1 /* 2 * Copyright 2001-2006 Sun Microsystems, Inc. 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 Sun Microsystems, Inc., 4150 Network Circle, Santa Clara, 20 * CA 95054 USA or visit www.sun.com if you need additional information or 21 * have any questions. 22 * 23 */ 24 25 # include "incls/_precompiled.incl" 26 # include "incls/_compactibleFreeListSpace.cpp.incl" 27 28 ///////////////////////////////////////////////////////////////////////// 29 //// CompactibleFreeListSpace 30 ///////////////////////////////////////////////////////////////////////// 31 32 // highest ranked free list lock rank 33 int CompactibleFreeListSpace::_lockRank = Mutex::leaf + 3; 34 35 // Constructor 36 CompactibleFreeListSpace::CompactibleFreeListSpace(BlockOffsetSharedArray* bs, 37 MemRegion mr, bool use_adaptive_freelists, 38 FreeBlockDictionary::DictionaryChoice dictionaryChoice) : 39 _dictionaryChoice(dictionaryChoice), 40 _adaptive_freelists(use_adaptive_freelists), 41 _bt(bs, mr), 42 // free list locks are in the range of values taken by _lockRank 43 // This range currently is [_leaf+2, _leaf+3] 44 // Note: this requires that CFLspace c'tors 45 // are called serially in the order in which the locks are 46 // are acquired in the program text. This is true today. 47 _freelistLock(_lockRank--, "CompactibleFreeListSpace._lock", true), 48 _parDictionaryAllocLock(Mutex::leaf - 1, // == rank(ExpandHeap_lock) - 1 49 "CompactibleFreeListSpace._dict_par_lock", true), 50 _rescan_task_size(CardTableModRefBS::card_size_in_words * BitsPerWord * 51 CMSRescanMultiple), 52 _marking_task_size(CardTableModRefBS::card_size_in_words * BitsPerWord * 53 CMSConcMarkMultiple), 54 _collector(NULL) 55 { 56 _bt.set_space(this); 57 initialize(mr, true); 58 // We have all of "mr", all of which we place in the dictionary 59 // as one big chunk. We'll need to decide here which of several 60 // possible alternative dictionary implementations to use. For 61 // now the choice is easy, since we have only one working 62 // implementation, namely, the simple binary tree (splaying 63 // temporarily disabled). 64 switch (dictionaryChoice) { 65 case FreeBlockDictionary::dictionaryBinaryTree: 66 _dictionary = new BinaryTreeDictionary(mr); 67 break; 68 case FreeBlockDictionary::dictionarySplayTree: 69 case FreeBlockDictionary::dictionarySkipList: 70 default: 71 warning("dictionaryChoice: selected option not understood; using" 72 " default BinaryTreeDictionary implementation instead."); 73 _dictionary = new BinaryTreeDictionary(mr); 74 break; 75 } 76 splitBirth(mr.word_size()); 77 assert(_dictionary != NULL, "CMS dictionary initialization"); 78 // The indexed free lists are initially all empty and are lazily 79 // filled in on demand. Initialize the array elements to NULL. 80 initializeIndexedFreeListArray(); 81 82 // Not using adaptive free lists assumes that allocation is first 83 // from the linAB's. Also a cms perm gen which can be compacted 84 // has to have the klass's klassKlass allocated at a lower 85 // address in the heap than the klass so that the klassKlass is 86 // moved to its new location before the klass is moved. 87 // Set the _refillSize for the linear allocation blocks 88 if (!use_adaptive_freelists) { 89 FreeChunk* fc = _dictionary->getChunk(mr.word_size()); 90 // The small linAB initially has all the space and will allocate 91 // a chunk of any size. 92 HeapWord* addr = (HeapWord*) fc; 93 _smallLinearAllocBlock.set(addr, fc->size() , 94 1024*SmallForLinearAlloc, fc->size()); 95 // Note that _unallocated_block is not updated here. 96 // Allocations from the linear allocation block should 97 // update it. 98 } else { 99 _smallLinearAllocBlock.set(0, 0, 1024*SmallForLinearAlloc, 100 SmallForLinearAlloc); 101 } 102 // CMSIndexedFreeListReplenish should be at least 1 103 CMSIndexedFreeListReplenish = MAX2((uintx)1, CMSIndexedFreeListReplenish); 104 _promoInfo.setSpace(this); 105 if (UseCMSBestFit) { 106 _fitStrategy = FreeBlockBestFitFirst; 107 } else { 108 _fitStrategy = FreeBlockStrategyNone; 109 } 110 checkFreeListConsistency(); 111 112 // Initialize locks for parallel case. 113 if (ParallelGCThreads > 0) { 114 for (size_t i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) { 115 _indexedFreeListParLocks[i] = new Mutex(Mutex::leaf - 1, // == ExpandHeap_lock - 1 116 "a freelist par lock", 117 true); 118 if (_indexedFreeListParLocks[i] == NULL) 119 vm_exit_during_initialization("Could not allocate a par lock"); 120 DEBUG_ONLY( 121 _indexedFreeList[i].set_protecting_lock(_indexedFreeListParLocks[i]); 122 ) 123 } 124 _dictionary->set_par_lock(&_parDictionaryAllocLock); 125 } 126 } 127 128 // Like CompactibleSpace forward() but always calls cross_threshold() to 129 // update the block offset table. Removed initialize_threshold call because 130 // CFLS does not use a block offset array for contiguous spaces. 131 HeapWord* CompactibleFreeListSpace::forward(oop q, size_t size, 132 CompactPoint* cp, HeapWord* compact_top) { 133 // q is alive 134 // First check if we should switch compaction space 135 assert(this == cp->space, "'this' should be current compaction space."); 136 size_t compaction_max_size = pointer_delta(end(), compact_top); 137 assert(adjustObjectSize(size) == cp->space->adjust_object_size_v(size), 138 "virtual adjustObjectSize_v() method is not correct"); 139 size_t adjusted_size = adjustObjectSize(size); 140 assert(compaction_max_size >= MinChunkSize || compaction_max_size == 0, 141 "no small fragments allowed"); 142 assert(minimum_free_block_size() == MinChunkSize, 143 "for de-virtualized reference below"); 144 // Can't leave a nonzero size, residual fragment smaller than MinChunkSize 145 if (adjusted_size + MinChunkSize > compaction_max_size && 146 adjusted_size != compaction_max_size) { 147 do { 148 // switch to next compaction space 149 cp->space->set_compaction_top(compact_top); 150 cp->space = cp->space->next_compaction_space(); 151 if (cp->space == NULL) { 152 cp->gen = GenCollectedHeap::heap()->prev_gen(cp->gen); 153 assert(cp->gen != NULL, "compaction must succeed"); 154 cp->space = cp->gen->first_compaction_space(); 155 assert(cp->space != NULL, "generation must have a first compaction space"); 156 } 157 compact_top = cp->space->bottom(); 158 cp->space->set_compaction_top(compact_top); 159 // The correct adjusted_size may not be the same as that for this method 160 // (i.e., cp->space may no longer be "this" so adjust the size again. 161 // Use the virtual method which is not used above to save the virtual 162 // dispatch. 163 adjusted_size = cp->space->adjust_object_size_v(size); 164 compaction_max_size = pointer_delta(cp->space->end(), compact_top); 165 assert(cp->space->minimum_free_block_size() == 0, "just checking"); 166 } while (adjusted_size > compaction_max_size); 167 } 168 169 // store the forwarding pointer into the mark word 170 if ((HeapWord*)q != compact_top) { 171 q->forward_to(oop(compact_top)); 172 assert(q->is_gc_marked(), "encoding the pointer should preserve the mark"); 173 } else { 174 // if the object isn't moving we can just set the mark to the default 175 // mark and handle it specially later on. 176 q->init_mark(); 177 assert(q->forwardee() == NULL, "should be forwarded to NULL"); 178 } 179 180 debug_only(MarkSweep::register_live_oop(q, adjusted_size)); 181 compact_top += adjusted_size; 182 183 // we need to update the offset table so that the beginnings of objects can be 184 // found during scavenge. Note that we are updating the offset table based on 185 // where the object will be once the compaction phase finishes. 186 187 // Always call cross_threshold(). A contiguous space can only call it when 188 // the compaction_top exceeds the current threshold but not for an 189 // non-contiguous space. 190 cp->threshold = 191 cp->space->cross_threshold(compact_top - adjusted_size, compact_top); 192 return compact_top; 193 } 194 195 // A modified copy of OffsetTableContigSpace::cross_threshold() with _offsets -> _bt 196 // and use of single_block instead of alloc_block. The name here is not really 197 // appropriate - maybe a more general name could be invented for both the 198 // contiguous and noncontiguous spaces. 199 200 HeapWord* CompactibleFreeListSpace::cross_threshold(HeapWord* start, HeapWord* the_end) { 201 _bt.single_block(start, the_end); 202 return end(); 203 } 204 205 // Initialize them to NULL. 206 void CompactibleFreeListSpace::initializeIndexedFreeListArray() { 207 for (size_t i = 0; i < IndexSetSize; i++) { 208 // Note that on platforms where objects are double word aligned, 209 // the odd array elements are not used. It is convenient, however, 210 // to map directly from the object size to the array element. 211 _indexedFreeList[i].reset(IndexSetSize); 212 _indexedFreeList[i].set_size(i); 213 assert(_indexedFreeList[i].count() == 0, "reset check failed"); 214 assert(_indexedFreeList[i].head() == NULL, "reset check failed"); 215 assert(_indexedFreeList[i].tail() == NULL, "reset check failed"); 216 assert(_indexedFreeList[i].hint() == IndexSetSize, "reset check failed"); 217 } 218 } 219 220 void CompactibleFreeListSpace::resetIndexedFreeListArray() { 221 for (int i = 1; i < IndexSetSize; i++) { 222 assert(_indexedFreeList[i].size() == (size_t) i, 223 "Indexed free list sizes are incorrect"); 224 _indexedFreeList[i].reset(IndexSetSize); 225 assert(_indexedFreeList[i].count() == 0, "reset check failed"); 226 assert(_indexedFreeList[i].head() == NULL, "reset check failed"); 227 assert(_indexedFreeList[i].tail() == NULL, "reset check failed"); 228 assert(_indexedFreeList[i].hint() == IndexSetSize, "reset check failed"); 229 } 230 } 231 232 void CompactibleFreeListSpace::reset(MemRegion mr) { 233 resetIndexedFreeListArray(); 234 dictionary()->reset(); 235 if (BlockOffsetArrayUseUnallocatedBlock) { 236 assert(end() == mr.end(), "We are compacting to the bottom of CMS gen"); 237 // Everything's allocated until proven otherwise. 238 _bt.set_unallocated_block(end()); 239 } 240 if (!mr.is_empty()) { 241 assert(mr.word_size() >= MinChunkSize, "Chunk size is too small"); 242 _bt.single_block(mr.start(), mr.word_size()); 243 FreeChunk* fc = (FreeChunk*) mr.start(); 244 fc->setSize(mr.word_size()); 245 if (mr.word_size() >= IndexSetSize ) { 246 returnChunkToDictionary(fc); 247 } else { 248 _bt.verify_not_unallocated((HeapWord*)fc, fc->size()); 249 _indexedFreeList[mr.word_size()].returnChunkAtHead(fc); 250 } 251 } 252 _promoInfo.reset(); 253 _smallLinearAllocBlock._ptr = NULL; 254 _smallLinearAllocBlock._word_size = 0; 255 } 256 257 void CompactibleFreeListSpace::reset_after_compaction() { 258 // Reset the space to the new reality - one free chunk. 259 MemRegion mr(compaction_top(), end()); 260 reset(mr); 261 // Now refill the linear allocation block(s) if possible. 262 if (_adaptive_freelists) { 263 refillLinearAllocBlocksIfNeeded(); 264 } else { 265 // Place as much of mr in the linAB as we can get, 266 // provided it was big enough to go into the dictionary. 267 FreeChunk* fc = dictionary()->findLargestDict(); 268 if (fc != NULL) { 269 assert(fc->size() == mr.word_size(), 270 "Why was the chunk broken up?"); 271 removeChunkFromDictionary(fc); 272 HeapWord* addr = (HeapWord*) fc; 273 _smallLinearAllocBlock.set(addr, fc->size() , 274 1024*SmallForLinearAlloc, fc->size()); 275 // Note that _unallocated_block is not updated here. 276 } 277 } 278 } 279 280 // Walks the entire dictionary, returning a coterminal 281 // chunk, if it exists. Use with caution since it involves 282 // a potentially complete walk of a potentially large tree. 283 FreeChunk* CompactibleFreeListSpace::find_chunk_at_end() { 284 285 assert_lock_strong(&_freelistLock); 286 287 return dictionary()->find_chunk_ends_at(end()); 288 } 289 290 291 #ifndef PRODUCT 292 void CompactibleFreeListSpace::initializeIndexedFreeListArrayReturnedBytes() { 293 for (size_t i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) { 294 _indexedFreeList[i].allocation_stats()->set_returnedBytes(0); 295 } 296 } 297 298 size_t CompactibleFreeListSpace::sumIndexedFreeListArrayReturnedBytes() { 299 size_t sum = 0; 300 for (size_t i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) { 301 sum += _indexedFreeList[i].allocation_stats()->returnedBytes(); 302 } 303 return sum; 304 } 305 306 size_t CompactibleFreeListSpace::totalCountInIndexedFreeLists() const { 307 size_t count = 0; 308 for (int i = MinChunkSize; i < IndexSetSize; i++) { 309 debug_only( 310 ssize_t total_list_count = 0; 311 for (FreeChunk* fc = _indexedFreeList[i].head(); fc != NULL; 312 fc = fc->next()) { 313 total_list_count++; 314 } 315 assert(total_list_count == _indexedFreeList[i].count(), 316 "Count in list is incorrect"); 317 ) 318 count += _indexedFreeList[i].count(); 319 } 320 return count; 321 } 322 323 size_t CompactibleFreeListSpace::totalCount() { 324 size_t num = totalCountInIndexedFreeLists(); 325 num += dictionary()->totalCount(); 326 if (_smallLinearAllocBlock._word_size != 0) { 327 num++; 328 } 329 return num; 330 } 331 #endif 332 333 bool CompactibleFreeListSpace::is_free_block(const HeapWord* p) const { 334 FreeChunk* fc = (FreeChunk*) p; 335 return fc->isFree(); 336 } 337 338 size_t CompactibleFreeListSpace::used() const { 339 return capacity() - free(); 340 } 341 342 size_t CompactibleFreeListSpace::free() const { 343 // "MT-safe, but not MT-precise"(TM), if you will: i.e. 344 // if you do this while the structures are in flux you 345 // may get an approximate answer only; for instance 346 // because there is concurrent allocation either 347 // directly by mutators or for promotion during a GC. 348 // It's "MT-safe", however, in the sense that you are guaranteed 349 // not to crash and burn, for instance, because of walking 350 // pointers that could disappear as you were walking them. 351 // The approximation is because the various components 352 // that are read below are not read atomically (and 353 // further the computation of totalSizeInIndexedFreeLists() 354 // is itself a non-atomic computation. The normal use of 355 // this is during a resize operation at the end of GC 356 // and at that time you are guaranteed to get the 357 // correct actual value. However, for instance, this is 358 // also read completely asynchronously by the "perf-sampler" 359 // that supports jvmstat, and you are apt to see the values 360 // flicker in such cases. 361 assert(_dictionary != NULL, "No _dictionary?"); 362 return (_dictionary->totalChunkSize(DEBUG_ONLY(freelistLock())) + 363 totalSizeInIndexedFreeLists() + 364 _smallLinearAllocBlock._word_size) * HeapWordSize; 365 } 366 367 size_t CompactibleFreeListSpace::max_alloc_in_words() const { 368 assert(_dictionary != NULL, "No _dictionary?"); 369 assert_locked(); 370 size_t res = _dictionary->maxChunkSize(); 371 res = MAX2(res, MIN2(_smallLinearAllocBlock._word_size, 372 (size_t) SmallForLinearAlloc - 1)); 373 // XXX the following could potentially be pretty slow; 374 // should one, pesimally for the rare cases when res 375 // caclulated above is less than IndexSetSize, 376 // just return res calculated above? My reasoning was that 377 // those cases will be so rare that the extra time spent doesn't 378 // really matter.... 379 // Note: do not change the loop test i >= res + IndexSetStride 380 // to i > res below, because i is unsigned and res may be zero. 381 for (size_t i = IndexSetSize - 1; i >= res + IndexSetStride; 382 i -= IndexSetStride) { 383 if (_indexedFreeList[i].head() != NULL) { 384 assert(_indexedFreeList[i].count() != 0, "Inconsistent FreeList"); 385 return i; 386 } 387 } 388 return res; 389 } 390 391 void CompactibleFreeListSpace::reportFreeListStatistics() const { 392 assert_lock_strong(&_freelistLock); 393 assert(PrintFLSStatistics != 0, "Reporting error"); 394 _dictionary->reportStatistics(); 395 if (PrintFLSStatistics > 1) { 396 reportIndexedFreeListStatistics(); 397 size_t totalSize = totalSizeInIndexedFreeLists() + 398 _dictionary->totalChunkSize(DEBUG_ONLY(freelistLock())); 399 gclog_or_tty->print(" free=%ld frag=%1.4f\n", totalSize, flsFrag()); 400 } 401 } 402 403 void CompactibleFreeListSpace::reportIndexedFreeListStatistics() const { 404 assert_lock_strong(&_freelistLock); 405 gclog_or_tty->print("Statistics for IndexedFreeLists:\n" 406 "--------------------------------\n"); 407 size_t totalSize = totalSizeInIndexedFreeLists(); 408 size_t freeBlocks = numFreeBlocksInIndexedFreeLists(); 409 gclog_or_tty->print("Total Free Space: %d\n", totalSize); 410 gclog_or_tty->print("Max Chunk Size: %d\n", maxChunkSizeInIndexedFreeLists()); 411 gclog_or_tty->print("Number of Blocks: %d\n", freeBlocks); 412 if (freeBlocks != 0) { 413 gclog_or_tty->print("Av. Block Size: %d\n", totalSize/freeBlocks); 414 } 415 } 416 417 size_t CompactibleFreeListSpace::numFreeBlocksInIndexedFreeLists() const { 418 size_t res = 0; 419 for (size_t i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) { 420 debug_only( 421 ssize_t recount = 0; 422 for (FreeChunk* fc = _indexedFreeList[i].head(); fc != NULL; 423 fc = fc->next()) { 424 recount += 1; 425 } 426 assert(recount == _indexedFreeList[i].count(), 427 "Incorrect count in list"); 428 ) 429 res += _indexedFreeList[i].count(); 430 } 431 return res; 432 } 433 434 size_t CompactibleFreeListSpace::maxChunkSizeInIndexedFreeLists() const { 435 for (size_t i = IndexSetSize - 1; i != 0; i -= IndexSetStride) { 436 if (_indexedFreeList[i].head() != NULL) { 437 assert(_indexedFreeList[i].count() != 0, "Inconsistent FreeList"); 438 return (size_t)i; 439 } 440 } 441 return 0; 442 } 443 444 void CompactibleFreeListSpace::set_end(HeapWord* value) { 445 HeapWord* prevEnd = end(); 446 assert(prevEnd != value, "unnecessary set_end call"); 447 assert(prevEnd == NULL || value >= unallocated_block(), "New end is below unallocated block"); 448 _end = value; 449 if (prevEnd != NULL) { 450 // Resize the underlying block offset table. 451 _bt.resize(pointer_delta(value, bottom())); 452 if (value <= prevEnd) { 453 assert(value >= unallocated_block(), "New end is below unallocated block"); 454 } else { 455 // Now, take this new chunk and add it to the free blocks. 456 // Note that the BOT has not yet been updated for this block. 457 size_t newFcSize = pointer_delta(value, prevEnd); 458 // XXX This is REALLY UGLY and should be fixed up. XXX 459 if (!_adaptive_freelists && _smallLinearAllocBlock._ptr == NULL) { 460 // Mark the boundary of the new block in BOT 461 _bt.mark_block(prevEnd, value); 462 // put it all in the linAB 463 if (ParallelGCThreads == 0) { 464 _smallLinearAllocBlock._ptr = prevEnd; 465 _smallLinearAllocBlock._word_size = newFcSize; 466 repairLinearAllocBlock(&_smallLinearAllocBlock); 467 } else { // ParallelGCThreads > 0 468 MutexLockerEx x(parDictionaryAllocLock(), 469 Mutex::_no_safepoint_check_flag); 470 _smallLinearAllocBlock._ptr = prevEnd; 471 _smallLinearAllocBlock._word_size = newFcSize; 472 repairLinearAllocBlock(&_smallLinearAllocBlock); 473 } 474 // Births of chunks put into a LinAB are not recorded. Births 475 // of chunks as they are allocated out of a LinAB are. 476 } else { 477 // Add the block to the free lists, if possible coalescing it 478 // with the last free block, and update the BOT and census data. 479 addChunkToFreeListsAtEndRecordingStats(prevEnd, newFcSize); 480 } 481 } 482 } 483 } 484 485 class FreeListSpace_DCTOC : public Filtering_DCTOC { 486 CompactibleFreeListSpace* _cfls; 487 CMSCollector* _collector; 488 protected: 489 // Override. 490 #define walk_mem_region_with_cl_DECL(ClosureType) \ 491 virtual void walk_mem_region_with_cl(MemRegion mr, \ 492 HeapWord* bottom, HeapWord* top, \ 493 ClosureType* cl); \ 494 void walk_mem_region_with_cl_par(MemRegion mr, \ 495 HeapWord* bottom, HeapWord* top, \ 496 ClosureType* cl); \ 497 void walk_mem_region_with_cl_nopar(MemRegion mr, \ 498 HeapWord* bottom, HeapWord* top, \ 499 ClosureType* cl) 500 walk_mem_region_with_cl_DECL(OopClosure); 501 walk_mem_region_with_cl_DECL(FilteringClosure); 502 503 public: 504 FreeListSpace_DCTOC(CompactibleFreeListSpace* sp, 505 CMSCollector* collector, 506 OopClosure* cl, 507 CardTableModRefBS::PrecisionStyle precision, 508 HeapWord* boundary) : 509 Filtering_DCTOC(sp, cl, precision, boundary), 510 _cfls(sp), _collector(collector) {} 511 }; 512 513 // We de-virtualize the block-related calls below, since we know that our 514 // space is a CompactibleFreeListSpace. 515 #define FreeListSpace_DCTOC__walk_mem_region_with_cl_DEFN(ClosureType) \ 516 void FreeListSpace_DCTOC::walk_mem_region_with_cl(MemRegion mr, \ 517 HeapWord* bottom, \ 518 HeapWord* top, \ 519 ClosureType* cl) { \ 520 if (SharedHeap::heap()->n_par_threads() > 0) { \ 521 walk_mem_region_with_cl_par(mr, bottom, top, cl); \ 522 } else { \ 523 walk_mem_region_with_cl_nopar(mr, bottom, top, cl); \ 524 } \ 525 } \ 526 void FreeListSpace_DCTOC::walk_mem_region_with_cl_par(MemRegion mr, \ 527 HeapWord* bottom, \ 528 HeapWord* top, \ 529 ClosureType* cl) { \ 530 /* Skip parts that are before "mr", in case "block_start" sent us \ 531 back too far. */ \ 532 HeapWord* mr_start = mr.start(); \ 533 size_t bot_size = _cfls->CompactibleFreeListSpace::block_size(bottom); \ 534 HeapWord* next = bottom + bot_size; \ 535 while (next < mr_start) { \ 536 bottom = next; \ 537 bot_size = _cfls->CompactibleFreeListSpace::block_size(bottom); \ 538 next = bottom + bot_size; \ 539 } \ 540 \ 541 while (bottom < top) { \ 542 if (_cfls->CompactibleFreeListSpace::block_is_obj(bottom) && \ 543 !_cfls->CompactibleFreeListSpace::obj_allocated_since_save_marks( \ 544 oop(bottom)) && \ 545 !_collector->CMSCollector::is_dead_obj(oop(bottom))) { \ 546 size_t word_sz = oop(bottom)->oop_iterate(cl, mr); \ 547 bottom += _cfls->adjustObjectSize(word_sz); \ 548 } else { \ 549 bottom += _cfls->CompactibleFreeListSpace::block_size(bottom); \ 550 } \ 551 } \ 552 } \ 553 void FreeListSpace_DCTOC::walk_mem_region_with_cl_nopar(MemRegion mr, \ 554 HeapWord* bottom, \ 555 HeapWord* top, \ 556 ClosureType* cl) { \ 557 /* Skip parts that are before "mr", in case "block_start" sent us \ 558 back too far. */ \ 559 HeapWord* mr_start = mr.start(); \ 560 size_t bot_size = _cfls->CompactibleFreeListSpace::block_size_nopar(bottom); \ 561 HeapWord* next = bottom + bot_size; \ 562 while (next < mr_start) { \ 563 bottom = next; \ 564 bot_size = _cfls->CompactibleFreeListSpace::block_size_nopar(bottom); \ 565 next = bottom + bot_size; \ 566 } \ 567 \ 568 while (bottom < top) { \ 569 if (_cfls->CompactibleFreeListSpace::block_is_obj_nopar(bottom) && \ 570 !_cfls->CompactibleFreeListSpace::obj_allocated_since_save_marks( \ 571 oop(bottom)) && \ 572 !_collector->CMSCollector::is_dead_obj(oop(bottom))) { \ 573 size_t word_sz = oop(bottom)->oop_iterate(cl, mr); \ 574 bottom += _cfls->adjustObjectSize(word_sz); \ 575 } else { \ 576 bottom += _cfls->CompactibleFreeListSpace::block_size_nopar(bottom); \ 577 } \ 578 } \ 579 } 580 581 // (There are only two of these, rather than N, because the split is due 582 // only to the introduction of the FilteringClosure, a local part of the 583 // impl of this abstraction.) 584 FreeListSpace_DCTOC__walk_mem_region_with_cl_DEFN(OopClosure) 585 FreeListSpace_DCTOC__walk_mem_region_with_cl_DEFN(FilteringClosure) 586 587 DirtyCardToOopClosure* 588 CompactibleFreeListSpace::new_dcto_cl(OopClosure* cl, 589 CardTableModRefBS::PrecisionStyle precision, 590 HeapWord* boundary) { 591 return new FreeListSpace_DCTOC(this, _collector, cl, precision, boundary); 592 } 593 594 595 // Note on locking for the space iteration functions: 596 // since the collector's iteration activities are concurrent with 597 // allocation activities by mutators, absent a suitable mutual exclusion 598 // mechanism the iterators may go awry. For instace a block being iterated 599 // may suddenly be allocated or divided up and part of it allocated and 600 // so on. 601 602 // Apply the given closure to each block in the space. 603 void CompactibleFreeListSpace::blk_iterate_careful(BlkClosureCareful* cl) { 604 assert_lock_strong(freelistLock()); 605 HeapWord *cur, *limit; 606 for (cur = bottom(), limit = end(); cur < limit; 607 cur += cl->do_blk_careful(cur)); 608 } 609 610 // Apply the given closure to each block in the space. 611 void CompactibleFreeListSpace::blk_iterate(BlkClosure* cl) { 612 assert_lock_strong(freelistLock()); 613 HeapWord *cur, *limit; 614 for (cur = bottom(), limit = end(); cur < limit; 615 cur += cl->do_blk(cur)); 616 } 617 618 // Apply the given closure to each oop in the space. 619 void CompactibleFreeListSpace::oop_iterate(OopClosure* cl) { 620 assert_lock_strong(freelistLock()); 621 HeapWord *cur, *limit; 622 size_t curSize; 623 for (cur = bottom(), limit = end(); cur < limit; 624 cur += curSize) { 625 curSize = block_size(cur); 626 if (block_is_obj(cur)) { 627 oop(cur)->oop_iterate(cl); 628 } 629 } 630 } 631 632 // Apply the given closure to each oop in the space \intersect memory region. 633 void CompactibleFreeListSpace::oop_iterate(MemRegion mr, OopClosure* cl) { 634 assert_lock_strong(freelistLock()); 635 if (is_empty()) { 636 return; 637 } 638 MemRegion cur = MemRegion(bottom(), end()); 639 mr = mr.intersection(cur); 640 if (mr.is_empty()) { 641 return; 642 } 643 if (mr.equals(cur)) { 644 oop_iterate(cl); 645 return; 646 } 647 assert(mr.end() <= end(), "just took an intersection above"); 648 HeapWord* obj_addr = block_start(mr.start()); 649 HeapWord* t = mr.end(); 650 651 SpaceMemRegionOopsIterClosure smr_blk(cl, mr); 652 if (block_is_obj(obj_addr)) { 653 // Handle first object specially. 654 oop obj = oop(obj_addr); 655 obj_addr += adjustObjectSize(obj->oop_iterate(&smr_blk)); 656 } else { 657 FreeChunk* fc = (FreeChunk*)obj_addr; 658 obj_addr += fc->size(); 659 } 660 while (obj_addr < t) { 661 HeapWord* obj = obj_addr; 662 obj_addr += block_size(obj_addr); 663 // If "obj_addr" is not greater than top, then the 664 // entire object "obj" is within the region. 665 if (obj_addr <= t) { 666 if (block_is_obj(obj)) { 667 oop(obj)->oop_iterate(cl); 668 } 669 } else { 670 // "obj" extends beyond end of region 671 if (block_is_obj(obj)) { 672 oop(obj)->oop_iterate(&smr_blk); 673 } 674 break; 675 } 676 } 677 } 678 679 // NOTE: In the following methods, in order to safely be able to 680 // apply the closure to an object, we need to be sure that the 681 // object has been initialized. We are guaranteed that an object 682 // is initialized if we are holding the Heap_lock with the 683 // world stopped. 684 void CompactibleFreeListSpace::verify_objects_initialized() const { 685 if (is_init_completed()) { 686 assert_locked_or_safepoint(Heap_lock); 687 if (Universe::is_fully_initialized()) { 688 guarantee(SafepointSynchronize::is_at_safepoint(), 689 "Required for objects to be initialized"); 690 } 691 } // else make a concession at vm start-up 692 } 693 694 // Apply the given closure to each object in the space 695 void CompactibleFreeListSpace::object_iterate(ObjectClosure* blk) { 696 assert_lock_strong(freelistLock()); 697 NOT_PRODUCT(verify_objects_initialized()); 698 HeapWord *cur, *limit; 699 size_t curSize; 700 for (cur = bottom(), limit = end(); cur < limit; 701 cur += curSize) { 702 curSize = block_size(cur); 703 if (block_is_obj(cur)) { 704 blk->do_object(oop(cur)); 705 } 706 } 707 } 708 709 void CompactibleFreeListSpace::object_iterate_mem(MemRegion mr, 710 UpwardsObjectClosure* cl) { 711 assert_locked(); 712 NOT_PRODUCT(verify_objects_initialized()); 713 Space::object_iterate_mem(mr, cl); 714 } 715 716 // Callers of this iterator beware: The closure application should 717 // be robust in the face of uninitialized objects and should (always) 718 // return a correct size so that the next addr + size below gives us a 719 // valid block boundary. [See for instance, 720 // ScanMarkedObjectsAgainCarefullyClosure::do_object_careful() 721 // in ConcurrentMarkSweepGeneration.cpp.] 722 HeapWord* 723 CompactibleFreeListSpace::object_iterate_careful(ObjectClosureCareful* cl) { 724 assert_lock_strong(freelistLock()); 725 HeapWord *addr, *last; 726 size_t size; 727 for (addr = bottom(), last = end(); 728 addr < last; addr += size) { 729 FreeChunk* fc = (FreeChunk*)addr; 730 if (fc->isFree()) { 731 // Since we hold the free list lock, which protects direct 732 // allocation in this generation by mutators, a free object 733 // will remain free throughout this iteration code. 734 size = fc->size(); 735 } else { 736 // Note that the object need not necessarily be initialized, 737 // because (for instance) the free list lock does NOT protect 738 // object initialization. The closure application below must 739 // therefore be correct in the face of uninitialized objects. 740 size = cl->do_object_careful(oop(addr)); 741 if (size == 0) { 742 // An unparsable object found. Signal early termination. 743 return addr; 744 } 745 } 746 } 747 return NULL; 748 } 749 750 // Callers of this iterator beware: The closure application should 751 // be robust in the face of uninitialized objects and should (always) 752 // return a correct size so that the next addr + size below gives us a 753 // valid block boundary. [See for instance, 754 // ScanMarkedObjectsAgainCarefullyClosure::do_object_careful() 755 // in ConcurrentMarkSweepGeneration.cpp.] 756 HeapWord* 757 CompactibleFreeListSpace::object_iterate_careful_m(MemRegion mr, 758 ObjectClosureCareful* cl) { 759 assert_lock_strong(freelistLock()); 760 // Can't use used_region() below because it may not necessarily 761 // be the same as [bottom(),end()); although we could 762 // use [used_region().start(),round_to(used_region().end(),CardSize)), 763 // that appears too cumbersome, so we just do the simpler check 764 // in the assertion below. 765 assert(!mr.is_empty() && MemRegion(bottom(),end()).contains(mr), 766 "mr should be non-empty and within used space"); 767 HeapWord *addr, *end; 768 size_t size; 769 for (addr = block_start_careful(mr.start()), end = mr.end(); 770 addr < end; addr += size) { 771 FreeChunk* fc = (FreeChunk*)addr; 772 if (fc->isFree()) { 773 // Since we hold the free list lock, which protects direct 774 // allocation in this generation by mutators, a free object 775 // will remain free throughout this iteration code. 776 size = fc->size(); 777 } else { 778 // Note that the object need not necessarily be initialized, 779 // because (for instance) the free list lock does NOT protect 780 // object initialization. The closure application below must 781 // therefore be correct in the face of uninitialized objects. 782 size = cl->do_object_careful_m(oop(addr), mr); 783 if (size == 0) { 784 // An unparsable object found. Signal early termination. 785 return addr; 786 } 787 } 788 } 789 return NULL; 790 } 791 792 793 HeapWord* CompactibleFreeListSpace::block_start(const void* p) const { 794 NOT_PRODUCT(verify_objects_initialized()); 795 return _bt.block_start(p); 796 } 797 798 HeapWord* CompactibleFreeListSpace::block_start_careful(const void* p) const { 799 return _bt.block_start_careful(p); 800 } 801 802 size_t CompactibleFreeListSpace::block_size(const HeapWord* p) const { 803 NOT_PRODUCT(verify_objects_initialized()); 804 assert(MemRegion(bottom(), end()).contains(p), "p not in space"); 805 // This must be volatile, or else there is a danger that the compiler 806 // will compile the code below into a sometimes-infinite loop, by keeping 807 // the value read the first time in a register. 808 oop o = (oop)p; 809 volatile oop* second_word_addr = o->klass_addr(); 810 while (true) { 811 klassOop k = (klassOop)(*second_word_addr); 812 // We must do this until we get a consistent view of the object. 813 if (FreeChunk::secondWordIndicatesFreeChunk((intptr_t)k)) { 814 FreeChunk* fc = (FreeChunk*)p; 815 volatile size_t* sz_addr = (volatile size_t*)(fc->size_addr()); 816 size_t res = (*sz_addr); 817 klassOop k2 = (klassOop)(*second_word_addr); // Read to confirm. 818 if (k == k2) { 819 assert(res != 0, "Block size should not be 0"); 820 return res; 821 } 822 } else if (k != NULL) { 823 assert(k->is_oop(true /* ignore mark word */), "Should really be klass oop."); 824 assert(o->is_parsable(), "Should be parsable"); 825 assert(o->is_oop(true /* ignore mark word */), "Should be an oop."); 826 size_t res = o->size_given_klass(k->klass_part()); 827 res = adjustObjectSize(res); 828 assert(res != 0, "Block size should not be 0"); 829 return res; 830 } 831 } 832 } 833 834 // A variant of the above that uses the Printezis bits for 835 // unparsable but allocated objects. This avoids any possible 836 // stalls waiting for mutators to initialize objects, and is 837 // thus potentially faster than the variant above. However, 838 // this variant may return a zero size for a block that is 839 // under mutation and for which a consistent size cannot be 840 // inferred without stalling; see CMSCollector::block_size_if_printezis_bits(). 841 size_t CompactibleFreeListSpace::block_size_no_stall(HeapWord* p, 842 const CMSCollector* c) 843 const { 844 assert(MemRegion(bottom(), end()).contains(p), "p not in space"); 845 // This must be volatile, or else there is a danger that the compiler 846 // will compile the code below into a sometimes-infinite loop, by keeping 847 // the value read the first time in a register. 848 oop o = (oop)p; 849 volatile oop* second_word_addr = o->klass_addr(); 850 DEBUG_ONLY(uint loops = 0;) 851 while (true) { 852 klassOop k = (klassOop)(*second_word_addr); 853 // We must do this until we get a consistent view of the object. 854 if (FreeChunk::secondWordIndicatesFreeChunk((intptr_t)k)) { 855 FreeChunk* fc = (FreeChunk*)p; 856 volatile size_t* sz_addr = (volatile size_t*)(fc->size_addr()); 857 size_t res = (*sz_addr); 858 klassOop k2 = (klassOop)(*second_word_addr); // Read to confirm. 859 if (k == k2) { 860 assert(res != 0, "Block size should not be 0"); 861 assert(loops == 0, "Should be 0"); 862 return res; 863 } 864 } else if (k != NULL && o->is_parsable()) { 865 assert(k->is_oop(), "Should really be klass oop."); 866 assert(o->is_oop(), "Should be an oop"); 867 size_t res = o->size_given_klass(k->klass_part()); 868 res = adjustObjectSize(res); 869 assert(res != 0, "Block size should not be 0"); 870 return res; 871 } else { 872 return c->block_size_if_printezis_bits(p); 873 } 874 assert(loops == 0, "Can loop at most once"); 875 DEBUG_ONLY(loops++;) 876 } 877 } 878 879 size_t CompactibleFreeListSpace::block_size_nopar(const HeapWord* p) const { 880 NOT_PRODUCT(verify_objects_initialized()); 881 assert(MemRegion(bottom(), end()).contains(p), "p not in space"); 882 FreeChunk* fc = (FreeChunk*)p; 883 if (fc->isFree()) { 884 return fc->size(); 885 } else { 886 // Ignore mark word because this may be a recently promoted 887 // object whose mark word is used to chain together grey 888 // objects (the last one would have a null value). 889 assert(oop(p)->is_oop(true), "Should be an oop"); 890 return adjustObjectSize(oop(p)->size()); 891 } 892 } 893 894 // This implementation assumes that the property of "being an object" is 895 // stable. But being a free chunk may not be (because of parallel 896 // promotion.) 897 bool CompactibleFreeListSpace::block_is_obj(const HeapWord* p) const { 898 FreeChunk* fc = (FreeChunk*)p; 899 assert(is_in_reserved(p), "Should be in space"); 900 // When doing a mark-sweep-compact of the CMS generation, this 901 // assertion may fail because prepare_for_compaction() uses 902 // space that is garbage to maintain information on ranges of 903 // live objects so that these live ranges can be moved as a whole. 904 // Comment out this assertion until that problem can be solved 905 // (i.e., that the block start calculation may look at objects 906 // at address below "p" in finding the object that contains "p" 907 // and those objects (if garbage) may have been modified to hold 908 // live range information. 909 // assert(ParallelGCThreads > 0 || _bt.block_start(p) == p, "Should be a block boundary"); 910 klassOop k = oop(p)->klass(); 911 intptr_t ki = (intptr_t)k; 912 if (FreeChunk::secondWordIndicatesFreeChunk(ki)) return false; 913 if (k != NULL) { 914 // Ignore mark word because it may have been used to 915 // chain together promoted objects (the last one 916 // would have a null value). 917 assert(oop(p)->is_oop(true), "Should be an oop"); 918 return true; 919 } else { 920 return false; // Was not an object at the start of collection. 921 } 922 } 923 924 // Check if the object is alive. This fact is checked either by consulting 925 // the main marking bitmap in the sweeping phase or, if it's a permanent 926 // generation and we're not in the sweeping phase, by checking the 927 // perm_gen_verify_bit_map where we store the "deadness" information if 928 // we did not sweep the perm gen in the most recent previous GC cycle. 929 bool CompactibleFreeListSpace::obj_is_alive(const HeapWord* p) const { 930 assert (block_is_obj(p), "The address should point to an object"); 931 932 // If we're sweeping, we use object liveness information from the main bit map 933 // for both perm gen and old gen. 934 // We don't need to lock the bitmap (live_map or dead_map below), because 935 // EITHER we are in the middle of the sweeping phase, and the 936 // main marking bit map (live_map below) is locked, 937 // OR we're in other phases and perm_gen_verify_bit_map (dead_map below) 938 // is stable, because it's mutated only in the sweeping phase. 939 if (_collector->abstract_state() == CMSCollector::Sweeping) { 940 CMSBitMap* live_map = _collector->markBitMap(); 941 return live_map->isMarked((HeapWord*) p); 942 } else { 943 // If we're not currently sweeping and we haven't swept the perm gen in 944 // the previous concurrent cycle then we may have dead but unswept objects 945 // in the perm gen. In this case, we use the "deadness" information 946 // that we had saved in perm_gen_verify_bit_map at the last sweep. 947 if (!CMSClassUnloadingEnabled && _collector->_permGen->reserved().contains(p)) { 948 if (_collector->verifying()) { 949 CMSBitMap* dead_map = _collector->perm_gen_verify_bit_map(); 950 // Object is marked in the dead_map bitmap at the previous sweep 951 // when we know that it's dead; if the bitmap is not allocated then 952 // the object is alive. 953 return (dead_map->sizeInBits() == 0) // bit_map has been allocated 954 || !dead_map->par_isMarked((HeapWord*) p); 955 } else { 956 return false; // We can't say for sure if it's live, so we say that it's dead. 957 } 958 } 959 } 960 return true; 961 } 962 963 bool CompactibleFreeListSpace::block_is_obj_nopar(const HeapWord* p) const { 964 FreeChunk* fc = (FreeChunk*)p; 965 assert(is_in_reserved(p), "Should be in space"); 966 assert(_bt.block_start(p) == p, "Should be a block boundary"); 967 if (!fc->isFree()) { 968 // Ignore mark word because it may have been used to 969 // chain together promoted objects (the last one 970 // would have a null value). 971 assert(oop(p)->is_oop(true), "Should be an oop"); 972 return true; 973 } 974 return false; 975 } 976 977 // "MT-safe but not guaranteed MT-precise" (TM); you may get an 978 // approximate answer if you don't hold the freelistlock when you call this. 979 size_t CompactibleFreeListSpace::totalSizeInIndexedFreeLists() const { 980 size_t size = 0; 981 for (size_t i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) { 982 debug_only( 983 // We may be calling here without the lock in which case we 984 // won't do this modest sanity check. 985 if (freelistLock()->owned_by_self()) { 986 size_t total_list_size = 0; 987 for (FreeChunk* fc = _indexedFreeList[i].head(); fc != NULL; 988 fc = fc->next()) { 989 total_list_size += i; 990 } 991 assert(total_list_size == i * _indexedFreeList[i].count(), 992 "Count in list is incorrect"); 993 } 994 ) 995 size += i * _indexedFreeList[i].count(); 996 } 997 return size; 998 } 999 1000 HeapWord* CompactibleFreeListSpace::par_allocate(size_t size) { 1001 MutexLockerEx x(freelistLock(), Mutex::_no_safepoint_check_flag); 1002 return allocate(size); 1003 } 1004 1005 HeapWord* 1006 CompactibleFreeListSpace::getChunkFromSmallLinearAllocBlockRemainder(size_t size) { 1007 return getChunkFromLinearAllocBlockRemainder(&_smallLinearAllocBlock, size); 1008 } 1009 1010 HeapWord* CompactibleFreeListSpace::allocate(size_t size) { 1011 assert_lock_strong(freelistLock()); 1012 HeapWord* res = NULL; 1013 assert(size == adjustObjectSize(size), 1014 "use adjustObjectSize() before calling into allocate()"); 1015 1016 if (_adaptive_freelists) { 1017 res = allocate_adaptive_freelists(size); 1018 } else { // non-adaptive free lists 1019 res = allocate_non_adaptive_freelists(size); 1020 } 1021 1022 if (res != NULL) { 1023 // check that res does lie in this space! 1024 assert(is_in_reserved(res), "Not in this space!"); 1025 assert(is_aligned((void*)res), "alignment check"); 1026 1027 FreeChunk* fc = (FreeChunk*)res; 1028 fc->markNotFree(); 1029 assert(!fc->isFree(), "shouldn't be marked free"); 1030 assert(oop(fc)->klass() == NULL, "should look uninitialized"); 1031 // Verify that the block offset table shows this to 1032 // be a single block, but not one which is unallocated. 1033 _bt.verify_single_block(res, size); 1034 _bt.verify_not_unallocated(res, size); 1035 // mangle a just allocated object with a distinct pattern. 1036 debug_only(fc->mangleAllocated(size)); 1037 } 1038 1039 return res; 1040 } 1041 1042 HeapWord* CompactibleFreeListSpace::allocate_non_adaptive_freelists(size_t size) { 1043 HeapWord* res = NULL; 1044 // try and use linear allocation for smaller blocks 1045 if (size < _smallLinearAllocBlock._allocation_size_limit) { 1046 // if successful, the following also adjusts block offset table 1047 res = getChunkFromSmallLinearAllocBlock(size); 1048 } 1049 // Else triage to indexed lists for smaller sizes 1050 if (res == NULL) { 1051 if (size < SmallForDictionary) { 1052 res = (HeapWord*) getChunkFromIndexedFreeList(size); 1053 } else { 1054 // else get it from the big dictionary; if even this doesn't 1055 // work we are out of luck. 1056 res = (HeapWord*)getChunkFromDictionaryExact(size); 1057 } 1058 } 1059 1060 return res; 1061 } 1062 1063 HeapWord* CompactibleFreeListSpace::allocate_adaptive_freelists(size_t size) { 1064 assert_lock_strong(freelistLock()); 1065 HeapWord* res = NULL; 1066 assert(size == adjustObjectSize(size), 1067 "use adjustObjectSize() before calling into allocate()"); 1068 1069 // Strategy 1070 // if small 1071 // exact size from small object indexed list if small 1072 // small or large linear allocation block (linAB) as appropriate 1073 // take from lists of greater sized chunks 1074 // else 1075 // dictionary 1076 // small or large linear allocation block if it has the space 1077 // Try allocating exact size from indexTable first 1078 if (size < IndexSetSize) { 1079 res = (HeapWord*) getChunkFromIndexedFreeList(size); 1080 if(res != NULL) { 1081 assert(res != (HeapWord*)_indexedFreeList[size].head(), 1082 "Not removed from free list"); 1083 // no block offset table adjustment is necessary on blocks in 1084 // the indexed lists. 1085 1086 // Try allocating from the small LinAB 1087 } else if (size < _smallLinearAllocBlock._allocation_size_limit && 1088 (res = getChunkFromSmallLinearAllocBlock(size)) != NULL) { 1089 // if successful, the above also adjusts block offset table 1090 // Note that this call will refill the LinAB to 1091 // satisfy the request. This is different that 1092 // evm. 1093 // Don't record chunk off a LinAB? smallSplitBirth(size); 1094 1095 } else { 1096 // Raid the exact free lists larger than size, even if they are not 1097 // overpopulated. 1098 res = (HeapWord*) getChunkFromGreater(size); 1099 } 1100 } else { 1101 // Big objects get allocated directly from the dictionary. 1102 res = (HeapWord*) getChunkFromDictionaryExact(size); 1103 if (res == NULL) { 1104 // Try hard not to fail since an allocation failure will likely 1105 // trigger a synchronous GC. Try to get the space from the 1106 // allocation blocks. 1107 res = getChunkFromSmallLinearAllocBlockRemainder(size); 1108 } 1109 } 1110 1111 return res; 1112 } 1113 1114 // A worst-case estimate of the space required (in HeapWords) to expand the heap 1115 // when promoting obj. 1116 size_t CompactibleFreeListSpace::expansionSpaceRequired(size_t obj_size) const { 1117 // Depending on the object size, expansion may require refilling either a 1118 // bigLAB or a smallLAB plus refilling a PromotionInfo object. MinChunkSize 1119 // is added because the dictionary may over-allocate to avoid fragmentation. 1120 size_t space = obj_size; 1121 if (!_adaptive_freelists) { 1122 space = MAX2(space, _smallLinearAllocBlock._refillSize); 1123 } 1124 space += _promoInfo.refillSize() + 2 * MinChunkSize; 1125 return space; 1126 } 1127 1128 FreeChunk* CompactibleFreeListSpace::getChunkFromGreater(size_t numWords) { 1129 FreeChunk* ret; 1130 1131 assert(numWords >= MinChunkSize, "Size is less than minimum"); 1132 assert(linearAllocationWouldFail() || bestFitFirst(), 1133 "Should not be here"); 1134 1135 size_t i; 1136 size_t currSize = numWords + MinChunkSize; 1137 assert(currSize % MinObjAlignment == 0, "currSize should be aligned"); 1138 for (i = currSize; i < IndexSetSize; i += IndexSetStride) { 1139 FreeList* fl = &_indexedFreeList[i]; 1140 if (fl->head()) { 1141 ret = getFromListGreater(fl, numWords); 1142 assert(ret == NULL || ret->isFree(), "Should be returning a free chunk"); 1143 return ret; 1144 } 1145 } 1146 1147 currSize = MAX2((size_t)SmallForDictionary, 1148 (size_t)(numWords + MinChunkSize)); 1149 1150 /* Try to get a chunk that satisfies request, while avoiding 1151 fragmentation that can't be handled. */ 1152 { 1153 ret = dictionary()->getChunk(currSize); 1154 if (ret != NULL) { 1155 assert(ret->size() - numWords >= MinChunkSize, 1156 "Chunk is too small"); 1157 _bt.allocated((HeapWord*)ret, ret->size()); 1158 /* Carve returned chunk. */ 1159 (void) splitChunkAndReturnRemainder(ret, numWords); 1160 /* Label this as no longer a free chunk. */ 1161 assert(ret->isFree(), "This chunk should be free"); 1162 ret->linkPrev(NULL); 1163 } 1164 assert(ret == NULL || ret->isFree(), "Should be returning a free chunk"); 1165 return ret; 1166 } 1167 ShouldNotReachHere(); 1168 } 1169 1170 bool CompactibleFreeListSpace::verifyChunkInIndexedFreeLists(FreeChunk* fc) 1171 const { 1172 assert(fc->size() < IndexSetSize, "Size of chunk is too large"); 1173 return _indexedFreeList[fc->size()].verifyChunkInFreeLists(fc); 1174 } 1175 1176 bool CompactibleFreeListSpace::verifyChunkInFreeLists(FreeChunk* fc) const { 1177 if (fc->size() >= IndexSetSize) { 1178 return dictionary()->verifyChunkInFreeLists(fc); 1179 } else { 1180 return verifyChunkInIndexedFreeLists(fc); 1181 } 1182 } 1183 1184 #ifndef PRODUCT 1185 void CompactibleFreeListSpace::assert_locked() const { 1186 CMSLockVerifier::assert_locked(freelistLock(), parDictionaryAllocLock()); 1187 } 1188 #endif 1189 1190 FreeChunk* CompactibleFreeListSpace::allocateScratch(size_t size) { 1191 // In the parallel case, the main thread holds the free list lock 1192 // on behalf the parallel threads. 1193 assert_locked(); 1194 FreeChunk* fc; 1195 { 1196 // If GC is parallel, this might be called by several threads. 1197 // This should be rare enough that the locking overhead won't affect 1198 // the sequential code. 1199 MutexLockerEx x(parDictionaryAllocLock(), 1200 Mutex::_no_safepoint_check_flag); 1201 fc = getChunkFromDictionary(size); 1202 } 1203 if (fc != NULL) { 1204 fc->dontCoalesce(); 1205 assert(fc->isFree(), "Should be free, but not coalescable"); 1206 // Verify that the block offset table shows this to 1207 // be a single block, but not one which is unallocated. 1208 _bt.verify_single_block((HeapWord*)fc, fc->size()); 1209 _bt.verify_not_unallocated((HeapWord*)fc, fc->size()); 1210 } 1211 return fc; 1212 } 1213 1214 oop CompactibleFreeListSpace::promote(oop obj, size_t obj_size, oop* ref) { 1215 assert(obj_size == (size_t)obj->size(), "bad obj_size passed in"); 1216 assert_locked(); 1217 1218 // if we are tracking promotions, then first ensure space for 1219 // promotion (including spooling space for saving header if necessary). 1220 // then allocate and copy, then track promoted info if needed. 1221 // When tracking (see PromotionInfo::track()), the mark word may 1222 // be displaced and in this case restoration of the mark word 1223 // occurs in the (oop_since_save_marks_)iterate phase. 1224 if (_promoInfo.tracking() && !_promoInfo.ensure_spooling_space()) { 1225 return NULL; 1226 } 1227 // Call the allocate(size_t, bool) form directly to avoid the 1228 // additional call through the allocate(size_t) form. Having 1229 // the compile inline the call is problematic because allocate(size_t) 1230 // is a virtual method. 1231 HeapWord* res = allocate(adjustObjectSize(obj_size)); 1232 if (res != NULL) { 1233 Copy::aligned_disjoint_words((HeapWord*)obj, res, obj_size); 1234 // if we should be tracking promotions, do so. 1235 if (_promoInfo.tracking()) { 1236 _promoInfo.track((PromotedObject*)res); 1237 } 1238 } 1239 return oop(res); 1240 } 1241 1242 HeapWord* 1243 CompactibleFreeListSpace::getChunkFromSmallLinearAllocBlock(size_t size) { 1244 assert_locked(); 1245 assert(size >= MinChunkSize, "minimum chunk size"); 1246 assert(size < _smallLinearAllocBlock._allocation_size_limit, 1247 "maximum from smallLinearAllocBlock"); 1248 return getChunkFromLinearAllocBlock(&_smallLinearAllocBlock, size); 1249 } 1250 1251 HeapWord* 1252 CompactibleFreeListSpace::getChunkFromLinearAllocBlock(LinearAllocBlock *blk, 1253 size_t size) { 1254 assert_locked(); 1255 assert(size >= MinChunkSize, "too small"); 1256 HeapWord* res = NULL; 1257 // Try to do linear allocation from blk, making sure that 1258 if (blk->_word_size == 0) { 1259 // We have probably been unable to fill this either in the prologue or 1260 // when it was exhausted at the last linear allocation. Bail out until 1261 // next time. 1262 assert(blk->_ptr == NULL, "consistency check"); 1263 return NULL; 1264 } 1265 assert(blk->_word_size != 0 && blk->_ptr != NULL, "consistency check"); 1266 res = getChunkFromLinearAllocBlockRemainder(blk, size); 1267 if (res != NULL) return res; 1268 1269 // about to exhaust this linear allocation block 1270 if (blk->_word_size == size) { // exactly satisfied 1271 res = blk->_ptr; 1272 _bt.allocated(res, blk->_word_size); 1273 } else if (size + MinChunkSize <= blk->_refillSize) { 1274 // Update _unallocated_block if the size is such that chunk would be 1275 // returned to the indexed free list. All other chunks in the indexed 1276 // free lists are allocated from the dictionary so that _unallocated_block 1277 // has already been adjusted for them. Do it here so that the cost 1278 // for all chunks added back to the indexed free lists. 1279 if (blk->_word_size < SmallForDictionary) { 1280 _bt.allocated(blk->_ptr, blk->_word_size); 1281 } 1282 // Return the chunk that isn't big enough, and then refill below. 1283 addChunkToFreeLists(blk->_ptr, blk->_word_size); 1284 _bt.verify_single_block(blk->_ptr, (blk->_ptr + blk->_word_size)); 1285 // Don't keep statistics on adding back chunk from a LinAB. 1286 } else { 1287 // A refilled block would not satisfy the request. 1288 return NULL; 1289 } 1290 1291 blk->_ptr = NULL; blk->_word_size = 0; 1292 refillLinearAllocBlock(blk); 1293 assert(blk->_ptr == NULL || blk->_word_size >= size + MinChunkSize, 1294 "block was replenished"); 1295 if (res != NULL) { 1296 splitBirth(size); 1297 repairLinearAllocBlock(blk); 1298 } else if (blk->_ptr != NULL) { 1299 res = blk->_ptr; 1300 size_t blk_size = blk->_word_size; 1301 blk->_word_size -= size; 1302 blk->_ptr += size; 1303 splitBirth(size); 1304 repairLinearAllocBlock(blk); 1305 // Update BOT last so that other (parallel) GC threads see a consistent 1306 // view of the BOT and free blocks. 1307 // Above must occur before BOT is updated below. 1308 _bt.split_block(res, blk_size, size); // adjust block offset table 1309 } 1310 return res; 1311 } 1312 1313 HeapWord* CompactibleFreeListSpace::getChunkFromLinearAllocBlockRemainder( 1314 LinearAllocBlock* blk, 1315 size_t size) { 1316 assert_locked(); 1317 assert(size >= MinChunkSize, "too small"); 1318 1319 HeapWord* res = NULL; 1320 // This is the common case. Keep it simple. 1321 if (blk->_word_size >= size + MinChunkSize) { 1322 assert(blk->_ptr != NULL, "consistency check"); 1323 res = blk->_ptr; 1324 // Note that the BOT is up-to-date for the linAB before allocation. It 1325 // indicates the start of the linAB. The split_block() updates the 1326 // BOT for the linAB after the allocation (indicates the start of the 1327 // next chunk to be allocated). 1328 size_t blk_size = blk->_word_size; 1329 blk->_word_size -= size; 1330 blk->_ptr += size; 1331 splitBirth(size); 1332 repairLinearAllocBlock(blk); 1333 // Update BOT last so that other (parallel) GC threads see a consistent 1334 // view of the BOT and free blocks. 1335 // Above must occur before BOT is updated below. 1336 _bt.split_block(res, blk_size, size); // adjust block offset table 1337 _bt.allocated(res, size); 1338 } 1339 return res; 1340 } 1341 1342 FreeChunk* 1343 CompactibleFreeListSpace::getChunkFromIndexedFreeList(size_t size) { 1344 assert_locked(); 1345 assert(size < SmallForDictionary, "just checking"); 1346 FreeChunk* res; 1347 res = _indexedFreeList[size].getChunkAtHead(); 1348 if (res == NULL) { 1349 res = getChunkFromIndexedFreeListHelper(size); 1350 } 1351 _bt.verify_not_unallocated((HeapWord*) res, size); 1352 return res; 1353 } 1354 1355 FreeChunk* 1356 CompactibleFreeListSpace::getChunkFromIndexedFreeListHelper(size_t size) { 1357 assert_locked(); 1358 FreeChunk* fc = NULL; 1359 if (size < SmallForDictionary) { 1360 assert(_indexedFreeList[size].head() == NULL || 1361 _indexedFreeList[size].surplus() <= 0, 1362 "List for this size should be empty or under populated"); 1363 // Try best fit in exact lists before replenishing the list 1364 if (!bestFitFirst() || (fc = bestFitSmall(size)) == NULL) { 1365 // Replenish list. 1366 // 1367 // Things tried that failed. 1368 // Tried allocating out of the two LinAB's first before 1369 // replenishing lists. 1370 // Tried small linAB of size 256 (size in indexed list) 1371 // and replenishing indexed lists from the small linAB. 1372 // 1373 FreeChunk* newFc = NULL; 1374 size_t replenish_size = CMSIndexedFreeListReplenish * size; 1375 if (replenish_size < SmallForDictionary) { 1376 // Do not replenish from an underpopulated size. 1377 if (_indexedFreeList[replenish_size].surplus() > 0 && 1378 _indexedFreeList[replenish_size].head() != NULL) { 1379 newFc = 1380 _indexedFreeList[replenish_size].getChunkAtHead(); 1381 } else { 1382 newFc = bestFitSmall(replenish_size); 1383 } 1384 } 1385 if (newFc != NULL) { 1386 splitDeath(replenish_size); 1387 } else if (replenish_size > size) { 1388 assert(CMSIndexedFreeListReplenish > 1, "ctl pt invariant"); 1389 newFc = 1390 getChunkFromIndexedFreeListHelper(replenish_size); 1391 } 1392 if (newFc != NULL) { 1393 assert(newFc->size() == replenish_size, "Got wrong size"); 1394 size_t i; 1395 FreeChunk *curFc, *nextFc; 1396 // carve up and link blocks 0, ..., CMSIndexedFreeListReplenish - 2 1397 // The last chunk is not added to the lists but is returned as the 1398 // free chunk. 1399 for (curFc = newFc, nextFc = (FreeChunk*)((HeapWord*)curFc + size), 1400 i = 0; 1401 i < (CMSIndexedFreeListReplenish - 1); 1402 curFc = nextFc, nextFc = (FreeChunk*)((HeapWord*)nextFc + size), 1403 i++) { 1404 curFc->setSize(size); 1405 // Don't record this as a return in order to try and 1406 // determine the "returns" from a GC. 1407 _bt.verify_not_unallocated((HeapWord*) fc, size); 1408 _indexedFreeList[size].returnChunkAtTail(curFc, false); 1409 _bt.mark_block((HeapWord*)curFc, size); 1410 splitBirth(size); 1411 // Don't record the initial population of the indexed list 1412 // as a split birth. 1413 } 1414 1415 // check that the arithmetic was OK above 1416 assert((HeapWord*)nextFc == (HeapWord*)newFc + replenish_size, 1417 "inconsistency in carving newFc"); 1418 curFc->setSize(size); 1419 _bt.mark_block((HeapWord*)curFc, size); 1420 splitBirth(size); 1421 return curFc; 1422 } 1423 } 1424 } else { 1425 // Get a free chunk from the free chunk dictionary to be returned to 1426 // replenish the indexed free list. 1427 fc = getChunkFromDictionaryExact(size); 1428 } 1429 assert(fc == NULL || fc->isFree(), "Should be returning a free chunk"); 1430 return fc; 1431 } 1432 1433 FreeChunk* 1434 CompactibleFreeListSpace::getChunkFromDictionary(size_t size) { 1435 assert_locked(); 1436 FreeChunk* fc = _dictionary->getChunk(size); 1437 if (fc == NULL) { 1438 return NULL; 1439 } 1440 _bt.allocated((HeapWord*)fc, fc->size()); 1441 if (fc->size() >= size + MinChunkSize) { 1442 fc = splitChunkAndReturnRemainder(fc, size); 1443 } 1444 assert(fc->size() >= size, "chunk too small"); 1445 assert(fc->size() < size + MinChunkSize, "chunk too big"); 1446 _bt.verify_single_block((HeapWord*)fc, fc->size()); 1447 return fc; 1448 } 1449 1450 FreeChunk* 1451 CompactibleFreeListSpace::getChunkFromDictionaryExact(size_t size) { 1452 assert_locked(); 1453 FreeChunk* fc = _dictionary->getChunk(size); 1454 if (fc == NULL) { 1455 return fc; 1456 } 1457 _bt.allocated((HeapWord*)fc, fc->size()); 1458 if (fc->size() == size) { 1459 _bt.verify_single_block((HeapWord*)fc, size); 1460 return fc; 1461 } 1462 assert(fc->size() > size, "getChunk() guarantee"); 1463 if (fc->size() < size + MinChunkSize) { 1464 // Return the chunk to the dictionary and go get a bigger one. 1465 returnChunkToDictionary(fc); 1466 fc = _dictionary->getChunk(size + MinChunkSize); 1467 if (fc == NULL) { 1468 return NULL; 1469 } 1470 _bt.allocated((HeapWord*)fc, fc->size()); 1471 } 1472 assert(fc->size() >= size + MinChunkSize, "tautology"); 1473 fc = splitChunkAndReturnRemainder(fc, size); 1474 assert(fc->size() == size, "chunk is wrong size"); 1475 _bt.verify_single_block((HeapWord*)fc, size); 1476 return fc; 1477 } 1478 1479 void 1480 CompactibleFreeListSpace::returnChunkToDictionary(FreeChunk* chunk) { 1481 assert_locked(); 1482 1483 size_t size = chunk->size(); 1484 _bt.verify_single_block((HeapWord*)chunk, size); 1485 // adjust _unallocated_block downward, as necessary 1486 _bt.freed((HeapWord*)chunk, size); 1487 _dictionary->returnChunk(chunk); 1488 } 1489 1490 void 1491 CompactibleFreeListSpace::returnChunkToFreeList(FreeChunk* fc) { 1492 assert_locked(); 1493 size_t size = fc->size(); 1494 _bt.verify_single_block((HeapWord*) fc, size); 1495 _bt.verify_not_unallocated((HeapWord*) fc, size); 1496 if (_adaptive_freelists) { 1497 _indexedFreeList[size].returnChunkAtTail(fc); 1498 } else { 1499 _indexedFreeList[size].returnChunkAtHead(fc); 1500 } 1501 } 1502 1503 // Add chunk to end of last block -- if it's the largest 1504 // block -- and update BOT and census data. We would 1505 // of course have preferred to coalesce it with the 1506 // last block, but it's currently less expensive to find the 1507 // largest block than it is to find the last. 1508 void 1509 CompactibleFreeListSpace::addChunkToFreeListsAtEndRecordingStats( 1510 HeapWord* chunk, size_t size) { 1511 // check that the chunk does lie in this space! 1512 assert(chunk != NULL && is_in_reserved(chunk), "Not in this space!"); 1513 assert_locked(); 1514 // One of the parallel gc task threads may be here 1515 // whilst others are allocating. 1516 Mutex* lock = NULL; 1517 if (ParallelGCThreads != 0) { 1518 lock = &_parDictionaryAllocLock; 1519 } 1520 FreeChunk* ec; 1521 { 1522 MutexLockerEx x(lock, Mutex::_no_safepoint_check_flag); 1523 ec = dictionary()->findLargestDict(); // get largest block 1524 if (ec != NULL && ec->end() == chunk) { 1525 // It's a coterminal block - we can coalesce. 1526 size_t old_size = ec->size(); 1527 coalDeath(old_size); 1528 removeChunkFromDictionary(ec); 1529 size += old_size; 1530 } else { 1531 ec = (FreeChunk*)chunk; 1532 } 1533 } 1534 ec->setSize(size); 1535 debug_only(ec->mangleFreed(size)); 1536 if (size < SmallForDictionary) { 1537 lock = _indexedFreeListParLocks[size]; 1538 } 1539 MutexLockerEx x(lock, Mutex::_no_safepoint_check_flag); 1540 addChunkAndRepairOffsetTable((HeapWord*)ec, size, true); 1541 // record the birth under the lock since the recording involves 1542 // manipulation of the list on which the chunk lives and 1543 // if the chunk is allocated and is the last on the list, 1544 // the list can go away. 1545 coalBirth(size); 1546 } 1547 1548 void 1549 CompactibleFreeListSpace::addChunkToFreeLists(HeapWord* chunk, 1550 size_t size) { 1551 // check that the chunk does lie in this space! 1552 assert(chunk != NULL && is_in_reserved(chunk), "Not in this space!"); 1553 assert_locked(); 1554 _bt.verify_single_block(chunk, size); 1555 1556 FreeChunk* fc = (FreeChunk*) chunk; 1557 fc->setSize(size); 1558 debug_only(fc->mangleFreed(size)); 1559 if (size < SmallForDictionary) { 1560 returnChunkToFreeList(fc); 1561 } else { 1562 returnChunkToDictionary(fc); 1563 } 1564 } 1565 1566 void 1567 CompactibleFreeListSpace::addChunkAndRepairOffsetTable(HeapWord* chunk, 1568 size_t size, bool coalesced) { 1569 assert_locked(); 1570 assert(chunk != NULL, "null chunk"); 1571 if (coalesced) { 1572 // repair BOT 1573 _bt.single_block(chunk, size); 1574 } 1575 addChunkToFreeLists(chunk, size); 1576 } 1577 1578 // We _must_ find the purported chunk on our free lists; 1579 // we assert if we don't. 1580 void 1581 CompactibleFreeListSpace::removeFreeChunkFromFreeLists(FreeChunk* fc) { 1582 size_t size = fc->size(); 1583 assert_locked(); 1584 debug_only(verifyFreeLists()); 1585 if (size < SmallForDictionary) { 1586 removeChunkFromIndexedFreeList(fc); 1587 } else { 1588 removeChunkFromDictionary(fc); 1589 } 1590 _bt.verify_single_block((HeapWord*)fc, size); 1591 debug_only(verifyFreeLists()); 1592 } 1593 1594 void 1595 CompactibleFreeListSpace::removeChunkFromDictionary(FreeChunk* fc) { 1596 size_t size = fc->size(); 1597 assert_locked(); 1598 assert(fc != NULL, "null chunk"); 1599 _bt.verify_single_block((HeapWord*)fc, size); 1600 _dictionary->removeChunk(fc); 1601 // adjust _unallocated_block upward, as necessary 1602 _bt.allocated((HeapWord*)fc, size); 1603 } 1604 1605 void 1606 CompactibleFreeListSpace::removeChunkFromIndexedFreeList(FreeChunk* fc) { 1607 assert_locked(); 1608 size_t size = fc->size(); 1609 _bt.verify_single_block((HeapWord*)fc, size); 1610 NOT_PRODUCT( 1611 if (FLSVerifyIndexTable) { 1612 verifyIndexedFreeList(size); 1613 } 1614 ) 1615 _indexedFreeList[size].removeChunk(fc); 1616 debug_only(fc->clearNext()); 1617 debug_only(fc->clearPrev()); 1618 NOT_PRODUCT( 1619 if (FLSVerifyIndexTable) { 1620 verifyIndexedFreeList(size); 1621 } 1622 ) 1623 } 1624 1625 FreeChunk* CompactibleFreeListSpace::bestFitSmall(size_t numWords) { 1626 /* A hint is the next larger size that has a surplus. 1627 Start search at a size large enough to guarantee that 1628 the excess is >= MIN_CHUNK. */ 1629 size_t start = align_object_size(numWords + MinChunkSize); 1630 if (start < IndexSetSize) { 1631 FreeList* it = _indexedFreeList; 1632 size_t hint = _indexedFreeList[start].hint(); 1633 while (hint < IndexSetSize) { 1634 assert(hint % MinObjAlignment == 0, "hint should be aligned"); 1635 FreeList *fl = &_indexedFreeList[hint]; 1636 if (fl->surplus() > 0 && fl->head() != NULL) { 1637 // Found a list with surplus, reset original hint 1638 // and split out a free chunk which is returned. 1639 _indexedFreeList[start].set_hint(hint); 1640 FreeChunk* res = getFromListGreater(fl, numWords); 1641 assert(res == NULL || res->isFree(), 1642 "Should be returning a free chunk"); 1643 return res; 1644 } 1645 hint = fl->hint(); /* keep looking */ 1646 } 1647 /* None found. */ 1648 it[start].set_hint(IndexSetSize); 1649 } 1650 return NULL; 1651 } 1652 1653 /* Requires fl->size >= numWords + MinChunkSize */ 1654 FreeChunk* CompactibleFreeListSpace::getFromListGreater(FreeList* fl, 1655 size_t numWords) { 1656 FreeChunk *curr = fl->head(); 1657 size_t oldNumWords = curr->size(); 1658 assert(numWords >= MinChunkSize, "Word size is too small"); 1659 assert(curr != NULL, "List is empty"); 1660 assert(oldNumWords >= numWords + MinChunkSize, 1661 "Size of chunks in the list is too small"); 1662 1663 fl->removeChunk(curr); 1664 // recorded indirectly by splitChunkAndReturnRemainder - 1665 // smallSplit(oldNumWords, numWords); 1666 FreeChunk* new_chunk = splitChunkAndReturnRemainder(curr, numWords); 1667 // Does anything have to be done for the remainder in terms of 1668 // fixing the card table? 1669 assert(new_chunk == NULL || new_chunk->isFree(), 1670 "Should be returning a free chunk"); 1671 return new_chunk; 1672 } 1673 1674 FreeChunk* 1675 CompactibleFreeListSpace::splitChunkAndReturnRemainder(FreeChunk* chunk, 1676 size_t new_size) { 1677 assert_locked(); 1678 size_t size = chunk->size(); 1679 assert(size > new_size, "Split from a smaller block?"); 1680 assert(is_aligned(chunk), "alignment problem"); 1681 assert(size == adjustObjectSize(size), "alignment problem"); 1682 size_t rem_size = size - new_size; 1683 assert(rem_size == adjustObjectSize(rem_size), "alignment problem"); 1684 assert(rem_size >= MinChunkSize, "Free chunk smaller than minimum"); 1685 FreeChunk* ffc = (FreeChunk*)((HeapWord*)chunk + new_size); 1686 assert(is_aligned(ffc), "alignment problem"); 1687 ffc->setSize(rem_size); 1688 ffc->linkNext(NULL); 1689 ffc->linkPrev(NULL); // Mark as a free block for other (parallel) GC threads. 1690 // Above must occur before BOT is updated below. 1691 // adjust block offset table 1692 _bt.split_block((HeapWord*)chunk, chunk->size(), new_size); 1693 if (rem_size < SmallForDictionary) { 1694 bool is_par = (SharedHeap::heap()->n_par_threads() > 0); 1695 if (is_par) _indexedFreeListParLocks[rem_size]->lock(); 1696 returnChunkToFreeList(ffc); 1697 split(size, rem_size); 1698 if (is_par) _indexedFreeListParLocks[rem_size]->unlock(); 1699 } else { 1700 returnChunkToDictionary(ffc); 1701 split(size ,rem_size); 1702 } 1703 chunk->setSize(new_size); 1704 return chunk; 1705 } 1706 1707 void 1708 CompactibleFreeListSpace::sweep_completed() { 1709 // Now that space is probably plentiful, refill linear 1710 // allocation blocks as needed. 1711 refillLinearAllocBlocksIfNeeded(); 1712 } 1713 1714 void 1715 CompactibleFreeListSpace::gc_prologue() { 1716 assert_locked(); 1717 if (PrintFLSStatistics != 0) { 1718 gclog_or_tty->print("Before GC:\n"); 1719 reportFreeListStatistics(); 1720 } 1721 refillLinearAllocBlocksIfNeeded(); 1722 } 1723 1724 void 1725 CompactibleFreeListSpace::gc_epilogue() { 1726 assert_locked(); 1727 if (PrintGCDetails && Verbose && !_adaptive_freelists) { 1728 if (_smallLinearAllocBlock._word_size == 0) 1729 warning("CompactibleFreeListSpace(epilogue):: Linear allocation failure"); 1730 } 1731 assert(_promoInfo.noPromotions(), "_promoInfo inconsistency"); 1732 _promoInfo.stopTrackingPromotions(); 1733 repairLinearAllocationBlocks(); 1734 // Print Space's stats 1735 if (PrintFLSStatistics != 0) { 1736 gclog_or_tty->print("After GC:\n"); 1737 reportFreeListStatistics(); 1738 } 1739 } 1740 1741 // Iteration support, mostly delegated from a CMS generation 1742 1743 void CompactibleFreeListSpace::save_marks() { 1744 // mark the "end" of the used space at the time of this call; 1745 // note, however, that promoted objects from this point 1746 // on are tracked in the _promoInfo below. 1747 set_saved_mark_word(BlockOffsetArrayUseUnallocatedBlock ? 1748 unallocated_block() : end()); 1749 // inform allocator that promotions should be tracked. 1750 assert(_promoInfo.noPromotions(), "_promoInfo inconsistency"); 1751 _promoInfo.startTrackingPromotions(); 1752 } 1753 1754 bool CompactibleFreeListSpace::no_allocs_since_save_marks() { 1755 assert(_promoInfo.tracking(), "No preceding save_marks?"); 1756 guarantee(SharedHeap::heap()->n_par_threads() == 0, 1757 "Shouldn't be called (yet) during parallel part of gc."); 1758 return _promoInfo.noPromotions(); 1759 } 1760 1761 #define CFLS_OOP_SINCE_SAVE_MARKS_DEFN(OopClosureType, nv_suffix) \ 1762 \ 1763 void CompactibleFreeListSpace:: \ 1764 oop_since_save_marks_iterate##nv_suffix(OopClosureType* blk) { \ 1765 assert(SharedHeap::heap()->n_par_threads() == 0, \ 1766 "Shouldn't be called (yet) during parallel part of gc."); \ 1767 _promoInfo.promoted_oops_iterate##nv_suffix(blk); \ 1768 /* \ 1769 * This also restores any displaced headers and removes the elements from \ 1770 * the iteration set as they are processed, so that we have a clean slate \ 1771 * at the end of the iteration. Note, thus, that if new objects are \ 1772 * promoted as a result of the iteration they are iterated over as well. \ 1773 */ \ 1774 assert(_promoInfo.noPromotions(), "_promoInfo inconsistency"); \ 1775 } 1776 1777 ALL_SINCE_SAVE_MARKS_CLOSURES(CFLS_OOP_SINCE_SAVE_MARKS_DEFN) 1778 1779 ////////////////////////////////////////////////////////////////////////////// 1780 // We go over the list of promoted objects, removing each from the list, 1781 // and applying the closure (this may, in turn, add more elements to 1782 // the tail of the promoted list, and these newly added objects will 1783 // also be processed) until the list is empty. 1784 // To aid verification and debugging, in the non-product builds 1785 // we actually forward _promoHead each time we process a promoted oop. 1786 // Note that this is not necessary in general (i.e. when we don't need to 1787 // call PromotionInfo::verify()) because oop_iterate can only add to the 1788 // end of _promoTail, and never needs to look at _promoHead. 1789 1790 #define PROMOTED_OOPS_ITERATE_DEFN(OopClosureType, nv_suffix) \ 1791 \ 1792 void PromotionInfo::promoted_oops_iterate##nv_suffix(OopClosureType* cl) { \ 1793 NOT_PRODUCT(verify()); \ 1794 PromotedObject *curObj, *nextObj; \ 1795 for (curObj = _promoHead; curObj != NULL; curObj = nextObj) { \ 1796 if ((nextObj = curObj->next()) == NULL) { \ 1797 /* protect ourselves against additions due to closure application \ 1798 below by resetting the list. */ \ 1799 assert(_promoTail == curObj, "Should have been the tail"); \ 1800 _promoHead = _promoTail = NULL; \ 1801 } \ 1802 if (curObj->hasDisplacedMark()) { \ 1803 /* restore displaced header */ \ 1804 oop(curObj)->set_mark(nextDisplacedHeader()); \ 1805 } else { \ 1806 /* restore prototypical header */ \ 1807 oop(curObj)->init_mark(); \ 1808 } \ 1809 /* The "promoted_mark" should now not be set */ \ 1810 assert(!curObj->hasPromotedMark(), \ 1811 "Should have been cleared by restoring displaced mark-word"); \ 1812 NOT_PRODUCT(_promoHead = nextObj); \ 1813 if (cl != NULL) oop(curObj)->oop_iterate(cl); \ 1814 if (nextObj == NULL) { /* start at head of list reset above */ \ 1815 nextObj = _promoHead; \ 1816 } \ 1817 } \ 1818 assert(noPromotions(), "post-condition violation"); \ 1819 assert(_promoHead == NULL && _promoTail == NULL, "emptied promoted list");\ 1820 assert(_spoolHead == _spoolTail, "emptied spooling buffers"); \ 1821 assert(_firstIndex == _nextIndex, "empty buffer"); \ 1822 } 1823 1824 // This should have been ALL_SINCE_...() just like the others, 1825 // but, because the body of the method above is somehwat longer, 1826 // the MSVC compiler cannot cope; as a workaround, we split the 1827 // macro into its 3 constituent parts below (see original macro 1828 // definition in specializedOopClosures.hpp). 1829 SPECIALIZED_SINCE_SAVE_MARKS_CLOSURES_YOUNG(PROMOTED_OOPS_ITERATE_DEFN) 1830 PROMOTED_OOPS_ITERATE_DEFN(OopsInGenClosure,_v) 1831 1832 1833 void CompactibleFreeListSpace::object_iterate_since_last_GC(ObjectClosure* cl) { 1834 // ugghh... how would one do this efficiently for a non-contiguous space? 1835 guarantee(false, "NYI"); 1836 } 1837 1838 bool CompactibleFreeListSpace::linearAllocationWouldFail() const { 1839 return _smallLinearAllocBlock._word_size == 0; 1840 } 1841 1842 void CompactibleFreeListSpace::repairLinearAllocationBlocks() { 1843 // Fix up linear allocation blocks to look like free blocks 1844 repairLinearAllocBlock(&_smallLinearAllocBlock); 1845 } 1846 1847 void CompactibleFreeListSpace::repairLinearAllocBlock(LinearAllocBlock* blk) { 1848 assert_locked(); 1849 if (blk->_ptr != NULL) { 1850 assert(blk->_word_size != 0 && blk->_word_size >= MinChunkSize, 1851 "Minimum block size requirement"); 1852 FreeChunk* fc = (FreeChunk*)(blk->_ptr); 1853 fc->setSize(blk->_word_size); 1854 fc->linkPrev(NULL); // mark as free 1855 fc->dontCoalesce(); 1856 assert(fc->isFree(), "just marked it free"); 1857 assert(fc->cantCoalesce(), "just marked it uncoalescable"); 1858 } 1859 } 1860 1861 void CompactibleFreeListSpace::refillLinearAllocBlocksIfNeeded() { 1862 assert_locked(); 1863 if (_smallLinearAllocBlock._ptr == NULL) { 1864 assert(_smallLinearAllocBlock._word_size == 0, 1865 "Size of linAB should be zero if the ptr is NULL"); 1866 // Reset the linAB refill and allocation size limit. 1867 _smallLinearAllocBlock.set(0, 0, 1024*SmallForLinearAlloc, SmallForLinearAlloc); 1868 } 1869 refillLinearAllocBlockIfNeeded(&_smallLinearAllocBlock); 1870 } 1871 1872 void 1873 CompactibleFreeListSpace::refillLinearAllocBlockIfNeeded(LinearAllocBlock* blk) { 1874 assert_locked(); 1875 assert((blk->_ptr == NULL && blk->_word_size == 0) || 1876 (blk->_ptr != NULL && blk->_word_size >= MinChunkSize), 1877 "blk invariant"); 1878 if (blk->_ptr == NULL) { 1879 refillLinearAllocBlock(blk); 1880 } 1881 if (PrintMiscellaneous && Verbose) { 1882 if (blk->_word_size == 0) { 1883 warning("CompactibleFreeListSpace(prologue):: Linear allocation failure"); 1884 } 1885 } 1886 } 1887 1888 void 1889 CompactibleFreeListSpace::refillLinearAllocBlock(LinearAllocBlock* blk) { 1890 assert_locked(); 1891 assert(blk->_word_size == 0 && blk->_ptr == NULL, 1892 "linear allocation block should be empty"); 1893 FreeChunk* fc; 1894 if (blk->_refillSize < SmallForDictionary && 1895 (fc = getChunkFromIndexedFreeList(blk->_refillSize)) != NULL) { 1896 // A linAB's strategy might be to use small sizes to reduce 1897 // fragmentation but still get the benefits of allocation from a 1898 // linAB. 1899 } else { 1900 fc = getChunkFromDictionary(blk->_refillSize); 1901 } 1902 if (fc != NULL) { 1903 blk->_ptr = (HeapWord*)fc; 1904 blk->_word_size = fc->size(); 1905 fc->dontCoalesce(); // to prevent sweeper from sweeping us up 1906 } 1907 } 1908 1909 // Support for concurrent collection policy decisions. 1910 bool CompactibleFreeListSpace::should_concurrent_collect() const { 1911 // In the future we might want to add in frgamentation stats -- 1912 // including erosion of the "mountain" into this decision as well. 1913 return !adaptive_freelists() && linearAllocationWouldFail(); 1914 } 1915 1916 // Support for compaction 1917 1918 void CompactibleFreeListSpace::prepare_for_compaction(CompactPoint* cp) { 1919 SCAN_AND_FORWARD(cp,end,block_is_obj,block_size); 1920 // prepare_for_compaction() uses the space between live objects 1921 // so that later phase can skip dead space quickly. So verification 1922 // of the free lists doesn't work after. 1923 } 1924 1925 #define obj_size(q) adjustObjectSize(oop(q)->size()) 1926 #define adjust_obj_size(s) adjustObjectSize(s) 1927 1928 void CompactibleFreeListSpace::adjust_pointers() { 1929 // In other versions of adjust_pointers(), a bail out 1930 // based on the amount of live data in the generation 1931 // (i.e., if 0, bail out) may be used. 1932 // Cannot test used() == 0 here because the free lists have already 1933 // been mangled by the compaction. 1934 1935 SCAN_AND_ADJUST_POINTERS(adjust_obj_size); 1936 // See note about verification in prepare_for_compaction(). 1937 } 1938 1939 void CompactibleFreeListSpace::compact() { 1940 SCAN_AND_COMPACT(obj_size); 1941 } 1942 1943 // fragmentation_metric = 1 - [sum of (fbs**2) / (sum of fbs)**2] 1944 // where fbs is free block sizes 1945 double CompactibleFreeListSpace::flsFrag() const { 1946 size_t itabFree = totalSizeInIndexedFreeLists(); 1947 double frag = 0.0; 1948 size_t i; 1949 1950 for (i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) { 1951 double sz = i; 1952 frag += _indexedFreeList[i].count() * (sz * sz); 1953 } 1954 1955 double totFree = itabFree + 1956 _dictionary->totalChunkSize(DEBUG_ONLY(freelistLock())); 1957 if (totFree > 0) { 1958 frag = ((frag + _dictionary->sum_of_squared_block_sizes()) / 1959 (totFree * totFree)); 1960 frag = (double)1.0 - frag; 1961 } else { 1962 assert(frag == 0.0, "Follows from totFree == 0"); 1963 } 1964 return frag; 1965 } 1966 1967 #define CoalSurplusPercent 1.05 1968 #define SplitSurplusPercent 1.10 1969 1970 void CompactibleFreeListSpace::beginSweepFLCensus( 1971 float inter_sweep_current, 1972 float inter_sweep_estimate) { 1973 assert_locked(); 1974 size_t i; 1975 for (i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) { 1976 FreeList* fl = &_indexedFreeList[i]; 1977 fl->compute_desired(inter_sweep_current, inter_sweep_estimate); 1978 fl->set_coalDesired((ssize_t)((double)fl->desired() * CoalSurplusPercent)); 1979 fl->set_beforeSweep(fl->count()); 1980 fl->set_bfrSurp(fl->surplus()); 1981 } 1982 _dictionary->beginSweepDictCensus(CoalSurplusPercent, 1983 inter_sweep_current, 1984 inter_sweep_estimate); 1985 } 1986 1987 void CompactibleFreeListSpace::setFLSurplus() { 1988 assert_locked(); 1989 size_t i; 1990 for (i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) { 1991 FreeList *fl = &_indexedFreeList[i]; 1992 fl->set_surplus(fl->count() - 1993 (ssize_t)((double)fl->desired() * SplitSurplusPercent)); 1994 } 1995 } 1996 1997 void CompactibleFreeListSpace::setFLHints() { 1998 assert_locked(); 1999 size_t i; 2000 size_t h = IndexSetSize; 2001 for (i = IndexSetSize - 1; i != 0; i -= IndexSetStride) { 2002 FreeList *fl = &_indexedFreeList[i]; 2003 fl->set_hint(h); 2004 if (fl->surplus() > 0) { 2005 h = i; 2006 } 2007 } 2008 } 2009 2010 void CompactibleFreeListSpace::clearFLCensus() { 2011 assert_locked(); 2012 int i; 2013 for (i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) { 2014 FreeList *fl = &_indexedFreeList[i]; 2015 fl->set_prevSweep(fl->count()); 2016 fl->set_coalBirths(0); 2017 fl->set_coalDeaths(0); 2018 fl->set_splitBirths(0); 2019 fl->set_splitDeaths(0); 2020 } 2021 } 2022 2023 void CompactibleFreeListSpace::endSweepFLCensus(size_t sweep_count) { 2024 setFLSurplus(); 2025 setFLHints(); 2026 if (PrintGC && PrintFLSCensus > 0) { 2027 printFLCensus(sweep_count); 2028 } 2029 clearFLCensus(); 2030 assert_locked(); 2031 _dictionary->endSweepDictCensus(SplitSurplusPercent); 2032 } 2033 2034 bool CompactibleFreeListSpace::coalOverPopulated(size_t size) { 2035 if (size < SmallForDictionary) { 2036 FreeList *fl = &_indexedFreeList[size]; 2037 return (fl->coalDesired() < 0) || 2038 ((int)fl->count() > fl->coalDesired()); 2039 } else { 2040 return dictionary()->coalDictOverPopulated(size); 2041 } 2042 } 2043 2044 void CompactibleFreeListSpace::smallCoalBirth(size_t size) { 2045 assert(size < SmallForDictionary, "Size too large for indexed list"); 2046 FreeList *fl = &_indexedFreeList[size]; 2047 fl->increment_coalBirths(); 2048 fl->increment_surplus(); 2049 } 2050 2051 void CompactibleFreeListSpace::smallCoalDeath(size_t size) { 2052 assert(size < SmallForDictionary, "Size too large for indexed list"); 2053 FreeList *fl = &_indexedFreeList[size]; 2054 fl->increment_coalDeaths(); 2055 fl->decrement_surplus(); 2056 } 2057 2058 void CompactibleFreeListSpace::coalBirth(size_t size) { 2059 if (size < SmallForDictionary) { 2060 smallCoalBirth(size); 2061 } else { 2062 dictionary()->dictCensusUpdate(size, 2063 false /* split */, 2064 true /* birth */); 2065 } 2066 } 2067 2068 void CompactibleFreeListSpace::coalDeath(size_t size) { 2069 if(size < SmallForDictionary) { 2070 smallCoalDeath(size); 2071 } else { 2072 dictionary()->dictCensusUpdate(size, 2073 false /* split */, 2074 false /* birth */); 2075 } 2076 } 2077 2078 void CompactibleFreeListSpace::smallSplitBirth(size_t size) { 2079 assert(size < SmallForDictionary, "Size too large for indexed list"); 2080 FreeList *fl = &_indexedFreeList[size]; 2081 fl->increment_splitBirths(); 2082 fl->increment_surplus(); 2083 } 2084 2085 void CompactibleFreeListSpace::smallSplitDeath(size_t size) { 2086 assert(size < SmallForDictionary, "Size too large for indexed list"); 2087 FreeList *fl = &_indexedFreeList[size]; 2088 fl->increment_splitDeaths(); 2089 fl->decrement_surplus(); 2090 } 2091 2092 void CompactibleFreeListSpace::splitBirth(size_t size) { 2093 if (size < SmallForDictionary) { 2094 smallSplitBirth(size); 2095 } else { 2096 dictionary()->dictCensusUpdate(size, 2097 true /* split */, 2098 true /* birth */); 2099 } 2100 } 2101 2102 void CompactibleFreeListSpace::splitDeath(size_t size) { 2103 if (size < SmallForDictionary) { 2104 smallSplitDeath(size); 2105 } else { 2106 dictionary()->dictCensusUpdate(size, 2107 true /* split */, 2108 false /* birth */); 2109 } 2110 } 2111 2112 void CompactibleFreeListSpace::split(size_t from, size_t to1) { 2113 size_t to2 = from - to1; 2114 splitDeath(from); 2115 splitBirth(to1); 2116 splitBirth(to2); 2117 } 2118 2119 2120 void CompactibleFreeListSpace::print() const { 2121 tty->print(" CompactibleFreeListSpace"); 2122 Space::print(); 2123 } 2124 2125 void CompactibleFreeListSpace::prepare_for_verify() { 2126 assert_locked(); 2127 repairLinearAllocationBlocks(); 2128 // Verify that the SpoolBlocks look like free blocks of 2129 // appropriate sizes... To be done ... 2130 } 2131 2132 class VerifyAllBlksClosure: public BlkClosure { 2133 const CompactibleFreeListSpace* _sp; 2134 const MemRegion _span; 2135 2136 public: 2137 VerifyAllBlksClosure(const CompactibleFreeListSpace* sp, 2138 MemRegion span) : _sp(sp), _span(span) { } 2139 2140 size_t do_blk(HeapWord* addr) { 2141 size_t res; 2142 if (_sp->block_is_obj(addr)) { 2143 oop p = oop(addr); 2144 guarantee(p->is_oop(), "Should be an oop"); 2145 res = _sp->adjustObjectSize(p->size()); 2146 if (_sp->obj_is_alive(addr)) { 2147 p->verify(); 2148 } 2149 } else { 2150 FreeChunk* fc = (FreeChunk*)addr; 2151 res = fc->size(); 2152 if (FLSVerifyLists && !fc->cantCoalesce()) { 2153 guarantee(_sp->verifyChunkInFreeLists(fc), 2154 "Chunk should be on a free list"); 2155 } 2156 } 2157 guarantee(res != 0, "Livelock: no rank reduction!"); 2158 return res; 2159 } 2160 }; 2161 2162 class VerifyAllOopsClosure: public OopClosure { 2163 const CMSCollector* _collector; 2164 const CompactibleFreeListSpace* _sp; 2165 const MemRegion _span; 2166 const bool _past_remark; 2167 const CMSBitMap* _bit_map; 2168 2169 public: 2170 VerifyAllOopsClosure(const CMSCollector* collector, 2171 const CompactibleFreeListSpace* sp, MemRegion span, 2172 bool past_remark, CMSBitMap* bit_map) : 2173 OopClosure(), _collector(collector), _sp(sp), _span(span), 2174 _past_remark(past_remark), _bit_map(bit_map) { } 2175 2176 void do_oop(oop* ptr) { 2177 oop p = *ptr; 2178 if (p != NULL) { 2179 if (_span.contains(p)) { // the interior oop points into CMS heap 2180 if (!_span.contains(ptr)) { // reference from outside CMS heap 2181 // Should be a valid object; the first disjunct below allows 2182 // us to sidestep an assertion in block_is_obj() that insists 2183 // that p be in _sp. Note that several generations (and spaces) 2184 // are spanned by _span (CMS heap) above. 2185 guarantee(!_sp->is_in_reserved(p) || _sp->block_is_obj((HeapWord*)p), 2186 "Should be an object"); 2187 guarantee(p->is_oop(), "Should be an oop"); 2188 p->verify(); 2189 if (_past_remark) { 2190 // Remark has been completed, the object should be marked 2191 _bit_map->isMarked((HeapWord*)p); 2192 } 2193 } 2194 else { // reference within CMS heap 2195 if (_past_remark) { 2196 // Remark has been completed -- so the referent should have 2197 // been marked, if referring object is. 2198 if (_bit_map->isMarked(_collector->block_start(ptr))) { 2199 guarantee(_bit_map->isMarked((HeapWord*)p), "Marking error?"); 2200 } 2201 } 2202 } 2203 } else if (_sp->is_in_reserved(ptr)) { 2204 // the reference is from FLS, and points out of FLS 2205 guarantee(p->is_oop(), "Should be an oop"); 2206 p->verify(); 2207 } 2208 } 2209 } 2210 }; 2211 2212 void CompactibleFreeListSpace::verify(bool ignored) const { 2213 assert_lock_strong(&_freelistLock); 2214 verify_objects_initialized(); 2215 MemRegion span = _collector->_span; 2216 bool past_remark = (_collector->abstract_state() == 2217 CMSCollector::Sweeping); 2218 2219 ResourceMark rm; 2220 HandleMark hm; 2221 2222 // Check integrity of CFL data structures 2223 _promoInfo.verify(); 2224 _dictionary->verify(); 2225 if (FLSVerifyIndexTable) { 2226 verifyIndexedFreeLists(); 2227 } 2228 // Check integrity of all objects and free blocks in space 2229 { 2230 VerifyAllBlksClosure cl(this, span); 2231 ((CompactibleFreeListSpace*)this)->blk_iterate(&cl); // cast off const 2232 } 2233 // Check that all references in the heap to FLS 2234 // are to valid objects in FLS or that references in 2235 // FLS are to valid objects elsewhere in the heap 2236 if (FLSVerifyAllHeapReferences) 2237 { 2238 VerifyAllOopsClosure cl(_collector, this, span, past_remark, 2239 _collector->markBitMap()); 2240 CollectedHeap* ch = Universe::heap(); 2241 ch->oop_iterate(&cl); // all oops in generations 2242 ch->permanent_oop_iterate(&cl); // all oops in perm gen 2243 } 2244 2245 if (VerifyObjectStartArray) { 2246 // Verify the block offset table 2247 _bt.verify(); 2248 } 2249 } 2250 2251 #ifndef PRODUCT 2252 void CompactibleFreeListSpace::verifyFreeLists() const { 2253 if (FLSVerifyLists) { 2254 _dictionary->verify(); 2255 verifyIndexedFreeLists(); 2256 } else { 2257 if (FLSVerifyDictionary) { 2258 _dictionary->verify(); 2259 } 2260 if (FLSVerifyIndexTable) { 2261 verifyIndexedFreeLists(); 2262 } 2263 } 2264 } 2265 #endif 2266 2267 void CompactibleFreeListSpace::verifyIndexedFreeLists() const { 2268 size_t i = 0; 2269 for (; i < MinChunkSize; i++) { 2270 guarantee(_indexedFreeList[i].head() == NULL, "should be NULL"); 2271 } 2272 for (; i < IndexSetSize; i++) { 2273 verifyIndexedFreeList(i); 2274 } 2275 } 2276 2277 void CompactibleFreeListSpace::verifyIndexedFreeList(size_t size) const { 2278 guarantee(size % 2 == 0, "Odd slots should be empty"); 2279 for (FreeChunk* fc = _indexedFreeList[size].head(); fc != NULL; 2280 fc = fc->next()) { 2281 guarantee(fc->size() == size, "Size inconsistency"); 2282 guarantee(fc->isFree(), "!free?"); 2283 guarantee(fc->next() == NULL || fc->next()->prev() == fc, "Broken list"); 2284 } 2285 } 2286 2287 #ifndef PRODUCT 2288 void CompactibleFreeListSpace::checkFreeListConsistency() const { 2289 assert(_dictionary->minSize() <= IndexSetSize, 2290 "Some sizes can't be allocated without recourse to" 2291 " linear allocation buffers"); 2292 assert(MIN_TREE_CHUNK_SIZE*HeapWordSize == sizeof(TreeChunk), 2293 "else MIN_TREE_CHUNK_SIZE is wrong"); 2294 assert((IndexSetStride == 2 && IndexSetStart == 2) || 2295 (IndexSetStride == 1 && IndexSetStart == 1), "just checking"); 2296 assert((IndexSetStride != 2) || (MinChunkSize % 2 == 0), 2297 "Some for-loops may be incorrectly initialized"); 2298 assert((IndexSetStride != 2) || (IndexSetSize % 2 == 1), 2299 "For-loops that iterate over IndexSet with stride 2 may be wrong"); 2300 } 2301 #endif 2302 2303 void CompactibleFreeListSpace::printFLCensus(size_t sweep_count) const { 2304 assert_lock_strong(&_freelistLock); 2305 FreeList total; 2306 gclog_or_tty->print("end sweep# " SIZE_FORMAT "\n", sweep_count); 2307 FreeList::print_labels_on(gclog_or_tty, "size"); 2308 size_t totalFree = 0; 2309 for (size_t i = IndexSetStart; i < IndexSetSize; i += IndexSetStride) { 2310 const FreeList *fl = &_indexedFreeList[i]; 2311 totalFree += fl->count() * fl->size(); 2312 if (i % (40*IndexSetStride) == 0) { 2313 FreeList::print_labels_on(gclog_or_tty, "size"); 2314 } 2315 fl->print_on(gclog_or_tty); 2316 total.set_bfrSurp( total.bfrSurp() + fl->bfrSurp() ); 2317 total.set_surplus( total.surplus() + fl->surplus() ); 2318 total.set_desired( total.desired() + fl->desired() ); 2319 total.set_prevSweep( total.prevSweep() + fl->prevSweep() ); 2320 total.set_beforeSweep(total.beforeSweep() + fl->beforeSweep()); 2321 total.set_count( total.count() + fl->count() ); 2322 total.set_coalBirths( total.coalBirths() + fl->coalBirths() ); 2323 total.set_coalDeaths( total.coalDeaths() + fl->coalDeaths() ); 2324 total.set_splitBirths(total.splitBirths() + fl->splitBirths()); 2325 total.set_splitDeaths(total.splitDeaths() + fl->splitDeaths()); 2326 } 2327 total.print_on(gclog_or_tty, "TOTAL"); 2328 gclog_or_tty->print_cr("Total free in indexed lists " 2329 SIZE_FORMAT " words", totalFree); 2330 gclog_or_tty->print("growth: %8.5f deficit: %8.5f\n", 2331 (double)(total.splitBirths()+total.coalBirths()-total.splitDeaths()-total.coalDeaths())/ 2332 (total.prevSweep() != 0 ? (double)total.prevSweep() : 1.0), 2333 (double)(total.desired() - total.count())/(total.desired() != 0 ? (double)total.desired() : 1.0)); 2334 _dictionary->printDictCensus(); 2335 } 2336 2337 // Return the next displaced header, incrementing the pointer and 2338 // recycling spool area as necessary. 2339 markOop PromotionInfo::nextDisplacedHeader() { 2340 assert(_spoolHead != NULL, "promotionInfo inconsistency"); 2341 assert(_spoolHead != _spoolTail || _firstIndex < _nextIndex, 2342 "Empty spool space: no displaced header can be fetched"); 2343 assert(_spoolHead->bufferSize > _firstIndex, "Off by one error at head?"); 2344 markOop hdr = _spoolHead->displacedHdr[_firstIndex]; 2345 // Spool forward 2346 if (++_firstIndex == _spoolHead->bufferSize) { // last location in this block 2347 // forward to next block, recycling this block into spare spool buffer 2348 SpoolBlock* tmp = _spoolHead->nextSpoolBlock; 2349 assert(_spoolHead != _spoolTail, "Spooling storage mix-up"); 2350 _spoolHead->nextSpoolBlock = _spareSpool; 2351 _spareSpool = _spoolHead; 2352 _spoolHead = tmp; 2353 _firstIndex = 1; 2354 NOT_PRODUCT( 2355 if (_spoolHead == NULL) { // all buffers fully consumed 2356 assert(_spoolTail == NULL && _nextIndex == 1, 2357 "spool buffers processing inconsistency"); 2358 } 2359 ) 2360 } 2361 return hdr; 2362 } 2363 2364 void PromotionInfo::track(PromotedObject* trackOop) { 2365 track(trackOop, oop(trackOop)->klass()); 2366 } 2367 2368 void PromotionInfo::track(PromotedObject* trackOop, klassOop klassOfOop) { 2369 // make a copy of header as it may need to be spooled 2370 markOop mark = oop(trackOop)->mark(); 2371 trackOop->clearNext(); 2372 if (mark->must_be_preserved_for_cms_scavenge(klassOfOop)) { 2373 // save non-prototypical header, and mark oop 2374 saveDisplacedHeader(mark); 2375 trackOop->setDisplacedMark(); 2376 } else { 2377 // we'd like to assert something like the following: 2378 // assert(mark == markOopDesc::prototype(), "consistency check"); 2379 // ... but the above won't work because the age bits have not (yet) been 2380 // cleared. The remainder of the check would be identical to the 2381 // condition checked in must_be_preserved() above, so we don't really 2382 // have anything useful to check here! 2383 } 2384 if (_promoTail != NULL) { 2385 assert(_promoHead != NULL, "List consistency"); 2386 _promoTail->setNext(trackOop); 2387 _promoTail = trackOop; 2388 } else { 2389 assert(_promoHead == NULL, "List consistency"); 2390 _promoHead = _promoTail = trackOop; 2391 } 2392 // Mask as newly promoted, so we can skip over such objects 2393 // when scanning dirty cards 2394 assert(!trackOop->hasPromotedMark(), "Should not have been marked"); 2395 trackOop->setPromotedMark(); 2396 } 2397 2398 // Save the given displaced header, incrementing the pointer and 2399 // obtaining more spool area as necessary. 2400 void PromotionInfo::saveDisplacedHeader(markOop hdr) { 2401 assert(_spoolHead != NULL && _spoolTail != NULL, 2402 "promotionInfo inconsistency"); 2403 assert(_spoolTail->bufferSize > _nextIndex, "Off by one error at tail?"); 2404 _spoolTail->displacedHdr[_nextIndex] = hdr; 2405 // Spool forward 2406 if (++_nextIndex == _spoolTail->bufferSize) { // last location in this block 2407 // get a new spooling block 2408 assert(_spoolTail->nextSpoolBlock == NULL, "tail should terminate spool list"); 2409 _splice_point = _spoolTail; // save for splicing 2410 _spoolTail->nextSpoolBlock = getSpoolBlock(); // might fail 2411 _spoolTail = _spoolTail->nextSpoolBlock; // might become NULL ... 2412 // ... but will attempt filling before next promotion attempt 2413 _nextIndex = 1; 2414 } 2415 } 2416 2417 // Ensure that spooling space exists. Return false if spooling space 2418 // could not be obtained. 2419 bool PromotionInfo::ensure_spooling_space_work() { 2420 assert(!has_spooling_space(), "Only call when there is no spooling space"); 2421 // Try and obtain more spooling space 2422 SpoolBlock* newSpool = getSpoolBlock(); 2423 assert(newSpool == NULL || 2424 (newSpool->bufferSize != 0 && newSpool->nextSpoolBlock == NULL), 2425 "getSpoolBlock() sanity check"); 2426 if (newSpool == NULL) { 2427 return false; 2428 } 2429 _nextIndex = 1; 2430 if (_spoolTail == NULL) { 2431 _spoolTail = newSpool; 2432 if (_spoolHead == NULL) { 2433 _spoolHead = newSpool; 2434 _firstIndex = 1; 2435 } else { 2436 assert(_splice_point != NULL && _splice_point->nextSpoolBlock == NULL, 2437 "Splice point invariant"); 2438 // Extra check that _splice_point is connected to list 2439 #ifdef ASSERT 2440 { 2441 SpoolBlock* blk = _spoolHead; 2442 for (; blk->nextSpoolBlock != NULL; 2443 blk = blk->nextSpoolBlock); 2444 assert(blk != NULL && blk == _splice_point, 2445 "Splice point incorrect"); 2446 } 2447 #endif // ASSERT 2448 _splice_point->nextSpoolBlock = newSpool; 2449 } 2450 } else { 2451 assert(_spoolHead != NULL, "spool list consistency"); 2452 _spoolTail->nextSpoolBlock = newSpool; 2453 _spoolTail = newSpool; 2454 } 2455 return true; 2456 } 2457 2458 // Get a free spool buffer from the free pool, getting a new block 2459 // from the heap if necessary. 2460 SpoolBlock* PromotionInfo::getSpoolBlock() { 2461 SpoolBlock* res; 2462 if ((res = _spareSpool) != NULL) { 2463 _spareSpool = _spareSpool->nextSpoolBlock; 2464 res->nextSpoolBlock = NULL; 2465 } else { // spare spool exhausted, get some from heap 2466 res = (SpoolBlock*)(space()->allocateScratch(refillSize())); 2467 if (res != NULL) { 2468 res->init(); 2469 } 2470 } 2471 assert(res == NULL || res->nextSpoolBlock == NULL, "postcondition"); 2472 return res; 2473 } 2474 2475 void PromotionInfo::startTrackingPromotions() { 2476 assert(_spoolHead == _spoolTail && _firstIndex == _nextIndex, 2477 "spooling inconsistency?"); 2478 _firstIndex = _nextIndex = 1; 2479 _tracking = true; 2480 } 2481 2482 void PromotionInfo::stopTrackingPromotions() { 2483 assert(_spoolHead == _spoolTail && _firstIndex == _nextIndex, 2484 "spooling inconsistency?"); 2485 _firstIndex = _nextIndex = 1; 2486 _tracking = false; 2487 } 2488 2489 // When _spoolTail is not NULL, then the slot <_spoolTail, _nextIndex> 2490 // points to the next slot available for filling. 2491 // The set of slots holding displaced headers are then all those in the 2492 // right-open interval denoted by: 2493 // 2494 // [ <_spoolHead, _firstIndex>, <_spoolTail, _nextIndex> ) 2495 // 2496 // When _spoolTail is NULL, then the set of slots with displaced headers 2497 // is all those starting at the slot <_spoolHead, _firstIndex> and 2498 // going up to the last slot of last block in the linked list. 2499 // In this lartter case, _splice_point points to the tail block of 2500 // this linked list of blocks holding displaced headers. 2501 void PromotionInfo::verify() const { 2502 // Verify the following: 2503 // 1. the number of displaced headers matches the number of promoted 2504 // objects that have displaced headers 2505 // 2. each promoted object lies in this space 2506 debug_only( 2507 PromotedObject* junk = NULL; 2508 assert(junk->next_addr() == (void*)(oop(junk)->mark_addr()), 2509 "Offset of PromotedObject::_next is expected to align with " 2510 " the OopDesc::_mark within OopDesc"); 2511 ) 2512 // FIXME: guarantee???? 2513 guarantee(_spoolHead == NULL || _spoolTail != NULL || 2514 _splice_point != NULL, "list consistency"); 2515 guarantee(_promoHead == NULL || _promoTail != NULL, "list consistency"); 2516 // count the number of objects with displaced headers 2517 size_t numObjsWithDisplacedHdrs = 0; 2518 for (PromotedObject* curObj = _promoHead; curObj != NULL; curObj = curObj->next()) { 2519 guarantee(space()->is_in_reserved((HeapWord*)curObj), "Containment"); 2520 // the last promoted object may fail the mark() != NULL test of is_oop(). 2521 guarantee(curObj->next() == NULL || oop(curObj)->is_oop(), "must be an oop"); 2522 if (curObj->hasDisplacedMark()) { 2523 numObjsWithDisplacedHdrs++; 2524 } 2525 } 2526 // Count the number of displaced headers 2527 size_t numDisplacedHdrs = 0; 2528 for (SpoolBlock* curSpool = _spoolHead; 2529 curSpool != _spoolTail && curSpool != NULL; 2530 curSpool = curSpool->nextSpoolBlock) { 2531 // the first entry is just a self-pointer; indices 1 through 2532 // bufferSize - 1 are occupied (thus, bufferSize - 1 slots). 2533 guarantee((void*)curSpool->displacedHdr == (void*)&curSpool->displacedHdr, 2534 "first entry of displacedHdr should be self-referential"); 2535 numDisplacedHdrs += curSpool->bufferSize - 1; 2536 } 2537 guarantee((_spoolHead == _spoolTail) == (numDisplacedHdrs == 0), 2538 "internal consistency"); 2539 guarantee(_spoolTail != NULL || _nextIndex == 1, 2540 "Inconsistency between _spoolTail and _nextIndex"); 2541 // We overcounted (_firstIndex-1) worth of slots in block 2542 // _spoolHead and we undercounted (_nextIndex-1) worth of 2543 // slots in block _spoolTail. We make an appropriate 2544 // adjustment by subtracting the first and adding the 2545 // second: - (_firstIndex - 1) + (_nextIndex - 1) 2546 numDisplacedHdrs += (_nextIndex - _firstIndex); 2547 guarantee(numDisplacedHdrs == numObjsWithDisplacedHdrs, "Displaced hdr count"); 2548 } 2549 2550 2551 CFLS_LAB::CFLS_LAB(CompactibleFreeListSpace* cfls) : 2552 _cfls(cfls) 2553 { 2554 _blocks_to_claim = CMSParPromoteBlocksToClaim; 2555 for (size_t i = CompactibleFreeListSpace::IndexSetStart; 2556 i < CompactibleFreeListSpace::IndexSetSize; 2557 i += CompactibleFreeListSpace::IndexSetStride) { 2558 _indexedFreeList[i].set_size(i); 2559 } 2560 } 2561 2562 HeapWord* CFLS_LAB::alloc(size_t word_sz) { 2563 FreeChunk* res; 2564 word_sz = _cfls->adjustObjectSize(word_sz); 2565 if (word_sz >= CompactibleFreeListSpace::IndexSetSize) { 2566 // This locking manages sync with other large object allocations. 2567 MutexLockerEx x(_cfls->parDictionaryAllocLock(), 2568 Mutex::_no_safepoint_check_flag); 2569 res = _cfls->getChunkFromDictionaryExact(word_sz); 2570 if (res == NULL) return NULL; 2571 } else { 2572 FreeList* fl = &_indexedFreeList[word_sz]; 2573 bool filled = false; //TRAP 2574 if (fl->count() == 0) { 2575 bool filled = true; //TRAP 2576 // Attempt to refill this local free list. 2577 _cfls->par_get_chunk_of_blocks(word_sz, _blocks_to_claim, fl); 2578 // If it didn't work, give up. 2579 if (fl->count() == 0) return NULL; 2580 } 2581 res = fl->getChunkAtHead(); 2582 assert(res != NULL, "Why was count non-zero?"); 2583 } 2584 res->markNotFree(); 2585 assert(!res->isFree(), "shouldn't be marked free"); 2586 assert(oop(res)->klass() == NULL, "should look uninitialized"); 2587 // mangle a just allocated object with a distinct pattern. 2588 debug_only(res->mangleAllocated(word_sz)); 2589 return (HeapWord*)res; 2590 } 2591 2592 void CFLS_LAB::retire() { 2593 for (size_t i = CompactibleFreeListSpace::IndexSetStart; 2594 i < CompactibleFreeListSpace::IndexSetSize; 2595 i += CompactibleFreeListSpace::IndexSetStride) { 2596 if (_indexedFreeList[i].count() > 0) { 2597 MutexLockerEx x(_cfls->_indexedFreeListParLocks[i], 2598 Mutex::_no_safepoint_check_flag); 2599 _cfls->_indexedFreeList[i].prepend(&_indexedFreeList[i]); 2600 // Reset this list. 2601 _indexedFreeList[i] = FreeList(); 2602 _indexedFreeList[i].set_size(i); 2603 } 2604 } 2605 } 2606 2607 void 2608 CompactibleFreeListSpace:: 2609 par_get_chunk_of_blocks(size_t word_sz, size_t n, FreeList* fl) { 2610 assert(fl->count() == 0, "Precondition."); 2611 assert(word_sz < CompactibleFreeListSpace::IndexSetSize, 2612 "Precondition"); 2613 2614 // We'll try all multiples of word_sz in the indexed set (starting with 2615 // word_sz itself), then try getting a big chunk and splitting it. 2616 int k = 1; 2617 size_t cur_sz = k * word_sz; 2618 bool found = false; 2619 while (cur_sz < CompactibleFreeListSpace::IndexSetSize && k == 1) { 2620 FreeList* gfl = &_indexedFreeList[cur_sz]; 2621 FreeList fl_for_cur_sz; // Empty. 2622 fl_for_cur_sz.set_size(cur_sz); 2623 { 2624 MutexLockerEx x(_indexedFreeListParLocks[cur_sz], 2625 Mutex::_no_safepoint_check_flag); 2626 if (gfl->count() != 0) { 2627 size_t nn = MAX2(n/k, (size_t)1); 2628 gfl->getFirstNChunksFromList(nn, &fl_for_cur_sz); 2629 found = true; 2630 } 2631 } 2632 // Now transfer fl_for_cur_sz to fl. Common case, we hope, is k = 1. 2633 if (found) { 2634 if (k == 1) { 2635 fl->prepend(&fl_for_cur_sz); 2636 } else { 2637 // Divide each block on fl_for_cur_sz up k ways. 2638 FreeChunk* fc; 2639 while ((fc = fl_for_cur_sz.getChunkAtHead()) != NULL) { 2640 // Must do this in reverse order, so that anybody attempting to 2641 // access the main chunk sees it as a single free block until we 2642 // change it. 2643 size_t fc_size = fc->size(); 2644 for (int i = k-1; i >= 0; i--) { 2645 FreeChunk* ffc = (FreeChunk*)((HeapWord*)fc + i * word_sz); 2646 ffc->setSize(word_sz); 2647 ffc->linkNext(NULL); 2648 ffc->linkPrev(NULL); // Mark as a free block for other (parallel) GC threads. 2649 // Above must occur before BOT is updated below. 2650 // splitting from the right, fc_size == (k - i + 1) * wordsize 2651 _bt.mark_block((HeapWord*)ffc, word_sz); 2652 fc_size -= word_sz; 2653 _bt.verify_not_unallocated((HeapWord*)ffc, ffc->size()); 2654 _bt.verify_single_block((HeapWord*)fc, fc_size); 2655 _bt.verify_single_block((HeapWord*)ffc, ffc->size()); 2656 // Push this on "fl". 2657 fl->returnChunkAtHead(ffc); 2658 } 2659 // TRAP 2660 assert(fl->tail()->next() == NULL, "List invariant."); 2661 } 2662 } 2663 return; 2664 } 2665 k++; cur_sz = k * word_sz; 2666 } 2667 // Otherwise, we'll split a block from the dictionary. 2668 FreeChunk* fc = NULL; 2669 FreeChunk* rem_fc = NULL; 2670 size_t rem; 2671 { 2672 MutexLockerEx x(parDictionaryAllocLock(), 2673 Mutex::_no_safepoint_check_flag); 2674 while (n > 0) { 2675 fc = dictionary()->getChunk(MAX2(n * word_sz, 2676 _dictionary->minSize()), 2677 FreeBlockDictionary::atLeast); 2678 if (fc != NULL) { 2679 _bt.allocated((HeapWord*)fc, fc->size()); // update _unallocated_blk 2680 dictionary()->dictCensusUpdate(fc->size(), 2681 true /*split*/, 2682 false /*birth*/); 2683 break; 2684 } else { 2685 n--; 2686 } 2687 } 2688 if (fc == NULL) return; 2689 // Otherwise, split up that block. 2690 size_t nn = fc->size() / word_sz; 2691 n = MIN2(nn, n); 2692 rem = fc->size() - n * word_sz; 2693 // If there is a remainder, and it's too small, allocate one fewer. 2694 if (rem > 0 && rem < MinChunkSize) { 2695 n--; rem += word_sz; 2696 } 2697 // First return the remainder, if any. 2698 // Note that we hold the lock until we decide if we're going to give 2699 // back the remainder to the dictionary, since a contending allocator 2700 // may otherwise see the heap as empty. (We're willing to take that 2701 // hit if the block is a small block.) 2702 if (rem > 0) { 2703 size_t prefix_size = n * word_sz; 2704 rem_fc = (FreeChunk*)((HeapWord*)fc + prefix_size); 2705 rem_fc->setSize(rem); 2706 rem_fc->linkNext(NULL); 2707 rem_fc->linkPrev(NULL); // Mark as a free block for other (parallel) GC threads. 2708 // Above must occur before BOT is updated below. 2709 _bt.split_block((HeapWord*)fc, fc->size(), prefix_size); 2710 if (rem >= IndexSetSize) { 2711 returnChunkToDictionary(rem_fc); 2712 dictionary()->dictCensusUpdate(fc->size(), 2713 true /*split*/, 2714 true /*birth*/); 2715 rem_fc = NULL; 2716 } 2717 // Otherwise, return it to the small list below. 2718 } 2719 } 2720 // 2721 if (rem_fc != NULL) { 2722 MutexLockerEx x(_indexedFreeListParLocks[rem], 2723 Mutex::_no_safepoint_check_flag); 2724 _bt.verify_not_unallocated((HeapWord*)rem_fc, rem_fc->size()); 2725 _indexedFreeList[rem].returnChunkAtHead(rem_fc); 2726 smallSplitBirth(rem); 2727 } 2728 2729 // Now do the splitting up. 2730 // Must do this in reverse order, so that anybody attempting to 2731 // access the main chunk sees it as a single free block until we 2732 // change it. 2733 size_t fc_size = n * word_sz; 2734 // All but first chunk in this loop 2735 for (ssize_t i = n-1; i > 0; i--) { 2736 FreeChunk* ffc = (FreeChunk*)((HeapWord*)fc + i * word_sz); 2737 ffc->setSize(word_sz); 2738 ffc->linkNext(NULL); 2739 ffc->linkPrev(NULL); // Mark as a free block for other (parallel) GC threads. 2740 // Above must occur before BOT is updated below. 2741 // splitting from the right, fc_size == (n - i + 1) * wordsize 2742 _bt.mark_block((HeapWord*)ffc, word_sz); 2743 fc_size -= word_sz; 2744 _bt.verify_not_unallocated((HeapWord*)ffc, ffc->size()); 2745 _bt.verify_single_block((HeapWord*)ffc, ffc->size()); 2746 _bt.verify_single_block((HeapWord*)fc, fc_size); 2747 // Push this on "fl". 2748 fl->returnChunkAtHead(ffc); 2749 } 2750 // First chunk 2751 fc->setSize(word_sz); 2752 fc->linkNext(NULL); 2753 fc->linkPrev(NULL); 2754 _bt.verify_not_unallocated((HeapWord*)fc, fc->size()); 2755 _bt.verify_single_block((HeapWord*)fc, fc->size()); 2756 fl->returnChunkAtHead(fc); 2757 2758 { 2759 MutexLockerEx x(_indexedFreeListParLocks[word_sz], 2760 Mutex::_no_safepoint_check_flag); 2761 ssize_t new_births = _indexedFreeList[word_sz].splitBirths() + n; 2762 _indexedFreeList[word_sz].set_splitBirths(new_births); 2763 ssize_t new_surplus = _indexedFreeList[word_sz].surplus() + n; 2764 _indexedFreeList[word_sz].set_surplus(new_surplus); 2765 } 2766 2767 // TRAP 2768 assert(fl->tail()->next() == NULL, "List invariant."); 2769 } 2770 2771 // Set up the space's par_seq_tasks structure for work claiming 2772 // for parallel rescan. See CMSParRemarkTask where this is currently used. 2773 // XXX Need to suitably abstract and generalize this and the next 2774 // method into one. 2775 void 2776 CompactibleFreeListSpace:: 2777 initialize_sequential_subtasks_for_rescan(int n_threads) { 2778 // The "size" of each task is fixed according to rescan_task_size. 2779 assert(n_threads > 0, "Unexpected n_threads argument"); 2780 const size_t task_size = rescan_task_size(); 2781 size_t n_tasks = (used_region().word_size() + task_size - 1)/task_size; 2782 assert((used_region().start() + (n_tasks - 1)*task_size < 2783 used_region().end()) && 2784 (used_region().start() + n_tasks*task_size >= 2785 used_region().end()), "n_task calculation incorrect"); 2786 SequentialSubTasksDone* pst = conc_par_seq_tasks(); 2787 assert(!pst->valid(), "Clobbering existing data?"); 2788 pst->set_par_threads(n_threads); 2789 pst->set_n_tasks((int)n_tasks); 2790 } 2791 2792 // Set up the space's par_seq_tasks structure for work claiming 2793 // for parallel concurrent marking. See CMSConcMarkTask where this is currently used. 2794 void 2795 CompactibleFreeListSpace:: 2796 initialize_sequential_subtasks_for_marking(int n_threads, 2797 HeapWord* low) { 2798 // The "size" of each task is fixed according to rescan_task_size. 2799 assert(n_threads > 0, "Unexpected n_threads argument"); 2800 const size_t task_size = marking_task_size(); 2801 assert(task_size > CardTableModRefBS::card_size_in_words && 2802 (task_size % CardTableModRefBS::card_size_in_words == 0), 2803 "Otherwise arithmetic below would be incorrect"); 2804 MemRegion span = _gen->reserved(); 2805 if (low != NULL) { 2806 if (span.contains(low)) { 2807 // Align low down to a card boundary so that 2808 // we can use block_offset_careful() on span boundaries. 2809 HeapWord* aligned_low = (HeapWord*)align_size_down((uintptr_t)low, 2810 CardTableModRefBS::card_size); 2811 // Clip span prefix at aligned_low 2812 span = span.intersection(MemRegion(aligned_low, span.end())); 2813 } else if (low > span.end()) { 2814 span = MemRegion(low, low); // Null region 2815 } // else use entire span 2816 } 2817 assert(span.is_empty() || 2818 ((uintptr_t)span.start() % CardTableModRefBS::card_size == 0), 2819 "span should start at a card boundary"); 2820 size_t n_tasks = (span.word_size() + task_size - 1)/task_size; 2821 assert((n_tasks == 0) == span.is_empty(), "Inconsistency"); 2822 assert(n_tasks == 0 || 2823 ((span.start() + (n_tasks - 1)*task_size < span.end()) && 2824 (span.start() + n_tasks*task_size >= span.end())), 2825 "n_task calculation incorrect"); 2826 SequentialSubTasksDone* pst = conc_par_seq_tasks(); 2827 assert(!pst->valid(), "Clobbering existing data?"); 2828 pst->set_par_threads(n_threads); 2829 pst->set_n_tasks((int)n_tasks); 2830 }