1 /*
2 * Copyright (c) 2001, 2013, 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/g1Log.hpp"
33 #include "gc_implementation/g1/g1OopClosures.inline.hpp"
34 #include "gc_implementation/g1/g1RemSet.hpp"
35 #include "gc_implementation/g1/heapRegion.inline.hpp"
36 #include "gc_implementation/g1/heapRegionRemSet.hpp"
37 #include "gc_implementation/g1/heapRegionSeq.inline.hpp"
38 #include "gc_implementation/shared/vmGCOperations.hpp"
39 #include "memory/genOopClosures.inline.hpp"
40 #include "memory/referencePolicy.hpp"
41 #include "memory/resourceArea.hpp"
42 #include "oops/oop.inline.hpp"
43 #include "runtime/handles.inline.hpp"
44 #include "runtime/java.hpp"
45 #include "services/memTracker.hpp"
46
47 // Concurrent marking bit map wrapper
48
49 CMBitMapRO::CMBitMapRO(int shifter) :
50 _bm(),
51 _shifter(shifter) {
52 _bmStartWord = 0;
53 _bmWordSize = 0;
54 }
55
56 HeapWord* CMBitMapRO::getNextMarkedWordAddress(HeapWord* addr,
57 HeapWord* limit) const {
58 // First we must round addr *up* to a possible object boundary.
59 addr = (HeapWord*)align_size_up((intptr_t)addr,
60 HeapWordSize << _shifter);
61 size_t addrOffset = heapWordToOffset(addr);
62 if (limit == NULL) {
63 limit = _bmStartWord + _bmWordSize;
64 }
65 size_t limitOffset = heapWordToOffset(limit);
66 size_t nextOffset = _bm.get_next_one_offset(addrOffset, limitOffset);
67 HeapWord* nextAddr = offsetToHeapWord(nextOffset);
68 assert(nextAddr >= addr, "get_next_one postcondition");
69 assert(nextAddr == limit || isMarked(nextAddr),
70 "get_next_one postcondition");
71 return nextAddr;
72 }
73
74 HeapWord* CMBitMapRO::getNextUnmarkedWordAddress(HeapWord* addr,
75 HeapWord* limit) const {
76 size_t addrOffset = heapWordToOffset(addr);
77 if (limit == NULL) {
78 limit = _bmStartWord + _bmWordSize;
79 }
80 size_t limitOffset = heapWordToOffset(limit);
81 size_t nextOffset = _bm.get_next_zero_offset(addrOffset, limitOffset);
82 HeapWord* nextAddr = offsetToHeapWord(nextOffset);
83 assert(nextAddr >= addr, "get_next_one postcondition");
84 assert(nextAddr == limit || !isMarked(nextAddr),
85 "get_next_one postcondition");
86 return nextAddr;
87 }
88
89 int CMBitMapRO::heapWordDiffToOffsetDiff(size_t diff) const {
90 assert((diff & ((1 << _shifter) - 1)) == 0, "argument check");
91 return (int) (diff >> _shifter);
92 }
93
94 #ifndef PRODUCT
95 bool CMBitMapRO::covers(ReservedSpace heap_rs) const {
96 // assert(_bm.map() == _virtual_space.low(), "map inconsistency");
97 assert(((size_t)_bm.size() * ((size_t)1 << _shifter)) == _bmWordSize,
98 "size inconsistency");
99 return _bmStartWord == (HeapWord*)(heap_rs.base()) &&
100 _bmWordSize == heap_rs.size()>>LogHeapWordSize;
101 }
102 #endif
103
104 bool CMBitMap::allocate(ReservedSpace heap_rs) {
105 _bmStartWord = (HeapWord*)(heap_rs.base());
106 _bmWordSize = heap_rs.size()/HeapWordSize; // heap_rs.size() is in bytes
107 ReservedSpace brs(ReservedSpace::allocation_align_size_up(
108 (_bmWordSize >> (_shifter + LogBitsPerByte)) + 1));
109 if (!brs.is_reserved()) {
110 warning("ConcurrentMark marking bit map allocation failure");
111 return false;
112 }
113 MemTracker::record_virtual_memory_type((address)brs.base(), mtGC);
114 // For now we'll just commit all of the bit map up front.
115 // Later on we'll try to be more parsimonious with swap.
116 if (!_virtual_space.initialize(brs, brs.size())) {
117 warning("ConcurrentMark marking bit map backing store failure");
118 return false;
119 }
120 assert(_virtual_space.committed_size() == brs.size(),
121 "didn't reserve backing store for all of concurrent marking bit map?");
122 _bm.set_map((uintptr_t*)_virtual_space.low());
123 assert(_virtual_space.committed_size() << (_shifter + LogBitsPerByte) >=
124 _bmWordSize, "inconsistency in bit map sizing");
125 _bm.set_size(_bmWordSize >> _shifter);
126 return true;
127 }
128
129 void CMBitMap::clearAll() {
130 _bm.clear();
131 return;
132 }
133
134 void CMBitMap::markRange(MemRegion mr) {
135 mr.intersection(MemRegion(_bmStartWord, _bmWordSize));
136 assert(!mr.is_empty(), "unexpected empty region");
137 assert((offsetToHeapWord(heapWordToOffset(mr.end())) ==
138 ((HeapWord *) mr.end())),
139 "markRange memory region end is not card aligned");
140 // convert address range into offset range
141 _bm.at_put_range(heapWordToOffset(mr.start()),
142 heapWordToOffset(mr.end()), true);
143 }
144
145 void CMBitMap::clearRange(MemRegion mr) {
146 mr.intersection(MemRegion(_bmStartWord, _bmWordSize));
147 assert(!mr.is_empty(), "unexpected empty region");
148 // convert address range into offset range
149 _bm.at_put_range(heapWordToOffset(mr.start()),
150 heapWordToOffset(mr.end()), false);
151 }
152
153 MemRegion CMBitMap::getAndClearMarkedRegion(HeapWord* addr,
154 HeapWord* end_addr) {
155 HeapWord* start = getNextMarkedWordAddress(addr);
156 start = MIN2(start, end_addr);
157 HeapWord* end = getNextUnmarkedWordAddress(start);
158 end = MIN2(end, end_addr);
159 assert(start <= end, "Consistency check");
160 MemRegion mr(start, end);
161 if (!mr.is_empty()) {
162 clearRange(mr);
163 }
164 return mr;
165 }
166
167 CMMarkStack::CMMarkStack(ConcurrentMark* cm) :
168 _base(NULL), _cm(cm)
169 #ifdef ASSERT
170 , _drain_in_progress(false)
171 , _drain_in_progress_yields(false)
172 #endif
173 {}
174
175 bool CMMarkStack::allocate(size_t capacity) {
176 // allocate a stack of the requisite depth
177 ReservedSpace rs(ReservedSpace::allocation_align_size_up(capacity * sizeof(oop)));
178 if (!rs.is_reserved()) {
179 warning("ConcurrentMark MarkStack allocation failure");
180 return false;
181 }
182 MemTracker::record_virtual_memory_type((address)rs.base(), mtGC);
183 if (!_virtual_space.initialize(rs, rs.size())) {
184 warning("ConcurrentMark MarkStack backing store failure");
185 // Release the virtual memory reserved for the marking stack
186 rs.release();
187 return false;
188 }
189 assert(_virtual_space.committed_size() == rs.size(),
190 "Didn't reserve backing store for all of ConcurrentMark stack?");
191 _base = (oop*) _virtual_space.low();
192 setEmpty();
193 _capacity = (jint) capacity;
194 _saved_index = -1;
195 _should_expand = false;
196 NOT_PRODUCT(_max_depth = 0);
197 return true;
198 }
199
200 void CMMarkStack::expand() {
201 // Called, during remark, if we've overflown the marking stack during marking.
202 assert(isEmpty(), "stack should been emptied while handling overflow");
203 assert(_capacity <= (jint) MarkStackSizeMax, "stack bigger than permitted");
204 // Clear expansion flag
205 _should_expand = false;
206 if (_capacity == (jint) MarkStackSizeMax) {
207 if (PrintGCDetails && Verbose) {
208 gclog_or_tty->print_cr(" (benign) Can't expand marking stack capacity, at max size limit");
209 }
210 return;
211 }
212 // Double capacity if possible
213 jint new_capacity = MIN2(_capacity*2, (jint) MarkStackSizeMax);
214 // Do not give up existing stack until we have managed to
215 // get the double capacity that we desired.
216 ReservedSpace rs(ReservedSpace::allocation_align_size_up(new_capacity *
217 sizeof(oop)));
218 if (rs.is_reserved()) {
219 // Release the backing store associated with old stack
220 _virtual_space.release();
221 // Reinitialize virtual space for new stack
222 if (!_virtual_space.initialize(rs, rs.size())) {
223 fatal("Not enough swap for expanded marking stack capacity");
224 }
225 _base = (oop*)(_virtual_space.low());
226 _index = 0;
227 _capacity = new_capacity;
228 } else {
229 if (PrintGCDetails && Verbose) {
230 // Failed to double capacity, continue;
231 gclog_or_tty->print(" (benign) Failed to expand marking stack capacity from "
232 SIZE_FORMAT"K to " SIZE_FORMAT"K",
233 _capacity / K, new_capacity / K);
234 }
235 }
236 }
237
238 void CMMarkStack::set_should_expand() {
239 // If we're resetting the marking state because of an
240 // marking stack overflow, record that we should, if
241 // possible, expand the stack.
242 _should_expand = _cm->has_overflown();
243 }
244
245 CMMarkStack::~CMMarkStack() {
246 if (_base != NULL) {
247 _base = NULL;
248 _virtual_space.release();
249 }
250 }
251
252 void CMMarkStack::par_push(oop ptr) {
253 while (true) {
254 if (isFull()) {
255 _overflow = true;
256 return;
257 }
258 // Otherwise...
259 jint index = _index;
260 jint next_index = index+1;
261 jint res = Atomic::cmpxchg(next_index, &_index, index);
262 if (res == index) {
263 _base[index] = ptr;
264 // Note that we don't maintain this atomically. We could, but it
265 // doesn't seem necessary.
266 NOT_PRODUCT(_max_depth = MAX2(_max_depth, next_index));
267 return;
268 }
269 // Otherwise, we need to try again.
270 }
271 }
272
273 void CMMarkStack::par_adjoin_arr(oop* ptr_arr, int n) {
274 while (true) {
275 if (isFull()) {
276 _overflow = true;
277 return;
278 }
279 // Otherwise...
280 jint index = _index;
281 jint next_index = index + n;
282 if (next_index > _capacity) {
283 _overflow = true;
284 return;
285 }
286 jint res = Atomic::cmpxchg(next_index, &_index, index);
287 if (res == index) {
288 for (int i = 0; i < n; i++) {
289 int ind = index + i;
290 assert(ind < _capacity, "By overflow test above.");
291 _base[ind] = ptr_arr[i];
292 }
293 NOT_PRODUCT(_max_depth = MAX2(_max_depth, next_index));
294 return;
295 }
296 // Otherwise, we need to try again.
297 }
298 }
299
300 void CMMarkStack::par_push_arr(oop* ptr_arr, int n) {
301 MutexLockerEx x(ParGCRareEvent_lock, Mutex::_no_safepoint_check_flag);
302 jint start = _index;
303 jint next_index = start + n;
304 if (next_index > _capacity) {
305 _overflow = true;
306 return;
307 }
308 // Otherwise.
309 _index = next_index;
310 for (int i = 0; i < n; i++) {
311 int ind = start + i;
312 assert(ind < _capacity, "By overflow test above.");
313 _base[ind] = ptr_arr[i];
314 }
315 NOT_PRODUCT(_max_depth = MAX2(_max_depth, next_index));
316 }
317
318 bool CMMarkStack::par_pop_arr(oop* ptr_arr, int max, int* n) {
319 MutexLockerEx x(ParGCRareEvent_lock, Mutex::_no_safepoint_check_flag);
320 jint index = _index;
321 if (index == 0) {
322 *n = 0;
323 return false;
324 } else {
325 int k = MIN2(max, index);
326 jint new_ind = index - k;
327 for (int j = 0; j < k; j++) {
328 ptr_arr[j] = _base[new_ind + j];
329 }
330 _index = new_ind;
331 *n = k;
332 return true;
333 }
334 }
335
336 template<class OopClosureClass>
337 bool CMMarkStack::drain(OopClosureClass* cl, CMBitMap* bm, bool yield_after) {
338 assert(!_drain_in_progress || !_drain_in_progress_yields || yield_after
339 || SafepointSynchronize::is_at_safepoint(),
340 "Drain recursion must be yield-safe.");
341 bool res = true;
342 debug_only(_drain_in_progress = true);
343 debug_only(_drain_in_progress_yields = yield_after);
344 while (!isEmpty()) {
345 oop newOop = pop();
346 assert(G1CollectedHeap::heap()->is_in_reserved(newOop), "Bad pop");
347 assert(newOop->is_oop(), "Expected an oop");
348 assert(bm == NULL || bm->isMarked((HeapWord*)newOop),
349 "only grey objects on this stack");
350 newOop->oop_iterate(cl);
351 if (yield_after && _cm->do_yield_check()) {
352 res = false;
353 break;
354 }
355 }
356 debug_only(_drain_in_progress = false);
357 return res;
358 }
359
360 void CMMarkStack::note_start_of_gc() {
361 assert(_saved_index == -1,
362 "note_start_of_gc()/end_of_gc() bracketed incorrectly");
363 _saved_index = _index;
364 }
365
366 void CMMarkStack::note_end_of_gc() {
367 // This is intentionally a guarantee, instead of an assert. If we
368 // accidentally add something to the mark stack during GC, it
369 // will be a correctness issue so it's better if we crash. we'll
370 // only check this once per GC anyway, so it won't be a performance
371 // issue in any way.
372 guarantee(_saved_index == _index,
373 err_msg("saved index: %d index: %d", _saved_index, _index));
374 _saved_index = -1;
375 }
376
377 void CMMarkStack::oops_do(OopClosure* f) {
378 assert(_saved_index == _index,
379 err_msg("saved index: %d index: %d", _saved_index, _index));
380 for (int i = 0; i < _index; i += 1) {
381 f->do_oop(&_base[i]);
382 }
383 }
384
385 bool ConcurrentMark::not_yet_marked(oop obj) const {
386 return _g1h->is_obj_ill(obj);
387 }
388
389 CMRootRegions::CMRootRegions() :
390 _young_list(NULL), _cm(NULL), _scan_in_progress(false),
391 _should_abort(false), _next_survivor(NULL) { }
392
393 void CMRootRegions::init(G1CollectedHeap* g1h, ConcurrentMark* cm) {
394 _young_list = g1h->young_list();
395 _cm = cm;
396 }
397
398 void CMRootRegions::prepare_for_scan() {
399 assert(!scan_in_progress(), "pre-condition");
400
401 // Currently, only survivors can be root regions.
402 assert(_next_survivor == NULL, "pre-condition");
403 _next_survivor = _young_list->first_survivor_region();
404 _scan_in_progress = (_next_survivor != NULL);
405 _should_abort = false;
406 }
407
408 HeapRegion* CMRootRegions::claim_next() {
409 if (_should_abort) {
410 // If someone has set the should_abort flag, we return NULL to
411 // force the caller to bail out of their loop.
412 return NULL;
413 }
414
415 // Currently, only survivors can be root regions.
416 HeapRegion* res = _next_survivor;
417 if (res != NULL) {
418 MutexLockerEx x(RootRegionScan_lock, Mutex::_no_safepoint_check_flag);
419 // Read it again in case it changed while we were waiting for the lock.
420 res = _next_survivor;
421 if (res != NULL) {
422 if (res == _young_list->last_survivor_region()) {
423 // We just claimed the last survivor so store NULL to indicate
424 // that we're done.
425 _next_survivor = NULL;
426 } else {
427 _next_survivor = res->get_next_young_region();
428 }
429 } else {
430 // Someone else claimed the last survivor while we were trying
431 // to take the lock so nothing else to do.
432 }
433 }
434 assert(res == NULL || res->is_survivor(), "post-condition");
435
436 return res;
437 }
438
439 void CMRootRegions::scan_finished() {
440 assert(scan_in_progress(), "pre-condition");
441
442 // Currently, only survivors can be root regions.
443 if (!_should_abort) {
444 assert(_next_survivor == NULL, "we should have claimed all survivors");
445 }
446 _next_survivor = NULL;
447
448 {
449 MutexLockerEx x(RootRegionScan_lock, Mutex::_no_safepoint_check_flag);
450 _scan_in_progress = false;
451 RootRegionScan_lock->notify_all();
452 }
453 }
454
455 bool CMRootRegions::wait_until_scan_finished() {
456 if (!scan_in_progress()) return false;
457
458 {
459 MutexLockerEx x(RootRegionScan_lock, Mutex::_no_safepoint_check_flag);
460 while (scan_in_progress()) {
461 RootRegionScan_lock->wait(Mutex::_no_safepoint_check_flag);
462 }
463 }
464 return true;
465 }
466
467 #ifdef _MSC_VER // the use of 'this' below gets a warning, make it go away
468 #pragma warning( disable:4355 ) // 'this' : used in base member initializer list
469 #endif // _MSC_VER
470
471 uint ConcurrentMark::scale_parallel_threads(uint n_par_threads) {
472 return MAX2((n_par_threads + 2) / 4, 1U);
473 }
474
475 ConcurrentMark::ConcurrentMark(G1CollectedHeap* g1h, ReservedSpace heap_rs) :
476 _g1h(g1h),
477 _markBitMap1(MinObjAlignment - 1),
478 _markBitMap2(MinObjAlignment - 1),
479
480 _parallel_marking_threads(0),
481 _max_parallel_marking_threads(0),
482 _sleep_factor(0.0),
483 _marking_task_overhead(1.0),
484 _cleanup_sleep_factor(0.0),
485 _cleanup_task_overhead(1.0),
486 _cleanup_list("Cleanup List"),
487 _region_bm((BitMap::idx_t)(g1h->max_regions()), false /* in_resource_area*/),
488 _card_bm((heap_rs.size() + CardTableModRefBS::card_size - 1) >>
489 CardTableModRefBS::card_shift,
490 false /* in_resource_area*/),
491
492 _prevMarkBitMap(&_markBitMap1),
493 _nextMarkBitMap(&_markBitMap2),
494
495 _markStack(this),
496 // _finger set in set_non_marking_state
497
498 _max_worker_id(MAX2((uint)ParallelGCThreads, 1U)),
499 // _active_tasks set in set_non_marking_state
500 // _tasks set inside the constructor
501 _task_queues(new CMTaskQueueSet((int) _max_worker_id)),
502 _terminator(ParallelTaskTerminator((int) _max_worker_id, _task_queues)),
503
504 _has_overflown(false),
505 _concurrent(false),
506 _has_aborted(false),
507 _restart_for_overflow(false),
508 _concurrent_marking_in_progress(false),
509
510 // _verbose_level set below
511
512 _init_times(),
513 _remark_times(), _remark_mark_times(), _remark_weak_ref_times(),
514 _cleanup_times(),
515 _total_counting_time(0.0),
516 _total_rs_scrub_time(0.0),
517
518 _parallel_workers(NULL),
519
520 _count_card_bitmaps(NULL),
521 _count_marked_bytes(NULL),
522 _completed_initialization(false) {
523 CMVerboseLevel verbose_level = (CMVerboseLevel) G1MarkingVerboseLevel;
524 if (verbose_level < no_verbose) {
525 verbose_level = no_verbose;
526 }
527 if (verbose_level > high_verbose) {
528 verbose_level = high_verbose;
529 }
530 _verbose_level = verbose_level;
531
532 if (verbose_low()) {
533 gclog_or_tty->print_cr("[global] init, heap start = "PTR_FORMAT", "
534 "heap end = "PTR_FORMAT, _heap_start, _heap_end);
535 }
536
537 if (!_markBitMap1.allocate(heap_rs)) {
538 warning("Failed to allocate first CM bit map");
539 return;
540 }
541 if (!_markBitMap2.allocate(heap_rs)) {
542 warning("Failed to allocate second CM bit map");
543 return;
544 }
545
546 // Create & start a ConcurrentMark thread.
547 _cmThread = new ConcurrentMarkThread(this);
548 assert(cmThread() != NULL, "CM Thread should have been created");
549 assert(cmThread()->cm() != NULL, "CM Thread should refer to this cm");
550
551 assert(CGC_lock != NULL, "Where's the CGC_lock?");
552 assert(_markBitMap1.covers(heap_rs), "_markBitMap1 inconsistency");
553 assert(_markBitMap2.covers(heap_rs), "_markBitMap2 inconsistency");
554
555 SATBMarkQueueSet& satb_qs = JavaThread::satb_mark_queue_set();
556 satb_qs.set_buffer_size(G1SATBBufferSize);
557
558 _root_regions.init(_g1h, this);
559
560 if (ConcGCThreads > ParallelGCThreads) {
561 warning("Can't have more ConcGCThreads (" UINT32_FORMAT ") "
562 "than ParallelGCThreads (" UINT32_FORMAT ").",
563 ConcGCThreads, ParallelGCThreads);
564 return;
565 }
566 if (ParallelGCThreads == 0) {
567 // if we are not running with any parallel GC threads we will not
568 // spawn any marking threads either
569 _parallel_marking_threads = 0;
570 _max_parallel_marking_threads = 0;
571 _sleep_factor = 0.0;
572 _marking_task_overhead = 1.0;
573 } else {
574 if (!FLAG_IS_DEFAULT(ConcGCThreads) && ConcGCThreads > 0) {
575 // Note: ConcGCThreads has precedence over G1MarkingOverheadPercent
576 // if both are set
577 _sleep_factor = 0.0;
578 _marking_task_overhead = 1.0;
579 } else if (G1MarkingOverheadPercent > 0) {
580 // We will calculate the number of parallel marking threads based
581 // on a target overhead with respect to the soft real-time goal
582 double marking_overhead = (double) G1MarkingOverheadPercent / 100.0;
583 double overall_cm_overhead =
584 (double) MaxGCPauseMillis * marking_overhead /
585 (double) GCPauseIntervalMillis;
586 double cpu_ratio = 1.0 / (double) os::processor_count();
587 double marking_thread_num = ceil(overall_cm_overhead / cpu_ratio);
588 double marking_task_overhead =
589 overall_cm_overhead / marking_thread_num *
590 (double) os::processor_count();
591 double sleep_factor =
592 (1.0 - marking_task_overhead) / marking_task_overhead;
593
594 FLAG_SET_ERGO(uintx, ConcGCThreads, (uint) marking_thread_num);
595 _sleep_factor = sleep_factor;
596 _marking_task_overhead = marking_task_overhead;
597 } else {
598 // Calculate the number of parallel marking threads by scaling
599 // the number of parallel GC threads.
600 uint marking_thread_num = scale_parallel_threads((uint) ParallelGCThreads);
601 FLAG_SET_ERGO(uintx, ConcGCThreads, marking_thread_num);
602 _sleep_factor = 0.0;
603 _marking_task_overhead = 1.0;
604 }
605
606 assert(ConcGCThreads > 0, "Should have been set");
607 _parallel_marking_threads = (uint) ConcGCThreads;
608 _max_parallel_marking_threads = _parallel_marking_threads;
609
610 if (parallel_marking_threads() > 1) {
611 _cleanup_task_overhead = 1.0;
612 } else {
613 _cleanup_task_overhead = marking_task_overhead();
614 }
615 _cleanup_sleep_factor =
616 (1.0 - cleanup_task_overhead()) / cleanup_task_overhead();
617
618 #if 0
619 gclog_or_tty->print_cr("Marking Threads %d", parallel_marking_threads());
620 gclog_or_tty->print_cr("CM Marking Task Overhead %1.4lf", marking_task_overhead());
621 gclog_or_tty->print_cr("CM Sleep Factor %1.4lf", sleep_factor());
622 gclog_or_tty->print_cr("CL Marking Task Overhead %1.4lf", cleanup_task_overhead());
623 gclog_or_tty->print_cr("CL Sleep Factor %1.4lf", cleanup_sleep_factor());
624 #endif
625
626 guarantee(parallel_marking_threads() > 0, "peace of mind");
627 _parallel_workers = new FlexibleWorkGang("G1 Parallel Marking Threads",
628 _max_parallel_marking_threads, false, true);
629 if (_parallel_workers == NULL) {
630 vm_exit_during_initialization("Failed necessary allocation.");
631 } else {
632 _parallel_workers->initialize_workers();
633 }
634 }
635
636 if (FLAG_IS_DEFAULT(MarkStackSize)) {
637 uintx mark_stack_size =
638 MIN2(MarkStackSizeMax,
639 MAX2(MarkStackSize, (uintx) (parallel_marking_threads() * TASKQUEUE_SIZE)));
640 // Verify that the calculated value for MarkStackSize is in range.
641 // It would be nice to use the private utility routine from Arguments.
642 if (!(mark_stack_size >= 1 && mark_stack_size <= MarkStackSizeMax)) {
643 warning("Invalid value calculated for MarkStackSize (" UINTX_FORMAT "): "
644 "must be between " UINTX_FORMAT " and " UINTX_FORMAT,
645 mark_stack_size, 1, MarkStackSizeMax);
646 return;
647 }
648 FLAG_SET_ERGO(uintx, MarkStackSize, mark_stack_size);
649 } else {
650 // Verify MarkStackSize is in range.
651 if (FLAG_IS_CMDLINE(MarkStackSize)) {
652 if (FLAG_IS_DEFAULT(MarkStackSizeMax)) {
653 if (!(MarkStackSize >= 1 && MarkStackSize <= MarkStackSizeMax)) {
654 warning("Invalid value specified for MarkStackSize (" UINTX_FORMAT "): "
655 "must be between " UINTX_FORMAT " and " UINTX_FORMAT,
656 MarkStackSize, 1, MarkStackSizeMax);
657 return;
658 }
659 } else if (FLAG_IS_CMDLINE(MarkStackSizeMax)) {
660 if (!(MarkStackSize >= 1 && MarkStackSize <= MarkStackSizeMax)) {
661 warning("Invalid value specified for MarkStackSize (" UINTX_FORMAT ")"
662 " or for MarkStackSizeMax (" UINTX_FORMAT ")",
663 MarkStackSize, MarkStackSizeMax);
664 return;
665 }
666 }
667 }
668 }
669
670 if (!_markStack.allocate(MarkStackSize)) {
671 warning("Failed to allocate CM marking stack");
672 return;
673 }
674
675 _tasks = NEW_C_HEAP_ARRAY(CMTask*, _max_worker_id, mtGC);
676 _accum_task_vtime = NEW_C_HEAP_ARRAY(double, _max_worker_id, mtGC);
677
678 _count_card_bitmaps = NEW_C_HEAP_ARRAY(BitMap, _max_worker_id, mtGC);
679 _count_marked_bytes = NEW_C_HEAP_ARRAY(size_t*, _max_worker_id, mtGC);
680
681 BitMap::idx_t card_bm_size = _card_bm.size();
682
683 // so that the assertion in MarkingTaskQueue::task_queue doesn't fail
684 _active_tasks = _max_worker_id;
685
686 size_t max_regions = (size_t) _g1h->max_regions();
687 for (uint i = 0; i < _max_worker_id; ++i) {
688 CMTaskQueue* task_queue = new CMTaskQueue();
689 task_queue->initialize();
690 _task_queues->register_queue(i, task_queue);
691
692 _count_card_bitmaps[i] = BitMap(card_bm_size, false);
693 _count_marked_bytes[i] = NEW_C_HEAP_ARRAY(size_t, max_regions, mtGC);
694
695 _tasks[i] = new CMTask(i, this,
696 _count_marked_bytes[i],
697 &_count_card_bitmaps[i],
698 task_queue, _task_queues);
699
700 _accum_task_vtime[i] = 0.0;
701 }
702
703 // Calculate the card number for the bottom of the heap. Used
704 // in biasing indexes into the accounting card bitmaps.
705 _heap_bottom_card_num =
706 intptr_t(uintptr_t(_g1h->reserved_region().start()) >>
707 CardTableModRefBS::card_shift);
708
709 // Clear all the liveness counting data
710 clear_all_count_data();
711
712 // so that the call below can read a sensible value
713 _heap_start = (HeapWord*) heap_rs.base();
714 set_non_marking_state();
715 _completed_initialization = true;
716 }
717
718 void ConcurrentMark::update_g1_committed(bool force) {
719 // If concurrent marking is not in progress, then we do not need to
720 // update _heap_end.
721 if (!concurrent_marking_in_progress() && !force) return;
722
723 MemRegion committed = _g1h->g1_committed();
724 assert(committed.start() == _heap_start, "start shouldn't change");
725 HeapWord* new_end = committed.end();
726 if (new_end > _heap_end) {
727 // The heap has been expanded.
728
729 _heap_end = new_end;
730 }
731 // Notice that the heap can also shrink. However, this only happens
732 // during a Full GC (at least currently) and the entire marking
733 // phase will bail out and the task will not be restarted. So, let's
734 // do nothing.
735 }
736
737 void ConcurrentMark::reset() {
738 // Starting values for these two. This should be called in a STW
739 // phase. CM will be notified of any future g1_committed expansions
740 // will be at the end of evacuation pauses, when tasks are
741 // inactive.
742 MemRegion committed = _g1h->g1_committed();
743 _heap_start = committed.start();
744 _heap_end = committed.end();
745
746 // Separated the asserts so that we know which one fires.
747 assert(_heap_start != NULL, "heap bounds should look ok");
748 assert(_heap_end != NULL, "heap bounds should look ok");
749 assert(_heap_start < _heap_end, "heap bounds should look ok");
750
751 // Reset all the marking data structures and any necessary flags
752 reset_marking_state();
753
754 if (verbose_low()) {
755 gclog_or_tty->print_cr("[global] resetting");
756 }
757
758 // We do reset all of them, since different phases will use
759 // different number of active threads. So, it's easiest to have all
760 // of them ready.
761 for (uint i = 0; i < _max_worker_id; ++i) {
762 _tasks[i]->reset(_nextMarkBitMap);
763 }
764
765 // we need this to make sure that the flag is on during the evac
766 // pause with initial mark piggy-backed
767 set_concurrent_marking_in_progress();
768 }
769
770
771 void ConcurrentMark::reset_marking_state(bool clear_overflow) {
772 _markStack.set_should_expand();
773 _markStack.setEmpty(); // Also clears the _markStack overflow flag
774 if (clear_overflow) {
775 clear_has_overflown();
776 } else {
777 assert(has_overflown(), "pre-condition");
778 }
779 _finger = _heap_start;
780
781 for (uint i = 0; i < _max_worker_id; ++i) {
782 CMTaskQueue* queue = _task_queues->queue(i);
783 queue->set_empty();
784 }
785 }
786
787 void ConcurrentMark::set_phase(uint active_tasks, bool concurrent) {
788 assert(active_tasks <= _max_worker_id, "we should not have more");
789
790 _active_tasks = active_tasks;
791 // Need to update the three data structures below according to the
792 // number of active threads for this phase.
793 _terminator = ParallelTaskTerminator((int) active_tasks, _task_queues);
794 _first_overflow_barrier_sync.set_n_workers((int) active_tasks);
795 _second_overflow_barrier_sync.set_n_workers((int) active_tasks);
796
797 _concurrent = concurrent;
798 // We propagate this to all tasks, not just the active ones.
799 for (uint i = 0; i < _max_worker_id; ++i)
800 _tasks[i]->set_concurrent(concurrent);
801
802 if (concurrent) {
803 set_concurrent_marking_in_progress();
804 } else {
805 // We currently assume that the concurrent flag has been set to
806 // false before we start remark. At this point we should also be
807 // in a STW phase.
808 assert(!concurrent_marking_in_progress(), "invariant");
809 assert(_finger == _heap_end, "only way to get here");
810 update_g1_committed(true);
811 }
812 }
813
814 void ConcurrentMark::set_non_marking_state() {
815 // We set the global marking state to some default values when we're
816 // not doing marking.
817 reset_marking_state();
818 _active_tasks = 0;
819 clear_concurrent_marking_in_progress();
820 }
821
822 ConcurrentMark::~ConcurrentMark() {
823 // The ConcurrentMark instance is never freed.
824 ShouldNotReachHere();
825 }
826
827 void ConcurrentMark::clearNextBitmap() {
828 G1CollectedHeap* g1h = G1CollectedHeap::heap();
829 G1CollectorPolicy* g1p = g1h->g1_policy();
830
831 // Make sure that the concurrent mark thread looks to still be in
832 // the current cycle.
833 guarantee(cmThread()->during_cycle(), "invariant");
834
835 // We are finishing up the current cycle by clearing the next
836 // marking bitmap and getting it ready for the next cycle. During
837 // this time no other cycle can start. So, let's make sure that this
838 // is the case.
839 guarantee(!g1h->mark_in_progress(), "invariant");
840
841 // clear the mark bitmap (no grey objects to start with).
842 // We need to do this in chunks and offer to yield in between
843 // each chunk.
844 HeapWord* start = _nextMarkBitMap->startWord();
845 HeapWord* end = _nextMarkBitMap->endWord();
846 HeapWord* cur = start;
847 size_t chunkSize = M;
848 while (cur < end) {
849 HeapWord* next = cur + chunkSize;
850 if (next > end) {
851 next = end;
852 }
853 MemRegion mr(cur,next);
854 _nextMarkBitMap->clearRange(mr);
855 cur = next;
856 do_yield_check();
857
858 // Repeat the asserts from above. We'll do them as asserts here to
859 // minimize their overhead on the product. However, we'll have
860 // them as guarantees at the beginning / end of the bitmap
861 // clearing to get some checking in the product.
862 assert(cmThread()->during_cycle(), "invariant");
863 assert(!g1h->mark_in_progress(), "invariant");
864 }
865
866 // Clear the liveness counting data
867 clear_all_count_data();
868
869 // Repeat the asserts from above.
870 guarantee(cmThread()->during_cycle(), "invariant");
871 guarantee(!g1h->mark_in_progress(), "invariant");
872 }
873
874 class NoteStartOfMarkHRClosure: public HeapRegionClosure {
875 public:
876 bool doHeapRegion(HeapRegion* r) {
877 if (!r->continuesHumongous()) {
878 r->note_start_of_marking();
879 }
880 return false;
881 }
882 };
883
884 void ConcurrentMark::checkpointRootsInitialPre() {
885 G1CollectedHeap* g1h = G1CollectedHeap::heap();
886 G1CollectorPolicy* g1p = g1h->g1_policy();
887
888 _has_aborted = false;
889
890 #ifndef PRODUCT
891 if (G1PrintReachableAtInitialMark) {
892 print_reachable("at-cycle-start",
893 VerifyOption_G1UsePrevMarking, true /* all */);
894 }
895 #endif
896
897 // Initialise marking structures. This has to be done in a STW phase.
898 reset();
899
900 // For each region note start of marking.
901 NoteStartOfMarkHRClosure startcl;
902 g1h->heap_region_iterate(&startcl);
903 }
904
905
906 void ConcurrentMark::checkpointRootsInitialPost() {
907 G1CollectedHeap* g1h = G1CollectedHeap::heap();
908
909 // If we force an overflow during remark, the remark operation will
910 // actually abort and we'll restart concurrent marking. If we always
911 // force an oveflow during remark we'll never actually complete the
912 // marking phase. So, we initilize this here, at the start of the
913 // cycle, so that at the remaining overflow number will decrease at
914 // every remark and we'll eventually not need to cause one.
915 force_overflow_stw()->init();
916
917 // Start Concurrent Marking weak-reference discovery.
918 ReferenceProcessor* rp = g1h->ref_processor_cm();
919 // enable ("weak") refs discovery
920 rp->enable_discovery(true /*verify_disabled*/, true /*verify_no_refs*/);
921 rp->setup_policy(false); // snapshot the soft ref policy to be used in this cycle
922
923 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set();
924 // This is the start of the marking cycle, we're expected all
925 // threads to have SATB queues with active set to false.
926 satb_mq_set.set_active_all_threads(true, /* new active value */
927 false /* expected_active */);
928
929 _root_regions.prepare_for_scan();
930
931 // update_g1_committed() will be called at the end of an evac pause
932 // when marking is on. So, it's also called at the end of the
933 // initial-mark pause to update the heap end, if the heap expands
934 // during it. No need to call it here.
935 }
936
937 /*
938 * Notice that in the next two methods, we actually leave the STS
939 * during the barrier sync and join it immediately afterwards. If we
940 * do not do this, the following deadlock can occur: one thread could
941 * be in the barrier sync code, waiting for the other thread to also
942 * sync up, whereas another one could be trying to yield, while also
943 * waiting for the other threads to sync up too.
944 *
945 * Note, however, that this code is also used during remark and in
946 * this case we should not attempt to leave / enter the STS, otherwise
947 * we'll either hit an asseert (debug / fastdebug) or deadlock
948 * (product). So we should only leave / enter the STS if we are
949 * operating concurrently.
950 *
951 * Because the thread that does the sync barrier has left the STS, it
952 * is possible to be suspended for a Full GC or an evacuation pause
953 * could occur. This is actually safe, since the entering the sync
954 * barrier is one of the last things do_marking_step() does, and it
955 * doesn't manipulate any data structures afterwards.
956 */
957
958 void ConcurrentMark::enter_first_sync_barrier(uint worker_id) {
959 if (verbose_low()) {
960 gclog_or_tty->print_cr("[%u] entering first barrier", worker_id);
961 }
962
963 if (concurrent()) {
964 ConcurrentGCThread::stsLeave();
965 }
966 _first_overflow_barrier_sync.enter();
967 if (concurrent()) {
968 ConcurrentGCThread::stsJoin();
969 }
970 // at this point everyone should have synced up and not be doing any
971 // more work
972
973 if (verbose_low()) {
974 gclog_or_tty->print_cr("[%u] leaving first barrier", worker_id);
975 }
976
977 // let the task associated with with worker 0 do this
978 if (worker_id == 0) {
979 // task 0 is responsible for clearing the global data structures
980 // We should be here because of an overflow. During STW we should
981 // not clear the overflow flag since we rely on it being true when
982 // we exit this method to abort the pause and restart concurent
983 // marking.
984 reset_marking_state(concurrent() /* clear_overflow */);
985 force_overflow()->update();
986
987 if (G1Log::fine()) {
988 gclog_or_tty->date_stamp(PrintGCDateStamps);
989 gclog_or_tty->stamp(PrintGCTimeStamps);
990 gclog_or_tty->print_cr("[GC concurrent-mark-reset-for-overflow]");
991 }
992 }
993
994 // after this, each task should reset its own data structures then
995 // then go into the second barrier
996 }
997
998 void ConcurrentMark::enter_second_sync_barrier(uint worker_id) {
999 if (verbose_low()) {
1000 gclog_or_tty->print_cr("[%u] entering second barrier", worker_id);
1001 }
1002
1003 if (concurrent()) {
1004 ConcurrentGCThread::stsLeave();
1005 }
1006 _second_overflow_barrier_sync.enter();
1007 if (concurrent()) {
1008 ConcurrentGCThread::stsJoin();
1009 }
1010 // at this point everything should be re-initialised and ready to go
1011
1012 if (verbose_low()) {
1013 gclog_or_tty->print_cr("[%u] leaving second barrier", worker_id);
1014 }
1015 }
1016
1017 #ifndef PRODUCT
1018 void ForceOverflowSettings::init() {
1019 _num_remaining = G1ConcMarkForceOverflow;
1020 _force = false;
1021 update();
1022 }
1023
1024 void ForceOverflowSettings::update() {
1025 if (_num_remaining > 0) {
1026 _num_remaining -= 1;
1027 _force = true;
1028 } else {
1029 _force = false;
1030 }
1031 }
1032
1033 bool ForceOverflowSettings::should_force() {
1034 if (_force) {
1035 _force = false;
1036 return true;
1037 } else {
1038 return false;
1039 }
1040 }
1041 #endif // !PRODUCT
1042
1043 class CMConcurrentMarkingTask: public AbstractGangTask {
1044 private:
1045 ConcurrentMark* _cm;
1046 ConcurrentMarkThread* _cmt;
1047
1048 public:
1049 void work(uint worker_id) {
1050 assert(Thread::current()->is_ConcurrentGC_thread(),
1051 "this should only be done by a conc GC thread");
1052 ResourceMark rm;
1053
1054 double start_vtime = os::elapsedVTime();
1055
1056 ConcurrentGCThread::stsJoin();
1057
1058 assert(worker_id < _cm->active_tasks(), "invariant");
1059 CMTask* the_task = _cm->task(worker_id);
1060 the_task->record_start_time();
1061 if (!_cm->has_aborted()) {
1062 do {
1063 double start_vtime_sec = os::elapsedVTime();
1064 double start_time_sec = os::elapsedTime();
1065 double mark_step_duration_ms = G1ConcMarkStepDurationMillis;
1066
1067 the_task->do_marking_step(mark_step_duration_ms,
1068 true /* do_stealing */,
1069 true /* do_termination */,
1070 false /* is_serial*/);
1071
1072 double end_time_sec = os::elapsedTime();
1073 double end_vtime_sec = os::elapsedVTime();
1074 double elapsed_vtime_sec = end_vtime_sec - start_vtime_sec;
1075 double elapsed_time_sec = end_time_sec - start_time_sec;
1076 _cm->clear_has_overflown();
1077
1078 bool ret = _cm->do_yield_check(worker_id);
1079
1080 jlong sleep_time_ms;
1081 if (!_cm->has_aborted() && the_task->has_aborted()) {
1082 sleep_time_ms =
1083 (jlong) (elapsed_vtime_sec * _cm->sleep_factor() * 1000.0);
1084 ConcurrentGCThread::stsLeave();
1085 os::sleep(Thread::current(), sleep_time_ms, false);
1086 ConcurrentGCThread::stsJoin();
1087 }
1088 double end_time2_sec = os::elapsedTime();
1089 double elapsed_time2_sec = end_time2_sec - start_time_sec;
1090
1091 #if 0
1092 gclog_or_tty->print_cr("CM: elapsed %1.4lf ms, sleep %1.4lf ms, "
1093 "overhead %1.4lf",
1094 elapsed_vtime_sec * 1000.0, (double) sleep_time_ms,
1095 the_task->conc_overhead(os::elapsedTime()) * 8.0);
1096 gclog_or_tty->print_cr("elapsed time %1.4lf ms, time 2: %1.4lf ms",
1097 elapsed_time_sec * 1000.0, elapsed_time2_sec * 1000.0);
1098 #endif
1099 } while (!_cm->has_aborted() && the_task->has_aborted());
1100 }
1101 the_task->record_end_time();
1102 guarantee(!the_task->has_aborted() || _cm->has_aborted(), "invariant");
1103
1104 ConcurrentGCThread::stsLeave();
1105
1106 double end_vtime = os::elapsedVTime();
1107 _cm->update_accum_task_vtime(worker_id, end_vtime - start_vtime);
1108 }
1109
1110 CMConcurrentMarkingTask(ConcurrentMark* cm,
1111 ConcurrentMarkThread* cmt) :
1112 AbstractGangTask("Concurrent Mark"), _cm(cm), _cmt(cmt) { }
1113
1114 ~CMConcurrentMarkingTask() { }
1115 };
1116
1117 // Calculates the number of active workers for a concurrent
1118 // phase.
1119 uint ConcurrentMark::calc_parallel_marking_threads() {
1120 if (G1CollectedHeap::use_parallel_gc_threads()) {
1121 uint n_conc_workers = 0;
1122 if (!UseDynamicNumberOfGCThreads ||
1123 (!FLAG_IS_DEFAULT(ConcGCThreads) &&
1124 !ForceDynamicNumberOfGCThreads)) {
1125 n_conc_workers = max_parallel_marking_threads();
1126 } else {
1127 n_conc_workers =
1128 AdaptiveSizePolicy::calc_default_active_workers(
1129 max_parallel_marking_threads(),
1130 1, /* Minimum workers */
1131 parallel_marking_threads(),
1132 Threads::number_of_non_daemon_threads());
1133 // Don't scale down "n_conc_workers" by scale_parallel_threads() because
1134 // that scaling has already gone into "_max_parallel_marking_threads".
1135 }
1136 assert(n_conc_workers > 0, "Always need at least 1");
1137 return n_conc_workers;
1138 }
1139 // If we are not running with any parallel GC threads we will not
1140 // have spawned any marking threads either. Hence the number of
1141 // concurrent workers should be 0.
1142 return 0;
1143 }
1144
1145 void ConcurrentMark::scanRootRegion(HeapRegion* hr, uint worker_id) {
1146 // Currently, only survivors can be root regions.
1147 assert(hr->next_top_at_mark_start() == hr->bottom(), "invariant");
1148 G1RootRegionScanClosure cl(_g1h, this, worker_id);
1149
1150 const uintx interval = PrefetchScanIntervalInBytes;
1151 HeapWord* curr = hr->bottom();
1152 const HeapWord* end = hr->top();
1153 while (curr < end) {
1154 Prefetch::read(curr, interval);
1155 oop obj = oop(curr);
1156 int size = obj->oop_iterate(&cl);
1157 assert(size == obj->size(), "sanity");
1158 curr += size;
1159 }
1160 }
1161
1162 class CMRootRegionScanTask : public AbstractGangTask {
1163 private:
1164 ConcurrentMark* _cm;
1165
1166 public:
1167 CMRootRegionScanTask(ConcurrentMark* cm) :
1168 AbstractGangTask("Root Region Scan"), _cm(cm) { }
1169
1170 void work(uint worker_id) {
1171 assert(Thread::current()->is_ConcurrentGC_thread(),
1172 "this should only be done by a conc GC thread");
1173
1174 CMRootRegions* root_regions = _cm->root_regions();
1175 HeapRegion* hr = root_regions->claim_next();
1176 while (hr != NULL) {
1177 _cm->scanRootRegion(hr, worker_id);
1178 hr = root_regions->claim_next();
1179 }
1180 }
1181 };
1182
1183 void ConcurrentMark::scanRootRegions() {
1184 // scan_in_progress() will have been set to true only if there was
1185 // at least one root region to scan. So, if it's false, we
1186 // should not attempt to do any further work.
1187 if (root_regions()->scan_in_progress()) {
1188 _parallel_marking_threads = calc_parallel_marking_threads();
1189 assert(parallel_marking_threads() <= max_parallel_marking_threads(),
1190 "Maximum number of marking threads exceeded");
1191 uint active_workers = MAX2(1U, parallel_marking_threads());
1192
1193 CMRootRegionScanTask task(this);
1194 if (use_parallel_marking_threads()) {
1195 _parallel_workers->set_active_workers((int) active_workers);
1196 _parallel_workers->run_task(&task);
1197 } else {
1198 task.work(0);
1199 }
1200
1201 // It's possible that has_aborted() is true here without actually
1202 // aborting the survivor scan earlier. This is OK as it's
1203 // mainly used for sanity checking.
1204 root_regions()->scan_finished();
1205 }
1206 }
1207
1208 void ConcurrentMark::markFromRoots() {
1209 // we might be tempted to assert that:
1210 // assert(asynch == !SafepointSynchronize::is_at_safepoint(),
1211 // "inconsistent argument?");
1212 // However that wouldn't be right, because it's possible that
1213 // a safepoint is indeed in progress as a younger generation
1214 // stop-the-world GC happens even as we mark in this generation.
1215
1216 _restart_for_overflow = false;
1217 force_overflow_conc()->init();
1218
1219 // _g1h has _n_par_threads
1220 _parallel_marking_threads = calc_parallel_marking_threads();
1221 assert(parallel_marking_threads() <= max_parallel_marking_threads(),
1222 "Maximum number of marking threads exceeded");
1223
1224 uint active_workers = MAX2(1U, parallel_marking_threads());
1225
1226 // Parallel task terminator is set in "set_phase()"
1227 set_phase(active_workers, true /* concurrent */);
1228
1229 CMConcurrentMarkingTask markingTask(this, cmThread());
1230 if (use_parallel_marking_threads()) {
1231 _parallel_workers->set_active_workers((int)active_workers);
1232 // Don't set _n_par_threads because it affects MT in proceess_strong_roots()
1233 // and the decisions on that MT processing is made elsewhere.
1234 assert(_parallel_workers->active_workers() > 0, "Should have been set");
1235 _parallel_workers->run_task(&markingTask);
1236 } else {
1237 markingTask.work(0);
1238 }
1239 print_stats();
1240 }
1241
1242 void ConcurrentMark::checkpointRootsFinal(bool clear_all_soft_refs) {
1243 // world is stopped at this checkpoint
1244 assert(SafepointSynchronize::is_at_safepoint(),
1245 "world should be stopped");
1246
1247 G1CollectedHeap* g1h = G1CollectedHeap::heap();
1248
1249 // If a full collection has happened, we shouldn't do this.
1250 if (has_aborted()) {
1251 g1h->set_marking_complete(); // So bitmap clearing isn't confused
1252 return;
1253 }
1254
1255 SvcGCMarker sgcm(SvcGCMarker::OTHER);
1256
1257 if (VerifyDuringGC) {
1258 HandleMark hm; // handle scope
1259 gclog_or_tty->print(" VerifyDuringGC:(before)");
1260 Universe::heap()->prepare_for_verify();
1261 Universe::verify(/* silent */ false,
1262 /* option */ VerifyOption_G1UsePrevMarking);
1263 }
1264
1265 G1CollectorPolicy* g1p = g1h->g1_policy();
1266 g1p->record_concurrent_mark_remark_start();
1267
1268 double start = os::elapsedTime();
1269
1270 checkpointRootsFinalWork();
1271
1272 double mark_work_end = os::elapsedTime();
1273
1274 weakRefsWork(clear_all_soft_refs);
1275
1276 if (has_overflown()) {
1277 // Oops. We overflowed. Restart concurrent marking.
1278 _restart_for_overflow = true;
1279 // Clear the marking state because we will be restarting
1280 // marking due to overflowing the global mark stack.
1281 reset_marking_state();
1282 if (G1TraceMarkStackOverflow) {
1283 gclog_or_tty->print_cr("\nRemark led to restart for overflow.");
1284 }
1285 } else {
1286 // Aggregate the per-task counting data that we have accumulated
1287 // while marking.
1288 aggregate_count_data();
1289
1290 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set();
1291 // We're done with marking.
1292 // This is the end of the marking cycle, we're expected all
1293 // threads to have SATB queues with active set to true.
1294 satb_mq_set.set_active_all_threads(false, /* new active value */
1295 true /* expected_active */);
1296
1297 if (VerifyDuringGC) {
1298 HandleMark hm; // handle scope
1299 gclog_or_tty->print(" VerifyDuringGC:(after)");
1300 Universe::heap()->prepare_for_verify();
1301 Universe::verify(/* silent */ false,
1302 /* option */ VerifyOption_G1UseNextMarking);
1303 }
1304 assert(!restart_for_overflow(), "sanity");
1305 // Completely reset the marking state since marking completed
1306 set_non_marking_state();
1307 }
1308
1309 // Expand the marking stack, if we have to and if we can.
1310 if (_markStack.should_expand()) {
1311 _markStack.expand();
1312 }
1313
1314 #if VERIFY_OBJS_PROCESSED
1315 _scan_obj_cl.objs_processed = 0;
1316 ThreadLocalObjQueue::objs_enqueued = 0;
1317 #endif
1318
1319 // Statistics
1320 double now = os::elapsedTime();
1321 _remark_mark_times.add((mark_work_end - start) * 1000.0);
1322 _remark_weak_ref_times.add((now - mark_work_end) * 1000.0);
1323 _remark_times.add((now - start) * 1000.0);
1324
1325 g1p->record_concurrent_mark_remark_end();
1326 }
1327
1328 // Base class of the closures that finalize and verify the
1329 // liveness counting data.
1330 class CMCountDataClosureBase: public HeapRegionClosure {
1331 protected:
1332 G1CollectedHeap* _g1h;
1333 ConcurrentMark* _cm;
1334 CardTableModRefBS* _ct_bs;
1335
1336 BitMap* _region_bm;
1337 BitMap* _card_bm;
1338
1339 // Takes a region that's not empty (i.e., it has at least one
1340 // live object in it and sets its corresponding bit on the region
1341 // bitmap to 1. If the region is "starts humongous" it will also set
1342 // to 1 the bits on the region bitmap that correspond to its
1343 // associated "continues humongous" regions.
1344 void set_bit_for_region(HeapRegion* hr) {
1345 assert(!hr->continuesHumongous(), "should have filtered those out");
1346
1347 BitMap::idx_t index = (BitMap::idx_t) hr->hrs_index();
1348 if (!hr->startsHumongous()) {
1349 // Normal (non-humongous) case: just set the bit.
1350 _region_bm->par_at_put(index, true);
1351 } else {
1352 // Starts humongous case: calculate how many regions are part of
1353 // this humongous region and then set the bit range.
1354 BitMap::idx_t end_index = (BitMap::idx_t) hr->last_hc_index();
1355 _region_bm->par_at_put_range(index, end_index, true);
1356 }
1357 }
1358
1359 public:
1360 CMCountDataClosureBase(G1CollectedHeap* g1h,
1361 BitMap* region_bm, BitMap* card_bm):
1362 _g1h(g1h), _cm(g1h->concurrent_mark()),
1363 _ct_bs((CardTableModRefBS*) (g1h->barrier_set())),
1364 _region_bm(region_bm), _card_bm(card_bm) { }
1365 };
1366
1367 // Closure that calculates the # live objects per region. Used
1368 // for verification purposes during the cleanup pause.
1369 class CalcLiveObjectsClosure: public CMCountDataClosureBase {
1370 CMBitMapRO* _bm;
1371 size_t _region_marked_bytes;
1372
1373 public:
1374 CalcLiveObjectsClosure(CMBitMapRO *bm, G1CollectedHeap* g1h,
1375 BitMap* region_bm, BitMap* card_bm) :
1376 CMCountDataClosureBase(g1h, region_bm, card_bm),
1377 _bm(bm), _region_marked_bytes(0) { }
1378
1379 bool doHeapRegion(HeapRegion* hr) {
1380
1381 if (hr->continuesHumongous()) {
1382 // We will ignore these here and process them when their
1383 // associated "starts humongous" region is processed (see
1384 // set_bit_for_heap_region()). Note that we cannot rely on their
1385 // associated "starts humongous" region to have their bit set to
1386 // 1 since, due to the region chunking in the parallel region
1387 // iteration, a "continues humongous" region might be visited
1388 // before its associated "starts humongous".
1389 return false;
1390 }
1391
1392 HeapWord* ntams = hr->next_top_at_mark_start();
1393 HeapWord* start = hr->bottom();
1394
1395 assert(start <= hr->end() && start <= ntams && ntams <= hr->end(),
1396 err_msg("Preconditions not met - "
1397 "start: "PTR_FORMAT", ntams: "PTR_FORMAT", end: "PTR_FORMAT,
1398 start, ntams, hr->end()));
1399
1400 // Find the first marked object at or after "start".
1401 start = _bm->getNextMarkedWordAddress(start, ntams);
1402
1403 size_t marked_bytes = 0;
1404
1405 while (start < ntams) {
1406 oop obj = oop(start);
1407 int obj_sz = obj->size();
1408 HeapWord* obj_end = start + obj_sz;
1409
1410 BitMap::idx_t start_idx = _cm->card_bitmap_index_for(start);
1411 BitMap::idx_t end_idx = _cm->card_bitmap_index_for(obj_end);
1412
1413 // Note: if we're looking at the last region in heap - obj_end
1414 // could be actually just beyond the end of the heap; end_idx
1415 // will then correspond to a (non-existent) card that is also
1416 // just beyond the heap.
1417 if (_g1h->is_in_g1_reserved(obj_end) && !_ct_bs->is_card_aligned(obj_end)) {
1418 // end of object is not card aligned - increment to cover
1419 // all the cards spanned by the object
1420 end_idx += 1;
1421 }
1422
1423 // Set the bits in the card BM for the cards spanned by this object.
1424 _cm->set_card_bitmap_range(_card_bm, start_idx, end_idx, true /* is_par */);
1425
1426 // Add the size of this object to the number of marked bytes.
1427 marked_bytes += (size_t)obj_sz * HeapWordSize;
1428
1429 // Find the next marked object after this one.
1430 start = _bm->getNextMarkedWordAddress(obj_end, ntams);
1431 }
1432
1433 // Mark the allocated-since-marking portion...
1434 HeapWord* top = hr->top();
1435 if (ntams < top) {
1436 BitMap::idx_t start_idx = _cm->card_bitmap_index_for(ntams);
1437 BitMap::idx_t end_idx = _cm->card_bitmap_index_for(top);
1438
1439 // Note: if we're looking at the last region in heap - top
1440 // could be actually just beyond the end of the heap; end_idx
1441 // will then correspond to a (non-existent) card that is also
1442 // just beyond the heap.
1443 if (_g1h->is_in_g1_reserved(top) && !_ct_bs->is_card_aligned(top)) {
1444 // end of object is not card aligned - increment to cover
1445 // all the cards spanned by the object
1446 end_idx += 1;
1447 }
1448 _cm->set_card_bitmap_range(_card_bm, start_idx, end_idx, true /* is_par */);
1449
1450 // This definitely means the region has live objects.
1451 set_bit_for_region(hr);
1452 }
1453
1454 // Update the live region bitmap.
1455 if (marked_bytes > 0) {
1456 set_bit_for_region(hr);
1457 }
1458
1459 // Set the marked bytes for the current region so that
1460 // it can be queried by a calling verificiation routine
1461 _region_marked_bytes = marked_bytes;
1462
1463 return false;
1464 }
1465
1466 size_t region_marked_bytes() const { return _region_marked_bytes; }
1467 };
1468
1469 // Heap region closure used for verifying the counting data
1470 // that was accumulated concurrently and aggregated during
1471 // the remark pause. This closure is applied to the heap
1472 // regions during the STW cleanup pause.
1473
1474 class VerifyLiveObjectDataHRClosure: public HeapRegionClosure {
1475 G1CollectedHeap* _g1h;
1476 ConcurrentMark* _cm;
1477 CalcLiveObjectsClosure _calc_cl;
1478 BitMap* _region_bm; // Region BM to be verified
1479 BitMap* _card_bm; // Card BM to be verified
1480 bool _verbose; // verbose output?
1481
1482 BitMap* _exp_region_bm; // Expected Region BM values
1483 BitMap* _exp_card_bm; // Expected card BM values
1484
1485 int _failures;
1486
1487 public:
1488 VerifyLiveObjectDataHRClosure(G1CollectedHeap* g1h,
1489 BitMap* region_bm,
1490 BitMap* card_bm,
1491 BitMap* exp_region_bm,
1492 BitMap* exp_card_bm,
1493 bool verbose) :
1494 _g1h(g1h), _cm(g1h->concurrent_mark()),
1495 _calc_cl(_cm->nextMarkBitMap(), g1h, exp_region_bm, exp_card_bm),
1496 _region_bm(region_bm), _card_bm(card_bm), _verbose(verbose),
1497 _exp_region_bm(exp_region_bm), _exp_card_bm(exp_card_bm),
1498 _failures(0) { }
1499
1500 int failures() const { return _failures; }
1501
1502 bool doHeapRegion(HeapRegion* hr) {
1503 if (hr->continuesHumongous()) {
1504 // We will ignore these here and process them when their
1505 // associated "starts humongous" region is processed (see
1506 // set_bit_for_heap_region()). Note that we cannot rely on their
1507 // associated "starts humongous" region to have their bit set to
1508 // 1 since, due to the region chunking in the parallel region
1509 // iteration, a "continues humongous" region might be visited
1510 // before its associated "starts humongous".
1511 return false;
1512 }
1513
1514 int failures = 0;
1515
1516 // Call the CalcLiveObjectsClosure to walk the marking bitmap for
1517 // this region and set the corresponding bits in the expected region
1518 // and card bitmaps.
1519 bool res = _calc_cl.doHeapRegion(hr);
1520 assert(res == false, "should be continuing");
1521
1522 MutexLockerEx x((_verbose ? ParGCRareEvent_lock : NULL),
1523 Mutex::_no_safepoint_check_flag);
1524
1525 // Verify the marked bytes for this region.
1526 size_t exp_marked_bytes = _calc_cl.region_marked_bytes();
1527 size_t act_marked_bytes = hr->next_marked_bytes();
1528
1529 // We're not OK if expected marked bytes > actual marked bytes. It means
1530 // we have missed accounting some objects during the actual marking.
1531 if (exp_marked_bytes > act_marked_bytes) {
1532 if (_verbose) {
1533 gclog_or_tty->print_cr("Region %u: marked bytes mismatch: "
1534 "expected: " SIZE_FORMAT ", actual: " SIZE_FORMAT,
1535 hr->hrs_index(), exp_marked_bytes, act_marked_bytes);
1536 }
1537 failures += 1;
1538 }
1539
1540 // Verify the bit, for this region, in the actual and expected
1541 // (which was just calculated) region bit maps.
1542 // We're not OK if the bit in the calculated expected region
1543 // bitmap is set and the bit in the actual region bitmap is not.
1544 BitMap::idx_t index = (BitMap::idx_t) hr->hrs_index();
1545
1546 bool expected = _exp_region_bm->at(index);
1547 bool actual = _region_bm->at(index);
1548 if (expected && !actual) {
1549 if (_verbose) {
1550 gclog_or_tty->print_cr("Region %u: region bitmap mismatch: "
1551 "expected: %s, actual: %s",
1552 hr->hrs_index(),
1553 BOOL_TO_STR(expected), BOOL_TO_STR(actual));
1554 }
1555 failures += 1;
1556 }
1557
1558 // Verify that the card bit maps for the cards spanned by the current
1559 // region match. We have an error if we have a set bit in the expected
1560 // bit map and the corresponding bit in the actual bitmap is not set.
1561
1562 BitMap::idx_t start_idx = _cm->card_bitmap_index_for(hr->bottom());
1563 BitMap::idx_t end_idx = _cm->card_bitmap_index_for(hr->top());
1564
1565 for (BitMap::idx_t i = start_idx; i < end_idx; i+=1) {
1566 expected = _exp_card_bm->at(i);
1567 actual = _card_bm->at(i);
1568
1569 if (expected && !actual) {
1570 if (_verbose) {
1571 gclog_or_tty->print_cr("Region %u: card bitmap mismatch at " SIZE_FORMAT ": "
1572 "expected: %s, actual: %s",
1573 hr->hrs_index(), i,
1574 BOOL_TO_STR(expected), BOOL_TO_STR(actual));
1575 }
1576 failures += 1;
1577 }
1578 }
1579
1580 if (failures > 0 && _verbose) {
1581 gclog_or_tty->print_cr("Region " HR_FORMAT ", ntams: " PTR_FORMAT ", "
1582 "marked_bytes: calc/actual " SIZE_FORMAT "/" SIZE_FORMAT,
1583 HR_FORMAT_PARAMS(hr), hr->next_top_at_mark_start(),
1584 _calc_cl.region_marked_bytes(), hr->next_marked_bytes());
1585 }
1586
1587 _failures += failures;
1588
1589 // We could stop iteration over the heap when we
1590 // find the first violating region by returning true.
1591 return false;
1592 }
1593 };
1594
1595
1596 class G1ParVerifyFinalCountTask: public AbstractGangTask {
1597 protected:
1598 G1CollectedHeap* _g1h;
1599 ConcurrentMark* _cm;
1600 BitMap* _actual_region_bm;
1601 BitMap* _actual_card_bm;
1602
1603 uint _n_workers;
1604
1605 BitMap* _expected_region_bm;
1606 BitMap* _expected_card_bm;
1607
1608 int _failures;
1609 bool _verbose;
1610
1611 public:
1612 G1ParVerifyFinalCountTask(G1CollectedHeap* g1h,
1613 BitMap* region_bm, BitMap* card_bm,
1614 BitMap* expected_region_bm, BitMap* expected_card_bm)
1615 : AbstractGangTask("G1 verify final counting"),
1616 _g1h(g1h), _cm(_g1h->concurrent_mark()),
1617 _actual_region_bm(region_bm), _actual_card_bm(card_bm),
1618 _expected_region_bm(expected_region_bm), _expected_card_bm(expected_card_bm),
1619 _failures(0), _verbose(false),
1620 _n_workers(0) {
1621 assert(VerifyDuringGC, "don't call this otherwise");
1622
1623 // Use the value already set as the number of active threads
1624 // in the call to run_task().
1625 if (G1CollectedHeap::use_parallel_gc_threads()) {
1626 assert( _g1h->workers()->active_workers() > 0,
1627 "Should have been previously set");
1628 _n_workers = _g1h->workers()->active_workers();
1629 } else {
1630 _n_workers = 1;
1631 }
1632
1633 assert(_expected_card_bm->size() == _actual_card_bm->size(), "sanity");
1634 assert(_expected_region_bm->size() == _actual_region_bm->size(), "sanity");
1635
1636 _verbose = _cm->verbose_medium();
1637 }
1638
1639 void work(uint worker_id) {
1640 assert(worker_id < _n_workers, "invariant");
1641
1642 VerifyLiveObjectDataHRClosure verify_cl(_g1h,
1643 _actual_region_bm, _actual_card_bm,
1644 _expected_region_bm,
1645 _expected_card_bm,
1646 _verbose);
1647
1648 if (G1CollectedHeap::use_parallel_gc_threads()) {
1649 _g1h->heap_region_par_iterate_chunked(&verify_cl,
1650 worker_id,
1651 _n_workers,
1652 HeapRegion::VerifyCountClaimValue);
1653 } else {
1654 _g1h->heap_region_iterate(&verify_cl);
1655 }
1656
1657 Atomic::add(verify_cl.failures(), &_failures);
1658 }
1659
1660 int failures() const { return _failures; }
1661 };
1662
1663 // Closure that finalizes the liveness counting data.
1664 // Used during the cleanup pause.
1665 // Sets the bits corresponding to the interval [NTAMS, top]
1666 // (which contains the implicitly live objects) in the
1667 // card liveness bitmap. Also sets the bit for each region,
1668 // containing live data, in the region liveness bitmap.
1669
1670 class FinalCountDataUpdateClosure: public CMCountDataClosureBase {
1671 public:
1672 FinalCountDataUpdateClosure(G1CollectedHeap* g1h,
1673 BitMap* region_bm,
1674 BitMap* card_bm) :
1675 CMCountDataClosureBase(g1h, region_bm, card_bm) { }
1676
1677 bool doHeapRegion(HeapRegion* hr) {
1678
1679 if (hr->continuesHumongous()) {
1680 // We will ignore these here and process them when their
1681 // associated "starts humongous" region is processed (see
1682 // set_bit_for_heap_region()). Note that we cannot rely on their
1683 // associated "starts humongous" region to have their bit set to
1684 // 1 since, due to the region chunking in the parallel region
1685 // iteration, a "continues humongous" region might be visited
1686 // before its associated "starts humongous".
1687 return false;
1688 }
1689
1690 HeapWord* ntams = hr->next_top_at_mark_start();
1691 HeapWord* top = hr->top();
1692
1693 assert(hr->bottom() <= ntams && ntams <= hr->end(), "Preconditions.");
1694
1695 // Mark the allocated-since-marking portion...
1696 if (ntams < top) {
1697 // This definitely means the region has live objects.
1698 set_bit_for_region(hr);
1699
1700 // Now set the bits in the card bitmap for [ntams, top)
1701 BitMap::idx_t start_idx = _cm->card_bitmap_index_for(ntams);
1702 BitMap::idx_t end_idx = _cm->card_bitmap_index_for(top);
1703
1704 // Note: if we're looking at the last region in heap - top
1705 // could be actually just beyond the end of the heap; end_idx
1706 // will then correspond to a (non-existent) card that is also
1707 // just beyond the heap.
1708 if (_g1h->is_in_g1_reserved(top) && !_ct_bs->is_card_aligned(top)) {
1709 // end of object is not card aligned - increment to cover
1710 // all the cards spanned by the object
1711 end_idx += 1;
1712 }
1713
1714 assert(end_idx <= _card_bm->size(),
1715 err_msg("oob: end_idx= "SIZE_FORMAT", bitmap size= "SIZE_FORMAT,
1716 end_idx, _card_bm->size()));
1717 assert(start_idx < _card_bm->size(),
1718 err_msg("oob: start_idx= "SIZE_FORMAT", bitmap size= "SIZE_FORMAT,
1719 start_idx, _card_bm->size()));
1720
1721 _cm->set_card_bitmap_range(_card_bm, start_idx, end_idx, true /* is_par */);
1722 }
1723
1724 // Set the bit for the region if it contains live data
1725 if (hr->next_marked_bytes() > 0) {
1726 set_bit_for_region(hr);
1727 }
1728
1729 return false;
1730 }
1731 };
1732
1733 class G1ParFinalCountTask: public AbstractGangTask {
1734 protected:
1735 G1CollectedHeap* _g1h;
1736 ConcurrentMark* _cm;
1737 BitMap* _actual_region_bm;
1738 BitMap* _actual_card_bm;
1739
1740 uint _n_workers;
1741
1742 public:
1743 G1ParFinalCountTask(G1CollectedHeap* g1h, BitMap* region_bm, BitMap* card_bm)
1744 : AbstractGangTask("G1 final counting"),
1745 _g1h(g1h), _cm(_g1h->concurrent_mark()),
1746 _actual_region_bm(region_bm), _actual_card_bm(card_bm),
1747 _n_workers(0) {
1748 // Use the value already set as the number of active threads
1749 // in the call to run_task().
1750 if (G1CollectedHeap::use_parallel_gc_threads()) {
1751 assert( _g1h->workers()->active_workers() > 0,
1752 "Should have been previously set");
1753 _n_workers = _g1h->workers()->active_workers();
1754 } else {
1755 _n_workers = 1;
1756 }
1757 }
1758
1759 void work(uint worker_id) {
1760 assert(worker_id < _n_workers, "invariant");
1761
1762 FinalCountDataUpdateClosure final_update_cl(_g1h,
1763 _actual_region_bm,
1764 _actual_card_bm);
1765
1766 if (G1CollectedHeap::use_parallel_gc_threads()) {
1767 _g1h->heap_region_par_iterate_chunked(&final_update_cl,
1768 worker_id,
1769 _n_workers,
1770 HeapRegion::FinalCountClaimValue);
1771 } else {
1772 _g1h->heap_region_iterate(&final_update_cl);
1773 }
1774 }
1775 };
1776
1777 class G1ParNoteEndTask;
1778
1779 class G1NoteEndOfConcMarkClosure : public HeapRegionClosure {
1780 G1CollectedHeap* _g1;
1781 int _worker_num;
1782 size_t _max_live_bytes;
1783 uint _regions_claimed;
1784 size_t _freed_bytes;
1785 FreeRegionList* _local_cleanup_list;
1786 OldRegionSet* _old_proxy_set;
1787 HumongousRegionSet* _humongous_proxy_set;
1788 HRRSCleanupTask* _hrrs_cleanup_task;
1789 double _claimed_region_time;
1790 double _max_region_time;
1791
1792 public:
1793 G1NoteEndOfConcMarkClosure(G1CollectedHeap* g1,
1794 int worker_num,
1795 FreeRegionList* local_cleanup_list,
1796 OldRegionSet* old_proxy_set,
1797 HumongousRegionSet* humongous_proxy_set,
1798 HRRSCleanupTask* hrrs_cleanup_task) :
1799 _g1(g1), _worker_num(worker_num),
1800 _max_live_bytes(0), _regions_claimed(0),
1801 _freed_bytes(0),
1802 _claimed_region_time(0.0), _max_region_time(0.0),
1803 _local_cleanup_list(local_cleanup_list),
1804 _old_proxy_set(old_proxy_set),
1805 _humongous_proxy_set(humongous_proxy_set),
1806 _hrrs_cleanup_task(hrrs_cleanup_task) { }
1807
1808 size_t freed_bytes() { return _freed_bytes; }
1809
1810 bool doHeapRegion(HeapRegion *hr) {
1811 if (hr->continuesHumongous()) {
1812 return false;
1813 }
1814 // We use a claim value of zero here because all regions
1815 // were claimed with value 1 in the FinalCount task.
1816 _g1->reset_gc_time_stamps(hr);
1817 double start = os::elapsedTime();
1818 _regions_claimed++;
1819 hr->note_end_of_marking();
1820 _max_live_bytes += hr->max_live_bytes();
1821 _g1->free_region_if_empty(hr,
1822 &_freed_bytes,
1823 _local_cleanup_list,
1824 _old_proxy_set,
1825 _humongous_proxy_set,
1826 _hrrs_cleanup_task,
1827 true /* par */);
1828 double region_time = (os::elapsedTime() - start);
1829 _claimed_region_time += region_time;
1830 if (region_time > _max_region_time) {
1831 _max_region_time = region_time;
1832 }
1833 return false;
1834 }
1835
1836 size_t max_live_bytes() { return _max_live_bytes; }
1837 uint regions_claimed() { return _regions_claimed; }
1838 double claimed_region_time_sec() { return _claimed_region_time; }
1839 double max_region_time_sec() { return _max_region_time; }
1840 };
1841
1842 class G1ParNoteEndTask: public AbstractGangTask {
1843 friend class G1NoteEndOfConcMarkClosure;
1844
1845 protected:
1846 G1CollectedHeap* _g1h;
1847 size_t _max_live_bytes;
1848 size_t _freed_bytes;
1849 FreeRegionList* _cleanup_list;
1850
1851 public:
1852 G1ParNoteEndTask(G1CollectedHeap* g1h,
1853 FreeRegionList* cleanup_list) :
1854 AbstractGangTask("G1 note end"), _g1h(g1h),
1855 _max_live_bytes(0), _freed_bytes(0), _cleanup_list(cleanup_list) { }
1856
1857 void work(uint worker_id) {
1858 double start = os::elapsedTime();
1859 FreeRegionList local_cleanup_list("Local Cleanup List");
1860 OldRegionSet old_proxy_set("Local Cleanup Old Proxy Set");
1861 HumongousRegionSet humongous_proxy_set("Local Cleanup Humongous Proxy Set");
1862 HRRSCleanupTask hrrs_cleanup_task;
1863 G1NoteEndOfConcMarkClosure g1_note_end(_g1h, worker_id, &local_cleanup_list,
1864 &old_proxy_set,
1865 &humongous_proxy_set,
1866 &hrrs_cleanup_task);
1867 if (G1CollectedHeap::use_parallel_gc_threads()) {
1868 _g1h->heap_region_par_iterate_chunked(&g1_note_end, worker_id,
1869 _g1h->workers()->active_workers(),
1870 HeapRegion::NoteEndClaimValue);
1871 } else {
1872 _g1h->heap_region_iterate(&g1_note_end);
1873 }
1874 assert(g1_note_end.complete(), "Shouldn't have yielded!");
1875
1876 // Now update the lists
1877 _g1h->update_sets_after_freeing_regions(g1_note_end.freed_bytes(),
1878 NULL /* free_list */,
1879 &old_proxy_set,
1880 &humongous_proxy_set,
1881 true /* par */);
1882 {
1883 MutexLockerEx x(ParGCRareEvent_lock, Mutex::_no_safepoint_check_flag);
1884 _max_live_bytes += g1_note_end.max_live_bytes();
1885 _freed_bytes += g1_note_end.freed_bytes();
1886
1887 // If we iterate over the global cleanup list at the end of
1888 // cleanup to do this printing we will not guarantee to only
1889 // generate output for the newly-reclaimed regions (the list
1890 // might not be empty at the beginning of cleanup; we might
1891 // still be working on its previous contents). So we do the
1892 // printing here, before we append the new regions to the global
1893 // cleanup list.
1894
1895 G1HRPrinter* hr_printer = _g1h->hr_printer();
1896 if (hr_printer->is_active()) {
1897 HeapRegionLinkedListIterator iter(&local_cleanup_list);
1898 while (iter.more_available()) {
1899 HeapRegion* hr = iter.get_next();
1900 hr_printer->cleanup(hr);
1901 }
1902 }
1903
1904 _cleanup_list->add_as_tail(&local_cleanup_list);
1905 assert(local_cleanup_list.is_empty(), "post-condition");
1906
1907 HeapRegionRemSet::finish_cleanup_task(&hrrs_cleanup_task);
1908 }
1909 }
1910 size_t max_live_bytes() { return _max_live_bytes; }
1911 size_t freed_bytes() { return _freed_bytes; }
1912 };
1913
1914 class G1ParScrubRemSetTask: public AbstractGangTask {
1915 protected:
1916 G1RemSet* _g1rs;
1917 BitMap* _region_bm;
1918 BitMap* _card_bm;
1919 public:
1920 G1ParScrubRemSetTask(G1CollectedHeap* g1h,
1921 BitMap* region_bm, BitMap* card_bm) :
1922 AbstractGangTask("G1 ScrubRS"), _g1rs(g1h->g1_rem_set()),
1923 _region_bm(region_bm), _card_bm(card_bm) { }
1924
1925 void work(uint worker_id) {
1926 if (G1CollectedHeap::use_parallel_gc_threads()) {
1927 _g1rs->scrub_par(_region_bm, _card_bm, worker_id,
1928 HeapRegion::ScrubRemSetClaimValue);
1929 } else {
1930 _g1rs->scrub(_region_bm, _card_bm);
1931 }
1932 }
1933
1934 };
1935
1936 void ConcurrentMark::cleanup() {
1937 // world is stopped at this checkpoint
1938 assert(SafepointSynchronize::is_at_safepoint(),
1939 "world should be stopped");
1940 G1CollectedHeap* g1h = G1CollectedHeap::heap();
1941
1942 // If a full collection has happened, we shouldn't do this.
1943 if (has_aborted()) {
1944 g1h->set_marking_complete(); // So bitmap clearing isn't confused
1945 return;
1946 }
1947
1948 HRSPhaseSetter x(HRSPhaseCleanup);
1949 g1h->verify_region_sets_optional();
1950
1951 if (VerifyDuringGC) {
1952 HandleMark hm; // handle scope
1953 gclog_or_tty->print(" VerifyDuringGC:(before)");
1954 Universe::heap()->prepare_for_verify();
1955 Universe::verify(/* silent */ false,
1956 /* option */ VerifyOption_G1UsePrevMarking);
1957 }
1958
1959 G1CollectorPolicy* g1p = G1CollectedHeap::heap()->g1_policy();
1960 g1p->record_concurrent_mark_cleanup_start();
1961
1962 double start = os::elapsedTime();
1963
1964 HeapRegionRemSet::reset_for_cleanup_tasks();
1965
1966 uint n_workers;
1967
1968 // Do counting once more with the world stopped for good measure.
1969 G1ParFinalCountTask g1_par_count_task(g1h, &_region_bm, &_card_bm);
1970
1971 if (G1CollectedHeap::use_parallel_gc_threads()) {
1972 assert(g1h->check_heap_region_claim_values(HeapRegion::InitialClaimValue),
1973 "sanity check");
1974
1975 g1h->set_par_threads();
1976 n_workers = g1h->n_par_threads();
1977 assert(g1h->n_par_threads() == n_workers,
1978 "Should not have been reset");
1979 g1h->workers()->run_task(&g1_par_count_task);
1980 // Done with the parallel phase so reset to 0.
1981 g1h->set_par_threads(0);
1982
1983 assert(g1h->check_heap_region_claim_values(HeapRegion::FinalCountClaimValue),
1984 "sanity check");
1985 } else {
1986 n_workers = 1;
1987 g1_par_count_task.work(0);
1988 }
1989
1990 if (VerifyDuringGC) {
1991 // Verify that the counting data accumulated during marking matches
1992 // that calculated by walking the marking bitmap.
1993
1994 // Bitmaps to hold expected values
1995 BitMap expected_region_bm(_region_bm.size(), false);
1996 BitMap expected_card_bm(_card_bm.size(), false);
1997
1998 G1ParVerifyFinalCountTask g1_par_verify_task(g1h,
1999 &_region_bm,
2000 &_card_bm,
2001 &expected_region_bm,
2002 &expected_card_bm);
2003
2004 if (G1CollectedHeap::use_parallel_gc_threads()) {
2005 g1h->set_par_threads((int)n_workers);
2006 g1h->workers()->run_task(&g1_par_verify_task);
2007 // Done with the parallel phase so reset to 0.
2008 g1h->set_par_threads(0);
2009
2010 assert(g1h->check_heap_region_claim_values(HeapRegion::VerifyCountClaimValue),
2011 "sanity check");
2012 } else {
2013 g1_par_verify_task.work(0);
2014 }
2015
2016 guarantee(g1_par_verify_task.failures() == 0, "Unexpected accounting failures");
2017 }
2018
2019 size_t start_used_bytes = g1h->used();
2020 g1h->set_marking_complete();
2021
2022 double count_end = os::elapsedTime();
2023 double this_final_counting_time = (count_end - start);
2024 _total_counting_time += this_final_counting_time;
2025
2026 if (G1PrintRegionLivenessInfo) {
2027 G1PrintRegionLivenessInfoClosure cl(gclog_or_tty, "Post-Marking");
2028 _g1h->heap_region_iterate(&cl);
2029 }
2030
2031 // Install newly created mark bitMap as "prev".
2032 swapMarkBitMaps();
2033
2034 g1h->reset_gc_time_stamp();
2035
2036 // Note end of marking in all heap regions.
2037 G1ParNoteEndTask g1_par_note_end_task(g1h, &_cleanup_list);
2038 if (G1CollectedHeap::use_parallel_gc_threads()) {
2039 g1h->set_par_threads((int)n_workers);
2040 g1h->workers()->run_task(&g1_par_note_end_task);
2041 g1h->set_par_threads(0);
2042
2043 assert(g1h->check_heap_region_claim_values(HeapRegion::NoteEndClaimValue),
2044 "sanity check");
2045 } else {
2046 g1_par_note_end_task.work(0);
2047 }
2048 g1h->check_gc_time_stamps();
2049
2050 if (!cleanup_list_is_empty()) {
2051 // The cleanup list is not empty, so we'll have to process it
2052 // concurrently. Notify anyone else that might be wanting free
2053 // regions that there will be more free regions coming soon.
2054 g1h->set_free_regions_coming();
2055 }
2056
2057 // call below, since it affects the metric by which we sort the heap
2058 // regions.
2059 if (G1ScrubRemSets) {
2060 double rs_scrub_start = os::elapsedTime();
2061 G1ParScrubRemSetTask g1_par_scrub_rs_task(g1h, &_region_bm, &_card_bm);
2062 if (G1CollectedHeap::use_parallel_gc_threads()) {
2063 g1h->set_par_threads((int)n_workers);
2064 g1h->workers()->run_task(&g1_par_scrub_rs_task);
2065 g1h->set_par_threads(0);
2066
2067 assert(g1h->check_heap_region_claim_values(
2068 HeapRegion::ScrubRemSetClaimValue),
2069 "sanity check");
2070 } else {
2071 g1_par_scrub_rs_task.work(0);
2072 }
2073
2074 double rs_scrub_end = os::elapsedTime();
2075 double this_rs_scrub_time = (rs_scrub_end - rs_scrub_start);
2076 _total_rs_scrub_time += this_rs_scrub_time;
2077 }
2078
2079 // this will also free any regions totally full of garbage objects,
2080 // and sort the regions.
2081 g1h->g1_policy()->record_concurrent_mark_cleanup_end((int)n_workers);
2082
2083 // Statistics.
2084 double end = os::elapsedTime();
2085 _cleanup_times.add((end - start) * 1000.0);
2086
2087 if (G1Log::fine()) {
2088 g1h->print_size_transition(gclog_or_tty,
2089 start_used_bytes,
2090 g1h->used(),
2091 g1h->capacity());
2092 }
2093
2094 // Clean up will have freed any regions completely full of garbage.
2095 // Update the soft reference policy with the new heap occupancy.
2096 Universe::update_heap_info_at_gc();
2097
2098 // We need to make this be a "collection" so any collection pause that
2099 // races with it goes around and waits for completeCleanup to finish.
2100 g1h->increment_total_collections();
2101
2102 // We reclaimed old regions so we should calculate the sizes to make
2103 // sure we update the old gen/space data.
2104 g1h->g1mm()->update_sizes();
2105
2106 if (VerifyDuringGC) {
2107 HandleMark hm; // handle scope
2108 gclog_or_tty->print(" VerifyDuringGC:(after)");
2109 Universe::heap()->prepare_for_verify();
2110 Universe::verify(/* silent */ false,
2111 /* option */ VerifyOption_G1UsePrevMarking);
2112 }
2113
2114 g1h->verify_region_sets_optional();
2115 }
2116
2117 void ConcurrentMark::completeCleanup() {
2118 if (has_aborted()) return;
2119
2120 G1CollectedHeap* g1h = G1CollectedHeap::heap();
2121
2122 _cleanup_list.verify_optional();
2123 FreeRegionList tmp_free_list("Tmp Free List");
2124
2125 if (G1ConcRegionFreeingVerbose) {
2126 gclog_or_tty->print_cr("G1ConcRegionFreeing [complete cleanup] : "
2127 "cleanup list has %u entries",
2128 _cleanup_list.length());
2129 }
2130
2131 // Noone else should be accessing the _cleanup_list at this point,
2132 // so it's not necessary to take any locks
2133 while (!_cleanup_list.is_empty()) {
2134 HeapRegion* hr = _cleanup_list.remove_head();
2135 assert(hr != NULL, "the list was not empty");
2136 hr->par_clear();
2137 tmp_free_list.add_as_tail(hr);
2138
2139 // Instead of adding one region at a time to the secondary_free_list,
2140 // we accumulate them in the local list and move them a few at a
2141 // time. This also cuts down on the number of notify_all() calls
2142 // we do during this process. We'll also append the local list when
2143 // _cleanup_list is empty (which means we just removed the last
2144 // region from the _cleanup_list).
2145 if ((tmp_free_list.length() % G1SecondaryFreeListAppendLength == 0) ||
2146 _cleanup_list.is_empty()) {
2147 if (G1ConcRegionFreeingVerbose) {
2148 gclog_or_tty->print_cr("G1ConcRegionFreeing [complete cleanup] : "
2149 "appending %u entries to the secondary_free_list, "
2150 "cleanup list still has %u entries",
2151 tmp_free_list.length(),
2152 _cleanup_list.length());
2153 }
2154
2155 {
2156 MutexLockerEx x(SecondaryFreeList_lock, Mutex::_no_safepoint_check_flag);
2157 g1h->secondary_free_list_add_as_tail(&tmp_free_list);
2158 SecondaryFreeList_lock->notify_all();
2159 }
2160
2161 if (G1StressConcRegionFreeing) {
2162 for (uintx i = 0; i < G1StressConcRegionFreeingDelayMillis; ++i) {
2163 os::sleep(Thread::current(), (jlong) 1, false);
2164 }
2165 }
2166 }
2167 }
2168 assert(tmp_free_list.is_empty(), "post-condition");
2169 }
2170
2171 // Supporting Object and Oop closures for reference discovery
2172 // and processing in during marking
2173
2174 bool G1CMIsAliveClosure::do_object_b(oop obj) {
2175 HeapWord* addr = (HeapWord*)obj;
2176 return addr != NULL &&
2177 (!_g1->is_in_g1_reserved(addr) || !_g1->is_obj_ill(obj));
2178 }
2179
2180 // 'Keep Alive' oop closure used by both serial parallel reference processing.
2181 // Uses the CMTask associated with a worker thread (for serial reference
2182 // processing the CMTask for worker 0 is used) to preserve (mark) and
2183 // trace referent objects.
2184 //
2185 // Using the CMTask and embedded local queues avoids having the worker
2186 // threads operating on the global mark stack. This reduces the risk
2187 // of overflowing the stack - which we would rather avoid at this late
2188 // state. Also using the tasks' local queues removes the potential
2189 // of the workers interfering with each other that could occur if
2190 // operating on the global stack.
2191
2192 class G1CMKeepAliveAndDrainClosure: public OopClosure {
2193 ConcurrentMark* _cm;
2194 CMTask* _task;
2195 int _ref_counter_limit;
2196 int _ref_counter;
2197 bool _is_serial;
2198 public:
2199 G1CMKeepAliveAndDrainClosure(ConcurrentMark* cm, CMTask* task, bool is_serial) :
2200 _cm(cm), _task(task), _is_serial(is_serial),
2201 _ref_counter_limit(G1RefProcDrainInterval) {
2202 assert(_ref_counter_limit > 0, "sanity");
2203 assert(!_is_serial || _task->worker_id() == 0, "only task 0 for serial code");
2204 _ref_counter = _ref_counter_limit;
2205 }
2206
2207 virtual void do_oop(narrowOop* p) { do_oop_work(p); }
2208 virtual void do_oop( oop* p) { do_oop_work(p); }
2209
2210 template <class T> void do_oop_work(T* p) {
2211 if (!_cm->has_overflown()) {
2212 oop obj = oopDesc::load_decode_heap_oop(p);
2213 if (_cm->verbose_high()) {
2214 gclog_or_tty->print_cr("\t[%u] we're looking at location "
2215 "*"PTR_FORMAT" = "PTR_FORMAT,
2216 _task->worker_id(), p, (void*) obj);
2217 }
2218
2219 _task->deal_with_reference(obj);
2220 _ref_counter--;
2221
2222 if (_ref_counter == 0) {
2223 // We have dealt with _ref_counter_limit references, pushing them
2224 // and objects reachable from them on to the local stack (and
2225 // possibly the global stack). Call CMTask::do_marking_step() to
2226 // process these entries.
2227 //
2228 // We call CMTask::do_marking_step() in a loop, which we'll exit if
2229 // there's nothing more to do (i.e. we're done with the entries that
2230 // were pushed as a result of the CMTask::deal_with_reference() calls
2231 // above) or we overflow.
2232 //
2233 // Note: CMTask::do_marking_step() can set the CMTask::has_aborted()
2234 // flag while there may still be some work to do. (See the comment at
2235 // the beginning of CMTask::do_marking_step() for those conditions -
2236 // one of which is reaching the specified time target.) It is only
2237 // when CMTask::do_marking_step() returns without setting the
2238 // has_aborted() flag that the marking step has completed.
2239 do {
2240 double mark_step_duration_ms = G1ConcMarkStepDurationMillis;
2241 _task->do_marking_step(mark_step_duration_ms,
2242 false /* do_stealing */,
2243 false /* do_termination */,
2244 _is_serial /* is_serial */);
2245 } while (_task->has_aborted() && !_cm->has_overflown());
2246 _ref_counter = _ref_counter_limit;
2247 }
2248 } else {
2249 if (_cm->verbose_high()) {
2250 gclog_or_tty->print_cr("\t[%u] CM Overflow", _task->worker_id());
2251 }
2252 }
2253 }
2254 };
2255
2256 // 'Drain' oop closure used by both serial and parallel reference processing.
2257 // Uses the CMTask associated with a given worker thread (for serial
2258 // reference processing the CMtask for worker 0 is used). Calls the
2259 // do_marking_step routine, with an unbelievably large timeout value,
2260 // to drain the marking data structures of the remaining entries
2261 // added by the 'keep alive' oop closure above.
2262
2263 class G1CMDrainMarkingStackClosure: public VoidClosure {
2264 ConcurrentMark* _cm;
2265 CMTask* _task;
2266 bool _is_serial;
2267 public:
2268 G1CMDrainMarkingStackClosure(ConcurrentMark* cm, CMTask* task, bool is_serial) :
2269 _cm(cm), _task(task) {
2270 assert(!_is_serial || _task->worker_id() == 0, "only task 0 for serial code");
2271 }
2272
2273 void do_void() {
2274 do {
2275 if (_cm->verbose_high()) {
2276 gclog_or_tty->print_cr("\t[%u] Drain: Calling do_marking_step - serial: %s",
2277 _task->worker_id(), BOOL_TO_STR(_is_serial));
2278 }
2279
2280 // We call CMTask::do_marking_step() to completely drain the local
2281 // and global marking stacks of entries pushed by the 'keep alive'
2282 // oop closure (an instance of G1CMKeepAliveAndDrainClosure above).
2283 //
2284 // CMTask::do_marking_step() is called in a loop, which we'll exit
2285 // if there's nothing more to do (i.e. we'completely drained the
2286 // entries that were pushed as a a result of applying the 'keep alive'
2287 // closure to the entries on the discovered ref lists) or we overflow
2288 // the global marking stack.
2289 //
2290 // Note: CMTask::do_marking_step() can set the CMTask::has_aborted()
2291 // flag while there may still be some work to do. (See the comment at
2292 // the beginning of CMTask::do_marking_step() for those conditions -
2293 // one of which is reaching the specified time target.) It is only
2294 // when CMTask::do_marking_step() returns without setting the
2295 // has_aborted() flag that the marking step has completed.
2296
2297 // It only makes sense to allow stealing in CMTask::do_marking_step()
2298 // if this closure is being instantiated for parallel reference
2299 // processing.
2300
2301 _task->do_marking_step(1000000000.0 /* something very large */,
2302 !_is_serial /* do_stealing */,
2303 true /* do_termination */,
2304 _is_serial /* is_serial */);
2305 } while (_task->has_aborted() && !_cm->has_overflown());
2306 }
2307 };
2308
2309 // Implementation of AbstractRefProcTaskExecutor for parallel
2310 // reference processing at the end of G1 concurrent marking
2311
2312 class G1CMRefProcTaskExecutor: public AbstractRefProcTaskExecutor {
2313 private:
2314 G1CollectedHeap* _g1h;
2315 ConcurrentMark* _cm;
2316 WorkGang* _workers;
2317 int _active_workers;
2318
2319 public:
2320 G1CMRefProcTaskExecutor(G1CollectedHeap* g1h,
2321 ConcurrentMark* cm,
2322 WorkGang* workers,
2323 int n_workers) :
2324 _g1h(g1h), _cm(cm),
2325 _workers(workers), _active_workers(n_workers) { }
2326
2327 // Executes the given task using concurrent marking worker threads.
2328 virtual void execute(ProcessTask& task);
2329 virtual void execute(EnqueueTask& task);
2330 };
2331
2332 class G1CMRefProcTaskProxy: public AbstractGangTask {
2333 typedef AbstractRefProcTaskExecutor::ProcessTask ProcessTask;
2334 ProcessTask& _proc_task;
2335 G1CollectedHeap* _g1h;
2336 ConcurrentMark* _cm;
2337
2338 public:
2339 G1CMRefProcTaskProxy(ProcessTask& proc_task,
2340 G1CollectedHeap* g1h,
2341 ConcurrentMark* cm) :
2342 AbstractGangTask("Process reference objects in parallel"),
2343 _proc_task(proc_task), _g1h(g1h), _cm(cm) {
2344 ReferenceProcessor* rp = _g1h->ref_processor_cm();
2345 assert(rp->processing_is_mt(), "shouldn't be here otherwise");
2346 }
2347
2348 virtual void work(uint worker_id) {
2349 CMTask* task = _cm->task(worker_id);
2350 G1CMIsAliveClosure g1_is_alive(_g1h);
2351 G1CMKeepAliveAndDrainClosure g1_par_keep_alive(_cm, task, false /* is_serial */);
2352 G1CMDrainMarkingStackClosure g1_par_drain(_cm, task, false /* is_serial */);
2353
2354 _proc_task.work(worker_id, g1_is_alive, g1_par_keep_alive, g1_par_drain);
2355 }
2356 };
2357
2358 void G1CMRefProcTaskExecutor::execute(ProcessTask& proc_task) {
2359 assert(_workers != NULL, "Need parallel worker threads.");
2360 assert(_g1h->ref_processor_cm()->processing_is_mt(), "processing is not MT");
2361
2362 G1CMRefProcTaskProxy proc_task_proxy(proc_task, _g1h, _cm);
2363
2364 // We need to reset the phase for each task execution so that
2365 // the termination protocol of CMTask::do_marking_step works.
2366 _cm->set_phase(_active_workers, false /* concurrent */);
2367 _g1h->set_par_threads(_active_workers);
2368 _workers->run_task(&proc_task_proxy);
2369 _g1h->set_par_threads(0);
2370 }
2371
2372 class G1CMRefEnqueueTaskProxy: public AbstractGangTask {
2373 typedef AbstractRefProcTaskExecutor::EnqueueTask EnqueueTask;
2374 EnqueueTask& _enq_task;
2375
2376 public:
2377 G1CMRefEnqueueTaskProxy(EnqueueTask& enq_task) :
2378 AbstractGangTask("Enqueue reference objects in parallel"),
2379 _enq_task(enq_task) { }
2380
2381 virtual void work(uint worker_id) {
2382 _enq_task.work(worker_id);
2383 }
2384 };
2385
2386 void G1CMRefProcTaskExecutor::execute(EnqueueTask& enq_task) {
2387 assert(_workers != NULL, "Need parallel worker threads.");
2388 assert(_g1h->ref_processor_cm()->processing_is_mt(), "processing is not MT");
2389
2390 G1CMRefEnqueueTaskProxy enq_task_proxy(enq_task);
2391
2392 _g1h->set_par_threads(_active_workers);
2393 _workers->run_task(&enq_task_proxy);
2394 _g1h->set_par_threads(0);
2395 }
2396
2397 void ConcurrentMark::weakRefsWork(bool clear_all_soft_refs) {
2398 ResourceMark rm;
2399 HandleMark hm;
2400
2401 G1CollectedHeap* g1h = G1CollectedHeap::heap();
2402
2403 // Is alive closure.
2404 G1CMIsAliveClosure g1_is_alive(g1h);
2405
2406 // Inner scope to exclude the cleaning of the string and symbol
2407 // tables from the displayed time.
2408 {
2409 if (G1Log::finer()) {
2410 gclog_or_tty->put(' ');
2411 }
2412 TraceTime t("GC ref-proc", G1Log::finer(), false, gclog_or_tty);
2413
2414 ReferenceProcessor* rp = g1h->ref_processor_cm();
2415
2416 // See the comment in G1CollectedHeap::ref_processing_init()
2417 // about how reference processing currently works in G1.
2418
2419 // Set the soft reference policy
2420 rp->setup_policy(clear_all_soft_refs);
2421 assert(_markStack.isEmpty(), "mark stack should be empty");
2422
2423 // Instances of the 'Keep Alive' and 'Complete GC' closures used
2424 // in serial reference processing. Note these closures are also
2425 // used for serially processing (by the the current thread) the
2426 // JNI references during parallel reference processing.
2427 //
2428 // These closures do not need to synchronize with the worker
2429 // threads involved in parallel reference processing as these
2430 // instances are executed serially by the current thread (e.g.
2431 // reference processing is not multi-threaded and is thus
2432 // performed by the current thread instead of a gang worker).
2433 //
2434 // The gang tasks involved in parallel reference procssing create
2435 // their own instances of these closures, which do their own
2436 // synchronization among themselves.
2437 G1CMKeepAliveAndDrainClosure g1_keep_alive(this, task(0), true /* is_serial */);
2438 G1CMDrainMarkingStackClosure g1_drain_mark_stack(this, task(0), true /* is_serial */);
2439
2440 // We need at least one active thread. If reference processing
2441 // is not multi-threaded we use the current (VMThread) thread,
2442 // otherwise we use the work gang from the G1CollectedHeap and
2443 // we utilize all the worker threads we can.
2444 bool processing_is_mt = rp->processing_is_mt() && g1h->workers() != NULL;
2445 uint active_workers = (processing_is_mt ? g1h->workers()->active_workers() : 1U);
2446 active_workers = MAX2(MIN2(active_workers, _max_worker_id), 1U);
2447
2448 // Parallel processing task executor.
2449 G1CMRefProcTaskExecutor par_task_executor(g1h, this,
2450 g1h->workers(), active_workers);
2451 AbstractRefProcTaskExecutor* executor = (processing_is_mt ? &par_task_executor : NULL);
2452
2453 // Set the degree of MT processing here. If the discovery was done MT,
2454 // the number of threads involved during discovery could differ from
2455 // the number of active workers. This is OK as long as the discovered
2456 // Reference lists are balanced (see balance_all_queues() and balance_queues()).
2457 rp->set_active_mt_degree(active_workers);
2458
2459 // Process the weak references.
2460 rp->process_discovered_references(&g1_is_alive,
2461 &g1_keep_alive,
2462 &g1_drain_mark_stack,
2463 executor);
2464
2465 // The do_oop work routines of the keep_alive and drain_marking_stack
2466 // oop closures will set the has_overflown flag if we overflow the
2467 // global marking stack.
2468
2469 assert(_markStack.overflow() || _markStack.isEmpty(),
2470 "mark stack should be empty (unless it overflowed)");
2471
2472 if (_markStack.overflow()) {
2473 // This should have been done already when we tried to push an
2474 // entry on to the global mark stack. But let's do it again.
2475 set_has_overflown();
2476 }
2477
2478 assert(rp->num_q() == active_workers, "why not");
2479
2480 rp->enqueue_discovered_references(executor);
2481
2482 rp->verify_no_references_recorded();
2483 assert(!rp->discovery_enabled(), "Post condition");
2484 }
2485
2486 // Now clean up stale oops in StringTable
2487 StringTable::unlink(&g1_is_alive);
2488 // Clean up unreferenced symbols in symbol table.
2489 SymbolTable::unlink();
2490 }
2491
2492 void ConcurrentMark::swapMarkBitMaps() {
2493 CMBitMapRO* temp = _prevMarkBitMap;
2494 _prevMarkBitMap = (CMBitMapRO*)_nextMarkBitMap;
2495 _nextMarkBitMap = (CMBitMap*) temp;
2496 }
2497
2498 class CMRemarkTask: public AbstractGangTask {
2499 private:
2500 ConcurrentMark* _cm;
2501 bool _is_serial;
2502 public:
2503 void work(uint worker_id) {
2504 // Since all available tasks are actually started, we should
2505 // only proceed if we're supposed to be actived.
2506 if (worker_id < _cm->active_tasks()) {
2507 CMTask* task = _cm->task(worker_id);
2508 task->record_start_time();
2509 do {
2510 task->do_marking_step(1000000000.0 /* something very large */,
2511 true /* do_stealing */,
2512 true /* do_termination */,
2513 _is_serial);
2514 } while (task->has_aborted() && !_cm->has_overflown());
2515 // If we overflow, then we do not want to restart. We instead
2516 // want to abort remark and do concurrent marking again.
2517 task->record_end_time();
2518 }
2519 }
2520
2521 CMRemarkTask(ConcurrentMark* cm, int active_workers, bool is_serial) :
2522 AbstractGangTask("Par Remark"), _cm(cm), _is_serial(is_serial) {
2523 _cm->terminator()->reset_for_reuse(active_workers);
2524 }
2525 };
2526
2527 void ConcurrentMark::checkpointRootsFinalWork() {
2528 ResourceMark rm;
2529 HandleMark hm;
2530 G1CollectedHeap* g1h = G1CollectedHeap::heap();
2531
2532 g1h->ensure_parsability(false);
2533
2534 if (G1CollectedHeap::use_parallel_gc_threads()) {
2535 G1CollectedHeap::StrongRootsScope srs(g1h);
2536 // this is remark, so we'll use up all active threads
2537 uint active_workers = g1h->workers()->active_workers();
2538 if (active_workers == 0) {
2539 assert(active_workers > 0, "Should have been set earlier");
2540 active_workers = (uint) ParallelGCThreads;
2541 g1h->workers()->set_active_workers(active_workers);
2542 }
2543 set_phase(active_workers, false /* concurrent */);
2544 // Leave _parallel_marking_threads at it's
2545 // value originally calculated in the ConcurrentMark
2546 // constructor and pass values of the active workers
2547 // through the gang in the task.
2548
2549 CMRemarkTask remarkTask(this, active_workers, false /* is_serial */);
2550 // We will start all available threads, even if we decide that the
2551 // active_workers will be fewer. The extra ones will just bail out
2552 // immediately.
2553 g1h->set_par_threads(active_workers);
2554 g1h->workers()->run_task(&remarkTask);
2555 g1h->set_par_threads(0);
2556 } else {
2557 G1CollectedHeap::StrongRootsScope srs(g1h);
2558 uint active_workers = 1;
2559 set_phase(active_workers, false /* concurrent */);
2560
2561 // Note - if there's no work gang then the VMThread will be
2562 // the thread to execute the remark - serially. We have
2563 // to pass true for the is_serial parameter so that
2564 // CMTask::do_marking_step() doesn't enter the sync
2565 // barriers in the event of an overflow. Doing so will
2566 // cause an assert that the current thread is not a
2567 // concurrent GC thread.
2568 CMRemarkTask remarkTask(this, active_workers, true /* is_serial*/);
2569 remarkTask.work(0);
2570 }
2571 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set();
2572 guarantee(satb_mq_set.completed_buffers_num() == 0, "invariant");
2573
2574 print_stats();
2575
2576 #if VERIFY_OBJS_PROCESSED
2577 if (_scan_obj_cl.objs_processed != ThreadLocalObjQueue::objs_enqueued) {
2578 gclog_or_tty->print_cr("Processed = %d, enqueued = %d.",
2579 _scan_obj_cl.objs_processed,
2580 ThreadLocalObjQueue::objs_enqueued);
2581 guarantee(_scan_obj_cl.objs_processed ==
2582 ThreadLocalObjQueue::objs_enqueued,
2583 "Different number of objs processed and enqueued.");
2584 }
2585 #endif
2586 }
2587
2588 #ifndef PRODUCT
2589
2590 class PrintReachableOopClosure: public OopClosure {
2591 private:
2592 G1CollectedHeap* _g1h;
2593 outputStream* _out;
2594 VerifyOption _vo;
2595 bool _all;
2596
2597 public:
2598 PrintReachableOopClosure(outputStream* out,
2599 VerifyOption vo,
2600 bool all) :
2601 _g1h(G1CollectedHeap::heap()),
2602 _out(out), _vo(vo), _all(all) { }
2603
2604 void do_oop(narrowOop* p) { do_oop_work(p); }
2605 void do_oop( oop* p) { do_oop_work(p); }
2606
2607 template <class T> void do_oop_work(T* p) {
2608 oop obj = oopDesc::load_decode_heap_oop(p);
2609 const char* str = NULL;
2610 const char* str2 = "";
2611
2612 if (obj == NULL) {
2613 str = "";
2614 } else if (!_g1h->is_in_g1_reserved(obj)) {
2615 str = " O";
2616 } else {
2617 HeapRegion* hr = _g1h->heap_region_containing(obj);
2618 guarantee(hr != NULL, "invariant");
2619 bool over_tams = _g1h->allocated_since_marking(obj, hr, _vo);
2620 bool marked = _g1h->is_marked(obj, _vo);
2621
2622 if (over_tams) {
2623 str = " >";
2624 if (marked) {
2625 str2 = " AND MARKED";
2626 }
2627 } else if (marked) {
2628 str = " M";
2629 } else {
2630 str = " NOT";
2631 }
2632 }
2633
2634 _out->print_cr(" "PTR_FORMAT": "PTR_FORMAT"%s%s",
2635 p, (void*) obj, str, str2);
2636 }
2637 };
2638
2639 class PrintReachableObjectClosure : public ObjectClosure {
2640 private:
2641 G1CollectedHeap* _g1h;
2642 outputStream* _out;
2643 VerifyOption _vo;
2644 bool _all;
2645 HeapRegion* _hr;
2646
2647 public:
2648 PrintReachableObjectClosure(outputStream* out,
2649 VerifyOption vo,
2650 bool all,
2651 HeapRegion* hr) :
2652 _g1h(G1CollectedHeap::heap()),
2653 _out(out), _vo(vo), _all(all), _hr(hr) { }
2654
2655 void do_object(oop o) {
2656 bool over_tams = _g1h->allocated_since_marking(o, _hr, _vo);
2657 bool marked = _g1h->is_marked(o, _vo);
2658 bool print_it = _all || over_tams || marked;
2659
2660 if (print_it) {
2661 _out->print_cr(" "PTR_FORMAT"%s",
2662 o, (over_tams) ? " >" : (marked) ? " M" : "");
2663 PrintReachableOopClosure oopCl(_out, _vo, _all);
2664 o->oop_iterate_no_header(&oopCl);
2665 }
2666 }
2667 };
2668
2669 class PrintReachableRegionClosure : public HeapRegionClosure {
2670 private:
2671 G1CollectedHeap* _g1h;
2672 outputStream* _out;
2673 VerifyOption _vo;
2674 bool _all;
2675
2676 public:
2677 bool doHeapRegion(HeapRegion* hr) {
2678 HeapWord* b = hr->bottom();
2679 HeapWord* e = hr->end();
2680 HeapWord* t = hr->top();
2681 HeapWord* p = _g1h->top_at_mark_start(hr, _vo);
2682 _out->print_cr("** ["PTR_FORMAT", "PTR_FORMAT"] top: "PTR_FORMAT" "
2683 "TAMS: "PTR_FORMAT, b, e, t, p);
2684 _out->cr();
2685
2686 HeapWord* from = b;
2687 HeapWord* to = t;
2688
2689 if (to > from) {
2690 _out->print_cr("Objects in ["PTR_FORMAT", "PTR_FORMAT"]", from, to);
2691 _out->cr();
2692 PrintReachableObjectClosure ocl(_out, _vo, _all, hr);
2693 hr->object_iterate_mem_careful(MemRegion(from, to), &ocl);
2694 _out->cr();
2695 }
2696
2697 return false;
2698 }
2699
2700 PrintReachableRegionClosure(outputStream* out,
2701 VerifyOption vo,
2702 bool all) :
2703 _g1h(G1CollectedHeap::heap()), _out(out), _vo(vo), _all(all) { }
2704 };
2705
2706 void ConcurrentMark::print_reachable(const char* str,
2707 VerifyOption vo,
2708 bool all) {
2709 gclog_or_tty->cr();
2710 gclog_or_tty->print_cr("== Doing heap dump... ");
2711
2712 if (G1PrintReachableBaseFile == NULL) {
2713 gclog_or_tty->print_cr(" #### error: no base file defined");
2714 return;
2715 }
2716
2717 if (strlen(G1PrintReachableBaseFile) + 1 + strlen(str) >
2718 (JVM_MAXPATHLEN - 1)) {
2719 gclog_or_tty->print_cr(" #### error: file name too long");
2720 return;
2721 }
2722
2723 char file_name[JVM_MAXPATHLEN];
2724 sprintf(file_name, "%s.%s", G1PrintReachableBaseFile, str);
2725 gclog_or_tty->print_cr(" dumping to file %s", file_name);
2726
2727 fileStream fout(file_name);
2728 if (!fout.is_open()) {
2729 gclog_or_tty->print_cr(" #### error: could not open file");
2730 return;
2731 }
2732
2733 outputStream* out = &fout;
2734 out->print_cr("-- USING %s", _g1h->top_at_mark_start_str(vo));
2735 out->cr();
2736
2737 out->print_cr("--- ITERATING OVER REGIONS");
2738 out->cr();
2739 PrintReachableRegionClosure rcl(out, vo, all);
2740 _g1h->heap_region_iterate(&rcl);
2741 out->cr();
2742
2743 gclog_or_tty->print_cr(" done");
2744 gclog_or_tty->flush();
2745 }
2746
2747 #endif // PRODUCT
2748
2749 void ConcurrentMark::clearRangePrevBitmap(MemRegion mr) {
2750 // Note we are overriding the read-only view of the prev map here, via
2751 // the cast.
2752 ((CMBitMap*)_prevMarkBitMap)->clearRange(mr);
2753 }
2754
2755 void ConcurrentMark::clearRangeNextBitmap(MemRegion mr) {
2756 _nextMarkBitMap->clearRange(mr);
2757 }
2758
2759 void ConcurrentMark::clearRangeBothBitmaps(MemRegion mr) {
2760 clearRangePrevBitmap(mr);
2761 clearRangeNextBitmap(mr);
2762 }
2763
2764 HeapRegion*
2765 ConcurrentMark::claim_region(uint worker_id) {
2766 // "checkpoint" the finger
2767 HeapWord* finger = _finger;
2768
2769 // _heap_end will not change underneath our feet; it only changes at
2770 // yield points.
2771 while (finger < _heap_end) {
2772 assert(_g1h->is_in_g1_reserved(finger), "invariant");
2773
2774 // Note on how this code handles humongous regions. In the
2775 // normal case the finger will reach the start of a "starts
2776 // humongous" (SH) region. Its end will either be the end of the
2777 // last "continues humongous" (CH) region in the sequence, or the
2778 // standard end of the SH region (if the SH is the only region in
2779 // the sequence). That way claim_region() will skip over the CH
2780 // regions. However, there is a subtle race between a CM thread
2781 // executing this method and a mutator thread doing a humongous
2782 // object allocation. The two are not mutually exclusive as the CM
2783 // thread does not need to hold the Heap_lock when it gets
2784 // here. So there is a chance that claim_region() will come across
2785 // a free region that's in the progress of becoming a SH or a CH
2786 // region. In the former case, it will either
2787 // a) Miss the update to the region's end, in which case it will
2788 // visit every subsequent CH region, will find their bitmaps
2789 // empty, and do nothing, or
2790 // b) Will observe the update of the region's end (in which case
2791 // it will skip the subsequent CH regions).
2792 // If it comes across a region that suddenly becomes CH, the
2793 // scenario will be similar to b). So, the race between
2794 // claim_region() and a humongous object allocation might force us
2795 // to do a bit of unnecessary work (due to some unnecessary bitmap
2796 // iterations) but it should not introduce and correctness issues.
2797 HeapRegion* curr_region = _g1h->heap_region_containing_raw(finger);
2798 HeapWord* bottom = curr_region->bottom();
2799 HeapWord* end = curr_region->end();
2800 HeapWord* limit = curr_region->next_top_at_mark_start();
2801
2802 if (verbose_low()) {
2803 gclog_or_tty->print_cr("[%u] curr_region = "PTR_FORMAT" "
2804 "["PTR_FORMAT", "PTR_FORMAT"), "
2805 "limit = "PTR_FORMAT,
2806 worker_id, curr_region, bottom, end, limit);
2807 }
2808
2809 // Is the gap between reading the finger and doing the CAS too long?
2810 HeapWord* res = (HeapWord*) Atomic::cmpxchg_ptr(end, &_finger, finger);
2811 if (res == finger) {
2812 // we succeeded
2813
2814 // notice that _finger == end cannot be guaranteed here since,
2815 // someone else might have moved the finger even further
2816 assert(_finger >= end, "the finger should have moved forward");
2817
2818 if (verbose_low()) {
2819 gclog_or_tty->print_cr("[%u] we were successful with region = "
2820 PTR_FORMAT, worker_id, curr_region);
2821 }
2822
2823 if (limit > bottom) {
2824 if (verbose_low()) {
2825 gclog_or_tty->print_cr("[%u] region "PTR_FORMAT" is not empty, "
2826 "returning it ", worker_id, curr_region);
2827 }
2828 return curr_region;
2829 } else {
2830 assert(limit == bottom,
2831 "the region limit should be at bottom");
2832 if (verbose_low()) {
2833 gclog_or_tty->print_cr("[%u] region "PTR_FORMAT" is empty, "
2834 "returning NULL", worker_id, curr_region);
2835 }
2836 // we return NULL and the caller should try calling
2837 // claim_region() again.
2838 return NULL;
2839 }
2840 } else {
2841 assert(_finger > finger, "the finger should have moved forward");
2842 if (verbose_low()) {
2843 gclog_or_tty->print_cr("[%u] somebody else moved the finger, "
2844 "global finger = "PTR_FORMAT", "
2845 "our finger = "PTR_FORMAT,
2846 worker_id, _finger, finger);
2847 }
2848
2849 // read it again
2850 finger = _finger;
2851 }
2852 }
2853
2854 return NULL;
2855 }
2856
2857 #ifndef PRODUCT
2858 enum VerifyNoCSetOopsPhase {
2859 VerifyNoCSetOopsStack,
2860 VerifyNoCSetOopsQueues,
2861 VerifyNoCSetOopsSATBCompleted,
2862 VerifyNoCSetOopsSATBThread
2863 };
2864
2865 class VerifyNoCSetOopsClosure : public OopClosure, public ObjectClosure {
2866 private:
2867 G1CollectedHeap* _g1h;
2868 VerifyNoCSetOopsPhase _phase;
2869 int _info;
2870
2871 const char* phase_str() {
2872 switch (_phase) {
2873 case VerifyNoCSetOopsStack: return "Stack";
2874 case VerifyNoCSetOopsQueues: return "Queue";
2875 case VerifyNoCSetOopsSATBCompleted: return "Completed SATB Buffers";
2876 case VerifyNoCSetOopsSATBThread: return "Thread SATB Buffers";
2877 default: ShouldNotReachHere();
2878 }
2879 return NULL;
2880 }
2881
2882 void do_object_work(oop obj) {
2883 guarantee(!_g1h->obj_in_cs(obj),
2884 err_msg("obj: "PTR_FORMAT" in CSet, phase: %s, info: %d",
2885 (void*) obj, phase_str(), _info));
2886 }
2887
2888 public:
2889 VerifyNoCSetOopsClosure() : _g1h(G1CollectedHeap::heap()) { }
2890
2891 void set_phase(VerifyNoCSetOopsPhase phase, int info = -1) {
2892 _phase = phase;
2893 _info = info;
2894 }
2895
2896 virtual void do_oop(oop* p) {
2897 oop obj = oopDesc::load_decode_heap_oop(p);
2898 do_object_work(obj);
2899 }
2900
2901 virtual void do_oop(narrowOop* p) {
2902 // We should not come across narrow oops while scanning marking
2903 // stacks and SATB buffers.
2904 ShouldNotReachHere();
2905 }
2906
2907 virtual void do_object(oop obj) {
2908 do_object_work(obj);
2909 }
2910 };
2911
2912 void ConcurrentMark::verify_no_cset_oops(bool verify_stacks,
2913 bool verify_enqueued_buffers,
2914 bool verify_thread_buffers,
2915 bool verify_fingers) {
2916 assert(SafepointSynchronize::is_at_safepoint(), "should be at a safepoint");
2917 if (!G1CollectedHeap::heap()->mark_in_progress()) {
2918 return;
2919 }
2920
2921 VerifyNoCSetOopsClosure cl;
2922
2923 if (verify_stacks) {
2924 // Verify entries on the global mark stack
2925 cl.set_phase(VerifyNoCSetOopsStack);
2926 _markStack.oops_do(&cl);
2927
2928 // Verify entries on the task queues
2929 for (uint i = 0; i < _max_worker_id; i += 1) {
2930 cl.set_phase(VerifyNoCSetOopsQueues, i);
2931 CMTaskQueue* queue = _task_queues->queue(i);
2932 queue->oops_do(&cl);
2933 }
2934 }
2935
2936 SATBMarkQueueSet& satb_qs = JavaThread::satb_mark_queue_set();
2937
2938 // Verify entries on the enqueued SATB buffers
2939 if (verify_enqueued_buffers) {
2940 cl.set_phase(VerifyNoCSetOopsSATBCompleted);
2941 satb_qs.iterate_completed_buffers_read_only(&cl);
2942 }
2943
2944 // Verify entries on the per-thread SATB buffers
2945 if (verify_thread_buffers) {
2946 cl.set_phase(VerifyNoCSetOopsSATBThread);
2947 satb_qs.iterate_thread_buffers_read_only(&cl);
2948 }
2949
2950 if (verify_fingers) {
2951 // Verify the global finger
2952 HeapWord* global_finger = finger();
2953 if (global_finger != NULL && global_finger < _heap_end) {
2954 // The global finger always points to a heap region boundary. We
2955 // use heap_region_containing_raw() to get the containing region
2956 // given that the global finger could be pointing to a free region
2957 // which subsequently becomes continues humongous. If that
2958 // happens, heap_region_containing() will return the bottom of the
2959 // corresponding starts humongous region and the check below will
2960 // not hold any more.
2961 HeapRegion* global_hr = _g1h->heap_region_containing_raw(global_finger);
2962 guarantee(global_finger == global_hr->bottom(),
2963 err_msg("global finger: "PTR_FORMAT" region: "HR_FORMAT,
2964 global_finger, HR_FORMAT_PARAMS(global_hr)));
2965 }
2966
2967 // Verify the task fingers
2968 assert(parallel_marking_threads() <= _max_worker_id, "sanity");
2969 for (int i = 0; i < (int) parallel_marking_threads(); i += 1) {
2970 CMTask* task = _tasks[i];
2971 HeapWord* task_finger = task->finger();
2972 if (task_finger != NULL && task_finger < _heap_end) {
2973 // See above note on the global finger verification.
2974 HeapRegion* task_hr = _g1h->heap_region_containing_raw(task_finger);
2975 guarantee(task_finger == task_hr->bottom() ||
2976 !task_hr->in_collection_set(),
2977 err_msg("task finger: "PTR_FORMAT" region: "HR_FORMAT,
2978 task_finger, HR_FORMAT_PARAMS(task_hr)));
2979 }
2980 }
2981 }
2982 }
2983 #endif // PRODUCT
2984
2985 // Aggregate the counting data that was constructed concurrently
2986 // with marking.
2987 class AggregateCountDataHRClosure: public HeapRegionClosure {
2988 G1CollectedHeap* _g1h;
2989 ConcurrentMark* _cm;
2990 CardTableModRefBS* _ct_bs;
2991 BitMap* _cm_card_bm;
2992 uint _max_worker_id;
2993
2994 public:
2995 AggregateCountDataHRClosure(G1CollectedHeap* g1h,
2996 BitMap* cm_card_bm,
2997 uint max_worker_id) :
2998 _g1h(g1h), _cm(g1h->concurrent_mark()),
2999 _ct_bs((CardTableModRefBS*) (g1h->barrier_set())),
3000 _cm_card_bm(cm_card_bm), _max_worker_id(max_worker_id) { }
3001
3002 bool doHeapRegion(HeapRegion* hr) {
3003 if (hr->continuesHumongous()) {
3004 // We will ignore these here and process them when their
3005 // associated "starts humongous" region is processed.
3006 // Note that we cannot rely on their associated
3007 // "starts humongous" region to have their bit set to 1
3008 // since, due to the region chunking in the parallel region
3009 // iteration, a "continues humongous" region might be visited
3010 // before its associated "starts humongous".
3011 return false;
3012 }
3013
3014 HeapWord* start = hr->bottom();
3015 HeapWord* limit = hr->next_top_at_mark_start();
3016 HeapWord* end = hr->end();
3017
3018 assert(start <= limit && limit <= hr->top() && hr->top() <= hr->end(),
3019 err_msg("Preconditions not met - "
3020 "start: "PTR_FORMAT", limit: "PTR_FORMAT", "
3021 "top: "PTR_FORMAT", end: "PTR_FORMAT,
3022 start, limit, hr->top(), hr->end()));
3023
3024 assert(hr->next_marked_bytes() == 0, "Precondition");
3025
3026 if (start == limit) {
3027 // NTAMS of this region has not been set so nothing to do.
3028 return false;
3029 }
3030
3031 // 'start' should be in the heap.
3032 assert(_g1h->is_in_g1_reserved(start) && _ct_bs->is_card_aligned(start), "sanity");
3033 // 'end' *may* be just beyone the end of the heap (if hr is the last region)
3034 assert(!_g1h->is_in_g1_reserved(end) || _ct_bs->is_card_aligned(end), "sanity");
3035
3036 BitMap::idx_t start_idx = _cm->card_bitmap_index_for(start);
3037 BitMap::idx_t limit_idx = _cm->card_bitmap_index_for(limit);
3038 BitMap::idx_t end_idx = _cm->card_bitmap_index_for(end);
3039
3040 // If ntams is not card aligned then we bump card bitmap index
3041 // for limit so that we get the all the cards spanned by
3042 // the object ending at ntams.
3043 // Note: if this is the last region in the heap then ntams
3044 // could be actually just beyond the end of the the heap;
3045 // limit_idx will then correspond to a (non-existent) card
3046 // that is also outside the heap.
3047 if (_g1h->is_in_g1_reserved(limit) && !_ct_bs->is_card_aligned(limit)) {
3048 limit_idx += 1;
3049 }
3050
3051 assert(limit_idx <= end_idx, "or else use atomics");
3052
3053 // Aggregate the "stripe" in the count data associated with hr.
3054 uint hrs_index = hr->hrs_index();
3055 size_t marked_bytes = 0;
3056
3057 for (uint i = 0; i < _max_worker_id; i += 1) {
3058 size_t* marked_bytes_array = _cm->count_marked_bytes_array_for(i);
3059 BitMap* task_card_bm = _cm->count_card_bitmap_for(i);
3060
3061 // Fetch the marked_bytes in this region for task i and
3062 // add it to the running total for this region.
3063 marked_bytes += marked_bytes_array[hrs_index];
3064
3065 // Now union the bitmaps[0,max_worker_id)[start_idx..limit_idx)
3066 // into the global card bitmap.
3067 BitMap::idx_t scan_idx = task_card_bm->get_next_one_offset(start_idx, limit_idx);
3068
3069 while (scan_idx < limit_idx) {
3070 assert(task_card_bm->at(scan_idx) == true, "should be");
3071 _cm_card_bm->set_bit(scan_idx);
3072 assert(_cm_card_bm->at(scan_idx) == true, "should be");
3073
3074 // BitMap::get_next_one_offset() can handle the case when
3075 // its left_offset parameter is greater than its right_offset
3076 // parameter. It does, however, have an early exit if
3077 // left_offset == right_offset. So let's limit the value
3078 // passed in for left offset here.
3079 BitMap::idx_t next_idx = MIN2(scan_idx + 1, limit_idx);
3080 scan_idx = task_card_bm->get_next_one_offset(next_idx, limit_idx);
3081 }
3082 }
3083
3084 // Update the marked bytes for this region.
3085 hr->add_to_marked_bytes(marked_bytes);
3086
3087 // Next heap region
3088 return false;
3089 }
3090 };
3091
3092 class G1AggregateCountDataTask: public AbstractGangTask {
3093 protected:
3094 G1CollectedHeap* _g1h;
3095 ConcurrentMark* _cm;
3096 BitMap* _cm_card_bm;
3097 uint _max_worker_id;
3098 int _active_workers;
3099
3100 public:
3101 G1AggregateCountDataTask(G1CollectedHeap* g1h,
3102 ConcurrentMark* cm,
3103 BitMap* cm_card_bm,
3104 uint max_worker_id,
3105 int n_workers) :
3106 AbstractGangTask("Count Aggregation"),
3107 _g1h(g1h), _cm(cm), _cm_card_bm(cm_card_bm),
3108 _max_worker_id(max_worker_id),
3109 _active_workers(n_workers) { }
3110
3111 void work(uint worker_id) {
3112 AggregateCountDataHRClosure cl(_g1h, _cm_card_bm, _max_worker_id);
3113
3114 if (G1CollectedHeap::use_parallel_gc_threads()) {
3115 _g1h->heap_region_par_iterate_chunked(&cl, worker_id,
3116 _active_workers,
3117 HeapRegion::AggregateCountClaimValue);
3118 } else {
3119 _g1h->heap_region_iterate(&cl);
3120 }
3121 }
3122 };
3123
3124
3125 void ConcurrentMark::aggregate_count_data() {
3126 int n_workers = (G1CollectedHeap::use_parallel_gc_threads() ?
3127 _g1h->workers()->active_workers() :
3128 1);
3129
3130 G1AggregateCountDataTask g1_par_agg_task(_g1h, this, &_card_bm,
3131 _max_worker_id, n_workers);
3132
3133 if (G1CollectedHeap::use_parallel_gc_threads()) {
3134 assert(_g1h->check_heap_region_claim_values(HeapRegion::InitialClaimValue),
3135 "sanity check");
3136 _g1h->set_par_threads(n_workers);
3137 _g1h->workers()->run_task(&g1_par_agg_task);
3138 _g1h->set_par_threads(0);
3139
3140 assert(_g1h->check_heap_region_claim_values(HeapRegion::AggregateCountClaimValue),
3141 "sanity check");
3142 _g1h->reset_heap_region_claim_values();
3143 } else {
3144 g1_par_agg_task.work(0);
3145 }
3146 }
3147
3148 // Clear the per-worker arrays used to store the per-region counting data
3149 void ConcurrentMark::clear_all_count_data() {
3150 // Clear the global card bitmap - it will be filled during
3151 // liveness count aggregation (during remark) and the
3152 // final counting task.
3153 _card_bm.clear();
3154
3155 // Clear the global region bitmap - it will be filled as part
3156 // of the final counting task.
3157 _region_bm.clear();
3158
3159 uint max_regions = _g1h->max_regions();
3160 assert(_max_worker_id > 0, "uninitialized");
3161
3162 for (uint i = 0; i < _max_worker_id; i += 1) {
3163 BitMap* task_card_bm = count_card_bitmap_for(i);
3164 size_t* marked_bytes_array = count_marked_bytes_array_for(i);
3165
3166 assert(task_card_bm->size() == _card_bm.size(), "size mismatch");
3167 assert(marked_bytes_array != NULL, "uninitialized");
3168
3169 memset(marked_bytes_array, 0, (size_t) max_regions * sizeof(size_t));
3170 task_card_bm->clear();
3171 }
3172 }
3173
3174 void ConcurrentMark::print_stats() {
3175 if (verbose_stats()) {
3176 gclog_or_tty->print_cr("---------------------------------------------------------------------");
3177 for (size_t i = 0; i < _active_tasks; ++i) {
3178 _tasks[i]->print_stats();
3179 gclog_or_tty->print_cr("---------------------------------------------------------------------");
3180 }
3181 }
3182 }
3183
3184 // abandon current marking iteration due to a Full GC
3185 void ConcurrentMark::abort() {
3186 // Clear all marks to force marking thread to do nothing
3187 _nextMarkBitMap->clearAll();
3188 // Clear the liveness counting data
3189 clear_all_count_data();
3190 // Empty mark stack
3191 reset_marking_state();
3192 for (uint i = 0; i < _max_worker_id; ++i) {
3193 _tasks[i]->clear_region_fields();
3194 }
3195 _has_aborted = true;
3196
3197 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set();
3198 satb_mq_set.abandon_partial_marking();
3199 // This can be called either during or outside marking, we'll read
3200 // the expected_active value from the SATB queue set.
3201 satb_mq_set.set_active_all_threads(
3202 false, /* new active value */
3203 satb_mq_set.is_active() /* expected_active */);
3204 }
3205
3206 static void print_ms_time_info(const char* prefix, const char* name,
3207 NumberSeq& ns) {
3208 gclog_or_tty->print_cr("%s%5d %12s: total time = %8.2f s (avg = %8.2f ms).",
3209 prefix, ns.num(), name, ns.sum()/1000.0, ns.avg());
3210 if (ns.num() > 0) {
3211 gclog_or_tty->print_cr("%s [std. dev = %8.2f ms, max = %8.2f ms]",
3212 prefix, ns.sd(), ns.maximum());
3213 }
3214 }
3215
3216 void ConcurrentMark::print_summary_info() {
3217 gclog_or_tty->print_cr(" Concurrent marking:");
3218 print_ms_time_info(" ", "init marks", _init_times);
3219 print_ms_time_info(" ", "remarks", _remark_times);
3220 {
3221 print_ms_time_info(" ", "final marks", _remark_mark_times);
3222 print_ms_time_info(" ", "weak refs", _remark_weak_ref_times);
3223
3224 }
3225 print_ms_time_info(" ", "cleanups", _cleanup_times);
3226 gclog_or_tty->print_cr(" Final counting total time = %8.2f s (avg = %8.2f ms).",
3227 _total_counting_time,
3228 (_cleanup_times.num() > 0 ? _total_counting_time * 1000.0 /
3229 (double)_cleanup_times.num()
3230 : 0.0));
3231 if (G1ScrubRemSets) {
3232 gclog_or_tty->print_cr(" RS scrub total time = %8.2f s (avg = %8.2f ms).",
3233 _total_rs_scrub_time,
3234 (_cleanup_times.num() > 0 ? _total_rs_scrub_time * 1000.0 /
3235 (double)_cleanup_times.num()
3236 : 0.0));
3237 }
3238 gclog_or_tty->print_cr(" Total stop_world time = %8.2f s.",
3239 (_init_times.sum() + _remark_times.sum() +
3240 _cleanup_times.sum())/1000.0);
3241 gclog_or_tty->print_cr(" Total concurrent time = %8.2f s "
3242 "(%8.2f s marking).",
3243 cmThread()->vtime_accum(),
3244 cmThread()->vtime_mark_accum());
3245 }
3246
3247 void ConcurrentMark::print_worker_threads_on(outputStream* st) const {
3248 if (use_parallel_marking_threads()) {
3249 _parallel_workers->print_worker_threads_on(st);
3250 }
3251 }
3252
3253 // We take a break if someone is trying to stop the world.
3254 bool ConcurrentMark::do_yield_check(uint worker_id) {
3255 if (should_yield()) {
3256 if (worker_id == 0) {
3257 _g1h->g1_policy()->record_concurrent_pause();
3258 }
3259 cmThread()->yield();
3260 return true;
3261 } else {
3262 return false;
3263 }
3264 }
3265
3266 bool ConcurrentMark::should_yield() {
3267 return cmThread()->should_yield();
3268 }
3269
3270 bool ConcurrentMark::containing_card_is_marked(void* p) {
3271 size_t offset = pointer_delta(p, _g1h->reserved_region().start(), 1);
3272 return _card_bm.at(offset >> CardTableModRefBS::card_shift);
3273 }
3274
3275 bool ConcurrentMark::containing_cards_are_marked(void* start,
3276 void* last) {
3277 return containing_card_is_marked(start) &&
3278 containing_card_is_marked(last);
3279 }
3280
3281 #ifndef PRODUCT
3282 // for debugging purposes
3283 void ConcurrentMark::print_finger() {
3284 gclog_or_tty->print_cr("heap ["PTR_FORMAT", "PTR_FORMAT"), global finger = "PTR_FORMAT,
3285 _heap_start, _heap_end, _finger);
3286 for (uint i = 0; i < _max_worker_id; ++i) {
3287 gclog_or_tty->print(" %u: "PTR_FORMAT, i, _tasks[i]->finger());
3288 }
3289 gclog_or_tty->print_cr("");
3290 }
3291 #endif
3292
3293 void CMTask::scan_object(oop obj) {
3294 assert(_nextMarkBitMap->isMarked((HeapWord*) obj), "invariant");
3295
3296 if (_cm->verbose_high()) {
3297 gclog_or_tty->print_cr("[%u] we're scanning object "PTR_FORMAT,
3298 _worker_id, (void*) obj);
3299 }
3300
3301 size_t obj_size = obj->size();
3302 _words_scanned += obj_size;
3303
3304 obj->oop_iterate(_cm_oop_closure);
3305 statsOnly( ++_objs_scanned );
3306 check_limits();
3307 }
3308
3309 // Closure for iteration over bitmaps
3310 class CMBitMapClosure : public BitMapClosure {
3311 private:
3312 // the bitmap that is being iterated over
3313 CMBitMap* _nextMarkBitMap;
3314 ConcurrentMark* _cm;
3315 CMTask* _task;
3316
3317 public:
3318 CMBitMapClosure(CMTask *task, ConcurrentMark* cm, CMBitMap* nextMarkBitMap) :
3319 _task(task), _cm(cm), _nextMarkBitMap(nextMarkBitMap) { }
3320
3321 bool do_bit(size_t offset) {
3322 HeapWord* addr = _nextMarkBitMap->offsetToHeapWord(offset);
3323 assert(_nextMarkBitMap->isMarked(addr), "invariant");
3324 assert( addr < _cm->finger(), "invariant");
3325
3326 statsOnly( _task->increase_objs_found_on_bitmap() );
3327 assert(addr >= _task->finger(), "invariant");
3328
3329 // We move that task's local finger along.
3330 _task->move_finger_to(addr);
3331
3332 _task->scan_object(oop(addr));
3333 // we only partially drain the local queue and global stack
3334 _task->drain_local_queue(true);
3335 _task->drain_global_stack(true);
3336
3337 // if the has_aborted flag has been raised, we need to bail out of
3338 // the iteration
3339 return !_task->has_aborted();
3340 }
3341 };
3342
3343 // Closure for iterating over objects, currently only used for
3344 // processing SATB buffers.
3345 class CMObjectClosure : public ObjectClosure {
3346 private:
3347 CMTask* _task;
3348
3349 public:
3350 void do_object(oop obj) {
3351 _task->deal_with_reference(obj);
3352 }
3353
3354 CMObjectClosure(CMTask* task) : _task(task) { }
3355 };
3356
3357 G1CMOopClosure::G1CMOopClosure(G1CollectedHeap* g1h,
3358 ConcurrentMark* cm,
3359 CMTask* task)
3360 : _g1h(g1h), _cm(cm), _task(task) {
3361 assert(_ref_processor == NULL, "should be initialized to NULL");
3362
3363 if (G1UseConcMarkReferenceProcessing) {
3364 _ref_processor = g1h->ref_processor_cm();
3365 assert(_ref_processor != NULL, "should not be NULL");
3366 }
3367 }
3368
3369 void CMTask::setup_for_region(HeapRegion* hr) {
3370 // Separated the asserts so that we know which one fires.
3371 assert(hr != NULL,
3372 "claim_region() should have filtered out continues humongous regions");
3373 assert(!hr->continuesHumongous(),
3374 "claim_region() should have filtered out continues humongous regions");
3375
3376 if (_cm->verbose_low()) {
3377 gclog_or_tty->print_cr("[%u] setting up for region "PTR_FORMAT,
3378 _worker_id, hr);
3379 }
3380
3381 _curr_region = hr;
3382 _finger = hr->bottom();
3383 update_region_limit();
3384 }
3385
3386 void CMTask::update_region_limit() {
3387 HeapRegion* hr = _curr_region;
3388 HeapWord* bottom = hr->bottom();
3389 HeapWord* limit = hr->next_top_at_mark_start();
3390
3391 if (limit == bottom) {
3392 if (_cm->verbose_low()) {
3393 gclog_or_tty->print_cr("[%u] found an empty region "
3394 "["PTR_FORMAT", "PTR_FORMAT")",
3395 _worker_id, bottom, limit);
3396 }
3397 // The region was collected underneath our feet.
3398 // We set the finger to bottom to ensure that the bitmap
3399 // iteration that will follow this will not do anything.
3400 // (this is not a condition that holds when we set the region up,
3401 // as the region is not supposed to be empty in the first place)
3402 _finger = bottom;
3403 } else if (limit >= _region_limit) {
3404 assert(limit >= _finger, "peace of mind");
3405 } else {
3406 assert(limit < _region_limit, "only way to get here");
3407 // This can happen under some pretty unusual circumstances. An
3408 // evacuation pause empties the region underneath our feet (NTAMS
3409 // at bottom). We then do some allocation in the region (NTAMS
3410 // stays at bottom), followed by the region being used as a GC
3411 // alloc region (NTAMS will move to top() and the objects
3412 // originally below it will be grayed). All objects now marked in
3413 // the region are explicitly grayed, if below the global finger,
3414 // and we do not need in fact to scan anything else. So, we simply
3415 // set _finger to be limit to ensure that the bitmap iteration
3416 // doesn't do anything.
3417 _finger = limit;
3418 }
3419
3420 _region_limit = limit;
3421 }
3422
3423 void CMTask::giveup_current_region() {
3424 assert(_curr_region != NULL, "invariant");
3425 if (_cm->verbose_low()) {
3426 gclog_or_tty->print_cr("[%u] giving up region "PTR_FORMAT,
3427 _worker_id, _curr_region);
3428 }
3429 clear_region_fields();
3430 }
3431
3432 void CMTask::clear_region_fields() {
3433 // Values for these three fields that indicate that we're not
3434 // holding on to a region.
3435 _curr_region = NULL;
3436 _finger = NULL;
3437 _region_limit = NULL;
3438 }
3439
3440 void CMTask::set_cm_oop_closure(G1CMOopClosure* cm_oop_closure) {
3441 if (cm_oop_closure == NULL) {
3442 assert(_cm_oop_closure != NULL, "invariant");
3443 } else {
3444 assert(_cm_oop_closure == NULL, "invariant");
3445 }
3446 _cm_oop_closure = cm_oop_closure;
3447 }
3448
3449 void CMTask::reset(CMBitMap* nextMarkBitMap) {
3450 guarantee(nextMarkBitMap != NULL, "invariant");
3451
3452 if (_cm->verbose_low()) {
3453 gclog_or_tty->print_cr("[%u] resetting", _worker_id);
3454 }
3455
3456 _nextMarkBitMap = nextMarkBitMap;
3457 clear_region_fields();
3458
3459 _calls = 0;
3460 _elapsed_time_ms = 0.0;
3461 _termination_time_ms = 0.0;
3462 _termination_start_time_ms = 0.0;
3463
3464 #if _MARKING_STATS_
3465 _local_pushes = 0;
3466 _local_pops = 0;
3467 _local_max_size = 0;
3468 _objs_scanned = 0;
3469 _global_pushes = 0;
3470 _global_pops = 0;
3471 _global_max_size = 0;
3472 _global_transfers_to = 0;
3473 _global_transfers_from = 0;
3474 _regions_claimed = 0;
3475 _objs_found_on_bitmap = 0;
3476 _satb_buffers_processed = 0;
3477 _steal_attempts = 0;
3478 _steals = 0;
3479 _aborted = 0;
3480 _aborted_overflow = 0;
3481 _aborted_cm_aborted = 0;
3482 _aborted_yield = 0;
3483 _aborted_timed_out = 0;
3484 _aborted_satb = 0;
3485 _aborted_termination = 0;
3486 #endif // _MARKING_STATS_
3487 }
3488
3489 bool CMTask::should_exit_termination() {
3490 regular_clock_call();
3491 // This is called when we are in the termination protocol. We should
3492 // quit if, for some reason, this task wants to abort or the global
3493 // stack is not empty (this means that we can get work from it).
3494 return !_cm->mark_stack_empty() || has_aborted();
3495 }
3496
3497 void CMTask::reached_limit() {
3498 assert(_words_scanned >= _words_scanned_limit ||
3499 _refs_reached >= _refs_reached_limit ,
3500 "shouldn't have been called otherwise");
3501 regular_clock_call();
3502 }
3503
3504 void CMTask::regular_clock_call() {
3505 if (has_aborted()) return;
3506
3507 // First, we need to recalculate the words scanned and refs reached
3508 // limits for the next clock call.
3509 recalculate_limits();
3510
3511 // During the regular clock call we do the following
3512
3513 // (1) If an overflow has been flagged, then we abort.
3514 if (_cm->has_overflown()) {
3515 set_has_aborted();
3516 return;
3517 }
3518
3519 // If we are not concurrent (i.e. we're doing remark) we don't need
3520 // to check anything else. The other steps are only needed during
3521 // the concurrent marking phase.
3522 if (!concurrent()) return;
3523
3524 // (2) If marking has been aborted for Full GC, then we also abort.
3525 if (_cm->has_aborted()) {
3526 set_has_aborted();
3527 statsOnly( ++_aborted_cm_aborted );
3528 return;
3529 }
3530
3531 double curr_time_ms = os::elapsedVTime() * 1000.0;
3532
3533 // (3) If marking stats are enabled, then we update the step history.
3534 #if _MARKING_STATS_
3535 if (_words_scanned >= _words_scanned_limit) {
3536 ++_clock_due_to_scanning;
3537 }
3538 if (_refs_reached >= _refs_reached_limit) {
3539 ++_clock_due_to_marking;
3540 }
3541
3542 double last_interval_ms = curr_time_ms - _interval_start_time_ms;
3543 _interval_start_time_ms = curr_time_ms;
3544 _all_clock_intervals_ms.add(last_interval_ms);
3545
3546 if (_cm->verbose_medium()) {
3547 gclog_or_tty->print_cr("[%u] regular clock, interval = %1.2lfms, "
3548 "scanned = %d%s, refs reached = %d%s",
3549 _worker_id, last_interval_ms,
3550 _words_scanned,
3551 (_words_scanned >= _words_scanned_limit) ? " (*)" : "",
3552 _refs_reached,
3553 (_refs_reached >= _refs_reached_limit) ? " (*)" : "");
3554 }
3555 #endif // _MARKING_STATS_
3556
3557 // (4) We check whether we should yield. If we have to, then we abort.
3558 if (_cm->should_yield()) {
3559 // We should yield. To do this we abort the task. The caller is
3560 // responsible for yielding.
3561 set_has_aborted();
3562 statsOnly( ++_aborted_yield );
3563 return;
3564 }
3565
3566 // (5) We check whether we've reached our time quota. If we have,
3567 // then we abort.
3568 double elapsed_time_ms = curr_time_ms - _start_time_ms;
3569 if (elapsed_time_ms > _time_target_ms) {
3570 set_has_aborted();
3571 _has_timed_out = true;
3572 statsOnly( ++_aborted_timed_out );
3573 return;
3574 }
3575
3576 // (6) Finally, we check whether there are enough completed STAB
3577 // buffers available for processing. If there are, we abort.
3578 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set();
3579 if (!_draining_satb_buffers && satb_mq_set.process_completed_buffers()) {
3580 if (_cm->verbose_low()) {
3581 gclog_or_tty->print_cr("[%u] aborting to deal with pending SATB buffers",
3582 _worker_id);
3583 }
3584 // we do need to process SATB buffers, we'll abort and restart
3585 // the marking task to do so
3586 set_has_aborted();
3587 statsOnly( ++_aborted_satb );
3588 return;
3589 }
3590 }
3591
3592 void CMTask::recalculate_limits() {
3593 _real_words_scanned_limit = _words_scanned + words_scanned_period;
3594 _words_scanned_limit = _real_words_scanned_limit;
3595
3596 _real_refs_reached_limit = _refs_reached + refs_reached_period;
3597 _refs_reached_limit = _real_refs_reached_limit;
3598 }
3599
3600 void CMTask::decrease_limits() {
3601 // This is called when we believe that we're going to do an infrequent
3602 // operation which will increase the per byte scanned cost (i.e. move
3603 // entries to/from the global stack). It basically tries to decrease the
3604 // scanning limit so that the clock is called earlier.
3605
3606 if (_cm->verbose_medium()) {
3607 gclog_or_tty->print_cr("[%u] decreasing limits", _worker_id);
3608 }
3609
3610 _words_scanned_limit = _real_words_scanned_limit -
3611 3 * words_scanned_period / 4;
3612 _refs_reached_limit = _real_refs_reached_limit -
3613 3 * refs_reached_period / 4;
3614 }
3615
3616 void CMTask::move_entries_to_global_stack() {
3617 // local array where we'll store the entries that will be popped
3618 // from the local queue
3619 oop buffer[global_stack_transfer_size];
3620
3621 int n = 0;
3622 oop obj;
3623 while (n < global_stack_transfer_size && _task_queue->pop_local(obj)) {
3624 buffer[n] = obj;
3625 ++n;
3626 }
3627
3628 if (n > 0) {
3629 // we popped at least one entry from the local queue
3630
3631 statsOnly( ++_global_transfers_to; _local_pops += n );
3632
3633 if (!_cm->mark_stack_push(buffer, n)) {
3634 if (_cm->verbose_low()) {
3635 gclog_or_tty->print_cr("[%u] aborting due to global stack overflow",
3636 _worker_id);
3637 }
3638 set_has_aborted();
3639 } else {
3640 // the transfer was successful
3641
3642 if (_cm->verbose_medium()) {
3643 gclog_or_tty->print_cr("[%u] pushed %d entries to the global stack",
3644 _worker_id, n);
3645 }
3646 statsOnly( int tmp_size = _cm->mark_stack_size();
3647 if (tmp_size > _global_max_size) {
3648 _global_max_size = tmp_size;
3649 }
3650 _global_pushes += n );
3651 }
3652 }
3653
3654 // this operation was quite expensive, so decrease the limits
3655 decrease_limits();
3656 }
3657
3658 void CMTask::get_entries_from_global_stack() {
3659 // local array where we'll store the entries that will be popped
3660 // from the global stack.
3661 oop buffer[global_stack_transfer_size];
3662 int n;
3663 _cm->mark_stack_pop(buffer, global_stack_transfer_size, &n);
3664 assert(n <= global_stack_transfer_size,
3665 "we should not pop more than the given limit");
3666 if (n > 0) {
3667 // yes, we did actually pop at least one entry
3668
3669 statsOnly( ++_global_transfers_from; _global_pops += n );
3670 if (_cm->verbose_medium()) {
3671 gclog_or_tty->print_cr("[%u] popped %d entries from the global stack",
3672 _worker_id, n);
3673 }
3674 for (int i = 0; i < n; ++i) {
3675 bool success = _task_queue->push(buffer[i]);
3676 // We only call this when the local queue is empty or under a
3677 // given target limit. So, we do not expect this push to fail.
3678 assert(success, "invariant");
3679 }
3680
3681 statsOnly( int tmp_size = _task_queue->size();
3682 if (tmp_size > _local_max_size) {
3683 _local_max_size = tmp_size;
3684 }
3685 _local_pushes += n );
3686 }
3687
3688 // this operation was quite expensive, so decrease the limits
3689 decrease_limits();
3690 }
3691
3692 void CMTask::drain_local_queue(bool partially) {
3693 if (has_aborted()) return;
3694
3695 // Decide what the target size is, depending whether we're going to
3696 // drain it partially (so that other tasks can steal if they run out
3697 // of things to do) or totally (at the very end).
3698 size_t target_size;
3699 if (partially) {
3700 target_size = MIN2((size_t)_task_queue->max_elems()/3, GCDrainStackTargetSize);
3701 } else {
3702 target_size = 0;
3703 }
3704
3705 if (_task_queue->size() > target_size) {
3706 if (_cm->verbose_high()) {
3707 gclog_or_tty->print_cr("[%u] draining local queue, target size = %d",
3708 _worker_id, target_size);
3709 }
3710
3711 oop obj;
3712 bool ret = _task_queue->pop_local(obj);
3713 while (ret) {
3714 statsOnly( ++_local_pops );
3715
3716 if (_cm->verbose_high()) {
3717 gclog_or_tty->print_cr("[%u] popped "PTR_FORMAT, _worker_id,
3718 (void*) obj);
3719 }
3720
3721 assert(_g1h->is_in_g1_reserved((HeapWord*) obj), "invariant" );
3722 assert(!_g1h->is_on_master_free_list(
3723 _g1h->heap_region_containing((HeapWord*) obj)), "invariant");
3724
3725 scan_object(obj);
3726
3727 if (_task_queue->size() <= target_size || has_aborted()) {
3728 ret = false;
3729 } else {
3730 ret = _task_queue->pop_local(obj);
3731 }
3732 }
3733
3734 if (_cm->verbose_high()) {
3735 gclog_or_tty->print_cr("[%u] drained local queue, size = %d",
3736 _worker_id, _task_queue->size());
3737 }
3738 }
3739 }
3740
3741 void CMTask::drain_global_stack(bool partially) {
3742 if (has_aborted()) return;
3743
3744 // We have a policy to drain the local queue before we attempt to
3745 // drain the global stack.
3746 assert(partially || _task_queue->size() == 0, "invariant");
3747
3748 // Decide what the target size is, depending whether we're going to
3749 // drain it partially (so that other tasks can steal if they run out
3750 // of things to do) or totally (at the very end). Notice that,
3751 // because we move entries from the global stack in chunks or
3752 // because another task might be doing the same, we might in fact
3753 // drop below the target. But, this is not a problem.
3754 size_t target_size;
3755 if (partially) {
3756 target_size = _cm->partial_mark_stack_size_target();
3757 } else {
3758 target_size = 0;
3759 }
3760
3761 if (_cm->mark_stack_size() > target_size) {
3762 if (_cm->verbose_low()) {
3763 gclog_or_tty->print_cr("[%u] draining global_stack, target size %d",
3764 _worker_id, target_size);
3765 }
3766
3767 while (!has_aborted() && _cm->mark_stack_size() > target_size) {
3768 get_entries_from_global_stack();
3769 drain_local_queue(partially);
3770 }
3771
3772 if (_cm->verbose_low()) {
3773 gclog_or_tty->print_cr("[%u] drained global stack, size = %d",
3774 _worker_id, _cm->mark_stack_size());
3775 }
3776 }
3777 }
3778
3779 // SATB Queue has several assumptions on whether to call the par or
3780 // non-par versions of the methods. this is why some of the code is
3781 // replicated. We should really get rid of the single-threaded version
3782 // of the code to simplify things.
3783 void CMTask::drain_satb_buffers() {
3784 if (has_aborted()) return;
3785
3786 // We set this so that the regular clock knows that we're in the
3787 // middle of draining buffers and doesn't set the abort flag when it
3788 // notices that SATB buffers are available for draining. It'd be
3789 // very counter productive if it did that. :-)
3790 _draining_satb_buffers = true;
3791
3792 CMObjectClosure oc(this);
3793 SATBMarkQueueSet& satb_mq_set = JavaThread::satb_mark_queue_set();
3794 if (G1CollectedHeap::use_parallel_gc_threads()) {
3795 satb_mq_set.set_par_closure(_worker_id, &oc);
3796 } else {
3797 satb_mq_set.set_closure(&oc);
3798 }
3799
3800 // This keeps claiming and applying the closure to completed buffers
3801 // until we run out of buffers or we need to abort.
3802 if (G1CollectedHeap::use_parallel_gc_threads()) {
3803 while (!has_aborted() &&
3804 satb_mq_set.par_apply_closure_to_completed_buffer(_worker_id)) {
3805 if (_cm->verbose_medium()) {
3806 gclog_or_tty->print_cr("[%u] processed an SATB buffer", _worker_id);
3807 }
3808 statsOnly( ++_satb_buffers_processed );
3809 regular_clock_call();
3810 }
3811 } else {
3812 while (!has_aborted() &&
3813 satb_mq_set.apply_closure_to_completed_buffer()) {
3814 if (_cm->verbose_medium()) {
3815 gclog_or_tty->print_cr("[%u] processed an SATB buffer", _worker_id);
3816 }
3817 statsOnly( ++_satb_buffers_processed );
3818 regular_clock_call();
3819 }
3820 }
3821
3822 if (!concurrent() && !has_aborted()) {
3823 // We should only do this during remark.
3824 if (G1CollectedHeap::use_parallel_gc_threads()) {
3825 satb_mq_set.par_iterate_closure_all_threads(_worker_id);
3826 } else {
3827 satb_mq_set.iterate_closure_all_threads();
3828 }
3829 }
3830
3831 _draining_satb_buffers = false;
3832
3833 assert(has_aborted() ||
3834 concurrent() ||
3835 satb_mq_set.completed_buffers_num() == 0, "invariant");
3836
3837 if (G1CollectedHeap::use_parallel_gc_threads()) {
3838 satb_mq_set.set_par_closure(_worker_id, NULL);
3839 } else {
3840 satb_mq_set.set_closure(NULL);
3841 }
3842
3843 // again, this was a potentially expensive operation, decrease the
3844 // limits to get the regular clock call early
3845 decrease_limits();
3846 }
3847
3848 void CMTask::print_stats() {
3849 gclog_or_tty->print_cr("Marking Stats, task = %u, calls = %d",
3850 _worker_id, _calls);
3851 gclog_or_tty->print_cr(" Elapsed time = %1.2lfms, Termination time = %1.2lfms",
3852 _elapsed_time_ms, _termination_time_ms);
3853 gclog_or_tty->print_cr(" Step Times (cum): num = %d, avg = %1.2lfms, sd = %1.2lfms",
3854 _step_times_ms.num(), _step_times_ms.avg(),
3855 _step_times_ms.sd());
3856 gclog_or_tty->print_cr(" max = %1.2lfms, total = %1.2lfms",
3857 _step_times_ms.maximum(), _step_times_ms.sum());
3858
3859 #if _MARKING_STATS_
3860 gclog_or_tty->print_cr(" Clock Intervals (cum): num = %d, avg = %1.2lfms, sd = %1.2lfms",
3861 _all_clock_intervals_ms.num(), _all_clock_intervals_ms.avg(),
3862 _all_clock_intervals_ms.sd());
3863 gclog_or_tty->print_cr(" max = %1.2lfms, total = %1.2lfms",
3864 _all_clock_intervals_ms.maximum(),
3865 _all_clock_intervals_ms.sum());
3866 gclog_or_tty->print_cr(" Clock Causes (cum): scanning = %d, marking = %d",
3867 _clock_due_to_scanning, _clock_due_to_marking);
3868 gclog_or_tty->print_cr(" Objects: scanned = %d, found on the bitmap = %d",
3869 _objs_scanned, _objs_found_on_bitmap);
3870 gclog_or_tty->print_cr(" Local Queue: pushes = %d, pops = %d, max size = %d",
3871 _local_pushes, _local_pops, _local_max_size);
3872 gclog_or_tty->print_cr(" Global Stack: pushes = %d, pops = %d, max size = %d",
3873 _global_pushes, _global_pops, _global_max_size);
3874 gclog_or_tty->print_cr(" transfers to = %d, transfers from = %d",
3875 _global_transfers_to,_global_transfers_from);
3876 gclog_or_tty->print_cr(" Regions: claimed = %d", _regions_claimed);
3877 gclog_or_tty->print_cr(" SATB buffers: processed = %d", _satb_buffers_processed);
3878 gclog_or_tty->print_cr(" Steals: attempts = %d, successes = %d",
3879 _steal_attempts, _steals);
3880 gclog_or_tty->print_cr(" Aborted: %d, due to", _aborted);
3881 gclog_or_tty->print_cr(" overflow: %d, global abort: %d, yield: %d",
3882 _aborted_overflow, _aborted_cm_aborted, _aborted_yield);
3883 gclog_or_tty->print_cr(" time out: %d, SATB: %d, termination: %d",
3884 _aborted_timed_out, _aborted_satb, _aborted_termination);
3885 #endif // _MARKING_STATS_
3886 }
3887
3888 /*****************************************************************************
3889
3890 The do_marking_step(time_target_ms, ...) method is the building
3891 block of the parallel marking framework. It can be called in parallel
3892 with other invocations of do_marking_step() on different tasks
3893 (but only one per task, obviously) and concurrently with the
3894 mutator threads, or during remark, hence it eliminates the need
3895 for two versions of the code. When called during remark, it will
3896 pick up from where the task left off during the concurrent marking
3897 phase. Interestingly, tasks are also claimable during evacuation
3898 pauses too, since do_marking_step() ensures that it aborts before
3899 it needs to yield.
3900
3901 The data structures that it uses to do marking work are the
3902 following:
3903
3904 (1) Marking Bitmap. If there are gray objects that appear only
3905 on the bitmap (this happens either when dealing with an overflow
3906 or when the initial marking phase has simply marked the roots
3907 and didn't push them on the stack), then tasks claim heap
3908 regions whose bitmap they then scan to find gray objects. A
3909 global finger indicates where the end of the last claimed region
3910 is. A local finger indicates how far into the region a task has
3911 scanned. The two fingers are used to determine how to gray an
3912 object (i.e. whether simply marking it is OK, as it will be
3913 visited by a task in the future, or whether it needs to be also
3914 pushed on a stack).
3915
3916 (2) Local Queue. The local queue of the task which is accessed
3917 reasonably efficiently by the task. Other tasks can steal from
3918 it when they run out of work. Throughout the marking phase, a
3919 task attempts to keep its local queue short but not totally
3920 empty, so that entries are available for stealing by other
3921 tasks. Only when there is no more work, a task will totally
3922 drain its local queue.
3923
3924 (3) Global Mark Stack. This handles local queue overflow. During
3925 marking only sets of entries are moved between it and the local
3926 queues, as access to it requires a mutex and more fine-grain
3927 interaction with it which might cause contention. If it
3928 overflows, then the marking phase should restart and iterate
3929 over the bitmap to identify gray objects. Throughout the marking
3930 phase, tasks attempt to keep the global mark stack at a small
3931 length but not totally empty, so that entries are available for
3932 popping by other tasks. Only when there is no more work, tasks
3933 will totally drain the global mark stack.
3934
3935 (4) SATB Buffer Queue. This is where completed SATB buffers are
3936 made available. Buffers are regularly removed from this queue
3937 and scanned for roots, so that the queue doesn't get too
3938 long. During remark, all completed buffers are processed, as
3939 well as the filled in parts of any uncompleted buffers.
3940
3941 The do_marking_step() method tries to abort when the time target
3942 has been reached. There are a few other cases when the
3943 do_marking_step() method also aborts:
3944
3945 (1) When the marking phase has been aborted (after a Full GC).
3946
3947 (2) When a global overflow (on the global stack) has been
3948 triggered. Before the task aborts, it will actually sync up with
3949 the other tasks to ensure that all the marking data structures
3950 (local queues, stacks, fingers etc.) are re-initialised so that
3951 when do_marking_step() completes, the marking phase can
3952 immediately restart.
3953
3954 (3) When enough completed SATB buffers are available. The
3955 do_marking_step() method only tries to drain SATB buffers right
3956 at the beginning. So, if enough buffers are available, the
3957 marking step aborts and the SATB buffers are processed at
3958 the beginning of the next invocation.
3959
3960 (4) To yield. when we have to yield then we abort and yield
3961 right at the end of do_marking_step(). This saves us from a lot
3962 of hassle as, by yielding we might allow a Full GC. If this
3963 happens then objects will be compacted underneath our feet, the
3964 heap might shrink, etc. We save checking for this by just
3965 aborting and doing the yield right at the end.
3966
3967 From the above it follows that the do_marking_step() method should
3968 be called in a loop (or, otherwise, regularly) until it completes.
3969
3970 If a marking step completes without its has_aborted() flag being
3971 true, it means it has completed the current marking phase (and
3972 also all other marking tasks have done so and have all synced up).
3973
3974 A method called regular_clock_call() is invoked "regularly" (in
3975 sub ms intervals) throughout marking. It is this clock method that
3976 checks all the abort conditions which were mentioned above and
3977 decides when the task should abort. A work-based scheme is used to
3978 trigger this clock method: when the number of object words the
3979 marking phase has scanned or the number of references the marking
3980 phase has visited reach a given limit. Additional invocations to
3981 the method clock have been planted in a few other strategic places
3982 too. The initial reason for the clock method was to avoid calling
3983 vtime too regularly, as it is quite expensive. So, once it was in
3984 place, it was natural to piggy-back all the other conditions on it
3985 too and not constantly check them throughout the code.
3986
3987 If do_stealing is true then do_marking_step will attempt to steal
3988 work from the other CMTasks. It only makes sense to enable
3989 stealing when being called by multiple threads.
3990
3991 If do_termination is true then do_marking_step will enter its
3992 termination protocol.
3993
3994 The value of is_serial must be true when do_marking_step is being
3995 called serially (i.e. by the VMThread) and do_marking_step should
3996 skip any synchronization in the termination and overflow code.
3997 Examples include the serial remark code and the serial reference
3998 processing closures.
3999
4000 The value of is_serial must be false when do_marking_step is
4001 being called by any of the worker threads in a work gang.
4002 Examples include the concurrent marking code (CMMarkingTask),
4003 the MT remark code, and the MT reference processing closures.
4004
4005 *****************************************************************************/
4006
4007 void CMTask::do_marking_step(double time_target_ms,
4008 bool do_stealing,
4009 bool do_termination,
4010 bool is_serial) {
4011 assert(time_target_ms >= 1.0, "minimum granularity is 1ms");
4012 assert(concurrent() == _cm->concurrent(), "they should be the same");
4013
4014 G1CollectorPolicy* g1_policy = _g1h->g1_policy();
4015 assert(_task_queues != NULL, "invariant");
4016 assert(_task_queue != NULL, "invariant");
4017 assert(_task_queues->queue(_worker_id) == _task_queue, "invariant");
4018
4019 assert(!_claimed,
4020 "only one thread should claim this task at any one time");
4021
4022 // OK, this doesn't safeguard again all possible scenarios, as it is
4023 // possible for two threads to set the _claimed flag at the same
4024 // time. But it is only for debugging purposes anyway and it will
4025 // catch most problems.
4026 _claimed = true;
4027
4028 _start_time_ms = os::elapsedVTime() * 1000.0;
4029 statsOnly( _interval_start_time_ms = _start_time_ms );
4030
4031 double diff_prediction_ms =
4032 g1_policy->get_new_prediction(&_marking_step_diffs_ms);
4033 _time_target_ms = time_target_ms - diff_prediction_ms;
4034
4035 // set up the variables that are used in the work-based scheme to
4036 // call the regular clock method
4037 _words_scanned = 0;
4038 _refs_reached = 0;
4039 recalculate_limits();
4040
4041 // clear all flags
4042 clear_has_aborted();
4043 _has_timed_out = false;
4044 _draining_satb_buffers = false;
4045
4046 ++_calls;
4047
4048 if (_cm->verbose_low()) {
4049 gclog_or_tty->print_cr("[%u] >>>>>>>>>> START, call = %d, "
4050 "target = %1.2lfms >>>>>>>>>>",
4051 _worker_id, _calls, _time_target_ms);
4052 }
4053
4054 // Set up the bitmap and oop closures. Anything that uses them is
4055 // eventually called from this method, so it is OK to allocate these
4056 // statically.
4057 CMBitMapClosure bitmap_closure(this, _cm, _nextMarkBitMap);
4058 G1CMOopClosure cm_oop_closure(_g1h, _cm, this);
4059 set_cm_oop_closure(&cm_oop_closure);
4060
4061 if (_cm->has_overflown()) {
4062 // This can happen if the mark stack overflows during a GC pause
4063 // and this task, after a yield point, restarts. We have to abort
4064 // as we need to get into the overflow protocol which happens
4065 // right at the end of this task.
4066 set_has_aborted();
4067 }
4068
4069 // First drain any available SATB buffers. After this, we will not
4070 // look at SATB buffers before the next invocation of this method.
4071 // If enough completed SATB buffers are queued up, the regular clock
4072 // will abort this task so that it restarts.
4073 drain_satb_buffers();
4074 // ...then partially drain the local queue and the global stack
4075 drain_local_queue(true);
4076 drain_global_stack(true);
4077
4078 do {
4079 if (!has_aborted() && _curr_region != NULL) {
4080 // This means that we're already holding on to a region.
4081 assert(_finger != NULL, "if region is not NULL, then the finger "
4082 "should not be NULL either");
4083
4084 // We might have restarted this task after an evacuation pause
4085 // which might have evacuated the region we're holding on to
4086 // underneath our feet. Let's read its limit again to make sure
4087 // that we do not iterate over a region of the heap that
4088 // contains garbage (update_region_limit() will also move
4089 // _finger to the start of the region if it is found empty).
4090 update_region_limit();
4091 // We will start from _finger not from the start of the region,
4092 // as we might be restarting this task after aborting half-way
4093 // through scanning this region. In this case, _finger points to
4094 // the address where we last found a marked object. If this is a
4095 // fresh region, _finger points to start().
4096 MemRegion mr = MemRegion(_finger, _region_limit);
4097
4098 if (_cm->verbose_low()) {
4099 gclog_or_tty->print_cr("[%u] we're scanning part "
4100 "["PTR_FORMAT", "PTR_FORMAT") "
4101 "of region "HR_FORMAT,
4102 _worker_id, _finger, _region_limit,
4103 HR_FORMAT_PARAMS(_curr_region));
4104 }
4105
4106 assert(!_curr_region->isHumongous() || mr.start() == _curr_region->bottom(),
4107 "humongous regions should go around loop once only");
4108
4109 // Some special cases:
4110 // If the memory region is empty, we can just give up the region.
4111 // If the current region is humongous then we only need to check
4112 // the bitmap for the bit associated with the start of the object,
4113 // scan the object if it's live, and give up the region.
4114 // Otherwise, let's iterate over the bitmap of the part of the region
4115 // that is left.
4116 // If the iteration is successful, give up the region.
4117 if (mr.is_empty()) {
4118 giveup_current_region();
4119 regular_clock_call();
4120 } else if (_curr_region->isHumongous() && mr.start() == _curr_region->bottom()) {
4121 if (_nextMarkBitMap->isMarked(mr.start())) {
4122 // The object is marked - apply the closure
4123 BitMap::idx_t offset = _nextMarkBitMap->heapWordToOffset(mr.start());
4124 bitmap_closure.do_bit(offset);
4125 }
4126 // Even if this task aborted while scanning the humongous object
4127 // we can (and should) give up the current region.
4128 giveup_current_region();
4129 regular_clock_call();
4130 } else if (_nextMarkBitMap->iterate(&bitmap_closure, mr)) {
4131 giveup_current_region();
4132 regular_clock_call();
4133 } else {
4134 assert(has_aborted(), "currently the only way to do so");
4135 // The only way to abort the bitmap iteration is to return
4136 // false from the do_bit() method. However, inside the
4137 // do_bit() method we move the _finger to point to the
4138 // object currently being looked at. So, if we bail out, we
4139 // have definitely set _finger to something non-null.
4140 assert(_finger != NULL, "invariant");
4141
4142 // Region iteration was actually aborted. So now _finger
4143 // points to the address of the object we last scanned. If we
4144 // leave it there, when we restart this task, we will rescan
4145 // the object. It is easy to avoid this. We move the finger by
4146 // enough to point to the next possible object header (the
4147 // bitmap knows by how much we need to move it as it knows its
4148 // granularity).
4149 assert(_finger < _region_limit, "invariant");
4150 HeapWord* new_finger = _nextMarkBitMap->nextObject(_finger);
4151 // Check if bitmap iteration was aborted while scanning the last object
4152 if (new_finger >= _region_limit) {
4153 giveup_current_region();
4154 } else {
4155 move_finger_to(new_finger);
4156 }
4157 }
4158 }
4159 // At this point we have either completed iterating over the
4160 // region we were holding on to, or we have aborted.
4161
4162 // We then partially drain the local queue and the global stack.
4163 // (Do we really need this?)
4164 drain_local_queue(true);
4165 drain_global_stack(true);
4166
4167 // Read the note on the claim_region() method on why it might
4168 // return NULL with potentially more regions available for
4169 // claiming and why we have to check out_of_regions() to determine
4170 // whether we're done or not.
4171 while (!has_aborted() && _curr_region == NULL && !_cm->out_of_regions()) {
4172 // We are going to try to claim a new region. We should have
4173 // given up on the previous one.
4174 // Separated the asserts so that we know which one fires.
4175 assert(_curr_region == NULL, "invariant");
4176 assert(_finger == NULL, "invariant");
4177 assert(_region_limit == NULL, "invariant");
4178 if (_cm->verbose_low()) {
4179 gclog_or_tty->print_cr("[%u] trying to claim a new region", _worker_id);
4180 }
4181 HeapRegion* claimed_region = _cm->claim_region(_worker_id);
4182 if (claimed_region != NULL) {
4183 // Yes, we managed to claim one
4184 statsOnly( ++_regions_claimed );
4185
4186 if (_cm->verbose_low()) {
4187 gclog_or_tty->print_cr("[%u] we successfully claimed "
4188 "region "PTR_FORMAT,
4189 _worker_id, claimed_region);
4190 }
4191
4192 setup_for_region(claimed_region);
4193 assert(_curr_region == claimed_region, "invariant");
4194 }
4195 // It is important to call the regular clock here. It might take
4196 // a while to claim a region if, for example, we hit a large
4197 // block of empty regions. So we need to call the regular clock
4198 // method once round the loop to make sure it's called
4199 // frequently enough.
4200 regular_clock_call();
4201 }
4202
4203 if (!has_aborted() && _curr_region == NULL) {
4204 assert(_cm->out_of_regions(),
4205 "at this point we should be out of regions");
4206 }
4207 } while ( _curr_region != NULL && !has_aborted());
4208
4209 if (!has_aborted()) {
4210 // We cannot check whether the global stack is empty, since other
4211 // tasks might be pushing objects to it concurrently.
4212 assert(_cm->out_of_regions(),
4213 "at this point we should be out of regions");
4214
4215 if (_cm->verbose_low()) {
4216 gclog_or_tty->print_cr("[%u] all regions claimed", _worker_id);
4217 }
4218
4219 // Try to reduce the number of available SATB buffers so that
4220 // remark has less work to do.
4221 drain_satb_buffers();
4222 }
4223
4224 // Since we've done everything else, we can now totally drain the
4225 // local queue and global stack.
4226 drain_local_queue(false);
4227 drain_global_stack(false);
4228
4229 // Attempt at work stealing from other task's queues.
4230 if (do_stealing && !has_aborted()) {
4231 // We have not aborted. This means that we have finished all that
4232 // we could. Let's try to do some stealing...
4233
4234 // We cannot check whether the global stack is empty, since other
4235 // tasks might be pushing objects to it concurrently.
4236 assert(_cm->out_of_regions() && _task_queue->size() == 0,
4237 "only way to reach here");
4238
4239 if (_cm->verbose_low()) {
4240 gclog_or_tty->print_cr("[%u] starting to steal", _worker_id);
4241 }
4242
4243 while (!has_aborted()) {
4244 oop obj;
4245 statsOnly( ++_steal_attempts );
4246
4247 if (_cm->try_stealing(_worker_id, &_hash_seed, obj)) {
4248 if (_cm->verbose_medium()) {
4249 gclog_or_tty->print_cr("[%u] stolen "PTR_FORMAT" successfully",
4250 _worker_id, (void*) obj);
4251 }
4252
4253 statsOnly( ++_steals );
4254
4255 assert(_nextMarkBitMap->isMarked((HeapWord*) obj),
4256 "any stolen object should be marked");
4257 scan_object(obj);
4258
4259 // And since we're towards the end, let's totally drain the
4260 // local queue and global stack.
4261 drain_local_queue(false);
4262 drain_global_stack(false);
4263 } else {
4264 break;
4265 }
4266 }
4267 }
4268
4269 // If we are about to wrap up and go into termination, check if we
4270 // should raise the overflow flag.
4271 if (do_termination && !has_aborted()) {
4272 if (_cm->force_overflow()->should_force()) {
4273 _cm->set_has_overflown();
4274 regular_clock_call();
4275 }
4276 }
4277
4278 // We still haven't aborted. Now, let's try to get into the
4279 // termination protocol.
4280 if (do_termination && !has_aborted()) {
4281 // We cannot check whether the global stack is empty, since other
4282 // tasks might be concurrently pushing objects on it.
4283 // Separated the asserts so that we know which one fires.
4284 assert(_cm->out_of_regions(), "only way to reach here");
4285 assert(_task_queue->size() == 0, "only way to reach here");
4286
4287 if (_cm->verbose_low()) {
4288 gclog_or_tty->print_cr("[%u] starting termination protocol", _worker_id);
4289 }
4290
4291 _termination_start_time_ms = os::elapsedVTime() * 1000.0;
4292
4293 // The CMTask class also extends the TerminatorTerminator class,
4294 // hence its should_exit_termination() method will also decide
4295 // whether to exit the termination protocol or not.
4296 bool finished = (is_serial ||
4297 _cm->terminator()->offer_termination(this));
4298 double termination_end_time_ms = os::elapsedVTime() * 1000.0;
4299 _termination_time_ms +=
4300 termination_end_time_ms - _termination_start_time_ms;
4301
4302 if (finished) {
4303 // We're all done.
4304
4305 if (_worker_id == 0) {
4306 // let's allow task 0 to do this
4307 if (concurrent()) {
4308 assert(_cm->concurrent_marking_in_progress(), "invariant");
4309 // we need to set this to false before the next
4310 // safepoint. This way we ensure that the marking phase
4311 // doesn't observe any more heap expansions.
4312 _cm->clear_concurrent_marking_in_progress();
4313 }
4314 }
4315
4316 // We can now guarantee that the global stack is empty, since
4317 // all other tasks have finished. We separated the guarantees so
4318 // that, if a condition is false, we can immediately find out
4319 // which one.
4320 guarantee(_cm->out_of_regions(), "only way to reach here");
4321 guarantee(_cm->mark_stack_empty(), "only way to reach here");
4322 guarantee(_task_queue->size() == 0, "only way to reach here");
4323 guarantee(!_cm->has_overflown(), "only way to reach here");
4324 guarantee(!_cm->mark_stack_overflow(), "only way to reach here");
4325
4326 if (_cm->verbose_low()) {
4327 gclog_or_tty->print_cr("[%u] all tasks terminated", _worker_id);
4328 }
4329 } else {
4330 // Apparently there's more work to do. Let's abort this task. It
4331 // will restart it and we can hopefully find more things to do.
4332
4333 if (_cm->verbose_low()) {
4334 gclog_or_tty->print_cr("[%u] apparently there is more work to do",
4335 _worker_id);
4336 }
4337
4338 set_has_aborted();
4339 statsOnly( ++_aborted_termination );
4340 }
4341 }
4342
4343 // Mainly for debugging purposes to make sure that a pointer to the
4344 // closure which was statically allocated in this frame doesn't
4345 // escape it by accident.
4346 set_cm_oop_closure(NULL);
4347 double end_time_ms = os::elapsedVTime() * 1000.0;
4348 double elapsed_time_ms = end_time_ms - _start_time_ms;
4349 // Update the step history.
4350 _step_times_ms.add(elapsed_time_ms);
4351
4352 if (has_aborted()) {
4353 // The task was aborted for some reason.
4354
4355 statsOnly( ++_aborted );
4356
4357 if (_has_timed_out) {
4358 double diff_ms = elapsed_time_ms - _time_target_ms;
4359 // Keep statistics of how well we did with respect to hitting
4360 // our target only if we actually timed out (if we aborted for
4361 // other reasons, then the results might get skewed).
4362 _marking_step_diffs_ms.add(diff_ms);
4363 }
4364
4365 if (_cm->has_overflown()) {
4366 // This is the interesting one. We aborted because a global
4367 // overflow was raised. This means we have to restart the
4368 // marking phase and start iterating over regions. However, in
4369 // order to do this we have to make sure that all tasks stop
4370 // what they are doing and re-initialise in a safe manner. We
4371 // will achieve this with the use of two barrier sync points.
4372
4373 if (_cm->verbose_low()) {
4374 gclog_or_tty->print_cr("[%u] detected overflow", _worker_id);
4375 }
4376
4377 if (!is_serial) {
4378 // We only need to enter the sync barrier if being called
4379 // from a parallel context
4380 _cm->enter_first_sync_barrier(_worker_id);
4381
4382 // When we exit this sync barrier we know that all tasks have
4383 // stopped doing marking work. So, it's now safe to
4384 // re-initialise our data structures. At the end of this method,
4385 // task 0 will clear the global data structures.
4386 }
4387
4388 statsOnly( ++_aborted_overflow );
4389
4390 // We clear the local state of this task...
4391 clear_region_fields();
4392
4393 if (!is_serial) {
4394 // ...and enter the second barrier.
4395 _cm->enter_second_sync_barrier(_worker_id);
4396 }
4397 // At this point everything has bee re-initialised and we're
4398 // ready to restart.
4399 }
4400
4401 if (_cm->verbose_low()) {
4402 gclog_or_tty->print_cr("[%u] <<<<<<<<<< ABORTING, target = %1.2lfms, "
4403 "elapsed = %1.2lfms <<<<<<<<<<",
4404 _worker_id, _time_target_ms, elapsed_time_ms);
4405 if (_cm->has_aborted()) {
4406 gclog_or_tty->print_cr("[%u] ========== MARKING ABORTED ==========",
4407 _worker_id);
4408 }
4409 }
4410 } else {
4411 if (_cm->verbose_low()) {
4412 gclog_or_tty->print_cr("[%u] <<<<<<<<<< FINISHED, target = %1.2lfms, "
4413 "elapsed = %1.2lfms <<<<<<<<<<",
4414 _worker_id, _time_target_ms, elapsed_time_ms);
4415 }
4416 }
4417
4418 _claimed = false;
4419 }
4420
4421 CMTask::CMTask(uint worker_id,
4422 ConcurrentMark* cm,
4423 size_t* marked_bytes,
4424 BitMap* card_bm,
4425 CMTaskQueue* task_queue,
4426 CMTaskQueueSet* task_queues)
4427 : _g1h(G1CollectedHeap::heap()),
4428 _worker_id(worker_id), _cm(cm),
4429 _claimed(false),
4430 _nextMarkBitMap(NULL), _hash_seed(17),
4431 _task_queue(task_queue),
4432 _task_queues(task_queues),
4433 _cm_oop_closure(NULL),
4434 _marked_bytes_array(marked_bytes),
4435 _card_bm(card_bm) {
4436 guarantee(task_queue != NULL, "invariant");
4437 guarantee(task_queues != NULL, "invariant");
4438
4439 statsOnly( _clock_due_to_scanning = 0;
4440 _clock_due_to_marking = 0 );
4441
4442 _marking_step_diffs_ms.add(0.5);
4443 }
4444
4445 // These are formatting macros that are used below to ensure
4446 // consistent formatting. The *_H_* versions are used to format the
4447 // header for a particular value and they should be kept consistent
4448 // with the corresponding macro. Also note that most of the macros add
4449 // the necessary white space (as a prefix) which makes them a bit
4450 // easier to compose.
4451
4452 // All the output lines are prefixed with this string to be able to
4453 // identify them easily in a large log file.
4454 #define G1PPRL_LINE_PREFIX "###"
4455
4456 #define G1PPRL_ADDR_BASE_FORMAT " "PTR_FORMAT"-"PTR_FORMAT
4457 #ifdef _LP64
4458 #define G1PPRL_ADDR_BASE_H_FORMAT " %37s"
4459 #else // _LP64
4460 #define G1PPRL_ADDR_BASE_H_FORMAT " %21s"
4461 #endif // _LP64
4462
4463 // For per-region info
4464 #define G1PPRL_TYPE_FORMAT " %-4s"
4465 #define G1PPRL_TYPE_H_FORMAT " %4s"
4466 #define G1PPRL_BYTE_FORMAT " "SIZE_FORMAT_W(9)
4467 #define G1PPRL_BYTE_H_FORMAT " %9s"
4468 #define G1PPRL_DOUBLE_FORMAT " %14.1f"
4469 #define G1PPRL_DOUBLE_H_FORMAT " %14s"
4470
4471 // For summary info
4472 #define G1PPRL_SUM_ADDR_FORMAT(tag) " "tag":"G1PPRL_ADDR_BASE_FORMAT
4473 #define G1PPRL_SUM_BYTE_FORMAT(tag) " "tag": "SIZE_FORMAT
4474 #define G1PPRL_SUM_MB_FORMAT(tag) " "tag": %1.2f MB"
4475 #define G1PPRL_SUM_MB_PERC_FORMAT(tag) G1PPRL_SUM_MB_FORMAT(tag)" / %1.2f %%"
4476
4477 G1PrintRegionLivenessInfoClosure::
4478 G1PrintRegionLivenessInfoClosure(outputStream* out, const char* phase_name)
4479 : _out(out),
4480 _total_used_bytes(0), _total_capacity_bytes(0),
4481 _total_prev_live_bytes(0), _total_next_live_bytes(0),
4482 _hum_used_bytes(0), _hum_capacity_bytes(0),
4483 _hum_prev_live_bytes(0), _hum_next_live_bytes(0) {
4484 G1CollectedHeap* g1h = G1CollectedHeap::heap();
4485 MemRegion g1_committed = g1h->g1_committed();
4486 MemRegion g1_reserved = g1h->g1_reserved();
4487 double now = os::elapsedTime();
4488
4489 // Print the header of the output.
4490 _out->cr();
4491 _out->print_cr(G1PPRL_LINE_PREFIX" PHASE %s @ %1.3f", phase_name, now);
4492 _out->print_cr(G1PPRL_LINE_PREFIX" HEAP"
4493 G1PPRL_SUM_ADDR_FORMAT("committed")
4494 G1PPRL_SUM_ADDR_FORMAT("reserved")
4495 G1PPRL_SUM_BYTE_FORMAT("region-size"),
4496 g1_committed.start(), g1_committed.end(),
4497 g1_reserved.start(), g1_reserved.end(),
4498 HeapRegion::GrainBytes);
4499 _out->print_cr(G1PPRL_LINE_PREFIX);
4500 _out->print_cr(G1PPRL_LINE_PREFIX
4501 G1PPRL_TYPE_H_FORMAT
4502 G1PPRL_ADDR_BASE_H_FORMAT
4503 G1PPRL_BYTE_H_FORMAT
4504 G1PPRL_BYTE_H_FORMAT
4505 G1PPRL_BYTE_H_FORMAT
4506 G1PPRL_DOUBLE_H_FORMAT,
4507 "type", "address-range",
4508 "used", "prev-live", "next-live", "gc-eff");
4509 _out->print_cr(G1PPRL_LINE_PREFIX
4510 G1PPRL_TYPE_H_FORMAT
4511 G1PPRL_ADDR_BASE_H_FORMAT
4512 G1PPRL_BYTE_H_FORMAT
4513 G1PPRL_BYTE_H_FORMAT
4514 G1PPRL_BYTE_H_FORMAT
4515 G1PPRL_DOUBLE_H_FORMAT,
4516 "", "",
4517 "(bytes)", "(bytes)", "(bytes)", "(bytes/ms)");
4518 }
4519
4520 // It takes as a parameter a reference to one of the _hum_* fields, it
4521 // deduces the corresponding value for a region in a humongous region
4522 // series (either the region size, or what's left if the _hum_* field
4523 // is < the region size), and updates the _hum_* field accordingly.
4524 size_t G1PrintRegionLivenessInfoClosure::get_hum_bytes(size_t* hum_bytes) {
4525 size_t bytes = 0;
4526 // The > 0 check is to deal with the prev and next live bytes which
4527 // could be 0.
4528 if (*hum_bytes > 0) {
4529 bytes = MIN2(HeapRegion::GrainBytes, *hum_bytes);
4530 *hum_bytes -= bytes;
4531 }
4532 return bytes;
4533 }
4534
4535 // It deduces the values for a region in a humongous region series
4536 // from the _hum_* fields and updates those accordingly. It assumes
4537 // that that _hum_* fields have already been set up from the "starts
4538 // humongous" region and we visit the regions in address order.
4539 void G1PrintRegionLivenessInfoClosure::get_hum_bytes(size_t* used_bytes,
4540 size_t* capacity_bytes,
4541 size_t* prev_live_bytes,
4542 size_t* next_live_bytes) {
4543 assert(_hum_used_bytes > 0 && _hum_capacity_bytes > 0, "pre-condition");
4544 *used_bytes = get_hum_bytes(&_hum_used_bytes);
4545 *capacity_bytes = get_hum_bytes(&_hum_capacity_bytes);
4546 *prev_live_bytes = get_hum_bytes(&_hum_prev_live_bytes);
4547 *next_live_bytes = get_hum_bytes(&_hum_next_live_bytes);
4548 }
4549
4550 bool G1PrintRegionLivenessInfoClosure::doHeapRegion(HeapRegion* r) {
4551 const char* type = "";
4552 HeapWord* bottom = r->bottom();
4553 HeapWord* end = r->end();
4554 size_t capacity_bytes = r->capacity();
4555 size_t used_bytes = r->used();
4556 size_t prev_live_bytes = r->live_bytes();
4557 size_t next_live_bytes = r->next_live_bytes();
4558 double gc_eff = r->gc_efficiency();
4559 if (r->used() == 0) {
4560 type = "FREE";
4561 } else if (r->is_survivor()) {
4562 type = "SURV";
4563 } else if (r->is_young()) {
4564 type = "EDEN";
4565 } else if (r->startsHumongous()) {
4566 type = "HUMS";
4567
4568 assert(_hum_used_bytes == 0 && _hum_capacity_bytes == 0 &&
4569 _hum_prev_live_bytes == 0 && _hum_next_live_bytes == 0,
4570 "they should have been zeroed after the last time we used them");
4571 // Set up the _hum_* fields.
4572 _hum_capacity_bytes = capacity_bytes;
4573 _hum_used_bytes = used_bytes;
4574 _hum_prev_live_bytes = prev_live_bytes;
4575 _hum_next_live_bytes = next_live_bytes;
4576 get_hum_bytes(&used_bytes, &capacity_bytes,
4577 &prev_live_bytes, &next_live_bytes);
4578 end = bottom + HeapRegion::GrainWords;
4579 } else if (r->continuesHumongous()) {
4580 type = "HUMC";
4581 get_hum_bytes(&used_bytes, &capacity_bytes,
4582 &prev_live_bytes, &next_live_bytes);
4583 assert(end == bottom + HeapRegion::GrainWords, "invariant");
4584 } else {
4585 type = "OLD";
4586 }
4587
4588 _total_used_bytes += used_bytes;
4589 _total_capacity_bytes += capacity_bytes;
4590 _total_prev_live_bytes += prev_live_bytes;
4591 _total_next_live_bytes += next_live_bytes;
4592
4593 // Print a line for this particular region.
4594 _out->print_cr(G1PPRL_LINE_PREFIX
4595 G1PPRL_TYPE_FORMAT
4596 G1PPRL_ADDR_BASE_FORMAT
4597 G1PPRL_BYTE_FORMAT
4598 G1PPRL_BYTE_FORMAT
4599 G1PPRL_BYTE_FORMAT
4600 G1PPRL_DOUBLE_FORMAT,
4601 type, bottom, end,
4602 used_bytes, prev_live_bytes, next_live_bytes, gc_eff);
4603
4604 return false;
4605 }
4606
4607 G1PrintRegionLivenessInfoClosure::~G1PrintRegionLivenessInfoClosure() {
4608 // Print the footer of the output.
4609 _out->print_cr(G1PPRL_LINE_PREFIX);
4610 _out->print_cr(G1PPRL_LINE_PREFIX
4611 " SUMMARY"
4612 G1PPRL_SUM_MB_FORMAT("capacity")
4613 G1PPRL_SUM_MB_PERC_FORMAT("used")
4614 G1PPRL_SUM_MB_PERC_FORMAT("prev-live")
4615 G1PPRL_SUM_MB_PERC_FORMAT("next-live"),
4616 bytes_to_mb(_total_capacity_bytes),
4617 bytes_to_mb(_total_used_bytes),
4618 perc(_total_used_bytes, _total_capacity_bytes),
4619 bytes_to_mb(_total_prev_live_bytes),
4620 perc(_total_prev_live_bytes, _total_capacity_bytes),
4621 bytes_to_mb(_total_next_live_bytes),
4622 perc(_total_next_live_bytes, _total_capacity_bytes));
4623 _out->cr();
4624 }