1 /* 2 * Copyright (c) 2001, 2011, Oracle and/or its affiliates. All rights reserved. 3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. 4 * 5 * This code is free software; you can redistribute it and/or modify it 6 * under the terms of the GNU General Public License version 2 only, as 7 * published by the Free Software Foundation. 8 * 9 * This code is distributed in the hope that it will be useful, but WITHOUT 10 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or 11 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License 12 * version 2 for more details (a copy is included in the LICENSE file that 13 * accompanied this code). 14 * 15 * You should have received a copy of the GNU General Public License version 16 * 2 along with this work; if not, write to the Free Software Foundation, 17 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. 18 * 19 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA 20 * or visit www.oracle.com if you need additional information or have any 21 * questions. 22 * 23 */ 24 25 #include "precompiled.hpp" 26 #include "classfile/symbolTable.hpp" 27 #include "gc_implementation/g1/concurrentMark.inline.hpp" 28 #include "gc_implementation/g1/concurrentMarkThread.inline.hpp" 29 #include "gc_implementation/g1/g1CollectedHeap.inline.hpp" 30 #include "gc_implementation/g1/g1CollectorPolicy.hpp" 31 #include "gc_implementation/g1/g1ErgoVerbose.hpp" 32 #include "gc_implementation/g1/g1OopClosures.inline.hpp" 33 #include "gc_implementation/g1/g1RemSet.hpp" 34 #include "gc_implementation/g1/heapRegionRemSet.hpp" 35 #include "gc_implementation/g1/heapRegionSeq.inline.hpp" 36 #include "gc_implementation/shared/vmGCOperations.hpp" 37 #include "memory/genOopClosures.inline.hpp" 38 #include "memory/referencePolicy.hpp" 39 #include "memory/resourceArea.hpp" 40 #include "oops/oop.inline.hpp" 41 #include "runtime/handles.inline.hpp" 42 #include "runtime/java.hpp" 43 44 // 45 // CMS Bit Map Wrapper 46 47 CMBitMapRO::CMBitMapRO(ReservedSpace rs, int shifter): 48 _bm((uintptr_t*)NULL,0), 49 _shifter(shifter) { 50 _bmStartWord = (HeapWord*)(rs.base()); 51 _bmWordSize = rs.size()/HeapWordSize; // rs.size() is in bytes 52 ReservedSpace brs(ReservedSpace::allocation_align_size_up( 53 (_bmWordSize >> (_shifter + LogBitsPerByte)) + 1)); 54 55 guarantee(brs.is_reserved(), "couldn't allocate CMS bit map"); 56 // For now we'll just commit all of the bit map up fromt. 57 // Later on we'll try to be more parsimonious with swap. 58 guarantee(_virtual_space.initialize(brs, brs.size()), 59 "couldn't reseve backing store for CMS bit map"); 60 assert(_virtual_space.committed_size() == brs.size(), 61 "didn't reserve backing store for all of CMS bit map?"); 62 _bm.set_map((uintptr_t*)_virtual_space.low()); 63 assert(_virtual_space.committed_size() << (_shifter + LogBitsPerByte) >= 64 _bmWordSize, "inconsistency in bit map sizing"); 65 _bm.set_size(_bmWordSize >> _shifter); 66 } 67 68 HeapWord* CMBitMapRO::getNextMarkedWordAddress(HeapWord* addr, 69 HeapWord* limit) const { 70 // First we must round addr *up* to a possible object boundary. 71 addr = (HeapWord*)align_size_up((intptr_t)addr, 72 HeapWordSize << _shifter); 73 size_t addrOffset = heapWordToOffset(addr); 74 if (limit == NULL) { 75 limit = _bmStartWord + _bmWordSize; 76 } 77 size_t limitOffset = heapWordToOffset(limit); 78 size_t nextOffset = _bm.get_next_one_offset(addrOffset, limitOffset); 79 HeapWord* nextAddr = offsetToHeapWord(nextOffset); 80 assert(nextAddr >= addr, "get_next_one postcondition"); 81 assert(nextAddr == limit || isMarked(nextAddr), 82 "get_next_one postcondition"); 83 return nextAddr; 84 } 85 86 HeapWord* CMBitMapRO::getNextUnmarkedWordAddress(HeapWord* addr, 87 HeapWord* limit) const { 88 size_t addrOffset = heapWordToOffset(addr); 89 if (limit == NULL) { 90 limit = _bmStartWord + _bmWordSize; 91 } 92 size_t limitOffset = heapWordToOffset(limit); 93 size_t nextOffset = _bm.get_next_zero_offset(addrOffset, limitOffset); 94 HeapWord* nextAddr = offsetToHeapWord(nextOffset); 95 assert(nextAddr >= addr, "get_next_one postcondition"); 96 assert(nextAddr == limit || !isMarked(nextAddr), 97 "get_next_one postcondition"); 98 return nextAddr; 99 } 100 101 int CMBitMapRO::heapWordDiffToOffsetDiff(size_t diff) const { 102 assert((diff & ((1 << _shifter) - 1)) == 0, "argument check"); 103 return (int) (diff >> _shifter); 104 } 105 106 bool CMBitMapRO::iterate(BitMapClosure* cl, MemRegion mr) { 107 HeapWord* left = MAX2(_bmStartWord, mr.start()); 108 HeapWord* right = MIN2(_bmStartWord + _bmWordSize, mr.end()); 109 if (right > left) { 110 // Right-open interval [leftOffset, rightOffset). 111 return _bm.iterate(cl, heapWordToOffset(left), heapWordToOffset(right)); 112 } else { 113 return true; 114 } 115 } 116 117 void CMBitMapRO::mostly_disjoint_range_union(BitMap* from_bitmap, 118 size_t from_start_index, 119 HeapWord* to_start_word, 120 size_t word_num) { 121 _bm.mostly_disjoint_range_union(from_bitmap, 122 from_start_index, 123 heapWordToOffset(to_start_word), 124 word_num); 125 } 126 127 #ifndef PRODUCT 128 bool CMBitMapRO::covers(ReservedSpace rs) const { 129 // assert(_bm.map() == _virtual_space.low(), "map inconsistency"); 130 assert(((size_t)_bm.size() * (size_t)(1 << _shifter)) == _bmWordSize, 131 "size inconsistency"); 132 return _bmStartWord == (HeapWord*)(rs.base()) && 133 _bmWordSize == rs.size()>>LogHeapWordSize; 134 } 135 #endif 136 137 void CMBitMap::clearAll() { 138 _bm.clear(); 139 return; 140 } 141 142 void CMBitMap::markRange(MemRegion mr) { 143 mr.intersection(MemRegion(_bmStartWord, _bmWordSize)); 144 assert(!mr.is_empty(), "unexpected empty region"); 145 assert((offsetToHeapWord(heapWordToOffset(mr.end())) == 146 ((HeapWord *) mr.end())), 147 "markRange memory region end is not card aligned"); 148 // convert address range into offset range 149 _bm.at_put_range(heapWordToOffset(mr.start()), 150 heapWordToOffset(mr.end()), true); 151 } 152 153 void CMBitMap::clearRange(MemRegion mr) { 154 mr.intersection(MemRegion(_bmStartWord, _bmWordSize)); 155 assert(!mr.is_empty(), "unexpected empty region"); 156 // convert address range into offset range 157 _bm.at_put_range(heapWordToOffset(mr.start()), 158 heapWordToOffset(mr.end()), false); 159 } 160 161 MemRegion CMBitMap::getAndClearMarkedRegion(HeapWord* addr, 162 HeapWord* end_addr) { 163 HeapWord* start = getNextMarkedWordAddress(addr); 164 start = MIN2(start, end_addr); 165 HeapWord* end = getNextUnmarkedWordAddress(start); 166 end = MIN2(end, end_addr); 167 assert(start <= end, "Consistency check"); 168 MemRegion mr(start, end); 169 if (!mr.is_empty()) { 170 clearRange(mr); 171 } 172 return mr; 173 } 174 175 CMMarkStack::CMMarkStack(ConcurrentMark* cm) : 176 _base(NULL), _cm(cm) 177 #ifdef ASSERT 178 , _drain_in_progress(false) 179 , _drain_in_progress_yields(false) 180 #endif 181 {} 182 183 void CMMarkStack::allocate(size_t size) { 184 _base = NEW_C_HEAP_ARRAY(oop, size); 185 if (_base == NULL) { 186 vm_exit_during_initialization("Failed to allocate " 187 "CM region mark stack"); 188 } 189 _index = 0; 190 _capacity = (jint) size; 191 _oops_do_bound = -1; 192 NOT_PRODUCT(_max_depth = 0); 193 } 194 195 CMMarkStack::~CMMarkStack() { 196 if (_base != NULL) { 197 FREE_C_HEAP_ARRAY(oop, _base); 198 } 199 } 200 201 void CMMarkStack::par_push(oop ptr) { 202 while (true) { 203 if (isFull()) { 204 _overflow = true; 205 return; 206 } 207 // Otherwise... 208 jint index = _index; 209 jint next_index = index+1; 210 jint res = Atomic::cmpxchg(next_index, &_index, index); 211 if (res == index) { 212 _base[index] = ptr; 213 // Note that we don't maintain this atomically. We could, but it 214 // doesn't seem necessary. 215 NOT_PRODUCT(_max_depth = MAX2(_max_depth, next_index)); 216 return; 217 } 218 // Otherwise, we need to try again. 219 } 220 } 221 222 void CMMarkStack::par_adjoin_arr(oop* ptr_arr, int n) { 223 while (true) { 224 if (isFull()) { 225 _overflow = true; 226 return; 227 } 228 // Otherwise... 229 jint index = _index; 230 jint next_index = index + n; 231 if (next_index > _capacity) { 232 _overflow = true; 233 return; 234 } 235 jint res = Atomic::cmpxchg(next_index, &_index, index); 236 if (res == index) { 237 for (int i = 0; i < n; i++) { 238 int ind = index + i; 239 assert(ind < _capacity, "By overflow test above."); 240 _base[ind] = ptr_arr[i]; 241 } 242 NOT_PRODUCT(_max_depth = MAX2(_max_depth, next_index)); 243 return; 244 } 245 // Otherwise, we need to try again. 246 } 247 } 248 249 250 void CMMarkStack::par_push_arr(oop* ptr_arr, int n) { 251 MutexLockerEx x(ParGCRareEvent_lock, Mutex::_no_safepoint_check_flag); 252 jint start = _index; 253 jint next_index = start + n; 254 if (next_index > _capacity) { 255 _overflow = true; 256 return; 257 } 258 // Otherwise. 259 _index = next_index; 260 for (int i = 0; i < n; i++) { 261 int ind = start + i; 262 assert(ind < _capacity, "By overflow test above."); 263 _base[ind] = ptr_arr[i]; 264 } 265 } 266 267 268 bool CMMarkStack::par_pop_arr(oop* ptr_arr, int max, int* n) { 269 MutexLockerEx x(ParGCRareEvent_lock, Mutex::_no_safepoint_check_flag); 270 jint index = _index; 271 if (index == 0) { 272 *n = 0; 273 return false; 274 } else { 275 int k = MIN2(max, index); 276 jint new_ind = index - k; 277 for (int j = 0; j < k; j++) { 278 ptr_arr[j] = _base[new_ind + j]; 279 } 280 _index = new_ind; 281 *n = k; 282 return true; 283 } 284 } 285 286 287 CMRegionStack::CMRegionStack() : _base(NULL) {} 288 289 void CMRegionStack::allocate(size_t size) { 290 _base = NEW_C_HEAP_ARRAY(MemRegion, size); 291 if (_base == NULL) { 292 vm_exit_during_initialization("Failed to allocate CM region mark stack"); 293 } 294 _index = 0; 295 _capacity = (jint) size; 296 } 297 298 CMRegionStack::~CMRegionStack() { 299 if (_base != NULL) { 300 FREE_C_HEAP_ARRAY(oop, _base); 301 } 302 } 303 304 void CMRegionStack::push_lock_free(MemRegion mr) { 305 assert(mr.word_size() > 0, "Precondition"); 306 while (true) { 307 jint index = _index; 308 309 if (index >= _capacity) { 310 _overflow = true; 311 return; 312 } 313 // Otherwise... 314 jint next_index = index+1; 315 jint res = Atomic::cmpxchg(next_index, &_index, index); 316 if (res == index) { 317 _base[index] = mr; 318 return; 319 } 320 // Otherwise, we need to try again. 321 } 322 } 323 324 // Lock-free pop of the region stack. Called during the concurrent 325 // marking / remark phases. Should only be called in tandem with 326 // other lock-free pops. 327 MemRegion CMRegionStack::pop_lock_free() { 328 while (true) { 329 jint index = _index; 330 331 if (index == 0) { 332 return MemRegion(); 333 } 334 // Otherwise... 335 jint next_index = index-1; 336 jint res = Atomic::cmpxchg(next_index, &_index, index); 337 if (res == index) { 338 MemRegion mr = _base[next_index]; 339 if (mr.start() != NULL) { 340 assert(mr.end() != NULL, "invariant"); 341 assert(mr.word_size() > 0, "invariant"); 342 return mr; 343 } else { 344 // that entry was invalidated... let's skip it 345 assert(mr.end() == NULL, "invariant"); 346 } 347 } 348 // Otherwise, we need to try again. 349 } 350 } 351 352 #if 0 353 // The routines that manipulate the region stack with a lock are 354 // not currently used. They should be retained, however, as a 355 // diagnostic aid. 356 357 void CMRegionStack::push_with_lock(MemRegion mr) { 358 assert(mr.word_size() > 0, "Precondition"); 359 MutexLockerEx x(CMRegionStack_lock, Mutex::_no_safepoint_check_flag); 360 361 if (isFull()) { 362 _overflow = true; 363 return; 364 } 365 366 _base[_index] = mr; 367 _index += 1; 368 } 369 370 MemRegion CMRegionStack::pop_with_lock() { 371 MutexLockerEx x(CMRegionStack_lock, Mutex::_no_safepoint_check_flag); 372 373 while (true) { 374 if (_index == 0) { 375 return MemRegion(); 376 } 377 _index -= 1; 378 379 MemRegion mr = _base[_index]; 380 if (mr.start() != NULL) { 381 assert(mr.end() != NULL, "invariant"); 382 assert(mr.word_size() > 0, "invariant"); 383 return mr; 384 } else { 385 // that entry was invalidated... let's skip it 386 assert(mr.end() == NULL, "invariant"); 387 } 388 } 389 } 390 #endif 391 392 bool CMRegionStack::invalidate_entries_into_cset() { 393 bool result = false; 394 G1CollectedHeap* g1h = G1CollectedHeap::heap(); 395 for (int i = 0; i < _oops_do_bound; ++i) { 396 MemRegion mr = _base[i]; 397 if (mr.start() != NULL) { 398 assert(mr.end() != NULL, "invariant"); 399 assert(mr.word_size() > 0, "invariant"); 400 HeapRegion* hr = g1h->heap_region_containing(mr.start()); 401 assert(hr != NULL, "invariant"); 402 if (hr->in_collection_set()) { 403 // The region points into the collection set 404 _base[i] = MemRegion(); 405 result = true; 406 } 407 } else { 408 // that entry was invalidated... let's skip it 409 assert(mr.end() == NULL, "invariant"); 410 } 411 } 412 return result; 413 } 414 415 template<class OopClosureClass> 416 bool CMMarkStack::drain(OopClosureClass* cl, CMBitMap* bm, bool yield_after) { 417 assert(!_drain_in_progress || !_drain_in_progress_yields || yield_after 418 || SafepointSynchronize::is_at_safepoint(), 419 "Drain recursion must be yield-safe."); 420 bool res = true; 421 debug_only(_drain_in_progress = true); 422 debug_only(_drain_in_progress_yields = yield_after); 423 while (!isEmpty()) { 424 oop newOop = pop(); 425 assert(G1CollectedHeap::heap()->is_in_reserved(newOop), "Bad pop"); 426 assert(newOop->is_oop(), "Expected an oop"); 427 assert(bm == NULL || bm->isMarked((HeapWord*)newOop), 428 "only grey objects on this stack"); 429 // iterate over the oops in this oop, marking and pushing 430 // the ones in CMS generation. 431 newOop->oop_iterate(cl); 432 if (yield_after && _cm->do_yield_check()) { 433 res = false; 434 break; 435 } 436 } 437 debug_only(_drain_in_progress = false); 438 return res; 439 } 440 441 void CMMarkStack::oops_do(OopClosure* f) { 442 if (_index == 0) return; 443 assert(_oops_do_bound != -1 && _oops_do_bound <= _index, 444 "Bound must be set."); 445 for (int i = 0; i < _oops_do_bound; i++) { 446 f->do_oop(&_base[i]); 447 } 448 _oops_do_bound = -1; 449 } 450 451 bool ConcurrentMark::not_yet_marked(oop obj) const { 452 return (_g1h->is_obj_ill(obj) 453 || (_g1h->is_in_permanent(obj) 454 && !nextMarkBitMap()->isMarked((HeapWord*)obj))); 455 } 456 457 #ifdef _MSC_VER // the use of 'this' below gets a warning, make it go away 458 #pragma warning( disable:4355 ) // 'this' : used in base member initializer list 459 #endif // _MSC_VER 460 461 ConcurrentMark::ConcurrentMark(ReservedSpace rs, 462 int max_regions) : 463 _markBitMap1(rs, MinObjAlignment - 1), 464 _markBitMap2(rs, MinObjAlignment - 1), 465 466 _parallel_marking_threads(0), 467 _sleep_factor(0.0), 468 _marking_task_overhead(1.0), 469 _cleanup_sleep_factor(0.0), 470 _cleanup_task_overhead(1.0), 471 _cleanup_list("Cleanup List"), 472 _region_bm(max_regions, false /* in_resource_area*/), 473 _card_bm((rs.size() + CardTableModRefBS::card_size - 1) >> 474 CardTableModRefBS::card_shift, 475 false /* in_resource_area*/), 476 _prevMarkBitMap(&_markBitMap1), 477 _nextMarkBitMap(&_markBitMap2), 478 _at_least_one_mark_complete(false), 479 480 _markStack(this), 481 _regionStack(), 482 // _finger set in set_non_marking_state 483 484 _max_task_num(MAX2(ParallelGCThreads, (size_t)1)), 485 // _active_tasks set in set_non_marking_state 486 // _tasks set inside the constructor 487 _task_queues(new CMTaskQueueSet((int) _max_task_num)), 488 _terminator(ParallelTaskTerminator((int) _max_task_num, _task_queues)), 489 490 _has_overflown(false), 491 _concurrent(false), 492 _has_aborted(false), 493 _restart_for_overflow(false), 494 _concurrent_marking_in_progress(false), 495 _should_gray_objects(false), 496 497 // _verbose_level set below 498 499 _init_times(), 500 _remark_times(), _remark_mark_times(), _remark_weak_ref_times(), 501 _cleanup_times(), 502 _total_counting_time(0.0), 503 _total_rs_scrub_time(0.0), 504 505 _parallel_workers(NULL) { 506 CMVerboseLevel verbose_level = (CMVerboseLevel) G1MarkingVerboseLevel; 507 if (verbose_level < no_verbose) { 508 verbose_level = no_verbose; 509 } 510 if (verbose_level > high_verbose) { 511 verbose_level = high_verbose; 512 } 513 _verbose_level = verbose_level; 514 515 if (verbose_low()) { 516 gclog_or_tty->print_cr("[global] init, heap start = "PTR_FORMAT", " 517 "heap end = "PTR_FORMAT, _heap_start, _heap_end); 518 } 519 520 _markStack.allocate(MarkStackSize); 521 _regionStack.allocate(G1MarkRegionStackSize); 522 523 // Create & start a ConcurrentMark thread. 524 _cmThread = new ConcurrentMarkThread(this); 525 assert(cmThread() != NULL, "CM Thread should have been created"); 526 assert(cmThread()->cm() != NULL, "CM Thread should refer to this cm"); 527 528 _g1h = G1CollectedHeap::heap(); 529 assert(CGC_lock != NULL, "Where's the CGC_lock?"); 530 assert(_markBitMap1.covers(rs), "_markBitMap1 inconsistency"); 531 assert(_markBitMap2.covers(rs), "_markBitMap2 inconsistency"); 532 533 SATBMarkQueueSet& satb_qs = JavaThread::satb_mark_queue_set(); 534 satb_qs.set_buffer_size(G1SATBBufferSize); 535 536 _tasks = NEW_C_HEAP_ARRAY(CMTask*, _max_task_num); 537 _accum_task_vtime = NEW_C_HEAP_ARRAY(double, _max_task_num); 538 539 // so that the assertion in MarkingTaskQueue::task_queue doesn't fail 540 _active_tasks = _max_task_num; 541 for (int i = 0; i < (int) _max_task_num; ++i) { 542 CMTaskQueue* task_queue = new CMTaskQueue(); 543 task_queue->initialize(); 544 _task_queues->register_queue(i, task_queue); 545 546 _tasks[i] = new CMTask(i, this, task_queue, _task_queues); 547 _accum_task_vtime[i] = 0.0; 548 } 549 550 if (ConcGCThreads > ParallelGCThreads) { 551 vm_exit_during_initialization("Can't have more ConcGCThreads " 552 "than ParallelGCThreads."); 553 } 554 if (ParallelGCThreads == 0) { 555 // if we are not running with any parallel GC threads we will not 556 // spawn any marking threads either 557 _parallel_marking_threads = 0; 558 _sleep_factor = 0.0; 559 _marking_task_overhead = 1.0; 560 } else { 561 if (ConcGCThreads > 0) { 562 // notice that ConcGCThreads overwrites G1MarkingOverheadPercent 563 // if both are set 564 565 _parallel_marking_threads = ConcGCThreads; 566 _sleep_factor = 0.0; 567 _marking_task_overhead = 1.0; 568 } else if (G1MarkingOverheadPercent > 0) { 569 // we will calculate the number of parallel marking threads 570 // based on a target overhead with respect to the soft real-time 571 // goal 572 573 double marking_overhead = (double) G1MarkingOverheadPercent / 100.0; 574 double overall_cm_overhead = 575 (double) MaxGCPauseMillis * marking_overhead / 576 (double) GCPauseIntervalMillis; 577 double cpu_ratio = 1.0 / (double) os::processor_count(); 578 double marking_thread_num = ceil(overall_cm_overhead / cpu_ratio); 579 double marking_task_overhead = 580 overall_cm_overhead / marking_thread_num * 581 (double) os::processor_count(); 582 double sleep_factor = 583 (1.0 - marking_task_overhead) / marking_task_overhead; 584 585 _parallel_marking_threads = (size_t) marking_thread_num; 586 _sleep_factor = sleep_factor; 587 _marking_task_overhead = marking_task_overhead; 588 } else { 589 _parallel_marking_threads = MAX2((ParallelGCThreads + 2) / 4, (size_t)1); 590 _sleep_factor = 0.0; 591 _marking_task_overhead = 1.0; 592 } 593 594 if (parallel_marking_threads() > 1) { 595 _cleanup_task_overhead = 1.0; 596 } else { 597 _cleanup_task_overhead = marking_task_overhead(); 598 } 599 _cleanup_sleep_factor = 600 (1.0 - cleanup_task_overhead()) / cleanup_task_overhead(); 601 602 #if 0 603 gclog_or_tty->print_cr("Marking Threads %d", parallel_marking_threads()); 604 gclog_or_tty->print_cr("CM Marking Task Overhead %1.4lf", marking_task_overhead()); 605 gclog_or_tty->print_cr("CM Sleep Factor %1.4lf", sleep_factor()); 606 gclog_or_tty->print_cr("CL Marking Task Overhead %1.4lf", cleanup_task_overhead()); 607 gclog_or_tty->print_cr("CL Sleep Factor %1.4lf", cleanup_sleep_factor()); 608 #endif 609 610 guarantee(parallel_marking_threads() > 0, "peace of mind"); 611 _parallel_workers = new FlexibleWorkGang("G1 Parallel Marking Threads", 612 (int) _parallel_marking_threads, false, true); 613 if (_parallel_workers == NULL) { 614 vm_exit_during_initialization("Failed necessary allocation."); 615 } else { 616 _parallel_workers->initialize_workers(); 617 } 618 } 619 620 // so that the call below can read a sensible value 621 _heap_start = (HeapWord*) rs.base(); 622 set_non_marking_state(); 623 } 624 625 void ConcurrentMark::update_g1_committed(bool force) { 626 // If concurrent marking is not in progress, then we do not need to 627 // update _heap_end. This has a subtle and important 628 // side-effect. Imagine that two evacuation pauses happen between 629 // marking completion and remark. The first one can grow the 630 // heap (hence now the finger is below the heap end). Then, the 631 // second one could unnecessarily push regions on the region 632 // stack. This causes the invariant that the region stack is empty 633 // at the beginning of remark to be false. By ensuring that we do 634 // not observe heap expansions after marking is complete, then we do 635 // not have this problem. 636 if (!concurrent_marking_in_progress() && !force) return; 637 638 MemRegion committed = _g1h->g1_committed(); 639 assert(committed.start() == _heap_start, "start shouldn't change"); 640 HeapWord* new_end = committed.end(); 641 if (new_end > _heap_end) { 642 // The heap has been expanded. 643 644 _heap_end = new_end; 645 } 646 // Notice that the heap can also shrink. However, this only happens 647 // during a Full GC (at least currently) and the entire marking 648 // phase will bail out and the task will not be restarted. So, let's 649 // do nothing. 650 } 651 652 void ConcurrentMark::reset() { 653 // Starting values for these two. This should be called in a STW 654 // phase. CM will be notified of any future g1_committed expansions 655 // will be at the end of evacuation pauses, when tasks are 656 // inactive. 657 MemRegion committed = _g1h->g1_committed(); 658 _heap_start = committed.start(); 659 _heap_end = committed.end(); 660 661 // Separated the asserts so that we know which one fires. 662 assert(_heap_start != NULL, "heap bounds should look ok"); 663 assert(_heap_end != NULL, "heap bounds should look ok"); 664 assert(_heap_start < _heap_end, "heap bounds should look ok"); 665 666 // reset all the marking data structures and any necessary flags 667 clear_marking_state(); 668 669 if (verbose_low()) { 670 gclog_or_tty->print_cr("[global] resetting"); 671 } 672 673 // We do reset all of them, since different phases will use 674 // different number of active threads. So, it's easiest to have all 675 // of them ready. 676 for (int i = 0; i < (int) _max_task_num; ++i) { 677 _tasks[i]->reset(_nextMarkBitMap); 678 } 679 680 // we need this to make sure that the flag is on during the evac 681 // pause with initial mark piggy-backed 682 set_concurrent_marking_in_progress(); 683 } 684 685 void ConcurrentMark::set_phase(size_t active_tasks, bool concurrent) { 686 assert(active_tasks <= _max_task_num, "we should not have more"); 687 688 _active_tasks = active_tasks; 689 // Need to update the three data structures below according to the 690 // number of active threads for this phase. 691 _terminator = ParallelTaskTerminator((int) active_tasks, _task_queues); 692 _first_overflow_barrier_sync.set_n_workers((int) active_tasks); 693 _second_overflow_barrier_sync.set_n_workers((int) active_tasks); 694 695 _concurrent = concurrent; 696 // We propagate this to all tasks, not just the active ones. 697 for (int i = 0; i < (int) _max_task_num; ++i) 698 _tasks[i]->set_concurrent(concurrent); 699 700 if (concurrent) { 701 set_concurrent_marking_in_progress(); 702 } else { 703 // We currently assume that the concurrent flag has been set to 704 // false before we start remark. At this point we should also be 705 // in a STW phase. 706 assert(!concurrent_marking_in_progress(), "invariant"); 707 assert(_finger == _heap_end, "only way to get here"); 708 update_g1_committed(true); 709 } 710 } 711 712 void ConcurrentMark::set_non_marking_state() { 713 // We set the global marking state to some default values when we're 714 // not doing marking. 715 clear_marking_state(); 716 _active_tasks = 0; 717 clear_concurrent_marking_in_progress(); 718 } 719 720 ConcurrentMark::~ConcurrentMark() { 721 for (int i = 0; i < (int) _max_task_num; ++i) { 722 delete _task_queues->queue(i); 723 delete _tasks[i]; 724 } 725 delete _task_queues; 726 FREE_C_HEAP_ARRAY(CMTask*, _max_task_num); 727 } 728 729 // This closure is used to mark refs into the g1 generation 730 // from external roots in the CMS bit map. 731 // Called at the first checkpoint. 732 // 733 734 void ConcurrentMark::clearNextBitmap() { 735 G1CollectedHeap* g1h = G1CollectedHeap::heap(); 736 G1CollectorPolicy* g1p = g1h->g1_policy(); 737 738 // Make sure that the concurrent mark thread looks to still be in 739 // the current cycle. 740 guarantee(cmThread()->during_cycle(), "invariant"); 741 742 // We are finishing up the current cycle by clearing the next 743 // marking bitmap and getting it ready for the next cycle. During 744 // this time no other cycle can start. So, let's make sure that this 745 // is the case. 746 guarantee(!g1h->mark_in_progress(), "invariant"); 747 748 // clear the mark bitmap (no grey objects to start with). 749 // We need to do this in chunks and offer to yield in between 750 // each chunk. 751 HeapWord* start = _nextMarkBitMap->startWord(); 752 HeapWord* end = _nextMarkBitMap->endWord(); 753 HeapWord* cur = start; 754 size_t chunkSize = M; 755 while (cur < end) { 756 HeapWord* next = cur + chunkSize; 757 if (next > end) { 758 next = end; 759 } 760 MemRegion mr(cur,next); 761 _nextMarkBitMap->clearRange(mr); 762 cur = next; 763 do_yield_check(); 764 765 // Repeat the asserts from above. We'll do them as asserts here to 766 // minimize their overhead on the product. However, we'll have 767 // them as guarantees at the beginning / end of the bitmap 768 // clearing to get some checking in the product. 769 assert(cmThread()->during_cycle(), "invariant"); 770 assert(!g1h->mark_in_progress(), "invariant"); 771 } 772 773 // Repeat the asserts from above. 774 guarantee(cmThread()->during_cycle(), "invariant"); 775 guarantee(!g1h->mark_in_progress(), "invariant"); 776 } 777 778 class NoteStartOfMarkHRClosure: public HeapRegionClosure { 779 public: 780 bool doHeapRegion(HeapRegion* r) { 781 if (!r->continuesHumongous()) { 782 r->note_start_of_marking(true); 783 } 784 return false; 785 } 786 }; 787 788 void ConcurrentMark::checkpointRootsInitialPre() { 789 G1CollectedHeap* g1h = G1CollectedHeap::heap(); 790 G1CollectorPolicy* g1p = g1h->g1_policy(); 791 792 _has_aborted = false; 793 794 #ifndef PRODUCT 795 if (G1PrintReachableAtInitialMark) { 796 print_reachable("at-cycle-start", 797 VerifyOption_G1UsePrevMarking, true /* all */); 798 } 799 #endif 800 801 // Initialise marking structures. This has to be done in a STW phase. 802 reset(); 803 } 804 805 806 void ConcurrentMark::checkpointRootsInitialPost() { 807 G1CollectedHeap* g1h = G1CollectedHeap::heap(); 808 809 // If we force an overflow during remark, the remark operation will 810 // actually abort and we'll restart concurrent marking. If we always 811 // force an oveflow during remark we'll never actually complete the 812 // marking phase. So, we initilize this here, at the start of the 813 // cycle, so that at the remaining overflow number will decrease at 814 // every remark and we'll eventually not need to cause one. 815 force_overflow_stw()->init(); 816 817 // For each region note start of marking. 818 NoteStartOfMarkHRClosure startcl; 819 g1h->heap_region_iterate(&startcl); 820 821 // Start weak-reference discovery. 822 ReferenceProcessor* rp = g1h->ref_processor(); 823 rp->verify_no_references_recorded(); 824 rp->enable_discovery(); // enable ("weak") refs discovery 825 rp->setup_policy(false); // snapshot the soft ref policy to be used in this cycle 826 827 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); 828 // This is the start of the marking cycle, we're expected all 829 // threads to have SATB queues with active set to false. 830 satb_mq_set.set_active_all_threads(true, /* new active value */ 831 false /* expected_active */); 832 833 // update_g1_committed() will be called at the end of an evac pause 834 // when marking is on. So, it's also called at the end of the 835 // initial-mark pause to update the heap end, if the heap expands 836 // during it. No need to call it here. 837 } 838 839 /* 840 * Notice that in the next two methods, we actually leave the STS 841 * during the barrier sync and join it immediately afterwards. If we 842 * do not do this, the following deadlock can occur: one thread could 843 * be in the barrier sync code, waiting for the other thread to also 844 * sync up, whereas another one could be trying to yield, while also 845 * waiting for the other threads to sync up too. 846 * 847 * Note, however, that this code is also used during remark and in 848 * this case we should not attempt to leave / enter the STS, otherwise 849 * we'll either hit an asseert (debug / fastdebug) or deadlock 850 * (product). So we should only leave / enter the STS if we are 851 * operating concurrently. 852 * 853 * Because the thread that does the sync barrier has left the STS, it 854 * is possible to be suspended for a Full GC or an evacuation pause 855 * could occur. This is actually safe, since the entering the sync 856 * barrier is one of the last things do_marking_step() does, and it 857 * doesn't manipulate any data structures afterwards. 858 */ 859 860 void ConcurrentMark::enter_first_sync_barrier(int task_num) { 861 if (verbose_low()) { 862 gclog_or_tty->print_cr("[%d] entering first barrier", task_num); 863 } 864 865 if (concurrent()) { 866 ConcurrentGCThread::stsLeave(); 867 } 868 _first_overflow_barrier_sync.enter(); 869 if (concurrent()) { 870 ConcurrentGCThread::stsJoin(); 871 } 872 // at this point everyone should have synced up and not be doing any 873 // more work 874 875 if (verbose_low()) { 876 gclog_or_tty->print_cr("[%d] leaving first barrier", task_num); 877 } 878 879 // let task 0 do this 880 if (task_num == 0) { 881 // task 0 is responsible for clearing the global data structures 882 // We should be here because of an overflow. During STW we should 883 // not clear the overflow flag since we rely on it being true when 884 // we exit this method to abort the pause and restart concurent 885 // marking. 886 clear_marking_state(concurrent() /* clear_overflow */); 887 force_overflow()->update(); 888 889 if (PrintGC) { 890 gclog_or_tty->date_stamp(PrintGCDateStamps); 891 gclog_or_tty->stamp(PrintGCTimeStamps); 892 gclog_or_tty->print_cr("[GC concurrent-mark-reset-for-overflow]"); 893 } 894 } 895 896 // after this, each task should reset its own data structures then 897 // then go into the second barrier 898 } 899 900 void ConcurrentMark::enter_second_sync_barrier(int task_num) { 901 if (verbose_low()) { 902 gclog_or_tty->print_cr("[%d] entering second barrier", task_num); 903 } 904 905 if (concurrent()) { 906 ConcurrentGCThread::stsLeave(); 907 } 908 _second_overflow_barrier_sync.enter(); 909 if (concurrent()) { 910 ConcurrentGCThread::stsJoin(); 911 } 912 // at this point everything should be re-initialised and ready to go 913 914 if (verbose_low()) { 915 gclog_or_tty->print_cr("[%d] leaving second barrier", task_num); 916 } 917 } 918 919 #ifndef PRODUCT 920 void ForceOverflowSettings::init() { 921 _num_remaining = G1ConcMarkForceOverflow; 922 _force = false; 923 update(); 924 } 925 926 void ForceOverflowSettings::update() { 927 if (_num_remaining > 0) { 928 _num_remaining -= 1; 929 _force = true; 930 } else { 931 _force = false; 932 } 933 } 934 935 bool ForceOverflowSettings::should_force() { 936 if (_force) { 937 _force = false; 938 return true; 939 } else { 940 return false; 941 } 942 } 943 #endif // !PRODUCT 944 945 void ConcurrentMark::grayRoot(oop p) { 946 HeapWord* addr = (HeapWord*) p; 947 // We can't really check against _heap_start and _heap_end, since it 948 // is possible during an evacuation pause with piggy-backed 949 // initial-mark that the committed space is expanded during the 950 // pause without CM observing this change. So the assertions below 951 // is a bit conservative; but better than nothing. 952 assert(_g1h->g1_committed().contains(addr), 953 "address should be within the heap bounds"); 954 955 if (!_nextMarkBitMap->isMarked(addr)) { 956 _nextMarkBitMap->parMark(addr); 957 } 958 } 959 960 void ConcurrentMark::grayRegionIfNecessary(MemRegion mr) { 961 // The objects on the region have already been marked "in bulk" by 962 // the caller. We only need to decide whether to push the region on 963 // the region stack or not. 964 965 if (!concurrent_marking_in_progress() || !_should_gray_objects) { 966 // We're done with marking and waiting for remark. We do not need to 967 // push anything else on the region stack. 968 return; 969 } 970 971 HeapWord* finger = _finger; 972 973 if (verbose_low()) { 974 gclog_or_tty->print_cr("[global] attempting to push " 975 "region ["PTR_FORMAT", "PTR_FORMAT"), finger is at " 976 PTR_FORMAT, mr.start(), mr.end(), finger); 977 } 978 979 if (mr.start() < finger) { 980 // The finger is always heap region aligned and it is not possible 981 // for mr to span heap regions. 982 assert(mr.end() <= finger, "invariant"); 983 984 // Separated the asserts so that we know which one fires. 985 assert(mr.start() <= mr.end(), 986 "region boundaries should fall within the committed space"); 987 assert(_heap_start <= mr.start(), 988 "region boundaries should fall within the committed space"); 989 assert(mr.end() <= _heap_end, 990 "region boundaries should fall within the committed space"); 991 if (verbose_low()) { 992 gclog_or_tty->print_cr("[global] region ["PTR_FORMAT", "PTR_FORMAT") " 993 "below the finger, pushing it", 994 mr.start(), mr.end()); 995 } 996 997 if (!region_stack_push_lock_free(mr)) { 998 if (verbose_low()) { 999 gclog_or_tty->print_cr("[global] region stack has overflown."); 1000 } 1001 } 1002 } 1003 } 1004 1005 void ConcurrentMark::markAndGrayObjectIfNecessary(oop p) { 1006 // The object is not marked by the caller. We need to at least mark 1007 // it and maybe push in on the stack. 1008 1009 HeapWord* addr = (HeapWord*)p; 1010 if (!_nextMarkBitMap->isMarked(addr)) { 1011 // We definitely need to mark it, irrespective whether we bail out 1012 // because we're done with marking. 1013 if (_nextMarkBitMap->parMark(addr)) { 1014 if (!concurrent_marking_in_progress() || !_should_gray_objects) { 1015 // If we're done with concurrent marking and we're waiting for 1016 // remark, then we're not pushing anything on the stack. 1017 return; 1018 } 1019 1020 // No OrderAccess:store_load() is needed. It is implicit in the 1021 // CAS done in parMark(addr) above 1022 HeapWord* finger = _finger; 1023 1024 if (addr < finger) { 1025 if (!mark_stack_push(oop(addr))) { 1026 if (verbose_low()) { 1027 gclog_or_tty->print_cr("[global] global stack overflow " 1028 "during parMark"); 1029 } 1030 } 1031 } 1032 } 1033 } 1034 } 1035 1036 class CMConcurrentMarkingTask: public AbstractGangTask { 1037 private: 1038 ConcurrentMark* _cm; 1039 ConcurrentMarkThread* _cmt; 1040 1041 public: 1042 void work(int worker_i) { 1043 assert(Thread::current()->is_ConcurrentGC_thread(), 1044 "this should only be done by a conc GC thread"); 1045 ResourceMark rm; 1046 1047 double start_vtime = os::elapsedVTime(); 1048 1049 ConcurrentGCThread::stsJoin(); 1050 1051 assert((size_t) worker_i < _cm->active_tasks(), "invariant"); 1052 CMTask* the_task = _cm->task(worker_i); 1053 the_task->record_start_time(); 1054 if (!_cm->has_aborted()) { 1055 do { 1056 double start_vtime_sec = os::elapsedVTime(); 1057 double start_time_sec = os::elapsedTime(); 1058 double mark_step_duration_ms = G1ConcMarkStepDurationMillis; 1059 1060 the_task->do_marking_step(mark_step_duration_ms, 1061 true /* do_stealing */, 1062 true /* do_termination */); 1063 1064 double end_time_sec = os::elapsedTime(); 1065 double end_vtime_sec = os::elapsedVTime(); 1066 double elapsed_vtime_sec = end_vtime_sec - start_vtime_sec; 1067 double elapsed_time_sec = end_time_sec - start_time_sec; 1068 _cm->clear_has_overflown(); 1069 1070 bool ret = _cm->do_yield_check(worker_i); 1071 1072 jlong sleep_time_ms; 1073 if (!_cm->has_aborted() && the_task->has_aborted()) { 1074 sleep_time_ms = 1075 (jlong) (elapsed_vtime_sec * _cm->sleep_factor() * 1000.0); 1076 ConcurrentGCThread::stsLeave(); 1077 os::sleep(Thread::current(), sleep_time_ms, false); 1078 ConcurrentGCThread::stsJoin(); 1079 } 1080 double end_time2_sec = os::elapsedTime(); 1081 double elapsed_time2_sec = end_time2_sec - start_time_sec; 1082 1083 #if 0 1084 gclog_or_tty->print_cr("CM: elapsed %1.4lf ms, sleep %1.4lf ms, " 1085 "overhead %1.4lf", 1086 elapsed_vtime_sec * 1000.0, (double) sleep_time_ms, 1087 the_task->conc_overhead(os::elapsedTime()) * 8.0); 1088 gclog_or_tty->print_cr("elapsed time %1.4lf ms, time 2: %1.4lf ms", 1089 elapsed_time_sec * 1000.0, elapsed_time2_sec * 1000.0); 1090 #endif 1091 } while (!_cm->has_aborted() && the_task->has_aborted()); 1092 } 1093 the_task->record_end_time(); 1094 guarantee(!the_task->has_aborted() || _cm->has_aborted(), "invariant"); 1095 1096 ConcurrentGCThread::stsLeave(); 1097 1098 double end_vtime = os::elapsedVTime(); 1099 _cm->update_accum_task_vtime(worker_i, end_vtime - start_vtime); 1100 } 1101 1102 CMConcurrentMarkingTask(ConcurrentMark* cm, 1103 ConcurrentMarkThread* cmt) : 1104 AbstractGangTask("Concurrent Mark"), _cm(cm), _cmt(cmt) { } 1105 1106 ~CMConcurrentMarkingTask() { } 1107 }; 1108 1109 void ConcurrentMark::markFromRoots() { 1110 // we might be tempted to assert that: 1111 // assert(asynch == !SafepointSynchronize::is_at_safepoint(), 1112 // "inconsistent argument?"); 1113 // However that wouldn't be right, because it's possible that 1114 // a safepoint is indeed in progress as a younger generation 1115 // stop-the-world GC happens even as we mark in this generation. 1116 1117 _restart_for_overflow = false; 1118 1119 size_t active_workers = MAX2((size_t) 1, parallel_marking_threads()); 1120 force_overflow_conc()->init(); 1121 set_phase(active_workers, true /* concurrent */); 1122 1123 CMConcurrentMarkingTask markingTask(this, cmThread()); 1124 if (parallel_marking_threads() > 0) { 1125 _parallel_workers->run_task(&markingTask); 1126 } else { 1127 markingTask.work(0); 1128 } 1129 print_stats(); 1130 } 1131 1132 void ConcurrentMark::checkpointRootsFinal(bool clear_all_soft_refs) { 1133 // world is stopped at this checkpoint 1134 assert(SafepointSynchronize::is_at_safepoint(), 1135 "world should be stopped"); 1136 G1CollectedHeap* g1h = G1CollectedHeap::heap(); 1137 1138 // If a full collection has happened, we shouldn't do this. 1139 if (has_aborted()) { 1140 g1h->set_marking_complete(); // So bitmap clearing isn't confused 1141 return; 1142 } 1143 1144 SvcGCMarker sgcm(SvcGCMarker::OTHER); 1145 1146 if (VerifyDuringGC) { 1147 HandleMark hm; // handle scope 1148 gclog_or_tty->print(" VerifyDuringGC:(before)"); 1149 Universe::heap()->prepare_for_verify(); 1150 Universe::verify(/* allow dirty */ true, 1151 /* silent */ false, 1152 /* option */ VerifyOption_G1UsePrevMarking); 1153 } 1154 1155 G1CollectorPolicy* g1p = g1h->g1_policy(); 1156 g1p->record_concurrent_mark_remark_start(); 1157 1158 double start = os::elapsedTime(); 1159 1160 checkpointRootsFinalWork(); 1161 1162 double mark_work_end = os::elapsedTime(); 1163 1164 weakRefsWork(clear_all_soft_refs); 1165 1166 if (has_overflown()) { 1167 // Oops. We overflowed. Restart concurrent marking. 1168 _restart_for_overflow = true; 1169 // Clear the flag. We do not need it any more. 1170 clear_has_overflown(); 1171 if (G1TraceMarkStackOverflow) { 1172 gclog_or_tty->print_cr("\nRemark led to restart for overflow."); 1173 } 1174 } else { 1175 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); 1176 // We're done with marking. 1177 // This is the end of the marking cycle, we're expected all 1178 // threads to have SATB queues with active set to true. 1179 satb_mq_set.set_active_all_threads(false, /* new active value */ 1180 true /* expected_active */); 1181 1182 if (VerifyDuringGC) { 1183 HandleMark hm; // handle scope 1184 gclog_or_tty->print(" VerifyDuringGC:(after)"); 1185 Universe::heap()->prepare_for_verify(); 1186 Universe::verify(/* allow dirty */ true, 1187 /* silent */ false, 1188 /* option */ VerifyOption_G1UseNextMarking); 1189 } 1190 assert(!restart_for_overflow(), "sanity"); 1191 } 1192 1193 // Reset the marking state if marking completed 1194 if (!restart_for_overflow()) { 1195 set_non_marking_state(); 1196 } 1197 1198 #if VERIFY_OBJS_PROCESSED 1199 _scan_obj_cl.objs_processed = 0; 1200 ThreadLocalObjQueue::objs_enqueued = 0; 1201 #endif 1202 1203 // Statistics 1204 double now = os::elapsedTime(); 1205 _remark_mark_times.add((mark_work_end - start) * 1000.0); 1206 _remark_weak_ref_times.add((now - mark_work_end) * 1000.0); 1207 _remark_times.add((now - start) * 1000.0); 1208 1209 g1p->record_concurrent_mark_remark_end(); 1210 } 1211 1212 #define CARD_BM_TEST_MODE 0 1213 1214 class CalcLiveObjectsClosure: public HeapRegionClosure { 1215 1216 CMBitMapRO* _bm; 1217 ConcurrentMark* _cm; 1218 bool _changed; 1219 bool _yield; 1220 size_t _words_done; 1221 size_t _tot_live; 1222 size_t _tot_used; 1223 size_t _regions_done; 1224 double _start_vtime_sec; 1225 1226 BitMap* _region_bm; 1227 BitMap* _card_bm; 1228 intptr_t _bottom_card_num; 1229 bool _final; 1230 1231 void mark_card_num_range(intptr_t start_card_num, intptr_t last_card_num) { 1232 for (intptr_t i = start_card_num; i <= last_card_num; i++) { 1233 #if CARD_BM_TEST_MODE 1234 guarantee(_card_bm->at(i - _bottom_card_num), "Should already be set."); 1235 #else 1236 _card_bm->par_at_put(i - _bottom_card_num, 1); 1237 #endif 1238 } 1239 } 1240 1241 public: 1242 CalcLiveObjectsClosure(bool final, 1243 CMBitMapRO *bm, ConcurrentMark *cm, 1244 BitMap* region_bm, BitMap* card_bm) : 1245 _bm(bm), _cm(cm), _changed(false), _yield(true), 1246 _words_done(0), _tot_live(0), _tot_used(0), 1247 _region_bm(region_bm), _card_bm(card_bm),_final(final), 1248 _regions_done(0), _start_vtime_sec(0.0) 1249 { 1250 _bottom_card_num = 1251 intptr_t(uintptr_t(G1CollectedHeap::heap()->reserved_region().start()) >> 1252 CardTableModRefBS::card_shift); 1253 } 1254 1255 // It takes a region that's not empty (i.e., it has at least one 1256 // live object in it and sets its corresponding bit on the region 1257 // bitmap to 1. If the region is "starts humongous" it will also set 1258 // to 1 the bits on the region bitmap that correspond to its 1259 // associated "continues humongous" regions. 1260 void set_bit_for_region(HeapRegion* hr) { 1261 assert(!hr->continuesHumongous(), "should have filtered those out"); 1262 1263 size_t index = hr->hrs_index(); 1264 if (!hr->startsHumongous()) { 1265 // Normal (non-humongous) case: just set the bit. 1266 _region_bm->par_at_put((BitMap::idx_t) index, true); 1267 } else { 1268 // Starts humongous case: calculate how many regions are part of 1269 // this humongous region and then set the bit range. It might 1270 // have been a bit more efficient to look at the object that 1271 // spans these humongous regions to calculate their number from 1272 // the object's size. However, it's a good idea to calculate 1273 // this based on the metadata itself, and not the region 1274 // contents, so that this code is not aware of what goes into 1275 // the humongous regions (in case this changes in the future). 1276 G1CollectedHeap* g1h = G1CollectedHeap::heap(); 1277 size_t end_index = index + 1; 1278 while (end_index < g1h->n_regions()) { 1279 HeapRegion* chr = g1h->region_at(end_index); 1280 if (!chr->continuesHumongous()) break; 1281 end_index += 1; 1282 } 1283 _region_bm->par_at_put_range((BitMap::idx_t) index, 1284 (BitMap::idx_t) end_index, true); 1285 } 1286 } 1287 1288 bool doHeapRegion(HeapRegion* hr) { 1289 if (!_final && _regions_done == 0) { 1290 _start_vtime_sec = os::elapsedVTime(); 1291 } 1292 1293 if (hr->continuesHumongous()) { 1294 // We will ignore these here and process them when their 1295 // associated "starts humongous" region is processed (see 1296 // set_bit_for_heap_region()). Note that we cannot rely on their 1297 // associated "starts humongous" region to have their bit set to 1298 // 1 since, due to the region chunking in the parallel region 1299 // iteration, a "continues humongous" region might be visited 1300 // before its associated "starts humongous". 1301 return false; 1302 } 1303 1304 HeapWord* nextTop = hr->next_top_at_mark_start(); 1305 HeapWord* start = hr->top_at_conc_mark_count(); 1306 assert(hr->bottom() <= start && start <= hr->end() && 1307 hr->bottom() <= nextTop && nextTop <= hr->end() && 1308 start <= nextTop, 1309 "Preconditions."); 1310 // Otherwise, record the number of word's we'll examine. 1311 size_t words_done = (nextTop - start); 1312 // Find the first marked object at or after "start". 1313 start = _bm->getNextMarkedWordAddress(start, nextTop); 1314 size_t marked_bytes = 0; 1315 1316 // Below, the term "card num" means the result of shifting an address 1317 // by the card shift -- address 0 corresponds to card number 0. One 1318 // must subtract the card num of the bottom of the heap to obtain a 1319 // card table index. 1320 // The first card num of the sequence of live cards currently being 1321 // constructed. -1 ==> no sequence. 1322 intptr_t start_card_num = -1; 1323 // The last card num of the sequence of live cards currently being 1324 // constructed. -1 ==> no sequence. 1325 intptr_t last_card_num = -1; 1326 1327 while (start < nextTop) { 1328 if (_yield && _cm->do_yield_check()) { 1329 // We yielded. It might be for a full collection, in which case 1330 // all bets are off; terminate the traversal. 1331 if (_cm->has_aborted()) { 1332 _changed = false; 1333 return true; 1334 } else { 1335 // Otherwise, it might be a collection pause, and the region 1336 // we're looking at might be in the collection set. We'll 1337 // abandon this region. 1338 return false; 1339 } 1340 } 1341 oop obj = oop(start); 1342 int obj_sz = obj->size(); 1343 // The card num of the start of the current object. 1344 intptr_t obj_card_num = 1345 intptr_t(uintptr_t(start) >> CardTableModRefBS::card_shift); 1346 1347 HeapWord* obj_last = start + obj_sz - 1; 1348 intptr_t obj_last_card_num = 1349 intptr_t(uintptr_t(obj_last) >> CardTableModRefBS::card_shift); 1350 1351 if (obj_card_num != last_card_num) { 1352 if (start_card_num == -1) { 1353 assert(last_card_num == -1, "Both or neither."); 1354 start_card_num = obj_card_num; 1355 } else { 1356 assert(last_card_num != -1, "Both or neither."); 1357 assert(obj_card_num >= last_card_num, "Inv"); 1358 if ((obj_card_num - last_card_num) > 1) { 1359 // Mark the last run, and start a new one. 1360 mark_card_num_range(start_card_num, last_card_num); 1361 start_card_num = obj_card_num; 1362 } 1363 } 1364 #if CARD_BM_TEST_MODE 1365 /* 1366 gclog_or_tty->print_cr("Setting bits from %d/%d.", 1367 obj_card_num - _bottom_card_num, 1368 obj_last_card_num - _bottom_card_num); 1369 */ 1370 for (intptr_t j = obj_card_num; j <= obj_last_card_num; j++) { 1371 _card_bm->par_at_put(j - _bottom_card_num, 1); 1372 } 1373 #endif 1374 } 1375 // In any case, we set the last card num. 1376 last_card_num = obj_last_card_num; 1377 1378 marked_bytes += (size_t)obj_sz * HeapWordSize; 1379 // Find the next marked object after this one. 1380 start = _bm->getNextMarkedWordAddress(start + 1, nextTop); 1381 _changed = true; 1382 } 1383 // Handle the last range, if any. 1384 if (start_card_num != -1) { 1385 mark_card_num_range(start_card_num, last_card_num); 1386 } 1387 if (_final) { 1388 // Mark the allocated-since-marking portion... 1389 HeapWord* tp = hr->top(); 1390 if (nextTop < tp) { 1391 start_card_num = 1392 intptr_t(uintptr_t(nextTop) >> CardTableModRefBS::card_shift); 1393 last_card_num = 1394 intptr_t(uintptr_t(tp) >> CardTableModRefBS::card_shift); 1395 mark_card_num_range(start_card_num, last_card_num); 1396 // This definitely means the region has live objects. 1397 set_bit_for_region(hr); 1398 } 1399 } 1400 1401 hr->add_to_marked_bytes(marked_bytes); 1402 // Update the live region bitmap. 1403 if (marked_bytes > 0) { 1404 set_bit_for_region(hr); 1405 } 1406 hr->set_top_at_conc_mark_count(nextTop); 1407 _tot_live += hr->next_live_bytes(); 1408 _tot_used += hr->used(); 1409 _words_done = words_done; 1410 1411 if (!_final) { 1412 ++_regions_done; 1413 if (_regions_done % 10 == 0) { 1414 double end_vtime_sec = os::elapsedVTime(); 1415 double elapsed_vtime_sec = end_vtime_sec - _start_vtime_sec; 1416 if (elapsed_vtime_sec > (10.0 / 1000.0)) { 1417 jlong sleep_time_ms = 1418 (jlong) (elapsed_vtime_sec * _cm->cleanup_sleep_factor() * 1000.0); 1419 os::sleep(Thread::current(), sleep_time_ms, false); 1420 _start_vtime_sec = end_vtime_sec; 1421 } 1422 } 1423 } 1424 1425 return false; 1426 } 1427 1428 bool changed() { return _changed; } 1429 void reset() { _changed = false; _words_done = 0; } 1430 void no_yield() { _yield = false; } 1431 size_t words_done() { return _words_done; } 1432 size_t tot_live() { return _tot_live; } 1433 size_t tot_used() { return _tot_used; } 1434 }; 1435 1436 1437 void ConcurrentMark::calcDesiredRegions() { 1438 _region_bm.clear(); 1439 _card_bm.clear(); 1440 CalcLiveObjectsClosure calccl(false /*final*/, 1441 nextMarkBitMap(), this, 1442 &_region_bm, &_card_bm); 1443 G1CollectedHeap *g1h = G1CollectedHeap::heap(); 1444 g1h->heap_region_iterate(&calccl); 1445 1446 do { 1447 calccl.reset(); 1448 g1h->heap_region_iterate(&calccl); 1449 } while (calccl.changed()); 1450 } 1451 1452 class G1ParFinalCountTask: public AbstractGangTask { 1453 protected: 1454 G1CollectedHeap* _g1h; 1455 CMBitMap* _bm; 1456 size_t _n_workers; 1457 size_t *_live_bytes; 1458 size_t *_used_bytes; 1459 BitMap* _region_bm; 1460 BitMap* _card_bm; 1461 public: 1462 G1ParFinalCountTask(G1CollectedHeap* g1h, CMBitMap* bm, 1463 BitMap* region_bm, BitMap* card_bm) 1464 : AbstractGangTask("G1 final counting"), _g1h(g1h), 1465 _bm(bm), _region_bm(region_bm), _card_bm(card_bm) { 1466 if (ParallelGCThreads > 0) { 1467 _n_workers = _g1h->workers()->total_workers(); 1468 } else { 1469 _n_workers = 1; 1470 } 1471 _live_bytes = NEW_C_HEAP_ARRAY(size_t, _n_workers); 1472 _used_bytes = NEW_C_HEAP_ARRAY(size_t, _n_workers); 1473 } 1474 1475 ~G1ParFinalCountTask() { 1476 FREE_C_HEAP_ARRAY(size_t, _live_bytes); 1477 FREE_C_HEAP_ARRAY(size_t, _used_bytes); 1478 } 1479 1480 void work(int i) { 1481 CalcLiveObjectsClosure calccl(true /*final*/, 1482 _bm, _g1h->concurrent_mark(), 1483 _region_bm, _card_bm); 1484 calccl.no_yield(); 1485 if (G1CollectedHeap::use_parallel_gc_threads()) { 1486 _g1h->heap_region_par_iterate_chunked(&calccl, i, 1487 HeapRegion::FinalCountClaimValue); 1488 } else { 1489 _g1h->heap_region_iterate(&calccl); 1490 } 1491 assert(calccl.complete(), "Shouldn't have yielded!"); 1492 1493 assert((size_t) i < _n_workers, "invariant"); 1494 _live_bytes[i] = calccl.tot_live(); 1495 _used_bytes[i] = calccl.tot_used(); 1496 } 1497 size_t live_bytes() { 1498 size_t live_bytes = 0; 1499 for (size_t i = 0; i < _n_workers; ++i) 1500 live_bytes += _live_bytes[i]; 1501 return live_bytes; 1502 } 1503 size_t used_bytes() { 1504 size_t used_bytes = 0; 1505 for (size_t i = 0; i < _n_workers; ++i) 1506 used_bytes += _used_bytes[i]; 1507 return used_bytes; 1508 } 1509 }; 1510 1511 class G1ParNoteEndTask; 1512 1513 class G1NoteEndOfConcMarkClosure : public HeapRegionClosure { 1514 G1CollectedHeap* _g1; 1515 int _worker_num; 1516 size_t _max_live_bytes; 1517 size_t _regions_claimed; 1518 size_t _freed_bytes; 1519 FreeRegionList* _local_cleanup_list; 1520 HumongousRegionSet* _humongous_proxy_set; 1521 HRRSCleanupTask* _hrrs_cleanup_task; 1522 double _claimed_region_time; 1523 double _max_region_time; 1524 1525 public: 1526 G1NoteEndOfConcMarkClosure(G1CollectedHeap* g1, 1527 int worker_num, 1528 FreeRegionList* local_cleanup_list, 1529 HumongousRegionSet* humongous_proxy_set, 1530 HRRSCleanupTask* hrrs_cleanup_task); 1531 size_t freed_bytes() { return _freed_bytes; } 1532 1533 bool doHeapRegion(HeapRegion *r); 1534 1535 size_t max_live_bytes() { return _max_live_bytes; } 1536 size_t regions_claimed() { return _regions_claimed; } 1537 double claimed_region_time_sec() { return _claimed_region_time; } 1538 double max_region_time_sec() { return _max_region_time; } 1539 }; 1540 1541 class G1ParNoteEndTask: public AbstractGangTask { 1542 friend class G1NoteEndOfConcMarkClosure; 1543 1544 protected: 1545 G1CollectedHeap* _g1h; 1546 size_t _max_live_bytes; 1547 size_t _freed_bytes; 1548 FreeRegionList* _cleanup_list; 1549 1550 public: 1551 G1ParNoteEndTask(G1CollectedHeap* g1h, 1552 FreeRegionList* cleanup_list) : 1553 AbstractGangTask("G1 note end"), _g1h(g1h), 1554 _max_live_bytes(0), _freed_bytes(0), _cleanup_list(cleanup_list) { } 1555 1556 void work(int i) { 1557 double start = os::elapsedTime(); 1558 FreeRegionList local_cleanup_list("Local Cleanup List"); 1559 HumongousRegionSet humongous_proxy_set("Local Cleanup Humongous Proxy Set"); 1560 HRRSCleanupTask hrrs_cleanup_task; 1561 G1NoteEndOfConcMarkClosure g1_note_end(_g1h, i, &local_cleanup_list, 1562 &humongous_proxy_set, 1563 &hrrs_cleanup_task); 1564 if (G1CollectedHeap::use_parallel_gc_threads()) { 1565 _g1h->heap_region_par_iterate_chunked(&g1_note_end, i, 1566 HeapRegion::NoteEndClaimValue); 1567 } else { 1568 _g1h->heap_region_iterate(&g1_note_end); 1569 } 1570 assert(g1_note_end.complete(), "Shouldn't have yielded!"); 1571 1572 // Now update the lists 1573 _g1h->update_sets_after_freeing_regions(g1_note_end.freed_bytes(), 1574 NULL /* free_list */, 1575 &humongous_proxy_set, 1576 true /* par */); 1577 { 1578 MutexLockerEx x(ParGCRareEvent_lock, Mutex::_no_safepoint_check_flag); 1579 _max_live_bytes += g1_note_end.max_live_bytes(); 1580 _freed_bytes += g1_note_end.freed_bytes(); 1581 1582 // If we iterate over the global cleanup list at the end of 1583 // cleanup to do this printing we will not guarantee to only 1584 // generate output for the newly-reclaimed regions (the list 1585 // might not be empty at the beginning of cleanup; we might 1586 // still be working on its previous contents). So we do the 1587 // printing here, before we append the new regions to the global 1588 // cleanup list. 1589 1590 G1HRPrinter* hr_printer = _g1h->hr_printer(); 1591 if (hr_printer->is_active()) { 1592 HeapRegionLinkedListIterator iter(&local_cleanup_list); 1593 while (iter.more_available()) { 1594 HeapRegion* hr = iter.get_next(); 1595 hr_printer->cleanup(hr); 1596 } 1597 } 1598 1599 _cleanup_list->add_as_tail(&local_cleanup_list); 1600 assert(local_cleanup_list.is_empty(), "post-condition"); 1601 1602 HeapRegionRemSet::finish_cleanup_task(&hrrs_cleanup_task); 1603 } 1604 double end = os::elapsedTime(); 1605 if (G1PrintParCleanupStats) { 1606 gclog_or_tty->print(" Worker thread %d [%8.3f..%8.3f = %8.3f ms] " 1607 "claimed %d regions (tot = %8.3f ms, max = %8.3f ms).\n", 1608 i, start, end, (end-start)*1000.0, 1609 g1_note_end.regions_claimed(), 1610 g1_note_end.claimed_region_time_sec()*1000.0, 1611 g1_note_end.max_region_time_sec()*1000.0); 1612 } 1613 } 1614 size_t max_live_bytes() { return _max_live_bytes; } 1615 size_t freed_bytes() { return _freed_bytes; } 1616 }; 1617 1618 class G1ParScrubRemSetTask: public AbstractGangTask { 1619 protected: 1620 G1RemSet* _g1rs; 1621 BitMap* _region_bm; 1622 BitMap* _card_bm; 1623 public: 1624 G1ParScrubRemSetTask(G1CollectedHeap* g1h, 1625 BitMap* region_bm, BitMap* card_bm) : 1626 AbstractGangTask("G1 ScrubRS"), _g1rs(g1h->g1_rem_set()), 1627 _region_bm(region_bm), _card_bm(card_bm) 1628 {} 1629 1630 void work(int i) { 1631 if (G1CollectedHeap::use_parallel_gc_threads()) { 1632 _g1rs->scrub_par(_region_bm, _card_bm, i, 1633 HeapRegion::ScrubRemSetClaimValue); 1634 } else { 1635 _g1rs->scrub(_region_bm, _card_bm); 1636 } 1637 } 1638 1639 }; 1640 1641 G1NoteEndOfConcMarkClosure:: 1642 G1NoteEndOfConcMarkClosure(G1CollectedHeap* g1, 1643 int worker_num, 1644 FreeRegionList* local_cleanup_list, 1645 HumongousRegionSet* humongous_proxy_set, 1646 HRRSCleanupTask* hrrs_cleanup_task) 1647 : _g1(g1), _worker_num(worker_num), 1648 _max_live_bytes(0), _regions_claimed(0), 1649 _freed_bytes(0), 1650 _claimed_region_time(0.0), _max_region_time(0.0), 1651 _local_cleanup_list(local_cleanup_list), 1652 _humongous_proxy_set(humongous_proxy_set), 1653 _hrrs_cleanup_task(hrrs_cleanup_task) { } 1654 1655 bool G1NoteEndOfConcMarkClosure::doHeapRegion(HeapRegion *hr) { 1656 // We use a claim value of zero here because all regions 1657 // were claimed with value 1 in the FinalCount task. 1658 hr->reset_gc_time_stamp(); 1659 if (!hr->continuesHumongous()) { 1660 double start = os::elapsedTime(); 1661 _regions_claimed++; 1662 hr->note_end_of_marking(); 1663 _max_live_bytes += hr->max_live_bytes(); 1664 _g1->free_region_if_empty(hr, 1665 &_freed_bytes, 1666 _local_cleanup_list, 1667 _humongous_proxy_set, 1668 _hrrs_cleanup_task, 1669 true /* par */); 1670 double region_time = (os::elapsedTime() - start); 1671 _claimed_region_time += region_time; 1672 if (region_time > _max_region_time) { 1673 _max_region_time = region_time; 1674 } 1675 } 1676 return false; 1677 } 1678 1679 void ConcurrentMark::cleanup() { 1680 // world is stopped at this checkpoint 1681 assert(SafepointSynchronize::is_at_safepoint(), 1682 "world should be stopped"); 1683 G1CollectedHeap* g1h = G1CollectedHeap::heap(); 1684 1685 // If a full collection has happened, we shouldn't do this. 1686 if (has_aborted()) { 1687 g1h->set_marking_complete(); // So bitmap clearing isn't confused 1688 return; 1689 } 1690 1691 g1h->verify_region_sets_optional(); 1692 1693 if (VerifyDuringGC) { 1694 HandleMark hm; // handle scope 1695 gclog_or_tty->print(" VerifyDuringGC:(before)"); 1696 Universe::heap()->prepare_for_verify(); 1697 Universe::verify(/* allow dirty */ true, 1698 /* silent */ false, 1699 /* option */ VerifyOption_G1UsePrevMarking); 1700 } 1701 1702 G1CollectorPolicy* g1p = G1CollectedHeap::heap()->g1_policy(); 1703 g1p->record_concurrent_mark_cleanup_start(); 1704 1705 double start = os::elapsedTime(); 1706 1707 HeapRegionRemSet::reset_for_cleanup_tasks(); 1708 1709 // Do counting once more with the world stopped for good measure. 1710 G1ParFinalCountTask g1_par_count_task(g1h, nextMarkBitMap(), 1711 &_region_bm, &_card_bm); 1712 if (G1CollectedHeap::use_parallel_gc_threads()) { 1713 assert(g1h->check_heap_region_claim_values( 1714 HeapRegion::InitialClaimValue), 1715 "sanity check"); 1716 1717 int n_workers = g1h->workers()->total_workers(); 1718 g1h->set_par_threads(n_workers); 1719 g1h->workers()->run_task(&g1_par_count_task); 1720 g1h->set_par_threads(0); 1721 1722 assert(g1h->check_heap_region_claim_values( 1723 HeapRegion::FinalCountClaimValue), 1724 "sanity check"); 1725 } else { 1726 g1_par_count_task.work(0); 1727 } 1728 1729 size_t known_garbage_bytes = 1730 g1_par_count_task.used_bytes() - g1_par_count_task.live_bytes(); 1731 g1p->set_known_garbage_bytes(known_garbage_bytes); 1732 1733 size_t start_used_bytes = g1h->used(); 1734 _at_least_one_mark_complete = true; 1735 g1h->set_marking_complete(); 1736 1737 ergo_verbose4(ErgoConcCycles, 1738 "finish cleanup", 1739 ergo_format_byte("occupancy") 1740 ergo_format_byte("capacity") 1741 ergo_format_byte_perc("known garbage"), 1742 start_used_bytes, g1h->capacity(), 1743 known_garbage_bytes, 1744 ((double) known_garbage_bytes / (double) g1h->capacity()) * 100.0); 1745 1746 double count_end = os::elapsedTime(); 1747 double this_final_counting_time = (count_end - start); 1748 if (G1PrintParCleanupStats) { 1749 gclog_or_tty->print_cr("Cleanup:"); 1750 gclog_or_tty->print_cr(" Finalize counting: %8.3f ms", 1751 this_final_counting_time*1000.0); 1752 } 1753 _total_counting_time += this_final_counting_time; 1754 1755 if (G1PrintRegionLivenessInfo) { 1756 G1PrintRegionLivenessInfoClosure cl(gclog_or_tty, "Post-Marking"); 1757 _g1h->heap_region_iterate(&cl); 1758 } 1759 1760 // Install newly created mark bitMap as "prev". 1761 swapMarkBitMaps(); 1762 1763 g1h->reset_gc_time_stamp(); 1764 1765 // Note end of marking in all heap regions. 1766 double note_end_start = os::elapsedTime(); 1767 G1ParNoteEndTask g1_par_note_end_task(g1h, &_cleanup_list); 1768 if (G1CollectedHeap::use_parallel_gc_threads()) { 1769 int n_workers = g1h->workers()->total_workers(); 1770 g1h->set_par_threads(n_workers); 1771 g1h->workers()->run_task(&g1_par_note_end_task); 1772 g1h->set_par_threads(0); 1773 1774 assert(g1h->check_heap_region_claim_values(HeapRegion::NoteEndClaimValue), 1775 "sanity check"); 1776 } else { 1777 g1_par_note_end_task.work(0); 1778 } 1779 1780 if (!cleanup_list_is_empty()) { 1781 // The cleanup list is not empty, so we'll have to process it 1782 // concurrently. Notify anyone else that might be wanting free 1783 // regions that there will be more free regions coming soon. 1784 g1h->set_free_regions_coming(); 1785 } 1786 double note_end_end = os::elapsedTime(); 1787 if (G1PrintParCleanupStats) { 1788 gclog_or_tty->print_cr(" note end of marking: %8.3f ms.", 1789 (note_end_end - note_end_start)*1000.0); 1790 } 1791 1792 1793 // call below, since it affects the metric by which we sort the heap 1794 // regions. 1795 if (G1ScrubRemSets) { 1796 double rs_scrub_start = os::elapsedTime(); 1797 G1ParScrubRemSetTask g1_par_scrub_rs_task(g1h, &_region_bm, &_card_bm); 1798 if (G1CollectedHeap::use_parallel_gc_threads()) { 1799 int n_workers = g1h->workers()->total_workers(); 1800 g1h->set_par_threads(n_workers); 1801 g1h->workers()->run_task(&g1_par_scrub_rs_task); 1802 g1h->set_par_threads(0); 1803 1804 assert(g1h->check_heap_region_claim_values( 1805 HeapRegion::ScrubRemSetClaimValue), 1806 "sanity check"); 1807 } else { 1808 g1_par_scrub_rs_task.work(0); 1809 } 1810 1811 double rs_scrub_end = os::elapsedTime(); 1812 double this_rs_scrub_time = (rs_scrub_end - rs_scrub_start); 1813 _total_rs_scrub_time += this_rs_scrub_time; 1814 } 1815 1816 // this will also free any regions totally full of garbage objects, 1817 // and sort the regions. 1818 g1h->g1_policy()->record_concurrent_mark_cleanup_end( 1819 g1_par_note_end_task.freed_bytes(), 1820 g1_par_note_end_task.max_live_bytes()); 1821 1822 // Statistics. 1823 double end = os::elapsedTime(); 1824 _cleanup_times.add((end - start) * 1000.0); 1825 1826 // G1CollectedHeap::heap()->print(); 1827 // gclog_or_tty->print_cr("HEAP GC TIME STAMP : %d", 1828 // G1CollectedHeap::heap()->get_gc_time_stamp()); 1829 1830 if (PrintGC || PrintGCDetails) { 1831 g1h->print_size_transition(gclog_or_tty, 1832 start_used_bytes, 1833 g1h->used(), 1834 g1h->capacity()); 1835 } 1836 1837 size_t cleaned_up_bytes = start_used_bytes - g1h->used(); 1838 g1p->decrease_known_garbage_bytes(cleaned_up_bytes); 1839 1840 // We need to make this be a "collection" so any collection pause that 1841 // races with it goes around and waits for completeCleanup to finish. 1842 g1h->increment_total_collections(); 1843 1844 if (VerifyDuringGC) { 1845 HandleMark hm; // handle scope 1846 gclog_or_tty->print(" VerifyDuringGC:(after)"); 1847 Universe::heap()->prepare_for_verify(); 1848 Universe::verify(/* allow dirty */ true, 1849 /* silent */ false, 1850 /* option */ VerifyOption_G1UsePrevMarking); 1851 } 1852 1853 g1h->verify_region_sets_optional(); 1854 } 1855 1856 void ConcurrentMark::completeCleanup() { 1857 if (has_aborted()) return; 1858 1859 G1CollectedHeap* g1h = G1CollectedHeap::heap(); 1860 1861 _cleanup_list.verify_optional(); 1862 FreeRegionList tmp_free_list("Tmp Free List"); 1863 1864 if (G1ConcRegionFreeingVerbose) { 1865 gclog_or_tty->print_cr("G1ConcRegionFreeing [complete cleanup] : " 1866 "cleanup list has "SIZE_FORMAT" entries", 1867 _cleanup_list.length()); 1868 } 1869 1870 // Noone else should be accessing the _cleanup_list at this point, 1871 // so it's not necessary to take any locks 1872 while (!_cleanup_list.is_empty()) { 1873 HeapRegion* hr = _cleanup_list.remove_head(); 1874 assert(hr != NULL, "the list was not empty"); 1875 hr->par_clear(); 1876 tmp_free_list.add_as_tail(hr); 1877 1878 // Instead of adding one region at a time to the secondary_free_list, 1879 // we accumulate them in the local list and move them a few at a 1880 // time. This also cuts down on the number of notify_all() calls 1881 // we do during this process. We'll also append the local list when 1882 // _cleanup_list is empty (which means we just removed the last 1883 // region from the _cleanup_list). 1884 if ((tmp_free_list.length() % G1SecondaryFreeListAppendLength == 0) || 1885 _cleanup_list.is_empty()) { 1886 if (G1ConcRegionFreeingVerbose) { 1887 gclog_or_tty->print_cr("G1ConcRegionFreeing [complete cleanup] : " 1888 "appending "SIZE_FORMAT" entries to the " 1889 "secondary_free_list, clean list still has " 1890 SIZE_FORMAT" entries", 1891 tmp_free_list.length(), 1892 _cleanup_list.length()); 1893 } 1894 1895 { 1896 MutexLockerEx x(SecondaryFreeList_lock, Mutex::_no_safepoint_check_flag); 1897 g1h->secondary_free_list_add_as_tail(&tmp_free_list); 1898 SecondaryFreeList_lock->notify_all(); 1899 } 1900 1901 if (G1StressConcRegionFreeing) { 1902 for (uintx i = 0; i < G1StressConcRegionFreeingDelayMillis; ++i) { 1903 os::sleep(Thread::current(), (jlong) 1, false); 1904 } 1905 } 1906 } 1907 } 1908 assert(tmp_free_list.is_empty(), "post-condition"); 1909 } 1910 1911 // Support closures for reference procssing in G1 1912 1913 bool G1CMIsAliveClosure::do_object_b(oop obj) { 1914 HeapWord* addr = (HeapWord*)obj; 1915 return addr != NULL && 1916 (!_g1->is_in_g1_reserved(addr) || !_g1->is_obj_ill(obj)); 1917 } 1918 1919 class G1CMKeepAliveClosure: public OopClosure { 1920 G1CollectedHeap* _g1; 1921 ConcurrentMark* _cm; 1922 CMBitMap* _bitMap; 1923 public: 1924 G1CMKeepAliveClosure(G1CollectedHeap* g1, ConcurrentMark* cm, 1925 CMBitMap* bitMap) : 1926 _g1(g1), _cm(cm), 1927 _bitMap(bitMap) {} 1928 1929 virtual void do_oop(narrowOop* p) { do_oop_work(p); } 1930 virtual void do_oop( oop* p) { do_oop_work(p); } 1931 1932 template <class T> void do_oop_work(T* p) { 1933 oop obj = oopDesc::load_decode_heap_oop(p); 1934 HeapWord* addr = (HeapWord*)obj; 1935 1936 if (_cm->verbose_high()) { 1937 gclog_or_tty->print_cr("\t[0] we're looking at location " 1938 "*"PTR_FORMAT" = "PTR_FORMAT, 1939 p, (void*) obj); 1940 } 1941 1942 if (_g1->is_in_g1_reserved(addr) && _g1->is_obj_ill(obj)) { 1943 _bitMap->mark(addr); 1944 _cm->mark_stack_push(obj); 1945 } 1946 } 1947 }; 1948 1949 class G1CMDrainMarkingStackClosure: public VoidClosure { 1950 CMMarkStack* _markStack; 1951 CMBitMap* _bitMap; 1952 G1CMKeepAliveClosure* _oopClosure; 1953 public: 1954 G1CMDrainMarkingStackClosure(CMBitMap* bitMap, CMMarkStack* markStack, 1955 G1CMKeepAliveClosure* oopClosure) : 1956 _bitMap(bitMap), 1957 _markStack(markStack), 1958 _oopClosure(oopClosure) 1959 {} 1960 1961 void do_void() { 1962 _markStack->drain((OopClosure*)_oopClosure, _bitMap, false); 1963 } 1964 }; 1965 1966 // 'Keep Alive' closure used by parallel reference processing. 1967 // An instance of this closure is used in the parallel reference processing 1968 // code rather than an instance of G1CMKeepAliveClosure. We could have used 1969 // the G1CMKeepAliveClosure as it is MT-safe. Also reference objects are 1970 // placed on to discovered ref lists once so we can mark and push with no 1971 // need to check whether the object has already been marked. Using the 1972 // G1CMKeepAliveClosure would mean, however, having all the worker threads 1973 // operating on the global mark stack. This means that an individual 1974 // worker would be doing lock-free pushes while it processes its own 1975 // discovered ref list followed by drain call. If the discovered ref lists 1976 // are unbalanced then this could cause interference with the other 1977 // workers. Using a CMTask (and its embedded local data structures) 1978 // avoids that potential interference. 1979 class G1CMParKeepAliveAndDrainClosure: public OopClosure { 1980 ConcurrentMark* _cm; 1981 CMTask* _task; 1982 CMBitMap* _bitMap; 1983 int _ref_counter_limit; 1984 int _ref_counter; 1985 public: 1986 G1CMParKeepAliveAndDrainClosure(ConcurrentMark* cm, 1987 CMTask* task, 1988 CMBitMap* bitMap) : 1989 _cm(cm), _task(task), _bitMap(bitMap), 1990 _ref_counter_limit(G1RefProcDrainInterval) 1991 { 1992 assert(_ref_counter_limit > 0, "sanity"); 1993 _ref_counter = _ref_counter_limit; 1994 } 1995 1996 virtual void do_oop(narrowOop* p) { do_oop_work(p); } 1997 virtual void do_oop( oop* p) { do_oop_work(p); } 1998 1999 template <class T> void do_oop_work(T* p) { 2000 if (!_cm->has_overflown()) { 2001 oop obj = oopDesc::load_decode_heap_oop(p); 2002 if (_cm->verbose_high()) { 2003 gclog_or_tty->print_cr("\t[%d] we're looking at location " 2004 "*"PTR_FORMAT" = "PTR_FORMAT, 2005 _task->task_id(), p, (void*) obj); 2006 } 2007 2008 _task->deal_with_reference(obj); 2009 _ref_counter--; 2010 2011 if (_ref_counter == 0) { 2012 // We have dealt with _ref_counter_limit references, pushing them and objects 2013 // reachable from them on to the local stack (and possibly the global stack). 2014 // Call do_marking_step() to process these entries. We call the routine in a 2015 // loop, which we'll exit if there's nothing more to do (i.e. we're done 2016 // with the entries that we've pushed as a result of the deal_with_reference 2017 // calls above) or we overflow. 2018 // Note: CMTask::do_marking_step() can set the CMTask::has_aborted() flag 2019 // while there may still be some work to do. (See the comment at the 2020 // beginning of CMTask::do_marking_step() for those conditions - one of which 2021 // is reaching the specified time target.) It is only when 2022 // CMTask::do_marking_step() returns without setting the has_aborted() flag 2023 // that the marking has completed. 2024 do { 2025 double mark_step_duration_ms = G1ConcMarkStepDurationMillis; 2026 _task->do_marking_step(mark_step_duration_ms, 2027 false /* do_stealing */, 2028 false /* do_termination */); 2029 } while (_task->has_aborted() && !_cm->has_overflown()); 2030 _ref_counter = _ref_counter_limit; 2031 } 2032 } else { 2033 if (_cm->verbose_high()) { 2034 gclog_or_tty->print_cr("\t[%d] CM Overflow", _task->task_id()); 2035 } 2036 } 2037 } 2038 }; 2039 2040 class G1CMParDrainMarkingStackClosure: public VoidClosure { 2041 ConcurrentMark* _cm; 2042 CMTask* _task; 2043 public: 2044 G1CMParDrainMarkingStackClosure(ConcurrentMark* cm, CMTask* task) : 2045 _cm(cm), _task(task) 2046 {} 2047 2048 void do_void() { 2049 do { 2050 if (_cm->verbose_high()) { 2051 gclog_or_tty->print_cr("\t[%d] Drain: Calling do marking_step", 2052 _task->task_id()); 2053 } 2054 2055 // We call CMTask::do_marking_step() to completely drain the local and 2056 // global marking stacks. The routine is called in a loop, which we'll 2057 // exit if there's nothing more to do (i.e. we'completely drained the 2058 // entries that were pushed as a result of applying the 2059 // G1CMParKeepAliveAndDrainClosure to the entries on the discovered ref 2060 // lists above) or we overflow the global marking stack. 2061 // Note: CMTask::do_marking_step() can set the CMTask::has_aborted() flag 2062 // while there may still be some work to do. (See the comment at the 2063 // beginning of CMTask::do_marking_step() for those conditions - one of which 2064 // is reaching the specified time target.) It is only when 2065 // CMTask::do_marking_step() returns without setting the has_aborted() flag 2066 // that the marking has completed. 2067 2068 _task->do_marking_step(1000000000.0 /* something very large */, 2069 true /* do_stealing */, 2070 true /* do_termination */); 2071 } while (_task->has_aborted() && !_cm->has_overflown()); 2072 } 2073 }; 2074 2075 // Implementation of AbstractRefProcTaskExecutor for G1 2076 class G1RefProcTaskExecutor: public AbstractRefProcTaskExecutor { 2077 private: 2078 G1CollectedHeap* _g1h; 2079 ConcurrentMark* _cm; 2080 CMBitMap* _bitmap; 2081 WorkGang* _workers; 2082 int _active_workers; 2083 2084 public: 2085 G1RefProcTaskExecutor(G1CollectedHeap* g1h, 2086 ConcurrentMark* cm, 2087 CMBitMap* bitmap, 2088 WorkGang* workers, 2089 int n_workers) : 2090 _g1h(g1h), _cm(cm), _bitmap(bitmap), 2091 _workers(workers), _active_workers(n_workers) 2092 { } 2093 2094 // Executes the given task using concurrent marking worker threads. 2095 virtual void execute(ProcessTask& task); 2096 virtual void execute(EnqueueTask& task); 2097 }; 2098 2099 class G1RefProcTaskProxy: public AbstractGangTask { 2100 typedef AbstractRefProcTaskExecutor::ProcessTask ProcessTask; 2101 ProcessTask& _proc_task; 2102 G1CollectedHeap* _g1h; 2103 ConcurrentMark* _cm; 2104 CMBitMap* _bitmap; 2105 2106 public: 2107 G1RefProcTaskProxy(ProcessTask& proc_task, 2108 G1CollectedHeap* g1h, 2109 ConcurrentMark* cm, 2110 CMBitMap* bitmap) : 2111 AbstractGangTask("Process reference objects in parallel"), 2112 _proc_task(proc_task), _g1h(g1h), _cm(cm), _bitmap(bitmap) 2113 {} 2114 2115 virtual void work(int i) { 2116 CMTask* marking_task = _cm->task(i); 2117 G1CMIsAliveClosure g1_is_alive(_g1h); 2118 G1CMParKeepAliveAndDrainClosure g1_par_keep_alive(_cm, marking_task, _bitmap); 2119 G1CMParDrainMarkingStackClosure g1_par_drain(_cm, marking_task); 2120 2121 _proc_task.work(i, g1_is_alive, g1_par_keep_alive, g1_par_drain); 2122 } 2123 }; 2124 2125 void G1RefProcTaskExecutor::execute(ProcessTask& proc_task) { 2126 assert(_workers != NULL, "Need parallel worker threads."); 2127 2128 G1RefProcTaskProxy proc_task_proxy(proc_task, _g1h, _cm, _bitmap); 2129 2130 // We need to reset the phase for each task execution so that 2131 // the termination protocol of CMTask::do_marking_step works. 2132 _cm->set_phase(_active_workers, false /* concurrent */); 2133 _g1h->set_par_threads(_active_workers); 2134 _workers->run_task(&proc_task_proxy); 2135 _g1h->set_par_threads(0); 2136 } 2137 2138 class G1RefEnqueueTaskProxy: public AbstractGangTask { 2139 typedef AbstractRefProcTaskExecutor::EnqueueTask EnqueueTask; 2140 EnqueueTask& _enq_task; 2141 2142 public: 2143 G1RefEnqueueTaskProxy(EnqueueTask& enq_task) : 2144 AbstractGangTask("Enqueue reference objects in parallel"), 2145 _enq_task(enq_task) 2146 { } 2147 2148 virtual void work(int i) { 2149 _enq_task.work(i); 2150 } 2151 }; 2152 2153 void G1RefProcTaskExecutor::execute(EnqueueTask& enq_task) { 2154 assert(_workers != NULL, "Need parallel worker threads."); 2155 2156 G1RefEnqueueTaskProxy enq_task_proxy(enq_task); 2157 2158 _g1h->set_par_threads(_active_workers); 2159 _workers->run_task(&enq_task_proxy); 2160 _g1h->set_par_threads(0); 2161 } 2162 2163 void ConcurrentMark::weakRefsWork(bool clear_all_soft_refs) { 2164 ResourceMark rm; 2165 HandleMark hm; 2166 G1CollectedHeap* g1h = G1CollectedHeap::heap(); 2167 ReferenceProcessor* rp = g1h->ref_processor(); 2168 2169 // See the comment in G1CollectedHeap::ref_processing_init() 2170 // about how reference processing currently works in G1. 2171 2172 // Process weak references. 2173 rp->setup_policy(clear_all_soft_refs); 2174 assert(_markStack.isEmpty(), "mark stack should be empty"); 2175 2176 G1CMIsAliveClosure g1_is_alive(g1h); 2177 G1CMKeepAliveClosure g1_keep_alive(g1h, this, nextMarkBitMap()); 2178 G1CMDrainMarkingStackClosure 2179 g1_drain_mark_stack(nextMarkBitMap(), &_markStack, &g1_keep_alive); 2180 // We use the work gang from the G1CollectedHeap and we utilize all 2181 // the worker threads. 2182 int active_workers = g1h->workers() ? g1h->workers()->total_workers() : 1; 2183 active_workers = MAX2(MIN2(active_workers, (int)_max_task_num), 1); 2184 2185 G1RefProcTaskExecutor par_task_executor(g1h, this, nextMarkBitMap(), 2186 g1h->workers(), active_workers); 2187 2188 2189 if (rp->processing_is_mt()) { 2190 // Set the degree of MT here. If the discovery is done MT, there 2191 // may have been a different number of threads doing the discovery 2192 // and a different number of discovered lists may have Ref objects. 2193 // That is OK as long as the Reference lists are balanced (see 2194 // balance_all_queues() and balance_queues()). 2195 rp->set_active_mt_degree(active_workers); 2196 2197 rp->process_discovered_references(&g1_is_alive, 2198 &g1_keep_alive, 2199 &g1_drain_mark_stack, 2200 &par_task_executor); 2201 2202 // The work routines of the parallel keep_alive and drain_marking_stack 2203 // will set the has_overflown flag if we overflow the global marking 2204 // stack. 2205 } else { 2206 rp->process_discovered_references(&g1_is_alive, 2207 &g1_keep_alive, 2208 &g1_drain_mark_stack, 2209 NULL); 2210 2211 } 2212 2213 assert(_markStack.overflow() || _markStack.isEmpty(), 2214 "mark stack should be empty (unless it overflowed)"); 2215 if (_markStack.overflow()) { 2216 // Should have been done already when we tried to push an 2217 // entry on to the global mark stack. But let's do it again. 2218 set_has_overflown(); 2219 } 2220 2221 if (rp->processing_is_mt()) { 2222 assert(rp->num_q() == active_workers, "why not"); 2223 rp->enqueue_discovered_references(&par_task_executor); 2224 } else { 2225 rp->enqueue_discovered_references(); 2226 } 2227 2228 rp->verify_no_references_recorded(); 2229 assert(!rp->discovery_enabled(), "should have been disabled"); 2230 2231 // Now clean up stale oops in StringTable 2232 StringTable::unlink(&g1_is_alive); 2233 // Clean up unreferenced symbols in symbol table. 2234 SymbolTable::unlink(); 2235 } 2236 2237 void ConcurrentMark::swapMarkBitMaps() { 2238 CMBitMapRO* temp = _prevMarkBitMap; 2239 _prevMarkBitMap = (CMBitMapRO*)_nextMarkBitMap; 2240 _nextMarkBitMap = (CMBitMap*) temp; 2241 } 2242 2243 class CMRemarkTask: public AbstractGangTask { 2244 private: 2245 ConcurrentMark *_cm; 2246 2247 public: 2248 void work(int worker_i) { 2249 // Since all available tasks are actually started, we should 2250 // only proceed if we're supposed to be actived. 2251 if ((size_t)worker_i < _cm->active_tasks()) { 2252 CMTask* task = _cm->task(worker_i); 2253 task->record_start_time(); 2254 do { 2255 task->do_marking_step(1000000000.0 /* something very large */, 2256 true /* do_stealing */, 2257 true /* do_termination */); 2258 } while (task->has_aborted() && !_cm->has_overflown()); 2259 // If we overflow, then we do not want to restart. We instead 2260 // want to abort remark and do concurrent marking again. 2261 task->record_end_time(); 2262 } 2263 } 2264 2265 CMRemarkTask(ConcurrentMark* cm) : 2266 AbstractGangTask("Par Remark"), _cm(cm) { } 2267 }; 2268 2269 void ConcurrentMark::checkpointRootsFinalWork() { 2270 ResourceMark rm; 2271 HandleMark hm; 2272 G1CollectedHeap* g1h = G1CollectedHeap::heap(); 2273 2274 g1h->ensure_parsability(false); 2275 2276 if (G1CollectedHeap::use_parallel_gc_threads()) { 2277 G1CollectedHeap::StrongRootsScope srs(g1h); 2278 // this is remark, so we'll use up all available threads 2279 int active_workers = ParallelGCThreads; 2280 set_phase(active_workers, false /* concurrent */); 2281 2282 CMRemarkTask remarkTask(this); 2283 // We will start all available threads, even if we decide that the 2284 // active_workers will be fewer. The extra ones will just bail out 2285 // immediately. 2286 int n_workers = g1h->workers()->total_workers(); 2287 g1h->set_par_threads(n_workers); 2288 g1h->workers()->run_task(&remarkTask); 2289 g1h->set_par_threads(0); 2290 } else { 2291 G1CollectedHeap::StrongRootsScope srs(g1h); 2292 // this is remark, so we'll use up all available threads 2293 int active_workers = 1; 2294 set_phase(active_workers, false /* concurrent */); 2295 2296 CMRemarkTask remarkTask(this); 2297 // We will start all available threads, even if we decide that the 2298 // active_workers will be fewer. The extra ones will just bail out 2299 // immediately. 2300 remarkTask.work(0); 2301 } 2302 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); 2303 guarantee(satb_mq_set.completed_buffers_num() == 0, "invariant"); 2304 2305 print_stats(); 2306 2307 #if VERIFY_OBJS_PROCESSED 2308 if (_scan_obj_cl.objs_processed != ThreadLocalObjQueue::objs_enqueued) { 2309 gclog_or_tty->print_cr("Processed = %d, enqueued = %d.", 2310 _scan_obj_cl.objs_processed, 2311 ThreadLocalObjQueue::objs_enqueued); 2312 guarantee(_scan_obj_cl.objs_processed == 2313 ThreadLocalObjQueue::objs_enqueued, 2314 "Different number of objs processed and enqueued."); 2315 } 2316 #endif 2317 } 2318 2319 #ifndef PRODUCT 2320 2321 class PrintReachableOopClosure: public OopClosure { 2322 private: 2323 G1CollectedHeap* _g1h; 2324 outputStream* _out; 2325 VerifyOption _vo; 2326 bool _all; 2327 2328 public: 2329 PrintReachableOopClosure(outputStream* out, 2330 VerifyOption vo, 2331 bool all) : 2332 _g1h(G1CollectedHeap::heap()), 2333 _out(out), _vo(vo), _all(all) { } 2334 2335 void do_oop(narrowOop* p) { do_oop_work(p); } 2336 void do_oop( oop* p) { do_oop_work(p); } 2337 2338 template <class T> void do_oop_work(T* p) { 2339 oop obj = oopDesc::load_decode_heap_oop(p); 2340 const char* str = NULL; 2341 const char* str2 = ""; 2342 2343 if (obj == NULL) { 2344 str = ""; 2345 } else if (!_g1h->is_in_g1_reserved(obj)) { 2346 str = " O"; 2347 } else { 2348 HeapRegion* hr = _g1h->heap_region_containing(obj); 2349 guarantee(hr != NULL, "invariant"); 2350 bool over_tams = false; 2351 bool marked = false; 2352 2353 switch (_vo) { 2354 case VerifyOption_G1UsePrevMarking: 2355 over_tams = hr->obj_allocated_since_prev_marking(obj); 2356 marked = _g1h->isMarkedPrev(obj); 2357 break; 2358 case VerifyOption_G1UseNextMarking: 2359 over_tams = hr->obj_allocated_since_next_marking(obj); 2360 marked = _g1h->isMarkedNext(obj); 2361 break; 2362 case VerifyOption_G1UseMarkWord: 2363 marked = obj->is_gc_marked(); 2364 break; 2365 default: 2366 ShouldNotReachHere(); 2367 } 2368 2369 if (over_tams) { 2370 str = " >"; 2371 if (marked) { 2372 str2 = " AND MARKED"; 2373 } 2374 } else if (marked) { 2375 str = " M"; 2376 } else { 2377 str = " NOT"; 2378 } 2379 } 2380 2381 _out->print_cr(" "PTR_FORMAT": "PTR_FORMAT"%s%s", 2382 p, (void*) obj, str, str2); 2383 } 2384 }; 2385 2386 class PrintReachableObjectClosure : public ObjectClosure { 2387 private: 2388 G1CollectedHeap* _g1h; 2389 outputStream* _out; 2390 VerifyOption _vo; 2391 bool _all; 2392 HeapRegion* _hr; 2393 2394 public: 2395 PrintReachableObjectClosure(outputStream* out, 2396 VerifyOption vo, 2397 bool all, 2398 HeapRegion* hr) : 2399 _g1h(G1CollectedHeap::heap()), 2400 _out(out), _vo(vo), _all(all), _hr(hr) { } 2401 2402 void do_object(oop o) { 2403 bool over_tams = false; 2404 bool marked = false; 2405 2406 switch (_vo) { 2407 case VerifyOption_G1UsePrevMarking: 2408 over_tams = _hr->obj_allocated_since_prev_marking(o); 2409 marked = _g1h->isMarkedPrev(o); 2410 break; 2411 case VerifyOption_G1UseNextMarking: 2412 over_tams = _hr->obj_allocated_since_next_marking(o); 2413 marked = _g1h->isMarkedNext(o); 2414 break; 2415 case VerifyOption_G1UseMarkWord: 2416 marked = o->is_gc_marked(); 2417 break; 2418 default: 2419 ShouldNotReachHere(); 2420 } 2421 bool print_it = _all || over_tams || marked; 2422 2423 if (print_it) { 2424 _out->print_cr(" "PTR_FORMAT"%s", 2425 o, (over_tams) ? " >" : (marked) ? " M" : ""); 2426 PrintReachableOopClosure oopCl(_out, _vo, _all); 2427 o->oop_iterate(&oopCl); 2428 } 2429 } 2430 }; 2431 2432 class PrintReachableRegionClosure : public HeapRegionClosure { 2433 private: 2434 outputStream* _out; 2435 VerifyOption _vo; 2436 bool _all; 2437 2438 public: 2439 bool doHeapRegion(HeapRegion* hr) { 2440 HeapWord* b = hr->bottom(); 2441 HeapWord* e = hr->end(); 2442 HeapWord* t = hr->top(); 2443 HeapWord* p = NULL; 2444 2445 switch (_vo) { 2446 case VerifyOption_G1UsePrevMarking: 2447 p = hr->prev_top_at_mark_start(); 2448 break; 2449 case VerifyOption_G1UseNextMarking: 2450 p = hr->next_top_at_mark_start(); 2451 break; 2452 case VerifyOption_G1UseMarkWord: 2453 // When we are verifying marking using the mark word 2454 // TAMS has no relevance. 2455 assert(p == NULL, "post-condition"); 2456 break; 2457 default: 2458 ShouldNotReachHere(); 2459 } 2460 _out->print_cr("** ["PTR_FORMAT", "PTR_FORMAT"] top: "PTR_FORMAT" " 2461 "TAMS: "PTR_FORMAT, b, e, t, p); 2462 _out->cr(); 2463 2464 HeapWord* from = b; 2465 HeapWord* to = t; 2466 2467 if (to > from) { 2468 _out->print_cr("Objects in ["PTR_FORMAT", "PTR_FORMAT"]", from, to); 2469 _out->cr(); 2470 PrintReachableObjectClosure ocl(_out, _vo, _all, hr); 2471 hr->object_iterate_mem_careful(MemRegion(from, to), &ocl); 2472 _out->cr(); 2473 } 2474 2475 return false; 2476 } 2477 2478 PrintReachableRegionClosure(outputStream* out, 2479 VerifyOption vo, 2480 bool all) : 2481 _out(out), _vo(vo), _all(all) { } 2482 }; 2483 2484 static const char* verify_option_to_tams(VerifyOption vo) { 2485 switch (vo) { 2486 case VerifyOption_G1UsePrevMarking: 2487 return "PTAMS"; 2488 case VerifyOption_G1UseNextMarking: 2489 return "NTAMS"; 2490 default: 2491 return "NONE"; 2492 } 2493 } 2494 2495 void ConcurrentMark::print_reachable(const char* str, 2496 VerifyOption vo, 2497 bool all) { 2498 gclog_or_tty->cr(); 2499 gclog_or_tty->print_cr("== Doing heap dump... "); 2500 2501 if (G1PrintReachableBaseFile == NULL) { 2502 gclog_or_tty->print_cr(" #### error: no base file defined"); 2503 return; 2504 } 2505 2506 if (strlen(G1PrintReachableBaseFile) + 1 + strlen(str) > 2507 (JVM_MAXPATHLEN - 1)) { 2508 gclog_or_tty->print_cr(" #### error: file name too long"); 2509 return; 2510 } 2511 2512 char file_name[JVM_MAXPATHLEN]; 2513 sprintf(file_name, "%s.%s", G1PrintReachableBaseFile, str); 2514 gclog_or_tty->print_cr(" dumping to file %s", file_name); 2515 2516 fileStream fout(file_name); 2517 if (!fout.is_open()) { 2518 gclog_or_tty->print_cr(" #### error: could not open file"); 2519 return; 2520 } 2521 2522 outputStream* out = &fout; 2523 out->print_cr("-- USING %s", verify_option_to_tams(vo)); 2524 out->cr(); 2525 2526 out->print_cr("--- ITERATING OVER REGIONS"); 2527 out->cr(); 2528 PrintReachableRegionClosure rcl(out, vo, all); 2529 _g1h->heap_region_iterate(&rcl); 2530 out->cr(); 2531 2532 gclog_or_tty->print_cr(" done"); 2533 gclog_or_tty->flush(); 2534 } 2535 2536 #endif // PRODUCT 2537 2538 // This note is for drainAllSATBBuffers and the code in between. 2539 // In the future we could reuse a task to do this work during an 2540 // evacuation pause (since now tasks are not active and can be claimed 2541 // during an evacuation pause). This was a late change to the code and 2542 // is currently not being taken advantage of. 2543 2544 class CMGlobalObjectClosure : public ObjectClosure { 2545 private: 2546 ConcurrentMark* _cm; 2547 2548 public: 2549 void do_object(oop obj) { 2550 _cm->deal_with_reference(obj); 2551 } 2552 2553 CMGlobalObjectClosure(ConcurrentMark* cm) : _cm(cm) { } 2554 }; 2555 2556 void ConcurrentMark::deal_with_reference(oop obj) { 2557 if (verbose_high()) { 2558 gclog_or_tty->print_cr("[global] we're dealing with reference "PTR_FORMAT, 2559 (void*) obj); 2560 } 2561 2562 HeapWord* objAddr = (HeapWord*) obj; 2563 assert(obj->is_oop_or_null(true /* ignore mark word */), "Error"); 2564 if (_g1h->is_in_g1_reserved(objAddr)) { 2565 assert(obj != NULL, "null check is implicit"); 2566 if (!_nextMarkBitMap->isMarked(objAddr)) { 2567 // Only get the containing region if the object is not marked on the 2568 // bitmap (otherwise, it's a waste of time since we won't do 2569 // anything with it). 2570 HeapRegion* hr = _g1h->heap_region_containing_raw(obj); 2571 if (!hr->obj_allocated_since_next_marking(obj)) { 2572 if (verbose_high()) { 2573 gclog_or_tty->print_cr("[global] "PTR_FORMAT" is not considered " 2574 "marked", (void*) obj); 2575 } 2576 2577 // we need to mark it first 2578 if (_nextMarkBitMap->parMark(objAddr)) { 2579 // No OrderAccess:store_load() is needed. It is implicit in the 2580 // CAS done in parMark(objAddr) above 2581 HeapWord* finger = _finger; 2582 if (objAddr < finger) { 2583 if (verbose_high()) { 2584 gclog_or_tty->print_cr("[global] below the global finger " 2585 "("PTR_FORMAT"), pushing it", finger); 2586 } 2587 if (!mark_stack_push(obj)) { 2588 if (verbose_low()) { 2589 gclog_or_tty->print_cr("[global] global stack overflow during " 2590 "deal_with_reference"); 2591 } 2592 } 2593 } 2594 } 2595 } 2596 } 2597 } 2598 } 2599 2600 void ConcurrentMark::drainAllSATBBuffers() { 2601 CMGlobalObjectClosure oc(this); 2602 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); 2603 satb_mq_set.set_closure(&oc); 2604 2605 while (satb_mq_set.apply_closure_to_completed_buffer()) { 2606 if (verbose_medium()) { 2607 gclog_or_tty->print_cr("[global] processed an SATB buffer"); 2608 } 2609 } 2610 2611 // no need to check whether we should do this, as this is only 2612 // called during an evacuation pause 2613 satb_mq_set.iterate_closure_all_threads(); 2614 2615 satb_mq_set.set_closure(NULL); 2616 assert(satb_mq_set.completed_buffers_num() == 0, "invariant"); 2617 } 2618 2619 void ConcurrentMark::markPrev(oop p) { 2620 // Note we are overriding the read-only view of the prev map here, via 2621 // the cast. 2622 ((CMBitMap*)_prevMarkBitMap)->mark((HeapWord*)p); 2623 } 2624 2625 void ConcurrentMark::clear(oop p) { 2626 assert(p != NULL && p->is_oop(), "expected an oop"); 2627 HeapWord* addr = (HeapWord*)p; 2628 assert(addr >= _nextMarkBitMap->startWord() || 2629 addr < _nextMarkBitMap->endWord(), "in a region"); 2630 2631 _nextMarkBitMap->clear(addr); 2632 } 2633 2634 void ConcurrentMark::clearRangeBothMaps(MemRegion mr) { 2635 // Note we are overriding the read-only view of the prev map here, via 2636 // the cast. 2637 ((CMBitMap*)_prevMarkBitMap)->clearRange(mr); 2638 _nextMarkBitMap->clearRange(mr); 2639 } 2640 2641 HeapRegion* 2642 ConcurrentMark::claim_region(int task_num) { 2643 // "checkpoint" the finger 2644 HeapWord* finger = _finger; 2645 2646 // _heap_end will not change underneath our feet; it only changes at 2647 // yield points. 2648 while (finger < _heap_end) { 2649 assert(_g1h->is_in_g1_reserved(finger), "invariant"); 2650 2651 // Note on how this code handles humongous regions. In the 2652 // normal case the finger will reach the start of a "starts 2653 // humongous" (SH) region. Its end will either be the end of the 2654 // last "continues humongous" (CH) region in the sequence, or the 2655 // standard end of the SH region (if the SH is the only region in 2656 // the sequence). That way claim_region() will skip over the CH 2657 // regions. However, there is a subtle race between a CM thread 2658 // executing this method and a mutator thread doing a humongous 2659 // object allocation. The two are not mutually exclusive as the CM 2660 // thread does not need to hold the Heap_lock when it gets 2661 // here. So there is a chance that claim_region() will come across 2662 // a free region that's in the progress of becoming a SH or a CH 2663 // region. In the former case, it will either 2664 // a) Miss the update to the region's end, in which case it will 2665 // visit every subsequent CH region, will find their bitmaps 2666 // empty, and do nothing, or 2667 // b) Will observe the update of the region's end (in which case 2668 // it will skip the subsequent CH regions). 2669 // If it comes across a region that suddenly becomes CH, the 2670 // scenario will be similar to b). So, the race between 2671 // claim_region() and a humongous object allocation might force us 2672 // to do a bit of unnecessary work (due to some unnecessary bitmap 2673 // iterations) but it should not introduce and correctness issues. 2674 HeapRegion* curr_region = _g1h->heap_region_containing_raw(finger); 2675 HeapWord* bottom = curr_region->bottom(); 2676 HeapWord* end = curr_region->end(); 2677 HeapWord* limit = curr_region->next_top_at_mark_start(); 2678 2679 if (verbose_low()) { 2680 gclog_or_tty->print_cr("[%d] curr_region = "PTR_FORMAT" " 2681 "["PTR_FORMAT", "PTR_FORMAT"), " 2682 "limit = "PTR_FORMAT, 2683 task_num, curr_region, bottom, end, limit); 2684 } 2685 2686 // Is the gap between reading the finger and doing the CAS too long? 2687 HeapWord* res = (HeapWord*) Atomic::cmpxchg_ptr(end, &_finger, finger); 2688 if (res == finger) { 2689 // we succeeded 2690 2691 // notice that _finger == end cannot be guaranteed here since, 2692 // someone else might have moved the finger even further 2693 assert(_finger >= end, "the finger should have moved forward"); 2694 2695 if (verbose_low()) { 2696 gclog_or_tty->print_cr("[%d] we were successful with region = " 2697 PTR_FORMAT, task_num, curr_region); 2698 } 2699 2700 if (limit > bottom) { 2701 if (verbose_low()) { 2702 gclog_or_tty->print_cr("[%d] region "PTR_FORMAT" is not empty, " 2703 "returning it ", task_num, curr_region); 2704 } 2705 return curr_region; 2706 } else { 2707 assert(limit == bottom, 2708 "the region limit should be at bottom"); 2709 if (verbose_low()) { 2710 gclog_or_tty->print_cr("[%d] region "PTR_FORMAT" is empty, " 2711 "returning NULL", task_num, curr_region); 2712 } 2713 // we return NULL and the caller should try calling 2714 // claim_region() again. 2715 return NULL; 2716 } 2717 } else { 2718 assert(_finger > finger, "the finger should have moved forward"); 2719 if (verbose_low()) { 2720 gclog_or_tty->print_cr("[%d] somebody else moved the finger, " 2721 "global finger = "PTR_FORMAT", " 2722 "our finger = "PTR_FORMAT, 2723 task_num, _finger, finger); 2724 } 2725 2726 // read it again 2727 finger = _finger; 2728 } 2729 } 2730 2731 return NULL; 2732 } 2733 2734 bool ConcurrentMark::invalidate_aborted_regions_in_cset() { 2735 bool result = false; 2736 for (int i = 0; i < (int)_max_task_num; ++i) { 2737 CMTask* the_task = _tasks[i]; 2738 MemRegion mr = the_task->aborted_region(); 2739 if (mr.start() != NULL) { 2740 assert(mr.end() != NULL, "invariant"); 2741 assert(mr.word_size() > 0, "invariant"); 2742 HeapRegion* hr = _g1h->heap_region_containing(mr.start()); 2743 assert(hr != NULL, "invariant"); 2744 if (hr->in_collection_set()) { 2745 // The region points into the collection set 2746 the_task->set_aborted_region(MemRegion()); 2747 result = true; 2748 } 2749 } 2750 } 2751 return result; 2752 } 2753 2754 bool ConcurrentMark::has_aborted_regions() { 2755 for (int i = 0; i < (int)_max_task_num; ++i) { 2756 CMTask* the_task = _tasks[i]; 2757 MemRegion mr = the_task->aborted_region(); 2758 if (mr.start() != NULL) { 2759 assert(mr.end() != NULL, "invariant"); 2760 assert(mr.word_size() > 0, "invariant"); 2761 return true; 2762 } 2763 } 2764 return false; 2765 } 2766 2767 void ConcurrentMark::oops_do(OopClosure* cl) { 2768 if (_markStack.size() > 0 && verbose_low()) { 2769 gclog_or_tty->print_cr("[global] scanning the global marking stack, " 2770 "size = %d", _markStack.size()); 2771 } 2772 // we first iterate over the contents of the mark stack... 2773 _markStack.oops_do(cl); 2774 2775 for (int i = 0; i < (int)_max_task_num; ++i) { 2776 OopTaskQueue* queue = _task_queues->queue((int)i); 2777 2778 if (queue->size() > 0 && verbose_low()) { 2779 gclog_or_tty->print_cr("[global] scanning task queue of task %d, " 2780 "size = %d", i, queue->size()); 2781 } 2782 2783 // ...then over the contents of the all the task queues. 2784 queue->oops_do(cl); 2785 } 2786 2787 // Invalidate any entries, that are in the region stack, that 2788 // point into the collection set 2789 if (_regionStack.invalidate_entries_into_cset()) { 2790 // otherwise, any gray objects copied during the evacuation pause 2791 // might not be visited. 2792 assert(_should_gray_objects, "invariant"); 2793 } 2794 2795 // Invalidate any aborted regions, recorded in the individual CM 2796 // tasks, that point into the collection set. 2797 if (invalidate_aborted_regions_in_cset()) { 2798 // otherwise, any gray objects copied during the evacuation pause 2799 // might not be visited. 2800 assert(_should_gray_objects, "invariant"); 2801 } 2802 2803 } 2804 2805 void ConcurrentMark::clear_marking_state(bool clear_overflow) { 2806 _markStack.setEmpty(); 2807 _markStack.clear_overflow(); 2808 _regionStack.setEmpty(); 2809 _regionStack.clear_overflow(); 2810 if (clear_overflow) { 2811 clear_has_overflown(); 2812 } else { 2813 assert(has_overflown(), "pre-condition"); 2814 } 2815 _finger = _heap_start; 2816 2817 for (int i = 0; i < (int)_max_task_num; ++i) { 2818 OopTaskQueue* queue = _task_queues->queue(i); 2819 queue->set_empty(); 2820 // Clear any partial regions from the CMTasks 2821 _tasks[i]->clear_aborted_region(); 2822 } 2823 } 2824 2825 void ConcurrentMark::print_stats() { 2826 if (verbose_stats()) { 2827 gclog_or_tty->print_cr("---------------------------------------------------------------------"); 2828 for (size_t i = 0; i < _active_tasks; ++i) { 2829 _tasks[i]->print_stats(); 2830 gclog_or_tty->print_cr("---------------------------------------------------------------------"); 2831 } 2832 } 2833 } 2834 2835 class CSMarkOopClosure: public OopClosure { 2836 friend class CSMarkBitMapClosure; 2837 2838 G1CollectedHeap* _g1h; 2839 CMBitMap* _bm; 2840 ConcurrentMark* _cm; 2841 oop* _ms; 2842 jint* _array_ind_stack; 2843 int _ms_size; 2844 int _ms_ind; 2845 int _array_increment; 2846 2847 bool push(oop obj, int arr_ind = 0) { 2848 if (_ms_ind == _ms_size) { 2849 gclog_or_tty->print_cr("Mark stack is full."); 2850 return false; 2851 } 2852 _ms[_ms_ind] = obj; 2853 if (obj->is_objArray()) { 2854 _array_ind_stack[_ms_ind] = arr_ind; 2855 } 2856 _ms_ind++; 2857 return true; 2858 } 2859 2860 oop pop() { 2861 if (_ms_ind == 0) { 2862 return NULL; 2863 } else { 2864 _ms_ind--; 2865 return _ms[_ms_ind]; 2866 } 2867 } 2868 2869 template <class T> bool drain() { 2870 while (_ms_ind > 0) { 2871 oop obj = pop(); 2872 assert(obj != NULL, "Since index was non-zero."); 2873 if (obj->is_objArray()) { 2874 jint arr_ind = _array_ind_stack[_ms_ind]; 2875 objArrayOop aobj = objArrayOop(obj); 2876 jint len = aobj->length(); 2877 jint next_arr_ind = arr_ind + _array_increment; 2878 if (next_arr_ind < len) { 2879 push(obj, next_arr_ind); 2880 } 2881 // Now process this portion of this one. 2882 int lim = MIN2(next_arr_ind, len); 2883 for (int j = arr_ind; j < lim; j++) { 2884 do_oop(aobj->objArrayOopDesc::obj_at_addr<T>(j)); 2885 } 2886 2887 } else { 2888 obj->oop_iterate(this); 2889 } 2890 if (abort()) return false; 2891 } 2892 return true; 2893 } 2894 2895 public: 2896 CSMarkOopClosure(ConcurrentMark* cm, int ms_size) : 2897 _g1h(G1CollectedHeap::heap()), 2898 _cm(cm), 2899 _bm(cm->nextMarkBitMap()), 2900 _ms_size(ms_size), _ms_ind(0), 2901 _ms(NEW_C_HEAP_ARRAY(oop, ms_size)), 2902 _array_ind_stack(NEW_C_HEAP_ARRAY(jint, ms_size)), 2903 _array_increment(MAX2(ms_size/8, 16)) 2904 {} 2905 2906 ~CSMarkOopClosure() { 2907 FREE_C_HEAP_ARRAY(oop, _ms); 2908 FREE_C_HEAP_ARRAY(jint, _array_ind_stack); 2909 } 2910 2911 virtual void do_oop(narrowOop* p) { do_oop_work(p); } 2912 virtual void do_oop( oop* p) { do_oop_work(p); } 2913 2914 template <class T> void do_oop_work(T* p) { 2915 T heap_oop = oopDesc::load_heap_oop(p); 2916 if (oopDesc::is_null(heap_oop)) return; 2917 oop obj = oopDesc::decode_heap_oop_not_null(heap_oop); 2918 if (obj->is_forwarded()) { 2919 // If the object has already been forwarded, we have to make sure 2920 // that it's marked. So follow the forwarding pointer. Note that 2921 // this does the right thing for self-forwarding pointers in the 2922 // evacuation failure case. 2923 obj = obj->forwardee(); 2924 } 2925 HeapRegion* hr = _g1h->heap_region_containing(obj); 2926 if (hr != NULL) { 2927 if (hr->in_collection_set()) { 2928 if (_g1h->is_obj_ill(obj)) { 2929 _bm->mark((HeapWord*)obj); 2930 if (!push(obj)) { 2931 gclog_or_tty->print_cr("Setting abort in CSMarkOopClosure because push failed."); 2932 set_abort(); 2933 } 2934 } 2935 } else { 2936 // Outside the collection set; we need to gray it 2937 _cm->deal_with_reference(obj); 2938 } 2939 } 2940 } 2941 }; 2942 2943 class CSMarkBitMapClosure: public BitMapClosure { 2944 G1CollectedHeap* _g1h; 2945 CMBitMap* _bitMap; 2946 ConcurrentMark* _cm; 2947 CSMarkOopClosure _oop_cl; 2948 public: 2949 CSMarkBitMapClosure(ConcurrentMark* cm, int ms_size) : 2950 _g1h(G1CollectedHeap::heap()), 2951 _bitMap(cm->nextMarkBitMap()), 2952 _oop_cl(cm, ms_size) 2953 {} 2954 2955 ~CSMarkBitMapClosure() {} 2956 2957 bool do_bit(size_t offset) { 2958 // convert offset into a HeapWord* 2959 HeapWord* addr = _bitMap->offsetToHeapWord(offset); 2960 assert(_bitMap->endWord() && addr < _bitMap->endWord(), 2961 "address out of range"); 2962 assert(_bitMap->isMarked(addr), "tautology"); 2963 oop obj = oop(addr); 2964 if (!obj->is_forwarded()) { 2965 if (!_oop_cl.push(obj)) return false; 2966 if (UseCompressedOops) { 2967 if (!_oop_cl.drain<narrowOop>()) return false; 2968 } else { 2969 if (!_oop_cl.drain<oop>()) return false; 2970 } 2971 } 2972 // Otherwise... 2973 return true; 2974 } 2975 }; 2976 2977 2978 class CompleteMarkingInCSHRClosure: public HeapRegionClosure { 2979 CMBitMap* _bm; 2980 CSMarkBitMapClosure _bit_cl; 2981 enum SomePrivateConstants { 2982 MSSize = 1000 2983 }; 2984 bool _completed; 2985 public: 2986 CompleteMarkingInCSHRClosure(ConcurrentMark* cm) : 2987 _bm(cm->nextMarkBitMap()), 2988 _bit_cl(cm, MSSize), 2989 _completed(true) 2990 {} 2991 2992 ~CompleteMarkingInCSHRClosure() {} 2993 2994 bool doHeapRegion(HeapRegion* r) { 2995 if (!r->evacuation_failed()) { 2996 MemRegion mr = MemRegion(r->bottom(), r->next_top_at_mark_start()); 2997 if (!mr.is_empty()) { 2998 if (!_bm->iterate(&_bit_cl, mr)) { 2999 _completed = false; 3000 return true; 3001 } 3002 } 3003 } 3004 return false; 3005 } 3006 3007 bool completed() { return _completed; } 3008 }; 3009 3010 class ClearMarksInHRClosure: public HeapRegionClosure { 3011 CMBitMap* _bm; 3012 public: 3013 ClearMarksInHRClosure(CMBitMap* bm): _bm(bm) { } 3014 3015 bool doHeapRegion(HeapRegion* r) { 3016 if (!r->used_region().is_empty() && !r->evacuation_failed()) { 3017 MemRegion usedMR = r->used_region(); 3018 _bm->clearRange(r->used_region()); 3019 } 3020 return false; 3021 } 3022 }; 3023 3024 void ConcurrentMark::complete_marking_in_collection_set() { 3025 G1CollectedHeap* g1h = G1CollectedHeap::heap(); 3026 3027 if (!g1h->mark_in_progress()) { 3028 g1h->g1_policy()->record_mark_closure_time(0.0); 3029 return; 3030 } 3031 3032 int i = 1; 3033 double start = os::elapsedTime(); 3034 while (true) { 3035 i++; 3036 CompleteMarkingInCSHRClosure cmplt(this); 3037 g1h->collection_set_iterate(&cmplt); 3038 if (cmplt.completed()) break; 3039 } 3040 double end_time = os::elapsedTime(); 3041 double elapsed_time_ms = (end_time - start) * 1000.0; 3042 g1h->g1_policy()->record_mark_closure_time(elapsed_time_ms); 3043 3044 ClearMarksInHRClosure clr(nextMarkBitMap()); 3045 g1h->collection_set_iterate(&clr); 3046 } 3047 3048 // The next two methods deal with the following optimisation. Some 3049 // objects are gray by being marked and located above the finger. If 3050 // they are copied, during an evacuation pause, below the finger then 3051 // the need to be pushed on the stack. The observation is that, if 3052 // there are no regions in the collection set located above the 3053 // finger, then the above cannot happen, hence we do not need to 3054 // explicitly gray any objects when copying them to below the 3055 // finger. The global stack will be scanned to ensure that, if it 3056 // points to objects being copied, it will update their 3057 // location. There is a tricky situation with the gray objects in 3058 // region stack that are being coped, however. See the comment in 3059 // newCSet(). 3060 3061 void ConcurrentMark::newCSet() { 3062 if (!concurrent_marking_in_progress()) { 3063 // nothing to do if marking is not in progress 3064 return; 3065 } 3066 3067 // find what the lowest finger is among the global and local fingers 3068 _min_finger = _finger; 3069 for (int i = 0; i < (int)_max_task_num; ++i) { 3070 CMTask* task = _tasks[i]; 3071 HeapWord* task_finger = task->finger(); 3072 if (task_finger != NULL && task_finger < _min_finger) { 3073 _min_finger = task_finger; 3074 } 3075 } 3076 3077 _should_gray_objects = false; 3078 3079 // This fixes a very subtle and fustrating bug. It might be the case 3080 // that, during en evacuation pause, heap regions that contain 3081 // objects that are gray (by being in regions contained in the 3082 // region stack) are included in the collection set. Since such gray 3083 // objects will be moved, and because it's not easy to redirect 3084 // region stack entries to point to a new location (because objects 3085 // in one region might be scattered to multiple regions after they 3086 // are copied), one option is to ensure that all marked objects 3087 // copied during a pause are pushed on the stack. Notice, however, 3088 // that this problem can only happen when the region stack is not 3089 // empty during an evacuation pause. So, we make the fix a bit less 3090 // conservative and ensure that regions are pushed on the stack, 3091 // irrespective whether all collection set regions are below the 3092 // finger, if the region stack is not empty. This is expected to be 3093 // a rare case, so I don't think it's necessary to be smarted about it. 3094 if (!region_stack_empty() || has_aborted_regions()) { 3095 _should_gray_objects = true; 3096 } 3097 } 3098 3099 void ConcurrentMark::registerCSetRegion(HeapRegion* hr) { 3100 if (!concurrent_marking_in_progress()) return; 3101 3102 HeapWord* region_end = hr->end(); 3103 if (region_end > _min_finger) { 3104 _should_gray_objects = true; 3105 } 3106 } 3107 3108 // Resets the region fields of active CMTasks whose values point 3109 // into the collection set. 3110 void ConcurrentMark::reset_active_task_region_fields_in_cset() { 3111 assert(SafepointSynchronize::is_at_safepoint(), "should be in STW"); 3112 assert(parallel_marking_threads() <= _max_task_num, "sanity"); 3113 3114 for (int i = 0; i < (int)parallel_marking_threads(); i += 1) { 3115 CMTask* task = _tasks[i]; 3116 HeapWord* task_finger = task->finger(); 3117 if (task_finger != NULL) { 3118 assert(_g1h->is_in_g1_reserved(task_finger), "not in heap"); 3119 HeapRegion* finger_region = _g1h->heap_region_containing(task_finger); 3120 if (finger_region->in_collection_set()) { 3121 // The task's current region is in the collection set. 3122 // This region will be evacuated in the current GC and 3123 // the region fields in the task will be stale. 3124 task->giveup_current_region(); 3125 } 3126 } 3127 } 3128 } 3129 3130 // abandon current marking iteration due to a Full GC 3131 void ConcurrentMark::abort() { 3132 // Clear all marks to force marking thread to do nothing 3133 _nextMarkBitMap->clearAll(); 3134 // Empty mark stack 3135 clear_marking_state(); 3136 for (int i = 0; i < (int)_max_task_num; ++i) { 3137 _tasks[i]->clear_region_fields(); 3138 } 3139 _has_aborted = true; 3140 3141 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); 3142 satb_mq_set.abandon_partial_marking(); 3143 // This can be called either during or outside marking, we'll read 3144 // the expected_active value from the SATB queue set. 3145 satb_mq_set.set_active_all_threads( 3146 false, /* new active value */ 3147 satb_mq_set.is_active() /* expected_active */); 3148 } 3149 3150 static void print_ms_time_info(const char* prefix, const char* name, 3151 NumberSeq& ns) { 3152 gclog_or_tty->print_cr("%s%5d %12s: total time = %8.2f s (avg = %8.2f ms).", 3153 prefix, ns.num(), name, ns.sum()/1000.0, ns.avg()); 3154 if (ns.num() > 0) { 3155 gclog_or_tty->print_cr("%s [std. dev = %8.2f ms, max = %8.2f ms]", 3156 prefix, ns.sd(), ns.maximum()); 3157 } 3158 } 3159 3160 void ConcurrentMark::print_summary_info() { 3161 gclog_or_tty->print_cr(" Concurrent marking:"); 3162 print_ms_time_info(" ", "init marks", _init_times); 3163 print_ms_time_info(" ", "remarks", _remark_times); 3164 { 3165 print_ms_time_info(" ", "final marks", _remark_mark_times); 3166 print_ms_time_info(" ", "weak refs", _remark_weak_ref_times); 3167 3168 } 3169 print_ms_time_info(" ", "cleanups", _cleanup_times); 3170 gclog_or_tty->print_cr(" Final counting total time = %8.2f s (avg = %8.2f ms).", 3171 _total_counting_time, 3172 (_cleanup_times.num() > 0 ? _total_counting_time * 1000.0 / 3173 (double)_cleanup_times.num() 3174 : 0.0)); 3175 if (G1ScrubRemSets) { 3176 gclog_or_tty->print_cr(" RS scrub total time = %8.2f s (avg = %8.2f ms).", 3177 _total_rs_scrub_time, 3178 (_cleanup_times.num() > 0 ? _total_rs_scrub_time * 1000.0 / 3179 (double)_cleanup_times.num() 3180 : 0.0)); 3181 } 3182 gclog_or_tty->print_cr(" Total stop_world time = %8.2f s.", 3183 (_init_times.sum() + _remark_times.sum() + 3184 _cleanup_times.sum())/1000.0); 3185 gclog_or_tty->print_cr(" Total concurrent time = %8.2f s " 3186 "(%8.2f s marking, %8.2f s counting).", 3187 cmThread()->vtime_accum(), 3188 cmThread()->vtime_mark_accum(), 3189 cmThread()->vtime_count_accum()); 3190 } 3191 3192 void ConcurrentMark::print_worker_threads_on(outputStream* st) const { 3193 _parallel_workers->print_worker_threads_on(st); 3194 } 3195 3196 // Closures 3197 // XXX: there seems to be a lot of code duplication here; 3198 // should refactor and consolidate the shared code. 3199 3200 // This closure is used to mark refs into the CMS generation in 3201 // the CMS bit map. Called at the first checkpoint. 3202 3203 // We take a break if someone is trying to stop the world. 3204 bool ConcurrentMark::do_yield_check(int worker_i) { 3205 if (should_yield()) { 3206 if (worker_i == 0) { 3207 _g1h->g1_policy()->record_concurrent_pause(); 3208 } 3209 cmThread()->yield(); 3210 if (worker_i == 0) { 3211 _g1h->g1_policy()->record_concurrent_pause_end(); 3212 } 3213 return true; 3214 } else { 3215 return false; 3216 } 3217 } 3218 3219 bool ConcurrentMark::should_yield() { 3220 return cmThread()->should_yield(); 3221 } 3222 3223 bool ConcurrentMark::containing_card_is_marked(void* p) { 3224 size_t offset = pointer_delta(p, _g1h->reserved_region().start(), 1); 3225 return _card_bm.at(offset >> CardTableModRefBS::card_shift); 3226 } 3227 3228 bool ConcurrentMark::containing_cards_are_marked(void* start, 3229 void* last) { 3230 return containing_card_is_marked(start) && 3231 containing_card_is_marked(last); 3232 } 3233 3234 #ifndef PRODUCT 3235 // for debugging purposes 3236 void ConcurrentMark::print_finger() { 3237 gclog_or_tty->print_cr("heap ["PTR_FORMAT", "PTR_FORMAT"), global finger = "PTR_FORMAT, 3238 _heap_start, _heap_end, _finger); 3239 for (int i = 0; i < (int) _max_task_num; ++i) { 3240 gclog_or_tty->print(" %d: "PTR_FORMAT, i, _tasks[i]->finger()); 3241 } 3242 gclog_or_tty->print_cr(""); 3243 } 3244 #endif 3245 3246 void CMTask::scan_object(oop obj) { 3247 assert(_nextMarkBitMap->isMarked((HeapWord*) obj), "invariant"); 3248 3249 if (_cm->verbose_high()) { 3250 gclog_or_tty->print_cr("[%d] we're scanning object "PTR_FORMAT, 3251 _task_id, (void*) obj); 3252 } 3253 3254 size_t obj_size = obj->size(); 3255 _words_scanned += obj_size; 3256 3257 obj->oop_iterate(_cm_oop_closure); 3258 statsOnly( ++_objs_scanned ); 3259 check_limits(); 3260 } 3261 3262 // Closure for iteration over bitmaps 3263 class CMBitMapClosure : public BitMapClosure { 3264 private: 3265 // the bitmap that is being iterated over 3266 CMBitMap* _nextMarkBitMap; 3267 ConcurrentMark* _cm; 3268 CMTask* _task; 3269 // true if we're scanning a heap region claimed by the task (so that 3270 // we move the finger along), false if we're not, i.e. currently when 3271 // scanning a heap region popped from the region stack (so that we 3272 // do not move the task finger along; it'd be a mistake if we did so). 3273 bool _scanning_heap_region; 3274 3275 public: 3276 CMBitMapClosure(CMTask *task, 3277 ConcurrentMark* cm, 3278 CMBitMap* nextMarkBitMap) 3279 : _task(task), _cm(cm), _nextMarkBitMap(nextMarkBitMap) { } 3280 3281 void set_scanning_heap_region(bool scanning_heap_region) { 3282 _scanning_heap_region = scanning_heap_region; 3283 } 3284 3285 bool do_bit(size_t offset) { 3286 HeapWord* addr = _nextMarkBitMap->offsetToHeapWord(offset); 3287 assert(_nextMarkBitMap->isMarked(addr), "invariant"); 3288 assert( addr < _cm->finger(), "invariant"); 3289 3290 if (_scanning_heap_region) { 3291 statsOnly( _task->increase_objs_found_on_bitmap() ); 3292 assert(addr >= _task->finger(), "invariant"); 3293 // We move that task's local finger along. 3294 _task->move_finger_to(addr); 3295 } else { 3296 // We move the task's region finger along. 3297 _task->move_region_finger_to(addr); 3298 } 3299 3300 _task->scan_object(oop(addr)); 3301 // we only partially drain the local queue and global stack 3302 _task->drain_local_queue(true); 3303 _task->drain_global_stack(true); 3304 3305 // if the has_aborted flag has been raised, we need to bail out of 3306 // the iteration 3307 return !_task->has_aborted(); 3308 } 3309 }; 3310 3311 // Closure for iterating over objects, currently only used for 3312 // processing SATB buffers. 3313 class CMObjectClosure : public ObjectClosure { 3314 private: 3315 CMTask* _task; 3316 3317 public: 3318 void do_object(oop obj) { 3319 _task->deal_with_reference(obj); 3320 } 3321 3322 CMObjectClosure(CMTask* task) : _task(task) { } 3323 }; 3324 3325 G1CMOopClosure::G1CMOopClosure(G1CollectedHeap* g1h, 3326 ConcurrentMark* cm, 3327 CMTask* task) 3328 : _g1h(g1h), _cm(cm), _task(task) { 3329 assert(_ref_processor == NULL, "should be initialized to NULL"); 3330 3331 if (G1UseConcMarkReferenceProcessing) { 3332 _ref_processor = g1h->ref_processor(); 3333 assert(_ref_processor != NULL, "should not be NULL"); 3334 } 3335 } 3336 3337 void CMTask::setup_for_region(HeapRegion* hr) { 3338 // Separated the asserts so that we know which one fires. 3339 assert(hr != NULL, 3340 "claim_region() should have filtered out continues humongous regions"); 3341 assert(!hr->continuesHumongous(), 3342 "claim_region() should have filtered out continues humongous regions"); 3343 3344 if (_cm->verbose_low()) { 3345 gclog_or_tty->print_cr("[%d] setting up for region "PTR_FORMAT, 3346 _task_id, hr); 3347 } 3348 3349 _curr_region = hr; 3350 _finger = hr->bottom(); 3351 update_region_limit(); 3352 } 3353 3354 void CMTask::update_region_limit() { 3355 HeapRegion* hr = _curr_region; 3356 HeapWord* bottom = hr->bottom(); 3357 HeapWord* limit = hr->next_top_at_mark_start(); 3358 3359 if (limit == bottom) { 3360 if (_cm->verbose_low()) { 3361 gclog_or_tty->print_cr("[%d] found an empty region " 3362 "["PTR_FORMAT", "PTR_FORMAT")", 3363 _task_id, bottom, limit); 3364 } 3365 // The region was collected underneath our feet. 3366 // We set the finger to bottom to ensure that the bitmap 3367 // iteration that will follow this will not do anything. 3368 // (this is not a condition that holds when we set the region up, 3369 // as the region is not supposed to be empty in the first place) 3370 _finger = bottom; 3371 } else if (limit >= _region_limit) { 3372 assert(limit >= _finger, "peace of mind"); 3373 } else { 3374 assert(limit < _region_limit, "only way to get here"); 3375 // This can happen under some pretty unusual circumstances. An 3376 // evacuation pause empties the region underneath our feet (NTAMS 3377 // at bottom). We then do some allocation in the region (NTAMS 3378 // stays at bottom), followed by the region being used as a GC 3379 // alloc region (NTAMS will move to top() and the objects 3380 // originally below it will be grayed). All objects now marked in 3381 // the region are explicitly grayed, if below the global finger, 3382 // and we do not need in fact to scan anything else. So, we simply 3383 // set _finger to be limit to ensure that the bitmap iteration 3384 // doesn't do anything. 3385 _finger = limit; 3386 } 3387 3388 _region_limit = limit; 3389 } 3390 3391 void CMTask::giveup_current_region() { 3392 assert(_curr_region != NULL, "invariant"); 3393 if (_cm->verbose_low()) { 3394 gclog_or_tty->print_cr("[%d] giving up region "PTR_FORMAT, 3395 _task_id, _curr_region); 3396 } 3397 clear_region_fields(); 3398 } 3399 3400 void CMTask::clear_region_fields() { 3401 // Values for these three fields that indicate that we're not 3402 // holding on to a region. 3403 _curr_region = NULL; 3404 _finger = NULL; 3405 _region_limit = NULL; 3406 3407 _region_finger = NULL; 3408 } 3409 3410 void CMTask::set_cm_oop_closure(G1CMOopClosure* cm_oop_closure) { 3411 if (cm_oop_closure == NULL) { 3412 assert(_cm_oop_closure != NULL, "invariant"); 3413 } else { 3414 assert(_cm_oop_closure == NULL, "invariant"); 3415 } 3416 _cm_oop_closure = cm_oop_closure; 3417 } 3418 3419 void CMTask::reset(CMBitMap* nextMarkBitMap) { 3420 guarantee(nextMarkBitMap != NULL, "invariant"); 3421 3422 if (_cm->verbose_low()) { 3423 gclog_or_tty->print_cr("[%d] resetting", _task_id); 3424 } 3425 3426 _nextMarkBitMap = nextMarkBitMap; 3427 clear_region_fields(); 3428 assert(_aborted_region.is_empty(), "should have been cleared"); 3429 3430 _calls = 0; 3431 _elapsed_time_ms = 0.0; 3432 _termination_time_ms = 0.0; 3433 _termination_start_time_ms = 0.0; 3434 3435 #if _MARKING_STATS_ 3436 _local_pushes = 0; 3437 _local_pops = 0; 3438 _local_max_size = 0; 3439 _objs_scanned = 0; 3440 _global_pushes = 0; 3441 _global_pops = 0; 3442 _global_max_size = 0; 3443 _global_transfers_to = 0; 3444 _global_transfers_from = 0; 3445 _region_stack_pops = 0; 3446 _regions_claimed = 0; 3447 _objs_found_on_bitmap = 0; 3448 _satb_buffers_processed = 0; 3449 _steal_attempts = 0; 3450 _steals = 0; 3451 _aborted = 0; 3452 _aborted_overflow = 0; 3453 _aborted_cm_aborted = 0; 3454 _aborted_yield = 0; 3455 _aborted_timed_out = 0; 3456 _aborted_satb = 0; 3457 _aborted_termination = 0; 3458 #endif // _MARKING_STATS_ 3459 } 3460 3461 bool CMTask::should_exit_termination() { 3462 regular_clock_call(); 3463 // This is called when we are in the termination protocol. We should 3464 // quit if, for some reason, this task wants to abort or the global 3465 // stack is not empty (this means that we can get work from it). 3466 return !_cm->mark_stack_empty() || has_aborted(); 3467 } 3468 3469 void CMTask::reached_limit() { 3470 assert(_words_scanned >= _words_scanned_limit || 3471 _refs_reached >= _refs_reached_limit , 3472 "shouldn't have been called otherwise"); 3473 regular_clock_call(); 3474 } 3475 3476 void CMTask::regular_clock_call() { 3477 if (has_aborted()) return; 3478 3479 // First, we need to recalculate the words scanned and refs reached 3480 // limits for the next clock call. 3481 recalculate_limits(); 3482 3483 // During the regular clock call we do the following 3484 3485 // (1) If an overflow has been flagged, then we abort. 3486 if (_cm->has_overflown()) { 3487 set_has_aborted(); 3488 return; 3489 } 3490 3491 // If we are not concurrent (i.e. we're doing remark) we don't need 3492 // to check anything else. The other steps are only needed during 3493 // the concurrent marking phase. 3494 if (!concurrent()) return; 3495 3496 // (2) If marking has been aborted for Full GC, then we also abort. 3497 if (_cm->has_aborted()) { 3498 set_has_aborted(); 3499 statsOnly( ++_aborted_cm_aborted ); 3500 return; 3501 } 3502 3503 double curr_time_ms = os::elapsedVTime() * 1000.0; 3504 3505 // (3) If marking stats are enabled, then we update the step history. 3506 #if _MARKING_STATS_ 3507 if (_words_scanned >= _words_scanned_limit) { 3508 ++_clock_due_to_scanning; 3509 } 3510 if (_refs_reached >= _refs_reached_limit) { 3511 ++_clock_due_to_marking; 3512 } 3513 3514 double last_interval_ms = curr_time_ms - _interval_start_time_ms; 3515 _interval_start_time_ms = curr_time_ms; 3516 _all_clock_intervals_ms.add(last_interval_ms); 3517 3518 if (_cm->verbose_medium()) { 3519 gclog_or_tty->print_cr("[%d] regular clock, interval = %1.2lfms, " 3520 "scanned = %d%s, refs reached = %d%s", 3521 _task_id, last_interval_ms, 3522 _words_scanned, 3523 (_words_scanned >= _words_scanned_limit) ? " (*)" : "", 3524 _refs_reached, 3525 (_refs_reached >= _refs_reached_limit) ? " (*)" : ""); 3526 } 3527 #endif // _MARKING_STATS_ 3528 3529 // (4) We check whether we should yield. If we have to, then we abort. 3530 if (_cm->should_yield()) { 3531 // We should yield. To do this we abort the task. The caller is 3532 // responsible for yielding. 3533 set_has_aborted(); 3534 statsOnly( ++_aborted_yield ); 3535 return; 3536 } 3537 3538 // (5) We check whether we've reached our time quota. If we have, 3539 // then we abort. 3540 double elapsed_time_ms = curr_time_ms - _start_time_ms; 3541 if (elapsed_time_ms > _time_target_ms) { 3542 set_has_aborted(); 3543 _has_timed_out = true; 3544 statsOnly( ++_aborted_timed_out ); 3545 return; 3546 } 3547 3548 // (6) Finally, we check whether there are enough completed STAB 3549 // buffers available for processing. If there are, we abort. 3550 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); 3551 if (!_draining_satb_buffers && satb_mq_set.process_completed_buffers()) { 3552 if (_cm->verbose_low()) { 3553 gclog_or_tty->print_cr("[%d] aborting to deal with pending SATB buffers", 3554 _task_id); 3555 } 3556 // we do need to process SATB buffers, we'll abort and restart 3557 // the marking task to do so 3558 set_has_aborted(); 3559 statsOnly( ++_aborted_satb ); 3560 return; 3561 } 3562 } 3563 3564 void CMTask::recalculate_limits() { 3565 _real_words_scanned_limit = _words_scanned + words_scanned_period; 3566 _words_scanned_limit = _real_words_scanned_limit; 3567 3568 _real_refs_reached_limit = _refs_reached + refs_reached_period; 3569 _refs_reached_limit = _real_refs_reached_limit; 3570 } 3571 3572 void CMTask::decrease_limits() { 3573 // This is called when we believe that we're going to do an infrequent 3574 // operation which will increase the per byte scanned cost (i.e. move 3575 // entries to/from the global stack). It basically tries to decrease the 3576 // scanning limit so that the clock is called earlier. 3577 3578 if (_cm->verbose_medium()) { 3579 gclog_or_tty->print_cr("[%d] decreasing limits", _task_id); 3580 } 3581 3582 _words_scanned_limit = _real_words_scanned_limit - 3583 3 * words_scanned_period / 4; 3584 _refs_reached_limit = _real_refs_reached_limit - 3585 3 * refs_reached_period / 4; 3586 } 3587 3588 void CMTask::move_entries_to_global_stack() { 3589 // local array where we'll store the entries that will be popped 3590 // from the local queue 3591 oop buffer[global_stack_transfer_size]; 3592 3593 int n = 0; 3594 oop obj; 3595 while (n < global_stack_transfer_size && _task_queue->pop_local(obj)) { 3596 buffer[n] = obj; 3597 ++n; 3598 } 3599 3600 if (n > 0) { 3601 // we popped at least one entry from the local queue 3602 3603 statsOnly( ++_global_transfers_to; _local_pops += n ); 3604 3605 if (!_cm->mark_stack_push(buffer, n)) { 3606 if (_cm->verbose_low()) { 3607 gclog_or_tty->print_cr("[%d] aborting due to global stack overflow", 3608 _task_id); 3609 } 3610 set_has_aborted(); 3611 } else { 3612 // the transfer was successful 3613 3614 if (_cm->verbose_medium()) { 3615 gclog_or_tty->print_cr("[%d] pushed %d entries to the global stack", 3616 _task_id, n); 3617 } 3618 statsOnly( int tmp_size = _cm->mark_stack_size(); 3619 if (tmp_size > _global_max_size) { 3620 _global_max_size = tmp_size; 3621 } 3622 _global_pushes += n ); 3623 } 3624 } 3625 3626 // this operation was quite expensive, so decrease the limits 3627 decrease_limits(); 3628 } 3629 3630 void CMTask::get_entries_from_global_stack() { 3631 // local array where we'll store the entries that will be popped 3632 // from the global stack. 3633 oop buffer[global_stack_transfer_size]; 3634 int n; 3635 _cm->mark_stack_pop(buffer, global_stack_transfer_size, &n); 3636 assert(n <= global_stack_transfer_size, 3637 "we should not pop more than the given limit"); 3638 if (n > 0) { 3639 // yes, we did actually pop at least one entry 3640 3641 statsOnly( ++_global_transfers_from; _global_pops += n ); 3642 if (_cm->verbose_medium()) { 3643 gclog_or_tty->print_cr("[%d] popped %d entries from the global stack", 3644 _task_id, n); 3645 } 3646 for (int i = 0; i < n; ++i) { 3647 bool success = _task_queue->push(buffer[i]); 3648 // We only call this when the local queue is empty or under a 3649 // given target limit. So, we do not expect this push to fail. 3650 assert(success, "invariant"); 3651 } 3652 3653 statsOnly( int tmp_size = _task_queue->size(); 3654 if (tmp_size > _local_max_size) { 3655 _local_max_size = tmp_size; 3656 } 3657 _local_pushes += n ); 3658 } 3659 3660 // this operation was quite expensive, so decrease the limits 3661 decrease_limits(); 3662 } 3663 3664 void CMTask::drain_local_queue(bool partially) { 3665 if (has_aborted()) return; 3666 3667 // Decide what the target size is, depending whether we're going to 3668 // drain it partially (so that other tasks can steal if they run out 3669 // of things to do) or totally (at the very end). 3670 size_t target_size; 3671 if (partially) { 3672 target_size = MIN2((size_t)_task_queue->max_elems()/3, GCDrainStackTargetSize); 3673 } else { 3674 target_size = 0; 3675 } 3676 3677 if (_task_queue->size() > target_size) { 3678 if (_cm->verbose_high()) { 3679 gclog_or_tty->print_cr("[%d] draining local queue, target size = %d", 3680 _task_id, target_size); 3681 } 3682 3683 oop obj; 3684 bool ret = _task_queue->pop_local(obj); 3685 while (ret) { 3686 statsOnly( ++_local_pops ); 3687 3688 if (_cm->verbose_high()) { 3689 gclog_or_tty->print_cr("[%d] popped "PTR_FORMAT, _task_id, 3690 (void*) obj); 3691 } 3692 3693 assert(_g1h->is_in_g1_reserved((HeapWord*) obj), "invariant" ); 3694 assert(!_g1h->is_on_master_free_list( 3695 _g1h->heap_region_containing((HeapWord*) obj)), "invariant"); 3696 3697 scan_object(obj); 3698 3699 if (_task_queue->size() <= target_size || has_aborted()) { 3700 ret = false; 3701 } else { 3702 ret = _task_queue->pop_local(obj); 3703 } 3704 } 3705 3706 if (_cm->verbose_high()) { 3707 gclog_or_tty->print_cr("[%d] drained local queue, size = %d", 3708 _task_id, _task_queue->size()); 3709 } 3710 } 3711 } 3712 3713 void CMTask::drain_global_stack(bool partially) { 3714 if (has_aborted()) return; 3715 3716 // We have a policy to drain the local queue before we attempt to 3717 // drain the global stack. 3718 assert(partially || _task_queue->size() == 0, "invariant"); 3719 3720 // Decide what the target size is, depending whether we're going to 3721 // drain it partially (so that other tasks can steal if they run out 3722 // of things to do) or totally (at the very end). Notice that, 3723 // because we move entries from the global stack in chunks or 3724 // because another task might be doing the same, we might in fact 3725 // drop below the target. But, this is not a problem. 3726 size_t target_size; 3727 if (partially) { 3728 target_size = _cm->partial_mark_stack_size_target(); 3729 } else { 3730 target_size = 0; 3731 } 3732 3733 if (_cm->mark_stack_size() > target_size) { 3734 if (_cm->verbose_low()) { 3735 gclog_or_tty->print_cr("[%d] draining global_stack, target size %d", 3736 _task_id, target_size); 3737 } 3738 3739 while (!has_aborted() && _cm->mark_stack_size() > target_size) { 3740 get_entries_from_global_stack(); 3741 drain_local_queue(partially); 3742 } 3743 3744 if (_cm->verbose_low()) { 3745 gclog_or_tty->print_cr("[%d] drained global stack, size = %d", 3746 _task_id, _cm->mark_stack_size()); 3747 } 3748 } 3749 } 3750 3751 // SATB Queue has several assumptions on whether to call the par or 3752 // non-par versions of the methods. this is why some of the code is 3753 // replicated. We should really get rid of the single-threaded version 3754 // of the code to simplify things. 3755 void CMTask::drain_satb_buffers() { 3756 if (has_aborted()) return; 3757 3758 // We set this so that the regular clock knows that we're in the 3759 // middle of draining buffers and doesn't set the abort flag when it 3760 // notices that SATB buffers are available for draining. It'd be 3761 // very counter productive if it did that. :-) 3762 _draining_satb_buffers = true; 3763 3764 CMObjectClosure oc(this); 3765 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set(); 3766 if (G1CollectedHeap::use_parallel_gc_threads()) { 3767 satb_mq_set.set_par_closure(_task_id, &oc); 3768 } else { 3769 satb_mq_set.set_closure(&oc); 3770 } 3771 3772 // This keeps claiming and applying the closure to completed buffers 3773 // until we run out of buffers or we need to abort. 3774 if (G1CollectedHeap::use_parallel_gc_threads()) { 3775 while (!has_aborted() && 3776 satb_mq_set.par_apply_closure_to_completed_buffer(_task_id)) { 3777 if (_cm->verbose_medium()) { 3778 gclog_or_tty->print_cr("[%d] processed an SATB buffer", _task_id); 3779 } 3780 statsOnly( ++_satb_buffers_processed ); 3781 regular_clock_call(); 3782 } 3783 } else { 3784 while (!has_aborted() && 3785 satb_mq_set.apply_closure_to_completed_buffer()) { 3786 if (_cm->verbose_medium()) { 3787 gclog_or_tty->print_cr("[%d] processed an SATB buffer", _task_id); 3788 } 3789 statsOnly( ++_satb_buffers_processed ); 3790 regular_clock_call(); 3791 } 3792 } 3793 3794 if (!concurrent() && !has_aborted()) { 3795 // We should only do this during remark. 3796 if (G1CollectedHeap::use_parallel_gc_threads()) { 3797 satb_mq_set.par_iterate_closure_all_threads(_task_id); 3798 } else { 3799 satb_mq_set.iterate_closure_all_threads(); 3800 } 3801 } 3802 3803 _draining_satb_buffers = false; 3804 3805 assert(has_aborted() || 3806 concurrent() || 3807 satb_mq_set.completed_buffers_num() == 0, "invariant"); 3808 3809 if (G1CollectedHeap::use_parallel_gc_threads()) { 3810 satb_mq_set.set_par_closure(_task_id, NULL); 3811 } else { 3812 satb_mq_set.set_closure(NULL); 3813 } 3814 3815 // again, this was a potentially expensive operation, decrease the 3816 // limits to get the regular clock call early 3817 decrease_limits(); 3818 } 3819 3820 void CMTask::drain_region_stack(BitMapClosure* bc) { 3821 if (has_aborted()) return; 3822 3823 assert(_region_finger == NULL, 3824 "it should be NULL when we're not scanning a region"); 3825 3826 if (!_cm->region_stack_empty() || !_aborted_region.is_empty()) { 3827 if (_cm->verbose_low()) { 3828 gclog_or_tty->print_cr("[%d] draining region stack, size = %d", 3829 _task_id, _cm->region_stack_size()); 3830 } 3831 3832 MemRegion mr; 3833 3834 if (!_aborted_region.is_empty()) { 3835 mr = _aborted_region; 3836 _aborted_region = MemRegion(); 3837 3838 if (_cm->verbose_low()) { 3839 gclog_or_tty->print_cr("[%d] scanning aborted region " 3840 "[ " PTR_FORMAT ", " PTR_FORMAT " )", 3841 _task_id, mr.start(), mr.end()); 3842 } 3843 } else { 3844 mr = _cm->region_stack_pop_lock_free(); 3845 // it returns MemRegion() if the pop fails 3846 statsOnly(if (mr.start() != NULL) ++_region_stack_pops ); 3847 } 3848 3849 while (mr.start() != NULL) { 3850 if (_cm->verbose_medium()) { 3851 gclog_or_tty->print_cr("[%d] we are scanning region " 3852 "["PTR_FORMAT", "PTR_FORMAT")", 3853 _task_id, mr.start(), mr.end()); 3854 } 3855 3856 assert(mr.end() <= _cm->finger(), 3857 "otherwise the region shouldn't be on the stack"); 3858 assert(!mr.is_empty(), "Only non-empty regions live on the region stack"); 3859 if (_nextMarkBitMap->iterate(bc, mr)) { 3860 assert(!has_aborted(), 3861 "cannot abort the task without aborting the bitmap iteration"); 3862 3863 // We finished iterating over the region without aborting. 3864 regular_clock_call(); 3865 if (has_aborted()) { 3866 mr = MemRegion(); 3867 } else { 3868 mr = _cm->region_stack_pop_lock_free(); 3869 // it returns MemRegion() if the pop fails 3870 statsOnly(if (mr.start() != NULL) ++_region_stack_pops ); 3871 } 3872 } else { 3873 assert(has_aborted(), "currently the only way to do so"); 3874 3875 // The only way to abort the bitmap iteration is to return 3876 // false from the do_bit() method. However, inside the 3877 // do_bit() method we move the _region_finger to point to the 3878 // object currently being looked at. So, if we bail out, we 3879 // have definitely set _region_finger to something non-null. 3880 assert(_region_finger != NULL, "invariant"); 3881 3882 // Make sure that any previously aborted region has been 3883 // cleared. 3884 assert(_aborted_region.is_empty(), "aborted region not cleared"); 3885 3886 // The iteration was actually aborted. So now _region_finger 3887 // points to the address of the object we last scanned. If we 3888 // leave it there, when we restart this task, we will rescan 3889 // the object. It is easy to avoid this. We move the finger by 3890 // enough to point to the next possible object header (the 3891 // bitmap knows by how much we need to move it as it knows its 3892 // granularity). 3893 MemRegion newRegion = 3894 MemRegion(_nextMarkBitMap->nextWord(_region_finger), mr.end()); 3895 3896 if (!newRegion.is_empty()) { 3897 if (_cm->verbose_low()) { 3898 gclog_or_tty->print_cr("[%d] recording unscanned region" 3899 "[" PTR_FORMAT "," PTR_FORMAT ") in CMTask", 3900 _task_id, 3901 newRegion.start(), newRegion.end()); 3902 } 3903 // Now record the part of the region we didn't scan to 3904 // make sure this task scans it later. 3905 _aborted_region = newRegion; 3906 } 3907 // break from while 3908 mr = MemRegion(); 3909 } 3910 _region_finger = NULL; 3911 } 3912 3913 if (_cm->verbose_low()) { 3914 gclog_or_tty->print_cr("[%d] drained region stack, size = %d", 3915 _task_id, _cm->region_stack_size()); 3916 } 3917 } 3918 } 3919 3920 void CMTask::print_stats() { 3921 gclog_or_tty->print_cr("Marking Stats, task = %d, calls = %d", 3922 _task_id, _calls); 3923 gclog_or_tty->print_cr(" Elapsed time = %1.2lfms, Termination time = %1.2lfms", 3924 _elapsed_time_ms, _termination_time_ms); 3925 gclog_or_tty->print_cr(" Step Times (cum): num = %d, avg = %1.2lfms, sd = %1.2lfms", 3926 _step_times_ms.num(), _step_times_ms.avg(), 3927 _step_times_ms.sd()); 3928 gclog_or_tty->print_cr(" max = %1.2lfms, total = %1.2lfms", 3929 _step_times_ms.maximum(), _step_times_ms.sum()); 3930 3931 #if _MARKING_STATS_ 3932 gclog_or_tty->print_cr(" Clock Intervals (cum): num = %d, avg = %1.2lfms, sd = %1.2lfms", 3933 _all_clock_intervals_ms.num(), _all_clock_intervals_ms.avg(), 3934 _all_clock_intervals_ms.sd()); 3935 gclog_or_tty->print_cr(" max = %1.2lfms, total = %1.2lfms", 3936 _all_clock_intervals_ms.maximum(), 3937 _all_clock_intervals_ms.sum()); 3938 gclog_or_tty->print_cr(" Clock Causes (cum): scanning = %d, marking = %d", 3939 _clock_due_to_scanning, _clock_due_to_marking); 3940 gclog_or_tty->print_cr(" Objects: scanned = %d, found on the bitmap = %d", 3941 _objs_scanned, _objs_found_on_bitmap); 3942 gclog_or_tty->print_cr(" Local Queue: pushes = %d, pops = %d, max size = %d", 3943 _local_pushes, _local_pops, _local_max_size); 3944 gclog_or_tty->print_cr(" Global Stack: pushes = %d, pops = %d, max size = %d", 3945 _global_pushes, _global_pops, _global_max_size); 3946 gclog_or_tty->print_cr(" transfers to = %d, transfers from = %d", 3947 _global_transfers_to,_global_transfers_from); 3948 gclog_or_tty->print_cr(" Regions: claimed = %d, Region Stack: pops = %d", 3949 _regions_claimed, _region_stack_pops); 3950 gclog_or_tty->print_cr(" SATB buffers: processed = %d", _satb_buffers_processed); 3951 gclog_or_tty->print_cr(" Steals: attempts = %d, successes = %d", 3952 _steal_attempts, _steals); 3953 gclog_or_tty->print_cr(" Aborted: %d, due to", _aborted); 3954 gclog_or_tty->print_cr(" overflow: %d, global abort: %d, yield: %d", 3955 _aborted_overflow, _aborted_cm_aborted, _aborted_yield); 3956 gclog_or_tty->print_cr(" time out: %d, SATB: %d, termination: %d", 3957 _aborted_timed_out, _aborted_satb, _aborted_termination); 3958 #endif // _MARKING_STATS_ 3959 } 3960 3961 /***************************************************************************** 3962 3963 The do_marking_step(time_target_ms) method is the building block 3964 of the parallel marking framework. It can be called in parallel 3965 with other invocations of do_marking_step() on different tasks 3966 (but only one per task, obviously) and concurrently with the 3967 mutator threads, or during remark, hence it eliminates the need 3968 for two versions of the code. When called during remark, it will 3969 pick up from where the task left off during the concurrent marking 3970 phase. Interestingly, tasks are also claimable during evacuation 3971 pauses too, since do_marking_step() ensures that it aborts before 3972 it needs to yield. 3973 3974 The data structures that is uses to do marking work are the 3975 following: 3976 3977 (1) Marking Bitmap. If there are gray objects that appear only 3978 on the bitmap (this happens either when dealing with an overflow 3979 or when the initial marking phase has simply marked the roots 3980 and didn't push them on the stack), then tasks claim heap 3981 regions whose bitmap they then scan to find gray objects. A 3982 global finger indicates where the end of the last claimed region 3983 is. A local finger indicates how far into the region a task has 3984 scanned. The two fingers are used to determine how to gray an 3985 object (i.e. whether simply marking it is OK, as it will be 3986 visited by a task in the future, or whether it needs to be also 3987 pushed on a stack). 3988 3989 (2) Local Queue. The local queue of the task which is accessed 3990 reasonably efficiently by the task. Other tasks can steal from 3991 it when they run out of work. Throughout the marking phase, a 3992 task attempts to keep its local queue short but not totally 3993 empty, so that entries are available for stealing by other 3994 tasks. Only when there is no more work, a task will totally 3995 drain its local queue. 3996 3997 (3) Global Mark Stack. This handles local queue overflow. During 3998 marking only sets of entries are moved between it and the local 3999 queues, as access to it requires a mutex and more fine-grain 4000 interaction with it which might cause contention. If it 4001 overflows, then the marking phase should restart and iterate 4002 over the bitmap to identify gray objects. Throughout the marking 4003 phase, tasks attempt to keep the global mark stack at a small 4004 length but not totally empty, so that entries are available for 4005 popping by other tasks. Only when there is no more work, tasks 4006 will totally drain the global mark stack. 4007 4008 (4) Global Region Stack. Entries on it correspond to areas of 4009 the bitmap that need to be scanned since they contain gray 4010 objects. Pushes on the region stack only happen during 4011 evacuation pauses and typically correspond to areas covered by 4012 GC LABS. If it overflows, then the marking phase should restart 4013 and iterate over the bitmap to identify gray objects. Tasks will 4014 try to totally drain the region stack as soon as possible. 4015 4016 (5) SATB Buffer Queue. This is where completed SATB buffers are 4017 made available. Buffers are regularly removed from this queue 4018 and scanned for roots, so that the queue doesn't get too 4019 long. During remark, all completed buffers are processed, as 4020 well as the filled in parts of any uncompleted buffers. 4021 4022 The do_marking_step() method tries to abort when the time target 4023 has been reached. There are a few other cases when the 4024 do_marking_step() method also aborts: 4025 4026 (1) When the marking phase has been aborted (after a Full GC). 4027 4028 (2) When a global overflow (either on the global stack or the 4029 region stack) has been triggered. Before the task aborts, it 4030 will actually sync up with the other tasks to ensure that all 4031 the marking data structures (local queues, stacks, fingers etc.) 4032 are re-initialised so that when do_marking_step() completes, 4033 the marking phase can immediately restart. 4034 4035 (3) When enough completed SATB buffers are available. The 4036 do_marking_step() method only tries to drain SATB buffers right 4037 at the beginning. So, if enough buffers are available, the 4038 marking step aborts and the SATB buffers are processed at 4039 the beginning of the next invocation. 4040 4041 (4) To yield. when we have to yield then we abort and yield 4042 right at the end of do_marking_step(). This saves us from a lot 4043 of hassle as, by yielding we might allow a Full GC. If this 4044 happens then objects will be compacted underneath our feet, the 4045 heap might shrink, etc. We save checking for this by just 4046 aborting and doing the yield right at the end. 4047 4048 From the above it follows that the do_marking_step() method should 4049 be called in a loop (or, otherwise, regularly) until it completes. 4050 4051 If a marking step completes without its has_aborted() flag being 4052 true, it means it has completed the current marking phase (and 4053 also all other marking tasks have done so and have all synced up). 4054 4055 A method called regular_clock_call() is invoked "regularly" (in 4056 sub ms intervals) throughout marking. It is this clock method that 4057 checks all the abort conditions which were mentioned above and 4058 decides when the task should abort. A work-based scheme is used to 4059 trigger this clock method: when the number of object words the 4060 marking phase has scanned or the number of references the marking 4061 phase has visited reach a given limit. Additional invocations to 4062 the method clock have been planted in a few other strategic places 4063 too. The initial reason for the clock method was to avoid calling 4064 vtime too regularly, as it is quite expensive. So, once it was in 4065 place, it was natural to piggy-back all the other conditions on it 4066 too and not constantly check them throughout the code. 4067 4068 *****************************************************************************/ 4069 4070 void CMTask::do_marking_step(double time_target_ms, 4071 bool do_stealing, 4072 bool do_termination) { 4073 assert(time_target_ms >= 1.0, "minimum granularity is 1ms"); 4074 assert(concurrent() == _cm->concurrent(), "they should be the same"); 4075 4076 assert(concurrent() || _cm->region_stack_empty(), 4077 "the region stack should have been cleared before remark"); 4078 assert(concurrent() || !_cm->has_aborted_regions(), 4079 "aborted regions should have been cleared before remark"); 4080 assert(_region_finger == NULL, 4081 "this should be non-null only when a region is being scanned"); 4082 4083 G1CollectorPolicy* g1_policy = _g1h->g1_policy(); 4084 assert(_task_queues != NULL, "invariant"); 4085 assert(_task_queue != NULL, "invariant"); 4086 assert(_task_queues->queue(_task_id) == _task_queue, "invariant"); 4087 4088 assert(!_claimed, 4089 "only one thread should claim this task at any one time"); 4090 4091 // OK, this doesn't safeguard again all possible scenarios, as it is 4092 // possible for two threads to set the _claimed flag at the same 4093 // time. But it is only for debugging purposes anyway and it will 4094 // catch most problems. 4095 _claimed = true; 4096 4097 _start_time_ms = os::elapsedVTime() * 1000.0; 4098 statsOnly( _interval_start_time_ms = _start_time_ms ); 4099 4100 double diff_prediction_ms = 4101 g1_policy->get_new_prediction(&_marking_step_diffs_ms); 4102 _time_target_ms = time_target_ms - diff_prediction_ms; 4103 4104 // set up the variables that are used in the work-based scheme to 4105 // call the regular clock method 4106 _words_scanned = 0; 4107 _refs_reached = 0; 4108 recalculate_limits(); 4109 4110 // clear all flags 4111 clear_has_aborted(); 4112 _has_timed_out = false; 4113 _draining_satb_buffers = false; 4114 4115 ++_calls; 4116 4117 if (_cm->verbose_low()) { 4118 gclog_or_tty->print_cr("[%d] >>>>>>>>>> START, call = %d, " 4119 "target = %1.2lfms >>>>>>>>>>", 4120 _task_id, _calls, _time_target_ms); 4121 } 4122 4123 // Set up the bitmap and oop closures. Anything that uses them is 4124 // eventually called from this method, so it is OK to allocate these 4125 // statically. 4126 CMBitMapClosure bitmap_closure(this, _cm, _nextMarkBitMap); 4127 G1CMOopClosure cm_oop_closure(_g1h, _cm, this); 4128 set_cm_oop_closure(&cm_oop_closure); 4129 4130 if (_cm->has_overflown()) { 4131 // This can happen if the region stack or the mark stack overflows 4132 // during a GC pause and this task, after a yield point, 4133 // restarts. We have to abort as we need to get into the overflow 4134 // protocol which happens right at the end of this task. 4135 set_has_aborted(); 4136 } 4137 4138 // First drain any available SATB buffers. After this, we will not 4139 // look at SATB buffers before the next invocation of this method. 4140 // If enough completed SATB buffers are queued up, the regular clock 4141 // will abort this task so that it restarts. 4142 drain_satb_buffers(); 4143 // ...then partially drain the local queue and the global stack 4144 drain_local_queue(true); 4145 drain_global_stack(true); 4146 4147 // Then totally drain the region stack. We will not look at 4148 // it again before the next invocation of this method. Entries on 4149 // the region stack are only added during evacuation pauses, for 4150 // which we have to yield. When we do, we abort the task anyway so 4151 // it will look at the region stack again when it restarts. 4152 bitmap_closure.set_scanning_heap_region(false); 4153 drain_region_stack(&bitmap_closure); 4154 // ...then partially drain the local queue and the global stack 4155 drain_local_queue(true); 4156 drain_global_stack(true); 4157 4158 do { 4159 if (!has_aborted() && _curr_region != NULL) { 4160 // This means that we're already holding on to a region. 4161 assert(_finger != NULL, "if region is not NULL, then the finger " 4162 "should not be NULL either"); 4163 4164 // We might have restarted this task after an evacuation pause 4165 // which might have evacuated the region we're holding on to 4166 // underneath our feet. Let's read its limit again to make sure 4167 // that we do not iterate over a region of the heap that 4168 // contains garbage (update_region_limit() will also move 4169 // _finger to the start of the region if it is found empty). 4170 update_region_limit(); 4171 // We will start from _finger not from the start of the region, 4172 // as we might be restarting this task after aborting half-way 4173 // through scanning this region. In this case, _finger points to 4174 // the address where we last found a marked object. If this is a 4175 // fresh region, _finger points to start(). 4176 MemRegion mr = MemRegion(_finger, _region_limit); 4177 4178 if (_cm->verbose_low()) { 4179 gclog_or_tty->print_cr("[%d] we're scanning part " 4180 "["PTR_FORMAT", "PTR_FORMAT") " 4181 "of region "PTR_FORMAT, 4182 _task_id, _finger, _region_limit, _curr_region); 4183 } 4184 4185 // Let's iterate over the bitmap of the part of the 4186 // region that is left. 4187 bitmap_closure.set_scanning_heap_region(true); 4188 if (mr.is_empty() || 4189 _nextMarkBitMap->iterate(&bitmap_closure, mr)) { 4190 // We successfully completed iterating over the region. Now, 4191 // let's give up the region. 4192 giveup_current_region(); 4193 regular_clock_call(); 4194 } else { 4195 assert(has_aborted(), "currently the only way to do so"); 4196 // The only way to abort the bitmap iteration is to return 4197 // false from the do_bit() method. However, inside the 4198 // do_bit() method we move the _finger to point to the 4199 // object currently being looked at. So, if we bail out, we 4200 // have definitely set _finger to something non-null. 4201 assert(_finger != NULL, "invariant"); 4202 4203 // Region iteration was actually aborted. So now _finger 4204 // points to the address of the object we last scanned. If we 4205 // leave it there, when we restart this task, we will rescan 4206 // the object. It is easy to avoid this. We move the finger by 4207 // enough to point to the next possible object header (the 4208 // bitmap knows by how much we need to move it as it knows its 4209 // granularity). 4210 assert(_finger < _region_limit, "invariant"); 4211 HeapWord* new_finger = _nextMarkBitMap->nextWord(_finger); 4212 // Check if bitmap iteration was aborted while scanning the last object 4213 if (new_finger >= _region_limit) { 4214 giveup_current_region(); 4215 } else { 4216 move_finger_to(new_finger); 4217 } 4218 } 4219 } 4220 // At this point we have either completed iterating over the 4221 // region we were holding on to, or we have aborted. 4222 4223 // We then partially drain the local queue and the global stack. 4224 // (Do we really need this?) 4225 drain_local_queue(true); 4226 drain_global_stack(true); 4227 4228 // Read the note on the claim_region() method on why it might 4229 // return NULL with potentially more regions available for 4230 // claiming and why we have to check out_of_regions() to determine 4231 // whether we're done or not. 4232 while (!has_aborted() && _curr_region == NULL && !_cm->out_of_regions()) { 4233 // We are going to try to claim a new region. We should have 4234 // given up on the previous one. 4235 // Separated the asserts so that we know which one fires. 4236 assert(_curr_region == NULL, "invariant"); 4237 assert(_finger == NULL, "invariant"); 4238 assert(_region_limit == NULL, "invariant"); 4239 if (_cm->verbose_low()) { 4240 gclog_or_tty->print_cr("[%d] trying to claim a new region", _task_id); 4241 } 4242 HeapRegion* claimed_region = _cm->claim_region(_task_id); 4243 if (claimed_region != NULL) { 4244 // Yes, we managed to claim one 4245 statsOnly( ++_regions_claimed ); 4246 4247 if (_cm->verbose_low()) { 4248 gclog_or_tty->print_cr("[%d] we successfully claimed " 4249 "region "PTR_FORMAT, 4250 _task_id, claimed_region); 4251 } 4252 4253 setup_for_region(claimed_region); 4254 assert(_curr_region == claimed_region, "invariant"); 4255 } 4256 // It is important to call the regular clock here. It might take 4257 // a while to claim a region if, for example, we hit a large 4258 // block of empty regions. So we need to call the regular clock 4259 // method once round the loop to make sure it's called 4260 // frequently enough. 4261 regular_clock_call(); 4262 } 4263 4264 if (!has_aborted() && _curr_region == NULL) { 4265 assert(_cm->out_of_regions(), 4266 "at this point we should be out of regions"); 4267 } 4268 } while ( _curr_region != NULL && !has_aborted()); 4269 4270 if (!has_aborted()) { 4271 // We cannot check whether the global stack is empty, since other 4272 // tasks might be pushing objects to it concurrently. We also cannot 4273 // check if the region stack is empty because if a thread is aborting 4274 // it can push a partially done region back. 4275 assert(_cm->out_of_regions(), 4276 "at this point we should be out of regions"); 4277 4278 if (_cm->verbose_low()) { 4279 gclog_or_tty->print_cr("[%d] all regions claimed", _task_id); 4280 } 4281 4282 // Try to reduce the number of available SATB buffers so that 4283 // remark has less work to do. 4284 drain_satb_buffers(); 4285 } 4286 4287 // Since we've done everything else, we can now totally drain the 4288 // local queue and global stack. 4289 drain_local_queue(false); 4290 drain_global_stack(false); 4291 4292 // Attempt at work stealing from other task's queues. 4293 if (do_stealing && !has_aborted()) { 4294 // We have not aborted. This means that we have finished all that 4295 // we could. Let's try to do some stealing... 4296 4297 // We cannot check whether the global stack is empty, since other 4298 // tasks might be pushing objects to it concurrently. We also cannot 4299 // check if the region stack is empty because if a thread is aborting 4300 // it can push a partially done region back. 4301 assert(_cm->out_of_regions() && _task_queue->size() == 0, 4302 "only way to reach here"); 4303 4304 if (_cm->verbose_low()) { 4305 gclog_or_tty->print_cr("[%d] starting to steal", _task_id); 4306 } 4307 4308 while (!has_aborted()) { 4309 oop obj; 4310 statsOnly( ++_steal_attempts ); 4311 4312 if (_cm->try_stealing(_task_id, &_hash_seed, obj)) { 4313 if (_cm->verbose_medium()) { 4314 gclog_or_tty->print_cr("[%d] stolen "PTR_FORMAT" successfully", 4315 _task_id, (void*) obj); 4316 } 4317 4318 statsOnly( ++_steals ); 4319 4320 assert(_nextMarkBitMap->isMarked((HeapWord*) obj), 4321 "any stolen object should be marked"); 4322 scan_object(obj); 4323 4324 // And since we're towards the end, let's totally drain the 4325 // local queue and global stack. 4326 drain_local_queue(false); 4327 drain_global_stack(false); 4328 } else { 4329 break; 4330 } 4331 } 4332 } 4333 4334 // If we are about to wrap up and go into termination, check if we 4335 // should raise the overflow flag. 4336 if (do_termination && !has_aborted()) { 4337 if (_cm->force_overflow()->should_force()) { 4338 _cm->set_has_overflown(); 4339 regular_clock_call(); 4340 } 4341 } 4342 4343 // We still haven't aborted. Now, let's try to get into the 4344 // termination protocol. 4345 if (do_termination && !has_aborted()) { 4346 // We cannot check whether the global stack is empty, since other 4347 // tasks might be concurrently pushing objects on it. We also cannot 4348 // check if the region stack is empty because if a thread is aborting 4349 // it can push a partially done region back. 4350 // Separated the asserts so that we know which one fires. 4351 assert(_cm->out_of_regions(), "only way to reach here"); 4352 assert(_task_queue->size() == 0, "only way to reach here"); 4353 4354 if (_cm->verbose_low()) { 4355 gclog_or_tty->print_cr("[%d] starting termination protocol", _task_id); 4356 } 4357 4358 _termination_start_time_ms = os::elapsedVTime() * 1000.0; 4359 // The CMTask class also extends the TerminatorTerminator class, 4360 // hence its should_exit_termination() method will also decide 4361 // whether to exit the termination protocol or not. 4362 bool finished = _cm->terminator()->offer_termination(this); 4363 double termination_end_time_ms = os::elapsedVTime() * 1000.0; 4364 _termination_time_ms += 4365 termination_end_time_ms - _termination_start_time_ms; 4366 4367 if (finished) { 4368 // We're all done. 4369 4370 if (_task_id == 0) { 4371 // let's allow task 0 to do this 4372 if (concurrent()) { 4373 assert(_cm->concurrent_marking_in_progress(), "invariant"); 4374 // we need to set this to false before the next 4375 // safepoint. This way we ensure that the marking phase 4376 // doesn't observe any more heap expansions. 4377 _cm->clear_concurrent_marking_in_progress(); 4378 } 4379 } 4380 4381 // We can now guarantee that the global stack is empty, since 4382 // all other tasks have finished. We separated the guarantees so 4383 // that, if a condition is false, we can immediately find out 4384 // which one. 4385 guarantee(_cm->out_of_regions(), "only way to reach here"); 4386 guarantee(_aborted_region.is_empty(), "only way to reach here"); 4387 guarantee(_cm->region_stack_empty(), "only way to reach here"); 4388 guarantee(_cm->mark_stack_empty(), "only way to reach here"); 4389 guarantee(_task_queue->size() == 0, "only way to reach here"); 4390 guarantee(!_cm->has_overflown(), "only way to reach here"); 4391 guarantee(!_cm->mark_stack_overflow(), "only way to reach here"); 4392 guarantee(!_cm->region_stack_overflow(), "only way to reach here"); 4393 4394 if (_cm->verbose_low()) { 4395 gclog_or_tty->print_cr("[%d] all tasks terminated", _task_id); 4396 } 4397 } else { 4398 // Apparently there's more work to do. Let's abort this task. It 4399 // will restart it and we can hopefully find more things to do. 4400 4401 if (_cm->verbose_low()) { 4402 gclog_or_tty->print_cr("[%d] apparently there is more work to do", 4403 _task_id); 4404 } 4405 4406 set_has_aborted(); 4407 statsOnly( ++_aborted_termination ); 4408 } 4409 } 4410 4411 // Mainly for debugging purposes to make sure that a pointer to the 4412 // closure which was statically allocated in this frame doesn't 4413 // escape it by accident. 4414 set_cm_oop_closure(NULL); 4415 double end_time_ms = os::elapsedVTime() * 1000.0; 4416 double elapsed_time_ms = end_time_ms - _start_time_ms; 4417 // Update the step history. 4418 _step_times_ms.add(elapsed_time_ms); 4419 4420 if (has_aborted()) { 4421 // The task was aborted for some reason. 4422 4423 statsOnly( ++_aborted ); 4424 4425 if (_has_timed_out) { 4426 double diff_ms = elapsed_time_ms - _time_target_ms; 4427 // Keep statistics of how well we did with respect to hitting 4428 // our target only if we actually timed out (if we aborted for 4429 // other reasons, then the results might get skewed). 4430 _marking_step_diffs_ms.add(diff_ms); 4431 } 4432 4433 if (_cm->has_overflown()) { 4434 // This is the interesting one. We aborted because a global 4435 // overflow was raised. This means we have to restart the 4436 // marking phase and start iterating over regions. However, in 4437 // order to do this we have to make sure that all tasks stop 4438 // what they are doing and re-initialise in a safe manner. We 4439 // will achieve this with the use of two barrier sync points. 4440 4441 if (_cm->verbose_low()) { 4442 gclog_or_tty->print_cr("[%d] detected overflow", _task_id); 4443 } 4444 4445 _cm->enter_first_sync_barrier(_task_id); 4446 // When we exit this sync barrier we know that all tasks have 4447 // stopped doing marking work. So, it's now safe to 4448 // re-initialise our data structures. At the end of this method, 4449 // task 0 will clear the global data structures. 4450 4451 statsOnly( ++_aborted_overflow ); 4452 4453 // We clear the local state of this task... 4454 clear_region_fields(); 4455 4456 // ...and enter the second barrier. 4457 _cm->enter_second_sync_barrier(_task_id); 4458 // At this point everything has bee re-initialised and we're 4459 // ready to restart. 4460 } 4461 4462 if (_cm->verbose_low()) { 4463 gclog_or_tty->print_cr("[%d] <<<<<<<<<< ABORTING, target = %1.2lfms, " 4464 "elapsed = %1.2lfms <<<<<<<<<<", 4465 _task_id, _time_target_ms, elapsed_time_ms); 4466 if (_cm->has_aborted()) { 4467 gclog_or_tty->print_cr("[%d] ========== MARKING ABORTED ==========", 4468 _task_id); 4469 } 4470 } 4471 } else { 4472 if (_cm->verbose_low()) { 4473 gclog_or_tty->print_cr("[%d] <<<<<<<<<< FINISHED, target = %1.2lfms, " 4474 "elapsed = %1.2lfms <<<<<<<<<<", 4475 _task_id, _time_target_ms, elapsed_time_ms); 4476 } 4477 } 4478 4479 _claimed = false; 4480 } 4481 4482 CMTask::CMTask(int task_id, 4483 ConcurrentMark* cm, 4484 CMTaskQueue* task_queue, 4485 CMTaskQueueSet* task_queues) 4486 : _g1h(G1CollectedHeap::heap()), 4487 _task_id(task_id), _cm(cm), 4488 _claimed(false), 4489 _nextMarkBitMap(NULL), _hash_seed(17), 4490 _task_queue(task_queue), 4491 _task_queues(task_queues), 4492 _cm_oop_closure(NULL), 4493 _aborted_region(MemRegion()) { 4494 guarantee(task_queue != NULL, "invariant"); 4495 guarantee(task_queues != NULL, "invariant"); 4496 4497 statsOnly( _clock_due_to_scanning = 0; 4498 _clock_due_to_marking = 0 ); 4499 4500 _marking_step_diffs_ms.add(0.5); 4501 } 4502 4503 // These are formatting macros that are used below to ensure 4504 // consistent formatting. The *_H_* versions are used to format the 4505 // header for a particular value and they should be kept consistent 4506 // with the corresponding macro. Also note that most of the macros add 4507 // the necessary white space (as a prefix) which makes them a bit 4508 // easier to compose. 4509 4510 // All the output lines are prefixed with this string to be able to 4511 // identify them easily in a large log file. 4512 #define G1PPRL_LINE_PREFIX "###" 4513 4514 #define G1PPRL_ADDR_BASE_FORMAT " "PTR_FORMAT"-"PTR_FORMAT 4515 #ifdef _LP64 4516 #define G1PPRL_ADDR_BASE_H_FORMAT " %37s" 4517 #else // _LP64 4518 #define G1PPRL_ADDR_BASE_H_FORMAT " %21s" 4519 #endif // _LP64 4520 4521 // For per-region info 4522 #define G1PPRL_TYPE_FORMAT " %-4s" 4523 #define G1PPRL_TYPE_H_FORMAT " %4s" 4524 #define G1PPRL_BYTE_FORMAT " "SIZE_FORMAT_W(9) 4525 #define G1PPRL_BYTE_H_FORMAT " %9s" 4526 #define G1PPRL_DOUBLE_FORMAT " %14.1f" 4527 #define G1PPRL_DOUBLE_H_FORMAT " %14s" 4528 4529 // For summary info 4530 #define G1PPRL_SUM_ADDR_FORMAT(tag) " "tag":"G1PPRL_ADDR_BASE_FORMAT 4531 #define G1PPRL_SUM_BYTE_FORMAT(tag) " "tag": "SIZE_FORMAT 4532 #define G1PPRL_SUM_MB_FORMAT(tag) " "tag": %1.2f MB" 4533 #define G1PPRL_SUM_MB_PERC_FORMAT(tag) G1PPRL_SUM_MB_FORMAT(tag)" / %1.2f %%" 4534 4535 G1PrintRegionLivenessInfoClosure:: 4536 G1PrintRegionLivenessInfoClosure(outputStream* out, const char* phase_name) 4537 : _out(out), 4538 _total_used_bytes(0), _total_capacity_bytes(0), 4539 _total_prev_live_bytes(0), _total_next_live_bytes(0), 4540 _hum_used_bytes(0), _hum_capacity_bytes(0), 4541 _hum_prev_live_bytes(0), _hum_next_live_bytes(0) { 4542 G1CollectedHeap* g1h = G1CollectedHeap::heap(); 4543 MemRegion g1_committed = g1h->g1_committed(); 4544 MemRegion g1_reserved = g1h->g1_reserved(); 4545 double now = os::elapsedTime(); 4546 4547 // Print the header of the output. 4548 _out->cr(); 4549 _out->print_cr(G1PPRL_LINE_PREFIX" PHASE %s @ %1.3f", phase_name, now); 4550 _out->print_cr(G1PPRL_LINE_PREFIX" HEAP" 4551 G1PPRL_SUM_ADDR_FORMAT("committed") 4552 G1PPRL_SUM_ADDR_FORMAT("reserved") 4553 G1PPRL_SUM_BYTE_FORMAT("region-size"), 4554 g1_committed.start(), g1_committed.end(), 4555 g1_reserved.start(), g1_reserved.end(), 4556 (size_t)HeapRegion::GrainBytes); 4557 _out->print_cr(G1PPRL_LINE_PREFIX); 4558 _out->print_cr(G1PPRL_LINE_PREFIX 4559 G1PPRL_TYPE_H_FORMAT 4560 G1PPRL_ADDR_BASE_H_FORMAT 4561 G1PPRL_BYTE_H_FORMAT 4562 G1PPRL_BYTE_H_FORMAT 4563 G1PPRL_BYTE_H_FORMAT 4564 G1PPRL_DOUBLE_H_FORMAT, 4565 "type", "address-range", 4566 "used", "prev-live", "next-live", "gc-eff"); 4567 } 4568 4569 // It takes as a parameter a reference to one of the _hum_* fields, it 4570 // deduces the corresponding value for a region in a humongous region 4571 // series (either the region size, or what's left if the _hum_* field 4572 // is < the region size), and updates the _hum_* field accordingly. 4573 size_t G1PrintRegionLivenessInfoClosure::get_hum_bytes(size_t* hum_bytes) { 4574 size_t bytes = 0; 4575 // The > 0 check is to deal with the prev and next live bytes which 4576 // could be 0. 4577 if (*hum_bytes > 0) { 4578 bytes = MIN2((size_t) HeapRegion::GrainBytes, *hum_bytes); 4579 *hum_bytes -= bytes; 4580 } 4581 return bytes; 4582 } 4583 4584 // It deduces the values for a region in a humongous region series 4585 // from the _hum_* fields and updates those accordingly. It assumes 4586 // that that _hum_* fields have already been set up from the "starts 4587 // humongous" region and we visit the regions in address order. 4588 void G1PrintRegionLivenessInfoClosure::get_hum_bytes(size_t* used_bytes, 4589 size_t* capacity_bytes, 4590 size_t* prev_live_bytes, 4591 size_t* next_live_bytes) { 4592 assert(_hum_used_bytes > 0 && _hum_capacity_bytes > 0, "pre-condition"); 4593 *used_bytes = get_hum_bytes(&_hum_used_bytes); 4594 *capacity_bytes = get_hum_bytes(&_hum_capacity_bytes); 4595 *prev_live_bytes = get_hum_bytes(&_hum_prev_live_bytes); 4596 *next_live_bytes = get_hum_bytes(&_hum_next_live_bytes); 4597 } 4598 4599 bool G1PrintRegionLivenessInfoClosure::doHeapRegion(HeapRegion* r) { 4600 const char* type = ""; 4601 HeapWord* bottom = r->bottom(); 4602 HeapWord* end = r->end(); 4603 size_t capacity_bytes = r->capacity(); 4604 size_t used_bytes = r->used(); 4605 size_t prev_live_bytes = r->live_bytes(); 4606 size_t next_live_bytes = r->next_live_bytes(); 4607 double gc_eff = r->gc_efficiency(); 4608 if (r->used() == 0) { 4609 type = "FREE"; 4610 } else if (r->is_survivor()) { 4611 type = "SURV"; 4612 } else if (r->is_young()) { 4613 type = "EDEN"; 4614 } else if (r->startsHumongous()) { 4615 type = "HUMS"; 4616 4617 assert(_hum_used_bytes == 0 && _hum_capacity_bytes == 0 && 4618 _hum_prev_live_bytes == 0 && _hum_next_live_bytes == 0, 4619 "they should have been zeroed after the last time we used them"); 4620 // Set up the _hum_* fields. 4621 _hum_capacity_bytes = capacity_bytes; 4622 _hum_used_bytes = used_bytes; 4623 _hum_prev_live_bytes = prev_live_bytes; 4624 _hum_next_live_bytes = next_live_bytes; 4625 get_hum_bytes(&used_bytes, &capacity_bytes, 4626 &prev_live_bytes, &next_live_bytes); 4627 end = bottom + HeapRegion::GrainWords; 4628 } else if (r->continuesHumongous()) { 4629 type = "HUMC"; 4630 get_hum_bytes(&used_bytes, &capacity_bytes, 4631 &prev_live_bytes, &next_live_bytes); 4632 assert(end == bottom + HeapRegion::GrainWords, "invariant"); 4633 } else { 4634 type = "OLD"; 4635 } 4636 4637 _total_used_bytes += used_bytes; 4638 _total_capacity_bytes += capacity_bytes; 4639 _total_prev_live_bytes += prev_live_bytes; 4640 _total_next_live_bytes += next_live_bytes; 4641 4642 // Print a line for this particular region. 4643 _out->print_cr(G1PPRL_LINE_PREFIX 4644 G1PPRL_TYPE_FORMAT 4645 G1PPRL_ADDR_BASE_FORMAT 4646 G1PPRL_BYTE_FORMAT 4647 G1PPRL_BYTE_FORMAT 4648 G1PPRL_BYTE_FORMAT 4649 G1PPRL_DOUBLE_FORMAT, 4650 type, bottom, end, 4651 used_bytes, prev_live_bytes, next_live_bytes, gc_eff); 4652 4653 return false; 4654 } 4655 4656 G1PrintRegionLivenessInfoClosure::~G1PrintRegionLivenessInfoClosure() { 4657 // Print the footer of the output. 4658 _out->print_cr(G1PPRL_LINE_PREFIX); 4659 _out->print_cr(G1PPRL_LINE_PREFIX 4660 " SUMMARY" 4661 G1PPRL_SUM_MB_FORMAT("capacity") 4662 G1PPRL_SUM_MB_PERC_FORMAT("used") 4663 G1PPRL_SUM_MB_PERC_FORMAT("prev-live") 4664 G1PPRL_SUM_MB_PERC_FORMAT("next-live"), 4665 bytes_to_mb(_total_capacity_bytes), 4666 bytes_to_mb(_total_used_bytes), 4667 perc(_total_used_bytes, _total_capacity_bytes), 4668 bytes_to_mb(_total_prev_live_bytes), 4669 perc(_total_prev_live_bytes, _total_capacity_bytes), 4670 bytes_to_mb(_total_next_live_bytes), 4671 perc(_total_next_live_bytes, _total_capacity_bytes)); 4672 _out->cr(); 4673 }