1 /*
   2  * Copyright (c) 1997, 2015, 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_OOPS_MARKOOP_HPP
  26 #define SHARE_VM_OOPS_MARKOOP_HPP
  27 
  28 #include "oops/oop.hpp"
  29 
  30 // The markOop describes the header of an object.
  31 //
  32 // Note that the mark is not a real oop but just a word.
  33 // It is placed in the oop hierarchy for historical reasons.
  34 //
  35 // Bit-format of an object header (most significant first, big endian layout below):
  36 //
  37 //  32 bits:
  38 //  --------
  39 //             hash:24 ------------>| age:5    biased_lock:1 lock:2 (normal object)
  40 //             JavaThread*:22 epoch:2 age:5    biased_lock:1 lock:2 (biased object)
  41 //             size:32 ------------------------------------------>| (CMS free block)
  42 //             PromotedObject*:29 ---------->| promo_bits:3 ----->| (CMS promoted object)
  43 //
  44 //  64 bits:
  45 //  --------
  46 //  unused:24 hash:31 -->| unused:1   age:5    biased_lock:1 lock:2 (normal object)
  47 //  JavaThread*:53 epoch:2 unused:1   age:5    biased_lock:1 lock:2 (biased object)
  48 //  PromotedObject*:61 --------------------->| promo_bits:3 ----->| (CMS promoted object)
  49 //  size:64 ----------------------------------------------------->| (CMS free block)
  50 //
  51 //  unused:24 hash:31 -->| cms_free:1 age:5    biased_lock:1 lock:2 (COOPs && normal object)
  52 //  JavaThread*:53 epoch:2 cms_free:1 age:5    biased_lock:1 lock:2 (COOPs && biased object)
  53 //  narrowOop:32 unused:23 cms_free:1 unused:5 promo_bits:3 ----->| (COOPs && CMS promoted object)
  54 //  unused:20 size:35 -->| cms_free:1 unused:8 ------------------>| (COOPs && CMS free block)
  55 //
  56 //  - hash contains the identity hash value: largest value is
  57 //    31 bits, see os::random().  Also, 64-bit vm's require
  58 //    a hash value no bigger than 32 bits because they will not
  59 //    properly generate a mask larger than that: see library_call.cpp
  60 //    and c1_CodePatterns_sparc.cpp.
  61 //
  62 //  - the biased lock pattern is used to bias a lock toward a given
  63 //    thread. When this pattern is set in the low three bits, the lock
  64 //    is either biased toward a given thread or "anonymously" biased,
  65 //    indicating that it is possible for it to be biased. When the
  66 //    lock is biased toward a given thread, locking and unlocking can
  67 //    be performed by that thread without using atomic operations.
  68 //    When a lock's bias is revoked, it reverts back to the normal
  69 //    locking scheme described below.
  70 //
  71 //    Note that we are overloading the meaning of the "unlocked" state
  72 //    of the header. Because we steal a bit from the age we can
  73 //    guarantee that the bias pattern will never be seen for a truly
  74 //    unlocked object.
  75 //
  76 //    Note also that the biased state contains the age bits normally
  77 //    contained in the object header. Large increases in scavenge
  78 //    times were seen when these bits were absent and an arbitrary age
  79 //    assigned to all biased objects, because they tended to consume a
  80 //    significant fraction of the eden semispaces and were not
  81 //    promoted promptly, causing an increase in the amount of copying
  82 //    performed. The runtime system aligns all JavaThread* pointers to
  83 //    a very large value (currently 1024 bytes (32bVM) or 2048 bytes (64bVM))
  84 //    to make room for the age bits & the epoch bits (used in support of
  85 //    biased locking), and for the CMS "freeness" bit in the 64bVM (+COOPs).
  86 //
  87 //    [JavaThread* | epoch | age | 1 | 01]       lock is biased toward given thread
  88 //    [0           | epoch | age | 1 | 01]       lock is anonymously biased
  89 //
  90 //  - the two lock bits are used to describe three states: locked/unlocked and monitor.
  91 //
  92 //    [ptr             | 00]  locked             ptr points to real header on stack
  93 //    [header      | 0 | 01]  unlocked           regular object header
  94 //    [ptr             | 10]  monitor            inflated lock (header is wapped out)
  95 //    [ptr             | 11]  marked             used by markSweep to mark an object
  96 //                                               not valid at any other time
  97 //
  98 //    We assume that stack/thread pointers have the lowest two bits cleared.
  99 
 100 class BasicLock;
 101 class ObjectMonitor;
 102 class JavaThread;
 103 
 104 class markOopDesc: public oopDesc {
 105  private:
 106   // Conversion
 107   uintptr_t value() const { return (uintptr_t) this; }
 108 
 109  public:
 110   // Constants
 111   enum { age_bits                 = 5,
 112          lock_bits                = 2,
 113          biased_lock_bits         = 1,
 114          max_hash_bits            = BitsPerWord - age_bits - lock_bits - biased_lock_bits,
 115          hash_bits                = max_hash_bits > 31 ? 31 : max_hash_bits,
 116          cms_bits                 = LP64_ONLY(1) NOT_LP64(0),
 117          epoch_bits               = 2
 118   };
 119 
 120   // The biased locking code currently requires that the age bits be
 121   // contiguous to the lock bits.
 122   enum { lock_shift               = 0,
 123          biased_lock_shift        = lock_bits,
 124          age_shift                = lock_bits + biased_lock_bits,
 125          cms_shift                = age_shift + age_bits,
 126          hash_shift               = cms_shift + cms_bits,
 127          epoch_shift              = hash_shift
 128   };
 129 
 130   enum { lock_mask                = right_n_bits(lock_bits),
 131          lock_mask_in_place       = lock_mask << lock_shift,
 132          biased_lock_mask         = right_n_bits(lock_bits + biased_lock_bits),
 133          biased_lock_mask_in_place= biased_lock_mask << lock_shift,
 134          biased_lock_bit_in_place = 1 << biased_lock_shift,
 135          age_mask                 = right_n_bits(age_bits),
 136          age_mask_in_place        = age_mask << age_shift,
 137          epoch_mask               = right_n_bits(epoch_bits),
 138          epoch_mask_in_place      = epoch_mask << epoch_shift,
 139          cms_mask                 = right_n_bits(cms_bits),
 140          cms_mask_in_place        = cms_mask << cms_shift
 141 #ifndef _WIN64
 142          ,hash_mask               = right_n_bits(hash_bits),
 143          hash_mask_in_place       = (address_word)hash_mask << hash_shift
 144 #endif
 145   };
 146 
 147   // Alignment of JavaThread pointers encoded in object header required by biased locking
 148   enum { biased_lock_alignment    = 1 << (epoch_shift + epoch_bits)
 149   };
 150 
 151 #ifdef _WIN64
 152     // These values are too big for Win64
 153     const static uintptr_t hash_mask = right_n_bits(hash_bits);
 154     const static uintptr_t hash_mask_in_place  =
 155                             (address_word)hash_mask << hash_shift;
 156 #endif
 157 
 158   enum { locked_value             = 0,
 159          unlocked_value           = 1,
 160          monitor_value            = 2,
 161          marked_value             = 3,
 162          biased_lock_pattern      = 5
 163   };
 164 
 165   enum { no_hash                  = 0 };  // no hash value assigned
 166 
 167   enum { no_hash_in_place         = (address_word)no_hash << hash_shift,
 168          no_lock_in_place         = unlocked_value
 169   };
 170 
 171   enum { max_age                  = age_mask };
 172 
 173   enum { max_bias_epoch           = epoch_mask };
 174 
 175   // Biased Locking accessors.
 176   // These must be checked by all code which calls into the
 177   // ObjectSynchronizer and other code. The biasing is not understood
 178   // by the lower-level CAS-based locking code, although the runtime
 179   // fixes up biased locks to be compatible with it when a bias is
 180   // revoked.
 181   bool has_bias_pattern() const {
 182     return (mask_bits(value(), biased_lock_mask_in_place) == biased_lock_pattern);
 183   }
 184   JavaThread* biased_locker() const {
 185     assert(has_bias_pattern(), "should not call this otherwise");
 186     return (JavaThread*) ((intptr_t) (mask_bits(value(), ~(biased_lock_mask_in_place | age_mask_in_place | epoch_mask_in_place))));
 187   }
 188   // Indicates that the mark has the bias bit set but that it has not
 189   // yet been biased toward a particular thread
 190   bool is_biased_anonymously() const {
 191     return (has_bias_pattern() && (biased_locker() == NULL));
 192   }
 193   // Indicates epoch in which this bias was acquired. If the epoch
 194   // changes due to too many bias revocations occurring, the biases
 195   // from the previous epochs are all considered invalid.
 196   int bias_epoch() const {
 197     assert(has_bias_pattern(), "should not call this otherwise");
 198     return (mask_bits(value(), epoch_mask_in_place) >> epoch_shift);
 199   }
 200   markOop set_bias_epoch(int epoch) {
 201     assert(has_bias_pattern(), "should not call this otherwise");
 202     assert((epoch & (~epoch_mask)) == 0, "epoch overflow");
 203     return markOop(mask_bits(value(), ~epoch_mask_in_place) | (epoch << epoch_shift));
 204   }
 205   markOop incr_bias_epoch() {
 206     return set_bias_epoch((1 + bias_epoch()) & epoch_mask);
 207   }
 208   // Prototype mark for initialization
 209   static markOop biased_locking_prototype() {
 210     return markOop( biased_lock_pattern );
 211   }
 212 
 213   // lock accessors (note that these assume lock_shift == 0)
 214   bool is_locked()   const {
 215     return (mask_bits(value(), lock_mask_in_place) != unlocked_value);
 216   }
 217   bool is_unlocked() const {
 218     return (mask_bits(value(), biased_lock_mask_in_place) == unlocked_value);
 219   }
 220   bool is_marked()   const {
 221     return (mask_bits(value(), lock_mask_in_place) == marked_value);
 222   }
 223   bool is_neutral()  const { return (mask_bits(value(), biased_lock_mask_in_place) == unlocked_value); }
 224 
 225   // Special temporary state of the markOop while being inflated.
 226   // Code that looks at mark outside a lock need to take this into account.
 227   bool is_being_inflated() const { return (value() == 0); }
 228 
 229   // Distinguished markword value - used when inflating over
 230   // an existing stacklock.  0 indicates the markword is "BUSY".
 231   // Lockword mutators that use a LD...CAS idiom should always
 232   // check for and avoid overwriting a 0 value installed by some
 233   // other thread.  (They should spin or block instead.  The 0 value
 234   // is transient and *should* be short-lived).
 235   static markOop INFLATING() { return (markOop) 0; }    // inflate-in-progress
 236 
 237   // Should this header be preserved during GC?
 238   inline bool must_be_preserved(oop obj_containing_mark) const;
 239   inline bool must_be_preserved_with_bias(oop obj_containing_mark) const;
 240 
 241   // Should this header (including its age bits) be preserved in the
 242   // case of a promotion failure during scavenge?
 243   // Note that we special case this situation. We want to avoid
 244   // calling BiasedLocking::preserve_marks()/restore_marks() (which
 245   // decrease the number of mark words that need to be preserved
 246   // during GC) during each scavenge. During scavenges in which there
 247   // is no promotion failure, we actually don't need to call the above
 248   // routines at all, since we don't mutate and re-initialize the
 249   // marks of promoted objects using init_mark(). However, during
 250   // scavenges which result in promotion failure, we do re-initialize
 251   // the mark words of objects, meaning that we should have called
 252   // these mark word preservation routines. Currently there's no good
 253   // place in which to call them in any of the scavengers (although
 254   // guarded by appropriate locks we could make one), but the
 255   // observation is that promotion failures are quite rare and
 256   // reducing the number of mark words preserved during them isn't a
 257   // high priority.
 258   inline bool must_be_preserved_for_promotion_failure(oop obj_containing_mark) const;
 259   inline bool must_be_preserved_with_bias_for_promotion_failure(oop obj_containing_mark) const;
 260 
 261   // Should this header be preserved during a scavenge where CMS is
 262   // the old generation?
 263   // (This is basically the same body as must_be_preserved_for_promotion_failure(),
 264   // but takes the Klass* as argument instead)
 265   inline bool must_be_preserved_for_cms_scavenge(Klass* klass_of_obj_containing_mark) const;
 266   inline bool must_be_preserved_with_bias_for_cms_scavenge(Klass* klass_of_obj_containing_mark) const;
 267 
 268   // WARNING: The following routines are used EXCLUSIVELY by
 269   // synchronization functions. They are not really gc safe.
 270   // They must get updated if markOop layout get changed.
 271   markOop set_unlocked() const {
 272     return markOop(value() | unlocked_value);
 273   }
 274   bool has_locker() const {
 275     return ((value() & lock_mask_in_place) == locked_value);
 276   }
 277   BasicLock* locker() const {
 278     assert(has_locker(), "check");
 279     return (BasicLock*) value();
 280   }
 281   bool has_monitor() const {
 282     return ((value() & monitor_value) != 0);
 283   }
 284   ObjectMonitor* monitor() const {
 285     assert(has_monitor(), "check");
 286     // Use xor instead of &~ to provide one extra tag-bit check.
 287     return (ObjectMonitor*) (value() ^ monitor_value);
 288   }
 289   bool has_displaced_mark_helper() const {
 290     return ((value() & unlocked_value) == 0);
 291   }
 292   markOop displaced_mark_helper() const {
 293     assert(has_displaced_mark_helper(), "check");
 294     intptr_t ptr = (value() & ~monitor_value);
 295     return *(markOop*)ptr;
 296   }
 297   void set_displaced_mark_helper(markOop m) const {
 298     assert(has_displaced_mark_helper(), "check");
 299     intptr_t ptr = (value() & ~monitor_value);
 300     *(markOop*)ptr = m;
 301   }
 302   markOop copy_set_hash(intptr_t hash) const {
 303     intptr_t tmp = value() & (~hash_mask_in_place);
 304     tmp |= ((hash & hash_mask) << hash_shift);
 305     return (markOop)tmp;
 306   }
 307   // it is only used to be stored into BasicLock as the
 308   // indicator that the lock is using heavyweight monitor
 309   static markOop unused_mark() {
 310     return (markOop) marked_value;
 311   }
 312   // the following two functions create the markOop to be
 313   // stored into object header, it encodes monitor info
 314   static markOop encode(BasicLock* lock) {
 315     return (markOop) lock;
 316   }
 317   static markOop encode(ObjectMonitor* monitor) {
 318     intptr_t tmp = (intptr_t) monitor;
 319     return (markOop) (tmp | monitor_value);
 320   }
 321   static markOop encode(JavaThread* thread, uint age, int bias_epoch) {
 322     intptr_t tmp = (intptr_t) thread;
 323     assert(UseBiasedLocking && ((tmp & (epoch_mask_in_place | age_mask_in_place | biased_lock_mask_in_place)) == 0), "misaligned JavaThread pointer");
 324     assert(age <= max_age, "age too large");
 325     assert(bias_epoch <= max_bias_epoch, "bias epoch too large");
 326     return (markOop) (tmp | (bias_epoch << epoch_shift) | (age << age_shift) | biased_lock_pattern);
 327   }
 328 
 329   // used to encode pointers during GC
 330   markOop clear_lock_bits() { return markOop(value() & ~lock_mask_in_place); }
 331 
 332   // age operations
 333   markOop set_marked()   { return markOop((value() & ~lock_mask_in_place) | marked_value); }
 334   markOop set_unmarked() { return markOop((value() & ~lock_mask_in_place) | unlocked_value); }
 335 
 336   uint    age()               const { return mask_bits(value() >> age_shift, age_mask); }
 337   markOop set_age(uint v) const {
 338     assert((v & ~age_mask) == 0, "shouldn't overflow age field");
 339     return markOop((value() & ~age_mask_in_place) | (((uintptr_t)v & age_mask) << age_shift));
 340   }
 341   markOop incr_age()          const { return age() == max_age ? markOop(this) : set_age(age() + 1); }
 342 
 343   // hash operations
 344   intptr_t hash() const {
 345     return mask_bits(value() >> hash_shift, hash_mask);
 346   }
 347 
 348   bool has_no_hash() const {
 349     return hash() == no_hash;
 350   }
 351 
 352   // Prototype mark for initialization
 353   static markOop prototype() {
 354     return markOop( no_hash_in_place | no_lock_in_place );
 355   }
 356 
 357   // Helper function for restoration of unmarked mark oops during GC
 358   static inline markOop prototype_for_object(oop obj);
 359 
 360   // Debugging
 361   void print_on(outputStream* st) const;
 362 
 363   // Prepare address of oop for placement into mark
 364   inline static markOop encode_pointer_as_mark(void* p) { return markOop(p)->set_marked(); }
 365 
 366   // Recover address of oop from encoded form used in mark
 367   inline void* decode_pointer() { if (UseBiasedLocking && has_bias_pattern()) return NULL; return clear_lock_bits(); }
 368 
 369   // These markOops indicate cms free chunk blocks and not objects.
 370   // In 64 bit, the markOop is set to distinguish them from oops.
 371   // These are defined in 32 bit mode for vmStructs.
 372   const static uintptr_t cms_free_chunk_pattern  = 0x1;
 373 
 374   // Constants for the size field.
 375   enum { size_shift                = cms_shift + cms_bits,
 376          size_bits                 = 35    // need for compressed oops 32G
 377        };
 378   // These values are too big for Win64
 379   const static uintptr_t size_mask = LP64_ONLY(right_n_bits(size_bits))
 380                                      NOT_LP64(0);
 381   const static uintptr_t size_mask_in_place =
 382                                      (address_word)size_mask << size_shift;
 383 
 384 #ifdef _LP64
 385   static markOop cms_free_prototype() {
 386     return markOop(((intptr_t)prototype() & ~cms_mask_in_place) |
 387                    ((cms_free_chunk_pattern & cms_mask) << cms_shift));
 388   }
 389   uintptr_t cms_encoding() const {
 390     return mask_bits(value() >> cms_shift, cms_mask);
 391   }
 392   bool is_cms_free_chunk() const {
 393     return is_neutral() &&
 394            (cms_encoding() & cms_free_chunk_pattern) == cms_free_chunk_pattern;
 395   }
 396 
 397   size_t get_size() const       { return (size_t)(value() >> size_shift); }
 398   static markOop set_size_and_free(size_t size) {
 399     assert((size & ~size_mask) == 0, "shouldn't overflow size field");
 400     return markOop(((intptr_t)cms_free_prototype() & ~size_mask_in_place) |
 401                    (((intptr_t)size & size_mask) << size_shift));
 402   }
 403 #endif // _LP64
 404 };
 405 
 406 #endif // SHARE_VM_OOPS_MARKOOP_HPP