1 /*
   2  * Copyright (c) 2014, 2019, 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_GC_SHARED_STRINGDEDUP_STRINGDEDUP_HPP
  26 #define SHARE_GC_SHARED_STRINGDEDUP_STRINGDEDUP_HPP
  27 
  28 //
  29 // String Deduplication
  30 //
  31 // String deduplication aims to reduce the heap live-set by deduplicating identical
  32 // instances of String so that they share the same backing character array.
  33 //
  34 // The deduplication process is divided in two main parts, 1) finding the objects to
  35 // deduplicate, and 2) deduplicating those objects. The first part is done as part of
  36 // a normal GC cycle when objects are marked or evacuated. At this time a check is
  37 // applied on each object to check if it is a candidate for deduplication. If so, the
  38 // object is placed on the deduplication queue for later processing. The second part,
  39 // processing the objects on the deduplication queue, is a concurrent phase which
  40 // starts right after the stop-the-wold marking/evacuation phase. This phase is
  41 // executed by the deduplication thread, which pulls deduplication candidates of the
  42 // deduplication queue and tries to deduplicate them.
  43 //
  44 // A deduplication hashtable is used to keep track of all unique character arrays
  45 // used by String objects. When deduplicating, a lookup is made in this table to see
  46 // if there is already an identical character array somewhere on the heap. If so, the
  47 // String object is adjusted to point to that character array, releasing the reference
  48 // to the original array allowing it to eventually be garbage collected. If the lookup
  49 // fails the character array is instead inserted into the hashtable so that this array
  50 // can be shared at some point in the future.
  51 //
  52 // Candidate selection criteria is GC specific.
  53 //
  54 // Interned strings are a bit special. They are explicitly deduplicated just before
  55 // being inserted into the StringTable (to avoid counteracting C2 optimizations done
  56 // on string literals), then they also become deduplication candidates if they reach
  57 // the deduplication age threshold or are evacuated to an old heap region. The second
  58 // attempt to deduplicate such strings will be in vain, but we have no fast way of
  59 // filtering them out. This has not shown to be a problem, as the number of interned
  60 // strings is usually dwarfed by the number of normal (non-interned) strings.
  61 //
  62 // For additional information on string deduplication, please see JEP 192,
  63 // http://openjdk.java.net/jeps/192
  64 //
  65 
  66 #include "gc/shared/stringdedup/stringDedupQueue.hpp"
  67 #include "gc/shared/stringdedup/stringDedupStat.hpp"
  68 #include "gc/shared/stringdedup/stringDedupTable.hpp"
  69 #include "memory/allocation.hpp"
  70 #include "runtime/thread.hpp"
  71 
  72 class ThreadClosure;
  73 
  74 //
  75 // Main interface for interacting with string deduplication.
  76 //
  77 class StringDedup : public AllStatic {
  78 private:
  79   // Single state for checking if string deduplication is enabled.
  80   static bool _enabled;
  81 
  82 public:
  83   // Returns true if string deduplication is enabled.
  84   static bool is_enabled() {
  85     return _enabled;
  86   }
  87 
  88   // Stop the deduplication thread.
  89   static void stop();
  90 
  91   // Immediately deduplicates the given String object, bypassing the
  92   // the deduplication queue.
  93   static void deduplicate(oop java_string);
  94 
  95   static void parallel_unlink(StringDedupUnlinkOrOopsDoClosure* unlink, uint worker_id);
  96 
  97   static void threads_do(ThreadClosure* tc);
  98   static void print_worker_threads_on(outputStream* st);
  99   static void verify();
 100 
 101   // GC support
 102   static void gc_prologue(bool resize_and_rehash_table);
 103   static void gc_epilogue();
 104 
 105 protected:
 106   // Initialize string deduplication.
 107   // Q: String Dedup Queue implementation
 108   // S: String Dedup Stat implementation
 109   template <typename Q, typename S>
 110   static void initialize_impl();
 111 };
 112 
 113 //
 114 // This closure encapsulates the closures needed when scanning
 115 // the deduplication queue and table during the unlink_or_oops_do() operation.
 116 //
 117 class StringDedupUnlinkOrOopsDoClosure : public StackObj {
 118   AlwaysTrueClosure   _always_true;
 119   DoNothingClosure    _do_nothing;
 120   BoolObjectClosure*  _is_alive;
 121   OopClosure*         _keep_alive;
 122 
 123 public:
 124   StringDedupUnlinkOrOopsDoClosure(BoolObjectClosure* is_alive,
 125                                    OopClosure* keep_alive);
 126 
 127   bool is_alive(oop o) { return _is_alive->do_object_b(o); }
 128 
 129   void keep_alive(oop* p) { _keep_alive->do_oop(p); }
 130 };
 131 
 132 #endif // SHARE_GC_SHARED_STRINGDEDUP_STRINGDEDUP_HPP