolTable::hash_symbol(const char* s, int len) {
coleenp@3865:   return use_alternate_hashcode() ?
coleenp@3865:            AltHashing::murmur3_32(seed(), (const jbyte*)s, len) :
brutisso@4335:            java_lang_String::hash_code(s, len);
coleenp@3865: }
coleenp@3865: 
duke@435: 
duke@435: // We take care not to be blocking while holding the
duke@435: // SymbolTable_lock. Otherwise, the system might deadlock, since the
duke@435: // symboltable is used during compilation (VM_thread) The lock free
duke@435: // synchronization is simplified by the fact that we do not delete
duke@435: // entries in the symbol table during normal execution (only during
duke@435: // safepoints).
duke@435: 
coleenp@2497: Symbol* SymbolTable::lookup(const char* name, int len, TRAPS) {
duke@435:   unsigned int hashValue = hash_symbol(name, len);
duke@435:   int index = the_table()->hash_to_index(hashValue);
duke@435: 
coleenp@2497:   Symbol* s = the_table()->lookup(index, name, len, hashValue);
duke@435: 
duke@435:   // Found
duke@435:   if (s != NULL) return s;
duke@435: 
coleenp@3875:   // Grab SymbolTable_lock first.
coleenp@3875:   MutexLocker ml(SymbolTable_lock, THREAD);
coleenp@3875: 
duke@435:   // Otherwise, add to symbol to table
coleenp@3682:   return the_table()->basic_add(index, (u1*)name, len, hashValue, true, CHECK_NULL);
duke@435: }
duke@435: 
coleenp@2497: Symbol* SymbolTable::lookup(const Symbol* sym, int begin, int end, TRAPS) {
duke@435:   char* buffer;
duke@435:   int index, len;
duke@435:   unsigned int hashValue;
duke@435:   char* name;
duke@435:   {
duke@435:     debug_only(No_Safepoint_Verifier nsv;)
duke@435: 
duke@435:     name = (char*)sym->base() + begin;
duke@435:     len = end - begin;
duke@435:     hashValue = hash_symbol(name, len);
duke@435:     index = the_table()->hash_to_index(hashValue);
coleenp@2497:     Symbol* s = the_table()->lookup(index, name, len, hashValue);
duke@435: 
duke@435:     // Found
duke@435:     if (s != NULL) return s;
duke@435:   }
duke@435: 
duke@435:   // Otherwise, add to symbol to table. Copy to a C string first.
duke@435:   char stack_buf[128];
duke@435:   ResourceMark rm(THREAD);
duke@435:   if (len <= 128) {
duke@435:     buffer = stack_buf;
duke@435:   } else {
duke@435:     buffer = NEW_RESOURCE_ARRAY_IN_THREAD(THREAD, char, len);
duke@435:   }
duke@435:   for (int i=0; i<len; i++) {
duke@435:     buffer[i] = name[i];
duke@435:   }
duke@435:   // Make sure there is no safepoint in the code above since name can't move.
duke@435:   // We can't include the code in No_Safepoint_Verifier because of the
duke@435:   // ResourceMark.
duke@435: 
coleenp@3875:   // Grab SymbolTable_lock first.
coleenp@3875:   MutexLocker ml(SymbolTable_lock, THREAD);
coleenp@3875: 
coleenp@3682:   return the_table()->basic_add(index, (u1*)buffer, len, hashValue, true, CHECK_NULL);
duke@435: }
duke@435: 
coleenp@2497: Symbol* SymbolTable::lookup_only(const char* name, int len,
duke@435:                                    unsigned int& hash) {
duke@435:   hash = hash_symbol(name, len);
duke@435:   int index = the_table()->hash_to_index(hash);
duke@435: 
coleenp@2497:   Symbol* s = the_table()->lookup(index, name, len, hash);
coleenp@2497:   return s;
duke@435: }
duke@435: 
phh@3427: // Look up the address of the literal in the SymbolTable for this Symbol*
phh@3427: // Do not create any new symbols
phh@3427: // Do not increment the reference count to keep this alive
phh@3427: Symbol** SymbolTable::lookup_symbol_addr(Symbol* sym){
phh@3427:   unsigned int hash = hash_symbol((char*)sym->bytes(), sym->utf8_length());
phh@3427:   int index = the_table()->hash_to_index(hash);
phh@3427: 
zgu@3900:   for (HashtableEntry<Symbol*, mtSymbol>* e = the_table()->bucket(index); e != NULL; e = e->next()) {
phh@3427:     if (e->hash() == hash) {
phh@3427:       Symbol* literal_sym = e->literal();
phh@3427:       if (sym == literal_sym) {
phh@3427:         return e->literal_addr();
phh@3427:       }
phh@3427:     }
phh@3427:   }
phh@3427:   return NULL;
phh@3427: }
phh@3427: 
jrose@1100: // Suggestion: Push unicode-based lookup all the way into the hashing
jrose@1100: // and probing logic, so there is no need for convert_to_utf8 until
coleenp@2497: // an actual new Symbol* is created.
coleenp@2497: Symbol* SymbolTable::lookup_unicode(const jchar* name, int utf16_length, TRAPS) {
jrose@1100:   int utf8_length = UNICODE::utf8_length((jchar*) name, utf16_length);
jrose@1100:   char stack_buf[128];
jrose@1100:   if (utf8_length < (int) sizeof(stack_buf)) {
jrose@1100:     char* chars = stack_buf;
jrose@1100:     UNICODE::convert_to_utf8(name, utf16_length, chars);
jrose@1100:     return lookup(chars, utf8_length, THREAD);
jrose@1100:   } else {
jrose@1100:     ResourceMark rm(THREAD);
jrose@1100:     char* chars = NEW_RESOURCE_ARRAY(char, utf8_length + 1);;
jrose@1100:     UNICODE::convert_to_utf8(name, utf16_length, chars);
jrose@1100:     return lookup(chars, utf8_length, THREAD);
jrose@1100:   }
jrose@1100: }
jrose@1100: 
coleenp@2497: Symbol* SymbolTable::lookup_only_unicode(const jchar* name, int utf16_length,
jrose@1100:                                            unsigned int& hash) {
jrose@1100:   int utf8_length = UNICODE::utf8_length((jchar*) name, utf16_length);
jrose@1100:   char stack_buf[128];
jrose@1100:   if (utf8_length < (int) sizeof(stack_buf)) {
jrose@1100:     char* chars = stack_buf;
jrose@1100:     UNICODE::convert_to_utf8(name, utf16_length, chars);
jrose@1100:     return lookup_only(chars, utf8_length, hash);
jrose@1100:   } else {
jrose@1100:     ResourceMark rm;
jrose@1100:     char* chars = NEW_RESOURCE_ARRAY(char, utf8_length + 1);;
jrose@1100:     UNICODE::convert_to_utf8(name, utf16_length, chars);
jrose@1100:     return lookup_only(chars, utf8_length, hash);
jrose@1100:   }
jrose@1100: }
jrose@1100: 
coleenp@4037: void SymbolTable::add(ClassLoaderData* loader_data, constantPoolHandle cp,
coleenp@3682:                       int names_count,
duke@435:                       const char** names, int* lengths, int* cp_indices,
duke@435:                       unsigned int* hashValues, TRAPS) {
coleenp@3875:   // Grab SymbolTable_lock first.
coleenp@3875:   MutexLocker ml(SymbolTable_lock, THREAD);
coleenp@3875: 
duke@435:   SymbolTable* table = the_table();
coleenp@4037:   bool added = table->basic_add(loader_data, cp, names_count, names, lengths,
duke@435:                                 cp_indices, hashValues, CHECK);
duke@435:   if (!added) {
duke@435:     // do it the hard way
duke@435:     for (int i=0; i<names_count; i++) {
duke@435:       int index = table->hash_to_index(hashValues[i]);
coleenp@4037:       bool c_heap = !loader_data->is_the_null_class_loader_data();
coleenp@3682:       Symbol* sym = table->basic_add(index, (u1*)names[i], lengths[i], hashValues[i], c_heap, CHECK);
duke@435:       cp->symbol_at_put(cp_indices[i], sym);
duke@435:     }
duke@435:   }
duke@435: }
duke@435: 
coleenp@3682: Symbol* SymbolTable::new_permanent_symbol(const char* name, TRAPS) {
coleenp@3682:   unsigned int hash;
coleenp@3682:   Symbol* result = SymbolTable::lookup_only((char*)name, (int)strlen(name), hash);
coleenp@3682:   if (result != NULL) {
coleenp@3682:     return result;
coleenp@3682:   }
coleenp@3875:   // Grab SymbolTable_lock first.
coleenp@3875:   MutexLocker ml(SymbolTable_lock, THREAD);
coleenp@3875: 
coleenp@3682:   SymbolTable* table = the_table();
coleenp@3682:   int index = table->hash_to_index(hash);
coleenp@3682:   return table->basic_add(index, (u1*)name, (int)strlen(name), hash, false, THREAD);
coleenp@3682: }
coleenp@3682: 
coleenp@3875: Symbol* SymbolTable::basic_add(int index_arg, u1 *name, int len,
coleenp@3865:                                unsigned int hashValue_arg, bool c_heap, TRAPS) {
duke@435:   assert(!Universe::heap()->is_in_reserved(name) || GC_locker::is_active(),
duke@435:          "proposed name of symbol must be stable");
duke@435: 
coleenp@3875:   // Don't allow symbols to be created which cannot fit in a Symbol*.
coleenp@3875:   if (len > Symbol::max_length()) {
coleenp@3875:     THROW_MSG_0(vmSymbols::java_lang_InternalError(),
coleenp@3875:                 "name is too long to represent");
coleenp@3875:   }
coleenp@3875: 
coleenp@3875:   // Cannot hit a safepoint in this function because the "this" pointer can move.
coleenp@3875:   No_Safepoint_Verifier nsv;
duke@435: 
coleenp@3865:   // Check if the symbol table has been rehashed, if so, need to recalculate
coleenp@3875:   // the hash value and index.
coleenp@3875:   unsigned int hashValue;
coleenp@3875:   int index;
coleenp@3875:   if (use_alternate_hashcode()) {
coleenp@3875:     hashValue = hash_symbol((const char*)name, len);
coleenp@3875:     index = hash_to_index(hashValue);
coleenp@3875:   } else {
coleenp@3875:     hashValue = hashValue_arg;
coleenp@3875:     index = index_arg;
coleenp@3875:   }
coleenp@3865: 
duke@435:   // Since look-up was done lock-free, we need to check if another
duke@435:   // thread beat us in the race to insert the symbol.
coleenp@2497:   Symbol* test = lookup(index, (char*)name, len, hashValue);
duke@435:   if (test != NULL) {
coleenp@3682:     // A race occurred and another thread introduced the symbol.
coleenp@2497:     assert(test->refcount() != 0, "lookup should have incremented the count");
duke@435:     return test;
duke@435:   }
duke@435: 
coleenp@3682:   // Create a new symbol.
coleenp@3682:   Symbol* sym = allocate_symbol(name, len, c_heap, CHECK_NULL);
coleenp@3682:   assert(sym->equals((char*)name, len), "symbol must be properly initialized");
coleenp@3682: 
zgu@3900:   HashtableEntry<Symbol*, mtSymbol>* entry = new_entry(hashValue, sym);
duke@435:   add_entry(index, entry);
coleenp@2497:   return sym;
duke@435: }
duke@435: 
coleenp@3682: // This version of basic_add adds symbols in batch from the constant pool
coleenp@3682: // parsing.
coleenp@4037: bool SymbolTable::basic_add(ClassLoaderData* loader_data, constantPoolHandle cp,
coleenp@3682:                             int names_count,
duke@435:                             const char** names, int* lengths,
duke@435:                             int* cp_indices, unsigned int* hashValues,
duke@435:                             TRAPS) {
coleenp@3682: 
coleenp@3682:   // Check symbol names are not too long.  If any are too long, don't add any.
coleenp@3682:   for (int i = 0; i< names_count; i++) {
coleenp@3682:     if (lengths[i] > Symbol::max_length()) {
coleenp@3682:       THROW_MSG_0(vmSymbols::java_lang_InternalError(),
coleenp@3682:                   "name is too long to represent");
coleenp@3682:     }
duke@435:   }
duke@435: 
coleenp@3875:   // Cannot hit a safepoint in this function because the "this" pointer can move.
coleenp@3875:   No_Safepoint_Verifier nsv;
duke@435: 
coleenp@2497:   for (int i=0; i<names_count; i++) {
coleenp@3865:     // Check if the symbol table has been rehashed, if so, need to recalculate
coleenp@3865:     // the hash value.
coleenp@3875:     unsigned int hashValue;
coleenp@3875:     if (use_alternate_hashcode()) {
coleenp@3875:       hashValue = hash_symbol(names[i], lengths[i]);
coleenp@3875:     } else {
coleenp@3875:       hashValue = hashValues[i];
coleenp@3875:     }
duke@435:     // Since look-up was done lock-free, we need to check if another
duke@435:     // thread beat us in the race to insert the symbol.
coleenp@3865:     int index = hash_to_index(hashValue);
coleenp@3865:     Symbol* test = lookup(index, names[i], lengths[i], hashValue);
duke@435:     if (test != NULL) {
twisti@1040:       // A race occurred and another thread introduced the symbol, this one
duke@435:       // will be dropped and collected. Use test instead.
duke@435:       cp->symbol_at_put(cp_indices[i], test);
coleenp@2497:       assert(test->refcount() != 0, "lookup should have incremented the count");
duke@435:     } else {
coleenp@3682:       // Create a new symbol.  The null class loader is never unloaded so these
coleenp@3682:       // are allocated specially in a permanent arena.
coleenp@4037:       bool c_heap = !loader_data->is_the_null_class_loader_data();
coleenp@3682:       Symbol* sym = allocate_symbol((const u1*)names[i], lengths[i], c_heap, CHECK_(false));
coleenp@3682:       assert(sym->equals(names[i], lengths[i]), "symbol must be properly initialized");  // why wouldn't it be???
zgu@3900:       HashtableEntry<Symbol*, mtSymbol>* entry = new_entry(hashValue, sym);
duke@435:       add_entry(index, entry);
duke@435:       cp->symbol_at_put(cp_indices[i], sym);
duke@435:     }
duke@435:   }
duke@435:   return true;
duke@435: }
duke@435: 
duke@435: 
duke@435: void SymbolTable::verify() {
duke@435:   for (int i = 0; i < the_table()->table_size(); ++i) {
zgu@3900:     HashtableEntry<Symbol*, mtSymbol>* p = the_table()->bucket(i);
duke@435:     for ( ; p != NULL; p = p->next()) {
coleenp@2497:       Symbol* s = (Symbol*)(p->literal());
duke@435:       guarantee(s != NULL, "symbol is NULL");
duke@435:       unsigned int h = hash_symbol((char*)s->bytes(), s->utf8_length());
duke@435:       guarantee(p->hash() == h, "broken hash in symbol table entry");
duke@435:       guarantee(the_table()->hash_to_index(h) == i,
duke@435:                 "wrong index in symbol table");
duke@435:     }
duke@435:   }
duke@435: }
duke@435: 
coleenp@3865: void SymbolTable::dump(outputStream* st) {
iklam@5144:   the_table()->dump_table(st, "SymbolTable");
coleenp@3865: }
coleenp@3865: 
duke@435: 
duke@435: //---------------------------------------------------------------------------
duke@435: // Non-product code
duke@435: 
duke@435: #ifndef PRODUCT
duke@435: 
duke@435: void SymbolTable::print_histogram() {
duke@435:   MutexLocker ml(SymbolTable_lock);
duke@435:   const int results_length = 100;
duke@435:   int results[results_length];
duke@435:   int i,j;
duke@435: 
duke@435:   // initialize results to zero
duke@435:   for (j = 0; j < results_length; j++) {
duke@435:     results[j] = 0;
duke@435:   }
duke@435: 
duke@435:   int total = 0;
duke@435:   int max_symbols = 0;
duke@435:   int out_of_range = 0;
coleenp@2497:   int memory_total = 0;
coleenp@2497:   int count = 0;
duke@435:   for (i = 0; i < the_table()->table_size(); i++) {
zgu@3900:     HashtableEntry<Symbol*, mtSymbol>* p = the_table()->bucket(i);
duke@435:     for ( ; p != NULL; p = p->next()) {
coleenp@4037:       memory_total += p->literal()->size();
coleenp@2497:       count++;
coleenp@2497:       int counter = p->literal()->utf8_length();
duke@435:       total += counter;
duke@435:       if (counter < results_length) {
duke@435:         results[counter]++;
duke@435:       } else {
duke@435:         out_of_range++;
duke@435:       }
duke@435:       max_symbols = MAX2(max_symbols, counter);
duke@435:     }
duke@435:   }
duke@435:   tty->print_cr("Symbol Table:");
coleenp@2497:   tty->print_cr("Total number of symbols  %5d", count);
coleenp@2497:   tty->print_cr("Total size in memory     %5dK",
coleenp@2497:           (memory_total*HeapWordSize)/1024);
coleenp@2497:   tty->print_cr("Total counted            %5d", symbols_counted);
coleenp@2497:   tty->print_cr("Total removed            %5d", symbols_removed);
coleenp@2497:   if (symbols_counted > 0) {
coleenp@2497:     tty->print_cr("Percent removed          %3.2f",
coleenp@2497:           ((float)symbols_removed/(float)symbols_counted)* 100);
coleenp@2497:   }
coleenp@2497:   tty->print_cr("Reference counts         %5d", Symbol::_total_count);
coleenp@3682:   tty->print_cr("Symbol arena size        %5d used %5d",
coleenp@3682:                  arena()->size_in_bytes(), arena()->used());
coleenp@2497:   tty->print_cr("Histogram of symbol length:");
duke@435:   tty->print_cr("%8s %5d", "Total  ", total);
duke@435:   tty->print_cr("%8s %5d", "Maximum", max_symbols);
duke@435:   tty->print_cr("%8s %3.2f", "Average",
duke@435:           ((float) total / (float) the_table()->table_size()));
duke@435:   tty->print_cr("%s", "Histogram:");
duke@435:   tty->print_cr(" %s %29s", "Length", "Number chains that length");
duke@435:   for (i = 0; i < results_length; i++) {
duke@435:     if (results[i] > 0) {
duke@435:       tty->print_cr("%6d %10d", i, results[i]);
duke@435:     }
duke@435:   }
coleenp@2497:   if (Verbose) {
coleenp@2497:     int line_length = 70;
coleenp@2497:     tty->print_cr("%s %30s", " Length", "Number chains that length");
coleenp@2497:     for (i = 0; i < results_length; i++) {
coleenp@2497:       if (results[i] > 0) {
coleenp@2497:         tty->print("%4d", i);
coleenp@2497:         for (j = 0; (j < results[i]) && (j < line_length);  j++) {
coleenp@2497:           tty->print("%1s", "*");
coleenp@2497:         }
coleenp@2497:         if (j == line_length) {
coleenp@2497:           tty->print("%1s", "+");
coleenp@2497:         }
coleenp@2497:         tty->cr();
duke@435:       }
coleenp@2497:     }
coleenp@2497:   }
coleenp@2497:   tty->print_cr(" %s %d: %d\n", "Number chains longer than",
coleenp@2497:                     results_length, out_of_range);
coleenp@2497: }
coleenp@2497: 
coleenp@2497: void SymbolTable::print() {
coleenp@2497:   for (int i = 0; i < the_table()->table_size(); ++i) {
zgu@3900:     HashtableEntry<Symbol*, mtSymbol>** p = the_table()->bucket_addr(i);
zgu@3900:     HashtableEntry<Symbol*, mtSymbol>* entry = the_table()->bucket(i);
coleenp@2497:     if (entry != NULL) {
coleenp@2497:       while (entry != NULL) {
coleenp@2497:         tty->print(PTR_FORMAT " ", entry->literal());
coleenp@2497:         entry->literal()->print();
coleenp@2497:         tty->print(" %d", entry->literal()->refcount());
coleenp@2497:         p = entry->next_addr();
zgu@3900:         entry = (HashtableEntry<Symbol*, mtSymbol>*)HashtableEntry<Symbol*, mtSymbol>::make_ptr(*p);
duke@435:       }
duke@435:       tty->cr();
duke@435:     }
duke@435:   }
duke@435: }
duke@435: #endif // PRODUCT
duke@435: 
duke@435: // --------------------------------------------------------------------------
duke@435: 
duke@435: #ifdef ASSERT
duke@435: class StableMemoryChecker : public StackObj {
duke@435:   enum { _bufsize = wordSize*4 };
duke@435: 
duke@435:   address _region;
duke@435:   jint    _size;
duke@435:   u1      _save_buf[_bufsize];
duke@435: 
duke@435:   int sample(u1* save_buf) {
duke@435:     if (_size <= _bufsize) {
duke@435:       memcpy(save_buf, _region, _size);
duke@435:       return _size;
duke@435:     } else {
duke@435:       // copy head and tail
duke@435:       memcpy(&save_buf[0],          _region,                      _bufsize/2);
duke@435:       memcpy(&save_buf[_bufsize/2], _region + _size - _bufsize/2, _bufsize/2);
duke@435:       return (_bufsize/2)*2;
duke@435:     }
duke@435:   }
duke@435: 
duke@435:  public:
duke@435:   StableMemoryChecker(const void* region, jint size) {
duke@435:     _region = (address) region;
duke@435:     _size   = size;
duke@435:     sample(_save_buf);
duke@435:   }
duke@435: 
duke@435:   bool verify() {
duke@435:     u1 check_buf[sizeof(_save_buf)];
duke@435:     int check_size = sample(check_buf);
duke@435:     return (0 == memcmp(_save_buf, check_buf, check_size));
duke@435:   }
duke@435: 
duke@435:   void set_region(const void* region) { _region = (address) region; }
duke@435: };
duke@435: #endif
duke@435: 
duke@435: 
duke@435: // --------------------------------------------------------------------------
duke@435: StringTable* StringTable::_the_table = NULL;
duke@435: 
coleenp@3865: bool StringTable::_needs_rehashing = false;
coleenp@3865: 
coleenp@3865: // Pick hashing algorithm
coleenp@3875: unsigned int StringTable::hash_string(const jchar* s, int len) {
coleenp@3865:   return use_alternate_hashcode() ? AltHashing::murmur3_32(seed(), s, len) :
brutisso@4335:                                     java_lang_String::hash_code(s, len);
coleenp@3865: }
coleenp@3865: 
duke@435: oop StringTable::lookup(int index, jchar* name,
duke@435:                         int len, unsigned int hash) {
coleenp@3865:   int count = 0;
zgu@3900:   for (HashtableEntry<oop, mtSymbol>* l = bucket(index); l != NULL; l = l->next()) {
coleenp@3865:     count++;
duke@435:     if (l->hash() == hash) {
duke@435:       if (java_lang_String::equals(l->literal(), name, len)) {
duke@435:         return l->literal();
duke@435:       }
duke@435:     }
duke@435:   }
coleenp@3865:   // If the bucket size is too deep check if this hash code is insufficient.
zgu@3900:   if (count >= BasicHashtable<mtSymbol>::rehash_count && !needs_rehashing()) {
coleenp@3865:     _needs_rehashing = check_rehash_table(count);
coleenp@3865:   }
duke@435:   return NULL;
duke@435: }
duke@435: 
duke@435: 
coleenp@3875: oop StringTable::basic_add(int index_arg, Handle string, jchar* name,
coleenp@3865:                            int len, unsigned int hashValue_arg, TRAPS) {
duke@435: 
duke@435:   assert(java_lang_String::equals(string(), name, len),
duke@435:          "string must be properly initialized");
coleenp@3875:   // Cannot hit a safepoint in this function because the "this" pointer can move.
coleenp@3875:   No_Safepoint_Verifier nsv;
duke@435: 
coleenp@3865:   // Check if the symbol table has been rehashed, if so, need to recalculate
coleenp@3875:   // the hash value and index before second lookup.
coleenp@3875:   unsigned int hashValue;
coleenp@3875:   int index;
coleenp@3875:   if (use_alternate_hashcode()) {
coleenp@3875:     hashValue = hash_string(name, len);
coleenp@3875:     index = hash_to_index(hashValue);
coleenp@3875:   } else {
coleenp@3875:     hashValue = hashValue_arg;
coleenp@3875:     index = index_arg;
coleenp@3875:   }
coleenp@3865: 
duke@435:   // Since look-up was done lock-free, we need to check if another
duke@435:   // thread beat us in the race to insert the symbol.
duke@435: 
duke@435:   oop test = lookup(index, name, len, hashValue); // calls lookup(u1*, int)
duke@435:   if (test != NULL) {
duke@435:     // Entry already added
duke@435:     return test;
duke@435:   }
duke@435: 
zgu@3900:   HashtableEntry<oop, mtSymbol>* entry = new_entry(hashValue, string());
duke@435:   add_entry(index, entry);
duke@435:   return string();
duke@435: }
duke@435: 
duke@435: 
coleenp@2497: oop StringTable::lookup(Symbol* symbol) {
duke@435:   ResourceMark rm;
duke@435:   int length;
duke@435:   jchar* chars = symbol->as_unicode(length);
mgerdin@4850:   return lookup(chars, length);
mgerdin@4850: }
mgerdin@4850: 
mgerdin@4850: 
mgerdin@4850: oop StringTable::lookup(jchar* name, int len) {
mgerdin@4850:   unsigned int hash = hash_string(name, len);
mgerdin@4850:   int index = the_table()->hash_to_index(hash);
mgerdin@4850:   return the_table()->lookup(index, name, len, hash);
duke@435: }
duke@435: 
duke@435: 
duke@435: oop StringTable::intern(Handle string_or_null, jchar* name,
duke@435:                         int len, TRAPS) {
coleenp@3865:   unsigned int hashValue = hash_string(name, len);
duke@435:   int index = the_table()->hash_to_index(hashValue);
coleenp@3875:   oop found_string = the_table()->lookup(index, name, len, hashValue);
duke@435: 
duke@435:   // Found
coleenp@3875:   if (found_string != NULL) return found_string;
coleenp@3875: 
coleenp@3875:   debug_only(StableMemoryChecker smc(name, len * sizeof(name[0])));
coleenp@3875:   assert(!Universe::heap()->is_in_reserved(name) || GC_locker::is_active(),
coleenp@3875:          "proposed name of symbol must be stable");
coleenp@3875: 
coleenp@3875:   Handle string;
coleenp@3875:   // try to reuse the string if possible
coleenp@4037:   if (!string_or_null.is_null()) {
coleenp@3875:     string = string_or_null;
coleenp@3875:   } else {
coleenp@4037:     string = java_lang_String::create_from_unicode(name, len, CHECK_NULL);
coleenp@3875:   }
coleenp@3875: 
coleenp@3875:   // Grab the StringTable_lock before getting the_table() because it could
coleenp@3875:   // change at safepoint.
coleenp@3875:   MutexLocker ml(StringTable_lock, THREAD);
duke@435: 
duke@435:   // Otherwise, add to symbol to table
coleenp@3875:   return the_table()->basic_add(index, string, name, len,
duke@435:                                 hashValue, CHECK_NULL);
duke@435: }
duke@435: 
coleenp@2497: oop StringTable::intern(Symbol* symbol, TRAPS) {
duke@435:   if (symbol == NULL) return NULL;
duke@435:   ResourceMark rm(THREAD);
duke@435:   int length;
duke@435:   jchar* chars = symbol->as_unicode(length);
duke@435:   Handle string;
duke@435:   oop result = intern(string, chars, length, CHECK_NULL);
duke@435:   return result;
duke@435: }
duke@435: 
duke@435: 
duke@435: oop StringTable::intern(oop string, TRAPS)
duke@435: {
duke@435:   if (string == NULL) return NULL;
duke@435:   ResourceMark rm(THREAD);
duke@435:   int length;
duke@435:   Handle h_string (THREAD, string);
hseigel@4987:   jchar* chars = java_lang_String::as_unicode_string(string, length, CHECK_NULL);
duke@435:   oop result = intern(h_string, chars, length, CHECK_NULL);
duke@435:   return result;
duke@435: }
duke@435: 
duke@435: 
duke@435: oop StringTable::intern(const char* utf8_string, TRAPS) {
duke@435:   if (utf8_string == NULL) return NULL;
duke@435:   ResourceMark rm(THREAD);
duke@435:   int length = UTF8::unicode_length(utf8_string);
duke@435:   jchar* chars = NEW_RESOURCE_ARRAY(jchar, length);
duke@435:   UTF8::convert_to_unicode(utf8_string, chars, length);
duke@435:   Handle string;
duke@435:   oop result = intern(string, chars, length, CHECK_NULL);
duke@435:   return result;
duke@435: }
duke@435: 
stefank@5196: void StringTable::unlink_or_oops_do(BoolObjectClosure* is_alive, OopClosure* f) {
coleenp@2497:   // Readers of the table are unlocked, so we should only be removing
coleenp@2497:   // entries at a safepoint.
coleenp@2497:   assert(SafepointSynchronize::is_at_safepoint(), "must be at safepoint");
coleenp@2497:   for (int i = 0; i < the_table()->table_size(); ++i) {
zgu@3900:     HashtableEntry<oop, mtSymbol>** p = the_table()->bucket_addr(i);
zgu@3900:     HashtableEntry<oop, mtSymbol>* entry = the_table()->bucket(i);
coleenp@3875:     while (entry != NULL) {
stefank@5195:       assert(!entry->is_shared(), "CDS not used for the StringTable");
stefank@5195: 
stefank@5195:       if (is_alive->do_object_b(entry->literal())) {
stefank@5196:         if (f != NULL) {
stefank@5196:           f->do_oop((oop*)entry->literal_addr());
stefank@5196:         }
coleenp@2497:         p = entry->next_addr();
coleenp@2497:       } else {
coleenp@2497:         *p = entry->next();
coleenp@2497:         the_table()->free_entry(entry);
coleenp@2497:       }
stefank@5195:       entry = *p;
coleenp@2497:     }
coleenp@2497:   }
coleenp@2497: }
coleenp@2497: 
coleenp@2497: void StringTable::oops_do(OopClosure* f) {
coleenp@2497:   for (int i = 0; i < the_table()->table_size(); ++i) {
zgu@3900:     HashtableEntry<oop, mtSymbol>* entry = the_table()->bucket(i);
coleenp@2497:     while (entry != NULL) {
stefank@5195:       assert(!entry->is_shared(), "CDS not used for the StringTable");
stefank@5195: 
coleenp@2497:       f->do_oop((oop*)entry->literal_addr());
coleenp@2497: 
stefank@5195:       entry = entry->next();
coleenp@2497:     }
coleenp@2497:   }
coleenp@2497: }
coleenp@2497: 
duke@435: void StringTable::verify() {
duke@435:   for (int i = 0; i < the_table()->table_size(); ++i) {
zgu@3900:     HashtableEntry<oop, mtSymbol>* p = the_table()->bucket(i);
duke@435:     for ( ; p != NULL; p = p->next()) {
duke@435:       oop s = p->literal();
duke@435:       guarantee(s != NULL, "interned string is NULL");
never@2700:       unsigned int h = java_lang_String::hash_string(s);
duke@435:       guarantee(p->hash() == h, "broken hash in string table entry");
duke@435:       guarantee(the_table()->hash_to_index(h) == i,
duke@435:                 "wrong index in string table");
duke@435:     }
duke@435:   }
duke@435: }
coleenp@3865: 
coleenp@3865: void StringTable::dump(outputStream* st) {
iklam@5144:   the_table()->dump_table(st, "StringTable");
coleenp@3865: }
coleenp@3865: 
coleenp@3865: 
coleenp@3865: // Create a new table and using alternate hash code, populate the new table
coleenp@3865: // with the existing strings.   Set flag to use the alternate hash code afterwards.
coleenp@3865: void StringTable::rehash_table() {
coleenp@3865:   assert(SafepointSynchronize::is_at_safepoint(), "must be at safepoint");
coleenp@3875:   // This should never happen with -Xshare:dump but it might in testing mode.
coleenp@3875:   if (DumpSharedSpaces) return;
coleenp@3865:   StringTable* new_table = new StringTable();
coleenp@3865: 
coleenp@3865:   // Rehash the table
coleenp@3865:   the_table()->move_to(new_table);
coleenp@3865: 
coleenp@3865:   // Delete the table and buckets (entries are reused in new table).
coleenp@3865:   delete _the_table;
coleenp@3865:   // Don't check if we need rehashing until the table gets unbalanced again.
coleenp@3865:   // Then rehash with a new global seed.
coleenp@3865:   _needs_rehashing = false;
coleenp@3865:   _the_table = new_table;
coleenp@3865: }